Name ///sr-nova/Dclabs_wip/High Temp/5208_1.

pdf/Chap_01/ 26/10/2007 1:51PM Plate # 0 pg 1

High-Temperature Corrosion And Materials Applications Copyright © 2007 ASM International®

George Y. Lai, editor, p1 All rights reserved.
DOI: 10.1361/hcma2007p001 www.asminternational.org

CHAPTER 1

Introduction
METALS AND ALLOYS will react during cracking,” “stress-relaxation cracking,” or
high-temperature service with the surrounding “strain-age cracking” (for nickel-base alloys).
environment, resulting in high-temperature Both of these subjects are covered in Chapter 14,
corrosion. In gaseous environments, high-tem- “Stress-Assisted Corrosion and Cracking.” The
perature corrosion is defined as the corrosion that subject of erosion and erosion/corrosion is also
takes place above the maximum temperature reviewed with an attempt to offer readers general
at which acids condense and dew-point corro- guidance on materials selection and application.
sion takes place. Although a majority of high- Discussion also includes hydrogen attack of
temperature corrosion reactions take place at carbon steels in boilers and refinery equipment.
temperatures above 500 °C (930 °F), severe Finally, extensive discussion on the materials
high-temperature corrosion has been encoun- problems in coal-fired boilers, oil-fired boilers,
tered in many cases at temperatures below waste-to-energy boilers, and black liquor re-
500 °C (930 °F). In waste-to-energy boilers, for covery boilers is included. In summary, the sub-
example, carbon and low-alloy steels have jects covered extensively in this book include:
experienced severe fireside corrosion problems
 Oxidation
in the waterwalls of the boilers at the tube metal
 Nitridation
temperatures of approximately 260 to 315 °C
 Carburization and metal dusting
(500 to 600 °F).
 Corrosion by halogen and hydrogen halides
This book is intended primarily for engineers
 Sulfidation
and metallurgists who are concerned with high-
 Hot corrosion
temperature materials problems in the following
 Molten salt corrosion
industries: aerospace/gas turbine, chemical pro-
 Liquid metal corrosion and embrittlement
cessing, refining and petrochemical, fossil-fired
 Erosion and erosion/corrosion
power generation, coal gasification, waste-to-
 Stress-assisted corrosion and cracking
energy industry, pulp and paper, heat treating,
 Hydrogen attack
mineral and metallurgical processing, and
 Coal-fired boilers
others. The technical data presented in this book
 Oil-fired boilers and furnaces
are pertinent to “real” materials problems related
 Waste-to-energy boilers and waste inci-
to the aforementioned industries. The book will
nerators
also be useful for both undergraduate and grad-
 Black-liquor recovery boilers
uate students who are interested in studying or
pursuing research on the subject of high- The focus of this book is on commercial
temperature corrosion. alloys, including both generic and proprietary
The book covers eight basic modes of high- alloys. Most data are presented to reveal alloy
temperature corrosion. A brief description of ranking and thus serve as a general guide to
thermodynamics is included for most chapters to materials selection and application. Engineers
help readers to understand the corrosion reac- can thus use the data and information to com-
tions. The external stresses (or strains) can cause pare alloys that are commercially available.
alloys to suffer preferential corrosion penetration The effects of alloying elements, temperature,
attack in a certain corrosive environment, such as and environmental conditions on the corrosion
sulfidizing environments. In addition, external behavior of alloys are also discussed, providing
stresses or residual stresses can cause the alloy information about the capability of an alloy in
to suffer brittle, intergranular cracking when terms of useful temperature limitation. Trade-
exposed to the lower end of the intermediate marks for alloys and alloy manufacturers are
temperatures for certain alloys. This type of listed in Appendix 1. The compositions of alloys
cracking is frequently referred to as “reheat are tabulated in Appendix 2.
Name ///sr-nova/Dclabs_wip/High Temp/5208_3-4.pdf/Chap_02/ 26/10/2007 12:10PM Plate # 0 pg 3

High-Temperature Corrosion And Materials Applications Copyright © 2007 ASM International®

George Y. Lai, editor, p3-4 All rights reserved.
DOI: 10.1361/hcma2007p003 www.asminternational.org

CHAPTER 2

Challenges in Materials Applications

for High-Temperature Service

IN MANY INDUSTRIAL SYSTEMS, plant analysis needs to be conducted to balance the

operating conditions can be quite complex; it is material cost with the expected life for the
rather difficult to use laboratory tests to simulate component to ensure the alloy is cost effective.
plant conditions. However, laboratory tests can Often the life-cycle cost is a better criterion than
provide good general guidance for making pre- the initial material cost in making an alloy sel-
liminary alloy selection. In situ field testing or ection.
field trials of candidate alloys in the operating Selection of an appropriate filler metal for
plant provides the best way of obtaining the welding is important for component fabrication
corrosion information that can be reliably used involving welding. Normally, it is a simple pro-
for final materials selection. cess when the candidate alloy has a filler metal
During the preliminary alloy selection pro- with matching chemical composition. However,
cess, it is important to evaluate not only the many high-temperature alloys do not have filler
high-temperature corrosion resistance of the metals with matching chemistries. The widely
alloy, but also its mechanical properties such as used Fe-Ni-Cr alloy 800H is a good example.
tensile and creep-rupture strengths. The micro- Many heat-resistant cast alloys also do not have
structural changes at the application tempera- matching filler metals. Thus, when no matching
tures such as thermal stability of the alloy, should chemistry filler metal is available for welding, it
also be considered. For example, duplex stainless is critical to select a filler metal that not only
steels are known to suffer 475 °C (885 °F) possesses excellent weldability but, also exhibits
embrittlement caused by the formation of alpha comparable or better high-temperature corrosion
prime (α0 ) coherent precipitates. Accordingly, resistance along with comparable strengths,
these stainless steels should be avoided for thermal stability, and other relevant properties.
use as a structural component at temperatures Some fabricators sometimes use weldability to
approximately, above 340 °C (650 °F). ASME select a filler metal without considering the
Codes may have lower maximum service tem- resistance of the weld metal to the specific high-
perature limits for these alloys. Consideration temperature corrosive environment in the end
should also be given to fabrication issues, such application. This can lead to premature failures.
as weldability and welding procedures, anneal- For example, because of their good weldability,
ing heat treatments, postweld heat treatment high nickel filler metals, such as filler metal alloy
(PWHT) and stress relieving, and codes and 82 (ERNiCr-3), are sometimes used for welding
standards requirements. the alloys that are to be in service in sulfidizing
The availability of the alloy can also be an environments. This can cause preferential sulfi-
issue. It is not uncommon to find that some dation attack at the weld joint because of the
alloys are no longer commercially available relatively poor sulfidation resistance of high
in stock due to a number of reasons, which may nickel alloys.
include poor market demands in the past, diffi- Welding can still be an issue for some high-
culty in manufacturing, and so forth. In some temperature alloys even with matching filler
cases, the alloy may only be available on order metals. This is because some high-temperature
for a whole production heat, which can be tens alloys contain many alloying elements for var-
of thousands of pounds of material. Another ious metallurgical reasons, such as improving the
important factor is the alloy price. A cost resistance to a certain mode of high-temperature
Name ///sr-nova/Dclabs_wip/High Temp/5208_3-4.pdf/Chap_02/ 26/10/2007 12:10PM Plate # 0 pg 4

4 / High-Temperature Corrosion and Materials Applications

corrosion, increasing tensile and creep-rupture diameter have long been available for construc-
strengths, or increasing wear resistance. In- tion of waterwalls as well as superheaters in
creasing the levels of some of these alloying boilers. Composite tubes manufactured by a
elements can increase the difficulty in the weld- spiral weld overlaying process have been made
ability of the alloy. For example, an alloy con- available in recent years. These composite tubes
taining high silicon, high aluminum, high use the outer diameter cladding for providing
carbon, or very high chromium can be difficult to corrosion protection and the substrate base tube
weld even though a matching filler metal is for the load-bearing structural part. Most of these
available. composite tubes are used in superheaters and
For construction of a component, engineers reheaters in boilers with metal temperatures
have the option to consider whether a wrought being likely less than about 650 °C (1200 °F).
alloy or a cast alloy will be more suitable metal- Many furnace tubes used in petrochemical pro-
lurgically and/or economically for the intended cessing, such as ethylene cracking furnace tubes,
high-temperature application. Engineers may are exposed to temperatures higher than 980 °C
also consider a totally different approach to (1800 °F) and carburizing gas streams on the
address the high-temperature corrosion issue internal diameter (ID) of the tube, application
for some existing plant equipment that has of composite tubes with a carburization- and
suffered corrosion. In refineries, many reactor coking-resistant alloy cladding on the tube ID
vessels, such as crude towers, hydrocrackers, can potentially increase the operating tempera-
and hydrodesulfurizers, are made of clad plates ture and/or prolong the tube life.
with a corrosion-resistant cladding in original Aluminide coatings reportedly have been used
installations. Cladding can be corroded after in ethylene cracking furnace tubes. At the writing
years of operation. One common approach is to of this book, it appears no commercial companies
refurbish the corroded vessels by applying a in the United States provide aluminizing coating
corrosion-resistant weld overlay instead of re- services for ethylene furnace tubes or pipes.
placing it with a new construction. This Another diffusion coating, chromized coating,
approach has been adopted in the boiler industry has also reportedly been used in boilers. Both of
in recent years to address the severe corrosion these diffusion coatings are very thin. Coatings
problems with the waterwalls of boilers in have been highly successful in providing pro-
waste-to-energy boilers, coal-fired boilers, basic tection against oxidation and hot corrosion for
oxygen furnace hoods in steel mills, and so forth. the high-temperature components, such as air-
With automatic controls for gas metal arc weld- foils, in gas turbines. The coatings used involve
ing machines, a large scale of weld overlay can aluminide coatings, overlay MCrAlY coatings
be applied in vessels or boilers with engineering by vapor deposition processes (e.g., electron
quality. Laser cladding can also be applied in the beam physical vapor deposition), and ceramic
shop on large equipment such as waterwall thermal barrier coatings (e.g., stabilized ZrO2).
panels. Coextruded composite tubes with a Coatings are considered sacrificial and are to be
corrosion-resistant alloy cladding on the outer replaced periodically.
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:43PM Plate # 0 pg 5

High-Temperature Corrosion And Materials Applications Copyright © 2007 ASM International®

George Y. Lai, editor, p5-66 All rights reserved.
DOI: 10.1361/hcma2007p005 www.asminternational.org

CHAPTER 3

Oxidation

3.1 Introduction sluggish for most alloys. As a result, the effects of

corrosive contaminants can become more pro-
nounced, resulting in other modes of high-
Oxidation is the most important high- temperature corrosion. For example, if the sulfur
temperature corrosion reaction. Metals or alloys level in the environment is high, sulfidation then
are oxidized when heated to elevated tempera- becomes the predominant mode of corrosion,
tures in air or in highly oxidizing environments, even though oxidation also takes part in the
such as combustion atmospheres with excess air corrosion reaction. Thus, a majority of high-
or oxygen. Many metallic components, such as temperature corrosion problems in reducing
retorts in heat treat furnaces, furnace heater tubes environments are caused by modes of corrosion
and coils in chemical and petrochemical plants, attack other than oxidation. Most industrial
waterwalls and superheater/reheater tubes in environments have sufficient oxygen activities
boilers, and combustors and transition ducts in (or potentials) to allow oxidation to participate in
gas turbines, are subject to oxidation. the high-temperature corrosion reaction regard-
For many industrial processes, combustion less of the predominant mode of corrosion. In
involves relatively “clean” fuels such as natural fact, the alloy often relies on the oxidation reac-
gas or No. 1 or No. 2 fuel oil. These fuels gen- tion to develop a protective oxide scale to resist
erally have low concentrations of contaminants, corrosion attack, such as sulfidation, carburiza-
such as sulfur, chlorine, alkali metals, and tion, hot corrosion, and so forth. The oxidation
vanadium. In many cases, excess air is used to behavior under these conditions is discussed in
ensure complete combustion of the fuel. The other chapters dealing with different modes of
combustion products thus consist primarily of corrosion attack.
O2, N2, CO2, and H2O. Although alloys in these There is a large spectrum of engineering alloys
environments are oxidized by oxygen, other available for applications in different tempera-
combustion products, such as H2O, may play an ture ranges. This chapter presents a large oxida-
important role in affecting the oxidation behavior tion database for a wide spectrum of engineering
of the alloy. The presence of N2 in the combus- alloys, ranging from carbon and Cr-Mo steels,
tion gas stream can cause significant internal which serve the low end of the temperature
nitridation attack under certain conditions, which spectrum, to superalloys serving the highest
is discussed in Chapter 4 “Nitridation.” temperature regime. The data are organized by
Oxidation can also take place in a “reducing” alloy groups to help readers to compare alloys
environment (i.e., the environment with a low within the same alloy group and also to compare
oxygen potential created by the combustion alloys between different alloy groups. The focus
under a substoichiometric condition). When is to present comparison data, thus allowing
combustion takes place under stoichiometric or readers to consider candidate alloys for applica-
substoichiometric conditions, the resultant tions in the temperature regime of interest. Also
environment becomes “reducing.” This type of included are some important metallurgical and
environment is generally characterized by low environmental factors that can affect the oxida-
oxygen potentials. Under this condition, the tion behavior of the alloy.
oxygen potential of the environment is typically It should be noted that a majority of the data-
controlled by pH2 =pH2 O or pCO =pCO2 ratio, and base has been generated in laboratory tests that
the oxidation kinetic is generally slow. The were conducted at temperatures higher than those
development of a protective oxide scale can be at which the tested alloys would normally be
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:43PM Plate # 0 pg 6

6 / High-Temperature Corrosion and Materials Applications

used. The intent for this approach was to deter- 3.2 Thermodynamic Considerations
mine the oxidation behavior of alloys within
relatively short test durations. For example, 3.2.1 Formation of Oxides
many tests have been conducted at 980 to Thermodynamically, an oxide is likely to form
1200 °C (1800 to 2200 °F) for stainless steels, on a metal surface when the oxygen potential
Fe-Ni-Cr alloys, and Ni-Cr alloys. This may (pO2 ) in the environment is greater than the
result in the alloy performance ranking based oxygen partial pressure in equilibrium with the
more on scaling (metal loss) than on internal oxide. The oxygen partial pressure in equilibrium
oxidation attack which the alloys would most with the oxide can be determined from the stan-
likely have encountered at lower application dard free energy of formation of the oxide.
temperatures. An example is given in Fig. 3.1, Consider the reaction:
which illustrates an actual field experience with a
furnace heater coil made of a Ni-Cr alloy that had M+O2 ÐMO2 ð3:1Þ
been in service for about 4 to 5 years at tem-
peratures less than 900 °C (1650 °F), suffering


aMO2
extensive internal oxidation attack with very lit- DG =7RT ln ð3:2Þ
aM  pO2
tle scaling (metal loss) (Ref 1). Few long-term
tests have been conducted at temperatures of Assuming the activities of the metal and the
650 to 980 °C (1200 to 1800 °F) where most oxide are unity, Eq 3.2 becomes:
stainless steels, Fe-Ni-Cr alloys, and Ni-Cr
alloys are used in high-temperature applications. DG
=RT ln pO2 ð3:3Þ
Furthermore, many test results were presented
as weight changes instead of actual measure- Then


ments of the damage to the metal, such as the pO2 =eDG =RT
ð3:4Þ
total depth of oxidation attack including both
metal loss (thickness reduction) and internal Equation 3.4 permits the determination of the
penetration. oxygen partial pressure in equilibrium with the

0.5 mm

Fig. 3.1 A Ni-Cr alloy furnace heater coil suffering extensive internal oxidation attack with little surface scaling after service for 4 to 5
years at temperatures below 900 °C (1650 °F). Source: Ref 1
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:43PM Plate # 0 pg 7

Chapter 3: Oxidation / 7

oxide from the standard free energy of formation. partial pressure of oxygen in equilibrium with
The standard free energies of formation of Cr2O3 at 1000 °C (1830 °F) is about 10−21 atm
selected oxides as a function of temperature are from Fig. 3.2. This implies that the formation of
shown in Fig. 3.2. The figure also allows quick Cr2O3 is favored thermodynamically at 1000 °C
determination of the oxygen partial pressure in environments with oxygen potentials higher
(pO2 ) in equilibrium with the oxide. This oxygen than 10−21 atm.
partial pressure can be read by drawing a straight When the environment is “reducing” (e.g., the
line from the point marked “O” on the left ver- environment generated by stoichiometric or
tical axis of Fig. 3.2 through the free-energy line substoichiometric combustion), the oxygen
of the oxide at the intersecting point with the potential is controlled by either pH2 =pH2 O or
temperature of interest. This line continues to pCO =pCO2 ratio. The oxygen potential can be
extend until it intersects with the pO2 scale loca- determined by the reaction:
ted at the right-hand side and bottom of the 2H2 +O2 Ð2H2 O ð3:5Þ
Fig. 3.2. The intersecting point shows the oxygen
partial pressure in equilibrium with the oxide The standard free energy of formation is related
of interest. If the oxygen partial pressure in the to the partial pressures of hydrogen, oxygen, and
environment is greater than the oxygen partial water by:
!
pressure in equilibrium with the oxide, the oxide p2H2 O
is likely to form on the metal surface. Conversely, DG
=7RT ln ð3:6Þ
pH2  pO2
2
the oxide is not likely to form. For example, the

H2/H2O ratio pO
10–8 10–6 10–4 2
CO/CO2 ratio 10–8 10–6 10–4 10–2
10–2
O 0 1
O3 1
e2
–100 6F oO
= = 2C 10–2
+O
2 M O2
o+ 1
–200 O4 NiO 2C
4 Fe 3 =2
+ O2 M 10–4
2Ni
–300 102
10–6
–400 102
∆G °=RT In pO2 (kJ/mole O2)

M
10–8
H –500 2 Cr2O 3
=- 104
O2 3
C 4- Cr +
–600 3 M B 10–10
iO 2 104
=S
O2
–700 Si + 10–12
106
2- A I 2O 3
–800
= 3 10–14
O 2
I+ 106
4- A B
–900 3 O
Mg
O Ca
M = 2 2
=2 B 108 10–16
2 + O
–1000 g + O 2Ca
2M Change of state Element Oxide
M
M 10–18
–1100 Melting point M 108
M
Boiling point B B
–1200 1010 10–20
0 200 400 600 800 1000 1200 1400 1600 1800 2000 2200 2400
Temperature, °C 10
10–22
CO/CO2 ratio 10–14 10–12 10
OK H2/H2O ratio

10–20010–100 10–70 10–60 10–50 10–42 10–38 10–34 10–30 10–28 10–26 10–24
pO
2

Fig. 3.2 Standard free energies of formation of selected oxides as a function of temperature. Source: Ref 2
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:43PM Plate # 0 pg 8

8 / High-Temperature Corrosion and Materials Applications

Rearranging the equation results in: that protects the alloy from rapid oxidation. This
process is known as “selective oxidation.” The

1
pO2 =eDG =RT
ð3:7Þ majority of iron-, nickel-, and cobalt-base alloys
(pH2 =pH2 O )2 rely on selective oxidation of chromium to form a
Cr2O3 scale for oxidation resistance. Some high-
Thus, the oxygen partial pressures at various temperature alloys use aluminum to form an
temperatures can be determined as a function of Al2O3 scale for oxidation resistance.
pH2 =pH2 O values. The pH2 =pH2 O value in equili- Most oxides exhibit high melting points and
brium with the oxide can be read from Fig. 3.2, remain in a solid state for the temperature range
using the method discussed previously, except in which the alloys are used. If the oxide is pre-
that the starting point for the straight line is “H” sent as a liquid state, catastrophic oxidation can
and the pH2 =pH2 O value is determined from the occur. Since many engineering alloys contain
H2/H2O scale. For example, the oxygen poten- many alloying elements for various metallurgical
tial, in terms of pH2 =pH2 O , in equilibrium with reasons, formation of oxides that become liquid
Cr2O3 at 1000 °C, is about 5 × 103 from Fig. 3.2. at the service temperature should be prevented.
Thus, Cr2O3 is likely to form at 1000 °C Table 3.1 shows the melting points of selected
when the pH2 =pH2 O ratio in the environment oxides of alloying elements commonly found in
is less than 5×103. The equilibrium reaction high-temperature alloys. Most oxides remain
for an environment whose oxygen potential is solid until they reach extremely high tempera-
controlled by pCO =pCO2 is: tures. Oxides of molybdenum (MoO3) and
2CO+O2 Ð 2CO2 ð3:8Þ vanadium (V2O5), however, exhibit very low
melting points. Vanadium (V), which is a strong
The corresponding oxygen potential is: carbide former, is often used in alloy steels for
increasing the strength of the material. However,

1 the amount used typically is quite small and is not
pO2 =eDG =RT
ð3:9Þ
(pCO =pCO2 )2 likely to form V2O5. Molybdenum (Mo) is also a
strong carbide former and is used in a small
The pCO =pCO2 value can be read from Fig. 3.2 amount to strengthen low-alloy steels (e.g., Cr-
using the method discussed previously, with the Mo steels). It is unlikely these steels will be
exception that a straight line is drawn from point affected by MoO3-related oxidation problems.
“C” to the CO/CO2 scale. However, molybdenum is an effective alloying
Thus, it is possible to obtain the oxygen element for improving the resistance of the alloy
potential of the environment in terms of to aqueous corrosion. Some stainless steel grades
pO2 , pH2 =pH2 O , pCO =pCO2 , and the oxygen partial contain molybdenum, with superaustenitic
pressure in equilibrium with the oxide of interest stainless steels containing much higher levels of
from Fig. 3.2, to determine whether or not the molybdenum. Some nickel-base alloys contain
oxide is likely to form thermodynamically. very high levels of molybdenum for either
Figure 3.2 also illustrates the relative stability of aqueous corrosion resistance or solid-solution
various oxides. The most stable oxides have the
largest negative values of ΔG°, or the lowest
value of pO2 , or the highest values of pH2 =pH2 O Table 3.1 Melting points of selected oxides
and pCO =pCO2 . for alloying elements commonly found in
It is clear from Fig. 3.2 that oxides of iron, high-temperature alloys
nickel, and cobalt, which are the alloy bases for Qxide Melting point, °C (°F)
the majority of engineering alloys, are sig- αAl2O3 2015 (3659)
nificantly less stable than the oxides of some CoO 1935 (3515)
solutes (e.g., chromium, aluminum, silicon, etc.) Cr2O3 2435 (4415)
FeO 1420 (2588)
in engineering alloys. When one of these solute Mn3O4 1705 (3101)
elements is added to iron, nickel, or cobalt, MoO3 795 (1463)
internal oxidation of the solute is expected to Nb2O5 1460 (2660)
NiO 1990 (3614)
occur if the concentration of the solute is rela- SiO2 1713 (3115)
tively low. As the solute concentration increases TiO2 1830 (3326)
to a sufficiently high level, oxidation of the solute V2O5 690 (1274)
WO3 1473 (2683)
will be changed from internal oxidation to
Source: Ref 3
external oxidation, resulting in an oxide scale
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:43PM Plate # 0 pg 9

Chapter 3: Oxidation / 9

strengthening. The formation of MoO3 and its volatile CrO3 (Ref 5–8). Table 3.2 shows the
effect on oxidation are discussed in section 3.2.2 weight-loss data for Cr2O3 pellets after heating to
“Volatility of Oxides” and section 3.4.17 “Cata- 1000 to 1200 °C (1830 to 2190 °F) in dry O2 due
strophic Oxidation.” to formation of gaseous CrO3 by oxidation of
Cr2O3 (Ref 5). Caplan and Cohen (Ref 5) also
observed that moisture promoted volatilization
3.2.2 Volatility of Oxides of Cr2O3. Asteman et al. (Ref 9) indicated that
Some oxides exhibit high vapor pressures at high vapor pressure of CrO2(OH)2 can form by
very high temperatures (e.g., above 1000 °C, or reacting Cr2O3 with H2O in O2-containing
1830 °F). Oxide scales become less protective environments. The theoretical calculated partial
when their vapor pressures are high. Figure 3.3 pressure of CrO2(OH)2 as a function of tem-
shows vapor pressures of several refractory metal perature for the O2-containing environment with
oxides exhibiting high vapor pressures at tem- pO2 =0:9 atm and pH2 O =0:1 atm is shown in
peratures above 1000 °C (1830 °F) (Ref 4). Fig. 3.4 (Ref 9).
Vanadium is typically used in small quantities as
a carbide former in alloy steels. Thus, the vola-
tility of VO2 is generally of no concern in oxi-
dation of alloys. Molybdenum (Mo) and tungsten Table 3.2 Weight loss of Cr2O3 on heating in
(W) are often used as alloying elements in sig- dry O2 and Ar environments
nificant amounts in Ni- or Co-base alloys as Temperature, Gas flow, Weight
Run °C (°F) Time, h Gas mL/min loss, mg
solution-strengthening elements. Formation of
1 1100 (2010) 20 Dry O2 200 0.6
WO3 or MoO3 may occur under certain condi- 2 1200 (2190) 20 Dry O2 10 2.1
tions in some alloy systems, particularly in alloy 3 1200 (2190) 20 Dry O2 10 1.3
4 1200 (2190) 20 Dry O2 20 1.8
systems containing insufficient chromium for 5 1200 (2190) 20 Dry O2 200 2.3
forming a protective Cr2O3 scale. 6 1200 (2190) 20 Dry O2 200 2.6
A majority of engineering alloys rely on the 7 1200 (2190) 42 Dry O2 200 8.0
8 1200 (2190) 66 Dry Ar 200 0
Cr2O3 scale to provide resistance to oxidation. 9 1200 (2190) 115 Dry Ar 192 0
When heated to very high temperatures (i.e., Source: Ref 5
above 1000 °C), Cr2O3 can react with O2 to form

Temperature, °C
1500 1000 600
1

10–2 WO3

CrO3
10–4
MoO3
10–6
Vapor pressure, atm

VO2
–8
10

10–10

10–12

10–14

10–16

10–18
0.5 0.6 0.7 0.8 0.9 1 1.1 1.2
Inverse temperature, 103/T (K–1)

Fig. 3.3 Vapor pressures of several refractory metal oxides exhibiting high vapor pressures at temperatures above 1000 °C (1830 °F).
Source: Ref 4
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:43PM Plate # 0 pg 10

10 / High-Temperature Corrosion and Materials Applications

–2

–4

CrO2(OH)2, g
–6
Log p, X, atm

–8

–10

–12

CrO3, g
–14

–16
600 700 800 900 1000 1100 1200 1300
Temperature, K

Fig. 3.4 Theoretical partial pressures of CrO2(OH)2 and CrO3 as a function of temperature for the environment with pO2 =0:9 atm and
pH2 O =0:1 atm: Source: Asteman et al. (Ref 9)

For service temperatures above 1200 °C

(2190 °F), the increasing volatility of oxides can Inverse log
progressively cause the oxide to lose its protec-
tive capability. SiO2 and Al2O3 are the only two Parabolic
Log
Oxide mass, m

oxides that are capable of forming a very pro-

tective barrier against oxidation at temperatures
above 1200 °C (Ref 9). However, the SiO2 scale
may lose some protective capability by forming
Linear
gaseous SiO at low oxygen partial pressures
(Ref 10).
Time, t

Fig. 3.5 Different oxidation kinetics. Source: Ref 11

3.3 Kinetic Considerations
The kinetics of oxidation of metals and alloys decreases with increasing time due to the
generally follow several reaction rates. Most increasing diffusion distance for ions. The oxi-
reactions follow a parabolic rate. Some reactions dation rate is, thus, inversely proportional to the
follow a linear rate. Some other reaction kinetics thickness of the oxide scale:
may include logarithmic and inverse logarithmic
rates. These reaction kinetics are illustrated X 2 =k′t ð3:10Þ
schematically in Fig. 3.5 (Ref 11), and a brief where X is the oxide scale thickness, t is the
summary, based on the article by Danielewski exposure time, and k′ is the parabolic constant;
(Ref 11), is presented in sections 3.3.1 to 3.3.3. when t = 0, X = 0.

3.3.1 Parabolic Kinetics

When the oxide scale forms on the metal sur- 3.3.2 Linear Kinetics
face, the oxidation reaction is controlled by the When the oxide scale forming on the metal
diffusion of ions through the oxide scale, which surface provides no protection barrier due to
is in turn controlled by the chemical potential oxide cracking and spalling, volatile oxides, and
gradient as a driving force. As the thickness of molten oxidation products, the oxidation rate
the oxide scale increases, the rate of oxidation generally remains constant with increasing
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:43PM Plate # 0 pg 11

Chapter 3: Oxidation / 11

time. The linear oxidation kinetic rate can be The inverse logarithmic rate can be expressed
expressed by: by the following equation:
X =kl t ð3:11Þ 1=X =b7ki log t ð3:13Þ

where X is the mass (or thickness) of the oxide, t where b and ki are constants.
is the exposure time, and kl is the linear rate
constant; when t = 0, X = 0.
3.4 Oxidation in Air, O2, and “Clean”
3.3.3 Logarithmic and Inverse Combustion Atmospheres
Logarithmic Kinetics 3.4.1 Carbon and Cr-Mo Steels
At very low temperatures when the oxide film Carbon and Cr-Mo steels are the most widely
forms on the metal surface, the oxidation rate used engineering materials and are used exten-
usually follows either a logarithmic or inverse sively for high-temperature applications in
logarithmic rate. The driving force for the oxi- power generation, chemical and petrochemical
dation is the electric field across the oxide film. processing, petroleum refining, pulp and paper
The logarithmic rate can be expressed by: industry, industrial heating, and metallurgical
X =ke log (at+1) ð3:12Þ processing.
At temperatures below 570 °C (1060 °F), iron
where ke and a are constants. (Fe) oxidizes to form Fe3O4 and Fe2O3. Above

120

110

100

90 1400 °F

80
Weight-loss, mg/cm2

70
1200 °F
60

50

40

30

20 1000 °F

10 800 °F

0
0 100 200 300 400 500 600 700
Time, h

Fig. 3.6 Oxidation behavior of plain low-carbon steel in air at 430, 540, 650, and 760 °C (800, 1000, 1200, and 1400 °F). Source:
Ref 12
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:43PM Plate # 0 pg 12

12 / High-Temperature Corrosion and Materials Applications

570 °C (1060 °F), it oxidizes to form FeO, 1000 °F), carbon steel showed very little weight
Fe3O4, and Fe2O3. The oxidation behavior of gain after exposure for 500 h. As the temperature
carbon steel in air at 430, 540, 650, and 760 °C was increased to 650 °C (1200 °F), the oxidation
(800, 1000, 1200, and 1400 °F) is summarized in rate was significantly increased. Carbon steel
Fig. 3.6 (Ref 12). At 430 and 540 °C (800 and suffered rapid oxidation at 760 °C (1400 °F),
exhibiting essentially a linear rate of oxidation
attack. Vrable et al. (Ref 13) reported oxidation
data for carbon steel and high-strength low-alloy
(HSLA) steel, as shown in Fig. 3.7 (Ref 14). At
Calculated continuing 53 mpy*
650 °C (1200 °F), carbon steel suffered an oxi-
10.5 penetration rate
dation rate of about 1.3 mm/yr, or 53 mpy (mils
per year). The oxidation rate is expected to be
Average penetration/side, mils

9.0
much higher when exposed to temperatures
7.5 Carbon steel - 1200 °F higher than 650 °C (1200 °F). Recent test results
by John (Ref 15) showed that carbon steel
6.0 exhibited about 0.25 mm/yr (10 mpy) of oxida-
tion at 604 °C (1120 °F). Figure 3.7 also illus-
4.5 trates that HSLA steel is more oxidation resistant
5.3 mpy*
A242 type 1 HSLA steel - 1200 °F than carbon steel, presumably due to minor
3.0
alloying elements such as manganese, silicon,
Carbon steel - 1000 °F chromium, and nickel.
1.5 2.3 mpy* Cr-Mo steels are used at higher temperatures
A242 type 1 HSLA steel - 1000 °F 1.0 mpy*
than carbon steel because of higher tensile and
0 200 400 600 800 1000 creep-rupture strengths as well as better micro-
Exposure time, h structural stability. Molybdenum and chromium
provide not only solid-solution strengthening but
Fig. 3.7 Oxidation of carbon steel and high-strength low-alloy
also carbide strengthening. Low-alloy steels with
(HSLA) steel in air. Source: Ref 13, reproduced
from Ref 14 chromium and silicon additions exhibit better
oxidation resistance than carbon steel. The ben-
eficial effects of chromium and silicon additions
to carbon steel are summarized in Fig. 3.8 (Ref
16). Silicon is very effective in improving the
oxidation resistance of Cr-Mo steels. Addition of
1.5% Si to 5Cr-0.5Mo steel significantly
improved its oxidation resistance. The most
important alloying element for improving oxi-
dation resistance is chromium. As shown in
Fig. 3.8, for 0.5% Mo steels, increasing chro-
mium from 1 to 9% significantly increases oxi-
dation resistance. The 7Cr-0.5Mo and 9Cr-1Mo
steels showed negligible oxidation rates at tem-
peratures up to 680 °C (1250 °F) and 700 °C
(1300 °F), respectively. Further increases in
chromium improve oxidation resistance even
more. Alloys become martensitic or ferritic
grades of stainless steels (400 series) when
chromium content is increased to 12% or higher.

3.4.2 Martensitic, Ferritic, and

Austenitic Stainless Steels
The superior oxidation resistance of marten-
sitic and ferritic stainless steels to that of carbon
and Cr-Mo steels is illustrated in Fig. 3.9
Fig. 3.8 Effects of chromium and/or silicon on the oxidation
resistance of steels in air. Source: Ref 16 (Ref 17). As chromium content in the straight
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:44PM Plate # 0 pg 13

Chapter 3: Oxidation / 13

Fig. 3.9 Oxidation resistance of carbon, low-alloy, and stainless steels in air after 1000 h at temperatures from 590 to 930 °C (1100 to
1700 °F). Source: Ref 17

chromium steels increases from 9 to 25%, resis- exhibited a thin, adherent (Fe,Cr)2O3 scale, as
tance to oxidation improves significantly. The observed by Walter et al. (Ref 20). The growth of
25Cr steel (Type 446) is the most oxidation the (Fe,Cr)2O3 scale as a function of the accu-
resistant among the 400 series stainless steels, mulated isothermal hold time up to 1000 h is
due to the development of a continuous Cr2O3 shown in Fig. 3.12 (Ref 20). John (Ref 15)
scale on the metal surface. In Fe-Cr alloys, it reported that Fe-12Cr steel (Type 410) exhibited
appears that a minimum of approximately 18wt an air oxidation rate of 0.25 mm/yr (10 mpy) at
% Cr is needed to develop a continuous Cr2O3 832 °C (1530 °F).
scale against further oxidation attack (Fig. 3.10) Another ferritic stainless steel, 18SR (about
(Ref 18). Cyclic oxidation studies conducted by 18% Cr), was found to be as good as, and
Grodner (Ref 19) also revealed that Type 446 sometimes better than, Type 446 (25% Cr), as
was the best performer in the 400 series stainless illustrated in Tables 3.3 and 3.4 (Ref 21). This
steels, followed by Types 430 (14–18Cr), 416 was attributed to the addition of 2% Al and 1% Si
(12–14Cr), and 410 (11.5–13.5Cr) (Fig. 3.11). to the alloy. Furthermore, both of these ferritic
Figure 3.11 shows that Fe-12Cr steels, such as stainless steels, Type 446 and 18SR, showed
Types 410 and 416, showed increased rates of better cyclic oxidation resistance than some
cyclic oxidation above 760 °C (1400 °F). At austenitic stainless steels, such as Type 309 and
650 °C (1200 °F) in air, cycling from 650 to 310, when cycled to 980 to 1040 °C (1800 to
300 °C, 12Cr-1Mo steel (X20 CrMoV 12 1) steel 1900 °F), as shown in Table 3.4.
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:44PM Plate # 0 pg 14

14 / High-Temperature Corrosion and Materials Applications

Fe2O3
Fe3O4
FeO
Fe

10–6
Fe2O3
Fe3O4
FeO
Parabolic rate constant, g2 · cm4 · s1

10–7 Fe – 2Cr Fe/Cr oxide

Fe2O3

(Fe,Cr)2O3
10–8
Fe – 9Cr

Fe3O4 Fe2O3
10–9
Cr2O3

Fe – 16Cr FeFe(2 ... x)CrxO4

10–10
Cr2O3

Fe – 28Cr

10–11
0 10 20 30 40 50 60 70 80 90 100
Alloy chromium content, wt%

Fig. 3.10 Effect of chromium content on oxidation of Fe-Cr alloys at 1000 °C (1830 °F) in 0.13 atm O2. Source: Ref 18

When the service temperature is above strength than do ferritic stainless steels. Fur-
650 °C (1200 °F), ferritic stainless steels, which thermore, they do not suffer 475 °C (885 °F)
have a body-centered cubic (bcc) crystal struc- embrittlement or ductility-loss problems in thick
ture, drastically lose their elevated-temperature sections and in heat-affected zones as do ferritic
strength (both tensile and creep-rupture stainless steels. Nevertheless, some austenitic
strength). As a result, the application of ferritic stainless steels can suffer some ductility loss
stainless steels becomes limited at higher tem- upon long-term exposure to intermediate tem-
peratures. At these temperatures, alloys with a peratures (e.g., 540 to 800 °C, or 1000 to
face-centered cubic (fcc) crystal structure are 1470 °F) due to sigma-phase formation.
preferred because of their higher creep-rupture The oxidation resistance of two austenitic
strength. Nickel is added to Fe-Cr steels to sta- stainless steels, Types 309 and 310, is compared
bilize the austenitic structure. The austenitic with that of a number of ferritic stainless steels in
structure is inherently stronger and more creep Fig. 3.11. Several austenitic stainless steels are
resistant than the ferritic structure (Ref 22). compared in Fig. 3.13 (Ref 23). Nickel improves
The 300 series austenitic stainless steels have the resistance of alloys to cyclic oxidation.
been widely used for high-temperature compo- Moccari and Ali (Ref 24) also observed the
nents in various industries because of their similar beneficial effects of nickel in improving
strength and high-temperature corrosion resis- the oxidation resistance of alloys. Brasunas et al.
tance, including oxidation resistance. These (Ref 25) studied the oxidation behavior of
alloys exhibit higher elevated-temperature about 80 experimental Fe-Cr-Ni alloys exposed
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:44PM Plate # 0 pg 15

Chapter 3: Oxidation / 15

Fig. 3.11 Oxidation resistance of several stainless steels as a function of temperature. Source: Ref 19

40
Thickness of the oxide layer (d ), µm

35
30
25
20
15
10
5
0
0 100 200 300 400 500 600 700 800 900 1000
Testing time (t ), h

Fig. 3.12 Oxidation behavior of 12Cr-1Mo steel (X20 CrMoV 12 1) at 650 °C (1200 °F) in air with every 8 h of exposure specimens
being cycled from 650 to 300 °C. Source: Ref 20

to air-H2O mixture at 870 to 1200 °C (1600 to internal oxidation penetration) at 893 and 982 °C
2190 °F) for 100 and 1000 h. They observed that (1640 and 1800 °F), respectively, in air. In
increases in nickel in excess of 10% in alloys Fig. 3.13, several high-nickel alloys were found
containing 11 to 36% Cr improved the oxidation to be more resistant to oxidation than austenitic
resistance of the alloys. John (Ref 15) reported stainless steels. The oxidation behavior of high-
that Type 304 and 310 exhibited 0.25 mm/yr nickel Fe-Ni-Cr alloys and Ni-base alloys is
(10 mpy) of oxidation attack (both metal loss and discussed in later sections of this chapter.
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:44PM Plate # 0 pg 16

16 / High-Temperature Corrosion and Materials Applications

Table 3.3 Cyclic oxidation resistance of several Kado et al. (Ref 26) also investigated oxida-
stainless steels in air cycling to 870 to 930 °C tion behavior in a combustion environment that
(1600 to 1700 °F) temperature range simulated the gasoline engine. Their test
Specimen weight changes after indicated cycles, mg/cm2 involved air-to-fuel ratios of 9 to 1 and 14.5 to 1
Alloy 288 cycles 480 cycles 750 cycles 958 cycles and regular gasoline that contained 0.01wt% S.
409 + Al Destroyed … … … Exhaust gas taken from the exhaust manifold was
430 9.9 Destroyed … …
22-13-5 0.5 −3.0 −18.8 −35.7
mixed with air before being piped into a furnace
442 0.7 1.2 1.5 1.5 retort where tests were performed. Test speci-
446 0.3 0.4 0.2 0.1 mens were exposed to the mixture of exhaust gas
309 0.3 −4.6 −23.7 −32.6
18SR 0.3 0.4 0.5 0.6 and air. The gas mixture consisted of 72.4% N2,
Note: 15 min in furnace and 15 min out of furnace. Source: Ref 21
9.7% H2O, 9.93% O2, 8% CO2, and 507 ppm
NOx, when an air-to-fuel ratio of 14.5 to 1 was
used for combustion, while that coming from the
combustion using an air-to-fuel ratio of 9 to 1
Table 3.4 Cyclic oxidation resistance of several consisted of 70.6% N2, 13.7% H2O, 3.21% O2,
stainless steels in air cycling to 980 to 1040 °C 12.5% CO2, and 34 ppm NOx. The total accu-
(1800 to 1900 °F) temperature range mulated test duration was 200 h (400 cycles
Specimen weight changes after indicated cycles, mg/cm2 with 30 min in the hot zone and 30 min out of
Alloy 130 cycles 368 cycles 561 cycles 753 cycles 1029 cycles the hot zone). Test results along with air oxida-
446 0.4 0.5 −0.2 7.0 −19.4 tion data are summarized in Fig. 3.16. There were
18SR 0.7 1.1 1.5 2.2 3.0 no significant differences between air and
309 −24.2 −77.5 −178.3 −242 −358
310 1.5 −11.3 −29.3 −62.8 −107
exhaust gas test environments when tested at
Note: 15 min in furnace and 15 min out of furnace. Source: Ref 21
800 °C (1470 °F). All the alloys tested showed
negligible attack except Type 304, which
exhibited much more severe attack in the exhaust
environment. When the test temperature was
increased to 1000 °C (1830 °F), all the 400
In evaluating materials for automobile emis- series stainless steels with less than 17% Cr (i.e.,
sion-control devices, such as thermal reactors Types 409, 405, 410, and 430) and Type 304
and catalytic converters, Kado et al. (Ref 26) and exhibited significantly more oxidation attack in
Michels (Ref 27) have carried out cyclic oxida- the exhaust environment. Type 310, Type 446,
tion tests on various stainless steels. In cyclic DIN 4828 (Fe-19Cr-12Ni-2Si), F-1 alloy (Fe-
oxidation tests performed by Kado et al. (Ref 26) 15Cr-4Al), A-1 alloy (Fe-16Cr-13Ni-3.5Si), and
in still air at 1000 °C (1830 °F) for 400 cycles A-2 alloy (Fe-20Cr-13Ni-3.5Si) performed well.
(30 min in the furnace and 30 min out of the At 1200 °C (2190 °F), all the alloys tested suf-
furnace), Types 409 (12Cr), 420 (13Cr), and 304 fered severe oxidation attack with the exhaust
(18Cr-8Ni) suffered severe attack. Type 420 environment being more aggressive than air.
(13Cr) was completely oxidized after only 100 These authors attributed the enhanced attack to
cycles, although the sample did not show any the presence of sulfur in the exhaust gas envi-
weight changes. Alloys that performed well ronment, although low-sulfur (0.01%) gasoline
under these conditions were Types 405 (14Cr), was used for testing. Sulfur segregation to the
430 (17Cr), 446 (25Cr), 310 (25Cr-20Ni), and scale/metal interface was detected.
DIN 4828 (19Cr-12Ni-2Si), as illustrated in In a study by Michels (Ref 27), the engine
Fig. 3.14. When cycled to 1200 °C (2190 °F) for combustion atmosphere was also found to be
400 cycles (30 min in the furnace and 30 min out significantly more corrosive than the air-10%
of the furnace), all the alloys tested except F-1 H2O environment. The engine combustion
alloy (Fe-15Cr-4Al) suffered severe oxidation exhaust gas contained about 10% H2O along
attack (Fig. 3.15). This illustrates the superior with 2% CO, 0.33 to 0.55% O2, 0.05 to 0.24%
oxidation resistance of alumina formers (i.e., hydrocarbon, and 0.085% NOx. The balance was
alloys that form Al2O3 scales when oxidized at presumably N2 (not reported in the paper). The
elevated temperatures). The data also illustrate engine exhaust gas was piped into a furnace
that for temperatures as high as 1200 °C retort where the tests were performed. The
(2190 °F), Cr2O3 oxide scales can no longer results, which were generated in the air-10%H2O
provide adequate oxidation resistance. Oxidation and the engine exhaust environment, are shown
of Fe-Cr-Al alloys is discussed in Section 3.4.7. in Fig. 3.17. After exposure to the air-10%H2O
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:44PM Plate # 0 pg 17

Chapter 3: Oxidation / 17

10

625
600
0 80Ni-20Cr
60Ni-15Cr
800
22Cr-32Ni

–10

Type 310
25Cr-20Ni

– 20

19Cr-14Ni
20Cr-25Ni
Change in weight, %

Type 309
23Cr-13Ni

– 30

– 40
Type 347
18Cr-8Ni(Cb)

– 50

Type 304
18Cr-8Ni

– 60

– 70
0 200 400 600 800 1000
Hours of cyclic exposure
(15 min heating – 5 min cooling)

Fig. 3.13 Cyclic oxidation resistance of several stainless steels and nickel-base alloys in air at 980 °C (1800 °F). Source: Ref 23

environment at 980 °C (1800 °F) for 102 h, Type in the gasoline used in this test was not reported.
309, Type 310, 18SR, alloy OR-1 (Fe-13Cr-3Al), The relatively high gas velocity, about 6.1 to 9.2
alloy 800, and alloy 601 were all relatively m/s (20 to 30 ft/s) was considered by the author to
unaffected. On the other hand, only 18SR and be one of the possible factors responsible for
alloy 601 were relatively unaffected by the much higher oxidation attack in the engine
engine exhaust gas environment, with alloy exhaust gas test.
OR-1, Type 309, Type 310, and alloy 800 suf- Oxidation data generated in combustion
fering severe oxidation attack. The sulfur content atmospheres is relatively limited. No systematic
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:44PM Plate # 0 pg 18

18 / High-Temperature Corrosion and Materials Applications

Fig. 3.14 Cyclic oxidation resistance of several ferritic and austenitic stainless steels in still air at 1000 °C (1830 °F) for up to 400 cycles
(30 min in furnace and 30 min out of furnace). Source: Ref 26

409

0
F–1
(Fe-15Cr-4AI)

430
Weight change, mg/cm2

–100

DIN 4828
420 (19Cr-12Ni-2Si)

–200

310

304 446
–300

0 100 200 300 400

Number of cycles

Fig. 3.15 Cyclic oxidation resistance of several ferritic and austenitic stainless steels in still air at 1200 °C (2190 °F) for up to 400 cycles
(30 min in furnace and 30 min out of furnace). Source: Ref 26
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:44PM Plate # 0 pg 19

Chapter 3: Oxidation / 19

1200 °C (2190 °F)/400 cycles

100

80

60

40

20
Thickness loss, %

0
1000 °C (1830 °F)/400 cycles
100

80

60

40

20

0
800 °C (1470 °F)/400 cycles
20

0
)

)
)
9

0S
28

Si

Si
AI
40

40

41

43

44

30

31

.5

.5
31
r-4

48

i-3

i-3
C

IN

3N

3N
15

D
e-

r-1

r-1
(F

C
16

20
1
F–

e-

e-
(F

(F
1

In air
A–

A–

In exhaust gas (R = 9)

Fig. 3.16 Comparison of cyclic oxidation resistance between air and gasoline engine exhaust gas environments at 800, 1000, and
1200 °C (1470, 1830, 2190 °F) for 400 cycles (30 min in hot zone and 30 min out of hot zone). Alloy F-1 suffered localized
attack at 1200 °C in engine exhaust gas. Source: Ref 26

studies have been reported that varied combus- Chapter 4 “Nitridation.” The presence of water
tion conditions, such as air-to-fuel ratios. In vapor can be an important factor in affecting
combustion atmospheres, the oxidation of metals oxidation behavior of alloys. The effect of water
or alloys is not controlled by oxygen only. Other vapor on the oxidation resistance of alloys is
combustion products, such as H2O, CO, CO2, covered in Section 3.4.15.
N2, hydrocarbon, and others, are expected to
influence oxidation behavior. When air is used
for combustion, nitridation in conjunction with 3.4.3 Surface Chemistry versus
oxidation can occur in combustion atmospheres Bulk Chemistry
under certain conditions. This nitridation/ It is important to point out that the surface
oxidation behavior of alloys is discussed in chemistry may not be the same as the reported
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:44PM Plate # 0 pg 20

20 / High-Temperature Corrosion and Materials Applications

Fig. 3.17 Cyclic oxidation resistance of several ferritic and austenitic stainless steels in (a) air-10H2O at 980 °C (1800 °F) cycled every
2 h, and (b) gasoline engine exhaust gas at 980 °C (1800 °F) cycled every 6 h. Source: Ref 27

“bulk” chemistry. Some of the manufacturing becomes more critical for sheet products or thin-
processes involved in the production of an alloy wall tubular products because of much higher
product, such as plate, sheet, or tubular products, percentage of the surface-depletion zone in the
may result in lower chromium contents at or overall thickness of the component. In alloy
near the surface of the product. This is particu- manufacturing, annealing is required after each
larly important for austenitic stainless steels, cold-rolling step for sheet product manufacturing
since the chromium specification range for aus- (or cold pilgering for reduction in thickness and
tenitic stainless steels can be at or near the bor- sizes for tubular product manufacturing) to
derline for forming a continuous Cr2O3 scale. For soften the metal for further cold-reduction steps
example, ASTM A 213/A 213M (or ASME SA- until a final product is produced. Stainless steels
213/SA-213M) specification for the chromium are typically annealed in the temperature range of
range is 18.0 to 20.0% for Type 304, 16.0 to 1010 to 1120 °C (1850 to 2050 °F) (Ref 28).
18.0% for Type 316, 17.0 to 20.0% for Type 321, When annealing is performed in air, heavy
and 17.0 to 20.0% for Type 347. The lower end of chromium oxide scales form on the metal sur-
the chromium content in these alloys is essen- face. As a result, the matrix immediately under-
tially at the minimum level that is considered to neath the oxide scales can be depleted in
be required for forming a continuous Cr2O3 scale chromium. Figure 3.18 shows the concentration
when exposed to elevated temperatures. A slight profile of chromium near the surface of the plate
surface depletion in chromium due to some of alloy AL-6XN (Fe-21Cr-24Ni-6.5Mo-0.2N)
manufacturing processes may result in the after annealing in air at 1120 and 1175 °C (2050
condition that a continuous, protective Cr2O3 and 2150 °F), respectively (Ref 29). Oxidation
scale cannot be formed, thus resulting in accel- during annealing at either temperature resulted
erated oxidation due to formation of non- in a significant chromium depletion near the
protective iron oxides. surface of the plate. It is quite common to per-
Some manufacturing processes are prone to form annealing in air during manufacturing of
producing the final finished product with surface stainless steels. This is commonly referred to
depletion of chromium. This surface depletion as “black annealing,” as opposed to “bright
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:44PM Plate # 0 pg 21

Chapter 3: Oxidation / 21

annealing,” which is performed in a protective decreases. Figure 3.19 shows an example of

atmosphere, such as hydrogen atmosphere. The surface depletion of chromium in a thin-gage
oxide scales on the alloy plate, sheet, or tubing commercial heat-exchanger tube made from
are generally removed by acid pickling. This Type 321 in the as-fabricated condition (Ref 30).
manufacturing process can often produce flat The manufacturing process involved was not
products (plate or sheet) as well as tubular pro- known. The Type 321 tube is shown to exhibit
ducts with surface depletion of chromium. When depletion in chromium near the tube surface
bright annealing is performed, the alloy surface is when analyzed by energy-dispersive x-ray
protected from oxide scale formation. As a result, spectroscopy (EDX). The analysis was termi-
chromium depletion at and near the product nated at approximately about 0.5 μm from the
surface is minimized. tube surface. It is expected that the chromium
Surface depletion in chromium becomes content would decrease further if the analysis
increasingly critical as the thickness of the com- was performed at locations closer to the surface.
ponent, such as sheet or tubular product, After service for 6 months as a recuperator tube at

22
As hot rolled

20
Chromium content, wt%

18
Annealed 1175 °C

Annealed 1120 °C
16

AL-6XN

14
0 2 4 6 8 10 12 14 16 18 20

Distance, µm

Fig. 3.18 Surface deletion of chromium near the surface of the plate of alloy AL-6XN (Fe-21Cr-24Ni-6.5Mo-0.2N) after annealing in
air at 1120 and 1175 °C (2050 and 2150 °F), respectively. Source: Ref 29

18.4

18.2

18
Cr concentration, wt%

17.8

17.6

17.4

17.2

17

16.8
0 1 2 3 4 5 6 7 8 9 10
Distance from tube OD surface, µm

Fig. 3.19 Surface depletion of chromium observed in a thin-gage commercial heat-exchanger tube in the as-fabricated condition
made from Type 321. Source: Ref 30
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:44PM Plate # 0 pg 22

22 / High-Temperature Corrosion and Materials Applications

metal temperatures of approximately 620 to by a thin, adherent oxide scale, which was not
670 °C (1150 to 1240 °F) for preheating air, clearly revealed in the figure at about 400×
significant oxide spalling and scaling was magnification. Once the formation of the light,
observed on the air side of the heat-exchanger grayish outer oxide scale formed on the metal
tube. Figure 3.20 shows heavy oxide scales surface, the inner oxidation took place with
formed on the side of the tube exposed to the accelerated growth. The outer oxide scale was
incoming air after 6 months of service (Ref 30). found to be iron-rich oxides with little chromium,
The oxide scales were analyzed by scanning as shown in Fig. 3.23 (Ref 30). With the forma-
electron microscopy/energy-dispersive x-ray tion of the iron-rich oxide scale at the outer layer,
spectroscopy (SEM/EDX) analysis; the results which failed to provide protection, the inner
are shown in Fig. 3.21. The analysis showed that Fe-Cr oxides were found to penetrate into the
the outermost oxide layer was essentially iron metal, causing a relatively massive internal oxi-
oxides with very little chromium. Thus, the dation penetration (Fig. 3.23). In the same test,
stainless steel recuperator no longer exhibited the other Type 321 heat-exchanger tube (pro-
adequate resistance to oxidation because of its duced by supplier B), which was subjected to the
failure to form and maintain a protective Cr2O3 same test condition and duration, was found to
scale. Once a protective Cr2O3 scale is no longer exhibit a thin, adherent oxide scale with no evi-
present on the stainless steel surface, iron oxides dence of mushroom-type oxide nodules, as
then take over. This results in scaling and accel- shown in Fig. 3.22(b). The oxide scale was
erating oxidation. This is often referred to as analyzed by SEM/EDX, showing an Fe-Cr oxide
“breakaway” oxidation (or corrosion). “Break- scale formed on the metal surface (Fig. 3.24).
away” oxidation is discussed in Section 3.4.13. Bulk chemical compositions of two tubes were
Stainless steels manufactured by different analyzed; results are shown in Table 3.5. The
producers can exhibit different chemical com- chemical analysis results showed that the chro-
positions and different surface characteristics. mium in the Type 321 tube from supplier A was
This is illustrated in Fig. 3.22 (Ref 30). Two Type essentially at the lower end of the specification
321 heat-exchanger tubes, manufactured by range. Any depletion in chromium near the sur-
two different suppliers (supplier A and B), were face could result in a chromium level that is
tested in the field in the same recuperator as below the specification limit. On the other hand,
described previously for preheating air at the tube from supplier B contained much higher
approximate metal temperature of 620 to 670 °C chromium and was found to be much more
(1150 to 1240 °F) for about 1008 h. Figure 3.22 resistant to oxidation under the same test condi-
(a) shows the formation of mushroom-type oxide tions. It appears that some stainless steel manu-
nodules on the surface of the Type 321 tube, facturers produce their products with leaner
produced by Supplier A. This is considered the chemistry in terms of major alloying elements,
initiation of breakaway oxidation. As shown in such as chromium and nickel. For high-
Fig. 3.22(a), the mushroom-type oxide nodule temperature oxidation and other modes of cor-
consisted of two layers of oxides, light grayish rosion, stainless steel with the chromium in the
outer oxide scale and darkish inner oxide layer. lower end of the specification range can be
The rest of the metal surface was still protected potentially more prone to breakaway oxidation

25 µm

Fig. 3.20 Heavy oxide scales formed on the side of Type 321 recuperator tube that was exposed to the incoming air after 6 months of
service with the metal temperatures approximately 620 to 670 °C (1150 to 1240 °F). This tube was from the same batch of
tubes that shows surface chromium depletion (Fig. 3.19). Source: Ref 30
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:44PM Plate # 0 pg 23

Chapter 3: Oxidation / 23

D
E
25 µm

(a)

30 µm

25 µm
Fig. 3.21 Scanning electron micrograph (backscattered
(b)
electron image) showing the oxide scales formed
on the outside diameter of the heat-exchanger tube (from the same
batch of tubes that showed surface chromium depletion) exposed Fig. 3.22 Type 321 heat-exchanger tubes, which were
to air for 6 months. Energy-dispersive x-ray spectroscopy (EDX) manufactured by two different alloy suppliers,
analysis was performed to determine the chemical compositions were tested in the same facility as described previously for pre-
at different locations, marked “A” to “E.” The results (wt%) of the heating air at approximate metal temperature of 620 to 670 °C
EDX analysis are summarized below (minor elements not inclu- (1150 to 1240 °F) for about 1008 h. (a) Supplier A. (b) Supplier B.
ded). Source: Ref 30 Note the tube from supplier A showed the initiation of accelerated
oxidation attack (a) as opposed to the tube from supplier B
A: 1.9% Cr, 97.6% Fe. showing no sign of accelerated oxidation attack (b).
B: 44.2% Cr, 44.4% Fe, 6.8% Ni.
C: 48.8% Cr, 40.4% Fe, 4.4% Ni.
D: 28.3% Cr, 46.4% Fe, 21.0% Ni.
E: 38.7% Cr, 45.3% Fe, 10.2% Ni. Manufacturing processes can greatly influence
the surface chemistry of an alloy product.
or corrosion. A brief discussion of this important Stainless steels can be finished into the final
issue is presented in Section 3.4.5. product by bright annealing (i.e., annealing is
performed in a protective atmosphere, such as
hydrogen environment or dissociated ammonia
3.4.4 Surface Conditions environment). This process generally produces
As discussed in Section 3.4.3 “Surface a product with minimal depletion of chromium
Chemistry versus Bulk Chemistry,” the con- at or near the surface. On the other hand, when
centration of chromium at and near the surface of the alloy product is finished by black annealing
the alloy product plays a significant role in the (i.e., annealing is performed in air or com-
oxidation of stainless steels such as Types 304, bustion atmosphere in the furnace) and followed
316, 321, 347, and so forth. This is because the by acid pickling, there is a good chance that
chromium concentration of these stainless steels the alloy product may exhibit a surface depletion
is at the lower end of the chromium range that is of chromium. This is particularly important
generally required to form a continuous Cr2O3 for thin-gage sheet products or thin tubular pro-
scale when heated to elevated temperatures. ducts.
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:44PM Plate # 0 pg 24

24 / High-Temperature Corrosion and Materials Applications

In most laboratory oxidation tests, the test chromium oxide scales when heated to elevated
specimens are typically prepared by grinding and temperatures, thus increasing the oxidation
polishing with different grits of emery papers resistance of the alloy. For some critical appli-
prior to testing. The objective of grinding and cations involving a thin-gage sheet (or foil)
polishing the test specimens to a certain surface product or a thin tubular product, testing should
finish condition is to keep the surface condition be carried out on the specimen that retains the
of all test specimens constant in order to compare surface condition of the product without prior
the oxidation behavior of different alloys. How- surface grinding or mechanical polishing.
ever, the mechanical forces of grinding and Electropolishing, which is not commonly used to
polishing can produce a thin cold-worked layer improve the surface finish of the alloy product for
on the specimen surface. This cold-worked high-temperature services, may cause surface
structure at the surface layer can significantly depletion of chromium for the product. Never-
enhance the diffusion of chromium from the theless, some investigators may use electro-
interior to the surface of the metal to form polishing to prepare the surface condition of the
test specimens. Table 3.6 shows some compar-
ison oxidation data generated in wet O2 between
the wet ground and electropolished surface
1 conditions for several stainless steels (Ref 31).

3.4.5 Today’s Stainless Steels

Some stainless steel producers may manu-
2
facture stainless steels at the bottom of the

1
3

2
4

13 µm

13 µm
Fig. 3.23 Scanning electron micrograph (backscattered
electron image) showing the oxide scales formed
on the outside diameter of Type 321 tube (from supplier A) Fig. 3.24 Scanning electron micrograph (backscattered
exposed to air at approximately 620 to 670 °C (1150 to 1240 °F) image) showing the oxide scales formed on the
for 1008 h. Energy-dispersive x-ray spectroscopy (EDX) analysis outside diameter of Type 321 tube (from supplier B) exposed to air
was performed to determine the chemical compositions at dif- at approximately 620 to 670 °C (1150 to 1240 °F) for 1008 hours.
ferent locations as marked 1 to 4 in the oxides. The results (wt%) of EDX analysis was performed to determine the chemical compo-
the EDX analysis are (minor elements not included): sitions at different locations, marked as No. 1 and 2, in the oxides.
The results (wt%) of the EDX analysis are (minor elements not
1: 12% Cr, 86% Fe. included):
2: 52% Cr, 24% Fe, 18% Ni.
3: 32% Cr, 25% Fe, 38% Ni. 1: 35% Cr, 50% Fe, 6% Ni.
4: 37% Cr, 52% Fe, 7% Ni. 2: 21% Cr, 59% Fe, 4% Ni.

Table 3.5 Chemical compositions of Type 321 tubes from Suppliers A and B
Composition, wt%
Supplier C Cr Ni Ti Mn Si Mo Cu P S Fe
A 0.043 17.01 8.85 0.32 1.04 0.54 0.35 0.26 0.032 <0.005 bal
B 0.072 17.77 8.81 0.38 0.94 0.76 0.21 0.09 0.035 0.018 bal
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:44PM Plate # 0 pg 25

Chapter 3: Oxidation / 25

specification range for key alloying elements, oxide scale during service. As a result, iron oxi-
such as chromium, to reduce materials cost. des and isolated nonprotective Fe-Cr oxide
Kelly (Ref 32) discusses this point in his 1992 nodules can develop on the metal surface, thus
paper, as shown in Tables 3.7 and 3.8. Accord- resulting in breakaway oxidation.
ingly, the chromium content can be insufficient
to maintain a continuous chromium oxide scale 3.4.6 Special, Proprietary Austenitic
during prolonged service, or when subjected to Stainless Steels
thermal cycling, or overheating conditions, thus
promoting breakaway oxidation. These “lean” There are several commercial proprietary heat-
stainless steels can be further “aggravated” by resistant alloys that belong to the austenitic
the surface depletion of chromium resulting stainless steel group in terms of chromium
from manufacturing processes that may involve and nickel contents, but with the addition of
excessive pickling after “black” annealing silicon for increasing the resistance of the alloy
(annealing in air or combustion atmosphere), to oxidation and other high-temperature corro-
during successive reductions in cold rolling in sion attack. These alloys include 253MA® (Fe-
flat product manufacturing, or pilgering in tub- 21Cr-11Ni-1.7Si-0.17N-0.04Ce) and 85H® (Fe-
ular manufacturing. This can result in a product, 18.5Cr-14.5Ni-3.5Si-0.2C-1.0Al). Figure 3.25
particularly a thin-gage sheet or tube, in which shows the cyclic oxidation resistance of 253MA
the chromium concentration at the surface is too in air at 1090 °C (2000 °F) compared with that
low to form or maintain a continuous chromium of Type 309 and some higher-alloyed Fe-Ni-
Cr alloys (e.g., 800H and RA330) (Ref 33).
Figure 3.26 shows more cyclic oxidation tests
Table 3.6 Metal losses of several stainless steels comparing 253MA with 800H and 353MA at
tested at 648 °C (1200 °F) for 168 h in wet O2
1093 and 1150 °C (2000 and 2100 °F) (Ref 34).
Alloy Surface finish
Metal loss after
descaling, mg/cm2
Figure 3.27 shows the oxidation resistance of
304 Wet ground 3.6
253MA compared with austenitic stainless
Electropolished 5.9 steels, such as 18-10Ti (Type 321), 20-12Si (DIN
321 Wet ground 1.2 1.4828, 20Cr-12Ni-2Si), and 25-20 (Type 310),
Electropolished 6.5
316 Wet ground 1.8 and 353MA and Ni-Cr alloy 601 as a function of
Electropolished 6.6 temperature (Ref 35). Figure 3.28 shows that
347 Wet ground 2.3
Electropolished 6.4
Source: Ref 31 12
1090 °C
(2000 °F)
20 h cycles in air
10
Table 3.7 Chromium and nickel contents for RA309 800H
Type 304 between pre-1965 heats and current RA310
production heats
8
Weight gain, mg/cm2

Composition, wt% RA330

Cr Ni RA333
ASTM A 240 18.00–20.00 8.00–10.50 6
Pre-1965 heats 18.7 9.9 RA601
Current production 18.3 9.0
Source: Ref 32
4

RA 253 MA
Table 3.8 Chromium, nickel, and molybdenum 2
contents for Type 316 between pre-1965 heats
and current production heats
Composition, wt% 0 100 200 300 400 500
Cr Ni Mo Exposure time, h
ASTM A 240 16.00–18.00 10.00–14.00 2.00–3.00
Pre-1965 heats 17.9 12.4 2.4
Current production 16.3 10.2 2.1
Fig. 3.25 Cyclic oxidation resistance of 253MA in air at
1090 °C (2000 °F) with 20 h cycles compared
Source: Ref 32 with that of Type 309 and some higher-alloyed Fe-Ni-Cr alloys
(e.g., 800H and RA330) up to 500 h testing. Source: Ref 33
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:44PM Plate # 0 pg 26

26 / High-Temperature Corrosion and Materials Applications

160 specimen was cathodically descaled, instead of

870 °C weight gain. The oxidation resistance of RA330
(1600 °F) along with other alloys in the 20–25Cr/30–40Ni
140
980 °C group of Fe-Ni-Cr alloys is discussed in Section
(1800 °F)
120 3.4.8.
1090 °C
(2000 °F)

Weight gain, mg/cm2

100
1150 °C 3.4.7 Fe-Cr-Al Alloys
(2100 °F)
80 Aluminum is a very effective alloying element
in improving the resistance of the alloy to oxi-
60 dation and other high-temperature corrosion
attack. The alloy normally requires about 4% Al
40 to form a continuous Al2O3 scale. The Al2O3
scale provides exceptional protection against
20 oxidation attack. Figure 3.29 shows parabolic
rate constants of some important oxides, such as
0 Al2O3, Cr2O3, SiO2, FeO, NiO, and CoO (Ref 2).
RA 253 MA RA 353 MA RA800H/AT
Al2O3 exhibits the lowest parabolic rate con-
Fig. 3.26 Cyclic oxidation tests performed on 253MA, stants. When the metal is heated to 1200 °C
353MA, and 800H at 871, 982, 1093, and 1150 °C (2200 °F) and higher, the Cr2O3 scale, which
(1600, 1800, 2000, and 2100 °F) for 1640 h with specimens
cycling to room temperature every 160 h. Source: Ref 34
exhibits high growth rates as well as forming
volatile CrO3, becomes essentially non-
protective. Under these very high temperature
conditions, Al2O3 scale provides excellent pro-
3
18 - 10Ti 20 - 12Si 253 MA tection against oxidation. Because of extremely
slow growth rates at low and intermediate tem-
Weight increase, g/m2 h

peratures, Al2O3 scale provides less protection at

2 these temperatures. As a result, high-temperature
153 MA alloys that are designed to form Al2O3 scale for
very high temperatures also contain adequate
25 - 20 Alloy 601 chromium to form Cr2O3 scale for intermediate
1
temperatures.
Some commercial electrical resistance heating
353 MA elements are made of Fe-Cr-Al alloys, such as
0 Kanthal® alloys, which rely on the formation of
1000 1100 1200
the Al2O3 scale for applications up to 1400 °C
Temperature, °C
(2550 °F) (Ref 37). For example, some of the
Isothermal oxidation rates (g/m2 h) obtained by
Kanthal alloys that are available in wire, strip,
Fig. 3.27
testing for 45 h at various temperatures for 253MA and ribbon product forms are Kanthal A-1
and 353MA compared with some austenitic stainless steels, such (Fe-22Cr-5.8Al), AF (Fe-22Cr-5.3Al), and D
as 18-10Ti (Type 321), 20-12Si (DIN 1,4828, 20Cr-12Ni-2Si), 25-
20 (Type 310), and nickel alloy 601. 1.0 g/m2 = 0.1 mg/cm2. (Fe-22Cr-4.8Al). Since these wrought alloy
Source: Ref 35 products are essentially ferritic alloys, they
exhibit low creep-rupture strengths when the
temperature exceeds 650 °C (1200 °F) and can-
RA85H was better than alloy 800H, but not as not be used for high-temperature structural
good as some higher-alloyed Fe-Ni-Cr alloys components. Thus, the electrical resistance
(e.g., RA330, 353MA) at 1150 °C (2100 °F) in heating elements made of these alloys have to be
air with 164 h cycles (Ref 33). In a long-term properly supported to avoid creep deformation,
cyclic oxidation test in air at 1090 °C (2000 °F) such as sagging. These Kanthal wires can be used
for times up to 1 year with specimens cycling to in arc or flame spraying to produce an oxidation-
room temperature after every 20 h of exposure, resistant coating or in weld overlay cladding by
both 253MA and 85H were not as good as using gas metal arc welding (GMAW) process.
RA330, as shown in Table 3.9 (Ref 36). The A powder metallurgy (P/M) process was used
oxidation resistance in this test was measured to produce a new alloy product, Kanthal APM.
by metal loss, which was determined after the This P/M-produced Kanthal APM alloy was
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:44PM Plate # 0 pg 27

Chapter 3: Oxidation / 27

120

1150 °C (2100 °F)

164 h cycles in air

164
100

80
80
74
72
Weight gain, mg/cm2

60

40
36

23
20
20

RA85H RA310 RA330 RA333 RA 353MA RA600 800H

Total exposure time, 1640 h

Fig. 3.28 Cyclic oxidation tests at 1150 °C (2100 °F) for 1640 h in air, with specimens cycling to room temperature every 164 h for
Type 310, RA85H, 800H, 353MA, RA330, and nickel-base alloys 600 and RA333. Source: Ref 33

Table 3.9 Isothermal and cyclic oxidation tests Fecralloy® (Fe-16Cr-4Al-0.3Y) developed by
in air at 1090 °C (2000 °F) U.K. Atomic Energy Authority (Ref 40). Alle-
Metal loss, μm (mils) gheny Ludlum (Ref 39) conducted cyclic oxi-
Alloy Exposure time Cyclic(a) Isothermal dation tests in air by repeatedly resistance heating
85H 3000 h 883 (34.8) 41 (1.63) a thin ALFA IV foil (0.050 mm, or 0.02 in.,
1 year 2665 (105) … thick) to various temperatures for 2 min until
253MA 3000 h 1033 (40.7) 22 (0.88)
1 year 3730 (147) … it failed. The cycles to failure of the foil are
RA330 3000 h 683 (26.9) 44 (1.74) summarized in Table 3.10. No definition of the
1 year 2005 (79) … cycle to failure was given in the report. It is
310 3000 h … 34 (1.34)
believed that the failure of the foil was defined as
(a) Specimens were cycled to room temperature after every 20 h of exposure.
Source: Ref 36 when oxidation penetrated through the thin foil.
ALFA IV uses a rare earth element, cerium, for
enhancement of the Al2O3 scale. Very few stu-
reported to exhibit higher creep-rupture strengths dies have been conducted on the effectiveness of
(Ref 38). cerium on the adhesion of the Al2O3 scale. More
Other commercial Fe-Cr-Al alloys include studies have been conducted on the effects of
ALFA-I™ (Fe-13Cr-3Al), ALFA-II™ (Fe-13Cr- yttrium, zirconium, and other reactive elements.
4Al), and ALFA-IV™ (Fe-20Cr-5Al-Ce) devel- In the Fecralloy alloy, the rare earth element
oped by Allegheny Ludlum (Ref 39), and yttrium is added to increase the adhesion of the
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:45PM Plate # 0 pg 28

28 / High-Temperature Corrosion and Materials Applications

Al2O3 scale. Yttrium and reactive elements such 3.4.8 Fe-Ni-Cr Alloys (20–25Cr/30–40Ni
as zirconium have been found to increase the Alloys)
adhesion of the Al2O3 scale formed on Fe-Cr-Al
As the nickel content in the Fe-Ni-Cr system
alloys, thus improving the oxidation resistance
increases from austenitic stainless steels to a
of the alloy (Ref 41–44). An Fe-Cr-Al-Y alloy
group of iron-base alloys with 20–25Cr and
can be strengthened by the oxide-dispersion
30–40Ni, the alloys become more stable in terms
strengthening (ODS) mechanism to significantly
of metallurgical structure and more resistant to
increase its elevated-temperature strengths by the
creep deformation (i.e., higher creep-rupture
mechanical alloying process (Ref 45). One such
strengths). In general, this group of alloys also
commercial ODS alloy is MA956 (Fe-20Cr-
exhibits better oxidation resistance. Some of
4.5Al-0.5Y2O3). Oxidation resistance of some
ODS alloys is discussed in Section 3.4.10 on the wrought alloys in this group are 800H/800HT
(Fe-21Cr-32Ni-Al-Ti), RA330 (Fe-19Cr-35Ni-
ODS alloys as well as in Section 3.4.12 on oxi-
1.2Si), HR120 (Fe-25Cr-37Ni-0.7Nb-N), AC66
dation in high-velocity combustion gas streams.
(Fe-27Cr-32Ni-0.8Nb-Ce), 353MA (Fe-25Cr-
35Ni-1.5Si-Ce), and 803 (Fe-26Cr-35Ni-Al-Ti).
The oxidation resistance of 353MA compared
Temperature, K
with RA330, 800H, several stainless steels, and
1400 1350 1300 1250
nickel-base alloys is shown in Fig. 3.26 to 3.28.
Figure 3.30 shows the comparison of 353MA
–6 FeO
Log k, g2 cm–4s–1

CoO
–8
Table 3.10 Cycles to failure of a thin foil
NiO
–10 (0.050 mm, or 0.002 in., thick) of ALFA IV
Cr2O3 repeatedly resistance heated to the indicated
–12 temperatures for 2 min in still air from room
SiO2
AI2 O temperature
3
Temperature, °C (°F) Cycles to failure
–14
7.0 7.2 7.4 7.6 7.8 8.0 8.2
1090 (2000) 2600
1 × 104, K–1 1150 (2100) 1000
T 1200 (2200) 460
1260 (2300) 200
Fig. 3.29 Parabolic rate constants of several oxides. Source: Source: Ref 39
Ref 2

4
253MA
2
Weight change, mg/cm2

353MA
–2 HK 4M

–4

–6

–8
Inco 803
HP45-Nb
–10
500 1500 2500 3500
Time, h

Fig. 3.30 Isothermal oxidation tests in air at 1000 °C (1830 °F) for 353MA, alloy 803, HK4M (Fe-25Cr-25Ni-0.3Al-0.4Ti), and
HP45Nb (Fe-24Cr-37Ni-1.4Si-1.2Nb). Source: Ref 46
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:45PM Plate # 0 pg 29

Chapter 3: Oxidation / 29

and 803 in comparison with some casting alloys for a Ni-Cr alloy to form a protective Al2O3
with similar chromium and nickel contents when scale, the addition of less than 4% Al can also
tested in air at 1000 °C (1830 °F) (Ref 46). significantly improve the oxidation resistance of
Additional oxidation data on 353MA, 803, the alloy. Alloy 601 with only about 1.3% Al
HR120, 800H, and RA330 are summarized in shows excellent oxidation resistance, as shown in
Fig. 3.31 to 3.35 (Ref 47–49). Particular Fig. 3.32 to 3.34 and Fig. 3.36 to 3.38. In both
emphasis should be placed on long-term Fig. 3.37 and 3.38, alloy 601GC was included in
oxidation tests (Fig. 3.33 and 3.35), where some the testing to compare with alloy 601. Alloy
alloys are shown to suffer breakaway oxidation 601GC is essentially alloy 601 with small addi-
after some incubation time of relatively little or tions of titanium, zirconium, and nitrogen for
mild weight losses. More discussion on break- grain-size control involving nitrides and carbo-
away oxidation is presented in Section 3.4.13. nitrides. The development of alloy 601GC was
intended to prevent excessive grain coarsening
when exposed to 1100 °C (2010 °F) and higher
3.4.9 Ni-Cr/Co-Cr Superalloys with the formation of nitrides and carbonitrides
In many Ni-Cr alloys, many alloying of zirconium, titanium, and Zr + Ti (Ref 51). In
elements, such as those for solid-solution long-term oxidation testing (for up to close to
strengthening (e.g., Mo, W) and precipitation 500 days) in air at 1000 °C (1830 °F), alloy
strengthening (e.g., Al, Ti, Nb), are added into 601GC was found to be better than alloy 601, as
the alloys to provide strengthening of the alloy shown in Fig. 3.38. The improvement of oxi-
at elevated temperatures. Many of these alloys dation resistance in alloy 601GC over alloy 601
are commonly referred to as “superalloys.” may be the result of finer grain sizes as well as the
The superalloys also include oxide dispersion presence of zirconium, a reactive element that
strengthened (ODS) alloys, which are briefly has been found to be beneficial in improving
discussed in Section 3.4.10. oxidation resistance for Fe-Cr-Al alloys by
Similar to Fe-Cr-Al alloys, aluminum is also several investigators cited previously.
used as an alloying element in Ni-Cr alloys to Even though the presence of about 1.4% Al in
improve the oxidation resistance. Although it alloy 601 is beneficial in improving oxidation
generally requires a minimum of 4% Al in order resistance, the oxide scales formed on alloy 601

50

N08811 R30566

40

N08330
Weight gain, mg/cm2

30

S35315

20

N06333
10 HR-120

0
0 500 1000 1500 2000 2500 3000
Time, h

Fig. 3.31 Cyclic oxidation tests for 353MA (S35315), HR120, RA330 (N08330), 800HT (N08811), RA333 (N06333), and 556
(R30556) in air at 1090 °C (2000 °F) with specimens cycling to room temperature once a week. Source: Ref 47
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:45PM Plate # 0 pg 30

30 / High-Temperature Corrosion and Materials Applications

350

300
N08810

250
Weight gain, mg/cm2

HR-120
200

150

100

N06601 S35315
50 N06045

0
0 500 1000 1500 2000 2500 3000
Time, h

Fig. 3.32 Cyclic oxidation tests for 353MA (S35315), HR120, 800H (N08810), 601 (N06601), and 45TM (N06045) in air at 1150 °C
(2100 °F) with specimens cycling to room temperature once a week. Source: Ref 47

100 100
0 UNS N06601 Incoloy alloy DS
Incoloy alloy DS 0 UNS N06601
–100
Mass change, mg/cm2

Mass change, mg/cm2

Incoloy alloy 803 –100 UNS N06600

–200 Incoloy alloy 803
–300 UNS N06600 –200 UNS N08330

–400
–300
–500 UNS N08330
– 400
–600
UNS N08810 – 500
–700
UNS N08810
–800 – 600
0 5000 10,000 15,000 20,000 0 1000 2000 3000 4000 5000 6000
Exposure time, h Exposure time, h

Fig. 3.33 Oxidation tests for 803, RA330 (N08330), 800H

Fig. 3.34 Oxidation tests for 803, RA330 (N08330), 800H
(N08810), alloy DS, and 600 (N06600) and 601 (N08810), and alloy DS in compared with Ni-Cr
(N06601) in air + 5% H2O at 1000 °C (1830 °F). Source: Ref 48 alloys 600 (N06600) and 601 (N06601) in air + 5% H2O at
1100 °C (2010 °F). Source: Ref 48
are primarily chromium-rich oxides. Chromium-
rich oxide scales are prone to scaling, cracking,
and spalling, particularly under cyclic conditions temperature every 16 h, with 2 h of heating and
at very high temperatures, such as 1100 °C 6 h of cooling. For 1200 °C testing, specimens
(2010 °F) and higher. This is illustrated in were removed from the hot zone every 16 h. In
Fig. 3.39 for alloy 601 tested in air for 1056 h at addition to scaling and spalling of external oxide
850 °C (1560 °F), 1000 °C (1830 °F), 1100 °C scales when exposed to 1100 and 1200 °C (2010
(2010 °F), and 1200 °C (2190 °F) (Ref. 52). For and 2190 °F), significant internal oxide pene-
testing at 850, 1000, and 1100 °C, specimens tration was also observed. However, with slightly
were cycled to room temperature from the test higher aluminum content (i.e., about 2%), the
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:45PM Plate # 0 pg 31

Chapter 3: Oxidation / 31

HR120
–50.0
2
Weight change, mg/cm

–100.0

RA85H

–150.0

800HT

–200.0

0 90 180 270 360 450 540 630 720

Time, days

Fig. 3.35 Oxidation tests for HR120, 800H, and RA85H in still air at 980 °C (1800 °F) for times up to 720 days with specimens being
cooled to room temperature every 30 days for weight measurement. Source: Ref 49

alloy can significantly reduce its scaling and Fig. 3.42, where alloy 214 (Ni-16Cr-4.5Al-Y)
spalling of external oxide scales as well as was compared with alloy 601 and alloy 800H in
internal oxide penetration. This is illustrated in cyclic oxidation tests performed in still air at
Fig. 3.40, which shows the oxide scales formed 1150 °C (2100 °F) with specimens cycling to
on alloy 602CA (Ni-25Cr-9Fe-2.2Al-0.2C- room temperature once a day except weekends
0.06Zr-0.08Y) after the same test conditions as (Ref 54). Alloy 214 showed essentially no
alloy 601 as shown in Fig. 3.39. The improved weight loss after 42 days of testing, while alloy
oxidation resistance of alloy 602CA may also be 601 suffered a linear weight loss. Figure 3.43
attributed to the presence of yttrium and zirco- shows the cross section of the specimen between
nium. The beneficial effects of yttrium and zir- alloy 214 and alloy 601 oxidation tested in air at
conium along with other reactive elements in 1090 °C (2000 °F) for 1008 h. The Al2O3 oxide
aluminum-containing alloys are discussed later scale formed on alloy 214 remains very protec-
in the chapter. The cyclic oxidation resistance tive even when the temperature is increased to
of alloy 602CA compared with alloy 800H close to its incipient melting point, as shown in
and alloy 601 as well as several nickel- and Fig. 3.44 and 3.45 (Ref 55). Figure 3.44 shows
cobalt-base superalloys at temperatures from 750 the alloy 214 coupon after exposure to flowing
to 1200 °C (1380 to 2190 °F) are summarized in air at 1320 °C (2400 °F) for 200 h with the
Table 3.11 (Ref 53). With further increase in specimen being cycled to room temperature
aluminum to 2.8% in the same alloy composition every 24 h. The incipient melting temperature
series, alloy 603GT (Ni-25Cr-9Fe-2.8Al-0.22C- for alloy 214 is believed to be slightly higher
0.1Zr-0.1Y) exhibits more compact, adherent than 1345 °C (2450 °F) (Ref 56). The specimen
oxide scales (Fig. 3.41) under the same test showed no evidence of oxidation attack
conditions as Fig. 3.40 (Ref 52). (Fig. 3.44). SEM/EDX examination of the cross
Ni-Cr alloys containing about 4% Al or higher section of the tested specimen showed a thin
form a very protective Al2O3 scale when heated aluminum-rich oxide scale formed on the metal
to very high temperatures. This is illustrated in surface (Fig. 3.45).
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:45PM Plate # 0 pg 32

32 / High-Temperature Corrosion and Materials Applications

As shown in Fig. 3.29, in addition to Cr2O3 Birks and Pettit (Ref 57) suggested that both
and Al2O3, SiO2 also exhibits very low parabolic Al2O3 and SiO2 are capable of providing ade-
rate constants and can provide an effective bar- quate oxidation resistance above 1200 °C
rier to oxidation attack at very high temperatures. (2200 °F). There are a number of commercial
high-temperature alloys that contain relatively
high levels of silicon for improving high-
temperature corrosion resistance. However, the
amount of silicon addition to an alloy that can be
manufactured into a wrought product or casting
product is quite limited. Silicon can make an
alloy difficult to cast and also very difficult to
weld. For some commercial, wrought alloys,
silicon levels are up to 3.5%. The formation of a
continuous, external SiO2 scale was not observed
for these commercial Fe-Ni-Cr alloys or Ni-Cr
alloys in a way that a continuous, external Al2O3
scale forms on Fe-Cr-Al alloys or Ni-Cr-Al
alloys. A Ni-Cr-Co-Si alloy, HR160, was
developed in late 1980s for applications in severe
sulfidizing environments (Ref 58). The oxidation
resistance of HR160 alloy (Ni-28Cr-30Co-
2.75Si-0.5Ti-0.5Nb) was found to be quite
comparable to alloy 601 and significantly better
than alloy 800HT when tested for time up to
about 1 year in air at 1090 °C (2000 °F), as
shown in Fig. 3.46 (Ref 59). However, in contrast
to Ni-Cr-Al alloy 214 that forms an external
Al2O3 with essentially no internal oxide or void
formation when exposed to very high tempera-
tures, such as 1200 °C (2200 °F), HR160 forms
external Cr2O3/SiO2 oxides scales with internal
Fig. 3.36 Cyclic oxidation resistance of alloy 601 (Ni-23Cr- oxide/void formation (Ref 49). Figure 3.47
14Fe-1.4Al) compared with alloy 600 (Ni-16Cr- shows oxidation data for alloys 214 and HR160
8Fe) and alloy 800 (Fe-20Cr-32Ni-0.4Al-0.4Ti) in air at 1090 °C
(2000 °F) with specimens being in hot zone for 15 min and out of along with several other nickel- and iron-base
hot zone for 5 min. Source: Ref 50 alloys. The oxidation tests were conducted at

100

0
Mass change, mg/cm2

601
–100 601GC

–200
330

–300
RA85H
–400
0 50 100 150 200
Exposure time, days

Fig. 3.37 Oxidation tests in air + 5% H2O at 1100 °C (2010 °F) for up to 200 days for alloys 601 (Ni-23Cr-14Fe-1.4Al), 601GC
(Ni-23Cr-14Fe-1.3Al-0.2Ti-0.2Zr-0.05N), RA85H (Fe-18Cr-14Ni-3.5Si-1Al), and RA330 (Fe-19Cr-35Ni-1.2Si). Specimens
were cooled to room temperature and weighed at the indicated data points. Source: Ref 51
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:45PM Plate # 0 pg 33

Chapter 3: Oxidation / 33

10

0
Mass change, mg/cm2

–10
601GC

–20

–30
330
601
–40

–50
0 100 200 300 400 500 600
Exposure time, days

Fig. 3.38 Oxidation tests in air + 5% H2O at 1000 °C (1830 °F) for times up to close to 500 days for alloys 601 (Ni-23Cr-14Fe-1.4Al),
601GC (Ni-23Cr-14Fe-1.3Al-0.2Ti-0.2Zr-0.05N), and RA330 (Fe-19Cr-35Ni-1.2Si). Specimens were cooled to room
temperature and weighed at the indicated data points. Source: Ref 51

(a) (c)
20 µm

(b) (d)

Fig. 3.39 Alloy 601 tested in air for 1056 h at (a) 850 °C (1560 °F), (b) 1000 °C (1830 °F), (c) 1100 °C (2010 °F), and (d) 1200 °C
(2190 °F). For testing at 850, 1000, and 1100 °C, specimens were cycled to room temperature from the test temperature
every 16 h, with 2 h of heating and 6 h of cooling. For 1200 °C testing, specimens were removed from the hot zone every 16 h.
Magnification bar represents 20 μm for all micrographs. Source: Ref 52. Courtesy of ThyssenKrupp VDM
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:45PM Plate # 0 pg 34

34 / High-Temperature Corrosion and Materials Applications

(a) (c)
20 µm

(b) (d)

Fig. 3.40 Alloy 602CA (Ni-25Cr-9Fe-2.2Al-0.2C-0.06Zr-0.08Y) tested in air for 1056 h at (a) 850 °C (1560 °F), (b) 1000 °C (1830 °F),
(c) 1100 °C (2010 °F), and (d) 1200 °C (2190 °F). For testing at 850, 1000, and 1100 °C, specimens were cycled to room
temperature from the test temperature every 16 h, with 2 h of heating and 6 h of cooling. For 1200 °C testing, specimens were removed
from the hot zone every 16 h. Magnification bar represents 20 μm for all micrographs. Source: Ref. 52. Courtesy of ThyssenKrupp VDM

Table 3.11 Weight change data (mg/m2h) 214. The internal attack observed in alloy HR160
from cyclic oxidation tests in air for 1200 h at was caused by formation of internal oxides and
indicated temperatures voids. Alloy 214, on the other hand, showed
Test temperature essentially no metal loss due to the formation of a
750 °C 850 °C 1000 °C 1100 °C 1200 °C compact external Al2O3 scale.
Alloy (1380 °F) (1560 °F) (1830 °F) (2010 °F) (2190 °F)
Also shown in Fig. 3.47 is a high-silicon iron-
602CA +0.4 +3 +12 +7 −310
X +1 +8 +5 −5 …
base alloy, RA85H (Fe-19Cr-15Ni-3.5Si-1Al),
800H +7 +8 −24 −162 … suffering much more extensive internal attack
625 +1 +6 −100 −1410 … under the same test conditions. The oxidation
601 +1 +10 +7 −24 −820
617 +4 +12 +19 −19 … resistance data for alloy RA85H is also shown in
188 +1 +4 +7 −302 … Table 3.9 and Fig. 3.28 and 3.37. Figure 3.48
Note: Specimens were held at the test temperature for 16 h followed by cooling to shows the oxide morphology of a high-silicon
room temperature. For testing at 750 to 1100 °C, specimens were cycled by Ni-Cr-Fe alloy, 45TM (Ni-27Cr-23Fe-2.7Si),
furnace cooling and furnace heat up (about 1.5 hours heat up). Cycling for
1200 °C testing involved air cooling and inserting the specimen directly to the after oxidation testing in air for 1154 h at
furnace hot zone. Source: Ref 53
850 to 1200 °C (1560 to 2190 °F) (Ref 52).
Extensive internal void formation was observed
1200 °C (2200 °F) in air for almost 1 year. in the specimen tested at 1100 °C (2010 °F).
HR160 showed extensive internal attack with The oxidation data in comparing 45TM with
relatively little metal loss due to external oxida- other high-temperature alloys is shown in
tion, compared with the alumina-former alloy Fig. 3.32.
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:45PM Plate # 0 pg 35

Chapter 3: Oxidation / 35

(a) (c)
20 µm

(b) (d)

Fig. 3.41 Alloy 603GT (Ni-25Cr-9Fe-2.8Al-0.22C-0.13Zr-0.01Y) tested in air for 1056 h at (a) 850 °C (1560 °F), (b) 1000 °C (1830 °F),
(c) 1100 °C (2010 °F), and (d) 1200 °C (2190 °F). For testing at 850, 1000, and 1100 °C, specimens were cycled to room
temperature from the test temperature every 16 h, with 2 h of heating and 6 h of cooling. For 1200 °C testing, specimens were removed
from the hot zone every 16 h. Magnification bar represents 20 μm for all micrographs. Source: Ref. 52. Courtesy of ThyssenKrupp VDM

In addition to maximizing the oxidation take advantage of the precipitation strengthen-

resistance by adjusting the level of chromium ing. Most heat treating procedures are performed
and/or aluminum, or silicon, a majority of in the intermediate temperature range. As a
high-temperature alloys have been developed to result, the applications for these alloys are
attain elevated-temperature strengths by alloying in the intermediate temperature range to prevent
with many other elements. A large number of overaging of the strengthened precipitates.
superalloys have been developed in response to Oxidation of these alloys at intermediate tem-
the demands of gas turbine engines for critical peratures generally does not present a significant
operating conditions involving high stresses issue in terms of the performance.
and high temperatures. In response to the Another important group of wrought
demands for high stresses at intermediate tem- superalloys is generally classified as solid-
peratures, one group of wrought superalloys is solution-strengthened alloys. Solid-solution-
strengthened by precipitation strengthening with strengthening alloying elements typically are
Ni3X precipitates (X represents aluminum, tita- molybdenum and tungsten. These alloys are also
nium, niobium, etc.) along with solid-solution strengthened with carbides. This group of alloys
strengthening using molybdenum or tungsten. is typically used in stationary components with
These alloys include 718, R-41, X-750, Wasp- lower mechanical stresses than rotating parts
alloy, and Nimonic 80A. Some of the applica- such as disks and blades. However, operating
tions for these alloys in gas turbines include temperatures for these alloys are generally
compressors, diffusers, turbine disks and cases, higher. Typical applications in gas turbines are
heat shields, exhaust system, thrust reversers, combustors, transition ducts, and afterburners.
and turbine shroud rings. Most alloys in this The alloys require good combined properties
group are used in the heat treated conditions to of creep and fatigue strengths, fabricability,
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:45PM Plate # 0 pg 36

36 / High-Temperature Corrosion and Materials Applications

weldability, thermal stability, and oxidation For cast nickel-base superalloys, substantial
resistance. Some of these alloys are being used amounts of aluminum and titanium are used to
increasingly in nongas turbine industries. produce a large volume fraction of gamma prime
(γ′) precipitates to further increase strengthening
of the alloy. They also contain large amounts of
+20
refractory elements, such as molybdenum and
214 tungsten for solid-solution strengthening, along
0
with boron, zirconium, carbon, and hafnium for
grain-boundary strengthening. Since these alloys
–20 do not require hot and cold working during
manufacturing, they can be designed to contain
–40 maximum amounts of these alloying elements to
attain the maximum strength requirements.
–60 These cast nickel-base superalloys include
Weight change, mg/cm2

601 IN713C, IN713LC, IN738X, IN100, B-1900,

Rene 80, IN792, MAR-M246, and MAR-M247.
–80
Lacking precipitation-strengthening phases in
cobalt-base alloys, cast cobalt-base superalloys
–100
derive their strengths from solid-solution
strengthening as well as carbide strengthening.
–120 These alloys typically contain high levels of
carbon. Some of these alloys are X-40 (or alloy
–140 31), MAR-M509, WI-52, and MAR-M302.
They are widely used for high-pressure vanes.
–160 800H Another group of superalloys are referred to as
the ODS alloys. These alloys are produced by the
–180
mechanical alloying process and are strength-
ened by oxide dispersoids (Ref 45). MA956 and
MA754 were originally developed for gas tur-
–200
10 20 30 40 50 bine combustors and stator vanes, respectively
Exposure time, days (Ref 60), and MA6000 was developed for
rotor blades (Ref 61). ODS alloys are much
Fig. 3.42 Cyclic oxidation resistance of alloy 214 compared more difficult to fabricate than conventional
to alloys 601 and 800H in still air at 1150 °C
(2100 °F) cycled once a day every day except weekends. wrought alloys. Joining can present a particularly
Source: Ref 54 significant challenge to this group of alloys. ODS

sample sample
surface surface

Fig. 3.43 Cross sections of the specimens for alloy 601 (a) and alloy 214 (b) after oxidation testing in flowing air at 1090 °C (2000 °F) for
1008 h (samples were cycled to room temperature once every week). Samples were cathodically descaled to remove oxide
scale prior to metallographic mounting. Top edge of the micrograph represents the original specimen surface. Source: Ref 55
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:45PM Plate # 0 pg 37

Chapter 3: Oxidation / 37

0 HR160
–100
601

Weight change, mg/cm2

–200
–300
–400
–500
–600
–700
–800 800HT
–900

100 200 300 400

Time, days

Fig. 3.46 Oxidation resistance of HR160 alloy (Ni-28Cr-

30Co-2.75Si-0.5Ti-0.5Nb) compared with alloys
601 and 800HT when tested for up to about 1 year in air at
Fig. 3.44 Alloy 214 specimen tested in air at 1320 °C
1090 °C (2000 °F). Source: Ref 59
(2400 °F), which was about 25 °C below the
incipient melting point of the alloy, for 200 h with the specimen
being cycled to room temperature every 24 h. Source: Ref 55
low chromium content, such as nickel-base alloy
B (Ni-28Mo) and alloy N (Ni-7Cr-16.5Mo) and
Type 409 (Fe-11Cr), can suffer severe oxidation.
It was surprising to find that Type 321 exhibited
poor oxidation resistance compared with Types
304, 316, and 347. The author offered no
explanation in the paper. As is discussed in other
sections, austenitic stainless steels, such as Types
304, 347, and 321, containing a borderline level
of chromium, are very sensitive to the chromium
content for the bulk chemistry as well as the
surface chemistry in the alloy. When the surface
depletion of chromium occurs for a stainless steel
product whose bulk chromium content is at
the low end of the specification, breakaway
oxidation is likely to take place, thus resulting in
Fig. 3.45 Scanning electron micrograph showing the adher- severe oxidation attack.
ent aluminum-rich oxide scale formed on alloy 214
after exposure in flowing air at 1320 °C (2400 °F) for 200 h with
The majority of the oxidation data reported so
the specimen being cycled to room temperature every 24 h. EDX far has been presented in terms of weight changes
analysis was performed at three different locations, marked 1, 2, (mg/cm2) as a function of times or temperatures.
and 3, in the aluminum oxide scale. The results of the analysis
(wt%) are shown in: Area. 1: 69.9% Al, 17.0% Cr, 10.2% Ni, 1.1% However, it is impossible to use the weight
Fe, and 1.8% Zr. Area 2: 97.8% Al, 1.3% Cr, 0.7% Ni, and 0.2% change data (either weight gain or weight loss) to
Zr. Area 3: 98.3% Al, 0.2% Cr, and 1.5% Ni. Source: Ref 55 estimate the life of the component due to oxida-
tion attack. The oxidation data that is of engi-
alloys are briefly described in Section 3.4.10. neering importance is the depth of oxidation
More information about superalloys for gas attack, which includes the depth of metal loss and
turbines is available in Ref 62 to 66. the depth of internal oxidation attack. The total
Long-term oxidation data in air at an inter- depth of the oxidation attack (or the depth of the
mediate temperature were generated by Barrett metal affected) is responsible for the reduction
(Ref 67) for 33 alloys, from ferritic stainless of the load-bearing capability for the component.
steels to superalloys. Tests were performed at A large oxidation database in terms of the depth
815 °C (1500 °F) for 10,000 h with 1000 h of oxidation attack for commercial alloys from
cycles (a total of 10 cycles). The results are stainless steels to high-alloy Fe-Ni-Cr alloys
shown in Fig. 3.49. Alloys, which contain no or (20–25Cr/30–40Ni alloys), and nickel- and
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:45PM Plate # 0 pg 38

38 / High-Temperature Corrosion and Materials Applications

7 >6.35 >6.35 >6.35 >6.35

Average internal attack
Average metal affected, mm
Metal loss
6

0
214 HR160 230 601 RA85H 617 HR120 800HT

Fig. 3.47 Oxidation data in terms of metal loss, resulting from external oxide scales, and internal attack, resulting from internal oxide
and/or void formation, for alumina-former alloy 214 and chromia/silica-former alloy HR160 along with several other nickel-
and iron-base alloys, generated at 1200 °C (2200 °F) in air for 360 days. Source: Ref 49

(a) (c)
20 µm

(b) (d)

Fig. 3.48 A high-silicon Ni-Cr-Fe alloy, 45TM (Ni-27Cr-23Fe-2.7Si), after oxidation testing in air for 1056 h at (a) 850 °C (1560 °F),
(b) 1000 °C (1830 °F), (c) 1100 °C (2010 °F), and (d) 1200 °C (2190 °F). For testing at 850, 1000, and 1100 °C, specimens
were cycled to room temperature from the test temperature every 16 h, with 2 h of heating and 6 h of cooling. For 1200 °C testing,
specimens were removed from the hot zone every 16 h. Magnification bar represents 20 μm for all micrographs. Source: Ref. 52.
Courtesy of ThyssenKrupp VDM

cobalt-base superalloys, is summarized in 1008 h. The specimens were cooled to room

Table 3.12 (Ref 68). Tests were conducted in temperature for visual examination once a week
flowing air (30 cm/min) at 980, 1095, 1150, and (168 h). Specimens from sheet products were
1200 °C (1800, 2000, 2100, and 2200 °F) for surface ground to maintain the uniform surface
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:45PM Plate # 0 pg 39

Chapter 3: Oxidation / 39

Fig. 3.49 Long-term oxidation tests (10,000 h) in air at 815 °C (1500 °F) with 1000 h cycles (a total of 10 cycles to room temperature for
the entire test) for iron-, nickel-, and cobalt-base alloys. Also included is the upper limit of the metal loss for isothermal tests
(i.e., 10,000 h without cycling to room temperature) for same alloys. Source: Ref 67

condition for all test specimens. Some of the (Fe-20Ni-18Co-22Cr-3Mo-2.5W-0.6Ta-0.02La-

observations of the data are summarized: 0.02Zr), are good examples. However, the oxides
of both molybdenum and tungsten (MO3 and
 Types 304 and 316 specimens were con-
WO3) exhibit high vapor pressures at very high
sumed completely at 1095, 1150, and
temperatures, as shown in Fig. 3.3. Multimet
1200 °C (2000, 2100, and 2200 °F).
alloy suffered rapid oxidation attack at 1150 and
 Type 304 was found to be much more resis-
1200 °C (2100 and 2200 °F), with specimens
tant than Type 316 at 980 °C (1800 °F).
completely consumed at both temperatures.
 Type 446 was not as good as Type 310 at all
However, formation of the volatile oxides of
test temperatures.
MO3 and WO3 can be minimized by modifica-
 Type 446 specimens were consumed at 1150
tion of some key alloying elements in Multimet
and 1205 °C (2100 and 2200 °F).
alloy. The development of alloy 556 was aimed at
 Type 310 appeared to be slightly better than
improving the oxidation resistance of Multimet
RA330 and 800H.
alloy without losing the elevated-temperature
For applications at high temperatures, many strength by making some modification of alloy-
superalloys contain numerous alloying elements ing elements in Multimet alloy. The modification
for increasing the elevated-temperature strength involved a slight increase in chromium, a
of the alloy. Molybdenum and tungsten are decrease in cobalt, replacement of niobium with
common alloying elements for providing solid- tantalum, and addition of a rare-earth element,
solution strengthening for increasing the creep- lanthanum, and a reactive element, zirconium,
rupture strength of the alloy. Two iron-base but the same amounts of molybdenum and
superalloys, Multimet alloy (Fe-20Ni-20Co- tungsten were kept. The result was a much
21Cr-3Mo-2.5W-1.0Nb+Ta) and alloy 556 more oxidation-resistant alloy, alloy 556, at 1095
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:45PM Plate # 0 pg 40

40 / High-Temperature Corrosion and Materials Applications

and 1150 °C (2000 and 2100 °F), although the For nickel-base alloys containing high levels
alloy still suffered rapid oxidation at 1200 °C of molybdenum and/or tungsten, it is believed
(2200 °F). that increasing chromium is probably the most
As shown in the Table 3.12 nickel- and cobalt- important factor in suppressing rapid oxidation
base alloys containing molybdenum and/or involving molybdenum and/or tungsten. Some
tungsten were also found to suffer rapid oxida- nickel-base precipitation-strengthened alloys
tion at very high temperatures (i.e., specimens containing high titanium as well as molybdenum
were consumed during the tests). Specimens of that were consumed at 1200 °C (2200 °F) were
nickel-base alloys that were consumed at Waspaloy (4.3% Mo, 3.0 Ti), René 41 (10% Mo,
1200 °C (2200 °F) were alloy S (14% Mo), alloy 3.0 Ti), and alloy 263 (6% Mo, 2.2Ti). Titanium
X (9% Mo, 0.6% W), and alloy 625 (9% Mo, was found to be very active in oxide-scale
3.5% Nb). Some of those nickel-base alloys formation. Figure 3.50 illustrates the oxide
containing molybdenum and/or tungsten that scale formed on alloy 263 after exposure to air
were not consumed at 1200 °C (2200 °F) were for 1 h at 1200 °C (2200 °F), showing mainly
alloy 230 (14% W), alloy 617 (9% Mo), and titanium-rich oxides and Cr-Ti oxides. The
RA333 (3% Mo, 3% W). From these two dif- formation of titanium-rich oxides apparently
ferent sets of oxidation behavior at very high disrupts the Cr2O3 scale. Nagai et al. (Ref 69)
temperatures, one can design nickel-base alloys found that titanium was detrimental to the oxi-
(relying on chromium oxide scales) contain- dation resistance of Ni-20Cr alloy. In the Fe-Cr-
ing molybdenum and tungsten for elevated- Al system, however, the addition of 1% Ti to Fe-
temperature strengthening to resist oxidation 18Cr-6Al was found to improve resistance in
resistance at very high temperatures by adjusting cyclic oxidation in air at 950 °C (1740 °F)
other alloying elements. (Ref 70). It is not clear whether the beneficial

Table 3.12 Results of oxidation tests for various alloys at indicated temperatures in flowing air
(30 cm/min) for 1008 h
980 °C (1800 °F) 1095 °C (2000 °F) 1150 °C (2100 °F) 1205 °C (2200 °F)
Average Average Average Average
metal metal metal metal
Metal loss, affected, Metal loss, affected, Metal loss, affected, Metal loss, affected,
Alloy mm (mils) mm (mils) mm (mils) mm (mils) mm (mils) mm (mils) mm (mils) mm (mils)
214 0.0025 (0.1) 0.005 (0.2) 0.0025 (0.1) 0.0025 (0.1) 0.005 (0.2) 0.0075 (0.3) 0.005 (0.2) 0.018 (0.7)
601 0.013 (0.5) 0.033 (1.3) 0.03 (1.2) 0.067 (2.6) 0.061 (2.4) 0.135 (5.3) 0.11 (4.4) 0.19 (7.5)
600 0.0075 (0.3) 0.023 (0.9) 0.028 (1.1) 0.041 (1.6) 0.043 (1.7) 0.074 (2.9) 0.13 (5.1) 0.21 (8.9)
230 0.0075 (0.3) 0.018 (0.7) 0.013 (0.5) 0.033 (1.3) 0.058 (2.3) 0.086 (3.4) 0.11 (4.5) 0.20 (7.9)
S 0.005 (0.2) 0.013 (0.5) 0.01 (0.4) 0.033 (1.3) 0.025 (1.0) 0.043 (1.7) >0.8I (31.7) >0.8I (31.7)
617 0.0075 (0.3) 0.033 (1.3) 0.015 (0.6) 0.046 (1.8) 0.028 (1.1) 0.086 (3.4) 0.27 (10.6) 0.32 (12.5)
333 0.0075 (0.3) 0.025 (1.0) 0.025 (1.0) 0.058 (2.3) 0.05 (2.0) 0.1 (4.0) 0.18 (7.1) 0.45 (17.7)
X 0.0075 (0.3) 0.023 (0.9) 0.038 (1.5) 0.069 (2.7) 0.11 (4.5) 0.147 (5.8) >0.9 (35.4) >0.9 (35.4)
671 0.0229 (0.9) 0.043 (1.7) 0.038 (1.5) 0.061 (2.4) 0.066 (2.6) 0.099 (3.9) 0.086 (3.4) 0.42 (16.4)
625 0.0075 (0.3) 0.018 (0.7) 0.084 (3.3) 0.12 (4.8) 0.41 (16.0) 0.46 (18.2) >1.2 (47.6) >1.2 (47.6)
Waspa- 0.0152 (0.6) 0.079 (3.1) 0.036 (1.4) 0.14 (5.4) 0.079 (3.1) 0.33 (13.0) >0.40 (15.9) >0.40 (15.9)
loy
R-4I 0.0178 (0.7) 0.122 (4.8) 0.086 (3.4) 0.30 (11.6) 0.21 (8.2) 0.44 (17.4) >0.73 (28.6) >0.73 (28.6)
263 0.0178 (0.7) 0.145 (5.7) 0.089 (3.5) 0.36 (14.2) 0.18 (6.9) 0.41 (16.1) >0.91 (35.7) >0.91 (35.7)
188 0.005 (0.2) 0.015 (0.6) 0.01 (0.4) 0.033 (1.3) 0.18 (7.2) 0.2 (8.0) >0.55 (21.7) >0.55 (21.7)
25 0.01 (0.4) 0.018 (0.7) 0.23 (9.2) 0.26 (10.2) 0.43 (16.8) 0.49 (19.2) >0.96 (37.9) >0.96 (37.9)
150 0.01 (0.4) 0.025 (1.0) 0.058 (2.3) 0.097 (3.8) >0.68 (26.8) >0.68 (26.8) >1.I7 (46.1) >1.I7 (46.1)
6B 0.01 (0.4) 0.025 (1.0) 0.35 (13.7) 0.39 (15.2) >0.94 (36.9) >0.94 (36.9) >0.94 (36.8) >0.94 (36.8)
556 0.01 (0.4) 0.028 (1.1) 0.025 (1.0) 0.067 (2.6) 0.24 (9.3) 0.29 (11.6) >3.8 (150.0) >3.8 (150.0)
Multi- 0.01 (0.4) 0.033 (1.3) 0.226 (8.9) 0.29 (11.6) >1.2 (47.2) >1.2 (47.2) >3.7 (146.4) >3.7 (146.4)
met
800H 0.023 (0.9) 0.046 (1.8) 0.14 (5.4) 0.19 (7.4) 0.19 (7.5) 0.23 (8.9) 0.29 (11.3) 0.35 (13.6)
RA330 0.01 (0.4) 0.11 (4.3) 0.02 (0.8) 0.17 (6.7) 0.041 (1.6) 0.22 (8.7) 0.096 (3.8) 0.21 (8.3)
310 0.01 (0.4) 0.028 (1.1) 0.025 (1.0) 0.058 (2.3) 0.075 (3.0) 0.11 (4.4) 0.2 (8.0) 0.26 (10.3)
316 0.315 (12.4) 0.36 (14.3) >1.7 (68.4) >1.7 (68.4) >2.7 (105.0) >2.7 (105.0) >3.57 (140.4) >3.57 (140.4)
304 0.14 (5.5) 0.21 (8.1) >0.69 (27.1) >0.69 (27.1) >0.6 (23.6) >0.6 (23.6) >1.7 (68.0) >1.73 (68.0)
446 0.033 (1.3) 0.058 (2.3) 0.33 (13.1) 0.37 (14.5) >0.55 (21.7) >0.55 (21.7) >0.59 (23.3) >0.59 (23.3)
Note: 3304 cm3/min of flow rate in a 1.75 in. diam furnace tube. The moisture was removed from the air by a filter prior to entering into the furnace tube. Specimens were
cathodically descaled for measurement of the metal loss. The average metal affected is the sum of the metal loss and the depth of internal attack. The depth of internal attack
was measured by metallography. Source: Ref 68
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:45PM Plate # 0 pg 41

Chapter 3: Oxidation / 41

effect of titanium for oxidation resistance is only 1205 °C (2200 °F). The oxidation resistance of
for alumina formers such as in this case (Fe- alloy 214 is also presented in Fig. 3.42 to 3.45
18Cr-6Al), but not for chromia formers in Ni- and 3.47.
20Cr alloy. Niobium is another alloying element The oxidation data presented in Table 3.12
that may be detrimental to alloy oxidation resis- were generated with a weekly cycle (168 h).
tance at very high temperatures. The relatively When the cyclic frequency was increased to 25 h
poor oxidation resistance of alloy 625 at 1095 cycles, oxidation rates were increased for all the
and 1150 °C (2000 and 2100 °F) can be attrib- alloys tested. However, some alloys are more
uted to niobium. sensitive to cyclic oxidation than others. The
Cobalt-base alloys with tungsten, such as effect of thermal cycling on the oxidation resis-
alloy 188 (Co-22Cr-22Ni-14W-0.04La), alloy 25 tance of various alloys in air at 1095 °C
(Co-20Cr-10Ni-15W), and alloy 6B (Co-30Cr- (2000 °F) is illustrated in Table 3.13, which
4.5W-1.2C), suffered rapid oxidation at 1205 °C compares once-a-week (168 h) cycle data (Ref
(2200 °F). A cobalt-base alloy, alloy 150 (Co- 68) with 25 h cycle data (Ref 71). All the data are
27Cr-18Fe), containing no tungsten also suffered presented in terms of the average depth of metal
rapid oxidation attack at 1205 °C (2200 °F). affected, which represents the metal loss plus the
Again, the oxidation of a cobalt-base alloy can be depth of internal oxidation attack. Both sets of
significantly improved with some modification the data were generated under the same test
of alloying elements. Alloy 25 with 15% W conditions using the same test furnaces and test
exhibits excellent creep-rupture strengths at high procedures except the differences in cyclic fre-
temperatures. However, because of the high level quencies. Some chromia formers, such as alloys
of tungsten, the alloy suffers high oxidation rates 230, S, 188, 556, and 310, showed good resis-
at very high temperatures, such as 1095 and tance to thermal cycling.
1150 °C (2000 and 2100 °F). With slight A long-term oxidation test program was
increase in chromium and nickel along with the undertaken to test alloys up to 2 years at 980,
addition of lanthanum, the result of the mod- 1095, and 1150 °C (1800, 2000, and 2100 °F)
ification was alloy 188. As shown in Table 3.12,
alloy 188 exhibits significantly better oxidation
resistance than alloy 25 at 1095 and 1150 °C Table 3.13 Comparative oxidation resistance of
(2000 and 2100 °F). various alloys in flowing air between 168 h and
The best alloy among those investigated was 25 h cycles at 1095 °C (2000 °F)
an alumina former, alloy 214 (Ni-16Cr-4.5Al-Y). Total depth of attack, Extrapolated oxidation
The depth of oxidation attack (average metal mm (mils) rate, mm/yr (mpy)

affected) was found to be less than 0.025 mm Alloy 1008 h/168 h 1050 h/25 h 168 h cycles 25 h cycles

(1.0 mils) after 1008 h at temperatures up to 214 0.003 (0.1) 0.025 (1.0) 0.025 (1) 0.20 (8)
601 0.066 (2.6) 0.297 (11.7) 0.58 (23) 2.49 (98)
600 0.041 (1.6) 0.185 (7.3) 0.36 (14) 1.55 (61)
671 0.061 (2.4) 0.584 (23.0) 0.53 (21) 4.88 (192)
230 0.003 (1.3) 0.086 (3.4) 0.28 (11) 0.71 (28)
S 0.003 (1.3) 0.061 (2.4) 0.28 (11) 0.51 (20)
G-30 0.122 (4.8) 0.203 (8.0) 1.07 (42) 1.70 (67)
617 0.046 (1.8) 0.267 (10.5) 0.41 (16) 2.24 (88)
RA333 0.058 (2.3) 0.130 (5.1) 0.51 (20) 1.09 (43)
625 0.122 (4.8) 0.414 (16.3) 1.07 (42) 3.45 (136)
Waspaloy 0.137 (5.4) 0.414 (16.3) 1.19 (47) 3.45 (136)
263 0.361 (14.2) 0.478 (18.8) 3.12 (123) 3.99 (157)
188 0.003 (1.3) 0.058 (2.3) 0.28 (11) 0.48 (19)
25 0.259 (10.2) 0.490 (19.3) 2.26 (89) 4.09 (161)
150 0.097 (3.8) 0.353 (13.9) 0.84 (33) 2.95 (116)
6B 0.394(15.5) >0.800 (31.5) 3.43 (135) >6.68 (263)
556 0.066 (2.6) 0.117 (4.6) 0.58 (23) 0.97 (38)
Multimet 0.295 (11.6) 0.381 (15.0) 2.57 (101) 3.18 (125)
800H 0.188 (7.4) 0.406 (16.0) 1.63 (64) 3.40 (134)
RA330 0.170 (6.7) 0.442 (17.4) 1.47 (58) 3.68 (145)
310 0.058 (2.3) 0.112 (4.4) 0.51 (20) 0.94 (37)
446 0.368 (14.5) 0.655 (25.8) 3.20 (126) 5.46 (215)
Note: 3304 cm3/min of flow rate in a 1.75 in. diam furnace tube. The moisture was
removed from the air by a filter prior to entering into the furnace tube. Specimens
Fig. 3.50 Scanning electron micrograph showing the early were cathodically descaled for measurement of the metal loss. The total depth of
stage of oxidation in air at 1200 °C (2200 °F) for attack is the sum of the metal loss and the depth of internal attack. Metal loss was
1 hour for alloy 263, revealing titanium-rich and Cr-Ti oxides on measured by cathodically descaling the oxide scale prior to measurement of the
the outermost oxide scale. Area 1: 28.1% Cr, 70.9% Ti, 0.8% Co, specimen thickness. Source: Ref 71
0.2% Ni. Area 2: 57.0% Cr, 36.8% Ti, 2.8% Co, 25% Ni, 1.0% Fe.
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:45PM Plate # 0 pg 42

42 / High-Temperature Corrosion and Materials Applications

(Ref 49). Test specimens, which were cut from thinner), no “cracking” noises were heard during
plate products, had dimensions of 1.27 cm specimen cooling from the hot zone, and oxide
(thick) by 2.54 cm (width) by 2.54 cm (length). scales remained on the specimen surface for
Tests were conducted in a box furnace with still these thin test coupons. Figure 3.51 shows
air; specimens were removed from the furnace weight change data for alloy 800H in this long-
every 30 days to allow air cooling. Weights were term testing at 1095 °C (2000 °F) using thick,
then measured to determine the time to initiation blocky specimens, showing breakaway oxida-
of breakaway oxidation. For block specimens tion with linear weight loss after 60 days of
(e.g., 1.25 cm, or 0.5 in., thick specimens), oxide exposure. Some alloys showed no breakaway
scales were cracking, breaking, and spalling with oxidation even after 2 years of testing. Figure
“bing” noises as soon as the specimens were 3.35 shows no breakaway oxidation for HR120
removed from the hot zone in the furnace. after 2 years of testing at 980 °C (1800 °F),
Cracking noises would not stop until almost all while 800H and RA85H suffered breakaway
the oxide scales were broken and spalled off. oxidation. HR160 and 601 showed no break-
Thus, no cathodic descaling was necessary to away oxidation after one year of testing at
remove the oxide scales for measurement of the 1095 °C (2000 °F), while alloy 800H suffered
specimen thickness at the end of the test. When breakaway oxidation (Fig. 3.46). At the end of
similar oxidation testing was performed with testing for 2 years (720 days) at 980 °C
thin test coupons (about 3.2 mm, (0.125 in.) or (1800 °F) and 1 year (360 days) at 1093, 1150,
and 1200 °C (2000, 2100, and 2200 °F), speci-
mens were cut, mounted, and polished for
200
metallographic determination of the depth of
internal attack. Metal loss was determined by
0
subtracting the original specimen thickness from
the thickness after testing. The data generated
Weight gain, mg/cm2

–200
from blocky specimens are summarized in
Tables 3.14 to 3.16. The annual oxidation rates in
–400 terms of the total depth of oxidation attack are
included in Tables 3.14 to 3.16 at 980, 1090, and
–600 1150 °F (1800, 2000, and 2100 °F), respec-
tively. These oxidation rate values would be
–800 considered to be quite reasonable, since the test
duration was almost 2 years for 980 °C
–1000 (1800 °F) and about 1 year for 1090 and 1150 °C
0 30 60 90 120 150 180 210 240 270 300 330 360
testing. At 980 °C (1800 °F), alloys 230, 617,
Days HR120, 556, and HR160 exhibited oxidation
rates of less than 10 mpy. At 1090 °C (2000 °F),
Fig. 3.51 The oxidation behavior of alloy 800H tested in still
only alloy 230 exhibited about 10 mpy of oxi-
air at 1095 °C (2000 °F) involving a thick, blocky
specimen (1.25 cm, or 0.5 in., thick) cycling to room temperature dation rate, while other alloys tested exhibited
every 30 days for weight measurement, showing the alloy was
under protective scales initially for about 30 days and then suf- more than 20 mpy. At 1150 °C (2100 °F), all
fered breakaway oxidation. Courtesy of Haynes International, Inc. alloys tested exhibited more than 30 mpy of

Table 3.14 Oxidation of several high temperature alloys in still air at 980 °C (1800 °F) for 720 days
with specimens cycling to room temperature every 30 days
Weight change, Metal loss, Total depth of Oxidation rate,
Alloy mg/cm2 mm (mils) attack, mm (mils) mm (mpy)
230 −1.4 0.00254 (0.1) 0.14732 (5.8) 0.0508 (2)
617 1.0 0 0.23876 (9.4) 0.127 (5)
HR120 −33.7 0.04060 (1.6) 0.30988 (12.2) 0.1524 (6)
556 −19.8 0.02286 (0.9) 0.38608 (15.2) 0.2032 (8)
HR160 −51.2 0.0635 (2.5) 0.42418 (16.7) 0.2286 (9)
601 −9.9 0.0127 (0.5) 0.56896 (22.4) 0.2794 (11)
RA85H −122.2 0.16002 (6.3) 1.36398 (53.7) 0.6858 (27)
800HT −417.8 0.52578 (20.7) 2.02692 (79.8) 1.0414 (41)
Note: Total depth of attack = metal loss + internal attack. Source: Ref 49
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:45PM Plate # 0 pg 43

Chapter 3: Oxidation / 43

Table 3.15 Oxidation of several high-temperature alloys in still air at 1095 °C (2000 °F) for
360 days with specimens cycling to room temperature every 30 days
Alloy Weight change, mg/cm2 Metal loss, mm (mils) Total depth of attack, mm (mils) Oxidation rate, mm/yr (mpy)
230 −42.1 0.04826 (1.9) 0.27178 (10.7) 0.27940 (11)
556 −298.1 0.36322 (14.3) 0.53848 (21.2) 0.55880 (22)
RA330 −405.0 0.50800 (20.0) 0.60706 (23.9) 0.60960 (24)
HR160 −73.1 0.09144 (3.6) 0.73660 (29.0) 0.73660 (29)
HR120 −665.7 0.82804 (32.6) 0.96520 (38.0) 0.99060 (39)
601 −110.5 0.13716 (5.4) 1.14554 (45.1) 1.16840 (46)
800HT −893.9 1.12522 (44.3) 1.29540 (51.0) 1.32080 (52)
RA85H −348.1 0.45466 (17.9) 2.03962 (80.3) 2.05740 (81)
Note: Total depth of attack = metal loss + internal attack. Source: Ref 49

Table 3.16 Oxidation of several high-temperature alloys in still air at 1150 °C (2100 °F) for
360 days with specimens cycling to room temperature every 30 days
Alloy Weight change, mg/cm2 Metal loss, mm (mils) Total depth of attack, mm (mils) Oxidation rate, mm/yr (mpy)
230 −249.7 0.28194 (11.1) 0.83680 (34.0) 0.88900 (35)
617 −452.7 0.54102 (21.3) 0.94488 (37.2) 0.96520 (38)
HR120 −894.2 1.10998 (43.7) 1.34620 (53.0) 1.37160 (54)
HR160 −155.5 0.19304 (7.6) 1.49098 (58.7) 1.52400 (60)
800H −1315.6 1.65608 (65.2) 1.78562 (70.3) 1.80340 (71)
601 −258.7 0.32004 (12.6) 1.84912 (72.8) 1.87960 (74)
RA85H −389.2 0.50800 (20.0) 2.40792 (94.8) 2.4384 (96)
Note: Total depth of attack = metal loss + internal attack. Source: Ref 49

oxidation rate. At 1200 °C (2200 °F), alumina- RA85H. These alloys are primarily chromia
former alloy 214 showed little or no oxidation formers. This correlation may be useful in mak-
attack (Fig. 3.47). ing rough estimates of the metal loss for an alloy
These data are valuable in providing readers that showed only weight-loss data.
with the oxidation data in terms of the total depth The depth of oxidation attack was also inves-
of oxidation attack in air based on the actual tigated by John (Ref 15) for a wide variety of
measurements of the specimens after 1 to 2 years commercial alloys in isothermal air oxidation
of testing. Since the data were generated from testing. Table 3.17 summarizes his data in terms
thick, blocky specimens, caution should be used of the temperature at which the oxidation rate
when the data are being considered for applica- reaches 10 mpy. Figure 3.53 illustrates some
tion in thin-gage sheets or foils. This is related to oxidation data in terms of oxide penetration as a
the reservoir effect of a solute alloying element function of test temperature in air after 1 year for
for the formation of a protective oxide scale. some alloys (Ref 15). Table 3.18 shows the depth
The discussion of the reservoir issue and the of oxidation attack of various heat-resistant
oxidation in thin foils is presented later. alloys after cyclic oxidation tests at 1100 °C
In this test program (Ref 49), in addition to the (2010 °F) in air + 5% H2O for 504 h with spe-
metal loss caused by formation of external oxide cimens cycling out of the furnace every 15 min
scales and internal attack caused by formation of (Ref 72).
internal oxides and/or voids, the weight-loss Lai et al. (Ref 73) reported the oxidation data
values were also determined. The weight loss of generated from a field test inside a radiant tube
the specimens is related mostly to the metal loss fired with natural gas with an average tempera-
resulting from the removal of the external oxide ture of 1010 °C (1850 °F) (Table 3.19). The test
scales and is not significantly affected by the rack containing coupons of various alloys was
formation of internal oxides and/or voids. The exposed for about 3000 h. Many chromia for-
weight-loss values of the alloys tested are plotted mers, such as alloys 601, 230, 556, 310, 600, and
against their corresponding metal-loss values RA330, were found to perform well, with
at 980, 1090, and 1150 °C (1800, 2000, and extrapolated oxidation rates of less than 0.5 mm/
2100 °F), revealing a nice straight line correla- yr (20 mpy). Type 304, however, suffered severe
tion (Fig. 3.52). Alloys tested were 230, 617, oxidation attack with an extrapolated oxidation
601, 556, HR160, HR120 RA330, 800HT, and rate of more than 4.4 mm/yr (>175 mpy). The
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:45PM Plate # 0 pg 44

44 / High-Temperature Corrosion and Materials Applications

70

60 1.50

50
Metal loss, mils

Metal loss, mm
40 1.00

30

20 0.50

10

0
0 200 400 600 800 1000 1200 1400
Weight loss, mg/cm2

Fig. 3.52 Correlation between weight loss (mg/cm2) and the depth of metal loss (mils) for commercial alloys that are primarily chromia
formers tested in air at 980 °C (1800 °F)/720 days, 1095 °C (2000 °F)/360 days, and 1150 °C (2100 °F)/360 days. Alloys
tested were 230, 617, 601, 556, HR160, HR120, RA330, 800HT, and RA85H. 1.0 mil = 0.0254 mm

Table 3.17 The oxidation rate of total depth of 102 2.5

attack was reached after 1 year in air 9Cr -1Mo
Carbon steel
Maximum temperature for
Alloy UNS No. 0.25 mm/yr (10 mpy), °C (°F) AISI 410
Carbon steel … 604 (1120)
Copper C11000 677 (1250) AISI 304
Nickel N02270 782 (1440)
9Cr-1Mo S50400 799 (1470)
10 0.3
410 S41000 832 (l530)
Penetration, mils

Penetration, mils
304 S30400 893 (1640) Alloy 617
617 N06617 938 (1720) Nickel
803 … 954 (1750)
625 N06625 960 (1760)
800H N08810 966 (1770) AISI 310
601GC … 977 (1790)
Alloy 800 H
DS … 977 (1790)
230 N06230 982 (1800) 1
310 S31000 982 (l800)
RA330 S33000 999 (1830) 0.03
446 S44600 1010 (1850)
556 R30556 1010 (1850)
HR120 … 1010 (1850)
253MA S30815 1082 (1980)
602CA … 1121 (2050)
MA956 S67956 >1150 (>2100) 0.1
214 N07214 >1150 (>2100) 1000 1200 1400 1600 1800 2000
Source: Ref 15 Temperature, °F

Fig. 3.53 Oxidation penetration (metal loss + internal attack)

as a function of test temperature for 1 year in air for
alumina former, alloy 214, showed little or no a variety of commercial alloys. Source: Ref 15
oxidation attack with an extrapolated oxidation
rate of about 0.076 mm/yr (3 mpy). In another
field test (Ref 74), a test rack containing coupons (1000 °F) during furnace idling. The results are
of various alloys was placed in a natural-gas-fired summarized in Table 3.20. All the chromia for-
furnace used for reheating ingots and slabs of mers tested suffered severe oxidation attack. The
nickel- and cobalt-base alloys. The test was alumina former (alloy 214), however, exhibited
conducted for about 113 days at temperatures little attack. Examination of the oxide scale
varying from 1090 to 1230 °C (2000 to formed on alloy 214 was found to consist of
2250 °F), with frequent cycles to 540 °C essentially aluminum-rich oxides.
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:45PM Plate # 0 pg 45

Chapter 3: Oxidation / 45

Table 3.18 Cyclic oxidation resistance of various heat-resistant alloys at 1100 °C (2010 °F) for 504 h
in air-5H2O
Specific weight change (descaled), mg/cm2
Metal loss, Maximum attack,
Alloy Mean Range mm (mils) mm (mils)
ACI grade HK −105.8 −98 to −124 0.25 (9.9) 0.35 (13.8)
310SS −149.0 −92 to −235 0.31 (12.2) 0.38 (15.0)
800 −168.6 −83 to −223 0.39 (15.4) 0.59 (23.2)
601 −11.0 −6.3 to −17.2 <0.02 (0.8) 0.12 (4.7)
617 −13.5 −6.5 to −17.5 … …
X −20.0 −10.0 to −29.5 0.05 (2.0) 0.25 (9.9)
RA333 −30.5 … … …
IN-814 −3.5 −2.5 to −4.9 0 <0.01 (0.4)
188 −25.0 −13.0 to −40.5 0.03 (1.2) 0.15 (5.9)
MA-956 −1.0 −0.3 to −1.5 0.01 (0.4) 0.02 (0.8)
Note: 15 min in furnace and 5 min out of furnace. Source: Ref 72

Table 3.19 Results of field test in a Table 3.20 Results of field test in a
natural-gas-fired radiant tube at 1010 °C natural-gas-fired furnace for reheating nickel-
(1850 °F) for 3000 h and cobalt-base alloy ingots and slabs for
Metal loss, Maximum metal Oxidation rate, 113 days at 1090 to 1230 °C (2000 to 2250 °F)
Alloy mm (mils) affected(a), mm (mils) mm/yr (mpy) with frequent cycles to 540 °C (1000 °F)
214 0.003 (0.1) 0.025 (1) 0.076 (3) Maximum metal
601 0.023 (0.9) 0.076 (3) 0.23 (9) Metal loss, affected(a),
230 0.028 (1.1) 0.10 (4) 0.30 (12) Alloy mm (mils) mm (mils)
556 0.018 (0.7) 0.10 (4) 0.30 (12) 214 0.013 (0.5) 0.11 (4.5)
310 0.041 (1.6) 0.10 (4) 0.30 (12) RA330 0.39 (15.5) 0.65 (25.5)
600 0.018 (0.7) 0.15 (6) 0.46 (18) 601 0.18 (7.2) 0.95 (37.2)
RA330 0.048 (1.9) 0.15 (6) 0.46 (18) 600 0.64 (25.0) 1.1 (45.0)
800H 0.12 (4.7) 0.30 (12) 0.89 (35) 800H >0.79 (31.0)(b) >0.79 (31.0)(b)
309 0.50 (19.7) 0.50 (20) 1.5 (58) 310SS >1.0 (41.0)(b) >1.0 (41.0)(b)
304 >1.5 (60)(b) >1.5 (60)(b) >4.4 (175) 304SS >1.5 (60.0)(b) >1.5 (60.0)(b)
(a) Metal loss + maximum internal penetration. (b) Sample was consumed. Source: 316SS >1.6 (63.0)(b) >1.6 (63.0)(b)
Ref 73 446SS >0.61 (24.0)(b) >0.61 (24.0)(b)
(a) Metal loss + internal penetration. (b) Samples were consumed. Source: Ref 74

3.4.10 Oxide-Dispersion-Strengthened
(ODS) Alloys annealing to produce a textured microstructure
Oxide-Dispersion Strengthened alloys use (Ref 75). Alloys are available in mill products
very fine oxide particles that are uniformly dis- such as bar, plate, sheet, and so forth, or custom
tributed throughout the matrix to provide exces- forgings. Some ODS alloys are shown in
sive strengthening at very high temperatures. Table 3.21 (Ref 75). The oxidation behavior
These oxide particles, typically yttrium oxide, do of some of these ODS alloys tested in air con-
not react with the alloy matrix so no coarsening taining 5% H2O at 1200 °C (2190 °F) is shown
or dissolution occurs during the exposure to in Fig. 3.54 (Ref 75). The oxidation behavior
very high temperatures, thus maintaining the of MA956 compared with those of several iron-
strengthening of the alloy. This group of super- and nickel-base alloys at 1100 °C (2010 °F)
alloys is produced using specialty powders that is shown in Fig. 3.55 (Ref 76). Additional oxi-
are manufactured by the mechanical alloying dation data for some ODS alloys is presented in
process. These powders are essentially compo- Section 3.4.12.
site powders with each particle containing a
uniform distribution of submicron oxide particles
3.4.11 Effect of Oxygen Concentration
in an alloy matrix. The process of producing
these ODS powders involves repeated fracturing on Oxidation
and rewelding of a mixture of powder particles in Air atmosphere consists primarily of oxygen
vertical attritors or horizontal ball mills (Ref 75). and nitrogen with some water vapor and small
Alloy powders are then canned, degassed, and amounts of inert gases, such as argon, neon, and
hot extruded, followed by hot working and helium. Dry air consists of essentially 21% O2
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:45PM Plate # 0 pg 46

46 / High-Temperature Corrosion and Materials Applications

Table 3.21 Nominal chemical compositions of several ODS alloys

Composition, wt%
Alloy Ni Fe Cr Al Ti W Mo Ta Y2O3 B Zr
MA754 bal … 20 0.3 0.5 … … … 0.6 … …
MA758 bal … 30 0.3 0.5 … … … 0.6 … …
MA760 bal … 20 6.0 … 3.5 2.0 … 0.95 0.01 0.15
MA6000 bal … 15 4.5 2.5 4.0 2.0 2.0 1.1 0.01 0.15
MA956 … bal 20 4.5 0.5 … … … 0.5 … …
Note: All alloys contain 0.05% C. Source: Ref 75

50
MA 956 0.5
0 0 Mass change, Ib/in.2 × 10–3
Mass change, mg/cm2

MA 760 –0.5
–50 MA 754
–1.0
–100 –1.5

–150 –2.0
MA 6000 –2.5
–200
–3.0
–250 –3.5
0 10 20 30 40 50 60
Exposure time, days

Fig. 3.54 Oxidation behavior of several ODS alloys in

air containing 5% H2O at 1200 °C (2190 °F).
Source: Ref 75
Fig. 3.55 Cyclic oxidation resistance of ODS alloy MA956
compared with alloy 601, HK alloy, alloy 800, and
Type 310. Source: Ref 76
and 78% N2 with about 1% inert gases. In com-
bustion, the concentration of oxygen may vary.
Also, in some processes, oxygen may be the only concentration included 9Cr-1Mo steel, 410, 304,
gaseous component to which the equipment is and 617, while carbon steel, nickel, 800H, and
exposed. Thus, the effect of oxygen concentra- 310 were relatively unaffected by oxygen con-
tion on the oxidation behavior of alloys may need centrations. In Fig. 3.57, Type 304 was found to
to be evaluated for some applications. John (Ref exhibit approximately 0.25 mm/yr (10 mils) of
15, 77) investigated the oxidation behavior of a attack at 927 °C (1700 °F) after 1 year in 100%
wide variety of commercial alloys in N2-O2 O2, while close to 2.5 mm/yr (100 mpy) of
mixtures with the concentration of oxygen attack was extrapolated based on 1152 h expo-
varying from 1 to 100%. The data generated at sure at 871 °C (1600 °F) in 100% O2, as shown
871 °C (1600 °F) are shown in Fig. 3.56 (Ref in Fig. 3.56. Extrapolation from short-term tests
77), and data generated at 927 °C (1700 °F) are here showing a higher oxidation rate at lower
shown in Fig. 3.57 (Ref 15). The data presented temperature could be an issue here. More long-
in Fig. 3.56 were based on the 1152 h testing, term tests are needed. It is also of practical
while the data in Fig. 3.57 were based on tests interest to perform long-term tests to evaluate the
after 1 year. At 871 °C (1600 °F), Type 304 was oxidation behavior in 100% O2 environments for
the only alloy that showed significant increase in some alloys that are to be used in chemical pro-
oxidation attack from about 0.25 mm/yr cesses involving 100% O2.
(10 mpy) in N2-21%O2 to close to 2.5 mm/yr
(100 mpy) in 100% O2. All other alloys in the
3.4.12 High-Velocity Combustion
figure showed about 10 mpy or less of oxidation
attack at three different levels of O2 concentra- Gas Streams
tions (1%, 21%, and 100%). Figure 3.57 shows Oxidation of alloys can significantly increase
the oxidation behavior of a number of alloys at under high-velocity gas streams. Combustors
927 °C (1700 °F). The alloys that were found to and transition ducts in the gas turbine are subject
increase oxidation attack with increasing oxygen to such conditions. These components are also
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:45PM Plate # 0 pg 47

Chapter 3: Oxidation / 47

103 AISI 304

AISI 310
102
956MA

10 253MA

HR120
Penetration rate, mpy

1
800HT

10–1 AISI 446

617
10–2
230

214
10–3
601GC
10–4
10–2 10–1 1 602CA

Partial pressure of O2, atm 556

Fig. 3.56 Effect of oxygen concentration in the N2-O2 mixture on the oxidation penetration (metal loss + internal attack) at 871 °C
(1600 °F) for 1152 h. 1.0 mil = 0.025 mm. Source: Ref 77

subject to severe thermal cycling, particularly for attack. The alloy showed no sign of breakaway
gas turbines in airplane engines. Laboratory oxidation after 500 h. Figure 3.58 shows the
burner rigs have been developed to evaluate this oxide scale formed after testing for 500 h
type of oxidation, often referred to as “dynamic (1000 cycles) (Ref 82). The scale consisted of
oxidation,” under the condition of very high gas aluminum-rich oxides. After 1000 h (2000
velocities. Some of these dynamic oxidation cycles) of testing, the scale remained aluminum-
burner rigs are described elsewhere (Ref 78–83). rich. The maximum metal affected (metal loss
Lai (Ref 82) investigated a wide range of +maximum internal penetration) remained about
alloys—from stainless steels to superalloys—in a the same after 1000 h compared to after 500 h
burner rig that generated a combustion gas (Ref 82).
stream with 0.3 Mach (100 m/s) velocity. The The test results generated at 980 °C (1800 °F)
specimens were held in a carousel-type holder for 1000 h (2000 cycles) are shown in Table 3.23
rotating at 30 rpm with respect to the combustion (Ref 82). Unlike 1090 °C (2000 °F) testing
gas stream. Every 30 min the carousel was (Table 3.22), testing at 980 °C (1800 °F) resulted
withdrawn from the hot zone, and quenched to in internal oxidation and nitridation in addition
less than 260 °C (500 °F) by a blast of cold air, to metal loss. Internal nitridation penetrated
and then automatically reinserted back into the deeper into metal interior than internal oxidation
hot zone. The specimens were subject to severe penetration. Table 3.23 included only internal
thermal cycling. Combustion was generated oxidation penetration data (i.e., maximum
using No. 2 fuel oil with an air-to-fuel ratio of 50 metal affected = metal loss + internal oxidation
to 1, producing a high-velocity (0.3 Mach, or penetration). Limited tests were conducted to
100 m/s) test gas. The tests were conducted at determine the effect of thermal cycling by testing
1090 °C (2000 °F) with 30 min cycles, and the at 980 °C (1800 °F) for 1000 h with 30 min
results are tabulated in Table 3.22. cycling and without thermal cycling in dynamic
At 1090 °C (2000 °F) with a high-velocity oxidation testing (Ref 82). As expected, thermal
gas stream plus severe thermal cycling, most cycling primarily contributed metal loss portion
alloys suffered significant metal loss, which of the oxidation attack. The results are summar-
constituted a large portion of the total depth of ized in Table 3.24 (Ref 82). Total attack pre-
oxidation attack. With protection by an alumi- sented in Table 3.24 was based on metal loss and
num oxide scale, alloy 214 suffered very little internal oxidation penetration. Since internal
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:46PM Plate # 0 pg 48

48 / High-Temperature Corrosion and Materials Applications

3
10
Carbon steel

10

2 9Cr-1Mo
10
Penetration, mils

AISI 410

Penetration, mm
1

Nickel

Alloy 800 H

10

–1
AISI 310 10
AISI 304

Alloy 617

–2
1 10
–3 –2 –1 1
10 10 10
pO , atm
2

Fig. 3.57 Effect of oxygen concentration in the N2-O2 mixture on the oxidation penetration (metal loss + internal attack) after 1 year at
927 °C (1700 °F) for various commercial alloys. 1.0 mils = 0.025 mm. Source: Ref 15

nitridation attack was found to penetrate deeper MA956 along with some ODS alloys was tested
into the alloy than internal oxidation attack does by Lowell et al. (Ref 78) with 0.3 Mach gas
for many alloys, the total depth of attack for velocity at 1100 °C (2010 °F) with 60 min
many alloys under dynamic oxidation test con- cycles. ODS alloys tested included MA956 (Fe-
ditions was more than that reported in Tables 19Cr-4.4Al-0.6Y2O3), HDA8077 (Ni-16Cr-
3.23 and 3.24. The oxidation/nitridation beha- 4.2Al-1.6Y2O3), TD-NiCr (Ni-20Cr-2.2ThO2)
vior of various alloys under dynamic oxidation and STCA264 (Ni-16Cr-4.5Al-1Co-1.5Y2O3).
test conditions is discussed in Chapter 4 “Nitri- Also included in the test was physical vapor
dation.” deposition (PVD) coating of Ni-15Cr-17Al-0.2Y
Hicks (Ref 83) performed dynamic oxidation on MAR-M-200 alloy (Ni-9Cr-10Co-12W-1Nb-
tests with 170 m/s gas velocity at 1100 °C 5Al-2Ti). Their results are shown in Fig. 3.60.
(2010 °F) with 30 min cycles for several MA956 and HDA8077 as well as PVD Ni-Cr-Al-
wrought chromia-former superalloys and an Y coating were found to perform well. No
ODS alumina-former (MA956). Alumina former explanation was offered in the paper for
MA956 was found to be considerably better than STCA264, which did not perform as well as
chromium formers, such as alloys 230, 86, 617, HDA8077 although both alloys had similar
188, and 263. His results are shown in Fig. 3.59. chemical compositions.
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:46PM Plate # 0 pg 49

Chapter 3: Oxidation / 49

Table 3.22 Dynamic oxidation resistance

of iron-, nickel-, and cobalt-base alloys in
high-velocity combustion gas stream (0.3 mach
velocity) with 30 min cycles at 1090 °C (2000 °F)
for 500 h
Metal loss, Maximum metal
Alloy mm (mils) affected(a), mm (mils)
214 0.013 (0.5) 0.046 (1.8)
230 0.056 (2.2) 0.15 (5.7)
RA333 0.10 (4.0) 0.22 (8.7)
188 0.19 (7.5) 0.27 (10.7)
556 0.22 (8.7) 0.30 (11.7)
X 0.23 (9.0) 0.34 (13.5)
RA330 0.28 (10.9) 0.35 (13.6)
S 0.30 (11.8) 0.39 (15.2)
600 0.44 (17.2) 0.53 (20.7)
310 0.54 (21.2) 0.61 (24.1)
601 0.27 (10.7) 0.61 (24.0)
617 0.32 (12.4) 0.61 (24.0)
800H 0.77 (30.5)(b) 0.86 (34.0)(b)
625 >0.79 (31.0)(c) >0.79 (31.0)(c)
Multimet 1.25 (49.1)(d) 1.42 (55.8)(d)
Note: Gas velocity was 0.3 mach (100 m/s, or 225 mph); samples were cycled to
less than 260 °C (500 °F) once every 30 min; 50 to 1 air-to-fuel ratio; two parts
No. 1 fuel oil and one part No. 2 fuel oil. Internal nitridation occurred in some
alloys, but is not included in the current data. See section 4.3.3 in Chapter 4.
(a) Metal loss + maximum internal, penetration. (b) Extrapolated from 400 h;
sample was about to be consumed after 400 h. (c) Sample was consumed in
500 h. (d) Extrapolated from 225 h; sample was about to be consumed after
225 h. Source: Ref 82

Table 3.23 Dynamic oxidation resistance

of iron-, nickel-, and cobalt-base alloys in
high-velocity combustion gas stream at 980 °C
(1800 °F) for 1000 h
Metal loss, Maximum metal
Alloy mm (mils) affected(a), mm (mils)
214 0.010 (0.4) 0.031 (1.2)
230 0.020 (0.8) 0.089 (3.5)
Fig. 3.58 Oxide scales formed on alloy 214 in a high-velocity
gas stream (0.3 Mach velocity) with 30 min cycles
188 0.028 (1.1) 0.107 (4.2) at 1090 °C (2000 °F) for 500 h. Area 1: 96.5% Al, 1.5% Cr, 0.1%
556 0.043 (1.7) 0.158 (6.2) Fe, 1.9% Ni. Area 2: 75.2% Al, 6.2% Cr, 2.6% Fe, 16.0% Ni. Area
X 0.069 (2.7) 0.163 (6.4) 3: 95.8% Al, 1.0% Cr, 0.1% Fe, 3.1% Ni. Area 4: 53.0% Al, 2.8%
S 0.079 (3.1) 0.17 (6.6) Cr, 9.2% Fe, 35.0% Ni. Source: Ref 82
RA333 0.064 (2.5) 0.18 (7.0)
625 0.12 (4.9) 0.19 (7.6)
617 0.069 (2.7) 0.27 (10.7)
RA330 0.20 (7.8) 0.30 (11.8) Table 3.24 Effect of thermal cycling in
Multimet 0.30 (11.8) 0.38 (14.8) dynamic oxidation behavior of several
800H 0.31 (12.3) 0.39 (15.3)
310 0.35 (13.7) 0.42 (16.5) nickel-base alloys at 980 °C (1800 °F) for 1000 h
600 0.31 (12.3)(b) 0.45 (17.8)(b) No thermal cycling Thermal cycling
601 0.076 (3.0) 0.51 (20.0)
Metal loss, Total attack(a), Metal loss, Total attack(a)
304 >9.0 (354)(c) >9.0 (354)(c)
Alloy mm mm mm mm
316 >9.0 (354)(c) >9.0 (354)(c)
230 0.04 0.11 0.07 0.16
Note: Gas velocity was 0.3 mach (100 m/s, or 225 mph); samples were cycled to
617 0.03 0.16 0.17 0.24
less than 260 °C (500 °F) once every 30 min; 50 to 1 air-to-fuel ratio; two parts
No. 1 fuel oil and one part No. 2 fuel oil. Internal nitridation occurred in some X 0.03 0.12 0.16 0.23
alloys, but is not included in the current data. See section 4.3.3 in Chapter 4. 263 0.07 0.21 0.32 0.42
(a) Metal loss + maximum internal penetration. (b) Extrapolated from 917 h;
(a) Metal loss + internal oxidation. Source: Ref 82
sample was about to be consumed after 917 h. (c) Extrapolated from 65 h; sample
was consumed in 65 h. Source: Ref 82

the alloy. The growth of the Cr2O3 oxide scale

3.4.13 Breakaway Oxidation
follows a parabolic rate law as the exposure time
In Fe-Cr, Fe-Ni-Cr, Ni-Cr, and Co-Cr alloy increases. As the temperature increases, the
systems, the formation of an external Cr2O3 oxide scale grows faster. The growth of the
oxide scale provides the oxidation resistance for Cr2O3 oxide scale requires a continuous supply
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:46PM Plate # 0 pg 50

50 / High-Temperature Corrosion and Materials Applications

0 or thermal cycling. Once this occurs, fast-

MA 956 growing, nonprotective iron oxides, or nickel
oxides or cobalt oxides (i.e., oxides of base
metal) form and grow on the alloy surface.
0.1 230 Breakaway oxidation, thus, initiates, and the
Total depth of oxidation, mm/side

alloy begins to undergo oxidation at a rapid rate.

This is illustrated in Fig. 3.51. The alloy thus
617 requires the level of chromium immediately
0.2 underneath the chromium oxide scale to have a
86 critical level to allow the chromium oxide scale
to reheal.
188
Gleeson (Ref 84) presented air cyclic oxida-
0.3 (1971 Cast) tion data for three chromia formers tested at
X 982 °C (1800 °F) for up to 360 days. Also pre-
263 sented were the corresponding chromium con-
188
(1980 Cast) centration analyzed by EDX on the surface of the
0.4 metal when the oxide scale was spalled off from
(C)
the test specimen. The data are presented in
0 200 400 600 800 Fig. 3.61. In this test program, thick, blocky
Number of cycles specimens (13 mm thick, 25 mm wide, and
25 mm long) instead of typically thin coupons
Fig. 3.59 Dynamic oxidation resistance of several wrought were used. The specimens were cycled to room
superalloys including MA956 alloy in high-
velocity combustion gas stream (170 m/s) at 1100 °C (2010 °F) temperature once every 30 days for weight
with 30 min cycles. Source: Ref 83 measurement. Oxide scales were found to com-
pletely spall off while the specimens were
removed from the furnace for cooling to room
10
Coated MAR-M-200
temperature with “popping” noises being heard
Specific weight change, mg/cm2

during cooling (Ref 85). Figure 3.61 shows that

0
alloy 230 exhibited very little weight loss with no
MA 956
–10 evidence of breakaway oxidation after 360 days
HDA- at 982 °C (1800 °F). The chromium concentra-
8077
–20 tion of the alloy immediately underneath the
spalled oxide scale was found to remain at about
–30
16% with no sign of decreasing with increasing
–40 exposure time. For HR120 alloy, the weight loss
STCA-264 data also revealed no evidence of breakaway
–50 oxidation up to 360 days of exposure. The cor-
0 400 800 1200 1600 2000 2400 2800 3200
responding chromium concentration of the alloy
Cycles
on the surface underneath the spalled oxide scale
remained approximately about 18 to 20% up to
Fig. 3.60 Dynamic oxidation tests at 1100 °C (2010 °F) in a
Mach 0.3 gas stream with each cycle consisting of 240 days, and then dropped to about 13% after
1 h at temperature followed by quenching to ambient tempera- 360 days. Alloy 800HT, on the other hand,
ture for 3 min. Source: Ref 78
showed breakaway oxidation after 180 days. The
corresponding chromium concentration of the
alloy on the surface immediately underneath of
of chromium from the alloy interior diffusing to the spalled oxide scale after 180 days was found
the oxide/metal interface. Continued oxidation to be about 10%.
can eventually deplete chromium in the alloy With continuing oxidation, alloy 800HT suf-
matrix immediately underneath the oxide scale. fered linear weight loss and continued the
When the chromium concentration in the alloy decrease in chromium concentration to 8% when
matrix immediately beneath the oxide scale is the exposure reached to 360 days. However,
reduced to below a critical concentration, the when alloy 800HT was oxidized after 90 days
alloy matrix no longer has adequate chromium to showing no sign of weight loss, the chromium
reform a protective Cr2O3 oxide scale when the concentration of the alloy underneath the oxide
scale cracks or spalls due to oxide growth stresses scale was about 11%. The data appear to suggest
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:46PM Plate # 0 pg 51

Chapter 3: Oxidation / 51

that the critical chromium concentration of the spalled oxide scale to the alloy interior after 360
alloy underneath the oxide scale is approxi- days of testing was determined using EDX ana-
mately 11% for alloy 800HT. When the chro- lysis with the results shown in Fig. 3.63. The
mium concentration underneath the oxide scale chromium concentration immediately under-
is below 11%, the reformation of a protective neath the spalled oxide scale for both alloys 230
chromium-rich oxide scale is not possible, thus and HR120 was well below 10%, while that of
resulting in breakaway oxidation. alloy HR160 was about 10%. Alloy HR120
In the same oxidation test program at 1150 °C showed sign of breakaway oxidation after
(2100 °F), similar analysis on the chromium 90 days of exposure. Continuing oxidation test-
concentration profile underneath the oxide ing resulted in a linear weight-loss rate. Alloy
scale was performed for alloys 230, HR160 230 showed signs of breakaway oxidation after
and HR120 with the data presented in 240 days. HR160, however, showed a linear
Fig. 3.62 and 3.63 (Ref 86). Figure 3.62 shows weight loss up to 360 days, although suffering
the weight-loss data for three alloys up to the least weight-loss rate, with no clear sign of
360 days. The chromium concentration profile breakaway oxidation. Oxidation of alloy HR160
from the surface immediately underneath the is involved the formation of Cr2O3 and SiO2.
This may explain that HR160, although con-
tinuing to lose weight, still showed no sign of
breakaway oxidation after 360 days at 1150 °C
230 (2100 °F). In a long-term oxidation study of
0.0
Fe-20Cr-25Ni alloy in CO2 containing 1% CO,
HR120
300 ppm H2O, and 300 ppm CH4 at 1023 to
Weight change, mg/cm2

1173 K (750 to 900 °C), Evans et al. (Ref 87)

–100.0 found the critical chromium level for rehealing of
chromium oxide scales to be about 16%. The
minimum level of chromium needed to maintain
a protective chromium oxide scale to prevent
–200.0 breakaway oxidation may vary from environ-
800HT ment to environment and from alloy to alloy.
To prolong the time for the initiation of
–300.0
breakaway oxidation, it is necessary to have an
0 100 200 300 400 adequate reservoir for chromium immediately
(a) Time, days below the oxide scale to provide adequate chro-
mium to maintain a protective chromium oxide
28.0

0
HR160
Surface Cr concentration, wt%

24.0
Weight change, mg/cm2

HR120 –250.0
20.0 230

230
–500.0
16.0

HR120

12.0 –750.0
800HT

8.0 –1000.0
0 100 200 300 400
0 60 120 180 240 300 360
(b) Time, days
Time, days
Fig. 3.61 Weight changes as a function of exposure time in
long-term cyclic oxidation tests in air at 982 °C Fig. 3.62 Weight changes as a function of exposure time for
(1800 °F) for alloys 800HT, HR120, and 230 (a), and the corre- alloys 230, HR160, and HR120 in air oxidation
sponding changes in the surface chromium concentration (mea- tests at 1150 °C (2100 °F) with thermal cycling to room tem-
sured after the scale was spalled off) as a function of exposure time perature for weight measurement once every 30 days. Source:
(b). Source: Ref 84 Ref 86
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:46PM Plate # 0 pg 52

52 / High-Temperature Corrosion and Materials Applications

80.0 scale or to reheal the oxide scale that suffered

local cracking or failure. Brady et al. (Ref 88)
70.0 proposed that chromium carbides might provide
Ni
a reservoir of chromium for maintaining the
60.0
growth and rehealing of chromium oxide scales.
Concentration, wt%

50.0 These authors observed that an as-cast Fe-15Cr-

0.5C formed a nonprotective iron-rich oxide
40.0 scale when exposed in O2 at 850 °C (1562 °F).
30.0
When the alloy was forged at 1150 °C (2100 °F)
Cr to produce a uniformly distributed fine carbide
20.0 phase, the forged alloy showed a thin protective
W Cr2O3 scale under the same test condition (Ref
10.0
88). During oxidation testing of this forged
0 sample, the fine chromium-rich carbides are
0 200 400 600 800 1000 1200 dissolved into the underlying alloy substrate to
Distance from surface, µm continue supplying chromium to maintain and
(a) reheal the chromium oxide scales (Ref 88).
For alumina formers, such as Fe-Cr-Al alloys,
50.0
and Fe-Cr-Al-base and Ni-Cr-Al-base ODS
alloys, breakaway oxidation occurs when alu-
Ni minum concentration underneath the Al2O3 scale
40.0
has reduced to a critical level such that healing of
Concentration, wt%

Co
the Al2O3 is no longer possible, thus resulting in
30.0 the formation of nonprotective, fast-growing
oxides of base metals (e.g., iron oxides or nickel
oxides). The breakaway oxidation due to rapid
20.0
Cr growth of iron oxides or nickel oxides becomes
essentially a life-limiting factor. This critical
10.0 aluminum concentration was found to be about
1.0 to 1.3% for Fe-Cr-Al-base ODS alloys (e.g.,
MA956, ODM751) at 1100 to 1200 °C (2012
0 to 2192 °F) (Ref 89, 90). These values were
0 400 800 1200 1600 2000 obtained from foil specimens (0.2 to 2 mm thick)
Distance from surface, µm tested in still air at 1100 to 1200 °C. For the non-
(b)
ODS Fe-20Cr-5Al alloy, this critical aluminum
60.0 concentration was found to be higher (about
2.5%) at 1200 °C (Ref 89). Since the breakaway
50.0 oxidation is related to aluminum reservoir in the
alloy, and the aluminum reservoir becomes a
Ni critical issue when the component is made of thin
Concentration, wt%

40.0
sheet or foil. Because of excellent oxidation
Fe
resistance at very high temperatures, there is
30.0
increasing interest in looking at alumina formers
for products that require thin foils, such as hon-
Cr
20.0 eycomb seals in gas turbines, metallic substrates
for automobile catalyst converters, and recup-
10.0 erators in microturbines. The oxidation behavior
of several commercial alumina formers in thin
0 foils is summarized in Section 3.4.14.
0 200 400 600 800 1000 For alumina formers to improve their resis-
Distance from surface, µm
tance to breakaway oxidation, yttrium is fre-
(c) quently used to increase the adhesion of the
aluminum oxide scale. Other alloying elements
Fig. 3.63 Concentration profiles for (a) alloy 230, (b) HR160,
that are known to increase the adhesion of the
and (c) HR120 after air oxidation tests for 360 days
at 1150 °C (2100 °F), as shown in Fig. 3.62. Source: Ref 86 aluminum oxide scale include zirconium and
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:46PM Plate # 0 pg 53

Chapter 3: Oxidation / 53

hafnium. Quadakkers (Ref 91) shows that both normal purity Ni-Cr-Al alloys (approximately 30
MA956 (Fe-20Cr-4.5Al-0.5Y2O3) and Alu- to 40 ppm S) with the high-purity Ni-Cr-Al
chrom (Fe-20Cr-5Al-0.01Y) exhibited much alloys (approximately 1 to 2 ppm S), showing a
more cyclic oxidation resistance than Fe-20Cr- significant improvement in cyclic oxidation
5Al when tested at 1100 °C in synthetic air with a resistance when sulfur in the alloy was sig-
hourly cycle to room temperature (Fig. 3.64). nificantly reduced. This is illustrated in Fig. 3.66
Addition of Y2O3 to an alumina former has a (Ref 96). Also demonstrated in the figure is the
similar beneficial effect as yttrium alloying ele- beneficial effect of yttrium addition to the normal
ment. Klower and Li (Ref 92) studied the oxi- purity Ni-20Cr-12Al alloy, showing significant
dation resistance of Fe-20Cr-5Al alloys in 10 improvement in the cyclic oxidation resistance of
different compositions containing various the alloy without reducing the sulfur content in
amounts of yttrium ranging from 0.045 to 0.28%. the alloy. Sulfur has been found to segregate to
All 10 compositions contained 0.002% S, and the oxide/alloy interface during oxidation in Fe-
eight compositions contained 0.04 to 0.06% Zr Cr-Al alloys (Ref 97, 98). The role of yttrium is
with two compositions containing no zirconium. believed to tie up sulfur at the oxide/metal
Cyclic oxidation tests were performed at 1100 interface, thus improving the oxide-scale adhe-
and 1200 °C (2012 and 2192 °F), respectively, sion (Ref 96).
with each cycle consisting of 96 h at temperature
and rapid air cooling to room temperature. These
authors concluded that the yttrium addition 3.4.14 Thin Foils
of about 0.045% was sufficient to prevent the
There are some industrial applications that
oxide scales from spalling and when the yttrium
require thin-gage sheet materials or thin foils
concentration was increased to more than
for construction of some critical components.
0.08%, substantial internal oxidation could
As the component thickness decreases, oxidation
occur, resulting in rapid metal wastage, as shown
becomes a major limiting factor for its service
in Fig. 3.65 (Ref 92).
life. When the component is made of thin foil,
Sulfur in the alloy is known to play a very
prolonging the incubation time before the
significant role in the adhesion of the aluminum
initiation of breakaway oxidation is the control-
oxide scale to the alloy substrate for alumina
ling factor for extending the service life of the
formers. The role of yttrium is believed to pre-
component. Thus, as applications are being
vent the preferential segregation of sulfur in the
pushed toward higher and higher temperatures,
alloy to the scale/metal interface to weaken the
alloys that form aluminum oxide scales can offer
adhesion of the oxide scale (Ref 93–95). Redu-
tremendous advantages in performance over
cing the concentration of sulfur in a Ni-Cr-Al
those alloys that form chromium oxide scales.
alloy can significantly improve the oxidation
In gas turbine applications, one important
resistance of the alloy. Smeggil (Ref 96) com-
component made of a thin foil is a turbine seal
pared cyclic oxidation resistance between the
ring assembly that controls the turbine tip

4
Depth of internal oxidation, µm

Fe-20Cr-AI 800
1100 °C
Weight change, mg/cm2

3 MA956
1200 °C
Aluchrom
600
2

1 400

0 200

–1 0
0 200 400 600 800 1000 1200 0 0.1 0.2 0.3
Time, h Yttrium, wt%

Fig. 3.64 Cyclic oxidation resistance of MA956, Aluchrom, Fig. 3.65 Maximum internal oxidation depth as a function of
and Fe-20Cr-5Al tested in synthetic air at 1100 °C yttrium content in the alloys after 3000 h of cyclic
(2012 °F) with an hourly cycle (each cycle consisted of 56 min oxidation tests at 1100 and 1200 °C (2012 and 2192 °F) in air.
heating and 4 min cooling). Source: Ref 91 Source: Ref 92
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:46PM Plate # 0 pg 54

54 / High-Temperature Corrosion and Materials Applications

High-purity NiCrAI
Mass change/unit area, mg/cm2

–8

Normal-purity
NiCrAIY

–16

–24 Normal-purity
NiCrAI

–32

0 20 40 60 80 100
Number of 1 h cycles

Fig. 3.66 Cyclic oxidation resistance of the normal purity Ni-20Cr-12Al (30 to 40 ppm S), the high-purity Ni-20Cr-12Al (1 to 2 ppm S)
and the normal purity Ni-20Cr-12Al-Y at 1180 °C. Source: Ref 96

clearance for improving thermal efficiency. contained in an individual alumina crucible so

The seal ring assembly is typically constructed the spalled oxides could be included in the
out of a honeycomb seal brazed onto a super- weight measurement. All the specimens were
alloy casing. The traditional alloys used for preoxidized in air at 1050 °C (1922 °F) for 1 h
honeycomb seals are chromia formers, such as prior to oxidation testing. The alloys tested
nickel-base alloy X. Lai (Ref 99) evaluated included Kanthal AF (Fe-20Cr-5Al-0.05Y-
honeycomb samples made of wrought alloy 214 0.08Zr wrought alloy), Aluchrom YHf (Fe-20Cr-
(Ni-16Cr-3Fe-4.5Al-Y) and alloy X (Ni-22Cr- 5.8Al-0.04Y-0.05Zr-<0.1Hf wrought alloy),
18.5Fe-9Mo) in dynamic burner rig testing that PM2000 (Fe-20Cr-5.5Al-0.5Y2O3 ODS alloy),
simulated gas turbine hot gas conditions. Tests and Haynes 214 alloy (Ni-3Fe-16Cr-4.5Al-
were conducted at 954 °C (1750 °F) in a high- 0.01Y wrought alloy). The 0.127 mm thick
velocity combustion gas stream (0.3 Mach or Haynes 214 foil specimens were the only foils
100 m/s) with cycling every 30 min. Both alloy exhibiting no breakaway oxidation over the test
X and alloy 214 honeycomb samples were duration of 300 h (15 cycles) at 1250 °C
made of 0.076 mm (3 mils) foils. The alloy X (2282 °F). Kanthal AF, Aluchrom YHf and
honeycomb sample was completely oxidized PM2000 foil specimens with 0.127 mm thick
(destroyed) after 154 h of testing, while the suffered breakaway oxidation well before com-
alloy 214 honeycomb sample was unaffected pleting 300 h (15 cycles) test duration. Thinner
after 317 h when the test was terminated. foils suffered breakaway oxidation in shorter
Figure 3.67 shows the condition of both samples times than thicker foils. As the test temperature
after tests. was lowered, the life of the foil was progressively
Simms et al. (Ref 100) conducted extensive longer. At 950 °C, tests were conducted only on
oxidation studies on several commercial foil alloys 214 and PM2000. Both alloys exhibited
materials with different thicknesses (from 0.05 to no breakaway oxidation for up to 1500 h (15
0.127 mm) in a simulated combustion environ- cycles) for all the thicknesses (0.05 to 0.127 mm)
ment (nominally N2-14%O2-3%CO2-7.8%H2O) tested.
at 950 to 1250 °C (1742 to 2282 °F). Test spe- Klower (Ref 101) studied the effect of the
cimens were cycled to room temperature every foil thickness varying from 0.049 to 0.25 mm
100 h when tested at 950 and 1050 °C (1742 and (2 to 10 mils) on breakaway oxidation of Fe-
1922 °F), every 40 h at 1150 °C (2102 °F), every 20Cr-5Al alloys, which contained rare-earth
20 h at 1250 °C (2280 °F). Each specimen was metals (mischmetal, such as Ce), in air at
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:46PM Plate # 0 pg 55

Chapter 3: Oxidation / 55

(a)

(b)

Fig. 3.67 Honeycomb samples after testing at 950 °C (1750 °F) for 154 h for the alloy X honeycomb sample (a) and 317 h for the 214
honeycomb sample (b) in a high-velocity combustion gas stream (0.3 Mach or 100 m/s) generated by a dynamic burner rig.
The samples were also subjected to rapid quenching from the test temperature to less than 260 °C (500 °F) for 2 min every 30 min. Both
honeycomb samples were made of 0.076 mm (3 mil) foils. Courtesy of Haynes International, Inc.

