Soil Dynamics and Earthquake Engineering 65 (2014) 181–188

Contents lists available at ScienceDirect

Soil Dynamics and Earthquake Engineering

journal homepage: www.elsevier.com/locate/soildyn

Factors affecting shear modulus degradation of cement treated clay

P. Subramaniam a,n, Subhadeep Banerjee b,1
a
Geotechnical Engineering Division, Department of Civil Engineering, Indian Institute of Technology Madras, Chennai 600036, India
b
Department of Civil Engineering, Indian Institute of Technology Madras, Chennai 600036, India

art ic l e i nf o a b s t r a c t

Article history: Cement stabilization is often used to improve the bearing capacity and compressibility of soft clays.
Received 22 November 2013 The present paper aims to investigate the shear modulus degradation of cement treated clay during
Received in revised form cyclic loading. A series of cyclic triaxial test was conducted on artificially cement treated marine clay to
4 June 2014
study the factors affecting the shear modulus degradation. The parameters considered for the study are
Accepted 14 June 2014
cement content (2.5–7.5%), curing days (7–28), cyclic shear strain amplitude (0.3–1%), number of loading
cycles (1–100) and loading frequency (0.1–0.5 Hz). As in the case of natural clays, cement treated clays
Keywords: exhibit stiffness degradation which depends on mix ratio, curing days and loading conditions. The
Shear modulus results show that the shear modulus degradation decreases with increase in the shear strain amplitude,
Cyclic loading
cement content and curing days. It is also noted that irrespective of the mix ratio and curing conditions,
Cement treated clay
the degradation decreases with increase in loading frequency. An empirical relationship is proposed to
Empirical model
predict the shear modulus degradation based on Idriss's degradation model. The performance of the
proposed empirical model is validated with the present experimental results.
& 2014 Elsevier Ltd. All rights reserved.

1. Introduction as a function of plasticity index, overconsolidation ratio and cyclic

While many studies have been carried out on cement-treated
marine clay subjected to static loadings [1–3], the dynamic properties
of such treated soils are largely unknown. However in light of rapidly
rising demand for cement stabilization in soft soils for urban devel-
opment in conjunction with the relatively large number of recent
earthquakes, a thorough study of the dynamic properties of cement-
treated marine clay is of urgent need.
In the past, majority of the studies on dynamic properties of
cement treated marine clay was focused to determine the maximum
shear modulus at small strains either by the bender element technique
[4,5] or using the resonant column apparatus [6–8]. However the
cyclic degradation behaviour of cement treated clay subjected to
medium to large cyclic strains was not rigorously addressed.
Though there is a lack of understanding on the dynamic pro-
perties of cement treated clay, there are a large number of studies
reported on nonlinear, hysteretic and stiffness degrading behaviour of
natural clay subjected to cyclic loading [9–11]. Idriss
et al. [12] introduced the concept of the degradation index to model
the reduction in shear modulus with the progression of loading cycles.
Vucetic and Dobry [13] proposed the design charts to find degradation
shear strain. Zhou and Gong [14] developed the stiffness degradation
model of saturated clay considering cyclic shear strain amplitude,
cyclic stress ratio, overconsolidation ratio and loading frequency. Li
et al. [15] studied the post cyclic strength degradation behaviour of
marine clay under long-term cyclic loading. More recently Subrama-
niam and Banerjee [16] proposed a shear modulus degradation model
for cohesive soil considering plasticity index, overconsolidation ratio,
loading frequency and cyclic shear strain amplitude.
In summary, preceding discussion shows that whilst the modelling
of dynamic behaviour of natural clay is well-established, there are
limited numbers of studies conducted on dynamic properties of
cement treated clay. Moreover most of those studies concentrated
on small strain behaviour of the cement treated clay. Hence there is a
need to study the effect of medium to large cyclic shear strains on the
behaviour of the cement treated clay. The present paper aims to study
the shear modulus degradation behaviour of the cement treated
marine clay. Series of undrained cyclic triaxial test (UCTT) were
performed and effects of various parameters such as, cement content
(c), cyclic shear strain amplitude (γ), curing days (d) and loading
frequency (f) on the stiffness degradation were studied. Based on the
experimental observations, an empirical model was proposed to
compute the degradation index of the cement treated marine clay.

