Applied Clay Science 142 (2017) 60–68

Contents lists available at ScienceDirect

Applied Clay Science

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

Research paper

Swelling properties of natural and modified bentonites by

rheological description
Gilles Barast a,b,1, Andry-Rico Razakamanantsoa c, Irini Djeran-Maigre a,⁎, Timothy Nicholson d, David Williams b
a
University of Lyon, INSA Lyon, Laboratory of Civil and Environmental Engineering, 69621 Villeurbanne Cedex, France
b
School of Civil Engineering, The University of Queensland, St Lucia, Brisbane, QLD 4072, Australia
c
Institut Français des Sciences et Technologies des Transports, de l'Aménagement et des Réseaux, CS 4 Route Bouaye, 44340 Bouguenais, France
d
School of Chemical Engineering, The University of Queensland, St Lucia, Brisbane, QLD 4072, Australia

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

Article history: Bentonite is used for landfill application because of its high sealing ability. Following the swelling phenomenon
Received 12 August 2015 the water reacts with the bentonite and expands the solid particles. The voids are filled and this reduces perme-
Received in revised form 5 January 2016 ability of the soil. However, the drawback is the long time needed to assess the hydraulic performance, whereas
Accepted 10 January 2016
the engineers require simple and rapid characterization parameters to design the hydraulic barrier. Present paper
Available online 9 February 2016
investigates the possibility of use of rheology testing to describe the swelling properties of bentonite used for geo-
Keywords:
technical engineering issues. Rheology testing is proposed to investigate the swelling properties of different na-
Modified clays ture of bentonite. Different types of bentonite are tested: natural sodium, natural calcium and sodium-activated-
Free swelling index calcium bentonites. Rheological measurements are performed in steady and dynamic conditions. Free swell and
Oedometer swelling index oedometer swell tests are also performed and compared.
Rheology steady state swelling ratio A simple model is proposed to unify the output parameters of the liquid-state bentonite from two rheology tests,
Rheology dynamic state swelling ratio with the ones obtained from two swelling tests. This approach propose a single parameter: the swelling ratio β
that describes the layer of water surrounding the particles based on the aspect ratio of bentonite packed particles.
The method highlights the correlation between swelling characteristics and rheological parameters of bentonites.
The results of swelling ratio underline distinctly the difference of mineralogy between the groups of bentonites. A
linear equation fits the values of β obtained from the two swelling tests and the values of β obtained from the
rheology tests, for all bentonites.
Low correlations are observed between the oedometer swelling ratio and the rheology ratio. But, the free swell-
ing ratio is in good agreement with rheological ratio and demonstrates a stronger correlation between the steady
state rheology ratio than the dynamic state rheology ratio.
© 2016 Elsevier B.V. All rights reserved.

1. Introduction The commercial bentonites are natural or chemical-treated clays

Geomaterials are defined as natural or enhanced (compacted,
blended, etc.) soils designed for engineering applications such as
hydraulic barriers (Koch, 2002) where the hydromechanics properties
of the soil on site must satisfy to the regulation requirements. Use of
high hydraulic performance geomaterials composed of bentonite is an
alternative. Bentonite, a high-moisture reactive clay, is applied as soil
binder for CCL (Compacted Clay Liner) or as sandwiched material be-
tween geotextiles for GCL (Geosynthetic Clay Liners). The bentonite
plays the role of water–absorbent, resulting in low permeability of the
soil and contaminant stabilisation.
⁎ Corresponding author at: University of Lyon, INSA Lyon, LGCIE, Coulomb Building, 34
av. des Arts, 69621 Villeurbanne CX, France.
E-mail address: irini.djeran-maigre@insa-lyon.fr (I. Djeran-Maigre).
1
Currently at University of Quebec in Abitibi-Temiscamingue, Rouyn-Noranda, Québec
J9X 5E4, Canada.
(Alther, 1986; Yildiz et al., 1999). Many factors affect the clay behaviour
(Van Olphen, 1977), but the bentonite is generally described as Ca ben-
tonite or Na bentonite, depending on the main exchangeable cation
(Ca2 + or Na+ respectively) on the clay surface sites. Gleason et al.,
1997 have performed permeability tests on a Ca clay and Na clay and
showed that Na clay have lower hydraulic conductivity than Ca clay.
The Na+ cation induces a higher water uptake than Ca2 + cation.
Gleason et al. (1997) state that to reach the plastic limit, the Na-
bentonite requires more than 500% of water compared to its initial vol-
ume, instead of Ca-bentonite which requires only around 100% of water.
Internally, this leads to the extension of the clay particle unit and pro-
duces the phenomenon commonly known as swelling (Jo et al., 2001;
Katsumi et al., 2008; Shackelford et al., 2000). The dispersion of the
bentonite particles is an effect of the swelling. As of immediate
consequence, the porosity is reduced and the fluid flow path is then re-
strained. The seepage is reduced (Mitchell, 1977). This permits there-
fore the correlation between the low permeability of the geomaterials

http://dx.doi.org/10.1016/j.clay.2016.01.008
0169-1317/© 2016 Elsevier B.V. All rights reserved.
 G. Barast et al. / Applied Clay Science 142 (2017) 60–68 61