1100 °C (2012 °F). The specimens were cooled Thin foils of oxidation-resistant alloys
to room temperature for weighing every 96 h. including stainless steels, Fe-Ni-Cr alloys and
The time to breakaway oxidation increased nickel-base alloys have been extensively eval-
with increasing foil thickness, as illustrated in uated for high-temperature recuperators in
Fig. 3.68 (Ref 101). microturbines (Ref 102–104). The use of a
The incubation time for breakaway oxidation recuperator for preheating the incoming air for
is dependent on the amount of aluminum in the combustion can significantly increase the effi-
reservoir for alumina formers. The total amount ciency of the microturbine. The oxidation tests
of aluminum in the reservoir of the alloy were carried out on foils with about 100 µm
increases with increasing foil thickness. Thus, it (0.1 mm, or 4 mils) thick in air containing 10%
takes much longer for a thicker foil to reduce the H2O, which was to simulate the level of H2O that
concentration of aluminum below the critical would be present in the exhaust gas stream in
level such that rehealing of the protective oxide microturbines. It was found that air containing
scale is not possible, resulting in breakaway 10% H2O was significantly more aggressive than
oxidation. dry air (i.e., laboratory air) (Ref 102–104). The
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:46PM Plate # 0 pg 56

56 / High-Temperature Corrosion and Materials Applications

60
0.072
Foil thick-
0.058 ness, mm
50
0.25
0.165
Mass change, g/m2

40
0.05 0.125
0.091 0.091
30 0.072
0.049 0.058
20 0.05
0.049

10

0
0 500 1000 1500 2000 2500 3000 3500
Exposure time, h

Fig. 3.68 Weight gain as a function of exposure time in air at 1100 °C (2012 °F) for Fe-20Cr-5Al foils with various thicknesses. The
alloy also contained 0.015% Mischmetal. Source: Ref 101

effect of water vapor on oxidation behavior of accelerated oxidation (Ref 105). The effect of
alloys is discussed in next section. water vapor on oxidation of alloys is discussed in
Pint (Ref 104) reported that Type 347 foil the next section.
(100 µm, or 4 mils, thick) suffered accelerated
oxidation (or breakaway oxidation) after less
than 2000 h of exposure at 650 °C (1200 °F) in 3.4.15 Effect of H2O on Oxidation
air containing 10% H2O. The foil was also sub- In high-temperature combustion atmospheres,
jected to thermal cycling every 100 h. Under the water vapor is invariably present in the envi-
same test condition, alloy HR120, and alloy 625 ronment. In some cases, the level of water vapor
showed no accelerated oxidation (or breakaway in the environment can be significant. The effect
oxidation) after more than 8000 h. Both alloys of water vapor on the oxidation of alloys is thus
HR120 and 625 exhibited a slight mass loss an important factor in the alloy selection process.
of about less than 0.2 mg/cm2 after 8000 h of Most oxidation data are generated in laboratory
exposure. The author attributed this small mass air, which generally contain low levels of water
loss to volatilization of CrO2(OH)2 that formed vapor.
during exposure of the water vapor in the envir- Tuck et al. (Ref 106) investigated the effect of
onment (Ref 104). Both alloys HR120 and 625 water vapor on oxidation of iron in air and air
showed that the oxidation behavior at 700 °C with 5%, 10%, 15%, and 20% H2O at 800 and
(1292 °F) was similar to that at 650 °C 1000 °C (1472 and 1832 °F) for times up to
(1200 °F), exhibiting very little mass loss after 200 min. Figure 3.69 shows the weight changes
more than 8000 h (Ref 104). At 800 °C as a function of time at 800 °C (1472 °F), indi-
(1472 °F), alloy HR120 showed no sign of cating essentially no effects on oxidation by the
accelerated oxidation after 7500 h. Alloy 625 presence of water vapor up to 20% (Ref 106).
again showed excellent oxidation resistance with Also, no water vapor effects were observed when
little mass loss at 800 °C (1472 °F) for times up tested at 1000 °C (1832 °F). The tests, however,
to 6000 h when the test was terminated (Ref were conducted at such high temperatures that
104). Both HR120 and 625 exhibited excellent carbon steels would never be used. In boilers
oxidation resistance in air containing 10% H2O burning fuels containing large amount of moist-
at 650 to 800 °C (1200 to 1472 °F). However, ure (e.g., 15 to 40% in coal) (Ref 107), a large
Type 347 performed poorly at 650 °C (1200 °F) amount of H2O is expected to be present in the
in air containing 10% H2O. Nevertheless, when combustion atmosphere in a coal-fired boiler.
Type 347 foil was tested at 650 °C (1200 °F) in The flue gas analysis from a utility pulverized
laboratory air (i.e. dry air) for up to 40,000 h, the coal-fired boiler showed approximately 10%
alloy showed a thin oxide scale with no sign of H2O (Ref 108). Carbon and low-alloy steels,
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:46PM Plate # 0 pg 57

Chapter 3: Oxidation / 57

40

30
Weight gain, mg/cm2

20

10
Air 20% steam
Air 15% steam
Air 10% steam
Air 5% steam
Dry air
0 50 100 150 200 250
Time, min

Fig. 3.69 Effect of water vapor on oxidation of iron in air and air with 5%, 10%, 15%, and 20% H2O at 800 °C (1472 °F) for times up to
200 min. Source: Ref 106

which are used for the construction of the boiler and O2 with 10% H2O. In the O2 environment,
waterwalls in the combustion zone, have the weight of the alloy increased initially with
performed well under normal oxidizing con- increasing time, which was then followed by a
ditions. More detailed discussion on oxidation decrease in the rate of increase with further
and high-temperature corrosion of carbon and exposure time, indicating the formation of a
low-alloy steels in coal-fired boilers is presented protective oxide scale. As for the environment
in Chapter 10. consisting of O2 with 10% H2O, the weight
Segerdahl et al. (Ref 109, 110) studied the of the alloy initially increased and then decreased
oxidation behavior of 11Cr steel (X20 with increasing exposure time. The authors
11Cr1MoV) in O2 and O2 with 10% and 40% observed a brownish layer of deposits in the
H2O at 600 °C (1112 °F). In tests at 600 °C exhaust end of the furnace tube downstream from
(1112 °F) for 700 h at a gas flow rate of 0.5 cm/s, the test specimen. The analysis of the deposits
the steel showed no weight changes for 100% O2 revealed a high concentration of chromium.
and O2-10H2O, while it suffered breakaway They proposed the weight loss in the O2-10%
oxidation after 336 h when tested in O2-40% H2O environment was due to the evaporation
H2O (Ref 109). The steel exposed to dry O2 of CrO2(OH)2 that formed on the alloy as a
exhibited a protective (Cr,Fe)2O3 scale after result of water vapor. The author showed that
testing, while the steel showed hematite in the the theoretical partial pressure of CrO2(OH)2
outer layer and FeCr spinel in the inner layer in the O2-10%H2O environment was quite
when exposed to O2-40H2O mixtures (Ref 109). high at low temperatures, as shown in Fig. 3.4
When tests were conducted at 600 °C (1112 °F) (Ref 9).
for 168 h with test gas velocity at 0.5, 1.0, 2.5, 5, The effect of water vapor on the oxidation
and 10 cm/s, the steel showed no weight changes behavior of Type 321 at 800 °C (1472 °F) was
in 100% O2 and O2-10% H2O at all gas velo- illustrated in Fig. 3.70, showing Type 321 foil
cities, while it suffered accelerated weight gain suffered severe oxidation attack in air containing
at 5 and 10 cm/s when tested in O2-40% H2O 10% H2O compared with the sample tested in dry
(Ref 110). air (laboratory air) (Ref 111). The figure also
Asteman et al. (Ref 9) studied oxidation shows a severe oxidation attack for Type 347
behavior of Type 304L at 873 K (600 °C) in O2 foil tested at 800 °C (1472 °F) in air with 10%
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:46PM Plate # 0 pg 58

58 / High-Temperature Corrosion and Materials Applications

10

347, 800 °C 10% H2O

9

8 321, 800 °C 10% H2O

7
Mass change, mg/cm2

6
321, 700 °C 10% H2O

2 347, 700 °C 10% H2O

1 321, 800 °C dry air

0
0 100 200 300 400 500 600 700 800 900 1000
Time, h

Fig. 3.70 Oxidation behavior of Type 321 and 347 foils (100 µm thick) at 700 and 800 °C (1292 and 1472 °F) in dry air and air with
10% H2O. Specimens were cycled to room temperature every 100 h. Source: Ref 111

H2O. The detrimental effect of water vapor in compare some promising nickel aluminides
air on the oxidation resistance of Type 347 is (Ni3Al) or nickel aluminide-base materials with
clearly illustrated in Fig. 3.71 (Ref 112). Water some commercial nickel-base alloys. Readers are
vapor has been found to be detrimental to referred to Ref 117 for more information about
the oxidation resistance of Type 310 (Fig. 3.72) the oxidation and corrosion of various types of
(Ref 114), alloy X (Fig. 3.73) (Ref 114), and intermetallic alloys.
CMSX-4 (Fig. 3.74) (Ref 115). Both Type Comparison oxidation tests were performed
310 and alloy X are chromia formers, while in air at 1150 and 1200 °C (2100 and 2200 °F)
CMSX-4 (Ni-9.5Co-6.4Cr-5.7Al-6.3W-6.5Ta- between two promising Ni3Al-base nickel
2.9Re) is an alumina former. Onal et al. (Ref 116) aluminides, IC-50 and IC-218 (developed by
investigated the cyclic oxidation for a number of Oak Ridge National Laboratory), and nickel-
alumina-formers cyclic oxidation in both dry air base alloys, alloy 214 (alumina former) and
and air containing 30% H2O at temperatures alloys X and 230 (chromia formers) (Ref
from 700 to 1000 °C (1292 F to 1832 °F). The 55). Test results for IC-50 (Ni-11.3Al-0.6Zr-
authors proposed that water vapor reduced the 0.02B), IC-218 (Ni-8.5Al-7.8Cr-0.8Zr-0.02B),
adhesion of aluminum oxide scale, thus causing alloy 214 (Ni-4.5Al-16Cr-3Fe-0.01Y), alloy X
oxide spallation (Ref 116). (Ni-22Cr-18Fe-9Mo-0.6W), and alloy 230
(Ni-22Cr-14W-2Mo-0.02La) are summarized
in Table 3.25. Test results on other alloys
3.4.16 Intermetallic Compounds under the same test conditions are shown in
Table 3.12.
Tremendous interest has been generated Samples of IC-218 were completely turned
among academic researchers in intermetallic into oxides after 1008 h at 1150 and 1200 °C
materials and their structures, properties, and (2100 and 2200 °F), although they still main-
high-temperature corrosion behaviors. However, tained the sample shape. Their specimen weight
industrial applications of these materials are still gain data did not reveal the complete oxidation of
limited. Limited data are presented here to briefly the specimen. IC-50 suffered internal oxidation
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:46PM Plate # 0 pg 59

Chapter 3: Oxidation / 59

347 foil, 40,000 h, air

50 µm
(a) Cu-plating

Cu-plating

347 foil, 40,000 h, air 10 µm

(b)

347 foil, 10,000 h, air+10% H2O

50 µm
(c)

Fig. 3.71 Optical micrographs showing a thin oxide scale formed on Type 347 foil when tested in laboratory air after 40,000 h at
650 °C (1200 °F) (a) and (b) and thick unprotective oxide scales formed on Type 347 foil when tested in air containing 10%
H2O after only 10,000 h at the same temperature (c). Source: Ref 112. Courtesy of Oak Ridge National Laboratory

10 that penetrated through the specimen thickness,

0 as illustrated in Fig. 3.75. Again, the weight
change data failed to reveal the severity of the
oxidation attack for IC-50. This example
emphasizes the fact that specimen weight gain
(or weight change) data can be misleading.
Weight change, mg/cm2

Metallographic examination of the test speci-

–50 mens is also required to determine the total depth
of attack, which included both metal loss and
internal penetration.
Comparing IC-50 and IC-218 with alloy 214,
it is interesting to note that both nickel alumi-
nides, with significantly higher aluminum con-
–100 tents (twice of that in alloy 214), failed to
produce a protective aluminum oxide scale.
0.01% H2O
5% H2O
SEM/EDX analyses of the scales formed on both
10% H2O nickel aluminide samples after testing at 1150 °C
(2100 °F) revealed both nickel- and aluminum-
rich oxides for IC-50 and mostly nickel-rich
–150 oxides for IC-218 (Fig. 3.76) (Ref 55). It is
0 200 400 600 800 1000
believed that chromium is needed to increase
Time, h Descaled
aluminum activities in the alloy for forming a
continuous aluminum oxide scale. In his studies
Fig. 3.72 Effect of water vapor in air on the oxidation resis-
on oxidation of IC-50 in air at 1000 °C
tance of Type 310 at 1100 °C (2010 °F) cycled
every 100 h. Source: Ref 113 (1832 °F), Natesan (Ref 118) found that NiO was
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:46PM Plate # 0 pg 60

60 / High-Temperature Corrosion and Materials Applications

10 +1
CMSX4 in dry air
0 0

–1

–2
∆W/A, mg/cm2

∆m/A, mg/cm2
–3 CMSX4
0.01% H20 –4 in wet air
–50 5% H20
–5
10% H20
–6

–7

–8
–100 –9
0 200 400 600 800 1000 0 100 200 300 400
Time, h Descaled
Time, h

Fig. 3.73 Oxidation of alloy X at 1100 °C (2010 °F) in air Fig. 3.74 Cyclic oxidation behavior of CMSX4 (Ni-9.5Co-
containing various amounts of H2O. Cycling to 6.4Cr-5.7Al-6.3W-6.5Ta-2.9Re) in dry air and air
room temperature every 100 h for weighing. Source: Ref 114 containing 0.1 atm H2O (10% H2O in 1.0 atm test gas), cycling
once an hour with 45 min at 1100 °C (2012 °F) and 15 min
cooling. Source: Ref 115

Table 3.25 Results of oxidation tests on nickel aluminides and nickel-base alloys in air for 1008 h at
indicated temperatures
1150 °C (2100 °F) 1200 °C (2200 °F)
Average metal Average metal
Weight gain, Metal loss, affected(a), Weight gain, Metal loss, affected,
Material mg/cm2 mm (mils) mm (mils) mg/cm2 mm (mils) mm (mils)
IC-50 2.7 0.008 (0.3) >0.38 (15)(b) 8.8 0.02 (0.8) >0.38 (15)(b)
IC-218 16.7 >0.36 (14)(c) >0.36 (14)(c) 61.5 >0.36 (14)(c) >0.36 (14)(c)
214 1.2 0.005 (0.2) 0.0075 (0.3) 1.3 0.005 (0.2) 0.018 (0.7)
230 33.5 0.058 (2.3) 0.086 (3.4) 81.2 0.11 (4.5) 0.20 (7.9)
X 68.5 0.11 (4.5) 0.147 (5.8) 169(d) >0.9 (35.4)(e) >0.9 (35.4)(e)
Note: Flowing air 30 cm/min (472 cm3 min in a 1.75 in. diam tube); cycled to room temperature once a week (168 h cycles). (a) Metal loss + average internal penetration.
(b) Internal penetration through thickness. (c) Sample was completely oxidized. (d) Specimen weight gain after 504 h. (e) Extrapolated from 504 h; specimen was
completely oxidized (consumed) in 504 h. Source: Ref 55

the only oxide formed on the sample; no Al2O3 50 h tested sample revealed mainly nickel-rich
was detected. oxides.
Burner rig dynamic oxidation tests under a
high-velocity combustion gas stream (0.3 Mach
or 100 m/s) were also performed on IC-50 3.4.17 Catastrophic Oxidation
compared with commercial alloys (Ref 55). As temperature increases, metals and alloys
The nickel aluminide IC-50 suffered sig- generally suffer increasingly higher rates of
nificantly more severe oxidation attack than oxidation. When the temperature is excessively
nickel-base alloys 214 (alumina former) and 230 high, metals and alloys can suffer rapid oxida-
(chromia former) after testing at 1090 °C tion. There is, however, another mode of rapid
(2000 °F) for 500 h with 30 min cycles. IC-50 oxidation that takes place at relatively lower
suffered more than 0.38 mm (15 mils) of oxi- temperatures. This mode of rapid oxidation,
dation attack, compared with 0.05 mm (1.8 mils) which is often referred to as “catastrophic oxi-
for alloy 214 and 0.14 (5.7 mils) for alloy 230. dation,” is associated with the formation of a
The data for other commercial alloys tested liquid oxide. The liquid oxide disrupts and dis-
under the same conditions are shown in Table solves the protective oxide scale, causing the
3.22. The nickel aluminide was very susceptible alloy to suffer rapid oxidation at relatively low
to internal oxidation. Significant internal oxida- temperatures. Leslie and Fontana (Ref 119)
tion was observed after only 50 h of testing. observed an unusually rapid oxidation for Fe-
SEM/EDX analysis of the scale formed on the 25Ni-16Cr alloy containing 6% Mo when heated
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:46PM Plate # 0 pg 61

Chapter 3: Oxidation / 61

Original
Sample Thickness

33.3 µm

Fig. 3.76 Nickel-rich oxides formed on nickel aluminide IC-

218 after 1008 h at 1150 °C (2100 °F) in air with
168 h cycles. Area 1: 11.7% Al, 80.0% Ni, 8.3% Cr. Area 2:
18.8% Al, 49.0% Ni, 32.2% Cr. Area 3: 22.7% Al, 57.1% Ni,
20.2% Cr. Source: Ref 55

Other mixed oxides involving MoO3 that exhibit

low melting points are MoO2-MoO3 (778 °C,
or 1435 °F), MoO2-MoO3-NiO (764 °C, or
1400 °F), Fe2O3-MoO3 (730 °C, or 1345 °F),
Fe2O3-Fe-MoO3 (725 °C, or 1335 °F), V2O5-
MoO3 (610 to 718 °C, or 1130 to 1330 °F),
Na2O-MoO3 (499 °C, or 930 °F), and Cu2O-
MoO2-MoO3 (470 °C, or 880 °F) (Ref 123).
Other oxides, such as PbO and V2O5, can also
cause metals or alloys to suffer catastrophic
oxidation in air at intermediate temperatures
200 µm of 640 to 930 °C (1200 to 1700 °F) (Ref 124).
PbO and V2O5 melt at 888 and 690 °C (1630
Fig. 3.75 Specimen cross section of IC-50 after oxidation
and 1270 °F), respectively. The deleterious
testing in air at 1150 °C (2100 °F) for 1008 h with
168 h cycles. Samples were descaled prior to metallographic effect of lead oxide was believed to be related to
mounting. Source: Ref 55
exhaust-valve failures in gasoline engines.
Gasoline additives were a primary source for lead
in static air to 900 °C (1650 °F). The same compounds. Vanadium is an important con-
alloy exhibited good oxidation resistance when taminant in residual or heavy fuel oils. Therefore,
heated to the same temperature in flowing V2O5 plays a significant role in oil ash
air. They postulated that the rapid oxidation corrosion in oil-fired boilers, which is discussed
was due to the accumulation of gaseous MoO3 in Chapter 11.
on the metal surface. The oxidation is accelerated Sawyer (Ref 124) indicated that accelerated
by the thermal dissociation of MoO3 into MoO2 oxidation of Type 446 stainless steel in the
and O. Meijering and Rathenau (Ref 120), presence of lead oxide can proceed at tempera-
Brasunas and Grant (Ref 121), and Brennor (Ref tures where the liquid phase does not exist.
122), however, attributed this to the presence Experiments carried out by Brasunas and Grant
of a liquid oxide phase. The MoO3 oxide (Ref 125) showed that 16Cr-25Ni-6Mo alloy
melts at about 795 °C (1463 °F). The 19Cr-9Ni specimens placed adjacent to, but not in contact
steel suffered catastrophic oxidation attack in with, 0.5 g samples of WO3 oxides suffered
the presence of MoO3 at 770 °C (1420 °F). accelerated oxidation attack when tested in
This temperature was very close to the eutectic air at 868 °C (1585 °F), which is well below
temperature of MoO2-MoO3-Cr2O3, which was the melting point of WO3 (i.e., 1473 °C, or
reported to be 772 °C (1420 °F) (Ref 123). 2683 °F).
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:46PM Plate # 0 pg 62

62 / High-Temperature Corrosion and Materials Applications

Because molybdenum and tungsten are very the service condition. Once the chromium supply
important solid-solution-strengthening alloying drops below the critical level to maintain and
elements, many superalloys containing either reheal the chromium oxide scale, breakaway
or both have been developed since Leslie and oxidation is initiated followed by rapid growth of
Fontana first observed catastrophic oxidation oxides of base metals, such as iron, nickel, or
in 1948. Some of these alloys, including alloys cobalt oxides. Detailed discussion on the effect
X, R-41, 625, 617, S, 230, 25, and 188, have of chromium on the oxidation behavior and
been used successfully in service for the hot breakaway oxidation of various alloys is pre-
section of gas turbine engines. Many of them sented. Effects of other minor elements on the
have also been used successfully in heat treating, oxidation behavior of chromia formers are also
chemical processing, and related industries. The discussed.
most effective way to alleviate the potential A relatively small number of commercial
catastrophic oxidation problem is to avoid a high-temperature alloys rely on aluminum for
stagnant condition for the gaseous atmosphere. forming a protective aluminum oxide scale to
resist oxidation attack at very high temperatures.
This group of alloys that form aluminum oxide
3.5 Oxidation/Nitridation in Air and scales are typically referred to as alumina
Combustion Atmospheres formers. The oxidation behavior of alumina
formers that include Fe-Cr-Al, Ni-Cr-Al, and
Nitridation can take place in conjunction with ODS alloys is presented. The effect of aluminum
oxidation in oxidizing atmospheres including air as well as yttrium and sulfur on the oxidation
and combustion environments. This can result in resistance of alumina formers is discussed.
significant internal nitridation penetration and Discussion also includes oxidation under
affect the mechanical properties of the alloy. This high-velocity gas streams, oxidation of thin foils,
topic is covered in detail in Chapter 4 “Nitrida- effect of surface depletion of chromium in aus-
tion.” tenitic stainless steels, effect of water vapor, and
catastrophic oxidation (oxidation under molten
oxides). Most oxidation data were generated at
3.6 Summary 980 to 1200 °C (1800 to 2200 °F). However,
many industrial applications are in the tempera-
Oxidation data for carbon and low-alloy steels, ture range of 650 to 980 °C (1200 to 1800 °F),
ferritic stainless steels, austenitic stainless steels, which are below the test temperatures at which
Fe-Cr-Ni alloys with 20-25Cr/30-40Ni, Fe-Cr-Al most data were generated. More long-term oxi-
alloys, and superalloys including ODS alloys are dation data need to be generated at 650 to 980 °C
presented. The data are presented in such a way (1200 to 1800 °F) for stainless steels, Fe-Ni-Cr
for readers to make comparisons between alloys and some simple Ni-Cr alloys to provide a more
within the same alloy group or between alloys reliable database at intended application tem-
from different alloy groups. The chapter focuses peratures.
on long-term oxidation behavior. Oxidation
resistance of some nickel aluminides is also
compared with that of several commercial REFERENCES
nickel-base superalloys.
A large portion of commercial high- 1. G.Y. Lai, unpublished data, 2004
temperature alloys relies on chromium for 2. N. Birks and G.H. Meier, Introduction to
forming a protective chromium oxide scale to High Temperature Oxidation of Metals,
resist oxidation attack. The temperature range in Edward Arnold Ltd., London, 1983
which the chromium oxide scale is effective in 3. R.C. Weast, Ed., Physical Constants of
providing adequate oxidation resistance varies Inorganic Compounds, Handbook of
from 540 to 1090 °C (1000 to 2000 °F). This Chemistry and Physics, 65th ed., The
group of alloys is frequently referred to as Chemical Rubber Company, 1984, p B68–
chromia formers. An adequate supply of chro- B161
mium from the alloy interior to the metal surface 4. C.E. Ramberg, P. Beatrice, K. Kurokawa,
to form and maintain a continuous, protective and W.L. Worrell, Mater. Res. Soc. Proc.,
chromium oxide scale is necessary for an alloy to Vol 322, C.L. Briant, J.J. Petrovic, B.P.
maintain its oxidation-resistant capability under Bewlay, A.K. Vasudevan, and H.H. Lipsitt,
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:46PM Plate # 0 pg 63

Chapter 3: Oxidation / 63

Ed., Materials Research Society, 1994, 19. A. Grodner, Weld. Res. Counc. Bull.,
p 243 No. 31, 1956
5. D. Caplan and M. Cohen, The Volatiliza- 20. M. Walter, M. Schutze, and A. Rahmel,
tion of Chromium Oxide, J. Electrochem. Behavior of Oxide Scales on 12Cr-1Mo
Soc., Vol 108 (No. 5), 1961, p 438 Steel During Thermal Cycling, Oxid. Met.,
6. E.A. Gulbranson and S.A. Jansson, in Vol 39 (No. 5/6), 1993, p 389
Heterogenous Kinetics at Elevated Tem- 21. S.B. Lasday, Ind. Heat., March 1979, p 12
peratures, G.R. Belton and W.L. Warrell, 22. O.D. Sherby, Acta Metall., Vol 10, 1962,
Ed., Plenum Press, 1970, p 181 p 135
7. H.C. Graham and H.H. Davis, Oxidation/ 23. H.E. Eiselstein and E.N. Skinner, STP 165,
Vaporization Kinetics of Cr2O3, J. Am. ASTM, 1954, p 162
Ceram. Soc., Vol 54 (No. 2), 1971, p 89 24. A. Moccari and S.I. Ali, Br. Corros. J.,
8. C.A. Stearns, F.J. Kohl, and G.C. Vol 14 (No. 2), 1979, p 91
Fryburg, Oxidative Vaporization Kinetics 25. A.S. Brasunas, J.T. Gow, and O.E. Harder,
of Cr2O3 in Oxygen from 1000 to 1300 °C, Proc. ASTM, Vol 46, 1946, p 870
J. Electrochem. Soc.: Solid-State Sci. 26. S. Kado, T. Yamazaki, M. Yamazaki,
Technol., Vol 121 (No. 7), 1974, p 945 K. Yoshida, K. Yabe, and H. Kobayashi,
9. H. Asteman, J.E. Svensson, L.G. Johans- Trans. Iron Steel Inst. Jpn., Vol 18 (No. 7),
son, and M. Norell, Indication of Chro- 1978, p 387
mium Oxide Hydroxide Evaporation 27. H.T. Michels, Met. Eng. Q., Aug 1974, p 23
During Oxidation of 304L at 873 K in the 28. J. Douthett, Heat Treating of Stainless
Presence of 10% Water Vapor, Oxid. Met., Steels, in Heat Treating, Vol 4, ASM
Vol 52 (No. 1/2), 1999, p 95 Handbook, ASM International, 1991,
10. N. Birks and F.S. Pettit, Environmental p 769
Effects During Application of Materials at 29. J.F. Grubb, in Proc. International Conf. on
Temperatures Above 1200 °C, Mater. Sci. Stainless Steels, Iron and Steel Institute of
Eng., A143, 1991, p 187 Japan, Tokyo, 1991, p 944
11. M. Danielewski, Kinetics of Gaseous 30. G.Y. Lai, Unpublished data, 2005
Corrosion Processes, in Corrosion: 31. W.E. Ruther and S. Greenberg, J. Electro-
Fundamentals, Testing and Protection, chem. Soc., Vol 111, 1964, p 1116
Vol 13A, ASM Handbook, ASM Interna- 32. J.C. Kelly, Today’s Versus Yesterday’s
tional, 2003, p 97 Stainless Steels, Mater. Perform., March
12. W.R. Patterson, in Designing for Auto- 1992
motive Corrosion Prevention, Proceedings 33. “Heat Resistant Alloy Specifications and
P-78 (Troy, MI), Nov 8–10, 1978, Society Operating Data,” Bulletin 150, Rolled
of Automotive Engineers, p 71 Alloys, Temperance, MI, 1997
13. J.B. Vrable, R.T. Jones, and E.H. Phelps, 34. “RA 353MA alloy,” Bulletin No. 1069,
“The Application of Corrosion-Resistant Rolled Alloys, Temperance, MI
High-Strength Low-Alloy Steels in the 35. “Avesta 353MA,” Information 9270,
Chemical Industry,” presented at Fall Avesta AB, S-77480 Avesta, Sweden
Meeting, American Society for Metals, 36. G.R. Rundell, Evaluation of A High Sili-
Chicago, 1977 con-Aluminum Alloy for Waste Incinera-
14. R.T. Jones, in Process Industry Corrosion, tion and Fluidized Bed Combustion
NACE, 1986, p 373 Systems, Performance of High Tempera-
15. R.C. John, Compilation and Use of ture Materials in Fluidized Bed Combus-
Corrosion Data for Alloys in Various High- tion Systems and Process Industries, P.
Temperature Gases, Paper No. 73, Corro- Ganesan and R.A. Bradley, Ed., ASM
sion/99, NACE International, 1999 International, 1987, p 193
16. A.W. Zeuthen, Heating, Piping and Air 37. “Kanthal® Electrical Resistance and High
Conditioning, Vol 42 (No. 1), 1970, p 152 Temperature Alloys,” Kanthal Alloys Data
17. H.E. McGarrow, Ed., The Making, Shaping Sheet, Sandvik, Sweden
and Treating of Steel, United States Steel 38. R. Berglund and B. Jonsson, New Powder
Corp., 1971, p 1136 Metallurgical Fe-Cr-Al Alloy for High
18. I.G. Wright, in Corrosion, Vol 13, Metals Temperature and Corrosion Resistance in
Handbook, ASM International, 1987, p 97 Furnace Uses, Ind. Heat., Oct 1989, p 21
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:46PM Plate # 0 pg 64

64 / High-Temperature Corrosion and Materials Applications

39. “Allegheny Ludlum Data Sheet,” Alle- 50. “Inconel Alloy 601 Brochure,” Huntington
gheny Ludlum, Pittsburgh, PA, 1990 Alloys, Inc., 1969
40. P.T. Moseley, K.R. Hyde, B.A. Bellamy, 51. P. Ganesan, G.D. Smith, and C.S. Tassen,
and G. Tappin, The Microstructure of the Performance of A New Alloy in High
Scale Formed During the High Tempera- Temperature Service, Paper No. 234, Cor-
ture Oxidation of a Fecralloy® Steel, rosion/93, NACE, 1993
Corros. Sci., Vol 24 (No. 6), 1984, p 547 52. D.C. Agarwal, ThyssenKrupp VDM
41. E. Tsuzi, The Role of Yttrium on the Oxide unpublished data
Adherence of Fe-24Cr Base Alloys, Metall. 53. D.C. Agarwal and U. Brill, Performance of
Trans. A, Vol 11A, Dec 1980, p 1965 Alloy 602CA (UNS N06025) in High
42. T.A. Ramanarayanan, M. Raghavan, Temperature Environments up to 1200 °C,
and R. Petkovic-Luton, Metallic Yttrium Paper No. 521, Corrosion/2000, NACE
Additions to High Temperature Alloys: International, 2000
Influence on Al2O3 Scale Properties, Oxid. 54. G.Y. Lai, J. Met., Vol 37 (No. 7), July 1985,
Met., Vol 22 (No. 3/4), 1984, p 83 p 14
43. J.L. Pandey, S. Prakash, and M.L. Mehta, 55. G.Y. Lai, unpublished results, Haynes
Effect of Zirconium Concentration on High International, Inc., 1988
Temperature Cyclic Oxidation Behavior of 56. “Haynes Alloy No. 214,” H-3008B, Haynes
Fe-15Cr-4Al at 1150 °C, J. Electrochem. International, Inc. Kokomo, IN
Soc.: Solid-State Sci. Technol., Vol 135 57. N. Birks and F.S. Pettit, Environmental
(No. 1), Jan 1988, p 209 Effects During Application of Materials at
44. J.L. Pandey, S. Prakash, and M.L. Mehta, Temperatures above 1200 °C, Mater. Sci.
Effects of Varying the Zirconium Con- Eng., Vol A143, 1991, p 187
centration and 1 wt.% Y on High Tem- 58. G.Y. Lai, Sulfidation-Resistant Co-Cr-Ni
perature Oxidation of Fe-15wt.%Cr-4wt.% Alloys with Critical Contents of Silicon and
Al Alloy under Isothermal and Cyclic Cobalt, U.S. Patent No. 4711763, Dec 1987
Conditions, J. Less-Common Met., Vol 59. G.Y. Lai, Meeting the Challenge of Mate-
159, 1990, p 23 rials Development for Coal Combustion
45. J.S. Benjamin, Metall. Trans., Vol 1, 1970, Plants, Mater. High Temp., Vol 11
p 2943 (No. 1–4), 1993, p 143
46. M. Lundberg, L.-P. Bergmark, and M. 60. W. Crawford, in Proc. Conf. Frontiers
Ramberg, Mechanical and Chemical of High Temperature Materials II, London,
Properties of 353MA—A Seamless Tube Inco Alloys International, May 1983, p 272
for High-Temperature Petrochemical 61. R.F. Singer, in Proc. Conf. Frontiers of
Applications, 1999 Stainless Steel World High Temperature Materials II, London,
Conf. Proc., Book 2, KCI Publishing BV, Inco Alloys International, May 1983, p 336
Zutphen, The Netherlands, 1999, p 563 62. Superalloys Source Book, M.J. Donachie,
47. J.C. Kelly and J.D. Wilson, Oxidation Jr., Ed., American Society for Metals, 1984
Rates of Some Heat Resistant Alloys, 63. C.T. Sims, in Proc. Fifth International
Heat-Resistant Materials II: Conf. Proc. Symposium on Superalloys (Seven
Second International Conference on Heat- Springs, Champion, PA), Metallurgical
Resistant Materials, K. Natesan, P. Gane- Society of AIME, 1984, p 399
san, and G. Lai, Ed., ASM International, 64. W. Betteridge and W.W.K. Shaw, Mater.
1995, p 53 Sci. Technol., Vol 3, 1987, p 682
48. P. Ganesan, G.D. Smith and C.S. Tassen, 65. B.H. Kear and E.R. Thompson, Science,
Mechanical Properties and Corrosion Vol 208, May 23, 1980, p 847
Resistance of Incoloy Alloy 803, Applica- 66. M.J. Donachie and S.J. Donachie, Super-
tions and Materials Performance: Proc. alloys: A Technical Guide, 2nd ed., ASM
Nickel-Cobalt 97 International Symposium, International, 2002
F.N. Smith, J.F. McGurn, G.Y. Lai, and 67. C.A. Barrett, in Proc. Conf. Environmental
V.S. Sastri, Ed., The Metallurgical Society Degradation of Engineering Materials,
of CIM, Montreal, Canada, 1997, p 97 M.R. Louthan, Jr. and R.P. McNitt, Ed.,
49. M.A. Harper, J.E. Barnes, and G.Y. Lai, Virginia Polytechnic Institute, 1977, p 319
Paper No. 132, Corrosion/97, NACE 68. M.F. Rothman, Cabot Corporation internal
International, 1997 report, 1985
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:46PM Plate # 0 pg 65

Chapter 3: Oxidation / 65

69. H. Nagai, M. Okaboyashi, and H. Mitani, No. 522, Corrosion/2000, NACE Interna-
Trans. Jpn. Inst. Met., Vol 21, 1980, p 341 tional, 2000
70. G.M. Kim, E.A. Gulbranson, and G.H. 82. G.Y. Lai, unpublished results, Haynes
Meier, in Proc. Conf. Fossil Energy International, Inc., 1988
Materials Program, ORNL/FMP 87/4, 83. B. Hicks, Mater. Sci. Technol., Vol 3,
May 19–21, 1987, R.R. Judkins, Ed., Oak Sept 1987, p 772
Ridge National Laboratory, 1987, p 343 84. B. Gleeson, High-Temperature Corrosion
71. G.Y. Lai, unpublished results, Haynes of Metallic Alloys and Coatings, in Cor-
International, Inc., 1988 rosion and Environmental Degradation,
72. R.H. Kane, J.W. Schultz, H.T. Michels, Vol II, Materials Science and Technology,
R.L. McCarron, and F.R. Mazzotta, Ref 30 M. Schutze, Ed., Wiley-VCH, Weinheim,
of the paper by R.H. Kane in Process Germany, 2000, p 173
Industries Corrosion, B.J. Moniz and W.I. 85. G.Y. Lai, unpublished results, Haynes
Pollock, Ed., NACE, 1986, p 45 International, Inc., 1996
73. G.Y. Lai, M.F. Rothman, and D.E. Fluck, 86. B. Gleeson and M.A. Harper, The Long-
Paper No. 14, Corrosion/85, NACE, 1985 Term, Cyclic-Oxidation Behavior of
74. J.J. Barnes and S.K. Srivastava, Paper No. Selected Chromia-Forming Alloys, Oxid.
527, Corrosion/89, NACE, 1989 Met., Vol 49 (No. 3/4), 1998, p 373
75. J.J. deBarbadillo and J.J. Fischer, Disper- 87. H.E. Evans, D.A. Hilton, R.A. Holm, and
sion-Strengthened Nickel-Base and Iron- S.J. Webster, Oxid. Met., Vol 14, 1980,
Base Alloys, in Properties and Selection: p 235
Nonferrous Alloys and Special-Purpose 88. M.P. Brady, B. Gleeson, and I.G. Wright,
Materials, Vol 2, Metals Handbook, ASM Alloy Design Strategies for Promoting
International, 1990, p 943 Protective Oxide-Scale Formation, JOM,
76. R.H. Kane, G.M. McColvin, T.J. Kelly, and Jan 2000, p 16
J.M. Davidson, Paper No. 12, Corrosion/ 89. W.J. Quadakkers and K. Bongartz, The
84, NACE, 1984 Prediction of Breakaway Oxidation for
77. R.C. John, Oxidation Studies of Commer- Alumina forming ODS Alloys Using Oxi-
cial Alloys at 871–1093 °C (1600– dation Diagrams, Werkst. Korros., Vol 45,
2000 °F), in Heat-Resistant Materials 1994, p 232
II—Conf. Proc. Second International 90. I. Gurrappa, S. Weinruch, D. Naumenko,
Conference on Heat-Resistant Materials, and W.J. Quadakkers, Factors Governing
K. Natesan, P. Ganesan, and G. Lai, Ed., Breakaway Oxidation of FeCrAl-Based
ASM International, 1995, p 41. Alloys, Mater. Corros., Vol 51, 2000, p 224
78. C.E. Lowell, D.L. Deadmore, and J.D. 91. W.J. Quadakkers, Growth Mechanisms
Whittenberger, Long-Term High-Velocity of Oxide Scales on ODS Alloys in
Oxidation and Hot Corrosion Testing of the Temperature Range 1000–1100 °C,
Several NiCrAl and FeCrAl Base Oxide Werkst. Korros., Vol 41, 1990, p 659
Dispersion Strengthened Alloys, Oxid. 92. J. Klower and J.G. Li, Effects of Yttrium on
Met., Vol 17 (No. 3/4), 1982, p 205 the Oxidation Behavior of Iron-Chromium-
79. M.F. Rothman, “Oxidation Resistance of Aluminum Alloys, Mater. Corros., Vol 47,
Gas Turbine Combustion Materials,” 1996, p 545
Paper No. 85-GT-10, presented at the Gas 93. J.G. Smeggil, A.W. Funkenbusch, and
Turbine Conference (Houston, TX), March N.S. Bornstein, High Temp. Sci., Vol 20,
18–21, 1985, ASME, 1985 1985, p 163
80. J.V. Wright, “The Effects of Gas Velocity 94. A.W. Funkenbusch, J.G. Smeggil, and
and of Temperature on the Oxidative N.S. Bornstein, Met. Trans. A, Vol 16,
Response of Selected Sheet Superalloys,” 1985, p 1164
Paper No. 88-GT-281, presented at the 95. J.G. Smeggil, A.W. Funkenbusch, and
Gas Turbine and Aeroengine Congress N.S. Bornstein, Met. Trans. A, Vol 17,
(Amsterdam, The Netherlands), June 6–9, 1986, p 923
1988, ASME, 1988 96. J.G. Smeggil, Some Comments on the
81. U. Brill and T.I. Haubold, Corrosion Role of Yttrium in Protective Oxide Scale
Behaviour of Some Gas Turbine Alloys Adherence, Mater. Sci. Eng., Vol 87, 1987,
under High Velocity Burnt Fuels, Paper p 261
Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/ 31/10/2007 12:46PM Plate # 0 pg 66

66 / High-Temperature Corrosion and Materials Applications

97. P.Y. Hou and J. Stringer, Oxide Scale I—Influence of pH2 O , Mater. Corros., Vol
Adhesion and Impurity Segregation at the 53, 2002 p 247
Scale/Metal Interface, Oxid. Met., Vol 38 110. K. Segerdahl, J.E. Svensson, and L.G.
(No. 5/6), 1992, p 323 Johansson, The High Temperature Oxida-
98. P.Y. Hou, Compositions at Al2O3/ tion of 11% Chromium Steel: Part II—
FeCrAl Interfaces after High Temperature Influence of Flow Rate, Mater. Corros.,
Oxidation, Mater. Corros., Vol 51, 2000, Vol 53, 2002, p 479
p 329 111. B.A. Pint, R. Peraldi, and P.F. Tortorelli,
99. G.Y. Lai, “Several Modern Wrought The Effect of Alloy Composition on the
Superalloys for Gas Turbine Applications,” Performance of Stainless Steels in Exhaust
Paper 96-TA-030, presented at ASME Gas Environments, Paper No. 03499,
Turbo Asia ’96 (Jakarta, Indonesia), Corrosion/2003, NACE International,
Nov 5–7, 1996 2003
100. N.J. Simms, R. Newton, J.F. Norton, A. 112. B.A. Pint, private communication, 2007
Encinas-Oropesa, J.E. Oakey, J.R. 113. R.L. McCarron and J.W. Schultz, in Proc.
Nicholls, and J. Wilber, Mater. High Temp., Symp. High Temperature Gas-Metal
Vol 20 (No. 3), 2003, p 439 Reactions in Mixed Environments, AIME,
101. J. Klower, Factors Affecting the Oxidation 1973, p 360
Behaviour of Thin Fe-Cr-Al Foils, Mater. 114. C.C. Clark and W.R. Hulsizer, Superalloys
Corros., Vol 49, 1998, p 758 Development for Gas Turbines Operating
102. P.J. Maziasz, B.A. Pint, R.W. Swindeman, in the Marine Environment, Conf. Proc.,
K.L. More, and E. Lara-Curzio, “Advanced Gas Turbine Materials Conference, Naval
Stainless Steels and Alloys for High Ship Engineering Center, 1972, p 35
Temperature Recuperators,” DOE/CETC/ 115. C. Sarioglu et al., The Adhesion of
CANDRA Workshop on Microturbine Alumina Films to Metallic Alloys and
Applications (Calgary, Alberta, Canada), Coatings, Mater. Corros., Vol 51, 2000,
Jan 21–23, 2003 p 358
103. B.A. Pint, “The Effect of Water Vapor on 116. K. Onal, M.C. Maris-Sida, G.H. Meier, and
Cr Depletion in Advanced Recuperator F.S. Pettit, Water Vapor Effects on the
Alloys,” GT2005-68495, ASME Turbo Cyclic Oxidation Resistance of Alumina
Expo 2005 (Reno-Tahoe, Nevada), Forming Alloys, Mater. High Temp., Vol 20
June 6–9, 2005 (No. 3), 2003, p 327
104. B.A. Pint, Stainless Steels with Improved 117. G. Welsch and P.D. Desai, Ed., Oxidation
Oxidation Resistance for Recuperators, and Corrosion of Intermetallic Alloys,
J. Eng. Gas Turbines Power, Vol 128, Purdue University, 1996
2006, p 1 118. K. Natesan, Oxid. Met., Vol 30 (No. 1/2),
105. B.A. Pint, “The Effect of Water Vapor on 1988, p 53
Cr Depletion in Advanced Recuperator 119. W.C. Leslie and M.C. Fontana, Paper
Alloys, GT2005-68495,” ASME Turbo No. 26, 30th Annual Convention of
Expo 2005, (Reno-Tahoe, Nevada), June ASM (Philadelphia, PA), Oct 25–29,
6–9, 2005 1948
106. C.W. Tuck, M. Odgers, and K. Sachs, 120. J.K. Meijering and G.W. Rathenau, Nature,
Scaling Rates of Pure Iron and Mild Steel in Vol 165, Feb 11, 1950, p 240
Oxygen, Steam, Carbon Dioxide in the 121. A.D. Brasunas and N.J. Grant, Iron Age,
Range 850°–1000 °C, Anti-Corrosion, Aug 17, 1950, p 85
June 1966, p 14 122. S.S. Brennor, J. Electrochem. Soc., Vol 102
107. S.C. Stultz and J.B. Kitto, Ed., Steam and (No. 1), Jan 1955, p 16
Its Generation and Use, 40th ed., Babcock 123. J.H. DeVan, “Catastrophic Oxidation of
& Wilcox, 1992, p 13-2 High Temperature Alloys,” ORNL-TM-51,
108. S.C. Stultz and J.B. Kitto, Ed., Steam and Oak Ridge National Laboratory, Oak
Its Generation and Use, 40th ed., Babcock Ridge, TNn, Nov 10, 1961
& Wilcox, 1992, p T-17 124. J.W. Sawyer, Trans. TMS-AIME, Vol 221,
109. K. Segerdahl, J.E. Svensson, and 1961, p 63
L.G. Johansson, The High Temperature 125. A. de S. Brasunas and N.J. Grant, Trans.
Oxidation of 11% Chromium Steel: Part ASM, Vol 44, 1950, p 1133
Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/ 30/10/2007 3:55PM Plate # 0 pg 67

High-Temperature Corrosion And Materials Applications Copyright © 2007 ASM International®

George Y. Lai, editor, p67-96 All rights reserved.
DOI: 10.1361/hcma2007p067 www.asminternational.org

CHAPTER 4

Nitridation

4.1 Introduction Cracked ammonia (i.e., ammonia that is com-

pletely dissociated into H2 and N2) provides an
In air or combustion atmospheres containing economical protective atmosphere for processing
nitrogen, nitridation can take place under certain metals and alloys. Many bright annealing opera-
exposure conditions. In most cases, oxidation tions for stainless steels use a protective atmo-
dominates the high-temperature corrosion re- sphere consisting of N2 and H2, generated by
action. However, nitridation can take place for dissociation of ammonia. With three parts H2
some alloys when oxide scales no longer provide and one part N2 produced in cracked ammonia,
protection. The alloys that are particularly sus- nitridation is less critical for the heat treating
ceptible to oxidation/nitridation attack are those equipment.
containing strong nitride formers, such as tita- In the chemical processing industry, nitriding
nium and aluminum. Many high-temperature environments are generated by processes em-
nickel-base alloys containing both aluminum and ployed for production of ammonia, nitric acid,
titanium are strengthened by γ′ phase, Ni3(Al,Ti). melamine, and nylon 6-6 (Ref 3, 4). Ammonia is
For these alloys, nitridation by forming internal produced by reacting nitrogen with hydrogen
nitrides of aluminum and titanium can deplete the over a catalyst at temperatures of typically 500
surface layer with aluminum and titanium, thus to 550 °C (930 to 1020 °F) and pressures of 200
weakening the alloy. Under a high-velocity com- and 400 atm. Commercial processes for ammo-
bustion gas stream with severe thermal cycling, nia synthesis are discussed in detail in Ref 5. The
similar to the conditions in “flying” gas turbines converter, where the ammonia synthesis reaction
(aircraft engines), nitridation can be particularly takes place, may suffer nitridation attack. Brittle
severe in oxidation/nitridation attack. failure of the welds for the waste heat boiler of
In nitrogen-base atmospheres, such as N2 or an ammonia plant has been reported by Van der
N2-H2, metals and alloys can also suffer nitri- Horst (Ref 6). These tube-to-tube sheet welds
dation attack. This type of atmosphere is often were high-nickel alloy 182, which suffered
used as a protective atmosphere in heat treating nitridation attack.
and sintering operations. Molecular nitrogen can Production of nitric acid involves the oxida-
be severely nitriding for many metals and alloys, tion of ammonia over a platinum gauze catalyst
particularly when temperatures are sufficiently at temperatures of about 900 °C (1650 °F)
high. (Ref 5). The catalyst grid support structure and
Ammonia (NH3) is a commonly used nitriding other processing components in contact with
gas for case hardening at temperatures from ammonia may also be susceptible to nitridation
500 to 590 °C (925 to 1100 °F) (Ref 1). Furnace attack. Figure 4.1 shows the nitrided structure of
equipment and components repeatedly subjected a nickel-base alloy catalyst grid support after
to these service conditions frequently suffer two years of service in a nitric acid plant.
brittle failures as a result of nitridation attack.
Carbonitriding is another important method of
case hardening that produces a surface layer of 4.2 Thermodynamic Considerations
both nitrides and carbides. The process is typi-
cally carried out at 700 to 900 °C (1300 to 1650 ° When metal is exposed to nitrogen gas at
F) in ammonia, with additions of carbonaceous elevated temperatures, nitridation proceeds
gases, such as CH4 (Ref 2). Thus, the heat treat according to:
retort, fixtures, and other furnace equipment are
subject to both nitridation and carburization. 1=2N2 (gas)=[N] (dissolved in metal) ð4:1Þ
Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/ 30/10/2007 3:56PM Plate # 0 pg 68

68 / High-Temperature Corrosion and Materials Applications

[%N] = k(pN2 )1=2 ð4:2Þ nitrogen molecules and absorption of nitrogen

atoms by the metal following Reaction 4.1.
where k is the equilibrium constant and pN2 is the Nitridation often takes place in the metal at the
partial pressure of N2 in the atmosphere. vicinity of cracks developed under creep condi-
In nitrogen-base atmospheres, the nitriding tions in air or N2-containing combustion atmo-
potential is proportional to (pN2 )1=2 . Increasing spheres. In this case, oxides are often associated
the nitrogen partial pressure (or nitrogen con- with the crack under creep deformation in air.
centration) increases the thermodynamic poten- Oxidation consumes the oxygen molecules from
tial for nitridation. Molecular nitrogen is less air, which penetrates into the crack, thus deplet-
aggressive than ammonia in terms of nitridation ing oxygen and increasing nitrogen potential
of metals. However, when metal is heated to (or concentration) around the crack. As a result,
excessively high temperatures (e.g., 1000 °C nitridation takes place in the vicinity of the creep
(1830 °F) or higher), nitridation by molecular crack.
nitrogen can become a serious material issue. When the environment is ammonia (NH3) or
Under oxidation/nitridation conditions, nitro- contains NH3, metals or alloys may undergo rapid
gen molecules permeate through cracks and nitridation reactions. It is precisely for this reason
pores and reach the metal underneath the oxide that NH3 is frequently used for case hardening.
scales when the oxide scale is no longer protec- Ammonia is metastable and dissociates into
tive. Nitridation then proceeds by dissociation of molecular N2 and molecular H2 when heated to
elevated temperatures. Once NH3 is completely
dissociated into N2 and H2, the nitriding potential
is defined by Reaction 4.3. To increase nitrogen
absorption by steel, molecular NH3 should be
allowed to dissociate on the steel surface, thus
allowing dissociated atomic nitrogen to be dis-
solved at the metal surface (Ref 7, 8). Thus, to
increase nitridation reactions, it is necessary to
bring as much fresh, uncracked NH3 as possible
in contact with the surface of the metal to mini-
mize the production of molecular nitrogen.
At temperatures below 600 °C (1110 °F) and at
high gas flow rates, the production of nitrogen
is minimized and the nitrogen solubility at the
100 µm
surface of iron is determined by (Ref 8):
(a)
NH3 $ 3=2 H2 +[N] (dissolved in Fe) ð4:3Þ

[%N]=k(pNH3 =(pH2 ))3=2 ð4:4Þ

where k is the equilibrium constant, and pNH3 ,

and pH2 are partial pressures of NH3 and H2,
respectively.
The nitriding potential is proportional to
pNH3 =( pH2 )3=2 . Increasing ammonia partial
pressure (or concentration) in the atmosphere
increases the thermodynamic potential for nitri-
dation. When nitrogen in the metal exceeds its
100 µm solubility limit, nitrides will then precipitate
(b) out. The nitrides of important alloying elements
for engineering alloys are tabulated in Table 4.1
Fig. 4.1 Alloy X (Ni-22Cr-18.5Fe-9Mo-0.6W) catalyst grid (Ref 9). For iron, nickel, and cobalt, three im-
support structure bar after 2 years of service in a nitric
acid plant. (a) Internal nitride precipitates (about 20 mils in depth) portant alloy bases for high-temperature alloys,
containing mainly chromium-rich nitrides along with some car- only iron forms stable nitrides. No nitrides of
bides formed during thermal aging. (b) Microstructure in the
unaffected interior containing mainly carbides due to thermal
nickel and cobalt have been reported. The
aging at the service temperature relative stabilities among various nitrides can
Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/ 30/10/2007 3:56PM Plate # 0 pg 69

Chapter 4: Nitridation / 69

be compared in terms of their free energies of alloys with nickel ranging from 0 to 35%
formation, as illustrated in Fig. 4.2 (Ref 10). (Ref 19). In Fe-18Cr-Ni-N system, increasing
Physical-chemical properties of some of these nickel reduces the solubility of nitrogen, as
nitrides can be found in Ref 9. The types of shown in Fig. 4.3 (Ref 11). It is also shown in
nitrides that are likely to form in the alloy can Table 4.2 that the solubility of nitrogen in stain-
be predicted by examining the phase-stability less steels increased with increasing temperature.
diagram. Figure 4.3 shows a phase-stability Alloys with higher nitrogen solubilities generally
diagram of Fe-18Cr-Ni-N system at 900 °C exhibit less resistance to nitridation attack.
(1650 °F), indicating phase regions of Cr2N
in γ (or α+γ) phase as a function of nickel +10 e 4N
and nitrogen contents (Ref 11). Nitride phases 2F
0

3
formed in alloys are also dependent on nitrogen

H
N
N

2
partial pressure ( pN2 ), as shown in Fig. 4.4 for –10 2M
o2
Ni-Cr-N system at 1000 °C (1830 °F) (Ref 12). N
–20 Cr
2
Nitrogen solubility in the alloy is important in
affecting the nitridation resistance of the alloy. –30 r 2N N4
2C
Si 3
Table 4.2 summarizes some nitrogen solubility –40 1/
2
data for iron, stainless steels, and nickel alloy
–50 bN
(Ref 13–18). Iron and stainless steels exhibit 2
N

significantly higher nitrogen solubility than a

∆G°, Kcal
–60
BN
nickel-base alloy (Ni-20Fe). Nickel has been VN 2
–70 2

2
3N
N2
found to decrease the nitrogen solubility in Fe-Ni
a
g3
–80 C M

N
AI
N
–90 Ta

2
Table 4.1 Nitrides of important alloying 2
elements for engineering alloys –100
Element Nitrides
–110
N
Iron Fe4N Fe2N Ti
2
Chromium CrN Cr2N –120
N
Molybdenum MoN Mo2N Zr
Tungsten WN W 2N –130 2
Aluminum AlN 2M + N2 = 2MN
Titanium TiN Ti2N –140
Niobium NbN Nb2N Nb4N3
Tantalum TaN Ta2N Ta3N5 0 500 1000 1500 2000
Zirconium ZrN
Hafnium HfN Hf3N2 Hf4N3 Temperature, °C
Silicon Si3N4
Vanadium VN V 2N
Boron BN
Fig. 4.2 Standard free energy of formation for selected
nitrides. Source: Ref 10
Manganese Mn4N Mn2N Mn3N
Magnesium Mg3N
Source: Ref 9
102
1000 °C
10 γ + CrN (1830 °F)
N2 partial pressure, bar

1
20
γ + Cr2N
0.1
Nickel, %

γ γ + Cr2N
10 –2 γ
10 γ+π
10 –3
α+γ γ+α
α + γ + Cr2N 10 –4
0 α 0 10 20 30 40 50 60
0 0.1 0.2 0.3 0.4
Cr concentration, wt %
Nitrogen, %
Fig. 4.4 Phase stability diagram for the Ni-Cr-N system as a
Fig. 4.3 Phase stability diagram for Fe-18Cr-Ni-N system at function of N2 partial pressure at 1000 °C (1830 °F).
900 °C (1650 °F). Source: Ref 11 Source: Ref 12
Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/ 30/10/2007 3:56PM Plate # 0 pg 70

70 / High-Temperature Corrosion and Materials Applications

Table 4.2 Nitrogen solubility in metals and alloys underneath the external oxide scale and internal
Temperature, oxides in alloy 601 exposed to a furnace oxi-
Nitrogen, wt% Metal or alloy °C (°F) Ref dizing atmosphere for about 4 to 5 years at
0.06 α-Fe 502 (936) 13 temperatures probably between 760 and 870 °C
0.26 max γFe γ region of Fe-C 14
0.02 (at pN2 : 1 atm) Fe-10%Ni 1000 (1832) 15 (1400 and 1600 °F), as shown in Fig. 4.5.
0.125 Type 304 538 (1000) 16 Severe thermal cycling that causes cracking
0.177 Type 304 593 (1100) 17
0.190 Type 304 927 (1700) 17
and spalling of oxide scales can also result in
0.258 Type 304 954 (1749) 17 severe internal nitridation. Han and Young (Ref
0.281 Type 304 981 (1800) 17 23) conducted cyclic oxidation tests by heating
0.18 Fe-18Cr-12Ni-2Ti 985 (1805) 18
0.18 Fe-18Cr-12Ni-2Ti 1040 (1905) 18 the specimens to 1100 °C (2010 °F) in still air
0.21 Fe-18Cr-12Ni-2Ti 1093 (2000) 18 for 1 h followed by cooling to room temperature
0.26 Fe-18Cr-12Ni-2Ti 1150 (2100) 18 for 15 min then repeating the cycle again for 260
0.26 Fe-18Cr-12Ni-2Ti 1210 (2210) 18
0.0001 (at pN2 : 1 atm) Ni-20%Fe 1000 (1832) 15 cycles. The alloys investigated were Ni-24 to
38%Cr-14 to 25%Al. The specimens suffered
severe oxide scale spallation. The internal nitri-
dation attack was found to be extensive, and the
4.3 Internal Nitridation in Oxidizing nitridation zone consisted of AlN beneath Cr2O3
and Al2O3, then AlN+Cr2N, and then AlN in the
Environments deepest region (Ref 23).
Douglas (Ref 24) indicated that the diffusivity
In air or oxidizing combustion environments,
of nitrogen appears to be two orders of magni-
oxidation usually dominates high-temperature
tude greater than that of oxygen in nickel or
corrosion reactions. However, under certain
nickel alloys. Table 4.3 summarizes the diffusion
conditions, alloys can suffer internal nitridation
coefficients of nitrogen in nickel and iron alloys
attack along with oxidation. Internal nitridation
compared with those of oxygen and carbon in
attack, when it occurs, can penetrate farther
nickel, based on the diffusivity data from Rubly
into the metal interior than oxidation, thus sig-
and Douglas (Ref 25, 26), Grabke and Peterson
nificantly affecting the creep-rupture behavior
(Ref 27), Park and Alstetter (Ref 28), and Gruzin
of the alloy by accelerating the creep crack
et al. (Ref 29). The diffusivity of nitrogen is also
growth. Discussion of internal nitridation under
on the same order of magnitude as that of carbon
no external stresses and under creep conditions
as shown in Table 4.3. It is thus not surprising to
is presented in sections 4.3.1 and 4.3.2.
find internal nitrides were advancing in front of
internal oxides.
4.3.1 Internal Nitridation in Air under
No External Stress
4.3.2 Internal Nitridation at Creep Cracks
Some high-temperature alloys that contain in Air Environment
strong nitride formers such as aluminum and
titanium can suffer internal nitridation even in air During creep testing in air, extensive internal
environments. In an air oxidation study for two nitridation can develop in the vicinity of cracks.
nickel-base alloys, IN939 (Ni-22Cr-20Co-3.8Ti- Brickner et al. (Ref 30) found that types 302,
1.4Al-2W-1Nb-1.3Ta) and IN738LC (Ni-16Cr- 304, and 310 stainless steels showed significant
9Co-3.5Ti-3.3Al-1.8Mo-1Nb-1.8Ta) at 700, nitridation after creep-rupture testing in air at
900, and 1100 °C (1290, 1650, and 2010 °F). 870 °C (1600 °F) in less than 1000 h. Acicular
Litz et al. (Ref 20) observed internal titanium nitrides (believed to be chromium nitrides) in
nitrides (needle shape) formed in front of internal a Widmanstätten pattern were found to form
aluminum oxides that formed underneath the extensively in the vicinity of microcracks, as
external oxide scales. In an oxidation study of shown in Fig. 4.6 (Ref 30). Extensive nitridation
alloy 800HT (Fe-21Cr-32Ni-0.5Al-0.5Ti) in air was confirmed by the chemical analysis of the
at 980 °C (1800 °F) for about 2 years (720 days), tested specimens for nitrogen, which showed
Harper et al. (Ref 21) observed Widmanstätten the nitrogen content was increased from about
acicular chromium-rich nitrides along with 0.058% before testing to 0.30 to 0.53% after
aluminum nitrides that formed below chromium- creep-rupture testing (Ref 30). Extensive internal
rich oxides. Lai (Ref 22) observed internal nitrides were also observed in the vicinity of
aluminum nitrides (needle shape) that formed creep cracks in alloy 253MA (Fe-21Cr-11Ni)
Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/ 30/10/2007 3:56PM Plate # 0 pg 71

Chapter 4: Nitridation / 71

(a)
0.1 mm

(c)
10 µm

(b)
0.0010 in
Fig. 4.5 Formation of internal aluminum nitrides beneath external oxide scales and internal oxides in alloy 601 after exposing to a
furnace oxidizing atmosphere for approximately 4 to 5 years in a temperature range of 760 to 870 °C (1400 to 1600 °F). (a)
Optical micrograph showing the external oxide scales and the internal oxides, and then the chromium denuded zone immediately below,
followed by internal nitrides underneath the denuded zone. (b) Optical micrograph at higher magnification showing internal nitrides. (c)
SEM (backscattered electron image) showing internal aluminum nitrides and the EDX analysis of nitrides. Results of the semiquantitative
EDX analysis (at.%) on internal aluminum nitrides are summarized as:
Phase 1 41.5% Al, 24.7% Ni, 10.6% Cr, 6.8% Fe, 5.5% Ti, and 10.0% N
Phase 2 58.0% Al, 13.1% Ni, 6.1% Cr, 3.6% Fe, 0.8% Ti, and 17.3% N

Table 4.3 Diffusion coefficients of nitrogen, oxygen, and carbon.

Diffusion coefficient, cm2/s
Temperature, °C (°F) N in Ni-Alloys(a) N in Fe-20Ni(b) O in Ni(c) C in Ni(d)
700 (1290) 9.5 × 10−9 to 2.3 × 10−8 1.17 × 10−8 7.4 × 10−11 3.19 × 10−9
800 (1470) 3.2 × 10−8 to 8.5 × 10−8 3.86 × 10−8 5.05 × 10−10 1.47 × 10−8
900 (1650) 1.4 × 10−7 to 4.0 × 10−7 1.47 × 10−7 2.3 × 10−9 0.55 × 10−7
(a) Ref 25, 26. (b) Ref 27. (c) Ref 28. (d) Ref 29

after creep-rupture testing at 900 °C (1652 °F) (Ref 32), and Cr2N and AlN in 800H creep-crack
for 11,800 h in air (Ref 31), in 800H after growth specimens (Ref 33).
creep-rupture testing at 900 to 1000 °C (1650 to Hoffman and Lai (Ref 34) investigated an
1830 °F) in air (Ref 32), and in alloy 800H alloy 800HT pigtail that suffered cracking after
during the creep crack growth testing at 1000 °C about 7.5 years of service in a hydrogen reformer.
(1830 °F) (Ref 33). The nitrides identified were The pigtail section, which was exposed to air at
Cr2N in 253MA (Ref 31), Cr2N (major) and CrN approximately 850 °C (1565 °F) at the outside
(minor) in 800H creep-ruptured specimens of the reformer furnace, was found to show
Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/ 30/10/2007 3:56PM Plate # 0 pg 72

72 / High-Temperature Corrosion and Materials Applications

extensive blocky precipitates along grain the bend section of the pigtail from another
boundaries and acicular precipitates in the matrix hydrogen reformer (Ref 34).
in the vicinity of cracks at the tube outer-diameter When creep cracks initially develop at the
side (exposed to air). Samples were cut from metal surface during creep testing in air, oxi-
this pigtail section and solution-annealed in a dation occurs at the crack surface including
furnace for 1 h at 1093, 1149, and 1204 °C the crack tip. The oxide scales formed on the
(2000, 2100, and 2200 °F), respectively. Micro- crack surface become nonprotective due to creep
structural examination of these samples indicated deformation, thus causing the oxygen potential
that both blocky, grain-boundary phases and to decrease significantly with concurrent increase
acicular phases in the matrix remained in the in nitrogen potentials at the oxide/metal inter-
microstructure and were not put back into sol- face. As a result, nitrogen is absorbed by the
ution, suggesting those phases were nitrides metal and is diffused into the metal in the vicinity
instead of carbides. Also, chemical analysis of of cracks to form internal nitrides.
the samples from the pigtail indicated that carbon
content remained about the same as that of the
material before service (about 0.07%), while
nitrogen content was about 0.27 wt% with a
nominal nitrogen content of about 0.02% prior to
service. The process gas in the tube contained
essentially no nitrogen (typically about 0.01%).
Thus, the nitrogen ingress into the tube was pri-
marily from air from the outside diameter side
of the pigtail. Using scanning electron micro-
scopy with energy-dispersive x-ray spectroscopy
(SEM/EDX) analysis, acicular phases were
found to be enriched in aluminum, while blocky
phases were enriched in chromium; the former
was believed to be aluminum nitride and the
latter chromium nitride. Figure 4.7 shows the
acicular aluminum nitrides and blocky chromium
nitrides that remained in the microstructure after 100 µm
solution annealing at 1150 °C (2100 °F) for 1 h
for the sample from the straight section of the Fig. 4.7 Acicular aluminum nitrides and blocky chromium
nitrides, which formed in the vicinity of the creep
pigtail (Ref 34). Figure 4.8 shows extensive cracks in alloy 800HT pigtail in a hydrogen reformer, were not
nitride formation in the vicinity of creep cracks in dissolved into solution after the sample was resolution annealed at
1150 °C (2100 °F) for 1 h. Source: Ref 34

100 µm
Fig. 4.6 Acicular nitrides (believed to be chromium nitrides)
in a Widmanstätten pattern formed in the vicinity of Fig. 4.8 Extensive aluminum and chromium nitrides formed
creep cracks in Type 302SS after creep-rupture testing at 870 °C in the vicinity of creep cracks in the bend section
(1600 °F) in less than 1000 h. Original magnification, 500×. of an alloy 800H pigtail in another hydrogen reformer. Source:
Source: Ref 30 Ref 34
Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/ 30/10/2007 3:56PM Plate # 0 pg 73

Chapter 4: Nitridation / 73

4.3.3 Oxidation and Nitridation in high-velocity combustion gas stream with about
Combustion Atmospheres 0.3 Mach (100 m/s) was generated. Specimens
were loaded in a carousel specimen holder
High-temperature alloys that are exposed to a that rotated at 30 rpm during testing to ensure all
high-velocity, oxidizing combustion gas stream the specimens were subjected to the same test
at high temperatures are susceptible to internal conditions. Furthermore, the specimens were
nitridation attack. In investigating transition duct subjected to severe thermal cycling once every
component failures in a land-based gas turbine, 30 min by lowering the carousel from the test
Swaminathan and Lukezich (Ref 35) observed chamber followed by rapid fan-air cooling to
that alloy 617 (Ni-22Cr-12.5Co-9Mo-1.2Al) had below 260 °C (500 °F) for 2 min before return-
suffered severe oxidation and nitridation attack ing the carousel back to the test chamber. A
from both the air side (outside diameter side schematic of this dynamic burner rig is shown
of the transition duct) and the combustion side in Fig. 4.10. The combustion gas was determined
(inside diameter side of the duct) after service for to consist of 76% N2, 13% O2, 6% CO2, and 5%
slightly less than 2 years (14,000 h). Extensive H2O.
internal nitridation from both air and combustion
The test on alloy 617 produced severe internal
gas sides of alloy 617 transition duct is shown in
nitridation, with the microstructure very similar
Fig. 4.9 (Ref 35). Alloy 230 (Ni-22Cr-14W-
to that observed by Swaminathan and Lukezich
2Mo-0.3Al-La) was also tested for 16,000 h as a
(Ref 35) from the transition duct in a land-based
transition duct, suffering similar oxidation/nitri-
gas turbine power plant. Figure 4.11 shows the
dation attack (Ref 35). However, no aluminum
microstructure of an alloy 617 specimen after
nitrides were observed in alloy 230. Significant
testing at 980 °C (1800 °F) for 1000 h with
nitrogen pickup was observed from both transi-
30 min thermal cycling. Extensive needle-shape
tion ducts. Results of the chemical analyses of
aluminum nitrides were observed. Some blocky
nitrogen from samples at the exit end of the chromium nitrides were observed to form right
transition duct and at the location far away from
below the external oxide scales. Aluminum
the exit for both alloy 617 and 230 transition
nitrides were found to penetrate farther into the
ducts are shown in Table 4.4 (Ref 35).
metal interior than chromium nitrides. The oxide
Lai (Ref 34) used a high-velocity dynamic
scales were found to be porous and non-
burner rig test to simulate a gas turbine com-
protective.
bustion environment. The simulated combustion
Alloy 230 was included in the test and found
gas stream was generated by burning fuel oil
to show less nitridation attack under the same
(a mixture of two parts No. 1 fuel and one
test condition. Nitridation in alloy 230 involved
part No. 2 fuel) with an air-to-fuel ratio of ap-
only the formation of internal chromium nitrides
proximately 50 to 1 in a laboratory burner rig.
below internal chromium oxides and chromium
Most of the air for combustion was from a
denuded zone with no aluminum nitrides. Two
compressor. When combusted with fuel oil, a
other common combustor alloys, alloys X and
263, were also included in the test. Figures 4.12
and 4.13 show the microstructures of alloys
X and 263, respectively, after testing at 980 °C
(1800 °F) for 1000 h with 30 min thermal
cycling. Alloy X showed mainly internal chro-
OD ID mium nitrides, while alloy 263 showed mainly
tiny needle-shaped nitrides, presumably titanium

Table 4.4 Nitrogen contents at different

locations of alloy 617 and alloy 230 transition
ducts
Transition duct/service Location Nitrogen, wt%
Fig. 4.9 Cross section (2.5 mm, or 0.1 in.) of an alloy 617 (Ni-
Alloy 617/14,000 h Exit 0.24
22Cr-12.5Co-9Mo-1.2Al) transition duct after service
for less than 2 years (about 14,000 h) in a land-based gas turbine, Far away from exit 0.004
showing extensive formation of both aluminum and chromium Alloy 230/16,000 h Exit 0.22
nitrides from both air side (outside diameter of the transition duct) Far away from exit 0.05
and the combustion gas side (inside diameter of the duct). Source: Source: Ref 35
Ref 35
Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/ 30/10/2007 3:56PM Plate # 0 pg 74

74 / High-Temperature Corrosion and Materials Applications

Thermocouple for
steady-state control Thermocouple for recording
specimen temperature history

50 mm square insulated
flame tunnel
Specimen temperature measured
by pyrometer Compressed
inlet air 425 °C (800 °F)

Fuel

Combustor
Blower
Rotating shaft (thermal shock)

Thermal cycle

Fig. 4.10 The dynamic burner rig used by Lai (Ref 36) for simulating a gas turbine combustion environment in evaluating the
oxidation/nitridation behavior of gas turbine combustor alloys. Courtesy of Haynes International, Inc.

50 µm
50 µm

Fig. 4.12 Extensive internal chromium nitrides formed in

Fig. 4.11 Some blocky chromium nitrides and extensive alloy X after testing in the dynamic burner rig at
acicular aluminum nitrides formed in alloy 617 980 °C (1800 °F) for 1000 h with 30 min thermal cycling. The
after testing in the dynamic burner rig at 980 °C (1800 °F) for combustion gas stream with about mach 0.3 (100 m/s) consisted
1000 h with 30 min thermal cycling. The combustion gas stream of 76% N2, 13% O2, 6% CO2, and 5% H2O. Source: Ref 36
with about mach 0.3 (100 meter/s) consisted of 76% N2, 13% O2,
6% CO2, and 5% H2O. Source: Ref 36

X-ray diffraction analysis of the oxide scales

nitrides. Similar to both alloy 617 and 230, alloys formed on alloys 230, 617, and X was performed
X and 263 also exhibited nonprotective, porous with the results summarized in Table 4.5. The
oxide scales. results showed that NiO and NiCr2O4 along with
Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/ 30/10/2007 3:56PM Plate # 0 pg 75

Chapter 4: Nitridation / 75

Cr2O3 oxides made up the external oxide scales contents for the original samples (before test-
for alloy 230, and NiO and Cr2O3 made up the ing) and those after testing are summarized in
oxide scales for alloys 617 and X. With the for- Table 4.6, showing significant nitrogen adsorp-
mation of NiO oxides, the alloys were no longer tion for alloys 263, 617, and X. Alloy 230
protected by Cr2O3 oxide scales. An electrolytic showed little nitrogen adsorption. Bulk carbon
extraction technique was used to extract preci- contents for the samples before and after testing
pitate phases in the tested specimens for analysis, were also determined, and the results clearly
and the results are summarized in Table 4.5, showed that carburization was not involved
revealing essentially nitride phases in all three (Table 4.7). The overall test results in terms of
alloys. Alloy 230 also showed M6C carbides, weight loss (due to oxidation), metal loss (due
which were the carbides in the alloy in the to oxidation), internal oxidation, internal nitri-
as-solution-annealed condition. Bulk nitrogen dation, and total depth of attack are summarized
in Table 4.8.
In continuing his testing program for the same
simulated gas turbine environment involving the
same four combustor alloys (i.e., 230, 617, 263,
and X) at the same test temperature and duration

Table 4.6 Results of bulk nitrogen analysis before

and after testing at 980 °C (1800 °F) for 1000 h
with 30 min thermal cycling
Alloy Original nitrogen, wt% Nitrogen after testing, wt%
230 0.05 0.06
263 0.004 0.42
617 0.03 0.52
X 0.04 0.57
Source: Ref 36

Table 4.7 Results of bulk carbon analysis before

and after testing in the dynamic burner rig at
980 °C (1800 °F) for 1000 h with 30 minute
thermal cycling
50 µm
Alloy Original carbon, wt% Carbon after testing, wt%
230 0.09 0.09
Fig. 4.13 Extensive internal nitrides (believed to be titanium
263 0.06 0.03
nitrides) formed in alloy 263 after testing in the
dynamic burner rig at 980 °C (1800 °F) for 1000 h with 30 min 617 0.05 0.04
thermal cycling. The combustion gas stream with about mach 0.3 X 0.08 0.01
(100 m/s) consisted of 76% N2, 13% O2, 6% CO2, and 5% H2O. Source: Ref 36
Source: Ref 36

Table 4.5 Results of x-ray diffraction analysis Table 4.8 Test results in terms of weight loss,
on oxide scales and extracted precipitate depth of oxidation penetration, depth of
phases for alloys 230, 617, and X after testing nitridation, and total depth of attack after
at 980 °C (1800 °F) for 1000 h with 30 min testing at 980 °C (1800 °F) for 1000 h with
thermal cycling 30 min thermal cycling
Weight Internal Internal Total
Alloy Surface oxide scales Extraction residues change, Metal loss, oxidation, nitridation, attack(a),
230 NiCr2O4 (strong) M6C (strong) Alloy mg/cm2 mm mm mm mm
Cr2O3 (medium) Cr2N (medium) 230 −6.8 0.07 0.09 0.17 0.24
NiO (medium) 617 −80.1 0.17 0.07 >0.41(b) >0.58(b)
617 NiO (strong) AlN (strong) 263 −219.4 0.32 0.10 0.29 0.61
Cr2O3 (medium) TiN (medium weak) X −107.1 0.16 0.07 >0.40(b) >0.56(b)
X NiO (strong) CrN (medium strong)
Cr2O3 (weak) Cr2N (medium strong) Note: 1.0 mm=39.4 mils. (a) Metal loss + internal oxidation or internal
nitridation (whichever is greater). (b) Internal nitridation through thickness.
Source: Ref 36 Source: Ref 36
Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/ 30/10/2007 3:56PM Plate # 0 pg 76

76 / High-Temperature Corrosion and Materials Applications

(i.e., 980 °C for 1000 h), Lai (Ref 37) examined at 980 °C (1800 °F) for 1000 h with 30 min
the effect of the thermal cycling on the internal thermal cycling. The nitrogen content was found
nitridation. In this test, no thermal cycling was to increase to 1.27% after testing from the
involved. The results of this test (Ref 37) were original 0.13% before testing. Figure 4.16(b)
then compared with those in the earlier test shows the microstructure of Type 310 stainless
(Ref 36). The results of the high-velocity steel (SS) after testing at 980 °C (1800 °F)
dynamic burner rig test at 980 °C (1800 °F) for for 1000 h with 30 min thermal cycling, reveal-
1000 h without thermal cycling are summarized ing significant internal nitridation attack with
in Table 4.9. The results of chemical analysis formation of blocky chromium nitrides. The
showing nitrogen content before and after testing
are summarized in Table 4.10. In comparing
the test results with thermal cycling and those 0
230 X 617 263
without thermal cycling, thermal cycling sig-
nificantly accelerated oxidation attack by causing
oxide spallation, as shown in Fig. 4.14. Thermal –20
cycling was also found to accelerate nitridation
attack, as shown in Fig. 4.15. Among the four
Weight change, mg/cm2
combustor alloys, alloy 230, however, was least –40
affected by thermal cycling.
The test program was extended to include
–60
some iron-base alloys under the same test
conditions using the same dynamic burner rig
(Ref 38). The results showed that iron-base –80
alloys suffered significantly more nitridation
attack than nickel-base alloys. Figure 4.16(a)
shows the microstructure of alloy 556 (Fe-22Cr- –100
20Ni-18Co-3Mo-2.5W-0.6Ta-0.2N-La), reveal-
ing significant internal nitridation attack with –219
formation blocky chromium nitrides after testing
Non cyclic test
Cyclic test

Table 4.9 Test results in terms of weight loss,

depth of oxidation penetration, depth of
Fig. 4.14 Comparison weight change data between the
thermal cycling test (30 min cycles) and no ther-
nitridation, and total depth of attack after testing mal cycle test during the dynamic burner rig testing at 980 °C
at 980 °C (1800 °F) for 1000 h without thermal (1800 °F) for 1000 h for alloys 230, X, 617, and 263. Source:
Ref 36, 37
cycling
Weight Metal Internal Internal Total
change, loss, oxidation, nitridation, attack(a),
Alloy mg/cm2 mm mm mm mm Cyclic test
0.6
Nitrogen, wt%

230 −3.7 0.039 0.065 0.18 0.22

617 −12.1 0.031 0.126 >0.55(b) >0.58(b) 0.4
263 −19.3 0.072 0.135 0.16 0.23
X −5.0 0.030 0.086 0.10 0.13 0.2
Note: 1.0 mm = 39.4 mils. (a) Metal loss +internal oxidation or internal nitridation
(whichever is greater). (b) Internal nitridation through thickness. Source: Ref 37 0
230 X 617 263
Nitrogen, wt%

0.4 Non cyclic test

Table 4.10 Results of bulk nitrogen analysis

0.2
before and after testing at 980 °C (1800 °F) for
1000 h without thermal cycling 0
Original Nitrogen after
230 X 617 263
Alloy nitrogen, wt% testing, wt% Original
230 0.05 0.065 After testing
263 0.004 0.097
617 0.03 0.179
X 0.04 0.075 Fig. 4.15 Comparison nitrogen gain data between the thermal
cycling test (30 min cycles) and no thermal cycle
Source: Ref 37 test during the dynamic burner rig testing at 980 °C (1800 °F)
for 1000 h for alloys 230, X, 617, and 263. Source: Ref 36, 37
Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/ 30/10/2007 3:56PM Plate # 0 pg 77

Chapter 4: Nitridation / 77

nitrogen content increased from the original Figure 4.17 shows the microstructures of alloys
0.03% before testing to 1.69% after testing. 230, 617, and X after testing at 870 °C (1600 °F)
Alloy 800H was also shown to suffer severe for 2000 h with 30 min cycles. Some aluminum
nitridation attack with chromium nitrides and nitrides and chromium nitrides were observed
needle-shaped aluminum nitrides (Fig. 4.16c). in alloy 617, and only some chromium nitrides
The oxide scales formed on these iron-base were observed in alloy X, while no nitrides were
alloys after testing were porous and non- observed in alloy 230, as shown in Fig. 4.17
protective.
At 870 °C (1600 °F), internal nitridation
attack was found to be significantly reduced.

(a) 20 µm

(a)
50 µm

(b)

(b)

(c)

(c) Fig. 4.17 Alloys 230 (a), 617 (b), and X (c) after the dynamic
burner rig testing at 870 °C (1600 °F) for 2000 h
with 30 min cycles. Alloy 230 revealed no nitrides, alloy 617
Fig. 4.16 Extensive internal blocky chromium nitrides showed both chromium nitrides (blocky phases) and aluminum
formed in alloy 556 (a), Type 310 (b), and nitrides (needle phases), and alloy X showed only blocky chro-
alloy 800H (c) after the dynamic burner rig testing at 980 °C mium nitrides. Internal oxides were observed for all three alloys,
(1800 °F) for 1000 h with 30 min cycles. Courtesy of Haynes and all three alloys showed porous and nonprotective external
International, Inc. oxide scales. Courtesy of Haynes International, Inc.
Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/ 30/10/2007 3:56PM Plate # 0 pg 78

78 / High-Temperature Corrosion and Materials Applications

(Ref 38). Oxide scales for three alloys were alloy 214 tested specimen, and Fig. 4.20 shows
found to be porous and nonprotective. Iron-base the cross section of alloy MA956 tested speci-
alloys, such as Type 310SS, on the other hand, men. Both alloys showed fingerlike preferential
were found to suffer severe internal nitridation oxidation penetration. The preferential oxidation
attack, as shown in Fig. 4.18 (Ref 38). The dif- penetration was the result of thermal stresses
ference in the internal nitridation attack between developed from severe thermal cycling from
nickel-base and iron-base alloys is likely to be 1150 to less than 260 °C (2100 to <500 °F) every
caused by the differences in nitrogen solubilities 30 min. No preferential oxidation penetration
in two different alloy systems with much lower was observed under the same test condition
nitrogen solubilities in nickel-base alloys. without thermal cycling. Scanning electron
The above test data were generated from Ni-Cr microscopy with energy-dispersive x-ray spec-
and Fe-Ni-Cr alloys. These alloys are chromia troscopy (SEM/EDX) analysis of the oxide
formers (i.e., alloys forming Cr2O3 oxide scale).
These chromia formers are susceptible to oxi-
dation/nitridation attack in varying degrees under 2 1
3
gas turbine combustion conditions.
For very high temperatures and harsh oxi-
dizing combustion conditions, alumina formers
(i.e., alloys forming Al2O3 oxide scale) are
better performers. Lai (Ref 38) investigated two
alumina formers; one was wrought alloy 214
(Ni-16Cr-3Fe-4.5Al-Y) and the other oxide-
dispersion-strengthened alloy (produced by
powder metallurgy) MA956 (Fe-20Cr-4.5Al-
0.5Y2O3). Due to much more tenacious alumi-
num oxide scales, these two alloys were tested at
1150 °C (2100 °F) for 200 h with 30 min cycles.
The test results showed no nitridation in either 10 µm
alloy. Figure 4.19 shows the cross section of
Fig. 4.19 Scanning electron micrograph showing the oxide
scale of alloy 214 after testing in the dynamic
burner rig at 1150 °C (2100 °F) with 30 min cycle. The results of
the energy-dispersive x-ray spectroscopy (EDX) analysis of the
oxide scale are summarized: 1, aluminum oxide; 2, aluminum
oxide; and 3, Al-rich (Ni,Cr) oxide

2
1

10 µm

20 µm
Fig. 4.20 Scanning electron micrograph showing the oxide
scale of alloy MA956 after testing in the dynamic
Fig. 4.18 Extensive blocky chromium nitrides formed in Type burner rig at 1150 °C (2100 °F) with 30 min cycle. The results of
310SS after testing in the dynamic burner rig testing the energy-dispersive x-ray spectroscopy (EDX) analysis of the
at 870 °C (1600 °F) for 2000 h with 30 min cycles. Courtesy of oxide scale are summarized: 1, Fe-Al-rich oxide; 2–4, aluminum
Haynes International, Inc. oxide
Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/ 30/10/2007 3:56PM Plate # 0 pg 79

Chapter 4: Nitridation / 79

scales showed that both alloys exhibited an alu- increasing exposure times eventually leads
minum oxide scale. For MA956 some Fe-Al-rich to the formation of nitrides in the alloy once
oxide phases were observed to form on the top of the solubility limits for CrN, Cr2N, AlN,
the aluminum oxide scale. and/or TiN are exceeded.
A model for oxidation/nitridation reactions
for chromia formers (Ni-Cr and Fe-Ni-Cr alloys)
in oxidizing combustion atmospheres was pro- 4.4 Nitridation in NH3-H2O
posed by Lai (Ref 36). This model, as schema- Environments
tically illustrated in Fig. 4.21, involves the
following reaction steps: A NH3-H2O mixture has been considered in
the Kalina cycle for power generation (Ref 39).
1. Chromium oxides form a protective scale Very little published data are available for the
initially on the alloy surface in oxidizing corrosion behavior of engineering alloys in this
combustion atmospheres. type of environment. Grabke et al. (Ref 40)
2. Cracks, pores, and other defects develop in investigated the corrosion behavior of a number
the chromium oxide scales after thermal of commercial alloys in the NH3-30%H2O gas
cycling and/or long-term exposure. mixture at 500 °C (930 °F). Several interesting
3. Chromium oxide scales become porous and results were obtained from this investigation.
nonprotective with nickel oxides forming in One of the most interesting observations was
Ni-Cr alloys and iron oxides forming in iron- that significant nitridation attack and severe
base alloys. intergranular cracking were observed in alloy
4. Both O2 and N2 molecules from the com- 600 (Ni-16Cr-8Fe) after only 200 h of exposure,
bustion gas stream permeate through the as shown in Fig. 4.22. Alloy 600 has been
oxide scales and reach the metal underneath. known to be one of the most nitridation-resistant
5. Oxidation of the metal surface results in alloys in ammonia environments, and the
lower oxygen potential with a concurrent alloy has been widely used in ammonia plants
increase in nitrogen potential. (data are presented in Section 4.5). Alloy 800
6. Nitridation then occurs following the reac- (Fe-22Cr-32Ni-Al-Ti) was also found to suffer
tion: 1=2N2 (gas) $ N (solution); the concen- severe nitridation attack and intergranular
tration of nitrogen absorbed in the metal is cracking (Fig. 4.23). For the ferritic stainless
then proportional to the nitrogen potential steel, Fe-18Cr (Sicromal), a combination of
( pN2 ) by: severe intergranular nitridation, cracking, and
[%N]=k( pN2 )1=2 oxidation attack caused rapid disintegration of
the alloy in 200 h (Fig. 4.24). An austenitic
where k is an equilibrium constant. stainless steel (Fe-18Cr-9Ni) was also found to
7. Nitrogen dissolves into the alloy and dif- suffer severe nitridation attack. However, no
fuses into the metal interior. Increasing
nitrogen concentration in the alloy with

2
CrN
–2 Cr2N Test
environment
–6
Log pN , atm

AIN Cr2O3
2

–10

Cr
–14
AI2O3
–18 AI
50 µm
–22
–55 –45 –35 –25 –15 –5 5
Log pO , atm
2
Fig. 4.22 Optical micrograph showing the cross section
of alloy 600 (Ni-16Cr-8Fe) after exposure to the
Fig. 4.21 Schematic showing a model for internal nitridation NH3-30%H2O gas mixture at 500 °C (930 °F) for only 200 h,
attack in high-temperature alloys in a simulated revealing severe nitridation attack and intergranular cracking.
combustion environment. Source: Ref 36 Source: Ref 40
Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/ 30/10/2007 3:56PM Plate # 0 pg 80

80 / High-Temperature Corrosion and Materials Applications

cracking was observed in the austenitic stainless 4.5 Nitridation in NH3 and H2-N2-NH3
steel. They found formation of very fine CrN Environments
precipitates in these alloys. The authors (Ref 40)
pointed out that in these tests (low pressure) NH3 Ammonia (NH3) is thermally unstable. It can
should decompose largely to N2 and H2, while readily dissociate into N2 and H2 at elevated
in the Kalina process (under high pressures) temperatures. There appears to be no published
decomposition of NH3 is considered to be data available on the dissociation rate as func-
negligible. It is believed that at the test tem- tions of temperature and pressure. However,
perature of 500 °C, a significant amount of several laboratory measurements have indicated
NH3 is believed to be retained without decom- that dissociation rates were extremely fast at
position. Robo (Ref 4) found that about 60% high temperatures. Table 4.11 shows some dis-
NH3 was decomposed, leaving about 40% sociation data generated in laboratory test fur-
NH3 at the exhaust end at 525 °C (980 °F) in naces by measuring the amount of NH3 at the
his laboratory testing using 100% NH3 in the exhaust end of the tube furnace when 100% NH3
inlet gas. Barnes and Lai (Ref 41) found a larger was entered into the furnace tube. Both sets of
decomposition of NH3 (100% NH3 in the inlet data were measured in laboratory setups with
test gas and 30% NH3 in the exhaust) when tested pressures being close to 1 atm. In order to avoid
at a slightly higher temperature, 650 °C. The catalyst reactions by metals, the measurement of
corrosion behavior of alloys in the NH3-H2O ammonia dissociation at 650, 980, and 1090 °C
mixture as observed by Grabke et al. (Ref 40) was made with no metallic samples in the fur-
appears to be different from that of the NH3-H2 nace tube, which was composed of high-purity
mixture, as is discussed in the next section. alumina (Ref 38). No published high-pressure
dissociation values are available. Thus, for
laboratory test data even with 100% NH3 as the

(a)
30 µm 200 µm

Fig. 4.24 Optical micrograph showing the cross section of

an Fe-18Cr (Sicromal alloy) tested specimen that
suffered combination of severe nitridation, oxidation, and inter-
granular cracking after the exposure to the NH3-30%H2O gas
mixture at 500 °C (930 °F) for 200 h. Source: Ref 40

Table 4.11 NH3 content measured at

50 µm
the exhaust end of the tube furnace at different
temperatures when 100% NH3 was injected
NH3 at inlet, vol% Temperature, °C (°F) NH3 at exhaust, vol%
(b)
100 525 °C (980 °F) 40(a)
650 °C (1200 °F) 30(b)
Fig. 4.23 Severe nitridation attack and intergranular cracking 980 °C (1800 °F) <5(b)
in alloy 800 (Fe-22Cr-32Ni-Al-Ti) after exposure
to the NH3-30%H2O gas mixture at 500 °C (930 °F) for 200 h. (a) 1090 °C (2000 °F) <5(b)
Surface appearance of the tested specimen showing intergranular (a) Ref 4. (b) Ref 41
cracks. (b) Cross section of the tested specimen. Source: Ref 40
Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/ 30/10/2007 3:56PM Plate # 0 pg 81

Chapter 4: Nitridation / 81

inlet gas, the corrosion reactions generally converter. The converters operate at high pres-
involve both N2 and NH3. At very high tem- sures (130–350 atm or 800–1000 atm) and
peratures, such as 980° and 1090 °C (1800 and temperatures up to 650 °C (1200 °F) (Ref 43).
2000 °F), nitridation is most likely involved in Cihal (Ref 43) discussed the major corrosion
the reaction with N2 because of rapid dissocia- problems—hydrogen attack and nitridation—for
tion of NH3. the ammonia converter. The converter usually
Verma et al. (Ref 42) reported that an ammonia consists of a vessel with a catalyst basket and
cracker unit, used to produce nitrogen and an interchanger inside the vessel. Because of
hydrogen, failed after 1000 h of operation. The high-pressure, high-temperature hydrogen in the
preheater tubes (operating at 350 to 400 °C, converter, early converters were constructed out
or 660 to 750 °F) were made of Type 304SS, of a thick-wall steel vessel with an inner carbon
while the furnace tubes (operating at about steel lining and vent holes through the vessel
600 °C, or 1110 °F) were made of Type 310SS. wall. Thus, the inner carbon steel lining was
Both suffered severe nitridation attack. To select the only part suffering hydrogen attack, while the
an alternate alloy, nitriding tests were performed main thick-wall vessel was unaffected by high-
on various alloy samples at 600 °C (1110 °F) in pressure, high-temperature hydrogen (Ref 43).
an environment consisting of 6 to 8% NH3, 75.77 Hydrogen attack is the damage of steel by the
to 77.5 wt% N2, and 16.25 to 16.5 wt% H2. Test reaction of hydrogen with cementite (Fe3C) in
results are summarized in Table 4.12. The alloys steel to form methane gas (CH4), resulting in
that performed well include Types 347, 316, formation of microcracks and fissures as well
321, SLX-254, and HV-9A. Type 347 was the as decarburization in steel. (Hydrogen attack is
best performer, having a linearly extrapolated reviewed and discussed in Chapter 17.) Later
penetration rate of about 0.13 mm/yr (5 mpy). designs of the converter allowed the cold inlet
Alloy 800, which contains more nickel than any gas flowing along the vessel wall to keep the
of the above stainless steels, did not perform vessel cold, thus eliminating the potential hydro-
as well. Furthermore, Type 304 was found to gen attack problem for the vessel (Ref 43).
suffer attack two orders of magnitudes higher Cihal (Ref 43) indicated that the internal
than that of Type 316L. The results also showed components made of carbon steels exhibited a
that titanium suffered severe nitridation attack, short life due to hydrogen attack. Alloy steels
which resulted in severe sample cracking. Both containing chromium were more resistant to
carbon steel and 1Cr-0.5Mo steel suffered de- hydrogen attack, but had suffered severe em-
carburization after only 50 h. brittlement problems due to nitridation attack.
Ammonia (NH3) is produced by synthesis Figure 4.25 shows intergranular cracking in
from hydrogen and nitrogen at high pressures the nitrided layer of an alloy steel (0.12C-
and elevated temperatures. The “heart” of the 5.6Cr-0.42Mo) after exposure to the synthesis
process is the ammonia “converter,” where gas inside the converter at 325 atm and 450 to
hydrogen and nitrogen combine. Significant 500 °C (840 to 930 °F) for 4380 h (Ref 43).
corrosion issues are associated with the con- An alloy steel containing a strong nitride former
verter and the internal components inside the such as titanium, such as alloy steel with

Table 4.12 Nitridation attack of various alloys in an ammonia-bearing environment at 600 °C

(1110 °F) for indicated exposure times
Penetration depth of nitrtdiltion attack, mm (mils)
Alloy 50 h 100 h 300 h 600 h 1000 h 1500 h
Carbon steel Decarb. Decarb. Decarb. Decarb. Decarb. Decarb.
1Cr-0.5Mo steel Decarb. Decarb. 0.033 (1.3) 0.033 (1.3) 0.033 (1.3) 0.3 (11.8)
Titanium 0.0066 (0.3) 0.0133 (0.5) 0.233 (9.2) 0.266 (10.5) Cracked(a) Cracked(a)
304 … 0.013 (0.5) 0.013 (0.5) 0.03 (1.2) 0.06 (2.4) 4.2 (165)
316L 0.02 (0.8) 0.02 (0.8) 0.02 (0.8) 0.03 (1.2) 0.04 (1.6) 0.04 (1.6)
329 … 0.066 (2.6) 0.10 (3.9) 0.10 (3.9) 0.20 (7.9) 0.40 (15.7)
310 … 0.03 (1.2) 0.13 (5.1) 0.16 (6.3) 0.33 (13.0) 0.40 (15.7)
321 0.013 (0.5) 0.013 (0.5) 0.013 (0.5) 0.016 (0.6) 0.06 (2.4) 0.06 (2.4)
347 … 0.013 (0.5) 0.013 (0.5) 0.013 (0.5) 0.02 (0.8) 0.02 (0.8)
SLX-254(b) 0.013 (0.5) 0.013 (0.5) 0.026 (1.0) 0.026 (1.0) 0.06 (2,4) 0.06 (2.4)
HV-9A(c) 0.01 (0.4) 0.10 (3.9) 0.10 (3.9) 0.10 (3.9) 0.10 (3.9) 0.10 (3.9)
800 0.02 (0.8) 0.10 (3.9) 0.20 (7.9) … 0.20 (7.9) 0.20 (7.9)
Decarb.: decarburized. (a) Nitridation through thickness. (b) SLX-254: Fe-19.7Cr-24.5Ni-4.35Mo-1.43Cu. (c) HV-9A: Fe-21.2Cr-24.6Ni-3.8Mo-1.5Cu. Source: Ref 42
Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/ 30/10/2007 3:56PM Plate # 0 pg 82

82 / High-Temperature Corrosion and Materials Applications

0.05% C, 2.9% W, and 0.54% Ti, exposed to the 4 mpy). One Type 304 sample showed a slightly
same converter environment under the same higher corrosion rate (10 mpy), presumably due
test conditions as the 0.12C-5.6Cr-0.42Mo steel to a higher temperature. Alloy 600 (Ni-Cr-Fe
(as shown in Fig. 4.25) was found to show no alloy) had significantly better nitridation resis-
cracking. tance than stainless steels, with corrosion rates
Nitridation resistance of various alloys was 1 or 2 orders of magnitude lower. The results are
studied by Moran et al. (Ref 44) in an ammonia summarized in Table 4.14 (Ref 4).
converter and preheater line. The results are McDowell (Ref 45) reported field test results
summarized in Table 4.13. Corrosion rates were performed in a Casale converter (540 °C, or
found to depend strongly on the concentration 1000 °F, and 11 ksi) for 1 and 3 years. These
of ammonia. Type 304, for example, suffered results are summarized in Table 4.15. AISI 502
corrosion rates that increased from 0.02 to (5Cr steel) was extremely susceptible to nitrida-
2.5 mm/yr (0.6 to 99 mpy) as the concentration tion attack, with more than 2.54 mm (0.1 in., or
of NH3 was increased from 5 to 6% (in the 100 mils) of nitridation depth in a year. Results
ammonia converter) to 99% (in the ammonia showed a general trend of increased resistance
preheater line) at about 500 °C (930 °F). In an to nitridation as nickel content in the alloy
ammonia converter with about 5 to 6% NH3 and increased. One striking observation was that after
490 to 550 °C (910 to 1020 °F), all stainless 3 years of exposure, the alloys showed essen-
steels tested (i.e., 430, 446, 302B, 304, 316, tially similar depths of nitridation attack as they
321, 309, 314, 310, and 330) showed negligible did after 1 year.
nitridation attack, with corrosion rates of Robo (Ref 4) reported kinetic data for
about 0.03 mm/yr (1 mpy) or less. For the plant Type 304 and alloys 600 and 625 in laboratory
ammonia line (preheater exit), which was tests performed with pure ammonia as the inlet
exposed to 99% NH3, stainless steels, such as test gas. The nitridation for Type 304 was found
446, 304, 316, and 309, suffered severe nitrida- to follow a linear rate law at 525 °C (980 °F) for
tion attack, with corrosion rates of about up to 1000 h. The maximum thickness of the
2.54 mm/yr (100 mpy) or more. Moran et al. nitride layer (in the form of scale) measured
(Ref 44) found that Type 316 suffered sig- metallographically is shown in Fig. 4.26 as a
nificantly more attack than Type 304, contrary function of time. A growth rate of about 0.37
to the observations of Verma et al. (Ref 42). µm/h was observed. This corresponds to about
Robo (Ref 4) reported the performance of 3240 µm/yr (128 mpy). This rate is signifi-
several alloys in a Topsoe-type ammonia con- cantly higher than those observed in ammonia
verter. Most of the components made of converters. The ammonia concentration in this
Type 304, exposed to temperatures up to 500 °C test (reportedly, 40% ammonia was dissociated
(930 °F) with ammonia concentration up to 20%,
exhibited negligible nitridation rates (0.4 to
Table 4.13 Corrosion behavior of various
alloys in an ammonia converter and plant
ammonia line
Corrosion rate. mm/yr (mpy)
Ammonia Plant ammonia
Alloy (converter)(a) line(b)
430 0.022 (0.90) …
446 0.028 (1.12) 4.18 (164.5)
302B 0.019 (0.73) …
304 0.015 (0.59) 2.53 (99.5)
316 0.012 (0.47) >13.21 (520)
321 0.012 (0.47) …
309 0.006 (0.23) 2.41 (95)
314 0.003 (0.10) …
310 0.004 (0.14) …
330 (0.47Si) 0.002 (0.06) …
330 (1.00Si) 0.001 (0.02) 0.43 (17.1)
600 0.16 (6.3)
80Ni-20Cr 0.19 (7.4)
Ni 2.01 (79.0)
Fig. 4.25 Intergranular cracking in the nitrided layer of an
(a) 5 to 6% NH3, 29164 h at 490 to 550 °C (910 to 1020 °F), and 354 atm
alloy steel (0.12C-5.6Cr-0.42Mo) after exposure (5200 psi) ( Haber-Bosch converter). (b) 99. 1% NH3, 1540 h at 500 °C (930 °F).
to the synthesis gas inside the converter at 325 atm and 450 to Source: Ref 44
500 °C (840 to 930 °F) for 4380 h. Source: Ref 43
Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/ 30/10/2007 3:56PM Plate # 0 pg 83

Chapter 4: Nitridation / 83

Table 4.14 Corrosion behavior of various alloys in an ammonia converter

Gas Thickness
temperature, Time of of nitride, Average nitriding,
Converter No. Component Material °C NH3 operation, yr μm (mils) µm/yr (mpy)
1 Lining 304 525 15–20 4 1000 (39.4) 250 (9.8)
Plate, 2nd bed 304 475 15–20 7 100 (3.9) 14 (0.6)
2 Bolt 302 … … 7 375 (14.8) 54 (2.1)
Wire mesh, 2nd bed Alloy 600 520 … 7 8 (0.3) 1 (0.04)
3 Perforated plate. 1st bed 304 500 13 5 270 (10.6) 54 (2.1)
Inner shell, 2nd bed 304 440 8–10 5 45 (1.8) 9 (0.4)
Perforated plate. 2nd bed 304 440 8–10 5 60 (2.4) 12 (0.5)
Center tube, 2nd bed 304 485 16 5 440 (17.3) 88 (3.5)
Nut, bottom 304 480 16 5 260 (10.2) 52 (2.0)
Bolt, bottom 403 480 16 5 540 (21.3) 108 (4.3)
Wire mesh Alloy 600 c.500 4 6 (0.2) 1.5 (0.06)
4 ThermoweII 304 c.500 3.5 8 200 (7.9) 25 (1.0)
Note: Topsoe-type ammonia converter operated at 22 MPa (3.2 ksi). Source: Ref 4

Table 4.15 Depth of nitridation for various alloys 500

after 1 and 3 years in a Casale ammonia converter
Nitridation depth, mm (mils)
Alloy 1 yr 3 yr
502 (5Cr steel) 2. 88 (113.2) Completely nitrided 400
446 1.06 (41.7) 1.15 (45.3)
Thickness of nitride layer, µm

304 1.08 (42.7) 1.12 (44.0)

316 0.46 (18. 2) 0.48 (18.7)
321 0.46 (18.3) 0.60 (23.6)
347 0.49 (19.2) 0.45 (17.6) 300
309 0.24 (9.5) 0.24 (9.6)
310 0.22 (8.8) 0.23 (9.2)
800 0.14 (5.4) 0.13 (5.3)
804 (30Cr-42Ni) 0.03 (1.2) 0.03 (1.2)
600 0.16 (6.4) 0.16 (6.4) 200
Nickel 200 None None
Note: Operated at 538 °C (1000 °F) and 76 MPa (11 ksi). Source: Ref 45

100
in the test furnace) was significantly higher than
in the ammonia converters. The rate of about
3.25 mm/yr (128 mpy) was, however, of the
same order of magnitude as that observed by 0
Moran et al. (Ref 44) in the plant ammonia line 0 500 1000
(about 100 mpy for Type 304). Exposure time, h
At 700 °C (1290 °F), Robo (Ref 4) found that
the reaction rates for Type 304, alloy 600, and Fig. 4.26 Nitriding depth of Type 304SS in ammonia (100%
in the inlet gas and 60% in the exhaust) at 525 °C
alloy 625 can be described by: (980 °F) as a function of exposure time. Source: Ref 4

X =kt n ð4:5Þ
where X is thickness of the nitrided layer in µm, (at.%), γ is the ratio of nitrogen to alloy element
k is reaction constant, t is time in hours, n is 0.66 in the nitride phase, D is diffusivity of nitrogen,
for Type 304 and 0.26 for alloys 600 and 625. and t is time.
Jack (Ref 46) developed a kinetic model based This model predicts that fast nitriding rates
on the models for internal oxidation, to describe can be achieved by increasing the ammonia
the growth of internal penetration (X) in the content in the gas mixture and thus the surface
absence of iron nitride formation: nitrogen concentration. Another important factor
2[N] in nitriding kinetics is the concentration of the
X 2= Dt ð4:6Þ alloy element that forms internal nitrides. The
c[m]
model predicts that nitriding depth is inversely
where [N] is the surface nitrogen concentration proportional to the concentration of the nitride-
(at.%), [m] is the alloy element concentration forming alloy element.
Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/ 30/10/2007 3:56PM Plate # 0 pg 84

84 / High-Temperature Corrosion and Materials Applications

The nitridation behavior of a wide variety of Table 4.17 Nitridation resistance of various
commercial alloys in ammonia was extensively alloys in ammonia at 980 °C (1800 °F) for
investigated by Barnes and Lai (Ref 41). The 168 h
alloys tested included stainless steels, Fe-Ni-Cr Nitrogen Depth of
absorption nitride penetration,
alloys, and nickel- and cobalt-base superalloys. Alloy Alloy base mg/cm2 mm (mils)
Many alloys tested contained alloying elements 214 Nickel 0.3 0.04 (1.4)
(e.g., Al, Ti, Zr, Nb, Cr, Mo, W, Fe, etc.) that 600 Nickel 0.9 0.12 (4.8)
S Nickel 0.9 0.18 (7.2)
form nitrides. The test results generated at 650, 601 Nickel 1.2 0.17 (6.6)
980, and 1090 °C (1200, 1800, and 2000 °F) 230 Nickel 1.4 0.12(4.9)
are summarized in Tables 4.16 to 4.18. It was 617 Nickel 1.5 0.38 (15.0)
HR-160 Nickel 1.7 0.18 (7.2)
found that nickel-base alloys are generally more 188 Cobalt 2.3 0.19(7.4)
nitridation resistant than iron-base alloys. In- 625 Nickel 2.5 0.17 (6.9)
creasing nickel content generally improves 6B Cobalt 3.1 0.15 (5.8)
253MA Iron 3.3 0.48 (19.0)
the resistance of the alloy to nitridation attack. 25 Cobalt 3.6 0.26(10.4)
Increasing cobalt content appears to have the X Nickel 3.2 0.19 (7.4)
RA333 Nickel 3.7 0.42(16.4)
same effect. When nitrogen absorption was RA330 Iron 3.9 0.52 (20.6)
plotted against Ni + Co content in the alloy for 800H Iron 4.0 0.28 (11.1)
the 650 °C (1200 °F) test data, resistance to 825 Nickel 4.3 0.58 (23.0)
150 Cobalt 5.3 0.38 (15.1)
nitridation improves with increasing Ni + Co MULTIMET Iron 5.6 0.35 (13.6)
content up to about 50 wt%, as shown in 316 Iron 6.0 0.52 (20.3)
Fig. 4.27. Further increases up to about 75% did 556 Iron 6.7 0.37 (14.7)
304 Iron 7.3 >0.58 (23.0)
not seem to affect the nitridation resistance of the 310 Iron 7.7 0.38 (15.1)
alloy. For maximum resistance to nitridation 446 Iron 12.9 >0.58 (23.0)
attack at 650 °C (1200 °F), it appears that alloys Note: 100% NH3 in the inlet gas and less than 5% NH3 (detection limit) in the
exhaust gas. Source: Ref 41
with at least 50% Ni or Ni + Co are most suitable.
This is in general agreement with the results
reported by Moran et al. (Ref 44). Their results
suggested that improvement in nitridation resis-
tance began to level off at about 40% Ni, with no
improvement resulting from further increases in
Table 4.18 Nitridation resistance of various
nickel up to about 80%. Pure nickel, however,
alloys in ammonia at 1090 °C (2000 °F) for
showed significantly lower nitridation resistance 168 h
(Ref 43). At 980 °C (1800 °F), a slightly differ-
Nitrogen Depth of nitride
ent relationship was observed, as shown in absorption, penetration,
Alloy Alloy base mg/cm2 mm (mils)
600 Nickel 0.2 0
Table 4.16 Nitridation resistance of various 214 Nickel 0.2 0.02 (0.7)
alloys in ammonia at 650 °C (1200 °F) for 168 h S Nickel 1.0 0.34 (13.4)
230 Nickel 1.5 0.39 (15.3)
Nitrogen Depth of nitride 25 Cobalt 1.7 >0.65 (25.5)
absorption, penetration,
617 Nickel 1.9 >0.56 (22)
Alloy Alloy base mg/cm2 mm (mils)
188 Cobalt 2.0 >0.53 (21)
C-276 Nickel 0.7 0.02 (0.6) HR-160 Nickel 2.5 0.46 (18)
230 Nickel 0.7 0.03 (1.2) 601 Nickel 2.6 >0.58 (23)
HR-160 Nickel 0.8 0.01 (0.5) RA330 Iron 3.1 >0.56 (22)
600 Nickel 0.8 0.03(1.3) 625 Nickel 3.3 >0.56 (22)
625 Nickel 0.9 0.01 (0.5) 316 Iron 3.3 >0.91 (36)
RA333 Nickel 1.0 0.03 (1.0) 304 Iron 3.5 >0.58 (23)
601 Nickel 1.1 0.03 (1.0) X Nickel 3.8 >0.58 (23)
188 Cobalt 1.2 0.02 (0.6) 150 Cobalt 4.1 0.51 (20)
S Nickel 1.3 0.03 (1.1) 556 Iron 4.2 >0.51 (20)
617 Nickel 1.3 0.03 (1.0) 446 Iron 4.5 >0.58 (23)
214 Nickel 1.5 0.04 (1.5) 6B Cobalt 4.7 >0.64 (25)
X Nickel 1.7 0.04 (1.5) MULTIMET Iron 5.0 >0.64 (25)
825 Nickel 2.5 0.06 (2.2) 825 Nickel 5.2 0.58 (23)
800H Iron 4.3 0.10 (4.1) RA333 Nickel 5.2 >0.71 (28)
556 Iron 4.9 0.09 (3.5) 800H Iron 5.5 >0.76 (30)
316 Iron 6.9 0.19 (7.3) 253MA Iron 6.3 >1.5 (60)
310 Iron 7.4 0.15 (6.0) 310 Iron 9.5 >0.79 (31)
304 Iron 9.8 0.21 (8.4)
Note: 100% NH3 in the inlet gas and less than 5% (detection limit) in the exhaust
Note: 100% NH3 in the inlet gas and 30% NH3 in the exhaust gas. Source: Ref 41 gas. Source: Ref 41
Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/ 30/10/2007 3:56PM Plate # 0 pg 85

Chapter 4: Nitridation / 85

10 15

8
Nitrogen absorption, mg/cm2

Nitrogen absorption, mg/cm2

10
6

4
5

0 0
0 20 40 60 80 0 20 40 60 80
Ni + Co, wt% Ni + Co, wt%

Fig. 4.27 Effect of the Ni + Co content in iron-, nickel-, Fig. 4.28 Effect of the Ni + Co content in iron-, nickel-,
and cobalt-base alloys on nitridation resistance at and cobalt-base alloys on nitridation resistance at
650 °C (1200 °F) for 168 h in ammonia (100% NH3 in the inlet 980 °C (1800 °F) for 168 h in ammonia (100% NH3 in the inlet
gas and 30% NH3 in the exhaust). Source: Ref 41 gas and <5% NH3 in the exhaust). Source: Ref 41

Fig. 4.28. Nitrogen absorption was reduced Table 4.19 Phases detected from the x-ray
drastically with an initial 15% Ni (or Ni + Co). As diffraction analysis performed on the surfaces
Ni + Co content increased from 15 to 50%, no of test specimens after exposure to NH3 at
drastic improvement in nitridation resistance was temperatures indicated for 168 h
noted. Further increases in Ni + Co content in Phases
excess of about 50% caused a sharp improve- Alloy 650 °C (1200 °F) 980 °C (1800 °F) 1090 °C (2000 °F)

ment. Alloys with Ni (or Ni + Co) in excess of 304SS Fe2N CrN (Cr,Fe)2N1-x
800H (Fe3Ni)N CrN (Cr,Fe)2N1-x
about 60% showed the most resistance to nitri- 556 (Fe3Ni)N CrN (Cr,Fe)2N1-x
dation. In this test program (Ref 41), no alloys 230 CrN CrN (Cr,Mo)12(Fe, Ni)8-xN4-z
with 80% Ni (or Ni + Co) and higher were tested. 188 CrN CrN Cr2N, CrN
It is generally believed that the beneficial Note: At 650 °C (1200 °F), 100% NH3 in the inlet gas and 30% NH3 in the
exhaust. At 980 and 1090 °C (1800 and 2000 °F), 100% NH3 in the inlet gas and
effect of nickel or cobalt in increasing nitridation <5% NH3 in the exhaust. Source: Ref 41
resistance is caused by the reduced solubility of
nitrogen in the alloy. Nickel and cobalt were
found to reduce the solubility of nitrogen in iron
(Ref 15, 47). 650 °C (1200 °F) for iron-base alloys, Type
Morphology of nitrides formed in alloys as a 446SS (Fe-25Cr) and Type 304SS (Fe-18Cr-
result of exposure to NH3 is widely different 8Ni), and nickel-base alloys, alloys 600 (Ni-
between low- and high-temperature exposures. 16Cr-8Fe), 625 (Ni-22Cr-9Mo-3.5Nb-3Fe), X
Nitridation at low temperatures (e.g., 650 °C, or (Ni-22Cr-18.5Fe-9Mo-0.6W), and C-276 (Ni-
1200 °F) generally results in a surface nitride 16Cr-5Fe-16Mo-4W). Type 446SS containing
layer. For iron-base alloys, the surface nitride about 25% Cr with no nickel exhibited a thick
layer consists of mostly iron nitrides (Fe2N or nitride layer (about 0.72 mm, or 28 mils, thick)
Fe4N), while for nickel- and cobalt-base alloys, formed on the alloy surface, as shown in
the nitride layer consists of mainly CrN. High- Fig. 4.29(a). With about 8% Ni in the alloy, Type
temperature exposures, on the other hand, result 304SS showed a significantly thinner nitride
in formation of internal nitrides, which are layer (about 0.2 mm, or 0.008 in., 8 mils thick)
mostly CrN, Cr2N, (Fe,Cr)2N, AlN, and TiN. formed on the alloy surface. For four nickel-
Table 4.19 summarizes the results of the x-ray base alloys, alloys 600, 625, X, and C-276, an
diffraction analysis performed on the surfaces of extremely thin nitride layer (about 1.2 to 2.2 µm
the selected test specimens tested at different thick) was found to form on the alloy surface.
temperatures. Figure 4.29 illustrates the mor- The iron content in these four nickel-base alloys
phology of the surface nitride layer formed at varies from about 3 to 19%.
Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/ 30/10/2007 3:56PM Plate # 0 pg 86

86 / High-Temperature Corrosion and Materials Applications

When exposed to high temperatures, the alloy alloys (Table 4.19). As nitrogen diffuses farther
forms internal nitrides due to increased diffu- into metal interior, nitrogen activities become
sivity of nitrogen. Figure 4.30 shows internal lower, thus forming Cr2N. Thus, the nitrides
nitrides formed in Type 446SS, Type 304SS, (etched lighter) formed in the metal interior are
alloy 800H (Fe-21Cr-32Ni-0.4Al-0.4Ti), and believed to be Cr2N.
alloy 188 (Co-22Cr-22Ni-14W-La). For Type Ni-Cr alloys containing aluminum or titanium
304SS, alloy 800H, and alloy 188, nitrides (or both) form internal nitrides of not only
appeared to be etched differently with the nitrides chromium but also aluminum or titanium (or
formed near the surface etched darker than those both) at high temperatures. Both aluminum and
in the interior. The nitrides (etched darker) that titanium are stronger nitride formers, and AlN
formed in the surface zone are believed to be and TiN can form and penetrate farther into the
CrN. The x-ray diffraction analysis of the test metal interior than CrN and Cr2N. When alloys
specimen surface showed CrN for these three contain relatively low aluminum (e.g., about

(a) 200 µm (d)

200 µm

(b) 200 µm (e)

200 µm

(c) 200 µm (f)

200 µm

Fig. 4.29 Optical micrographs showing typical nitride morphology of a surface nitride layer that formed on the alloy surface when
exposed to NH3 (100% NH3 in the inlet gas and 30% NH3 in the exhaust) for 168 h at 650 °C (1200 °F) for (a) Type 446, (b)
Type 304, (c) alloy 600, (d) alloy 625, (e) alloy X, and (f) alloy C-276. Courtesy of Haynes International, Inc.
Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/ 30/10/2007 3:56PM Plate # 0 pg 87

Chapter 4: Nitridation / 87

1%), such as alloys 601 and 617, a significant 4.6 Nitridation in N2 Atmosphere
amount of internal aluminum nitrides formed in
the alloy, as shown in Fig. 4.31(a). Also observed Metals and alloys are also susceptible to
in alloy 601 (Fig. 4.31a) are chromium nitrides nitridation attack in N2 or N2-H2 environments,
(blocky-type phases) that formed near the alloy particularly at high temperatures. The N2 or
surface. For alloys containing high concen- N2-H2 atmosphere is commonly used as a pro-
trations of aluminum (e.g., 4.5% Al in alloy 214), tective atmosphere in heat treating and sinter-
aluminum nitrides formed on the alloy surface, ing operations. Figure 4.32 shows extensive
as shown in Fig. 4.31(b). Al2O3 oxide is also nitridation attack of Type 314 wire mesh belt
believed to form on the alloy 214 surface. in a sintering furnace after 2 to 3 months of

(a) 50 µm (c) 50 µm

(b) 50 µm (d) 50 µm

Fig. 4.30 Optical micrographs showing typical nitride morphology in form of internal nitrides penetrating into the metal interior when
exposed to NH3 for 168 h at 980 °C (1800 °F) for (a) Type 446, (b) Type 304, (c) alloy 800H, and (d) alloy 188. Courtesy of
Haynes International, Inc.
Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/ 30/10/2007 3:56PM Plate # 0 pg 88

88 / High-Temperature Corrosion and Materials Applications

service at 1120 °C (2050 °F) in the N2-10%H2 For alloys containing 20% or more chromium,
atmosphere. Type 314SS (Fe-26Cr-20Ni-2Si) is Cr2N-type nitrides were found in the region near
commonly used for wire mesh furnace belts. the surface, except alloy 600, which contains
Smith and Bucklin (Ref 48) investigated nitri- only about 16% Cr. This is in agreement with the
dation reactions in 100% N2 for several iron- and phase stability diagram at 1000 °C in terms of
nickel-base alloys. Their results, generated pN2 versus Cr content in Ni-Cr alloys as shown in
at 980, 1090, and 1200 °C (1800, 2000, and Fig. 4.4, which shows CrN is the most likely
2200 °F), are tabulated in Table 4.20. As shown nitride in Ni-16Cr alloy (Ref 12).
in the table, the nitridation kinetics in 100% N2 is Barnes and Lai (Ref 50) conducted an exten-
extremely rapid. Even nickel-base alloys were sive nitridation study in pure nitrogen atmo-
found to suffer severe nitridation attack even sphere for iron-, nickel-, and cobalt-base alloys
when the temperature was reduced to 980 °C at 1090 °C (2000 °F) for 168 h. Test results
(1800 °F). Both AlN and Cr2N were found in in terms of nitrogen absorption (mg/cm2) and
alloys 600 and 800 after exposure to 100% N2 at the depth of nitridation are summarized in
1200 °C (2200 °F) for 100 h. For RA330, only Table 4.22. As a result of rapid nitridation
Cr2N was detected after exposure to the same test kinetics under the test condition, nitridation
conditions. Ganesan and Smith (Ref 49) identi- attack penetrated through the thickness of the
fied the nitride phases formed near the surface of test specimen for many alloys. Due to different
the test specimens after exposure at 980 °C thicknesses for different alloys, the ranking of
(1800 °F) for 1008 h in pure nitrogen atmo- alloy performance in terms of nitridation depths
sphere using x-ray diffraction. The major phases became difficult for most alloys tested. (The
identified are summarized in Table 4.21 (Ref 49). thickness of the test specimen varied from alloy
to alloy, because of the use of whatever sheet
products were available for preparation of test
specimens.) Iron-base alloys, the last group from
RA330 to Type 310SS, suffered the worst nitri-
dation attack. Two cobalt-base alloys, alloys 188
(Co-22Cr-22Ni-14W-La) and 150 (Co-27Cr-
18Fe), exhibited poor resistance, with alloy 150
(high Cr and no Ni) showing extremely poor
nitridation resistance similar to iron-base alloys.
The nitride phases formed in alloys were
analyzed using x-ray diffraction performed on
the chemical extraction residues obtained from
the test specimens. Selected alloys (six nickel-
200 µm base alloys, two cobalt-base alloys, and one
(a) iron-base alloy) were analyzed, and the x-ray
diffraction analysis results are summarized in
Table 4.23. All the alloys except alloy 214
exhibited Cr2N nitrides. No internal nitrides
were observed in alloy 214, which contains 4.5%
Al. The alloy 214 specimen showed only surface
Al2O3 and AlN phases, as analyzed by x-ray
diffraction analysis performed on the surface
scales of the specimen. For nickel-base alloys
containing low levels of aluminum, such as
alloys 601 and 617 (both contain about 1.3% Al),
extensive AlN nitrides formed in metal interior.
Figure 4.33 shows a through-thickness nitri-
200 µm ded alloy 617 specimen, exhibiting extensive
(b) needle-shaped internal AlN nitrides along with
Cr2N nitrides. Figure 4.34 shows needle-shaped
Fig. 4.31 (a) Extensive internal aluminum nitride (long internal AlN nitrides as well as Cr2N nitrides at
needle phase) formation in alloy 601 and (b)
insignificant AlN formation in alloy 214 after exposure to NH3
high magnification in alloy 601. The addition
at 1090 °C (2000 °F) for 168 h of 4.5% Al to a nickel-base alloy can provide
Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/ 30/10/2007 3:56PM Plate # 0 pg 89

Chapter 4: Nitridation / 89

a very effective protection against nitridation was 0.3 mg/cm2, and no further increase in
attack at high temperatures. The nitride formed nitrogen absorption was observed after 504 h.
on alloy 214 surface after 168 h was too thin to Comparing alloy 214 with another nickel-base
be identified. However, after 500 h of exposure, alloy containing little aluminum, such as alloy
the surface scale was found to be composed 230, in 100% N2 at 1090 °C (2000 °F) for up to
of AlN and Al2O3, with AlN predominating 500 h of exposure clearly showed the superior
(Ref 50). The nitrogen absorbed after 168 h resistance of alloy 214 against nitridation attack
1.3 mm (50 mils)

100 µm

Fig. 4.32 Optical micrograph showing extensive internal chromium nitrides that formed in the entire cross section of a wire sample
obtained from a Type 314 wire mesh belt in a sintering furnace after service for 2 to 3 months at 1120 °C (2050 °F) in N2-10%
H2. Courtesy of Haynes International, Inc.
Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/ 30/10/2007 3:56PM Plate # 0 pg 90

90 / High-Temperature Corrosion and Materials Applications

Table 4.20 Nitridation resistance of iron- and nickel-base alloys in pure nitrogen
982 °C (1800 °F)/1008 h(a) 1093 °C (2000 °F)/900 h(b) 1204 °C (2200 °F)/100 h(c)
Alloy nitrided depth, mm (mils) nitrided depth, mm (mils) nitrided depth, mm (mils)
600 1.30 (51) 1.85 (73) 2.16 (85)
601 1.55 (61) 2.79 (110) >3.81 (150)
800 1.85 (73) >3.81 (150) >3.81 (150)
520 … >3.81 (150) …
330 2.57 (101) >3.81 (150) >3.81 (150)
DS … >3.81 (150) …
314SS >3.81 (150) >3.81 (150) …
(a) Specimens were cycled to room temperature once every 24 h for the first 3 days and then weekly for the remainder of the test. (b) Specimens were cycled to room
temperature once every 96 h (4 days). (c) Isothermal exposure. Source: Ref 48

Table 4.21 Major phases formed in the near- Table 4.23 Results of x-ray diffraction analysis of
surface region of the test specimens after exposure extraction residues obtained from specimens after
to 100% N2 at 980 °C (1800 °F) for 1008 h, as exposure to 100% N2 at 1090 °C (2000 °F) for
determined by x-ray diffraction 168 h
Alloy Major phases Alloy Phases detected
Type 314SS (Cr,Fe)2N 214 (a) AlN, Al2O3
Type 330SS (Cr,Fe)2N 230 Cr2N, (Cr,Mo)12(Fe, Ni)8-xN4-z, M6C
Alloy 800 (Cr,Fe)2N, AlN 600 Cr2N, TiN
Alloy 601 (Cr,Fe)2N, AlN 601 Cr2N, AlN
Alloy 600 CrN 617 Cr2N, AlN
Source: Ref 49 HR160 CrN, Cr2N
188 Cr2N
150 Cr2N
RA85H Cr2N, AlN
(a) Surface analysis. Source: Ref 50
Table 4.22 Nitrogen absorbed (mg/cm2) and the
average depth of internal nitridation for iron-,
nickel-, and cobalt-base alloys after exposure in
100% N2 at 1090 °C (2000 °F) for 168 h
Nitrogen absorbed, Depth of internal
Alloy mg/cm2 nitridation, mm
214 0.2 0.0
600 1.1 0.41
230 2.7 0.46
HR160 3.9 1.19
X 6.0 0.63
617 5.1 >0.58
601 7.2 >0.59
188 3.7 >0.51
150 9.0 >0.80
RA330 6.6 >1.52
RA85H 8.5 >1.44
556 9.0 >1.52
HR120 9.6 >0.86
253MA 10.0 >1.50
800H 10.3 >1.50
800HT 11.4 >1.46
Type 310 SS 12.3 >0.79
Source: Ref 50

Fig. 4.33 Optical micrographs showing a through-thickness

nitridation attack for alloy 617, a nickel-base alloy
with AlN/Al2O3 surface scales, as illustrated in containing about 1.3%Al, after exposure to 100% N2 at 1090 °C
Fig. 4.35. (2000 °F) for 168 h. Note extensive blocky chromium nitrides and
needle-shaped aluminum nitrides. Magnification bar represents
Alloy 150 (Co-27Cr-18Fe) suffered nitri- 200 μm. Courtesy of Haynes International, Inc.
dation attack as severe as that experienced by
some iron-base alloys. Figure 4.36 shows a
through-thickness nitrided alloy 150 compared observed throughout the alloy 150 specimen
with Type 310SS. Extensive Cr2N nitrides were cross section.
Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/ 30/10/2007 3:56PM Plate # 0 pg 91

Chapter 4: Nitridation / 91

Nickel-base alloys are in general more resis- showed the nitridation rate constants decreased
tant to nitridation attack than iron-base alloys. with increasing nickel concentration, as illu-
This is illustrated in Fig. 4.37 comparing alloy X strated in Fig. 4.38.
with 253MA after 168 h in 100% N2 at 1090 °C
(2000 °F). Similar findings were observed in
nitridation studies in nitrogen atmospheres by
Smith and Bucklin (Ref 48) and Tjokro and
Young (Ref 51). Tjokro and Young (Ref 51)
investigated a number of commercial alloys
in N2-5%H2 at 1100 and 1200 °C. Their results

(a)
200 µm
10 µm

Fig. 4.34 Scanning electron micrograph (backscattered

image) showing internal chromium nitrides
(blocky phases) and aluminum nitrides (long needle-shaped
phases) formed in alloy 601, a nickel-base alloy containing about
1.3% Al, after exposure to 100% N2 at 1090 °C (2000 °F) for
168 h. Source: Ref 50

4.50
230 alloy
4.00

3.50
N absorbed, mg/cm2

3.00

2.50

2.00

1.50

1.00

0.50
214 alloy
0.00
0.00 200.00 400.00 600.00 (b)
200 µm
Time, h
Fig. 4.36 Optical micrographs showing through-thickness
Fig. 4.35 Nitridation kinetic data for alloy 214 (nickel-base nitridation attack for (a) Type 310SS (Fe-25Cr-
alloy containing 4.5% Al) and alloy 230 (nickel- 20Ni) and (b) alloy 150 (Co-27Cr-18Fe) after exposure to 100%
base alloy containing little aluminum) after exposure to 100% N2 N2 at 1090 °C (2000 °F) for 168 h. Courtesy of Haynes Interna-
at 1090 °C (2000 °F) for 168 h. Source: Ref 50 tional, Inc.
Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/ 30/10/2007 3:57PM Plate # 0 pg 92

92 / High-Temperature Corrosion and Materials Applications

4.7 Nitridation Kinetics between NH3 (2000 °F) for 168 h in 100% NH3 (Ref 41) and
and N2 Atmospheres 100% N2 (Ref 50) are tabulated in Table 4.24.
The data are also presented in terms of nitrogen
Nitridation data generated in 100% NH3 absorption as a function of Ni + Co content in the
(Ref 41) and those generated in 100% N2 alloy (Fig. 4.39). The results clearly indicated
(Ref 50) were compared. Since both test pro- that the nitrogen atmosphere was a more severe
grams were generated using the same test nitriding environment than the ammonia environ-
apparatus and procedures, and both tests were ment at 1090 °C (2000 °F). The amount of
carried out by same technicians, the laboratory- nitrogen absorbed in N2 environment was more
to-laboratory variation was significantly mini- than double that in NH3 environment for many
mized. Thus, the comparison between these two alloys. Figure 4.40 shows two nitrided alloy
sets of test results could yield a more meaningful 601 specimens, one exposed to NH3 environ-
comparison in terms of the difference in environ- ment and the other to N2 environment. The
ments. The test results generated at 1090 °C N2 environment caused significantly more
internal nitride formation than for the NH3
environment. More chromium nitrides (blocky
shaped) and aluminum nitrides (needle shaped)
formed in the N2 environment than in the NH3
environment.
Ammonia readily dissociates to one part N2
and three parts H2 at 1090 °C (2000 °F). With
the test system used in the study by Barnes and
Lai (Ref 41), 100% NH3 was fed into the alumina
test tube with no test specimens inside, and the
exhaust gas was measured to contain less than
5% NH3, which was the detection limit of the the
apparatus used for measuring NH3 (Table 4.11).
It is believed most, if not all, of the ammonia
had been dissociated into H2 and N2 before the
test gas was in contact with the test specimens.
The NH3 test environment was essentially a
cracked ammonia, which was dissociated into
(a)
200 µm H2 and N2. Thus, the nitridation potential ( pN2 )
in the NH3 test environment (0.25 atm) was
much lower than that in the N2 test environment
(1.0 atm). As a result, the N2 test environment
was found to produce more severe nitridation
attack for most of the alloys tested (Table 4.24
and Fig. 4.39).

4.8 Summary
Nitridation behavior of metals and alloys in
(a) air, (b) gas-turbine combustion gas, (c)
NH3-H2O, (d) NH3, and (e) N2 environments
is reviewed. Nitridation attack can occur in air
and oxidizing, combustion environments. Under
certain conditions, alloys can suffer oxidation/
(b) 200 µm nitridation attack. Internal nitridation attack is
much more prevalent in a high-velocity com-
Fig. 4.37 Optical micrographs showing a through-thickness bustion gas stream with thermal cycling. In NH3-
nitrided Fe-20Cr-10Ni-1.7Si-Ce alloy 253MA (a)
and a better nitridation resistant Ni-22Cr-9Mo-18Fe-0.6W alloy X
H2O environments, alloys appear to behave
after exposure to 100% N2 at 1090 °C (2000 °F) for 168 h. differently under nitridation attack. Extensive
Courtesy of Haynes International, Inc. review is carried out on the behavior of metals
Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/ 30/10/2007 3:57PM Plate # 0 pg 93

Chapter 4: Nitridation / 93

6000

Intragranular
309S 1000 °C (1830 °F)
1100 °C (2010 °F)

4000

310S
kp, µm2/h

2000 153MA
RA330
253MA AC66
800
353MA
IN601
0
0 0.200 0.400 0.600 0.800 1.000
XNi'
(a)

12
Intergranular
153MA
1000 °C (1830 °F)
1100 °C (2010 °F)
253MA
800
kp, 103 µm2/h

309S

6 310S

RA330
AC66
353MA
IN601

0
0 0.200 0.400 0.600 0.800 1.000
XNi'
(b)

Fig. 4.38 Nitridation rate constants as a function of the alloy’s nickel concentration when tested in N2-5%H2 at 1000 and 1100 °C
(1830 and 2010 °F). Source: Ref 51

and alloys in NH3 and N2 environments. Com- 1979 Petten International Conference,
parative resistance to nitridation attack for a wide I. Kirman et al., Ed., The Metals Society,
variety of alloys is presented. London, 1980, p 45
4. K. Rorbo, Environmental Degradation
of High Temperature Materials, Series 3,
REFERENCES No. 13, Vol 2, The Institution of Metal-
lurgists, London, 1980, p 147
1. Metals Handbook, Vol 2, 8th ed., American 5. R.N. Shreve, The Chemical Process Indus-
Society For Metals, 1964, p 149 tries, McGraw-Hill, 1956
2. Metals Handbook, Vol 2, 8th ed., American 6. J.M.A. Van der Horst, Corrosion Problems
Society For Metals, 1964, p 119 in Energy Conversion and Generation,
3. G.L. Swales, Behavior of High Temperature C.S. Tedmon, Jr., Ed., The Electrochemical
Alloys in Aggressive Environments, Proc. Society, 1974
Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/ 30/10/2007 3:57PM Plate # 0 pg 94

94 / High-Temperature Corrosion and Materials Applications

Table 4.24 Nitrogen absorption in NH3 and N2

environments at 1090 °C (2000 °F) for 168 h
Nitrogen absorption, mg/cm2
Alloy NH3 N2
214 0.2 0.0
600 0.3 1.1
230 1.5 2.7
617 1.9(a) 5.1(a)
601 2.6(a) 7.2(a)
X 3.8(a) 6.0(a)
188 2.0(a) 3.7(a)
150 4.1 9.0(a)
556 4.2(a) 9.0(a)
RA330 3.1(a) 6.6(a)
800H 5.5 10.3(a)
253MA 6.3(a) 10.0(a)
(a) Nitrided all the way through specimens. Source: Ref 41, 50

13
1090 °C (2000 °F) / 168 h
N2 (a)
200 µm
12 NH3

10
Nitrogen absorption, mg/cm2

0
0 20 40 60 80
Ni or Ni + Co, wt% (b)
200 µm

Fig. 4.39 Nitrogen absorption as a function of Ni + Co con-

tent in the alloy for 100% NH3 and 100% N2 Fig. 4.40 Optical micrographs showing both chromium
environments at 1090 °C (2000 °F) for 168 h. Source: Ref 41, 50 nitride and aluminum nitride (needle-shaped
phase) formed in alloy 601 after exposure to (a) 100% NH3 and (b)
100% N2 at 1090 °C (2000 °F) for 168 h. Courtesy of Haynes
International, Inc.
7. M.B. Bever and C.F. Floe, Source Book on
Nitriding, American Society For Metals,
1977, p 125
8. B.J. Lightfoot and D.H. Jack, Source Book 11. T. Masumoto and Y. Imai, J. Jpn. Inst. Met.,
on Nitriding, American Society For Metals, Vol 33, 1969, p 1364
1977, p 248 12. H.J. Christ, S.Y. Chang, and U. Krupp,
9. K.N. Strafford, Corros. Sci., Vol 19, 1979, Thermodynamic Characteristics and Nu-
p 49 merical Modeling of Internal Nitridation of
10. T. Rosenquist, Principles of Extractive Nickel Base Alloys, Mater. Corros., Vol 54
Metallurgy, McGraw-Hill, 1974 (No. 11), 2003, p 887
Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/ 30/10/2007 3:57PM Plate # 0 pg 95

Chapter 4: Nitridation / 95

13. N.S. Corney and E.T. Turkdogan, The Effect Advances in the Technology of Stainless
of Alloying Elements on the Solubility Steels and Related Alloys, STP 369, ASTM,
of Nitrogen in Iron, J. Iron Steel Inst., Aug 1965, p 99
1955, p 344 31. M. Yu, R. Sandstrom, B. Lehtinen, and
14. D.H. Jack and K.H. Jack, Carbides and C. Westman, Scand. J. Metall., Vol 16, 1987,
Nitrides in Steel, Mater. Sci. Eng., Vol 11, p 154
1973, p 1 32. V. Guttmann and R. Burgel, Creep-
15. H.A. Wriedt and O.D. Gonzalez, Trans. Structural Relationship in Steel Alloy 800H
AIME, Vol 221, 1961, p 532 at 900–1000 °C, Met. Sci., Vol 17, 1983,
16. J.F. Eckel, T.P. Floridis, and B.N. Ferry, p 549
Nitrides in Type 304 Stainless Steel, Virginia 33. M. Welker, A. Rahmel, M. Schutze, Oxi-
J. Sci., Vol 17, 1966, p 325 dation and Nitridation of Alloy 800H at a
17. J.F. Eckel and T.B. Cox, J. Mater., Vol 3, Growing Creep Crack and for Unstressed
1968, p 605 Samples, Metall. Trans. A, Vol 20A, 1989,
18. L.E. Kindlimann and G.S. Ansell, Kinetics p 1541
of the Internal Nitridation of Austenitic 34. J.J. Hoffman and G.Y. Lai, Paper No. 5402,
Fe-Cr-Ni-Ti Alloys, Metall. Trans., Vol 1, Corrosion 2005, NACE International,
1970, p 163 2005
19. A.J. Heckler and J.A. Peterson, The Effect 35. V.P. Swaminathan and S.J. Lukezich,
of Nickel on the Activity of Nitrogen in Degradation of Transition Duct Alloys in
Fe-Ni-N Austenite, Trans. Metall. Soc. Gas Turbines, Advanced Materials and
AIME, Vol 245, 1969, p 2537 Coatings for Combustion Turbines, Proc.
20. J. Litz, A. Rahmel, M. Schorr, and J. Weiss, ASM 1993 Materials Congress Materials
Scale Formation on the Ni-Base Superalloys Week (Pittsburgh, PA), Oct 17–21, 1993,
IN 939 and IN 738LC, Oxid. Met., Vol 32, V.P. Swaminathan and N.S. Cheruvu, Ed.,
1989, p 167 ASM International, 1994, p 99
21. M.A. Harper, J.E. Barnes, and G.Y. Lai, 36. G.Y. Lai, Nitridation of Several Combustor
Long-Term Oxidation Behavior of Selected Alloys in a Simulated Gas Turbine Com-
High Temperature Alloys, Paper No. 132, bustion Environment, Advanced Materials
Corrosion/97, NACE International, 1997 and Coatings for Combustion Turbines,
22. G.Y. Lai, unpublished results, 2003 Proc. ASM 1993 Materials Congress
23. S. Han and D.J. Young, Simultaneous Materials Week (Pittsburgh, PA), Oct
Internal Oxidation and Nitridation of Ni- 17–21, 1993, V.P. Swaminathan and
Cr-Al Alloys, Oxid. Met., Vol 55, 2001, N.S. Cheruvu, Eds., ASM International,
p 223 1994, p 113
24. D.L. Douglass, Anomalous Behavior 37. G.Y. Lai, Nitridation Attack in a Simulated
During Internal Oxidation and Nitridation, Gas Turbine Combustion Environment,
JOM, Nov 1991, p 74 Materials for Advanced Power Engineering,
25. R.P. Rubly and D.L. Douglass, Oxid. Met., Part II, D. Coutsouradis et al., Ed., Kluwer
Vol 35, 1991, p 269 Academic Publishers, The Netherlands,
26. R.P. Rubly and D.L. Douglass, Internal 1994, p 1263
Nitridation of Ni-Cr-Al Alloys, Proc. Int. 38. G.Y. Lai, unpublished results, Haynes
Symp. On Solid-State Chemistry of Ad- International, Inc., 1995
vanced Materials: High-Temperature Cor- 39. Y.M. Park and R.E. Sonntag, Int. J. Energy
rosion Workshop, 1992 Res., Vol 14, 1990, p 153
27. H.J. Grabke and E.M. Peterson, Scr. Met., 40. H.J. Grabke, S. Strauss, and D. Vogel,
Vol 12, 1978, p 1111 Nitridation in NH3-H2O Mixtures, Mater.
28. J.-W. Park and C. J. Alstetter, Metall. Trans. Corros., Vol 54 (No. 11), 2003, p 895
A, Vol 18A, 1987, p 43 41. J.J. Barnes and G.Y. Lai, High Temperature
29. P.L. Gruzin, Y.A. Polikarpov, and G.B. Nitridation of Fe-, Ni-, and Co-base Alloys,
Federov, Fiz. Metal. I Metalloved., Vol 4 Corrosion & Particle Erosion at High
(No. 1), 1957, p 94 Temperatures, Proc. TMS-ASM Sympo-
30. K.G. Brickner, G.A. Ratz, and R.F. sium, V. Srinivasan and K. Vedula, Ed., The
Domagala, Creep-Rupture Properties of Minerals, Metals & Materials Society, 1989,
Stainless Steels at 1600, 1800, and 2000 °F, p 617
Name ///sr-nova/Dclabs_wip/High Temp/5208_67-96.pdf/Chap_04/ 30/10/2007 3:57PM Plate # 0 pg 96

96 / High-Temperature Corrosion and Materials Applications

42. K.M. Verma, H. Ghosh, and J.S. Rai, Brit. 48. G.D. Smith and P.J. Bucklin, Some
Corros. J., Vol 13 (No. 4), 1978, p 173 Observation on the Performance of
43. V. Cihal, Corrosion Mechanisms in Nickel-Containing Commercial Alloys in
Ammonia Synthesis Equipment, Conf. Nitrogen-Based Atmospheres, Paper No.
Proc., First International Congress on 375, Corrosion/86, NACE, 1986
Metallic Corrosion (London, U.K.), April 49. P. Ganesan and G.D. Smith, Performance of
10–15, 1961, L. Kenworthy, Ed., Butter- Selected Commercial Alloys in Nitrogen
worths, London, 1962, p 591 Based Sintering Atmospheres, Paper
44. J.J. Moran, J.R. Mihalisin, and E.N. Skinner, No. 278, Corrosion/90, NACE, 1990
Corrosion, Vol 17 (No. 4), 1961, p 191t 50. J.J. Barnes and G.Y. Lai, Factors Affecting
45. D.W. McDowell, Jr., Mater. Protect., Vol 1 the Nitridation Behavior of Fe-Base, Ni-
(No. 7), 1962, p 18 Base and Co-Base Alloys in Pure Nitrogen,
46. K.H. Jack, High Temperature Gas-Metal J. Physique IV, Colloque C9, supplemental
Reactions in Mixed Environments, au Journal de Physique III, Vol 3, 1993,
S.A. Jansson and Z.A. Foroulis, Ed., The p 167
Metallurgical Society of AIME, 1973, 51. K. Tjokro and D.J. Young, Comparison of
p 182 Internal Nitridation Reactions in Ammonia
47. H. Schenck, M.G. Frohberg, and F. and in Nitrogen, Oxid. Met., Vol 44, 1995,
Reinders, Stahl Eisen, Vol 83, 1963, p 93 p 453
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:21PM Plate # 0 pg 97

High-Temperature Corrosion And Materials Applications Copyright © 2007 ASM International®

George Y. Lai, editor, p97-145 All rights reserved.
DOI: 10.1361/hcma2007p097 www.asminternational.org

CHAPTER 5

Carburization and Metal Dusting

5.1 Introduction Production of carbon fibers also generates

carburizing atmospheres in a furnace. As a result,
Metals and alloys are susceptible to carbur- the furnace’s retorts, fixtures, and other compo-
ization when exposed to an environment con- nents require frequent replacement because of
taining CO, or CH4 or other hydrocarbon gases, carburization attack.
such as ethane (C2H6), propane (C3H8), and so Metal dusting, a form of catastrophic carbur-
forth, at elevated temperatures. Carburization ization, can occur at intermediate temperatures
attack generally results in formation of internal when a process gas stream consists of primarily
carbides, which often cause the alloy to suffer H2, CO, and CO2 along with some hydrocarbons
embrittlement as well as other mechanical with high carbon potentials (ac > 1). Metals or
property degradation. alloys can suffer rapid metal wastage in a form of
Carburization problems are quite common to pitting or general thinning of the cross-sectional
heat treating equipment, particularly furnace thickness of a metallic component. Metal dusting
retorts, baskets, fans, and other components used typically occurs at temperatures between 430
for case hardening of steels by gas carburizing. and 900 °C (800 and 1650 °F). Materials failures
A common commercial practice for control of associated with metal dusting have been en-
gas carburizing is to use an endothermic gas as a countered in refining and petrochemical proces-
carrier enriched with one of the hydrocarbon sing, such as production of syngas in hydrogen,
gases, such as CH4, C3H8, and so forth (Ref 1). ammonia, and methanol plants, heat treating, and
An endothermic gas enriched with about 10% other industrial processes (Ref 5–10).
natural gas (CH4) is a commonly used atmos-
phere (Ref 2). The typical endothermic gas
consists of 39.8% N2, 20.7% CO, 38.7% H2, and 5.2 Carburization
0.8% CH4, with a dew point of −20 to −4 °C
(−5 to +25 °F) (Ref 1). Gas carburizing occurs 5.2.1 Carburization—Thermodynamic
typically at 840 to 930 °C (1550 to 1700 °F). Considerations
Furnace equipment and components repeatedly
Whether an alloy is likely to be carburized or
subjected to these service conditions frequently
decarburized depends on the carbon activity (ac)
suffer brittle failures as a result of carburization
in the environment and that of the alloy. The
attack.
thermodynamic condition that dictates either
In the petrochemical industry, carburization is carburization or decarburization can be described
one of the major modes of high-temperature
simply.
corrosion for processing equipment. The pyro-
The alloy is likely to be carburized when:
lysis furnace tubes for production of ethylene and
olefins are a good example (Ref 3–5). Ethylene is (ac )environment 4 (ac )metal
formed by cracking petroleum feedstock, such as
ethane and naphtha, at temperatures up to 1150 ° The alloy is likely to be decarburized when:
C (2100 °F). This generates a strong carburizing (ac )environment 5 (ac )metal
gas stream inside the tubes. As a result, carbur-
ization was found to be a major mode of tube Thus, in order to predict whether an alloy
failure in a survey of ethylene and olefin pyro- will be carburized, one needs to know the
lysis furnaces conducted by Moller and Warren carbon activities of both the environment and
(Ref 3). the alloy.
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:21PM Plate # 0 pg 98

98 / High-Temperature Corrosion and Materials Applications

Carburization can proceed by one of the fol- Similarly, if carburization follows Reaction
lowing reactions when the environment contains 5.2, the carbon activity of the environment can
CH4, CO, or H2 and CO: also be calculated:


CO+H2 =C+H2 O ðEq 5:1Þ ac  pCO2
DG
=7RT ln ðEq 5:6Þ
p2CO
2CO=C+CO2 ðEq 5:2Þ



p2CO
ac =e7DG =RT
ðEq 5:7Þ
CH4 =C+2H2 ðEq 5:3Þ pCO2

Assuming that carburization follows Reaction Plots of carbon activities as a function of gas
5.1, the carbon activity in the environment can be compositions in terms of (p2CO =pCO2 ) for various
calculated by: temperatures are shown in Fig. 5.2.

 When carburization follows Reaction 5.3, the
ac  pH2 O
DG
=7RT ln ðEq 5:4Þ carbon activity in the environment is:
pCO  pH2 !
7DG
=RT pCH4
Rearranging the equation changes it to: ac =e ðEq 5:8Þ

 p2H2

pCO  pH2
ac =e7DG =RT
ðEq 5:5Þ
pH2 O Carbon activities as a function of ( pCH4 =p2H2 ) are
plotted in Fig. 5.3.
From Eq 5.5, one can construct graphs of carbon Reactions 5.1 and 5.2 have a similar charac-
activity as a function of gaseous composition teristic, showing lower carbon activities with
in terms of ( pCO  pH2 =pH2 O ) ratios for various increasing temperature (Fig. 5.1 and 5.2).
temperatures, as shown in Fig. 5.1.

H2 = C + H2O 2CO = CO2 + C

Fig. 5.1 Carbon activity (ac) as a function of gaseous com- Fig. 5.2 Carbon activity (ac) as a function of gaseous com-
2
position in terms of (pCO  pH2 =pH2 O ) ratios based on position in terms of (pCO =pCO2 ) based on Eq 5.2 for
Eq 5.1 for various temperatures. Also plotted are carbon activities various temperatures. Also plotted are carbon activities for carbon
for carbon steel (in equilibrium with Fe3C), and for 2.25Cr-1Mo steel (in equilibrium with Fe3C), and for 2.25Cr-1Mo and auste-
and austenitic stainless steels (both measured ac). nitic stainless steels (both measured ac).
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:21PM Plate # 0 pg 99

Chapter 5: Carburization and Metal Dusting / 99

Reaction 5.3, on the other hand, shows increased H2O, and under very dynamic conditions with a
carbon activities with increasing temperature high gas velocity such that the gaseous com-
(Fig. 5.3). If the environment contains CH4, ponents do not have time to react to reach a
the carbon activity of the environment at higher thermodynamic equilibrium (i.e., nonequilibrium
temperatures is likely to be dominated by Re- conditions), the gas-metal reaction can be rea-
action 5.3. When no CH4 is present in the sonably assumed to follow the dominating
environment, Reaction 5.1 and/or 5.2 will dictate reaction from one of those shown in Reaction
the carbon activity. 5.1, 5.2, or 5.3, and thus, one of the activity maps
The carbon activity maps shown in Fig. 5.1 to shown in Fig. 5.1, 5.2, or 5.3.
5.3 were previously described by Mazandarany Data for carbon activities of commercial alloys
and Lai (Ref 11) in assessing the carburization- at temperatures below 1200 °C (2190 °F) is very
decarburization behavior of alloys in high- limited. Natesan (Ref 12) reported that ac for
temperature gas-cooled helium environments 2.25Cr-1Mo steel is in the range of 1×10−1 to
containing H2, CO, CO2, CH4, and H2O. These 10−2 from 550 to 750 °C (1020 to 1380 °F).
activity maps provide a simple means of esti- Natesan and Kassner (Ref 13) reported the
mating an environment’s carbon activity for carbon activities of Fe-18Cr-8Ni alloys. These
predicting whether or not the environment is values are superimposed in Fig. 5.1 to 5.3.
thermodynamically capable of carburizing an For carbon steels, carbon activity can be esti-
alloy. mated by assuming that it is in equilibrium with
When the gas stream contains many gaseous cementite (Fe3C):
components, such as H2, CO, CO2, CH4, and
3Fe+C=Fe3 C ðEq 5:9Þ
 
aFe3 C
DG
=7RT ln ðEq 5:10Þ
(ac )  (aFe )3


1
DG
=7RT ln ðEq 5:11Þ
ac

where aFe3 C and aFe are assumed to be unity.

ac =eDG =RT
ðEq 5:12Þ

The ac values for carbon steel based on Eq 5.12

are plotted in Fig. 5.1 to 5.3. Using Fig. 5.1, 5.2,
or 5.3 one can make a quick determination as to
whether an environment has a carbon potential
(or activity) high enough to carburize the alloy of
interest.
Even though in cases where the gas mixture
may not reach an equilibrium condition, it will be
of great benefit to better understand the gas-metal
reaction in terms of the thermodynamic equili-
brium condition in multicomponent gases
environments. The thermodynamic equilibrium
CH4 = 2H2 + C gaseous composition along with its thermo-
dynamic potentials, such as carbon activity (ac),
oxygen potential ( pO2 ), and other potentials, can
be determined using a commercial software
program such as “HSC Chemistry for Windows”
(Ref 14) and ChemSage (Ref 15).
The environment can also be characterized in
Fig. 5.3 Carbon activity (ac) as a function of gaseous com-
2 terms of ac and pO2 to determine the relative
position in terms of (pCH4 =pH 2
) based on Eq 5.3 for
various temperatures. Also plotted are carbon activities for carbon severity of its carburization potential. Both car-
steel (in equilibrium with Fe3C), and for 2.25Cr-1Mo (measured
ac). Carbon activities of austenitic stainless steels are below 10−2 at
bon activity and oxygen potential can be calcu-
800 to 1000 °C (1470 to 1830 °F). lated by a computer program. The environment
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:21PM Plate # 0 pg 100

100 / High-Temperature Corrosion and Materials Applications

can then be presented in a stability diagram of During carburization, the relative stabilities of
a metal-carbon-oxygen system. The stability these carbides can be best described by a stability
diagrams of Fe-C-O and Cr-C-O systems are diagram, such as the one shown in Fig. 5.5. If the
shown in Fig. 5.4 and 5.5 (Ref 16). From the carbon and oxygen activities of the environment
stability diagram, the possible phases that the are in the Cr3C2 region, conditions will favor
alloy may form at the gas/metal interface can be formation of Cr3C2 on the surface and/or in the
predicted. As the activities of both carbon and underlying metal. As carbon diffuses farther into
oxygen are decreasing from the gas/metal inter- the alloy’s interior, carbon activities will be
face to the metal interior, the possible phases lowered, thus favoring Cr7C3. Moving even
that the alloy may form beneath the gas/metal farther into the interior, carbon activities will
interface can also be predicted. be further reduced, favoring the formation of
For carbon and alloy steels with low con- Cr23C6. These chromium carbides can incorpo-
centrations of chromium, ingress of carbon into rate other alloying elements depending on alloy
the metal or alloy may result in the formation of system. For example, in Fe-Ni-Cr system, iron
iron carbides. Several forms of iron carbides with very little nickel can be incorporated into
have been reported (Ref 17), with compositions these chromium carbides. The combined metal
ranging from Fe4C to Fe2C. They are ξ phase elements in the carbide are then represented by
(Fe4C), θ phase (Fe3C), χ phase (Fe2.2C), and ε “M,” as M3C2, M7C3, and M23C6. An example
phase (Fe2-3C). Fe3C (cementite) is the most of the metallic compositions of M7C3 and M23C6
stable iron carbide. Other iron carbides are less formed in carburization of Type 304L is
stable. Browning et al. (Ref 18) found that χ illustrated in Fig. 5.6 (Ref 20). Both M7C3 and
phase, which formed by carburizing αFe with M23C6 contain essentially Cr and Fe with
butane at 275 °C (530 °F), was converted to negligible amount of Ni.
Fe3C when heated to 500 °C (930 °F). The ε For many high-temperature alloys, parti-
phase (Fe2-3C) is a transition phase that forms in cularly superalloys, there are other alloying
martensite during tempering of steel (Ref 19). elements, such as Ti, Ta, Nb (or Cb), Mo, and W,
In ferritic and austenitic stainless steels and that can form carbides. The carbides of these
nickel- and cobalt-base alloys, ingress of carbon alloying elements are important to the physical
into the alloy results in the formation of mainly metallurgy of high-temperature alloys in
chromium carbides. There are three forms of that they provide an important strengthening
chromium carbides: Cr23C6, Cr7C3, and Cr3C2. mechanism. A general review of binary metallic
2
Log PCH /PH 2
2
Log PCO /PCO
4
2

Log PCO /PCO –18 –16 –14 –12 –10 –8 –6 –4 –2 0 2 4 6 8 10 12

2
Log P H2O/P –18 –16 –14 –12 –10 –8 –6 –4 –2
H2
0 2 4 6 8 10 12
0
2
0 C(s)
–2
0
–2
Fe3C(s) –4
–2
–4
–6
–4
–6
Log aC

–8
Fe(s) Fe3C4(s) Fe2O3(s) –6
–8 –10
Fe0.95O(s) –8
–10 –12
–10
–12 –14
–12
–14 –16
–14
–16
–50 –45 –40 –35 –30 –25 –20 –15 –10 –5 0 5
Log PO , atm
2

Fig. 5.4 Stability diagram of Fe-C-O system at 870 °C (1600 °F). Source: Ref 16
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:21PM Plate # 0 pg 101

Chapter 5: Carburization and Metal Dusting / 101

2
Log PCH /PH2
2
Log PCO /PCO
4
2
Log PCO /PCO –14 –12 –10 –8 –6 –4 –2 0 2 4 6 8 10 12 14
2
Log P /P –14 –12 –10 –8 –6 –4 –2 0 2 4 6 8 10 12 14
H2O H2
Cr3C2(s) 0 0
0 C(s)
–2 –2
–2 Cr7C3(s)

–4 –4
–4
Cr23C6(s) –6 –6
–6
Log aC

–8 –8
–8
Cr2O3(s)
–10 –10 –10
Cr(s)

–12 –12 –12

–14 –14 –14

–16 –16 –16

–50 –45 –40 –35 –30 –25 –20 –15 –10 –5 0 5
(a) Log PO , atm
2

2
Log PCH /PH 2
2
Log PCO /PCO
4
2
Log PCO /PCO –18 –16 –14 –12 –10 –8 –6 –4 –2 0 2 4 6 8 10 12
2
Log P /P –18 –16 –14 –12 –10 –8 –6 –4 –2 0 2 4 6 8 10 12
H2O H2
0
Cr3C2(s) 2
0 C(s) –2
Cr7C3(s) 0
–2 –4
–2
–4 –6
Cr23C6(s) –4
–6 –8
Log aC

–6
–8 –10
Cr2O3(s) –8
–10 Cr(s)
–12
–10
–12 –14
–12
–14 –16
–14
–16
–50 –45 –40 –35 –30 –25 –20 –15 –10 –5 0 5
(b) Log PO , atm
2

Fig. 5.5 Stability diagrams of Cr-C-O system at (a) 620 °C (1150 °F), (b) 870 °C (1600 °F), and (c) 1090 °C (2000 °F). Source: Ref 16

carbides can be found elsewhere (Ref 21, 22). to keep the reaction going from left to right (i.e.,
The relative stabilities of some binary carbides keeping Cr3C2 stable), the pO2 of the environ-
are shown in Fig. 5.7 (Ref 22). ment shall be lower than the equilibrium pO2
When the environment contains oxygen and associated with the above reaction. On the other
carbon activities, temperature is an important hand, if the pO2 of the environment is higher than
factor in determining whether the oxide or car- the equilibrium pO2 associated with the above
bide will be thermodynamically stable. Con- reaction, Cr2O3 will become stable. Temperature
sidering a chemical reaction, such as 3Cr2O3 + can be a significant factor in determining whether
4C = 2Cr3C2 + 9/2 O2, Cr3C2 will remain stable chromium oxide or chromium carbide is stable,
when the reaction goes from left to right. In order and thus significantly affects the carburization
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:21PM Plate # 0 pg 102

102 / High-Temperature Corrosion and Materials Applications

2
Log PCH /PH 2
2
Log PCO /PCO
4
2
Log PCO /PCO –20 –18 –16 –14 –12 –10 –8 –6 –4 –2 0 2 4 6 8 10
2
Log P /P –20 –18 –16 –14 –12 –10 –8 –6 –4 –2
H2O H2
0 2 4 6 8 10
4
Cr3C2(s) –2
0 C(s)
2
Cr7C3(s) –4
–2
0
–4 Cr23C-6(s) –6
–2
–8
Log aC

–6
–4
Cr2O3(s) –10
–8
Cr(s) –6
–12
–10
–8
–14
–12
–10
–16
–14
–12
–16 –18
–50 –45 –40 –35 –30 –25 –20 –15 –10 –5 0 5
(c) Log PO , atm
2

Fig. 5.5 (continued) (c) 1090 °C (2000 °F). Source: Ref 16

100 Thus, Cr2O3 is stable. On the left side of the

1123 k intersections (i.e., higher temperatures), pO2
80 150 h (environment) is lower, thus favoring the for-
Concentration, wt.%

mation of carbides, but not oxides. Accordingly,

60 higher temperatures favor carburization thermo-
M23C6
Cr Fe
dynamically and lower temperatures favor
Ni
40 formation of oxides thus retarding carburization.
M7C3
Nishiyama et al. (Ref 24) examined the effect
20 of the temperature on the stability of chromium
oxide and carbides based on the ethylene
0 pyrolysis environment, which is generated by the
0 0.2 0.4 0.6 0.8 1
reaction of naphtha with steam for the production
Distance to surface, mm
of ethylene (C2H4) and propylene (C3H6) at
Fig. 5.6 Compositions of the metallic components of M7C3 temperatures of approximately 900 to 1100 °C
and M23C6 formed in Type 304L after carburizing at (1650 to 2012 °F). In their calculation of the
1123 K (850 °C) in H2-2.6CH4 (ac = 0.9) for 150 h. Source: Ref 20
oxygen potential for the reaction of naphtha with
steam, three naphtha feedstocks were used with
behavior of an alloy. This can be nicely illu- the steam/naphtha weight ratios of 0.4 and 0.5.
strated in a plot that contains the oxygen poten- The calculated pO2 for the pyrolysis environment
tials of the environment (Boudouard reaction as a function of temperature is plotted in Fig. 5.9.
[2CO = C + CO2] is used for this example) and Also plotted in Fig. 5.9 are pO2 values in equili-
those in equilibrium with Cr2O3/Cr3C2 as a brium with Cr3C2/Cr2O3 and those in equili-
function of temperature. This is shown in Fig. 5.8 brium with Cr7C3/Cr2O3. The results in Fig. 5.9
(Ref 23). The figure shows the equilibrium pO2 are very similar to those shown in Fig. 5.8, where
line of Cr2O3/Cr3C2 intersecting with the en- the environment was calculated from CO-CO2
vironment’s pO2 lines (pCO = 1.0 bar, pCO = reaction (Boudouard reaction). The calculation
0.5 bar, and pCO = 0.25 bar). The intersections by Nishiyama et al. showed that the chromium
are between 1000 and 1200 °C. On the right side oxide was stable up to 1030 to 1040 °C (1886
of the intersections (i.e., lower temperatures), pO2 to 1904 °F) and became unstable above those
(environment) is greater than pO2 (Cr2O3/Cr3C2). temperatures. At temperatures above 1030 to
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:21PM Plate # 0 pg 103

Chapter 5: Carburization and Metal Dusting / 103

Fig. 5.7 Standard free energies of formation for carbides. Source: Ref 22

1040 °C (1886 to 1904 °F), carbides such as the common operating ratios of 0.35 to 0.5, the
Cr3C2 and Cr7C3 became stable. This oxide- oxide-carbide transition temperature is decreased
carbide transition temperature can vary depend- to about 970 °C (1778 °F).
ing on the steam/naphtha ratio, as illustrated in In Reactions 5.1 to 5.3, carbon deposition
Fig. 5.10 (Ref 24). As shown in the figure, when (coking) can occur when the carbon activity (ac)
the steam/naphtha ratio is decreased to 0.1 from in the environment is greater than 1.0 (ac = 1
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:22PM Plate # 0 pg 104

104 / High-Temperature Corrosion and Materials Applications

in equilibrium with graphite). Many laboratory (Ref 25) recommended that the H2-1%CH4
carburization tests have been conducted in mixture with ac <1.0 be used for laboratory
H2-CH4 mixtures. Figure 5.11 shows the carbon testing to correspond to the industrial process that
activities of H2-1%CH4 and H2-2%CH4 as does not form coking. He also recommended the
a function of temperature (Ref 25). Grabke use of the H2-2%CH4 mixture with ac > 1 for
testing to correspond to the processes where
coking is taking place. The test temperature is
T, °C thus recommended to be higher than 1000 °C
(1832 °F) (see Fig. 5.11). Ethylene pyrolysis
1800
1600
1400
1200

1000

800

600
environment is known to develop coking on
the internal surface of the pyrolysis furnace
–10 Cr3C2/Cr2O3 tubes. The deposition of carbon is the result of
decomposition of ethylene (C2H4) in the reaction
as described in Eq 5.13 (Ref 26). Significant
Cr3C2 PCO–1 bar coking in the internal surface of the ethylene
–15 pyrolysis furnace tube and the repeated decoking
Log PO , bar

PCO–0.5 bar
PCO–0.25 bar
2

–20
C/CO-CO2

–25 Cr2O3

5.10–4 75.10–4 10.10–4 12.5.10–4

1/T, K–4

Fig. 5.8 The oxygen potentials of the environment based on

the Boudouard reaction (2CO = C + CO2) and those
in equilibrium with Cr2O3/Cr3C2 as a function of temperature.
Source: Ref 23

T, °C
0,150 1,100 1,050 1,000 950 900 850
–17 Fig. 5.10 Equilibrium pO2 of the environment based on
ethylene pyrolysis of naphtha I with various steam/
naphtha weight (S/O) ratios as a function of temperature, and pO2
Cr-carbides + metastable Cr2O3 in equilibrium with Cr3C2/Cr2O3 and Cr7C3/Cr2O3 as a function of
–18 temperature. Source: Ref 24

Cr2O3+ metastable Cr-carbides

Log PO .atm

–19
ac
2

Carbon activity in H2-CH4

5 1 bar
–20 II (0
Cr 7

Cr 3

.5)
and
C3

C2

Na III ( 4 2% CH4
/C

/C

pht 0.5
r 2O

r 2O

–21 ha )
I (0
.4) 3
3

3
(a c

Cr-carbides
(a c
=1

=1

2 1% CH4
–22
)

7 7.5 8 8.5 9
1/T × 104, 1/K 1

Fig. 5.9 Equilibrium pO2 of the environment based on ethy-

lene pyrolysis of naphtha with steam/naphtha ratios 800 850 900 950 1000 1050 1100 1150 °C
of 0.4 and 0.5 (in parentheses) as a function of temperature,
and pO2 in equilibrium with Cr3C2/Cr2O3 and Cr7C3/Cr2O3 as a Fig. 5.11 Carbon activities (ac) of H2-1%CH4 and H2-2%CH4
function of temperature. Naphtha I, II, and III represent different are plotted as a function of temperature. Source:
naphtha feed stocks. Source: Ref 24 Ref 25
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:22PM Plate # 0 pg 105

Chapter 5: Carburization and Metal Dusting / 105

operation to remove this coke deposit can the carburized alloy. Each evaluation method
significantly degrade the tube life (Ref 27). has its merit. It certainly will be beneficial to
use as many evaluation methods as possible for
C2 H4 =2C+2H2 ðEq 5:13Þ characterizing the carburized alloy.
With respect to the impact of carburization on
an alloy’s performance, Krikke et al. (Ref 28)
5.2.2 Resistance to Carburization believed that not only the total amount of carbon
Carburization attack generally results in the absorbed but also the maximum carbon level and
formation of internal carbides in the alloy matrix the maximum carbon concentration gradient are
as well as at grain boundaries. The gravimetric the most important factors. Figure 5.12 illustrates
method has been widely used for studying car- three possible carbon concentration profiles,
burization kinetics. This method can sometimes as suggested by Krikke et al. (Ref 28). They
produce a misleading result when the environ- considered profile A with a steep concentration
ment exhibits an oxygen potential high enough to profile to be the most damaging. Heubner (Ref
form oxides of some active alloying elements. 29) tested various commercial alloys in H2-CH4
The weight gain, in this case, is the result of both gas mixture (ac = 0.8) at 1000 °C (1832 °F) and
carbon ingress and oxide formation. Measure- observed a steep carbon concentration profile in
ments of carburization depth have also been used Fe-Ni-Cr alloys and a low flat concentration
by some investigators. Different alloy systems profile in Ni-Cr alloys, as shown in Fig. 5.13.
can produce significant differences in the con- One Ni-Cr alloy (alloy 45TM), which contained
centration profile for the carburized layer. Thus, relatively high Fe and high Si, was an exception
one alloy may exhibit a large carburization depth showing a steep carbon concentration profile
with only a slight concentration gradient, while similar to Fe-Ni-Cr alloys such as 800H, AC66,
another alloy may show a narrow carburization and DS. The Fe-Ni-Cr alloys were found to have
depth with a steep concentration profile. Fur- suffered more room-temperature impact tough-
thermore, measurements of carburization depth ness drop in general than Ni-Cr alloys (Ref 29).
by the metallographic method can be difficult However, the relative room-temperature impact
when separating the carbides formed by carbur- toughness loss (%) was found to increase with
ization from those formed by thermal aging. increasing total carbon pickup (Ref 29). For
Some investigators measured the total amount of carburization, the real issue is the effect of car-
carbon in the alloy after the exposure. The mea- burization on the alloy’s mechanical properties,
surement of the carbon concentration profile as such as creep-rupture properties and toughness
a function of distance from the metal surface or ductility. This type of data, however, is quite
may be an excellent method for characterizing limited and is inadequate as a basis for making an
informed materials selection. Thus, this chapter
reviews mainly the carburization data in terms of
mass gain, mass of carbon absorption, carbur-
5
A ization depth, and concentration profile of the
carburized layer.
When the environment is such that no pro-
Carbon content, %

4
tective oxide scale (e.g., Cr2O3 scale) is formed
3 B on the metal surface, carburization is controlled
by diffusivity and solubility (Ref 30). The ingress
of carbon will be greatly reduced when a chro-
2 mium oxide scale is developed. Carburization
C kinetics in this case are then controlled by the
1 diffusion of carbon through the oxide scale. Wolf
and Grabke (Ref 31) demonstrated that there was
no detectable solubility of carbon in Cr2O3 oxi-
0 des by equilibrating the oxides with CO2-CO
ID Distance from bore of tube, mm OD mixtures tagged with radioactive 14C at 1000 °C
Wall thickness (1832 °F). Thus, carbon permeation is not pos-
sible through the perfectly dense Cr2O3 oxide
Fig. 5.12 Three possible carbon concentration profiles of layer, unless the oxide layer contains pores and
carburized alloys. Source: Ref 28 fissures (Ref 31).
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:22PM Plate # 0 pg 106

106 / High-Temperature Corrosion and Materials Applications

Resistance to carburization is an important alloy compared with the as-cast surface finish
factor in the performance of pyrolysis furnace (Fig. 5.17).
tubes as well as pigtails for ethylene and olefin Norton and his colleagues (Ref 35–37) at
plants. Furnace tubes are typically constructed of Petten Laboratories have conducted a series of
Fe-Ni-Cr cast alloys, such as HK (Fe-25Cr-20Ni studies on the effects of silicon, niobium, chro-
or 25/20), HP (Fe-25Cr-35Ni or 25/35) and their mium, and iron in Fe-Ni-Cr alloys, including
variants. Some of these variants involved addi- four commercial alloys (HK-40, HP-40Nb, Type
tions of niobium (or columbium), tungsten, 314 SS, and alloy 800H) and three experimental
molybdenum, silicon, and titanium. Some also alloys. Their test environments had fixed carbon
involved increases in nickel and/or chromium. activities (ac) of 0.3 and 0.8, with various oxygen
Some of the modified alloys are referred to as potentials (pO2 ) at temperatures from 825 to
“microalloyed” castings. It has been found that 1050 °C (1520 to 1920 °F), as illustrated in
these additions and increases improve carbur- Fig. 5.18. The oxygen potentials of the test
ization resistance as well as creep-rupture environments were below that in equilibrium
strengths. Figures 5.14 to 5.17 illustrate the car- with Cr2O3 (Fig. 5.18). That means that no
burization resistance of some of these modified chromium oxide scales should have formed on
alloys compared with HK alloy (Ref 32–34). the metal surface. However, a SiO2 scale was
Also shown in Fig. 5.17 is the effect of the likely to form at 825 °C (1520 °F), but not at
surface finish on carburization resistance of the 1000 °C (1830 °F). The test results at 825 °C
alloy. The machined-finished surface signifi- (1520 °F) showed that Type 314 stainless steel
cantly reduced the carbon ingress into the (2.04% Si) was significantly more carburization

3.5

3 AC 66
Carbon concentration, %

2.5 alloy 800 H

2 alloy DS

1.5

0.5

0
0 1 2 3 4 5 ∞
Distance to surface, mm
(a)

2.5
45 TM

2
Carbon concentration, %

1.5

1
alloy 600H

0.5
alloy 617
alloy 602 CA
0
0 1 2 3 4 5 ∞
Distance to surface, mm
(b)

Fig. 5.13 Carbon concentration profile for alloys after testing at 1000 °C (1832 °F) for 1008 h in a H2-CH4 mixture (ac = 0.8) for (a)
Fe-Ni-Cr and (b) Ni-Cr alloys. Source: Ref 29
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:22PM Plate # 0 pg 107

Chapter 5: Carburization and Metal Dusting / 107

resistant than HK-40 (1.35% Si), HP-40Nb formers, constituting the major elements in M7C3
(1.29% Si), and alloy 800H (0.4% Si), as shown and M23C6 carbides resulting from carburization.
in Fig. 5.19. Also revealed by the test results was Harrison et al. (Ref 37) found that the surface
that the model alloys were significantly less carbides removed from HK-40 sample after
resistant to carburization than the commercial testing at 1000 °C (1830 °F) contained 55 to
alloys with the same chromium, nickel, and iron. 58% Cr and 41 to 43% Fe, with very little nickel
The model alloys had much lower silicon levels (approximately Cr4Fe3C3). Nickel reduces the
(0.13–0.28%) as well as manganese, aluminum, diffusivity of carbon in Fe-Ni-Cr alloys, as
titanium, and so forth. The beneficial effect of demonstrated by Demel et al. (Ref 38) in
silicon on carburization resistance was clearly Fig. 5.23. Nickel also decreases the solubility of
demonstrated when the data were replotted carbon in Fe-Ni alloy system as shown in
(Fig. 5.20). The presence of a SiO2 scale was Fig. 5.24 (Ref 39). Decreases in carbon diffu-
confirmed by Van der Biest et al. (Ref 36). When sivity and solubility can result in increases in
the test temperature was increased to 1000 °C carburization resistance. Grabke et al. (Ref 40)
(1832 °F), where the pO2 was below that in observed that increasing nickel improved car-
equilibrium of SiO2 (Fig. 5.18), Type 314 SS was burization resistance in Fe-Ni-Cr alloys, with the
found to be similar to HK-40, alloy 800H, and maximum resistance achieved when the ratio of
HP-40Nb (Fig. 5.21). Under these conditions, Ni to Fe was 4 to 1. This is in general agreement
SiO2 was no longer thermodynamically stable. with the product of carbon solubility and diffu-
Thus, the silicon effect was diminished. The sivity (Ref 41). High-nickel alloys are generally
model alloy 50/50 (50Ni-50Cr with very low more resistant to carburization than Fe-Ni-Cr
silicon level) was among the best performers in alloys. This is illustrated by the test results of
the alloys tested. Klower and Heubner (Ref 29), as shown in
The ratio of Ni to Cr + Fe is an important factor Fig. 5.25. The beneficial effect of nickel on car-
(Ref 35) in governing carburization resistance, as burization resistance can also be clearly revealed
shown in Fig. 5.22. Decreases in Cr+ Fe in in Fig. 5.26 when the data of Fe-Ni-Cr alloys
Fe-Ni-Cr alloys improved carburization resis-
tance. Both chromium and iron are carbide

Fig. 5.14 Carburization resistance of HK (25Cr-20Ni) and Fig. 5.15 Carbon concentration profiles for HK (25Cr-20Ni)
several HP alloys (Cr/Ni) as a function of tem- and several HP alloys (Cr/Ni) carburized at 1100 °C
perature in pack carburization tests. Source: Ref 32 (2010 °F) for 520 h in pack carburization tests. Source: Ref 32
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:22PM Plate # 0 pg 108

108 / High-Temperature Corrosion and Materials Applications

and Ni-Cr alloys generated at 1000 °C for containing 26 to 45% Ni. For alloys containing
1008 h were plotted as a function of iron content 46 to 70% Ni, increasing chromium resulted in an
(i.e., decreasing nickel content) (Ref 29). The increase in carbon pickup. In this study, although
oxygen partial pressure (pO2 ) of the test en- there is no mention of the oxygen potential of the
vironment—which was 10−27.92, 10−25.25, test environment, it is believed that the chromium
and 10−23.88 for 850, 1000, and 1100 °C, oxide scale was not involved in the carburization
respectively—was below the pO2 in equilibrium reaction. Wolfe (Ref 43) also found that a higher
of Cr2O3 (Ref 29). Accordingly, no chromium Ni-containing Fe-Cr-Ni cast alloy, alloy HU
oxide scale would be present to provide protec- (Fe-19Cr-39Ni-0.5C), was significantly better
tion for the alloys in these tests. than Type 304 (19Cr-9Ni) and Type 321 (18Cr-
In an extensive study undertaken by Steel and 12Ni) with similar amounts of chromium but
Engel (Ref 42) on the relative influence of nickel lower nickel. Even a low Cr-containing Fe-Cr-Ni
and chromium on the carburization resistance of cast alloy, alloy HT (Fe-15Cr-35Ni-0.5C) was
Fe-Ni-Cr alloys, standard heats of ASTM grades significantly better than both Types 304 and 321
varying from HC to HX cast alloys were (Ref 43). These data are shown in Fig. 5.29.
investigated, along with many experimental cast Small additions of some minor elements,
alloys. They found nickel to be beneficial, as such as titanium, niobium, tungsten, and rare
illustrated in Fig. 5.27. The role of chromium, earth elements, may also improve an alloy’s
however, appeared to be different for different resistance to carburization in test environments
levels of nickel, as shown in Fig. 5.28. For iron- of H2-8.6%CH4-7%H2O and H2-12%CH4-10%
base alloys with 25% or less nickel, increasing H2O. This is illustrated in Table 5.1 (Ref 44).
chromium significantly reduced carbon pickup. The superior carburization resistance of TMA
A slight decrease in carbon pickup with 4750 alloy to HK-40 (2% Si) can be attributed to
increasing chromium was noted for alloys small additions of titanium, niobium, tungsten,

Fig. 5.16 Carbon concentration profiles of HK and HP alloys tested at 1050 °C (1920 °F) for 1200 h in 37%N2-40%H2-20%CO-3%
CH4 (ac = 1.0, pO2 = 3:4 · 10720 atm). Source: Ref 33
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:22PM Plate # 0 pg 109

Chapter 5: Carburization and Metal Dusting / 109

and rare earth elements. In both test environ- that in equilibrium with SiO2 (i.e., SiO2 could not
ments, the oxygen potentials, although not dis- form). Thus, silicon played no role in carbur-
cussed, are believed to be high enough to form ization resistance. At high oxygen potentials
oxide scales. Thus, improvements in carburiza- (i.e., 1% and 10% H2O injections) where SiO2
tion resistance may be partly due to the improved was stable, silicon improved carburization
oxide scale. resistance.
Silicon has also been found to be very effec- The focus thus far has been primarily on alloys
tive in improving carburization resistance. The used in ethylene cracking and steam hydrocarbon
beneficial effect of silicon on carburization reforming operations. Most of the alloys are
resistance has been reported by Norton and cast alloys used for furnace tubes. A variety of
Barnes (Ref 35), Steinkusch (Ref 45), Wolfe wrought alloys of stainless steels, Fe-Ni-Cr
(Ref 46), and Van den Bruck and Schillmoller
(Ref 47), and the data are illustrated in Fig. 5.19,
and Fig. 5.30 to 5.32. Kane (Ref 48) investigated
a large number of centrifugally cast tubes of HK
(Fe-25Cr-20Ni) and HP (Fe-25Cr-35Ni) alloys
from four producers. The tubes contained silicon
varying from about 1% to more than 2%. Tests
were conducted at 980 and 1090 °C (1800 and
2000 °F). A unit carbon activity (ac = 1.0) was
maintained for all the test environments. Oxygen
potentials of the test environments were varied
by injecting different levels of H2O. With no
H2O injection, the environment’s pO2 was below

Fig. 5.18 Oxygen potentials of the test environments used by

Norton and his colleagues in carburization studies
at Petten Laboratories. Source: Ref 37

Fig. 5.17 Carbon concentration profiles of several cen- Fig. 5.19 Carburization rate constants of several Fe-Ni-Cr
trifugally cast alloys in (a) the as-cast surface alloys at 825 °C (1520 °F) in the test environment
condition and (b) the machined surface condition after 1 year of with a carbon activity of 0.8 and an oxygen potential such that
field testing in an ethylene cracking furnace. Source: Ref 34 SiO2 is stable (but not Cr2O3), as shown in Fig. 5.18. Source: Ref 35
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:22PM Plate # 0 pg 110

110 / High-Temperature Corrosion and Materials Applications

Fig. 5.20 Carburization rate constants as a function of silicon content in the alloy for several Fe-Ni-Cr alloys tested at 825 °C (1520 °F)
in the test environment with a carbon activity of 0.8 and an oxygen potential such that SiO2 is stable (but not Cr2O3), as
shown in Fig. 5.18. Source: Ref 35

Fig. 5.22 Carburization rate constants as a function of Ni to

Fig. 5.21 Carburization rate constants of several Fe-Ni-Cr Cr + Fe ratio [Ni/(Cr + Fe)] for several Fe-Ni-Cr
alloys at 1000 °C (1830 °F) in the test environment alloys tested at 1000 °C (1830 °F) in the test environment with a
with a carbon activity of 0.8 and an oxygen potential such that carbon activity of 0.8 and an oxygen potential such that SiO2 is not
SiO2 is not stable as shown in Fig. 5.18. Source: Ref 35 stable as shown in Fig. 5.18. Source: Ref 35

alloys, and Ni-Cr alloys have been widely used in noted for its beneficial effect, as illustrated by
various industries, including heat treating and Type 330 (0.47% Si) versus Type 330 (1.0% Si)
chemical processing. Mason et al. (Ref 49) and Type 304 (0.39% Si) versus Type 302B
investigated various stainless steels by perform- (2.54% Si). Chromium was found to be bene-
ing pack carburization tests. Their results are ficial in Fe-Cr alloys, as shown by Type 446
summarized in Table 5.2. Silicon was again (27% Cr) versus 430 (16% Cr). Small additions
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:22PM Plate # 0 pg 111

Chapter 5: Carburization and Metal Dusting / 111

Fig. 5.23 Effect of nickel content on the diffusion coefficient of carbon in Fe-15Cr-Ni alloys. Source: Ref 38

of titanium or niobium appeared to be beneficial activity was maintained at unity, alloys contain-
when comparing Type 321 and Type 347 to ing aluminum, such as alloys 601 and 617,
Type 304. were much more resistant to carburization than
Several nickel-base alloys along with HK-40 HK-40. The authors (Ref 50) attributed this to the
were investigated by Kane and Hosier (Ref 50). formation of an Al2O3 scale, although alloys
Tests were conducted in environments with unit 601 and 617 contain only about 1.3% Al. Other
carbon activity and various oxygen potentials. investigators (Ref 24, 51) also showed beneficial
Different rankings were obtained at different effect of silicon in 25Cr-30/35Ni type alloys.
oxygen potentials. Test results for two environ- Nishiyama et al. (Ref 24) performed their
ments are summarized in Table 5.3. In the study of carburization in a simulated ethylene
test environment of H2-12%CH4-10%H2O with pyrolysis environment (H2-15%CH4-3%CO2)
1.3 ×10−20 atm of pO2 , where SiO2 was thermo- for Fe-Ni-Cr alloys with high Si (1.7%) and low
dynamically stable, HK-40 (1.19% Si) perform- Si (0.3–0.5%). When tested at 1000 °C, where
ed the best. When pO2 was reduced to 1.9 × 10−24 Cr2O3 was stable (see Fig. 5.9), chromium,
atm in H2-0.1%C7H13OH, where SiO2 was not silicon, was more effective in improving
not thermodynamically stable, and the carbon carburization resistance as shown in Fig. 5.33
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:22PM Plate # 0 pg 112

112 / High-Temperature Corrosion and Materials Applications

(Ref 24). Alloy B (26Cr-35Ni-0.5Si) was sig- of alloy 214 compared with those of alloys X,
nificantly better than alloy D (21Cr-31Ni-0.3Si), 601, and 150 after exposure to the test environ-
while alloy B (26Cr-35Ni-0.5Si) was similar to ment at 980 °C (1800 °F) for 55 h (Ref 52). The
alloy A (25Cr-37Ni-1.8Si). At 1150 °C, where
Cr2O3 oxide was not stable, high-Si alloys × alloy 800H
×
(alloy A: 25Cr-37Ni-1.8Si, alloy C: 32Cr-43Ni- 150 AC66
1.7Si) exhibited significantly better carburization ×

Mass change, g/m2

resistant than low-Si alloys (alloy B: 26Cr-35Ni- 100 alloy 602CA
0.5Si, alloy D: 21Cr-31Ni-0.3Si), as shown in × alloy 617
Fig. 5.34. ∗ alloy 601
50 ∗ 45TM
Aluminum is the most effective alloying ∗
element in improving an alloy’s carburization ∗
∗
resistance at high temperatures. Lai (Ref 52) 0
0 200 400 600 800 1000 1200
showed that when tested at 870, 930, and 980 °C (a)
Time, h
(1600, 1700, and 1800 °F) alloy 214 (Ni-16Cr-
3Fe-4.5Al-Y) was the most carburization- 350
resistant alloy among more than 20 commercial Mass change, g/m2 × AC66
300 alloy 800H
×
wrought alloys, ranging from stainless steels and 250 × • • alloy DS
ו • •
Fe-Ni-Cr alloys to nickel- and cobalt-base 200
× ×
alloy 600H
× ∗ ∗ alloy 601
superalloys in the test environments, which were 150 ∗ 45TM
∗ ∗
characterized by a unit carbon activity and oxy- 100 alloy 602CA
50 ו
gen potentials such that Cr2O3 was not expected alloy 617
to form on the metal surface. Oxides of silicon, 0
0 200 400 600 800 1000 1200 1400
titanium, and aluminum were expected to be (b)
Time, h
stable under the test conditions. The excellent
carburization resistance of this alloy was attrib-
× AC66
uted to the Al2O3 oxide scale formed on the metal 320 ×
×
Mass change, g/m2

surface. Figure 5.35 illustrates the microstructure ∗ 45TM

240 × ∗
× ∗
∗
• alloy DS
160 × ∗ •
∗ • •
80 × alloy 601
• alloy 617
∗• • alloy 602CA
alloy 600H
0
0 200 400 600 800 1000 1200
(c)
Time, h

Fig. 5.25 Results of carburization tests in H2-CH4 mixtures

(ac = 0.8) at (a) 850 °C, (b) 1000 °C, and (c) 1100 °C
for Fe-Ni-Cr alloys (800H, AC66, and DS) and nickel-base alloys
(alloys 600H, 601, 602CA, 617, and 45TM). Source: Ref 29
of exposure at 1000 °C g/cm2
Carbon pick-up after 1000 h

400
alloy 800H
AC66
• •
300
alloy 600H 45TM • •
200 • • • • HPM
alloy DS
alloy 625 alloy 601
100 • alloy 602CA
•alloy 617
0
0 10 20 30 40 50
Fe, %

Fig. 5.26 Weight gain as a function of iron content (i.e., a

Fig. 5.24 Carbon solubility as a function of nickel content at function of nickel) in Fe-Ni-Cr and Ni-Cr alloys
different carbon activities in Fe-Ni alloy system at tested at 1000 °C for 1008 h in a H2-CH4 mixture (ac = 0.8 and
725:25
1000 °C (1830 °F). Source: Ref 39 pO2 =10 ). Source: Ref 29
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:22PM Plate # 0 pg 113

