Computational Study on Removal of Naphthenic Acids from Petroleum-Based Systems

University of Calgary PRISM: University of Calgary's Digital Repository

Graduate Studies The Vault: Electronic Theses and Dissertations

2017-12 Computational Study on Removal of Naphthenic Acids from Petroleum-based Systems

Wu, Chongchong

Wu, C. (2017). Computational Study on Removal of Naphthenic Acids from Petroleum-based Systems (Unpublished doctoral thesis). University of Calgary, Calgary, AB. http://hdl.handle.net/1880/106332 doctoral thesis

University of Calgary graduate students retain copyright ownership and moral rights for their thesis. You may use this material in any way that is permitted by the Copyright Act or through licensing that has been assigned to the document. For uses that are not allowable under copyright legislation or licensing, you are required to seek permission. Downloaded from PRISM: https://prism.ucalgary.ca UNIVERSITY OF CALGARY

Computational Study on Removal of Naphthenic Acids from Petroleum-based Systems

Chongchong Wu

A THESIS

SUBMITTED TO THE FACULTY OF GRADUATE STUDIES

IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE

DEGREE OF DOCTOR OF PHILOSOPHY

GRADUATE PROGRAM IN CHEMICAL AND PETROLEUM ENGINEERING

CALGARY, ALBERTA

JANUARY, 2018

Naphthenic acids (NAs) are toxic compounds found in crude oil and products of petroleum processing including produced water. Reactions of hydroxyl radicals with NAs and ionic liquids

(ILs) extraction have been reported to be effective in removing NAs from petroleum-based systems. However, the reaction and interaction mechanisms are not fully understood. In this thesis, a density functional theory study was conducted to explore mechanisms and kinetics of reactions between benzoic acid (BA), benzoate (BZ) and hydroxyl radicals as well as the extraction mechanisms of model NAs by ILs. The relationships between physicochemical properties of ILs and their intramolecular and intermolecular interactions were also investigated. The results show that all reaction pathways between BA, BZ and hydroxyl radicals involve the formation of pre- reactive complexes. The reaction rate constants for the addition reactions are highest for BZ in the aqueous phase, followed by BA in the aqueous phase, then by BA in the gas phase. For interactions between ILs and NAs, the main extraction mechanism by 1-butyl-3-methylimidazolium tetrafluoroborate ([BMIM][BF4]) for NAs without long alkyl chain is hydrogen bonding, whereas van der Waals interaction and hydrogen bonding are the dominant extraction mechanisms for NAs with long alkyl chain. With the incorporation of biodegradable substitutional groups, hydrogen bonding is still the main extraction mechanism, however, the interaction energy is higher than that of [BMIM][BF4]. More intermolecular hydrogen bonds occur when model NAs are absorbed by

[BMIM][BF4] with -COOH or -COOCH3. Additionally, the interaction energy between model NAs and ILs with -COOH or -COOCH3 is higher than that with -OH, -NH2, or -OCH3. Aggregation behaviour and hydrogen bonds in ILs influence their densities and self-diffusion coefficients.

ii Preface

The thesis is based on five manuscripts that have been published or in preparation for submission.

(1) Chapter 2 is modified from a review paper “Wu, C., De Visscher, A., and Gates, I. D. Removal

of naphthenic acids from oil sands process-affected water and crude oil” in preparation for

submission to Critical Reviews in Environmental Science and Technology.

(2) Chapter 4 is based on “Wu, C., De Visscher, A., and Gates, I. D. (2017). Reactions of hydroxyl

radicals with benzoic acid and benzoate. RSC Advances, 7(57), 35776-35785.”

(3) Chapter 5 is published as “Wu, C., De Visscher, A., and Gates, I. D. (2017). Molecular

interactions between 1-butyl-3-methylimidazolium tetrafluoroborate and model naphthenic

acids: A DFT study. Journal of Molecular Liquids, 243, 462-471.”

(4) Chapter 6 is published as “Wu, C., De Visscher, A., and Gates, I. D. Interactions of

biodegradable ionic liquids with a model naphthenic acid (2018).” Scientific Reports, 8, 176.

(5) Chapter 7 is based on a paper “Wu, C., De Visscher, A., and Gates, I. D. Comparison of

electronic and physicochemical properties between imidazolium-based and pyridinium-based

ionic liquids” in preparation for submission to Journal of Physical Chemistry B.

iii Acknowledgements

First and foremost, I would like to express my gratitude to my supervisor Dr. Ian Donald Gates.

He is a trustworthy supervisor and provides me enormous confidence whenever I discuss my research with him. I appreciate him working really hard to revise my manuscripts and thesis even in the midnight and on weekends. I feel so touched whenever I receive his comments on 2 AM. In addition, he inspires me to carry on different areas of research and I feel quite excited about his brilliant research ideas. Moreover, he gives me freedom to make my own research plan and his big smile always motivates me to work hard. Thank you for creating the fantastic research group!

Secondly, I want to convey my sincere thanks to my co-supervisor, Dr. Alex De Visscher. He is a knowledgeable professor and provides a lot of useful insights to my research. I will never forget the bubble dynamics equation, different Matlab codes, notes, thick chalk powders, and discussions on the blackboard. No matter how I perform, he always says I have done very well. Although he is in Montreal now, he keeps on instructing me through skype chat and email. His valuable suggestions improve our manuscripts and thesis. Thank you for the continuous support from you!

I would like to express my gratitude to my supervisory committee members Dr. Kunal Karan and

Dr. Hector De la Hoz Siegler for their time and precious suggestions. I also want to acknowledge

Dr. Craig Johansen, Dr. Edward Roberts, and Dr. Phillip Choi for being examiners for my candidacy exams and final defense. I am grateful for the computational support from Doug

Phillips, Jin Lin, Weijie Yang, China University of Petroleum (Qsing dao Campus), and Westgrid as well as the help from Arthur De Vera and Suha Abusalim.

iv My sincere thanks go to all the members of CCIT 313 for creating such a harmonious and hilarious working environment: Antonio Vazquez Zamora, Gaurav Patel, Maureen Austin-Adigo, Pachari

Detpunyawat, Samaneh Ashoori, Temilola Famakinwa, Dr. Mahta Sadeghvishkaei, and Dr. Yi Su.

I want to specially thank Jacky Wang for setting up lab equipment, office computers and software for me. I am thankful to the understanding from Dr. Ranjani Kannaiyan. I am also grateful for all the encouragements from my friends: Chunyang Yu, Jicong Cao, Shuying Wu, Yan Xiang, and

Yani Dong.

I am indebted to the endless and unconditional love from my parents and my godfather. You are the most perfect parents! I really appreciate what you have done and all your sacrifices to make a better life for me. Without your encouragement and support, I could not finish my graduate study.

I also want to convey my gratitude for my younger brother and other family members. A strong and united family is always the source of my motivation. Special thanks to my talented fiancé for his guidance, encouragement, support, being my biggest fan for my crazy research and life ideas, and being my doraemon with magic pocket.

Dedication

To my wonderful parents

vi Table of Contents

Abstract ...... ii Preface ...... iii Acknowledgements ...... iv Dedication ...... vi Table of Contents ...... vii List of Tables ...... x List of Figures and Illustrations ...... xi List of Schemes ...... xiii List of Symbols, Abbreviations and Nomenclature ...... xiv

Chapter One: Introduction ...... 1 1.1 Background ...... 1 1.2 Ultrasonic Treatment ...... 4 1.3 Ionic Liquid Extraction ...... 6 1.4 Research Objectives ...... 7 1.5 Organization of Thesis...... 9 1.6 References ...... 10

Chapter Two: Literature Review ...... 16 2.1 Introduction ...... 16 2.2 Properties of NAs ...... 17 2.3 NAs in OSPW ...... 21 2.4 NAs in Crude Oil ...... 23 2.5 NAs Removal from OSPW ...... 25 2.5.1 Oxidation ...... 25 2.5.2 Membrane Filtration ...... 32 2.5.3 Adsorption ...... 34 2.5.4 Coagulation and Flocculation ...... 36 2.5.5 Biodegradation ...... 37 2.6 NAs Removal from Crude Oil ...... 42 2.6.1 Decarboxylation ...... 42 2.6.2 Hydrogenation ...... 45 2.6.3 Thermal Decomposition ...... 46 2.6.4 Esterification ...... 48 2.6.5 Neutralization ...... 50 2.6.6 Adsorption ...... 51 2.6.7 Solvent Extraction ...... 53 2.6.8 Ionic Liquid Extraction ...... 54 2.7 Knowledge Gaps ...... 57 2.8 References ...... 58

Chapter Three: Computational Methods ...... 74 3.1 Computational Chemistry ...... 74 3.2 Density Functional Theory ...... 76 3.2.1 Basis Sets ...... 78

vii 3.2.2 Methods ...... 79 3.2.3 Solvation Models ...... 80 3.3 Molecular Dynamics ...... 81 3.4 Structure Analyses ...... 82 3.4.1 NBO Analysis ...... 83 3.4.2 AIM Analysis ...... 84 3.4.3 NCI Analysis ...... 85 3.5 References ...... 86

Chapter Four: Reactions of Hydroxyl Radicals with Benzoic Acid and Benzoate ....91 4.1 Introduction ...... 91 4.2 Computational Methods ...... 93 4.3 Results and Discussion...... 94 4.3.1 Reaction Pathways ...... 94 4.3.2 Pre-reactive Complexes and Transition States ...... 97 4.3.3 Energetics of the Reaction Paths ...... 101 4.3.4 Reaction Rate Constants ...... 105 4.3.5 Influence of Explicit Water Molecule ...... 109 4.4 Conclusions ...... 111 4.5 References ...... 111

Chapter Five: Molecular Interactions between 1-Butyl-3-Methylimidazolium Tetrafluoroborate and Model Naphthenic Acid: A DFT Study ...... 117 5.1 Introduction ...... 117 5.2 Computational Methods ...... 119 5.3 Results and Discussion...... 120 5.3.1 Optimized Geometries ...... 120 5.3.2 Interaction Energies ...... 125 5.3.3 NBO Analyses ...... 126 5.3.4 The Topological Properties of Interactions ...... 128 5.3.5 NCI Analyses ...... 133 5.3.6 The HOMO-LUMO Overlap Integral Analyses ...... 137 5.3.7 Electron Density Difference Analyses ...... 140 5.4 Conclusions ...... 142 5.5 References ...... 142

Chapter Six: Interactions of Biodegradable Ionic Liquids with a Model Naphthenic Acid...... 147 6.1 Introduction ...... 147 6.2 Computational Methods ...... 149 6.3 Results and Discussion...... 150 6.3.1 Optimized Geometries ...... 150 6.3.2 Interaction Energies ...... 157 6.3.3 NBO Analyses ...... 158 6.3.4 Topological Properties of Interactions ...... 162 6.3.5 NCI Analyses ...... 166 6.3.6 Electron Density Difference Analyses ...... 172

viii 6.4 Conclusions ...... 173 6.5 References ...... 174

Chapter Seven: Comparison of Electronic and Physicochemical Properties between Imidazolium-based and Pyridinium-based Ionic Liquids ...... 178 7.1 Introduction ...... 178 7.2 Computational Methods ...... 181 7.3 Results and Discussion...... 182 7.3.1 Characterization of the Cations ...... 182 7.3.2 Geometric Structures ...... 184 7.3.3 Intermolecular Interaction Energy ...... 187 7.3.4 Averaged Noncovalent Interaction Analyses ...... 189 7.3.5 Densities ...... 194 7.3.6 Self-diffusion Coefficients Analyses ...... 196 7.4 Conclusions ...... 198 7.5 References ...... 199

Chapter Eight: Conclusions and Recommendations ...... 204 8.1 Conclusions ...... 204 8.2 Recommendations ...... 206

Appendix A: Supplementary Materials for Chapter Four ...... 208

Appendix B: Supplementary Materials for Chapter Five...... 218

Appendix C: Supplementary Materials for Chapter Six ...... 237

Appendix D: Copyright Permissions ...... 251

ix List of Tables

Table 2.1 Concentration and elementary analyses of different NAs samples (Grewer et al. 2012)...... 18

Table 2.2 MW, density, and pKa of model NAs...... 21

Table 2.3 Saturation capacities for different adsorbents...... 34

Table 2.4 Microbial degradation of NAs...... 39

Table 4.1 Reaction rate constants of BA and BZ with hydroxyl radical...... 107

Table 5.1 Electron densities () and Laplacian of electron density (2) of BCPs in [BMIM][BF4] and [BMIM][BF4]-NAs complexes...... 131

Table 6.1 Bond lengths (Å) of ILs-CHCA complexes...... 154

Table 6.2 Interaction energies between ILs and CHCA...... 158

Table 6.3 The donor-acceptor interaction in ILs-CHCA complexes, and their stabilization energies, E(2)(kcal/mol)...... 160

x List of Figures and Illustrations

Figure 1.1 The formation and collapse of cavitation bubbles as well as three reaction zones in the cavitation process (Wang & Xu 2012)...... 5

Figure 2.1 Samples of NA structures (Chemente et al. 2005)...... 19

Figure 3.1 Perturbation donor-acceptor interaction...... 84

Figure 4.1 Reactants optimization...... 96

Figure 4.2 Transition states optimization (BA gas phase)...... 99

Figure 4.3 Transition states optimization (BA aqueous phase)...... 100

Figure 4.4 Transition states optimization (BZ aqueous phase)...... 101

Figure 4.5 Relative energies of six possible reaction paths (BA gas phase)...... 102

Figure 4.6 Relative energies of six possible reaction paths (BA aqueous phase)...... 103

Figure 4.7 Relative energies of three possible reaction paths (BZ aqueous phase)...... 104

Figure 4.8 Electrostatic potential analysis...... 109

Figure 5.1 Structures of [BMIM][BF4] and six types of model NAs compounds...... 119

Figure 5.2 The electrostatic potential (a.u.) of [BMIM][BF4] and model NAs...... 122

Figure 5.3 The optimized structures of [BMIM][BF4] and [BMIM][BF4]-NAs complexes...... 123

Figure 5.4 The electrostatic potential (a.u.) of [BMIM][BF4]-NAs complexes...... 125

Figure 5.5 BCPs and bond paths in [BMIM][BF4] and model NAs...... 130

Figure 5.6 The sign(λ2)ρ vs RDG (left) and the gradient isosurfaces (right) for [BMIM][BF4] and [BMIM][BF4]-NAs...... 136

Figure 5.7 HOMO and LUMO analyses of [BMIM][BF4] and model NAs...... 139

Figure 6.1 Chemical structures of ILs with biodegradable groups and CHCA...... 149

Figure 6.2 The electrostatic potential (a.u.) of ILs and CHCA...... 152

Figure 6.3 The optimized structures of ILs-CHCA complexes...... 156

Figure 6.4 BCPs and bond paths in ILs and ILs-CHCA complexes...... 165

Figure 6.5 The sign(λ2)ρ vs RDG (left) and the gradient isosurfaces (right) for ILs...... 168

xi Figure 6.6 The sign(λ2)ρ versus RDG (left) and the gradient isosurfaces (right) for ILs-CHCA complexes...... 172

Figure 7.1 The structures of cations and anions...... 180

Figure 7.2 Charges (black), Mulliken bond orders (red), and electrostatic potential of [BMIM]+ and [BMPy]+...... 183

Figure 7.3 Typical arrangement of cations and anions in ILs...... 186

Figure 7.4 Intermolecular interaction energy of ILs...... 188

Figure 7.5 vdW interaction energy of ILs...... 189

Figure 7.6 Plots of RDG versus sign(2) for ILs...... 192

Figure 7.7 RDG isosurfaces for ILs...... 194

Figure 7.8 Calculated densities of ILs...... 196

Figure 7.9 Self-diffusion coefficients of ILs...... 198

xii List of Schemes

Scheme 2.1 Corrosion mechanisms in oil refining process (Alvisi and Lins 2011)...... 23

Scheme 2.2 CHCA degradation by UV/H2O2 (Drzewicz et al. 2010)...... 28

Scheme 2.3 Sonolytic degradation of BA (Singla et al. 2013)...... 31

Scheme 2.4 α- and β- oxidation of CHAA (Whitby 2010)...... 41

Scheme 2.5 NA decarboxylation reaction mechanisms using MgO as catalyst (Dias et al. 2015)...... 44

Scheme 2.6 NAs decarboxylation reaction mechanisms with the presence of supercritical water (Mandal et al. 2012)...... 45

Scheme 2.7 Hydrogenation mechanisms of BA (Yokoyama & Yamagata 2001)...... 46

Scheme 2.8 Thermal decomposition mechanisms of BA (Winter and Barton 1970)...... 47

Scheme 2.9 Main reaction pathways for BA reaction...... 49

Scheme 2.10 The adsorption of NAs by acid-ion exchange resin (Mediaas et al. 2003)...... 52

Scheme 2.11 Chemical interactions between NAs and amino acid-based ILs (Anderson et al. 2013)...... 55

Scheme 4.1 Reaction pathways of BA and BZ with hydroxyl radicals...... 96

xiii List of Symbols, Abbreviations and Nomenclature

Symbol Definition AIM atoms in molecules avgNCI averaged noncovalent interaction avgRDG averaged reduced density gradient BA benzoic acid bcp bond critical point [BMIM][BF4] 1-butyl-3-methylimidazolium tetrafluoroborate [BMIM][PF6] 1-butyl-3-methylimidazolium hexafluorophosphate [BMIM][HSO4] 1-butyl-3-methylimidazolium hydrogen sulfate [BMIM][MSO4] 1-butyl-3-methylimidazolium methylsulfate [BMIM][ESO4] 1-butyl-3-methylimidazolium ethylsulfate [BMPy][BF4] 1-butyl-1-methylpyrrolidinium tetrafluoroborate [BMPy][PF6] 1-butyl-1-methylpyrrolidinium hexafluorophosphate [BMPy][HSO4] 1-butyl-1-methylpyrrolidinium hydrogen sulfate [BMPy][MSO4] 1-butyl-1-methylpyrrolidinium methylsulfate [BMPy][ESO4] 1-butyl-1-methylpyrrolidinium ethylsulfate BO bond order BZ benzoate ccp cage critical pont CCSD(T) coupled-cluster with single, double, and perturbative triple excitations CHAA cyclohexaneacetic acid CHBA cyclohexanebutyric acid CHCA cyclohexanecarboxylic acid CHDCA 1,4-cyclohexanedicarboxylic acid CHPA cyclohexanepentanoic acid CMC critical micelle concentration COD chemical oxygen demand COMPASS condensed-phase optimized molecular potentials for atomistic simulation studies CPCA cyclopentanecarboxylic acid DA decanoic acid DCHA dicyclohexylacetic acid DFT density functional theory DFT-D3 dispersion corrected density functional theory DNP double numerical plus polarization E(2) stabilization energy 퐸푒푒 electron-electron repulsion energy 퐸푥푐 exchange-correction energy GGA generalized gradient approximation Ψ wave function Ĥ Hamiltonian operator H-abs hydrogen-abstraction HOMA harmonic oscillator measure of aromaticity

xiv HOMO highest occupied molecular orbital HPLC high performance liquid chromatography HRMS high resolution mass spectrometry IER ion exchange resin ILs ionic liquids ipso-add ipso-addition IRC intrinsic reaction coordinate LP lone pair LUMO lowest unoccupied molecular orbital m-add meta-addition MDs molecular dynamics MEUF micellar-enhanced ultrafiltration MP2 Møller–Plesset perturbation MSD mean square displacement MW molecular weight NAs naphthenic acids NBO natural bond orbital n-BPBA (4´-n-butylphenyl)-4-butanoic acid NCI noncovalent interactions ncp nuclear critical point o-add ortho-addition OSPW oil sands process-affected water p-add para-addition polyDADMAC polydiallyldimethylammonium chloride QST3 quasi-Newton synchronous transit rcp ring critical point RDG reduced density gradient SMD solvation model based on density STO slater-type orbital TAN total acid number TOC total organic carbon t-BPBA (4´-t-butylphenyl)-4-butanoic acid UV ultraviolet vdW van der Waals 푉푒푓푓 effective potential 푉푥푐 exchange-correlation potential VMD visual molecular dynamics

xv Chapter One: Introduction

1.1 Background

The oil sands resources in Alberta are the third largest crude oil reserves in the world next to Saudi

Arabia and Venezuela (Nikakhtari et al. 2013). According to Natural Resources Canada, the total proven oil reserves in Canada are estimated to be about 173 billion barrels and of this amount, 168 billion barrels are in Alberta (Natural Resources Canada, 2016). It is estimated that the production of oil sands will increase to 4 million barrels per day by 2024 compared to 2.3 million barrels in

2014 (Alberta Energy, 2015). Most of the oil sands in Alberta are located in the Athabasca, Peace

River, and Cold Lake deposits. They are composed of 6 to 16 weight percent bitumen, 1-8 weight percent water, and 80 to 87 weight percent sand, silt, and clay (Kannel & Gan 2012). The oil sands deposits in Alberta are a huge asset for Canada.

Hot water extraction is the most common process used to separate bitumen from oil sands (Martin et al. 2010; Painter et al. 2010). A large amount of oil sands process-affected water (OSPW) is generated in this process. Typically, about three barrels of water are used to produce one barrel of oil (Allen 2008a, 2008b). OSPW consists of unrecovered bitumen, sands, silts, heavy metals, and organic and inorganic compounds (Morandi et al. 2015; Quesnel et al. 2015). The organic compounds in OSPW are mixtures of naphthenic acids (NAs), benzene, humic and fulvic acids, phenols, and polycyclic aromatic hydrocarbons (Wang et al. 2013). Due to a zero-discharge policy in Alberta, OSPW cannot be discharged into the environment directly and thus, it is accumulated in large tailing ponds (Martin et al. 2010; Quesnel et al. 2015). Part of OSPW is recycled in the

1 mining process. However, the amount of OSPW has still been increasing over the past four decades

(Beck et al. 2014). The surface area of tailing water in northeastern Alberta grew by 422% between

1992 and 2008 (Timoney & Ronconi 2010). It is estimated that over 1 billion m3 of OSPW will be accumulated in the Athabasca by the year 2025 (Kannel & Gan 2012). Due to high costs and environmental problems of OSPW, OSPW treatment is an important challenge for the Canadian oil sands industry (Pérez-Estrada et al. 2011).

NAs are the primary toxic components in OSPW and among its most persistent organic compounds

(Drzewicz et al. 2012; Frank et al. 2008; He et al. 2012; Mohammed et al. 2009). They are the primary targets for OSPW treatment and the most important indicators of potential downstream effects after release (Giesy et al. 2010). NAs are a mixture of monocyclic and polycyclic carboxylic acids with smaller amount of acyclic acids (Drzewicz et al. 2012; He et al. 2011), and they are also loosely defined including all aromatic (Hsu et al. 2000), cyclic, and acyclic acids (Colati et al.

2013). It is reported that NAs have acute and chronic toxicity to goldfish, larval zebrafish,

Pimephales promelas, Vibrio fischeri, and the mammalian immune system (Mohseni et al. 2015).

NAs with concentration between 40 and 65 mg/L caused a 50% suppression of chemiluminescence in the bacterium Vibrio fischeri (Reinardy et al. 2013). Futhermore, results have shown that NAs extracted from OSPW are acutely toxic to fathead minnow and rats (Scarlett et al. 2012). Yellow perch and Japanese medaka had some developmental abnormalities, including increased incidence of deformities and decreased larval length regardless of whether they were exposed to NAs extracted from OSPW or commercial NAs (Anderson et al. 2012).

2 NAs are acidic components of crude oil and petroleum (Clemente & Fedorak 2005). More specifically, NAs are the most abundant carboxylic acids in petroleum (Jones et al. 2001). NAs in the Athabasca oil sands ores are reported to originate from the biodegradation of petroleum hydrocarbons (Scott et al. 2005). The presence of NAs in crude oil induces a variety of concerns for the oil refining industry. In addition to leading to generation of substantial amount of hazardous

OSPW during the bitumen extraction process, NAs cause serious corrosion problems to refinery equipment, transportation pipelines, and storage tanks, increasing costs of these systems (Khan et al. 2017). Moreover, the structures of NAs enable them to act as surfactants (Poteau et al. 2005) and bring about crude oil emulsion formation and stabilization during the refining process, rendering it difficult to separate water from oil (Varadaraj & Brons 2007). Furthermore, they can cause foam in refinery units and leach cations (such as Ca2+, Na+) in the desalting process, thereby deactivating catalysts (Shah et al. 2014). Additionally, the existence of high concentration NAs in crude oil reduces crude oil quality and sales price.

Considering that the presence of NAs in OSPW and crude oil has serious adverse effects on the environment, organisms, and refining processes, it is highly desirable to develop efficient, cost- effective, and environmentally friendly approaches to remove NAs from OSPW and crude oil to decrease the negative effects associated with NAs. To improve water quality, protect aquatic ecosystems, and make the oil sands industry sustainable, researchers have focused on OSPW treatment, especially NAs removal from OSPW. In addition, isolating NAs from heavy oil can help reduce their negative effects during the refining process, increase oil quality, and prevent the formation of NAs in OSPW. Therefore, NAs removal is considered to be an important process of crude oil upgrading process (Zhang et al. 2006) and the oil sands industry (Scott et al. 2005).

1.2 Ultrasonic Treatment

Ultrasound is sound waves resulting from mechanical vibrations and having frequency above the hearing limit of the human ear, typically from above 15 kHz to l0 MHz (Award et al. 2012; Riesz

& Kondo 1992). A growing interest has begun to use ultrasound to destroy organic contaminants in water since 1990 (Petrier et al. 1998). Ultrasonic treatment is a promising technology for NAs degradation because it has signiﬁcant advantages such as safety, cleanliness, modular design, absence of treatment chemicals, and energy conservation without causing secondary pollution in comparison with other physical methods such as high voltage corona and ultraviolet light (Hao et al. 2003).

As depicted in Figure 1.1, when water is treated by ultrasonic waves, the cavitation effect of ultrasound generates the formation and collapse of gas bubbles (De Visscher et al. 2003), which contain water vapors, gases, and volatile compounds (Weavers et al. 1998). During bubble collapse, the compressed gases and vapors in the cavity generate intense heat and raise the temperature of the liquid surrounding the cavity immediately while hardly changing the bulk liquid temperature

(Suslick 1989; Weavers et al. 1998). Due to the high temperatures and pressures in the center of the collapsed bubbles, compounds in the cavitation bubbles undergo chemical reactions through pyrolysis (De Visscher et al. 1997). The organic compounds will also be oxidized by hydroxyl radicals (OH), hydrogen peroxide (H2O2), and other oxidants generated in cavitation bubbles during ultrasonic treatment (De Visscher & Van Langenhove 1998).

Figure 1.1 The formation and collapse of cavitation bubbles as well as three reaction zones in the cavitation process (Wang & Xu 2012).

Sonochemical reaction mainly occurs in three different regions in aqueous solution (Cravotto &

Cintas 2006; Krishna et al. 1989), as shown in Figure 1.1. The first region is the interior of collapsing cavitation bubbles. In this region, extremely high temperature and pressure exist transiently, and volatile solutes are pyrolyzed. For instance, the sonolysis of water molecules produces H and OH radicals (Okuno et al. 2000; Tu and Yen 2000). The second region is the interfacial liquid region between the cavitation bubbles and bulk solution. Although the temperature here is lower than that of the interior bubbles, it is still high enough to rupture chemical bonds, induce free radical reactions, and cause thermal decomposition (Kumar et al. 2000; Okuno et al. 2000). The third region is the bulk solution with ambient temperature, where the solutes react with radicals that are escaped from the interior or interfacial region, but have not yet recombined, disproportionated or been scavenged (Dewulf et al. 2001). Generally speaking, when irradiated with ultrasound, organic compounds in water can be degraded through thermal decomposition or radical reaction in the interior and interface region, and undergo oxidation reaction with hydroxyl radicals and hydrogen peroxide in the bulk solution (Ince et al. 2001).

1.3 Ionic Liquid Extraction

Ionic liquids (ILs) are salts existing as liquids at room temperatures and below (as low as -96oC), and they are entirely composed of cations and anions (Huddleston et al. 1998). The properties of

ILs can be tuned to specific capabilities by changing cations and anions, thus, ILs are also known as “designer solvents” (Wasserscheid and Keim 2000). Owing to their attractive physical and chemical properties, such as no detectable vapor pressure, environmentally friendly, thermal stability, nonflammability, and relatively favorable viscosities and densities, they have promising applications in research and industry (Pinkert et al. 2009; Rogers and Seddon 2003), and are also referred to as “green solvents” (Earle and Seddon 2000).

It is noted that they are potential alternative solvents for volatile organic compounds in liquid- liquid extraction processes (Huddleston et al. 1998). A variety of organic cations such as 1-butyl-

3-methylimidazolium and 1-butyl-3-methylpyridinium, and anions such as tetrafluoroborate, hexafluorophosphate, hydrogen sulfate, methylsulfate, and ethylsulfate can be used for the synthesis of ILs. Therefore, taking account of the structures of ILs, most types of interactions, including dispersive, π‐π, hydrogen bonding, and dipole-dipole, can occur when ILs interact with other compounds (Lü et al. 2017).

Recently, researchers have focused on the separation of NAs from crude oil by ILs through experimental study, and significant progress has been made. Shi et al. utilized in situ synthesized imidazole-based ILs to separate NAs from Beijiang crude oil (Shi et al. 2008). Anderson et al.

6 (2003) used amino acid-based ILs and carbonate-based ILs to remove NAs from Doba crude oil

(Anderson et al. 2003). Sun et al. attempted to use imidazole anion-based ILs to separate NAs from crude oil (Sun and Shi 2012). Shah et al. investigated the isolation of NAs from highly acidic model oil using imidazolium-based ILs, hydroxide-based ILs, and phenolate-based ILs (Shah et al. 2014; Shah et al. 2016a; Shah et al. 2016c). Furthermore, the separation of NAs from crude oil was also realized through employing thiocyanate-based ILs (Najmuddin et al. 2016; Shah et al.

2016b). Additionally, Duan et al. attempted nitrogen-containing alkaline ILs to isolate NAs (Duan et al. 2013). Moreover, it is demonstrated that 1-butyl-3-methylimidazolium octylsulfate combing with subcritical methanol could help decrease total acid number (TAN) of NAs under mild conditions (Zafar et al. 2016).

1.4 Research Objectives

Based on the literature review presented in Chapter 2, the knowledge gaps are listed below:

 There are limited studies about the application of ultrasound to treat NAs especially on the mechanisms and kinetics. Ultrasonic degradation of organic compounds could happen both inside the bubbles and in the bulk liquid phase, however, few detailed investigations have been conducted to compare the reaction mechanisms and kinetics differences of model NAs with hydroxyl radicals in the gas and liquid phases without the addition of catalysts.

 Although researchers have focused on ILs extraction of NAs through experimental studies, no comprehensive theoretical study has been reported to inquire into the interaction mechanisms between ILs and model NAs.

 The most commonly used ILs have low biodegradability, however, there are few studies

7 exploring the design of biodegradable ILs and the influence of biodegradable ILs on the extraction of NAs.

 Few fundamental articles have been published to compare electronic and physicochemical properties of ILs with similar structures, and explore the influence of intermolecular and intramolecular interactions on physicochemical properties of ILs.

The main objectives of the research reported in this thesis are to fill the knowledge gaps and investigate the removal of NAs from OSWP and crude oil through ultrasonic irradiation and ILs extraction. The detailed objectives are listed as follows:

 Select benzoic acid (BA) as a model NA. Use Gaussian software to determine the reaction mechanisms and rate constants between BA and hydroxyl radicals both in the gas and aqueous phases, and decide whether the reactions inside the bubbles or in the water phase are more beneficial for BA degradation.

 Explore the extraction mechanisms of six different types of NAs by ILs, and provide theoretical support for experimental research.

 Design ILs that have high biodegradability and removal efficiency for NAs, investigate the influence of biodegradable substitutional groups on the extraction mechanisms and removal efficiency, and determine properties of substitutional groups that are more favorable for NAs removal.

 Compare the electronic and physicochemical properties differences between imidazolium- based ILs and pyridinium-based ILs, and explore the influence of intermolecular and intramolecular interactions on the physicochemical properties of ILs.

8 The outcomes of the research will provide guidance for the degradation of NAs by hydroxyl radicals, especially through ultrasonic irradiation. In addition, it will contribute to the understanding of the extraction mechanisms of NAs by ILs, and assist researchers to design ILs with high biodegradability and high extraction efficiency. Moreover, the current study will be conducive to synthesize ILs with favorable physicochemical properties.

1.5 Organization of Thesis

The thesis is composed of eight chapters. Chapter 2 is a literature review that introduces chemical structures, physical properties of NAs, states the adverse effects of NAs in OSPW and crude oil, and summarizes the technologies that are widely used to remove NAs from petroleum-based systems. Chapter 3 mainly focuses on the methodology used in this thesis, describing density functional theory, molecular dynamics simulation, simulation methods, and analytical techniques.

Chapter 4 presents the study of the kinetics and mechanisms reactions of hydroxyl radicals with

BA and benzoate (BZ) using density functional theory both in the gas and aqueous phases.

Chapter 5 concentrates on the extraction mechanisms of six model NAs by ILs, and compares the interaction differences between different types of NAs. Chapter 6 emphasizes the design of biodegradable ILs for the removal of model NAs considering both biodegradability and extraction efficiency. It describes the influence of the incorporation of biodegradable substitutional groups on the structures of ILs, and investigates the extraction mechanisms of NAs using those biodegradable ILs. Chapter 7 mainly investigates the influences of intermolecular and intramolecular interactions on the physicochemical properties of ILs combining density functional

9 theory with molecular dynamics simulation. Chapter 8 summarizes the main research conclusions and provides recommendations for future research directions.

1.6 References

Anderson, J., Wiseman, S. B., Moustafa, A., El-Din, M. G., Liber, K., & Giesy, J. P. (2012). Effects of exposure to oil sands process-affected water from experimental reclamation ponds on Chironomus dilutus. Water Research, 46(6), 1662-1672. Anderson, K., Goodrich, P., Hardacre, C., Hussain, A., Rooney, D. W., & Wassell, D. (2013). Removal of naphthenic acids from crude oil using amino acid ionic liquids. Fuel, 108, 715-722. Allen, E. W. (2008a). Process water treatment in Canada's oil sands industry: I. Target pollutants and treatment objectives. Journal of Environmental Engineering and Science, 7(2), 123-138. Allen, E. W. (2008b). Process water treatment in Canada's oil sands industry: II. A review of emerging technologies. Journal of Environmental Engineering and Science, 7(5), 499-524. Awad, T. S., Moharram, H. A., Shaltout, O. E., Asker, D., & Youssef, M. M. (2012). Applications of ultrasound in analysis, processing and quality control of food: A review. Food Research International, 48(2), 410-427. Beck, E. M., Smits, J. E., & St. Clair, C. C. (2014). Health of domestic mallards (Anas platyrhynchos domestica) following exposure to oil sands process-affected water. Environmental Science & Technology, 48(15), 8847-8854. Clemente, J. S., & Fedorak, P. M. (2005). A review of the occurrence, analyses, toxicity, and biodegradation of naphthenic acids. Chemosphere, 60(5), 585-600. Colati, K. A., Dalmaschio, G. P., de Castro, E. V., Gomes, A. O., Vaz, B. G., & Romão, W. (2013). Monitoring the liquid/liquid extraction of naphthenic acids in brazilian crude oil using electrospray ionization FT-ICR mass spectrometry (ESI FT-ICR MS). Fuel, 108, 647-655. Cravotto, G., & Cintas, P. (2006). Power ultrasound in organic synthesis: moving cavitational chemistry from academia to innovative and large-scale applications. Chemical Society Reviews, 35(2), 180-196. De Visscher, A., Van Langenhove, H., & Van Eenoo, P. (1997). Sonochemical degradation of

10 ethylbenzene in aqueous solution: a product study. Ultrasonics Sonochemistry, 4(2), 145-151. De Visscher, A., & Van Langenhove, H. (1998). Sonochemistry of organic compounds in homogeneous aqueous oxidising systems. Ultrasonics Sonochemistry, 5(3), 87-92. De Visscher, A. (2003). Kinetic model for the sonochemical degradation of monocyclic aromatic compounds in aqueous solution: new insights. Ultrasonics Sonochemistry, 10(3), 157-165. Drijvers, D., De Baets, R., De Visscher, A., & Van Langenhove, H. (1996). Sonolysis of trichloroethylene in aqueous solution: volatile organic intermediates. Ultrasonics Sonochemistry, 3(2), S83-S90. Drzewicz, P., Perez-Estrada, L., Alpatova, A., Martin, J. W., & El-Din, M. G. (2012). Impact of peroxydisulfate in the presence of zero valent iron on the oxidation of cyclohexanoic acid and naphthenic acids from oil sands process-affected water. Environmental Science & Technology, 46(16), 8984-8991. Duan, J., Sun, Y., & Shi, L. (2013). Three different types of heterocycle of nitrogen-containing alkaline ionic liquids treatment of acid oil to remove naphthenic acids. Catalysis Today, 212, 180-185 Earle, M. J., & Seddon, K. R. (2000). Ionic liquids. Green solvents for the future. Pure and Applied Chemistry, 72(7), 1391-1398. Frank, R. A., Fischer, K., Kavanagh, R., Burnison, B. K., Arsenault, G., Headley, J. V., Kerry, M. P., Kraak, G.V.D., & Solomon, K. R. (2008). Effect of carboxylic acid content on the acute toxicity of oil sands naphthenic acids. Environmental Science & Technology, 43(2), 266-271. Giesy, J. P., Anderson, J. C., & Wiseman, S. B. (2010). Alberta oil sands development. Proceedings of the National Academy of Sciences, 107(3), 951-952. Hao, H., Wu, M., Chen, Y., Yin, Y., & Lü, Z. (2003). Cavitation-induced pyrolysis of toxic chlorophenol by high frequency ultrasonic irradiation. Environmental Toxicology, 18(6), 413- 417. He, Y., Wiseman, S. B., Hecker, M., Zhang, X., Wang, N., Perez, L. A., Jones, P.D., El-Din, M. G., Martin, J.W., & Giesy, J. P. (2011). Effect of ozonation on the estrogenicity and androgenicity of oil sands process-affected water. Environmental Science & Technology, 45(15), 6268-6274. He, Y., Wiseman, S. B., Wang, N., Perez-Estrada, L. A., El-Din, M. G., Martin, J. W., & Giesy, J. P. (2012). Transcriptional responses of the brain–gonad–liver axis of fathead minnows exposed

11 to untreated and ozone-treated oil sands process-affected water. Environmental Science & Technology, 46(17), 9701-9708. Hsu, C. S., Dechert, G. J., Robbins, W. K., & Fukuda, E. K. (2000). Naphthenic acids in crude oils characterized by mass spectrometry. Energy & Fuels, 14(1), 217-223. Huddleston, J. G., Willauer, H. D., Swatloski, R. P., Visser, A. E., & Rogers, R. D. (1998). Room temperature ionic liquids as novel media for ‘clean’ liquid–liquid extraction. Chemical Communications, (16), 1765-1766. Ince, N. H., Tezcanli, G., Belen, R. K., & Apikyan, İ. G. (2001). Ultrasound as a catalyzer of aqueous reaction systems: the state of the art and environmental applications. Applied Catalysis B: Environmental, 29(3), 167-176. Jones, D. M., Watson, J. S., Meredith, W., Chen, M., & Bennett, B. (2001). Determination of naphthenic acids in crude oils using nonaqueous ion exchange solid-phase extraction. Analytical Chemistry, 73(3), 703-707. Kannel, P. R., & Gan, T. Y. (2012). Naphthenic acids degradation and toxicity mitigation in tailings wastewater systems and aquatic environments: A review. Journal of Environmental Science and Health, Part A, 47(1), 1-21. Khan, M. K., Riaz, A., Yi, M., & Kim, J. (2017). Removal of naphthenic acids from high acid crude via esterification with methanol. Fuel Processing Technology, 165, 123-130. Krishna, C. M., Kondo, T., & Riesz, P. (1989). Sonochemistry of alcohol-water mixtures. Spin- trapping evidence for thermal decomposition and isotope-exchange reactions. The Journal of Physical Chemistry, 93(13), 5166-5172. Kumar, R. V., Mastai, Y., & Gedanken, A. (2000). Sonochemical synthesis and characterization of nanocrystalline paramelaconite in polyaniline matrix. Chemistry of Materials, 12(12), 3892- 3895. Lü, R., Wu, C., Lin, J., Xiao, Y., Wang, F., & Lu, Y. (2017). The study on interactions between 1‐ ethyl‐3‐methylimidazolium chloride and benzene/pyridine/pyrrole/thiophene. Journal of Physical Organic Chemistry, 30(8), e3663. Martin, J. W., Barri, T., Han, X., Fedorak, P. M., El-Din, M. G., Perez, L., Scott, A.C., & Jiang, J. T. (2010). Ozonation of oil sands process-affected water accelerates microbial bioremediation. Environmental Science & Technology, 44(21), 8350-8356. Mohammed, M. A., & Sorbie, K. S. (2009). Naphthenic acid extraction and characterization from

12 naphthenate field deposits and crude oils using ESMS and APCI-MS. Colloids and Surfaces A: Physicochemical and Engineering Aspects, 349(1), 1-18. Mohseni, P., Hahn, N. A., Frank, R. A., Hewitt, L. M., Hajibabaei, M., & Van Der Kraak, G. (2015). Naphthenic acid mixtures from oil sands process-affected water enhance differentiation of mouse embryonic stem cells and affect development of the heart. Environmental Science & Technology, 49(16), 10165-10172. Morandi, G. D., Wiseman, S. B., Pereira, A., Mankidy, R., Gault, I. G., Martin, J. W., & Giesy, J. P. (2015). Effects-directed analysis of dissolved organic compounds in oil sands process- affected water. Environmental Science & Technology, 49(20), 12395-12404. Najmuddin, R. A., Mutalib, M. A., Shah, S. N., Suleman, H., Lethesh, K. C., Pilus, R. B. M., & Maulud, A. S. (2016). Liquid-liquid extraction of naphthenic acid using thiocyanate based ionic liquids. Procedia Engineering, 148, 662-670. Nikakhtari, H., Vagi, L., Choi, P., Liu, Q., & Gray, M. R. (2013). Solvent screening for non‐ aqueous extraction of Alberta oil sands. The Canadian Journal of Chemical Engineering, 91(6), 1153-1160. Okuno, H., Yim, B., Mizukoshi, Y., Nagata, Y., & Maeda, Y. (2000). Sonolytic degradation of hazardous organic compounds in aqueous solution. Ultrasonics Sonochemistry, 7(4), 261-264. Painter, P., Williams, P., & Lupinsky, A. (2010). Recovery of bitumen from Utah tar sands using ionic liquids. Energy & Fuels, 24(9), 5081-5088. Pérez-Estrada, L. A., Han, X., Drzewicz, P., Gamal El-Din, M., Fedorak, P. M., & Martin, J. W. (2011). Structure–reactivity of naphthenic acids in the ozonation process. Environmental Science & Technology, 45(17), 7431-7437. Petrier, C., Jiang, Y., & Lamy, M. F. (1998). Ultrasound and environment: sonochemical destruction of chloroaromatic derivatives. Environmental Science & Technology, 32(9), 1316- 1318. Pinkert, A., Marsh, K. N., Pang, S., & Staiger, M. P. (2009). Ionic liquids and their interaction with cellulose. Chemical Reviews, 109(12), 6712-6728. Poteau, S., Argillier, J. F., Langevin, D., Pincet, F., & Perez, E. (2005). Influence of pH on stability and dynamic properties of asphaltenes and other amphiphilic molecules at the oil−water interface. Energy & Fuels, 19(4), 1337-1341. Quesnel, D. M., Oldenburg, T. B., Larter, S. R., Gieg, L. M., & Chua, G. (2015). Biostimulation

13 of oil sands process-affected water with phosphate yields removal of sulfur-containing organics and detoxification. Environmental Science & Technology, 49(21), 13012-13020. Reinardy, H. C., Scarlett, A. G., Henry, T. B., West, C. E., Hewitt, L. M., Frank, R. A., & Rowland, S. J. (2013). Aromatic naphthenic acids in oil sands process-affected water, resolved by GCxGC-MS, only weakly induce the gene for vitellogenin production in zebrafish (Danio rerio) larvae. Environmental Science & Technology, 47(12), 6614-6620. Riesz, P., & Kondo, T. (1992). Free radical formation induced by ultrasound and its biological implications. Free Radical Biology and Medicine, 13(3), 247-270. Rogers, R. D., & Seddon, K. R. (2003). Ionic liquids--solvents of the future? Science, 302(5646), 792-793. Scarlett, A. G., West, C. E., Jones, D., Galloway, T. S., & Rowland, S. J. (2012). Predicted toxicity of naphthenic acids present in oil sands process-affected waters to a range of environmental and human endpoints. Science of the Total Environment, 425, 119-127. Scott, A. C., Mackinnon, M. D., & Fedorak, P. M. (2005). Naphthenic acids in Athabasca oil sands tailings waters are less biodegradable than commercial naphthenic acids. Environmental Science & Technology, 39(21), 8388-8394. Shah, S. N., Mutalib, M. I. A., Pilus, R. B. M., & Lethesh, K. C. (2014). Extraction of naphthenic acid from highly acidic oil using hydroxide-based ionic liquids. Energy & Fuels, 29(1), 106- 111. Shah, S. N., Chellappan, L. K., Gonfa, G., Mutalib, M. I. A., Pilus, R. B. M., & Bustam, M. A. (2016a). Extraction of naphthenic acid from highly acidic oil using phenolate based ionic liquids. Chemical Engineering Journal, 284, 487-493. Shah, S. N., Ismail, M., Mutalib, M. I. A., Pilus, R. B. M., & Chellappan, L. K. (2016b). Extraction and recovery of toxic acidic components from highly acidic oil using ionic liquids. Fuel, 181, 579-586. Shah, S. N., Mutalib, M. A., Ismail, M. F., Suleman, H., Lethesh, K. C., & Pilus, R. B. M. (2016c). Thermodynamic modelling of liquid-liquid extraction of naphthenic acid from dodecane using imidazolium based phenolate ionic liquids. Journal of Molecular Liquids, 219, 513-525. Shi, L. J., Shen, B. X., & Wang, G. Q. (2008). Removal of naphthenic acids from Beijiang crude oil by forming ionic liquids. Energy & Fuels, 22(6), 4177-4181.

