Page 1
International Journal of Civil & Environmental Engineering IJCEE-IJENS Vol: 11 No: 04 34
119104-5858 IJCEE-IJENS © August 2011 IJENS I J E N S
Experimental Investigation on Curing Time and
Stress Dependency of Strength and Deformation
Characteristics of Cement-treated Sand and it's
Degradation Phenomena Abu Taher Md. Zillur Rabbi
1, Md. Kamal Hossain
2, Jiro Kuwano
3, Wee Boon Tay
3
Abstract— Soil stabilization with cement is a good solution for
the construction of subgrades for roadway and railway lines,
especially in transition zones between embankments and rigid
structures, where the mechanical properties of supporting soil
are very influential. In order to optimize the design of cement-
mixed soil structures, their behaviors need to be well understood;
especially the strength and deformation characteristics at very
small strains are of great importance. Similar to concrete
material, the strength of cement-mixed soil continues to increase
with time, thereby improving its strength and deformation
properties with time. On the other hand, in the field cementation
bonds in cement-mixed soils are formed under stress in case of
in-situ soil. However, in the past researches the cementation
bonds under stress was considered only by a few researchers.
This is an underestimation of the stress-strain behavior of
cement-mixed soil. This study investigates the influence of long
curing period (e.g. up to 180 days) and the stress condition
during curing stage on the strength and deformation
characteristics of cement-mixed sand. A series of consolidated
drained (CD) triaxial compression (TC) tests were performed
along with the small strain cyclic loading and bender element
tests at intervals during monotonic loading to determine the
elastic Young's modulus (Ev) at extremely small strain range and
shear moduli (Ghh, Ghv and Gvh) respectively. The test results
show that the curing stress and curing period both have a
significant influence in the peak strength, stiffness, Ev, and Gvh
value. Curing period also influences the value of shear moduli in
the two horizontal directions Ghh and Ghv. However, the influence
of curing stress on the Ghh and Ghv is not very clear. The
degradation phenomena of cementation bond were discussed
according to the test results obtained from the cyclic loading and
bender element test during shearing of cement-mixed sand.
This work was done in the Geosphere Research Institute of Saitama
University (GRIS) in order to update the Triaxial testing system and also as a part of first author’s Master’s by research. The financial support for his
Master’s Degree and therefore this research is from the Asian Development
Bank – Japan Scholarship Program (ADB-JSP) which is greatly acknowledged.
1 Abu Taher Md. Zillur Rabbi is an Assistant Professor of Department of Civil
Engg., Dhaka University of Engg. & Technology, Gazipur-1700, Bangladesh. (Corresponding author, Phone: +880-1712-526634, e-mail:
[email protected] ) 2 Md. Kamal Hossain is a Professor of Department of Civil Engg., Dhaka University of Engg. & Technology, Gazipur-1700, Bangladesh. ( e-mail:
[email protected] ) 3 Jiro Kuwano is a Professor of Geosphere Research Institute of Saitama University (GRIS), 255, Shimo-okubo, Sakura-ku, Saitama-shi, Saitama
University, Saitama 338-8570, Japan. 4 Wee Boon Tay is a Government employee of Singapore (formerly graduate
student of Tokyo Institute of Technology)
Index Term— cementation bond, triaxial test, buoyancy,
curing overburden stress, bender element, shear wave velocity,
Young’s modulus, shear moduli, phase transformation.
I. INTRODUCTION
Ground improvement by cement treatment has been widely
applied for structural foundations, excavation control,
reinforced soil wall construction, bridge embankments,
highway embankments and liquefaction mitigation. One of the
new cost effective methods, to construct important permanent
soil structures that allow only a limited amount of deformation
such as bridge abutment etc. is the use of compacted cement-
mixed soil as the backfill. After successful construction of the
first new type bridge abutment having the backfill of well-
compacted cement-mixed gravelly soil for a bullet train line
(Shinkansen) in 2003 at Kyushu, Japan [1]-[4], the use of
cement-mixed soil is gaining more acceptance throughout the
different parts of the world and it is widely used in several
ground improvement projects such as highway and railway
embankments. An example of the use of cement-mixed
geogrid reinforced embankments on both sides of a highway
flyover bridge in Utsunomya, Japan and the schematic
illustration of the reinforced soil embankment is shown in Fig.
1 [5]. The use of cement-mixed soil is found to be relatively
simple and economical compared to deep piling and the use of
reinforced concrete structures since less concrete is required.
Reliable evaluation of strength and deformation
characteristics of compacted cement-mixed soil is one of the
essential factors in order to design effectively and confidently
design such soil structures. Though there have been many
studies on cement-mixed soil using different types of soils [6]-
[10], the behavior has yet to be generalized. The effect of
curing conditions in terms of curing time and stress conditions
during curing are still poorly understood because sufficient
evaluation of these effects is extremely time-consuming and
so very difficult [5], [6], [9]. Moreover, with the use of
different testing techniques, there are discrepancies in the test
results.
Hydration of cement in cement-mixed soil continues over a
very long period [9] which therefore gives more resistance to
shearing. On the other hand, the cementation bonds in in situ
soil are formed under stress. However, it was found in the
literature that cementation bonds under stress has been
considered by only very few researchers. This leads to an
underestimation of the stress-strain-strength behavior of
cement-treated soil [6], [11].
Page 2
International Journal of Civil & Environmental Engineering IJCEE-IJENS Vol: 11 No: 04 35
119104-5858 IJCEE-IJENS © August 2011 IJENS I J E N S
II. OBJECTIVE
The objective of this study is to investigate the influence of
curing period (up to 180 days) and applied stress during
curing stage on the strength and deformation characteristics of
cement-mixed sand. To investigate the change of dynamic
shear modulus in the vertical direction (Gvh) and void ratio (e)
with curing period and application of curing stress is also
another objective of this study. The composition of cement-
mixed sand used in this study is the same as those used by
Kuwano [12] and Rabbi et al., [5].
