Construction and Building Materials 154 (2017) 61–71

Contents lists available at ScienceDirect

Construction and Building Materials

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

Improvement of soft clay with cement and bagasse ash waste

Pitthaya Jamsawang a,⇑, Hatairat Poorahong b, Naphol Yoobanpot c, Smith Songpiriyakij d,
Pornkasem Jongpradist e
a
Soil Engineering Research Center, Department of Civil Engineering, King Mongkut’s University of Technology, North Bangkok, Thailand
b
Department of Civil Engineering, Bangkok Thonburi University, Thailand
c
Department of Civil Engineering, King Mongkut’s University of Technology, North Bangkok, Thailand
d
Construction and Building Materials Research Center, King Mongkut’s University of Technology, North Bangkok, Thailand
e
Department of Civil Engineering, Faculty of Engineering, King Mongkut’s University of Technology Thonburi, Bangkok, Thailand

h i g h l i g h t s

The construction material ‘‘CBA-admixed soft clay” is created.

The replacement of OPC with BA can enhance the strength as much as OPC alone.

20% BA is optimal because of appropriate proportions of SiO2 (from BA) and Ca(OH)2.

Average strength losses of 42.5 and 22% are observed due to soaking.

The void ratio, curing time, CSH and Et products affect strength development.

a r t i c l e i n f o a b s t r a c t

Article history: The main objective of this study is to assess the effectiveness of bagasse ash (BA), a type of agricultural
Received 25 January 2017 waste from the sugar industry, as an admixture to improve the unconfined compressive strength, chem-
Received in revised form 24 July 2017 ical composition and microstructural properties of soft clay. Ordinary Portland cement (OPC) type I is par-
Accepted 25 July 2017
tially replaced with BA and mixed with soft clay to generate CBA-admixed soft clay. The test results show
Available online 1 August 2017
that the replacement of OPC with BA can enhance the strength by as much as that of OPC alone, and 20%
BA is considered optimum. The correlation between the modulus of elasticity and the unconfined com-
Keywords:
pressive strength can be expressed as a linear function. The loss of strength due to soaking during the
Bagasse ash
Cement
early stage is higher than that during the final stage. Multiple regression analysis is conducted to predict
Soil improvement the strength while accounting for four parameters that affect strength development, namely, the after-
Unconfined compressive strength curing void ratio, the curing time, and the concentrations of the calcium silicate hydrate and the ettringite
Waste products.
Ó 2017 Elsevier Ltd. All rights reserved.

1. Introduction improve the properties of soft clay because of its compressive

The improvement of soft clay properties has been performed
widely to mitigate problems in civil engineering projects that
result from construction on top of soft clay deposits. It is well
known that soft clay has a low shear strength and high compress-
ibility [1]. Chemical stabilization is used extensively in deep soil
mixing techniques in which a cement-based slurry is mixed with
the existing soft clay [2]. The resistance of stabilized soft clay to
compression and the resulting increase in the strength of stabilized
soft clay will increase as the curing time increases. Type I ordinary
Portland cement (OPC) is commonly used as a cementing agent to
⇑ Corresponding author.
E-mail address: pitthaya_kmutnb@hotmail.com (P. Jamsawang).
strength, high bearing capacity and low compressibility for various
applications [3,4]. However, OPC is both valuable and expensive,
resulting in higher construction costs. The feasibility of using cheap
pozzolanic materials to replace OPC has been studied [5]. Poz-
zolanic materials are SiO2- and Al2O3-based materials that react
with the calcium hydroxide generated by hydrating cement to
form additional cementitious materials that, in themselves, gener-
ate little to no cementing functionality. Industrial waste is increas-
ingly being considered as a soil stabilizer. By-products such as
sludge and bottom ash have been observed and found to be useful
as pozzolanic materials for the improvement of the properties of
fine-grained soils [6,7]. Another interesting material that could
be used to replace OPC is bagasse ash (BA). BA is a by-product of
natural agricultural waste that is generated during bagasse

