Effect of pH in clay particle – solution systems
The negative charge at face sides of clay particles is permanent, whereas the charge at edge sides varies as a function of the pH of the ambient solution. The conditional variation in charge is due to the formation of hydroxyl groups. Those groups are characterized by either a positive charge, when pH-values are low, or a negative charge, if pH-values are high (e.g. van Olphen, 1980). The point of zero charge defines the pH-value, at which the charge at the edge sides is zero (Sposito, 1992). In slurried samples, a reorientation of layers can ensue as the pH varies. They can position themselves face-to-face, edge-to-edge or edge-to-face, which leads to a sig-nificant alteration of the rheological properties of the slurry (e.g. van Olphen, 1980).
By definition, a solution is composed of a liquid solvent, in which solutes are dissolved. Ac-cordingly, groundwater is a solution. Water is the liquid solvent, whereas different ions are the solutes. In this regard, water refers to water molecules in the liquid- and gas phase. Inorganic salts are the dominant electrolytes in groundwater. Once dissociated, their ions are hydrated by the polar water molecules. The preference to be hydrated depends mainly on the nature of the ions, such as the ion charge and the ion radius (Adamson and Gast, 1997).
According to Sposito (1992), the higher hydration potential accounts for the migration of water molecules from the ambient solution into the interlayer space of smectites, where they either hydrate the exchangeable cations or interact with the interlayer surface. The system approaches electroneutrality by these means. He indicated that those interactions are mainly governed by the activity of water molecules in the clay particle – solution system, the hydration potential of the exchangeable cations and the polarization of water molecules by the exchangeable cations. In general, cation hydration involves the adsorption of water molecules on the exchangeable cations and the formation of hydration complexes. Cations are preferentially hydrated due to their positive charge and their smaller ion radius. Hydration complexes differ in their mode of interaction with the interlayer surface (e.g. Sposito, 1992):
Inner sphere complexes are characterized by the direct connection of the cation to the in-terlayer surface on the one side and to a number of water molecules on the other side.
The cation of outer sphere complexes interacts with the interlayer surface through its water ligands as it is completely surrounded by water molecules. A discrete layer of water ligands is generally referred to as hydration shell. The formation of outer sphere complexes is a key process accounting for the mobilization of cations.
Water molecules somehow interacting with the interlayer surface are henceforth referred to as interlayer water (e.g. Bradbury and Baeyens, 2003).
Crystalline swelling mechanism
As described above, water molecules hydrate the exchangeable cations and interact with the interlayer surface once they enter the interlayer space. The major interaction is the adsorption of water molecules on the interlayer surface, where they can constitute a discrete layer. Upon saturation, discrete layers intercalate, which leads to a stepwise increase in the basal spacing of particles. The increase in the basal spacing due to the intercalation of discrete layers of water molecules is referred to as crystalline swelling (Norrish, 1954). The mechanism contributes to soil expansion, predominantly when soils are characterized by a low water content.
Several studies aimed to quantify the discrete layers of water molecules in Na- and Ca-mont-morillonites under different compaction and confining conditions upon increasing the relative humidity (RH) (e.g. Chipera et al., 1998; Caballero et al., 2004; Saiyouri et al., 2004; Ferrage et al., 2005; Devineau et al., 2006; Villar, 2007). The number of layers was deduced from the basal spacing, which was measured by means of X-ray diffractometry once samples were in equilibrium with the imposed RH. The results are compiled in Figure 2.3a and Figure 2.3b.
In general, different numbers of discrete layers of water molecules can coexist as differences in the layer charge location cause spatial variations in the surface potential on particle face sides. Following, values correspond to average values of basal spacing. In the case of Na-montmoril-lonite, basal spacing of about 10 Å, 12.5 Å, 15.5 Å and 18.5 Å indicated no, one, two and three discrete layer(s) of water molecules, respectively, regardless of confinement and initial dry den-sity. One, two and three discrete layer(s) of water molecules were adsorbed upon imposing RH greater than 30 %, 80 % and 95 %, respectively (Devineau et al., 2006). Saiyouri et al. (2004) reported that the basal spacing even expanded to about 21.5 Å corresponding to a four discrete layers of water molecules. The value was obtained by imposing a relative humidity higher than 99 %. In the case of Ca-montmorillonite, the adsorption of no, one and two discrete layer(s) of water molecules induced basal spacing of about 11 Å, 15.5 Å and 18 Å, respectively. One and two distinct layers evolved upon imposing RH greater than 20 % and 95 %, respectively. Ac-cording to Michot et al. (2013), the maximum basal spacing of Ca-montmorillonite might be limited to 18 Å due to ion correlation forces.