14 Sun, Y., & Shi, L. (2012). Basic ionic liquids with imidazole anion: New reagents to remove naphthenic acids from crude oil with high total acid number. Fuel, 99, 83-87. Suslick, K. S. (1989). The chemical effects of ultrasound. Scientific American, 260(2), 80-86. Timoney, K. P., & Ronconi, R. A. (2010). Annual bird mortality in the bitumen tailings ponds in northeastern Alberta, Canada. The Wilson Journal of Ornithology, 122(3), 569-576. Tu, S. P., & Yen, T. F. (2000). The feasibility studies for radical-induced decomposition and demetalation of metalloporphyrins by ultrasonication. Energy & Fuels, 14(6), 1168-1175. Varadaraj, R., & Brons, C. (2007). Molecular origins of heavy crude oil interfacial activity part 2: Fundamental interfacial properties of model naphthenic acids and naphthenic acids separated from heavy crude oils. Energy & Fuels, 21(1), 199-204. Wang, J. L., & Xu, L. J. (2012). Advanced oxidation processes for wastewater treatment: formation of hydroxyl radical and application. Critical Reviews in Environmental Science and Technology, 42(3), 251-325. Wang, N., Chelme-Ayala, P., Perez-Estrada, L., Garcia-Garcia, E., Pun, J., Martin, J. W., Belosevic, M., & El-Din, M. G. (2013). Impact of ozonation on naphthenic acids speciation and toxicity of oil sands process-affected water to Vibrio fischeri and mammalian immune system. Environmental Science & Technology, 47(12), 6518-6526. Wasserscheid, P., & Keim, W. (2000). Ionic liquids—new “solutions” for transition metal catalysis. Angewandte Chemie International Edition, 39(21), 3772-3789. Weavers, L. K., Ling, F. H., & Hoffmann, M. R. (1998). Aromatic compound degradation in water using a combination of sonolysis and ozonolysis. Environmental Science & Technology, 32(18), 2727-2733. Zafar, F., Mandal, P. C., & Moniruzzaman, M. (2016). Total acid number reduction of naphthenic acid using subcritical methanol and 1-butyl-3-methylimidazolium octylsulfate. Procedia Engineering, 148, 1074-1080. Zhang, A., Ma, Q., Wang, K., Liu, X., Shuler, P., & Tang, Y. (2006). Naphthenic acid removal from crude oil through catalytic decarboxylation on magnesium oxide. Applied Catalysis A: General, 303(1), 103-109.

15 Chapter Two: Literature Review

2.1 Introduction

Naphthenic Acids (NAs) are naturally present in crude oil, oil sands, and groundwater associated with bitumen deposits (Hughes et al. 2017). Almost all crude oils globally, including China,

Eastern Europe, the US, and Canada, contain NAs (Slavcheva et al. 1999). The composition and concentration of NAs varies with the source of the oil. The concentration of NAs in crude oil typically can reach as high as 4% by weight (Headley et al. 2007). More specifically, crude oils from Russia, Romania, and Poland contain on the order of 3 weight percent NAs (Rudzinski et al.

2002). The production of Athabasca bitumen is estimated to reach 3.8 million barrels per day

(McKenzie et al. 2014). NAs account for 2 weight percent of the bitumen samples with 90% being tricyclic acids in Athabasca bitumen (Headley et al. 2004). The existence of NAs in crude oil induces a variety of problems for refining, upgrading, separation, and water treatment processes, hence, it is a major concern for the petroleum industry to isolate NAs from crude oil.

Apart from issues triggered by NAs in crude oil refining processes, exploitation of Canada oil sands to produce bitumen through the commonly used Clarke hot alkaline water extraction process brings about the release and solubilization of NAs in oil sands process-affected water (OSPW)

(Clemente et al. 2003a), especially under highly alkaline conditions. The concentration of NAs in

OSPW is estimated to range from 40 to 120 mg/L (Clemente et al. 2003a). It is noted that the toxicity of OSPW is mainly ascribed to the high concentration of NAs (Scarlett et al. 2012). NAs in OSPW cause serious environmental problems, and they exhibit acute and chronic toxicity to a

16 range of organisms. Moreover, they are recalcitrant and difficult to be naturally biodegraded. For surface mining, most of the NAs end up in the tailings ponds. The area of tailing ponds has been estimated to be more than 176 km2 (Sohrabi et al. 2013), taking up a large area of unused land and water. The volume of mature fine tailings stored in Athabasca ponds will be on the order of 1 billion m3 by the year 2025 (Kannel and Gan 2012). Consequently, removing NAs is a primary target for treating OSPW and tailings ponds to make the oil sands operations sustainable.

Given that the removal of NAs is critical to the petroleum industry, a comprehensive literature review about NAs and their separation techniques is necessary and essential. This literature review provides a thorough overview of chemical and physical properties of NAs, toxicity of NAs, to elucidate the negative effects of NAs when they are present in OSPW and crude oil. Moreover, existing techniques to isolate and treat NAs will be summarized and compared. Furthermore, the mechanisms of those approaches will be provided.

2.2 Properties of NAs

NAs are traditionally defined as a mixture of monocyclic and polycyclic carboxylic acids with smaller amounts of acyclic acids (Drzewicz et al. 2012). The general chemical formula for NAs is

CnH2n+ZO2, where n indicates the number of carbon atoms, and Z is zero or a negative even integer representing hydrogen deficiency as a result of ring formation or double bonds (Reinardy et al.

2012; Ross et al. 2012). Figure 2.1 shows examples of NAs with different Z number. NAs with

Z=0 are acyclic, and their structures are more common with branched-chains than straight-chains

(Del Rio et al. 2006). Rings with 5 or 6 carbons are the predominant structures (approximately

17 95%), while the majority of NAs in OSPW have 2 (Z=-4) or 3 (Z=-6) rings (Clemente et al. 2003b;

Del Rio et al. 2006). For NAs with 7 to 12 carbons, monocyclic acids are the main structures

(Brient et al. 1995). The carboxylic group is usually bonded or attached to a side chain instead of to the cycloaliphatic ring. It should be noted that aromatic acids are also minor components of NAs in a broader sense (Colati et al. 2013; Hsu et al. 2000). Moreover, the general chemical formula for NAs are not adequate to describe acids with more than one carboxylic group, nitrogen-, and sulfur- containing organic acids that are also belongs to NAs (Rowland et al. 2011). According to the elemental analyses in Table 2.1, nitrogen- and sulfur- containing acids exist almost in all NAs extracted from different OSPW samples and tailing ponds.

Table 2.1 Concentration and elementary analyses of different NAs samples (Grewer et al. 2012). NAs Sample Concentration C H O N S sources mg/L Syncrude Mildred Lake Setting 44 70.10 8.68 15.98 0.68 3.80 Basin West In-Pit 60 69.82 8.70 16.40 0.51 3.38 Pond 9 20 59.71 6.82 20.44 0.24 1.06 Suncor Pond 2/3 63 71.12 9.13 16.34 0.69 3.85 Pond 5 38 67.34 8.45 17.68 0.38 2.91 Steam assisted gravity 130 68.53 7.66 15.12 0.56 6.90 drainage Albian fresh Tailings ponds 35 69.28 8.32 16.24 0.00 4.30 waters

18 Z=0 CH3(CH2)mCOOH

Z=-2 R R (CH2)mCOOH

(CH2)mCOOH

Z=-4 R (CH2)mCOOH (CH2)mCOOH

(CH2)mCOOH

R Z=-6

R (CH2)mCOOH (CH2)mCOOH

(CH2)mCOOH R

Figure 2.1 Samples of NA structures (Chemente et al. 2005).

Note: R is an alkyl chain, Z is the hydrogen deficiency, and m is the number of CH2 units.

NAs with same n and Z have different isomers, and different molecular structures. The composition of NAs is extremely complicated and varies with the source of crude oil. It is demonstrated that there are more than 200,000 different NAs structures in oil sands (Anderson et al. 2011). The value of n is typically between 5 and 33, therefore, the molecular weight (MW) of

NAs commonly ranges from 100 g/mol to 500 g/mol. (Clemente 2003a, 2003b). Nevertheless, the

19 MW of NAs can reach as high as 1,200 g/mol. For low MW NAs, the composition of NAs includes cyclopentane carboxylic acid (CPCA) and cyclohexane carboxylic acid (CHCA); for high MW

NAs, they are thought to contain 4 carboxylic acids and 4 to 8 unsaturated rings (Mohammed and

Sorbie 2009). However, NAs with more branched and complex structures are more difficult to degrade.

NAs are viscous liquids with unique odor due to the presence of phenolic and sulfur impurities.

The viscosity of NAs depends on the grade of oil, and it is usually between 40 mPa.s to 100 mPa.s

(Whitby 2010). The volatility of NAs is low with around 8.56×10-6 atm.m3/mole as the Henry’s law constant and approximate 2.35×10-6 atm as vapour pressure, and their boiling points are between 250C and 350oC (Kannel and Gan 2012). The acidity of NAs is weaker than that of low

MW carboxylic acids, but is higher than that of phenol and cresylic acid (Brient et al. 1995). In addition, the density of NAs is between 0.97 and 0.99 g/cm3 (Brient et al. 1995). NAs are soluble in water at neutral or alkaline pH with dissociation constant around 10-5 to 10-6 (Xu et al. 2017).

In fact, they mainly exist in the form of sodium naphthenate in OSPW. The MW, density, and pKa of some commonly used model NAs were provided in Table 2.2. The solubility of NAs in water is higher than that of alkanes, alkenes, and polycyclic aromatic hydrocarbons. In addition, their solubility is more influenced by pH than temperature, and alkaline pH promotes the solubility of

NAs (Janfada et al. 2006). On the other hand, they are soluble in organic solvents and oils. The concentration of NAs in different tailing ponds has a big variation (Table 2.1), and it can reach as high as 130 mg/L in steam assisted gravity drainage water (Grewer et al. 2012).

20 Table 2.2 MW, density, and pKa of model NAs. Model NAs Structure MW Density(g/cm3) pKa/25C DA O 172.27 0.8858/40C (Lide 4.9 (Kessenich OH & Milne 1993) 2015)

BA 122.12 1.320/25C 4.2 (Nair 2001) HOOC (Hamilton et al. 1978) 1-naphthoic COOH 172.18 1.398/25C 3.64 (Bouchard et acid (McCrone et al. al. 2002) 1995)

CHCA 128.17 1.0334/22C (Lide 4.88 (Hoefnagel HOOC & Milne 1993) & Wepster 1990)

CPCA 114.14 1.053/20C (Lide & 4.99 (Eftink et al. HOOC 1989) Milne 1993) CHAA 142.20 1.0423/18C (Lide 4.80 (Eftink et al. HOOCH2C & Milne 1993) 1989)

Note: DA represents decanoic acid, BA stands for benzoic acid, CHAA is cyclohexaneacetic acid.

2.3 NAs in OSPW

Although the concentration of NAs in OSPW is not high, it is reported that NAs are the primary toxic components in OSPW (Frank et al. 2008). When keeping the concentrations of other pollutants in tailing ponds from Mildred Lake Settling Basin essentially the same and only removing NAs, the toxicity of OSPW to rainbow trout and water fleas decreased considerably

(Rogers et al. 2002). The toxicity of NAs is influenced by their structures, polarity, relative proportions of different individual acids, and surfactant characteristics (Clemente & Fedorak 2005).

Frank et al. (2008) found that the toxicity of NAs decreased with more carboxylic groups, whereas it increased with higher MW when the carboxylic group was the same. Additionally, more carbon atoms and cyclic rings enhanced the toxicity of NAs, whereas branched chains reduced their toxicity (Xu et al. 2017). Concentration is another factor that influences the toxicity of NAs. NAs

21 with low solubility and strong sorption in soils have lower toxicity (Barrow et al. 2009). On the other hand, NAs in fresh OSPW are more toxic than those in aged OSPW (Bartlett et al. 2017).

Owing to their polar carboxylic groups and nonpolar aliphatic ends, NAs possess both hydrophilic and hydrophobic groups, which cause them to have surfactant characteristics. Consequently, it is more easily for them to penetrate cell walls and threaten different organisms including mammals, fish, and zooplankton even at 100 ppb aqueous concentration (Quagraine et al. 2005a; Smith et al.

2008a). When the hydrophobic group of NAs enters the lipid bilayer, the cell membrane is disrupted and its properties (fluidity, thickness, and surface tension) are changed, leading to the cell death (Frank et al. 2008).

NAs are reported to be toxic to invertebrates, such as Hyalella azteca, Vibrio fischeri, and

Lampsilis cardium (Bartlett et al. 2017). They induce negative effects to the development and endocrine function of zebrafish and decrease birthing rates of fathead minnow (Wang et al. 2015;

Marentette et al. 2015). In addition, the presence of NAs in tailing ponds poses threats to water- associated birds. It is claimed that OSPW decreased the reproductive success of the tree swallow and reduced growth rates of the mallard (Beck et al. 2015). Apart from the negative impacts caused by the toxicity of NAs, the number of birds' mortalities that are oiled and sink in the tailing ponds can reach as high as 10,000 per year (Beck et al. 2015). Exposure to NAs reduced the growth and development of amphibians such as the embryos of Lithobates pipiens and Silurana tropicalis, and leaded to an increase in deformities and mortalities, decreasing their populations (Melvin and

Trudeau 2012). Moreover, NAs are detrimental to mammals, for instance it caused hepatotoxicity,

22 brain hemorrhage, and periarteriolar necrosis and fibrosis in rats (Rogers et al. 2002). When the amount of NAs in water exceeds 11 g/kg, they are also lethal to humans (Headley et al. 2007).

2.4 NAs in Crude Oil

NAs in crude oil cause serious corrosion problems in the oil refining process. This issue has been noted in the oil and gas industry since the 1920s (Meredith et al. 2000). When the temperature is above 230C, corrosion induced by NAs in distillation units is enhanced (Quagraine et al. 2005a).

Usually the rise of the temperature aggravates corrosion due to the augmentation of acid transfer to the metal surface (Alvisi and Lins 2011). The iron reacts with NAs to generate iron naphthenates, which are soluble in oil and are responsible for corrosion of equipment and pipelines. The unit with a significant degree of corrosion is the vacuum distillation unit. The mechanism of NAs corrosion is presented in Scheme 2.1 (Alvisi and Lins 2011, Slavcheva et al. 1999):

Fe + 2RCOOH = Fe(RCOO)2 + H2

Fe + H2S = FeS + H2

Fe(RCOO)2 + H2S = FeS + 2RCOOH

Scheme 2.1 Corrosion mechanisms in oil refining process (Alvisi and Lins 2011).

Sulfur-containing compounds are another source of corrosion in the refining process. However, it is noted that iron reacts with the hydrogen sulfide to produce iron sulfide, which coats the steel surface and acts as a protective film to prevent corrosion (see Scheme 2.1). In addition, owing to more localized attachment areas, corrosion caused by NAs is more serious than sulfidic corrosion

(Kane and Cayard 2002). On the other hand, NAs are able to form metal naphthenate precipitates

23 and salt deposits, blocking pipelines and equipment, and consequently leading to equipment failures (Johnson et al. 2011) and huge expenses for repairs.

NAs lead to the deactivation of heterogeneous catalysts in the refining process due to foaming in various operational units and leaching of cations carried by NAs in the desalting process (Shah et al. 2014). Moreover, the surfactant characteristics of NAs promote the formation and stabilization of oil/water emulsions, making it more difficult to separate water from oil (Saab et al. 2005). In addition, NAs also decrease the thermal stabilities of crude oil products (Wang et al. 2014a).

Furthermore, the presence of NAs decreases crude oil quality and its economic value. The amount of NAs in crude oil can be quantified by the amount of potassium hydroxide (KOH) used to neutralize one gram of oil, i.e. the total acid number (TAN). For crude oil with TAN greater than

0.5 mg KOH/g, the oil is categorized as an acidic oil.

NAs in oil sands ores can reach surface water systems through effluent discharge, ground water mixing, crude oil spills, or erosion of riverbank oil sands deposits (Headley et al. 2007). The concentration of NAs in the surface water sources in the Athabasca oil sands region is less than 1 mg/L, whereas it ranges from 2 to 5 mg/L in near-surface aquifers (Janfada et al. 2006). The contamination of ground water by NAs is another concern in this region.

Pure NAs are reported to be important raw materials in chemical industry, and they can be used as antiseptics, paint drying reagents, and additives in petroleum fluids (Shi et al. 2010). Naphthenate salts are often used as catalysts, preservatives, corrosion inhibitors, emulsifiers, and dispersants

(Wang et al. 2006). The removal of NAs would improve oil quality, solve problems in

24 transportation and refining processes, and provide valuable raw materials. Therefore, there is a need to separate NAs from crude oil and oil products.

2.5 NAs Removal from OSPW

Large volumes of OSPW are generated during the bitumen extraction process, which are usually stored in tailing ponds. Tailing ponds are toxic to a variety of organisms, and they occupy a large area of land that cannot be used for other functions. Additionally, water in the tailing ponds remains unused as well, leading to stranding of valuable water resources. Even worse, contaminated water in the tailing ponds can seep from ponds, leading to pollution of surface and ground water.

Therefore, it is imperative to treat OSPW to decrease its toxicity so that the oil sands industry can be sustainable.

Given that NAs are mainly responsible for the toxicity of OSPW, isolating NAs from OSPW is the priority. Although the acute toxicity of NAs in OSPW may decrease through evaporation or biodegradation, those processes are extremely slow and are not able to remove NAs completely.

NAs are recalcitrant and their concentration in wastewater has hovered at 19 mg/L even after several decades of storage (Johnson et al. 2011). Many options have been tried to remove NAs from OSPW including oxidation, filtration, adsorption, coagulation, flocculation, and biodegradation. A detailed introduction of these treatment options is listed below.

2.5.1 Oxidation

25 Advanced oxidation has been proved to be an efficient approach for treatment of recalcitrant compounds in water. Organic compounds can be partially or completely degraded through reactions with hydroxyl radicals. Advanced oxidation treatment methods for OSPW usually use ozone, titanium dioxide, hydrogen peroxide, activated persulfate, or Fenton reagents as oxidizing agents (Deng & Zhao 2015). These agents can help to remove NAs from OSPW, or degrade NAs to less recalcitrant and more biodegradable organic compounds. Hence, oxidation is also a complementary process for microbial biodegradation treatment (Zhang et al. 2016).

Ozonation is a chemical oxidation process that employs ozone to remove NAs in OSPW and reduce the toxicity of OSPW. Under acidic conditions, ozone directly reacts with NAs; whereas under alkaline conditions, ozone is decomposed to hydroxyl radicals, and the hydroxyl radicals react with NAs. Scott et al. (2008) reported that ozonation of OSPW for 50 min lowered the concentration of NAs by 70%. However, the results also revealed that ozonation could not totally degrade NAs, and that it led to breaking of highly branched chains or cyclic fractions (Kumar et

- - al. 2014). As for degradation of NAs by persulfate (S2O8 ) and manganate (MnO4 ), the results

- from experiments concluded that S2O8 completely mineralized acid-extractable organics and that

- 100% of oxidized NAs were removed. MnO4 mainly transformed the acid-extractable organics into dissolved organic carbon and only 93% of oxidized NAs were oxidized (Sohrabi et al. 2013).

For Fenton reagents, the reaction between hydroxide and ferrous iron generates hydroxyl radicals at acidic pH, resulting in the oxidation of NAs. Zhang et al. indicated that nitrilotriacetic acid- modified Fenton reagent provided a potential approach for NAs removal (Zhang et al. 2016).

26 Photolysis is another oxidation process that is used to treat organic solutes in water. Photolysis of hydroxide under ultraviolet (UV) light at 254 nm can also generate hydroxyl radicals (Afzal et al.

2012a). It has been demonstrated that an UV/H2O2 process was able to degrade CHCA, a model compound for NAs (Afzal et al. 2012a). TiO2 is a widely used photocatalyst that is relatively cheap and abundantly available with high chemical stability (Binas et al. 2017). It is capable of producing highly reactive free radicals (hydroxyl and superoxide radicals) under UV light, enabling the degradation of organic pollutants. Research confirmed that the acute toxicity of NAs in OSPW was completely removed under UV/TiO2 treatment (Mishra et al. 2010a). In addition, TiO2- graphene catalyst has been applied to degrade organic compounds and 90% removal rate of NAs was achieved (Liu et al. 2016). Some studies have been carried out to investigate the different

- - methods of photolysis on the removal of NAs. The comparison of UV/TiO2, UV/IO4 , UV/S2O8 ,

- and UV/H2O2 to degrade NAs in OSPW containing solids illustrated that UV/S2O8 (20 mM) at pH 8 and 10, and UV/H2O2 (50 mM) at pH 8 with quartz immersion removed 5 mg/L NAs from

OSPW (Liang et al. 2011).

Oxidation techniques to degrade NAs from OSPW mainly involve the reactions between NAs and hydroxyl radicals. Drzewicz et al. (2010) investigated the mechanisms of CHCA degradation under UV/H2O2 (Scheme 2.2), and concluded that some intermediates of CHCA also went through reactions with peroxyl radicals. However, the degradation of CHCA depended on the amount of hydroxyl radicals formed. Additionally, the hexane ring was degraded to acyclic byproduct.

27 COOH COOH COOH COOH

-H + O2 + OH + H2O

COOH COOH HO O2 OH COOH COOH

+ O2

+ OH + O2 + H2O

O2 COOH COOH HO O 2 O COOH COOH

• + + HO2 • + HO2 COOH COOH

O OH HO O2 COOH COOH 2 O 2 + H2O2

+ H2O + OH + H2O2 O2 COOH O

O + O2 OH 2 COOH

COOH COOH

+ OH O •

COOH COOH

OH COOH

O OH H

COOH

Scheme 2.2 CHCA degradation by UV/H2O2 (Drzewicz et al. 2010).

28 Photolysis of NAs under natural and artificial light is not effective for NAs removal since NAs are not able to absorb light in that wavelength region (Afzal et al. 2012b). However, with the addition of TiO2, it is noted that OSPW was no longer toxic to Vibiro fischeri under natural sunlight treatment for 14 h (Leshuk et al. 2016). An empirical kinetic model was proposed to better describe the photocatalysis of NAs under solar light with the addition of TiO2 (Leshuk et al. 2016). In addition, under microwave treatment, commercial NAs and NAs in OSPW could be degraded, and the toxicity of commercial NAs was changed from high to moderate with the presence of TiO2

(Mishra et al. 2010b).

Ultrasonic treatment is an emerging technique to treat organic contaminants. Due to the recalcitrance and complex composition of NAs, it is difficult to use ultrasonic treatment alone to totally degrade NAs. When a dicyclohexyl acetic acid (a high MW NA) solution underwent ultrasonic treatment with ozone or hydrogen peroxide, the chemical oxygen demand (COD) of the solution decreased by 84%±2.2% and 81%±2.1%, respectively (Kumar et al. 2014). Optimizing the reaction conditions for ultrasonic treatment and ozone resulted in the complete removal of dicyclohexyl acetic acid (Kumar et al. 2014). Furthermore, combining ultrasonic treatment with ozone successfully removed 89.3±1.1% of NAs from OSPW (Kumar et al. 2014).

The mechanisms of ultrasonic degradation of organic compounds ultimately rely on the properties of organic compounds themselves. For hydrophobic compounds with high vapor pressure, they can enter cavitation microbubbles created through the ultrasonic stimulation and the associated high temperature and pressure at the bubbles break their chemical bonds, therefore, they are mainly degraded by thermal decomposition in the gas phase (Drijvers et al. 1999). For low vapor pressure

29 hydrophobic compounds, it is difficult for them to enter the microbubbles. Pyrolysis in the interfacial region is the predominant reaction when they are at high concentrations (eg, in the case of surface active compounds). If they are at low concentration, their main degradation pathway is by reacting with OH radicals in the bulk solution or in the interfacial region. For nonvolatile compounds, they are mainly degraded by reacting with OH radicals in the bulk solution

(Kotronarou et al. 1991).

Singla et al. (2013) studied the mechanisms of the sonochemical degradation of BA in aqueous solution. It was concluded that the sonolytic degradation of BA was rather efficient and it could undergo both pyrolysis and reaction with OH radicals at pH

30 COOH

COOH

COOH OH OH

OH OH

O OH

COOH CHOOH + + CH3COOH CO2 + H2O COOH CHOOH

Scheme 2.3 Sonolytic degradation of BA (Singla et al. 2013).

Although oxidation is deemed as an effective process for the treatment of NAs, it also has several drawbacks. The presence of inorganic anions, such as chloride and carbonate, can scavenge hydroxyl radicals, and the oxidation of NAs is negatively influenced (Afzal et al. 2012a). In addition, the degradation of NAs is adversely affected by suspended particles in the solution since

31 they lessen the transmission of UV light (Liang et al. 2011). These factors lead to the partial or incomplete degradation of NAs. Moreover, oxidation processes are considered to be relatively expensive treatment techniques (Kannel and Gan 2012).

2.5.2 Membrane Filtration

Membrane filtration is effective in water purification, removal of organic and inorganic compounds without the addition of chemicals, and recently it has been used to treat OSPW. A membrane is a thin film that acts as a barrier to separate specific components from one or more phases. Electrostatic interaction or size exclusion is the mostly used membrane filtration separation mechanism (Quinlan and Tam 2015). Separation efficiency is related to the trans-membrane pressure and cross-flow velocity (Alpatova et al. 2014). Based on the pore size of membrane, membrane filtration is categorized into microfiltration, ultrafiltration, nanofiltration, and reverse osmosis (Peng et al. 2004).

For removal of NAs from OSPW, the ideal pore size of the membrane should only enable water to pass through, whereas other molecules, ions, and solids should be rejected. As the pore size of microfiltration is typically greater than 0.1 μm, it is inefficient for NAs removal and is more suited as a pretreatment process (Munirasu et al. 2016). With regards to ultrafiltration membrane treatment of OSPW, up to 38.6±2.7% of COD was removed. However, since the molecular sizes of NAs are smaller than ultrafiltration membrane pores, it is not effective for removal of NAs

(Alpatova et al. 2014). On the other hand, it was demonstrated that through membrane nanofiltration, the concentration of NAs in tailing ponds could be reduced by up to 95% (Peng et

32 al. 2004). A pilot-scale study was carried out, confirming that reverse osmosis was capable of completely removing all solids, and reducing the amount of ions to negligible levels after pretreatment (Loganathan et al. 2015).

Membrane fouling, which is caused by colloids, organic matter, and bitumen residues adhering to the surface and pores of the membrane, lowers the membrane permeate flux and rejection capability, and shortens membrane life (Kim et al, 2012). Therefore, membrane replacement is required which raises both capital and operating cost. Due to the presence of solids, particles, and unrecovered bitumen in OSPW, separation of NAs through membrane filtration leads to membrane fouling. To prevent membrane fouling, removal of solids, particles, and unrecovered bitumen from

OSPW is necessary and important before membrane treatment. Kim et al. (2012) reported that coagulation-flocculation pretreatment of OSPW could reduce membrane fouling, maintaining the membrane permeate flux.

Micellar-enhanced ultrafiltration (MEUF) is a separation technique based on the enlargement of

MW of organic pollutants by using surfactants. Surfactants form micelles with organic pollutants when their concentration is above critical micelle concentration (CMC). Due to the larger size of micelles, they were rejected by the membrane and consequently, organic pollutants were separated in this treatment process (Schwarze et al. 2017). Interestingly, it was noticed that the presence of

NAs in OSPW reduced the CMC of cetylpyridinium chloride (a widely used surfactant), which in turn promoted the formation of micelles and increased micelle sizes, leading to the separation of

99.8% of NAs (Deriszadeh et al. 2009).

33 2.5.3 Adsorption

Adsorption is a cost-effective approach to isolate NAs. The substance being adsorbed is the adsorbate, and the adsorbing material is the adsorbent. The adsorption process can be more attractive by regenerating adsorbents and reducing the potential of secondary pollution (Xiao and

Josephine 2017). The most commonly used adsorbents for NAs removal are zeolites, aluminosilicates from catalyst manufacturing waste, silica-gels, clays, alumina, mixtures of magnesium and aluminum oxides, and ion-exchange resins (Silva et al. 2013). In addition, biochar is a potential adsorbent for NAs removal owing to its low cost (Bhuiyan et al. 2017). Modified biopolymers from chicken feathers also emerge as promising adsorbents for NAs separation

(Arshad et al. 2016). The efficiency of the adsorption relies largely on the adsorbents (Wang &

Peng 2010). Table 2.3 summarizes different adsorbents used for NAs isolation, demonstrating that different adsorbents have different saturation capacities.

Table 2.3 Saturation capacities for different adsorbents. NAs sources Adsorbents Saturation capacities Reference Commercial NAs Activated carbon and 20 mg of TOC/1 mg Azad et al. nickel based alumina of adsorbent (2013) 1,4-cyclohexanedicarboxylic Commercial granular 452.6 mg/g; Martinez- acid; 2-naphthoic acid; activated carbon 357.9 mg/g; Iglesias et diphenylacetic acid 317.7 mg/g al. (2015) OSPW Wheat straw; 0.59 mg/g for biochar Bhuiyan et switchgrass; mountain from wheat straw al. (2017) pine; hemp shives; aspen wood 2-naphthoxyacetic acid Quaternized chitosan 315 mg/g Quinlan et hydrogels al. (2017)

Petroleum coke is a by-product of the bitumen upgrading process. Owing to its low price and abundance, petroleum coke has been reported to be an attractive adsorbent for NAs removal from

34 OSPW. The adsorption kinetics of acid-extractable organics by petroleum coke follow both

Langmuir and Langmuir-Freundlich isotherm models. Zubot et al. (2012) reported that the adsorption capacities could reach 0.46 mg/g when the concentration of acid-extractable organics was 60 mg/L. Although petroleum coke only adsorbed 84% of NAs in OSPW, combining petroleum coke adsorption and ozone treatment offered a potential technique for OSPW treatment

(El-Din et al. 2011). On the other hand, by adding nitrogen-containing groups to the surface of petroleum coke, the surface then exhibits basic properties which promote the removal of NAs. The removal of total organic carbon (TOC) in OSPW reached 99% at pH 3.5 (Niasar et al. 2016).

Some theoretical research has been performed to investigate adsorption mechanisms. Ma and Chen

(2016) conducted a density functional theory study that showed that the adsorption of NAs on four-nitrogen coordinated transition-metal (Mn, Fe, Co, Ni, Cu, and Zn) embedded graphene was realized mainly through the interactions between carbonyl oxygen atoms and active sites of adsorbents. The principal interactions between NAs and carbonaceous materials were hydrophobic interactions, hydrogen bonding, and electrostatic interactions (Bhuiyan et al. 2017).

Additionally, it should be noted that granular activated carbon can also act as a biofilm substrate, and a synergetic effect between adsorption and microbial biofilm-based degradation promotes the removal of NAs. Islam et al. (2015) showed that the contact time between microbial and organic compounds was increased by biofilm growth on granular activated carbon, further contributing to the degradation of NAs. Moreover, it was concluded that the presence of biofilms (in biochar) increased the NA removal rate.

35 2.5.4 Coagulation and Flocculation

Wastewater usually contains negatively charged tiny particles, and repulsive forces exist when particles are similarly charged, which prevent particle agglomerations and give rise to stable suspensions. Coagulation is a complex process to destabilize colloidal materials stimulating particle agglomeration, which enhances subsequent particle removal through segregation.

Flocculation helps agglomerate the small particles to form larger flocs so that they are more easily removed from suspension (Semerjian & Ayoub. 2003). Trivalent aluminum salts such as aluminum sulfate, iron salts such as ferric chloride are the most commonly used coagulants (Wang et al. 2015). Organic polymers such as cationic polydiallyldimethylammonium chloride

(polyDADMAC) could be used as flocculants (Pourrezaei et al. 2011). The formation of large flocs is due to the charge neutralization caused by the electrostatic interactions between coagulants/flocculants and negatively charged particles (Quinlan & Tam 2015), adsorption, and sweep flocculation (Pourrezaei et al. 2011).

Despite that coagulation/flocculation is mainly suitable for separating suspended solids from

OSPW, recently, researchers have applied it for the removal of NAs. Pourrezaei et al. (2011) demonstrated through an experiment employing cationic polyDADMAC as flocculants, that the concentration of NAs and oxidized NAs (CnH2n+ZO3) from OSPW provided by Syncrude Canada

Ltd. decreased by 37% and 86%, respectively. The reason that accounts for the failure to remove

NAs completely through coagulation/flocculation is that NAs generally exist in the form of salts and they are soluble in alkaline OSPW. This problem can be addressed by adding iron oxide and water soluble acidic-group-containing polymer so that metal ions can form ionic bonds with the

36 carboxylic groups of NAs and with the acidic-group-containing polymer, generating the formation of flocs (Sasaki et al. 2012).

Coagulation/flocculation is a promising pretreatment method to remove solid particles from

OSPW, and a potential technique to isolate NAs owing to the possible electrostatic interactions between coagulants/flocculants and NAs. It is not influenced by the toxicity of OSPW, and requires low energy consumption since it does not need high operating temperatures and pressures.

However, coagulants/flocculants chemicals are expensive, and there are serious environmental consequences by utilizing coagulants/flocculants chemicals. The addition of metal coagulants/flocculants results in the increase of metal concentration in water, which may have human health implications. It also produces large volumes of toxic sludge and further treatment of the sludge is required (Renault et al. 2009).

2.5.5 Biodegradation

NAs can serve as carbon sources for bacterial cultures. Herman et al. (1994) and Lai et al. (1996) demonstrated that the toxicity of NAs dropped through microbial degradation. However, natural microbial degradation of NAs is not sufficient enough to remove NAs in tailing ponds, and the degradation rate is slow. The complete biodegradation of NAs generates carbon dioxide and water, whereas incomplete biodegradation induces the formation of new intermediates (Quinlan and Tam

2015). In addition, NAs are toxic to a wide range of microbial species, hampering the biodegradation process. To remove NAs through biodegradation, a lot of research has been carried out to investigate the influence of different microbial species on the biodegradation of NAs

37 (Clothier and Gieg 2016; Headley et al. 2008; Johnson et al. 2013; Mahdavi et al. 2015; McKenzie et al. 2014; Quesnel et al. 2011; Toor et al. 2013; Yue et al. 2015).

Table 2.4 lists studies where microbial communities were used to remove surrogate NAs and commercial NAs. The results show that microbial-based methods are less effective for NAs biodegradation in OSPW. The differences are attributed to the fact that NAs in OSPW consist of more branched alkyl chains (less bioavailable for microbial biodegradation) than commercial NAs

(Whitby 2010). Therefore, commercial NAs and surrogate NAs are not appropriate to predict biodegradation of NAs in OSPW. The biodegradation rates for NAs with different structures are also different. Quagraine et al. (2005b) showed that chain length, branching, the number of carbon atoms in the alkyl chain that contain carboxylic acids, the position of alkyl groups to the cyclic ring, and the number of cyclic rings influenced the biodegradation of NAs. Rogers et al. (2002) demonstrated that biodegradation was more efficient to remove NAs with carbon numbers less than 21. It should be noted that NAs in tailing ponds can be adsorbed to clays through hydrogen bonding, electrostatic-dipole, and van der Waals (vdW) interaction. Hence, the biodegradation of

NAs is also influenced by the adsorption of NAs to clays (Quagraine et al. 2015a).