http://dx.doi.org/10.1016/j.conbuildmat.2017.07.188
0950-0618/Ó 2017 Elsevier Ltd. All rights reserved.
62 P. Jamsawang et al. / Construction and Building Materials 154 (2017) 61–71
Notation
A and B dimensionless constants
BA bagasse ash
CBA cement bagasse ash-admixed soft clay
CSH calcium silicate hydrate gel
D curing time (day)
D50 diameter at which 50% of the particles have a smaller
diameter
eot void ratio after the curing time
Et ettringite
E50 modulus of elasticity at 50% of the qu value
combustion in the sugar industry. BA can be used as a mineral
admixture in mortar and concrete because its chemical composi-
tion is appropriate for use as a pozzolan, mainly due to its high sil-
ica content [8,9]. Several studies have been conducted to
investigate the chemical effects and pozzolanic activity of BA and
have indicated that BA is a pozzolan that improves the perfor-
mance of concrete when mixed with cement [9–12].
Many researchers have reported that BA exhibits satisfactory
behavior as a blended cementitious material in concrete and has
great potential for use in other applications [13]. BA can be ground
to obtain fine particles and then added to concrete to produce
high-strength concrete. The use of 10–30% BA in concrete (by dry
weight) was observed to increase the concrete strength and
decrease the concrete porosity and resistance to chloride penetra-
tion because the fineness of BA contributes to a finer pore struc-
ture, which reduces chloride permeation and diffusion [14]. The
replacement of cement with 20% BA by dry weight was optimum,
and the strength of the soil increased with curing time. The main
advantage of using BA is that it reduces chloride diffusion by more
than 50% without negatively affecting the other properties of the
hardened concrete [15]. The addition of 10% BA increased the com-
pressive strength of cement paste at all hydration ages [10]. As a
construction material, clay and BA can be combined to produce
bricks with properties similar to the properties of original clay
bricks. Previously, it was found that BA can be used as a filler mate-
rial in clay bricks to reduce the use of natural raw materials [16].
The chemical deterioration of blended cement is also reduced
due to the pozzolanic nature of BA and the reduced permeability
of BA-containing mixtures [15,17]. In summary, the most common
pozzolanic materials used for the improvement of the engineering
properties of soft clay are OPC, lime, fly ash, bottom ash and indus-
trial waste.
However, a limited number of studies on the use of BA as a sta-
bilizer for soft clay have been reported. The main objective of the
study is to determine the compressive strength of soft clay stabi-
lized using OPC alone and OPC partially replaced with BA. Based
on compressive strength tests, an optimum ratio of BA replace-
ment is proposed. The development of strength in the stabilized
soil is investigated relative to curing time, the changes of the struc-
ture of the stabilized soft clay are observed using scanning electron
microscopy (SEM), and the products resulting from the chemical
reactions of the soil and cement are observed using X-ray diffrac-
tion (XRD).
2. Materials and Methods
2.1. Materials
The clay used in this study is typical soft Bangkok clay that was excavated from
the site of a condominium building project near the Chaophraya River in Non-
thaburi Province, Thailand. Soil samples were collected at depths of 3–8 m from
OPC ordinary Portland cement
pa atmospheric pressure
qu strength of CBA-stabilized soft clay
qu,D strength of CBA-stabilized soft clay after D days of cur-
ing
qu,28 strength after 28 days
SEM scanning electron microscopy
XRD X-ray diffraction
XRF X-ray fluorescence
the original ground level. The basic physical and engineering properties of the soft
clay are listed in Table 1. The natural water content, wet unit weight and specific
gravity were 81%, 15.18 kN/m3 and 2.68, respectively. The liquid and plastic limits
were 78 and 34, respectively, which correspond to a plasticity index of 44. The
undrained shear strength obtained using the torvane test was 14.5 kPa, which indi-
cated that the soil is classified as soft clay. Based on the Unified Soil Classification
System, the soil is classified as inorganic clay with high plasticity (CH). This soft clay
is too soft and weak to support upper infrastructures in construction projects,
which makes it an excellent candidate for soil improvement. The cement used in
this study was type I OPC according to the standard industrial cement TIS 15-
2547 [18] and ASTM standard C-150 [19]. The specific gravity and fineness of
OPC are 3.15 and 2900 cm2/g, respectively. The BA employed in this study was
obtained from a cleaning boiler at a sugar production factory in Singburi Province,
Thailand. The BA was burned at a temperature of approximately 600–800 °C to gen-
erate electricity. No standard for burning temperatures exists. However, ranges of
such temperatures are optimum to obtain BA that contains amorphous silica, which
is able to react with calcium hydroxide. The specific gravity of the BA was 2.35, and
the fineness of the BA ranged from 2800 to 3000 cm2/g after grinding for 30 min.
Table 2 shows the chemical composition of the soft clay and BA compared with
that of OPC according to X-ray fluorescence (XRF) spectrometry. The soft clay con-
tains 62.5% silicon dioxide (SiO2), 18.14% aluminum oxide (Al2O3), 8.49% ferric oxide
(Fe2O3), 2.35% magnesium oxide (MgO), and 1.07% calcium oxide (CaO). Note that
all compositions are considered inert relative to those of OPC and BA. The total
amount of the major components (SiO2, Al2O3 and Fe2O3) in the BA was 86.1%.
The SiO2 + Al2O3 + Fe2O3 content of the BA accounts for more than 70% of the total
composition, which is higher than the limited value specified by ASTM C618 [20] for
a class N pozzolan material. Thus, the BA used in this study had a high pozzolanic
composition and could be used for the partial replacement of OPC to improve the
quality of soft clay. A small amount of CaO was found in BA (5.9%). In the ash col-
lecting process, lime powder (CaCO3) was sprayed to transform SO3 from the gas-
eous to the solid form. The by-product was gypsum (CaSO4) intermixed in the BA.
Based on chemical composition, OPC mainly consists of CaO (69.5%), with minor
amounts of SiO2 (16.3%), Al2O3 (3.72%) and Fe2O3 (3.55%) oxides. In addition, chem-
ical analysis indicated that the silica content of BA is three times higher than the
silica content of OPC. BA contains considerable amounts of Al2O3 and Fe2O3 but
lower concentrations of CaO relative to OPC.
The grain size distribution curves of the OPC and BA were obtained via laser
particle size analysis, and the grain size distribution curves of the soft clay were
obtained using a hydrometer. These curves are plotted together for comparison,
as shown in Fig. 1. The D50 (i.e., the diameter at which 50% of the particles have a
smaller diameter) of the OPC and BA were 0.015 and 0.018 mm (15 and 18 lm),
respectively, while the D50 of the soft clay was 0.005 mm (5 lm). Thus, the average
particle sizes of OPC and BA were similar, and the average particle size of the soft
clay was approximately three times smaller than the average particle sizes of the
OPC and BA.
Table 1
Properties of unstabilized soft clay soils.
Property Value
Natural moisture content (%) 81
Specific gravity 2.68
Initial void ratio 2.16
Wet unit weight (kN/m3) 15.18
Undrained shear strength (kPa) 14.5
Liquid limit (%) 78
Plastic limit (%) 34
Plasticity index (%) 44
Unified soil classification CH
 P. Jamsawang et al. / Construction and Building Materials 154 (2017) 61–71 63

Table 2
Chemical compositions of soft clay, OPC and BA.

Compound Soft clay BA OPC

(%) (%) (%)
SiO2 62.5 71.63 16.3
Al2O3 18.4 9.37 3.72
Fe2O3 8.49 5.1 3.55
CaO 1.07 5.9 69.5
MgO 2.35 1.67 0.728
SO3 1.86 0.7 5.04
K2O 2.96 2.5 0.37
Other 2.37 3.13 0.792

Fig. 1. Particle size distribution curves of unstabilized soft clay, OPC and BA.

