Applied Clay Science 99 (2014) 24–34

Contents lists available at ScienceDirect

Applied Clay Science

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

Research paper

Rheology of concentrated bentonite dispersions treated with sodium

pyrophosphate for application in mitigating
earthquake-induced liquefaction
Marika Santagata a,⁎, Julia P. Clarke a,1, Antonio Bobet a, Vincent P. Drnevich a, Chadi S. El Mohtar a,2,
Pao-Tsung Huang a, Cliff T. Johnston b
a
Lyles School of Civil Engineering, Purdue University, 550 Stadium Mall, West Lafayette, IN 47907, USA
b
Department of Agronomy, Purdue University, 915 W State Street, West Lafayette, IN 47907, USA

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

Article history: The paper investigates the effectiveness of sodium pyrophosphate (SPP) to engineer the rheology of concentrat-
Received 20 January 2013 ed bentonite dispersions for their use in treating sands susceptible to earthquake-induced liquefaction. This ap-
Received in revised form 19 May 2014 plication requires that on the short term the rheological properties of the dispersion allow its permeation inside a
Accepted 21 May 2014
porous medium, but that once inside the sand pores, the dispersion regains its gel structure.
Available online 30 July 2014
The testing program focuses on 10% bentonite dispersions (expressed in terms of mass of bentonite by total mass
Keywords:
of the dispersion) prepared using a commercial Wyoming sodium-bentonite, pure water and SPP at dosages
Bentonite ranging from 0% to 2% by mass of the bentonite. Rheological tests are conducted using a Physica MCR 301 rota-
Sodium pyrophosphate tional rheometer to probe the flow and viscoelastic properties of the dispersions over time. Tests conducted at
Rheology early ages suggest that dispersions with SPP ≥ 0.5% are likely to be suitable permeation materials. Tests conduct-
Soil liquefaction ed over a period of two years show a significant evolution of the rheology, with all dispersions returning to ex-
Storage modulus hibit response typical of gel-like materials, including thixotropic behavior; the greater the SPP%, the more
Thixotropy delayed the formation of the gel. Custom rheological tests establish the independence of the behavior of the
SPP dispersions on the frequency of the applied load. These results suggest the ability of these dispersions to mit-
igate liquefaction during subsequent seismic events, and further support their use in the application considered.
© 2014 Elsevier B.V. All rights reserved.

1. Introduction
Bentonite, a clay composed primarily of montmorillonite, is used in
many industries due to the thixotropic properties of water bentonite
dispersions. In the geotechnical engineering field, bentonite dispersions
are used as drilling fluids, for underground sealing, for excavation sup-
port, in slurry cut-off walls to form impermeable barriers, and for
grouting. Control of their rheological properties is critical in all these ap-
plications. It is well established that a number of factors, including clay
concentration, pH, water ionic strength, type of cation and anion affect
the rheology of bentonite dispersions (e.g. Bekkour et al., 2005; Durán
et al., 2000; Harvey and Lagaly, 2013; Kelessidis et al., 2007; Laribi
et al., 2006; Luckham and Rossi, 1999; Malfoy et al., 2003; Penner and
Lagaly, 2001; Tombácz and Szekeres, 2004; Yildiz et al., 1999), which,
as a result of changes in these parameters, exist in different “states”:
⁎ Corresponding author.
E-mail address: mks@purdue.edu (M. Santagata).
1
Present address: Fugro Consultants, Inc., 6100 Hillcroft, Houston, TX 77081, USA.
2
Present address: Department of Civil, Architectural and Environmental Engineering,
The University of Texas at Austin, 301 E. Dean Keeton Stop C1792, Austin, TX 78712, USA.
sol, repulsive gel, attractive gel, and sediment (Abend and Lagaly,
2000; Lagaly and Dekany, 2013; Michot et al., 2004).
Recent research by El Mohtar et al. (2013, 2014) has explored the
use of bentonite for treatment of sand deposits susceptible to
earthquake-induced liquefaction. Soil liquefaction, which occurs in sat-
urated loose-medium granular deposits, is caused by the pore pressure
increase, and associated effective stress loss, produced by rapid loading
(e.g. see Ishihara, 1993; Seed and Lee, 1966). Liquefaction is an impor-
tant cause of damage to civil infrastructures during earthquakes, and ex-
amples of its occurrence have been documented in all recent
earthquakes, including the 2007 Pisco, Peru, the 2011 Tohoku, Japan,
the 2010 Maule, Chile, and the 2010 and 2011 Christchurch, New
Zealand earthquakes (e.g. see Bray and Frost, 2010; O'Connor et al.,
2007; Cubrinovski et al., 2011; PEER, 2011). Thus, there continues to
be interest in the geotechnical community to identify improved
methods to mitigate liquefaction, particularly in proximity of existing
structures, where traditional approaches relying on densification of
the soil cannot be always implemented.
Both field observations and laboratory measurements indicate that
the presence of plastic fines increases the liquefaction resistance of
sands (e.g. Ishihara, 1993; Sancio et al., 2002; Tang et al., 2013; Yasuda

