Clay–Water/Cement Ratio Identity for Cement Admixed Soft

Clays
Suksun Horpibulsuk1; Norihiko Miura2; and T. S. Nagaraj3

Abstract: The in situ deep mixing technique is an established means of enhancing bearing capacity and reducing settlement by placing
columnar inclusions in soft ground. Developments in equipment and field techniques to implement this method have surpassed the basic
understanding of strength development in soft clays admixed with cementing agents at water contents close to their liquid limit or higher.
Laboratory tests on soft clays admixed with cementing agents are a prerequisite for engineering decisions on the proper mix proportions
for intended field strength. This paper presents and analyzes test results on cement admixed soft clays within the basic framework of
induced cementation of soft clays. It is possible to propose a clay–water/cement ratio formulation that enables calculation of cement
contents consistent with changes in in situ water content.
DOI: 10.1061/(ASCE)1090-0241(2005)131:2(187)
CE Database subject headings: Mixing; Marine clays; Soil strength; Deformation; Bearing capacity; Soft soils.

Introduction method was initiated and implemented mainly by the Port and
Harbour Research Institute, Tokyo. Developments in the plant and
In recent years, the scale of design and construction of infrastruc- machinery and associated field techniques have stayed ahead of
ture in natural soft ground has increased tremendously as a result our basic understanding of strength development in high water
of extensive urbanization and industrialization. Soil strengthening content clays admixed with cementing agents (e.g., Nagaraj et al.
is required in many land reclamation projects. The desired prop- 1996, 1998; Yamadera et al. 1998; Miura et al. 2001). Kamalud-
erties of the improved soil are increased strength, reduced com- din and Balasubramaniam (1995); Uddin et al. (1997) and Yin
pressibility, and appropriate permeability to solve stability, settle- and Lai (1998), among others, have focused on the strength and
ment, ground water, and other environmental-related problems. deformation characteristics of cemented soft clays.
Soft clay formations, especially those with high in situ water Despite the availability of information on the factors control-
contents, are susceptible to large settlements and possess low ling the strength development of in situ deep mixed soft clays,
shear strength unless they are naturally cemented. Precompres- additional experiments are necessary to identify the dominant pa-
sion of such deposits with geodrains can prevent this large settle- rameters controlling strength development and to arrive at a
ment and thus enhance shear strength. But this mode of attacking means of combining them to formulate appropriate phenomeno-
the problem often requires more time than is practically available. logical models. Such models would facilitate engineering deci-
An alternative to this is cementation of the soft clay with supple- sions on the proportions of cementing agents and on curing peri-
mentary cementing materials such as lime and cement. Such ad- ods for specific target strengths of cement-admixed soft clays.
mixtures impart resistance to compression and develop adequate This paper reports on an attempt at meeting these objectives. A
shear strength during time periods typically much less than those brief discussion of the basic characteristics of the materials pre-
required by precompression. cedes details of the experiment.
In situ deep mixing was introduced in 1975 as a viable means
to place columnar inclusions in the soil (Broms and Boman 1975;
Okumura and Terashi 1975). Since this seminal work there have Material Characteristics of Cement Admixed Soft
been numerous innovations focusing on implementation technolo- Clays
gies. In Japan, research and development dealing with this
While cement and clays enter into physicochemical interactions
1
Assistant Professor, School of Civil Engineering, Suranaree Univ. of with fluids, fine and coarse aggregates are noninteractive with
Technology, Nakhon-Ratchasima, Thailand. water. There is little or no time-related change in the physico-
2
Professor of Civil Engineering, Dept. of Civil Engineering, Saga chemical potential of clay–water interactions because clay water
Univ., Saga, Japan. mixes do not by themselves harden into a nonparticulate material.
3
Adjunct Professor, University of Massachusetts, Amherst, MA and In contrast, fresh mixes of cement based composites transform
Distinguished Professor, RV College of Engineering, Bangalore, India. over time into a coherent mass of nonparticulate material of
Note. Discussion open until July 1, 2005. Separate discussions must higher strength. The coarse constituents contribute to the dimen-
be submitted for individual papers. To extend the closing date by one
sional stability of the mix.
month, a written request must be filed with the ASCE Managing Editor.
The manuscript for this paper was submitted for review and possible Since cement is the only reacting material in cement-based
publication on June 4, 2002; approved on May 17, 2004. This paper is composites, the hardened cement paste provides continuity to the
part of the Journal of Geotechnical and Geoenvironmental Engineer- mix structure, with the coarse constituents remaining embedded.
ing, Vol. 131, No. 2, February 1, 2005. ©ASCE, ISSN 1090-0241/2005/ In soft clays, which already contain a certain degree of moisture,
2-187–192/$25.00. microfabrics form due to interaction with water. If a cementing

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / FEBRUARY 2005 / 187

Downloaded 16 Jun 2012 to 180.211.192.67. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
agent is admixed with such a system, the strength will increase
with time due to the transformation of the mix into a nonparticu-
late state; that is a combination of two materials interacting with
the pore fluid. The experiments reported here were carried out to
determine the parameters responsible for the strength and defor-
mation characteristics of soft clays admixed with cementing
agents.

