566 ENGINEERING GEOLOGY/Rock Properties and Their Assessment

Rock Properties and Their Assessment

F G Bell, British Geological Survey, Keyworth, UK periodic agitation to remove any trapped air. This
ß 2005, Elsevier Ltd. All Rights Reserved. method is not suitable for rocks that slake when
immersed in water. The bulk density simply requires
the mass of the specimen as obtained from the field,
Introduction that is, with its natural moisture content. The volume
Those properties that can be used to describe rock of specimens can be determined by the caliper method
materials in terms of engineering classification are or the buoyancy method. Regular shaped specimens
referred to as index properties. Index properties fre- are required for the caliper method so that accu-
quently show a good correlation one with another. In rate measurements of dimensions are obtained; the
order for the test of an index property to be useful it volume is obtained by immersion in water with the
is desirable that it should be simple to obtain, rapidly buoyancy method.
performed, and inexpensive. The test results must The porosity of a rock can be defined as the volume
be reproducible and the index properties must be of the pore space divided by the total volume expres-
relevant to the engineering requirement. sed as a percentage (Figure 1). Grades of dry density
and porosity were suggested by the International
Association of Engineering Geology (IAEG) and are
Density and Porosity provided in Table 1. Total or absolute porosity in-
volves the total pore volume, that is, it includes the
The density of a rock is one of its most fundamental
occluded pores (occluded pores are isolated and not
properties. It is influenced principally by its mineral
interconnected with other pores). However, the ef-
composition on the one hand and the amount of pore
fective or net porosity is a more practical measure-
space on the other; as the proportion of pore space
ment of porosity and it may be regarded as the pore
increases so the density decreases. Four different types
space from which water or fluid can be removed. The
of density are recognised. Firstly, grain density is the
mass of the mineral aggregate per volume of
solid material. The grain density is similar to the spe-
cific gravity (or relative density) of a rock except that
the specific gravity is not expressed in units, it being
the ratio of solid rock to that of an equal volume
of water at a specified temperature. Secondly, dry
density is the mass of the mineral aggregate per volume.
Thirdly, bulk density is the mass of mineral aggregate
and natural water content per volume. Fourthly, sat-
urated density is the mass of mineral aggregate and
saturated water content per volume. The unit weight
can be used instead of density, and is expressed in
terms of stress (kN m
3). It can be derived from
density simply by dividing by 98.8 (e.g. density of
2200 kg m
3 98.8 ¼ 22.26 kN m
3 unit weight; dens-
ity is usually expressed in Mg m
3).
The specific gravity can be determined by grinding
a rock to powder and using a density bottle or alter-
natively by the immersion in water method. Deter-
mination of dry density, bulk density, and saturated
density is dependent upon accurately weighing the
rock specimen concerned and upon the accurate
measurement of its volume. In the case of the dry
density the rock specimen is dried in a ventilated
oven at 105 C until a constant mass is reached and
then allowed to cool in a desiccator. In order to obtain Figure 1 Different values of porosity attributable to different
the saturated density, the rock specimen is first satur- degrees of sorting. Numbers beneath diagrams refer to percent-
ated by immersing in water under vacuum with age porosity.
 ENGINEERING GEOLOGY/Rock Properties and Their Assessment 567

Table 1 Dry density and porosity

Dry density Porosity

Class (Mg m
3) Description (%) Description

1 Less than 1.8 Very low Over 30 Very high

2 1.8–2.2 Low 30–15 High
3 2.2–2.55 Moderate 15–5 Medium
4 2.55–2.75 High 5–1 Low
5 Over 2.75 Very high Less than Very low
1

Reproduced from Bulletin International Association Engineering

Geology, No. 19, 364–371, 1979.

latter generally is determined by using the standard

saturation method. Alternatively, an air porosimeter
can be used. In both methods, the pore volume is
obtained by saturating with water or with air, the
total volume being found by the caliper or bouyancy
method. The porosity as determined by these two
tests does not provide an indication of the way in
which the pore space is distributed within a rock, or
whether it consists of many fine pores or a smaller
number of coarse pores. Two tests have been used to
investigate the distribution of pore sizes and the
microporosity of a rock specimen, namely, the suction
plate test and the mercury porosimeter test. Micro-
porosity in the suction plate method is defined as the
volume of water retained (expressed as a percentage
of the total available pore space) when a suction
equivalent to 6.4 m of head of water is applied to
Figure 2 The Shore scleroscope.
the specimen. In effect, this measures the percentage
of pores with an effective diameter of less than 5 mm.
Such pores are able to retain water against applied instance, the hardness of a rock can be considered as
suction and influence the amount of damage that can its resistance to a penetrating force, whether static or
be caused by frost or by the crystallization of soluble dynamic, or the resistance to displacement of surface
salts that a rock used for building stone may undergo. particles by tangential abrasive force. As such, hard-
Mercury is forced to penetrate the pores of the speci- ness is controlled by the efficiency of the bond be-
men under an applied pressure in the mercury poro- tween minerals or grains, as well as the strength of
simeter test. Obviously, the finer the pores, the higher these two components.
the pressure that must be used to bring about pene- A number of tests have been used to assess indenta-
tration. In this way it is possible to derive the dimen- tion hardness, of which the two most often used in
sions and pore size distribution from a graph showing rock mechanics testing are the Shore scleroscope and
the distribution of pores sizes. A line is drawn on the Schmidt hammer tests. The Shore scleroscope is a non-
curve at the position where 10% of the pore space has destructive hardness measuring device that indicates
been filled with mercury. The pores below this size the relative values of hardness from the height of
limit can be regarded as the microporosity. rebound of a small diamond-pointed hammer that is
dropped vertically onto a securely clamped test sur-
face from a height of 250 mm (Figure 2). Because a
Hardness rock is not a homogeneous material, several hardness
Hardness is one of the most investigated properties of tests must be made over the surface of the specimen
materials, yet it is one of the most complex to under- and the results averaged. Hence, at least 20 hardness
stand. It does not lend itself to exact definition in determinations should be taken and each point of test
terms of physical concepts. The numerical value of should be at least 5 mm from any other. The Shore
hardness is as much a function of the type of test used hardness value can be used to derive an approximate
as a material property. The concept of hardness is value of uniaxial compressive strength from Figure 3
usually associated with the surface of a material. For for rocks with strengths in excess of 35 MPa.
568 ENGINEERING GEOLOGY/Rock Properties and Their Assessment