n
Corresponding author. 2. Experimental programme
E-mail addresses: psmani100@gmail.com (P. Subramaniam),
subhadeep@iitm.ac.in (S. Banerjee).
1
Tel.: þ91 44 2257 4304 (off), þ 91 44 2257 6304 (res); Series of strain controlled cyclic triaxial test was conducted on
fax: þ91 44 2257 4252. the cement treated marine clay specimens with varying curing

http://dx.doi.org/10.1016/j.soildyn.2014.06.013
0267-7261/& 2014 Elsevier Ltd. All rights reserved.
182 P. Subramaniam, S. Banerjee / Soil Dynamics and Earthquake Engineering 65 (2014) 181–188

days, cement content, loading frequencies and cyclic shear strain

amplitude. The complete list of different testing conditions is given 1
in Table 1. The low range of cement content was selected to 10000
represent the typical stabilized dredged marine clay [17]. Curing
1. Quartz
days were chosen based on the details available in the literature
8000 2. Kaolinite

Intensity (Counts)
[5,17]. The cement treated specimens (50 mm in diameter and
3. Feldspar
100 mm in height) were isotropically consolidated under the
confining pressure of 100 kPa. As the effect of the confining 6000 4. Illite
pressure on secant shear modulus of cohesive soil is reported to
be negligible [18], all the tests are carried out at the same 4000
confining pressure. The specimens were subjected to undrained 2
3
two-way sinusoidal cyclic loading for 100 cycles.
2000 1
4 1
2.1. Specimen preparation
0
The marine clay samples were collected from Ennore coast of 0 10 20 30 40 50 60 70 80
Tamil Nadu, India. The physical and mechanical properties of the 2θ (in degree)
clay are shown in Table 2.
Fig. 1a shows the X-ray diffraction pattern of the marine clay
used. The plot shows the presence of quartz and feldspar as the silt 6000
fraction. It also indicates kaolinite and illite as the clay fraction.
1
1. Calcium silicate
Commercially available 53 grade ordinary Portland cement was
2. Calcite
used for the present study. The X-ray diffraction pattern of the
3. Alumina
cement (Fig. 1b) shows a clear dominance of calcium silicate.
Intensity (Counts) 4000 4. Silica
The cement slurry was gradually added to the remoulded base clay
5. Fluorite
and mixed in a mixer for 10 min at a speed of 61 rpm. The final water
1
content of the clay–cement mix can be obtained from Eq. (1) [19]. 2
C w ¼ wn þ ðw=cÞAw ð1Þ 1
2000 1
n
where, Cw is the final water content of the clay–cement mix (%), w is
the water content of the base clay, w/c is the water–cement ratio by 1
weight of the cement slurry and Aw is the cement content as the 4 2 5 3

percentage by weight of the dry base clay. In the present study the
water–cement ratio was generally maintained as 0.8. The water 0
content of the base clay was 102%. It should be noted that the effect 0 10 20 30 40 50 60 70 80
of water content is significant on the properties of cement treated soil. 2θ (in degree)
The cement admixed clay was consolidated in the fabricated sampling Fig. 1. (a) Powder XRD patterns of the ( o 75 μm fraction) marine clay used for the
tubes with loading frame as shown in Fig. 2. The sampling tubes, made study, (b) Powder XRD patterns of the ordinary Portland cement used for the study.
of Fe 316 grade stainless steel, are 50 mm in inner diameter and
200 mm in height. The inner surface of the tubes is polished to
minimize the friction between the specimen and the wall of the tube. In order to facilitate the ejection of the specimen, a thin layer of
silicone grease was applied in the inner wall of the tube. The cement
Table 1 admixed clay was filled in the sampling tube in 3 stages. After each
Different conditions studied in cyclic triaxial tests. stage of filling, the cement admixed clay is tamped to remove the
entrapped air. In each batch nine specimens were prepared in order to
d – Curing (days) 7, 14, 28 avoid the quick setting. The entire process of filling in the sampling
f – Loading Frequency (Hz) 0.1, 0.5
tubes was done in less than 10 min. A detachable porous stone is
c – Cement content (%) 2.5, 5, 7.5
γ – Cyclic shear strain amplitude (%) 0.3, 0.7, 1 mounted at the bottom of the sampling tube for bottom drainage.
N – Number of Cycles 100 Porous stones are boiled after each and every single use to ensure the
free drainage. Filter papers are kept at the top and bottom of the
sampling tube to avoid clogging in the drainage path. To allow top
drainage, drainage holes are provided in the top cap. Once the top cap
is placed, water has been poured at a regular interval (every four days)
Table 2 above the top cap to avoid drying of the specimen. For curing,
Physical and mechanical properties of Chennai minimum height of 10 mm water level is maintained above the top
marine clay used for the study.
cap. To represent the shallow depth, the specimens were loaded for
Liquid limit, LL (%) 54 the stress of 30 kPa after 8 h. After the required curing period, the
Plastic limit, PL (%) 30 specimen was ejected from the sampling tube. The tube was kept
Specific gravity, Gs 2.69 upside down in the sampling extruder in order to maintain a unique
Sand (%) (4.75–0.075 mm) 10
loading direction. To ensure the mixing quality and homogeneity of
Silt size (%) (0.075–0.002 mm) 33
Clay size (%) ( o0.002 mm) 57 the specimens, three specimens in each batch were tested for
pH 7.8 unconfined compression test (ASTM D2166) [20] after 7 days of
Compression Index, Cc 0.67 curing period. If the deviation was greater than 75 kPa from the
Permeability kv (mm/sec) 5  10  7 mean value of the compressive strengths of the specimens, new
Unconfined compressive strength 28
(kPa)
batch of specimens were prepared. All the specimens were cured in
the mist room at the temperature of 2773 1C and the humidity of
 P. Subramaniam, S. Banerjee / Soil Dynamics and Earthquake Engineering 65 (2014) 181–188 183

