University of kirkuk – College of engineering

Civil department – Third stage A

Triaxial shear behavior of a cement-treated

sand and gravel mixture

Supervised by: Dr.Aram Mohammed

Prepared by: Mahmood Younis Mohammed (A)
Abstract:
A number of parameters, e.g. cement content, cement type, relative density, and grain
size distribution, can influence the mechanical behaviors of cemented soils. In the
present study, a series of conventional triaxial compression tests were conducted on a
cemented poorly graded sand gravel mixture containing 30% gravel and 70% sand in
both consolidated drained and untrained conditions. Portland cement used as the
cementing agent was added to the soil at 0%, 1%, 2%, and 3% (dry weight) of sand
gravel mixture. Samples were prepared at 70% relative density and tested at confining
pressures of 50 kPa, 100 kPa, and 150 kPa. Comparison of the results with other
studies on well graded gravely sands indicated more dilation or negative pore pressure
in poorly graded samples. Untrained failure envelopes determined using zero
Skempton’s pore pressure coefficient ðA ¼ 0Þ criterion were consistent with the
drained ones. Energy absorption potential was higher in drained condition than
untrained condition, suggesting that more energy was required to induce deformation
in cemented soil under drained state. Energy absorption increased with increase in
cement content under both drained and untrained conditions. _ 2014 Institute of Rock
and Soil Mechanics, Chinese Academy of Sciences. Production and hosting by
Elsevier B.V. All rights reserved.

1. Introduction
Cementation occurs due to the various geological processes that create bonds between
soil particles like aging, chemical reactions, carbonates, silicates, iron oxides, and
natural cementing agents. Due to the difficulties of in situ sampling, the mechanical
characteristics of cemented soils are usually studied using artificial samples prepared
in laboratory and cured by different cementing agents. The mechanical behavior of
cemented soils is influenced by a number of parameters including cement content,
cement type, density, confining stress, grain size, and stress-strain history (Saxena and
Lastrico, 1978; Clough et al., 1979, 1981, 1989; Acar and El Tahir, 1986; Leroueil and
Vaughan, 1990; Airey, 1993; Coop and Atkinson, 1993; Das et al., 1995; Malandraki
and Toll, 1996; Cuccovillo and Coop, 1997, 1999; Huang and Airey, 1998; Consoli et
al., 2000, 2007, 2009, 2010, 2011; Schnaid et al., 2001; Ismail et al., 2002; Rotta et al.,
2003; Lee et al., 2010; Park, 2010; Baxter et al., 2011; Hamidi and Hooresfand, 2013;
Shahnazari and Rezvani, 2013). According to the previous studies, the cementation
can increase brittleness, shear strength, and dilative behavior of sands. However, it
should be noted that most of the previous studies have focused on the mechanical
behavior of cemented fine sands, rather than the mechanical behavior of coarse
grained gravels or gravely sands. In the last decade, a number of studies have been
performed to investigate the mechanical behavior of cemented gravely sands or sandy
gravels (Haeri et al., 2005a,b, 2006), in which a series of triaxial compression tests
were performed on a representative gradation of the Tehran coarse grained alluvium.
Cemented samples were prepared with different cementing agents including lime,
1
Portland cement, and gypsum. They concluded that the strain associated with the peak
deviatory stress decreases as the cementation increases. Also it was indicated that the
maximum rate of dilation and negative pore water pressure occur after the maximum
shear strength is obtained. Review of the literature shows that there is rarely particular
study on investigation of the mechanical behavior of cemented poorly graded sand
gravel mixtures. Indeed, all previous studies on the behavior of cemented gravely
sands concern fine sands or well graded gravely sands as the base soil. Therefore, the
objective of present research is to investigate the mechanical behavior of a cement-
treated poorly graded sand gravel mixture. In this regard, a number of new features of
the mechanical behavior of cemented soils are reported.

2. Testing program
Thirty groups of conventional triaxial compression tests were performed, among
which 24 groups are considered in this study. Six groups of additional tests (25% of
the total) were performed to check repeatability of the experiments and results. Also,
12 groups of unconfined compression tests were conducted and reported. Cement
content and confining pressure were considered as the variables of testing program,
and the triaxial compression tests were performed in the consolidated drained and
untrained conditions.