Figure 3 Correlation chart for Shore hardness, relating unit weight of rock, unconfined compressive strength and hardness value
(After Deere DU and Miller RP (1966) Engineering Classification of Engineering Properties for Intact Rock. Technical Report AFWL-TR-65-116,
Air Force Weapons Laboratory, Kirkland Air Base, New Mexico).

The Schmidt hammer is a portable non-destructive

device that expends a known amount of stored energy
from a spring and indicates the degree of rebound of a
hammer mass, following impact, within the instrument
(Figure 4). Tests are made by placing the specimen in a
rigid cradle and impacting the hammer at a series of
points along its upper surface. The hammer is held
vertically at right angles to the axis of the specimen.
The specimens should have a flat smooth surface where
tested and the rock beneath this area should be free
Figure 4 The Schmidt hammer. Reproduced from Engineering
from cracks. Test locations should be separated by at Geology, vol. 4, 1979, Elsevier.
least the diameter of the plunger. At least 20 readings
should be taken from each specimen. The lower 50%
of the test values should be discarded and the average values can be used to derive approximate values of
obtained from the upper 50%. This average is multi- uniaxial compressive strength, as shown in Figure 5.
plied by the correction factor of the Schmidt hammer Abrasion tests measure the resistance of rocks to
to obtain the hardness. However, the Schmidt ham- wear. The two abrasion tests most frequently used are
mer test is not a satisfactory method for the determin- the Dorry and Los Angeles tests. As both these tests
ation of the hardness of very soft or very hard rocks. are used to assess the resistance to wear of aggregate
Like the Shore scleroscope test, Schmidt hardness for road making, they are not dealt with here.
 ENGINEERING GEOLOGY/Rock Properties and Their Assessment 569

Figure 5 Correlation chart for Schmidt hardness, relating unit weight, unconfined compressive strength and hardness value (After
Deere DU and Miller RP (1966) Engineering Classification of Engineering Properties for Intact Rock. Technical Report AFWL-TR-65-116, Air
Force Weapons Laboratory, Kirkland Air Base, New Mexico).

Deformation of Rocks The composition and texture of a rock are gover-

ned by its origin and geological history. Few rocks are
The factors that influence the deformation characteris- composed of only one mineral species and even when
tics and the failure of rock can be divided into two they are the properties of that species vary slightly
categories: internal and external. The internal category from mineral to mineral. Such variations within min-
includes the inherent properties of the rock itself, whilst erals may be due to cleavage, twinning, inclusions,
the external category is that of its environment at a cracking, and alteration, as well as to slight differ-
particular point in time. As far as the internal factors ences in composition. This, in turn, is reflected in the
are concerned, the mineralogical composition and tex- mechanical behaviour of the resultant rocks. As a
ture are obviously important, but fractures within a consequence, few rocks can be regarded as homoge-
rock and the degree of mineral alteration are frequently neous isotropic substances. The size and shape rela-
more important. The temperature-pressure conditions tionships of the component minerals/grains are also
under which a rock exists, significantly affect its mech- significant in this respect, generally the smaller the
anical behaviour, as does its pore water content. size, the stronger the rock. One of the most important
Importantly, the length of time that a rock experiences features of texture as far as mechanical behaviour,
a change in stress, and the rate at which this is imposed, particularly strength, is concerned, is the degree of
significantly affects its deformation characteristics.
570 ENGINEERING GEOLOGY/Rock Properties and Their Assessment