Chapter 5: Carburization and Metal Dusting / 113

Fig. 5.27 Effect of nickel content on the carburization resistance of Fe-Ni-Cr alloys. Source: Ref 42

Fig. 5.29 Carburization resistance of several wrought and

cast Fe-Cr-Ni alloys (Type 304, 321, HT, HU, and
HK) after testing in dry ethane (C2H6) for 24 h at temperatures from
Fig. 5.28 Effect of chromium on the carburization resistance 880 to 1000 °C. Source: Ref 43
of Fe-Ni-Cr alloys after testing in H2-2.5%CH4 at
1050 °C (1920 °F) for 100 h. Source: Ref 42

H2-2%CH4 at 982 °C (1800 °F) for 96 h. The

existence of this oxide scale was confirmed by carbon activity for the test environment at the test
Auger analysis (Ref 52). Similar test results were temperature was greater than 1.0 (see Fig. 5.11).
also observed by Lai et al. (Ref 53) when tested in Their results are summarized in Fig. 5.36. Alloy
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:22PM Plate # 0 pg 114

114 / High-Temperature Corrosion and Materials Applications

214 showed no evidence of carburization, Table 5.2 Results of pack carburization tests
while Type 310 and alloys 800H, 25-35NbMA, at 980 °C (1800 °F)(a) for various stainless steels
25-35Nb, 803, and 602CA showed different Nominal Si content, Increase in C
Alloy composition % content(b), %
degrees of carburization attack. Alloy 602CA
800 Fe-21Cr-34Ni 0.34 0.04
with about 2% Al (Ni-25Cr-10Fe-2Al-Y-Zr), 330 Fe-15Cr-35Ni 0.47 0.23
although not as good as alloy 214, was signifi- 330 Fe-15Cr-35Ni-Si 1.00 0.08
cantly better than other alloys. A similar obser- 310 Fe-25Cr-20Ni 0.38 0.02
314 Fe-25Cr-20Ni-Si 2.25 0.03
vation was also made by Kane et al. (Ref 54), 309 Fe-25Cr-12Ni 0.25 0.12
showing that the alumina-forming MA956 347 Fe-18Cr-8Ni-Nb 0.74 0.57
321 Fe-18Cr-8Ni-Ti 0.49 0.59
304 Fe-18Cr-8Ni 0.39 1.40
302B Fe-18Cr-8Ni-Si 2.54 0.22
Table 5.1 Weight gain (mg/cm2) for several 446 Fe-28Cr 0.34 0.07
cast alloys after 100 h at 1090 °C (2000 °F) in 430 Fe-16Cr 0.36 1.03
H2-CH4-H2O mixtures (a) 40 cycles of 25 h each cycle at 980 °C (1800 °F). Carburizer was renewed after
each cycle. (b) Bulk analysis. Source: Ref 49
Weight gain, mg/cm2
Alloy(a) H2-8.6CH4-7H2O H2-12CH4-10H2O
HK-40 (1% Si) 25.0 21.8
HK-40 (2% Si) 16.8 10.2
TMA-4750 2.0 1.0
HP-45 19.0 4.3
TMA-6350 3.8 2.3
(a) HK-40 (1% Si): 0.43C-0.60Mn-0.96Si-25.4Cr-20.7Ni. HK-40 (2% Si): 0.41C-
0.60Mn-1.98Si-25.0Cr-20.7Ni. TMA-4750: 0.44C-0.69Mn-1.99Si-24.9Cr-
20.8Ni-0.11Ti-0.29Nb-0.30W-REM. HP-45: 0.51C-0.54Mn-1.65Si-25.5Cr-
36.1Ni. TMA-6350: 0.50C-0.70Mn-1.84Si-25.1Cr-38.4Ni-0.13Ti-0.28Nb-0.27W-
REM. REM denotes rare earth metals. Source: Ref 44

Fig. 5.31 Effect of silicon on the carburization resistance of

cast Fe-20Ni-Cr alloys tested at 1090 °C (2000 °F)
for 24 h in wet ethane (C2H6). Source: Ref 46

Fig. 5.32 Effect of silicon on the carburization resistance

Fig. 5.30 Effect of silicon on the carburization resistance of HK-40 with different silicon levels tested at
of 25Cr-20Ni and 35Cr-25Ni-Nb alloys. Source: 1100 °C (2010 °F) for 520 h in carbon granulate (pack carbur-
Ref 45 ization test). Source: Ref 47
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:22PM Plate # 0 pg 115

Chapter 5: Carburization and Metal Dusting / 115

alloy (Fe-20Cr-4.5Al-0.5Y2O3) performed sig- to allow decoking, typically with steam and air to
nificantly better than alloys 601, 800H, Type remove the coked layer. This decoking operation
310, and HK alloy (Table 5.4). can have detrimental effects on the tube proper-
Aluminum, which has been used for alloy ties, thus reducing the tube’s service life.
addition to form external Al2O3 scale in iron- Other factors such as surface finish have been
and nickel-base wrought alloys, has been used found to be very important in affecting carbur-
in centrifugally cast alloys for oxidation or ization reactions. Machining the metal surface to
carburization resistance. Recently a commercial, improve the surface finish can significantly
centrifugally cast nickel-base alumina-forming increase an alloy’s carburization resistance. It is
alloy, alloy 60HT, containing approximately common practice to bore or hone the internal
25% Cr, 11% Fe, 0.4% C, and Al, was developed diameter of a centrifugally cast tube to remove
(Ref 55). In the paper published by Kirchheiner surface shrinkage pores. Figures 5.17, 5.37, and
et al. (Ref 55), no carburization data were 5.38 illustrate the significant improvement in
reported. However, the resistance to coking was carburization resistance as a result of surface
studied on alloy 60HT containing three levels of machining (Ref 34, 56). A cast metal surface with
aluminum (i.e., 2.35, 3.55, and 4.81%). They shrinkage pores can generate stagnant conditions
found significant reduction in coking rates for the in crevices, which are very conducive to car-
samples containing 3.55 and 4.81% Al, with the burization attack. In addition, a machined surface
sample containing 2.35% Al exhibiting only exhibits a cold-worked layer, which tends to
slight reduction of coking rates compared with accelerate the diffusion process and results in
the conventional HP-40 alloy (4852). Coking, rapid formation of oxide scale or film, thus
which is an important phenomenon in ethylene slowing subsequent carbon ingress.
cracking, develops on the internal surface of the Compared to the machined or ground surface,
pyrolysis tube and reduces the heat transfer. The the electropolished surface exhibited accelerated
ethylene cracking operation has to be interrupted carburization (Fig. 5.39). Norton and Barnes
(Ref 35) considered the electropolished surface a
Table 5.3 Weight gain (mg/cm2) for several work-free surface that failed to develop a surface
Fe-Ni-Cr and Ni-base alloys after ten 24 h cycles oxide film as readily as the cold-worked ground
at 1100 °C (2010 °F) in H2-12%CH4-10%H2O surface. Not mentioned by Norton and Barnes
and H2-0.1%C7H13OH environments in their paper (Ref 35), electropolishing can
Weight gain, mg/cm2 produce a surface layer that may be depleted
Alloy H2-12CH4-10H2O(a) H2-0.1C7H13OH(b) in chromium, and the surface depletion of
HK-40 (1.19% Si) 6.97 54.38 chromium may result in the observed accelerated
800H 23.46 46.60 carburization. In their study of oxidation/
600 12.00 4.85
617 16.84 0.50 carburization of a high-temperature gas-cooled
601 22.74 1.81 reactor (HTGR) helium environment containing
690 25.54 54.38 parts per million (ppm) levels of H2, CO, CO2,
(a) Inlet gas mixture: ac = 1.0, pO2 ¼ 1:3 · 10720 atm at 1100 °C. (b) Inlet gas CH4, and H2O, Mazandarany and Lai (Ref 57)
mixture: ac = 1.0, pO2 =1:9 · 10724 atm at 1100 °C. Source: Ref 50
observed that Type 316SS specimen (obtained

Fig. 5.33 Carbon profiles of high-Si (HSi) and low-Si (LSi) Fig. 5.34 Carbon profiles of high-Si (HSi) and low-Si (LSi)
Fe-Ni-Cr alloys after testing at 1000 °C (1832 °F) Fe-Ni-Cr alloys after testing at 1150 °C (2100 °F)
for 96 h in H2-15%CH4-3%CO. Source: Ref 24 for 48 h in H2-15%CH4-3%CO. Source: Ref 24
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:22PM Plate # 0 pg 116

116 / High-Temperature Corrosion and Materials Applications

from sheet material) with the as-received surface observation for the 316 specimen with the as-
condition suffered severe carburization after received surface condition after testing at 649 °C
testing at 649 °C (1200 °F) for 5000 h in (1200 °F) in He-1500 µatm H2-450 µatm CO-50
He-1500 µatm H2-450 µatm CO-50 µatm CH4-50 µatm CH4-50 µatm H2O was the formation of an
µatm H2O. The 316 specimen with the as-ground Fe-Ni metallic scale on the specimen surface and
specimen, on the other hand, showed no evi- internal oxides underneath the metallic scale.
dence of carburization tested in the same This is shown in Fig. 5.41 (Ref 57). The 316SS
retort under the same test condition. This obser- specimen with the as-ground surface condition
vation was supported by both microstructure after testing under the same condition showed
and microhardness profile results, as shown in surface oxide scales with no outer metallic scale
Fig. 5.40 (Ref 57). The oxygen potentials of and internal oxides. The authors thus believed
the test environment were reported to be high that the initial surface of the 316SS sheet was
enough to form chromium oxides and too low depleted in chromium at a significant degree,
to form iron and nickel oxides. The unusual such that not enough chromium was available

Fig. 5.35 Optical micrographs showing the microstructures of (a) alumina-former alloy 214 (Ni-16Cr-3Fe-4.5Al-Y), and several
chromia-former alloys (b) 601 (Ni-23Cr-14Fe-1.4Al), (c) X (Ni-22Cr-18Fe-9Mo), and (d) 150 (Co-27Cr-18Fe) after testing at
980 °C (1800 °F) for 55 h in Ar-5%H2-5%CO-5%CH4 (ac = 1.0,pO2 =9 · 10722 atm). Source: Ref 52
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:23PM Plate # 0 pg 117

Chapter 5: Carburization and Metal Dusting / 117

Fig. 5.36 Weight gain as a function of cycles (24 h cycle) tested at 982 °C (1800 °F) for 96 h with 24 h cycles in H2-2%CH4 for Type
310SS, 800H (N08810), 25-35Nb (25Cr-35Ni-1.3Nb-0.4C-2.0max Si), 25-35NbMA (microalloyed, 25Cr-35Ni-1.1Nb-
0.4C-1.4Si, Ti, REM), 803 (S35045: Fe-25Cr-35Ni), 602CA (N06025: Ni-25Cr-10Fe-2.1Al-Y-Zr), and 214 (N07214: Ni-16Cr-3Fe-4.5Al-
Y). Note: All four alloy 800H (N08810) specimens were control samples. Source: Ref 53

Table 5.4 Carburization resistance of Fe-Ni-Cr,

Ni-base, and ODS alloys in H2-2%CH4 at
1000 °C (1830 °F) for 100 h
Alloy Weight gain, mg/cm2
MA956 <0.3
601 10
800 19
310SS 36
HK (1.07% Si) 34
HK (2.54% Si) 29
Source: Ref 54

metal on the metal surface. When the surface

of the test specimen from the same sheet
material was ground in specimen preparation, the
chromium-depleted surface layer was removed.
When tested under the same condition, this
ground specimen formed an external chromium
oxide scale on the metal surface without the
Fig. 5.37 Comparative carburization resistance of as-cast and formation of an Fe-Ni metallic surface layer. The
machined tubes of alloys 30Cr-30Ni, 36X (Fe-0.4C-
25Cr-34Ni-1.2Nb), and 36XS (Fe-0.4C-25Cr-34Ni-1.5Nb-1.5W)
Fe-Ni metallic layer formed on the as-received
after 3 years of field testing in an ethylene pyrolysis furnace. surface of the 316SS specimen provided rapid
Source: Ref 34 absorption of carbon and a rapid diffusion path
of carbon in the metal, resulting in significant
to form a continuous external chromium oxide carburization. When tested at 870 °C (1600 °F)
scale. Then, the nucleation of internal chromium- in the same environment, an external chromium
rich oxides resulted in a chromium-depleted oxide scale was observed to form on the
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:23PM Plate # 0 pg 118

118 / High-Temperature Corrosion and Materials Applications

as-received surface with no evidence of carbur- Figure 5.45 shows the carburization of alloy
ization attack. No explanation was offered by 800H in H2-CH4 (ac=1.0) with injection of dif-
the authors on this. It is believed that higher ferent levels of H2S (H2S/H2) (Ref 59). When
temperatures caused a rapid homogenization of this approach is used for controlling carburiza-
the chromium concentration at the initially tion, it is important to determine the optimum
depleted surface layer due to faster diffusion rate. level of sulfur compounds needed to achieve
Since this type of environment contains only maximum effectiveness. Too much can lead to
ppm levels of corrosive gaseous components, the accelerated corrosion due to sulfidation, which
reaction kinetics are believed to be relatively may be worse than carburization. Figure 5.46
slow. This slow reaction kinetics allowed the shows that a low H2S injection failed to reduce
homogenization of the chromium concentration carburization and a high H2S injection caused
at the metal surface when tested at 870 °C sulfidation attack for alloy 800 (Ref 59). It
(1600 °F). has been proposed by Grabke et al. (Ref 60)
Injecting sulfur compounds into the carbur- and Ramanarayanan and Srolovitz (Ref 61) that
ization gas stream can significantly reduce the sulfur absorbed on the metal surface blocks
carburization kinetics. Sulfur compounds, such the potential sites for carbon absorption from the
as H2S, can retard carburization kinetics. This is carburizing gas, thus significantly reducing the
illustrated by the test results shown in Fig. 5.42, carbon concentration in the metal surface layer
obtained by Ramanarayanan et al. (Ref 58). and retarding the overall carbon transfer into the
Beneficial effect of sulfur in reducing carbur- metal.
ization attack was demonstrated by Norton
and Barnes (Ref 35) in their laboratory tests.
5.2.3 Effect of Carburization on Mechanical
Figure 5.43 shows that carburization of HK-40
alloy was significantly reduced when 100 ppm Properties
H2S was injected into the test environment. The Carburization results in formation of internal
effect of different levels of H2S injection on the carbides in the alloy. The room-temperature
carburization of HK-40 is illustrated in Fig. 5.44. ductility or toughness can be severely degraded

Fig. 5.39 Detrimental effect of the electropolished surface

condition on the carburization resistance of HK-40
Fig. 5.38 Effect of machining on the carburization resistance alloy tested at 825 °C (1520 °F) in an environment with 0.8 ac and
of cast HK-40 alloy. Source: Ref 56 the oxygen potential shown in Fig. 5.18. Source: Ref 35
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:23PM Plate # 0 pg 119

Chapter 5: Carburization and Metal Dusting / 119

Oxalic etch

Exposed at 649 °C (1200 °F) for 5000 h, as-received surface

(649 °C)1200 °F
As-received surface
500

400
Microhardness, DPH

300
649 °C(1200 °F)
and Oxalic etch
760 °C(1400 °F) Exposed at 649 °C (1200 °F) for 5000 h, ground surface
Ground surface
200

871 °C(1600 °F)

100

0
0 100 200 300 400 500 600 700
Distance from exposed surface, µm

Fig. 5.40 Microhardness profile and optical micrograph showing severe carburization attack on the 316 specimen with the original
as-received surface (solid circle data point) after testing at 649 °C (1200 °F) for 5000 h in He-1500 µatm H2-450 µatm
CO-50 µatm CH4-50 µatm H2O. Also shown are microhardness profile and optical micrograph of the 316 specimen with the as-ground
surface condition (solid square data point) after testing under the same condition showing no sign of carburization attack. The specimens
with the ground surface condition showed no carburization attack after testing at a higher temperature (i.e., 760 °C) in the same
environment for the same test duration. The specimen with the as-received surface condition showed no evidence of carburization attack
when tested at 871 °C (1600 °F). Source: Ref 57
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:23PM Plate # 0 pg 120

120 / High-Temperature Corrosion and Materials Applications

when the alloy is heavily carburized (Ref 29). For (Ref 62) performed a metallurgical analysis of a
high-temperature components that are subjected failed ethylene pyrolysis furnace tube made of
to severe carburizing environments at very high HP40Nb and found the tube was carburized
temperatures, such as ethylene pyrolysis furnace
tubes, the component can be carburized through 4
the thickness of the component with significant
amount of carbon pickup. Grabke and Jakobi 3.5 T = 1000 °C

3
H2S Off
2.5
H2S On
2

1.5 Carburization

1
Cr2O3 Growth Cr2O3 to Cr7C3 Conversion
0.5

0
0 20 40 60 80 100 120

Fig. 5.42 Effect of sulfur on the carburization kinetics of Ni-

30Cr alloy. Vertical axis is mass gain per unit area,
and horizontal axis is exposure time in hour. The test began with
oxidation at 1000 °C (1832 °F) in a CO-CO2 environment for
about 40 h, and then switched to carburization in a H2-CH4
mixture with a carbon activity of 0.9, thus prompting the con-
Fig. 5.41 As-received surface 316 specimen after testing at
version of oxide scale to carbides and a rapid carburization. This
649 °C (1200 °F) for 2000 h in He-1500 µatm H2-
was then followed by injecting 100 ppm of H2S into the carbur-
450 µatm CO-50 µatm CH4-50 µatm H2O. Area 1: Fe-Ni metallic
izing environment, thus immediately stopping carburization
phase, Area 2: Cr-Mn-Si oxides, and Area 3: Fe-Ni metallic phase.
reaction until the injection of H2S was turned off, which set off a
Unetched condition. Source: Ref 57
rapid carburization again. Source: Ref 58

Fig. 5.43 Carbon concentration profiles for HK-40 alloy after

testing at 1000 °C (1830 °F) for 100 h in a carbur-
izing environment (ac = 0.8, pO2 is shown in Fig. 5.18) with and
without injection of 100 ppm H2S. The specimen was a flat
coupon with both sides being exposed to the carburizing gas. Fig. 5.44 Effect of H2S on the carburization behavior of
Source: Ref 35 HK-40 alloy. Source: Ref 35
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:23PM Plate # 0 pg 121

Chapter 5: Carburization and Metal Dusting / 121

throughout the tube wall with 3.14 to 3.3% C, as the volume of the carburized zone, thus devel-
opposed to the original carbon content of about oping internal stresses. When the alloy is
0.5%. Furthermore, carburization can increase severely carburized, these internal stresses can be

Fig. 5.45 Weight gain of alloy 800 after 100 h in H2-CH4-H2S environment (ac = 1.0) at 900, 1000, and 1100 °C (1652, 1832, and
2012 °F) for different H2S/H2 ratios. Source: Ref 59

Fig. 5.46 Effect of H2S on the carburization behavior of alloy 800 in H2-CH4-H2S environments (ac = 1.0) at 1000 °C (1832 °F) for
different H2S/H2 ratios. Also shown are corresponding microstructures of the tested specimens. Source: Ref 59
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:23PM Plate # 0 pg 122

122 / High-Temperature Corrosion and Materials Applications

significant enough to develop cracking when the as the result of either testing in a carburization
carburized alloy is cooled to room temperature. environment or precarburization can signifi-
This is illustrated in Fig. 5.47 (Ref 25). cantly reduce the 1% creep strengths of both
When the alloy is in the high-temperature HK-40 and HK-30. The authors attributed this
creep range, a thoroughly carburized alloy can to both the internal stresses developed due to
retain reasonable ductility at elevated tempera-
tures. In fact, a thoroughly carburized Fe-Ni-Cr
alloy can exhibit a better creep ductility than the
same alloy tested in air. Guttmann and Schonherr
(Ref 63) studied the creep-rupture behavior of
HK alloys in a heavily precarburized condition
and found that the creep ductility was sig-
nificantly improved by carburization when creep
tested at 1000 °C (1832 °F). In their study, they
precarburized both HK-40 and HK-30 specimens
at 1000 °C (1832 °F) in H2-CH4 to a complete
saturation of M7C3 carbides (about 55 vol%) in
the material with about 4% C. The creep-rupture
testing was conducted at 1000 °C (1832 °F) in
H2-1%CH4 (ac = 0.8) to avoid decarburization
during testing. The rupture ductility of pre-
carburized specimens was compared with that of
HK-40 and HK-30 in the as-cast condition
tested in air at the same temperature, as shown
in Fig. 5.48. The authors also made a similar
observation in comparing the rupture ductility
when tested in H2-1%CH4 (ac = 0.8) at 1000 °C
(1832 °F) (i.e., specimens were not precar-
Fig. 5.48 Elongation to fracture as a function of rupture life
burized prior to testing) with that tested in air, as of HK-40 and HK-30 comparing the as-cast
shown in Fig. 5.49. Carburization-accelerated specimens tested in air and the precarburized (thoroughly car-
burized) specimens tested in H2-1%CH4 (ac = 0.8) (to avoid
creep deformation at high temperatures is clearly decarburization) after creep-rupture testing at 1000 °C (1832 °F).
illustrated by the two comparison creep curves Source: Ref 63
tested at 1000 °C (1832 °F) in H2-1%CH4 (ac =
0.8) and air reported by Guttmann and Schonherr
(Ref 63), as shown in Fig. 5.50. Because car-
burization accelerates creep deformation, creep
strength is significantly reduced. Carburization

Fig. 5.47 Cracking developed in alloy 800 due to internal Fig. 5.49 Elongation to fracture as a function of rupture life of
stresses resulting from heavy carburization in the HK-40 and HK-30 comparing the data from tests in
test coupon with no external loading during the carburization test. the carburizing environment (i.e., H2-1%CH4 [ac = 0.8]) and that
Source: Ref 25 from air tests at 1000 °C (1832 °F). Source: Ref 63
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:23PM Plate # 0 pg 123

Chapter 5: Carburization and Metal Dusting / 123

volume increase in carburization and carbide air. Precarburization was carried out at 1000 °C
coarsening during creep testing at high tem- (1832 °F) for 1200 h in a H2-CH4 mixture with
peratures. This is illustrated in Fig. 5.51. ac being 0.8, which resulted in a fully carburized
Ethylene pyrolysis furnace tubes that are condition containing only M7C3 (about 30 vol
subjected to severe carburization during service %). Partially carburized specimens were carbur-
at high temperatures typically experience sig- ized at 1000 °C (1832 °F) for 100 h in a H2-CH4
nificant creep deformation as a result of carbur- mixture with ac being 0.8, which produced a
ization. Accordingly, the tube suffers stretching carburized depth of about 1 mm (equivalent to
due to creep elongation, and frequent shortening 56% of cross-sectional area being carburized).
of tube coils is required in order to avoid the Pre-aging treatment consisted of heating the
tube from reaching to the furnace floor to cause specimen at 1000 °C (1832 °F) for 100 h in air to
tube damages (Ref 64). Jakobi and Gommans produce a comparable structure to the uncarbur-
(Ref 64) also indicated brittle fracture was one of ized core of the partially carburized specimens.
the failure modes for centrifugally cast pyrolysis Their creep-rupture data are summarized in
furnace tubes. The brittle fracture was primarily Fig. 5.52. Fully carburized specimens were
due to the furnace trip that caused a large found to exhibit significantly higher creep-
temperature drop of 500 to 1000 °C (932 to rupture strengths than the as-received specimens
1832 °F). Such a furnace trip can produce a strain tested in air. Partially carburized or pre-aged
of 0.75 to 1.5%, which is equivalent to about the specimens showed strengths comparable to those
rupture ductility of aged, carburized, and nitrided of the as-received specimens. As for rupture
cast furnace tubes at room temperature to about ductility, the fully carburized specimens showed
600 °C (1112 °F) (Ref 64). Thus, the furnace lower rupture ductility for shorter rupture times
tubes suffer brittle fracture at low temperatures (at high stresses), but comparable to those of
due to the large temperature drop. as-received specimens and pre-aged as well as
Taylor et al. (Ref 65) studied the effect of partially carburized specimens for longer rupture
carburization on the creep-rupture behavior of times (i.e., at lower stresses). This is illustrated in
alloy 800H at 800 °C (1472 °F) and found that Fig. 5.53. The authors also investigated the effect
carburization increased rupture strength and of the temperature on creep-rupture ductility of
life. They compared the creep-rupture data of alloy 800H that was fully carburized. The alloy’s
precarburized specimens (partially carburized rupture ductility (when fully carburized) was
specimens and fully carburized specimens) with found to decrease significantly with decreasing
that of as-received specimens as well as pre-aged temperature. The carburized specimen failed in
specimens. The data for precarburized specimens a completely brittle manner on loading when
were tested in a H2-CH4 environment (ac > 0.5) tested at 600 °C (1112 °F). This is illustrated in
to avoid decarburization during testing, while as- Fig. 5.54.
received and pre-aged specimens were tested in

Fig. 5.50 Comparative creep curves of HK-40 tested at Fig. 5.51 1% creep strengths of HK-40 and HK-30 tested at
1000 °C (1832 °F) and 15 MPa in air and H2-1% 1000 °C (1832 °F) in air, H2-1%CH4 (ac = 0.8), and
CH4 (ac = 0.8). Source: Ref 63 for precarburized specimens tested in H2-1%CH4. Source: Ref 63
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:23PM Plate # 0 pg 124

124 / High-Temperature Corrosion and Materials Applications

5.3 Metal Dusting reactions (Eq 5.1–5.3) that can cause carburiza-
tion can also cause metal dusting attack. Metal
5.3.1 Metal Dusting—Thermodynamic dusting most often occurs in syngas environ-
Considerations ments, which consist of primarily H2 and CO.
Metal dusting occurs in carburizing atmo- In the earlier discussion of Eq 5.1: H2 + CO =
spheres at intermediate temperatures (approxi- C + H2O, where graphical presentation of the
mately between 430 and 900 °C (800 and chemical reaction is presented as a function of
1650 °F) with the maximum rate of attack carbon activity (ac), temperature, and the gas
occurring at about 600 to 700 °C (1112 to composition in terms of (pCO  pH2 = pH2 O ), as
1292 °F), depending on alloys and environ- shown in Fig. 5.1. Figure 5.1 illustrates that the
ments. It is now well understood that metal carbon activity of the environment decreases
dusting occurs in environments that exhibit high- with increasing temperature. Thus, at high tem-
carbon activities (i.e., ac > 1) (Ref 66–75). peratures, many gaseous compositions are not
As discussed in section 5.2.1, three chemical thermodynamically favored for causing metal
dusting. Conversely, lower temperatures with
increasing carbon activity are thermodynamic-
ally favored for causing metal dusting. This is
why metal dusting attack diminishes as the
temperature increases after passing the peak
reaction temperature due to decreasing carbon
activity. For example, for the gas mixture with
(pCO  pH2 =pH2 O ) ratio being 1.0, its carbon
activity varies from more than 102 to approxi-
mately 10−1 as the temperature increases from
400 to 800 °C (752 to 1472 °F). Metal dusting
attack diminishes as the temperature decreases
after dropping below the peak temperature. This
is due to the reaction’s kinetics that decreases
with decreasing temperature.
The reaction Eq 5.2, as graphically presented
in Fig. 5.2, shows a similar fashion. When the
Fig. 5.52 Stress rupture data of precarburized (fully carbur-
environment consists of CO and CO2, metal
ized and partially carburized) specimens tested at
800 °C (1472 °F) in a carburizing environment (to prevent dec- dusting attack follows that involves H2 and CO.
arburization) is compared with that of as-received specimens and
pre-aged specimens tested in air at the same temperature. Source: However, when the environment consists of H2,
Ref 65 CO, and CO2, the Boudouard reaction (CO-CO2

Fig. 5.53 Comparison rupture ductility of alloy 800H tested

at 800 °C (1472 °F) for fully and partially carbur- Fig. 5.54 Effect of temperature on rupture ductility of fully
ized specimens in comparison with as-received and pre-aged carburized alloy 800H tested at different tem-
specimens. Source: Ref 65 peratures with stresses to cause rupture in 200 h. Source: Ref 65
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:23PM Plate # 0 pg 125

Chapter 5: Carburization and Metal Dusting / 125

reaction) is believed to be slower in reaction 5.3.2 Mechanisms of Metal Dusting Attack

kinetics than that of Eq 5.1, thus making Eq 5.1 to
Significant advancement in understanding the
be the dominant reaction for the metal dusting
mechanisms of metal dusting has been achieved
(Ref 76, 77). The reaction 5.3 (Eq 5.3), on the
in the past decade, with most of the contribution
other hand, increases the carbon activity with
being made by Professor Grabke and his group
increasing temperature. The graphic presentation
(Ref 68–75). Two models have been proposed
of this reaction is shown in Fig. 5.3. The carbon
with one model explaining the metal dusting
activity is generally too low for many environ-
behavior of iron and low-alloy steels and
ments containing CH4 where metal dusting the other model explaining the behavior of Cr-
occurs at 600 or 700 °C (1112 or 1292 °F). The containing high-temperature alloys (i.e., Fe-Cr,
presence of CH4 in the gas mixture containing Fe-Ni-Cr, and Ni-Cr alloys). These two models,
CO and CO2 may help bring up the carbon which are summarized by Grabke (Ref 73), are
activity at higher temperatures, thus increasing briefly described here. The model for iron and
the upper temperature limit at which metal low-alloy steels is summarized in Fig. 5.55.
dusting can take place. Based on the model, the prerequisite for metal
dusting to occur is a high carbon activity en-
vironment (ac > 1) that causes the oversaturated
ac (gas)
metal with the carbon activity being more
than 1. This condition promotes the formation of
CO + H2 cementite (Fe3C) on the metal surface. As shown
H2O + C (diss.) in Fig. 5.1, the calculated carbon activities in
equilibrium with cementite are higher than 1.0
ac″ (Fe3C1Fe)
(a) ac=1 at 700 °C (1292 °F) and lower temperatures.
Figure 5.56 shows a metastable Fe-C-O phase
Fe3C C (diss.) + Fe diagram at 600 °C indicating that metastable
cementite forms at ac > 1 (Ref 78). Once graphite
CO + H2 deposits on the cementite with the carbon activity
H2O + C (in Fe3C) of the graphite at 1, cementite in contact with
graphite becomes unstable and then decomposes
(b) into iron and carbon with iron migrating outward
into the graphite layer. As a result, metal wastage
takes place as a repeated reaction of formation
of metastable cementite and decomposition of
CO + H2 cementite into iron and carbon. For iron and
H2O + C (graphite) low-alloy steels with no protective oxide scales,
metal wastage attack in general follows uniform
(c)

CO + H2
H2O + C
(graphite)

(d) 3 Fe + C Fe3C

Fig. 5.55 Schematic showing mechanism of metal dusting for

iron and low alloy steels with the following steps:
(a) The metal is oversaturated with carbon (ac > 1) due to carbon
transfer from a high carbon activity environment (ac > 1) to the
metal, (b) thus resulting in the formation of cementite (Fe3C) at the
surface, (c) and later the formation of graphite on top of cementite
and then decreases in ac to one (ac = 1 for graphite) at the
cementite in contact with graphite, (d) thus, cementite becomes
unstable and decomposes into iron and carbon, with iron
migrating into graphite layer to form iron particles (embedded in Fig. 5.56 Fe-C-O metastable phase stability diagram showing
graphite), which act as catalysts for more carbon deposition and that metastable cementite (Fe3C) forms under high
coking (d). Source: Ref 73 carbon activities (ac > 1) at 700 °C and lower. Source: Ref 78
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:23PM Plate # 0 pg 126

126 / High-Temperature Corrosion and Materials Applications

metal thinning (Ref 73). A layer of cementite (or direct graphite formation in high nickel
(Fe3C) was found to occur on 1Cr-0.5Mo steel alloys), thus resulting in lower carbon activity
after exposure in metal dusting conditions with a (ac = 1) and decomposition of metastable car-
layer of coke on top of the cementite (Ref 71). bides into metal particles and carbon. As a result,
For chromia formers in Fe-Cr, Fe-Ni-Cr, and the attack starts locally and often leads to the
Ni-Cr alloys, a different mechanism was pro- formation of hemispherical pits (Ref 73). This
posed, which was summarized in Fig. 5.57 type of morphology is shown in Fig. 5.58.
(Ref 73). These high-chromium alloys form a Pippel et al. (Ref 79) found graphite on
protective chromium oxide scale. Metal dusting cementite that formed on Fe-5Ni alloy, when
attack is initiated when the oxide scale develops exposed to metal dusting conditions. For
local defects, allowing carbon transfer from a Fe-10Ni and Fe-25Ni to Fe-30Ni alloys, graphite
high carbon activity environment (ac > 1) to was found to grow into metal with no metastable
metal causing oversaturation of carbon, and carbide formed on the alloy (Ref 79). Graphite
subsequent formation of graphite after formation was also observed to grow into metal directly
of metastable M3C carbides in low nickel alloys in high-nickel Fe-Ni alloys, such as Fe-40Ni,
Fe-50Ni, and Fe-80Ni (Ref 72).

Cr2O3 alloy
5.3.3 Alloy Resistance to Metal Dusting
Process equipment failures due to metal dust-
Graphite
ing were reported in the refinery industry during
the 1950s. Eberle and Wylie (Ref 80) reported
metal wastage of uncooled components, such as
soot blower elements, made of Types 347SS and
(a) 310SS in the waste heat boiler of a synthesis
gas reactor. The synthesis gas, predominantly
(d) CO and H2 with some water vapor and carbon
particles, was produced by combustion of
Internal methane with oxygen. The wastage took place
C carbides 3M+C M3C at temperatures between 480 and 900 °C (900
and 1650 °F). Both 347SS and 310SS suffered
severe metal wastage after only 3 weeks of
'Coke' service.
and C Prange (Ref 8) reported that tubes containing
(b)
a chromia-alumina catalyst at a temperature of
about 590 °C (1100 °F) in a butane dehydro-
genation system, where butane was converted to
(e)
butene, suffered metal loss problems. The oxide

C Metastable
carbide

(c)

Fig. 5.57 Mechanism of metal dusting for chromia formers

with the following steps: (a) Development of local
defects in the oxide scale, allowing carbon transfer from the
environment to the metal, (b) carbon ingress into the metal
resulting in the formation of stable carbides, M23C6 and M7C3, (c)
continued carbon ingress increases the carbon activity to more
than one (ac > 1), resulting in the formation of metastable M3C 1cm
carbides in low-nickel alloys, or resulting in direct growth of
graphite in high-nickel alloys, (d) graphite deposition occurs
decreasing the carbon activity to one, thus, causing the decom-
position of metastable M3C into metal particles and carbon with Fig. 5.58 Metal dusting attack in alloy 800 in synthesis gas
metal particles catalyzing more coking (e). Source: Ref 73 environment at 550 °C (1022 °F). Source: Ref 75
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:23PM Plate # 0 pg 127

Chapter 5: Carburization and Metal Dusting / 127

dusts resulting from metal dusting contaminated (or N-155). Metal dusting typically occurred at
the catalyst and caused undesirable side effects temperatures between 540 and 820 °C (1000 and
for the operation. The alloys reported to perform 1500 °F). Severe attack frequently took place
poorly were 12Cr steel, 18Cr-11Ni, and alloy after 1 year of service and at locations where
600. The alloys that performed well included the gaseous environment became stagnant.
Fe-27Cr, 25Cr-20Ni, and Type 302B (18Cr-8Ni Favorite locations included the interface with
with 2.4% Si). Severe metal loss in the form of refractories, where small gaps or “dead” spaces
pitting was also observed by Hoyt and Caughey were created. Figure 5.59 (Ref 84) shows a
(Ref 81) for Type 310SS equipment exposed to a sample of Multimet alloy obtained from a
gas mixture rich in H2 and CO at temperatures of furnace fan housing in a carburizing furnace. The
650 to 700 °C (1200 to 1300 °F) in a plant that fan housing suffered metal dusting at the metal
converted coal to gasoline and other products.
Metal dusting problems were also encountered
in reforming plants (Ref 5), where a synthesis
gas (i.e., H2 + CO) for methanol manufacturing 1 2 3 4
was produced. Type 304SS and 310SS reformer Cm
outlet tubes were perforated by severe pitting
attack at 650 to 725 °C (1200 to 1340 °F).
Hydrodealkylation units (Ref 5), acetic acid
cracking furnaces (Ref 5), and coal gasification
plants (Ref 6) were reported to suffer metal
dusting problems. Perkins et al. (Ref 6) reported
pitting attack of alloy 800 tubes in a preheater for
gasifier recycle gas rich in H2 and CO with some
H2O. The tubes, with a wall thickness of 0.38 cm
(0.15 in.), were perforated in a few thousand
hours. The attack occurred at 540 to 870 °C
(1000 to 1600 °F).
Dunmore (Ref 82) reported a failure of the
waste heat boiler in an ammonia plant. The alloy
800 exit ferrules suffered severe metal wastage,
with large uniform circular pits penetrating
through the tubes and black sooty deposits on the
pitted surface. The attack occurred at a location
where the metal temperature was about 600 °C
(1110 °F). The environment was highly enriched
in H2 and CO, along with large quantities of
steam. This was unusual in that steam was con-
sidered an effective additive to mitigate the metal
dusting problem (Ref 7). Grabke and Spiegel
(Ref 83) discussed several industrial failure
cases that were caused by metal dusting. These
cases involved an alloy 800 heat exchanger for
synthesis gas production, HK-40 gas heaters in a
direct iron reduction plant, 5Cr and 9Cr steels in
catalyst regeneration units in a refinery, and an
alloy 601 heat exchanger in an ammonia plant.
Metal dusting has also been encountered
in the heat treating industry (Ref 84). Refractory
anchors, fan housing assemblies, and other
components in carburizing furnaces frequently Fig. 5.59 Multimet alloy fan box suffering metal dusting
attack in a carburizing furnace. (a) General view of
suffer metal dusting problems. Alloys typically failed the sample. Note the perforated edge of the fan box. (b)
used include stainless steels, such as Type 310, Cross section of the sample showing pitting attack and severe
metal thinning. (c) Severe carburization attack beneath the pitted
nickel-base alloys, such as alloys X and 333, and wasted area. Attack was initiated in the “dead” spaces created
and iron-base alloys, such as Multimet alloy by the refractory and the fan box. Source: Ref 84
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:23PM Plate # 0 pg 128

128 / High-Temperature Corrosion and Materials Applications

surface that was in contact with the furnace Model alloy (Type 410 + 2.75Si), and Type
refractory lining. The 1.9 cm (0.75 in.) thick fan 310 (Fe-25Cr-20Ni) in H2-24.4CO-2.4H2O at
housing was perforated in about a year. The 560 °C (ac > 1). The specimens were annealed in
nearby metal surface that was exposed to the H2 at 1000 °C (1832 °F) for 1 h to eliminate
circulating carburizing gas, on the other hand, surface deformation structure and produced grain
showed no sign of metal dusting (Fig. 5.60). A coarsening. The test specimens were ground.
chromium-rich oxide scale was observed on the Their results are summarized in Fig. 5.62. In this
metal surface. One common feature associated short-term test, alloys with low chromium con-
with metal dusting is carburization beneath the tents (9Cr for P91 and 12Cr for 410SS) showed
pitted area (Fig. 5.59). metal dusting attack. Type 310SS (25Cr-20Ni)
Manufacturing of carbon fibers for carbon showed no sign of metal dusting attack up to
composite materials can also generate high 200 h. Silicon was found to be a very effective
temperature carburizing environments for caus- alloying element in improving the resistance to
ing metal dusting problems to some furnace metal dusting. This is illustrated by the data
components. Carbon fibers are manufactured comparing “model” alloy (410SS + 2.75Si) with
from polyacrylonitrile (PAN) (Ref 85). The 410SS.
last step of manufacturing, “carbonization,” is Grabke et al. (Ref 69) investigated commercial
carried out at about 900 °C (1650 °F) in an alloys, which included Fe-Cr, Fe-Ni-Cr, Ni-base
environment enriched with CO, CO2, CH4, N2, alloys, and silicon-containing alloys, at 650 °C
HCN, H2O, and so forth. Furnace components (1200 °F) for 7 days in H2-24.7CO-1.9H2O (ac =
made of nickel- and iron-base alloys were found 15). All specimens were annealed in dry hydro-
to suffer general metal thinning and pitting gen at 1000 °C (1832 °F) for 1 h to achieve large
attack. An example is shown in Fig. 5.61. grains, followed by grinding and polishing. All
The general characteristics of metal dusting test specimens were in as-ground surface condi-
observed in the industrial environments were tions. The data are summarized in Fig. 5.63. The
reproduced in many laboratory tests involving
primarily gas mixtures of H2-CO-H2O. Accord-
ingly, laboratory testing has become an important
tool for generating relevant data to allow
performance ranking among engineering alloys.
Laboratory data generated by various labora-
tories are summarized below.
Maier and Norton (Ref 86) investigated
9Cr-1Mo steel (P91), 12Cr steel (Type 410),

Fig. 5.60 Oxide scale and no evidence of metal dusting on

the surface of a Multimet alloy fan box exposed to Fig. 5.61 Type 310SS furnace component suffering metal
flowing carburizing gas (same fan box as that shown in Fig. 5.59). dusting in a furnace used for manufacturing carbon
The sample was plated with nickel before mounting for metallo- fibers. (a) General view of the failed component. (b) Cross section
graphic examination. of the sample showing pitting and thinning
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:23PM Plate # 0 pg 129

Chapter 5: Carburization and Metal Dusting / 129

hatched data show mass gain of coke deposits on to the presence of some aluminum and silicon
the test specimen, while the solid data represent in the alloy. Both aluminum and silicon are
the either weight loss due to metal dusting or effective alloying additions in improving the
weight gain due to oxidation and/or carbon gain metal dusting resistance. X18 CrN28 (Fe-28Cr)
(no metal dusting attack). Sicromal (Fe-18Cr- showed significantly better performance than
1Al-1Si) contained not very high chromium Sanicro 28 (Fe-27Cr-30Ni). Both alloys con-
level, but showed no metal dusting. This was due tained about the same level of chromium, but one

250
560 °C, 1.5 bar

200
Mean metal loss, µm

150

100

50

0
0 50 100 150 200
Exposure time, h

Fig. 5.62 Average metal loss as a function of exposure time (hours) for P91, 410, 310, and model alloy (410 + 2.75Si). Source: Ref 86

1000

100

10

1
2
Mass gain, mg/cm

0.1

0.01

0.001

–0.01

–0.1

–1

–10

–100
Sicromal

X18 CrN28

Alloy 410

12 CrMoV

253 MA

HK 40

Sanicro 28

AC 66

Inconel 600
X15 CrNiSi 2012

Removable deposits Cleaned specimen

Fig. 5.63 Mass gain (hatched data) due to coke deposits and metal loss of specimen (solid data) due to metal dusting as well as
mass gain of specimen (solid data) due to possible oxidation/carburization after exposure at 650 °C (1200 °F) in
H2-24.7CO-1.9H2O (ac = 15). X18 CrN28 (Fe-28Cr); Sanicro 28 (Fe-27Cr-30Ni); 253MA (Fe-21Cr-11Ni-2Si-0.05Ce); X15 CrNiSi 2012
(Fe-20Cr-12Ni-2Si). Source: Ref 69
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:23PM Plate # 0 pg 130

130 / High-Temperature Corrosion and Materials Applications

Fe25Cr2.5Ni
0
Fe60Cr
–0.02 Fe25Cr

–0.04
Mass change, mg/mm2

–0.06

–0.08 Fe25Cr5Ni

–0.1

–0.12

–0.14 Fe25Cr25Ni
–0.15
Fe25Cr10Ni
–0.18
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210
Exposure cycles, h

Fig. 5.64 Weight loss of specimens after removing carbon deposits as a function of thermal cycles for Fe-Cr and Fe-Cr-Ni alloys in
metal dusting tests at 680 °C (1256 °F) in H2-68%CO-6%H2O (ac = 2.9 and pO2 =2 · 10723 atm) cycling specimens to room
temperature every 60 min in the test environment. Source: Ref 90

was a ferritic alloy (Fe-Cr) while the other was an with 2.5% Ni appeared to perform well. It was
austenitic alloy (Fe-Ni-Cr). The diffusivity of likely that the alloy with only 2.5% Ni remained a
chromium in ferritic alloys is generally about two ferritic structure. The test also included Fe-60Cr,
orders of magnitude higher than that in austenitic which showed good metal dusting resistance.
Fe-Ni-Cr alloys (Ref 87–89). As a result, ferritic Ferritic Fe-Cr alloys are more resistant to
stainless steels form chromium oxide scales metal dusting than austenitic Fe-Ni-Cr alloys
much more readily than austenitic stainless presumably due to much faster chromium diffu-
steels, as was observed in Fig. 5.63. Both 12Cr sion rates, thus resulting in faster formation of
steels (410SS and 12CrMoV) showed some chromium oxide scales, and then better metal
metal dusting attack. Again, silicon was found dusting resistance. Rapid diffusion of chromium
to greatly improve metal dusting resistance. to the metal surface to form a protective chro-
Both alloy 253MA (Fe-21Cr-11Ni-2Si-0.05Ce) mium oxide scale is important in retarding metal
and X15 CrNiSi 2012 (Fe-20Cr-12Ni-2Si) dusting attack. Accordingly, for the same alloy, a
showed no metal dusting attack. Another silicon- fine-grained material can form a surface oxide
containing alloy, HK-40 (Fe-25Cr-20N-2Si), scale more readily than a coarse-grained mate-
showed some metal dusting attack. The alloys rial, thus better metal dusting resistance. This is
that showed the worst attack in this test were illustrated in Fig. 5.65, where the fine-grained
Fe-Ni-Cr alloys (Sanicro 28 and AC66), and a Type 304SS (as-received from the supplier)
low-chromium nickel alloy (alloy 600). with a grain size of ASTM No. 10 (average size
The results of Toh et al. (Ref 90) further con- of 10 µm) showed no metal dusting attack while
firmed that Fe-Cr alloys exhibit better metal the coarse-grained material with a grain size
dusting resistance than Fe-Ni-Cr alloys. They of about ASTM No. 3 to 7 (about 30–100 µm)
investigated Fe-25Cr alloy and Fe-25Cr with 5, suffered severe attack (Ref 91). The coarse-
10, and 25% Ni. All alloys were annealed in grained material was obtained by annealing
Ar-10%H2 at 1050 °C (1922 °F) for 100 h. the sample to 1000 °C (1832 °F) for 1 h.
Test specimens were then cut, ground, and Metal dusting testing was conducted at 600 °C
polished to 3 µm before being electrolytically (1112 °F) in H2-24CO-2H2O. More discussion
polished. Testing was conducted at 680 °C on the effects of surface conditions on the metal
(1256 °F) in H2-68%CO-6%H2O (ac = 2.9 and dusting behavior is presented later.
pO2 =2 · 10723 atm). Specimens were cycled by Fe-Ni-Cr alloys, which showed less resistance
heating them to the test temperature for 60 min to metal dusting than Fe-Cr alloys due to lower
for each cycle. Their test results are summarized chromium diffusivity, are also much less resis-
in Fig. 5.64. Fe-25Cr alloy was found to be sig- tance to metal dusting as compared with Ni-base
nificantly more resistant to metal dusting than alloys. This is illustrated in Fig. 5.66, which were
Fe-25Cr alloys with 5, 10, and 25% Ni. Fe-25Cr generated on the commercial alloys without prior
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:23PM Plate # 0 pg 131

Chapter 5: Carburization and Metal Dusting / 131

annealing treatments. Specimens were in the as- Type 304, 310, and 800H—which are Fe-Ni-Cr
received condition from the supplier and were alloys—were found to suffer metal dusting attack
ground to a 120-grit finish prior to testing in (i.e., the reaction rates in weight loss) while two
H2-90%CO for 672 days at 482, 566, 649, and nickel-base alloys (alloy 601 and RA333)
732 °C (900, 1050, 1200, and 1350 °F) (Ref 92). showed no metal dusting attack (i.e., the reaction
rates in weight gain). The figure also shows that
alloy 85H (Fe-18.5Cr-14.5Ni-3.5Si-1Al) showed
10
no metal dusting attack. This was due to bene-
ficial effect of silicon along with aluminum in the
18Cr-8Ni-Steel
alloy. Beneficial effects of silicon and aluminum
8 are demonstrated by additions of these two ele-
ments to alloy 800 (Fe-20Cr-32Ni), as shown in
Fig. 5.67 (Ref 93). Also in this study by Strauss
Mass gain, mg/cm2

6 and Grabke (Ref 93), other alloying elements,

such as niobium, molybdenum, and tungsten,
Coarse grain were found to have some beneficial effects as
4 well.
Klower et al. (Ref 94) investigated a large
number of commercial alloys including several
2
Fe-Ni-Cr alloys, such as alloys 800H, HK-40,
Fine grain HP-40, and DS, and many nickel-base alloys.
0
Test results are summarized in Table 5.5.
0 50 100 150 200 Fe-Ni-Cr alloys were much less resistant to metal
Time, h dusting than nickel-base alloys. High-silicon-
containing alloy DS (Fe-18Cr-35Ni-2.2Si) was
Fig. 5.65 Effect of grain size on the metal dusting behavior
much better than other Fe-Ni-Cr alloys. This
of Type 304SS tested at 600 °C (1112 °F) in H2-
24CO-2H2O. Source: Ref 91 is nicely illustrated in Fig. 5.68. In Ni-Cr-Fe

Fig. 5.66 Metal dusting resistance of several Fe-Ni-Cr alloys and Ni-base alloys tested in H2-90%CO for 672 days at 482, 566, 649,
and 732 °C (900, 1050, 1200, and 1350 °F). The alloys tested were Type 304 (S30403), Type 310 (S31000), alloy 85H
(S30615), alloy 800H (N08810), alloy 601 (N06601), and RA333 (N06333). Source: Ref 92
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:23PM Plate # 0 pg 132

132 / High-Temperature Corrosion and Materials Applications

system, increases in chromium along with 2.3% Al in alloy 602CA, alloy 602CA specimens
aluminum additions can significantly improve might have preformed an Cr/Al-rich oxide scales
the alloy’s metal dusting resistance, as illustrated prior to metal dusting testing. The effects of
in Fig. 5.69. Alloy 601 was more resistant than surface finishes are discussed further.
alloy 600 because of higher chromium and alu- In Table 5.5, it is surprising to find that alloy
minum addition in alloy 601. Alloy 602CA was 214 with about 4.5% Al did not perform well
better than alloy 601 because of higher aluminum compared with 602CA (2.3% Al), 617 (1.3% Al),
content. There were two different surface finishes and 690 (30% Cr, no Al). It is believed that with
involved in this test, with alloys 600 and 601 in as only about 16% Cr, alloy 214 did not contain
ground surface, and alloy 602CA in “black” enough chromium to rapidly develop a protective
surface finish. There was no discussion about chromium oxide scale at such a low temperature.
the procedure for preparing the “black” surface And, at this low temperature, development of an
finish in the paper. If the 602CA specimens were exclusive Al2O3 is difficult in short time. Thus, a
black annealed (heat treated in air), with about

Table 5.5 Metal wastage rates after exposure

20 to H2-CO-H2O gas mixtures at 650 °C (1200 °F)
+4.5%AI +2.5%Si for various commercial alloys
+ +
Surface Wastage rates,
+3%Mo Alloy condition Total time, h mg/cm2h
+3%Nb
Mass loss, mg/cm2

–20 800H Ground 95 0.21

+3%W HK-40 … 190 0.04
HP-40 … 190 0.038
DS Ground 1,988 4.3 × 10−3
Fe-32%Ni-20%Cr 600H Ground 5,000 3.3 × 10−2
601 Black 6,697 7.3 × 10−3
–60 601 Polished 1,988 4.9 × 10−3
601 Ground 10,000 5.8 × 10−4
C-4 Ground 10,000 1.1 × 10−3
+0.05%Ce 214 Ground 9,665 1.2 × 10−3
HR160 Ground 9,665 6.3 × 10−4
–100 45TM Black 10,000 1.0 × 10−5
0 500 1000 1500 602CA Black 10,000 1.1 × 10−5
Time, h 617 Ground 7,000 3.7 × 10−6
690 Ground 10,000 2.0 × 10−6
Fig. 5.67 Effects of Al, Si, Nb, Mo, and W additions to a model Note: Gas mixtures H2-24CO-2H2O (ac = 14 at temperature) for the first 5000 h of
alloy (alloy 800, Fe-20Cr-32Ni) on the metal dust- testing and H2-49CO-2H2O for the subsequent 5000 h. Source: Ref 94
ing behavior of modified experimental alloys. Source: Ref 93

102

10
Metal wastage rate, mg/cm2h

10–1

10–2

10–3

10–4

10–5

Fig. 5.68 Metal dusting resistance of Fe-Ni-Cr alloys (800H and HP40) in comparison with Ni-base alloy 600H tested at 650 °C
(1200 °F) in H2-24CO-2H2O (ac = 14). Source: Ref 94
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:23PM Plate # 0 pg 133

Chapter 5: Carburization and Metal Dusting / 133

high-chromium nickel-base alloy, alloy 690 alloy 45TM specimen was in a “black” surface
(30% Cr), performed the best in this test. It is also condition. The “black” surface condition might
surprising to find that HR-160 alloy with both be resulted from a preoxidation treatment that
high Cr (28%) and high Si (2.75%) did not per- might have produced chromium- and silicon-rich
form as well as some other nickel-base alloys. oxide scales prior to metal dusting testing.
Alloy 45TM, also with high Cr (27%) and high Si In conducting metal dusting testing that
(2.7%) performed better than HR-160 alloy. The involved only the specimens with as-ground

10–1
Alloy 600

10–2
Metal wastage rate, mg/cm2h

Alloy 601

10–3

10–4

10–5 Alloy 602CA

10–6
0 2,000 4,000 6,000 8,000 10,000
Exposure time, h

Fig. 5.69 Increasing Cr along with addition of aluminum improves the metal dusting resistance of the alloy in Ni-Cr-Fe alloys. Source:
Ref 94

Fig. 5.70 Metal dusting behavior of various Ni-base alloys tested at 593 °C (1100 °F) in H2-18.4CO-5.7CO2-22.5H2O at 14.3 atm of
pressure. Source: Ref 95
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:23PM Plate # 0 pg 134

134 / High-Temperature Corrosion and Materials Applications

surface finish, Natesan and Zeng (Ref 95) per- scale. The Raman spectra of the alloy 214 tested
formed tests at 593 °C (1100 °F) in H2-18.4CO- specimen showed low intensity for the Cr2O3
5.7CO2-22.5H2O at 14.3 atm of pressure. Alloy band (Ref 95).
45TM suffered rapid metal dusting attack with In manufacturing of sheet products (or thin
about 59.1 mg/cm2 of weight loss after only gage tubular products) in nickel-base alloys, the
3300 h while HR-160 alloy exhibited only manufacturing process in the later stage typically
7.3 mg/cm2 of weight loss after 9700 h. Their consists of cold working (cold pilgering) and
results are summarized in Fig. 5.70. The best- bright annealing (i.e., annealing in H2 atmo-
performing alloy was alloy 693 (30% Cr, 3.3Al), sphere). The hydrogen atmosphere in bright
followed by alloy 602CA (25% Cr, 2.3% Al), annealing typically contains some moisture, thus
and HR-160 (28% Cr, 2.75% Si) and alloy 690 it is typically characterized with a certain dew
(30% Cr). The worst performer was alloy 45TM, point (i.e., a certain oxygen potential, pO2 ). In
followed by alloys 617, 214, and 601. These tests general, annealing furnaces exhibit low enough
conducted in Argonne National Laboratory were dew points such that chromium oxide scales do
under a high pressure (14.3 atm, or 210 psi), and not form on the metal surface when processing
all other data generated in other laboratories were stainless steels or nickel-base alloys. Sheet pro-
under gas pressure close to 1 atm (atmosphere ducts of these alloys typically look shiny. How-
pressure). Natesan and Zeng (Ref 95) observed ever, the dew points or oxygen potentials in
that high pressure could significantly reduce hydrogen-annealing furnaces generally are high
the time to initiate metal dusting attack for enough that Al2O3 or SiO2 will form on the metal
some alloys. Tests were performed at 593 °C surface of a bright-annealed sheet of a nickel
(1100 °F) for 246 h in 1, 14.3, and 40.8 atm alloy containing a sufficient level of aluminum or
pressures. Specimens were then examined for silicon. Klarstrom and Grabke (Ref 96) tested
signs of metal dusting attack. Their results are bright-annealed sheet specimens of alloy 214
summarized in Table 5.6, showing alloys 601, (alumina former) and alloy HR160 (silica
690, 617, and 214 suffering metal dusting attack former) along with alloy 230 and HR120 in the
at high pressures but not at 1 atm pressure after bright-annealed condition. Also included in
246 h. On the other hand, alloys 45TM, 602CA, testing were black annealed and pickled sheet
and HR160 showed no metal dusting attack at specimens of alloys 800H and 601 (no bright-
both low and high pressures after 246 h. They annealed sheet samples were available for these
also measured the maximum pit size and average two alloys at the time of testing). No surface
pit depth in addition to weight loss. This is grinding for bright-annealed sheet specimens
illustrated in Table 5.7. In terms of the pitting (alloys 214, HR160, 230, and HR120). Surface
depth, HR160 was found to perform the best, grinding to a 120-grit surface finish was per-
while alloy 690 with little weight loss showed formed for black-annealed and pickled sheet
significant pit depth. The data suggested that the specimens of alloys 800H and 601. Testing
total weight loss was not correlated well with was performed at 650 °C (1200 °F) for up to
pitting depth. Alloy 214 was found to show 10,000 hours in H2-49CO-2H2O (ac =18.9).
uniform metal wastage. This was apparently due Test results are summarized in Table 5.8. HR160
to insufficient chromium content (16%) at such a alloy was found to perform very well, showing
low temperature to form a continuous Cr2O3 no evidence of metal dusting after 10,000 h

Table 5.6 Surface conditions of alloys Table 5.7 Pit size, pit depth, and weight loss
after testing at 593 °C (1100 °F) for 246 h in for alloys after testing at 593 °C (1100 °F) for
H2-18.4CO-5.7CO2-22.5H2O at 1, 14.3, and 9700 h in H2-18.4CO-5.7CO2-22.5H2O
40.8 atm pressures at 14.3 atm pressure
Condition at pressure Alloy Weight loss, mg/cm2 Pit depth, µm Pit diameter, µm
Alloy 1 atm 14.3 atm 40.8 atm 601 19.5 110 450
601 Clean surface Pits Pits 690 6.5 147 440
690 Clean surface Pits Pits 617 35.1 201 887
617 Clean surface Pits Pits 602CA 2.1 96 374
602CA Clean surface Clean surface Clean surface 214 25.6 (a) (a)
214 Clean surface Pits Pits 45TM(b) 59.1 141 600
45TM Clean surface Clean surface Clean surface HR160 7.3 13 210
HR160 Clean surface Clean surface Clean surface 693 0.1 37 99
Source: Ref 95 (a) Specimen uniformly corroded. (b) Exposed for only 3300 h. Source: Ref 95
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:23PM Plate # 0 pg 135

Chapter 5: Carburization and Metal Dusting / 135

of exposure. The alloy, however, showed slight dusting attack with fairly high wastage rates.
carburization underneath the metal surface. Both Fe-Ni-Cr alloys, alloys 800H, and HR120,
Alloy 230 showed slight metal dusting attack performed the worst among the alloys tested.
with a few spots 1 to 2 mm in diameter on each Figure 5.72 shows typical surface conditions of
side of the specimen after 10,000 h of exposure. alloys 214, HR120, and 800H after 5707 h,
Alloy 601 suffered significantly more metal 190 h, and 925 h, respectively.
dusting attack than alloy 230. Typical surface Baker and Smith (Ref 97, 98) and Baker et al.
conditions for these three alloys after 10,000 h (Ref 99) investigated about 20 commercial alloys
of testing are shown in Fig. 5.71. Alloy 214, in H2-80CO at 621 °C (1150 °F) for times up to
however, did not perform well, suffering metal 16,000 h. All materials were cold rolled and
annealed sheets (i.e., bright annealed products)
Table 5.8 Final metal wastage rates for various
except alloys K-500 and 617, which were hot-
alloys tested at 650 °C (1200 °F) for up to rolled, annealed plates (i.e., black annealed and
10,000 h in H2-49CO-2H2O (ac = 18.9) pickled products), and alloy DS, which was
Alloy Total exposure time, h Wastage rate, mg/cm2h
extruded and annealed tubing. Nevertheless, test
HR120 190 4.1 × 10−2
coupons were all ground to a 120-grit finish.
800H 925 2.7 × 10−3 For exposure times less than 10,000 h, weight
214 5,707 1.0 × 10−3 changes as a function of exposure time for
601 10,000 2.5 × 10−3
230 10,000 3.2 × 10−4
Fe-Ni-Cr alloys as well as 9Cr steel and Monel in
HR160 10,000 0.0(a) comparison with some nickel-base alloys are
(a) Attack too small for analysis. Source: Ref 96

Fig. 5.71 Optical micrographs showing typical surface con-

ditions for (a) HR160 alloy, (b) alloy 230, and (c) Fig. 5.72 Typical surface conditions for (a) alloy 214
alloy 601 after testing at 650 °C (1200 °F) for 10,000 h in H2- after 5707 h, (b) alloy HR120 after 190 h, and
49CO-2H2O (ac = 18.9). Magnification bar represents 20 µm for (c) alloy 800H after 925 h at 650 °C (1200 °F) in H2-49CO-2H2O
(a), 200 µm for (b) and (c). Source: Ref 96 (ac = 18.9). Source: Ref 96
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:23PM Plate # 0 pg 136

136 / High-Temperature Corrosion and Materials Applications

summarized in Fig. 5.73. Nickel-base alloys maximum pitting depth as a function of exposure
except alloy 600 performed much better than time. Alloys 690, 617, and MA956 showed the
Fe-Ni-Cr alloys. Nickel-base alloy 600 per- minimum pitting attack. Figure 5.75 shows
formed poorly because of its low chromium the test results for some additional alloys
content (15–16%). MA956 (Fe-20Cr-4.5Al- and some better performing alloys tested up
Y2O3 ODS alloy) performed very well. The test to 16,000 hours (Ref 99). Alloys 690, 263, and
results are also presented in Fig. 5.74 in terms of MA956 were observed to show pits with

20

617 263
0
DS 690 MA 956
803
825

864 601
–20
Mass change, mg/cm2

600
–40
800
K-500
9Cr-1Mo

–60

–80

–100

330
–120
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Exposure time, h

Fig. 5.73 Mass change as a function of exposure time for various alloys including 9Cr steel, Ni-Cr alloy (Monel K-500), Fe-Ni-Cr
alloys, and Ni-base alloys tested at 621 °C (1150 °F) in H2-80CO. All surfaces were ground to a 120-grit finish prior to
testing. Source: Ref 98

30 762

800 635
25
DS
Maximum pitting depth, mils

803 601
Maximum pitting depth, µm

508
o

20
-1M

330
9Cr

15 381

10 864 254
5
82
5 127
690
600 617 MA956
0 0

K500
–5 –127
0 1000 2000 3000 4000 5000 6000 7000 8000 9000
Exposure time, h

Fig. 5.74 Maximum pit depths as a function of exposure time for various alloys including 9Cr steel,Fe-Ni-Cr alloys and Ni-base alloys
tested at 621 °C (1150 °F) in H2-80CO. All surfaces were ground to a 120-grit finish prior to testing. Source: Ref 98
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:24PM Plate # 0 pg 137

Chapter 5: Carburization and Metal Dusting / 137

significant pit depths after 14,000–16,000 h of than that of alloy 693 (Fig. 5.76). It should be
exposure. Alloys 617 and 693, on the other hand, cautioned in interpreting pit depth data, when the
still exhibited pits with insignificant pit depths. specimen surface has shown general wastage
Alloy 693 was a recently developed commercial without the original specimen surface surround-
alloy targeting for metal dusting environments by ing the pit, pitting depth data thus measured then
adding about 3% Al to alloy 690 with same becomes questionable. This would particularly
amount of Cr (about 30%) but slightly lowered be the case after very long exposure when pits
Fe. The alloy was found to perform much better have spread throughout the specimen surface.
than alloy 690. It is understandable that high Cr In reviewing the data summarized in Fig. 5.75
and high Al in nickel-base alloys should perform and 5.76, one alloy of interest was alloy 671
well in metal dusting environments. However, it (Ni-46Cr), which was found to exhibit excellent
was surprising to find that alloy 617 with only metal dusting resistance. Unfortunately, the test
22% Cr (quite normal level chromium for high- on this alloy was terminated at close to 10,000 h.
temperature alloys) and fairly low aluminum This alloy exhibited low pit depths and very low
content (about 1.3%) exhibited a pitting depth metal loss rates (lower than alloy 693). Grabke et
similar to that of alloy 693. In metal dusting al. (Ref 100) also found that Ni-50Cr alloy was
testing by Natesan and Zeng (Ref 95) involving very resistant to metal dusting, showing a thin
several nickel-base alloys including alloy 617, chromium oxide scale (about 3 µm thick) after
the test results, which were presented as weight testing in H2-49CO-2H2O at 650 °C (1200 °F)
loss, showed alloy 617 was worse than alloy 601, (ac = 18.9) for 10,000 h with no evidence of
602CA, and several other nickel-base alloys metal dusting or carburization. Only minute
(Fig. 5.70). When Baker and Smith (Ref 98) amounts of coke were observed on the specimen
presented their test results in terms of weight loss surface. Under this test condition, alloy 601
rate as a function of exposure time, alloy 617 was had already suffered metal dusting attack after
found to exhibit metal loss rate (in the range of only 1993 h of exposure. Similarly, a very
alloys 263 and MA956) significantly higher thin chromium oxide scale was observed in

25
24 800 DS
610
23
22 803 559
21
20 601
508
19 956
18
0

457
33

17 263
16 406
15 602CA
Pit depth, mils

Pit depth, µm

14 356
690

13
12 305
11
10 4 254
86
9
8 5 203
82 TD 758
7 LCE
625
400

6 152
5
4 C–276 671 102
3
754
2 617 693 51
1
0 600 0
–1
0 2000 4000 6000 8000 10000 12000 14000 16000 18000
Exposure time, h

Fig. 5.75 Maximum pit depths as a function of exposure time for various alloys including 9Cr steel, Fe-Ni-Cr alloys, and Ni-base alloys
tested at 621 °C (1150 °F) in H2-80CO for exposure time up to 16,000 h. All surfaces were ground to a 120-grit finish prior to
testing. Source: Ref 99
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:24PM Plate # 0 pg 138

138 / High-Temperature Corrosion and Materials Applications

100
0

0
40

33

10–1
6 0
80
82 00
5

10–2
Mass loss rate, mg/cm2h

3
75

80

1
864

60
DS 625
690
10–3 C 276
7 58 602CA 263
TD MA
617
10–4 956

10–5

693

671

0 2000 4000 6000 8000 10000 12000 14000 16000 18000

Exposure time, h

Fig. 5.76 Metal loss rate as a function of the exposure time for various alloys tested at 621 °C (1150 °F) in H2-80CO for exposure time
up to 16,000 h. All surfaces were ground to a 120-grit finish prior to testing. Source: Ref 99

Cr-5Fe-1Y2O3 after 10,000 h of exposure in the Fig. 5.65. This is because grain boundaries pro-
same test, showing no evidence of metal dusting vide fast diffusion paths for chromium to reach
attack. to the metal surface. A fine-grain-sized material
will have more grain boundaries and, thus, more
chromium reaching to the metal surface to form a
5.3.4 Effects of Surface Conditions and
better chromium oxide scale faster than a coarse-
Finish grained material, thus, resisting metal dusting
From the discussion so far on metal dusting, attack much better.
formation of a good, protective chromium oxide The surface of the metal can also be prepared
scale is a very effective way to provide protection by grinding or machining to produce a thin cold-
against metal dusting attack. Since the tempera- worked layer with high density of dislocations,
ture range for metal dusting attack is quite low which also provide fast diffusion paths for
and chromium diffusivities are relatively low at chromium to reach to the metal surface to form a
these temperatures, rapid formation of a good, protective chromium oxide scale. As a result,
protective chromium oxide scale is critical. metal dusting is greatly improved when the sur-
Accordingly, the concentration of chromium at face is ground or machined. The beneficial effect
the surface of the metal becomes important. of the ground surface condition is illustrated
Higher chromium concentration at the metal Fig. 5.77 and 5.78 in a thermogravimetric study
surface provides faster formation of a protective by Grabke et al. (Ref 101). For both alloy
chromium oxide scale. Thus, alloys with higher 800 and Type 310SS, the as-ground specimen
bulk chromium concentration (higher surface was most resistant to metal dusting compared
chromium concentration) are more resistant to with the as-received surface (cold rolled) and
metal dusting as was discussed in the previous electropolished surface. Electropolished surface
section. Also, for the same alloy, the fine-grained condition was the worst because of possible
structure exhibits better metal dusting resistance surface depletion in chromium during electro-
than the coarse-grained structure, as illustrated in polishing. Both alloy 800 and Type 310SS were
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:24PM Plate # 0 pg 139

Chapter 5: Carburization and Metal Dusting / 139

tested under the same condition; Type 310 sample. The black-annealed sample, as in
was much more resistant to metal dusting than electropolished specimen, exhibited surface
alloy 800. This is primarily because Type chromium depletion due to annealing in air (i.e.,
310SS contains more chromium. Specimens black annealing). The formation of chromium
with ground surface finish, alloy 601, showed oxide scales during annealing in air results in
better metal dusting resistance than those with some chromium depletion in the surface layer.
electropolished surface finish, as shown in Similar results were obtained by Baker
Fig. 5.79 (Ref 94). In this study, the “black” and Smith (Ref 98) in their study on alloy 601
sample (believed to be a black-annealed sample) with different surface conditions, as shown in
performed as poorly as the electropolished Fig. 5.80 and 5.81. In this study, samples from
as-received sheet manufactured from black-
annealing and pickling (i.e., annealing in air
10 followed by acid pickling to remove oxide
Alloy 800 scales), which were identified as as-produced
(annealed+ pickled), were slightly better than
8 those black-annealed samples and electro-
Electropolished polished samples. Black annealing and acid
Mass gain, mg/cm2

Cold rolled pickling can often result in surface depletion in

6
chromium. These processes are often involved
in producing sheet products by hot rolling.
4
When sheet products are produced via cold
Ground
rolling, the sheet is typically bright annealed.
Under this condition, no chromium oxide scales
2 formed on the sheet metal; thus no chromium
depletion occurred at the metal surface. For high-
aluminum or high-silicon alloys, bright-anneal-
0 ing atmospheres typically exhibit high enough
0 5 10 15 20 25
Time, h oxygen potentials (or dew points) that a thin
aluminum or silicon oxide film can form, which
Fig. 5.77 Thermogravimetric testing in H2-24CO-2H2O at can be exceedingly beneficial to the subsequent
600 °C (1112 °F) for alloy 800 (Fe-21Cr-32Ni) in
three different surface conditions: as-ground surface (to 600-grit),
service in metal dusting environments. This
cold-rolled (as-received surface), and electropolished surface. was discussed earlier for HR160. Alloy 214,
Source: Ref 101 unfortunately with too low a chromium level,
showed no beneficial effect from bright anneal-
ing in its resistance to metal dusting attack.
6

25Cr-20Ni - Steel
5.3.5 Metal Dusting Behavior of Weldments
Selection of an appropriate filler metal for
Mass gain, mg/cm2

Electropolished
4 metal dusting environments is also critical, since

2
Cold rolled

Ground
0
0 5 10 15 20 25
Time, h

Fig. 5.78 Thermogravimetric testing in H2-24CO-2H2O at

600 °C (1112 °F) for Type 310SS (Fe-25Cr-20Ni) Fig. 5.79 Metal dusting behavior at 650 °C (1200 °F) in
in three different surface conditions: as-ground surface (to 600- H2-49CO-2H2O for alloy 601 in three different
grit), cold-rolled (as-received surface), and electropolished sur- surface conditions: as-ground surface, elecropolished surface,
face. Source: Ref 101 and black-annealed surface. Source: Ref 94
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:24PM Plate # 0 pg 140

140 / High-Temperature Corrosion and Materials Applications

not every wrought alloy has a matching filler weldability. Also, in many cases, machining or
metal. Thus, it is necessary to select a filler metal grinding the weld joint may not be possible,
that is at least as good as, but preferably better particularly in tubular butt welds where inside
than, the wrought alloy selected for the applica- diameter (ID) grinding or machining is not
tion. The suitable filler metal requires not only feasible after joining. Furthermore, alloys that
good resistance to metal dusting but also good resist metal dusting contain high aluminum or

Ground

Fig. 5.80 Metal dusting behavior in terms of weight change tested at 621 °C (1150 °F) in H2-80CO for alloy 601 in various surface
conditions. Source: Ref 98

Fig. 5.81 Metal dusting behavior in terms of weight loss rate tested at 621 °C (1150 °F) in H2-80CO for alloy 601 in various surface
conditions. Source: Ref 98
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:24PM Plate # 0 pg 141

Chapter 5: Carburization and Metal Dusting / 141

silicon along with high chromium, as discussed were “brushed,” with one weldment being
in previous sections. If matching filler metals are ground, one being sandblasted, and one being
available for these alloys, the weldability of these pickled. Tests were conducted at 600 and 650 °C
high-aluminum or high-silicon filler metals can (1112 and 1200 °F) in H2-24CO-2H2O. In
be an issue. almost all cases, metal dusting was initiated at the
Grabke et al. (Ref 101) investigated the metal interface between the weld metal and the base
dusting behavior of different weldments invol- metal (i.e., the heat-affected zone surface). Alloy
ving base metals of alloys 800H, 600H, 601, and 82 weld metal was found to be more resistant to
602CA using different filler metals, such as alloy metal dusting than alloy 800H and alloy 600H.
82 (Nicrofer S 7020 or UTP 068 HH) and alloy The authors (Ref 101) also concluded that TIG
602CA (Nicrofer 6025 or UTP 6225 Al). Weld- welding led to a better resistance to metal dusting
ment specimens included alloy 800H to alloy than “hand-welding,” and grinding led to a
800H with alloy 82 filler metal, alloy 600H to modest delay in initiation of metal dusting attack.
alloy 600H with alloy 82 filler metal, alloy 601H
to alloy 601H with alloy 602CA, and alloy 800H 5.3.6 Effects of Sulfur on Metal Dusting
to alloy 602CA with alloy 602CA as a cap layer
and alloy 82 filler metal for root pass and filler. Sulfur, which has a strong tendency to segre-
Welding processes involved were listed as TIG gate to the surface and grain boundaries, can
welding (i.e., gas-tungsten arc welding) and act as an inhibitor to metal dusting. Hochman
MMAW (probably shielded metal arc welding). (Ref 67) first reported beneficial effect of sulfur
The surface finish for most weldment specimens against metal dusting attack. Extensive investi-
gations on the effect of sulfur on metal dusting
behavior of iron were carried out by Grabke
1.0
and Muller-Lorenz (Ref 102), Schneider et al.
no H2S 0.1 ppm H2S (Ref 103), Schneider et al. (Ref 104), and
0.3 ppm H2S 0.5 ppm H2S 0.75 ppm H2S
0.8
Schneider and Grabke (Ref 105). The mechan-
ism for sulfur to inhibit metal dusting attack,
∆m/A, mg/cm2

0.6 as proposed by Grabke and Muller-Lorenz

(Ref 102), is the absorption of sulfur on the
0.4 surface of cementite suppresses the nucleation of
1 ppm H2S
graphite, thus inhibiting metal dusting attack.
0.2 Figure 5.82 shows the beneficial effect of H2S on
the kinetics of metal dusting of iron at 500 °C
0.0
0 50 100 150 200 250 300
Time, h

Fig. 5.82 Effect of H2S on metal dusting behavior of iron

at 500 °C (932 °F) in H2-CO-H2O-H2S gas mixture
(ac = 100). Source: Ref 103

1.50
no H2S
0.1 ppm
0.7 ppm H2S T=700 °C
1.25
H2S 1 ppm H2S ac = 100
∆m/A, mg/cm2

1.00 0.5 ppm

H2S
0.75
5 ppm H2S
0.50
15 ppm H2S
0.25

0.00
0 50 100 150 200 250 300 350
Time, h
Fig. 5.84 Effect of H2S on metal dusting behavior of iron in
terms of pH2 S =pH2 versus 1/T. Open data points
represent that the onset of metal dusting was retarded for more
Fig. 5.83 Effect of H2S on metal dusting behavior of iron at than 48 h, while the solid data points represent metal dusting
700 °C (1292 °F) in H2-CH4-H2S gas mixture (ac = without retardation. The hatched region represents a transition to
100). Source: Ref 104 an iron surface saturated with sulfur. Source: Ref 74
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:24PM Plate # 0 pg 142

142 / High-Temperature Corrosion and Materials Applications

(932 °F) in H2-CO-H2O-H2S gas mixture (ac = Alumina formers (i.e., alloys forming Al2O3
100) (Ref 103), and Fig. 5.83 shows the similar scale, such as alloy 214 and MA956) are sig-
effect for iron at 700 °C (1292 °F) in H2-CH4- nificantly better than chromia formers (i.e., alloys
H2S gas mixture (ac = 100) (Ref 103). Grabke forming Cr2O3 scale).
(Ref 74) provides metal dusting behavior of iron Surface finish plays an important role in
in terms of pH2 S =pH2 ratio as a function of tem- carburization resistance. For cast products,
perature (1/T) in Fig. 5.84. The figure shows the machining the metal surface can significantly
region when metal dusting was avoided by sulfur reduce carburization attack. Injecting sulfur
injection under certain pH2 S =pH2 ratios at differ- compounds (e.g., 50 to 100 ppm) into the pro-
ent temperatures from about 500 to 1000 °C (932 cessing gas stream is also effective in reducing
to 1832 °F). Injecting a right amount of sulfur carburization attack.
into the environment to retard metal dusting Metal dusting is another form of carburization
without causing accelerated sulfidation attack is a attack; it typically causes an alloy to suffer pitting
balancing act. attack and/or thinning. The metal beneath the
pitted area generally shows carburization. The
corrosion products typically consist of carbon
5.4 Summary soots, metal particles, carbides, and oxides. The
environment in which metal dusting occurs
Metals and alloys are generally susceptible to generally contains H2, CO, CO2, and H2O with
carburization when exposed to environments high carbon activities (i.e., aC > 1). Stagnant gas
containing CO, CH4, or other hydrocarbon gases conditions can be conducive in initiating metal
at elevated temperatures. Carburization typically dusting attack. The metal temperatures at which
results in the formation of internal carbides in the metal dusting occurs are between 430 and 900 °C
matrix as well as boundaries, causing the alloy (800 and 1650 °F). Metal dusting data for various
to lose its room-temperature ductility and/or commercial alloys are presented. Nickel-base
creep-rupture strengths. alloys containing high chromium and high
Fe-Ni-Cr alloys are widely used for processing aluminum (e.g., alloys 602CA and 693) or con-
equipment to resist carburization in the petro- taining high chromium and high silicon (e.g.,
chemical industry. The cast 25Cr-20Ni, HK40, alloy HR160) showed excellent resistance
was once the workhorse of pyrolysis furnace to metal dusting attack. Surface conditions
tubes in ethylene cracking operations. Many also play an important role in metal dusting
modifications based on HK40 have been devel- resistance. Sulfur may also retard metal dusting
oped and used now with improved carburization attack.
resistance as well as increased creep-rupture
strengths. These alloy modifications involve the
use of alloying elements such as, titanium, REFERENCES
niobium, tungsten, molybdenum, and silicon, as
well as increases in nickel and/or chromium. 1. Metals Handbook, 8th ed., Vol 2, American
Increasing nickel in Fe-Ni-Cr alloys improves Society for Metals, 1964, p 93
carburization resistance. Nickel reduces the dif- 2. L.J. Haga, Heat Treat., Dec 1986, p 6
fusivity of carbon in Fe-Ni-Cr alloys. Nickel also 3. G.E. Moller and C.W. Warren, Paper No. 237,
reduces the solubility of carbon in Fe-Ni alloys. Corrosion/81, NACE, 1981
Among these alloying elements, silicon is the 4. A.J. McNab, Hydrocarbon Process., Dec
most effective in improving carburization re- 1987, p 43
sistance. This is attributed to the formation of 5. G.L. Swales, in Behavior of High Tempera-
SiO2 scale, which is more impervious to carbon ture Alloys in Aggressive Environments, Proc.
ingress than Cr2O3 scale. However, when the 1979 Petten International Conference, I.
silicon content in the alloy is too high, the Kirman et al., Ed., The Metals Society,
weldability of the alloy can become a serious London, 1980, p 45
issue. 6. R.A. Perkins, W.C. Coons, and F.J. Radd, in
Aluminum is another alloying element that Properties of High Temperature Alloys, Proc.
can significantly improve carburization resis- 1976 Fall Meeting of the Electrochemical
tance. However, effectiveness generally requires Society, The Electrochemical Society
about 4% or higher aluminum, the amount 7. R.C. Schueler, Hydrocarbon Process., Aug
needed to form a continuous Al2O3 scale. 1972, p 73
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:24PM Plate # 0 pg 143

Chapter 5: Carburization and Metal Dusting / 143

8. F.A. Prange, Corrosion, Vol 15 (No. 12), 26. S. Forseth and P. Kofstad, Carburization
Dec 1959, p 619t of Fe-Ni-Cr Steels in CH4-H2 Mixtures at
9. G.Y. Lai, M.F. Rothman, and D.E. 850–1000 °C, Mater. Corros., Vol 49, 1998,
Fluck, Paper No. 14, Corrosion/85, NACE, p 266
1985 27. D. Jakobi and R. Gommans, Typical Failures
10. O.J. Dunmore, Proc. UK Corrosion Conf. in Pyrolysis Coils for Ethylene Cracking,
1982, Institution of Corrosion Science and Mater. Corros., Vol 54 (No. 11), 2003, p 881
Technology, Birmingham, UK, 1982, p 101 28. R.H. Krikke, J. Horing, and K. Smit, Mater.
11. F.N. Mazandarany and G.Y. Lai, Nucl. Perform., Aug 1976, p 9
Technol., Vol 43, 1979, p 349 29. J. Klower and U. Heubner, Carburization of
12. K. Natesan, Nucl. Technol., Vol 28, 1976, Nickel-Base Alloys and its Effects on the
p 441 Mechanical Properties, Mater. Corros., Vol
13. K. Natesan and T.F. Kassner, Metall. Trans., 49, 1998, p 237
Vol 4, 1973, p 2557 30. H.J. Grabke, Metallic Corrosion, Proc.
14. HSC, Chemistry for Windows, Version 6.0, Eighth Int. Cong. Metallic Corrosion, Vol
A. Roine, Outokumpu Technology, Finland, III, Mainz, Germany, Sept 6–11, 1981
www.outokumputechnology.com, accessed 31. I. Wolf and H.J. Grabke, Solid State Com-
Dec 2006 mun., Vol 54, 1985, p 5
15. ChemSage, Version 4.16, GTT-Technolo- 32. C.M. Schillmoller, Chem. Eng., Jan 6, 1986,
gies, Aachen, 1998 p 87
16. P.L. Hemmings and R.A. Perkins, “Thermo- 33. D.J. Hall, M.K. Hossain, and J.J. Jones,
dynamic Phase Stability Diagrams for Mater. Perform., Jan 1985, p 25
the Analysis of Corrosion Reactions in 34. J.A. Thuillier, Mater. Perform., Nov 1976,
Coal Gasification/Combustion Atmospheres,” p9
Report FP-539, EPRI, Palo Alto, CA, 1977 35. J.F. Norton and J. Barnes, in Corrosion in
17. R. Hultgren, P. Desai, D. Hawkins, M. Fossil Fuel Systems, I.G. Wright, Ed., The
Gleiser, and K.K. Kelley, Selected Values of Electrochemical Society, 1983, p 277
the Thermodynamic Properties of Elements 36. O. Van der Biest, J.M. Harrison, and J.F.
and Binary Alloys, ASM, 1973 Norton, in Behavior of High Temperature
18. L.C. Browning, T.W. Dewitt, and P.H. Alloys in Aggressive Environments, Proc.
Emmett, J. Am. Chem. Soc., Vol 72, 1950, International Conference (Patten, The
p 4211 Netherlands), Oct 15–18, 1979, The Metal
19. K.H. Jack, J. Iron Steel Inst., Vol 169, 1951, Society, London, 1980, p 681
p 26 37. J.M. Harrison, J.F. Norton, R.T. Derricott,
20. H.J. Christ, Experimental Characterization and J.B. Marriott, Werkst. Korros., Vol 30,
and Computer-Based Description of the 1979, p 785
Carburization Behaviour of the Austenitic 38. O. Demel, E. Keil, and P. Kostecki, SGAW
Stainless Steel AISI 304L, Mater. Corros., Report No. 2538, Osterreichische, Studieng-
Vol 49, 1998, p 258 esellschaft fur Atomenergie, Lenaugasse
21. R.G. Coltters, Mater. Sci. Eng., Vol 76, 10, A-1082 Wien, Forschungszentrum Sei-
1985, p 1 bersdorf, Institut fur Metallurgie
22. S.R. Shatynski, Oxid. Met., Vol 13 (No. 2), 39. R.P. Smith, Trans. Met. AIME, Vol 224,
1979, p 105 1962, p 10
23. W.F. Chu and A. Rahmel, Oxid. Met., Vol 40. H.J. Grabke, U. Gravenhorst, and W. Stein-
15, 1981, p 331 kusch, Werkst. Korros., Vol 27, 1976, p 291
24. Y. Nishiyama, N. Otsuka, and T. Nishizawa, 41. S.K. Bose and H.J. Grabke, Z. Metallkde.,
Carburization Resistance of Austenitic Vol 69, 1978, p 8
Alloys in CH4 -CO2 -H2 Gas Mixtures at 42. C. Steel and W. Engel, AFS Int. Cast Metals
Elevated Temperatures, Corrosion, Vol 59 J., Sept 1981, p 28
(No. 8), 2003, p 688 43. L.H. Wolfe, Laboratory Investigation of
25. H.J. Grabke, “Carburization: A High High Temperature Alloy Failure Mechan-
Temperature Corrosion Phenomenon,’’ MTI isms, Paper No. 12, Corrosion/77, NACE,
Publications No. 52, Materials Technology 1977
Institute of the Chemical Process Industries, 44. I.Y. Khandros, R.G. Bayer, and C.A. Smith,
St. Louis, Missouri, 1998 Paper No. 10, Corrosion/84, NACE, 1984
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:24PM Plate # 0 pg 144

144 / High-Temperature Corrosion and Materials Applications

45. W. Steinkrsch, Werkst. Korros., Vol 30, State Sci. Technol., Vol 132 (No. 9), 1985,
1979, p 837 p 2268
46. L.H. Wolfe, Mater. Perform., April 1978, 62. H.J. Grabke and D. Jakobi, High Tempera-
p 38 ture Corrosion of Cracking Tubes, Mater.
47. U. Van den Bruck and C.M. Schillmoller, Corros., Vol 53, 2002, p 494
Paper No. 23, Corrosion/85, NACE, 1985 63. V. Guttmann and H. Schonherr, Creep
48. R.H. Kane, Paper No. 266, Corrosion/83, Properties of Two 25Cr-20Ni Cast Alloys
NACE, 1983 in Air and under Carburizing Conditions,
49. J.F. Mason, J.J. Moran, and E.N. Skinner, High Temp. Technol., Vol 3 (No. 2), 1985,
Corrosion, Vol 16, 1960, p 593t p 79
50. R.H. Kane and J.C. Hosier, “Carburization 64. D. Jakobi and R. Gommans, Typical Failures
Resistance of Some Wrought Nickel- in Pyrolysis Coils for Ethylene Cracking,
Containing Alloys in Simulated Industrial Mater. Corros., Vol 54, 2003, p 881
Environments,” Inco Alloys International 65. N.G. Taylor, V. Guttmann, and R.C. Hurst,
Technical Report, Inco Alloys International, The Creep Ductility and Fracture of Car-
Huntington, WV, 1985 burized Alloy 800H at High Temperatures,
51. D.R.G. Mitchell, D.J. Young, and W. in High Temperature Alloys: Their Exploi-
Kleemann, Carburization of Heat-Resistant table Potential, J.B. Marriott, M. Merz,
Steels, Mater. Corros., Vol 49, 1998, p 231 J. Nihoul, and J. Ward, Ed., Elsevier Applied
52. G.Y. Lai, in High Temperature Corrosion in Science, 1987, p 475
Energy Systems, Proc. TMS-AIME Sympo- 66. R.F. Hochman, Proc. Fourth International
sium, M.F. Rothman, Ed., The Metallurgical Congress on Metal Corrosion, NACE, 1972,
Society of AIME, 1985, p 551 p 258
53. G.Y. Lai, B. Li, B. Gleeson, and H.L. 67. R.F. Hochman, Proc. Symposium on
Craig, Proposed Standard Carburization Properties of High Temperature Alloys with
Test Method, Paper No. 3473, Corrosion/ Emphasis on Environmental Effects, Z.A.
2003, NACE International, 2003 Foroulis and F.S. Pettit, Ed., The Electro-
54. R.H. Kane, G.M. McColvin, T.J. Kelly, and chemical Society, 1977, p 715
J.M. Davison, Paper No. 12, Corrosion/84, 68. J.C. Nava Paz and H.J. Grabke, Metal
NACE, 1984 Dusting, Oxid. Met., Vol 39, 1993, p 437
55. R. Kirchheiner, D.J. Young, P. Becker, and 69. H.J. Grabke, R. Krajak, E.M. Muller-
R.N. Durham, Improved Oxidation and Lorenz, Metal Dusting of High Temperature
Coking Resistance of a New Alumina Alloys, Werkst. Korros., Vol 44, 1993, p 89
Forming Alloy 60HT for the Petrochemical 70. H.J. Grabke, C.B. Bracho-Troconis, and
Industry, Paper No. 5428, Corrosion/2005, E.M. Muller-Lorenz, Metal Dusting of Low
2005 Alloy Steels, Werkst. Korros., Vol 45, 1994,
56. F. Pons and M. Hugo, Paper No. 272, p 215
Corrosion/81, NACE, 1981 71. H.J. Grabke, Metal Dusting of Low- and
57. F.N. Mazandarany and G.Y. Lai, High-Alloy Steels, Corrosion, Vol 51, 1995,
“Corrosion Behavior of Selected Structural p 711
Materials in a Simulated Steam-Cycle 72. H.J. Grabke, R. Krajak, E.M. Muller-
HTGR Helium Environment,” GA-A14446, Lorenz, and S. Straub, Metal Dusting of
General Atomic Company, San Diego, CA, Nickel-Base Alloys, Mater. Corros., Vol 47,
Oct 1977 1996, p 495
58. T.A. Ramanarayanan, R.A. Petkovic, J.D. 73. H.J. Grabke, Thermodynamics, Mechan-
Mumford, and A. Ozekcin, Carburization isms and Kinetics of Metal Dusting, Mater.
of High Chromium Alloys, Mater. Corros., Corros., Vol 49, 1998, p 303
Vol 49, 1998, p 226 74. H.J. Grabke, Corrosion by Carbonaceous
59. H.J. Grabke, R. Moller, and A. Schnaas, Gases, Carburization and Metal Dusting,
Werkst. Korr., Vol 30, 1979, p 794 and Methods of Prevention, Mater. High
60. H.J. Grabke, E.M. Peterson, and S.R. Temp., Vol 17 (No. 4), 2000, p 483
Srinivasan, Surf. Sci., Vol 67, 1977, p 501 75. H.J. Grabke, Metal Dusting, Mater. Corros.,
61. T.A. Ramanarayanan and D.J. Srolovitz, Vol 54, 2003, p 736
Carburization Mechanisms of High Chro- 76. H.J. Grabke and G. Tauber, Arch. Eisen-
mium Alloys, J. Electrochem. Soc.: Solid- hüttenwes., Vol 46, 1975, p 215
Name ///sr-nova/Dclabs_wip/High Temp/5208_97-145.pdf/Chap_05/ 26/10/2007 12:24PM Plate # 0 pg 145

Chapter 5: Carburization and Metal Dusting / 145

77. S.R. Shatynski and H.J. Grabke, Arch. Alloys, Paper No. 532, Corrosion/2000,
Eisenhüttenwes., Vol 49, 1978, p 129 NACE International, 2000
78. F. Bonnet, F. Ropital, Y. Berthier, and 93. S. Strauss and H.J. Grabke, Mater. Corros.,
P. Marcus, Filamentous Carbon Formation Vol 49, 1998, p 321
Caused by Catalytic Metal Particles from 94. J. Klower, H.J. Grabke, E.M. Muller-
Iron Oxide, Mater. Corros., Vol 54, 2003, Lorenz, and D.C. Agarwal, Metal Dusting
p 870 and Carburization Resistance of Nickel-
79. E. Pippel, J. Woltersdorf, and H.J. Grabke, Base Alloys, Paper No. 139, Corrosion/97,
Microprocesses of Metal Dusting on Iron- NACE International, 1997
Nickel Alloys and Their Dependence on the 95. K. Natesan and Z. Zeng, Metal Dusting
Alloy Composition, Mater. Corros., Vol 54, Performance of Structural Alloys, Paper
2003, p 747 No. 5409, Corrosion/2005, NACE Inter-
80. F. Eberle and R.D. Wylie, Corrosion, Vol 15 national, 2005
(No. 12), 1959, p 622t 96. D.L. Klarstrom and H.J. Grabke, The Metal
81. W.B. Hoyt and R.H. Caughey, Corrosion, Dusting Behavior of Several High Tem-
Vol 15 (No. 12), 1959, p 627t perature Alloys, Paper No. 1379, Corrosion/
82. O.J. Dunmore, A Case History of a Metal 2001, NACE International, 2001
Dusting Problem Which Led to a Boiler 97. B.A. Baker and G.D. Smith, Metal Dusting
Failure, presented at UK Corrosion/82 Behavior of High-Temperature Alloys,
(London), Nov 16–18, 1982 Paper No. 54, Corrosion/99, NACE Inter-
83. H.J. Grabke and M. Spiegel, Mater. Corros., national, 1999
Vol 54, 2003, p 799 98. B.A. Baker and G.D. Smith, Alloy Selection
84. G.Y. Lai, J. Met., Vol 37 (No. 7), 1985, for Environments Which Promote Metal
p 14 Dusting, Paper No. 257, Corrosion/2000,
85. E. Fitzer, W. Frohs, and M. Heine, Carbon, NACE International, 2000
Vol 24 (No. 4), 1986, p 387 99. B.A. Baker, G.D. Smith, V.W. Hartmann, L.
86. M. Maier and J.F. Norton, Studies Con- E. Shoemaker, and S.A. McCoy, Nickel-
cerned with the Metal Dusting of Fe-Cr-Ni Base Material Solutions to Metal Dusting
Materials, Paper No. 75, Corrosion/99, Problems, Paper No. 2394, Corrosion/2002,
NACE International, 1999 NACE International, 2002
87. R.A. Perkins, R.A. Padgett, and N.K. Tunali, 100. H.J. Grabke, H.P. Martinz, and E.M.
Met. Trans. AIME, Vol 4, 1973, p 2535 Muller-Lorenz, Metal Dusting Resistance
88. P.J. Albery and C.W. Haworth, Met. Sci., of High Chromium Alloys, Mater. Corros.,
Vol 8, 1974, p 407 Vol 54, 2003, p 860
89. A.F. Smith, Met. Sci., Vol 9, 1975, p 375, 101. H.J. Grabke, E.M. Muller-Lorenz, and M.
425 Zinke, Metal Dusting Behaviour of Welded
90. C.H. Toh, P.R. Munroe, and D.J. Young, Ni-Base Alloys with Different Surface
Metal Dusting of Fe-Cr and Fe-Ni-Cr Alloys Finish, Mater. Corros., Vol 34, 2003, p 785
under Cyclic Conditions, Oxid. Met., Vol 58 102. H.J. Grabke and E.M. Muller-Lorenz, Steel
(No. 1/2), 2002, p 1 Res., Vol 66, 1995, p 252
91. H.J. Grabke, E.M. Muller-Lorenz, S. 103. A. Schneider, H. Viefhaus, G. Inden, H.J.
Strauss, E. Pippel, and J. Woltersdorf, Grabke, and E.M. Muller-Lorenz, Mater.
Effects of Grain Size, Cold Working, and Corros., Vol 49, 1998, p 330
Surface Finish on the Metal-Dusting Resis- 104. A. Schneider, H. Viefhaus, and G. Inden,
tance of Steels, Oxid. Met., Vol 50 (No 3/4), Surface Analytical Studies of Metal
1998, p 241 Dusting of Iron in CH4-H2-H2S Mixtures,
92. A.S. Fabiszewski, W.R. Warkins, J.J. Mater. Corros., Vol 51, 2000, p 338
Hoffman, and S.W. Dean, The Effect of 105. A. Schneider and H.J. Grabke, Effect of
Temperature and Gas Composition on the H2S on Metal Dusting of Iron, Mater.
Metal Dusting Susceptibility of Various Corros., Vol 54, 2003, p 793
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:25PM Plate # 0 pg 147

High-Temperature Corrosion And Materials Applications Copyright © 2007 ASM International®

George Y. Lai, editor, p147-200 All rights reserved.
DOI: 10.1361/hcma2007p147 www.asminternational.org

CHAPTER 6

Corrosion by Halogen and

Hydrogen Halides

6.1 Introduction refining operation in the production of uranium,

UO2 is fluorinated at elevated temperatures
Many metals react readily with halogen gases (e.g., 500 to 600 °C) with HF to produce UF4 or
at elevated temperatures to form volatile metal UF6 for separation of U235 (Ref 6). The mat-
halides. Some metal halides also exhibit low erials of construction for this processing equip-
melting points, and some even sublime at rela- ment must resist corrosion by HF at both high
tively low temperatures. As a result, alloys and low temperatures.
containing elements that form volatile or low- This chapter reviews data on primarily gas-
melting-point halides may suffer severe high- eous corrosion by halogen and hydrogen halides.
temperature corrosion. In many high-temperature industrial processes
Industrial environments often contain halogen where fuels and/or feedstocks are often con-
gases. Because of high vapor pressures of many taminated with impurities, such as alkaline
metal chlorides, the chlorination process is an metals, halogen may react readily with these
important step in processing metallurgical ores metals to form halide salts. Corrosion reactions
for production of titanium, zirconium, tantalum, under these conditions are to be discussed in later
niobium, and tungsten (Ref 1–3). Chlorination is chapters dealing with high-temperature corro-
also used for extraction of nickel from iron sion in gas turbines, coal-fired boilers, oil-fired
laterites (Ref 4) and for detinning of tin plate boilers, waste-to-energy boilers, black liquor
(Ref 5). Production of TiO2 and SiO2 involves recovery boilers, and so forth.
processing environments containing Cl2 and/or
HCl, along with O2 and other combustion pro-
ducts. Calcining operations in the production
of (a) lanthanum, cerium, and neodymium for 6.2 Thermodynamic Considerations
electronic and magnetic materials and (b) cera-
mic ferrites for permanent magnets, frequently The relative stabilities of various chlorides,
generate environments contaminated with fluorides, bromides, and iodides are presented in
chlorine. In the chemical process industry, many Fig. 6.1 to 6.3, in terms of standard free energies
processing streams also contain chlorine. Manu- of formation (∆G°) versus temperature (Ref 7).
facturing of ethylene dichloride (EDC), which is The figures also include the information about
an intermediate for the production of vinyl the melting points and boiling points of halides.
chloride monomer, generates chlorine-bearing Some of the halides exhibit low melting and
environments. The reactor vessels, calciners, boiling points. Tables 6.1 to 6.4 list melting
and other processing equipment for the above and boiling points of some important halides and
operations require alloys that are resistant to oxyhalides (Ref 8, 9). Since industrial environ-
high-temperature chloridation attack. ments typically have finite oxygen activities, the
In the manufacture of fluorine-containing thermodynamic phase stability diagram in terms
compounds, such as fluorocarbon plastics, re- of the M-O-Cl (M represents metal, O represents
frigerants, and fire-extinguishing agents, the oxygen, and Cl represents chlorine) system is
processing equipment requires alloys with good quite useful in describing the possible corrosion
resistance to corrosion by fluorine and hydrogen products that may form on the metal. These
fluorides at elevated temperatures. During the diagrams can be constructed for major alloying
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:25PM Plate # 0 pg 148

148 / High-Temperature Corrosion and Materials Applications

elements of high-temperature alloys (e.g., Fe, diagram defines the boundary ( pO2  10715 atm)
Ni, Co, Cr, Mo, W, Al, Si, etc.). Commercial between Ni and NiO, the boundary (pCl2  1078 )
computer programs, such as HSC (Ref 10) and between Ni and NiCl2, and the boundary be-
Chemsage (Ref 11), are available for construc- tween NiO and NiCl2. This means that if the
tion of these phase-stability diagrams. Figure 6.4 equilibrium gas mixture of the environment
shows a Ni-O-Cl stability diagram in terms of exhibits an oxygen potential ( pO2 ) higher than
pO2 and pCl2 at 723 °C (1333 °F) (Ref 12). The about 10−15 atm, NiO may form. Similarly,

Fig. 6.1 Standard free energies of formation for chlorides. Source: Ref 7
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:25PM Plate # 0 pg 149

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 149

NiCl2 may form if the environment exhibits a metal. Let’s consider an example where nickel
a chlorine potential ( pCl2 ) higher than about is exposed to an environment consisting of air
10−8 atm when the oxygen potential ( pO2 ) is with 0.1% Cl2 at 723 °C (1333 °F). The envir-
below about 10−15 atm. onment would be at the location that identifies
With phase-stability diagrams, one can predict pO2 being very close to 100 and pCl2 being 10−3 in
the possible phases that are likely to form on the 723 °C Ni-O-Cl stability diagram (Fig. 6.4).

Fig. 6.2 Standard free energies of formation for fluorides. Source: Ref 7
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:25PM Plate # 0 pg 150

150 / High-Temperature Corrosion and Materials Applications

Based on the stability diagram shown in Fig. 6.4, be below the value needed to form NiCl2, thus
the environment is in the NiO regime. NiO preventing the formation of NiCl2. However,
oxide is expected to form on the surface of when defects and cracks develop in the NiO
nickel when exposed to this environment. If NiO scale, Cl2 can permeate through the oxide scale
oxide scale formed on nickel is defect free, the and reach the nickel with high enough pCl2 to
pCl2 at the interface between NiO and Ni would form NiCl2, initiating chloridation attack. The

Fig. 6.3 Standard free energies of formation for bromides and iodides. Source: Ref 7
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:25PM Plate # 0 pg 151

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 151

NiCl2 formed at 723 °C (1333 °F) would be Table 6.2 Melting points, temperatures at which
a solid phase. As shown in Table 6.1, NiCl2 fluoride vapor pressure reaches 10−4 atm, and
melts at relatively high temperature (1030 °C) boiling points of various fluorides
compared with other metal chlorides, such as Melting Temperature Boiling
Fluorides point, °C (°F) at 10−4 atm, °C (°F) point, °C (°F)
CoCl2 (740 °C), FeCl2 (676 °C), and FeCl3
FeF2 1020 (1868) 906 (1663) 1800 (3272)
(303 °C). FeF3 1027 (1881) 673 (1243) …
The phase-stability diagram for Co-O-Cl at NiF2 1450 (2642) 939 (1722) …
723 °C (1333 °F) is shown in Fig. 6.5 (Ref 12), CoF2 1250 (2282) 962 (1764) …
CrF2 894 (1641) 928 (1702) …
and those for Cr-O-Cl and Fe-O-Cl at 600 °C CrF3 1404 (2559) 855 (1571) …
(1112 °F) are shown in Fig. 6.6 and 6.7, CuF 908 (1666) … …
MoF5 64 (147) 24 (75) …
respectively (Ref 13). In addition to low melting MoF6 17 (63) −82 (−116) 36 (97)
points, some chlorides have low boiling points WF6 2 (36) −91 (−132) 17 (63)
also. In Fig. 6.7, FeCl3 is in a gaseous state at TiF3 1200 (2192) … …
TiF4 … 108 (226) 283 (541)
600 °C (1112 °F). Figure 6.8 shows WCl4 and AlF3 … 825 (1517) 1270 (2318)
WO2Cl2 in a gaseous state at 900 °C (1650 °F). SiF4 −90 (−130) −160 (−256) −95 (−139)
Some fluorides, bromides, and iodides also have MnF2 920 (1688) 992 (1818) …
ZrF4 932 (1710) 583 (1081) …
either low melting points or low boiling points NbF5 79 (174) … 233 (451)
(see Tables 6.2–6.4). Corrosion reactions can be HfF4 … 615 (1139) …
TaF5 97 (207) 37 (99) …
significantly increased when the corrosion pro- NaF 992 (1818) 928 (1702) 1704 (3099)
duct is in either a liquid state or a gaseous state. KF 857 (1575) 788 (1450) 1502 (2736)
Furthermore, many halides, although in a solid LiF 848 (1558) 908 (1666) 1681 (3058)
MgF2 1263 (2305) 1257 (2295) 2230 (4046)
state, may exhibit high vapor pressures. When CaF2 1418 (2584) 1429 (2604) 2500 (4532)
the corrosion product in a solid state exhibits a BaF2 1290 (2354) 1581 (2878) 2215 (4019)
ZnF2 875 (1607) 806 (1483) 1500 (2732)
PbF2 822 (1512) 664 (1227) 1293 (2359)
Source: Ref 8, Ref 9

Table 6.1 Melting points, temperatures at

which chloride vapor pressure reaches 10−4 atm,
and boiling points of various chlorides
Melting Temperature Boiling
Chlorides point, °C (°F) at 10 −4 atm, °C (°F) point, °C (°F) Table 6.3 Melting points, temperatures at which
FeCl2 676 (1249) 536 (997) 1026 (1879) bromide vapor pressure reaches 10−4 atm, and
FeCl3 303 (577) 167 (333) 319 (606) boiling points of various bromides
NiCl2 1030 (1886) 607 (1125) 987 (1809)
CoCl2 740 (1364) 587 (1089) 1025 (1877) Melting Temperature Boiling
CrCl2 820 (1508) 741 (1366) 1300 (2372) Bromides point, °C (°F) at 10−4 atm, °C (°F) point, °C (°F)
CrCl3 1150 (2102) 611 (1132) 945 (1733) FeBr2 689 (1272) 509 (948) 974 (1785)
CrO2Cl2 −95 (−139) … 117 (243) FeBr3 … 156 (313) …
CuCl 430 (806) 387 (729) 1690 (3074) NiBr2 965 (1769) 580 (1076) 919 (1686)
MoCl5 194 (381) 58 (136) 268 (514) CoBr2 678 (1252) … …
WCl5 240 (464) 72 (162) … CrBr2 842 (1548) 716 (1321) …
WCl6 280 (536) 11 (52) 337 (639) CrBr3 >800 (1472) 615 (1139) …
TiCl2 1025 (1877) 921 (1690) … CrBr4 … 516 (961) …
TiCl3 730 (1346) 454 (849) 750 (1382) CuBr 488 (910) 435 (815) 1318 (2404)
TiCl4 −23 (−9.4) −38 (−36) 137 (279) WBr5 276 (529) … …
AlCl3 193 (379) 76 (169) … WBr6 309 (588) … …
SiCl4 −70 (−94) −87 (−125) 58 (136) AlBr3 97 (207) 53 (127) 255 (491)
MnCl2 652 (1206) 607 (1125) 1190 (2174) SiBr4 5 (41) … 153 (307)
ZrCl4 483 (901) 146 (295) … MnBr2 695 (1283) … …
NbCl5 205 (401) … 250 (482) ZrBr4 450 (842) 169 (336) …
NbCl4 … 239 (462) 455 (851) NbBr5 267 (513) … 361 (682)
TaCl5 216 (421) 80 (176) 240 (464) HfBr4 424 (795) 137 (279) …
HfCl4 434 (813) 132 (297) … TiBr4 … … 232 (450)
CCl4 −24 (−11) −80 (−112) 77 (171) TaBr5 267 (513) 145 (293) 347 (657)
NaCl 801 (1474) 742 (1368) 1465 (2669) NaBr 750 (1382) 690 (1274) 1393 (2539)
KCl 772 (1422) 706 (1303) 1407 (2565) KBr 740 (1364) 671 (1240) 1383 (2521)
LiCl 610 (1130) 665 (1229) 1382 (2520) LiBr 550 (1022) 630 (1166) 1310 (2390)
MgCl2 714 (1317) 663 (1225) 1418 (2584) MgBr2 710 (1310) 626 (1159) 1230 (2246)
CaCl2 772 (1422) 1039 (1902) 2000 (3632) CaBr2 742 (1368) … 1800 (3272)
BaCl2 962 (1764) … 1830 (3326) BaBr2 854 (1569) … …
ZnCl2 318 (604) 349 (660) 732 (1350) ZnBr2 398 (748) 320 (608) 650 (1202)
PbCl2 498 (928) 484 (903) 954 (1749) PbBr2 373 (703) 432 (810) 914 (1677)
Source: Ref 8, Ref 9 Source: Ref 8, Ref 9
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:25PM Plate # 0 pg 152

152 / High-Temperature Corrosion and Materials Applications

high vapor pressure, the corrosion reaction can

increase as well. It is generally considered that
when the vapor pressure of a corrosion product
reaches to 10−4 atm, the corrosion reaction can
become significant. Tables 6.1 to 6.4 also include
the temperature at which a halide’s vapor pres-
sure reaches to 10−4 atm.
The chlorine partial pressure (pCl2 ) needed to
form a chloride corrosion product with 10−4 atm
will be lower than that needed to form either
solid or liquid chloride. This is illustrated in
Fig. 6.9 (Ref 14). Bender and Schutze (Ref 15)

Table 6.4 Melting points, temperatures at Fig. 6.5 Phase stability diagram for Co-O-Cl system at 723 °C
which iodide vapor pressure reaches 10−4 atm, (1333 °F). All the corrosion products (i.e., CoO,
Co3O4, and CoCl2) are solid phases at this temperature. Source:
and boiling points of various iodides Ref 12
Melting Temperature Boiling
Iodides point, °C (°F) at 10−4 atm, °C (°F) point, °C (°F)
FeI2 594 (1101) 476 (889) 935 (1715)
NiI2 780 (1436) … …
CoI2 515 (959) … …
CrI2 869 (1596) 702 (1296) …
CrI3 >600 (1112) … …
CuI 588 (1090) 529 (984) 1207 (2205)
AlI3 191 (376) 144 (291) 385 (725)
SiI4 122 (252) 55 (131) 301 (574)
MnI2 613 (1135) … …
ZrI4 499 (930) 227 (441) …
NbI4 503 (937) … …
HfI4 449 (840) 244 (471) …
TaI5 496 (925) 208 (406) 545 (1013)
NaI 660 (1220) 651 (1204) 1304 (2379)
KI 685 (1265) 629 (1164) 1330 (2426)
LiI 469 (876) 621 (1150) 1170 (2138)
MgI2 650 (1202) 425 (797) …
CaI2 740 (1364) … …
BaI2 712 (1314) … …
ZnI2 446 (835) 316 (601) 730 (1346)
PbI2 412 (774) 397 (747) 872 (1602) Fig. 6.6 Phase stability diagram for Cr-O-Cl system at 600 °C
(1112 °F). All the corrosion products (i.e., Cr2O3,
Source: Ref 8, Ref 9 CrCl2, and CrCl3) are solid phases at this temperature. Source:
Ref 13

Fig. 6.4 Phase stability diagram for Ni-O-Cl system at 723 °C Fig. 6.7 Phase stability diagram for Fe-O-Cl system at 600 °C
(1333 °F). Both corrosion products (NiO and NiCl2) (1112 °F). All the corrosion products are solid phases
are solid phases at this temperature. Source: Ref 12 except FeCl3 at this temperature. Source: Ref 13
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:26PM Plate # 0 pg 153

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 153

Fig. 6.8 Phase stability diagram for the W-O-Cl system at 900 °C (1650 °F), showing tungsten chloride (WCl4) and tungsten
oxychloride (WO2Cl2) in a gaseous state.

and higher, NiCl2 with vapor pressures of 10−4

atm and higher (not NiO) is to form on nickel.
Molybdenum oxychlorides also exhibit high
vapor pressures. Figure 6.11 shows the quasi-
stability diagram for Mo-O-Cl system at 500 °C
(Ref 15). The figure also shows that MoO2Cl2
with vapor pressures of 10−4 atm and higher is to
form in the environments of air + 0.1% Cl2, and
air + 2% Cl2.
The metal-halogen reaction differs from
other reactions, such as oxidation, in that most
reaction products are characteristic of high vapor
pressures and, in some cases, low melting points.
The volatile halides (reaction products) formed
on the metal surface can no longer provide
protection against further corrosion. This is in
contrast to most oxides, which generally exhibit
very low vapor pressures and high melting
points. Furthermore, many halides exhibit low
Fig. 6.9 Phase stability diagram for Fe-O-Cl system at 700 °C melting points. Once the reaction products
(1292 °F). The solid line represents the boundary for
forming solid FeCl2, while the dotted line represents the boundary become molten, the alloy loses all protec-
for forming FeCl2 with 10−4 atm pressure. Source: Ref 14 tion against further corrosion, leading to rapid
attack.

extensively examined the phase-stability dia-

grams involving the vapor pressures of chlorides
reaching 10−4 atm for various alloying elements. 6.3 Corrosion in Cl2- and HCl-Bearing
The authors termed these phase-stability dia- Environments
grams “quasi-stability diagrams.” Figure 6.10
shows the pCl2 values that are needed to form
6.3.1 Corrosion in Cl2 Environments
NiCl2 with 10−4 atm at various temperatures
(Ref 15). Also included in the figure are two (No O2)
environments (air + 0.1% Cl2, and air + 2% Cl2) This section focuses on the corrosion behavior
for illustration purpose. For both of these of alloys in environments containing essentially
environments, NiO, not NiCl2 with vapor Cl2 with no oxygen (O2) present. An excellent
pressures of 10−4 atm or higher, is likely to form review of halogen corrosion data up to the
on nickel at 500 and 650 °C. However, at 800 °C mid-1970s was presented by Daniel and Rapp
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:26PM Plate # 0 pg 154

154 / High-Temperature Corrosion and Materials Applications

(Ref 8). They summarized the test results Above 250 °C, corrosion rates abruptly in-
obtained by various authors (Ref 16–19) on creased. Iron forms two types of chlorides:
chloridation of iron (Table 6.5). It should be FeCl2 and FeCl3. The melting and boiling
noted that the data in Table 6.5 were obtained points of FeCl2 are 676 and 1026 °C (1249 and
from short-duration tests, ranging from a few 1879 °F), respectively. FeCl3, on the other hand,
minutes to hours. Use of these test results for is extremely unstable. Its melting and boiling
extrapolation to 1 year could result in significant points are 303 and 319 °C (577 and 606 °F),
errors. They should not be used to estimate the respectively. Bohlken et al. (Ref 20, 21) sug-
service life of equipment, but should instead be gested that the abrupt increase in the corro-
used for comparison purposes. As illustrated in sion rate of iron in Cl2 at temperatures above
Table 6.5, iron exhibited little corrosion attack 250 °C (482 °F) was related to the formation
in Cl2 at temperatures up to 250 °C (480 °F). of FeCl3.

0
p(NiCl2) ≥ 10–4 bar
500 °C
–2

–4 650 °C
Log p(Cl2), bar

NiCl2
–6 850 °C

1000 °C
–8

–10
NiO

–12
Ni
–14
–30 –25 –20 –15 –10 –5 0
Log p(O2), bar

Fig. 6.10 Quasi-stability diagram for Ni-O-Cl system for NiCl2 with vapor pressures of 10−4 atm (bar) and higher at temperatures from
500 to 1000 °C (932 to 1832 °F). Source: Ref 15

0
p(MoOxCly) ≥ 10–4 bar
MoOCl4
–2

–4 MoOCl3
Log p(Cl2), bar

–6
MoCl4
–8 MoO2Cl2

–10

Mo MoO2 MoO3
–12

–40 –35 –30 –25 –20 –15 –10 –5 0

Log p(O2), bar

Fig. 6.11 Quasi-stability diagram for Mo-O-Cl system for vapor pressures of chlorides and oxychlorides being 10−4 atm (bar) and
higher at 800 °C (1472 °F). Source: Ref 15
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:26PM Plate # 0 pg 155

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 155

Adding chromium and/or nickel to iron and small amounts of FeCl3 and FeCl2 formed at
improves the alloy’s resistance to chloridation the “breakaway” stage. Also, at the breakaway
attack. Ferritic and austenitic stainless steels can stage, the specimen showed aluminum depletion
resist chloridation attack at higher temperatures at the metal/scale interface. The thin aluminum
than cast iron and carbon steels. Brown et al. oxide scale was observed to form on the speci-
(Ref 22) reported a corrosion rate of about men during heating to the test temperature with
600 mpy at 232 °C (450 °F) for both carbon argon gas flowing through the test chamber
steel and cast iron. A ferritic stainless steel (approximately 2 h) prior to switching to the test
(Fe-17Cr) showed a corrosion rate of about gas. The test gas (i.e., Ar-1%Cl2) was found to
79 mpy at 360 °C (680 °F), and a titanium- contain 1 ppm O2. Thus, under the test condition,
stabilized austenitic stainless steel showed a rate Al2O3 could form on the metal as well as FeCl2,
of 24 mpy at 418 °C (784 °F) (Ref 8). The as shown in Fig. 6.13.
results of studies on several stainless steels by Nickel and nickel-base alloys are widely used
Tseitlin and Strunkin (Ref 16) and Brown et al. in chlorine-bearing environments. The corrosion
(Ref 22) were summarized by Daniel and Rapp behavior of nickel in chlorine at various tem-
(Ref 8) and are presented in Table 6.6. peratures was analyzed by Daniel and Rapp (Ref
Adequate aluminum when added to iron to 8), using the test results of Downey et al. (Ref
form aluminum oxide scales is also beneficial in 24), Tseitlin and Strunkin (Ref 16), and McKin-
improving the chloridation resistance. Han and ley and Shuler (Ref 25) (see Table 6.7). At tem-
Cho (Ref 23) studied corrosion behavior of peratures up to 500 °C (930 °F), nickel showed
Fe3Al (Fe-12.11%Al) in Ar-1%Cl2 at 750, 800, relatively low corrosion rates. Corrosion rates
and 900 °C using a thermogravimetric method. became suddenly and significantly higher at
The alloy behaved similarly at three different temperatures over 500 °C (930 °F).
temperatures, showing an initial stage of an Nickel reacts with chlorine to form NiCl2,
“incubation” time before the breakaway corro- which exhibits relatively high melting point
sion showing a drastic weight loss, as shown in (1030 °C) compared to FeCl2 and FeCl3 (676
Fig. 6.12. The authors observed a thin protective and 303 °C, respectively). This may be an im-
Al2O3 scale during the initial “incubation” stage, portant factor, making nickel much more resis-
and nonprotective oxide scales (Al2O3, Fe2O3) tant to chloridation attack than iron. Brown et al.

Table 6.5 Corrosion of iron in chlorine

Temperature, Flow rate, pCl2, Diluent Test duration, Linear rate constant(a), Corrosion rate(b),
°C (°F) (L/min) atm gas min μm/min mm/yr (mpy)
−4
77 (170) 100 1 None 480 3 × 10 0.16 (6.3)
166 (330) 100 1 None 0–15, 15–480 3.3 × 10−3, 3.8 × 10−4 0.20 (7.9)
198 (388) 100 1 None 0–15, 15–480 5.2 × 10−3, 3.7 × 10−4 0.19 (7.5)
200 (392) 15 1 None 360 2 × 10−4 0.11 (4.3)
230 (446) 100 1 None 0–15, 15–480 8.5 × 10−3, 2.2 × 10−4 0.12 (4.7)
240 (464) 100 1 None 0–15, 15–480 9.5 × 10−3, 2.4 × 10−4 0.13 (5.1)
240 (464) 15 1 None 360 2 × 10-4 0.11 (4.3)
247 (477) 100 I None 0–15, 15–480 1 × 10−2, 1.9 ×10−4 0.10 (3.9)
251 (484) 100 1 None 480 1.55 820 (32 in.)
255 (491) 120 (c) Ar … 0.94 490 (19 in.)
−4
260 (500) 15 1 None 360 2 × 10 0.11 (4.3 mils)
268 (514) 120 (c) He … 1.78 940 (37 in.)
279 (534) 120 (c) Ar … 2.45 1,300 (51 in.)
−4
285 (545) 15 1 None 360 4 × 10 0.21 (8 mils)
285 (545) 15 1 None 60 20.4 11,000 (433 in.)
302 (576) 120 (c) He … 4.04 2,100 (83 in.)
304 (579) 120 (c) Ar … 4.17 2,200 (87 in.)
310 (590) 15 1 None 60 3.9 1,700 (67 in.)
323 (613) 120 (c) Ar … 8.94 4,700 (185 in.)
327 (621) 120 (c) He … 11.4 6,000 (236 in.)
381 (718) 120 (c) Ar … 40.4 21,000 (827 in.)
381 (718) 120 (c) He … 65.3 34,000 (1,338 in.)
540 (1004) 15 (c) N2 … 6.6 3,500 (138 in.)
540 (1004) 15 (c) N2 … 1.5 830 (33 in.)
595 (1103) 120 (c) Ar … 187 98,000 (3,858 in.)
599 (1110) 120 (c) He … 624 328,000 (13,000 in.)
(a) Rate constants are given as metal loss rates, μm/min. (b) Estimated metal loss after 1 yr of exposure. (c) Tests were conducted with chlorine partial pressures up to 0.3 atm
and total pressures of 1.0 atm. However, the metal loss rates were extrapolated to 1.0 atm chlorine pressure. Source: Ref 8
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:26PM Plate # 0 pg 156

156 / High-Temperature Corrosion and Materials Applications

Table 6.6 Corrosion of stainless steels in chlorine(a)

Linear rate
Temperature, Flow rate, constant(b), Corrosion rate(c),
Alloy °C (°F) L/min µm/min mm/yr (mpy)
Ferritic stainless 300 (572) 15 4 × 10−4 0.2 (7.9)
(Fe-17Cr) 360 (680) 15 3.8 × 10−3 2 (79)
440 (824) 15 6.7 × 10−2 40 (1.6 in.)
540 (1,004) 15 1.35 700 (28 in.)
Austenitic stainless 418 (784) 15 1.1 × 10−3 0.6 (24)
(Fe-18Cr-9Ni-Ti) 450 (842) 15 4.3 × 10−2 20 (787)
480 (896) 15 0.13 70 (2.8 in.)
535 (995) 15 0.47 200 (7.9 in.)
640 (1,184) 15 46 20,000 (787 in.)
Austenitic stainless 315 (599) 28 1.4 × 10−3 0.8 (31)
(Fe-18Cr-8Ni-Mo) 340 (644) 28 2.9 × 10−3 1.5 (59)
400 (752) 28 5.9 × 10−3 3 (118)
450 (842) 28 2.9 × 10−2 15 (590)
480 (896) 28 5.9 × 10−2 30 (1.2 in.)
Austenitic stainless 290 (554) 28 1.5 × 10−3 0.8 (31)
(Fe-18Cr-8Ni) 315 (599) 28 2.9 × 10−3 1.5 (59)
340 (644) 28 5.9 × 10−3 3 (118)
400 (752) 28 2.9 × 10−2 15 (590)
450 (842) 28 5.9 × 10−2 30 (1.2 in.)
(a) Chlorine pressure was approximately 1.0 atm. (b) Duration of these tests was 60–360 min for the first two alloys and 120–1200 min for the last two alloys. (c) Estimated
metal loss after one year of exposure. Source: Rcf 8

5 10
1% Cl2/Ar 750 °C
5 FeCl3
0
0
Weight change, mg/cm2

–5
–5 FeCl2
Log p Cl , g
2

–10 –10
AlCl3 Al2O3
–15 Fe2O3
–15
900 °C 800 °C 750 °C Fe3O4
–20 Fe
–20 Al
–25
FeO
–25 –30
0 10 20 30 40 50 60 70 –60 –50 –40 –30 –20 –10 0 10
Exposure time, h Log p O , g
2

Fig. 6.12 Thermogravimetric results for Fe3Al (Fe-12.11%Al)

tested in Ar-1%Cl2 at 750, 800, and 900 °C. Fig. 6.13 Phase stability for Al-O-Cl system at 750 °C. The
Source: Ref 23 solid circle indicates the test environment. Source:
Ref 23

(Ref 22) conducted short-term laboratory tests illustrated in Fig. 6.14 (Ref 26). This trend is
in chlorine on various commercial alloys. The also reflected in long-term tests (Table 6.9).
results (see Table 6.8) suggested that, in an Alloy 600 is the most commonly used alloy for
environment of 100% Cl2, carbon steel and high-temperature services in Cl2 environments.
cast iron are useful at temperatures up to 150 to Figure 6.15 shows the corrosion rates of alloy
200 °C (300 to 400 °F) only. The 18-8 stainless 600 in dry chlorine gas as a function of tem-
steels can be used at higher temperatures—up to perature (Ref 22). MTI Publication MS-3
320 to 430 °C (600 to 800 °F). Nickel and (Ref 27) suggests corrosion guidelines for Ni200,
nickel-base alloys (e.g., Ni-Cr-Fe, Ni-Mo, and alloy 600, alloy 400, Type 304, and steel in dry
Ni-Cr-Mo alloys) were most resistant. The chlorine gas applications, as shown in Fig. 6.16.
beneficial effect of nickel on the resistance Tu et al. (Ref 28) performed phase analysis
of chloridation attack in Cl2 environments is using x-ray diffraction on the external corrosion
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:26PM Plate # 0 pg 157

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 157

Table 6.7 Corrosion of nickel in chlorine

Temperature, pCl2, Test duration, Corrosion rate(a),
°C (°F) atm min mm/yr (mpy)
350 (662) 0.13 3000–5000 0.0012 (0.05)(b)
375 (707) 0.13 … 0.0014 (0.06)(b)
400 (752) 0.13 … 0.0036 (0.14)(b)
407 (765) 0.13 … 0.005 (0.2)(b)
433 (811) 0.13 … 0.0058 (0.2)(b)
465 (869) 0.13 … 0.107 (4.2)
485 (905) 0.13 … 0.142 (5.6)
500 (932) 0.13 … 0.319 (12.6)
500 (932) ~1 atm at 16 cm3/min 3600 0.492 (19.4)
525 (977) 0.13 … 2.41 (95)
540 (1004) ~1 atm at 16 cm3/min 3600 2.07 (82)
550 (1022) 0.13 … 6.02 (237)
550 (1022) 0.13 … 5.85 (230)
660 (1220) ~1 atm at 16 cm3/min 3600 40.3 (1,587)
740 (1364) ~1 atm at 16 cm3/min 60 120 (4,724)
770 (1418) ~1 atm at 16 cm3/min 60 2,150 (84,646)
(a) Estimated metal loss after 1 year of exposure to chlorine. (b) These estimates are probably low. Source: Ref 8

Table 6.8 Corrosion of selected alloys follow either a paralinear rate law (a combination
in chlorine of weight gain due to oxidation and weight loss
Approximate temperature, °C (°F), due to chlorination) or a linear rate law due to
at which given corrosion rate is exceeded chlorination. This is illustrated by the results of
0.8 mm/yr 1.5 mm/yr 3.0 mm/yr 15 mm/yr
Alloy (30 mpy) (60 mpy) (120 mpy) (600 mpy)
Maloney and McNallan (Ref 29) on corrosion
Nickel 510 (950) 538 (1000) 593 (1100) 650 (1200)
of cobalt in Ar-50O2-Cl2 mixtures (Fig. 6.17).
Alloy 600 510 (950) 538 (1000) 565 (1050) 650 (1200) As shown in the figure, at high Cl2 levels, the
Alloy B 510 (950) 538 (1000) 593 (1100) 650 (1200) corrosion products are primarily cobalt chloride
Alloy C 480 (900) 538 (1000) 565 (1050) 650 (1200)
Chromel A 425 (800) 480 (900) 538 (1000) 620 (1150) vapor, causing the weight loss to follow a linear
Alloy 400 400 (750) 455 (850) 480 (900) 538 (1000) rate law. Because of volatile corrosion products,
18-8 Mo 315 (600) 345 (650) 400 (750) 455 (850) the reaction rate can be highly dependent on the
18-8 288 (550) 315 (600) 345 (650) 400 (750)
Carbon steel 120 (250) 175 (350) 205 (400) 230 (450) gas flow rate. McNallan and Liang (Ref 30)
Cast iron 93 (200) 120 (250) 175 (350) 230 (450) showed that CoO specimens exhibited increased
Source: Ref 22 linear weight loss rates with increasing gas
velocity when exposed in the Ar-O2-1Cl2 mix-
tures with pO2 being 0.01 and 0.15 atm pressures
products formed on Ni-4Cr alloy when exposed at 723 °C (1000 K). Furthermore, the oxygen
in 105 Pa (1 atm) Cl2 at 575 and 700 °C. The partial pressure (pO2 ) of 0.01 atm resulted in
authors found that the scales consisted of mainly higher weight loss rates than that of 0.15 atm.
NiCl2, CrCl3, and CrCl2. The deposits on the This is also illustrated in Fig. 6.18 (Ref 31),
quartz test assembly during testing of Ni-4Cr showing the corrosion of cobalt in Ar-O2-1Cl2
alloy were also analyzed. These deposits were mixtures with three different concentrations
mainly NiCl2 and CrCl3 with very little CrCl2, (1, 10, and 50% O2) of oxygen at 650 °C
indicating both NiCl2 and CrCl3 are major vapor (1200 °F). When the environment contained 10
phases during testing Ni-4Cr alloy. and 50% O2, the corrosion reaction involved
mainly the formation of cobalt oxide, thus fol-
lowing an approximate parabolic rate. When the
6.3.2 Corrosion in O2-Cl2 Environments oxygen concentration reduced to 1%, the corro-
Many industrial environments may contain sion reaction involved mainly volatile CoCl2,
both chlorine and oxygen. Metals generally fol- thus following an approximate linear weight
low a parabolic rate law by forming condensed loss with time. The figure also shows that
phases of oxides, if the environment is free of the linear weight loss agreed very well with the
chlorine. With the presence of both oxygen and volatilization of CoCl2. It should be noted that in
chlorine, corrosion of metals then involves a the above test environments containing 10 and
combination of condensed oxides and volatile 50% O2, the corrosion reaction, which followed a
chlorides. Depending on the relative amounts of parabolic rate due to formation of condensed
oxides and chlorides formed, corrosion can cobalt oxides (Fig. 6.18), involved only a very
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:26PM Plate # 0 pg 158

158 / High-Temperature Corrosion and Materials Applications

mg/cm2

690

Fig. 6.14 Effect of nickel on the corrosion resistance of alloys in Ar-30Cl2 at 704 °C (1300 °F) for 24 h. Source: Ref 26

short-term test (2 h). It is extremely likely that

Table 6.9 Corrosion of several alloys upon longer exposure times the cobalt oxide
in Ar-30Cl2 after 500 h at 400 to 704 °C scales may eventually crack (or spall), thus
(750 to 1300 °F)(a)
allowing chlorine gas to reach the underlying
Descaled weight loss, mg/cm2 metal and causing formation of volatile CoCl2
400 °C 500 °C 600 °C 704 °C
Alloy (750 °F) (930 °F) (1110 °F) (1300 °F)(a)
and resulting in a linear corrosion rate.
Ni-201 0.2 0.3 47–101 97
Corrosion of pure nickel in O2-Cl2 environ-
600 0.02 5 127–180 160 ments was found to be dominated by the for-
601 0.3 3 85–200 215 mation of volatile NiCl2 corrosion product, and
625 0.7 7 … 180
617 0.6 7 … 190 the weight loss essentially followed a linear rate
800 6 13 200–270 890 law (Ref 32, 33). Figure 6.19 shows thermo-
310 28 370 … 820 gravimetric results at 927 °C in Ar-50O2 con-
304 108 1100 … >1000
347 215 Total … Total taining various amounts of Cl2 from 0 to 5%
(a) 24 h test period. Source: Ref 26 (Ref 32). The figure also shows that the experi-
mental data were in good agreement with the
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:26PM Plate # 0 pg 159

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 159

theoretical calculation based on vapor pressures

of NiCl2. In their testing of nickel in Ar-50O2-Cl2
containing 0, 0.5, 1.0, 1.75, 3, and 5% Cl2 at 627
to 927 °C, Lee and McNallan (Ref 32) observed
a very rapid reaction referred to as “ignition” at
727 and 827 °C in the environments containing
higher concentrations of Cl2. McNallan (Ref 33)
suggested that this rapid corrosion reaction was
caused by the reaction of chlorine with nickel
at the metal/oxide scale interface to form nickel
/month

chloride vapor, which then diffuses out and is

converted into powdery nickel oxide. This rapid
corrosion reaction causes metal temperature
to increase. He further indicated that this rapid
reaction (ignition) can be prevented at tempera-
tures below 727 °C due to low vapor pressures
of nickel chloride and can also be prevented at
temperatures higher than 827 °C due to form-
ation of a protective oxide scale by rapid oxida-
tion of nickel to reduce the ingress of chlorine
through the oxide scale and thus the formation of
nickel chloride vapor at the metal/oxide scale
interface.
In an O2-Cl2 containing environment, oxi-
dation and chloridation can take place. As dis-
cussed in section 6.2, a protective oxide scale can
lower the pCl2 below the value for forming metal
chlorides at the metal/oxide scale interface.
Fig. 6.15 Corrosion rate of alloy 600 in Cl2 as a function of However, cracking and spalling of this oxide
temperature. Source: Ref 22 scale resulting from, for example, thermal

Fig. 6.16 MTI corrosion guidelines for Ni200, alloy 600, alloy 400, Type 304SS and steel in dry chlorine (Cl2) as a function of
temperature. Source: Ref 27. Courtesy of Materials Technology Institute
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:26PM Plate # 0 pg 160

160 / High-Temperature Corrosion and Materials Applications

Fig. 6.17 Corrosion of cobalt in Ar-50O2-Cl2 at 927 °C (1700 °F). Source: Ref 29

Fig. 6.18 Thermogravimetric results of corrosion of cobalt in

Ar-O2-1Cl2 with 1, 10, and 50% O2 at 650 °C
(1200 °F). Source: Ref 31 Fig. 6.19 Thermogravimetric results of corrosion of nickel at
927 °C in Ar-50O2-Cl2 (0, 0.5, 1.0, 1.75, 3, and 5%
Cl2). Source: Ref 32
cycling can allow permeation of chlorine to reach
the metal underneath to initiate chloridation
attack. Figure 6.20 illustrates the effect of ther- runs, showing similar behavior in initiating an
mal cycling on the initiation of accelerated accelerated chloridation attack right after a ther-
chloridation attack for Fe-20Cr at 927 °C in mal cycle, which involved cooling the specimen
Ar-20O2-0.5Cl2 (Ref 34). The figure shows the for 30 min by lowering it from the test tem-
thermogravimetric results for three separate test perature (927 °C) to 100 °C in 5 min. As soon
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:27PM Plate # 0 pg 161

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 161

as the test was resumed after the thermal cycle, gravimetric results for representative alloys
specimens (three separate test runs) showed are summarized in Fig. 6.21 (Ref 36). Thermo-
accelerated chloridation attack. dynamic phase stability diagrams showing high
A significant amount of data have been gen- vapor pressures of oxychlorides of tungsten and
erated for commercial alloys in environments molybdenum were presented earlier in Fig. 6.8
containing both oxygen and chlorine. Table 6.10 and 6.11, respectively.
summarizes short-term test results generated In order to determine whether molybdenum
in Ar-20O2-2Cl2 at 900 °C (1650 °F) for 8 h oxychlorides would contribute to high corrosion
(Ref 35). The results revealed several interest- rates in high-molybdenum-containing nickel
ing trends. The best-performing alloy was an alloys, such as alloy S (Ni-16Cr-14.5Mo) in
aluminum-containing alloy 214 (Fe-16Cr-3Fe- O2-Cl2 environments, Jacobson et al. (Ref 37)
4.5Al) with a small amount iron. Two worst- used a high-pressure sampling mass spectro-
performing alloys were cobalt-base alloys (alloys meter to measure volatile species produced from
188 and 6B) containing large amounts of tung- the preoxidized specimen of alloy S with 14.5%
sten. Molybdenum-containing nickel-base alloys Mo in comparison with alloy 600 (Ni-16Cr-9Fe)
also did not perform well. Oh et al. (Ref 36) with no Mo during the exposure of Ar-50O2-
attributed this to the formation of oxychlorides 1Cl2. Thermogravimetric data for these two
of molybdenum and tungsten, which have very preoxidized alloys under the test condition are
high vapor pressures. The partial pressures of shown in Fig. 6.22, showing a significantly
WO2Cl2 and MoO2Cl2 in equilibrium with the higher weight loss rate for alloy S (14.5Mo)
oxides (WO3 and MoO3, respectively) and the than alloy 600 (no Mo) during the exposure of
test environment (Ar-20O2-2Cl2) at 900 °C the preoxidized specimens to Ar-50O2-1Cl2 at
(1650 °F) were 7.52 ×10−2 and 2.1× 100 atm, 900 °C (1650 °F). The mass spectrometer re-
respectively. Accordingly, alloy 188 (14% W), sults indicated that MoO2Cl2 along with NiCl2
alloy C-276 (16% Mo, 4% W), alloy 6B (4.5% and CrO2Cl2 were major vapor phases in the
W, 1.5% Mo), alloy X (9% Mo), and alloy S case of alloy S. For alloy 600, NiCl2 and CrO2Cl2
(14.5% Mo) suffered relatively high rates of were detected.
corrosion attack. Simple Fe-Ni-Cr (Type 310 SS) Alloy R-41 (nickel-base alloy with 1.5% Al,
and Ni-Cr-Fe (alloy 600) performed better than 3% Ti and 10% Mo) suffered less chloridation
molybdenum- or tungsten-containing alloys. The attack than other nickel-base alloys containing
molybdenum despite high molybdenum content
in a short-term test presented in Table 6.10.
100
However, the results of long-term tests in
Ar-20O2-0.25Cl2 by Rhee et al. (Ref 38) and
80
McNallan et al. (Ref 39) showed that these
nickel-base alloys with molybdenum, such
Mass change, mg/cm2

36 h as R-41 (Ni-19Cr-11Co-10Mo-1.5Al-3Ti) and

60
18 h alloy 263 (Ni-20Cr-20Co-5.8Mo-0.5Al-2.2Ti),
eventually suffered severe attack despite the
presence of aluminum and titanium. Figure 6.23
40
24 h

20 Table 6.10 Corrosion of selected alloys in

Ar-20O2-2Cl2 at 900 °C (1650 °F) for 8 h
Pure oxygen
Average metal
0 Metal loss, affected(a),
Alloy mm (mils) mm (mils)

4 8 12 16 20 24 28 32 36
214 0 0.012 (0.48)
R-41 0.004 (0.16) 0.028 (1.12)
Time, h 600 0.012 (0.48) 0.035 (1.36)
310SS 0.012 (0.48) 0.041 (1.60)
S 0.053 (2.08) 0.063 (2.48)
Fig. 6.20 Thermogravimetric results of three test runs for Fe-
X 0.020 (0.80) 0.071 (2.80)
20Cr alloy tested in Ar-20O2-0.5Cl2 at 927 °C iso-
C-276 0.079 (3.12) 0.079 (3.12)
thermally for the first 12 h, followed by a thermal cycle by cooling
6B 0.014 (0.56) 0.098 (3.84)
the specimen to 100 °C for 30 min and raising the specimen to the
188 0.014 (0.56) 0.116 (4.56)
test temperature to resume testing. Note that the thermal cycle
resulted in the initiation of an accelerated chloridation attack. (a) Metal loss + average internal penetration. Source: Ref 35
Source: Ref 34
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:27PM Plate # 0 pg 162

162 / High-Temperature Corrosion and Materials Applications

310SS

Fig. 6.21 Gravimetric results for selected Fe-, Ni-, and Co-base alloys in Ar-20O2-2Cl2 at 900 °C (1650 °F). Source: Ref 36

tested at 900 °C (1650 °F) (Ref 39). Test results

for all the alloys tested at 900 and 1000 °C (1650
and 1830 °F) are summarized in Tables 6.11 and
6.12 (Ref 39).
The beneficial effect of aluminum, as well as
the detrimental effect of molybdenum and tung-
sten, on the resistance to chloridation attack in
oxidizing environments is further substantiated
by the results of long-term tests in another
environment with a higher concentration of Cl2,
as shown in Fig. 6.24 (Ref 35). Similar results
were obtained by Elliott et al. (Ref 40) from tests
conducted in air-2Cl2 at 900 °C (1650 °F) for
50 h (Fig. 6.25).
Chloridation attack in O2-Cl2 environments at
these high temperatures primarily consisted of
metal wastage and internal penetration for most
alloys, with the exception of Ni-Cr-Mo alloys
containing high levels of molybdenum, such as
Fig. 6.22 Thermogravimetric data showing weight loss of
alloys S and C-276, which showed mainly metal
two preoxidized specimens during the exposure
in Ar-50O2-1Cl2 at 900 °C (1650 °F). Source: Ref 37 wastage with little or no internal penetration. In
general, the scales formed on the alloy surface
shows corrosion test results for aluminum- were loose when the test specimens were cooled
containing nickel-base alloys with and without to room temperature after the exposure test, as
molybdenum, such as 214, 601, R-41, and 263, illustrated in Fig. 6.26. Scales were basically
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:27PM Plate # 0 pg 163

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 163

oxides, as shown in Fig. 6.27 and 6.28. Fe-Cr and thus were not collected. As discussed earlier,
oxides were found to form on Type 310 SS, while Jacobson et al. (Ref 37) used a high-pressure
aluminum-rich oxides were the major oxide sampling mass spectrometer to measure volatile
phase along with little nickel-rich oxides that species from alloy S with 14.5% Mo and alloy
formed on alloy 214. Internal attack consisted of 600 (Ni-16Cr-9Fe) with no Mo during the
voids for some alloys and of oxides for other exposure of Ar-50O2-1Cl2 at 900 °C (1650 °F)
alloys. Some alloys appeared to contain both and found that MoO2Cl2 along with NiCl2 and
internal voids and oxides. Figure 6.29 shows CrO2Cl2 were major vapor phases for alloy S
internal void formation in alloy R-41 (Ni-Cr-Co- and NiCl2 and CrO2Cl2 for alloy 600.
Mo-Al-Ti) and alloy 25 (Co-Cr-Ni-W) after McNallan et al. (Ref 39) reported corro-
exposure for 50 h at 900 °C (1650 °F) in air- sion behavior in Ar-20O2-0.25Cl2 at 900 and
2Cl2. It was suggested that the internal pene- 1000 °C (1650 and 1830 °F). This was followed
tration may involve halogen-carbide reactions by a study (Ref 41) using the same environment
and simultaneous void formation (Ref 25). to investigate the same alloys at lower tempera-
Figure 6.29 shows some evidence of carbides tures (i.e., 700, 800, and 850 °C). The results of
being converted into voids during chloridation the tests at 700, 800, and 850 °C (1290, 1470,
attack. Some alloys, however, showed internal and 1560 °F) are summarized in Table 6.14
oxides instead of voids, as illustrated in Fig. 6.30.
Elliott et al. (Ref 40) have identified volatile
species of the condensed products removed from Table 6.11 Corrosion of various alloys in
the exit end of the test apparatus during their Ar-20O2-0.25Cl2 for 400 h at 900 and 1000 °C
investigation in air-2Cl2. Their results are shown (1650 and 1830 °F)
in Table 6.13. No oxychlorides were detected. Weight loss, mg/cm2

Since the analysis of the volatile species was Alloy 900 °C (1650 °F) 1000 °C (1830 °F)

performed on the condensed phases collected at 214 4.28 9.05

601 20.67 124.99
the exit end of the test apparatus, oxychlorides 600 72.08 254.96
were likely to be in a gaseous state at the exit end, 800H 26.91 87.05
310SS 47.15 97.40
556 40.29 82.74
X 54.41 153.49
625 99.07 220.09
R-41 63.83 207.32
263 82.57 229.53
188 139.77 156.30
S 228.21 248.98
C-276 132.05 298.85
Source: Ref 39

Table 6.12 Depth of attack after 400 h

at 900 and 1000 °C (1650 and 1830 °F) in
Ar-20O2-0.25Cl2
900 °C (1650 °F) 1000 °C (1830 °F)
Metal loss, Total depth(a), Metal loss, Total depth(a),
Alloy mm (mils) mm (mils) mm (mils) mm (mils)
214 0.023 (0.9) 0.150 (5.9) 0.013 (0.5) 0.051 (2.0)
601 0.061 (2.4) 0.264 (10.4) 0.203 (8.0) 0.295 (11.6)
600 0.127 (5.0) 0.252 (9.9) 0.330 (13.0) 0.386 (15.2)
800H 0.043 (1.7) 0.191 (7.5) 0.203 (8.0) 0.424 (16.7)
310SS 0.086 (3.4) 0.152 (6.0) 0.191 (7.5) 0.246 (9.7)
556 0.046 (1.8) 0.152 (6.0) 0.152 (6.0) 0.300 (11.8)
X 0.099 (3.9) 0.218 (8.6) 0.318 (12.5) 0.434 (17.1)
625 0.208 (8.2) 0.272 (10.7) 0.356 (14.0) 0.437 (17.2)
R-41 0.114 (4.5) 0.244 (9.6) 0.381 (15.0) 0.457 (18.0)
263 0.130 (5.1) 0.193 (7.6) 0.368 (14.5) 0.424 (16.7)
188 0.216 (8.5) >0.356 (14.0) 0.254 (10.0) 0.417 (16.4)
S 0.315 (12.4) 0.353 (13.9) 0.419 (16.5) 0.472 (18.6)
C-276 0.300 (11.8) 0.320 (12.6) 0.419 (16.5) 0.450 (17.7)
Fig. 6.23 Corrosion of several aluminum-containing nickel-
base alloys with and without molybdenum in (a) Metal loss + internal penetration. Source: Ref 39
Ar-20O2-0.25Cl2 at 900 °C (1650 °F). Source: Ref 39
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:27PM Plate # 0 pg 164

164 / High-Temperature Corrosion and Materials Applications

Fig. 6.24 Corrosion of several nickel- and cobalt-base alloys in Ar-20O2-1Cl2 at 900 °C (1650 °F). Source: Ref 35

in oxidizing environments at high temperatures.

Alloy S was found to be less corrosion resistant
than alloy 800H. However, both alloys S and
800H suffered increasing corrosion attack with
increasing temperatures. This represents a typical
trend for most alloys in oxidizing environments.
One exception is the Ni-Cr-Al system. As illus-
trated in Fig. 6.31, alloy 214 showed a sudden
decrease in corrosion attack as the test tempera-
ture was increased from 900 to 1000 °C (1650
to 1830 °F). This sharp reduction in corrosion
attack at 1000 °C (1830 °F) was attributed to the
formation of a protective Al2O3 scale. At lower
temperatures, such as 900 °C or less, the kinetics
of Al2O3 formation was not fast enough to form a
protective oxide scale in O2-Cl2 environments.
Formation of a protective Al2O3 scale is favored
at higher temperatures (e.g., 1000 °C or higher).
Most of the data presented so far were gener-
Fig. 6.25 Corrosion of several iron- and nickel-base alloys
in air-2Cl2 at 900 and 1000 °C (1650 and 1830 °F) ated at fairly high temperatures (i.e., 900 °C and
for 50 h. Source: Ref 40 higher). Chloridation was quite aggressive at
those high temperatures. Schwalm and Schutze
(Ref 42–44) investigated a large number of
(Ref 41). The corrosion behavior of alloys as a commercial alloys at significantly lower tem-
function of temperature from 700 to 1000 °C peratures, varying from 300 to 800 °C (572 to
(1290 to 1830 °F) can best be summarized in 1472 °F) for times up to 300 h in air-2Cl2. The
Fig. 6.31 using three different alloy systems corrosion behavior of various alloys were gen-
(Ni-Cr-Mo alloys S, Fe-Ni-Cr alloy 800H, and erated in terms of the decrease in specimen cross-
Ni-Cr-Al alloy 214). As discussed earlier, re- section thickness (i.e., metal loss) and the depth
fractory metals, such as molybdenum and tung- of internal attack. Extensive characterization of
sten, are detrimental to chloridation resistance the corrosion products formed on the alloys was
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:27PM Plate # 0 pg 165

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 165

investigated using SEM/EDX analyses, which a function of temperature. The data on the depth
showed distribution of elements including oxy- of internal corrosion attack are presented in
gen and chlorine in the corrosion products. Their Fig. 6.32(b). The corrosion attack at 300 °C
results are briefly summarized below. (572 °F) was quite negligible after 300 h for
Figure 6.32(a) shows the decrease in metal the alloys tested including 2.25Cr-1Mo steel.
cross-section thickness for 2.25Cr-1Mo steel Nevertheless, the Cr-Mo steel was found to
(10CrMo9 10), alloy 800H, alloy AC66, alloy exhibit a fragile oxide scale, which contained
45TM, and alloy 690 after testing for 300 h as Fe, O, and Cl. All other alloys including two

Fig. 6.26 Loose scales on samples of several nickel-base alloys after testing at 900 °C (1650 °F) in Ar-20O2-1Cl2 for 100 h

Semiquantitative EDX analysis, wt%

Area Fe Cr Si
1 85.1 14.9 …
2 69.5 29.9 0.6

Fig. 6.27 Scanning electron micrograph showing oxide scale formed on Type 310SS sample exposed at 900 °C (1650 °F) for 400 h
in Ar-20O2-0.25Cl2. The results of the EDX analysis of the corrosion products on the areas, as marked No. 1 and No. 2,
are tabulated. Magnification bar represents 33.3 µm
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:27PM Plate # 0 pg 166

166 / High-Temperature Corrosion and Materials Applications

nickel alloys (alloys 45TM and 690) showed attack was observed underneath the oxide scale,
pitting type of attack. The phase in the pit was and these internal particles are believed to be
heavily enriched in Cl and O with some Cr, Fe, aluminum oxides. The morphology of the cor-
and Ni. Figure 6.33 shows the morphology of rosion products formed in alloy 690 is shown in
the pit along with an x-ray map for Cl. At 500 °C Fig. 6.35.
(932 °F), 2.25Cr-1Mo steel suffered both sig- Corrosion behavior for alloys 59, C-2000, and
nificant thickness reduction and internal attack, HR160 after testing for 300 h as a function of
while alloys AC66, 800H, 45TM, and 690 temperature is shown in Fig. 6.36. All three
showed little attack. At 650 °C, alloy AC66 alloys exhibited little corrosion attack at 300 and
suffered significantly more thickness loss than 500 °C. At 650 °C, both alloys 59 and HR160
alloys 800H, 45TM, and 690. The elemental continued to exhibit little corrosion attack, while
distribution in the corrosion products for alloy alloy C-2000 suffered much more corrosion
AC66 (worst alloy in this group) is shown in attack. Both C-2000 (Ni-23Cr-16Mo-1.6Cu) and
Fig. 6.34. The chromium oxides with a layer of 59 (Ni-23Cr-16Mo-0.3Al) exhibit similar chemi-
iron-rich oxide that formed on AC66 became cal compositions except C-2000 contains addi-
convoluted. Chlorine was detected at the metal/ tional 1.6% Cu and alloy 59 contains additional
oxide scale interface and in the metal underneath 0.3Al. It is not clear whether Cu in alloy C-2000
the metal/oxide scale interface, where internal was responsible. At 650 °C, both alloys 59 and
attack was observed. Alloy 690, on the other HR160 were found to perform better than alloys
hand, showed a continuous chromium oxide 690 and 45TM (Fig. 6.32). Alloy 59 was found to
scale. No chlorine was detected. Some internal exhibit a thin, continuous chromium-rich oxide

Semiquantitative EDX analysis, wt%

Area Ni Cr Al Fe
1 72 20 4 4
2 6 5 88 1

Fig. 6.28 Scanning electron micrograph showing oxide scale formed on alloy 214 sample tested at 900 °C (1650 °F) for 400 h in
Ar-20O2-0.25Cl2. The results of the EDX analysis of the corrosion products on the areas, as marked No. 1 and No. 2,
are tabulated. Magnification bar represents 33.3 µm
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:27PM Plate # 0 pg 167

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 167

scale after exposure at 800 °C for 300 h in air- high concentration of molybdenum (16%)
2Cl2. as shown in Fig. 6.37. Also observed was showed no detrimental effect on the alloy’s
an outer Fe-, Ni-, and Cr-rich oxide scale. Also corrosion resistance in this oxidizing environ-
observed were a small amount of chlorine and ment containing 2% Cl2. Thermodynamically,
slight molybdenum enrichment at the metal/ this environment at 800 °C, molybdenum oxy-
chromium-oxide scale interface. At 800 °C, all chloride (MoO2Cl2) would be stable, as shown
three alloys showed higher corrosion attack. in Fig. 6.11. Molybdenum oxychloride was be-
Alloy 59 continued to perform the best. The lieved to contribute to high corrosion rates for

Fig. 6.29 Scanning electron micrographs showing internal void formation in (a) alloy R-41 and (b) alloy 25 after exposure for 50 h at
900 °C (1650 °F) in air-2Cl2
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:27PM Plate # 0 pg 168

168 / High-Temperature Corrosion and Materials Applications

high-molybdenum-containing alloys, such as layer of silicon-rich oxide scale underneath the

alloys C-276, S, and so forth when exposed to chromium-rich oxide scale with slight internal
O2-Cl2 environments at 900 °C (1650 °F) and silicon oxides after exposure at 800 °C for 300 h
higher (Ref 35, 39, 40). The chloride phases in air-2Cl2, as shown in Fig. 6.38. The figure also
exhibiting 10−4 atm pressure that are predicted shows that silicon oxides contained what
thermodynamically under this test condition are appeared to be tiny internal voids.
NiCl2, CoCl2, CrO2Cl2, FeCl3, and MoO2Cl2 Schwalm and Schutze (Ref 44) also tested
(Table 6.1), while the oxides are Al2O3, SiO2, alumina former alloy 214 (Ni-16Cr-3Fe-4.5Al-
Cr2O3, and Fe2O3. HR160 alloy showed an inner Y-Zr), intermetallic Fe3Al (Fe-5.5Cr-15.9Al-
0.2Zr), and intermetallic TiAl (Ti-36Al). The
corrosion data are shown in terms of cross-
section thickness loss (Fig. 6.39a) and internal
attack (Fig. 6.39b) after exposure for 300 h at
temperatures from 300 to 800 °C (572 to
1472 °F) in air-2Cl2. Alumina former alloy 214
performed very well, comparable to alloy 59
(Fig. 6.36). Fe3Al, while showing little corrosion
attack at 300 and 500 °C, suffered severe cor-
rosion at 650 °C. At 800 °C, Fe3Al showed very
little corrosion again. The oxide scales formed
on Fe3Al at 650 °C consisted of nonprotective,
multilayers of Cr-containing Fe2O3, Al2O3,
and Fe2O3. At 800 °C, a thin, continuous
aluminum-rich oxide scale was found to form
on Fe3Al when tested at 800 °C. It is clear that

Table 6.13 Volatile species of condensed

products(a) after testing at 900 °C (1650 °F)
Semiquantitative EDX analysis, wt% for 50 h in air-2Cl2
Area Ni Cr Fe Al Ti Alloy Major constituents
1 1 62 1 27 9 Alloy 214 …
2 19 66 7 6 2 Alloy 601(b) NiCl2, AlCl3
3 25 60 12 2 1 Type 310SS FeCl3·2H·2O, NiCl2·6H2O
4 29 45 13 10 3 Alloy 800H(b) FeCl3·2H2O, NiCl2·6H2O
5 56 26 18 … … Alloy 25 CoCl2, NiCl2·6H2O, WCl6
Alloy 625 NiCl2·6H2O, FeCl3·2H2O, MoCl5
Alloy 617 NiCl2·6H2O, FeCl3·2H2O, CoCl2
Fig. 6.30 Scanning electron micrograph showing oxide Alloy 263 NiCl2·6H2O, CoCl2, MoCl5, FeCl3·2H2O
scales and internal oxides for alloy 601 exposed at Alloy C-276 NiCl2·6H2O, MoCl5
900 °C (1650 °F) for 400 h in Ar-20O2-0.25Cl2. The results of the
EDX analysis of the corrosion products on the areas, as marked (a) Collected at the downstream, cooler section of the test apparatus. (b) Tested at
No. 1, No. 2, No. 3, No. 4, and No. 5, are listed. 1000 °C (1830 °F) for 50 h. Source: Ref 40

Table 6.14 Depth of attack for various alloys after 400 h at 700, 800, and 850 °C (1290, 1470,
and 1560 °F) in Ar-20O2-0.25Cl2
700 °C (1290 °F) 800 °C (1470 °F) 850 °C (1560 °F)
Metal loss, Total depth, Metal loss, Total depth, Metal loss. Total depth,
Alloy mm (mils) mm (mils) mm (mils) mm (mils) mm (mils) mm (mils)
214 0.010 (0.4) 0.010 (0.4) 0.018 (0.7) 0.061 (2.4) 0.018 (0.7) 0.066 (2.6)
600 … … 0.020 (0.8) 0.086 (3.4) 0.038 (1.5) 0.132 (5.2)
800H 0.025 (1.0) 0.033 (1.3) 0.023 (0.9) 0.046 (1.8) 0.031 (1.2) 0.097 (3.8)
310SS … … 0.036 (1.4) 0.053 (2.1) 0.031 (1.2) 0.061 (2.4)
556 … … 0.020 (0.8) 0.051 (2.0) 0.020 (0.8) 0.079 (3.1)
S 0.079 (3.1) 0.081 (3.2) 0.145 (5.7) 0.150 (5.9) 0.224 (8.8) 0.257 (10.1)
C-276 0.033 (1.3) 0.046 (1.8) 0.066 (2.6) 0.071 (2.8) 0.163 (6.4) 0.175 (6.9)
188 … … 0.058 (2.3) 0.074 (2.9) 0.025 (1.0) 0.264 (10.4)
Source: Ref 41
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:28PM Plate # 0 pg 169

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 169

higher temperatures are favored for forming chlorides, thus more metal loss. In addition to
an aluminum oxide scale. Unlike Fe3Al, TiAl metal loss, the internal attack for alloy 800H was
suffered severe corrosion attack at 500, 650, more severe in the CO2-Cl2 environment than
and 800 °C. the O2-Cl2 environment (Fig. 6.41). The authors
In many combustion processes involving hypothesized the mechanism of the internal
burning fuels or feedstock containing hydro- corrosion by forming internal chromium car-
carbon, combustion environments under sub- bides, which are then converted to chromium
stoichiometric combustion conditions most often chlorides with carbon reacting with chlorine to
would contain CO2. McNallan et al. (Ref 45) form more chromium carbides. Thus, internal
examined the effect of CO2 on the chloridation corrosion is the result of formation of internal
resistance of Fe-Cr and Fe-Ni-Cr alloys and voids and pores in the metal. Fe-20Cr alloy, on
found that CO2 significantly increased the the other hand, suffered no internal corrosion.
alloy’s metal wastage and internal attack. The The authors attributed this lack of internal corro-
authors compared the environments between sion to the alloy’s low carbon content and ferritic
Ar-20O2-2500 ppm (0.25%) Cl2 and Ar-20CO2- structure.
2500 ppm Cl2 for Fe-20Cr and 800H at 927 °C In another paper by McNallan et al. (Ref 46),
(1700 °F). Figure 6.40 shows metal loss data the CO2-containing environments were further
for both environments as a function of exposure examined. The authors observed internal car-
time. For both alloys, the CO2-Cl2 environment burization of Type 310SS in Ar-20CO2 at 800 °C
caused more metal wastage than the O2-Cl2 for 24 h. When the test was conducted in
environment with Fe-20Cr being more severely Ar-20CO2-Cl2 at 800 °C for 24 h, Type 310SS
affected than alloy 800H. The authors did not suffered internal corrosion attack in forms of
offer an explanation in the paper. It was likely voids and carbides. Similar internal attack was
that when the environment was switched from observed for alloy 800. The corrosion, suggested
Ar-20O2-Cl2 to Ar-20CO2-Cl2 the environment by the authors, proceeded in two stages with
changed from oxidizing to reducing with its carburization preceding the chlorine-accelerated
oxygen potential being reduced to closer to, or in, oxidation.
the CrCl3 and FeCl2 regimes in the Cr(Fe)-O-C In combustion environments, H2O is invar-
diagrams, resulting in formation of more volatile iably present among the combustion products
produced. The oxygen potential is dictated by
partial pressures of oxygen and hydrogen.
Hydrogen reacts with Cl2 to form more stable
HCl molecule. The chloridation behavior in HCl-
containing environments is discussed in the next
two sections.

6.3.3 Corrosion in O2-HCl Environments

In some combustion environments, chlorine is
present as HCl instead of Cl2. This section covers
the corrosion data in oxidizing environments
containing HCl. In an oxidizing environment
containing HCl, chlorine partial pressure, pCl2 ,
can be calculated from the equilibrium condition
of the reaction below. The oxide and chloride
phases that are likely to form thermodynamically
can then be determined from the thermodynamic
phase stability diagrams discussed earlier in
section 6.2.
4HCl(g)+O2 (g)=2Cl2 (g)+2H2 O(g)
The effects of adding O2 to HCl environment
Fig. 6.31 Corrosion behavior of alloy 214 (Ni-Cr-Al-Y), alloy were extensively studied by Ihara et al. (Ref 47)
S (Ni-Cr-Mo) and alloy 800H (Fe-Ni-Cr) in Ar-
20O2-0.25Cl2 for 400 h at 700–1000 °C (1290–1830 °F). Source: on iron, by Ihara et al. (Ref 48) on nickel, and
Ref 39 and Ref 41 by Ihara et al. (Ref 49) on chromium. Adding O2
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:28PM Plate # 0 pg 170

170 / High-Temperature Corrosion and Materials Applications

to HCl significantly accelerates the corrosion mixtures at different temperatures is shown in

of iron (Ref 47), as shown in Fig. 6.42. The Fig. 6.44 (Ref 49).
increased corrosion was attributed to the for- Devisme et al. (Ref 50) investigated the effect
mation of highly volatile FeCl3 (Ref 47). For of O2 and CO2 in HCl-bearing environments on
nickel, adding O2 to HCl does not significantly the corrosion behavior of four nickel-base alloys,
affect the metal’s corrosion rate, as shown in alloys C-276, 600, 601, and 214. Corrosion data
Fig. 6.43 (Ref 48). This is due to the formation in terms of metal loss, which was determined
of primarily NiCl2 whether it is 100% HCl or by weight change and metallographic exami-
O2-HCl mixtures (Ref 48). Adding O2 to HCl nation of the test specimen, are summarized
suppresses the corrosion rate of chromium at in Tables 6.15 and 6.16. Adding 2% O2 to Ar-
lower temperatures (400–600 °C) by forming 20HCl significantly increased corrosion attack
Cr2O3 and accelerates the corrosion rate at for alloy C-276 (Ni-Cr-Mo alloy), but signifi-
higher temperatures (700 and 800 °C) by form- cantly decreased corrosion attack for alloy 214
ing highly volatile CrCl3 (Ref 49). The corrosion (Ni-Cr-Al alloy). Both alloys 600 and 601 were
rate of chromium as a function of O2-HCl not significantly affected. Nevertheless, adding

1600

1575 Alloy AC66

500
Decrease, µm

400 Alloy 800H

10CrMo 9 10*

300
Alloy 45TM
200

100 Alloy 690

0
300 350 400 450 500 550 600 650 700 750 800

Temperature, °C
(a)

225
Alloy AC66
200
10CrMo 9 10*
175

150
Depth, µm

125
Alloy 800H
100

75

50
Alloy 45TM
25 Alloy 690

0
300 350 400 450 500 550 600 650 700 750 800

Temperature, °C
(b)

Fig. 6.32 Corrosion behavior of 2.25Cr-1Mo steel (10CrMo9 10), alloy 800H, alloy AC66, alloy 45TM and alloy 690 tested for 300 h
at temperatures from 300 to 800 °C (572 to 1472 °F) in air-2Cl2; (a) decrease in thicknesses as a function of temperature, and
(b) depth of internal corrosion attack as a function of temperature. Source: Ref 42
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:28PM Plate # 0 pg 171

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 171

10O2 to Ar-5HCl significantly increased corro- were better than alloy 825. Their test results are
sion attack for alloys C-276, 600, and 601. For shown in Fig. 6.45 to 6.47.
Ar-5HCl, adding 0.5% H2O to the environment Smith and Ganesan (Ref 52) conducted further
caused increased corrosion for all four alloys, extensive studies on the corrosion behavior of
particularly severe for alloys 600 and 601. The iron-base alloys (Type 316SS, Type 347SS, and
test results obtained by Devisme et al. (Ref 50) alloy 800HT) and nickel-base alloys (alloys 825,
showed that overall, alloy C-276 (Ni-Cr-Mo) 600, and 625) in simulated combustion environ-
performed better than alumina former alloy 214 ments consisting of N2, O2, SO2 and various
in reducing environments, such Ar-HCl and Ar- amounts of HCl at 426, 593, and 704 °C (800,
HCl-H2 (Table 6.15). Addition of 10% CO2 was 1100, and 1300 °F). Also included in their
found have less effect on corrosion attack for studies was the effect of H2O in the environment
four alloys in Ar-20HCl environment at 600 and on the alloys’ corrosion behavior. Table 6.17
700 °C, as shown in Table 6.16. summarizes the test results generated from tests
Ganesan et al. (Ref 51) investigated the conducted in N2-10O2-50 ppm SO2-500 ppm
corrosion behavior of nickel-base alloys (625, HCl at 426, 482, and 593 °C (800, 900, and
825, and 600) and iron-base alloys (800HT, 1100 °F) for 1008 h. HCl is known to be more
316SS, and 347SS) in a combustion environment corrosive than SO2 in high-temperature corro-
consisting of N2, 4 and 9O2, 12CO2, 500 ppm sion. Thus, in this environment, corrosion attack
SO2 with two levels of HCl (1% HCl and 4% is primarily from HCl. The test results show that
HCl). For both levels of HCl, three nickel- the environment that contained about 500 ppm
base alloys were significantly more resistant to HCl was not corrosive at all for Type 316SS,
chloridation attack than iron-base alloys. For Type 347SS, 800HT, 825, 600, and 625 at
three nickel-base alloys, alloys 625 and 600 temperatures up to 593 °C (1100 °F). When the
HCl content was increased to 4%, both Types
316 and 347 showed much higher corrosion rates
at 593 °C (1100 °F), while alloys 800HT, 825,
600, and 625 showed little corrosion attack at
593 °C (1100 °F) after 1008 h, as shown in
Table 6.18. However, when the temperature was
increased to 704 °C (1300 °F), only high-nickel
alloys, such as alloys 600 and 625 exhibited
good corrosion resistance in the environment
containing 4% HCl (Table 6.19). The environ-
ment containing 10% HCl became very corrosive
to nickel-base alloys 825, 600, and 625 even at
593 °C (1100 °F) (Table 6.18).
At high temperatures, nickel- and cobalt-base
alloys, while exhibiting low metal loss, suffered
more internal attack in O2-HCl environments.
Elliott et al. (Ref 53) examined the corrosion
attack in terms of the metal loss and internal
penetration for nickel-base alloys (alloys 214,
600, and 601) and cobalt-base alloys (alloys 25
and 188) along with Fe-Ni-Co-Cr alloy (alloy
556), Fe-Ni-Cr alloy (alloy 800H), and Type
310SS after testing in Ar-5.5O2-1HCl-1SO2 at
900 °C (1650 °F) for 800 h with thermal cycling
to 200 °C (390 °F) every 100 h. Although the
environment contained SO2, hydrogen chloride
(HCl), which is known to be more corrosive
than SO2, was the primary corrodent causing
the corrosion attack. All alloys, while exhibiting
Fig. 6.33 Scanning electron backscattered image (a) and low metal losses except Type 310SS, suffered
an x-ray map for Cl (b), showing a typical pit on
alloy 800H tested for 300 h at 300 °C (572 °F) in air-2Cl2. Source: significant internal penetration attack. These test
Ref 42 results are shown in Fig. 6.48.
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:28PM Plate # 0 pg 172

172 / High-Temperature Corrosion and Materials Applications

6.3.4 Corrosion in HCl and HCl-Bearing from 2 to 20 h. Thus, extrapolation to a year

Reducing Environments would yield an unreliable corrosion rate. Hossain
et al. (Ref 54) performed long-term tests in HCl
Brown et al. (Ref 22) reported corrosion rates on several nickel alloys and one stainless steel
of various commercial alloys in a dry HCl (Type 310SS). Their results are summarized in
environment (Table 6.20). Test duration varied Table 6.21 and Fig. 6.49. Type 310SS was found

(a) (d)

(b) (e)

(c) (f)

Fig. 6.34 (a) Scanning electron backscattered image of the corrosion products formed on alloy AC66 tested at 800 °C for 300 h in air-
2Cl2 and the x-ray maps showing elemental distribution for (b) chlorine, (c) chromium, (d) oxygen, (e) iron, and (f) nickel.
Source: Ref 42
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:28PM Plate # 0 pg 173

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 173

to be the worst among the alloys tested. All nickel best performers. In a study by Hossain et al.
alloys (Ni-201, 601, 625, C-4, B-2, and 600) (Ref 54), nickel performed reasonably well in
were much better than Type 310SS. In this HCl until the temperature reached 700 °C
HCl test environment containing no O2, the (1290 °F). At 700 °C, nickel was inferior to
molybdenum-containing nickel-base alloys, many nickel-base alloys, such as alloys 600,
such as alloys 625 (Ni-22Cr-9Mo-3.5Nb) and 625, and C-4 (Table 6.21). Alloy 400 was
C-4 (Ni-16Cr-15.5Mo), were found to be the found to be very susceptible to chloridation

(a) (d)

(b) (e)

(c) (f)

Fig. 6.35 (a) Scanning electron backscattered image of the corrosion products formed on alloy 690 tested at 800 °C for 300 h in air-
2Cl2 and the x-ray maps showing elemental distribution for (b) chlorine, (c) chromium, (d) iron, (e) oxygen, and (f ) aluminum.
Source: Ref 42
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:28PM Plate # 0 pg 174

174 / High-Temperature Corrosion and Materials Applications

attack in HCl (Ref 54). Alloy 400 specimens were significantly better than alloys 600, 625,
(6 mm diam × 12 mm length) were completely 188, and X. Their test results generated from the
destroyed after exposure for 100 h at 400 °C 8 h tests at 900 °C (1650 °F) in Ar-4H2-4HCl
(750 °F) in HCl. Alloy 600 is commonly used are shown in Fig. 6.52. In general, alloys suffered
for high-temperature service in HCl environ- very little metal losses, but suffered significant
ments. Figure 6.50 shows the corrosion rates internal penetration attack. Figure 6.53 shows
of alloy 600 in HCl gas as a function of the cross sections of the specimens in various
temperature (Ref 22). MTI Publication MS-3 iron- and nickel-base alloys after testing in Ar-
(Ref 27) suggests corrosion guidelines for 4H2-4HCl at 900 °C (1650 °F) for 8 h (Ref 55).
Ni200, alloy 600, alloy 400, Type 304, and steel A study was performed by Brill et al. (Ref 56)
in HCl environments, as shown in Fig. 6.51. investigating the chlorination resistance of nickel
In reducing environments, such as Ar-4H2- and nickel-base alloys in H2-10HCl. Pure nickel,
4HCl, investigated by Baranow et al. (Ref 35), Ni-Mo, and Ni-Cr-Mo alloys containing little
Ni-Cr-Mo alloys, such as alloys C-276 and S, or no iron were found to be more resistant

Alloy C-2000
500

400
Decrease, µm

300 Alloy HR160

40

30

20 Alloy 59
10

0
300 350 400 450 500 550 600 650 700 750 800
Temperature, °C
(a)

22
20 Alloy C-2000
18 Alloy 59
16
14 Alloy HR160
12
Depth, µm

10
8
6
4
2
0

300 350 400 450 500 550 600 650 700 750 800
Temperature, °C
(b)

Fig. 6.36 Corrosion behavior of alloys 59, C-2000, and HR160; tested for 300 h at temperatures from 300 to 800 °C (572 to 1472 °F)
in air-2Cl2; (a) decrease in thicknesses as a function of temperature, and (b) depth of internal corrosion attack as a function of
temperature. Source: Ref 43
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:28PM Plate # 0 pg 175

Fig. 6.37 (a) Scanning electron backscattered image of the corrosion products formed on alloy 59 tested at 650 °C for 300 h in air-2Cl2
and the x-ray maps showing elemental distribution for (b) chromium, (c) chlorine, (d) molybdenum, (e) iron, (f ) nickel, and
(g) oxygen. Source: Ref 43
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:28PM Plate # 0 pg 176

176 / High-Temperature Corrosion and Materials Applications

than some nickel-base alloys with higher iron In another study by Devisme et al. (Ref 50)
contents. This is illustrated in Fig. 6.54, show- on chlorination of Ni-Cr-Mo alloy C-276, Ni-
ing alloy 205 (pure nickel), B-2 (Ni-28Mo), C-4 Cr-Fe alloy 600, Ni-Cr-Fe-Al alloy 601 (1.4Al),
(Ni-16Cr-16Mo), and 59 (Ni-23Cr-16Mo) being and Ni-Cr-Al-Fe alloy 214 in Ar-HCl environ-
much more resistant than alloys 625 (Ni-22Cr- ments. Their test results conducted at 600 °C
9Mo-3Fe) and 600H (Ni-16Cr-9Fe) (Ref 56). in Ar-5HCl, Ar-10HCl, and Ar-20HCl are

Fig. 6.38 (a) Scanning electron backscattered image of the corrosion products formed on alloy HR160 tested at 800 °C for 300 h
in air-2Cl2, and the x-ray maps showing elemental distribution for (b) chromium, (c) chlorine, (d) silicon, (e) oxygen, and
(f ) nickel. Source: Ref 43
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:28PM Plate # 0 pg 177

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 177

shown in Fig. 6.55. Barnes (Ref 57) investigated (Ref 57). The figure shows that there is no sig-
nickel-base alloys in Ar-HCl mixtures with nificant difference in pCl2 between the 100HCl
concentrations of HCl varying from 13 to 100% and Ar-33HCl environments. However, with the
and found that Ni201 (pure nickel) was slightly
more resistant than alloy 600 (8% Fe) for all the
concentrations tested except 100% HCl, as
shown in Table 6.22. Other nickel-base alloys
were also included in his test program, with the
test results summarized in Table 6.23. Except for
pure nickel (Ni201), nickel-base alloys suffered
internal attack by formation of internal voids.
This surface zone with extensive internal voids
was found to be highly depleted in chromium in
the matrix near voids. For example, alloy 600
showed about 1.7% Cr and 4.7% Fe in the
matrix in the porous zone after testing at 735 °C
(1355 °F) for 100 h in Ar-33HCl (Ref 57).
Barnes’s test environments were strictly HCl
(inlet gas), thus making pCl2 in the environment
significantly higher than that if the initial inlet
gas contain both H2 and HCl. This is shown in
Fig. 6.56, where pCl2 was plotted as a function
of temperature for 100% HCl, Ar-33HCl, and
H2-30HCl along with several metal chlorides

400
350
300
250
200
Decrease, µm

150
50 TiAl Fe3Al
40
30
20
10 Alloy 214
0
300 350 400 450 500 550 600 650 700 750 800
Temperature, °C
(a)

105
100
95
30
Depth, µm

25
20 Fe3Al Alloy 214
15
10
5
TiAl
0
300 350 400 450 500 550 600 650 700 750 800
Temperature, °C
(b)

Fig. 6.39 Corrosion behavior of alumina former alloy 214

and intermetallics Fe3Al and TiAl; tested for 300 h
at temperatures from 300 to 800 °C (572 to 1472 °F) in air-2Cl2; Fig. 6.40 Metal loss as a function of time for alloys Fe-20Cr
(a) decrease in thicknesses as a function of temperature, and and 800H tested at 927 °C (1700 °F) in (a) Ar-
(b) depth of internal corrosion attack as a function of temperature. 20O2-2500 ppm (0.25%) Cl2 and (b) Ar-20CO2-2500 ppm Cl2.
Source: Ref 44 Source: Ref 45
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:28PM Plate # 0 pg 178

178 / High-Temperature Corrosion and Materials Applications

introduction of H2 in the H2-HCl gas mixture,

pCl2 becomes significantly lower, as shown in
H2-30HCl. Even though chlorine partial pressure
(pCl2 ) stays relatively unchanged (Fig. 6.56)
thermodynamically, the kinetic of the HCl-metal
reaction (corrosion rate) in Ar-HCl mixtures
was found to increase with increasing HCl con-
centration (HCl partial pressure). This is shown
in Fig. 6.57 for both Ni201 and alloy 600 tested
at 735 °C (1355 °F). It should also be noted
that Barnes’s test environments involving
mainly HCl were extremely corrosive for nickel
alloys tested. In addition to high pCl2 values, the
environments were so reducing that no oxides
were thermodynamically stable. Thus, in those
Fig. 6.41 Depth of internal penetration as a function of time
test environments, the gas-metal reaction mainly
(h0.5) for alloy 800H tested at 927 °C (1700 °F) in
(a) argon 20% O2, 0.25% Cl2 and (b) argon 20% CO2, 0.25% Cl2. involved formation of chlorides (highly volatile
Source: Ref 45
corrosion products) and no oxides (nonvolatile
solid phases). Figure 6.58 shows nickel chlorides
that formed on a Ni201 coupon after testing at
735 °C (1355 °F) in Ar-33HCl (Ref 57).
In a simulated waste incineration environment
consisting of N2, 12% CO2, 500 ppm SO2, and

Fig. 6.42 Effect of oxygen in O2-HCl mixtures on the corro- Fig. 6.43 Effect of oxygen in O2-HCl mixtures on the corro-
sion rate of iron at 300–700 °C (570–1290 °F). sion rate of nickel at 400–700 °C (750–1290 °F).
Source: Ref 47 Source: Ref 48
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:28PM Plate # 0 pg 179

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 179

1% HCl, Ganesan et al. (Ref 51) observed that

nickel-base alloys, particularly alloy 625, per-
formed significantly better than iron-base alloys,
such as alloy 800HT, Type 316SS, and Type
347SS, as shown in Fig. 6.59 and 6.60. Although
not discussed in the paper, it is believed that
the oxygen potential (pO2 ) was high enough to
form some oxides, such as Cr2O3. Thus, little
corrosion attack was detected for nickel-base
alloys even when the test temperature was in-
creased to 704 °C (1300 °F) (Ref 51). Formation
of chromium oxides can significantly reduce
the corrosion attack in reducing, HCl-bearing
1

800HT

316SS

347SS

Fig. 6.44 Effect of oxygen in O2-HCl mixtures on Fig. 6.45 Weight change as a function of exposure time
for nickel-base alloys (alloys 625, 600, and 825)
the corrosion rate of chromium at 400–800 °C
and iron-base alloys (alloy 800HT, 316SS, and 347SS) in N2-
(750–1470 °F). Source: Ref 49
4O2-12CO2-1HCl-500 ppm SO2. Testing was initially performed
at 649 °C, then increased to 704 °C, and finally to 760 °C as
indicated. Source: Ref 51
Table 6.15 Corrosion of nickel-base alloys in
terms of metal loss µm (mils) after 500 h at 600 °C
(1112 °F) in the indicated test environments
Metal loss, μm (mils)
Environment C-276 600 601 214
Ar-20HCl 60 (2.4) 150 (5.9) 150 (5.9) 260 (10.2)
Ar-20HCl-2O2 330 (13.0) 185 (7.3) 120 (4.7) 65 (2.6)
Ar-5HCl 35 (1.4) 50 (2.0) 90 (3.5) 30 (1.2)
Ar-5HCl-10O2 120 (4.7) 140 (5.5) 160 (6.3) 55 (2.2)
Ar-5HCl-0.5H2O 90 (3.5) 240 (9.5) 255 (10.0) 80 (3.2) 800HT
Ar-5HCl-3H2 5 (0.2) 15 (0.6) 15 (0.6) 20 (0.8)
Source: Ref 50

347SS

Table 6.16 Metal loss after 500 h at 600 and

316SS
700 °C in the indicated test environments
Metal loss, μm
Environment Temperature, °C C-276 600 601 214
Ar-20HCl 600 60 150 150 260
Ar-20HCl-10CO2 600 110 100 120 230 Fig. 6.46 Weight change as a function of exposure time
for nickel-base alloys (alloys 625, 600, and 825)
Ar-20HCl 700 200 280 225 360 and iron-base alloys (alloy 800HT, 316SS, and 347SS) in N2-
Ar-20HCl-10CO2 700 280 195 250 190 9O2-12CO2-1HCl-500 ppm SO2. Testing was initially performed
Source: Ref 50 at 649 °C, then increased to 704 °C, and finally to 760 °C as
indicated. Source: Ref 51
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:29PM Plate # 0 pg 180

180 / High-Temperature Corrosion and Materials Applications

800HT
316SS

704 347SS

Fig. 6.48 The metal loss and internal penetration for nickel-
base alloys (alloys 214, 600, and 601) and cobalt-
Fig. 6.47 Weight change as a function of exposure time
base alloys (alloys 25 and 188) along with Fe-Ni-Co-Cr alloy
for nickel-base alloys (alloys 625, 600, and 825)
and iron-base alloys (alloys 800HT, 316SS, and 347SS) in N2- (alloy 556), Fe-Ni-Cr alloy (alloy 800H), and Type 310SS tested in
9O2-12CO2-4HCl-100 ppm SO2. Testing was initially performed Ar-5.5O2-1HCl-1SO2 at 900 °C (1650 °F) for 800 h with thermal
at 593 °C, then increased to 704 °C, and to 816 °C, and finally cycling to 200 °C (390 °F) every 100 h. Source: Ref 53
to 927 °C as indicated. Source: Ref 51

Table 6.19 Corrosion rates in terms of metal

loss for iron- and nickel-base alloys at 593
Table 6.17 Corrosion rates in terms of metal loss and 704 °C (1100 and 1300 °F) in N2-10O2-
for iron- and nickel-base alloys at 426, 482 and 50 ppm SO2-4HCl
593, 900 °C (800, 900, and 1300 °F) for 1008 h
in N2-10O2-50 ppm SO2-500 ppm HCl Metal loss
593 °C(a) 704 °C(b)
Metal loss
Alloy µm/yr mpy µm/yr mpy
426 °C 482 °C 593 °C
Type 316 914 36 3,810 149
Alloy µm/yr mpy µm/yr mpy µm/yr mpy
Type 347 1245 49 … …
Type 316 0.07 0.003 2.54 0.1 5.08 0.2 Alloy 800HT 74 2.9 12,039 470
Type 347 0.09 0.004 2.29 0.09 7.11 0.28 Alloy 825 20 0.8 2,083 81
Alloy 800HT 0.07 0.003 1.02 0.04 5.33 0.21 Alloy 600 25 1.0 41 1.6
Alloy 825 0.02 0.0008 1.27 0.05 2.54 0.1 Alloy 625 16 0.6 152 5.9
Alloy 600 0.04 0.002 2.03 0.08 3.05 0.12
(a) Test duration: 72 h. (b) Test duration: 192 h. 1.0 µm = 0.001 mm = 0.0394 mil.
Alloy 625 0.02 0.0008 1.78 0.07 2.79 0.11 Source: Ref 52
1.0 µm = 0.001 mm = 0.0394 mil. Source: Ref 52

Table 6.20 Corrosion of alloys in dry HCl(a)

Approximate temperature, °C (°F),
Table 6.18 Corrosion rates in terms of at which given corrosion rate is exceeded
metal loss for iron- and nickel-base alloys at 0.8 mm/yr 1.5 mm/yr 3.0 mm/yr 15 mm/yr
593 °C (1100 °F) for 72 h in N2-9O2-12CO2- Alloy (30 mpy) (60 mpy) (120 mpy) (600 mpy)
100 ppm SO2-4 and 10HCl Nickel 455 (850) 510 (950) 565 (1050) 675 (1250)
600 425 (800) 480 (900) 538 (1000) 675 (1250)
Metal loss
B 370 (700) 425 (800) 480 (900) 650 (1200)
4% HCl 10% HCl C 370 (700) 425 (800) 480 (900) 620 (1150)
Alloy µm/yr mpy µm/yr mpy D 288 (550) 370 (700) 455 (850) 650 (1200)
l8-8Mo 370 (700) 370 (700) 480 (900) 593 (1100)
Type 316 914 36 … … 25-12Cb 345 (650) 400 (750) 455 (850) 565 (1050)
Type 347 1245 49 … …
18-8 345 (650) 400 (750) 455 (850) 593 (1100)
Alloy 800HT 74 2.9 … … Carbon steel 260 (500) 315 (600) 400 (750) 565 (1050)
Alloy 825 20 0.8 1066 42
Ni-resist 260 (500) 315 (600) 370 (700) 538 (1000)
Alloy 600 25 1.0 1219 48 400 230 (450) 260 (500) 345 (650) 480 (900)
Alloy 625 16 0.6 1549 60
Cast iron 205 (400) 260 (500) 315 (600) 455 (850)
1.0 µm = 0.001 mm = 0.0394 mil. Source: Ref 52 Copper 93 (200) 148 (300) 205 (400) 315 (600)
(a) Based on short-term laboratory tests. Source: Ref 22

environments. This is illustrated by the test

results generated by Strafford et al. (Ref 58). H2-H2O-HCl mixture giving pO2 value of
The authors tested Fe-Cr alloys containing 2, 5, 1.5 × 10−16 atm and pCl2 value of 10−8 atm at the
9, 14, and 25% Cr at 1000 °C (1832 °F) in a test temperature. Based on the M-O-Cl stability
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:29PM Plate # 0 pg 181

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 181

diagram at the test temperature, the environment reduced or prevented. The results also indicate
was in the location where Cr2O3, but not FeO, that Fe-Cr alloys with low chromium contents,
was to form, and FeCl2 was to form. Test results such as 2%, 5%, and 9%, suffered significant
show that Fe-25Cr alloy exhibited no weight loss corrosion attack (Fig. 6.61).
for the exposure time up to 100 h. This suggests In an investigation into possible candidate
that once the alloy contains sufficient chromium alloys for a process developed by the Bureau
to form a continuous Cr2O3 scale, as in the case of Mines for extracting alumina from Kaolini-
of Fe-25Cr, corrosion can be significantly tic clay, Carter et al. (Ref 59) tested various

Table 6.21 Corrosion of selected alloys in HCl at 400, 500, 600, and 700 °C (750, 930, 1110,
and 1290 °F)
Metal loss mg/cm2
400 °C (750 °F) 500 °C (930 °F) 600 °C (1110 °F) 700 °C (1290 °F)
Alloy 300 h 1000 h 100 h 300 h 1000 h 100 h 300 h 96 h
Ni-201 1.19 0.91 1.60 2.89 4.86 11.46 37.7 377
601 1.58 1.47 2.57 4.14 9.38 9.01 19.46 102.5
310 3.26 5.16 6.74 13.65 46.60 15.65 32.6 1025
625 0.74 1.1 2.42 3.78 8.64 6.79 14.6 26.5
C-4 0.55 1.12 2.09 3.36 7.24 7.31 19.14 34.9
B-2 0.75 0.76 2.10 2.65 5.87 12.93 62.3 126.4
600 0.93 0.81 1.69 3.31 7.81 7.67 17.3 49.6
Source: Ref 54

Fig. 6.49 Corrosion rates of several iron- and nickel-base alloys in HCl at 400 to 700 °C (750 to 1290 °F). Source: Ref 54
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:29PM Plate # 0 pg 182

182 / High-Temperature Corrosion and Materials Applications

commercial alloys in an environment of 40% alloy. Long-term tests (1584 h) at 260 to 380 °C
HCl and 60% H2O at various temperatures up to (500 to 720 °F) showed corrosion rates of 0.3
500 °C (930 °F). The results of iron- and nickel- to 8 µm/yr for alloy 625 and 4 to 10 µm/yr for
base alloys are summarized in Tables 6.24 alloy R-41. The results were in agreement with
and 6.25. Corrosion data at temperatures below those obtained from short-term tests. Titanium
200 °C (390 °F) are not included because of dew and Ti-0.2Pd were also tested in the same
point corrosion. Stainless steels and nickel alloys environment for 15 days. The corrosion rates
tested showed low corrosion rates at all test for titanium were found to be 246, 28, and 13
temperatures in this environment. A cobalt-base µm/yr at 400, 300, and 200 °C, respectively.
alloy (alloy 188) and an Fe-Ni-Co-Cr alloy Ti-0.2Pd was found to corrode at 1900, 108, 0,
(Multimet alloy) were also tested in the tem- and 2 µm/yr at 500, 400, 300, and 200 °C,
perature range of 315 to 375 °C (600 to 710 °F), respectively.
showing no measurable attack for alloy 188 and In a study by Reeve (Ref 60) involving 80%
a corrosion rate of only 0 to 3 µm/yr for Multimet HCl and 20% H2O, corrosion rates were obtained
for steel and stainless steels (Fig. 6.62). Mild
steel was found to suffer corrosion rates of
less than 0.76 mm/yr (30 mpy) up to 400 °C
(752 °F), and the 18-8 stainless steel, less than
month

1.0 mm/yr (40 mpy) up to 500 °C (930 °F).

