Molecular dynamics study of the bentonite barrier : effects of salinity and temperature on the swelling pressure

(1)

uef.fi

PUBLICATIONS OF

THE UNIVERSITY OF EASTERN FINLAND Dissertations in Forestry and Natural Sciences

ISBN 978-952-61-3332-4 ISSN 1798-5668

Dissertations in Forestry and Natural Sciences

DISSERTATIONS | BUKUNMI AKINWUNMI | MOLECULAR DYNAMICS STUDY OF THE BENTONITE BARRIER:… | No 37

BUKUNMI AKINWUNMI

MOLECULAR DYNAMICS STUDY OF THE BENTONITE BARRIER: EFFECTS OF SALINITY AND TEMPERATURE ON THE SWELLING PRESSURE

PUBLICATIONS OF

THE UNIVERSITY OF EASTERN FINLAND

The long-term performance of the nuclear waste disposal repository is dependent on the behaviour of constituent materials. Therefore, it is important to

understand the impact of environmental changes on these constituent barriers.

The thesis focuses on the impact of environmental factors on the swelling behaviour and especially the

swelling pressure of bentonite clays. With the use of molecular dynamics simulations, the influence of

salt solutions and temperature on montmorillonite clays have been investigated. The results are applicable in choosing appropriate bentonite clays with respect to the present and future environmental

conditions and changes of the repository.

BUKUNMI AKINWUNMI

(2)

(3)

Bukunmi Akinwunmi

MOLECULAR DYNAMICS STUDY OF THE BENTONITE BARRIER: EFFECTS OF SALINI- TY AND TEMPERATURE ON THE SWELLING

PRESSURE

Publications of the University of Eastern Finland Dissertations in Forestry and Natural Sciences

No 373

University of Eastern Finland Joensuu

2020

Academic dissertation

To be presented by the permission of the faculty of science and forestry for public examination in the Auditorium F100 in the Futura Building at the University of

Eastern Finland, Joensuu, on February 21, 2020, at 12 o’clock noon

(4)

Grano Oy Jyväskylä, 2020 Editor: Nina Hakulinen

Distribution: University of Eastern Finland/ Sales of publications www.uef.fi/kirjasto

ISBN: 978-952-61-3332-4 (print) ISSNL: 1798-5668

ISSN: 1798-5668 ISBN: 978-952-61-3333-1 (pdf)

ISSNL: 1798-5668 ISSN: 1798-5676

(5)

Author’s address: Bukunmi Akinwunmi University of Eastern Finland Department of Chemistry P.O BOX 111

80101 JOENSUU, FINLAND email: bukunmi.akinwunmi@uef.fi

Supervisors: Professor Emeritus Tapani Pakkanen, Ph.D.

University of Eastern Finland Department of Chemistry P.O BOX 111

80101 JOENSUU, FINLAND email: tapani.pakkanen@uef.fi Docent Janni Hirvi, Ph.D.

University of Eastern Finland Department of Chemistry P.O BOX 111

80101 JOENSUU, FINLAND email: janne.hirvi@uef.fi

Reviewers: Professor Emeritus Matti Hotokka, Ph.D.

Åbo Akademi

Department of Chemistry P.O BOX 780

20500 TURKU, FINLAND matti.hotokka@abo.fi

Professor Risto Laitinen, Ph.D.

University of Oulu

Environmental and Chemical Engineering unit P.O BOX 8000

90014 OULU, FINLAND risto.laitinen@oulu.fi

Opponent: Professor Markus Olin, Ph.D.

VTT Technical Research Center of Finland LTD, P.O. Box 1000

02044 ESPOO, FINLAND markus.olin@vtt.fi

(6)

(7)

Akinwunmi, Bukunmi

Molecular dynamics study of the bentonite barrier: Effects of salinity and temperature on the swelling pressure

Joensuu: University of Eastern Finland, 2020 Publications of the University of Eastern Finland Dissertations in Forestry and Natural Sciences 2020;

ABSTRACT

Bentonite clay is a potential buffer material in nuclear waste geological repositories, partly due to its swelling ability. While factors influencing the swelling behavior (especially the swelling pressure) of bentonite clays have been reported, some environmental factors have received less attention. This thesis focuses on smectite clays in homoionic and mixed salt solutions. Molecular dynamics simulation was used to examine the influence of different salt solutions and temperatures on the interlayer swelling pressure of the smectite clays.

Compared to calcium montmorillonite clay, the swelling pressure of sodium montmorillonite clay was found to be more sensitive to the salinity of the surrounding solution. Therefore, the swelling pressure and ion exchange process in the interlayer of sodium montmorillonite clay were further studied in detail in mixed- valence salt solutions containing both NaCl and CaCl2. The results show that both the swelling and the extent of cation exchange depend on the total ionic strength and the composition of the mixed salt solutions.

In addition, the interlayer swelling pressure increases slightly as the temperature increases from ambient. The introduction of ice in the environment leads to complete loss of the swelling pressure. When increasing the temperature to the ice melting point and beyond, the interlayer swelling pressure appears due to the increased entropy of water molecules and the hydration of both the clay surface and interlayer cations. The results of the study are applicable to the selection of buffer materials and the long-term performance of nuclear repositories under varying environmental conditions.

Universal Decimal Classification:544.022.84, 553.611, 621.039.7, 628.4.047

Library of Congress Subject Headings: Geological repositories; Radioactive waste repositories; Clay; Smectite; Montmorillonite; Bentonite – Expansion and contrac- tion; Molecular dynamics; Computer simulation; Salinity; Saltwater solutions; Ion exchange; Temperature; Ice; Entropy; Hydration

(8)

ACKNOWLEDGEMENTS

This work was carried out at the Department of Chemistry, University of East- ern Finland during the years 2017-2019. Financial support provided by KYT 2018 and Posiva Oy is gratefully acknowledged. I also acknowledge grants of computing capacity from the Finnish grid and cloud infrastructure for the computational work.

Foremost, my profound gratitude goes to Professor Emeritus Tapani Pakkanen for accepting me into his research group during my master’s studies, and for giving me the opportunity to proceed in the research work. I am grateful for your patience, guidance, and constructive criticism over the years. My appreciation also goes to Docent Janne Hirvi. I am grateful for your research advice, suggestions, patience, technical contributions, and insight during this work.

My appreciation goes to the Posiva Oy team for their influence during this study.

I specially thank Seppo Kasa for his valuable ideas, suggestions, and contributions from the beginning of this study. Mika Niskanen and Eveliina Muuri are well appreciated for their suggestions and comments, especially on Articles I and II. I acknowledge Paul Wersin and Andreas Jenni (University of Bern), and Vincente Navarro and Ángel Yustres (Universidad de Castilla-La Mancha) for their contributions in the planning phase of some of the works in this thesis.

I appreciate the staff members at the Department of Chemistry, University of Eastern Finland for their kind help, especially Sari Suvanto, Urpo Ratinen, Mari Heiskanen, and Professor Mika Suvanto. Associate Professor Jarkko Saarinen is appreciated for the interesting short discussions.