and the swelling mechanism of the clay (Norrish and Quirk, 1954; Van 2. Background
Olphen, 1977).
The characterization of the hydraulic performance is an impor- A simple model based on the general clay particle fabric is used to
tant key for the qualification and optimization of the bentonite, describe the clay swelling. The swelling is interpreted by a single
required for the design of the hydraulic barrier system (Koch, value that represents the extension of a fictive layer around the solid
2002). This characterization must be done at laboratory scale prior particle. This description is illustrated in Fig. 1 and is known to be com-
any site design to ensure the appropriate choice of bentonite. The monly approached in other works (Larson, 1999; Russel et al., 1989;
hydraulic properties are mainly determined with permeability tests Tadros, 1996).
carried out with de-aired water as percolating fluid. Those perme- The shape of particle is an influential factor to the swelling mecha-
ability tests usually ran for several months (Katsumi et al., 2008; nisms. Therefore, a schematic parameter, namely aspect ratio, is given
Razakamanantsoa, 2009) because the hydraulic conductivity of ben- in order to approach the general mechanism of the swelling. Clay parti-
tonite geomaterials is very low, with values often observed in the cles are structured as platelets, or groups of platelets (Mitchell, 1977).
range 10 − 9 m/s–10− 11 m/s (Couradin et al., 2008; Gleason et al., The shape of clay particle is elongated with high aspect ratio value be-
1997; Katsumi et al., 2008; Razakamanantsoa, 2009; Shackelford cause its length dimension is much greater than its width dimension.
et al., 2000). Suspended clay particles pack into group of platelets and remain
Considering that the swelling mechanisms are intimately linked to difficult to be separated into singular particle due to the electrostatic
the hydraulic conductivity of clayey soils, investigations on the swelling interactions. But the group of platelets are aggregated particles with
properties of clay in the laboratory by oedometer tests and free swell significant width. Its morphology is of spheroid shape.
tests are therefore legitimate. A strong correlation has yet to be demon- Let R be the longitudinal particle radius and r the transversal particle
strated to improve the understanding of the relationship between radius. The unitary solid bentonite volume is in Eq. (1):
swelling and permeability. These two properties depend on the clay
nature, the fluid and the loading conditions (Mitchell, 1977). The V b ¼ 4πRr 2 =3: ð1Þ
engineers' works would be greatly simplified if fast and reliable tests
could accurately approach the hydraulic performance of the material The bentonite particles adsorb water and swell. A diffuse double
and predict the hydraulic conductivity values of the clay (and specifically) layer is formed by ion exchanges around the particle, due to the differ-
bentonite geomaterials. ent charge position (Van Olphen, 1977). Let us associate a fictive water
Plasticity properties of soil determined by Atterberg limit tests layer, of thickness a, that includes the water mobilized and bound by
are also used as index for soil mechanics parameters (Mitchell, electrostatic forces to the clay surface. This fictive layer is simply the
1977). The plasticity parameter values vary in accordance with consequence of the diffuse double layer. The water layer is supposed
the effects of electrolyte fluids and clay mineralogy. Both parame- to be immobile, relatively to the clay particle. For the simplification of
ters are thus associated with the change in clay particles interac- the model, the particles are considered to be spread homogenously in
tions (Gleason et al., 1997; Katsumi et al., 2008). However, the the total volume VT and their size is assumed to be identical and con-
preliminary observations made by Razakamanantsoa et al. (2012) stant. The unitary fictive volume is hence described in Eq. (2):
with modified bentonites showed divergence from the expected
Atterberg test results. The modified bentonite mixtures were too V p ¼ 4πðR þ aÞðr þ aÞ2 =3: ð2Þ
viscous to be rolled in a cylindrical shape as recommended by the
standard. The packing fraction Φp represents the clay particle fraction for
The rheological characterization of clay is extensively used for which the solid particles are packed in the total volume VT (Eq. (3)):
colloid state materials (Coussot, 2005; Goodwin, 2004; Russel et al.,
1989; Van Olphen, 1977). A few recent studies have tried to apply Φp ¼ ΣV b =V T : ð3Þ
rheology testing to understand and solve geotechnical problems
(Ghezzehei and Or, 2001; Markgraf et al., 2006), highlighting its poten- The packing fraction Φp is function of the particles shape and size.
tial application. The clay particles occur as platelets, non-spherical forms. Due to the
For landfill application, the swelling properties of clays via the rheo-
logical characterization for hydraulic barrier designing, is not common
(Khandal and Tadros, 1988; Paumier et al., 2008). A precise method to
link the rheology parameters of bentonite to its swelling properties is
still missing.
This paper highlights the possibility to study precisely the swelling
mechanism using a complementary, simple and rapid test which is
the rheology. It has the advantage of being precise and given qualitative
results traduced to swelling parameters. This kind of parameters is very
important for landfill designing, giving practical tools for engineers to
optimize the design of hydraulic barriers.
The present study focuses on the liquid–solid continuum between
colloid liquid-state and solid-state of the bentonite geomaterials. A sin-
gle parameter deduced from rheology on liquid-state materials is used
to understand and describe the swelling properties of different types
of bentonite at solid state. The discussion will be based on the proposed
parameter, declined in both liquid and solid domains. For that purpose,
the swelling will be unified by an unique comparative parameter,
interpreted by spheroids model of clay particles. First, the swelling
parameter β is calculated with two swelling tests at solid state:
Fig. 1. Scheme of the proposed model of swollen particle configuration, with R the
oedometer swell test and free swell test. It is then compared to the β cal- longitudinal particle radius, r the transversal particle radius, Vb the solid bentonite
culated with two rheological tests at liquid state (dynamic and steady volume of a particle, a the thickness of a fictive water layer, VT the total volume and Vp
conditions). the fictive volume of a particle.
62 G. Barast et al. / Applied Clay Science 142 (2017) 60–68

water molecules mainly attached to the particle longitudinal surface
rather than the edge, and the particle flow in the fluid, the shape can
be assumed as spheroidal. A factor ratio p = R/r is introduced to take
into account the shape of the particle in the modelling. Three configura-
tions are envisaged and reported in Table 1. The simplest model is for
p = 1 where the particles are considered as spheres. For more realistic
description of particles observation at micro-scale (Mitchell, 1977),
the aspect ratio was selected for p = 3 and p = 50. The intrinsic viscosity
and maximal volume fraction in accordance to the particle aspect ratio
have been evaluated in other works for clay and other materials
(Guven and Pollastro, 1992; Mitchell, 1977; Philipse, 1996; Wouterse
et al., 2007).
Coefficient β is introduced to describe the Swelling Ratio (Larson,
1999; Russel et al., 1989; Tadros, 1996;). β defines the swelling of parti-
cles as expressed in Eq. (4):
∑V p V p
β¼ ¼ :
∑V b V b
for 20 min. For hydric equilibrium of the mixture, the suspension was
kept, for 24 h, in a temperature controlled room, at 22 ± 1 °C. This
time was followed by 1 min stirring to re-homogenize. The specimen
preparation remained identical for all tests to minimize discrepancy in
the results (Laribi et al., 2005). For all bentonite suspensions, the dry
mass concentration in demineralized water (DW) was in the range
from 4% to 8%. For each bentonite suspension, a volume inferior to
30 ml suspension was required for testing the material in the
rheometer.
4. Methods
The comparison between the different mechanical and rheological
tests, in this section, will be unified with a key parameter β that is
defined for each test as a function of the interpretation of the swelling
occurring during the test.
ð4Þ
4.1. Free swell test

Remembering Eq. (2), it is therefore possible to write the Eq. (4) as Standard free swell tests are carried out following the French stan-
follows in Eq. (5): dard NF XP 84-703. Two grams of dried powdered clay, ground and
passed through a 160 μm European standard sieve, was progressively
4
a 2 poured into a 0.1 mL graduated cylinder, containing 100 mL of aqueous
2
∑V p V p 3 π ðR þ aÞðr þ aÞ Vb 1 þ
R
1 þ ar solution. The height of swollen clay in the cylinder was measured after
β¼ ¼ ¼ ¼
∑V Vb Vb Vb 24 h with mL/2 g as unit. Tests were reproduced several times so that

b a
a 2
uncertainties of ± 0.5 mL/2 g were accounted for the operator visual
¼ 1þ 1þ : ð5Þ
R r reading. The result of the test is given as a free swelling index (FSI)
value. FSI represents the total volume of the swollen material VT, inside
In the following subsections, a simple approach to determine the the testing cylinder, and written as in Eq. (6):
swelling ratio will be described for each test, complying with the
model assumed in Fig. 1. The assumption of spheroidal shape will be FSI ¼ V T : ð6Þ
used to describe the particle and aggregate shapes.
Following the definition of the clay fraction Φp (Eq. (3)), the Eq. (6)
3. Materials is transformed to Eq. (7):

