Effect of water chemistry on the hydro-mechanical behaviour of compacted mixtures of claystone and Na/Ca-bentonites for deep geological repositories

Zhixiong Zeng, Yu-Jun Cui, Jen Tlndier

a Laboratoire Navier/CERMES, Ecole des Ponts ParisTech, Marne La Vallée, 77455, France

b French National Radioactive Waste Management Agency (Andra), Châtenay-Malabry, 92298, France

Keywords:Bentonite/claystone mixture Synthetic site solution Cement solution Bentonite type Swelling pressure Hydraulic conductivity

ABSTRACT In the French deep geological disposal for radioactive wastes, compacted bentonite/claystone mixtures have been considered as possible sealing materials.After emplacement in place, such mixtures are hydrated by the site solution as well as the cement solution produced by the degradation of concrete.In this study, the effects of synthetic site solution and cement solution on the hydro-mechanical behaviour of compacted mixtures of claystone and two types of bentonites (MX80 Na-bentonite and Sardinia Cabentonite) were investigated by carrying out a series of swelling pressure, hydraulic conductivity and mercury intrusion porosimetry (MIP) tests.It was found that for the MX80 bentonite/claystone mixture hydrated with synthetic site solution, the swelling capacity was reduced compared to the case with deionised water owing to the transformation of Na-montmorillonite to multi-cation dominant montmorillonite by cation exchanges.For the Sardinia bentonite/claystone mixture,the similar increasing rate of swelling pressure was observed during the crystalline swelling process for different solutions, suggesting insignificant cation exchanges.Additionally, the cations in the synthetic site solution could reduce the thickness of diffuse double layer and the osmotic swelling for both MX80 bentonite/claystone and Sardinia bentonite/claystone mixtures.The large-pore volume increased consequently and enhanced water flow.In the cement solution, the hydroxide could also dissolve the montmorillonite, reducing the swelling pressure, and increase the large-pore volume, facilitating the water flow.Furthermore, the decrease of swelling pressure and the increase of hydraulic conductivity were more significant in the case of low dry density because of more intensive interaction between montmorillonite and hydroxide due to the high permeability.

1.Introduction

Deep geological repository has been accepted as a solution for the long-term storage of high-level radioactive wastes in many countries (Sellin and Leupin, 2013).To ensure the long-term safe isolation of radioactive wastes in deep geological formations,bentonite-based materials are commonly used as sealing materials(Pusch,1982; Dixon et al.,1985; Villar and Lloret, 2008; Ye et al.,2010), while cement (in mortar or concrete) is used to construct the tunnel retaining structures and tunnel plugs/seals (Sánchez et al., 2006).After the repositories are closed, the bentonitebased materials will be hydrated by the pore water of the host rock and then swell, filling the technological voids and restricting the release of radionuclides from the radioactive wastes to the surrounding environment.It has been reported that the pore water of the host rock considered in several countries is of certain salinity(Deng et al.,2011;Nguyen et al.,2013;Wang et al.,2014;Sun et al.,2018).In addition, the cement in contact with groundwater for a long period can degrade and release a large number of Ca2+and OH-, resulting in a high-pH alkaline solution (Savage et al., 2002;Karnland et al., 2007; Anh et al., 2017; Sun et al., 2018).In the performance assessment of the storage repositories, it appears essential to account for the effect of such water chemistry on the hydro-mechanical behaviour of bentonite-based materials.

In the past decades, the influence of saline solution on the hydro-mechanical behaviour of bentonite-based materials was studied by many researchers(Karnland et al.,2006;Rao et al.,2006;Castellanos et al.,2008;Komine et al.,2009;Zhu et al.,2013).It was found that the pore water salinity could significantly decrease the swelling pressure (Komine et al., 2009; Zhu et al., 2013) and increase the hydraulic conductivity (Castellanos et al., 2008; Zhu et al., 2013).Yukselen-Aksoy et al.(2008) and Zhu et al.(2013)reported that the cations in the pore water might alter the mineralogical compositions through cation exchanges and influence the crystalline swelling process,resulting in a lower swelling pressure.Additionally, the cations in the pore water could decrease the thickness of diffuse double layer, affecting the osmotic swelling behaviour (Castellanos et al., 2008; Schanz and Tripathy, 2009;Siddiqua et al.,2011;Zhu et al.,2013;Du et al.,2021).At a given void ratio, a larger large-pore volume could be expected for the specimen hydrated with a saline solution,leading to a higher hydraulic conductivity.