These values were significantly lower than those
predicted by Brown et al. (Ref 22) in dry HCl
(Table 6.20). It appears that addition of signifi-
cant amount of H2O into HCl environment
makes HCl less corrosive.

6.4 Corrosion in F2- and HF-Bearing

Environments
6.4.1 Corrosion in F2 Environments
Fig. 6.50 Corrosion rate of alloy 600 in HCl as a function of Early studies of corrosion in fluorine gas for
temperature. Source: Ref 22 various metals and alloys were carried out by

Corrosion rate, mm/yr

0.025 0.05 0.10 0.25 0.50 1.00 2.5

700 1290

Tubes / Internals Vessels / Pipes

600 1110

500 Nickel 200 and Alloy 600 1020

Temperature, °C

Temperature, °F

400 Type 304 930

300 Alloy 400 750

200 Carbon steel 572

100 390

0
1 2 4 6 10 20 40 60 100

Corrosion rate, mpy

Fig. 6.51 MTI corrosion guidelines for Ni200, alloy 600, alloy 400, Type 304SS and steel in HCl as a function of temperature. Source:
Ref 27. Courtesy of Materials Technology Institute
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:29PM Plate # 0 pg 183

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 183

Hauffe (Ref 64) summarized the results of

Metal loss corrosion tests by Myers and DeLong (Ref 61),
214
Jarry et al. (Ref 62), and Lukyanchev et al.
Continuous
internal penetration
(Ref 65) in Table 6.27. The tests were very short
C-276
in duration, from 30 min to 32 h. The corrosion
rates obtained by Myers and DeLong were two
orders of magnitude higher than those observed
S by Jarry et al. and Lukyanchev et al. Based on
the corrosion rates observed by Jarry et al. and
Lukyanchev et al., nickel exhibited good resis-
601 tance in fluorine at temperatures up to 810 °C
(1490 °F) (about 12 mpy).
Hale et al. (Ref 66) tested six different types
R-41
of nickel with various degrees of purity at 590 °C
(1100 °F) for 95 h in fluorine and found no
600
measurable corrosion for all the samples, except
one with a slightly higher silicon content (about
8 mils of general corrosion). The results by
625 Hale et al. (Ref 66) are shown in Table 6.28.
They also found that the material’s purity
played an important role when tested at 700 °C
188 (1290 °F) for 210 h (Ref 66). Both high-purity
vacuum-melted nickel and electrolytic nickel
showed negligible attack. Low-carbon nickel
X I (about 0.015% Si), carbonyl nickel, and “A”
0 40 80 120 160 200 240 280 320 360 nickel (commercial grade pure nickel) began to
Corrosion rate, mils/month show intergranular void formation at 700 °C
(1290 °F). At the same temperature, low-carbon
Fig. 6.52 Corrosion rates in terms of metal loss and internal nickel II (about 0.05% Si) suffered significant
penetration for nickel- and cobalt-base alloys at general corrosion and intergranular attack (void
900 °C (1650 °F) in Ar-4HCl-4H2. Data was based on 8 h tests.
Source: Ref 35 formation). Silicon appears to be a detrimental
impurity in pure nickel in resisting corrosion
attack by fluorine. Results obtained by Steindler
Myers and DeLong (Ref 61). Their test duration and Vogel (Ref 67) showed that “A” nickel
was short, normally 4 h. The longest test run (commercial pure nickel) suffered significant
was 15 h. The corrosion rates they obtained are corrosion at 750 °C (1380 °F), as shown in
summarized in Table 6.26. These rates, extra- Table 6.29. A general review on the kinetic
polated from short-term tests, are not recom- aspects of nickel-fluorine reactions can be found
mended for estimating the service life of process in Ref 8.
equipment. Rather, the data are useful for Additions of alloying elements to nickel
making performance comparisons among var- generally are detrimental to fluorine corrosion
ious alloys. resistance. Many nickel-base alloys were found
The results obtained by Myers and DeLong to be significantly more susceptible than
(Ref 61) indicated that nickel had good corrosion nickel to fluorine corrosion (Ref 61, 66, 67).
resistance in fluorine gas at temperatures up to Alloy 600 (Ni-15.5Cr-8Fe), Inco 61 weld wire
500 °C (930 °F). Even at temperatures higher (Ni-1.5Al-3Ti), Ni-O-Nel (Ni-21Cr-31Fe-3
than 500 °C (930 °F), corrosion rates for nickel Mo-1.75Cu), INOR-1 (Ni-20Mo), INOR-2 (Ni-
were significantly lower than those of other 5Cr-16Mo), INOR-3 (Ni-16Mo-1Al-1.6Ti),
alloys. Nickel is commonly used for plant equip- INOR-4 (Ni-17Mo-1.7Fe-2Al-1.7Ti), INOR-5
ment handling fluorine at temperatures up to (Ni-13Mo-2.7W-2.2Cb+Ta-1Mn), Hastelloy W
500 °C (930 °F) (Ref 62). The resistance of (Ni-25Mo-5.5Fe-2.5Co), Hastelloy B (Ni-
nickel to fluorine gas was attributed to the for- 25Mo-6Fe), HyMa 80 (Ni-16Fe-4Mo), and
mation of an adherent nickel fluoride scale Monel (Ni-30Cu) suffered significantly more
(Ref 62). The reaction of nickel and fluorine corrosion than nickel in fluorine at 590 °C
was found to follow a parabolic rate law (Ref 63). (1100 °F), as shown in Table 6.28 (Ref 66).
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:29PM Plate # 0 pg 184

184 / High-Temperature Corrosion and Materials Applications

Fig. 6.53 Internal penetration in terms of voids for various iron- and nickel-base alloys after testing in 900 °C (1650 °F) for 8 h in
Ar-4H2-4HCl. Source: Ref 55

–50

–100
Metal loss, µm

–150

–200

–250 5% HCl
Fig. 6.54 Corrosion of nickel-base alloys in H2-10HCl
10% HCl
at 850 °C with 24 h cycles. Source: Ref 56
20% HCl
–300
C276 600 601 214
Dura-Nickel (Ni-3Ti-1.5Al), 70Cu-30Ni, and Alloy
Ni-10Co suffered slightly more corrosion than
nickel at 590 °C (1095 °F) (Ref 66). Fluorides of Fig. 6.55 Corrosion in terms of loss of sound metal (µm) of
alloys C-276, 600, 601, and 214 in Ar-5HCl, Ar-
molybdenum, tungsten, titanium, and other ele- 10HCl, and Ar-20HCl at 600 °C for 500 h. Source: Ref 50
ments have low melting points and/or high vapor
pressures (Table 6.2).
Iron is significantly less resistant to fluorine at temperatures up to 250 °C (480 °F). Steels
attack than nickel. Myers and DeLong (Ref 61) were found to be less resistant than Armco Iron
observed that Armco Iron is resistant to fluorine (Ref 61). The concentration of silicon appears
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:29PM Plate # 0 pg 185

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 185

Table 6.22 Metal loss rate(a) in Ar-HCl to be important in affecting the steel’s fluor-
and 100HCl environments at 735 °C (1355 °F)(b) idation resistance, as shown in Table 6.26
Metal loss (Ref 61). SiF4 has a very low melting point
Alloy Environment mm/yr mpy and high vapor pressure (Table 6.2). Ferritic
Ni201 Ar-13HCl 2.8 110 and austenitic stainless steels, except Type
Alloy 600 Ar-13HCl 3.5 138 347SS, showed negligible corrosion at 200 and
Ni201 Ar-33HCl 5.3 209
Alloy 600 Ar-33HCl 7.9 311 250 °C (390 and 480 °F) (Table 6.26). At
Ni201 Ar-52HCl 9.6 378 higher temperatures, corrosion of these alloys
Alloy 600 Ar-52HCl 11.4 449
Ni201 100HCl 18.4 724
became significant. Jackson (Ref 68) also
Alloy 600 100HCl 14.2 559 reported significant corrosion rates for several
(a) Metal loss rate did not include internal void penetration. Internal void pene- austenitic stainless steels at 370 °C (700 °F), as
tration was a small portion of the total metal loss. (b) Test durations from 15–100 h.
Source: Ref 57

Table 6.23 Estimated metal loss rates for

pure nickel and nickel-base alloys tested at 685,
735, and 785 °C (1265, 1355, and 1445 °F)
in 100HCl
Metal loss
Alloy Temperature, °C (°F) mm/yr mpy
Ni201 685 (1265) 4.2 165
Alloy 600 685 (1265) 7.5 295
Alloy HR160 685 (1265) 7.0 276
Alloy 214 686 (1265) 5.7 224
Alloy 602CA 686 (1265) 28.6 1130
Alloy HR160 735 (1355) 45.6 1800
Ni201 785 (1445) 16.7 657
Alloy 600 785 (1445) 26.4 1040
Alloy HR160 785 (1445) 96.0 3780
Fig. 6.57 Metal loss rate (mm/yr) as a function of HCl con-
Source: Ref 57 centrations (pHCl) in Ar-HCl mixtures at 735 °C
(1355 °F). Source: Ref 57

H2-30% HCI

Fig. 6.56 Thermodynamic equilibrium chlorine partial pressure (pCl2 ) as a function of temperature for several environments (100HCl,
Ar-33HCl, and H2-30HCl) and several chlorides. Source: Ref 57
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:29PM Plate # 0 pg 186

186 / High-Temperature Corrosion and Materials Applications

Fig. 6.58 Nickel chlorides formed on Ni201 after testing at 735 °C (1355 °F) for 15 h in Ar-33HCl. Source: Ref 57

347SS

316SS
316SS
347SS

Fig. 6.59 Mass change as a function of time for nickel-

Fig. 6.60 Mass change as a function of time for nickel- and
and iron-base alloys tested initially at 593 °C
(1100 °F), then increased to 649 °C (1200 °F), and finally to iron-base alloys tested at 593 °C (1100 °F) in N2-
704 °C (1300 °F) in N2-12%CO2-500ppm SO2-1%HCl. Source: 12%CO2-500ppmSO2-1%HCl. Source: Ref 51
Ref 51

shown in Table 6.30. Cobalt and cobalt-base volatile fission product fluorides, which were
alloys are not as resistant to fluorine as nickel associated with the development of a process to
(Ref 66) (Table 6.28). Limited data for other recover uranium and plutonium from partially
metals, such as copper, aluminum, and mag- spent nuclear reactor fuels. Most of these fission
nesium, are shown in Tables 6.26, 6.29, and product fluorides were much less corrosive than
6.31. Aluminum was resistant to fluorine at fluorine gas. However, whenever fluorine gas
temperatures up to approximately 500 °C was present along with the fluoride, the corrosion
(930 °F) (Table 6.26 and 6.31). AlF3 has a rate was generally more aggressive. Nickel and
relatively high melting point (1197 °C or alloy 400 exhibited relatively low corrosion rates
2187 °F) (Ref 64). for all the volatile fluorides at 500 °C (930 °F),
Corrosion of nickel, alloy 400, and alloy 600 except TeF6, as shown in Table 6.32. Alloy 600,
by various volatile metallic fluorides was inves- on the other hand, suffered extremely high cor-
tigated by Vogel et al. (Ref 69) in their studies on rosion rates.
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:29PM Plate # 0 pg 187

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 187

6.4.2 Corrosion in HF Environments alloys in HF based on short-term tests were

reported by Myers and DeLong (Ref 61), as
Hydrogen fluoride is generally less corrosive
shown in Table 6.33. Nickel was the most resis-
than fluorine for most metals and alloys. How-
tant among the alloys tested. Copper, alloy 400,
ever, it is still very corrosive at elevated tem-
and alloy 600 were slightly worse than nickel.
peratures. Corrosion rates of various metals and
Carbon steels and stainless steels showed poor
resistance. Tyreman and Elliott (Ref 70) found
that nickel, cobalt, copper, and molybdenum
were more resistant to HF corrosion than iron,
chromium, niobium (columbium), and tantalum
(Fig. 6.63). Chromium was found to be detri-
mental to resistance to corrosion by HF for Ni-Cr
alloys (Fig. 6.64) (Ref 71). Chromium fluorides
were the major corrosion products for Ni-Cr
alloys tested in HF (Ref 71).
Marsh (Ref 72) did an extensive investigation
for his Ph.D. thesis on the resistance of several
nickel-base alloys to corrosion attack in HF
environments at elevated temperatures. In Ni-
Cr alloys tested in HF at 650 °C (1200 °F),
Fig. 6.61 Weight change as a function of exposure time at chromium fluorides were found to form on the
1000 °C (1832 °F) in a H2-H2O-HCl mixture giving metal surface as well as internally as internal
pO2 value of 1.5 × 10−16 atm and pCl2 value of 10−8 atm for Fe-Cr
alloys containing various amounts of chromium. Source: Ref 58 phases. This is illustrated in Fig. 6.65, showing
elemental distribution for Ni-40Cr alloy tested at
650 °C (1200 °F) in anhydrous HF environment
Table 6.24 Corrosion of iron-base alloys in (Ref 72). Surface corrosion products were found
40HCl-60H2O to be enriched in chromium and fluorine, but
depleted in nickel, and similar elemental
Corrosion rate(a) μm/yr
Temperature, E-Brite 1020 4130
distribution was observed for the internal phases.
°C (°F) 316LSS 29-4SS 430SS 26-1 steel steel Several commercial nickel-base alloys were also
500 (930) 483 … … 363 2100 1700 tested. Figure 6.66 shows corrosion attack of
400 (750) 15 … … 5 326 406 alloy 600 in HF after 92 h at 650 °C (1200 °F).
300 (570) 3 10 … 2 700 870
210 (410) … 3 8 … … … In this case, both the corrosion products formed
200 (390) 7 … … 3 46 51 on the metal surface and the internal phases
(a) Linearly extrapolated from 15 d (360 h) laboratory tests. Note: mpy = (µm) × were found to consist of both CrF3 and FeF2,
0.0394. Source: Ref 59
as determined by x-ray diffraction analysis
(Ref 72). In Ni-Cr alloys, it appears that Ni-
Cr-Mo alloy 625 was more resistant to HF
corrosion attack than Ni-Cr alloys X750 and
Table 6.25 Corrosion of nickel-base alloys in
40HCl-60H2O
Corrosion rate(a), μm/yr
Temperature,
°C (°F) Ni-200 600 625 R-41 B-2
500 (930) … … 132 … …
380 (720) 13 58 … 3 15
380 (720) 15 46 13 3 18
380 (720) 15 58 23 3 20
375 (710) 25 48 10 5 13
375 (710) 13 48 15 0 28
365 (690) 20 46 13 3 33
350 (660) 3 30 8 3 …
350 (660) 18 33 15 5 36
315 (600) 13 15 8 0 …
310 (590) … … 0 0 …
290 (550) … … … … …
260 (500) 30 13 8 3 28
(a) Linearly extrapolated from 15 d (360 h) laboratory tests. Note: mpy = (µm) × Fig. 6.62 Corrosion rates of mild steel, Type 304SS, and alloy
0.0394. Source: Ref 59 800 in 80HCl-20H2O at 300 to 600 °C (570 to
1110 °F). Source: Ref 60
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:29PM Plate # 0 pg 188

188 / High-Temperature Corrosion and Materials Applications

Table 6.26 Corrosion of various metals and alloys in fluorine

Corrosion rates(a), mpy
Materials 200 °C 250 °C 300 °C 350 °C 400 °C 450 °C 500 °C 600 °C 650 °C 700 °C
Nickel … … … … 8 23 61 348 192 408
400 … … … … 6 18 24 720 960 1,800
600 … … … … 456 1,152 744 2,040 1,560 6,120
Copper … … … … 1,920 … 1,440 11,880 … …
Aluminum … … … … Nil Nil 156 216 … …
Magnesium Nil Nil Nil … … … … … … …
430SS 8.4 Nil 3,060 936 936 … … … … …
347SS Nil 1,740 2,556 6,204 9,540 … … … … …
309CbSS Nil Nil 900 5,544 7,980 … … … … …
310SS Nil Nil 372 4,248 6,732 … … … … …
Armco Fe Nil 24 108 96 288 3600 139 in. … … …
Steel (0.007% Si) Nil 192 48 2.4 144 … … … … …
1020 (0.22% Si) 456 5,760 7,920 1,764 6,480 18 in. … … … …
1030 (trace Si) 24 96 108 Nil 180 6,480 238 in. … … …
1030 (0.18% Si) … … 9,000 … … … … … … …
1015 (0.07% Si) … … 9,960 … … … … … … …
Music wire (0.13% Si) … … 4,800 … … … … … … …
(a) Note: mpy × 0.0254 = mm/yr. Source: Ref 61

Table 6.27 Corrosion of nickel in fluorine Table 6.28 Results of corrosion tests in
at various temperatures fluorine at 590 °C (1100 °F) for 95 h
Temperature, Pressure, Test Corrosion rate, Depth of corrosion
°C (°F) atm duration, h mm/yr (mpy) Ref attack(a),
Alloy mm (mils) Comments
300 (570) 0.9 8 0.005 (0.2) 62
300 (570) 0.9 32 0.00073 (0.03) 62 High-purity nickel sheet NM …
400 (750) 0.9 0.5 0.015 (0.59) 62 High-purity nickel rod NM …
400 (750) 0.9 28 0.0012 (0.05) 62 Low-carbon nickel (II) 0.20 (8) General attack
400 (750) 0.99 … 0.21 (8.3) 61 Low-carbon nickel (I) NM …
500 (930) 0.9 0.25 0.018 (0.7) 62 Carbonyl nickel NM …
500 (930) 0.9 30 0.003 (0.12) 62 Electrolytic nickel NM …
500 (930) 0.99 … 1.5 (59.1) 61 “A” nickel NM …
550 (1020) 0.99 2 0.036 (1.4) 65 70-30 Cupronickel 0.06 (2.5) General attack
600 (1110) 0.9 32 0.017 (0.7) 62 Inconel (low carbon) 0.64 (25) General attack
660 (1220) 0.99 1.5 0.150 (5.9) 65 Inco “61” weld wire 0.64 (25) General attack
720 (1330) 0.99 2 0.240 (9.4) 65 Duranickel 0.10 (4) General attack
810 (1490) 0.99 1.7 0.310 (12.2) 65 Ni-O-NEL >0.42 (16.5) Completely converted
to fluorides
Source: Ref 64 INOR-1 >0.83 (32.5) Completely converted
to fluorides
INOR-2 >0.83 (32.5) Completely converted
to fluorides
601 containing no significant amounts of INOR-3 >0.83 (32.5) Completely converted
to fluorides
molybdenum, as illustrated in Fig. 6.67 (Ref 71). INOR-4 >0.81 (32) Completely converted
Corrosion attack in HF for alloys 601, 625, and to fluorides
N is shown in Fig. 6.68 to 6.71. Alloy N (Ni-5Cr- INOR-5 0.76 (30) General attack
Hastelloy alloy B >0.48 (19) Completely converted
16Mo) was found to be significantly more to fluorides
resistant to HF corrosion than nickel-base alloys HyMa 80 0.37 (14.5) General attack
600, 601, and 625. This suggests that nickel-base 90Ni-10Co 0.06 (2.5) General attack
80Ni-20Co >0.13 (5) General attack
alloys containing low chromium and high Monel >0.80 (31.5) Completely converted
molybdenum would be more resistant to HF to fluorides
Hastelloy alloy W >0.79 (31) Completely converted
corrosion attack than Ni-Cr alloys. to fluorides
Molybdenum showed little corrosion attack 310SS >0.60 (23.5) Completely converted
in HF after 10 h at 850 °C (1560 °F), as shown to fluorides
Carpenter 20 >0.66 (26) Completely converted
in Fig. 6.63 (Ref 70). The results of Zotikov and to fluorides
Semenyuk (Ref 73) indicated that both molyb- Haynes alloy No. 25 >0.34 (13.5) Completely converted
denum and tungsten were quite resistant to to fluorides
Cobalt >1.52 (60) Completely converted
corrosion attack by HF at temperatures from to fluorides
300 to 800 °C (up to 700 °C for tungsten), as (a) NM, not measurable. Source: Ref 66
illustrated in Table 6.34. Molybdenum appears
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:29PM Plate # 0 pg 189

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 189

to improve the corrosion resistance of nickel- conducted in Ar-5HF, Ar-15HF, and Ar-35HF
base alloys in HF. The results published by mixtures. In nickel-base alloys, alloys B-3 (Ni-
International Nickel Company (Ref 74) showed 28Mo-1.5Cr) and 242 (Ni-25Cr-8Mo) were
that alloy N (Ni-16Mo-5Cr) and alloy B (Ni- found to be significantly better than alloys 600,
25Mo) were slightly better than nickel in HF 617, and 602CA. His test results are summarized
(Table 6.35). Field testing in an HF-bearing in Table 6.36 and Fig. 6.72. Both Ni-Mo alloys
environment showed that alloy N and alloy S containing low Cr formed thin surface scales
(Ni-16Cr-15Mo) were better than alloys 625 and with very little internal fluoride penetration.
C-22 at 900 °C (1650 °F) (Ref 75). Surprisingly, alloy 188, a cobalt-base alloy with
Barnes (Ref 76) investigated the corrosion high chromium (22%) and high tungsten (14%),
behavior of a wide range of materials in HF also exhibited good corrosion resistance with
environments at 1000 °C (1832 °F). Tests were a thin surface scale and little internal fluoride
penetration. Nickel aluminide intermetallic
Table 6.29 Corrosion of several metals and (Ni-8Al-8Cr-1.4Mo-1.7Zr) was found to exhibit
alloys in fluorine(a) good corrosion resistance in Ar-5HF. However,
Test temperature, Weight gain, Corrosion rate, when tested in Ar-35HF, the nickel-aluminide
Metal °C (°F) Time, h mg/cm2 mm/y (mils/y)
Ni(b) 550 (1,020) 6.2 1.9 0.11 (4.4)
600 550 (1,020) 5.0 706 81 (3,200) Table 6.32 Corrosion of nickel and nickel-base
Cu 550 (1,020) 2.42 17.4 2.8 (110) alloys by various volatile fluorides at 500 °C
Ni(b) 650 (1,200) 5.28 21.5 1.5 (59) (930 °F)
400 650 (1,200) 6.0 16.0 1.0 (41)
600 650 (1,200) 5.6 1,743 180 (7,100) Test Corrosion rate(a), mm/yr (mpy)
Corrosive duration,
Cu 650 (1,200) 4.8 134 10.9 (430)
environment h Ni-200 Alloy 400 Alloy 600
Ni(b) 750 (1,380) 4.1 100 9.0 (353)
400 750 (1,380) 6.1 −1, 831 74 (2,900) GeF4 + F2 8.4 Nil(b) 0.22 (8.8) 29 (1139)
600 750 (1,380) 4.7 4, 907 610 (24,000) AsF5 7.0 0.22 (8.8) Nil Nil
600 750 (1,380) 4.7 −12, 220 660 (26,000) AsF5 + F2 7.2 0.44 (17.5) 0.67 (26.3) 45 (1770)
Cu 750 (1,380) 5.8 278 19 (750) SeF6 7.0 0.22 (8.8) 0.22 (8.8) 0.44 (17.5)
SeF6 29.5 (c) (c) …
(a) Tests were conducted in flowing gas (30 to 130 cc/min). (b) Nickel “A”
SeF6 + F2 6.6 0.22 (8.8) 0.22 (8.8) 22 (850)
(commercial grade pure nickel). Source: Ref 67
MoF6 9.2 0.22 (8.8) Nil 0.22 (8.8)
MoF6 + F2 6.0 0.22 (8.8) Nil 44 (1726)
MoF6 + F2 7.0 (c) (c) 63 (2488)
TeF6 18.9 8.9 (350) 1.8 (70) 2.4 (96)
Table 6.30 Corrosion of several alloys TeF6 5.7 135 (5326) 56 (2190) 5.6 (219)
TeF6 + F2 7.8 0.22 (8.8) 0.22 (8.8) 40 (1568)
in fluorine(a) SF6 28.7 Nil Nil Nil
Corrosion rate, mm/yr (mpy) UF6 28.8 0.22 (8.8) … …
F2 7.0 0.66 (26) 0.22 (8.8) 9.8 (385)
Exposure 200 °C 370 °C 540 °C
Alloy time, h (400 °F) (700 °F) (1000 °F)
F2 8.6 … … 31 (1209)

400 5 0.013 (0.5) 0.048 (1.9) 0.76 (29.8) (a) Calculated from weight loss after descaling. (b) Rates reported as nil are
less than 0.001 mils/h. (c) Scale was not completely removed by descaling.
24 0.013 (0.5) 0.043 (1.7) 0.29 (11.3)
Source: Ref 69
120 0.003 (0.1) 0.031 (1.2) 0.18 (7.2)
Ni-200 5 0.084 (3.3) 0.043 (1.7) 0.62 (24.5)
24 0.013 (0.5) 0.031 (1.2) 0.41 (16.1)
120 0.003 (0.1) 0.010 (0.4) 0.35 (13.8)
304 5 0.155 (6.1) 40 (1565) …
304L 24 0.191 (7.5) 153 (6018) …
Table 6.33 Corrosion of various metals and
120 0.65 (25.4) … … alloys in anhydrous HF
347 5 0.102 (4.0) 108 (4248) … Corrosion rate, mm/yr (mpy)
Illium “R” 5 0.152 (6.0) 0.32 (12.7) 103 (4038)
500 °C 550 °C 600 °C
600 5 0.015 (0.6) 2.0 (78.0) 88 (3451)
Material (930 °F) (1020 °F) (1110 °F)
(a) Tests were conducted in flowing fluorine. Source: Ref 68 Nickel 0.9 (36) … 0.9 (36)
400 1.2 (48) 1.2 (48) 1.8 (72)
600 1.5 (60) … 1.5 (60)
Copper 1.5 (60) … 1.2 (48)
Aluminum 4.9 (192) … 14.6 (576)
Table 6.31 Corrosion of aluminum in fluorine Magnesium 12.8 (504) … …
Carbon steel (1020) 15.5 (612) 14.6 (576) 7.6 (300)
Test temperature, °C (°F) Corrosion rate, mm/yr (mpy) 304 … … 13.4 (528)
26 (79) 0.00087 (0.03) 347 183 (7,200) 457 (18,000) 177 (6,960)
201 (394) 0.0003 (0.01) 309Cb 5.8 (228) 43 (1,680) 168 (6,600)
356 (673) 0.57 (22) 310 12. 2 (480) 100 (3,960) 305 (12,000)
543 (1009) 2.2 (87) 430 1.5 (60) 9.1 (360) 11.6 (456)
Source: Ref 68 Source: Ref 61
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:29PM Plate # 0 pg 190

190 / High-Temperature Corrosion and Materials Applications

suffered extensive corrosion attack, while alloy Ni270, and copper. The test results on pure
242 remained resistant under the same test con- metals are summarized in Table 6.37 (Ref 76).
dition. The results indicated that Ni-Al system Precious metals (Au, Pt, and Pd) showed no
was not as good as Ni-Mo system in resisting HF weight changes and little changes in specimen
corrosion attack. surface appearance after testing. However, both
Barnes (Ref 76) also tested various pure gold and platinum specimens were found to
metals, which included gold (Au), platinum exhibit surface etching on the specimen surface.
(Pt), palladium (Pd), chromium, Ni200, Ni201, When tested in a higher concentration of HF
(Ar-35HF), palladium was heavily corroded with
some molten corrosion product formed on the
specimen surface. There was no discussion in
the paper about the performance of gold and
platinum in Ar-35HF mixture. Copper was also
found to exhibit few changes in specimen
appearance after testing in Ar-5HF. In fact, the
copper specimen retained its original metallic
luster after testing in Ar-5HF. Chromium was
found to be quite susceptible to fluoridation
attack when tested in Ar-5HF. The metal suffered
extensive internal chromium fluoride penetration
and extensive weight loss due to vaporization
of chromium fluorides.
Nickel was found to be extremely resistant
to HF corrosion attack. Three different types of
nickel were tested: Ni200 and Ni201 (99% pure
nickel), and Ni270 (99.9% pure nickel) (Ref 76).
Ni201 is a low carbon nickel (0.02% C), while
Ni200 is a high carbon nickel (0.15% C). The
Fig. 6.63 Corrosion of various metals in HF and HF-50H2O impurities and minor elements in these three
at 850 °C (1562 °F) for 10 h. Source: Ref 70 nickel specimens in the test program are: 0.15C,
0.4Fe, 0.35Si, 0.35Mn, and 0.25Cu for Ni200;
0.02C, 0.4Fe, 0.35Si, 0.35Mn, and 0.25Cu for
Ni201; and 0.02C, 0.001Cu, and 0.05 max Fe
for Ni270. The test results after 15 h in Ar-5HF
are summarized in Table 6.37 (Ref 76). All nickel
specimens retained metallic luster appearance
after testing. A nickel wire that had been used
for holding test specimens during testing of
specimens in the Ar-5HF test gas at 1000 °C for
1600 h was sectioned for metallographic exam-
ination, showing only a penetration of about
20 µm (0.8 mils) of internal nickel fluoride
precipitates. These nickel fluorides, which were
randomly distributed, were found to be less than
2 µm in diameter. Some nickel fluorides were
also detected on the metal surface.
Ceramic materials and graphite were also
tested by Barnes (Ref 76) under the same test
conditions. The results are summarized in
Table 6.38. Graphite was found to be extremely
resistant to HF corrosion. Ceramic materials,
such as alumina (polycrystalline or single
crystal), sintered silicon carbide, chemical vapor
deposited (CVD) silicon carbide, and silicon
Fig. 6.64 Effect of chromium on resistance to HF at 650 °C
(1200 °F) for Ni-Cr alloys. Source: Ref 71 nitride, suffered high corrosion rates.
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:29PM Plate # 0 pg 191

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 191

Fig. 6.65 Scanning electron backscattered image (a) and x-ray maps for Cr (b), Ni (c), and F (d) for Ni-40Cr alloy tested at 650 °C
(1200 °F) for 22 h in anhydrous HF environment. Source: Ref 72. Courtesy of Glyn Marsh

Fig. 6.66 Optical micrograph showing corrosion attack of alloy 600 after testing in HF for 92 h at 650 °C (1200 °F). Source: Ref 72.
Courtesy of Glyn Marsh
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:29PM Plate # 0 pg 192

192 / High-Temperature Corrosion and Materials Applications

Adding H2O to HF appears to make HF less

corrosive. Figure 6.63 shows the corrosion data
generated by Tyreman and Elliott (Ref 70) by

Fig. 6.69 Optical micrograph showing corrosion attack

on alloy 601 after testing in HF for 16 h at 650 °C
(1200 °F). Source: Ref 72. Courtesy of Glyn Marsh

Fig. 6.67 Corrosion kinetics of alloys 625, X750, and 601 as a

function of time in HF at 650 °C (1200 °F). Source:
Ref 71

Fig. 6.70 Optical micrograph showing very little corrosion

attack (about 1 to 2 µm in depth) on alloy N
(Ni-5Cr-16Mo) after testing in HF for 46 h at 650 °C (1200 °F).
Source: Ref 72. Courtesy of Glyn Marsh

Fig. 6.71 Scanning electron micrograph showing the

corrosion scale (mainly NiF2 with some CrF3, as
Fig. 6.68 Optical micrograph showing corrosion attack on determined by x-ray diffraction analysis) formed on alloy N after
alloy 625 after testing in HF for 142 h at 650 °C testing in HF for 115 h at 650 °C (1200 °F). Source: Ref 72.
(1200 °F). Source: Ref 72. Courtesy of Glyn Marsh Courtesy of Glyn Marsh
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:29PM Plate # 0 pg 193

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 193

comparing the 100% HF and HF-50%H2O removal rates (or corrosion rates) were more than
environments for nickel, copper, cobalt, iron, 1 mm/h (40 mils per h (Ref 77). Macheteau et al.
chromium, molybdenum, niobium, and tanta- (Ref 78) also found that oxygen contamination
lum. All metals except molybdenum showed accelerated fluoridation attack of iron. Marsh and
more corrosion attack in HF than in HF-50H2O. Elliott (Ref 79) observed that cobalt exhibited a
Corrosion data generated in HF-10H2O and HF- protective CoF2 film with a very low corrosion
50%H2O are summarized in Tables 6.39 and rate when exposed to HF at 650 °C (1200 °F).
6.40, respectively. However, once air was introduced to mix with
the test gas of HF (i.e., HF-O2-N2 mixture), the
corrosion rate increased significantly. This is
6.4.3 Corrosion in O2-HF Environments illustrated in Fig. 6.73 (Ref 79). The rapid
Oxygen appears to make HF more corrosive. corrosion attack of cobalt when air was mixed
During nuclear fuel reprocessing, stainless steel with HF was associated with the formation of
fuel cladding was being chemically removed by numerous Co3O4 oxide protrusions, as shown in
reacting it with an O2-HF mixture (40–60% HF) Fig. 6.74. At the Co3O4 oxide and the cobalt
at 377 to 627 °C (710 to 1160 °F) (Ref 77). The metal interface, no protective fluoride films were
observed except remnant segments of CoF2 films
were still present as indicated the lower line in
Table 6.34 Corrosion of molybdenum and Fig. 6.74. It is believed that as soon as the air
tungsten in HF containing 0.6% H2O was introduced to mix with HF gas stream, a
Corrosion rate, mm/yr (mpy) protective CoF2 film was “broken,” prompting
Test temperature,
°C (°F) Molybdenum Tungsten the formation of Co3O4 oxide protrusions.
300 (570) 0.003 (0.1) 0.003 (0.1) Schutze and Simon (Ref 80) tested a number of
400 (750) 0.011 (0.4) 0.009 (0.4) alloys in N2-5O2-3HF at 1100 °C for 75 h in
500 (930) 0.014 (0.6) 0.017 (0.7)
600 (1110) 0.023 (0.9) 0.022 (0.9)
700 (1290) 0.144 (5.7) 0.91 (36)
800 (1470) 1.3 (51) …
Source: Ref 73

Table 6.35 Corrosion of various nickel-base

alloys in HF at 500 to 600 °C (930 to 1110 °F)
Alloy Corrosion rate, mm/yr (mpy)
Alloy C 0.008 (0.3)
Alloy 600 0.018 (0.7)
Alloy B 0.051 (2)
Ni-200 0.229 (9)
Ni-201 0.356 (14)
Alloy 400 0.330 (13)
Alloy K-500 0.406 (16)
70Cu-30Ni 0.406 (16) Fig. 6.72 Depth of internal fluoride penetration for alloys
tested in Ar-5HF at 1000 °C (1832 °F) for 15 h.
Tests were performed in HF (7 lb/h) for 36 h. Source: Ref 74 Source: Ref 76

Table 6.36 Corrosion behavior of nickel-base alloys along with one cobalt-base alloy (alloy 188)
and nickel aluminide Intermetallic (IC221M) in Ar-5HF at 1000 °C (1832 °F) for 15 h
Maximum attack(a),
Alloy mm/side (mils/side) Specimen appearance Comments
Alloy 242 0.015 (0.6) Surface scale Good resistance at 35HF
Alloy B-3 0.02 (0.8) Surface scale …
IC221M 0.025 (1.0) Surface scale Extensive corrosion at 35HF(b)
Alloy 188 0.051 (2.0) Surface scale …
Alloy 617 0.13 (5.3) Nodules on surface Extensive internal fluorides
Alloy 600 0.15 (5.9) Nodules on surface Extensive internal fluorides
Alloy 602CA 0.18 (7.1) Molten corrosion product Extensive internal fluorides
(a) Surface metal loss + internal fluoride penetration. (b) IC221M exhibited good resistance in Ar-5HF, but poor resistance in Ar-35HF. The specimen was almost completely
converted to fluorides. Source: Ref 76
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:29PM Plate # 0 pg 194

194 / High-Temperature Corrosion and Materials Applications

Table 6.37 Corrosion behavior of pure metals in Ar-5HF at 1000 °C (1832 °F) for 15 h
Materials Weight change, mg/cm2 Corrosion rate, mpy Specimen appearance Comments
Gold 0 0.0 Unchanged …
Platinum −0.2 1.1 Matte grey …
Palladium 0.2 … Unchanged Heavily corroded in 35HF
Chromium(a) −35.25 1454(b) Green Total depth of attack = 8.3 mils
Ni200 −0.1 1.4 Unchanged No internal attack
Ni201 −0.1 1.4 Unchanged No internal attack
Ni270 0.0 0.0 Unchanged No internal attack
(a)Tested for 50 h at 1000 °C in Ar-5HF. (b) Based on total depth of attack (maximum metal affected) of 8.3 mils/side in a 50 h test. Source: Ref 76

Table 6.38 Corrosion behavior of ceramic Table 6.39 Corrosion of several alloys in
materials and graphite in Ar-5HF at 1000 °C HF-10H2O at 850 °C (1560 °F)
(1832 °F) for 15 h Alloy Corrosion rate, mm/yr (mpy)
Weight 304LSS 100–130 (4000–5000)
change, Corrosion 310SS 100–130 (4000–5000)
Materials mg/cm2 rate, mpy Comments
800H 100–130 (4000–5000)
Graphite (CZR-1) 0 0.0 Unchanged 600 23 (900)
Graphite (DFP-2) 0 0.0 Unchanged 625 6.7 (265)
Alumina (polycrystalline) −2.4 141 General wastage
Source: Ref 70
Alumina (sapphire) −0.9 52 Frosted appearance
CVD SiC −5.7 411 General wastage
Sintered SiC −6.4 458 General wastage
Si3N4 −2.2 155 General wastage
Source: Ref 76
Table 6.40 Corrosion of nickel and Alloy 400 in
50HF-50H2O at various temperatures
Corrosion rate, mm/yr (mpy)
their search for a candidate alloy for air nozzles 550 °C 600 °C 650 °C 700 °C 750 °C
in a circulating fluidized-bed combustor for re- (1020 °F) (1110 °F) (1200 °F) (1290 °F) (1380 °F)
Nickel 0.79 (31) 1.83 (72) 2.74 (108) 3.66 (144) 3.05 (120)
processing spent pot lining material in aluminum Alloy 400 … 0.61( 24) 1.52 (60) 3.96 (156) 5.18 (204)
production. Among the metallic materials tested, Source: Ref 61
alloy 242 was found to exhibit the smallest
metal loss (50 µm). Although the alloy suffered
extensive nitridation attack, the authors con-
sidered that nitridation attack would not affect iodine environments at elevated temperatures.
the alloy’s performance as air nozzles. In a low- Miller et al. (Ref 82) investigated the corrosion
temperature test (450 °C) in N2-8O2-10CO2- behavior of copper, nickel, and nickel-base
15H2O-5HF, Crum et al. (Ref 81) observed no alloys in bromine at 300 and 500 °C (470 and
measurable corrosion attack for many nickel- 930 °F) (Table 6.42). Copper suffered rapid
base alloys and a superaustenitic stainless steel attack by bromine at 300 °C (570 °F). The CuBr
after 155 h of exposure. Their test results are compound exhibits very high vapor pressures.
summarized in Table 6.41. Stress-corrosion The vapor pressure of CuBr reaches 1× 10−4 atm
cracking (SCC) resistance of these alloys was (a value considered high enough to cause high-
also included in the test program using the same temperature corrosion) when the temperature is
test environment with U-bend test specimens. as low as 435 °C (815 °F) (Table 6.3). Nickel
All alloys showed no cracking after 100 and exhibited good corrosion resistance in Br2 at
155 h except alloy 600. Authors did not explain 300 and 500 °C (470 and 930 °F). Bromine
why alloy 600 suffered SCC while other alloys reacts with nickel to form NiBr2, which melts at
including many nickel-base alloys and one 965 °C (1769 °F), significantly higher than the
superaustenitic stainless steel (alloy 25-6MO) melting point of CuBr. Duranickel 301 (Ni-5Al)
showed no cracking. had a corrosion resistance similar to nickel.
Alloy 400 was less resistant than nickel and
Duranickel 301.
6.5 Corrosion in Bromine and Iodine Smith and Ganesan (Ref 83) investigated the
Environments corrosion behavior of various commercial alloys
in a simulated combustion environment con-
Very little data have been reported on the taining a very high level of HBr (about 4%) at
performance of metals and alloys in bromine and 593 and 927 °C (1100 and 1700 °F). The test
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:30PM Plate # 0 pg 195

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 195

gas (the inlet gas) consisted of 9% O2, 12% CO2, 347), alloy 800, and nickel-base alloys. At
4% HBr, 100 ppm SO2, and balance N2. The 927 °C (1700 °F), nickel-base alloys showed
equilibrium partial pressures of the test environ- little weight loss, while austenitic stainless steels,
ment at 593 and 927 °C (1100 and 1700 °F) are Fe-Ni-Cr alloy 800, and high-iron-containing
listed in Table 6.43. The test results, which are nickel alloy 825 suffered high weight losses.
summarized in Table 6.44, showed surprisingly The authors (Ref 83) analyzed the corrosion
very little corrosion attack at 593 °C (1100 °F) products spalled from test specimens as well as
for all the alloys tested, which included
austenitic stainless steels (Type 309, 316, and
Table 6.41 Corrosion rates generated from
tests in N2-8O2-10CO2-15H2O-5HF at 450 °C
(842 °F) for 100 and 155 h
Corrosion rate, Corrosion rate,
mm/yr (mpy) mm/yr (mpy)
Alloy from 100 h tests from 155 h tests
Alloy 600 0.01 (0.4) 0
Alloy 400 0 0.002 (0.1)
Alloy 825 0.01 (0.4) 0
Alloy C-276 0 0
Alloy 686 0 0
Alloy 622 0 0
Alloy 25-6MO 0 0
Source: Ref 81

Table 6.42 Corrosion of several alloys in

bromine at 300 and 500 °C (570 and 930 °F)
Temperature, Time, Weight loss, Corrosion rate,
Alloy °C (°F) days mg/cm2 mm/yr (mpy)
Ni-201 300 (570) 11 0.28 <0.03 (1.2)
300 (570) 11 0.16 <0.03 (1.2)
300 (570) 11 0.14 <0.03 (1.2)
Fig. 6.73 Weight change of cobalt at 650 °C (1200 °F) when Duranickel 301 300 (570) 10 0.17 <0.03 (1.2)
exposed to HF during the first 12 h, showing a very 300 (570) 10 0.07 <0.03 (1.2)
low corrosion rate. At time “X,” air was introduced into the test gas 400 300 (570) 2 7.07 1.5 (60)
to mix with HF, rapid corrosion attack was observed. Source: 300 (570) 2 6.65 1.5 (60)
Ref 79 Copper 300 (570)(a) <1 Rapid attack …
300 (570)(a) <1 Rapid attack …
Ni-201 500 (930) 10 3.83 0.15 (6.0)
500 (930) 10 1.40 0.06 (2.4)
Duranickel 301 500 (930) 10 2.14 0.09 (3.6)
500 (930) 10 3.64 0.15 (6.0)
(a) Tests were terminated before 300 °C was reached because of rapid corrosion
attack. Source: Ref 82

Table 6.43 Thermodynamic equilibria in terms

of partial pressures (atm) for the inlet gas of
N2-9O2-12CO2-4HBr-100 ppm SO2 at 593 and
927 °C (1100 and 1700 °F)
Partial pressure Partial pressure
Gas component at 593 °C, atm at 927 °C, atm
N2 0.78 0.78
CO2 0.12 0.12
O2 0.81 × 10−1 0.81 × 10−1
Br2 0.20 × 10−1 0.19 × 10−1
H 2O 0.20 × 10−1 0.19 × 10−1
HBr 0.11 × 10−3 0.22 × 10−2
Fig. 6.74 A Co3O4 oxide protrusion formed on cobalt
SO3 0.90 × 10−4 0.82 × 10−5
immediately after test gas was switched from HF
to air-HF mixture at 650 °C. The upper line indicates the Co3O4 SO2 0.63 × 10−6 0.82 × 10−4
oxide protrusion, and the lower line indicates the CoF2 phase H2 0.39 × 10−14 0.82 × 10−9
formed at the interface between the Co3O4 oxide and the substrate Source: Ref 83
cobalt. Source: Ref 72. Courtesy of Glyn Marsh
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:30PM Plate # 0 pg 196

196 / High-Temperature Corrosion and Materials Applications

the extraction residue obtained from the test phases were detected. The results by Smith and
specimens using x-ray diffraction technique to Ganesan (Ref 83) suggest that nickel-base alloys
determine the phases present. The results of the appear to be suitable for applications in HBr-
x-ray diffraction analysis are summarized in containing environments up to about 4%. It is
Table 6.45 for the corrosion scales spalled from also suggested that HBr is not as corrosive as HCl
tested specimens and in Table 6.46 for the at elevated temperatures. More tests, particularly
extraction residue obtained from tested speci- long-term tests, are needed to confirm the current
mens by electrolytic extraction of phases. In both findings.
cases, the phases found were oxides. No bromide Corrosion of various alloys in I2 was investi-
gated by Shapiro (Ref 84) when determining the
most suitable alloy for construction of vessels
Table 6.44 Mass change data for alloys tested for iodine zirconium processing in the pro-
in N2-9O2-12CO2-4HBr-100 ppm SO2 at 593 duction of zirconium. Tests were conducted at
and 927 °C (1100 and 1700 °F) for 300 h 300 and 450 °C (570 and 840 °F) in iodine vapor
Mass change, mg/cm2 with 400 mm Hg (0.53 atm) pressure. Results
Alloy 593 °C (1100 °F) 927 °C (1700 °F) are shown in Table 6.47 and 6.48. Platinum, gold,
309 0.19 −218.70 tungsten, and molybdenum were very resistant
316 0.75 −83.35 to corrosion attack in I2. Alloy B (Ni-Mo), alloy
347 0.59 −67.57
800 0.55 −53.25
C (Ni-Cr-Mo), alloy 600 (Ni-Cr-Fe), and nickel
825 −0.11 −28.57 were quite resistant to I2. Stainless steels and
625 0.10 −8.77 alloy 400 were less resistant. In the production
601 0.56 −7.78
600 0.55 −6.04 of zirconium using the iodide process, closure
617 0.15 −5.57 gaskets made of gold are used for handling dry
690 0.50 −3.88 iodine vapors at 500 °C (939 °F) (Ref 85).
Source: Ref 83

Table 6.47 Corrosion rates of several metals

Table 6.45 Results of x-ray diffraction analysis and alloys in iodine at 300 °C (570 °F)(a)
on the spalled corrosion products for the test Alloy Corrosion rate, mm/yr (mpy)
specimens after testing in N2-9O2-12CO2-4HBr- Platinum 0
100 ppm SO2 at 927 °C (1700 °F) for 300 h Gold 0
Tungsten 0
Alloy Major phases Minor phases Molybdenum 0.003 (0.11)
316 Fe2O3 NiCr2O4, M3O4 Tantalum 0.004 (0.16)
800 NiFe2O4, Cr2O4 Fe2O3 Alloy B 0.044 (1.7)
825 Cr2O3, NiCr2O4 … Alloy C 0.057 (2.2)
600 Cr2O3, NiFe2O4 … Alloy 600 0.107 (4.2)
601 Cr2O3 M 3O 4 Nickel 0.27 (10.6)
617 Cr2O3 M 3O 4 Alloy 400 0.56 (22.0)
625 Cr2O3 M 3O 4 (a)400 mm Hg (0.53 atm) iodine pressure. Source: Ref 84
690 Cr2O3, Fe2O3 M 3O 4
Source: Ref 83

Table 6.48 Corrosion rates of several metals

and alloys in iodine at 450 °C (840 °F)(a)
Table 6.46 Results of x-ray diffraction Alloy Corrosion rate, mm/yr (mpy)
analysis on the extracted residue from the test Platinum 0.0055 (0.2)
specimens after testing in N2-9O2-12CO2-4HBr- Tungsten 0.008 (0.3)
100 ppm SO2 at 927 °C (1700 °F) for 300 h Gold 0.024 (0.9)
Molybdenum 0.033 (1.3)
Alloy Major phases Minor phases
Alloy B 0.46 (18)
316 Cr2O3, M3O4 … Alloy 600 0.54 (21)
800 NiFe2O4, Cr2O3 NiO Tantalum 0.88 (35)
825 Cr2O3 NiFe2O4 Nickel 1.2 (47)
600 NiCrO3 NiO Type 310SS 1.8 (71)
601 Al2O3, Cr2O3 … Type 316SS 2.1 (83)
617 Cr2O3 M3O4 Type 347SS 2.1 (83)
625 Cr2O3, NiCr2O4 … Alloy 400 2.6 (102)
690 Cr2O3 NiFe2O4 Type 304SS 3.2 (126)
Source: Ref 83 (a)400 mm Hg (0.53 atm) iodine pressure. Source: Ref 84
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:30PM Plate # 0 pg 197

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 197

6.6 Summary 7. T. Rosenquist, Principles of Extractive

Metallurgy, McGraw-Hill, 1974
High-temperature corrosion behavior of 8. P.L. Daniel and R.A. Rapp, Advances in
metals and alloys in gaseous environments con- Corrosion Science and Technology, Vol 5,
taining (a) chlorine and hydrogen chloride, (b) M.G. Fontana and R.W. Staehle, Ed., Ple-
fluorine and hydrogen fluoride, (c) bromine and num Press, 1970, p 55
hydrogen bromide, and (d) iodine and hydrogen 9. O. Kubaschewski and E. Evans, Metallur-
iodide is reviewed. Metals or alloys in these gical Thermochemistry, Pergamon Press,
environments form metal halides as corro- 1958
sion products. Many metal halides exhibit high 10. HSC, Chemistry for Windows, Version 6.0,
vapor pressures; the corrosion products can A. Roine, Outokumpu Technology, Finland,
thus be in a gaseous phase. Some metal halides www.outokumpotechnology.com, accessed
also exhibit low melting points; the corrosion Dec 2006
products can thus be in a liquid phase. As a result, 11. ChemSage, Version 4.16, GTT-Techno-
the metal-halogen or metal-hydrogen-halide logies, Aachen, 1998
reactions can proceed at a rapid rate at elevated 12. M.J. McNallan, W.W. Liang, J.M. Oh,
temperatures. and C.T. Kang, Morphology of Corrosion
The corrosion behavior of alloys can be Products Formed on Cobalt and Nickel
significantly different between “oxidizing” in Argon-Oxygen-Chlorine Mixtures at
environments that contain O2 and reducing 1000 °K, Oxid. Met., Vol 17, 1982, p 371
environments that contain no O2. Also, corrosion 13. A. Zahs, M. Spiegel, and H.J. Grabke, The
by halogen or by hydrogen halide can proceed at Influence of Alloying Elements on the
different rates. Accordingly, the corrosion data Chlorine-Induced High Temperature Corro-
are categorized in different environment types. sion of Fe-Cr Alloys in Oxidizing Atmo-
The arrangement of the corrosion data in this way spheres, Mater. Corros., Vol 50, 1999, p 561
may also help readers to determine the type of 14. F.H. Stott and C.Y. Shih, The Influence of
data that may be more pertinent to their intended HCl on the Oxidation of Iron at Elevated
applications or conditions. For the corrosion Temperatures, Mater. Corros., Vol 51, 2000,
behavior of alloys in environments containing p 277
Cl2 and HCl, the data are presented in (a) Cl2 15. R. Bender and M. Schutze, The Role of
environments containing no O2, (b) O2-Cl2 Alloying Elements in Commercial Alloys
mixtures, (c) O2-HCl mixtures, and (d) HCl and for Corrosion Resistance in Oxidizing-
HCl-bearing, reducing environments. For corro- Chloridizing Atmospheres—Part I: Litera-
sion in F2- and HF-bearing environments, the ture Evaluation and Thermodynamic
data are presented in (a) F2 environments, (b) Calculations on Phase Stabilities, Mater.
HF environments, and (c) O2-HF mixtures. For Corros., Vol 54, 2003, p 567
bromine- and iodine-containing environments 16. K.L. Tseitlin and V.A. Strunkin, J. Appl.
the corrosion data are very limited, and a brief Chem. USSR, Vol 31, 1958, p 1832
review is presented. 17. K.L. Tseitlin, J. Appl. Chem. USSR, Vol 28
(No. 5), 1955, p 467
18. G. Heinemann, F.G. Garrison, and P.A.
REFERENCES Haber, Ind. Eng. Chem., Vol 38, 1946, p 497
19. R.J. Fruehan, Metall. Trans., Vol 3 (No. 10),
1. W.J. Kroll, Method of Manufacturing Ti 1972, p 2585
and Alloys Thereof, U.S. Patent 2205854, 20. S.F. Bohlken, A. Klinkenberg, and
Jan 1940 H.W. Nicolai, Ind. Chim. Belge., Vol 20,
2. W.A. Henderson, J. Met., Vol 16, 1964, p 155 1955, p 579
3. S.M. Shelton, The Metallurgy of Zirconium, 21. S.F. Bohlken, A. Klinkenberg, and H.W.
B. Lustman and F. Kerze, Jr., Ed., McGraw- Nicolai, C. R. Cong. Int. Chim. Ind., Vol 27,
Hill, 1955, p 59 1954, p 2
4. I. Iwasaki, Y. Takahasi, and H. Kahata, 22. M.H. Brown, W.B. DeLong, and J.R. Auld,
Trans. SME AIME, Vol 243, 1966, p 308 Ind. Eng. Chem., Vol 39 (No. 7), 1947, p 839
5. C.L. Mantell, Tin, Reinhold, 1949 23. G. Han and W.D. Cho, High-Temperature
6. W.E. Berry, Corrosion in Nuclear Applica- Corrosion of Fe3Al in 1% Cl2/Ar, Oxid.
tions, John Wiley & Sons, 1971 Met., Vol 58, 2002, p 390
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:30PM Plate # 0 pg 198

198 / High-Temperature Corrosion and Materials Applications

24. B.J. Downey, J.C. Bermel, and P.J. Energy Systems, M.F. Rothman, Ed., The
Zimmer, Corrosion, Vol 25 (No. 12), 1969, Metallurgical Society of AIME, 1985,
p 502 p 483
25. J.D. McKinley, Jr. and K.E. Shuler, J. Chem. 39. M.J. McNallan, M.H. Rhee, S. Thongtem,
Phys., Vol 28, 1958, p 1207 and T. Hensler, Paper No. 11, Corrosion/85,
26. R.H. Kane, Process Industries Corrosion, NACE, 1985
B.J. Moritz and W.I. Pollock, Ed., NACE, 40. P. Elliott, A.A. Ansari, R. Prescott, and
1986, p 45 M.F. Rothman, Paper No. 13, Corrosion/85,
27. MTI Publication MS-3, Materials Selector NACE, 1985
for Hazardous Chemicals, Vol 3, Hydro- 41. S. Thongtem, M.J. McNallan, and G.Y. Lai,
chloric Acid, Hydrogen Chloride and Paper No. 372, Corrosion/86, NACE, 1986
Chlorine, Materials Technology Institute 42. C. Schwalm and M. Schutze, The Corro-
of the Chemical Process Industries, Inc., sion Behavior of Several Heat Resistant
C.P. Dillon and W.I. Pollock, Ed., 1999 Materials in Air + 2 vol-% Cl2 at 300 to
28. J.P. Tu, Z.Z. Li, and Z.Y. Mao, Internal 800 °C: Part 1—Fe-Base and Fe-Containing
Chlorination of Ni-Based Alloys and Its Alloys, Mater. Corros., Vol 51, 2000, p 34
Relation to Volatilization Corrosion, Mater. 43. C. Schwalm and M. Schutze, The Corrosion
Corros., Vol 48, 1997, p 441 Behavior of Several Heat Resistant Materi-
29. M.J. Maloney and M.J. McNallan, Met. als in Air +2% Cl2 at 300 to 800 °C: Part
Trans. B, Vol 16, 1985, p 751 2—Nickel Base Alloys, Mater. Corros.,
30. M.J. McNallan and W.W. Liang, Gaseous- Vol 51, 2000, p 73
Transport-Controlled Chlorination of CoO 44. C. Schwalm and M. Schutze, The Corrosion
in Flowing Oxygen-Chlorine-Argon Mix- Behavior of Several Heat Resistant Mat-
tures at 1000 °K, J. Am. Ceram. Soc., Vol 64 erials in Air + 2% Cl2 at 300 to 800 °C: Part
(No. 5), 1981, p 302 3—Alumina Formers and Intermetallics,
31. N.S. Jacobson, M.J. McNallan, and Mater. Corros., Vol 51, 2000, p 161
Y.Y. Lee, The Formation of Volatile Corro- 45. M.J. McNallan, S. Thongtem, J.C. Liu,
sion Products During the Mixed Oxidation- Y.S. Park, and P. Shyu, Corrosion of Chro-
Chlorination of Cobalt at 650 °C, Metall. mium Containing Alloys in Non-steady
Trans., Vol 17A, 1986, p 1223 State Environments Containing Oxygen,
32. Y.Y. Lee and M.J. McNallan, Ignition of Carbon, and Chlorine, J. Phys. IV, Coll. C9,
Nickel in Environments Containing Oxygen Suppl. J. Phys. III, Vol 3, Dec 1993, p 143
and Chlorine, Metall. Trans., Vol 18A, 1987, 46. M.J. McNallan, Z. Niemczura, H.H. Lu,
p 1099 and Y.S. Park, Carbide Formation and Oxi-
33. M.J. McNallan, High-Temperature Corro- dation of Austenitic Alloys in Oxidizing/
sion in Halogen Environments, MP, Sept. Carburizing Environments Contaminated
1994, p 54 with Chlorine, Paper No. 252, Corrosion/93,
34. J.C. Liu and M.J. McNallan, Effects of NACE, 1993
Temperature Variations on Oxidation of 47. Y. Ihara, H. Ohgame, K. Sakiyama, and
Iron-20% Chromium Alloys at 1200 °K K. Hashimoto, Corros. Sci., Vol 21 (No. 12),
in Ar-20% O2-Cl2 Gas Mixtures, Mater. 1981, p 805
Corros., Vol 50, 1999, p 253 48. Y. Ihara, H. Ohgame, and K. Sakiyama,
35. S. Baranow, G.Y. Lai, M.F. Rothman, J.M. Corros. Sci., Vol 22 (No. 10), 1982, p 901
Oh, M.J. McNallan, and M.H. Rhee, Paper 49. Y. Ihara, H. Ohgame, and K. Sakiyama,
No. 16, Corrosion/84, NACE, 1984 Corros. Sci., Vol 23 (No. 2), 1983, p 167
36. J.M. Oh, M.J. McNallan, G.Y. Lai, and M.F. 50. F. Devisme, P. Falgoux, F. Lefebvre, and
Rothman, Metall. Trans. A, Vol 17A, 1986, T. Flament, “High Temperature Corrosion
p 1087 in Atmospheres Containing Hydrogen
37. N.S. Jacobson, M.J. McNallan, and Y.Y. Chloride,” 11th International Incinera-
Lee, Mass Spectrometric Observations of tion Conference (Albuquerque, NM), May
Metal Oxychlorides Produced by Oxidation- 11–15, 1992
Chlorination Reactions, Metall. Trans. A, 51. P. Ganesan, G.D. Smith, and L.E.
Vol 20A, 1989, p 1566 Shoemaker, The Effects of Excursions of
38. M.H. Rhee, M.J. McNallan, and M.F. Oxygen and Water Vapor Contents on
Rothman, High Temperature Corrosion in Nickel-Containing Alloy Performance in
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:30PM Plate # 0 pg 199

Chapter 6: Corrosion by Halogen and Hydrogen Halides / 199

A Waste Incineration Environment, Paper 64. K. Hauffe, Z. Werkstofftech., Vol 15, 1984,
No. 248, Corrosion/91, NACE, 1991 p 427
52. G.D. Smith and P. Ganesan, Metallic 65. Y.A. Lukyanchev, I.I. Astakhov, and N.S.
Corrosion in Waste Incineration: A Look at Nikolaev, Mechanism of the Formation of
Selected Environmental and Alloy Funda- Fluoride Films on Nickel and Their Proper-
mentals, Heat-Resistant Materials II, Conf. ties, Bull. Acad. SSSR. Div. Chem. Sci.,
Proc. Second International Conference 1965, p 577
on Heat-Resistant Materials, K. Natesan, 66. C.F. Hale, E.J. Barber, H.A. Bernhardt, and
P. Ganesan, G. Lai, Ed., ASM International, K.E. Rapp, “High Temperature Corrosion of
1995, p 631 Some Metals and Ceramics in Fluorinating
53. P. Elliott, R. Prescott, and C.J. Tyreman, Atmospheres,” Report K-1459, Union Car-
Aspects of Halogen Corrosion at High bide Nuclear Co., Sept 1960
Temperatures, Paper No. 373, Corrosion/86, 67. M.J. Steindler and R.C. Vogel, “Corrosion
NACE, 1986 of Materials in the Presence of Fluorine
54. M.K. Hossain, J.E. Rhoades-Brown, S.R.J. at Elevated Temperatures,” ANL-5662,
Saunders, and K. Ball, Proc. UK Corrosion/ Argonne, IL, Jan 1957
83, 1983, p 61 68. R.B. Jackson, “Corrosion of Metals and
55. G.Y. Lai, M.F. Rothman, S. Baranow, and Alloys by Fluorine,” NP-8845, Allied
R.B. Herchenroeder, Recuperator Alloys Chemical Corp., Morristown, PA, 1960
for High-Temperature Waste Heat Recovery, 69. R.C. Vogel, J.H. Schraidt, and J. Royal,
J. Met., Vol 35 (No. 7), 1983, p 24 “Chemical Engineering Division Semi-
56. U. Brill, J. Klower, and D.C. Agarwal, Annual Report,” ANL-7125, Argonne
Influence of Alloying Elements on the National Laboratory, Argonne, IL, May
Chlorination Behavior of Nickel- and 1966
Iron-Based Alloys, Paper No. 439, Corro- 70. C.J. Tyreman and P. Elliott, Paper No. 135,
sion/96, NACE International, 1996 Corrosion/88, NACE, 1988
57. J.J. Barnes, Materials Limitations in High 71. G. Marsh and P. Elliott, High Tem-
Temperature Chlorination Environments, perature Corrosion in Energy Systems, M.F.
Paper No. 446, Corrosion/96, NACE Inter- Rothman, Ed., The Metallurgical Society of
national, 1996 AIME, 1985, p 467
58. K.N. Strafford, P.K. Datta, and G. Forster, 72. G. Marsh, Ph.D. thesis, UMIST, Manche-
High-Temperature Chloridation of Binary ster, U.K., Dec 1982
Fe-Cr Alloys at 1000 °C, High Temperature 73. V.S. Zotikov and E.Ya, Semenyuk, Metall-
Corrosion 2: Advanced Materials and Coat- schutz Moscow, Vol 6, 1970, p 218
ings, Proc. Second International Symposium 74. “Corrosion Resistance of Nickel-Containing
on High Temperature Corrosion of Ad- Alloys in Hydrofluoric Acid, Hydrogen
vanced Materials and Coatings (Les Embiez, Fluoride and Fluorine,” CEB-5, Inter-
France), May 22–26, 1989, R. Streiff, national Nickel Co., 1968
J. Stringer, R.C. Krutenat, and M. Caillet, 75. Haynes International, Inc., unpublished
Ed., Elsevier Science Publishers, London, results, 1988
1989, p 61 76. J.J. Barnes, Materials Behavior in High
59. J.P. Carter, B.S. Covino, Jr., T.J. Driscoll, Temperature HF-Containing Environments,
W.D. Riley, and M. Rosen, Corrosion, Paper No. 518, Corrosion/2000, NACE
Vol 40 (No. 5), 1984, p 205 International, 2000
60. L. Reeve, J. Iron Steel Inst., Sept 1955, p 26 77. G. Dumont and W. Goossens, Chemical
61. W.R. Myers and W.B. DeLong, Chem. Eng. Decladding of Fuel Elements with HF/O2
Prog., Vol 44 (No. 5), 1948, p 359 Mixture, Chem. Eng. Technol., Vol 43, 1971,
62. R.L. Jarry, J. Fischer, and W.H. Gunther, p 800
J. Electrochem. Soc., Vol 110 (No. 4), 1963, 78. Y. Macheteau, J. Gillardeau, P. Plurien, and
p 346 J. Oudar, Oxid. Met., Vol 4 (No. 3), 1972,
63. R.L. Jarry, W.H. Gunther, and J. Fischer, p 141
“The Mechanism and Kinetics of the 79. G. Marsh and P. Elliott, Aspects of High
Reaction Between Nickel and Fluorine,” Temperature Hydrofluorination of Cobalt,
ANL-6684, Argonne National Laboratory, High Temp. Technol., Vol 3 (No. 4), 1985,
Argonne, IL, Aug 1963 p 215
Name ///sr-nova/Dclabs_wip/High Temp/5208_147-200.pdf/Chap_06/ 26/10/2007 12:30PM Plate # 0 pg 200

200 / High-Temperature Corrosion and Materials Applications

80. M. Schutze and P. Simon, Materials for Use Bromine,” BMI-X-489, Battelle Memorial
in Atmospheres Containing HF in a Pyro- Institute, Columbus, Ohio, Jan 1968
hydrolysis Plant for Re-processing Spent Pot 83. G.D. Smith and P. Ganesan, Performance
Lining from Aluminum Production, Werkst. and Scale Formation of Selected High
Korros., Vol 45, 1994, p 435 Temperature Alloys in Simulated Waste
81. J.R. Crum, G.D. Smith, M.J. McNallan, and Incineration Environments Containing
S. Himyj, Characterization of Corrosion Gaseous Bromides and Chlorides, Paper No.
Resistant Materials in Low and High Tem- 406, Corrosion/87, NACE, 1987
perature HF Environments, Paper No. 382, 84. Z.M. Shapiro, Metallurgy of Zirconium, B.
Corrosion/99, NACE International, 1999 Lustman and F. Kerze, Jr., Ed., McGraw-
82. P.D. Miller, E.F. Stephan, W.E. Berry, and Hill, 1955, p 135
W.K. Boyd, “Corrosion Resistance of 85. Handbook of Corrosion Data, B.D. Craig,
Nickel and Two Nickel Alloys to Gaseous Ed., ASM International, 1989
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:13AM Plate # 0 pg 201

High-Temperature Corrosion And Materials Applications Copyright © 2007 ASM International®

George Y. Lai, editor, p201-234 All rights reserved.
DOI: 10.1361/hcma2007p201 www.asminternational.org

CHAPTER 7

Sulfidation

7.1 Introduction (2100 to 2200 °F). The product gas typically

contains about 10 to 15% SO2 along with 5 to
Sulfur is one of the most common corrosive 10% O2 (balance N2), which is then converted to
contaminants in high-temperature industrial SO3 for sulfuric acid.
environments. Sulfur is generally present as an In many industrial processes, combustion is
impurity in fuels or feedstocks. Typically, fuel carried out under stoichiometric or substoichio-
oils are contaminated with sulfur varying from metric conditions to convert feedstock to process
fractions of 1% (No. 1 or No. 2 fuel oil) to about gases consisting of H2, CO, CH4, and other
3% (No. 6 fuel oil). United States coal may hydrocarbons. Sulfur is converted to H2S. The
contain from 0.5 to 5% S, depending on where it environment, in this case, is reducing and is
is mined. Feedstocks for calcining operations in characterized by low oxygen potentials. Coal
mineral and chemical processing are frequently gasification, which converts coal to substitute
contaminated with various amounts of sulfur. natural gas or medium- and low-BTU fuel gases,
Sulfur is one of the corrosive species in petro- is an example of a process that generates this type
leum-refining operations. of atmosphere. Reducing conditions may also
When combustion takes place with excess air prevail in localized areas, in some cases, even
to ensure complete combustion of fuel for gen- when combustion is taking place with excess air.
erating heat in many industrial processes, sulfur A coal-fired boiler equipped with low NOx
in the fuel reacts with oxygen to form SO2 and combustion burners is an example. More dis-
SO3. A combustion atmosphere of this type is cussion on this issue is presented in Chapter 10
oxidizing in nature. Oxidizing environments on high-temperature corrosion and other
are usually much less corrosive than reducing materials issues in coal-fired boilers. Further-
environments, where sulfur is in form H2S. more, ash deposits on the metal surface can
However, sulfidation in oxidizing environments sometimes turn an oxidizing condition in the
(as well as in reducing environments) is fre- gaseous environment into a reducing condition
quently accelerated by other fuel impurities, such beneath the deposits.
as sodium, potassium, and chlorine, which may In most cases, alloys rely on oxide scales to
react among themselves and/or with sulfur dur- resist sulfidation attack; most high-temperature
ing combustion to form salt vapors. These salt alloys rely on chromium oxide scales. In oxi-
vapors may then deposit at lower temperatures dizing environments, oxide scales form readily
on metal surfaces, resulting in accelerated sulfi- because of high oxygen activities. Thus, oxida-
dation attack. This accelerated sulfidation attack tion dominates the corrosion reaction. When the
due to salt deposits is frequently referred to environment is reducing (i.e., low oxygen
as “hot corrosion” in gas turbines (Chapter 9), potentials), the corrosion reaction becomes a
“coal ash corrosion” in coal-fired boilers competition between oxidation and sulfidation.
(Chapter 10), and “oil ash corrosion” in oil-fired Thus, lowering the oxygen activity tends to make
boilers (Chapter 11). They are discussed in these the environment more sulfidizing, resulting in
chapters. increased domination by sufidation. Conversely,
Oxidizing environments that contain high increasing the oxygen activity generally results
concentrations of SO2 can be produced in the in a less-sulfidizing environment with increased
chemical process used to manufacture sulfuric domination by oxidation. Sulfidation reaction is
acid. Sulfur, in this case, is used as a feedstock. thus controlled by both sulfur and oxygen
Combustion of sulfur with excess air takes place activities (also referred to as “potentials”). When
in a sulfur furnace at about 1150 to 1200 °C corrosion involves more than one mode,
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:13AM Plate # 0 pg 202

202 / High-Temperature Corrosion and Materials Applications

including sulfidation, sulfidation generally dic- oil ash corrosion) are covered in Chapters 9 to 11,
tates materials selection. respectively.
Sulfidation attack, in many cases, can be quite
localized. Figure 7.1 shows a high-nickel alloy
that suffered sulfidation attack at about 930 °C 7.2 Thermodynamic Considerations
(1700 °F) in a furnace firing ceramic tiles. The
cross section at the corroded area showed sul- Iron, nickel, and cobalt are the alloy bases for a
fides through the cross section of the component. majority of high-temperature alloys. Most high-
The breakdown of a protective oxide scale (i.e., temperature alloys rely on chromium to form a
Cr2O3 scale for most high-temperature alloys) protective chromium oxide scale to resist oxi-
usually signifies the initiation of breakaway dation and other high-temperature corrosion
corrosion, which is generally followed by rapid attack. When the sulfur potential is high enough,
corrosion attack. Another example of component sulfides of iron, nickel, cobalt, and chromium are
failure due to sulfidation attack is illustrated in likely to form under sulfidation attack. An
Fig. 7.2. Alloy 800H suffered breakaway corro- Ellingham diagram such as the one shown in
sion. The chromium oxide scale broke down and
was replaced by iron-rich oxides, nickel-rich
sulfides, and chromium-rich sulfides, as illus-
trated in Fig. 7.2(b). The alloy matrix beneath
the corrosion products was found to be severely
depleted of chromium, as shown in Fig. 7.2(b).
Materials problems related to sulfidation
attack have been encountered in various indus-
tries. Some of these problems have been reported
in calcining of mineral and chemical feedstock
(Ref 1), petrochemical processing (Ref 2), fossil-
fired boilers (Ref 2), petroleum refining (Ref 3),
coal gasification (Ref 4, 5), waste incineration
(Ref 6, 7), fluidized-bed coal combustion (Ref
8–10), and oil-fired boilers.
This chapter reviews corrosion data involving
mainly gaseous environments. The data are
grouped into three different types typical of
industrial environments: (1) H2/H2S mixtures
and sulfur vapor with extremely low oxygen
activities such that a Cr2O3 scale is not thermo-
dynamically stable, (2) reducing, mixed-gas
environments with sufficiently high oxygen
activities such that Cr2O3 is thermodynamically
stable, and (3) SO2-bearing environments.
Sulfidation accelerated by salt deposits (hot
corrosion in gas turbines, coal ash corrosion, and

Fig. 7.2 Alloy 800H recuperator suffering severe sulfidation

attack in a nonferrous metal scrap melting furnace.
The 9.5 mm (0.4 in) thick recuperator was perforated in less than
2 years at metal temperatures of about 650 to 760 °C (1200 to
1400 °F). (a) General view of a corroded sample. (b) Cross section
Fig. 7.1 Alloy 601 tube suffering localized sulfidation attack. of the sample showing corrosion products. 1, Fe-rich oxide (57Fe-
The tube was in service at about 930 °C (1700 °F) in 18Ni-16Cr-5S-4K); 2, Fe-rich oxide (80Fe-8Ni-12Cr); 3, Ni sulfide
a natural gas-fired furnace making ceramic tiles. Sulfur was (light gray phase) (72Ni-2Cr-27S); 4, Cr depleted zone (72Ni-
believed to come from the ceramic feedstock. 28Fe); 5, Cr-rich sulfide (internal phase) (40Cr-10 Fe-50S).
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:13AM Plate # 0 pg 203

Chapter 7: Sulfidation / 203

Fig. 7.3 (Ref 11) can help determine whether an can be read from Fig. 7.3 by drawing a straight
environment has a sulfur potential high enough line from point “S” through the free-energy line
to form sulfides. The sulfur potential is repre- of the sulfide phase through the temperature of
sented by either pS2 or pH2 S =pH2 . The sulfur interest, and intersecting with the pS2 scale. The
partial pressure (pS2 ) in equilibrium with a sulfide intersection at the pS2 scale gives the sulfur

H2S/H2 ratio

Fig. 7.3 Standard free energies of formation of selected sulfides. Source: Ref 11
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:13AM Plate # 0 pg 204

204 / High-Temperature Corrosion and Materials Applications

partial pressure in equilibrium with the sulfide Metal-sulfur-oxygen stability diagrams at

phase. The pH2 S =pH2 value in equilibrium with a 870 °C (1600 °F) for iron, nickel, cobalt, and
sulfide can be obtained similarly by using the chromium are shown in Fig. 7.4 to 7.7 (Ref 15).
starting point “H” for the straight line; the inter- The diagrams map out the regions within which
section with H2S/H2 scale gives the equilibrium different phases will be in equilibrium at the gas/
H2S/H2 value. It is clear from Fig. 7.3 that the metal interface. Gordon and Nagarajan (Ref 16)
sulfide of chromium is more stable than those of constructed stability diagrams for the Fe-Cr-Ni
iron, nickel, and cobalt. system at 870 °C (1600 °F) as shown in Fig. 7.8.
Most sulfidizing environments exhibit both The spinel phases were deleted from the original
sulfur and oxygen activities. It is thus better to figure for better clarity. In an environment
characterize the environment in terms of pS2 and marked “X,” for example, phases of Cr2O3, FeS,
pO2 . Both sulfur and oxygen potentials of an and NiS are likely to form at the gas/metal
environment containing various gaseous com- interface for the Fe-Cr-Ni system.
ponents in an equilibrium condition can be cal- Perkins (Ref 14) proposed that in the envi-
culated using commercial software, such as HSC ronment within the upper right region of the Cr-
(Ref 12) and Chemsage (Ref 13). The environ- S-O diagram (Fig. 7.9) both CrS and Cr2O3 will
ment in terms of pS2 and pO2 can be presented in a form initially on the metal surface of chromium
stability diagram of a metal-sulfur-oxygen sys- and high-chromium alloys. The Cr2O3 will grow
tem. Perkins (Ref 14) and Hemmings and Perkins and overtake the sulfide because it is a stable
(Ref 15) discuss briefly how stability diagrams phase in this phase stability region. The alloying
can be used to aid understanding corrosion elements that form stable sulfides can diffuse
behavior. Various metal-sulfur-oxygen stability through the oxide scale and eventually form
diagrams at different temperatures can be found sulfides on the surface of the oxide scale, leading
in Ref 15. to breakaway corrosion. It is particularly serious
Stability diagrams allow one to predict the when the sulfides become liquid. Some metal-
phases that are likely to form on pure metals. metal sulfide eutectics have low melting

Log pSO /pSO –26 –24 –22 –20 –18 –16 –14 –12 –10 –8 –6 –4 –2 0 2 4
3 2
Log pCO /pCO –18 –16 –14 –12 –10 –8 –6 –4 –2 0 2 4 6 8 10 12
2
Log pH O /pH –18 –16 –14 –12 –10 –8 –6 –4 –2 0 2 4 6 8 10 12
2 2

S(l ) 2
0
Fe2(SO4)3(s)
FeS1+x (s) 0
–5
–2
–10 FeSO4 (s)
–4
2
Log pH S/pH

–15 –6
Log pS , atm

–8
2

–20 Fe3O4 (s)

–10
Fe(s)
–25 Fe0.950(s)
–12

–30 –14
Fe2O3(s)
–35 –16

–18
–50 –45 –40 –35 –30 –25 –20 –15 –10 –5 0 5
Log pO , atm
2

Fig. 7.4 Stability diagram of the Fe-S-O system at 870 °C (1600 °F). Source: Ref 15
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:13AM Plate # 0 pg 205

Chapter 7: Sulfidation / 205

Log pSO /pSO –26 –24 –22 –20 –18 –16 –14 –12 –10 –8 –6 –4 –2 0 2 4
3 2
Log pCO /pCO –18 –16 –14 –12 –10 –8 –6 –4 –2 0 2 4 6 8 10 12
2
Log pH O /pH –18 –16 –14 –12 –10 –8 –6 –4 –2 0 2 4 6 8 10 12
2 2

S(l ) 2
0

NiSy (l) 0
–5
–2
NiSO4 (s)
–10
–4

2
Log pH S/pH
–15 –6
Log pS , atm

2
–8
2

–20
–10
Ni(s)
–25
–12
NiO(s)
–30 –14

–35 –16

–18
–50 –45 –40 –35 –30 –25 –20 –15 –10 –5 0 5
Log pO , atm
2

Fig. 7.5 Stability diagram of the Ni-S-O system at 870 °C (1600 °F). Source: Ref 15

Log pSO /pSO –26 –24 –22 –20 –18 –16 –14 –12 –10 –8 –6 –4 –2 0 2 4
3 2
Log pCO /pCO –18 –16 –14 –12 –10 –8 –6 –4 –2 0 2 4 6 8 10 12
2
Log pH O /pH –18 –16 –14 –12 –10 –8 –6 –4 –2 0 2 4 6 8 10 12
2 2

S(l ) 2
0
CoS1+x (s)
0
–5
–2
Co4S3+x (s) CoSO4 (s)
–10
–4
2
Log pH S/pH

–15 –6
Log pS , atm

–8
2

–20

CoO(s) –10
Co(s)
–25
–12

–30 –14

Co3O4 (s) –16

–35

–18
–50 –45 –40 –35 –30 –25 –20 –15 –10 –5 0 5
Log pO , atm
2

Fig. 7.6 Stability diagram of the Co-S-O system at 870 °C (1600 °F). Source: Ref 15
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:13AM Plate # 0 pg 206

206 / High-Temperature Corrosion and Materials Applications

points: 635 °C (1175 °F) for Ni-Ni3S2, 880 °C chromium oxide scale for protection against
(1616 °F) for Co-Co4S3, and 985 °C (1805 °F) sulfidation. Most industrial environments have
for Fe-FeS (Ref 17). Most high-temperature sufficient oxygen activities to form a chromium
alloys are chromia formers, which rely on oxide scale. The scale may eventually break

Log pSO /pSO –26 –24 –22 –20 –18 –16 –14 –12 –10 –8 –6 –4 –2 0 2 4
3 2
Log pCO /pCO –18 –16 –14 –12 –10 –8 –6 –4 –2 0 2 4 6 8 10 12
2
Log pH O /pH –18 –16 –14 –12 –10 –8 –6 –4 –2 0 2 4 6 8 10 12
2 2

S(l ) 2
0

0
–5 CrS (s)
–2
–10
–4

2
Log pH S/pH
–15 –6
Log pS , atm

2
Cr2O3 (s) –8
2

–20
–10
–25
Cr (s) –12

–30 –14

–35 –16

–18
–50 –45 –40 –35 –30 –25 –20 –15 –10 –5 0 5
Log pO , atm
2

Fig. 7.7 Stability diagram of the Cr-S-O system at 870 °C (1600 °F). Source: Ref 15

Cr2 S3
FeS Cr2 O3
NiS
–4
FeO

NiO
–8
2
Log pS

Fe Ni
NiO
CrS

FeO
–12
Fe
Cr Cr2 O3
Ni

–16

–36 –32 –28 –24 –20 –16 –12 –8 –4 0

Log pO
2

Fig. 7.8 Stability diagram for the M-S-O systems at 870 °C (1600 °F), where M stands for metal. Source: Ref 16
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:13AM Plate # 0 pg 207

Chapter 7: Sulfidation / 207

down, leading to breakaway corrosion. Perkins H2-H2S environments. Sulfidation of pure

(Ref 14) referred to this region as “limited metals, such as iron, nickel, and cobalt, and iron-,
life”; see the upper right region of the Cr-S-O nickel-, and cobalt-base alloys was found to
diagram shown schematically in Fig. 7.9. The proceed with the formation of sulfides (Ref 18,
environments within this region typically contain 19) in these environments. High sulfur potentials
H2, H2O, CO, CO2, H2S, and so forth, referred combined with extremely low oxygen potentials
to in this chapter as “reducing, mixed-gas have resulted in formation of sulfides. Thus, data
environments,” which are characterized by low generated in these types of environments are
oxygen and high sulfur potentials. Another type applicable only to industrial environments where
of sulfidizing environment with common char- sulfides, but not oxides, are thermodynamically
acteristics for the sulfidation of various alloys is stable.
the upper left region in Fig. 7.9. In this region, Various investigations of sulfidation of
sulfides are stable phases. Alloys in these envi- iron (247–977 °C, or 475–1790 °F), nickel
ronments form sulfide scales. The environments (400–640 °C, or 750–1185 °F), and cobalt (497–
include sulfur vapor and H2/H2S mixtures with 1000 °C, 925–1830 °F) in sulfur vapor and
extremely low oxygen activities such that Cr2O3 H2-H2S mixtures were summarized by Mrowec
is not thermodynamically stable. and Przybylski (Ref 18). Sulfidation of these
metals was found to follow a parabolic rate law,
with sulfide scales forming on metal surfaces in
7.3 Sulfidation in Sulfur Vapor, all cases. Chromium also sulfidizes at a parabolic
Hydrocarbon Streams (No H2), rate in sulfur vapor and H2-H2S mixtures with
sulfur potentials of 10−7 to 1.0 atm at 700 to
H2S, and H2-H2S Mixtures 1000 °C (1290 to 1830 °F), with only sulfide
scales forming on metal surfaces (Ref 18).
Sulfur vapor, hydrocarbon streams with no Data for Fe-Cr, Ni-Cr, and Co-Cr alloys were
hydrogen (H2), H2S, and H2-H2S mixtures have generated by several studies (Ref 20–26).
a common feature in sulfidation reactions. This Mrowec and Przybylski (Ref 18) summarized
common feature is that the corrosion products in these data in Fig. 7.10, which shows that at sulfur
these environments are sulfides. The data pre- potentials greater than 10−2 atm, all three alloy
sented in this section are not applicable to other systems follow a similar trend of decreasing
environments where oxides and sulfides coexist. sulfidation rate with increasing chromium con-
Sulfidation behavior of metals and alloys in tent. Sulfidation rates were expressed in terms
sulfur vapor environments has been studied gen- of parabolic rate constants since all three alloy
erally at sulfur pressures higher than 10−3 atm. systems were found to follow a parabolic rate
Studies of H2-H2S mixtures have been typically law. It is surprising to find that there is no sig-
carried out at sulfur partial pressures less than nificant difference in sulfidation resistance
10−2 atm. Mrowec and Przybylski (Ref 18) and among the three alloy systems. It appears that the
Young (Ref 19) gave excellent reviews on sulfi- binary alloys with less than 15% Cr showed more
dation of metals and alloys in sulfur vapor and scattering. Improved sulfidation resistance for
alloys with less than 40% Cr (at.%) was attri-
buted to the formation of an inner sulfide layer:
Fe(Fe2−xCrx)S4 for Fe-Cr alloys, chromium sul-
fides with nickel for Ni-Cr alloys, and chromium
sulfides with cobalt for Co-Cr alloys (Ref 18). In
alloys with higher chromium (40% at.), a single
layer of chromium sulfide (Cr2S3) was observed
(Ref 18). Davin and Coutsouradis (Ref 20)
p

investigated the relative sulfidation resistance

of Fe-Cr, Ni-Cr, and Co-Cr alloys with 20 and
35% Cr in H2S at 800 °C (1470 °F). For 20% Cr
alloys, the Co-Cr alloy was most resistant, fol-
p lowed by the Fe-Cr alloy. The Ni-Cr alloy was
least resistant. When chromium was increased to
35%, however, all three alloys showed similar
Fig. 7.9 Stability diagram for the Cr-S-O system. Source:
Ref 14 resistance.
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:13AM Plate # 0 pg 208

208 / High-Temperature Corrosion and Materials Applications

Aluminum is a beneficial alloying element Fe-Cr-Al system was summarized in Fig. 7.11 by
in Fe-Cr and Co-Cr systems in both sulfur vapor Mrowec and Przybylski (Ref 18) using data
and H2-H2S environments. The effect of alumi- generated by various studies (Ref 27–30). The
num on the parabolic rate constant of the improved sulfidation resistance of Fe-Cr alloys

Fig. 7.10 Parabolic rate constants for Fe-Cr, Ni-Cr, and Co-Cr alloys as a function of chromium content. Source: Ref 18

10–6

800 °C (1470 °F) --Fe-19 Cr-AI, MROWEC,WEDRYCHOWSKA

(1979)
--Fe-17 Cr-AI, JALLOULI et al. (1979)
10–7 --Co-25 Cr-AI, BIEGUN, BRÜCKMAN
(1980)
" g2cm4 s–1

--Fe-25 Cr-AI, SMELTZER et al. (1982)

pS = 1 atm
2

10–8
kp,

pS = 1 atm
2
pS = 8.10–5atm
2
pS = 10–7atm
Fe-Cr-AI 2
10–9
Co-Cr-AI

10 20 30
% at AI

Fig. 7.11 Parabolic rate constants for Fe-Cr-Al and Co-Cr-Al alloys as a function of aluminum content. Source: Ref 18
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:13AM Plate # 0 pg 209

Chapter 7: Sulfidation / 209

due to aluminum additions was attributed to the stainless steels in H2-50%H2S at 500 to 700 °C
formation of an inner sulfide layer of presumably (932 to 1292 °F) is shown in Table 7.2 (Ref 32).
Fe(FexAlyCrz-x-y)S4 (Ref 27). This sulfide Corrosion becomes much more aggressive in
decomposed at room temperature (Ref 27). The 100% H2S. Malinowski et al. (Ref 33) reported
outer layer consisted of Fe1−yS (Ref 27). Biegun corrosion rates of iron and iron-aluminum alloys
and Bruckman (Ref 29) reported a similar bene- in 100 H2S at 400, 600, and 700 °C (752, 1112,
ficial effect in Co-25Cr alloys when aluminum and 1292 °F) (Fig. 7.12). Although aluminum
content was varied 1 to 22% (at.%) at tempera- was found to be beneficial in corrosion rates of
tures of 700 to 1100 °C (1290 to 2010 °F) in Fe-Al system, the content of aluminum that was
sulfur vapor at 1 atm. However, the influence of needed to significantly reduce the alloy’s corro-
aluminum in the Co-Cr-Al system was not as sion rates was found to be about 10% at 600 and
effective as in the Fe-Cr-Al system (Fig. 7.11). 700 °C (1112 and 1292 °F), and slightly lower at
The Ni-Cr-Al system behaved differently from 400 °C (752 °F). Bruns (Ref 34) tested many
both the Fe-Cr-Al and Co-Cr-Al systems commercial alloys in 100% H2S at 593 °C
(Ref 31). Biegun and Bruckman (Ref 31) inves- (1100 °F). Also included in the tests were
tigated Ni-25Cr alloys containing 1 to 10 at.% experimental alloys based on Fe-32Ni-20Cr with
aluminum in sulfur vapor (1 atm) at 580 to various levels of aluminum (Ref 34). His test
950 °C (1080 to 1740 °F) with erratic results. results are summarized in Fig. 7.13 in terms of
The kinetics rates observed ranged from para- the percent of either chromium or aluminum
bolic to linear. content in the alloy (Ref 34). The data generated
Corrosion rates of several austenitic stainless in 100% H2S by Bruns indicate that increasing
steels in sulfur vapor at 571 °C (1060 °F) are chromium significantly improves the alloy’s
shown in Table 7.1 (Ref 32). Higher-chromium sulfidation resistance. It also indicated that add-
alloys, such as 314SS (24Cr, 2Si), 310SS (25Cr), ing 8.5% Mn to Fe-18Cr alloy failed to improve
and 309SS (23Cr), were found to perform the the alloy’s corrosion resistance. For Fe-32Ni-
best. Silicon may not be very effective at this 20Cr alloy (the base composition is similar to
relatively low temperature. In addition to Type alloy 800/800H) with various aluminum con-
314SS, which contains about 2% Si, there was tents, Fig. 7.13 shows strong beneficial effect for
another silicon-containing stainless steel, Type aluminum in increasing the alloy’s sulfidation
302B (18Cr, 2.5Si). Type 302B performed quite resistance. Furthermore, the aluminum content
similarly to Type 304SS (19Cr, no Si). For that significantly increased the sulfidation resis-
comparison, the corrosion rates of austenitic tance of Fe-32Ni-20Cr alloy was found to be in a
range of only 2 to 4%.
In refineries, sulfidation, which is commonly
Table 7.1 Corrosion rates of several referred to as “sulfidic corrosion” in refinery
austenitic stainless steels in sulfur vapor at industry, is a common materials problem. The
571 °C (1060 °F) for 1295 h temperatures of the refinery equipment having
Corrosion rates,
sulfidation problems are typically between about
Alloy mm/yr (mpy) 260 and 540 °C (500 and 1000 °F). Sulfur
314 0.43 (16.9) compounds originating from crude oils include
310 0.48 (18.9) polysulfides, hydrogen sulfide, mercaptans, ali-
309 0.57 (22.3)
304 0.69 (27.0) phatic sulfides, disulfides, and so forth (Ref 35).
302B 0.76 (29.8) Sulfidation occurs in some processing units
316 0.79 (31.1)
321 1.39 (54.8) where no hydrogen is present in the system, such
Source: Ref 32
as crude distillation units. The crude distillation
units that process mostly sweet crude oils (less

Table 7.2 Corrosion rates of Types 310, 304, and 316 in H2-50%H2S at 500–700 °C (932–1292 °F)
Corrosion rate 500 °C Corrosion rate 550 °C Corrosion rate 600 °C Corrosion rate 700 °C
Alloy (932 °F), mm/yr (mpy) (1022 °F), mm/yr (mpy) (1112 °F), mm/yr (mpy) (1292 °F), mm/yr (mpy)
Type 310 0.91 (36) 1.45 (57) 2.79 (110) 8.94 (352)
Type 304 1.12 (44) 1.57 (62) 2.95 (116) 10.2 (400)
Type 316 1.50 (59) 2.44 (96) 4.45 (175) 10.8 (424)
Source: Ref 32
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:13AM Plate # 0 pg 210

210 / High-Temperature Corrosion and Materials Applications

than 0.6% total sulfur, with essentially no the hydrocarbon streams (Ref 39). Hydrocrack-
hydrogen sulfide) experience relatively less sul- ing processes combine desulfurization and
fidation problems. More sulfidation problems are cracking operation that can convert hydrocarbon
encountered in the distillation units that process feedstocks into various products (Ref 39). Sul-
mostly sour crudes (Ref 36). The main fractio- fidation in these processing units is dictated by
nation tower is usually made of carbon steel the H2S concentration in the H2-H2S environ-
with the lower part lined with a stainless steel ment. Sulfidation in H2-H2S environments is
containing about 12% Cr, such as Type 405 severe since the corrosion products are sulfides
(Ref 37). Type 309L weld overlay cladding is with no protection by oxide scales in these
most commonly used for replacing the ferritic environments. Austenitic stainless steels are
stainless steel lining or cladding that suffered generally used as a cladding or weld overlay for
sulfidation attack. The corrosion rates of the sulfidation resistance. Because of high hydrogen
crudes and their liquid fractions can be predicted temperatures and pressures in the system, the
by the so-called modified McConomy curves. vessels generally are made of low-alloy steels for
Figure 7.14 shows the modified McConomy resisting hydrogen attack, and the selection of
curves for liquid hydrocarbon streams containing materials to avoid hydrogen attack problem
0.6 wt.% S (Ref 38). For hydrocarbon streams should follow the recommendations of the API
containing more than 0.6% sulfur, the corrosion Publication 941 (Ref 40). Materials issues related
rate multiplier as shown in Fig. 7.15 can be used to hydrogen attack are discussed in Chapter 17:
in conjunction with the data shown in Fig. 7.14 Hydrogen Attack. The reactors in hydrotreating
for making the prediction of the corrosion rate and hydrocracking units are normally made of
for various steels at different temperatures low-alloy steels with Type 347 cladding or weld
(Ref 38). overlay (Ref 41). The use of Type 347 (a stabi-
Hydrogen in hydrotreating, hydrocracking, lized grade of austenitic stainless steel) for
and hydrodesulfurizing processes is used to cladding or weld overlay is to avoid intergranular
remove sulfur (to convert it to hydrogen sulfide) stress-corrosion cracking by polythionic acid
and nitrogen (to ammonia) for separation from during downtime (Ref 39, 41).
Corrosion rates of metals and alloys in H2-H2S
environments are dependent on the H2S con-
centration. Corrosion of austenitic stainless steels
was very aggressive in H2-50%H2S and 100%
H2S environments, as shown in Table 7.2 and
Fig. 7.13. However, the concentrations
of hydrogen sulfide (H2S) in H2-H2S environ-
ments in hydrotreating, hydrocracking, hydro-
desulfurizing, and catalytic reforming processes
are significantly lower, thus making carbon
steels, low-alloy steels and stainless steels cap-
able of resisting sulfidation attack in different
temperature ranges. For example, Neumair and
Schillmoller (Ref 42) reported the corrosion rates
of steels and austenitic stainless steels as a
function of hydrogen sulfide concentration for
a hydrodesulfurization equipment at 415 °C
(780 °F) and 27.2 atm (400 psig) (Fig. 7.16) and
for a catalytic-reformer equipment at 510 °C
(950 °F) and 27.2 atm (400 psig) (Fig. 7.17).
The corrosion data in catalytic reforming units
generated by different refinery companies was
summarized by Sorell (Ref 43) and are shown in
Table 7.3 to illustrate the sulfidation of various
steels in H2-H2S mixtures. Catalytic reforming is
used in petroleum refineries to upgrade the
Fig. 7.12 Corrosion rates of iron and iron-aluminum alloys in
octane number of gasoline (Ref 44). Severe
100% H2S at 400, 600, and 700 °C (752, 1112, and
1292 °F). Source: Ref 33 corrosion attack on processing equipment by
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:13AM Plate # 0 pg 211

Chapter 7: Sulfidation / 211

hydrogen sulfide has been encountered in several Sorell and Hoyt (Ref 45) reported the sulfi-
catalytic reforming and desulfurizing units dation resistance of Fe-Cr and Fe-Cr-Ni alloys
(Ref 43, 45, 46). In Table 7.3, Sorell (Ref 43) in H2-16H2S at 315 to 480 °C (600 to 900 °F)
reported hydrogen sulfide corrosion data gene- (Fig. 7.18). Slight to moderate addition of nickel
rated by various refinery operators in laboratory (e.g., 8 or 40%) to Fe-20Cr improved the alloy
tests, pilot plant testing, field testing in com- sulfidation resistance. Greater nickel addition
mercial units, and inspection of commercial (e.g., 65% or more) reduced sulfidation resis-
operating equipment. Austenitic stainless steels tance. Figure 7.18 also surprisingly revealed that
(18Cr-8Ni) were most resistant, followed by Fe-20Cr-65Ni exhibited a sulfidation resistance
straight chromium stainless steels (12–16% Cr), similar to that of Fe-20Cr (Ref 45). Sulfidation of
with low-chromium steels (0–9% Cr) being alloys in H2-H2S mixtures has been described by
worst. isocorrosion rate curves, which show corrosion

Fig. 7.13 Corrosion rates of commercial alloys as well as experimental Fe-30Ni-20Cr with various aluminum contents in 100% H2S at
593 °C (1100 °F). Note the upper curve shows the corrosion rates as a function of chromium content in the alloy, while the
lower curve show the corrosion rates as function of Al content in the alloy. Note: alloy 202 (Fe-18Cr-8.5Mn), CL (Fe-32Ni-20Cr-0.25C), CI
(Fe-32Ni-21Cr-0.025C), 804 (Fe-41Ni-29Cr-0.97Ti), 2.24% Si alloy (Fe-32Ni-21Cr-2.24Si). Source: Ref 34
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:13AM Plate # 0 pg 212

212 / High-Temperature Corrosion and Materials Applications

rate as a function of H2S concentration and concentration of H2S for various steels (Ref 38).
temperature. Backensto and Sjoberg (Ref 48) These curves were based on a survey conducted
determined isocorrosion rate curves for chro- by NACE Committee T-8 on Refining Industry
mium steels (0 to 5% Cr) and austenitic stainless Corrosion (Ref 49). The modified Couper-Gor-
steels, shown in Fig. 7.19 and 7.20, respectively. man curves (i.e., the original curves were
“Couper-Gorman” curves are generally extended to higher concentrations of H2S with
believed to provide the most practical correlation dashed lines) for 5Cr-0.5Mo steels (also applic-
among corrosion rates, temperature and able to carbon steels) and 18Cr-8Ni steel are
shown in Fig. 7.21 and 7.22, respectively. It was
found that total pressure between 1 and 18 MPa
(150 and 2650 psig) was not a significant vari-
able. The authors also observed that no sulfida-
tion occurred at very low H2S concentrations and
at temperatures above 315 °C (600 °F) because
formation of iron sulfides in that regime is not
feasible thermodynamically.

7.4 Sulfidation in Environments with

Low-Oxygen and High-Sulfur Potentials

The environments covered in this section

typically contain H2, CO, CO2, H2O, H2S, and
other gaseous components with oxygen and
sulfur potentials high enough to form oxides and
sulfides for most high-temperature alloys. These
types of reducing environments are generated by
combustion under substoichiometric conditions
with insufficient air or oxygen. The corrosion
reactions in these environments generally in-
Fig. 7.14 Modified McConomy curves providing predicted
corrosion rates as a function of temperatures for volve oxidation and sulfidation. Sulfidation is
various steels in a liquid hydrocarbon stream containing 0.6% S. usually the determining factor in the alloy’s
Source: Ref 38
performance capability. In most cases, the alloy
remains protected during the oxidation period
until the initiation of breakaway corrosion, which
is then followed by rapid sulfidation attack by
forming primarily sulfides.
Sulfidation behavior of alloys in this type of
environments, where oxidation and sulfidation
reactions are involved, has been extensively
studied during the 1970s, 1980s, and early 1990s
when coal gasification programs had been
actively pursued. Coal gasification atmospheres
typically have sulfur potentials (pS2 ) of 10−5 to
10−10 atm and oxygen potentials (pO2 ) of 10−15 to
10−20 atm (Ref 50). Significant understanding of
corrosion reactions in these environments has
been achieved (Ref 51–67). In addition, a large
engineering database on commercial alloys has
been generated. Major contributions to this
database came from Metal Properties Council
Fig. 7.15 Effect of sulfur content in hydrocarbon streams (MPC) programs from 1972 to 1985 that eval-
on corrosion rates predicted by the modified
McConomy curves in the 290–400 °C (550–750 °F) range. uated more than 80 commercial alloys and
Source: Ref 38 coatings for coal gasification. The tests were
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:14AM Plate # 0 pg 213

Chapter 7: Sulfidation / 213

carried out mostly by IIT Research Institute, and by Howes (Ref 60) and Verma (Ref 63). The
results were documented in MPC annual reports test gas mixtures used are tabulated in Table 7.4,
(Ref 51–59). All these results were summarized and their thermodynamic potentials (pS2 and

Fig. 7.16 Maximum anticipated corrosion rates for alloys in hydrodesulfurization equipment at 415 °C (780 °F) and 27.2 atm
(400 psig). Source: Ref 42

Fig. 7.17 Maximum anticipated corrosion rates for alloys in catalytic reforming equipment at 510 °C (950 °F) and 27.2 (400 psig).
Source: Ref 42
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:14AM Plate # 0 pg 214

214 / High-Temperature Corrosion and Materials Applications

Table 7.3 Summary of corrosion data in H2-H2S mixtures typical of catalytic reforming generated
by various petroleum refining companies
Type of Test Hydrogen H2S
corrosion Temperature, duration, pressure, concentration, Corrosion rate,
Source data(a) °C (°F) days atm (psi) vol% Materials mm/yr (mpy)
American Oil I 510 (950) 36–89 18–27 (265–400) 0.03–0.07 1¼Cr, 2¼Cr 1.5 (59)
C 510–550 (950–1025) 89 18–27 (265–400) 0.03–0.07 0–9Cr 1.5 (59)
12Cr 0.76 (30)
18Cr–8Ni 0.33 (13)
Atlantic L, P, C, I 480 (900) 1–365 34 (500) 0.013–0.074 0–5Cr 0.1–2.5 (5–100)
0.011–0.13 11½–13½Cr 0.03–1.5 (1–60)
0.015–0.27 18Cr–8Ni 0.03–0.2 (1–8)
I 480 (900) 180 34 (500) 0.036 2¼Cr 1.0 (38)
0.016 12Cr 0.18 (7)
D-X Sunray I 465 (870) 180 27 (400) 0.04 5Cr 1.0 (40)
Canadian Petrofina C 480–490 (890–920) 90 21–22 (310–330) 0.008 0–9Cr 0.4 (16)
12Cr 0.2 (8)
18Cr–8Ni 0.06 (2.5)
Humble C 480–510 (900–956) 1–4 20 (300) 0.013–0.14 0–12Cr 1.9–10 (73–398)
18Cr–8Ni 0.1–1.1 (5–42)
P 540 (1000) 4 20 (300) 0.007–0.15 0–12Cr 0.03–10 (1–400)
18Cr–8Ni 0.01–1.5 (0.4–60)
I 450–520 (850–975) 127 20 (300) 0.035 1¼Cr 2.3 (90)
480–550 (900–1025) CS, 2¼Cr 5.1 (200)
Major U.S. C 470–520 (875–960) 139–577 34–36 (500–535) 0.035–0.09 2¼Cr, 5Cr 0.3–1.0 (13.5–41.5)
Ref. Comp.
12Cr 0.28–0.7 (11–27)
18Cr–8Ni 0.1 (3.7)
Major West Coast C 480–490 (900–920) 72 37–38 (540–565) 0.015 0–5Cr 0.5 (18.9)
Ref. Comp.
12Cr 0.2 (8.3)
18Cr–8Ni 0.1 (4.5)
Pure Oil I 480–540 (900–1000) 240 34 (500) 0.026 2Cr 0.8 (33)
Richfield L, P, C, I 510 (950) 16.7 27 (400) 0.011–1.0 0–9Cr 0.4–7.9 (15–310)
0.014–1.0 12Cr 0.3–6.1 (10–240)
0.025–1.0 18Cr–8Ni 0.1–0.5 (3.5–20)
Shell Oil C 510 (950) 71 43 (625) 0.015 0–9Cr 0.5 (18)
12Cr 0.2 (8)
I 510 (950) 148–758 46 (680) 0.05–0.08 5Cr 1.4–3.4 (56–135)
510–610 (955–1135) 758 46 (670) 0.08 9Cr 1.7 (68)
C 500 (925) 148–251 46 (680) 0.05–0.08 0–9Cr 0.9–1.2 (35–47)
12Cr 0.4–0.6 (15–22)
18Cr–8Ni 0.1–0.2 (4.3–7.2)
Sinclair L 510 (950) 4.2 34 (500) 0.08–1.0 0–7Cr 2.0–11 (80–450)
0.1–1.0 13Cr 0.5–5.3 (20–210)
0.1–1.0 18Cr–8Ni 0.1–1.2 (3–48)
C 490–500 (910–940) 212–382 34 (500) 0.001–0.0089 0–7Cr 0.1–0.3 (2.4–10)
13Cr 0.05–0.2 (1.9–6.5)
18Cr–8Ni 0.01–0.02 (0.55–0.6)
I 480–550 (900–1025) 120–270 34 (500) 0.005 5Cr 0.2 (90).
480–510 (900–950) 12Cr 0.2 (6.5)
Socony Mobil L 530 (985) 6.3–20.8 33 (485) 0.009–0.2 0–5Cr 0.03–7.2 (1–285)
0.017–0.2 7–16Cr 0.03–2.6 (1–102)
0.06–0.2 Cr–Ni 0.03–0.2 (1–8.7)
L 490 (905) 21 31 (460) 0.085 5Cr 1.1 (44)
C 470–500 (885–935) 55 12 (175) 0.5 0–9Cr 2.7 (108)
12–16Cr 0.5 (21)
Cr–Ni 0.1 (5)
C 470–490 (885–920) 50.5–213 37–39 (550–575) 0.018–0.04 0–9Cr 0.5–1.6 (18–62)
12Cr 0.3–0.6 (13–23)
Cr-Ni 0.02–0.1 (0.7–4.3)
Standard Oil (Ind.) L 480 (900) 0.7–11.7 19 (285) 0.013–0.44 0–9Cr 0.3–13 (10–500)
12Cr 0.1–5 (4–200)
18Cr–8Ni 0.03–1.6 (1.3–62.5)
C 510 (950) 16.5–46 20 (300) 0.023 CS 1.8 (70)
510–540 (950–1000) 0.032 1.1 (45)
C 550 (1030) 47 23 (340) 0.029 0–9Cr 1.0 (37.5)
18Cr–8Ni 0.1 (3.3)
Sun Oil C 500–520 (925–960) 188–395 20–41 (300–600) 0.017–0.04 0–5Cr 0.2–0.8 (6.1–29.7)
12 Cr 0.1–0.3 (5.8–11.7)
18Cr–8Ni 0.02–0.05 (0.7–1.9)
Texas C 500 (940) 133–668 41 (600) 0.0016 0–5Cr 0.07 (2.9)
12Cr 0.05 (1.9)
18Cr–8Ni 0.01 (0.3)
(a) L, laboratory corrosion tests; P, pilot plant corrosion tests; C, commercial unit corrosion tests; I, inspection of commercial operating equipment. Source: Ref 43
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:14AM Plate # 0 pg 215

Chapter 7: Sulfidation / 215

pO2 ) at various test temperatures and pressures sulfide slag over the chromium oxide scale is
are shown in Table 7.5. The test environments most damaging. Alloying elements, such as
in terms of pO2 and pS2 in the phase-stability manganese, iron, cobalt, nickel, and so forth,
diagrams for two test temperatures—650 and diffuse through the chromium oxide scale and
980 °C (1200 and 1800 °F)—are shown in react with the environment on top of the oxide
Fig. 7.23 and 7.24 (Ref 63). scale to form external sulfides. This was pro-
Most alloys suffered breakaway corrosion in posed by Perkins (Ref 14) as a possible mech-
several thousands of hours or less. Breakaway anism for breakaway corrosion, schematically
corrosion is illustrated in Fig. 7.25 (Ref 60). Both illustrated in Fig. 7.27. Manganese is the fastest
Type 310SS and alloy 800H followed a parabolic diffusing element, followed by iron, cobalt,
reaction rate prior to rapid corrosion attack. nickel, and chromium (Ref 14). In addition to
However, a few alloys did not exhibit breakaway outward diffusion of alloying elements to form
corrosion within the test duration of up to external sulfides, the corrosion reaction also
10,000 h, including the Co-Cr-W alloy 6B as involves sulfur penetrating through the oxide
shown in Fig. 7.25. Verma (Ref 63) discussed in scale to form discrete particles of sulfides in the
detail the performance of various alloys in the matrix (Ref 14). Natesan (Ref 62) showed that
MPC environments. Chromium was the most for a given sulfur potential, there exists a
important alloying element in resisting sulfida- threshold value for oxygen potential beyond
tion attack. which a continuous protective oxide scale is
Most alloys were protected by chromium developed. This threshold oxygen partial pres-
oxide scales. The scale may eventually break sure (kinetic boundary) is about 103 times the
down, leading to breakaway corrosion oxygen partial pressure for chromium oxide and
(Fig. 7.26) (Ref 63). Breakdown of the protective
chromium oxide scale results from the sulfides
formed over it (Ref 14, 61). Formation of liquid

Fig. 7.19 Corrosion rates of chromium steels (0–5% Cr)

Fig. 7.18 Corrosion rates of Fe-Cr and Fe-Cr-Ni alloys in H2- generated from laboratory tests in H2-H2S at
16H2S at 315 to 480 °C (600 to 900 °F) and 75 atm hydrogen pressures of 12 to 34 atm (175 to 500 psig) as a function
(1100 psig) pressure. Source: Ref 45, summarizing results of of H2S concentration and temperature. IPY, inch per year. Source:
Ref 47 Ref 48
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:14AM Plate # 0 pg 216

216 / High-Temperature Corrosion and Materials Applications

chromium sulfide equilibrium (thermodynamic In the MPC data summarized in Figs. 7.28 to
boundary) (Ref 62). 7.31, many stainless steels and nickel alloys,
The sulfidation resistance of various com- while showing very high wastage rates at high
mercial alloys generated in the MPC programs temperatures (e.g., 816 °C and higher), showed
is summarized in Fig. 7.28 to 7.31 (Ref 60) and low corrosion rates at 482 and 650 °C (900 and
Fig. 7.32 and 7.33 (Ref 67). For nickel-base 1200 °F). Review of the test data generated
alloys, increasing nickel generally increases at 650 °C (1200 °F) for 10,000 h in the MPC test
susceptibility to sulfidation attack (Fig. 7.28). programs summarized in Howes’s report
High-nickel alloys, such as alloys 600 and 601, (Ref 60) showed that alloys containing at least
are particularly susceptible to sulfidation attack, 18% Cr, which included 304, 309, 310, 312, 329,
as illustrated in Fig. 7.32 (Ref 67). The Ni-Ni3S2 22-13-5, 29-4-2, Ebrite 26-1, 446, 20Cb-3,
eutectic melts at 635 °C (1175 °F). Molten sul- 800, 253MA, Crump 25, 825, RA333, 617,
fide slag can easily destroy the chromium oxide and Multimet, exhibited corrosion rates of
scale and cause catastrophic sulfidation attack. 0.02 mm/yr or less (1 mpy or less) when tested in
Increasing chromium generally improves sul- a 500 psig gas pressure and 1% H2S in the MPC
fidation resistance in Fe-Cr-Ni alloys (Fig. 7.29) coal gasification environment. Type 410 (12Cr)
(Ref 60). This also holds true for iron- and nickel- showed the corrosion rate of 0.05 mm/yr
base alloys. Type 446SS (Fe-27Cr) and 50Ni- (2 mpy), and 9Cr-1Mo steel suffered the corro-
50Cr alloys (alloys 671 and 657) are good sion rate of 1.01 mm/yr (40 mpy) (Ref 60).
examples (Fig. 7.30) (Ref 60). Despite its high John et al. (Ref 68) reported the corrosion data
nickel content, the 50Ni-50Cr alloy exhibited generated at 700 °C (1292 °F) and lower (1 atm
good sulfidation resistance due to its extremely pressure) for various commercial alloys in syn-
high chromium content. gas environments that were rich in CO and H2
with low water. The corrosion data for carbon

Fig. 7.20 Corrosion rates of Cr-Ni austenitic stainless steels Fig. 7.21 Modified Couper-Gorman curves showing corro-
generated from laboratory tests in H2-H2S at sion rates as a function of H2S concentration (mol.
hydrogen pressures of 12 to 34 atm (175 to 500 psig) as a function %) and temperatures for 5Cr-0.5Mo steel. The data also is
of H2S concentration and temperature. IPY, inch per year. Source: applicable to carbon steels and alloy steels with less than 5% Cr
Ref 48 (naphtha desulfurizers). Source: Ref 38
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:14AM Plate # 0 pg 217

Chapter 7: Sulfidation / 217

steel and 1Cr-0.5Mo steel tested in CO-51.2H2- Table 7.5 Oxygen and sulfur potentials ( pO2 and
0.8CO2-1.7H2S-0.2HCl at 330, 400, and 500 °C pS2 ) in the MPC-CGA test gas mixtures at 68 atm
(626, 752, and 932 °F) are summarized in (1000 psig)
Fig. 7.34 (Ref 68). The corrosion behavior of Temperature, H2S in test gas,
°C (°F) vol% pO2 (atm) pS2 (atm)
alloy 625 in the same test environment at 330
480 (900) 0.1 1.2 × 10−23 1.5 × 10−10
to 700 °C (626 to 1292 °F) is summarized in 650 (1200) 0.1 7.0 × 10−21 8.8 × 10−10
Fig. 7.35 (Ref 68). The data for various Fe-, Ni-, 816 (1500) 0.1 3.0 × 10−18 5.4 × 10−9
and Co-base alloys at 700 °C (1292 °F) and 900 (1650) 0.1 6.4 × 10−17 1.3 × 10−8
980 (1800) 0.1 1.3 × 10−15 3.1 × 10−8
lower in CO-26H2-1.0CO2-0.9 H2S-2H2O-10N2 480 (900) 0.5 1.2 × 10−23 3.5 × 10−9
is summarized in Fig. 7.36 (Ref 68). It was sur- 650 (1200) 0.5 7.0 × 10−21 2.2 × 10−8
816 (1500) 0.5 3.0 × 10−18 1.4 × 10−7
prising that Co-base alloys, such as alloys 25 900 (1650) 0.5 6.4 × 10−17 3.1 × 10−7
980 (1800) 0.5 1.3 × 10−15 7.6 × 10−7
480 (900) 1.0 1.2 × 10−23 1.5 × 10−8
650 (1200) 1.0 7.0 × 10−21 8.8 × 10−9
816 (1500) 1.0 3.0 × 10−18 5.5 × 10−7
900 (1650) 1.0 6.4 × 10−17 1.3 × 10−6
980 (1800) 1.0 1.3 × 10−15 3.1 × 10−6
Source: Ref 60

NiO
p

Co/CoO

Fig. 7.23 The phase-stability diagram and the test environ-

ment (star) in terms of equilibrium pO2 and pS2 in
Fig. 7.22 Modified Couper-Gorman curves showing corro- the MPC coal gasification test programs at 650 °C (1200 °F).
sion rates as a function of H2S concentration (mol. Source: Ref 63
%) and temperatures for 18Cr-8Ni steel. Source: Ref 38

Table 7.4 Inlet and equilibrium gas compositions at different temperatures for the MPC-CGA test gas
mixtures
Equilibrium gas composition(a), vol%
Gaseous component Inlet gas, vol% 480 °C (900 °F) 650 °C (1200 °F) 816 °C (1500 °F) 900 °C (1650 °F) 980 °C (1800 °F)
H2 24 4 11 23 27 31
CO 18 5 8 11 14 17
CO2 12 25 22 19 17 15
CH4 5 19 14 9 6 3
NH3(b) 1 1 1 1 1 1
H2S 0–1.5 0.5 1.5 0.1–1.5 0.5–1.0 0–1.0
H2O bal bal bal bal bal bal
(a) Computer calculations at 6.9 MPa (1000 psig). (b) NH3 will decompose to N2 and H2 on increasing temperature. Source: Ref 60
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:14AM Plate # 0 pg 218

218 / High-Temperature Corrosion and Materials Applications

(Co-20Cr-10Ni-115W) and 150 (Co-27Cr- (Fe-Ni-Cr alloys). In general, cobalt-base alloys

18Fe), were not significantly better than Type and cobalt-containing alloys have higher tem-
310, and but not as good as Type 446. The sul- perature capabilities and are more resistant to
fidation behavior of Co-base alloys at higher breakaway corrosion. Results generated by Lai
temperatures is discussed later. (Ref 71) at 760, 870, and 980 °C (1400, 1600,
Bakker (Ref 69) and Bakker and Bonvallet and 1800 °F) also revealed that cobalt-base
(Ref 70) investigated the effect of H2O con-
centration on the corrosion behavior of Type 310
and alloy 800 in coal gasification environments
at 540 °C (1004 °F) for 600 h. The test gas
environments with different levels of H2O from
0 to 15% along with oxygen partial pressures
(pO2 ) at 540 °C (1004 °F) are tabulated in
Table 7.6. The test results are summarized in
Fig. 7.37. Both oxides and sulfides were ob-
served to have formed on both alloys in all test
environments. Type 310 was found to suffer
significantly less metal loss than alloy 800, and
its corrosion was relatively unaffected by H2O
concentrations. Alloy 800 suffered much higher
metal loss, with highest metal loss surprisingly at
about 3% H2O. Long-term tests are needed to
confirm this “abnormal” observation.
Cobalt-base alloys as well as cobalt-
containing alloys (e.g., Fe-Ni-Co-Cr) generally
have better sulfidation resistance than nickel-
base alloys and Fe-Cr-Ni alloys (Fig. 7.33) Fig. 7.25 Corrosion behavior of Type 310 stainless steel,
(Ref 67). Alloys 6B and 188 (cobalt-base alloys) alloy 800, and alloy 6B at 980 °C (1800 °F) in the
MPC coal gasification atmosphere, showing breakaway corrosion
and alloy N155 (Fe-Ni-Co-Cr alloy) are better for Type 310SS and alloy 800. Inlet gas: 24H2-18CO-12CO2-
than alloys RA333 and X (nickel-base alloys) 5CH4-1NH3-0.5H2S (balance H2O) at 6.9 MPa (100 psig). pO2 =
and alloys 800H, HK-40, and Type 310SS 1.3 × 10−15 atm. pS2 = 7.6 × 10−7 atm. Source: Ref 60

Fe/FeS
–5
Ni/NiS
Cr/CrS
–10
Log pS , atm

Al/AlS
FeO/Fe3O4
2

–15
Fe3O4/Fe2O3

–20
Cr2O3/Cr
Al2O3/Al

FeO/Fe

Ni/NiO

–25

–30
–40 –35 –30 –25 –20 –15 –10 –5
Log pO , atm
2

Fig. 7.24 The phase stability diagram and the test environment (star) in terms of equilibrium pO2 and pS2 in the MPC coal gasification
test programs at 980 °C (1800 °F). Source: Ref 63
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:14AM Plate # 0 pg 219

Chapter 7: Sulfidation / 219

Fig. 7.26 Corrosion behavior of alloy 800 in the MPC coal gasification atmosphere (0.5% H2S at 900 °C or 1650 °F, and 6.9 MPa
or 1000 psig) showing oxide scales during the protective stage and oxides/sulfides after breakaway corrosion. Source:
Ref 63

alloys were more sulfidation resistant than

nickel-base and Fe-Ni-Cr alloys with similar
chromium contents (Fig. 7.38).
Higher chromium (preferably more than 25%)
is needed for an alloy to form a protective
oxide scale in coal gasification environments.
Verma (Ref 63) suggested that chromium oxide
scale would be more protective if doped with
cobalt, as was the case in alloy 6B (Co-Cr-W)
that formed a (Cr,Co)2O3 scale. Bradshaw et al.
(Ref 65) believed that titanium modified the
Cr2O3 scale, thus improving the alloy’s sulfida-
tion resistance. They compared 3% Ti-modified
Fig. 7.27 Formation of external sulfides on top of the chro-
Type 310SS with Type 310SS after exposure in mium oxide scale and the formation of internal
an MPC gas mixture with 1% H2S at about sulfides. Source: Ref 14
1.0 atm pressure for 100 h at 980 °C (1800 °F).
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:14AM Plate # 0 pg 220

220 / High-Temperature Corrosion and Materials Applications

Fig. 7.28 Corrosion rates of high-nickel alloys in the MPC coal gasification atmosphere with 1.0 and 1.5% H2S (see Table 7.4 and 7.5
for gas composition). Source: Ref 60

330

Fig. 7.29 Corrosion of Fe-Cr-Ni alloys in the MPC coal gasification atmosphere with 1.0 and 1.5% H2S (see Tables 7.4 and 7.5 for gas
composition). Source: Ref 60
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:14AM Plate # 0 pg 221

Chapter 7: Sulfidation / 221

Type 310 sample showed some sulfide nodules Matl % H2S

as well as some spalled oxides, while Ti-mod- 446
0.1
0.5
ified Type 310 sample showed an adherent oxide 1.0
0.1
302 0.5 NT
scale with no sulfide nodules. Their analysis 0.0
0.1
indicated a significant amount of titanium in the 304 0.5
1.0
Cr2O3 scale (Ref 65). They also tested alloys 309
0.1
0.5
1.0 Scaling loss
800 and 801 in the same environment. Alloy 310
0.1
0.5 Internal penetration
800 sample was totally corroded after 100 h at 1.0
0.1
NT Not tested
314 0.5% H2S Welded specimens
980 °C (1800 °F), while the alloy 801 sample 0.5
1.0
NT

0.1
showed an adherent oxide scale with only about 316 0.5
1.0
1.4 mg/cm2 weight gain. The only apparent dif- 800
0.1
0.5
1.0
ference in the chemical composition between the 0.1
793 0.5
two alloys is titanium (about 0.4% for alloy 800 1.0
0.1
and about 1.1% for alloy 801). Additional data 600 0.5
1.0 NT
Completely corroded 125

0.1
indicating the beneficial effect of titanium are 601 0.5
1.0
Completely corroded
43.8
125
75.1
shown in Table 7.7 (Ref 64). Results obtained by 671
0.1
0.5
1.0
310 0.1
0.5
(Al) 1.0
0.1
Temperature, °C 800
(Al)
0.5
1.0
482 650 816 900 982 310 0.1
(Cr) 0.5 NT
1.0
Completely 800 0.1
0.5 NT
(Cr)
Sound metal loss, mils/yr

corroded 1.0

>50 446 0 2 4 6 8 10 12 14 16 18 20
Total corrosion, mil

0 50 100 150 200 250 300 350 400 450 500

20–50 Total corrosion, µ m

2–20 671 Fig. 7.32 Corrosion of stainless steels and nickel-base alloys
657 at 816 °C (1500 °F) for 100 h in the MPC coal
446 gasification atmosphere with 0.1, 0.5, and 1.0% H2S. Also
<2 671 included were aluminized Type 310 and alloy 800 [310 (Al) and
657 800 (Al], and chromized Type 310 and alloy 800 [310 (Cr) and
800 (Cr)] (see Tables 7.4 and 7.5 for gas composition). Source:
900 1200 1500 1650 1800
Ref 67
Temperature, °F

Fig. 7.30 Corrosion of Fe-27Cr (Type 446) and 50Ni-50Cr

alloys (alloy 671 and 657) in the MPC coal gasifi- Alloy Time,
cation stmosphere with 1.0 and 1.5% H2S (see Tables 7.4 and 7.5 h
1000
310 0.9 1.5
for gas composition). Source: Ref 60 2000
1000
2000
HK-40 5000
10,000
1000
2000
Cru-25 5000
10,000
Temperature, °C N155
1000
2000
5000
482 650 816 900 982 10,000
1000
2000
T63wC 5000
Completely 10,000
1000
Sound metal loss, mils/yr

corroded 188-Breakway occurs RA333 2000

5000 2.0
in 2000–10,000 h 7278
1000
>50 2000
Alloy X 5000
10,000
1000
20–50 556 6B
2000 Legend
5000
10,000 Scaling
1000
2000 Penetration
188 5000
2–50 N155 10,000
1000
6B 6B 2000
800 5000
556 188 7000
8000 0.67
<2 N155 RV-18 & 19 1000
RV-18 & 19 671 2000
5000
10,000
900 1200 1500 1650 1800
0 2 4 6 8 10 12 14 16 18 20
Temperature, °F Total corrosion, mil

0 50 100 150 200 250 300 350 400 450 500

Fig. 7.31 Corrosion of cobalt-base alloys (alloys 188 and 6B) Total corrosion, µm
and cobalt-containing alloys (alloys 556, N155, Fig. 7.33 Corrosion of Fe-Cr-Ni, nickel-base, and cobalt-base
RV-18, and RV-19) in the MPC coal gasification atmosphere with alloys at 980 °C (1800 °F) in the MPC coal gasifi-
1.0 and 1.5% H2S (see Tables 7.4 and 7.5 for gas composition). cation atmosphere with 0.5% H2S (see Tables 7.4 and 7.5 for gas
Source: Ref 60 composition). Source: Ref 67
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:14AM Plate # 0 pg 222

222 / High-Temperature Corrosion and Materials Applications

Lai (Ref 71) also revealed the beneficial effect of nickel-base alloys tested. In fact, they approa-
titanium in resisting sulfidation attack. As shown ched some cobalt-base alloys. Alloy 263 (2.5%
in Fig. 7.38, alloy R-41 and Waspaloy alloy Ti), while performing well at 760 °C (1400 °F),
(both contain about 3% Ti) were the best of the suffered severe sulfidation attack at 870 and
980 °C (1600 and 1800 °F) (Ref 71).
Bradshaw et al. (Ref 64) also examined the
effects of molybdenum and aluminum. Addition
of 6% Mo to high-purity Type 310SS

Temperature, °C
700 600 500
100

Corrosion after 1 year, mils

Corrosion after 1 year, mm

Alloy 600 1

Alloy 625
Alloy 25
10
Alloy DS
AISI 310 0.1
AISI 446

1
1.00 1.04 1.08 1.12 1.16 1.20 1.24 1.28
1000/(Temperature), 1/K

5000 Temperature, °C
700 600 500
100
HK4M
Fig. 7.34
Corrosion after 1 year, mils

Total surface recession including internal attack for

Corrosion after 1 year, mm

Cr30A
carbon steel and 1Cr-0.5Mo steel at 330, 400, and 1
500 °C (626, 752, and 932 °F) in CO-51.2H2-0.8CO2-1.7H2S- Alloy 556
0.2HCl. Source: Ref 68
10
Alloy 188
Alloy 150
0.1
Cr35A

1
1.00 1.04 1.08 1.12 1.16 1.20 1.24 1.28
1000/(Temperature), 1/K

Fig. 7.36 Total surface recession including internal attack for

Fe-, Ni-, and Co-base alloys at 700 °C (1292 °F)
and lower in CO-26H2-1.0CO2-0.9H2S-2H2O-10N2. Source:
Ref 68

Table 7.6 Test gas environments (non-

equilibrium) used for investigating the effect of
water content on corrosion behavior of Type 310
and alloy 800 at 540 °C (1004 °F) for 600 h
Gas composition, vol %
Component No. 1 No. 2 No. 3 No. 4 No. 5 No. 6 No. 7
H 2O 0.0 1.0 2.0 3.0 5.0 10.0 15.0
H2 32.0 30.2 29.2 28.8 28.2 28.2 30.0
CO 64.0 64.0 64.0 64.0 60.0 52.0 45.0
CO2 3.2 4.0 4.0 4.0 6.0 9.0 9.2
H 2S 0.8 0.8 0.8 0.8 0.8 0.8 0.8
HCl 0.04 0.04 0.04 0.04 0.04 0.04 0.04
pO2 at 10−29 10−28.4 10−27.8 10−27.4 10−7 10−26.4 10−26
540 °C
Fig. 7.35 Total surface recession including internal attack
for alloy 625 at 330 to 700 °C (626 to 1292 °F) in Source: Ref 70
CO-51.2H2-0.8CO2-1.7H2S-0.2HCl. Source: Ref 68
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:15AM Plate # 0 pg 223

Chapter 7: Sulfidation / 223

significantly improved sulfidation resistance stability issue when used at high temperatures
(Table 7.7). However, a layer of (Fe, Ni} sulfides because of formation of intermetallic phases
was observed on top of the chromium oxide during the long-term, high temperature expo-
scale. Thus, formation of molten sulfide slag is sures. However, application of these alloys at
expected upon further exposure. No explanation temperatures below 650 °C (1200 °F) may not
was given in the paper by Bradshaw et al. raise this thermal stability issue.
(Ref 64) on the beneficial effect of molybdenum. Addition of 3% Al to high-purity Type 310SS
The test temperature used by Bradshaw et al. significantly improved sulfidation resistance, as
(Ref 64) was excessively high (i.e., 1000 °C). shown in Table 7.7. However, this composition
It would be of interest if the tests were conducted was not considered worthwhile for further
at much lower temperatures. Natesan (Ref 72) development because of expected fabrication
found that molybdenum and TZM (molybdenum problems (Ref 64).
with 0.5Ti and 0.04Zr) formed very thin adherent Manganese has a deleterious effect on an
sulfide scales when tested at 871 °C (1600 °F) in alloy’s sulfidation resistance. In their tests in
a sulfidizing environment with low oxygen and Ar-40H2-15H2O-1.4H2S at 1000 °C (1830 °F)
high sulfur potentials. Test results are shown in for 24 h, Bradshaw et al. (Ref 64) found
Fig. 7.39 (Ref 72). He et al. (Ref 73) showed that that addition of 2% Mn adversely affected sulfi-
Ni-10Mo was more resistant to sulfidation than dation resistance of Ni-30Cr and Type 310SS,
nickel (Fig. 7.40) in short-term tests. The bene- particularly a high-purity 310SS (Table 7.8).
ficial effect of molybdenum in improving the Manganese diffuses the quickest through chro-
alloy sulfidation resistance in the environments mium oxide scale to form external sulfides on
with low oxygen, high sulfur potentials would top of it, thus causing accelerating corrosion
be of interest, since molybdenum has been used (Ref 14).
in many high-temperature nickel-base alloys to The effect of silicon was investigated by
provide solid solution strengthening for the Nagarajan et al. (Ref 74) in Fe-18Cr alloys in
alloy. Some of these wrought alloys include 24H2-39H2O-18CO-12CO2-5CH4-1H2S-1NH3.
alloys X, S, 617, 625, and R-41. Furthermore, The Fe-18Cr-2Si alloy exhibited sulfidation
molybdenum is also a major alloying element resistance significantly better than that of
in Ni-Cr-Mo corrosion-resistant alloys, such as Fe-18Cr-0.5Si alloy (i.e., 20 mg/cm2 versus
C-276, C-22, (622), 59, C-2000, and 686. These slightly more than 200 mg/cm2) after 120 h at
Ni-Cr-Mo alloys contain molybdenum in a range 980 °C (1800 °F) (pO2 =9:9 · 10716 atm, pS2 =
of 13 to 16%. High levels of molybdenum 2:4· 1076 atm, and ac = 0.3 at the test tempera-
in these Ni-Cr-Mo alloys can cause a thermal ture). A nickel-base wrought alloy (HR160 alloy)

Metal loss (µm)

200

Alloy 800

100

SS 310

20-35-3.25
0
0 5 10 15
% H2O

Fig. 7.37 Effect of H2O on the metal loss of Type 310, alloy 800, and experimental alloy Fe-20Cr-35Ni-3.25Si (20-35-3.25) tested at
540 °C (1004 °F) for 600 h. The test gas environments are shown in Table 7.6. Source: Ref 70
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:15AM Plate # 0 pg 224

224 / High-Temperature Corrosion and Materials Applications

was developed by Lai (Ref 75) using a combi- at 870 °C (1600 °F) (Ref 77). The sulfidation
nation of silicon, chromium, and cobalt to max- resistance of alloy HR160 was quite comparable
imize the alloy’s sulfidation resistance as well as to that of cobalt-base alloy 6B and significantly
its metallurgical stability, creep-rupture proper- better than other cobalt-base alloys, such as
ties, and weldability. With high chromium (28%) alloys 188, 25, and 150 (Table 7.9). Norton et al.
and high silicon (2.75%), a protective oxide scale (Ref 78) conducted corrosion tests on several
consisting of a chromium oxide scale with Fe-Ni-Cr and Ni-base alloys in comparison
underlying silicon enrichment formed on the with alloy HR160 at 700 °C (1290 °F) for
alloy surface (Ref 76, 77). In addition, cobalt up to 1000 h in H2-7CO-1.5H2O-0.6H2S
provided further improvement in the alloy’s (pO2 =10723 atm, pS2 =1079 atm, and ac = 0.3
sulfidation resistance. Figure 7.41 shows the to 0.4). Their test environment, as located in the
sulfidation resistance of alloy HR160 compared phase stability diagram at the test temperature, is
to some familiar alloys, such as alloys 556, shown in Fig. 7.42 (Ref 78), and their test results
800H, and 600 in Ar-5H2-5CO-1CO2-0.15H2S are summarized in Fig. 7.43 (Ref 78).

Fig. 7.38 Corrosion of iron-, nickel-, and cobalt-base alloys after 215 h at (a) 760 °C (1400 °F), (b) 870 °C (1600 °F), and (c) 980 °C
(1800 °F) in Ar-5H2-5CO-1CO2-0.15H2S. Source: Ref 71
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:15AM Plate # 0 pg 225

Chapter 7: Sulfidation / 225

Norton et al. (Ref 79) also conducted corrosion the oxide scale. On the other hand, a thin, con-
tests on an alumina-forming alloy MA956 tinuous and adherent Si-rich oxide scale enriched
(Fe-20Cr-4.5Al-0.5Y2O3), which is an in titanium in the outer region of the oxide scale
oxide-dispersion-strengthened alloy. Also inclu- formed on HR160, and Cr-rich sulfides grew on
ded in the tests were HR160, 45TM (Ni-Cr-Fe- the top of the oxide scale (Ref 80). The corrosion
Si), HR 3C (Fe-25Cr-20Ni-0.5Nb-0.2N), and an
experimental alloy (Fe-26Cr-39Ni-3.2V). Tests
were conducted in CO-32H2-3.8CO2-0.2H2S at
600 °C (1112 °F) and the test environment in
terms of pO2 and pS2 potentials in the phase-
stability diagram as shown in Fig. 7.44. Their
test results are summarized in Fig. 7.45 (Ref 80).
MA956 was found to be slightly less resistant
compared with HR160 and 45TM. Norton and
Levi (Ref 80) reported that a mixed Cr-Al oxide
scale with Cr-rich oxide on the outer layer and
Al-rich oxide on the inner layer formed on
MA956, and Fe-rich sulfides grew on the top of

(a)
Table 7.7 Results of corrosion tests at 1000 °C
(1832 °F) for 100 h in Ar-30H2-30H2O-1H2S(a)
Total affected
Alloy depth, μm (mils) Comments
310SS 92 (3.6) Internal sulfidation
810 (31.9) Liquid sulfides
605 (23.8) Liquid sulfides
310HP(b) >1500 (59.1) Liquid sulfides
650 (25.6) Sulfide penetration
310HP + 2% Ti 330 (13.0) Sulfide penetration
62 (2.4) Adherent oxide
38 (1.5) Adherent oxide
310HP + 3% Ti 38 (1.5) Adherent oxide
34 (1.3) Adherent oxide
310HP + 6% Mo 12 (0.5) Spalling oxides and (Fe.Ni)
sulfides over Cr2O3
310HP + 3% Al 5 (0.2) Spalling oxides (b)
5 (0.2) Spalling oxides
(a) Test gas was at 1 atm; pO2 =3 · 10715 atm and pS2 =3 · 1076 atm: (b) HP Fig. 7.40 Weight change data comparing nickel (a) with Ni-
indicates high-purity material. Source: Ref 64 10Mo alloy (b) when tested at 550 and 600 °C
(1022 and 1112 °F) in Ar-13.56H2-0.6H2O-1.89H2S (pO2 = 2.7 ×
10 , pS2 = 4.1 × 10−8 at 600 °C; pO2 = 4.5 × 10−29, pS2 = 1 × 10−8 at
−27

550 °C). Source: Ref 73

2.0
Weight change, mg/mm2

Test temperature 871 °C Table 7.8 Results of corrosion tests at 1000 °C

1.6
Alloy 800 p = 4.1 x 10–18 atm (1832 °F) for 24 h in Ar-40H2-15H2O-1.4H2S(a)
O
p 2 = 9.4 x 10–7 atm
1.2 S2 Total affected
Alloy depth, µm (mils) Comments

0.6 310SS 56 (2.2) Sulfide layer

310 SS Oxide layer
Intergranular sulfides
0.4 Ta Internal sulfides
Nb TZM Mo
310SS + 2% Mn 72 (2.8) Sulfide layer
v
0 Oxide layer
0 30 60 90 120 150 Internal sulfides
Exposure time, h 310HP(b) + 2% Mn 420 (16.5) Liquid sulfides
Ni-30Cr 38 (1.5) Oxides
Internal sulfides
Fig. 7.39 Thermogravimetric data for refractory metals, such
Ni-30Cr-2Mn 1220 (48.0) Liquid sulfides
as, Mo, TZM, Nb, Ta, and V in comparison with
conventional alloys in a sulfidizing environment with high sulfur (a) Test gas was gas at 1 atm; pO2 =3 · 10716 and pS2 =3 · 1076 atm: (b) HP
and low oxygen potentials tested at 871 °C (1600 °F). Source: indicates high-purity material. Source: Ref 64
Ref 72
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:15AM Plate # 0 pg 226

226 / High-Temperature Corrosion and Materials Applications

morphologies for both alloys after exposure for binary alloys in SO2 and SO2-O2 environments.
2000 h are schematically illustrated in Fig. 7.46 Metals and alloys exposed to these environments
(Ref 80). Additional low temperature sulfidation generally form oxides and/or sulfides as corro-
data generated in reducing environments can be sion products. Corrosion products and corros-
found in Tables 10.5 and 10.6 and Figures 10.17 ion rates depend strongly on temperature. The
and 10.18.

Table 7.9 Sulfidation resistance of Alloy HR160

compared to cobalt-base alloys at 870 °C (1600 °F)
7.5 Corrosion in SO2-Bearing for 500 h(a)
Environments Weight change, Metal loss, Maximum metal
Alloy mg/cm2 mm (mils) affected(b), mm (mils)
Most corrosion studies of SO2-bearing en- HR160 −0.5 0.01 (0.2) 0.13 (5.2)
vironments have been conducted in either SO2 or 556 183.6 0.52 (20.6) 0.90 (35.6)
188 40.6 0.19 (7.6) 0.60 (23.6)
O2-SO2 mixtures. Materials investigated include 25 33.0 0.10 (4.1) 0.37 (14.6)
nickel (Ref 81–84), iron (Ref 85–87), cobalt 150 131.7 0.26 (10.3) 0.72 (28.3)
(Ref 88–89), chromium (Ref 90), Co-Cr alloys 6B 10.7 0.01 (0.3) 0.08 (3.3)

(Ref 91), and Ni-Cr alloys (Ref 92–97). Kofstad (a) Ar-5H2-5CO-lCO2-0.15H2S; pO2 =3 · 10719 atm, pS2 =0:9 · 1076 atm:
(b) Metal loss + maximum internal penetration. Source: Ref 76
(Ref 98) reviews the corrosion of pure metals and

Fig. 7.41 Sulfidation resistance of alloy HR160 compared to those of alloys 556, 800H, and 600 after 215 h at 870 °C (1600 °F) in
Ar-5H2-5CO-1CO2-0.15H2S (pO2 = 3 × 10−19 atm, pS2 = 0.9 × 10−6 atm). Samples were cathodically descaled before
being mounted for metallographic examination. Source: Ref 77
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:15AM Plate # 0 pg 227

Chapter 7: Sulfidation / 227

reaction rate generally peaks at a certain tem- reaction between SO2 and O2 leads to formation
perature, then decreases with increasing tem- of SO3. The kinetics for this reaction are rela-
perature. The highest corrosion rate is normally tively slow without catalysts such as platinum.
related to the formation of sulfides. Sulfides Most tests were conducted in conjunction with
provide paths for rapid outward diffusion of platinum catalysts to obtain an equilibrium SO3
metals, such as nickel, iron, and chromium, and partial pressure. SO3 is an important reactant in
so forth, resulting in rapid corrosion attack the corrosion reaction involving SO2 (Ref 83,
(Ref 98). The rate is particularly rapid when 93). Corrosion rate was found to be dependent on
liquid sulfide eutectics are formed. The Ni-Ni3S2 the ratio of SO2:O2. When that ratio is 2:1, which
eutectic melts at 635 °C (1175 °F), the Fe-FeS gives the highest partial pressure of SO3, the
eutectic at 985 °C (1805 °F), and the Co-Co4S3 corrosion rate is generally fastest, as illustrated in
eutectic at 880 °C (1616 °F) (Ref 17). Fig. 7.47 (Ref 83). The corrosion reaction is also
The temperature at which the corrosion rate is temperature dependent. Corrosion rate increases
highest varies from metal to metal. For nickel, it with increasing temperature until a maximum is
is around 600 °C (1110 °F). At temperatures reached. Further increases in temperature result
above 800 °C (1470 °F), the rate decreases with in a reduced corrosion rate (Ref 83). With an
increasing temperature (Ref 98). Sulfides SO2:O2 ratio of 2:1, nickel corroded fastest at
become less and less stable as temperature 700 to 800 °C (1290 to 1470 °F) (Ref 83).
increases. Eventually, NiO becomes the only
corrosion product (Ref 98). For cobalt, the
corrosion rate is highest at 920 °C (1688 °F), 250
Type 321 SS
and decreases above that temperature (Ref 98). 200
The corrosion rate for iron is high above
∆W, mg/cm2

940 °C (1724 °F) (Ref 98), which corresponds 150

Type 347 SS
approximately to the iron sulfide eutectic tem- 100
perature. Chromium, on the other hand, does not Alloy 800H
form sulfides in SO2 at temperatures from 700 to 50
1000 °C (1290 to 1830 °F), but instead forms a 0
Cr2O3 scale (Ref 90). Thus, chromium as an 0 200 400 600 800 1000
alloying element improves sulfidation resistance Exposure, h
(a)
in Co-Cr (Ref 91) and Ni-Cr alloys (Ref 92).
Corrosion tests in environments containing 100
both O2 and SO2 were conducted primarily
80
in SO2-O2 mixtures with various ratios. The Alloy 625
∆W, mg/cm2

60 Alloy X
0 40 Alloy 617
FeS2
20
NiS
0
–5
0 200 400 600 800 1000
Co9S8 Exposure, h
Cr2S3 Test gas (b)
oO
Log pS , bar

C
2

–10 Ni3S2 FeS Oo Ni 10

Fe Type 347 SS
8
FeO Alloy 800H
∆W, mg/cm2

CrS
Fe3O4 6 Alloy HR-120
–15 Cr Cr2O3
NiO 4 Alloy 556

2 Alloy HR-160
–20
–35 –30 –25 –20 –15 0
Log pO , bar 0 200 400 600 800 1000
2
Exposure, h
(c)
Fig. 7.42 Test environment in terms of pO2 and pS2 is plotted
in the stability diagram at 700 °C (1202 °F) Source: Fig. 7.43 Resuts of corrosion tests in H2-7CO-1.5H2O-
Ref 78 0.6H2S at 700 °C (1292 °F). Source: Ref 78
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:15AM Plate # 0 pg 228

228 / High-Temperature Corrosion and Materials Applications

Fig. 7.44 The test environment in terms of pO2 and pS2 potentials at 600 °C in CO-32H2-3.8CO2-0.2H2S is plotted as an equilibrium
condition (identified as “E”) and a nonequilibrium (NE) is plotted in the phase-stability diagram. Source: Ref 79

8
7 26/37/3
weight gain, mg/cm2

6
HR3C
5
4
3 MA960 (ave)
2
1 45 TM
MA956 (low)
0
0 500 1000 1500 2000 HR160
time, h

Fig. 7.45 Corrosion behavior of MA956 in comparison with HR160, 45TM, HR3C, and an experimental alloy 26Cr/37Ni/3V (26/37/3)
in CO-32H2-3.8CO2-0.2H2S at 600 °C (1112 °F). Source: Ref 79

diffusion of nickel. At sufficiently high tem-

peratures, NiO becomes the only stable corrosion
product (Ref 98). Then the corrosion rate is
lowered. Ni-20Cr alloy in SO2-O2 mixture
behaved similarly to nickel (Ref 98). Vasantasree
and Hocking (Ref 93) investigated Ni-Cr alloys
with various chromium contents. Their results
are shown in Fig. 7.48. Rates were highest at
700 °C (1290 °F) for Ni-Cr alloys. Formation of
Fig. 7.46 Corrosion scales formed on MA956 (left picture) oxides was favored at high temperatures, parti-
and HR160 (right picture) after exposure at 600 °C cularly 900 and 1000 °C (1650 and 1830 °F).
(1112 °F) for 2000 h in CO-32H2-3.8CO2-0.2H2S. For MA956,
the oxide scale (in black) was a mixed Cr-Al-rich oxide scale with Therefore, corrosion rates became much lower at
the outer layer being chromium-rich oxide scale and the inner these temperatures.
layer being aluminum-rich oxide scale. For HR160, the oxide
scale (in black) was silicon-rich oxide enriched with titanium on
Very few investigators have tested commercial
the outer layer of the oxide scale. Source: Ref 80 alloys in SO2-bearing environments. Barnes and
Lai (Ref 99) examined the oxidation behavior of
two iron-base commercial alloys, Type 304SS
Nickel forms nickel oxides, nickel sulfides, and and alloy 556, in Ar-5O2-5CO2 with and without
NiSO4 when the corrosion rate is highest 10% SO2 at 980 °C (1800 °F). The equilibrium
(Ref 98). The sulfides provide paths for outward partial pressures for the Ar-5O2-5CO2-10SO2 at
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:15AM Plate # 0 pg 229

Chapter 7: Sulfidation / 229

the test temperature were 4.9 × 10−2 atm for pO2 ,

9.7 × 10−2 atm for pSO2 , 3.6 × 10−3 atm for pSO3 ,
and 1 × 10−22 atm for pS2 . In the SO2-bearing test
environment, Type 304 sample suffered severe
oxidation/sulfidation attack, with some areas
being completed consumed by corrosion attack.
On the other hand, alloy 556 formed a thin,
compact chromium-rich oxide scale with some
internal chromium-rich sulfides underneath the

p
oxide scale. Table 7.10 summarizes the corrosion
p attack for both alloys in both environments. The
results suggest that an alloy that forms a protec-
tive chromium oxide scale in a purely oxidizing
environment is likely to form a similar scale in
SO2-O2 environment. Many industrial processes
that generate SO2-bearing environments are
generally at much lower temperatures. Yates
et al. (Ref 100) tested alloys X (identified
as HX Ni-22Cr-9Mo-18Fe), 617 (Ni-22Cr-
Fig. 7.47 Initial linear rate constants of nickel in different 12Co-9Mo-1.2Al), 230 (Ni-22Cr-14W-La), 188
SO2-O2 mixtures at 1 atm and 800 °C (1470 °F). (Co-22Cr-20Ni-14W-La), and 214 (Ni-16Cr-
Also shown are equilibrium SO3 pressures at different SO2-O2 3Fe-4.5Al-Y) at 704 °C (1300 °F) in O2- 4%
mixtures. Source: Ref 83
SO2 for more than 40 days. Their test results are
summarized in Fig. 7.49. Based on these test
results, Ni-Cr and Co-Cr alloys containing
700° about 22% Cr are considered to have adequate
SO2:O2::2:1 corrosion resistance in O2-4%SO2 at 704 °C
Temp. °C
600 (1300 °F). All four 22Cr alloys (X, 617, 188, and
700 230) exhibited a parabolic reaction kinetics with
103 800 800°
900 low mass changes over more than 40 days of
1000
L=Linear exposure, indicating formation of protective
chromium-rich oxide scales. On the other hand,
Ni-Cr-Al alloy 214 with only about 16% Cr
Parabolic rate constant, mg2/cm4/h

102
showed some indication of breakaway corrosion.
The test temperature of 704 °C (1300 °F) was
600° likely too low for rapid formation of Al2O3 in
O2-SO2 mixtures. The amount of chromium
101 (about 16%) is considered to be inadequate in
forming protective Cr2O3 scales under the test
condition. Field test data, which was generated
900° in a chemical plant with the environment con-
100 taining about 18% SO2 in the temperature range
of 260 to 371 °C (500 to 700 °F), showed
1000° minimal corrosion rates (about 0.3 mpy and less)
for Type 316, 317, alloy 20, and alloy 825
10–1 (Ref 101).
The levels of SO2 used for most corrosion
studies in either SO2 environments or SO2-O2
mixtures are significantly higher than the
amounts expected in the combustion of sulfur-
10–2
100 50 20 10 1 0.1 0 bearing fuels such as coal or oil. Typical com-
Cr at.%, Cr Ni bustion flue gas produced in a coal-fired boiler,
for example, contains approximately 0.25% SO2
Fig. 7.48 Parabolic rate constants of Ni-Cr alloys as a func-
(Ref 102). Very few investigators have studied
tion of chromium concentration in SO2:O2 mixture
(2:1 ratio) at various temperatures. Source: Ref 93 SO2 corrosion at this low level (i.e., less than
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:15AM Plate # 0 pg 230

230 / High-Temperature Corrosion and Materials Applications

Table 7.10 Corrosion of Type 304SS and alloy 556 at 980 °C (1800 °F) for 550 h in an oxidizing
environment with and without SO2
Type 304SS Alloy 556
Metal loss, Maximum depth of Metal loss, Maximum depth of
Test environment mm (mils) attack, mm (mils)(a) mm (mils) attack, mm (mils)(a)
Ar-5O2-5CO2 0.31 (12.2) 0.44 (17.2) 0.005 (0.2) 0.056 (2.2)
Ar-5O2-5CO2-10SO2 > 0.61 (24) > 0.61 (24) 0.06 (2.5) 0.10 (4.0)
(a) Metal loss + maximum internal penetration. Source: Ref 99

Table 7.11 Corrosion of Ni-15Cr alloy in

SO2-bearing environments at 870 °C (1600 °F)
for 30 h
Weight loss, Depth of attack,
Test environment mg/cm2 mm (mils)
Pure SO2 0.6 0.008 (0.3)
N2-1.35SO2 1.8 …
N2-1SO2 2.0 …
N2-0.2SO2 86.0 0.52 (20.5)
N2-0.05SO2 13.7 …
N2-0.2SO2-0.01O2 6.5 …
N2-0.2SO2-0.1O2 0.3 …
N2-0.2SO2-2O2 0.3 …
N2-0.2SO2-3O2 … 0.008 (0.3)
Source: Ref 103

Fig. 7.49 Mass change as a function of exposure time of SO2 in the environment with no “free” oxygen
for alloys X (identified as HX) 617, 230, 188 and can cause Ni-Cr alloys to suffer severe sulfida-
214 at 704 °C (1300 °F) in O2-4% SO2. Source: Ref 100
tion attack even without ash/salt deposits.

