You are on page 1of 11

Construction and Building Materials 336 (2022) 127544

Contents lists available at ScienceDirect

Construction and Building Materials


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

Stabilization nature and unified strength characterization for cement-based


stabilized soils
Li Liu a, Tingting Deng a, Yongfeng Deng a, *, Liangtong Zhan b, Suksun Horpibulsuk c,
Qiong Wang d
a
Institute of Geotechnical Engineering, School of Transportation, Southeast University, Nanjing 211189, China
b
MOE Key Laboratory of Soft Soils and Geoenvironmental Engineering, Zhejiang University, Hangzhou 310058, China
c
Institute of Engineering, Suranaree University of Technology, Nakhon Ratchasima 30000, Thailand
d
School of Civil Engineering, Tongji University, Shanghai 200092, China

A R T I C L E I N F O A B S T R A C T

Keywords: Although extensive studies have been conducted on cement-based stabilized soils (CBS), there is limited research
Unconfined compressive strength focusing on the effect of water content, sand/silt fraction and particle size distribution of the original clays on the
Soil–cement, ground improvement unconfined compression strength (UCS) of CBS. In this study, experimental results were combined with the
Sand incorporation
available data collected from the literatures to clarify the effect of the water content, sand/silt fraction and size
Modified binder/water ratio
Mass ratio of clay fraction to water
distribution on the CBS strength and to develop a generalized predictive strength equation for clays with various
gradations, water contents and cement contents. The findings suggest that at a constant cement mass, original
dry clay and water, the addition of sand (<80% in this case) has no influence on the UCS. However, at the same
cement mass, water, and total soil (e.g., original clay with incorporated sands), the UCS decreases because of less
absorbed water and more available water with sand substitution. It can be concluded by the scanning electron
microscopy (SEM), that sand acts as hard inclusions within the CBS matrix. Based on the nature of the sand
function within the CBS, a modified binder/water ratio (new indicator) was proposed to characterize the UCS by
combining the traditional binder/water ratio and the mass ratio of the clay fraction to water. Using the modified
binder/water ratio, the strength predictive equation was developed, which is useful as a tool for mix design to
attain a target strength. The powerfulness of the equation is confirmed by the prediction results of the available
CBS data collected from the literatures. The slope of the equation is found to be dependent on the cement hy­
dration and pozzolanic reaction with clay minerals.

strength development, which can be further classified into clay mineral-


absorbed and cement-hydrated divisions [15,16]. Evidently, the cement
1. Introduction type and dosage in the soft clay are also leading factors affecting the
strength, according to previous cognitions of concrete and mortar [17].
The ultralow strength of soft clay deposits with high water content In terms of the water/cement ratio (e.g., mw/mc), it is confirmed that a
has been a serious challenge in constructions at the coastal and riverside higher mw/mc leads to a lower strength [12,13,15,18]. Miura et al. [18]
regions. The soft clay ground is usually strengthened with introducing first introduced the mw/mc parameter from concrete industry to evaluate
cement-based cementitious materials by grouting or deep mixing the CBS’s strength, and Horpibulsuk et al. [4,12,17,19] put forward a
methods. The unconfined compressive strength (UCS) is a convenient power function of mw/mc to describe the CBS’s strength development.
engineering parameter to evaluate the reinforcement quality, which However, the difference between CBS and mortar or concrete is obvious
depends on the curing period, water content, soil composition, clay for the intervention of various clay minerals. Note that, clay minerals
mineral content, and cement incorporation ratio [1-10]. It has been well have a significant influence on the performance of CBS, especially for the
recognized by many researchers [11-14] that the strength of cement- clays with more hydrophilic and active montmorillonite.
based stabilized soils (CBS) increases with curing time for the cement Owing to the heterogeneity of natural soils, the different fractions (e.
hydration process. In the hydration process, water has a great impact on

* Corresponding author at: Institute of Geotechnical Engineering, School of Transportation, Southeast University, SEU Avenue 2, Jiangning District, Nanjing
211189, China.
E-mail address: noden@seu.edu.cn (Y. Deng).

https://doi.org/10.1016/j.conbuildmat.2022.127544
Received 26 July 2021; Received in revised form 28 March 2022; Accepted 15 April 2022
Available online 22 April 2022
0950-0618/© 2022 Elsevier Ltd. All rights reserved.
L. Liu et al. Construction and Building Materials 336 (2022) 127544

Nomenclature

CBS Cement-based stabilized soils


UCS Unconfined compressive strength
SEM Scanning electron microscopy
PSD Particle size distribution
LL Liquid limit
PL Plastic limit
PI Plasticity index
w0 Mass ratio of water to original clay
c0 Mass ratio of cement to original clay
w1 Mass ratio of water to whole original clay and added
sand
c1 Mass ratio of cement to whole original clay and added
sand
mc/mw Mass ratio of binder/water
mf/mw Mass ratio of clay fraction to water
G Modified binder/water ratio
K Slope of the UCS vs the modified binder/water ratio
l Intercept of the fitting line between the UCS and the
modified binder/water ratio