http://dx.doi.org/10.1016/j.clay.2014.05.017
0169-1317/© 2014 Elsevier B.V. All rights reserved.
 M. Santagata et al. / Applied Clay Science 99 (2014) 24–34
et al., 1994; Yilmaz et al., 2004) and work by El Mohtar et al. (2013)
demonstrates that if approximately 3% bentonite by dry mass of the
sand can be delivered in the sand pore space, the liquefaction resistance
of the sand is greatly enhanced. This effect is ascribed to the rheology of
the gel-like pore fluid formed in the presence of bentonite inside the
sand pores, which effectively limits the mobility of the sand grains, re-
ducing the generation of pore pressure during cyclic loading (El
Mohtar et al., 2014). However, a soil treatment method that is based
on the delivery of this amount of bentonite into the sand pore space in
dispersion form is faced with significant practical challenges, as 3% ben-
tonite by mass of the sand corresponds to a concentration of clay in the
pore fluid of approximately 10%. At this concentration, bentonite disper-
sions exhibit a gel-like structure and rheological properties (e.g. the
presence of a yield stress), which do not permit their permeation in a
porous medium. Thus, modification of the rheology of the dispersion
is required. This modification, however, has to be reversible so that
once inside the sand, the bentonite dispersion can recover the gel struc-
ture and properties that ensure its effectiveness in mitigating
liquefaction.
This paper presents the results of a study that explored the use of
sodium pyro-phosphate decahydrate (herein referred to as SPP) to
engineer concentrated bentonite dispersions to be used in treating
liquefiable soils. Phosphates are known to be effective in dispersing
clays, and previous work (e.g. Goh et al., 2011; Lagaly, 1989; Lagaly
and Dekany, 2013; Penner and Lagaly, 2001; Shankar et al., 2010) has
shown that, through the neutralization of the positively charged sites
of the clay particles, the adsorption of pyrophosphate decreases the
yield stress of clay dispersions, permitting in some cases to reach
Newtonian behavior. It is also established that the addition of pyrophos-
phate can retard the gelation process of Laponite (e.g. Martin et al.,
2002; Mongondry et al., 2004), which like montmorillonite is a smec-
tite. Results obtained by Clarke (2008) demonstrate that this occurs
also in bentonite, with dosages much smaller than those required for
Laponite.
The study presented in the paper relies on rheological tests that
probed both the at rest and flow properties of concentrated (10% clay
by total mass of the dispersion) bentonite dispersions prepared with
dosages of SPP ranging from 0 to 2% by mass of the bentonite, and
that were conducted from immediately after mixing to after approxi-
mately two years of aging. The tests, performed under well-controlled
conditions using a state of the art rheometer, provide rheological data
that are directly relevant to the performance of the bentonite disper-
sions in the application of interest. In particular, the data presented
describe:
i) the initial structural breakdown of the dispersions due to the
addition of sodium pyrophosphate that is required for the “deliv-
ery” of the bentonite in a porous medium;
ii) the ability of the dispersions to recover gel structure and “solid”
like properties over time, that is critical for the bentonite to in-
hibit liquefaction;
iii) the small strain frequency independent response of the disper-
sions after the recovery of the gel structure that is relevant to
the performance under earthquake loading;
iv) the effectiveness of the dispersions to resist more than one seis-
mic event.
2. Experimental methods
2.1. Materials
Volclay CP-200, a Wyoming Na-bentonite commercialized by CETCO
and intended for geotechnical applications, was used to prepare all the
dispersions used in the rheological tests. While impurities present in
raw bentonite are known to potentially affect rheological results, the
clay was not pre-treated, with the exception of being sieved through a
25
#200 sieve to eliminate all non clay material coarser than 75 μm that
could cause blockage of the rheometer when using the cone-plate ge-
ometry (see below). The decision to test a commercial bentonite and
to avoid pre-treatment procedures stemmed from the desire to use a
material that was representative of one that could ultimately be used
in the field.
Sodium pyrophosphate decahydrate (Na4P2O7·10H2O) or SPP, with
molecular weight of 446.06 g/mol was purchased from the Sigma-
Aldrich Chemical Company.
Pure water with resistivity ranging between 17.8 and 18 MΩ was
used to prepare all clay dispersions. This was done to ensure that the
only variables controlling the rheological behavior were SPP percentage
and aging time.
2.2. Sample preparation and storage
A Hamilton Beach Model 936 mixer was used to prepare all benton-
ite dispersions. Half the amount of water needed for the dispersions was
first poured into the mixer's cup, and then bentonite was added, follow-
ed by the remaining mixing water. In the case of dispersions with SPP,
the mixing water was prepared appropriately diluting concentrated so-
lutions of SPP. The lowest speed (~11,000 RPM) available was selected
for mixing the dispersions for a total duration of 20 min. Mixing was
interrupted every 5 min to fold in the material adhering to the sides of
the vessel and promote homogeneity of the samples. Following mixing,
the dispersions were either tested or poured into 250 ml plastic con-
tainers that were sealed and stored in a cooler at 23 °C. This storage
method was found to be effective in avoiding drying of the sample, as
demonstrated by measurements of the clay content conducted on all
samples used for the tests after two years of storage. For all dispersions
these measurements show that the average values of the clay content
fall between 9.4% and 10%, with no trend with SPP%, and within the
range of values determined at earlier ages.
2.3. Rheological tests
All rheological tests were conducted using the Physica MCR 301 ro-
tational rheometer, an air bearing, stress-controlled device manu-
factured by Anton Paar GmbH, which can also operate in strain rate
controlled mode through a feedback control loop. The Physica MCR
301 is equipped with a permanent magnet synchronous drive (mini-
mum torque = 0.1 μN m, torque resolution = 0.001 μN m), and an op-
tical incremental encoder for measuring the shear strain (resolution b
1 μrad). It can be used to perform a variety of oscillatory and monotonic,
stress-controlled, and strain-rate controlled tests including amplitude
sweeps, frequency sweeps, time sweeps, rate ramps, stress ramps and
creep and recovery tests. The rheometer is equipped with a Peltier tem-
perature control system that can control the temperature in the −40 to
200 °C range. An automated computer software system is used to pro-
gram the testing variables and to compute and store the experimental
results.
Two different geometries (Fig. 1) were used for the tests: the coaxial
cylinder geometry for the low viscosity dispersions and the cone-plate
geometry for the higher viscosity dispersions (see Table 1). When
using the first geometry, the dispersion was poured in the cup, and
the bob was then automatically lowered into the cup to the gap position
(8 mm from the bottom of the cup). When performing tests with aging
up to 30 min the dispersion was poured into the coaxial cylinder cup
immediately after mixing, and a thin layer of vegetable oil was poured
on top of the dispersion to avoid exposure of the dispersion to the atmo-
sphere and minimize sample evaporation. It was established through
preliminary testing that the addition of the vegetable oil did not affect
the rheological results. For testing at all other aging times, the disper-
sion was kept in the storage container until immediately before testing.
No oil was used in this case. The dispersions were not stirred prior to
testing to minimize disturbance, as one of the main objectives of the
26 M. Santagata et al. / Applied Clay Science 99 (2014) 24–34

modulus, respectively (e.g. Barnes et al., 1989). These two moduli

Fig. 1. Schematic of testing geometries used for rheological tests (not drawn to scale).
work was to track the recovery of the gel structure and the evolution of
the mechanical properties of the dispersion over time.
When using the cone–plate measuring system special care was
placed in handling the dispersions and preparing the sample prior to
testing, as the reliability and reproducibility of the rheological results
hinge on these operations. After placing the required amount of disper-
b) Stress ramp tests — these tests consist of applying shear stress incre-
sion on the base plate using a small spatula, the cone was lowered under
computer control to the trimming height (slightly greater than the final
desired gap of 0.046 mm), and the excess material outside the cone
rim was trimmed. The cone was then lowered to the final position of
0.046 mm, equal to the cone truncation height. Note that this height is
over twenty times greater than the maximum size of the particles in dis-
persion, thus meeting minimum gap requirements recommended in the
literature (e.g. Mezger, 2006). Finally, the Peltier hood, fitted with a spe-
cial ring designed to minimize sample evaporation (Sato and Breedveld,
2005), was moved down and rested on the base. For all tests performed
with the cone and plate, prior to starting any test, an initial normal force
of 1 N was applied through the software for minimizing slipping
between the sample and the cone.
All tests presented in this paper were conducted at 25 °C. Because of
the small difference relative to room temperature, short equilibration
periods were required. Note that every test was performed on an inde-
pendent fresh sample.
Four different types of rheological tests were conducted:
c) Frequency sweep tests — these tests consist of applying an oscillato-
a) Amplitude or strain sweep tests — these tests consist of applying an
oscillatory shear strain (γ = γ0sin[ωt]) of increasing amplitude, γ0,
at constant frequency while measuring the resulting time-
dependent stress. The stress can be expressed as the sum of an elas-
tic component in phase with the strain and a viscous component in
quadrature with the strain, i.e. τ = γ0{G′sin(ωt) + G″cos(ωt)},
where G′ and G″ are termed the storage modulus and the loss
describe the elastic (solid) and the viscous (liquid) components of
the response. Hence, amplitude sweep tests represent one of the ap-
proaches to characterizing the visco-elastic behavior of dispersions.
All amplitude sweep tests presented in this paper were performed
immediately after sample setup (i.e. with no pre-shearing) varying
the applied shear strain amplitude between either 0.01% and 100%
(higher viscosity dispersions) or 0.1% and 100% (lower viscosity dis-
persions), using the same constant oscillatory frequency of 1 Hz.
It was of specific interest to determine the response of the samples at
1 Hz, as this is the frequency employed when measuring the lique-
faction resistance of soils in cyclic triaxial tests (ASTM D5311,
2011). Narrowing the shear strain range for the lower viscosity dis-
persions reduced noise and error in the data. Within the above
ranges the rheometer was programmed to ramp the shear strain
up on a log scale and conduct measurements at either 20 or 25 dif-
ferent shear strain values (6 points per decade). The average time
associated with completing this test varied between 3.8 and 6.1 min.
In this testing program the amplitude sweep tests are used to
measure the visco-elastic response of the material as a function of
SPP% and time, to identify the time required for the SPP dispersions
to develop a gel structure and to quantify the stiffness of the gel.
ments up to a limit value, and then applying shear stress decrements
with the same values as the increments. Stress ramp tests are gener-
ally performed to obtain measurements of the yield stress and of the
viscosity.
For these tests, the rheometer was programmed to increase the
shear stress from 5 Pa to a maximum value (between 60 and
140 Pa, depending on the test) in 25 steps and then decrease the
shear stress back to 5 Pa in the same number of steps. Also in this
case the tests were conducted with no pre-shearing. The average
time associated with completing this test was approximately
10 min in which each of the stress levels was held for 12 s to obtain
shear rates values closer to the steady-state shear rates (this was
determined based on preliminary tests described by Clarke, 2008).
In this paper, stress ramp tests are employed to quantify the changes
to short term flow behavior of the dispersion as function of SPP%,
and to describe changes in the degree of thixotropy with time for
two of the dispersions with SPP.
ry shear strain (γ = γ0sin[ωt]) of constant amplitude, γ0, varying
the oscillating frequency, while measuring the resulting time-
dependent stress.
In this testing program the oscillating frequency was varied between
0.5 and 10 Hz. The value of γ0 was selected based on the results of
the strain sweep tests and was equal to 0.5% for the dispersions
with SPP b 0.5%, and 0.1% for the dispersions with SPP% ≥ 0.5%.