interlocking of the component minerals or grains. elastic deformation is defined as one that disappears
Breakage is more likely to take place along grain when the stress responsible for it ceases. Ideal elasti-
boundaries (intergranular fracture) than through city exists if the deformation on loading and its disap-
grains (transgranular fracture) and therefore irregular pearance on unloading are both instantaneous. This is
boundaries make fracture more difficult. The bond never the case with rocks since there is always some
between grains in many sedimentary rocks is pro- retardation, known as hysteresis, in the unloading
vided by the cement and/or matrix, rather than by process. With purely elastic deformation the strain is
grains interlocking. The amount and, to a lesser a linear function of stress, that is, the material obeys
extent, the type of cement/matrix is important, not Hooke’s law. Therefore, the relationship between
only influencing strength and elasticity, but also dens- stress and strain is constant, and is referred to as
ity, porosity, and primary permeability. Rocks are Young’s modulus, E. Rock only approximates to an
not uniformly coherent materials, but contain defects ideal Hookean solid. In fact, Young’s modulus is not a
that include microfractures, grain boundaries, min- simple constant but is related to the level of applied
eral cleavage, twinning planes, inclusion trains, and stress. Just how closely rock approximates to an ideal
elongated shell fragments. Obviously, such defects material depends on its homogeneity, isotropy, and
influence the ultimate strength of a rock and may continuity. Homogeneity refers to the physical con-
act as surfaces of weakness that control the direction tinuity of a material, that is, the constituent particles
in which failure occurs. Crystal grain orientation in are evenly distributed throughout its volume so that
a particular direction facilitates breakage along that the elastic properties are the same at all points. Isot-
direction. ropy represents a measure of the directional properties
The presence of moisture in rocks adversely affects of a rock. Hence, a rock is only isotropic if it is mono-
their engineering behaviour. For instance, moisture mineralic and the crystals/grains have a random orien-
content increases the strain velocity and lowers tation. Since most rocks are composed of two or more
the strength. More specifically the angle of internal essential minerals, which may possess preferred orien-
friction is not affected significantly by changes in tation, they are generally anisotropic. Continuity
moisture content whereas the cohesion undergoes a refers to the pore space and fractures within a rock.
notable reduction. It has therefore been suggested The degree of continuity affects the cohesion and so
that the reduction in strength with increasing mois- the transmission of stress throughout a rock.
ture content is due primarily to a lowering of the The change in deformability at the elastic limit
tensile strength, which is a function of the molecular from elastic to plastic deformation is referred to as
cohesive strength of the material. the yield point or yield strength. If the stress acting on
Although all rock types undergo a decrease in a rock exceeds its elastic limit, then it becomes per-
strength with increasing temperature and an increase manently strained, the latter being brought about by
in strength as the confining pressure is increased, the plastic flow. Within the zone of plastic flow there is a
combined effect of these is notably different for dif- region where elastic stress is still important, referred
ferent rock types. With increasing temperature there to as the field of elastico-viscous flow. Plasticity
is a reduction in yield stress and strain hardening may be regarded as time-independent, non-elastic,
decreases. Heating enhances the ductility of rocks non-recoverable, stress-dependent deformation under
and their ability to deform permanently without loss uniform sustained load. Solids are classified as brittle
of integrity. The transition from brittle to ductile or ductile according to the amount of plastic deform-
deformation in porous rocks is characterized by an ation they exhibit. In brittle materials the amount of
abrupt change from dilational behaviour at low stress plastic deformation is zero or very little, whereas it is
to compaction during inelastic axial strain at high large in ductile substances. Rupture occurs when the
stress. This type of behaviour differs from that of stress exceeds the strength of the material involved.
rocks with low porosity. With the latter, dilatancy It represents the maximum stress a rock is able to
persists well into the ductile zone. The compaction withstand prior to loss of cohesion by fracturing.
that occurs during ductile deformation in porous The initiation of rupture is marked by an increase in
rocks at high confining stress is due to collapse of strain velocity.
the pore space and the rearrangement of grains to Most strong rocks exhibit little time-dependent
give more compact packing. strain or creep. However, creep in evaporitic rocks,
Four stages of deformation have been recognized, notably salt, may greatly exceed the instantaneous
namely: elastic, elastico-viscous, plastic, and rupture. elastic deformation. The time-strain pattern exhibi-
The stages are dependent upon the elasticity, viscosity, ted by such rocks, when subjected to a constant uni-
and rigidity of the rock, as well as on stress history, axial stress, can be represented diagramatically as
temperature, time, pore water, and anisotropy. An shown in Figure 6. The instantaneous elastic strain,
 ENGINEERING GEOLOGY/Rock Properties and Their Assessment 571

Figure 7 Representative stress–strain curve for rock in uncon-

fined compression, showing hysteresis.
Figure 6 Theoretical strain curve at constant stress (creep
curve).