When the products resulting from the chemical reaction were determined by
using XRD, continuous intensity changes and gradual two theta changes were
observed, confirming that continuous and gradual changes of the internal atomic
arrangement of the materials occurred. The continuous change in the intensity
based on the diffraction angle shows that an abrupt energy reaction occurs and that
the characteristics of the material change with the structural properties of the min-
erals. Fig. 2a, b and c show the XRD analysis results from the soft clay, BA and OPC
used in this study, respectively. The XRD patterns obtained from the Bangkok clay
indicate the presence of mainly quartz, with some kaolinite, montmorillonite and
illite that likely affect the behaviors and properties of the clay. The cation exchange
capacity, specific surface area, plasticity, and retention capacity depend on the clay
mineralogy. The microstructure of the soft clay was observed by SEM, as shown in
Fig. 3a, in which the Bangkok clay shows groups of clusters. The main component of
the BA was quartz [21]. For BA, most researchers [13] worldwide have observed
diffraction angles between 15° and 35° (2h); a similar observation is made when
considering Fig. 2b, which indicates the presence of quartz and cristobalite. For
the BA used in this study, the diffraction angles were between 20.8° and 26.6°
(2h). However, the most common Si minerals observed are quartz and cristobalite.
The presence of silica in the form of cristobalite indicates that a portion of the silica
in BA is in the form of elongated spheres. Cristobalite has a silica structure with
small quantities of crystal phases that can be identified as quartz [22]. Several other
intense peaks occur after 35° (2h) that indicate the presence of quartz. Large
amounts of quartz are available because sand can adhere to the sugarcane [23] or
occur within the fibrous waste itself. Based on the SEM analysis shown in Fig. 3b,
the ground BA particles are angular and irregular in shape, with rough surfaces
and highly porous textures. Regarding the OPC, 90–95% is composed of the four
main cement minerals show in Fig. 2c, which are C3S, C2S, C3A, and C4AF. Calcium
silicates contribute to the strength of cement, with tri-calcium silicate (C3S)
accounting for most of the early strength (first 7 days) and di-calcium silicate
(C2S), which reacts more slowly, accounting for strength at later times. C2S reacts
with water similar to C3S but at a much slower rate. The products that result from
the hydration of C2S are the same as those that result from C3S Two other major
compounds, tri-calcium aluminate (C3A) and tetra-calcium aluminoferrite (C4AF),
also react with water. These reactions do not contribute significantly to the strength
of the cement, so they will be neglected in this discussion. The morphologies of the
OPC particles are shown in the SEM images in Fig. 3c and were angular and
irregular.

2.2. Methods and mix proportions

Unconfined compression tests were carried out according to ASTM D, 2166-06
[24] and Güllü [7]. A compression machine with a maximum load capacity of 100
kN was used for the testing. The testing machine can obtain a relatively sufficient
amount of data to generate very sensitive stress-strain curves for estimating a reli-
able elasticity modulus and to interpret the test results effectively. A vertical dis-
placement rate of 1.14 mm/min was applied until sample failure. The unconfined
compression tests were performed using a total of 200 samples. The unconfined
Fig. 2. XRD analysis of: (a) soft clay; (b) BA; (c) OPC samples.
compression strength (UCS) was determined based on the maximum (peak) stress
attained or on the stress at 15% axial strain, whichever was obtained first. The axial
strain is the value at which the UCS is obtained. The modulus of elasticity of the
samples could be estimated from the slope of the stress-strain curve obtained from
64 P. Jamsawang et al. / Construction and Building Materials 154 (2017) 61–71
Fig. 3. SEM of: (a) soft clay; (b) bagasse ash (BA); (c) OPC samples.
the UCS test, particularly for the routine test requirements of soil improvement pro-
jects. The estimation is acceptable because of the conservative nature of the design,
as the UCS test produces smaller values of the modulus of elasticity over field values
by a factor of 4 or 5 [25]. The secant modulus is reported to be more appropriate for
the estimation of the modulus of elasticity in the general range of field loading [8].
Based on the aforementioned reasons, E50 (i.e., the secant modulus) was estimated
by following the procedure suggested by Holtz et al. [26], in which E50 is deter-
mined from the slope of a straight line drawn from the origin to a predetermined
stress as 50% of the maximum stress (i.e., the UCS).
A series of unconfined compressive strength tests was conducted using OPC
alone and OPC partially replaced with BA. The induced microstructural changes
were also traced using XRD and SEM analyses to obtain additional information. In
this study, additional mixtures were prepared with OPC alone for comparative pur-
poses. The amounts of OPC and the water-to-cement ratio used in this study were
200 kg/m3 of wet soft clay volume and 1.0, respectively, which were recommended
by the Department of Highways [27] for improving soft clay. In this study, a mate-
rial, which is produced by mixing OPC, a waste material (BA), and soft clay, called
cement bagasse ash (CBA)-admixed soft clay was used. The cement-admixed soft
clay was designated as CBA0, and the CBA-admixed soft clays with the OPC partially
replaced with BA, containing 10, 20, 30, 40 and 50% BA by weight of cement, were
Table 3
Mixture proportion.
Symbol %OPC:%BA OPC (kg/m3) BA (kg/m3)
of wet soil volume of wet soil volume
CBA0 100:0 200 0
CBA10 90:10 180 20
CBA20 80:20 160 40
CBA30 70:30 140 60
CBA40 60:40 120 80
CBA50 50:50 100 100
designated as CBA10, CBA20, CBA30, CBA40 and CBA50, respectively. The details of
all the mixture proportions are listed in Table 3. The clay-cement admixture was
prepared according to the JGS T821-1990 standard [28] by using the kneading tech-
nique. Cylindrical specimens that were 50 mm in diameter and 100 mm in height
were prepared for the UCS tests. Then, the samples were kneaded inside a PVC plas-
tic mold that had a split lengthwise on the side and a clamp around the perimeter to
resist lateral expansion. The specimen was allowed to cure in the mold for approx-
imately 24 h to achieve equilibrium of the free moisture [7]. Then, the specimens
were moved and rapidly wrapped with a plastic sheet to prevent moisture loss to
maintain the unsoaked conditions. Unsoaked sample conditions (where adequate
drainage and unsaturated conditions are assumed to be valid for the applications
in practice) were applied in the experimental work based on the recommendations
of past studies [7]. Therefore, some results of this paper are applicable only to cases
for which there is no danger of direct contact with groundwater. For the soaked
condition, the samples were submerged under tap water throughout the curing
time. After that, the samples were carefully removed from the plastic mold, and
the dimensions and weights of the samples were measured. Finally, unconfined
compression tests were performed after the samples were cured for 3, 7, 14, 28,
90 and 120 days. The water contents of the unsoaked and soaked samples before
testing for all curing periods are presented in Table 4. After the unconfined com-
pressive strength tests, selected block-shaped samples were collected to further
investigate the changes in the microstructure of the stabilized soft clay. The block
samples were approximately 3–7 mm long and were removed from the failure
plane of the sample and placed in a container with desiccant to dry.
3. Results and discussion
3.1. Unconfined compressive strength versus curing time
When OPC and BA are blended with soft clay in the presence of
water, a set of reactions occur that results in the dissociation of
CaO in the cement and BA and the formation of cementitious and
pozzolanic gels (calcium silicate hydrate gel, CSH). First, the OPC
hydration reaction occurs. SiO2 and Al2O3 are dissolved with a high
concentration of CaO. C3S, C2S and C3A provide the cementing
property. The pozzolanic reaction is a secondary reaction that
immediately follows the hydration of OPC. SiO2 from the BA grad-
ually sequesters the calcium hydroxide, Ca(OH)2, remaining after
CSH formation. These reactions are expressed in Eqs. (1) and (2)
below:
OPC hydration reaction : C 3 S; C 2 S; C 3 A þ H2 O ¼ CSH þ CaðOHÞ2
by-product
ð1Þ
Pozzolanic reaction : CaðOHÞ2 þ SiO2 ¼ CSH
by-product from hydration compounds in pozzolan new product
ð2Þ
When the Ca(OH)2 is consumed and its concentration reduced, the
pozzolanic reaction no longer occurs. The leftover BA powder does
not contribute any cementing property to the matrix.
The increase in strength was found to be roughly related to the
type and quantity of possible reaction products (i.e., the OPC
hydration reaction product, CSH for short-term strength and poz-
zolanic reaction product for long-term strength gain). The BA as
a pozzolan (a siliceous material) contributes strength to stabilize
the soil matrix in two ways. First, the plasticity of the soil is
reduced by the exchange of calcium ions in the pore water with
 P. Jamsawang et al. / Construction and Building Materials 154 (2017) 61–71 65