38 Table 2.4 Microbial degradation of NAs. NAs Microbe Degradation Rate Incuba- References tion time CHCA;CHAA; Dunaliella CHCA, CHAA, 35 d Quesnel et cyclohexanepropionic tertiolecta cyclohexanepropionic 42 d al. (2011) acid; acid, and CHBA were 21 d cyclohexanebutyric completely transformed. acid (CHBA); No degradation for 1, 2, 35 d 1, 2, 3, 4-tetrahydro-2- 3, 4-tetrahydro-2- naphthoic acid naphthoic acid. 42 d OSPW Simulated 64%-74% NAs removal 52 weeks Toor et al. wetland (2013) OSPW Sediment 38±7% NAs removal 16 McKenzie months et al. (2014) NAs in tailing ponds Algae-bacterial 24.06% NAs removal 120 d Mahdavi et al. (2015) CHCA; CHAA; Tailings The degradation rate is -- Clothier & cyclohexanepropionic microorganisms influenced by anoxic Gieg (2016) acid; CHBA; conditions. Individual cyclohexanepentanoic NAs is susceptible to acid biodegradation by tailings microorganisms. (4´-n-butylphenyl)-4- Pseudomonas For n-BPBA, the removal 49 d Johnson et butanoic acid (n- putida efficiency was 70, 57, al. (2013) BPBA); KT2440 and 14% for 2, 3, and 4 (4´-t-butylphenyl)-4- mg/L of concentration. butanoic acid (t- For t-BPBA, it cannot be BPBA) degraded. 4- Blue-green 100% removal of 4- 14 d Headley et methylcyclohexaneace algae; green methylcyclohexaneacetic al. (2008) tic acid; algae; diatoms acid Oil sand NAs mixture No removal for oil sand NAs mixture NAs surrogates Ochrobactrum; Ochrobactrum and 21 d Yue et al. Bacillus; Bacillus can degrade (2015) Brevundimonas wide range of NAs surrogates; Ochrobactrum and Brevundomonas have higher degradation rate for polycyclic aromatic compounds.

39 Three reaction pathways have been proposed between NAs and microorganisms: β-oxidation, combined α- and β- oxidation, and aromatic oxidation (Quagraine et al. 2005b). The majority of microorganisms such as Acinetobacter anitratum, Alcaligenes faecalis, and Pseudomonas putida mainly degrade NAs through β-oxidation pathways (Johnson et al. 2011). For NAs with an odd number of carbons in the side chain, they are biodegraded through β-oxidation (Quagraine et al.

2005b). Therefore, it is more difficult for NAs with an even number of carbons in side chain to be biodegraded (Quagraine et al. 2005b). CHAA, which has two carbons in side chain, was biodegraded by Alcaligenes species through both α and β-oxidation pathways (Whitby 2010).

Owing to the hindrance of the carboxyl group and the cyclohexane ring, CHAA firstly metabolized through α-oxidation to generate cyclohexylcarboxaldehyde. Then cyclohexylcarboxaldehyde was oxidized through β-oxidation pathway (Scheme 2.4). With regards to CHCA, it followed aromatic oxidation reaction pathways, and formed p-hydroxybenzoic acid before cleavage (Blakley 1974).

40 O

CH2COOH CHOHCOOH CCOOH CHO

-oxidation

COOH COOH COOH COOH

O OH -oxidation

COOH COOH

-oxidation -oxidation COOH COOH

Scheme 2.4 α- and β- oxidation of CHAA (Whitby 2010).

Researchers have attempted to adopt microbial communities that are indigenous to ores and tailings, and outside microbial to degrade NAs (Clothier and Gieg 2016; Mahdavi et al. 2015).

However, due to the toxicity of NAs and formation of the persistent cage-like tricyclic diamondoid

NAs, microbial degradation of NAs is not viable and is shown to be an extremely slow process

(Xu et al. 2017). The concentration of NAs decreased by 16% per year over the first five years and beyond that time, biodegradation became less significant (Allen 2008). Through high performance liquid chromatography (HPLC)/high resolution mass spectrometry (HRMS) analysis, Han et al.

(2009) estimated that in situ biodegradation half-lives of NAs was 12.8 to 13.6 years. The incomplete degradation of NAs induced by the lack of specific enzymes or bacteria is also another disadvantage of microbial methods.

41 Gamma irradiation treatment of OSPW leads to hydrocarbons fragmentation and makes them more biodegradable. VanMensel et al. (2017) found that gramma irradiation treatment of fresh and aged tailings accelerated biodegradation both under aerobic and anaerobic conditions. Huang et al.

(2017) proposed the application of non-toxic biosurfactants to enhance the treatment of OSPW through biodegradation. In addition, ozone pretreatment and nitrifying aerobic promoted NA biodegradation achieved 76.5% removal rate for NAs and 23.6% for oxidized NAs (Xue et al.

2016).

2.6 NAs Removal from Crude Oil

By mixing crude oil that has high TAN with one that has a low TAN, the result is an intermediate

TAN of the mixture. Despite that the TAN of the high TAN oil is decreased in the process, NAs still remain in the crude oil. To deal with the problems induced by NAs in the petroleum refining process, and lower the NAs content in OSPW, it is necessary to investigate efficient and cost- effective methods to remove NAs from crude oil. There are two main approaches to separate NAs from crude oil. One is to destroy the carboxylic group through chemical reactions such as esterification, hydrogenation, and thermal decomposition. The other is to separate NAs for other uses such as alkali washing, alcohol and ammonia processing, solvent extraction, and adsorption

(Wang et al. 2014b). A detailed review of treatment methods is described in the following section.

2.6.1 Decarboxylation

42 Decarboxylation can be used to successfully remove NAs from high acidity crude oils. This process depends on the adsorption of NAs on solid catalysts, and the formation of CO2 and conversion of NAs. Generally, decarboxylation reactions require a temperature above 250C, which in turn leads to serious corrosion problems (Shah et al. 2014). Therefore, it is imperative to develop effective catalysts so that the reaction rate is increased at milder reaction conditions.

Zhang et al. (2006) employed MgO as the catalyst for the decarboxylation reaction and observed that it was effective at removing model NAs compounds and NAs from crude oil. The decarboxylation of NAs using alkaline earth metal oxides (MgO, CaO, BaO, and SrO) demonstrated that MgO and CaO had better activities than BaO and SrO for NAs removal (Oh et al. 2011). Metal oxides not only acted as decarboxylation reaction catalysts, but also neutralized acids and promoted cracking reactions. When MgO and ZnO was employed as catalysts, the main reaction pathway was catalytic decarboxylation. For BaO, the main reaction pathway was neutralization. Whereas for CaO, catalytic decarboxylation, neutralization, and thermal cracking reaction all contributed to the NAs removal from crude oil (Ding et al. 2009). Dias et al. (2015) attempted to use steel slag catalyst (CaCO3, SiO2, and MgO) to remove NAs and noticed that MgO was mainly responsible for the catalytic decarboxylation reaction, and the main reaction mechanisms are explained by Scheme 2.5.

43 MgO 2RCOOH Mg(RCOO)2 + H2O MgO RCOOH RH + CO2

MgO + CO2 MgCO3

Scheme 2.5 NA decarboxylation reaction mechanisms using MgO as catalyst (Dias et al. 2015).

Wang et al. (2014b) demonstrated that the use of Mg-Al hydrotalcite/γ-Al2O3 catalyst reduced the

TAN from 2.7 mg KOH/g to 0.5 mg KOH/g after reacting at 330C and 1.013×105 Pa for 60 min with a 0.17-0.20 catalyst/oil ratio. Different types of compounds in NAs have various removal efficiencies under the same reaction condition using same catalysts. Riahi et al. (2010), in a density functional theory study, concluded that the adsorption energy of aromatic NAs to MgO was higher than that of nonaromatic NAs, and NAs with longer alkaline chain had higher adsorption energy.

Considering the existence of sulfur-containing compounds in crude oil, transition metal oxides can be poisoned, and are not suitable for being the catalysts for decarboxylation reaction. Utilizing supercritical water provides a potential technique to remove NAs through decarboxylation without the addition of catalysts. It was demonstrated that supercritical water could act as hydrogen donor, and separate 83% of NAs through decarboxylation reaction at 490C and 45 MPa for 90 min without the addition of catalysts (Mandal et al. 2012). The removal reactions followed first order kinetics and the mechanisms are denoted in Scheme 2.6 (Mandal et al. 2012):

44 O O

R C OH R C OH

H2O H + OH O OH

R C OH + H R C OH

R C OH Hydrocarbons + gas product + H

H + OH H2O

Scheme 2.6 NAs decarboxylation reaction mechanisms with the presence of supercritical water (Mandal et al. 2012).

2.6.2 Hydrogenation

Hydrogenation is an important crude oil treatment technique in petroleum refining. Apart from saturating double bonds, and reducing nitrogen and sulfur-containing compounds, hydrogenation can also eliminate oxygen containing compounds such as NAs. Le Roy (1960) found that molybdenum oxide-silica alumina catalyst was an efficient catalyst in the removal of NAs by hydrogenation. When hydrogenated at 20 bars and 230C using Ni-Mo deposited on alumina,

Grande and Sorlie (2000) showed that the TAN number of crude oil was reduced to less than 0.5 mg KOH/g. In addition, NAs with MW lower than 450 g/mol were removed from crude oil by adding one of the hydrotreating catalysts from Co, Mo, Ni, and W at temperature from 200 to

370C (Trachte and Robbins 1999). Furthermore, hydrogenation reactions were capable of selectively transforming carboxylic acids to alcohols by using titania-supported platinum catalyst

(Manyar et al. 2010). The hydrogenation mechanisms of BA by using chromium/zirconium 45 dioxide catalyst followed Scheme 2.7. The hydrogenation of BA was a reversible reaction, nevertheless, the rate constant of hydrogenation reaction was higher than the reverse reaction

(Yokoyama & Yamagata 2001).

COOH CHO CH2OH CH3

+ H + H2 + H2 2

+ H2O - H2

- CO 2

Scheme 2.7 Hydrogenation mechanisms of BA (Yokoyama & Yamagata 2001).

Despite the critical role hydrogenation plays in the oil refining process, for treatment of NAs, it remains controversial. Hydrogenation requires great investments with respect to facilities and high consumption of hydrogen. The high operating temperature, pressure, and expensive cost of catalysts are other disadvantages that inhibit the commercial application of this treatment.

Meanwhile, NAs structures are destroyed in the process, and cannot be used for other applications.

2.6.3 Thermal Decomposition

Under high temperatures, the carboxylic groups in NAs are decomposed to carbon oxide, water, and corresponding hydrocarbons (Wang et al. 2014b). Except for the thermal decomposition reaction of NAs and a drop of the TAN, thermal degradation also resulted in the upgrading of heavy oil to light hydrocarbons, a rise of the API gravity, and a reduction of the pour point, viscosity, total sulfur, and nitrogen content of the crude oil (Barros et al. 2017). Yang et al. (2013) 46 showed that temperatures between 350 and 400C were efficient for thermal decomposition of

NAs in crude oil from Liaohe oilfield and that NAs were decomposed to smaller acids such as acetic acid, propanoic acid, and butyric acid. Dias et al. (2014) demonstrated that conversion efficiencies of 80% could be achieved when the crude oil was under thermal treatment at 350C for 6 hours. Smith et al. (2008b) reported that thermal treatment of Athabasca bitumen heavy vacuum gas oil under a nitrogen atmosphere between 300 and 400C promoted carboxylic acids decomposition. Carbon dioxide and benzene were the main products for the thermal decomposition of BA with a smaller proportion of carbon monoxide, hydrogen, and biphenyl (Scheme 2.8)

(Winter and Barton 1970).

C6H5COOH C6H5 + COOH

COOH + M CO2 + H + M COOH + M CO + OH + M

C6H5COOH + H C6H5CO2 + H2

C6H5COOH + OH C6H5CO2 + H2O

C6H5COOH + C6H5 C6H5CO2 + C6H6

C6H5CO2 C6H5 + CO2

2C6H5 (C6H5)2

C6H5 + H C6H6

C6H5 + OH C6H5OH

C6H5COOH C6H6 + CO2

Scheme 2.8 Thermal decomposition mechanisms of BA (Winter and Barton 1970).

47 However, it should be noted that the smaller acids generated in the thermal decomposition process lead to serious corrosion problems. In addition, despite that thermal decomposition of NAs does not need catalysts, a temperature as high as 400C is required and the process is energy intensive

(Anderson et al. 2013). Moreover, the potential for use of NAs as products is not possible since they are destroyed in this process. Although basic metal oxides, such as MgO or CaO, are reported to promote thermal decarboxylation of NAs, insoluble metal-organic components can be formed

(Lee et al. 2016).

2.6.4 Esterification

Esterification between NAs and alcohols utilizing metal carboxylates and oxides catalysts are a promising approach to remove NAs, especially from high acidity crude oil (Anderson et al. 2013).

Esterification usually requires high reaction temperature and relatively long reaction time. Wang et al. (2017) attempted to use layered double hydroxides with various Ni/Al molar ratios as catalysts for the reaction between NAs and ethylene glycol to remove NAs through intracrystalline catalytic esterification from crude oil. Wang et al. (2014a) investigated NAs removal through catalytic esterification by methanol using SnO-Al2O3 catalyst and the TAN was reduced from 2.8 mg KOH/g to 0.5 mg KOH/g.

To decrease reaction time and pressure of esterification, supercritical methanol was used to isolate

NAs and reduce the TAN of crude oils (Mandal et al. 2013; Mandal and Nagarajan 2016). In methanol esterification, supercritical methanol reacts with the carboxylic groups in NAs, generating methyl esters. Mandal et al. (2013) successfully reduced 99.77% of NAs when reacting

48 with supercritical methanol at 350C and 10 MPa for 60 min and derived a first order kinetic model

(Mandal et al. 2013). Supercritical methanol was reported to be capable of reducing CPCA to cyclopentane, formaldehyde, methyl acetate, and 3-pentanol at 180-220C for up to 30 min with an initial pressure of 3 MPa (Mandal and Nagarajan 2016). Khan et al. (2017) also showed that

96.9% of NAs was removed from highly acids crude oil at 250C and 6.4 MPa for 90 min using supercritical methanol.

Uncatalyzed esterification reactions with supercritical methanol require high temperature and pressure (Khan et al. 2017; Mandal et al. 2013; Mandal and Nagarajan 2016). Recently ILs were employed as catalysts for supercritical methanol esterification of NAs to overcome these problems.

Zafar et al. (2016) demonstrated that the addition of 1-butyl-3-methylimidazolium octylsulfate as catalyst reduced the TAN of crude oil by 56% at 150 oC and 0.20 MPa for 120 min. Zafar et al.

(2017) showed that when butyl-methylimidazoilum hydrogen sulfate was used as catalyst, 90%

TAN reduction was obtained at 150C and 2 MPa for 180 min, and moreover the catalyst remained

95% efficient after being recycled four times. Recently, Lee et al. (2016) used metal-modified H- form ferrierite catalyst to remove BA both through catalytic esterification and thermal decomposition. The main reaction pathways are expressed in Scheme 2.9:

C6H5COOH + CH3OH C6H5COOCH3 + H2O

C H COOH C H + CO 6 5 6 6 2

Scheme 2.9 Main reaction pathways for BA reaction.

49 2.6.5 Neutralization

Anderson et al. (2013) illustrated that a dilute caustic wash using alkali/alkaline earth metals can be utilized to wash kerosene/diesel fractions. The NAs were recovered after the water-soluble calcium or sodium naphthenates were acidified with a mineral acid (Anderson et al. 2013). The disadvantage of dilute caustic water wash is that a huge volume of wastewater is generated.

Moreover, NAs removal are incomplete, and emulsions are formed during this treatment process.

Neutralization extraction of NAs is to use the neutralization reaction between NAs and alkaline to separate NAs. Kumara et al. (2014) showed that 95.16% extraction efficiency was achieved when sodium hydroxide was used to separate NAs, however, the emulsion formed inhibited sodium naphthenates separation. To overcome this problem, a non-aqueous solution extraction of NAs was attempted. Shi et al. (2010) employed a sodium hydroxide solution of ethanol to separate NAs from crude oil, and almost 92% of acids could be removed. Since no water was added to crude oil, no emulsion was formed during the process, which made it relatively easy for NAs separation. The main reaction between sodium hydroxide and NAs is denoted below:

R (CH2)mCOOH + NaOH R (CH2)mCOONa + H2O

In addition, an ammonia solution of ethylene glycol has been demonstrated to be effective in separating NAs, and more than 85% of NAs were removed from heavy fractions of petroleum

(Wang et al. 2006). NAs were easily recovered by heating the ammonia solution to release NH3 and decomposing NAs ammonia salts (Wang et al. 2006).

50 Recently, a catalytic neutralization technique was developed to exploit catalysts to promote the neutralization of NAs. Shohaimi et al. (2013) used Cu/Mg(10:90)/Al2O3 and Ni/Mg(10:90)/Al2O3 catalysts to promote the removal of NAs through ammonia solution in ethylene glycol neutralization method. The presence of Cu/Mg(10:90)/Al2O3 reduced the TAN of the heavy oil by

84.8%, whereas Ni/Mg(10:90)/Al2O3 decreased the TAN by 66.7%. When utilizing ammonia solution in ethylene glycol neutralization method to separate NAs from crude oil, Shohaimi et al.

(2014) showed that Cu/Ca(10:90)/Al2O3 was proven to be the catalyst with the most potential.

Further investigation on the mechanism indicated that the reactants needed to be adsorbed on the catalysts surfaces and form active species before interaction, following the Langmuir-

Hinshelwood mechanism (Shohaimi et al. 2017). Furthermore, Cu/Ce(10:90)/Al2O3 catalyst resulted in 93.3% removal of NAs from highly acidity Korean crude oil when a basic chemical consisting of ammonia solution in polyethylene glycol was added (Shukri et al. 2015).

2.6.6 Adsorption

Adsorption is a cost-effective technique to remove NAs from crude oil and oil fractions. It was demonstrated that commercial clay was effective for removing CPCA, CHAA, and cyclohexane butyric acid from diesel sample (Silva et al. 2013). The adsorption capacities for cyclohexane butyric acid and CPCA were 19.4 g/kg and 25.3 g/kg, respectively (Silva et al. 2013). Li et al.

(2017) also investigated the separation of NAs from dewaxed vacuum gas oil by utilizing commercial clay. He concluded that the equilibrium isotherms followed the Dubinin-

Radushkevich adsorption model, the kinetics followed a pseudo-first order kinetic equation, and adsorption was endothermic. In addition, the adsorption enthalpies and saturated adsorption

51 amounts of NAs for three different clays followed the order of Na-montmorillonite>Na-illite>Na- kaolinite (Zou et al. 1997).

Nonaqueous ion exchange solid phase extraction with ionic phases is an emerging technique to separate NAs, and it has a variety of advantages such as good selectivity, low consumption of solvents and stationary phases. De Conto et al. (2014) demonstrated that more than three times as much NAs are separated from crude oil as in liquid-liquid extraction. It has been shown that this method is effective at separating NAs from both light and heavy oil (Jones et al. 2001). A sugar- based QAE Sephadex A-25 ion exchange resin (Acid-IER) was found efficient for isolating NAs from crude oil. Firstly, the commercial form of Acid-IER (Acid-IER-Cl), which was deactivated by chloride, was activated by bicarbonate ions. Then, NAs were extracted on to Acid-IER. After that, volatile solvents were used to recover NAs. The mechanisms can be described in Scheme

2.10 Scheme 2.10(Mediaas et al. 2003):

- - Acid-IER-Cl + HCO3 Acid-IER-HCO3 + Cl

Acid-IER-HCO3 + HA Acid-IER-A + H2CO3

Acid-IER-A + HCOOH Acid-IER-OOCH + HA

Scheme 2.10 The adsorption of NAs by acid-ion exchange resin (Mediaas et al. 2003).

For the isolation of NAs from crude oil, the separation efficiency is governed by the adsorption of ions from NAs by the solid phase. Solid phase alumina functionalized with 1,4-bis(n- propyl)diazoniabicyclo[2,2,2]octane chloride silsesquioxane was proven to be more efficient than commercial solid phase such as SAX and NH2 (De Conto et al. 2012). An ionic silica based hybrid

52 material containing the pyridinium group was synthesized and showed satisfactory results for extracting NAs (De Conto et al. 2014).

Mesoporous molecular sieve type MCM-41 modified with Sr was also demonstrated to be useful at separating NAs from aviation kerosene with a maximum adsorption capacity of 2.0 g/g

(Nascimento et al. 2017). Moreover, Zhu et al. (2017) developed a micro-solid phase extraction method employing amino-functionalized silica as sorbent to separate petroleum acids in crude oil.

However, isolation of NAs from oil through adsorption requires adsorption, desorption, and solvent recovery process (Wang et al. 2014a), rendering it to be slow and complicated. Moreover, it is reported only effective for low temperature distillate fractions (Anderson et al. 2013).

2.6.7 Solvent Extraction

The principle of solvent extraction is to utilize solvents to dissolve NAs so that NAs can be removed from crude oil instead of converting them to salts or other hydrocarbons and gases. The first step of the solvent extraction process is to use solvents to extract NAs from crude oil. After extraction, the solvent phase is vaporized to separate the extracting solvents and isolate NAs.

Solvents such as a methanol-water mixture (V:V=88:12), methanol, anhydrous acid, and water mixture (V:V:V=73.5:22.7:3.8), and acetic acid and water mixture (V:V=94.3:5.7) were demonstrated to have high efficiency and selectivity in NAs removal (Honeycutt et al. 1952). A solvent mixture composed of methanol, water and ammonia was found to be effective to extract high purity NAs from petroleum distillates with high TAN, and the volume ratio of solvent to oil is from 0.01 to 1 (Danzik et al. 1987). Solvent extraction of NAs consumes large volumes of

53 solvent, which is expensive and harmful to environment. Therefore, it is not a feasible approach for NAs separation from crude oil. Nowadays, researchers have focused on developing more environmentally friendly methods to isolate NAs.

2.6.8 Ionic Liquid Extraction

Owing to their attractive properties, and abilities to form different interaction types, Ionic liquids

(ILs) have many applications in petroleum refining processes. Apart from being utilized in extracting sulfur- and nitrogen-containing compounds in crude oil, IL extraction has been emerged as an alternative technique for NAs isolation since they have excellent solubility for organic compounds. IL extraction of NAs from crude oil involves the formation of IL-naphthenate complexes, and they are capable of isolating NAs below the boiling point of crude oil. Anderson et al. (2015) demonstrated that NAs in Doba crude oil could be separated efficiently and the TAN was reduced to 0.30 mg KOH/g by using carbonate-based IL extraction. Imidazolium chloride- based ILs removed 48% of TAN at 40C for 60 min with a 3 IL/NA ratio (Mandal et al. 2015).

Moreover, the in situ formation of IL, which was realized through the reaction between NAs and imidazole derivatives using ethanol as solvent, reduced the volume of NAs in crude oil by 67.0%

(Shi et al. 2008). It should be noted that the interactions of amino acid-based ILs and NAs mainly proceeded in two reaction pathways according to Scheme 2.11. The first reaction represents the formation of IL-NAs complexes, whereas the second reaction is the protonation of the amino acid anion (Anderson et al. 2013).

54 Cation Cation H2N-C(R)-CO2 + R'-CO2H H2N-C(R)-CO2 O2C-R'

Cation O C-R' + NH -C(R)-CO H Cation H2N-C(R)-CO2 + R'-CO2H 2 2 2

Scheme 2.11 Chemical interactions between NAs and amino acid-based ILs (Anderson et al. 2013).

Sun & Shi (2012) deduced that ILs with good alkaline properties such as 1-alkyl-3- methylimidazolium imidazolide-based ILs could remove NAs from crude oil. The hydroxide anions of hydroxide-based ILs reacted with NAs, therefore, naphthenate salts were formed and

NAs were removed (Shah et al. 2014). Phenolate-based ILs are reported to be efficient in isolating

NAs and they could be reused for at least three times (Shah et al. 2016a). Duan et al. also demonstrated that the stronger alkalinity of ILs was, the easier it was to de-acidify NAs. The NAs removal efficiency reached 100% when 1-octyl-3-methylimidazolium imidazolide was employed and the IL/oil ratio was 0.008 (Duan et al. 2013). Although ILs with stronger alkaline properties promote the extraction of NAs, it has been verified that the interaction sites between ILs and NAs also determine the removal efficiency. The diazabicyclo undecane-based cation and thiocyanate anion effectively isolated NAs from crude oil, and the possible explanation was the two coordination sites of the thiocyanate anion (Shah et al. 2016b). Judging from liquid-liquid equilibrium of dodecane, NAs, and n-alkyl imidazolium-based ILs, n-alkyl imidazolium-based ILs with longer chain had higher extraction capacity for NAs because of higher solubility for NAs

(Shah et al. 2016c).

To efficiently remove NAs from crude oil and separate NAs from ILs, researchers began to combine ILs with solid support materials. ILs were grafted onto solids combining the properties

55 of ILs and the solids (Shah et al. 2015). Shah et al. (2017) showed that around 92% and 94% of

NAs were isolated by using solid support phases with a 0.06 ILs/oil ratio and longer alkyl chain.

Imidazolium-based ILs are among the most studied ILs because of their solubility for a broad range of inorganic and organic compounds (Liwarska-Bizukojc & Gendaszewska 2013). However, ILs are toxic to microorganisms, which negatively influence their biodegradation after being released to the environment (Gathergood et al. 2004). With their low biodegradability and high solubility and stability, imidazolium-based ILs are persistent pollutants, causing serious environmental problems (Romero et al. 2008). However, existing research about ILs extraction of NAs mainly focuses on the extraction efficiency of NAs with little of them paying attention to the biodegradation of ILs after usage.

Except for extracting NAs from crude oil, ILs have other potential applications in petroleum industry, such as desulfurization, denitrogenation, dearomatization, dehydration, and desalting

(José-Alberto et al. 2011). To select suitable ILs for different applications, it is desirable to understand their physicochemical properties. ILs are designable solvents with physicochemical properties being alterable to some extent by varying the cation, anion, or substituent groups

(Herrera et al. 2016). Since a variety of cations and anions can be used to synthesize ILs, the relationships between physicochemical properties and intramolecular and intermolecular interactions should be understood so that they can be tuned to desirable physicochemical properties

(Marium et al. 2017).

56 2.7 Knowledge Gaps

Ultrasonic treatment of NAs in OSPW can provoke reactions both inside the microbubbles and liquid phases. The generation of hydroxyl radicals leads to the degradation of organic compounds.

If the rate constants between model NAs and hydroxyl radicals in the gas and aqueous phases can be compared, it can provide insights on how to increase the degradation rates of NAs. However, little research has been conducted to compare the reaction rates of NAs degradation in the gas and aqueous phases as well as the degradation mechanisms.

In the petroleum industry, imidazolium-based ILs have been widely used for the extraction of NAs from crude oil. A lot of experimental studies have been performed to explore the suitable imidazolium-based ILs for NAs isolation. Nevertheless, the interaction mechanisms between NAs and imidazolium-based ILs remains unclear. Since ILs are capable of forming different types of interactions with organic compounds, it is necessary and critical to conduct a comprehensive study to explore the interaction mechanisms in order to provide guidance for the design of ILs to efficiently remove NAs.

In addition, the biodegradation of imidazolium-based ILs is still a challenge. To solve the environmental problems caused by ILs, it is imperative to design ILs with high biodegradability and high extraction efficiency for NAs. Few experimental and theoretical studies have been performed to design biodegradable ILs for NAs removal, compare the mechanisms between imidazolium-based ILs and biodegradable imidazolium-based ILs, and investigate the extraction efficiency differences.

Moreover, the properties of ILs are influenced by the cations and anions. To tune ILs with specific characteristics, it is fundamental to gain insights into the relationships between physicochemical properties and intermolecular and intramolecular interactions. Few theoretical studies have been performed to compare the physicochemical differences between imidazolium-based ILs and pyridinium-based ILs at the molecular level.

2.8 References

Afzal, A., Drzewicz, P., Martin, J. W., & El-Din, M. G. (2012a). Decomposition of cyclohexanoic

acid by the UV/H2O2 process under various conditions. Science of the Total Environment, 426, 387-392. Afzal, A., Drzewicz, P., Pérez-Estrada, L. A., Chen, Y., Martin, J. W., & El-Din, M. G. (2012b).

Effect of molecular structure on the relative reactivity of naphthenic acids in the UV/H2O2 advanced oxidation process. Environmental Science & Technology, 46(19), 10727-10734. Allen, E. W. (2008). Process water treatment in Canada's oil sands industry: I. Target pollutants and treatment objectives. Journal of Environmental Engineering and Science, 7(2), 123-138. Alpatova, A., Kim, E. S., Dong, S., Sun, N., Chelme-Ayala, P., & El-Din, M. G. (2014). Treatment of oil sands process-affected water with ceramic ultrafiltration membrane: Effects of operating conditions on membrane performance. Separation and Purification Technology, 122, 170-182. Alvisi, P. P., & Lins, V. F. (2011). An overview of naphthenic acid corrosion in a vacuum distillation plant. Engineering Failure Analysis, 18(5), 1403-1406. Anderson, J. C., Wiseman, S. B., Wang, N., Moustafa, A., Perez-Estrada, L., El-Din, M. G., Martin, J.W., Liber, K., & Giesy, J. P. (2011). Effectiveness of ozonation treatment in eliminating toxicity of oil sands process-affected water to Chironomus dilutus. Environmental Science & Technology, 46(1), 486-493. Anderson, K., Goodrich, P., Hardacre, C., Hussain, A., Rooney, D. W., & Wassell, D. (2013). Removal of naphthenic acids from crude oil using amino acid ionic liquids. Fuel, 108, 715-722.

58 Anderson, K., Atkins, M. P., Goodrich, P., Hardacre, C., Hussain, A. S., Pilus, R., & Rooney, D. W. (2015). Naphthenic acid extraction and speciation from Doba crude oil using carbonate- based ionic liquids. Fuel, 146, 60-68. Arshad, M., Khosa, M. A., Siddique, T., & Ullah, A. (2016). Modified biopolymers as sorbents for the removal of naphthenic acids from oil sands process affected water (OSPW). Chemosphere, 163, 334-341. Azad, F. S., Abedi, J., & Iranmanesh, S. (2013). Removal of naphthenic acids using adsorption process and the effect of the addition of salt. Journal of Environmental Science and Health, Part A, 48(13), 1649-1654. Barros, E. V., Dias, H. P., Gomes, A. O., Rodrigues, R. R., Moura, R. R., Sad, C. M., Freitas, J.C., Neto, A.C., Aquije, G.M., & Romão, W. (2017). Study of degradation of acid crude oil by high resolution analytical techniques. Journal of Petroleum Science and Engineering, 154, 194-203. Barrow, M. P., Headley, J. V., Peru, K. M., & Derrick, P. J. (2009). Data visualization for the characterization of naphthenic acids within petroleum samples. Energy & Fuels, 23(5), 2592- 2599. Beck, E. M., Smits, J. E., & St Clair, C. C. (2015). Evidence of low toxicity of oil sands process- affected water to birds invites re-evaluation of avian protection strategies. Conservation Physiology, 3(1), cov038. Bhuiyan, T. I., Tak, J. K., Sessarego, S., Harfield, D., & Hill, J. M. (2017). Adsorption of acid- extractable organics from oil sands process-affected water onto biomass-based biochar: Metal content matters. Chemosphere, 168, 1337-1344.

Binas, V., Venieri, D., Kotzias, D., & Kiriakidis, G. (2017). Modified TiO2 based photocatalysts for improved air and health quality. Journal of Materiomics, 3(1), 3-16. Blakley, E. R. (1974). The microbial degradation of cyclohexanecarboxylic acid: a pathway involving aromatization to form p-hydroxybenzoic acid. Canadian Journal of Microbiology, 20(10), 1297-1306. Bouchard, G., Carrupt, P. A., Testa, B., Gobry, V., & Girault, H. H. (2002). Lipophilicity and solvation of anionic drugs. Chemistry-A European Journal, 8(15), 3478-3484. Brient, J. A., Wessner, P. J., & Doyle, M. N. (1995). Naphthenic acids. In Kirk-Othmer Encyclopedia of Chemical Technology. New York: John Wiley & Sons.

59 Clemente, J. S., Yen, T. W., & Fedorak, P. M. (2003a). Development of a high performance liquid chromatography method to monitor the biodegradation of naphthenic acids. Journal of Environmental Engineering and Science, 2(3), 177-186. Clemente, J. S., Prasad, N. G. N., MacKinnon, M. D., & Fedorak, P. M. (2003b). A statistical comparison of naphthenic acids characterized by gas chromatography–mass spectrometry. Chemosphere, 50(10), 1265-1274. Clemente, J. S., & Fedorak, P. M. (2005). A review of the occurrence, analyses, toxicity, and biodegradation of naphthenic acids. Chemosphere, 60(5), 585-600. Clothier, L. N., & Gieg, L. M. (2016). Anaerobic biodegradation of surrogate naphthenic acids. Water Research, 90, 156-166. Danzik, M. (1987). U.S. Patent No. 4,634,519. Washington, DC: U.S. Patent and Trademark Office. De Conto, J. F., Nascimento, J. D. S., de Souza, D. M. B., Da Costa, L. P., Egues, S. M. D. S., Freitas, L. D. S., & Benvenutti, E. V. (2012). Solid phase extraction of petroleum carboxylic acids using a functionalized alumina as stationary phase. Journal of Separation Science, 35(8), 1044-1049. De Conto, J. F., Santos, M. R., Carvalho, A. S., Campos, K. V., Freitas, L. S., Benvenutti, E. V., de Menezes, E.W., Santana, C.C., & Egues, S. M. (2014). Naphthenic acids recovery from petroleum using ionic silica based hybrid material as stationary phase in solid phase extraction (SPE) process. Adsorption, 20(8), 917-923. Del Rio, L. F., Hadwin, A. K. M., Pinto, L. J., MacKinnon, M. D., & Moore, M. M. (2006). Degradation of naphthenic acids by sediment micro‐organisms. Journal of Applied Microbiology, 101(5), 1049-1061. Deng, Y., & Zhao, R. (2015). Advanced oxidation processes (AOPs) in wastewater treatment. Current Pollution Reports, 1(3), 167-176. Deriszadeh, A., Harding, T. G., & Husein, M. M. (2009). Improved MEUF removal of naphthenic acids from produced water. Journal of Membrane Science, 326(1), 161-167. Dias, H. P., Pereira, T. M., Vanini, G., Dixini, P. V., Celante, V. G., Castro, E. V., Vaz, B.G., Fleming, F.P., Gomes, A.O., Aquije, G.M., & Romão, W. (2014). Monitoring the degradation and the corrosion of naphthenic acids by electrospray ionization Fourier transform ion cyclotron resonance mass spectrometry and atomic force microscopy. Fuel, 126, 85-95

60 Dias, H. P., Gonçalves, G. R., Freitas, J. C., Gomes, A. O., de Castro, E. V., Vaz, B. G., Aquije, G.M., & Romão, W. (2015). Catalytic decarboxylation of naphthenic acids in crude oils. Fuel, 158, 113-121. Ding, L., Rahimi, P., Hawkins, R., Bhatt, S., & Shi, Y. (2009). Naphthenic acid removal from heavy oils on alkaline earth-metal oxides and ZnO catalysts. Applied Catalysis A: General, 371(1), 121-130. Drijvers, D., Van Langenhove, H., & Beckers, M. (1999). Decomposition of phenol and

trichloroethylene by the ultrasound/H2O2/CuO process. Water Research, 33(5), 1187-1194. Drzewicz, P., Afzal, A., El-Din, M. G., & Martin, J. W. (2010). Degradation of a model naphthenic

acid, cyclohexanoic acid, by vacuum UV (172 nm) and UV (254 nm)/H2O2. The Journal of Physical Chemistry A, 114(45), 12067-12074. Drzewicz, P., Perez-Estrada, L., Alpatova, A., Martin, J. W., & El-Din, M. G.(2012). Impact of peroxydisulfate in the presence of zero valent iron on the oxidation of cyclohexanoic acid and naphthenic acids from oil sands process-affected water. Environmental Science & Technology, 46(16), 8984-8991. Duan, J., Sun, Y., & Shi, L. (2013). Three different types of heterocycle of nitrogen-containing alkaline ionic liquids treatment of acid oil to remove naphthenic acids. Catalysis Today, 212, 180-185. Eftink, M. R., Andy, M. L., Bystrom, K., Perlmutter, H. D., & Kristol, D. S. (1989). Cyclodextrin inclusion complexes: studies of the variation in the size of alicyclic guests. Journal of the American Chemical Society, 111(17), 6765-6772. El-Din, M. G., Fu, H., Wang, N., Chelme-Ayala, P., Pérez-Estrada, L., Drzewicz, P., Martin, J.W., Zubot, W., & Smith, D. W. (2011). Naphthenic acids speciation and removal during petroleum- coke adsorption and ozonation of oil sands process-affected water. Science of the Total Environment, 409(23), 5119-5125. Frank, R. A., Fischer, K., Kavanagh, R., Burnison, B. K., Arsenault, G., Headley, J. V., Peru, K.M., Kraak, G.V.D., & Solomon, K. R. (2008). Effect of carboxylic acid content on the acute toxicity of oil sands naphthenic acids. Environmental Science & Technology, 43(2), 266-271. Gathergood, N., Garcia, M. T., & Scammells, P. J. (2004). Biodegradable ionic liquids: Part I. Concept, preliminary targets and evaluation. Green Chemistry, 6(3), 166-175.

61 Grande, K., & Sorlie, C. (2000). U.S. Patent No. 6,063,266. Washington, DC: U.S. Patent and Trademark Office. Grewer, D. M., Young, R. F., Whittal, R. M., & Fedorak, P. M. (2010). Naphthenic acids and other acid-extractables in water samples from Alberta: what is being measured? Science of the Total Environment, 408(23), 5997-6010. Hamilton, W. S., Benton, S., French, J., McCormick, D., Pustejovsky, S., & Thompson, P. (1978). Enthalpies of combustion and formation of 3-methylisoxazole and 5-methylisoxazole. Journal of Chemical and Engineering Data, 23(3), 201-203. Han, X., MacKinnon, M. D., & Martin, J. W. (2009). Estimating the in situ biodegradation of naphthenic acids in oil sands process waters by HPLC/HRMS. Chemosphere, 76(1), 63-70. Headley, J. V., & McMartin, D. W. (2004). A review of the occurrence and fate of naphthenic acids in aquatic environments. Journal of Environmental Science and Health, Part A, 39(8), 1989-2010. Headley, J. V., Peru, K. M., Barrow, M. P., & Derrick, P. J. (2007). Characterization of naphthenic acids from Athabasca oil sands using electrospray ionization: the significant influence of solvents. Analytical Chemistry, 79(16), 6222-6229. Headley, J. V., Du, J. L., Peru, K. M., Gurprasad, N., & McMartin, D. W. (2008). Evaluation of algal phytodegradation of petroleum naphthenic acids. Journal of Environmental Science and Health, Part A, 43(3), 227-232. Herman, D. C., Fedorak, P. M., MacKinnon, M. D., & Costerton, J. W. (1994). Biodegradation of naphthenic acids by microbial populations indigenous to oil sands tailings. Canadian Journal of Microbiology, 40(6), 467-477. Herrera, C., García, G., Atilhan, M., & Aparicio, S. (2016). A molecular dynamics study on aminoacid-based ionic liquids. Journal of Molecular Liquids, 213, 201-212. Hoefnagel, A. J., & Wepster, B. M. (1990). Substituent effects. 13. Anomalous dissociation constants in water‐organic solvent mixtures: Alicyclic and aliphatic carboxylic acids. Recueil des Travaux Chimiques des Pays-Bas, 109(9), 455-462. Honeycutt, E. M. (1952). U.S. Patent No. 2,610,209. Washington, DC: U.S. Patent and Trademark Office.