Experimental Investigation

Ariake clay samples collected from Fukodomi town, Saga, Japan

were used in these experiments. After an excavator removed
weathered surface soil to a depth of 2 m, the representative clay
Fig. 1. Typical compressibility characteristics of Ariake clay
samples were collected by an excavator and transferred into plas-
admixed with cement with a curing period of 7 days 共wc / C = 15兲
tic containers. The Atterberg limits of this clay are wL = 120% and
w P = 57%, with specific gravity, G, pH of the pore fluid, and so-
dium chloride concentration values of 2.61, 8.8, and 3.2 g / L, Test Results
respectively.
The natural water content of this clay, 130%, was higher than Compressibility Characteristics
its liquid limit water content and it had a sensitivity value of 8.
The clay paste was passed through a 2 mm sieve for removal of Figs. 1 and 2 show typical compression paths of cement admixed
soft clays covering the clay water content, wc, range of 120–250%
shell pieces and was adjusted to water contents to reach liquidity
The cement content, C, was in the range of 8–33%, resulting in
indices ranging up to 3.0. This intentional increase in water con-
wc / C values of 7.5, 10, and 15. The cement content, C, expressed
tent simulated either the increase that may take place during the in percent, is defined as the ratio of the dry weight of cement to
wet method of dispensing the cement admixture, or the water the dry weight of clay.
content increase that occurs in jet grouting. Fig. 3 depicts typical stress-strain relationships of cemnet ad-
In concrete technology, strength development in cement-based mixed soft clay specimens tested in unconfined compression tests.
composites, in which cement is the only interacting material, is These three sets of specimens differ in their initial wc values
often analyzed using Abrams’ law (Abrams 1918). The law states admixed with cement contents of different percentages, C, such
that for a given set of ingredients, strength development is solely that wc / C = 7.5, 10, and 15.
dependent upon the water-to-cement ratio. Considering this and Triaxial test results are presented in Figs. 4–7. Figs. 4 and 5
the material characteristics of clay and cement discussed in the present plots of typical deviator stress and pore pressure versus
previous section, the clay–water/cement ratio was taken as a com- shear strain during consolidated undrained shearing utilizing con-
bined parameter that incorporates the effects of interactions be- solidation pressures of 100 and 400 kPa, respectively, at wc / C
tween these materials, both expressed in percent with respect to = 15. Typical plots of deviator stress and volumetric strain versus
shear strain for consolidated drained shearing are presented in
the dry weight of clay.
Figs. 6 and 7. Fig. 6 shows the results for consolidation pressure
Using the clay with water contents corresponding to levels of
of 100 kPa and a clay–water/cement ratio of 15, while Fig. 7 is
liquidity indices stated above along with the required quantity of for a consolidation pressure of 400 kPa and a clay–water/cement
cement, clay–water/cement ratios of 7.5, 10, and 15 were ob- ratio of 7.5. Undrained and drained q-p stress paths for a wc / C
tained by thorough mixing for 10 min to ensure a uniform disper- = 15 are shown in Fig. 8.
sion of the cementing agent. The uniform paste was transferred to
split cylinder molds 50 mm in diameter and 100 mm in height, as
well as to consolidometer rings, taking care to prevent any air
entrapment. After curing for 1 day, the cylindrical specimens
were dismantled and these specimens and consolidometer rings
were wrapped in vinyl bags and stored at constant temperature
and humidity conditions for different curing periods.
Unconfined compression tests and consolidated-undrained and
consolidated-drained triaxial compression tests were carried out
on specimens cured for different periods. A backpressure of
190 kPa was maintained to ensure saturation at all levels of test-
ing. The rate of loading for the unconfined compression tests was
1 mm/ min. The effective confining pressure for isotropically con-
solidated drained tests was 50, 100, 200, and 400 kPa. The rates
of displacement were fixed at 0.0075 and 0.0025 mm/ min for
undrained and drained tests, respectively. The axial strain mea-
surements were made using a displacement transducer with re- Fig. 2. Typical compressibility characteristics of Ariake clay
duced end restraint using lubricated end platens. admixed with cement with a curing period of 28 days 共wc / C = 7.5兲