III. MATERIAL AND METHODOLOGY
A. Specimen Preparation
In this study, Toyoura sand is mixed with high-early
strength Portland cement to improve its mechanical
properties.The amount of high-early-strength Portland cement
used is 60 kg/m3 of Toyoura sand, to achieve an unconfined
compressive strength of 500 kPa after 7 days of curing which
is typical in a method used for the highway embankment
introduced in Fig. 1. The component ratio is calculated such
that the wet density of cement-mixed sand ρt =1.6g/cm3, which
is in accordance with the soil characteristics of that used for
strengthening embankments using cement-mixed sand and
geogrids by Itoh et al., [13]. Wet density is the moist density
of the specimen just after preparing the specimen before
curing. It is the ratio of the total mass of cement, sand and
water to the total volume of the specimen. The properties and
the ratios of the raw materials used in cement-mixed sand are
shown in Table I. The composition of cement-mixed sand
used in this study is the same as that used in the study of
cement-mixed sand by Rabbi et al., [5]. As noted, the amount
of cement used is only 4.13% of the total weight of sand. This
is small compared to normal cement mortar (C/S=50%).
Moreover, the water-cement ratio used is about 242%, which
is higher compared to W/C<100% for cement mortar. This
amount of water was used to spread out the small amount of
cement and ensure that the hydration of cement occurs
throughout the specimen.
Specimens were cured for 4 different curing periods of 7,
28, 90 and 180 days in order to investigate the curing period
dependency of the stress-strain characteristics of cement-
mixed sand. For each curing period specimens were cured
under 2 different curing overburden stresses, σv of 0 kPa and
98 kPa, to investigate the stress dependency during the
formation of cementation bonds in curing stage on the
mechanical properties of cement-mixed sand. To compare the
test results of cement-treated sand with clean Toyoura sand,
one specimen prepared with untreated clean Toyoura sand was
tested in triaxial compression testing machine. The specimen
with clean Toyoura sand was prepared by pouring sand from a
funnel with a constant falling height to control uniform
density all through the specimen height. Density of the
specimen prepared with clean sand is 1.54 g/cm3 (Mg/m
3). All
the test cases are shown in Table II.
TABLE II
TEST CASES TO STUDY THE INFLUENCE OF
CURING TIME AND CURING STRESS.
Specimen
type
Curing
Overburden
stress
Curing
time
(days)
Effective
confining
stress (kPa)
Cement-
mixed sand
0 kPa
7 98
28 98
90 98
180 98
98 kPa
7 98
28 98
90 98
180 98
Clean sand - - 98
TABLE I PROPERTIES AND RATIO OF RAW MATERIAL
Specific gravity, Gs
Toyoura sand (S) 2.645
High-early-strength Portland
cement (C)
3.130
Water (W) 1.000
S C W C/S W/C
Ratio (%) 87.62 3.62 8.76 4.13 242
(b)
Fig. 1. Example of use of cement-mixed sand as reinforced soil wall (a) highway flyover bridge using reinforced soil wall in Utsunomya, Japan, (b)
Cross-section used for the reinforced soil wall in both side embankment.
(a)
Page 3
International Journal of Civil & Environmental Engineering IJCEE-IJENS Vol: 11 No: 04 36
119104-5858 IJCEE-IJENS © August 2011 IJENS I J E N S
Specimen was prepared using a method similar to that used
by Kuwano [12] and Rabbi [5]. Sand and cement were mixed
thoroughly together in a dry state. After that water was added
to the mixture and they were mixed thoroughly again. The
mixture was then compacted into moulds of height 170mm
and diameter 77mm to control the wet density of the specimen
1.6 g/cm3 (Mg/m
3). Compaction was done in 5 stages of about
the same amount of cement-mixed sand each time in order to
make uniform density although the specimen height. The
specimens were then wrapped with plastic wrapping sheet and
stored under constant temperature of 20ºC and humidity of
50% for the specified number of days before they were used
for experimental purposes. Specimens to be cured under stress
were set in a specially made consolidation apparatus as shown
in Fig. 2. The specimens were then loaded with the desired
amount of overburden stress immediately after the moulds
were filled. The loading was applied using air pressure in the
bellofram cylinder mounted on that special apparatus. The
whole process of preparation of specimen and the application
of curing stress is finished within 25 to 30 minute from the
time when water was first added to the sand and cement
mixture in order to avoid the disturbance of cementation
bonds beyond the setting time of cement. The change in shear
modulus in the vertical direction (Gvh) during the curing stage
for both the specimens cured without and under stress was
monitored using a pair of bender elements attached at the top
and bottom ends of the specially made consolidation apparatus
as shown in Fig. 2.
B. Triaxial Testing System
The triaxial testing machine used in this study has an
automatically control and measurement system. The triaxial
compression test system consists of 2 main parts, triaxial test
and bender element test. In triaxial test, output voltages from
all measuring sensors are converted into digital signals which
were recorded on the PC through a 16-bit AD converter. In
turn, the control signals from the PC are converted into
voltages for each control sensor, through a 12-bit DA
converter as shown schematically in Fig. 3 [14]. The vertical
strain was measured both externally using an outer LVDT and
locally using a pair of LDT and a pair of inner LVDT as
shown in Fig. 3. The Young’s modulus was determined from
small strain cyclic loading from the measurement of the LDT
as the LDT has a lower electrical noise level. Since, however,
the measuring range of the LDT is only 2.5%, the
measurements of inner LVDT are used as a supplement to the
LDT beyond the range of 2.5%. The measuring range of the
inner LVDT was 15%.
Introduced by Shirley and Hampton [15], bender elements
are currently a standard technique for deriving the elastic
Fig. 2. Consolidation mould for applying curing stress with BE monitoring.
Fig. 3. Schematic illustration of triaxial testing system (not in scale). (Chowaudhary et al., 2004)
Page 4
International Journal of Civil & Environmental Engineering IJCEE-IJENS Vol: 11 No: 04 37
119104-5858 IJCEE-IJENS © August 2011 IJENS I J E N S
shear modulus of soil at very small strains. In bender element
tests, the maximum shear strain was estimated by Dyvik and
Madshus [16] to be less than 10-5
so that the shear modulus G
determined is relevant to very small strains [17]. Bender
element systems can be set up in most laboratory apparatus,
but are particularly versatile when used in triaxial test as
described by Dyvik and Madshus [16]. Shear wave velocity is
calculated from the time taken for the shear wave to travel
from transmitter to receiver bender element and the distance
between the tip of the transmitter and receiver bender element
[19]. In the bender element test, a function generator sends out
a sine pulse wave, as proposed by Viggani & Atkinson [18], to
the transmitter at one end of the specimen. The receiver at the
other end of the specimen receives this wave. Both wave
patterns at the transmitter and receiver ends are outputted into
an oscilloscope screen. The time taken for sine pulse wave to
pass through the specimen is the time lapse between the start
of transmitting wave and the start of the receiving wave
pattern, as shown in Fig. 4.