In addition to the saline solution, alkaline solution can also significantly influence the hydro-mechanical behaviour of bentonite-based materials (Karnland et al., 2007; Herbert et al.,2008; Lee et al., 2012; Chen et al., 2016; Sun et al., 2018, 2019,2020; Liu et al., 2020).Chen et al.(2016) experimentally determined the swelling pressures and hydraulic conductivities of GMZ bentonite hydrated with various NaOH solutions and found that the final swelling pressure decreased and the hydraulic conductivity increased significantly with the increase of NaOH solution concentration.Karnland et al.(2007)investigated the constant-volume swelling pressures of compacted Wyoming MX80 bentonite hydrated with NaOH and Ca(OH)2solutions and found that 0.1 mol/L NaOH (pH = 12.9) or saturated Ca(OH)2solutions (pH = 12.4) did not significantly change the swelling pressure of compacted MX80 bentonite.In contrast,Sun et al.(2018,2019)compared the swelling pressures and mineralogical compositions of GMZ bentonite hydrated with Beishan site solution(pH=8.5),young concrete water(pH = 13) and evolved concrete water (pH = 12), and observed a remarkably lower swelling pressure in the case of young concrete water.They explained this phenomenon by the dissolution of montmorillonite upon saturation with alkaline solutions and the formation of some less active zeolites and calcium hydrated silicates.Additionally,the montmorillonite content decreased linearly as the pH increased (Sun et al., 2019).

Fig.1.Grain size distributions of MX80 and Sardinia bentonites and crushed COx claystone.

In France, Callovo-Oxfordian (COx) claystone has been considered as a potential geological host rock for high-level radioactive waste disposal.To reduce the excavation waste and ensure the mineralogical compatibility, a mixture of crushed COx claystone and bentonite in the form of pre-compacted blocks has been proposed as a possible sealing material by the French National Radioactive Waste Management Agency(Andra)(Wang et al.,2014;Zeng et al., 2019, 2021a; Middelhoff, 2020).Wang et al.(2014)carried out a series of constant-volume swelling pressure tests on compacted MX80 bentonite/COx claystone mixture with a bentonite fraction of 70%(dry density of 1.7 Mg/m3)and observed a slight decrease of swelling pressure with the synthetic site solution after 700 d of hydration.Cuisinier et al.(2009) studied the microstructure of compacted COx claystone at a dry density of 1.61 Mg/m3hydrated with portlandite-saturated solution for 1 year using mercury intrusion porosimetry (MIP) and scanning electron microscopy (SEM) techniques and found an increase of macro-pore void ratio after the fluid circulation.By contrast, Middelhoff(2020) indicated that the swelling pressure after about 7 d and the hydraulic conductivity after 1 year were hardly affected by the site solution and cement solution for the compacted MX80 bentonite/COx claystone mixture (a proportion of 30/70 in dry mass) at a dry density of 1.72 Mg/m3.Up to now, there was no consensus regarding the effects of site solution and cement solution on the hydro-mechanical behaviour of bentonite/claystone mixture.Additionally, to the authors’ knowledge, the responses of the mixtures of COx claystone and different bentonites upon hydration with site solution and cement solution have not been systematically studied.

In this study,a series of infiltration tests using deionised water,synthetic site solution and cement solution were performed on compacted mixtures of COx claystone and MX80 Na-bentonite and Sardinia Ca-bentonite at different dry densities.The effects of synthetic site solution and cement solution on the swelling pressure and hydraulic conductivity were analysed.Moreover, the microstructure features before and after hydration were also determined using MIP technique,enabling the interpretation of the physico-chemical interaction between different bentonite/claystone mixtures and different fluids involved.

