On the hydro-mechanical behaviour of MX80 bentonite-based materials

Accepted Manuscript On the hydro-mechanical behaviour of MX80 bentonite-based materials Yu-Jun Cui

PII:

S1674-7755(16)30055-5

DOI:

10.1016/j.jrmge.2016.09.003

Reference:

JRMGE 275

To appear in:

Journal of Rock Mechanics and Geotechnical Engineering

Received Date: 15 July 2016 Revised Date:

1 September 2016

Accepted Date: 4 September 2016

Please cite this article as: Cui Y-J, On the hydro-mechanical behaviour of MX80 bentonitebased materials, Journal of Rock Mechanics and Geotechnical Engineering (2016), doi: 10.1016/ j.jrmge.2016.09.003. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

On the hydro-mechanical behaviour of MX80 bentonite-based materials

RI PT

Yu-Jun Cui

Ecole des Ponts ParisTech, Laboratoire Navier/CERMES, 6 – 8 av. Blaise Pascal, Cité Descartes, Champs – sur – Marne, 77455 Marne – la – Vallée cedex 2, France

AC C

EP

TE D

M AN U

SC

Telephone: +33 1 64 15 35 50; Fax: +33 1 64 15 35 62; E-mail: [email protected]

1

ACCEPTED MANUSCRIPT Abstract Bentonite-based materials have been considered in many countries as engineered barrier/back-filling materials in deep geological disposal of high level radioactive waste. During the long period of waste storage, these materials will play an essential role of ensuring

RI PT

the integrity of the storage system that consists of the waste canisters, the engineered barrier/back-fill, the retaining structures as well as the geological barrier. Thus, it is essential

SC

to well understand the hydro-mechanical behaviour of these bentonite-based materials. This review paper presents the recent advances of knowledge MX80 bentonite-based materials, in

M AN U

terms of water retention properties, hydraulic behaviour and mechanical behaviour. Emphasis is put on the effect of technological voids and the role of the dry density of bentonite. The swelling anisotropy is also discussed based on the results from swelling tests with measurements of both axial and radial swelling pressures on a sand/bentonite mixture

TE D

compacted at different densities. Microstructure observation was used to help the interpretation of macroscopic hydro-mechanical behaviour. Also, the evolution of soil

EP

microstructure thus the soil density over time is discussed based on the results from mock-up tests. This evolution is essential for understanding the long-term hydro-mechanical behaviour

AC C

of the engineered barrier/back-fill.

Key words: Bentonite-based materials; water retention; hydraulic conductivity; mechanical behaviour; microstructure; dry density evolution

2

ACCEPTED MANUSCRIPT INTRODUCTION

Deep geological disposal has been adopted in many countries for the high level radioactive waste (HLW). Mostly, multi-barrier concept is adopted, including the canisters, the natural

RI PT

geological barrier (host rock) and the engineered barriers made up of compacted bentonite-based materials. When backing filling is needed, the bentonite-based materials are

SC

also used for this purpose. The use of bentonite-based materials in engineered barrier/back-filling is justified by their low permeability, high swelling and high radionuclide

Watanabe 2010, Cui et al. 2011).

M AN U

retardation capacities (Pusch 1979, Yong et al. 1986, Villar and Lloret 2008, Komine and

Different bentonites have been studied over the world, among which we can cite the most

TE D

famous: MX80 used in the Sweden program (Pusch et al. 2001a) and considered in the French program (Wang et al. 2012), FoCa7 that was considered initially in the French program (Cui et al. 2002, Imbert and Villar 2006), FEBEX in the Spanish program (Villar 2002), Kunigel

EP

V1 considered in the Canadian program (Dixon et al. 1996) and GMZ considered in the

AC C

Chinese program (Ye et al. 2009). Among these bentonites, only FoCa7 is a calcium bentonite, others being sodium bentonite. The general trend in the field of deep geological disposal of high level radioactive waste is the abandon of calcium bentonite and the use of sodium bentonite. The main reason is that the latter is more expansive than the former. Ben Rhaïem et al. (1987) investigated the structures of a calcium bentonite gel and a sodium bentonite gel and observed more numerous and smaller pores in the latter, suggesting a higher hydration capacity thus a higher swelling capacity of sodium bentonite. This is also related to the lower

3

ACCEPTED MANUSCRIPT hydration energy of exchangeable monovalent cation such as Na+ as compared with divalent cations such as Ca++. Sposito (1984) reported a solvation complex made up of a single water shell of 6 water molecules surrounding a monovalent cation, but a solvation complex made up

RI PT

of 2 water shells composed of 8 and 15 water molecules surrounding a divalent cation, respectively. Marcial et al. (2002) investigated the compression behaviours of calcium FoCa7 bentonite and sodium MX80 bentonite, and observed that FoCa7 presents a larger void ratio

SC

than MX80 under high vertical stresses, confirming the larger diameters of hydration shells of

M AN U

divalent cations like Ca++.

For the engineered barrier/back filling, either pure bentonite or bentonite/sand mixture can be used. Sometimes, the mixture of bentonite and crashed in situ geological barrier material is considered in order to reduce the excavation waste on one hand and to better ensure the

TE D

compatibility of chemistry between the engineered barrier/back filling and the geological barrier on the other hand. Figure 1 shows the grain size distribution curves of MX80 and a

EP

sand used to form the mixture considered in the French program (Wang et al. 2013a), and Figure 2 shows the grain size distribution curves of MX80 and crashed Collovo-Oxfordian

AC C

(Cox) claystone used to form a mixture also considered in the French program (Wang et al. 2012).

The initial state of such bentonite based materials are usually controlled by the maximum swelling pressure - this welling pressure must be lower than the in situ minor stress. For instance, this minor stress is 7 MPa in the Underground Research Laboratory (URL) of Bure site, France (Delege et al. 2010, Tang et al. 2011a, 2011b); 3-4 MPa at Tournemire site,

4

ACCEPTED MANUSCRIPT France (Barnichon and Deleruyelle 2009) and 4-5 MPa at Mol site, Belgium (Li et al. 2010). The use of bentonite/sand mixture is often preferred for the reason of increasing heat transfer and mechanical strength.

RI PT

The compacted bentonite-based materials has been widely investigated, in terms of swelling properties (Pusch 1982, Komine and Ogata 1994, 2003, 2004, Delage et al. 1998, Agus and

state (Kenney et

SC

Schanz 2005, Komine et al. 2009, Ye et al. 2009, 2013a ), hydraulic behaviour under saturated al. 1992, Dixon et al. 1999, Komine 2004) and unsaturated state

M AN U

(Borgesson et al. 1985, 2001, Kröhn 2003, Lemaire el al. 2004, Loiseau et al. 2002, Cui et al. 2008, Ye et al. 2014a). There are also many studies on the effect of temperature (Komine and Ogata 1998, Romero et al. 2001, Villar and Lloret 2004, Tang and Cui 2005, Tang et al. 2007, 2008, Tang and Cui 2009, Ye et al. 2013b, 2014b, Wan et al. 2015) and water chemistry of the

TE D

saturating fluid on the hydro-mechanical (HM) behaviour (Pusch 2001b, Karnland et al. 2005, Suzuki et al. 2005, Komine et al. 2009, Zhu et al. 2013, 2015, Ye et al. 2014c, Chen et al.