Table 3
Summary of the specimens tested in the present study.
Weights
Type of Cement Curing Water Density γ (%) N f
test content days content (g/cc) (Hz)
(%) (%)

CIU 0 – 105.1 1.338 – – –

CIU 7.5 28 82.3 1.426 – – –
Loading piston UCTT 2.5 7 101.5 1.342 0.3 100 0.1
UCTT 2.5 7 100.1 1.345 0.3 100 0.5
UCTT 2.5 7 99.8 1.351 0.7 100 0.1
UCTT 2.5 7 98.1 1.355 0.7 100 0.5
UCTT 2.5 7 99.3 1.352 1 100 0.1
UCTT 2.5 7 98.2 1.355 1 100 0.5
Loading cap
UCTT 2.5 14 96.2 1.359 0.3 100 0.1
UCTT 2.5 14 95.4 1.365 0.3 100 0.5
UCTT 2.5 14 97.2 1.352 0.7 100 0.1
UCTT 2.5 14 95.9 1.362 0.7 100 0.5
Top drainage UCTT 2.5 14 96.5 1.360 1 100 0.1
UCTT 2.5 14 97.0 1.351 1 100 0.5
UCTT 2.5 28 89.5 1.382 0.3 100 0.1
Stainless steel UCTT 2.5 28 90.3 1.387 0.3 100 0.5
UCTT 2.5 28 91.2 1.382 0.7 100 0.1
Sampling tube
UCTT 2.5 28 90.8 1.389 0.7 100 0.5
200 mm

