9

Corrosion

9.1 Introduction
In general terms, corrosion is defined as the degradation of a structural material
by chemical reactions with the environment. From a chemical point of view,
any type of corrosion is basically an oxidation reaction where a metal atom M
loses a number x of negatively charged electrons e to become a positively charged
ion, i.e.,
M ! Mþxe þ xe
Metals in particular are the most susceptible to the above type of reaction because
of their characteristic electronic structure. It is recalled from Chap. 2 that metals
have a few valency electrons in the outermost electron shells, which are loosely
bound to the nucleus. As a result, these electrons can be easily separated from the
respective atoms, leaving behind positive ions as indicated by the above reaction.
Naturally, metals exist in the form of oxides or salts corresponding to the most
stable or lowest energy state. Therefore, a metal recovered from its naturally
existing ore always exhibits a tendency to transform to its most stable state result-
ing in corrosion; i.e., corrosion is basically the result of metals seeking their
lowest energy state. It is then evident that corrosion products consist of oxides
or salts.
Structural materials always function in some kind of an environment, e.g.,
normal atmosphere, aqueous medium, gaseous medium, etc. During service, a
structural material always tends to react with its environment. Depending upon
344
Copyright © 2004 by Marcel Dekker, Inc.
Corrosion 345

the temperature of the environment as well as its nature, two main classes of
corrosion are identified: (i) low-temperature aqueous corrosion and (ii) high-
temperature gaseous corrosion. Corrosion is considered to be one of the most
important causes of failure of structural materials during service. As a result of
metal wastage to a corrosion product by the above basic reaction, both the struc-
tural integrity of the material as well as its mechanical strength can be severely
degraded. In this chapter, the two main classes of corrosion are discussed.
Examples are given to illustrate the specific types of corrosion within each
class and how they can be characterized using various investigative tools.

9.2 Low-Temperature Aqueous Corrosion

Low-temperature aqueous corrosion refers to corrosion attack of a structural
material in the presence of a liquid medium. Corrosion proceeds when the corrod-
ing material has a significant solubility in the liquid. It is recalled that surface
atoms are in a higher energy state than bulk atoms because of their fewer nearest
neighboring atoms. As a result, surface atoms always exhibit a tendency to react
with other species in the environment to lower their energy. Therefore, corrosion
can only proceed if the associated chemical reaction results in a net decrease in
free energy. By definition, free energy is a thermodynamic quantity representing
the energy released or absorbed during a chemical reaction or any other physical
change. To illustrate the principle involved, consider the reaction of magnesium
(a reactive metal) and gold (a noble metal) with the normal atmosphere:
 
Mg þ H2 O(water) þ 12O2 (oxygen) ! Mg(OH)2 DG ¼ 142 kcal=mol
 
Au þ 32H2 O(water) þ 34O2 (oxygen) ! Au(OH)3 DG ¼ þ15:7 kcal=mol
As can be seen, the reaction leading to the formation of Mg(OH)2 is accompanied
by a decrease in free energy DG, as indicated by the minus sign, and therefore it is
thermodynamically favorable. In contrast, formation of Au(OH)3 is accompanied
by an increase in free energy, and therefore it is thermodynamically unfavorable.
It is evident that the reverse reaction, i.e., decomposition of Au(OH)3 into Au,
H2O, and O2, is accompanied by a decrease in free energy, and therefore it is
thermodynamically favorable.
Although thermodynamics is able to predict whether a given reaction can
occur, it is inherently limited in that it does not provide any information about
the reaction rate, which falls under the subject of kinetics. It is possible that a
given reaction is accompanied by a decrease in free energy as determined from ther-
modynamic data. On the other hand, the reaction rate can be extremely sluggish to
the extent that, for practical purposes, the reaction is considered not to occur. Thus,
in principle, corrosion tendencies can be predicted from thermodynamic data.
However, this approach cannot explain the “selectivity” of the environment, i.e.,

Copyright © 2004 by Marcel Dekker, Inc.

346 Chapter 9

specific metal in specific environment. In contrast, the electrochemical approach

can explain the selectivity of the environment as described below.
Most corrosion processes occur by an electrochemical mechanism invol-
ving the interconversion of electrical and chemical energy. Basically, corrosion
proceeds by removal of electrons from an atom converting it into a positive
ion. Electrons released by this oxidation reaction provide an electric current either
between two electronically different regions of the same material or between two
different materials. Differences in electronic structure are reflected by an electro-
motive force or a potential acting as the driving force of the corrosion reaction. An
electrolyte must be present to complete the reaction path. Positively charged ions
dissolve preferentially in the electrolyte and migrate to another region where they
can regain the lost electrons balancing their electronic structure.
It is evident from the above discussion that certain requirements must be
satisfied before electrochemical corrosion can occur. Such requirements define
the components of an electrochemical cell, which are (i) a liquid environment act-
ing as an electrolyte, (ii) both anodic and cathodic regions acting as a galvanic
couple, and (iii) a metallic conductor between the anode and cathode. An electro-
lyte is a liquid solution of a compound, salt, or an acid capable of conducting
electricity. By definition, an anode is a positive electrode, i.e., an electric conduc-
tor by means of which electric current enters the electrolyte. A cathode is a nega-
tive electrode by means of which electric current leaves the electrolyte.
Galvanic couples initiating electrochemical corrosion attack can be
established by various processes described later. To illustrate the process of
electrochemical corrosion, consider the example of Fig. 9.1, where a piece of

Figure 9.1 Illustration of electrochemical corrosion where anodic and cathodic

regions are formed at the surface of Fe in contact with an electrolyte.

Copyright © 2004 by Marcel Dekker, Inc.

Corrosion 347

iron is placed in contact with an electrolyte in which it can dissolve. It is

assumed that a galvanic couple, i.e., anodic and cathodic regions, is established
within the piece of iron. At the anodic region, the following oxidation reaction
occurs:
Fe ! Fe2þ þ 2e
At the cathodic region, another chemical reaction occurs resulting in either
the reduction of oxygen or liberation of hydrogen dissolved in the electrolyte
depending upon whether the electrolyte is aerated, i.e., containing dissolved oxy-
gen such as alkaline solutions, or deaerated (acidic solution).
Aerated (alkaline): O2 (in solution) þ 4e þ H2 O ! 4OH (in solution)
In an aerated solution, Fe2þ is further oxidized by
Fe2þ ! Fe3þ þ e
and then combines with (OH2) by
Fe3þ þ 3OH ! Fe(OH)3
resulting in the formation of rust deposited at the cathodic region.
In contrast, if the electrolyte is deaerated, i.e., an acidic solution, the
following reaction occurs at the cathodic region:
2Hþ (in solution) þ 2e ! H2 (gas)
To summarize, corrosion by an electrochemical mechanism consists of an
oxidation reaction at the anode and a reduction reaction at the cathode. It is evi-
dent from the above discussion that actual corrosion occurs at the anode where
the metal dissolves in the electrolyte by an oxidation reaction. Very little or no
corrosion occurs at the cathode.
In view of the electrochemical mechanism of corrosion, the tendency of
different materials to corrode can be expressed in terms of electromotive force
(EMF) or electrode potential, and the rate of corrosion can be expressed in
terms of an electric current as described below.

9.2.1 Electrode Potential: Corrosion Tendency

In the above example, a galvanic couple is assumed to be established within the
same metal. It is also possible to establish a galvanic couple between two dissim-
ilar metals. By definition, electrode potential determines which metal acts as an
anode, and which acts as a cathode in a galvanic cell. To define electrode poten-
tial, consider a simple electrochemical cell such as that shown in Fig. 9.2. It con-
sists of a Zn electrode immersed in a compartment containing a solution of zinc
salt, and a Cu electrode immersed in another compartment containing a solution

Copyright © 2004 by Marcel Dekker, Inc.

348 Chapter 9

Figure 9.2 Schematic diagram showing an electrochemical cell consisting of Zn

(anode) and Cu (cathode) electrodes.

of Cu salt. In this cell, the Zn electrode acts as an anode, and the Cu electrode as a
cathode. At the Zn anode, an oxidation reaction occurs such that a Zn atom loses
two electrons and becomes a positive ion, i.e.,

Zn ! Zn2þ þ 2e
and at the Cu cathode, a reduction reaction occurs whereby a positive ion of Cu
(Cu2þ) in solution gains the two electrons lost by the Zn, and deposits as metallic
Cu on the cathode, i.e.,

Cu2þ þ 2e ! Cu
and the net cell reaction is obtained by adding the above two reactions:

Zn þ Cu2þ ! Zn2þ þ Cu
By definition, an electric current is a directed flow of electrons, and therefore, for
the above reaction to proceed, the electrons released by the Zn atoms are made to
flow into the cathode by means of an external electric circuit, as shown in Fig. 9.2.
To complete the circuit, ions must be transferred from the cathode into the anode
by means of the electrolyte, and this can achieved by using what is known as a
salt bridge, e.g., an inverted U tube containing an inert electrolyte, which does
not take part in the electrode reactions. If an electric charge Q flows through
the cell across a potential difference or electromotive force E, the maximum
work W done on the cell is given by
W ¼ EQ

Copyright © 2004 by Marcel Dekker, Inc.

Corrosion 349

By convention, work done on the cell is indicated by a minus sign, to distinguish

it from work done by the cell, which is considered positive. For each mole of the
cell reaction, Q is given by
Q ¼ NAe
where N is the ionic charge or the number of electrons transferred in the cell per
mole of reaction, A is Avogadro’s number, and e is the charge on the electron.
Since the product Ae is by definition the Faraday F, the work W can be expressed
as
W ¼ NEF
Under equilibrium conditions, the chemical free energy change DG of the cell
reaction must supply the energy required for the electrical work, and therefore
DG ¼ NEF
However, under nonequilibrium conditions, the chemical free energy change of
the reaction DG becomes
DG ¼ DG0 þ RT ln K
where DG0 is called the standard free energy change corresponding to a reaction
occurring under standard conditions corresponding to unit activity, R is the uni-
versal gas constant, T is the temperature in degrees Kelvin, and K is the equili-
brium constant of the reaction (K ¼ ap/ar where ap and ar are the activities of
the products and reactants; activity is proportional to concentration). Similar to
DG, the standard free energy change DG0 can be written as
DG0 ¼ NE0 F
Combining the above equations results in
E ¼ E0 þ RT=NF ln K
where E is the electrode potential for a given concentration of the electrolyte and E0
is called the standard electrode potential corresponding to unit activity of ions in
the electrolyte. By using a potentiometer, the standard values of the electrode
potential can be determined under equilibrium conditions when no net current is
flowing in the electrochemical cell. It is not possible to measure the value of Eo
on an absolute scale; therefore, it is determined with reference to a standard elec-
trode. By convention, a hydrogen electrode represented by the reaction
2Hþ þ 2e ! H2
is arbitrarily selected as a reference with E ¼ 0. A hydrogen electrode consists of a
platinum wire in a 1 normal hydrochloric acid solution (1 g equivalent of HCl per
liter) in contact with hydrogen gas at 1 atmospheric pressure. Table 9.1 summarizes

Copyright © 2004 by Marcel Dekker, Inc.

350 Chapter 9

TABLE 9.1 Standard EMF Series of Metals

Metal reaction E0 (V)

Au ¼ Au3þ þ 3e2 þ1.498 Cathodic

Pt ¼ Pt2þ þ 2e2 þ1.20
Pd ¼ Pd2þ þ 2e2 þ0.987
Ag ¼ Ag3þ þ 3e2 þ0.799
Fe3þ þ e ¼ Fe2þ þ0.771
Cu ¼ Cu2þ þ 2e2 þ0.337
Sn4þ þ 2e2 ¼ Sn2þ þ0.15
2Hþ þ 2e2 ¼ H2 0 (arbitrary reference)
Pb ¼ Pb2þ þ 2e2 20.126
Sn ¼ Sn4þ þ 4e2 20.136
Ni ¼ Ni2þ þ 2e2 20.250
Co ¼ Co2þ þ 2e2 20.277
Cd ¼ Cd2þ þ 2e2 20.403
Fe ¼ Fe2þ þ 2e2 20.440
Cr ¼ Cr3þ þ 3e2 20.744
Zn ¼ Zn2þ þ 2e2 20.763
Al ¼ Al3þ þ 3e2 21.662
Mg ¼ Mg2þ þ 2e2 22.363 Anodic

the standard electrode potentials of selected metals, defining what is known as the
standard electromotive (EMF) series. When two metals are coupled in a galvanic
cell, the values of their electrode potentials determine which metal acts as cathode,
and which is anodic. As a general rule, the metal having the most positive electrode
potential acts as cathode, and the difference in electrode potential is a measure of the
driving force of corrosion of the anodic metal. With increasing potential difference,
corrosion of the anodic metal is accelerated, and vice versa.
It is a common practice to illustrate an electrode in the form MjMþNe. To
illustrate the physical significance of the standard electromotive series shown in
Table 9.1, consider the case of FejFe2þ electrode with a standard electrode poten-
tial of 20.444 V. If the potential of Fe is maintained at 20.444 V, the Fe2þ ions
in the electrolyte at unit activity remain in equilibrium. However, if Fe is electri-
cally connected to a SnjSn2þ electrode (E0 ¼ 20.136), the potential of Fe tends
to rise. In this case, the Fe2þ ions can be envisioned as moving down the potential
gradient, out of the metal and into the electrolyte, which by definition is oxidation
or corrosion. Conversely, if Fe is electrically connected to a ZnjZn2þ electrode
(E0 ¼ 20.736 V), the Fe2þ can be envisioned to move up the potential differ-
ence, resulting in reduction of Fe2þ into Fe, and therefore the Fe is protected
against corrosion.

Copyright © 2004 by Marcel Dekker, Inc.