Villar et al. (2012) emphasized that the expansion of the basal spacing is time-dependent and the probability of coexisting hydration states reduces over time as water molecules redistribute from micro- and macropores into the interlayer space. Further, redistribution processes ap-peared to be accelerated through confined conditions.
Solution retention behavior of expansive soils
According to the reference concept, backfill materials are compacted directly inside the drifts in unsaturated state and saturate through the long-term percolation of solutions originating in the surrounding geological formation. This section covers a brief description of the basic prin-ciples of solution retention in expansive soils. It aims to emphasize the difference of the hydro-mechanical behaviour in expansive soil under saturated and unsaturated conditions.
Concept of suction in expansive soils
Partially saturated soils are characterized by a negative pore solution pressure being equal to the total free energy of the pore solution. In this regard, pore solution comprises the interlayer water, the double layer water, the free water, and the solutes, whereas pore air refers to the gas phase in the different pores. The energy of the pore solution is related to soil-solution potential which portrays the mechanisms triggered by the different thermodynamic forces in soils (Yong and Warkertin, 1975; Yong, 1999). The soil-solution potential comprises five components: 𝜓𝑡𝑜𝑡= 𝜓𝑚+ 𝜓𝜋+ 𝜓𝑔+ 𝜓𝑎+ 𝜓𝑝 (2.5).
Where tot is the total soil-solution potential, m is the matric soil-solution potential, is the osmotic soil-solution potential, g is the gravimetric soil-solution potential, a is the pneumatic soil-solution potential and p is the external pressure soil-solution potential. Matric potential is caused by capillarity in conjunction with adsorption forces, whereas osmotic potential is at-tributable to differences in the chemical potential of solutes and solvent. In expansive soils, hydration potentials in the interlayer pores, adsorption of ionic and molecular species on clay particles and the formation of diffuse double layers particularly promote capillarity. Matric and osmotic potentials are often considered to be sufficiently precise to portray the total energy of the pore solution. The other components are the gravitational potential being related to the po-sition of the soil in the gravitational field, the pneumatic potential being attributable the pressure in the pore air, and the pressure potential induced by externally applied pressures. Although the latter three potentials exhibit a minor impact compared to the matric and osmotic potential, they still contribute to the total free energy of the pore solution.
The conclusions of Yong and Warkertin (1975) are in accordance with the definition of Aitchison (1965). By his definition, which bases on the principles of capillary condensation in porous media, the total free energy of the pore solution is characterized by its total suction stot. It is defined as the ratio of the partial vapor pressure of the solution in capillary-like pores pv to the saturation vapor pressure of the solvent in unconfined conditions pv0. The ratio of the partial vapor pressure to the saturation vapor pressure is commonly referred to as relative humidity RH and can be related to the total suction b Kelvin’s law: 𝑠𝑡𝑜𝑡=−𝑅𝑇𝑀𝑔 𝑙𝑛(𝑝𝑣𝑝𝑣0) ⇒ 𝑠𝑡𝑜𝑡=−𝑅𝑇𝑀𝑔 𝑙𝑛(𝑅𝐻) (2.6).
Where stot is the total suction, R is the universal gas constant, T is the absolute temperature, M is the molar mass of water and g is the gravity acceleration. The components of total suction are matric suction sm and osmotic suction sπ. Matric suction is mainly caused by capillarity promoted by other physico-chemical interactions. Since the partial vapor pressure is smaller than the saturation vapor pressure, vapor condenses in the capillaries. The pressure difference is equilibrated by the formation of menisci at the interface of the pore air and pore solution, whose shape changes as a function of the surface tension T of the solvent and the radius r of the capillary. This functional relation can be defined by Young-Laplace’s equation for completel wettable solvents: 𝑢𝑎− 𝑢𝑤= 2𝑇𝑟 (2.7).
Impact of physical and physico-chemical parameters
Among others, Villar (2007) investigated MX80- and FEBEX-bentonites with regard to their solution retention behavior. Her major aim was to assess the impact of the dominant exchange-able cation species and confinement on the solution retention behavior. She compacted the sam-ples to target dry densities at hygroscopic water contents and inserted them into special constant volume test cells. Placed in desiccators, samples were subjected to different values of RH by means of the vapor equilibrium method. In addition, materials were prepared to different water contents before compacting them to the target dry densities. The suction of those blocks was determined later by means of a humidity sensor. Her results indicated that the water content of FEBEX-bentonite was higher than that of MX80-bentonite for any given suction in the meas-ured range. She attributed her observations to the higher hydration potential of Ca2+-ions in the interlayer pores (e.g. Caballero et al., 2004).