2.Materials and methods

2.1.Materials

According to the preliminary results of Zeng et al.(2020a), the bentonite/claystone mixtures with a proportion of 30/70 in dry mass were used in this study to fulfill the requirements of Andra in terms of swelling pressure and hydraulic conductivity with a dry density of 1.72-1.82 Mg/m3.To investigate the effect of bentonitetype, two bentonites, MX80 and Sardinia bentonites, were considered.The MX80 bentonite is a Na-bentonite,extracted from Wyoming in the USA,while the Sardinia bentonite is a Ca-bentonite taken from Monte Furros, Italy.The main physical and chemical characteristics of the two bentonites are presented in Table 1.The COx claystone was sampled from the Andra Underground Research Laboratory (URL) in Bure, France.It is composed of 40%-45%interstratified illite/smectite,30%carbonates,and 25%-30%quartz and feldspar.The COx claystone is characterised by a liquid limit of 41%, a plastic limit of 24% and a specific gravity of 2.7.Prior to testing, the bentonites and claystone were crushed to have a maximum grain size of 2 mm.The grain size distributions of the bentonites and claystone determined by dry sieving are presented in Fig.1.The mean grain diameters(D50)of the MX80 and Sardinia bentonites and COx claystone are 0.55 mm,1.06 mm and 0.58 mm,respectively.

Table 1Physical and chemical properties of MX80 and Sardinia bentonites.

Table 2Recipe for the synthetic site solution and cement solution (g/L) preparation.

Table 3Composition (mmol/L) of the synthetic site solution and cement solution.

Table 4Experimental programme.

Fig.2.Schematic diagram of the constant-volume cell for swelling pressure and hydraulic conductivity tests.

Synthetic site solution and cement solution were used for the infiltration tests.These solutions respectively represent the site water at a depth of 490 m and the alkaline solution after the degradation of concrete,and were prepared according to the recipe provided by Andra (Table 2).The chemical compositions of the solutions are summarised in Table 3.The pH values of the two solutions were measured, equal to 8.6 and 12.5, respectively.In addition, deionised water was also employed as a reference case.

2.2.Experimental methods

The bentonite and claystone powders with a proportion of 30/70 in dry mass were first mixed for more than 10 min.The gravimetric water contents of the MX80 bentonite, Sardinia bentonite and claystone under laboratory conditions(a relative humidity of about 60%) were determined by oven-drying at 105◦C for 24 h, equal to 11.4%,16%and 6.1%,respectively.Accordingly,the respective water contents of the MX80 bentonite/claystone and Sardinia bentonite/claystone mixtures were 7.7% and 9.1%, respectively.Based on the target dry density, a pre-determined amount of mixture was poured into a rigid steel ring with an inner diameter of 50 mm and statically compacted at a constant displacement rate of 0.05 mm/min.The final height of compacted specimens was 10 mm.For the MX80 bentonite/claystone mixture, nine specimens with dry densities of 1.6 Mg/m3,1.8 Mg/m3and 2 Mg/m3were compacted,three for each dry density(Table 4);for the Sardinia bentonite/claystone mixture,six specimens with dry densities of 1.6 Mg/m3and 1.8 Mg/m3were prepared.Afterwards, the compacted specimens were transferred into a testing cell with an inner diameter of 50 mm(Fig.2)and placed between two porous stones and filter papers.A circular cap with a diameter of 50 mm was placed at the top of the specimens and then blocked using a screw to prevent axial swelling.After that, deionised water, synthetic site solution or cement solution was injected into the specimens from the bottom of the cell under a water head of about 1 m,and the axial swelling force was monitored using a force transducer installed under the testing cell.For all the specimens,the hydration lasted for 90 d.All the tests were performed at constant ambient temperature(20◦C ± 1◦C).