g., clay, silt, and sand fractions) can also dominate the CBS’s strength.
Taki and Yang [20] found that cement-based stabilized sand shows
higher strength than cement-based stabilized silty soil and clay. Bell
[21] observed that more stabilizing materials (e.g., cement) were
consumed when increasing the clay fraction at a given strength. A
similar conclusion was also drawn by Bergado et al. [22], who suggested
that the strength of the CBS decreases with the clay fraction and plastic
index. Chian et al. [23,24] intentionally added sand to cement-stabilized
Singapore marine clay and kaolin, and concluded that sand involution
reduced the strength of cement-based stabilized clayey soil due to more
free water available for hydration. However, He and Shen [25] found
that the strength of CBS could be significantly enhanced by the sand
addition to the raw specimen.
For fluvial deposition, the particle size distribution (PSD) of the Fig. 1. Soils’ PSD in the Yangtze delta. (a) Sample locations in Yangtze River
Delta, (b) PSD curve.
upstream soil is rather coarse, while that of the downstream soil is finer
due to hydraulic sorting. This hydraulic sorting also occurs in the land
reclamation project by hydraulic filling. In Fig. 1, the soils with the same experimental results from riverside and coastal regions were collected to
substance source were collected from the Yangtze River Delta [26-31], confirm the powerfulness of the developed strength predictive equation.
where the PSD from the literatures was summarized. Note that the PSD
varies greatly, and the clay fraction ranges from 5% to 30%, causing the 2. Materials and methods
distinct physical and mechanical properties when stabilization. Taking
the CBS’s UCS at Nantong and Shanghai as examples, the UCS with the 2.1. Materials
same cement incorporation ratio (e.g., 12%) had a large difference, i.e.
the 28-day UCS at Nantong ranged from 1.5 to 2.0 MPa [32], but the 28- The soil used was the upper marine clay sampled at Lianyungang city
day UCS at Shanghai was between 0.2 and 0.3 MPa [33]. Notably, the on the east coast of China where this original marine clay at the delta
PSD of Shanghai soft clay was finer than that of Nantong clay (fewer silt edge of ancient Yellow River is distinguished for very fine depositional
and sand fractions). matter. The basic properties of the original marine clay are listed in
The above contradictory observation indicates that the role of sand Table 1. The liquid limit (LL), plastic limit (PL), and plasticity index (PI)
incorporation in the strength formation of CBS is still contentious and
uncertain and requires further investigation. This research therefore
attempts to examine the role of sand incorporation (with various PSD Table 1
and contents) and water content on the strength change in CBS as well as Basic properties of Lianyungang soft clay.
to propose a prime parameter controlling the strength development of Property Value
CBS, which can take the effect of sand incorporation and water content Natural water content, w0 (%) 56.6
into account. The proposed prime parameter will be further adopted to Specific density, Gs 2.71
develop a generalized strength predictive equation for CBS with various Liquid limit, LL (%) 54.7
PSD and water contents. This research is novel and innovation in terms Plastic limit, PL (%) 26.4
Plasticity index, PI 28.3
of construction and building materials for civil infrastructure engineers.
Sand fraction (>0.075 mm, %) 2.5
A series of experiments were designed by adding sand to adjust the PSD Silt fraction (0.002–0.075 mm, %) 41.7
of CBS. Portland cement was used as a cementing binder to investigate Clay fraction (<0.002 mm, %) 55.8
the CBS’s strength performance. Finally, the extensive available USCS classification High-plasticity clay (CH)

2
L. Liu et al. Construction and Building Materials 336 (2022) 127544

were 54.7%, 26.4%, and 28.3%, respectively [34]. According to the During specimen preparation, the mixing procedure of the clay
Unified Soil Classification System [35], the original marine clay was slurry, cement, and sand was completed within 10 min to minimize
classified as high-plasticity clay (CH). The PSD [36] as shown in Fig. 2 water loss. After mixing, the paste was carefully poured into cylindrical
indicated that the original marine clay was classified poorly graded with molds with a diameter of 50 mm and height of 100 mm, and then tapped
57.8% clay, 39.6% silt, and 2.6% sand. The specific gravity of the soil to remove trapped air bubbles. Note that three parallel specimens were
was 2.71 [37]. prepared for each paste type at the same PSD curve. All specimens were
The sand employed in this study was a quartz sand, which was then sealed in a plastic bag in a curing room at a temperature of 20 ±
classified by a standard sieve, and then adjusted according to designed 2 ◦ C and a relative humidity of 95%. Prior to testing, the top and bottom
gradations. The cement used in this study was ordinary Portland cement surfaces of the specimens were carefully trimmed to maintain a height of
of Hailuo brand (#42.5), whose chemical compounds are presented in 100 ± 5 mm and a height/diameter ratio of 2:1. Each specimen was then
Table 2. weighed to maintain the mass variation within 3 g. At the designed
curing period (e.g., 7 and 28 days), the specimens were tested for UCS at
a constant shearing rate of 1.00 mm/min to obtain the peak strength or
2.2. Sample preparations and tests the strength at 15% axial strain [38].
In addition to macroscopic observation, scanning electron micro­
The original marine clay was first air-dried and then passed through scopy (SEM) was employed to investigate the microstructures of the CBS
sieve #10 (2 mm openings) to remove larger particles and impurities. To with 36% sand incorporation (specimen #1 in Group 2). After the UCS
explore the effect of the sand fraction on the strength of the CBS, the tests, small pieces of the specimens were trimmed to appropriate di­
sample was first thoroughly mixed with distilled water to reach a preset mensions and then immersed in liquid nitrogen (with a temperature of
water content. After curing for 24 h, cement and sand with different − 196 ◦ C) for instant freezing to terminate the ongoing hydration pro­
gradations and proportions were added to the clay sample to investigate cess. The frozen specimens were then transferred to a freeze dryer for
the effect of sand incorporation on the strength enhancement. In this sublimation for approximately 24 h. Next, the dried specimens were
study, four groups of experiments on sand fraction incorporation were polished and vacuum coated with a thin layer of gold with a thickness of
conducted. For Group 1, sands with four diameter ranges (e.g., 1–2 mm, 200–300 Å to enhance electrical conduction and prevent electric
0.5–1 mm, 0.25–0.5 mm, and 0.075–0.25 mm) were added into the stacking on the specimen surface.
original clay at original clay’s water content (e.g., w0, the mass ratio of
water to original clay) of 77.0% to study the sand fraction and size effect, 3. Results
such as 20%, 40%, 60%, and 80% (wt. of dry original clay). Subse­
quently, the 15% wt. cement of dry original clays (e.g., c0, the mass ratio 3.1. Effects of sand content, size, and gradation
of cement to original clay) was incorporated into the pastes, as shown in
Table 3. For Group 2, sand particles with different gradations were The CBS’s UCS of specimens with sand incorporation at different
mixed with the original marine clays with w0 of 60.0%, 70.0%, and sizes and proportions, after curing for 7 and 28 days, compared with the
80.0% to investigate the sand gradation effect at the same sand fraction UCS of the referential specimens, are summarized in Fig. 3 (e.g., Group
(36% wt. of sand in terms of the new specimens), and c0 was preset as 1). It is evident that the UCS varies slightly regardless of the proportion
15%. When mixing the original clay with sand, the new gradations are and size of sand particles added to the original clays. The mean UCS
listed in Fig. 2(a). Specimen #1 was composed of coarser sand, while values of all 17 specimens at 7 and 28 days are about 0.64 MPa and 0.84
specimen #5 included finer sands, where the involved sand sizes MPa, respectively. After curing for 7 days, the mean UCS with 20% sand
decreased with the specimen number, and specimen #6 was the original is 0.63 MPa with a maximum value of 0.67 MPa (particle size of added
soft clay as the reference. For Group 3, sand particles with a diameter of sand is 0.25–0.5 mm) and a minimum value of 0.61 MPa (particle size of
0.25–0.5 mm were introduced to the original clay to examine the effect added sand is 1.0–2.0 mm). After curing for 28 d, the mean UCS with
of sand fraction (e.g., 0%, 10%, 20%, and 30% wt. of dry sand and 20% sand is 0.84 MPa, with a maximum value of 0.86 MPa (particle size
original clay) when keeping the total water content of 87% (e.g., w1, the of added sand is 1.0–2.0 mm) and a minimum value of 0.81 MPa (par­
mass ratio of water to whole original clay and added sand). And the ticle size of added sand is 0.25–0.5 mm). In this case, the masses of the
cement content (e.g., c1, the mass ratio of cement to whole original clay original clay, cement, and water within the matrix are all maintained the
and added sand) was set at 15%. For Group 4, the total water content w1 same except for the PSD of the sand incorporation. This indicates that
and cement incorporation ratio c1 were designed to verify the indicator the CBS’s strength did not benefit from the addition of sand particles as
proposed in this study considering the fraction effect, as shown in long as the original clay content and water content remain the same.
Table 4. The PSD curves after mixing are presented in Fig. 2(b).