62 Huang, G., Yu, H., Li, G., An, C., & Wei, J. (2017). Development of an innovative bioremediation technology for oil-sands tailings waste treatment. Report 002-00017-UOR to Petroleum Technology Research Centre, 2017. Hughes, S. A., Huang, R., Mahaffey, A., Chelme-Ayala, P., Klamerth, N., Meshref, M. N., Ibrahim, M.D., Brown, C., Peru, K.M., Headley, J.V., & El-Din, M. G. (2017). Comparison of methods for determination of total oil sands-derived naphthenic acids in water samples. Chemosphere, 187, 376-384. Islam, M. S., Zhang, Y., McPhedran, K. N., Liu, Y., & El-Din, M. G. (2015). Granular activated carbon for simultaneous adsorption and biodegradation of toxic oil sands process-affected water organic compounds. Journal of Environmental Management, 152, 49-57. Janfada, A., Headley, J. V., Peru, K. M., & Barbour, S. L. (2006). A laboratory evaluation of the sorption of oil sands naphthenic acids on organic rich soils. Journal of Environmental Science and Health Part A, 41(6), 985-997. Johnson, R. J., Smith, B. E., Sutton, P. A., McGenity, T. J., Rowland, S. J., & Whitby, C. (2011). Microbial biodegradation of aromatic alkanoic naphthenic acids is affected by the degree of alkyl side chain branching. The ISME Journal, 5(3), 486. Johnson, R. J., Smith, B. E., Rowland, S. J., & Whitby, C. (2013). Biodegradation of alkyl branched aromatic alkanoic naphthenic acids by Pseudomonas putida KT2440. International Biodeterioration & Biodegradation, 81, 3-8. Jones, D. M., Watson, J. S., Meredith, W., Chen, M., & Bennett, B. (2001). Determination of naphthenic acids in crude oils using nonaqueous ion exchange solid-phase extraction. Analytical Chemistry, 73, 703–707. José-Alberto, M. H., & Jorge, A. (2011). Current knowledge and potential applications of ionic liquids in the petroleum industry. In Ionic liquids: applications and perspectives. Rijeka, Croatia: InTech. Kane, R. D., & Cayard, M. S. (2002). A comprehensive study on naphthenic acid corrosion. In NACE Corrosion. Houston, Texas. Paper No. 2555; 2002. Kannel, P. R., & Gan, T. Y. (2012). Naphthenic acids degradation and toxicity mitigation in tailings wastewater systems and aquatic environments: a review. Journal of Environmental Science and Health, Part A, 47(1), 1-21.

63 Kessenich, B. (2015). Surface thermodynamics of decanoic acid (Undergraduate Honors Theses, University of Colorado at Boulder). Khan, M. K., Riaz, A., Yi, M., & Kim, J. (2017). Removal of naphthenic acids from high acid crude via esterification with methanol. Fuel Processing Technology, 165, 123-130. Kim, E. S., Liu, Y., & El-Din, M. G. (2012). Evaluation of membrane fouling for in-line filtration of oil sands process-affected water: the effects of pretreatment conditions. Environmental Science & Technology, 46(5), 2877-2884. Kotronarou, A., Mills, G., & Hoffmann, M. R. (1991). Ultrasonic irradiation of p-nitrophenol in aqueous solution. Journal of Physical Chemistry, 95(9), 3630-3638. Kumar, P., Headley, J., Peru, K., Bailey, J., & Dalai, A. (2014). Removal of dicyclohexyl acetic acid from aqueous solution using ultrasound, ozone and their combination. Journal of Environmental Science and Health, Part A, 49(13), 1512-1519. Kumara, R. B., Shindeb, S. N., & Gaikwadc, S. G. (2014). Reactive extraction of naphthenic acid by using sodium hydroxides as an extractant. International Journal of Advanced Engineering Technology, 5, 103-106. Lai, J. W., Pinto, L. J., Bendell‐Young, L. I., Moore, M. M., & Kiehlmann, E. (1996). Factors that affect the degradation of naphthenic acids in oil sands wastewater by indigenous microbial communities. Environmental Toxicology and Chemistry, 15(9), 1482-1491. Le Roy, W. H. (1960). U.S. Patent No. 2,921,023. Washington, DC: U.S. Patent and Trademark Office. Lee, Y. H., Park, J. Y., Park, S. Y., Kim, C. H., Nam, J. W., Kim, Y. J., & Bae, J. W. (2016). Removal of benzoic acid in heavy oils by esterification using modified ferrierite: roles of Brønsted and Lewis acid sites. Energy & Fuels, 30(7), 5391-5397. Leshuk, T., de Oliveira Livera, D., Peru, K. M., Headley, J. V., Vijayaraghavan, S., Wong, T., & Gu, F. (2016). Photocatalytic degradation kinetics of naphthenic acids in oil sands process- affected water: Multifactorial determination of significant factors. Chemosphere, 165, 10-17. Leshuk, T., Wong, T., Linley, S., Peru, K. M., Headley, J. V., & Gu, F. (2016). Solar photocatalytic degradation of naphthenic acids in oil sands process-affected water. Chemosphere, 144, 1854- 1861.

64 Li, X., Ma, R., Huang, F., & Li, Y. (2017). Adsorption of naphthenic acids from dewaxed vacuum gas oil by activated clay: kinetics, equilibrium and thermodynamic Studies. China Petroleum Processing & Petrochemical Technology, 19(1), 123-134. Liang, X., Zhu, X., & Butler, E. C. (2011). Comparison of four advanced oxidation processes for the removal of naphthenic acids from model oil sands process water. Journal of Hazardous Materials, 190(1), 168-176. Lide, D. R., & Milne, G. W. A. (1993). Handbook of Data on Organic Compounds. Boca Raton: CRC Press. Liu, J., Wang, L., Tang, J., & Ma, J. (2016). Photocatalytic degradation of commercially sourced

naphthenic acids by TiO2-graphene composite nanomaterial. Chemosphere, 149, 328-335. Loganathan, K., Chelme-Ayala, P., & El-Din, M. G. (2015). Pilot-scale study on the reverse osmosis treatment of oil sands tailings pond water: Impact of pretreatment on process performance. Desalination, 360, 52-60. Liwarska-Bizukojc, E., & Gendaszewska, D. (2013). Removal of imidazolium ionic liquids by microbial associations: study of the biodegradability and kinetics. Journal of Bioscience and Bioengineering, 115(1), 71-75. Ma, L., & Chen, X. (2016). Adsorption of naphthenic acids to the nitrogen-coordinated transition- metal embedded graphene: A DFT study. Russian Journal of Physical Chemistry B, 10(6), 1027-1031. Mahdavi, H., Prasad, V., Liu, Y., & Ulrich, A. C. (2015). In situ biodegradation of naphthenic acids in oil sands tailings pond water using indigenous algae–bacteria consortium. Bioresource Technology, 187, 97-105. Mandal, P. C., Sasaki, M., & Goto, M. (2012). Reduction of total acid number (TAN) of naphthenic acid (NA) using supercritical water for reducing corrosion problems of oil refineries. Fuel, 94, 620-623. Mandal, P. C., Sasaki, M., & Goto, M. (2013). Non-catalytic reduction of total acid number (TAN) of naphthenic acids (NAs) using supercritical methanol. Fuel Processing Technology, 106, 641- 644. Mandal, P. C., Abdalla, M. A., & Moniruzzaman, M. Acidity reduction of naphthenic acid using imidazolium chloride based ionic liquids. International Journal of Applied Engineering Research, 10(89), 2015.

65 Mandal, P. C., & Nagarajan, T. (2016). Kinetics and reaction pathways of total acid number reduction of cyclopentane carboxylic acid using subcritical methanol. Polish Journal of Chemical Technology, 18(3), 44-49. Manyar, H. G., Paun, C., Pilus, R., Rooney, D. W., Thompson, J. M., & Hardacre, C. (2010). Highly selective and efficient hydrogenation of carboxylic acids to alcohols using titania supported Pt catalysts. Chemical Communications, 46(34), 6279-6281. Marium, M., Auni, A., Rahman, M. M., Mollah, M. Y. A., & Susan, M. A. B. H. (2017). Molecular level interactions between 1-ethyl-3-methylimidazolium methanesulphonate and water: Study of physicochemical properties with variation of temperature. Journal of Molecular Liquids, 225, 621-630. Martinez-Iglesias, A., Niasar, H. S., Xu, C., & Ray, M. B. (2015). Adsorption of model naphthenic acids in water with granular activated carbon. Adsorption Science & Technology, 33(10), 881- 894. McCrone, W. C. (1953). Crystallographic data. 70. 1-naphthoic acid (α-naphthoic acid). Analytical Chemistry, 25(7), 1126-1127. McKenzie, N., Yue, S., Liu, X., Ramsay, B. A., & Ramsay, J. A. (2014). Biodegradation of naphthenic acids in oils sands process waters in an immobilized soil/sediment bioreactor. Chemosphere, 109, 164-172. Mediaas, H., Grande, K. V., Hustad, B. M., Rasch, A., Rueslåtten, H. G., & Vindstad, J. E. (2003). The acid-IER method-a method for selective isolation of carboxylic acids from crude oils and other organic solvents. In International Symposium on Oilfield Scale. Society of Petroleum Engineers. Melvin, S. D., & Trudeau, V. L. (2012). Growth, development and incidence of deformities in amphibian larvae exposed as embryos to naphthenic acid concentrations detected in the Canadian oil sands region. Environmental Pollution, 167, 178-183. Meredith, W., Kelland, S. J., & Jones, D. M. (2000). Influence of biodegradation on crude oil acidity and carboxylic acid composition. Organic Geochemistry, 31(11), 1059-1073. Mishra, S., Meda, V., Dalai, A. K., McMartin, D. W., Headley, J. V., & Peru, K. M. (2010a). Photocatalysis of naphthenic acids in water. Journal of Water Resource and Protection, 2(07), 644.

66 Mishra, S., Meda, V., Dalai, A. K., Headley, J. V., Peru, K. M., & McMartin, D. W. (2010b). Microwave treatment of naphthenic acids in water. Journal of Environmental Science and Health Part A, 45(10), 1240-1247 Mohammed, M. A., & Sorbie, K. S. (2009). Naphthenic acid extraction and characterization from naphthenate field deposits and crude oils using ESMS and APCI-MS. Colloids and Surfaces A: Physicochemical and Engineering Aspects, 349(1), 1-18. Munirasu, S., Haija, M. A., & Banat, F. (2016). Use of membrane technology for oil field and refinery produced water treatment—a review. Process Safety and Environmental Protection, 100, 183-202. Nair, B. (2001). Final report on the safety assessment of benzyl alcohol, benzoic acid, and sodium benzoate. International Journal of Toxicology, 20, 23-50. Nascimento, G. E. D., Duarte, M. M., Sales, D. C., & Barbosa, C. M. D. M. (2017). Kinetic and equilibrium adsorption studies for removal of naphthenic acids present in model mixture of aviation kerosene. Chemical Engineering Communications, 204(1), 105-110. Niasar, H. S., Li, H., Kasanneni, T. V. R., Ray, M. B., & Xu, C. C. (2016). Surface amination of activated carbon and petroleum coke for the removal of naphthenic acids and treatment of oil sands process-affected water (OSPW). Chemical Engineering Journal, 293, 189-199. Oh, H. Y., Park, J. H., Rhee, Y. W., & Kim, J. N. (2011). Decarboxylation of naphthenic acid using alkaline earth metal oxide. Journal of Industrial and Engineering Chemistry, 17(4), 788- 793. Peng, H., Volchek, K., MacKinnon, M., Wong, W. P., & Brown, C. E. (2004). Application on to nanofiltration to water management options for oil sands operation. Desalination, 170(2), 137- 150. Pourrezaei, P., Drzewicz, P., Wang, Y., Gamal El-Din, M., Perez-Estrada, L. A., Martin, J. W., Anderson, J., Wiseman, S., Liber, K., & Giesy, J. P. (2011). The impact of metallic coagulants on the removal of organic compounds from oil sands process-affected water. Environmental Science & Technology, 45(19), 8452-8459. Quagraine, E. K., Peterson, H. G., & Headley, J. V. (2005a). In situ bioremediation of naphthenic acids contaminated tailing pond waters in the Athabasca oil sands region—demonstrated field studies and plausible options: a review. Journal of Environmental Science and Health, 40(3), 685-722.

67 Quagraine, E. K., Headley, J. V., & Peterson, H. G. (2005b). Is biodegradation of bitumen a source of recalcitrant naphthenic acid mixtures in oil sands tailing pond waters? Journal of Environmental Science and Health, 40(3), 671-684. Quesnel, D. M., Bhaskar, I. M., Gieg, L. M., & Chua, G. (2011). Naphthenic acid biodegradation by the unicellular alga Dunaliella tertiolecta. Chemosphere, 84(4), 504-511. Quinlan, P. J., & Tam, K. C. (2015). Water treatment technologies for the remediation of naphthenic acids in oil sands process-affected water. Chemical Engineering Journal, 279, 696- 714. Quinlan, P. J., Grishkewich, N., & Tam, K. C. (2017). Removal of 2‐naphthoxyacetic acid from aqueous solution using quaternized chitosan beads. The Canadian Journal of Chemical Engineering, 95(1), 21-32. Reinardy, H. C., Scarlett, A. G., Henry, T. B., West, C. E., Hewitt, L. M., Frank, R. A., & Rowland, S. J. (2013). Aromatic naphthenic acids in oil sands process-affected water, resolved by GCxGC-MS, only weakly induce the gene for vitellogenin production in zebrafish (Danio rerio) larvae. Environmental Science & Technology, 47(12), 6614-6620. Renault, F., Sancey, B., Badot, P. M., & Crini, G. (2009). Chitosan for coagulation/flocculation processes–an eco-friendly approach. European Polymer Journal, 45(5), 1337-1348. Riahi, S., Pourhossein, P., & Ganjali, M. R. (2010). Removal of naphthenic acids from liquid petroleum: theoretical study. Petroleum Science and Technology, 28(1), 68-78. Rogers, V. V., Wickstrom, M., Liber, K., & MacKinnon, M. D. (2002). Acute and subchronic mammalian toxicity of naphthenic acids from oil sands tailings. Toxicological Sciences, 66(2), 347-355. Romero, A., Santos, A., Tojo, J., & Rodriguez, A. (2008). Toxicity and biodegradability of imidazolium ionic liquids. Journal of Hazardous Materials, 151(1), 268-273. Ross, M. S., Pereira, A. D. S., Fennell, J., Davies, M., Johnson, J., Sliva, L., & Martin, J. W. (2012). Quantitative and qualitative analysis of naphthenic acids in natural waters surrounding the Canadian oil sands industry. Environmental Science & Technology, 46(23), 12796-12805. Rowland, S. J., Scarlett, A. G., Jones, D., West, C. E., & Frank, R. A. (2011). Diamonds in the rough: identification of individual naphthenic acids in oil sands process water. Environmental Science & Technology, 45(7), 3154-3159.

68 Rudzinski, W. E., Oehlers, L., Zhang, Y., & Najera, B. (2002). Tandem mass spectrometric characterization of commercial naphthenic acids and a Maya crude oil. Energy & Fuels, 16(5), 1178-1185. Saab, J., Mokbel, I., Razzouk, A. C., Ainous, N., Zydowicz, N., & Jose, J. (2005). Quantitative extraction procedure of naphthenic acids contained in crude oils. Characterization with different spectroscopic methods. Energy & Fuels, 19(2), 525-531. Sasaki, H., Mochizuki, A., & Isogami, H. (2012). U.S. Patent Application No. 14/369,723. Washington, DC: U.S. Patent and Trademark Office. Scarlett, A. G., West, C. E., Jones, D., Galloway, T. S., & Rowland, S. J. (2012). Predicted toxicity of naphthenic acids present in oil sands process-affected waters to a range of environmental and human endpoints. Science of the Total Environment, 425, 119-127. Schwarze, M. (2017). Micellar-enhanced ultrafiltration (MEUF)–state of the art. Environmental Science: Water Research & Technology, 3, 598-624. Scott, A. C., Zubot, W., MacKinnon, M. D., Smith, D. W., & Fedorak, P. M. (2008). Ozonation of oil sands process water removes naphthenic acids and toxicity. Chemosphere, 71(1), 156- 160. Semerjian, L., & Ayoub, G. M. (2003). High-pH–magnesium coagulation–flocculation in wastewater treatment. Advances in Environmental Research, 7(2), 389-403. Shah, S. N., Mutalib, M. I. A., Pilus, R. B. M., & Lethesh, K. C. (2014). Extraction of naphthenic acid from highly acidic oil using hydroxide-based ionic liquids. Energy & Fuels, 29(1), 106- 111. Shah, S. N., Lethesh, K. C., Abdul Mutalib, M. I., Pilus, M., & Binti, R. (2015). Extraction and recovery of naphthenic acid from acidic oil using supported ionic liquid phases (SILPs). Chemical Product and Process Modeling, 10(4), 221-228. Shah, S. N., Chellappan, L. K., Gonfa, G., Mutalib, M. I. A., Pilus, R. B. M., & Bustam, M. A. (2016a). Extraction of naphthenic acid from highly acidic oil using phenolate based ionic liquids. Chemical Engineering Journal, 284, 487-493. Shah, S. N., Ismail, M., Mutalib, M. I. A., Pilus, R. B. M., & Chellappan, L. K. (2016b). Extraction and recovery of toxic acidic components from highly acidic oil using ionic liquids. Fuel, 181, 579-586.

69 Shah, S. N., Mutalib, M. A., Ismail, M. F., Suleman, H., Lethesh, K. C., & Pilus, R. B. M. (2016c). Thermodynamic modelling of liquid-liquid extraction of naphthenic acid from dodecane using imidazolium based phenolate ionic liquids. Journal of Molecular Liquids, 219, 513-525. Shah, S. N., Pranesh, M., Raj, J. J., Mutalib, M. A., Lethesh, K. C., Ghanem, O. B., & Ullah, Z. (2017). De-acidification of crude oil using supported ionic liquids phases. Separation and Purification Technology. Shi, L. J., Shen, B. X., & Wang, G. Q. (2008). Removal of naphthenic acids from Beijiang crude oil by forming ionic liquids. Energy & Fuels, 22(6), 4177-4181. Shi, L., Wang, G., & Shen, B. (2010). The removal of naphthenic acids from Beijiang crude oil with a sodium hydroxide solution of ethanol. Petroleum Science and Technology, 28(13), 1373- 1380. Shohaimi, N. A. M., Bakar, W. A. W. A., Jaafar, J., & Shukri, N. M. (2013). Treatment of acidic petroleum crude oil utilizing catalytic neutralization technique of magnesium oxide catalyst. Modern Chemistry & Applications, 1(3), 1–5. Shohaimi, N. A. M., Bakar, W. A. W. A., & Jaafar, J. (2014). Catalytic neutralization method for naphthenic acid removal in crude oil by alumina supported Ca and Ba catalysts. Petroleum Science and Technology, 32(19), 2365-2375. Shohaimi, N. A. M., Bakar, W. A. W. A., & Jaafar, J. (2017). The catalytic deacidification of acidic crude oil using Cu-doped alkaline earth metal oxide catalysts. Petroleum Science and Technology, 35(11), 1097-1103. Shukri, N. M., Bakar, W. A. W. A., Jaafar, J., & Majid, Z. A. (2015). Removal of naphthenic acids from high acidity Korean crude oil utilizing catalytic deacidification method. Journal of Industrial and Engineering Chemistry, 28, 110-116. Silva, J. P., Costa, A. L., Chiaro, S. S., Delgado, B. E., de Figueiredo, M. A., & Senna, L. F. (2013). Carboxylic acid removal from model petroleum fractions by a commercial clay adsorbent. Fuel Processing Technology, 112, 57-63. Singla, R., Ashokkumar, M., & Grieser, F. (2004). The mechanism of the sonochemical degradation of benzoic acid in aqueous solutions. Research on Chemical Intermediates, 30(7- 8), 723-733. Slavcheva, E., Shone, B., & Turnbull, A. (1999). Review of naphthenic acid corrosion in oilrefining. British Corrosion Journal, 34(2), 125-131.

70 Smith, B. E., Lewis, C. A., Belt, S. T., Whitby, C., & Rowland, S. J. (2008a). Effects of alkyl chain branching on the biotransformation of naphthenic acids. Environmental Science & Technology, 42(24), 9323-9328. Smith, D. F., Rodgers, R. P., Rahimi, P., Teclemariam, A., & Marshall, A. G. (2008b). Effect of thermal treatment on acidic organic species from Athabasca bitumen heavy vacuum gas oil, analyzed by negative-ion electrospray Fourier transform ion cyclotron resonance (FT-ICR) mass spectrometry. Energy & Fuels, 23(1), 314-319. Sohrabi, V., Ross, M. S., Martin, J. W., & Barker, J. F. (2013). Potential for in situ chemical oxidation of acid extractable organics in oil sands process affected groundwater. Chemosphere, 93(11), 2698-2703. Sun, Y., & Shi, L. (2012). Basic ionic liquids with imidazole anion: New reagents to remove naphthenic acids from crude oil with high total acid number. Fuel, 99, 83-87. Tatsi, A. A., Zouboulis, A. I., Matis, K. A., & Samaras, P. (2003). Coagulation–flocculation pretreatment of sanitary landfill leachates. Chemosphere, 53(7), 737-744. Toor, N. S., Franz, E. D., Fedorak, P. M., MacKinnon, M. D., & Liber, K. (2013). Degradation and aquatic toxicity of naphthenic acids in oil sands process-affected waters using simulated wetlands. Chemosphere, 90(2), 449-458. Trachte, K. L., & Robbins, W. K. (1999). U.S. Patent No. 5,897,769. Washington, DC: U.S. Patent and Trademark Office. VanMensel, D., Chaganti, S. R., Boudens, R., Reid, T., Ciborowski, J., & Weisener, C. (2017). Investigating the microbial degradation potential in oil sands fluid fine tailings using gamma irradiation: A metagenomic perspective. Microbial Ecology, 74, 362-372. Wang, Y., Chu, Z., Qiu, B., Liu, C., & Zhang, Y. (2006). Removal of naphthenic acids from a vacuum fraction oil with an ammonia solution of ethylene glycol. Fuel, 85(17), 2489-2493. Wang, S., & Peng, Y. (2010). Natural zeolites as effective adsorbents in water and wastewater treatment. Chemical Engineering Journal, 156(1), 11-24. Wang, Y. Z., Li, J. Y., Sun, X. Y., Duan, H. L., Song, C. M., Zhang, M. M., & Liu, Y. P. (2014a). Removal of naphthenic acids from crude oils by fixed-bed catalytic esterification. Fuel, 116, 723-728.

71 Wang, Y. Z., Duan, H. L., Song, C. M., Han, X. T., & Ma, X. R. (2014b). Removal of naphthenic

acids from crude oils by catalytic decomposition using Mg–Al hydrotalcite/γ-Al2O3 as a catalyst. Fuel, 134, 499-504. Wang, C., Alpatova, A., McPhedran, K. N., & El-Din, M. G. (2015). Coagulation/flocculation process with polyaluminum chloride for the remediation of oil sands process-affected water: Performance and mechanism study. Journal of Environmental Management, 160, 254-262. Wang, H., Liu, X., Wu, Y., Hou, C., Qiu, Y., & Guo, K. (2017). Microwave-assisted synthesis of ethylene glycol-intercalated NiAl LDHs and their application for intracrystalline catalytic esterification with naphthenic acids in crude oil. Energy & Fuels, 31(9), 9898-9904. Whitby, C. (2010). Microbial naphthenic acid degradation. Advances in Applied Microbiology, 70, 93-125. Winter, K., & Barton, D. (1970). The thermal decomposition of benzoic acid. Canadian Journal of Chemistry, 48(24), 3797-3801. Xiao, Y., & Hill, J. M. (2017). Impact of pore size on Fenton oxidation of methyl orange adsorbed on magnetic carbon materials: trade-off between capacity and regenerability. Environmental Science & Technology, 51(8), 4567-4575. Xu, X., Pliego, G., Zazo, J. A., Sun, S., García-muñoz, P., He, L., Casas, J.A., & Rodriguez, J. J. (2017). An overview on the application of advanced oxidation processes for the removal of naphthenic acids from water. Critical Reviews in Environmental Science and Technology, 47(15), 1337-1370. Xue, J., Zhang, Y., Liu, Y., & El-Din, M. G. (2016). Treatment of raw and ozonated oil sands process-affected water under decoupled denitrifying anoxic and nitrifying aerobic conditions: a comparative study. Biodegradation, 27(4-6), 247-264. Yang, B., Xu, C., Zhao, S., Hsu, C. S., Chung, K. H., & Shi, Q. (2013). Thermal transformation of acid compounds in high TAN crude oil. Science China Chemistry, 56(7), 848-855. Yokoyama, T., & Yamagata, N. (2001). Hydrogenation of carboxylic acids to the corresponding aldehydes. Applied Catalysis A: General, 221(1), 227-239. Yue, S., Ramsay, B. A., & Ramsay, J. A. (2015). Biodegradation of naphthenic acid surrogates by axenic cultures. Biodegradation, 26(4), 313-325.

72 Zafar, F., Mandal, P. C., & Moniruzzaman, M. (2016). Total acid number reduction of naphthenic acid using subcritical methanol and 1-butyl-3-methylimidazolium octylsulfate. Procedia Engineering, 148, 1074-1080. Zafar, F., Mandal, P. C., Shaari, K. Z. B. K., & Ullah, Z. (2017). Coupling of subcritical methanol with acidic ionic liquids for the acidity reduction of naphthenic acids. Polish Journal of Chemical Technology, 19(3), 68-74. Zhang, A., Ma, Q., Wang, K., Liu, X., Shuler, P., & Tang, Y. (2006). Naphthenic acid removal from crude oil through catalytic decarboxylation on magnesium oxide. Applied Catalysis A: General, 303(1), 103-109. Zhang, Y., Klamerth, N., & El-Din, M. G. (2016). Degradation of a model naphthenic acid by nitrilotriacetic acid–modified Fenton process. Chemical Engineering Journal, 292, 340-347. Zou, L., Han, B., Yan, H., Kasperski, K. L., Xu, Y., & Hepler, L. G. (1997). Enthalpy of adsorption and isotherms for adsorption of naphthenic acid onto clays. Journal of Colloid and Interface Science, 190(2), 472-475. Zhu, G. T., He, S., He, X. M., Zhu, S. K., & Feng, Y. Q. (2017). A micro-solid phase extraction in glass pipette packed with amino-functionalized silica for rapid analysis of petroleum acids in crude oils. RSC Advances, 7(64), 40608-40614. Zubot, W., MacKinnon, M. D., Chelme-Ayala, P., Smith, D. W., & El-Din, M. G. (2012). Petroleum coke adsorption as a water management option for oil sands process-affected water. Science of the Total Environment, 427, 364-372.

73 Chapter Three: Computational Methods

3.1 Computational Chemistry

Computational chemistry employs theoretical chemistry and computers to calculate the structures and properties of individual atoms or molecules, such as molecular geometry, energies of molecules and transition states, chemical reactivity, spectra, and other physical properties of substances (Lewars 2016). Today, with the increase of computational resources, computational chemistry has been applied to different areas of research including biomolecules, drugs, polymers, and inorganic and organic molecules.

The Schrödinger equation is fundamental to quantum theory, and is the basis for nearly all the computational methods. There are two versions of the Schrödinger equation: time-independent

(Eq. 3-1) and time-dependent (Eq. 3-2):

Ĥψ(푅̅, 푟̅) = Εψ(푅̅, 푟̅) Eq. 3-1

∂Ψ(푅̅, 푟̅, 푡) Eq. 3-2 ĤΨ(푅̅, 푟̅, 푡) = 푖ℏ ∂t

Ĥ is the Hamiltonian operator for a molecule with M nuclei and N electrons. ψ and Ψ are the time- independent and time-dependent wave functions for electrons, respectively, and E is the eigenvector or energy of the system. 푅̅ and 푟̅ are the positions for nuclei and electrons, respectively. The wave function is a description of the electron as a wave which determines the probability of electrons in certain locations (Young 2004). The atomic units of Hamiltonian operator for a molecule usually contains kinetic energy and potential energy terms, and it can be expressed by Eq. 3-3:

푁 푀 푁 푀 푁 푁 푀 푀 1 1 1 푍 1 푍 푍 Eq. 3-3 ̂ 2 2 퐴 퐴 퐵 H = − ∑ ∇푖 − ∑ 훻퐴 − ∑ ∑ + ∑ ∑ + ∑ ∑ 2 2 푀퐴 푟푖퐴 푟푖푗 푅퐴퐵 푖=1 퐴=1 푖=1 퐴=1 푖=1 푗>푖 퐴=1 퐵>퐴 where ∇2 is the Laplacian operator given by:

휕2 휕2 휕2 Eq. 3-4 ∇2= + + 휕푥2 휕푦2 휕푧2

In Eq. 3-3, A and B represent the M nuclei, i and j stand for the N electrons, and Z denotes atomic number. The first two items are the kinetic energy for electrons and nuclei, respectively. The last three items are the potential energy of the electron-nuclear attractions, electron-electron repulsions, and nuclear-nuclear repulsions (Koch and Holthausen 2015).

The time-dependent Schrödinger equation is a second-order differential equation, which is only analytically solvable for the hydrogen atom (Nakatsuji 2012). For other atoms, it cannot be solved exactly. However, since nuclei move much slower than electrons, their kinetic energy can be approximated to be zero, and the potential energy owing to nucleus-nucleus repulsion is almost constant (Koch and Holthausen 2015). Therefore, the Hamiltonian operator in the Schrödinger equation can be further simplified according to Born-Oppenheimer approximation, given by:

푁 푁 푀 푁 푁 1 푍 1 Eq. 3-5 ̂ 2 퐴 ̂ ̂ ̂ H = − ∑ ∇푖 − ∑ ∑ + ∑ ∑ = 푇 + 푉푒푥푡 + 푉푒−푒 2 푟푖퐴 푟푖푗 푖=1 푖=1 퐴=1 푖=1 푗>푖 where 푇̂, 푉̂푒푥푡 , and 푉̂푒−푒 represent the kinetic energy, the external potential, and the electron- electron repulsion, respectively (Young 2004).

75 3.2 Density Functional Theory

There are five main computational methods: molecular mechanics, ab initio calculations, semiempirical calculations, density functional theory (DFT), and molecular dynamics (MDs)

(Lewars 2016). DFT is the most widely used method thus far since it can be used to study large molecules with relatively high speed and accuracy (van Mourik et al. 2014). Compared to Hartree-

Fock theory, it generally provides more improved results (Jensen 2017).

There are two kinds of repulsive forces between electrons. According to the Pauli exclusion principle, two electrons with parallel spins are not allowed to occupy the same position. Therefore, the repulsive forces between electrons with parallel spins leads to the formation of exchange energy. In addition, repulsive electron-electron interactions, which are also referred to as the

Coulomb correction, can be formed between interacting electrons. The correction energy is the difference between Coulomb correlation energy and exchange energy (Mizutani 2001). Hartree-

Fock method and semi empirical calculations are not adequate to take account of electron correlation energy. Nevertheless, the electron correlation is included in the DFT method.

The principle of DFT is related to Hohenberg-Kohn theorem, specifically the electron density

휌(푟⃗) can determine the ground-states energy according to (Alcami el al. 2001):

퐸 = 퐸(휌(푟⃗)) Eq. 3-6

There are three reasons accounting for why the electron density can define the system. One is that the electron density 휌(푟⃗) determines the probability of finding electrons, which is zero at infinity, and can be integrated to obtain the total number of electrons given by (Koch and Holthausen 2015):

76 Eq. 3-7 ∫ 휌(푟⃗)푑푟⃗ = 푁

The other two reasons are that the position and corresponding charge of nuclei can be determined by the cusps in the density and the heights of cusps, respectively (Jensen 2017). For the Born-

Oppenheimer approximation, Eq. 3-5, the wave function of ab initio calculations has 3N coordinates, whereas DFT is a three-dimensional function since it is not based on the wave function and only needs electron density (Alcami el al. 2001). The electronic ground state energy can be calculated by solving the Schrödinger equation or through the Rayleigh-Ritz minimal principle, given by:

̃ ̃ 퐸 = 푚푖푛Ψ̃ (Ψ, 퐻Ψ) Eq. 3-8 where Ψ̃ is a trial function (Kohn 1999). The Kohn-Sham method, which maps the system of interacting electrons to a fictitious system of non-interacting particles, is the most successful way to implement and realize DFT so far (Yu et al. 2016). The total energy can be calculated by:

Eq. 3-9 퐸[휌(푟⃗)] = 푇[휌(푟⃗)] + 퐸푒푒[휌(푟⃗)] + 퐸푥푐[휌(푟⃗)] + ∫ 푉푒푥푡(푟⃗) 휌(푟⃗)푑푟

where 푇, 퐸푒푒, and 퐸푥푐 are the kinetic energy, electron-electron repulsion energy, and exchange- correction energy, respectively. According to this method, each electron moves in an effective single-particle potential, following the single particle Schrödinger equation:

ℏ2 Eq. 3-10 {− ∇2 + 푉 [휌(푟⃗)]} ψ (푟⃗) = ϵ ψ (푟⃗) 2푚 푒푓푓 푖 푖 푖 where 푉푒푓푓 denotes an effective potential and it can be expressed according to:

푉푒푓푓[휌(푟⃗)] = 푉푒푥푡[휌(푟⃗)] + 푉푒−푒[휌(푟⃗)] + 푉푥푐[휌(푟⃗)] Eq. 3-11

푉푥푐 in Eq. 3-11 represents the exchange-correlation potential, defined by:

77 휕퐸 [휌(푟⃗)] Eq. 3-12 푉 [휌(푟⃗)] = 푥푐 푥푐 휕휌(푟⃗)

The density of a system can be determined according to (Malet et al. 2013):

2 Eq. 3-13 ρ(푟⃗) = ∑|ψ푖(푟⃗)| 푖

In addition, 푇 and 퐸푥푐 are expressed in Eqs. 3-14 and 3-15, respectively (Cohen et al. 2012):

ℏ2 1 Eq. 3-14 푇[휌(푟⃗)] = ∑ 〈ψ |− ∇2| ψ 〉 푖 2푚 2 푖 푖

1 휌(푟)휌(푟,) Eq. 3-15 퐸 [휌(푟⃗)] = ∬ 푑푟푑푟′ 푒푒 2 |푟 − 푟′|

3.2.1 Basis Sets

A set of basis functions are used to describe the molecular orbitals (Foresman & Frisch 1996). The

STO-3G basis set contains the minimum number of basis functions required to describe each atom: specifically 1s, 2s, 2px, 2py, and 2pz orbitals are all represented by three Gaussian functions fitted to a Slater-type orbital (STO) (Withnall et al. 2007). To enlarge the basis set, a split valence basis set such as 3-21G and 6-31G are adopted (Foresman and Frisch 1996). For carbon 3-21G, it utilizes three Gaussian functions to represent 1s, two Gaussian functions to represent 2s and 2p, and one

Gaussian function to denote 2s´ and 2p´; with regards to carbon 6-31G, six, three, and one Gaussian functions are employed to represent 1s, 2s and 2p, and 2s´ and 2p´, respectively (Withnall et al.

2007).

The use of a polarized basis set such as 6-31G(d) increases the angular flexibility through adding orbitals, thereby permitting the polarization of electrons and leading to more accurate geometries

78 and frequencies (Young 2004). For instance, potassium 6-31G(d) means adding d-functions to the valence orbital of potassium (Rassolov et al. 2001). Although 6-31G(d) is quite popular, 6-

31G(d,p) basis set has better performance than 6-31G(d) when hydrogen atoms participate in hydrogen bonding or bridging (Lawars et al. 2016).

A basis set with diffuse functions allows orbitals to take up a larger area. It is suitable when the electron is located far away from the nucleus, such as anions, Rydberg states, and loose supermolecular complexes (Cramer 2013). 6-31+G(d) implies that additional s and p functions have been added to heavy atoms. Considering its accuracy and moderate computational time, 6-

311++G(d,p), which adds diffusion functions to hydrogen and heavy atoms, was chosen as the basis set used in the research to study interactions between ionic liquids (ILs) and naphthenic acids

(NAs).

3.2.2 Methods

For the study of weak interactions, although post Hartree-Fock methods such as the second-order

Møller–Plesset perturbation (MP2) method and the coupled-cluster with single, double, and perturbative triple excitations [CCSD(T)] can get accurate results, they are relatively time- consuming to achieve convergence (Sato and Nakai 2009). The hybrid density functional method

B3LYP can obtain fairly accurate chemical structures, however, it is not suitable to study intermolecular interactions since it does not reproduce weakly bonded complexes other than hydrogen-bonding between organic compounds and water (Tirado-Rives et al. 2008).

79 The M06-2X hybrid density functional method is suitable for nonmetals calculation and is recommended to calculate main-group thermochemistry, noncovalent interactions (NCI), and electronic excitation energies to valence and Rydberg states (Zhao and Truhlar 2008). It also has better performance than the B3LYP and PW91 methods for systems with dispersion and ionic hydrogen-bonding interactions (Walker et al. 2013), and is comparable to post Hartree-Fock methods CCSD(T) and MP2 in describing NCI (Lemke and Seward 2013). Therefore, the M06-

2X method was adopted for the research documented in this thesis to study the interactions between model NAs and ILs.

It should be noted that although DFT calculations provide better results for electrostatic interactions and hydrogen bonding, exchange-correlation functionals in DFT methods are not adequate to describe non-bonded interactions such as London dispersion forces (Doemer et al.

2012). Dispersion effects are reported to be important in describing IL interactions (Grimme et al.

2012). Consequently, an accurate and reliable study of IL interactions must account for dispersion effects. Dispersion corrected density functional theory (DFT-D3) has been demonstrated to be accurate to study interaction energies for large complexes, and it has higher accuracy than the

DFT-D1 and DFT-D2 methods (Grimme 2012). Hence, the M06-2X method was combined with dispersion correction DFT-D3 to explore the interactions between ILs and NAs.

3.2.3 Solvation Models

A huge number of molecules interact with each other in solutions. Solvent-solvent interactions influence molecular energies, structures, and properties (Cossi et al. 2003). To accurately calculate

80 the reactions in the aqueous phase, a suitable solvation model is necessary. Generally speaking, there are two commonly used solvation models. One model is the continuum model or implicit solvent model, where solvent is represented as a dielectric continuum. The other model is the explicit model, where the solvent is represented by simplified molecular mechanics (Da Silva and

Svendsen 2007). The implicit solvation model based on density (SMD) model represents solute by an electron charge density and can be applied to solute in any solvent or liquid medium (Marenich et al. 2009). It has been widely used to calculate the rate constants of reactions between radicals and other compounds in the aqueous phase (Galano and Alvarez‐Idaboy 2014; Minakata et al.

2011). In this thesis, the rate constants for the reaction between BA and hydroxyl radicals in the aqueous phase were calculated by using an implicit SMD model.

3.3 Molecular Dynamics

There are two main molecular simulation techniques: MDs and Monte Carlo molecular modelling

(Allen 2004). It was reported that MDs simulation of liquids or solvent-solute problems is able to obtain different properties of liquids or solutions, including diffusion coefficients, densities, viscosities, and solubility parameters (Choi 2000; Young 2004). MD models can predict the properties of ILs yielding a fundamental understanding of ILs more easily than through experimental studies (Borodin 2009). In fact, the first publication of MDs simulation of ILs was reported in 2001 for imidazolium-based ILs (Kowsari et al. 2008).

MD is a simulation of the transient molecular behavior including vibrational motion and Brownian motion, which requires a set of initial coordinates, velocities, and interaction potentials (Jensen

81 2017). The simulation process involves solving the equations of motion. Eq. 3-16 and Eq. 3-17 can provide information about particle positions, and the force of each particle, respectively.

푚푖푟푖̈ = 푓푖 Eq. 3-16

휕푈 Eq. 3-17 푓푖 = − 휕푟푖 where 푟 is the 3N atomic coordinates, and 푈 is the potential energy consisting of non-bonded potentials and bonded potentials.

The selection of the interaction potential (force field) is critical to accurately predict IL properties

(Borodin 2009). The condensed-phase optimized molecular potentials for atomistic simulation studies (COMPASS) force field, developed by Sun et al. (1998), has been reported to be an ab initio force field for condensed-phase applications (Li and Choi 2007). Moreover, it is accurate in predicting physical properties of different molecules (Li and Choi 2008). The COMPASS force field has been applied to predict the densities and solubility parameters of ILs with satisfactory results (Derecskei and Derecskei-Kovacs 2008). The COMPASS II force field, which has extended the coverage of COMPASS force field while preserving the original parameters, was used in this thesis to study the physicochemical properties of ILs (Sun et al. 2016).

3.4 Structure Analyses

After optimizing structures, analyses of the electrostatic potential, natural bond orbital (NBO), atoms in molecules (AIM), noncovalent interactions (NCI), and electron density differences, were conducted to analyze wave functions to investigate interactions comprehensively. An introduction of NBO, AIM, and NCI analysis is provided in this section.

3.4.1 NBO Analysis

NBO analysis provides insights of intramolecular and intermolecular orbital interactions in the complexes, such as atomic charge, bond type, hybridization, bond order, and charge transfer

(Glendening et al. 2012). Here, charge distributions and stabilization energy E(2) for donor- acceptor was analyzed. The stabilization energy E(2) of possible interactions between filled donors

(Lewis-type NBOs) and empty acceptors (Non-Lewis-type NBOs) is evaluated by second-order perturbation theory (Gangadharan and Krishnan 2014). Figure 3.1 depicts the interaction between filled donor and empty acceptor as well as the formation of stabilization energy E(2) (Reed et al.