Table 1
Overview of tests performed.

Time SPP dosage

0% 0.05% 0.1% 0.5% 1% 2%

0h AS(3), FS(8), SR(3)a AS(3)a – – – –

30 min – – AS(3), SR(2)b AS(2), SR(2)b AS(2), SR(4)b AS(3)b
10 h – – AS(3)b – AS(2)b AS(2)b
2 days AS(4)a AS(2)a AS(3)a AS(3)a AS(3)a AS(5)a
12 days – – AS(5)a – AS(3)a AS(4)a
25–27 days AS(3)a AS(2)a AS(3)a AS(3)a AS(3)a AS(3)a
2 months AS(3), FS(5)a AS(4)a AS(5)a AS(3), FS(6)a AS(4)a AS(4)a
4 months AS(4)a AS(4)a AS(3)a AS(5), FS(6), SR(4)a AS(3), SR(2)a AS(4)a
7 months AS(4)a – AS(6)a AS(3)a – –
2 years AS(2), Multi St.a – AS(2)a AS(2), Multi St.a AS(2)a AS(2)a

AS = amplitude sweep. SR = stress ramp. FS = frequency sweep. Multi St. = multi stage test.
(#) = Number of test repetitions.
a
Dispersions tested using cone–plate geometry.
b
Dispersions tested using coaxial cylinder geometry.
 M. Santagata et al. / Applied Clay Science 99 (2014) 24–34
These selections were determined by the results of the strain sweeps
to ensure that the dispersions displayed viscoelastic behavior during
the test duration. The rheometer was programmed to conduct a total
of 9 measurements (6 per decade of the frequency range). For the
pre-set measurement points, two tests were performed on separate
representative samples: one in which the frequency was increased
from the lowest value to the highest value of the range (ramp up);
the second conducted decreasing the frequency from the highest
to the lowest value (ramp down). This was done to ensure that stress
history effects did not control the results. The average times associ-
ated with completing these tests were 1.6 min and 1 min, respec-
tively.
In this paper, frequency sweep tests are performed to examine the
response of the dispersions for a range of frequencies (0.5–10 Hz)
that encompasses the frequency content of a typical earthquake
(e.g. see Kramer, 1996).
d) Multi stage tests — these tests were designed to assess whether dis-
persions could recover their ability to mitigate liquefaction after sig-
nificant straining. The tests consisted of three stages: the first was an
amplitude sweep test as the one described in a) which provided a
measure of the reference moduli (G′ and G″) for the particular test-
ing conditions examined (SPP% and time); during the second stage,
which lasted 4 min, oscillatory cycles of large amplitude (γ = 50%)
were applied with a frequency of 1 Hz to simulate the straining
that might be locally caused by a seismic event; finally, the third
stage consisted of a time sweep, that is the application of oscillations
of small amplitude (γ0 equal to 0.1% or 0.5% based on the same
criteria used for the frequency sweep tests) at constant frequency
(1 Hz) over 10 min to monitor the structure rebuilding process
after the destructuring stage, based on measurements of G′ and G″.
The average duration of these tests was 19.1 min.
These tests provide the means to: i) assess the degradation in the
mechanical properties (as measured by G′) occurring as a result of
the application of the large shear strain cycles; and ii) indicate the
ability of the dispersions to recover the pre-disturbance properties.
Table 1 summarizes the tests presented in this paper. The variables
examined are the dosage of SPP (0 to 2% by mass of the bentonite)
and the age of the dispersion (from after mixing to after two years of
aging). Initial measurements on the bentonite-only and 0.05% SPP dis-
persions were performed immediately after mixing (0 h, as stated in
Table 1). For all other SPP dispersions the first set of data was collected
30 min (0.5 h) after mixing, as studies of the kinetics of the sorption of
SPP to bentonite by ion chromatography indicate that equilibrium con-
ditions are most likely not reached until 30 min after mixing.
All tests shown in Table 1 were conducted on independent samples.
Moreover, for each testing condition several repetitions (see numbers in
parenthesis in Table 1) were performed on independent samples. Aver-
age results for these tests are reported in the following figures. Where
possible, compatibly with figure clarity, standard deviation bars are
also included.
3. Results and discussion
3.1. Effect of SPP on early age rheological behavior
As summarized in Table 1, amplitude sweep tests were conducted
on 10% bentonite dispersions with amounts of SPP equal to 0, 0.05,
0.1, 0.5, 1 and 2% SPP by bentonite mass. Representative results of strain
sweep tests performed at very early aging times (0 h or 30 min) on
these dispersions are shown in Fig. 2a–e. The data are presented in
terms of storage modulus (G′), loss modulus (G″) and phase angle (δ)
versus shear strain, where δ = arctan(G″/G′).
The plots for the reference 10% bentonite-only (i.e. 0% SPP) disper-
sion show the existence of a linear viscoelastic region, which extends
to a shear strain of about 2%. In this regions G′ is essentially constant,
27
and significantly greater than G″. This is reflected in the very small
phase angle (δ ~ 2°) measured in this region, which indicates that the
measured stress is almost in perfect phase with the applied oscillation.
This solid-like response is evidence of a gel type microstructure typical
of concentrated dispersions, which is broken down as the applied
shear strain increases. This breakdown is reflected in a decrease in G′.
In particular it is possible to identify a critical shear strain (γcrit) in
correspondence to which G′ becomes smaller than G″, i.e. the material's
response is dominated by the viscous behavior. For 0% SPP the average
value of γcrit is 30%.
Based on previous work by the authors (El Mohtar et al., 2013), the
behavior of the dispersion in Fig. 2a represents the target response in
terms of both type of behavior and mechanical properties of the pore
fluid that is effective in increasing the liquefaction resistance of a sand
deposit.
The dispersion's microstructure, and hence the response to the
strain sweep are clearly affected by the addition of SPP (Fig. 2b–f).
With the smallest addition of SPP (0.05% by bentonite mass), the disper-
sion continues to exhibit the same type of solid-like response at small
strains (δ ~ 3°), although G′ decreases to less than half the value mea-
sured on the bentonite only dispersion (note that because of the large
range of values obtained for G′, the plots do not have the same vertical
scale). As in the case of the bentonite-only dispersion, the solid to liquid
transition occurs in correspondence to a critical strain of about 30%.
With larger percentages of SPP, the dispersions experience severe
weakening of the microstructure, reflected in the decrease of G′ mea-
sured at small strains, which for SPP ≥ 0.5% is smaller than that of the
bentonite only dispersion by over two orders of magnitude. Additional-
ly, an increased viscous response is observed at the start of the test, with
the initial δ reaching values that exceed 30° for SPP% N 0.5%. Moreover, a
clear decrease in γcrit is observed with increasing SPP%: γcrit ~ 15%, 8.5%
and 7% for 0.5%, 1% and 2% SPP, respectively.
For SPP ≥ 0.5% the G′ versus shear strain curve does not show a clear
plateau of constant stiffness at smaller strains. Instead, a continuous de-
crease in G′ with increasing shear strain is observed. The slope of the
curve is relatively small at low strains and becomes larger at higher
strains. For these curves the G′ values reported in this paper are defined
in correspondence to the maximum curvature in the curve of G′ versus
shear strain, using the procedure described by Clarke (2008). Consistent
with data reported in the literature for some clay and polymer systems
(e.g. Geier, 2004; Markgraf et al., 2006, and Tirtaatmadja et al., 1997),
this type of behavior is observed for the dispersions with the weakest
micro-structural state, in which a continuous rupture of the material is
observed throughout the entire range of shear strains probed by the
test. Fig. 2 also shows that there is no significant difference between
the results for the 1% SPP and 2% SPP dispersions, indicating a limit in
the short-term effects of the sodium pyrophosphate.
The average values of the small strain G′ and δ (along with
+/− standard deviation bars) obtained from the analysis of all the
strain sweep tests conducted after mixing are synthesized in Fig. 3 as
a function of SPP%. Each of the data points shown represents the average
of values measured on independent samples (see Table 1), while the
horizontal bars represent the +/− standard deviation interval. For
each of the testing conditions the range in results is thought to reflect
slight changes in the stress history of the dispersions due to differences
in the duration of the trimming operations. Conventionally these effects
are minimized by applying a pre-shear stage prior to the rheological
measurements, to ensure that all tests start from a similar condition
and differences in stress history are eliminated. In this testing program
this was avoided, as the tests were aimed at capturing the “undisturbed”
response of the dispersions with SPP at a given aging time.
The data shown in Fig. 3 illustrate how the change in the rheological
properties due to the addition of SPP is not linearly proportional to the
percentage of SPP added. Rather, a significant reduction in G′ and an
increase in δ are associated with relatively small additions of SPP
(0.1%–0.5%). Higher additions of SPP cause a less significant decrease
28 M. Santagata et al. / Applied Clay Science 99 (2014) 24–34