1%). However, studies have been conducted of 7.6 Summary

corrosion in environments containing low con-
centrations of SO2 in conjunction with ash/salt The sulfidation behavior of a wide variety of
deposits in order to simulate fuel ash corrosion in alloys was reviewed. The data presented are
combustion systems, particularly fossil-fired primarily related to gaseous environments;
boilers. These data are discussed in later chapters sulfate-accelerated sulfidation (hot corrosion) is
covering specifically on coal-fired boilers and covered in Chapter 9. The corrosion data are
oil-fired boilers. Viswanathan and Spengler grouped into three different types of environ-
(Ref 103) observed that Ni-15Cr alloy suffered ments: (1) sulfur vapor, hydrocarbon streams
significantly more corrosion attack in 0.2% SO2 containing no hydrogen gas, H2S, and H2-H2S,
(balance N2) than in pure SO2 at 870 °C (2) reduced, mixed gas environments with low
(1600 °F). The sample exposed to N2-0.2SO2 at oxygen and high sulfur potentials, and (3) SO2-
870 °C (1600 °F) for 30 h showed about bearing environments.
0.5 mm (20 mils) of attack with internal globular Sulfur vapor, hydrocarbon streams (no H2),
sulfide phases, while the sample exposed to H2S, and H2-H2S environments have a common
100% SO2 for same temperature and time feature in that the corrosion products formed in
exhibited only about 0.008 mm (0.3 mil) of these environments are essentially sulfides. The
attack. Addition of oxygen significantly reduced reduced, mixed gas environments with low
the alloy’s corrosion rates. The weight loss for oxygen and high sulfur potentials typically con-
the sample after 30 h fell from 86 mg/cm2 in N2- tain H2, CO, CO2, H2O, H2S, and other gaseous
0.2SO2 to about 0.3 mg/cm2 in N2-0.2SO2- components where oxides and sulfides can form
0.1O2. The results of their tests (Ref 103) are on most high temperature alloys. In these envir-
summarized in Table 7.11. Low concentrations onments, alloys are protected by a protective
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:15AM Plate # 0 pg 231

Chapter 7: Sulfidation / 231

oxide scale before the initiation of breakaway 14. R.A. Perkins, in Environmental Degrada-
corrosion due to formation of fast-growing, tion of High Temperature Materials, Series
nonprotective sulfides. The SO2-bearing envir- 3, Vol 2 (No. 13), 1980, p 5/1
onments are generated by combustion of a sulfur- 15. P.L. Hemmings and R.A. Perkins,
containing fuel or feedstock with excess air or “Thermodynamic Phase Stability Dia-
oxygen. These environments are oxidizing, and grams for the Analysis of Corrosion
are generally less corrosive than reducing Reactions in Coal Gasification/Combus-
environments. tion Atmospheres,” EPRI Report FP-539,
Lockheed Palo Alto Research Labora-
tories, Palo Alto, CA, 1977
REFERENCES 16. B.A. Gordon and V. Nagarajan, Oxid. Met.,
Vol 13 (No. 2), 1979, p 197
1. G.Y. Lai, J. Met., July 1985, p 14 17. M. Hansen and K. Anderko, Constitution
2. G.L. Swales, in Behavior of High Tem- of Binary Alloys, McGraw-Hill, 1958
perature Alloys in Aggressive Environ- 18. S. Mrowec and K. Przybylski, High
ments, I. Kirman et al., Ed., Proc. Petten Temp. Mater. Proc., Vol 6 (No. 1 and 2),
International Conference, Oct 15–18, 1984, p 1
1979, The Metals Society, London, 1980, 19. D.J. Young, Rev. High Temp. Mater., Vol 4
p 45 (No. 4), 1980, p 299
3. G. Sorell, M.J. Humphries, E. Bullock, and 20. A. Davin and D. Coutsouradis, Cobalt, Vol
M. Van de Voorde, Int. Met. Rev., Vol 31 17, 1962, p 23
(No. 5), 1986, p 216 21. S. Mrowec, T. Walec, and T. Werber, Oxid.
4. J.F. Norton, Ed., High Temperature Mate- Met., Vol 1, 1969, p 93
rials Corrosion in Coal Gasification 22. T. Narita, W.W. Smeltzer, and K. Nihida,
Atmospheres, Elsevier, Amsterdam, 1984 Oxid. Met., Vol 17, 1982, p 299
5. K.J. Barton, V.L. Hill, and R. Yurkewycz, 23. T. Narita and K. Nihida, Oxid. Met., Vol 6,
in The Properties and Performance of 1973, p 157 and 181
Materials in the Coal Gasification Envir- 24. S.K. Mrowec, T. Werber, and M. Zas-
onments, V.L. Hill and H.L. Black, Ed., tawnik, Corros. Sci., Vol 6, 1966, p 47
American Society For Metals, 1981, p 65 25. D.P. Whittle, S.K. Verma, and J. Stringer,
6. S.K. Srivastave, G.Y. Lai, and D.E. Fluck, Corros. Sci., Vol 13, 1973, p 247
Paper No. 398, Corrosion/87, NACE, 1987 26. T. Biegun, A. Bruckman, and S. Mrowec,
7. J.A. Harris, W.G. Lipscomb, and G.D. Oxid. Met., Vol 12, 1978, p 157
Smith, Paper No. 402, Corrosion/87, 27. S. Mrowec and M. Wedrychowska, Oxid.
NACE, 1987 Met., Vol 13, 1979, p 481
8. J. Stringer, in High Temperature Corro- 28. E.M. Jallouli, J.P. Larpin, M. Lambertin,
sion, R.A. Rapp, Ed., Conference Pro- and J.C. Colson, J. Electrochem. Soc., Vol
ceedings (San Diego, CA) March 2–6, 126, 1979, p 2254
1981, NACE, 1981, p 389 29. T. Biegun and A. Bruckman, Bull. Acad.
9. A.J. Minchener, D.M. Lloyd, and P.T. Polon. Ser. Sci. Chim., Vol 28, 1980, p 377;
Sutcliffe, “Materials Evaluation for Flui- Vol 29, 1981, p 69
dized Bed Combustion Systems,” CS- 30. W.W. Smeltzer, T. Narita, and K. Przy-
3511, Final Report to EPRI on Research bylski, in Proc. Corrosion-Erosion, Wear
Project RP979-11, Electric Power of Materials in Emerging Fossil Energy
Research Institute, Palo Alto, CA, 1984 Systems, A.V. Levy, Ed., NACE, 1982,
10. J. Stringer, Paper No. 90, Corrosion/86, p 860
NACE, 1986 31. T. Biegun and A. Bruckman, “Sulfidation
11. S.R. Shatynski, Oxid. Met., Vol 11 (No. 6), of Ni-Cr-Al Alloys,” Report No. 2.34264,
1977, p 307 Institute of Physical Chemistry, Polish
12. HSC, Chemistry for Windows, Version 6.0, Academy of Science, Warsaw, 1980
A. Roine, Outokumpu Technology, Fin- 32. L.A. Morris, Chapter 17: Resistance to
land, www.outokumputechnology.com, Corrosion in Gaseous Atmospheres, in
accessed Dec 2006 Handbook of Stainless Steels, D. Peckner
13. ChemSage, Version 4.16, GTT-Techno- and I.M. Bernstein, McGraw-Hill, 1977,
logies, Aachen (1998) p 17.1
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:15AM Plate # 0 pg 232

232 / High-Temperature Corrosion and Materials Applications

33. E. Malinowski, R. Bigot, and E. Herzog, 45. G. Sorell and W.B. Hoyt, “Collection and
Le Metallurgic, Vol 94 (No. 4), 1962 Correlation of High Temperature Hydro-
34. F.J. Bruns, Corrosion of Ni-Cr-Al-Fe gen Sulfide Corrosion Data,” NACE
Alloys by Hydrogen Sulfide at 1100 to Technical Committee Report, Publication
1800 °F, Corrosion, Vol 25 (No. 3), 1969, 56-7, NACE, Houston, TX, 1956
p 119 46. E.B. Backensto, “Corrosion in Catalytic
35. Z.A. Foroulis, High Temperature Degra- Reforming and Associated Processes,
dation of Structural Materials in Environ- Summary Report of the Panel on Reformer
ments Encountered in the Petroleum and Corrosion to the Subcommittee on Corro-
Petrochemical Industries: Some Mechan- sion,” presented at the 22nd Midyear
istic Observations, Anti-Corrosion, Vol 32 Meeting of API’s Division of Refining
(No. 11), 1985, p 4–9 (Philadelphia, PA), May 13, 1957
36. J. Gutzeit, R.D. Merrick, and L.R. Scharf- 47. E. Dittrich, Chem. Fab., Vol 10 (No. 13/
stein, Corrosion in Petroleum Refining and 14), 1947, p 145
Petrochemical Operations, in Corrosion, 48. E.B. Backensto and J.W. Sjoberg,
Vol 13, 9th ed., Metals Handbook, ASM “Iso-Corrosion Rate Curves for High
International, 1987, p 1262 Temperature Hydrogen-Hydrogen Sul-
37. F.A. Hendershot and H.L. Valentine, fide,” Technical Committee Report,
Materials for Catalytic Cracking Equip- Publication 59-10, NACE, Houston, TX,
ment (Survey), Mater. Prot., Vol 6 (No. 10), 1958
1967, p 43 49. A.S. Couper and J.W. Gorman, Computer
38. J. Gutzeit, High Temperature Sulfidic Correlations to Estimate High Temperature
Corrosion of Steels, in Process Industries H2S Corrosion in Refinery Streams, Mater.
Corrosion—The Theory and Practice, Prot. Perform., Vol 10 (No. 4), 1971, p 31
NACE, 1986 50. V.L. Hill and H.S. Meyer, in High Tem-
39. The Role of Stainless Steels in Petroleum perature Corrosion in Energy Systems,
Refining, Originally published by the M.F. Rothman, Ed., The Metallurgical
Committee of Stainless Steel Producers, Society of AIME, 1985, p 29
AISI (1977), Nickel Development Insti- 51. A.O. Schaefer, C.H. Samans, M.A. Howes,
tute, Toronto, Ontario, Canada, April 1996 S. Bhattacharyya, E.R. Bangs, V.L. Hill,
40. Steels for Hydrogen Service at Elevated and F.C. Chang, “A Program to Discover
Temperatures and Pressures in Petroleum Materials Suitable for Service under
Refineries and Petrochemical Plants, Hostile Conditions Obtaining in Equip-
Publication 941, 3rd ed., American Petro- ment for the Gasification of Coal and
leum Institute, 1983 Other Solid Fuels,” 1975 Annual Report,
41. R.A. White and E.F. Ehmke, Materials for The Metal Properties Council, New York,
Refineries and Associated Facilities, 1976
NACE, 1991, p 51 52. A.O. Schaefer, “A Program to Discover
42. B.W. Neumaier and C.M. Schillmoller, Materials Suitable for Service Under
“How Richfield Plans to Combat High- Hostile Conditions Obtaining in Equip-
Temperature Sulfide Corrosion in Its ment for the Gasification of Coal and Other
New Catalytic Reformer,” presented at the Solid Fuels,” 1976 Annual Report, The
21st Midyear Meeting of the American Metal Properties Council, New York, 1977
Petroleum Institute’s Division of Refining 53. A.O. Schaefer, “A Program to Discover
(Montreal, Canada), May 14, 1956 Materials Suitable for Service Under Hos-
43. G. Sorell, “Compilation and Correlation of tile Conditions Obtaining in Equipment for
High Temperature Catalytic Reformer the Gasification of Coal and Other Solid
Corrosion Data,” Technical Committee Fuels,” 1977 Annual Report, The Metal
Report, Publication 58-2, NACE, Houston, Properties Council, New York, 1978
TX, 1957 54. A.O. Schaefer, “A Program to Discover
44. E.B. Backensto, R.E. Drew, J.E. Prior, and Materials Suitable for Service Under Hos-
J.W. Sjoberg, “High-Temperature Hydro- tile Conditions Obtaining in Equipment for
gen Sulfide Corrosion of Stainless Steels,” the Gasification of Coal and Other Solid
Technical Committee Report, Publication Fuels,” 1978 Annual Report, The Metal
58-3, NACE, Houston, TX, 1957 Properties Council, New York, 1979
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:15AM Plate # 0 pg 233

Chapter 7: Sulfidation / 233

55. A.O. Schaefer, “A Program to Discover 67. J.L. Blough, V.L. Hill, and B.A. Hum-
Materials Suitable for Service Under Hos- phreys, in The Properties and Performance
tile Conditions Obtaining in Equipment for of Materials in the Coal Gasification
the Gasification of Coal and Other Solid Environment, V.L. Hill and H.L. Black,
Fuels,” Supplement to 1978 Annual Ed., American Society For Metals, 1981,
Report, The Metal Properties Council, New p 225
York, 1979 68. R.C. John, W.C. Fort III, and R.A. Tait,
56. A.O. Schaefer, “A Program to Discover Prediction of Alloy Corrosion in the Shell
Materials Suitable for Service Under Hos- Coal Gasification Process, Mater. High
tile Conditions Obtaining in Equipment Temp., Vol 11 (No. 1–4), 1993, p 124
for the Gasification of Coal and Other Solid 69. W.T. Bakker, Effect of Gasifier Environ-
Fuels,” 1979 Annual Report, The Metal ment on Materials Performance, Mater.
Properties Council, New York, 1980 High Temp., Vol 11 (No. 1–4), 1993, p 81
57. A.O. Schaefer, “A Program to Discover 70. W.T. Bakker and J.A. Bonvallet, Corrosion
Materials Suitable for Service Under Hos- of Stainless Steels on the Wrong Side of
tile Conditions Obtaining in Equipment for the Kinetic Boundary, in Heat-Resistant
the Gasification of Coal and Other Solid Materials II, Conf. Proc. Second Inter-
Fuels,” 1980 Annual Report, The Metal national Conference on Heat-Resistant
Properties Council, New York, 1981 Materials, K. Natesan, P. Ganesan, G. Lai,
58. A.O. Schaefer, “A Program to Discover Ed., ASM International, 1995, p 121
Materials Suitable for Service Under 71. G.Y. Lai, in High Temperature Corrosion
Hostile Conditions Obtaining in Equip- in Energy Systems, M.F. Rothman, Ed.,
ment for the Gasification of Coal and Other The Metallurgical Society of AIME, 1985,
Solid Fuels,” 1981 Annual Report, The p 227
Metal Properties Council, New York, 1982 72. K. Natesan, Surface Modification for
59. A. Humphreys and A.O. Schaefer, “A Corrosion Resistance, Mater. High Temp.,
Program to Discover Materials Suitable for Vol 11 (No. 1–4), 1993, p 36
Service under Hostile Conditions Obtain- 73. Y.-R. He, D.L. Douglass, and F. Ges-
ing in Equipment for the Gasification of mundo, The Corrosion Behavior of
Coal and Other Solid Fuels,” 1982 Annual Ni-Mo Alloys in a H2/H2O/H2S Gas
Report, The Metal Properties Council, New Mixture, Oxid. Met., Vol 37 (No. 5/6),
York, 1983 1992, p 413
60. M.A.H. Howes, “High Temperature Cor- 74. V. Nagarajan, R.G. Miner, and A.V. Levy,
rosion in Coal Gasification Systems,” Final J. Electrochem. Soc., Vol 129 (No. 4),
Report GRI-8710152, Gas Research Insti- 1982, p 782
tute, Chicago, Aug 1987 75. G.Y. Lai, Sulfidation-Resistant Co-Cr-Ni
61. R.A. Perkins and S.J. Vonk, EPRI Report Alloy With Critical Contents of Silicon and
FP-1280, Electric Power Research Insti- Cobalt, U.S. Patent No. 4711763, Dec
tute, Palo Alto, CA, Dec 1979 1987
62. K. Natesan, in High Temperature Corro- 76. G.Y. Lai, Paper No. 209, Corrosion/89,
sion, R.A. Rapp, Ed., NACE, 1983, p 336 NACE, 1989
63. S.K. Verma, Corrosion of Commercial 77. G.Y. Lai, J. Met., Vol 41 (No. 7), 1989,
Alloys in A Laboratory-Simulated Me- p 21
dium-BTU Coal Gasification Environment, 78. J.F. Norton, F.G. Hodge, and G.Y. Lai,
Paper No. 336, Corrosion/85, NACE, 1985 A Study of the Corrosion Behavior of
64. R.W. Bradshaw, R.E. Stoltz, and D.R. Some Fe-Cr-Ni and Advanced Ni-Based
Adolphson, Report SAND 77-8277, Alloys Exposed to a Sulphidising/Oxidis-
Sandia Laboratories, Livermore, CA 1977 ing/Carburising Atmosphere at 700 °C, in
65. R.W. Bradshaw, A.N. Nagelberg, R.E. High Temperature Materials for Power
Stoltz, and D.R. Adolphson, Report SAND Engineering (Part I), Conf. Proc., E.
78-8260, Sandia Laboratories, Livermore, Bachelet et al., Ed., Kluwer Academic
CA, 1978 Publishers, Dordrecht, The Netherlands,
66. J.A. Kneeshaw, I.A. Menzies, and J.F. 1990, p 167
Norton, Werkst. Korros., Vol 38, 1987, 79. J.F. Norton, T.P. Levi, and W.T. Bakker,
p 473 High Temperature Corrosion of Candidate
Name ///sr-nova/Dclabs_wip/High Temp/5208_201-234.pdf/Chap_07/ 26/10/2007 11:15AM Plate # 0 pg 234

234 / High-Temperature Corrosion and Materials Applications

Heat Exchanger Alloys in a Dry-Feed 92. H. Lewis and J.E. Whittle, in Proc. Fourth
Entrained Slagging Gasifier Atmosphere, Int. Cong. Met. Corros., NACE, TX, 1972
in High Temperature Materials for Power 93. V. Vasantasree and M.G. Hocking, Corros.
Engineering (Part II), Conf. Proc., E. Sci., Vol 16, 1976, p 261
Bachelet et al., Ed., Kluwer Academic 94. M.G. Hocking and V. Vasantasree, Corros.
Publishers, Dordrecht, The Netherlands, Sci., Vol 16, 1976, p 279
1994, p 1617 95. K.N. Strafford, P.K. Datta, A.F. Hampton,
80. J.F. Norton and T.P. Levi, A Laboratory and P. Mistry, Corros. Sci., Vol 29 (No. 6),
Study of the Corrosion Behaviour of 1989, p 673
Alloys Exposed in a Non-Equilibrated 96. P.S. Sidky and M.G. Hocking, Corros. Sci.,
Coal-Gasification Atmosphere at 600 °C, Vol 27 (No. 2), 1987, p 183
Mater. Corros., Vol 46, 1995, p 286 97. M.G. Hocking and P.S. Sidky, Corros. Sci.,
81. M. Seiersten and P. Kofstad, Corros. Sci., Vol 27 (No. 2), 1987, p 205
Vol 22 (No. 5), 1982, p 487 98. P. Kofstad, High Temperature Corrosion,
82. M.R. Wootton and N. Birks, Corros. Sci., Elsevier Applied Science, 1988
Vol 12, p 829 99. J.J. Barnes and G.Y. Lai, Paper No. 90276,
83. B. Haflan and P. Kofstad, Corros. Sci., Corrosion/90, NACE, 1990
Vol 23 (No. 12), 1983, p 1333 100. D.H. Yates, P. Ganesan, and G.D. Smith,
84. A. Andersen, B. Haflan, P. Kofstad, and Recent Advances in the Enhancement of
P.K. Lillerud, Mater. Sci. Eng., Vol 87, Inconel Alloy 617 Properties to Meet the
1987, p 45 Needs of the Land Based Gas Turbine In-
85. J. Gilewicz-Wolter, Oxid. Met., Vol 11, dustry, in Advanced Materials and Coat-
1977, p 81 ings for Combustion Turbines Conference
86. A. Rahmel, Oxid. Met., Vol 9, 1975, p 491 Proceedings, V.P. Swaminathan and N.S.
87. A. Rahmel, Corros. Sci., Vol 13, 1975, Cheruvu, ASM International, 1994, p 89
p 125 101. A Guide to Corrosion Resistance, Climax
88. P. Singh and N. Birks, Oxid. Met., Vol 12, Molybdenum Company, Greenwich, CT,
1978, p 1 1981
89. P. Singh and N. Birks, Oxid. Met., Vol 12, 102. R.W. Borio, A.L. Plumley, and W.R. Syl-
1978, p 22 vester, in Ash Deposits and Corrosion Due
90. C.D. Asmundis, F. Gesmundo, and C. to Impurities in Combustion Gases, R.W.
Bottino, Oxid. Met., Vol 14 (No. 4), 1980, Bryers, Ed., Hemisphere Publishing, 1978,
p 351 p 163
91. P. Singh and N. Birks, Oxid. Met., Vol 13 103. R. Viswanathan and C.J. Spengler, Corro-
(No. 5), 1979, p 457 sion, Vol 26 (No. 1), 1970, p 29
Name ///sr-nova/Dclabs_wip/High Temp/5208_235-248.pdf/Chap_08/ 26/10/2007 12:33PM Plate # 0 pg 235

High-Temperature Corrosion And Materials Applications Copyright © 2007 ASM International®

George Y. Lai, editor, p235-248 All rights reserved.
DOI: 10.1361/hcma2007p235 www.asminternational.org

CHAPTER 8

Erosion and Erosion-Corrosion

8.1 Introduction and through the convection pass. The ash carried
by the flue gas stream is referred to as “fly-ash.”
Industrial plants are often involved in pro- With pulverized coal firing, about 70 to 90% of
cesses where gas streams are laden with particles. the ash in the coal is carried by the flue gas
Metallic components can suffer severe metal stream, while only about 40% of the ash is carried
wastage when subjected to constant impinge- by the flue gas in a stoker-fired furnace (Ref 3).
ment by this particle-laden gas stream under The cyclone-fired boiler generates about 15 to
high velocity and high particle loading. The 30% of the ash from the coal in the flue gas.
continual removal of the material from a com- Major constitutes of fly-ash are SiO2, Al2O3, and
ponent by this process is referred to as “erosion.” Fe2O3. (More information about ash constituents
When the component is exposed to elevated tem- in coal is available in Chapter 10 “Coal-Fired
peratures where oxidation and/or other modes of Boilers.”) This fly-ash-laden flue gas stream can
high-temperature corrosion are involved, it is pose fly-ash erosion problems to convection pass
frequently referred to as “erosion-corrosion.” tubes, such as superheater, reheater, boiler bank,
Materials problems that are caused by erosion or and economizer tubes. To reduce fly-ash erosion
erosion-corrosion in industries are numerous. problems for the convection pass tubes, boiler
In the petrochemical industry, the pyrolysis designers typically have set maximum flue gas
furnace tubes for the production of ethylene velocities in these areas. For example, Babcock
are a good example. Ethylene is produced by & Wilcox typically limits flue gas velocity to
cracking petroleum feedstocks, such as ethane 19.8 m/s (65 ft/s) or less for relatively non-
and naphtha, at temperatures up to 1150 °C abrasive low ash coal, and 13.7 m/s (45 ft/s) or
(2100 °F), thus making the process gas stream less for coals with high ash quantities and/or
inside the tube highly carburizing in nature. The abrasive ash (Ref 4). Combustion Engineering
furnace tubes suffer both carburization and (now Alstom Power) typically set the design
coking on the internal surface of the tube. In velocity in the range of 12 to 18 m/s (40 to 60 ft/
order to maintain the process efficiency, the coke s) (Ref 5). Lower velocities would be used for a
deposits have to be regularly removed from the boiler burning coals that yield heavy loading of
tube inner diameter (ID) surface by a process erosive ash, which is usually indicated by high
referred to as “decoking,” which involves in- silica content (Ref 5). For fluidized-bed coal-
jecting a mixture of steam and air into the furnace fired boilers, flue gas streams are laden with not
tube. Thus, during the decoking operation, the only fly-ash but also sands from the bed. The
return bends of the pyrolysis furnace tubes can convection pass tubes in these boilers are subject
suffer erosion or erosion-corrosion due to the to erosion attack.
coke particle-laden gas stream (Ref 1, 2), with Waste-to-energy (WTE) boilers burning muni-
average tube skin temperatures varying from 800 cipal and industrial waste can also experience fly-
to 1120 °C (1475 to 2050 °F) and gas velocities ash erosion problems for their convection pass
greater than 200 m/s (656 ft/s). tubes (e.g., superheater and economizer). Be-
In coal-fired boilers, it is well known that fly- cause of much more corrosive environments in
ash erosion can be a serious problem for boiler WTE boilers than coal-fired boilers, the boiler
tubes. When coal (which contains ash) is com- designers have used lower maximum design
busted in the lower furnace of the boiler, some of velocities in WTE boilers. Combustion Engi-
the ash drops out of the furnace from the bottom neering, for example, generally limits the flue gas
with remaining ash being carried by the com- velocities entering into the superheater or econ-
bustion flue gas stream to the top of the furnace omizer to 6 to 7.5 m/s (20 to 25 ft/s) (Ref 6).
Name ///sr-nova/Dclabs_wip/High Temp/5208_235-248.pdf/Chap_08/ 26/10/2007 12:33PM Plate # 0 pg 236

236 / High-Temperature Corrosion and Materials Applications

Babcock & Wilcox sets the maximum design research data generated can provide industry
velocity for superheaters and generating bank with some practical guidance for materials
at 9.1 m/s (30 ft/s) (Ref 7). In practice, lower selection based on either the mechanical prop-
velocities (3 to 4.6 m/s, or 10 to 15 ft/s) are used erties of the alloy (to select a more erosion-
(Ref 7). resistant alloy) or based on the resistance of the
Local flow disturbances can sometimes create alloy to oxidation or high-temperature corrosion
conditions that are more conducive to erosion (to select a more erosion-corrosion resistant
attack. This is illustrated in a case of the fly-ash alloy). The objective of this chapter is to answer
erosion in the backpass area of the economizer this question by reviewing the relevant erosion
when the flue gas stream makes a 90° directional and erosion-corrosion data that are mainly rela-
turn and creates nonuniform flow of gas stream ted to a particle-laden gas stream “jet” impacting
and uneven distribution of fly-ash caused by on the metal surface. The erosion or erosion-
centrifugal force (Ref 8). corrosion that is related to (a) in-bed components
Other systems where erosion can present a of bubbling fluidized-bed boilers and (b) water-
serious material issue include coal gasification, walls of circulating fluidized-bed boilers is not
combined cycles, and gas turbine. Gas turbines covered in this chapter. Discussion of erosion in
may involve extremely high gas velocities, up to fluidized-bed coal-fired boilers is presented in
250 m/s (820 ft/s), but with small particles Chapter 10 “Coal-Fired Boilers.”
(typically 5 µm) (Ref 9).
The erosion problems that have been
encountered in the aforementioned systems all 8.2 Erosion and Erosion-Corrosion
occur at elevated temperatures where the metal
surface not only experiences mechanical damage Finne (Ref 10) described the response of a
caused by erosion attack but also experiences material to erosion at room temperature in two
oxidation or other high-temperature corrosion distinctively different modes in terms of the
reactions at the same time. Although it is often particle incident angles (or particle impingement
referred to as “erosion” in industry, the damage angle). For a ductile material, such as aluminum,
reactions, in many cases, may involve a combi- the maximum erosion attack occurs at low inci-
nation of erosion and corrosion, and are more dent angles (less than 30°) with respect to the
appropriately referred to as “erosion-corrosion.” particle impingement direction and the minimum
Since industry continues to use a trial-and-error attack at about 90°, as shown in Fig. 8.1 (Ref 10).
approach to solve this practical erosion-related On the other hand, the maximum erosion attack
materials issue, the question is whether the on a brittle material, such as glass, occurs at close

Fig. 8.1 Effect of the particle incidence angle on the room-temperature erosion wear for a ductile material (aluminum) and a brittle
material (glass). Source: Ref 10
Name ///sr-nova/Dclabs_wip/High Temp/5208_235-248.pdf/Chap_08/ 26/10/2007 12:33PM Plate # 0 pg 237

Chapter 8: Erosion and Erosion-Corrosion / 237

to 90° (Ref 10). In both cases, no erosion occurs Carbon steel was also observed to behave
when the incident angle is 0° indicating the sur- similarly in air, showing increasing erosion (or
face of the object is in parallel with the particle erosion-corrosion) with advancing temperature,
impingement direction. as shown in Fig. 8.4 (Ref 12). Shida and Fuji-
The erosion model described by Finne (Ref kawa (Ref 13) examined the effect of the
10) is related to the room-temperature response impingement angle on the erosion rate for carbon
of materials. In industrial environments, erosion steel, 1.25Cr-1Mo-0.3V steel, and Type 304 at
problems are most often related to elevated ser- 300 °C (570 °F) in an argon environment (i.e.,
vice temperature. Levy (Ref 11) examined the elimination of oxidation participation). Their test
erosion response of Type 310 as a function of results are shown in Fig. 8.5 (Ref 13). The
temperature at incident angles of 30° and 90°. maximum erosion attack was found to occur at
The test environment consisted of nitrogen gas about a 30° impingement angle for all three
(N2), thus eliminating the effect of oxidation. The alloys with very little erosion attack at a 90°
data indicated that there was significantly less impingement angle. The test was carried out in an
erosion attack at a 90° impingement angle than inert environment—argon. Type 304 suffered
30° particularly at high temperatures, as shown in much more erosion attack than both carbon steel
Fig. 8.2 (Ref 11). The erosion behavior of Type and Cr-Mo-V steel in an inert environment. This
310 in N2 with no oxidation in the erosion suggests that austenitic stainless steels may not
reaction is quite similar to the room-temperature
erosion model of a ductile material proposed by
Finne. Figure 8.2 also shows that erosion attack
increased with advancing temperature at 30°
impingement angle. Similar temperature beha-
vior was also observed for Type 304 tested in N2
environment, as shown in Fig. 8.3 (Ref 11).
N2

Fig. 8.2 Erosion rate as a function of temperature in N2 for Fig. 8.3 Erosion rate as a function of temperature in N2 for
Type 310 steel for 30° and 90° impingement angles Type 304 steel for 30° impingement angle. Source:
(α). Source: Ref 11 Ref 11
Name ///sr-nova/Dclabs_wip/High Temp/5208_235-248.pdf/Chap_08/ 26/10/2007 12:33PM Plate # 0 pg 238

238 / High-Temperature Corrosion and Materials Applications

be as resistant to erosion attack as ferritic steels simulated fluidized-bed combustion atmosphere

when oxidation is not involved. This is further (N2-3%O2-15%CO2-0.026%SO2), which was
substantiated by the data that included carbon referred to as “FBC” environment in the paper.
steel, 2.25Cr-1Mo steel, and 12Cr steel com- For all three alloy systems, erosion rates were
pared with Type 304 and alloy 800 (Fig. 8.6). found to be significantly higher in the oxidizing
Alloy 800 suffered the worst erosion attack at all environment than in the inert environment.
temperatures from room temperature to 650 °C Figure 8.7 shows the behavior of Stellite No. 1
(1200 °F) when oxidation was not involved. (HS 1) and Stellite 6B (HS 6B) at 760 °C
Nagarajan and Wright (Ref 14) investigated (1400 °F) at both 30° and 90° impingement
iron-, nickel-, and cobalt-base alloys and the angles. Both alloys No. 1 and No. 6B are wear-
effect of an oxidizing environment on the erosion resistant alloys. With much more carbon and
resistance of the alloys. Their test environments tungsten, alloy No. 1 is generally more resistant
included argon (an inert atmosphere) and a to wear than alloy 6B. The data, however, show
that both alloys exhibited similar resistance to
erosion attack under the test condition in an inert
Test conditions: environment (Ar) at both 30° and 90° impinge-
11 1018 steel, erodent: 180 µm angular Al2O3 ment angles. When tested in an oxidizing envir-
v = 10 m/s, α = 30°,
10
t = 8 h, loading: 600 g, air
onment (FBC), both alloys also exhibited similar
erosion rates, although alloy 6B showed only
slightly higher erosion rates at the 90° impinge-
Thickness loss, µm

8
ment angle. However, both alloys showed higher
erosion rates in an oxidizing environment (FBC)
6 than in an inert environment (argon) for both
impingement angles. As for the effect of the
4
impingement angle, a 90° impingement angle
produced slightly higher erosion rates than did
30° in the oxidizing environment (FBC) while
2

25 100 200 300 400 500 600

Test temperature, °C

Fig. 8.4 Effect of temperature on erosion (or erosion-

corrosion) of carbon steel in air at 30° impingement 150
angle under the particle velocity of 10 m/s (32.8 ft/s) with 180 μm
alumina particles. Source: Ref 12 Alloy 800
Max thickness loss, µm/h

100
Max thickness loss, µm/h

C-steel
200
304
304

50 2.25Cr-1Mo
C-steel 12Cr-1Mo-V
100

1.25Cr-1Mo-0.3V
0 0
0 10 20 30 40 50 60 70 80 90 RT 300 500 650
Angle of impingement, degree Temperature, °C

Fig. 8.5 Effect of impingement angle on the erosion of Type Fig. 8.6 Effect of temperature on the erosion rates of various
304, carbon steel and Cr-Mo-V steel at 300 °C alloys in argon at the impingement angle of 20° with
(570 °F) in argon with 120 m/s (394 ft/s) particle velocity, 120 m/s (394 ft/s) particle velocity, 120 g/m3 particle concentra-
3
120 g/m particle concentration, and silica particles of 120 µm tion, and silica particles of 120 µm average particle size. Source:
average particle size. Source: Ref 13 Ref 13
Name ///sr-nova/Dclabs_wip/High Temp/5208_235-248.pdf/Chap_08/ 26/10/2007 12:33PM Plate # 0 pg 239

Chapter 8: Erosion and Erosion-Corrosion / 239

both 30° and 90° angles showed similar erosion environments (Fig. 8.8). For the nickel-base
rates in an inert environment (argon). superalloy IN-100, significantly higher erosion
In the same study by Nagarajan and Wright rates were observed in an oxidizing environment
(Ref 14), ferritic FeCrAlY (Fe-25Cr-4Al-1Y) (FBC) than in an inert environment (argon) at
and high-strength nickel-base superalloy IN-100 both impingement angles. The most interesting
(Ni-15Co-10Cr-5.5Al-4.7Ti-3Mo) were tested. finding from the study by Nagarajan and Wright
For the ferritic FeCrAlY alloy, no significant (Ref 14) (which included high particle velocity
difference in erosion rates was observed between of 42.7 m/s, or, 140 ft/s test conditions) was
argon and the FBC oxidizing environment. The that the low-strength ferritic FeCrAlY alloy was
alloy also exhibited similar erosion rates at significantly more resistant to erosion than the
both 30° and 90° impingement angles in both high-strength IN-100 superalloy in an oxidizing
environment (Fig. 8.8). FeCrAlY alloy was also
found to be much more erosion resistant than
wear-resistant cobalt-base alloys No. 1 and 6B in
an oxidizing environment.
There was no direct comparison of tensile
strength data between FeCrAlY and IN-100.
However, FeCrAlY (Fe-25Cr-4Al-1Y) is be-
lieved to exhibit tensile strengths similar to Type
446 (Fe-25Cr). For example, the ultimate tensile
HS 1, argon strength of Kanthal D (Fe-22Cr-4.8Al) at 900 °C
HS 6B, argon
(1650 °F) is 34.5 MPa (5 ksi) (Ref 15), while
that of Type 446 is 29 MPa (4.2 ksi) (Ref 16). At
700 °C (1300 °F), Type 446 exhibits ultimate
HS 1, FBC tensile strength of about 9.0 MPa (1.3 ksi)
(Ref 16) as opposed to 1270 MPa (184 ksi) for
IN-100 (Ref 17). Thus, IN-100 would be ex-
pected to exhibit significantly higher tensile
(a) strength than FeCrAlY at 700 °C (1300 °F) and
would likely be the same at the test temperature
of 760 °C (1400 °F) in the erosion study con-
ducted by Nagarajan and Wright. It is thus sur-
prising to see that IN-100 alloy with its tensile
HS 6B, argon strength approximately more than 30 times
HS 1, argon higher than that of FeCrAlY alloy suffered sig-
nificantly higher erosion rates than FeCrAlY
under a particle velocity of about 42.7 m/s
(140 ft/s) in an oxidizing environment. It is also
surprising to see that the ferritic FeCrAlY alloy
was much more resistant to erosion than Stellite
alloys No. 1 and No. 6B, which are considered to
be excellent wear-resistant alloys. Both alloys
No. 1 and No. 6B also exhibit much higher ten-
sile strengths compared with FeCrAlY. For
example, at 675 °C (1250 °F), the ultimate ten-
sile strength of Stellite 6B is about 793 MPa
(b)
(115 ksi) (Ref 18). Stellite No. 1 and No. 6B also
show much higher hardness than ferritic Type
Fig. 8.7 Effect of the test environments (argon and fluidized- 446, as shown in Fig. 8.9 (Ref 19).
bed combustion) on the erosion behavior of HS 1 Pettit and Birks along with their research
(Co-30Cr-12W-2.5C) and HS 6B (Co-30Cr-4W-1C) at 760 °C
(1400 °F), 42.7 m/s (140 ft/s), with 12 µm alumina (Al2O3) as
group have investigated erosion behavior of
erodent (a) 30° impingement angle and (b) 90° impingement nickel and cobalt (Ref 20, 21), and Cr2O3- and
angle. The argon test environment contained 1% H2, where H2 Al2O3-forming alloys (Ref 22, 23) at elevated
was used to remove O2, while fluidized-bed combustion (FBC)
environment was a simulated test gas consisting of N2, 3% O2, temperatures under very high particle velocities
15% CO2, and 0.026% SO2. Source: Ref 14 with 20 µm alumina particles. Kang et al.
Name ///sr-nova/Dclabs_wip/High Temp/5208_235-248.pdf/Chap_08/ 26/10/2007 12:33PM Plate # 0 pg 240

240 / High-Temperature Corrosion and Materials Applications

(Ref 20) indicated that air caused much higher 140 m/s (295 and 459 ft/s). They observed that
erosion rates than N2 for nickel under their test at 800 °C (1470 °F) there was essentially no
conditions involving particle velocities of 90 and erosion in N2 at both 90 and 140 m/s (295 and
459 ft/s) (Ref 20). Erosion rates for nickel were
1 significantly increased when the test environ-
ment was switched from N2 (an inert environ-
0 FeCrAlY, FBC ment) to air (oxidizing environment). This is
illustrated in Fig. 8.10 (Ref 20). Similar results
FeCrAlY (2541), argon
–1 were also observed for cobalt (Ref 20). The effect
of the impingement angle on erosion under the
–2
IN-100, argon same test condition for both nickel and cobalt
Specimen weight change, mg/cm2

was reported Chang et al. (Ref 21). The data for

–3
nickel are summarized in Fig. 8.11 and for cobalt
–4 in Fig. 8.12 (Ref 21). Erosion tests under the
same conditions also included Cr2O3- and
–5 Al2O3-forming alloys (Ref 22, 23). Figure 8.13
shows the effect of the impingement angle on
–6 erosion rates of MA754 (Ni-20Cr-0.6Y2O3, an
oxide-dispersion-strengthened alloy) at 780 °C
–7
(1435 °F) under 140 m/s (459 ft/s) particle
IN-100, FBC velocity (Ref 22). Erosion test results for nickel,
–8
cobalt, MA754, Ni30Cr, CoCrAlY (Co-22Cr-
–9
11Al-0.17Y), and Ni20Al (Ni-20Al) are sum-
marized in Fig. 8.14 (Ref 22). Both nickel and
–10 cobalt were found to suffer high erosion rates due
0 5 10 15 20 25 30 35 40
to formation of nickel and cobalt oxides,
Erodent impacted, g/cm2 respectively. Both MA754 and Ni30Cr alloys
(a)
that formed Cr2O3 scales showed better resis-
2 tance to erosion attack, and Al2O3-forming

0
IN-100, argon
Specimen weight change, mg/cm2

–2

–4
FeCrAlY, argon
FeCrAlY, FBC
–6

–8

–10

–12

–14 IN-100, FBC

–16
0 10 20 30 40 50 60 70 80
Erodent impacted, g/cm2
(b)

Fig. 8.8 Effect of the test environments (argon and fluidized-

bed combustion) on the erosion behavior of FeCrAlY
(Fe-25Cr-4Al-1Y) and IN-100 (Ni-15Co-10Cr-5.5Al-4.7Ti-3Mo)
at 760 °C (1400 °F), 42.7 m/s (140 ft/s), with 12 µm alumina
(Al2O3) as erodent (a) 30° impingement angle and (b) 90° impin-
gement angle. The argon test environment contained 1% H2, Temperature, °C (°F)
where H2 was used to remove residual O2, while fluidized-bed
combustion (FBC) environment was a simulated test gas consist- Fig. 8.9 Hot hardness data for various alloys as a function of
ing of N2, 3% O2, 15% CO2, and 0.026% SO2. Source: Ref 14 temperature. Source: Ref 19
Name ///sr-nova/Dclabs_wip/High Temp/5208_235-248.pdf/Chap_08/ 26/10/2007 12:33PM Plate # 0 pg 241

Chapter 8: Erosion and Erosion-Corrosion / 241

alloys (CoCrAlY and Ni20Al) were much more (Ref 24) conducted erosion tests in a combustion
resistant to erosion attack presumably due to gas stream using a propane-fired abrasive jet rig
formation of aluminum oxide scales. Ives at 975 °C (1790 °F) with 55 m/s (123 mph)
particle velocity of about 135 µm (5 mil) SiC

2.5

780 °C, 140 m/s

Erosion rate, ×10–5 g/cm2 s

1.5
600 °C, 140 m/s

0.5

780 °C, 70 m/s

0
0 30 60 90
Impact angle, degrees
Fig. 8.10 Effect of the test environments (N2 and air) on the
erosion behavior of nickel with particle velocities
of 90 and 140 m/s (295 and 459 ft/s) at 90° impingement angle. Fig. 8.12 Effect of the impingement angle on erosion rate of
cobalt in air. Source: Ref 21
Source: Ref 20

Fig. 8.13 Effect of the impingement angle on erosion rate of

Fig. 8.11 Effect of the impingement angle on erosion rate of MA754 (Ni-20Cr-0.6Y2O3, an oxide-dispersion-
nickel in air at 780 °C (1435 °F) and 140 m/s strengthened alloy) at 780 °C (1435 °F) in air under 140 m/s
(459 ft/s) particle velocity. Source: Ref 21 (459 ft/s) particle velocity. Source: Ref 22
Name ///sr-nova/Dclabs_wip/High Temp/5208_235-248.pdf/Chap_08/ 26/10/2007 12:33PM Plate # 0 pg 242

242 / High-Temperature Corrosion and Materials Applications

particles. The erosion test results in terms of the erosion tests conducted by Tabakoff et al., coal-
specimen thickness loss (µm/h) are summarized ash particles consisting of primarily SiO2, Al2O3,
in Table 8.1. The data indicate that nickel-base and Fe2O3 with a mean particle diameter of
alloys were better than Fe-Ni-Cr alloys, which about 15 µm were used. MAR-M246 (Ni-10Co-
were better than Fe-Cr alloys. 9Cr-10W-2.4Mo-1.5Ta-1.5Ti-5.5Al) and X40
Tabakoff et al. (Ref 25) found that aluminized (Co-25Cr-10Ni-7.5W) were tested.
coatings were capable of providing a significant Aluminized coatings tested were “C” coating
reduction in erosion rates under very high parti- on X40 and “N” coating on M246. Also included
cle velocities and high temperatures in the con- were platinum-modified aluminized coating,
ditions that were similar to gas turbines. This is RT22, and rhodium/platinum-modified alumi-
illustrated in Fig. 8.15 and 8.16 (Ref 25). In the nide coating, RT22B. Both RT22 and RT22B

(a)

Fig. 8.14 Erosion behavior of nickel, cobalt, Cr2O3-forming

alloys (MA754 and Ni30Cr alloys), and Al2O3-
forming alloys (CoCrAlY and Ni20Al alloys) in air at 600 and (b)
780 °C (1110 and 1435 °F) under 140 m/s (459 ft/s) particle
velocity and 30° impingement angle. Source: Ref 22

Table 8.1 Metal loss rates for alloys tested in

a combustion atmosphere using a propane-fired
abrasive jet rig at 975 °C (1790 °F), 55 m/s
(123 mph) particle velocity of about 135 µm
(5 mils) SiC abrasive particles
Metal loss rate, µm/h
Alloy 45° impinging angle 90° impinging angle
671 31 19
601 62 29
800 80 70
310 80 73 (c)
446 130 90
316 160 150
304 190 195 Fig. 8.15 Erosion behavior of M246 (a), X40 (b), and RT22
coating (c) under hot, oxidizing combustion gas
Source: Ref 24 stream at 815 °C (1500 °F) and 366 m/s (1200 ft/s) with fly-ash as
erodent. Source: Ref 25
Name ///sr-nova/Dclabs_wip/High Temp/5208_235-248.pdf/Chap_08/ 26/10/2007 12:33PM Plate # 0 pg 243

Chapter 8: Erosion and Erosion-Corrosion / 243

were deposited on M246. The thickness was plastic deformation developed on the metal
0.076 mm (3 mils) for aluminized coating “C,” surface. This was caused by the impinging par-
“N,” and RT22B, and 0.127 mm (5 mils) for ticles for nickel and cobalt tested in air at 800 °C
RT22. Test results showed that aluminized (1470 °F). In addition, thin, discontinuous oxide
coatings significantly reduced the erosion rates of scales were observed to form on the eroded sur-
the alloy with both platinum- and rhodium/pla- face (Ref 20, 21). The impinging particles not
tinum-modified aluminide coatings providing only removed the oxide scales, but also caused
the most resistance. significant plastic deformation on the metal sur-
The erosion and erosion-corrosion (oxidation) face, developing a rippled surface with mounds
data presented so far involve very high particle and valleys (Ref 20 to 23). However, the repeated
velocities. Erosion or erosion-corrosion rates of removal and reformation of oxide scales was
materials can be a strong function of particle attributed to the accelerated erosion-corrosion
velocities. Many of the industrial applications rates in oxidizing environments (Ref 20 to 23).
involve particle velocities that are much lower Since significant plastic deformation was
than the data presented earlier. Wright et al. (Ref observed under high particle velocities, the
26) investigated the erosion-corrosion behavior mechanical properties, hardness, and the micro-
of 10 commercial alloys as a function of particle structure of the alloy may play an important
velocities (27 to 52 m/s, or 88 to 170 ft/s) at role in affecting the erosion-dominated erosion-
760 °C (1400 °F) in an oxidizing environment corrosion behavior of the alloy. Various mech-
(N2-15CO2-3O2-0.03SO2), which was referred anisms for erosion have been proposed in the
to as “FBC” gas. Alumina particles (15 µm size) literature, such as, cutting by Finne (Ref 10),
with a loading in the gas stream of about 15,000 cutting and ploughing by Hutchings and Winter
ppm (by wt) were used in the tests with each test (Ref 27), flake formation by Brown et al.(Ref 28),
being about 4 to 6 h. Test results showed a sharp platelet formation by Bellman and Levy (Ref 29),
transition from low metal-loss rates to a regime and deformation wear by Bitter (Ref 30).
of rapid metal-loss rates at particle velocities in Hardness has often been used in industry as a
the range of about 27 to 34 m/s (90 to 110 ft/s) key material property for making materials
(Ref 26). The authors termed this regime of high selection for resisting wear. Hardness increases
metal-loss rates as erosion dominated. The test can arise from different hardening mechanisms,
results on some of the alloys are summarized in such as cold working, formation of fine-coherent
Fig. 8.17 (Ref 26). precipitates, martensitic transformation, and
Under the particle velocities of 90 and 140 m/s second-phase hard particles (e.g., eutectic car-
(295 and 459 ft/s), Kang et al. (Ref 20) and bides, tungsten carbides, etc.). In industrial trial-
Chang et al. (Ref 21) observed that significant and-error tests, hardening that resulted from cold
working, coherent precipitation, and martensitic
transformation is somewhat beneficial in resist-
4 ing sliding wear but not erosion or erosion-
X40
corrosion. Hardness increases by second-phase,
M246
hard particles appear to be beneficial in increas-
3
ing the erosion resistance of the alloy. Under
erosion-corrosion conditions at high particle
Erosion rate, mgm/gm

velocities and elevated temperatures, alloys

that exhibit good oxidation resistance (or high-
2 temperature corrosion resistance) and are har-
dened by second-phase hard particles are likely
C to be promising candidates for the application.
Increasing hardness can adversely affect alloy
1 toughness. Thus, application of a suitable coating
N or weld overlay hardfacing material appears to be
RT22B a more viable practical approach to the problem.
RT22
0 At lower particle velocities, the erosion-
corrosion behavior can be strongly dependent on
Fig. 8.16 Erosion behavior of alloys and coatings under the type of oxide scales formed on the alloy.
hot, oxidizing combustion gas stream at 815 °C
(1500 °F) and 366 m/s (1200 ft/s) with fly-ash as erodent. Source: Steels and low-alloy steels that form iron oxides
Ref 25 can suffer much more erosion-corrosion attack
Name ///sr-nova/Dclabs_wip/High Temp/5208_235-248.pdf/Chap_08/ 26/10/2007 12:33PM Plate # 0 pg 244

244 / High-Temperature Corrosion and Materials Applications

than stainless steels with adequate chromium for in the coal gasification environment was also
forming chromium oxide scales. This is illu- found to be less resistant to erosion-corrosion
strated by the test results generated by Levy and attack in the same environment. This is illu-
Man (Ref 31) in Fe-Cr alloys with various strated in Fig. 8.20 (Ref 32), showing Type 310
chromium contents at 850 °C (1560 °F) in air suffering significantly more erosion-corrosion
under the particle velocity of 35 m/s (115 ft/s) attack than alloy 6B in a MPC coal gasification
with alumina particles (130 µm particle size). environment (24H2-18CO-12CO2-39H2O-5CH4-
The results are summarized in Fig. 8.18 (Ref 31). 1NH3-1H2S) tested at 815 °C (1500 °F) with a
Levy and Man (Ref 31) also performed static particle velocity of 15.3 m/s (50 ft/s) and using
oxidation tests on the same alloys in air at the metallurgical coke as the erodent (300 to 600 µm
same temperature (850 °C, or 1560 °F). Their particle size). [MPC, Materials Property Council
oxidation tests showed decreasing weight gain coal-gasification test program (See Chapter 7
with increasing chromium content in the alloy, as “Sulfidation”).] Hard particles, such as alumina,
shown in Fig. 8.19. This behavior is quite similar can cause more erosion-corrosion attack than soft
to that generated under erosion-corrosion con- particles, such as coke. This is illustrated in
ditions (Fig. 8.18). Fig. 8.21 (Ref 33). In general, the alloys that
Erosion-corrosion behavior of alloys in a are more resistant to sulfidation are better at
simulated coal gasification environment was resisting erosion-corrosion in sulfidizing en-
studied by Agarwal and Howes (Ref 32). The vironments. Figure 8.22 shows the results of
alloy that was less resistant to sulfidation attack erosion-corrosion tests at 980 °C (1800 °F) in

(a) (b)

(c) (d)

Fig. 8.17 Effect of particle velocity on the erosion-corrosion rate for (a) Type 446, (b) alloy 671, (c) alloy 188, and (d) alloy 6B tested at
760 °C (1400 °F) in a simulated combustion gas stream (designated as FBC gas, N2-15CO2-3O2-0.03SO2) containing 15 µm
alumina particles with a particle loading of 15,000 ppm (by wt). 1 ft = 0.305 m. Source: Ref 26
Name ///sr-nova/Dclabs_wip/High Temp/5208_235-248.pdf/Chap_08/ 26/10/2007 12:33PM Plate # 0 pg 245

Chapter 8: Erosion and Erosion-Corrosion / 245

a coal-gasification environment eroded with resistant to sulfidation attack. (For more infor-
metallurgical coke at 30.5 m/s (100 ft/s) particle mation about the sulfidation behavior of alloys in
velocity as a function of exposure time up to coal-gasification environments, readers are
200 h (Ref 33). Some less-resistant alloys were referred to Chapter 7.)
found to suffer “breakaway” erosion-corrosion. In the regime where particle velocities are
This “breakaway” erosion-corrosion is quite lower, the erosion-corrosion behavior is likely to
similar to “breakaway” corrosion in reducing be dominated by corrosion. The alloys that are
sulfidizing environments. After 200 h, cobalt- more resistant to corrosion (i.e., high-tempera-
base alloy No. 1 (Co-30Cr-12W-2.5C) was found ture corrosion) are generally better at resisting
to be most resistant among alloys tested, showing erosion-corrosion in the same environment. This
no breakaway erosion-corrosion. Alloy 671 may provide a practical guide to materials
(Ni-48Cr-0.5Ti) also showed no breakaway selection for applications that are under erosion-
erosion-corrosion. Both of these were high- corrosion conditions. In this regime, it is believed
chromium alloys, which are known to be highly that impinging particles cause damage only
to oxides or corrosion products without sig-
nificantly affecting the underlying metal. The
120
most likely scenario in this regime involves the
Corrosion-erosion
110
impinging particles removing oxide scales or
corrosion products and allowing the fresh metal
to be exposed to the corrosive environment.
Oxide scales or corrosion products form and are
then removed by subsequent impinging particles.
This process involves repeated removal of the
corrosion products by eroding particles and
reformation of the fresh corrosion products, thus
resulting in erosion-induced accelerated corro-
sion. This is best described graphically by a
schematic showing a corrosion mode where

60
Alloy 310
Fig. 8.18 Effect of chromium in Fe-Cr alloys on the erosion-
corrosion resistance of the alloys at 850 °C
(1560 °F) in air with 35 m/s (115 ft/s) particle velocity (130 µm 50
alumina particles). Source: Ref 31
Maximum thickness loss, mils

40

30

20

Alloy 6B
10

0
20 40 60 80 100
Impingement angle, degrees

Fig. 8.20 Erosion-corrosion behavior of Type 310 and alloy

6B in MPC coal-gasification environment (24H2-
Fig. 8.19 Effect of chromium in Fe-Cr alloys on the oxidation 18CO-12CO2-39H2O-5CH4-1NH3-1H2S) tested for 100 h at
resistance of the alloys at 850 °C (1560 °F) in air. 250 psig, 1500 °F (815 °C), and 50 ft/s, with metallurgical coke
Source: Ref 31 as erodent (300 to 600 µm). Source: Ref 32
Name ///sr-nova/Dclabs_wip/High Temp/5208_235-248.pdf/Chap_08/ 26/10/2007 12:33PM Plate # 0 pg 246

246 / High-Temperature Corrosion and Materials Applications

Fig. 8.21 Erosion-corrosion behavior of various alloys tested at 980 °C (1800 °F) for 50 h in MPC coal-gasification environment at
1000 psig with 30.5 m/s (100 ft/s) at 45° impingement angle, using coke and alumina as erodents. S-1: Stellite No. 1: Co-
30Cr-12W-2.5C; Cru 25: Fe-25Cr-25Ni; LM 1866: Fe-18Cr-6Al-0.6Hf; Alloy 671: Ni-48Cr-0.5Ti; RA333: Ni-25Cr-18Fe-3Mo-3W; Alloy
188: Co-22Cr-22Ni-14W-0.04La; 800AL (aluminized alloy 800); and 310AL (aluminized 310). Source: Ref 33

the corrosion scale reaches a steady state and an temperatures, oxidizing environments (e.g., air)
erosion-corrosion mode where the corrosion significantly accelerate the erosion-corrosion
scale is continuously reformed after it is repeat- rates compared with an inert environment, such
edly removed by eroding particles, as shown in as N2 or argon. Under such high particle velo-
Fig. 8.23 (Ref 34). cities, thin, disconnected oxide scales (instead
of a continuous oxide scale) were observed to
form in addition to severe plastic deformation
on the underlying metal that caused the forma-
8.3 Summary tion of a rippled surface with mounds and val-
leys. Some limited data suggest alloys that form
Erosion and erosion-corrosion behavior of Cr2O3 or Al2O3 showing less scaling are better
alloys are reviewed. The data examined are gen- than those having high scaling rates for resisting
erated mainly in laboratory tests using a “jet”- erosion-corrosion attack. The data also suggest
type apparatus where the metal is impacted by a that aluminized coatings are capable of reducing
particle-laden gas stream. Under conditions erosion-corrosion rates in oxidizing environ-
involving very high particle velocities, such ments. Hardness has often been used as an
as 100 m/s (328 ft/s) or higher, at elevated important material property in resisting wear.
Name ///sr-nova/Dclabs_wip/High Temp/5208_235-248.pdf/Chap_08/ 26/10/2007 12:33PM Plate # 0 pg 247

Chapter 8: Erosion and Erosion-Corrosion / 247

Fig. 8.23 A corrosion mode where the corrosion rate is

Fig. 8.22 Erosion-corrosion behavior of various alloys tested diminishing in time and an erosion-corrosion (E-C)
at 980 °C (1800 °F) in MPC coal-gasification mode where the corrosion products are repeatedly removed by
environment, 500 psig, 100 ft/s, coke as erodent. S-1: Stellite No. impinging particles and reformed subsequently. Source: Ref 34
1: Co-30Cr-12W-2.5C; Cru 25: Fe-25Cr-25Ni; LM 1866: Fe-18Cr-
6Al-0.6Hf; Alloy 671: Ni-48Cr-0.5Ti; RA333: Ni-25Cr-18Fe-
3Mo-3W; Type 310: Fe-25Cr-20Ni. Source: Ref 33

REFERENCES

Limited laboratory data generated so far have 1. G.E. Moller and C.W. Warren, Survey of
failed to provide practical guidance on the Tube Experience in Ethylene and Olefins
hardness of the metal in resisting erosion- Pyrolysis Furnaces: T-5B-6 Task Group
corrosion attack at elevated temperatures. In Report, Corrosion/81, NACE
industrial trial-and-error tests, hardening of the 2. D. Jakobi and R. Gommans, Typical Failures
alloy by second-phase, hard particles appears to in Pyrolysis Coils for Ethylene Cracking,
be beneficial in improving resistance to erosion- Mater. Corros., Vol 54 (No. 11), 2003, p 881
corrosion attack at elevated temperatures. 3. S.C. Stultz and J.B. Kitto, Ed., Steam: Its
Erosion-corrosion studies conducted by Generation and Use, Babcock & Wilcox,
Wright et al. (Ref 26) suggest that an erosion- 1992, p 33-1
dominated erosion-corrosion regime occurs 4. S.C. Stultz and J.B. Kitto, Ed., Steam: Its
when the particle velocities are in excess of the Generation and Use, Babcock & Wilcox,
range of about 27 to 34 m/s (90 to 110 ft/s), and 1992, p 20-16
below that particle velocity range is a corrosion- 5. J.G. Singer, Ed., Combustion Fossil Power,
dominated erosion-corrosion regime. At lower Combustion Engineering, Inc. (now Alstom
particle velocities, that is, in a corrosion- Power), 1991, p 7–8
dominated erosion-corrosion regime, in both 6. J.G. Singer, Ed., Combustion Fossil Power,
oxidizing and sulfidizing environments, the data Combustion Engineering, Inc. (now Alstom
suggest that the alloys that provide better oxi- Power), 1991, p 8–14
dation or sulfidation resistance are likely to pro- 7. S.C. Stultz and J.B. Kitto, Ed., Steam: Its
vide better erosion-corrosion resistance in the Generation and Use, Babcock & Wilcox,
same environment. This may provide a general 1992, p 27-1
practical guide to materials selection for appli- 8. J.G. Singer, Ed., Combustion Fossil Power,
cation in the regime where the corrosion dom- Combustion Engineering, Inc. (now Alstom
inates the erosion-corrosion reactions. Power), 1991, p 23–21
Name ///sr-nova/Dclabs_wip/High Temp/5208_235-248.pdf/Chap_08/ 26/10/2007 12:34PM Plate # 0 pg 248

248 / High-Temperature Corrosion and Materials Applications

9. M.M. Stack, F.H. Stott, and G.C. Wood, 22. S.L. Chang, F.S. Pettit, and N. Birks, Some
Mater. High Temp., Vol 9 (No. 3), 1991, p 153 Interactions in the Erosion-Oxidation of
10. I. Finne, Wear, 1960, p 87 Alloys, Oxid. Met., Vol 34 (No. 1 & 2), 1990,
11. A.V. Levy, Chapter 5: Erosion and Erosion- p 71
Corrosion of Steels at Elevated 23. D.M. Rishel, F.S. Pettit, and N. Birks, Some
Temperatures, Solid Particle Erosion and Principle Mechanisms in the Simultaneous
Erosion-Corrosion of Materials, ASM Erosion and Corrosion Attack of Metals at
International, 1995 High Temperatures, Mater. Sci. Eng., Vol
12. A.V. Levy, Erosion-Corrosion of Tubing A143, 1991, p 197
Steels in Combustion Boiler Environments, 24. L.K. Ives, Erosion of 310 Stainless Steel
Paper No. 236, Corrosion 93, NACE, 1993 at 975 °C in Combustion Gas Atmospheres,
13. Y. Shida and H. Fujikawa, Particle Erosion Trans. ASME, April 1977, p 126
Behaviour of Boiler Tube Materials at 25. W. Tabakoff, A. Hamed, M. Metwally, and
Elevated Temperature, Wear, Vol 103, 1985, M. Pasin, High-Temperature Erosion
p 281 Resistance of Coatings for Gas Turbine,
14. V. Nagarajan and I.G. Wright, Influence of Trans. ASME, Vol 114, 1992, p 242
Oxide Scales on High Temperature Corro- 26. I.G. Wright, V. Nagarajan, W.E. Merz, and
sion Erosion Behavior of Alloys, High J. Stringer, The Kinetics of High-
Temperature Corrosion, R.A. Rapp, Ed., Temperature Erosion-Corrosion of Oxida-
Conf. Proc. (San Diego, CA), March 2–6, tion-Resistant Alloys, Corrosion/81, NACE,
1981, NACE, 1981, p 398 1981
15. Kanthal Handbook: Resistance Heating 27. I.M. Hutchings and R.E. Winter, Wear,
Alloys and Elements for Industrial Fur- Vol 27, 1974, p 121
naces, Kanthal Heating Systems, Hall- 28. R. Brown, E.J. Jun, and J.W. Edington,
stahammar, Sweden, p 7 Wear, Vol 70, 1981, p 347
16. AISI Type 446, Alloy Digest, Engineering 29. R. Bellman and A.V. Levy, Wear, Vol 70,
Alloys Digest, Inc., May 1982 1981, p 1
17. M.J. Donachie and S.J. Donachie, Super- 30. J.G.A. Bitter, Wear, Vol 6, 1963, p 5
alloys: A Technical Guide, 2nd ed., ASM 31. A.V. Levy and Y.F. Man, Erosion-Corrosion
International, 2002, p 249 Mechanisms and Rates in Fe-Cr Steels,
18. Haynes Stellite Alloy No. 6B, Alloy Digest, Wear, Vol 131 (No. 1), 1989, p 39
Engineering Alloys Digest, Inc., Sept 1960 32. S.C. Agarwal and M.A.H. Howes,
19. I.G. Wright, V. Nagarajan, and R.B. Herch- Erosion-Corrosion Materials in High-
enroeder, Some Factors Affecting Solid Temperature Environments: Impingement
Particle Erosion-Corrosion of Metals and Angle Effects in Alloys 310 and 6B under
Alloys, Corrosion-Erosion Behavior of Simulated Coal Gasification Atmosphere,
Materials, K. Natesan, Ed., Conf. Proc., Fall J. Mater. Energy Syst., Vol 7 (No. 4), 1986,
Meeting of The Metallurgical Society of p 370
AIME (St. Louis, MO), October 17–18, 1978 33. M.A.H. Howes, Elevated Temperature
20. C.T. Kang, F.S. Pettit, and N. Birks, Erosion-Corrosion of Alloys in Sulfidizing
Mechanisms in the Simultaneous Erosion- Gas/Solid Streams: Mechanistic Studies,
Oxidation Attack of Nickel and Cobalt Proc. Conf., Corrosion-Erosion-Wear of
at High Temperatures, Metall. Trans. A, Materials at Elevated Temperatures, NACE,
Vol 18, 1987, p 1785 1986, p 230
21. S.L. Chang, F.S. Pettit, and N. Birks, Effect 34. V.K. Sethi and I.G. Wright, Observations
of Angle of Incidence on the Combined on the Erosion-Oxidation Behavior of
Erosion-Oxidation Attack of Nickel and Alloys, in Proc. TMS Conf. on Corrosion
Cobalt, Oxid. Met., Vol 34 (No. 1 & 2), 1990, and Particle Erosion, V. Srinivasan and
p 47 K. Vedula, Ed., 1986, p 245
Name ///sr-nova/Dclabs_wip/High Temp/5208_249-258.pdf/Chap_09/ 26/10/2007 2:02PM Plate # 0 pg 249

High-Temperature Corrosion And Materials Applications Copyright © 2007 ASM International®

George Y. Lai, editor, p249-258 All rights reserved.
DOI: 10.1361/hcma2007p249 www.asminternational.org

CHAPTER 9

Hot Corrosion in Gas Turbines

9.1 Introduction although dependent on alloy composition, is

generally 800 to 950 °C (1470 to 1740 °F). The
During combustion in the gas turbine, sulfur lower threshold temperature is believed to be
from the fuel reacts with sodium chloride from the melting temperature of the salt deposit, and
ingested air at elevated temperatures to form the upper temperature is the salt dew point
sodium sulfate. The sodium sulfate then deposits (Ref 3). This type of corrosion process is some-
on the hot-section components, such as nozzle times referred to as Type I hot corrosion to dif-
guide vanes and rotor blades, resulting in accel- ferentiate it from Type II hot corrosion, which
erated oxidation (or sulfidation) attack. This occurs at lower temperatures (typically 670 to
is commonly referred to as “hot corrosion.” 750 °C, or 1240 to 1380 °F) (Ref 4). Type II
Sulfur in the fuel is generally limited to 0.3% hot corrosion is characterized by pitting attack
for commercial jet engines and to 1.0% for with little or no internal attack underneath the
marine gas turbines (Ref 1). Sodium chloride pit (Ref 4). Type II hot corrosion is rarely
comes from seawater (see Table 9.1) (Ref 2). observed in aeroengines because the blades
Seawater is also a source of sulfur. For aircraft are generally operated at higher temperatures
engines, Tschinkel (Ref 1) suggested that runway (Ref 5). However, marine and industrial gas
dust may be a source of salts. turbines, which operate at lower temperatures,
Gas turbines generally use large amounts can experience low-temperature Type II hot
of excess air for combustion (a large fraction of corrosion.
air, in fact, is also used to cool the combustor), Type I hot corrosion generally proceeds in two
with a typical air-to-fuel ratio from about 40 to 1 stages: an incubation period exhibiting a low
(during takeoff) to 100 to 1 (at cruising speed) corrosion rate, followed by accelerated corrosion
for aircraft gas turbine engines (Ref 2). These attack. The incubation period is related to the
air-to-fuel ratios correspond to about 0.12 to 0.18 formation of a protective oxide scale. Initiation
mole fractions of oxygen in the combustion zone of accelerated corrosion attack is believed to be
(Ref 2). Thus, the combustion gas atmosphere related to the breakdown of the protective oxide
is highly oxidizing. The sulfur partial pressure scale. Many mechanisms have been proposed
in the atmosphere can be extremely low, vary- to explain accelerated corrosion attack; the salt
ing from 10−40 to 10−26 atm over the range from fluxing model is probably the most widely
330 to 1230 °C (620 to 2240 °F) (Ref 2). These accepted. Oxides can dissolve in Na2SO4 as
sulfur partial pressures are well below those anionic species (basic fluxing) or cationic species
necessary to form chromium sulfides, which (acid fluxing), depending on the salt composition
are frequently observed in alloys suffering hot (Ref 6). Salt is acidic when it is high in SO3,
corrosion attack. and basic when low in SO3. The hot corrosion
High-temperature alloys that suffered hot
corrosion attack were generally found to exhibit Table 9.1 Composition of seawater
both oxidation and sulfidation. The hot corrosion Element Composition, ppm

morphology is typically characterized by a thick, Chlorine 18,980

Sodium 10,561
porous layer of oxides with the underlying Magnesium 1,272
alloy matrix depleted in chromium, followed Sulfur 884
by internal chromium-rich sulfides. It is gener- Calcium 400
Potassium 380
ally believed that the molten sodium sulfate Beryllium 65
deposit is required to initiate hot corrosion attack. Source: Ref 2
The temperature range for hot corrosion attack,
Name ///sr-nova/Dclabs_wip/High Temp/5208_249-258.pdf/Chap_09/ 26/10/2007 2:02PM Plate # 0 pg 250

250 / High-Temperature Corrosion and Materials Applications

mechanism by salt fluxing has been discussed data show a good correlation between alloy
in detail in Ref 7, 8, and 9. The topic of hot performance and chromium content. Increasing
corrosion has been extensively covered in chromium in the alloy significantly improves
several reports and conference proceedings resistance to hot corrosion. Alloys with 15% Cr
(Ref 2, 10–13). or less are very susceptible to hot corrosion
attack. Cobalt-base alloys are generally better
than nickel-base alloys. This may simply be due
9.2 Alloys Resistant to Hot Corrosion to higher chromium contents in cobalt-base
alloys. One nickel-base alloy (Hastelloy X) with
Various test methods have been used to study a chromium level similar to those of cobalt-base
hot corrosion. Immersion testing (or crucible test- alloys was found to behave similarly to cobalt-
ing), which was the first laboratory test method, is base alloys.
not considered reliable for simulating the gas Among the alloys tested (Ref 17), alloy X-40
turbine environment (Ref 14, 15). The salt-coated (Co-25Cr-10Ni-7.5W) performed best. This is in
method is quite popular in academia for studying good agreement with the operating experience
corrosion mechanisms. Engine manufacturers, obtained by Royal Navy Ship (U.K.), which has
however, use the burner rig test system to deter- demonstrated the superior hot corrosion resist-
mine relative alloy performance ranking. The rig ance of alloy X-40 in a marine environment
burns fuel with excess air to produce combustion (Ref 18). Alloy X-40 was also found to be sig-
gases with continuous injection of a synthetic nificantly better than nickel-base alloys (Ref 19),
sea-salt solution. This type of test system rep- such as B-1900, U-700, U-500, and IN738
resents the best laboratory apparatus for simu- (Table 9.3). After 240 h, alloy X-40 showed
lating the gas turbine environment. A special hardly any corrosion attack, while alloy B-1900
issue of High Temperature Technology published (Ni-10Co-8Cr-6Mo-4.3Ta-6Al-1Ti) suffered
in 1989 contained a number of papers discussing severe attack. Alloy U-500 (Ni-18Co-19Cr-
burner rig test procedures (Ref 16). The data 4Mo-2.9Al-2.9Ti) and IN738 (Ni-8.5Co-16Cr-
reviewed here are limited to those generated by 1.7Mo-2.6W-1.7Ta-0.9Nb-3.4Al-3.4Ti) were
burner rig test systems. similar, suffering only mild attack. Surprisingly,
alloy U-700 (15% Cr) was found to be slightly
worse than alloy B-1900 (8% Cr). Alloy B-1900
9.2.1 High Temperature or Type I
along with IN100 (10% Cr) and Nimonic
Hot Corrosion 100 (11% Cr) were considered to be poor in
Bergman et al. (Ref 17) studied hot corrosion hot corrosion and suggested that they not be
resistance of various nickel- and cobalt-base considered for use without coatings, even in
alloys at temperatures from 870 to 1040 °C mildly corrosive environments (Ref 20).
(1600 to 1900 °F) with 5 ppm sea-salt injection. Burner rig tests were conducted (Ref 21)
Their results are tabulated in Table 9.2. The using residual oil, containing 3% S and 325 ppm

Table 9.2 Results of burner rig hot corrosion tests on nickel- and cobalt-base alloys
Loss in sample diameter, mm (mils)
Chromium content 870 °C (1600 °F) 950 °C (1750 °F) 980 °C (1800 °F) 1040 °C (1900 °F)
Alloy in alloy, % 500 h 1000 h 1000 h 1000 h
SM-200 9.0 1.6 (64.4) 3.3+ (130+) … …
IN100 10.0 3.3+ (130+) 3.3+ (130+) … …
SEL-15 11.0 3.3+ (130+) 3.3+ (130+) … …
IN713 13.0 3.3+ (130+) 2.0+ (77+) … …
U-700 14.8 1.7+ (66+) 1.6 (63.9) … …
SEL 15.0 1.2 (45.8) 1.3 (51.8) 0.3 (11.4) …
U-500 18.5 0.2 (7.6) 0.8 (31.7) 0.7 (29.3) …
Rene 41 19.0 0.3 (10.3) … 0.8 (30.8)
Hastelloy alloy X 22.0 … 0.3 (12.0) 0.4 (15.2) …
L-605 (alloy 25) 20.0 … 0.4 (15.3) 0.3 (11.3) 1.1 (41.9)
WI-52 21.0 0.5 (21.4) 0.5 (18.2) … 1.9 (73.9)
MM-509 21.5 … 0.3 (10.9) … 0.8 (31.8)
SM-302 21.5 0.14 (5.4) 0.3 (10.0) … 0.6 (23.1)
X-40 25.0 0.11 (4.2) 0.3 (11.6) … 0.5 (18.5)
Note: 5 ppm sea salt injection. Source: Ref 17
Name ///sr-nova/Dclabs_wip/High Temp/5208_249-258.pdf/Chap_09/ 26/10/2007 2:02PM Plate # 0 pg 251

Chapter 9: Hot Corrosion in Gas Turbines / 251

NaCl (equivalent to 5 ppm NaCl in air), at In examining the effect of the third alloying
870 °C (1600 °F) for 600 h on several cobalt- element in Ni-Cr alloys, they found that (Ref 22):
base alloys, which were X-45 (Co-25Cr-10Ni-
7.5W), MAR-M302 (Co-21.5Cr-10W-9Ta-0.2Zr),
 Tungsten (8%) showed little effect at 910
and 950 °C (1675 and 1750 °F), but a
MAR-M509 (Co-21.5Cr-10Ni-7W-3.5Ta-0.2T-
slightly detrimental effect at 1040 °C
0.5Zr), S-816 (Co-20Cr-20Ni-4W-4Nb-4Mo),
(1900 °F).
FSX-418 (Co-30Cr-10Ni-7W-0.15Y), and FSX-
414 (Co-30Cr-10Ni-7W). The test results are  Cobalt was only slightly beneficial.
shown in Fig. 9.1. All six cobalt-base alloys with  Molybdenum was detrimental at 1040 °C
chromium varying from 20 to 30% suffered little (1900 °F), but had little effect at lower tem-
corrosion attack (about 0.04 to 0.12 mm, or peratures.
0.002 to 0.005 in., or 2 to 5 mils). Under the  Titanium (5%) showed significant improve-
same test condition, Udimet 700 (Ni-15Cr- ment at 1040 °C (1900 °F), but little effect at
18.5Co-5.2Mo-5.3Al-3.5T) suffered about 0.76 lower temperatures.
mm (0.03 in., or 30 mils) of attack (Ref 21).  Aluminum (8%) was detrimental, causing
Figure 9.2 summarizes the data for a group of severe hot corrosion attack at 1040 °C
nickel- and cobalt-base alloys at 870 to 1040 °C (1900 °F).
(1600 to 1900 °F) (Ref 22). Alloy U-700 was For Co-25Cr alloys, the effect of the third
found to be inferior to IN713 at 950 °C alloying element was summarized as (Ref 22):
(1750 °F). This was contrary to field experience.
Alloy U-700 has served most reliably in aircraft  Tungsten (8%) was detrimental at 1040 °C
jet engines, whereas IN713 has suffered severe (1900 °F), with little effect at lower tem-
hot corrosion in many applications (Ref 22). peratures.
The authors attributed the high corrosion rate
of alloy U-700 to the low chromium content in
this heat (i.e., 13.6% versus 15.0% for regular ATTACK, mm per side
heats). 0 0.04 0.08 0.12
A systematic study was conducted (Ref 22)
to determine the effects of alloying elements on
FSX– 414
hot corrosion resistance. In the Ni-10Co-15Cr-
4Al-2Ti system, decreasing chromium from 25
to 10% resulted in increases in hot corrosion
FSX–418
attack (Fig. 9.3). The data also suggest that
decreasing aluminum while increasing titanium
improves hot corrosion resistance. Furthermore, S–816
addition of 8% W to Ni-10Co-15Cr-4Al-2Ti
alloy resulted in no apparent change in hot
corrosion resistance. In binary and ternary alloy MAR–M 509
systems of nickel- and cobalt-base alloys, these
authors further observed the effectiveness of
chromium in improving hot corrosion resistance MAR–M 302
at 910, 950, and 1040 °C (1675, 1750, and
1900 °F) (Ref 22). Results are shown in Fig. 9.4.
X–45

Table 9.3 Results of burner rig tests at 874 °C 0 1 2 3 4 5 6

(1605 °F) for nickel- and cobalt-base alloys
ATTACK, mils per side
Penetration depth, mm/1000 h (mils/1000 h)
Exposure
time, h B-1900 U-700 U-500 IN738 X-40 Surface loss
100 2.8 (111) … … … …
170 2.5 (97) 3.3 (129) 0.7 (29) … … Maximum penetration
240 2.1 (83) … 0.5 (20) 0.9 (35) Slight
Note: Diesel fuel containing 1.0% S, 125 ppm Na, 15 ppm Mg, 4.8 ppm Ca, Fig. 9.1 Relative hot corrosion resistance of cobalt-base alloys
4.1 ppm K, and 225 ppm Cl, air-to-fuel ratio was 50 : 1, and 100 h cycle. Source: obtained from burner rig tests using 3% S residual oil
Ref 19 and 325 ppm NaCl in fuel (equivalent to 5 ppm NaCl in air) at
870 °C (1600 °F) for 600 h. Source: Beltran (Ref 21)
Name ///sr-nova/Dclabs_wip/High Temp/5208_249-258.pdf/Chap_09/ 26/10/2007 2:02PM Plate # 0 pg 252

252 / High-Temperature Corrosion and Materials Applications

 Molybdenum (6%) was detrimental at 950 at 30 rpm during testing to ensure that all the
and 1040 °C (1750 and 1900 °F), with little specimens were subjected to the same test con-
effect at 910 °C (1675 °F). dition. The specimens were cycled out of the
 Tantalum (7%) and nickel (10%) showed combustion gas stream once every hour for
little effect. 2 min, during which time the specimens were
cooled by forced air (fan cool) to less than
Burner rig tests were conducted (Ref 23) at 205 °C (400 °F). Superalloys tested were
900 °C (1650 °F) on several wrought super- alloy X (Ni-22Cr-18.5Fe-9Mo-0.5W), alloy S
alloys and nickel aluminides. The combustion (Ni-15.5Cr-14.5Mo-0.05La), alloy 230 (Ni-22Cr-
gas stream was generated by using No. 2 fuel oil 14W-2Mo-0.02La), alloy 625 (Ni-21.5Cr-9Mo-
containing about 0.4 wt% S with an air-to-fuel 3.6Nb), alloy 188 (Co-22Cr-22Ni-14W-0.04La),
ratio of 35 to 1 and injection of either 5 or 50 ppm alloy 25 (Co-20Cr-10Ni-15W), and alloy 150
sea salt into the combustion gas stream. The (Co-27Cr-18Fe). Two nickel aluminides,
specimens were loaded in a carousel, that rotated IC-50 (Ni-11.3Al-0.6Zr-0.02B) and IC-218

Surface loss
Maximum
penetration
1600 °F (870 °C)
1750 °F (950 °C)
1900 °F (1040 °C)

SL- Surface loss

MP- Maximum penetration
Alloy

U–700
75.6 SL
92.4 MP

SM–200 80.2

SEL

IN713

U–500

WI–52
109.2

SM–302

X–45

0 10 20 30 40 50 60 70
(250) (500) (750) (1000) (1250) (1500) (1750)
Loss in diameter, mils (µm)

Fig. 9.2 Relative hot corrosion resistance of nickel- and cobalt-base alloys obtained from burner rig tests at 870, 950, and 1040 °C
(1600, 1750, and 1900 °F) for 100 h, using 1% S diesel fuel, 30:1 air-to-fuel ratio, and 200 ppm sea-salt injection. Source:
Bergman et al. (Ref 22)
Name ///sr-nova/Dclabs_wip/High Temp/5208_249-258.pdf/Chap_09/ 26/10/2007 2:02PM Plate # 0 pg 253

Chapter 9: Hot Corrosion in Gas Turbines / 253

(Ni-7.8Cr-8.5Al-0.8Zr-0.02B), which were (1650 °F) for 200 h with 50 ppm sea salt, re-
developed by Oak Ridge National Laboratory, vealing the formation of nickel oxides and nickel
were included in the test program. sulfides. SEM/EDX analysis showed that a thin,
The results of tests at 900 °C (1650 °F) for protective chromium-rich oxide scale formed on
200 h with 50 ppm sea salt are summarized in alloys X, 230, and 188.
Table 9.4 (Ref 23). Both IC-50 and IC-218 nickel Alloy 25, although exhibiting little weight
aluminides suffered severe hot corrosion attack change (Table 9.4), showed evidence of initial
after 200 h at 900 °C (1650 °F) with 50 ppm sea breakdown of the chromium-rich oxide scale.
salt being injected into the combustion gas SEM/EDX analysis revealed the formation of
stream. Scanning electron microscopy with cobalt-rich oxide nodules on the outer oxide
energy-dispersive x-ray spectroscopy (SEM/ scale on alloy 25. This indicated the initiation of
EDX) analysis showed that both nickel alumi- the breakaway corrosion for alloy 25 after 200 h
nides exhibited porous nickel or nickel-rich at 900 °C (1650 °F) with 50 ppm sea salt. Long-
oxides with nickel sulfide penetrating through term test results under the same test condition
the remaining metal (Ref 23). Figure 9.5 shows clearly showed that alloy 25 suffered severe hot
the cross section of a corroded IC-218 specimen corrosion in excess of 200 h of testing, as shown
after hot corrosion burner rig testing at 900 °C in Fig. 9.6 (Ref 23).

Loss in diameters, µm
0 250 500 750 1000 1250 1500 1750 2000
Alloy

Ni-10Co-0.5Nb-4Al-2Ti-25Cr TEL-1
Ni-10Co-0.5Nb-4Al-2Ti-20Cr TEL-2
Ni-10Co-0.5Nb-4Al-2Ti-15Cr TEL-3
140
Ni-10Co-0.5Nb-4Al-2Ti-10Cr TEL-4
Ni-10Co-0.5Nb-6Al-0Ti-15Cr TEL-5
Ni-10Co-0.5Nb-5Al-3Ti-15Cr TEL-6
Ni-10Co-0.5Nb-2Al-4Ti-15Cr TEL-7
Ni-10Co-0.5Nb-8W-4Ai-2Ti-15Cr TEL-8
Ni-10Co-0.5Nb-4Mo-15Cr TEL-9
Ni-0Co-0.5Nb-2Mo-4W-15Cr TEL-10
Ni-25Co-0.5Nb-15Cr TEL-11
Ni-0Co-0.5Nb-6Al-4Ti-15Cr TEL-12

Ni-15Co-4Mo-3.7Al-1.8Ti-19 Cr MELINI-1

Ni-15Co-3.5Al-1.8Ti-4.5 Re-19Cr MELINI-2

Ni-16Co-3.6Al-2.1Ti-21Cr MELINI-3

Ni-15Co-4Mo-2.9Al-1.8Ti-0.2Y-19Cr MELNI-4 Surface Maximum

loss penetration

Ni-25Co-4Mo-3.5Al-2.2Ti-0.2Y-19Cr MELNI-5
950 °C (1750 °F)
1040 °C (1900 °F)
Ni-15Co-4Mo-3.7Al-1.8Ti-0.08Y-19Cr MELNI-6

0 10 20 30 40 50 60 70 80
Loss in diameters, mils

Fig. 9.3 Relative hot corrosion resistance of experimental alloys obtained from burner rig tests at 950 and 1040 °C (1750 and
1900 °F) for 100 h, using 1% S diesel fuel, 30:1 air-to-fuel ratio, and 200 ppm sea-salt injection. Source: Bergman et al.
(Ref 22)
Name ///sr-nova/Dclabs_wip/High Temp/5208_249-258.pdf/Chap_09/ 26/10/2007 2:02PM Plate # 0 pg 254

254 / High-Temperature Corrosion and Materials Applications

Figure 9.6 also shows alloys 230 and 188 are summarized in Table 9.5 (Ref 23). Three
exhibiting very little weight change for exposure nickel-base alloys (alloys S, X, and 625) and one
time up to 1000 h. The results of the 1000 h tests cobalt-base alloy (alloy 25) were completely
corroded before the test reached 1000 h, while
Maximum penetration
nickel-base alloy 230 and cobalt-base alloys 150
measurements and 188 exhibited little corrosion attack after
1675 °F (910 °C) 1000 h. Under the same test conditions in the
1750 °F (950 °C)
1900 °F (1040 °C)
same burner rig, nickel-base alloy HR-160 with
29% Co, 28% Cr and 2.75% Si was found to
CP-Complete penetration
Alloy through the specimen perform as well as alloys 230, 150, and 188.
Figure 9.7 shows the conditions of the test spe-
RL-1
UNALLOYED NICKEL } 130CP cimens comparing HR-160 with other wrought
alloys (Ref 24). Another burner rig test was
RL-2
Ni-10Cr conducted (Ref 23) with 5 ppm sea salt at 900 °C
RL-3 (1650 °F) under the same combustion conditions
Ni-15Cr
(i.e., No. 2 fuel oil with 0.4% S and 35-to-1 air-
RL-4
Ni-25Cr
to-fuel ratio). The results are summarized in
Table 9.6. Even at this low level of sea salt
RL-5
Ni-15Cr-8W (5 ppm) in the combustion gas stream, cobalt-
RL-6 base alloy 25 continued to exhibit very poor hot
Ni-15Cr-24Co
corrosion resistance compared with some nickel-
RL-7
Ni-15Cr-6Mo 130CP
base alloys.
RL-8
Ni-15Cr-5Ti
9.2.2 Low-Temperature or Type II
RL-9
Ni-15Cr-8Al 90 Hot Corrosion
RL-10 } 130CP “Low-temperature” or Type II hot corrosion
UNALLOYED COBALT
has been observed at temperatures lower than
RL-11 the temperature range where Type I hot corrosion
Co-25Cr
has been encountered. Severe hot corrosion
RL-12
Co-25Cr-8W of alloy S590 and Nimonic 80A after several
RL-13 thousand hours of operation in a gas turbine that
Co-25Cr-6Mo 130CP
burned blast-furnace gas with the 700 to 730 °C
RL-14
Co-25Cr-7Ta
(1290 to 1345 °F) turbine entry temperature was
reported in Ref 25. In 1976, a new form of hot
RL-15
Co-25Cr-10Ni corrosion attack of gas turbine airfoil materials
in a marine gas turbine was reported (Ref 26).
0 10 20 30 40 50 60 70
(250) (500) (750) (1000) (1250) (1500) (1750) The first-stage turbine blades coated with a
Loss in diameter, mils (µm)
CoCrAlY coating, which had exhibited satis-
factory hot corrosion resistance for metal tem-
Fig. 9.4 Relative hot corrosion resistance of experimental peratures in the range 800 to 1000 °C (1470 to
alloys obtained from burner rig tests at 910, 950, and 1830 °F), were found to suffer corrosion attack
1040 °C (1675, 1750, and 1900 °F) for 100 h, using 1% S diesel
fuel, 30:1 air-to-fuel ratio, and 200 ppm sea salt injection. Source: for metal temperatures at about 600 to 730 °C
Bergman et al. (Ref 22) (1110 to 1345 °F) (Ref 26). The corrosion

Table 9.4 Results of burner rig hot corrosion tests at 900 °C (1650 °F) for 200 h with 50 ppm
sea salt with specimens being cycled once every hour
Alloy Weight change, mg/cm2 Metal loss, mm (mils) Total depth of attack(a), mm (mils)
X −0.76 0.02 (0.6) 0.07 (2.8)
230 −1.35 0.02 (0.8) 0.06 (2.4)
25 −1.62 0.02 (0.9) 0.07 (2.8)
188 0.93 0.02 (0.7) 0.04 (1.6)
IC-50 72 >0.72 (28.3), completely corroded >0.72 (28.3), completely corroded
IC-218 83 >0.75 (29.5), completely corroded >0.75 (29.5), completely corroded
(a) Metal loss + internal penetration. Source: Ref 23
Name ///sr-nova/Dclabs_wip/High Temp/5208_249-258.pdf/Chap_09/ 26/10/2007 2:02PM Plate # 0 pg 255

Chapter 9: Hot Corrosion in Gas Turbines / 255

4
5

3
2
1

100 µm

Fig. 9.5 Scanning electron backscattered image showing the cross section of a corroded IC-218 nickel aluminide specimen after hot
corrosion burner rig testing at 900 °C (1650 °F) for 200 h with 50 ppm sea salt using No. 2 fuel oil (0.4% S) for combustion at
35:1 air-to-fuel ratio. The results (wt%) of EDX analysis are: 1: 100% Ni; 2: 74% Ni, 26% S; 3: 100% Ni; 4: 88% Ni, 8% Cr, 4% Al; and 5:
98% Ni, 2% Al. Areas 1, 4, and 5 were essentially nickel oxides, area 2 was nickel sulfide, and area 3 was pure nickel. Courtesy of Haynes
International, Inc.

products formed on the CoCrAlY coating were corrosion (705 °C, or 1300 °F) compared with
found to contain CoSO4 and NiSO4 (Ref 26). It CoCrAlY coating with about 20% Cr. A critical
was proposed (Ref 27) that the mechanism of chromium content of no less than 37% was
low-temperature hot corrosion attack of a required for cobalt-base coatings to provide
CoCrAlY coating involved the formation of the resistance to both low- and high-temperature hot
low melting Na2SO4-CoSO4 eutectic (melting corrosion (Ref 33).
point of 565 °C, or 1045 °F). Nickel-base alloys
were, in general, more resistant to Type II hot
corrosion than cobalt-base alloys as found in Ref 9.3 Summary
28. It was also found (Ref 29) that NiCrAlY and
NiCoCrAlY coatings were, in general, more High-temperature or Type I hot corrosion
resistant than CoCrAlY coatings. generally occurs in the temperature range of
Increases in chromium content in superalloys 800 to 950 °C (1470 to 1740 °F). It is believed
and coatings provided significant increases in that the molten sodium sulfate deposit is re-
low-temperature hot corrosion resistance quired to initiate hot corrosion attack. The Type I
(750 °C, or 1380 °F) for these materials hot corrosion morphology is typically charac-
(Ref 30). Both IN939 (23% Cr) and NiCrAlY terized by a thick, porous layer of oxides with the
coating (39% Cr) were found to exhibit good underlying alloy matrix depleted in chromium,
low-temperature hot corrosion resistance. followed by internal chromium-rich sulfides.
NiCrAlY coatings with 26, 34, and 42% Cr were Low-temperature or Type II hot corrosion gen-
tested and found significant resistance to low- erally occurs in the temperature range of 670
temperature hot corrosion for coatings with only to 750 °C (1238 to 1382 °F). Type II hot corro-
34 and 42% Cr (Ref 31). MCrAlY coatings sion is characterized by pitting attack with
(M = Co and/or Ni) with 30 to 35% Cr were little or no internal attack underneath the pit.
tested (Ref 32). All of the coatings showed Cobalt-base alloys are more susceptible to Type
improved resistance to low-temperature hot II hot corrosion, which generally involves
Name ///sr-nova/Dclabs_wip/High Temp/5208_249-258.pdf/Chap_09/ 26/10/2007 2:02PM Plate # 0 pg 256

256 / High-Temperature Corrosion and Materials Applications

20 Table 9.5 Results of burner rig hot corrosion

Haynes alloy 230 tests at 900 °C (1650 °F) for 1000 h with 50 ppm
10
sea salt with specimens being cycled once every
0 Haynes alloy 188 hour
–10 Weight change,
Alloy mg/cm2 Total depth of attack(a), mm (mils)
–20
S … Completely corroded in 350 h
–30 (1.3 mm, or 49.2 mil thick specimen)
X … Completely corroded in 500 h
–40 (1.2 mm, or 45.9 mil thick specimen)
625 … Completely corroded in 940 h
–50 (1.6 mm, or 63.7 mil thick specimen)
Weight change, mg/cm2

230 11.5 0.11 (4.4)

–60
25 … Completely corroded in 476 h
–70 (1.0 mm, or 37.8 mil thick specimen)
150 10.0 0.13 (5.2)
–80 188 9.9 0.08 (3.2)
–90 (a) Metal loss + internal penetration. Source: Ref 23

–100
–110 Haynes alloy 25

–120
–130 5. A.K. Koul, J.P. Immarigeon, R.V. Dainty,
–140
and P.C. Patnaik, Degradation of High Per-
formance Aero-Engine Turbine Blades,
–150 Advanced Materials and Coatings for
–160 Combustion Turbines, V.P. Swaminathan
–170 and N.S. Cheruvu, Ed., ASM International,
1994, p 69
–180
0 100 200 300 400 500 600 700 800 9001000 6. G.H. Meier, High Temperature Corrosion 2
Exposure time, h —Advanced Materials and Coatings (Les
Embiez, France), May 22–26, 1989, Else-
Fig. 9.6 Results of burner rig tests at 900 °C (1650 °F) with vier Science, 1989, p 1
50 ppm sea salt using No. 2 fuel oil (0.4% S) for
combustion at 35:1 air-to-fuel ratio for alloys 230, 188, and 25. 7. J.A. Goebel, F.S. Pettit, and G.W. Goward,
Source: Lai et al. (Ref 23) Metall. Trans., Vol 4, 1973, p 261
8. J. Stringer, Am. Rev. Mater. Sci., Vol 7, 1977,
p 477
Na2SO4 and CoSO4. Increasing chromium in 9. R.A. Rapp, Corrosion, Vol 42 (No. 10),
alloys or coatings will improve the resistance 1986, p 568
of the material to both Type I and Type II hot 10. J. Stringer, R.I. Jaffee, and T.F. Kearns, Ed.,
corrosion attack. High Temperature Corrosion of Aerospace
Alloys, Advisory Group for Aerospace
Research and Development, North Atlantic
REFERENCES Treaty Organizations, AGARD-CP-120,
Harford House, London, 1973
1. J.G. Tschinkel, Corrosion, Vol 28 (No. 5), 11. J.W. Fairbanks and I. Machlin, Ed., Proc.
1972, p 161 1974 Gas Turbine Materials in The Marine
2. J. Stringer, “Hot Corrosion in Gas Tur- Environment Conf., MCIC-75-27, Battelle
bines,” Report MCIC-72-08, Battelle Columbus Laboratories, 1974
Columbus Laboratories, Columbus, OH, 12. Hot Corrosion Problems Associated with
1972 Gas Turbines, STP 421, ASTM, 1967
3. J. Stringer, High Temperature Corrosion in 13. A.B. Hart and A.J.B. Cutler, Ed., Deposition
Energy Systems, M.F. Rothman, Ed., The and Corrosion in Gas Turbines, Applied
Metallurgical Society of AIME, 1985, p 3 Science, London, 1973
4. J. Stringer, Coatings in the Electricity Sup- 14. J.F.G. Conde and G.C. Booth, Deposition
ply Industry: Past, Present, and Opportu- and Corrosion in Gas Turbines, A.B. Hart
nities for the Future, Surf. Coat. Technol., and A.J.B. Cutler, Ed., John Wiley & Sons,
Vol 108/109, 1998, p 1 1973, p 278
Name ///sr-nova/Dclabs_wip/High Temp/5208_249-258.pdf/Chap_09/ 26/10/2007 2:02PM Plate # 0 pg 257

Chapter 9: Hot Corrosion in Gas Turbines / 257

310SS RA330 alloy Alloy 800H Alloy 625 HR160® alloy

1000 h 476 h 1000 h 940 h 1000 h

Fig. 9.7 Test specimens alloy HR-160, 625, 800H, RA330, and Type 310 at 900 °C (1650 °F) in the combustion gas stream generated
by a burner rig using No. 2 fuel oil (0.4% S) for combustion at 35:1 air-to-fuel ratio and with injection of 50 ppm sea salt into the
combustion gas stream. During testing, specimens were cycled once every hour. Source: Ref 24

Table 9.6 Results of burner rig hot corrosion 19. M.J. Zetlmeisl, D.F. Laurence, and K.J.
tests at 900 °C (1650 °F) for 1000 h with 5 ppm McCarthy, Mater. Perform., June 1984, p 41
sea salt with specimens being cycled once 20. J. Stringer, Proc. Symp. Properties of High
every hour Temperature Alloys with Emphasis on
Weight change, Metal loss, Total depth of attack(a), Environmental Effects, Z.A. Foroulis and
Alloy mg/cm2 mm (mils) mm (mils)
F.S. Pettit, Ed., The Electrochemical Society,
X −0.24 0.04 (1.6) 0.14 (5.5)
230 −0.79 0.03 (1.2) 0.13 (5.1)
1976, p 513
625 5.87 0.05 (1.9) 0.14 (5.3) 21. A.M. Beltran, Cobalt, Vol 46, 1970, p 3
25 … … Completely corroded 22. P.A. Bergman, C.T. Sims, and A.M. Beltran,
(1.09 mm, or 43 mils)
188 1.09 0.02 (0.8) 0.07 (2.8) Hot Corrosion Problems Associated with
(a) Metal loss + internal penetration. Source: Ref 23
Gas Turbines, STP 421, ASTM, 1967,
p 38
23. G.Y. Lai, J.J. Barnes, and J.E. Barnes, “A
Burner Rig Investigation of the Hot Corro-
sion Behavior of Several Wrought Super-
15. M.J. Donachie, R.A. Sprague, R.N. Russell, alloys and Intermetallics,” Paper 91-GT-21,
K.G. Boll, and E.F. Bradley, Hot Corrosion International Gas Turbine and Aeroengine
Problems Associated with Gas Turbines, Congress and Exposition (Orlando, FL),
STP 421, ASTM, 1967, p 85 June 3–6, 1991
16. High Temperature Technology, Special Issue 24. “Haynes HR-160 Alloy,” H-3129A, Haynes
on Hot-Salt Corrosion Standards Test Pro- International, Inc., Kokomo, IN
cedures and Performance, Vol 7 (No. 4), 25. W. Moller, Deposition and Corrosion in Gas
Nov 1989 Turbines, A.B. Hart and A.J.B. Cutler, Ed.,
17. P.A. Bergman, A.M. Beltran, and C.T. Sims, John Wiley and Sons, 1973, p 1
Development of Hot Corrosion-Resistant 26. D.J. Wortman, R.E. Fryxell, and I.I. Bessen,
Alloys for Marine Gas Turbine Service, A Theory for Accelerated Turbine Corrosion
Final Summary Report to Marine Engi- at Intermediate Temperatures, Proc. Third
neering Lab., Contract N600 (61533) 65661, Conference on Gas Turbine Materials in a
Navy Ship R&D Center, Annapolis, MD, Marine Environment, Session V, Paper 11
Oct 1, 1967 (Bath, England), 1976
18. J. Clelland, A.F. Taylor, and L. Wortley, 27. K.L. Luthra, Metall. Trans. A, Vol 13, 1982,
Proc. 1974 Gas Turbine Materials in the p 1853
Marine Environment Conf., MCIC-75-27, 28. D.J. Wortman, R.E. Fryxell, K.L. Luthra,
J.W. Fairbanks and I. Machlin, Ed., Battelle and P.A. Bergman, Proc. Fourth US/UK
Columbus Laboratories, 1974, p 397 Conf. on Gas Turbine Materials in a Marine
Name ///sr-nova/Dclabs_wip/High Temp/5208_249-258.pdf/Chap_09/ 26/10/2007 2:02PM Plate # 0 pg 258

258 / High-Temperature Corrosion and Materials Applications

Environment (Annapolis, MD), U.S. Naval Marine Gas Turbines, Proc. Conf. High Tem-
Academy, 1979, p 317 perature Alloys for Gas Turbines, COST 50,
29. L.F. Aprigliano, Proc. Fourth US/UK Conf. 1982, p 237
on Gas Turbine Materials in a Marine 32. J.A. Goebel, Advanced Coating Develop-
Environment (Annapolis, MD), U.S. Naval ment for Industrial/Utility Gas Turbine
Academy, 1979, p 151 Engines, Proc. First Conf. on Advanced
30. A.R. Taylor, B.A. Wareham, G.C. Booth, Materials for Alternative Fuel Capable
and J.F. Conde, Low and High Pressure Rig Directly Fired Heat Engines (Castine, ME),
Evaluation of Materials and Coatings, Proc. Department of Energy–Electric Power
Third Conference on Gas Turbine Materials Research Institute, 1979, p 473
in a Marine Environment, Session III, Paper 33. K.L. Luthra and J.H. Wood, High Chro-
No. 3 (Bath, England), 1976 mium Cobalt-Base Coatings for Low Tem-
31. J.F.G. Conde, G.C. Booth, A.F. Taylor, perature Hot Corrosion, Thin Solid Films,
and C.G. McGreath, Hot Corrosion in Vol 119, 1984, p 271
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:01PM Plate # 0 pg 259

High-Temperature Corrosion And Materials Applications Copyright © 2007 ASM International®

George Y. Lai, editor, p259-320 All rights reserved.
DOI: 10.1361/hcma2007p259 www.asminternational.org

CHAPTER 10

Coal-Fired Boilers

10.1 Introduction (3208 psig) critical point, while the supercritical

unit operates at pressures higher than 218.2 atm
A coal-fired electric power generating station (3208 psig).
is a highly complex plant. A major piece of the When heated at atmosphere pressure, water
equipment in the power plant is the boiler, which boils at 100 °C (212 °F). As the pressure of the
generates steam that is then delivered to the tur- water increases, the boiling (saturation) tem-
bine for generation of electricity. This chapter perature also increases. The saturation tempera-
reviews high-temperature corrosion problems ture becomes constant once the water is boiling at
associated with utility boilers. A brief description a given pressure. Accordingly, the saturation
of boilers is included to help readers understand temperature increases as the pressure increases.
some of the basics in the boiler. A brief discus- Thus, for subcritical units, the saturation tem-
sion of coal is also presented since different types perature is established by a given operating
of coal may present different issues in combus- waterwall water/steam pressure. For example,
tion, thus resulting in different types of materials for the pressure of 170 atm (2500 psig), the
issues and corrosion. More detailed discussion saturated temperature is 353.4 °C (668.11 °F)
and coverage of boilers and the fuel can be found (Ref 3). In a subcritical unit, the water (slightly
in Steam—Its Generation and Use (Babcock & undercooled) in the waterwall tubes rises from
Wilcox, Ref 1) and Combustion Fossil Power the bottom of the waterwall. As the water absorbs
(Combustion Engineering, Ref 2). heat, bubbles form. More bubbles form as the
water continues to rise in the waterwall tubes as
shown in Figs. 10.1 and 10.2 (Ref 1). At the top
of the waterwall tubes, a mixture of steam and
10.2 Boiler Description water leaves the waterwall tubes and enters the
steam drum where steam is separated to be heated
The primary function of the boiler is to gen- further in the convection section for higher
erate steam for the turbine to convert to elec- temperatures prior to delivery to the turbine for
tricity. The steam is converted from water in a electricity generation. Water at the bottom of the
furnace where heat is generated by combustion steam drum is then mixed with the replacement
of coal. The modern furnace in a square or water (feedwater) and is returned to the bottom of
rectangular horizontal cross section is enclosed the boiler through an unheated downcomer to
with water-cooled tubes with a narrow plate reenter the waterwall tubes.
(commonly referred to as a “membrane”) con- In a supercritical unit, water at pressures above
nected between two adjacent tubes. When the 218.2 atm (3208 psig) does not boil. The tem-
furnace is heated, the water under high pressure perature rises steadily, and the water changes
inside these water-cooled tubes (also referred completely to steam. There is no saturation tem-
to as waterwalls) is converted to steam. Some perature and no need for a steam drum. The
boilers use a tangent tube design for waterwalls supercritical unit operates on a once-through
where no membranes are used to connect adja- flow system with water flowing in the furnace
cent tubes. waterwall tubes from the bottom of the furnace
In terms of the operating pressure of the water/ and changing to 100% steam with a continuous
steam in the waterwalls, there are two classifi- temperature increase. The steam then leaves the
cations of boilers: subcritical and supercritical furnace waterwall tubes and is further heated
units. The subcritical unit is the boiler that in superheaters before entering into the steam
operates at pressures below the 218.2 atm turbines.
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:01PM Plate # 0 pg 260

260 / High-Temperature Corrosion and Materials Applications

There are several methods for burning coal. bottom of the furnace is used to feed fuel, such as
Firing with pulverized coal has been the most coal, onto a grate where coal is burnt. Fluidized-
dominant method for utility boilers. Coal is bed combustion (FBC) technology is considered
burned as fine powers suspended in the furnace, to be an emerging technology for power gen-
and almost all types of coal from anthracite eration. In a FBC boiler, combustion of coal (or
to lignite can be burned by pulverized firing other type of fuel) takes place in a fluidized bed.
(Ref 1). Pulverized coal in particle sizes of There are two types of bed: bubbling bed and
50 µm diameter or smaller can be completely circulating bed. The bed temperature is typically
combusted in a matter of 1 to 2 s (Ref 1). maintained in a range of 816 to 900 °C (1500 to
Babcock & Wilcox developed the Cyclone fur- 1650 °F) for maximum efficiency of both com-
nace for firing coal grades that have a low fusion bustion and sulfur capture. The bed consists of
temperature and are not suitable for pulverized essentially sand and limestone or dolomite. Air
coal firing because of potentially forming a enters the bed from the air distributor plate at the
molten slag, thus developing a severe slagging bottom and renders the bed into a fluidization
problem in superheaters (Ref 1). In Cyclone condition. Limestone or dolomite is added to the
boilers, the Cyclone barrels burn coal in such a bed as a sorbent to capture SO2 from the flue gas
way that most of the coal ash is captured to form a by forming calcium sulfate (CaSO4). In a bub-
molten slag that coats the inside surface of the bling bed, an evaporator tube bundle is typically
Cyclone barrels (Ref 1). The combustion flue gas immersed in the bed to help maintain the
from the Cyclone barrels then enters the main designed bed temperature. Figure 10.3 shows a
furnace to generate steam. schematic of a bubbling fluidized-bed boiler
Stoker-firing boilers are very versatile for (Ref 1).
burning a wide range of solid fuels including
various types of coals, municipal waste, wood
waste, and other types of biomass fuels. In a
stoker-firing boiler, a mechanical stoker at the 10.3 Coal and Coal Ash
Coal is generally classified by rank, which
indicates the geological formation history of the
coal. There are four different types of coal. The
youngest, or lowest-rank coal, is lignite, which is
then followed by an older, or higher-rank sub-
bituminous coal, then bituminous coal, and then
anthracitic coal. Heating values, moisture con-
tent, volatile matter content, ash content, and
sulfur content can be different among these dif-
ferent types of coals. Bituminous coal is the most
commonly used coal for utility boilers in the
United States (Ref 1). Subbituminous coal in the
United States generally contains very low sulfur,
with many deposits containing less than 1%.
Table 10.1 shows the properties of some U.S.
coal (Ref 1). There is a big variation in moisture,
ash content, ash softening temperatures, and
sulfur content from various grades of coals. The
material factors associated with coal can directly
or indirectly affect the boiler tube material per-
formance at different locations in the boiler.
Sulfur is the most important impurity in coal
for causing high-temperature corrosion in the
boiler. Sulfur is present in coal in forms of
Fig. 10.1 Water enters the waterwall tubes at the furnace
bottom and turns into a mixture of water and steam organic sulfur, pyritic sulfur (i.e., pyrite), and
that leaves the waterwall tubes at the top (C) and enters the steam iron sulfate. High-sulfur coals also cause SOx
drum where steam and water is separated. Water mixed with the
replacement water (A) is returned to the waterwall tubes at the emission problems and require expensive air
furnace bottom (B). Source: Ref 1. Courtesy of Babcock & Wilcox pollution control equipment. As a result of the
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:01PM Plate # 0 pg 261

Chapter 10: Coal-Fired Boilers / 261

U.S. Federal Clean Air Act emissions issues, the blowers (with steam), waterlances, or water
low-sulfur, Powder River Basin (PRB) coal (a cannons (with room-temperature water),
subbituminous coal) has become extremely depending on the nature of the ash deposits.
popular for the past few years (Ref 4). The PRB One of the important characteristics of ash is
coal, which is from mines in southern Montana its fusion temperature. Table 10.1 lists the ash
through northern Wyoming, contains less than fusion temperatures of several coals under
1.2 lb of sulfur per million btu, making it com- reducing conditions (Ref 1). These fusion tem-
pliant with Clean Air Act emissions limits peratures can affect the nature of the ash deposits,
without air pollution control equipment (Ref 4). whether in the form of “dust” or a tenacious slag.
Ash from coal after combustion can be If ash reaches the heat-absorbing surface at a
entrained in flue gas to cause fly-ash erosion on temperature near its softening temperature, the
heat-absorbing surfaces in the lower furnace, resulting deposits are likely to be porous and can
such as waterwalls, and in the convection pass, be easily removed by sootblowing. Also, if such
such as superheater, reheater, generating bank, a deposit is subjected to high gas temperature, the
and economizer. Ash can also deposit on the ash deposit can reach its melting point (due to the
furnace wall, causing a slagging problem, and on thermal insulating properties of the ash) and run
superheater and reheater bundles, causing a down the furnace wall surface (Ref 1). This
fouling problem. Ash deposits can cause heat solidified slag is tightly bonded and is difficult to
transfer problems, and they are required to be remove. This slag may require water lances or
removed regularly using devices such as soot water cannons to create thermal shock for the

Fig. 10.2 Same water and water/steam circulation in the furnace waterwall tubes as in Fig. 10.1 with illustration of the furnace
waterwall tubes. Source: Ref 1. Courtesy of Babcock & Wilcox
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:01PM Plate # 0 pg 262

262 / High-Temperature Corrosion and Materials Applications

removal of this slag deposit. The furnace walls (e.g., Fe2O3, CaO, MgO, Na2O, and K2O) to
that are subject to radiant heat are likely locations acidic oxides (e.g., SiO2, Al2O3, and TiO2) may
for developing this slagging problem. determine the fusion (fusibility) temperature of
The analyses of ash for several types of U.S. the reaction product. It was reported that the ash
coal are shown in Table 10.1. The compositions may exhibit low fusibility temperature with
shown in Table 10.1 are presented as oxides. higher slagging potential when its base/acid ratio
However, most ash constituents in coal are min- is in a range of 0.4 to 0.7 (Ref 2). Many other
erals. Typical minerals found in coal are shown parameters, such as SiO2/Al2O3 ratio, Fe2O3/
in Table 10.2 (Ref 2). When these minerals are CaO ratio, Fe2O3/(CaO+MgO) ratio, (Na2O +
exposed to oxidizing environments at appro- K2O), and so forth are also used for predicting the
priate high temperatures, oxides are likely to be fusibility temperature of the coal ash. It has been
the stable phases. Most of these oxides typically suggested that SiO2 is more likely than Al2O3 to
exhibit very high melting points. However, some form lower melting species (Ref 2). The fusibility
minerals may themselves react at combustion temperature of coal ash will be lowered when the
temperatures to form reaction products, most Fe2O3/CaO ratios are in a range of 0.2 to 10.
likely complex salts, that may exhibit much Alkalies are important in affecting the fusibility
lower melting temperatures. This is illustrated of coal ash and the furnace slagging potential
in Table 10.3 (Ref 2). If the ash constituents (Ref 2). Many sodium compounds melt at
are present as oxides, the ratio of basic oxides temperatures below 900 °C (1650 °F) (Ref 2).

Distributor
Plate Windbox

Fig. 10.3 Bubbling fluidized-bed boiler. Source: Ref 1. Courtesy of Babcock & Wilcox
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:02PM Plate # 0 pg 263

Chapter 10: Coal-Fired Boilers / 263

Table 10.1 Properties of several types of U.S. coal

Coal type
Illinois #6 Spring Creek Bryan
Bituminous Subbituminous Lignite
Property Anthracite Illinois Wyoming Texas
Proximate composition, %
Moisture 7.7 17.6 24.1 34.1
Volatile matter, dry 6.4 44.2 43.1 31.5
Fixed carbon, dry 83.1 45.0 51.2 18.1
Ash, dry 10.5 10.8 5.7 50.4
Composition, %
Carbon 83.7 69.0 70.3 33.8
Hydrogen 1.9 4.9 5.0 3.3
Nitrogen 0.9 1.0 1.0 0.4
Sulfur 0.7 4.3 0.4 1.0
Oxygen 2.3 10.0 17.7 11.1
Ash 10.5 10.8 5.7 50.4
Thermal properties
Heating value (as received), Btu/lb 11,890 10,300 9,190 3,930
Ash fusion temperatures (reducing atmosphere), °F
Initial deformation … 1930 2100 2370
Softening … 2040 2160 2580
Fluid … 2700 2700 2900+
Ash analysis(a), %
SiO2 51.0 41.7 32.6 62.4
Al2O3 34.0 20.0 13.4 21.5
Fe2O3 3.5 19.0 7.5 3.0
TiO2 2.4 0.8 1.6 0.5
CaO 0.6 8.0 15.1 3.0
MgO 0.3 0.8 4.3 1.2
Na2O 0.7 1.6 7.4 0.6
K 2O 2.6 1.6 0.9 0.9
P 2O 5 … … 0.4 …
SO3 1.4 4.4 14.6 3.5
(a) Elements present in the ash are determined and reported as oxides. Source: Ref 1

More detailed discussion about these parameters Table 10.2 Typical mineral species found in
can be found in Ref 1 and 2. Chlorine content coal
is also an important indication for fouling Mineral species Formula
potential. When chlorine in coal is greater than Kaolinite Al2O3·2SiO2·H2O
0.3%, fouling potential becomes high (Ref 2). Illite K2O·3Al2O3·6SiO2·2H2O
Biotite K2O·MgO·Al2O3·3SiO2·H2O
Alkali metals and chlorine can also play a sig- Orthoclase K2O·Al2O3·6SiO2
nificant role in high-temperature corrosion in Albite Na2O·Al2O3·6SiO2
boilers. Calcite CaCO3
Dolomite CaCO3·MgCO3
Siderite FeCO3
Pyrite FeS2
Gypsum CaSO4·2H2O
10.4 Combustion Environments Quartz SiO2
Hematite Fe2O3
Magnetite Fe3O4
Stoichiometric combustion is a complete Rutile TiO2
combustion of a fuel with a theoretically calcu- Halite NaCl
lated amount of oxygen in the combustion air. Sylvite KCl
Complete combustion of all combustible con- Source: Ref 2
stituents in coal requires excess air beyond the
theoretically calculated value needed in the
combustion air. Excess air is typically expressed
in terms of percentage above the theoretically Nitrogen oxides (NOx), which form during
calculated value. For pulverized coal, typically combustion under oxidizing conditions, are an
15 to 30% excess air is required to enssure ade- undesirable pollutant from the boiler and con-
quate combustion (Ref 2). This will make the tribute to acid rain and ozone formation. In order
combustion environment an oxidizing atmo- to reduce this NOx (i.e., NO and NO2) emission
sphere. from the coal-fired boiler, low NOx burner
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:02PM Plate # 0 pg 264

264 / High-Temperature Corrosion and Materials Applications

systems and staged combustion techniques have 3000 °F), some of the mineral constituents and
been developed and employed with great success compounds can be in a molten or plastic state
in significantly reducting NOx emission. The (see Tables 10.1 and 10.3). These ash con-
majority of NOx is formed by high-temperature stituents can then deposit on the heat-absorbing
oxidation of N2 in the combustion air under surfaces in the boiler. Two general types of ash
oxidizing conditions. Nitrogen in coal can also deposition, slagging and fouling, can take place.
lead to formation of NOx during combustion. Slagging is the deposition of molten, partially
Reducing the availability of air during initial fused deposits on the furnace walls and the upper
combustion can significantly reduce the forma- furnace radiant superheaters exposed to radiant
tion of NOx. This results in reducing environ- heat (Ref 1). Fouling is the deposition of more
ments. For staged combustion, the lower furnace loosely bonded deposits on the heat-absorbing
is under reducing conditions, and additional air is surfaces in the convection pass, such as super-
then introduced at a higher elevation in the fur- heater and reheater, that are not exposed to
nace to complete the combustion process. More radiant heat (Ref 1). These ash deposits, which
detailed discussion of NOx control in combustion are impeding the heat transfer, require their
is available in Ref 1 and 2. As a result of low NOx regular removal from the tube surfaces. Soot
combustion techniques, the furnace environment blowers using steam are generally adequate for
changes from an oxidizing to a reducing condi- removing the ash deposits on the fouled tube
tion. The location for this change is variable and surfaces, while waterlances or water cannons
will fluctuate as the load changes for the boiler using thermal shock by water may be needed
operation especially under non-base-load opera- to clean the slagged tube surfaces. Material
tion. Sulfur in the coal forms H2S in reducing behavior under the influence of ash deposits is
conditions instead of SO2 in oxidizing condi- discussed in later sections.
tions. High wastage rates for waterwall tubes
under reducing conditions and oxidizing/redu-
cing alternating conditions can become a very
serious materials issue. 10.5 Fireside Corrosion and Other
Ash released from coal during combustion can Materials Problems in Boilers
cause significant problems to the behavior of the
boiler components. During combustion, signifi- This section discusses the materials problems
cant amounts of ash are carried by the flue gas related to the heat-absorbing surfaces in the fur-
stream and cause fly-ash erosion problems par- nace combustion area (i.e., waterwalls) and also
ticularly for superheaters and reheaters in the in the convection pass, such as superheaters
convection path. In addition, it can produce and reheaters. A schematic of a coal-fired boiler
slagging on the furnace walls and fouling for showing these two areas is shown in Fig. 10.4.
superheaters and reheaters. The amount of ash Accelerated tube wall wastage attack resulting
being carried by the flue gas stream depends from the change of the combustion environment
on the type of firing, for example, approximately from oxidizing to reducing due to installation of
70 to 90% for pulverized coal units, approxi- low-NOx combustion technology intended to re-
mately 40% for stoker-firing units, and all the duce NOx emissions is discussed. Also discussed
ash along with some fluidized-bed material for in detail are circumferential cracking of the fur-
circulating fluidized-bed boilers (Ref 1). nace waterwalls, soot blower erosion/corrosion,
During combustion, the gas temperature thermal fatigue cracking induced by waterlances
can reach a range of 1370 to 1650 °C (2500 to and water cannons for deslagging the furnace

Table 10.3 Melting points of coal ash constituents

Element Oxide Melting point, °C (°F) Compound Melting point, °C (°F)
Si SiO2 1715 (3120) Na2SiO3 877 (1610)
Al Al2O3 2043 (3710) K2SiO3 977 (1790)
Ti TiO2 1838 (3340) Al2O3·Na2O·6SiO2 1099 (2010)
Fe Fe2O3 1565 (2850) Al2O3·K2O·6SiO2 1149 (2100)
Ca CaO 2521 (4570) FeSiO3 1143 (2090)
Mg MgO 2799 (5070) CaO·Fe2O3 1249 (2280)
Na Na2O Sublimes at 1277 (2330) CaO·MgO·2SiO2 1391 (2535)
K K2O Decomposes at 349 (660) CaSiO3 1540 (2804)
Source: Ref 2
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:02PM Plate # 0 pg 265

Chapter 10: Coal-Fired Boilers / 265

waterwalls, and superheater/reheater corrosion. (i.e., the side facing the combustion), are subject
Limited discussion includes erosion issues in to extreme radiant heat from the combustion. The
fluidized-bed boilers. Hydrogen attack encoun- tube wall exhibits temperature gradient across
tered in subcritical units as a result of water the tube wall thickness. The outer metal tem-
corrosion in the internal diameter of the water- perature depends on the temperature of the water
wall tubes is discussed in “Hydrogen Attack” in and/or steam inside the tube along with many
Chapter 17. other factors. For subcritical units of drum-type
boilers, the temperature of water/steam depends
on the pressure. For example, for the pressure of
10.5.1 Fireside Corrosion of Furnace
170 atm (2500 psig), the saturated temperature
Waterwalls is 353.4 ° C (668.11 °F) (Ref 3). In a supercritical
The furnace is enclosed by four walls (or unit with pressures above 218.2 atm (3208 psig),
waterwalls) with front, rear, and two side walls in the water inside the tube changes to 100% steam
a square or rectangular form for the horizontal with a continuous temperature increase. The tube
cross section of the furnace. The waterwall con- metal temperature for the supercritical unit is
sists of tubes with membranes (i.e., steel plates) generally higher than that of a subcritical unit.
connecting adjacent tubes (a tube-membrane- Other factors that can affect tube metal tem-
tube construction), or of tubes without mem- perature include oxide scales formed on the tube
branes connecting adjacent tubes (a tangent tube inside diameter (ID) surface, slag deposits on the
construction). These waterwalls, which enclose tube outside diameter (OD) surface, flame
the furnace combustion zone, are heat-absorbing impingement, and so forth. Typical temperature
surfaces that heat the rising water in the tubes range for the waterwall tube on the fireside has
(from bottom) to become steam. The gas tem- been given as 380 to 450 °C (716–842 °F) by
perature in the combustion inside the furnace Lees and Whitehead (Ref 6), below about 450 °C
can reach a range of 1370 to 1650 °C (2500 to (842 °F) by Hay (Ref 7), 427 to 482 °C (800 to
3000 °F). As a result, these waterwalls, particu- 900 °F) by Cunningham and Webster (Ref 8),
larly the crown location of the tube on the fireside and 400 to 500 °C (752 to 932 °F) by Stringer

Fig. 10.4 Coal-fired boiler, showing two main areas for discussion of materials problems in a boiler: the furnace combustion
area (i.e., furnace walls) and the heat-absorbing surfaces in the convection path, such as superheaters and reheaters.
Source: Ref 5
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:02PM Plate # 0 pg 266

266 / High-Temperature Corrosion and Materials Applications

(Ref 9). Eurich and Kramer (Ref 10) indicated heat suffers the worst metal wastage. One boiler
that the design waterwall tube surface metal designer suggests the maximum tube metal
temperature was between 482 and 510 °C (900 temperature limits based on oxidation for typical
and 950 °F) for each of the five 676 MW(e) ferritic steels used for waterwalls, as shown in
opposed-fired supercritical boilers. Yeager (Ref Table 10.4 (Ref 2). For example, carbon steel is
11) indicated that the design waterwall surface limited to 454 °C (850 °F), but without men-
temperature for the two supercritical, tangen- tioning the design service life. However, due to
tially fired boilers (each with 750 MWe) at
Montour Station was between 482 and 538 °C
(900 and 1000 °F). Figure 10.5 (Ref 12) shows a
schematic of the temperature gradient across the
tube wall and different oxide layers and ash/slag
deposits.
Wastage. Excess air is normally used for
combustion to ensure adequate combustion of
coal. For example, typically 15 to 30% excess
air is required for firing pulverized coal (Ref 2).
This produces an oxidizing condition with
major combustion gaseous products being N2,
O2, CO2, and H2O. Under this normal oxidizing
condition, waterwalls made of carbon steels
and low-alloy steels have been found to show
metal wastage rates of approximately 40 nm/h
(13 mpy) (Ref 12). A typical tube wastage profile
for the corroded waterwall tube is illustrated in
Fig. 10.6 (Ref 13). The maximum wastage typi-
cally occurred at the crown location (12 o’clock
position). Figure 10.7 shows a cross section of a
carbon steel waterwall tube removed from a
subcritical unit after 21 years of service with an
Fig. 10.6 Typical wastage profile of a corroded waterwall
observed wastage rate of about 0.19 mm/y tube. Source: Ref 13
(7.5 mpy) (Ref 14). The front face of the tube at
the crown location facing the combustion radiant

Fig. 10.7 Cross section of a carbon steel waterwall tube from

a subcritical unit in the United States after 21 years
Fig. 10.5 Temperature gradients through the inner oxide of service, showing the maximum wastage at the crown location
scale, tube wall, outer oxide scale, and ash/slag with about 1.9 mm/yr (7.5 mpy) of wastage rate. Source: Ref 14.
deposits. Source: Ref 12 Courtesy of Welding Services Inc.
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:02PM Plate # 0 pg 267

Chapter 10: Coal-Fired Boilers / 267

the complexity of the combustion in a large coal- oxidizing condition, sulfur oxidizes to form SO2
fired boiler it is almost impossible to establish a during combustion. The presence of SO2 in an
normal oxidizing condition uniformly in the flue oxidizing combustion atmosphere does not gen-
gas throughout the entire furnace cross section in erally cause accelerated corrosion. However,
a dynamic condition. As a result, reducing con- under reducing conditions, sulfur reacts to form
ditions are developed in some localized areas H2S. This can significantly increase corrosion
near the waterwall. Chemical analysis of ash rates by sulfidation. Lees and Whitehead (Ref 6)
deposits formed on tube surfaces revealed an examined several corroded carbon steel water-
appreciable amount of free carbon (Ref 15). The wall tube samples as well as Type 310 co-
presence of free carbon indicates the establish- extruded tube samples removed from several
ment of reducing conditions at the furnace wall power plants in the United Kingdom that suf-
surface (Ref 16). Localized reducing conditions fered severe wastage problems. They had
with up to 10% CO in the furnace atmosphere in observed sulfidation in some severely corroded
the vicinity of the furnace walls have been carbon steel tubes in addition to a chlorine-rich
observed (Ref 17, 18). phase at the boundary between the oxide/sulfide
Sulfidation. One of the most corrosive impu- phases and the metal. (Chlorine issues are dis-
rities in coals is sulfur, which is present in coals cussed next in this section). In one carbon steel
as pyrite (FeS) and organic sulfur. Under an sample that had suffered a wastage rate of about
530 nm/h (170 mpy), they found the corrosion
product to be predominantly sulfides. They even
Table 10.4 Maximum outer tube metal
temperature limits based on oxidation for
observed a substantial layer of oxides/sulfides on
common ferritic steels used in waterwalls, as the corroded Type 310 cladding in a Type 310/CS
suggested by a boiler designer composite tube that suffered more than 50 nm/h
Steel Oxidation limit,
(16 mpy) (Ref 6). Figure 10.8 shows a waterwall
designation Type °C (°F) carbon steel tube after only 1 year of service in a
SA210 A1 Carbon Steel 454 (850) subcritical unit in the United States, showing
T1 Carbon-0.5Mo 482 (900) pitting attack. The boiler was in the United
T11 1.25Cr-0.5Mo 552 (1025)
T22 2.25Cr-1Mo 593 (1100) States, burning a coal containing about 3.0 to
Source: Ref 2
3.5% S and about 300 to 400 ppm chlorine
(Ref 14). Metallurgical examination of the

Fig. 10.8 Close-up view of a waterwall carbon steel tube showing pitting attack after 1 year of service in a subcritical unit in the United
States, burning coal containing about 3.0 to 3.5% S and about 300 to 400 ppm chlorine. Source: Ref 14. Courtesy of
Welding Services Inc.
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:02PM Plate # 0 pg 268

268 / High-Temperature Corrosion and Materials Applications

sample indicated that pitting was associated with converts mostly to HCl during combustion. It is
sulfidation attack, as shown in Fig. 10.9, which believed that every 0.1% Cl in coal produces
shows initiation of internal sulfidation attack approximately 80 ppm HCl in the flue gas (Ref
(Ref 14). The tube was in the initiation stage of 21). Clarke and Morris (Ref 18) performed gas
the accelerated corrosion by sulfidation. There analysis on the gas samples obtained from the
were also areas in the same tube that showed only vicinity of the furnace wall at various locations in
oxidation attack, as illustrated in Fig. 10.10 a 120 MWe boiler that historically suffered high
(Ref 14). wastage rates (excess of 150 nm/h, or 48 mpy, or
Sulfidation of the waterwall tubes in the lower 1.2 mm/yr). This boiler burned high-chlorine
furnace causes a serious wastage issue when the coal containing about 0.4 to 0.6% Cl. The authors
furnace is fired intentionally under reducing observed that reducing conditions occurred in the
conditions (i.e., substoichiometric conditions) areas of high wastage, while oxidizing conditions
when retrofitted with low-NOx burners to reduce were observed in the areas that were outside of
the formation of NOx. This subject is discussed in these high-wastage areas. In addition, the con-
section 10.5.2. centration of HCl was found to be about
Chlorine. Another important impurity in coal 400 ppm in the high-wastage areas. The level of
that can significantly affect the corrosion of H2S was found to be about 300 to 400 ppm
metals in boilers is chlorine. Most coals from the only when CO concentration exceeded 3%, but
U.S. contain low chlorine (<0.1% Cl) (Ref 19). in some areas no H2S was observed under
Thus, there is very little experience and few those reducing conditions. In laboratory tests
studies have been conducted on the effects of in a simulated combustion gas environment
high-chlorine coals on corrosion of metals in
U.S. boilers. British coals contain much higher
chlorine. The average chlorine content in coals
used in British power plants is about 0.25% with
some as much as 0.8% (Ref 20). Chlorine in coal

19 µm

Fig. 10.9 Scanning electron microscopy backscattered Fig. 10.10 Scanning electron microscopy backscattered
electrons image of the corrosion products show- electrons image of the corrosion products
ing initiation of sulfidation attack on the tube, where pitting attack showing ash deposits and iron oxides with no evidence of sulfi-
was observed on the tube surface as shown in Fig. 10.8 Chemical dation attack on other area of the tube that did not suffer pitting
compositions of the phases of the corrosion products at different attack (Fig. 10.8). Chemical compositions of the phases of
locations were analyzed by energy dispersive x-ray spectroscopy the corrosion products at different locations were analyzed
(EDX) with the results summarized as: by energy dispersive x-ray spectroscopy (EDX) with the results
summarized as:
1: 1.1% S, 0.7% Al, 0.8% Si, 0.6% Mn, 95% Fe, and trace
elements 1: 8.9% Si, 1.5% Al, 86.9% Fe, and trace elements
2: 0.8% S, 0.3% Al, 0.6% Si, 0.5% Cl, 0.5% Ca, 1.9% Zn, 94% Fe, 2: 44.3% Si, 22.9% Al, 2.2% Mg, 6.7% Ca, 5.6% K, 14.7% Fe, and
and trace elements trace elements
3: 13.4% S, 0.8% Al, 0.4% Si, 0.8% Mn, 84.1% Fe, and trace 3: 1.7% Si, 1.0% Al, 1.0% S, 1.5% Zn, 93.7% Fe, and trace
elements elements
4: 9.2% S, 0.3% Al, 0.4% Si, 0.6% Mn, 89.3% Fe, and trace 4: 1.3% Si, 1.0% Cu, 1.3% Zn, 1.2% S, 91.7% Fe, and trace
elements elements
5: 1.7% S, 0.8% Al, 1.0% Si, 1.2% Cl, 0.2% Ca, 0.7% Mn, 91.8% 5: 2.9% Cu, 93.8% Fe, and trace elementsa
Fe, and trace elements 6: 1.0% Si, 1.3% Zn, 93.1% Fe, and trace elements
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:02PM Plate # 0 pg 269

Chapter 10: Coal-Fired Boilers / 269

(N2-10%CO-10%H2O-0.5%SO2), Brooks and However, after 27,000 h, both co-extruded tube

Meadowcroft (Ref 22) showed that carbon alloys showed a maximum corrosion rate of about
steel corroded at linear rates of about 130 and 60 nm/h (19 mpy). Both Type 304 and 310 alloys
550 nm/h (42 and 176 mpy) at 400 and 500 °C exhibited significant performance improvement
(752 and 932 °F), respectively, with 400 ppm compared with carbon steel. Examination of
HCl added to the test environment. Their test data corroded 304 and 310 samples revealed sub-
are shown in Fig. 10.11 and 10.12. The labora- stantial amounts of sulfides along with some
tory 400 °C (752 °F) test data of Brooks and chlorine.
Meadowcroft (Ref 22) and the field data of Clarke The effect of chlorine in coals on the furnace
and Morris obtained from a boiler (Ref 18) were wall corrosion was summarized by Daniel et al.
in the same order of magnitude. Figures 10.10 (Ref 19) based on CEGB data generated in boi-
and 10.11 also show that the corrosion of carbon lers in the United Kingdom. The effect is illu-
steel followed a parabolic reaction rate when the strated in Fig. 10.13 and 10.14. Coals with more
environment contained no chlorine. In the than 0.2% Cl may result in accelerated fireside
examination of waterwall tube samples from
several plants by Lees and Whitehead (Ref 6),
corrosion rates were found to vary from about
200 to 530 nm/h (64 to 170 mpy) in different
plants. The corrosion products the authors
observed were mainly sulfides with a chlorine-
rich phase along the metal/scale interface for
some samples. Other samples contained mainly
sulfides with no chlorine-rich phases. If iron
chlorides formed at the metal/scale interface, the
phases that were highly volatile with high vapor
pressures could easily escape into the environ-
ment (see Chapter 6). In a field testing of co-
extruded tubes with Type 304 and Type 310
stainless steel claddings in a Central Electricity
Generating Board (CEGB) boiler (500 MW) that Fig. 10.12 Metal loss data for carbon steel in a simulated
had suffered aggressive furnace wall corrosion combustion atmosphere consisting of N2-10CO-
problems, Flatley et al. (Ref 23) found that the 10H2O-0.5SO2 containing different levels of HCl (0, 400, and
2000 ppm) at 500 °C. Source: Ref 22
corrosion rates were about 77 nm/h (25 mpy)
for Type 304, 55 nm/h (17 mpy) for Type 310,
and 180 to 250 nm/h (58 to 80 mpy) for
carbon steel after 15,000 h of exposure.
Corrosion rate, mpy
Metal loss, µm

10

Fig. 10.11 Metal loss data for carbon steel in a simulated Fig. 10.13 Effect of chlorine in coals on the corrosion rate of
combustion atmosphere consisting of N2-10CO- carbon steel and low-alloy steels under reducing
10H2O-0.5SO2 containing different levels of HCl (0, 400, and conditions, based on CEGB laboratory data (Ref 16). Source:
2000 ppm) at 400 °C. Source: Ref 22 Ref 19
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:02PM Plate # 0 pg 270

270 / High-Temperature Corrosion and Materials Applications

corrosion for furnace wall tubes under reducing boiler designs and were under various operating
conditions (Ref 19). In 1997, James and Pinder parameters, along with many other variables that
(Ref 24) provided a different perspective on the were involved. The correlation would be best
effect of coal chlorine on the furnace wall cor- established when the data were obtained from the
rosion. They indicated that one source of data same boiler under constant operating parameters
that contributed to the correlation of furnace wall with the chlorine content in the coal as the only
tube corrosion and the chlorine concentrations variable.
in coals was obtained from a power station with
five front-fired boilers (each with 120 MW).
10.5.2 Fireside Corrosion of Furnace Walls
Between 1967 and 1977, the chlorine content in
coals for this station in United Kingdom was under Low NOx Combustion Conditions
increased from 0.35 to 0.65%. The maximum In order to comply with the Clean Air Act
corrosion rates over this period showed an Amendments of 1990, power plants are required
increase from 170 to 550 nm/h (58.5 to to reduce NOx emissions from boilers. Nitrogen
189.2 mpy) when chlorine was increased from oxides (NOx) are produced during combustion
0.35 to 0.65%, independent of sulfur content and under oxidizing conditions, and they are unde-
load factor (Ref 24). James and Pinder summar- sirable pollutants from the boiler because they
ize the data in Fig. 10.15 (Ref 24). Also included contribute to acid rain and ozone formation. In
are data from the other plant. The data show a order to reduce NOx emissions, staged firing is
strong dependence of furnace wall corrosion on used to produce a substoichiometric combustion
the chlorine content in coal. Figure 10.16 shows (i.e., combustion with insufficient oxygen or a
the data from many plants plotted together reducing condition) in the lower furnace, which
(Ref 25), showing a significant scattering. is followed by introduction of adequate air at the
This may not be unexpected, since there were so overfire air ports at a higher elevation to complete
many plants that were probably of different the combustion process. As a result, the lower

Fig. 10.14 Corrosion rate as a function of temperature for carbon steel and austenitic stainless steels obtained from laboratory testing
as well as field testing at Eggborough Power Station. Source: Ref 19
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:02PM Plate # 0 pg 271

Chapter 10: Coal-Fired Boilers / 271

furnace is under reducing conditions and/or content in the alloy produces a more protective
shifting alternatively from oxidizing to reducing chromium oxide scale with less probability of
conditions. Under reducing conditions, sulfur forming sulfides.
from coal forms H2S instead of SO2. H2S can Similar results (Table 10.6) were reported by
cause severe sulfidation attack on carbon steel Gilroy (Ref 27) using a simulated test environ-
furnace walls. ment with much lower H2S level, showing that
Chou and Daniel (Ref 26) conducted a simu- Type 310, Type 446, alloy 671 and chromized
lated test with a relatively high level of H2S for a coatings were significantly more resistant to
variety of materials. Their results are summar- sulfidation than carbon steel. The simulated
ized in Table 10.5 (Ref 26). Carbon and low- environment used by Gilroy contained about
alloy steels were found to suffer corrosion rates 600 ppm HCl. Where chloridation attack was
in a 1 to 1.3 mm/yr (40 to 50 mpy) range. Type involved in this testing, chromium was also
304, which is significantly more resistant than effective in increasing the alloy’s resistance to
carbon and low-alloy steels, showed corrosion chloridation attack. Kung and Eckhart (Ref 28)
rates of 0.2 to 0.3 mm/yr (8 to 12 mpy). This is conducted tests in H2S-containing environments
because the corrosion products were changed for carbon steel, low-alloy steel, and several
from iron oxides/sulfides on steels to chromium stainless steels. Figure 10.17 shows the corrosion
oxides (or possibly with chromium sulfides) on rates for carbon steel and 2.25Cr-1Mo steel tes-
Type 304. Both chromium oxides and/or chro- ted at 370 °C (700 °F) in N2-5.1CO-16.7CO2-
mium sulfides exhibit much slower growth rates 4.6H2O-0.55H2 containing several levels of H2S
than iron oxides/sulfides; thus much less metal is (0.05, 0.25, 0.5, 4.8% H2S) for 1000 h. Similar
consumed and accordingly metal wastage rates tests were conducted for Types 304L and 310
are much lower. Type 309 showed a very low (Fig. 10.18). Both 304L and 310 showed sig-
corrosion rate. Alloy 671 and chromized coatings nificant reduction in corrosion rates compared
showed extremely low corrosion rates. Increases with carbon and 2.25Cr-1Mo steels. At 480 °C
in corrosion resistance are mainly related to the (900 °F) in the gas mixture with 0.05% H2S, the
chromium content in the alloy. Higher chromium corrosion rate was about 0.05 mm/yr (2 mpy) for
Corrosion rate (max), nm/h

Weighted mean chlorine content, %

Fig. 10.15 Corrosion rates of furnace walls (carbon steel) as a function of chlorine content in coal from boilers in two plants. The
equation of the line is Rc =1380 (%Cl)-208, where Rc is corrosion rate (nm/h). Source: Ref 24 from original data (Ref 25)
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:02PM Plate # 0 pg 272

272 / High-Temperature Corrosion and Materials Applications

Corrosion rate, nm/h

Weighted mean coal chlorine content

Fig. 10.16 Corrosion rates of furnace walls (carbon steel) as a function of chlorine content in coal from boilers in many plants. Source:
Ref 24 from original data (Ref 25)

Table 10.5 Corrosion of various materials in Table 10.6 Weight gain (mg/cm2) of several
N2-5CO-16CO2-10H2O-0.5H2-2H2S at 482 °C alloys after testing in N2-10CO-5CO2-10H2O-
(900 °F) for 4000 h 0.1H2S-600 ppm HCl at 400 and 500 °C
Corrosion rate, (752 and 932 °F) for 3000 h
Alloy mm/yr (mpy)
400 °C 500 °C
Carbon steel 1.04 (41) Alloy (752 °F) (932 °F)
2.25Cr-1Mo 1.32 (52)
Type 304 0.2 (8.2) Carbon steel 25.0 90.0
Type 304L 0.3 (12.0) 50Ni/50Cr, sprayed coating … 9.5
Type 309 0.04 (1.6) Aluminized (A) 8.0 …
Alloy 800 0.33 (13.0) Aluminized (B) 5.5 8.8
Alloy 671 0.005 (0.18) Fe-27Cr-6Al-2Mo … 3.8 (2000 h
Chromized carbon steel 0.006 (0.25) exposure)
Chromized carbon steel 0.008 (0.32) FAL (Fe-13Cr-4Al) 3.1 0.1
Chromized 2.25Cr-1Mo 0.007 (0.28) Ferralium (Fe-25Cr-5Ni-4Mo) 0.7 2.1
44-LN (Fe-26Cr-5Ni-1.5Mo) 0.8 0.5
Source: Ref 26 Monit (Fe-25Cr-5Ni-4Mo) 0.3 0.7
29-4-C (Fe-28Cr-4Mo) 0.2 1.1
29-4-2 (Fe-29Cr-4Mo-2Ni) 0.2 0.9
E Brite (Fe-26Cr-1Mo) 0.7 0.8
Type 310 and about 0.2 mm/yr (8 mpy) for Fecralloy (Fe-16Cr-5Al-0.35Y) 0.6 0.6
Fecralloy (Fe-19Cr-5Al-0.32Y) 0.7 0.7
Type 304L (Ref 28). Fecralloy (Fe-20Cr-5Al-0.34Y) 0.7 0.4
Dooley et al. (Ref 29) indicated that most Fecralloy A (Fe-16Cr-4.5Al-0.26Al) 0.2 0.2
boilers experienced waterwall wastage rates of GE2541 (Fe-26Cr-5Al-0.45Y) 0.6 0.4
Type 310 0.7 0.6
29 nm/h (10 mpy) or less prior to installation of Type 310Nb (0.8Nb) 0.4 0.3
low NOx burners. The wastage rates have been Alloy 800H 0.3 (1500 h 0.3 (1500 h
dramatically increased to a range of 145 to 350 exposure) exposure)
Type 446 0.5 0.3
nm/h (50 to 120 mpy) for some boilers after Alloy 671 0.6 0.5
installing low NOx burners (Ref 29). In some Chromized 2.25Cr-1Mo 0.3 0.2
Chromized carbon steel 0.2 0.3
boilers, waterwall wastage rates, which were in a
range of 49 to 87 nm/h (17 to 30 mpy), were Source: Ref 27

found to increase to a range of 96 to 226 nm/h

(33 to 78 mpy) after retrofitting with low-NOx
burners (Ref 30). It was also reported that the the vicinity of the furnace walls that suffered
corrosion scales were found to consist of iron accelerated wastage rates was observed to have
sulfides and oxides. Furthermore, the flue gas in levels of CO greater than 5% and of H2S up to
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:02PM Plate # 0 pg 273

Chapter 10: Coal-Fired Boilers / 273

500 ppm (Ref 30). Also observed in the high 1996, an inspection revealed approximately
wastage areas were carbon-rich deposits and 370 m2 (4000 ft2) of the waterwall had suffered
pyrites. Urich and Kramer (Ref 10) reported that severe wastage with maximum wastage rates
four supercritical boilers (each with 676 MWe, being in excess of 2.2 mm/yr (85 mpy). The
opposed-fired) at Gibson Generating Station boiler, a tangentially fired unit (750 MWe),
were retrofitted with low NOx burners and burned Eastern bituminous coal with 1.9 to 2.2%
overfire air. Prior to the installation of low NOx S, had a waterwall wastage problem prior to
burners, tube wastage problem was only limited installation of low NOx burners with the average
to a small area on the side walls at locations wastage rates in the range of 0.25 to 0.38 mm/yr
where H2S concentrations were measured to be (10 to 15 mpy). Yeager indicated that the fireball
as high as 1000 ppm. Since the retrofit, a marked had tendency to be closer to certain burner cor-
increase in waterwall tube wastage was observed ners, thus causing higher wastage rates. The
for all four boilers. Severe corrosion was found to furnace wall gas sampling tests revealed high
occur on the side walls, with the most severe total reduced sulfurs (TRS) and CO levels at the
corrosion occurring in the top burner elevation burner corners that suffered high wastage rates
up to the level of the overfire air ports (Ref 10). In and an oxidizing atmosphere at the burner cor-
some cases, wastage rates were on the order of ners that showed little wastage (Ref 11). Clark
285 nm/h (100 mpy). Furthermore, the corro- and Morris (Ref 18) reported that a level of H2S
sion scales in these areas were mainly iron sul- in a range of 300 to 400 ppm and of CO in a
fides. Gas sampling from the side wall in one of range of greater than 3.5% were observed in the
the boilers showed the levels of CO as high as 11 vicinity of the furnace walls with accelerated
or 12% and those of H2S as high as 1000 ppm wastage rates in U.K. boilers. The level of H2S
(Ref 10). It was reported that the service life of observed in U.K. boilers was similar to what was
the waterwall panels has been reduced from 12 to observed in U.S. boilers.
15 years to 4 years after installation of the low The areas that suffer severe waterwall wastage
NOx burners in these boilers at Gibson Station can be highly dependent on boiler design and
(Ref 31). Two supercritical units with tangen- firing configuration, among other factors. James
tially firing (582 MWe each) were also reported and Pinder (Ref 32) indicated that the side and/or
to have a more serious waterwall wastage issue rear walls were most vulnerable in a front-fired
after installation of low NOx burners with over- boiler, while in a tangentially fired boiler, the
fire air (Ref 31). wastage was invariably highest on the front wall.
Yeager (Ref 11) reported that a supercritical Bakker and Kung (Ref 33) observed that FeS
boiler, Unit No. 2, at Montour Station was deposits can significantly increase the corrosion
retrofitted with low NOx burners in 1994. In rates of carbon and low-alloy steels. They found
that FeS deposits showed no effects in reducing
environments, but accelerated the corrosion in
alternating oxidizing and reducing environments
or oxidizing environments. This is shown in

Fig. 10.17 Corrosion rates of carbon steel and 2.25Cr-1Mo Fig. 10.18 Corrosion rates of Types 304L and 310 as a
steel (T-22) as a function of H2S in the N2-5.1CO- function of H2S in the N2-5.1CO-16.7CO2-
16.7CO2-4.6H2O-0.55H2 gas mixture at 370 °C (700 °F) for 4.6H2O-0.55H2 gas mixture at 370 °C (700 °F) for 1000 h.
1000 h. Source: Ref 28 Source: Ref 28
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:02PM Plate # 0 pg 274

274 / High-Temperature Corrosion and Materials Applications

Tables 10.7 and 10.8. The authors believe that and only 90 ppm chlorine in the ash deposits
FeS under oxidizing conditions can produce (maximum wastage rate of 1.5 mm/yr or
volatile reduced sulfur species including sulfur 60 mpy) (Ref 34). Higher chlorine content in the
vapor, H2S, COS, and so forth, thus making the ash deposits might play a role in causing higher
environment more aggressive and causing ac- wastage rates (Ref 34).
celerated wastage rates. Since FeS is stable only Bakker (Ref 34) reported preliminary results
under reducing conditions and causes accelerated of the laboratory test involving the effect of the
corrosion to carbon and low-alloy steels only KCl-NaCl (equal amounts) in the ash deposits of
under oxidizing conditions, the waterwall area char, FeS, and fly ash. The test involved flue gas
that is likely to suffer accelerated wastage rates temperature of 1000 °C (1830 °F), maximum
appears to be subject to alternating reducing and ash deposit temperature of 650 °C (1200 °F),
oxidizing conditions. These alternating reducing metal tube sample temperature of 450 to 470 °C
and oxidizing conditions may exist in staged (850 to 900 °F), and the temperature of the
firing boilers where the overfire air mixes with cooling air (inside the tube sample to produce the
substoichiometric flue gas from the burner zone temperature gradient) of 320 °C (610 °F) at the
(Ref 33). In one boiler where two locations at the exit end. The metal loss was found to be about 20
waterwall made of T2 steel were monitored for to 22 µm in 100 h when the ash deposits con-
their wastage rates as well as the ash deposits, the tained no chloride and about 44 to 50 µm in
maximum wastage rate of 1.5 mm/yr (60 mpy) 100 h when 0.2% alkali chloride (KCl-NaCl in
at one location and of 0.37 mm/yr (15 mpy) at equal amounts) was added to the ash deposits.
the other location were observed after 12,633 h Lees and Whitehead (Ref 6) observed sulfides
of service (Ref 34). The flue gas measurements as well as a chlorine-containing phase at the
near the waterwall at those two locations showed sulfide/metal interface when they examined the
highly reducing conditions with about 7% CO corroded carbon steel waterwall samples using
and 500 ppm H2S during full-load operation and SEM/EDX analyses. Similar observation was
slightly oxidizing conditions during low-load also found by Lai (Ref 14). Figure 10.19 shows
conditions. However, there are significant dif- a corroded carbon steel waterwall sample from
ferences in the ash deposits between the two a boiler fired with low NOx burners (Ref 14).
locations; the area suffering high wastage rates The boiler, a subcritical unit, had been burning
had an ash deposit consisting of 90% FeS and Western Kentucky bituminous coal containing
10% fly ash, while the low wastage rate area had about 3.5% S. The waterwalls started to experi-
an ash deposit of mainly Fe3O4 and fly ash ence accelerated wastage rates after the boiler
(Ref 34). This field experience was consistent began stage firing with low NOx burners and
with the laboratory test results observed by overfire air. SEM/EDX analysis of the corrosion
Bakker and Kung (Ref 33). In the same boiler as products formed on the sample revealed iron
discussed previously in this paragraph, another sulfide phases and a chlorine-containing phase
panel testing was conducted after 17,155 h of at the sulfide/metal interface, as shown in
exposure, the FeS content in the ash deposits was Fig. 10.20 (Ref 14).
in the 40 to 90% range, but with chlorine content Coating and Weld Overlay. As more boilers
in the range of 500 to 1800 ppm. The maximum experienced accelerated wastage rates for carbon
wastage rate in the area was about 3.5 mm/yr and low-alloy steels at waterwalls in boilers that
(140 mpy), a significantly higher wastage rate had been retrofitted with low-NOx burners,
than the earlier test panel with about 90% FeS

Table 10.8 Corrosion rates of T2 in alternating

reducing(a) and oxidizing environments(b) with
Table 10.7 Corrosion rates of T2 in
various ash deposits at 427 °C (800 °F) for 400 h
N2-7CO-12CO2-6H2O-0.13H2S-0.008HCl with
various ash deposits at 427 °C (800 °F) for 600 h Ash deposit Corrosion rate, mm/yr (mpy)
No deposit 0.37–0.5 (15–20)
Ash deposit Corrosion rate, mm/yr (mpy)
30% FeS, 70% fly ash 1.0 (41)
No deposit 0.42–0.57 (17–23) 60% FeS, 40% fly ash 0.82 (33)
30% FeS, 70% fly ash 0.1–0.3 (5–12) 60% FeS, 20% carbon, 20% fly ash 1.2 (49)
60% FeS, 40% fly ash 0.1–0.17 (5–7) 90% FeS, 10% fly ash 1.2 (49)
60% FeS, 20% carbon, 20% fly ash 0.2–0.6 (9–22) 50% carbon, 50% fly ash 0.5 (20)
90% FeS, 10% fly ash 0.1–0.27 (5–11)
(a) 16 h in N2-7CO-12CO2-6H2O-0.13H2S-0.008HCl. (b) 8 h in N2-1O2-17CO2-
Source: Ref 33 0.13SO2-0.008HCl. Source: Ref 33
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:02PM Plate # 0 pg 275

Chapter 10: Coal-Fired Boilers / 275

industry looked for a viable coating or weld of Ni-42Cr-2Si alloy also suffered minor
overlay to extend the waterwall life. Bakker et al. spallation. The arc sprayed coating of Ni-
(Ref 35) reported the results of in-plant tests of 45Cr-2Al was completely spalled off from
thermal sprayed coatings (arc sprayed coatings the tube. Metallurgical evaluation of all the
and HVOF coatings), diffusion coatings (chro- coated tubes showed some sulfur penetration
mized coating and Cr/Si diffused coating), and and corrosion attack at the coating/metal
weld overlays (stainless steels 410, 309, and interface. Furthermore, a thickness loss of 20
312, and alloy 625). Waterwall test panels, which to 50% of the original thickness (typically
consisted of uncoated T2 tubes, coated and weld about 0.5 mm, or 20 mils) was observed at
overlay tubes, were tested in No. 2 boiler at the areas where no spallation occurred after
Hatfield’s Ferry Station. The boiler, a super- 12,683 h of exposure. After additional
critical unit, was retrofitted with a low NOx cell 10,072 h of exposure, most coatings had lost
burner system. Test panels were installed at 35 to 75% of the original coating thickness.
North side wall at the burner elevation. Side The authors concluded that thermal sprayed
walls typically suffered the worst corrosion coatings, which might require replacement
attack for an opposed wall firing boiler such as after 21,000 h exposure, were not adequate
this one. During testing, the boiler burned East- in providing corrosion protection for water-
ern bituminous coals with about 2.2% S with no walls under low NOx combustion conditions.
measurements or control of chlorine content.  The thickness of the coating in the as-coated
However, the Pennsylvania coals typically con- condition for both diffusion coatings was
tained about 0.05 to 0.15% Cl. The test results about 0.25 mm (10 mils). The chromized
generated by Bakker et al. (Ref 35) are sum- coating exhibited a surface layer containing
marized: about 20 to 30% Cr in the as-coated condi-
tion. The Cr/Si diffusion coating contained
 After 12,683 h of exposure, T2 waterwall about 10% Cr, 14% Si, and balance Fe. In a
tubes were found to suffer a wastage rate of
about 1.6 mm/yr (62 mpy) at the burner
elevation. The arc sprayed coating of Ni-
44Cr alloy wire suffered minor spallation.
The high-velocity oxyfuel (HVOF) coating

Fig. 10.20 Scanning electron micrograph (backscattered

electron image) showing the corrosion scales
formed on the fireside of the tube sample (shown in Fig. 10.19).
The chemical compositions of the corrosion scales at different
locations were analyzed semiquantitatively by energy dispersive
x-ray spectroscopy (EDX) analysis. The results are summarized as:
Fig. 10.19 A corroded carbon steel tube sample from the
waterwall of a boiler (subcritical unit) retrofitted 1: 66.9% Fe, 26.6% S, 1.9% Al, 2.6% Si, and minor elements
with a low NOx burner system with overfire air ports. The water- 2: 69.5% Fe, 30.0% S, and minor elements
wall tube suffered accelerated wastage after the furnace was ret- 3: 64.3% Fe, 35.2% S, and minor elements
rofitted with NOx burner system. Courtesy of Welding Services 4: 74.6% Fe, 23.6% S, and minor elements
Inc. 5: 91.0% Fe, 3.2% S, 4.4% Cl, and minor elements
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:02PM Plate # 0 pg 276

276 / High-Temperature Corrosion and Materials Applications

second test panel in the same boiler after Weld overlay using the composition that
exposure for about 17,155 h, both chromized matches the boiler tube chemistry had often been
coating and Cr/Si diffusion coating had used in the past to make temporary repair of the
completely failed, with corrosion penetrating boiler tube in the field by manual welding tech-
into the substrate steel. niques. These manually applied weld overlays
 Four weld overlays were also included in the were often referred to as “pad welds.” In mid-
second test panel; they were Type 410 (12% 1980s, an automated weld overlay technology
Cr), Type 309 (23.5% Cr), Type 312 (31% using gas metal arc welding (GMAW) process
Cr), and alloy 625 (21% Cr). The chromium was developed to apply a corrosion-resistant
content in the weld wire, as reported in the overlay alloy onto a large area of waterwall in a
paper, is indicated in the parentheses. The waste-to-energy boiler that burned municipal
overlays were reported to exhibit about 15% solid waste (Ref 36). This modern, automated
dilution. Thus, the chromium content in the weld overlay technology has been expanded
overlay was, then, about 10% for Type 410 from relatively small waste-to-energy boilers to
overlay, 20.0% for Type 309 overlay, 26.4% large coal-fired boilers (Ref 37–40). A schematic
for Type 312 overlay, and 17.9% for alloy showing an example of overlay weld beads
625 overlay. After exposure for about covering the waterwall by a GMAW overlay
17,155 h, Type 410 overlay with only about welding process is shown in Fig. 10.21.
10% Cr suffered general wastage and Figure 10.22 shows an example of the cross
sulfidation at a wastage rate equivalent to half section of a corrosion-resistant overlay on a
of that of carbon steel, thus providing no waterwall sample that was cut from an overlaid
protection. Alloy 625 with about 18% Cr waterwall. The weld overlay was applied in the
showed wastage rates of about 0.125 to field using automated overlay welding process.
0.25 mm/yr (5 to 10 mpy). Both 309 and 312 The weld overlay was characterized with con-
overlays with 20% or higher Cr showed no sistency in weld bead sequence and overlay
measurable losses in the overlay thickness. thickness of each weld bead. This automated

Fig. 10.21 Overlay weld beads applied to cover the waterwall during GMAW overlay welding process. The indicated bead sequence
number is for illustration purposes. The actual overlay welding of the waterwall may not follow the bead sequence number
indicated here. The weld bead typically progresses in a vertical down mode from top (left side of the schematic) to the bottom (right side of
the schematic). Courtesy of Welding Services Inc.
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:02PM Plate # 0 pg 277

Chapter 10: Coal-Fired Boilers / 277

GMAW overlay welding process can be carried Figure 10.25 shows the heavy Fe-Cr oxide scales
out on-site in the boiler or in-shop for overlaid formed on this overlay during service.
panels (Ref 37–40). Figure 10.23 shows a partial The most common overlay alloys that have
view of a waterwall that was overlaid in the field been applied in both subcritical and supercritical
using a nickel-base alloy. The overlay cladding units to protect the waterwalls from severe fire-
can also be accomplished using a laser technol- side corrosion attack under the low NOx com-
ogy, but the laser cladding can only be performed bustion conditions are austentic stainless steel
in-shop (Ref 41–43). Type 309 and two nickel-base alloys 625 and 622
For the weld overlay to provide adequate (Ref 39–41). Chromium is the most important
resistance to fireside corrosion and erosion/cor- alloying element in the alloy (or overlay) to allow
rosion, dilution in chemistry must be low when
overlay welding is applied onto the waterwall. In
overlay welding, the surface layer of the sub-
strate steel must be melted to establish a metal-
lurgical bond between the overlay and the
substrate steel. If too much substrate steel is
melted and mixed with the molten overlay weld
wire, the overlay chemistry will thus be sig-
nificantly diluted. This can result in lowering the
concentration of chromium in the overlay too
low to form a protective chromium oxide scale,
thus losing the protective nature of a weld over-
lay. For example, one utility company selected
Type 430 stainless steel (18% Cr) for field
application of the overlay on the waterwall of a
boiler using a submerged arc welding process
(Ref 14). The “430” overlay was found to
experience severe wastage problems due to
significant dilution in overlay chemistry
(Fig. 10.24). The overlay was found to contain
only about 11% Cr, thus making the overlay
vulnerable to oxidation/sulfidation attack.
This overlay was, thus, not capable of forming
protective chromium oxide scales, instead
forming heavy scales in Fe-Cr oxides and sulfur-
rich Fe-Cr phases at the scale/metal interface.
Fig. 10.23 Partial view of a waterwall that was overlaid with
a nickel-base alloy using automated GMAW
process on-site. Courtesy of Welding Services Inc.

Fig. 10.22 An example of an overlaid waterwall sample that

was cut from the overlaid waterwall applied by
automatic overlay welding in the field, showing consistency in
weld bead sequence and overlay thickness of each weld bead. Fig. 10.24 A Type 430 stainless steel overlaid waterwall
The waterwall shown in the figure is of a tangent tube design (no sample, showing severe overlay wastage and
membranes). Courtesy of Welding Services Inc. cracking. Courtesy of Welding Services Inc.
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:02PM Plate # 0 pg 278

278 / High-Temperature Corrosion and Materials Applications

the overlay to form chromium oxide scales to carbon and Cr-Mo steels (Ref 44), Type 309 (Ref
provide adequate protection by preventing the 14), alloy 625 (Ref 45), and alloy 22 (or 622)
formation of sulfides and/or iron-rich oxides. The (Ref 46) at temperatures of interest for water-
specification range of chromium is 23.0 to 25.0% walls are shown in Table 10.9. There exists some
for Type 309 weld wire (ER309), 20 to 23% for thermal expansion coefficient mismatch between
alloy 625 weld wire (ERNiCrMo-3), and 20.0 to the 309 overlay and T-11 substrate steel. Never-
22.5% for alloy 622 (ERNiCrMo-10). The auto- theless, no cracks have been found to occur at the
mated GMAW process typically aims to achieve interface (or the fusion boundary) between the
a dilution in the overlay of approximately 10% or 309 overlay and the substrate steel waterwall
less. Figure 10.26 shows the chronology of the tube due to this thermal expansion mismatch.
total square feet of waterwall overlays from early Figure 10.27 shows typical interface structure
1990s to 2001 by a leading applicator of weld between the 309 overlay and the T11 substrate
overlay cladding in terms of the year of appli- steel waterwall tube in a supercritical unit after 10
cation for three major overlay alloys (Ref 14). years of service. The stresses developed due to
Beginning in 2000 and 2001, the industry this thermal expansion coefficient mismatch
began switching from alloy 625 to alloy 622 in appear to be too small to initiate cracking. Fur-
waterwall overlay cladding applications in coal- thermore, both the 309 overlay and the substrate
fired boilers, particularly in supercritical units. carbon and Cr-Mo steels exhibit excellent duc-
This switch was related to circumferential tility and toughness to prevent any development
grooving of the alloy 625 overlay on the water- of cracks at the interface.
walls of several supercritical units. (The issue of The concern of high stresses that can lead
circumferential grooving and cracking is dis- to cracking at the overlay/substrate interface
cussed in section 10.5.3.) during operation in the 309 overlay because of its
In fireside corrosion and erosion/corrosion, it higher thermal expansion coefficients prompted
is desirable to select an overlay alloy with ade- Coleman and Gandy (Ref 47) to examine Type
quate chromium to form protective chromium 312 (a duplex stainless steel) as an alternative
oxide scales. Another factor that is viewed by stainless steel overlay alloy. Type 312 exhibits
some boiler operators to be important is the thermal expansion coefficients that are much
thermal expansion coefficient mismatch between lower than those of Type 309, and the alloy
the overlay and the substrate ferritic steel. contains higher chromium (about 30% Cr in
Nickel-base alloys, such as alloys 625 and 622, the weld wire), making it a potentially good
generally have thermal expansion coefficients candidate overlay alloy. However, this alloy is a
that match quite well with carbon and Cr-Mo duplex stainless steel, which consists of austenite
steels. Austenitic stainless steels, such as Type and ferrite phases, and can suffer severe embrit-
309, exhibit thermal expansion coefficients that tlement when exposed to a temperature range of
are higher than those of carbon and Cr-Mo steels.
Mean coefficients of thermal expansion for
35,000
Type 309
30,000
Alloy 625
Total overlay area, ft2

25,000 Alloy 622

20,000

15,000

10,000

5,000
0
1993 1994 1995 1996 1997 1998 1999 2000 2001
Year overlay performed

Fig. 10.26 Chronology of the total area of the weld overlay

applied to the waterwalls of coal-fired boilers
(both subcritical and critical units) from early 1990s to 2001 as a
function of the year of application for three major overlay alloys—
Fig. 10.25 Optical micrograph showing nonprotective Fe-Cr Type 309, alloy 625, and alloy 622—by a weld overlay applica-
oxide scales formed on the “430SS” weld overlay. tion company. The data covers only field application overlays.
Courtesy of Welding Services Inc. 1.0 ft2 = 0.093 m2. Source: Ref 40
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:03PM Plate # 0 pg 279

Chapter 10: Coal-Fired Boilers / 279

Table 10.9 Mean coefficient of thermal expansion

Coefficient of thermal expansion, 10−6 in/in/°F
Temperature Carbon steel(a) 1.25Cr-0.5Mo(a) 309 SS(b) Alloy 625(c) Alloy 22(d)
70–600 °F 7.2 … 9.3 7.4 7.0
70–800 °F 7.6 7.7 9.5 7.6 7.4
70–1000 °F 8.0 7.8 9.7 7.8 7.7
(a) EPRI data Ref 44. (b) Weld overlay data Ref 14. (c) Wrought alloy data Ref 45. (d) Wrought alloy data Ref 46. Alloy 22 and alloy 622 meet UNS N06022 specification.
Unit conversions: 1.8 ×10−6 in/in/°F = 10−6/K = 10−6 m/m/°C; 70 °F = 21 °C; 600 °F =316 °C, 800 °F = 427 °C, 1000 °F = 538 °C

Overlay

Substrate

0.010 in.

Fig. 10.27 Optical micrograph showing typical interface

structure in longitudinal orientation between the
309 overlay and T11 substrate steel (waterwall tube) after 10 years
of service in a supercritical boiler, revealing no cracking at the
interface. Substrate steel was etched with nital to show the inter-
face. Courtesy of Welding Services Inc.

approximately 371 to 593 °C (700 to 1100 °F)

due to formation of ordered α‫ ׳‬phases. This
embrittlement phenomenon is commonly refer-
red to as 475 °C (885 °F). This embrittling
temperature range is within the waterwall metal
temperature range. Furthermore, Type 312 weld
overlay is susceptible to weld solidification
cracking when a weld overlay having a low
dilution is attempted (Ref 48). There have been
several boilers where Type 312 overlay was
applied to protect the waterwalls against fireside Fig. 10.28 Macro cross sections cut from a Type 312 overlay
corrosion. waterwall panel sample obtained from a super-
critical unit showing the overlay after 6.5 years of service in
In one supercritical unit where Type 312 (a) the transverse cross section and (b) the longitudinal cross
waterwall overlay was inspected recently after section. The macro samples were etched in nital to reveal the
about 6.5 years of service, and an overlay sample overlay. The scale is in inches. 1.0 in. = 25.4 mm. Source: Ref 49
was obtained from the waterwall for metallurgi-
cal examination. Surprisingly, the overlay was
found to exhibit excellent condition, showing metallographic cross section of the 312 overlay
no evidence of cracking or embrittlement, or after about 6.5 years of service, showing essen-
circumferential grooving/cracking during ser- tially no corrosion and no cracking.
vice (Ref 49). Typical macro cross sections of Bonnington and Brennan (Ref 43) reported
the overlay waterwall tubes after about 6.5 installation of laser-clad 312 overlay panels in
years of service in a supercritical unit are shown supercritical boilers at the Mirant Mid-Atlantic
in Fig. 10.28. Figure 10.29 shows typical station in 2000. Mirant also tested a laser-clad
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:03PM Plate # 0 pg 280

280 / High-Temperature Corrosion and Materials Applications

Type 446SS overlay in a high-wastage area of a steel) weld metal buildup to the same thickness of
supercritical unit in 1991. The test results showed about 0.090 in. (2.3 mm) on the bare steel tube
excellent performance of Type 446 overlay in the (Ref 50). For example, for a bare tube with an
boiler (Ref 43). However, the cost of producing a outer metal skin temperature of 482 °C (900 °F),
laser-clad panel with Type 446 was considered to the application of an overlay of T2 of about
be prohibitive at that time (Ref 43). Type 446 is a 90 mils (2.3 mm) thick to the bare tube increases
Fe-25Cr ferritic stainless steel, which is also the skin temperature by approximately 33 °C
susceptible to the 475 °C (885 °F) embrittle- (60 °F). In comparison with T2 overlay, Type
ment. 309 overlay would increase the skin temperature
Ferritic steels, such as carbon and low-alloy by about 22 °C (40 °F) higher than that of T2
steels, typically exhibit higher thermal con- overlay. Similarly, alloy 622 would increase the
ductivities than austenitic stainless steels and skin temperature by about 28 °C (50 °F) higher
nickel-base alloys. As a result, the outer skin than that of T2 overlay, and alloy 625 by about
metal temperature of the waterwall tube will be 36 °C (65 °F) higher than that of T2 overlay.
increased somewhat during service when a weld Type 312 overlay is expected to show lower outer
overlay cladding of an austenitic stainless steel skin metal temperature than Type 309 (or alloys
(e.g., Type 309) or a nickel-base alloy (e.g., 625 and 622) because of its duplex micro-
alloys 625 and 622) is applied onto a carbon steel structure containing a mixture of austenite and
or low-alloy steel. Blough (Ref 50) performed ferrite. Since the application of a weld overlay
heat transfer calculation using the thermal con- using Type 309, Type 312, or nickel-base alloy
ductivities of wrought alloys. His results are has essentially eliminated the general wastage
summarized in Fig. 10.30, which compares outer problem, it becomes desirable to apply a thinner
skin overlay temperatures for Type 309, alloy overlay to lower the outer skin overlay metal
625, and alloy 622 overlays with T2 (0.5 Mo temperature.
The overlays of Type 309, alloy 625, alloy
622, and Type 312 have so far accumulated
various service times with 309 and 625 overlays
staying in service the longest (about 9 to 10
years) in some boilers. These overlays have
essentially eliminated the fireside wastage pro-
blems caused by staged combustion with low
NOx burners and overfire air (Ref 35, 37–43).
However, several supercritical boilers have been
found to experience circumferential grooving
and cracking problem. The circumferential
grooving/cracking problem is discussed in sec-
tion 10.5.3.

09 in.

Fig. 10.29 Optical micrograph showing the transverse cross Fig. 10.30 Calculated overlay metal temperatures for T2
section of Type 312 overlay on the waterwall in a bare tube with 0.090 in weld metal build-up of
supercritical unit after about 6.5 years of service, revealing T2 in comparison with 0.090 in overlay of Type 309, alloy 625,
essentially no corrosion and no cracking and alloy 622. Source: Ref 50
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:03PM Plate # 0 pg 281

Chapter 10: Coal-Fired Boilers / 281

10.5.3 Circumferential Grooving of Conemaugh Station (Ref 52), Morgantown, and

Furnace Waterwalls Chalk Point Generating Stations (Ref 53).
In order to determine the root cause of cir-
Fireside corrosion of the furnace waterwalls cumferential grooving, it is best to examine the
particularly in supercritical units has caused initial stage of the groove formation. This is
carbon and low-alloy steel waterwall tubes to illustrated by several examples for carbon and
suffer circumferential (transverse) grooves that low-alloy steel tubes. Figure 10.32 shows the
resemble an alligator hide or an elephant hide. An initiation of several circumferential grooves on a
example of a waterwall tube exhibiting this type T22 tube in a supercritical boiler. Scanning
of circumferential grooving appearance is shown electron microscopy with energy dispersive
in Fig. 10.31. This type of waterwall corrosion x-ray spectroscopy (SEM/EDX) was used to
has been discussed by Wright (Ref 51) and determine semiquantitatively the chemical com-
French (Ref 15). Wright (Ref 51) indicated that positions of the corrosion products inside the
circumferential grooving has been found to occur groove. One of the grooves shown in Fig. 10.32
more frequently in supercritical boilers than was analyzed by SEM/EDX with the results
subcritical boilers. He also indicated that the shown in Fig. 10.33. The corrosion products
waterwall areas that suffered this type of attack were mainly enriched in iron and sulfur particu-
received the highest heat flux with superimposed larly in light grayish “stringers” or “channels”
thermal stresses. French (Ref 15) suggested that in the interior region containing higher sulfur.
the circumferential grooving was caused by a Areas 2 and 3 are located in the light grayish
corrosion-fatigue mechanism. He indicated that channels (or stringers), as shown in Fig. 10.33.
the corrosion products in the circumferential Similar observation was made for a T2 tube
groove contained sulfur and the ash deposits in another supercritical boiler, as shown in
on the waterwall tube surface contained carbon Fig. 10.34, revealing largely iron sulfides in the
(Ref 15). He thus suggested that the reducing corrosion products. Again, there were light
atmospheres were an essential part of the grayish phases in a form of stringers or “chan-
grooving-corrosion mechanism. The circumfer- nels” containing higher sulfur (Area 2, 3, and 4 in
ential grooves in the later stage exhibit a thermal- Fig. 10.34). These stringers or channels, which
fatigue crack appearance. Some supercritical tended to form in the interior region, extended
boilers historically had circumferential grooving/ from the surface of the corrosion products and
cracking problems even prior to the installation penetrated along with the overall corrosion
of weld overlay claddings, such as the ones in penetration. The one in the center tends to be
much wider and longer. These stringers, which
were found to be more enriched in sulfur than the
surrounding phases, appear to be gas paths

0.0010 in.

Fig. 10.32 Optical micrograph showing the initial develop-

Fig. 10.31 An example of the circumferential grooving ment of several circumferential grooves on a T22
encountered in the T2 steel waterwall. Courtesty waterwall tube in a supercritical unit. Courtesy of Welding Ser-
of Welding Services Inc. vices Inc.
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:03PM Plate # 0 pg 282

282 / High-Temperature Corrosion and Materials Applications

formed during the corrosion process. In some was depleted in chromium (Fig. 10.37b), but
cases, these stringers were quite complex and do enriched in sulfur, as shown in Fig. 10.37(c), and
not appear to be the cracks which are subse- also in Ni (the Ni dot map was not shown). Thus,
quently filled with another corrosion product or
ash deposit. Figure 10.35 shows the case of a
very dense group of stringers or channels formed
in the circumferential groove on T11 tube in a
supercritical boiler.
Several supercritical boilers have been found
to experience circumferential grooving for Type
309 overlay (Ref 39) and alloy 625 overlays (Ref
39, 54) on waterwalls. Figure 10.36 shows a
longitudinal cross section of an alloy 625 overlay
waterwall tube at the crown location after 1 year
of service in a supercritical boiler. Two tiny
circumferential grooves are visible. One of the
grooves was analyzed by SEM/EDX, showing
that the alloy 625 overlay had suffered essentially
sulfidation penetration. Figure 10.37 shows the
corrosion products to be essentially chromium
sulfides. There was a centerline channel which

(a) 140 µm

60 µm

Fig. 10.33 Scanning electron micrograph (backscattered

electron image) showing one of the circumfer- 60 µm
ential grooves formed on a T22 waterwall tube (2.25Cr-1Mo) as (b)
shown in Fig. 10.32. Semiquantative energy dispersive x-ray
spectroscopy (EDX) analysis in terms of weight percent at different Fig. 10.34 (a) Scanning electron micrograph (backscattered
locations of the corrosion products inside the groove is summar- electron image) showing typical morphology of
ized below. Courtesy of Welding Services Inc. the circumferential groove formed on a T2 tube (0.5Cr-0.5Mo) in a
supercritical boiler. The results (in wt%) obtained from the semi-
1: 83.9% Fe, 7.1% S, 4.6% Cr, 2.1% Mo, 0.5% Al, 0.7% Si, 0.9% quantitative analysis using EDX on the corrosion products are
Mn, and trace elements summarized as:
2: 72.2% Fe, 21.2% S, 3.6% Cr, 1.3% Mo, 0.5% Al, 0.9% Mn, and
trace elements 1: 86.8% Fe, 10.5% S, 1.4% Cr, and minor elements
3: 73.3% Fe, 22.8% S, 2.2% Cr, 0.6% Mn, and trace elements 2: 79.7% Fe, 18.2% S, 0.64% Cr, and minor elements
4: 80.6% Fe, 9.8% S, 5.1% Cr, 2.5% Mo, 0.9% Mn, 0.5% Al, and 3: 69.0% Fe, 30.4% S, and minor elements
trace elements 4: 67.3% Fe, 32.0% S, and minor elements
5: 87.6% Fe, 1.3% S, 6.7% Cr, 2.6% Mo, 1.2% Al, 0.8% Si, and 5: 93.7% Fe, 0.5% S, 2.5% Cr, 1.9% Mo, 1.2% Si, and minor
trace elements elements
6: 87.0% Fe, 0.8% S, 7.5% Cr, 2.8% Mo, 0.8% Si, 0.7% Mn, and 6: 91.8% Fe, 3.7% S, 1.3% Cr, 1.2% Mo, and minor elements
trace element (b) Detail of regions 1 to 4
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:03PM Plate # 0 pg 283

Chapter 10: Coal-Fired Boilers / 283

this centerline channel appeared to be nickel products were enriched in chromium and iron
sulfide. Similar observation of the corrosion with some sulfur. Sulfur was particularly high
products in the circumferential groove formed in within the phases inside the light grayish strin-
alloy 625 overlay cladding was also reported by gers that were clustered around the inner portion
Luer et al. (Ref 54). Figure 10.38 shows a cir- of the corrosion penetration attack. However, the
cumferential groove that formed on alloy 625 levels of sulfur were significantly lower than
overlay on a waterwall tube after 2 years of ser- those observed in alloy 625 overlay. It is thus
vice in another supercritical boiler. The corrosion believed that corrosion penetration was essen-
products were found to consist of essentially tially combined oxidation and sulfidation in the
chromium sulfides. The alloy 625 was etched to 309 overlay.
show the dendritic microstructure of the weld From the previous discussion of the initial
overlay. Figure 10.39 shows the formation of development of the circumferential grooving in
the initial preferential sulfidation penetration at both alloy 625 and 309 overlays, it is concluded
a high magnification, revealing in detail the that the circumferential grooves were initially
corrosion phases, which consisted of essentially developed by way of preferential sulfidation
chromium sulfides. penetration. For nickel-base alloy 625 overlay,
Type 309 overlay on the waterwall of a super- sulfidation was the major mode of preferential
critical unit was found to suffer circumferential attack, while for Type 309 overlay (an austenitic
grooving after 10 years of service. Figure 10.40 stainless steel), sulfidation along with oxidation
shows circumferential grooves at the early stage. was involved in the preferential attack. As the
The circumferential groove consisted of several preferential attack continued, the circumferential
branches of preferential penetrations. These groove, which penetrated farther into the overlay
several branches of preferential penetrations with increasing stresses at the penetration tip,
tended to converge into a groovelike attack. The eventually developed into a crack at a later stage.
morphology of the circumferential groove (at the This is illustrated in Fig. 10.42.
early stage of development) formed in Type 309 In another power station, Type 309 overlay has
overlay (an austenitic stainless steel) is some- provided protection for the waterwalls of two
what different from that formed in nickel- supercritical units for 8 years in one unit and
base alloy 625. The corrosion products in one of 10 years for the other unit. The 309 overlay in
the branches of preferential penetrations, as one unit was inspected using nondestructive
shown in Fig. 10.40, were analyzed by SEM/ testing including dye-penetrant testing in 2006
EDX with the results summarized in Fig. 10.41. after 8 years of service. The inspection revealed
The SEM/EDX results indicate the corrosion

0.5 mm

0.0010 in. Fig. 10.36 Optical micrograph showing the initial develop-
ment of two tiny circumferential grooves formed
on the alloy 625 overlay at the crown bead of a waterwall tube in a
Fig. 10.35 Optical micrograph showing a circumferential supercritical unit after 1 year of service. The metallographic
groove containing a complex “network” of mount was in the longitudinal cross section. One of the grooves
stringers formed on T11 tubes (1.25Cr-0.5Mo) in a supercritical formed on the overlay (left side of the micrograph) was analyzed
boiler. Courtesy of Welding Services Inc. by SEM/EDX with the results shown in Fig. 10.37. Source: Ref 40
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:03PM Plate # 0 pg 284

284 / High-Temperature Corrosion and Materials Applications

no sign of tube wastage and circumferential elevated temperatures can produce preferential
cracking (Ref 14). Typical weld overlay is shown sulfidation penetration on wrought alloys. Gutt-
in Fig. 10.43. The unit was equipped with low mann et al. (Ref 56) showed that several alloys
NOx burners and overfire air, burning bituminous were suffering preferential sulfidation penetra-
coal containing 3.0 to 3.5% sulfur. Type 309 tion under tensile strains in a sulfidizing envir-
waterwall overlay, which has accumulated so far onment. Their tests were conducted at 600 °C
for 10 years of service, was to be inspected dur- (1112 °F) in CO-32H2-4CO2-0.2H2S (inlet
ing the next maintenance shutdown. gas mixture) with equilibrium sulfur and
In laboratory testing of various alloys in oxygen potentials at the test temperature being
sulfidizing environments, several authors (Ref about 10−11 and 10−28 bar (or atm), respectively.
55–58) have found that applied tensile stresses Figure 10.44 shows alloy 45TM (Ni-27Cr-23Fe-
during exposure to a sulfidizing environment at 2.8Si) after exposure to the environment at

(a)
0.0010 in.

(b) 50 µm (c) 50 µm

Fig. 10.37 (a) Optical micrograph of a circumferential groove in Fig. 10.36. (b) Chromium x-ray dot map for the corrosion products
inside the groove. (c) Sulfur x-ray dot map for the corrosion products inside the groove. Source: Ref 40
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:03PM Plate # 0 pg 285

Chapter 10: Coal-Fired Boilers / 285

temperature for 2000 h under (a) no external For Fe-Cr-Ni alloys, such as austenitic stain-
strain and (b) a 2% strain. The specimen under no less steels, the stress-enhanced sulfidation pene-
external strain during exposure showed essen- tration attack takes a different morphology.
tially uniform corrosion attack, while the one Guttmann et al. (Ref 56) found that alloy HR3C
under a 2% strain showed preferential penetra- (Fe-25Cr-20Ni-0.5Nb-0.2N), which is a wrought
tion attack. The morphology of the preferential alloy, suffered stress-enhanced preferential
penetration attack is similar to the circumfer- attack along grain boundaries instead of a fin-
ential groove observed in nickel-base alloy 625 gerlike protrusion, as was observed in alloy
overlay on the waterwall. The preferential 45TM (a nickel-base alloy). Figure 10.47 shows
penetration was found to be essentially oxida- alloy HR3C after testing in CO-32H2-4CO2-
tion/sulfidation. Also observed in this pre- 0.2H2S (inlet gas mixture) at 600 °C (1112 °F)
ferential penetration attack were channels under (a) no external stresses showing general
that were similar to those observed in cir- uniform corrosion attack after 2100 h and (b)
cumferential groove. The channel and stringers a 1.3% strain after 250 h of exposure showing
contain phases with much lighter color, similar to grain-boundary attack. For Type 309 overlay
what was observed in circumferential grooves (approximately Fe-21Cr-12Ni) on the water-
formed in Cr-Mo steel waterwall tubes or alloy wall, preferential sulfidation penetration attack
625 overlay in few supercritical boilers. Coze followed individual branches of continuous
et al. (Ref 59) observed similar phenomena in penetrations during the initial stage of develop-
testing Fe-12Cr-3Al-3Ti experimental alloy in ment in the circumferential grooving, as shown
H2-34.3H2O-18.5CO2-3.8CH4-7.9CO-1.3H2S in Fig. 10.40. These individual branches of sul-
at 600 °C (1112 °F) 615 h under no external fidation penetrations were most likely to be
stresses and external stresses, as shown in interdendritic boundaries.
Fig. 10.45. The internal penetration attack under From the discussions in this section, it can be
stresses (upper micrograph of Fig. 10.45) is very concluded that circumferential grooving is a
similar to circumferential grooves formed in preferential sulfidation penetration attack under
Type 430 overlay on the waterwall tube in a tensile stresses. Tensile stresses can be varying.
boiler, as shown in Fig. 10.46. However, thermal cycling is not a prerequisite for

Fig. 10.38 A circumferential goove formed on alloy 625 overlay applied to the waterwall of a supercritical boiler after 2 years
of service. The overlay was etched to show the dendritic microstructure of the alloy 625 overlay. Courtesy of
Welding Services
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:03PM Plate # 0 pg 286

286 / High-Temperature Corrosion and Materials Applications

initiating circumferential grooving, even though supercritical units was mentioned to be 482
the circumferential groove at a later stage can to 510 °C (900 to 950 °F) (Ref 10) and 482 to
develop into a crack. 538 °C (900 to 1000 °F) (Ref 11), it is most
Wright (Ref 51) indicated that circumferential likely that the overheating at certain locations
grooving typically occurs in the areas of the of the waterwall is due to flame impingement.
waterwall receiving the highest heat flux and is Since alloy 625 overlay has been found to suffer
apparently a result of superimposed thermal circumferential grooving in several supercritical
stress. It is believed that the preferential sulfida- units, the outer overlay metal temperature of the
tion penetration that initiates the circumferential overlay can be estimated based on the aging
grooving at certain locations of the waterwall is characteristics of alloy 625.
due to overheating under sulfidizing or alternat- Alloy 625 is known to exhibit age hardening
ing sulfidizing/oxidizing conditions. Although a at intermediate temperatures of 538 to 760 °C
temperature range for waterwall tubes has been (1000 to 1400 °F). The alloy age hardens by
mentioned to be 400 to 500 °C (752 to 932 °F) forming γ″ (Ni3Nb) precipitates in this tempera-
(Ref 9) and the design temperature of some ture range. At 538 to 593 °C (1000 to 1100 °F)

(b) (c)

Fig. 10.39 Scanning electron micrograph (a) showing a very early stage of the circumferential grooving formed in alloy 625 overlay
on a waterwall tube in the same supercritical boiler, as shown in Fig. 10.38. EDX spectra (b) and (c) show the alloying
elements associated with the “light” phases (b) and grayish phases (c), indicating the corrosion phases were essentially chromium sulfides.
The sample surface was plated prior to mounting to protect the corrosion products for SEM examination. Courtesy of Welding Services Inc.
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:03PM Plate # 0 pg 287

Chapter 10: Coal-Fired Boilers / 287

for 1 to 2 years of aging, the alloy forms fine,

coherent precipitates causing significant hard-
ness increases particularly at 593 °C (1100 °F).
With aging at 650 °C (1200 °F), initial pre-
cipitates formed in the alloy are fine, coherent γ″,
but later become incoherent long, platelet Ni3Nb
precipitates, as shown in Fig. 10.48 (Ref 61).
The microhardness profile was determined on
the alloy 625 overlay that was found to show
initiation of circumferential grooving after about
1 year in service (shown in Fig. 10.36). The
results, which are shown in Fig. 10.49, indicate
0.0010 in. age hardening to 35 to 45 HRC at the outermost
surface layer of the overlay (Ref 40). The interior
Fig. 10.40 Optical micrograph showing circumferential part of the overlay exhibited a hardness level (23
grooves that formed in Type 309 overlay applied
to the waterwall of a supercritical boiler after 10 years of service. to 25 HRC) that essentially corresponded to the
Courtesy of Welding Services Inc. hardness of the as-overlay alloy 625. Based on
the hardness data, the metal temperature of the
overlay surface layer is believed to have in-
creased to above 538 °C (1000 °F), most likely
to about 593 °C (1100 °F), the temperature that
causes significant hardening for alloy 625.
In another supercritical boiler where alloy 625
overlay on the waterwall suffered circumferential

50 µm

Fig. 10.41 Scanning electron micrograph (backscattered

electron image) showing the corrosion products
in a circumferential groove formed in Type 309 overlay on the
waterwall of a supercritical boiler after 10 years of service. The
results of the semiquantative EDX analysis of the corrosion pro-
ducts at different locations are as summarized below. Courtesy of
Welding Services Inc.
1: 62.2% Cr, 29.2% Fe, 4.6% Mn, 1.7% Ni, 1.4% S, and trace
elements
2: 40.7% Cr, 39.5% Fe, 4.2% Mn, 4.1% Ni, 8.4% S, 2.4% Si, and
trace elements
3: 28.0% Cr, 59.8% Fe, 2.6% Mn, 4.5% Ni, 2.9% S, 1.5% Si, and Fig. 10.42 Some well-developed circumferential grooves,
trace elements which developed into cracks at a later stage,
4: 18.1% Cr, 62.4% Fe, 1.5% Mn, 4.7% Ni, 11.6% S, 1.0% Si, and were observed in the 309 overlay on the waterwall after 10 years
trace elements of service. There was no evidence of cracking developed at the
5: 46.8% Cr, 40.2% Fe, 2.3% Mn, 2.2% Ni, 5.8% S, 2.2% Si, and interface between the 309 overlay and the T11 substrate even
trace elements with some thermal expansion coefficient mismatch. The substrate
6: 45.8% Cr, 43.2% Fe, 3.2% Mn, 2.3% Ni, 3.5% S, 1.5% Si, and steel was etched with nital to reveal the fusion boundary. Original
trace elements magnification: 25×
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:03PM Plate # 0 pg 288

288 / High-Temperature Corrosion and Materials Applications

grooving after 2 years of operation, the overlay chromium carbides along grain boundaries, the
surface layer was also found to exhibit age overlay metal temperature was likely to be below
hardening to about 40 to 42 HRC. 538 °C (1000 °F). Microhardness measurements
Unlike alloy 625, Type 309 does not exhibit made across the overlay that corresponds to
age-hardening characteristics. As a result, the the microstructure containing carbides, as shown
possible overheated overlay surface temperature in Fig. 10.50, indicated about 250 HV (about 100
could not be estimated using microhardness data. HRB) as opposed to 200 HV (about 90 HRB)
Chromium carbides have been found to pre- for the overlay that corresponds to the
cipitate along grain boundaries in a Type 309 microstructure with no carbides, as shown in
waterwall overlay after 10 years of service in a Fig. 10.51. The 309 overlay is typically applied
supercritical unit, as shown in Fig. 10.50. Pre- using a low-carbon Type 309 weld wire
cipitation of these chromium carbides may be an (ER309LSi). Thus, age-hardening response is
indication that the overlay was overheated to generally minimal. Accordingly, the develop-
above 538 °C (1000 °F). In performing metal- ment of preferential grooving and later cracking
lurgical evaluation of a 309 overlay sample that was not related to the aging behavior of the 309
was removed from the same boiler in a waterwall weld overlay, which is expected to exhibit good
area (probably from a different location as the ductility, after exposure to 538 to 593 °C (1000
area discussed in Fig. 10.50) after 7 years of to 1100 °F) for 10 years.
service, the microstructure of the overlay showed
no chromium carbide precipitates formed along
grain boundaries (Fig. 10.51) (Ref 37). The
temperature of the steam in this area was esti-
mated by a plant engineer to be 371 to 427 °C
(700 to 800 °F) (Ref 14). Without the evidence of

Fig. 10.43 Type 309 overlay on the waterwall in another Fig. 10.44 Alloy 45TM (Ni-27Cr-23Fe-2.8Si) after exposure
supercritical unit equipped with low NOx bur- to the test gas of CO-32H2-4CO2-0.2H2S (inlet
ners and overfire air after service for 8 years, showing no sign of gas mixture) at 600 °C (1112 °F) for 2000 h under (a) no external
metal wastage or circumferential cracking. The boiler reportedly strain and (b) 2% strain. The equilibrium sulfur and oxygen
burned bituminous coal containing 3 to 3.5% S. Courtesy of potentials for the test environment at the test temperature were
Welding Services Inc. about 10−11 and 10−28 bar (or atm), respectively. Source: Ref 56
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:03PM Plate # 0 pg 289

Chapter 10: Coal-Fired Boilers / 289

Tensile ductility after long-term aging may circumferential grooving or cracking. The
not be an important factor in the development microhardness profile across the 312 overlay at
of circumferential cracking. An example is illu- the crown bead from two samples is shown in
strated by Type 312 waterwall overlay that Fig. 10.53. Both samples showed some hard-
has been in service in a supercritical unit for ening at the region close to the overlay surface
about 6.5 years with no evidence of circumfer- with one sample showing significantly more
ential cracking being revealed by dye-penetrant hardening (close to 50 HRC). Type 312 overlay
testing during a recent inspection (Ref 14). is a duplex stainless steel containing a mixture
Samples obtained from the overlaid waterwall of austenite and ferrite phases. When exposed to
for metallurgical examination also showed a temperature range of approximately 343 to
similar results. Figure 10.28(b) shows a long- 593 °C (650 to 1100 °F), the ferrite phase is
itudinal macro cross section of the overlay sam- subject to ordering reaction forming α′ ordered
ple, revealing no circumferential grooving or precipitates, resulting in significant hardening
cracking. Figure 10.52 shows typical overlay and causing the alloy to become brittle. This
surface condition in a longitudinal metallo- phenomenon is commonly referred to as the
graphic cross section, again, revealing no 475 °C (885 °F) embrittlement. Even under this
brittle condition, the 312 overlay showed no
evidence of circumferential grooving or crack-
ing. The overlay was found to contain about 29%
Cr. The corrosion products observed on the

Fig. 10.45 Fe-12Cr-3Al-3Ti experimental alloy after testing

in H2-34.3H2O-18.5CO2-3.8CH4-7.9CO-1.3H2S
at 600 °C (1112 °F) for 615 h under no external stresses (lower
micrograph) and external stresses (upper micrograph). The
external strain applied during test was unknown. Source: Ref 59

Fig. 10.47 Alloy HR3C after testing in CO-32H2-4CO2-

0.2H2S (inlet gas mixture) at 600 °C (1112 °F)
Fig. 10.46 Circumferential grooves formed in Type 430 weld under (a) no external stresses showing general uniform corrosion
overlay on the waterwall tube in a coal-fired attack after 2100 h and (b) a 1.3% strain after 250 h of exposure
boiler. Source: Ref 60 showing grain-boundary attack. Source: Ref 56
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:03PM Plate # 0 pg 290

290 / High-Temperature Corrosion and Materials Applications

overlay were found to be chromium-rich oxides, to 1%) the alloys could develop preferential
thus providing adequate protection against sulfidation penetration after longer exposure
preferential sulfidation penetrations, a precursor times, such as a year or longer). For circumfer-
of circumferential grooving and cracking. ential grooving to develop, the axial stresses
In the laboratory tests discussed above, would have to be large enough to develop tensile
applications of about 1 to 2% tensile strains on strains of approximately 0.5 to 1%. Axial stresses
the test specimens were capable of causing pre- generated on the waterwall tubes may come from
ferential sulfidation penetration at about 593 °C the dead load of the waterwall panels, thermal
(1100 °F) for less than a couple of thousands of stresses, pressures of water/steam inside the tube,
hours of test duration for various alloys including and bending stresses due to localized flame
nickel-base alloys and Fe-Cr-Ni alloys. It would impingement (overheating) due to constraints
appear that under lower tensile strains (e.g., 0.5 from the surrounding cooler waterwall areas as

Fig. 10.48 Aging behavior of alloy 625 (wrought alloy samples) at 650, 760, and 870 °C (1200, 1400, and 1600 °F) for 16,000 h in
terms of microstructure, losses in impact toughness and elongation, and increases in strength. Source: Ref 61
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:03PM Plate # 0 pg 291

Chapter 10: Coal-Fired Boilers / 291

well as buckstays. (Buckstays are structural scales in a sulfidizing environment, thus allowing
shapes or trusses that encircle and restrain the preferential sulfidation attack to take place. Three
movement of the furnace waterwalls caused by conditions are required to be met for developing
fluctuation in furnace pressure, or transient circumferential grooving for the weld over-
internal or external loads.) The combination of laid waterwall: (a) sulfidizing environment,
these stresses may be adequate to lead to the (b) adequate tensile strains (e.g., 0.5 to 1%), and
development of preferential sulfidation penetra- (c) overheating of tube outer skin metal to
tions and circumferential grooving when the approximately 593 °C (1100 °F).
overlay outer layer is heated to approximately For widely used weld overlays of Type 309,
593 °C (1100 °F) under localized sulfidizing alloy 625 and alloy 622, the chromium content
conditions. The development of preferential of the weld overlay at the crown location (i.e.,
sulfidation penetrations are believed to result the location subjected to the highest heat flux)
from local breakdown of the protective oxide was typically 20%. This level of chromium in the
Rockwell C, HRC

Fig. 10.49 Microhardness profile measured using Vickers hardness tester with a 500 g load as a function of the distance from the
overlay surface for alloy 625 weld overlay on the waterwall of a supercritical boiler after 1 year of operation when
circumferential grooves, as shown in Fig. 10.36, were initiated. Vickers hardness values (HV) were converted to Rockwell C (HRC) values.
Hardening of the overlay surface layer (within 0.5 mm, or 20 mils) is believed to result from age hardening of alloy 625 due to formation of
fine, coherent γ″ (Ni3Nb) precipitates when heated to probably 593 °C (1100 °F). Source: Ref 40

Fig. 10.50 Optical micrograph showing typical microstructure of the 309 overlay on the waterwall that suffered circumferential
grooving (Fig. 10.40) in a supercritical boiler after 10 years of service. Original magnification: 200×
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:03PM Plate # 0 pg 292

292 / High-Temperature Corrosion and Materials Applications

weld overlay may not be adequate when the over- and cracking. Alloy 52 overlay, which was ap-
laid waterwall is subject to the aforementioned plied to the waterwall of a supercritical unit, has
three conditions. To avoid the local breakdown now been in service for 3 years with no reported
of the chromium oxide scales under these con- degradation problems (Ref 14). Paul et al.
ditions, an overlay containing a higher level of (Ref 62) indicated some field testing of alloy 33
chromium is required to ensure the adequate (UNS R20033: Fe-33Cr-32Ni) overlay test
chromium content to maintain the protective panels in two supercritical boilers, showing
chromium oxide scales over a prolong service no cracking after slightly less than 2 years of
duration. Alloy 52 (AWS ERNiCrFe-7 or UNS testing. Both boilers were tangential firing with
N06052) with 30Cr, 9Fe and balance Ni provides low NOx and overfire air. Alloy 33 test panels
a good candidate overlay alloy with potentially were located in the overfire air region in both
improved resistance to circumferential grooving boilers. In both cases, alloy 622 overlay panels
were installed at the same time in the same area.
After slightly less than 2 years of exposure, both
alloy 33 and alloy 622 overlays tube samples
with similar exposure times were removed for
metallurgical evaluation, showing no evidence of
circumferential cracking for both overlay alloys
(Ref 62).