Corrosion 351

From a practical point of view, the above example illustrates that Zn provides
a better corrosion protection of iron or steel in comparison with Sn. Although Sn is
commonly used as a protective layer, any discontinuity can lead to corrosion of Fe,
which is anodic with respect to Sn. In contrast, for galvanized steel containing a sur-
face where Zn is anodic with respect Fe, Zn corrodes rather than Fe. However, this
requires the Zn layer to be replaced periodically.
Although the standard electrode potentials listed in Table 9.1 can serve as
useful guides in comparing corrosive tendencies of pure metals, considerable
deviation may occur in practice for various reasons. First, the oxidation
tendencies of metals are dependent upon the particular environment. Second, gal-
vanic corrosion occurring in practice usually involves alloys rather than pure
metals, and therefore it is highly unlikely that the corrosion reaction involves a
pure metal in equilibrium with its ions, as listed in Table 9.1. It is therefore of
more practical significance to compare the corrosive tendencies of various metals
under actual service conditions. For example, Table 9.2 lists a modified galvanic
series in sea water.
It is evident from the above discussion that in order to predict the corrosive
tendencies of different materials, a series such as that shown in Table 9.2 is
required for each specific environment. Obviously, this is impractical, because
it requires an endless number of corrosion tests.

9.2.2 Corrosion Current: Corrosion Rate

In practice, knowledge of corrosion rates of different materials in specific
environments is of utmost importance in (i) selection of proper materials for
certain applications, (ii) ranking the corrosion resistance of different materials,
and (iii) maximizing the efficiency of operation as well as planned shutdowns

TABLE 9.2 Galvanic Series of Commercial Metals and Alloys in Sea Water

Gold Nickel
Graphite Tin
Titanium Lead
Silver 18Cr-8Ni stainless steel
18Cr-8Ni stainless steel Steel or iron
11– 30% Cr stainless steel Cast iron
Silver Aluminium-4.5% Cu-1.5% Mg-0.6% Mn
Monel alloy (70% Ni, 30% Cu) Cadmium
Bronzes (Cu-Sn) Aluminium
Copper Zinc
Brasses (Cu-Zn) Magnesium

Copyright © 2004 by Marcel Dekker, Inc.

352 Chapter 9

for repair and maintenance. As explained in the previous section, electrochemical

corrosion involves the flow of an electric current from the anode into the cathode.
Obviously, a direct correlation exists between the current density, i.e., the flow of
current per unit area, and corrosion rate. Corrosion rates can be expressed by var-
ious methods such as milligrams per square decimeter per day (mdd), and mils
per year (mpy) where 1 mil ¼ 0.001 in. However, expressing corrosion rate in
terms of mpy is generally preferred because it provides the rate of penetration
or the thinning of the part as a result of corrosion attack.
To illustrate the correlation between corrosion rate and electric current,
consider the mdd expression, where

1 mdd ¼ 1 mg=(100 cm2 )(24 h) ¼ ½(103 g)(102 )=(cm2 )(24 h)

¼ ½(1 g)(105 )=(cm2 )(24 h)

By definition, when a mass of a metal equivalent to its atomic weight in grams is

dissolved in an electrolyte, the amount of electric charge Q0 (expressed in units of
coulomb) which flows in the electrochemical cell is given by

Q0 ¼ NF coulomb

where N is the ionic charge and F is a constant called the Faraday (F ¼ 96,485
coulomb). Therefore, the charge Q associated with dissolution of 1 g is

Q ¼ NF=atomic weight coulomb=g

Electric current I expressed in units of amperes (A) is defined as the rate of flow
of electric charge, i.e., electric current ¼ Q in coulomb/time in seconds (c/s).
Therefore, the electric current density equivalent to 1 mdd is given by

I ¼ ½96,485N(105 )=½(atomic weight)(24
60
60)

¼ 1:17N(105 )=atomic weight A=cm2

To summarize, 1 mdd is equivalent to (1.17 N/atomic weight) (1025). Exper-

imental techniques have been developed which can directly convert the measured
electric current associated with corrosion into corrosion rate. For most engineer-
ing purposes, however, the corrosion rate is determined from weight loss
measurements. Test specimens are exposed to a given environment for a prede-
termined period of time, and the measured weight loss is converted into mpy by
the following relationship:

mpy ¼ (534W)(DAT)

where W is the weight loss in mg (1 mg ¼ 1023 gm), D is the density of the

material in g/cm3, A is the surface area of the specimen in in2, and T is the

Copyright © 2004 by Marcel Dekker, Inc.

Corrosion 353

exposure time in hours. In terms of corrosion rate expressed as mpy, different

materials are in general classified into three classes as follows:
Less than 5 mpy: Good corrosion resistance
5 , mpy , 50: Acceptable
More than 50 mpy: Unsatisfactory
One of the most important factors influencing the rate of actual corrosion pro-
cesses is the deviation of the electrode potential from the equilibrium value E0
leading to a phenomenon known as polarization. By polarization is meant that
the rate of electrochemical reactions occurring at the surface is limited. If the
potential is increased above the equilibrium value, the reaction is said to be ano-
dically polarized. In this case, the material is said to be active and the corrosion
current or corrosion rate increases with potential, as shown in Fig. 9.3. Conver-
sely, if the potential is reduced below the equilibrium value, the reaction is said to
be cathodically polarized where the corrosion current decreases with increasing
potential, as shown in Fig. 9.3.
Electrochemical reactions can be polarized in two distinct manners: (i) acti-
vation polarization and (ii) concentration polarization. In the case of activation
polarization, the corrosion rate or current is limited by the rate of some reaction

Figure 9.3 A schematic illustration of anodic and cathodic polarization.

Copyright © 2004 by Marcel Dekker, Inc.

354 Chapter 9

involved in a corrosion process such as the adsorption of a species to the surface.

Concentration polarization is a process where the corrosion rate is limited by the
diffusion of a certain species in the electrolyte. Evidently, in the absence of
polarization, the corrosion rate increases. However, many metal-electrolyte sys-
tems undergo some type of polarization, limiting the corrosion rate. One of the
most important types of activation polarization is passivity, as described below.
By definition, passivity is the loss of chemical reactivity between a metal
and its environment. It results from the formation of a thin layer usually of an
oxide phase on the surface of the metal. Such a passive layer acts as a barrier
between the metal and its environment. Typically, when a metal inherently
capable of developing a protective passive layer is first placed in contact with
an electrolyte, the corrosion rate increases, exhibiting the behavior of an active
material. Subsequently, the corrosion rate decreases while the passive layer is
being developed. During this stage, the metal exhibits a cathodic polarization
behavior. Once the passive layer is developed, no further corrosion takes place
until the passive layer is disrupted by various means as described later. To sum-
marize, the corrosion rate or current of a passive metal as a function of potential
or the oxidizing power of the electrolyte can be represented by a diagram as
shown in Fig. 9.4.
Regardless of the mechanism contributing to corrosion, the corrosion rate is
influenced by a number of factors dependent upon the nature of the material and

Figure 9.4 Schematic illustration of passive behavior of a metal. The corrosion rate
drops sharply within the passive region.

Copyright © 2004 by Marcel Dekker, Inc.

Corrosion 355

the environment. Materials-related factors include (i) electrode potential, (ii) sur-
face homogeneity, and (iii) inherent formation of a surface protective film. When
anodic and cathodic regions are established within the material, the corrosion rate
is accelerated the farther apart are those regions in the galvanic series, and vice
versa. Metallic and nonmetallic inclusions in the material, as well as surface areas
of local stress concentration, surface irregularities such as machining marks, and
microcracks, contribute to accelerating the corrosion rate. If the material is inher-
ently capable of developing a continuous surface protective film, it is said to be
passivated; i.e., it becomes resistant to corrosion. For example, stainless steels are
passivated by a surface layer of a Cr-rich oxide. It is to be noted that this layer
results from a corrosion reaction; i.e., it is a corrosion product. Provided the sur-
face layer is thermodynamically stable, well adhered to the metal, and of a slow
growth rate, it acts as an effective barrier between the environment and the under-
lying metal. However, the corrosion resistance is degraded when the surface layer
becomes discontinuous, or is removed and the material becomes unable to rees-
tablish that layer. Environmental-related factors include (i) presence of a suitable
electrolyte, (ii) nature of corrosion product, and (iii) duration of exposure to the
environment. Both an electrolyte and a proper supply of oxygen are necessary
agents for galvanic corrosion to take place. Corrosion is minimized or inhibited
by the absence or limited access of either agent. In contrast with surface protec-
tive layers (passive layers), spongy and poorly adhered corrosion products
contribute to accelerating the corrosion rate. To some extent, the extent of
damage caused by corrosion is influenced by the exposure time to a corrosive
environment.

9.2.3 Types of Galvanic Couples

In practice, three types of galvanic couples can initiate electrochemical corrosion
attack: (i) composition couples, (ii) stress couples, and (iii) concentration
couples. Proper design can minimize the corrosive effects of these couples, as
described below.
A composition couple consists of regions of different chemical composition
or structure, such as two different metals in contact. Fasteners, e.g., a steel bolt hold-
ing together aluminum components, are a typical example for setting up galvanic
composition cells. It is possible to judge the electrode potential generated in such
a cell from the data of Table 9.2. As pointed out earlier, metals or alloys toward
the bottom of the electromotive or galvanic series behave in an anodic manner,
i.e., corrode, if placed in contact with another metal or alloy above them. Electrode
potential increases, and in turn the corrosion rate is accelerated as the two members
of a couple become farther apart in the galvanic series. Although aluminum is
anodic with respect to iron, the relatively small difference in electrode potential
limits the corrosion rate of aluminum when it is in contact with steel. Conversely,

Copyright © 2004 by Marcel Dekker, Inc.

356 Chapter 9

because copper and zinc are more widely separated in the series, zinc corrodes at a
higher rate when it is in contact with copper as typified by dezincification of brass
(an alloy of copper and zinc). When zinc corrodes, it leaves behind holes in the
alloy, rendering it porous, weak, and prone to complete breakdown. It is to be
noted that in this case the two metals are present in the form of phases within
the same heterogeneous alloy. In general, all multiphase alloys are prone to cor-
rosion by composition couples, as further illustrated in the following example.
Nickel-molybdenum alloys containing 26– 30 wt% Mo are known to be
highly resistant to reducing media such as hydrochloric acid provided Mo
remains in solid solution. However, precipitation of a Mo-rich phase, e.g., a car-
bide or an intermetallic, leads to establishing anodic and cathodic regions in the
alloy. In this example, the Mo-rich phase becomes cathodic with respect to the
surrounding Mo-depleted zones, which are anodic. When the material is placed
in contact with hydrochloric acid, preferential corrosion attack occurs in the
Mo-depleted zones (anodic regions). This behavior can be demonstrated by test-
ing specimens of the Ni-27 Mo alloy in hydrochloric acid after given different
types of heat treatments to produce different microstructures. In the heat-treated
condition, the alloy consists essentially of a single phase, i.e., a solid solution of
Ni and Mo, as shown in the light optical micrograph of Fig. 9.5a. Figure 9.5b
illustrates the microstructure of the alloy after 24 h of exposure in boiling 20%
hydrochloric acid solution. As can be seen, the microstructure remains similar
to that in the heat-treated condition, indicating that corrosion occurs uniformly
over the entire test specimen, i.e., by uniform dissolution of the metal.
Figure 9.5c illustrates the alloy microstructure after 1000 h of exposure at
7008C. It is observed that a secondary phase of a lamellar structure has precipi-
tated at the grain boundaries. Detailed analysis shows that this phase is Ni4Mo,
and its precipitation has resulted in Mo-depleted zones at the grain boundaries.
If the same corrosion test (24 h of exposure in boiling 20% hydrochloric acid
solution) is carried out in this condition, extensive corrosion attack at grain
boundaries is observed, as shown in Fig. 9.5d. It is clear from the above example
that the corrosion behavior of a given material in the same environment is a
sensitive function of its microstructure.
Stress couples are developed when parts or regions of the same structure
contain higher residual stresses in comparison with adjacent regions. Typical
examples include rivet heads, any part where localized stress concentration
occurs, and welded components. Residual stresses increase the energy of the
atoms, and therefore the stressed regions become anodic and prone to corrosion
attack. Just like grain boundaries provide a source of a composition couple as
described above, they also act as a source of a stress couple. It is recalled that
grain boundaries contain a larger concentration of lattice defects relative to the
matrix. As a result, the boundaries become locally stressed leading to the devel-
opment of a stress couple.

Copyright © 2004 by Marcel Dekker, Inc.

Corrosion 357

Figure 9.5 Optical micrographs of Ni-27 Mo alloy after (a) heat treatment showing a
single phase of Ni and Mo, (b) 24 h exposure in boiling 20% hydrochloric acid solution
showing uniform corrosion, (c) 1000 h exposure at 7008C showing the precipitation of
lamellar Ni4Mo phase (marked A) that results in Mo-depleted zones at the grain bound-
aries, and (d) 24 h exposure in boiling 20% hydrochloric acid solution of depleted alloy
resulting in extensive corrosion attack at grain boundaries.

Concentration couples usually develop locally in partly shielded areas as a

result of changes in environmental conditions from one surface region of the part
to another. For example, variation in the oxygen content of an electrolyte such as
water can develop anodic regions depleted in oxygen (corroding) and cathodic
regions enriched in oxygen (protected). In practice, variation in oxygen content
of an electrolyte can occur in such parts as gasket surfaces, lap weld joints, sur-
face deposits, and crevices under bolt heads and rivet heads. An electrolyte such

Copyright © 2004 by Marcel Dekker, Inc.

358 Chapter 9

as water can seep into the crevice where it is entrapped. Because of the low
oxygen concentration of the entrapped water, the crevice region becomes anodic
and preferentially corrodes. A concentration cell is also developed by a spec of
dirt or a scale on an exposed metal surface, which becomes a preferential site
for corrosion in the presence of atmospheric moisture, even a single drop of
water. Grain boundaries intersecting the exposed surface of a part can also act
as microscopic crevices conducive to the formation of concentration couples.

9.2.4 Nature of Corrosion Products

It is recalled that corrosion is basically the result of metals seeking their lowest
energy state corresponding to their naturally existing ores, such as oxides and
salts. Consequently, products resulting from corrosion reactions, which are
deposited at the cathodic sites are oxides or salts of the corroding metal. Most
of these compounds are porous and are built up as weak, spongy layers with
poor adherence to the metal substrate. Usually, corrosion products can be easily
separated from the metal surface by scraping, vibration, or friction.
Characterization of corrosion products is of particular importance in failure
analysis studies. It provides a great deal of information about the environmental
conditions leading to corrosion attack.