I would like to dedicate this work to my parents, Pastor and Deaconess Akinwunmi. Thank you for your prayers, love, and concern over the years. I could fly higher than an eagle because you are the wind beneath my wings. To my broth- ers, Korede and Kolade, thank you for being a part of my journey. You both have contributed to my growth in different ways during this study. To my friend, Ebunoluwa, words cannot express how much you mean to me. Thank you for your listening ears, my unpaid ‘therapist’. To my cousin, Ayomide, I appreciate your presence in my life. To other family members and friends, thank you all for your contributions in one way or another.

To my loving husband, Adeoluwa, I am grateful for the gift of you, being with you makes me better. You are the joy of my life. Thank you for cheering me on, for always being by my side, and for your immense support.

‘For I know the thoughts I have towards you, the thoughts of good and not of evil, to give you a future and a hope.’ Finally, I am thankful to God for being so mindful of me during this study.

Joensuu, 2020 Bukunmi Akinwunmi

(9)

LIST OF ABBREVIATIONS

Ca²⁺ Calcium ion CaCl² Calcium chloride

Ca-MMT Calcium montmorillonite

Cl Chloride ion

DDL Diffuse double layer

HLRW High-level radioactive waste

MD Molecular dynamics

Na⁺ Sodium ion

NaCl Sodium chloride

Na-MMT Sodium montmorillonite NPT Isobaric-isothermal ensemble

NVT Canonical ensemble

SPC Simple point charge

(10)

(11)

LIST OF ORIGINAL PUBLICATIONS

This thesis is based on data presented in the following articles, referred to by the Roman Numerals I–III.

I Akinwunmi B, Hirvi J. T, Kasa S, Pakkanen T. A. (2019). Swelling pressure of Na- and Ca-montmorillonites in saline environments: A molecular dynamics study,Chemical Physics, 528: 110511.

II Akinwunmi B, Kporha F. E. A, Hirvi J. T, Kasa S, Pakkanen T. A. Atomistic simulations of the swelling behavior of Na-montmorillonite in mixed NaCl and CaCl² solutions,Chemical Physics,submitted for publication.

III Akinwunmi B, Sun L, Hirvi J. T, Kasa S, Pakkanen T. A. (2019). Influence of temperature on the swelling pressure of bentonite clay,Chemical Physics, 516:

177–181.

(12)

AUTHOR’S CONTRIBUTION

The author carried out all the simulation work for article I, manuscript II, and article III. The author analyzed all the results in the articles and manuscript, except for some of the interlayer ion population analysis in manuscript II. The author also wrote the manuscripts for I–III under the guidance of the supervisors.

(13)

1 INTRODUCTION

1.1 Bentonite and nuclear waste management

Nuclear wastes are generated from a number of sources, especially the nuclear power industry¹. With the spread of general nuclear technology, the amount of generated harmful wastes also increases. Nuclear wastes at both low and high lev- els pose a threat to the environment, and therefore must be safely contained and appropriately disposed for optimum environmental conservation. At the moment, methods and facilities already exist for the disposal of low- and intermediate-level wastes². However, for high-level radioactive waste (HLRW), only temporary stor- ages are currently used in different countries. Many countries have ongoing plans and activities towards the permanent disposal of HLRW, while Finland is on the verge of making its first disposal^3–6. The final disposal repository design adopted by these countries is the geological multi-barrier system (figure 1), which is envisaged to provide adequate safety for the waste from the environment for a period of 100,000 years.

Figure 1. Finnish repository design showing (a) the fuel pellet, (b) fuel rod and fuel assembly (c) canister insert, (d) copper overpack, (e) bentonite buffer and tunnel backfill, and (f) 400–500 m of bedrock⁷.

The multi-barrier system consists of vitrified waste emplaced in copper canisters (figure 1d), while the canisters are deposited in the host rock approximately 500 m below ground surface. The space between the host rock and the canisters will be filled with compacted bentonite to adequately hold the canister in place and provide additional protection in case of canister failure⁸. Bentonite clay will also be used as the backfill for sealing the tunnel after waste deposition.

Bentonite clay was selected as the candidate buffer material because of its many desirable physico-chemical properties, including high swelling and sealing abilities, low hydraulic conductivity, durability, high absorbing power, high shear and com- pressive strengths, desirable rheological characters, and low compressibility. These

(16)

properties also led to the adoption of bentonite clay in other industrial applications shown in figure 2.

Figure 2. Industrial applications of bentonite clays.

1.2 From bentonite to montmorillonite

Generally, clays are defined as natural materials that exhibit plastic properties when hydrated and can be hardened by frying or drying⁹. Bentonites are naturally occurring clays formed from volcanic ashes after long-term changes or by the hy- drothermal weathering of igneous rocks and other materials^10–13. The major constituent of bentonite clay is the montmorillonite clay mineral¹².

Clay minerals are a group of phyllosilicates responsible for the characteristic behaviour of clays¹⁴. Their structure consists of tetrahedral and octahedral sheets arranged into layers. The arrangement of these sheets results in the two major classifications of clay minerals; the 1:1 clay minerals have one tetrahedral and one octahedral sheet in each clay layer, while the 2:1 clay minerals have one octahedral sheet sandwiched between two tetrahedral sheets in each layer (figure 3).

Smectites are typical 2:1 clay minerals^9,15. They are negatively charged with a total charge of 0.2–0.6 per half unit cell⁹, due to isomorphous substitution in either the tetrahedral sheet, octahedral sheet, or both. The Si⁴⁺ in the tetrahedral sheet can be replaced by lower valent Al³⁺ or Fe³⁺, while the Al³⁺ in the octahedral layer can be substituted with lower valent Mg²⁺ or Fe²⁺ ions⁹. The interlayers of smectites are

(17)

occupied by hydrated cations to achieve overall charge neutrality in the system.

The octahedral sheet of a smectite clay can contain either 3 divalent metals (triocta- hedral) or 2 trivalent ions (dioctahedral).

Figure 3. Schematic illustration of the classification of clay minerals.

A typical dioctahedral clay mineral is the montmorillonite clay, whose layer charge is mainly in the octahedral sheet (figure 4). The basic structure of montmorillonite clay is the tetrahedral silica and octahedral alumina. The Si⁴⁺ are tetrahe- drally coordinated with oxygen atoms, while the aluminum ions have six-fold co- ordination. The most abundant hydrated cations in the interlayer of montmorillonites are Na⁺ and Ca²⁺, while Na-montmorillonite (Na-MMT) is the dominating fraction. In addition to montmorillonites, bentonite may contain other accessory minerals such as calcite, feldspar, gypsum, and silica.

(18)

Figure 4. Montmorillonite clay structure. The colours red, orange, white, purple, and green represent oxygen, silicon, hydrogen, aluminum, and magnesium atoms respectively.

1.3 Bentonite swelling and swelling pressure

The characteristic swelling behavior of bentonite clay is an important factor in its use as a buffer and backfill material in constructing the nuclear waste repository.

When bentonite comes in contact with water, water is absorbed into its interlayer while the cations as well as the clay surface becomes hydrated, thereby increasing the volume of the clay¹⁶. If the volume of the clay is confined, a swelling pressure will be generated and exerted on the boundaries^17,18.