In this experiment, 3 pairs of bender elements are used to
measure the shear wave velocity in 3 directions. Fig. 5 shows
the arrangement of bender elements used in this experiment.
The transmitting and receiving ends are denoted by (T) and (R)
respectively. The shear wave is denoted by Sij, where i refers
to the direction of propagation of shear wave and j refers to
the direction of motion of soil particles. Here, a pair of vertical
and two pairs of lateral bender elements were used to measure
the shear waves Svh, Shh, and Shv respectively. The distance
traveled by the shear wave is taken as the distance between the
tips of the transmitter and receiver bender elements, proposed
by Dyvik & Madshus [16].
The shear modulus is then calculated using the following
equations: 2
vhvh VG (1)
2
hvhv VG (2)
2
hhhh VG (3)
where, ρ: Density of specimen
Vvh, Vhv and Vhh: Shear wave velocity in the corresponding
direction of wave propagation.
Volume Change Mesurement using Digital Balance:
The volume change of the specimen during isotropic
consolidation and during monotonic loading was measured
using a digital balance with an accuracy of 0.0001g of weight
instead of the conventional double burette volume change
apparatus and a differential pressure transducer (DPT). The
main part of the digital balance is placed in a glass chamber
and it was connected to the body of the digital balance placed
outside of the glass chamber. A plastic pot with water was
placed over the main part of the digital balance and it was
connected to the specimen through a pipeline as shown in the
schematic triaxial test system in Fig. 3. The volume change of
the saturated specimen was measured by measuring the weight
of water goes in from plastic pot to the specimen or came out
from specimen to the plastic pot. The change of the weight of
water in the pot is a measure of the volume change of the
specimen during isotropic consolidation & shearing. A larger
schematic representation of the whole arrangement of the
digital balance system is shown in Fig. 6.
Effect of Buoyancy Force on Volume Change Measurement:
Fig. 4. Measurement of time taken by shear wave to pass through the
specimen.
Input wave
Output wave
Time taken
Shh(T)
Svh(T)
Shh(R)
Svh(R)
Shv(T)
Shv(R)
Fig. 5. Arrangement of Bender elements
Sensor of Digital
balance
Back
Pressure
Connected to
pressure gauge
Pot & Water
Body of the
digital balance
Oil
Pipeline
connected to the
specimen.
Fig. 6. Schematic illustration of digital balance setup for volume
change measurement.
Page 5
International Journal of Civil & Environmental Engineering IJCEE-IJENS Vol: 11 No: 04 38
119104-5858 IJCEE-IJENS © August 2011 IJENS I J E N S
During the triaxial test a backpressure of 200kPa to 300kPa
was applied to the specimen. Since the specimen was
connected to the water in the plastic pot placed over the digital
balance in the glass chamber, the chamber pressure was also
increased equal to the backpressure in order to avoid water
movement to or from the specimen due to pressure difference.
However, when the pressure in the glass chamber is greater
than the atmospheric pressure, it affects the air density i.e., the
air density increases as well as there introduces an extra
buoyancy force. This extra buoyancy force affects the
readings of the digital balance i.e., digital balance gives the
weight measurement less than the actual weight. In order to
check the effect the digital balance reading was recorded by
increasing the chamber pressure from 0kPa to 300kPa and the
results are shown in Fig. 7. The result shows that the weight of
the plastic pot with water with an initial weight of 155.5g is
reduced to 151.75g when pressure increased from 0kPa to
300kPa i.e., around 2.5% weight reduced. Therefore, we need
to consider the error in digital balance reading due to
buoyancy force and need a correction to be made to the digital
balance reading. Since the buoyancy force has an effect on the
self weight of the main part of the digital balance, readings are
also taken from the digital balance when no weight is over the
balance by increasing the pressure form 0 to 300 kPa. By
subtracting these two values we can get the weight which
containing the actual error due to the buoyancy forces. The
theoretical buoyancy force over an object given by the air
pressure can be calculated by using the equation:
gVFbuoyancy (4)
where, V is the volume of metal piece in cm3 on which
buoyancy force is acting, is the mass density of the air in
g/cm3, g is the acceleration due to gravity and Fbuoyancy is the
buoyancy force in Newton. And its mass equivalent in gram (g)
is V can be obtained by dividing the Fbuoyancy value by the
value of g. Air density can be determined using the following
equation:
T
p
273*05.287
325.101 (5)
where,
is the air density in g/cm3, p is the chamber
pressure in kPa, T is the temperature in degree Celsius.
In order to determine the effect of buoyancy force in the
digital balance reading and compare it with the theoretical
buoyancy force a piece of metal of known volume was placed
over the digital balance and the pressure inside the glass
chamber was increased from 0kPa to 300kPa at an increment
of 20kPa. Readings from the digital balance were recorded at
each pressure increment. Readings from the digital balance
were also taken by increasing the chamber pressure from 0 to
300kPa at an increment of 20kPa when there is no weight over
the balance in order to take into consideration of the effect of
buoyancy force on the digital balance itself. Actual error in
weight measurement from the digital balance can be obtained
by combining these two readings. This actual error in the
measurement was almost equal to the difference of the
theoretical bouyancy force due to the increase in chamber
pressure from the bouyancy force at 0kPa pressure on the
metal piece calculated using eq. (4) due to the increase of
chamber pressure from 0 to 300kPa as shown in Fig. 8.
Therefore, the error in the digital balance is due to the
buoyancy force only.
Correction for Buoyancy Force:
The above discussion implies that the volume change
measurement from the digital balance needs a correction. We
can correct this value using the following equation derived
based on the buoyancy force calculation equations (4) and (5):
VT
pwww bcor *
273*05.287
325.101
(6)
where, corw and w is the corrected weight and the weight
reading of balance at backpressure p respectively both in gram
0 50 100 150 200 250 300151
152
153
154
155
156
Chamber pressure (kPa)
Dig
ital
bal
ance
rea
din
g (
g)
Effect when mass over balance
Effect of empty balance
Combined effect
Fig. 7. Effect of buoyancy force on the digital balance reading.