1998). The stabilization energy E(2) was calculated according to:

2 퐹푖,푗 Eq. 3-18 퐸(2) = 푞푖 휀푖 − 휀푗 where 푞푖 is the orbital occupancy, 휀푖 and 휀푗 are diagonal elements (orbital energies), and 퐹푖,푗 is the off-diagonal NBO Fock matrix element (Gangadharan and Krishnan 2014). A higher value of E(2) implies stronger donor-acceptor interaction.

Figure 3.1 Perturbation donor-acceptor interaction.

3.4.2 AIM Analysis

AIM theory was first developed by Bader and his co-workers in the early 1970s to employ the topology of electron densities to describe both intramolecular and intermolecular interactions

(Matta and Boyd 2007). The critical point for density is the point where the first derivatives of density is zero, given by:

휕휌 휕휌 휕휌 Eq. 3-19 ∇휌 = + + = 0 휕푥 휕푦 휕푧

All the nine second derivatives of density are arranged to obtain a Hessian matrix:

휕2휌 휕2휌 휕2휌 Eq. 3-20 휕푥2 휕푥휕푦 휕푥휕푧 휕2휌 휕2휌 휕2휌 퐴(푟 ) = 푐 휕푦휕푥 휕푦2 휕푦휕푧

휕2휌 휕2휌 휕2휌 (휕푧휕푥 휕푧휕푦 휕푧2 ) 푟=푟퐶

84 According to the number of nonzero eigenvalues of the Hessian matrix and the algebraic sum of the signs of eigenvalues, critical points in a molecule are classified into four types: (3, -3) nuclear critical point (ncp), (3, -1) bond critical point (bcp), (3, +1) ring critical point (rcp), and (3, +3) cage critical point (ccp) (Bader 1985). For ncp, it has three negative eigenvalues; as for bcp, it has two negative and one positive eigenvalues (Bader 1985).

The bcp is the point on the bond path with smallest electron density (Bader 1990). The electron density at the bcp can help illustrate the strength of chemical bonds through evaluating bond order

(BO) according to:

퐵푂 = exp[퐴(휌푏 − 퐵)] Eq. 3-21 where A and B are constants related to the nature of bonded atoms. Larger electron density at the bcp indicates a stronger chemical bond (Popelier 2000). The Laplacian of electron density(∇2휌) is evaluated in the AIM method: ∇2휌<0 means that the electron density is concentrated and higher at the given point than neighbors, whereas ∇2휌 >0 implies that the electron density is depleted and the electron density is higher at the neighbors of given points (Popelier 2000).

3.4.3 NCI Analysis

When molecules or atoms interact with each other, covalent interactions lead to the formation of new molecules; whereas NCI generate molecular complexes. For covalent interactions, the orbitals of interacting atoms overlap and a pair of electrons are shared by the interacting atoms, whereas the orbitals of NCI are not necessarily overlapped (Müller-Dethlefs and Hobza 2000). Hydrogen- bonding, electrostatic interactions, charge-transfer interactions, and dispersion interactions belong

85 to NCI (Müller-Dethlefs and Hobza 2000). In addition, the strength and stabilization energy of

NCI is much weaker than that of covalent interactions, therefore, it is difficult to measure it

(Müller-Dethlefs and Hobza 2000).

Johnson et al. (2010) presented an approach to analyze NCI through drawing the plots of the reduced density gradient (RDG) versus the electron density multiplied by the sign of the second

Hessian eigenvalue (sign(λ2)ρ). The RDG is a dimensionless quantity and calculated from:

|∇휌(푟)| Eq. 3-22 푠 = 1 4 2(3휋2) ⁄3휌(푟) ⁄3

For the eigenvalues (λ1< λ2< λ3) of the electron-density Hessian matrix (Eq. 3-20), λ3 depends on the nuclear position, and λ1 and λ2 are related to the electron density normal to the λ3 eigenvector

(Contreras-Garía et al. 2011). The sign of λ2 is decided by the interaction types: λ2<0 means bonding interactions such as hydrogen bonds; λ2>0 indicates non-bonded interactions such as steric repulsion; and λ2 ≲0 designates van der Waals (vdW) interactions (Johnson et al. 2010). The electron density of NCI can also illustrate the length of chemical bonds.

3.5 References

Alcami, M., Mo, O., & Yáñez, M. (2001). Computational chemistry: A useful (sometimes mandatory) tool in mass spectrometry studies. Mass Spectrometry Reviews, 20(4), 195-245. Allen, M. P. (2004). Introduction to molecular dynamics simulation. Computational Soft Matter: from Synthetic Polymers to Proteins, 23, 1-28. Bader, R. F. (1985). Atoms in molecules. Accounts of Chemical Research, 18(1), 9-15. Bader, R. F. (1990). Atoms in molecules: A quantum theory. Oxford: University of Oxford Press. Borodin, O. (2009). Polarizable force field development and molecular dynamics simulations of ionic liquids. The Journal of Physical Chemistry B, 113(33), 11463-11478.

86 Choi, P. (2000). Molecular dynamics studies of the thermodynamics of HDPE/butene-based LLDPE blends. Polymer, 41(24), 8741-8747. Cohen, A. J., Mori-Sánchez, P., & Yang, W. (2012). Challenges for density functional theory. Chemical Reviews, 112(1), 289-320. Contreras-García, J., Johnson, E. R., Keinan, S., Chaudret, R., Piquemal, J. P., Beratan, D. N., & Yang, W. (2011). NCIPLOT: a program for plotting noncovalent interaction regions. Journal of Chemical Theory and Computation, 7(3), 625-632. Cossi, M., Rega, N., Scalmani, G., & Barone, V. (2003). Energies, structures, and electronic properties of molecules in solution with the CPCM solvation model. Journal of Computational Chemistry, 24(6), 669-681. Cramer, C. J. (2013). Essentials of computational chemistry: theories and models. Chichester: John Wiley & Sons. Da Silva, E. F., & Svendsen, H. F (2007). Computational chemistry study of reactions, equilibrium

and kinetics of chemical CO2 absorption. International Journal of Greenhouse Gas Control, 1(2), 151-157. Derecskei, B., & Derecskei-Kovacs, A. (2008). Molecular modelling simulations to predict density and solubility parameters of ionic liquids. Molecular Simulation, 34(10-15), 1167-1175. Doemer, M., Tavernelli, I., & Rothlisberger, U. (2012). Intricacies of describing weak interactions involving Halogen atoms within density functional theory. Journal of Chemical Theory and Computation, 9(2), 955-964. Foresman, J. B., & Frisch, Æ. (1993). Exploring chemistry with electronic structure methods: a guide to using Gaussian. Pittsburgh, PA: Gaussian. Gangadharan, R. P., & Krishnan, S. S. (2014). Natural bond orbital (NBO) population analysis of 1-azanapthalene-8-ol. Acta Physica Polonica, A., 125(1), 18-22. Galano, A., & Alvarez‐Idaboy, J. R. (2014). Kinetics of radical‐molecule reactions in aqueous solution: A benchmark study of the performance of density functional methods. Journal of Computational Chemistry, 35(28), 2019-2026. Glendening, E. D., Landis, C. R., & Weinhold, F. (2012). Natural bond orbital methods. Wiley Interdisciplinary Reviews: Computational Molecular Science, 2(1), 1-42.

87 Grimme, S., Hujo, W., & Kirchner, B. (2012). Performance of dispersion-corrected density functional theory for the interactions in ionic liquids. Physical Chemistry Chemical Physics, 14(14), 4875-4883. Grimme, S. (2012). Supramolecular binding thermodynamics by dispersion‐corrected density functional theory. Chemistry-A European Journal, 18(32), 9955-9964. Johnson, E. R., Keinan, S., Mori-Sanchez, P., Contreras-Garcia, J., Cohen, A. J., & Yang, W. (2010). Revealing noncovalent interactions. Journal of the American Chemical Society, 132(18), 6498-6506. Jensen, F. (2017). Introduction to computational chemistry. Chichester: John wiley & sons. Kohn, W. (1999). Nobel Lecture: Electronic structure of matter—wave functions and density functionals. Reviews of Modern Physics, 71(5), 1253. Koch, W., & Holthausen, M. C. (2015). A chemist's guide to density functional theory. Chichester: John Wiley & Sons. Kowsari, M. H., Alavi, S., Ashrafizaadeh, M., & Najafi, B. (2008). Molecular dynamics simulation of imidazolium-based ionic liquids. I. Dynamics and diffusion coefficient. The Journal of Chemical Physics, 129(22), 224508. Lemke, K. H., & Seward, T. M. (2013). Thermodynamic properties of carbon dioxide clusters by M06-2X and dispersion-corrected B2PLYP-D theory. Chemical Physics Letters, 573, 19-23. Lewars, E. G. (2016). Computational chemistry: introduction to the theory and applications of molecular and quantum mechanics. Cham, Switzerland: Springer. Li, C., & Choi, P. (2007). Molecular dynamics study of the adsorption behavior of normal alkanes

on a relaxed α-Al2O3 (0001) surface. The Journal of Physical Chemistry C, 111(4), 1747-1753. Li, C., & Choi, P. (2008). Molecular dynamics study on the effect of solvent adsorption on the

morphology of glycothermally produced α-Al2O3 particles. The Journal of Physical Chemistry C, 112(27), 10145-10152. Malet, F., Mirtschink, A., Cremon, J. C., Reimann, S. M., & Gori-Giorgi, P. (2013). Kohn-Sham density functional theory for quantum wires in arbitrary correlation regimes. Physical Review B, 87(11), 115146. Marenich, A. V., Cramer, C. J., & Truhlar, D. G. (2009). Universal solvation model based on solute electron density and on a continuum model of the solvent defined by the bulk dielectric constant and atomic surface tensions. Journal of Physical Chemistry B, 113(18), 6378-6396.

88 Matta, C. F., & Boyd, R. J. (2007). The quantum theory of atoms in molecules: from solid state to DNA and drug design. Weinheim: Wiley-VCH. Minakata, D., Song, W., & Crittenden, J. (2011). Reactivity of aqueous phase hydroxyl radical with halogenated carboxylate anions: experimental and theoretical studies. Environmental Science & Technology, 45(14), 6057-6065. Mizutani, U. (2001). Introduction to the electron theory of metals. Cambridge: Cambridge University Press. Müller-Dethlefs, K., & Hobza, P. (2000). Noncovalent interactions: a challenge for experiment and theory. Chemical Reviews, 100(1), 143-168. Nakatsuji, H. (2012). Discovery of a general method of solving the Schrödinger and Dirac equations that opens a way to accurately predictive quantum chemistry. Accounts of Chemical Research, 45(9), 1480-1490. Popelier, P. L. A. (2000). On the full topology of the Laplacian of the electron density. Coordination Chemistry Reviews, 197(1), 169-189. Rassolov, V. A., Ratner, M. A., Pople, J. A., Redfern, P. C., & Curtiss, L. A. (2001). 6‐31G* basis set for third‐row atoms. Journal of Computational Chemistry, 22(9), 976-984. Reed, A. E., Curtiss, L. A., & Weinhold, F. (1988). Intermolecular interactions from a natural bond orbital, donor-acceptor viewpoint. Chemical Reviews, 88(6), 899-926. Sato, T., & Nakai, H. (2009). Density functional method including weak interactions: Dispersion coefficients based on the local response approximation. The Journal of Chemical Physics, 131(22), 224104. Sun, H. (1998). COMPASS: an ab initio force-field optimized for condensed-phase applications overview with details on alkane and benzene compounds. The Journal of Physical Chemistry B, 102(38), 7338-7364. Sun, H., Jin, Z., Yang, C., Akkermans, R. L., Robertson, S. H., Spenley, N. A., Miller, S., & Todd, S. M. (2016). COMPASS II: extended coverage for polymer and drug-like molecule databases. Journal of Molecular Modeling, 22(2), 47. Tirado-Rives, J., & Jorgensen, W. L. (2008). Performance of B3LYP density functional methods for a large set of organic molecules. Journal of Chemical Theory and Computation, 4(2), 297- 306

89 van Mourik, T., Bühl, M., & Gaigeot, M. P. (2014). Density functional theory across chemistry, physics and biology. Philosophical Transactions of the Royal Society A, 372, 20120488. Walker, M., Harvey, A. J., Sen, A., & Dessent, C. E. (2013). Performance of M06, M06-2X, and M06-HF density functionals for conformationally flexible anionic clusters: M06 functionals perform better than B3LYP for a model system with dispersion and ionic hydrogen-bonding interactions. The Journal of Physical Chemistry A, 117(47), 12590-12600. Withnall, R., Chowdhry, B. Z., Bell, S., & Dines, T. J. (2007). Computational chemistry using modern electronic structure methods. Journal of Chemical Education, 84(8), 1364-1370. Young, D. (2004). Computational chemistry: a practical guide for applying techniques to real world problems. New York: John Wiley & Sons. Yu, H. S., Li, S. L., & Truhlar, D. G. (2016). Perspective: Kohn-Sham density functional theory descending a staircase. The Journal of Chemical Physics, 145(13), 130901. Zhao, Y., & Truhlar, D. G. (2008). The M06 suite of density functionals for main group thermochemistry, thermochemical kinetics, noncovalent interactions, excited states, and transition elements: two new functionals and systematic testing of four M06-class functionals and 12 other functionals. Theoretical Chemistry Accounts, 120(1), 215-241.

90 Chapter Four: Reactions of Hydroxyl Radicals with Benzoic Acid and Benzoate

4.1 Introduction

Produced water from mining operations, in particular, oil sands mining operations, contain solids, salts, and other organic compounds such as benzene, phenols, naphthenic acids (NAs) which are toxic to wildlife (Shu et al. 2014; Wang et al. 2016a). Despite the small amount of aromatic alkanoic acids such as benzoic acid (BA) in NAs, they increase the toxicity and recalcitrance of

NAs in wastewater (Johnson et al. 2011). In addition, BA is a common recalcitrant contaminant which exists in domestic wastewater as well as other industrial wastewater, making it imperative to remove BA from waste water (Enami et al. 2016; Wahab 2012).

Ultrasound is effective for breaking down organic compounds in water (Perrone et al. 2016). When water is treated by ultrasonic waves, cavitation within the water leads to generation and collapse of microbubbles that result in high temperatures in a small region where organic compounds are degraded through pyrolysis. In addition, organic compounds will also be oxidized by hydroxyl radicals (OH) generated by the cavitation bubbles during ultrasonic treatment (De Visscher 2003;

Lin et al. 2016; Yusof et al. 2016). Furthermore, BA can be removed during electrochemical oxidation, photochemical oxidation, Fenton oxidation, and ozonation (Huang et al. 2015; Oliveira et al. 2014). These methods involve the reaction of BA with hydroxyl radicals in the liquid phase.

Moreover, BA could also react with hydroxyl radicals inside the bubbles under ultrasound irradiation (Singla et al. 2004). BA is a weak acid with a pKa of 4.2 and it can exist in the form of

91 benzoate (BZ) in the aqueous phase (Nair 2001). Therefore, the reaction between BZ and hydroxyl radicals can also take place in the oxidation process under neutral and alkaline pH.

Quantum chemical reaction computation using density functional theory (DFT) is a powerful tool to understand mechanisms. These methods enable the calculation of optimized geometries for reactants and products, assessment of intermediate and transition states, and estimation of reaction kinetics for different reaction pathways with reasonable accuracy (Alvarez-Idaboy et al. 2001).

Theoretical studies of the mechanisms and kinetics of BA reacting with hydroxyl radicals are important to improve techniques to remove BA completely from wastewater. The distribution of the intermediates is crucial to the detoxification of BA (Li et al. 2012). San et al. studied the transition states of photodegradation of BA in liquid phase (San et al. 2007). Minakata et al. analyzed the lowest unoccupied molecular orbital and singly occupied molecular orbital of hydroxyl radical addition to BZ at ortho, meta, and para positions in the aqueous phase (Minakata et al. 2015). Li et al. investigated the effects of OH substitution on the rates and mechanisms of decarboxylation of BA (Li & Brill 2003). Chuchev & BelBruno 2007 conducted a detailed theoretical study of the decarboxylation reaction for BA ortho-derivatives. Wang et al. calculated the catalytic decarboxylation mechanisms of BA (Wang et al. 2016b). These studies are useful but do not examine uncatalyzed reaction mechanisms and kinetics of BA with hydroxyl radicals in both gas and aqueous phases. Existing studies do not consider the difference between ortho and meta carbons on the side of carboxyl group and opposite the carboxyl group. In addition, these studies did not consider pre-reactive complexes, which can alter the kinetics of OH reactions dramatically. Furthermore, there are few theoretical studies that compare rate constant differences between BA and BZ in the aqueous phase.

The objective of this chapter is to employ DFT to study the transition states of six different pathways of BA with hydroxyl radicals both in gas and aqueous phases, BZ with hydroxyl radicals in the aqueous phase, and estimate the energy barriers and reaction rate constants to determine the possible distribution of intermediates including the potential effect of pre-reactive complexes.

Moreover, the influence of an individual explicit water molecule on rate constants was investigated. The results will help understand the ultrasound irradiation of BA and BZ, determine the rate constants differences inside the cavitation bubble and bulk solution, and provide guidance for controlling ultrasound reaction conditions to efficiently degrade BA and BZ.

4.2 Computational Methods

The evolution of the reactions between BA and hydroxyl radicals was evaluated by using the

Gaussian 09 molecular modelling package (Frisch et al. 2009). Restricted and unrestricted DFT was adopted for closed and open shell systems separately. The M06-2X method was reported to be appropriate for calculating hydroxyl radical reactions, predicting optimized geometries and vibrational frequencies (Bian et al. 2017; Dash & Rajakumar 2013). 6-311+G(d,p) is a widely used basis set by researchers for radical reaction calculation (Farkas & Schlegel 2003). Considering accuracy and computational costs, M06-2X/6-311+G(d,p) was chosen as the method and basis set for the optimizations, electronic, and vibrational properties both in the gas and aqueous phases in this study. After optimization of the reactants and products, the transition states, saddle point in the potential energy surface, were optimized by a quasi-Newton synchronous transit (QST3) calculation followed by the Berny optimization method calculation (Xie et al. 2014; Zhao et al.

93 2017). Frequency analyses were conducted to confirm the transition states by one imaginary frequency and obtain zero-point corrected energies as well as free energies. Intrinsic reaction coordinate (IRC) calculations were conducted to ensure that the transition states connected the original reactants and products (Bhoorasingh et al. 2015; Sabet-Sarvestani et al. 2017).

Comparison of the calculated rate constants with experimental data also provided confidence that the pre-reactive complexes and transition states determined were adequate. The implicit solvation model based on density (SMD) was adopted to investigate the effects of water solvent on the reaction (the dielectric constant was taken to be 78.3553 for water) (Kazmierczak et al. 2017; Papp et al. 2013; Tiwari & Mishra 2016). The relative energies and rate constants were calculated at

298.15K, and 1 atm. The single-point energies were further refined employing coupled cluster theory at the level of CCSD(T)/6-311++G(d,p) (Bartlett & Musiał 2007). The study on incorporating one water molecule around the carboxylic group of BZ and BA followed the same methods as described above.

4.3 Results and Discussion

4.3.1 Reaction Pathways

Due to the existence of unpaired electrons, the hydroxyl radical is an electrophilic radical (Medina et al. 2014). When reacting with BA, there are five possible reaction pathways: ortho-addition (o- add), para-addition (p-add), meta-addition (m-add), ipso-addition (ipso-add), and hydrogen- abstraction (H-abs), respectively. On the other hand, o-add, p-add, m-add, and ipso-add are the four reaction pathways between BZ and hydroxyl radicals, as shown in Scheme 4.1.

94 The optimized structures of the reactants are depicted in Figure 4.1, indicating that the geometry of BA and BZ were planar. In addition, the bond lengths of the optimized structures of BA in the gas and liquid phases were almost the same. However, it is noteworthy that the structure of BA was not symmetric, and the reaction of o-add and m-add could happen at two different sites of the benzene ring, leading to seven different reaction pathways. Furthermore, it is demonstrated that the reaction rate constant of ipso-add is negligible compared to other reaction pathways considering the significant steric effect of the carboxylic group (Minakata et al. 2009), and this pathway was excluded in the present study. Thus, the six reaction pathways for BA with hydroxyl radicals were calculated both in the gas and aqueous phases. The additions opposite the C=O group were denoted as o-add and m-add on ortho and meta positions, respectively, whereas the additions on the side of C=O group were denoted as o2-add and m2-add, separately. Due to the symmetric structures of BZ, the reactions between BZ and hydroxyl radicals were calculated only for the ortho, meta, and para positions.

O + H2O O HO O O HO OH H H-abs HO

ipso-add o-add O + OH para-add o2-add HO HO O O OH m-add m2-add H

H HO OH HO O HO O

HO H H OH

(a) Reaction pathways of BA with hydroxyl radicals

95 O

O OH

ipso-add H HO HO H O o-add O m-add O + OH O O O

p-add

O OH

O H

(b) Reaction pathways of BZ with hydroxyl radicals

Scheme 4.1 Reaction pathways of BA and BZ with hydroxyl radicals.

(a) BA gas phase (b) BA aqueous phase (c) BZ aqueous phase

Figure 4.1 Reactants optimization.

The optimization of adducts in the gas and aqueous phases are depicted in Appendix A Figs S1,

S2, and S3. When hydroxyl radicals reacted with BA through addition reaction, the π bonds of the

96 aromatic rings were broken, and the carbon-carbon lengths were altered. In the process of hydrogen abstraction reactions, the O-H bond broke and hydrogen was abstracted to hydroxyl radicals to form a new O-H bond, leading to the formation of a water molecule. On the other hand, it was deduced that the length of the newly formed carbon-oxygen bond was below 1.45 Å in the gas phase for all addition reaction pathways, which was the same reported by the addition of hydroxyl radicals to phenol (Wu et al. 2012). When comparing the bond lengths in the gas phase and aqueous phase, the bond lengths were prolonged for the reaction taking place in the aqueous phase. In addition, the bond lengths of BZ products were longer than that of BA products.

4.3.2 Pre-reactive Complexes and Transition States

Pre-reactive complexes have been identified in many radical-molecule reactions as the precursors for reaction (Alvarez-Idaboy et al. 2001), however, they have not been reported for reactions of

BA and hydroxyl radicals thus far. From the calculations, when hydroxyl radicals approached BA, weakly bonded pre-reactive complexes were formed due to van der Waals interactions and long range coulombic interactions. The optimized pre-reactive structures for the gas phase and liquid medium reactions for BA and BZ are drawn in Appendix A Figs S4 to S6, separately. For the addition reaction, the OH radicals interacted with the benzene ring, forming the O-H···π hydrogen bond complex, whereas for the hydrogen abstraction reaction, the OH radical approached the hydrogen atom of the carboxyl group, leading to a O14-H15···O16 hydrogen bond complex.

Interestingly, the orientation of hydroxyl radicals for the ortho site reaction was different from meta and para site reactions. The hydroxyl radicals were positioned above the benzene ring plane for the reactions that occurred at meta and para sites, whereas they were almost in the same plane

97 for the o-add and o2-add reactions. In addition, the o-add had a longer distance between the hydroxyl radical and benzoic ring compared with the m-add and p-add reactions. The above phenomena could be caused by steric hindrance of the carboxyl group as well as the hydrogen bond formation between hydroxyl radicals and carboxyl group, hindering the approach of OH radicals to benzoic acids.

The transition states in the gas and liquid phases are shown in Figure 4.2 to Figure 4.4. They were confirmed by a single imaginary frequency as well as the IRC analysis. There were remarkable changes between the structures of pre-reactive complexes and transition states. The main structural changes were located around the reacting carbon and oxygen. For the o-add reaction pathway, it was observed that the C2-C1, C1-C6, C3-C4, and C5-C4 bonds elongated whereas the C2-C3 and

C5-C6 bond lengths shortened. The same trends were also found in other hydroxyl radical addition reactions both in the gas and aqueous phases. In other words, when the hydroxyl radicals reacted with a particular carbon in the benzene ring, the bond lengths of the reacting carbon with adjacent carbons increased as well as the lengths of the bonds between the carbon opposite the reacting carbon on the aromatic ring with its neighboring carbons. The remaining two carbon-carbon bond lengths in the aromatic rings were shortened. This is explained by the electron density transfer of the aromatic ring.

(a) o-add (b) o2-add (c) m-add

(d) m2-add (e) p-add (g) H-abs

Figure 4.2 Transition states optimization (BA gas phase).

(a) o-add (b) o2-add (c) m-add

(d) m2-add (e) p-add (g) H-abs

Figure 4.3 Transition states optimization (BA aqueous phase).

100

(a) o-add (b) m-add (c) p-add

Figure 4.4 Transition states optimization (BZ aqueous phase).

Comparing the transition states of hydroxyl radical addition to BA in gas phase versus aqueous phase, it is obvious that the distances of oxygen and reacting carbon of the transition states in the aqueous phase were longer than that in the gas phase, which could possibly be due to the formation of hydrogen bonds between BA and water, therefore, hindering the approach of hydroxyl radicals to benzene rings. However, it is worth noting that the bond length of oxygen and the reacting carbon for BZ o-add was shorter than BA o-add, and the difference was caused by the larger steric effects of the carboxylic group than the carboxylic anion. As for the hydrogen abstraction reaction, the bond between O14 and H15 was broken, and H15 was attracted to hydroxyl radicals, generating water and a benzoic acid free radical as products.

4.3.3 Energetics of the Reaction Paths

101 The reaction energies for the six different reaction paths with reactants, pre-reactive complexes, transition states, and products were calculated and presented in Figure 4.5 and Appendix A Table

S1 for the gas phase. The results show that the energies of pre-reactive complexes were lower than the energy of reactant, which would influence the energy barriers for the reaction. It has also been detected in previous research that pre-reactive complexes are critical to the reaction path since they would influence energy barriers and the energy partitioning of the products (Uc et al. 2006;

Degirmenci & coote 2016). The reaction energy barriers from pre-reactive complexes to transition states in the gas phase followed the order of o2-add>H-abs>o-add>m2-add>m-add>p-add, indicating that the addition of hydroxyl radicals on the same side as the carboxyl group was more difficult than on the opposite side for this step reaction, which might be caused by steric effects. It is remarkable that the energy of H-abs product was higher than addition reaction products for the gas phase reaction, indicating its instability and relative ease to proceed to the next step reaction, whereas the products of o-add and o2-add were the most stable among the seven reaction pathways.

Figure 4.5 Relative energies of six possible reaction paths (BA gas phase).

102

The relative energies for the reaction between BA and hydroxyl radicals in the aqueous phase are displayed in Figure 4.6 and Appendix A Table S2. Compared with the reaction paths in the liquid and gas phases, it was deduced that the reactants needed to overcome smaller energy barriers for the reaction from pre-reactive complexes to transition states. The reaction energy barriers for the reaction from pre-reactive complexes to transition states in the aqueous medium was highest for

H-abs. In addition, the energy barrier for H-abs in the aqueous phase was much higher than in the gas phase, which was possibly caused by hydrogen bonding formation between the carboxylic group and water molecules. Therefore, the reaction in the gas phase was preferable for H-abs reaction compared to the aqueous phase reaction. According to the energy, o2-add products were most stable among the six different reaction pathways, making them less vulnerable to subsequent reaction.

)

(

Figure 4.6 Relative energies of six possible reaction paths (BA aqueous phase).

103 As for the reaction energies between BZ and hydroxyl radicals, displayed in Figure 4.7 and

Appendix A Table S3, the energy barriers from pre-reactive complexes to transitions states followed the order of o-add>m-add>p-add with p-add transition state having the longest bond length among the addition reactions. This is explained because the p-add transition state was the earliest transition state among the addition reaction pathways from pre-reactive complexes for the reaction between BZ and hydroxyl radicals (Uc et al. 2006). In addition, the product of o-add had the lowest energy among all the products, which was the same for the reaction between BA and hydroxyl radicals in the gas and aqueous phases, indicating that the ortho position was not susceptible for the following reactions.

)

(

Figure 4.7 Relative energies of three possible reaction paths (BZ aqueous phase).

Furthermore, the values of the open shell systems for all the reaction pathways in the gas and aqueous phases were summarized in Appendix A Table S4-S6 to analyze spin contamination. The spin contamination could be measured through analyzing the deviation of the values to the

104 exact values (0.75 for a doublet) (Menon & Radom 2008). It is deduced that the variation was within 10% for all the species, implying that spin contamination had insignificant influences on the reaction pathways (Young et al. 2001).

4.3.4 Reaction Rate Constants

The reaction rate constants for the elementary reaction was calculated by using transition state theory. Due to the formation of pre-reactive complex, the main reactant pathway is expressed as:

k 1 k2 C6H5COOH + OH Pre-reactive complex product + H2O Eq. 4.1 k-1 with

k1 k  k2  keq k2 Eq. 4.2 k1

Q Complex ((EC ER ) / RT ) keq  e Eq. 4.3 Q Q C6H5COOH OH

k T Q B TS ((ETS EC ) / RT ) k2   e Eq. 4.4 h QComplex and therefore

k T Q ((E E ) / RT ) k   B TS e TC R Eq. 4.5 h Q Q C6H5COOH OH

where 훤 is the tunneling factor, kB is Boltzmann 's constant, T is temperature, h is Planck 's constant,

QTS, QC6H5COOH, and QOH are molecular partition functions for transition states, BA and hydroxyl radicals, EC, ER, ETS is the energy including zero-point energy correction for pre-reactive complex, reactant, and transition state, respectively (Anglada 2004; Chen et al. 2015). The tunneling factor

105 was calculated by using Eq. 4.6 (Kazmierczak et al. 2017; Reisi-Vanani et al. 2015; Tiwari &

Mishra 2016):

1 h 2 Eq. 4-6   1 ( ) 24 k BT where ν is the imaginary frequency of transition state. The Wigner tunneling factor is one of the most commonly used tunneling corrections since it can be easily obtained and imaginary frequence is the only variable (Hoseinpour & Reisi-Vanani 2016; Zhou et al. 2016).

The reaction rate constants in the gas and aqueous phases are listed in Table 4.1. The rate constants in Table 4.1 were calculated from the simulation results by using the ωB97XD method since the rate constants by ωB97XD method had smaller deviation with experimental value than M06-2X method (Appendix A Table S7). As shown, the reaction rate constants of m-add, m2-add, p-add were higher than o-add and o2-add for the reactions between BA and hydroxyl radicals both in the gas and aqueous phases. There are three possible reasons to explain the phenomena, one is that the carboxyl group is an electron withdrawing group and it would decrease the electron density on the benzene ring through a resonance withdrawing effect at the ortho and para positions. Another reason is the steric influence at the ortho positions of the carboxyl group that impedes the approach of hydroxyl radicals. The formation of hydrogen bonds between hydroxyl radicals and hydrogen in the carboxyl group is another possible factor that would inhibit the ortho position reaction. For the gas phase reaction, the rate constant followed the order of H-abs>meta addition (m-add+m2- add)>para addition>ortho addition (o-add+o2-add). The reason why the rate constant for H-abs was highest is that benzene ring is capable of withdrawing electron density from carboxyl group by induction, making the hydrogen on the carboxyl group more active and reacting with hydroxyl

106 group to form water (DeRuiter 2005). The H-abs rate constant in the gas phase (2.66×10-11 cm3 molecule-1 s-1) was higher than in the aqueous phase (4.97×10-15 cm3 molecule-1 s-1), implying that the optimum reaction medium should be in gas phase or inside the cavitation bubbles when targeting the decarboxylation reaction. Furthermore, the tunneling factors of H-abs were much higher than all the addition reaction pathways both in the gas and aqueous phases.

Table 4.1 Reaction rate constants of BA and BZ with hydroxyl radical. Reaction path BA gas phase BA aqueous phase BZ aqueous phase k 훤 k 훤 k 훤 o-add 1.30×10-14 1.176 8.27×107 1.093 6.53×108 1.078 o2-add 3.17×10-14 1.185 1.15×108 1.094 6.53×108 1.078 m-add 2.01×10-13 1.163 4.70×108 1.059 2.55×109 1.051 m2-add 1.39×10-13 1.165 2.79×108 1.059 2.55×109 1.051 p-add 1.91×10-13 1.150 2.57×108 1.085 2.75×109 1.055 H-abs 2.66×10-11 3.122 2.99×106 3.183 Total 2.72×10-11 1.21×109 9.16×109 Note: Units for gas phase rate constants and aqueous phase rate constants are cm3 molecule-1 s-1, and M-1 s-1, respectively.

Except for H-abs, the rate constants of the other five reaction pathways in the aqueous phase were much higher than in the gas phase. Furthermore, the hydrogen abstraction rates of BA were much higher than for phenol (1.60×10-16 cm3 molecule-1 s-1) (Jayathilaka et al. 2014), since BA has an electronic withdrawing group, while phenol contains an electron donor group (Bellardita et al.

2012). San et al. calculated the reaction rate constants of BA with hydroxyl radical reaction in the aqueous phase, and the reaction rates were 8.38×10-11, 4.20×10-8, 1.65×10-20, 5.08×10-24 cm3 molecule-1 s-1 for meta addition reaction, ortho addition reaction, para addition reaction, and H-abs,

107 respectively (San et al. 2007). The reaction rates calculated were different due to the discovery of pre-reactive complexes as well as the different calculation levels. The overall rate constant of the six reaction pathways computed at 298 K, 1 atm in the aqueous medium was 1.21×109 M-1 s-1.

Assuming an uncertainty of 0.4 kcal mol-1 on the calculated energy levels, the expected uncertainty on the calculated rate constants was about a factor of 2, which was in the range of the previous reported experimental value (2.1±0.3×109 and 1.8×109 M-1 s-1) (Ashton et al. 1995; Dorfman et al.

1964), indicating the reliability of the calculations reported here.

The total rate constant for the reactions between BZ and hydroxyl radicals was approximately

9.16×109 M-1 s-1, which corresponds well with the experimental value of 5.9×109 M-1 s-1(Buxton et al. 1988). Furthermore, the overall rate constant for the reactions between BZ and hydroxyl radicals was higher than that between BA and hydroxyl radicals, implying that transferring BA to

BZ could promote the addition reaction. Therefore, the alkaline pH is more favorable for the degradation of BA in waste water. In addition, the rate constants for the addition reaction pathway followed the order of BZ aqueous phase>BA aqueous phase>BA gas phase (see Table 4.1 and

Appendix A Table S7). To further explain the trends, electrostatic potential analysis were carried out to display charge distributions three dimensionally by M06-2X method using 6-311+G(d,p) basis set, shown in Figure 4.8. It is worth noting that BZ in the aqueous phase had the biggest negative electrostatic potential, followed by BA in the aqueous phase, then by BA in the gas phase, consistent with the rate constants results. Hydroxyl radicals have strong electrophilic character and tend to react at negative regions, therefore, the electrostatic potential analysis results can help explain the lower energy barriers and higher rate constants for addition reactions in the aqueous phase than in the gas phase, further confirming the accuracy of the rate constants results.

108

(a) BA gas phase (b) BA aqueous phase (c) BZ aqueous phase

Figure 4.8 Electrostatic potential analysis.

After a single-point energy calculation at CCSD(T)/6-311++G(d,p) level of theory, the rate constants of all the reaction pathways were recalculated (Appendix A Table S8). Compared with rate constants obtained by the M06-2X method, it was deduced that the rate constants predicted by

CCSD(T) method have more obvious deviation from the experimental data. On the other hand,

CCSD(T) method is computationally more expensive than the M06-2X method. Therefore, the rate constants in this study were calculated without single-point energy correction.

4.3.5 Influence of Explicit Water Molecule

In the aqueous phase, water molecules form hydrogen bonds with the carboxylic group of BA and

BZ that might influence energy barriers. However, one of the significant drawbacks of using implicit SMD model is that it does not completely account for hydrogen bonding interactions

(Thapa & Schlegel 2016). To investigate the impact of solvation on transition states structures and rate constants, one explicit water molecule was incorporated around the carboxylic group of BA

109 and BZ. Due to the addition of water molecule to BZ, the structures of the reactants are not symmetric, therefore, there are five different reaction pathways between BZ and hydroxyl radicals.

The structures for the pre-reactive complexes, transition states, and products were optimized, and rate constants of the six reaction pathways between BA and hydroxyl radicals, and five reaction pathways between BZ and hydroxyl radicals, were calculated.

The transition states with the addition of one explicit water molecule for BA and BZ are depicted in Appendix A Fig. S7 and S8, respectively. It was deduced that the carbon-carbon lengths in the benzene ring of the transition states were essentially the same compared to the cases without one water molecule for both BA and BZ. The main difference was found in the variation in bond length between oxygen and the reacting carbon in benzene ring. Compared to the case without the explicit water molecule, bond lengths of the reacting carbons and oxygens were elongated for all the addition reactions between BA and hydroxyl radicals. On the other hand, they were longer for o- add and o2-add, and shorter for m-add, m2-add, and p-add for the addition reactions between BZ and hydroxyl radicals. As for the rate constants, listed in Appendix A Table S9, they were different from the case without a water molecule, which was possibly caused by electron redistribution on the benzene ring. The most substantial variation was the rate constant for H-abs, only accounting for 4.70% of that for implicit water molecule. Furthermore, the rate constant of H-abs was much smaller than for addition reaction. Therefore, the formation of hydrogen bonds between the carboxylic group and water molecules negatively influenced the rate constant for H-abs, and reaction medium should choose the gas phase or inside the bubbles for the H-abs pathway. In addition, it should be noted that the joint use of implicit solvation model and one explicit water molecule may not accurately reproduce the boundary conditions between the solute and bulk, and

110 it also requires the evaluation of the entropic effects with the explicit water molecule (Kamerlin et al. 2009).

4.4 Conclusions

An extensive theoretical study was conducted on the six possible reaction pathways of BA with hydroxyl radicals both in the gas phase and aqueous medium and BZ with hydroxyl radicals in the aqueous phase. The pre-reactive complexes identified affect the energy barriers. The rate constants calculated were 2.72×10-11 cm3 molecule-1 s-1 and 1.21×109 M-1 s-1 for the reactions between BA and hydroxyl radicals in the gas and aqueous phases, respectively, and 9.16×109 M-1 s-1 for the reactions between BZ and hydroxyl radicals in the aqueous phase. The rate constant of hydrogen abstraction was much higher in the gas phase than that in the liquid medium, however, the rate constants of addition reactions were much lower in the gas phase. Moreover, the energy barrier for hydrogen abstraction in the aqueous phase was higher than that in the gas phase. The incorporation of one water molecule was found to influence rate constants for both BA and BZ in the aqueous medium.

4.5 References

Anglada, J. M. (2004). Complex mechanism of the gas phase reaction between formic acid and hydroxyl radical. Proton coupled electron transfer versus radical hydrogen abstraction mechanisms. Journal of the American Chemical Society, 126(31), 9809-9820. Alvarez-Idaboy, J. R., Mora-Diez, N., Boyd, R. J., & Vivier-Bunge, A. (2001). On the importance of prereactive complexes in molecule-radical reactions: hydrogen abstraction from aldehydes by OH. Journal of the American Chemical Society, 123(9), 2018-2024.

111 Ashton, L., Buxton, G. V., & Stuart, C. R. (1995). Temperature dependence of the rate of reaction of OH with some aromatic compounds in aqueous solution. Evidence for the formation of a π- complex intermediate? Journal of the Chemical Society, Faraday Transactions, 91(11), 1631- 1633. Bartlett, R. J., & Musiał, M. (2007). Coupled-cluster theory in quantum chemistry. Reviews of Modern Physics, 79(1), 291. Bellardita, M., Augugliaro, V., Loddo, V., Megna, B., Palmisano, G., Palmisano, L., & Puma, M.