The overall swelling behavior of bentonite clay is governed by two main mechanisms, namely crystalline swelling and osmotic or continuous swelling^17,19–21. Crys- talline swelling occurs when exchangeable cations in the clay interlayer are hydrated. Meanwhile, water molecules are arranged between the clay layers, separating the clay layers and increasing the volume of montmorillonite. Crystalline swelling is the dominating swelling mechanism at a smaller interlayer spacing and contrib- utes the most to the total swelling. On the other hand, osmotic swelling occurs when the interlayer space is large, and it makes a rather minute contribution to the overall swelling. Crystalline swelling occurs in all types of swelling clays, while osmotic swelling is restricted to specific clays²². For instance, Na-MMT exhibits both swelling mechanisms, while only crystalline swelling is observed in Ca- montmorillonite (Ca-MMT).

(19)

Figure 5. Internal factors influencing the swelling pressure of montmorillonite clay.

Previous studies found that the degree of swelling and swelling pressure generated from smectites depend on a handful of internal and external factors. The internal factors are related to the clay structure, such as the d-spacing, dry density, layer charge, charge location, and the nature of exchangeable cations (figure 5)^23–27. On the other hand, the external factors are related to the environment, including the temperature, pressure, salinity of infiltrating solution, and pH (figure 6)^28–33. How- ever, complications in the clay structure, the degree of disorder, and the clay parti- cle sizes make it challenging to study these internal and external factors experimentally. Computer simulation, meanwhile, is a useful complimentary method to investigate the influence of these factors on the swelling behavior of bentonite clays^34–42.

(20)

Figure 6. External factors influencing the swelling pressure of montmorillonite clay.

1.4 Clay-water-electrolyte interactions

For an in-depth understanding of the swelling pressure generated by bentonite clay under varying conditions, it is important to know the detailed swelling phe- nomenon and interactions governing the swelling ability. When montmorillonite and water are in contact, water is readily and strongly adsorbed on the clay surface.

As a consequence, there will be interactions of clay, water, and other ions at the clay surface or in the water⁴³. Several attractive/repulsive forces influence the swelling properties of the clay and are important in this study²¹.

1.4.1 Water-ion interactions

As the water molecule has a permanent dipole with uneven charge distribution, it is attracted to ions in the solution, in a process known as the hydration of ions.

The structure of water molecules changes as they move into their position in the hydration shell of cations^43,44, as the negatively charged oxygen atoms in water are attracted to the cations. There are three major zones of water interaction around an ion as seen in figure 7. Zone 1 is the immediate surrounding of the ion, where the movement of water molecules is restricted due to their strong interaction with the cation.

(21)

Figure 7. Illustration of the zones of water structure near a cation.

There are weaker interactions between the ions and water molecules in zone 2.

In zone 3, the water structure is consistent with that of bulk water despite being polarized by the ionic field. Zone 1 is also known as the zone of immobilization or the primary solvation shell, while zone 2 is the secondary solvation shell⁴⁵. The behavior of water molecules in zones 1 and 2 depends on the valency and concentration of the cation. Small monovalent cations have approximately six to three water molecules in their primary solvation shell in dilute and concentrated solutions, respectively. The secondary solvation shell has not been reported for monovalent cations. Divalent cations have approximately eight to six water molecules in the primary shell in dilute and concentrated solutions, respectively, while approximately 15 water molecules are in the secondary shell⁴⁴.

1.4.2 Clay-water interactions

The interactions between clay and water can also change the structure of water, and such a change can partly explain the swelling and the swelling pressure. The mechanisms of clay-water interactions include osmosis, van der Waals attraction, hydrogen bonding, and cation hydration⁴⁶. The maximum ion concentration is lo- cated near the clay surface. Therefore, water molecules tend to move towards there to balance the concentration difference with respect to osmosis. In addition, hydrated cations are attracted to the negative charge on the clay surface, dragging their water of hydration toward the surface. The oxygen atoms and hydroxyls on the clay surface structure can also form hydrogen bonding by attracting the positive and negative charges in the water molecule, respectively⁴³.

(22)

1.4.3 Distribution of ions in the clay-water system

In dry clays, adsorbed cations occupy the hole positions on the clay surface, while during hydration, the ions move towards the central region between the clay layers. Meanwhile, cation diffusion away from the surface was balanced by an electrostatic attractive force from the negatively charged clay surface. These opposing interactions cause a pattern of ion distribution in the clay-water system.

1.4.3.1 Diffuse double layer theory

The distribution of cations near the clay surface according to the electrostatic attraction and thermal diffusion, with a higher concentration near the surface and a lower one towards the bulk, is called the diffuse double layer (DDL)13,21,43,47,48. The DDL theory particularly holds true for monovalent smectites at low electrolyte concentrations, and it has also been considered for divalent smectites¹⁷.

A thicker diffuse layer implies that a greater swelling pressure could be generated by the clay⁴³. The DDL thickness in turn depends on the temperature, cation valency, and electrolyte concentration^43,49. A higher environmental temperature leads to an increased double layer thickness, while a higher cation valency or a higher ion concentration makes the diffuse layer thinner. The DDL theory provides a foundation for understanding the interactions of swelling clay layers and could be used to predict the swelling properties. It was utilized to interpret the molecular dynamics (MD) simulation results of the water-mineral interface in nanopores, and credible correlations have been reported^50,51. However, a main limitation of the DDL theory is that it only holds when the clay dry density is lower than 1550 kgm^-3 and the ionic strength of the salt solution is less than 1.0 M, according to previous studies^17,49.

1.4.3.2 Cation exchange capacity

Montmorillonite clay is produced from the isomorphous substitution of octahedral Al³⁺ by lower valent cations such as Mg²⁺. This substitution leaves the clay layer with a total negative charge that is balanced with cations in the interlayer and on the surface of the clay. When the clay layer encounters saline water, the exchangeable cations can be readily replaced with the cations present in the water. The capacity of a type of cation to replace another one depends on its valence, the ion concentration, and the ionic radius^43,47,48. The order for the exchange capacity of cations is as follows⁴³

Na⁺ < Li⁺ < K⁺ < Rb⁺ < Cs⁺ < Mg²⁺ < Ca²⁺ < Ba²⁺ < Cu²⁺ < Al³⁺ < Fe³⁺ < Th⁴⁺

The rate of cation exchange depends on the properties of the clay, being the lowest in smectites compared to illites and kaolinites. This is because for smectites, the interlayer region is the major part of exchange⁴⁸. In contrast, the ion exchange regions in both kaolinites and illites are the broken bonds at the silica-alumina unit and the predominate edge hydrogen replacement in kaolinites.

(23)

1.5 Effects of salinity on the swelling behavior of bentonite clay During the long-term operational phase of the repository, the bentonite buffer is expected to encounter ground water, whose chemical compositions might affect various properties of the clay such as the swelling ability and capacity. Many researchers have reported on the swelling behavior of bentonite clays in saline solutions. All of them found a decrease in the swelling pressure with increasing ionic strength or concentration of the salt solution.

Most previous investigations on the effect of salinity on the swelling pressure of montmorillonite clays were carried out in homoionic solutions. Karnland reviewed and compared several models with experimental data by investigating the swelling pressure of bentonite clay in separate NaCl and CaCl2 solutions¹⁶. It was reported that the higher the salinity of the infiltrating solution, the lower the swelling pressure generated. Although the swelling pressure is overall sensitive to the salinity, at higher dry densities, NaCl and CaCl² solutions generate comparable pressures. In a further study, Na-bentonite and Ca-bentonite were infiltrated with alkaline and salt solutions. The swelling pressure drop observed in the salt solutions was stable and reversible, while a continuous drop was observed in the alkaline solutions²⁷.