EP

2015). The effect of technological voids existing either between the bricks of compacted bentonite-based materials or between the bricks and the canisters/the host rock were

AC C

investigated by Wang et al. (2013a) and the swelling anisotropy was studied by Saba et al. (2014a).

In this paper, a review of the recent advances of knowledge on the hydro-mechanical behaviour of MX80 bentonite-based materials is presented, starting with the water retention properties, then the hydraulic behaviour, and finishing by the mechanical behaviour. Emphasis is put on the effects of technological voids and the final bentonite density. The swelling

5

ACCEPTED MANUSCRIPT anisotropy and the evolution of microstructure are also discussed based on the mock-up test

WATER RETENTION PROPERTIES

RI PT

results.

The mixture of MX80 Na-bentonite and a sand was considered, in a proportion of 70/30 in

SC

dry mass. This bentonite is characterised by a high montmorillonite content (80%), a liquid limit of 575%, a plastic limit of 53% and a unit mass of the solid particles of 2.77 Mg/m3. The

M AN U

cation exchange capacity (CEC) is 76 meq/100g (83 % of Na). The mixture was compacted statically to a desired density ρd = 1.67 Mg/m3, corresponding to the final dry density adopted in the in situ experiment at Tournemire URL by the French Institution for Radiation Protection and Nuclear Safety (IRSN). To investigate the effect of compaction to the soil

TE D

microstructure, two samples of different dry densities (ρd = 1.67 and 1.97 Mg/m3) were prepared by compaction and freeze dried. Their microstructures were then observed through

EP

mercury intrusion porosimetry (MIP) tests (Wang et al. 2013a). The results obtained are

AC C

presented in Figure 3. A typical bimodal porosity can be observed in both samples, defining intra-aggregate pores (micro-pores) with a mean size of 0.02 µm that is independent of the soil dry density, and inter-aggregate pores (macro-pores) that depend on the soil dry density 10 µm for ρd = 1.67 Mg/m3 and 50 µm for ρd = 1.97 Mg/m3. This is consistent with the observation by Delage and Graham (1995) - compaction only affects the largest inter-aggregate pores while intra-aggregate pores remain unaffected.

6

ACCEPTED MANUSCRIPT The water retention properties of samples compacted at ρd = 1.67 Mg/m3 were determined through controlling suctions using both vapor equilibrium method for high suctions and osmotic method for low suctions (Wang et al. 2013a). Both free swell condition and restrained

RI PT

swell (constant volume) condition were considered. The results presented in Figure 4 shows that for suctions higher than 9 MPa, the two curves are very similar while a significant difference can be identified in the range of suction below 9 MPa. When suction reached

SC

0.01 MPa, the water content under free swell condition is 246%, much higher than that under

M AN U

restrained swell condition (25.4%). This confirms that the restrained swelling condition significantly affects the retention property only in the range of low suctions (Yahia-Aissa et al. 2001, Cui et al. 2008, Ye et al. 2009). For further analysis, the curves determined by Marcial (2003) for pure MX80 bentonite compacted at ρd = 1.7 Mg/m3 under both free swell and

TE D

constant volume conditions are also presented in Figure 4. It appears that in the high suction range (s > 9 MPa), all data fall on the same curve, regardless of the test conditions, soil type and soil density. By contrast, at lower suctions (s < 9MPa), the water content of the mixture

EP

under free swell condition is lower than that of the pure bentonite at the same suction.

AC C

Normally in a bentonite/sand mixture, all water is contained in bentonite because sand grains are not sensitive to water. This can be verified using a parameter namely water volume ratio (ew) defined as the ratio of water volume (Vw) to the bentonite volume (Vbs) (Romero et al. 2011). This parameter can be determined using the following equation:

ew =

wm G sb B

(1)

7

ACCEPTED MANUSCRIPT where wm is the water content of the mixture and Gsb is the specific density of bentonite.

Figure 5 shows the changes of ew of both the mixture and the pure bentonite with suction under free swell condition. Unlike in the water content/suction plot (Figure 4), an excellent

RI PT

agreement between the mixture and the pure bentonite is obtained, with a unique relationship between ew and suction. This confirms that water was only adsorbed in the bentonite (volume

M AN U

related to the lower volume of bentonite in the mixture.

SC

Vbs) and that the lower water content observed in the mixture at same suction (Figure 4) is

The results from the wetting tests on pure compacted MX80 bentonite under restrained swell condition with different void ratios (Marcial 2003, Tang and Cui 2010, Villar 2005) are also presented in Figure 5. All results are consistent and confirm that the confining conditions

TE D

affect the water retention property in the range of low suctions (<9 MPa), but not in the range of high suctions. Similar observations were made by Agus (2005) and Agus et al. (2010). Cui et al. (2002, 2008) and Ye et al. (2009) explained this phenomenon by the limited exfoliation

EP

of clay particles from the aggregates into inter-aggregate pores at high suctions, but

AC C

significant exfoliation of clay particles at lower suctions. In case of limited exfoliation, water adsorption is limited, explaining the same curve for both free swell and restrained swell conditions. By contrast, in case of significant exfoliation, the water intake is governed by the volume of initial macro-pores in the case of restrained condition because only the initial macro-pores can host exfoliated clay particles, while water intake is governed by both the initial macro-pores and the wetting-induced swell in the case of free swell condition.

8

ACCEPTED MANUSCRIPT HYDRAULIC BEHAVIOUR The hydraulic conductivity of unsaturated soils can be determined using the instantaneous profile method, which requires both suction and volumetric water content profiles (Daniel

RI PT

1982, Cui et al. 2008, Ye et al. 2009, Wang et al. 2013b). Very often, the suction profile is obtained by suction monitoring in an infiltration test, whereas the volumetric water content profile is deduced from the water retention curve determined separately. Figure 6 shows the

SC

relationship of hydraulic conductivity and suction for the bentonite/sand mixture compacted

M AN U

at ρd = 1.67 Mg/m3. The hydraulic conductivity determined by Karnland et al. (2008) for the same material at the same dry density (1.67 Mg/m3) under saturated state is also presented (10-13 m/s). Unlike the non-swelling soils for which the hydraulic conductivity is constantly increasing upon wetting, a decrease followed by a decrease is observed: the hydraulic

TE D

conductivity decreases with suction decreasing from 65 to about 15 MPa, then increases with further suction decrease. Similar observations were made on the Kunigel-V1/sand mixture

EP

(Cui et al. 2008) and GMZ bentonite (Ye et al. 2009).