UCTT 2.5 28 91.8 1.383 1 100 0.1

Cement admixed
UCTT 2.5 28 91.5 1.386 1 100 0.5
clay UCTT 5 7 98.8 1.368 0.3 100 0.1
UCTT 5 7 99.9 1.362 0.3 100 0.5
UCTT 5 7 100.3 1.352 0.7 100 0.1
Porous stone UCTT 5 7 101.8 1.348 0.7 100 0.5
UCTT 5 7 98.2 1.366 1 100 0.1
UCTT 5 7 99.3 1.359 1 100 0.5
50 mm Drainage path UCTT 5 14 95.1 1.375 0.3 100 0.1
UCTT 5 14 94.3 1.389 0.3 100 0.5
UCTT 5 14 98.1 1.371 0.7 100 0.1
UCTT 5 14 96.2 1.382 0.7 100 0.5
UCTT 5 14 96.2 1.384 1 100 0.1
Fig. 2. Schematic diagram of the mould used for the preparation of specimens. UCTT 5 14 95.3 1.385 1 100 0.5
UCTT 5 28 85.6 1.415 0.3 100 0.1
UCTT 5 28 84.1 1.420 0.3 100 0.5
90–99%. The specimen was transferred to the split moulds (50 mm UCTT 5 28 87.1 1.410 0.7 100 0.1
UCTT 5 28 86.3 1.395 0.7 100 0.5
diameter and 100 mm height) and the specimen was trimmed at
UCTT 5 28 87.8 1.397 1 100 0.1
both ends using a wire cutter to obtain the required specimen UCTT 5 28 85.2 1.419 1 100 0.5
height of 100 mm. UCTT 7.5 7 101.9 1.342 0.3 100 0.1
After the necessary curing days, a few standard isotropically UCTT 7.5 7 101.5 1.347 0.3 100 0.5
consolidated undrained (CIU ) tests were conducted on cement treated UCTT 7.5 7 98.3 1.356 0.7 100 0.1
UCTT 7.5 7 96.4 1.359 0.7 100 0.5
marine clay (5% cement content and 7 days of curing) specimens along UCTT 7.5 7 97.8 1.356 1 100 0.1
with the untreated marine clay specimens. The untreated marine clay UCTT 7.5 7 97.2 1.354 1 100 0.5
specimen (50 mm in diameter and 100 mm in height) was prepared UCTT 7.5 14 87.8 1.385 0.3 100 0.1
by a slurry consolidation method in which remoulded clay slurry was UCTT 7.5 14 87.3 1.391 0.3 100 0.5
UCTT 7.5 14 86.3 1.398 0.7 100 0.1
prepared to 1.5 times of liquid limit of the base clay in a mixer. The
UCTT 7.5 14 86.2 1.395 0.7 100 0.5
clay slurry was poured in the sampling tubes in 3 stages. After each UCTT 7.5 14 87.8 1.385 1 100 0.1
stage, the clay slurry is tamped using a tamping rod to remove UCTT 7.5 14 87.3 1.390 1 100 0.5
entrapped air. The consolidation pressure of 30 kPa was applied in UCTT 7.5 28 80.4 1.440 0.3 100 0.1
order to attain some initial shear strength. After 30 days of consolida- UCTT 7.5 28 80.2 1.443 0.3 100 0.5
UCTT 7.5 28 79.8 1.435 0.7 100 0.1
tion, the untreated specimens were ejected from the sampling tubes UCTT 7.5 28 79.0 1.431 0.7 100 0.5
and trimmed at both ends in order to get a specimen height of UCTT 7.5 28 81.2 1.456 1 100 0.1
100 mm. Summary of the specimens tested in the present experi- UCTT 7.5 28 80.8 1.451 1 100 0.5
mental study are listed in Table 3. Isotropically consolidated undrained
Confining pressure¼100 kPa, curing stress¼30 kPa.
tests (CIU ) were conducted on untreated and cement treated speci-
mens (5%c, 7d) as per ASTM D4767 -11 [21]. The specimens were
checked for Skempton's pore pressure parameter ‘B’ to ensure the
saturation prior to consolidation in all the static and cyclic tests. The strain (ε) are shown in Fig. 3a and b. Fig. 3c shows the stress paths for
untreated and cement treated specimens (5%c, 7d) were saturated at untreated and cement treated clay (5%c, 7d) in p0 – q plane where p0 ¼
(σ1 þ2σ3)/3 and q¼(σ1  σ3). As expected, Fig. 3a shows that the
0 0 0 0
the back pressure of 240 kPa. For treated specimens with higher
cement content and longer curing period (7.5%c, 28d), back pressure of natural marine clay exhibits a strain softening behaviour. However
340 kPa was required at the end of saturation procedure. After the unlike the natural clay, figure also shows that the cement treated
saturation, the specimens were consolidated at a mean effective specimens present a peak followed by strain softening. The type of
confining pressure of 100 kPa followed by undrained monotonic failure was often termed as brittle–ductile translational mode of failure
shearing. The shearing rate was obtained as 0.024 mm/min based on [23]. As can be seen from Fig. 3b, the pore pressure increases with
the consolidation results as per the procedure given by Head [22]. The strain until the failure and becomes constant in the subsequent
variations of deviatoric stress (q) and pore pressure (Δu) with shear softening regime.
184 P. Subramaniam, S. Banerjee / Soil Dynamics and Earthquake Engineering 65 (2014) 181–188

80 160

q (kPa)
70 120
60
80
50
40
q (kPa)

40
Untreated marine clay 0
30 1.2 -0.8 -0.4 0.0 0.4 0.8 1.2
Treated clay (5%c, 7d)
-40 γ (%)
20
-80 1st cycle
10
10th cycle
0 -120 100th cycle
0 2 4 6 8 10 12
ε (%) -160

0.16
80

70
0.14
60

50 0.12
u (kPa)

40 t
0.10
30
Untreated marine clay
20 Treated clay (5% c, 7d)
0.08 Cement treated marine clay
Natural chennai marine clay
10
0.06
0
0.2 0.4 0.6 0.8 1.0 1.2
0 2 4 6 8 10 12
ε (%)
γ (%)
Fig. 4. (a) Stress–strain plot of cement treated marine clay (5%c, 28d, 0.5 Hz f),
(b) degradation parameter vs. cyclic shear strain amplitude for untreated and
80
cement treated marine clay (5%c, 7d, 0.1 Hz f).

60
Fig. 4a shows that the secant shear modulus for 1st, 10th and
100th cycle was respectively 12.5 MPa, 11.5 MPa and 9.25 MPa.
As the figure suggests the gradual reductions in the secant shear
q (kPa)

40 modulus is an indication of cyclic shear modulus degradation.