0 50 100 150 200 250 300-0.02
0
0.02
0.04
0.06
0.08
Err
or
in w
eig
ht
dif
fere
nce
(g
)
Chamber pressure (kPa)
Actual error in weight difference
Theoretical weight difference due to bouyancy
Fig. 8. Comparison of actual error in measured weight in digital balance
and calculated theoretical buoyancy effect on weight.
e digital balance reading.
0 50 100 150 200 250 300-0.1
-0.08
-0.06
-0.04
-0.02
0
wb=0.0003(p)+0.0008
Chamber pressure (kPa)
Dig
ital
bal
ance
rea
din
g (
g)
No mass over Digital balance
Fig. 9. Effect of buoyancy force on the digital balance reading when no
weight over balance.
Page 6
International Journal of Civil & Environmental Engineering IJCEE-IJENS Vol: 11 No: 04 39
119104-5858 IJCEE-IJENS © August 2011 IJENS I J E N S
(g), p is the pressure in glass chamber or backpressure in kPa,
T is the experiment room temperature in degree Celsius (˚C),
V is the volume of the plastic pot and water containing it in
cm3, and bw is the balance reading in gram (g) at
backpressure p when balance is empty. The reading from the
digital balance when the balance is empty bw can be obtained
from the following equation which is obtained from the
readings obtained from digital balance when it is empty as
shown in Fig. 9:
0007.00003.0 pwb (7)
where, bw is in gram (g) and p is the back pressure in kPa.
C. Experimental Program
In this study, specimens were cured under two different
stresses of 0kPa and 98kPa in order to investigate the stress
dependency and for each curing stress 4 different curing
period of 7, 28, 90 and 180 days were considered in order to
investigate and time dependency of the strength and
deformation characteristics of cement-mixed sand. Specimens
were taken out of their moulds after their respective curing
days and saturated with de-aired water in the triaxial cell.
Backpressure and cell pressure were then applied up to
200kPa and 225kPa respectively, while the effective confining
stress was kept constant at 25kPa. In some cases, the
backpressure and cell pressure were increased up to 300 kPa
and 325kPa in order to get better saturation of the specimen.
Because, at high pressure the air bubbles inside the specimen
which cannot be removed from the specimen during the
application of vacuum pressure were diluted to water. The
specimens were then isotropically consolidated to an effective
confining stress of 98 kPa, followed by drained monotonic
loading. Young’s modulus Ev and shear modulli Ghh, Ghv and
Gvh were determined at each 50 kPa intervals during
monotonic loading using small strain cyclic loading and
bender element test respectively. In addition, Ghh, Ghv and Gvh
are also determined at each 10 kPa interval during isotropic
consolidation. In case of specimen prepared with clean
Toyoura sand, the specimen was isotropically consolidated to
an effective confining stress of 98 kPa, and then drained
monotonic loading was applied. The stress path for all the test
cases is shown in Fig. 10. The effective vertical and horizontal
stress are denoted by σv' and σh', respectively. The change in
shear modulus in vertical direction Gvh during curing stage is
also measured for specimens cured without and under stress
using a pair of bender elements attached in the special
consolidometer as shown in Fig. 2.
In all test cases, loading strain rate was kept constant at
0.01% /min. For cyclic loading, at each 50 kPa interval during
monotonic loading, 5 cycles with an amplitude of ±4 kPa were
used to produce small strain changes of about 10-5
to 10-4
.
Toyoura sand behalves elastically at strain levels of 10-6
to 10-
5. But at such low strain levels, interference of noise affects
the true value of Young’s modulus. Thus in this study, a
higher strain level is used as cement-mixed sand is much
stiffer than Toyoura sand. Properties of specimens are as
shown in Table III.
IV. MATERIAL AND METHODOLOGY
A. Stress-strain Relationship
Figs. 11 (a) and (b) shows the stress-strain relationships for
specimens cured without and under stress. It can be observed
that the deviator stress q increases with the axial strain εv,
reaches to peak and then reduces gradually. This brittle
behavior in the post peak region is more prominent in case of
specimens cured for longer periods of time. The cementation
bonds result in changes in the cement-mixed Toyoura sand
from a ductile to stiff brittle material, its strength increases by
a factor of approximately 4 over 180 days. The stiffness of
specimens increases notably with curing time regardless of the
availability of acting stresses during curing which is similar to
the results obtained by Rabbi [5]. On the other hand,
specimens cured under stress are noted to be stiffer than those
cured in the absence of stress, as shown in Fig. 12 which is
more clear than the results obtained by Rabbi [5]. This reflects
the coupled effect of the specimen becoming denser upon
loading during the initial curing stage and the hydration
process of the cement.
B. Peak Strength
TABLE III PROPERTIES OF SPECIMEN
Specimen
type
Curing
Overburden
stress
Curing time
(days)
Effective
confining
stress (kPa)
Wet density
before curing
(g/cm3)
Void ratio e
Cement-
mixed sand
0 kPa
7 98 1.612 0.795
28 98 1.630 0.794
90 98 1.622 0.794
180 98 1.622 0.793
98 kPa
7 98 1.620 0.791
28 98 1.630 0.790
90 98 1.640 0.790
180 98 1.630 0.789
Clean sand - - 98 1.54 0.718
Isotropic consolidation
σh′ = σv′
Monotonic loading +
Cyclic loading &
Bender element test
σv′
σh′ 98kPa
Fig. 10. Stress path for all the tests
Page 7
International Journal of Civil & Environmental Engineering IJCEE-IJENS Vol: 11 No: 04 40
119104-5858 IJCEE-IJENS © August 2011 IJENS I J E N S
Fig. 13 shows that the peak strength qmax increases with
curing time regardless of the availability of acting stresses
during curing. Peak strength increases with time and
specimens cured under stress are noted to have higher peak
strength during shearing than those cured in the absence of
stress. Although the peak strength is little less than that
obtained by Rabbi [5] the increasing trend shows similar
tendency as obtained by Rabbi [5]. This again shows the
acting stresses during curing enhances the formation of
cementation bonds and thus improves the specimen’s
resistance to shearing. The rate of increase in peak strength
with curing period for specimens cured without any stress and
under stress can be expressed by the following equations (2)
& (3) respectively, which are obtained directly from Fig. 13.