9.2.5 Types of Corrosion Attack

It is possible to classify the various types of corrosion into five main types: (i)
uniform corrosion, (ii) localized corrosion, (iii) environmental stress cracking,
(iv) erosion corrosion, and (v) selective leaching. An account for each of the
above types of corrosion is given below.
Uniform or General Corrosion
Uniform or general corrosion proceeds uniformly over the exposed metal surface,
resulting in uniform thinning of the part. Sometimes, this type of corrosion, which
prevails in all materials, is referred to as chemical corrosion. Corrosion occurs
when the material is in contact with a liquid solution in which it has a significant
solubility. It is evaluated and quantified by corrosion rates (mpy and mdd) as
described earlier.
As explained in the previous section, grain boundaries can lead to the
development of composition, stress, and concentration couples. Consequently,
grain boundaries can be expected to corrode more rapidly relative to the matrix.
In fact, this is the basis for metallographically revealing the grain structure of a
material by etching. This may lead to the conclusion that the uniform corrosion
rate of a given material in a certain environment increases with decreasing its
grain size or increasing the grain boundary area per unit volume. However, the
exact rate at which a grain boundary corrodes relative to the matrix is dependent

Copyright © 2004 by Marcel Dekker, Inc.

Corrosion 359

upon its structure and composition. Because of the larger degree of misorienta-
tion and in turn higher energy associated with high-angle grain boundaries,
they corrode more rapidly than low-angle boundaries. Furthermore, the bound-
aries can become extremely reactive (highly anodic) if they contain certain impu-
rities. In practice, corrosion-resistant materials are heat-treated to optimize their
resistance to corrosion. Within this context, grain size has little or no significant
effect on the uniform corrosion rate, as demonstrated in Table 9.3 for a Ni-Mo
alloy resistant to HCl. When the alloy is heat treated to produce a grain size cor-
responding to ASTM No. 5-6, the uniform corrosion rate in boiling 20% HCl sol-
ution is 203 mpy. At a grain size of ASTM No. 2-3, the corrosion rate becomes
229 mpy. In contrast, if the grain boundaries become highly anodic or sensitized
by impurities or precipitates, the boundaries can preferentially corrode at a high
rate relative to the matrix.
An isocorrosion diagram is a practical tool for presenting uniform cor-
rosion rate data for a given material, indicating its resistance to a specific environ-
ment such as an acid. Isocorrosion diagrams resemble the time-temperature-
transformation (TTT) diagram in that they contain three variables in a two-
dimensional plot. Variables contained in an isocorrosion diagram include tem-
perature, concentration of a certain corrosive species in an aqueous solution,
and corrosion rate. As shown in Fig. 9.6, the x axis represents concentration
and the y axis represents temperature. Each diagram indicates the effect of con-
centration on the boiling point of the aqueous solution (boiling point curve).
Other curves in the diagram indicates the effects of temperature and concen-
tration on given corrosion rates such as 5, 20, 50 mpy, etc.
To illustrate the usefulness of isocorrosion curves, it is recalled from Sec.
9.2.2 that from an engineering point of view, a material is considered to have a
good corrosion resistance to a specific environment if the corrosion rate is less
than 5 mpy. The plot of Fig. 9.6 represents the resistance of stainless steel type
316L to formic acid. It is evident that it has good resistance at all concentrations
of formic acid provided the service temperature is below that represented by boil-
ing point curve.
It is evident from the above discussion that uniform corrosion results in an
overall reduction in cross-sectional area. If the part is stressed, uniform corrosion
has the effect of increasing the applied stress. However, since the uniform

TABLE 9.3 Effect of Grain Size on the Corrosion Rate of a Ni-Mo Alloy in
boiling 20% HCl

Grain size (ATM number) Corrosion rate, mm/year (mpy)

5 –6 203
2 –3 229

Copyright © 2004 by Marcel Dekker, Inc.

360 Chapter 9

Figure 9.6 Isocorrosion diagram illustrating the corrosion behavior of SS316L in var-
ious concentrations of formic acid. The corrosion rate curve corresponds to 4 mpy.

corrosion rate is usually slow, its weakening effect on the cross section is rather
small. Furthermore, it can easily be detected in its early stages.

Localized Corrosion
When the anodic process becomes localized and limited to specific regions,
corrosion becomes of the localized type. Examples include (i) intergranular
corrosion of sensitized material (grain boundary attack), (ii) preferential cor-
rosion attack at precipitate-matrix interface, (iii) heat-affected zone attack in
welded components, (iv) pitting, which refers to corrosion attack at specific
sites of the material, and (v) crevice, which results from oxygen depleted sites
related to environmental conditions.
In comparison with uniform corrosion, all types of localized corrosion pose
a more serious problem, which can lead to deep pits and even perforations
severely weakening the cross section. For example, under fatigue loading con-
ditions, the pits act as notches or localized regions of stress concentration accel-
erating fatigue failure. Another problem of localized corrosion is that it
frequently occurs at spots hidden from direct view, and as a result the damage
is not detected until it has progressed into an advanced stage.
Localized intergranular corrosion can occur by three mechanisms: (i) pre-
ferential attack of grain boundary precipitates whose composition renders them

Copyright © 2004 by Marcel Dekker, Inc.

Corrosion 361

anodic in a specific environment, (ii) preferential attack of grain boundaries

resulting from segregation of certain impurities at grain boundaries, and (iii) pre-
ferential attack of alloy zones adjacent to grain boundary precipitates which
deplete those zones in certain alloying elements. Under severe conditions, inter-
granular corrosion can lead to grain separation causing the material to lose its
structural integrity.
Preferential attack of grain boundary precipitates can be exemplified by
precipitation of the Mo-rich mu phase in a Ni-Mo-Cr alloy. In an oxidizing
environment, the mu phase behaves in anodic manner, and therefore it is prefer-
entially attacked. Since free machining steel grades contain larger concentrations
of P and S in comparison with other steels, portions of these elements diffuse into
the grain boundaries, rendering them conducive to corrosion. Preferential attack
of alloy-depleted zones alongside grain boundaries can result from grain bound-
ary precipitation of the Cr-rich M23C6 carbide in some stainless steels and Ni-
based alloys. Since the smaller C atoms diffuse to the grain boundaries more
rapidly than the larger Cr atoms, precipitation of the Cr-rich phase leaves behind
a zone depleted in Cr. Because Cr in solid solution provides resistance to cor-
rosion in oxidizing environments, the Cr-depleted zones alongside grain bound-
aries become highly susceptible to preferential attack in an oxidizing
environment such as HNO3. Intergranular corrosion associated with precipitates
of a secondary phase becomes catastrophic if the precipitates form a continuous
grain boundary network.
Localized attack at a precipitate-matrix interface resembles the case of
alloy-depleted zones alongside grain boundaries. Precipitation of secondary
phases enriched in important alloying elements within the matrix can leave
behind a zone at the precipitate-matrix interface depleted in those elements.
Such a zone becomes highly anodic (corroding) relative to the matrix (protected).
A similar situation can be encountered in a weld heat – affected zone. Phases
enriched in alloying elements, providing resistance to corrosion, can precipitate
at grain boundaries in a weld heat – affected zone. If the part is used in the as-
welded condition, it becomes susceptible to localized corrosion in alloy-depleted
zones alongside grain boundaries, resulting in what is known as weld decay or
knife line attack. Knife line attack differs from weld decay in that it is confined
to a narrow band of the base metal immediately adjacent to the weld. In contrast,
weld decay occurs in regions of the base metal at a larger distance from the weld.
Pitting is another type of localized corrosion where the attack is confined to
specific areas relatively small in comparison with the total exposed area of the
part. It occurs whenever localized surface couples, such as those described in
Sec. 9.2.3, are developed. Engineering alloys which rely upon the formation of
a passive surface oxide film for corrosion protection are usually more prone to
pitting attack. In this case, pitting is promoted by localized disruption of the pro-
tective film, which can result from either the initial surface condition of the part

Copyright © 2004 by Marcel Dekker, Inc.

362 Chapter 9

and/or other conditions encountered during service. Localized surface deposits

such as inclusions can preclude the formation of the protective film. Under this
condition, pitting frequently occurs at locations below the surface deposit. If a
large surface area is covered with mill scale or an applied coating, pitting can
occur at discontinuities in those layers. Since chloride ions are particularly effec-
tive in disrupting the continuity of the passive film and penetrating into the metal,
pitting is commonly encountered in chloride-containing environments. Because
the pits retain the chloride-containing solution, they tend to grow along the direc-
tion of the gravitational force. Refractory transition elements, particularly Mo
and to a lesser extent W, are found to be effective alloying elements in reducing
the susceptibility to pitting of ferrous and nonferrous alloys in chloride-containing
environments. It is possible that these elements promote the ability of the alloy to
maintain a protective passive film. Although chloride ions are most common
source of a pitting attack, it is important to realize that pitting can still occur
even if the environment is free of chloride ions. Generally, if the alloy derives
its corrosion resistance from a surface passive film, pitting may occur at those
locations where the film is locally disrupted for any reason, e.g., mechanical
effects, change in environmental conditions, an increase in temperature, etc.
Crevice attack is a form of localized corrosion related to pitting. It usually
occurs in materials susceptible to pitting. However, it differs from pitting in that it
is caused by localized changes in environmental conditions. As explained in Sec.
9.2.3, crevices are partially shielded areas of the part where the oxygen concen-
tration of the environment becomes less than that of the environment present out-
side the crevice. This leads to the development of a concentration cell where the
crevice site becomes anodic (corroding) relative the region outside the crevice
(protected).
Localized corrosion can be judged from metallographic examinations at
small magnifications, e.g., 40
or 50
. When corrosion rates are measured
and expressed as mpy, they reflect both uniform and localized corrosion attack.
Environmental Stress Cracking
Environmental stress cracking is defined as the premature failure of materials
under the combined effects of stresses (static or cyclic; applied or residual),
and damaging environments (corrosive or noncorrosive). Various types of
environmental stress cracking include (i) stress corrosion cracking, (ii) hydrogen
damage, (iii) corrosion fatigue, and (iv) fretting corrosion.
Stress corrosion cracking results from the combined effect of a corrosive
environment and a tensile stress. It is recalled that the surface energy required
to extend a crack is considered to be provided totally by strain energy. However,
if an aggressive environment is present at the crack tip, some or all of the crack
surface energy can be provided by the free energy of the chemical reaction occur-
ring at the crack tip. As a result, the fracture stress is considerably lowered in the

Copyright © 2004 by Marcel Dekker, Inc.

Corrosion 363

presence of an aggressive environment. Usually, stress corrosion cracking is

initiated at a corrosion pit. Stress corrosion cracks can propagate either intergra-
nularly or transgranularly, depending up (i) nature of the material, (ii) the
environment, and (iii) the stress level, as described below.
Experiment shows that a correlation exists between the crack propagation
mechanism and the type of dislocation structure developed by materials having
fcc structure. If dislocations cannot cross-slip with relative ease, they tend to
form planar arrays. It is recalled that such a dislocation structure is promoted
by short-range order and/or low stacking fault energy. Under this condition,
the cracks tend to propagate intergranularly. Conversely, in the absence of
short-range order or if the material has a high stacking fault energy, dislocations
can cross-slip with relative ease, forming subgrain boundaries, and the material
becomes resistant to intergranular stress corrosion cracking. However, in this
case, the cracking can occur transgranularly. Sensitization of the material by
grain boundary precipitates localizing corrosion attack alongside grain bound-
aries promotes intergranular stress corrosion cracking. Grain size is another
material character influencing stress corrosion cracking. Whether the cracking
occurs intergranularly or transgranularly, the material becomes more resistant
to stress corrosion cracking as its grain size is reduced, which is related to the
higher strength associated with a finer grain size.
For a given material, the crack propagation mechanism can change from
transgranular to intergranular if the environment is altered. Also, the nature of
the environment can have a strong influence on the stress level at which the
cracks can propagate. In a given environment, the cracks can propagate under
an applied stress as low as 10% of the yield strength of the material. However,
in another environment, the stress required to propagate the cracks in the same
material can be as high as 90% of the yield strength.
A distinguishing feature of stress corrosion cracking is that the cracks grow
incrementally or in a stepwise manner, giving rise to branching as shown in
Fig. 9.7. However, exceptions to this rule do exist. Although various models
are proposed to explain the mechanism of stress corrosion cracking, the sequence
of events leading to crack propagation according to a commonly accepted model
are (i) A protective passive film at a crack tip is ruptured under the influence of
the local stress; (ii) once the protective film is ruptured, anodic dissolution occurs
at the crack tip; (iii) the material redevelops a protective passive film at the crack
tip; and so on. There is accumulating evidence that transgranular stress corrosion
cracks proceed by a series of cleavage events; i.e., the cracks propagate along cer-
tain crystallographic planes. For example, in the case of Ti, the cracking proceeds
along a plane inclined at a small angle from the basal plane.
Stress corrosion cracking is pronounced in aggressive environments,
particularly those containing chloride ions. Although austenitic stainless steels
particularly the 300 series are susceptible to stress corrosion cracking in

Copyright © 2004 by Marcel Dekker, Inc.

364 Chapter 9

Figure 9.7 Illustration of branching during stress corrosion cracking of a 321 stainless
steel material exposed to H2S environment. The EDS spectrum shows the presence of
Fe-sulfide within the crack.

Copyright © 2004 by Marcel Dekker, Inc.