AC C

The results of MIP tests shown in Figure 7 (Wang et al. 2013c) illustrate that suction decrease did not change the micro-pores family, whereas a decrease of macro-pores quantity occurred and this decrease was particularly significant when suction decreased from 65 MPa to 38 MPa and much less for the subsequent suction decrease (38-24.8-12.6 MPa). This supports the explanation by Cui et al. (2008) and Ye et al. (2009) that during suction decrease the macro-pores were clogged by the exfoliated clay particles whose interlaminar distance increases upon wetting.

9

ACCEPTED MANUSCRIPT The micro-pores changed only when saturation was approached (1 MPa and 0 MPa in Figure 7). This phenomenon is in agreement with the conclusion by Cui et al. (2002) for the Kunigel clay/sand mixture - the hydration of clay aggregates started first on the aggregate surfaces

RI PT

giving rise to the exfoliation of clay particles and resulting in the clogging of large-pores, whereas the small-pores remained almost un-affected; the small-pores started to change only when saturation was approached (suction lower than 9 MPa). When saturation was reached, a

SC

new pore family of about 0.4 µm diameter was formed. Audiguier et al. (2008) called this

M AN U

family a 2-D pore family, and attributed it to the division of clay particles within the aggregates.

In order to quantify the suction effect on the large-pores, the pores larger than 2 µm are defined as large-pores (see Figure 7). Using this criterion, the void ratio corresponding to the

TE D

large-pores was derived from the cumulative curves (Figure 7a). The relationship between the calculated large-pore void ratio (eL) and the applied suction are presented in Figure 8. It can

EP

be observed that the large-pores quantity was progressively reduced with decreasing suction in the range of 65 to 12.6 MPa (zone I). However, the subsequent decrease of suction to 4.2

AC C

MPa led to an increase of large-pores quantity (zone II). This can be explained by the creation of the 2-D pores described above. A significant decrease of large-pore void ratio occurred when suction decreased from 1 MPa to zero (zone II).

Comparison between the large-pore changes (Figure 8) and the hydraulic conductivity changes (Figure 6) shows that the hydraulic conductivity changed following the same trend as the large-pore void ratio. This suggests that water transfer was primarily governed by the

10

ACCEPTED MANUSCRIPT network of large-pores in the full suction range. The large-pores decreased progressively in quantity due to the clay particle exfoliation in the suction range from 65 to 12.6 MPa, leading to a decrease of hydraulic conductivity. This exfoliation was identified by Cui et al. (2002) on

RI PT

compacted Kunigel V1 bentonite using scanning electron microscope (SEM). When saturation was approached (4.2 - 1 MPa), water transfers through the 2-D pores appeared, leading to an increase in hydraulic conductivity. In saturated state the hydraulic conductivity

SC

reached as high as 10-13 m/s. It should be noted that after saturation, the hydraulic

M AN U

conductivity has been found to be time dependent (Loiseau et al. 2002, Cui et al. 2008). It decreased with time due to the re-organization of microstructure over time in the soil sample. After saturation, water in the inter-aggregate pores is not necessarily in equilibrium with the water inside the aggregates in terms of water potential. Thus, the water re-distribution

al. 2010).

TE D

occurred inside the soil, leading to a uniform microstructure in long term (Stroes-Gascoyne et

EP

In case of the presence of technological voids (Wang et al. 2013a), the hydraulic conductivity can be not uniform on a cross section. Indeed, the values of global hydraulic conductivities

AC C

presented in Figure 9 showed possible preferential flow pathway in the case of heterogeneous samples with initial technological voids. Note that the bentonite void ratio (eb) is defined as follows:

eb = ebi + etech

ebi =

Vi Vbs

etech =

Vtech Vbs

(2) (3)

(4)

11

ACCEPTED MANUSCRIPT where ebi is the intra-bentonite void ratio inside the soil and etech is the void ratio corresponding to the technological void, Vbs is the bentonite particle volume and Vi is the intra-void volume, Vtech is the volume of technological void.

RI PT

In Figure 9, the data of Wang et al. (2013a) are compared with other data of pure MX80 bentonite (Karnland et al. 2008, Dixon et al. 1996) and of the 70/30 bentonite-sand mixture

SC

(Gatabin et al. 2008). It appears clearly that in case of the presence of initial technological voids, the hydraulic conductivity is higher. Furthermore, the effect of sand seems to be

M AN U

negligible. In fact, the higher conductivity is due to the looser zone of gel corresponding to the initial technological void. This is a weak zone with poorer mechanical resistance, at least at the short term. Further studies are needed to investigate the long term change in hydraulic

TE D

conductivity of samples with initial technological voids.

MECHNAICAL BEHAVIOUR

EP

Figure 10 shows the variations of swelling pressure of a mixture of MX80/ and crashed Cox

AC C

claystone (see Figure 2) with the final dry density of bentonite. In the same figure, the results of Karnland et al. (2008) on the MX80/sand mixture, Dixon et al. (1996) and Karnland et al. (2008) on pure MX80 bentonite are also presented. A unique relationship is obtained, showing that the swelling pressure solely depends on the final dry density of bentonite, sand and claystone being both inert components in the mixtures for the swelling pressure development. Note that the dry density of bentonite (ρdb) is defined as follows:

12

ACCEPTED MANUSCRIPT ρ db =

( B / 100) ρ m Gsa Gsa (1 + wm / 100) − ρ m (1 − B / 100)(1 + wa )

(5)

where ρm (Mg/m3) is the mixture density, B (%) is the bentonite content (in dry mass) in the

RI PT

mixture, wm is the water content of the mixture, wa is the initial water content of coarse materials (claystone or sand), Gsa is the specific gravity of coarse materials.

SC

For further analysis, the data for other bentonites are collected in Figure 11 showing the variations of swelling pressure with the final dry density of MX80 (from Figure 10), FoCa7

M AN U

(Imbert and Villar 2006), FEBEX (Villar 2002) and Kunigel V1 (Dixon et al. 1996). A unique relationship is identified for each bentonite, with however different slopes: the largest for FoCa7 and the lowest for Kunigel V1. This illustrates that the relationship between swelling

TE D

pressure and final dry bentonite density depends on the soil mineralogy.

The swelling pressure is thus controlled by the final density of bentonite. But what happens in

EP

long term considering different possible interactions between different minerals and fluids in the mixture? In Figure 12 the results from three tests LT01, LT02 and LT03 on MX80/Cox

AC C

claystone mixture during 700 days, with synthetic water, distilled water and vapour-distilled water respectively are presented (Wang et al. 2014). It appears that over such long period, the swelling pressure decreased slightly in all tests, especially for the sample saturated with synthetic water (LT01). The final swelling pressure was 3.95 MPa (corresponding to a decrease of 9%), and 4.19 MPa (decrease of 3%) for samples saturated with synthetic water (LT01) and distilled water (LT02 and LT03), respectively.