At the end of the swelling pressure tests, the solution injection was continued for more than 24 h using a pressure/volume controller under constant pressures.To minimise the disturbance of the microstructure and avoid hydraulic fracturing, the water injection pressure was lower than 1/10 of the final swelling pressure.For the MX80 bentonite/claystone mixtures at 1.6 Mg/m3,1.8 Mg/m3and 2 Mg/m3dry densities,the injection pressures were respectively 0.04 MPa, 0.1 MPa and 0.2 MPa, while those for the Sardinia bentonite/claystone mixtures at 1.6 Mg/m3and 1.8 Mg/m3dry densities were 0.03 MPa and 0.1 MPa, respectively.During the solution injection,the solution volume injected into the specimens was recorded.After stabilisation of the flow rate, the specimens were regarded as saturated and the saturated hydraulic conductivity was calculated based on the Darcy’s law:

Fig.3.Evolution of axial swelling pressure of specimens at different dry densities: (a) 1.6 Mg/m3, (b) 1.8 Mg/m3, and (c) 2 Mg/m3.

where q is the flow rate(m3/s),i is the hydraulic gradient,and A is the cross-sectional area (m2).

After the infiltration tests,the specimens were taken out of the cell for microstructure observation.The specimens were first cut into small pieces (about 1 cm3in volume).Afterwards, the pieces were immersed in slush nitrogen(-210◦C)obtained by previously submitting it to vacuum for instantaneous freezing, and then lyophilised in a vacuumed chamber (Delage et al., 1996).For the MIP test, the freeze-dried pieces were put in a low pressure chamber with a working pressure from 3.6 kPa to 2000 kPa and then a high pressure chamber with a working pressure from 0.2 MPa to 228 MPa, enabling identification of pore entrance diameter of 350-0.006 μm (the diameter at the narrowest position).

3.Experimental results

3.1.Swelling pressure

The evolutions of swelling pressure for different dry densities are depicted in Fig.3.On the whole,the curves can be divided into two stages: a relatively large primary swell (stage I) and a small secondary swell(stage II).In stage I,the swelling pressure increased quickly.The overall increasing rate for the MX80 bentonite/claystone mixture hydrated with synthetic site solution and cement solution was slightly lower than that of the specimens hydrated with deionised water.Moreover, the synthetic site solution and cement solution reduced the elapsed time required to complete stage I.By contrast, the increasing rate for the Sardinia bentonite/claystone mixture was almost the same regardless of the permeating solutions.In stage II, for the MX80 bentonite/claystone mixture with large dry densities (1.8 Mg/m3and 2 Mg/m3), the swelling pressure tended to stabilisation.By contrast,for the MX80 bentonite/claystone mixture at a low dry density (1.6 Mg/m3) and the Sardinia bentonite/claystone mixtures at dry densities of 1.6 Mg/m3and 1.8 Mg/m3, the swelling pressure reached a peak firstly, then decreased sharply, and finally tented to stabilisation.The specimens at a low dry density and hydrated with cement solution exhibited a more remarkable decrease of swelling pressure after reaching a peak.Moreover,at a given dry density,the decrease of swelling pressure after reaching a peak was more significant for the Sardinia bentonite/claystone mixture compared to the MX80 bentonite/claystone mixture.The final swelling pressures of the specimens at various dry densities and hydrated with different solutions are summarised in Fig.4.There were reasonable linear relationships between the final swelling pressure and the drydensity for the specimens hydrated with different solutions.In case of high dry densities (1.8 Mg/m3and 2 Mg/m3), the final swelling pressure was almost independent of the permeating solution.In case of low dry density(1.6 Mg/m3),the synthetic site solution and cement solution more or less reduced the swelling pressure.Additionally, from Fig.4, a slightly lower final swelling pressure could be observed on the Sardinia bentonite/claystone mixture comparted to the MX80 bentonite/claystone mixture,especially in the case of low dry density of 1.6 Mg/m3.

Fig.4.Final swelling pressure versus dry density.