Among other aspects, Marcial et al. (2002) and Saiyouri et al. (2004) determined the solution retention properties of MX80-, Kunigel-V1- and FoCa-bentonites under free conditions. Kuni-gel-V1- and FoCa-bentonites are Na- and Ca-bentonites, respectively. Final values of water content and void ratio indicated that Na-bentonites absorbed considerably more water than Ca-bentonites when suctions went below a certain value. These findings were in accordance with section 2.3.3 since the expansion of Na-montmorillonite in its crystallographic c-direction was not limited by ion correlation forces. In addition, Saiyouri et al. (2004) put emphasis on the different textures. FoCa-bentonite is typically characterized by bulky particles, so its specific surface area is smaller.
Al-Mukhtar et al. (1999) investigated the impact of different parameters, in particular the spe-cific surface area (SSA), on the solution retention behavior of two artificially activated Na-bentonites under free and confined conditions by means of the vapor equilibrium technique. Bentonites mainly differed in their SSA. According to their results, bentonites reacted similarly to RH lower than 40 %. Upon imposing higher RH, their retention behavior diverged consider-ably from each other since the one characterized by the higher SSA adsorbed more water. This result might support the results presented in the paragraph above.
Impact of compaction conditions
Romero et al. (1999) aimed to determine the main wetting and drying paths of Boom-clay, whose smectite-content ranges between 10 % and 20 %. Samples were compacted to two dif-ferent dry densities at an initial water content of 15 % prior testing. They reported that the initial void ratio affected the main wetting and drying paths once water contents exceeded 15 %. As water contents ranged between 5 % and 15 %, the impact started vanishing and, as water con-tents went below 5 %, the main wetting and drying paths converged. Thus, densification af-fected the solution retention behavior, only if macropores controlled the capillary condensation, and became less pronounced when micropores kept the solution bound. The impact of the initial dry density on the distribution of micro- and macropores in expansive soils is emphasized in section 2.5.3.
Gatabin et al. (2016) confirmed the observations of Villar (2007) and Villar et al. (2012) by reporting that samples of MX80-bentonite/ sand-mixtures stored more water under free swell conditions than under confined conditions. However, this effect was less pronounced upon ex-ceeding suctions greater than 10 MPa. Elevated water contents were attributed to the facilitated development of interlayer and diffuse double layer water under free conditions. As highlighted in section 2.3.2, the confinement modified the redistribution of water molecules between the interlayer and the micro- and macropores.
Impact of physical and physico-chemical conditions
Masrouri et al. (2005) studied Manois-Argillite (MA) and its mixtures with MX80-bentonite with regard to their compaction behavior by means of standard Proctor experiments. The ben-tonite content in mixtures ranged from 5 % to 20 % in weight. Their results are compiled and depicted in Figure 2.7. As a consequence of elevated fraction of associated minerals, optimum water contents were shifted to lower values, whereas maximum dry densities were increased. They attributed the findings to the reduced amount of diffuse double layers and, in turn, to their limited repulsion. Komine and Ogata (1999) and Srikanth and Mishra (2016) conducted standard Proctor experi-ments on various sand/ bentonite-mixtures and showed that there was an optimum bentonite content (20% – 30%). It likely ensued when the void space between the sand grains was com-pletely occupied by bentonite particles.
Relations of the compaction conditions to the structure
Diamond (1971) studied kaolinite- and illite-dominated soils with special regard to the impact of the initial water content on their pore size distribution. His experimental protocol comprised the compaction of the samples to their individual maximum dry density at the optimum water content as well as at dry and wet side of optimum before analyzing them by means of mercury intrusion porosimetry (MIP) experiments. The pore size distribution function was bi-modal when the soil was compacted at the dry side and indicated the existence of cardhouse-like ag-gregations incorporating micro- and macropores. When the water content approached the opti-mum value, the pore size distribution became mono-modal as the macropores vanished. Their disappearance was due to the evolution of the cardhouse-like aggregations into bulkier struc-tures. Therein, the clay particles were in face-to-face contact. The further increase of the water content caused a shift of the single peak to greater pore diameter values. Similar results were reported by Tang et al. (2011b) for COXc, for instance. Some studies reported that soils might exhibit a bi-modal pore size distribution even when they are compacted at the wet side of opti-mum (e.g. Tarantino and Col, 2008; Monroy et al., 2010). The results of Monroy et al. (2010) are depicted in Figure 2.8a.