Basically, cation exchange between sodium and calcium may take place in bentonite, 13

ACCEPTED MANUSCRIPT transforming sodium montmorillonite to calcium montmorillonite. This transformation leads to a decrease of swelling capacity and is strongly dependent on the amount of available calcium (Ca2+) in the pore water (Muurinen and Lehikoinen 1999, Fernández and Villar 2010).

RI PT

For the sample saturated with distilled water, the pore water chemistry was being enriched by interaction with the Cox claystone minerals (the Cox claystone contains 20–30% carbonates, mainly calcite), and it was as if diluted Cox claystone site water infiltrated to the MX80

SC

bentonite, leading to the degradation of swelling pressure. However, with a much lower cation

M AN U

concentration (only 30% claystone in the mixture), the decrease of swelling pressure was less significant than in the case with synthetic water.

The pore size distribution curves for all the samples taken at the end of tests are presented in Figure 13, including a sample after a short term of 100 h saturation with synthetic water (ST).

TE D

It is observed clearly in Figure 13a that the amount of accessible porosity after 700 days is larger in both cases of distilled water (LT02 and LT03) and synthetic water (LT01) than after

EP

100 hours with synthetic water (ST). In addition, for samples hydrated for 700 days, more

AC C

quantity of accessible porosity is observed in the case of synthetic water (LT01).

Figure 13b shows the density function. For all samples a typical bimodal porosity can be identified, with a population of micro-pores of 0.02 µm and a population of macro-pores of 20

µm. After 700-day saturation (LT01, LT02 and LT03), the macro-pores and micro-pores quantity increased when comparing to the short-term case (ST). Change in macro-pores is relatively more significant in particular for the sample saturated with synthetic water (LT01). As far as changes in pores size are concerned, it can be observed that over time the population

14

ACCEPTED MANUSCRIPT of micro-pores has a size slightly decreased. On the contrary, the population of macro-pores has a size slightly increased.

Comparison between Figure 12 and Figure 13 shows that the more significant the swelling

RI PT

pressure decrease, the larger the amount of accessible porosity. For samples with higher swelling capacity, more interlayer hydration occurred, leading to a constriction of accessible

SC

pores. Both the macro-pores and micro-pores were involved, but most changes in microstructure occurred in macro-pores population. This is in agreement with the observation

M AN U

made by Mata (2003) on compacted sodium bentonite MX80 and sand mixture (70/30) saturated with distilled water and saline water.

Wang et al. (2013a) investigated the yield behaviour of compacted MX80/sand mixture in

TE D

oedometer and the results are presented in Figure 14. Relationship between yield stress and bentonite void ratioIt is observed that for both the mixture and pure bentonite, the yield stress increases sharply with decreasing bentonite void ratio. However, the curve of the mixture lies

EP

in the right of that of pure bentonite, evidencing the role of sand in the compression behaviour.

AC C

It appears that at the same bentonite void ratio, the mixture yields at a higher yield stress.

DISCUSSION

The swelling pressure of bentonite-based materials has been found to be controlled by the final dry density of bentonite: the higher the final dry density of bentonite, the higher the swelling pressure. The relationship between swelling pressure and final dry density of bentonite depends on the mineralogy of bentonite. In short term, the effect of pore water 15

ACCEPTED MANUSCRIPT chemistry has been found to be negligible. By contrast, in long term this effect can be significant: the higher the salinity (in calcium) of pore water the larger the decrease of swelling pressure. It is worth noting that these observations have been made by considering

RI PT

one direction (axial direction usually) swelling. However, in reality the swelling behaviour of soils can be anisotropic, depending on the arrangements of clay particles/aggregates. Saba et al. (2014a) investigated this swelling anisotropy using a cell with both axial and radial

SC

swelling pressure measurements. Different dry densities of MX80/sand mixture (70/30 in dry

M AN U

mass) was considered, and the results allowed an anisotropy coefficient Ca to be determined for each dry density. Note that Ca is defined as the ratio of radial swelling pressure to axial swelling pressure. This coefficient is presented versus the bentonite dry density in Figure 15. It can be observed that, in the range of medium bentonite dry densities, from 1.16 Mg/m3 to

TE D

1.3 Mg/m3, the swelling behaviour is mostly isotropic as the anisotropy coefficient is close to 1. At higher bentonite dry densities, the anisotropy coefficient becomes slightly smaller than 1 with values ranging between 0.90 and 0.76, indicating an anisotropic swelling behaviour.

EP

When the bentonite dry density is lower than 1.1 Mg/m3, the anisotropy becomes higher with

AC C

coefficients varying from 0.82 to 0.48. The sample preparation mode (In-cell or Out-cell) does not seem to affect the results. The anisotropy behaviour is also related to the microstructure changes of the compacted mixture. After the uniaxial compaction of the grains-powder mixture, an initial structural anisotropy is induced by the compaction process that results in a larger swelling reaction in the compaction direction. On the other hand, upon hydration, bentonite grains swell and split up, filling macro-pores. This process leads to a microstructure collapse, decreasing the initial 16

ACCEPTED MANUSCRIPT anisotropy. In the case of medium range of bentonite dry densities, collapse upon wetting was found to be significant, favouring the development of an isotropic microstructure. This explains the isotropic behaviour identified. In the case of high bentonite dry densities, a

RI PT

limited collapse of microstructure can be expected because of the macro-pore space is relatively smaller. This explains why the material remained anisotropic. In the case of very low bentonite dry densities (< 1.15 Mg/m3), the bentonite grains can swell but cannot fill up

SC

all the voids and thus microstructure collapse cannot take place. The initial anisotropy is then

M AN U

preserved.

The presence of technological voids makes compacted bentonite-based materials non-homogeneous after wetted. Even without technological voids this kind of materials can be also heterogeneous because of the non-homogenous wetting – the part close to the wetting

TE D

face swells first, compressing the rest. This is to say that even though the whole sample is fully saturated, it can be far from being homogenous. After saturation, a homogenisation

EP

process will take place over time, making the sample more and more homogenous. This phenomenon was investigated by Saba et al. (2014b) through the evolution of soil dry density.