Table 4
Water contents of the unsoaked and soaked samples before testing for all curing periods.

Symbol Curing time (days)

3 7 14 28 90 120
Unsoaked Soaked Unsoaked Soaked Unsoaked Soaked Unsoaked Soaked Unsoaked Soaked Unsoaked Soaked
CBA0 75.7 78.4 72.8 78.1 61.8 77.3 58.0 76.9 49.0 75.9 46.2 74.8
CBA10 76.2 78.5 74.3 78.3 60.2 77.2 56.4 76.8 47.9 75.7 44.7 74.4
CBA20 77.2 78.7 73.2 78.0 59.9 77.1 52.7 76.5 46.8 75.5 44.0 74.0
CBA30 77.4 79.0 74.9 78.2 63.3 77.3 54.1 76.6 51.9 76.2 48.8 75.6
CBA40 78.0 79.3 75.1 79.1 67.2 78.8 58.4 78.0 52.1 77.0 49.4 76.3
CBA50 78.6 79.6 75.8 79.3 70.3 79.2 62.9 78.7 59.1 78.0 57.5 77.7

monovalent cations on clay surfaces and by compression of the
adsorbed layer because of the elevated ionic strength of the pore
water. Second, the CSH gels formed by hydration and pozzolanic
reactions bind the solid particles together. This binding produces
a stronger soil matrix, resulting in an increase in compressive
strength [29].
Fig. 4a shows the effects of the curing time and stabilizer con-
tent on the qu of CBA-stabilized soft clay for the unsoaked samples.
The values of the qu of CBA0 for the unsoaked and soaked samples
after curing times of 3, 7, 14, 28, 90 and 120 days were used as a
reference to determine the effectiveness of the pozzolanic materi-
als (BA) used in this study. The results reveal that the strengths of
the CBA for all the mixtures increase as the curing time (D)
increases; thus, the development of strength during curing can
be represented using a nonlinear hyperbola function. The values
of the qu of CBA0 were found to be 0.318 MPa and 0.411 MPa after
3 and 7 days of curing, respectively, and increased rapidly to
0.947 MPa, 1.182 MPa and 1.834 MPa after 14, 28 and 90 days of
curing, respectively. The qu slightly increased to 2.138 MPa after
120 days of curing. The strength characteristics of CBA0 were
shown to rapidly increase in the short term, i.e., 3, 7 and 14 days,
continuously increase after 28 days, and slightly increase in the
long term, i.e., 120 days. After the partial replacement of the BA,
the qu value also progressively increased with curing time. During
the early stage at 3 and 7 days, the qu decreases as the amount of
BA increases. The qu values of the mixes containing BA are lower
than those of the control CBA0 (only OPC content). At 14 days,
CBA10 and CBA20 had compressive strengths of 0.995 and
1.074 MPa, respectively, indicating that these samples had higher
strengths than CBA0 and the other mixtures. At 28, 90 and
120 days of curing, CBA20 had the highest compressive strength
(1.435, 2.011 and 2.533 MPa, respectively) and CBA50 had the low-
est compressive strength (0.927, 1.083 and 1.128 MPa, respec-
tively). In addition, CBA50 had the lowest strength following all
curing periods. When considering the strength of the soaked sam-
ple, as shown in Fig. 4b, the strength increased with curing time in
all cases, which was similar to but lower than the strengths
observed in the unsoaked samples. CBA20 provided the highest
qu for all curing times, whereas CBA50 had the lowest strength in
all cases.