Dixon and Vieno separately reported the influence of salinity on the DDL of bentonite clay^52,53. At a low dry density, the salt concentration in the pore water will readily influence the swelling pressure generated by the clay, while at higher dry densities, the influence of pore water chemistry is negligible.

Pusch classified the swelling pressure generated by bentonite clay into two components⁵⁴: The interlamellar hydration pressure and the osmotic pressure. The interlamellar hydration pressure is generated from the hydration of the interlayer cations, and it is solely responsible for swelling at high dry density. At low dry density, the osmotic pressure is responsible for the swelling pressure, which in this case depends on the DDL interactions. It was reported that the effect of salinity on the swelling pressure of bentonite clay depends on the clay density.

Herbert et al. stated that the influence of salinity on swelling pressure is greater than the impact of the exchangeable cation present⁵⁵. Lee et al. studied the effect of NaCl solutions on the swelling pressure of Ca-bentonite⁵⁶. They reported an initial conversion of Ca-bentonite to Na-bentonite at low NaCl concentrations, which was explained by the initial increase in swelling pressure followed by a pressure drop as the concentration increases. Zhu et al. examined a bentonite clay with a dry density of 1700 kgm^-3, and stated that NaCl has a higher impact on its swelling pressure compared to CaCl²⁵⁷. At constant concentration, the swelling pressure generated in NaCl solution was lower than that in CaCl². In addition, Castellanos et al. reported a similar observation, namely that the swelling pressure of bentonite clay is higher in CaCl² solution compared to in NaCl solution at the same concentration⁵⁸.

(24)

1.6 Effects of temperature on the swelling behavior of bentonite clay

During the operational phase of the nuclear waste repository, the radioactive waste is expected to generate some heat, which will raise the temperature of both the canister and the buffer in the immediate environment. On the other hand, during permafrost periods, the buffer may become frozen. These temperature changes might influence the swelling pressure and other properties of the bentonite buffer.

Therefore, understanding how the buffer properties change with temperature can help maintain the integrity of the whole multi-barrier system.

Several experimental investigations have examined the influence of temperature on the swelling pressure of bentonite clay. However, experimental conditions such as the bentonite composition, compaction, and water content make it quite complicated to uniquely determine the temperature dependence. As a result, researchers have drawn different conclusions on the subject matter. A decrease in swelling pressure as the temperature increases was reported by several researchers. The observation was justified by the cement formation in water, the dissolution of some montmorillonite mineral, and the transfer of water molecules from the clay micro- structure into the macrostructure, all of which inhibit clay swelling^59–66. On the other hand, there were reports of an increased swelling pressure with increasing temperature. This was justified by the enhanced thermal motions of water molecules, an increase in the system entropy, and the influence of temperature on the DDL15,28,29,67–

69.

1.7 Application of MD in clay science

The structure, properties, and behavior of clay, its interlayer species, and water molecules in the vicinity have been extensively studied using MD simulations, and the results are comparable with experimental findings. For MD studies of clay swelling, the most prominent and simplest model is the periodic model. In this model, the clay layers are modeled from end to end in a simulation box, and water molecules are inserted stepwise between the clay layers. This model has been employed to investigate the swelling behavior and properties of bentonite clays.

Previous studies have examined the hydration mechanism and dynamics of different interlayer ions. Tao et al. investigated the influence of K⁺, Na⁺, and Ca²⁺ on the swelling ability and stability of montmorillonite clay⁷⁰. It was reported that the valance of the interlayer cation has a significant influence on the clay hydration and swelling behaviour. It was further reported that among the three cations, Ca²⁺ is the most strongly attached to the clay layers in pure water environment. Boek evaluat- ed the structure and mobility of Li⁺, Na⁺, and Ca²⁺ in the interlayer of montmorillonite clays⁷¹, while Yang et al. investigated the difference in the hydration of Na⁺ and Ca²⁺ in the interlayer of montmorillonite clays⁵¹. Both Li⁺ and Na⁺ readily diffuse in

(25)

the interlayer while K⁺ diffusion is restricted by the strong interaction with the clay surface.

Furthermore, the diffusion of cations into the interlayer of montmorillonite clays has been studied using the periodic model. For example, Malikova et al. examined the influence of temperature on the diffusion of both Na and Cs in montmorillonite clay⁷². The diffusion coefficients of both cations were reported to increase as temperature increased. However, different diffusion characteristic was observed for both cations. Holmboe and Bourg examined the influence of temperature and pore size on the diffusion of water and Na⁺ in the clay interlayer. They reported that the diffusion coefficients of water and Na⁺ in the interlayer increase as the temperature of the system increases⁷³ Anion exclusion in the clay interlayer was discussed by Tournassat et al., who reported that the tendency of Cl to enter the clay interlayer decreases as the interlayer space decreases⁷⁴. The presence of anions in the clay interlayer was also confirmed by Hedström and Karnland⁷⁵.

The effects of the interlayer cation, temperature, pressure, and charge location on the swelling properties of bentonite clays have also been investigated using the periodic model. Teich-McGoldrick et al. reported the formation of two water layers in montmorillonite at a low water content, while beidellite required more water to stabilize the two water layers³⁷. Sun et al. also discussed the roles the interlayer cations and water content in understanding the swelling behavior of Na-MMT⁴¹.

Despite the numerous MD studies on the different properties of smectite clays, the periodic model is evidently unable to predict the swelling pressure of these clays. To quantify the swelling pressure generated by bentonite clays, Sun et al.

proposed another simulation model, the spring model⁴⁰. In this model, two clay layers are centrally placed in a simulation box and surrounded by the solution molecules and ions. The model is based on Hooke’s law, where force constants are employed to determine the swelling pressure. The upper clay layer is attached to a spring with a known force constant, and this spring is compressed as the clay ex- pands. The vertical displacement of the upper clay layer and the spring constant are used to calculate the swelling pressure. The spring model has proven to be effective for predicting the swelling pressure of bentonite clays, especially with varying structural factors^38,40. The present work involves the modification of the spring model and its application in predicting the swelling pressure of smectite clays with varying environmental conditions.

1.8 Research aim

Existing experimental results on the temperature dependence of the swelling pressure of smectites have reached contradictory conclusions due to the different experimental conditions used. The influence of mixed salt solutions has been over- looked. At the same time, computational study of the impact of these external factors on the swelling pressure of smectite clays has received less attention. Therefore, studies summarized in this dissertation aimed to use MD simulations with the

(26)

spring model to gain in-depth information on how these environmental factors affect the swelling pressure of smectite clays. The specific goals of the research are as follows:

To understand the swelling behavior of Na-MMT and Ca-MMT in salt solutions;

To predict the interlayer swelling pressure of Na-MMT in mixed salt solutions;

To understand processes occurring in the interlayer of Na-MMT with respect to the ion population in an environment of mixed salt compositions;

To examine the interlayer swelling pressure of smectite clay at very high and low temperatures.

(27)

2 COMPUTATIONAL DETAILS

Computational chemistry is the use of models and computer calculations to study complex molecular systems, such as biomolecules, organic and inorganic molecules, polymers, and clay⁷⁶. With the recent improvement in technology, especially in computer software and hardware, computational chemistry has proven to be a very effective tool in solving complicated problems.