10.5.4 Sootblowing
One important by-product of the combustion
of coal is ash. During combustion, some of the
mineral constituents and compounds can be in a
molten or plastic state. These ash constituents
can then deposit on the heat-absorbing surfaces
0.10 mm causing slagging and fouling. If ash reaches the
heat-absorbing surface at a temperature near its
Fig. 10.51 Microstructure of Type 309 overlay on the softening temperature, the resulting deposits are
waterwall after 7 years of service from the same
boiler, but likely from different area in the boiler. No chromium likely to be porous and can be removed by
carbides along grain boundaries were detected in this area. The sootblowing. Slagging is the deposition of mol-
overlay showed no circumferential grooving or cracking. Source: ten, partially fused deposits on the furnace walls
Ref 37
and the upper furnace radiant superheaters
exposed to radiant heat (Ref 1). Fouling is the
deposition of more loosely bonded deposits on
the heat-absorbing surfaces in the convection
path, such as superheater and reheater, that

0.0010 in.
Fig. 10.53 Microhardness profile across the overlay from the
Fig. 10.52 Optical micrograph showing typical longitudinal overlay surface measured using Vickers hardness
cross section at the crown bead of the 312 tester with a 500 g load. Vickers hardness values (HV) are con-
waterwall overlay after about 6.5 years of service in a supercritical verted to Rockwell C (HRC) values. Data were obtained from two
unit. The corrosion scales were chromium-rich oxides as identi- different overlay samples (Series 1 and 2). 1 in. = 25.4 mm.
fied by SEM/EDX analysis. Courtesy of Welding Services Inc. Source: Ref 49
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:04PM Plate # 0 pg 293

Chapter 10: Coal-Fired Boilers / 293

are not exposed to radiant heat (Ref 1). Soot Nickel-base alloy overlays, such as alloy 622
blowers using steam are generally adequate overlay, that form protective chromium oxide
for removing the ash deposits on the fouled scales will have similar resistance to sootblower
tube surfaces, while waterlances or water can- erosion-corrosion. More discussion on erosion/
nons using thermal shock by room-temperature corrosion is covered in Chapter 8: Erosion and
water may be needed to clean the slagged tube Erosion-Corrosion.
surfaces.
The damage that causes the tube surface due to
steam sootblowers is generally referred to as
sootblower erosion. Under normal operations,
steam sootblowing is in the erosion/corrosion
(i.e., erosion/oxidation) regime. An example is
given here to illustrate this phenomenon. A
905 MW(e) supercritical unit had experienced
severe waterwall tube wastage problem with its
SA213 T11 tubes (1.25 in. outside diameter ×
0.220 in. minimum wall thickness) subjected to
steam sootblowing by wall blowers. These
waterwall tubes had to be replaced every 18
months because of severe tube wastage. The
boiler, which was not equipped with low NOx
burners, burned the Illinois basin coal with about
2.5% S. The corrosion products were most likely
iron oxides with possibly some iron sulfides
(if sulfidation also took place). No analysis was
done on the corrosion products formed on T11
waterwall tube. However, when the soot blower
affected area was weld overlaid with Type 309
using automatic GMAW process, the wastage
problem was essentially eliminated. A Type 309
overlay tube sample was removed after 7 year of
service for metallurgical evaluation. During this
7 year period, the boiler had burned the Illinois
basin coal (2.5% S) for about 4 years and an
Eastern Appalachian low sulfur coal for 3 years
with low NOx burners and separated overfired air
during these 3 years. The overlay showed no
evidence of metal wastage and erosion/corrosion
attack by steam sootblowing (Fig. 10.54 and
10.55).
The reason for significant life improvement
offered by the 309 overlay is the erosion/corro-
sion resistance of the 309 overlay. The 309
overlay forms chromium oxide scales, which are
much thinner and grow significantly more slowly
than iron oxides that form on T11 (1.25Cr-0.5Mo
steel). Steam impingement on the steel removes
iron oxide scales, which are thick, nonprotective,
and fast growing. The iron oxide scales reformed
repeatedly after they were removed by steam
impingement, thus resulting in accelerated Fig. 10.54 Field-applied Type 309 overlay on the waterwall
tube after 7 years of service in a 905 MW(e)
wastage. Once the tube was protected by the 309 supercritical boiler: (a) close-up view of the crown and side beads
overlay, metal wastage due to erosion/corrosion of the overlay, and (b) cross section of the crown bead of the
overlay showing both the overlay surface and the overlay/sub-
was significantly reduced due to the formation strate interface (as-polished condition, original magnification:
of thin, protective chromium oxide scales. 25×). Courtesy of Welding Services Inc.
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:04PM Plate # 0 pg 294

294 / High-Temperature Corrosion and Materials Applications

10.5.5 Deslagging by Waterlances and from the field was significantly lower than the
Water Cannons EPRI’s 288 °C (550 °F) drop in temperature
measured by a waterlance simulator used to test
One of the important characteristics of ash is the waterwall panels when sprayed without a slag
its ash fusion temperatures. If the ash fusion (no thermal insulating by a slag). The boiler at
temperatures are low enough, the ash deposit Plant Miller (Alabama Power), which burned
reaches its melting point (due to the ash’s thermal Powder River Basin (PRB) coal, used water-
insulating properties) and the molten slag runs lances for continuous waterwall cleaning with a
down the furnace wall surface (Ref 1). When 3 hour cycle time. This resulted in reduced
solidified, this slag is tightly bonded and is dif- waterwall tube life due to quench cracking (Ref
ficult to remove. This slag may require water- 64). Ray et al. (Ref 65) reported that a boiler
lances or water cannons to create thermal shock suffered waterwall tube damage (a waterwall
for the removal of this slag deposit. The furnace blowout) after 2½ years of “unrestricted”
walls are likely locations for developing this waterlance usage.
slagging problem when coal with low ash fusion The morphology of thermal fatigue cracking
temperatures is combusted. on carbon steel waterwall tube as a result of
When room-temperature water is used for water spraying from water cannons is illustrated
deslagging the waterwall through either water- in Fig. 10.56. Surface appearance of thermal
lances or water cannons, the waterwall tube is fatigue cracks on a waterwall tube due to
subjected to thermal shock. Repeated use of water spraying from waterlances is shown in
waterlances or water cannons for deslagging can Fig. 10.57. Figure 10.58 shows typical mor-
cause thermal fatigue cracking of the waterwall phology of the circumferential cracks in a long-
tubes. Thermal fatigue cracking of waterwall itudinal cross section from the sample shown in
tubes due to deslagging by waterlances or water Fig. 10.57. Figure 10.59 shows an SEM of a
cannons has been reported by Kessler (Ref 63), circumferential crack along with the energy dis-
Carlisle et al. (Ref 64), Ray et al. (Ref 65), and persive x-ray spectroscopy (EDX) analysis on
Blinka (Ref 66). Kessler (Ref 63) reported field the corrosion product inside the crack. The EDX
measurement data that indicated a rapid tem- analysis showed the corrosion product inside the
perature drop from approximately 379 to 317 °C crack to be essentially iron oxides.
(715 to 603 °F) for the waterwall tube right after One boiler operator decided to repair the
water spraying from water cannons. He indicated cracked waterwall tubes, which were made of
in Ref 63, that this temperature drop measured SA210 A1 steel, due to water spraying from
waterlances by applying alloy 622 weld overlay
cladding in the field after the cracks were ground

0.1 mm

0.5 mm
Fig. 10.55 Optical micrograph showing a very thin oxide
scale on the surface of the 309 overlay on a Fig. 10.56 Optical micrograph showing typical morphology
waterwall tube (T11) after 7 years of service subjected to steam of circumferential thermal fatige cracking on a
sootblowing in a supercritical boiler, revealing no evidence of carbon steel waterwall tube due to water spraying from water
erosion/corrosion damage or cracking. The 309 overlay was not cannons. The tube OD (fireside) is on the left side of the micro-
etched showing white portion of the micrograph with a magnifi- graph and the ID (water/steam side) is on the right side. Courtesy of
cation marker at the bottom of the micrograph. Source: Ref 40 Welding Services Inc.
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:04PM Plate # 0 pg 295

Chapter 10: Coal-Fired Boilers / 295

off from the tubes. The repaired waterwall area superheater, typically a horizontal heat exchan-
was more than 420 m2 (4500 ft2). After 4 years ger located above the economizer, and then
of service, dye penetrant testing (PT) of the through a secondary or finishing platen super-
whole overlaid waterwall area showed no evi- heater (i.e., tubes are in a flat arrangement and
dence of cracking. Figure 10.60 shows typical hung in the furnace roof) with a number of pla-
overlay area after PT testing. tens in parallel. In some boilers, this finishing
superheater is heated by radiant heat from the
furnace combustion. The steam leaving the sec-
10.5.6 Superheater and Reheater Corrosion
ondary or finishing superheater passes through a
Steam from either the steam drum in a sub- high-pressure steam turbine. After expanding
critical unit or from the furnace wall in a super- through the high-pressure steam turbine, the
critical unit first passes through a primary steam is returned to the boiler to be reheated in a
reheater. Steam from the reheater then passes
through an intermediate-pressure turbine, fol-
lowed by passing through a low-pressure turbine.
Typical superheated steam temperature in most
utility boilers in the United States is about
538 °C (1000 °F). The metal temperatures of
superheaters and reheaters may be up to 650 °C
(1200 °F).
Superheater and reheater tubes suffer oxida-
tion attack at lower temperatures. Oxidation
attack generally results in lower wastage rates.
When flue gas stream is entrained with fly
ash, the tubes can also suffer erosion/corrosion
attack. The wastage rates under erosion/corro-
sion conditions can be much higher. When
the metal temperature is approaching 650 °C
(1200 °F) or higher, superheater and reheater
tubes can suffer accelerated wastage rates due
to coal ash corrosion. The accelerated wastage
rates due to coal ash corrosion are the result
Fig. 10.57 Appearance of thermal fatigue cracks occurred
of molten salt (sulfate) corrosion. French
on a carbon steel waterwall tube (viewed from 12
o’clock crown position) due to water spraying from waterlances. (Ref 15) suggests that the corrosion follows a
Source: Ref 40 “hockey stick” type behavior with two dis-
tinctive corrosion regimes: low metal wastage
rate by oxidation and accelerated wastage
rates by molten salt corrosion, as shown in
Fig. 10.61. The molten salt corrosion behavior
exhibits a “bell-shaped” curve with respect to
temperature for austenitic stainless steels. The
rate increases with temperature to a maximum,
then decreases with increasing temperature.
The accelerated corrosion associated with this
bell-shaped curve is related to the formation
of molten alkali metal-iron-trisulfate [(Na,K)3Fe
(SO4)3] (Ref 67, 68). Variation of the sodium-to-
potassium ratio greatly affects the melting point
of the complex sulfate in an ash deposit (Ref 68).
Figure 10.62 illustrates a wide variation of
0.5 mm melting points of mixtures of sodium iron tri-
sulfate and potassium iron trisulfate (Ref 68).
Fig. 10.58 Optical micrograph showing circumferential
Corrosion rate was also affected by the sodium-
thermal fatigue cracks that developed on a car-
bon steel waterwall tube (shown in Fig. 10.57) due to water to-potassium ratio in the complex sulfate (Ref
spraying from waterlances. Source: Ref 40 68). Nelson and Cain (Ref 69) conducted a series
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:04PM Plate # 0 pg 296

296 / High-Temperature Corrosion and Materials Applications

10,000 Fe

9,000

8,000

7,000

6,000
Counts
EDS-1 corrosion products
5,000

4,000

3,000

2,000
O Fe
1,000 Fe
C Mn Si S Cr Mn
0
0.000 1.000 2.000 3.000 4.000 5.000 6.000 7.000 8.000 9.000 10.000 11.000 12.000 13.000 14.000 15.000
keV
120 µm
(a) (b)

Fig. 10.59 (a) Scanning electron micrograph (backscattered electron image) showing a circumferential thermal fatigue crack (from
the sample shown in Fig. 10.57) along with (b) an EDX spectrum showing the corrosion product inside the crack to be
essentially iron oxides. Source: Ref 40

Fig. 10.61 Tube wastage as a function of tube metal tem-

perature after 125,000 h, with low wastage rate
by oxidation and accelerated wastage rates by molten salt cor-
Fig. 10.60 Alloy 622 overlay (dye penetrant tested) after 4
rosion. Source: Ref 15
years of service involving the use of waterlances
for deslagging. The overlay was applied onto the carbon steel
waterwall after thermal fatigue cracks caused by waterlances
were ground off. The dye penetrant testing showed no cracking. The typical corroded superheater or reheater
Courtesy of Welding Services Inc.
tube is characterized by two wastage flats at
about the 2 o’clock and 10 o’clock positions, as
shown schematically in Fig. 10.66 (Ref 13). At
of laboratory tests with flowing synthetic flue gas these two locations, where the ash was thin and
(N2-15CO2-3.6O2-0.25SO2) over a mixture of the heat flux was high, a molten sulfate layer
potassium sulfate, sodium sulfate, and iron oxide formed, resulting in severe corrosion attack
(molecular ratio 1.5:1.5:1.0) that covered the test (Ref 13). On the other hand, the front face of the
coupons. Samples were exposed at different tem- tube, with a heavy ash deposit providing suffi-
peratures for 5 days. Results are summarized in cient insulation to keep the metal surface tem-
Fig. 10.63 (Ref 69). The corrosion rate was perature below the melting point of the sulfate,
greatly enhanced when the sulfate was molten. suffered significantly less corrosion attack
Corrosion rate also depends on SO2 concentra- (Ref 13).
tion in the flue gas and Na2SO4 + K2SO4 con- Figure 10.67 shows the cross section of a
centration in the ash deposit, as illustrated in Type 304H reheater tube after service for about
Fig. 10.64 and 10.65 (Ref 70). For alkali sulfates, 4 years with the reheated steam at about 538 °C
only the acid-soluble ones are of concern in coal- (1000 °F) suffering coal-ash corrosion attack in a
ash corrosion (Ref 71). subcritical boiler. The locations that suffered the
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:04PM Plate # 0 pg 297

Chapter 10: Coal-Fired Boilers / 297

worst coal-ash corrosion attack were at

approximately the 2 and 10 o’clock positions
(when the flue gas impinges at 12 o’clock posi-
tion). The boiler used staged combustion and
overfire air for NOx reduction control. The sur-
face appearance of the maximum wastage area is
shown in Fig. 10.68. SEM/EDX analysis was
performed on the corrosion scale formed on the
maximum wastage area with the results being
summarized in Fig. 10.69 (Ref 14). The analysis
shows sulfidation attack along with oxidation.

Fig. 10.64 Effect of SO2 content in flue gas on the corrosion

of several superheater/reheater materials ex-
posed to synthetic ash containing 5 wt% (Na2SO4 + K2SO4) at
650 °C (1200 °F). Source: Ref 70

Fig. 10.62 Melting point curve for the K3Fe(SO4)3-Na3Fe

(SO4)3 system. Source: Ref 68

Fig. 10.65 Effect of Na2SO4 + K2SO4 content in synthetic ash

on the corrosion of several superheater/reheater
materials at 650 °C (1200 °F) in flue gas containing 0.25% SO2.
Source: Ref 70

Fig. 10.63 Results of laboratory tests with flowing synthetic

flue gas (N2-15CO2-3.6O2-0.25SO2) over a
synthetic coal ash (K2SO4, Na2SO4, and Fe2O3 with a mol- Fig. 10.66 Typical wastage feature of a corroded superheater
ecular ratio of 1.5:1.5:1.0) that covered the test coupons. Source: and reheater tube from a coal-fired boiler. Source:
Ref 69 Ref 13
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:04PM Plate # 0 pg 298

298 / High-Temperature Corrosion and Materials Applications

It was further observed that carburization surface that did not suffer coal-ash corrosion
occurred at the locations (i.e., about 2 and 10 attack, such as at the 3 or 9 o’clock position,
o’clock positions) where coal-ash corrosion took showed no carburization (Fig. 10.71). It is
place, as shown in Fig. 10.70. However, the believed that carburization was not the cause that
resulted in coal ash corrosion because of the
formation of chromium carbides that reduced the
chromium level in the metal matrix. On the
contrary, coal-ash corrosion (or sulfidation)
caused the metal surface to form sulfides along
with unprotective chromium oxide scales, thus
resulting in carburization when the localized
environment was under reducing conditions with
the presence of CO and/or unburned carbon soot.
A superheater tube made of Type 304 suffer-
ing coal-ash corrosion attack is shown in
Fig. 10.72. Similar to the reheater tube discussed
earlier (Fig. 10.67), the worst attack occurred at
the 2 and 10 o’clock positions (when the flue gas
was impinging at the 12 o’clock position). The
morphology of corrosion attack at the worst
corrosion attack area is shown in Fig. 10.73,
showing general wastage and some internal
attack in the matrix as well as intergranular at-
tack. Figure 10.74 shows a scanning electron
micrograph (backscattered electron image),
showing both general corrosion scales and
Fig. 10.67 Cross section of a Type 304H reheater tube internal attack. The EDX analysis of the corro-
showing two wastage flats with the maximum sion products indicated sulfides (marked as 1)
wastage. Note the wastage flats on both sides of the tube surface
where the flue gas impinging at the 90° location (i.e., facing the and internal oxide phases (marked as 2). Area 1
ruler in the photo). Courtesy of Welding Services Inc. area was found to contain primarily 45.5%

Fig. 10.68 Surface appearance of one of the wastage flats with the maximum wastage of the 304H reheater tube shown in Fig. 10.67.
Courtesy of Welding Services Inc.
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:04PM Plate # 0 pg 299

Chapter 10: Coal-Fired Boilers / 299

Cr, 33.5% Fe, 9.2% Ni, and 8.6% S (in wt%) again removed by subsequent impinging fly-ash
and area 2 primarily 56.8% Fe, 27.5% Cr, and particles. This repeated oxide removal and
11.7% Ni. reformation process caused the wastage rate to be
In the same boiler, Type 304H reheater tubes increased due to erosion-enhanced corrosion
suffered maximum wastage rates at about 30° reaction. (Detailed discussion of erosion/corro-
locations on either side of the direct flue gas sion is covered in Chapter 8) Some internal attack
impingement position, as shown in Fig. 10.75. was observed underneath the flyash deposits.
The wastage pattern in this case is believed to Internal attack was essentially in form of oxides
be caused by fly-ash induced erosion/corrosion along grain boundaries (marked 2, 3, and 4).
instead of coal-ash corrosion. Figure 10.76 Trace of chlorine was detected in the phase
shows the cross section of the maximum wastage marked by 4. In adjacent area, which was still on
area where fly-ash deposits were in contact with the flat wasted face (but near the area shown in
the metal surface. The EDX analysis of the fly- Fig. 10.76) relatively thin oxide scales were
ash deposits, which are marked as 1, showed observed. These oxide scales appeared to be
primarily 46.4% Si, 21.6% Al, and 20.7% Fe. fragmented and fractured possibly by impinging
Impinging fly-ash particles removed oxide scales flyash (Fig. 10.77). The EDX analysis showed
from the metal surface that prompted the fresh presence of sulfur in these oxides. These oxides
metal surface to form oxides again, which were were likely the subsequently “reformed” oxides
that formed after preceding oxides were removed
by impinging fly-ash particles.
At the position where the flue gas directly
impinging on the tube in the same 304H reheater
tube sample, significant amount of ash deposits
was observed, as shown in Fig. 10.78. At this
location, the tube wall wastage was significantly
less. The EDX analysis was performed in the
areas in (a) the top portion of the deposits
(Fig. 10.79), (b) midsection of the deposits/
corrosion products (Fig. 10.80), and (c) the
deposits/corrosion products near and at the
interface (Fig. 10.81). The top layer was found to
be essentially ash deposits enriched in aluminum,
silicon, iron, sulfur and arsenic, and calcium.
The midsection also consisted of ash deposits
enriched primarily in silicon, arsenic and iron.
Corrosion products formed on the tube were
found to consist of essentially Cr-Fe-rich sulfides
and Fe-Ni-Cr-rich sulfides. Some internal sul-
21 µm fides were also observed.
Many investigators have been using labora-
Fig. 10.69 Scanning electron micrograph (backscattered
tory simulation tests involving synthetic ash
electron image) showing the corrosion products
formed on the maximum wastage area of Type 304H reheater (typically Na2SO4, K2SO4, and Fe2O3) to cover
shown in Fig. 10.67. Semiquantative EDX analysis shows the the test specimens with a flowing synthetic flue
compositions (wt%) at different locations as indicated below.
Courtesy of Welding Services Inc. gas (typically N2-O2-CO2-H2O-SO2) in a test
retort to rank alloy performance for resistance to
1: 34.4% Fe, 49.1% Cr, 5.6% Ni, 4.9% S, 3.7% Mn, 1.3% Si, and
trace elements coal-ash corrosion. Figure 10.82 shows the
2: 34.7% Fe, 17.6% Cr, 22.8% Ni, 22.3% S, 1.6% Mn, 0.6% Si, results of laboratory coal-ash corrosion tests
and trace elements involving various austenitic stainless steels,
3: 32.1% Fe, 45.5% Cr, 6.8% Ni, 7.6% S, 4.5% Mn, 0.9% Si, 0.7%
Na, and trace elements Fe-Ni-Cr alloy 800H, and “50Ni-50Cr” alloy
4: 28.9% Fe, 49.0% Cr, 6.3% Ni, 6.6% S, 5.0% Mn, and trace 671 (Ref 72). The best performer was the alloy
elements
5: 33.7% Fe, 17.6% Cr, 22.4% Ni, 23.1% S, and trace elements
with the highest chromium content (alloy 671).
6: 32.4% Fe, 49.1% Cr, 5.0% Ni, 5.7% S, 6.1% Mn, and trace Resistance to coal-ash corrosion has been found
elements to increase with increasing chromium content in
7: 37.7% Fe, 48.3% Cr, 1.8% Ni, 6.8% S, and trace elements
8: 47.2% Fe, 40.4% Cr, 4.0% Ni, 4.8% S, 2.4% Mn, and trace the alloy. Castello et al. (Ref 73) summarized the
elements coal-ash corrosion data generated by various
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:04PM Plate # 0 pg 300

300 / High-Temperature Corrosion and Materials Applications

investigators to show the beneficial effect of chromized T91. During the 16,000 h exposure,
chromium in resisting coal-ash corrosion in the boiler burned coal containing 3.08 to 3.26%
Fig. 10.83. S, 8 to 11% ash. Field data showed significant
Kihara et al. (Ref 78) also presented their scattering with alloys containing about 20% Cr.
laboratory simulated coal-ash corrosion tests as a Alloys containing about 25% Cr and higher were
function of chromium contents in the alloys in found to perform significantly better with less
Fig. 10.84. Synthetic ash used in laboratory data scattering.
testing consisted of 2.5, 5, and 20% Na2SO4 + At TVA’s Gallatin Station, previous 304 and
K2SO4 (equal amounts) with Fe2O3, Al2O3 and 321 stainless steel reheaters in Unit No. 2 (a
SiO2 (equal parts). Tests were performed at 600, subcritical boiler) typically required replacement
650, and 700 °C (1112, 1200, and 1292 °F) for after 7 years of service (Ref 79). The reheater
times up to 300 h in a synthetic flue gas con- made of higher chromium-containing alloy
taining 0.05, 0.1, 0.25, and 1.0% SO2. Plant HR3C (25Cr-20Ni-0.5Nb-0.25N) had been in
exposure data were also included in the figure. service for about 45,004 h (approximately 5
Plant exposure data were obtained using corro- years) without causing any forced outage due to
sion probes inserted into the operating boiler at corrosion-induced tube wastage problem
Tennessee Valley Authority’s Gallatin Station (Ref 79). In order to develop coal-ash corrosion
Unit No. 2. Alloys tested included 17-14 CuMo data for superheaters and reheaters at higher
(17Cr-14Ni austenitic stainless steel), Type steam temperatures in ultra-supercritical boilers,
347, 25Cr-20Ni-Nb, 21Cr-11Ni-Si-Ce austenitic a field exposure test program was initiated that
stainless steel, 20Cr-18Ni-Si-Al austenitic stain- involved using tube sleeves to increase metal
less steel, alloy 800, 30Cr-45Ni-2Mo, and temperatures. Samples were exposed at 521 to

Fig. 10.70 Optical micrograph showing Type 304H reheater (the same one shown in Fig. 10.67) that suffered carburization
on the surface where severe coal-ash corrosion took place (about 2 and 10 o’clock positions). Courtesy of Welding
Services Inc.
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:04PM Plate # 0 pg 301

Chapter 10: Coal-Fired Boilers / 301

Fig. 10.71 Type 304H reheater (the same one shown in Fig 10.67) at 3 or 9 o’clock positions, where no severe coal-ash corrosion
attack occurred, showing no carburization attack. Courtesy of Welding Services Inc.

685 °C (970 to 1265 °F) for about 45,004 h.

Blough et al. (Ref 79) summarized the test
results on the tube shields (higher metal tem-
peratures) in Fig. 10.85, showing metal wastage
rates as a function of metal temperature. Two
highest chromium-containing alloys, HR3C
(25Cr), and CR30A (30 Cr), performed the
best. The data in Fig. 10.85 show what appeared
to be low wastage rates for chromized T22 at
585 °C (1085 °F). In fact, the chromizing
layer with about 0.33 mm (0.013 in.) in thick-
ness was found to be completely corroded at
certain locations during the test duration, and
the corrosion had penetrated into the substrate
T22 steel (Ref 79). Thus, the authors concluded
that chromizing layer can only offer temporary
protection to the substrate steel (Ref 79).
A long-term plant exposure program to
determine suitable superheater and reheater
alloys against coal-ash corrosion for applications
in ultra-supercritical boilers was conducted by Fig. 10.72 Cross section of a superheater tube made of Type
Babcock & Wilcox (B&W), the U.S. Department 304SS suffering coal-ash corrosion attack. Note
the wastage flats on both sides of the tube surface where the flue
of Energy (DOE), and the Ohio Coal Develop- gas impinged at the 90° location (i.e., facing the ruler in the
ment Office (OCDO) under the “Coal Ash photo). Courtesy of Welding Services Inc.
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:04PM Plate # 0 pg 302

302 / High-Temperature Corrosion and Materials Applications

Corrosion Resistant Materials Testing Program.” overlay tube was produced by spiral overlay
The results of the test program were published welding method using gas metal arc welding
by McDonald (Ref 80) and McDonald and (GMAW) process.
Robitz (Ref 81). Plant exposure tests were con- The Central Electricity Generating Board
ducted in a B&W 120 MW(e) cyclone-fired (CEGB) in the United Kingdom tested and tried
boiler (a subcritical unit burning a 3 to 3.5% S co-extruded tubing with a high chromium alloy
Ohio coal) at Reliant Energy’s plant in as a cladding for superheaters and reheaters in
Niles, Ohio. Test sections, which were cooled by 1970s and 1980s. Latham et al. (Ref 82) reported
600 °F/315 psi reheat steam, were located within test trials of Type 310/Esshete 1250 co-extruded
the superheater bank. Figure 10.86 shows cross tubes and 671/800H co-extruded tubes in CEGB
sections of SAVE 25 (Sumitomo’s 25Cr-20Ni boilers. One test trial was conducted in a
steel) tube specimens tested at three different 550 MW(e) boiler burning coal with 0.45% Cl.
temperatures, showing increased wastage rates The test tubes were welded to the hottest section
with increasing exposure temperature. The of the reheater with maximum tube metal tem-
results summarizing the performance of various perature of approximately 650 °C (1200 °F), and
alloys are presented in Fig. 10.87. The alloys that flue gas temperature of about 1150 °C (2100 °F).
exhibit acceptable wastage rates were found to After 15,400 h of exposure, 671/800H co-
be alloy 671 cladding and alloy 72 weld overlay. extruded tubes along with Type 316 and
Both alloys 671 and 72 were nickel alloys with 347 tubes were removed for evaluation. In
very high chromium contents (47% Cr for former addition, Type 310/E1250 and 671/800H co-
and 44% Cr for the latter). They were both used as extruded tubes were removed after 18,000 h of
a cladding. Alloy 671 clad tubing was produced exposure, and 671/800H co-extruded tubes were
by coextrusion method, while alloy 72 weld removed after 34,000 h of exposure. The results

Fig. 10.73 Optical micrograph showing the morphology of the corrosion attack on the 304H superheater tube at the severely wasted
area. Courtesy of Welding Services Inc.
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:04PM Plate # 0 pg 303

Chapter 10: Coal-Fired Boilers / 303

in terms of wastage rates (nm/h or mpy) are Type 310/E1250 co-extruded tubes. After 17,500
summarized in Table 10.10. and 29,000 h of exposure, Type 310/E1250
In another test conducted in a 275 MW(e) co-extruded tubes showed a wastage rate of only
boiler burning coal with less than 0.15% Cl, 6 nm/h (1.9 mpy) versus a wastage rate of about
a complete final reheater was retubed with

30 µm
11 µm

Fig. 10.76 Scanning electron micrograph (backscattered

electron image) showing fly-ash deposits
Fig. 10.74 Scanning electron micrograph showing the cor-
(46.4 Si-21.6Al-20.7Fe) (marked 1) on the surface of Type 304H
rosion products (general corrosion scales and
internal corrosion phases) on Type 304H superheater at the reheater (Fig. 10.75) that suffered the maximum wastage
severely wasted area. The EDX analysis showed that the corrosion at location 30° away from the direct flue gas impingement point.
product (area 1) was Cr-Fe-rich sulfide (45.5Cr-33.5Fe-9.2Ni- The 304H also suffered intergranular corrosion attack (areas 2–4).
8.6S), and the internal corrosion phase (area 2) was Fe-Cr-rich Courtesy of Welding Services Inc.
oxide (56.8Fe-27.5Cr-11.7Ni). Area 3 is the alloy matrix. Courtesy
of Welding Services Inc.

50 µm

Fig. 10.77 Scanning electron micrograph showing the

Fig. 10.75 Cross section of a Type 304H reheater tube oxide scales formed on the nearby location of
suffering fly-ash erosion/corrosion damage. the one shown in Fig. 10.76 on the severely wasted area for Type
Courtesy of Welding Services Inc. 304H reheater tube. Courtesy of Welding Services Inc.
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:04PM Plate # 0 pg 304

304 / High-Temperature Corrosion and Materials Applications

50 nm/h (16 mpy) for previous Type 347 tubes.

A final superheater was retubed with Type 310/
E1250 co-extruded tubes for a 500 MW(e) boiler
burning a 0.3% Cl coal, Type 310/E1250 co-
extruded tubes showed maximum corrosion rate
of about 50 to 60 nm/h (16 to 19 mpy) versus
maximum corrosion rates of about 100 nm/h
(32 mpy) for the original Type 316 tubes.
Experiences in the United Kingdom with co-
extruded tubes for superheaters and reheaters
were summarized in a 1987 paper by Flatley et al.
(Ref 83). Flatley et al. (Ref 83) indicated that
boilers in several major power stations retubed
the hottest sections of superheaters and reheaters
with Type 310/E1250 co-extruded tubes with
excellent results. In one installation, however,
some of the hotter leading tubes were less satis-
factory in performance advantage of the 310/
11 µm
E1250 co-extruded tubes over the original Type
316 tubes with wastage rates being about 50 nm/
h (16 mpy) for Type 310 cladding versus about Fig. 10.79 Scanning electron micrograph showing only the
top portion of the ash deposits/corrosion scale
70 to 100 nm/h (22 to 32 mpy) for the original shown in Fig. 10.78 at higher magnification. The gray and white
Type 316 tubes. For alloy 671/800H co-extruded phases were analyzed by EDX with the results (in wt%) summar-
ized below. Courtesy of Welding Services Inc.
tubes installed in a complete reheater in a
500 MW(e) boiler burning high chlorine (0.5%) 1: 30.2% Fe, 21.0% As, 14.1% Si, 11.0% Al, 8.7% Ca, 9.7% S,
2.3% P, and residual elements
coal, the alloy 671 cladding showed a wastage 2: 80.7% Fe, 11.1% Si, 2.1% Al, 1.5% As, and residual elements
rate of less than 0.1 mm/yr (4 mpy) after
60,000 h of operation as opposed to a wastage

11 µm

Fig. 10.80 Scanning electron micrograph showing the mid-

30 µm dle section of the ash deposits/corrosion scale
shown in Fig. 10.78 at higher magnification. The dark gray (3) and
light gray (4) phases were analyzed by EDX with the results (in wt
Fig. 10.78 Scanning electron micrograph showing ash %) summarized below. Courtesy of Welding Services Inc.
deposits that formed at the 12 o’clock position
where the flue gas made a direct impingement on the Type 304H 3: 32.8% Si, 26.1% As, 22.6% Fe, 9.6% Al, 2.4% P, 1.5% S, and
reheater tube (Fig. 10.75). This location showed significantly less residual elements
tube wall wastage. Courtesy of Welding Services Inc. 4: 93.8% Fe, 2.4% Cr, and residual elements
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:04PM Plate # 0 pg 305

Chapter 10: Coal-Fired Boilers / 305

rate of about 2 mm/yr (79 mpy) for the original gas tungsten arc welding method was developed
Type 347 tubes. for manufacture of weld overlay bimetallic tub-
Some of the performance experiences for this ing for superheaters and reheaters (Ref 86).
671/800H co-extruded tubing have been reported Two spiral weld overlay tubes tested in the “Coal
by Fahrmann and Smith (Ref 84) and Kiser and
Orsini (Ref 85). Fahrmann and Smith (Ref 84)
reported the results of the metallurgical evalua-
tion of two superheater tubes and one reheater
tube after 18 years of service in a utility coal-fired
boiler. The alloy 671 cladding in the reheater tube
was found to exhibit very little corrosion attack, -
with the cladding thickness very close to that of
the original as-fabricated tube. For two super-
heater tubes, the alloy 671 cladding generally
showed good conditions except at few locations
where pitting attack was observed. In one
case, the pitting attack was almost penetrated
to the entire cladding thickness. Even at this
location, the corrosion rate would still be very
low (less than 0.1 mm/yr, or 4 mpy, assuming
the original cladding thickness being 2 mm, or
80 mils).
In late 1990s, a spiral overlay welding tech-
nology involving a tandem gas metal arc and

Fig. 10.82 Results of laboratory tests conducted in synthetic

flue gas (80N2-15CO2-4O2-1SO2, saturated with
H2O) with synthetic ash (37.5 mol% Na2SO4, 37.5 mol% K2SO4,
and 25 mol% Fe2O3) covering the samples. Exposure time was
50 h. Source: Ref 72

11 µm

Fig. 10.81 Scanning electron micrograph showing the bot-

tom portion of the corrosion products shown in
Fig. 10.78 at higher magnification. The phases (5–8) were ana- Fig. 10.83 Metal loss as a function of chromium contents in
lyzed by EDX with the results (in wt%) summarized below. the alloys generated by various investigators in
Courtesy of Welding Services Inc. laboratory coal-ash corrosion tests as well as plant exposure.
Source: Ref 73. Note: “This work” by Castello et al. in Lab tests:
5: 47.2% Cr, 35.9% Fe, 13.0% S, 2.5% Mn, and residual elements 10% alkali, 1% SO2, 700 °C (Ref 73); Plumley et al. (Ref 74): Plant
6: 54.2% Cr, 31.3% Fe, 5.7% S, 3.7% Mn, 2.8% Ni, and residual exposure, 677–727 °C, 1.7% S in coal, 6% alkali in ash; Kihara
elements et al. (Ref 75): Lab test, 7.3% alkali, 0.3% SO2, 700 °C; Van Weele
7: 45.9% Fe, 27.5% Ni, 15.9% Cr, 6.4% S, 3.7% Si, and residual et al. (Ref 76): Lab test, 10% alkali, 1% SO2, 10% H2O, 700 °C;
elements Blough (Ref 77): Plant exposure, 538–871 °C, 0.6–3.6% S, 0.4–
8: 50.9% Fe, 31.9% Ni, 10.3% Cr, 5.4% S, and residual elements 0.9% Na in coal.
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:04PM Plate # 0 pg 306

306 / High-Temperature Corrosion and Materials Applications

Ash Corrosion Resistant Materials Testing Pro- identified as IN72WO, the weld overlay was pro-
gram” (Ref 80, 81) were identified as IN72WO duced using alloy 72 weld wire (AWS ERNiCr-
and IN52WO in Fig. 10.87. For the sample tube 4: Ni-44Cr), while the alloy 52 weld overlay for
IN52WO was produced using alloy 52 weld
wire (AWS ERNiCrFe-7: Ni-30Cr-9Fe). In
Fig. 10.87, alloy 72 overlay was found to per-
form similarly to alloy 671 cladding. Alloy 52
overlay, although not performing as well as alloy
72 overlay and alloy 671 cladding, performed
significantly better than many austenitic alloys
particularly at higher temperatures.
In one subcritical boiler (255 MWe) burning
high-chlorine coal (about 0.3% Cl), Type 304H
reheater tubes typically lasted for about 4 years.
Test trials were performed for alloy 72/304H and
alloy 52/304H weld overlay tubes as part of the
reheater (538 °C/1000 °F outlet steam) for 3½
years. Typical tube cross sections are shown in
Fig. 10.88 for Alloy 72/304H tube and Fig. 10.89
for alloy 52/304H tube. Very little tube thinning
was observed for alloy 72 overlay. Slight corro-
sion was observed for alloy 52 overlay. The
maximum wastage rate was estimated to be less
than 0.05 mm/yr (2 mpy) for alloy 72 and about
0.2 mm/yr (8 mpy) for alloy 52. Kiser et al. (Ref
87) indicated that, in one boiler with low NOx
burners, alloy 72 overlay reheater tubes were
Fig. 10.84 Metal loss as a function of chromium contents in
the alloys in laboratory coal-ash corrosion tests found to show negligible wastage of about
(solid line) and plant exposure using corrosion probes inserted 0.076 mm (3 mils) after service for about 6 years
into the operating boiler at Tennessee Valley Authority’s Gallatin
Station Unit No. 2. The vertical axis on the right-hand side is in at the metal temperature of about 649 to 677 °C
mils. Source: Ref 78 (1200 to 1250 °F). The authors (Ref 87) also

Fig. 10.85 Metal wastage rates as a function of metal temperature for 304HSS, 347SS, alloy 800H, alloy NF709 (20Cr-25Ni-
1.5Mo-0.2Nb-Fe), alloy HR3C (25Cr-20Ni-0.5Nb-0.25N-Fe), alloy CR30A (30Cr-48Ni-2Mo-0.25Al-0.25Ti-
Fe), and chromized T22 (T22Cr). The data were generated from plant exposure tests at TVA’s Gallatin Station Unit No. 2 boiler.
Source: Ref 79
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:04PM Plate # 0 pg 307

Chapter 10: Coal-Fired Boilers / 307

indicated that, in another boiler, alloy 72 overlay stainless steels have adequate resistance, since
superheater tubes showed essentially no sign of the metal temperatures for these heat exchangers
corrosion or cracking after 6 years of service in normally do not exceed 650 °C (1200 °F)
sootblower lanes (Fig. 10.90). (Ref 88–90). Materials wastage due to erosion is
a major issue for the in-bed heat exchangers.
Stringer (Ref 88, 89, 91), Stallings and Stringer
10.6 Erosion in Fluidized-Bed Boilers (Ref 92), and Rademakers et al. (Ref 90) pro-
vided excellent reviews on erosion related to in-
In fixed-bed, fluidized-bed boilers, the major bed heat exchangers.
high-temperature materials issues are essentially The wastage problem for the in-bed heat
corrosion and erosion of in-bed heat-exchanger exchanger appears to be a low-temperature phe-
components. Corrosion is mainly sulfidation/ nomenon. Rademakers et al. (Ref 90) showed
oxidation. In general, the 300 series austenitic that the tube wastage rates were high at low
temperatures and decreased with increasing
temperature, as shown in Fig. 10.91. Stringer

Table 10.10 Results of field tests on austenitic

stainless steel tubes and co-extruded tubes in the
reheater of a 500 MW (e) boiler
Wastage rate

Material Exposure, h nm/h mpy

316 15,400 160 51

347 15,400 200 64
Fig. 10.86 Tube cross-section wastage profile for alloy Save
310/E 1250 18,000 60 19
25 (Sumitomo’s 25Cr-20Ni steel, similar to ASME
SA213TP310H or SA213TP310HCbN) after exposure to the 671/800H 34,000 Negligable
temperature for about 15.5 months. Wastage rates in each case Source: Ref 82
are also indicated. Source: Ref 81

Fig. 10.87 Tube metal wastage rates as a function of surface metal temperature for various alloys tested at Reliant Energy’s Niles plant
under the “Coal Ash Corrosion Resistant Materials Testing Program” conducted jointly by B&W, DOE, and the Ohio Coal
Development Office. Source: Ref 80
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:04PM Plate # 0 pg 308

308 / High-Temperature Corrosion and Materials Applications

(Ref 89) indicated that this tube wastage does not on the in-bed superheater tubes (Ref 92). This
appear to be a problem above approximately may suggest a temperature effect.
500 °C (932 °F). Wastage has taken place on The wastage damage is commonly referred to
the in-bed evaporator tubes, but seldom occurred be that of erosion. Stringer and Wright (Ref 88),
Stringer (Ref 89), and Stallings and Stringer (Ref
92) considered that mechanical processes, such
as erosion and abrasion wear, are involved in the
damage mechanisms. Stringer (Ref 91) defines
erosion and abrasion: erosion is material removal
by the impact of particles moving freely before
and after the impact; abrasion wear is material
removal by particles that are loaded onto the
surface and remain in contact for a period of time.
Possible wastage mechanisms are discussed in

Fig. 10.88 The cross section of alloy 72/304H overlay tube

after testing as part of the reheater (1000 °F
steam) for 3½ years in a 255 MW(e) subcritical boiler,
which burned high chlorine coal (about 0.3%). The 304H
reheater tubes exhibited a typical 4 year life. Courtesy of Welding
Services Inc.

Fig. 10.89 The cross section of alloy 52/304H overlay tube

after testing as part of the reheater (538 °C, or
1000 °F steam) for 3½ years in a 255 MW(e) subcritical boiler,
which burned high chlorine coal (about 0.3%), showing slight Fig. 10.90 Alloy 72 overlay superheater tubes in a soot-
pitting attack. The 304H reheater tubes exhibited a typical 4 year blower lane after 6 years of service. Original
life. Courtesy of Welding Services Inc. weld bead ripples are still clearly visible. Source: Ref 87
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:05PM Plate # 0 pg 309

Chapter 10: Coal-Fired Boilers / 309

detail by these authors in Ref 88, 89, 92, and 93. Type 304/SA210 A1 co-extruded tubes, two
Stringer and Wright (Ref 88) briefly summarize proprietary thermal sprayed coatings, sprayed
possible wastage mechanisms: (a) erosion by and fused WC-based coated (Extendalloy coat-
in-bed moving particles, (b) abrasion wear by ing) tubes, and chromized T11 tubes. The
“loaded” particles, (c) erosion by particles in the thickness of Extendalloy coating was about 25
wake of bubbles, (d) bubbles track along vertical mils. The in-bed evaporator tubes were operated
or sloping tubes, (e) erosion by particles thrown at approximately 343 to 371 °C (650 to 700 °F).
into the metal surface by bubble collapse, (f) Test tubes were removed for metallurgical eva-
erosion in the splash zone by bubble collapse at luation after exposure of about 6500 h. The
the bed surface, (g) erosion induced by in-bed findings, as reported by Lewis et al. (Ref 94),
jets, associated with coal or acceptor injection indicated that the maximum wastage rate was
ports, particle recirculation ports, and so forth, 8 mils/1000 h (70 mpy) for carbon steel tubes
and (h) erosion by “gulf stream” (long-range (SA210 A1 control samples) and 95 mils/1000 h
flow patterns) in the bed. (83 mpy) for Type 304 cladding (co-extruded
Materials wastage issues in fluidized-bed tubes). Extendalloy, chromized, and thermal
boilers have been extensively reviewed in Ref 88 sprayed coatings, however, showed no measur-
to 93. This section discusses the major erosion able wastage. Nevertheless, the authors observed
and/or abrasion problem (a) for in-bed eva- that oxide phases had formed at the coating/
porator tubes in bubbling fluidized-bed boilers substrate interface for the thermal sprayed coat-
and (b) at the refractory/waterwall interface ings and concluded that this could lead to coating
region in circulating fluidized-bed boilers. spallation (Ref 94). As for a chromized coating, it
is questionable whether such a thin coating
(about 400 µm thick) could sustain a long-term
10.6.1 In-Bed Evaporator Tubes in performance under fluidized-bed erosive condi-
Bubbling Fluidized-Bed Boilers tions.
The evaporator tubes in a bubbling fluidized- In examining evaporator tube wastage
bed boiler (TVA 20 MWe boiler) experienced experienced by the TVA 20-MW(e) pilot plant
severe wastage problems with carbon steel tubes (atmospheric bubbling fluidized-bed boiler)
during operation in 1986. The average wastage firing Pyro coal during operation in 1986,
rates for the carbon steel (SA210 A1) tubes were Stallings and Stringer (Ref 92) concluded that the
about 13 mils/1000 h with some tubes as high as tube wastage occurred mainly at the bottom of
35 mils/1000 h (Ref 94). A test program was the tube. This was because the larger, harder,
subsequently conducted to evaluate several more angular bed materials tend to migrate
cladding and coating materials that included toward the bottom region of the bed where they

Fig. 10.91 Tube wall wastage rates as a function of tube wall temperature from the in-bed tube tests in a 4 MW atmospheric fluidized-
bed combustor (AFBC). St. 35.8 is a carbon steel. Source: Ref 90
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:05PM Plate # 0 pg 310

310 / High-Temperature Corrosion and Materials Applications

were picked up in the wakes of bubbles and rose 82, and alloy 625 overlays were completely worn
upward. The tube wastage was primarily caused away at the underside of the tube samples. The
by the impact and motion of particles entrained in WC-base hardfacing overlay (HF60) showed no
bubble wakes on the underside surface of the sign of erosion or abrasive wear. The cross sec-
tubes. tions of all the test specimens are shown in
In a 130 MW(e) atmospheric bubbling flui- Fig. 10.93 (Ref 14). The wastage was found to be
dized-bed boiler, the in-bed evaporator was the greatest at the bottom of the tube surface
constructed out of sprayed and fused WC-based (6 o’clock position). Traditional stainless steels
hardfacing coated tubes. These coated tubes were and nickel alloys are not adequate in providing
in general satisfactory in performance. Never- erosion or abrasive wear protection. A proprie-
theless, occasional local failures did occur tary WC-based hardfacing weld overlay showed
according to the plant operator. Repair of the no evidence of wastage. The overlay surface
localized failures could be difficult for these showed a polished appearance. The micro-
sprayed and fused hardfacing coating. The plant structure of this hardfacing weld overlay and
operator was interested in weld overlay the analyses of the phases are shown in
hardfacing materials that were weld repairable Fig. 10.94. The hardness of this weld overlay
when local wear took place. A test tube consist- was found to be in the range of 50 to 60 HRC
ing of test sections of different weld overlay (converted from Vickers hardness values) across
alloys was tested as part of an evaporator tube. the overlay.
The steam in the evaporator tube element was A proprietary hardfacing weld overlay
about 336 °C (637 °F). The bed temperature was (HF35), based on chromium eutectic carbides,
typically 788 to 870 °C (1450 to 1600 °F). Test was developed and tested in the same bubbling
duration was 14,021 operating hours. The over- fluidized-bed boiler. Testing was conducted on
lay alloys tested included Type 309, Type 312, tubes in the loop section, which was historically a
alloy 82, alloy 625, a proprietary WC-based high wear area. Since it was a loop section of
hardfacing alloy (HF60), and a thermal sprayed the tube, the tube was weld overlaid using a
coating on alloy 625 overlay (butter layer). manual technique. Figure 10.95 shows one of the
Evaluation of test specimens after 14,021 oper- overlaid tubes in the loop section after exposure
ating hours showed that thermal sprayed coating for close to 3 years as part of the in-bed eva-
was gone (worn away). Figure 10.92 shows porator tube bundle. The manually applied
spallation of the coating. Type 309, 312, alloy longitudinal weld beads are still visible. Two
tubes in the loop section were later removed from
the boiler for metallurgical evaluation. The weld
overlay overall remained protective under the
erosive conditions for about 3 years. There was,
however, a localized wear area near the pin
studded carbon steel tube end. Both tubes
showed the similar locations that suffered loca-
lized wear.
An in-bed superheater tube bundle located
above this in-bed evaporator bundle reportedly
showed no erosion or abrasion problems. The
superheater was made of an austenitic stainless
steel. This appears to confirm earlier observation
by Rademakers et al. (Ref 90) that the erosion
or abrasion wear of the in-bed tubes were sig-
nificantly reduced at higher temperatures and
by Stringer (Ref 89) that the in-bed tube
0.1 mm wastage problem does not appear to be a prob-
lem above approximately 500 °C (932 °F).
Fig. 10.92 Spallation of the thermal sprayed coating on the
The plant engineer indicated that the in-bed
bottom surface of the test tube sample after
exposure for 14,021 h as part of the in-bed evaporator tubes in a superheater tube bundle was more tightly ar-
130 MW(e) bubbling fluidized-bed boiler. The coating was ranged than the in-bed evaporator tubes
applied to alloy 625 overlay, which acted as a butter layer
between the coating and the substrate steel tube. Courtesy of (Ref 14). It is not clear whether this could be a
Welding Services Inc. factor also.
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:05PM Plate # 0 pg 311

Chapter 10: Coal-Fired Boilers / 311

10.6.2 Waterwalls in Circulating 95–97). It is generally believed that in a circu-

Fluidized-Bed Boilers lating fluidized-bed boiler the bed particles flow
upward in the center of the boiler and migrate
One serious materials issue for circulating downward along the waterwall (Ref 98, 99). This
fluidized-bed boilers is the erosion or abrasion downward flow of particles is believed to be
wear that occurs on the waterwall at the refrac- responsible for the erosion wear above the
tory lining-waterwall transition region (Ref 90, refractory lining. The authors suggested that the

(a) (b)

(c) (d)

Fig. 10.93 Cross sections of test tube samples showing wear profiles for (a) Type 309 overlay tube, (b) alloy 625 overlay tube, (c) Type
312 overlay tube, and (d) HF60 hardfacing overlay tube after exposure for 14,021 h as part of the in-bed evaporator tubes
in a 130 MWe bubbling fluidized-bed boiler. Courtesy of Welding Services Inc.
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:05PM Plate # 0 pg 312

312 / High-Temperature Corrosion and Materials Applications

particles flowing downward along the waterwall surface. Slusser et al. (Ref 95) reported the
were forced to change flow direction at a tapered waterwall erosion wear problem at the transition
(or sloped) refractory lining forcing the particles region in a 50 MW(e) boiler. Figure 10.96
to flow from the membrane (or web) outward illustrates schematically the relationship between
against the tube walls on the tapered refractory the waterwall and the refractory lining at the
transition region (Ref 95). The figure also indi-
cates that the potentially high wear area was
protected by a weld overlay. No technical infor-
mation about the weld overlay alloy used was
discussed in the paper. However, the weld over-
lay was reported to have suffered severe metal
loss (Ref 95).
Some of the protective methods suggested by
the authors (Ref 95–97), although somewhat
helpful, may not be long-term solutions. For
example, installing a shelf above the refractory
lining by interrupting the downward particle flow
was found to be helpful in reducing wastage rates
at the waterwall-refractory transition region, but
induced wear at the shelf location (Ref 95).
Because of the interruption of the particle flow
direction by the tapered refractory surface, par-
ticles are then forced to flow against the tube
walls on the tapered refractory surface, thus
resulting in high waterwall tube wear at the
transition region. This is illustrated schematically
19 µm in Fig. 10.97. Nevertheless, Slusser et al. (Ref 96)
reported in their 1992 paper indicating that
Fig. 10.94 Scanning electron micrograph (backscattered shelves (7 in. wide sheets) installed perpendi-
electron image) showing various hardface par- cular to the waterwalls around the entire
ticles in the proprietary tungsten carbide based hardfacing weld
overlay, HF60. The results (wt%) of semiquantative EDX analyses
perimeter of the boiler at 2.2 m (7 ft) and 5.2 m
of various phases are summarized as: Light color phases (A, B, C, (17 ft) above the refractory transition area were
E, and F) are tungsten carbides (60–70W, 14–16Ni, 8Cr, 5Si), capable of reducing wastage rates from >25 mm/
grayish phase (H) is also tungsten carbide ( 36W, 39Ni, 18Cr,
2.5Si), and dark phases (I and G) are matrix (70–80Ni, 10Cr, yr (>1000 mpy) to a manageable rate, thus
5–10W). Courtesy of Welding Services Inc. allowing reasonable operating periods between

Fig. 10.95 One tube with a proprietary HF35 hardfacing weld overlay showing manually applied weld overlay beads along the tube
that are still visible after exposure for approximately 3 years as part of an in-bed evaporator tube bundle in a 130 MWe
bubbling fluidized-bed boiler. Courtesy of Welding Services Inc.
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:05PM Plate # 0 pg 313

Chapter 10: Coal-Fired Boilers / 313

inspections. Wastage rates at the shelves were flue gas stream occurs under certain conditions.
found to be accelerated. For example, the For example, excessive fly-ash erosion of the
waterwalls at the 17 ft shelf location suffered economizer can occur in the backpass area
wastage rates of about 0.76 to 1.78 mm/yr (30 to when the flue gas stream making the direct-
70 mpy) (Ref 96). Some other protection meth- ion change and creating uneven distribution
ods including cast erosion blocks made of Type of fly-ash particles due to centrifugal force
310 and thermal sprayed coatings of various (Ref 1).
alloys were mentioned under trials in a 1997 An example is given below on erosion or
paper by Solomon (Ref 97). Some coatings erosion/corrosion of the hanger tubes in the
such as thermal sprayed coatings of 50Ni-50Cr backpass area for the economizer of a circulating
alloy and Cr2O3-base coating were found to be fluidized-bed boiler (90 MWe) burning Eastern
unacceptable in performance (Ref 97). Some bituminous coal (Ref 14). The current hanger
success was attained in an approach involving tubes are protected with tube shields, which
bending the waterwall tubes out of the vertical typically last for a year. A Type 309 overlay tube
plane and into the wall at the interface with the and an HF35 hardfacing overlay tube were tested
refractory lining making a smooth, vertical as part of the hanger tubes at top of the backpass.
transition from the waterwall tubes to the The temperature of the steam inside the tubes was
refractory (Ref 93). reportedly about 315 °C (600 °F). The tubes
were removed for metallurgical evaluation after
10.6.3 Erosion in the Convection Pass 2 years of exposure. For both tubes, the leeward
side typically exhibited some thin scales,
Tubes in Fluidized-Bed Boilers
while the windward side showed no scales.
In fluidized-bed boilers, flue gas stream Figure 10.98 shows typical cross sections of
leaving the combustor and enters a cyclone, these tube samples after 2 years of exposure. The
where particles are removed, before exiting to Type 309 overlay tube showed more thinning
the convection pass. Typically, a superheater on the overlay on the windward side of the tube
and an economizer are in the convection pass (top side of the cross section in the photograph).
in the downstream of the cyclone. The flue This can be seen in tube cross sections as
gas stream is entrained with fine particles that shown in Fig. 10.98. The metallographic cross
are not removed by the cyclone. High tube sections showing the overlay in comparing
metal wastages could occur at superheater the windward side with the leeward side of the
tube banks and economizer tube banks tube for both 309 and HF35 overlay tubes, as
(Ref 93). Erosion by fly ash can be serious shown in Fig. 10.99 and 10.100. The surface
when the uneven distribution of fly ash in the condition as well as the microstructure of the
weld overlay at the windward side for both
overlays is shown in Figs. 10.101 and 10.102.
The 309 overlay exhibits austenitic structure with
some delta ferrite phases, with a hardness range
of about 91 to 92 HRB (converted from Vickers
hardness values). The HF35 overlay exhibits

Fig. 10.97 Particles flowing outward against the water-

Fig. 10.96 Waterwall in the transition region that suffers wall tube wall from the membrane due to inter-
severe erosion wear immediately above the ruption of the downward flow of particles on a shelf. Source:
tapered refractory surface. Source: Ref 95 Ref 96
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:05PM Plate # 0 pg 314

314 / High-Temperature Corrosion and Materials Applications

chromium eutectic carbides formed along inter- for the 309 overlay and about 20% Cr for the
dendritic boundaries, with a hardness range of HF35 overlay, and both should form chromium
about 35 to 40 HRC (converted from Vickers oxide scales. The reduced wastage rate exhibited
hardness values). The wastage rate was calcu- by the HF35 overlay over the 309 overlay is
lated to be about 0.38 mm/yr (15 mpy) for believed to be due to its erosion resistance
Type 309 overlay and 0.2 mm/yr (8 mpy) for resulting from hardfacing particles of chromium
HF35 hardfacing overlay. Both overlays con- eutectic carbides.
tained similar chromium contents, about 21% Cr

10.7 Summary

A general description of coal-fired boilers and

their combustion conditions is presented. Fire-
side corrosion of the furnace waterwalls is dis-
cussed with particular emphasis on the materials
problems associated with a staged firing by sub-
stoichiometric combustion (i.e., combustion with
insufficient oxygen) in the lower furnace for
reducing NOx emissions. As a result of the staged

(a) 0.010 in.

(b) 0.010 in.

Fig. 10.98 Cross sections of Type 309 overlay tube (a) and
HF35 overlay tube (b) showing the wastage
profile of the weld overlay after exposure of 2 years as part of the Fig. 10.99 Type 309 overlay tube in the unaffected leeward
hanger tube for the economizer in a circulating fluidized-bed side (a) and the windward side (b) after testing as
boiler. The top of the tube cross section was the windward side part of hanger tubes for the economizer in a circulating fluidized-
while the bottom (i.e., the ruler side) was the leeward side. bed boiler for 2 years. The substrate steel was etched with nital to
Courtesy of Welding Services Inc. reveal the fusion boundary. Courtesy of Welding Services Inc.
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:05PM Plate # 0 pg 315

Chapter 10: Coal-Fired Boilers / 315

firing, sulfidation dominates the corrosion pro- presumably due to overheating and preferential
cess for waterwall materials. The wastage rates sulfidation penetration. The possible causes to
for carbon and low alloy steels have been found circumferential grooving and cracking are dis-
to increase to an unacceptable rate for many cussed.
boilers, particularly those of supercritical units. Also, included in the discussion are erosion-
The current most widely used waterwall pro- corrosion caused by steam sootblowing and
tection method against tube wall wastage pro- thermal fatigue cracking due to the use of water
blems involves weld overlay cladding in the lances or water canons for deslagging. Super-
boiler on the affected waterwall area with a cor- heater/reheater wastage problems as well as
rosion-resistant alloy using automatic gas-metal- erosion-corrosion issues in convection pass tubes
arc welding (GMAW) process. Cladding of are discussed. Erosion, erosion-corrosion, or ab-
waterwall panels can also be applied in shop rasion wear issues encountered in fluidized-bed
using either GMAW or laser cladding techniques, boilers including both bubbling beds and circu-
and the cladded panels are then installed in lating beds are also discussed.
the boiler. The use of a corrosion-resistant
cladding has essentially eliminated the tube
wastage problems. For some supercritical
units, circumferential grooving and cracking has
been encountered on overlaid waterwall tubes

0.0010 in.

Fig. 10.101 The surface of Type 309 overlay tube at the

windward side after exposure for 2 years as part
of the hanger tube for the economizer in a circulating fluidized-
bed boiler. Microstructure of Type 309 overlay consists of delta
ferrite (dark phases) in austenite. Courtesy of Welding Services
(a) 0.010 in. Inc.

(b) 0.010 in.

0.0010 in.

Fig. 10.100 HF35 overlay tube in the unaffected leeward Fig. 10.102 The surface of HF35 weld overlay tube at the
side (a) and the windward side (b) after testing windward side after exposure for 2 years as part
as part of the hanger tube for the economizer in a circulating of the hanger tube for the economizer in a circulating fluidized-
fluidized-bed boiler for 2 years. The substrate steel was etched bed boiler. The overlay contained hardfacing chromium eutectic
with nital to reveal the fusion boundary. Courtesy of Welding carbides (dark phases) formed along interdendritic boundaries.
Services Inc. Courtesy of Welding Services Inc.
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:05PM Plate # 0 pg 316

316 / High-Temperature Corrosion and Materials Applications

REFERENCES from Laboratory Data, Corros. Sci., Vol 21

(No. 12), 1981, p 789
1. S.C. Stultz and J.B. Kitto, Ed., Steam and Its 13. A.J.B. Cutler, T. Flatley, and K.A. Hay,
Generation and Use, 40th ed., Babcock & Metall. Mater. Technol., Feb 1981, p 69
Wilcox, Barberton, OH, 1992 14. Welding Services Inc., Norcross, GA
2. J.G. Singer, Ed., Combustion Fossil Power, 15. D.N. French, Metallurgical Failures in
4th ed., Combustion Engineering, Inc., Fossil Fired Boilers, John Wiley & Sons,
Windsor, CT, 1991 1993
3. C.A. Meyer, et al., ASME Steam Tables, 16. D.N. French, Paper No. 133, Corrosion/88,
5th ed., The American Society of Mechan- NACE, 1988
ical Engineers, 1983 17. T. Flatley, E.P. Latham, and C.W. Morris, in
4. R. Peltier and K. Wicker, PRB Coal Makes Inst. Metallurgists Spring Residential Con-
the Grade, Power, Oct 2003, p 28 ference: Environmental Degradation of
5. J. Stringer, High Temperature Corrosion High Temperature Materials, Vol 2 (No. 13),
Problems in the Electric Power Industry Series 3, 1980
and Their Solution, in High Temperature 18. F. Clarke and C.W. Morris, Combustion
Corrosion, R.A. Rapp, Ed., Proc. Conf. Aspects of Furnace Wall Corrosion, in
(San Diego, CA, March 2–6, 1981, NACE, Corrosion Resistant Materials for Coal
1981 Conversion Systems, D.B. Meadowcroft and
6. D.J. Lees and M.E. Whitehead, Micro- M.I. Manning, Ed., Applied Science Pub-
analysis of Scales and Deposits Formed on lishers, London, 1982, p 47
Corroding Furnace Tubes in Coal-Fired 19. P.L. Daniel, L.D. Paul, J.M. Tanzosh, and
Boilers, in Corrosion Resistant Materials for R. Hubinger, Estimating Effects of Chlorine
Coal Conversion Systems, D.B. Meadow- on Fireside Corrosion of Furnace Walls in
croft and M.I. Manning, Ed., Applied Sci- Coal-Fired Boilers, Paper No. 138, Corro-
ence, London, 1982, p 63 sion/88, NACE, 1988
7. K.A. Hay, High Temperature Environmental 20. E. Latham, D.B. Meadowcroft, and L. Pin-
Degradation of Materials and Future der, The Effects of Coal Chlorine on Fireside
Requirements in The Power Generation Corrosion, Chlorine in Coal, J. Stringer and
Industry, in Environmental Degradation of D.D. Barnerjee, Ed., Elsevier Science Pub-
High Temperature Materials, Vol 2 (No. 3), lishers B.V., Amsterdam, 1991, p 225
Series 3, 1980, p 4/1 21. S. Brooks and D.B. Meadowcroft, The Role
8. G.T. Cunningham and T. Webster, “Design, of Chlorine in the Corrosion of Mild and
Manufacturing, and Testing of a Boiler Low Alloy Steels in Low pO2 and High pS2
Metal Surface Temperature Probe,” Paper Environments, in High Temperature Corro-
No. IJPCG 2000-15075, presented at the sion in Energy Systems, M.F. Rothman, Ed.,
2000 International Joint Power Generation The Metallurgical Society of AIME, 1985,
Conference (Miami Beach, FL), July 23–26, p 515
2000 22. S. Brooks and D.B. Meadowcroft, The
9. J. Stringer, Materials for Fossil Energy Influence of Chlorine on the Corrosion of
Systems—Past, Present, and Future, Paper Mild and Low Alloy Steels in Sub-stoi-
No. 225, Corrosion/93, NACE, 1993 chiometric Combustion Gases, in Corrosion
10. J.A. Urich and E. Kramer, Designing Solu- Resistant Materials for Coal Conversion
tions for Low NOx-Related Water Wall Systems, D.B. Meadowcroft and M.I. Man-
Corrosion, Conf. Proc., 1996 ASME ning, Ed., Applied Science, London, 1982,
Joint Power Generation Conference, p 105
S.M. Smouse and A. Gupta, Ed., ASME, 23. T. Flatley, E.P. Latham, and C.W. Morris,
1996, p 25 Co-Extruded Tubes Improve Resistance to
11. M.D. Yeager, “The Effect of Low NOx Sta- Fuel-Ash Corrosion in U.K. Utility Boilers,
ging on Waterwall Wastage,” presented at Paper No. 62, Corrosion/80, NACE, 1980
the EPRI International Conference on Boiler 24. P.J. James and L.W. Pinder, The Impact of
Tube Failures in Fossil Plants (Nashville, Coal Chlorine on the Fireside Corrosion
TN), Nov 11–13, 1997 Behavior of Boiler Tubing: A UK Perspec-
12. A.J.B. Cutler and E. Raask, External Cor- tive, Corrosion/97, NACE International,
rosion in Coal-Fired Boilers: Assessment 1997
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:05PM Plate # 0 pg 317

Chapter 10: Coal-Fired Boilers / 317

25. D.J. Lees, “A Summary of Observations Term Testing of Protective Coatings and
Relating Furnace Wall Fireside Corrosion to Weld Overlays in a Supercritical Boiler,
Chlorine Content of Coal,” SSD/MID/M26/ Retrofitted with Low NOx Burners, Paper
79, 1979 No. 2384, Corrosion/2002, NACE Interna-
26. S.F. Chou and P.L. Daniel, in High Tem- tional, 2002
perature Corrosion in Energy Systems, M.F. 36. P. Hulsizer, Paper No. 246, Corrosion/91,
Rothman, Ed., The Metallurgical Society of NACE, 1991
AIME, 1985 37. G. Lai, P. Hulsizer, and R. Lee, “Waterwall
27. K.S. Gilroy, Laboratory Evaluation of Can- Wastage Mitigation for Coal-Fired Boilers
didate Materials for Furnace Wall Applica- Using Automatic Pulse Spray GMAW
tions, in High Temperature Corrosion in Overlay Technology,” presented at 1999
Energy Systems, M.F. Rothman, Ed., EPRI Fossil Plant Maintenance Conference
The Metallurgical Society of AIME, 1985, (Atlanta, GA), 1999
p 345 38. G. Lai and P. Hulsizer, Corrosion & Erosion/
28. S.C. Kung and C.F. Eckhart, Corrosion of Corrosion Protection by Modern Weld
Iron-Base Alloys in Reducing Combustion Overlays in Low NOx Coal-Fired Boilers,
Gases, Paper No. 242, Corrosion/93, Paper No. 258, Corrosion/2000, NACE
NACE, 1993 International, 2000
29. B. Dooley, R. Tilley, T.P. Sherlock, and 39. G.Y. Lai, “Performance of Automatic
C.H. Wells, “State of Knowledge Assess- GMAW Overlays for Waterwall Protection
ment for Waterwall Wastage,” presented at in Coal-Fired Boilers,” presented at EPRI
EPRI International Conference on Boiler fifth International Conference on Welding
Tube Failures in Fossil Plants (Nashville, and Repair Technology for Power Plants
TN), Nov 11–13, 1997 (Point Clear, AL), June 26–28, 2002
30. Workshop on Materials Issues Associated 40. G.Y. Lai, Fireside Corrosion and Erosion/
with Low-NOx Combustion in Fossil-Fired Corrosion Protection in Coal-Fired Boilers,
Boilers, Summary of Workshop held during Paper No. 4522, Corrosion/2004, NACE
Advanced Research and Technology International, 2004
Development’s Tenth Annual Conference 41. M.S. Brennan and R.C. Gassmann, “Laser
on Fossil Energy Materials (Knoxville, TN), Cladding of Nickel- and Iron-Base Alloys
May 14–16, 1996, Materials & Components on Boiler Pressure Panels and Tubes,” pre-
in Fossil Energy Applications, Department sented at the EPRI Third International
of Energy and Electric Power Research Conference on Welding and Repair Tech-
Institute, No. 123, August 1, 1996 nology for Power Plants (Scottsdale, AZ),
31. C. Jones, Maladies of Low-NOx Firing June 9–12, 1998
Come Home to Roost, Power, Jan/Feb, 42. M.S. Brennan and R.C. Gassmann, Laser
1997, p 54 Cladding of Nickel and Iron Base Alloys on
32. P.J. James and L.W. Pinder, “Furnace Wall Boiler Waterwall Panels and Tubes, Paper
Fireside Corrosion: Taming the Beast No. 235, Corrosion/2000, NACE Interna-
Within,” presented at EPRI International tional, 2000
Conference on Boiler Tube Failures in 43. A.J. Bonnington and M.S. Brennan, “Type
Fossil Plants (Nashville, TN), Nov 11–13, 312 Stainless Steel Laser Cladding for
1997 Waterwalls in Supercritical Units,” pre-
33. W.T. Bakker and S.C. Kung, Waterwall sented at the EPRI/DOE Conference on
Corrosion in Coal-Fired Boilers—A New Advances in Life Assessment and Optimi-
Culprit: FeS, Paper No. 246, Corrosion/ zation of Fossil Power Plants (Orlando, FL)
2000, NACE International, 2000 March 11–13, 2002
34. W. Bakker, “Root Causes of Accelerated 44. “Boiler Tube Failure Metallurgical Guide,
Waterwall Wastage in Coal-Fired Boilers,” Volume 2: Appendices,” EPRI TR-102433-
presented at EPRI International Conference V2, Electric Power Research Institute, Oct
on Materials and Corrosion Experience for 1993
Fossil Power Plants (Isle of Palms, SC), Nov 45. “INCONEL alloy 625,” Inco Alloys Inter-
18–21, 2003 national Huntington, WV
35. W.T. Bakker, J.L. Blough, S.C. Kung, 46. “HASTELLOY C-22 Alloy,” H-2019D,
T.L. Banfield, and P. Cunningham, Long Haynes International Inc., Kokomo, IN
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:05PM Plate # 0 pg 318

318 / High-Temperature Corrosion and Materials Applications

47. K. Coleman and D. Gandy, “Corrosion 58. M.F. Stroosnijder, V. Guttmann, and
Resistant Waterwall Overlays Selection of R.J.N. Gommans, Influence of Creep
Alternative Materials,” presented at Inter- Deformation on the Corrosion Behavior of
national Conference on Boiler Tube Failures a CeO2 Surface-Modified Alloy 800H in a
and HRSG Tube Failures, and Inspections Sulfidizing-Oxidizing-Carburizing Envir-
(Phoenix, AZ), Nov 6–8, 2001 onment, Mater. Sci. Eng., Vol A121, 1989,
48. G. Lai, The Microstructure, “Properties and p 581
Corrosion Resistance of Type 312 Overlay 59. J.L. Coze, et al., The Development of High-
in Batch Digesters,” Conf. Proc., 2002 Temperature Corrosion-Resistant Alumi-
TAPPI Fall Technical Conference (San num-Containing Ferritic Steels, Mater. Sci.
Diego, CA), Sept 8–11, 2002 Eng., Vol A120, 1989, p 293
49. G.Y. Lai, presented at WSI Boiler Tube 60. E.C. Lewis and A.L. Plumley, Chromizing
Overlay Forum (Atlanta, GA) July 10–12, for Combating Fireside Corrosion, in
2006 Advances in Materials Technology for Fossil
50. J.L. Blough, private communication, Power Plants, R. Viswanathan and
First Energy, Mayfield Village, OH, Sept 29, R.I. Jaffee, Ed., ASM International, 1987,
2006 p 291
51. I.G. Wright, Hot Corrosion in Coal- and Oil- 61. H.M. Tawancy, Structure & Properties of
Fired Boilers, Metals Handbook, Vol 13, 9th High Temperature Alloys: Applications of
ed., Corrosion, ASM International, 1987, p Analytical Electron Microscopy, King Fahd
995 University of Petroleum & Minerals,
52. S. French, K. Rumbaugh, and P.N. Hulsizer, Dhahran, Saudi Arabia, 1993
Fireside Corrosion-Erosion Mitigation via 62. L. Paul, G. Clark, and A. Ossenberg-Engels,
the Application of Weld Metal Overlay, “Protection of Waterwall Tubes from Cor-
Proceedings: Welding and Repair Technol- rosion in Low NOx Coal Fired Boilers,”
ogy for Power Plants, EPRI, Charlotte, NC presented at Coal-Gen Conference (Cincin-
1997, p 477 nati, OH), Aug 16–18, 2006
53. A.J. Bonnington and T.M. Cullen, “Perfor- 63. R.E. Kessler, “Thermal Fatigue Cracking of
mance of Chromized Waterwall Panels in Waterwall Tubes from Waterlances and
Supercritical Units,” presented at EPRI Water Cannons,” presented at EPRI Intelli-
International Conference on Boiler Tube gent Sootblowing Workshop (Houston, TX),
Failures in Fossil Plants, (Nashville, TN), March 19–21, 2002
Nov 11–13, 1997 64. M. Carlisle, J. Sorge, and B. Mead, “Water
54. K. Luer, J. DuPont, A. Marder, and C. Cannon Application at Alabama Power’s
Skelonis, Corrosion Fatigue of Alloy 625 Plant Miller,” presented at EPRI Intelligent
Weld Claddings in Combustion Environ- Sootblowing Workshop (Houston, TX),
ments, Mater. High Temp., Vol 18 (No. 1), March 19–21, 2002
2001, p 11 65. B. Ray, R. Hemperley, and R. Courtney,
55. K. Stein, V. Guttmann, and W.T. Bakker, “W.A. Parish Units 7 & 8 ISB Project,”
The Influence of Deformation on High presented at EPRI Intelligent Sootblowing
Temperature Corrosion of CRONIFER Workshop (Houston, TX), March 19–21,
45TM, in Heat-Resistant Materials II, Conf. 2002
Proc. of the Second International Con- 66. H.S. Blinka, “W. A. Parish Unit 5 ISB Pro-
ference on Heat-Resistant Materials, ject,” presented at EPRI Intelligent Soot-
K. Natesan, P. Ganesan, G. Lai, Ed., ASM blowing Workshop (Houston, TX), March
International, 1995, p 367 19–21, 2002
56. V. Guttmann, K. Stein, and W.T. Bakker, 67. R.W. Borio, A.L. Plumley, and W.R. Syl-
Deformation-Corrosion Interactions in vester, in Ash Deposits and Corrosion Due
Selected Advanced High Temperature to Impurities in Combustion Gases,
Alloys, Mater. High Temp., Vol 14 (No. 2/3), R.W. Bryers, Ed., Hemisphere Publishing/
1997, p 61 McGraw-Hill, 1978, p 163
57. P. Castello, V. Guttmann, N. Farr, and 68. C. Cain, Jr. and W. Nelson, J. Eng. Power,
G. Smith, Simulated Coal Ash Corrosion of Trans. ASME, Oct 1961, p 468
Ni-Based Alloys, Mater. High Temp., Vol 19 69. W. Nelson and C. Cain, Jr., J. Eng. Power,
(No. 1), 2002, p 29 Trans. ASME, July 1960, p 194
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:05PM Plate # 0 pg 319

Chapter 10: Coal-Fired Boilers / 319

70. J.L. Blough and S. Kihara, Paper No. 129, Resistant Materials for Coal Conversion
Corrosion/88, NACE, 1988 Systems, D.B. Meadowcroft and M.I. Man-
71. R.W. Borio and R.P. Hensel, J. Eng. Power, ning, Ed., Applied Science Publishers,
Trans. ASME, Vol 94, 1972, p 142 London, U.K., 1982, p 137
72. J. Stringer, in, Corrosion, Vol 13, 9th ed., 83. T. Flatley, E.P. Latham, and C.W. Morris,
ASM International, 1987, p 998 CEGB Experience with Co-Extruded Tubes
73. P. Castello, V. Guttmann, N. Farr and for Superheated and Evaporative Sections of
G. Smith, Laboratory-Simulated Fuel-Ash PF Fired Boilers, in Advances in Materials
Corrosion of Superheater Tubes in Coal- Technology for Fossil Power Plants, Conf.
Fired Ultra-Supercrtical-Boilers, Mater. Proc., R. Viswanathan and R.I. Jaffee, Ed.,
Corros., Vol 51, 2000, p 786 ASM International, 1987, p 219
74. A.L. Plumley, J.I. Accort, and W.R. Rocz- 84. M.G. Fahrmann and G.D. Smith, Evaluation
niak, NACE-DOE Conference, Berkeley, of Clad Tubing after 18 Years of Service
CA, 1979 in a Coal-Fired Utility Boiler, Paper No.
75. S. Kihara, K. Nakagawa, A. Ohtomo, H. 232, Corrosion/2000, NACE International,
Aoki, and S. Ando, Simulated Test 2000
Results for Fireside Corrosion of Super- 85. S.D. Kiser and T. Orsini, Benefits of Chro-
heater and Reheater Tubes Operated at mium in Nickel for Corrosion Protection of
Advanced Steam Condition in Coal- Superheater and Reheater Tubes in Coal
Fired Boilers, in High Temperature Corro- Fired Boilers, Paper No. 5455, Corrosion/
sion in Energy Systems, M.F. Rothman, Ed., 2005, NACE International, 2005
The Metallurgical Society of AIME, 1985, 86. P.N. Hulsizer, Dual Pass Weld Overlay
p 361 Method and Apparatus, U.S. Patent No.
76. S. Van Weele, J.L. Blough, and J.H. DeVan, 6,013,890, Jan 11, 2000
Paper No. 182, Corrosion/94, NACE Inter- 87. S.D. Kiser, E.B. Hinshaw, and T. Orsini,
national, 1994 Extending the Life of Fossil Fired Boiler
77. J.L. Blough, ORNL/Sub/93-SM401/ Tubing with Cladding of Nickel Based
01, Foster Wheeler Corp., Livingston, NJ, Alloy Materials, Paper No. 06474, Corro-
1996 sion/2006, NACE International, 2006
78. S. Kihara, J.L. Blough, W. Wolowodiuk, and 88. J. Stringer and I.G. Wright, Materials Issues
W.T. Bakker, “Prediction of Corrosion Rate in Fluidized Bed Combustion, J. Mater.
of Superheater Tube in Boilers Burning Energy Systems, Vol 8 (No. 3), 1986, p 319
Various Kinds of Coals,” presented at the 89. J. Stringer, Alloys for Advanced Power
NACE International Conference on Life Systems, Heat-Resistant Materials, Proc. of
Prediction of Corrodible Structures (Kauai, the First International Conference, (Fontana,
HA), 1991 WI), Sept 23–26, 1991, K. Natesan and
79. J.L. Blough, G.J. Stanko, W.T. Bakker, and D.J. Tillack, Ed., ASM International, 1991
J.B. Brooks, “Superheater Corrosion in 90. P.L.F. Rademakers, D.M. Lloyd, and
Ultra-Supercritical Power Plants,” Paper No. V. Regis, AFBC’s: Bubbling, Circulating
250, Corrosion/2000, NACE International, and Shallow Beds, in High Temperature
2000 Materials for Power Engineering 1990,
80. D.K. McDonald, “Coal Ash Corrosion Part I, Conf. Proc. (Liege, Belgium), Sept
Resistant Materials Testing Program: Eva- 24–27, 1990, E. Bachelet, et al., Ed., Kluwer
luation of the First Section Removed in Academic Publishers, Dordrecht, Nether-
November 2001,” Babcock & Wilcox lands, 1990, p 43
Report, Barberton, OH 91. J. Stringer, Erosion/Corrosion in Fluidized
81. D.K. McDonald and E.S. Robitz, “Coal Bed Combustion Boilers, in Advances in
Ash Corrosion Resistant Materials Testing Materials Technology for Fossil Power
Program: Evaluation of the Second Section Plants, Conf. Proc., R. Viswanathan and
Removed in August 2003,” Babcock & R.I. Jaffee, Ed., ASM International, 1987,
Wilcox Report, Barberton, OH p 319
82. E.P. Latham, T. Flatley, and C.W. Morris, 92. J.W. Stallings and J. Stringer, Mechanisms
Comparative Performance of Superheated of In-Bed Tube Wastage in Fluidized-Bed
Steam Tube Materials in Pulverized Fuel Combustors, in Corrosion-Erosion-Wear of
Fired Plant Environments, in Corrosion Materials at Elevated Temperatures, Conf.
Name ///sr-nova/Dclabs_wip/High Temp/5208_259-320.pdf/Chap_10/ 30/10/2007 4:05PM Plate # 0 pg 320

320 / High-Temperature Corrosion and Materials Applications

Proc., A.V. Levy, Ed., NACE International, of A Circulating Fluidized Bed Combustor,
1991, p 19-1 Paper No. 138, Corrosion/92, NACE, 1992
93. I.G. Wright, A Review of Experience 97. N.G. Solomon, Erosion-Resistant Coatings
of Wastage in Fluidized-Bed Boilers, for Fluidized-Bed Boilers, Paper No. 135,
Mater. High Temp., Vol 14 (No. 2/3), 1997, Corrosion/97, NACE International, 1997
p 207 98. M.D. Mirolli and W.P. Bauer, “A Summary
94. E.C. Lewis, D.A. Canonico, and R.Q. Vin- of Combustion Engineering’s Programs to
cent, Metal Wastage Experiences in BFBC Control CFB Material Wastage,” presented
Environments, in Corrosion-Erosion-Wear at the EPRI-Argonne Workshop on Materi-
of Materials at Elevated Temperatures, als Issues in Circulating Fluidized Bed
Conf. Proc., A.V. Levy, Ed., NACE Inter- Combustors, Argonne, IL, 1989
national, 1991, p 20-1 99. J. Zhao, R. Wu, R. Senior, R. Legros,
95. J.W. Slusser, A.D. Bixler, and S.P. Bartlett, C. Brereton, J. Grace, and C. Lim, “Spatial
Materials Experience from A Circulating Variation Inside a Pilot Scale Circulating
Fluidized Bed Coal Combustor, Paper No. Fluidized Bed Combustion Unit,” presented
285, Corrosion/90, NACE, 1990 at the EPRI-Argonne Workshop on Materi-
96. J.W. Slusser, A.D. Bixler, and D.E. als Issues in Circulating fluidized Bed
Thompson, Four Years of Field Performance Combustors, Argonne, IL, 1989
Name ///sr-nova/Dclabs_wip/High Temp/5208_321-334.pdf/Chap_11/ 26/10/2007 12:44PM Plate # 0 pg 321

High-Temperature Corrosion And Materials Applications Copyright © 2007 ASM International®

George Y. Lai, editor, p321-334 All rights reserved.
DOI: 10.1361/hcma2007p321 www.asminternational.org

CHAPTER 11

Oil-Fired Boilers and Furnaces

11.1 Introduction deposits collected from corrosion probes at metal

temperatures of 475 to 675 °C (890 to 1250 °F)
Fuel oils primarily consist of residues from from a utility boiler at Marchwood Power Station
distillation of crude oil in petroleum refining. (United Kingdom) is shown in Table 11.1 (Ref
Sulfur in fuel oils can vary from a fraction of a 4). Table 11.2 shows the results of the analysis
percent for lighter oils to 3.5% for some residual of the deposits from another utility boiler burn-
oils (Ref 1). Two other corrosive constituents in ing Buncker “C” fuel oil (Ref 5). Ash deposits
fuel oils are vanadium and sodium. Vanadium formed on metallic components whether cooled
exists in certain crude oils as an oil-soluble por- (e.g., superheaters/reheaters) or uncooled (e.g.,
phyrin complex (Ref 2). Fuel oils from some tube hangers) can exhibit low melting points.
crudes contain low levels of vanadium, while Table 11.3 lists a number of oil-ash constituents
certain Middle Eastern crudes and Venezuelan and their melting points (Ref 6). It is clear from
crudes contain high levels of vanadium with up the table that the melting point of the ash salt
to several hundred parts per million (ppm)
(Ref 2). Sodium can originate from crudes, the
neutralizer used in crude distillation during the Table 11.1 Analysis of deposits in terms
refining process, and contamination with sea- of percent water soluble collected from the
water in transportation and storage (Ref 2). corrosion probe at metal temperatures of
During combustion, compounds formed by these 475–675 °C (890–1250 °F) at 0.4% boiler
three constituents make up a major part of the ash oxygen in a utility boiler fired with fuel oil
that deposits on the metallic components. For containing 2.65% S, 49 ppm V, and 44 ppm Na
example, superheater and reheater tubes in the Content, %

boiler, and tube hangers and other uncooled parts Constituent 1-h exposure 5-h exposure

in boilers or refinery and petrochemical furnaces. S as water soluble SO3 43.4 44.1
Na as water soluble Na2O 31.8 32.6
When the metal temperature of the component V as water soluble V2O5 10.9 6.8
reaches 540 °C (1000 °F) or higher, compounds V as water insoluble V2O5 1.0 4.1
in the ash may become molten and corrosion Fe as water insoluble Fe2O3 2.6 4.5
Ni as water insoluble NiO 1.3 2.0
attack can become severe. This mode of cor- Other metal oxides 9.0 5.9
rosion is frequently referred to as “oil-ash cor- Source: Ref 4
rosion” for boilers or furnaces fired with fuel oils.

Table 11.2 Analysis of deposits/scales

11.2 Oil-Ash Corrosion removed from the test specimens exposed in the
superheater area at temperatures from 565
During combustion of the fuel oils that contain to 850 °C (1050 to 1560 °F) in a utility boiler
vanadium, sulfur, and sodium, low-melting Constituent Content, wt %
compounds can form. These compounds are Si as silicon oxide 7.8
mixtures of vanadium pentoxide (V2O5), sodium Al as aluminum oxide 2.0
V as vanadium pentoxide 36.37
oxide (Na2O), or sodium sulfate (Na2SO4) (Ref S as sulfuric anhydride 32.80
3). Significant amounts of vanadium, sodium, Ni as nickel oxide 2.16
and sulfur were observed in the deposits col- Na as sodium oxide 5.12
Mg as magnesium oxide 2.35
lected from superheater/reheater tubes, or from P as phosphate anhydride 0.5
the uncooled tube hangers for the superheater/ Source: Ref 5
reheater tubes in boilers. The analysis of the
Name ///sr-nova/Dclabs_wip/High Temp/5208_321-334.pdf/Chap_11/ 26/10/2007 12:44PM Plate # 0 pg 322

322 / High-Temperature Corrosion and Materials Applications

deposit can vary widely, depending on compo- Kawamura and Harada (Ref 8) observed that
sition. The phase diagram for V2O5-Na2O sys- when the value of the (Na +S)/V in at.% ratio
tem showing a series of low-melting eutectic was 20 or higher, the melting point of the
compounds is shown in Fig. 11.1 (Ref 7). Low- deposits on the superheater tubes was more than
melting compounds in the V2O5-Na2SO4 system 800 °C (1470 °F). However, this melting point
are shown in Fig. 11.2 (Ref 7). Some of the could be reduced to about 500 °C (930 °F) when
eutectics become molten at 538 °C (1000 °F) or the ratio became less than 3 to 4. This is
even lower. illustrated in Fig. 11.3 (Ref 8). The analysis of the
chemical compositions (wt%) of the deposits that
Table 11.3 Melting points of some oil-ash formed on the superheater tubes when the boiler
constituents was fired with fuel oils with various ratios of
Compound Melting point, °C (°F)
(Na + S)/V showed that the amounts of S as SO3,
Ferric oxide, Fe2O3 1565 (2850)
V as V2O5, and Na as Na2O varied as a function
Ferric sulfate, Fe2(SO4)3 Decomposes at 480 of the (Na +S)/V ratios in the fuel, as shown in
(895) to Fe2O3 Table 11.4 (Ref 8).
Magnesium oxide, MgO 2500 (4530)
Magnesium sulfate, MgSO4 Decomposes at 1125 It has been well accepted that the oil-ash cor-
(2060) to MgO rosion is the result of the formation of molten
Nickel oxide, NiO 2090 (3795) vanadate compounds involving V2O5-Na2O and/
Nickel sulfate, NiSO4 Decomposes at 840
(1545) to NiO or V2O5-Na2SO4 systems. Molten vanadate
Silicon oxide, SiO2 1720 (3130) compounds flux away the oxide scales formed on
Sodium sulfate, Na2SO4 880 (1615)
Sodium bisulfate, NaHSO4 250 (480)
the metal, thus causing rapid corrosion attack.
Sodium pyrosulfate, Na2S2O7 400 (750) The corrosion mechanism is essentially hot cor-
Vanadium trioxide, V2O3 1970 (3580) rosion involving a fused salt in an oxidizing
Vanadium tetraoxide, V2O4 1970 (3570)
Vanadium pentoxide, V2O5 675 (1250) environment. The hot corrosion mechanism
Sodium metavanadate, Na2O.V2O5 630 (1165) involving fused Na2SO4 has been extensively
Sodium pyrovanadate, 2Na2O.V2O5 640 (1185) studied in high-temperature corrosion of gas
Sodium orthovanadate, 3Na2O.V2O5 850 (1560)
Nickel pyrovanadate, 2NiO.V2O5 >900 (>1650) turbine components (see Chapter 9). The sol-
Nickel orthovanadate, 3NiO.V2O5 >900 (>1650) ubility of an oxide in a fused salt can vary as the
Ferric metavanadate, Fe2O3.V2O5 860 (1580)
Ferric vanadate, Fe2O3.2V2O5 855 (1570)
melt chemistry changes. A salt can dissociate
Sodium vanadic vanadate, 625 (1160) into a base (Na2O) and acid (SO3). The melt is
Na2O.V2O4.V2O5 generally characterized by Na2O concentration
Sodium vanadic vanadate, 535 (995)
5Na2O.V2O4.11V2O5 (melt basicity). When the melt exhibits high
Source: Ref 6
solubility of an oxide formed on an alloy, a high
corrosion rate will occur. Rapp (Ref 9) provides

Fig. 11.1 Phase diagram for V2O5-Na2O system showing a Fig. 11.2 Melting points of low melting compounds in the
series of low melting eutectic compounds. Source: V2O5-Na2SO4 system during heating and cooling.
Ref 7 Source: Ref 7
Name ///sr-nova/Dclabs_wip/High Temp/5208_321-334.pdf/Chap_11/ 26/10/2007 12:44PM Plate # 0 pg 323

Chapter 11: Oil-Fired Boilers and Furnaces / 323

Fig. 11.3 Melting of the deposits formed on the superheater tubes as a function of the value of (Na + S)/ V ratio (in atomic percent) in the
fuel oils used in firing a boiler (375 MW) producing superheated steam of 570 °C (1060 °F). Source: Ref 8

Table 11.4 Chemical compositions (wt %) of the deposits formed on the superheater tubes in a boiler
when fired with fuel oils with different concentrations of vanadium
0.2–0.3% S 2.7–2.8% S 1.6–1.8% S 2.4–2.5% S
Fuel oil(a) 1–3 ppm V 45–65 ppm V 130–150 ppm V 200–250 ppm V
S as SO3 51.8 24.4 21.6 0.89
V as V2O5 0.85 30.0 49.7 83.0
Na as Na2O 34.4 17.6 17.8 2.69
Fe as Fe2O3 4.70 13.0 11.2 6.48
Ni as NiO 3.38 6.42 2.24 7.45
Ca as CaO 2.06 2.25 1.17 0.22
Mg as MgO 1.92 1.41 0.88 0.20
SO3+V2O5+Na2O 87.1 72.0 72.0 86.6
(a) Sodium in fuel oils was in a range of 8–15 ppm. Source: Ref 8

a general review on the subject of the hot cor-

rosion mechanism in fused salt. Figure 11.4
shows the solubilities of various oxides in fused
Na2SO4 at 927 °C (1700 °F) and 1.0 atm O2 as a
function of the activity of Na2O (aNa2O) (Ref 9).
Oil-ash corrosion may involve fused salt of
sodium vanadate and/or sodium sulfate. The
partial pressure of SO3 plays an important role in
the stability of the fused salt, such as sodium
vanadate, or sodium vanadate/sodium sulfate
mixture, or sulfate, formed on the metal. Seier-
sten and Kofstad (Ref 10) discuss the effect of
SO3 on vanadate-induced hot corrosion.
Fig. 11.4 The solubilities of various oxides in fused Na2SO4 at
927 °C (1700 °F) and 1.0 atm O2 (The solubility of
SiO2 is at 900 °C, or 1650 °F). Source: Ref 9
11.3 Corrosion Problems in Oil-Fired
Boilers and Furnaces uncooled components, such as tube hangers, in
oil-fired boilers are subject to much higher
Oil-ash corrosion problems in boilers are temperatures and thus can suffer more serious
typically associated with superheaters and re- corrosion problems. Many furnaces in refinery
heaters because the tube metal temperature may and petrochemical plants often use residual fuel
reach the melting point of the ash deposit. The oils for heating high-temperature processing
Name ///sr-nova/Dclabs_wip/High Temp/5208_321-334.pdf/Chap_11/ 26/10/2007 12:45PM Plate # 0 pg 324

324 / High-Temperature Corrosion and Materials Applications

equipment. The uncooled components in these waterwall tube had the dimensions of 11/8 in.
furnaces can suffer severe oil-ash corrosion outside diameter (OD) by 0.220 in. MWT.
attacks. This section discusses oil-ash corrosion Several panels that were aluminized by a com-
problems in both power boilers and refinery mercial company were tried in the boiler. A tube
furnaces. leakage developed after about 2.5 years of
service. Thus, the aluminized coating failed to
extend the life of the waterwall tube. A close-up
11.3.1 Waterwall Corrosion of Oil-Fired view of the waterwall tube showing circumfer-
Boilers ential grooves at the location near the localized
rupture area is shown in Fig. 11.5. The alumin-
Because of much lower metal temperatures, ized coating was completely destroyed. Oxi-
furnace waterwalls generally do not suffer seri- dation attack penetrated through the aluminized
ous corrosion problems in oil-fired boilers. The coating and into the steel tube wall (Fig. 11.6).
temperature of the fireside metal surface for
the furnace waterwalls for fossil-fired boilers in-
cluding oil-fired boilers in the United Kingdom
is generally limited to about 450 °C (840 °F)
(Ref 11). Reichel (Ref 12) indicates the metal
temperatures for the furnace waterwalls were in
the range of 300 to 460 °C (570 to 860 °F) for
fossil-fired boilers including oil-fired boilers in
Germany. However, in case of flame impinge-
ment, the furnace waterwall may be subjected
to higher-temperature exposure. Formation of
oxide scales or corrosion products on the internal
diameter (ID) of the waterwall tubes can also
significantly increase the outer tube metal tem-
perature. French (Ref 13) indicated that a thin
internal deposit can raise the tube metal tem-
perature into the ash-corrosion range, into the
creep-failure range, or into the rapid-oxidation
range, leading to serious furnace-tube problems.
In some cases, oil-fired boilers are used as a
“peaking” unit.* As a result, the load in these Fig. 11.5 Close-up view of the tube surface near the rupture
units is often cycled daily or weekly, thus sub- area showing numerous circumferential grooves
jecting the boiler tubes to thermal cycling. The and cracks. Courtesy of Welding Services Inc.

thermal cycling can lead to circumferential

grooving and cracking of the waterwall tubes
under corrosive conditions such as under the oil-
ash deposits and higher tube metal temperatures.
The development of circumferential grooving/
cracking on the waterwall tubes in coal-fired
boilers is discussed in Chapter 10 “Coal-Fired
Boilers.” The example that is given below shows
the circumferential grooving/cracking of the
waterwall tubes that is believed to be the result of
thermal cycling, higher metal temperatures, and
oil-ash corrosive conditions (Ref 14).
A supercritical oil-fired boiler (590 MWe)
required replacement of waterwall panels, which
were made of 2.25Cr-1Mo steel (T22), every 2
to 3 years because of severe corrosion. The
Fig. 11.6 Optical micrograph showing the oxidation pene-
trated through the aluminized coating and into the
tube wall. Note the tube inside diameter surface is visible at
* The peaking unit is a boiler that is used mainly during high- the right side bottom of the micrograph and the fire side is on the
power demand seasons, such as summer. leftside of the micrograph. Courtesy of Welding Services Inc.
Name ///sr-nova/Dclabs_wip/High Temp/5208_321-334.pdf/Chap_11/ 26/10/2007 12:45PM Plate # 0 pg 325

Chapter 11: Oil-Fired Boilers and Furnaces / 325

The remnant material of the aluminized coating 11.3.2 Superheater and Reheater
can be seen in Fig. 11.7 (area No. 5 in Fig. 11.7). Corrosion in Oil-Fired Boilers
The deposits and corrosion phases were analyzed
by scanning electron microscopy with energy To avoid the formation of molten vanadate
dispersive x-ray spectroscopy (SEM/EDX) with compounds in the ash deposits formed on
the results summarized in Fig. 11.7. A significant superheaters or reheaters in boilers, European
amount of vanadium along with magnesium, practice in general has limited the steam tem-
phosphorus, sodium, zinc, and other elements peratures to 540 °C (1000 °F) (Ref 15). This
was observed in the ash deposit and corrosion steam-outlet temperature limit implies that the
products. The preferential corrosion penetration outer tube metal temperature is not greater than
that forms a groove into the metal was essentially 580 °C (1080 °F) (Ref 11). In the United States,
iron with little sulfur and is believed to consist of oil-fired boilers ordered after 1965 have design
iron oxides. Thermal cycling of the waterwall steam-exit temperatures limited to 540 °C
caused by load changes and oil-ash corrosion (1005 °F) to prevent oil-ash corrosion problems
with possible higher waterwall tube temperatures (Ref 16). The changes in steam-exit temperatures
caused by flame impingement are believed to for oil-fired boilers constructed during the first
have caused preferential oxidation penetrations, eight decades of the 20th century are summarized
which developed into circumferential grooves in Fig. 11.8 by Paul and Seeley (Ref 16). In
and cracks. Kawamura and Harada (Ref 8) also general, oil-ash corrosion problems tend to occur
observed circumferential grooving or cracking at the secondary superheaters and reheaters
on furnace waterwall tubes made of 2.25Cr-1Mo where the metal temperatures are the highest.
steel in an oil-fired supercritical unit. For austenitic stainless steels, corrosion rates
were found to be strongly dependent on the flue
gas temperature. Figure 11.9 shows the corrosion
rates of austenitic stainless steels as a function
of the tube surface temperature with flue gas
temperatures of 800 and 1150 °C (1470 and
2100 °F) (Ref 11). Austenitic stainless steels
suffered a rapid increase in wastage rates with
increasing tube surface temperature when the
flue gas was 1150 °C (2100 °F) compared with
the 800 °C (1470 °F) flue gas. Ferritic steels, on
the other hand, were less sensitive to the flue gas
temperature.
Holland et al. (Ref 15) conducted corrosion
probe tests in an oil-fired boiler at Marchwood
Power Station for three austenitic stainless steels
(Type 316, 321, and 347) and a 12Cr ferritic steel
(12Cr, 0.5Mo, 0.25V). The boiler was operated
under base-load conditions with 0.5% O2 in flue
gas. The analysis of the fuel oil used during
testing showed 3.54% S, 91 ppm V, and 71 ppm
Fig. 11.7 Scanning backscattered electron image showing
Na. The test results are summarized in Fig. 11.10
the oxidation attack of the diffusion coating and the
substrate steel. The chemical analysis at different locations was and 11.11 (Ref 15). The results confirmed that the
performed using EDX. The results of the analysis (wt %) are sum- 12Cr ferritic steel performed much better than all
marized as:
No. 1: 23% V, 20% Ni, 17% Si, 10% Fe, 8% Mg, 9% Al, 5% P, 3% three austenitic stainless steels. Among three
Na, 2% Zn, and trace elements. austenitic stainless steels, Type 347 was found
No. 2: 53% Fe, 19% Ni, 12% Zn, 11% Al, 2% Mg, 2% V, and trace to be much better than Type 316 and 321. The
elements.
No. 3: 34% Si, 24% Al, 12% V, 8% Fe, 7% Mg, 5% Na, 4% Ca, 2% authors did not offer any explanation on the
Ni, and trace elements. performance ranking of the three austenitic
No. 4: 54% Si, 20% Al, 8% Na, 6% Fe, 3% P, 4% Ca, and trace
elements.
stainless steels. However, the results of the
No. 5: 70% Fe, 28% Al, 1% Cr, and trace elements. chemical analyses of the three alloys showed
No. 6: 74% Fe, 14% Al, 10% V, and trace elements. 18.2% Cr for Type 347, 17.1% Cr for Type 321,
No. 7: 90% Fe, 4% Cr, 2% Mo, 1% S, and trace elements.
No. 8: 90% Fe, 4% Cr, 2% Mo, 2% S, and trace elements. and 16.7% Cr for Type 316. It is, thus, believed
No. 9: 97% Fe and trace elements. that the performance of three austenitic stainless
Name ///sr-nova/Dclabs_wip/High Temp/5208_321-334.pdf/Chap_11/ 26/10/2007 12:45PM Plate # 0 pg 326

326 / High-Temperature Corrosion and Materials Applications

Steam exit temperature, °C

Steam exit temperature, °F

Fig. 11.8 Summary of the design steam-exit temperatures for oil-fired boilers built during the 20th century. Source: Ref 16

chromium. The authors compared several ferritic

steels with several austenitic stainless steels by
welding the tube samples together to form a test
tube. This test tube was then installed in the
secondary superheater tube bank in a boiler
fired with fuel oil containing 4.0% S, 0.007% Na,
0.007% V, and 0.017% Cl at Bromborough
Power Station (United Kingdom). During the
exposure test, the metal temperatures ranged 315
to 680 °C (600 to 1225 °F) for about 20,026 h.
The tube was exposed to 590 to 620 °C (1100
to 1150 °F) for about 8000 h. Test tube samples
were measured for maximum metal loss caused
by the loss in thickness by corrosion (measured
after the corrosion scales were removed by de-
scaling) and for the total metal wastage that
included maximum metal loss and intergranular
penetration (measured by metallography). Ferri-
Fig. 11.9 Corrosion rates of austenitic stainless steels and tic steels showed very little internal penetration,
ferritic steels as a function of metal temperature while austenitic stainless steels suffered some
and flue gas temperatures. Source: Ref 11
internal penetration. The data were plotted in
terms of the chromium concentration in alloys
steels followed the chromium concentration for both ferritic steels and austenitic stainless
levels in the alloys. The metal-loss data appear steels and are summarized in Fig. 11.13 (Ref 17).
to follow a linear kinetic when plotted as a More data comparing ferritic steels with aus-
function of exposure time, as shown in Fig. 11.12 tenitic stainless steels were generated by Parker
(Ref 15). This figure clearly shows the superior et al. (Ref 4) using corrosion probes tested in a
performance of the 12Cr ferritic steel compared boiler fired with fuel oil containing 2.65% S, 49
with austenitic stainless steels containing 16 to ppm V, and 44 ppm Na at Marchwood Power
18% Cr. Station (United Kingdom). Tube metal loss data
Alexander et al. (Ref 17) found the similar test for ferritic steels and austenitic stainless steels are
results showing ferritic steels with less chromium summarized in Fig. 11.14 and 11.15 (Ref 4). The
being significantly more resistant to oil-ash cor- test results showed both 9Cr and 12Cr ferritic
rosion than austenitic stainless steels with more steels were significantly more corrosion resistant
Name ///sr-nova/Dclabs_wip/High Temp/5208_321-334.pdf/Chap_11/ 26/10/2007 12:45PM Plate # 0 pg 327

Chapter 11: Oil-Fired Boilers and Furnaces / 327

than austenitic stainless steels including Type in a boiler at Bromborough Power Station,
310 (25Cr-20Ni). It is somewhat surprising to showing that increasing chromium concentration
find that 2.25Cr-1Mo steel was much more significantly increased the corrosion resistance
resistant than austenitic stainless steels (Fig. for ferritic test steels 2.25Cr-1Mo, 9Cr, and 12Cr
11.14). It is more surprising to find that 2.25Cr- steels (Fig. 11.13). Among the four austenitic
1Mo steel was comparable to 9Cr and 12Cr steels stainless steels, Type 310 (25Cr-20Ni) was more
at approximately 550 to 625 °C (1020 to resistant than Type 316, 321, and 347 due to
1155 °F). This was in contrast with test results higher chromium content. The test results gen-
that were generated by Alexander et al. (Ref 17) erated at the upper furnace are summarized

Fig. 11.10 Results of corrosion probe tests for 1000 h as a function of the tube metal temperature in a boiler fired with fuel oil
containing 3.54% S, 91 ppm V, and 71 ppm Na. Source: Ref 15

Fig. 11.11 Results of corrosion probe tests for 2000 h as a function of the tube metal temperature in a boiler fired with fuel oil
containing 3.54% S, 91 ppm V, and 71 ppm Na. Source: Ref 15
Name ///sr-nova/Dclabs_wip/High Temp/5208_321-334.pdf/Chap_11/ 26/10/2007 12:45PM Plate # 0 pg 328

328 / High-Temperature Corrosion and Materials Applications

in Fig. 11.15 (Ref 4), again showing both ferri- The plant exposure tests have shown that 9Cr
tic steels were more resistant than austenitic and 12Cr ferritic steels are much more resistant to
stainless steels. In this upper furnace test, 2.25Cr- oil-ash corrosion than austenitic stainless steels,
1Mo steel was not included. such as Type 316, 321, and 347, typically used

Exposure time, h

Fig. 11.12 Metal loss as a function of exposure time for 12Cr ferritic steel and three austenitic stainless steels. Source: Ref 15

Fig. 11.13 Tube wastage data, which included both metal loss (tube thickness loss) and total wastage (metal loss+intergranular
penetration), for ferritic steels and austenitic stainless steels in terms of chromium concentration in alloys. The data were
generated in a boiler at Bromborough Power Station (United Kingdom) fired with fuel oil containing 4.0% S, 0.007% Na, 0.007% V, and
0.017% Cl. Open data points: total wastage (maximum metal loss+intergranular penetration). Solid data points: Maximum metal loss.
Source: Ref 17
Name ///sr-nova/Dclabs_wip/High Temp/5208_321-334.pdf/Chap_11/ 26/10/2007 12:45PM Plate # 0 pg 329

Chapter 11: Oil-Fired Boilers and Furnaces / 329

for superheaters and reheaters in utility boilers. cladding for providing corrosion resistance offer
Nevertheless, the ferritic steels may not have better alternatives for higher-temperature appli-
adequate creep strengths at higher temperatures. cations. Bolt (Ref 18) evaluated some of the
Composite tubes with austenitic stainless steels composite tubes in an oil-fired experimental
for providing materials strength and outer tube boiler heated with fuel oil containing 2.2% S, 200
ppm V, and 50 ppm Na at metal temperatures
from 500 to 700 °C (930 to 1290 °F) and two
flue gas temperatures 1000 and 1125 °C (1830
and 2060 °F). The tests were conducted in
corrosion probes for 2000 h of exposure with a
continuous load (i.e., constant metal tempera-
ture) and a discontinuous load (varying tem-
peratures—8 h at higher and 16 h at lower levels).
Four coextruded composite tubes tested were
Esshete 1250/Type 310, 800H/Type 446, 1714
CuMo/35CrA, and 800H/671; Type 310, Type
446, 35CrA, and 671 were cladding alloys. The
35CrA cladding was 35Cr-45Ni-Fe alloy, and
the 671 cladding was 46Cr-Ni alloy. The test
results are summarized in Fig. 11.16 and 11.17.
A bell-shaped curve with the maximum attack at
about 650 °C was observed for Type 347 for both
continuous and discontinuous loads. Type 347
and 310 showed unacceptable corrosion rates (>1
mm/yr, or >39 mpy) at 630 to 675 °C (1165
to 1250 °F). Type 446 cladding appeared to be

Fig. 11.14 Maximum metal loss as a function of metal tem-

perature tested for 10,000 h in the superheater
inlet zone in a boiler fired with fuel oil containing 2.65% S, 49
ppm V, and 44 ppm Na. Source: Ref 4

Type 347H

Fig. 11.16 Corrosion rates of Type 347 in comparison

Fig. 11.15 Maximum metal loss as a function of metal tem- with 310, 446, 35CrA, and 671 claddings as a
perature tested for 10,000 h in the combustion function of temperature (°C) with flue gas temperatures of 1000
zone in a boiler fired with fuel oil containing 2.65% S, 49 ppm V, and 1125 °C (1830 and 2060 °F) in a constant thermal load.
and 44 ppm Na. Source: Ref 4 Source: Ref 18
Name ///sr-nova/Dclabs_wip/High Temp/5208_321-334.pdf/Chap_11/ 26/10/2007 12:45PM Plate # 0 pg 330

330 / High-Temperature Corrosion and Materials Applications

sufficiently corrosion resistant. Both 35CrA and it is believed that the observed carburization
671 claddings performed extremely well even was not caused by oil-ash deposits. Internal
though alloy 35CrA contained about 45% Ni and carburization can occur in CO2-containing
alloy 671 contained about 54% Ni with 35% Cr environments. McNallan et al. (Ref 21) observed
in the former and 46% Cr in the latter. internal carburization of Type 310 in Ar-20CO2
Carburization has been observed at times in at 800 °C (1470 °F) for 24 h. Furthermore, it is
superheaters and reheaters. Lopez-Lopez et al. believed that the observed carburization was not
(Ref 19) found severe carburization in a heavily the major factor causing the severe superheater
corroded Type 321H superheater tube after wastage problem. Carburization was often ob-
107,000 h of service in a boiler that had been served to take place after the alloy suffered severe
used as a “peaking” unit. The tube had suffered wastage with no protective oxide scales on the
severe metal wastage. However, underneath the corroded surface.
corroded surface was a shallow, severely car- In addition to selecting a better-performing
burized layer about 80 µm deep and a peak alloy to resist oil-ash corrosion, methods that
hardness of about 400 HV (Vickers hardness could reduce superheater and reheater corrosion
number). In this study, the authors failed to report include (a) decreasing excess air for combustion
(or analyze) the chemistry of the ash deposits and (b) the use of additives to raise the oil-ash
including carbon collected from the superheater melting point. Excess air promotes the formation
tube. In a separate study, Wong-Moreno et al. of SO3 and of V2O5, which melts at 675 °C
(Ref 20) analyzed nine different ash deposits (1250 °F), instead of V2O3 or V2O4, which melts
collected from superheaters in different oil-fired at 1970 °C (3580 °F) (Ref 6). However, keeping
boilers ranging from 84 to 350 MW. Most de- very low excess air for combustion can probably
posits contained about 0.03 to 0.06% C, with be difficult to maintain in practice. The Central
only two deposits showing about 0.14% C. This Electricity Generating Board (CEGB) in the
level of carbon in the deposits is unlikely to cause United Kingdom claimed to have made sig-
severe carburization in austenitic stainless steels nificant achievements in combustion in oil-fired
at the superheater tube temperatures. Therefore, boilers with low excess air (Ref 22). Boilers at
Merksem Power Station were reported to have
Discontinuous load, operated at 1% excess air (0.2% O2) as a regular
operating practice (Ref 23).
The use of additives to increase the melting
Discontinuous load, point of the vanadate or sulfate is likely to be
a more practical approach. Useful additives in-
clude (Ref 8):
Type 347H
 Magnesium compounds: MgO, Mg(OH)2,
MgCO3, MgSO4, MgCO3·CaCO3, organic
magnesium compounds
 Calcium compounds: CaO, Ca(OH)2, CaCO3
 Barium compounds: BaO, Ba(OH)2
 Aluminum and silicon compounds: Al2O3,
Fe SiO2, 3Al2O3·2SiO2

The additive reacts with vanadium compounds to

form reaction products with higher melting
points. When magnesium compounds are used,
some of the reaction products and their melting
points are:

 MgO·V2O5: 671 °C (1240 °F)

 2MgO·V2O5: 835 °C (1535 °F)
 3MgO·V2O5: 1191 °C (2175 °F)
Fig. 11.17 Corrosion rates of Type 347 in comparison with
310, 446, 35CrA, and 671 claddings as a function
of temperature (°C) with flue gas temperatures of 1000 and
With injection of magnesium compounds into
1125 °C (1830 and 2060 °F) in varying loads. Source: Ref 18 the fuel, the higher Mg/V ratios increase the
Name ///sr-nova/Dclabs_wip/High Temp/5208_321-334.pdf/Chap_11/ 26/10/2007 12:45PM Plate # 0 pg 331

Chapter 11: Oil-Fired Boilers and Furnaces / 331

melting point of the oil-ash deposits (Fig. 11.18) heater tubes used to process hydrocarbon fluid
(Ref 6). Increasing the melting point of oil-ash in the furnace are uncooled and thus can be
deposits would result in lowering the corrosion subjected to temperatures as high as 900 °C
rates. Figure 11.19 shows the effectiveness of the (1650 °F) or higher. Severe materials problems
Mg(OH)2 additive injection in reducing the cor- due to oil-ash corrosion were illustrated by
rosion rate of Type 321 superheater tubes (Ref 8). numerous case histories presented in a 1958
Disadvantages of the additive injection approach NACE Technical Committee Report (Ref 25).
include additional operating costs and a For example, Type 309 tube supports suffered a
substantial increase in ash volume that may corrosion rate of about 12.7 mm/yr (500 mpy) in
require additional furnace downtime for tube an oil-fired heater in a refinery plant (Ref 25).
cleaning (Ref 24). This heater increased the temperature of the oil
charge to 350 to 470 °C (670 to 880 °F) in a
furnace operating at 940 to 980 °C (1700 to
11.3.3 Tube Supports or Hangers in Boilers
1800 °F) fired with residual fuel containing
and Refinery/Petrochemical Furnaces about 4.1% S and 0.05% ash composed of about
16.2% V2O5. Another example involving
Many refinery and petrochemical furnaces are
uncooled superheater spacers made of cast HH
fired with residual fuel oils containing significant
stainless steel (25Cr-12Ni) failed after only
amounts of sulfur, vanadium, and sodium.
7 months of service in a boiler fired with a
Structural supports and tube hangers for the
Buncker “C” oil, producing 565 °C (1050 °F)
steam (Ref 25). The corrosion of the HH tube
spacers was found to proceed by sulfidation/
oxidation attack (Ref 25).
McDowell et al. (Ref 5) conducted field test
racks in a utility boiler fired with Buncker “C”
fuel oil. The test racks were exposed to the
flue gas in the second bank of the superheater.
During testing, the boiler operated at full load
during the day for 5 days a week and at half
and low loads during early morning hours and
weekends. The temperature of the flue gas was
about 850 °C (1560 °F) during full load of
power, 660 °C (1220 °F) during half load, and
565 to 590 °C (1050 to 1100 °F) during mini-
mum loads. No fuel analysis was reported in the
paper. Analysis of the deposits/corrosion scales
showed a significant amount of V2O5 (Table
11.2). The test results, which were extrapolated
Fig. 11.18 Effect of MgO addition on the melting point of oil- from three test racks exposed for 504, 648, and
ash deposit on superheater tubing of an oil-fired 707 h, respectively, are summarized in Table
boiler. Source: Ref 6 11.5. None of the alloys tested showed accept-
able wastage rates.
More rack tests in boilers fired with Buncker
“C” oils containing high concentrations of
vanadium (150 to 450 ppm) were reported by
McDowell and Mihalisin (Ref 26). Test racks
were exposed to the flue gas in the superheater
section. Alloys ranging from low-alloy steels
to iron- and nickel-base alloys suffered severe
corrosion attack. The results of one test rack are
shown in Table 11.6 (Ref 26). Specimens
included 5Cr steel, stainless steels (both 400 and
300 series), Fe-Ni-Cr alloys, Ni-Cr-Fe alloy, and
Fig. 11.19 Corrosion of Type 321 superheater tubes with and
50Ni-50Cr alloy, along with two cast stainless
without Mg(OH)2 injection in an oil-fired boiler.
Source: Ref 8 steels (HE and HH alloys). All the alloys tested
Name ///sr-nova/Dclabs_wip/High Temp/5208_321-334.pdf/Chap_11/ 26/10/2007 12:45PM Plate # 0 pg 332

332 / High-Temperature Corrosion and Materials Applications

exhibited unacceptable corrosion rates. Even the hangers and tube supports made of cast HH alloy
best performer (50Ni-50Cr alloy) suffered a (25Cr-12Ni steel) suffered severe corrosion
corrosion rate of 3.1 mm/yr (121 mpy) (Ref 26). attack. Metal temperatures were in the range of
Some investigators have reported excellent 730 to 890 °C (1350 to 1630 °F). The highest
performance of “50Cr-50Ni” alloy as tube sup- corrosion rates of cast HH alloy were 6.4 to 9.5
ports in furnaces or boilers fired with fuel oils mm/yr (250 to 375 mpy). Replacements with
containing sulfur and vanadium. Spafford (Ref alloy 657 (a cast 50Ni-50Cr alloy) were reported
27) reported good performance of the 50Ni-50Cr to perform very well, with minimal maintenance
alloy in refinery heaters for coking and catalytic and repair (Ref 27).
reformer units. The heaters were fired with heavy Swales and Ward (Ref 2) reported the results
fuel oil containing 2.5 to 4% S and 50 to 70 ppm of a field test that showed alloy 657 performed 10
V (occasionally up to 150 ppm). The original times better than HH and HK alloys, as illustrated
in Fig. 11.20. There are additional cases where
Table 11.5 Results of field rack tests conducted alloy 657 tube supports provided excellent per-
in the second bank of the high-temperature formance; two such examples (Ref 2) are given
superheater at the location about 3 ft from the below. The authors also concluded that at tem-
brick side wall in a utility boiler fired with peratures higher than 900 °C (1650 °F), alloy
Buncker “C” fuel oil with flue gas stream 657 often suffered severe corrosion attack, but
fluctuating from 565 to 850 °C (1050 to 1560 °F) provided good protection up to that temperature
while the boiler changed from half to full loads (Ref 2).
during the exposure test Catalytic Reformer Heaters in a Refinery.
Alloy Wastage rate, mm/yr (mpy) The furnace was fired with a mixture of fuel oil
Carbon steel 17.7 (695) (about 50%) and gas. The ash analysis showed
2.25Cr-1Mo 17.0–19.1 (670–750)
502 (5Cr steel) 6.2–14.2 (244–557)
48% V2O5 and 2% Na2O. Tube supports were
410 (12Cr steel) 14.1 (555) exposed to temperatures up to 900 °C (1650 °F).
406 (12Cr-3Al steel) 2.7 (108) The tube supports made of cast HH stainless steel
430 9.5–26.9 (374–1060)
446 4.8 (189) (25Cr-12Ni) suffered severe corrosion after
302 10.3 (406) 2 years of service. Tube supports were replaced
321 17.5 (690) with cast alloy 657 (Ni-48Cr) showing no note-
309 6.2–7.7 (244–305)
800 5.5 (217) able corrosion after 7 years of service.
HW (12Cr-60Ni) 5.8–15.3 (230–601)
HF (21Cr-9Ni) 10.7 (422)
HE (28Cr-10Ni) 3.1 (124)
Inconel 600 (Ni-15Cr-7Fe) 2.9 (113)
Source: Ref 5

Table 11.6 Results of a field test (uncooled

specimens) exposed in the superheater section
at 815 °C (1500 °F) in a boiler fired with high
vanadium (150–450 ppm) Buncker “C” fuel
Alloy Wastage rate, mm/yr (mpy)
5Cr steel 32–45 (1270–1775)
406 5.7 (224)
431 17–23 (655–925)
446 3.8 (149)
302B 14–16 (533–644)
309 5.0 (196)
321 9–13 (346–505)
347 6.2 (243)
310 4.7 (187)
Incoloy(a) 9–12 (364–458)
Incoloy 804 13–18 (495–710)
Inconel(a) 5.0 (196)
50Cr-50Ni 3.1 (121)
HE 4.8 (187)
HH 12–16 (467–645)
(a) Incoloy may refer to Incoloy alloy 800, while Inconel may refer to Inconel alloy Fig. 11.20 Results of a field test at 700 °C (1290 °F) for
600. Source: Ref 26 6 months in a refinery (crude oil) heater com-
paring alloy 657 to HH and HK alloys. Source: Ref 2
Name ///sr-nova/Dclabs_wip/High Temp/5208_321-334.pdf/Chap_11/ 26/10/2007 12:45PM Plate # 0 pg 333

Chapter 11: Oil-Fired Boilers and Furnaces / 333

Catalytic Reformer Heater in a Refinery. 4. J.C. Parker, D.F. Rosborough, and M.J. Virr,
The furnace was fired with fuel oil containing High Temperature Corrosion Trials at
3 to 4% S, 40 to 50 ppm V (90 to 150 ppm Marchwood Power Station—10,000 Hour
occasionally for short periods). Tube hangers Corrosion Probe Trials, J. Inst. Fuel, Feb
were exposed to 700 to 900 °C (1290 to 1972, p 95
1650 °F). The original HH cast roof tube hangers 5. D.W. McDowell, R.J. Raudebaugh, and
suffered severe wastage problems after 1 to W.E. Somers, High-Temperature Corrosion
2 years of service. All tube hangers were then of Alloys Exposed in the Superheater of
replaced with alloy 657. No failures occurred an Oil-Fired Boiler, Trans. ASME, Feb 1957,
after more than 4 years of service for alloy 657 p 319
hangers. 6. M. Fichera, R. Leonardi, and C.A. Farina,
Fuel Ash Corrosion and Its Prevention with
MgO Addition, Electrochim. Acta, Vol 32
(No. 6), 1987, p 955
11.4 Summary 7. W.T. Reid, External Corrosion and Deposits
—Boiler and Gas Turbines, Elsevier Pub-
Fireside corrosion can present a serious prob- lishing, 1971
lem in oil-fired boilers or refinery/petrochemical 8. T. Kawamura and Y. Harada, “Control of
furnaces fired with low-grade fuels with high Gas Side Corrosion in Oil Fired Boilers,”
concentrations of vanadium, sulfur, and sod- Mitsubishi Technical Bulletin No. 139,
ium. This corrosion is frequently referred to as Mitsubishi Heavy Industries, Ltd., May
“oil-ash corrosion.” Accelerated attack by oil-ash 1980
corrosion is related to the formation of low- 9. R.A. Rapp, Hot Corrosion of Materials, Pure
melting point molten vanadium pentoxide and Appl. Chem., Vol 62 (No. 1), 1990, p 113
sodium sulfate eutectics, which flux the pro- 10. M. Seiersten and P. Kofstad, The Effect of
tective oxide scale from the metal surface. In SO3 on Vanadate-Induced Hot Corrosion,
boilers, superheater and reheater tubes are sus- High Temp. Technol., Vol 5 (No. 3), 1987,
ceptible to oil-ash corrosion attack. Uncooled p 115
components in the boilers, such as tube supports 11. A.J.B. Cutler, T. Flatley, and K.A. Hay, Fire-
and spacers, can suffer severe corrosion attack Side Corrosion in Power Station Boilers,
because of higher temperatures. Oil-ash corro- Metall. Mater. Technol., Feb 1981, p 69
sion can also occur in refinery and petrochemical 12. H.H. Reichel, Fireside Corrosion in German
furnaces burning low-grade fuels. The resistance Fossil-Fuel Fired Power Plants: Appearance,
of oil-ash corrosion for various alloys in both Mechanism and Causes, Werkst. Korros.,
boilers and refinery/petrochemical furnaces is Vol 39, 1988, p 54
reviewed. 13. D.N. French, Metallurgical Failures in
Fossil Fired Boilers, 2nd ed., John Wiley &
Sons, 1993, p 370
REFERENCE 14. Welding Services Inc. unpublished data
15. N.H. Holland, D.F. O’Dwyer, D.F. Rosbor-
1. S.C. Stultz and J.B. Kitto, Ed., Steam and Its ough, and W. Wright, High-Temperature
Generation and Use, 40th ed., Babcock & Corrosion Investigations on an Oil-Fired
Wilcox, Barberton, OH, 1992 Boiler at Marchwood Power Station, J. Inst.
2. G.L. Swales and D.M. Ward, Strengthened Fuel, May 1968, p 206
50% Chromium, 50% Nickel Alloy (IN657) 16. L.D. Paul and R.R. Seeley, Oil Ash
Refinery Heater Tube Supports to Combat Corrosion—A Review of Utility Boiler
Fuel Ash Corrosion—A Review of Service Experience, Paper No. 267, Corrosion/90,
Caswe Histories, Paper No. 126, Corrosion/ NACE International, 1990
79, NACE, 1979 17. P.A. Alexander, R.A. Marsden, J.M. Nelson-
3. D.N. French, Corrosion of Superheaters Allen, and W.A. Stewart, Operational Trials
and Reheaters in Fossil Fired Boilers, of Superheater Steels in a C.E.G.B. Oil-
Environmental Degradation of Engineering Fired Boiler at Bromborough Power Station,
Materials in Aggressive Environments, J. Inst. Fuel, Feb 1964, p 59
Conf. Proc., Virginia Polytechnic Institute, 18. N. Bolt, Fireside Corrosion Phenomena in
Blacksburg, VA, 1981, p 407 Superheaters of Coextruded Materials with
Name ///sr-nova/Dclabs_wip/High Temp/5208_321-334.pdf/Chap_11/ 26/10/2007 12:45PM Plate # 0 pg 334

334 / High-Temperature Corrosion and Materials Applications

18, 25, 35 and 50% Cr at Metal Tempera- Air,” CEGB Report RD/H/MI, Central
tures of up to 700 °C, International Con- Electricity Generating Board
gress on Metallic Corrosion, Conf. Proc., 23. J. Remeysen, “Operation of Large Boilers at
Vol IV, Oxford & IBH Publishing Co. Pvt. Very Low Excess Air Levels,” Proceedings
Ltd, 1987, p 3593 of Brussels Conference of Institute of Fuel,
19. D. Lopez-Lopez, A. Wong-Noreno, and Paper 1, Sept 1964
L. Martinez, Unusual Superheater Tube 24. J.R. Wilson, Understanding and Preventing
Wastage Associated with Carburization, Fuel Ash Corrosion, Paper No. 12, Corro-
Mater. Perform., Dec 1994, p 45 sion/76, NACE, 1976
20. A. Wong-Moreno, Y.M. Martinez, and L. 25. The Present Status of the Oil Ash Corrosion
Martinez, High Temperature Corrosion Problem, NACE Technical Committee
Enhanced by Residual Fuel Oil Ash Report, Pub. 58-11, Corrosion, 1958,
Deposits, Paper No. 185, Corrosion/94, p 369t
NACE International, 1994 26. D.W. McDowell, Jr. and J.R. Mihalisin,
21. M.J. McNallan, S. Thongtem, J.C. Liu, Y.S. Paper No. 60-WA-260, presented at ASME
Park, and P. Shyu, Corrosion of Chromium Winter Annual Meeting, Nov 27–Dec 2,
Containing Alloys in Non-steady State 1960
Environments Containing Oxygen, Carbon, 27. B.F. Spafford, in UK Corrosion ’82, Conf.
and Chlorine, J. Phys. IV, Coll. C9, Suppl. J. Proc. (Birmingham, U.K.), Nov 15–17
Phys. III, Vol 3, Dec 1993, p 143 1982, Institution of Corrosion Science &
22. E.M. Hamilton and G.R. Stern, “Operation Technology, Birmingham, U.K., 1982,
of Oil Fired Boilers with Very Low Excess p 67
Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/ 30/10/2007 4:09PM Plate # 0 pg 335

High-Temperature Corrosion And Materials Applications Copyright © 2007 ASM International®

George Y. Lai, editor, p335-358 All rights reserved.
DOI: 10.1361/hcma2007p335 www.asminternational.org

CHAPTER 12

Waste-to-Energy Boilers and

Waste Incinerators

12.1 Introduction chlorine (Ref 4). Rubbers such as Hypalon,

chloroprene, and neoprene also contain high
Municipal solid waste (MSW) is a combina- concentrations of chlorine (Ref 4). Other waste
tion of residential, commercial, and industrial constituents that can contribute to corrosive
refuse. Historically, the typical means of dis- environments under combustion are batteries
posal of this refuse has been landfilling. A novel (automobile batteries, button-type batteries in
approach in recovering the energy from this watches and calculators, alkaline, zinc, nickel,
“resource” in boilers for electricity generation and cadmium batteries) and consumer electro-
and minimizing the landfill requirements began nics. Batteries contribute lead and cadmium;
in the 1960s in Europe. It became apparent household batteries contribute cadmium; and
from these early European boilers that corro- consumer electronics contribute both lead and
sion of the boiler tube materials was a major cadmium (Ref 5). Rademakers et al. (Ref 6)
obstacle for operating these so-called waste- reported their calculation of the heavy metal
to-energy (WTE) boilers (Ref 1, 2). Managing composition of Dutch solid waste based on 1994
the corrosion problems continues to be a big data in g/kg waste: 0.01 Cd, 0.0004 to 0.0006 Hg,
challenge for operators of modern boilers 0.22 to 0.56 Pb, 0.02 to 0.04 Sb, 0.63 to 1.04 Zn,
worldwide. This chapter focuses on the corrosion 0.01 As, 3.96 to 6.89 Cl, 0.11 to 0.19 Br, 0.1 to
issues related to boilers burning municipal solid 0.2 F, and 1.48 to 2.96 S. During combustion,
waste. these heavy metals can react with chlorine to
form some low-melting-point chlorides, thus
causing severe boiler tube material corrosion
problems. Corrosion mechanisms are discussed
12.2 Fuels, Combustion Environments, in the next section.
and Boilers There are primarily two types of combustion
technologies, namely, mass-burning and refuse-
Municipal solid waste typically contains paper derived fuel (RDF) burning. For mass-burning
and paperboard, plastics, rubber, textile, leather, units, the fuel, which is not segregated except for
batteries, food waste, yard waste, metal, glass, appliances, furniture, and other large articles,
and other miscellaneous materials. This type burns in the as-received condition. In RDF units,
of fuel is very heterogeneous and varies greatly on the other hand, the fuel is segregated, classi-
by geographic location, country, and “culture.” fied, and shredded to size. Metals and glass are
In North America, typical constituents of typically removed from RDF fuel prior to com-
the MSW in 1990 were paper and paperboard bustion. Thus, RDF fuel has a higher heating
(34.2–40.0%), plastics (7.2–9.2%), food waste value than that of a mass-burning unit. A small
(7.3–8.5%), yard waste (17.6–19.9%), metal number of boilers in use today are based on the
(8.1–9.6%), glass (7.7–9.7%), and other (10.3– fluidized-bed technology. The mass-burning unit
13.2%) (Ref 3). Polyvinyl chloride (PVC) plastic uses a mechanical stoker to feed the fuel onto a
is a dominant source of chlorine, which is a major grate where the fuel is combusted. For most RDF
corrosive constituent making the combustion units, the fuel is blown into the furnace sidewalls
environment corrosive to boiler tube materials. through an RDF burner with a fuel-distribution
Polyvinyl chloride contains about 36 wt% impeller (Ref 3). Most RDF units burn the fuel in
Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/ 30/10/2007 4:09PM Plate # 0 pg 336

336 / High-Temperature Corrosion and Materials Applications

suspension (Ref 3). RDF units also employ boilers developed in the 1960s were designed
moving grates. to operate at higher steam pressures and tem-
Combustion gas temperatures are generally peratures, particularly those in Germany, with
at or below approximately 1090 °C (2000 °F) in one in Mannheim operating at 12.5 MPa/530 °C
mass-burning units and approximately 1315 to (1800 psig/980 °F) steam and another one in
1370 °C (2400 to 2500 °F) in RDF units (Ref 7). Munich at 18 MPa/540 °C (2650 psig/1000 °F)
The guidelines for solid waste combustion steam (Ref 1). Higher steam temperatures result
introduced by the European authorities require in higher tube metal temperatures and thus more
that the temperature should be above 850 °C serious corrosion problems.
(1560 °F) for a 2 s residence time at the level Combustion of the MSW fuel is significantly
1 m above the secondary or tertiary air (Ref 6). different from that of coal or oil because the solid
The MSW boilers operate at much lower steam waste is a heterogeneous fuel. In addition, the
pressures and temperatures compared with coal- fuel contains numerous impurities that include
fired utility boilers, as discussed in Chapter 10. chlorine, sulfur, sodium, potassium, cadmium,
Kubin (Ref 7) reported that Ogden Martin boilers zinc, lead, and other heavy metals. Because of
in the United States typically operated at steam these impurities, combustion of this fuel gen-
pressures of 4 to 6 MPa (615 to 880 psig) and erates a very corrosive environment that causes
temperatures of 370 to 440 °C (700 to 830 °F) serious corrosion problems for boiler tube
for mass-burning units, and steam pressures of materials. Gaseous combustion products include
4.5 to 6 MPa (665 to 900 psig) and temperatures N2, O2, CO2, H2O, SO2, HCl, HF, and other
of 370 to 440 °C (700 to 830 °F) for RDF units. gaseous impurities such as CO and HBr. These
These pressures and temperatures appear to be gaseous constituents are often measured by
representative of other WTE boilers in the United plant operators. Examples of these flue gas
States. For example, two boilers at the Wheel- compositions measured in mass-burning units at
abrator Concord facility operate at a steam pres- different plants are shown in Table 12.1 (Ref 7,
sure of 650 psig (4.5 MPa) and a temperature of 11–13). However, vapors of metal chlorides and
400 °C (750 °F), and two at the Wheelabrator sulfates are also produced during combustion.
Spokane facility operate at 6.2 MPa (900 psig) These compounds are normally not quantified by
and 440 °C (830 °F) steam (Ref 8), with some plant personnel. Many of these metal chlorides
other Wheelabrator boilers operating at higher exhibit high vapor pressures and/or low melting
pressures and temperatures (Ref 9). Some of the points. Some physical properties of many metal
European boilers operate at higher steam pres- chlorides can be found in Chapter 6 “Corrosion
sures and temperatures. For example, the boilers by Halogen and Halides.” As is discussed in this
at Ivry Paris (France) operate at 7.5 MPa chapter, some of these metal chlorides are pri-
(1103 psig, or 75 bar) and 480 °C (900 °F) marily responsible for the corrosion of the boiler
steam, at HKW Mannheim (Germany) at 8 MPa tube materials.
(1176 psig, or 80 bar) and 500 °C (930 °F) The furnace is typically enclosed by four walls
steam, at BSR Berlin (Germany) at 7.5 MPa (rear, front, and two side walls) using a tube-
(1088 psig, or 74 bar) and 420 °C (790 °F) membrane construction (i.e., individual tubes
steam, and at EVO Oberhausen (Germany) are connected by narrow plates). These tube-
at 7 MPa (1029 psig, or 70 bar) and 480 °C membrane walls, which are commonly referred
(896 °F) steam (Ref 10). Early European WTE to as “waterwalls,” provide heat-absorbing

Table 12.1 Examples of flue gas compositions generated by different WTE mass-burning boilers
Composition, vol%
Gas Ogden-Martin (US)(a) Saint Ouen (France)(b) Richtlinien (Germany)(c) Kuririn (Japan)(d)
O2 8.5–9.5 9–10 NR 9.1
CO2 8–9 9–12 NR 11.1
SO2 100–200 ppm 90–130 ppm 100–2000 ppm 40 ppm
HCl 400–600 ppm 600–1200 ppm 560–2240 ppm 810 ppm
HF 5–20 ppm 20–40 ppm … …
CO … <20 ppm 64–640 ppm 42 ppm
H 2O 13–15 NR NR 21.4
N2 bal bal NR bal
NR, not reported. (a) Source: Ref 7. (b) Source: Ref 11. (c) Source: Ref 12. (d) Source: Ref 13
Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/ 30/10/2007 4:09PM Plate # 0 pg 337

Chapter 12: Waste-to-Energy Boilers and Waste Incinerators / 337

surfaces for converting the water inside the tubes Figure 12.1 shows a schematic of a mass-
to steam as the water rises from the bottom of the burning unit. For some mass-burning units,
waterwall. At the top of the waterwall tubes, a screen tubes are installed in front of the super-
mixture of steam and water leaves the waterwall heater to lower the temperature of the flue gas
tubes and enters the steam drum where steam is entering into the superheater section. In some
separated from water. The furnace waterwall other mass-burning units, there are additional gas
typically operates with water at a saturation passes (one or more additional walls to allow flue
temperature. Thus, the temperature of the water/ gas to pass through in the convection path) to
steam in the waterwall tubes, which depends on allow the flue gas stream to lower its temperature
the pressure of the water/steam, is generally at or before entering into the superheater section.
below 277 °C (530 °F) (Ref 14). The ranges of This type of design is shown schematically in
the waterwall tube metal temperatures were cited Fig. 12.2 (Ref 17). Licata et al. (Ref 17) indicated
to be 260 to 293 °C (500 to 560 °F) (Ref 7), 260 that passing flue gas through two 180° turns and
to 290 °C (500 to 550 °F) (Ref 15), and 260 to one 90° turn prior to entering the superheaters
315 °C (500 to 600 °F) (Ref 16). For higher- can not only lower the flue gas temperature but
pressure boilers, the temperature of the steam also remove particles from the flue gas stream. In
will be higher. The steam from the steam drum addition, the flue gas stream can achieve better
after separating from water is then further heated mixing to minimize local reducing conditions
in superheaters (typically two superheaters— (Ref 17). Figure 12.3 shows a schematic of an
primary and secondary or final superheaters) to RDF unit. Carbon steels and sometimes low-
higher temperature in the convection path before alloy steels are typical construction materials
it is delivered to turbines for electricity gener- for waterwalls, screen tubes, the boiler bank,
ation. The combustion flue gas exits from the and economizer. For superheaters, carbon and
furnace at the top and then enters into the con- low-alloy steels are typically the materials of
vection path, flowing typically through the construction.
superheater, then the boiler bank, and the eco- This chapter focuses mainly on waterwalls
nomizer. The flue gas temperature entering into and superheaters that have experienced severe
the superheater may vary from approximately corrosion problems. Also discussed are the
650 to 900 °C (1200 to 1650 °F). latest protection methods against corrosion and

Refuse
feed
hopper

Boiler

Refuse Refuse
receiving area pit

Plunger ash
extractor
Ash removal

Fig. 12.1 Mass-burning unit, with a grate where the fuel burns and the superheater right above the arch (or bull nose) in the upper
furnace, followed by a boiler bank and an economizer in the convection pass downstream of the superheater. Source: Ref 3
Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/ 30/10/2007 4:10PM Plate # 0 pg 338

338 / High-Temperature Corrosion and Materials Applications

13
Drum

Radiant SH. SH. Evap. Econ.

pass 2 1
1 2ND 3RD
pass pass

6 14
10
2
5
4
4

12
1 Refuse feed hopper
3
2 Refuse chute
A 11 3 Refuse incineration grate
B 4 Secondary air supply
C 5 Furnace
D 6 Auxiliary burner
E
15 7 Primary air hopper
7 8 Ash extractor
9 Scraper conveyor
10 Fly-ash hopper
9 11 Fly-ash conveyor
8 12 Boiler ash
13 Superheater steam outlet
14 Boiler feed water inlet
15 Primary air supply

Fig. 12.2 Mass-burning unit of different design involving multiple passes for the flue gas stream before entering the superheaters. Also
shown is the grate where the fuel is combusted. Source: Ref 17

erosion/corrosion at these two critical areas in the

boiler.

12.3 Boiler Tube Fireside Corrosion

Problems
Prepared
fuel
storage Materials problems and high wastage rates
experienced by furnace waterwalls and super-
Boiler heaters in numerous waste-to-energy plants
worldwide have been extensively reviewed by
Krause (Ref 1), Sorell (Ref 2), and Krause and
Wright (Ref 18). Many of the early boilers, par-
ticularly European boilers, suffered severe boiler
tube failure problems. A boiler (410 °C, or
770 °F, and 5 MPa, or 725 psig steam) at the Issy
plant in Paris suffered tube failures in the side
Ash removal wall after only about 5000 h of operation, and the
system same boiler suffered serious superheater corro-
sion problems from the start of the operation
Fig. 12.3 Refuse-derived fuel unit, with superheaters (right (Ref 1). In a boiler (525 °C, or 980 °F, and
above the nose arch or bull nose) and a boiler
(generating) bank in the convection pass. The bottom of the fur- 12 MPa, or 1800 psig steam) in Mannheim,
nace is a traveling grate. Source: Ref 3 Germany, the furnace wall tubes failed only after
Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/ 30/10/2007 4:10PM Plate # 0 pg 339

Chapter 12: Waste-to-Energy Boilers and Waste Incinerators / 339

3000 h of operation, and superheater tube failed Alloy 625 overlay has been found to provide
after only 2000 h of operation (Ref 1). The first dramatic reduction in metal loss in areas where
U.S. WTE boiler with relatively low steam carbon or low-alloy steels have suffered unac-
temperature (320 °C, or 610 °F) in Nashville, ceptable wastage rates (Ref 7, 16, 22–25). For
TN, suffered both waterwall and superheater mass-burning units, the waterwalls are typically
failure within the first year of operation (Ref 18). protected by alloy 625 weld overlay above the
Many other tube failures have been cited in Ref 1, refractory lining at the lower, high-radiant sec-
2, 18–21). All these initial boiler tubes materials tion of the boiler (Ref 9). The waterwalls in an
were carbon and low-alloy steels with no corro- RDF boiler are generally fully protected with
sion protection. alloy 625 weld overlay (Ref 9). Kubin (Ref 7)
Modern WTE boilers have increasingly used indicated that all RDF boiler waterwalls for
various corrosion protection methods to pro- Ogden plants were virtually fully covered with
tect waterwalls, superheaters, and boiler banks alloy 625 weld overlay.
against corrosion and erosion/corrosion. Many In many plants, alloy 625 as an outer tube
boilers with no corrosion protection for their cladding for corrosion protection of superheaters
waterwalls and superheaters continue to suffer (with higher metal temperatures) has also proved
premature failures. Figure 12.4 shows an exam- to provide significant reduction in tube-wall
ple of a tube failure at the waterwall in one boiler wastage rates compared to carbon and low-alloy
after only 8 months of operation (Ref 22). One steels (Ref 10, 22, 24). Excellent performance
effective corrosion protection method for the was observed for alloy 625 overlay superheater
furnace waterwalls was developed in the 1985 to tubes in a boiler in Europe, as shown in Fig. 12.5.
1986 period by using alloy 625 overlay cladding, The cladding can either be in a form of weld
which was applied on-site in an RDF boiler in overlay (Ref 24) or coextruded tube cladding
Lawrence, MA using an automatic gas metal arc (Ref 10). In some plants, however, alloy 625
welding (GMAW) process (Ref 23). A total of cladding in superheaters was found to experience
about 21,000 lb of alloy 625 weld overlay metal high wastage rates (Ref 10, 12, 22, 24). This is
was applied to this boiler. The performance of the illustrated in Fig. 12.6 (Ref 10, 22). The wastage
waterwall overlay of alloy 625 in this first over- rates ranged from less than 10 mpy to hundreds
laid waterwall proved to be very successful of mpy. Superheater corrosion appears to be
during the next 3 to 4 years of boiler operation. highly dependent on the individual boiler.
Between 1989 and 1990, a total of about Possible corrosion mechanisms involved in
260,000 lb of alloy 625 weld overlay metal was WTE environments are discussed in the next
applied to the waterwalls of 29 boilers (Ref 23). section.

Fig. 12.4 Carbon steel waterwall suffering blown tubes due to high wastage rates resulting in significant tube-wall thinning after
8 months of service in a WTE boiler. Courtesy of Welding Services Inc.
Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/ 30/10/2007 4:10PM Plate # 0 pg 340

340 / High-Temperature Corrosion and Materials Applications

12.4 Corrosion Mechanisms if present, is significantly lower than that of HCl

(see Table 12.1). Thus, HF plays no significant
As described in Section 12.3, the furnace role in corrosion of boiler tubes.
waterwalls and superheaters made of carbon or Carbon and low-alloy steels are very suscep-
low-alloy steels suffered severe wastage pro- tible to corrosion by HCl at elevated tempera-
blems in numerous boilers, experiencing tube tures. The level of HCl generated in combustion
failures after only thousands of hours of opera- of MSW fuel is typically in the range of 100 to
tion. Corrosion was the result of exposure to the 2000 ppm. The tube metal temperatures of the
combustion products generated during the com- waterwall, although depending on the operating
bustion of MSW. Typical gaseous components pressure of the boiler along with other factors, are
generated during combustion are O2, CO2, H2O, generally in the range of 260 to 315 °C (500 to
SO2, and HCl with HF in some units. In some 600 °F). The superheater tube metal tempera-
cases, local areas at the waterwall may be in tures are generally in the range of 370 to 540 °C
reducing conditions. In those local furnace wall (700 to 1000 °F), depending on the steam tem-
areas, there could be some CO present. Among perature, along with other factors. A brief review
these gaseous components, HCl is considered to of the corrosion data in terms of the direct HCl
be the most corrosive. In general, the level of HF, corrosion is discussed below to see whether HCl

Fig. 12.5 Alloy 625 overlay superheater tubes (on 15Mo3 steel substrate) after 4.5 years of service in a superheater producing 405 °C
(760 °F)/42 bar (609 psi) superheated steam, showing no evidence of corrosion or erosion/corrosion. Source: Ref 24

Steam temperature, °C
–18 95 205 315 425 535 Series 1
350 Series 2

300
Wastage rate, mpy

250

200

150

100

50

0
0 200 400 600 800 1000
Steam temperature, °F

Fig. 12.6 Wastage rates as a function of steam temperature for alloy 625 cladding in weld overlay tubes and coextruded tubes tested as
part of superheater tube bundles at various WTE boilers. Source: Ref 10, 22
Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/ 30/10/2007 4:10PM Plate # 0 pg 341

Chapter 12: Waste-to-Energy Boilers and Waste Incinerators / 341

Table 12.2 Corrosion rates in terms of metal loss for iron- and nickel-base alloys at indicated
temperatures for 1008 h in N2-10O2-50ppmSO2-500ppm HCl
Metal loss
at 425 °C (800 °F) at 480 °C (900 °F) at 590 °C (1100 °F)
Alloy µm/yr mpy µm/yr mpy µm/yr mpy
Type 316 0.07 0.003 2.54 0.1 5.08 0.2
Type 347 0.09 0.004 2.29 0.09 7.11 0.28
Alloy 800HT 0.07 0.003 1.02 0.04 5.33 0.21
Alloy 825 0.02 0.0008 1.27 0.05 2.54 0.1
Alloy 600 0.04 0.002 2.03 0.08 3.05 0.12
Alloy 625 0.02 0.0008 1.78 0.07 2.79 0.11
1.0 µm = 0.001 mm = 0.0394 mil. Source: Ref 27

corrosion can be responsible for severe waterwall Table 12.3 Corrosion rates in terms of metal
and superheater corrosion. loss for iron- and nickel-base alloys at 590 °C
Paul and Daniel (Ref 26) conducted laboratory (1100 °F) for 72 h in N2-9O2-12CO2-
tests at 315 °C (600 °F) for 720 h in simulated 100ppmSO2-4 and 10HCl
flue gas environments with one being an oxidiz- Metal loss

ing atmosphere (N2-2.6O2-15.0CO2-11.9H2O- 4% HCl 10% HCl

1000ppmSO2-1000ppmHCl) and the other Alloy µm/yr mpy µm/yr mpy

a reducing atmosphere (N2-11.4CO2-7.0CO- Type 316 914.4 35.7 … …

Type 347 1244.6 48.5 … …
4.4H2-11.9H2O-1000ppmHCl). Specimens were Alloy 800HT 73.66 2.9 … …
not coated with salts nor covered with ash de- Alloy 825 20.32 0.8 1066 42.0
posits. Both environments, containing about Alloy 600 25.4 1.0 1219 47.5
Alloy 625 15.75 0.6 1549 60.4
1000 ppm HCl, showed that the corrosion rates
1.0 µm = 0.001 mm = 0.0394 mil. Source: Ref 27
for both carbon steel (SA178C) and Type 304
were less than 0.25 mm/yr (<10 mpy). Smith
and Ganesan (Ref 27) conducted extensive tests
in simulated flue gas environments with HCl con- et al. (Ref 28) conducted tests in Ar-5HCl
centrations varying from 500 ppm to 10% and and Ar-5HCl-10O2 environments at 600 °C
temperatures from 425 °C (800 °F) to 590 °C (1110 °F), showing much higher corrosion rates
(1100 °F). Corrosion rates were found to be quite in an oxidizing environment. However, the test
low for austenitic stainless steels and nickel-base tempered was much higher than the current
alloys at temperatures up to 590 °C (1100 °F) superheater metal temperature. Their test results
in flue gas environment containing 500 ppm are summarized in Table 12.5. More corrosion
HCl (Table 12.2). Even for the environment data in halogen and halide environments are
containing 4% HCl, the nickel-base alloys (alloys available in Chapter 6.
825, 600, and 625) continued to show low cor- It is generally agreed that chlorine in the fuel is
rosion rates. However, austenitic stainless steels an important cause of corrosion problems of
exhibited high corrosion rates. This is illustrated boiler tubes. However, from the previous dis-
in Table 12.3. A level of 4% HCl was more than cussion, the direct corrosion of HCl, a gaseous
an order of magnitude higher than what can be component that formed when the chlorine in the
expected in a refuse-fired combustion environ- fuel is combusted in the boiler, is not likely to
ment, and the temperature of 590 °C (1100 °F) be primarily responsible for the boiler tube cor-
was much higher than the superheater metal rosion. In Section 12.2, it was explained that
temperature of current WTE boilers. Never- there are a number of heavy metals, such as lead
theless, alloy 625 as either coextruded cladding (Pb), zinc (Zn), cadmium (Cd), arsenic (As), and
or weld overlay was observed to suffer wastage tin (Sb), along with alkali metals (i.e., Na, K) that
rates of more than 1.3 mm/yr (>50 mpy) in some are present in MSW. These metallic elements
aggressive refuse-fired boilers (see Fig. 12.6). can react with chlorine to form various chlor-
Table 12.3 shows that alloys 625 suffered a ides that have very low melting points. Many
corrosion rate of about 1.53 mm/yr (60.4 mpy) investigators (Ref 12, 14, 16, 20, 24, 29, 30) have
when HCl was increased to 10% in the en- considered that corrosion attack by low melting
vironment. Additional data for nickel-base alloys point chlorides is a principal mode of corrosion
are shown in Table 12.4 (Ref 27). Devisme for boiler tubes in WTE plants.
Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/ 30/10/2007 4:10PM Plate # 0 pg 342

342 / High-Temperature Corrosion and Materials Applications

Kawahara and Kira (Ref 31) observed that exhibited melting points from approximately 330
the corrosion rate of carbon steel changed to a to 650 °C (625 to 1200 °F), as shown in Fig. 12.8
much higher rate above approximately 300 °C (Ref 32). The figure also indicates that larger
(150 °F), which corresponded to the melting amounts of salt deposits melt at around 380 and
point of the deposits collected from the tube 500 °C (715 and 930 °F). In the deposits col-
samples in field testing. This is illustrated in lected from the 3000 h exposure tests, sulfur
Fig. 12.7. They also performed laboratory (as SO3) as high as about 40% and chlorine as
tests, which involved actual deposits collected high as 9%, along with Na, K, Pb, and Zn were
from the boiler to which ZnCl2 was added as 8% detected. The authors observed that the corrosion
ZnO. The laboratory data conformed well with attack increased with higher fused salt content
the field data. The temperature at which carbon in the deposits. This is illustrated in Fig. 12.9
steel suffered accelerated corrosion attack was (Ref 32) for Type 347 and alloy 625 in the
found to coincide with the melting point of the corrosion probe tests. The figure shows the cor-
boiler tube deposits. The data strongly suggest rosion attack increases as the sum of the heats of
that accelerated corrosion attack in the WTE fusion (or the amount of fused salt) of the deposit
boiler combustion environment was caused by increases.
the formation of molten phases in the tube The salts in the deposits are mainly chlorides
deposits. and sulfates. Chlorides exhibit much lower
Otsuka et al. (Ref 32) performed an extensive melting points than sulfates. Under the operating
analysis on the salt deposits collected from three conditions of WTE boilers, many metal chlorides
commercial WTE boilers (stoker type furnaces). can be in a molten state in the deposits on
A total of 23 salt deposits were collected from waterwall/screen tubes (evaporator tubes) and
three boilers using corrosion probes at 550 °C superheaters. This is illustrated in Fig. 12.10,
(1020 °F) (probe metal temperature) for expo- showing many low melting point chlorides.
sure of 700 and 3000 h, respectively. Each col- Approximate steam temperatures at the water-
lected deposit was analyzed for the heat of fusion wall and superheater are shown for general
and the melting point by differential scanning comparison (Ref 30).
calorimeter measurements. The heat of fusion is In Fig. 12.9, it is shown that increasing the
related to the amount of the fused salt in the amount of salts in the deposit could increase
deposit, with higher heat of fusion indicating
a larger amount of fused salt. The deposits

: Field test data

Table 12.4 Corrosion rates in terms of metal 0.30 400 h, gas temp.: 735 °C (1355 °F)
loss for iron- and nickel-base alloys at 590 °C : Laboratory test data
Average corrosion thickness loss, mm/400 h

(1100 °F) for 72 h in N2-9O2-12CO2- 40-fold 100 h thickness loss

100ppmSO2-4HCl with 0% and 10% H2O 0.25
Melting point of deposits
Metal loss
on field test tube
0% H2O 10% H2O
0.20 301 °C (570 °F)
Alloy µm/yr mpy µm/yr mpy
Type 316 914.4 35.7 1.85 0.07
Alloy 825 20.32 0.8 3.81 0.15
0.15
Alloy 600 25.4 1.0 8.85 0.35
Alloy 625 15.75 0.6 No measurable corrosion
1.0 µm = 0.001 mm = 0.0394 mil. Source: Ref 27
0.10

0.05
Table 12.5 Corrosion of nickel-base alloys in
terms of metal loss after 500 h at 600 °C (1110 °F)
in the indicated test environments 0
0 200 250 300 350 400
Metal loss, μm (mils)
Metal temperature, °C
Environment C-276 600 601 214
Ar-5HCl 35 (1.4) 50 (2.0) 90 (3.5) 30 (1.2)
Ar-5HCl-10O2 120 (4.7) 140 (5.5) 160 (6.3) 55 (2.2) Fig. 12.7 Corrosion in terms of thickness loss for carbon steel
(SA178) as a function of the metal temperature in
Source: Ref 28 field exposure tests. Also superimposed are data generated from
laboratory tests. Source: Ref 31
Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/ 30/10/2007 4:10PM Plate # 0 pg 343

Chapter 12: Waste-to-Energy Boilers and Waste Incinerators / 343

the corrosion attack on the boiler tube, chloride that forms on the deposit, as well as the
and Fig. 12.10 shows that there are many low- type of the corrosion products formed between
melting salts, particularly metal chlorides, that the chloride deposit and the underlying tube
can become molten at the operating temperatures metal, can significantly affect the corrosion
of furnace walls and superheaters. The type of attack. For example, some of the eutectic chlor-
ide salts involving FeCl3 can become molten
at 150 to 205 °C (300 to 400 °F) as shown in
35
Fig. 12.10. At higher metal temperatures, some
chloride salts exhibit high vapor pressures.
Figure 12.11 shows vapor pressures of some
30
alkali metal salts (KCl, NaCl, KOH, and NaOH)
and heavy metal chlorides (ZnCl2 and PbCl2)
25 (Ref 26). K, Na, Zn, and Pb are among the most
Heat of fusion, J/g

frequently detected elements in the tube deposits.

20 It is generally believed that vapor pressures of
about 10−4 atm are adequate to cause significant
15 high-temperature corrosion attack. The tem-
perature at which the vapor pressure of ZnCl2
is at 10−4 atm is about 350 °C (660 °F) (see
10
Chapter 6). At temperatures of 350 °C (660 °F)
5

0
300 400 500 600 700 PbCl2
NaCl/CaCl2
Melting point temperature, °C 900
(480) PbCl2/CaCl2
Max. superheater PbCl2/MgCl2
Fig. 12.8 Melting point temperatures versus the heat of fusion
steam temperature
for the deposits collected at the corrosion probes
(with the probe metal temperature of about 550 °C, or 1020 °F) in
three commercial WTE boilers. The open data points are from the 800
(425) PbCl2/FeCl2
outer portion of the deposit, and the solid data points are from the ZnSO4 KCl/PbCl2
inner portion of the deposit. Source: Ref 32 Na2S2O7 NaCl/PbCl2

700 NaCl/FeCl2
2 K2SO4·Na2SO4·ZnSO4 (370)
KCl/FeCl2
(30–50%:10–30%:40–60%)
Maximum corrosion thickness loss, mm/3000 h

600
1.5 (315)
K2S2O7
ZnCl2
Max. waterwall
fluid temperature 500 NaCl/ZnCl2
1 (260)
SnCl2
ZnCl2/KCl
Temperature °F (°C)

400 SnCl2/ZnCl2
0.5 (205) ZnCl2/FeCl3
NaCl/SnCl2
PbCl2/FeCl3 KCl/SnCl2

0 300 NaCl/FeCl3
0 2 4 6 8 10 12 (150)
Total sum of heat of fusion up to 550 °C, J/S

Fig. 12.9 Maximum thickness loss of corrosion probes

exposed at metal temperature 550 °C (1020 °F)
for 3000 h in WTE boilers as a function of the sum of heat of fusion Fig. 12.10 Melting temperatures of various salts that are
(corresponding to the amount of fused salt) of the deposits. The likely to form in WTE boilers. The temperatures
open data points are Type 347H, and the solid data points are of waterwall saturated fluid and superheater steam are also indi-
alloy 625. Source: Ref 32 cated for comparison. Source: Ref 30
Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/ 30/10/2007 4:10PM Plate # 0 pg 344

344 / High-Temperature Corrosion and Materials Applications

and higher, vapor-phase corrosion by ZnCl2 can second zone to 540 and 510 °C (1000 and
be significant. Below this temperature, ZnCl2 950 °F) in the third zone where specimens were
condenses, thus causing its vapor pressure below tested, the salt deposition rate onto the test spe-
10−4 atm and making vapor-phase corrosion less cimen was found to be about 0.07 mg/cm2 h of
likely. Accordingly, waterwall corrosion is less molten ZnCl2. The test specimens were thus
likely to be caused by vapor-phase corrosion coated with molten ZnCl2 during the testing.
involving ZnCl2. The temperature at which the Their test results are summarized in Table 12.6
vapor pressure of PbCl2 is at 10−4 atm is about (Ref 33). The test temperature of 510 °C (950 °F)
485 °C (900 °F) (see Chapter 6). If the vapor is very close to some superheaters with
phase of PbCl2 is involved in the corrosion steam temperatures of about 455 to 480 °C (850
attack, it is likely to be for superheaters. Both to 900 °F). The corrosion rates of alloy 625 were
KCl and NaCl will be at much higher temper- on the same order of magnitude as those of alloy
atures when these chlorides reach the vapor 625 (as a weld overlay or coextruded cladding)
pressure of 10−4 atm. Thus, both KCl and NaCl
are most likely to be involved in molten salt
corrosion not vapor corrosion. Table 12.6 Corrosion rates after testing for
Gleeson et al. (Ref 33) examined the effect of 190 h in N2-3.6O2-14CO2-0.25SO2 containing
ZnCl2 on the corrosion of several iron- and ZnCl2 vapor, which condensed onto the test
nickel-base alloys using a modified Dean test specimens as molten ZnCl2 at a rate of
with a three-zone furnace arrangement. The test 0.07 mg/cm2 h
involved a flowing flue gas (N2-3.6O2-14CO2- Corrosion rate, mm/yr (mpy)

0.25SO2) passing first through a crucible Material 540 °C (1000 °F) front row 510 °C (950 °F) back row

containing molten ZnCl2 salt at 425 to 540 °C C-2000 1.83 (72) …

625 3.07 (121) 2.97 (117)
(800 to 1000 °F) to pick up ZnCl2 vapor C-22 2.9 (114) 2.11 (83)
(1.3×10−3 atm at 425 °C, or 800 °F, and 214 1.42 (56) 2.18 (86)
2.8×10−2 at 540 °C, or 1000 °F), followed by the 230 4.24 (167) 2.87 (113)
G-30 … 4.17 (164), 4.42 (174)
second zone heated to about 815 °C (1500 °F) 825 4.72 (186) 3.15 (124)
with a catalyst to equilibrate the gas mixture, and HR-160 2.18 (86) 3.45 (136)
to the third zone where specimens were exposed HR-120 8.94 (352) 5.26 (207)
556 7.11 (280) 3.63 (143)
to a 540 °C (1000 °F) gas mixture containing Type 310 7.77 (306) …
ZnCl2 vapor. Because the gas mixture drops in Source: Ref 33
temperature from 815 °C (1500 °F) in the

Temperature, °C
200 400 600 800 1000 1200
1
ZnCI2
KOH

10–1 Condensation
occurs
PbCI2 KCI
Vapor pressure, atm

10–2

10–3

10–4 Species remains

as a vapor
NaOH
10–5
NaCI

10–6
600 800 1000 1200 1400 1600
Temperature, K

Fig. 12.11 Vapor pressures of some alkali and heavy metal salts as a function of temperatures. Source: Ref 26
Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/ 30/10/2007 4:10PM Plate # 0 pg 345

Chapter 12: Waste-to-Energy Boilers and Waste Incinerators / 345

in some very aggressive boilers, as shown in and corrosion products. Wright et al. (Ref 16)
Fig. 12.6. It is thus believed that the severe cor- reported that the deposits (on a severely corroded
rosion of the alloy 625 overlay (or cladding) waterwall tube) were found to contain 10 to 30%
superheater tubes was most likely due to molten chloride, 33 to 47% Pb, 3 to 13% Zn, 3 to 14%
ZnCl2 and/or molten PbCl2. Because of con- Na, and 2 to 11% K. Spiegel (Ref 12) analyzed
siderable scattering in the data, the test results two overlay superheater tube samples that were
were not adequate for making an alloy ranking severely corroded in German boilers. One tube
among nickel-base alloys. X-ray diffraction sample was overlaid with Ni-16.7Cr-8Fe-4.6Si
analysis of the scale and corrosion products was alloy, while the other overlaid with alloy 625.
performed on alloys C-22, 625, G-30, 825, HR- In both cases, the overlays were severely cor-
120, and Type 310, showing mainly Cr2O3 and roded. Analysis of the deposits in both cases
NiCr2O4 for all alloys, with NiSO4 being detec- showed salt melts were sulfates in contact with
ted for C-22, 625, 825, and Type 310 (Ref 33). the flue gas and chlorides at the deposit/scale
From analysis of severely corroded tube phase boundary (Ref 12). The sulfate smelt
samples obtained from operating boilers, it was was a CaSO4-K2SO4-Na2SO4 system including
learned that Cl, Zn, Pb, Cd, Na, K, and S, along ZnSO4 and PbSO4. The chloride melt was KCl-
with major alloying elements from the tube (e.g., ZnCl2 and also NaCl.
Fe if the tube was steel, or Cr and Ni if alloy 625 Heavy metals, such as Zn and Pb, along with
overlay) were often detected in the deposit Cl, have been detected in the corrosion front

60 µm

Fig. 12.12 Scanning electron micrograph (backscattered electron image) showing the deposits and corrosion scales formed on a
carbon steel (SA178A) superheater tube suffering severe tube-wall wastage. Chemical compositions at different locations
were analyzed by energy-dispersive x-ray spectroscopy (EDX) analysis (trace elements not reported here):
1: 31% Ca, 29% Si, 14% Mg, 15% Fe, 9% S, and 2% Zn 5: 72% Fe, 6% Cl, 7% Zn, 4% S, 4% Na, and 2% K
2: 63% Fe, 16% Cl, 9% Zn, 4% Pb, and 2% S 6: 88% Fe, 2% Cl, 2% Zn, 2% Cd, and 1% Na
3: 20% Fe, 13% Cl, 3% Zn, 41% Pb, 11% S, 4% Na, 3% K, and 2% Ca 7: 76% Fe, 8% Cl, 5% Zn, 2% Cd, 2% Na, and 2% S
4: 67% Fe, 12% Cl, 7% Zn, 4% S, and 6% Na
Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/ 30/10/2007 4:10PM Plate # 0 pg 346

346 / High-Temperature Corrosion and Materials Applications

of superheater tubes from operating boilers The composition of the deposit and corrosion
(Ref 22). This is illustrated in Fig. 12.12 (Ref 22) products can vary significantly from plant to
for a carbon steel superheater tube that suffered plant. This is shown in another boiler described
severe corrosion attack. The steam temperature in Fig. 12.13 and 12.14 (Ref 22), which involved
and pressure of the superheater were reported an alloy 625 overlay superheater tube (410 °C, or
to be 400 °C (750 °F) and 4.5 MPa (625 psig), 770 °F, steam) after service for about 6.5 months.
respectively. The tube wall was reportedly re- Figure 12.13(a) shows the full cross section of
duced from the original 5.6 mm (0.220 in.) to the weld overlay with slight surface pitting at-
about 1.02 mm (0.040 in.) after 11 months of tack, and Fig. 12.13(b) shows one of the corro-
service, with a wastage rate of approximately sion pits at a higher magnification. The chemical
5 mm/yr (196 mpy). Figure 12.12 shows the compositions of the deposit and the corrosion
deposit and corrosion products. The corrosion products, which were analyzed by SEM/EDX,
front (location No. 5, 6, and 7 in Fig. 12.12) was are shown in Fig. 12.14. Significant amounts
found to contain Cl, Zn, Cd, Na, S, and Fe. The of lead were found throughout the deposit
compounds are believed to contain iron chlorides and corrosion products. It is also significant
with Zn, Cd, and Na. that Pb, Zn, and Cl, along with Cr, Ni, and

(a)

0.025 mm
(b)

Fig. 12.13 (a) Optical micrograph showing slight pitting corrosion attack on alloy 625 overlay in an alloy 625 spiral overlay
superheater tube after about 6.5 months of service. Micrograph (a) also shows the fusion boundary and substrate carbon
steel. (b) Higher-magnification view of one of the corrosion pits. SEM/EDX analysis on the corrosion products in one of the surface pits is
summarized in Fig. 12.14.
Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/ 30/10/2007 4:10PM Plate # 0 pg 347

Chapter 12: Waste-to-Energy Boilers and Waste Incinerators / 347

Mo (major alloying elements from the weld penetration) is in a way similar to internal oxi-
overlay) were detected in the corrosion front, dation attack or intergranular oxidation attack
as indicated in locations No. 10 and 12 in following general oxidation (or general metal
Fig. 12.14. wastage by oxidation) for wrought alloys. Simi-
Both the morphology and the corrosion front larly, internal corrosion attack (or intergranular
that contained heavy metals, such as Zn and/or corrosion attack) also takes place in wrought
Pb, as were observed in the above two cases, alloys under gaseous chloridation attack. Fig-
suggest that the corrosion mechanism is fluxing ure 12.15, shows general material wastage
by molten chloride salts. followed by internal dendrite corrosion penetra-
In some cases, the corrosion morphology for a tion in the overlay for an alloy 622 overlay
weld overlay of a Ni-Cr-Mo alloy involves gen- superheater tube after 225 days of exposure.
eral metal wastage followed by internal corrosion This weld overlay tube was still in excellent
penetration along dendrites of the weld over- condition, as shown in Fig. 12.16, showing the
lay. This internal dendrite corrosion attack (or entire cross section of this alloy 622 overlay (on

30 µm

Fig. 12.14 Scanning electron micrograph (backscattered electron image) showing a localized corrosion pit on alloy 625 overlay of a
superheater tube. SEM examination using energy-dispersive x-ray (EDX) spectroscopy showed the chemical compositions
(wt%) at different phases, marked as No. 1 through No. 12. The chemical compositions (wt%) at different phases are:
1: 68% Pb, 11% Mo, 6% Cr, 3% Fe, 3% Ni, 4% S, 3% Cl, and trace elements
2: 63% Pb, 9% S, 6% Cl, 7% Cr, 5% Mo, 2% Fe, 3% Ni, 2% Na, and trace elements
3: 31% Cr, 24% Ni, 2% Fe, 27% Pb, 6% Zn, 5% S, 1% Cl, and trace elements
4: 32% Pb, 14% Cr, 8% Ni, 5% Fe, 10% Mo, 7% Zn, 9% Cl, 5% K, 3% Na, 3% S, and trace elements
5: 52% Pb, 11% K, 13% S, 7% Cl, 6% Mo, 4% Na, 4% Cr, 2% Ni, and trace elements
6: 34% Ni, 23% Cr, 3% Fe, 23% Pb, 7% Zn, 6% Mo, 2% Cl, and trace elements
7: 49% Pb, 12% K, 3% Na, 13% S, 7% Cl, 5% Cr, 6% Mo, 3% Ni, and trace elements
8: 30% Cr, 28% Pb, 13% Ni, 11% Zn, 7% Mo, 6% Cl, 2% K, and trace elements
9: 26% Pb, 22% Ni, 16% Cr, 11% Zn, 7% Cl, 5% Mo, 7% Na, 3% Fe, and trace elements
10: 27% Pb, 19% Cr, 14% Ni, 13% Zn, 9% Mo, 5% Cl, 4% Na, 2% K, and trace elements
11: 54% Ni, 24% Cr, 6% Pb, 5% S, 6% Fe, and trace elements
12: 27% Cr, 19% Ni, 19% Pb, 15% Zn, 9% Mo, 5% Cl, 2% K, 2% Fe, and trace elements
Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/ 30/10/2007 4:10PM Plate # 0 pg 348

348 / High-Temperature Corrosion and Materials Applications

a carbon steel tube). The morphology of this Spiegel (Ref 12) observed similar internal den-
internal dendritic corrosion penetration is drite corrosion attack on an alloy 625 overlay
believed to be the result of gaseous corrosion superheater tube where alloy 625 overlay suf-
reactions, but not by the molten salt fluxing fered severe corrosion attack. He found that
mechanism. In the same exposure test, an alloy those internal dendrite corrosion phases were
625 overlay tube was installed at the next platen chlorides of mainly iron, chromium, and nickel,
on the same row as the alloy 622 overlay along with the corresponding oxides. He sug-
tube discussed in Fig. 12.15 and 12.16. After gested that the molten chlorides reacted with
152 days of exposure, the tube was removed oxides (formed on the metal) to produce chlorine,
for metallurgical examination. The alloy 625 which then reacted with the metal to form metal
overlay showed slight surface corrosion pitting chlorides.
attack (similar to alloy 622 overlay), but The internal dendrite corrosion penetration is
no internal dendritic corrosion attack, as illu- similar to internal corrosion attack within the
strated in Fig. 12.17 (Ref 22). In another boiler alloy matrix or internal grain-boundary corrosion
where alloy 625 overlay superheater tube attack in wrought alloys under high-temperature
suffered severe corrosion attack, the overlay gaseous corrosion, such as oxidation and
showed only general metal wastage with no chloridation. In high-temperature gaseous cor-
internal dendritic corrosion attack (Ref 22). rosion, the extent of internal corrosion attack
Montgomery and Larsen (Ref 34) also found can vary from alloy to alloy. Some alloys may
that alloy 622 weld overlay, which was applied exhibit more internal attack than others. It is
to the waterwall (the rear wall of the boiler in important to note, however, that the alloys with
the first pass) in a WTE boiler (Haderslev plant) less internal oxidation attack are not necessarily
in Denmark, showed internal dendrite corrosion more corrosion resistant. There is no clear
penetration after 8000 h of exposure, while indication that the weld overlay with internal
alloy 625 weld overlay at a similar location dendrite corrosion penetration would be less
tested during the same time period showed resistant in terms of the alloy’s overall corrosion
no internal dendrite corrosion penetration after attack. It is believed that the corrosion rate due
8000 h of exposure. Both overlays showed only to general metal wastage by molten salt fluxing
narrow pitting corrosion attack. Nevertheless, is the rate-controlling factor.