3.2. Optimum BA replacement content

Fig. 5 presents the relationship between the qu of unsoaked

Fig. 4. Strength development with time for the CBA-admixed soft clay under: (a)
unsoaked; (b) soaked conditions.
samples and the BA replacement content for all curing periods.
As mentioned in the previous section, CBA20 provides the highest
qu for all curing times, and CBA50 had the lowest strength in all
cases. Normally, BA can be considered as a pozzolanic material,
and its small particles fill the voids in the concrete structure, which
increases the compressive strength of the concrete [17]. However,
the BA particles are far larger than the clay particles. The packing
effect in the OPC matrix does not apply in this circumstance. The
replacement of OPC with up to 30, 40 and 50% BA by weight
reduces the soil strength at all ages. When a large amount of BA
was used to replace the OPC in CBA-admixed soft clay, the OPC
content in the mixture decreased, resulting in less CSH. Conse-
quently, the Ca(OH)2 content also decreased from the hydration
reaction. Therefore, the quantity of Ca(OH)2 for the pozzolanic
reaction was insufficient. This finding has also been reported in
other works [30]. Low qu values of the soil stabilized using cement
and BA contents of 30, 40 and 50% (by weight of binder) were
observed in this study. Based on the test results mentioned above,
the replacement of cement with 20% BA is the optimal limit for
improving soft soils. In conclusion, a suitable BA replacement ratio
for improving the compressive strengths of soils stabilized with
cement and BA is 20% by weight of the binder. The soil strength
obtained when using BA and cement to stabilize soft clay soils in
this study is consistent with that reported by Dang et al. [31],
who used BA and hydrated lime to stabilize expansive soil.
66 P. Jamsawang et al. / Construction and Building Materials 154 (2017) 61–71
Fig. 5. Relationship between qu and the BA replacement content for: (a) unsoaked;
(b) soaked samples.
3.3. Strength loss due to soaking
The strength loss of CBA specimens due to soaking can be
defined, based on concepts presented by Horpibulsul et al. [32]
related to the decrease in strength between soaked and unsoaked
soils that are improved using cementitious materials, using Eq. (3).
ðUnsoaked strength  Soaked strengthÞ
Strength loss ¼  100%
Unsoaked strength
ð3Þ
The strength loss can be observed over the curing time, as shown in
Fig. 6. The overall percentage of strength loss in the CBA decreased
as the curing time increased. The CBA with low strength had low
durability during the early stages of curing. Therefore, when the
CBA was incubated in water, the strength decreased greatly. The
strength of the CBA increased as the curing time increased, resulting
Fig. 6. Relationship between strength loss and curing time.
in a decrease in the strength loss. In the early stages (after mixing to
28 days), strength losses were apparent for all the mixtures. How-
ever, the strength losses slightly decreased and remained relatively
constant.
After 3 days, the strength losses were relatively high, with
approximate values from 40% to 45%. After 7 days, the strength loss
decreased to values from 28 to 32% and then continued to decrease
until 14 and 28 days had passed, with values from 24 to 32% and
from 22 to 28%, respectively. After 90 days, the strength of the
CBA decreased to values of approximately 19–27%, and after
120 days, the strengths decreased by values from 17% to 27%. Note
that the strength loss decreased rapidly during the first 14 days of
curing and was insignificantly different for CBA0 (OPC alone),
CBA10 and CBA20 because the amount of OPC content was suffi-
cient to give these clay mixtures short-term strength and durability,
which indicates that CBA 10 and CBA 20 are applicable for short-
term works. In addition, the strength loss decreased slowly until
it became nearly constant after 120 days because the long-term
strengths increased with increasing BA contents of 10 and 20%.
The lowest percentage of strength loss after curing was observed
for CBA20. When 20% of the cement was replaced with BA, the soft
clay amended with a sufficient amount of OPC and BA contents was
durable and very strong. These test results are consistent with the
strength results, which showed that CBA20 had the greatest com-
pressive strength. These findings indicate that CBA20 provided
the maximum strength values and exhibited the lowest loss of
strength among all the CBA specimens for each curing time. In addi-
tion, CBA50, which has the lowest strength, exhibited the highest
loss of strength for all curing times, which is consistent with previ-
ous research by Horpibulsuk et al. [32], who found that soft clay sta-
bilized using lime, calcium carbide and fly ash lost 32–37% of its
strength under soaked conditions from 7 days to 2 months.
3.4. Modulus of elasticity
The modulus of elasticity (E50) at 50% of the qu in the stabilized
soil has frequently been used in geotechnical studies [2,33–35]. The
E50 of the unsoaked cement and BA-stabilized soil samples were
investigated at 3, 7, 14, 28, 90 and 120 days, as shown in Fig. 7a.
At 3 and 7 days, CBA0 had the highest E50 values among all the mix-
tures. At 14 days, CBA10 and CBA20 had higher E50 values than
CBA0 and the other mixtures. Following 28, 90 and 120 days of cur-
ing, CBA20 had the highest E50 values and CBA50 had the lowest E50
values. The value of E50 tends to decrease when the replacement
content is higher than 20% BA. This observation agrees with the
changes of qu in the CBA-stabilized soils. The E50 values of the
soaked samples are shown in Fig. 7b. The E50 tends to increase with
curing time in all cases, which is similar to the results observed for
the unsoaked samples but with lower E50 values.
The variations of E50 with curing time and CBA content are very
consistent with the development of qu for the designed materials.
Therefore, a quantitative correlation potentially exists between qu
and E50. In terms of this relationship, many authors [1,36–39] have
reached valuable conclusions. The results of previous studies and
this study are summarized in Table 3, and the correlation obtained
using the curve-fitting method is plotted in Fig. 8. E50 tends to
increase with qu in all cases for the unsoaked and soaked samples,
which is consistent with the findings of past studies [32]. In this
study, the linear function is expressed as E50 = 90qu and yields an
R2 value of 0.9890, and the value of E50 mainly varies between 20
and 240 MPa (75–110qu) for the unsoaked samples. The E50 values
for the soaked samples vary between 11 and 132 MPa (50–92qu),
and the linear function is expressed as E50 = 60qu and yields an R2
value of 0.9894. E50 correlates well with qu for a large value of R2.
Thus, the fitting curves (Fig. 8) are useful for predicting the strength
responses of CBA samples, thereby providing experimenters with a
 P. Jamsawang et al. / Construction and Building Materials 154 (2017) 61–71
Fig. 7. E50 development with curing time for different stabilized soils: (a) unsoaked