Models employed in computer calculations can be grouped into four types based on their specific functions. First, simple models help in the simplification of systems and the analysis of the most important factors of interest. Didactical models are used to illustrate very complicated systems that may seem inaccessible oth- erwise. Mechanical models operate according to classical mechanics such as Hooke’s law. Finally, mathematical models are used in the study of processes such as chemical reaction steps on enzymes⁷⁷.

Molecular calculations methods can also be grouped into two major classes governed by quantum mechanics and molecular mechanics, as seen in figure 8. Quan- tum mechanics simulations, which are based strictly on theoretical principles, ex- plicitly describe the electrons, the breakage and formation of bonds, and chemical reactions. Despite its very high accuracies, however, the approach requires longer simulation times and is restricted to small molecules.

Figure 8. Basics classifications of methods in computational chemistry.

(28)

Molecular mechanics is based on simplified interactions in a system, such as the bond stretching, angle bending, and non-bonded interactions among the atoms. The system energy is calculated using empirical force fields based on the nuclear positions, while the properties and motion of the electrons are ignored. Therefore, this method can easily reduce the simulation time and accommodate larger systems, as compared to that based on quantum mechanics.

2.1 MD method and force field

As a technique based on molecular mechanics, MD simulations are based on Newton’s laws of motion. The calculated trajectories describe deterministic change in the positions and velocities of the atoms with respect to time. MD can be employed to model both the transport and equilibrium properties of a system. Forces in the system are calculated by integrating the equations of motion. MD simulations require a molecular mechanics force field to properly describe the potential energy of the system by providing the interaction energy (potential) of the atoms. These potentials are derived from the parameterization of experimental and spectroscopic data and/or quantum mechanics simulation.

A force field is specifically designed to predict and reproduce various properties and structures for a given molecular system, while being general enough to model some other related systems. The force field that is generated by partial charges and the Lennard-Jones potential determines the intermolecular interaction.

The Clayff force field was developed to simulate the structures, properties, and dynamics of hydrated clay, hydroxide, and oxyhydroxide systems⁷⁸. Both spectroscopic data and density functional theory calculations were employed in its devel- opment. The Clayff force field treats the interatomic interactions as non-bonded, which affords the use of this force field for a wide range of systems. It also enables the correct amount of energy transfer between clay surfaces and fluids. In the Clayff force field, metal-oxygen interactions are described by the Lennard-Jones potential, while the flexible simple point charge (SPC) water model is incorporated to describe the behavior of water⁷⁹.

(29)

3 RESULTS AND DISCUSSION

3.1 Effect of salinity on the swelling pressure of montmorillonite clay^I,II

3.1.1 Model setup

To study the influence of salinity on the interlayer swelling pressure of montmorillonite clay, a low-charge montmorillonite clay with the unit cell formula M0.5/nSi8(Al3.50Mg0.50)O20(OH)4was used for the study. M is the exchangeable cation on the clay surface, and n is the charge of the cation. The two clay layers were cen- tered in a box and separated by a varying d-spacing. Each layer of this clay consists of 8-unit cells. The clay layers were terminated along the (010) plane by protona- tion, hydration, and hydroxylation. The octahedral sheet of the layer was subjected to isomorphic substitution of Al³⁺ with Mg²⁺ to achieve a total layer charge of 0.5 e, and Na⁺ or Ca²⁺ was placed for charge balancing in Na-MMT or Ca-MMT, respectively.

The simulation set up was adapted from the previously reported spring model⁴⁰. For the swelling pressure calculation, the lower clay layer was restrained in all di- rections while the upper clay layer was also restrained in all but one direction. The vertical movement of the upper layer of the clay was allowed while the movement was constrained by springs attached to the layer. The d-spacing and force constants were modified to improve the accuracy of the simulation. The modified model is discussed in detail in article I. A total of 17 values from 1.4–3.0 nm were used for the d-spacing, while the force constant was within 0.5–16.5 kJ mol¹nm² with 0.5 kJ mol¹nm² increments. Both the d-spacing and force constant of the springs were varied to generate the swelling pressure curves.

The impact of salinity on the swelling pressure of montmorillonite clay was studied in two major stages. The first stage was to understand the swelling behavior of Na-MMT and Ca-MMT in pure NaCl and CaCl² solutions, respectively. The ionic strengths of interest are presented in table 1. However, the real environment in the nuclear waste repository will include both NaCl and CaCl² in various propor- tions and concentration gradients. As Na-bentonite is the preferred candidate buffer material, the second stage involves further investigation on the swelling behavior of Na-MMT in mixed salt solutions of varying ionic strength percentage ratios of NaCl and CaCl², as presented in table 2. The ionic strength percentage ratio is defined as I^NaCl(%)/I^CaCl²(%), whereas I^NaCl(%) equals the percentage molarity of NaCl and I^CaCl² (%) is three times the percentage molarity of CaCl². In the bulk saline environment, the ionic strength percentage ratios were chosen to be 100/0, 90/10, 75/25, 50/50, 25/75, 10/90, and 0/100 for each ionic strength, except that at 0.1 M the ratios of 90/10, 50/50, and 10/90 were not included in order to maintain a uniform volume

(30)

for the surrounding solutions. The 100/0 and 0/100 ratios represent pure NaCl and CaCl² bulk solutions, respectively.

Table 1. Clay models and ionic strengths (I) of surrounding solutions.

Clay model I (M)

Na-MMT 0 (Pure water)

0.125 NaCl 0.250 NaCl 0.500 NaCl 1.000 NaCl

Ca-MMT 0 (Pure water)

0.125 CaCl2

0.250 CaCl2

0.500 CaCl² 1.000 CaCl²

All simulations were performed with the Gromacs package⁸⁰, using the Clayff force field⁷⁸ and SPC water model⁷⁹ for all interactions of the clay, ions, and water molecules. The simulations were carried out at 300 K and 1 bar. After the structures were initially minimized, the system was equilibrated in the isothermal-isobaric ensemble (NPT). The production run was a continuation of the equilibration stage using the canonical ensemble (NVT). Simulation time for each production run was 50 ns and the last 40 ns were analyzed for the swelling pressure calculations.

(31)

Table 2. Bulk solution constituents for each simulation system, with the total ionic strength (I) and ionic strength percentage ratio of NaCl to CaCl² in the salt solutions (INaCl(%)/ICaCl2(%)).cNaCl andcCaCl2 represent the concentration of NaCl and CaCl² in the system, respectively.