0.025 mm

Fig. 12.15 Internal dendrite corrosion attack that followed general wastage was observed in the overlay of an alloy 622 weld overlay
in an alloy 622 overlay superheater tube. Lower-magnification micrographs showing the through-thickness overlay is
shown in Fig. 12.16(a) and general surface corrosion morphology is shown in Fig. 12.16(b).
Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/ 30/10/2007 4:10PM Plate # 0 pg 349

Chapter 12: Waste-to-Energy Boilers and Waste Incinerators / 349

12.5 Corrosion Protection for temperature is reached. For corrosion protection

Furnace Waterwalls of the waterwall above the refractory, the current
prevailing method is the application of alloy
In mass-burning units, the lower part of the 625 weld overlay using automatic gas metal
waterwall is generally protected by refractories arc welding (GMAW) on site for existing boilers.
(e.g., SiC) (Ref 3, 7, 8, 16). The refractories The waterwall can also be constructed with shop-
provide protection for the waterwall against fabricated weld overlay panels with alloy 625
mechanical damage from abrasion of sliding overlay or with coextruded composite tubes with
large articles on the moving grate and also alloy 625 cladding.
against corrosion from the combustion products. In RDF units, the same refractory design used
Relatively lengthy installation and upkeep time for mass-burning units was tried initially by a
as well as reducing the heat-absorbing surface of major boiler designer resulting in slagging pro-
the waterwalls can be an issue for refractory blems for the refractory wall surface (Ref 3).
linings (Ref 20). Nevertheless, Licata et al. This was caused by the insulation of the lower
(Ref 17) indicated that refractory linings are furnace waterwalls by the refractory that resulted
needed to provide thermal insulation for ensur- in higher flame temperatures and caused sig-
ing that the designed combustion flue gas nificant refractory wall slagging (Ref 3).

(a)

0.010 mm
(b)

Fig. 12.16 Alloy 622 overlay superheater tube after 225 days of exposure in a boiler. (a) Cross section of the overlay showing slight
pitting attack. Micrograph (a) also shows the fusion boundary and substrate carbon steel. (b) Higher-magnification
micrograph showing surface corrosion morphology with general metal wastage and internal dendrite corrosion penetration
Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/ 30/10/2007 4:10PM Plate # 0 pg 350

350 / High-Temperature Corrosion and Materials Applications

Following rapid corrosion attack of the bare overlay on the waterwalls of the boiler was
carbon steel waterwalls in an RDF unit in Lawr- found to perform well in both mass-burning and
ence, MA, the lower furnace waterwall was weld RDF units. Generally, the wastage rates of alloy
overlaid with Inconel material in 1986, and this 625 overlay on waterwalls in U.S. boilers were
overlay protection proved to be very effective quite low, typically approximately 125 μm/yr
(Ref 3). The overlay alloy applied in the Lawr- (5 mpy) (Ref 22, 25). The boiler tube metal
ence unit was alloy 625 (Ref 23). The furnace temperatures are in a range of 260 to 315 °C
waterwalls of modern RDF units are protected by (500 to 600 °F) for most U.S. boilers. For some
alloy 625 cladding either as weld overlay applied European boilers running higher water/steam
on-site, or shop-applied panels, or as coextruded pressures, the waterwall tube metal temperatures
composite tubes with no refractory. Kubin (Ref could be higher.
7) indicated in 1990 that the waterwalls of all Figure 12.19 shows the cross section of an
Ogden RDF boilers were virtually fully covered overlaid waterwall tube sample obtained from a
with alloy 625 weld overlay. RDF boiler in Lawrence, MA after 16 years of
Figure 12.18 shows a general view of an service. The metallographic cross section of the
automatic GMAW-applied alloy 625 weld over- alloy 625 weld overlay of that overlay waterwall
lay on the waterwall in a boiler. Alloy 625 weld tube is shown in Fig. 12.20, revealing the full

(a)

0.010 mm
(b)

Fig. 12.17 Alloy 625 overlay superheater tube after 152 days of exposure in the superheater platen next to the alloy 622 superheater
tube, which is shown in Fig. 12.16. (a) Cross section of the overlay showing slight pitting attack. Micrograph (a) also shows
the fusion boundary and substrate steel. (b) Higher-magnification micrograph showing surface corrosion morphology with general metal
wastage and no internal dendrite corrosion penetration.
Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/ 30/10/2007 4:10PM Plate # 0 pg 351

Chapter 12: Waste-to-Energy Boilers and Waste Incinerators / 351

cross section of the overlay with little evidence of 35). DeVincentis et al. (Ref 35) conducted cor-
corrosion attack. Figure 12.21 shows a close- rosion probe tests on sprayed coatings of three
up view of alloy 625 overlaid waterwall after nickel-base alloys (Ni-Cr-Co-Si alloy HF-160,
about 10 years of service in another RDF boiler.
The common mode of corrosion attack on alloy
625 weld overlay on the waterwall has been
found to be pitting attack. This is illustrated in
Fig. 12.22. When corrosion or pitting attack
becomes extensive, the corroded waterwall area
can be repaired by first grinding off the corroded
metal prior to performing overlay welding, as
shown in Fig. 12.23.
It has been reported by Vrchota (Ref 25) that
the bare carbon steel waterwall above the weld
overlaid waterwall had experienced higher
wastage rates as the boiler continued to increase
its service duration. As a result, the weld overlay
area has been “creeping” upward gradually,
leading to application of weld overlay at
increasingly higher elevation in the boiler. This
situation has also been experienced in other
plants (Ref 22). However, there is no consensus
on a technical explanation about this “pheno- Fig. 12.19 Cross section of an overlaid waterwall tube
menon.” showing alloy 625 overlay after 16 years of
service in a RDF unit in Lawrence, MA. Courtesy of Welding
Thermal sprayed coatings have not yet been Services Inc.
used on a large scale in WTE boilers. One of
the major issues is related to the intrinsic char-
acteristics of the sprayed coating in terms of
interconnecting pores that allow the corrosive to
permeate through the coating and cause corro-
sion attack at the coating/substrate interface, thus
leading to spallation. Another issue is lack
of automatic application system that can cover a
large waterwall area in achieving a consistent
quality over the large coating area. In the 1990s,
some tests on sprayed coatings were performed
in boilers without satisfactory results (Ref 7, 16,

Fig. 12.20 Optical micrograph showing the cross section of

the alloy 625 overlay of an overlaid waterwall
sample (Fig. 12.19) obtained from a RDF boiler (Lawrence, MA)
Fig. 12.18 General view of an overlaid waterwall with alloy after 16 years of service. Micrograph also shows the fusion
625 overlay in a WTE boiler. Courtesy of Welding boundary and substrate carbon steel. Courtesy of Welding Ser-
Services Inc. vices Inc.
Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/ 30/10/2007 4:10PM Plate # 0 pg 352

352 / High-Temperature Corrosion and Materials Applications

Ni-Cr-Fe-Al-Y alloy 214, and Ni-Cr-Mo alloy C- found to be an arduous and time-consuming
22). The corrosion probe tests were conducted in operation (Ref 7).
the first pass of the front wall in a boiler at
Hempstead plant for 76 days of exposure with the
metal temperature maintained between 180 and
230 °C (350 and 450 °F). Coatings were applied 12.6 Corrosion Protection for
by the inert gas electric arc spraying method. Superheaters
Disbonding of the coating during exposure was
found to be a major problem. Alloy C-22 and 214 Metal temperatures for superheater tubes can
coatings were found to suffer disbonding after be 370 to 480 °C (700 to 900 °F) or higher. In
exposure. Kubin (Ref 7) indicated that a sprayed this temperature range, there will be more
coating of alloy 50Ni-50Cr was tested in the chloride salts that become molten, as shown in
upper furnace of a mass-burning unit and failed Fig. 12.10. This may also result in a higher
after 2.2 years of service. Furthermore, the concentration of molten chloride salts in the ash
removal of all the coating material in preparation deposits forming on the tube surface. Presence of
for overlay welding with alloy 625 later was more chloride salts will result in more corrosion
attack (Fig. 12.9). At higher temperatures, some
of the heavy metal chlorides, such as ZnCl2 and
PbCl2, exhibit higher vapor pressures, as shown
in Fig. 12.11. This can make the environment
more corrosive for superheater tube materials. As
discussed in Section 12.3, premature failures of
superheater tubes made of carbon or low-alloy
steels were quite common. Figure 12.24 shows
the cross section of a carbon steel (SA178A)
superheater tube removed after 11 months of
service in a mass-burning unit (Ref 22). The
superheated steam temperature and pressure for
this boiler were 400 °C (750 °F) and 4.5 MPa
(625 psig), respectively.
Blough et al. (Ref 36) conducted corrosion
probe tests in a mass-burning unit at Charleston
Resource Recovery. A wide variety of commer-
Fig. 12.21 Close-up view of alloy 625 overlay on the cial alloys were tested, including carbon and
waterwall of another RDF boiler after 10 years of low-alloy steels, austenitic stainless steels, Fe-
service. Shown in the photograph are two tubes and a membrane.
Courtesy of Welding Services Inc. Ni-Cr alloys, and nickel-base alloys. Tests were

Fig. 12.23 Repair welding can be performed on the corroded

waterwall overlay by grinding followed by over-
Fig. 12.22 Pitting attack is the common mode of corrosion of lay welding. An alloy 625 weld overlay on the waterwall after
alloy 625 weld overlay on the waterwall. Shown 10 years of service in a RDF unit. Shown in the photograph are two
in the photograph are three tubes and three membranes. tubes and three membranes.
Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/ 30/10/2007 4:10PM Plate # 0 pg 353

Chapter 12: Waste-to-Energy Boilers and Waste Incinerators / 353

conducted for 4492 h. The results are summar-

ized in Fig. 12.25 (Ref 36). Alloy 625 was found
to be the best at all test temperatures. Austenitic
stainless steels and Fe-Ni-Cr alloys and some
nickel-base alloys except alloy 625 form a big
scattering band between alloy 625 and ferritic
steels. In examining the wastage profiles of the
probe sample cross sections, some alloys showed
the maximum wastage at the 90° position, where
the flue gas stream impinged upon, while others
showed the maximum wastage at about 45° from
either side from the 90° position. In another
field test conducted in an RDF unit at Elk River
Station, Blough et al. (Ref 37) reported that
alloy 625 was found to perform significantly
better than T-22, chromized T-22, Type 304H,
HR3C, and 825 at both 470 and 500 °C (880
and 935 °F). Tests were conducted with tube
Fig. 12.24 Carbon steel (SA178A) superheater tube after 11 samples of different alloys welded together as
months of service in a mass-burning unit. The part of the lead superheater tubes after about
superheated steam temperature and pressure were 400 °C
(750 °F) and 4.5 MPa (625 psig), respectively. Courtesy of 1180 h of exposure. The results are summarized
Welding Services Inc. in Fig. 12.26 and 12.27. The chromized layer was

Temperature, °C
205 260 315 370 425 480 535
0.160
Ferrite band
Ferritic steels
0.140 Austenitic steels

625
0.120 825

800H Austenite
0.100 600 band
Max. wastage, in.

HR160

0.080

0.060

0.040
625

0.020

0.000
400 450 500 550 600 650 700 750 800 850 900 950 1000
Temperature, °F

Fig. 12.25 Results of corrosion probe tests for various alloys in a boiler at Charleston Resource Recovery. The corrosion probes were
installed in the convection path with the exposure time of 4492 h. Ferritic steels included carbon steel, T-22, and T-91.
Austenitic steels included Type 304, 347, 310, and 27Cr-31Ni-3.5Mo (N08028). Source: Ref 36
Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/ 30/10/2007 4:11PM Plate # 0 pg 354

354 / High-Temperature Corrosion and Materials Applications

found to be consumed in large areas and thus cladding in coextruded composite tubes has also
provided no protection. Kubin (Ref 7) had tested been widely used in WTE boilers (Ref 10).
chromized coatings for superheater applications For superheaters, the corrosiveness of the
with mixed results, showing some success at one environment varies greatly from boiler to boiler.
facility, but not at another facility. Furthermore, In some boilers, alloy 625 has performed well
chromizing was not found to be cost effec- as an overlay or cladding in superheaters. One
tive (Ref 7). Other diffusion coatings, such as example is shown in Fig. 12.5. In some other
aluminizing and aluminum-silicon codiffusion boilers, the environment was so corrosive that
coatings, were tested and found to be inadequate alloy 625 cladding lasted for only about 1 year.
in performance (Ref 7). For example, in boilers with superheated steam
Alloy 625 as a weld overlay in spiral overlay temperatures between 400 and 455 °C (750 and
tubing or a cladding in coextruded tubing offers a 850 °F), the alloy 625 cladding was found to
viable solution to the superheater corrosion pro- exhibit a wide range of wastage rates from a
blems. Table 12.7 summarizes the comparative negligible rate to 7.5 mm/yr (300 mpy), as
performance between a bare carbon steel tube shown in Fig. 12.6.
and an alloy 625 overlay tube in a finishing The factors that affect the corrosiveness of the
superheater in a side-by-side field test in a boiler. environment can be very complex. In his corro-
The wastage rate was found to be about 2.8 mm/ sion probe testing in an operating boiler, Krause
yr (110 mpy) for carbon steel and 0.46 mm/yr (Ref 29) found that the flue gas temperature
(18.3 mpy) for alloy 625 overlay. Based on these could be important in affecting the corrosion rate
data, the expected replacement interval for car- of carbon steel. His data are shown in Fig. 12.28.
bon steel tubes and alloy 625 overlay tubes are The figure shows that when the flue gas
1.4 and 5.8 years, respectively (Table 12.7).
Figure 12.5 shows alloy 625 overlay superheater
tubes (405 °C, or 760 °F, and 42 bar, or 609 psi, 80 2
superheated steam) exhibiting excellent overlay 45° wastage 315° wastage
condition with no sign of corrosion or erosion/
corrosion after 4.5 years of service in a boiler in 60 1.5
the Netherlands (Ref 24). Alloy 625 overlay
Corrosion, mils

Corrosion, mm
tubing has been widely used for superheater
applications in WTE boilers. Alloy 625 as a 40 1

20 0.5
400 10
45° wastage 315° wastage

8 0 0
304 HR-3C T-22Cr 825 T-22 625
300
Alloys
Corrosion, mils

Corrosion, mm

6
Fig. 12.27 Tube metal loss of various alloys tested at 470 °C
200 (880 °F) for 1180 h in an RDF unit at Elk River
Station. T-22CR represents the chromized T-22 tube sample. The
4 chromized layer (on T-22) was found to be consumed in a large
area. Source: Ref 37

100
2
Table 12.7 Wastage rates for the carbon
steel tube and the alloy 625 overlay tube in a
0 0 side-by-side test as part of a finishing superheater
304 HR-3C T-22CR 825 T-22 625 in a boiler in New York
Alloys
Wastage rate
Tube construction mm/yr mpy Expected tube life, years
Fig. 12.26 Tube metal loss of various alloys tested at 500 °C
Carbon steel 2.8 110 1.4
(935 °F) for 1180 h in an RDF unit at Elk River
Station. T-22CR represents the chromized T-22 tube sample. The 625 overlay 0.46 18.3 5.8(a)
chromized layer (on T-22) was found to be consumed in a large (a) Overlay (0.080″ thick) + steel tube. Source: Ref 24
area. Source: Ref 37
Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/ 30/10/2007 4:11PM Plate # 0 pg 355

Chapter 12: Waste-to-Energy Boilers and Waste Incinerators / 355

temperature increased from 760 to 845 °C (1400 way to judge the severity of the superheater
to 1550 °F), the corrosion rate of carbon steel wastage issue in a particular boiler is to examine
was found to accelerate as the metal temperature the historical data of that particular boiler. For
exceeded 430 °C (800 °F). The author (Ref 29) “aggressive” boilers, efforts have been underway
believed that the increased corrosion rate was the in the industry to identify an alloy that can out-
result of the formation of volatile FeCl3, since perform alloy 625. The results, however, have
below 430 °C (800 °F) the corrosion product not been encouraging thus far. The findings of
was primarily FeCl2, which would not volatilize some of the comparison tests are summarized in
at these temperatures. This flue gas temperature the next paragraph.
increase is not likely to significantly affect the In a side-by-side field test on alloys 625 and
corrosion rate of alloy 625, which is a nickel-base 622 overlay tubes (five-tube platen panel each)
alloy and not likely to form either FeCl2 or FeCl3.
Metal temperature can significantly affect the
corrosion rate of the alloy. The corrosion reaction
is a thermally activated process, thus increasing
metal temperature can result in a higher corrosion
rate. Furthermore, increasing metal temperature
can increase the range of various chloride salts
that become molten (Fig. 12.10). Increased
amounts of molten chloride salts can also in-
crease corrosion rates. Furthermore, increasing
temperature can increase vapor pressures of
chloride salts, thus resulting in increased corro-
sion by chloride vapors.
The concentration of the chloride deposits and
the type of chlorides can vary from boiler to
boiler. Both the concentration of chloride salts
and the type of chlorides are not normally mon-
itored in plants. Furthermore, flue gas velocity
can be an important factor in the superheater
wastage. In some cases, the superheater wastage (a)
can be the result of erosion/corrosion. The only

Metal temperature, °C
200 400 600
0.6

760 °C (1400 °F) gas temperature

0.5 855 °C (1550 °F) gas temperature 13

11
Corrosion rate, mils/h

Corrosion rate, µm/h

0.4
9
0.3
7

0.2 5

3
0.1
(b)
1
0 Fig. 12.29 Alloys 72 overlay superheater tube (a) and alloy
0 200 400 600 800 1000 1200 C276 overlay superheater tube (b) after 7200
operating hours (10 months) in a RDF unit. The windward side of
Metal temperature, °F the tube, where the flue gas impinged upon the tube surface was
the top side of the tube cross section as shown in the figure. The
Fig. 12.28 Effect of flue gas temperature on the corrosion rate tube cross section was polished and etched with nital to reveal
of carbon steel in short-time corrosion probe tests alloy 625 overlay (white portion of the metal) which was not
(10 h exposure) in an operating boiler. Source: Ref 29 etched by nital. Courtesy of Welding Services Inc.
Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/ 30/10/2007 4:11PM Plate # 0 pg 356

356 / High-Temperature Corrosion and Materials Applications

with the steam temperature of about 340 °C and the tube. As a result, the shields can experi-
(640 °F) inside the tube and about 815 °C ence temperatures as high as those of flue gas
(1500 °F) flue gas temperature, alloy 625 over- streams. The shields can thus suffer warping,
lay tubes were exposed for 152 days and alloy distortion, and creep damage in addition to high-
622 tubes for 225 days. The wastage rate was temperature corrosion. Accumulation of ash/salt
estimated to be about 175 μm/yr (7 mpy) for deposits in the crevice behind the shields can
alloy 625 and about 355 μm/yr (14 mpy) for further accelerate the corrosion attack. Loosened
alloy 622. Alloy 622 containing 13% Mo and 3% or fallen shields can cause problems by impeding
W was found to be not as good as alloy 625 with the gas flow. Tube shields are generally con-
about 9% Mo and no W but about 3.5% Nb. Both sidered to be a “sacrificial” part and are replaced
alloys contain about 21 to 22% Cr. Overlay tubes regularly during the plant maintenance shut-
made of C-276 (Ni-16Cr-16Mo) and alloy 72 down. Tube shields are sometimes made of cast
(Ni-44Cr) were also tested in an RDF unit, where stainless steels. A cast tube shield can be made
alloy 625 overlay lasted for about 11 months. much thicker than a wrought alloy sheet, thus
Both overlays were found to be not as good as enabling the tube shield to last until the next
alloy 625 overlay. The overlays of both alloys annual maintenance shutdown. Vrchota (Ref 25)
were consumed after about 10 months as shown reported that a 7.5 mm (0.3 in.) thick cast tube
in Fig. 12.29 (Ref 22). Alloy 686 (Ni-21Cr- shield made of HD stainless steel (Fe-27Cr-5Ni)
16Mo-4W) was compared with alloy 625 in a lasted for 12 months, which coincided with the
side-by-side test as overlays applied to the maintenance shutdown cycle in an RDF boiler.
waterwall (72 bar and 290 °C, or 555 °F) of a To improve the heat transfer of the tube shield,
boiler in Denmark (Ref 34). Evaluation of the silicon carbide cement was used to fill the gap
overlay waterwall samples removed from the between the shield and the superheater tube. This
boiler after 8000 h of exposure showed both 625 allows for the reinstallation of new tube shields
and 686 overlays exhibited shallow pitting with every 12 months.
pitting depth of about 50 µm.
An alternate protection method for super-
heaters is the use of tube shields, which are
typically made of Type 309, 310, and 253MA. 12.7 Summary
Figure 12.30 shows carbon steel superheater
tubes protected by metallic tube shields awaiting Materials issues related to waste-to-energy
installation at a WTE plant. The shields are boilers for burning municipal solid waste (MSW)
typically attached to the tube by mechanical for electricity generation are presented. Com-
straps and clamps with fillet welds. The tube bustion of MSW generates a very hostile envir-
shields generally do not receive adequate heat onment for waterwall tubes and superheater
transfer through the air gap between the shield tubes. The wastage rates of waterwall tubes made

Fig. 12.30 Carbon steel superheater tubes protected by metallic tube shields awaiting installation at one WTE plant.
Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/ 30/10/2007 4:11PM Plate # 0 pg 357

Chapter 12: Waste-to-Energy Boilers and Waste Incinerators / 357

of carbon or low-alloy steels have been found to Refuse Boilers, Paper No. 219, Corrosion/
be unacceptably high if no protection method is 93, NACE, 1993
used. The corrosion is believed to result from 9. R.L. Anderson, Wheelabrator Technologies
molten chloride salts. The current, widely used Inc., private communication, 2006
method for protecting the waterwalls is the use of 10. A. Wilson, U. Forsberg, M. Lundberg, and
alloy 625 overlay cladding applied by automatic L. Nylof, Composite Tubes in Waste Incin-
gas metal arc welding process. The waterwalls eration Boilers, Stainless Steel World 99
can also be constructed out of alloy 625/carbon Conference on Corrosion-Resistant Alloys
steel coextruded tubes. Superheater tubes also (Conf. Proc.), Book 2, KCl Publishing BV,
require some methods of protection against cor- The Netherlands, 1999, p 669
rosion attack. Alloy 625 overlay tubes and 11. F. Soutrel, C. Rapin, P. Steinmetz, and G.
coextruded tubes have been used for super- Pierotti, Corrosion of Fe, Ni, Cr and Their
heaters successfully in many boilers. However, Alloys in Simulated Municipal Waste
for some boilers with higher steam temperatures Incineration Conditions, Paper No. 428,
and/or more corrosive environments, alloy 625 Corrosion/98, NACE International, 1998
overlay or cladding was found to be inadequate. 12. M. Spiegel, Salt Melt Induced Corrosion of
An alternate corrosion protection method is the Metallic Materials in Waste Incineration
use of metallic tube shields or refractories. Tube Plants, Mater. Corros., Vol 50, 1999, p 373
shields are generally considered to be a “sacrifi- 13. M. Noguchi et al., Experience of Super-
cial” part and are replaced regularly during heater Tubes in Municipal Waste Incinera-
the plant maintenance shutdown. Refractories tion Plant, Mater. Corros., Vol 51, 2000,
require regular repair or replacement. p 774
14. P.L. Daniel, L.D. Paul, and J. Barna, Fire-
REFERENCES Side Corrosion in Refuse-Fired Boilers,
Mater. Perform., May 1988, p 20
1. H.H. Krause, Historical Perspective of 15. J.G. Singer, Ed., Combustion Fossil Power,
Fireside Corrosion Problems in Refuse- 4th ed., Combustion Engineering, Inc.,
Fired Boilers, Paper No. 200, Corrosion/93, Windsor, CT, 1991
NACE, 1993 16. I.G. Wright, H.H. Krause, and R.B. Dooley,
2. G. Sorell, The Role of Chlorine in High A Review of Materials Problems and Solu-
Temperature Corrosion in Waste-To-Energy tions in U.S. Waste-Fired Steam Boilers,
Plants, Mater. High Temp., Vol 14 (No. 2/3), Paper No. 562, Corrosion/95, NACE Inter-
1997, p 137 national, 1995
3. S.C. Stultz and J.B. Kitto, Ed., Steam and Its 17. A.J. Licata, L.A. Terracciano, R.W. Herbert,
Generation and Use, 40th ed., Babcock & and U. Kaiser, Design Features for Super-
Wilcox, 1992 heater Corrosion Control in Municipal
4. C.F. Knights, I.W. Cavell, and B.A. Phillips, Waste Combustors, Materials Performance
Corrosion During Incineration of a Sulfur in Waste Incineration Systems, G.Y. Lai and
and Chlorine Bearing Mixture of Rubbers G. Sorell, Ed., NACE, 1992, p 5-1
and Plastics, Werkst. Korros., Vol 40, 1989, 18. H.H. Krause and I.G. Wright, Boiler Tube
p 163 Failures in Municipal Waste-To-Energy
5. E.A. Bretz, Energy from Wastes, Power, Plants: Case Histories, Paper No. 561, Cor-
March 1990, p S-1 rosion/95, 1995
6. P. Rademakers, W. Hesseling, L.A. Tange, 19. A. Pourbaix, Corrosion of a Waste Incin-
and R. Montaigne, “Review on Corrosion in erator: Effects of Design and Operating
Waste Incinerators, and Possible Effect of Conditions, Werkst. Korros., Vol 40, 1989,
Bromine,” CEF-12, Laan van Westenenk, p 157
The Netherlands, 2002 20. A.L. Plumley, W.R. Roczniak, and E.C.
7. P.Z. Kubin, Materials Performance and Lewis, Materials Performance of Heat
Corrosion Control in Modern Waste-To- Transfer Surfaces in A MSW-Fired Incin-
Energy Boilers Applications and Experi- erator, Materials Performance in Waste
ence, Paper No. 90, Corrosion/99, NACE Incineration Systems, G.Y. Lai and G.
International, 1999 Sorell, Ed., NACE, 1992, p 7-1
8. L. Strach and D.T. Wasyluk, Experience 21. W.G. Schuetzenduebel, I.E. Johnson, and
with Silicon-Carbide Tiles in Mass-Fired C.W. Clemons, Accelerated Tube Metal
Name ///sr-nova/Dclabs_wip/High Temp/5208_335-358.pdf/Chap_12/ 30/10/2007 4:11PM Plate # 0 pg 358

358 / High-Temperature Corrosion and Materials Applications

Wastage in Municipal Solid Waste Fired Waste Incineration Systems, G.Y. Lai and
Furnaces, Materials Performance in Waste G. Sorell, Ed., NACE, 1992, p 1–1
Incineration Systems, G.Y. Lai and G. 30. I.G. Wright, V. Nagarajan, and H.H. Krause,
Sorell, Ed., NACE, 1992, p 10-1 Paper No. 201, Corrosion/93, NACE, 1993
22. Welding Services Inc., unpublished data, 31. Y. Kawahara and M. Kira, Corrosion Pre-
Norcross, Georgia vention of Waterwall Tube by Field Metal
23. P.N. Hulsizer, Problems and Solutions Spraying in Municipal Waste Incineration
in Applying Weld Overlay to Waste Boiler Plants, Corrosion, Vol 53 (No. 3), 1997,
Incinerators, Materials Performance in p 241
Waste Incineration Systems, G.Y. Lai and 32. N. Otsuka, Y. Tsukaue, K. Nakagawa, Y.
G. Sorell, Ed., NACE, 1992, p 11-1 Kawahara, and K. Yukawa, A Corrosion
24. G.Y. Lai, “Corrosion Mechanisms and Alloy Mechanism for the Fireside Wastage of
Performance in Waste-To-Energy Boiler Superheater Materials in Waste Incinerators,
Combustion Environments,” presented at Paper No. 157, Corrosion/97, NACE Inter-
the 12th North American Waste To Energy national, 1997
Conference (NAWTEC 12) (Savannah, GA), 33. B. Gleeson, J.E. Barnes, and M.A. Harper,
May 17–19, 2004 Corrosion Behavior of Various Commercial
25. S. Vrchota, Fireside Corrosion Management Alloys in a Simulated Combustion Environ-
in RDF Waste-To-Energy Boilers, Paper No. ment Containing ZnCl2, Paper No. 196,
5317, Corrosion/2005, NACE International, Corrosion/98, NACE International, 1998
2005 34. M. Montgomery and O.H. Larsen, Field
26. L.D. Paul and P.L. Daniel, Corrosion Investigation of Various Weld Overlays
Mechanisms in Oxidizing, Reducing, and in a Waste Incineration Plant, Paper No.
Alternating Combustion Gases in Refuse- 5309, Corrosion/2005, NACE International,
Fired Boiler Environments, Paper No. 216, 2005
Corrosion/93, NACE, 1993 35. D.M. DeVincentis, S.P. Goff, J.W. Slusser,
27. G.D. Smith and P. Ganesan, Metallic Z. Zurecki, and J.T. Rooney, Solving Fire-
Corrosion in Waste Incineration: A Look side Corrosion in MSW Incinerators with
at Selected Environmental and Alloy Thermal Spray Coatings, Paper No. 198,
Fundamentals, Heat-Resistant Materials Corrosion/93, 1993
II (Conf. Proc.), Second International Con- 36. J.L. Blough, G.J. Stanko, and M.T. Kraw-
ference on Heat-Resistant Materials, K. chuk, In Situ Materials Testing in a Waste-
Natesan, P. Ganesan, and G. Lai, Ed., ASM To-Energy Power Plant, Mater. High Temp.,
International, 1995, p 631 Vol 14 (No. 2/3), 1997, p 181
28. F. Devisme, P. Falgoux, F. Lefebvre, and 37. J.L. Blough, G.J. Stanko, W.T. Bakker, and
T. Flament, “High Temperature Corrosion T. Steinbeck, Superheater Corrosion in a
in Atmospheres Containing Hydrogen Boiler Fired with Refuse-Derived Fuel,
Chloride,” presented at the 11th Interna- Heat-Resistant Materials II (Conf. Proc.),
tional Incineration Conference (Albuquer- Second International Conference on Heat-
que, NM), May 11–15, 1992 Resistant Materials, K. Natesan, P. Ganesan,
29. H.H. Krause, Chlorine Corrosion in Waste and G. Lai, Ed., ASM International, 1995,
Incineration, Materials Performance in p 645
Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/ 31/10/2007 12:47PM Plate # 0 pg 359

High-Temperature Corrosion And Materials Applications Copyright © 2007 ASM International®

George Y. Lai, editor, p359-377 All rights reserved.
DOI: 10.1361/hcma2007p359 www.asminternational.org

CHAPTER 13

Black Liquor Recovery Boilers

in the Pulp and Paper Industry

13.1 Introduction The boiler construction is similar to that of

coal-fired boilers (Chapter 10), oil-fired boilers
The black liquor recovery boiler is an (Chapter 11), and waste-to-energy (WTE) boilers
integral part of the kraft pulping process in (Chapter 12) using the tube-membrane water-
the pulp and paper industry. In the kraft process, walls for the furnace enclosure, with the excep-
wood chips are cooked in an aqueous solution, tion that the furnace floor is also enclosed with a
known as “white liquor,” containing primarily tube-membrane construction. However, the
sodium hydroxide (NaOH) and sodium sulfide combustion characteristics and problems are
(Na2S) under pressure in a digester for manu- uniquely different from those of fossil-fired boi-
facturing pulp. After cooking, the pulp is lers and WTE boilers. The first and most sig-
separated from residual liquor for paper manu- nificant issue is the formation of molten smelt (or
facturing. The residual liquor, which contains molten salts) on the furnace floor, and thus a
wood lignins, organic materials, and inorganic boiler tube leak (under certain conditions) may
compounds, then undergoes a series of evapora- lead to a smelt-water explosion (Ref 4). Recov-
tion processes to increase the concentration of ery boilers have the potential for devastating
solids from approximately 15 to 70 wt% or explosions causing fatalities and injuries (Ref 4).
higher. This liquor with high concentrations of Combustion of the black liquor in the boiler
solid is known as “black liquor.” The black liquor generates a sulfur-containing, reducing furnace
is burned in a boiler to produce molten inorganic environment in the lower furnace and a pile of
material, known as molten smelt, which is rich in char bed with molten smelt (or molten salts) on
sodium carbonate (Na2CO3) and sodium sulfide the floor tube bed. The major materials problems
(Na2S). The molten smelt forms in a bed on the involved tube-wall thinning due to sulfidation
furnace floor and is discharged from the furnace when carbon steels were used and subsequent
floor into a water tank outside the furnace to form severe cracking problems for Type 304L/carbon
“green liquor” containing primarily sodium car- steel coextruded tubes when the furnace floor
bonate (Na2CO3) and sodium sulfide (Na2S). tubes and waterwall tubes were upgraded with
This green liquor undergoes the causticizing these coextruded tubes. A review of the current
reaction to convert it to white liquor consisting of understanding on the tube cracking issues and
sodium hydroxide (NaOH) and sodium sulfide improved alloys is presented in this chapter.
(Na2S) for use in cooking wood chips to manu-
facture pulp again. This chemical recovery pro-
cess in the kraft pulping is described in detail in
Ref 1 and 2. Because of its function of recovering 13.2 Fuel, Combustion, and Boilers
cooking chemicals, the boiler is often referred to
as a “chemical recovery boiler” or simply “re- Black liquor contains dissolved wood com-
covery boiler.” In addition to recovering cooking ponents, inorganic constituents (NaOH, Na2S,
chemicals for the pulping process, the boiler also Na2CO3, Na2SO4, Na2S2O3, and NaCl), and
recovers energy by producing process steam. water. An example of the elemental compositions
Grant (Ref 3) reported in 1997 that about 216 re- of black liquor from North American wood
covery boilers which were operated by 130 mills species is shown in Table 13.1 (Ref 5). Black
in the United States produced approximately liquor is injected by spraying into the boiler
41% of the total energy used by these mills. through liquor guns at about 4.6 m (15 ft) above
Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/ 31/10/2007 12:47PM Plate # 0 pg 360

360 / High-Temperature Corrosion and Materials Applications

the hearth for some boilers (Ref 6). Air for the boiler at the location above the secondary
combustion is injected into the furnace separately air ports. The liquor droplets injected into the
from primary and secondary air ports for sub- boiler are generally 0.5 to 5 mm in size (Ref 7).
stoichiometric combustion and also from tertiary The droplets should be small enough so they
air ports to complete the combustion. Figure 13.1 become dry or nearly dry before reaching the
shows a schematic diagram of a recovery boiler char bed to aviod smelt-water contact. However,
with primary, secondary, and tertiary air ports they should not be too small, otherwise the dro-
(Ref 6). The liquor is shown to be sprayed into plets can be entrained in the flue gas stream
resulting in fouling and plugging in the convec-
tion path. (Ref 7). Large droplets may travel
Table 13.1 Elemental compositions of black directly to the walls (Ref 7). In general, black
liquor from North American wood species liquor burns in approximately four stages,
Element Concentration, wt% which are drying, devolatilization (pyrolysis),
Carbon 34–39 char burning, and smelt coalescence and reac-
Hydrogen 3–5
Oxygen 33–38
tions (Ref 7). The furnace can typically be divi-
Sodium 17–25 ded into three zones: the reducing zone at the
Sulfur 3–7 bottom, the drying zone where liquor is fired, and
Potassium 0.1–2
Chlorine 0.2–2 the oxidizing zone in the upper furnace, as
Nitrogen 0.04–0.2 schematically illustrated in Fig. 13.2 (Ref 8). The
Others 0.1–0.3 figure also shows the char bed that forms on the
Source: Ref 5 furnace hearth during combustion of black
liquor.

Steam flow
to mill
Stack gas
Feedwater
Sootblowing steam

Induced
draft
fan Boiler
bank
Super-
Econo-
heater
mizer
Waterwalls

Bullnose

Electrostatic Forced draft

precipitator fan
Dust Blowdown
Furnace
Recycle
Steam coil
air heater
Liquor
Tertiary air
guns
Black
liquor Liquor
mix Secondary air
heater
tank Primary air

Smelt spouts
Strong black Makeup Direct Smelt to
liquor from salt cake heating dissolving tank
concentrator steam

Fig. 13.1 Recovery boiler. Source: Ref 6

Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/ 31/10/2007 12:48PM Plate # 0 pg 361

Chapter 13: Black Liquor Recovery Boilers in the Pulp and Paper Industry / 361

Formation of a char bed on the furnace hearth

during combustion of black liquor makes the
recovery boiler uniquely different from fossil-
Flue fired boilers and WTE boilers discussed in pre-
gas vious chapters. The char bed generally consists
of carbon, partially pyrolyzed black liquor solids,
and molten and frozen smelt (Ref 9). The char-
acteristics of char beds are different for different
furnace floor designs. The sloped (or inclined)
Oxidizing furnace floor, such as for B&W boilers, and the
decanting floor, such as for Alstom CE boilers,
exhibit different characteristics of char beds,
and their characteristics are schematically illu-
strated in Fig. 13.3 and 13.4 for the sloped floor
Secondary
air
and decanted floor, respectively (Ref 9). For the
sloped floor design, molten smelt flows from the
active char bed and is collected in troughs around
the perimeter of the bed and then discharged
Black liquor Drying through smelt spouts (Ref 9). For the decanting
floor design, molten smelt is collected and con-
tained by the decanting bottom of the boiler
(Ref 9).
Primary
air
Since the floor tubes and lower furnace
Reducing
waterwall tubes are cooled with water, a layer
of the frozen (or solidified) smelt is formed on
the tube surface, thus preventing rapid corrosion
of tube metal by the molten salts. With studded
Smelt carbon steel tubes, a thicker frozen layer is
Fig. 13.2 Recovery boiler illustrating three zones of reactions formed (Ref 10). The frozen smelt is not corro-
in the furnace. Source: Ref 8 sive as long as it remains solid on the tube
surface. However, once this frozen smelt layer
suffers localized disturbances or instability,

Active layer
• Pyrolysis, combustion, reduction
• 10–25 cm (4–10 in.) thick
Secondary • 800–1200 °C (1500–2200 °F)
air

Inactive core
• Solidified smelt
• Carbon
• <760 °C (1400 °F)

Primary
air

Hearth 230–290 °C (450–550 °F)

Solid smelt

Smelt spout Liquid smelt

Fig. 13.3 Char beds formed in the sloped floor boiler. Source: Ref 9
Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/ 31/10/2007 12:48PM Plate # 0 pg 362

362 / High-Temperature Corrosion and Materials Applications

Chemically active layer

• Pyrolysis, combustion, reduction
Secondary
• 800–1200 °C (1500–2200 °F)
air

Physically
active
Primary
layer
air

Liquid smelt
Solid smelt
Smelt spout Hearth 260–340 °C (500–650 °F)

Fig. 13.4 Char beds formed in the decanting hearth boiler. Source: Ref 9

molten smelt may come within close proximity Table 13.2 The compositions of major gas
to, and possibly in direct contact with, the floor species detected in gas samples collected from the
tubes, prompting severe tube overheating and different locations of the lower furnace waterwall
thus premature tube failures (Ref 11). Hogan between the secondary and tertiary air ports
(Ref 11) reported several floor tube overheating Composition, %
incidents; in one case the tubes were overheated Gas species At waterwall 2.5 cm (1 in.) 30 cm (12 in.)
surface away away
in excess of 675 °C (1250 °F) and in another
N2 24.8 32.9 47.6
case in excess of 720 °C (1330 °F). The author CO2 29.4 23.9 18.3
(Ref 11) attributed the overheating to localized CO 7.5 9.1 16.8
disturbances or instability of the frozen melt on H2 2.4 2.3 1.5
O2 2.9 1.0 1.1
the tube surface. The furnace combustion con- CH4 4.0 2.0 0.9
ditions can be greatly affected by many boiler H 2S 18 24.2 0.6
SO2/COS 0.2 0.27 0.04
operating factors. For example, plugging of air Methyl mercaptan 0.17 0.51 0.002
ports by unburned material or unreacted smelt
as well as plugging of liquor gun nozzles can Source: Ref 13

greatly influence the furnace combustion condi-

tions (Ref 12).
Above the char bed up to the tertiary air ports, excluded from the data in the table. In addition,
the gaseous environment remains under reducing the authors indicated that during the present
conditions. Singh et al. (Ref 13) analyzed gas study, the black liquor was intentionally sprayed
samples collected from the areas near the water- on the sidewalls for drying, thus leading to higher
wall between the secondary and tertiary air ports concentrations of H2S. It should be noted that the
in an operating boiler. The samples collected samples collected were not in equilibrium; thus,
were analyzed using the gas chromatograph oxygen was detected. The results show that the
technique. The samples were collected from the concentration of H2S was significantly reduced
area of the carbon steel waterwall that had as the probe moved away from the waterwall.
experienced high corrosion. Their results are Hydrogen sulfide (H2S) is the major sulfur
summarized in Table 13.2 (Ref 13). The amount species of the gaseous sulfur compounds gener-
of water vapor (% H2O) in the environment was ated in the lower furnace. Other gaseous sulfur
Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/ 31/10/2007 12:48PM Plate # 0 pg 363

Chapter 13: Black Liquor Recovery Boilers in the Pulp and Paper Industry / 363

compounds include methyl mercaptan, dimethyl boilers in Scandinavia (34 in Sweden and
sulfide, and dimethyl disulfide. These com- Norway and 22 in Finland) had coextruded tube
pounds are frequently referred to as “total re- walls and more than half had coextruded tube
duced sulfur” (TRS) gases. In the lower furnace, floors, and in North America about 65 recovery
the concentration of TRS gases is typically in a boilers out of approximately 340 kraft recovery
range of few hundred to a thousand ppm (Ref 14). boilers had coextruded tube floors and many
As the flue gas stream rises to the upper fur- others had coextruded tube walls. One major
nace and enters into the superheater bundles, the U.S. boiler manufacturer (Babcock & Wilcox)
oxidizing atmosphere converts H2S to SO2. The installed coextruded tubes in the smelt flow areas
flue gas stream is also entrained with fly-ash adjacent to the sidewalls with carbon steel stud-
particles, which include carryover of black liquor ded tubes in the center of the floor in 22 boilers
droplets and fume that is formed by condensation and in the complete floor for 26 boilers (Ref 18).
of volatilized inorganic salts (Ref 15). These fly- In the 1990s, cracking of the 304L cladding of
ash particles deposit on the cooler surfaces of the composite tubes had become a major mate-
the heat-exchanger tubes, such as superheaters rials issue in recovery boilers, although some
and generating banks, in the convection path. cracking was reported in the 1980s (Ref 19, 20).
The deposits consist mainly of sodium sulfate In 1995, a United States Department of Energy
(Na2SO4) and sodium carbonate (Na2CO3) with (DOE) program was established to determine
small amounts of sodium sulfide (Na2S), sodium the cause of the tube cracking and to identify
chloride (NaCl), and potassium salts. The first alternative materials or process changes to pre-
melting temperature (FMT) of the deposits typi- vent this type of cracking. This project, coordi-
cally varies between 520 and 580 °C (970 and nated by J.R. Keiser of the Oak Ridge National
1075 °F), depending on composition (Ref 16). Laboratory (ORNL), was carried out by re-
The superheater tube can suffer accelerated searchers at ORNL, the Pulp and Paper Research
wastage when the deposit in contact with the tube Institute of Canada (PAPRICAN), the Institute of
surface exceeds its FMT (Ref 17). Paper Science and Technology (IPST) with
support of more than a dozen paper companies,
and boiler and tubing manufacturers (Ref 20).
The results of the research projects under this
13.3 Materials Problems in Lower program, which were presented at regular ORNL
Furnace and Superheaters review meetings and published in various
technical journals, provided a significant under-
The first Tomlinson recovery boiler built in standing on the root causes of the 304L compo-
1929 had refractory furnace walls that were too site tube cracking problems. This chapter
costly to maintain (Ref 1). The furnace design reviews (a) the cracking problems of the 304L
was later changed to a completely water-cooled cladding in composite tubes used as floor tubes,
furnace enclosure including a furnace floor made smelt openings, and primary air port openings
of carbon steel, with the first such unit built in and (b) alternative materials. Superheater corro-
1934 (Ref 1). Lower furnace waterwalls and floor sion is also included in the discussion.
tubes were protected by pin studs that held frozen
smelt providing protection against molten smelt
13.3.1 Furnace Floor Tubes
(Ref 1). There have been several incidents of pin
stud floor tube wastage and failure involving Cracking of the 304L cladding has been found
both sloped floor and decanting floor boilers (Ref to occur in (a) both high-pressure (8.3 to
18). Clement and Blue (Ref 18) characterized the 10.3 MPa, or 1200 to 1500 psi) and low-pres-
failure of the pin stud carbon steel floor tubes to sure (4 to 6.2 MPa, or 600 to 900 psi) boilers, (b)
be stud wastage, or burn-back, and tube-wall both sloped floor and decanting floor designs,
thinning. Since the 1970s, an increasing number and (c) boilers by different boiler manufacturers
of mills, first in Scandinavia and then in North (Ref 19). Cracks typically initiated from the
America, gradually upgraded carbon steel outer diameter of the clad tube and propagated
waterwalls and floors in the lower furnace to inward to the substrate steel; in most cases, the
coextruded tubes with Type 304L outer cladding cracks terminated at the cladding/steel interface
to provide corrosion protection for the inner with some cracks changing the direction of
carbon steel tube (Ref 19). Singbeil et al. (Ref 19) propagation along the cladding/steel interface
reported in 1997 that most of the 56 recovery (Ref 19–21). This is illustrated in Fig. 13.5 for the
Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/ 31/10/2007 12:48PM Plate # 0 pg 364

364 / High-Temperature Corrosion and Materials Applications

1250 µm
500 µm

Fig. 13.5 Cracks initiated on the outer diameter of the 304L Fig. 13.6 Cracks initiated on the outer surface of the 304L
cladding, propagated inward to the substrate steel
clad tube, propagated inward to the substrate steel
of the membrane and terminated at the cladding-steel interface.
and terminated at the cladding-steel interface. Courtesy of Oak
Courtesy of Oak Ridge National Laboratory.
Ridge National Laboratory.

clad tube and Fig. 13.6 for the clad membrane. Thermal fatigue data for Type 304H and 304L
These cracks were found to be essentially trans- cycling to 500, 550, and 600 °C (930, 1020, and
granular in nature (Ref 19, 20). In order to 1110 °F), respectively, were generated and com-
determine the cracking mechanisms, the DOE pared with the ASME Sec. III Subsec. NH de-
program included the following studies: (a) sign curve for Type 304H at 430 °C (800 °F)
measurement and modeling of residual stresses in isothermal fatigue. All the thermal fatigue
and the stress states of the composite tubes, (b) data fell near or above the ASME design curve
tube metal temperature measurements during (Ref 20, 23). Assuming thermal cycling to
operation, (c) thermal fatigue testing of tube 450 °C (840 °F) is considered, the cyclic strains
materials, and (d) the effect of waterwashing would be about 0.25% and the fatigue life for the
during shutdowns (Ref 20). stainless steel cladding is expected to be in excess
To determine whether thermal fatigue was a of 100,000 cycles (Ref 23).
possible cause of floor tube cracking, floor tube There is some thermal expansion mismatch
metal temperature was measured in a boiler at the between the 304L cladding and the carbon steel
Weyerhaeuser mill in Prince Albert, Saskatch- substrate in the composite tube due to differences
ewan, Canada, during September 1, 1998 in mean coefficients of thermal expansion
through February 28, 1999 (Ref 22). This boiler between these two alloys (e.g., 7.6× 10−6 and
with a sloped floor had experienced cracking of 10.0 × 10−6 in./in. · °F from 70 to 800 °F for
Type 304L cladding in coextruded floor tubes. In carbon steel and Type 304, respectively). The
many boilers, cracking had occurred within 2 m residual stresses formed in the cladding in both
of the rear (spout) wall. Keiser et al. (Ref 22) the as-fabricated condition and after service were
installed 25 thermocouples on the floor and studied under the DOE program. Also investi-
recorded the temperature spikes over a 6-month gated in the program was finite element modeling
period, showing an average of about one thermal of stresses in the cladding. Keiser et al. (Ref 20)
spike per thermocouple per day. A thermal spike summarized the results of both residual stress
was defined as a temperature spike of more measurements and finite element modeling of the
than 50 °C (90 °F) above approximately 275 °C 304L/SA210 coextruded composite tubes. Resi-
(530 °F). The authors observed that the thermo- dual stress measurements using x-ray and neutron
couples located near the spout wall that suffered diffraction showed compressive axial and tan-
tube cracking showed the fewest thermal spikes, gential stresses on the outer surface of the clad-
while the areas farther from the spout wall that ding of the as-fabricated tube, and tensile axial
exhibited little or no cracking experienced the and hoop stresses (up to a maximum of 300 MPa)
greatest numbers of thermal spikes (Ref 22). The on the crown location of the coextruded tube after
authors concluded that the frequency of thermal service in the boiler floor (Ref 20). Finite element
spikes was far too low for thermal fatigue to be modeling studies showed that the 304L cladding
the cause for the cracking (Ref 22). developed tensile stresses when cooled down
Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/ 31/10/2007 12:48PM Plate # 0 pg 365

Chapter 13: Black Liquor Recovery Boilers in the Pulp and Paper Industry / 365

from the “operating” temperature (e.g., 300 °C, (modified alloy 825) coextruded tube (on carbon
or 570 °F) to room temperature (Ref 20). Fur- steel), Type 309 overlay tube, alloy 625 overlay
thermore, the 304 cladding developed tensile tube, Type 310 coextruded tube, and carbon
stresses when cooled down from the thermal steel (Ref 20). The test results of these tubes are
spike to the “operating” temperature (e.g., cool- summarized: (a) Type 309 overlay tube was
ing down from 550 to 300 °C, or 1020 to 570 °F) found to be as bad as the 304L coextruded tube,
(Ref 20). As a result, tensile stresses can be pre- (b) Type 310 coextruded tube also suffered
sent in the 304 cladding at “operating” tempera- SCC cracking with the longest crack being
ture after a thermal spike as well as when floor about 325 µm, (c) Sanicro 38 cladding (modified
tubes are cooled to room temperature. alloy 825) exhibited cracks, but less than
The presence of tensile stresses in the 304L 50 µm deep, (d) alloy 625 overlay showed no
cladding when the composite floor tubes are at cracking, and (e) carbon steel showed no crack-
low temperatures is significant in that it can make ing (Ref 20).
stress-corrosion cracking (SCC) a potential One of the conditions required for developing
cracking mechanism since the morphology of stress-corrosion cracking is the presence of
cracking resembles that of SCC. Prescott and tensile stress. In this regard, alloys 825 and 625
Singbeil (Ref 24) reported that Type 304L suf- have advantages over Type 304L in that both
fered stress-corrosion cracking under tensile exhibit coefficients of thermal expansion much
stresses when exposed to concentrated solutions closer to those of carbon steels. Instead of
of alkali compounds containing Na2S at tem- developing axial tensile stresses in Type 304L
peratures from 160 to 200 °C (320 to 390 °F). In when cooled from the operating temperature
C-ring tests of Type 304L/SA210A1 coextruded to room temperature, both alloys 825 and
tubes in a salt consisting of 90% Na2S·9H2O and 625 develop axial compressive stresses (Ref
10% NaOH at about 170 °C (340 °F) for 48 h, 23). This is important in that cracking pre-
cracks developed at the outer diameter of the dominantly occurred in the transverse direction
cladding and propagated inward and terminated (Ref 23). The alloy 625 overlay in the as-overlaid
at the cladding/substrate interface (Ref 20). tube typically exhibits residual tensile stresses
Stress-corrosion cracking of Type 304L takes that can be eliminated by annealing the tube
place in this type of environment within a certain at 900 °C (1650 °F) for 20 min (Ref 28). The
temperature range. For example, in constant load overlay tubing is produced by a spiral overlay
tests with 275 MPa (40 ksi) in Na2S-10%NaOH, welding mode using gas metal arc and gas
stress-corrosion cracking of Type 304L readily tungsten arc welding processes (Ref 29), and its
occurred at 150 to 250 °C (300 to 480 °F) structure and properties are reported elsewhere
(Ref 23). It is generally believed that water- (Ref 30, 31).
washing of the boiler when there are significant Keiser et al. (Ref 26) reported the results of the
remnants of smelt on the floor results in aqueous tests comparing austenitic stainless steel clad-
solutions that contain Na2S along with NaOH or dings of Type 304L coextruded tubes and Type
Na2CO3. These aqueous solutions are believed 309L overlay tubes with nickel-base alloy clad-
(Ref 20, 23, 25, 26) to be an essential requirement dings of alloy 825 coextruded tubes, alloy 625
for stress-corrosion cracking of the 304L clad- coextruded tubes, and alloy 625 overlay tubes in
ding either during the boiler shutdown or startup a recovery boiler in North America. Both 304L
when the floor tubes are covered with these salt coextruded tubes and 309L overlay tubes
solutions and subject to this SCC temperature showed cracking during the second year of ser-
range. The critical temperature range is believed vice, while alloy 825 coextruded, 625 coex-
to be 150 to 200 °C (300 to 400 °F) (Ref 23). truded, and 625 overlay tubes showed no
C-ring tests were conducted in a salt con- cracking after 5 years of service. Barna and
sisting of 90% Na2S·9H2O and 10% NaOH Rivers (Ref 21) reported in their 1999 paper
at about 170 °C (340 °F) for 48 h for Type that several floors have been constructed with
304L/SA210A1 coextruded tube, Sanicro 38* alloy 825 coextruded tubes since 1996. Wilson
et al. (Ref 32) reported good performance
results of Sanicro 38 coextruded tubes as floor
tubes in recovery boilers. The authors (Ref 32)
* Sanicro 38 meets UNS N08825 (alloy 825) according to
Sandvik Steel's technical bulletin on Sanicro 38/4L7 coex-
indicated that the outer cladding alloy in Sanicro
truded tubes (Ref 27), thus the cladding should be considered 38 coextruded tube was a modified alloy 825.
alloy 825, not modified alloy 825. Nevertheless, the chemical composition of
Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/ 31/10/2007 12:48PM Plate # 0 pg 366

366 / High-Temperature Corrosion and Materials Applications

the outer cladding published in the paper 13.3.2 Smelt Spout Openings
was within the specification of UNS N08825
(alloy 825). Sandvik Steel’s technical bulletin Cracking of Type 304L coextruded tubes
(Ref 27) on Sanicro 38/4L7 composite tubes was initially observed at smelt openings in the
also indicates that the cladding meets UNS early 1980s before the observation of cracking of
N08825. Type 304L coextruded floor tubes (Ref 21).
Lai and Wensley (Ref 31) reported the per- Cracks can occur on adjacent tubes that form the
formance experience of alloy 625 overlay floor spout wall (Ref 21). The circumferential cracks
tubes in several boilers where good results were that occur in the 304L coextruded tubes at the
observed. One boiler (a Babcock & Wilcox smelt openings have been found to propagate
design) operated reportedly at 1420 psig and from the outer stainless steel cladding through
produced 883,400 lb/h steam. After severe the interface and into the carbon steel substrate
cracking of Type 304L coextruded tubes in 1995, (Ref 19). Smelt openings are subject to thermal
alloy 625 overlay tubes were installed in the fluctuations due to intermittent exposure of the
smelt runs in 1997 replacing the 304L coex- flowing smelt through the spout (Ref 21, 33). The
truded tubes. Figure 13.7 shows a general view of wall tubes including the tubes adjacent to spout
the smelt run floor tubes in the boiler (Ref 31). openings are also subject to thermal fluctuations
The alloy 625 overlay tubes had been in service due to rising and falling of the surface of the
for about 7 years with no reported material pro- smelt pool (Ref 18). It is believed that thermal
blems when the paper was published in 2005. fatigue cracking may be a contributing factor for
Barna and Rivers (Ref 21) also reported the cracking of smelt opening tubes made of Type
installation of alloy 625 overlay tubes for boiler 304L coextruded tubes (Ref 18). Dykstra et al.
floors in several boilers. One installation (Ref 34) attributed the cracking of Type 304L
involved covering the entire width of about one- coextruded smelt spout opening tubes to thermal
half the boiler floor in 1996, and three other fatigue. The authors (Ref 34) indicated that
boilers were fitted with partial floor installations fractographic examination of the crack surfaces
of alloy 625 overlay tubes in 1995 and 1997. No showed evidence of cyclic crack progression.
cracks or other damage were reported when the In addition to the thermal fatigue type cracks,
paper was presented in 1999 (Ref 21). craze cracks were also observed in Type 304L

Fig. 13.7 General view of the alloy 625 overlay smelt run floor tubes in the boiler. Source: Ref 31
Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/ 31/10/2007 12:48PM Plate # 0 pg 367

Chapter 13: Black Liquor Recovery Boilers in the Pulp and Paper Industry / 367

coextruded spout opening tubes (Ref 34). These 30,000 cycles and showed no cracking. The test
craze cracks (typically branched) were found to results showed that both alloy 825 coextruded
terminate at the cladding/steel interface (Ref 34). tubes and alloy 625 overlay tubes were very
In order to investigate alternate cladding resistant to thermal fatigue cracking.
alloys, Clement and Blue (Ref 18) conducted Barna and Rivers (Ref 21) reported in 1999
thermal fatigue testing of alloy 825 coextruded that there were about 60 smelt openings that were
tube and alloy 625 overlay tube compared with made of alloy 825 coextruded tubes. However,
Type 304L coextruded tube and Type 304L it has been reported that cracking of alloy 825
monolithic tube. The substrate tubes for the cladding was encountered in some boilers
above three composite tubes were all SA210 A1 (Ref 21). Furthermore, the cracks were found to
steel. The test involved a water-cooled test tube terminate at the fusion boundary or turn at 90°
that was rotated to make the tube-wall tempera- and proceed along the fusion boundary (Ref 21).
ture fluctuate from the cladding temperature at The authors (Ref 21) also reported that alloy
the point of flame impingement to a low of about 625 overlay tubing was first installed in a smelt
40 °C (105 °F). The temperature of the cladding opening in 1987 in a boiler where Type 304L
at the point of flame impingement was measured coextruded tube smelt openings required repla-
inside the cladding at the point about 0.25 mm cement approximately every 6 months as well as
(10 mils) from the cladding surface. The clad- several other installations including one boiler
ding metal temperature cycled from about where alloy 625 overlay smelt opening tubes
650 °C (1200 °F) (high) to 40 °C (105 °F) were installed in 1994. Lai and Wensley (Ref 31)
(low). Solid 304L tubes failed in less than 5000 reported that a smelt spout wall (including
cycles, and the 304L clad tubes failed in slightly opening tubes) made of alloy 625 overlay tubes
over 10,000 cycles to less than 10,000 cycles. was installed in a B&W 1050 psi boiler in 2000
Alloy 825 coextruded tubes cycled for 25,000 replacing Type 304L coextruded tubes that had
cycles showing no cracks, and alloy 625 overlay experienced cracking problems. At the time the
tubes cycled for 30,000 cycles showing no authors (Ref 31) presented their paper in
cracks. Both 825 coextruded tubes and 625 2005, no cracking problems had been reported.
overlay tubes were then cycled from 815 °C Figure 13.8 shows that the alloy 625 smelt spout
(1500 °F) (high) to 40 °C (105 °F) (low) for wall including opening tubes revealed no

Fig. 13.8 Alloy 625 overlay smelt spout wall including opening tubes after 2 years of service showing no indication of cracking by
liquid dye penetrant testing. Source: Ref 31
Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/ 31/10/2007 12:48PM Plate # 0 pg 368

368 / High-Temperature Corrosion and Materials Applications

cracking in dye penetrant testing after 2 years of temperature excursions (spikes) and corrosion
service (Ref 31). reactions with the tube surface.
Shenassa et al. (Ref 39) indicated that the air
port design and fabrication may be a factor in
causing primary air port cracking problems. The
13.3.3 Air Port Openings authors (Ref 39) discussed two air port opening
Cracking in both craze (or mosaic) and cir- designs (the casting design and the nonwelded
cumferential patterns has been observed in Type insert design) provided by a major European
304L coextruded tubes at the air port openings boiler designer. The opening is formed by
(Ref 35–38). These authors observed that some bending one or two composite tubes out of the
of the circumferential cracks penetrated into the plane of the waterwall toward the cold side of the
carbon steel tube. Most cracks, however, did not boiler. It was observed that the casting design,
progress into the carbon steel tube (Ref 36). which involved a cast iron casting being bolted to
Almost all the cracking was found to occur at the the bent tubes from the cold side, appeared to
bottom half of the air port openings (Ref 36, 38). show better resistance to cracking because of a
Temperature measurements of the air port low bending angle and the optimum bending
opening tubes in several mills have indicated that radius of the opening tubes (Ref 39). This
those air port opening tubes that had historical observation was consistent with what was
cracking problems showed the greatest tem- observed by Keiser et al. (Ref 36), who observed
perature fluctuations in frequency and ampli- that opening tubes with a shorter, wider design
tudes (Ref 36). The air port opening tubes with appeared to suffer more frequent cracking pro-
no historical cracking problems exhibited no blems than the ones with a longer, narrow
significant thermal fluctuations (Ref 36). In one opening design. In a search for alternate alloys in
example, based on the 5-day temperature mea- coextruded tubes to replace Type 304L coex-
surements, the air port opening tubes with truded tubes in primary air port openings with the
no history of cracking problems in one mill same air port design, alloy 625 coextruded tubes
showed temperature fluctuations between 300 were installed. After 1 year of operation, crack-
and 400 °C (570 and 750 °F), while the other ing was observed on many of the primary air port
mill with a history of cracking problems showed opening tubes (Ref 39). Alloy 825 coextruded
extensive thermal fluctuations from about tubes have been installed in primary air port
350 °C (660 °F) to temperatures exceeding openings in several boilers including one unit
500 °C (930 °F) (Ref 36). with alloy 825 coextruded primary air port
In making mill visits to inspect the air port openings having been in operation for 3 years; no
openings, the authors (Ref 36) observed that cracks had been observed on these air port
the air port opening tubes with a shorter, wider opening tubes (Ref 39).
design appeared to suffer more frequent cracking Keiser et al. (Ref 36) reported that the
problems than the ones with a longer, narrow inspection of a mill revealed extensive thinning
opening design. This certainly suggests that at the fireside of alloy 825 coextruded air port
the cold-worked conditions of the tube bends opening tubes with no evidence of cracking. Lai
and/or the increased stresses due to increased and Wensley (Ref 31) reported severe tube
constraint might be a factor. The results of the thinning for primary air port openings fabricated
residual stress measurements using x-ray and from spiral overlay Fe-20Cr-1Nb composite
neutron diffraction methods showed that the tubes after 1 year of service in a mill in South
304L coextruded tube typically exhibits com- America. A close-up view of the ferritic alloy
pressive axial stresses on the surface of the overlay tubes is shown in Fig. 13.9. Keiser et al.
cladding at the unbent section and tensile axial (Ref 37) reported a case where the air port
stresses in some areas of the bent section (Ref opening made of alloy 625 overlay tubes was
37). One interesting observation was made with a found to suffer overlay thinning of about 1 mm
video camera used to record the air port opening (0.04 in.) after the first 6 months of service.
activities in operating boilers. What appeared to However, no further additional wastage was
be molten smelt, which was found on the opening observed during subsequent exposure (Ref 37).
tube surfaces, flowed down the tube surfaces In fact, Wensley (Ref 40) reported that there was
(Ref 37). This suggests that molten smelt could no further overlay thinning after an additional 5
likely be on air port opening tube surfaces during years of operation for this boiler. It is believed
operation, thus resulting in possible local that the initial corrosion was likely to result from
Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/ 31/10/2007 12:48PM Plate # 0 pg 369

Chapter 13: Black Liquor Recovery Boilers in the Pulp and Paper Industry / 369

changes in liquor firing practices (Ref 40). Fur- tubes (Sanicro 63) in a mill suffered cracking
thermore, no cracking has been encountered for after only 1 year of operation. Cracking was
these air port opening tubes (Ref 40). Lai and observed in many of the primary air port open-
Wensley (Ref 31) summarized the performance ings (Ref 39). Keiser et al. (Ref 37) observed
experience of alloy 625 overlays in primary air that the circumferential cracks that occurred in
port openings. In a B&W boiler operating at alloy 625 coextruded tubes were intergranular
1420 psig and producing 883,400 lb/h steam, in nature (i.e., cracking along grain boundaries)
about 29 air port openings were replaced with (Fig. 13.11). Intergranular cracking was also
alloy 625 overlay tubes in 1999. These tubes
have been inspected annually using dye pene-
trant testing (PT). No cracking was reported
at the time the authors presented their paper
(Ref 31). Figure 13.10 shows some of these alloy
625 overlay air port opening tubes. In another
boiler, alloy 625 overlay air port opening tubes
were installed in 2000, and the inspection of
these tubes in 2002 did not reveal any cracks or
corrosion attack (Ref 31). However, inspection a
few years later showed cracks were developed in
few of these overlay tubes, and these cracks had
advanced at least to the carbon steel (Ref 41).
The performance experience of alloy 625
coextruded tubes used in air port openings
has been reported by several authors (Ref 36,
37, 39). Shenassa et al. (Ref 39) reported that air
port openings made of alloy 625 coextruded

Fig. 13.10 Primary air port openings fabricated from alloy

625 overlay tubes have been in service
since 1999. No cracking or corrosion has been observed.
The photo was taken after liquid dye penetrant testing. Source:
Ref 31

Fig. 13.9 Primary air port opening made of Fe-20Cr-1Nb

spiral overlay tubes suffered severe fireside Fig. 13.11 Intergranular cracking that occurred in the alloy
wastage after 1 year of service in a mill in South America. Source: 625 coextruded air port opening tube. Courtesy
Ref 31 of Oak Ridge National Lab.
Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/ 31/10/2007 12:48PM Plate # 0 pg 370

370 / High-Temperature Corrosion and Materials Applications

observed in alloy 625 coextruded floor tubes and Southeast (U.S.) was removed from the boiler
smelt opening tubes (Ref 42). and examined for the cold-work condition at
Primary air port openings made of alloy 625 different locations of the tube. The examination
coextruded tubes in a boiler in the Southeast was carried out by conducting microhardness
(U.S.) were also found to suffer primarily circum-
ferential cracking after 1 year of service (Ref 43).
The cracking was also found to be intergranular
(Fig. 13.12). However, these circumferential
cracks were found to terminate at the cladding/
steel interface and to change the direction of
propagation to follow the cladding/steel inter-
face, as shown in Fig. 13.13 (Ref 43). Corrosion
products formed inside the crack at the top portion
(Fig. 13.14) and at the crack tip (Fig. 13.13b)
were analyzed using energy-dispersive x-ray
spectroscopy (EDX). In the top portion of the
crack, the corrosion products appeared to be pri-
marily Ni-rich oxides (Fig. 13.14). The oxides
formed at the crack tip (Fig. 13.13b) were Ni-Cr
rich (area No.3) before the crack reached the
cladding/steel interface and Fe-Ni rich (area No.1
and 2) when the crack propagated along the
cladding/steel interface. It is more important to
note that both Na and K were detected in these
corrosion products, although it is not clear what
(a) 600 µm
role these alkali metals played in cracking.
Very little data have been reported on the cold-
work condition of the alloy 625 coextruded tube
that was formed for the air port opening. A pri-
mary air port opening tube made of alloy 625
coextruded tube that suffered circumferential
cracking after 1 year of service in a boiler in

(b) 11 µm

Fig. 13.13 Scanning electron micrographs (backscattered

electron image) showing a crack penetrated
through the 625 cladding (of a coextruded tube) and then termi-
nated at the cladding/steel interface at low magnification (original
micrograph 35×) (a), and at high magnification showing the crack
changed direction and followed the cladding-steel interface (b).
Semiquantitative EDX analysis (wt%) of the corrosion products on
the crack tip is summarized below. Courtesy of Welding Services
Inc.
Fig. 13.12 Intergranular cracking of the alloy 625 coex- 1: 52.0% Fe, 29.4% Ni, 9.5% Cr, 3.4% K, 1.4% Na, 0.8% S, and
truded air port opening tube in a mill in Southeast trace elements
(U.S.) after 1 year of service. The microstructure was not etched, 2: 61.3% Fe, 25.8% Ni, 7.6% Cr, 2.1% K, and trace elements
but cracking could be traced to follow austenitic grain bound- 3: 14.0% Fe, 50.3% Ni, 24.4% Cr, 4.0% K, 2.2% Na, and trace
aries. Courtesy of Welding Services Inc. elements
Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/ 31/10/2007 12:48PM Plate # 0 pg 371

Chapter 13: Black Liquor Recovery Boilers in the Pulp and Paper Industry / 371

measurements across the cross section of the residual stress of the alloy. This condition
cladding at 0.25, 0.50, 0.75, 1.00, and 1.25 mm prompts the alloy to undergo local plastic
(0.010, 0.020, 0.030, 0.040, and 0.050 in.) from deformation to relieve the stresses. However,
the cladding surface (Ref 43). Vickers hardness when the grain matrix is strengthened by fine
tester with a 500 g load was used for micro- precipitates, it cannot allow plastic deformation
hardness measurements. Vickers hardness values to occur to relieve those stresses, thus causing the
(HV) were then converted to Rockwell C scale alloy to develop cracking at grain boundaries,
(HRC). The schematic of the air port opening which are often the weakest locations. Austenitic
tube sample along with measured hardness at stainless steels and nickel-base alloys that form
different locations is shown in Fig. 13.15. homogeneous, coherent precipitates, such as γ 0
According to the plant personnel, the air port (Ni3Al) or γ 00 (Ni3Nb), are more susceptible to
opening tubes were reportedly not stress relieved this type of brittle cracking. Detailed discussion
or annealed after cold forming and prior to on this subject is covered in Chapter 14 “Stress-
installation (Ref 43). Circumferential cracks Assisted Corrosion and Cracking.”
were found to develop at the crown location (i.e.,
area with the highest heat flux) where the tube 13.3.4 Waterwalls above “Cut Line”
section was bent in the bottom portion of the air
port opening. This area showed hardness of The waterwalls above the butt weld joints with
about 36 to 39 HRC near the cladding surface composite tubes typically are bare carbon steel.
(about 0.25 mm, or 0.01 in., from the cladding Corrosion of these waterwall carbon steel tubes is
surface), which were slightly lower than either primarily caused by sulfidation by reduced sulfur
the extrados (the exterior bend section) (40 HRC) gases, primarily H2S. The corrosion rate is typi-
or the intrados (the inner bend section) (42 HRC) cally 0.2 mm/yr (8 mpy), but can be as high as
of the bend. The straight section was found to be
about 30 HRC (about 0.25 mm, or 0.01 in., from
the cladding surface). The microhardness mea-
surements of this air port opening sample showed
that the area that suffered cracking was in a cold-
worked condition.
1
From the above discussion, it appears that
cracking of the primary air port openings tends to
2
occur at the locations where the tube is bent. The
cladding in these locations is under a cold-
worked condition. The material at these locations
is characterized by high residual stresses. Some 3
alloys with high residual stresses (e.g., in a highly
cold-worked condition) can suffer brittle, inter-
4
granular cracking when heated to or in service at
temperatures of 425 to 650 °C (800 to 1200 °F),
depending on the alloy. The morphology of this
type of intergranular cracking, often referred to
as “stress-relaxation cracking,” is similar to the
morphology of cracking observed in Type 304L 21 µm

and alloy 625 claddings of the coextruded pri- Fig. 13.14 Scanning electron micrographs (backscattered
mary air port tubes. This cracking mechanism, electron image) showing the top portion of the
crack (shown in Fig. 13.13a) that penetrated through the 625
which has not been discussed by any of the cladding (of a coextruded tube) and then terminated at the clad-
authors, may play a role in the air port opening ding/steel interface and then changed direction and followed the
cracking. The process of stress-relaxation cladding/steel interface. Semiquantitative EDX analysis (wt%) of
the corrosion products on the top portion of the crack is sum-
cracking can be described in a simple way: when marized below. Courtesy of Welding Services Inc.
a heavily cold-worked alloy with high residual 1: 55.6% Ni, 20.9% Cr, 4.7% Fe, 4.6% S, 2.6% Na, 7.1% Si,
1.9% Ca, 1.5% Al, and trace elements
stresses (i.e., the residual stress in the alloy in the 2: 84.6% Ni, 5.2% Cr, 3.1% Fe, 2.8% Na, 1.6% Si, 1.0% S, and
cold-worked condition is essentially its room- trace elements
temperature yield strength) is heated to 540 °C 3: 85.6% Ni, 5.7% Cr, 3.1% Fe, 2.2% Na, 1.0% Si, 0.7% S, and
trace elements
(1000 °F), for example, the yield strength of the 4: 71.1% Ni, 18.3% Cr, 3.5% Fe, 2.3% Na, 1.2% Si, 1.6% S, and
alloy at that temperature will be lower than the trace elements
Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/ 31/10/2007 12:48PM Plate # 0 pg 372

372 / High-Temperature Corrosion and Materials Applications

0.8 mm/yr (32 mpy) (Ref 10). When sulfidation potentially contribute to the deposits on super-
becomes excessive, one effective approach is to heater platens. The deposits form primarily
apply Type 309 overlay using automatic overlay on the windward side of the tube. When the
welding process. Instead of forming FeS on tube metal temperature reaches the melting
unprotected carbon steel, protective Cr2O3 scales temperature of the deposits, corrosion of the
form on the 309 overlay containing typically superheater can become serious. Corrosion is
20% Cr. Numerous papers have been published more serious on the windward side than the lee-
on the beneficial effects of chromium on sulfi- ward side of the superheater tube. This is illu-
dation resistance of alloys (see Chapter 7 “Sul- strated Fig. 13.19 (Ref 46). Sulfidation/oxidation
fidation”). Moberg et al. (Ref 44) illustrates the is the major corrosion mechanism for super-
beneficial effect of chromium on the sulfidation heaters (Ref 10). Common superheater alloys are
resistance of steels containing chromium in a T-11 or T-22. When T-11 or T-22 superheater
sulfidizing environment at 400 °C (750 °F) (Fig. tubes suffer high wastage rates, one cost-effec-
13.16). Lai and Hulsizer (Ref 45) reported that tive solution will be to switch to austenitic
Type 309 overlay was applied on the waterwalls stainless steels that are capable of forming pro-
of a recovery boiler in a Midwest mill (U.S.) in tective chromium oxide scales instead of for-
1987, 1989, and 1992. Figure 13.17 shows the mation of iron oxides and/or sulfides on
cross section of a Type 309 overlay sample unprotected T-11 or T-22 tubes.
obtained from the overlaid waterwall after 13 An example was given by Lai and Wensley
years of service, showing essentially the original (Ref 31) for a boiler in South America, where
overlay thickness with only tiny surface corro- T-11 superheater tubes suffered severe wastage at
sion pits (Ref 45). A close-up view of the 309 the bend sections and required annual replace-
overlay on the front wall after 8 years of boiler ment. A T-11 superheater tube bend sample was
operation in the same boiler is shown in Fig. removed for metallurgical evaluation after 6
13.18. months of service. The sample is shown in
Fig. 13.20. The wastage rate was estimated to be
approximately 3.9 mm/yr (154 mpy). Super-
13.3.5 Superheater Tube Corrosion heater steam temperature and pressure were
As the combustion flue gas rises to the upper reportedly 435 °C (815 °F) and 6.4 MPa (63.5
furnace, black liquor droplets entrained in the bars, or 930 psig), respectively. Examination of
flue gas stream as carryover particles can the sample revealed that the corrosion products

33 HRC at 0.01 in. 39 HRC at 0.01 in.

(31–33 HRC) (36–39 HRC)
37 HRC at 0.01 in. 36 HRC at 0.01 in.
(37–39 HRC) (35–37 HRC)

42 HRC at 0.01 in.

(40–42 HRC)

Cracks

40 HRC at 0.01 in.

(40–42 HRC)
40 HRC at 0.01 in.
(38–40 HRC)
30 HRC at 0.01 in.
(28–30 HRC)

Fig. 13.15 Air port opening made of alloy 625 coextruded tube that suffered cracking after 1 year of service in a boiler in Southeast
(U.S.) with hardness data at different locations of the tube. Cracking occurred at the bottom half of the opening (right-hand
side of the tube). Rockwell C hardness values (converted from Vickers microhardness values) were reported in a range at different locations
of the tube as well as the hardness at 0.25 mm (0.01 in.) from the cladding surface. Courtesy of Welding Services Inc.
Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/ 31/10/2007 12:48PM Plate # 0 pg 373

Chapter 13: Black Liquor Recovery Boilers in the Pulp and Paper Industry / 373

14
400 °C (750 °F)

12
Rate of weight gain, g/m2/h

10

0
0 5 10 15 20 25 30
Chromium, %

Fig. 13.16 Effect of chromium on sulfidation resistance of steels containing various amounts of chromium tested at 400 °C (750 °F) in
N2-15H2O-10CO2-10H2-0.1O2-0.1H2S. Source: Ref 44

were essentially iron oxides but with the presence

of chlorine. Figure 13.21 and 13.22 show the
corrosion products and the energy-dispersive
x-ray spectroscopy (EDX) analysis of the cor-
rosion products. The results of the EDX analysis
indicated that the corrosion products were
essentially iron oxides with no sulfur. However, a
relatively high chlorine (Cl) peak was detected in
the corrosion scales at the scale/metal interface. It
is, thus, believed the corrosion was caused by
chlorine-accelerated oxidation. Type 310 overlay
tubes were thus selected to replace numerous
T-11 tube bends in this boiler in 2000, 2001, and
2003. So far, no corrosion or other degradation
has been encountered with these overlay tubes.
Figure 13.23 shows the 310 overlay superheater
tubes after 2 years of operation in the boiler,
revealing no evidence of corrosion attack.

13.4 Summary
A brief description of a black liquor recovery
boiler along with its fuel and combustion con-
ditions is presented. Materials problems with
0.5 mm floor tubes, smelt spout openings, and air port
openings in the lower furnace are discussed.
Fig. 13.17 Optical micrograph showing the cross section of
Type 309 overlay at the crown location of the rear The materials that have been tested and tried
waterwall tube after 13 years of boiler operation in a boiler in a with good results are reported. The corrosion
Midwest mill (U.S.). The metallographic mount was etched with
nital to reveal the substrate steel including the fusion boundary issue for the waterwalls above the “cut line” (i.e.,
and the heat-affected zone (HAZ). Source: Ref 45 the butt weld joints where carbon steel waterwall
Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/ 31/10/2007 12:49PM Plate # 0 pg 374

374 / High-Temperature Corrosion and Materials Applications

tubes join with the composite tubes at the lower 4. J. Gommi, Root Causes of Recovery Boiler
part of the furnace) is discussed. In addition, the Leaks, TAPPI Engineering & Papermakers
corrosion of superheater tubes is also discussed. Conference (Conf. Proc.), Book 2, TAPPI,
(1997), p 509
5. W.J. Frederick, Chapter 3, Black Liquor
REFERENCES Properties, Kraft Recovery Boilers, T.N.
Adams, Ed., TAPPI Press, (1997), p 61
1. S.C. Stultz and J.B. Kitto, Ed., Steam and Its 6. T.M. Grace, Chapter 5, Chemical Recovery
Generation and Use, 40th ed., Babcock & Process Chemistry, Chemical Recovery in
Wilcox, (1992), p 26–1 the Alkaline Pulping Processes, 3rd ed., R.P.
2. G.A. Smook, Handbook for Pulp & Paper Green and G. Hough, Ed., TAPPI Press,
Technologists, 2nd ed., Angus Wilde Pub- (1992), p 57
lications, Vancouver, Canada, (1992), p 74 7. W.J. Frederick and M. Hupa, Chapter 5,
3. T.J. Grant, Update of the American Forest Black Liquor Droplet Burning Processes,
& Paper Association’s Recovery Boiler Kraft Recovery Boilers, T.N. Adams, Ed.,
Program, TAPPI Engineering & Paper- TAPPI Press, (1997), p 131
makers Conference (Conf. Proc.), Book 2, 8. G.A. Smook, Handbook for Pulp &
TAPPI, (1997), p 589 Paper Technologists, 2nd ed., Angus Wilde
Publications, Vancouver, Canada, (1992),
p 133
9. T. Grace and W.J. Frederick, Chapter 6, Char
Bed Processes, Kraft Recovery Boilers,
T.N. Adams, Ed., TAPPI Press, (1997), p 163
10. H. Tran, Chapter 10, Recovery Boiler Cor-
rosion, Kraft Recovery Boilers, T.N. Adams,
Ed., TAPPI Press, (1997), p 285
11. E.F. Hogan, Investigation of Chemical
Recovery Unit Floor Tube Overheating
Failures, TAPPI Engineering & Paper-
makers Conference (Conf. Proc.), Book 2,
TAPPI, (1997), p 567
12. C.M. Wells, Chapter VIII, Chemical
Recovery Area, Pulp and Paper Manu-
facturing, Vol 10, Mill-Wide Process Con-
Fig. 13.18 A general view of the 309 overlay on the front
wall after about 8 years of boiler operation in trol & Information Systems, D.B. Brewster
Midwest mill (U.S.). Source: Ref 45 and M.J. Kocurek, Ed., Joint Textbook

5
T-22 probe

4 Windward
Metal loss, mm

2
Leeward

0
440 460 480 500 520 540 560 580 600 620 640 660
Temperature, °C

Fig. 13.19 Results of corrosion probe tests for T-22 in the lower superheater region for 840 h in a boiler. Source: Ref 46
Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/ 31/10/2007 12:49PM Plate # 0 pg 375

Chapter 13: Black Liquor Recovery Boilers in the Pulp and Paper Industry / 375

Fig. 13.20 T-11 superheater tube sample showing severe

corrosion attack after 6 months of operation.
Source: Ref 31

(a) 160 µm
Committee of the Paper Industry, Canada,
(1993), p 124
13. P.M. Singh, S.J. Al-Hassan, S. Stalder, and
G. Fonder, Corrosion in Kraft Recovery
Boilers—In-Situ Characterization of Cor-
rosive Environments, 1999 TAPPI
Engineering/Process and Product Quality
Conference (Conf. Proc.), Vol 3, TAPPI,
(1999), p 1047
14. A. Borg, A. Teder, and B. Warnqvist, TAPPI,
Vol 57 (No. 1), (1974), p 126
15. H. Tran, Chapter 9, Upper Furnace Deposi-
tion and Plugging, Kraft Recovery Boilers,
T.N. Adams, Ed., TAPPI Press, (1997),
p 247
16. H. Tran, M. Gonsko, and X. Mao, Effect
of Composition on the First Melting Tem-
perature of Fireside Deposits in Recovery
Boilers, TAPPI J., Vol 82 (No. 9), (1999), (b) 24 µm
p 93
17. H. Tran, Recovery Boiler Plugging and Fig. 13.21 Scanning electron micrographs showing (a) the
corrosion products formed on the T-11 super-
Prevention, TAPPI Kraft Recovery Opera- heater tube sample at the tube bend (Fig. 13.20) and (b)
tions Short Course Notes, TAPPI Press, the corrosion products on area C at high magnification.
EDX analysis showed essentially iron with tiny chromium
1992, p 209–218 peak. Area C showed presence of chlorine (Cl) in addition to
18. J.L. Clement and J.D. Blue, Recovery Fur- iron. The x-ray spectra from area C are shown in Fig. 13.22.
Source: Ref 31
nace Floor Design and Alternative Mate-
rials, presented at Tenth Latin American
Recovery Congress (Concepcion, Chile),
Aug 26–30, 1996 Tubes, 1997 Engineering & Papermakers
19. D. Singbeil, R. Prescott, J. Keiser, and Conference (TAPPI Conf. Proc.), Book 3,
R. Swindeman, Composite Tube Cracking TAPPI Press, (1997), p 1025
in Kraft Recovery Boilers—A State-Of- 21. J.L. Barna and K.B. Rivers, Improving
The-Art Review, 1997 Engineering & Recovery Boiler Furnace Reliability with
Papermakers Conference (TAPPI Conf. Advanced Materials and Application Meth-
Proc.), Book 3, TAPPI Press, (1997), p 1001 ods, presented at Canadian Pulp and Paper
20. J.R. Keiser et al., Analysis of Cracking Association Meeting, (Montreal, Quebec,
of Co-Extruded Recovery Boiler Floor Canada), Jan 25–29, 1999
Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/ 31/10/2007 12:49PM Plate # 0 pg 376

376 / High-Temperature Corrosion and Materials Applications

8000 Fe

7000

6000

5000
Counts

4000

3000

2000
C1
0
Fe Fe
1000
Cr Si Cr
C S Cr Mn
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
keV

Fig. 13.22 X-ray spectra from area C, as shown in Fig. 13.21, showing a relatively high chlorine (Cl) peak. Source: Ref 31

(a) (b)

Fig. 13.23 Type 310 overlay superheater tubes (a) and in close-up (b) after 2 years of operation in the boiler at a mill in South America.
Source: Ref 31

22. J.R. Keiser, L.M. Hall, K.A. Choudhury, 23. J.R. Keiser et al., Status Report on Studies of
G.B. Sarma, J.P. Gorog, and R.E. Baker, Recovery Boiler Composite Floor Tube
Thermal Behavior of Floor Tubes in a Kraft Cracking, 1999 TAPPI Engineering/Process
Recovery Boiler, 1999 TAPPI Engineering/ and Product Quality Conference (TAPPI
Process and Product Quality Conference Conf. Proc.), TAPPI Press, 1999, p 1099
(TAPPI Conf. Proc.), TAPPI Press, (1999), 24. R. Prescott and D.L. Singbeil, Stress Cor-
p 1109 rosion Cracking of Type 304L Stainless
Name ///sr-nova/Dclabs_wip/High Temp/5208_359-377.pdf/Chap_13/ 31/10/2007 12:49PM Plate # 0 pg 377

Chapter 13: Black Liquor Recovery Boilers in the Pulp and Paper Industry / 377

Steel in Kraft Recovery Boiler Environ- Quality Conference (TAPPI Conf. Proc.),
ments, Ninth International Symposium on TAPPI Press, (1999), p 1071
Corrosion in the Pulp and Paper Industry 35. A. Wensley, Alternative Materials for Floor
(Conf. Proc.), CPPA (Montreal, Quebec, and Lower Waterwall Tubes in Black Liquor
Canada), (1998), p 185 Recovery Boilers, 2000 TAPPI Engineering
25. H. Tran, B. Habibi, and C. Jia, Drying Conference (Conf. Proc.), TAPPI Press,
Behavior of Waterwash Solution and the 2000
Effect on Composite Floor Tube Cracking in 36. J.R. Keiser et al., Recent Observations
Recovery Boilers, 1999 TAPPI Engineering/ of Recovery Boiler Primary Air Port
Process and Product Quality Conference Cracking and Characterization of Environ-
(TAPPI Conf. Proc.), TAPPI Press, (1999), mental Conditions, 2001 TAPPI Engineer-
p 1061 ing/Finishing & Converting Conference
26. J.R. Keiser et al., Why Do Recovery Boiler (Conf. Proc.), TAPPI Press, 2001
Composite Floor Tubes Crack? 2000 TAPPI 37. J.R. Keiser et al., Relationship of Recovery
Engineering Conference (TAPPI Conf. Boiler Parameters and Primary Air Port
Proc.), TAPPI Press, 2000 Cracking, 2002 TAPPI Fall Conference
27. “Sandvik Sanicro 38/4L7,” S-12126-Eng, (Conf. Proc.), TAPPI Press, 2002
Sandvik Steel, Sweden, June 1996 38. A. Wensley and B. Woit, Inspection of
28. X.L. Wang, E.A. Payzant, B. Taljat, Primary Air Port Opening Tubes in Re-
C.R. Hubbard, J.R. Keiser, and M.J. Jirinec, covery Boilers, 2001 TAPPI Engineering/
Experimental Determination of the Re- Finishing & Converting Conference (Conf.
sidual Stresses in a Spiral Weld Overlay Proc.), TAPPI Press, 2001
Tube, Mater. Sci. Eng. Vol A232, (1997), 39. R. Shenassa, K. Haaga, and J. Tuiremo,
p 31 Primary Air Port Tube Integrity—A Critical
29. P.N. Hulsizer, Dual Pass Weld Overlay Review of Primary Air Port Design and
Method and Apparatus, U.S. Patent No. the Effect of Boiler Design Parameters,
6013890, Jan 2000 2002 TAPPI Fall Conference (Conf. Proc.),
30. G. Lai, M. Jirinec, and P. Hulsizer, The TAPPI Press, 2002
Properties and Characteristics of Unifuse 40. A. Wensley, presented at the International
625 Overlay Tubing for Recovery Boiler Symposium on Corrosion in the Pulp and
Applications, 1998 TAPPI Engineering Paper Industry, Charleston, SC, 2004
Conference (Conf. Proc.), Book 2, TAPPI 41. J.R. Keiser, private communication, 2006
Press, 1998, p 417 42. J.R. Keiser et al., Causes and Solutions for
31. G.Y. Lai and A. Wensley, Metallurgical Cracking of Co-extruded and Weld Overlay
Characteristics and Performance Experience Floor Tubes in Black Liquor Recovery
of Spiral Overlay Tubes in Black Liquor Boilers, Ninth International Symposium on
Recovery Boilers, 2005 TAPPI Engineering, Corrosion in the Pulp and Paper Industry
Pulping and Environmental Conference (Conf. Proc.), TAPPI Press, 1998
(Conf. Proc.), TAPPI Press, 2005 43. Welding Services Inc., unpublished data,
32. A. Wilson, M. Lundberg, and U. Forsberg, Norcross, GA
Alloy 825 Mod/SA210-A1 Composite Tube 44. O. Moberg, P.E. Ahlers, and L. Dahl,
for Black Liquor Recovery Boiler Floors, Recovery Boiler Corrosion, Pulp and Paper
1997 Engineering & Papermakers Con- Industry Corrosion Problems, NACE,
ference (TAPPI Conf. Proc.), Book 3, TAPPI (1974), p 125
Press, (1997), p 1043 45. G. Lai and P. Hulsizer, Performance of Type
33. D. Singbeil, Inspection for Cracking of 309 SS Overlay in the Lower Furnace of A
Composite Tubes in Black Liquor Recovery Black Liquor Recovery Boiler, 2001 TAPPI
Boilers, 2002 TAPPI Fall Conference (Conf. Engineering Conference (Conf. Proc.),
Proc.), TAPPI Press, 2002 TAPPI Press, 2001
34. H. Dykstra, N. Risebrough, and A. Wensley, 46. H.N. Tran, D.C. Pryke, and D. Barham,
Corrosion and Cracking of Lower Furnace Local Reducing Atmosphere—A Cause of
Wall Tubes in Recovery Boilers, 1999 Superheater Corrosion in Kraft Recovery
TAPPI Engineering/Process and Product Units, TAPPI, Vol 68 (No. 6), 1985, p 102
Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/ 31/10/2007 4:06PM Plate # 0 pg 379

High-Temperature Corrosion And Materials Applications Copyright © 2007 ASM International®

George Y. Lai, editor, p379-408 All rights reserved.
DOI: 10.1361/hcma2007p379 www.asminternational.org

CHAPTER 14

Stress-Assisted Corrosion and Cracking

14.1 Introduction The other subject area discussed in this chapter

is stress-induced cracking. This is a brittle
Most high-temperature components are under fracture with cracks propagating along grain
stress during service. The stress can be residual, boundaries. This is often referred to as “stress-
resulting from welding or forming operations relaxation cracking,” “reheat cracking,” or
prior to service. When the temperature is not “strain-age cracking” in the literature. Typically,
high enough, these stresses can remain in the a highly constrained component, such as a
component during service. The stress can also heavy wall construction, or a welded component,
come from externally imposed thermal stresses or a cold-worked structure, can be susceptible to
or mechanical loading to which the component is this type of intergranular, brittle cracking for
subjected during service. The stresses imposed some alloys when exposed to lower end of the
upon the component during service can affect the intermediate temperature range, such as 480 to
behavior of high-temperature corrosion attack 700 °C (900 to 1290 °F). Alloys that are sus-
or cause intergranular cracking. ceptible to this type of intergranular fracture
One important subject area is the effect of include ferritic steels, austenitic stainless steels,
tensile stresses (or strains) on high-temperature Fe-Ni-Cr alloys, and nickel-base alloys. Some γ′
corrosion attack. Tensile stresses that are im- [Ni3Al] strengthened nickel-base alloys can
posed on the component during service can cause be susceptible to this type of embrittlement at
preferential corrosion penetration. As this pre- temperatures higher than 700 °C (1290 °F). This
ferential corrosion penetration grows into the type of cracking can also occur during reheating
metal interior, it can eventually develop into a of the component in a postweld heat treatment
crack. One important industrial example in- or annealing/stress-relieving. Weldments or weld
volving this phenomenon is the circumferential joints are particularly susceptible to this type of
cracking that occurs on the waterwall tubes of cracking.
some supercritical coal-fired boilers, which are Some components can be subject to thermal
fired under low NOx combustion conditions. cycling during service. The components that are
The low NOx combustion produces sulfidizing subject to thermal cycling are generally more
environments at or near the waterwall of the easily understood from the operating standpoint.
boiler. The waterwall tubes, which are made of When both stress magnitudes (or strain magni-
a carbon or low-alloy steel, or even with a stain- tudes) and frequencies are high enough, the metal
less steel or nickel-base alloy weld overlay that can eventually fail by thermal fatigue cracking.
protects the waterwall steel tubes, can suffer The morphology of thermal fatigue cracking is
circumferential cracking in some boilers. This generally very well defined and easily identified.
particular phenomenon (i.e., circumferential The subject of thermal fatigue cracking has been
cracking of the waterwall) is covered in Chapter extensively discussed in the literature and is not
10 “Coal-Fired Boilers.’’ In that chapter, the covered in this chapter.
correlation between the circumferential cracking
and the stress-assisted preferential corrosion
penetration is discussed. The current chapter
discusses the phenomenon of the preferential 14.2 Stress-Assisted Preferential
high-temperature corrosion penetration that Corrosion Penetration
forms under the combination of stresses and cor-
rosive conditions, primarily in sulfidizing envir- In most industrial environments, alloys rely
onments. on the formation of a continuous oxide scale
Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/ 31/10/2007 4:07PM Plate # 0 pg 380

380 / High-Temperature Corrosion and Materials Applications

for protection against accelerated corrosion penetration during the exposure for times up
attack during service at elevated temperatures. to 1000 h. The internal oxidation penetration
Most commercial alloys rely on chromium oxide in the deforming Fe-18Cr-0.8Al alloy specimen
(Cr2O3) scales. The majority of the corrosion was found to be essentially in the form of nit-
data have been generated from test specimens ridation attack involving aluminum nitrides
that were not subjected to external stresses (Ref 1). (Nitridation attack in air or combustion
during exposure. However, as discussed earlier, environments is discussed in Chapter 4 “Nitri-
most high-temperature components are subject dation.”)
to external stresses during plant operation. Oxidation in air or oxidation in general is
When a metal with a protective oxide scale considered to be the least corrosive among
is under tensile stress, the oxide scale may crack various high-temperature corrosion modes. As
as the tensile stresses or strains become high the corrosivity of the environment increases, the
enough. Figure 14.1 shows cracking of the effect of the external tensile stresses on prefer-
oxide scale on alloy 800 after testing at 800 °C ential corrosion penetration becomes increas-
(1470 °F) in air under the strain rate of 10−6 s−1 ingly more significant. This section focuses on
(Ref 1). Under certain conditions, these cracks the stress-induced preferential corrosion pene-
can be rehealed. This is illustrated in Fig. 14.2, tration in sulfidizing environments.
showing a healed oxide crack on alloy 800 at Smolik and Flinn (Ref 3) examined the effects
800 °C (1470 °F) in air under the strain rate of stresses on sulfidation of alloy 800H in coal
of 10−8 s−1 (Ref 2). Schütze (Ref 2) observed gasification environments, which were charac-
that cracking of the oxide scale on a deforming terized with high sulfur and low oxygen poten-
metal can cause significant internal corrosion tials ( pS2 and pO2 ). Tests involved exposing the
penetration, as shown in Fig. 14.3 for Fe-18Cr outer diameter of the alloy 800H tube to three
steel (with 0.8Al and 1.5Si) deforming at 800 °C test environments, which were Ar-23%H2O-
(1470 °F) in air under a strain rate of 10−8 s−1. 10%H2 with 0.1, 0.2, and 0.4% H2S, respec-
The figure shows that once the oxide scale tively. The corresponding partial pressures of
reached the initial “threshold” strains for devel- oxygen ( pO2 ) and sulfur ( pS2 ) at the test tem-
oping cracking, further deforming of the steel perature of 870 °C (1600 °F) were 2 ×10−18 and
caused significant internal oxidation penetration. 1 × 10−8 atm for the 0.1 H2S test environ-
On the other hand, the nondeformed specimen ment, 2× 10−18 and 5 × 10−8 atm for the 0.2 H2S
(i.e., the specimen that was not under creep environment, and 2 ×10−18 and 2 × 10−7 atm for
deformation during the exposure) showed no the 0.4 H2S environment. The circumferential
changes in the depth of the internal oxidation strain was created by pressurizing the test
tube internally with argon. Preferential corrosion
penetration was observed to take place along

Fig. 14.1 Cracking of the oxide scale on alloy 800 during Fig. 14.2 Rehealed oxide-scale crack on alloy 800H during
deformation at the strain rate of 10−6 s−1 at 800 °C deformation at the strain rate of 10−8 s−1 at 800 °C
(1470 °F) in air. Source: Ref 1 (1470 °F) in air. Source: Ref 2
Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/ 31/10/2007 4:07PM Plate # 0 pg 381

Chapter 14: Stress-Assisted Corrosion and Cracking / 381

grain boundaries. Both oxides and sulfides were coal gasification environments, Guttmann and
found to form along grain boundaries. This inter- Timm (Ref 4) observed that creep deformation
granular oxidation/sulfidation penetration even- markedly accelerated preferential corrosion
tually affected the circumferential strains to penetration. Creep tests were conducted at
failure for the test tubes. The higher H2S con- 800 °C (1470 °F) in H2-1.2H2O-7CO-0.4H2S
centration produced deeper penetrations, thus (5 ×10−22 atm pO2 , 5×10−9 atm pS2 , and 0.3 ac)
resulting in smaller strains to failure. The cir- and H2-1.2H2O-0.4H2S (5×10−22 atm pO2 ,
cumferential strains to failure were found to be 4 ×10−9 atm pS2 ). Exposure tests were also con-
about 8 to 10% for the 0.1% H2S, about 6 to 7% ducted in the same test environments with
for the 0.2% H2S, and 1 to 4% for the 0.4% H2S. no applied stresses for comparison of the corro-
Researchers at the Joint Research Centre, sion behavior. Under the stressed condition pre-
Petten Establishment, The Netherlands, have ferential corrosion penetration was found to be
conducted a series of extensive studies on the deepest at grain boundaries that were normal to
effects of deformation on preferential corro- the stress direction. Samples in the stress-free
sion in reducing, sulfidizing environments (i.e., condition showed shallow corrosion depths
simulated coal gasification environments) (Ref following grain or twin boundaries without pre-
4–8). Some of their results are summarized in this ferential penetration. Preferential intergranular
section. corrosion penetration was found to consist of
In their study of the effect of the creep de- essentially sulfidation/oxidation. It was found
formation on the corrosion of alloy 800H in that the sulfidation/oxidation penetration in the

ε = 10–8s–1

Without deformation
150

Scale
cracking
ε = 0.7%

100
x i, µm

50

0
0 500 1000
Time, h

Fig. 14.3 Depth of internal corrosion penetration for Fe-18Cr-0.8Al-1.5Si between the undeformed specimens (open data points) and
the deforming specimens (solid data points) under the strain rate of 10−8 s−1 at 800 °C (1470 °F) in air. Source: Ref 2
Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/ 31/10/2007 4:07PM Plate # 0 pg 382

382 / High-Temperature Corrosion and Materials Applications

stressed specimens was about 2 to 3 times more structural components during service at high
than that of the unstressed specimens. This temperatures, the strains imposed on the com-
is illustrated in Fig. 14.4 (Ref 4). For the CO- ponent are likely to be much smaller. The design
containing environment, specimens in both strain rate (or creep rate) of ASME Boiler and
stressed and unstressed conditions suffered Pressure Vessel Codes is 10−5% h−1 (2.8 ×
carburization as well as sulfidation/oxidation. 10−11 s−1) (Ref 9). It is therefore important to
Under these conditions, carburization dominated examine the effects of low tensile stress (or
the corrosion penetration depth; the authors strain) on the preferential corrosion penetration.
found no significant effect on the corrosion depth Stroosnijder et al. (Ref 6) investigated the
when the corrosion attack involved carburization effect of strain on preferential corrosion of
(Fig. 14.4). In examination of the surface cracks alloy 800H when tested at 800 °C (1470 °F)
of the specimens tested in air and those tested in H2-1.2H2O-7CO-0.4H2S (5 × 10−22 atm pO2 ,
in the sulfidizing environments, Guttmann and 5 × 10−9 atm pS2 , and 0.3 carbon activity, ac)
Timm (Ref 4) found that the surface cracks, with (a) no external stress (or strain) during
which became blunted in air, were deeply pene- testing and (b) strains up to 4% and strain
trating and long and sharp with severe corrosion rates down to 10−9 s−1. In this study, alloy 800H
at the crack tip and in the crack vicinity. Stroos- specimens were surface treated with CeO2 using
nijder et al. (Ref 5) made similar observations cerium sol-gel techniques. The authors observed
when comparing the creep deformation behavior that the unstressed specimens showed some
of alloy 800 at 700 °C (1290 °F) between air external and internal corrosion with no evidence
and H2-1.2H2O-7CO-0.2H2S (2 ×10−25 atm pO2 , of preferential corrosion penetration after 663 h
2 ×10−10 atm pS2 ). These authors observed of exposure (Fig. 14.5a). For stressed specimens
blunted surface cracks when tested in air and after 663 h and subjected to about 2.3% strain
deeply penetrating sharp cracks along grain under a strain rate of 10−8 s−1, preferential cor-
boundaries with deep corrosion paths in front of rosion penetration along the grain boundary
cracks when tested in the sulfidizing environ- began to take place (Fig. 14.5b). After testing
ment. The corrosion products formed at the grain for 1247 h with about 5% strain (10−8 s−1 strain
boundary in the sulfidizing environment were rate), intergranular corrosion attack along grain
found to consist of oxides and sulfides. boundaries was found to penetrate farther into
The above discussion focuses on the effects the metal interior (Fig. 14.5c).
of relatively large deformation (i.e., large strain Guttmann et al. (Ref 7) examined the effect
or high strain rate) during creep testing. For many of a much lower strain range (typically 1 to 2%)
on the preferential corrosion penetration in a
sulfidizing environment. The strains the authors
examined were much closer to the strains that
might be experienced by operating high-
103
temperature components. Tests were conducted
at 600 °C (1110 °F) in CO-32H2-4CO2-0.2H2S.
Depth of corrosion, µm

The strain rates during creep were 10−10 s−1 to

10−8 s−1. Similar to what was observed earlier
in alloy 800H (Fe-33Ni-2Cr-0.4Al-0.4Ti)
102 (Ref 4–6), Guttmann et al. (Ref 7) observed
that alloy HR3C (Fe-20Ni-25Cr-0.5Nb-0.2N)
No load Load
showed stress-assisted intergranular corrosion
Sulfidation
Oxidation attack under the test conditions. This is illustrated
Carburization in Fig. 14.6 (Ref 7). The corrosion products
10 formed at grain boundaries were enriched in
10 102 103 104 chromium, oxygen, and sulfur, but depleted in
Exposure time, h nickel and iron. The intergranular corrosion
attack in alloy HR3C tested after 1810 h with
Fig. 14.4 Corrosion depth for unstressed and stressed speci-
2.2% strain is shown in Fig. 14.7 for the scanning
mens of alloy 800H tested at 800 °C (1470 °F)
in H2-1.2H2O-0.4H2S (5 × 10−22 atm pO2 , 4 × 10−9 atm pS2). Also electron microscopy (SEM) backscattered elec-
included was the data generated from H2-1.2H2O-7CO-0.4H2S tron image of the intergranular corrosion attack
(5 × 10−22 atm pO2 , 5 × 10−9 atm pS2 , and 0.3 ac). The latter
environment caused carburization in addition to sulfidation/ and the x-ray maps for chromium, oxygen, and
oxidation. Source: Ref 4 sulfur. The intergranular corrosion attack was
Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/ 31/10/2007 4:07PM Plate # 0 pg 383

Chapter 14: Stress-Assisted Corrosion and Cracking / 383

(a) (b) (c)

Fig. 14.5 Corrosion morphology of alloy 800H tested at 800 °C (1470 °F) in H2-1.2H2O-7CO-0.4H2S (5 × 10−22 atm pO2 , 5 × 10−9 atm
pS2 , and 0.3 ac) (a) after 663 h under no external stresses, (b) after 663 h at about 2.3% strain under the strain rate of 10−8 s−1,
and (c) after 1247 h under strain rate of 10−8 s−1. Note: the specimen surface was plated prior to the mounting of the sample to retain the
corrosion products. Source: Ref 6

(a) (b)
25 µm 25 µm

Fig. 14.6 Strain-assisted intergranular corrosion attack in alloy HR3C after testing at 600 °C (1110 °F) for 250 h in CO-32H2-4CO2-
0.2H2S with (a) 1.3% strain and (b) 2% strain. Corrosion products formed on the metal surface were also observed. Note:
the tested specimen surface was plated prior to the mounting of the metallographic sample to retain the surface corrosion products. Source:
Ref 7

oxidation/sulfidation, similar to the earlier completely different morphology in preferential

observations in alloy 800H. For nonstressed corrosion penetration under stress. In fact, the
specimens, the alloy suffered external sulfidation morphology of stress-assisted preferential cor-
with Fe- or Ni-rich sulfides ([Ni,Fe]9S8) and rosion attack for nickel-base alloys tends to be
Ni3S2. Underneath the surface sulfide scales similar to that for iron-base alloys having a
were fine internal corrosion phases randomly ferritic structure. Guttmann et al. (Ref 7) tested
distributed within the grain matrix with no deep nickel-base alloy 45TM (Ni-27Cr-23Fe-2.76Si)
penetrating intergranular corrosion attack, as under the same test condition as for alloy HR3C.
shown in Fig. 14.8. Alloy 45TM exhibited a fingerlike preferential
Both alloys 800H and HR3C are Fe-Ni-Cr corrosion penetration, as shown in Fig. 14.9
alloys with austenitic microstructure. Both alloys (Ref 7). The corrosion products formed in the
were found to suffer preferential corrosion preferential corrosion penetration were found to
penetration along grain boundaries. It is inter- be enriched in chromium, oxygen, sulfur, and
esting to note that nickel-base alloys with the silicon, as shown in Fig. 14.10. By contrast, the
same austenitic microstructure would follow a unstressed specimens were found to exhibit a
Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/ 31/10/2007 4:07PM Plate # 0 pg 384

384 / High-Temperature Corrosion and Materials Applications

Fig. 14.7 Scanning electron backscattered images showing intergranular corrosion penetration along with the x-ray maps for Cr, O,
and S for alloy HR3C after testing at 600 °C (1110 °F) for 1810 h in CO-32H2-4CO2-0.2H2S with 2.2% strain. Source: Ref 7

Fig. 14.8 Corrosion products formed on the metal surface as

well as internal corrosion products formed in alloy
HR3C with no external strain imposed on the test specimen when
tested at 600 °C (1110 °F) for 2100 h in CO-32H2-4CO2-0.2H2S.
Note: the tested specimen surface was plated prior to the
mounting of the metallographic sample to retain the surface cor- Fig. 14.9 Stress-assisted preferential corrosion penetration
for alloy 45TM tested to 2% strain at 600 °C
rosion products. Source: Ref 7
(1110 °F) for 1820 h in CO-32H2-4CO2-0.2H2S. Note: the tested
specimen surface was plated prior to the mounting of the metal-
lographic sample to retain the surface corrosion products. Source:
fairly uniform surface corrosion attack with no Ref 7
preferential corrosion penetration, as shown in
Fig. 14.11. (ODS) alloy, MA956 (Fe-20Cr-4.6Al-0.5Y2O3),
In the same study, the authors tested an under the same condition and found that this
iron-base ferritic oxide-dispersion-strengthened iron-base alloy with ferritic structure exhibited
Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/ 31/10/2007 4:07PM Plate # 0 pg 385

Chapter 14: Stress-Assisted Corrosion and Cracking / 385

a fingerlike preferential corrosion penetration condition but with no external stresses, alloy
under stresses, as shown in Fig. 14.12. The result MA956 showed only uniform corrosion attack,
of x-ray elemental mapping (Fig. 14.12) reveals as shown in Fig. 14.13. For the unstressed
that (a) the external surface layer was iron- specimen, the external corrosion layer was found
rich sulfides, (b) the inner surface layer was to be iron-rich sulfides and inner layer consisted
chromium- and aluminum-rich oxides along of (Fe,Cr)1−xS, Cr2O3, and Al2O3. Ferritic
with some sulfides, and (c) the inner surface stainless steels appeared to follow the similar
layer (Cr-Al rich oxides) also extended to the fingerlike preferential corrosion penetration
preferential corrosion penetration as an outer when the specimen was under creep deformation,
layer with a center core of iron-rich sulfides that as shown in Fig. 14.14 (Ref 8). This figure also
were extended from the external surface layer shows that the specimen suffered essentially
discussed in (a). When tested under a similar uniform attack when under no external stress.

Fig. 14.10 Scanning electron micrograph for the stress-assisted preferential corrosion penetration in alloy 45TM tested to 4%
strain at 600 °C (1110 °F) for 2000 h in CO-32H2-4CO2-0.2H2S and elemental x-ray maps for the corrosion products
in chromium, oxygen, sulfur, silicon, and iron. Source: Ref 7
Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/ 31/10/2007 4:07PM Plate # 0 pg 386

386 / High-Temperature Corrosion and Materials Applications

The authors did not report the percent of strain The waterwalls in some coal-fired boilers,
when the test was terminated and the specimen particularly some supercritical units, have been
was examined; however, the test duration was found to encounter preferential corrosion pene-
only 615 h, and the strain was likely to be high. tration. These supercritical boilers are typically
Upon closer examination of Fig. 14.14, it appears equipped with low NOx burners and overfire
that a surface crack formed in the center of each air, thus creating localized reducing, sulfidizing
corrosion penetration, further indicating that the environments at or near the waterwall in the
strain might be large. lower furnace. Furthermore, the waterwalls can
Preferential corrosion penetration under also be subjected to high heat flux and possibly
strain (or stress) was also observed in other flame impingement, thus resulting in higher
aggressive environments. The stress-assisted metal temperatures and higher thermal stresses.
preferential corrosion attack is quite sensitive to The combination of these conditions leads to
the environment. Le Calvar et al. (Ref 10) the development of stress-assisted preferential
performed low strain rate tests in air + 4CO2 + sulfidation penetration. As the preferential sul-
8H2O at 610 °C (1130 °F). Figure 14.15 shows fidation penetration continues to grow into
Type 304H suffering preferential oxidation the metal interior, it eventually develops into a
penetration under this test condition at a strain crack. At later stages, these cracks resemble
rate of 3×10−8/s with total strain of about 2%. thermal fatigue cracks. These cracks, which
The authors referred to this preferential oxida- are in the transverse direction with respect to
tion attack as “cracking” in their paper (Ref 10). the tube axis, are commonly referred to as
Similar tests were also performed in a vacuum “circumferential cracks.” Typical alloys of
environment, and the results showed construction have experienced circumferential
no “cracking.” The figure suggests that the cracking. The commonly used weld overlay
phenomenon Le Calvar et al. observed was cladding alloys including Type 309 and alloy 625
more like stress-assisted preferential corrosion have also been observed to suffer circumferential
penetration than stress-assisted cracking fol- cracking. This phenomenon is covered in detail
lowed by formation of the corrosion products in Chapter 10 “Coal-Fired Boilers” (Section
inside the crack. 10.5.3).
Rorbo (Ref 11) reported that the external The morphologies of the preferential sulfida-
stresses in Type 304 caused cracks to develop tion penetration are strikingly similar between
in the nitrided layer, and the cracks resulted in what has been observed in the waterwall of coal-
increased nitridation attack in front of the crack. fired boilers and that observed in test specimens
This is illustrated in Fig. 14.16. The nitrided under tensile stresses (or low strain rate creep
layer is believed to be very brittle, and very little tests) in laboratory sulfidizing environments.
strain is required to develop cracking in the Figure 10.32 in Chapter 10 shows an example
nitrided layer. In this case, cracking could of preferential sulfidation penetrations observed
have developed in the nitrided layer and caused on T-22 (2.25Cr-1Mo) in a supercritical coal-
preferential nitridation penetration in front of fired boiler equipped with low NOx burners
the crack. and overfire air. These preferential sulfidation
penetrations are considered to be precursors to
circumferential cracks observed from tubes that
failed or were about to fail. The preferential
sulfidation penetration often exhibits “channels”
in its core, as shown in Fig. 14.17. These
channels, which generally exhibited lighter color
(Fig. 14.17), were found to consist of sulfides
(see Fig. 10.33 and 10.34 in Chapter 10). In some
cases, these channels were much more pro-
nounced and numerous, such as the one shown
in Fig. 14.18.
The preferential sulfidation penetration
formed in steels was found to consist of essen-
tially iron oxides in the outer region of the
Fig. 14.11 Corrosion morphology for alloy 45TM tested
penetration with the inner region (core) includ-
under no external strains at 600 °C (1110 °F)
for 2000 h in CO-32H2-4CO2-0.2H2S. Source: Ref 7 ing channels being essentially iron sulfides
Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/ 31/10/2007 4:07PM Plate # 0 pg 387

Chapter 14: Stress-Assisted Corrosion and Cracking / 387

(see the SEM/EDX analysis in Fig. 10.33 and 2. Tensile stresses cause the breakdown of the
10.34 in Chapter 10). This is illustrated sche- iron oxide scales.
matically in Fig. 14.19. 3. The surface oxide scale breakdown initiates
Proposed reaction steps involved in develop- the development of a preferential oxidation
ing preferential sulfidation penetration in carbon penetration.
or low-alloy steel waterwall tubes are: 4. As the preferential oxidation penetration
continues to grow inward into the metal
1. Iron oxide scales initially form on the steel interior, the inner region of the oxide pene-
surface. tration (i.e., penetration core) becomes more

Fig. 14.12 Scanning electron micrograph in a backscattered electron image along with x-ray maps for Cr, O, S, Al and Fe showing
the stress-assisted preferential corrosion penetration (fingerlike penetration attack) on alloy MA956 tested to 5% strain at
600 °C for 1830 h in CO-32H2-4CO2-0.2H2S. Source: Ref 7
Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/ 31/10/2007 4:07PM Plate # 0 pg 388

388 / High-Temperature Corrosion and Materials Applications

deficient in oxygen, thus lowering oxygen sulfidation penetration grows deeper into the
potentials in the core region. metal interior.
5. Iron sulfides begin to form in the core region
due to insufficient oxygen. For a Ni-Cr-Fe alloy, the preferential sulfida-
6. As the penetration continues to grow farther tion penetration under large strains could develop
into the metal interior, oxidation/sulfidation a large core of iron sulfides in the inner region
processes continue to penetrate inward into with chromium oxides formed in the region
the metal interior with oxidation forming in next to the unaffected metal, such as in the case
the outer region and sulfidation in the inner for alloy 45TM tested to 4% strain at 600 °C
region of the corrosion penetration. (1110 °F) for 2000 h in CO-32H2-4CO2-0.2H2S,
7. Channels are likely created during the growth as shown in Fig. 14.10 (Ref 7). Because of a
of sulfide phases in the core region under large tensile strain (4%), the entire center core
the tensile stresses. These channels may became a one big channel. However, when the
also serve as gas passages as the oxidation/ strain was reduced to about 2%, alloy 45TM
showed one big channel through the external
surface scale followed by a number of fine
channels in the preferential sulfidation pene-
tration, as shown in Fig. 14.9. For MA956

Fig. 14.13 Alloy MA956 tested under no external strain

(or stress) at 600 °C (1110 °F) for 2000 h in
CO-32H2-4CO2-0.2H2S, showing uniform corrosion with no
preferential corrosion penetration. Source: Ref 7

10 µm

Fig. 14.14 Fe-12Cr-3Al-3Ti showing preferential corrosion Fig. 14.15 Type 304H tested in air + 4CO2 + 8H2O at 610 °C
penetration for the specimen under creep defor- (1130 °F) under the strain rate of 3 × 10−8/s with
mation at 600 °C (1110 °F) in H2-34.3H2O-18.5CO2-3.8CH4- about 2% strain, showing preferential oxidation penetration
7.9CO-1.3H2S for 615 h (upper figure) and uniform corrosion (the authors referred to as “cracking”). SEM/EDX analysis showed
attack for the specimen under no external stress (or strain) the oxide in area A was Fe-3Cr-4Si-0.5Mo, in area B was
after exposure to the same test environment and duration (lower Fe-29Cr-5Ni-0.8Mo-0.8Si, and in area C was Fe-37Cr-21Ni-
figure). Source: Ref 8 1.5Si. Source: Ref 10
Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/ 31/10/2007 4:07PM Plate # 0 pg 389

Chapter 14: Stress-Assisted Corrosion and Cracking / 389

(a Fe-Cr-Al alloy), the entire core became a one

big channel consisting of iron sulfides with the
outer region consisting of a mixture of chromium
oxides and aluminum oxides at 5% strain, as
shown in Fig. 14.12 (Ref 7). For operating
components, such as waterwalls in boilers, the
strains imposed onto the component during
operation are likely to be much lower. As shown
in Fig. 14.20, alloy 625 weld overlay on the
waterwall of a supercritical unit suffered prefer-
ential sulfidation penetration, which exhibited
fine channels. The figure also shows no evidence
of the formation of a crack near or at the tip
Fig. 14.16 Cracks developed in the nitrided layer formed in
Type 304 due to external stresses and accelerated
nitridation attack in front of the crack. Source: Ref 11 Iron sulfides (higher sulfur)
Iron oxides

Iron sulfides

Iron oxides

Carbon or low-alloy steel

Fig. 14.19 Preferential sulfidation penetration formed in

waterwall steel tubes
Fig. 14.17 Preferential sulfidation penetration (a precursor of
the circumferential cracking) observed on a T-22
(2.25Cr-1Mo) wingwall tube of a supercritical boiler, showing
“channels” (light color stringers) in the core of the penetration.
Courtesy of Welding Services Inc.

Fig. 14.18 Optical micrograph showing numerous “chan- Fig. 14.20 Scanning electron (backscattered electron) image
nels” in a circumferential groove formed on a showing a circumferential groove formed in alloy
T-11 (1.25Cr-0.5Mo) tube in a supercritical coal-fired boiler. 625 weld overlay on a waterwall tube of a supercritical coal-fired
Courtesy of Welding Services Inc. boiler. Courtesy of Welding Services Inc.
Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/ 31/10/2007 4:07PM Plate # 0 pg 390

390 / High-Temperature Corrosion and Materials Applications

of the preferential sulfidation penetration. The penetration attack along dendritic boundaries,
corrosion products were essentially chromium as shown in Fig. 10.40 in Chapter 10. This is
sulfides with nickel-rich sulfides forming the quite similar to the wrought alloys, such as
channels. The results of the SEM/EDX analysis HR3C (Fe-Ni-Cr alloy), that suffered preferen-
of the corrosion products are shown in Fig. 10.37 tial sulfidation penetrations along grain bound-
in Chapter 10. Figure 14.21 shows another ex- aries when tested in laboratory environments
ample of a circumferential groove that formed (Fig. 14.6).
in the alloy 625 weld overlay on the waterwall
tube after 6 years of service in another super-
critical coal-fired boiler. The circumferential 14.3 Stress-Assisted Intergranular
groove was found to contain essentially nickel- Cracking
and chromium-rich sulfides with numerous light
grayish channels. Some light grayish channels When external stress is applied to a metallic
were nickel-rich sulfides. Sulfides in the groove component, the metal first undergoes elastic
were also found to contain lead (Pb), arsenic deformation. With increasing applied stress, the
(As), and zinc (Zn), which are common im- metal then deforms plastically and eventually
purities in coal ash. Fine channels were also reaches the limit of its plastic deformation. At
observed in the preferential sulfidation penetra- this point, with further increase of the applied
tions formed in the 309SS weld overlay on the stress, the metal can no longer deform to relieve
waterwall, as shown in Fig. 10.40 and 10.41 in the stresses built up in the metal, thus resulting
Chapter 10. In the waterwall, the 309SS overlay in the initiation of cracks at the weakest locations
(Fe-Cr-Ni alloy) suffered preferential sulfidation to relieve those stresses. An engineering alloy
can normally undergo a large plastic deformation
before developing cracks. When the metallic
component is subject to external stresses at
elevated temperatures, the metal undergoes creep
deformation, which is a time-dependent defor-
mation governed by diffusion. An engineering
alloy typically exhibits a large creep deformation
prior to developing cracks and final rupture.
However, as the temperature decreases, the creep
strain at rupture decreases.
Under certain conditions, a metal can only
sustain very little deformation (plastic and/or
creep deformation) before developing brittle
fracture. The fracture typically takes the form
of intergranular cracking (i.e., cracking along
grain boundaries). This brittle, intergranular
fracture has occurred during heat treatments or
service for some engineering alloys, including
ferritic steels, austenitic stainless steels, Fe-Ni-Cr
80 µm alloys, and nickel-base alloys. This type of
brittle, intergranular cracking phenomenon has
Fig. 14.21 Scanning electron (backscattered electron) image
been described by different names, such as
showing a circumferential groove formed in alloy
625 weld overlay on a waterwall tube after 6 years of service in a “reheat cracking,” “stress-relaxation cracking,”
supercritical coal-fired boiler under low NOx combustion. The
compositions (wt%) of the phases at different locations were “strain-age cracking,” and “gamma-prime em-
analyzed by EDX, showing either nickel-rich or chromium-rich brittlement.” This phenomenon typically occurs
sulfides. Some sulfides were also found to contain Pb, As, and Zn at the lower end of the intermediate temperature
that are commonly found in coal ash. Note numerous light grayish
channels. Some of these channels (No. 3 and 4) were nickel-rich range, such as 480 to 700 °C (900 to 1290 °F).
sulfides. Courtesy of Welding Services Inc.
1: 56% Ni, 13.9% Cr, 6.2% Fe, 18.7% S, 2.5% Zn, 1.4% As
2: 53.8% Cr, 10.2% Ni, 7.0% Fe, 14.6% S, 9.4% Pb, 2.6% Zn 14.3.1 General Conditions That Cause
3: 58.9% Ni, 5.8% Cr, 4.6% Fe, 26.1% S, 1.9% As, 1.0% Pb
4: 56.4% Ni, 5.1% Fe, 3.3% Cr, 6.6% Mo, 24.4% S, 1.9% Pb Stress-Assisted Cracking Embrittlement
5: 52.5% Cr, 18.4% Mo, 7.5% Fe, 3.8% Ni, 9.5% Mo, 5.5% S
6: 18.7% Ni, 17.8% Cr, 13.7% Nb, 13.7% Mo, 21.3% Pb, Several material conditions can increase the
22.1% S, 4.6% Fe susceptibility of the stress-assisted cracking
Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/ 31/10/2007 4:07PM Plate # 0 pg 391

Chapter 14: Stress-Assisted Corrosion and Cracking / 391

embrittlement for an alloy. An alloy in a cold- Another factor that is important in increas-
worked condition, which exhibits higher yield ing the susceptibility of stress-assisted cracking
strength and higher hardness, and lower ductility, embrittlement is the surface geometry of the
is much more susceptible to this type of brittle component, which can lead to stress risers such
intergranular cracking. Figure 14.22 shows an as an abrupt thickness change in a component. A
example of the effect of cold work on hardness component that is under a highly constrained
and yield strength of several nickel-base alloys condition, such as a heavy section part, is most
and the corresponding effect on its tensile susceptible to this type of cracking embrittlement
elongation (Ref 12). The hardness and yield since the additional stresses imposed on the
strength of an alloy can be significantly in- component render it incapable of accomodating
creased by increasing the amount of cold work. any dimensional change or deformation, and it
The yield strength of a cold-worked metal also develops cracks to relieve those stresses.
represents the residual stress in the cold-worked A majority of cracking embrittlement cases
structure. An alloy with a larger amount of cold occur under two different conditions. The crack-
work exhibits a higher residual stress. Higher ing can occur during heat treatment, such as
residual stresses in a structural component can a postweld heat treatment (PWHT) for a
increase the susceptibility to stress-relaxation weldment, which involves a short duration. The
cracking. After the alloy is strengthened by cracking can also occur during service, which
cold working, the capability of the grain matrix involves a relatively longer duration, but still
to plastically deform to relieve additional exter- relatively short in terms of the design life of the
nal stresses imposed on the metal during service component, with most failures occurring after
at intermediate temperatures is significantly 1 to 2 years of service or less. In both cases, the
reduced. When these external stresses cannot metal (or the metallic component) has undergone
be relieved through deformation, cracking then very little deformation prior to intergranular
develops at grain boundaries to relieve those cracking fracture.
stresses. An example of heat-treatment-induced crack-
The matrix of an alloy can also be strength- ing embrittlement is described as follows. When
ened by precipitates, particularly fine, coherent welding is performed on a metal, residual
precipitates (e.g., γ′ [Ni3(Al,Ti)], γ″ [Ni3Nb], stresses are developed in the heat-affected zone
Ni2Mo ordered phases). Figure 14.23 shows (HAZ) of the base metal due to the volume
fine, coherent γ″ [Ni3Nb] precipitates formed shrinkage of adjacent weld metal changing from
in the alloy matrix (i.e., grain interior) of alloy the molten state to the solid state. The maximum
625 (Ni-22Cr-9Mo-3.5Nb) at 650 °C (1200 °F) residual stress is approximately the room-
for 24 h (Ref 13). Examples of fine precipitates temperature yield strength of the metal. When
formed in other alloy systems are shown in this weldment receives a PWHT, the yield
Fig. 14.24 for Ni2(Cr,Mo) ordered phases formed strength of the metal at this particular PWHT
in Ni-16Cr-15Mo-3Fe alloy, Fig. 14.25 for γ′ temperature would be lower than the residual
(Ni3Al) precipitates formed in alloy 214, and stresses (i.e., the room-temperature yield strength
Fig. 14.26 for γ′ (Ni3Al) precipitates in alloy 601. of the metal), thus causing the metal to under-
Precipitation of these fine, coherent precipitates go deformation to accommodate this additional
not only strengthens the grain matrix but also stress. However, when the metal cannot deform
causes volume contraction that causes additional to “relax” or “relieve” this additional stress
internal stresses in the alloy, making it more through deformation, cracks develop at the
susceptible to stress-relaxation cracking. Car- weakest locations, typically grain boundaries,
bides, such as TiC, NbC, and Cr23C6, also to relieve this additional stress, thus resulting
produce some alloy strengthening and can in this intergranular cracking embrittlement.
increase the susceptibility to stress-relaxation The service-related cracking embrittlement
cracking. The combination of a cold-worked typically occurs when the service temperatures
structure and fine precipitates in the matrix can are relatively low such that creep deformation
cause the material to be extremely susceptible to is too low to relieve the applied stresses, and the
stress-relaxation cracking. Formation of preci- alloy matrix is significantly strengthened such
pitates, such as Cr23C6, at grain boundaries can that these stresses are not able to be “relieved”
further promote grain-boundary cracking, thus through deformation, thus resulting in cracking
increasing the susceptibility of stress-relaxation for the relief of this applied stress. Detailed dis-
cracking. cussion on stress-relaxation cracking is presented
Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/ 31/10/2007 4:07PM Plate # 0 pg 392

392 / High-Temperature Corrosion and Materials Applications

0 10 20 30 40 50 60
50 50

40 40
Hardness, HRC

30 30

20 20
Alloy 25
Alloy 188
Alloy 625
10 10
230 Alloy
Alloy X

0 0
0 10 20 30 40 50 60
Cold work, %
(a)

0 10 20 30 40 50 60

1350 200

175
0.2 % yield strength, MPa

0.2 % yield strength, ksi

1100
150

125
850

100
Alloy 25
Alloy 188
600 Alloy 625
2230 Alloy 75
Alloy X

350 50
0 10 20 30 40 50 60
Cold work, %
(b)

0 10 20 30 40 50 60
70 70

60 60

50 50
Elongation, %

40 Alloy 25 40
Alloy 188
Alloy 625
230 Alloy 30
30
Alloy X

20 20

10 10

0 0
0 10 20 30 40 50 60
Cold work, %
(c)

Fig. 14.22 Effects of cold work on hardness (a) room-temperature yield strength (b) and the corresponding tensile elongation (c) for
several nickel-base alloys. Source: Ref 12
Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/ 31/10/2007 4:07PM Plate # 0 pg 393

Chapter 14: Stress-Assisted Corrosion and Cracking / 393

in the following sections based on the alloy referred to as “underclad cracking,” were en-
systems. countered in nuclear pressure vessels in the
1970s (Ref 17). Other publications discussing
reheat and underclad cracking of ferritic steels
14.3.2 Cracking in Ferritic Steels
in the 1970s can be found in Ref 18 to 21.
Cracking of CrMo and CrMoV steel weld- Dhooge and Vinckier (Ref 17) provided a
ments in steam pipework and valve assemblies detailed review of the factors that might be
during stress-relief heat treatment (e.g., postweld responsible for causing reheat cracking of
heat treatment) and high-temperature service ferritic steels. These factors included intra-
have been encountered in power-generating granular precipitation hardening that strengthens
equipment (Ref 17). Similar reheat cracking has the grain matrix and segregation of impurities to
also been reported in the heat-affected zone grain boundaries that weakens grain boundaries.
(HAZ) of low-alloy steels for pressure vessels Segregation of impurities to grain boundaries
during hydrostatic pressure tests (Ref 17). The may play a significant role in reheat cracking
phenomenon of reheat cracking also received
much attention when small cracks in the HAZ
under the austenitic alloy cladding, which were

Fig. 14.25 Transmission electron micrograph showing a

dark-field image of fine, coherent c 0 (Ni3Al)
Fig. 14.23 Transmission electron micrograph showing precipitates formed in the grain matrix of alloy 214 (Ni-16Cr-
fine, coherent c 00 (Ni3Nb) precipitates formed in 4.5Al-3Fe-Y) at 800 °C (1470 °F) for 8 h. Source: Ref 15
the grain matrix of alloy 625 at 650 °C (1200 °F) for 24 h. Source:
Ref 13

Fig. 14.26 Transmission electron micrograph showing a

Fig. 14.24 Transmission electron micrograph showing a dark-field image of fine, coherent c 0 (Ni3Al)
dark-field image of fine, Ni2(Cr,Mo) ordered precipitates formed in the grain matrix of alloy 601 at about
phases formed in the grain matrix of Ni-16Cr-15Mo-3Fe alloy at 590 °C (1100 °F) for 2.5 years. Original magnification: 97,000×.
540 °C (1005 °F) for 16,000 h. Source: Ref 14 Source: Ref 16
Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/ 31/10/2007 4:07PM Plate # 0 pg 394

394 / High-Temperature Corrosion and Materials Applications

50
2.25Cr-1Mo, 2 kJ/mm
2.25Cr-1Mo, 3 kJ/mm
40
2.25Cr-1Mo, 4 kJ/mm
HCM2S, 2 kJ/mm
Reduction in area, %

HCM2S, 3 kJ/mm
30
HCM2S, 4 kJ/mm

20

10

0
550 600 650 700 750
PWHT, °C

Fig. 14.27 Reduction in area for 2.25Cr-1Mo and HCM2S at various postweld heat treatment (PWHT) temperatures with an initial
applied tensile stress of 325 MPa (47 ksi). Source: Ref 22

in ferritic steels. However, reheat cracking (or

stress-relief cracking or strain-age cracking)
also occurs in austenitic stainless steels, alloy
800H, and gamma prime (γ′) strengthened alloys;
however, the impurity segregation is not likely
to be as critical a factor in these austenitic alloys.
Reheat cracking (or stress-relief cracking) of
ferritic steels can be described using the test
results generated by Nawrocki et al. (Ref 22, 23)
when comparing 2.25Cr-1Mo steel with a modi-
fied 2.25Cr steel (HCM2S). HCM2S contains
similar chromium amounts, but lower carbon
contents (to reduce hardness of HAZ) along 20 µm
with the substitution of molybdenum (Mo) with
tungsten (W), vanadium (V), and niobium (Nb) Fig. 14.28 Scanning electron micrograph of fracture surface
of HCM2S ruptured air at 675 °C (1250 °F) under
to increase creep strength, compared with an initial stress of 325 MPa (47 ksi). Source: Ref 22
2.25Cr-1Mo steel (Ref 22). Comparing these
two alloy steels in terms of their respective
resistance to stress-relief cracking will allow specimen was then subjected to a simulated,
a better understanding of the important metal- programmed postweld heat treatment tempera-
lurgical factors that contribute to reheat cracking. ture and held at constant temperature and load
Nawrocki et al. (Ref 22, 23) used Gleeble until failure (Ref 22). The rupture ductility of
simulation techniques to evaluate stress-relief HCM2S was found to be significantly lower
cracking. Specimens were subjected to single- than that of 2.25Cr-1Mo, as shown in Fig. 14.27
pass weld thermal simulation cycles with a (Ref 22). HCM2S was found to be very brittle,
peak temperature of 1315 °C (2400 °F) followed as shown in Fig. 14.28 which reveals essentially
by cooling to room temperature. A tensile stress complete intergranular fracture (Ref 22). The
was imposed on the specimen during cooling authors (Ref 22) concluded that HCM2S was
and held for the duration of the test to simulate more susceptible to stress-relief cracking than
the residual stresses present in an actual weld- 2.25Cr-1Mo steel. Nawrocki et al. (Ref 23)
ment. After cooling to room temperature, the attributed the brittle, intergranular fracture of
Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/ 31/10/2007 4:07PM Plate # 0 pg 395

Chapter 14: Stress-Assisted Corrosion and Cracking / 395

HCM2S to the formation of fine (5 to 40 µm) tube carrying a synthesis gas stream at about
carbides (W/Fe-rich carbides) in the grain matrix 600 °C (1110 °F) with the outer diameter
that were resistant to plastic deformation and being thermally insulated. The failure oc-
the formation of grain-boundary Fe-rich M3C curred at the intrados of the tube bend.
carbides that nucleated microcavities. With
Type 321 tubes imbedded in a catalytic bed of
more carbide-forming alloying elements, such as aluminum silicate of a cracking reactor and
tungsten, vanadium, and niobium, HCM2S is carrying saturated steam failed at the bend
expected to form more matrix-strengthening section after 3 months of operation. The tube
carbides, thus resulting in brittle, intergranular metal temperatures were between 650 and
fracture due to significantly reduced deformation 750 °C (1200 and 1380 °F).
capability of the grain matrix.
Type 321 serpentine heat-exchanger tubes,
which were exposed to synthesis gases at 700
14.3.3 Cracking in Austenitic Stainless to 750 °C (1290 to 1380 °F) with internal
Steels and Fe-Ni-Cr Alloys temperature of about 300 °C (570 °F), failed
at the bend section after 3 months of service.
Among austenitic stainless steels, Type 321
and 347 have been frequently found to suffer The authors (Ref 24) observed that these failures
stress-relaxation cracking (or reheat cracking). exhibited a common characteristic in which
Titanium is used in Type 321 to “stabilize” the intergranular cracks were oxidized, leaving a
alloy by forming titanium carbides (TiCs) in metallic film in the center of the crack. The
the grain interior to prevent sensitization (i.e., metallic film was found to be essentially iron
formation of chromium carbides along grain with nickel varying from 0 to about 15% and
boundaries, thus creating a chromium-denuded little chromium. In addition to oxides observed
zone along grain boundaries) during heat treat- in the crack, the authors also observed in most
ment or cooling following welding. Type 347, on cases iron and nickel sulfides.
the other hand, uses niobium (referred to as Both Type 321 and 347 are known to be sus-
columbium until 1968) to form NbC in the grain ceptible to intergranular cracking in the heat-
interior to avoid sensitization during heat treat- affected zone (HAZ) during services at a certain
ment or cooling following welding. Stress- temperature range (Ref 18). It is generally be-
relaxation cracking has been reported for these lieved (Ref 25–30) that titanium carbides (in
two grades of austenitic stainless steels during Type 321) and niobium carbides (in Type 347) in
service at temperatures between approximately the region next to the fusion are put back into
500 and 750 °C (930 and 1380 °F) (Ref 24). De solution during welding and subsequent repre-
Santis et al. (Ref 24) summarized a number of cipitation of these fine carbides (during cooling)
failure cases involving these two grades of that strengthens the grain matrix in the HAZ
stainless steels. These authors indicated that the (particularly the coarse-grain HAZ) during ser-
failures were the result of intergranular cracking, vice or reheating during PWHT. Stress relaxation
and the components that failed were either in a in the matrix-strengthened HAZ during service
cold-worked structure or were associated with or PWHT causes cracking to develop at grain
welding. Also, all failures were found to have boundaries, thus resulting in brittle, intergranular
occurred after relatively short service times. fracture at the HAZ.
Cases of service failures related to Type 321 and Alloy 800H (Fe-32Ni-21Cr-0.4Al-0.4Ti) is
347 reported by De Santis et al. (Ref 24) are the most widely used Fe-Ni-Cr alloy for high-
summarized: temperature applications in the petrochemical

A straight Type 347 tube, which was ther- industry. Stress-relaxation cracking has also been
mally insulated externally and carried a gas- observed in this alloy under certain conditions.
eous hydrocarbon stream with traces of Kohut (Ref 31) reported intergranular cracking
hydrogen sulfide at 650 to 680 °C (1200 to failures in a transfer line made of alloy 800H after
1255 °F), failed after only 96 h of service. several months of service in a temperature range
A similar failure also occurred in the same of 540 to 705 °C (1000 to 1300 °F). The transfer
plant after only 15 days of service involving line ran between a pressure vessel and a feed-
a heavier wall tube made of Type 347 at the effluent exchanger in a petrochemical unit. The
bend near the weld. pipe was made from solution-annealed alloy

Failures of Type 321 tubes occurred after 800H plate. Fabrication of pipe involved press
service of about 1 month at the bend with the breaking the plate into two pipe halves followed
Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/ 31/10/2007 4:07PM Plate # 0 pg 396

396 / High-Temperature Corrosion and Materials Applications

by seam welding using shielded metal arc pipe seam weld. The alloy 800H header was
welding. Press breaking produced cold-worked 508 mm in diameter, had 34.5 mm wall thick-
structure in the pipe halves. The 30 cm (12 in.) ness, and was 4720 mm long. The header was
(diameter) transfer line failed at about 1 m (3 ft) fabricated by cold bending the plate material into
from the flange of the feed-effluent exchanger. pipe halves that were then seam welded using a
Four months later, a second failure occurred at nickel-base alloy filler metal. No heat treatment
about 10 m (30 ft) from the first failure. Both of this cold-formed and seam-welded pipe was
failures were found to occur in the areas with performed prior to installation as a header.
some deformation, which was believed to be the A joint industry sponsored project address-
result of the local repair. A third failure occurred ing stress-relaxation cracking failures in aus-
in the heat-affected zone of a weld initiating from tenitic welded joints was undertaken by Van
the inside diameter of tube surface after 1 year Wortel (Ref 33) at TNO Institute of Industrial
of service. Electron microprobe analysis showed Technology in Apeldoorn, The Netherlands.
no chlorine, sulfur, or other contaminants that Van Wortel reported major observations and
would cause embrittlement of the alloy. The conclusions derived from the test program. Since
material near the fracture still exhibited good the brittle, intergranular cracking failures were
room-temperature tensile ductility but much mainly associated with cold working and/or
higher yield and tensile strengths. Table 14.1 welding, the program mainly focused on the
shows the room-temperature tensile test results effect of cold working and welding on the
of the alloy 800H material obtained from the susceptibility of stress-relaxation cracking. A
30 cm (12 in.) transfer line near the fracture area specially developed three-point bending test
compared with the tensile test results of the rig, which was capable of producing a cracking
material that was obtained from the same area but failure with a similar crack morphology to the
was re-solution annealed. The stress-rupture tests service-induced failure, was used to determine
at 595 °C (1100 °F), however, revealed that the the susceptibility of the alloy to stress-relaxation
material near the fracture was very brittle, as cracking. Testing procedures were not described
shown in Table 14.2 (Ref 31). The sample that in the paper. The test results generated by Van
was obtained from the material near the fracture Wortel are summarized in Table 14.3 (Ref 33).
exhibited extremely low elongation under the Type 321 is generally considered to be
test condition, while the sample obtained from susceptible to stress-relaxation cracking in the
the same area, but was given a re-solution anneal- industry. In Van Wortel’s test program, however,
ing treatment, was found to exhibit excellent Type 321 was found to suffer no cracking under
ductility under the same test condition. the test condition. It is believed that the 321H
Korkhaus (Ref 32) observed similar brittle, material used in Van Wortel’s test program was
intergranular cracking in an alloy 800H pipe a fine-grained material with ASTM grain size
that was part of an outlet header of a process No. 7. Fine-grained materials are less susceptible
air preheater coil operated at about 600 °C to stress-relaxation cracking. All other materials
(1110 °F). The leakage of the header was tested were coarse-grained materials with ASTM
detected after 2 years of service. Cracks were grain sizes of No. 2 and coarser. Van Wortel
observed to occur primarily near the pipe-to-tube (Ref 33) reported that several heat treatments
weld and in the vicinity of the longitudinal were beneficial in improving the resistance of
the alloy to stress-relaxation cracking based on

Table 14.1 Results of room-temperature

tensile tests on alloy 800H samples obtained Table 14.2 Results of stress-rupture tests
from the area near the fracture as well as the performed at 595 °C (1100 °F) and 276 MPa
material from the same area but was re-solution (40 ksi) on alloy 800H samples obtained from
annealed the area near the fracture as well as the
Yield Ultimate tensile material from the same area but was re-solution
strength, strength, Elongation, Reduction
Material MPa (ksi) MPa (ksi) % in area, %
annealed
Material near 460 (68) 765 (111) 23.5 46.5 Rupture Elongation, Reduction
Material life, h % in area, %
fracture
Re-solution 205 (30) 506 (74) 55.0 72.2 Material near fracture 385 1.2 12.5
annealed Re-solution annealed 261 15.8 18.6
Source: Ref 31 Source: Ref 31
Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/ 31/10/2007 4:07PM Plate # 0 pg 397

Chapter 14: Stress-Assisted Corrosion and Cracking / 397

Table 14.3 Results of relaxation tests on cold-worked plates and welded joints
304H 321H AC66 800H 617
Condition at 575 °C at 650 °C at 575 °C at 650 °C at 575 °C at 650 °C at 650 °C
Weldment Cracks No cracks No cracks No cracks No cracks Cracks Cracks
PWHT(a) No cracks No cracks … … No cracks No cracks No cracks
Aged(b) No cracks … … No cracks … Cracks Cracks
CW(c) Cracks … No cracks No cracks … Cracks Cracks
HT(1)(d) No cracks … … … … No cracks …
HT(2)(e) No cracks … … … … No cracks …
HT(3)(f) … … … … … No cracks …
Note: Grain sizes for the alloys under testing were ASTM No. 2 for 304H, No. 7 for 321H, No. 2 for AC66, No. 0 for 800H, and No. 1 and No. 0 for 617. (a) PWHT at 875 °C/
3 h for 304H and 800H; 980 °C/3 h for 617 after welding and before testing. (b) Material aged at 650 °C for 16,000 h before testing. (c) Cold worked (CW) the metal from
0 to 15% before testing. (d) Heat treated the metal at 875 °C/3 h for 304H and 980 °C/3 h for 800H before cold working and testing. (e) Heat treated the metal after cold
working and before testing at 875 °C/3 h for 304H and 980 °C/3 h for 800H. (f ) Heat treated the metal from in-service cracked material at 980 °C/3 h before testing. Source:
Ref 33

his three-point bending testing method. How- Testing temperature, °C

ever, no examples were given to illustrate the 50
100 200 300 400 500 600 700
effectiveness of these heat treatments in actual Solution annealed
applications. 45 Aged 4000 h at 540 °C (1000 °F)
Van Wortel (Ref 33) indicated that the alloy Aged 4000 h at 595 °C (1100 °F) 300
Aged 4000 h at 650 °C (1200 °F)
is most susceptible to stress-relaxation cracking 40
when the alloy matrix is in an age-hardened
250
condition with a high density of fine precipitates

0.2% offset yield strength, MPa

0.2% offset yield strength, ksi

35
within the grain. This condition can significantly
reduce the deformation capability of the alloy. 30
200
Alloys under this condition are expected to suffer
cracking after only 0.1 to 0.2% relaxation strain 25
(Ref 33). Cold working can accelerate the 150
precipitation process within the grain, making 20
the alloy more susceptible to stress-relaxation
Room temperature

15
cracking. Van Wortel (Ref 33) proposed that the 100
microstructure in alloy 800H susceptible to
10
stress-relaxation cracking consisted of very fine
50
matrix carbides, which strengthen the grain
5
matrix, and grain-boundary carbides along with a
denuded carbide zone along the grain boundaries 0 0
that weaken the grain boundaries. Van Wortel, 200 400 600 800 1000 1200 1400
Testing temperature, °F
however, did not mention the formation of γ′ [Ni3
(Al,Ti)] precipitates in the grain matrix in the
age-hardened condition for Alloy 800 or 800H. Fig. 14.29 Strengthening at room temperature compared
with strengthening at the aging temperatures
The major strengthening mechanism for after aging at 540, 595, and 650 °C (1000, 1100, and 1200 °F) for
4000 h for alloy 800H containing 0.39% Al and 0.44% Ti.
alloy 800H at temperatures at which the alloy is Source: Ref 37
susceptible to stress-relaxation cracking is, in
fact, the precipitation of fine, coherent γ′ [Ni3(Al,
Ti)] precipitates (Ref 34–36). Lai and Kimball room temperature and approximately 25 to 35%
(Ref 37) observed that the maximum age hard- tested at respective aging temperatures) (Ref 37).
ening for alloy 800H was at 595 and 650 °C It is well known that age hardening in alloy 800
(1100 and 1200 °F). These authors further indi- or 800H is a strong function of combined Al+Ti
cated that the strengthening was more pro- content in the alloy. Alloys with higher content of
nounced at the elevated temperatures than Al + Ti exhibit higher degree of age hardening
at room temperature. This is illustrated in due to formation of more fine γ′ [Ni3(Al,Ti)]
Fig. 14.29 (Ref 37). The retained tensile elon- precipitates. An example is given in Fig. 14.30,
gation at both room-temperature and aging tem- which shows the aging behavior of two heats
peratures after 4000 h of aging was found to be of alloy 800H with Heat A containing higher
quite good (approximately 35 to 45% tested at combined Al + Ti content (0.83%) than Heat E
Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/ 31/10/2007 4:07PM Plate # 0 pg 398

398 / High-Temperature Corrosion and Materials Applications

100

95
Heat A 650 °C
(1200 °F)
90 Heat E
595 °C
(1100 °F)
540 °C (1000 °F)
85
540 °C (1000 °F)
Hardness, HRB

595 °C (1100 °F)

80
650 °C (1200 °F)
75
760 °C (1400 °F)

70

760 °C (1400 °F) 815 °C

65 815 °C (1500 °F) (1500 °F)

60
1 10 102 103 104 105

Unaged Aging time, h

Fig. 14.30 Aging behavior of two heats of alloy 800H with Heat A containing higher combined Al + Ti (0.83%) showing significantly
more age hardening than Heat E with much lower combined Al + Ti (0.58%). Source: Ref 37

(0.58% of combined Al + Ti) (Ref 37). Heat A 150

showed significantly more age hardening than
Heat E with both materials having similar grain RP1/10,000
sizes. Increasing the combined Al + Ti content in 650 °C (1200 °F)
alloy 800H causes significant strengthening RP1/30,000
in creep-rupture strength. This is illustrated in 100
Stress (N/mm2)

Fig. 14.31 (Ref 38). Increasing the combined

Al +Ti content in alloy 800H increases the 1%
creep strength at 650 °C (1200 °F), which was
the maximum age-hardening temperature. At 50
850 °C, γ′ [Ni3(Al,Ti)] precipitates coarsen in
alloy 800H; thus the effect of the combined 850 °C (1560 °F) RP1/10,000
Al +Ti content significantly diminishes, as
shown in Fig. 14.31 (Ref 38). RP1/30,000
Accompanying the strengthening of the 0
alloy produced by γ′ [Ni3(Al,Ti)] precipitates, 0.4 0.6 0.8 1.0 1.2 1.4
Σ Ti + AI, %
alloy 800H exhibits creep embrittlement at the
maximum age-hardening temperature, particu-
larly for the heats containing higher levels of Fig. 14.31 Effect of combined Al + Ti content in alloy 800H
on the creep strength (1% in 10,000 h and
combined Al +Ti content. This is illustrated in 30,000 h) at 650 and 850 °C (1200 and 1560 °F). Source: Ref 38
Fig. 14.32 (Ref 38). Figure 14.33 also shows
the similar effect of drastic reduction of stress- volume of these fine, coherent precipitates, thus
rupture ductility for alloys containing higher resulting in significant reduction in stress-rupture
levels of Al+Ti in alloy 800H when tested ductility at these temperatures. The rupture duc-
at 650 °C (1200 °F)—the maximum age- tility of alloy 800 can be further reduced to close
hardening temperature (Ref 39). The data sug- to nil with prior cold work when tested at 600 or
gest that the low rupture ductility in alloy 800 650 °C (1110 or 1200 °F). Stone et al. (Ref 39)
at 600 and 650 °C (1110 and 1200 °F) is the showed the effect of cold work on stress-rupture
result of γ′ [Ni3(Al,Ti)] precipitates with higher ductility when tested at 600 °C (1110 °F) for
combined Al +Ti content producing a larger alloy 800. Their test results showed that the
Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/ 31/10/2007 4:07PM Plate # 0 pg 399

Chapter 14: Stress-Assisted Corrosion and Cracking / 399

rupture elongation was reduced from about 10% prestrained specimens, which was so low it could
in the solution-annealed condition to almost not be obtained by the traditional method of
nil with 10 to 20% prior cold work, as shown fitting the ends of the ruptured specimen together
in Fig. 14.34 (Ref 39). Similar results were for determination of the percent of elongation,
observed by Smith (Ref 40) when he conducted was taken from the creep curves.
stress-rupture tests at temperatures from 540 to
650 °C (1000 to 1200 °F). Smith’s test results
14.3.4 Cracking in Nickel-Base Alloys
are summarized in Table 14.4 (Ref 40). All the
prestrained specimens were quite brittle under Some wrought nickel-base chromia-forming
the test conditions, exhibiting rupture elong- alloys (i.e., nickel-base alloys forming chromium
ation of less than 1%. The elongation for the oxide scales for high-temperature corrosion re-
sistance) that are designed for high-temperature
applications contain a relatively low level of alu-
60 minum, typically 1 to 2%, to further improve
the oxidation resistance of the alloy. Some of
the more familiar alloys in this group includes
50
alloys 617 (1.2% Al) and 601 (1.4% Al). As
0.1 h discussed earlier, Van Wortel (Ref 33) reported
40 that alloy 617 was susceptible to stress-relaxation
Elongation, %

cracking at 650 °C (1200 °F), as shown in

30 Table 14.3. The author, however, provided little
1000 h
information about the microstructure and aging
20 characteristics of alloy 617. Alloy 617 also is
known to age harden at 600 and 650 °C (1110
10,000 h
and 1200 °F), as shown in Fig. 14.35 (Ref 41).
10 30,000 h Alloy 617, similar to alloy 800H, exhibits
excellent room-temperature ductility after long-
0 term aging, as shown in Fig. 14.36 (Ref 41).
0.4 0.6 0.8 1.0 1.2 1.4
Σ Ti + AI, % Nevertheless, alloy 617, similar to alloy 800H,
was susceptible to stress-relaxation cracking
Fig. 14.32 Creep rupture ductility of alloy 800H as a function at 650 °C (Table 14.3). Bassford and Schill
of combined Al + Ti content in the alloy tested at (Ref 42) believed that γ′ precipitates are a
650 °C (1200 °F). Source: Ref 38
major contributor to age hardening in alloy 617
at 650 °C as well as 600 °C. Both alloys 800H
and 617 are strengthened primarily by γ′ preci-
300 pitates at temperatures when stress-relaxation
cracking has been observed. For nickel-base
0.4% AI 0.6% Ti
alloys, this type of cracking is generally referred
to as “strain-age cracking.”
Stress, MPa

0.2% AI 0.4% Ti
100 Stahl et al. (Ref 43) reported a reformer tube
failure in a steam reformer for hydrogen pro-
duction. The reformer tube was made of alloy
50 601. The product gas cooled from 900 to 600 °C
(1650 to 1110 °F) in the tube and then flowed
30 through the outlet manifold system. Cracking
50 was observed to occur near the outlet header
Elongation, %

40 in the heat-affected zone parallel to the weld

30 0.2% AI 0.4% Ti fusion line after about 1 year of operation. Their
20
repair method involved cutting out all the pro-
10
0
0.4% AI 0.6% Ti blem connections between the reformer tube and
10 102 103 104 105 the outlet header, which were replaced with
Time, h new outlet manifold welded connections. These
welded connections were then heat treated at
Fig. 14.33 Effects of combined Al + Ti content in alloy 800 950 °C (1740 °F) for 1 h before they were con-
on stress rupture strength and rupture ductility at
650 °C (1200 °F). Source: Ref 39 nected to the reformer tubes. The postfabrication
Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/ 31/10/2007 4:07PM Plate # 0 pg 400

400 / High-Temperature Corrosion and Materials Applications

500
5% CW
10% CW
20% CW
300 25% CW
Stress, MPa

200
ST. 1120 °C (2050 °F)

100

40
Elongation, %

30

20

10 ST. 1120 °C (2050 °F)

0
10 102 103 104 105
Time, h

Fig. 14.34 Effects of cold work (cw) on the rupture ductility of alloy 800 at 600 °C (1110 °F). Solid lines represent the data generated
from the specimens in the as-solution heat treated condition. The specimens were solution heat treated (ST) at 1120 °C
(2050 °F). Source: Ref 39

Table 14.4 Effect of 20% prestrain in Exposure temperature, °C

compression on stress rupture ductility of alloy 595 650 705 760
800H tested at 540, 565, 595, and 650 °C (1000, 90 600
Exposure time, h
1050, 1100, and 1200 °F) using bar specimens 48,000 h
in both solution-annealed (SA) and prestrained 80
(PS) conditions 12,000 h
Yield strength, MPa
4,000 h 500
Yield strength, ksi

70
Rupture Rupture 1,000 h
Applied
life, h elongation, %
Test temperature, stress, 60 400
°C (°F) MPa (ksi) SA PS SA PS
100 h
540 (1000) 345 (50) 1214 129 19 0.26 50 Not exposed
310 (45) 3350 319 13 0.20
565 (1050) 310 (45) 615 92 15 0.2 300
275 (40) 1953 546 11 0.3 40
595 (1100) 241 (35) 772 247 13 0.2
207 (30) 2670 721 14 0.28
30
207(a) (30(a)) … 7342(a) … 0.3(a) 1100 1200 1300 1400
650 (1200) 172 (25) 546 861 16 0.45 Exposure temperature, °F
155 (22.5) 1152 1565 20 0.4
138 (20) 2352 2589 22 0.33
(a) Prestrained 20% in tension; broke at extensomer weld. Source: Ref 40
Fig. 14.35 Room-temperature yield strengths of alloy 617
after aging at different temperatures for various
times up to 48,000 h. Source: Ref 41

heat treatment, according to the authors (Ref 43), which contains about 1.4% Al and is known to
was (a) to eliminate the effect of retained cold- form γ′ (Ni3Al) precipitates. The heat treatment
work and welding stresses and (b) to coarsen at 950 °C is believed to mainly coarsen both
both the grain-boundary precipitates and grain- grain-boundary and intragranular carbides in
matrix precipitates to reduce age-hardening addition to relieving residual stresses resulting
effects. These authors did not mention in their from cold working and welding. It should pro-
paper about γ′ (Ni3Al) precipitates being a vide some improvement to the resistance of
possible strengthening mechanism in alloy 601, the alloy to strain-age cracking. Nevertheless, the
Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/ 31/10/2007 4:07PM Plate # 0 pg 401

Chapter 14: Stress-Assisted Corrosion and Cracking / 401

temperature of 950 °C (1740 °F) in the heat room temperature. At 590 °C (1100 °F), how-
treatment would be above the solvus temperature ever, the ductility was found to be much lower
of the γ′ (Ni3Al) precipitates, and γ′ (Ni3Al) and the yield strength was very high, which was
precipitates can form again during the service at close to that at room temperature. Transmission
strain-age cracking temperatures of 600 and electron microscopy examination of the recup-
650 °C (1110 and 1200 °F). erator shell showed the presence of fine, γ′ pre-
Lai (Ref 16) reported a case of strain-age cipitates, as shown in Fig. 14.26. Tensile blanks
cracking involving a recuperator shell made were obtained from the recuperator shell and
of alloy 601 that was used for preheating air re-solution annealed at 1150 °C (2100 °F) for
for combustion. The recuperator shell suffered 30 min followed by water quenching. Tensile
through-thickness cracking after 2.5 years of properties of these re-solution annealed speci-
service. The failure occurred at the location mens at both room temperature and 590 °C
where the metal temperature was approximately (1100 °F) are included in Table 14.5 for com-
590 °C (1100 °F). No cracking was observed at parison. The tensile properties of re-solution
other locations with temperatures either higher or annealed material were similar to those of sol-
lower than 590 °C (1100 °F). Both the air side ution-annealed material reported in the literature
and flue gas side of the recuperator shell showed (Ref 44). Assuming that the re-solution annealed
little oxidation or corrosion attack. The recup- material has the properties of the original plate
erator shell (7.9 mm, or 5/16 in. thick) failed by material prior to service, the data indicated that
intergranular fracture, as shown in Fig. 14.37. the alloy 601 recuperator shell increased its yield
Figure 14.38 shows a crack propagating along strength at 590 °C (1100 °F) by almost three
grain boundaries. The figure shows some evi- times (but only two times at room temperature)
dence of cold work that is likely to result from the as a result of the service exposure at about
original fabrication of the recuperator shell. 590 °C (1100 °F). In order to determine that
The tensile property of alloy 601 from the
recuperator shell near the main fracture was
evaluated. Tensile blanks were obtained and
machined into round tensile specimens with
adequate material being machined off from both
internal and external surfaces (ID and OD) of the
shell. Table 14.5 shows the tensile properties of
the material from the area that was near the main
fracture (Ref 16). The material from the recup-
erator shell exhibited good tensile ductility at

(a)

Exposure temperature, °C
600 650 700 750
Not exposed
70

60
Elongation, %

100 h
50 Exposure 1000 h
time, h
4,000 h
40 12,000 h

30 8,000 h
48,000 h

20
1100 1200 1300 1400 (b)
Exposure temperature, °F
Fig. 14.37 Scanning electron micrographs, (a) low magni-
Fig. 14.36 Room-temperature tensile elongation of alloy 617 fication and (b) high magnification, showing
after aging at different temperatures for various intergranular fracture surface of the alloy 601 recuperator shell.
times up to 48,000 h. Source: Ref 41 Original magnification: (a) 13× and (b) 100×. Source: Ref 16
Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/ 31/10/2007 4:07PM Plate # 0 pg 402

402 / High-Temperature Corrosion and Materials Applications

the recuperator shell material was in a brittle shell material and the shell material after re-
condition (extremely low creep ductility) under solution annealing at 1150 °C (2100 °F). The
creep condition, stress rupture testing was per- test results are shown in Table 14.6 (Ref 16). The
formed at 590 °C (1100 °F) involving smooth- brittle nature of the alloy under the creep test
notch specimens obtained from the recuperator condition at 590 °C (1100 °F) was substantiated
from these tests. Furthermore, the test results
indicated that a full recovery of the original
ductile material condition can be obtained by
re-solution annealing the material.
Lai (Ref 16) presented additional test data to
further substantiate that alloy 601 creep embrit-
tlement was related to the age-hardened con-
dition produced during service at 590 °C
(1100 °F). An annealed alloy 601 plate of dif-
ferent heat number was obtained from an alloy
supplier for evaluation. Specimen blanks were
aged at 590 °C (1100 °F) for 10,000 h. Stress-
rupture tests were performed at 590 °C (1100 °F)
and 310 MPa (45 ksi) on both annealed and
aged specimens using smooth-notch speci-
mens. Results of these tests are summarized in
Table 14.7. The aged specimen, although frac-
tured at the smooth section, clearly showed an
embrittled condition with very low creep duc-
tility and intergranular fracture. The as-received
specimen showed much better creep ductility
with dimple rupture, which is characteristic of
a ductile rupture.
Many high-temperature, nickel-base alloys
are solid-solution strengthened using primarily
Fig. 14.38 Optical micrograph showing intergranular
molybdenum and/or tungsten to increase their
cracking near the main fracture. Original mag-
nification: 100×. Source: Ref 16 tensile and creep rupture strengths. These alloys

Table 14.5 Tensile properties of alloy 601 specimens obtained from the area near the main fracture
in the alloy 601 recuperator shell that suffered intergranular through-wall cracking after 2.5 years
of service with the metal temperature at about 590 °C (1100 °F) during service. For each test condition,
two tests were run and both values are reported.
Test temperature, 0.2% yield strength, Ultimate tensile strength, Elongation, Reduction in
Material °C (°F) MPa (ksi) MPa (ksi) % area, %
From recuperator Room temperature 426, 414 (61.8, 60.0) 883, 865 (128.0, 125.5) 36, 34 32, 34
590 (1100) 372, 386 (54.0, 56.0) 606, 600 (87.9, 87.0) 18, 13 15, 15
RSA(a) Room temperature 232, 224 (33.7, 32.5) 616, 614 (89.3, 89.1) 56, 53 66, 53
590 (1100) 154, 142 (22.4, 20.6) 481, 478 (69.7, 69.3) 63, 62 55, 54
(a) Test blanks from the recuperator shell near the main fracture were re-solution annealed (RSA) at 1150 °C (2100 °F) for 30 min followed by water quenching. Source:
Ref 16

Table 14.6 Results of stress rupture tests performed at 590 °C (1100 °F) and 310 MPa (45 ksi)
on smooth-notch round specimens obtained from the 601 recuperator shell and from the
re-solution-annealed recuperator shell material
Specimen condition Failure location Fracture mode
From recuperator shell Notch, rupture life: 3110 h Intergranular fracture
From recuperator shell + 1150 °C (2100 °F) Smooth section, rupture life: 175 h, Dimple rupture
for 30 min and water quench 22% elongation, 14% reduction in area
Source: Ref 16
Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/ 31/10/2007 4:07PM Plate # 0 pg 403

Chapter 14: Stress-Assisted Corrosion and Cracking / 403

are also strengthened by chromium-rich carbides. is actually strengthened by fine, coherent γ″

Among the most widely used solid-solution- (Ni3Nb) precipitates, and alloy X-750 containing
strengthened alloys for high-temperature appli- Al, Ti, and Nb is strengthened by both γ′ and γ″
cations is alloy X (Ni-22Cr-18Fe-9Mo), which precipitates. Many of the applications for these
among other applications is widely used for alloys are in sheet products. As a result, fabri-
gas turbine combustion chambers in both land- cated products are under much less constraint
based gas turbine power generation and aircraft than are heavy section components. Further-
engines. In smooth-notch stress rupture tests more, the grain sizes for sheet products are
to investigate the susceptibility of strain-age typically ASTM No. 4 or finer. Accordingly,
cracking of alloy 601, Lai (Ref 16) in- components made of sheet products are less
cluded alloy X to examine this solid-solution- susceptible to strain-age cracking. Nevertheless,
strengthened alloy, which is also strengthened in some cases components made of sheet pro-
by carbides (primarily M6C carbides) but with ducts for these alloys can be heavily formed
no γ′ precipitates. The results of his tests on with severe cold-worked conditions, thus making
alloy X are summarized in Table 14.8 (Ref 16). them susceptible to strain-age cracking even
Alloy X aged for 10,000 h at 590 °C (1100 °F), during postfabrication heat treatments.
under the same test conditions as were used for Rowe (Ref 45) studied strain-age cracking of
alloy 601, was found to exhibit excellent creep wrought γ′- or γ″-strengthened sheet alloys by
ductility (about 55% rupture elongation) with using a controlled heating rate testing (CHRT)
ductile dimple rupture. method that was developed in 1960s (Ref 46).
Some nickel-base alloys use γ′ [Ni3(Al,Ti)] Testing involves heating a tensile specimen at a
precipitates for providing strengthening in ad- controlled heating rate to a test temperature in
dition to solid-solution strengthening using the γ′ or γ″ precipitation temperature range, and
elements such as molybdenum and/or tungsten. then pulling the specimen in tension to failure.
Some of the widely known wrought alloys in The heating rate of the postweld annealing heat
this group include alloys 263 (Ni-20Cr-20Co- treatment of a fabricated component could be
6Mo-0.5Al-2.2Ti), 718 (Ni-18Cr-19Fe-3Mo- used for the CHRT in the test. The minimum
5Nb), X-750 (Ni-16Cr-7Fe-0.7Al-2.5Ti-1Nb), tensile elongation from the test results is then
Waspaloy (Ni-19Cr-14Co-4Mo-1.5Al-3Ti), and used as a guide for the susceptibility to strain-age
R-41 (Ni-19Cr-11Co-10Mo-1.5Al-3Ti). Alloy cracking for the alloy. Figure 14.39 summarizes
718, which contains Nb instead of Al and Ti, the test results generated by Rowe (Ref 45). The
results indicated that increasing the combined
Al + Ti + Nb content in the alloy decreases the
Table 14.7 Results of stress rupture tests at minimum elongation, thus increasing the sus-
590 °C (1100 °F) and 310 MPa (45 ksi) on ceptibility to strain-age cracking.
smooth-notch specimens of as-received and
aged alloy 601 specimens from alloy 601 plate
of different heat number obtained from an alloy
supplier 20
263
Minimum CHRT elongation, %

Failure Rupture Elongation, Reduction Fracture 18

Specimen location life, h % in area, % mode 16
As-received Smooth 963 19 21 Dimple 14
Aged(a) Smooth 5348 4 4 Intergranular
12
(a) Aged at 590 °C (1100 °F) for 10,000 h. Source: Ref 16 X-750
10
718
8
6
Waspaloy
4
Table 14.8 Results of stress rupture tests at PK33
590 °C (1100 °F) and 310 MPa (45 ksi) on 2
R-41
smooth-notch specimens of as-received and aged 0
2 3 4 5 6 7 8
alloy X specimens
Al + Ti + Nb, at.%
Failure Rupture Elongation, Reduction Fracture
Specimen location life, h % in area, % mode
As-received Smooth 1574 42 44 Dimple
Fig. 14.39 Results of controlled heating rate tests (CHRT)
in terms of minimum elongation (%) as a func-
Aged(a) Smooth 1919 55 63 Dimple tion of combined Al + Ti + Nb content in atomic percent for
(a) Aged at 590 °C (1100 °F) for 10,000 h. Source: Ref 16 selected c - and/or c 00 -strengthened wrought sheet alloys. Source:
0

Ref 45
Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/ 31/10/2007 4:07PM Plate # 0 pg 404

404 / High-Temperature Corrosion and Materials Applications

Nickel-base alloys containing high levels of It is quite likely that a long-range ordered Ni-
aluminum or aluminum plus titanium can be Cr-Mo alloy may also be susceptible to strain-age
highly susceptible to strain-age cracking, par- cracking. Alloy C-22 (Ni-22Cr-13Mo-3W) is
ticularly in plate or thick-wall tubing products. selected to be a candidate nuclear waste container
Some of these alloys include alloy 214 (4.5% material for the Yucca Mountain site in Nevada
Al), which forms a tenacious Al2O3 oxide scale (Ref 49). Studies on the changes in micro-
to resist oxidation at very high temperatures, structure, mechanical properties, and corrosion
and alloy 693 (3% Al), which forms Cr2O3/ resistance of C-22 alloy after long-term aging
Al2O3 oxide scales to resist metal dusting in the have been carried out in recent years (Ref 50–
temperature range of 480 to 700 °C (900 to 52). Rebak et al. (Ref 52) reported the obser-
1300 °F). vation of long-range ordered phases, [Ni2(Cr,
Ni-Cr-Mo and Ni-Mo alloys have the potential Mo)], in C-22 alloy after aging at 595 °C
for developing a long-range ordered Ni2(Cr,Mo) (1100 °F) for 1000 and 16,000 h, at 540 °C
phase after long aging times at a certain tem- (1000 °F) for 1000 h, and at 430 °C (800 °F) for
perature range. Tawancy (Ref 47) observed the 30,000 and 40,000 h. Room-temperature tensile
formation of [Ni2(Cr,Mo)] phases in alloy S reduction in area of C-22 was reported to have
(Ni-15.5Cr-14.5Mo) after aging for 8000 h been about 80% in annealed condition to 75%
at 540 °C (1000 °F), as shown in Fig. 14.40 after aging for 40,000 h at 430 °C (800 °F),
(Ref 47). The ordered phase caused the alloy which was apparently in the early stage of long-
to significantly increase its room-temperature range ordered condition (Ref 52). In a similar
tensile strengths, particularly yield strength, as study on aging behavior of C-22 alloy, the alloy
shown in Table 14.9 (Ref 47). Tawancy and after aging at 595 °C (1100 °F) for 16,000 h
Asphahani (Ref 48) reported that C-276 alloy showed no strengthening for one heat but slight
(Ni-16Cr-16Mo-4W) exhibited significant room- strengthening for the other heat, as shown in
temperature strengthening when the alloy was Table 14.11 (Ref 51). The authors (Ref 51)
aged to develop a fully ordered structure, as observed continuous grain-boundary carbides
shown in Table 14.10. The Ni2(Cr,Mo) ordered in the aged samples and mentioned about the
phase exhibits a morphology similar to γ′ and γ″ formation of Ni2(Cr,Mo) ordered phases, but
precipitates and produces similar strengthening without presenting positive identification of
effects as γ′- and γ″-strengthened alloys. the ordered phases using transmission electron
microscopy. With either no strengthening or
slight strengthening as presented in Table 14.11,

Table 14.9 Room-temperature tensile properties

of alloy S comparing annealed material with the
material aged for 8000 h at 540 °C (1000 °F)
0.2% yield strength, Ultimate tensile Elongation,
Condition MPa (ksi) strength, MPa (ksi) %
Annealed specimen 436 (63) 889 (129) 58
Aged specimen(a) 824 (119) 1284 (186) 42
(a) Aged at 540 °C (1000 °F) for 8000 h. Source: Ref 47

Table 14.10 Room-temperature tensile

properties of C-276 alloy comparing annealed
material (disordered structure) with aged material
with a long-range ordered structure
0.2% yield Ultimate
strength, tensile strength, Elongation,
Condition MPa (ksi) MPa (ksi) %
Annealed 356 (51.6) 792 (114.8) 61
Fig. 14.40 Transmission electron micrograph showing long-
Ordered structure(a) 770 (111.7) 1240 (179.8) 28
range ordered phases [Ni2(Cr,Mo)] in a dark field
image using a h220i reflection in alloy S after 8000 h at 540 °C (a) Aging condition was not reported in the paper. Source: Ref 48
(1000 °F). Source: Ref 47
Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/ 31/10/2007 4:07PM Plate # 0 pg 405

Chapter 14: Stress-Assisted Corrosion and Cracking / 405

it is unlikely the ordered phases were fully 14.3.5 Mitigation of Reheat Cracking
developed. The slight reduction in ductility as and Stress-Relaxation Cracking or
shown in Table 14.11 is most likely caused by Strain-Age Cracking
the formation of continuous grain-boundary
carbides. The mechanical properties that were As was discussed in previous sections, the
under investigation in these studies (Ref 50–52) reheat cracking due to heat treatment or the
included tensile tests and impact toughness stress-relaxation cracking (or strain-age crack-
testing. When long-range ordered phases are ing) that occurs during service is primarily re-
fully developed, it may be imperative to perform lated to strengthening of the grain matrix by fine
stress-relaxation tests or smooth-notch stress precipitates that formed at the intermediate
rupture tests at the ordering temperatures to temperatures. Thus, to minimize the chances of
determine the susceptibility of the alloy to stress- developing reheat cracking during postweld heat
relaxation cracking (or strain-age cracking). treatment of a welded component or annealing/
A relatively new age-hardenable Ni-Cr-Mo stress-relieving of a cold-formed component, it
alloy (C-22HS) with 21Cr and 17Mo in nickel is recommended that the component be heated
was developed using long-range ordered phases through that temperature range quickly to mini-
[Ni2(Cr,Mo)] to produce strengthening through mize the formation of those detrimental pre-
a heat treatment (Ref 53). The heat treatment cipitates. As for the stress-relaxation cracking
produces a fully developed ordered structure that (or strain-age cracking) that occurs during
produces significant strengthening, particularly service, one effective method for mitigating
at the ordering temperature (i.e., 595 °C, or the cracking is to remove the residual stresses in
1100 °F). Tensile properties of Alloy C-22HS in the component by performing a postweld heat
both as-annealed and as-heat-treated conditions treatment for a welded component or an anneal-
are summarized in Table 14.12 (Ref 53). This ing/stress-relieving heat treatment for a cold-
behavior is quite similar to the alloys strength- formed component (e.g., pipe U-bend section)
ened by γ′ precipitates, such as Waspaloy, R-41, prior to service. Other factors that may be bene-
and so forth. Thus, the alloys that are strength- ficial in reducing the cracking tendency during
ened by the ordered Ni2(Cr,Mo) phase may be service include (a) minimizing stress risers, (b)
equally susceptible to strain-age cracking as γ′ the use of a fine-grained material, and (c) a heat
alloys. treatment involving heating the component to
a temperature high enough to coarsen preci-
pitates (e.g., carbides) to reduce the matrix
strengthening.
Table 14.11 Room temperature tensile
properties of C-22 alloy comparing annealed
material with the material aged for 16,000 h at
595 °C (1100 °F) 14.4 Summary
Ultimate
0.2% yield tensile Most high-temperature components are under
strength, strength, Elongation, Reduction
Heat Condition MPa (ksi) MPa (ksi) % in area, % stress during service. The tensile stresses (or
No. 1 Annealed 340 (49.3) 768 (111.3) 66.2 76.1 strains) that are imposed on the component
Aged(a) 381 (55.2) 806 (116.9) 55.4 50.7 can cause the alloy to suffer preferential corro-
No. 2 Annealed 342 (49.6) 757 (109.7) 70.0 77.0
Aged(a) 485 (70.3) 931 (135.0) 39.0 31.1
sion penetration in high-temperature, corrosive
environments, particularly low-oxygen and high-
(a) Aged at 595 °C (1100 °F) for 16,000 h. Source: Ref 51
sulfur partial pressure conditions. The effects

Table 14.12 Tensile properties of C-22HS (a heat treatable alloy) comparing the annealed material
and heat treated material tested at both room temperature and 595 °C (1100 °F)
0.2% yield strength, Ultimate tensile Reduction
Condition Test temperature, °C (°F) MPa (ksi) strength, MPa (ksi) Elongation, % in area, %
Annealed Room temperature 406 (59) 822 (119) 61 75
HT(a) Room temperature 742 (108) 1232 (179) 40 50
Annealed 595 (1100) 222 (32) 640 (93) 72 67
HT(a) 595 (1100) 542 (79) 934 (135) 48 66
(a) Heat treated at 705 °C (1300 °F)/16 h/furnace cool to 605 °C (1120 °F)/32 h followed by air cool. Source: Ref 53
Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/ 31/10/2007 4:07PM Plate # 0 pg 406