FSI ¼ V T ¼ ΣV p =Φp : ð7Þ

All the tested clays are commercial bentonites commonly used as
component for Compacted Clay Liner (CCL) or in Geosynthetic Clay
Liner (GCL) applications (Couradin et al., 2008). The B1, B2 and B3 are It is possible to introduce the parameter β, namely βFSI for the appli-
Na-activated Ca bentonites from France. The B4 clay is a natural Na cation to FSI, from its definition (Eq. (5)) in order to obtain Eq. (8):
bentonite from the Wyoming state in USA. The B5 and B6 are natural
FSI ¼ V T ¼ ΣV p =Φp ¼ βFSI ΣV b =Φp : ð8Þ
Ca bentonites from France. The testing fluid is demineralized water
(DW). Several characteristics of the bentonites are summarized in
Since the grains are supposed of equal size and homogenously
Table 2, as moisture content, geotechnical identification with liquid
spread, Vs is the total volume of solid and Vb the volume of bentonite
and plastic limits and blue methylene adsorption value, electrical con-
of one particle, thus ΣVb = Vs, and the Eq. (8) becomes Eq. (9):
ductivity and pH obtained with demineralized water. For the natural
Na-bentonite, the Atterberg limits are difficult to perform because of FSI ¼ V T ¼ ΣV p =Φp ¼ βFSI V s =Φp : ð9Þ
the high ability of the bentonite to retain water. As the B1 clay, is an
activated B5 and the B2 clay an activated B6, the activation permits to The total volume of solid bentonite Vs in the cylindrical tube corre-
exhibit higher values of blue methylene adsorption as excepted. sponds to Vs = ms/ρs. The bentonite solid mass is ms and the associated
The bentonite specimens are used in their as-received commercial specific density is ρs = 2.65 g/cm3. The free swell test is performed with
state in order to directly refer to the industrial applications. The benton- ms = 2 g of clay. The Eq. (9) is then rewritten as Eq. (10):
ite suspensions were prepared by mixing air-dried bentonite and water
FSI ¼ V T ¼ ΣV p =Φp ¼ βFSI V s =Φp ¼ βFSI mS =Φp ρs : ð10Þ
Table 1
Packing fraction and intrinsic viscosity chosen for bentonite mixtures, as function of the
factor ratio of the particles.
Based on Eq. (10) the swelling ratio parameter β is deduced,
(Eq. (11)):
Configuration Aspect Maximum packing Intrinsic viscosity [η]Φp
ratio p fraction Φp [η] βFSI ¼ FSI ρs Φp =ms : ð11Þ
Sphere 1 0.641 2.52 1.6
Spheroid 3 0.643 3.434 1.92
Ellipsoid 50 0.1155 35.14 4.03
4.2. Oedometer swell test
1
Maximum packing fraction for non-deformable hard spheres.
2
According to Einstein equation for non-deformable hard spheres. Oedometer swell tests are performed, following the standard XP P
3
According to the observations made by Paumier et al. (2008).
4
According to equations found in the works of Guven and Pollastro (1992) and
94-090-1. The powder bentonite is placed inside a ring cell with as-
citations therein. received moisture content. The cell is greased to avoid wall friction dur-
5
According to equations found in the works of Philipse (1996). ing the swelling phase on wetting. The specimen is mounted between
 G. Barast et al. / Applied Clay Science 142 (2017) 60–68 63

Table 2
Physical and geotechnical characteristics of different bentonites.

Bentonite Type w (%) LL (%) PL (%) BMV (mg/g) daverage (μm) EC (μS/cm) pH FSI (mL/2 g)

B1 Na-activated Ca 11.3 267 61 397 18.4 25.1 10.61 30

B2 Na-activated Ca 9.5 452 51 305 8 24.6 11.06 28
B3 Na-activated Ca 13.2 657 66 250 10 25 10.63 25
B4 Natural Na 10.6 671 113 – 7.6 24.6 10.24 31
B5 Natural Ca 9.95 468 56 320 32.5 24.4 9.1 6
B6 Natural Ca 10.35 423 62 250 22.4 24.7 9.3 7

w: Moisture content.
LL and PL: Liquid Limit and Plastic Limit, determined from Atterberg limits.
BMV: Blue Methylene adsorption value.
daverage: average assumed spherical particle diameter, obtained from low scattering laser Malver Mastersizer 2000.
EC: electrical conductivity.
pH obtained with demineralized water.
FSI: free swelling index.