Table 2 Classification of deformability

which takes place when a load is applied, is repre-
Class Deformability (MPa  10
3) Description
sented by OA. There follows a period of primary or
transient creep (AB) in which the rate of deforma- 1 Less than 5 Very high
2 5–15 High
tion decreases with time. If the stress is removed, 3 15–30 Moderate
the specimen recovers. At first this is instantaneous 4 30–60 Low
(BC), but this is followed by a time elastic recovery, 5 Over 60 Very low
illustrated by curve CD. On the other hand, if
Reproduced from Bulletin International Association Engineering
the loading continues, the specimen begins to exhibit Geology, No. 19, 364–371, 1979.
secondary or pseudo-viscous creep. This type of creep
represents a phase of deformation in which the rate of
strain is constant. The deformation is permanent and the tangent modulus or secant modulus. At this stress
is proportional to the length of time over which the level the secant modulus has a lower value than
stress is applied. If the loading is continued further, the tangent modulus because it includes the initial
then the specimen suffers tertiary creep during which ‘plastic’ history of the curve. A classification of
the strain rate accelerates with time and ultimately deformability has been suggested by the IAEG and is
leads to failure. given in Table 2.
In addition to their non-elastic behaviour, most
rocks exhibit hysteresis. Under uniaxial stress the
Elastic Properties slope of the stress–strain curve during unloading ini-
Young’s modulus, E, is the most important of the tially is greater than during loading for all stress
elastic constants and can be derived from the slope values (Figure 7). As stress is decreased to zero a
of the strain–stress curve obtained when a rock speci- residual strain, OR, is often exhibited. On reloading
men is subjected to unconfined compression (i.e., the curve RS is produced that, in turn, is somewhat
static loading), it being the ratio of stress to strain. steeper than OP. Further cycles of unloading and
The strains are measured by attaching strain gauges reloading to the same maximum stress give rise to
to the test specimens, or by displacement transducers, hysteresis loops, which are shifted slightly to the
and recording their outputs. Strain measurements on right. The non-linear elastic behaviour and elastic
specimens less than 50 mm in diameter, however, are hysteresis of brittle rocks under uniaxial compression
high and not representative of the material behaviour. is due to the presence of flaws or minute cracks in the
Most crystalline rocks have S-shaped stress–strain rock. At low stresses these cracks are open but they
curves (Figure 7). At low stresses the curve is non- close as the stress is increased and the rock becomes
linear and concave upwards, that is, Young’s modulus elastically stiffer, that is, E increases with stress. Once
increases as the stress increases. The initial tangent the cracks are closed the stress–strain curve becomes
modulus is given by the slope of the stress-strain curve linear.
at the origin. Gradually a level of stress is reached When a specimen undergoes compression it is
where the slope of the curve becomes approximately shortened and this generally is accompanied by an
linear. In this region Young’s modulus is defined as increase in its cross-sectional area. The ratio of lateral
572 ENGINEERING GEOLOGY/Rock Properties and Their Assessment

unit deformation to linear unit deformation, with- ð1 þ nÞð1

2nÞ
in the elastic range, is known as Poisson’s ratio, n. E ¼ rn2p ½3
ð1
nÞ
This similarly can be obtained by monitoring strains
during an unconfined compression test. The ideal or
geometrical value of Poisson’s ratio is 0.333.
Another elastic constant is compressibility, K, E ¼ 2n2s rð1
nÞ ½4
which is the ratio of change in volume of an elastic
or
solid to change in hydrostatic pressure. A further
measure of elasticity is rigidity, G, which refers to ½3ðnp =ns Þ2
4
the resistance of a body to shear. These four elastic E ¼ n2s r ½5
½ðnp =ns Þ2
1
constants (E, n, K, G) are not independent of each
other and, if any two are known, it is possible to
derive the other two from the following expressions: 0:5ðnp =ns Þ2
1
n¼ ½6
ðnp =ns Þ2
1
G ¼ E=2ð1 þ nÞ ½1
where r is density.
and

K ¼ E=3ð1
2nÞ ½2 Strength
Uniaxial Compression
Of the four constants, Young’s modulus and Poisson’s
ratio are more readily determined experimentally. The uniaxial strength, also known as the unconfined
Methods used to determine the dynamic values (as compressive strength, of a rock may be regarded as
opposed to static values, see above) of Young’s modu- the highest stress that a rock specimen can carry when
lus and Poisson’s ratio generally depend upon deter- a unidirectional stress is applied, normally in an axial
mining the velocities of propagation of elastic waves direction to the ends of a cylindrical specimen. It
through a specimen of rock. These can be measured represents the maximum load supported by a speci-
by using the high frequency ultrasonic pulse method, men during the test divided by the cross-sectional area
the low-frequency ultrasonic pulse technique, or the of the specimen. Grades of unconfined compressive
resonant method. For example, the high-frequency strength are shown in Table 3. Although its applica-
ultrasonic pulse method is used to determine the tion is limited, the uniaxial compressive strength
velocities of compressional, vp, and shear, vs, waves allows comparisons to be made between rocks and
in rock specimens of effectively infinite extent com- affords some indication of rock behaviour under more
pared to the wavelength of the pulse used. The condi- complex stress systems.
tion of infinite extent is satisfied if the average grain The behaviour of rock in uniaxial compression is
size is less than the wavelength of the pulse that, in influenced to some extent by the test conditions. The
turn, is less than the minimum dimensions of the most important of these is the length-diameter or
specimen. These two velocities can be substituted in slenderness ratio of the specimen, the most satisfac-
the following expressions to derive the dynamic tory slenderness ratio being 2.5 since it provides a
values of Young’s modulus and Poisson’s ratio: reasonably good distribution of stress throughout

Table 3 Grades of unconfined compressive strength

Geological Society (1977) IAEG (1979) ISRM (1981)

Term Strength (MPa) Term Strength (MPa) Term Strength (MPa)

Very weak Less than 1.25 Weak Under 15 Very low Under 6
Weak 1.25–5.00 Moderately strong 15–50 Low 6–10
Moderately weak 5.00–12.50 Strong 50–120 Moderate 20–60
Moderately strong 12.50–50 Very strong 120–230 High 60–200
Strong 50–100 Extremely strong Over 230 Very high Over 200
Very strong 100–200
Extremely strong Over 200

International Association of Engineering Geology.