406 / High-Temperature Corrosion and Materials Applications

of the tensile stresses and strains on the prefer- 5. M.F. Stroosnijder, V. Guttmann, and J.H.W.
ential sulfidation penetration attack on various de Wit, Corrosion and Creep Behaviour
alloys in sulfidizing environments are reviewed. of Alloy 800H in Sulphidizing/Oxidizing/
The preferential sulfidation penetration is con- Carburizing Environments at 700 °C—Part
sidered to be a precursor to the circumferential II: Creep Behaviour, Werkst. Korros., Vol
cracking that has been observed in the waterwall 41, 1990, p 508
tubes for some supercritical coal-fired boilers 6. M.F. Stroosnijder, V. Guttmann, and R.J.N.
fired under low NOx combustion conditions. Gommans, Influence of Creep Deformation
A detailed discussion on the preferential sulfi- on the Corrosion Behaviour of a CeO2
dation penetration and the circumferential crack- Surface-Modified Alloy 800H in a Sulphi-
ing of the waterwall tubes observed in some dising-Oxidixing-Carburising Environment,
supercritical coal-fired boilers is presented. Mater. Sci. Eng., Vol A121, 1989, p 581
Additional discussion on preferential sulfidation 7. V. Guttmann, K. Stein, and W.T. Bakker,
penetration and circumferential cracking ob- Deformation-Corrosion Interactions in Sel-
served in waterwall tubes is also presented in ected Advanced High Temperature Alloys,
Chapter 10 “Coal-Fired Boilers.” Mater. High Temp., Vol 14 (No. 2/3), 1997,
Stresses that are either residual stresses in the p 61
component or externally applied stresses can 8. J. Le Coze, U. Franzoni, O. Cayla, F.
cause brittle, intergranular cracking when the Devisme, and A. Lefort, The Development
component is under service at the low end of the of High-Temperature Corrosion-Resistant
intermediate temperature range. This cracking Aluminium-Containing Ferritic Steels,
phenomenon is frequently referred to as “reheat Mater. Sci. Eng., Vol A120, 1989, p 293
cracking,” or “stress-relaxation cracking,” or 9. Materials, Part D—Properties 2005
“strain-age cracking.” Reheat cracking occurs Addenda, ASME Boiler and Pressure Vessel
when a welded component is subjected to a Code, ASME, New York, July 1, 2005
postweld heat treatment (PWHT) or when a 10. M. Le Calvar, P.M. Scott, T. Magnin, and
cold-formed component is subjected to an P. Rieus, Strain Oxidation Cracking of
annealing or stress-relieving heat treatment. Austenitic Stainless Steels at 610 °C, Cor-
Stress-relaxation cracking (or strain-age crack- rosion, Vol 54 (No. 2), 1998, p 101
ing) occurs when the component is under service. 11. K. Rorbo, Experience with Nitriding of
This stress-induced cracking behavior of ferritic Austenitic Stainless Steel and Inconel in
steels, stainless steels, Fe-Ni-Cr alloys, and Ammonia Environments, Environmental
nickel-base alloys is reviewed. Degradation of High Temperature Mate-
rials, Series 3, No. 13, Vol 2, The Institution
of Metallurgists, London, 1980, p 147
12. D.L. Klarstrom, Metallurgical Factors that
REFERENCES Promote Cracking During the Heat Treat-
ment of High Performance Alloys, Heat-
1. M. Schütze, Deformation and Cracking Resistant Materials II (Conf. Proc.) Second
Behavior of Protective Oxide Scales on International Conference on Heat-
Heat-Resistant Steels under Tensile Strain, Resistant Materials, K. Natesan, P. Ganesan,
Oxid. Met., Vol 24 (No. 3/4), 1985, p 199 and G. Lai, Ed., ASM International, 1995,
2. M. Schütze, The Healing Behavior of Pro- p 487
tective Oxide Scales on Heat-Resistant 13. H.M. Tawancy, Structure & Properties of
Steels After Cracking under Tensile Strain, High Temperature Alloys: Applications of
Oxid. Met., Vol 25 (No. 5/6), 1986, p 409 Analytical Electron Microscopy, King Fahd
3. G.R. Smolik and J.E. Flinn, Stress and University of Petroleum & Minerals,
Environmental Interactions for INCOLOY Dhahran, Saudi Arabia, 1993, p 144
800H in Coal Gasification Environments, 14. H.M. Tawancy, Structure & Properties of
J. Mater. Energy Systems, Vol 8 (No. 3), High Temperature Alloys: Applications of
1986, p 297 Analytical Electron Microscopy, King Fahd
4. V. Guttmann and J. Timm, Corrosion and University of Petroleum & Minerals,
Creep of Alloy 800H under Simulated Coal Dhahran, Saudi Arabia, 1993, p 176
Gasification Conditions, Werkst. Korros., 15. R.B. Herchenroeder, G.Y. Lai, and K.V.
Vol 39, 1988, p 322 Rao, A New, Wrought, Heat-Resistant
Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/ 31/10/2007 4:07PM Plate # 0 pg 407

Chapter 14: Stress-Assisted Corrosion and Cracking / 407

Ni-Cr-Al-Fe-Y Alloy, J. Met., Vol 35 Heavy Section18-12-Nb Austenitic Stain-

(No. 11), 1983, p 16 less Steel Welded Joints, JISI, 1961, p 29
16. G.Y. Lai, presented at ASM Materials Week 30. R.N. Younger and R.G. Baker, Heat Af-
(Cincinnati, OH), Oct 20–24, 1991 fected Zone Cracking in Welded Austenitic
17. A. Dhooge and A. Vinckier, Reheating Steels During Treatment, Brit. Weld. J.,
Cracking—A Review of Recent Studies, 1961, p 579
Int. J. Pressure Vessels Piping, Vol 27, 1987, 31. G.B. Kohut, A Case History of Cracking
p 239 Failures in Deformed Austenitic Alloy Pipe
18. C.F. Meitzner, Stress-Relief Cracking in Operating in the Temperature Range of
Steel Weldments, WRC Bull, Vol 211, Nov 1000–1300 Deg F (538–705 Deg C), Trans.
1975 ASME, 1975, p 316
19. A. Dhooge, R.E. Dolby, J. Sebille, R. 32. J. Korkhaus, Application of Corrosion-
Steinmetz, and A.G. Vinckier, A Review Resistant Steels in the Chemical Industry,
of Work Related to Reheat Cracking in Stainless Steels World 1999 Conference,
Nuclear Reactor Pressure Vessel Steels, Int. Conf. Proc., Book 1, KCI Publishing BV,
J. Pressure Vessels Piping, Vol 6, 1978, The Netherlands, 1999, p 27
p 329 33. J.C. Van Wortel, Relaxation Cracking in the
20. A.G. Vinckier and A.W. Pense, “A Review Process Industry, an Underestimated Pro-
of Underclad Cracking in Pressure Vessel blem, BALTICA IV: Plant Maintenance
Components,” WRC Bulletin No. 197, for Managing Life & Performance, Vol 2,
Welding Research Council, Aug 1974 S. Hietanen and P. Auerkari, Ed., VTT
21. R.E. Dolby and G.G. Saunders, “A Review Technical Research Centre of Finland,
of the Problem of Reheat Cracking in Espoo, Finland, 1998, p 637
Nuclear Vessel Steels,” Report 3453/1/75, 34. P.G. Stone, J. Orr, and J.C. Guest, Status
The Welding Institute, April 1975 Review of Alloy 800, Proc. British Nuclear
22. J.G. Nawrocki, J.N. DuPont, C.V. Robino, Energy Society Conference, S.F. Pugh, Ed.,
and A.R. Marder, The Stress-Relief Crack- AERE Harwell, 1975, p 15
ing Susceptibility of A New Ferritic Steel— 35. J. Orr, “A Review of the Structural Char-
Part 1: Single-Pass Heat-Affected Zone acteristics of Alloy 800,” Paper No. 2-1,
Simulations, Weld. J., Dec 2000, p 355s Proc. Petten Conference, North-Holland
23. J.G. Nawrocki, J.N. DuPont, C.V. Robins, Publishing, 1978
J.D. Puskar, and A.R. Marder, The Mech- 36. N. Persson, “Mechanical Properties of Alloy
anism of Stress-Relief Cracking in a Ferritic 800 Above 600 C,” Paper No. 3.2-1, Proc.
Alloy Steel, Weld. J., Feb 2003, p 25s Petten Conference, North-Holland Publish-
24. R. De Santis, W. Dumini, and G. Casarini, ing, 1978
“Service Failure of Stressed Stabilized 37. G.Y. Lai and O.F. Kimball, “Aging Behavior
Stainless Steels at Elevated Temperatures,” of Alloy 800H and Associated Mechanical
WRC Bulletin No. 421, Welding Research Property Changes,” Report GA-A15194,
Council, 1997, p 17 General Atomic Company, San Diego, CA,
25. M. Chabaud, L. Allais, P. Dubuisson, Nov 1978
C. Hogrel, and A. Pineau, Ageing and 38. V. Coppolecchia, J. Bryant, F. Hofmann, and
Cracking of Simulated Heat Affected Zones K. Drefahl, Loss of Creep Ductility in Alloy
in 321 Stainless Steel, Mater. High Temp., 800H with High Levels of Titanium and
Vol 15 (No. 3/4), 1998, p 395 Aluminum, Performance of High Tempera-
26. L. Li and R.W. Messler, Stress Relaxation ture Materials in Fluidized Bed Combustion
Study of HAZ Reheat Cracking in Type 347 Systems and Process Industries, Conf. Proc.,
Stainless Steel, Weld. J., 2000, p 137s P. Ganesan and R.A. Bradley, Ed., ASM
27. R.J. Christoffel, Notch-Rupture Strength International, 1987, p 201
of Type 347 HAZ, Weld. J., July 1960, 39. P.G. Stone, J. Orr, and J.C. Guest, A Status
p 315 Review of Alloy 800, Proc. British Nuclear
28. R.J. Truman and H.W. Kirkby, Some Duc- Energy Society Conference, University of
tility Aspects of 18-12-1Nb Steel, JISI, Reading, Sept 25–26, 1974, S.F. Pugh, Ed.,
1960, p 180 AERE Harwell, UK, 1975, p 15
29. N.E. Moore and J.A. Griffiths, Micro- 40. A.B. Smith, Characterization of Materials
structural Causes of HAZ Cracking in for Service at Elevated Temperatures,
Name ///sr-nova/Dclabs_wip/High Temp/5208_379-408.pdf/Chap_14/ 31/10/2007 4:07PM Plate # 0 pg 408

408 / High-Temperature Corrosion and Materials Applications

MPC-7, G.V. Smith, Ed., ASME, 1978, C.T. Liu, and N.S. Stoloff, Ed., Materials
p 159 Research Society, 1985, p 485
41. W.G. Lipscomb, J.R. Crum, and P. Ganesan, 49. “1997 Findings and Recommendations,”
Mechanical Properties and Corrosion Re- Report to The U.S. Congress and The
sistance of INCONEL Alloy 617 for Refin- Secretary of Energy, U.S. Nuclear Waste
ery Services, Paper No. 259, Corrosion/89, Technical Review Board, Arlington, VA,
NACE, 1989 April 1998
42. T.H. Bassford and T.V. Schill, “A Review of 50. R.B. Rebak and N.E. Koon, Localized
INCONEL Alloy 617 and Its Properties Corrosion Resistance of High Nickel Alloys
after Long-Time Exposure to Intermediate as Candidate Materials for Nuclear Waste
Temperatures,” Special Metals, Inc., Hun- Repository—Effect of Alloy and Weldment
tington, WV Aging at 427 °C for up to 40,000 H, Paper
43. H. Stahl, G. Smith, and S. Wastiaux, Strain- No. 153, Corrosion/98, NACE Inter-
Age Cracking of Alloy 601 Tubes at 600 °C, national, 1998
Practical Failure Analysis, Vol 1 (No. 1), 51. T.S.E. Summers, M.A. Wall, M. Kumar,
Feb 2001 S.J. Matthews, and R.B. Rebak, Phase
44. “INCONEL Alloy 601,” Special Metals Stability and Mechanical Properties of C-22
Product Literature, Special Metals Com- Alloy Aged in the Temperature Range 590 to
pany, Huntington, WV 760 C for 16,000 Hours, Scientific Basis for
45. M.D. Rowe, Ranking the Resistance of Nuclear Waste Management XXII, (Sym-
Wrought Superalloys to Strain-Age Crack- posium Proc.), Vol 556, D.J. Wronkiewicz
ing, Weld. J., Feb 2006, p 27s and J.H. Lee, Ed., Materials Research
46. R.W. Fawley, M. Prager, J.B. Carlton, and Society, 1999, p 919
G. Sines, “Recent Studies of Cracking 52. R.B. Bebak, T.S.E. Summers, and R.M.
During Postwelding Heat Treatment of Carranza, Mechanical Properties, Micro-
Nickel-Base Alloys,” WRC Bulletin No. structure and Corrosion Performance of
150, Welding Research Council, 1970 C-22 Alloy Aged at 260 to 800 °C, Scientific
47. H.M. Tawancy, Order-Strengthening in a Basis for Nuclear Waste Management XXI
Nickel-Base Superalloy (Hastelloy Alloy S), (Symposium Proc.), Vol 608, R.W. Smith
Metall. Trans., Vol 11A, 1980, p 1764 and D.W. Shoesmith, Ed., Materials Re-
48. H.M. Tawancy and A.I. Asphahani, Order- search Society, 2000, p 109
ing Behavior and Corrosion Properties of 53. L.M. Pike and D.L. Klarstrom, A New
Ni-Mo and Ni-Mo-Cr Alloys, High- Corrosion-Resistant Ni-Cr-Mo Alloy with
Temperature Ordered Intermetallic Alloys High Strength, Paper No. 04239, Corrosion/
(Symposium Proc.), Vol 39, C.C. Koch, 2004, NACE International, 2004
Name ///sr-nova/Dclabs_wip/High Temp/5208_409-421.pdf/Chap_15/ 26/10/2007 12:42PM Plate # 0 pg 409

High-Temperature Corrosion And Materials Applications Copyright © 2007 ASM International®

George Y. Lai, editor, p409-421 All rights reserved.
DOI: 10.1361/hcma2007p409 www.asminternational.org

CHAPTER 15

Molten Salt Corrosion

15.1 Introduction metal surface. The corrosion reaction proceeds

primarily by oxidation, which is then followed
Molten salt technology plays an important role by dissolution of metal oxides in the melt. Oxy-
in various industries. In the heat treating industry, gen and water vapor in the molten salt thus often
molten salts are commonly used as a medium for accelerate molten salt corrosion.
heat treatment of metals and alloys (e.g., Corrosion can also take place through mass
annealing, tempering, hardening, quenching, and transfer due to thermal gradient in the melt. This
cleaning) as well as for surface treatment (e.g., mode of corrosion involves dissolution of an
case hardening). In nuclear and solar energy alloying element at hot spots and deposition
systems, they have been used as a medium for of that alloying element at cooler spots. This
heat transfer and energy storage. Other applica- can result in severe fouling and plugging in a
tions include extraction of aluminum, magne- circulating system. Corrosion is also strongly
sium, sodium, and other reactive metals; refining dependent on temperature and velocity of the
of refractory metals; and high-temperature bat- salt.
teries and fuel cells. Table 15.1 summarizes the Corrosion can take the form of uniform thin-
general applications of molten salt technology in ning, pitting, or internal or intergranular attack.
several industries (Ref 1). In general, molten salt corrosion is quite similar
The containment material, which is in contact to aqueous corrosion. More complete discussion
with the molten salt, is subject to molten salt cor- on the mechanisms of molten salt corrosion can
rosion. This chapter reviews the data relevant to be found in Ref 3 to 5.
the corrosion of containment materials. Although
the literature related to studies of corrosion in
molten salts is extensive, as can be seen in an an-
notated bibliography prepared by Janz and Tom- 15.3 Corrosion in Molten Chlorides
kins (Ref 2), corrosion data useful in selecting
materials are rather limited and fragmented. Chloride salts are widely used in the heat
treating industry for annealing and normalizing
of steels. These salts are commonly referred to as
15.2 Corrosion Process neutral salt baths. The most common neutral salt
baths are barium, sodium, and potassium chlor-
Molten salts generally are a good fluxing ides, used separately or in combination in the
agent, effectively removing oxide scales from a temperature range of 760 to 980 °C (1400 to
Table 15.1 General applications of molten salt technology in several industries
Power Metals/materials Chemicals
Solar/thermal: collection, Extraction: refractory metals, actinides, lanthanides, Fuels: cracking, catalysts
storage, transfer transition, and light metals
Nuclear: homogeneous Processing: heat treatment, annealing, quenching, cleaning, Plastics: curing, etching, vulcanizing
reactors, reprocessing cementation, electroforming
Batteries Surface finishing: anodizing, plating Pyrolysis: recycling, scrap treatment,
hazardous materials disposal
Fuel cells Joining: fluxes and slags for welding, brazing, soldering, and Synthesis: organics, gases
electroslag refining
Composites: glasses, ceramics, slags Special applications: liquid crystals,
single-crystal growing, matrix
Recycling
Source: Ref 1
Name ///sr-nova/Dclabs_wip/High Temp/5208_409-421.pdf/Chap_15/ 26/10/2007 12:42PM Plate # 0 pg 410

410 / High-Temperature Corrosion and Materials Applications

1800 °F). Compositions of some common neu- Thus, lowering carbon from 0.4% to about
tral salt baths are (Ref 6): 0.07% resulted in a threefold improvement.
Decreasing grain size also improved alloy resis-
 50NaCl-50KCl tance to intergranular attack. Five different neu-
 50KCl-50Na2CO3 tral salt baths were compared for HW, HT, and
 20NaCl-25KCl-55BaCl2 three Fe-Cr alloys, as shown in Fig. 15.3. In
 25NaCl-75BaCl2 general, the four chloride salt baths were quite
 21NaCl-31BaCl2-48CaCl2 similar. The KCl-Na2CO3 salt bath was sig-
Jackson and LaChance (Ref 6) performed an nificantly less aggressive than pure chloride
extensive study on the corrosion of cast Fe-Ni-Cr baths. It is also interesting to note that Fe-17Cr
alloys in the NaCl-KCl-BaCl2 salt bath. They alloy was better than HW (Fe-12Cr-60Ni) and
found that alloys suffered intergranular attack HT (Fe-15Cr-35Ni) alloys in NaCl-KCl, NaCl-
more than metal loss. Corrosion data in terms of KCl-BaCl2, and NaCl-BaCl2-CaCl2 salt baths.
metal loss and intergranular attack are shown in Lai et al. (Ref 7) evaluated various wrought
Fig. 15.1 and 15.2, respectively. The figures also iron-, nickel-, and cobalt-base alloys in a NaCl-
indicate that resistance to the molten salt (NaCl- KCl-BaCl2 salt bath at 840 °C (1550 °F) for
KCl-BaCl2) increases with decreasing chromium 1 month (Fig. 15.4). Surprisingly, two high-
and increasing nickel in Fe-Ni-Cr alloys. HW nickel alloys (alloys 600 and 601) suffered more
alloy (Fe-12Cr-60Ni) was consistently the best corrosion attack than stainless steels such as
performer among the four commercial cast alloys Types 304 and 310. Co-Ni-Cr-W, Fe-Ni-Co-Cr,
(HW, HT, HK, and HH alloys) studied. These and Ni-Cr-Fe-Mo alloys performed best.
authors further noted that intergranular attack Laboratory testing in a simple salt bath failed to
generally followed grain-boundary carbides. reveal the correlation between alloying elements

Fig. 15.1 Corrosion rates in terms of metal loss for four commercial cast Fe-Ni-Cr alloys in a 20NaCl-25KCl-55BaCl2 salt bath under
different conditions of rectification at 870 °C (1600 °F) for 60 h. Source: Ref 6
Name ///sr-nova/Dclabs_wip/High Temp/5208_409-421.pdf/Chap_15/ 26/10/2007 12:43PM Plate # 0 pg 411

Chapter 15: Molten Salt Corrosion / 411

Fig. 15.2 Corrosion rates in terms of intergranular attack for four commercial cast Fe-Ni-Cr alloys in a 20NaCl-25KCl-55BaCl2 salt bath
under different conditions of rectification at 870 °C (1600 °F) for 60 h. Source: Ref 6

Fig. 15.3 Comparison of different neutral salt baths for HW, HT, and Fe-Cr alloys at 870 °C (1600 °F) for 60 h. Source: Ref 6
Name ///sr-nova/Dclabs_wip/High Temp/5208_409-421.pdf/Chap_15/ 26/10/2007 12:43PM Plate # 0 pg 412

412 / High-Temperature Corrosion and Materials Applications

and performance. Tests were conducted at (Table 15.3). Fifteen alloys, including iron-,
840 °C (1550 °F) for 100 h in a NaCl salt bath nickel-, and cobalt-base alloys, were evaluated.
with fresh salt for each test run. Results are After 144 h of exposure, specimens of
tabulated in Table 15.2 (Ref 7, 8). Similar to the eight alloys were consumed. The remaining
field test results, Co-Ni-Cr-W and Fe-Ni-Co-Cr seven alloys disintegrated after 456 h of expo-
alloys performed best. sure. The authors concluded that the chloride salt
Evaluating a eutectic sodium-potassium-mag- was too aggressive to be used at 900 °C
nesium chloride (33NaCl-21.5KCl-45.5MgCl2, (1650 °F).
mol%) as a possible heat-transfer and energy- At lower temperatures, molten salts generally
storage medium in solar thermal energy systems become less aggressive. Susskind et al. (Ref 10)
for power generation, Coyle et al. (Ref 9) con- conducted corrosion tests at 450 to 500 °C (840
ducted corrosion tests on various commercial to 930 °F) in molten NaCl-KCl-MgCl2 eutectic
alloys at 900 °C (1650 °F) for 144 and 456 h and found many alloys resistant to molten salt
corrosion (Table 15.4). Low corrosion rates
may also be attributed to the vacuum environ-
ment used in these tests. Investigating the cor-
rosion behavior of alloys at 400 and 500 °C

Table 15.3 Results of corrosion tests in molten

eutectic NaCl-KCl-MgCl2 salt at 900 °C (1650 °F)
Weight change, mg/cm2
Alloy 144 h 456 h
304 Disintegrated …
316 Disintegrated …
800 Disintegrated …
800H −310 Disintegrated
556 −250 Disintegrated
Nickel Disintegrated …
600 −280 Disintegrated
214 −120 Disintegrated
X Disintegrated …
N Disintegrated …
S −400 Disintegrated
Fig. 15.4 Results of a field rack test in a NaCl-KCl-BaCl2 salt
230 −300 Disintegrated
bath at 840 °C (1550 °F) for 1 month. Source: Ref 7
X-750 Disintegrated …
R-41 −150 Disintegrated
188 Disintegrated …
Table 15.2 Results of laboratory tests in a NaCl N2-(0.1–1H2O)-(1–10O2) was used for the cover gas. Source: Ref 9
salt bath at 840 °C (1550 °F) for 100 h
Total depth of attack(a),
Alloy mm (mils)
188 0.051 (2.0) Table 15.4 Corrosion of alloys in molten
25 0.064 (2.5)
556 0.066 (2.6)
eutectic NaCl-KCl-MgCl2 salt at 450 to 500 °C
601 0.066 (2.6) (840 to 930 °F) with 50 °C temperature
Multimet 0.069 (2.7) differential under vacuum for 1000 h
150 0.076 (3-0)
Maximum penetration,
214 0.079 (3.1)
Alloy mm/yr (mpy)
304 0.081 (3.2)
446 0.081 (3.2) 1020 0
316 0.081 (3.2) 2.25Cr-lMo 0.08 (3)
X 0.097 (3.8) 304 <0.01 (0.4)
310 0.107 (4.2) 310 0
800H 0.109 (4.3) 316 <0.01 (0.4)
625 0.112 (4.4) 347 0.12 (4.7)
RA330 0.117 (4.6) 410 0.03 (1.0)
617 0.122 (4.8) 430 0.05 (2)
230 0.140 (5.5) 446 <0.01 (0.4)
S 0.168 (6.6) 600 0.05 (1.8)
RA330 0.191 (7.5) N 0.05 (2)
600 0.196 (7.7) Molybdenum 0
Tantalum 0.07 (2.9)
A fresh salt bath was used for each test run; air was used for the cover gas.
(a) Mainly intergranular attack; no metal wastage. Source: Ref 7 and 8 Source: Ref 10
Name ///sr-nova/Dclabs_wip/High Temp/5208_409-421.pdf/Chap_15/ 26/10/2007 12:43PM Plate # 0 pg 413

Chapter 15: Molten Salt Corrosion / 413

(750 and 930 °F) in the molten LiCl-KCl 1100 °F). It was not clear what type of cover
eutectic, which was being considered as an gas was involved in these tests. The results
electrolyte for lithium-sulfur fuel cells, Battles from both salt mixtures are summarized in
et al. (Ref 11) also found many alloys resistant Fig. 15.5 and 15.6. Steels and Fe-Cr alloys suf-
to molten salt (Table 15.5). Tests were con- fered severe corrosion in both types of salts.
ducted in closed quartz crucibles. All the alloys Chromium in Fe-Cr alloys and nickel in Fe-Ni
tested showed negligible corrosion rates. Alu- alloys improved performance. Fe-Cr-Ni alloys
minum in the aluminum-clad Type 434 SS performed significantly better than steels and
sample corroded at a higher rate due to the Fe-Cr alloys.
galvanic couple between aluminum and stain- Intergranular corrosion is the major corrosion
less steel (Ref 11). morphology by molten chloride salts. Fig-
Takehara and Ueshiba (Ref 12) investigated ures 15.7 and 15.8 show typical intergranular
the corrosion behavior of steel, Fe-Cr, Fe-Ni, and corrosion by molten chloride salt. Figure 15.7
Fe-Cr-Ni alloys in molten 20NaCl-30BaCl2- shows the intergranular attack of a Ni-Cr-Fe
50CaCl2 and molten 25LiCl-25ZnCl2-16BaCl2- alloy (alloy 600) coupon welded to a heat treat
24CaCl2-10NaCl at 500 and 600 °C (930 and basket that underwent heat treat cycles involving

Table 15.5 Corrosion rates of several metals and alloys in molten LiCl-KCl eutectic
Corrosion rate, µm/yr (mpy)
Annealed samples Sensitized samples(a)
Material 400 °C (750 °F) 500 °C (930 °F) 400 °C (750 °F) 500 °C (930 °F)
304 2 (0.08) 6 (0.24) 3 (0.12) 5 (0.2)
316 2 (0.08) … … …
347 … 2 (0.08) 1 (0.04) 1 (0.04)
430 2 (0.08) … … …
E-Brite 8 (0.32) 6 (0.24) 4 (0.16) 5 (0.2)
Al-Clad Type 434 130 (5.1) … … …
Iron (99.999% Fe) 41 (1.6) … … …
Armco electromagnet iron 12 (0.47) … … …
Tests were conducted in closed quartz crucibles. (a) Samples were sensitized at 650 °C (1200 °F) for 120 h. Source: Ref 11

Fig. 15.5 Corrosion rates of steel, Fe-Cr, Fe-Ni, and Fe-Cr-Ni alloys in molten 20NaCl-30BaCl2-50CaCl2 at 500 and 600 °C (930 and
1110 °F). Source: Ref 12
Name ///sr-nova/Dclabs_wip/High Temp/5208_409-421.pdf/Chap_15/ 26/10/2007 12:43PM Plate # 0 pg 414

414 / High-Temperature Corrosion and Materials Applications

a molten KCl salt bath at 870 °C (1600 °F) and a Cold work may significantly affect corrosion
quenching salt bath of molten sodium nitrate and of alloys in molten chloride salts. Lai (Ref 14)
sodium nitrite at 430 °C (800 °F) for 1 month discovered severe intergranular corrosion at the
(Ref 13). Figure 15.8 shows the intergranular sheared edge of a Ni-Cr-Fe-Mo alloy (alloy X)
attack of a heat treat basket made of the same coupon after exposure in a molten CaCl2-NaCl
alloy after service for 6 months in the same heat salt bath at 570 °C (1050 °F). The sample edge
treat cycling operation (Ref 13).
Another frequently observed corrosion mor-
phology is internal attack by void formation
(Fig. 15.9) (Ref 13). Voids tend to form at grain
boundaries as well as in the grain interior (Ref 4).
The continuing formation and growth of chro-
mium compounds at the metal surface causes
outward migration of chromium and inward
migration of vacancies, thus leading to internal
void formation (Ref 4).

Fig. 15.6 Corrosion rates of steel, Fe-Cr, Fe-Ni, and Fe-Cr-

Ni alloys in molten 25LiCl-25ZnCl2-16BaCl2-
24CaCl2-10NaCl at 500 and 600 °C (930 and 1110 °F). Source:
Ref 12

Fig. 15.7 Intergranular attack of a Ni-Cr-Fe alloy coupon

welded to a heat treat basket after service for
1 month in a heat treat operation cycling between a molten KCl Fig. 15.8 Intergranular attack of a Ni-Cr-Fe alloy heat treat
bath at 870 °C (1600 °F) and a quenching salt bath of molten basket after service for 6 months in the same heat
sodium nitrate-nitrite at 430 °C (800 °F). Source: Ref 13 treat cycling operation described in Fig. 15.7. Source: Ref 13
Name ///sr-nova/Dclabs_wip/High Temp/5208_409-421.pdf/Chap_15/ 26/10/2007 12:43PM Plate # 0 pg 415

Chapter 15: Molten Salt Corrosion / 415

was cold worked during sample shearing prior to molten NaNO3-KNO3 salt. Carbon steel and
exposure. There was no evidence of intergranular 2.25Cr-1Mo steel exhibited low corrosion rates
corrosion in the area away from the sheared edge (<0.13 mm/yr, or <5 mpy) at 460 °C (860 °F).
(Fig. 15.10). The prior cold work also resulted in At 500 °C (932 °F), 2.25Cr-1Mo steel exhibited
a thicker oxide scale during exposure in the a corrosion rate of about 0.026 mm/yr (1 mpy).
molten salt. A Co-Ni-Cr-W alloy (alloy 188) Aluminized Cr-Mo steel showed higher resis-
exhibited similar behavior. tance, with a corrosion rate of less than
0.004 mm/yr (<0.2 mpy) at 600 °C (1110 °F).
Austenitic stainless steels, alloy 800, and alloy
600 were more resistant than carbon steel and
15.4 Corrosion in Molten Nitrates/ Cr-Mo steels. Nickel, however, suffered high
Nitrides corrosion rates.
Slusser et al. (Ref 23) evaluated the cor-
Molten nitrates or nitrate-nitride mixtures are
rosion behavior of a variety of alloys in molten
widely used for heat treat salt baths, typically
NaNO3-KNO3 (equimolar volume) salt with an
operating from 160 to 590 °C (325 to 1100 °F).
equilibrium nitrite concentration (about 6 to
They are also used as a medium for heat transfer
12 wt%) at 675 °C (1250 °F) for 336 h. A
or energy storage. Molten drawsalt (NaNO3-
KNO3) is being considered as a heat-transfer
and energy-storage medium for a solar central
receiver for power generation from solar energy.
Numerous studies (Ref 15–21) have been
carried out to determine potential candidate
containment materials for handling molten
drawsalt. Bradshaw and Carling (Ref 22)
recently summarized these studies as well as the
results of their study (Table 15.6) (Ref 22). The
data suggest that, for temperatures up to 630 °C
(1170 °F), many alloys are adequate for handling

Fig. 15.10 Corrosion of Ni-Cr-Fe-Mo alloy in a molten

CaCl2-NaCl salt bath at 570 °C (1050 °F) for
Fig. 15.9 Corrosion attack consisting of voids in a nickel-base 6 months. (a) Little corrosion at the area not cold worked (away
alloy after 2 months at 870 °C (1600 °F) in a molten from the sheared edge). (b) Intergranular attack at the cold worked
BaCl2 salt bath. Source: Ref 13 area (specimen sheared edge). Source: Ref 14
Name ///sr-nova/Dclabs_wip/High Temp/5208_409-421.pdf/Chap_15/ 26/10/2007 12:43PM Plate # 0 pg 416

416 / High-Temperature Corrosion and Materials Applications

constant purge of air in the melt was main- resistant to molten NaOH (Ref 26–29), particu-
tained during testing. Nickel-base alloys were larly low-carbon nickel such as Ni 201 (Ref 30).
generally much more resistant than iron-base Gregory et al. (Ref 29) reported corrosion rates of
alloys. Increasing nickel content improved several nickel-base alloys obtained from
alloy corrosion resistance to molten nitrate- static tests at 400 to 680 °C (750 to 1256 °F)
nitrite salt. However, pure nickel suffered rapid (Table 15.8). Molybdenum and silicon appear to
corrosion attack. Figure 15.11 shows the cor- be detrimental alloying elements in molten
rosion rates of various alloys as a function of NaOH salt. Iron may also be detrimental.
nickel content (Ref 23). Silicon-containing Molybdenum and iron were found to be selec-
alloys, such as RA330 and Nicrofer 3718, tively removed from nickel-base alloys with less
performed poorly. A long-term test (1920 h) at than 90% nickel, leading to the formation of
675 °C (1250 °F) was performed on selected internal voids (Ref 31).
alloys, showing corrosion rates similar to Molten sodium hydroxide becomes increas-
those obtained from 336 h exposure tests ingly aggressive with increasing temperature.
(Table 15.7). Alloy 800, however, exhibited a Coyle et al. (Ref 9) evaluated a variety of
higher corrosion rate in the 1920 h test than in alloys for a possible containment material for
the 336 h test. As the temperature was molten sodium hydroxide operating at 900 °C
increased to 700 °C (1300 °F), corrosion rates (1650 °F) for a solar power generation system.
became much higher, particularly for iron-base Test results are tabulated in Table 15.9. Many
alloy 800, which suffered an unacceptably high
corrosion rates (Table 15.7). Boehme and
Bradshaw (Ref 24) attributed the increased
corrosion rate with increasing temperature to
higher alkali oxide concentration. Slusser et al.
(Ref 23) found that adding sodium peroxide
(Na2O2) to the salt increased the salt cor-
rosivity.

15.5 Corrosion in Molten Sodium

Hydroxide (Caustic Soda)

The reaction of metals with molten sodium

hydroxide (NaOH) leads to metal oxide, sodium
oxide, and hydrogen (Ref 25). Nickel is most

Table 15.6 Corrosion rates of selected

metals and alloys in molten NaNO3-KNO3
Temperature Corrosion rate,
Alloy °C (°F) mm/yr (mpy)
Carbon steel 460 (860) 0.120 (4.7) Fig. 15.11 Corrosion rates of various alloys as a function of
nickel content in the alloy tested in molten
2.25Cr-lMo 460 (860) 0.101 (4.0)
NaNO3-KNO3 salt at 675 °C (1250 °F). Source: Ref 23
500 (932) 0.026 (1.0)
9Cr-lMo 550 (1020) 0.006 (0.2)
600 (1110) 0.023 (0.9)
Aluminized Cr-Mo steel 600 (1110) <0.004 (0.2) Table 15.7 Corrosion rates of selected alloys
12Cr steel 600 (1110) 0.022 (0.9)
304SS 600 (1110) 0.012 (0.5) at 675 and 700 °C (1250 and 1300 °F) in
316SS 600 (1110) 0.007–0.010 (0.3–0.4) sodium-potassium nitrate-nitrite salt
630 (1170) 0.106 (4.2) Corrosion rate, mm/yr (mpy)
800 565 (1050) 0.005 (0.2)
600 (1110) 0.006–0.01 (0.2–0.4) 675 °C (1250 °F) 700 °C (1300 °F)
Alloy 1920 h 720 h
630 (1170) 0.075 (3.0)
600 600 (1110) 0.007–0.01 (0.3–0.4) 214 0.41 (16) 0.53 (21)
630 (1170) 0.106 (4.2) 600 0.25 (10) 0.99 (39)
Nickel 565 (1050) >0.5 (20) N 0.23 (9.1) 1.22 (48)
Titanium 565 (1050) 0.04 (1.6) 601 0.48 (19) 1.25 (49)
Aluminum 565 (1050) <0.004 (0.2) 800 1.85 (73) 6.6 (259)
Source: Ref 22 Source: Ref 23
Name ///sr-nova/Dclabs_wip/High Temp/5208_409-421.pdf/Chap_15/ 26/10/2007 12:43PM Plate # 0 pg 417

Chapter 15: Molten Salt Corrosion / 417

iron-, nickel-, and cobalt-base alloys disin- conditions to about 8 mm/yr (320 mpy) at a
tegrated in 84 h. Samples of the alloys that sur- rotational speed of 600 rpm (Ref 32).
vived the 84 h exposure test were severely Metals or alloys in molten sodium hydroxide
corroded. Scales that formed on these samples are susceptible to mass transfer due to thermal
were reportedly cracked and spalled. The weight- gradients in the melt. This causes corrosion in the
gain or weight-loss data of surviving samples hot zone, and potential tube plugging in the
were no longer indicative of alloy performance cold zone, of a circulating system. For example,
ranking. No metallographic examination was 6.35 mm (0.25 in.) nickel tubing was plugged
performed on these samples. The authors con- after 5000 h at 440 to 480 °C (830 to 900 °F),
cluded that no further studies on molten sodium and after 50 h at 690 to 730 °C (1280 to 1350 °F)
hydroxide were necessary, because the salt was (Ref 27).
too aggressive to metallic materials operating at
900 °C (1650 °F). The marked influence of
temperature on the corrosiveness of molten 15.6 Corrosion in Molten Fluorides
sodium hydroxide is also demonstrated by the
results shown in Table 15.10 (Ref 26). Corrosion of alloys in molten fluoride salts has
Corrosion of metals and alloys in molten been extensively studied for nuclear reactor
NaOH depends strongly on the velocity of the applications. The molten salt nuclear reactor uses
salt. Gregory et al. (Ref 32) showed that corro- a LiF-BeF2 base salt as a fuel salt, containing
sion of nickel under dynamic conditions was various amounts of UF4, ThF4, and ZrF4
enhanced by as much as several times at 540 °C (Ref 33). The reactor coolant salt is a NaBF4-NaF
(1000 °F) and higher. The corrosion rate for mixture (Ref 33). A nickel-base alloy, Hastelloy
nickel at 680 °C (1250 °F), for example, varied alloy N, has proved to be the most corrosion
from about 1 mm/yr (40 mpy) under static resistant in molten fluoride salts (Ref 34). The

Table 15.8 Corrosion rates of selected nickel-base alloys obtained from static tests in molten
sodium hydroxide
Corrosion rate, mm/yr (mpy)
Alloy 400 °C (750 °F) 500 °C (930 °F) 580 °C (1080 °F) 680 °C (1260 °F)
Ni-201 0.023 (0.9) 0.033 (1.3) 0.06 (2.5) 0.96 (37.8)
C … 2.54 (100) (a) …
D 0.018 (0.7) 0.056 (2.2) 0.25 (9.9) (a)
400 0.046 (1.8) 0.13 (5.1) 0.45 (17.6) …
600 0.028 (1.1) 0.06 (2.4) 0.13 (5.1) 1.69 (66.4)
301SS 0.043 (1.7) 0.08 (3.2) 0.26 (10.4) 1.03 (40.7)
75 0.028 (1.1) 0.36 (14.3) 0.53 (20.8) 1.21 (47.6)
(a) Severe corrosion. Source: Ref 29

Table 15.10 Corrosion of various metals

and alloys in molten sodium hydroxide
Table 15.9 Results of corrosion tests in molten Weight change rate(a), mg/cm2-day
sodium hydroxide at 900 °C (1650 °F) for 84 h 538 °C 816 °C
Material (1000 °F) (1500 °F)
Alloy Weight change, mg/cm2
Ni-201 +0.12 +1.7
304 Disintegrated Copper +1.54 −2.2
316 Disintegrated Chromium −0.12 −40
800 +60
800H +65
Aluminum −0.34 …
556 Disintegrated Silver +0.02 +2.9
Nickel −50 Gold −0.33 Broke
600 −27 Platinum −0.32 Broke
214 +160 70Au-30Pt +0.36 Broke
X +22 Palladium +6.59 +175
N Fractured Colmonoy No. 5 +0.63 +70
S 4
230 Disintegrated
Colmonoy No. 6 −0.48 −72
X-750 +35 Colmonoy No. 9 −0.23 Dissolved in 24 h
R-41 Disintegrated Chromel P +0.09 Dissolved in 24 h
188 Disintegrated Zirconium +0.56 …
N2-(0.1–1H2O)-(1–10O2) was used for the cover gas. Source: Ref 9 (a) Based on 24 h exposure tests. Source: Ref 26
Name ///sr-nova/Dclabs_wip/High Temp/5208_409-421.pdf/Chap_15/ 26/10/2007 12:43PM Plate # 0 pg 418

418 / High-Temperature Corrosion and Materials Applications

alloy was the primary containment material Molten fluorides are generally used in a closed
for a molten salt test reactor successfully oper- system under vacuum or an inert atmosphere.
ated from 1965 to 1969 (Ref 35). Koger However, hydrogen fluoride may be present
(Ref 33) reported a corrosion rate of less than in the system, resulting in increased corrosion
0.0025 mm/yr (0.1 mpy) at 704 °C (1300 °F) in rates. Moisture, a common impurity in fluoride
the LiF-BeF2 base salt (fuel salt) and about salts, can react with fluorides to produce gaseous
0.015 mm/yr (0.6 mpy) at 607 °C (1125 °F) in HF (Ref 39), some of which may dissolve in the
the NaBF4-NaF coolant salt for alloy N. melt. Corrosion by HF will also be involved,
Iwamoto et al. (Ref 36) performed corrosion leading to production of hydrogen (Ref 39).
tests in eutectic LiF-NaF-KF salt in a test loop Therefore, it is important to reduce the moisture
with a 750 °C (1380 °F) hot leg and a 685 °C level in the salt to reduce corrosion attack
(1265 °F) cold leg for 500 h. Alloy N exhibited (Ref 39).
only 2.06 mg/cm2 of maximum weight loss
at the hot leg.
At lower temperatures, austenitic stainless
steels showed good performance. Type 316 15.7 Corrosion in Molten Carbonates
exhibited 0.015 mm/yr (0.59 mpy) at 650 °C
(1200 °F) in 66LiF-34BeF2 (mol%), and about Molten carbonates are generally less corrosive
0.002 mm/yr (0.08 mpy) at 530 °C (986 °F) in than molten chlorides or hydroxides. Coyle et al.
22LiF-31LiCl-47LiBr (mol%) (Ref 37). In a LiF- (Ref 9) evaluated three different molten salts for
BeF2 fuel salt containing UF4, ThF4, and ZrF4, a possible heat-transfer and energy-storage
Type 304 suffered a corrosion rate of only about medium capable of operating 900 °C (1650 °F)
0.028 mm/yr (1.1 mpy) at 690 °C (1270 °F) for a solar power generation system. Both the
(Ref 33). eutectic sodium-potassium-magnesium chloride
Corrosion can become more aggressive as (33NaCl-21.5KCl-45.5MgCl2, mol%) and the
temperature increases. It is particularly severe sodium hydroxide were found to be too corro-
for stainless steels because of tube-plugging sive for many commercial alloys. The eutectic
problems due to mass transfer. Adamson et al. sodium-potassium carbonate (58Na2CO3-
(Ref 38) conducted corrosion tests in a thermal 42K2CO3, mol%), on the other hand, showed
convection loop involving 43.5KF-10.9NaF- promise because of the much lower corrosion
44.5LiF-1.1UF4 (mol%) with an 815 °C rates exhibited by many commercial alloys.
(1500 °F) hot leg and a 704 °C (1300 °F) cold Corrosion data generated in molten carbonate
leg. Types 410, 430, 316, 310, and 347 suffered salt at 900 °C (1650 °F) for 504 h are
severe tube-plugging problems at the cold leg summarized in Table 15.12. The best performer
within short test durations. Nickel and nickel- was Ni-Cr-Fe-Mo alloy (alloy X), followed
base alloys, on the other hand, showed no plug- by Ni-Cr-Fe-Al alloy (alloy 214) and Co-
ging even after 500 h of testing. However, these Ni-Cr-W alloy (alloy 188). The Ni-Cr-Mo
alloys suffered corrosion at the hot leg after
500 h of exposure. Alloy 600 suffered internal Table 15.11 Results of corrosion tests in
attack consisting of voids about 0.30 to 0.38 mm LiF-19.5CaF2 at 797 °C (1467 °F) for 500 h
(12 to 15 mils) deep. Nimonic alloy 75 suffered Depth of attack, µm (mils)
intergranular pitting about 0.20 to 0.33 mm (8 to Alloy General(a) Grain boundary(b)
13 mils) deep, and nickel suffered even metal Mild steel … 155 (6.1)
removal of about 0.23 mm (9 mils). 304 … 185 (7.3)
310 … 130 (5.1)
Misra and Whittenberger (Ref 39) reported 316 … 165 (6.5)
corrosion data for a variety of commercial alloys RA330 … 270 (10.6)
in molten LiF-19.5CaF2, which was being con- B 30 (1.2) …
N 15 (0.6) 15 (0.6)
sidered for a heat-storage medium in an advanced S 90 (3.5) …
solar space power system, at 797 °C (1467 °F) X … 140 (5.5)
for 500 h. The tests were conducted in alumina 600 90 (3.5) 30 (1.2)
718 45 (1.8) 120 (4.7)
crucibles with argon as a cover gas. Results are 75 30 (1.2) 135 (5.3)
tabulated in Table 15.11. For nickel-base alloys, 25 … 95 (3.7)
188 … 105 (4.1)
chromium was detrimental. No influence of
Tests were conducted in alumina crucibles under argon. (a) Intragranular voids
chromium, however, was noted on iron-base near surface. (b) Intergranular voids. Source: Ref 39
alloys.
Name ///sr-nova/Dclabs_wip/High Temp/5208_409-421.pdf/Chap_15/ 26/10/2007 12:43PM Plate # 0 pg 419

Chapter 15: Molten Salt Corrosion / 419

alloy (alloy S) was severely corroded. No general, corrosion rates of metals and alloys are
evidence of a systematic trend for the correlation strongly dependent on temperature and can
between alloying elements and performance is generally be reduced by decreasing the tem-
noted. A significant difference in corrosion was perature. Reducing oxidizing impurities, such as
observed between two samples of alloy 800 oxygen and water vapor, in the melt can also
obtained from different suppliers. However, two significantly reduce the corrosiveness of the
samples of alloy 600 showed good agreement. molten salt. Thermal gradients in the melt, in the
At lower temperatures, molten carbonate cor- case of circulating systems, may cause dissolu-
rosion generally becomes less severe. In molten tion of an alloying element at the hot leg and
eutectic alkali metal carbonate, Grantham et al. deposition of that element at the cold leg, leading
(Ref 40) found that many commercial alloys, to potential tube-plugging problems.
including Types 304L, 310, and 347, and alloys Corrosion data generated from different
600, C, N, X, and 25, exhibited low corrosion chloride salts at temperatures ranging from 400
rates (about 0.01mg/cm2/h or less) at 600 °C to 900 °C (750 to 1650 °F) for various cast and
(1110 °F). Nonoxidizing N2 was used for the wrought alloys were presented. The data are
cover gas in their corrosion tests, which may rather limited and fail to show a definitive cor-
also have contributed to low corrosion rates. relation between alloying elements and perfor-
The temperature dependence of corrosion rate mance. Intergranular attack is the major
can also be seen in the results generated by corrosion morphology in molten chlorides. Cold
Grantham and Ferry (Ref 41) in the eutectic work may significantly accelerate intergranular
Li2CO3-Na2CO3-K2CO3 (about equal weight) at attack. If the cold work effect is confirmed to be a
500, 600, and 700 °C (930, 1110, and 1290 °F). general phenomenon in molten chloride salts,
The cover gas in this case was not reported. annealing of fabricated components prior to
Corrosion rates were found to be less than 0.025, service is recommended.
0.025, and 2.54 mm/yr (1, 1, and 100 mpy) at The corrosion behavior of commercial alloys
500, 600, and 700 °C (930, 1110, and 1290 °F), in molten NaNO3-KNO3 is reasonably well
respectively. understood. For applications at temperatures up
to 630 °C (1170 °F), many commercial alloys,
including austenitic stainless steels, are capable
15.8 Summary of handling the molten salt. At higher tempera-
tures, nickel-base alloys are preferred because of
The corrosion behavior of alloys in molten the increased salt corrosivity. The resistance of
chlorides, nitrates/nitrites, sodium hydroxide, alloys to molten NaNO3-KNO3 improves with
fluorides, and carbonates was reviewed. In increasing nickel content. Nickel, however, per-
formed poorly.
Table 15.12 Results of corrosion tests in molten Corrosion data related to molten sodium
eutectic sodium-potassium carbonate at 900 °C hydroxide (caustic soda) are rather limited. At
(1650 °F) for 504 h 900 °C (1650 °F), the salt is probably too cor-
Total depth of attack(a), rosive for metallic materials to handle. Metals
Alloy mm (mils)
and alloys may have adequate resistance to
X 0.12 (4.7)
214 0.19 (7.5) molten sodium hydroxide at 680 °C (1260 °F)
188 0.22 (8.7) and lower temperatures. Nickel, particularly
556 0.26 (10.2)
X-750 0.27 (10.6)
low-carbon nickel such as Ni201, is most resis-
600(b) 0.34 (13.4) tant to the molten salt. Corrosion rates can be
600(b) 0.44 (17.3) significantly increased under dynamic condi-
R-41 0.42 (16.5)
N 0.51 (20.1) tions.
304SS 0.54 (21.3) Corrosion of alloys in molten fluorides has
316SS 0.63 (24.8) been extensively studied for nuclear reactor ap-
230 0.77 (30.3)
Nickel >0.30 (11.8) plications. A nickel-base alloy, Hastelloy alloy N,
800(b) 0.25 (9.8) has been found to be most suitable in molten
800(b)
S
>0.8 (31.5)
>1.43 (56.3)
fluoride salts. Since molten fluorides are used in
a circulating system with temperature differ-
N2-0.1CO2-(1–10O2) was used for the cover gas. (a) All alloys showed metal loss
only except nickel, which suffered 0.2 mm (7.9 mils) metal loss and more than entials, mass transfer induced corrosion due to
0.11 mm (4.3 mils) intergranular attack. (b) Two samples from two different
suppliers. Source: Ref 9
thermal gradients in the loop becomes an impor-
tant issue. It creates the potential for tube fouling
Name ///sr-nova/Dclabs_wip/High Temp/5208_409-421.pdf/Chap_15/ 26/10/2007 12:43PM Plate # 0 pg 420

420 / High-Temperature Corrosion and Materials Applications

and plugging at the cold leg of the loop. 81-8210, Sandia Laboratory, Livermore,
Stainless steels are generally worse than nickel CA, Feb 1982
and nickel-base alloys in terms of tube-plugging 16. P.F. Tortorelli and J.E. DeVan, “Thermal
problems. Convection Loop Study of the Corrosion of
Molten carbonates are generally less corrosive Fe-Ni-Cr Alloys by Molten NaNO3-KNO3,”
than molten chlorides or hydroxides. Some of the ORNL TM-8298, Oak Ridge National
alloys that perform best in molten eutectic Laboratory, Oak Ridge, TN, Dec 1982
Na2CO3-K2CO3 at high temperatures such as 17. R.W. Bradshaw, “A Thermal Convection
900 °C (1650 °F) include Ni-Cr-Fe-Mo alloy Loop Study of Corrosion of Alloy 800 in
(alloy X), Ni-Cr-Fe-Al alloy (alloy 214), and Molten NaNO3-KNO3,” SAND 82-8911,
Co-Ni-Cr-W alloy (alloy 188). Sandia Laboratory, Livermore, CA, Jan
1983
18. R.W. Bradshaw, “Kinetic Oxidation and
Elemental Depletion of Austenitic and Fer-
REFERENCES ritic Steels in Molten Nitrate Salt,” SAND
87-8011, Sandia Laboratory, Livermore,
1. D.G. Lovering, in Molten Salt Technology, CA, 1987
D.G. Lovering, Ed., Plenum Press, 1982, p 1 19. R.W. Bradshaw, “Oxidation of Chromium-
2. G.J. Janz and R.P.T. Tomkins, Corrosion, Molybdenum Steels by Molten Sodium
Vol 35 (No. 11), 1979, p 485 Nitrate-Potassium Nitrate,” SAND 87-8012,
3. J.W. Koger, in Corrosion, Vol 13, 9th ed., Sandia Laboratory, Livermore, CA 1987
Metals Handbook, ASM International, 20. R.W. Carling, R.W. Bradshaw, and R.W.
1987, p 51 Mar, J. Mater. Energy Sys., Vol 4 (No. 4),
4. J.W. Koger and S.L. Pohlman, in Corrosion, 1983, p 229
Vol 13, 9th ed., Metals Handbook, ASM 21. R.W. Bradshaw, “Oxidation and Chromium
International, 1987, p 88 Depletion of Alloy 800 and Type 316SS in
5. A. Rahmel, in Molten Salt Technology, D.G. Molten NaNO3-KNO3 at Temperatures
Lovering, Ed., Plenum Press, 1982, p 265 above 600 °C,” SAND 86-9009, Sandia
6. J.H. Jackson and M.H. LaChance, Trans. Laboratory, Livermore, CA, Jan 1987
ASM, Vol 46, 1954, p 157 22. R.W. Bradshaw and R.W. Carling, “A
7. G.Y. Lai, M.F. Rothman, and D.E. Fluck, Review of the Chemical and Physical
Paper No. 14, Corrosion/85, NACE, 1985 Properties of Molten Alkali Nitrate Salts
8. G.Y. Lai, unpublished results, Haynes and Their Effect on Materials Used for Solar
International, Inc., 1986 Central Receivers,” SAND 87-8005, Sandia
9. R.T. Coyle, T.M. Thomas, and G.Y. Lai, in Laboratory, Livermore, CA, April 1987
High Temperature Corrosion in Energy 23. J.W. Slusser, J.B. Titcomb, M.T. Heffelfin-
Systems, M.F. Rothman, Ed., The Metallur- ger, and B.R. Dunbobbin, J. Met., July 1985,
gical Society of AIME, 1985, p 627 p 24
10. H. Susskind, F.B. Hill, L. Green, S. Kalish, 24. D.R. Boehme and R.W. Bradshaw, High
L. Kukacka, W.E. McNulty, and E. Wirsing, Temp. Sci., Vol 18, 1984, p 39
Chem. Eng. Prog., Vol 56 (No. 3), 1960, p 57 25. G.P. Smith, “Corrosion of Materials in Fused
11. J.E. Battles, F.C. Mrazek, W.D. Tuohig, and Hydroxides,” USAEC Report ORNL-2048,
K.M. Myles, in Corrosion Problems in March 1956
Energy Conversion and Generation, C.S. 26. C.M. Craighead, L.A. Smith, and R.I. Jaffee,
Tedmon, Jr., Ed., The Electrochemical “Screening Tests on Metals and Alloys in
Society, 1974, p 20 Contact with Sodium Hydroxide at 1000 and
12. K. Takehara and T. Ueshiba, J. Soc. Mater. 1500 °F,” USAEC Report BMI-705, Nov
Sci. Jpn., Vol 179, 1968, p 755 1951
13. S.K. Srivastava, unpublished results, 27. E.M. Simmons, N.E. Miller, J.H. Stang, and
Haynes International, Inc., 1989 C. Weaver, “Corrosion and Components
14. G.Y. Lai, unpublished results, Haynes Studies on Systems Containing Fused
International, Inc., 1989 NaOH,” USAEC Report BMI-1118, July
15. R.W. Bradshaw, “Thermal Convection Loop 1956
Corrosion Tests of Type 316SS and Alloy 28. R.A. Lad and S.L. Simon, A Study of Cor-
800 in Molten Nitrate Salts,” SAND rosion and Mass Transfer of Nickel by
Name ///sr-nova/Dclabs_wip/High Temp/5208_409-421.pdf/Chap_15/ 26/10/2007 12:43PM Plate # 0 pg 421

Chapter 15: Molten Salt Corrosion / 421

Molten Sodium Hydroxide, Corrosion, Vol J.P. Pemsler et al., Ed., The Electrochemical
10 (No. 12), 1954, p 435 Society, 1976, p 315
29. J.N. Gregory, N. Hodge, and J.V.G. Iredale, 36. N. Iwamoto, Y. Makino, K. Furukawa,
“The Static Corrosion of Nickel and Other Y. Katoh, and H. Katsuta, Trans. JWRI, Vol 9
Materials in Molten Caustic Soda,” AERE- (No. 2), 1980, p 117
C/M-272, Atomic Energy Research Estab- 37. J.R. Keiser, J.H. DeVan, and E.J. Lawrence,
lishment, Harwell, U.K., March 1956 J. Nucl. Mater., Vol 85/86, 1979, p 295
30. R.R. Miller, “Thermal Properties of Sodium 38. G.M. Adamson, R.S. Crouse, and W.D.
Hydroxide and Lithium Metal,” Quarterly Manly, “Interim Report on Corrosion by
Progress Report May 1–Aug 1, 1952, NRL- Alkali-Metal Fluorides,” ORNL-2337, Oak
3230-201/52, 1952 Ridge National Laboratory, Oak Ridge, TN,
31. G.P. Smith and E.E. Hoffman, Corrosion, 1959
Vol 13, 1957, p 627t 39. A.K. Misra and J.D. Whittenberger,
32. J.N. Gregory, N. Hodge, and J.V.G. Iredale, “Fluoride Salts and Container materials for
“The Corrosion and Erosion of Nickel by Thermal Energy Storage Applications in
Molten Caustic Soda and Sodium Uranate Temperature Range 973 to 1400 °K,” NASA
Suspensions under Dynamic Conditions,” Tech. Memo. 89913, NASA Lewis Research
AERE Report C/M 273, Atomic Energy Center, Cleveland, OH, 1987
Research Establishment, Harwell, U.K., 40. L.F. Grantham, P.H. Shaw, and R.D.
March 1956 Oldenkamp, in High Temperature Metallic
33. J.W. Koger, Corrosion, Vol 29 (No. 3), Corrosion of Sulfur and Its Compounds,
1973, p 115 Z.A. Foroulis, Ed., The Electrochemical
34. J.W. Koger, Corrosion, Vol 30 (No. 4), Society, 1970, p 253
1974, p 125 41. L.F. Grantham and P.B. Ferry, in Proc. Int.
35. J.R. Keiser, D.L. Manning, and R.E. Symp. Molten Salts, J.P. Pemsler et al. Eds.,
Clausing, in Proc. Int. Symp. Molten Salts, The Electrochemical Society, 1976, p 270
Name ///sr-nova/Dclabs_wip/High Temp/5208_423-435.pdf/Chap_16/ 31/10/2007 12:52PM Plate # 0 pg 423

High-Temperature Corrosion And Materials Applications Copyright © 2007 ASM International®

George Y. Lai, editor, p423-435 All rights reserved.
DOI: 10.1361/hcma2007p423 www.asminternational.org

CHAPTER 16

Liquid Metal Corrosion and

Embrittlement

16.1 Introduction major alloying elements) with the molten metal

to form an intermetallic compound. This requires
Liquid metals are frequently used as a heat- some solubility of the liquid metal in the con-
transfer medium because of their excellent tainment metal. When a molten metal is used as
heat-transfer properties, such as high thermal a heat-transfer medium in a loop, mass transfer
conductivities, high heat capacities, and low due to thermal gradients in the melt can be a
vapor pressures. Various metals have been critical issue (Ref 6). It involves dissolution of
investigated for use as coolants in nuclear re- an alloying element in hot zones and deposition
actors. Sodium, for example, has been used as a of that alloying element in cold zones, resulting
coolant in fast breeder nuclear reactors. Most in tube plugging. The corrosion can also involve
corrosion studies of molten metals have been impurity and interstitial interactions, which are
carried out in conjunction with nuclear reactor generally associated with light elements such
applications (Ref 1–4). Other applications of as carbon, nitrogen, and oxygen. This type of
liquid metals include heat treat baths (e.g., reaction dictates the corrosion process when
molten lead) and power generation (Ref 5). the solubilities of the major alloying elements
The containment material, which is in contact in the molten metal are low. An example is car-
with the molten metal, is subject to molten metal burization-decarburization behavior of alloys in
corrosion. This chapter reviews data relevant a sodium loop. More complete discussions on the
to the corrosion of containment materials. Dis- mechanisms of molten metal corrosion can be
cussion of liquid metal embrittlement is also found in articles by numerous authors (Ref 1, 2,
included. 7–10).

16.2 Corrosion Process 16.3 Corrosion in Molten Aluminum

Molten metal corrosion of a containment metal Aluminum melts at 660 °C (1220 °F). Iron,
is most often related to its solubility in the molten nickel, and cobalt, along with their alloys, are
metal. This type of corrosion is simply a dis- readily attacked by molten aluminum. Extremely
solution-type attack. A containment metal with high corrosion rates of iron-, nickel-, and cobalt-
higher solubility in the molten metal generally base alloys in molten aluminum are illustrated
exhibits a higher corrosion rate. In the case of by the laboratory test results shown in Table 16.1
an alloy, the solubilities of the major alloying (Ref 11). Samples of carbon steel and iron-
elements could dictate the corrosion rate. The and nickel-base alloys were consumed in 4 h
solubility of an alloying element in a molten at 760 °C (1400 °F). Cobalt-base alloys, which
metal typically increases with increasing tem- appeared to be better than iron- and nickel-base
perature. As the temperature increases, the alloys, were corroding at rates too high to be con-
diffusion rate also increases. Thus, the alloy sidered for containment materials. In addition,
corrosion rate increases with increasing tem- titanium, although exhibiting a corrosion rate
perature. lower than iron-, nickel-, and cobalt-base alloys,
Corrosion by molten metal can also proceed should not be considered for use as a containment
by alloying of the containment metal (or its material because of its rapid corrosion rate.
Name ///sr-nova/Dclabs_wip/High Temp/5208_423-435.pdf/Chap_16/ 31/10/2007 12:52PM Plate # 0 pg 424

424 / High-Temperature Corrosion and Materials Applications

16.4 Corrosion in Molten Zinc containing up to 33% Ni, such as alloy 800H
(Fig. 16.2). The results of static tests in molten
Zinc melts at 420 °C (790 °F). Molten zinc is zinc for selected iron-, nickel-, and cobalt-base
widely used in the hot dip galvanizing process to alloys are summarized in Table 16.2. Nickel-base
coat steel for corrosion protection. Galvanizing alloys suffered the worst attack, followed by
tanks, along with baskets, fixtures, and other austenitic stainless steels, Fe-Ni-Cr alloys, and
accessories, require materials resistant to molten Fe-Cr alloys. Cobalt-base alloys generally per-
zinc corrosion. formed better. However, an Fe-Ni-Co-Cr alloy
Nickel and high-nickel alloys react readily
with molten zinc by direct alloying. This is illus-
trated in Fig. 16.1, which shows a nickel-base
alloy coupon after immersion in molten zinc at
455 °C (850 °F). Iron- and cobalt-base alloys
are generally corroded by dissolution, even those

Table 16.1 Results of static immersion tests in

molten aluminum at 760 °C (1400 °F) for 4 h
Maximum depth of corrosion
Alloy attack, mm (mils)
Titanium 0.22 (8.5)
6B 0.43 (16.8)
188 0.51 (20.2)
150 0.73 (28.9) (a)
556 >0.5 (20.6)(a) 200 µm
X >0.6 (23.8)(a)
671 >0.7 (26.3)(a)
Carbon steel >1.6 (63.1)(a)
(a) Sample was consumed. Source: Ref 11

(b)
200 µm

(c)
200 µm

Fig. 16.2 Sample cross sections for (a) alloy 556, (b) alloy
800H, and (c) Type 446 after immersion testing in
molten zinc at 455 °C (850 °F) for 50 h. The top edge of each
Fig. 16.1 Nickel-base alloy coupon after immersion testing photograph represents the original surface of the sample prior to
in molten zinc at 455 °C (850 °F) for 50 h. Source: immersion testing. Uniform dissolution of metal from the sample
Ref 11 surface is noted for all three alloys. Source: Ref 12
Name ///sr-nova/Dclabs_wip/High Temp/5208_423-435.pdf/Chap_16/ 31/10/2007 12:52PM Plate # 0 pg 425

Chapter 16: Liquid Metal Corrosion and Embrittlement / 425

(556 alloy) performed as well as cobalt-base (Ref 15). For example, cast iron centrifugal
alloys. pumps are used to pump liquid lead (Ref 15).
Intergranular cracking was observed for some Ali-Khan (Ref 16) performed extensive cor-
alloys in contact with molten zinc. Field testing rosion tests in molten lead. The results of his
in a galvanizing tank, with test coupons im- static corrosion tests at temperatures from 575
mersed in molten zinc at approximately 455 °C to 750 °C (1070 to 1380 °F) for chromium steels
(850 °F) for 19 runs (8 h immersion per run) for (7 to 17Cr) and austenitic stainless steels (18Cr-8
a total of 152 h, revealed intergranular cracking to 16Ni) are shown in Tables 16.3 and 16.4,
for coupons of alloy 25 (Co-Ni-Cr-W alloy) and respectively. Both exhibited similar perfor-
Type 316, as shown in Fig. 16.3 (Ref 12). The mance, with corrosion rates of less than 0.15 mm
other two alloys in the same field test (cobalt- (6 mils) after 3250 h of exposure.
base alloy 188 and Fe-Ni-Co-Cr alloy 556) did Wilkinson et al. (Ref 17) investigated molten
not suffer cracking. The metallurgical factors lead corrosion at a much higher temperature.
responsible for the cracking of alloy 25 and Type These authors performed static tests at 1000 °C
316 are unclear. Liquid metal embrittlement (1830 °F) for up to 408 h. Their test results
(LME) is discussed in Section 16.9. (Table 16.5) showed that the 17Cr steel (Type
Molten zinc corrosion is expected to become 430) was the most resistant. A sample of cast
more severe at higher temperatures. However, iron, on the other hand, completely dissolved in
very little data, particularly at 500 °C (930 °F) 408 h.
and higher, have been reported in the literature.

16.5 Corrosion in Molten Lead

Lead melts at 327 °C (620 °F). Nickel and
nickel-base alloys generally have poor resistance
to molten lead corrosion (Ref 13, 14). The solu-
bility of nickel in molten lead is higher than that
of iron. Cast iron, steels, and stainless steels
are commonly used for handling molten lead

Table 16.2 Results of static immersion tests

in molten zinc at 455 °C (850 °F) for 50 h
Alloy Depth of corrosion attack, mm (mils)
556 0.04 (1.6) (a) 100 µm
25 0.06 (2.3)
188 0.06 (2.5)
446SS 0.24 (9.3)
800H 0.28 (11.0)
304SS 0.36 (14.1)
625 >0.61 (24.0)(a)
X >0.61 (24.0)(a)
(a) Complete alloying. Source: Ref 11

Table 16.3 Results of static corrosion tests

at various temperatures for 3250 h in molten
lead for chromium steels
Depth of attack, μm (mils)
Alloy 575 °C (1067 °F) 650 °C (1200 °F) 750 °C (1382 °F)
7Cr-0.8Al 25 (1.0) (a) 30 (1.2)
(b)
13Cr 10 (0.4) 15 (0.6) 100 (3.9) 100 µm
13Cr-1.0Al 10 (0.4) 15 (0.6) …
13Cr-1.0Mo 10 (0.4) 30 (1.2) 150 (5.9)
17Cr (a) (a) 35 (1.4) Fig. 16.3 Intergranular cracking for (a) alloy 25 and (b) Type
316 coupons after field testing in a hot dip galva-
(a) Little or no corrosion. Source: Ref 16 nizing tank at 455 °C (850 °F) for a total of 152 h (19 runs).
Source: Ref 12
Name ///sr-nova/Dclabs_wip/High Temp/5208_423-435.pdf/Chap_16/ 31/10/2007 12:52PM Plate # 0 pg 426

426 / High-Temperature Corrosion and Materials Applications

The corrosion behavior of both chromium molten lead. Croloy 2.25Cr steel revealed no
steels and austenitic stainless steels in a thermal detectable corrosion attack in the Brookhaven
convection loop was studied by Ali-Khan loop after 27,765 h of exposure, while the same
(Ref 16). The results are shown in Tables 16.6 alloy showed 0.2 mm (8 mils) of attack in the
and 16.7. Compared with tests under static con- ORNL loop containing no inhibitors, although
ditions (Tables 16.3 and 16.4), alloys under the ORNL loop was operated 26 °C (47 °F)
thermal convection conditions corroded at faster higher. The beneficial effect of inhibitors has
rates. Tolson and Taboada (Ref 18) also reported been confirmed by Asher et al. (Ref 19). It is
corrosion data generated from a thermal con- believed that inhibitors are instrumental in the
vection loop (Table 16.8). The results of the formation of a protective film, which reduces the
Brookhaven loop, which used magnesium and solubility of the containment metal in molten
zirconium as inhibitors in the melt, are included lead (Ref 20).
in the table for comparison. It appears that the
inhibitors significantly reduced corrosion by

Table 16.6 Results of corrosion tests in a

molten lead thermal convection loop for 1002 h
Table 16.4 Results of static corrosion tests at 600 °C (1112 °F, hot leg) with a temperature
at various temperatures for 3250 h in molten differential of 232 °C (450 °F) for chromium steels
lead for austenitic stainless steels
Alloy Depth of attack, µm (mils)
Depth of attack, μm (mils)
7Cr-0.8Al 130–180 (5.1–7.1)
Alloy 575 °C 650 °C 750 °C 13Cr steel 273–300 (11–12)
(1067 °F) (1200 °F) (1382 °F)
13Cr-1.0Al 187–200 (7.4–7.9)
18Cr-9Ni 40 (1.6) 10 (0.4) 25 (1.0) 13Cr-1.0Mo 143–240 (5.6–9.5)
17Cr-8Ni 60–90 (2.4–3.5) 15 (0.6) 10 (0.4) 17Cr 130–140 (5.1–5.5)
18Cr-8Ni 10 (0.4) 20 (0.8) (a)
Source: Ref 16
18Cr-10Ni 40 (1.6) 25 (1.0) 150 (5.9)
18Cr-11Ni-2Mo 55 (2.2) 15 (0.6) 10 (0.4)
18Cr-12Ni-2.5Mo 25 (1.0) (a) 70 (2.8)
18Cr-12Ni-2Mo 40 (1.6) 10 (0.4) 15 (0.6)
18Cr-16Ni-3.5Mo 80 (3.2) 35 (1.4) 20 (0.8)
18Cr-10Ni 110 (4.3) 35 (1.4) 20 (0.8)
Table 16.7 Results of corrosion tests in a
(a) Little or no corrosion. Source: Ref 16 molten lead thermal convection loop for 1008 h
at 600 °C (1112 °F, hot leg) with a temperature
differential of 151 °C (304 °F) for austenitic
stainless steels
Table 16.5 Corrosion of several alloys in molten Alloy Depth of attack, µm (mils)
lead at 1000 °C (1830 °F) for up to 408 h 18Cr-9Ni 80–140 (3.2–5.5)
17Cr-8Ni 53–80 (2.1–3.2)
Alloy Corrosion rate, mm/yr (mpy)
18Cr-8Ni 53–80 (2.1–3.2)
Cast iron Complete dissolution in 408 h 18Cr-10Ni 62–100 (2.4–3.9)
ARMCO iron 4.0 (157) 18Cr-11Ni-2Mo 80–140 (3.2–5.5)
1020 steel 2.5 (100) 18Cr-12Ni-2.5Mo 80 (3.2)
430SS 0.7 (27) 18Cr-12Ni-2Mo 130–200 (5.1–7.9)
302SS 10.0 (393) 18Cr-16Ni-3.5Mo 120–140 (4.7–5.5)
347SS 9.8 (387) 18Cr-10Ni 55–70 (2.2–2.8)
Source: Ref 17 Source: Ref 16

Table 16.8 Results of thermal-convection loop tests in molten lead

Maximum Maximum
temperature, Operated depth of attack,
Loop material °C (°F) DT, °C (°F) time, (h) µm (mils) Inhibitors
Croloy 2.25Cr steel 654 (1210) 149 (300) 266 25 (1.0) None
410 654 (1210) 149 (300) 1,346 50 (2.0) None
Croloy 2.25Cr steel 593 (1100) 93 (200) 5,156 203 (8.0) None
Carbon steel (A-106) 593 (1100) 93 (200) 5,064 254 (10) None
Nb-1Zr (cladded on type 446) 760 (1400) 204 (400) 5,280 None None
Croloy 2.25Cr steel(a) 550 (1022) 105 (221) 27,765 None Mg+Zr
(a) Brookhaven loop. Source: Ref 18
Name ///sr-nova/Dclabs_wip/High Temp/5208_423-435.pdf/Chap_16/ 31/10/2007 12:52PM Plate # 0 pg 427

Chapter 16: Liquid Metal Corrosion and Embrittlement / 427

16.6 Corrosion in Molten Lithium Figure 16.5 illustrates the performance of

several iron-, nickel-, and cobalt-base alloys in
Liquid lithium is an attractive candidate molten lithium at 890 °C (1635 °F) (Ref 28).
blanket material for fusion reactors. Numerous Two nickel-base alloys (RA333 and Hastelloy
studies (Ref 21–27) have evaluated the compat- alloy X) suffered the worst corrosion attack in
ibility of various containment materials with terms of weight loss, followed by a cobalt-base
liquid lithium. This section gives a brief sum- alloy (Airesist 213) with two iron-base alloys
mary of the comparative performance of alloys (E-Brite alloy and Type 304L) being the best
in molten lithium. performers. However, in terms of penetration
The solubilities of some important alloying depths, Airesist 213, E-Brite, and Type 304L
elements in liquid lithium are shown in Fig. 16.4 behaved similarly (Fig. 16.6). TZM (Mo-0.5Ti-
(Ref 25). Nickel has the highest solubility 0.1Zr) was also included in the investigation,
among the metals investigated by Cleary et al. revealing no corrosion attack at all.
(Ref 25). Thus, nickel and nickel-base alloys Freed and Kelly (Ref 21) performed corrosion
are generally not considered good candidates tests in circulating molten lithium at 705 to
for handling liquid lithium. Cobalt-base alloys 815 °C (1300 to 1500 °F) for iron and stainless
are only slightly more resistant than nickel-base steels (Table 16.9). Both suffered considerable
alloys (Ref 21, 22, 28). metal loss due to dissolution. For austenitic

Temperature, °F
1700 1600 1500 1400 1300 1200
10–1

Nickel Nickel
Chromium
Iron
10–2 Titanium

Titanium results reported as being <10 ppm

Molybdenum
Solubility, atomic percent solute

Columbium

10–3

Chromium

Iron

10–4 Titanium

Columb
ium
Molybdenum

10–5
0.82 0.86 0.90 0.94 0.98 1.02 1.03 1.10
1000 / T , K

Fig. 16.4 Solubilities of some metals in molten lithium. Note: Columbium is the former (pre-1968) name of niobium. Source: Ref 25
Name ///sr-nova/Dclabs_wip/High Temp/5208_423-435.pdf/Chap_16/ 31/10/2007 12:52PM Plate # 0 pg 428

428 / High-Temperature Corrosion and Materials Applications

stainless steels, intergranular attack was also Fig. 16.7, are presented in terms of an Arrhenius
observed. The stabilized stainless steels, such as plot of dissolution rates for ferritic steels (HT-9
Types 321 and 347, were more resistant to inter- and 9Cr-1Mo) and austenitic stainless steels
granular attack. Also observed were significant (Type 316, cold-worked Type 316, and Type
amounts of deposits due to mass transfer. This 304). HT-9 exhibits excellent compatibility with
can result in flow restrictions, or eventual tube molten lithium.
plugging, in a circulating system. Chopra et al.
(Ref 24) summarized the data generated from
different circulating loops (forced circulation 16.7 Corrosion in Sodium
loop, FCL, and thermal convection loop, TCL)
at different laboratories. The data, illustrated in Liquid sodium is used as a coolant for fast
breeder reactors. Extensive studies (Ref 29–35)
on the corrosion of various alloys in liquid
1

10

RA-333

10–1 Hastelloy x
Weight loss, gm/cm2

1
Airesist-213
Penetration, mm

RA-333

Hastelloy x

10–2
E-Brite 26-1

10–1 Airesist-213

304L s.s E-Brite 26-1

304L stainless steel

10–3
TZM (no attack) TZM (no attack)
10–2
104 105 106 104 105 106
Time, s Time, s

Fig. 16.5 Weight loss as a function of time for selected alloys Fig. 16.6 Total penetration as a function of time for selected
in molten lithium (Ti-gettered) at 890 °C (1635 °F). alloys in molten lithium (Ti-gettered) at 890 °C
Source: Ref 28 (1635 °F). Source: Ref 28

Table 16.9 Results of corrosion tests in molten lithium at 705 to 815 °C (1300 to 1500 °F) in a
forced-convection loop
Alloy Exposure Maximum depth of Maximum thickness
time, h attack(a), mm (mils) of deposits, mm (mils)
Iron(b) 108–138 IG 0.0 0.32–0.38 (12.5–15.0)
ML 0.05–0.11 (2.0–4.5)
Iron 138–187 IG 0.0–0.02 (0.0–0.6) 0.36–0.46 (14.0–18.0)
ML 0.13–0.17 (5.0–6.6)
304 105–138 IG 0.03–0.13 (1.0–5.0) 0.47–0.51 (18.5–20.0)
ML 0.10–0.15 (3.8–6.0)
310 64–96 IG 0.08–0.10 (3.0–4.0) 0.49–0.61 (19.5–24.0)
ML 0.06–0.12 (2.2–4.7)
321 69–200 IG 0.0–0.05 (0.0–2.0) 0.64–0.81 (25.0–32.0)
ML 0.15–0.16 (6.2–6.4)
347 82–160 IG 0.01–0.02 (0.5–0.6) 0.84–1.02 (33.0–40.0)
ML 0.11–0.12 (4.3–4.9)
(a) ML, metal loss due to apparent solution attack, decrease in wall thickness; IG, intergranular attack. (b) Titanium getter in lithium flow stream. Source: Ref 21
Name ///sr-nova/Dclabs_wip/High Temp/5208_423-435.pdf/Chap_16/ 31/10/2007 12:52PM Plate # 0 pg 429

Chapter 16: Liquid Metal Corrosion and Embrittlement / 429

sodium have been carried out in support of corrosion database for a wide variety of alloys
sodium breeder reactor programs. In addition to has been published in Nuclear Systems Materials
numerous papers and reports discussing the Handbook (Ref 36) to provide guidance to
behavior of alloys in liquid sodium, a large corrosion allowance for design calculation. This

Temperature, °C
650 600 550 500 450 400 350
103

Type 316 stainless steel TCL Weight loss in lithium

and
HT-9 TCL and FCL

ORNL
ANL
WARD
UW
102 SU
JAERI
Dissolution rate, mg/m2.h

10 Type 316 stainless

steel FCL

Scatter band
TCL type 316 stainless
steel FCL
Type 316 stainless steel
PCA
HT-9
1
FCL

Type 316 stainless steel

Type 316 CW stainless steel
Type 304 stainless steel
PCA
HT-9
9 Cr-1Mo

10–1
1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7
1000 /T, K

Fig. 16.7 Corrosion rates of Types 304 and 316, PCA (primary candidate alloy), and HT-9 and 9Cr-1Mo steels in flowing lithium. CW,
cold worked; FCL, forced-convection loop; TCL, thermal-convection loop. Source: Ref 24
Name ///sr-nova/Dclabs_wip/High Temp/5208_423-435.pdf/Chap_16/ 31/10/2007 12:52PM Plate # 0 pg 430

430 / High-Temperature Corrosion and Materials Applications

section gives a very brief summary of the com- Table 16.10 Corrosion of carbon steel,
parative performance of alloys in liquid sodium. chromium-molybdenum steels, and stainless steels
Berry (Ref 1) summarized the data gener- in liquid sodium under isothermal conditions
ated by numerous authors (Ref 29–34). The data Weight
were generated in both static and flowing sys- Test
temperature, Exposure Test
change rate,
mg/cm2
tems at temperatures from 550 to 595 °C (1025 Materials °C (°F) time, h system per month

to 1100 °F), with some data generated at 1010 steel 593 (1100) 1000 Flowing −0.49
593 (1100) 1000 Static −0.37
1000 °C (1830 °F). Carbon steel, Cr-Mo steels,
2.25Cr-1Mo 552 (1026) 943 Flowing −0.12
and ferritic and austenitic stainless steels exhib- 556 (1033) 902 Static −0.12
ited low corrosion rates at temperatures up to 593 (1100) 1000 Flowing −0.14
593 (1100) 1000 Static −0.09
595 °C (1100 °F) (Table 16.10). The major prob-
5Cr-0.5Mo 552 (1026) 943 Flowing +0.22
lem for low-chromium alloy steels in sodium is 566 (1033) 1913 Static -0.06
decarburization and resultant loss of strength 593 (1100) 500 Flowing +0.23
(Ref 35). For austenitic stainless steels and 593 (1100) 500 Static −0.08
9Cr-1Mo 552 (1026) 943 Flowing +0.35
nickel-base alloys, the reaction between the alloy 566 (1033) 902 Static −0.05
and the sodium leads to carburization (Ref 35). 593 (1100) 500 Flowing +0.70
Corrosion of alloys in liquid sodium can be 593 (1100) 500 Static +0.29
severe at higher temperatures. Ferritic and aus- 410 593 (1100) 1000 Flowing +0.38
593 (1100) 1000 Static +0.35
tenitic stainless steels suffered rapid corrosion 420 593 (1100) 1000 Flowing +0.33
attack at 1000 °C (1830 °F) (Table 16.10). A 593 (1100) 1000 Static +0.31
nickel-base alloy (alloy 600) also exhibited a 304 593 (1100) 1000 Flowing +0.17
rapid corrosion rate (Table 16.10). Borgstedt 593 (1100) 1000 Static +0.15
310 593 (1100) 500 Flowing +0.75
et al. (Ref 37) investigated the corrosion behavior
593 (1100) 500 Static +0.27
of several Fe-Ni-Cr and nickel-base alloys in
316 593 (1100) 1000 Flowing +0.10
liquid sodium at 1000 °C (1830 °F). Their 593 (1100) 1000 Static +0.13
results are summarized in Table 16.11. Nickel- 347 593 (1100) 500 Flowing +1.46
base alloys suffered more attack than Fe-Ni-Cr 593 (1100) 500 Static +0.22
alloys. 410 1000 (1830) 400 Static +29.8
430 1000 (1830) 400 Static +46.8
446 1000 (1830) 400 Static +28.2
16.8 Corrosion in Other Molten Metals 304 1000 (1830) 400 Static +25.5
316 1000 (1830) 400 Static +29.6
310 1000 (1830) 400 Static +28.2
Corrosion data for other liquid metals, such
347 1000 (1830) 400 Static +44.2
as magnesium (melting point of 651 °C, or
600 1000 (1830) 400 Static +18.7
1205 °F); cadmium (melting point of 321 °C,
or 610 °F); mercury (melting point of −39 °C, Note: Sodium contained a maximum of 100 ppm oxygen. Source: Ref 1, based on
Ref 29–34
or −38 °F); tin (melting point of 232 °C, or
450 °F); antimony (melting point of 631 °C, or
1165 °F); and bismuth (melting point of 271 °C,

Table 16.11 Corrosion of several iron-nickel-chromium and nickel-base alloys in liquid sodium
at 1000 °C (1830 °F)
Test Alloy Cr Ni Fe Others Weight change, mg/cm2
Run No. 1, 1000 h Thermon 22 bal 30 W, Nb −13.0
617 21 bal 1.5 Co, Mo, Al −13.0
X 21 bal 18 Mo −4.5
AC-66 27 32 bal Nb, Ce −2.35
ASL71 20 20 bal Co, W −1.94
Pyrotherm 20 33 bal Nb −2.0
800 20 33 bal Al, Ti −0.7
Run No. 2, 1100 h 625 22 bal 2.5 Mo, Nb −36.01
625 22 bal 2.5 Mo, Nb −35.35
617 21 bal 1.5 Co, Mo, Al −25.11
X 21 bal 18 Mo −16.32
Pyrotherm 20 33 bal Nb +1.43
253-MA 21 11 bal Si +9.95
Source: Ref 37
Name ///sr-nova/Dclabs_wip/High Temp/5208_423-435.pdf/Chap_16/ 31/10/2007 12:52PM Plate # 0 pg 431

Chapter 16: Liquid Metal Corrosion and Embrittlement / 431

or 520 °F); are rather limited. Frequently, it is (Ref 15) indicated that cast iron has frequently
possible to use phase diagrams to determine been used for handling molten cadmium.
which metals may react readily with the liquid Nickel and nickel-base alloys are not suitable
metal. It is clear from the Ni-Mg phase diagram for handling molten tin because of relatively high
that magnesium reacts readily with nickel to solubilities of nickel in molten tin. Cast iron is
form a low-melting-point eutectic, with a large often used in the laboratory for handling molten
liquid phase field extending to low temperatures. tin (Ref 15). Cobalt, nickel, and iron have high
More than 50% Ni, for example, can be in solu- solubilities in molten antimony. These metals, as
tion with magnesium in liquid at 800 °C well as their alloys, are expected to have poor
(1470 °F). Nickel and nickel-base alloys are corrosion resistance in molten antimony.
therefore not suitable for use as a containment The solubilities of a number of metals in liquid
material for molten magnesium. Cobalt also mercury are illustrated in Fig. 16.8 (Ref 40).
forms a low-melting-point eutectic, but with a Both iron and cobalt have low solubilities in
much smaller liquid-phase field. The solubility of liquid mercury. Table 16.12 illustrates corrosion
iron in liquid magnesium is very low. It is thus rates of carbon steels and stainless steels in liquid
reasonable to assume that iron-base alloys with mercury (Ref 8).
low nickel are more suitable than cobalt- and
nickel-base alloys to handle molten magnesium.
The solubility of iron increases with increasing 16.9 Liquid Metal Embrittlement
temperature. Thus, the corrosion rate is expected
to increase with temperature as well. Metallic components may suffer embrittle-
The solubility of iron in molten cadmium ment when in contact with liquid metal. Com-
is low. Daniels (Ref 38) indicated the inactivity ponents are more susceptible to liquid metal
of steel in molten cadmium. Tammann and embrittlement (LME) when under stresses.
Oelsen (Ref 39) reported a solubility of 2 to McDonalt (Ref 41) indicated that stainless steels,
3× 10−4 wt% Fe in molten cadmium at 400 and nickel alloys and Cu-Ni alloys may suffer LME
700 °C (750 and 1290 °F). Jackson and Adams by the molten brazing alloy while the component
is stressed under brazing. Since the brazing
Temperature, °C operation involves contact of the metallic com-
780 727 679 636 596 560 527 496 ponent with a molten brazing alloy, LME can
1000
occur during brazing operation. Heiple et al.
Ti
Zr (Ref 42) reported that austenitic stainless steels
Ni
can be severely embrittled by copper and high
Zr
Ni copper braze alloys. However, austenitic stain-
100
less steels are not affected by silver-base braze
alloys (Ref 42). It was also found that Type
Ti
430 (a ferritic stainless steel) was not embrittled
10
by copper-base braze alloys, silver-base braze
Cr
alloys, aluminum, or gold (Ref 42). Small
ppm in Hg

Co
Table 16.12 Corrosion of metals and alloys in
Fe V
1.0 flowing liquid mercury
Material Maximum Test Corrosion rate,
temperature, duration, h mm/yr (mpy)
°C (°F)
Mild steel 482 (900) (a) 0.10 (4)
0.10 Nb 538 (1000) (a) 0.23 (9)
593 (1100) (a) 0.56 (22)
649 (1200) (a) 1.35 (53)
Ta < 0.002 5Cr steel 482 (900) (a) 0.05 (2)
538 (1000) (a) 0.10 (4)
0.01 593 (1100) (a) 0.25 (10)
0.95 1.0 1.05 1.1 1.15 1.2 1.25 1.3 649 (1200) (a) 0.64 (25)
1000/ T, K –1 304 652 (1205) 460 0.51 (20)
310 650 (1200) 400–500 1.19 (47)
Fig. 16.8 Solubilities of some metals in liquid mercury. (a) Average of a large number of laboratory tests as well as samples from large-
Note: Columbium is the former (pre-1968) name scale boiler operations; exposures up to 10,000 h. Source: Ref 8
of niobium. Source: Ref 40
Name ///sr-nova/Dclabs_wip/High Temp/5208_423-435.pdf/Chap_16/ 31/10/2007 12:52PM Plate # 0 pg 432

432 / High-Temperature Corrosion and Materials Applications

amounts of copper contaminant on the surface glass-metal seals. Most silver solders (for higher-
of the alloy that is under welding can cause LME strength applications) melt in a temperature
in the heat-affected zone of cobalt-base alloys range of 610 to 800 °C (1125 to 1475 °F) and
(Ref 43, 44). Savage and Mushala (Ref 44) may contain copper, cadmium, zinc, lithium, and
believed that copper and cobalt are essentially tin. Alloys are more susceptible to LME under
insoluble in each other and thus form a classic stress. Many brazing alloys that are used for
LME couple. many stainless steel plumbing systems contain
Korb (Ref 45) summarized some of the LME copper. For example, brazed manifold tube joints
issues of various alloy combinations involved in using Nicoro 80 braze alloy (81.5Au-16.5Cu-
the construction of manned spacecraft as follows. 2Ni) to join 21-6-9 stainless steels for the shuttle
Many structural alloys are embrittled by low- orbiter were found to suffer cracking. The
melting-point metals. Aluminum is embrittled cracking also was observed for brazed joints for
by mercury, indium, tin, and zinc; steel by tin, 304L and alloy 718.
cadmium, zinc, lead, copper, and lithium; stain- Ebert (Ref 46) indicated that cadmium-plated
less steels by cadmium, aluminum, lead, and Cr-Mo steel (ASTM A 193 grade B) studs from a
copper; titanium by cadmium and mercury; and steam line connector associated with a power tur-
nickel by zinc, cadmium, and mercury. Many of bine fractured during service at 315 °C (600 °F)
these embrittling elements are used in low-tem- by cadmium-induced LME. Figure 16.9(a) shows
perature solders for electrical and avionics the fractured studs, and Fig. 16.9(b) shows inter-
applications, for example, lead-tin for electrical granular cracks in the cross section of a fractured
soldered connections, indium-base solders for stud.
Zinc has been widely used in the industry as
a corrosion-resistant coating for carbon steels
(e.g., hot dip galvanizing, electroplating, and
spray painting) (Ref 47). Korbrin (Ref 47) indi-
cated (a) carbon steels are susceptible to LME by
zinc particularly under stresses or cold-worked
conditions, (b) austenitic stainless steels and
nickel alloys can suffer LME when in contact
with molten zinc, or when welded to galvanized
steels or parts contaminated with zinc. Inter-
granular cracking was observed in Type 316 and
alloy 25 (Co-20Cr-10Ni-15W) coupons after
exposure to molten zinc in a hot dip galvanizing
(a) tank at 455 °C (850 °F) for 50 h, as shown in
Fig. 16.3 (Ref 12).
Dillon (Ref 48) found that a Type 321 nozzle
suffered LME due to molten zinc contamination
of welds from zinc-pigmented painting over-
spray during initial fabrication. Dillon (Ref 49)
observed the failure of ASTM A193 2H nuts
during service at 370 °C (700 °F) and attributed
the failure to LME by zinc due to cadmium-
plated/zinc phosphate coated nuts. The Cd-Zn
eutectic melts at 270 °C (515 °F). Gutzeit et al.
(Ref 50) indicated that austenitic stainless
steels are susceptible to LME by zinc when
welding or during heat treatment of stainless steel
components contaminated with zinc-rich paint.
Figure 16.10 shows the formation of inter-
(b) granular cracking at the heat-affected zone of a
Type 304 pipe weld joint when the area was
Fig. 16.9 (a) Fractured studs due to cadmium-induced contaminated with zinc-rich paint during weld-
LME during service at 315 °C (600 °F), and
(b) optical micrograph showing intergranular cracks at the cross ing. Zinc-rich paints containing only metallic
section of a fractured stud. Source: Ref 46 zinc powders as a principal component can cause
Name ///sr-nova/Dclabs_wip/High Temp/5208_423-435.pdf/Chap_16/ 31/10/2007 12:52PM Plate # 0 pg 433

Chapter 16: Liquid Metal Corrosion and Embrittlement / 433

Cast iron, steels, and stainless steels are com-

monly used for handling molten lead. Nickel and
nickel-base alloys have poor resistance to molten
lead corrosion because of the high solubility of
nickel in molten lead.
Nickel also has relatively higher solubility
than iron in molten lithium, as well as in molten
sodium. Thus, nickel-base alloys are not good
candidates for handling either metal. Iron-base
alloys are more suitable.
Nickel reacts more readily with molten mag-
nesium. The solubility of iron in molten mag-
nesium is low. Thus, iron-base alloys with low
nickel contents are preferred as containment
materials. Due to low solubilities in the molten
metal, iron and iron-base alloys may be suitable
for molten cadmium, while iron, cobalt, and
iron- and cobalt-base alloys may be suitable for
molten mercury. Iron, nickel, and cobalt exhibit
high solubilities in molten antimony. These
metals and their alloys are thus not recommended
100 µm for use as containment materials for molten
antimony.
Liquid metal embrittlement of various alloys
Fig. 16.10 Intergranular cracking in the heat-affected zone
by low-melting-point metals, such as copper,
of a weld joint for Type 304 pipe that was con-
taminated with zinc-rich paint when welding was performed. zinc, cadmium, mercury, lead, tin, and so forth, is
Source: Ref 50
reviewed. LME cases are often associated with
welding, brazing, and soldering. Most brazing
and soldering alloys contain alloying elements
LME, while paints containing zinc oxide or zinc that can induce LME. The components are more
chromates do not cause LME (Ref 50). susceptible to LME when they are under stress
during welding or soldering.

16.10 Summary
REFERENCES
The corrosion behavior of alloys in molten
aluminum, zinc, lead, lithium, sodium, magne- 1. W.E. Berry, Corrosion in Nuclear Applica-
sium, mercury, and other molten metals is tions, John Wiley & Sons, 1971
reviewed. Corrosion data useful in assisting 2. H.U. Borgstedt, Ed., Materials Behavior
selection of materials are presented. Also pre- and Physical Chemistry in Liquid Metal
sented is the information about the LME caused Systems, Plenum Press, 1982
by a wide variety of low-melting-point metals 3. J.E. Draley and J.R. Weeks, Ed., Corrosion
during welding or heat treatment. by Liquid Metals, Plenum Press, 1970
Molten aluminum is extremely aggressive. 4. C. Bagnall and W.F. Brehm, Corrosion,
Iron-, nickel-, and cobalt-base alloys are readily Vol 13, 9th ed., Metals Handbook, ASM
attacked by molten aluminum. Molten zinc is International, 1987, p 91
less aggressive. Nickel and nickel-base alloys, 5. D.L. Katz, Liquid-Metals Handbook,
however, react readily with molten zinc and R.N. Lyon, Ed., NAVEXOS, P-733 (Rev.),
are not recommended for use. Cast iron, steels, U.S. Government Printing Office, Wash-
and iron- and cobalt-base alloys are generally ington, D.C., 1952, p 1
suitable for containment applications. Cobalt- 6. J.V. Cathcart and W.D. Manly, Corrosion,
base alloys are generally more resistant than iron- Vol 12, 1956, p 87t
base alloys. However, under some conditions, 7. P.F. Tortorelli, Corrosion, Vol 13, 9th ed.,
various metals or alloys may be susceptible to Metals Handbook, ASM International,
LME by molten zinc. 1987, p 56
Name ///sr-nova/Dclabs_wip/High Temp/5208_423-435.pdf/Chap_16/ 31/10/2007 12:52PM Plate # 0 pg 434

434 / High-Temperature Corrosion and Materials Applications

8. E.C. Miller, Liquid-Metals Handbook, 24. O.K. Chopra, D.L. Smith, P.F. Tortorelli,
R.N. Lyon, Ed., NAVEXOS, P-733 (Rev.), J.H. DeVan, and D.K. Sze, Fusion Technol.,
U.S. Government Printing Office, Vol 8, 1985, p 1956
Washington, D.C., 1952 25. R.E. Cleary, S.S. Blecherman, and J.E.
9. W.D. Manly, Corrosion, Vol 12 (No. 7), Corliss, “Solubility of Refractory Metals
1956, p 336t in Lithium and Potassium,” USAEC Report
10. A. Brasunas, Corrosion, Vol 9, 1953, p 78 TIM-850, Nov 1965
11. G.Y. Lai, unpublished results, Haynes 26. D.A. Bates, G.R. Edwards, and D.L.
International, Inc., 1985 Olson, An Evaluation of Engineering
12. S.K. Srivastava, Proc. Conf. Performance of Alloys for High Temperature Lithium Con-
High Temperature Materials in Fluidized tainment, Mater. Perform., March 1980,
Bed Combustion Systems and Process p 41
Industries, P. Ganesan and R.A. Bradley, 27. V. Coen, H. Kolbe, L. Orecchia, and
Ed., ASM International, 1987, p 161 T. Sasaki, Materials Behavior and Physi-
13. F.R. Morrall, Wire Wire Prod., Vol 23, 1948, cal Chemistry in Liquid Metal Systems,
p 484, 571 H.U. Borgstedt, Ed., Plenum Press, 1982,
14. L.R. Kelman, W.D. Wilkinson, and F.L. p 121
Yaggee, “Resistance of Materials to Attack 28. P.A. Steinmeyer, D.L. Olson, G.R. Edwards,
by Liquid Metals,” Report ANL-4417, and D.K. Matlock, Rev. Coatings Corros.,
Argonne National Laboratory, 1950 Vol 4 (No. 4), 1981, p 349
15. C.B. Jackson and R.M. Adams, Liquid- 29. R.F. Koening and E.G. Brush, Mater.
Metals Handbook, R.N. Lyon, Ed., Methods, Vol 42, 1955, p 112
NAVEXOS, P-733 (Rev.), U.S. Government 30. A. Brasunas, “Interim Report on Static-
Printing Office, Washington, D.C., 1952 Liquid Metal Corrosion,” USAEC Report
16. I. Ali-Khan, Materials Behavior and Physi- ORNL-1647, Oak Ridge National Labora-
cal Chemistry in Liquid Metal Systems, tory, Oak Ridge, TN, 1954
H.U. Borgstedt, Ed., Plenum Press, 1982, 31. R.F. Dudek and K.M. Ferguson, “The
p 243 Corrosion Testing of Various Materials
17. W.D. Wilkinson, E.W. Hoyt, and H.V. in Sodium: Part I and II,” USAEC Re-
Rhude, “Attack on Materials by Lead at port BW-7020, Babcock & Wilcox, April
1000 °C,” USAEC Report ANL-5449, 1957
Argonne National Laboratory, 1955 32. V.W. Eldred, “Interactions Between Solid
18. G.M. Tolson and A. Taboada, “A Study of and Liquid Metals and Alloys,” British
Lead and Lead-Salt Corrosion in Thermal Report AERE-X/R-1806, Nov 1955
Convection Loops,” ORNL-TM-1437, Oak 33. W. Markert, Jr., “The Corrosion Testing
Ridge National Laboratory, 1966 of Various Materials in Sodium,” USAEC
19. R.C. Asher, D. Davies, and S.A. Beetham, Report BW-3792, Babcock & Wilcox, Aug
Corros. Sci., Vol 17, 1977, p 545 1954
20. O.F. Kammerer et al., Trans. AIME, Vol 212, 34. W.C. Hayes and O.C. Shepard, “Corro-
1958, p 20 sion and Decarburization of the Ferritic
21. M.S. Freed and K.J. Kelly, “Corrosion Chromium-Molybdenum Steels in Sodium
of Columbium Base and Other Structural Coolant Systems,” USAEC Report NAA-
Alloys in High Temperature Lithium,” SR-2973, North American Aviation, Dec
Report No. PWAC-355, Pratt and Whitney 1958
Aircraft—CANEL, Division of United Air- 35. J.H. Stang, E.M. Simons, J.A. DeMastry,
craft Corp., June 1961 (declassified in June and J.M. Genco, “Compatibility of Liquid
1965) and Vapor Alkali Metals with Construction
22. E.E. Hoffman, “Corrosion of Materials by Materials,” DMIC Report 227, Defense
Lithium at Elevated Temperatures,” USAEC Metals Information Center, Battelle Mem-
Report ORNL-2924, Oak Ridge National orial Institute, April 1966
Laboratory, Oak Ridge, TN, 1960 36. Nuclear Systems Materials Handbook,
23. J.R. DiStefano, “Corrosion of Refractory Hanford Engineering Laboratory, Richland,
Metals by Lithium,” USAEC Report ORNL- WA, 1976
3551, Oak Ridge National Laboratory, Oak 37. H.U. Borgstedt, G. Frees, and H. Jesper,
Ridge, TN, 1964 Werkst. Korros., Vol 40, 1989, p 525
Name ///sr-nova/Dclabs_wip/High Temp/5208_423-435.pdf/Chap_16/ 31/10/2007 12:52PM Plate # 0 pg 435

Chapter 16: Liquid Metal Corrosion and Embrittlement / 435

38. E.J. Daniels, J. Inst. Met., Vol 46, 1931, 46. H.E. Ebert, Liquid Metal Embrittlement
p 87 of Flange Connector Studs in Contact with
39. G. Tammann and W. Oelsen, Z. Anorg. Cadmium, Handbook of Case Histories
Chem., Vol 186, 1930, p 277 in Failure Analysis, Vol 1, R.C. Uhl, et al.,
40. J.R. Weeks, Corrosion, April 1967, p 98 Ed., ASM International, 1992, p 335
41. M.M. McDonalt, Corrosion of Brazed 47. G. Korbrin, Materials Selection, Corrosion,
Joints, Corrosion, Vol 13, 9th ed., Metals Vol 13, 9th ed., Metals Handbook, ASM
Handbook, ASM International, 1987, p 876 International, 1987, p 321
42. C. Heiple, W. Bennett, and T. Rising, 48. C.P. Dillon, Unusual Corrosion Problems
Embrittlement of Several Stainless Steels by in the Chemical Industry, MTI Publication
Liquid Copper and Liquid Braze Alloys, No. 54, Materials Technology Institute of
Mater. Sci. Eng., Vol 52, 1982, p 177 the Chemical Process Industries, Inc., 2000,
43. S.J. Matthews, M.O. Maddock, and p 161
W.F. Savage, How Copper Surface Con- 49. C.P. Dillon, Unusual Corrosion Problems
tamination Affects Weldability of Cobalt in the Chemical Industry, MTI Publication
Superalloys, Weld. J., May 1972 No. 54, Materials Technology Institute of
44. W.F. Savage and M. Mushala, Copper the Chemical Process Industries, Inc., 2000,
Contamination Cracking in the Weld Heat p 204
Affected Zone, Weld. J., May 1978, p 145 50. J. Gutzeit, R.D. Merrick, and L.R. Scharf-
45. L.J. Korb, Corrosion of Manned Spacecraft, stein, Corrosion in Petroleum Refining
Corrosion, Vol 13, 9th ed., Metals Hand- and Petrochemical Operations, Corrosion,
book, ASM International, Metals Park, Vol 13, 9th ed., Metals Handbook, ASM
Ohio, 1987, p 1059 International, 1987, p 1263
Name ///sr-nova/Dclabs_wip/High Temp/5208_437-441.pdf/Chap_17/ 31/10/2007 4:06PM Plate # 0 pg 437

High-Temperature Corrosion And Materials Applications Copyright © 2007 ASM International®

George Y. Lai, editor, p437-441 All rights reserved.
DOI: 10.1361/hcma2007p437 www.asminternational.org

CHAPTER 17

Hydrogen Attack

17.1 Introduction When steel is in contact with hydrogen

molecules (H2) at elevated temperatures, hydro-
Hydrogen attack can result in brittle fracture gen atoms can be absorbed at the steel surface
of a steel component during high-temperature and then diffuse into the metal. Hydrogen atoms
service. It has also been referred to as “hydrogen in the metal then react with iron carbide (Fe3C) to
damage.” Carbon steels or low-alloy steels form methane (CH4):
can suffer severe hydrogen attack during high-
temperature service, resulting in brittle failure Fe3 C+4H ! 3Fe+CH4 (17:1)
due to severe loss in tensile and rupture strengths
as well as ductility. Hydrogen attack should As Reaction 17.1 continues with ongoing
not be confused with hydrogen embrittlement. ingress of hydrogen atoms into the steel, methane
Hydrogen attack occurs only when steel is gas is increasingly generated in the steel. With its
in service at elevated temperatures, while low diffusivity in steel, methane gas is accumul-
hydrogen embrittlement occurs near or at room ating at grain boundaries and other interfaces,
temperature. causing the methane gas pressure to build up in
Carbon and Cr-Mo steels derive strength pri- the steel. This eventually results in development
marily from carbon in the metal. In carbon steels, of microcracks and microfissures. At the same
carbon in the metal combines with iron to form time, carbon is being removed from iron carbide
iron carbides, primarily cementite (Fe3C), which (Fe3C) as a result of this chemical reaction,
along with ferrite lamellae, constitutes to pearlite leading to decarburization in the affected region.
phases in the matrix of ferrite, as shown in The formation of microcracks and microfissures
Fig. 17.1. For low-alloy steels containing chro- as well as decarburization causes the steel to lose
mium and molybdenum, these two alloying ele- its tensile and rupture strengths as well as duc-
ments can form carbides, such as M3C, M2C, and tility, thus resulting in rupture of the steel com-
M23C6 where M can represent Fe, Cr, and Mo. ponent during service.
There are two sources of hydrogen that con-
tribute to hydrogen attack of steels in industries.
One source of hydrogen is from the corrosion
of steel by boiler water in the waterwall tube
for coal-fired boilers, while the other source
of hydrogen is from high-pressure, high-
temperature hydrogen-containing atmospheres
used in petroleum refining. Hydrogen attack
problems related to these two separate areas are
discussed in the following sections.

17.2 Hydrogen Attack in Coal-Fired

Boilers

Hydrogen attack occurs more often in sub-

Fig. 17.1 Typical microstructure of carbon steel, consisting of
critical drum boilers that use a recirculating
ferrite (white grains) and pearlite (dark regions). The
pearlite consists of cementite (Fe3C) and ferrite lamellae. steam-generating system. Adequate control of
Name ///sr-nova/Dclabs_wip/High Temp/5208_437-441.pdf/Chap_17/ 31/10/2007 4:06PM Plate # 0 pg 438

438 / High-Temperature Corrosion and Materials Applications

water chemistry is critical to avoid internal 17.1, the affected area is decarburized. The
deposition and corrosion of boiler tubes. The microstructure of the steel suffering hydrogen
water chemistry is typically controlled by (a) attack is shown in Fig. 17.2. Hydrogen attack
water purification to remove impurities using typically is associated with heavy scales
makeup water, polishing of returned condensate, (Fig. 17.2). Figure 17.3 shows microcracks and
deaeration, and blowdown, and (b) chemical microfissures that formed along grain boundaries
treatments to control pH, electrochemical for the waterwall steel tube (ASME SA192)
potential, and dissolved oxygen concentration suffering hydrogen attack in a subcritical coal-
(Ref 1). In the waterwall tubes, as steam forms, fired boiler. Also shown in the figure is the
dissolved solids concentrate in the boiler water; complete decarburization of the steel in the
once the solubility limit of an impurity is affected area. Figure 17.4 shows the typical
exceeded, deposits can precipitate out (Ref 1). pearlite-ferrite microstructure in the unaffected
Typical boiler deposits are hardness deposits of waterwall tube in the same boiler as discussed in
calcium and magnesium salts and metal oxides Fig. 17.2 and 17.3. Large cracks developed in the
(Ref 1). These deposits formed on the surface of waterwall steel tube through the linking of
the waterwall tube can provide a location for numerous microcracks and microfissures. This is
accumulation of corrosive boiler water con- illustrated in Fig. 17.5. Large cracks were often
taminants and/or chemical additives and their found to show oxides formed on the surface of
reaction compounds (Ref 2). The deposits may these cracks (Fig. 17.5). The tube eventually
lead to low pH water chemistry in the local area, ruptured from the internal fluid pressure. The
resulting in acidic corrosion attack (Ref 3). locations that are often susceptible to hydrogen
With adequate water chemistry control, the attack in the boiler are the burner zone and
waterwall steel tube forms a protective magnetite the bull nose area, as illustrated in Fig. 17.6
(Fe3O4) scale when steel is corroded by water (Ref 4). Monitoring and controlling boiler water
under normal operating conditions:

3Fe+4H2 O ! Fe3 O4 +4H2 (17:2)

The by-product of the reaction is hydrogen.

The growth of the magnetite (Fe3O4) scale that
forms on the steel surface follows a parabolic
rate law with the scale growth rate diminishing
with time. Accordingly, the generation of
hydrogen likewise diminishes with time once
a steady state is reached. Atomic hydrogen
produced combines with other hydrogen atoms
to form molecular hydrogen that is then dis-
persed in the boiler water as a gas or in solution
(Ref 2).
However, when water chemistry is poorly
controlled, heavy internal tube deposits may
develop. This condition may lead to acidic cor-
rosion attack on the tube under the deposits,
resulting in large amounts of hydrogen atoms
being absorbed by the steel. Hydrogen atoms
then react with iron carbide (Fe3C) in the steel to
form methane gas (CH4) according to Reaction
17.1. Methane gas then accumulates at grain
boundaries and other interfaces due to its low
diffusivity. Microcracks and microfissures are
eventually developed by increasing local gas
Fig. 17.2 Low-magnification optical micrograph showing
pressure created by the increasing amount of numerous microcracks and microfissures formed
methane gas produced by the iron-carbide/ in the carbon steel (ASME SA192) of the waterwall tube that suf-
fered hydrogen attack in a subcritical coal-fired boiler. Note
hydrogen reaction. Furthermore, because of numerous microcracks and microfissures. Also shown are heavy
iron carbides being reduced to iron by Reaction corrosion scales formed on the tube internal surface. Unetched
Name ///sr-nova/Dclabs_wip/High Temp/5208_437-441.pdf/Chap_17/ 31/10/2007 4:06PM Plate # 0 pg 439

Chapter 17: Hydrogen Attack / 439

chemistry is critical in preventing internal tube partial pressure of hydrogen in the system. For
deposits and hydrogen attack (Ref 3). hydrotreating, reforming, and hydrocracking
units, steel can suffer hydrogen attack at tem-
peratures above approximately 260 °C (500 °F)
17.3 Hydrogen Attack in Petroleum and hydrogen partial pressures above 0.689 MPa
(100 psig) (Ref 7).
Refining The mechanism for hydrogen attack of steels
In petroleum refining, some refinery equip- in petroleum refining is essentially the same as
ment, such as reactors in hydrotreating, reform- that described in Section 17.2. Atomic hydrogen
ing, and hydrocracking units, is exposed to is first absorbed by the steel at the surface of the
a high-temperature, high-pressure hydrogen refinery equipment that is in contact with the
atmosphere. Hydrogen attack is potentially a high-temperature, high-pressure hydrogen gas
very serious materials issue in the design and stream. Hydrogen atoms then diffuse into the
operation of this type of equipment in hydrogen steel and react with iron carbide (Fe3C) in the
service (Ref 5, 6). Steels are susceptible to steel to form methane gas (CH4) according to
hydrogen attack as functions of temperature and Reaction 17.1. Methane gas then accumulates at
grain boundaries and other interfaces due to its
low diffusivity. Continuous ingress of hydrogen
atoms into the steel results in an increasing
amount of methane gas being generated by the
iron-carbide/hydrogen reaction, eventually caus-
ing microcracks and microfissures in the steel.
Cracks then develop by linking microcracks and
microfissures. Furthermore, because iron car-
bides are being reduced to iron by Reaction 17.1,
the affected area is decarburized. As a result,
the steel eventually ruptures from the internal
pressure of the reactor or vessel. The micro-
structure of the steel that suffers hydrogen attack
is characterized by numerous microcracks and
microfissures (or cracks) and decarburized
microstructure, which is essentially the same as
Fig. 17.3 Optical micrograph showing the formation of that described in the waterwall tube that suffers
microcracks and microfissures along grain hydrogen attack in the boiler. Typical optical
boundaries and decarburization of carbon steel (ASME SA192) in
a waterwall tube that suffered hydrogen attack in a subcritical microstructure of the hydrogen-attacked steel is
coal-fired boiler similar to that shown in Fig. 17.3.

Fig. 17.5 Optical micrograph showing a large crack that was

Fig. 17.4 Optical micrograph showing typical microstructure developed by linking microcracks and micro-
(pearlite and ferrite) in the unaffected carbon steel fissures in the steel (ASME SA192) of the waterwall tube that suf-
waterwall tube in the same boiler as discussed in Fig. 17.2 and fered hydrogen attack. Oxides were often observed to form
17.3 around the surface of the crack.
Name ///sr-nova/Dclabs_wip/High Temp/5208_437-441.pdf/Chap_17/ 31/10/2007 4:06PM Plate # 0 pg 440

440 / High-Temperature Corrosion and Materials Applications

Adding chromium and/or molybdenum to the “hydrogen damage.” The chemical reaction
steel to increase the stability of iron carbides can involves atomic hydrogen reacting with iron
increase the resistance of the steel to hydrogen carbide in the steel to form methane gas. Con-
attack. Some concerns have been raised for the tinuing ingress of atomic hydrogen into the metal
long-term performance of C-0.5Mo steel in high- causes an increasing amount of methane gas to be
temperature hydrogen environments (Ref 8). generated and accumulated at grain boundaries
Thus, Cr-Mo steels are much more resistant to and other interfaces, resulting in the formation of
hydrogen attack than carbon and C-0.5Mo steels. microcracks and microfissures in the steel. In
The conditions under which carbon and Cr-Mo addition, the steel is decarburized. Continuing
steels can be used in high-temperature hydrogen growth of microcracks and microfissures along
service are described in detail in API 941 (Ref 9). with decarburization of steel eventually result in
The behavior of carbon and Cr-Mo steels with rupture of the steel component during service at
respect to their resistance to hydrogen attack elevated temperatures.
is summarized in Nelson curves in Fig. 17.7 The source of atomic hydrogen can be from
(Ref 9). the rapid waterside corrosion at the internal
diameter of the waterwall tubes in a coal-fired
boiler when water chemistry is not properly
controlled, thus resulting in hydrogen attack.
17.4 Summary Hydrogen attack can also occur in some refinery
equipment, such as reactors in hydrotreating,
During service at elevated temperatures, reforming, and hydrocracking units, which is
carbon steel can react with atomic hydrogen exposed to a high-temperature, high-pressure
and result in brittle fracture. This phenomenon hydrogen atmosphere. Hydrogen attack in both
is often referred to as “hydrogen attack” or of these systems is discussed.

Steam

Risers
1

3 Heat flux
1 profile
2

Downcomer 1

Furnace
tubes

Steam-
water 1
mixture 1

Burners

1
1

Heat flux

Steam-free subcooled water in

Fig. 17.6 Locations, marked as 1, in the boiler that are susceptible to hydrogen attack. The area marked as 2 shows other modes of
waterside corrosion that are outside of the current discussion topic. Source: Stultz and Kitto (Ref 4) Courtesy of Babcock and
Wilcox
Name ///sr-nova/Dclabs_wip/High Temp/5208_437-441.pdf/Chap_17/ 31/10/2007 4:06PM Plate # 0 pg 441

Chapter 17: Hydrogen Attack / 441

Hydrogen partial pressure, MPa absolute

34.5 62.1
3.45 6.90 10.34 13.79 17.24 20.7 48.3 75.8
1500
800

1400

1300 700

1200

1100 600
6.0Cr-0.5Mo steel

Temperature, °C
Temperature, °F

1000
1.25Cr-0.5Mo steel
3.0Cr-0.5Mo steel 500
900

1.0Cr-0.5Mo steel 2.25Cr-1.0Mo steel

800
400
2.0Cr-0.5Mo steel
700

600
1.25Cr-0.5Mo or 1.0Cr-0.5Mo steel 300

500

Carbon Steel
400 200

300
500 1000 1500 2000 2500 3000 7000 11,000
Legend: Hydrogen partial pressure, 1b/psia 5000 9000 13,000
Surface decarburization
Internal decarburization Scale Change
(Hydrogen attack)

Fig. 17.7 Nelson curves showing the temperature and hydrogen partial pressure conditions under which carbon and Cr-Mo steels are
susceptible to hydrogen attack. Source: API 421 (Ref 9). Courtesy of American Petroleum Institute.

REFERENCES Refinery Equipment, Mater. Perform., Vol 17

(No. 8), 1978, p 33
1. P. Daniel, Corrosion of Steam- and Water- 6. A.R. Ciuffreda and W.R. Rowland, Hydrogen
Side of Boilers, Corrosion: Environments and Attack of Steel in Reformer Service, Proc.
Industries, Vol 13C, ASM Handbook, ASM API, Vol 37 (No. III), 1957, p 116
International, 2006, p 466 7. R.D. Merrick and A.R. Ciuffreda, Hydro-
2. H.A. Grabowski, Chapter 17, Management of gen Attack of Carbon-0.5 Molybdenum
Cycle Chemistry, ASME Handbook on Water Steels, Proc. API, Vol 61 (No. III), 1982,
Technology for Thermal Power Systems, p 101
P. Cohen, Ed., ASME, 1989, p 1379 8. R. Chiba, K. Ohnishi, K. Ishii, and K. Maeda,
3. R.B. Dooley, Corrosion of Steam/Water- Effect of Heat Treatment on Hydrogen Attack
side Boilers, Corrosion, Vol 13, 9th ed., Resistance of C-0.5Mo Steels for Pressure
Metals Handbook, ASM International, 1987, Vessels, Heat Exchangers, and Piping, Cor-
p 990 rosion, Vol 41 (No. 7), 1985, p 415
4. S.C. Stultz and J.B. Kitto, Ed., Steam and Its 9. Steels for Hydrogen Service at Elevated
Generation and Use, 40th ed., Babcock & Temperatures and Pressures in Petroleum
Wilcox, 1992, p 42-22 Refineries and Petrochemical Plants, Pub-
5. G. Sorell and M.J. Humphries, High Tem- lication 941, 3rd ed., American Petroleum
perature Hydrogen Damage in Petroleum Institute, 1983
Name ///sr-nova/Dclabs_wip/High Temp/5208_445-453.pdf/Appendix_2/ 26/10/2007 1:55PM Plate # 0 pg 445

High-Temperature Corrosion And Materials Applications Copyright © 2007 ASM International®

George Y. Lai, editor, p445-453 All rights reserved.
DOI: 10.1361/hcma2007p445 www.asminternational.org

APPENDIX 2

Chemical Compositions of Alloys and

Filler Metals
Table 1 Chemical compositions (wt.%) of wrought ferritic steels
Grade* C Cr Mo Mn Si Others
T2 0.10–0.20 0.50–0.81 0.44–0.65 0.30–0.61 0.10–0.30 …
T11 0.05–0.15 1.00–1.50 0.44–0.65 0.30–0.60 0.50–1.00 …
T12 0.05–0.15 0.80–1.25 0.44–0.65 0.30–0.61 0.50(a) …
T17 0.15–0.25 0.80–1.25 … 0.30–0.61 0.15–0.35 V: 0.15 min
T21 0.05–0.15 2.65–3.35 0.80–1.06 0.30–0.60 0.50(a) …
T22 0.05–0.15 1.90–2.60 0.87–1.13 0.30–0.60 0.50(a) …
T23 0.04–0.10 1.90–2.60 0.05–0.30 0.10–0.60 0.50(a) W: 1.45–1.75, V: 0.20–0.30, Nb: 0.02–0.08,
B: 0.0005–0.006
T5 0.15(a) 4.00–6.00 0.45–0.65 0.30–0.60 0.50(a) …
T9 0.15(a) 8.00–10.00 0.90–1.10 0.30–0.60 0.25–1.00 …
T91 0.08–0.12 8.00–9.50 0.85–1.05 0.30–0.60 0.20–0.50 V: 0.18–0.25, Nb: 0.06–0.10, N: 0.03–0.07
T92 0.07–0.13 8.50–9.50 0.30–0.60 0.30–0.60 0.50(a) V: 0.15–0.25, W: 1.5–2.0, Nb: 0.04–0.09,
B: 0.001–0.006, N: 0.03–0.07
T122 0.07–0.14 10.00–12.50 0.25–0.60 0.70(a) 0.50(a) V: 0.15–0.30, W: 1.50–2.50, Cu: 0.30–1.70,
Nb: 0.04–0.10, B: 0.0005–0.005,
N: 0.04–0.10
(a) Maximum, Nb = Cb. * For tubes

Table 2 Chemical compositions (wt.%) of wrought stainless steels

Alloy UNS No. C Cr Ni Fe Others
403 S40300 0.15(a) 11.5–13.0 … Bal …
410 S41000 0.15(a) 11.5–13.5 … Bal …
414 S41400 0.15(a) 11.5–13.5 1.25–2.50 Bal …
416 S41600 0.15(a) 12.0–14.0 … Bal S: 0.15 min
416 (Se) S41623 0.15(a) 12.0–14.0 … Bal Se: 0.15 min
420 S42000 0.15(a) 12.0–14.0 … Bal …
431 S43100 0.20(a) 15.0–17.0 1.25–2.50 Bal …
440A S44002 0.60–0.75 16.0–18.0 … Bal Mo: 0.75(a)
440B S44003 0.75–0.95 16.0–18.0 … Bal Mo: 0.75(a)
440C S44004 0.95–1.20 16.0–18.0 … Bal Mo: 0.75(a)
GREEK ASCOLOY S41880 0.12 12.6 2.0 Bal W: 3.0
154CM … 1.05 14.0 … Bal Mo: 4.0
405 S40500 0.08(a) 11.5–14.5 … Bal At: 0.10–0.30
430 S43000 0.12(a) 14.0–18.0 … Bal …
430F S43020 0.12(a) 14.0–18.0 … Bal S: 0.15 min
430F (Se) S43023 0.12(a) 14.0–18.0 … Bal Se: 0.15 min
446 S44600 0.20(a) 23.0–27.0 … Bal N: 0.25(a)
409 S40900 0.08(a) 10.5–11.75 0.5(a) Bal Ti: (6 × C) min, 0.75(a)
439 S43035 0.07(a) 17.0–19.0 0.5(a) Bal Ti: 0.2 + 4(C + N) min, 1.10(a)
444 S44400 0.025(a) 17.5–19.5 1.0(a) Bal Mo: 1.75–2.50, N: 0.035(a), Ti + Cb:
0.2 + 4(C + N) min, 0.8(a)
E-BRITE S44627 0.01(a) 25.0–27.5 0.5(a) Bal Mo: 0.75–1.50, N: 0.015(a), Cb: 0.05–0.20
MONIT S44635 0.025(a) 24.5–26.0 3.5–4.5 Bal Mo: 3.4–4.5, N: 0.035(a), Ti + Cb:
0.2 + 4(C + N) min, 0.80(a)
SEA-CURE S44660 0.03(a) 25.0–28.0 1.0–3.5 Bal Mo: 3.0–3.5, N: 0.04(a), Ti + Cb:
0.2–1.0, 6(C + N) min
AL 29-4C S44735 0.03(a) 28.0–30.0 1.0 Bal Mo: 3.6–4.2, N: 0.045(a), Ti + Cb:
0.2–1.0, 6(C + N) min
AL 29-4-2 S44800 0.01(a) 28.0–30.0 2.0–2.5 Bal Mo: 3.5–4.2, N: 0.02(a), C + N: 0.025(a)
18SR … 0.03 18.0 … Bal Al: 1.8, Ti: 0.4
12SR … 0.02 12.0 … Bal Al: 1.2, Cb: 0.6, Ti: 0.3
(continued)
(a) Maximum. Cb = Nb
Name ///sr-nova/Dclabs_wip/High Temp/5208_445-453.pdf/Appendix_2/ 26/10/2007 1:55PM Plate # 0 pg 446

446 / High-Temperature Corrosion and Materials Applications

Table 2 (Continued)
Alloy UNS No. C Cr Ni Fe Others
ALFA-I … 0.025 13.0 … Bal Al: 3.0, Ti: 0.4
ALFA-II … 0.025 13.0 … Bal Al: 4.0, Ti: 0.4
329 S32900 0.08(a) 23.0–28.0 2.5–5.0 Bal Mo: 1.0–2.0
URANUS 50 S32404 0.04(a) 20.5–22.5 5.8–8.5 Bal Mo: 2.0–3.0, Cu: 1.5, N: 0.2(a)
CD-4MCu J93370 0.04(a) 24.5–26.5 4.75–6.0 Bal Mo: 1.75–2.25, Cu: 3.0
44LN S31200 0.03(a) 24.0–26.0 5.5–6.5 Bal Mo: 1.2–2.0, N: 0.14–0.20
DP-3 S31260 0.03(a) 24.0–26.0 5.5–7.5 Bal Mo: 2.5–3.5, Cu: 0.2–0.8, N: 0.1–0.3, W: 0.1–0.4
3RE60 S31500 0.03(a) 18.0–19.0 4.25–5.25 Bal Mo: 2.5–3.0, N: 0.08–0.15, Si: 1.4–2.0
2205 S31803 0.03(a) 21.0–23.0 4.5–6.5 Bal Mo: 2.5–3.5, N: 0.08–0.2
FERRALIUM 255 S32550 0.04(a) 24.0–27.0 4.5–6.5 Bal Mo: 2.0–4.0, Cu: 1.4–2.5, N: 0.1–0.25
7-Mo: PLUS S32950 0.03(a) 26.0–29.0 3.5–5.2 Bal Mo: 1.0–2.5, N: 0.15–0.35
SUPEER-FERRIT … … 28.0 3.2 Bal Mo: 2.1
201 S20100 0.15(a) 16.0–18.0 3.5–5.5 Bal Mn: 5.5–7.5, N: 0.25(a)
202 S20200 0.15(a) 17.0–19.0 4.0–6.0 Bal Mn: 7.5–10.0, N: 0.25(a)
301 S30100 0.15(a) 16.0–18.0 6.0–8.0 Bal …
302 S30200 0.15(a) 17.0–19.0 8.0–10.0 Bal …
302B S30215 0.15(a) 17.0–19.0 8.0–10.0 Bal Si: 2.0–3.0
303 S30300 0.15(a) 17.0–19.0 8.0–10.0 Bal S: 0.15 min
303(Se) … 0.15(a) 17.0–19.0 8.0–10.0 Bal Se: 0.15 min
304 S30400 0.08(a) 18.0–20.0 8.0–10.5 Bal …
304L S30403 0.03(a) 18.0–20.0 8.0–12.0 Bal …
304H S30409 0.04–0.10 18.0–20.0 8.0–10.5 Bal …
305 S30500 0.12(a) 17.0–19.0 10.0–13.0 Bal …
308 S30800 0.08(a) 19.0–21.0 10.0–12.0 Bal …
309 S30900 0.2(a) 22.0–24.0 12.0–15.0 Bal …
309S S30908 0.08(a) 22.0–24.0 12.0–15.0 Bal …
310 S31000 0.25(a) 24.0–26.0 19.0–22.0 Bal …
310S S31008 0.08(a) 24.0–26.0 19.0–22.0 Bal …
314 S31400 0.25(a) 23.0–26.0 19.0–22.0 Bal Si: 1.5–3.0
316 S31600 0.08(a) 16.0–18.0 10.0–14.0 Bal Mo: 2.0–3.0
316L S31603 0.03(a) 16.0–18.0 10.0–14.0 Bal Mo: 2.0–3.0
316H S31609 0.04–0.10 16.0–18.0 10.0–14.0 Bal Mo: 2.0–3.0
317 S31700 0.08(a) 18.0–20.0 11.0–15.0 Bal Mo: 3.0–4.0
321 S32100 0.08(a) 17.0–19.0 9.0–12.0 Bal Ti: 5 × C min
321H S32109 0.04–0.10 17.0–19.0 9.0–12.0 Bal Ti: 5 × C min
347 S34700 0.08(a) 17.0–19.0 9.0–13.0 Bal Cb + Ta: 10 × C min
347H S34709 0.04–0.10 17.0–19.0 9.0–13.0 Bal Cb + Ta: 10 × C min
348 S34800 0.08(a) 17.0–19.0 9.0–13.0 Bal Cb + Ta: 10 × C min, Ta: 0.1(a)
253MA S30815 0.08 21.0 11.0 Bal Si: 1.7, N: 0.17, Ce: 0.04
RA85H S30615 0.2 18.5 14.5 Bal Si: 3.5, Al: 1.0
17-4PH S17400 0.04 16.5 4.25 Bal Cb: 0.25, Cu: 3.6
17-7PH S17700 0.07 17.0 7.0 Bal Al: 1.15
PH15-7Mo S15700 0.07 15.0 7.0 Bal Mo: 2.25, A1: 1.15
A286 K66286 0.08 13.5–16.0 24.0–27.0 Bal Mo: 1.0–1.5, Ti: 1.9–2.35, V: 0.1–0.5
AM 350 S35000 0.1 16.5 4.25 Bal Mo: 2.75, N: 0.1
AM 355 S35000 0.13 15.5 4.25 Bal Mo: 2.75, N: 0.12
16-18 … 0.05 16.0 19.0 Bal …
17-14CuMo … 0.12 16.0 14.0 Bal Mo: 2.5, Cb: 0.4, Ti: 0.3, Cu: 3.0
20-29CuMo … 0.05 20.0 29.0 Bal Mo: 2.20, Cu: 3.20
17-10P … 0.12 17.0 10.5 Bal P: 0.28
HMN … 0.30 18.5 9.5 Bal Mn: 3.5, P: 0.25
TENELON S21400 0.08 17.0 … Bal Mn: 14.5, N: 0.4
254SMO S31254 0.02 19.5–20.5 17.5–18.5 Bal Mo: 6.0–6.5, Cu: 0.5–1.0, N: 0.18–0.22
19-9 DL K63198 0.3 19.0 9.0 Bal Mo: 1.25, W: 1.25, Cb: 0.4
904L N08904 0.02 19.0–23.0 23.0–28.0 Bal Mo: 4.0–5.0, Cu: 1.0–2.0, V: 0.1–0.5
16-25-6 … 0.06 16.0 25.0 Bal Mo: 6.0
15-5PH S15500 0.07 15.0 4.5 Bal Cb: 0.3, Cu: 3.5
CUSTOM 450 S45000 0.05(a) 15.5 6.0 Bal Mo: 0.75, Cb: 8 × C min, Cu: 1.5
CUSTOM 455 S45500 0.03 11.75 8.5 Bal Cb: 0.3, Ti: 1.2, Cu: 2.25
AL-6XN N08367 0.03 20.0–22.0 23.5–25.5 Bal Mo: 6.0–7.0, N: 0.18–0.25
MVMA S30415 0.05 18.5 9.5 Bal Si: 1.3, N: 0.15, Ce: 0.04
22-4-9 … 0.5 21.5 4.0 Bal N: 0.4, S: 0.1
NITRONIC 60 S21800 0.05 17.0 8.5 Bal Si: 4.0, Mn: 8.0, N: 0.13
NITRONIC 50 S20910 0.03 22.0 12.5 Bal Mo: 2.0, Mn: 5.0, N: 0.3, Cb: 0.2, V: 0.2
NITRONIC 40 S21900 0.05 21.0 6.0 Bal Mn: 9.0
(21-6-9)
SNR-4 S31753 0.03 18.5 13.5 Bal Mo: 3.6
317LM S31725 0.03 18.0 15.0 Bal Mo: 4.1
17-14-4LM S31726 0.03 17.0 13.0 Bal Mo: 4.2, N: 0.15
JS700 N08700 0.03 21.0 25.0 Bal Mo: 4.5, Mn: 1.7
CRUTEMP: 25 … 0.05 25.0 25.0 Bal …
(a) Maximum. Cb = Nb
Name ///sr-nova/Dclabs_wip/High Temp/5208_445-453.pdf/Appendix_2/ 26/10/2007 1:55PM Plate # 0 pg 447

Appendix 2: Chemical Compositions of Alloys and Filler Metals / 447

Table 3 Chemical compositions (wt.%) of cast corrosion- and heat-resistant alloys

Alloy UNS No. C Cr Ni Fe Others
CA-15 … 0.15(a) 11.5–14.0 1.0(a) Bal Mo: 0.5(a)
CA-40 … 0.40(a) 11.5–14.0 1.0(a) Bal Mo: 0.5(a)
CB-30 … 0.30(a) 18–22 2.0(a) Bal …
CC-50 … 0.50(a) 26–30 4.0(a) Bal …
CE-30 … 0.30(a) 26–30 8–11 Bal …
CF-3 … 0.03(a) 18–21 8–11 Bal …
CF-3M … 0.03(a) 18–21 9–12 Bal Mo: 2.0–3.0
CF-8 … 0.08(a) 18–21 8–11 Bal …
CF-20 … 0.20(a) 18–21 8–11 Bal …
CF-8M … 0.08(a) 18–21 9–12 Bal Mo: 2.0–3.0
CF-12M … 0.12(a) 18–21 9–12 Bal Mo: 2.0–3.0
CF-8C … 0.08(a) 18–21 9–12 Bal Cb: 8 × C min, 1.0(a)
CF-16F … 0.16(a) 18–21 9–12 Bal Mo: 1.5(a), Se: 0.2–0.35
CG-8M … 0.08(a) 18–21 8–11 Bal Mo: 3.0 min
CH-20 … 0.20(a) 22–26 12–15 Bal …
CK-20 … 0.20(a) 23–27 19–22 Bal …
CN-7M … 0.07 19–22 28–31 Bal Mo: 2.0–3.0, Cu: 3.0–4.0
HA … 0.2(a) 8–10 … Bal …
HB … 0.3(a) 18–22 2(a) Bal …
HC J92605 0.5(a) 26–30 4(a) Bal …
HD J93005 0.5(a) 26–30 4–7 Bal …
HE J93403 0.2–0.5 26–30 8–11 Bal …
HF J92603 0.2–0.4 19–23 9–12 Bal …
HH J93503 0.2–0.5 24–28 11–14 Bal …
HI J94003 0.2–0.5 26–30 14–18 Bal …
HK J94224 0.2–0.6 24–28 18–22 Bal …
HL J94604 0.2–0.6 28–32 18–22 Bal …
HN J94213 0.2–0.5 19–23 23–27 Bal …
HT J94605 0.35–0.75 13–17 33–37 Bal …
HU J95405 0.35–0.75 17–21 37–41 Bal …
HW N08001 0.35–0.75 10–14 58–62 Bal …
HX N06006 0.35–0.75 15–19 64–68 Bal …
HP J95705 0.4 25 35 Bal …
THERMAX 40B … 0.4 25 13 Bal W: 0.5
THERMAX 47 … 0.45 25 20 Bal W: 0.5
THERMAX 63 … 0.45 23 35 Bal W: 0.5
THERMAX 63WC … 0.45 25 35 Bal Co: 15.0, W: 5.0
THERMALLOY 63W … 0.45 25 36 Bal W: 5.0
THERMALLOY 63WC … 0.5 28 36 Bal Co: 15.0, W: 5.0
(SUPERTHERM)
MANAURITE XU … 0.35–0.45 24–27 32–36 Bal Nb: 0.5, W: 0.5
MANAURITE 36X … 0.4 25 34 Bal Cb: 1.2
MANAURITE 36XS … 0.4 25 34 Bal Cb: 1.5, W: 1.5
MANAURITE 900 … 0.13 21 33 Bal Cb: 1.2
MANAURITE XT … 0.4 35 44 Bal Cb: 1.5, W: 1.5
MO-RE 1 … 0.45 26 Bal 33 W: 1.6
HOM-3 … 0.45 26 Bal 16 Mo: 3.0, W: 3.0, Co: 3.0
22-H (NA 22H or HOM-5) … 0.5 28 Bal 16 W: 5.0
SUPER 22-H … 0.5 28 Bal 13 Co: 3.0, W: 5.0
IN-657 … 0.06 48 Bal … Cb: 1.5
(a) Maximum. Cb = Nb
Name ///sr-nova/Dclabs_wip/High Temp/5208_445-453.pdf/Appendix_2/ 26/10/2007 1:55PM Plate # 0 pg 448

448 / High-Temperature Corrosion and Materials Applications

Table 4 Chemical compositions (wt.%) of wrought iron-, nickel-, and cobalt-base alloys
Alloy UNS No. C Cr Ni Co Fe Mo W Others
INCOLOY 800 N08800 0.05 21 32.5 … Bal … … Al: 0.3, Ti: 0.3
INCOLOY 800H N08810 0.08 21 32.5 … Bal … … Al: 0.4, Ti: 0.4
INCOLOY 800HT N08811 0.08 21 32.5 … Bal … … Al + Ti: 1.0
INCOLOY 802 N08802 0.4 21 32.5 … Bal … … …
INCOLOY 903 N19903 … … 38 15 Bal … … Ti: 1.4, Al: 0.9, Cb: 3.0
INCOLOY 904 … … … 32.5 14.5 Bal … … Ti: 1.6
INCOLOY 907 N19907 … … 38 13 Bal … … Ti: 1.5, Cb: 4.7, Si: 0.15
INCOLOY 909 N19909 … … 38 13 Bal … … Ti: 1.5, Cb: 4.7, Si: 0.4
INCOLOY DS … 0.06 17 35 … Bal … … Si: 2.3
RA 330 N08330 0.05 19 35 … Bal … … Si: 1.2
RA 330HC … 0.4 19 35 … Bal … … Si: 1.2
AC66 N33228 0.05 28 32 … Bal … … Nb: 0.8, Ce: 0.07
SANICRO 28 N08028 0.01 27 31 … Bal 3.5 … Cu: 1.0
20Cb-3 N08020 0.02 20 33 … Bal 2.2 … Cu: 3.3, Cb: 0.5
20Mo-4 N08024 0.02 23.5 37 … Bal 3.8 … Cu: 1.0, Cb: 0.25
20Mo-6 N08026 0.02 24 36 … Bal 5.6 … Cu: 3.0
Nicrofer 3033 (alloy 33) R20033 0.015(a) 33 32 … Bal 1.25 … Cu: 0.75, N: 0.5
HAYNES HR-120 … 0.05 25 37 … Bal … … Cb: 0.7, N: 0.2
HAYNES 556 R30556 0.1 22 20 18 Bal 3.0 2.5 Ta: 0.6, La: 0.02, N: 0.2, Zr: 0.02
MULTIMENT alloy R30155 0.1 21 20 20 Bal 3.0 2.5 Cb + Ta: 1.0, N: 0.15
(N-155)
V-57 … 0.8(a) 14.8 27.0 … Bal 1.25 … Al: 0.25, Ti: 3.0, V: 0.5(a), B: 0.01
W-545 K66545 0.08 13.5 26.0 … Bal 1.5 … Al: 0.2, Ti: 2.85, B: 0.05
DISCALOY K66220 0.06 14.0 26.0 … Bal 3.0 … Al: 0.25, Ti: 1.7
PYROMET CTX-1 … 0.03 … 37.7 16.0 Bal … … Cb: 3.0, Al: 1.0, Ti: 1.7
CHROMEL D … … 18.5 36.0 … Bal … … Si: 1.5
KANTHAL Al K92500 … 22.0 … … Bal … … Al: 5.8
KANTHAL AF … … 22.0 … … Bal … … Al: 5.3, Y
FECRALLOY A … 0.03 15.8 … … Bal … … Al: 4.8, Y: 0.3
Aluchrom S … 0.08(a) 20 … … Bal … … Al: 4.5, Zr: 0.3(a)
Aluchrom ISE … 0.10(a) 20 … … Bal … … Al: 5.0
Aluchrom Y … 0.01–0.1 21 … … Bal … … Al: 5.5, Zr: 0.01–0.1, Y: 0.05–0.15
Aluchrom O … 0.08(a) 22 … … Bal … … Al: 5.5, Zr: 0.3(a)
Aluchrom PSI …- 0.015–0.03 22.5 … … Bal … … Al: 5.6, Hf: 0.3
Ni 200 N02200 0.08 … 99.6 … Bal … … …
Ni 201 N02201 0.02(a) … 99.6 … Bal … … …
Ni 270 N02270 0.01 … 99.98 … Bal … … …
MONEL 400 N04400 … … Bal … 1.2 … … Cu: 31.5, Mn: 1.1
MONEL 401 N04401 … … Bal … 0.3 … … Cu: 55.5, Mn: 1.63
MONEL R-405 N04405 … … Bal … 1.2 … … Cu: 31.5, Mn: 1.1
MONEL K-500 N05500 … … Bal … 1.0 … … Cu: 29.5, Ti: 0.6, Al: 2.7
MONEL 450 C71500 … … Bal … 0.7 … … Cu: 68.0, Mn: 0.7
FERRY alloy … … … Bal … … … … Cu: 55.0
CUPRO 107 … … … Bal … 0.8 … … Cu: 68.0, Mn: 1.1
INCONEL 600 N06600 0.08(a) 15.5 Bal … 8.0 … … …
INCONEL 601 N06601 0.10(a) 23.0 Bal … 14.4 … … Al: 1.4
Nicrofer 6025HT N06025 0.2 25 Bal … 10 … … Al: 2.1, Ti: 0.15, Y: 0.05–0.12,
(602CA) Zr: 0.01–0.1
INCONEL 617 N06617 0.07 22.0 Bal 12.5 1.5 9.0 … A1: 1.2
INCONEL 625 N06625 0.10(a) 21.5 Bal … 2.5 9.0 … Cb: 3.6
INCONEL 690 N06690 0.02 29.0 Bal … 9.0 … … …
INCONEL 693 N06693 0.2 29.0 Bal … 4.0 … … Al: 2.5–4.0, Nb: 0.5–2.5
INCOLOY 825 N08825 0.03 21.5 Bal … 30.0 3.0 … Cu: 2.2
INCOLOY 890 N08890 0.1 25 42.5 … Bal 1.5 … Ta: 0.2
INCOLOY 925 N09925 0.01 21.0 Bal … 28.0 3.0 … Cu: 1.8, Ti: 2.1, Al: 0.3
INCONEL 706 N09706 0.03 16.0 Bal … 37.0 … … Ti: 1.8, Al: 0.2, Cb: 2.9
INCONEL 718 N07718 0.04 18.0 Bal … 18.5 3.0 … Cb: 5.1
INCONEL X-750 N07750 0.04 15.5 Bal … 7.0 … … Ti: 2.5, Al: 0.7, Cb: 1.0
INCONEL 751 N07751 0.05 15.0 Bal … 7.0 … … Ti: 2.5, Al: 1.1, Cb: 1.0
INCONEL 671 … 0.05 48.0 Bal … … … … Ti: 0.35
INCONEL 686 N06686 0.01(a) 21.0 Bal … 5.0(a) 16.0 4.0 Ti: 0.02–0.25
Nicrofer 5923 (59) N06059 0.01(a) 23 Bal … 1.5(a) 15.5 … …
HAYNES 214 … 0.04 16.0 Bal … 3.0 … … Al: 4.5, Y
HAYNES 230 N06230 0.1 22.0 Bal … 3.0(a) 2.0 14.0 La: 0.02, B: 0.015(a)
HAYNES 242 … 0.03(a) 8.0 Bal … … 25.0 … …
HAYNES HR-160 N12160 0.05 28.0 Bal 29.0 1.5 … … Si: 2.75, Ti: 0.5, Nb: 1.0(a)
Nicrofer 45TM N06045 0.1 27 Bal … 23 … … Si: 2.5–3.0, Ce: 0.03–0.09
HASTELLOY X N06002 0.1 22.0 Bal 1.5 18.5 9.0 0.6 …
HASTELLOY W N10004 0.12(a) 5.0 Bal 2.5 6.0 24.0 … …
HASTELLOY S N06635 0.02(a) 15.5 Bal … 3.0(a) 14.5 … La: 0.05, B: 0.015(a)
HASTELLOY N N10003 0.06 7.0 Bal … 5.0(a) 16.5 … …
HASTELLOY C-22 N06022 0.01(a) 22.0 Bal 2.5(a) 3.0 13.0 3.0 …
(continued)
(a) Maximum. Cb = Nb
Name ///sr-nova/Dclabs_wip/High Temp/5208_445-453.pdf/Appendix_2/ 26/10/2007 1:55PM Plate # 0 pg 449

Appendix 2: Chemical Compositions of Alloys and Filler Metals / 449

Table 4 (Continued)
Alloy UNS No. C Cr Ni Co Fe Mo W Others
HASTELLOY C-276 N10276 0.01(a) 15.5 Bal 2.5(a) 5.5 16.0 4.0 …
HASTELLOY C-2000 N06200 0.01(a) 23 Bal … 3.0(a) 16 … Cu: 1.6
HASTELLOY C-4 N06455 0.01(a) 16.0 Bal 2.0(a) 3.0(a) 15.5 … …
HASTELLOY C N10002 0.08(a) 15.5 Bal 2.5(a) 6.0 17.0 4.0 …
HASTELLOY B N10001 0.05(a) … Bal 2.5(a) 5.0 28.0 … V: 0.03
HASTELLOY B-2 N10665 0.01(a) … Bal … 2.0(a) 28.0 … …
HASTELLOY B-3 N10675 0.01(a) 2 Bal … 2 27–32 … …
NICROFER 6224 (B-10) … 0.01(a) 8 Bal … 7 23 … …
HASTELLOY G N06007 0.05(a) 22.0 Bal … 19.5 6.5 … Cb + Ta: 2.0, Cu: 2.0
HASTELLOY G-3 N06985 0.015(a) 22.0 Bal 5.0(a) 19.5 7.0 1.5(a) Cb + Ta: 0.3, Cu: 2.0
HASTELLOY G-30 N06030 0.03(a) 29.5 Bal 5.0(a) 15.0 5.0 2.5 Cu: 2.0
HASTELLOY G-35 N06035 0.05(a) 33 Bal … 2.0(a) 8 … …
HASTELLOY G-50 N06950 0.02(a) 20 Bal 2.5(a) 17 9 … …
HASTELLOY H-9M … 0.03(a) 22.0 Bal 5.0(a) 19.0 9.0 2.0 …
RA 333 N06333 0.05 25.0 Bal 3.0 18.0 3.0 3.0 …
CHROMEL A (or … … 20.0 Bal … … … … Si: 1.0
NICHROME 80)
NA 224 … 0.5 27.0 Bal … 18.5 … 6.0 …
NIMONIC 70 … … 20.0 Bal … 25.0 … … Al: 1.0, Ti: 1.25, Cb: 1.5
NIMONIC 75 … 0.10 19.5 Bal … … … … …
NIMONIC 80A N07080 0.06 19.5 Bal … … … … Al: 1.4, Ti: 2.4
NIMONIC 81 … 0.03 30.0 Bal … … … … Al: 0.9, Ti: 1.8
NIMONIC 86 … … 25.0 Bal … … 10.0 … Ce: 0.03
NIMONIC 90 N07090 0.07 19.5 Bal 16.5 … … … Al: 1.5, Ti: 2.5
NIMONIC 91 … … 28.5 Bal 20.0 … … … Al: 1.2, Ti: 2.3
NIMONIC 105 … 0.08 15.0 Bal 20.0 … 5.0 … Al: 4.7, Ti: 1.3, B: 0.005
NIMONIC 115 … 0.15 15.0 Bal 15.0 … 4.0 … Al: 5.0, Ti: 4.0
NIMONIC 901 … … 12.5 Bal … 36.0 5.8 … Ti: 2.9
NIMONIC AP 1 … … 15.0 Bal 17.0 … 5.0 … Al: 4.0, Ti: 3.5
NIMONIC PE 11 … 0.05 18.0 Bal … 34.0 5.2 … Al: 0.8, Ti: 2.3
NIMONIC PE 16 … 0.05 16.5 Bal … 34.0 3.3 … Al: 1.2, Ti: 1.2
NIMONIC PK 31 … … 20.0 Bal 14.0 … 4.5 … Al: 0.4, Ti: 2.35, Cb: 5.0
NIMONIC PK 33 … 0.04 18.0 Bal 14.0 … 7.0 … Al: 2.1, Ti: 2.4
NIMONIC PK 50 … … 19.5 Bal 13.5 … 4.25 … Al: 1.4, Ti: 3.0
NIMONIC PK 37 … … 19.5 Bal 16.5 … … … Al: 1.5, Ti: 2.5
WASPALOY alloy N07001 0.08 19.0 Bal 14.0 … 4.3 … Al: 1.5, Ti: 3.0, Zr: 0.05, B: 0.006
263 … 0.06 20.0 Bal 20.0 … 5.8 … Al: 0.5, Ti: 2.2
HAYNES 282 … 0.06 19.5 Bal 10 1.5(a) 8.5 … Al: 1.5, Ti: 2.1, B: 0.005
RENÉ 41 N07041 0.09 19.0 Bal 11.0 5.0(a) 10.0 … Al: 1.5, Ti: 3.0, B: 0.006
RENÉ 95 … 0.15 14.0 Bal 8.0 … 3.5 3.5 Cb: 3.5, Al: 3.5, Ti: 2.5, Zr: 0.05
RENÉ 100 … 0.16 9.5 Bal 15.0 … 3.0 … Al: 5.5, Ti: 4.2, Zr: 0.06, B: 0.015
UDIMET 400 … 0.06 17.5 Bal 14.0 … 4.0 … Cb: 0.5, Al: 1.5, Ti: 2.5, Zr: 0.06, B: 0.008
UDIMET 500 … 0.08 18.0 Bal 18.5 … 4.0 … Al: 2.9, Ti: 2.9, Zr: 0.05, B: 0.006
UDIMET 520 … 0.05 19.0 Bal 12.0 … 6.0 1.0 Al: 2.0, Ti: 3.0, B: 0.005
UDIMET 630 … 0.03 18.0 Bal … 18.0 3.0 3.0 Cb: 6.5, Al: 0.5, Ti: 1.0
UDIMET 700 … 0.03 15.0 Bal 18.5 … 5.2 … Al: 5.3, Ti: 3.5, B: 0.03
UDIMET 710 … 0.07 18.0 Bal 15.0 … 3.0 1.5 Al: 2.5, Ti: 5.0
UDIMET 720 … 0.03 17.9 Bal 14.7 … 3.0 1.3 Al: 2.5, Ti: 5.0, Zr: 0.03, B: 0.033
UNITEMP AF2-IDA … 0.35 12.0 Bal 10.0 … 3.0 6.0 Ta: 1.5, Al: 4.6, Ti: 3.5, Zr: 0.1
UNITEMP AF2-ID6 … 0.04 12.0 Bal 10.0 … 2.7 6.5 Ta: 1.5, Al: 4.0, Ti: 2.8, Zr: 0.1, B: 0.015
ASTROLOY … 0.06 15.0 Bal 17.0 … 5.3 … Al: 4.0, Ti: 3.5, B: 0.03
D-979 N09979 0.05 15.0 Bal … 27.0 4.0 4.0 Al: 1.0, Ti: 3.0
IN 100 N13100 0.15 10.0 Bal 15.0 … 3.0 … Al: 5.5, Ti: 4.7, Zr: 0.06, V: 1.0, B: 0.015
IN 102 N06102 0.06 15.0 Bal … 7.0 3.0 3.0 Cb: 3.0, Al: 0.4, Ti: 0.6, Mg: 0.02, Zr: 0.03
IN 587 … 0.05 28.5 Bal 20.0 … … … Cb: 0.7, Al: 1.2, Ti: 2.3, Zr: 0.5
IN 597 … 0.05 24.5 Bal 20.0 … 1.5 … Cb: 1.0, Al: 1.5, Ti: 3.0, Zr: 0.5
M 252 N07252 0.15 20.0 Bal 10.0 … 10.0 … Al: 1.0, Ti: 2.6, B: 0.005
PYROMET 31 N07031 0.04 22.5 Bal … 15.0 2.0 … Al: 1.4, Ti: 2.3, Cu: 0.9, B: 0.005
PYROMET 860 … 0.05 13.0 Bal 4.0 28.9 6.0 … Al: 1.0, Ti: 3.0, B: 0.01
REFRACTORY 26 … 0.03 18.0 Bal 20.0 16.0 3.2 … Al: 0.2, Ti: 2.6, B: 0.015
625 PLUS N07716 0.02 20.0 Bal … 5.0 9.0 … Cb: 3.1, Al: 0.2, Ti: 1.3
IN 100 GATORIZE … 0.07 12.4 Bal 18.5 … 3.2 … Al: 5.0, Ti: 4.3, Zr: 0.06, B: 0.02, V: 0.8
HAYNES 188 R30188 0.10 22.0 22.0 Bal 3.0(a) … 14.0 La: 0.04
HAYNES 25 (L-605) R30605 0.10 20.0 10.0 Bal 3.0(a) … 15.0 …
HAYNES 150 (UMCo-50) … 0.06 27.0 … Bal 18.0 … … …
HAYNES 6B … 1.20 30.0 … Bal … 1.5(a) 4.5 …
S-816 R30816 0.38 20.0 20.0 Bal 4.0 4.0 4.0 Cb: 4.0
MAR-M 918 … 0.05 20.0 20.0 Bal … … … Ta: 7.5, Zr: 0.10
MP 35N R30035 … 20.0 35.0 Bal … 10.0 … …
MP 159 … … 19.0 25.5 Bal 9.0 7.0 … Cb: 0.6, Al: 0.2, Ti: 3.0
AR 213 … 0.17 19.0 … Bal … … 4.5 Al: 3.5, Ta: 6.5, Zr: 0.15, Y: 0.1
ULTIMET alloy R31233 0.06 26.0 9.0 Bal 3.0 5.0 2.0 N: 0.08
(a) Maximum. Cb = Nb
Name ///sr-nova/Dclabs_wip/High Temp/5208_445-453.pdf/Appendix_2/ 26/10/2007 1:55PM Plate # 0 pg 450

450 / High-Temperature Corrosion and Materials Applications

Table 5 Chemical compositions (wt.%) of cast nickel- and cobalt-base alloys

Alloy C Cr Ni Co Fe Mo W Ta Zr Others
B-1900 0.10 8.0 Bal 10.0 … 6.0 … 4.0 0.10 Al: 6.0, Ti: 1.0, B: 0.015
IN 100 0.18 10.0 Bal 15.0 … 3.0 … … 0.06 Al: 5.5, Ti: 5.0, B: 0.01, V: 1.0
IN 162 0.12 10.0 Bal … … 4.0 2.0 2.0 0.10 Al: 6.5, Ti: 1.0, Cb: 1.0, B: 0.02
IN 731 0.18 9.5 Bal 10.0 … 2.5 … … 0.06 Al: 5.5, Ti: 4.6, B: 0.015, V: 1.0
IN 738 0.17 16.0 Bal 8.5 … 1.7 2.6 1.7 0.10 Al: 3.4, Ti: 3.4, Cb: 0.9, B: 0.01
IN 792 0.12 12.4 Bal 9.0 … 1.9 3.8 3.9 0.10 Al: 3.1, Ti: 4.5, B: 0.02
IN 713C 0.12 12.5 Bal … … 4.2 … … 0.10 Al: 6.0, Ti: 0.8, Cb: 2.0, B: 0.012
IN 713LC 0.05 12.0 Bal … … 4.5 … … 0.10 Al: 6.0, Ti: 0.6, Cb: 2.0, B: 0.01
M 21 0.13 5.7 Bal … … 2.0 11.0 … 0.12 Al: 6.0, Cb: 1.5, B: 0.02
M 22 0.13 5.7 Bal … … 2.0 11.0 3.0 0.60 Al: 6.3
M 252 0.15 20.0 Bal 10.0 … 10.0 … … … Al: 1.0, Ti: 2.6, B: 0.005
MAR-M 200 0.15 9.0 Bal 10.0 1.0 … 12.5 … 0.05 Al: 5.0, Ti: 2.0, B: 0.015, Cb: 1.0
MAR-M 246 0.15 9.0 Bal 10.0 … 2.5 10.0 1.5 0.05 Al: 5.5, Ti: 1.5, B: 0.015
MAR-M 247 0.15 8.25 Bal 10.0 0.5 0.7 10.0 3.0 0.05 Al: 5.5, Ti: 1.0, Hf: 1.5, B: 0.015
MAR-M 421 0.15 15.8 Bal 9.5 … 2.0 3.8 … 0.05 Al: 4.3, Ti: 1.8, Cb: 2.0, B: 0.015
MAR-M 432 0.15 15.5 Bal 20.0 … … 3.0 2.0 0.05 Al: 2.8, Ti: 4.3, Cb: 2.0, B: 0.015
MC 102 0.04 20.0 Bal … … 6.0 2.5 0.6 … Cb: 6.0
RENÉ 77 0.07 15.0 Bal 15.0 … 4.2 … … 0.04 Al: 4.3, Ti: 3.3, B: 0.015
RENÉ 80 0.17 14.0 Bal 9.5 … 4.0 4.0 … 0.03 Al: 3.0, Ti: 5.0, B: 0.015
TAZ-8A 0.12 6.0 Bal … … 4.0 4.0 8.0 1.0 Al: 6.0, Cb: 2.5, B: 0.004
TAZ-8B: (DS) 0.12 6.0 Bal 5.0 … 4.0 4.0 8.0 1.0 Al: 6.0, Cb: 1.5, B: 0.004
TRW-NASA-VIA 0.13 6.0 Bal 7.5 … 2.0 6.0 9.0 0.13 Al: 5.5, Ti: 1.0, Cb: 0.5, Hf: 0.4,
Re: 0.5, B: 0.02
WAZ-20 (DS) 0.20 … Bal … … … 20.0 … 1.5 Al: 6.5
DELORO 40 0.2 7.5 Bal 1.5(a) 2.5 … … … … Si: 3.5
DELORO 50 0.45 10.5 Bal … 3.5 … … … … Si: 4.0, B: 2.0
DELORO 60 0.7 14.5 Bal … 4.0 … … … … Si: 2.0–4.5, B: 3.3
AIRESIST 13 0.45 21.0 … Bal … … 11.0 2.0 … Al: 3.4, Y: 0.1
AIRESIST 213 0.20 20.0 0.5 Bal 0.5 … 4.5 6.5 0.1 Al: 3.5, Y: 0.1
AIRESIST 215 0.35 19.0 0.5 Bal 0.5 … 4.5 7.5 0.1 Al: 4.3, Y: 0.1
FSX 414 0.25 29.0 10.0 Bal 1.0 … 7.5 … … B: 0.01
J-1650 0.20 19.0 27.0 Bal … … 12.0 2.0 … Ti: 3.8, B: 0.02
MAR-M 302 0.85 21.5 … Bal 0.5 … 10.0 9.0 0.2 B: 0.005
MAR-M 322 1.00 21.5 … Bal 0.5 … 9.0 4.5 2.0 Ti: 0.75
MAR-M 509 0.60 23.5 10.0 Bal … … 7.0 3.5 0.5 Ti: 0.2
MAR-M 918 0.05 20.0 20.0 Bal … … … 7.5 0.1 …
NASA Co-W-Re 0.40 3.0 … Bal … … 25.0 … 1.0 Re: 2.0, Ti: 1.0
S-816 0.40 20.0 20.0 Bal 4.0 4.0 4.0 … … Cb: 4.0
V-36 0.27 25.0 20.0 Bal 3.0 4.0 2.0 … … Cb: 2.0
WI-52 0.45 21.0 … Bal 2.0 … 11.0 … … Cb + Ta: 2.0
X-40 (HAYNES 31) 0.50 22.0 10.0 Bal 1.5 … 7.5 … … …
Stellite 1 2.45 31 3(a) Bal 2.5(a) … 13 … … …
Stellite 6 1.2 28 3(a) Bal 3(a) … 4.5 … … …
Stellite 12 1.6 29.5 3(a) Bal 2.5(a) … 8.5 … … …
Stellite 21 0.25 27 2.5 Bal 3(a) 5.5 … … … …
Stellite 6K 1.6 31 3.0(a) Bal 3(a) … 4.5 … … …
Stellite 704 1 30 2(a) Bal 2(a) 14 … … … …
Stellite 706 1.2 29 3(a) Bal 3(a) 4.5 … … … …
Stellite 712 1.6 29 3(a) Bal 3(a) 8.5 … … … …
Stellite 706K 1.6 31 3(a) Bal 3(a) 4.5 … … … …
Tribaloy T-400C 0.1(a) 14 1(a) Bal 1(a) 27 … … … Si: 2.6
Tribaloy T-401 0.2 17 0.8(a) Bal 0.8(a) 22 … … … Si: 1.3
(a) Maximum. Cb = Nb

Table 6 Chemical compositions (wt.%) of oxide-dispersion-strengthened (ODS) alloys

Alloy C Ni Cr Fe Ti Al Others
INCOLOY MA 956 … … 20.0 Bal 0.5 4.5 Y2O3: 0.5
INCONEL MA 754 0.05 Bal 20.0 1.0 0.5 0.3 Y2O3: 0.6
INCONEL MA 758 0.05 Bal 30.0 … 0.5 0.3 Y2O3: 0.6
INCONEL MA 6000 0.05 Bal 15.0 … 2.5 4.5 Y2O3: 1.1, Mo: 2.0, W: 4.0, Ta: 2.0
Name ///sr-nova/Dclabs_wip/High Temp/5208_445-453.pdf/Appendix_2/ 26/10/2007 1:55PM Plate # 0 pg 451

Appendix 2: Chemical Compositions of Alloys and Filler Metals / 451

Table 7 Chemical compositions of iron-, nickel- and cobalt-base filler metals

Filler Metal AWS UNS C Cr Ni Fe Co Others
409 ER409 S40900 0.08(a) 10.5–13.5 0.6(a) Bal … Ti: 10 × C min/1.5 max
409Nb ER409Nb S40940 0.08(a) 10.5–13.5 0.6 (a) Bal … Nb: 10 × C min/0.75 max
410 ER410 S41080 0.12(a) 11.5–13.5 0.6 (a) Bal … …
410NiMo ER410NiMo S41086 0.06(a) 11.0–12.5 4.0–5.0 Bal … …
420 ER420 … 0.25–0.40 12.0–14.0 0.6(a) Bal … …
430 ER430 … 0.1(a) 15.5–17.0 0.6(a) Bal … …
26-1 ER26-1 … 0.01(a) 25.0–27.5 … Bal … N: 0.015
630 ER630 … 0.05(a) 16.0–16.75 4.5–5.0 Bal … Cu: 3.25–4.0,
Nb + Ta: 0.15–0.30,
16-8-2 ER16-8-2 … 0.1(a) 14.5–16.5 7.5–9.5 Bal … Mo: 1.0–2.0
Nitronic 50W ER209 S20980 0.05(a) 20.5–24.0 9.5–12.0 Bal … Mo: 1.5–3.0, Mn: 4.0–7.0,
N: 0.2
Nitronic 60W ER218 S21880 0.1(a) 16.0–18.0 8.0–9.0 Bal … Mn: 8.0, Si: 3.5–4.5,
N: 0.08–0.18
219 ER219 S21980 0.05(a) 19.0–21.5 5.5–7.0 Bal … Mn: 9.0, N: 0.2
240 ER240 S24080 0.05(a) 17.0–19.0 4.0–6.0 Bal … Mn: 10.5–13.5
307 ER307 S30780 0.04–0.14 19.5–22.0 8.0–10.7 Bal … Mo: 0.5–1.5, Mn: 3.3–4.75
308 ER308 S30880 0.08(a) 19.5–22.0 9.0–11.0 Bal … …
308H ER308H S30880 0.04–0.08 19.5–22.0 9.0–11.0 Bal … …
308L ER308L S30883 0.03(a) 19.5–22.0 9.0–11.0 Bal … …
308Mo ER308Mo S30882 0.08(a) 18.0–21.0 9.0–12.0 Bal … Mo: 2.0–3.0
308LMo ER308LMo S30886 0.04(a) 18.0–21.0 9.0–12.0 Bal … Mo: 2.0–3.0
308Si ER308Si S30881 0.08(a) 19.5–22.0 9.0–11.0 Bal … Si: 0.65–1.0
308LSi ER308LSi S30888 0.03(a) 19.5–22.0 9.0–11.0 Bal … Si: 0.65–1.0
309 ER309 S30980 0.12(a) 23.0–25.0 12.0–14.0 Bal … …
309L ER309L S30983 0.03(a) 23.0–25.0 12.0–14.0 Bal … …
309Mo ER309Mo S30982 0.12(a) 23.0–25.0 12.0–14.0 Bal … Mo: 2.0–3.0
309LMo ER309LMo S30986 0.03(a) 23.0–25.0 12.0–14.0 Bal … Mo: 2.0–3.0
309Si ER309Si S30981 0.12(a) 23.0–25.0 12.0–14.0 Bal … Si: 0.65–1.0
309LSi ER309LSi S30988 0.03(a) 23.0–25.0 12.0–14.0 Bal … Si: 0.65–1.0
310 ER310 S31080 0.08–0.15 25.0–28.0 20.0–22.5 Bal … …
312 ER312 S31280 0.15(a) 28.0–32.0 8.0–10.5 Bal … …
316 ER316 S31680 0.08(a) 18.0–20.0 11.0–14.0 Bal … Mo: 2.0–3.0
316H ER316H S31680 0.04–0.08 18.0–20.0 11.0–14.0 Bal … Mo: 2.0–3.0
316L ER316L S31683 0.03(a) 18.0–20.0 11.0–14.0 Bal … Mo: 2.0–3.0
316Si ER316Si S31681 0.08(a) 18.0–20.0 11.0–14.0 Bal … Mo: 2.0–3.0
316LSi ER316LSi S31688 0.03(a) 18.0–20.0 11.0–14.0 Bal … Mo: 2.0–3.0
317 ER317 S31780 0.08(a) 18.5–20.5 13.0–15.0 Bal … Mo: 3.0–4.0
317L ER317L S31783 0.03(a) 18.5–20.5 13.0–15.0 Bal … Mo: 3.0–4.0
321 ER321 S32180 0.08(a) 18.5–20.5 9.0–10.5 Bal … Ti: 9 × C min/1.0 max
347 ER347 S34780 0.08(a) 19.0–21.5 9.0–11.0 Bal … Nb: 10 × C min/1.0 max
347Si ER347Si S34788 0.08(a) 19.0–21.5 9.0–11.0 Bal … Nb: 10 × C min/1.0 max,
Si: 0.65–1.0
318 ER318 S31880 0.08(a) 18.0–20.0 11.0–14.0 Bal … Mo: 2.0–3.0,
Nb: 8 × C min/1.0 max
320 ER320 N08021 0.07(a) 19.0–21.0 32.0–36.0 Bal … Mo: 2.0–3.0, Cu: 3.0–4.0,
Nb: 8 × C min/1.0 max
320LR ER320LR N08022 0.025(a) 19.0–21.0 32.0–36.0 Bal … Mo: 2.0–3.0, Cu: 3.0–4.0,
Nb: 8 × C min/0.4 max
383 ER383 N08028 0.025(a) 26.5–28.5 30.0–33.0 Bal … Mo: 3.2–4.2, Cu: 0.7–1.5
904L ER385 N08904 0.025(a) 19.5–21.5 24.0–26.0 Bal … Mo: 4.2–5.2, Cu: 1.2–2.0
2209 ER2209 S39209 0.02(a) 22.5 8 Bal … Mo: 3, N: 0.14
2507 … S32750 0.02(a) 25 10 Bal … Mo: 4, N: 0.25
253MA … S30815 0.07 21 10 Bal … Si: 1.6, N: 0.16, Ce: 0.05
RA85H … S30615 0.2 18.5 14.5 Bal … Si: 3.7, Al: 1.0
RA330 … N08334 0.25 19 35 Bal … Mn: 5.25
Marathon … … 0.12 21.5 32 Bal … Nb: 1.7
21/33
353MA … … 0.03 28 34 Bal … N: 0.15, Ce: 0.03
33 … R20033 0.015(a) 31.0–35.0 30.0–33.0 Bal … Mo: 0.5–2.0, Cu: 0.3–1.2,
N: 0.35–0.6
556 ER3556 R30556 0.1 22 20 Bal 18 Mo: 3, W: 2.5, Ta: 0.6,
N: 0.2, Zr: 0.02, La: 0.02
MULTIMET … R30155 0.12 21 20 Bal 20 Mo: 3, W: 2.5, Nb + Ta: 1,
(N-155) N: 0.15
NOREM 02A … … 1.10–1.35 23.0–26.0 3.7–5.0 Bal … Mn: 4.5, Si: 3.3, Mo: 2.0
NOREM 03A … … 0.9–1.2 20.0–23.5 4.0–5.0 Bal … Mn: 4.0, Si: 2.8,
Mo: 1.7–2.2
61 ERNi-1 N02061 0.15(a) … 93.0 min 1(a) … Ti: 2.0–3.5, Al: 1.5(a)
67 ERCuNi C71581 0.04(a) … 29.0–32.0 0.40–0.70 … Cu: Bal, Ti: 0.20–0.50
(continued)
(a) Maximum. Nb = Cb
Name ///sr-nova/Dclabs_wip/High Temp/5208_445-453.pdf/Appendix_2/ 26/10/2007 1:55PM Plate # 0 pg 452

452 / High-Temperature Corrosion and Materials Applications

Table 7 (Continued)
Filler Metal AWS UNS C Cr Ni Fe Co Others
60 ERNiCu-7 N04060 0.15(a) … 62.0–69.0 2.5(a) … Cu: Bal, Mn: 4.0(a),
Ti: 1.5–3.0, Al: 1.25(a)
… ERNiCu-8 N05504 0.25(a) … 63.0–70.0 2(a) … Cu: Bal, Ti: 0.25–1.00,
Al: 2.0–4.0
82 ERNiCr-3 N06082 0.1(a) 18.0–22.0 67.0 min 3(a) … Nb + Ta: 2.5, Mn: 3.0
72 ERNiCr-4 N06072 0.01–0.10 42.0–46.0 Bal 0.5(a) … Ti: 0.3–1.0
76 ERNiCr-6 N06076 0.08–0.15 19.0–21.0 75.0 min 2(a) … Ti: 0.15–0.50
62 ERNiCrFe-5 N06062 0.08(a) 14.0–17.0 70.0 min 6.0–10.0 … Nb + Ta: 1.5–3.0
92 ERNiCrFe-6 N07092 0.08(a) 14.0–17.0 67.0 min 8(a) … Ti: 3.0, Mn: 2.0–2.7
52 ERNiCrFe-7 N06052 0.04(a) 28.0–31.5 Bal 7.0–11.0 … Ti: 1.0(a), Al: 1.1(a)
52M ERNiCrFe-7A N06054 0.04(a) 28.0–31.5 Bal 7.0–11.0 … Ti: 1.0(a), Al: 1.1(a),
Nb + Ta: 0.50–1.0
… ERNiCrFe-8 N07069 0.08(a) 14.0–17.0 70.0 min 5.0–9.0 … Ti: 2.0–2.75, Al: 0.4–1.0,
Nb + Ta: 0.7–1.2
53MD ERNiCrFeAl-1 N06693 0.15(a) 27.0–31.0 Bal 2.5–6.0 … Al: 2.5–4.0, Nb + Ta:
0.5–2.5, Ti: 1.0(a)
601 ERNiCrFe-11 N06601 0.1(a) 21.0–25.0 58.0–63.0 14 … Al: 1.0–1.7
602CA … N06025 0.2 25 Bal 10 … Al: 2.1, Ti: 0.15,
Y: 0.05–0.12, Zr: 0.01–0.1
214 … … 0.05(a) 16 Bal 3 … Al: 4.5, Y: 0.01
65 ERNiFeCr-1 N08065 0.05(a) 19.5–23.5 38.0–46.0 22.0 min … Mo: 3, Cu: 1.5–3.0,
Ti: 0.6–1.2
718 ERNiFeCr-2 N07718 0.08(a) 17.0–21.0 Bal 18.5 … Ti: 0.65–1.15,
Al: 0.20–0.80,
Nb + Ta: 4.75–5.50
B ERNiMo-1 N10001 0.08(a) 1(a) Bal 4.0–7.0 2.5(a) Mo: 26.0–30.0, V: 0.3
N ERNiMo-2 N10003 0.04–0.08 6.0–8.0 Bal 5(a) … Mo: 15.0–18.0
W ERNiMo-3 N10004 0.12(a) 4.0–6.0 Bal 4.0–7.0 … Mo: 23.0–26.0
242 … … 0.03(a) 8 Bal 2(a) … Mo: 25.0
B-2 ERNiMo-7 N10665 0.02(a) 1(a) Bal 2(a) … Mo: 26.0–30.0
… ERNiMo-8 N10008 0.1(a) 0.5–3.5 60.0 min 10(a) … Mo: 18.0–21.0, W: 3.0
… ERNiMo-9 N10009 0.1(a) … 65.0 min 5(a) … Mo: 19.0–22.0, W: 3.0
B-3 ERNiMo-10 N10675 0.01(a) 1.0–3.0 65.0 min 1.0–3.0 … Mo: 27.0–32.0, Mn: 3.0(a)
G ERNiCrMo-1 N06007 0.05(a) 21.0–23.5 Bal 18.0–21.0 2.5(a) Mo: 6.5, Cu: 2.0,
Nb + Ta: 1.75–2.5,
Mn: 1.5
X ERNiCrMo-2 N06002 0.05–0.15 20.5–23.0 Bal 17.0–20.0 0.5–2.5 Mo: 9.0, W: 0.2–1.0
625 ERNiCrMo-3 N06625 0.1(a) 20.0–23.0 58.0 min 5(a) … Mo: 9.0, Nb + Ta: 3.15–4.15
C-276 ERNiCrMo-4 N10276 0.02(a) 14.5–16.5 Bal 4.0–7.0 2.5(a) Mo: 16.0, W: 3.0–4.5
C-4 ERNiCrMo-7 N06455 0.015(a) 14.0–18.0 Bal 3.0(a) 2.0(a) Mo: 16.0
S … … 0.02(a) 14.5–17.0 Bal 3.0(a) 2.0(a) Mo: 14.0–16.5, Al: 0.1–0.5,
La: 0.01–0.1
… ERNiCrMo-8 N06975 0.03(a) 23.0–26.0 47.0–52.0 16 … Mo: 6.0, Ti: 0.7–1.5,
Cu: 1.0
G-3 ERNiCrMo-9 N06985 0.015(a) 21.0–23.5 Bal 18.0–21.0 5.0(a) Mo: 7.0, W: 1.5(a), Cu: 2.0
G-30 ERNiCrMo-11 N06030 0.03(a) 28.0–31.5 Bal 15 5.0(a) Mo: 5.0, W: 1.5–4.0,
Nb + Ta: 0.3–1.5,
Cu:1.0–2.4
G-35 … N06035 0.05(a) 33 Bal 2.0(a) … Mo: 8.0
RA333 … N06333 0.05 25 Bal 17 3 Mo: 3.0, W: 3.0, Mn: 2.5
C-22/622 ERNiCrMo-10 N06022 0.015(a) 20.0–22.5 Bal 2.0–6.0 2.5(a) Mo: 12.5–14.5, W: 2.5–3.5
C-22HS … … 0.01(a) 21 Bal 2.0(a) … Mo: 17.0
59 ERNiCrMo-13 N06059 0.01(a) 22.0–24.0 Bal 1.5(a) … Mo: 16.0, Al: 0.1–0.4
686CPT ERNiCrMo-14 N06686 0.01(a) 19.0–23.0 Bal 5.0(a) … Mo: 16.0, W: 3.0–4.4
C-2000 ERNiCrMo-17 N06200 0.01(a) 23 Bal 3.0(a) … Mo: 16.0, Cu: 1.6
725 ERNiCrMo-15 N07725 0.03(a) 19.0–22.5 55.0–59.0 7 … Mo: 7.0–9.5, Nb + Ta:
2.75–4.0, Ti: 1.0–1.7
617 ERNiCrCoMo-1 N06617 0.05–0.15 20.0–24.0 Bal 3.0(a) 10.0–15.0 Mo: 9.0, Al: 0.8–1.5
230-W ERNiCrWMo-1 N06231 0.05–0.15 20.0–24.0 Bal 3.0(a) 5.0(a) W: 14.0, Mo: 2.0,
Al: 0.2–0.5
HR-160 ERNiCoCrSi-1 N12160 0.05 28 Bal 2.0(a) 29 Si: 2.75, Ti: 0.5, Nb: 1.0(a)
R-41 … … 0.05–0.12 19 Bal 5.0(a) 11 Mo: 10.0, Ti: 3.1, Al: 1.5
WASPALOY … … 0.08 19 Bal 2.0(a) 13.5 Mo: 4.3, Al: 1.5, Ti: 3.0,
Zr: 0.05
DELORO 40 … … 0.2 7.5 Bal 2.5 1.5(a) Si: 3.5
DELORO 50 … … 0.45 10.5 Bal 3.5 … Si: 4.0, B: 2.0
DELORO 60 … … 0.7 14.5 Bal 4 … Si: 2.0–4.5, B: 3.3
Colmonoy 88 … … 0.8 15 Bal 3 … W: 17.0, Si: 4.0, B: 3.0
188 … … 0.05–0.15 22 22 3.0(a) Bal W: 14.0, La: 0.02–0.12
25 (L-605) … … 0.1 20 10 3.0(a) Bal W: 15.0
(continued)
(a) Maximum. Nb = Cb
Name ///sr-nova/Dclabs_wip/High Temp/5208_445-453.pdf/Appendix_2/ 26/10/2007 1:55PM Plate # 0 pg 453

Appendix 2: Chemical Compositions of Alloys and Filler Metals / 453

Table 7 (Continued)
Filler Metal AWS UNS C Cr Ni Fe Co Others
ULTIMET … R31233 0.06 26 9 3 Bal Mo: 5.0, W: 2.0
Stellite 1 … … 2.45 31 3.0(a) 2.5(a) Bal W: 13.0
Stellite 6 … … 1.2 28 3.0(a) 3.0(a) Bal W: 4.5
Stellite 12 … … 1.6 29.5 3.0(a) 2.5(a) Bal W: 8.5
Stellite 21 … … 0.25 27 2.5(a) 3.0(a) Bal W: 5.5
Stellite 6K … … 1.6 31 3.0(a) 3.0(a) Bal W: 4.5
Stellite 704 … … 1 30 2.0(a) 2.0(a) Bal Mo: 14.0
Stellite 706 … … 1.2 29 3.0(a) 3.0(a) Bal Mo: 4.5
Stellite 712 … … 1.6 29 3.0(a) 3.0(a) Bal Mo: 8.5
Stellite 706K … … 1.6 31 3.0(a) 3.0(a) Bal Mo: 4.5
Tribaloy T-400C … … 0.1(a) 14 1.0(a) 1.0(a) Bal Mo: 27.0, Si: 2.6
Tribaloy T-401 … … 0.2 17 0.8(a) 0.8(a) Bal Mo: 22.0, Si: 1.3
(a) Maximum. Nb = Cb