two drain porous stones and filter papers to allow uniform water infil- Finally, the definition of the maximum swelling strain (Eq. (13))
tration and prevent particles clogging in the porous stone. The loose leads to the Eq. (18):
state bentonite is placed with 5 kg/m2 of weight per surface and
X
0.9 g/cm3 of dry density in order to approach the density state of field Vp V T Φp V T Φp V T ρs Φp ρ
GCL materials (Badv and Farsimadan, 2009). The upper porous stone βoedo ¼ X ¼X ¼ ¼ ¼ ðεs max þ 1ÞΦp s : ð18Þ
Vb Vb Vs V i ρd ρd
and top plate initially apply 2 kPa stress on top of the specimen prior
water soaking for at least seven days. The thickness variation of the
specimen was recorded with a displacement sensor of 5 μm precision. The value of initial dry density ρd will be obtained during the test.
The measurement of the final moisture content by 105 °C oven dry
method confirmed that the specimen was fully saturated at the end of 4.3. Rheology measurements: dynamic test (RDT)
the test. At the start of the test, the material occupies an initial volume
Vi with an initial dry density ρd described in Eq. (12): Small amplitude with dynamic oscillatory shear test is frequently
used for investigations that are restricted to the linear viscoelasticity
ρd ¼ ms =V i ¼ ρs V s =V i : ð12Þ (Goodwin, 2004; Larson, 1999; Tadros, 1996). The theory can be applied
for low strain γ, where the constitutive equation, for shear rate γ and
The tested specimens are placed within the same range of dry den- shear stress σ, are governed by Newtonian law for viscosity (η = σ/γ)
sity values than the in situ applications (Cho et al., 1999; Komine and and Hooke law for elasticity (G = σ/γ). The stress per unit strain, at
Ogata, 1994; Sridharan and Gurtug, 2004). The test usually defines the an oscillatory frequency Ω, is called complex modulus G* (Pa) and is de-
swelling in term of volumetric strain ε = dV/Vi as the ratio of the incre- fined by the complex notation (i2 = −1) as G*(Ω) = G′(Ω)cos δ + iG
ment swollen volume dV over the initial sample volume Vi. The maxi- ″(Ω)sin δ. This complex modulus is composed of the storage elastic G′
mum swelling strain εsmax obtained after saturation stage is calculated (Pa) term and the loss viscous G″ (Pa) term. The phase shift δ is the
as in Eq. (13): difference in phase between the applied strain and the response stress
wave, for each frequency, providing information on the G′ and G″
εsmax ¼ ðV T− V i Þ=V i ¼ ðV T =V i Þ−1: ð13Þ contributions.
The dynamic rheological testing was performed with an ARES rate-
The parameter β (namely βoedo for the application to oedometer controlled rheometer. The geometry used for data is a coaxial Couette
test) is introduced from its definition (Eq. (4)) and expressed as cell of 32 mm bob diameter and 34 mm length. Preliminary testing
(Eq. (14)): with cone-plate at 1° angle and 70 mm diameter on a Thermo Haake
X rheometer confirmed the non-wall slip condition for the bentonite
Vp specimens at the tested solid concentration. The temperature was main-
βoedo ¼ X : ð14Þ tained constant in the controlled room to 22 ± 1 °C. Once the specimen
Vb
was poured into the cup, the bob was computer-assisted driven into the
suspension with automatic assistance to reach the correct gap position.
Implementing the Eq. (3) into the Eq. (14), the parameter can also be An excess of material and plastic cap were placed on top of the cell
written in Eq. (15): to cover the tested specimens and avoid great effect of evaporation.
X The testing process started with 1 rad/s pre-shear for one minute,
Vp V T Φp followed by one minute rest time, for the structure to have arbitrary
βoedo ¼ X ¼X : ð15Þ
Vb Vb restructuration as basis of comparison.
The rheological tests are dynamic frequency tests at 1% strain γ, over
Remembering that ΣVb = Vs, this Eq. (15) continues into Eq. (16): a frequency Ω range of 0.1 rad/s and 100 rad/s. Preliminary dynamic
strain tests at 1 Hz over 0.01% to 100% strain confirmed that, performed
X
strain value of 1% was low enough to maintain the material in the linear
Vp V T Φp V T Φp
βoedo ¼ X ¼X ¼ : ð16Þ viscoelastic region. The elastic modulus G′, the viscous modulus G″ and
Vb Vb Vs
the free-frequency elastic modulus G′o (defined as the value for which
the elastic modulus G′ remains approximately constant over a frequency
Coupling Eq. (12) and Eq. (16), it is possible to write the Eq. (17): range in the semi-log scale plot) are determined in accordance to the
viscoelasticity theory.
X
Vp V T Φp V T Φp V T ρs Φp The rheological parameters are expressed as function of the
βoedo ¼ X ¼X ¼ ¼ : ð17Þ clay fraction due to the strong dependence on the flow regime and the
Vb Vb Vs V i ρd
deformation properties in regard to the free bulk fluid in the system.
64 G. Barast et al. / Applied Clay Science 142 (2017) 60–68

Other researchers also pointed out that for moderate volume fractions, The swelling ratio β (namely βRST for application on rheology steady
the free frequency-dependent elastic modulus G′o is function of the test) is introduced by the relationship Eq. (4), and coupled with the
clay fraction as in Eq. (19) (Buscall, 1982; Larson, 1999; Ramos-Tejada Eqs. (3) and (20), to obtain the Eq. (23):
et al., 2001; Tadros, 1996):
X
G0o ¼ αΦn ð19Þ Vp Φp V T Φp
βRST ¼ X ¼ ¼ : ð23Þ
Vb Φm V T Φm
with α the proportional factor and n the power index. As expressed by
the power law function, the value of the rheological parameter tends
toward infinite value within the equipment measurement scale. The Finally, Eq. (22) is inserted into Eq. (23) to obtain the expression of
proposed model will consider that the maximum solid fraction of the βRST in Eq. (24):
bentonite, ΣVb, in a given total volume, VT, is represented by Φm. For
any solid fraction above this limit value, the material undergoes a tran-
 −ðηÞΦ !
sition in its behaviour since other types of interactions overlap (friction, Φp η
1
p

compression, etc.) and therefore the simple modelling approach will βRST ¼ 1− : ð24Þ
Φ ηs
not correctly describe the swelling mechanism (Coussot, 2005). As a
limit value, the value of the clay fraction Φm was defined as the measur-
able value of Φ for which the parameter G'o tends to an infinite value.
Knowing the viscosity of water ηs = 0.001 Pa·s, the swelling ratio
Obviously, this determination relies heavily on the equipment upper
βRST can be calculated from one rheology steady test, provided that
limit value and on the ability to place the homogeneous material in
the concentration of the bentonite suspension is low enough to main-
the testing geometry. From the previous considerations, the maximum
tain its Newtonian behaviour and that the curve drawn from Eq. (22)
clay fraction is defined as in Eq. (20):
shows a linear data fit.
X
Vb
Φm ¼ ¼ limΦ : ð20Þ 5. Results and discussions
VT 0
G o →∞