Corrosion 365

chloride-containing environments, most of the high-performance Ni-based alloys

are resistant to cracking in a wide variety of chloride-containing environments.
However, it is important to realize that no alloy is completely immune to this
type of failure. For example, Ni-based alloys are susceptible to stress corrosion
cracking in chloride and hydrogen sulfide environments, as well as in fluoride-
containing environments. In general, it is rather difficult to predict the incidence
of stress corrosion cracking particularly if it is realized that it can occur at an
extremely low stress level.
Caustic cracking or embrittlement is a phenomenon related to stress cor-
rosion cracking particularly encountered in boilers. It originates in riveted and
welded components, where small leaks allow soluble salts to build up high
local concentrations of caustic soda (NaOH) and silica. Although the concen-
tration of soda required to cause this cracking is usually 15– 30% in addition
to a small concentration of oxygen, it can occur at smaller concentrations
reaching 5%. Usually, caustic cracks proceed intergranularly and hence the
name caustic embrittlement.
Hydrogen damage is a term reserved to describe a type of stress corrosion
in an environment containing hydrogen. In applications involving hydrogen-rich
environments, the resistance of the material to the effects of high-pressure hydro-
gen is of extreme importance in ensuring the safety of the product. It is possible to
classify the damage produced by hydrogen into four main types: (i) hydrogen-
induced cracking (HIC) or hydrogen embrittlement, (ii) hydrogen blistering,
(iii) decarburization of steels, and (iv) hydrogen attack.
Hydrogen-induced cracking resembles stress corrosion cracking in that the
corrodant (atomic hydrogen in this case) diffuses to the crack tip and lowers the
stress required to propagate the crack. Usually, hydrogen-induced cracks propa-
gate discontinuously in regions of high triaxial stress (small shear stress) at a rate
dependent upon both the stress level and hydrogen concentration. Various chemi-
cal environments containing hydrogen atoms such as H2S plus moisture or high
pressure H2, as well as poor welding or electroplating procedures, can charge
hydrogen into the material.
Hydrogen blistering is a surface phenomenon, which occurs when atomic
hydrogen is retained by the material as it solidifies from the melt or as a result
of welding operations; acid pickling, electrolytic plating, and surface corrosion
reactions. When the material becomes supersaturated in hydrogen, it precipitates
as molecular gaseous hydrogen at internal interfaces producing voids. Under
pressure, the material is blistered, as shown in Fig. 9.8. Such blisters are also
referred to as holidays or fish eyes. They affect both the structural integrity of
the part as well as its cosmetic appearance.
Decarburization of steels is a high-temperature process where hydrogen in
the environment reacts with carbon in the steel, causing severe loss in strength.
Hydrogen attack is another type of damage produced by hydrogen at elevated

Copyright © 2004 by Marcel Dekker, Inc.

366 Chapter 9

Figure 9.8 Illustration of blistering/holidays formed due to the presence of gaseous

hydrogen at the internal interface of a material.

temperatures, which is related to decarburization. In this case, however, atomic

hydrogen diffuses into the steel and reacts with carbide phases to form gaseous
methane. Generation of methane within the steel causes the formation of
“bubbles” associated with high internal pressure, which can initiate and propa-
gate cracks.
Another problem related to hydrogen pickup is the precipitation of brittle
secondary phases, which in combination with dissolved hydrogen can lead to cat-
astrophic failure. This occurs in alloy systems (e.g., Ti and Zr alloys) where metal
halides are thermodynamically stable.
As its name implies, corrosion fatigue results from the combined effect of a
corrosive environment and fatigue loading. Typically, this combination consider-
ably shortens the fatigue life in comparison with the sum in whatever order of
corrosive damage and mechanical damage incurred separately. Because of the
lack of specificity to the corrosive environment, many industries consider cor-
rosion fatigue as a more pervasive problem than stress corrosion cracking. In gen-
eral, all engineering alloys susceptible to corrosion are expected to be prone to
corrosion fatigue. A characteristic feature of cracks produced by corrosion fati-
gue is their wide-mouthed character and tendency to be tapered to a blunt tip
with corrosion product filling much of the crack space. Some of the applications
where corrosion fatigue can be a problem include (i) shafts, (ii) agitator blades,
(iii) fans, and (iv) suspension wires and ropes.
Fretting corrosion is defined as the damage occurring at the interface of two
contacting surfaces undergoing small-amplitude frictional vibrations and/or

Copyright © 2004 by Marcel Dekker, Inc.

Corrosion 367

slight relative displacement in a corrosive atmosphere. Sometimes, it is referred

to as friction corrosion. It occurs at contact areas of engine components, bolted
parts, automotive parts, etc. shipped by railroads or sea. Damage produced
by fretting corrosion is manifested by discoloration or pits filled with corrosion
product, which is usually an oxide. Since pits can act as stress raisers, fretting
corrosion can lead to fatigue failure.
Erosion Corrosion
Erosion corrosion is defined as the accelerated corrosion of a material as a result
of relative motion between the material and the corrosive environment. It is
recalled from earlier discussions that when a material is subjected to general cor-
rosion, it may develop a passive surface film, reducing the rate of attack. If, how-
ever, the corrosive fluid in contact with the metal surface is in a state of motion,
maintaining the surface protective film becomes dependent upon the speed of the
corrosive fluid. When the speed of the fluid reaches the turbulent regime, it can
locally remove the protective film. Such a process is referred to as erosion. Metal
surface exposed by erosion then becomes subject to accelerated corrosion.
Because of the high speed of fluid involved in erosion, complete repassivation
becomes rather difficult.
When a metal surface is damaged by erosion corrosion, it exhibits a distinc-
tive sculpted or carved morphology. Examples of parts which can fail by erosion
corrosion include (i) elbow in steam condensate line, (ii) exhaust or wet-steam
ends of steam turbine blades, (iii) external components of aircrafts, (iv) parts
in front of inlet pipes in tanks, and (v) bends.
Selective Leaching
By definition, selective leaching is the selective removal or dissolution of an
element from an alloy by means of a composition couple. A classical example
of selective leaching is dezincification of brass described in Sec. 9.2.3. Another
example is graphitic corrosion of gray cast irons, where its graphite flakes are
selectively removed in mildly corrosive environments.

9.2.6 Corrosion Resistant Alloys

Before describing the various classes of corrosion resistant alloys, it is instructive
to know the corrosion resistance of selected metals. Table 9.4 summarizes the
performance of various metals in specific media representing reducing environ-
ments (HCl), oxidizing environments (HNO3), and alkaline environments
(NaOH/NH4OH).
It is possible to classify the corrosion resistant alloys into ferrous and
nonferrous alloys. Ferrous alloys can be either wrought or cast; they include
carbon steels, low-alloy steels, and stainless steels. Nonferrous alloys include

Copyright © 2004 by Marcel Dekker, Inc.

368 Chapter 9

TABLE 9.4 Corrosion Resistance of Various Metals

Metal HCl (reducing) HNO3 (oxidizing) NaOH/NH4OH

Fe Bad Excellent Bad

Ni Bad Bad Excellent
Co Bad Excellent Bad
Cr Bad Excellent —
Mo Excellent Bad Bad
W Good Bad Acceptable

low-grade Ni-based alloys, high grade Ni-based alloys, Co-based alloys, Al-
based alloys, Cu and Cu-based alloys, and various metals such as Ti, Zr, and
Ta. A survey of various corrosion-resistant alloys and suitable areas of appli-
cations are given below.

Wrought Ferrous Alloys

It is recalled that carbon steels are Fe-based alloys containing carbon, manganese,
silicon, sulfur, and phosphorus. All of these alloying elements are present in frac-
tional percentages (,1%), with the exception of manganese which can be close
to 1% or up to 1.5% in certain grades. Corrosion resistance of carbon steels is
essentially dependent upon the formation of an oxide surface film acting as a bar-
rier between the metal and the environment. However, in general, carbon steels
have limited corrosion resistance. Atmospheric corrosion dependent upon the
location is of particular importance in structural applications. In relatively
clean air, the products of corrosion are either oxides or carbonates. Corrosion
is accelerated in industrial areas where sulfuric acid is present in the atmosphere.
A higher corrosion rate is encountered near cities as well as near the ocean.
Because of the higher electrical conductivity of the rain and the tendency to
form soluble chlorides or sulfates in such environments, the protective oxide
film is destroyed. Low-alloy steels have better resistance to atmospheric cor-
rosion. Elements such as copper, nickel, and chromium in small concentrations
(,1% each) can considerably decelerate rusting. Phosphorus also provides pro-
tection against atmospheric corrosion. Presence of oxygen and/or acidic
conditions generally promote atmospheric corrosion of carbon steels; however,
corrosion is inhibited under alkaline conditions, e.g., steel embedded in concrete.
Generally, carbon steels should not be used in contact with dilute acids. Although
they can be used in concentrated sulfuric acid (90 – 98%) up to the boiling point,
however, they are not suitable in the presence of hydrochloric, phosphoric, and
nitric acids. Carbon steels are corroded at relatively slow rates in brines and
sea water, and therefore they may be used in such environments. Also, they
are slightly affected by neutral water and most organic chemicals.

Copyright © 2004 by Marcel Dekker, Inc.

Corrosion 369

Stainless steels are widely used in structural applications requiring

corrosion resistance. All types of stainless steels are Fe-based alloys containing
12 –30% Cr, 0– 22% Ni, and minor concentrations of C, Nb, Cu, Mo, Ta,
Ti, and Se. Depending upon their structures, stainless steels are classified into
(i) austenitic, (ii) ferritic, (iii) martensitic, and (iv) duplex stainless steels.
Table 9.5 summarizes the performance capabilities of various stainless steels
in different environments.
It is to be noted that austenitic stainless steels cannot be hardened by heat
treatment; however, their strength can significantly be increased by cold working.

TABLE 9.5 Chemical Stability of Various Grades of Stainless Steels

Environments of good
Grade Condition service performance

Austenitic
301 Annealed, and fully General atmosphere
hardened including milder
industrial atmospheres
302 Annealed, and fully Industrial atmospheres,
hardened freshwater exposure
304 Annealed Marine atmosphere, good
stress corrosion cracking
316 Annealed Sea water; acetic, nitric, and
phosphoric acids, nitrates
347 Annealed Dilute chemicals; excellent
stress corrosion resistance
Ferritic
405 Annealed General atmosphere
430 Annealed Marine atmosphere, dilute
acids, and nitrates
446 Annealed Industrial atmospheres,
high-temperature sulfur-
bearing media
Martensitic
416 Hardened and tempered General stability, fair
chemical stability
431 Hardened and tempered Same as 416
440B Annealed, hardened, and Fair chemical stability
tempered
Duplex stainless steels
Ferralium alloy Annealed Sulfuric acid production, hot
organic acids, crude oil
treatments

Copyright © 2004 by Marcel Dekker, Inc.

370 Chapter 9

All stainless steels derive their corrosion resistance from an adherent surface film
of chromium oxide. Basic 18Cr-8Ni steels (301 and 302 grades) have good
chemical stability under normal atmospheric conditions and they are weldable.
Also, they are characterized by good corrosion resistance to industrial atmos-
pheres as well as marine environments. However, pitting may occur with
extended exposure. Salt water applications require higher grades. Solutions con-
taining chlorine ions in addition to hydrochloric and hydrofluoric acid cause
severe corrosion at all concentrations and temperatures. Another disadvantage
of the lower grades of stainless steels is sensitization, referring to grain boundary
precipitation of Cr-rich carbide phases. Sensitization promotes intergranular cor-
rosion attack. This problem, however, is avoided in higher grades of stainless
steels such as 304, 316, and 347. In these steels, the carbon content is kept at a
very low level (0.08 maximum). Carbide stabilizing element such as Nb and
Ta are added to grade 347, minimizing the tendency to form Cr-rich carbides
at grain boundaries. A similar result is obtained by adding Ti to grade 321.
Ferritic grades of stainless steels derive their corrosion resistance from the
presence of Cr requiring at least 12% Cr in the alloy. Grade 405, containing a lit-
tle more than 12% Cr, has good corrosion resistance under common atmospheric
conditions as well as in fresh water and even in the presence of mild acids. Higher
Cr content increases the chemical stability of grade 430, and a better performance
is obtained by using grade 440 particularly at moderately elevated temperatures.
Similar to ferritic stainless steels, martensitic steels derive their corrosion
resistance from the presence of Cr. These grades develop maximum resistance
to atmospheric conditions in the fully hardened condition.
Duplex stainless steels consist of a mixture of ferrite and austenite. They
combine the high strength of ferritic steels and good ductility of austenitic steels.
Such steels can find many applications in the chemical process, petrochemical,
and oil industries.
Cast Ferrous Alloys
Some structural applications require that chemical stability in certain environ-
ments must be combined with good castability. Corrosion resistant cast ferrous
alloys are available in three classes depending upon the main alloying element:
(i) high Cr alloys, (ii) high Ni alloys, and (iii) high Si alloys.
High-chromium cast alloys contain 25–30% Cr and up to 3% C. Other alloy-
ing elements such as Ni are added to enhance their corrosion resistance. Although
these alloys are rather expensive and have poor ductility, they are characterized
by an outstanding corrosion resistance up to a temperature of about 10008C. Fre-
quently, these alloys are used in applications requiring both chemical stability and
oxidation resistance.
High-nickel cast alloys maintain an austenitic structure at room tempera-
ture. Their Ni content varies from 13 to 22%, and the C content can reach 3%

Copyright © 2004 by Marcel Dekker, Inc.