3.2.Saturated hydraulic conductivity

At the end of swelling pressures tests, the saturated hydraulic conductivities were determined or various dry densities, and the results are presented in Fig.5.For the specimens hydrated with a given solution, the hydraulic conductivity decreased with the increase of dry density.The synthetic site solution and cement solution increased the hydraulic conductivity for both the MX80 bentonite/claystone and the Sardinia bentonite/claystone mixtures.This is consistent with the results of Villar(2006)and Chen et al.(2016) while studying the effects of saline and alkaline solutions on the hydraulic conductivity of MX80 bentonite/granite mixture (a proportion of 30/70 in dry mass) and GMZ bentonite, respectively.Additionally, larger increases of hydraulic conductivity due to the water chemistry were observed on the specimens hydrated with cement solution.The lower the dry density of specimens, the more significant the influences of synthetic site and cement solutions on the saturated hydraulic conductivity.In addition, compared to the MX80 bentonite/claystone mixture, the compacted Sardinia bentonite/claystone mixture exhibited a larger hydraulic conductivity at the same dry density.

3.3.Microstructure observation

Fig.5.Hydraulic conductivity versus dry density.

Figs.6 and 7 illustrate the pore size distributions of the specimens before and after hydration.From the cumulative curves(Figs.6a and 7a), it could be observed that the final value of intruded mercury void ratio was lower than the global void ratio.This difference indicated the existence of a large amount of porosity which was inaccessible with the maximum pressure applied in the MIP tests.From the density curves(Figs.6b and 7b),the as-compacted specimens exhibited a typical bimodal porosity,allowing two main pore populations to be defined:a population of intra-aggregate pores(small-pores)with a mean pore diameter of 0.02 μm and a population of inter-aggregate pores (large-pores)with a mean pore diameter of 5 μm.This is consistent with the observation of Delage et al.(1996,2006)who found that the soils compacted dry of optimum could be described by a typical aggregate microstructure.After hydration, the mean pore diameter of large-pores increased to about 25 μm.By contrast, the mean pore diameter of small-pores remained unchanged but the peak value decreased remarkably.For the specimens at low dry densities(1.6 Mg/m3and 1.8 Mg/m3),a new pore population with a mean diameter of 0.2-0.5 μm (defined as medium-pores)appeared between the large and small-pores.Following the suggestion of Bian et al.(2019)and Zeng et al.(2020b),the respective delimiting diameters between the large and medium-pores and between the medium and small-pores were taken as 2 μm and 0.04 μm, respectively.To further investigate the effects of synthetic site solution and cement solution on the pore structure of bentonite/claystone mixtures, the void ratios of four populations were determined based on the delimiting values and the cumulative curves.The results are summarised in Figs.8 and 9.It can be clearly observed that the synthetic site solution and cement solution increased the larger- and small-pore void ratios and decreased the medium-pore void ratio.Similar phenomenon was observed by Wang et al.(2014) and Liu et al.(2020) while investigating the effect of saline solution on the pore structure of compacted MX80 bentonite/claystone(70/30 in dry mass)and the effect of NaOH solution on the pore structure of compacted GMZ bentonite,respectively.Moreover,the cement solution had a more remarkable influence on the pore structure changes than the synthetic site solution.Comparison of the pore structures of two bentonite/claystone mixtures shows that at a given dry density,the Sardinia bentonite/claystone mixture exhibited a lower largepore void ratio and larger medium- and small-pore void ratios than the MX80 bentonite/claystone mixture.

Fig.6.Pore size distribution of MX80 bentonite/claystone mixtures:(a)cumulative curves of specimens at a dry density of 1.6 Mg/m3;(b)density function curves of specimens at a dry density of 1.6 Mg/m3;(c)cumulative curves of specimens at a dry density of 1.8 Mg/m3;(d)density function curves of specimens at a dry density of 1.8 Mg/m3;(e)cumulative curves of specimens at a dry density of 2 Mg/m3; and (f) density function curves of specimens at a dry density of 2 Mg/m3.