I (M) INaCl(%)/ICaCl2(%) cNaCl (M) cCaCl2 (M)

0.1 100/0 0.100 0.000

75/25 0.075 0.008

25/75 0.025 0.025

0/100 0.000 0.033

0.3 100/0 0.300 0.000

90/10 0.270 0.010

75/25 0.225 0.025

50/50 0.150 0.050

25/75 0.075 0.075

10/90 0.030 0.090

0/100 0.000 0.100

1.0 100/0 1.000 0.000

90/10 0.900 0.033

75/25 0.750 0.083

50/50 0.500 0.167

25/75 0.250 0.250

10/90 0.100 0.300

0/100 0.000 0.333

3.1.2 Swelling pressure of Na-MMT and Ca-MMT in pure water

For comprehensive understanding and effective comparison, the swelling pressures of Na-MMT and Ca-MMT in pure water were initially compared. Figure 9 shows that at low and moderate dry densities, Na-MMT typically generates a higher swelling pressure than Ca-MMT. This can be attributed to the absence of osmotic or continuous swelling in Ca-MMT. However, beyond a given dry density, Ca- MMT generates a higher swelling pressure instead. At a low dry density, Na-MMT exhibits both osmotic and crystalline swelling, while crystalline swelling dominates the high dry density regime. Crystalline swelling is affected by the hydration of cations. As Ca²⁺ tends to have more water of hydration than Na⁺, more severe swelling is observed in Ca-MMT at a higher dry density. The steep swelling pressure curve of Ca-MMT at high dry density had been previously reported⁵⁶.

(32)

Figure 9. Comparison of the swelling pressure of Na-MMT and Ca-MMT in pure water.

3.1.3 Swelling pressure of Na-MMT in NaCl solutions

The effect of ionic strength on the swelling pressure of Na-MMT in NaCl solutions is shown in figure 10. It is evident from the pressure curves that the highest pressures occur in pure water, while the swelling pressure decreases significantly upon increasing the ionic strength, especially in the intermediate dry density region. Similar observations have been previously reported10,52,55,81,82.

The influence of ionic strength on the swelling pressure has been attributed to a few factors. Crystalline swelling is the dominant swelling mechanism at higher dry density. It has only minimal influence on the pressure as it is independent of salinity^55,57,81. However, as the dry density decreases, the double layer repulsion becomes the main contributor to the swelling pressure. A higher salinity significantly reduc- es the repulsion and thereby the swelling pressure. In addition, the influence of salinity has been rationalized by the changing chemical potential of water, as water moves from a region at higher potential to one at a lower potential. The larger the potential gradient, the higher the swelling pressure. The potential gradient between the interlayer water and the bulk solution is the highest in pure water, and so the highest swelling pressure is generated. Meanwhile, the swelling pressure decreases as the salinity increases.

(33)

Figure 10. Swelling pressure of Na-MMT in pure NaCl solutions.

3.1.4 Swelling pressure of Ca-MMT in CaCl2 solutions

The influence of CaCl² solutions on the swelling pressure of Ca-MMT is presented in figure 11. Compared to Na-MMT, only minimal change in swelling pressure was observed in Ca-MMT as the ionic strength increased. As mentioned earli- er, the swelling of Ca-MMT is restricted to the crystalline regime due to the absence of DDL swelling, and the hydration of ions and clay surface are the dominant sources of swelling^21,83–85. The crystalline regime is less affected by the ionic strength, which accounts for the small change in pressure as the ionic strength increases. Across the ionic strength considered, there is practically no swelling pressure when the dry density is below 1000 kgm³ due to the absence of osmotic swelling regime in Ca-MMT⁸⁶.

(34)

Figure 11. Swelling pressure of Ca-MMT in pure CaCl2 solutions.

3.1.5 Swelling pressure of Na-MMT in mixed salt solutions

The abovementioned swelling pressures of Na-MMT and Ca-MMT in pure solutions are important for understanding of the behavior of Na-MMT in mixed solutions. Figure 12 presents the swelling pressure curves of Na-MMT in single-salt solutions at low, moderate, and high ionic strengths (0.1, 0.3, and 1.0 M, respectively). At 0.1 M ionic strength, the swelling pressure curves for both solutions mimic the trend in figure 9, and the effect of salt composition is minimal. At moderate ionic strength, the two curves are distinguishable, implying significant influence of salt composition. In pure NaCl, the swelling pressure continuously increases with increasing dry density, while that in pure CaCl2 the curve turns up sharply at 1400 kgm³. As the ionic strength increases to 1.0 M, the swelling pressure curves in both NaCl and CaCl² systems are completely different from those at 0.1 M, but rather resemble the characteristic behavior of Ca-MMT (figure 9 and 11). This could be a result of excess cations present in the interlayer at this high ionic strength.

The swelling pressures at 900, 1200, 1400, 1500, and 1550 kgm³ were calculated and presented in table 3, based on the swelling pressure curves for all the systems presented in table 2. The results show the dependence of the swelling pressure on the ionic strength and salt composition of the surrounding solutions. The impact of salt composition on the swelling pressure is found to depend on the total ionic strength of the surrounding solutions and the dry density of the clay.

(35)

Figure 12. Swelling pressure of Na-MMT in pure NaCl and pure CaCl2 solutions at ionic strengths: (a) 0.1 M, (b) 0.3 M, and (c) 1.0 M.

At the dry densities of 900 and 1200 kgm³, the highest swelling pressures were recorded at a high ionic strength percentage ratio for NaCl. The general influence of salt composition is minimal at 900 kgm³, while a slight decreasing trend is observed at 1200 kgm³as the NaCl ionic strength percentage ratio decreases. At 1400 kgm³, a significant decreasing trend is seen as the NaCl ionic strength percentage ratio decreases. The dry density of 1500 kgm³ appears to be a critical value indicat- ing changes in the swelling pressure behavior and the impact of salt composition.

For low and moderate ionic strengths, the swelling pressure decreases as the NaCl ratio decreases, while a contrary behavior is found at high ionic strength. The contrary behavior extends to 1550 kgm³. At 1550 kgm³, for all ionic strengths the swelling pressure increases as the NaCl ionic strength percentage ratio decreases and the highest swelling pressure was recorded in pure CaCl² solutions.

The observed impact of salt composition on the swelling pressure of Na-MMT can be explained by the previously discussed clay-ion and clay-water interactions,

(36)

such as the double layer theory, its limitations, and the cation exchange capacity13,17,43,48,49. They were discussed in detail in manuscript II.

Table 3. Swelling pressure of Na-MMT in mixed salt solutions with varying ionic strength percentage ratios of NaCl and CaCl². (Pressures are in MPa)

I (M) INaCl(%)/ICaCl2(%) Dry density (kgm³)

900 1200 1400 1500 1550

0.1 100/0 2.6 5.4 9.6 15.0 20.5

75/25 2.3 5.2 10.1 15.8 20.8

25/75 2.3 4.7 8.9 15.6 23.1

0/100 2.4 4.2 7.4 14.4 24.0

0.3 100/0 2.0 5.0 9.9 14.8 18.5

90/10 2.1 4.5 8.9 14.5 19.5

75/25 1.5 3.6 6.8 11.4 18.3

50/50 1.4 3.7 7.2 12.5 20.5

25/75 0.4 2.4 8.0 14.6 19.7

10/90 1.6 3.0 6.2 13.3 22.6

0/100 1.0 2.2 4.4 11.3 25.8

1.0 100/0 2.2 3.3 4.6 8.2 15.6

90/10 1.4 2.6 4.1 7.4 14.2

75/25 1.7 2.8 4.6 8.9 15.3

50/50 1.6 2.6 4.0 7.4 14.5

25/75 1.7 2.1 3.8 10.4 21.1

10/90 1.5 1.6 3.4 10.0 19.6

0/100 1.4 1.4 2.7 11.3 27.9

3.1.6 Ion population analysis for the interlayer of Na-MMT

Activities in the interlayer space of montmorillonite are mainly responsible for the characteristic behavior of Na-MMT. The diffusion of both cations and anions have been reported to take place in the interlayer⁸⁷. The hydration of the interlayer cations is solely responsible for crystalline swelling⁵⁷. Therefore, to further understand the behavior of Na-MMT in mixed salt solutions, the interlayer charge and ion composition of the clay in salt environment were analyzed, using the interlayer space for representative cases in table 2.