sample; (b) soaked sample.
Fig. 8. Relationship between E50 and qu.
better design. These relationships are defined for soft Bangkok clay
with CBA.
The ratio E50/qu for soaked samples was smaller than that for
unsoaked samples by a factor of approximately 1.50, which
indicates that the soaked samples exhibit ductility rather than brit-
tleness, in contrast to the unsoaked samples [7]. As shown in
Table 5, the correlation between E50 and qu is generally linear
Table 5
Comparison of fitting curves between E50 and qu.
Materials
Cement and metakaolin agent mixed with marine clay
Soils with cement
Sediments with cement, lime and fly ash
Soft clay with cement kiln dust and fly ash
Soft clay with cement and bagasse ash
67
and is within the wide range of results obtained in previous stud-
ies. However, the linear relationship obtained in this study is dif-
ferent from the relationships observed in previous studies due to
differences in the soil properties (such as the soil heavy metal
and organic matter contents) and the use of stabilizers with differ-
ent components (such as lime, cement, fly ash and their amounts).
3.5. After-curing void ratio
From this section on, only the test results of the unsoaked sam-
ples are presented because of the recommendations of previous
studies [7]. Soft Bangkok clay has a very low strength because of
its high initial void ratio, which is approximately 2.16 for the soft
clay used in this study. Lorenzo and Begado [40] stated that the
void ratios of stabilized soils change with curing time from the ini-
tial conditions and proposed the after-curing void ratio, eot, which
accounts for the effects of curing time, clay water content, and
cement content and characterizes the engineering behaviors of
cement-admixed clay. In this study, the eot values were determined
for all mixtures of CBA-stabilized soft clay during the curing period,
as shown in Fig. 9a. The eot values of specimens with the same mix-
tures decreased significantly as the curing time increased, varying
between 2.19 and 1.02. Fig. 9a also shows that the relationships
between eot and curing time can be expressed as a function of an
exponential decay function for all mixtures. The eot value
decreased with increasing curing time because the amount of
water in the cement decreased following the OPC hydration and
pozzolanic reactions, resulting in strength increasing with time,
as shown in Fig. 9b. The decrease in void ratios implies that the
amount of contact points between soil particles increases, resulting
in an increase in frictional forces (e.g., increase in strength). In
addition, the eot values of the specimens after 3 and 7 days of incu-
bation slightly increased because the amount of BA replacement
increased. As the curing time increased to 14, 28, 90 and 120 days,
the eot value decreased from CBA0 to CBA20 and increased from
CBA20 to CBA50. This trend is similar to the observed increase in
strength with time, which is shown in Fig. 4a. Moreover, CBA20
provided the lowest eot, which corresponded to the highest qu (as
discussed above). Therefore, eot can characterize the strength of
CBA-admixed clay in this study. Lorenzo and Begado [40] proposed
the following empirical relationship between qu and eot:
qu ¼ Apa eBðeot Þ ð4Þ
where A and B are dimensionless constants and pa is the atmo-
spheric pressure.
To adapt Eq. (4), qu was plotted against eot, as shown in Fig. 9b.
The values of the dimensionless constants in this study were
A = 186.97 and B = -1.932. In the similar studies of cement-
admixed soft Bangkok clay conducted by Lorenzo and Begado
[40] and Jongpradist et al. [41], the following values were
obtained: A = 10.33, B = 0.046 and A = 40.8, B = -3.34, respectively.
The constant A depends on the type of admixture, while B is the
slope of the mean function and depends on the clay type and min-
eralogy [40]. Higher values correspond to greater strength
improvement of the cement-admixed clay.
Correlation Sources
E50 = 159qu Tongwei et al. [1]
E50 = (100–200) qu Heilli [36] and
Qiao et al. [37]
E50 = (60–170)qu Wang et al. [38]
E50 = 113qu Yoobanpot et al. [39]
E50 = (50–110)qu Present study
68 P. Jamsawang et al. / Construction and Building Materials 154 (2017) 61–71
Fig. 9. Relationship between: (a) eot and curing time; (b) unsoaked qu and eot.
3.6. Microscopic investigations of the reaction products in relation to
strength development
3.6.1. Chemical compositions
In Fig. 10a to d, XRD diffractograms of the unstabilized and sta-
bilized soils after 28 days of curing are shown. A new indication of
Fig. 10. XRD results for CBA-admixed soft clay samples after 28 days.
CSH and ettringite (Et) phases appear in the stabilized soft clay. A
small peak corresponding to CSH can be observed in the spectra
from CBA50 and the unstabilized soils, which indicates that there
was an insufficient amount of CaO from OPC and BA to produce
CSH, whereas additional pozzolanic reactions were produced for
CBA 20 because the CSH peak was the highest and the most dis-
tinct. The replacement of cement with 20% BA provided appropri-
ate proportions of SiO2 (from BA) and Ca(OH)2 (from the hydration
reactions) that result in the formation of the highest amount of
CSH during pozzolanic reactions [42]. Thus, CBA50 and CBA20 pro-
vided the lowest and highest qu, respectively, as discussed in Sec-
tion 3.2. However, no distinct peaks for Ca(OH)2 are identified
due to the consumption of the pozzolanic reaction. The distinct
peaks for CSH in all CBA-stabilized soils correspond to the poz-
zolanic reactions, while Et is attributed to the reactions between
Ca(OH)2 and reactive SO3[43]. The Et peak may originate because
of gypsum (CaSO4), as mentioned earlier. Both CSH and Et are
reportedly attributed to the qu of stabilized soft clay [44].
Figs. 11a and 12a show the development of the intensities of
CSH and Et versus curing time, respectively, for the stabilized soils
CBA0, CBA10 and CBA20, which can be expressed using power
equations. Fig. 11a also shows that the formation curves were sim-
ilar to the strength characteristic curves. In addition, the developed
strength generally increased in proportion with the amounts of
CSH and Et. Therefore, it could be concluded that these reaction
products mainly contribute to the development of strength in
CBA-stabilized soils, which can be expressed by a second-degree
polynomial equation, as shown in Figs. 11b and 12b for the stabi-
lized soils CBA0, CBA10 and CBA20.
From the laboratory test results and the XRD analysis, various
parameters were found to affect the qu value of CBA-admixed soft
clay, including the void ratio after the curing time (eot), the curing
time (D), and the concentrations of the CSH and Et products. All
these influence parameters have nonlinear relationships with the
qu value of CBA-admixed soft clay. To combine the influence
parameters, multiple regression analysis was performed. The
resulting regression for CBA-admixed soft clay with the replace-
ment of BA at 0–20% is as follows:
 P. Jamsawang et al. / Construction and Building Materials 154 (2017) 61–71 69