4.Interpretation and discussion

4.1.Effects of synthetic site solution and cement solution on hydromechanical behaviour

When the clay minerals in the bentonite/claystone mixtures were wetted with water, their hydration could be described by crystalline and osmotic swelling processes (Mitchell, 1993).At a low relative humidity, crystalline hydration is the predominant mechanism with water molecules being progressively adsorbed on the clay surfaces layer by layer,leading to an increase of the intraaggregate pore volume.Meanwhile, the swollen aggregates invaded the inter-aggregate pores and reduced the inter-aggregate pore volume.Obviously, this process depended on the global dry density.The lower the dry density, the more significant the swelling of aggregates.With the transformation of intra-and interaggregate pores to the medium-pore population, a remarkable increase of medium-pore void ratio occurred (Figs.8 and 9).During the crystalline swelling process,the clay particles would be divided into smaller ones and fissure-like two-dimensional(2D)pores with a mean pore size of 20 μm appeared.The variation of large-pore volume after hydration was due to the combined effects of theinvasion of swollen clay particles and the appearance of 2D pores.As the suction decreased to 4 MPa,the crystalline swelling reached a nearly steady state and a significant amount of water would be stored in inter-particle pores (Saiyouri et al., 2000).Due to the difference of cation concentration between the pore water and the clay particle surface, diffuse double layer would develop on a parallel assembly of clay particles (Liu, 2013) and osmotic swelling took place.

Fig.7.Pore size distribution of Sardinia bentonite/claystone mixtures:(a)cumulative curves of specimens at a dry density of 1.6 Mg/m3;(b)density function curves of specimens at a dry density of 1.6 Mg/m3; (c) cumulative curves of specimens at a dry density of 1.8 Mg/m3; and (d) density function curves of specimens at a dry density of 1.8 Mg/m3.

Generally, the maximum number of water molecule layers intercalated between the unit layers depends on the clay type and the available cations in the permeating water(Marcial et al.,2002;Liu,2013).If the cation in the pore water is characterised by a high replacing power,cation exchanges can take place and the clay type will change.A typically competitive order that has been commonly admitted is:Na+

Apart from the cations, the hydroxide in the cement solution could also greatly influence the hydro-mechanical behaviour of the bentonite/claystone mixtures.Upon contact with hydroxide, the montmorillonite might be dissolved, forming secondary minerals with much lower swelling capacity, such as analcime and calcium hydrated silicate gels (Sun et al., 2018, 2019).As a result, the swelling capacity was reduced (Fig.4).The produced calcium hydrated silicate gels were characterised by a large amount of small pores(Wang et al.,2017),leading to an increase of small-pore void ratio of the specimens hydrated with cement solution (Figs.8 and 9) (Sun et al., 2019; Liu et al., 2020).By contrast, the dissolution of montmorillonite would lead to an increase of the large-pore volume (Chen et al., 2016).Compared to the specimens hydrated with synthetic site solution,the specimens could be affected by the hydroxide in addition to the cations.Thus,the specimens hydrated with cement solution exhibited a larger large-pore volume and a higher hydraulic conductivity compared with those hydrated with synthetic site solution (Fig.5).

Fig.8.Changes in void ratios of inaccessible pore,and small-,medium-and large-pores for the compacted MX80 bentonite/claystone mixtures:(a)1.6 Mg/m3,(b)1.8 Mg/m3,and(c) 2 Mg/m3.

Fig.9.Changes in void ratios of inaccessiblepore pore,and small-,medium-and large-pores for the compacted Sardinia bentonite/claystone mixtures:(a)1.6 Mg/m3 and(b)1.8 Mg/m3.