Reminding Eqs. (3) and (20), the swelling ratio β, (namely βRDT for 5.1. Free swelling index results
application to the dynamic rheology test) defined in Eq. (4) is now
obtained in Eq. (21): The FSI results and the swelling ratio associated to all bentonite
specimens (Eq. (11)) are reported in the Fig. 2 and Table 2.
X The FSI values are in the same order of magnitude than what other
Vp Φp V T Φp
βRDT ¼ X ¼ ¼ : ð21Þ researchers measured for bentonite materials (Jo et al., 2001; Katsumi
Vb Φm V T limΦ
0
G o →∞ et al., 2008; Shackelford et al., 2000). The swelling potential of the
Ca-bentonites B5 and B6 are similar, with a low FSI value around
6 mL/2 g, which is relevant of low swelling clay properties. These ben-
4.4. Rheology measurements: steady test (RST) tonites mainly contain Ca2+ cations in their interfolliaire space so that
the clay layer extension, and hence the swelling, are minimal. All
The steady shear rate tests were performed on a Thermo Haake Na-activated Ca bentonites are within close values range (24 mL/
rheoscope stress-controlled, with cone-plate geometry of 70 mm cone 2 g–30 mL/2 g), expressing better swelling properties than the natural
diameter and 1° cone angle. A few millilitre of the bentonite suspension Ca bentonites. Their chemical treatment activation by Na2CO3 enhanced
were poured on the plate and constrained in the gap with assisted com- the presence of exchangeable cation Na+ that provides greater clay
puter mechanism. The room temperature was maintained constant at sheet layer extension and macroscopic swelling, in contact with water
22 ± 1 °C. The testing time was short enough to ensure minimal effect (Shackelford et al., 2000). The natural Na bentonite has higher swelling
of evaporation as performed by other researchers (Markgraf et al., potential with FSI more than 31 mL/2 g, as it contains more sodium
2006). The testing procedure started with 1 rad/s pre-shear rate for cations than the chemical treated-activated clays.
1 min and followed by 1 min rest time. The steady shear test was per-
formed for multi steps for 0.5 rad/s up to 500 rad/s; with the values of
the viscosity and stress for each rate obtained after 60 s shear time. A
preliminary step rate test was performed on the bentonite suspension
beforehand to ensure the achievement of shear equilibrium within the
60 s shear time. The bentonite particles are charged particles where
the electrostatics interactions take place. Russel et al., 1989 have
expressed the viscosity η, for charged spheres, as function of the clay
fraction Φ. Larson, 1999 underlined that for moderate volume fractions,
the extension of the Krieger–Dougherty (Woods and Krieger, 1970)
equation should be more accurate, with Eq. (22) rewritten with corre-
sponding nomenclature:

 
η Φm −½ηΦp
¼ 1− ð22Þ
ηs Φp

with ηs the solvent viscosity, [η] the intrinsic viscosity, Φp the effective
clay fraction taking into account the charged interactions surrounding Fig. 2. Free swelling index (FSI) and associated swelling ratio βFSI values for bentonites, in
the particles and Φm the maximum effective clay volume fraction possi- contact with demineralized water after 24 h rest time. Note that the determination is
ble in the volume. presented for spheres (p = 1).
 G. Barast et al. / Applied Clay Science 142 (2017) 60–68 65

Table 3
Swelling ratio determined as a function of aspect ratio for all bentonite suspension deter-
mined for free swell test (βFSI) and oedometer swell test (βoedo) with adjusted values
based on ρd value of 0.9 g/cm3.

βFSI βoedo

Materials p=1 p=3 p = 50 p=1 p=3 p = 50

B1 25.2 25.2 4.5 4.3 4.3 0.8

B2 23.8 23.8 4.3 5.0 5.0 0.9
B3 21.0 21.0 3.8 4.8 4.8 0.9
B4 26.1 26.1 4.7 6.6 6.6 1.2
B5 5.1 5.1 0.9 2.6 2.6 0.5
B6 5.9 5.9 1.1 2.6 2.6 0.5

The natural Na bentonite B4 had the highest swelling final value,

Fig. 3. Swelling strain ε(t) as function of time for all bentonites, respecting their raw dry
density ρd, in contact with demineralized water.
5.2. Oedometer swell test
The Fig. 3 presents the volumetric strain as function of time, for all
bentonites. The observed swelling kinetic of the bentonites are similar
to what other researchers found for pure bentonite in GCL (Badv
and Farsimadan, 2009; Cho et al., 1999; Didier and Comeaga, 1997;
Sridharan and Gurtug, 2004) or CCL (Komine and Ogata, 1994). The
volumetric swelling strain ε as function of time t is hyperbolic and is
described by ε(t) = t/(a + bt) with a and b constants.
The initial dry density at the beginning of the test was determined
between 0.77 and 1.07 g/cm3. However, the swelling of saturated
clays depends on the dry density (Cho et al., 1999; Komine and Ogata,
1994; Sridharan and Gurtug, 2004). Hence, in order to describe the
swelling properties on a similar initial state basis for all bentonites, a
correction was applied on the final swelling strain values. The correction
is based on the proportional coefficient α of 2.623 cm3/g between the
volumetric swelling strain ε and the dry density ρd obtained by some
authors (Komine and Ogata, 1994) for clay specimens of 1.2 g/cm3
to 2 g/cm3 dry density under low vertical pressure (6 kPa). The
corrected values of infinite volumetric strain εmax and the swelling
ratio β are reported in the Fig. 4 and Table 3 for a corrected fixed dry
density ρd equal to 0.9 g/cm3, corresponding to the mean value for all
bentonites.
whereas the natural Ca bentonites B5 and B6 had the lowest. The Na-
activated Ca bentonites B1, B2 and B3 had intermediate swelling values
that are much higher than the natural Ca bentonites but lower than the
natural Na bentonite B4, similar to the findings of the FSI test (Fig. 2)
and of other authors (Komine, 2004). The close aspect ratio p = 1
and p = 3 give the same values for swelling ratio, different values are
obtained for the aspect ratio equal to 50.
5.3. Rheological test — dynamic frequency
Fig. 5 presents the results of dynamic frequency tests for the B1
bentonite suspension in five different concentrations. The values of
elastic modulus G′ are in the same range that other authors observed
(Khandal and Tadros, 1988). For low bentonite concentration (2%), the
parameter values were too low to be measured by the equipment.
Above 3% bentonite concentration, the elastic modulus G′ is constant
over the studied frequency range, similar to other researcher's results
for clay suspensions (Ramos-Tejada et al., 2001). The linear part of the
curve was extrapolated to determine the unique value of the free
frequency-dependent elastic modulus G′o for each bentonite concentra-
tion (and associated bentonite volume fraction). This rheological
response was observed for all tested bentonites.
The maximum effective packing fractions are determined from the
Fig. 6, by plotting the free frequency-dependent elastic modulus G′o as
function of the bentonite volume fraction (or the swelling ratio as in
Eq. (21) summarized in Table 4). The parameters n and α are obtained
by data fitting the experimental results (G′o vs. Φ) to the power law
equation (Eq. (19)). The power law exponent n varied around 3–10 as
measured by other authors (Khandal and Tadros, 1988; Larson, 1999;
Ramos-Tejada et al., 2001; Tadros, 1996). In Fig. 6, the curves are
conformed to the expected approach expressed in Eq. (19). The natural
Ca bentonites have the lowest βRDT. Since the rheological parameter G′

Fig. 4. Swelling ratio βoedo for oedometer swell test for all bentonites, respect their testing
dry density ρd (closed marks) and the fixed dry density ρd = 0.9 g/cm3 through correction Fig. 5. Elastic modulus G′ as function of the frequency for the B1 bentonite suspensions for
(open marked). Note that the determination is presented for spheres (p = 1). various clay concentrations in demineralized water.
66 G. Barast et al. / Applied Clay Science 142 (2017) 60–68