Fig. 2. PSD curves with sand incorporation. (a) with 36% sand incorporation ratio of Group 2, (b) with different sand incorporations of Group 3.

3
L. Liu et al. Construction and Building Materials 336 (2022) 127544

Table 2
Cement composition.
Oxide content (%) SiO2 Al2O3 CaO Fe2O3 MgO Na2O SO3 K2O Loss of ignition

Cement 21.7 7.5 54.7 2.9 1.7 0.5 3.5 0.4 4.4

Table 3
Mix proportion design of Group 1.
Sand fraction/ Sand diameter Clay water Clay cement
S0 (mm) content/w0 content/c0

0% —— 77.0% 15.0%

20% 0.075–0.25 77.0% 15%


0.25–0.5
0.5–1.0
1.0–2.0

40% 0.075–0.25 77.0% 15%


0.25–0.5
0.5–1.0
1.0–2.0

60% 0.075–0.25 77.0% 15%


0.25–0.5
0.5–1.0
Fig. 3. UCS of CBS when sand incorporation (Group 1) at same w0 (87%).
1.0–2.0

80% 0.075–0.25 77.0% 15%


inclusion and UCS depends on the cementitious binder.
0.25–0.5 On the contrary, when keeping the total water content w1 constant
0.5–1.0 (suggesting the same cement/water ratio), the incorporation of sand
1.0–2.0 particle decreases CBS’s strength (e.g., Group 3) in Fig. 5., i.e., the 28-
day strength decreases from 0.82 MPa to 0.60 MPa when adding 30% of
sand particles. This finding is consistent with the observation by Chian
Table 4 and Bi [23], who also conducted experiments on cement-stabilized
Mix proportion design of Group 4. Singapore marine clay by replacing 0%, 20%, and 40% sand, and
Mix/soil- Original clay Sand c1 w1 found that the strength with 40% sand was the lowest. Overall, if
sand (g) (g) keeping the same w1 and c1 in CBS, the strength decreases with sand
1–0 1014 0 20% 86.6% content due to more water in the soil matrix.
8–2 812 200 20%/15%/ 71.2%
10%
7–3 710 300 20% 66.6%/63.6%/ 3.2. Sand fraction function in the cement-based stabilized soils
60.6%
6–4 608 400 20%/15%/ 51.9%
From the experimental observation above, sand introduction into the
10%
5–5 507 500 20%/15%/ 44.3%
CBS by the method of Group 1 and Group 2 has no influence on the
10% UCS, whlist the addition of sand by Group 3 reduces the strength. As
4–6 405 600 20%/15%/ 36.6% shown in Fig. 6, there are two ways of adding sand to CBS, leading to the
10% questionable definition of water content: (1) mixing sand in cement-
3–7 304 700 20%/15%/ 30.0%
based stabilized clay where water and cement incorporation ratio (e.
10%
2–8 203 800 20%/15%/ 25.3% g., c0 and w0, respectively) were the same as those of the original clay.
10% This method was adopted in Group 1 and Group 2 in this investigation.
(2) Replacing part of the original clay with sand, where the cement
incorporation ratio and water content (e.g., c1 and w1, respectively)
To further clarify the function of sand incorporation on the CBS’s were the same as those of the original clay, which was adopted in Group
strength enhancement, the UCS with five different gradations (e.g., 3. It is well known that clay minerals within the soil with a large specific
Group 2) but the same proportions (e.g., 36%) of sand at different w0 (e. surface area can adsorb an abundance of water, while there is almost no
g., 60.0%, 70.0%, and 80.0%) and referential specimens (original clay water absorbed on the sand and silt surface. As demonstrated in the
without any sand particles) are summarized in Fig. 4. In general, when literatures and in this study, the UCS greatly depends on w0 or essentially
w0 remains constant, the UCS with different sand fractions remains free water content [15], which equals the total water minus that
relatively constant. Furthermore, the UCS with and without sand adsorbed by clay minerals. When sand particles are embedded in the
decreased with increasing w0 increments. For example, when w0 is 60%, CBS, the free water content remains constant by the first method;
the UCS values with sand are approximately 0.81 MPa with deviations however, the free water content can greatly increase with the second
< 6.5% after 7 days of curing, while after 28 days of curing, the UCS method. In other words, the microstructure and void porosity in the
values are approximately 1.43 MPa with deviations < 7.0%. The same is matrix of the water-clay mineral-binder do not change by the first
also true when w0 = 70.0 and 80.0%. From the above observations, it method, whereas they will be altered by the second method owing to
can be concluded that sand added to clay slurry with different fractions more water. Hence, it can also imagine that the stabilization nature is
and gradations has a slight influence on the UCS as long as the w0 and c0 the matrix of the water-clay mineral-binder, and silt/sand fraction is just
in the paste are the same, indicating that the sand only acts as a hard embedded in this matrix.