a) b)
600 90 300 90

500 75 250 75

G' and G" (Pa)

G' and G" (Pa)

400 60 200 60

Phase Angle,

Phase Angle,
300 45 150 45
0% SPP - 0 hr 0.05% SPP - 0 hr
200 30 100 30
G' G" G' G"
100 15 50 15

0 0 0 0
0.01 0.1 1 10 100 0.001 0.01 0.1 1 10 100
Shear Strain, (%) Shear Strain, (%)
c) d)
120 90 6 90

100 75 5 75
G' and G" (Pa)

G' and G" (Pa)

80 60 4 60

Phase Angle,

Phase Angle,
60 45 3 45

40 30 2 30

20 0.1% SPP - 0.5 hrs 15 1 0.5% SPP - 0.5 hrs 15

G' G" G' G"
0 0 0 0
0.001 0.01 0.1 1 10 100 0.01 0.1 1 10 100
Shear Strain, (%) Shear Strain, (%)
e) f)
6 90 6 90

5 1% SPP - 0.5 hrs 75 5 2% SPP - 0.5 hrs 75

G' G" G' G"
G' and G" (Pa)

G' and G" (Pa)

4 60 4 60
Phase Angle,

Phase Angle,
3 45 3 45

2 30 2 30

1 15 1 15

0 0 0 0
0.01 0.1 1 10 100 0.01 0.1 1 10 100
Shear Strain, (%) Shear Strain, (%)

Fig. 2. Results of amplitude sweep tests conducted after mixing 10% bentonite dispersions with: a) 0%, b) 0.05%, c) 0.1%, d) 0.5%, e) 1%, and f) 2% SPP by mass of bentonite.

in G′ and increase in δ. Moreover, as already observed in Fig. 2, there is
no noticeable difference between the data for 1% and 2% SPP.
Additional information on the short-term effects of the addition of
SPP is provided by the flow curves obtained from the stress ramp
tests. There exist significant differences in the short term behavior of
the dispersions prepared with SPP compared to the bentonite-only dis-
persions (see Fig. 4). Consistent with previous studies on bentonite (e.g.
Abend and Lagaly, 2000; Geier, 2004; van Olphen, 1977), the curve for
the 10% bentonite-only dispersion is characterized by the existence of
a yield stress and significant hysteresis between the ramp-up and
ramp-down curves, which reflects thixotropy of the dispersion. In con-
trast, the addition of SPP produces a decrease in the yield stress, which
eventually disappears for 0.5% and 1% SPP. Along with a reduction in the
yield stress, an increasing percentage of SPP is also associated with a re-
duction of the thixotropic behavior. There is no difference in the flow
curves obtained with 0.5% and 1% SPP, and both dispersions exhibit
zero yield stress and no thixotropy.
The effect of SPP on the rheology of clay dispersions illustrated in
Figs. 2–4 is consistent with previous observations reported in the liter-
ature (e.g. Goh et al., 2011; Lagaly and Dekany, 2013; Mongondry
et al., 2004; Penner and Lagaly, 2001; Shankar et al., 2010) for both
bentonite and other clays treated with phosphate additives. The action
of SPP can be attributed to the high affinity for the broken edge sites
of clay minerals of the anion of SPP, which is strongly negative with a
charge of −4. Once adsorbed onto these positively charged sites, it im-
parts them a negative charge, which given the negative charge of the
basal surfaces of the bentonite, effectively limits edge to face interac-
tions. This explains the reduction in storage modulus, yield stress and
viscosity of the dispersion.
3.2. Short term aging (0 to 2 days)
Tests were conducted to investigate the impact of up to two days of
aging on the rheological properties of dispersions prepared with differ-
ent amounts of SPP. These tests were aimed at: a) evaluating the period
over which the SPP dispersions continued to exhibit behavior that made
them still suited as permeation materials, and b) at characterizing the
kinetics of the gel formation.
Fig. 5a shows the small strain average phase angle (δ) values obtain-
ed from strain sweep tests performed on 10% bentonite dispersions
with 0–2% SPP by bentonite mass over a period of two days. The figure
also presents best-fit lines through each of the data sets based on a
 M. Santagata et al. / Applied Clay Science 99 (2014) 24–34 29

a) a)
Storage Modulus, G' (Pa) 50
1000 1% SPP ( )
30 2% SPP ( )

100
0.5% SPP

Phase Angle,
10
10 0.1% SPP

1
0.0 0.5 1.0 1.5 2.0 0.05% SPP

% SPP by mass of bentonite

0% SPP
b) 1
50
0.1 1 10 100
40 Time (hours)

b)
Phase Angle,

30
1000
0% SPP
20
0.05% SPP

Storage Modulus, G' (Pa)

10

0 100
0.0 0.5 1.0
% SPP by mass of bentonite
Fig. 3. Small strain a) phase angle and b) storage modulus as a function of SPP% from
amplitude sweep tests conducted after mixing.
power equation (note that the 1% and 2% SPP data are grouped togeth-
er). Note that in this figure the data collected for 0% and 0.05% SPP are
plotted in correspondence to 0.1 h (6 min). This represents an estimate
of the time required to remove the dispersion from the mixing cup and
set up the sample.
Changes in δ illustrate the degree to which the nature of each disper-
sion evolves over the two-day period. Both the reference bentonite dis-
persion and the 0.05% SPP dispersion show no significant change in the
δ over two days (as reflected by the close to zero slope of the regression
lines though the data). This is not surprising as both these dispersions
display highly solid-like behavior (δ b 3° for both) from the onset. In
contrast, the dispersions with higher % SPP (0.1%, 0.5%, 1%, and 2%) all
show, to a different degree, a reduction in δ with time which reflects
an evolution in the microstructure. The most significant reduction in δ
is observed for the 0.1% SPP dispersion: δ drops from 27° to 4.5° within
the first 10 h. This is evidence of the formation of a gel structure over
this time period. For the other dispersions the decrease in δ is less
marked.
While the formation of a gel with solid-like properties (reflected by
small values of δ) is critical to the effectiveness of bentonite dispersions
60
0.5% & 1% SPP
0.5 hrs
50
Shear Stress, (Pa)
0.1% SPP
1.5 2.0
0.5% SPP
10
1% SPP ( )
2% SPP ( )
1
0.1 1 10 100
Time (hours)
Fig. 5. Variation of small strain a) phase angle and b) storage modulus from mixing to 2 h
of aging from amplitude sweep tests on 10% bentonite dispersions with 0–2% SPP.
in mitigating liquefaction, the mechanical properties of the gel also play
a role. For example, it has been shown (El Mohtar et al., 2013) that for
the same saturated sand–bentonite mixture the resistance to liquefac-
tion increases with time, i.e. as the stiffness of the bentonite gel inside
the sand pores increases. Measurements of the storage modulus obtain-
ed from strain sweep tests conducted over a period of two days can be
used to illustrate the evolution of the modulus of the dispersion with
time, as done in Fig. 5b. This figure summarizes average values (with
standard deviation bars) of the storage modulus G′ obtained from the
small strain region of the amplitude sweep tests. Again a power law
function is used in the figure to best fit each of the datasets (as above,
the data for 1% and 2% SPP are grouped together). For all the dispersions
the storage modulus increases with time. For the percentage of SPP in-
creasing from 0 to 0.5%, the greater the SPP percentage, the greater is
the increase in G′ with time, as reflected by the slopes of the regression
lines. For the 1%–2% SPP data the rate of increase in G′ appears to be