Fig. 6. Low frequency elastic modulus G′o as function of the swelling ratio βRDT and
bentonite fraction for the bentonite suspension in demineralized water. Note that the
determination is presented for spheres (p = 1). Fig. 7. Shear stress as function of shear rate for various clay concentrations of B1 bentonite
suspensions in demineralized water.

can represent the interactions energy between particles (Tadros, 1996),

the low values are expected for Ca-clay, B5 and B6. Low swelling parti-
cles are particles surrounded by weak electrostatic field forces. The
Na-activated Ca bentonite B1, B2 and B3 values are in the same narrow
range. It is not possible to establish a refined distinction between them.
The natural Na bentonite B4 possesses the higher value of swelling ratio
but slightly lower Φm in accordance to other authors (Paumier et al.,
2008) that found that homoionic saturated Na clay reaches lightly
lower maximum packing clay fraction than homoionic saturated Ca
clay due to their greater swelling.
5.4. Rheological test — steady shear
The Fig. 7 presents the results of steady shear tests for B1 bentonite
suspension with different concentrations. For low bentonite concentra-
tion (b 3%), the measured values are fitted by linear regression starting
from the origin. It expresses the well-known Newtonian behaviour of
low clay concentration suspension as observed by other researchers
(Paumier et al., 2008 and citations therein). The viscosity is determined
from the slope of the curve in good agreement with Newton's law. The
increase in clay concentration in the suspension changes the flow be-
haviour to progressively become pseudoplastic. For 3% bentonite sus-
pension, the data plotting is not rigorously linear and a quantifiable
yield stress σc of value 0.8 Pa appears. These curves fit better with
Bingham model or Herschel Buckley model, as observed also in other
works (Laribi et al., 2005; Kelessidis et al., 2007). Other authors
(Günster et al., 2006) observed the same rheological behaviour, within
the same shear stress and apparent viscosity range values, for their
Na-activated Ca bentonite tested from 1% to 6% mass concentration
with water. For higher clay concentration, the specimen becomes too
viscous to be properly set up in the testing geometry cup. An overshoot
peak of shear stress is observed. This suggests the transitory state
(Malkin and Isayev, 2006) from a gel-like material to a more solid-like
material with complex interactions. This rheological response was
observed for all tested bentonites.
The relative viscosity η/ηs of each bentonite, for low concentration is
deduced from the mean data and the formulae in Eq. (22). The Fig. 8
presents the relative viscosity as function of the clay fraction and con-
firms linear data fitting expected for Newtonian fluids (Eq. (22)). The
Table 4 summarizes the swelling ratio βRST for all bentonites, according
to Eq. (24). The data for each test was obtained for the following clay
fractions 0.19%, 0.38%, 0.56%, 0.75%, 0.93% and 1.12%, as shown in
Fig. 8. For the lowest clay fraction, the data accurately match the
Krieger–Dougherty equation due to the low torque values below the
equipment precision. For highest clay fraction, the bentonite behaviour
was not Newtonian any longer, as observed in Fig. 7. The data do not
match the Eq. (24). The interactions become too complex to be
modelled by the Krieger–Dougherty equation (Eq. (22)).
As shown in the Table 4, the natural Ca bentonites have the lowest
swelling ratio, and the B5 bentonite has the lowest value. The Na-
activated Ca bentonites (B1 and B2) possess the highest swelling ratio
(28.4 and 28.7), while the value for Na-natural bentonite (B4) is posi-
tioned in the same range (28.5). Several authors observed that the Na
clays have overall higher viscosity than Ca clays (Ece et al., 1999;
Gungor and Karaoglan, 2001). Other authors confirmed it for Newtoni-
an behaviours (Paumier et al., 2008; Radojević and Mitrovića, 2007). It
was also found that these statements are valid for Na-activated Ca ben-
tonites (Yildiz et al., 1999). The Na+ cation in the inner layer reduces the
free bulk volume, and thus increases the effective volume occupied by
clay. Although the distinction of performance between B1 and B2

Table 4
Swelling ratio as a function of aspect ratio for all bentonite suspension determined from
dynamic frequency test by asymptotic value (βRDT) and from steady shear test (βRST).

βRDT βRST

Materials p=1 p=3 p = 50 p=1 p=3 p = 50

B1 13.9 14.1 2.5 28.4 21.7 5.3

B2 13.9 14.1 2.5 28.7 22.1 5.4
B3 13.9 14.1 2.5 19.0 14.5 3.6
B4 16.8 17.1 3.0 28.5 21.8 5.3
Fig. 8. Relative viscosity η/ηs as function of the bentonite fraction for all bentonites
B5 6.1 6.2 1.1 12.0 9.0 2.3
suspension in demineralized water. Note that the determination is presented for
B6 8.6 8.7 1.5 14.7 11.0 2.7
spheres (p = 1).
 G. Barast et al. / Applied Clay Science 142 (2017) 60–68 67

bentonites was not observed, the bentonite B3 showed the lowest βoedo ¼ 0:17 βRST þ 0:73 with R2 ¼ 0:67 ð25bÞ
swelling ratio among the Na-activated Ca bentonites.