Untreated marine clay From Fig. 4a the degradation index can be calculated based on
Treated clay (5% c, 7d) Idriss's [12] formulation as follows:
20
GsecN
δ¼ ¼ Nt ð2Þ
Gsec1
0
where δ is the degradation index, GsecN and Gsec1 are the secant
20 40 60 80 100 120
p' (kPa) shear modulus at Nth cycle and 1st cycle respectively. N is the
Fig. 3. (a) Consolidated undrained tests: Variation of deviatoric stress (q) with number of loading cycles and (t) is the degradation parameter.
shear strain (ε), (b) consolidated undrained tests: variation of pore water pressure Eq. (2) can be rearranged to calculate the degradation para-
(u) with shear strain (ε), (c) consolidated undrained tests: tress path for untreated meter as shown in Eq. (3)
and treated (5%c, 7d) marine clay in p0 – q plane.

log ðGsecN =Gsec1 Þ

t¼ ð3Þ
log N
3. Results and discussion

3.1. Shear modulus degradation
Fig. 4a shows typical stress–strain loops of cement treated
marine clay plotted for 1st, 10th and 100th cycle. For this
particular test, the specimen with 5% cement content and 28 days
of curing period subjected to cyclic shear strain amplitude of 1%
and loading frequency of 0.5 Hz.
The degradation parameter (t) represents the rate of degradation.
The stiffness degradation decreases with an increase in parameter
(t). Fig. 4b plots the degradation parameter (t) against cyclic shear
strain amplitude for untreated and cement treated marine clay.
The figure shows that the cyclic degradation in cement treated
marine clay, particularly for large strains, is significantly lower
than that of the untreated marine clay.
 P. Subramaniam, S. Banerjee / Soil Dynamics and Earthquake Engineering 65 (2014) 181–188 185

3.2.1. Effects of cement content and curing days

In the present study, the tests are conducted on the clay treated
with low cement content (maximum 7.5%). Fig. 5a shows the
variation of degradation parameter (t) with cement content (c)
0.125 and cyclic shear strain amplitude (γ) for 28 days of curing and
0.1 Hz loading frequency. This figure shows that the degradation
parameter decreases with increase in the cement content. For 0.3%
0.100 cyclic shear strain, with an increase in cement content from 2.5% to
t
7.5% the degradation parameter is reduced by 92%. Further it is also
0.075 noted that the effect of shear strain amplitude on degradation
parameter is predominant at a lower cement content. As can be
1.0 seen, for 2.5% cement content with an increase in cyclic shear strain
0.050 0.8 amplitude from 0.3% to 1%, the increase in degradation parameter is
0.6 %)
2
4 0.4 γ( about 27%. Whereas, for 7.5% cement content the degradation
6 8 parameter is increased by 20% for the same strain increment.
c (%)
Fig. 5b shows the effect of curing days along with the cement
content on degradation parameter at 1% cyclic shear strain and
0.1 Hz loading frequency. It can be seen from the figure that the
degradation parameter, in general, decreases with increase in
curing days. However for lower cement content, degradation
0.14
parameter may not significantly vary with the curing days. For
2.5% cement content, t decreases about 7.7% between 7th and
0.12 14th days of curing and 6% between 14thand 28th days of
curing, respectively. On the other hand for 7.5% cement content,
t t decreases about 22.4% between 7th and 14th day of curing and
0.10 6.5% between 14th and 28th day of curing, respectively. This
may be due to the reduction in pore size with respect to curing
days [1]. Fig. 5c shows the effect of curing days and cyclic shear
0.08
strain amplitude on degradation parameter for 7.5% cement
content and 0.1 Hz loading frequency. The figure also shows
0.06 28 that in all curing days the degradation parameter increases with
2 21 an increase in the shear strain amplitude. Similar trend can be
4 14 seen for natural soil with wide range of plasticity and over-
d consolidation ratio [24].
c (%) 6 7