Corrosion 371

or slightly higher. Other alloying elements include Cr, Mn, and Si added in
smaller amounts. Corrosion resistance of these alloys is comparable to that of
wrought austenitic stainless steels in such environments as dilute acids, alkalies,
and salts.
High-silicon alloys contain 13 – 17% Si; however, the C content is not more
than 1%. Although these alloys have poor mechanical strength and fair castabil-
ity, they have outstanding resistance to nitric, sulfuric, and phosphoric acids.
However, they are readily attacked by hydrochloric acid. Resistance to hydro-
chloric acid is improved in some grades containing Mo and Ni.
Nonferrous Metals and Alloys
Nonferrous materials resistant to low-temperature corrosion include a wide
variety of alloys and metals such as low-grade Ni-based alloys, high-grade
Ni-based alloys, Cu and Cu-based alloys, Al-based alloys, Ti and Ti-based alloys,
Zr, Ta, and Nb. A brief account for the characteristic corrosion properties of these
materials is given below.
Commercially pure Ni is considered to be a general-purpose material used
in corrosion applications where the special properties of Ni-based alloys
described later are not required. It is known for its outstanding resistance to
hot or cold alkalies as well as its resistance to stress corrosion cracking. Also,
it is resistant to dilute nonoxidizing acids such as sulfuric, hydrochloric, and
phosphoric, as well as anhydrous ammonia and ammonium hydroxide solutions
of less than 1% concentration.
Low-grade Ni-based alloys such as Ni-Cu alloys (monel series) are known
for their ability to handle sea and brackish water. Also, in comparison with
commercially pure Ni, they have improved resistance to nonoxidizing acids.
However, these alloys are not resistant to oxidizing media such as nitric acid,
chromic acid, wet chlorine, and ammonia.
High-grade Ni-based alloys include a series of Ni-Mo, Ni-Mo-Cr, Ni-Mo-
W-Cr, Ni-Cr-Fe, and Ni-Mo-Cr-Fe alloys. Most of these alloys are more resistant
to stress corrosion cracking than the stainless steels. In general, Mo and W are
important alloying elements for corrosion resistance in reducing media while
Cr is essential for oxidizing media. Also, it is recalled that Mo promotes the
resistance to pitting. Alloys based upon the Ni-Mo alloys are known for their out-
standing resistance to reducing media such as HCl; however, they are not resist-
ant to oxidizing media. Resistance to both reducing and oxidizing environments
is provided by a series of Ni-Mo-Cr alloys. Other elements are added to Ni-Mo-
Cr alloys to enhance their resistance to certain media. For example, addition of
small concentrations of Cu promotes the resistance of Ni-Mo-Cr alloys to redu-
cing acids. Some of the Ni-Mo-Cr alloys contain W for further improvement of
corrosion resistance in reducing media. Alloys based upon the Ni-Cr-Fe system
are particularly useful in handling water environments because of their improved

Copyright © 2004 by Marcel Dekker, Inc.

372 Chapter 9

resistance to stress corrosion cracking in comparison with stainless steels. Also,

these alloys are highly resistant to ammonia. Their resistance to oxidizing media
is dependent upon the exact Cr concentrations. However, due to the lack of Mo,
they are not resistant to reducing media.
Commercially pure Cu is highly resistant to various atmospheric environ-
ments, such as rural, marine, and industrial. However, if sulfur is present in the
environment, it undergoes discoloration due to the formation of various oxides.
Furthermore, Cu is resistant to dry gases; fresh and sea water; pure steam;
most soils; deaerated nonoxidizing acids; organic acids such as acetic, formic,
lactic, citric, and maleic; alkaline solutions except those containing ammonium
compounds or cyanides; salt solutions such as sulfates and nitrates of sodium,
potassium, magnesium, and calcium. An added advantage of Cu is that alloying
elements added to improve its mechanical strength either have no effect on its
corrosion resistance or improve this resistance in certain environments. For
example, the addition of Zn to Cu improves its resistance to sulfur compounds.
In addition to commercially pure Cu, commercial bronze (Cu-Pb-Zn alloys),
red brass (Cu-Zn alloys), Muntz metal of the brass grades, and nickel silver
are the most widely used in applications requiring resistance to general atmos-
pheric environments. Both Cu and its alloys, however, are attacked by oxidizing
acids, salts, moist ammonia, halogen gases, sulfides, high-velocity sea water, and
a number of liquid metals such as tin and lead.
Commercially pure Al is characterized by excellent chemical stability
under most atmospheric conditions owing to formation of a passive surface
layer of Al2O3 . Specifically, Al is resistant to rural, industrial, and marine atmos-
pheres as well as neutral or nearly neutral fresh waters, sea waters, organic acids,
oils, greases, waxes, ammonia and ammonium compounds, concentrated nitric
acid (.82%), and many neutral aqueous inorganic salt solutions. Its mechanical
strength, however, is inadequate for most structural and mechanical applications.
Most of the alloying elements added to improve the mechanical strength of Al
degrades its chemical stability. However, Mg, Mn, Cr, and Si are considered to
be valuable alloying additions to Al. Addition of Mg to Al preserves its chemical
stability, and therefore Al-Mg alloys combine fairly good strength and excellent
corrosion resistance particularly in salt water and alkaline solutions. When these
alloys, however, contain Mg-rich precipitates, the general corrosion resistance is
lowered and the susceptibility to stress corrosion cracking is increased. A better
combination of mechanical strength is found in Al-Mg-Si alloys. In this case,
strengthening is provided by Mg3Si, which does not affect the electrode potential
of the alloy.
Commercially pure Ti has a chemical stability comparable to or better than
that of the best grades of stainless steels. It has an outstanding corrosion resist-
ance to most media including nitric acid in all concentrations and at all tempera-
tures, inorganic acids, organic salts, and low-temperature dilute hydrochloric,

Copyright © 2004 by Marcel Dekker, Inc.

Corrosion 373

sulfuric, and phosphoric acids, and aqua regia. Also, it is resistant to pitting attack
by chloride solutions (sea water), and to alkalies. However, Ti is attacked by
aluminum chloride and boiling concentrated potassium hydroxide, and hydro-
fluoric acid. Although Ti alloys are generally less resistant to corrosion than
the commercially pure metal, the high-strength, heat-treatable grades maintain
adequate chemical stability in many environments. Industrial applications of Ti
and its alloys are, however, limited to some extent due to their high cost.
Commercially pure zirconium (Zr) has excellent resistance to hydrochloric
and nitric acids at all concentrations and temperatures up to boiling point. Also, it
resists sulfuric acid up to 50% concentration as well as alkalies in all concen-
trations and at all temperatures. However, it is attacked by hydrofluoric acid
and aqua regia. Commercially pure tantalum (Ta) is resistant to all acids except
hydrofluoric and hot sulfuric acids. Also it is attacked by alkalies.

9.2.7 Corrosion Control and Protection

Design engineers can control the extent of damage caused by corrosion using
four methods: (i) proper selection of corrosion-resistant materials, (ii) application
of surface protective coatings, (iii) closer control of the environment, and
(iv) proper design.

Material Selection
As described in Sec. 9.2.6, many materials are available to designers for appli-
cations requiring resistance to low-temperature aqueous corrosion. Selection of
the proper material requires an exact knowledge of the environmental conditions
to which the material is exposed to during service, particularly the composition of
the corrosive medium and the operating temperature. Some of the corrosive
environments encountered in practice and the corresponding materials require-
ments are described below.
In many industrial processes, HCl results from the hydrolysis of catalysts,
i.e., their chemical decomposition by reaction with water. Evidently, structural
materials for such applications require adequate resistance to the corrosive effect
of HCl. It is recalled from earlier discussions that Mo is a very effective element
in such an environment. Therefore, Mo-containing alloys must be considered for
such applications. It is important to realize that the resistance to HCl is a function
of its concentration in the environment, as well as the Mo content of the alloy.
Stainless steels, in general, are not adequate because of their low Mo content.
In a dilute HCl environment (1% concentration) up to the boiling point, a number
of nonferrous alloys containing 5.5 – 27% Mo provide similar resistance. How-
ever, in concentrated HCl environments, only the high Mo-containing alloys
must be considered. Frequently, the environment is not purely HCl. Rather,
it contains other ionic species, e.g., aluminum chloride (nonoxidizing) from the

Copyright © 2004 by Marcel Dekker, Inc.

374 Chapter 9

catalyst and ferric chloride (oxidizing) from the steel vessels containing the
feedstock. In this case, alloys resistant to both oxidizing and reducing media
can have better performance, e.g., alloys based upon the Ni-Mo-Cr system.
Environments containing hydrofluoric acid present a different case than
those containing hydrochloric acid. In this case, is rather difficult to correlate
the resistance to hydrofluoric acid with certain alloying elements. Although the
Mo-containing alloys are not necessarily the most resistant, Ni-based alloys
usually have better performance than the stainless steels. In particular, Ni-Cu
alloys exhibit a high resistance to hydrofluoric acid.
In practice, mixtures of different acids or acids and salts are usually
encountered rather than a single corrosive species. Frequently, the corrosion
resistance of a given alloy in a mixture of corrosive species cannot be predicted
from its resistance in a single species. For mixtures of inorganic acids,
the addition of HCl or other chloride salts to sulfuric acid tends to increase
the corrosion rate. Under this condition, alloys containing a high concentration
of Mo exhibit a relatively better performance. However, the effect of adding
hydrofluoric acid is dependent upon the concentration of sulfuric acid. Addition
of hydrofluoric acid to a dilute sulfuric acid increases the corrosion rate. Conver-
sely, the corrosion rate is decreased in concentrated sulfuric acid. Severe
corrosion results if Fe-based alloys are used in an environment consisting of
dilute sulfuric acid and hydrofluoric acid. Some of the Ni-Mo-Cr-Fe alloys
have, however, good resistance to this environment. Alloys containing Cu are
recommended in environments consisting of hydrofluoric acid and concentrated
sulfuric acid.
Mixtures of organic acids are encountered in such processes as crude oil
distillation units. In this case, the acid mixture consists of saturated and unsatu-
rated fatty acids. For such an application, the selection of material is based upon
the acidity of the environment. Experience, however, shows that the presence of
Mo in the alloy is highly beneficial in this application. Although many alloys are
resistant to acetic acid, the addition of chloride as oxidizing chloride salts tends to
increase the corrosion rate. In this environment, alloys containing low Mo con-
tent are more resistant to corrosion.
A corrosive environment usually encountered in oil refineries and petro-
chemical operations is polythionic acid (H2SnO6 , where n ¼ 2 – 5). It results
from the combination of sulfur compounds, moisture condensation, and oxy-
gen. This acid is found to result in stress corrosion cracking of both stainless
steels and Ni-based alloys, particularly in the sensitized condition. In this case,
the cracking is intergranular. However, it can also occur in some materials
in the annealed condition, where it becomes transgranular. It is believed that
polythionic acid forms in refinery units during downtime on surface covered
with a sulfide in the presence of air (or oxygen) and moistures at ambient
temperatures.

Copyright © 2004 by Marcel Dekker, Inc.

Corrosion 375

Surface Protection
Although there are many materials which are inherently resistant to corrosion in
a wide variety of environments, their use in structural applications can be limi-
ted by high cost, inadequate strength, or fabrication difficulties. In applications
requiring the use of materials whose inherent resistance to corrosion is inade-
quate, the vulnerable surface is covered with a thin layer of a highly corrosion-
resistant material. Therefore, the underlying base material or substrate provides
the desired mechanical or other functional requirements, and the surface coating
provides the necessary environmental protection.
Surface protective coatings applied by hot dipping are the most commonly
used for applications requiring resistance to low-temperature aqueous corrosion.
This method of protection is based upon using metals of lower melting points in
comparison with the base metal. Among those metals are Al, Zn, Sn, Pb, and their
alloys. Zinc coating by hot dipping, commonly referred to as galvanizing, is the
most important method used to protect ferrous alloys. The coating layer may be
applied manually by dipping the part in molten Zn, particularly for small parts or
irregularly shaped objects such as castings. Galvanizing of sheet and wire pro-
ducts is carried out automatically where the thickness of the coating is closely
monitored. Although the intermetallic compound formed between Zn and the
base iron or steel provides good adherence, it can also lead to brittleness, requir-
ing a close control of the process variables to maintain a thin transition zone
between the coating and base metal. Galvanized products are characterized by
excellent resistance to general atmospheric attack, even in highly contaminated
environments. Protective coatings produced by hot dipping in Sn are also used.
Although Sn coatings provide chemical stability in severely corrosive environ-
ments such as dilute acids, alkalies, and several organic compounds, their use
is rather limited by higher cost. It is important to realize that surface protection
by hot dipping is limited to steels in the annealed condition. Because of the
high temperature required to apply the coating, the mechanical strength of steels
hardened by heat treatment or cold work can be seriously degraded.
In addition to hot dipping, surface protection of steels can be provided by
formation of an oxide layer at the surface, which is particularly effective in mildly
corrosive environments. Such an oxide layer acts as a barrier between the
environment and the base metal. It is important that the oxide layer must be con-
tinuous and well adhered to the base metal. Since the natural iron oxide Fe2O3
(common rust) is porous and soft, it is not suitable for that purpose. A protective
oxide layer of Fe3O4 (black oxide) is developed by immersing the steel in a salt
baths at a temperature of about 1508C. A somewhat better protection can be
achieved by forming a surface layer of iron phosphate at the surface of the
steel part by dipping it into a processing solution at a temperature slightly
above ambient.

Copyright © 2004 by Marcel Dekker, Inc.

376 Chapter 9

Electroplating is another method of surface protection against corrosion

attack. Both ferrous and nonferrous metals can be protected by this method.
Cadmium, zinc, nickel, and chromium are the major plating elements for protec-
tion against corrosion attack. In an electrolytic cell, the plating metal is made the
anode and the part to be plated acts as a cathode. Protection of steels against
mildly corrosive environments, e.g., ordinary atmospheric conditions, is achieved
by cadmium plating. Absorption of hydrogen by the steel during electroplating
resulting in hydrogen embrittlement is a major problem associated with electro-
plating. However, this problem can be controlled most effectively by baking the
component immediately after electroplating at a temperature of about 1908C.
Mechanical cladding is another method of surface protection. It consists of
firmly attaching a surface layer of a corrosion-resistant material to the base metal
by casting, forging, welding, or brazing followed by rolling, drawing, or extru-
sion operation. Therefore, various product shapes such as sheets, rods, wires
can be protected by a solidly bonded thin surface layer of a corrosion-resistant
material. Subsequent heat treatment can be used to optimize the mechanical prop-
erties of the base metal.

Cathodic Protection
Cathodic protection is a protective technique based upon reversing the chemical
reaction leading to corrosion. It is recalled that electrochemical corrosion occurs
as follows:
Oxidation reaction (anode) : M ! Mnþ þ ne
Reduction reaction (cathode)þ : ne þ nHþ ! n=2H2
It is evident that supplying excess electrons to the above cell suppresses the
anodic dissolution of the metal and accelerates the cathodic evolution of hydro-
gen. In practice, cathodic protection is accomplished by connecting a more active
metal (anodic) to the metal to be protected. Therefore, the anodic metal dissolves
preferentially and must be replaced periodically. Steels are cathodically protected
by Zn and Mg anodes, and the method is particularly effective in marine struc-
tures, pipelines, bridge decks, equipment, and tanks of all types, particularly
below water or underground.
Buried and submerged metal structures in the oil, gas, and waterworks
industries are increasingly made resistant to corrosion by cathodic protection.
Pipelines are now routinely designed to ensure the electrical continuity required
for cathodic protection to function.

