One of important subject in geotechnical engineering that has been evaluated
extensively over long period of time is shear strength of soils (Mesri
and Shahien, 2003, 2004; Hamel, 2004; Yudbir, 2004; Duncan, 2001a, b;
Christian and Baecher, 2001; Gilbert et al., 1998). The application
and importance of shear strength of soils in practice have also been challenged
in many geotechnical and geoenvironmental projects around the world (Filz
et al., 2001; Eid et al., 2000; Zornberg et al.,
1998a, b). Of many factors affecting shear strength of soils directly
or indirectly, density, effective stress and soil structure are the most
important (Terzaghi et al., 1996). Obtainable density under geological
conditions is related mainly to the size, shape, surface characteristics
and strength of particles constituent of soil. The effect of type of mineral
(composition) of soil particles and physiochemical environment on shear
strength is indirectly through their control of these important particle
characteristics (Terzaghi et al., 1996).
The relationship between shear strength and effective normal stress for
natural clays and clay shales is curved and concave to the effective normal
stress axis (Stark and Eid, 1997; Mesri and Shahien, 2003) and there is
no shear strength at zero effective normal stress (Terzaghi et al.,
1996). The intact strength envelope of over consolidated clays and clay
shales displays a pronounced curvature because swelling and softening
intensify as effective normal stress decreases toward zero (Mesri and
Shahien, 2003). The empirical equation by Mesri and Abdelghafar (1993)
for intact strength envelope of over consolidated soils is:
which means the intact strength is higher than normally consolidated
strength of the same constituent assuming linear relationship between
shear strength effective normal stress for the latter. In
Eq. 1, σ`p is over consolidation pressure, σ`n
is effective normal stress on sliding surface and m is slope of failure
line log τf Vs log σ`n. The magnitude
of m depends on structure and composition of soil (Mesri and Abdelghafar,
1993). High curvature of failure envelope means lower m and for very low
curvature the magnitude of m nears one. Recently, low curvature failure
envelopes has been also reported for fully softened and residual states
(Stark and Eid, 1997; Mesri and Shahien, 2003). The fully softened strength
envelope displays a curvature because, even for a random arrangement of
particles, high effective normal stresses promote face to face interaction
of plate shaped particles. The residual strength envelope is curved because
a higher degree of particle orientation in direction of shearing is possible
at high effective normal stresses (Mesri and Shahien, 2003).
In this study, the effect of composition, structure and degree of cementation
on parameter m using results of direct shear tests on natural and compacted
laboratory over consolidated soil samples is evaluated.
MATERIALS AND METHODS
In this research, both natural over consolidated samples from Khuzestan
province and compacted laboratory over consolidated soil samples at optimum
water content were tested.
Natural Soil Specimen Preparation
Natural soils tested were Shelby tubes samples of Behbahan clay obtained
from depth of 4 m below the surface and core samples of Ahwaz clay from
depth of 24 m. Samples of Behbahan clay were obtained from the site of
faculty of Engineering of Behbahan and those of Ahwaz clay from site of
Ahwaz metro in Golestan area. Consolidation tests were performed on samples
to determine over consolidatation pressure of soil. Direct shear tests
were performed to determine failure envelope. Consolidation specimens
were carefully placed into the ring using surgical blade and after transferring
to loading frame were loaded incrementally with load increment ratio until
the end of primary consolidation to maximum pressure of 1600 kPa.
Direct shear specimens were placed into the shear box and consolidated
under normal stress of 24, 48, 192 and 347 kPa before they were sheared
at the rate of 1 mm min-1. the lowest rate of the machine.
Mesri and Abdelghafar (1993) showed that c` (cohesion intercept) obtained
from triaxial and direct shear tests for 25 natural clays compared well.
This and the fact that direct shear test would give the result of drained
test in a faster time than triaxial test were the primary reasons to choose
this type of equipment for our testing program.
In order to obtain normally consolidated state failure envelope of natural
soil constituent, samples were prepared following the procedure given
by Terzaghi et al. (1996). In this procedure samples were dried
ball milled and passed from the No. 200 US sieve before they were mixed
with water at water content equal to liquid limit. Then the samples were
transferred into direct shear box and were sheared under normal stress
of 24, 48, 96, 192 and 347 kPa.
Preparation of Artificial Compacted Soil Samples
Compacted soil samples tested were mixtures of 90% sand 10% bentonite,
85% sand, 15% bentonite, 80% sand, 20% bentonite and 70% sand, 30% bentonite.