8 0
3.2.2. Effect of loading frequency
The shear modulus degradation of clay is strongly influenced by
the loading frequency [12]. Fig. 6a shows the effect of loading
frequency, together with cyclic shear strain amplitude on degrada-
tion parameter. The figure shows that the degradation parameter t
0.09
decreases with increase in loading frequency. Similar observation
was reported for natural cohesive soil with a wide range of plasticity
0.08 index [16]. Moreover it can be seen that the effect of loading
t 0.07 frequency on degradation parameter is dominant at higher cyclic
shear strains. For 0.3% shear strain amplitude, t decreases about 17%
between 0.1 Hz and 0.5 Hz of loading frequency. On the other hand, t
0.06 decreases about 25% between 0.1 Hz and 0.5 Hz for the cyclic shear
1.0 strain amplitude of 1%. It is also observed that the degradation
0.05 0.8 parameter increases with increase in the shear strain amplitude.
0.6 )
5
(%
10 15 0.4 During cyclic loading, stiffness degrades due to the generation of
20
d
25 30 0.2 γ pore water pressure. This pore pressure generation depends on the
type of soil and loading conditions [25]. The effect of loading frequency
Fig. 5. (a) Experimental results of the effects of cement content and cyclic shear and cyclic shear strain amplitude on the development of pore pressure
strain amplitude on degradation parameter ‘t’ (28d, 0.1 Hz f); (b) experimental
during cyclic loading can be explained using Fig. 6b. The figure shows
results of the effects of cement content and curing days on degradation parameter
‘t’ (0.1 Hz f, 0.7%γ), (c) Experimental results of the effects of curing days and cyclic
that the number of loading cycles required to stabilize excess pore
shear strain amplitude on degradation parameter ‘t’ (7.5% c, 0.1 Hz f). pressure is generally less for low shear strain amplitudes [26,27].
Furthermore figure also shows that for all three different cyclic shear
3.2. Factors affecting degradation parameter (t) strain amplitudes, compared to 0.5 Hz loading frequency, the devel-
opment of excess pore pressure (Δu) is higher for 0.1 Hz loading
As mentioned earlier, degradation parameter of natural clay frequency. However the effect of cyclic shear strain amplitude is more
depends on various parameters such as plasticity index, overconsoli- significant than the loading frequency. Fig. 6c and d shows the effect of
dation ratio, loading frequency and cyclic shear strain [13,14,16]. frequency along with the cement content and curing days on
Besides the above mentioned parameters, following section discusses degradation parameter respectively. It is noted that the effect of
the additional factors that affect the degradation parameter of cement frequency on degradation parameter decreases with increase in
treated marine clay. cement content and curing days.
186 P. Subramaniam, S. Banerjee / Soil Dynamics and Earthquake Engineering 65 (2014) 181–188

0.065
0.125
0.060
t 0.100
0.055 t
0.075
0.050 1.0
0.8 0.55
0.6
) 0.050 0.40
0.10
(%
0.25 0.4 0.25 z)
0.40 0.2 γ 2 4 H
f (Hz)
0.55 6 8
0.10 f(
c (%)

60

50

40 0.09
Δu (kPa)

30 0.08
t
0.3% γ, 0.1Hz f 0.07
20 0.3% γ, 0.5Hz f
0.7% γ, 0.1Hz f 0.55
10 0.7% γ, 0.5Hz f 0.06 0.40
1% γ, 0.1Hz f 0.25 z)
7 15 H
1% γ, 0.5Hz f 20 25 0.10 f(
0
0 20 40 60 80 100 120
d
N
Fig. 6. (a) Experimental results of the effects of loading frequency and cyclic shear strain amplitude on degradation parameter ‘t’ (7.5%c, 28d), (b) pore pressure generation (u)
vs. number of loading cycles (N) of cement treated marine clay specimen (5%c, 28d), (c) experimental results for the effects of cement content and loading frequency on
degradation parameter (t) (28d, 1%γ), (d) experimental results for the effects of curing days and frequency on degradation parameter (t) (7.5%c, 1%γ).

4. Degradation model
0.14

In this section, based on Idriss's [12] concept of degradation

index, a shear modulus degradation model for cement treated clay 0.12
is proposed. As shown earlier, the degradation parameter (t)
of cement treated clay depends on cement content (c), curing
days (d), cyclic shear strain amplitude (γ) and loading frequency
Predicted 't'

0.10
(f). Moreover it is also noted that these parameters are interde-
pendent. According to Myers et al. [28], the response surface
model can be constructed as 0.08
4 4 3 4
Y K ¼ β k0 þ ∑ βki X i þ ∑ βkii X 2i þ ∑ ∑ β kij X i X j ð4Þ
i¼1 i¼1 i ¼ 1 j ¼ iþ1 0.06
where Yk is the dependent variable or measured response, βk0, βkii, βkij
are the regression coefficients and Xi, Xj are independent variables
0.04
such as cement content (%), curing days, cyclic shear strain amplitude 0.04 0.06 0.08 0.10 0.12 0.14
(%) and loading frequency (Hz). The response surface design was
Experimental ' t '
created using central composite design and response surface regres-
sion analysis was performed using MINITAB 14s software [29]. From Fig. 7. Comparison of degradation parameter (t) obtained from tests with that
the analysis, regression coefficients were obtained and a second order predicted from Eq. (5).