For dry clay minerals, the cations are adsorbed on the clay surface. In the presence of water molecules, the adsorbed ions become hydrated and migrate to the midpoint of the clay layers^17,21,43. The total ion distribution between the layers and surrounding solution is dependent on the chemical composition of the surrounding solution. Considering the amount of Na⁺ in the interlayer at the beginning of the

(37)

simulation, the total interlayer charge of the clay was used to normalize that of Na- MMT clay at equilibrium in different solutions.

In 0.1 M solutions, the chemical potential of the surrounding water is relatively high compared to that of the interlayer water. Therefore, water molecules diffuse into the clay interlayers and some ions are lost in the surroundings. As presented in figure 13, the total interlayer charge becomes less than 100 % for all systems considered. In addition, the total interlayer charge is found to depend on the d-spacing of the clay layers.

As the d-spacing increases, the percentage charge of the cation decreases. The attractive forces between the clay surface and the cations is more dominant at lower d-spacings. Consequently, a smaller number of cations are lost to the surroundings.

As the NaCl ratio in the surrounding solution decreases, Na⁺ in the interlayer space is gradually replaced by Ca²⁺. At 0.1 M, the solutions are invariably dilute, therefore even for pure CaCl² solution there is a significant amount of Na⁺ in the interlayer.

Finally, the presence of negative charge in the interlayer is negligible at this ionic strength.

At moderate ionic strength (0.3 M, figure 14), the total interlayer charge is higher than the case at 0.1 M (figure 13). At equilibrium, the percentage charge in the interlayer is approximately 100 %, which shows that the net movement of cations is negligible. The replacement of Na⁺ by Ca²⁺ in the interlayer increases significantly as the ratio NaCl decreases. At 25/75 (figure 14d), the charge percentage contribution of Ca²⁺ is readily dominant. Moreover, the contribution of Cl is observed at higher d-spacing.

Figure 15 shows that at high ionic strength (1.0 M), the total interlayer positive charge is higher than 100 % for all systems irrespective of the salt composition. This excess positive charge in the interlayer can be accounted for by the influx of ions into the interlayer driven by the diffusion gradient. The excess positive charge in the interlayer also provides an explanation of the swelling pressure trend in 1.0 M pure NaCl (figure 12c). The increased positive charge in the interlayer generates a swelling pressure trend similar to the characteristics for Ca-MMT. Furthermore, the total interlayer charge increases with increasing d-spacing of the clay layers, because at a larger d-spacing more cations will be required to balance the concentration gradient. As the ratio of NaCl decreases, the percentage contribution Ca²⁺ to the interlayer charge also increases significantly.

(38)

Figure 13. Total interlayer charge for Na-MMT in mixed salt solutions at the total ionic strength of 0.1 M, with the ionic strength percentage ratios of (a) 100/0, (b) 75/25, (c) 25/75, and (d) 0/100. The blue, pink, and cyan bars denote Na⁺, Ca²⁺, and Cl respectively.

At 75/25, the interlayer charge is predominantly from Ca²⁺. Furthermore, the contribution from negative charge (Cl ) is visible at all d-spacings and shows a pro- nounced dependence on the d-spacing. This added negative charge balances the excess positive charge in the interlayer. The ion exchange process as well as the migration of anion into the clay interlayer is supported by experimental and computational findings43,75,87–89.

(39)

Figure 14. Total interlayer charge for Na-MMT in mixed salt solutions at a total ionic strength of 0.3 M, with the ionic strength percentage ratios of (a) 100/0, (b) 75/25, (c) 50/50, (d) 25/75, and (e) 0/100. The blue, pink, and cyan bars denote Na⁺, Ca²⁺, and Cl respectively.

(40)

Figure 15. Total interlayer charge for Na-MMT in mixed salt solutions at 1.0 M total ionic strength, with the ionic strength percentage ratios of (a) 100/0, (b) 75/25, (c) 50/50, (d) 25/75, and (e) 0/100. The blue, pink, and cyan bars denote Na⁺, Ca²⁺, and Cl respectively.

(41)

3.2 Effect of temperature on the swelling pressure of Na-smectite clay^III

3.2.1 Model setup

The model set up described in section 3.1.1 was slightly modified to examine the influence of temperature on the swelling pressure of bentonite clay. A smectite clay model with the chemical formula Na^0.5(Si^7.75Al^0.25)(Al^3.75Mg^0.25)O²⁰(OH)⁴ was used in this study. A layer charge of -0.5 e was evenly distributed in the tetrahedral and octahedral sheets, and it was balanced with Na⁺ placed on the clay surface. The clay layers were surrounded by water or ice crystals for the high- and low-temperature analysis, respectively (figure 16).

Figure 16. Simulation setup for the swelling pressure of Na-smectite clay in (a) water environment and (b) bulk ice.

MD simulations were first carried out at 300 K to serve as a reference point. The total studied temperature range was 150–600 K with an increment of 50 K. Different simulation methods were used in the low- and high-temperature regimes. For the high-temperature study, the system was first equilibrated at 300 K, and then in the production run the temperature was increased with simulated annealing using a stepwise approach from 300–600 K. For the low-temperature analysis, first the melting point of the SPC ice was investigated. Then, the clay layers were surrounded with ice, and the temperature was gradually increased to 150, 200 and 250 K to represent the scenarios before, during and after ice melting, respectively.

(42)

3.2.2Swelling pressure of Na-smectite at low and high tempera- tures

Figure 17 shows the swelling pressure of Na-smectite clay at high temperatures (300–600 K). The swelling pressure increases as the temperature increases, and it is more evident at a higher dry density. This trend can be attributed to the increased pore water pressure as well as the osmotic pressure with respect to the interlayer of the clay when the temperature becomes higher. In addition, the thermal motion of water molecules in the interlayer is also enhanced, thereby increasing the expansion of the DDL28,29,67,69,90.

Figure 17. Swelling pressure of Na-smectite at high temperatures. (a) Swelling pressure as a function of dry density for temperatures from 300 K to 600 K. (b) Swelling pressure as a function of temperature for the dry densities of 1000, 1250, and 1500 kgm³.

For the low-temperature regime, the bulk ice was initially melted to understand its melting behavior. The swelling pressure of smectite clay at lower temperatures is presented in figure 18. At 150 K, no swelling was observed as the bulk ice had not melted yet, albeit a little disorderliness was seen in the structure. At 200 K, there was some melting, while the movements of both water molecules and ions were still restricted. The onset of swelling was observed at approximately 225 K. Howev- er, the presence of an ice-water mixture led to a lot of scatter in the data, and the results could not be used to draw qualitative conclusions. At 250 K, the ice was completely melted, the interlayer ion hydration took place, and a reasonable swelling pressure was generated. The overall influence of temperature is reported in table 4.

(43)

Figure 18. Swelling pressure of Na-smectite as a function of temperature in the low- temperature regime.

Table 4. Swelling pressures of Na-smectite at 150–600 K and three dry densities.