AC C

Figure 16 shows the evolution of dry density profile in an infiltration column. The mean initial dry density is the one measured after compaction (1.69 Mg/m3). During the infiltration test, the sample was first wetted under constant volume condition with monitoring of radial swelling pressure at different heights. Then the piston was released, allowing 20% axial deformation. Afterwards, the piston was blocked again to ensure the constant volume condition. More details can be found in Saba et al. (2014b). The estimated final dry density is also plotted in Figure 16. By considering the swelling pressure profile at 58 days before the 17

ACCEPTED MANUSCRIPT piston release as the final swelling pressures, and the values at 120 days as the final swelling pressures after the piston release, the bentonite dry density profile was estimated based on the relationship between swelling pressure and final bentonite dry density shown in Figure 10. It

RI PT

appears after 58 days the dry density profile is close to the initial one. This is normal because the sample volume was kept constant. As the dry density values were estimated from the measured swelling pressures, the relatively uniform dry density suggests uniform swelling

SC

pressure. In turn, the uniform swelling pressure suggests full saturation of the sample. After

M AN U

the piston release, the dry density of the upper part of sample decreased as indicated by the profile at 120 days. In order to further analyse the evolution of dry density profile, a previous similar mock-up test conducted by Wang et al. (2013d) on the same material with similar test condition was considered. The test was dismantled at 350 days and dry density profile was

TE D

determined by measurements. It is observed that after a long period of 350 days the dry density profile is much closer to the expected final one, but the sample was still not homogeneous and a density gradient still existed along the sample. This clearly shows that

EP

over time the homogenisation of soil microstructure took place, but this process was quite

AC C

long.

CONCLUSION

Bentonite-based materials have been widely considered as engineered barrier/back filling materials in deep geological disposal of high level radioactive waste. Both sodium and calcium bentonites have been investigated over the world, but the current general trend is abandon of calcium bentonite and preference for the sodium one. This is mainly due to the 18

ACCEPTED MANUSCRIPT high swelling capacity of the sodium bentonite. The coarse materials used to form the mixture with bentonite can be either sand or crushed in situ materials. It appeared that with crushed in situ materials the water salinity (in calcium) is

RI PT

increased, leading to a decrease of swelling pressure in long term. But this decreased swelling pressure reflects what will happen in a real geological repository. Indeed, with infiltration of site water, the bentonite swelling pressure is expected to decrease. In this regard, it is

SC

important to determine the swelling pressure in the laboratory using site water.

M AN U

In bentonite-based materials, water is contained in bentonite as indicated by the water retention curve in the plot of water volume ratio versus suction. The confining conditions (free swell or restrained swell) affect the water retention curve in the low suction range. At higher suction range, the water retention curves are independent of the confining conditions

TE D

because of the limited clay exfoliation phenomenon. The hydraulic conductivity upon wetting decreases first and then increases with suction

EP

decrease. This is consistent with the variation of macro-pore volume identified by the MIP tests. During wetting, the macro-pore volume decreases first, evidencing a pore clogging

AC C

phenomenon and inducing a decrease of hydraulic conductivity. When the suction becomes quite low, the macro-pore volume starts to increase due to the creation of 2-D pores. This leads to an increase of hydraulic conductivity. Using sand as coarse element to form the mixture with bentonite, in addition to the well-known advantage of increasing the thermal conductivity of the mixture, there is the advantage of mechanical strength increase as illustrated by the higher yield stress of

19

ACCEPTED MANUSCRIPT compacted bentonite/sand mixture as compared to that of compacted pure bentonite. For a given pore water salinity (in calcium), the swelling pressure of a bentonite-based material depends solely on the final bentonite dry density. It is worth noting that the swelling

collapse upon wetting decreases the initial anisotropy of soil.

RI PT

pressure can be anisotropic, depending on the sample state. In general, microstructure

SC

Even though under fully saturated conditions, the microstructure still evolves as illustrated the changes of dry density profile in an infiltration test. This evolution corresponds to a

ACKNOWLEDGEMENTS

M AN U

homogenisation process and can take long time.

TE D

The author would like to address his acknowledgements to his co-workers: Dr Qiong Wang, Dr Anh Minh Tang, Dr Simona Saba and Prof. Pierre Delage at Ecole des Ponts ParisTech,

AC C

REFERENCES

EP

and Dr Jean-Dominique Barnichon at IRSN.

Agus S.S. 2005. An Experimental study on hydro-mechanical characteristics of compacted bentonite-sand mixtures. PhD thesis. Weimar. Agus S.S., Schanz T. 2005. Swelling pressures and wetting-drying curves of a highly compacted bentonite-sand mixture. Unsaturated Soils: Experimental Studies, 241–256. Agus S.S., Schanz T., Fredlund D.G. 2010. Measurements of suction versus water content for bentonite-sand mixtures. Can. Geotech. J 47, 583-594. Audiguier M., Geremew Z., Cojean R. 2008. Relations enter les microstructures de deux sols argileux de la région parisienne et leur sensibilité au retrait gonflement. SEC2008, Paris, France, 1-3 sept. Editions du LCPC.2008, 235-243. Barnichon J.D., Deleruyelle F. 2009. Sealing Experiments at the Tournemire URL. Towards convergence of technical nuclear safety practices in Europe, EUROSAFE.

20

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

Ben Rhaïem H., Pons C.H., Tessier D. 1987. Factors affecting the microstructure of smectites. Role of cations and history of applied stresses. In Proceedings of the International Clay Conference, Denver. The Clay Mineral Society, 292–297. Borgesson L. 1985. Water flow and swelling pressure in non-saturated bentonite-based clay barriers. Engineering Geology 21(3-4), 229–237. Borgesson L., Chijimatsu M., Fujita T., Nguyen T.S., Rutqvist J., Jing L. 2001. Thermo-hydro-mechanical characterisation of a bentonite-based buffer material by laboratory tests and numerical back analyses. International Journal of Rock Mechanics and Mining Sciences 38(1), 95–104. Chen Y.G., Zhu C.M., Ye W.M., Cui Y.J., Wang Q., 2015. Swelling pressure and hydraulic conductivity of compacted GMZ01 bentonite under salinization-desalinization cycle conditions. Applied Clay Science 114, 454-460. Cui Y.J., Loiseau C., Delage P. 2002. Microstructure changes of a confined swelling soil due to suction controlled hydration. Unsaturated soils: proceedings of the Third International Conference on Unsaturated Soils, 10-13, March 2002, Recife, Brazil, 593-598. Cui Y.J. Tang A.M., Loiseau C., Delage P. 2008. Determining the unsaturated hydraulic conductivity of a compacted sand-bentonite mixture under constant-volume and free-swell conditions. Physics and Chemistry of the Earth, Parts A/B/C, 33(Supplement 1), 462 – S471. Cui Y.J., Tang A.M., Qian L.X., Ye W.M., Chen B. 2011. Thermal-mechanical behavior of compacted GMZ Bentonite. Soils and Foundations 51(6), 1065-1074. Daniel D.E. 1982. Measurement of hydraulic conductivity of unsaturated soils with thermocouple psychrometers. Soil Science Society of America Journal 20(6), 1125–1129. Delage P., Graham J. 1995. The mechanical behaviour of unsaturated soils. Proceedings of the 1st International Conference on Unsaturated soils, Vol. 3, 1223-1256, Paris, E.E. Alonso and P. Delage eds, Balkema. Delage P., Howat M. D., Cui Y. J. 1998. The relationship between suction and swelling properties in a heavily compacted unsaturated clay. Engineering Geology 50, 31-48. Delage P., Cui Y.J., Tang A.M. 2010. Clays in radioactive waste disposal, Journal of Rock Mechanics and Geotechnical Engineering 2, 111–123. Dixon D.A., Gray M.N., Graham J. 1996. Swelling and hydraulic properties of bentonites from Japan, Canada and USA. In Proceedings of the second International Congress on Environmental Geotechnics, Osaka, Japan, 5–8. Dixon D.A., Graham J., Gray M.N. 1999. Hydraulic conductivity of clays in confined tests under low hydraulic gradients. Canadian Geotechnical Journal, 36(5), 815–825. Fernández A.M., Villar M.V. 2010. Geochemical behaviour of a bentonite barrier in the laboratory after up to 8 years of heating and hydration. Applied Geochemistry 25, 809–824 Gatabin C., Touze G., Imbert C., Guillot W., Billaud P. 2008. ESDRED Project, Module 1-Selection and THM characterization of the buffer material. In International conference underground disposal unit design&emplacement processes for a deep geological repository, 16-18 June, Prague. Imbert C., Villar M.V. 2006. Hydro-mechanical response of a bentonite pellets-powder mixture upon infiltration. Applied Clay Science 32(3-4), 197–209. Karnland O., Muurinen A., Karlsson F. 2005. Bentonite swelling pressure in NaCl solutions-Experimentally determined data and model calculations. Advances in Understanding Engineering Clay Barriers. Page 241. Karnland O., Nilsson U., Weber H., Wersin P. 2008. Sealing ability of Wyoming bentonite pellets foreseen