International Society for Rock Mechanics.
 ENGINEERING GEOLOGY/Rock Properties and Their Assessment 573

are obtained (Figure 10). The shear strength (t) is then

derived from the Mohr–Coulomb criterion:

t ¼ c þ s tanf ½7

where s is the normal stress.

Direct Shear

A number of other tests can be used to deduce the

Figure 8 Stages in the development of fracturing with
shear strength of intact rock and of rock fracture
increasing unconfined compressive loading. surfaces, in addition to the triaxial test. These include
the shear box test, the direct single and double shear
tests, the punch shear test, and the torsion test. The
most commonly used method is the shear box test,
the specimen. Secondly, the rate at which loading
which is particularly useful for assessing the shear
occurs also influences the compressive strength, a
strength of weaker rocks and also for obtaining the
loading rate of between 0.5 and 1.0 MPa normally is
shear strength along discontinuity surfaces. In this
recommended. Thirdly, the ends of a specimen should
test a constant normal force is applied to the speci-
be lapped so that they are exactly perpendicular to the
men, which then is sheared. A number of tests on the
long axis.
same material are undertaken, each at a higher normal
The onset of failure in a rock specimen subjected to
stress, so that a shear strength versus normal stress
compressive loading is marked initially by the forma-
graph can be drawn from which the value of cohesion
tion of a large number of isolated fractures, which
and angle of friction are derived (Figure 11A). The
characterizes the relief of stress concentration pro-
other shear tests are used infrequently but are illus-
duced by the mechanical inhomogeneities in the
trated in Figure 11B.
rock. Most cracks are orientated parallel to the ap-
plied stress (Figure 8). This is quickly followed by the
development of two groups of macroscopic shear Tensile Strength
failures, at the boundary and in the interior of the Rocks have a much lower tensile strength than com-
specimen, which suggests that most of the major pressive strength. Brittle failure theory predicts a ratio
sources of induced lateral tensile stresses have been of compressive strength to tensile strength of about
eliminated. The interior macroscopic shear failures 8:1 but in practice it is generally between 15:1 and
are extended and become interconnected to form a 25:1. The direct tensile strength of a rock can be
conjugate set of open shear fractures. Two central obtained by attaching metal end caps with epoxy
cones are formed, which either abrade during the resin to the specimen, which are then pulled into
large shear displacement or produce major fractures tension by wires. In direct tensile tests the slenderness
in the remaining rock material. If one shear failure ratio of cylindrical specimens should be 2.5 to 3.0 and
surface becomes dominant, then cones are not de- the diameter preferably should not be less than
veloped and the sample ultimately fails in two parts 54 mm. The ratio of the diameter of the specimen to
along a diagonal plane of shear. the largest crystal/grain in the rock should be at least
10:1. Unfortunately, the determination of the direct
Triaxial Compression Strength
tensile strength often proves difficult since a satisfac-
A triaxial test is necessary if the complete nature of tory method has not yet been devised to grip the
the failure of a rock is required. In this test a constant specimen without introducing bending stresses. Ac-
hydraulic pressure (the confining pressure) is applied cordingly, most tensile tests have been carried out by
to the cylindrical surface of the rock specimen, whilst indirect methods.
applying an axial load to the ends of the sample. The The Brazilian test is an indirect method of asses-
axial load is increased up to the point where the sing the tensile strength of rocks, based on the obser-
specimen fails. Testing of the rock specimen is carried vation that most rocks in biaxial stress fields fail in
out within a special high pressure cell (Figure 9). tension when one principal stress is compressive. In
A series of tests, each at higher confining pressure, this test a cylindrical specimen of rock is loaded in a
are carried out on specimens from the same rock. diametrical plane along its axis. The sample usually
These enable Mohr circles and their envelope to be fails by splitting along the line of diametrical loading
drawn, from which the strength parameters, that is, and the indirect tensile strength (Tb) can be obtained
the angle of friction (f) and value of the cohesion (c) from:
574 ENGINEERING GEOLOGY/Rock Properties and Their Assessment

Figure 9 Cutaway view of a triaxial cell for testing rock.

Tb ¼ 2P=pLD ½8

where P is the load at failure, and L and D are the

length and diameter of the specimen respectively.
Disc-shaped specimens are used in the Brazilian
disc test. In this case curved jaw loading platens are
used to improve loading conditions. Uncertainties
associated with the premature development of failure
can be removed by drilling a hole in the centre of the
disc-shaped specimen (sometimes this has been re-
ferred to as a ring test). A disc-shaped specimen
should be wrapped around its periphery with a layer
Figure 10 Mohr envelope and circles. of masking tape and the specimen should not be less
 ENGINEERING GEOLOGY/Rock Properties and Their Assessment 575

Figure 11 (A) Graph derived from shear test, showing curves derived for both peak and residual strength. (B) Other forms of shear
strength tests.