Fig. 11. Relationship between CSH intensity and: (a) curing time or (b) unsoaked qu. Fig. 12. Relationship between Et intensity and: (a) curing time or (b) qu.

qu;D ðMPaÞ ¼ 0:00106ðCSHÞ þ 0:000667ðEtÞ þ 0:00554ðDÞ

 1:144ðeot Þ þ 2:296 ð5Þ

where qu,D is the strength of the CBA-stabilized soft clay after D

days of curing. Comparisons between the experimental data and
the values predicted using Eq. (5) are presented in Fig. 13. The mea-
sured qu,D was 0.998 times the predicted qu,D, with R2 = 0.9970. For
curing times of 3 and 7 days, the absolute error percentage of qu
obtained from the multiple regression analysis was 11–25%. For
14, 28, 90 and 120 days, the absolute error percentage of qu was
0.2–2.9%. Over a short period, the absolute error percentage was
greater than that observed over a long period because the incuba-
tion strength of the soil-cement mixture changed rapidly and was
initially unstable. Thus, the prediction equations have high absolute
error percentages for short-term strengths. In addition, long-term
strength development consistently resulted in more accurate
equations.

3.6.2. Microstructure of CBA-admixed soft clay
This study focuses on understanding the changes in microstruc-
tures of CBA-admixed soft clay. Fig. 14a to c show the changes in
the microstructures of samples CBA0, CBA20 and CBA50 after
28 days of curing. The SEM photo of CBA0 is similar to that of
CBA20 because the eot value is only slightly different. Many hydra-
tion products can be observed, the surfaces of the samples were
rough, and the products contained CSH gel and ettringite crystals.
The cementitious compounds (as pozzolanic reaction products)
enhance the inter-cluster bonding strength and fill the pore spaces.
Consequently, the strength of CBA-admixed soft clay significantly
Fig. 13. Comparisons between the measured and predicted qu values based on
multiple regression analysis.
increases as the BA content increases, as described above. The
SEM photo of the 50% BA sample (Fig. 14c) is similar to that of
the unstabilized soil sample (Fig. 3a). The microstructure appears
highly porous, with the pore spaces being relatively larger than
those observed in CBA0 and CBA20. Thus, CBA50 provided the
highest eot, which corresponded to the lowest qu (as discussed
above in Section 3.5).
70 P. Jamsawang et al. / Construction and Building Materials 154 (2017) 61–71
Fig. 14. SEM images of the samples after 28 days of curing: (a) CBA0; (b) CBA20; (c)
CBA50 at a magnification 3500.
4. Conclusions
As a natural type of agricultural waste, BA was used as an
admixture for the partial replacement of type I OPC to improve soft
clay. This new material is referred to as CBA-admixed soft clay. The
amounts of OPC replaced by BA were 10–50% (by dry weight). The
UCS test was used to determine the optimum replacement content.
Strength loss due to soaking was observed during curing. XRD and
SEM analyses were conducted to determine the chemical reactions
that occurred during the experiment and to investigate how the
microstructure of the CBA-admixed soft clay changed. The follow-
ing conclusions were drawn from the results of this study.

The curves of strength development versus time for the CBA-
admixed and cement-admixed soft clays were similar for both
unsoaked and soaked conditions because the same hydration
and pozzolanic reactions occurred. The use of 10–20% BA can
enhance the value of qu by as much as approximately 1.12–
1.21 times that of OPC alone, whereas the use of 30–50% BA
results in approximately 0.822–0.47 times lower strength than
that of OPC alone. The correlation between E50 and qu is linear
(E50 = 60 and 90qu for soaked and unsoaked conditions,
respectively).

Strength losses of 40%–45% were observed due to soaking dur-
ing the early stage and due to low strength development, and
the strength losses observed after long-term curing were
approximately 17–27%. The minimum and maximum strength
losses were observed for CBA20 and CBA50, respectively.

The after-curing void ratio is a physical property, and the con-
centrations of CSH and Et are chemical properties. All these
properties can be used to assess the strength of CBA-admixed
soft clay.

The replacement of 20% of the OPC with BA was optimal for
improving the strength of the soft clay because it provided
the highest strength and appropriate proportions of SiO2 (from
BA) and Ca(OH)2 that were sufficient for the pozzolanic reac-
tions that resulted in the formation of CSH. When 50% of the
OPC was replaced by BA, the lowest strength was observed
because the amount of Ca(OH)2 resulting from the pozzolanic
reactions was not sufficient to improve the soil strength.

CBA20 provided the lowest eot and maximum qu because many
products, (i.e., CSH gels and ettringites) developed in CBA20,
which fills the pore spaces in CBA-admixed soft clay and pro-
duces high inter-cluster bonding strengths. However, large pore
spaces were found in CBA50, which resulted in the highest eot
corresponding to the minimum qu.