4
L. Liu et al. Construction and Building Materials 336 (2022) 127544

Fig. 4. UCS of CBS with 5 kinds of sand gradations (Group 2) at different w0. (a) t = 7 days, (b) t = 28 days.

SEM observation was used to reveal the microstructures of the CBS


with sand (specimen #1 in Group 2) and confirm the above assumption,
as shown in Fig. 7. Fig. 7(a) highlights the distribution of sand in the CBS
and suggests that the sand particles are randomly distributed and
embedded in the clay and hydration-product matrix. Fig. 7(b-d) further
show the interface of sand and stabilized clay matrix, implying that sand
particles do not take part in the hydration and the pozzolanic reaction
between the hydrated products and clay minerals. Note that sand and
silt particles are only embedded in a stabilized clay matrix where the
cement hydration products (e.g., CSH) adhere to the sand surface and
link sand particles to form a stabilized paste. As mentioned above, it is
confirmed that the sand/silt particles should only work as hard and inert
inclusions.

3.3. Modified binder/water ratio considering incorporation nature

The principle of the binder/water ratio was first proposed to predict


the strength of concrete by Lyse [39] via a large number of experimental
results. The UCS can be described by the mass ratio of binder/water (mc/
Fig. 5. UCS when sand incorporation (Group 3) at the same w1 (87%) and mw), as shown below:
c1 (15%).
mc
UCS = a +b (1)
mw

where a and b are experimental parameters correlated with the

Fig. 6. Two types of replacement models considering sand incorporation.

5
L. Liu et al. Construction and Building Materials 336 (2022) 127544

Fig. 7. SEM figures of CBS when sand incorporation (#1 of Group 2).

cementation strength and gradation of the coarse aggregate, respec­ the clay fraction to water. Thus, the strength prediction can be rewritten
tively. This shows that the strength depends linearly on the binder/ as follows:
water ratio. Hereinafter, Horpibulsuk et al. [12] introduced this ratio to ( )
mc mf
predict the strength of the CBS. Note that, for concrete, aggregates are UCS = K exp + l = KG + l (3)
mw mw
not involved in hydration process and only work as a skeleton or an
inclusion, in which all the water is free to participate in binder hydra­
where l is an empirical parameter, which is always less than zero
tion. However, in CBS, the water is polymorphic, including free water
because the threshold content of cement incorporation. As Horpibulsuk
and absorption water, where free water fully participates in hydration,
et al. [12] mentioned, when the cement content is less than 5%, the
and absorption water absorbed by clay minerals, is just partially
addition has almost no contribution on the strength CBS strength gen­
involved in this reaction [15,16]. As mentioned above, the CBS’s
eration K is the slope of the UCS versus the modified indicator,
strength does not benefit from the sand and silt incorporation because
depending on the reaction course between hydration products and clay
there is no water on the surface of the silt/sand particles, and sand/silt
minerals. Additionally, with this expression, the threshold cement con­
particles do not take part in the hydration process. It is essential to
tent (i.e., the lowest modified binder/water ratio) can be back analyzed
modify the binder/water ratio to address the water and binder functions
when UCS equal to zero.
in this type of mixture and provide a universal expression to better
The UCS results from the experiments in this study were replotted
characterize its strength performance considering the complex compo­
versus the binder/water ratio and the modified binder/water ratio, as
sitions (e.g., sand, silt, and clay fraction) in soils under depositional
shown in Fig. 8. In the relationship between UCS and the traditional
regions. A new indicator to reveal water and binder functions reflecting
indicator (binder/water ratio), a linear relation can be found for binder/
the incorporation nature to express the UCS behavior is necessary. After
water ratios smaller than 0.3, after 7 days of curing, but with significant
many attempts, a new indicator G considering the incorporation to
deviation at higher binder/water ratios after 28 days of curing, as shown
better reflect the interaction among the soil, cement, and water was
in Fig. 8a and b. On the contrary, when the UCS was replotted versus the
proposed as follows:
modified binder/water ratio, a nearly perfect linear relationship with
( )
mc mf slopes of 5.19 and 5.21 can be found after 7 and 28 days of curing in
G= exp (2)
mw mw Fig. 8c and d, respectively.
In addition to the experimental studies, a detailed literature review
where mc/mw is the binder/water ratio, and mf/mw is the mass ratio of and data analysis were used to further verify the applicability and