54 mm in diameter, with a thickness that is approxi- correction chart. Once determined, the point load
mately equal to the radius of the specimen. The tensile index can be used to grade the indirect tensile strength
strength of the specimen is obtained as follows: of rocks, as shown in Table 4.
Finally, in the flexural test a cylindrical specimen of
Tb ¼ 0:636P=DH ½9 rock is loaded between one lower and two upper
supports until the sample fails. The flexural strength
where H is the thickness of the specimen. The Brazil-
gives a higher value of tensile strength than that
ian test is useful for brittle materials but for other
determined in direct tension.
materials it may give erroneous results.
In the point load test the specimen is placed be-
tween opposing cone-shaped platens and subjected to Durability of Rocks
compression. This generates tensile stresses normal to
the axis of loading and the indirect tensile strength Durability refers to the resistance that a rock offers to
(Tp) is then derived from: the various processes that lead to its breakdown and
therefore durability tests can be used to provide a
Tp ¼ P=D2 ½10 general impression of how a rock will behave in rela-
tion to weathering, especially mechanical weathering.
Loading can take place across the diameter of the Durability tests most frequently are used to assess the
specimen, as in Figure 12, or along the axis. Rocks behaviour of suspect rocks, that is, those that tend
that are anisotropic should be tested along and paral- to breakdown more readily such as mudrocks, some
lel to the lineation. Irregular-shaped specimens can chalks, and certain basalts and dolerites. There are a
also be tested, but at least 20 tests should be made large number of tests that have been used to assess the
on the same sample material and the results averaged durability and many of them are used to determine
to obtain a value. The point load test is limited to the durability of rock as a material for building pur-
rocks with uniaxial compressive strengths exceeding poses, for aggregate, or for armourstone. The latter
25 MPa (i.e., point load index above 1 MPa). types of test are not dealt with here.
The effect of the size of specimens is greater in Some of the more simple tests include the water
tensile than compression testing because in tension, absorption test, the wet and dry test, the freeze-thaw
cracks open and give rise to large strength reductions, test, and soak tests. The water absorption test in-
whilst in compression the cracks close and so disturb- volves oven drying a rock specimen at 105 C until
ances are appreciably reduced. This is especially the it has attained a constant weight and then saturat-
case in the axial and irregular lump point load tests. ing it under vacuum. The percentage saturation is
Accordingly, a standard distance between the two determined and reflects porosity. As rocks break
cones of 50 mm has been recommended, to which down, their porosity increases and so the water ab-
other sizes should be corrected by reference to a sorption test has been used to indicate the degree of
576 ENGINEERING GEOLOGY/Rock Properties and Their Assessment

Figure 12 Point load test apparatus.

Table 4 Point load strength classification the same way as for the wet and dry test. The freeze-
thaw test has been used to assess the frost resistance
Point load Equivalent uniaxial
strength compressive of building stone. However, it is no longer used for
Description index (MPa) strength (MPa) this purpose in Britain, it being regarded as unsatis-
Extremely high strength Over 10 Over 160
factory primarily because of the difficulty of inter-
Very high strength 3–10 50–160 preting the results in relation to a period of time
High strength 1–3 15–60 over which a rock will perform as required. Soak
Medium strength 0.3–1 5–16 tests are used to assess the breakdown of rocks as a
Low strength 0.1–0.3 1.6–5 result of swelling brought about by the absorption of
Very low strength 0.03–0.1 0.5–1.6
Extremely low strength Less than 0.03 Less than 0.5
water, especially of those rocks that contain swelling
minerals. The rock specimen either may be soaked in
water or ethylene glycol (CH2OH)2 for a given
number of days. Ethylene glycol is much more effect-
weathering a rock, especially a crystalline rock such ive than water as far as assessment of those rocks
as granite, has undergone. In other words, is it fresh, that contain swelling minerals are concerned. The
slightly weathered, moderately weathered, or high- soak test allows five classes of rock disintegration
ly weathered. However, some moderately and highly and the time when the worst condition occurs to be
weathered rocks may break down before becoming recognized:
saturated. Similarly, in the wet and dry test the rock Degree of disintegration:
specimen is first dried and then saturated, but this Class 1 : No obvious effects, or only very minor
time for a given number of cycles. The effect that spalling of sand-sized particles.
cyclic wetting and drying has on the specimen is Class 2 : Flaking and/or swelling.
recorded (i.e., no effect, softening, minor spalling, Class 3 : Fracturing without extensive spalling.
minor hairline cracking, severe hairline cracking, Class 4 : Fracturing with extensive spalling.
breakdown before last cycle noting cycle number). Class 5 : Complete disintegration.
In the freeze-thaw test the specimen is saturated and Time required to reach worst condition:
then frozen for 24 h. The specimen is subjected to a Class 6 : 0–1 day
given number of cycles and the effects are recorded in Class 5 : 2–3 days
 ENGINEERING GEOLOGY/Rock Properties and Their Assessment 577