Control of the Environment

Since corrosion is very sensitive to the environmental conditions, a closer control
of the environment provides an effective means for protection against corrosion

Copyright © 2004 by Marcel Dekker, Inc.

Corrosion 377

attack. Various methods of controlling the environment include (i) a change in

concentration of corrosive species, (ii) lowering the service temperature, (iii)
decreasing the velocity of the corrosive fluid, and (iv) removing oxidizing agents.
To illustrate the role of changing concentration, it is recalled that the cor-
rosion rate is dependent upon the concentration of corrosive species in the
environment. Many acids become noncorrosive or inert at high concentrations,
e.g., sulfuric acid and phosphoric acid. In such cases, it becomes possible to
reduce the corrosion rate by increasing the concentration of the respective acid
in the environment.
Lowering the service temperature as a means for corrosion protection is
dependent upon the particular environment. Although, in many cases, lowering
the temperature results in a decrease in corrosion rate, in other cases, changing
temperature can have a little or no effect. Also, for certain environments, increas-
ing the temperature lowers the corrosion rate, e.g., sea water.
A further reduction in corrosion rate can be obtained by decreasing the
speed of the corrosive fluid. It is important to realize, however, that metals inher-
ently protected by a surface passive layer exhibit higher corrosion rates in stag-
nant fluids. Extremely high fluid speed should always be avoided to minimize the
incident of erosion-corrosion. Removing oxygen and oxidizing agents from the
environment is another method contributing to a reduction in corrosion rate.
Addition of corrosion inhibitors to the environment is also used to reduce
the corrosion rate. By definition, a corrosion inhibitor is a substance added to the
environment in small concentration to reduce its corrosive effect. Examples of
inhibitors include organic amines, arsenic and antimony ions, sodium sulfide,
hydrazine, nitrates, chromates, and ferric salts.
Design Considerations
It is possible to control the extent of corrosion damage by better selection of plant
layout and location, as well as by proper design of equipment. Whenever pos-
sible, coastal location of plants must be avoided. Also, the plant layout must
take into consideration the direction of winds, and possible fallout of corrosive
species. Both galvanic and crevice corrosion can be controlled by proper design
of equipment.

9.3 High-Temperature Corrosion

High-temperature corrosion differs from low-temperature aqueous corrosion
in that the corrosive species are primarily in the gaseous state. In this case, cor-
rosion proceeds by a direct chemical reaction between gaseous species in the
environment and the alloy. Corrosive environments responsible for materials
degradation at elevated temperatures can be classified into two main types:
(i) gas turbine environments and (ii) nongas turbine environments. Resistance

Copyright © 2004 by Marcel Dekker, Inc.

378 Chapter 9

to high-temperature corrosive environments is provided either inherently or by

means of applying a surface protective coating. Inherently protected alloys derive
their resistance to high-temperature corrosion from a protective surface scale
based upon chromia (Cr2O3), alumina (Al2O3), or a combination of both. Protec-
tion provided by Cr2O3 is only adequate up to a temperature of about 9508C.
Above that temperature, Cr2O3 tends to convert into volatile CrO3 . Because of
the higher thermodynamic stability of Al2O3 , it provides better protection at
extreme temperatures above 10008C. Alloys with inadequate inherent resistance
to high-temperature corrosion, such as those used in gas turbine blade appli-
cations, are protected by surface coatings. It is the primary function of the coating
to develop a protective Al2O3 scale upon exposure to elevated temperatures.
High-temperature alloys used in gas turbine applications are commonly
known as the superalloys, most of which are Ni-based. To satisfy the mechanical
strength requirements of these applications, superalloys used in the hottest sec-
tions such as turbine blades must contain about 5% Al. Because of their insuffi-
cient ductility, these alloys are produced as castings. In a gas turbine where the
environment is highly oxidizing, salts deposited on the surface of structural alloys
and protective coatings can considerably accelerate the oxidation reaction, lead-
ing to a type of corrosion attack known as hot corrosion. It particularly affects the
Ni- and Co-based superalloys and protective surface coatings used in gas turbine
applications. Alkaline sulfates such as Na2SO4 are the main aggressive com-
pounds formed in the environment of a gas turbine. Deposition of Na2SO4 can
result from reaction between S in the fuel and NaCl in the combustion air. It is
also possible that Na2SO4 results from combustion of fuels containing both S
and Na. When Na2SO4 is deposited on the surface of the alloy or coating, it
degrades the protective nature of the oxide scale resulting in sulfidation attack,
which accelerates the oxidation reaction. Severe hot corrosion occurs at tempera-
tures in the range of 700 –9008C, and in the presence of sea salt it can occur at
temperatures as low as 4508C.
Many of the nongas turbine applications at high temperatures require the
use of wrought alloys. Except in a very few cases, the amount of Al required
to develop a continuous protective layer of Al2O3 considerably reduces the duct-
ility of the alloy precluding the processing of wrought products. Therefore, most
of the high-temperature alloys used in nongas turbine applications, including
the Fe-, Ni-, and Co-based alloys, are inherently protected against envi-
ronmental degradation by Cr2O3 scale. Commonly encountered corrodants in
high-temperature environments of nongas turbine industrial processes are sulfur,
chlorine, and carbon. Primary modes of corrosion attack include sulfidation, car-
burization, and chlorination. It is recalled that from a chemical point of view, any
type of corrosion is an oxidation reaction.
At high temperatures, the presence of corrosive species in the environment,
such as sulfur, carbon, and chlorine, can degrade the protective nature of the

Copyright © 2004 by Marcel Dekker, Inc.

Corrosion 379

oxide scale resulting in an accelerated oxidation attack. For both gas turbine and
nongas turbine environments, the rate of attack varies with exposure time at a
given temperature, as schematically illustrated in Fig. 9.9. Initially, the alloy
develops a protective oxide scale corresponding to a very small rate of
attack. Subsequently, the protective nature of the scale is degraded by various
processes described later, which accelerates the attack. Since oxidation plays
an important role in high temperature corrosion, a brief account for this reaction
is given below.

9.3.1 High-Temperature Oxidation

When an alloy is exposed to a high temperature, the initial rapid up take of oxy-
gen from the environment converts the surface of the alloy into an oxide whose
composition is a function of the alloy composition. During the early stages of the
oxidation reaction, various oxides of the reactive components of the alloy are
formed, e.g., FeO, Cr2O3 , NiO, CoO, Ti2O3 , MnO, Al2O3 , SiO2 , La2O5 ,
Y2O3 , etc. Each oxide of an element is formed in a proportion dependent upon
the concentration of that element in the alloy. As the oxidation reaction pro-
gresses, the sequence of events becomes dependent on the relative stabilities of
various oxides and their growth rate, as described below.

Figure 9.9 Effect of exposure time at 11508C on the oxidation behavior of Al2O3 form-
ing (1) and Cr2O3 (2 and 3) forming alloys in still air.

Copyright © 2004 by Marcel Dekker, Inc.

380 Chapter 9

Less stable oxides may be converted into more stable oxides, and oxides of
higher growth rate overgrow those of slower growth rate. From a thermodynamic
point of view, the relative stability of solid oxides is well reflected by their enthal-
pies of formation. Similar to free energy, enthalpy is a thermodynamic quantity
representing the heat content of a substance. By convention, the heat of formation
of a given compound is considered negative, and the more negative is the heat
of formation or enthalpy of a given compound, the greater is its stability.
Figure 9.10 summarizes the relative stability of various oxides. Among these
oxides, NiO and CoO and FeO are the least stable. Alumina can be seen to
have a relatively high stability, and Cr2O3 has an intermediate stability. Growth
rate of various oxides can be determined from an equation of the type:

K ¼ K0 expðQ=RTÞ

where K is the temperature-dependent reaction rate constant, K0 is a constant,

Q is the activation energy of the reaction, R is the universal gas constant, and
T is the temperature in degrees Kelvin (K). A plot of log K vs. 1/T yields a
straight line whose slope is proportional to the activation energy Q , as shown
in Fig. 9.11. Alumina has a comparatively slower growth rate. Oxides such as
NiO, CoO, and FeO have higher growth rates, and Cr2O3 has an intermediate
growth rate.

Figure 9.10 Relative stabilities of various oxides expressed in terms of their enthalpies
of formation.

Copyright © 2004 by Marcel Dekker, Inc.

Corrosion 381

Figure 9.11 Effect of temperature on the parabolic oxidation rate constant K for
(1) alumina-forming [Ni-16Cr-4.5Al-4Fe-0.05C-0.01Y] and (2) chromia-forming
[Ni-22Cr-14W-5Co-3Fe-0.5Mn-0.4Si-0.1C-0.02La] alloys.

Consider the oxidation of a simple Ni-22Cr alloy that leads to the establish-
ment of a protective scale. During the early stages of oxidation characterized by
rapid kinetics, oxides of both Ni and Cr are formed on the surface. Since the con-
centration of Ni is higher than that of Cr, the proportion of NiO formed at the sur-
face is greater than that of Cr2O3. Nickel oxide is expected to grow more rapidly
than Cr2O3. With continued exposure, NiO overgrows Cr2O3. To increase its
stability, i.e., lower its free energy, NiO reacts with Cr2O3 to form a solid solution
by the following reaction:

NiO þ Cr2 O3 ! NiCr2 O4

An oxide such as NiCr2O4 , containing more than one metallic element, is called a
spinel. Because the oxygen activity established at the alloy surface by oxides
such as NiCr2O4 and NiO is less than that required to oxidize Cr in the alloy, oxy-
gen continues to diffuse inward, precipitating Cr2O3. Eventually, particles of the
slower growing Cr2O3 are spread laterally to cover the entire surface of the alloy
forming a protective layer. Once the protective scale is formed, further growth of
other oxides is blocked decelerating the kinetics of the reaction. At this stage,
further growth of the oxide must occur by outward diffusional transport through

Copyright © 2004 by Marcel Dekker, Inc.

382 Chapter 9

the protective scale, inward diffusion of oxygen or a combination of both. With

extended exposure to elevated temperatures, the oxide scale can lose its protec-
tive nature by various processes accelerating the reaction. Examples of such
processes include growth stresses leading to spallation of the scale, and diffusion
of other corrosive species into the surface, converting the protective scale into
nonprotective oxide.
Other than Cr, structural alloys for high-temperature service contain
several alloying elements, complicating the simplified scheme of oxidation
mechanics. However, the principles involved are similar. To be protective, the
oxide scale developed by an alloy upon exposure to elevated temperatures
must satisfy the following requirements:
1. Upon exposure to elevated temperatures, the oxide scale must initi-
ally form rapidly to cover the surface of the alloy, and its subsequent
growth must proceed at a slow rate, satisfying a parabolic behavior.
2. It must be thermodynamically stable, i.e., maintain its structure and
composition with continued exposure to elevated temperatures.
3. It must form a continuous tenacious layer, well adhered to the alloy and
resistant to spallation upon thermal cycling.
4. It must have a low vapor pressure.
5. It must form an effective barrier to inward diffusion of oxygen and out-
ward diffusion of metallic atoms.
6. It must be dense and free of cracks and interconnected pores.
7. It must be strong, tough, and resistant to abrasion.
Although Al2O3 has a greater thermodynamic stability and slower growth rate in
comparison with Cr2O3 , its adherence is rather inferior. However, the adherence
of Al2O3 can be significantly improved by critical additions of active elements
such as Hf and Y. Also, the protective nature of Cr2O3 can be improved by critical
additions of La.
To summarize, during the initial stages of oxidation, the reaction is charac-
terized by rapid kinetics. Commonly, this stage is referred to as primary oxi-
dation. Upon the development of a protective oxide scale, the kinetics are
decelerated and the reaction reaches the stage of steady-state oxidation. During
this stage, the kinetics of the reaction at a given temperature are governed by a
parabolic relationship of the type:

x2 ¼ Kt

where x is weight gain per unit surface area, K is the temperature-dependent

reaction rate constant, and t is the exposure time. When the oxide scale becomes
nonprotective, the oxidation rate is accelerated during a stage called breakaway

Copyright © 2004 by Marcel Dekker, Inc.

Corrosion 383

oxidation. Within this stage, the oxidation rate is usually governed by an equation
of the type:
x ¼ Kt

9.3.2 Hot Corrosion

Hot corrosion is a term reserved to describe a type of high temperature corrosion
encountered in gas turbine engines. As pointed out earlier, hot corrosion is
induced by the formation of Na2SO4 and its deposition on the surface of the com-
ponent. Two types of hot corrosion are encountered during service depending
upon the service temperature. At relatively higher temperatures above about
8008C , where Na2SO4 is partially or totally molten, corrosion is referred to as
type I. When the temperature is between 650 and 8008C , the salt deposit becomes
solid, and type II corrosion is identified.
Depending upon the alloy composition and environmental parameters,
type I hot corrosion can occur by two mechanisms: (i) basic fluxing and (ii) acidic
fluxing. Basic fluxing occurs when the concentration or activity of oxide ions
(Na2O) in the molten salt deposited on the alloy is increased. Conversely, acidic
fluxing is associated with a decreased concentration of the oxide ions.
Basic fluxing affects the alloys primarily protected by Al2O3 scale, which
usually contains ,15% Cr. When the alloy surface becomes covered with an
oxide scale, the oxygen pressure at the oxide-salt interface is reduced. As a result,
sulfate ions (SO22
4 ) in the salt are partially dissociated by the following reaction:

4 
SO2 ! S þ 32 O2 þ O2
Protective scales on the alloy surface particularly Al2O3 react with O22 to
form anions soluble in the salt according to the following reaction:
Al2 O3 þ O2 ! 2AlO2
As the protective oxide scale is disrupted by the above reaction, S resulting
from the dissociation of SO224 diffuses into the alloy where it reacts with reactive
elements particularly Cr to form Cr-rich sulfide phases. With continued loss of
Cr, the attack becomes more severe.
Acidic fluxing occurs when the alloy contains refractory elements, e.g.,
Mo, W, and V forming acidic oxides (MoO3 , WO3, and V2O5). It can also
occur if the fuel contains vanadium. When these oxides dissolve in the molten
salt, the oxide ion concentration is decreased by the following reaction:
MoO3 þ O2 ! MoO2
4

and thus, the salt becomes acidic. When the salt becomes sufficiently acidic,
it permits the dissolution of protective oxides such as Al2O3 and Cr2O3 by the

Copyright © 2004 by Marcel Dekker, Inc.