0% SPP - 0 hr
comparable to that of the 0.5% SPP dispersion (the slight difference in
40
slope is thought to be a result of the limited number of data available).
These results are consistent with the observations made for the phase
30
angle, and suggest that for the dispersions with greater SPP percentage,
0.1% SPP - 0.5 hrs the microstructure and hence the mechanical properties are evolving
20
more rapidly, compared to the bentonite only dispersion. Despite this,
the values of G′ measured on all dispersions prepared with SPP remain
10
lower than the initial value of the 10% bentonite-only dispersion. This
means that for all dispersions with SPP, there is no full recovery of the
0
0 200 400 600 800 1000 1200 stiffness over a two-day period. Moreover, significant differences are
. observed between the dispersions with SPP percentage b 0.5% and
Shear Rate, (s-1)
those with SPP percentage ≥ 0.5%. For the 0.05% and 0.1% SPP the values
Fig. 4. Flow curves from stress ramp tests conducted after mixing on 10% bentonite disper- of G′ measured at two days are of the same order of magnitude as the
sions with 0%, 0.1%, 0.5% and 1% SPP. initial value for the 10% bentonite-only dispersion (they differ by a
30 M. Santagata et al. / Applied Clay Science 99 (2014) 24–34

factor of approximately 2 and 3, respectively). On the other hand, at two 1000 0% SPP
days, G′ of the dispersions with SPP percentage ≥ 0.5% is over 30 times

Storage Modulus, G' (Pa)

smaller than the reference initial value measured on the 10% bentonite
0.05% SPP
only dispersion.
100
0.1% SPP

3.3. Long term aging (2 days to 2 years) 0.5% SPP

The effects of long term aging on the sodium pyrophosphate–
bentonite dispersions were investigated to establish which dispersions
had the potential over time to develop a gel structure with mechanical
properties comparable to those of a reference 10% bentonite-only dis-
persion. The recovery of the gel structure and of its mechanical proper-
ties was monitored through rheological tests conducted over a period of
two years.
The phase angle (δ) and the initial storage modulus (G′) derived
from strain sweep tests as a function of time (Figs. 6 and 7) complement
the results from to the shorter aging period (0 to two days) presented
above. As for the shorter aging periods, data for dispersions prepared
using different SPP percentages as well as for the reference 10%
bentonite-only dispersion are presented. Again, a power law function
is used to fit each data set.
Over the short term, δ of the 10% bentonite-only dispersion remains
essentially constant over the two-year period (Fig. 6). For all the SPP dis-
persions, again as seen over the short term, δ continuously decreases
with time. For the 0.5%, 1% and 2% SPP dispersions the rate at which δ
decreases (as measured by the slope of the regression line) remains
on the order of that observed over the first two days of aging (Fig. 5a).
Despite this, for the 1% and 2% SPP dispersions, even after two years of
aging the phase angle of these dispersions has not reached the low
value (~ 2°) characteristic of the 10% bentonite-only dispersion. This
value is approached by the 0.5% SPP dispersion at two years (δ = 2.5°).
The data of G′ (Fig. 7) obtained from amplitude sweep tests per-
formed over a two year period further illustrate the recovery of the mi-
crostructure of the dispersions with time. This figure, which also
includes the data presented in Fig. 5b for the shorter term aging,
shows that for all the dispersions G′ increases over time, and that, over-
all, the average rate of increase in G′ increases with SPP percentage
(compare the slopes of the lines drawn through the different data
sets). Additionally, the time at which G′ recovers the value measured
on the 10% bentonite only dispersion at 0 h increases with SPP percent-
age, going from ~2 days for 0.05% SPP to ~3.5 months for 0.1% SPP. After
two years of aging the 0.5% SPP dispersion is just approaching the refer-
ence value for the untreated dispersion, while for the 1% and 2% SPP dis-
persions the modulus remains close to 50% of the 0 h bentonite-only
value.
10
1% SPP ( )
2% SPP ( )
1
0.1 1 10 100 1000 10000
Time (hours)
Fig. 7. Variation of small strain storage modulus from mixing to two years of aging from
amplitude sweep tests on 10% bentonite dispersions with 0–2% SPP.
The data obtained by Clarke (2008) from the stress ramp tests are
generally consistent with the modulus results, showing the develop-
ment of a yield stress with time. As an example, after two months of
aging 10% bentonite dispersions with 0.5% and 1% SPP exhibit a yield
stress and thixotropic behavior (Fig. 8). The yield stress measured on
both dispersions exceeds the value measured at 0 h on the bentonite
only dispersion, indicating that the recovery of the modulus is delayed
compared to that of the yield stress.
As discussed above, the effect of phosphate additives on the short
term rheology of clay dispersions is well documented, and the primary
mechanism responsible for this behavior is generally accepted. Howev-
er, less evidence is presented in the literature on the recovery of the gel
structure and the associated rheological properties over time. For the
most part these data are limited to Laponite dispersions (e.g. Martin
et al., 2002; Mongondry et al., 2004). Moreover, the molecular mecha-
nisms that underlie this behavior are not understood. It is plausible
that the phosphate group that is not attached to the edge site attaches
to another clay particle over time, resulting in the observed recovery
of the storage modulus.
3.4. Sensitivity of response to frequency of loading
All amplitude sweep tests performed for this paper were conducted
at a frequency of 1 Hz, consistent with standard geotechnical practice
and with the recommendations contained in ASTM D5311 (2011) for
conducting cyclic triaxial tests to evaluate the liquefaction resistance

30 140
0.5% SPP - 2 months
20 120
Shear Stress, (Pa)

1% SPP ( )
2% SPP ( ) 100
10
Phase Angle,

80
5 1% SPP
60 2 months
0.5% SPP
0.1% SPP
3 40

0% SPP 20

1 0
100 1000 10000 0 500 1000 1500 2000
.
Time (hours) Shear Rate, (s-1)