Table 5 Ethylene glycol soak test index values

Time class

Degree of 6 5 4 3 2 1 0
disintegration class 0–1 day 2–3 days 4–10 days 11–15 days 16–20 days 21–30 days Over 30 days

1 7 6 5 4 3 2 1
2 8 7 6 5 4 3 2
3 9 8 7 6 5 4 3
4 10 9 8 7 6 5 4
5 11 10 9 8 7 6 5

Class 4 : 4–10 days

Class 3 : 11–15 days
Class 2 : 16–20 days
Class 1 : 21–30 days
Class 0 : Over 30 days
An index value can be derived from the integration
of the degree of disintegration and time taken to reach
the worst condition. In other words, a scale of relative
durability values can be developed, ranging from 1 in
which specimens are more or less unaffected to 11
when they are totally disintegrated (Table 5). The
ethylene glycol soak test has been used to distinguish
certain unsound basalts and dolerites, commonly
described as slaking types, from sound types, Figure 13 Slake-durability test apparatus.
particularly in relation to their use as road aggregate.
Slaking refers to the breakdown of rocks, especially
mudrocks, by alternate wetting and drying. If a frag-
ment of mudrock is allowed to dry out, air is drawn percentage. The following scale of slake-durability is
into the outer pores and high suction pressures de- used:
velop. When the mudrock is next saturated the en- under 25%, very low
trapped air is pressurised as water is drawn into the 25–50%, low
rock by capillary action. This slaking process causes 50–75%, medium
the internal arrangement of grains to be stressed. 75–90%, high
Given enough cycles of wetting and drying, break- 90–95%, very high
down can occur as a result of air breakage, the pro- over 95%, extremely high.
cess ultimately reducing the mudrock to gravel-sized However, it has been suggested that durable mudrocks
fragments. The slake-durability test estimates the re- may be better distinguished from non-durable types
sistance to wetting and drying of a rock sample, and is on the basis of compressive strength and three-cycle
particularly suitable for mudrocks and shales. The slake-durability index (i.e. those mudrocks with a
sample, which consists of 10 pieces of rock, each compressive strength of over 3.6 MPa and a three
weighing about 40 g, is placed in a test drum, oven cycle slake-durability index in excess of 60% are
dried, and weighed. After this, the drum, with regarded as durable). In fact, the value of the slake
sample, is half immersed in a tank of water and durability test as a means of assessing mudrock dur-
attached to a rotor arm, which rotates the drum for ability has been questioned as the results obtained
a period of 10 min at 20 revolutions per minute frequently do not compare well with those of other
(Figure 13). The cylindrical periphery of the drum is durability tests. This has led to a number of adapta-
formed using a 2 mm sieve mesh, so that broken down tions being made to the test, for example, ethylene
material can be lost whilst the test is in progress. After glycol has been used instead of water, the time taken
slaking, the drum and the material retained are dried to carry out the test has been extended and the number
and weighed. The slake-durability index is then of cycles has been increased.
obtained by dividing the weight of the sample retained Failure of weak rocks occurs during saturation
by its original weight, and expressing the answer as a when the swelling pressure (or internal saturation
578 ENGINEERING GEOLOGY/Rock Properties and Their Assessment

swelling stress, ss) developed by capillary suction pres- number of tests may be conducted that provide an
sures, exceeds their tensile strength. An estimate of ss indication of the swelling behaviour of such rock.
can be obtained from the modulus of deformation (E): These include the swelling strain index test performed
under unconfined conditions, and the swelling pres-
E ¼ ss =eD ½11 sure index test, carried out under conditions of zero
volume change. Swelling strain measurements can
where eD is the free swelling coefficient. The latter is
also be undertaken on radially confined specimens
determined by a sensitive dial gauge recording the
under various conditions of axial loading. The swell-
amount of swelling of an oven-dried core specimen
ing strain represents maximum expansion of an un-
per unit height along the vertical axis during saturation
confined rock specimen when it is submerged in
in water for 12 h, eD being obtained as follows:
water. Test specimens may either be oven-dried or
change in length after swelling retain their natural moisture content. As a test speci-
eD ¼ ½12 men is submerged in water it is advisable to prevent
initial length
rock prone to slaking from collapsing before the ul-
A durability classification has been developed based on timate swelling strain has developed by wrapping the
the free-swelling coefficient and uniaxial compressive specimen in muslin. During the test the specimen is
strength (Figure 14). supported by a frame between a fixed point and a
When the free expansion of rocks liable to swell measuring point. The latter may consist of a dial
is inhibited, stresses of sufficient magnitude to cause gauge or a deformation transducer. The specimen
damage to engineering structures may develop. A and frame are placed in a container that is filled