21

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

as buffer material-Laboratory results. Physics and Chemistry of the Earth, Parts A/B/C, 33, S472–S475. Kenney T.C., Veen W.A.V., Swallow M.A., Sungaila. M.A. 1992. Hydraulic conductivity of compacted bentonite-sand mixtures. Canadian Geotechnical Journal, 29(3), 364–374. Komine H. Ogata N. 1994. Experimental study on swelling characteristics of compacted bentonite. Canadian geotechnical journal, 31(4), 478–490. Komine H., Ogata N. 1998. Thermal influence on compacted bentonite for nuclear waste disposal. In Proceedings of the 3rd International Congress on Environmental Geotechnics, volume 1, pages 34–39. Komine H., Ogata N. 2003. New equations for swelling characteristics of bentonite-based buffer materials. Canadian Geotechnical Journal, 40(2), 460–475. Komine. H., Ogata N. 2004. Predicting swelling characteristics of bentonites. Journal of Geotechnical and Geoenvironmental engineering 130, 818. Komine H. 2004. Simplified evaluation for swelling characteristics of bentonites, Engineering geology, 71(3-4), 265-279. Komine H. Yasuhara K., Murakami S. 2009. Swelling characteristics of bentonites in artificial seawater. Canadian Geotechnical Journal, 46(2), 177-189. Komine H., Watanabe Y. 2010. The past, present and future of the geo-environment in Japan. Soils and Foundations, Vol. 50(6), 977-982. Kröhn K.P. 2003. Results and interpretation of bentonite resaturation experiments with liquid water and water vapour. In: Schanz, T. (Ed.), Proceedings of the International Conference from Experimental Evidence towards Numerical Modeling of Unsaturated Soils, Weimar, Germany, vol. 1. 598 Springer, Berlin, 257–272. Lemaire T., Moyne C., Stemmelen D. 2004. Imbibition test in a clay powder (MX-80 bentonite). Applied Clay Science, 26(1-4), 235–248. Li X.L., Bastiaens W., Van Marcke P., Verstricht J., Chen G.J., Weetjens E., Sillen X. 2010. Design and development of large-scale in-situ PRACLAY heater test and horizontal high-level radioactive waste disposal gallery seal test in Belgian HADES, Journal of Rock Mechanics and Geotechnical Engineering 2 (2), 103–110. Loiseau C., Cui Y.J., Delage, P., 2002. The gradient effect on the water flow through a compacted swelling soil, In: Proc. 3rd Int Conf Unsaturated Soils, UNSAT 2002, Recife, Brazil, Balkema 1, 395–400. Marcial D., Delage P., Cui Y.J. 2002. On the high stress compression of bentonites. Canadian Geotechnical Journal 39, 812-820. Marcial D. 2003. Comportement hydromécanique et microstructural des matériaux de barrière ouvragée, thèse ENPC. Mata C. 2003. Hydraulic behaviour of bentonite based mixtures in engineered barriers: The Backfill and Plug Test at the A¨ spo¨ HRL (Sweden). Ph. D. Thesis. Universitat Polite`cnica de Catalunya. 257 pp. Barcelona. Muurinen A., Lehikoinen J. 1999. Porewater chemistry in compacted bentonite, Engineering Geology 54, 207–214. Pusch R. 1979. Highly compacted sodium bentonite for isolating rock-deposited radioactive waste products. Nucl. Technol. (United States), 45(2). Pusch R. 1982. Mineral-water interactions and their influence on the physical behavior of highly compacted Na bentonite. Canadian Geotechnical Journal, 19(3), 381–387. Pusch R. 2001a. The microstructure of MX-80 clay with respect to its bulk physical properties under