188 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / FEBRUARY 2005

Downloaded 16 Jun 2012 to 180.211.192.67. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
Fig. 3. Stress-strain characteristics of cement admixed soft clay
samples in unconfined compression

Analysis of Test Results

Fig. 5. Stress-strain–pore pressure response of cement admixed soft
An examination of the compressibility characteristics (Figs. 1 and clay during undrained shearing (wc / C = 7.5, ␴⬘ = 400 kPa)
2) reveals that the void ratio of the clay does not exactly corre-
spond to e = wcG for initial levels of the clay water contents indi-
cated. Upon admixing with cement, the clay–water content de- of intersection of two straight lines extended from the linear por-
creases because hydration products increase the solid weight tions on either end of the compression curve plotted as log共1
while the water weight decreases due to utilization of the water by + e兲 against log ␴⬘v (Butterfield 1979; Sridharan et al. 1991).
the cement. Hence the e-log ␴⬘v curves for uncemented clay inter- Table 1 presents the compression indices in the pre- and post-
sect the initial e-log ␴⬘v curves for cement admixed with soft clay. transitional stress levels along with the transitional stress levels. It
The resistance to compression prevails up to a certain stress level is interesting to note that the resistance up to the transitional stress
equal to cementation bond strength at initial water content. Be- level is practically negligible in spite of the initial high water
yond this stress level, cement-admixed clays undergo sudden content of the clay compared to the resistance of posttransitional
compression. This stress level is different from preconsolidation compression indices. It is also noticeable that the transitional
stress since the compression path of a cemented system is above stress level at a specific curing period is of the same order as long
the compression path of an uncemented system devoid of any as the wc / C value is the same, but irrespective of the combination
stress history. This transitional stress can be obtained as the point of initial water content of clay and the cement content. The same

Fig. 4. Stress-strain–pore pressure response of cement admixed soft Fig. 6. Stress-strain–volumetric strain response of cement admixed
clay during undrained shearing (wc / C = 15, ␴⬘ = 100 kPa) soft clay during drained shearing (wc / C = 15, ␴⬘ = 100 kPa)

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / FEBRUARY 2005 / 189

Downloaded 16 Jun 2012 to 180.211.192.67. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
 Table 1. Values of Compression Index and Transitional Stress
Corresponding to Wc / C Initial Clay–Water Content, Curing Time
Initial water Curing time Transitional
Wc / C content (%) (days) CS Cc stress (kPa)
15 120 7 0.021 0.802 725
28 0.018 0.807 820
150 7 0.015 1.050 720
28 0.011 1.122 810
180 7 0.027 1.139 730
28 0.023 1.226 850
250 7 0.030 1.442 720
28 0.031 1.492 900

10 120 7 0.007 0.915 1,900

28 0,010 0.850 2,100
150 7 0.018 0.804 1,900
28 0.020 1.011 2,300
180 7 0.019 1.000 1,700
28 0.021 1.431 2,150
Fig. 7. Stress-strain–volumetric strain response of cement admixed 250 7 0.020 1.603 1,800
soft clay during drained shearing (wc / C = 7.5, ␴⬘ = 400 kPa) 28 0.025 1.559 2,000