2.1. Soil and cementing agent

Clean and uniform quartz beach sand with sub-round to sub angular particles from the
shores of the Caspian sea (specifically Babolsar, Iran) was first sieved using a #30
sieve and then was mixed with 30% uni-sized (9.5e12.5 mm) gravel grains. The mixed
soil can be named as SP in unified soil classification system and was used as the base
material. Gradation curves and physical properties of the base material are shown in
Fig. 1 and Table 1, respectively. In Table 1, Gs is the specific gravity; D10 is the
effective diameter; CU and CC are the uniformity and curvature coefficients,
respectively; and gd, min and gd, max are the minimum and maximum unit weights,
respectively. All physical characteristics were determined according to the ASTM
(1998) standard methods. Portland cement (Type II) with a setting time lasting for
about 4 h was used as the cementing agent. It was first sieved using a #100 sieve and
then added to the base soil.

2.2. Sample preparation and test procedure

The undercompaction method was used for sample preparation proposed by Ladd
(1978). Required weight of the soil was mixed with desired cement content and about
7% distilled water. Samples were prepared using a split mold, 100mmin diameter and
200mm in height, and were compacted in eight layers. Each layer was poured into the
mold and compacted using metal hammer until the desired height was reached.
Cemented samples were stored at a (25 3) C humid room with >90% relative

2
humidity for 24 h. After that, samples were extracted from the mold and were kept for
6 d at humid room. On the 7th day, the diameter, height and weight of the samples
were measured. For unconfined compression tests, samples were prepared with curing
times of 7 d and 28 d. The variables considered in sampling process are listed in Table
2. A computer-controlled triaxial cell was used to test the samples at the confining
pressures (CPs) of 50 kPa,100 kPa, and 150 kPa. The outer surface of samples was
soft enough to minimize the effect of membrane penetration. As a result, flexible
membranes do not affect pore pressure generation in saturated condition. Membranes
with average thickness of 1 mm were used and corrections such as

membrane thickness and cross-sectional area were considered according to Bishop and
Henkel (1969). All samples were fully saturated in two stages prior to shearing. At the
first stage, de-aired water was flushed from the bottom of sample under a very low
pressure difference of 10 kPa for 24 h. After that both cell and back pressure were
ramped simultaneously to 310 kPa and 300 kPa for complete saturation at the second
stage. Saturation procedure was considered to be completed until Skempton’s B value
of 0.9 was reached. The samples were consolidated up to the desired confining
pressures. Shear loading was applied at an axial strain rate of 0.1 mm/min for drained
tests and 0.3 mm/min for undrained ones. Cell pressure, volume change, pore pressure,
load and displacements were measured during triaxial compression tests by electronic
transducers and a calibrated data acquisition system. All the variables considered in
testing program are listed in Table 3.

3
3. Unconfined compression tests
Some specifications of the unconfined compression tests are described in Table 3. Fig. 2 indicates the
variation of unconfined compressive strength (UCS) with cement content at different curing times. Peak
strength occurred at small strains between 0.2% and 0.7%. It can be observed from Fig. 2 that the UCS
increases with increasing cement content and elapsed curing time. The lines intersect horizontal axis at cement
content about 0.5%, which is the minimum cement content to mobilize the shear strength of cemented soil and
formation of cemented bonds.

4. Analysis of the results

A summary of triaxial test results at failure and residual state is shown in Table 4.
Deviatoric stress (q), mean effective stress (p’) and specific volume (n) are defined
using the following equations:

Fig. 2. Variation of the unconfined compressive strength with cement content at

different curing times.

where s0 1 is the major effective principal stress, s03 is the minor effective principal
stress, and e is the void ratio.