6
L. Liu et al. Construction and Building Materials 336 (2022) 127544

Fig. 8. Relationship of the UCS of CBS with binder/water ratio, and modified binder/water ratio.

effectiveness of the newly proposed modified binder/water ratio. Feng simple and unique linear tendency with a minimal R2 = 0.81, at any
et al. [40] conducted a series of experiments on soils with PSD curves, as curing period, as shown in Fig. 11b, d, and f.
shown in Fig. 9, and the corresponding clay fractions ranged from 12.9%
to 45.2%. The data were re-analyzed, and the relationships between the 4. Discussion
UCS, the binder/water ratio, and the modified binder/water ratio are
shown in Fig. 10. Note that for the soil with the same clay fraction, the The solid particles within soils mainly include primary and second­
UCS increases almost linearly with the binder/water ratio; however, ary minerals. The chemical compositions of primary minerals remain
with the modified binder/water ratio, all UCS can be better fitted to a unchanged for the coarse dimension, weak hydrophilic and steady
straight line with a slope of 1.71. physical/chemical properties, where the main components of the
Furthermore, the data of cement-based stabilized Singapore marine pebble, gravel, sand and silt fraction are primary minerals. When the
clay with the addition of different proportions of sand particles (e.g., parent rock suffers from bio/chemical weathering, such as oxidation,
0%, 20%, and 40%) [24] were also collected to validate the applicability hydration, hydrolysis, and dissolution, secondary minerals are formed in
of the strength prediction using the modified binder/water ratio. the clay fraction. The clay fraction has properties distinct from those of
Fig. 11a, c, and e show the relationship between UCS versus binder/ sand and silt fraction. Clay minerals originate from chemical weathering
water ratios with poor linear correlations (e.g., maximum R2 = 0.73) at of primary minerals (e.g., feldspar, mica, and other silicate minerals)
different curing periods (e.g., 3, 28, and 91 d). However, the relationship forming flake or chain crystal lattices, which have strong hydrophilicity
between the UCS versus modified binder/water ratio can achieve a and colloidal properties. The clay mineral can also form ultrafine
colloidal particles with a larger specific surface area and special surface
properties. Overall, the sand and silt fractions have similar properties
and are essentially different from the clay fraction.
In summary, the sand and silt particles just work as hard inclusions
for CBS. It is feasible to adjust the sand fraction in CBS by sieving, but it
is difficult to change the silt particles owing to the dimension limitation
of the sieve opening. Therefore, the gradation of soil in this study was
changed by adding sand to reflect the non-clay fraction. Note that
adding sand and silt fraction can change in the friction angle of the CBS
for the blocking behavior within the shearing band [41,42].
As mentioned above, the modified binder/water ratio works more
efficiently than the traditional binder/water ratio for predicting the
UCS, because it can reflect two important factors dominating the UCS, i.
e., the binder/water ratio and clay fraction/water ratio. The clay frac­
tion/water ratio was introduced to reflect free water, excluding the
absorbed water by clay minerals.
As summarized in Table 5, the binder/water ratio has been studied
by many researchers, and many models have been proposed to predict
the UCS of concrete, mortar, and CBS [1,12,15,17,18,39,43-47]. In
concrete and mortar, coarse aggregates and sands act as skeletons to
Fig. 9. PSD of soils with different clay fractions.

7
L. Liu et al. Construction and Building Materials 336 (2022) 127544

y x- R2
y x- R2

Fig. 10. UCS of CBS with six different clay contents:(a) UCS vs binder/water ratio, (b) UCS vs modified binder/water ratio.

y x- R2= y x- R2=

y x- R2=

y x R2=

y x- R2=

y x- R2=

Fig. 11. Relationship of UCS of cemented Singapore marine clay with 0%, 20%, 40% sand with binder/water ratio, and modified binder/water ratio.

bear loads, which do not participate in cement hydration. However, for The primary hydration products harden into high-strength additives and
CBS, the strength development mostly depends on the hydration prod­ bonds connecting the solid phases, while the secondary products further
ucts of the binder and pozzolanic reactions between the productions and change the bond strength between the clay minerals.
clay minerals, where the sand/silt fractions in the matrix also work as It is necessary to understand the effect of clay minerals on calculating
filling inclusions. The hydration reactions occur immediately once the values of parameters K and G because clay minerals have a more specific
binder comes into contact with water within the soil, and generate surface area and hydrophilicity. The UCS and mineralogy of 12 types of
porous hydrated calcium silicates (C2SHx, C3S2Hx), hydrated calcium CBSs were collected, the properties of CBSs were shown in Table 6. And
aluminates (C3AHx, C4AHx), and hydrated lime (Ca(OH)2). The calcium the relationship between them is shown in Fig. 12 and Table 7
hydroxyl is dissolved in the pore water, forming a highly alkaline [16,24,33,40,48-51]. The relationship between the UCS and the tradi­
environment. In the following secondary pozzolanic reaction, hydrous tional binder/water ratio is also presented in Fig. 12(a). This suggests a
silica and alumina in the platy clay minerals react with the calcium ions distinct strength after the same binder incorporation ratio for the soils’
derived from cement hydration to form secondary pozzolanic products. mineralogy. For the soil in the same region, the UCS increases almost