polynomial (as shown in Eq. (5)) was proposed. The working range of
Eq. (5) is mentioned through Eq. (6).
t ¼ 0:154058 cð0:022863  0:001915c þ 0:005231γ %
8 9
þ 0:000138d  0:006413f Þ > 2:5 r cð%Þ r 7:5 >
>
> >
 0:065635f  dð0:001285  0:000083f < 0:1 r f ðHzÞ r 0:5 >
=
t¼F ð6Þ
þ 0:001130γ %  0:000051dÞ >
>
>
0:3 r γ ð%Þ r 1 >
>
>
: ;
þ γ %ð0:047614 þ 0:023977γ %  0:012841f Þ ð5Þ 7 r d r 28
 P. Subramaniam, S. Banerjee / Soil Dynamics and Earthquake Engineering 65 (2014) 181–188 187

data points form a relatively tight cluster around the line of unity
1.00
gradient, indicating that the cyclic triaxial test results generally
0.95 agree well with the predictions.
Fig. 8a plots the degradation index computed from Eqs. (2) and (5)
0.90 along with the experimental results for different curing days with 7.5%
0.85 cement content, 0.5 Hz loading frequency and 0.7% cyclic shear strain
amplitude. Figure shows that the proposed degradation model
0.80 matches favourably with the experimental results.
Similarly Fig. 8b and c shows the degradation index obtained
δ

7d - Experiment
0.75
7d - model from the proposed formulation along with the test results for
0.70 14d - Experiment varying loading frequencies and cement content respectively. The
14d - model respective tests conditions are mentioned in the figures. The
0.65 figures show that the proposed model compares satisfactorily
28d- Experiment
28d - model with the tests results.
0.60

0.55
1 10 100 5. Conclusion
N
The present paper aimed to study the shear modulus degrada-
tion of cement treated marine clay subjected to cyclic loading.
1.00 There are several notable conclusions that can be obtained from
0.95 the present study. Like ordinary clays, cement treated clays
experiences shear modulus degradation during cyclic loadings.
0.90
From the test results, it is evident that the degradation of shear
0.85 modulus of cement treated clay depends on various factors such
as, number of loading cycles, cement content, curing days, cyclic
0.80
shear strain amplitude and loading frequency. The shear modulus
0.75 degradation decreases with increase in cement content and curing
δ

0.70 days. Moreover, at lower loading frequency the shear modulus

degradation of cement treated clay is found to be higher. It is also
0.65 0.1Hz, 28days-Experimental
noted that the factors affecting degradation parameter of cement
0.5Hz, 28days-Experimental treated marine clay are interdependent. For example, the effect of
0.60
0.1Hz, 28days-model frequency on degradation parameter depends on the applied cyclic
0.55 0.5Hz, 28days-model shear strain amplitude. An empirical model is proposed based on
0.50 cement content (2.5–7.5%), curing days (7–28), and loading fre-
1 10 100 quencies (0.1–0.5 Hz). The proposed empirical model is able to
N match the shear modulus degradation obtained from cyclic triaxial
tests. However, one should note that the proposed model is
obtained from the cyclic tests carried out on Chennai marine clay
1.0
with the specified curing conditions and mix ratios. The perfor-
mance of the model is not validated for soils with very high or low
PI, higher curing stress, longer curing period or other mix ratios.
0.9

References
0.8
[1] Chew SH, Kamruzzaman AHM, Lee FH. Physicochemical and engineering
δ