4.1. Mode of failure

Fig. 3 shows the typical failure modes of the uncemented and cemented samples.
Although dilation occurred at different confining pressures, all uncemented samples in
the drained and undrained tests showed barreling mode without shear plane formation.
In lightly cemented samples (CC ¼ 1%), failure mode was a combination of barreling
shape and shear plane, although shear band was not obvious and barreling was the
predominant mode. Increase in cement content increased the thickness of the shear
band. Cemented samples with more cement content (CC > 1%) experienced a mode of
brittle failure and underwent significant dilation with apparent peak point in stresses
train curve. In the
drained and undrained conditions, cemented soils showed a shear band with a
thickness of 3e6 cm. Inclination of the shear band with horizontal axis decreased from
70_ to 55_ with increase in confining pressure from 50 kPa to 150 kPa.

4.2. Stress train behavior

Deviatoric stress axial strain curves are depicted in Figs. 4a, 5a, 6a and 7a for the
consolidated drained and undrained conditions. The drained and undrained tests are
indicated by the symbols “D” and “U”, respectively, and are followed by a number
that represents the value of confining pressure in kPa. Six tests were repeated twice to

4
check the repeatability of experiments, and the average stresse strain diagram is
plotted in these figures. All cemented samples in drained and undrained conditions
showed an apparent peak point

5
associated with the failure. After that, slope of the stress strain curve decreased
to its residual value in an axial strain of about 20%. For the uncemented samples
under undrained condition, the peak stress was not obvious and the softening
6
behavior was not as clear as the drained tests. Increase in cement content and
reduction in confining pressure caused more softening in stress-strain curve.
Comparison of the results in the drained and undrained conditions showed that the
strain associated with the peak strength was larger

7
in undrained tests than that in drained ones. It can be concluded that the cemented soil
behavior is more brittle in the drained condition and bond degradation occurs easier
when volumetric strains can freely occur in the soil (Malandraki and Toll, 2001).
In order to understand the effect of confining pressure on the shear strength of
cemented soil, the maximum deviatoric stress (qmax) is normalized to the uncemented
one in drained tests in Fig. 8. It can be seen from Fig. 8 that, for lightly cemented
sample (CC ¼ 1%), normalized shear strength was approximately constant under
different confining pressures. However, the effect of confining pressure on the shear
strength ratio increased for other cement contents.

8
confining pressure increased contractive behavior at the start of the test. Figs. 4c, 5c,
6c and 7c indicate the variation of the excess pore pressure with the mean effective
stress in the consolidated undrained tests. Positive pore pressure occurred at the
beginning of loading, followed by significant negative pore pressure at the final state.
Same as the volume change, increase in cement content and decrease in confining
pressure increased negative suction at the end of loading process. Haeri et al. (2005b)
reported the results of consolidated drained triaxial tests on a well graded gravely sand
(containing 45% gravel) cemented with Portland cement. The gradation curve is
shown in Fig. 1 and the maximum gravel size is 12.5 mm. Fig. 9 shows the
comparison between the dilation and excess pore pressure values for the present study
and well graded gravely sand tested by Haeri et al. (2005b). According to the previous
studies, increase in gravel content increases dilation in sandegravel mixtures (Evans
and Zhou, 1995; Simoni and Houlsby, 2006; Hamidi et al., 2009). Although gravel
content is larger inwell graded samples, dilation in drained state or negative pore
pressure in undrained condition is lower compared with the poorly graded mixture
tested. For the same gravel content, it can be concluded that dilation in drained
condition or negative pore pressure in undrained state is larger in poorly graded
sandegravel mixtures compared with the cemented well graded gravely sands.

4.4. Stress paths

Stress paths for different cement contents are shown in Figs. 4d, 5d, 6d and 7d. The
stress path moved linearly with a slope of 3:1 in q-p’ space in the consolidated drained
tests. It reached a peak point which has been marked on the figures. After that,
softening caused reversal of the stress path with the same slope until residual stress
state was reached. In undrained condition, the peak point of the stress path was higher
than the drained one due to generation of significant negative pore pressure. However,
the difference between peak points of stress paths for the drained and undrained

9
tests decreases by increase in confining pressure. The same trend has also been
reported for a gypsum cemented well graded gravely sand (Haeri et al., 2005a);
however, peak points of the stress path in undrained condition were lower than the
drained ones at high
confining pressures (over 300 kPa). Fig. 10 shows the axial strain contours of
cemented samples containing 2% cement in the drained and undrained conditions.
Peak shear stress has been observed in an axial strain level about 1% in drained
condition and 3% in undrained condition. Results of triaxial tests on the cemented well
graded gravely sand indicate that peak shear stress is associated with axial strains of
2.5% and 4.5% in drained and undrained states, respectively (Haeri et al., 2005b). It
confirms more brittle behavior of cemented poorly graded sandegravel mixture
compared with the cemented well graded gravely sand in the same conditions.