Figure 14 Geodurability classification chart (After Olivier HJ (1979) A new engineering-geological rock durability classification.
Engineering Geology 4: 255–279).
 ENGINEERING GEOLOGY/Rock Properties and Their Assessment 579

with water just above the level of the specimen. Strain particular rock is defined by its coefficient of permea-
is measured in all three perpendicular directions. No bility or hydraulic conductivity. Grades of permeability
strain is allowed to develop in the test specimen are given in Table 6.
during the swelling pressure test. This is accom- Determination of the permeability of many rock
plished by tightly mounting a cylindrical specimen types in the laboratory is made by using a falling-
inside a rigid ring that provides radial constraint. head permeameter (Figure 15A). The sample is placed
The specimen absorbs water through porous end in the permeameter, which is then filled with water
plates. Any axial strain that develops is monitored to a certain height in the standpipe. The stopcock is
and compensated for by increasing the axial load. then opened and the water allowed to infiltrate the
The stress required to prevent expansion when equi-
librium conditions are established is equal to the
swelling pressure index.
Table 6 Grades of permeability (IAEG, 1979)
Permeability Permeability

Permeability considers the ability of a rock to allow the Class m s

1 Description
passage of fluids into or through it without impairing 1 Greater than 10
2
Very highly
its fabric. In ordinary hydraulic usage, a substance is 2 10
2–10
4 Highly
called ‘permeable’ when it permits the passage of a 3 10
4–10
5 Moderately
measurable quantity of fluid in a finite period of time 4 10
5–10
7 Slightly
and ‘impermeable’ when the rate at which it transmits 5 10
7–10
9 Very slightly
6 Less than 10
9 Practically impermeable
that fluid is slow enough to be negligible under existing
temperature-pressure conditions. The permeability of a
International Association of Engineering Geology.

Figure 15 (A) Falling head permeameter. (B) Radial percolation test apparatus.
580 ENGINEERING GEOLOGY/Site and Ground Investigation

sample, the height of the water in the standpipe increase in permeability as the pressure attributable
falling. The times at the beginning, t1, and end, t2, to divergent flow is increased.
of the test are recorded and these, together with
the two corresponding heights, h1 and h2, the cross- See Also
sectional area of the standpipe, a, and cross-sectional
area, A, and length, l, of specimen are substituted in Aggregates. Engineering Geology: Codes of Practice;
the following expression to derive the coefficient of Natural and Anthropogenic Geohazards; Problematic
permeability, k: Rocks.

2:303al Further Reading

k¼  ðlog10 h1
log10 h2 Þ ½13
Aðt2
t1 Þ
Anon (1975) Methods of Sampling and Testing Mineral
Aggregates, Sands and Fillers, BS 812. London: British
Variations of permeability in rocks under stress Standards Institution.
can be obtained by using a radial percolation test. Anon (1982) Standard Test Methods for Absorption and Bulk
A cylindrical specimen, in which an axial hole is Specific Gravity of Natural Building Stone, C93–117.
drilled, is placed in the radial percolation cell. The Philadelphia: American Society for Testing Materials.
latter can either contain water under pressure where Bell FG (ed.) (1992) Engineering in Rock Masses. Oxford:
the axial hole is in contact with atmospheric pressure; Butterworth-Heinemann.
or water can be injected under pressure into the hole Bell FG (2000) Engineering Properties of Soils and Rocks.
(Figure 15B). The flow is radial over almost the whole Oxford: Blackwell Science.
height of the sample and is convergent when the Brown ET (ed.) (1981) Rock Characterization, Testing and
Monitoring. Oxford: Pergamon Press.
water pressure is applied to the outer face of the
Farmer IW (1983) Engineering Behaviour of Rocks, 2nd ed.
specimen, and divergent when the water is under
London: Chapman and Hall.
pressure within the axial hole. Porous rocks remain Goodman RE (1989) An Introduction to Rock Mechanics,
more or less unaffected by pressure changes. On 2nd ed. New York: Wiley.
the other hand, fissured rocks exhibit far greater Hudson JA and Harrison JP (1997) Engineering Rock
permeability in divergent flow than in convergent Mechanics: An Introduction to the Principles. Oxford:
flow. Moreover, fissured rocks exhibit a continuous Pergamon.

Site and Ground Investigation

J R Greenwood, Nottingham Trent University, site under consideration and the interaction with
Nottingham, UK the surrounding areas. It is not confined to obtaining
ß 2005, Elsevier Ltd. All Rights Reserved. information on geotechnical aspects but may in-
clude hydrological, meteorological, geological, and
environmental investigation.
Introduction and Terminology . ‘Ground investigation’ is more site-specific and
The procedure of ‘investigation’ is fundamental to aims to investigate ground and groundwater con-
any project or activity involving the ground. The ditions in and around the site of a proposed
historical records need to be reviewed, current condi- development or an identified post-construction
tions need to be established, and the consequences of problem.
the proposed activity, works, or construction need to
The term ‘site characterization’ is now also used; it
be carefully considered.
stems from the environmental specialist’s study of
Investigation is an on-going process of establishing
contaminated sites but is equally applicable to any
and reviewing the facts and processing the informa-
site. ‘Characterization’ perhaps implies the results of
tion to assist our future activities. With respect to
‘investigation’.
construction works the following definitions are used:
This article reviews the procedures necessary for
. ‘Site investigation’ is a continuous process, as the quality site investigation to be carried out and de-
construction project develops, involving both the scribes some of the ground investigation techniques