4.2.Effects of bentonite type and dry density on hydro-mechanical behaviour

Basically, the synthetic site solution and cement solution can influence the hydro-mechanical behaviour of bentonite-based materials mainly by means of cation exchanges, osmotic effect and montmorillonite dissolution.After the MX80 bentonite/claystone and Sardinia bentonite/claystone mixtures were in contact with synthetic site solution, cation exchanges could occur in the MX80 bentonite/claystone mixture,with the transformation of some Na-montmorillonite to Mg/Ca-montmorillonite, while the Sardinia bentonite/claystone mixture would be immune to the cation exchange.This would reduce the swelling capacity of MX80 bentonite/claystone mixture.Thus, a slightly smaller increasingrate of swelling pressure was found in stage I for MX80 bentonite/claystone mixture, as compared to that hydrated with deionised water (Fig.3a and b).Thanks to the higher permeability, the interaction between clay minerals and solution became more intensive for the MX80 bentonite/claystone mixture at a lower dry density upon the hydration with the synthetic site solution.By contrast, the cations in the synthetic site solution would exert an influence on the osmotic swelling of both the MX80 bentonite/claystone and the Sardinia bentonite/claystone mixtures at low dry densities.In general,the clay particles of Na-montmorillonite after crystalline swelling are thinner than those of Ca-montmorillonite(Saiyouri et al., 2004).It was reported that the clay particles are composed of 2-20 (and even more) aligned unit layers in Camontmorillonite, but only of 1-5 unit layers in Namontmorillonite (Liu, 2013).Therefore, during the osmotic swelling process,few double layer development could be expected in the Ca-montmorillonite, as compared to the Namontmorillonite.The weaker osmotic swelling of Sardinia bentonite/claystone mixture could not compensate the reduction of swelling pressure induced by the collapse of aggregate structure.Thus,a more significant decrease was observed in stage II at low dry densities (Fig.3a and b).Additionally, the cations in the synthetic site solution could reduce the repulsive force and the thickness of diffuse double layer, resulting in a lower swelling pressure and a higher hydraulic conductivity for the specimens hydrated with synthetic site solution(Figs.4 and 5).For the specimens at high dry densities,the low amount of inter-particle water was insufficient to form the diffuse double layer(Pusch and Yong,2006),and the effect of cations on the swelling pressure and hydraulic conductivity of the highly compacted MX80 bentonite/claystone and Sardinia bentonite/claystone mixtures was less significant.

5.Conclusions

The effects of synthetic site solution and cement solution on the hydro-mechanical behaviour of compacted MX80 Na-bentonite/claystone and Sardinia Ca-bentonite/claystone mixtures were investigated by carrying out a series of swelling pressure and hydraulic conductivity tests together with microstructure observation.The results obtained allow the following conclusions to be drawn:

(1) For the MX80 bentonite/claystone mixture,the cations in the synthetic site solution could transform the Namontmorillonite to multi-cation dominant montmorillonite, reducing the swelling capacity of the mixture.By contrast, for the Sardinia bentonite/claystone mixture, the cations in the synthetic site solution did not significantly change the Ca-montmorillonite during the crystalline swelling.Thus, an insignificant effect was identified on the swelling pressure.

(2) During the osmotic swelling,the presence of cations reduced the repulsive force by diffuse double layer and thus swelling pressure for both the MX80 bentonite/claystone and Sardinia bentonite/claystone mixtures.Additionally, a smaller thickness of diffuse double layer and a greater large-pore volume could be expected for the specimens hydrated with synthetic site solution.Thus, a slightly higher saturated hydraulic conductivity was obtained in that case, as compared to the case of deionised water.Furthermore, a more significant influence of synthetic site solution was observed on the specimens at a lower dry density thanks to the well-developed diffuse double layer.

(3) In addition to the cations, the hydroxide also reduced the swelling pressure of both the MX80 bentonite/claystone and Sardinia bentonite/claystone mixtures.The hydroxide led to a slight increase of large-pore volume and thus an increase of saturated hydraulic conductivity.Moreover, the lower the dry density, the more intensive the interaction between the montmorillonite and hydroxide,and the more significant the changes in swelling pressure and hydraulic conductivity,due to the higher permeability.

Over the experiment time scale of 90 d, only a slight deterioration of sealing performance of the bentonite/claystone mixture was identified due to the low cation concentration and the low ionic strength of permeating solutions.The effect of water chemistry in the very long-term lifespan of the underground radioactive waste repository needs to be investigated in further work.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

The research work was supported by Ecole des Ponts ParisTech and the French National Radioactive Waste Management Agency(Andra).The first author is grateful to the China Scholarship Council for the grant scholarship.