8
L. Liu et al. Construction and Building Materials 336 (2022) 127544

Table 5 linearly with the binder/water ratio, for which the slope varies from 2.7
Empirical strength prediction equation of cement-based stabilized soil. to 8.45. Furthermore, using the new indicator (modified binder/water
Reference Equation Notations ratio), the strength trends of different soils can be better merged by two
linear lines: the UCS of CBS from Lianyungang, Hangzhou, and Shanghai
Abrams, 1918 k1 k1, k2- experimental constants; W/C-
UCS = W/C water-binder ratio improves linearly with the same slope of 2.56, and the UCS for other
k2
Lyse, 1932; C a, b- constants related to cementation soils with a slope of 4.36, as shown in Fig. 12(b). Evidently, the modified
UCS = a⋅ +b
W strength and the gradation of coarse binder/water ratio can better describe the strength with different
aggregate; C/W- binder-water ratio mineralogy and provide reliable guidance for strength characterization.
Gallavresi, 1992 q0 q0, n-experimentally fitted
UCS = In Table 5, the major clay minerals of Shanghai and Lianyungang clays
(W/C)n coefficients, recommended a range of
5000–10,000 kPa and 1.5–3.0.
are illite/smectite or smectite/chlorite, while those from Dalian, Tianjin,
Horpibulsuk et A A, B- empirical constants; W/C- and Shenzhen are kaolinite and illite as well as Singapore marine clay
UCS =
al., 2011 (W/C)B water-binder ratio [52] for the matter source during sedimentation. Note that the strength
Lee et al., 2005 em(S/C) a, m, and n- experimentally fitted of the CBS, whose original clay minerals are dominated by kaolinite and
UCS = a
(W/C)n values; S/C- soil-binder ratio; W/C- illite, is higher than that dominated by smectite. This result agrees well
water-binder ratio
Chian and Bi, UCS = t- curing period; a, b, μ and σ- fitting
with the previous findings by Croft [53], which stated that cement-based
2020 a constants; WL- liquid limit cementitious materials work more efficiently in kaolinite and illite-rich
[a⋅ln(t) + b]
b(W/WL )/C soils for inert chemical activity. Additionally, the secondary pozzolanic
reaction proceeded with a long curing period, indicating a time-
dependent slope increase.
Table 6
The properties of CBS collected from literatures. 5. Conclusions
Original soil Liquid limit (%) Clay fraction (%) w1 (%) c1 (%)

Hangzhou 73% 38.7 60, 75, 100 5–20


The effect of the inert fraction (sand and silt fraction, in this case,
Shanghai 35, 39.4, 47.1 40, 41, 49 70 0–20 sand fraction adjustment was adopted) on the UCS of the CBS was
Wuxi 75 20 108 7.3–102 investigated. The stabilization nature of the cement-based stabilized
Guangzhou 62 30 97.4 6.7–94.4 soils was addressed, and a new indicator (i.e., modified binder/water
Shenzhen 73 32 92.6 6.5–90.4
ratio), was proposed to characterize the UCS performance. Additionally,
Dalian 53.1 40 58.2–164.6 3.8–24.2
Tianjin-1 / 11 80–140 12.4–68.6 literature data were collected to validate the applicability of the indi­
Tianjin-2 / 35.6 80 8–18 cator. The main conclusions drawn are as follows.
Baimahu 66.1 31.6 165.3–214.8 10.1–75.6
Huaian 18–32.3 11.8–69.5 93–99 5–15
1. The sand/silt fraction acts as a hard inclusion in cement-based sta­
Singapore 87 68 121.4–185.7 5–6.5
bilized soils for inert activity. The sizes, proportions, and gradations
of sand added to cemented soil have almost no influence on the UCS
as long as the w0 and c0 in the paste are the same. When replacing

Fig. 12. UCS development of CBS with binder/water and modified binder/water ratio.

Table 7
Clay minerals of different soils in literatures (%).
Soil illite kaolinite Chlorite smectite Illite/smectite Smectite/chlorite

Lianyungang 29 13 14 44
Shanghai 40.5 7.3 52.2
Dalian 63 10.9 11.9 14.2
Tianjin 44.7 12.8 12.8 29.7
Shenzhen 10.1 42 15 32.9