A. (2012). Selective oxidation of phenol and benzoic acid in water via home-prepared TiO2 photocatalysts: distribution of hydroxylation products. Applied Catalysis A: General, 441, 79- 89. Bhoorasingh, P. L., & West, R. H. (2015). Transition state geometry prediction using molecular group contributions. Physical Chemistry Chemical Physics, 17(48), 32173-32182. Bian, C., Li, Y., Wang, S., & Jing, X. (2017). Initial reaction mechanism between HO· and bisphenol‐F: Conformational dependence and the role of nonbond interactions. International Journal of Quantum Chemistry, 117(6). Buxton, G. V., Greenstock, C. L., Helman, W. P., & Ross, A. B. (1988). Critical review of rate constants for reactions of hydrated electrons, hydrogen atoms and hydroxyl radicals (⋅OH/⋅O− in aqueous solution. Journal of Physical and Chemical Reference Data, 17(2), 513-886. Chen, L., Wang, W., Wang, W., Li, C., Liu, F., & Lü, J. (2015). Kinetic and mechanistic investigations of the thermal decomposition of methyl-substituted cycloalkyl radicals. RSC Advances, 5(36), 28044-28053. Chuchev, K., & BelBruno, J. J. (2007). Mechanisms of decarboxylation of ortho-substituted benzoic acids. Journal of Molecular Structure: THEOCHEM, 807(1), 1-9. Dash, M. R., & Rajakumar, B. (2013). Experimental and theoretical rate coefficients for the gas phase reaction of β-Pinene with OH radical. Atmospheric environment, 79, 161-171. De Visscher, A. (2003). Kinetic model for the sonochemical degradation of monocyclic aromatic compounds in aqueous solution: new insights. Ultrasonics Sonochemistry, 10(3), 157-165. Degirmenci, I., & Coote, M. L. (2016). Comparison of thiyl, alkoxyl, and alkyl radical addition to double bonds: The unusual contrasting behavior of sulfur and oxygen radical chemistry. The Journal of Physical Chemistry A, 120(10), 1750-1755. Deruiter, J. (2005). Carboxylic acid structure and chemistry: part 1. Auburn university, Alabama.

112 Retrieved from http://www.auburn.edu/~deruija/pda1_acids1.pdf. Dorfman, L. M., Taub, I. A., & Harter, D. A. (1964). Rate constants for the reaction of the hydroxyl radical with aromatic molecules. The Journal of Chemical Physics, 41(9), 2954-2955. Enami, S., Hoffmann, M. R., & Colussi, A. J. (2016). Extensive H-atom abstraction from benzoate by OH-radicals at the air–water interface. Physical Chemistry Chemical Physics, 18(46), 31505-31512. Farkas, Ö., & Schlegel, H. B. (2003). Geometry optimization methods for modeling large molecules. Journal of Molecular Structure: THEOCHEM, 666, 31-39. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, Jr., J. A.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, Ö.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09, Revision E.01, Gaussian, Inc., Wallingford CT, 2013. Hoseinpour, S., & Reisi-Vanani, A. (2016). A mechanistic study of the decomposition and

reactivity of the C4F9OC2H4O• radical derived from HFE-7200 between 200 and 400 K. Progress in Reaction Kinetics and Mechanism, 41(3), 301-308. Huang, X., Li, X., Pan, B., Li, H., Zhang, Y., & Xie, B. (2015). Self-enhanced ozonation of benzoic acid at acidic pHs. Water Research, 73, 9-16. Li, J., & Brill, T. B. (2003). Spectroscopy of hydrothermal reactions 23: the effect of OH substitution on the rates and mechanisms of decarboxylation of benzoic acid. The Journal of Physical Chemistry A, 107(15), 2667-2673. Li, Y., Wen, B., Ma, W., Chen, C., & Zhao, J. (2012). Photocatalytic degradation of aromatic pollutants: a pivotal role of conduction band electron in distribution of hydroxylated intermediates. Environmental Science & Technology, 46(9), 5093-5099.

113 Lin, M., Ning, X. A., An, T., Zhang, J., Chen, C., Ke, Y., Wang, Y., Zhang, Y., Sun, J., & Liu, J. (2016). Degradation of polycyclic aromatic hydrocarbons (PAHs) in textile dyeing sludge with ultrasound and Fenton processes: Effect of system parameters and synergistic effect study. Journal of Hazardous Materials, 307, 7-16. Jayathilaka, P. B., Pathiraja, G. C., Bandara, A., Subasinghe, N. D., & Nanayakkara, N. (2014). Theoretical study of phenol and hydroxyl radical reaction mechanism in aqueous medium by the DFT/B3LYP/6-31+G(d, p)/CPCM model. Canadian Journal of Chemistry, 92(9), 809-813 Johnson, R. J., Smith, B. E., Sutton, P. A., McGenity, T. J., Rowland, S. J., & Whitby, C. (2011). Microbial biodegradation of aromatic alkanoic naphthenic acids is affected by the degree of alkyl side chain branching. The ISME Journal, 5(3), 486-496. Kamerlin, S. C., Haranczyk, M., & Warshel, A. (2009). Are mixed explicit/implicit solvation models reliable for studying phosphate hydrolysis? A comparative study of continuum, explicit and mixed solvation models. ChemPhysChem, 10(7), 1125-1134. Kazmierczak, L., Swiatla-Wojcik, D., & Wolszczak, M. (2017). Reaction of the hydrogen atom with nitrous oxide in aqueous solution–pulse radiolysis and theoretical study. RSC Advances, 7(15), 8800-8807. Medina, M. E., Iuga, C., & Álvarez-Idaboy, J. R. (2014). Antioxidant activity of fraxetin and its regeneration in aqueous media. A density functional theory study. RSC Advances, 4(95), 52920- 52932. Menon, A. S., & Radom, L. (2008). Consequences of spin contamination in unrestricted calculations on open-shell species: Effect of Hartree−Fock and Møller−Plesset contributions in hybrid and double-hybrid density functional theory approaches. The Journal of Physical Chemistry A, 112(50), 13225-13230. Minakata, D., Li, K., Westerhoff, P., & Crittenden, J. (2009). Development of a group contribution method to predict aqueous phase hydroxyl radical (HO•) reaction rate constants. Environmental Science & Technology, 43(16), 6220-6227. Minakata, D., Song, W., Mezyk, S. P., & Cooper, W. J. (2015). Experimental and theoretical studies on aqueous-phase reactivity of hydroxyl radicals with multiple carboxylated and hydroxylated benzene compounds. Physical Chemistry Chemical Physics, 17(17), 11796- 11812.

114 Nair, B. (2001). Final report on the safety assessment of benzyl alcohol, benzoic acid, and sodium benzoate. International Journal of Toxicology, 20, 23-50. Oliveira, R., Geraldo, D., & Bento, F. (2014). Electrogenerated HO radical reactions: the role of competing reactions on the degradation kinetics of hydroxy-containing aromatic compounds. Electrochimica Acta, 135, 19-26. Papp, T., Kollár, L., & Kégl, T. (2013). Employment of quantum chemical descriptors for Hammett constants: Revision suggested for the acetoxy substituent. Chemical Physics Letters, 588, 51-56. Perrone, O. M., Colombari, F. M., Rossi, J. S., Moretti, M. M. S., Bordignon, S. E., Nunes, C. D. C. C., Gomes, E., Boscolo, M., & Da-Silva, R. (2016). Ozonolysis combined with ultrasound as a pretreatment of sugarcane bagasse: effect on the enzymatic saccharification and the physical and chemical characteristics of the substrate. Bioresource Technology, 218, 69-76. Reisi-Vanani, A., Shahrokh, L., & Kokhdan, S. N. (2015). Theoretical study of the corannulene ozonolysis and evaluation of the various reaction paths. Computational and Theoretical Chemistry, 1051, 72-78. San, N., Kılıç, M., Tuiebakhova, Z., & Çınar, Z. (2007). Enhancement and modeling of the photocatalytic degradation of benzoic acid. Journal of Advanced Oxidation Technologies, 10(1), 43-50. Shu, Z., Li, C., Belosevic, M., Bolton, J. R., & El-Din, M. G. (2014). Application of a solar UV/chlorine advanced oxidation process to oil sands process-affected water remediation. Environmental Science & Technology, 48(16), 9692-9701. Singla, R., Ashokkumar, M., & Grieser, F. (2004). The mechanism of the sonochemical degradation of benzoic acid in aqueous solutions. Research on Chemical Intermediates, 30(7- 8), 723-733. Thapa, B., & Schlegel, H. B. (2016). Density functional theory calculation of pKa's of thiols in aqueous solution using explicit water molecules and the polarizable continuum model. The Journal of Physical Chemistry A, 120(28), 5726-5735. Tiwari, M. K., & Mishra, P. C. (2016). Catalytic role of iron-superoxide dismutase in hydrogen abstraction by super oxide radical anion from ascorbic acid. RSC Advances, 6(89), 86650- 86662.

115 Uc, V. H., Alvarez-Idaboy, J. R., Galano, A., García-Cruz, I., & Vivier-Bunge, A. (2006). Theoretical determination of the rate constant for OH hydrogen abstraction from toluene. The Journal of Physical Chemistry A, 110(33), 10155-10162. Wahab, H. S. (2012). Quantum chemical modeling study of adsorption of benzoic acid on anatase

TiO2 nanoparticles. Journal of Molecular Modeling, 18(6), 2709-2716. Yusof, N. S. M., Babgi, B., Alghamdi, Y., Aksu, M., Madhavan, J., & Ashokkumar, M. (2016). Physical and chemical effects of acoustic cavitation in selected ultrasonic cleaning applications. Ultrasonics Sonochemistry, 29, 568-576. Wander, R., Neta, P., & Dorfman, L. M. (1968). Pulse radiolysis studies. XII. Kinetics and spectra of the cyclohexadienyl radicals in aqueous benzoic acid solution. The Journal of Physical Chemistry, 72(8), 2946-2949. Wang, C., Klamerth, N., Huang, R., Elnakar, H., El-Din, & M. G. (2016a). Oxidation of oil sands process-affected water by potassium ferrate (VI). Environmental Science & Technology, 50(8), 4238-4247. Wang, M. F., Zuo, Z. J., Ren, R. P., Gao, Z. H., & Huang, W. (2016b). Theoretical study on catalytic pyrolysis of benzoic acid as a coal-based model compound. Energy & Fuels, 30(4), 2833-2840. Wu, P., Li, J., Li, S., Tao, & F. M. (2012). Theoretical study of mechanism and kinetics for the addition of hydroxyl radical to phenol. Science China Chemistry, 55(2), 270-276. Xie, H. B., Li, C., He, N., Wang, C., Zhang, S., & Chen, J. (2014). Atmospheric chemical reactions of monoethanolamine initiated by OH radical: Mechanistic and kinetic study. Environmental Science & Technology, 48(3), 1700-1706. Young, D. C. (2011). Computational chemistry: a practical guide for applying techniques to real world problems. Oxford: Wiley-Blackwell. Zhao, Q., Liu, F., Wang, W., Li, C., Lü, J., & Wang, W. (2017). Reactions between hydroxyl- substituted alkylperoxy radicals and Criegee intermediates: correlations of the electronic characteristics of methyl substituents and the reactivity. Physical Chemistry Chemical Physics, 19(23), 15073-15083. Zhou, H., Song, D., Zhong, C., & Ye, G. (2016). Theoretical and experimental study of light- assisted polymerization by multimechanism action. Scientific Reports, 6, 38473.

116 Chapter Five: Molecular Interactions between 1-Butyl-3-Methylimidazolium

Tetrafluoroborate and Model Naphthenic Acid: A DFT Study

5.1 Introduction

Naphthenic acids (NAs), a mixture of alkyl-substituted cyclic aliphatic carboxylic acids, are among the most common oxygen-containing compounds in crude oil (Quinlan & Tam 2015; Wu et al. 2017). The presence of NAs increases the acidity of crude oils, and causes serious corrosion to oil refining equipment, transportation pipelines, and storage tanks (Khan et al. 2017). NAs also induce other serious problems such as poisoning catalysts, forming coke, and emulsions (Colati et al., 2013). Therefore, it is important to remove NAs from crude oil, and NAs removal is critical for heavy oil upgrading (Zhang et al., 2006). Several methods are used to separate NAs from crude oil including aqueous solution washing, solvent extraction, adsorption, catalytic esterification, and decarboxylation (Shah et al., 2016a), but these methods have their drawbacks, making it critical to investigate alternative approaches to remove NAs from crude oil.

Ionic liquids (ILs) are nonvolatile, thermally stable, nonflammable, and environmentally friendly green solvents, and they are considered favorable solvents for many separation processes (El-

Nagar et al. 2017; Lin et al. 2017). Recently, significant progress has been made to utilize ILs to separate NAs from crude oil. Anderson et al. used tetraalkylammonium and tetraalkylphosphonium amino acid-based ILs to remove NAs from crude oil (Anderson et al.

2013). Sun et al. attempted to use imidazole anion-based ILs to separate NAs from crude oil (Sun

& Shi 2012). Shah et al. investigated the isolation of NAs from highly acidic model oil using

117 imidazolium-based ILs and hydroxide-based ILs (Shah et al. 2014; Shah et al. 2016b). In addition, the separation of NAs from crude oil was also realized through employing thiocyanate-based ILs

(Najmuddin et al. 2016).

Considering the structures and diversity of functionalities of ILs, most types of interactions, including dispersive, π‐π, n‐π, hydrogen bonding, dipolar, and ionic/charge‐charge, can occur between ILs and other compounds (Lü et al. 2017). Moreover, the composition of NAs is often complex and therefore, the interactions between ILs and NAs are different based on the structures of NAs. Although it is difficult and time-consuming to investigate IL-based extraction mechanisms for NAs through experiment, the interactions of ILs and compounds can be explored through quantum chemical computation (Babucci et al. 2016; Khan et al. 2014). Despite the numerous experimental investigations on the application of ILs to separate NAs from crude oil (Anderson et al. 2013; Najmuddin et al. 2016; Shah et al. 2014; Shah et al. 2016a; Sun & Shi 2012), few detailed computational studies have been reported on the direct interactions between ILs and NAs thus far.

Furthermore, current theoretical studies mainly focus on IL-based extraction of sulfur- and nitrogen-containing compounds from fuels (Ibrahim et al. 2015; Kianpour et al. 2016; Verdía et al. 2017). Due to structural differences between NAs and sulfur- and nitrogen-containing compounds, it is necessary to investigate the ILs extraction mechanisms of NAs to instruct the design of ILs for efficient NAs removal.

Here, we fill the gap by examining mechanistic interactions between ILs and six model NAs by using density functional theory. More specifically, we examine the extraction mechanisms between 1-butyl-3-methylimidazolium tetrafluoroborate ([BMIM][BF4]) as a model IL and

118 cyclohexanecarboxylic acid (CHCA), cyclopentanecarboxylic acid (CPCA), benzoic acid (BA), cyclohexanepentanoic acid (CHPA), 1,4-cyclohexanedicarboxylic acid (CHDCA), and dicyclohexylacetic acid (DCHA) as model NAs. BA is not strictly a NA, but the six compounds will be categorized collectively as NAs for convenience. The electronic structures, interaction energies, and the intermolecular interactions including van der Waals (vdW) interactions, hydrogen bonds, and electrostatic interactions, were analyzed. The structures for [BMIM][BF4] and six types of model NAs are shown in Figure 5.1.

(a) [BMIM][BF4] (b) CHCA (c) CPCA (d) BA

(e) CHPA (f) CHDCA (g) DCHA

Figure 5.1 Structures of [BMIM][BF4] and six types of model NAs compounds.

5.2 Computational Methods

The density functional computations were performed by using the Gaussian 09 program package

(Frisch et al. 2009). Many theoretical studies on ILs were carried out using M06-2X/6-311++g(d,p) level of theory (Allen et al. 2016; Liu et al. 2012; Marekha et al. 2015; Socha et al. 2014; Sun et

119 al. 2014). Therefore, the structures of [BMIM][BF4] and the six different model NAs were optimized by using the M06-2X method and 6-311++G(d,p) basis set. The [BMIM][BF4]-model

NAs structures were also optimized by using the same method and basis set. The interaction energies were calculated at M06-2X/6-311++G(d,p) level with the counterpoise method correction to the basis set superposition error (Kruse & Grimme 2012). The stabilization energy E(2) in the natural bond orbital (NBO) was investigated in Gaussian 09 at the M06-2X/6-311++G(d,p) level of theory. The highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) were calculated. Analyses of the noncovalent interactions (NCI) and electron density difference were carried out by analyzing the wave functions for the optimized structures using the Multiwfn software package (Lu & chen 2012a, 2012b). The coulomb-attenuating method

CAM-B3LYP, which considers long-range interactions, was used in combination with 6-

311++G(d,p) basis set to perform HOMO-LUMO overlap integral analysis (Yang et al. 2015).

Topological properties were analyzed using atoms in molecules (AIM) theory (Bader 1991).

5.3 Results and Discussion

5.3.1 Optimized Geometries

+ - - After optimization of [BMIM] and [BF4] , [BF4] was positioned at different orientations around

[BMIM]+, and those structures were optimized to determine the most stable cation-anion pair of

[BMIM][BF4]. Also, the structures of six types of model NAs were optimized, and electrostatic

- potential analysis was conducted; the results are displayed in Figure 5.2. It was observed that [BF4] has a strongly negative electrostatic potential, whereas [BMIM]+ has a high positive electrostatic potential. Regarding the six types of model NAs, the most negative electrostatic potentials are

120 located around the oxygen atoms in the carboxylic groups, and the most positive electrostatic potentials are located around the hydrogen atoms in the carboxylic groups. To obtain the most stable geometries for the interactions between [BMIM][BF4] and model NAs, model NAs were placed in different regions around the optimized [BMIM][BF4] to form complexes according to electrostatic potential analysis. Then, different structures were optimized, and the most stable geometry was the structure with the lowest energy and without imaginary frequency.

(a) [BMIM][BF4] (b) CHCA

(e) CHPA (f)CHDCA

121

(g) DCHA

Figure 5.2 The electrostatic potential (a.u.) of [BMIM][BF4] and model NAs.

The most stable molecular structures for [BMIM][BF4] and [BMIM][BF4]-model NAs are shown in Figure 5.3 with cartesian coordinates summarized in Appendix B Table S1. It was noted that the most stable geometry for [BMIM][BF4] calculated with the 6-311++G(d,p) basis set (Figure

5.3a) varies from the geometry calculated with the 6-311G(d,p) basis set (Figure 5.3b).

Additionally, the energy deviation between the two structures is 9.95 kcal/mol. With the addition of two diffusion functions, the 6-311++G(d,p) basis set can better represent the electron distributions (Wazzan et al. 2016), hence, it was chosen for the current study. The vdW radii for fluorine and hydrogen are 1.2 and 1.47 Å, respectively (Bondi 1964). Therefore, several hydrogen bonds are formed between the hydrogen atoms in the cation and fluorine atoms in the anion of

[BMIM][BF4], shown in Figure 5.3a. The hydrogen bond distances between H13 and F30, H13 and F26 are shorter than the hydrogen bond lengths formed by other hydrogen atoms (H18, H19) owing to the high electron potential of H13. H18 is closer to the imidazolium ring than H19, consequently, the H18···F26 hydrogen bond is shorter than H19···F26. In addition, the interaction between F28 and C5 in the imidazolium ring indicates the existence of lone pair (LP)-π interaction in [BMIM][BF4].

122

(a) BMIM][BF4] (6-311++G(d,p)) (b) [BMIM][BF4] (6-311G(d,p))

(e) [BMIM][BF4]-BA (f) [BMIM][BF4]-CHPA

(g) [BMIM][BF4]-CHDCA (h) [BMIM][BF4]-DCHA

Figure 5.3 The optimized structures of [BMIM][BF4] and [BMIM][BF4]-NAs complexes.

123 After forming complexes with model NAs, the cation and anion in [BMIM][BF4] form two

- prominent hydrogen bonds with model NAs: one is between the fluorine atom of [BF4] and the hydrogen atom in the carboxylic group of the model NAs, and the other is between the hydrogen atom in [BMIM]+ and the oxygen atom in the carboxylic group of the model NAs. The F51···H21 bond is shorter than the O19···H34 one, suggesting that the hydrogen bond between the fluorine atom and hydrogen atom is stronger, displayed in Figure 5.3c. It is noted that the F51···H34 bond length is longer after forming complexes because F51···H21 and H34···O19 hydrogen bonds formation result in an increase of the F51 electrostatic potential and a decrease of the H34 electrostatic potential. Hence, the F51···H34 bond length is longer and the interaction is weaker

(see Figure 5.3c and Figure 5.4), in agreement with the principle of bond order conservation.

Meanwhile, the formation of the F51···H21 and H34···O19 hydrogen bonds causes the F49···C26,

- F49···H40, and F47···H39 bonds to elongate, suggesting that the interactions between [BF4] and

[BMIM]+ are weaker (see Figure 5.3c). Interestingly, O19 and H21 in the carboxylic group form an intermolecular hydrogen bond in [BMIM][BF4]-CHCA. The interaction between [BMIM][BF4] and other five types of NAs follow the same trends as [BMIM][BF4]-CHCA. For the electrostatic potential, the highest negative values are located around the interaction region between

[BMIM][BF4] and model NAs, whereas the largest positive values are located around the hydrogen atoms in the imidazolium ring, as shown in Figure 5.4.

(a) [BMIM][BF4]-CHCA (b) [BMIM][BF4]-CPCA 124

(e) [BMIM][BF4]-CHDCA (f) [BMIM][BF4]-DCHA

Figure 5.4 The electrostatic potential (a.u.) of [BMIM][BF4]-NAs complexes.

5.3.2 Interaction Energies

Interaction energies between [BMIM][BF4] and model NAs can help to evaluate the stability of complexes, and moreover help to explain the different extraction efficiencies for NA extraction.

The interaction energies were calculated according to Eq. 5-1:

Eq. 5-1 ∆퐸 = (퐸[퐵푀퐼푀]][BF4] + 퐸푀표푑푒푙 푁퐴푠) − 퐸[퐵푀퐼푀][퐵퐹4]−푀표푑푒푙 푁퐴푠

The interaction energies between [BMIM][BF4] and CHCA/CPCA/BA/CHPA/CHDCA/DCHA are 15.22, 15.07, 15.63, 14.34, 16.36, and 16.57 kcal/mol, respectively. The interaction energy for

[BMIM][BF4]-DCHA is highest with [BMIM][BF4]-CHDCA ranking second among the interactions between [BMIM][BF4] and six types of model NAs, therefore, it was deduced that the interaction energy between [BMIM][BF4] and NAs with two cyclohexane rings is larger than that with one, and the interaction energy for NAs with two carboxylic groups is larger than that with

125 one carboxylic group. In addition, it is expected that the second carboxylic group in CHDCA is capable of forming complexes with a second [BMIM][BF4] molecule, which will result in higher interaction energy. Accordingly, [BMIM][BF4] is more favorable to extract NAs with polycyclic hydrocarbons or multiple carboxylic groups.

5.3.3 NBO Analyses

To obtain donor-acceptor interactions between [BMIM][BF4] and model NAs, NBO analysis and a second-order perturbation theory analysis were carried out. The results are listed in Appendix B

Table S2. According to the results, there are four kinds of donor-acceptor interactions for stable hydrogen bonds between CHCA/CPCA/CHPA/CHDCA/DCHA and [BMIM]+: (C-O)-*(C-H),

LP(O)-*(C-H), *(C-O)-*(C-H), and (C-H)-*(C-O). In addition, (C-O)-*(C-H), LP(O)-

*(C-H), *(C-O)-*(C-H), and (C-H)-*(C-O) interactions exist between BA and [BMIM]+.

- With regards to the interactions between CHCA/CPCA/ CHPA/CHDCA/DCHA and [BF4] , there are two main kinds of hydrogen bonding interactions: (O-H)-*(B-F) and LP(F)- *(O-H). On

- the other hand, only the LP(F)-*(O-H) interaction takes place between BA and [BF4] . The remarkable difference can be attributed to the fact that BA has an aromatic ring and the other five types of NAs are nonaromatic. Interestingly, the stabilization energy E(2) of LP(F)-*(O-H) and

LP(O)-*(C-H) is substantially greater than that of the other interactions, exhibiting that LP(F)-

*(O-H) and LP(O)-*(C-H) interactions are stronger. In addition, the stabilization energy E(2) for LP(F)-*(O-H) is larger than that of LP(O)-*(C-H). NBO analysis results are consistent with the findings from the geometry analysis.

126 The charge distributions of [BMIM][BF4], model NAs, and [BMIM][BF4]-NAs were summarized in Appendix B Table S3. For isolated model NAs, the hydrogen atom with the highest NBO charge is H21(0.48598), H18(0.48847), H15(0.49089), H33(0.48608), H24(0.48743), and H40(0.48313) in CHCA, CPCA, BA, CHPA, CHDCA, and DCHA, respectively, which is attributed to the electronegativity of oxygen atom in the carboxylic group. After being absorbed by [BMIM][BF4],

LP(F)-*(O-H) interactions between [BMIM][BF4] and CHCA/CPCA/BA/CHPA/DCHA, lead to

- a more negative charge of the fluorine atom in [BF4] and more positive charges for hydrogen atoms in model NAs. Similarly, LP(O)-*(C-H) interactions give rise to the decrease of charge for oxygen atom in the carboxylic group of model NAs and increase of charges for interreacting hydrogen atoms in [BMIM]+. Consequently, it is deduced that the complex formation, especially the strong hydrogen bonds formation, results in the charge redistribution.

The ,  occupancy and linear combination of some natural atomic orbitals of [BMIM]+ in isolated

[BMIM][BF4], and [BMIM][BF4]-NAs were exhibited in Appendix B Table S4 to investigate the influence of interactions on the orbital formation of the cation. The hybridization of N1, C2, C3, and N4 atoms in (N1-C2), (N1-C5), (C2-C3), (C3-N4), and (N4-C5) are essentially the same after adsorbing model NAs, whereas the hybridization of C5 atom in the imidazolium ring shows a clear deviation. As for the hybridization of (C5-H13), (C6-H14), and (C7-H18), it is noted that there are huge differences for C5(C26) and C6(C27) atoms after [BMIM][BF4] interacts with CHCA, changing from sp1.60 to sp1.53, and sp2.88 to sp2.79, respectively, whereas the hybridization of C7(C28) atom roughly remains the same. According to the geometry and donor- acceptor interactions analysis, LP(O19)-*(C26-H34)/*(C27-H35), (C18-O19)-*(C26-

H34)/*(C27-H35), *(C18-O19)-*(C26-H34)/*(C27-H35), and (C6-C18)-*(C26-H34)

127 interactions could account for the hybridization changes of C26(C5) and C27(C6) atoms for

[BMIM][BF4]-CHCA. On the other hand, no strong interactions are formed for the C28(C7) atom, therefore, its hybridization practically maintains the same. The hybridization changes of (C5-

H13), (C6-H14), and (C7-H18) for [BMIM][BF4]-CPCA/BA follow the same trends as

[BMIM][BF4]-CHCA. However, C5 and C7 atoms in [BMIM][BF4]-CHPA/CHDCA go through notable hybridization variation while hybridization of C6 atom virtually keeps unchanged. As for

[BMIM][BF4]-DCHA, only the C5 atom experiences significant hybridization alteration, compatible with geometry and donor-acceptor interactions analysis.

5.3.4 The Topological Properties of Interactions

Atoms in molecules (AIM) analysis was performed to investigate the topological properties of interactions (Bader 1991). The total electronic density (r) and Laplacian of electron density

2(r) at the bond critical point (BCP) are useful to illustrate the characteristics of the chemical bonds. The results are displayed in Figure 5.5 and listed in Table 5.1. The results show that

[BMIM][BF4]-CHCA, [BMIM][BF4]-CPCA, and [BMIM][BF4]-BA have 10 BCPs, whereas

[BMIM][BF4]-CHPA, [BMIM][BF4]-CHDCA, [BMIM][BF4]-DCHA have 15, 17, and 19 BCPs, respectively. The greater the alkyl chain length, number of carboxylic groups, and number of cyclohexane rings, the greater the opportunity for hydrogen-hydrogen (vdW) interactions as well as hydrogen bonding formation between model NAs and [BMIM]+, resulting in more BCPs.

128

(a) [BMIM][BF4] (b) [BMIM][BF4]-CHCA

(e) [BMIM][BF4]-CHPA (f) [BMIM][BF4]-CHDCA

129

(g) [BMIM][BF4]-DCHA

Figure 5.5 BCPs and bond paths in [BMIM][BF4] and model NAs.

The electron density (r) of BCP represents the strength of interactions; larger value means a stronger interaction (Bader 1998; Chęcińska et al. 2003). According to Table 5.1, the F51···H21 interaction has the largest (r) and the O19···H34 bond ranks second for [BMIM][BF4]-CHCA, implying that the F51···H21 and O19···H34 interactions are stronger than other interactions. Due to the larger electronegativity of fluorine than oxygen, F···H interaction is stronger than O···H interaction. Moreover, [BMIM][BF4]-CPCA, [BMIM][BF4]-BA, [BMIM][BF4]-CHDCA, and

[BMIM][BF4]-DCHA also follow the same trends as [BMIM][BF4]-CHCA. In summary, (r) is greatest for interactions between the hydrogen atom (in the carboxylic group) and the fluorine atom (closest to that hydrogen atom), and is second largest for the interactions between the oxygen atom (connecting with carbon through double bond in the carboxylic group) and the hydrogen atom in [BMIM]+ (closest to that oxygen atom). Nevertheless, the strongest interaction for

[BMIM][BF4]-CHPA is the H54···H22 bond. Therefore, it is concluded that hydrogen bonding is the main interaction for NAs without long alkyl chains, whereas vdW and hydrogen bonds

130 interactions are the main interactions for NAs with long alkyl chains. The hydrogen bonding extraction mechanism was confirmed by Shah et al. in their experimental study (Shah et al. 2014).

It is also correlated to the mechanism proposed by Anderson et al. (Anderson et al. 2013). The values of the Laplacian of the electron density are positive for interactions between [BMIM][BF4] and the six model NAs. This means that the electrons tend to segregate. The positive values also demonstrate that ionic bonds, hydrogen bonds, and vdW interactions exist in the six complexes formed between [BMIM][BF4] and model NAs (Casassa et al. 2015).

Table 5.1 Electron densities () and Laplacian of electron density (2) of BCPs in [BMIM][BF4] and [BMIM][BF4]-NAs complexes. CP label X···Y  (a.u.) 2 (a.u.)

(a) [BMIM][BF4] 35 F30···C6 0.1095E-01 0.4792E-01 40 F30···H13 0.1507E-01 0.6238E-01 43 F28···H19 0.8634E-02 0.3154E-01 45 F26···H19 0.1076E-01 0.4143E-01 46 F26···H13 0.1444E-01 0.6007E-01 47 F27···C5 0.1432E-01 0.5671E-01 53 F26···H18 0.1180E-01 0.4713E-01

(b) [BMIM][BF4]-CHCA 58 F51···H37 0.9475E-02 0.4102E-01 63 O19···H35 0.9788E-02 0.3213E-01 65 F49···C26 0.1159E-01 0.4781E-01 69 F51···C26 0.1005E-01 0.4198E-01 72 F51···H21 0.3651E-01 0.1528E+00 78 F49···H40 0.9481E-02 0.3439E-01 84 O19···H34 0.1857E-01 0.7570E-01 86 F47···C26 0.1245E-01 0.5105E-01 93 F47···H40 0.9369E-02 0.3651E-01 100 F47···H39 0.1093E-01 0.4413E-01

131 CP label X···Y  (a.u.) 2 (a.u.) 64 O46···H14 0.9306E-02 0.3004E-01 67 F30···C5 0.9844E-02 0.4089E-01 72 F28···H19 0.9010E-02 0.3272E-01 81 F26···C5 0.1273E-01 0.5199E-01 82 F30···H48 0.3703E-01 0.1544E+00 83 O46···H13 0.1879E-01 0.7666E-01 92 F26···H19 0.9255E-02 0.3599E-01 97 F26···H18 0.1074E-01 0.4358E-01

(d) [BMIM][BF4]-BA 61 F26···H19 0.9488E-02 0.3683E-01 65 F26···H18 0.1095E-01 0.4433E-01 78 F30···H45 0.3803E-01 0.1574E-01 80 F28···H19 0.9000E-02 0.3274E-01 84 F26···C5 0.1261E-01 0.5179E-01 93 O43···H13 0.1842E-01 0.7497E-01 94 F30···C5 0.1028E-01 0.4290E-01 95 F28···C5 0.1194E-01 0.3651E-01 100 F30···H16 0.9735E-02 0.4182E-01 101 O43···H14 0.9892E-02 0.3269E-01

(e) [BMIM][BF4]-CHPA 65 F30···H63 0.2655E-01 0.1104E+00 74 F26···C60 0.1134E-01 0.4830E-01 80 F30···C6 0.8701E-02 0.3843E-01 85 F30···H13 0.1388E-01 0.5633E-01 90 H59···H22 0.5359E-02 0.1577E-01 97 O61···H13 0.1861E-01 0.7702E-01 102 F26···H13 0.1010E-01 0.4101E-01 105 H54···H22 0.4101E-01 0.1680E-01 106 O61···H18 0.1241E-01 0.4106E-01 107 F28···N4 0.1306E-01 0.5395E-01 113 F26···H18 0.9531E-02 0.4037E-01 115 H56···H22 0.6331E-02 0.1952E-01 117 F26···H19 0.1161E-01 0.4555E-01 121 H56···H18 0.6853E-02 0.1955E-01 137 H51···H23 0.2406E-02 0.8204E-02 (f) [BMIM][BF4]-CHDCA 56 F30···H50 0.3074E-01 0.1281E+00 62 F30···C6 0.1070E-01 0.4659E-01

132 CP label X···Y  (a.u.) 2 (a.u.) 78 F26···O48 0.9720E-02 0.4184E-01 79 F30···H13 0.1242E-01 0.5048E-01 87 F26···H46 0.7176E-02 0.2978E-01 89 F28···N4 0.1241E-01 0.4960E-01 92 O48···H13 0.1837E-01 0.7461E-01 96 C5···F26 0.1169E-01 0.4640E-01 104 F26···H19 0.1274E-01 0.4993E-01 107 O48···H18 0.1235E-01 0.4264E-01 113 H46···H19 0.6175E-02 0.1825E-01 114 H39···H24 0.4357E-02 0.1403E-01 117 H46···H24 0.5538E-02 0.1669E-01 125 H46···H22 0.6611E-02 0.1938E-01 133 H22···H46 0.5357E-02 0.1347E-01 137 H40···H22 0.7454E-02 0.2083E-01 140 H40···C10 0.5644E-02 0.1963E-01 (g) [BMIM][BF4]-DCHA 74 F70···H56 0.8919E-02 0.3827E-01 81 F70···H40 0.3188E-01 0.1336E+00 84 F70···C45 0.1041E-01 0.4311E-01 88 C45···F68 0.1360E-01 0.5532E-01 101 O38···H53 0.2459E-01 0.1033E+00 103 C45···F66 0.1102E-01 0.4568E-01 108 O38···F66 0.4818E-02 0.2029E-01 110 H59···F68 0.7976E-02 0.2950E-01 122 F66···H59 0.1141E-01 0.4467E-01 128 F66···H58 0.1136E-01 0.4618E-01 129 F66···H11 0.6781E-02 0.2493E-01 134 O38···H8 0.1074E-01 0.3522E-01 143 H11···H58 0.4038E-02 0.1074E-01 147 H8···H58 0.2654E-02 0.7770E-02 157 H11···H62 0.3983E-02 0.1181E-01 163 H26···H9 0.8155E-02 0.2848E-01 167 H19···H58 0.4828E-02 0.1367E-01 174 H16···H62 0.5390E-02 0.1628E-01 180 H19···H61 0.6187E-02 0.1890E-01

5.3.5 NCI Analyses

133 NCI analyses can provide information about intramolecular and intermolecular interactions to distinguish interaction types and strength (Johnson et al. 2010). To investigate interaction types and strength, the plots of reduced density gradient (RDG) versus sign(2), and the gradient isosurface (s=0.7 a.u.) for [BMIM][BF4] and [BMIM][BF4]-model NAs are displayed in Figure

5.6. In the left figure, the peaks at the sign(2)<0 region suggest attractive interactions such as hydrogen bond interactions and π-π interactions, whereas the spikes close to sign(2)=0 exhibit vdW interactions. On the other hand, peaks at the sign(2)>0 region indicate steric effects. In the gradient isosurface figure, the interaction types and strength can be identified through analyzing the color and area.

(a) [BMIM][BF4]

(b) [BMIM][BF4]-CHCA

134

(d) [BMIM][BF4]-BA

(e) [BMIM][BF4]-CHPA

135

(f) [BMIM][BF4]-CHDCA

(g) [BMIM][BF4]-DCHA

Figure 5.6 The sign(λ2)ρ vs RDG (left) and the gradient isosurfaces (right) for [BMIM][BF4] and [BMIM][BF4]-NAs.

Note: red indicated sign(λ2)ρ>0, blue indicated sign(λ2)ρ<0.

As shown in Figure 5.6a, there is no peak in the hydrogen bond region of [BMIM][BF4], demonstrating no apparent hydrogen bonds in [BMIM][BF4]. As for complexes formed between

[BMIM][BF4] and model NAs, the spikes at -0.04 a.u. as well as the dark blue circle between the hydrogen atom and fluorine atom in the gradient isosurface figure indicate the existence of strong hydrogen bonding, as shown in Figure 5.6b, c, and d. The spikes at -0.04 a.u. correspond to the strong fluorine and hydrogen interactions according to electron density analysis (Table 5.1). On

136 the contrary, there is no spike at -0.04 a.u. in Figure 5.6e, f, and g, suggesting that the hydrogen bonds at -0.04 a.u between [BMIM] and CHPA/CHDCA/DCHA are weaker. Moreover, the existence of peaks at -0.02 a.u. relates to hydrogen bonds formed between the oxygen atom in the model NAs and hydrogen atom in [BMIM][BF4]. It is noteworthy that an oxygen atom can form several hydrogen bonds with hydrogen atoms in the ILs. The F-H hydrogen bond is stronger than

O-H hydrogen bonds, hence, the hydrogen bonds formed between the anion and model NAs is stronger than that formed between the cation and model NAs.

The area and color between the anion and cation in the [BMIM][BF4] are essentially the same after forming interactions with model NAs. It was deduced that the intramolecular LP-π interactions formed between the anion and cation in [BMIM][BF4] are not destroyed by the interactions with model NAs. Furthermore, due to the long alkyl chain of CHPA, the hydrogen atoms on the alkyl chain can form vdW interactions with [BMIM][BF4], as indicated by the peaks around 0.00 a.u. in

Figure 5.6e.

5.3.6 The HOMO-LUMO Overlap Integral Analyses

It is reported that overlap of the molecular orbitals occurs when ILs interact with compounds, and orbital overlap analysis is beneficial to understand the interactions (Ogunlaja et al. 2014; Youngs et al. 2008). In this section, we explored the HOMO and LUMO orbitals of [BMIM][BF4] and model NAs, with results displayed in Figure 5.7, as well as the HOMO-LUMO and LUMO-

HOMO integrals between [BMIM][BF4] and model NAs, with results shown in Appendix B Fig.

137 S1. A positive overlap integral indicates the attraction of two molecular orbitals, whereas a negative value means repulsion of molecular orbitals.

[BMIM][BF4] HOMO [BMIM][BF4] LUMO

CHCA HOMO CHCA LUMO

CPCA HOMO CPCA LUMO

138 BA HOMO BA LUMO

CHPA HOMO CHPA LUMO

CHDCA HOMO CHDCA LUMO

DCHA HOMO DCHA LUMO

Figure 5.7 HOMO and LUMO analyses of [BMIM][BF4] and model NAs. Note: isosurface for CHDCA LUMO is at 0.021 a.u., CHPA LUMO is at 0.022 a.u., and the others is 0.02 a.u.; green represents negative values, dark red represents positive values.

The results reveal that HOMO of CHCA, CPCA, CHPA, CHDCA, and DCHA have a four-leaf clover shape orbital near the carboxylic group. In addition, the LUMO of CHCA, CPCA, CHPA,

CHDCA, and DCHA occupy a large space near the hydrogen atom of the carboxylic group. On the other hand, neither the four-leaf clover shape nor large area are found for the HOMO and

LUMO of BA, as shown in Figure 5.7. The overlap integral between HOMO of [BMIM][BF4] and

LUMO of CHPA/CHDCA is negative, while the overlap integral between LUMO of

[BMIM][BF4] and HOMO of CHPA/CHDCA is positive, suggesting that LUMO-HOMO

139 interactions are beneficial to [BMIM][BF4]-CHPA/CHDCA complex formation. Furthermore, the overlap integral between HOMO of [BMIM][BF4] and LUMO of CHCA/CPCA, and between

LUMO of [BMIM][BF4] and HOMO of CHCA/CPCA are both positive, indicating that both

HOMO-LUMO and LUMO-HOMO interactions are beneficial to [BMIM][BF4] and

CHCA/CPCA interaction. On the other hand, the overlap integral between HOMO of

[BMIM][BF4] and LUMO of BA/DCHA are positive, and the overlap integral between LUMO of

[BMIM][BF4] and HOMO of BA/DCHA are negative, demonstrating that HOMO-LUMO is conducive to the interactions between [BMIM][BF4] and BA/DCHA.