behavior of cement treated clays. J Geotechn Geoenviron Eng 2004;130

2.5% c - Experiment (7):696–706.
0.7 [2] Xiao H. Yielding and Failure of Cement Treated Soil. Department of Civil
5% c - Experiment
Engineering National University of Singapore; 2009.
7.5% c - Experiment [3] Bushra I. Compressibility and Strength Characteristics of marine soil treated
2.5% c - Model with cement and fly ash. Department of Civil Engineering Indian Institute of
0.6
Technology Madras; 2011.
5% c - Model [4] Flores VRD, Emido DG. Hardening of kaolin clay mixed with different binders.
7.5% c - Model In: Proceedings of International Conference on Deformation Characteristics of
0.5 Geomaterials, vol. 2. Seoul; 2011. p. 669–76.
[5] Bahador M, Pak A. Small-strain shear modulus of cement-admixed kaolinite.
1 10 100
Geotech Geol Eng 2012;30:163–71.
N [6] Yang L, Woods RD. Modeling of dynamic properties of cemented clay. In:
Proceedings of Pan-American Geotechnical Conference; 2011.
Fig. 8. (a) Comparison of degradation index obtained from test results with that [7] Kim A, Chang I, Cho GC, Kang Y. Strain-dependent dynamic characteristics of
computed from Eqs. (5) and (2) for different curing days (7.5%c, 0.5 Hz f, 0.7%γ), cement treated clay. In: Proceedings of International Symposium on Deforma-
(b) comparison of degradation index obtained from test results with that computed tion Characteristics of Geomaterials. Seoul; 2011.p. 779–84.
from Eqs. (5) and (2) for different loading frequency and curing days (5%c, 0.7%γ), [8] Tsai PH, Ni SH. Effects of types of additives on dynamic properties of cement
(c) comparison of degradation index obtained from test results with that computed stabilized soils. Int J Appl Sci Eng 2012;10(2):131–44.
from Eqs. (5) and (2) for different cement contents (0.7%γ, 0.1 Hz f, 28d). [9] Hardin BO, Drnevich VP. Shear modulus and damping in soils: design
equations and curves. J Soil Mech Found Eng Div (ASCE) 1972;98:667–92.
[10] Vucetic M, Dobry R. Effect of soil plasticity on cyclic response. J Geotech Eng
Fig. 7 summarizes a comparison of the degradation parameter 1991;117(1):89–107.
[11] Puzrin A, Frydman S, Talesnick M. Normalising degrading behaviour of soft
measured from the different cyclic triaxial tests, along with the clay under cyclic simple shear loading. J Geotech Eng Div (ASCE) 1995;121
corresponding predicted values from Eq. (5). As can be seen, the (12):836–43.
188 P. Subramaniam, S. Banerjee / Soil Dynamics and Earthquake Engineering 65 (2014) 181–188
[12] Idriss IM, Dobry R, Singh RD. Nonlinear behaviour of soft clays during cyclic
loading. J Geotech Eng Div (ASCE) 1978;102(12):1427–47.
[13] Vucetic M, Dobry R. Degradation of marine clays under cyclic loading.
J Geotech Eng Div (ASCE) 1988;114(2):133–49.
[14] Zhou J, Gong X. Strain degradation of saturated clay under cyclic loading. Can
Geotech J 2001;38:208–12.
[15] Li LL, Dan HB, Wang LZ. Undrained behavior of natural marine clay under
cyclic loading. Ocean Eng 2011;38:1792–805.
[16] Subramaniam P, Banerjee S. Shear modulus degradation model for cohesive
soil. Soil Dyn Earthq Eng 2013;53:210–6.
[17] Zhang RJ, Santoso AM, Tan TS, Phoon KK. Strength of high water content
marine clay stabilized by low amount of cement. J Geotech Geoenviron Eng
2013;139(12):2170–81.
[18] Subramaniam P, Banerjee S. A correction to damping ratio for hyperbolic
hysteretic model for clayey soil. Int J Geotech Eng 2013;7(2):124–9.
[19] Lorenzo GA, Bergado DT. Fundamental parameters of cement admixed clay –
new approach. J Geotech Geoenviron Eng 2004;130(10):1042–50.
[20] ASTM D2166. Standard test method for unconfined compressive strength of
cohesive soil. American Society of Testing and Materials; 2006. p. 1–6.
[21] ASTM D4767. Standard test method for consolidated undrained triaxial
compression test for cohesive soils. American Society of Testing and Materials;
2011. p. 1–14.
[22] Head KH. Manual of Soil Laboratory Testing: Volume 3 Effective stress tests.
Chichester: Wiley Publishers; 1998.
[23] Panda AP, NarasimhaRao S. Undrained strength characteristics of artificially
cemented marine clay. Mar Georesour Geotechnol 1998;16:335–53.
[24] Ishihara K. Soil Behavior in Earthquake Geotechnics. Oxford: Clarendon Press;
1996.
[25] Matasovic N, Vucetic M. Generalized cyclic–degradation–pore–pressure gen-
eration model for clays. J Geotech Eng 1995;121(1):33–42.
[26] Vucetic M, Dobry R. 1986.
[27] Vucetic M. Normalized behavior of offshore clay under uniform cyclic loading.
Can Geotech J 1988;25:33–41.
[28] Myers RH, Montgomery DC, Cook CMA. Response Surface Methodology:
Process and Product Optimization Using Designed Experiments. New Jersey:
John Wiley and Sons; 2009.
[29] Minitab 14s. Meet Minitab: User Manual, Minitab Inc.; 2004.