Temperature (K)

Swelling pressure (MPa)

1000 kg m³ 1250 kg m³ 1500 kg m³

150 - - -

200 - - -

250 2.7 5.3 10.3

300 3.5 6.6 12.7

350 3.8 6.8 12.1

400 3.4 6.9 13.8

450 3.6 7.1 14.1

500 3.9 7.6 15.0

550 4.8 8.3 14.4

600 4.7 8.6 15.6

(44)

The table shows a general increasing trend in the swelling pressure, when the system changes from deep frozen to ambient and then to high temperatures. The influence of temperature is more significant when moving from the deep-frozen regime to ambient temperature than that above the ambient temperature, due to the significantly different entropy of water molecules in the different states. For the ice environment, the entropy of water molecules is completely lost, and kinetic energy is lost. Swelling is not possible in this case, as the water molecules are strongly held together by hydrogen bonding while the interlayer cations are not hydrated. As the temperature increases, the hydrogen bonds begin to break and the water molecules gain kinetic energy. This results in cation hydration and consequently clay swelling.

(45)

4 CONCLUSIONS

The research presented in the thesis focuses on the impacts of two external factors on the swelling pressure of bentonite clay. The influence of salt solutions and temperature on the swelling pressure of smectite clays were critically and extensively considered. These results help to fill the gap in the computational study of montmorillonite clay swelling pressure in certain environmental conditions.

The influence of salinity on the interlayer swelling pressure revealed that Na-MMT is more sensitive to a saline environment than Ca-MMT. In addition, the study of Na-MMT in mixed NaCl and CaCl² solutions shows the impact of cation exchange on the swelling pressure of Na-MMT. It was found that as the ionic strength percentage ratio of CaCl² increases, the swelling pressure of Na-MMT tends towards that of the typical Ca-MMT. At lower dry densities, the highest swelling pressures were recorded at high NaCl ionic strength percentage ratios, while at a high dry density they were found at the highest ionic strength percentage ratio of CaCl². At a critical dry density of 1500 kg m³, the influence of mixed salt solution on the interlayer swelling pressure of Na-MMT depends on the total ionic strength of the surrounding solution.

Based on the observed behavior of clays in mixed solutions, the migration of ions into the interlayer and the ion exchange process were systematically analyzed.

The total interlayer cation percentage was found to be dependent on the ionic strength of the surrounding solutions. The presence and population of anions in the interlayer is dependent on the total ionic strength of the surrounding solutions and the d-spacing of the clay layers.

Moreover, the temperature of water surrounding the clay layers was varied to understand the impact on the swelling behavior of a smectite clay (Na-smectite).

Above 300 K, the swelling pressure increases slightly as the temperature increases.

The swelling pressure of Na-smectite clay is highly sensitive to the temperature at low temperatures. In a frozen environment, the swelling pressure is lost due to the restricted motion of ions and water molecules. As the ice melts, cation hydration begins to take place, and the swelling pressure builds up. A more rapid increase in swelling pressure can be observed when the clay-water system temperature increases from the frozen state to ambient.

This research also verified that the modified spring model is effective for predicting the interlayer swelling pressure of smectite clay under different environmental conditions. Overall, the results presented here have practical applications in the construction and maintenance of repositories for disposing high-level radioactive wastes, which often employ bentonite clay as a barrier and/or backfill material.

(46)

5 BIBLOGRAPHY

1. Birkholzer, J., Houseworth, J. & Tsang, C.-F. Geologic disposal of high-level radioactive waste: Status, key issues, and trends.Annu. Rev. Environ. Resour.

37, 79–106 (2012).

2. Chapman, N. Geological disposal of radioactive wastes–concept, status and trends.J. Iber. Geol.32, 7–14 (2006).

3. Gordon, H. Journey Deep into the Finnish Caverns Where Nuclear Waste Will Be Buried for Millenia.

4. Raiko, H. Disposal canister for spent nuclear fuel-Design report.Posiva Rep.

2, 61 (2005).

5. Pastina, B. & Hellä, P. Expected evolution of a spent nuclear fuel repository at Olkiluoto. (Posiva Oy, 2006).

6. Vieno, T. & Ikonen, A. T. K. Plan for Safety Case of spent fuel repository at Olkiluoto. (Posiva Oy, 2005).

7. Johanna Hansen,Posiva and ONKALO status, (DOPAS. 2016).

8. Arthur, R., Apted, M. & Stenhouse, M.Comment on the long-term chemical and mineralogical stability of the buffer. (Swedish Nuclear Power Inspectorate, 2005).

9. Bergaya, F. & Lagaly, G.Handbook of clay science.5, (Newnes, 2013).

10. Karnland, O., Olsson, S. & Nilsson, U. Mineralogy and sealing properties of various bentonites and smectite-rich clay materials. Technical Report TR-06-30.

CM Digitaltryck AB (2006). doi:10.1109/56.9304

11. Christidis, G. E. & Huff, W. D. Geological aspects and genesis of bentonites.

Elements5, 93–98 (2009).

12. Pusch, R.Bentonite clay: environmental properties and applications. (CRC Press, 2015).

13. Mitchell, J. K. Fundamentals of Soil Behaviour. Wile)’.New York. NY (1976).

14. Guggenheim, S. & Martin, R. T. Definition of clay and clay mineral: joint report of the AIPEA nomenclature and CMS nomenclature committees.

Clays Clay Miner.43, 255–256 (1995).

15. Svensson, D. The Bentonite Barrier-Swelling Properties, Redox Chemistry and Mineral Evolution. (Lund University, 2015).

16. Karnland, O. Bentonite swelling pressure in strong NaCl solutions. Correlation between model calculations and experimentally determined data. (Swedish Nuclear Fuel and Waste Management Co., 1997).

17. Schanz, T. & Tripathy, S. Swelling pressure of a divalent-rich bentonite:

Diffuse double-layer theory revisited.Water Resour. Res.45, (2009).

18. Karnland, O. & Birgersson, M.Montmorillonite Stability with Special Respect to KBS-3 Conditions, Montmorillonite Stability with Special Respect to KBS-3 Conditions. (TR-06-11, Swedish Nuclear Fuel and Waste Management Co (SKB), Stockholm, Sweden, 2006).

Molecular dynamics study of the bentonite barrier : effects of salinity and temperature on the swelling pressure

uef.fi

Dissertations in Forestry and Natural Sciences

BUKUNMI AKINWUNMI

MOLECULAR DYNAMICS STUDY OF THE BENTONITE BARRIER: EFFECTS OF SALINITY AND TEMPERATURE ON THE SWELLING PRESSURE

BUKUNMI AKINWUNMI

Bukunmi Akinwunmi

MOLECULAR DYNAMICS STUDY OF THE BENTONITE BARRIER: EFFECTS OF SALINI- TY AND TEMPERATURE ON THE SWELLING

PRESSURE

ABSTRACT

ACKNOWLEDGEMENTS

LIST OF ABBREVIATIONS

LIST OF ORIGINAL PUBLICATIONS

AUTHOR’S CONTRIBUTION

CONTENTS

1 INTRODUCTION

2 COMPUTATIONAL DETAILS

3 RESULTS AND DISCUSSION

4 CONCLUSIONS

5 BIBLOGRAPHY