Fig. 6. Variation of small strain phase angle from 2 h to two years of aging from amplitude Fig. 8. Flow curves obtained at two months of aging from stress ramp tests on 10% benton-
sweep tests on 10% bentonite dispersions with 0–2% SPP. ite dispersions with 0.5% and 1% SPP.
 M. Santagata et al. / Applied Clay Science 99 (2014) 24–34
of soils. However, the frequency content of an earthquake is typically
quite broad, with the majority of the contribution between 0.5 and
10 Hz (e.g. Kramer, 1996). Thus, it is of interest to ascertain that the
response of the bentonite dispersions is not affected by frequency of
loading.
Fig. 9 shows the results of frequency sweeps performed on 10% ben-
tonite dispersions with the procedures outlined in Section 2.3 above.
The figure portrays the storage modulus (G′) as a function of the fre-
quency for three 10% bentonite dispersions: one with 0% SPP tested im-
mediately after mixing either ramping up or ramping down the
frequency; one with 0% SPP tested after two months, again following ei-
ther a ramp-up or a ramp-down procedure; and one with 0.5% SPP test-
ed after two months. For the latter, three sets of data are reported: two
from tests in which the frequency was ramped down, and one from a
test in which the frequency was ramped up. Each curve in Fig. 9 pertains
to an independent sample. Note that for two of the tests the measured
value of G′ greatly exceeds the average value (G′ = 187 Pa) measured
in the linear viscoelastic region in the amplitude sweep tests (see
Fig. 7). The reason for this inconsistency remains unexplained. While
differences in the values of G′ for the three dispersions reflect the role
played by aging and SPP addition, it is of interest to focus on changes
in G′ with frequency. For the 0% SPP dispersion, as expected, the re-
sponse is found to be “nominally” frequency independent with a less
than 5% increase in G′ per tenfold increase in frequency, for both aging
durations considered. Slightly higher sensitivity to frequency is ob-
served for the 0.5% SPP dispersion with a less than 15% average increase
in G′ per tenfold increase in frequency. This is a reduction in sensitivity
relative to the behavior observed for shorter aging durations. For all dis-
persions, the testing procedure (ramp up versus ramp down) does not
appear to play a significant role.
3.5. Structure recovery after large strain shearing
Another aspect of the behavior of SPP-bentonite dispersions that is
significant in the context of using these materials to treat liquefiable de-
posits is the ability of the dispersion to “heal”, i.e. to recover its behavior
after the straining that it might locally undergo as a result of a seismic
event, and mitigate liquefaction during a subsequent earthquake. To ad-
dress this aspect of performance the multi-stage tests described in
Sections 2.3 were conducted. As mentioned above, these tests involve
an amplitude sweep, followed by a period of large strain oscillations,
followed by a final stage during which the change in visco-elastic pa-
rameters is monitored applying small strain oscillations for 10 min.
Fig. 10 illustrates the results of two of these tests performed on disper-
sions with 0% (Fig. 10a–c) and 0.5% SPP (Fig. 10d–e) two years after
mixing. Fig. 10a and d show how the storage modulus (G′), the loss
modulus (G″) and the phase angle (δ) vary with time over the duration
1400
Ramp Down 0% SPP - 2 months
1200
Storage Modulus, G' (Pa)
Ramp Up
1000
Each test on an independent sample
800
Ramp Up 0% SPP - after mixing
600
Ramp Down
400 Ramp Up
200
Ramp Down 0.5% SPP - 2 months
0 In contrast with what is observed for the bentonite only dispersion,
1 10
Frequency, f (Hz)
Fig. 9. Variation of small strain storage modulus with oscillation frequency of 10% benton-
ite dispersions.
31
of the test, while the remaining figures show the change in these param-
eters with shear strain over the three stages of the test. Starting with the
test on the bentonite only dispersion (Fig. 10a–c), the following is
observed:
During the amplitude sweep, which comprises the first stage of the
test, the same behavior as that exhibited at early ages (Fig. 2a) is ob-
served: after an initial plateau characterized by constant values of
the moduli, G′ decreases with shear strain (Fig. 10b) while G″ and
δ (Fig. 10c) increase. In correspondence to a critical shear strain
γcrit of approximately 30%, δ exceeds 45°, i.e. G″ becomes greater
than G′. This value of γcrit is the same as that measured on the disper-
sion immediately after mixing (see Fig. 2a). This is consistent with
previous observations on bentonite dispersions (Santagata et al.,
2008) that, for a given clay concentration, the critical threshold is
not dependent on the stress history of the sample.
During stage 2 of the test, due to the large strain oscillations, G′ and G
″ decrease further, with δ increasing, and reaching ~61° at the end of
the stage. This behavior indicates continued destructuring of the mi-
crostructure, and demonstrates that the dispersion is in flow at the
end of stage 2.
As soon as the large strain oscillations stop, δ immediately drops to
4°, reflecting an immediate recovery of the gel's solid behavior.
Concurrently, G′ shows a sharp rise, which is also a reflection of
the decrease in the shear strain level from 50% to 0.1%.
During stage 3, G′ continuously increases, reflecting the rebuilding
of the microstructure. The rate of increase decreases with time ap-
proaching a constant value after about 2–3 min.
At the end of the 10 minute monitoring stage G′ has reached a value
slightly below 50% of that measured in the amplitude sweep con-
ducted during the first stage of the test. This suggests that in time a
full recovery of the mechanical properties of the gel may occur.
During the monitoring stage the phase angle continues to decrease
slightly, approaching an average value of 2° (equal to the original
value) at the end of stage 3.
Overall, the behavior observed in stages 2 and 3 of the test is a result
of the thixotropic nature of bentonite dispersions.
Generally, similar behavior is observed in the dispersion with 0.5%
SPP (Fig. 10d–f). In particular, the rapid drop in the phase angle (from
47.5° to ~7°) at the end of stage 2 indicates that also for the dispersion
with 0.5% SPP the solid-like response is almost immediately recov-
ered after the large strain oscillation stage; and the increase in mod-
ulus during the final monitoring stage indicates the ability of the
dispersion's microstructure to “heal” after the large oscillations stage,
demonstrating that the dispersion with SPP has regained its thixotropic
properties.
The data portrayed in Figs. 10d–f also indicates some important
differences between the bentonite only dispersion and the dispersion
with 0.5% SPP. In particular:
The dispersion with 0.5% SPP reaches δ = 45° at the very end of the
strain sweep, i.e. γcrit is approximately 50% (Fig. 10e), which exceeds
the value measured on the bentonite only dispersion at the same
aging time. This indicates that the dispersion with SPP can accom-
modate a greater shear strain before flowing. This is in contrast to
what was observed at early ages (see Fig. 2) where γcrit decreased
with the addition of SPP.
the value of γcrit measured at two years also exceeds the value mea-
sured on the same dispersion at earlier ages. This suggests an evolu-
tion of the arrangement of the clay particles within the dispersion's
microstructure over time.
32 M. Santagata et al. / Applied Clay Science 99 (2014) 24–34

Stage 1: Stage 2: Stage 3: Stage 1: Stage 2: Stage 3:

Amplitude Oscillations Time sweep Amplitude Oscillations Time sweep
a) sweep
=0.01-50%
w/ =50% w/ =0.5% d) sweep
=0.01-50%
w/ =50% w/ =0.5%

500 50

G" (Pa)

G" (Pa)
1200 0% SPP 60 0.5% SPP
2 years 2 years
1000 50 400 40

Phase Angle,
Phase Angle,
800 40 300 30
600 30
G' and

G' and
200 20
400 20
100 10
200 10

0 0 0 0
0 200 400 600 800 1000 0 200 400 600 800 1000
Time (s) Time (s)

b) e)
1200 500
Stage 1: 0.5% SPP
Stage 1:
1000 2 years
400
0% SPP

G' and G" (Pa)

G' and G" (Pa)

800 2 years Stage 2: 300 Stage 2:

600
200
400
Stage 3: Stage 3:
100
200
Stage 1: Stage 1:
0 0
0.01 0.1 1 10 0.01 0.1 1 10
Shear Strain, (%) Shear Strain, (%)

c) f)
60
60 Stage 2:
0% SPP Stage 2: 50 0.5% SPP
50 2 years 2 years
Phase Angle,

Phase Angle,

40 40
30 Stage 3: 30 Stage 3:

20 20
10 Stage 1: 10 Stage 1:

0 0
0.01 0.1 1 10 0.01 0.1 1 10
Shear Strain, (%) Shear Strain, (%)

Fig. 10. Results of multi stage tests conducted at two years on 10% bentonite dispersions with a–c) 0% and d–f) 0.5% SPP.