In addition, correlation curve between βFSI–βRST is located near to the

5.5. Comparison of swelling ratio
bisecting axis, with a parallel data line, giving more consistency to the
measured data. The correlation between βoedo–βRST is far from the corre-
The order of magnitude of the value of swelling ratio depends on the
sponding bisecting axis. This low correlation can be explained from the
type of test. The general distinction between the bentonite mineralogy
different states of tested specimens, that is suspension for rheology and
is highlighted by all tests. The approaches confirm that the natural Ca
solid state for oedometer.
clays B5 and B6 have the lowest value of swelling ratio, whereas the
Fig. 10, represents the variation of the swelling ratio measured by
natural Na bentonite B4 has the highest. The values of the β for Na-
rheology dynamic state βRDT as a function of oedometric swelling ratio
activated Ca bentonites are within the same range of intermediate
βoedo and the free swelling ratio βFSI. Two bisecting axes represent
values but no precise differentiation can be deduced.
the equality of values, on one hand, between oedometric swelling and
The βoedo are the lowest β among all swelling test ratio due to the
rheology dynamic state swelling ratio βoedo = βRDT, and on the other
loading and particles packing at the initial conditions in the testing
hand, between free swelling and rheology dynamic state swelling
cell. Results show a good match between the data plotted from the rhe-
ratio βFSI = βRDT.
ological tests and the FSI test. This is observed despise of the great dis-
Correlation curves between βoedo–βRDT and βFSI–βRDT give empirical
crepancy attributed to the equipment precision to measure very low
relations with correlations respectively equal to 0.9 and 0.92
and very high torque (Coussot, 2005) for the determination of the Φm.
(Eqs. (26b) and (26a)). However, the trends of the two curves are far
These observations are consistent with other authors that have found
from their respective bisecting axes. Thus, there is no direct correlation
linear relationships between a free swell index test value and the viscos-
between the three parameters, two by two.
ity parameters (Alther, 1986; Christidis, 1998). The βRST values are quite
similar. This implies that a simple and rapid test (free swell test) can
βFSI ¼ 2:33 βRDT −10:53 with R2 ¼ 0:92 ð26aÞ
predict the steady rheological behaviour of the bentonite suspension.
The values of swelling ratio βRDT given by the rheological dynamic test
are lower than βRST and βFSI, explained by the extension of clay particles βoedo ¼ 0:36 βRDT −0:08 with R2 ¼ 0:90 ð26bÞ
during the mixing preparation (Fig. 10).
The strong relationship between FSI and rheology are validated with
the results of βRST (0.87) and βRDT (0.92). The bisecting axis indicates a
5.6. Empirical correlations
perfect correlation between β obtained by two methods. For the
comparison of free swell versus rheology tests, one can observe that
Fig. 9 represents the variation of the swelling ratio measured by rhe-
the correlated data line, βFSI versus steady shear rheology test, is located
ology steady state βRST as a function of oedometric swelling ratio βoedo
near to the bisecting axis and almost parallel to this one.
and the free swelling ratio βFSI. Two bisecting axes represent the equal-
For the comparison of oedometer swelling ratio versus both rheolo-
ity of values, in one hand, between oedometric swelling and rheology
gy tests, the offset of data lines far from the bisecting axis suggest a re-
steady state swelling ratio βoedo = βRST, and on the other hand, between
lationship between the tests that is only partially described by the
free swelling and rheology steady state swelling ratio βFSI = βRST.
model. The difference between oedometric swelling result and rheology
Correlation curves between βoedo–βRST and βFSI–βRST give empirical
is explained by the nature of the two types of test performed in different
relations with high correlation (R2 near the value of 0.9) for free swell-
states: solid and liquid, respectively. The comparison chart confirms
ing index and rheology steady state (Eq. (25a)) whereas much lower
that the better correlation is obtained between βFSI and βRST.
correlation is calculated for oedometric swelling ratio and rheology
steady state (R2 = 0.67) (Eq. (25b)). This good correlation can be ex-
6. Conclusions
plained by the fact that both experiments are conducted on the same
state of clay that is a suspension.
The study investigated the application of rheology to describe the
swelling of different natures of bentonite using for geotechnical engi-
βFSI ¼ 1:20 βRST −8:29 with R2 ¼ 0:87 ð25aÞ neering applications. The different natures of bentonite can be divided