9
L. Liu et al. Construction and Building Materials 336 (2022) 127544

clay with sand at the same water mass, the sand addition weakens the [12] S. Horpibulsuk, N. Miura, T.S. Nagaraj, Assessment of strength development in
cement-admixed high water content clays with Abrams’ law as a basis,
UCS for more free water for the hydration process.
Geotechnique 53 (2003) 439–444, https://doi.org/10.1680/geot.2003.53.4.439.
2. By revealing the nature of sand/silt incorporation within the CBS [13] T.S. Nagaraj, P.P. Yaligar, N. Miura, A. Yamadera, Predicting strength development
paste, the modified binder/water ratio combining the binder/water by cement admixture based on water content, Grouting Deep Mix, Proc. Conf.
ratio and clay fraction/original clay water content ratio was pro­ Tokyo 1 (7) (1996) 1–9.
[14] N.C. Consoli, P.M. Vaz Ferreira, C.S. Tang, S.F. Veloso Marques, L. Festugato, M.
posed as a powerful indicator to characterize the UCS performance. B. Corte, A unique relationship determining strength of silty/clayey soils – Portland
3. A linear relationship between UCS and the new indicator was cement mixes, Soils Found 56 (2016) 1082–1088, https://doi.org/10.1016/j.
established, in which the slope depends on the curing period and the sandf.2016.11.011.
[15] L. Liu, A. Zhou, Y. Deng, Y. Cui, Z. Yu, C. Yu, Strength performance of cement/slag-
hydration/pozzolanic reaction. The data from the literature were based stabilized soft clays, Constr. Build. Mater. 211 (2019) 909–918, https://doi.
collected to validate the indicator, suggesting that the CBS’s UCS org/10.1016/j.conbuildmat.2019.03.256.
with similar original clay minerals can be unified. [16] W. Zhu, C.L. Zhang, A.C.F. Chiu, Soil-water transfer mechanism for solidified
dredged materials, J. Geotech. Geoenviron. Eng. 133 (2007) 588–598, https://doi.
org/10.1061/(ASCE)1090-0241(2007)133:5(588).
[17] S. Horpibulsuk, R. Rachan, A. Suddeepong, Assessment of strength development in
blended cement admixed Bangkok clay, Constr. Build. Mater. 25 (2011)
CRediT authorship contribution statement 1521–1531, https://doi.org/10.1016/j.conbuildmat.2010.08.006.
[18] N. Miura, S. Horpibulsuk, T.S. Nagaraj, Engineering behavior of cement stabilized
clay at high water content, Soils Found 41 (5) (2001) 33–45.
Li Liu: Data curation, Formal analysis, Writing – original draft, [19] Suksun Horpibulsuk, Norihiko Miura, T.S. Nagaraj, Clay–water/cement ratio
Investigation. Tingting Deng: Resources, Data curation, Investigation, identity for cement admixed soft clays, J. Geotech. Geoenviron. Eng. 131 (2005)
987–1003, https://doi.org/10.1061/(ASCE)1090-0241(2005)131.
Validation. Yongfeng Deng: Conceptualization, Methodology, Super­
[20] O. Taki, D.S. Yang, Soil-cement mixed wall technique, in: Geotech. Eng. Congr.,
vision, Writing – review & editing, Funding acquisition. Liangtong ASCE, 1991: pp. 298–309.
Zhan: Writing – review & editing, Methodology. Suksun Horpibulsuk: [21] F. Bell, Engineering treatment of soils, CRC Press, 1991.
[22] D.T. Bergado, L.R. Anderson, N. Miura, A.S. Balasubramaniam, Soft ground
Writing – review & editing, Formal analysis. Qiong Wang: Writing –
improvement in lowland and other environments., ASCE Press, New York, NY,
review & editing. USA: ASCE Press, 1996.
[23] S.C. Chian, J. Bi, Influence of grain size gradation of sand impurities on strength
behaviour of cement-treated clay, Acta Geotech. 16 (4) (2021) 1127–1145.
Declaration of Competing Interest [24] S.C. Chian, Y.Q. Chim, J.W. Wong, Influence of sand impurities in cement-treated
clays, Geotechnique 67 (2017) 31–41, https://doi.org/10.1680/jgeot.15.P.179.
[25] W. He, X. Shen, Mechanical behavior of soil-sand-cement admixture, Rock Soil
The authors declare that they have no known competing financial Mech. 32 (2011) 392–396, https://doi.org/10.16285/j.rsm.2011.s1.045.
interests or personal relationships that could have appeared to influence [26] Y. Wang, Research on the Mechanism of Shallow Sliding of Xiashu Soil Slope Under
the Condition of Wetting-Drying Cycles, Jiangsu University of Science and
the work reported in this paper. Technology, Zhenjiang, China, 2020.
[27] G. Cai, Experimental and Application Studies on Soft Soil Carbonated and
Stabilized by Reactive Magnesia (Doctoral thesis), Southeast University, Nanjing,
Acknowledgement
China, 2017.
[28] J. Hao, Study on the Stabilization Method of Silts for Roadbed and Slope in
This study was supported by the National Key R&D Program of China Nantong Area (Master’s thesis), Southeast University, Nanjing, China, 2006.
[29] H. Liu, X. Lu, A. Zhu, K. Wu, Experimental study on physical properties of raw
(Grant No. 2019YFC1806004), the National Natural Science Foundation
material of Nantong silt fired perforated brick, New Build. Mater. 53 (2011) 42–44.
of China (Grant No. 51878159) and Fundamental Research Funds for [30] Z. Wu, Depositional Environment and Geotechnical Properties for the Upper
Central Universities and Postgraduate Research & Practice Innovation Shanghai Clays (Doctoral thesis), Shanghai Jiaotong University, Shanghai, China,
Program of Jiangsu Province (Grant No. KYCX17-0131). 2016.
[31] J. Yin, Y. Miao, Laboratory tests on intrinsic compression behavior of
reconstitution clays, J. Eng. Geol. 20 (2012) 403–409.
References [32] H. Xu, J. Ma, Z. Hua, P. Zhao, Study on empirical formula of cement-soil
compressive strength, Site Invest. Sci. Technol. 1 (2009) 3–6.
[33] D. Che, Laboratory Investigation on the Behavior of Cement Mixed Soft Clayey Soil
[1] N.C. Consoli, D. Foppa, L. Festugato, K.S. Heineck, Key parameters for strength
of Shanghai (Master’s thesis), Shanghai Jiao Tong University, Shanghai, China,
control of artificially cemented soils, J. Geotech. Geoenviron. Eng. 133 (2007)
2012.
197–205, https://doi.org/10.1061/(asce)1090-0241(2007)133:2(197).
[34] ASTM, Standard Test Methods for Liquid Limit , Plastic Limit , and Plasticity Index
[2] J. Yang, M. Dong, T. Sun, M. Wang, Forecast formula for strength of cement-treated
of Soils 1, D4318. 04 (2010) 1–14. https://doi.org/10.1520/D4318-10.
clay, Soils Found 59 (2019) 920–929, https://doi.org/10.1016/j.
[35] ASTM, Standard Practice for Classification of Soils for Engineering Purposes
sandf.2019.03.006.
(Unified Soil Classification System), D2487-11. D5521-05 (2010) 1–5. https://doi.
[3] S. Sasanian, T.A. Newson, Basic parameters governing the behaviour of cement-
org/10.1520/D2487-11.
treated clays, Soils Found 54 (2014) 209–224, https://doi.org/10.1016/j.
[36] ASTM, Standard Test Method for Particle-Size Analysis of Soils, D422. D422-63
sandf.2014.02.011.
(2007) 1–8. https://doi.org/West Conshohocken, PA.
[4] S. Horpibulsk, R. Rachan, A. Suddeepong, A. Chinkulkijniwat, Strength
[37] ASTM, Standard Test Methods for Specific Gravity of Soil Solids by Water
development in cement admixed bangkok clay: Laboratory and field investigations,
Pycnometer, D854. 2458000 (2010) 1–7. https://doi.org/10.1520/D0854-10.2.
Soils Found 51 (2011) 239–251, https://doi.org/10.3208/sandf.51.239.
[38] ASTM, Standard Test Method for Unconfined Compressive Strength of Cohesive
[5] A.M. Santoso, K.K. Phoon, T.S. Tan, Estimating strength of stabilized dredged fill
Soil 1, ASTM Int. (2013) 1–7. https://doi.org/10.1520/D2166.
using multivariate normal model, J. Geotech. Geoenviron. Eng. 139 (2013)
[39] I. Lyse, Tests on consistency and strength of concrete having constant water
1944–1953, https://doi.org/10.1061/(ASCE)GT.1943-5606.0000910.
content, ASTM Proc. 32 (1932) 629–636. http://preserve.lehigh.edu/engr
[6] T. Tsuchida, Y.X. Tang, Estimation of compressive strength of cement-treated
-civil-environmental-fritz-lab-.
marine clays with different initial water contents, Soils Found 55 (2015) 359–374,
[40] Z. Feng, W. Zhu, C. Zhang, T. Ohki, Influence of clay content on mechanical
https://doi.org/10.1016/j.sandf.2015.02.011.
properties of solidified silt, Yanshilixue Yu Gongcheng Xuebao/Chinese, J. Rock
[7] L. David Suits, T.C. Sheahan, T.-S. Tan, T.-L. Goh, K.Y. Yong, Properties of
Mech. Eng. 26 (2007) 3052–3057.
Singapore marine clays improved by cement mixing, Geotech. Test. J. 25 (4)
[41] M. Xu, L. Liu, Y. Deng, A. Zhou, S. Gu, J. Ding, Influence of sand incorporation on
(2002) 9061.
unconfined compression strength of cement-based stabilized soft clay, Soils Found.
[8] G. Kang, T. Tsuchida, A.M.R.G. Athapaththu, Engineering behavior of cement-
61 (2021) 1132–1141, https://doi.org/10.1016/j.sandf.2021.06.008.
treated marine dredged clay during early and later stages of curing, Eng. Geol. 209
[42] X. Liu, X.Q. Fan, B.N. Hong, Experimental study of triaxial test of soils stabilized by
(2016) 163–174, https://doi.org/10.1016/j.enggeo.2016.05.008.
cement mortar, Yantu Lixue/Rock Soil Mech. 32 (2011) 1676–1682, https://doi.
[9] Y. Yi, L. Gu, S. Liu, Microstructural and mechanical properties of marine soft clay
org/10.16285/j.rsm.2011.06.050.
stabilized by lime-activated ground granulated blastfurnace slag, Appl. Clay Sci.
[43] F. Gallavresi, Grouting improvement of foundation soils, in: Soil Improv.
103 (2015) 71–76, https://doi.org/10.1016/j.clay.2014.11.005.
Geosynth., New orleans, LA, 1992: pp. 1–38.
[10] Y. Yi, X. Zheng, S. Liu, A. Al-Tabbaa, Comparison of reactive magnesia- and carbide
[44] D.A. Abrams, Design of concrete mixtures, Struct. Mater. Res. Lab. (1918), https://
slag-activated ground granulated blastfurnace slag and Portland cement for
doi.org/10.1093/schbul/sbl056.
stabilisation of a natural soil, Appl. Clay Sci. 111 (2015) 21–26, https://doi.org/
[45] S. Horpibulsuk, N. Miura, T.S. Nagaraj, Clay–water∕cement ratio identity for
10.1016/j.clay.2015.03.023.
cement admixed soft clays, J. Geotech. Geoenviron. Eng. 131 (2005) 187–192,
[11] K. Uddin, A.S. Balasubramaniam, D.T. Bergado, Engineering behavior of cement-
https://doi.org/10.1061/(asce)1090-0241(2005)131:2(187).
treated Bangkok soft clay, Geotech. Eng. 28 (1997) 89–119.