For specimens cured without stress
)log(16.3156.731max Tq (8)
And for specimens cured under stress
)log(60.4874.790max Tq (9)
where, maxq
Peak strength in kPa
and T Curing period in days.
The above equations show that the rate of increase in peak
strength is higher for the specimen cured under stress which
also coincides with the results obtained by Rabbi [5]. This
again shows that the curing stresses enhance the formation of
the cementation bonds and thus improve the shear strength of
specimen.
0 0.5 1 1.5 2 2.5 3 3.5 40
200
400
600
800
1000
1200
Axial strain, v (%)
Dev
iato
r st
ress
, q (
kP
a)
7 days 0kPa
28 days 0kPa
90 days 0kPa
180 days 0kPa
Curing stress = 98kPa
7 days 28 days90 days
180 days
Toyoura sand
Fig. 11(b). Stress-strain relationship for specimens cured under stress.
0 0.5 1 1.5 2 2.5 3 3.5 40
200
400
600
800
1000
1200
Axial strain, v (%)
Dev
iato
r st
ress
, q (
kP
a)7 days 0kPa
28 days 0kPa
90 days 0kPa
180 days 0kPa
7 days 90 days
180 days
28 days
Curing stress = 0kPa
Toyoura sand
Fig. 11(a). Stress-strain relationship for specimens cured without stress.
0 0.5 1 1.5 2 2.5 3 3.5 40
200
400
600
800
1000
1200
Axial strain, v (%)
Dev
iato
r st
ress
, q
(k
Pa)
7 days 0kPa
28 days 0kPa
90 days 0kPa
180 days 0kPa
7 days 0kPa
28 days 0kPa
90 days 0kPa
180 days 0kPa
Curing under stress
Curing without stress
Toyoura sand
Fig. 12. Stress-strain relationship for specimens cured without and
under stress.
1 5 10 50 100 500500
600
700
800
900
1000
1100
1200
qmax =731.56 + 31.16 ln (T)
qmax =790.74 + 48.60 ln (T)
Curing stress = 0kPaCuring stress = 98kPa
Curing period, T (Days)
Pea
k s
tren
gth
, q m
ax (
kP
a)
Fig. 13. Peak strength variation with curing time.
0 0.5 1 1.5 2 2.5 3 3.5 4-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
Axial strain v (%)
Vo
lum
etr
ic s
train
v (
%)
7 days 0kPa
28 days 0kPa
90 days 0kPa180 days 0kPa
7 days 98kPa28 days 98kPa90 days 98kPa180 days 98kPa
Cured without stress
Cured under stress
Fig. 14. Volume change behavior for specimens cured without and
stress.
Page 8
International Journal of Civil & Environmental Engineering IJCEE-IJENS Vol: 11 No: 04 41
119104-5858 IJCEE-IJENS © August 2011 IJENS I J E N S
C. Volume Change Behavior
As mentioned earlier, the volume change of specimen was
measured using a digital balance and corrected for the effect
of buoyancy force due to the increase in backpressure using
equations (6) and (7). Fig. 14 shows the relationship between
volumetric strain and axial strain during shearing for
specimens cured without stress and under stress. Here, the
positive and negative value of the volumetric strain indicates
the dilative and contractive behavior respectively. It is
observed that specimens become less compressive and more
dilatant with longer curing times.
As expected, specimens cured under stress are more
dilatant than those cured without stress. This also agrees with
the results obtained by Rabbi [5]. However, the reverse is seen
in specimens cured for 90 days and 180 days specimens. The
concentration of strain at the slip surface may result in the
concentration of the volume change around the slip surface,
making the specimen as a whole less dilatant.
D. Small Strain Cyclic Loading
Young's modulus Ev is dependent on the loading stress σ'v
and the void ratio e of specimen. The following equation can
be used to represent Ev (e.g. [20], [14]):
vh n
r
v
n
r
hv
r
v
ppeFC
p
E
'' (10)
where e
eeF
1
17.2)(
2
Cv, nh and nv are constants and can be obtained from
experimental data, while pr is the reference stress. In this study,
the reference stress pr is taken as 1 kPa. Since the void ratio, e
is almost same for all the specimens, F(e) is constant for all
specimens. Also, since the effective confining stress, σ′h is
constant for all the specimens, Equation (10) can be expressed
in the form n
vv AE , where A and n are constants. This is
a straight line when Ev and σ'v are plotted in logarithmic scale.
Therefore, Ev is plotted against vertical effective confining
stress, σ'v in a logarithmic scale, as shown in Fig. 11.
Fig. 15 shows the change in Ev during the monotonic
loading for specimens cured without stress and under stress. It
can be observed that Ev increases with σ'v and follows eq. (10).
Ev drops when vertical effective stress σ'v is about 50% to 60%
of its peak value which is similar to the Young’s modulus
obtained by Rabbi [5]. This may be explained as follows: Ev
increases when specimens become stiffer with increases in the
vertical contact forces between the sand particles during
monotonic loading, but it decreases when the cementation
bonds start to break down.
It is found that the Young's modulus increases notably with
the curing period, regardless of the availability of curing stress.
Also, it is quite apparent that specimens cured under stress
have a higher Young's modulus, regardless of the number of
curing days. The rate of increase of Ev with σ'v, however, is
higher for specimens cured in the absence of stress. This
shows that specimens cured without stress are more
compressive. However, the rate of increase in Ev obtained by
Rabbi [5] is lower than that obtained in the current study.
E. Bender Element Test
As is the case with the Young's modulus, the shear modulus,
G is dependent on the vertical effective stress σ'v, the
horizontal effective stress σ'h and the void ratio e of specimen.
The following equation can be used to represent G (e.g. [20],
[14]):
cba n
r
c
n
r
b
n
r
a
r pppeCF
p
G
''' (10)
C, na, nb and nc are calculated from experimental data,
while pr is the reference stress,σa’ is the principle stress in
the direction of shear wave,σb’ is the principle stress in the
direction of soil particle movement,σc’ is the principle stress
in the direction perpendicular to both directions of shear wave
and soil particle movement. Reference stress pr is taken as
1kPa similar to Young’s modulus. Rewritting the equations in
terms of σ’v and σ’h for triaxial test, we have the following
equations:
100 200 400 1000 2000
500
1000
5000
Effective vertical stress 'v (kPa)
Young's
modulu
s E
v (
MP
a)
7 days 0kPa28 days 0kPa90 days 0kPa180 days 0kPa7 days 98kPa28 days 98kPa90 days 98kPa180 days 98kPa
Fig. 15. Elastic Young’s modulus variation with effective vertical stress (log scale).