BA waste materials can be used to improve soft clay in civil
engineering projects to mitigate environmental pollution prob-
lems, which would increase the value of bagasse waste. Thus,
the use of BA could be very useful in sustainable construction
technologies.
Acknowledgments
The authors gratefully acknowledge financial support from the
Faculty of Engineering, King Mongkut’s University of Technology,
North Bangkok under Contract No. ENG-59-06. The authors also
extend their appreciation to the Office of the Higher Education
Commission for granting financial support through the Higher Edu-
cation Research Promotion and National Research University
(NRU) Project of Thailand.
References
[1] Z. Tongwei, Y. Xibing, D. Yongfeng, Z. Dingwen, L. Songyu, Mechanical
behaviour and micro-structure of cement-stabilised marine clay with a
metakaolin agent, Constr. Build. Mater. 73 (2014) 51–57.
[2] P. Jamsawang, P. Voottipruex, P. Jongpradist, D.T. Bergado, Parameters
affecting the lateral movements of compound deep cement mixing walls by
numerical simulations and parametric analyses, Acta Geotech. 10 (2015) 797–
812.
[3] A. Mohammadinia, A. Arulrajah, J. Sanjayan, M.M. Disfani, M. Win Bo, S.
Darmawan, Laboratory evaluation of the use of cement-treated construction
and demolition materials in pavement base/subbase applications, J. Mater. Civ.
Eng. 27 (6) (2014). 04014186.
[4] H. Güllü, H. Canakci, I.F. Al, Zangana. Use of cement based grout with glass
powder for deep mixing, Constr. Build. Mater. 137 (2017) 12–20.
[5] M.L. Garcia, J. Sousa-Coutinho, Strength and durability of cement with forest
waste bottom ash, Constr. Build. Mater. 41 (2013) 897–910.
[6] H. Güllü, S. Girisken, Performance of fine-grained soil treated with industrial
wastewater sludge, Environ Earth Sci. 70 (2013) 777–788.
[7] H. Güllü, A factorial experimental approach for effective dosage rate of
stabilizer: An application for fine-grained soil treated with bottom ash, Soils
Found. 54 (3) (2014) 462–477.
[8] T. Akram, S.A. Memon, H. Obaid, Production of low cost self compacting
concrete using bagasse ash, Constr. Build. Mater. 23 (2009) 703–712.
[9] K. Ganesan, K. Rajagopal, K. Thangavel, Evaluation of bagasse ash as
supplementary cementitious material, Cem. Concr. Compos. 29 (2007) 515–
524.
[10] N.B. Singh, V.D. Singh, S. Rai, Hydration of bagasse ash-blended portland
cement, Cem. Concr. Res. 30 (2000) 1485–1488.
 P. Jamsawang et al. / Construction and Building Materials 154 (2017) 61–71
[11] G.C. Cordeiro, R.D. Toledo Filho, L.M. Tavares, E.M.R. Fairbairn, Pozzolanic
activity and filler effect of sugar cane bagasse ash in Portland cement and lime
mortars, Cem. Concr. Compos. 30 (2008) 410–418.
[12] E.V. Morales, E. Villar-Cociña, M. Frías, S.F. Santos, H. Savastano Jr., Effects of
calcining conditions on the microstructure of sugar cane waste ashes (SCWA):
Influence in the pozzolanic activation, Cem. Concr. Compos. 31 (2009) 22–28.
[13] R. Alavéz-Ramíreza, P. Montes-García, J. Martínez-Reyes, D.C. Altamirano-
Juárez, Y. Gochi-Ponce, The use of sugarcane bagasse ash and lime to improve
the durability and mechanical properties of compacted soil blocks, Constr.
Build. Mater. 34 (2012) 296–305.
[14] S. Rukzon, P. Chindaprasirt, Utilization of bagasse ash in high-strength
concrete, Mater. Design. 34 (2012) 45–50.
[15] A. Noor-ul, Use of bagasse ash in concrete and its impact on the strength and
chloride resistivity, J. Mater. Civil. Eng. 23 (5) (2011) 717–720.
[16] K.C.P. Faria, R.F. Gurgel, J.N.F. Holanda, Recycling of sugarcane bagasse ash
waste in the production of clay bricks, J. Environ. Manage. 101 (2012) 7–12.
[17] N. Chusilp, C. Jaturapitakkul, K. Kiattikomol, Utilization of bagasse ash as a
pozzolanic material in concrete, Constr. Build. Mater. 23 (2009) 3352–3358.
[18] TIS 15-2547. Thai Industrial Standards specification for portland cement Part
1, 2004.
[19] ASTM C 150: Standard Specification for Portland Cement, Annual Book of
ASTM Standard, Section 4 Construction, Volume 04.01 Cement, Lime, Gypsum.
[20] ASTM C 618: Standard specification for coal fly ash and raw or calcined
natural pozzolan for use as a mineral admixture in concrete, Annual Book
of ASTM Standard, Section 4 Construction, Volume 04.02 Concrete and
Aggregate.
[21] N. Chusilp, C. Jaturapitakkul, K. Kiattikomol, Effects of LOI of ground bagasse
ash on the compressive strength and sulfate resistance of mortars, Constr.
Build. Mater. 23 (2009) 3523–3531.
[22] A. Abbasi, A. Zargar, Using Bagasse ash in concrete as pozzolan, Middle-East J.
Sci. Res. 13 (6) (2013) 716–719. ISSN 1990-9233.
[23] G.C. Cordeiro, R.D.T. Filho, L.M. Tavares, E.M.R. Fairbairn, Ultrafine grinding of
sugar cane bagasse ash for application as pozzolanic admixture in concrete,
Cem. Concr. Res. 39 (2009) 110–115.
[24] ASTM D 2166-06: Standard Test Method for Unconfined Compressive Strength
of Cohesive Soil. West Conshoshocken, PA, United States.
[25] J.E. Bowles, Foundation Analysis and Design, 5th ed., McGraw-Hill, 1996.
[26] R.D. Holtz, W.D. Kovacs, T.C. Sheaahan, An Introduction to Geotechnical
Engineering, 2nd ed., Prentice Hall, NJ, 2011.
[27] Department of Highways of Thailand (DOH) and Japan International
Cooperation Agency (JICA). (1998). Manual for Design and Construction of
Cement Column Method. DOH., pp.75.
[28] Japanese Geotechnical Society. Practice for Making and Curing Noncompacted
Stabilized Soil Specimens. JGS T821-1990. Tokyo, Japan, 1990.
71
[29] E.O. Tastan, T.B. Edil, C.H. Benson, A.H. Aydilek, Stabilization of organic soils
with fly ash, J. Geotech. Geoenviron. Eng. 137 (9) (2011) 189–833.
[30] L. Lam, Y.L. Wong, C.S. Poon, Degree of hydration and gel/space ratio of high
volume fly ash/cement systems, Cem. Concr. Res. 30 (2000) 747–756.
[31] L.C. Dang, H. Hasan, B. Fatahi, H. Khabbaz, Influence of Strength and
Mechanical Behaviour of Bagasse Ash and Hydrated Lime Stabilized
Expansive Soil, GEO Québec, At Québec City, Canada, 2015.
[32] S. Horpibulsukn, C. Phetchuay, A. Chinkulkijniwat, A. Cholaphatsorn, Strength
development in silty clay stabilized with calcium carbide residue and fly ash,
Soils Found. 53 (4) (2013) 477–486.
[33] P. Jamsawang, P. Voottipruex, P. Boathong, W. Mairaing, S. Horpibulsuk,
Threedimensional numerical investigation on lateral movement and factor of
safety of slopes stabilized with deep cement mixing column rows, Eng. Geol.
188 (2015) 159–167.
[34] P. Jamsawang, P. Boathong, W. Mairaing, P. Jongpradist, Undrained creep
failure of a drainage canal slope stabilized with deep cement mixing columns,
Landslides 13 (5) (2015) 939–955.
[35] P. Jamsawang, N. Yoobanpot, N. Thanasisathit, P. Voottipruex, P. Jongpradist,
Three-dimensional numerical analysis of a DCM column-supported highway
embankment, Comput. Geotech. 72 (2016) 42–56.
[36] K. Heilli, Stabilization of clay with inorganic by products, J. Mater. Civ. Eng. 12
(4) (2000) 307–309.
[37] X.C. Qiao, C.S. Poon, C.R. Cheeseman, Investigation into the
stabilization/solidification performance of Portland cement through cement
clinker phases, J. Hazard. Mater. 139 (2007) 238–243.
[38] D. Wang, N.E. Abriak, R. Zentar, Strength and deformation properties of
Dunkirk marine sediments solidified with cement lime and fly ash, Eng. Geol.
166 (2013) 90–99.
[39] N. Yoobanpot, P. Jamsawang, S. Horpibulsuk, Strength behavior and
microstructural characteristics of soft clay stabilized with cement kiln dust
and fly ash residue, Appl. Clay Sci. 141 (2017).
[40] G.A. Lorenzo, D.T. Bergado, Fundamental Characteristics of Cement-Admixed
Clay-New Approach, J. Geotech. Geoenviron. Eng. 130 (10) (2004) 1042–1050.
[41] P. Jongpradist, S. Youwai, C. Jaturapitakkul, Effective Void Ratio for Assessing
the Mechanical Properties of Cement-Clay Admixtures at High Water Content,
J. Geotech. Geoenviron. Eng. 137 (6) (2011) 621–627.
[42] A.L.A. Fraay, J.M. Bijen, Y.M. de Haan, The reaction of fly ash in concrete: A
critical examination, Cem. Concr. Res. 19 (1989) 235–246.
[43] J.M. Kinuthia, S. Wild, G.I. Jones, Effects of monovalent and divalentmetal
sulphates on consistency and compaction of lime-stabilised kaolinite, Appl.
Clay Sci. 14 (1–3) (1999) 27–45.
[44] S. Horpibulsuk, R. Rachan, Y. Raksachon, Role of fly ash on strength and
microstructure development in blended cement stabilized silty clay, Soils
Found 49 (1) (2009) 85–98.