22

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

different environmental conditions, SKBF/KBS technical report. No.TR01-08. Pusch R. 2001b. Experimental study of the effect of high porewater salinity on the physical properties of a natural smectitic clay, SKBF/KBS technical report. No.TR01-07. Romero E., Gens A., Lloret A. 2001. Temperature effects on the hydraulic behaviour of an unsaturated clay. Geotechnical and Geological Engineering, 19(3), 311–332. Romero E., Della Vecchia G., Jommi C. 2011. An insight into the water retention properties of compacted clayey soils. Géotechnique 6 (4), 313-328. Saba S., Barnichon J.D., Cui Y.J., Tang A.M., Delage P. 2014a. Microstructure and anisotropic swelling behaviour of compacted bentonite/sand mixture. Journal of Rock Mechanics and Geotechnical Engineering 6(2), 126-132. Saba S., Cui Y.J., Tang A.M., Barnichon J.D. 2014b. Investigation of the swelling behaviour of compacted bentonite/sand mixture by mock-up tests. Canadian Geotechnical Journal 51(12), 1399-1412. Sposito G. 1984. The surface chemistry of soils. Oxford Press, NewYork. Stroes-Gascoyne S. 2010. Microbial occurrence in bentonite-based buffer, backfill and sealing materials from large-scale experiments at AECL’s underground research laboratory. Applied Clay Science 47(1-2), 36–42. Suzuki S., Prayongphan S., Ichikawa Y., Chae B. 2005. In situ observations of the swelling of bentonite aggregates in NaCl solution, Applied Clay Science 29, 89-98. Tang A.M., Cui Y.J. 2005. Controlling suction by the vapour equilibrium technique at different temperatures and its application in determining the water retention properties of MX80 clay. Canadian Geotechnical Journal 42 (1), 287-296. Tang A.M., Cui Y.J. Barnel N. 2007. A new isotropic cell for studying the thermo-mechanical behavior of unsaturated expansive soil. Geotechnical Testing Journal 30 (5), 341 – 348. Tang A.M., Cui Y.J., Barnel N. 2008. Thermo-mechanical behaviour of a compacted swelling clay. Géotechnique 58 (1), 45-54. Tang A.M., Cui Y.J. 2009. Modelling the thermo-mechanical behaviour of compacted expansive clays. Géotechnique 59 (3), 185-195. Tang A.M, Cui, Y.J., 2010. Effects of mineralogy on thermo-hydro-mechanical parameters of MX80 bentonite, Journal of Rock Mechanics and Geotechnical Engineering 2(1), 91-96. Tang C.S., Tang A.M., Cui Y.J., Delage P., Barnichon J.D., Shi B. 2011a. A study of the hydro-mechanical behaviour of compacted crushed argillite. Engineering Geology 118 (3-4), 93-103. Tang C.S., Tang A.M., Cui Y.J., Delage P., Schroeder C., De Laure E. 2011b. Investigating the Swelling Pressure of Compacted Crushed-Callovo-Oxfordian Argillite. Physics and Chemistry of the Earth 36 (17-18), 1857–1866. Villar M.V. 2002. Thermo-hydro-mechanical characterisation of a bentonite from Cabo de Gata. A study applied to the use of bentonite as sealing material in high level radioactive waste repositories. Publicación Técnica ENRESA 01/2002, Madrid, 258 pp. Villar M.V., Lloret A. 2004. Influence of temperature on the hydro-mechanical behaviour of a compacted bentonite. Applied Clay Science 26(1-4), 337 – 350. Villar M.V. 2005. MX-80 Bentonite. Thermo-Hydro-Mechanical Characterisation Performed at CIEMAT in the Context of the Prototype Project. CIEMAT Technical Report: CIEMAT/DIAE/54540/2/04. Villar M.V., Lloret A. 2008. Influence of dry density and water content on the swelling of a compacted bentonite. Applied Clay Science 39(1-2), 38-49. Wan M., Ye W. M., Chen Y.G., Cui Y.J., Wang J., 2015. Influence of temperature on the water retention

23

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

properties of compacted GMZ01 Bentonite. Environmental Earth Science 73(8), 4053-4061. Wang Q., Tang A.M., Cui Y.J., Delage P., Gatmiri B. 2012. Experimental study on the swelling behaviour of bentonite/claystone mixture. Engineering Geology 124, 59-66. Wang Q., Tang A.M., Cui Y.J., Barnichon J.D. Delage P., Ye W.M. 2013a. The effects of technological voids on the hydro-mechanical behaviour of compacted bentonite-sand mixture. Soils and Foundations 53(2), 232-245. Wang Q., Cui Y.J., Tang A.M., Barnichon J.D., Saba S., Ye W.M. 2013b. Hydraulic conductivity and microstructure changes of compacted bentonite/sand mixture during hydration. Engineering Geology 164, 67-76. Wang Q., Tang A.M., Cui Y.J., Barnichon J.D. Ye W.M. 2013c. Investigation of the hydro-mechanical behaviour of compacted bentonite/sand mixture based on the BExM model. Computer and Geotechnics 54, 46-52. Wang Q., Tang A.M., Cui Y.J., Barnichon J.D. Ye W.M. 2013d. A comparative study on the hydro-mechanical behaviour of compacted bentonite/sand plug based on laboratory and field infiltration tests. Engineering Geology 162, 79-87. Wang Q., Cui Y.J., Tang A.M., Delage P., Gatmiri B., Ye W.M. 2014. Long-term effect of water chemistry on the swelling pressure of a bentonite-based material. Applied Clay Science 87, 157-162. Yahia-Aissa M., Delage P., Cui Y.J. 2001. Suction-water relationship in swelling clays. Clay science for engineering, IS-Shizuoka International Symposium on Suction, Swelling, Permeability and Structure of Clays, 65-68, Adachi & Fukue eds, Balkema. Ye W.M., Cui Y.J., Qian L.X., Chen B. 2009. An experimental study of the water transfer through compacted GMZ bentonite. Engineering Geology 108, 169 – 176. Ye W.M., Lai X.L., Liu Y., Chen Y.G., Cui Y.J. 2013a. Ageing effects on swelling behaviour of compacted GMZ01 bentonite. Nuclear Engineering and Design 265, 262-268. Ye W. M., Wan M., Chen B., Chen Y.G., Cui Y.J., Wang J. 2013b. Temperature effects on the swelling pressure and saturated hydraulic conductivity of the compacted GMZ01 bentonite. Environmental Earth Science 68, 281-288. Ye W. M., Wan M., Chen B., Chen Y.G., Cui Y.J., Wang J. 2014a. An unsaturated hydraulic conductivity model for compacted GMZ01 bentonite with consideration of temperature. Environmental Earth Science 71, 1937-1944. Ye W. M., Zheng Z. J., Chen B., Chen Y. G., Cui Y. J., Wang J. 2014b. Effects of pH and temperature on the swelling pressure and hydraulic conductivity of compacted GMZ01 bentonite. Applied Clay Science 101, 192-198. Ye W. M., Zhang F., Chen B., Chen Y.G., Wang Q., Cui Y.J. 2014c. Effects of salt solutions on the hydro-mechanical behaviour of compacted GMZ01 Bentonite. Environmental Earth Science 72(7), 2621-2630. Yong R.N., Boonsinsuk P., Wong G. 1986. Formulation of backfill material for a nuclear fuel waste disposal vault. Canadian Geotechnical Journal 23(2), 216-228. Zhu C.M., Ye W.M., Chen Y.G., Chen B., Cui Y.J. 2013. Influence of salt solutions on the swelling pressure and hydraulic conductivity of compacted GMZ01 bentonite. Engineering Geology 166, 74-80. Zhu C.M., Ye W.M., Chen Y.G., Chen B., Cui Y.J. 2015. Impact of cyclically infiltration of CaCl2 solution and de-ionised water on volume change behaviour of compacted GMZ01 bentonite. Engineering Geology 184, 104-110.