Page 9
International Journal of Civil & Environmental Engineering IJCEE-IJENS Vol: 11 No: 04 42
119104-5858 IJCEE-IJENS © August 2011 IJENS I J E N S
hn
r
hhh
r
hh
peFC
p
G
' (12)
vh n
r
v
n
r
hhv
r
hv
ppeFC
p
G
'' (13)
hv n
r
h
n
r
vvh
r
vh
ppeFC
p
G
'' (14)
In this study, the void ratio e is almost constant for all the
specimens and the effective confining stress, σ′h is also
constant for all the specimens. Therefore, equations (6) to (8)
can be generalized as mvBG , where B and m are
constants. This is also a straight line when G and σ'v are
plotted on a logarithmic scale, as shown in Figs. 16 to 18.
Change of Gvh during Curing
The change in the shear modulus in the vertical plane Gvh
and change in void ratio e during the application of curing
stress and also during the entire curing stage is shown in Fig.
19. The results show that as the void ratio decreases during the
loading stage, Gvh increases; however, as time passes, Gvh
continues to increase even when void ratio remains almost
constant. Similar to Rabbi et al., [5] this can be explained by
the strengthening of the cementation bonds due to the
hydration of cement. This is also true for specimens cured
without stress (clear dots). This observation, i.e. the Gvh of
specimen cured under stress is higher than that of cured
without stress, again shows the coupled effect of loading
during the initial curing stage and the hydration process of
cement which is similar to the results obtained by Rabbi [5].
In the specimens cured without stress, Gvh approaches the
Gvh of the specimen cured under stress as time passes.
Compared to the results obtained by Rabbi [5], the difference
between the Gvh value of two specimens are found lower in the
current study as curing time increases to 180 days. Due to the
time constraints of this study, whether or not the Gvh of the
specimen cured in the absence of stress actually increases to
the level of that cured under stress with time was not able to
be observed.
10 20 50 100 200 500 1000 2000
400
1000
20007 days 0kPa28 days 0kPa90 days 0kPa180 days 0kPa7 days 98kPa28 days 98kPa
'h = 'v 'h < 'v
7 days
28 days
90 days
180 days
Effective vertical stress 'v (kPa)
Shea
r m
od
ulu
s G
hh (
MP
a)
Fig. 16. Variation of Ghh with effective vertical stress.
10 20 50 100 200 500 1000 2000
400
1000
20007 days 0kPa28 days 0kPa90 days 0kPa180 days 0kPa7 days 98kPa28 days 98kPa90 days 98kPa180 days 98kPa
'h = 'v 'h < 'v
7 days
28 days
90 days
180 days
Effective vertical stress 'v (kPa)
Sh
ear
mo
du
lus
Gh
v (
MP
a)
Fig. 17. Variation of Ghv with effective vertical stress.
10 20 50 100 200 500 1000 2000
400
1000
2000
7 days 98kPa28 days 98kPa90 days 98kPa180 days 98kPa
'h = 'v 'h < 'v
7 days
28 days
90 days
180 days
Effective vertical stress 'v (kPa)
Shea
r m
od
ulu
s G
vh (
MP
a)
7 days 0kPa28 days 0kPa90 days 0kPa180 days 0kPa
Fig. 18. Variation of Gvh with effective vertical stress.
Gvh (
MP
a)
Curing stress of 0kPa Curing stress of 98kPa
0
1000
2000
30007 28 90180Days:
10-1 1 10 102 103 104 105 106 107
Void
rat
io e
Time (min)Curing stress (kPa)
0 20 40 60 800.785
0.79
0.795
0.8
Fig. 19.. Change of Gvh and void ratio during curing stage.
Page 10
International Journal of Civil & Environmental Engineering IJCEE-IJENS Vol: 11 No: 04 43
119104-5858 IJCEE-IJENS © August 2011 IJENS I J E N S
Change of Gvh during Shear Loading
Figs. 16 – 18 show the change in the shear moduli Ghh, Ghv
and Gvh respectively with vertical effective stress σ'v in both
isotropic consolidation and triaxial compression stages for
specimens cured in the absence of stress and under stress
during curing. For all the specimens, Ghh, Ghv and Gvh increase
with the n power of σ'v which follows eq. (12) to (14) and it
drops when σ'v reaches about 40% to 60% of its peak value.
As the specimens become denser during monotonic loading,
shear moduli increase and then drop when the cementation
bonds start to break. This is similar to what occurs in the case
of the Young's modulus. Shear moduli Ghh, Ghv and Gvh of
cement-mixed sand increase with time regardless of the
availability of curing stress, which is the same as that for
stiffness, peak strength and elastic modulus. By comparing the
difference in shear modulus Gvh between specimens cured
without and under stress, as shown in Fig. 15, it can be noted
that specimens cured under stress have a higher Gvh, than
those cured without stress, regardless of the number of curing
days. On the other, there is no apparent difference in Ghh or
Ghv, as observed in Figs. 17 and 18 respectively. The higher
value in Gvh of specimens cured under stress can be attributed
to the coupled effect of overburden stress and formation of
cementation bonds as discussed earlier. On the other hand,
since the consolidation curing apparatus only allowed
overburden stress to be exerted in the vertical plane,
significant effect might not be seen in Ghh and Ghv, as
compared to isotropic loading. It is also observed that the rate
of increase of Gvh with σ'v is slightly higher for specimens
cured without stress, which is also similar to what was found
in the case of the Young's modulus.
F. Phase Transformation Points
The points where Ev and Gvh, Ghv and Ghh changes their
phase from increasing to decreasing tendency is introduced
here as phase transformation points for Young's modulus
(PTPEv) and phase transformation for shear moduli (PTPGvh,
PTPGhv, PTPGhh) respectively. The phase transformation points
for Ev, Gvh, Ghv and Ghh is determined by drawing two straight
lines in the initial and final straight portion of the Ev- σ'v and G
- σ'v curves respectively when they are plotted in log-log scale
as shown in Fig. 15 and Figs. 16 to 18 respectively for Ev, Gvh,
Ghv and Ghh respectively. Crossing points of these two lines is
taken as phase transformation points for Ev and corresponding
G value. The detail description of determination of Phase
transformation points is described in Rabbi et al., [5].