5.3.7 Electron Density Difference Analyses

The formation of complexes between ILs and compounds generates electron density redistribution

(Noack et al. 2010). To understand the electron redistribution caused by the interaction between

[BMIM][BF4] and the six model NAs, electron density distribution maps were plotted shown in

Appendix B Fig. S2. The electron density change was calculated according to Eq. 5-2:

Eq. 5-2 ∆휌 = 휌[퐵푀퐼푀][퐵퐹4]−푀표푑푒푙 푁퐴푠 − (휌[퐵푀퐼푀][퐵퐹4) + 휌푀표푑푒푙 푁퐴푠)

According to the results in Appendix B Fig. S2, the difference of the electron density is located in the interaction region between IL and model NAs. The results reveal that the electron density of the carbon atom in the carboxylic group of model NAs is decreased. The reason is because of electron transfer from the carbon atom to the two adjacent oxygen atoms. Furthermore, the electron density of the oxygen atom connected with the carbon atom through double bond in the carboxylic group, is denoted purple in the exterior and green in the interior for all the six complexes, indicating the increase of electron density in the exterior of the complex and decrease in the interior of the

140 complex. The explanation is due to the formation of hydrogen bonds between the oxygen atom and the hydrogen in [BMIM][BF4] molecule leading to electron transfer from the interior of the complex to the exterior. Interestingly, the electron density of the oxygen atom connected with the hydrogen atom in the carboxylic group transfers from the center to the edges, induced by the redistribution of electrons during complex formation. Moreover, electron transfers from the hydrogen atom in the carboxylic group to the adjacent oxygen atom, giving rise to the decrease of electron density in the hydrogen atom.

Due to the hydrogen bonding formation between [BMIM][BF4] and model NAs, there are significant decreases of the electron density for hydrogen atoms both in [BMIM][BF4] and model

NAs. On the other hand, the electron transfer from those hydrogen atoms in the cation to the adjacent carbons causes the rise of the electron density between the carbon and hydrogen atoms.

After hydrogen bond formation between one of the fluorine atoms in [BMIM][BF4] and a hydrogen atom in model NAs, the electron density of the participating fluorine atom is raised, whereas the electron density between the fluorine atom and the boron atom is reduced, since electrons in that region are transferred to the neighboring fluorine atom, consistent with the principle of bond order conservation.

Overall, the electron density changes remarkably near the anion and carboxylic group for all six interactions, which is caused by the stronger hydrogen bonds of F···H and O···H. Compared with

[BMIM][BF4]-CHCA, [BMIM][BF4]-CPCA, and [BMIM][BF4]-BA, the hydrogen atoms in

[BMIM][BF4]-CHPA, [BMIM][BF4]-CHDCA, and [BMIM][BF4]-DCHA have pronounced

141 electron density changes due to vdW interactions of hydrogen atoms. The results are consistent with the AIM analysis.

5.4 Conclusions

The interactions between [BMIM][BF4] and CHCA/CPCA/BA/CHPA/CHDCA/DCHA were investigated by using the density functional theory approach. For NAs without a long alkyl chain, the main extraction mechanism is hydrogen bonding; For NAs with long alkyl chain, vdW interaction and hydrogen bonding are the dominant extraction mechanisms. [BMIM][BF4] is more efficient at extracting NAs with polycyclic hydrocarbons or multiple carboxylic groups. The bond characteristics and HOMO, LUMO orbitals for BA are different from the model NAs without

+ - aromatic ring. The LP-π interactions between [BMIM] and [BF4] are not destroyed by the adsorption of model NAs on [BMIM][BF4], whereas the electron density changes remarkably near

- [BF4] and carboxylic group upon complexes formation. The study on the interactions between

[BMIM][BF4] and six model NAs provides a comprehensive understanding of extraction mechanism by ILs, and will be beneficial for ILs design.

5.5 References

Allen, C., Ghebreab, R., Doherty, B., Li, B., & Acevedo, O. (2016). Examining ionic liquid eff ects on mononuclear rearrangement of heterocycles using QM/MM simulations. The Journal of Physical Chemistry B. Anderson, K., Goodrich, P., Hardacre, C., Hussain, A., Rooney, D. W., & Wassell, D. (2013). Removal of naphthenic acids from crude oil using amino acid ionic liquids. Fuel, 108, 715– 722.

142 Babucci, M., Balci, V., Akçay, A., & Uzun, A. (2016). Interactions of [BMIM][BF4 ] with metal oxides and their consequences on stability limits. The Journal of Physical Chemistry C, 120(36), 20089–20102. Bader, R. F. W. (1991). A quantum theory of molecular structure and its applications. Chemical Reviews, 91(5), 893–928. Bader, R. F. W. (1998). A bond path: a universal indicator of bonded interactions. The Journal of Physical Chemistry A, 102(37), 7314–7323. Bondi, A. (1964). van der Waals volumes and radii. The Journal of Physical Chemistry, 68(3), 441–451. Casassa, S., Erba, A., Baima, J., & Orlando, R. (2015). Electron density analysis of large (molecular and periodic) systems: A parallel implementation. Journal of Computational Chemistry, 36(26), 1940–1946. Checińska, L., Grabowski, S. J., & Małecka, M. (2003). An analysis of bifurcated H-bonds: Crystal and molecular structures of O,O-diphenyl 1-(3-phenylthioureido) pentanephosphonate and O,O-diphenyl 1-(3-phenylthioureido)butanephosphonate. Journal of Physical Organic Chemistry, 16(4), 213–219. Colati, K. A. P., Dalmaschio, G. P., De Castro, E. V. R., Gomes, A. O., Vaz, B. G., & Romão, W. (2013). Monitoring the liquid/liquid extraction of naphthenic acids in brazilian crude oil using electrospray ionization FT-ICR mass spectrometry (ESI FT-ICR MS). Fuel, 108, 647–655. El-Nagar, R. A., Nessim, M., Abd El-Wahab, A., Ibrahim, R., & Faramawy, S. (2017). Investigating the efficiency of newly prepared imidazolium ionic liquids for carbon dioxide removal from natural gas. Journal of Molecular Liquids, 237, 484–489. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, Jr., J. A.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin,

143 R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, Ö.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09, Revision E.01, Gaussian, Inc., Wallingford CT, 2013. Ibrahim, M. H., Hayyan, M., Hashim, M. A., & Hayyan, A. (2017). The role of ionic liquids in desulfurization of fuels: A review. Renewable and Sustainable Energy Reviews, 76, 1534–1549. Johnson, E. R., Keinan, S., Mori-Sánchez, P., Contreras-García, J., Cohen, A. J., & Yang, W. (2010). Revealing noncovalent interactions. Journal of the American Chemical Society, 132(18), 6498–6506. Khan, I., Kurnia, K. A., Mutelet, F., Pinho, S. P., & Coutinho, J. A. P. (2014). Probing the interactions between ionic liquids and water : experimental and quantum chemical approach. The Journal of Physical Chemistry. B, 118, 1848–1860. Khan, M. K., Riaz, A., Yi, M., & Kim, J. (2017). Removal of naphthenic acids from high acid crude via esterification with methanol. Fuel Processing Technology, 165, 123–130. Kianpour, E., Azizian, S., Yarie, M., Zolfigol, M. A., & Bayat, M. (2016). A task-specific phosphonium ionic liquid as an efficient extractant for green desulfurization of liquid fuel: An experimental and computational study. Chemical Engineering Journal, 295, 500–508. Kruse, H., & Grimme, S. (2012). A geometrical correction for the inter- and intra-molecular basis set superposition error in Hartree-Fock and density functional theory calculations for large systems. The Journal of Chemical Physics, 136, 154101. Lin, J., Lü, R., Wu, C., Xiao, Y., Liang, F., & Famakinwa, T. (2017). A density functional theory study on the interactions between dibenzothiophene and tetrafluoroborate-based ionic liquids. Journal of Molecular Modeling, 23(4). Liu, A., Ma, R., Song, C., Yang, Z., Yu, A., Cai, Y., He, L., Zhao, Y., Yu, B., & Song, Q. (2012).

Equimolar CO2 capture by N-substituted amino acid salts and subsequent conversion. Angewandte Chemie International Edition, 11306–11310. Lü, R., Wu, C., Lin, J., Xiao, Y., Wang, F., & Lu, Y. (2017). The study on interactions between 1‐ ethyl‐3‐methylimidazolium chloride and benzene/pyridine/pyrrole/thiophene. Journal of Physical Organic Chemistry, 30(8), e3663. Lu, T., Chen, F. (2012a). Multiwfn: A multifunctional wavefunction analyzer. Journal of Computational Chemistry, 33(5), 580–592. Lu, T., Chen, F. (2012b). Quantitative analysis of molecular surface based on improved Marching

144 Tetrahedra algorithm. Journal of Molecular Graphics and Modelling, 38, 314–323. Marekha, B. A., Kalugin, O. N., & Idrissi, A. (2015). Non-covalent interactions in ionic liquid ion pairs and ion pair dimers : a quantum chemical calculation analysis. Physical Chemistry Chemical Physics, 17, 16846–16857. Najmuddin, R. A., Mutalib, M. I. A., Shah, S. N., Suleman, H., Lethesh, K. C., Pilus, R. B. M., & Maulud, A. S. (2016). Liquid-liquid extraction of naphthenic acid using thiocyanate based ionic liquids. Procedia Engineering, 148, 662–670. Noack, K., Schulz, P. S., Paape, N., Kiefer, J., Wasserscheid, P., & Leipertz, A. (2010). The role of the C2 position in interionic interactions of imidazolium based ionic liquids: a vibrational and NMR spectroscopic study. Physical Chemistry Chemical Physics, 12(42), 14153. Ogunlaja, A. S., Hosten, E., & Tshentu, Z. R. (2014). Dispersion of asphaltenes in petroleum with ionic liquids: Evaluation of molecular interactions in the binary mixture. Industrial & Engineering Chemistry Research, 53(48), 18390–18401. Quinlan, P. J., & Tam, K. C. (2015). Water treatment technologies for the remediation of naphthenic acids in oil sands process-affected water. Chemical Engineering Journal, 279, 696– 714. Shah, S. N., Mutalib, M. I. A., Pilus, R. B. M., & Lethesh, K. C. (2014). Extraction of naphthenic acid from highly acidic oil using hydroxide-based ionic liquids. Energy & Fuels, 29(1), 106- 111. Shah, S. N., Kallidanthiyil Chellappan, L., Gonfa, G., Mutalib, M. I. A., Pilus, R. B. M., & Bustam, M. A. (2016a). Extraction of naphthenic acid from highly acidic oil using phenolate based ionic liquids. Chemical Engineering Journal, 284, 487–493. Shah, S. N., Mutalib, M. I. A., Ismail, M. F., Suleman, H., Lethesh, K. C., Pilus, & R. B. M. (2016b). Thermodynamic modelling of liquid-liquid extraction of naphthenic acid from dodecane using imidazolium based phenolate ionic liquids. Journal of Molecular Liquids, 219, 513–525. Socha, A. M., Parthasarathi, R., Shi, J., Pattathil, S., & Whyte, D. (2014). Efficient biomass pretreatment using ionic liquids derived from lignin and hemicellulose. Proceedings of the National Academy of Sciences, 3587–3595. Sun, N., Parthasarathi, R., Socha, A. M., Shi, J., Zhang, S., Stavila, V., Sale, K.L., Simmons, B.A., & Singh, S. (2014). Understanding pretreatment efficacy of four cholinium and imidazolium

145 ionic liquids by chemistry and computation. Green Chemistry, 2546–2557. Sun, Y., & Shi, L. (2012). Basic ionic liquids with imidazole anion: New reagents to remove naphthenic acids from crude oil with high total acid number. Fuel, 99, 83–87. Verdía, P., González, E. J., Moreno, D., Palomar, J., & Tojo, E. (2017). Deepening of the role of cation substituents on the extractive ability of pyridinium ionic liquids of N-Compounds from fuels. ACS Sustainable Chemistry & Engineering, 5(2), 2015–2025. Wazzan, N. A., Al-qurashi, O. S., & Faidallah, H. M. (2016). DFT/and TD-DFT/PCM calculations of molecular structure, spectroscopic characterization, NLO and NBO analyses of 4-(4- chlorophenyl) and 4-[4-(dimethylamino) phenyl]-2-oxo-1,2,5,6-tetrahydrobenzo[h] quinoline- 3-carbonitrile dyes. Journal of Molecular Liquids, 223, 29–47. Wu, C., Visscher, D., & Gates, I. D. (2017). Reactions of hydroxyl radicals with benzoic acid and benzoate. RSC Advances, 7, 35776–35785. Yang, J., Wang, X., Yim, W., & Wang, Q. (2015). Computational study on the intramolecular charge separation of D‑A‑π‑A organic sensitizers with different linker groups. The Journal of Physical Chemistry C, 119(47), 26355–26361. Youngs, T. G. A., & Hardacre, C. (2008). Application of static charge transfer within an ionic- liquid force field and its effect on structure and dynamics. ChemPhysChem, 9(11), 1548–1558. Zhang, A., Ma, Q., Wang, K., Liu, X., Shuler, P., &Tang, Y. (2006). Naphthenic acid removal from crude oil through catalytic decarboxylation on magnesium oxide. Applied Catalysis A: General, 303(1), 103–109.

146 Chapter Six: Interactions of Biodegradable Ionic Liquids with a

Model Naphthenic Acid

6.1 Introduction

Naphthenic acids (NAs) are complex mixtures of carboxylic acids with cyclic structures and aliphatic groups (Colati et al. 2013; Wu et al. 2017). In addition to increasing the acidity of crude oils, the presence of NAs also causes other serious problems such as poisoning catalysts, forming coke, and creating corrosion to equipment and pipelines (Zafar et al. 2016; Zhang et al. 2006).

Ionic liquids (ILs), recognized as “green solvents” and “solvents of the future”, provide us an alternative and promising approach to extract undesired products from fossil fuels, considering their outstanding properties such as non-volatility, low flammability, and reusability (Endres et al.

2006; Lü et al. 2017; Swatloski et al. 2003). It was reported that ILs are efficient at removing NAs from crude oil (Anderson et al. 2013; Shah S. N. 2016; Sasikumar et al. 2015).

Having excellent physicochemical properties and being easily synthesized, imidazolium-based ILs are the most extensively studied ILs (Sasikumar et al. 2015). However, the properties that make

ILs attractive result in low biodegradation of ILs (Liwarska-bizukojc et al. 2013; Phuong et al.

2010). Due to their high solubility, high stability, and low biodegradability, imidazolium-based

ILs are persistent pollutants that cause serious contamination after being released to aqueous solution (Romero et al. 2008; Domínguez et al. 2014). ILs are also reported to be toxic to a broad variety of organisms (Yoo et al. 2016). The most effective technique to remove organic pollutants from water is biodegradation (Liwarska-bizukojc et al. 2013), hence, a significant challenge

147 associated with the design and application of ILs is to increase the biodegradability of ILs (Jordan et al. 2015).

To improve the ultimate biodegradation of ILs and minimize the adverse environmental influences, researchers have paid more attention to change the framework of organic cations and anions of

ILs. It has been demonstrated that incorporating an ester in the side chain significantly enhanced imidazolium-based ILs biodegradation (Harjani et al. 2009). Furthermore, oxygenated and hydroxylated imidazolium-based ILs have a better biodegradability (Coleman et al. 2010). Amino groups are also reported to be able to increase the biodegradability of ILs (Hou et al. 2013).

Although there are numerous studies to investigate approaches to increase the biodegradability of

ILs, to the best of our knowledge, few theoretical studies are available to compare intramolecular and intermolecular interaction differences, and extraction mechanism variations of ILs with different biodegradable substitutional groups, especially for the removal of NAs from liquid oil.

The objective of the research was to fill the knowledge gap by exploring the influence of biodegradable substitutional groups on ILs intramolecular interactions, examining ILs extraction mechanisms of model NAs, and investigating the nature of the molecular interactions by using density functional theory (DFT) calculation. Five different biodegradable groups, including the hydroxyl group (-OH), amino group (-NH2), formate group (-COOH), methyl ester group (-

COOCH3), and methyl ether group (-OCH3) were incorporated to the cation of 1-butyl-3- methylimidazolium tetrafluoroborate ([BMIM][BF4]). The structures of the five ILs with these biodegradable groups are displayed in Figure 6.1. Cyclohexanecarboxylic acid (CHCA) was

148 selected to be the model NA. The results reported in this study will assist researchers to design ILs with high biodegradability and high extraction efficiency for NAs.

(a) [C4OHMIM][BF4] (b) [C4NHMIM][BF4]

(e) [C4OCMIM][BF4] (f) CHCA

Figure 6.1 Chemical structures of ILs with biodegradable groups and CHCA.

6.2 Computational Methods

The density functional computations were carried out using Gaussian 09 program package (Frisch et al. 2009). The geometries of CHCA and ILs with biodegradable groups were fully optimized by the M06-2X method in combination with the empirical dispersion-correction (DFT-D3) (Grimme et al. 2011) method and the 6-311++G(d,p) basis set. The ILs-CHCA interaction structures were

149 also optimized by employing the same method and basis set. Vibrational analyses were performed to confirm that the structures are at minimal energy without imaginary frequencies. The interaction energies were calculated with the correction by the counterpoise method for basis set superposition error (Kruse & Grimme 2012). The stabilization energy E(2) in the natural bond orbital (NBO) was determined by using the Gaussian 09 program package with the M06-2X/6-311++G(d,p) level of theory. The Multiwfn software package was adopted to analyze the wave functions of the optimized structures to obtain noncovalent interactions (NCI), and electron density differences (Lu

& Chen 2012a, 2012b). The interaction regions of NCI analysis were visualized and colored with the Visual Molecular Dynamics (VMD) software package (Humphrey et al. 1996). Topological properties were analyzed by using the atoms in molecules (AIM) theory (Bader 1991).

6.3 Results and Discussion

6.3.1 Optimized Geometries

The most stable geometries of CHCA and ILs with biodegradable groups are shown in Appendix

C Fig. S1, and some of the bond lengths are provided in Appendix C Table S1. It is illustrated that

- [BF4] is located above the imidazolium ring for all these stable structures of ILs, which is identical with [BMIM][BF4], demonstrating that the incorporation of biodegradation groups does not

- remarkably change the relative position of [BF4] to the cation (Wu et al. 2017). The van der Waals

(vdW) radii for nitrogen, oxygen, fluorine, and hydrogen are reported to be 1.55, 1.52, 1.47, and

1.2 Å, respectively (Bondi 1964). For [C4OHMIM][BF4], the distances of O30···H20, O30···H21,

O30···H28, and O30···H29 are 2.53, 2.61, 2.08, and 2.08 Å, respectively (see Appendix C Table

S1), which are shorter than the sum of the vdW radii for oxygen and hydrogen. Consequently, it

150 is concluded that hydrogen bonds are formed between the oxygen atom in –OH and hydrogen atoms of [C4OHMIM][BF4]. In addition, the nitrogen atom in –NH2 and oxygen atoms in -

COOCH3, -COOH, and -OCH3 form hydrogen bonds with hydrogen atoms in the ILs as well.

- Moreover, hydrogen bonding also occurs between fluorine atoms in [BF4] and hydrogen atoms in the cation for all these five type of ILs (Appendix C Fig. S1 and Table S1), which follows the same trends as [BMIM][BF4] (Wu et al. 2017). Interestingly, different from -OH and -OCH3, the distances of F24···H31(2.53 Å) in [C4NHMIM][BF4], F31···H28(1.99 Å) in

[C4COOHMIM][BF4], and F32···H31(2.20 Å) in [C4COOCMIM][BF4] are shorter than the vdW radii for fluorine and hydrogen, indicating the existence of intramolecular hydrogen bonds between

- a fluorine atom of [BF4] and a hydrogen atom of -NH2, -COOH, and -COOCH3 for those three

- type of ILs. On the other hand, the interaction between a fluorine atom in [BF4] and a carbon atom in the imidazolium ring corresponds to the presence of lone pair (LP)-π interaction.

To efficiently acquire the most stable interaction structures for ILs-CHCA, the electrostatic potential was analyzed for ILs and CHCA, shown in Figure 6.2. It was deduced that the highly positively charged regions for ILs are located around the imidazolium ring, whereas [BF4]- has a strongly negatively electrostatic potential. The addition of -OH and -OCH3 negatively influences the electrostatic potential, and those regions have stronger negative electrostatic potential compared with that for [BMIM][BF4] (Wu et al. 2017). With regards to CHCA, the most negative and positive electrostatic potentials are located around the oxygen atoms in the carboxylic group, and the hydrogen atom in the carboxylic group, separately.

151

(a) [C4OHMIM][BF4] (b) [C4NHMIM][BF4]

(e) [C4OCMIM][BF4] (f) CHCA

Figure 6.2 The electrostatic potential (a.u.) of ILs and CHCA.

Based on electrostatic potential analysis, CHCA were placed around different regions of ILs to obtain the most stable interaction structures as shown in Figure 6.3. There are three dominant

- hydrogen bonds for [C4OHMIM][BF4]-CHCA, one is between fluorine atom of [BF4] and the

152 hydrogen atom in the carboxylic group of CHCA; the other two are between oxygen atom in the carboxylic group of CHCA and hydrogen atoms of the cation, presented in Figure 6.3a and Table

6.1. With the formation of complexes, the distance of F25···H14 (displayed in Appendix C Fig.

S1 and Table S1) is lengthened from 2.61 Å to 2.79 Å (F26···H14), as displayed in Figure 6.3a and Table 6.1, longer than the sum of vdW radii for the fluorine and hydrogen atoms, indicating that F25 no longer forms a hydrogen bond with H14. In addition, the elongation of distances between H13, H16, and F25 implies weaker intramolecular interactions of [C4OHMIM][BF4] after it adsorbs CHCA. The above phenomena could be explained by the F25···H52 hydrogen bond formation. Therefore, the electronegativity of F25 is decreased and the interaction strengths between F25 and other hydrogen atoms are weaker (see Figure 6.3a and Table 6.1). F23···H13 and F26···C5 distances are also elongated after complex formation, suggesting that the interactions between anion and imidazolium ring are weaker as well. The interactions for [C4NHMIM][BF4]-

CHCA and [C4OCMIM][BF4]-CHCA follow the same trends as [C4OHMIM][BF4]-CHCA.

153

Table 6.1 Bond lengths (Å) of ILs-CHCA complexes. (a) (b) (c) (d) (e) F23···H13 2.485 F24···H13 2.463 F29···H16 2.341 F32···H13 2.389 F23···H13 2.482 F23···H18 2.414 F24···H18 2.485 F31···H13 2.953 F32···H14 2.698 F23···H18 2.425 F23···H19 2.500 F24···H19 2.533 F31···H14 2.611 F32···H31 3.326 F23···H19 2.496 F25···H13 2.624 F24···H22 2.415 F31···H28 2.987 F32···H57 1.761 F25···H13 2.638 F25···H14 2.786 F24···H31 2.726 F31···H54 1.679 F34···H16 2.443 F25···H14 2.764 F25···H16 2.535 F26···H13 2.763 F33···H19 2.292 F35···H19 2.633 F25···H16 2.530 F25···H52 1.689 F26···H14 2.872 F33···H24 2.519 F35···H24 3.698 F25···H55 1.687 F26···H19 2.389 F26···H16 2.565 F33···C5 2.811 F35···H57 2.630 F26···H19 2.387 F26···C5 2.823 F26···H53 1.688 O26···H13 2.427 F35···C5 2.823 F26···C5 2.829 O50···H13 2.065 F27···H19 2.348 O26···H18 2.260 O26···H13 4.073 O53···H13 2.053 O50···H14 2.447 F27···C5 2.803 O26···H47 2.700 O26···H18 2.423 O53···H14 2.486 O51···H13 2.060 O52···H13 2.131 O26···H50 2.530 O51···H14 2.449 O52···H14 2.395 O26···H52 2.620 O53···H28 1.865 O55···H13 1.970 O55···H18 2.434 O56···H31 2.539

Note: (a) [C4OHMIM][BF4]-CHCA, (b) [C4NHMIM][BF4]-CHCA, (c) [C4COOHMIM][BF4]- CHCA, (d) [C4COOCMIM][BF4]-CHCA, and (e) [C4OCMIM][BF4]-CHCA.

154

(a) [C4OHMIM][BF4]-CHCA (b) [C4NHMIM][BF4]-CHCA

(e) [C4OCMIM][BF4]-CHCA

Figure 6.3 The optimized structures of ILs-CHCA complexes.

156 With the exception of F32···H57, O55···H13, and O55···H18, four more hydrogen bonds are formed when [C4COOCMIM][BF4] interacts with CHCA, including F35···H57, O56···H31,

O26···H52, and O26···H50, as shown in Figure 6.3d and Table 6.1. As for [C4COOHMIM][BF4]-

CHCA, five hydrogen bonds (F31···H54, O52···H13, O52···H14, O53···H28, and O26···H47) are formed after interaction (see Figure 6.3c and Table 6.1). Judging from the length of hydrogen bonds, the two dominant hydrogen bonds for [C4COOHMIM][BF4]-CHCA are F31···H54 and

O53···H28. Distance analysis between the fluorine and hydrogen (carbon) atoms in isolated

[C4COOHMIM][BF4] and [C4COOHMIM][BF4]-CHCA also leads to the conclusion that the interactions between the anion and imidazolium ring are weaker upon complex formation.

6.3.2 Interaction Energies

The interaction energies are important to evaluate the stability of interactions. They are defined as the energy difference between complexes and the sum of isolated ILs and CHCA, and were calculated according to:

E  EILsCHCA  (EILs  ECHCA) Eq. 6-1

As shown in Table 6.2, the interaction energies between CHCA and [C4OHMIM][BF4]/

[C4NHMIM][BF4]/[C4COOHMIM][BF4]/[C4COOCMIM][BF4]/[C4OCMIM][BF4] follow the order of [C4COOCMIM][BF4]-CHCA>[C4COOHMIM][BF4]-CHCA>[C4OHMIM][BF4]-

CHCA>[C4OCMIM][BF4]-CHCA>[C4NHMIM][BF4]-CHCA. Hence, it is deduced that the interaction energies between CHCA and biodegradable ILs with two electronegative atoms are higher than that between CHCA and biodegradable ILs with one electronegative atom. In addition, the interaction energies between CHCA and five types of biodegradable ILs are higher than that

157 between CHCA and [BMIM][BF4] (Wu et al. 2017), which is 15.22 kcal/mol, demonstrating that the incorporation of biodegradable groups to ILs promote the extraction of CHCA from crude oil.

The interaction energy for [C4COOCMIM][BF4]-CHCA is highest among all complexes, which is attributed to the larger number of hydrogen bonding interactions in [C4COOCMIM][BF4]-

CHCA than other complexes. Similarly, the higher interaction energy of [C4COOHMIM][BF4]-

CHCA than [C4OHMIM][BF4]-CHCA, [C4OCMIM][BF4]-CHCA, and [C4NHMIM][BF4]-

CHCA is ascribed to the larger number of intermolecular hydrogen bonds which is consistent with the geometry analysis.

Table 6.2 Interaction energies between ILs and CHCA. Complexes ∆E(kcal/mol)

[C4OHMIM][BF4]-CHCA -15.89

[C4NHMIM][BF4]-CHCA -15.81

[C4COOHMIM][BF4]-CHCA -16.81

[C4COOCMIM][BF4]-CHCA -17.11

[C4OCMIM][BF4]-CHCA -15.88

6.3.3 NBO Analyses

To investigate electron distribution and bond characteristics after complex formation, NBO charge distributions of ILs and ILs-CHCA are summarized in Appendix C Table S2. For

[C4OHMIM][BF4], the charge of H31 (0.46291) is largest among all the hydrogen atoms, which is due to the higher electronegativity of the oxygen atom over that of the carbon atom. In addition,

H13 (0.2726) is more positively charged than other hydrogen atoms of imidazolium ring since it participates in two hydrogen bonding interactions with fluorine atoms. In [C4OHMIM][BF4]-

158 CHCA, owing to strong hydrogen bonding of F25···H52, F25 (-0.59611) is more negatively charged than the other fluorine atoms, and H52 (0.52920) is the most positively charged hydrogen atom. The pronounced increase of the charges of H13 (0.28877) and H14 (0.24554) after the formation of complexes is ascribed to the electron transfer from H13 and H14 to O50 because of

H13···O50 and H14···O50 interactions. The redistribution of NBO charges for

[C4NHMIM][BF4]-CHCA and [C4OCMIM][BF4]-CHCA are similar to that of

[C4OHMIM][BF4]-CHCA.

For [C4COOHMIM][BF4], the highest charge of H28 (0.51843) is due to the electronegativity of the oxygen atom as well as F31···H28 interaction. Hydrogen bonding between O26···H13 brings about electron transfer from H13 to O26, consequently, the charge of H13 is higher than the charge of other hydrogen atoms in the imidazolium ring. Moreover, F31···H28, F31···H14, and

F31···H13 interactions give rise to the more negative charge of F31 (-0.60678) than other fluorine atoms. With respect to [C4COOHMIM][BF4]-CHCA, the charge of H54 (0.53893) is highest among all hydrogen atoms, which results from the strong F31···H54 hydrogen bonding interaction and electron transfer from H54 to adjacent O53. The charge of H28 (0.51321) is second largest because of O53···H28 interaction. In addition, F31 (-0.61088) is far more negatively charged than other fluorine atoms since F31···H54 interaction leads to the electron transfer from H54 to F31.

Regarding [C4COOCMIM][BF4]-CHCA, the charge of the hydrogen atom (H57) in the carboxylic group of [C4COOCMIM][BF4]-CHCA is lower than that in the other four complexes. The reason is possibly due to weaker F32···H57 and F35···H57 formation of hydrogen bonds when

[C4COOCMIM][BF4] interacts with CHCA, whereas one strong F···H hydrogen bond occurs in

159 the other four ILs-CHCA complexes. F32···H57 and F35···H57 hydrogen bonds formation could also account for the largest positive charge of H57 among all hydrogen atoms.

The donor-acceptor interactions of ILs-CHCA and their stabilization energies E(2) were calculated to determine the extent of interaction. The interaction intensity is reflected by the value of E(2).

Higher values of E(2) indicate that electrons are more likely to migrate from donor to acceptor orbitals and stronger interaction exists between the donor and acceptor. As displayed in Table 6.3,

LP(F)-*(O-H) has the highest stabilization energy for all these five ILs-CHCA, suggesting that

LP(F)-*(O-H) is the strongest interaction among all donor-acceptor interactions. In addition, the stabilization energy of LP(O)-*(C-H) between ILs and CHCA is substantially higher than most of other donor-acceptor interactions. Except for LP(F)-*(C-H) and LP(O)-*(C-H) interactions, it is noteworthy that the stabilization energy of LP(O53)-*(O27-H28) in [C4COOHMIM][BF4]-

CHCA is 11.56 kcal/mol, demonstrating that LP(O53) and *(O27-H28) interaction is also extraordinarily strong, which is in agreement with geometry analysis and provides further explanation for the higher interaction energy between [C4COOHMIM][BF4] and CHCA.

Table 6.3 The donor-acceptor interaction in ILs-CHCA complexes, and their stabilization energies, E(2)(kcal/mol). Donor Acceptor E(2) Donor Acceptor E(2)

[C4OHMIM][BF4]-CHCA (C5-H13) *(C49-O50) 0.11 LP(O50) *(N1-C5) 0.25

(C6-H14) *(C49-O50) 0.07 LP(O50) *(C5-H13) 5.45

(B24-F25) *(O51-H52) 0.11 LP(O50) *(C6-H14) 0.83 LP(F25) *(O51-H52) 20.18 *(C49-O50) *(C5-H13) 0.14

160 (C37–C49) *(C5-H13) 0.25 *(C49-O50) *(C6-H14) 0.07 (C49-O50) *(C5-H13) 1.03 (O51-H52) *(B24-F25) 0.05 (C49-O50) *(C6-H14) 0.36 LP(O51) *(B24-F25) 0.09

[C4NHMIM][BF4]-CHCA (C5-H13) *(C50-O51) 0.12 LP(O51) *(C5-H13) 1.97

 (C6-H14) *(C49-O50) 0.07 LP(O51) *(C6-H14) 0.78

(B25-F26) *(O52-H53) 0.10 LP(O51) *(N1-C5) 0.13 LP(F26) *(O52-H53) 20.31 *(C50-O51) *(C5-H13) 0.23 (C38–C50) *(C5-H13) 0.25 *(C50-O51) *(C6-H14) 0.05 (C50-O51) *(C5-H13) 0.98  (O52-H52) *(B25-F26) 0.06 (C50-O51) *(C6-H14) 0.35 LP(O52) *(B25-F26) 0.09

[C4COOHMIM][BF4]-CHCA (C5-H13) *(C51-O52) 0.23 (C51-O52) *(O27-H28) 0.16 (O27-H28) *(C51-O53) 0.06 (C51-O52) *(N1-C5) 0.08 LP(O26) *(C34-C39) 0.20 (C51-O53) *(C5-H13) 0.09 LP(O26) *(C38-H47) 0.15 (O53-H54) *(O27-H28) 0.29 LP(O26) *(C51-O52) 2.44 LP(O52) *(N1-C5) 0.07 LP(O26) *(C51-O53) 0.14 LP(O52) *(C5-H13) 2.07 LP(F31) *(O53-H54) 18.09 LP(O52) *(C6-H14) 1.65 (C38-H47) *(C5-H13) 0.07 LP(O53) *(O27-H28) 11.56 (C39-C51) *(C5-H13) 0.18 *(C51-O52) *(C5-H13) 0.27 (C39-C51) *(C6-H14) 0.06 (O53-H54) *(B30-F31) 0.15 (C51-O52) *(C5-H13) 2.10 LP(O53) *(B30-F31) 0.12

[C4COOCMIM][BF4]-CHCA (C5-H13) *(C54-O55) 0.07 (C42-C54) *(C5-H13) 0.39 (C7-H18) *(O54-O55) 0.11 (C54-O55) *(C5-H13) 0.25 (C25-O26) *(C54-O55) 0.12 (C54-O55) *(N4-C5) 0.06 (C28-H29) *(C42-H52) 0.25 (C54-O55) *(C7-H18) 0.82

161 LP(O26) *(C37-C42) 0.37 (C54-O55) *(C25-O56) 0.08 LP(O26) *(C41-H50) 0.36 (C56-H57) *(C28-H31) 0.06 LP(O26) *(C41-H51) 0.06 LP(O55) *(N4-C5) 0.21 LP(O26) *(C42-H52) 0.07 LP(O55) *(C5-H13) 10.26 LP(O26) *(C54-O55) 1.14 LP(O55) *(C7-H18) 0.40 (F32-B33) *(O56-H57) 0.15 LP(O56) *(C28-H31) 0.55 LP(F32) *(O56-H57) 14.99 *(C54-O55) *(C7-H18) 0.08 (C42-H52) *(C28-H29) 0.12 LP(O56) *(F32-B33) 0.07

[C4OCMIM][BF4]-CHCA (C5-H13) *(C52-O53) 0.13 LP(O53) *(N1-C5) 0.23 (C6-H14) *(C52-O53) 0.05 LP(O53) *(C5-H13) 5.61 (B24-F25) *(O54-H55) 0.29 LP(O53) *(C6-H14) 0.80 LP(F25) *(O54-H55) 20.35 *(C52-O53) *(C5-H13) 0.17 (C40-C52) *(C5-H13) 0.26 *(C52-O53) *(C6-H14) 0.06 (C52-O53) *(C5-H13) 1.19 (O54-H55) *(B24-F25) 0.05 (C52-O53) *(C6-H14) 0.28 LP(O54) *(B24-F25) 0.09

6.3.4 Topological Properties of Interactions

Through analysis of the values of electron density (r) and Laplacian of electron density 2(r) at the bond critical points (BCPs) of the chemical bonds, the atoms in molecules (AIM) theory (Bader

1991) is valuable to characterize chemical bonds, especially hydrogen bonds. To identify bonding interactions for ILs and ILs-CHCA, the results of the AIM analysis are shown in Figure 6.4 and listed in Appendix C Table S3. The existence of BCPs indicates the formation of hydrogen bonds

(Popelier et al. 1998). [C4OHMIM][BF4] and [C4OCMIM][BF4] each have 7 BCPs. On the other hand, [C4NHMIM][BF4], [C4COOHMIM][BF4], and [C4COOCMIM][BF4] all have 10 BCPs, suggesting the presence of more intramolecular hydrogen bonds than other ILs, consistent with

162 geometry analysis. Interestingly, three more BCPs are found for [C4OHMIM][BF4],

[C4NHMIM][BF4], and C4OCMIM][BF4] after interacting with CHCA, whereas the increase of the number of BCPs are 9 and 5 for [C4COOCMIM][BF4] and [C4COOHMIM][BF4], respectively. It corresponds to geometry analysis and provides an explanation for the higher interaction energies of [C4COOCMIM][BF4]-CHCA and [C4COOHMIM][BF4]-CHCA.

Additionally, it is deduced that the incorporation of biodegradable groups with two electronegative atoms to ILs form more intermolecular hydrogen bonds with CHCA than those that have one electronegative atom.

(a) [C4OHMIM][BF4] (b) [C4NHMIM][BF4]

163 (c) [C4COOHMIM][BF4] (d) [C4COOCMIM][BF4]

(e) [C4OCMIM][BF4] (f) [C4OHMIM][BF4]-CHCA

(g) [C4NHMIM][BF4]-CHCA (h) [C4COOHMIM][BF4]-CHCA

164

(i) [C4COOCMIM][BF4]-CHCA (j) [C4OCMIM][BF4]-CHCA

Figure 6.4 BCPs and bond paths in ILs and ILs-CHCA complexes.

As shown in Appendix C Table S3, the values of the Laplacian of the electron density are positive for all ILs and ILs-CHCA, indicating that the electrons tend to segregate. It also suggests the existence of ionic bonds, hydrogen bonds, and vdW interactions in ILs and ILs-CHCA (Lu & Chen

2012b). Moreover, the strength of interactions can be evaluated by comparing the electron density

(r) value of BCPs. Larger values of (r) correspond to stronger interactions (Checińska et al.

2003). The electron densities of O26···H13 and F31···H28 in [C4COOHMIM][BF4] are larger than the electron densities of other ILs, which indicates the presence of stronger hydrogen bonds.

In addition, the electron density of F25···H52 is greatest for [C4OHMIM][BF4]-CHCA, implying that the hydrogen bonds between fluorine and hydrogen is the strongest hydrogen bond. Further analysis of the largest electron density for other ILs-CHCA also demonstrates that the F···H interaction is the strongest interaction. Compared to the distances of other hydrogen bonds, the

F···H distance is shortest for all five ILs-CHCA, therefore, it is inferred that electron density is correlated to intermolecular hydrogen bond distance. The electron density of O···H ranks second for [C4NHMIM][BF4]-CHCA, [C4OCMIM][BF4]-CHCA, [C4COOHMIM][BF4]-CHCA, and

165 [C4COOCMIM][BF4]-CHCA, which suggests that O···H hydrogen bond is also critical in the interaction between ILs and CHCA. Except for hydrogen bonding interactions, the F35···O55 and

F35···O26 pairs in [C4COOCMIM][BF4]-CHCA indicate the existence of anion-anion interactions.

6.3.5 NCI Analyses

Through performing NCI analyses, which is based on the reduced electron density gradient (RDG)

(Contreras-garcía et al. 2011), the intramolecular and intermolecular types and strength can be evaluated (Johnson et al. 2010). In the plots of RDG versus sign(2), the peaks in the sign(2)<0, sign(2)=0, and sign(2)>0 region suggest attractive interactions, vdW interactions, and steric effects, respectively. Furthermore, the interaction types and strength can be identified through analyzing the color and area in the gradient isosurface diagram; red indicates steric repulsions, green means weak interactions such as vdW interactions, and blue represents strong attractive interactions such as hydrogen bonds (Güryel et al. 2017; Wu et al. 2013). To investigate intramolecular and intermolecular interaction types and strength, the plots of RDG versus sign(2) and the gradient isosurface (s=0.6 a.u.) for ILs are shown in Figure 6.5. In addition, the plots of RDG versus sign(2) and the gradient isosurface (s=0.7 a.u.) for ILs-CHCA, displayed in Figure 6.6, are investigated.

166

(a) [C4OHMIM][BF4]

(b) [C4NHMIM][BF4]

167 (c) [C4COOHMIM][BF4]

(d) [C4COOCMIM][BF4]

(e) [C4OCMIM][BF4]

Figure 6.5 The sign(λ2)ρ vs RDG (left) and the gradient isosurfaces (right) for ILs.

Note: red indicates sign(λ2)ρ>0 and blue indicates sign(λ2)ρ<0.