Fig. 9. Swelling ratio βFSI from free swell test and swelling ratio βoedo from oedometer test Fig. 10. Swelling ratio βFSI from free swell test and swelling ratio βoedo from oedometer test
as function of βRST from Rheological Steady Test for all bentonites suspension in as function of βRDT from Rheological Dynamic Test for all bentonites suspension in
Demineralized water. demineralized water.
68 G. Barast et al. / Applied Clay Science 142 (2017) 60–68
to three groups: the natural Na, the natural Ca and the Na-activated Ca
bentonites.
A simple model was proposed to unify the output parameters from
rheology tests at liquid state geomaterials with the ones from the swell-
ing tests at solid state and liquid state. This approach was based on a sin-
gle parameter: the swelling ratio β that describes the effective layer of
water surrounding the particles and that assumes the particles as spher-
oids with three different aspect ratios. The experimental investigation
was supported by several natures of bentonite in contact with water.
- The proposed parameter β established a relationship between free
swell and rheology tests and allowed the identification of the
group clay mineralogy. Several empirical correlations were extract-
ed from the proposed approach, with a strong confidence (R2 ~ 0.9).
- The swelling deformation for 5 kg/m2 of Ca and Na bentonites at ρd
equal to 0.9 g/cm3 can be estimated, regardless of the bentonite na-
ture. Swelling ratios are in good agreement with the nature which is
strongly linked to the mineralogy of bentonite. The lowest swelling
ratio corresponds to the natural Ca bentonites, while the highest
value refers to the Na bentonite and the intermediate swelling
ratio values, closest to natural Na bentonite, are for Na-activated Ca
bentonites.
- βFSI and βoedo are compared separately with βRST and βRDT. Both rhe-
ological tests are in good agreement with βFSI than that of βoedo, the
goodest being the rheology steady shear test.
- Steady rheology test fits better with free swelling index compared to
dynamic rheology tests. It would therefore be possible to estimate
the potential of final free swell deformation of a bentonite with
fast and simple index tests: free swelling index or steady rheology
tests.
- The observations resulting from the study provide additional sup-
port for the use of FSI to characterize the bentonite geomaterials.
The FSI test takes into account the behaviour of the material, in con-
trast to the oedometer test that considers the solid state earth struc-
ture. Since the rheology highlighted the swelling characteristics of
the bentonite materials, then it offers the possibility for measuring
the swelling of the bentonite improved by polymer. This will be
very useful for measuring the swelling indexes of high viscosity
polymer blended bentonites as the FSI reaches its limits for measur-
ing the swelling of this kind of mixtures.
Acknowledgments
Support for this study was provided by the Education and Research
Ministry (INSA GRH-2007-2010-09GB) and the Expora'Doc Rhône
Alpes Region, France, (08-020994-02) in the framework of a cotutelle
between INSA Lyon, France and the University of Queensland, Australia.
References
Alther, G.R., 1986. The effect of the exchangeable cations on the physico-chemical proper-
ties of Wyoming bentonites. Appl. Clay Sci. 1, 273–284.
Badv, K., Farsimadan, R., 2009. Swelling and diffusion characteristics of the experimental
GCLs. Iran. J. Sci. Technol. 33, 15–30.
Buscal, R., Goodwin, J.W., Hawkins, M.W., Ottewell, R.H., 1982. Viscoelastic properties of
concentrated lattices I. Methods of examination. J. Chem. Soc. Faraday Trans. 78,
2873–2887.
Cho, W.J., Lee, J.O., Chun, K.S., 1999. Basic physicochemical and mechanical properties of
domestic bentonite for use as a buffer material in a high-level radioactive waste re-
pository. J. Kor. Nucl. Soc. 31, 39–50.
Christidis, G., 1998. Physical and chemical properties of some bentonite deposits of
Kimolos Island. Greece, Appl. Clay Sci. 13, 79–98.
Couradin, A., Razakamanantsoa, A.R., Didier, G., Djeran-Maigre, I., 2008. Etude compara-
tive de performance hydraulique d'adjuvant argileux en traitement de sol. Journees
nationales de geotechnique et de geologie de l'ingenieur, JNGG’08, June 18th-20th,
Nantes, France, pp. 53–60.
Coussot, P., 2005. Rheometry of pastes suspensions and granular materials. Applications
in Industry and Environment. John Wiley and Sons, New York.
Didier, G., Comeaga, L., 1997. Influence of initial hydration conditions on GCL leachate
permeability. In: Well, L.W. (Ed.), Testing and Acceptance Criteria for Geosynthetic
Clay Liners, West Conshohocken, pp. 181–195.
Ece, O.I., Güngör, N., Alemdar, A., 1999. Influences of electrolytes polymers and a
surfactant on rheological properties of bentonite–water systems. J. Incl. Phenom.
Macrocycl. 33, 155–168.
Ghezzehei, T.A., Or, D., 2001. Rheological properties of wet soils and clays under steady
and oscillatory stresses. Soil Sci. Soc. Am. J. 65, 624–637.
Gleason, M.H., Daniel, D.E., Eykholt, G.R., 1997. Calcium and sodium bentonite for
hydraulic containment applications. J. Geotech. Geoenviron. 123, 438–445.
Goodwin, J.W., 2004. Colloids and Interfaces With Surfactants and Polymers — An Intro-
duction. John Wiley and sons, Chichester.
Gungor, N., Karaoglan, S., 2001. Interactions of polyacrylamide polymer with bentonite in
aqueous systems. Mater. Lett. 48, 168–175.
Günster, E., İşçi, S., Öztekin, N., Erim, F.B., Ece, O.I., Güngör, N., 2006. Effect of cationic sur-
factant adsorption on the rheological and surface properties of bentonite dispersions.
J. Colloid Interface Sci. 303, 137–141.
Guven, N., Pollastro, R.M., 1992. Clay–water Interface and Its Rheological Implications. The
Clay Minerals Society, Boulder.
Jo, H.Y., Katsumi, T., Benson, C.H., Edil, T.B., 2001. Hydraulic conductivity and swelling of
nonprehydrated GCLs permeated with single-species salt solutions. J. Geotech.
Geoenviron. 127, 557–567.
Katsumi, T., Ishimori, H., Onikata, M., Fukagawa, R., 2008. Long-term barrier performance
of modified bentonite materials against sodium and calcium permeant solutions.
Geotext. Geomembr. 26, 14–30.
Kelessidis, V.C., Tsarnantaki, C., Dalamanis, P., 2007. Effect of pH and electrolyte on the
rheology of aqueous Wyoming bentonite dispersions. Appl. Clay Sci. 38, 86–96.
Khandal, R.K., Tadros, T.F., 1988. Application of viscoelastic measurements to the investi-
gation of the swelling of sodium montmorillonite suspensions. J. Colloid Interface Sci.
125, 122–128.
Koch, D., 2002. Bentonites as a basic material for technical base liners and site encapsula-
tion cut-off walls. Appl. Clay Sci. 21, 1–11.
Komine, H., 2004. Simplified evaluation for swelling characteristics of bentonites. Eng.
Geol. 71, 265–279.
Komine, H., Ogata, N., 1994. Experimental study on swelling characteristics of compacted
bentonite. Can. Geotech. J. 31, 478–490.
Laribi, S., Fleureau, J.M., Grossiord, J.L., Kbir-Ariguib, N., 2005. Comparative yield stress de-
termination for pure and interstratified smectite clays. Rheol. Acta 44, 262–269.
Larson, R.G., 1999. The Structure and Rheology of Complex Fluids. Oxford University
Press, New York.
Malkin, A.Y., Isayev, A.I., 2006. Rheology: Concepts, Methods and Applications. ChemTec
Publishing, Toronto.
Markgraf, W., Horna, R., Petha, S., 2006. An approach to rheometry in soil
mechanics—structural changes in bentonite clayey and silty soils. Soil Tillage Res.
91, 1–14.
Mitchell, J.K., 1977. Fundamentals of Soil Behaviour. second ed. John Wiley and Sons, New
York.
Norrish, K., Quirk, J.P., 1954. Crystalline swelling of montmorillonite: use of electrolytes to
control swelling. Nature 173, 255–256.
Paumier, S., Pantet, A., Monnet, P., 2008. Evaluation of the organization of the homoionic
smectite layers (Na+ or Ca2+) in diluted dispersions using granulometry microscopy
and rheometry. Adv. Colloid Interface 141, 66–75.
Philipse, A.P., 1996. The random contact equation and its implications for (colloidal) rods
in packings, suspensions, and anisotropic powders. Langmuir 12, 1127–1133.
Radojević, Z., Mitrovića, A., 2007. Study of montmorillonite and cationic activators system
rheological characteristic change mechanism. J. Eur. Ceram. Soc. 27, 1691–1695.
Ramos-Tejada, M.M., Arroyo, F.J., Perea, R., Duran, J.D.G., 2001. Scaling behaviour of the
rheological properties of montmorillonite suspensions: correlation between interpar-
ticle interaction and degree of flocculation. J. Colloid Interface Sci. 235, 251–259.
Razakamanantsoa, R., 2009. Etudes du comportement hydromécanique, chimique et de la
durabilité des géomatériaux d’étanchéité melangés avec des polymères, Ph D thesis.
INSA Lyon.
Razakamanantsoa, R., Barast, G., Djeran-Maigre, I., 2012. Hydraulic performance of
activated calcium bentonite treated by polyionic charged polymer. Appl. Clay Sci.
59-60, 103–114.
Russel, W.B., Saville, D.A., Schowalter, W.R., 1989. Colloidal Dispersions. Cambridge
University Press, Cambridge.
Shackelford, C.D., Benson, C.H., Katsumi, T., Edil, T.B., Lin, L., 2000. Evaluating the hydraulic
conductivity of GCLs permeated with non-standard liquids. Geotext. Geomembr. 18,
133–161.
Sridharan, A., Gurtug, Y., 2004. Swelling behaviour of compacted fine-grained soils. Eng.
Geol. 72, 9–18.
Tadros, T.F., 1996. Correlations of viscoelastic properties of stable and flocculated suspen-
sions with their interparticle interactions. Adv. Colloid Interface 68, 97–200.
Van Olphen, H., 1977. An Introduction to Clay Colloid Chemistry. second ed. John Wiley
and Sons, New York.
Woods, M.E., Krieger, I.M., 1970. Rheological studies on dispersions of uniform colloidal
spheres: I. Aqueous dispersions in steady shear flow. J. Colloid Interface Sci. 34,
91–99.
Wouterse, A., Williams, S.R., Philipse, A.P., 2007. Effect of particle shape on the density and
microstructure of random packings. J. Phys. Condens. Matter 19, 1–14.
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 ben-
tonite. Appl. Clay Sci. 14, 319–327.