384 Chapter 9

following reactions:

Al2 O3 ! 2Al3þ þ 3O2

Cr2 O3 ! 2Cr3þ þ 3O2

Type II hot corrosion occurs at temperatures between 650 and 8008C.

Cobalt-based alloys and coating systems containing less than about 20% Cr are
the most susceptible to this type of hot corrosion; however, it can also affect
Ni-based alloys. At the relatively lower temperatures where type II hot corrosion
occurs, the salt deposit is expected to be in the solid state (melting point of
Na2SO4 ¼ 8848C). However, the relatively high partial pressure of SO3 in the
salt at lower temperatures can promote the formation of mixed sulfates having
low melting points. As a result, a mixed sulfate of the type CoSO4-Na2SO4 in
the liquid state is deposited on the surface of Co-based alloys. In the case of
Ni-base alloys, the liquid deposit is of the type NiSO4-Na2SO4. When the liquid
sulfate is deposited on the oxide scale developed by the alloy, the partial pressure
of O2 is locally decreased, and the partial pressures of SO2 and S2 are increased.
Under these conditions, the alloy cannot maintain a protective scale. Instead,
either sulfides of Cr and/or Al are formed in the alloy, or a porous nonprotective
oxide is formed in regions of high partial pressure of oxygen. Sulfidation occurs
in Ni-based alloys or alloys containing a high concentration of Ni.
Protective scales based upon Cr2O3 are more resistant to hot corrosion
attack than Al2O3 scale, particularly at lower temperatures where the environ-
ment is severely sulfidizing. Under these conditions, Al2O3 tends to be inter-
mixed with deposited salt degrading its protective nature. However, at higher
temperatures where the environment becomes more oxidizing, Al2O3 scale offers
a greater protection. Therefore, alloys performing in hot corrosive environments
must be capable of establishing Cr2O3 scale.
Each of the two types of hot corrosion is distinguished by a characteristic
attack morphology. High-temperature or type I hot corrosion is distinguished by
a network of a sulfide phase beneath an external oxide layer. In contrast, low-
temperature or type II hot corrosion is distinguished by a pitting-type attack.
Each pit consisting of Cr2O3, Al2O3, or a mixture of both is covered by either
a Ni- or Co-rich oxide and is separated from the alloy or coating by an interfacial
layer of a sulfide phase.
An example illustrating the attack morphology of type I hot corrosion is
shown in Figure 9.12 where a Ni-based alloy protected by Cr2O3 scale is exposed
to a hot corrosive environment for 100 h at 9008C. It is observed that the alloy has
developed a surface layer of Cr2O3 scale. However, the attack can be seen to
occur more rapidly at certain locations resulting in the precipitation of a Cr-
rich sulfide phase at the grain boundaries of the alloy. When the composition
of the oxide scale is examined, it is found to contain some S in the regions

Copyright © 2004 by Marcel Dekker, Inc.

Corrosion 385

Figure 9.12 SEM micrograph of a Ni-based alloy upon exposure to a hot corrosive
environment for 100 h at 9008C. The region labeled 1 developed relatively pure Cr2O3,
region 2 formed Cr2O3 þ S while region 3 exhibited a Cr-rich sulfide.

where the attack occurs more rapidly. In contrast, at those locations where no or
little attack occurs, the scale is found to be free of S. It is possible that localized
pores or cracks developed in the scale have permitted S to diffuse into the alloy
and initiate the attack.
In general, hot corrosion attack at a given temperature occurs after an incu-
bation period. During this period, the alloy or coating system develops and main-
tains a protective oxide scale and therefore, the reaction rate is very slow and
similar to that observed in simple oxidation. Initiation of the attack at the end
of the incubation period occurs as a result of degradation of the protective nature
of the oxide scale by various processes, e.g., spallation of the scale due to growth
stresses and thermal cycling, erosion, fluxing of the scale into molten salt, and
diffusion of S into the oxide. Subsequently, the reaction rate is accelerated during
the propagation stage, and the attack becomes more severe.
It can be concluded from the above discussion that the length of the incu-
bation period at a given temperature can serve as a criterion for the resistance of
various materials to hot corrosion. This is influenced by a number of variables
some of which are related to service conditions, and others are related to the
material and design. Service conditions include such variables as temperature, ther-
mal cycling, composition and velocity of gases, composition and deposition rate of

Copyright © 2004 by Marcel Dekker, Inc.

386 Chapter 9

salts, physical state of the salts, and erosion. Variables related to the structural
material include composition, fabrication method, and precipitation of secondary
phases. Geometry of the part is the most important factor influencing hot corrosion
attack. A brief account for each of the above variables is given below.
As explained earlier, temperature determines whether the hot corrosion
attack is of type I or type II. Since the reactions involved in type II hot corrosion
(low temperature) occur more rapidly than those of type I (high temperature), the
time to initiate the attack can be decreased as the temperature is increased. Ther-
mal cycling during service can cause cracking and spallation of the oxide scale,
degrading its protective nature. Consequently, as the oxide scale becomes more
resistant to spallation, the time to initiate the attack is increased. Gas composition
can have a significant effect on hot corrosion particularly at lower temperatures. It
is recalled that the presence of SO3 promotes the formation of mixed sulfates in
the liquid state, accelerating the kinetics of type II hot corrosion. However, as the
temperature is increased, the effect of gas composition becomes increasingly less
important. Since most of the acidic oxides such as MoO3 are lost to the gas as its
velocity is increased, the time to initiate hot corrosion by acidic fluxing is
increased. Presence of certain substances in the liquid deposit can significantly
affect the kinetics of hot corrosion. For example, if NaCl is present in Na2SO4,
the time to initiate the attack at a given temperature is decreased. Chlorine can
convert the protective scale into a porous non-protective oxide. Also, if present
in larger concentrations, chlorine can promote the formation of volatile sub-
stances such as CrO2Cl2, depleting the alloy or coating system in Cr and in
turn reducing its ability to maintain a protective scale. Salt deposition rate affects
the kinetics of hot corrosion in a manner dependent upon the operative mechan-
ism. When hot corrosion occurs by the fluxing process, the incubation period is
increased with the deposition rate. In contrast, if the attack occurs as a result of
accumulation of elements from the alloy in the deposit, a slower deposition rate
decreases the time required to initiate the attack. Usually, a deposit in the liquid
state accelerates the kinetics of hot corrosion. However, certain solid deposits
such as dense layers of CaSO4 can prevent gases from reaching the alloy surface
accelerating the rate of attack by a sulfidation mechanism. Since particulate ero-
sion can cause cracking and spallation of the protective oxide scale, it decreases
the time to initiate the attack.
Effects produced by alloy composition are dependent on whether the hot
corrosion attack is of type I or type II. Under conditions favoring type I hot cor-
rosion, which primarily occurs by sulfidation, Co-based alloys are more resistant
than Ni-based alloys. Increasing the Cr content of both Co- and Ni-based alloys
improves the resistance to type I hot corrosion. Critical additions of Y also
improves the resistance to type I hot corrosion by improving the protective nature
of the oxide scale. Although Al can increase the resistance to this type of
attack, relatively small concentrations of Al in Ni-based alloys below which a

Copyright © 2004 by Marcel Dekker, Inc.

Corrosion 387

continuous layer of Al2O3 can promote sulfidation attack by increasing the sulfur
activity in the melt. Alloys with smaller concentrations of Cr (,20%) and rela-
tively high concentrations of refractory elements such as Mo and W become sus-
ceptible to a very severe attack. All types of alloys are susceptible to considerable
degradation by type II hot corrosion. Best resistance to this type of attack is
obtained at higher Cr concentrations in the range of 30– 35%. In general, alloys
produced as castings are more susceptible to hot corrosion attack in comparison
with wrought alloys and those produced by powder metallurgy techniques.
Part geometry influences the kinetics of hot corrosion by its effect on the
protective nature of the oxide scale. Edges and sharp corners can cause cracking
of the oxide scale, decreasing the time to initiate the attack.

9.3.3 High-Temperature Protective Coatings

Both the severe mechanical and thermal conditions encountered in gas turbine
applications are satisfied by a group of high-temperature alloys known as the
superalloys. Most of the superalloys used in gas turbine applications are Ni-
based, however, some Co-based alloys are also used. An inherent limitation of
superalloys, particularly those used in the hottest sections such as turbine blades,
is their inadequate environmental resistance at the service temperature, which is
usually below about 10008C, but can reach 11008C due to localized hot spot con-
ditions. To achieve an acceptable service life of gas turbine blades, they must be
protected by surface coatings. Basically, the objective of applying the coating is
to form an Al-rich layer at the surface of the alloy capable of developing and
maintaining a protective Al2O3 scale during service at elevated temperatures.
Various coatings used in practice can be classified into three main types:
(i) diffusion aluminides (ii) overlay coatings, and (iii) thermal barrier coatings.
Simple diffusion aluminide coatings are applied by enriching the surface in Al
and allowing the alloy to react with this layer to form an Al-rich compound
based upon the NiAl composition known as the b phase. Characteristically, the
coating layer contains various concentrations of alloy substrate elements. Other
modifications of the simple aluminides to improve their protective nature involve
the addition of Pt to produce a Pt-aluminide coating. Similar to the simple alumi-
nide coating, a Pt-aluminide coating contains various concentrations of substrate
elements; however, to a large extent, the presence of Pt excludes heavy or tramp
elements such as W and Mo. In this case, the coating consists of a mixture of b
phase and an intermetallic compound of Pt and Al, which can be PtAl, Pt2Al3, or
PtAl2 depending upon the type of coating. Since intermetallic compounds such as
the b phase are rather brittle, diffusion aluminide coatings lack sufficient duct-
ility. To improve the coating ductility, overlay coatings consisting of b phase dis-
persed in a ductile solid-solution matrix have been developed. They differ from
diffusion aluminides in that the alloy does not participate in forming the coating

Copyright © 2004 by Marcel Dekker, Inc.

388 Chapter 9

layer. An alloy of predetermined composition is deposited on the alloy surface by

various techniques. One of the most common types of overlay coatings is the
MCrAlY where M is Ni, Co, or Ni þ Co. Overlay coatings are characterized
by higher temperature capabilities than diffusion aluminide coatings.
To further improve the temperature capabilities of turbine blades, the met-
allic coating whether a diffusion aluminide or an overlay coating, can be further
protected by a ceramic layer acting as a thermal insulator. Conventionally, the
simple aluminide coatings have been used as standard surface protection systems
for gas turbine blades. However, the more advanced Pt-aluminide coatings and
overlay coatings are used in certain applications and expected to be more widely
used in future generations of gas turbine engines.
Typically, the thickness of a diffusion aluminide coating is 50 –70 mm;
however, the thickness of an overlay coating is about 120 – 150 mm.
Figure 9.13a – c illustrates characteristic microstructures of diffusion aluminide
and overlay coatings. As shown in Fig. 9.13a, the simple aluminide coating con-
sists of an outer layer and inner zone about 15– 20 mm in thickness and is called

Figure 9.13 Micrograph of (a) coating consisting of an ceramic top coat, outer bond
coat layer and an inner interdiffusion zone, (b) Pt-aluminide coating divided into an
outer zone consisting of a mixture of b phase and a Pt-Al compound and an inner zone
that consists of b phase, and (c) CoNiCrAlY overlay coating that consists of one layer
and a thin interdiffusion zone.

Copyright © 2004 by Marcel Dekker, Inc.

Corrosion 389

the interdiffusion zone. It results from interaction between the coating and alloy
substrate during processing. In contrast, the Pt-aluminide coating layer is divided
into two zones as illustrated in Fig. 9.13b: an outer zone consisting of a mixture-
of b phase and a Pt-Al compound (PtAl2 in this example), and the inner zone con-
sisting of b phase. However, the interdiffusion zone is observed to be similar to
that of a simple aluminide coating. Figure 9.13c illustrates the microstructure of a
CoNiCrAlY overlay coating. It consists of one layer and a thin interdiffusion
zone. Within the coating layer, particles of the b phase are dispersed in a ductile
Co-rich solid-solution matrix phase.
Performance capabilities of a given type of coating can significantly vary
from one alloy substrate to another depending upon its composition. In the
case of diffusion aluminide coatings, elements in the alloy substrate can diffuse
into the coating during heat treatment processes as well as during service. Con-
sequently, the initial coating composition is a function of alloy substrate compo-
sition. Diffusion of substrate elements into overlay coatings during processing is
insignificant; its initial composition is rather independent of alloy substrate
composition. However, significant interdiffusion can occur between an overlay
coating and the alloy substrate during service similar to the case of diffusion
aluminides.
During service, protective coatings can be degraded by various mechan-
isms, the most important of which are oxidation and interdiffusion between the
coating and alloy substrate. Oxidation plays an important role during the early
stages of coating degradation. Initially, the coating develops a surface layer of
protective Al2O3 scale. With extended exposure at elevated temperatures, the
protective nature of the scale can be degraded by environmental factors, as
well as service conditions as described in the previous section. Another important
factor contributing to coating degradation is interdiffusion between the coating
and alloy substrate. Elements from the alloy substrate can diffuse into the coating
and influence its performance dependent upon the service temperature and
exposure time at a given temperature. Interdiffusion degrades the thermal stab-
ility of the coating as well as the protective nature of the oxide scale. Although
some elements can have beneficial effects on oxidation resistance, others can pro-
duce adverse effects as described below.
A given coating remains to be protective as long as Al is selectively oxi-
dized to form a continuous layer of Al2O3 having a slow growth rate. Selective
oxidation of Al, however, can be significantly affected by interdiffusion between
the alloy substrate and coating. Outward diffusion of Ni and other elements from
the alloy into the coating and/or inward diffusion of Al dilute the coating surface
in Al. Consequently, the coating ability to maintain a continuous protective layer
of Al2O3 is reduced, and other less protective and nonprotective oxides can be
developed. However, these effects are dependent upon the extent of interdiffu-
sion. If Cr remains in solid solution within the coating, it increases the Al activity

Copyright © 2004 by Marcel Dekker, Inc.