10
4.5. Failure envelopes and shear strength parameters
Fig. 11 plots failure envelopes for all the tests. Numbers after the symbols “D” and “U” show the value of
cement content. Previousstudies have suggested a curved failure envelope for cemented materials (Malandraki
and Toll, 2001; Asghari et al., 2003; Baker, 2004; Sharma et al., 2011). However, the failure envelopes were
not curved for the studied soil. Moreover, the failure envelopes of drained tests were lower than those of
undrained ones which can be related to the large negative pore pressures induced
during undrained shearing. Different criteria can be used to determine the shear
strength parameters for the soil, including peak deviatoric stress at failure, maximum
principal stress ratio, maximum excess pore pressure, and zero Skempton’s pore
pressure parameter ðA ¼ 0Þ. The peak deviatoric stress is a commonly used method
for determination of the failure envelope in drained condition. In undrained state,
Baxter et al. (2011) recommended using A ¼ 0 as the failure criterion for cemented
soils, because it can eliminate the effects of large pore pressure gradients on the shear
strength. Fig. 12a plots failure envelopes for drained condition based on the peak
deviatoric stress and Fig. 12b shows undrained failure envelopes based on both peak
deviatoric stress and A ¼ 0 criteria. Table 5 also lists the shear strength parameters, c0
and 40.

According to Table 5, using A ¼ 0 criterion in undrained condition

yields consistent friction angle and cohesion intercept values with drained state.
Baxter et al. (2011) mentioned that increase in cement content increases the

11
differences between cohesion intercept values calculated in the drained and undrained
conditions using A ¼ 0 criterion. However, results of the present study show that
increase in cement content does not particularly influence the relation between shear
strength parameters in the

drained and undrained conditions. It can be related to the difference in gradation of the
cemented sandy gravel used in the present study and the fine cemented sand tested by
Baxter et al. (2011). Fig. 13 depicts the failure envelopes and residual state lines for
different cement contents in drained condition. Residual state line shows the relation
between deviatoric stress and mean effective stress in the final stage of loading. It can
be seen from Fig. 13 that the difference between the two lines increases with
increasing cement content, which confirms more brittle behavior as cement content
increases. Fig. 14 shows variation of the effective principal stress ratio at failure ðs0
12
1=s0 3 Þ f for drained and untrained conditions with confining pressure. The effective
principal stress ratio at failure decreases with increasing confining pressure and
increases with increasing cement content. Also, it is much higher in untrained state
than drained state, especially at lower confining pressures.

5. Stiffness and energy absorption

Stiffness of cemented soil was determined by calculating the secant modulus for half
shear strength in different confining pressures as shown in Fig. 15. We can note that
the stiffness increases with increase in cement content and confining pressure. The
drained stiffness is larger than the undrained one; however, the difference is less than
10%. Required energy to induce deformation in cemented soil can be calculated by the
area under stress-strain curve. In the present study, absorbed energy for different axial
strains is normalized to the absorbed energy at 10% axial strain. Fig. 16 shows the
variation of the normalized absorbed energy for different cement contents and
confining pressures in the drained and undrained conditions. A major change in the
slope of the curve can be observed at 2% axial strain in drained state. As the cement
content decreases, slope of the curve became flattened and reached to the curve of the
uncemented samples. For the undrained condition, the slope of the curve was lower
than that of the drained one. Also, the change in the slope of energy absorption curve
can be observed at axial strains more than 2%.