As a result of the increase in γcrit, less significant destructuring, as
measured by the decrease in G′ and increase in δ, is observed during
both stage 1 and stage 2 relative to the bentonite only dispersion. For
example, at the end of stage 2, G′ has decreased by 92% in the disper-
sion with 0.5% SPP, compared to 96% with no SPP.
During the final monitoring stage a higher percentage increase in G′
is measured in the dispersion with 0.5% SPP relative to the bentonite
only dispersion (e.g. G′ increases by ~80% over the 10 minute mon-
itoring stage compared to the ~34% increase measured on the ben-
tonite only dispersion).
The above indicates that after sufficient aging the dispersion with
0.5% SPP is at least as effective as the bentonite-only dispersion in recov-
ering its structure and properties after a shear process.
4. Summary and conclusions
The paper describes the effects of the addition of sodium pyrophos-
phate (SPP) at dosages of 0–2% by mass of the clay on the rheological be-
havior of concentrated (10%) bentonite dispersions, based on a program
of rheological tests designed to assess the potential of these materials to
be used to treat sand deposits susceptible to earthquake-induced lique-
faction. The results of the tests conducted from immediately after
mixing to after two years of aging highlight the effect of SPP on the
initial rheological behavior and the changes in response with time,
and provide guidance for the design of dispersions to be used in the
field.
Amplitude sweeps and stress ramps indicate the following changes
in the initial (post mixing) rheological behavior of concentrated benton-
ite dispersions with the addition of SPP: reduced elastic response (i.e.
increased phase angle δ), reduced storage modulus, reduced yield
stress, and reduced thixotropy. These effects reflect a breakdown in
the gel structure typical of concentrated dispersions, and, for the higher
dosages of SPP, the formation of a sol. The effects are not linear with SPP
dosage, and become less marked for SPP N 0.5%. For dosages exceeding
this threshold no significant difference is observed in the flow curves,
although the amplitude sweep tests indicate a further decrease in G′
and an increase in δ.
The results of the tests conducted immediately after mixing indicate
that a minimum dosage of 0.5% SPP per mass of bentonite is necessary to
produce a zero-yield stress and non-thixotropic material that will be
able to permeate a fine porous medium. This is consistent with the re-
sults of laboratory scale permeation tests presented by El Mohtar et al.
(2013) and Rugg et al. (2011).
With time all the SPP dispersions show an evolution in the rheology.
Significant from a practical perspective is the modest change in phase
angle observed in the amplitude sweep tests conducted over the first
 M. Santagata et al. / Applied Clay Science 99 (2014) 24–34
two days for dispersions with SPP ≥ 0.5%. This result suggests that these
materials may provide a time window of permeation of a few hours that
will facilitate their use in the field.
Over the two-year aging period probed through the amplitude
sweep tests all dispersions show a continuous decrease in the phase
angle, which reflects the formation of a gel structure, and an increase
in storage modulus of the gel. The change with time of these parameters
follows a power law, with exponent increasing with SPP percentage.
Despite this, the greater the SPP dosage, the longer the time required
for the dispersions to develop the solid-like response and the storage
modulus measured on the bentonite only dispersion after mixing. This
indicates that, while with time all the SPP dispersions may be expected
to recover the microstructure and properties that are critical to their
ability to mitigate liquefaction, minimizing the dosage of SPP expedites
the process. Therefore, from a practical perspective the 0.5% SPP disper-
sion is the prime candidate for treating a sand deposit to reduce its
susceptibility to liquefaction.
Frequency sweep tests conducted varying the frequency between
0.5 and 10 Hz on dispersions with 0.5% SPP after development of the
gel structure indicate only a slight frequency dependent response com-
parable to that of bentonite only dispersions. Thus, the effectiveness of
the SPP-bentonite dispersion in mitigating liquefaction, to date demon-
strated exclusively through cyclic laboratory tests conducted at 1 Hz (El
Mohtar et al., 2013), is expected to extend to the more general loading
conditions associated in the field with actual earthquake loading.
Finally, custom tests were performed to probe the ability of the SPP
dispersions to recover their properties after the large straining that
might locally occur as a result of seismic loading. These tests, conducted
on dispersions with 0.5% SPP, demonstrate an almost immediate recov-
ery of the solid-like response of the dispersion, and a gradual healing
with time of its mechanical properties, a reflection of the fact that
with time the dispersion regains its thixotropic behavior. This suggests
that in the field treatment with SPP-bentonite dispersions should be ef-
fective in mitigating the liquefaction resistance of a sand deposit during
subsequent earthquakes. This is consistent with laboratory cyclic tests
performed by El Mohtar et al. (2013), which show the ability of sand
specimens treated with bentonite to maintain their increased seismic
resistance even after reaching liquefaction under a previous loading.
Acknowledgments
This research was partially funded by the National Science Founda-
tion, Geomechanics and Geotechnical Systems Program, under grant
CMS-0408739. The second author's MS research was supported through
an NSF AGEP (Alliance for Graduate Education and the Professoriate)
Fellowship and the Fugro Graduate Fellowship Program. This support
is gratefully acknowledged.
References
Abend, S., Lagaly, G., 2000. Sol–gel transitions of sodium montmorillonite dispersions.
Appl. Clay Sci. 16, 201–227. http://dx.doi.org/10.1016/S0169-1317(99)00040-X.
ASTM Standard D5311, 2011. Standard Test Method for Load Controlled Cyclic Triaxial
Strength of Soil. ASTM International, West Conshohocken, PA.
Barnes, H.A., Hutton, J.F., Walters, K., 1989. An Introduction to Rheology. Elsevier Science
Publishers, Amsterdam, The Netherland.
Bekkour, K., Leyama, M., Benchabane, A., Scrivener, O., 2005. Time-dependent rheological
behavior of bentonite dispersions: an experimental study. J. Rheol. 49, 1329–1345.
http://dx.doi.org/10.1122/1.2079267.
Bray, J., Frost, D., 2010. Geo-Engineering Reconnaissance of the February 27, 2010 Maule,
Chile Earthquake. PEER Report. Retrieved January 25, 2014 from http://www.
geerassociation.org/GEER_Post%20EQ%20Reports/Maule_Chile_2010/Ver2_Cover_
Chile_2010.html.
Clarke, J.P., 2008. Investigation of Time-dependent Rheological Behavior of Sodium
Pyrophosphate-bentonite Suspensions. . M.S. Dissertation Purdue University, West
Lafayette, Indiana.
Cubrinovski, M., Bradley, B., Wotherspoon, L., Green, R.A., Bray, J.D., Wood, C., Pender, M.,
Allen, J., Bradshaw, A., Rix, G., Taylor, M., Robinson, K., Henderson, D., Giorgini, S., Ma,
K., Winkely, A., Zupan, J., O'Rourke, T., DePascale, G., Wells, D., 2011. Geotechnical as-
pects of the 22 February 2011 Christchurch earthquake. Bull. N. Z. Soc. Earthq. Eng.
44, 205–226.
33