24

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

List of Figures Figure 1. Grain size distribution of the MX80 bentonite and sand (modified from Wang et al. 2013a) Figure 2. Grain size distribution curves of MX80 bentonite and crushed COx claystone (modified from Wang et al. 2012) Figure 3. Pore size distribution of bentonite/sand mixture compacted to different dry density with a water content of about 11.0% (modified from Wang et al. 2013a) Figure 4. WRCs of bentonite-sand mixture and pure bentonite (modified from Wang et al. 2013a) Figure 5. Water volume ratio (ew) versus suction (modified from Wang et al. 2013a) Figure 6. Hydraulic conductivity versus suction (modified from Wang et al. 2013b) Figure 7. Pore size distribution changes (modified from Wang et al. 2013c) Figure 8. Changes of large-pores void ratio (diameter larger than 2 µm) with suction (modified from Wang et al. 2013b) Figure 9. Hydraulic conductivity versus bentonite void ratio (modified from Wang et al. 2013a) Figure 10. Results of various mixtures using MX80 bentonite - swelling pressure versus final dry density of bentonite (modified from Wang et al. 2012) Figure 11. Swelling pressure versus final dry density of bentonite (modified from Wang et al. 2012) Figure 12. Evolution of swelling pressure for tests LT01, LT02, and LT03 for 700 days; (a) scale from 0 to 5 MPa, (b) scale from 3.6 to 4.6 MPa (modified from Wang et al. 2014) Figure 13. Pore size distribution curves; (a) cumulative curves, (b) derived curves (modified from Wang et al. 2014) Figure 14. Relationship between yield stress and bentonite void ratio (modified from Wang et al. 2013a) Figure 15. Anisotropy coefficient changes with bentonite dry density for different test conditions (modified from Saba et al. 2014a) Figure 16. Evolution of dry density profile over time (modified from Saba et al. 2014b)

25

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

Figure 1. Grain size distribution of the MX80 bentonite and sand (modified from Wang et al. 2013a)

Figure 2. Grain size distribution curves of MX80 bentonite and crushed COx claystone (modified from Wang et al. 2012)

26

ACCEPTED MANUSCRIPT

0.4 0.35

3

Dry density (Mg/m )

0.25

RI PT

1.67 1.97

0.2 0.15

SC

dem/dlogd

0.3

0.1

0 0.001

0.01

M AN U

0.05 0.1

1

10

100

1000

d (µm)

AC C

EP

TE D

Figure 3. Pore size distribution of bentonite/sand mixture compacted to different dry density with a water content of about 11.0% (modified from Wang et al. 2013a)

Figure 4. WRCs of bentonite-sand mixture and pure bentonite (modified from Wang et al. 2013a)

27

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 5. Water volume ratio (ew) versus suction (modified from Wang et al. 2013a)

ksat (Karnland et al., 2008)

TE D

-13

1×10

AC C

EP

Hydraulic conductivity (m/s)

kunsat

-14

1×10

0

10

20

30

40

50

60

70

Suction (MPa) Figure 6. Hydraulic conductivity versus suction (modified from Wang et al. 2013b)

28

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

0.3

TE D

Figure 7. Pore size distribution changes (modified from Wang et al. 2013c)

0.28

Low suction High suction

Ⅰ

EP

Large-pore void ratio

Ⅱ

AC C

0.26 0.24 0.22

0.2

0

20

40

60

80

Suction (MPa)

Figure 8. Changes of large-pores void ratio (diameter larger than 2 µm) with suction (modified from Wang et al. 2013b)

29

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 9. Hydraulic conductivity versus bentonite void ratio (modified from Wang et al. 2013a)

TE D

10

1

EP

Swelling pressure (MPa)

100

AC C

0.1

0.01 0.4 0.8 1.2 1.6 2 3 Final dry density of bentonite (Mg/m ) Bentonite/claystone 70/30 (Wang et al. 2012) Bentonite/sand 70/30 (Karnland et al. 2008) Pure bentonite (Karnland et al. 2008) Pure bentonite (Dixon et al. 1996 )

Figure 10. Results of various mixtures using MX80 bentonite - swelling pressure versus final 30

ACCEPTED MANUSCRIPT dry density of bentonite (modified from Wang et al. 2012)

MX80 Foca Febex Kunigel V1

RI PT

10

SC

1

0.1

0.01 0.5

1

M AN U

Swelling pressure (MPa)

100

1.5

2

2.5 3

Final dry density of bentonite (Mg/m )

AC C

EP

TE D

Figure 11. Swelling pressure versus final dry density of bentonite (modified from Wang et al. 2012)

31

ACCEPTED MANUSCRIPT

3

2

SC

LT01 (Synthetic water) LT02 (Distilled water) LT03 (Vapour-distilled water)

RI PT

4

1

0 0

200

M AN U

Swelling pressure (MPa)

5

400 600 Elapsed time (day)

800

(a)

4.2

EP

4.0

TE D

4.4

3.8

AC C

Swelling pressure (MPa)

4.6

LT01 (Synthetic water) LT02 (Distilled water) LT03 (Vapour-distilled water)

3.6

0

200 400 600 Elapsed time (day)

800

(b)

Figure 12. Evolution of swelling pressure for tests LT01, LT02, and LT03 for 700 days; (a) scale from 0 to 5 MPa, (b) scale from 3.6 to 4.6 MPa (modified from Wang et al. 2014)

32

ACCEPTED MANUSCRIPT ST

0.5

(Synthetic water)

LT01 (Synthetic water) LT02 (Distilled water) LT03 (Vapour-Distilled water)

0.3

RI PT

0.2 0.1 0 0.001

0.01

0.1 1 10 100 1000 Entrance pore diameter (µm) (a)

M AN U

0.4

0.2 0.1

TE D

dem/dlogd (-)

0.3

0 0.001

SC

e (-)

0.4

0.01

0.1 1 10 100 Entrance pore diameter (µm)

1000

AC C

EP

(b) Figure 13. Pore size distribution curves; (a) cumulative curves, (b) derived curves (modified from Wang et al. 2014)

33

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

TE D

Figure 14. Relationship between yield stress and bentonite void ratio (modified from Wang et al. 2013a)

1.1 1.0

EP

0.9 0.8

AC C

C a 0.7 0.6

In-cell Out-cell 14% void

0.5

0.4 0.9

1.0

1.1

1.2

1.3

1.4

1.5

ρdb (Mg/m ) 3

Figure 15. Anisotropy coefficient changes with bentonite dry density for different test conditions (modified from Saba et al. 2014a)

34

ACCEPTED MANUSCRIPT 150

Initial (mean)

120

58 days (Saba et al., 2014)

60

RI PT

Height (mm)

Final (expected)

90

120 days (Saba et al., 2014)

30

350 days (measured)

0

0

0.5

1

1.5

2

SC

Dry density (Mg/m3)

AC C

EP

TE D

M AN U

Figure 16. Evolution of dry density profile over time (modified from Saba et al. 2014b)

35