Conclusions:
The present study deals with the engineering characteristics of cemented poorly
graded sand & gravel mixtures. The following conclusions can be drawn based on the
test results: (1) Under the similar conditions (the same cement type, cement content,
gravel content, and maximum gravel size), dilation and negative pore pressure induced
in cemented poorly graded sand & gravel mixture were larger than those of well
graded gravely sand. (2) Brittle behavior of cemented soils was more obvious in
drained condition than in untrained state. The axial strain at the peak shear strength
was smaller in drained condition than in untrained state. Failure envelopes were lower
in the drained condition than in the untrained state. (3) Shear strength parameters
calculated under the drained and untrained conditions were consistent when untrained
shear parameters were calculated based on zero Skempton’s coefficient ðA ¼ 0Þ
criterion. Consistency between the results can be observed for all cement contents. (4)
A major change in the slope of the absorbed energy curve was observed at axial strain
of about 2% in the drained condition. As cement content decreases, the slope of the
curve became flattened and reached to the curve of the uncemented soil. The change in
the slope of energy absorption curve was minor in untrained condition at axial strains
more than 2%.

Conflict of interest

13
The authors wish to confirm that there are no known conflicts of interest associated
with this publication and there has been no significant financial support for this work
that could have influenced its outcome.

References:
Acar Y, El-Tahir A. Low strain dynamic properties of artificially cemented sand.
Journal of Geotechnical Engineering 1986;112(11):1001e15.
Airey DW. Triaxial testing of naturally cemented carbonate soil. Journal of
Geotechnical Engineering 1993;119(9):1379e98.
Asghari E, Toll DG, Haeri SM. Triaxial behavior of a cemented gravelly sand Tehran
alluvium. Geotechnical and Geological Engineering 2003;21(1):1e28.
American Society for Testing and Materials (ASTM). Annual book of ASTM standards:
soils and rock. West Conshohocken, Philadelphia: ASTM; 1998.
Baxter CDP, Sharma MSR, Moran K, Vaziri H, Narayanasamy R. Use of ðA ¼ 0Þ as a
failure criterion for weakly cemented soils. Journal of Geotechnical and Geoenvironmental
Engineering 2011;137(2):161e70.
Baker R. Nonlinear Mohr envelopes based on triaxial data. Journal of Geotechnical
and Geoenvironmental Engineering 2004;130(5):498e506.
Bishop AW, Henkel DJ. The measurement of soil properties in triaxial tests. London:
Edward Arnold Ltd.; 1969.
Clough GW, Kuck WM, Kasali G. Silicate-stabilized sands. Journal of the Geotechnical
Engineering Division 1979;105(1):65e82.
Clough GW, Sitar N, Bachus RC, Rad NS. Cemented sands under static loading.
Journal of the Geotechnical Engineering Division 1981;107(6):799e817.
Clough GW, Iwabichi J, Rad NS, Kuppusamy T. Influence of cementation on liquefaction
of sands. Journal of Geotechnical Engineering 1989;115(8):1102e17.
Consoli NC, Rotta GV, Prietto PDM. Influence of curing under stress on the triaxial
response of cemented soils. Geotechnique 2000;50(1):99e105.
Consoli NC, Foppa D, Festugato L, Heineck KS. Key parameters for strength control
of artificially cemented soils. Journal of Geotechnical and Geoenvironmental
Engineering 2007;133(2):197e205.
Consoli NC, Viana da Fonseca A, Cruz RC, Heineck KS. Fundamental parameters for
the stiffness and strength control of artificially cemented sand. Journal of
Geotechnical Engineering 2009;135(9):1347e53.
Consoli NC, Cruz RC, Floss MF. Parameters controlling tensile and compressive
strength of artificially cemented sand. Journal of Geotechnical and Geoenvironmental
Engineering 2010;136(5):759e63.
Consoli NC, Cruz RC, Floss MF. Variables controlling strength of artificially cemented
sand: influence of curing time. Journal of Materials in Civil Engineering
2011;23(5):692e6.
Coop MR, Atkinson JH. The mechanics of cemented carbonate sands. Geotechnique
1993;43(1):53e67.
Cuccovillo T, Coop MR. Yielding and pre-failure deformation of structured sands.
Geotechnique 1997;47(3):481e508.
Cuccovillo T, Coop MR. On the mechanics of structured sands. Geotechnique
1999;49(6):741e60.
Amir Hamidi is an Associate Professor in the School of Engineering
at Kharazmi University of Tehran

14