Durán, J.D.G., Ramos-Tejada, M.M., Arroyo, F.J., González-Caballero, F., 2000. Rheological
and electrokinetic properties of sodium montmorillonite suspensions: I. Rheological
properties and interparticle energy of interaction. J. Colloid. Interf. Sci. 229,
107–117. http://dx.doi.org/10.1006/jcis.2000.6956.
El Mohtar, C.S., Bobet, A., Santagata, M., Drnevich, V.P., Johnston, C., 2013. Liquefaction
mitigation using bentonite suspensions. J. Geotech. Geoenviron. 139, 1369–1380.
http://dx.doi.org/10.1061/(ASCE)GT.1943-5606.0000865.
El Mohtar, C.S., Bobet, A., Drnevich, V.P., Johnston, C., Santagata, M., 2014. Pore pressure
generation in sands with bentonite: from small strains to liquefaction. Géotechnique
64, 108–117. http://dx.doi.org/10.1680/geot.12.P.169.
Geier, D., 2004. Rheological Investigation of Bentonite Based Suspensions for Geotechnical
Applications. M.S. Dissertation Purdue University, West Lafayette, Indiana.
Goh, R., Leong, Y.-K., Lehane, B., 2011. Bentonite slurries-zeta potential, yield stress,
adsorbed additive and time-dependent behaviour. Rheol. Acta 50, 29–38. http://dx.
doi.org/10.1007/s00397-010-0498-x.
Harvey, C.C., Lagaly, G., 2013. Industrial applications. In: Bergaya, F., Lagaly, G. (Eds.),
Handbook of Clay Science, 5B. Elsevier.
Ishihara, K., 1993. Liquefaction and flow failure during earthquakes. Géotechnique 43,
351–415. http://dx.doi.org/10.1680/geot.1993.43.3.351.
Kelessidis, V.C., Tsamantaki, C., Dalamarinis, P., 2007. Effect of pH and electrolyte on the
rheology of aqueous Wyoming bentonite dispersions. Appl. Clay Sci. 38, 86–96.
http://dx.doi.org/10.1016/j.clay.2007.01.011.
Kramer, S., 1996. Geotechnical Earthquake Engineering. Prentice-Hall Civil Engineering
and Engineering Mechanics Series, Upper Saddle River, N.J.
Lagaly, G., 1989. Principles of flow of kaolin and bentonite dispersions. Appl. Clay Sci. 4,
105–123. http://dx.doi.org/10.1016/0169-1317(89)90003-3.
Lagaly, G., Dekany, I., 2013. Chapter 8: Colloid Clay Science in Handbook of Clay Science.
Part A: Fundamentals, In: Bergaya, F., Lagaly, G. (Eds.), 2nd edition. Elsevier,
Amsterdam.
Laribi, S., Fleureau, J.-M., Grossiord, J.-L., Kbir-Ariguib, N., 2006. Effect of pH on the rheo-
logical behavior of pure and interstratified smectite clays. Clay Clay Miner. 54,
29–37. http://dx.doi.org/10.1346/CCMN.2006.0540104.
Luckham, P.F., Rossi, S., 1999. Colloidal and rheological properties of bentonite suspen-
sions. Adv. Colloid Interfac. 82, 43–92. http://dx.doi.org/10.1016/S0001-8686(99)
00005-6.
Malfoy, C., Pantet, A., Monnet, P., Righi, D., 2003. Effects of the nature of the exchangeable
cation and clay concentration on the rheological properties of smectite suspensions.
Clay Clay Miner. 51, 656–663. http://dx.doi.org/10.1346/CCMN.2003.0510608.
Markgraf, W., Horn, R., Peth, S., 2006. An approach to rheometry in soil mechanics —
structural changes in bentonite, clayey and silty soils. Soil Till. Res 91, 1–14. http://
dx.doi.org/10.1016/j.still.2006.01.007.
Martin, C., Pignon, F., Piau, J.-M., Magnin, A., Lindner, P., Cabane, B., 2002. Dissociation of
thixotropic clay gels. Phys. Rev. E. 66. http://dx.doi.org/10.1103/PhysRevE.66.021401.
Mezger, T.G., 2006. The Rheology Handbook: For Users of Rotational and Oscillatory
Rheometers, 2nd ed. Vincentz Network, Hannover, Germany.
Michot, L.J., Bihannic, I., Porsch, K., Maddi, S., Baravian, C., Mougel, J., Levitz, P., 2004. Phase
diagrams of Wyoming Na-montmorillonite clay. Influence of particle anisotropy.
Langmuir 20, 10829–10837. http://dx.doi.org/10.1021/la0489108.
Mongondry, P., Nicolai, T., Tassin, J.-F., 2004. Influence of pyrophosphate or polyethylene
oxide on the aggregation and gelation of aqueous laponite dispersions. J. Colloid
Interf. Sci. 275, 191–196. http://dx.doi.org/10.1016/j.jcis.2004.01.037.
O'Connor, J.S., Mesa, L., Nykamp, M., 2007. Damage to the Highway System from the Pisco,
Peru Earthquake of August 15, 2007. Technical Report MCEER-07-0021. Retrieved January
25, 2014, from http://mceer.buffalo.edu/publications/Reconnaissance/07-0021/.
PEER/EERI/GEER/Tsunami Field Investigation Team, 2011. Preliminary Briefing on Japan's
Tohoku Earthquake, Aftershocks, and Tsunami of March 11, 2011. PEER Report, Re-
trieved January 25, 2014 from http://peer.berkeley.edu/news/wp-content/uploads/
2011/04/Tohoku-short-interim-report.pdf.
Penner, D., Lagaly, G., 2001. Influence of anions on the rheological properties of clay min-
eral dispersions. Appl. Clay Sci. 19, 131–142. http://dx.doi.org/10.1016/S0169-
1317(01)00052-7.
Rugg, D.A., Yoon, J., Hwang, H., El Mohtar, C.S., 2011. Undrained shearing properties of sand
permeated with a bentonite suspension for static liquefaction mitigation. In: Hand,
Alzamora (Eds.), Advances in Geotechnical Engineering, GSP 211. ASCE, Resto, VA.
Sancio, R.B., Bray, J.D., Stewart, J.P., Youd, T.L., Durgunoglu, H.T., Onalp, A., Seed, R.B.,
Christensen, C., Baturay, M.B., Karadayilar, T., 2002. Correlation between ground fail-
ure and soil conditions in Adapazari, Turkey. Soil Dyn. Earthq. Eng. 22, 1093–1102.
http://dx.doi.org/10.1016/S0267-7261(02)00135-5.
Santagata, M.C., Dalmazzo, D., Santagata, E., 2008. Deformation behavior of clay-water
suspensions from rheological tests. In: Burns, et al. (Eds.), 4th International Sympo-
sium on Deformation Characteristics of Geomaterials (IS-Atlanta). IOS Press,
Netherlands, pp. 453–459 (1).
Sato, J., Breedveld, V., 2005. Evaporation blocker for cone-plate rheometry of volatile sam-
ples. Appl. Rheol. 15, 390–397. http://dx.doi.org/10.3933/ApplRheol-15-390.
Seed, H.B., Lee, K.L., 1966. Liquefaction of saturated sands during cyclic loading. J. Soil
Mech. Found. Div. ASCE 92, 105–134.
Shankar, P., Teo, J., Leong, Y.-K., Fourie, A., Fahey, M., 2010. Adsorbed phosphate additives
for interrogating the nature of interparticles forces in kaolin clay slurries via rheolog-
ical yield stress. Adv. Powder Technol. 21, 380–385. http://dx.doi.org/10.1016/j.apt.
2010.02.013.
Tang, X., Ma, L., Shao, Q., 2013. Experimental investigation on effect ofbentonite content
on the liquefaction potential in saturated sand. ElectronicJournal of Geotechnical En-
gineering. 18, 1409–1417.
Tirtaatmadja, V., Tam, K.C., Jenkins, R.D., 1997. Rheological properties of model alkali-
soluble associative (HASE) polymers: effect of varying hydrophobe chain length.
Macromolecules 30, 3271–3282. http://dx.doi.org/10.1021/ma961202b.
34 M. Santagata et al. / Applied Clay Science 99 (2014) 24–34

Tombàcz, E., Szekeres, M., 2004. Colloidal behavior of aqueous montmorillonite suspen-
sions: the specific role of pH in the presence of indifferent electrolytes. Appl. Clay
Sci. 27, 75–94. http://dx.doi.org/10.1016/j.clay.2004.01.001.
van Olphen, H., 1977. An Introduction to Clay Colloid Chemistry. Wiley, New York.
Yasuda, S., Wakamatsu, K., Nagase, H., 1994. Liquefaction of artificially filled silty sands.
Proc. Ground Failures under Seismic Conditions. Geotechnical Special Publication
44. ASCE, New York, pp. 91–104.
Yildiz, N., Sarikaya, Y., Çalimli, A., 1999. The effect of the electrolyte concentration and pH
on the rheological properties of the original and the Na2CO3-activated Kütahya
bentonite. Appl. Clay Sci. 14, 319–327. http://dx.doi.org/10.1016/S0169-1317(99)
00006-X.
Yilmaz, M.T., Pekcan, O., Bakir, B.S., 2004. Undrained cyclic shear anddeformation behav-
ior of silt–clay mixtures of Adapazarı. Turkey. Soil Dyn. Earthq. Eng. 24, 497–507.