In all mixtures sand portion was finer than No.40 US sieve. For mixture
of 80% sand, 20% bentonite sand finer than No.10 US sieve was also used.
All mixtures were compacted at their optimum water content using the standard
Proctor procedure and consolidated to the maximum pressure of about 1600
After consolidating to the maximum pressure and before dismantling the
test, each specimen was unloaded in two steps. In first step, the specimen
was unloaded to the pressure that was equal to the normal effective stress
at which it was going to be sheared and allowed to swell until primary
swelling. Then the water in the oedometer was emptied and the specimen
was unloaded to seating pressure, taken out from the cell and quickly
placed into shear box and the pressure at which the samples was unloaded
in the first stage was applied on the specimen before it was sheared to
failure. The over consolidated state failure envelopes for different mixtures
in the range of effective normal stresses of 24, 48, 96, 182 and 350 kPa
were determined in this way.
Sample preparation for bentonite-sand mixtures for determining failure
envelope of normally consolidated state was the same as for natural soil
samples explained earlier and it is not repeated here. However because
of long period of time required for performing these tests especially
for samples with higher bentonite content they were performed only on
mixture of 90% sand, 10% bentonite. Therefore, failure envelope for normally
consolidated state of other mixtures were obtained using empirical relationships
between φ` and Ip (Terzaghi et al., 1996).
Index Properties of Samples Tested
Results of liquid and plastic limits, natural water content and hydrometry
tests for natural soil samples are shown in Table 1
and for bentonite-sand mixtures are shown in Table 2.
In Table 2 optimum water content at which test samples
were prepared are also given.
Consolidation Tests Natural Samples
Results of consolidation tests on undisturbed samples of Ahwaz and
Behbahan clays in the from of end of primary e-log σ`v
are shown in Fig. 1 and 2. In Table
3, consolidation parameters such as recompression index Cr,
compression index Cc, initial void ratio e0 and
over consolidation pressure for
these two clays are given.
||Index properties of natural soils
||Index properties of bentonite-sand mixtures
||Consolidation properties of natural soils
||EOP e-logp` relation for Ahwaz clay
||EOP e-logp` relation for Behbahan clay
The values shown in Table 3
are in the range of values
reported in literature for over consolidated soils and shales (Terzaghi
., 1996). Over consolidation pressures are obtained using Cassagrandes
method. Over consolidation pressure of the Behbahan clay from depth of 4
m is estimated about 150 kPa and that for the Ahwaz clay from depth of 24
m is about 560 kPa.
Direct Shear Tests Natural Samples
Plots of horizontal displacement versus vertical displacement for
direct shear tests on the natural clays of Behhahan and Ahwaz are shown
in Fig. 3 and plots of τ versus δh
are shown in Fig. 4. These clays during shear at low
normal stresses expanded and at high normal stresses compressed, typical
behavior of over consolidated clays.
Failure Envelopes for Natural Samples
In order to plot intact failure envelopes of the natural clays of
Ahwaz and Behbahan values of τmax are plotted against
σ`n and are shown in Fig. 5. The best
curve is fitted to test data points. In Fig. 5 failure
envelopes for natural sample composition in normally consolidated state
are also shown. The normally consolidated state envelopes for both clays
is linear and those for natural samples are curved as expected. It is
noted that overconsolidated and normally consolidated states failure envelopes
for the Behbahan clay merged at a pressure of 150 kPa equal to overconsolidation
pressure obtained in oedometer test. This was only true for the Ahwaz
clay when the normally consolidated state failure envelope was obtained
using the empirical relation of Ip–φ` by Terzaghi
et al. (1996). Mesri and Abdelghafar (1993) indicated that overconsolidation
pressure obtained in oedometer and that obtained from failure envelope
were the same for most clays they reviewed but for some others they were
||δv-δh relation for (a) Ahwaz and
(b) Behbahan clay
||τ-δh relation for (a) Ahwaz and (b) Behbahan
||τ-σ`n relation for (a) Ahwaz and (b) Behbahan
Determination of Parameter m for Natural Samples
In order to determine parameter m for Ahwaz and Behbahan clays, log
τmax are plotted against log τ`n in Fig.