390 Chapter 9

near the surface, promoting the coating ability to develop a continuous layer of
Al2O3 at smaller concentration of Al. A similar effect can be produced by critical
concentrations of active elements such as Hf, Y, and Zr. However, excess Cr
concentrations leading to precipitation of Cr-rich phases at the coating surface
is detrimental because Al2O3 cannot develop upon phases containing little or
no Al. Also, excess concentrations of active elements can degrade the adherence
of the scale to the coating leading to its spallation. Titanium in concentrations
greater than about 0.5% can have adverse effects by forming TiO2, which
degrades the adherence of the scale to the coating. Refractory elements such as
W, Mo, V, Ta, and Re can produce deleterious effects; however, the exact effect
varies from one element to another and can be dependent upon concentration.
Among the refractory elements, W produces the most adverse effects by reducing
the diffusivity of Al to the surface, as well as by forming nonprotective oxides. To
summarize, the net effect of interdiffusion on the performance of a given coating
can significantly vary from one alloy substrate to another.
One of the most important factors affecting the diffusional stability of a
given coating is the structure and composition of the interdiffusion zone separ-
ating the coating from the alloy substrate. Typically, for diffusion aluminide coat-
ings, the interdiffusion zone consists of a mixture of b phase and sigma phase
enriched in refractory elements such as W and Mo.

9.3.4 Sulfidation
One of the most common contaminants found in the combustion products gener-
ated by many high temperature industrial processes is sulfur. It is usually present
in low-grade fuels, chemical additives, and fluxes. In a sulfidizing environment,
elements in structural alloys, including those to be selectively oxidized to
develop a protective scale, can react with sulfur and form sulfide phases. Some
of these sulfide phases have relatively low melting temperature and high growth
rate, degrading the structural integrity of the material. However, the severity of
sulfidation attack is dependent upon a number of variables, the most important
of which are (i) the exact nature of the environment, (ii) service temperature,
and (iii) alloy chemistry. An account for each of these variables is given below.
Sulfidizing environments encountered in practice can be classified into
three main types depending upon the relative activities of sulfur and oxygen.
When S is present in the gas stream as SO2 and SO3, the activity of oxygen is
higher than that of sulfur, and therefore the environment is predominantly oxidiz-
ing. In this type of an environment, sulfidation attack is less severe. However, if
the environment consists of H2S and H2, the activity of sulfur becomes consider-
ably higher than that of oxygen, and the environment becomes predominantly
reducing. Sulfidation attack is most severe in this environment because the
alloy becomes incapable of developing a protective oxide scale. An intermediate

Copyright © 2004 by Marcel Dekker, Inc.

Corrosion 391

type of environment is that of a mixed gas, usually consisting of H2, H2S, H2O,
CO, CO, and CH4, where the sulfur and oxygen activities are sufficient to cause
both sulfidation and oxidation.
Predominantly oxidizing environments containing SO2 and SO3 are
encountered in such processes where excess air is used to ensure complete com-
bustion of fuel, as in the case of gas turbines discussed earlier. Another example
is a coal-fired boiler. A typical example of a predominantly reducing environment
is that generated in catalytic reforming units used in petroleum refineries to
upgrade the octane number of gasoline. Mixed-gas environments are generated
by combustion under conditions where no excess air or oxygen is present, as
encountered in many industrial processes such as coal gasification, chemical and
petrochemical processing, glass manufacturing, waste incineration, and fossil-fired
power generation. It is to be noted that even in an oxidizing environment, sulfi-
dation attack may occur because of a localized decrease in oxygen activity.
Because of the low oxygen activity in reducing environments, a compe-
tition exists between sulfidation and oxidation. Sulfidation dominates oxidation
at reduced oxygen activity and lower temperatures. Conversely, at higher tem-
peratures, where the environment becomes more oxidizing, sulfidation is domi-
nated by oxidation. When the environment involves both sulfidation and
oxidation, structural materials are selected on the basis of their sulfidation resist-
ance, as described below. Usually, sulfidation attack is most severe at tempera-
tures below about 9008C. However, at higher temperatures where oxidation
dominates, the attack becomes less severe.
Sulfidation attack poses a serious problem when low-melting phases of
high growth rate, particularly Ni and Ni-Fe sulfides, are formed on the alloy sur-
face. In contrast, sulfide phases containing Co or Co þ Cr have significantly
higher melting temperature and a slower growth. Chromium is generally found
to increases the sulfidation resistance of all classes of high-temperature alloys.
Also, Al has beneficial effects on sulfidation resistance provided a protective
Al2O3 scale in developed. However, Al can have adverse effects in the case of
alloys protected by Al2O3 scale. In general, the performance capabilities of var-
ious alloys, and the effect of alloy chemistry, are dependent upon the exact
environment as described below.
In a predominantly reducing environment consisting of H2 and H2S, and the
service temperature is close to 5008C as encountered in catalytic reforming pro-
cesses in petroleum refineries, all classes of high-temperature alloys including
Co-, Ni-, and Fe-based alloys, are susceptible to severe sulfidation attack because
the very low oxygen activity precludes the formation of protective Cr2O3 scale.
Alloys with higher Cr content usually have better sulfidation resistance.
When the environment is of a mixed character, Co-based alloys are the
most resistant to sulfidation attack. Also, some of the Ni-based alloys containing
a relatively high concentration of Co can have adequate resistance. With respect

Copyright © 2004 by Marcel Dekker, Inc.

392 Chapter 9

to the major elements in the alloy, a correlation exists between the resistance to
sulfidation and the ratio of (Co/Ni þ Fe) at relatively lower temperatures below
about 9008C where the environment is sulfidizing. A considerable improvement
in sulfidation resistance occurs when this ratio exceeds unity. At temperatures
higher than about 9008C, where the environment becomes more oxidizing, the
above correlation is no longer valid. Instead, the resistance to sulfidation is con-
siderably increased as the ratio (Cr/Ni) in the alloy exceeds unity. Minor
elements which can have beneficial effects on sulfidation resistance include Si
and Ti, which can improve the protective nature of Cr2O3 scale. In contrast,
Mn can have adverse effects on sulfidation resistance because of its high diffusiv-
ity in Cr2O3 scale and tendency to form an external sulfide phase.
As pointed out earlier, sulfidation attack is less severe in an oxidizing
environment. Under this condition, the alloy is capable of developing a protective
oxide scale. At relatively lower temperature where the rate of sulfidation
increases, sulfide phases can form. Generally, increasing the Cr content of all
classes of alloys improves the resistance to sulfidation attack.
Sulfidation attack morphology is similar to that of type I hot corrosion.
When the alloy is capable of developing a surface oxide scale, the rate of attack
is very small as long as the scale remains protective. Any disruption of the oxide
scale, however, can promote the sulfidation attack resulting in sulfide phases
beneath the oxide layer, as shown in the example of Fig. 9.14 (see region marked
3), derived from an Fe-based alloy containing 33% Ni and 21% Cr. Sulfide phases
containing Fe and Cr are observed beneath the Cr2O3 scale. In extreme cases
where the low-melting sulfide phases overgrow the oxide scale, nodules of soli-
dified sulfides are observed at the surface, as shown at the regions marked 2 and 3
in the inset of Fig. 9.14.

Figure 9.14 Micrograph showing the formation of Fe-Cr rich sulfide phases above (regions
marked 1 and 2 in inset) and below (region marked 3) the Cr2O3 scale in Fe-33Ni-21Cr.

Copyright © 2004 by Marcel Dekker, Inc.

Corrosion 393

9.3.5 Carburization
Carbonaceous gases consisting of mixtures of CO, CO2, and hydrocarbons in
varying proportions such as CH4 are generated in many high-temperature proces-
ses, particularly in the petrochemical industry. High-temperature carburization
attack occurs by deposition of carbon on the alloy surface and its inward diffusion
to react with carbide forming elements, particularly Cr-precipitating carbide
phases. In addition to depleting the alloy in Cr, reducing its ability to develop
a protective oxide scale, carburization can cause severe grain boundary embrittle-
ment resulting in intergranular cracking. Another detrimental effect produced by
carburization is called metal dusting. In this case, the attack is the form of pitting
and/or thinning of the part resulting in metal wastage. Carburizing environments
can be classified into two main types depending upon the relative activities of
oxygen and carbon: (i) oxidizing environments and (ii) reducing environments.
In an oxidizing environment, C is generated by the following reaction:
2CO ! CO2 þ C
Carbon is generated in reducing environments containing H2 and a hydrocarbon
such as CH4 by the following reaction:
CH4 ! 2H2 þ C
Severely carburizing environments encountered in the petrochemical
industry are of the reducing type. Similar to the case of sulfidation, a competition
exists between oxidation and carburization in a reducing carburizing environ-
ment. If the environment is severely carburizing, i.e., the activity of C is consider-
ably higher than that of O, Cr2O3-forming alloys cannot develop a protective
scale. Instead, a carbide scale is formed. When massive coke is deposited on
the surface of, e.g., pyrolysis furnace tubes, it can act as an insulator. In contrast,
since Al2O3 scale can form even if the oxygen activity is very low, Al2O3-forming
alloys offer a greater protection against carburization attack. Usually, carburiza-
tion attack is accelerated at temperatures above about 9008C, and it proceeds as
follows.
Carbon is first adsorbed on the alloy surface, and then it diffuses inward
provided the temperature is sufficiently high. Since the carbon activity at the sur-
face is rather high, a carbide of the form Cr3C2 may form. However, as the C
activity decreases with distance from the surface, the carbide phase may change
from Cr3C2 to Cr7C3 and Cr23C6. Precipitation of these Cr-rich carbide phases
can have two detrimental effects on properties: (i) They considerably reduce
the ductility of the alloy, and (ii) they reduce the overall corrosion resistance
because of the localized depletion in Cr.
An example illustrating a typical carburization attack morphology derived
from 310 stainless steel is shown in Fig. 9.15. It is observed that the surface scale

Copyright © 2004 by Marcel Dekker, Inc.

394 Chapter 9

Figure 9.15 Illustration of carburization attack morphology in 310 stainless steel.

consists of a carbide phase determined to be of the type Cr3C2. In addition to Cr,

the carbide contains a significant concentration of Fe (a carbide-forming element)
and a very small concentration of Ni (Ni is not a carbide-forming element).
Beneath the carbide scale, an oxide phase of the type (Cr,Fe)2O3 is observed.
Evidently, during the carburization attack, the carbide phase has overgrown the
oxide phase. Below the surface scale a Cr-depleted zone is observed, which
can be explained in terms of the Cr-rich carbide scale formed at the surface.
As a result of Cr depletion near the surface, the alloy becomes incapable of devel-
oping a protective Cr2O3 scale. In the interior of the alloy, a high density of Cr-
rich precipitates determined to be a carbide of the type Cr23C6 is observed.
Initially, this carbide can be of the type Cr7C3; however, with extended exposure
it can transform into the more stable Cr23C6 carbide.
For the alloys protected by Cr2O3 scale, a correlation is found to exist
between the ratio (Ni/Fe þ Cr) and the resistance to carburization. When this
ratio is increased from 0.3 to 1, the resistance to carburization is improved by
a factor of 3. Other elements which can improve the resistance to carburization
are Si and refractory carbide-forming elements such as Mo, Nb, and
W. However, these elements can degrade the weldability of the material. Silicon
can either increase the stability of Cr2O3 and/or form a protective SiO2 scale. As
pointed out earlier, Al2O3-forming alloys are the most resistant to carburization.

9.3.6 Chlorination
Chlorinating environments are created in a number of high-temperature industrial
processes such as incineration of municipal wastes, combustion of chlorine con-
taminated coal, and waste heat recuperation from chemical and metallurgical pro-
cesses utilizing chlorine as a reactant. When Cl2, HCl, or NaCl is present in the gas
stream, it can significantly accelerate the oxidation kinetics. In the presence of Cl2

Copyright © 2004 by Marcel Dekker, Inc.

Corrosion 395

and/or HCl, the surface oxide scale can lose its protective nature by forming vol-
atile metal chlorides and/or metal oxychlorides accelerating the oxidation rate.
Generally, alloys protected by Cr2O3 and free of Al exhibit a relatively low
resistance to chlorination because of the tendency of Cr2O3 to transform into vol-
atile CrO2Cl2. Least resistant alloys are those containing relatively high concen-
trations of refractory elements, such as Mo and W, due to the formation of
oxychlorides of high vapor pressure and greater tendency to volatilize. Also,
chlorine tends to convert protective scale into porous nonprotective oxides.
Among the various classes of high temperature alloys, Fe-based alloys have
the greatest resistance to chlorination. This is related to the lower vapor pressure
of iron chloride in comparison with both nickel and cobalt chlorides. Alloys

Figure 9.16 (a) The scale formed on a Co-30Cr-4.5W-3Ni alloy after 400 h exposure at
9008C to a chlorinated environment. (b) EDS spectrum derived from the encircled region 1
showing the presence of Cl along with Cr and O.

Copyright © 2004 by Marcel Dekker, Inc.

396 Chapter 9

protected by Al2O3 have higher resistance to chlorination than Cr2O3-forming

alloys because of the very low-vapor pressure of AlCl3 in equilibrium with
Al2O3. Therefore, Cr2O3-forming alloys containing Al have improved resistance.
Usually, the scale developed in a chlorinating environment consists of an
oxide phase containing Cl as shown in Fig. 9.16a, b. It is observed that the
scale consists of Cr2O3 containing a marked concentration of Cl. In some
cases, it may also contain a chloride phase.

Copyright © 2004 by Marcel Dekker, Inc.