7.5 120 7 0.014 —a 2,900

order of increase is noticed in the same range as curing time 28 0.014 —a 3,500
increases. But the specimens with higher clay water content are 150 7 0.012 —a 3,000
stable at higher void ratios and provide higher compression indi- 28 0.013 —a 3,600
ces beyond the transition stress. This is especially true with speci- 250 7 0.014 1.575 2,800
mens at high water contents (of the order of 250%) due to the 28 0.014 1.407 3,400
compression required to sustain the stress level beyond the tran- Note: Cs and Cc are the slope of the 共e ⬃ log ␴⬘v兲 plot at pre- and post-
sition. The compression indices at the posttransition state of the transitional stress, respectively.
clay–cement mixtures having the same initial clay water content a
The data are not enough to obtain the corresponding values.
are of almost the same order of magnitude even if their cement
content differs. It is also clear that the lower the wc / C value, the
greater the enhancement of the transition stress, for the proportion oped will exhibit the same pattern and level. The postpeak re-
of cement is higher for the same range of clay–water content. sponse is also of the same pattern, except for water content of
The unconfined compressive strength of cemented soft clay 250%.
remains constant if wc / C is the same (Fig. 3) but the value in- The shear strength response under consolidated undrained
creases as wc / C increases. The value of unconfined compressive shearing (Figs. 4 and 5) shows that the deviator stress and excess
strength falls into three groups with wc / C as a variable. Within pore pressure versus shear strain follows the same paths as long
one set, the stress-strain response and maximum strength devel- as the wc / C value remains the same for various combinations of
clay wc and C. The maximum deviator stress level depends
mainly on clay–water/cement ratio. Both deviator stress and ex-
cess pore pressure trace the same path for all combinations of
clay–water content and cement content as long as the value of
wc / C is the same. However, the paths are different in the postpeak
stage of shearing. Even when consolidation pressure increases
fourfold (i.e., from 100 to 400 kPa), the maximum deviator
stress remains at the same level (Fig. 5). These consolidation
pressures are well within the range of yield stress levels in K0
compression. But the level of peak deviator stress increased from
500 to 2,000 kPa as the wc / C change from 15 to 7.5.
In the consolidated-drained shear response of specimens (Figs.
6 and 7), as shear strain is increased, the deviator stress and
volumetric strain follow the same paths irrespective of consolida-
tion pressure levels and clay–water/cement ratios; but the volu-
metric strain levels influence the peak deviator stress levels dur-
ing the shearing process. This accounts for the dilation or
contraction of specimens in accordance with the confining pres-
sures. Beyond the peak conditions, the responses due to increase
in shear strain are influenced by the clay–water content. It is
Fig. 8. Typical undrained and drained stress paths along with noticed that the undrained and drained stress paths culminate in a
strength envelope for wc / C = 15 unique strength envelope (Fig. 8).

190 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / FEBRUARY 2005

Downloaded 16 Jun 2012 to 180.211.192.67. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
 Fig. 9. Possible clay fabric and its cementation

The clay–water/cement ratio, wc / C, clearly affects the strength

and deformation characteristics of soft clays admixed with ce-
ment. Extensive investigations and analysis (Nagaraj et al. 1990;
Yamadera 1999; Nagaraj and Miura 2001 among others) have
revealed that clay microfabric consists of aggregated clay par-
ticles and the consequent enclosed capillary pore. Fig. 9(a) shows
the schematic diagram of the microfabric of clay. This is made up
of intraaggregate pores between individual aggregate with a pore
diameter less than 20 Å, and interaggregate pores between two
interacting aggregates with a diameter between 20 and 200 Å
depending upon the effective stress level. The enclosed large pore
would be far greater than 200 Å even up to 10,000 Å units when Fig. 10. Log e versus log k plots [data from Pane and Shiffman
the water content rises to the liquid limit. (1997)] and their normalization (Nagaraj and Narashimha Raju 2001)
The analysis of permeability data (Pane and Shiffman 1997) of
very high water content clays by Nagaraj and Narashimha Raju
(2001) (Fig. 10) suggests that the microfabric of clay at high
water content levels would be proportional to that at the liquid Once the clay–water/cement ratio is fixed in the field, if the
limit. The intercluster spacing would depend on the liquid limit, water content changes either due to wet mixing or to jet grouting,
resulting in capillary pores and the same fabric to give the same Eq. (1) can be used to estimate the corresponding changes to be
permeability for the same microfabric. made to the cement content, so as to realize not only the target
strength but also the desired levels of strength and compressibility
obtained in the laboratory test.
Clay–Water/Cement Ratio Identity

Based on the experimental observations and the basic consider- Conclusions

ations of clay microstructure in the analysis of test results, it is
possible to advance the following identity: The work reported in this paper indicates that clay–water/cement

再 冎再 冎
ratio is a microstructural parameter, which takes into account the
wc1 wc2 clay–water content responsible for the microfabric and the cement
= = constant 共1兲 content required to achieve cementation bonding for a specified
C1 C2
curing period. So the above parameter would take care of the
For a required target strength arrived at from a consideration of combined effect of both fabric and bonding. As long as this ratio
field parameters, the clay–water ratio can be calculated from the remains the same, even under different combinations of clay–
following relation (Horpibulsuk et al. 2003) which has been ad- water content and cement content, strength and deformation char-
vanced on the basis of Abrams’ law (1918). acteristics of cement admixed soft clay would be identical. Based