Table of contents :
1.1 Reference backfill and seal concept
1.2 Motivation and objectives
1.3 Thesis layout
2 Literature review
2.1 Scale dependence in expansive soil testing
2.2 Nature of expansive soils
2.2.1 Mineralogy of expansive soils
2.2.2 Structure of expansive soils
2.2.3 Experimental determination
2.3 Interactions in clay particle – solution systems
2.3.1 Effect of pH in clay particle – solution systems
2.3.2 Hydration mechanisms
2.3.3 Crystalline swelling mechanism
2.3.4 Osmotic swelling mechanism
2.3.5 Cation exchange processes
2.4 Solution retention behavior of expansive soils
2.4.1 Concept of suction in expansive soils
2.4.2 Solution retention in expansive soils
2.4.3 Experimental determination
2.4.4 Impact of physical and physico-chemical parameters
2.4.5 Impact of compaction conditions
2.5 Compaction behavior of expansive soils
2.5.1 Experimental determination
2.5.2 Impact of physical and physico-chemical conditions
2.5.3 Relations of the compaction conditions to the structure
2.6 Swelling behavior of expansive soils
2.6.1 Experimental determination
2.6.2 Swelling pressure evolution at experimental different scales
2.6.3 Impact of physical and physico-chemical parameters
2.6.4 Impact of compaction conditions
2.6.5 Impact of environmental conditions and stress history
2.7 Compression behavior of expansive soils
2.7.1 Experimental determination
2.7.2 Impact of physical and physico-chemical parameters
2.7.3 Impact of compaction conditions
2.7.4 Impact of environmental conditions and stress history
2.8 Hydraulic conductivity behaviour of expansive soils
2.8.1 Experimental determination
2.8.2 Impact of the hydraulic gradient
2.8.3 Impact of physical and physico-chemical parameters
2.8.4 Impact of compaction conditions
2.8.5 Impact of environmental conditions and stress history
2.9 Constitutive modelling of expansive soil behavior
2.9.1 Elastic behavior
2.9.2 Loading-collapse (LC) and suction increase (SI) yield curves
2.9.3 Plastic behavior
2.10 Extended constitutive modelling of expansive soils
2.10.1 Basic preliminary assumptions
2.10.2 Elastic behaviour
2.10.3 Suction increase (SI) and suction decrease (SD) yield curves
2.10.4 Interaction between micro- and macrostructural levels
2.10.5 Plastic behavior
2.11 Conclusions and anticipated contributions
3 Combined impact of selected material properties and environmental conditions on the swelling pressure of compacted claystone/ bentonite mixtures
3.2 Theoretical background
3.6.1 Introduction of expansive mineral dry density (EDD)
3.6.2 Impact of expansive mineral content and grain size distribution
3.6.3 Individual impact of EDD and solution chemistry
3.6.4 Combined impact of solution chemistry and EDD
4 Hydraulic conductivity, microstructure and texture of compacted claystone/ bentonite mixtures saturated with different solutions
4.1 Introduction and background
4.3.1 Sample preparation
4.3.2 Hydraulic conductivity
4.3.3 Microstructural analysis
4.3.4 Textural analysis
4.3.5 Experimental program
4.4.1 Hydraulic conductivity experiments
4.4.2 Microstructural analysis
4.4.3 Textural analysis
4.5.1 Textural analysis
4.5.2 Microstructural analysis
4.5.3 Hydraulic conductivity experiments
5 Hydro-mechanical path dependency of claystone/ bentonite mixture samples characterized by different initial dry densities
5.2.1 General characteristics
5.2.2 Microstructural characteristics
5.2.3 Water retention characteristics
5.3 Oedometer and constant-volume swelling pressure experiments
5.3.2 Sample preparation
5.3.3 Suction-controlled oedometer experiments
5.3.4 Suction-controlled constant-volume swelling pressure experiments
5.3.5 Hydro-mechanical paths and experimental program
5.4.1 Suction-controlled oedometer experiments
5.4.2 Suction-controlled constant-volume swelling pressure experiments
5.5.1 Compression behavior
5.5.2 Swelling behavior
5.5.3 Comparison of yield behavior
6 Conclusions and perspectives
6.2 Synthesizes and conclusions
A Supplementary material