6. Parameter m for the Ahwaz clay is about 0.61 and for the Behbahan
clay is 0.8. These values are in the range of values reported for stiff
clays and shales with intact and disturbed structure (Terzaghi et al.,
1996). The Behbahan clay sample from depth of 4 m may have acquired disturbed
structure in nature or during sampling or laboratory preparation; in fact
we had difficulty taking the sample out from Shelby tube. According to
Mesri and Shahien (2003) there can be a wide variation in the intact strength
at any effective normal stress because the stiff clay or clay shale may
experience different degree of fissuring and softening during its geological
log τ-log σ`n relation for (a) Ahwaz and
(b) Behbahan clay
||EOP e-logp` relation for bentonite sand mixtures
Consolidation Test Sand Mixtures
The results of consolidation tests on bentonite-sand mixtures in the
from of end of primary e-log σ`v are shown in Fig.
Direct Shear Tests Sand Mixtures
The plots δv versus δh and τ
vs δh for direct shear tests on compacted over consolidated
bentonite sand samples are shown in Fig. 8 and 9,
respectively. All samples at all effective normal stresses tested first
compressed and then expanded during shear.
Over Consolidated and Normally Consolidated Failure Envelopes
The values of τmax versus σ`n for
over consolidated bentonite-sand mixtures are plotted in Fig.
10. The best curve is fitted through the test data. In Fig.
10 also failure envelopes for normally consolidated state of these
samples using empirical relationship between Φ` and Ip
from Terzaghi et al. (1996) are shown. For mixture of 90% sand,
10% bentonite, normally consolidated state failure envelope is obtained
from direct shear tests on remolded samples at water content equal to
liquid limit. All overconsolidated and normally consolidated failure envelopes
merged at overconsolidation pressure of about 1600 kPa at which all overconsolidated
samples experienced in oedometer test.
Determination of Parameter m
Log τmax versus log σ` n from direct
shear testes on bentonite-sand samples are sown in Fig.
11. The value of m falls between 0.62 to 0.8. The value of m is increased
with amount of sand.
δv-δh relation for bentonite
||τ-δh relation for bentonite sand mixtures
||τ-σ`n relation for bentonite sand mixtures
in overconsolidated and normally consolidated state
||log τ-log σ`n relation for bentonite sand mixtures
in overconsolidated state
The Effect of Ip on m
In Fig. 12 values of m obtained for natural and
compacted samples are plotted against Ip. As indicated also
by Mesri and Abdelghafar (1993) value of m is decreased with Ip. The values
of m for natural samples falls below the trend for compacted samples.
Compacted samples lack bondings due to cementation, diagenetic and etc.
that exist for natural samples.
||Coefficient m-plasticity index relation for natural clays and compacted
overconsolidated bentonite sand mixtures
||Coefficient m-S1 relation for natural and compacted overconsolidated
bentonite sand mixtures
||Coefficient m-Friction angle, φ` for bentonite sand mixtures
Relation of m with s1
In Fig. 13, the coefficient of m is plotted
against s1, the intercept of log τ– logσ`n
for natural clays and bentonite sand mixtures. A decreasing trend in m
is observed with increasing s1.
Relation of m with φ`
In Fig. 14 the coefficient of m is plotted against
angle of friction ö` obtained from Ip- φ`
empirical relation from Terzaghi et al. (1996) for bentonite
sand mixtures. An increasing trend of m with φ` is observed.
||τ-σ`n relation for cemented compacted and overconsolidated
Determination of Effect of Cementation on m
In order to quantify the influence of cementation on parameter m,
mixtures of 70% sand, 30% bentonite was mixed with 5% of Portland cement
and compacted at optimum water content and over consolidated to maximum
pressure of 1600 kPa. Two series of samples were prepared, one series
was cured for a period of 7 days and second series was cured for 28 days.
After curing period, samples were tested in direct shear. Failure envelope
for these samples and samples without cement are compared in Fig.
15. Values of m for these samples were 0.59 and 0.57 for 7 and 28
days curing times respectively as compared to 0.62 of untreated samples.
As it is shown cementation causes a decrease in value of m.
Natural over consolidated soil samples and compacted over consolidated
bentonite- sand mixtures at optimum water content were tested in direct
shear test in order to determine the parameter m. The range of values
of m obtained for natural samples in this research was in the range that
has been reported by others for stiff clays and shales. It is also indicated
that as the plasticity of soil increases the value of m decreases, another
word as the amount of sand in soil decreases, the value of m decreases.
The results of test on cemented samples showed that cementation also decreases
the value of m.