冉 冊
on the experimental observations and the basic considerations of
q共wc/C兲,D clay microstructure, an identity was proposed [Eq. (1)] to calcu-
= 1.24兵共wc/C兲,28−共wc/C兲,D其共0.038 + 0.281 ln D兲 共2兲 late variations in cement content required to reach the target
q共wc/C兲,28
specifications without altering the clay–water/cement ratio. In in
where q共wc/C兲,D = target strength required; q共wc/C兲,28 = strength devel- situ deep mixing, after fixing the target strength with due consid-
oped after 28 days for reference at 共wc / C兲 , 28; and 共wc / C兲D eration to various field parameters, the clay–water/cement ratio is
= strength required at a curing time of D days. computed using Abrams’ law. For any subsequent changes in

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / FEBRUARY 2005 / 191

Downloaded 16 Jun 2012 to 180.211.192.67. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
clay–water content either due to wet mixing or jet grouting, com- eds., Tokyo, 1, 431–436.
mensurate changes can be brought about in cement content by use Nagaraj, T. S., Miura, N., and Yamadera, A. (1998). “Induced cementa-
of clay–water/cement ratio identity so as to realize the target tion of soft clays—Analysis and assessment.” 1st Int. Symposium on
strength and deformation characteristics. Lowland Technology, Institute of Lowland Technology, Saga Univer-
sity, Saga, Japan, 1, 267–278.
Nagaraj, T. S., and Narashimha Raju, P. S. R. (2001). “Discussion on
References ‘The permeability of clay suspensions,’ by V. Pane and R. L. Schiff-
Abrams, D. A. (1918). “Design of Concrete Mixtures.” Bulletin 1 Struc- man.” Geotechnique, 51(4) 375–376.
tural Materials Research Laboratory, Lewis Institute, Chicago. Nagaraj, T. S., Srinivasa Murthy, B. R., and Vatsala, A. (1990). “Discus-
Broms, B. B., and Boman, P. (1975). “Lime stabilized column.” Proc., sion on ‘Change in pore size distribution due to consolidation of
5th Asian Regional Conf., Indian Institute of Science, Bangalore, clays,’ by Griffiths Joshi.” Geotechnique, 40(2), 303–305.
India, 1, 227–234. Okamura, T., and Terashi, M. (1975). “Deep lime mixing method of
Butterfield, R. (1979). “A natural compression law for soils.” Geotech- stabilization for marine clays.” Proc., 5th Asian Regional Conf. on
nique, 29(4), 469–480. Soil Mech., and Found. Engineering, Bangalore, India, 1, 69–75.
Horpibulsuk, S., Miura, N., and Nagaraj, T. S. (2003). “Assessment of Pane, V., and Schiffman, R. L. (1997). “The permeability of clay suspen-
strength development in cement admixed high water content clays sions.” Geotechnique, 47(2), 273–288.
with Abram’s law as a basis.” Geotechnique, 53(4), 439–444. Sridharan, A., Abraham, B. M., and Jose, B. T. (1991). “Improved tech-
Kamaluddin, M., and Balasubramaniam, A. S. (1995). “Overconsolidated nique for estimation of preconsolidation pressure.” Geotechnique,
behaviour of cement treated soft clay.” Proc., 10th Asian Regional 41(2), 263–268.
Conf. on Soil Mechanics and Foundation Engineering, Thailand, 407– Uddin, K., Balasuramaniam, A. S., and Bergado, D. T. (1997). “Engineer-
412. ing behaviour of cement treated Bangkok soft clay.” Geotech. Eng. J.,
Miura, N., Horpibulsuk, S., and Nagaraj, T. S. (2001). “Engineering be- 28, 89–119.
haviour of cement stabilized clay at high water content.” Soils Found., Yamadera, A. (1999). “Microstructural study of geotechnical characteris-
41(5), 33–45. tics of marine clays.” PhD dissertation, Saga University, Saga, Japan.
Nagaraj, T. S., and Miura, N. (2001). Soft clay behaviour–Analysis and Yamadera, A., Nagaraj, T. S., and Miura, N. (1998). “Prediction of
assessment, AA Balkema, Rotterdam, The Netherlands. strength development in cement stabilized marine clay.” Proc., Geo-
Nagaraj, T. S., Miura, N., Yaligar, P., and Yamadera, A. (1996). “Predict- technical Engineering Conf., Asian Institute of Technology, Bangkok,
ing strength development by cement admixture based on water con- Thailand, 1, 141–153.
tent.” Proc., IS-Tokyo’96, The Second Int. Conf. on Ground Improve- Yin, J. H., and Lai, C. K. (1998). “Strength and stiffness of Hong Kong
ment Geosystems, Grouting and Deep Mixing, R. Yonekura et al., marine deposit mixed with cement.” Geotech. Eng. J., 29(1), 29–44.

192 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / FEBRUARY 2005

Downloaded 16 Jun 2012 to 180.211.192.67. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org