In Figs. 20 and 21 the phase transformation points for Ev
(PTPEv), Gvh (PTPGvh), Ghv (PTPGhv), Ghh (PTPGhh) volume
change (PTPv) and peak strength (qmax) is plotted in stress-
strain field for specimens cured without and under stress
respectively. With the increase in curing period, the deviator
stress for the PTPqmax and PTPv states is higher. The phase
transformation point for volume change is taken where the
volumetric strain curve changes their phase from compressive
to dilatant behavior. It can be observed that PTPEv and PTPG
obtained first during shear loading. Following that, phase
transformation points for volume change PTPv and qmax is
obtained which is similar to the case of Rabbi [5]. It can also
observe that PTPEv and PTPGvh are found almost at the same
axial strain level of 0.2% to 0.3% irrespective of curing stress
and curing period before PTPv and qmax. This result is almost
similar to that obtained by Rabbi [5] which strengthens the
point of discussion that the degradation of cementation bond
starts after a certain level of axial deformation due to the
relative displacement of particles. This amount of relative
displacement causes the cementation bond starts to break
down but this relative displacement level is not enough to
cause the specimen shows dilatancy. Breakage of cementation
bond occurs at relatively small strain level and is a
predominant factor to reduce stiffness of cement-mixed sand.
V. CONCLUSIONS
-Addition of a small amount of cement (4.13% by weight)
with Toyoura sand gives a much higher strength than fresh
sand alone, about 4 times after 180 days.
-Curing stress increases the stiffness, peak strength, Young's
modulus Ev and shear modulus in vertical plane Gvh
irrespective of curing period. However, shear moduli in other
two directions (Ghv and Ghh) do not show apparent increase
with the application of stress during curing stage.
-Curing period increases considerably the stiffness, peak
strength, Young's modulus and shear moduli (Gvh, Ghv and Ghh)
irrespective of the availability of curing stress.
0 0.5 1 1.5 20
200
400
600
800
1000
1200
Axial strain a (%)
Devia
tor
stre
ss q
max (
kP
a) PTPqmax ('h = 0kPa)
PTPv ('h = 0kPa)PTPEv ('h = 0kPa)PTPGvh ('h = 0kPa)PTPGhv ('h = 0kPa)PTPGhh ('h = 0kPa)
Curing stress = 0kPa
Curing period 7, 28, 90, 180 days
Fig. 20. Phase transformation points (PTP) plotted in stress-strain
field for specimens cured without stress.
0 0.5 1 1.5 20
200
400
600
800
1000
1200
Axial strain a (%)
Dev
iato
r st
ress
qm
ax (
kP
a)
PTPqmax ('h = 98kPa)PTPv ('h = 98kPa)PTPEv ('h = 98kPa)PTPGvh ('h = 98kPa)PTPGhv ('h = 98kPa)PTPGhh ('h = 98kPa)
Curing stress = 98kPa
Curing period7, 28, 90 & 180 days
Fig. 21. Phase transformation points (PTP) plotted in stress-strain field
for specimens cured under stress.
Page 11
International Journal of Civil & Environmental Engineering IJCEE-IJENS Vol: 11 No: 04 44
119104-5858 IJCEE-IJENS © August 2011 IJENS I J E N S
-Gvh increases during the application of curing stress as void
ratio decreasing. It is also noted to increase with time even
though void ratio remains constant. However, the curing
period should be further increased in order to investigate
whether the shear modulus value become equal or not for both
the specimens cured without and under stress. -Phase transformation of Ev and Gvh occurs almost at the
same level of axial strain of 0.3% to 0.4%. This might be due
to degradation of cementation bond starts from that level of
axial deformation due to the relative displacement of particles.
Further investigation is required and therefore recommended
in order to establish this point e.g., observation of microscopic
view of the cement bonds between particles at that level of
axial strain and also at initial and final stage of application of
loading to the specimens.
ACKNOWLEDGMENT
The first author acknowledge to the Foreign Student Office
(FSO) for giving assistance and guidelines for his Master’s
study in the Saitama University of Japan. The valuable
discussion and suggestions with Assistant Professor Dr.
Shinya Tachibana of Geosphere Research Institute of Saitama
University (GRIS) is also acknowledged. The first author
acknowledges the valuable time and the guidelines of Dr.
Jianglian Deng, JSPS Post Doctoral Fellow of IIS, Tokyo
University for teaching him the triaxial testing system with
very sensitive and modern automatic triaxial apparatus. He
also acknowledges the help of his colleagues Mr. Takeuchi
Yasunary, Mrs. July win and the undergraduate student of
Saitama University Mr. Tomonori Masaki for their help and
mental support during the triaxial test.
REFERENCES [1] H. Aoki, T. Yonezawa, O. Watanabe, M. Tateyama and F.
Tatsuoka, ‖Results from full-scale loading tests on a bridge abutment
with backfill of geogrid-reinforced cement-mixed gravel‖, Geosynth. Engg. Journal., Japanese chapter of International Geosynthetics
Society, 2003, Vol. 18, pp. 237-2428 (in Japanese).
[2] K. Watanabe, M. Tateyama, G. Jiang, F. Tatsuoka and T. N. Lohani, ‖Strength characteristics of cement-mixed gravel evaluated by
large triaxial compression tests‖, Proc. Of 3rd International Conference
on Pre-Failure Deformation Characteristics of Geomaterials (eds. By Di Bendetto et al.) Lyon, Balkema, 2003, Vol. 1, pp. 683-693.
[3] F. Tatsuoka, H. Nawir and R. Kuwano, ―A modeling procedure of
shear yielding characteristics affected by viscous properties of sand in
triaxial compression‖, Soils and Foundations, 2004, Vol. 44, No. 6, pp.
83-99.
[4] L. Kongsukprasert, F. Tatsuoka & H. Takahashi, ―Ageing and viscous effects on the deformation and strength characteristics of cement-mixed
gravelly soil in triaxial compression. Soils and Foundations, 2005(b),
Vol. 45, No. 6, pp. 55-74. [5] A. T. M. Z. Rabbi, J. Kuwano, J. Deng and W. B. Tay, ―Effect of
curing stress and period on the mechanical properties of cement-mixed
sand‖, Soils and Foundations IS-Seoul Special Issue, 2011, Vol. 51, No. 4 (to be Published).