10
L. Liu et al. Construction and Building Materials 336 (2022) 127544

[46] F.-H. Lee, Y. Lee, S.-H. Chew, K.-Y. Yong, Strength and modulus of marine clay- [50] J. Ding, Mechanical Properties and Microstructure Characteristics of Solidified
cement mixes, J. Geotech. Geoenviron. Eng. 131 (2005) 178–186, https://doi.org/ Dredged Clays at High Water Content (Doctoral thesis), Southeast University,
10.1061/(asce)1090-0241(2005)131:2(178). Nanjing, China, 2011.
[47] N.C. Consoli, D.A. Rosa, R.C. Cruz, A.D. Rosa, Water content, porosity and cement [51] B. Zhan, Experimental Study on Mechanical Behavior of Cement-stabilized
content as parameters controlling strength of artificially cemented silty soil, Eng. Dredged Seabed Clays in Dalian Bay (Master’s thesis), Southeast University,
Geol. 122 (2011) 328–333, https://doi.org/10.1016/j.enggeo.2011.05.017. Nanjing, China, 2018.
[48] L. Yan, Experimental Study on Solidification Characteristics of Dredging Sludge [52] A.H.M. Kamruzzaman, S.H. Chew, F.H. Lee, Structuration and destructuration
(Master’s thesis), Zhejiang University of technology, Hangzhou, China, 2013. behavior of cement-treated Singapore marine clay, J. Geotech. Geoenviron. Eng.
[49] J. Luan, Analysis of Laboratory Test Result of Cemented Soil of High Moisture 135 (2009) 573–589, https://doi.org/10.1061/(ASCE)1090-0241(2009)135:4
Content (Master’s thesis), Tianjin University, Tianjin, China, 2005. (573).
[53] J.B. Croft, The influence of soil mineralogical on cement stabilization,
Géotechnique 17 (1967) 119–135.

11

You might also like