[6] N. C. Consoli, G. V. Rotta and P. D. M. Prietto, ―Influence of curing
under stress on the triaxial response of cemented soils, Geotechnique,2000, Vol. 50, No. 1, pp. 99-105
[7] N. C. Consoli, D. Foppa, L. Festugato, & K. S. Heineck, ―Key
parameters for strength control of artificially cemented soils‖. J. Geotech. Geoenviron. Eng., 2007, Vol. 133, No. 2, pp. 197-205.
[8] L. Kongsukprasert, F. Tatsuoka & M. Tateyama, ―Several factors
affecting the strength and deformation characteristics of cement-mixed gravel‖. Soils and Foundations, 2005(a), Vol. 45, No. 3, pp. 107-124.
[9] L. Kongsukprasert, F. Tatsuoka, & M. Tateyama, ―Effects of curing
period and stress conditions on the strength and deformation characteristics of cement-mixed soil‖. Soils and Foundations, 2007,
Vol. 47, No. 3, pp. 577-596.
[10] T. N. Lohani, L. Kongsukprasert, K. Watanabe and F. Tatsuoka, ―Strength and deformation properties of compacted cement-mixed
gravel evaluated by triaxial compression test‖, Soils and Foundations.
2004, Vol. 44, No. 5, pp. 95-108.
[11] T. Taguchi, M. Suzuki, T. Yamamoto, H. Fujino, S. Okabayashi & T.
Fujimoto, "Influence of consolidation stress history on unconfined
compressive strength of cement-stabilized soil," Technical Report, Department of Engineering, Yamaguchi University, 2002, Vol. 52, No.
2, pp.87-92 (in Japanese).
[12] J. Kuwano and W. B. Tay, "Effects of curing time and stress on the strength and deformation characteristics of cement-mixed sand," Soil
Stress-Strain Behavior: Measurement, Modeling and Analysis (Ling,
H.I., Callisto, L., Leshchinsky, D. and Koseki, J. eds.), Springer, 2007, pp.413-418.
[13] H. Itoh, T. Saitoh, J. Kuwano & J. Izawa ―Development of
reinforcement wall using cement-mixed soil and geogrids.‖ Geosynthetics Technical Report, 2003, No. 11, pp.42-49.
[14] S. K. Chaudhary, J. Kuwano & Y. Hayano, ―Measurement of quasi-
elastic stiffness parameters of dense toyoura sand in hollow cylinder apparatus and triaxial apparatus with bender elements.‖ Geotechnical
Testing Journal, 2004, Vol. 27, No. 1, pp. 23-35.
[15] D. J. Shirley, and L. D. Hampton, ―Shear-wave measurements in laboratory sediments.‖ J. Acoust. Soc. Am., 1977, Vol. 63, No. 2, pp.
607-613.
[16] R. Dyvik and C. Madshus, ―Lab measurements of Gmax using bender elements.‖Proceedings of Advances in the Art of Testing Soils Under
Cyclic Conditions, V. Khosla, ed., ASCE Annual Convention, Detroit, Michigan, 1998, pp. 186-196.
[17] G. Viggani and J. H. Atkinson, ―Stiffness of fine-grained soils at very
small strains‖, Geotechnique, 1995(b), Vol. 45, No. 2, pp. 249-265. [18] G. Viggiani & J. H. Atkinsion, ―The interpretation of bender element
tests.‖ Geotechnique, 1995(a), Vol. 45, No. 1, pp. 149-155
[19] S. Mulmi, T. Sato, & R. Kuwano, ―Performance of plate type piezo-ceramic transducers for elastic wave measurements in laboratory soil
specimens‖. Seisan-Kenkyu, IIS, University of Tokyo, 2008, Vol.60,
No.6, pp. 43-47. [20] B. O. Hardin, and G. E. Blandford, ―Elasticity of particulate materials,‖
Journal of Geotechnical Engineering, ASCE, 1989, Vol. 115, No. 6, pp.
788-805.
ABOUT THE AUTHORS
Abu Taher Md. Zillur Rabbi is a member of Institute of Engineers,
Bangladesh (IEB), Japanese Geotechnical Society (JGS). He is an Assistant Professor in the department of civil engineering in Dhaka University of
Engineering & Technology (DUET), Bangladesh. He received B. Sc. Eng.
(Civil) from RUET, Rajshahi, Bangladesh in 2003 and M. Sc. Eng. (Civil & Geotechnical) from Saitama University, Japan in 2010. He has born in
Nilphamari district of Bangladesh in February 01, 1980. His research interest
includes soil improvement, laboratory based element and model tests of soil, small strain stiffness properties, dynamic properties of soil, geotechnical
earthquake engineering.
Dr. Md. Kamal Hossain is a Fellow of Institute of Engineers, Bangladesh (IEB). He is a Professor of department of civil engineering in Dhaka
University of Engineering & Technology (DUET), Bangladesh. He received
B. Sc. Eng. (Civil) from BUET, Bangladesh in 1993 and M. Sc. Engg. (Civil & Transportation) from the same University in 1997 and Ph. D. in 1999 from
UKM Malysia.. He has born in Khulna district of Bangladesh in December 01,
1968. His research interest includes soil improvement, laboratory based element and model tests of soil, small strain stiffness properties, dynamic
properties of soil, geotechnical earthquake engineering.
Dr. Jiro Kuwano is a Professor of the Department of Civil and
Environmental Engineering, Geosphere Research Institute of Saitama
University, Saitama, Japan. Prior to that he also have experience in teaching and research in Asian Institute of Technology (AIT), Tokyo Institute of
Technology (TIT), IIS-University of Tokyo. He is a member of ASCE, JSCE,
JGS, and many more renound organization. He supervised the first author’s Master’s thesis during Master’s study in Saitama University. His research
interest is Geotechnical earthquake engineering, reinforced soil, liquefaction
of soil etc.
Mr. Wee Boon Tay is a Government employee in Singapore. He is a national
of Singapore. He finished his B.Sc Engg. In Nanyang Institute of Technology, Singapore and finished his Master’s study in Tokyo Institute of Technology
(TIT), Tokyo, Japan. His research interest is the improved soil using binding
material.