The results in Figure 6.5 reveal that there is no peak in the sign(2)<0 region for

[C4OHMIM][BF4], [C4NHMIM][BF4], and [C4OCMIM][BF4], meaning that no apparent intramolecular hydrogen bonds exists in those ILs. Whereas prominent peaks at -0.02 a.u. in Figure

6.5c correspond to the F31···H28, O26···H18, and O26···H13 hydrogen bonds in

[C4COOHMIM][BF4], which is consistent with geometry analysis, NBO analysis, and AIM

168 analysis. The peaks at -0.02 a.u. for [C4COOCMIM][BF4] are not as strong as those for

[C4COOHMIM][BF4], which is attributed to the smaller electron densities of intramolecular hydrogen bonds for [C4COOCMIM][BF4]. On the other hand, the presence of peaks at 0.02 a.u., shown in Figure 6.5b and d, represents strong steric intramolecular interactions for

[C4NHMIM][BF4] and [C4COOCMIM][BF4]. Compared to the color and area of gradient isosurface for [BMIM][BF4], the gradient isosurface analyses of five biodegradable ILs imply that the addition of biodegradable groups do not destroy the LP-π interactions between the anion and cation. On the other hand, the areas of LP-π interactions are larger for [C4NHMIM][BF4],

[C4COOHMIM][BF4], and [C4COOCMIM][BF4] over that of [C4OHMIM][BF4] and

[C4OCMIM][BF4].

After interacting with CHCA, all the ILs-CHCA complexes have spikes at 0.02 a.u., as shown in

Figure 6.6, indicating the existence of steric effects for all these interactions. [C4OHMIM][BF4]-

CHCA, [C4NHMIM][BF4]-CHCA, and [C4OCMIM][BF4]-CHCA have spikes at -0.04 a.u., which corresponds to the dark blue circle between the fluorine atom and the hydrogen atom, further confirming the existence of strong hydrogen bonds. The peaks at -0.02 a.u. in Figure 6.6a, b, and e denote hydrogen bonds formed between oxygen and hydrogen atoms for [C4OHMIM][BF4]-

CHCA, [C4NHMIM][BF4]-CHCA, and [C4OCMIM][BF4]-CHCA. According to AIM analysis, the electron density of O55···H13 in [C4COOCMIM][BF4]-CHCA is larger than the electron density of O···H in [C4OHMIM][BF4]-CHCA, [C4NHMIM][BF4]-CHCA, and

[C4OCMIM][BF4]-CHCA, which can account for the deviation of the peaks from -0.2 a.u. (see

Figure 6.6d and Appendix C Table S3). On the other hand, the highest electron density of

F34···H16 for [C4COOCMIM][BF4]-CHCA is smaller than the highest electron density of F···H

169 in other complexes, hence, the strongest peak for [C4COOCMIM][BF4]-CHCA is located further from -0.4a.u., as shown in Figure 6.6d and Appendix C Table S3. In addition, there are three spikes for [C4COOHMIM][BF4]-CHCA in sign(2)<0 region, and can be explained by the strong electron density of: F31···H54, O53···H28, and O26···H18 in AIM analysis (see Figure 6.6c and

Appendix C Table S3).

(a) [C4OHMIM][BF4]-CHCA

(b) [C4NHMIM][BF4]-CHCA

170

(d) [C4COOCMIM][BF4]-CHCA

171

(e) [C4OCMIM][BF4]-CHCA

Figure 6.6 The sign(λ2)ρ versus RDG (left) and the gradient isosurfaces (right) for ILs- CHCA complexes.

Note: red indicates sign(λ2)ρ>0 and blue indicates sign(λ2)ρ<0.

6.3.6 Electron Density Difference Analyses

When ILs interact with CHCA, there is a transfer of electron density during the interaction process

(Noack et al. 2010). The electron density change was determined by subtracting the electron density of complexes from the sum of electron density of isolated ILs and CHCA:

  ILsCHCA  (ILs  CHCA) Eq. 6-2

To evaluate electron density redistribution caused by the interactions between ILs and CHCA, electron density distribution maps were plotted. As depicted in Appendix C Fig. S3, the obvious electron density change mostly locates around the interaction region between ILs and CHCA. The formation of F25···H52 hydrogen bond in [C4OHMIM][BF4]-CHCA, shown in Appendix C Fig.

S3a, increases the electron density of F25 and decreases that of H52 because of electron transfer from H52 to F25. In addition, due to O50···H13 and O50···H14 interactions, the electron density

172 of O50 increases whereas the electron densities of H13 and H14 decrease. The formation of hydrogen bonds between electronegative and hydrogen atoms provoke the increase of density for the electronegative atoms such as F and O, and induce the decrease of electron density for hydrogen atom.

In summary, the interaction energy between biodegradable ILs and CHCA is higher than that between [BMIM][BF4] and CHCA. Moreover, biodegradable ILs with two electronegative atoms have higher interaction energy with CHCA than that having one electronegative atom. Compared with the extraction mechanism for [BMIM][BF4], the main interaction is still hydrogen bonding.

However, biodegradable ILs form more hydrogen bonds with CHCA than [BMIM][BF4].

Therefore, it is deduced that the design of biodegradable ILs promote the extraction of CHCA.

Additionally, the greater the number of electronegative atoms in biodegradable group of ILs, the easier it is to extract CHCA.

6.4 Conclusions

Density functional theory was used to investigate the influence of biodegradable group addition on the intramolecular and intermolecular interactions of [BMIM][BF4]. In contrast with

[C4OHMIM][BF4] and [C4OCMIM][BF4], hydrogen bonds are formed between fluorine and hydrogen atoms of the substitutional group in [C4NHMIM][BF4], [C4COOHMIM][BF4], and

[C4COOCMIM][BF4]. The interaction energies of [C4COOHMIM][BF4]-CHCA and

[C4COOCMIM][BF4]-CHCA are higher than that of other biodegradable ILs-CHCA complexes, and they are all higher than that of [BMIM][BF4]-CHCA. When ILs interact with CHCA, hydrogen

173 bonding is the main interaction mechanism and steric repulsion is pronouncedly strong. ILs with the incorporation of -COOH or -COOCH3 form more intermolecular hydrogen bonds with CHCA than that with the addition of -OH, -NH2, or -OCH3. With the exception of LP(F)-*(C-H) and

LP(O)-*(C-H), LP(O)-*(O-H) interaction plays an important role for [C4COOHMIM][BF4]-

CHCA. On the other hand, anion-anion interactions exist in [C4COOCMIM][BF4]-CHCA. The formation of the complex induces charge transfer and electron density change of the interacting atoms.

6.5 References

Anderson, K., Goodrich, P., Hardacre, C., Hussain, A., Rooney, D. W., & Wassell, D. (2013). Removal of naphthenic acids from crude oil using amino acid ionic liquids. Fuel, 108, 715– 722. Bader, R. F. W. (1991). A quantum theory of molecular structure and its applications. Chemical Reviews, 91, 893–928. Bondi, A. (1964). van der Waals volumes and radii. The Journal of Physical Chemistry, 68, 441– 451. Checińska, L., Grabowski, S. J., & Małecka, M. (2003). An analysis of bifurcated H-bonds: Crystal and molecular structures of O,O-diphenyl 1-(3-phenylthioureido) pentanephosphonate and O,O-diphenyl 1-(3-phenylthioureido)butanephosphonate. Journal of Physical Organic Chemistry, 16(4), 213–219. Colati, K. A. P., Dalmaschio, G. P., De Castro, E. V. R., Gomes, A. O., Vaz, B. G., & Romão, W. (2013). Monitoring the liquid/liquid extraction of naphthenic acids in brazilian crude oil using electrospray ionization FT-ICR mass spectrometry (ESI FT-ICR MS). Fuel, 108, 647–655. Coleman, D., Gathergood, N., & Coleman, D. (2010). Biodegradation studies of ionic liquids. Chemical Society Reviews, (39), 600–637. Contreras-garcía, J., Yang, W., & Johnson, E. R. (2011). Analysis of hydrogen-bond interaction potentials from the electron density: integration of noncovalent interaction regions. The Journal

174 of Physical Chemistry A, 115, 12983–12990. Domínguez, C. M., Munoz, M., Quintanilla, A., Pedro, M. De, Ventura, S. P. M., Coutinho, J. A. P., Casas, J.A., & Rodriguez, J. J. (2014). Degradation of imidazolium-based ionic liquids in aqueous solution by Fenton oxidation. Journal of Chemical Technology and Biotechnology, 89(8), 1197–1202. Endres, F., Zein, S., & Abedin, E. (2006). Air and water stable ionic liquids in physical chemistry. Physical Chemistry Chemical Physics, 8, 2101–2116. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, Jr., J. A.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, Ö.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09, Revision E.01, Gaussian, Inc., Wallingford CT, 2013. Grimme, S., Ehrlich, S., & Goerigk, L. (2011). Effect of the damping function in dispersion corrected density functional theory. Journal of Computational Chemistry, 32, 1456–1465. Güryel, S., Alonso, M., Hajgató, B., Dauphin, Y., Lier, G. Van, & Geerlings, P. (2017). A computational study on the role of noncovalent interactions in the stability of polymer/graphene nanocomposites. Journal of Molecular Modeling, 23, 43. Harjani, J. R., Farrell, J., Garcia, M. T., Singer, D., & Scammells, P. J. (2009). Further investigation of the biodegradability of imidazolium ionic liquids. Green Chemistry, 11, 821–829. Hou, X., Liu, Q., Smith, T. J., Li, N., & Zong, M. (2013). Evaluation of toxicity and biodegradability of cholinium amino acids ionic liquids. PloS One, 8(3), e59145. Humphrey, W., Dalke, A., & Schulten, K. (1996). VMD : visual molecular dynamics. Journal of Molecular Graphics, 14, 33–38. Johnson, E. R., Keinan, S., Mori Sánchez, P., Contreras-García, J., Cohen, A. J., & Yang, W.

175 (2010). Revealing noncovalent interactions. Journal of the American Chemical Society, 132(18), 6498–6506. Jordan, A., & Gathergood, N. (2015). Biodegradation of ionic liquids–a critical review. Chemical Society Reviews, 44, 8200–8237. Kruse, H., & Grimme, S. (2012). A geometrical correction for the inter- and intra-molecular basis set superposition error in Hartree-Fock and density functional theory calculations for large systems. The Journal of Chemical Physics, 136, 154101. Liwarska-bizukojc, E., & Gendaszewska, D. (2013). Removal of imidazolium ionic liquids by microbial associations : Study of the biodegradability and kinetics. Journal of Bioscience and Bioengineering, 115(1), 71–75. Lü, R., Wu, C., Lin, J., Xiao, Y., Wang, F., & De la Hoz Siegler, H. (2017). Theoretical study on interactions between Trifluoromethanesulfonate (Triflate) based ionic liquid and thiophene. Journal of Molecular Liquids, 237, 289–294. Lu, T., & Chen, F. (2012a). Multiwfn: A multifunctional wavefunction analyzer. Journal of Computational Chemistry, 33(5), 580–592. Lu, T., & Chen, F. (2012b). Quantitative analysis of molecular surface based on improved Marching Tetrahedra algorithm. Journal of Molecular Graphics and Modelling, 38, 314–323. Noack, K., Schulz, P. S., Paape, N., Kiefer, J., Wasserscheid, P., & Leipertz, A. (2010). The role of the C2 position in interionic interactions of imidazolium based ionic liquids: a vibrational and NMR spectroscopic study. Physical Chemistry Chemical Physics, 12(42), 14153–14161. Phuong, T., Pham, T., Cho, C., & Yun, Y. (2010). Environmental fate and toxicity of ionic liquids : A review. Water Research, 44(2), 352–372. Popelier, P. L. A. (1998). Characterization of a dihydrogen bond on the basis of the electron density. The Journal of Physical Chemistry A, 102, 1873–1878. Romero, A., Santos, A., Tojo, J., & Rodr, A. (2008). Toxicity and biodegradability of imidazolium ionic liquids. Journal of Hazardous Materials, 151, 268–273. Sasikumar, Y., Adekunle, A. S., Olasunkanmi, L. O., Bahadur, I., Baskar, R., & Kabanda, M. M. (2015). Experimental, quantum chemical and Monte Carlo simulation studies on the corrosion inhibition of some alkyl imidazolium ionic liquids containing tetra fluoroborate anion on mild steel in acidic medium. Journal of Molecular Liquids, 211, 105–118. Shah, S. N., Mutalib, M. I. A., Binti, R., Pilus, M., & Lethesh, K. C. (2015). Extraction of

176 naphthenic acid from highly acidic oil using hydroxide-based ionic liquids. Energy & Fuels, 29, 106–111. Shah, S. N., Kallidanthiyil Chellappan, L., Gonfa, G., Mutalib, M. I. A., Pilus, R. B. M., & Bustam, M. A. (2016). Extraction of naphthenic acid from highly acidic oil using phenolate based ionic liquids. Chemical Engineering Journal, 284, 487–493. Swatloski, R. P., Holbrey, J. D., & Rogers, R. D. (2003). Ionic liquids are not always green: hydrolysis of 1-butyl-3-methylimidazolium hexafluorophosphate. Green Chemistry, 5, 361– 363. Wu, P., Chaudret, R., Hu, X., Yang, W. (2013). Noncovalent interaction analysis in fluctuating environments. Journal of Chemical Theory and Computation, 9(5), 2226–2234. Wu, C., De Visscher, A., & Gates, I. D. (2017). Molecular interactions between 1-butyl-3- methylimidazolium tetrafluoroborate and model naphthenic acids: A DFT study. Journal of Molecular Liquids, 243, 462–471. Wu, C., Visscher, D., & Gates, I. D. (2017). Reactions of hydroxyl radicals with benzoic acid and benzoate. RSC Advances, 7, 35776–35785. Yoo, B., Jing, B., Jones, S. E., Lamberti, G. A., Zhu, Y., Shah, J. K., & Maginn, E. J. (2016). Molecular mechanisms of ionic liquid cytotoxicity probed by an integrated experimental and computational approach. Scientific Reports, 6, 1–7. Zafar, F., Mandal, P. C., Shaari, K. Z. B. K., & Moniruzzaman, M. (2016). Total acid number reduction of naphthenic acid using subcritical methanol and 1-butyl-3-methylimidazolium octylsulfate. Procedia Engineering, 148, 1074–1080. Zhang, A., Ma, Q., Wang, K., Liu, X., Shuler, P., & Tang, Y. (2006). Naphthenic acid removal from crude oil through catalytic decarboxylation on magnesium oxide. Applied Catalysis A: General, 303(1), 103–109.

177 Chapter Seven: Comparison of Electronic and Physicochemical Properties between

Imidazolium-based and Pyridinium-based Ionic Liquids

7.1 Introduction

Ionic liquids (ILs), often referred to as “green solvents”, are salts in liquid state at room temperature with extremely low vapor pressure, low combustibility, and favorable solvating properties for a large range of polar and nonpolar compounds (Mahmood et al., 2017). They are alternative solvents for extraction of sulfur- and nitrogen-containing compounds from oil (Lin et al., 2017; Lü et al., 2017). In addition, they are also reported to be efficient in removing naphthenic acids (Shah et al., 2016; C. Wu et al., 2017; Wu et al. 2018). Furthermore, researchers have utilized

ILs to separate bioactive compounds, produce catalysts and electrodes, and pretreat cellulose (Cao et al., 2017; Ventura et al., 2017; Watanabe et al., 2017). It is also noteworthy that ILs have potential applications in pharmaceutics and medicine (Egorova et al., 2017).

ILs are obtained by combining ions with specific characteristics in order to tune them to specific capabilities, hence, they are called “designer solvents” (Niedermeyer et al., 2013). Bulky and asymmetric organic cations such as 1-butyl-3-methylimidazolium and 1-butyl-3- methylpyridinium, and common anions including tetrafluoroborate, hexafluorophosphate, hydrogen sulfate, methylsulfate, and ethylsulfate can be used for the synthesis of ILs (Gardas &

Coutinho, 2008). It has been demonstrated that the properties of ILs can be altered to some extent by varying the cation, anion, or substituent groups (Herrera et al., 2016). For instance, the density of imidazolium-based ILs increases with increasing molecular weight of the anion, whereas the

178 density drops with increasing alkyl-substituted chain length on the cation or by introducing a third alkyl substituent on the imidazolium ring at the C2 position (Fredlake et al., 2004).

Efficient application of ILs in multidisciplinary areas requires tuning of ILs so that they have desirable physicochemical properties (Marium et al., 2017). Considering that their physicochemical properties, such as density, self-diffusion coefficient, and vapor pressure, are tied to their intramolecular and intermolecular interactions, a fundamental issue in designing ILs is understanding the strength of intramolecular and intermolecular interactions, specifically, the intermolecular interactions in the bulk ionic fluid (Fumino et al., 2014; Kurnia et al., 2015). There are many kinds of ILs, making it time-consuming and laborious to select ILs for specific applications through experimental investigations (Chen & Bryantsev, 2017). Quantum chemistry and molecular simulations provide relatively easy means to investigate molecular interactions and characterize physicochemical properties of ILs so that the relationships between molecular structures and physicochemical properties can be understood at the molecular level (Červinka et al., 2016).

Imidazolium-based ILs and pyridinium-based ILs are commonly used ILs, and they are similar with respect to cationic ring structure and atomic composition ( Hou et al. 2013; Yeganegi &

Sokhanvaran 2013). Nevertheless, there are few studies that compare their electronic and physicochemical properties by using density functional theory (DFT) and molecular dynamics simulation to investigate the influence of intermolecular interactions on physicochemical properties of ILs.

179 Here, we use DFT to explore the differences between imidazolium-based ILs and pyridinium- based ILs, and in particular, the electronic structures of the imidazolium cation and pyridinium cation and the intermolecular interactions, including van der Waals (vdW) interactions, hydrogen bonds, π-stacking, and electrostatic interactions. Moreover, molecular dynamics simulations are conducted to investigate the influence of cations and anions on physicochemical properties such as density and self-diffusion coefficient. Specifically, 1-butyl-3-methylimidazolium tetrafluoroborate ([BMIM][BF4]), 1-butyl-3-methylimidazolium hexafluorophosphate

([BMIM][PF6]), 1-butyl-3-methylimidazolium hydrogen sulfate ([BMIM][HSO4]), 1-butyl-3- methylimidazolium methylsulfate ([BMIM][MSO4]), 1-butyl-3-methylimidazolium ethylsulfate

([BMIM][ESO4]), 1-butyl-3-methylpyridinium tetrafluoroborate ([BMPy][BF4]), 1-butyl-3- methylpyridinium hexafluorophosphate ([BMPy][PF6]), 1-butyl-3-methylpyridinium hydrogen sulfate ([BMPy][HSO4]), 1-butyl-3-methylpyridinium methylsulfate ([BMPy][MSO4]), and 1- butyl-3-methylpyridinium ethylsulfate ([BMPy][ESO4]) were selected for investigation. The structures of cations and anions are depicted in Figure 7.1.

N N N

(a) [BMIM]+ (b) [BMPy]+

F F O O O F F O O O F B F P S S S F F O O O F OH O O F

- - - - - (c) [BF4] (d) [PF6] (e) [HSO4] (f) [MSO4] (g) [ESO4]

Figure 7.1 The structures of cations and anions.

180 7.2 Computational Methods

The double numerical plus polarization (DNP) basis set, which adds the polarized d- and p- function to non-hydrogen atoms and hydrogen atoms, separately, has been reported to be accurate for investigating IL interactions (Martínez-Magadán et al. 2012). The generalized gradient approximation (GGA) functional by Perdrew and Wang (PW91) (Perdrew et al. 1992) has been found to be suitable to study exchange-correlations (Cui et al. 2014). Therefore, here, geometry optimizations were performed by using the GGA-PW91 functional and DNP basis set in the Dmol3 module of the Material Studio 8.0 software package (Delley, 1990, 2000). Electron density, electrostatic potential, and population analysis were conducted at the same level. The electrostatic potential was mapped on the electron density isosurface.

In this study, the amorphous cell module was used to construct amorphous cubic periodic boxes for bulk ILs including 100 cations and 100 anions at temperature 298.15 K and pressure 101.325 kPa. In the simulations, the COMPASS II force field (Sun et al., 2016), which was used by many researchers to conduct IL molecular dynamics study, was employed (Sasikumar et al. 2015; Shin et al. 2014). In addition, the Ewald summation method and atom based method was adopted to study electrostatic and vdW interactions during the construction (Darden et al., 1993). The cells were relaxed in a 200 ps molecular dynamics run using the isothermal-isobaric (NPT) ensemble.

The Berendsen barostat and the Andersen thermostat were used for all molecular dynamics calculations (Allen & Wilson, 1989). This was followed by a 200 ps molecular dynamics production run with frames outputted every 100 fs into the trajectory file. Electrostatic and vdW terms were determined by the Ewald summation method with 0.0001 kcal/mol accuracy by using

181 a 15.5 Å cut-off value and 0.5 Å buffer width. The trajectories calculated by NPT were used to analyze the density. After the 200 ps equilibration run, 200 ps canonical (NVT) molecular dynamics simulations were performed, with sampling every 100 fs, and resulting trajectories utilized for self-diffusion parameters calculations.

7.3 Results and Discussion

7.3.1 Characterization of the Cations

Charges, electrostatic potential, and Mulliken bond orders (Mulliken 1955) for [BMIM]+ and

[BMPy]+ cations are displayed in Figure 7.2. The positive charge of [BMIM]+ is “formally” carried by the quaternary nitrogen atoms, which is consistent with the resonance structures (Hunt et al.,

2006). Judging from electrostatic potential legends in Figure 7.2, the electrostatic potential of

[BMIM]+ is more positive than [BMPy]+. Interestingly, the electrostatic potential for the imidazolium and pyridinium ring of [BMIM]+ and [BMPy]+ is more positive than that of the butyl chain (Figure 7.2). Therefore, two conclusions can be deduced from electrostatic potential analysis. One is that electrostatic interaction is the main interaction between cation and anion, and the other is that anions prefer to locate nearer the imidazolium and pyridinium rings than the butyl chain owing to the more positive electrostatic potential of the aromatic ring than that of the butyl chain. The Mulliken bond orders of the pyridinium ring are more uniformly distributed in contrast to those in the imidazolium ring (Figure 7.2), indicating that the aromaticity of pyridinium ring is stronger. Moreover, the harmonic oscillator measure of aromaticity (HOMA) values of [BMIM]+ and [BMPy]+ are calculated to be 0.8870 and 0.9730, respectively. The closer HOMA is to unity,

182 the stronger the aromaticity of the ring (Andrzejak et al., 2013), further confirming that pyridinium ring has stronger aromaticity.

(a) [BMIM]+

(b) [BMPy]+

Figure 7.2 Charges (black), Mulliken bond orders (red), and electrostatic potential of [BMIM]+ and [BMPy]+.

183 7.3.2 Geometric Structures

Some typical arrangements of cations and anions in a cubic periodic box are presented in Figure

7.3. It was deduced that the anions tend to be located near hydrogen atoms of the pyridinium and imidazolium rings, consistent with electrostatic potential analysis. The T-shaped and parallel- displaced π-π stacking conformations are identified in the IL cluster. In [BMIM]+-based ILs, the

π+-π+ stacking conformations are common, and anion-π+ stacking is relatively uncommon, as illustrated in Figure 7.3a-e. On the contrary, for [BMPy]+-based ILs, anion-π+ stacking is more

+ + conspicuous than π -π stacking (Figure 7.3f-g). Concerning [BMIM][HSO4] and [BMPy][HSO4],

- - two [HSO4] anions form an octatomic ring ([HSO4] dimer) through two strong O···H hydrogen bonds (Figure 7.3c, h). With regards to [BMIM][ESO4] and [BMPy][ESO4], the ethyl chain in the

- + anion interacts with the butyl chain in the cations, and the [SO4] part of [ESO4] forms anion-π stacking with the positive ring in the cation or connects the cations through hydrogen bonding

(Figure 7.3e, j).

(a) [BMIM][BF4] (b) [BMIM][PF6]

184

(e) [BMIM][ESO4] (f) [BMPy][BF4]

185 (g) [BMPy][BF6] (h) [BMPy][HSO4]

(i) [BMPy][MSO4] (j) [BMPy][ESO4]

Figure 7.3 Typical arrangement of cations and anions in ILs.

It is reported that some ILs with chains longer than (or equal to) the butyl chain could cause aggregation behavior (Singh and Kumar 2007; Wang et al., 2007). For cations with a butyl chain, the influence of anions on aggregation is pronounced. It was deduced from Figure 7.3 that

- - - - prominent aggregation behavior does not exist in [BF4] , [PF6] , [MSO4] -based ILs. For [ESO4] -

- based ILs, the interactions between nonpolar butyl chains and ethyl chains in [ESO4] cause

- aggregation (Figure 7.3e, j). The long alkyl chain of [ESO4] makes it possible for them to aggregate through nonpolar interactions, which is one type of vdW interaction. With regards to

- - [HSO4] -based ILs, the [HSO4] dimer is surrounded by pyridinium or imidazolium rings, leading to aggregation behavior (Figure 7.3c, h). According to earlier molecular simulation and experimental studies, aggregation not only exists in nonpolar domains, but also between polar parts and charged parts (Bernardes et al., 2014; Saielli et al., 2015; Shimizu et al., 2014). Hence, aggregation behavior is formed between polar pyridinium or imidazolium ring and charged

- [HSO4] dimer.

186

7.3.3 Intermolecular Interaction Energy

The intermolecular interaction energy has significant impacts on physicochemical properties of

ILs. The average intermolecular interactional energy was calculated by molecular dynamics, and the results are presented in Figure 7.4 and 7.5. To understand the nonpolar and polar interactions separately in this study, the total non-bonding energy is divided into vdW interaction energy and electrostatic energy (Zhang et al. 2003). It is deduced that the total non-bonding energy and vdW interaction energy of [BMIM]+-based ILs are more negative than that of corresponding [BMPy]+- based ILs, demonstrating that the intermolecular interaction strength of [BMIM]+-based ILs is stronger than that of [BMPy]+-based ILs, as shown in Figure 7.4 and 7.5. Furthermore, the electrostatic energy accounts for the vast majority of the total non-bonding energy. However, the electrostatic energy is more negative for [BMPy]+-based ILs than that for [BMIM]+-based ILs.

According to geometry analysis, anion-π+ interactions are less common in [BMIM]+-based ILs than that in [BMPy]+-based ILs. Therefore, despite the higher electrostatic potential of [BMIM]+, the electrostatic energy is more negative for [BMPy]+-based ILs compared to [BMIM]+-based ILs.

187

Figure 7.4 Intermolecular interaction energy of ILs.

As depicted in Figure 7.5, the differences of vdW interaction energy between different types of

ILs are obvious. Moreover, it suggests that the strength of vdW interactions in descending order

+ + - - - - - are [BMIM] >[BMPy] , [BF4] >[PF6] , and [ESO4] >[MSO4] >[HSO4] . According to geometry structure analysis, the anion dimers are prone to be surrounded by the [BMPy]+ cation for

+ [BMPy][HSO4], and anions in [BMIM][HSO4] tend to interact with cations through anion-π interactions. Both [BMPy][HSO4] and [BMIM][HSO4] have two strong hydrogen bonds formation and obvious aggregative behavior. Compared with other [BMIM]+ and [BMPy]+-based ILs structures, there is smaller variation in the structure arrangements for [BMPy][HSO4] and

[BMIM][HSO4], which could help explain the smaller electrostatic energy and vdW interaction energy differences between these two type of ILs.

188

Figure 7.5 vdW interaction energy of ILs.

7.3.4 Averaged Noncovalent Interaction Analyses

In 2010, Yang et al. constructed a noncovalent interaction index based on the study of reduced density gradient (RDG) as a function of electron density ρ(r) (Johnson et al., 2010). In 2013, their group developed an averaged noncovalent interaction (avgNCI) index along with a fluctuation index to characterize the magnitude of interactions and fluctuations (Wu et al., 2013). The averaged reduced density gradient (avgRDG) was defined by using the RDG of each single structure obtained from a dynamics trajectory. By plotting avgRDG with respect to sign(λ2)ρ(r) (effective

189 density), noncovalent interaction regions could be identified through analyzing spikes in the figure.

The peaks in the sign(2)<0 region suggest attractive interactions such as hydrogen bonds; whereas spikes in sign(2)=0 region indicate weak interactions such as vdW interactions. With regards to peaks in sign(2)>0 region, repulsive interaction is indicated. The strength of interaction can be reflected by the absolute value of effective density, and larger absolute value means stronger interaction. In addition, to visually observe the strength and position of those noncovalent interactions, the low-gradient (RDG=0.25) isosufaces were colored according to the corresponding values of effective density. The surfaces were colored on a blue-green-red scale with blue indicating strong attractive interactions (such as dipole-dipole or hydrogen bonding), green demonstrating weak interactions (such as vdW interactions), and red representing steric interactions.

The plots of RDG versus sign(2) and gradient isosurfaces are displayed in Figure 7.6 and Figure

7.7, respectively. The existences of spikes from -0.02 a.u. to 0.02 a.u. in Figure 7.6a-e indicate that vdW interaction and electrostatic interaction are present in all those ten type of ILs. It is noted that there are two prominent red spikes for [BMIM][HSO4] and two black spikes for [BMPy][HSO4] between 0.07 a.u. and 0.10 a.u. in Figure 7.6c, demonstrating the presence of two types of strong hydrogen bonds in both [BMIM][HSO4] and [BMPy][HSO4]. In addition, as depicted in Figure

- 7.7e and f, there are two blue circular regions in the [HSO4] dimers of both [BMIM][HSO4] and

[BMPy][HSO4] RDG isosurfaces, further confirming the presence of two types of strong hydrogen

- bonds in the [HSO4] dimers. The location of the two red spikes does not overlap (Figure 7.6c),

- which suggests that the strength of the two types of hydrogen bonds in the [HSO4] dimers of

[BMIM][HSO4] are different. Similarly, the strength of the two types of hydrogen bonds in the

190 - [HSO4] dimers of [BMPy][HSO4] are also dissimilar. However, the distance between the two black spikes is shorter than that of the two red ones, meaning that the difference of the strength of

- hydrogen bonds in the [HSO4] dimers of [BMPy][HSO4] is smaller than that of [BMIM][HSO4].

(a) [BMIM][BF4] and [BMPy][BF4] (b) [BMIM][PF6] and [BMPy][BF6]

191

(e) [BMIM][ESO4] and [BMPy][ESO4]

Figure 7.6 Plots of RDG versus sign(2) for ILs.

(a) [BMIM][BF4] (b) [BMPy][BF4]

192

(e) [BMIM][HSO4] (f) [BMPy][HSO4]

(g) [BMIM][MSO4] (h) [BMPy][MSO4]

193

(i) [BMIM][ESO4] (j) [BMPy][ESO4]

Figure 7.7 RDG isosurfaces for ILs. Note: Blue indicates strong attractive interactions and red indicates strong non-bonded overlap.

7.3.5 Densities

To investigate the influences of intermolecular and intramolecular interactions on the densities of

ILs, the densities of these ten types of ILs at room temperature calculated by the COMPASS II force field are shown in Figure 7.8. It was reported that the densities for [BMIM][BF4],

[BMIM][PF6], [BMIM][HSO4], [BMIM][MSO4], [BMPy][BF4], and [BMPy][MSO4], based on experiments, are 1.205, 1.367, 1.277, 1.212, 1.182, and 1.19 g cm-3, respectively (Aguirre et al.

2016; Ortega et al., 2007; William 2016). The calculated densities from the analysis conducted here are 1.226, 1.328, 1.263, 1.234, 1.205, and 1.211 g cm-3, respectively. The maximum deviation between experimental density and calculated value is 3%, indicating the reliability of the results of the analysis conducted here.

The densities of ILs with same anions and different cations are slightly different; whereas the densities of ILs with same cations and different anions have much larger variation (Figure 7.8).

194 Therefore, it is deduced that the influence of the anions on the density of the ILs is greater than

+ + - - that of the cations. In general, the densities follow the order of [BMIM] >[BMPy] , [PF6] >[BF4]

- - - , and [HSO4] >[MSO4] >[ESO4] (Figure 7.8). It is noted that the experimental densities of ILs follows the order of [BMIM][PF6]>[BMIM][BF4], [BMIM][HSO4]>[BMIM][MSO4],

[BMIM][PF6]>[BMPy][PF6], and [BMIM][MSO4]>[BMPy][MSO4], which has the same tendency with the simulation data and further confirms the accuracy of the predicted densities.

There are several factors accounting for the density trend. One factor that rationalizes the higher density of [BMIM]+-based ILs than [BMPy]+-based ILs is the different attractive non-bonded interactions. [BMIM]+-based ILs have higher interaction energy than [BMPy]+-based ILs.

+ + Therefore, the interactions between [BMIM] and anions are stronger than that between [BMPy] and the same anions, leading to the more compressed distribution of [BMIM]+ and anions.

Consequently, the densities of [BMIM]+-based ILs are larger than that of [BMPy]+-based ILs. The

- - other factor is the steric hindrance; the steric hindrance of [ESO4] is larger than that of [MSO4]

- - - - and [HSO4] . Hence the densities of ILs with [ESO4] anions are smaller than [MSO4] and [HSO4]

-based ILs. Furthermore, based on geometry and avgNCI analysis, strong hydrogen bonds are

- - formed in [HSO4] -based ILs, which can explain the higher density of [HSO4] -based ILs than

- - - [MSO4] and [ESO4] -based ILs. Additionally, the higher density of [PF6] -based ILs in comparison

- - - with [BF4] -based ILs is ascribed to the relatively larger molecular mass of [PF6] over that of [BF4]

. Interestingly, the density differences between [BMIM][HSO4]/[BMPy][HSO4] and

+ [BMIM][ESO4]/[BMPy][ESO4] are smaller than that between other [BMIM] -based ILs and

[BMPy]+-based ILs. Geometry analysis demonstrates the existence of aggregation in

195 [BMIM][HSO4], [BMIM][ESO4], [BMPy][HSO4], and [BMPy][ESO4], implying that aggregation effects influence the densities of ILs.

Figure 7.8 Calculated densities of ILs.

7.3.6 Self-diffusion Coefficients Analyses

The self-diffusion coefficient is important for understanding mass transfer rates as well as transport properties such as viscosity and ionic conductivity (Harris et al., 2014; Harris & Kanakubo, 2015;

Tsuzuki et al., 2011). In this study, the mean square displacement (MSD) was used to calculate the diffusivity coefficient. After 200 ps NVT dynamics, MSD versus time data set can be obtained through forcite analysis. The MSD data set, excluding the first 20 ps and last 20 ps time intervals, was used for linear regression where the self-diffusion coefficient is equal to the slope of the

196 regressed fit divided by 6. The self-diffusion coefficients calculated in the thesis are used for comparison, not for predictive purposes.

The effect of structures on self-diffusion coefficients is shown in Figure 7.9. For ILs with the same anions and different cations, it is depicted that the variation of self-diffusion coefficients for

[BMIM][BF4]/[BMPy][BF4] and [BMIM][PF6]/[BMPy][PF6] are more obvious, exhibiting that

- - [BF4] and [PF6] anions impact the self-diffusion coefficient significantly. The differences of self-

- - - - diffusion coefficients for [BMIM][HSO4] /[BMPy][HSO4] , [BMIM][MSO4] /[BMPy][MSO4] ,

- - - - and [BMIM][ESO4] /[BMPy][ESO4] are smaller, especially for [BMIM][HSO4] /[BMPy][HSO4]

- - - - and [BMIM][ESO4] /[BMPy][ESO4] . Based on geometry analysis, [HSO4] and [ESO4] -based

ILs have conspicuous aggregations, which can account for the smaller self-diffusion coefficients

- - - - differences for [BMIM][HSO4] /[BMPy][HSO4] and [BMIM][ESO4] /[BMPy][ESO4] . On other

- other hand, with the same cations, the self-diffusion coefficients for [HSO4] -based ILs are smaller than other anions-based ILs. According to geometry and avgRDG analyses, two prominent

- - hydrogen bonds exist in both [BMIM][HSO4] /[BMPy][HSO4] . The presence of hydrogen bonds

- induces IL clustering, thereby, leading to the lowest self-diffusion coefficients of [HSO4] -based

ILs among these ILs.

197

Figure 7.9 Self-diffusion coefficients of ILs. Note: 1 A2/ps = 10-8 m2/s.

7.4 Conclusions

Density functional theory and molecular dynamics simulation were used to explore the interactions, densities, and self-diffusion coefficients of [BMIM]+-based ILs and [BMPy]+-based

ILs. The electrostatic interaction energy accounts for the vast majority of total non-bonding interaction energy. From geometry analysis, it is concluded that vdW interactions, hydrogen bonds, electrostatic interactions, anion-π+， and π+-π+ stacking widely occur in these type of ILs.

- Two strong hydrogen bonds are found in the [HSO4] dimer through geometry and avgNCI analyses. Anions have greater influence on the density and self-diffusion coefficient than cations.

198 Aggregation behavior and hydrogen bond formation can alter the density and self-diffusion coefficients of ILs.

7.5 References

Aguirre, C. L., Toro, N., Carvajal, N., Watling, H., & Aguirre, C. (2016). Leaching of chalcopyrite

(CuFeS2) with an imidazolium-based ionic liquid in the presence of chloride. Minerals Engineering, 99, 60-66. Allen, M. P., & Wilson, M. R. (1989). Computer simulation of liquid crystals. Journal of Computer-Aided Molecular Design, 3(4), 335–353. Andrzejak, M., Kubisiak, P., & Zborowski, K. K. (2013). Avoiding pitfalls of a theoretical approach: The harmonic oscillator measure of aromaticity index from quantum chemistry calculations. Structural Chemistry, 24(4), 1171–1184. Bernardes, C. E. S., Shimizu, K., Lobo Ferreira, A. I. M. C., Santos, L. M. N. B. F., & Canongia Lopes, J. N. (2014). Structure and aggregation in the 1,3-dialkyl-imidazolium bis(trifluoromethylsulfonyl)imide ionic liquid family: 2. from single to double long alkyl side chains. Journal of Physical Chemistry B, 118(24), 6885–6895. Cao, Y., Zhang, R., Cheng, T., Guo, J., Xian, M., & Liu, H. (2017). Imidazolium-based ionic liquids for cellulose pretreatment: recent progresses and future perspectives. Applied Microbiology and Biotechnology, 101(2), 521–532. Červinka, C., Pádua, A. A. H., & Fulem, M. (2016). Thermodynamic properties of selected homologous series of ionic liquids calculated using molecular dynamics. The Journal of Physical Chemistry. B, 120, 2362–2371. Chen, L., & Bryantsev, V. S. (2017). A density functional theory based approach for predicting melting points of ionic liquids. Physical Chemistry Chemical Physics, 19(5), 4114–4124.

Cui, Y., Chen, Y., Deng, D., Ai., Ning, &Zhao, Y. (2014). Difference for the absorption of SO2

and CO2 on [Pnnnm][Tetz](n=1, m=2, and 4) ionic liquids: a density functional theory investigation. Journal of Molecular Liquids, 199, 7-14. Darden, T., York, D., & Pedersen, L. (1993). Particle mesh Ewald: An Nlog(N) method for Ewald sums in large systems. The Journal of Chemical Physics, 98(12), 10089–10092.

199 Delley, B. (1990). An all electron numerical method for solving the local density functional for polyatomic molecules. The Journal of Chemical Physics, 92(1), 508–517. Delley, B. (2000). From molecules to solids with the DMol3 approach. The Journal of Chemical Physics, 113(18), 7756–7764. Egorova, K. S., Gordeev, E. G., & Ananikov, V. P. (2017). Biological activity of ionic liquids and their application in pharmaceutics and medicine. Chemical Reviews, 117(10), 7132–7189. Fredlake, C. P., Crosthwaite, J. M., Hert, D. G., Aki, S. N. V. K., & Brennecke, J. F. (2004). Thermophysical properties of imidazolium-based ionic liquids. Journal of Chemical & Engineering Data, 49(4), 954–964. Fumino, K., Reimann, S., & Ludwig, R. (2014). Probing molecular interaction in ionic liquids by low frequency spectroscopy: Coulomb energy, hydrogen bonding and dispersion forces. Physical Chemistry Chemical Physics, 16(40), 21903–21929. Gardas, R. L., & Coutinho, J. A. P. (2008). A group contribution method for heat capacity estimation of ionic liquids. Industrial & Engineering Chemistry Research, 47, 5751–5757. Harris, K. R., & Kanakubo, M. (2015). Self-diffusion, velocity cross-correlation, distinct diffusion

and resistance coefficients of the ionic liquid [BMIM][Tf2N] at high pressure. Physical Chemistry Chemical Physics, 17(37), 23977–23993. Harris, K. R., Makino, T., & Kanakubo, M. (2014). Viscosity scaling of the self-diffusion and velocity cross-correlation coefficients of two functionalised ionic liquids and of their non- functionalized analogues. Physical Chemistry Chemical Physics, 16(19), 9161–9170. Herrera, C., García, G., Atilhan, M., & Aparicio, S. (2016). A molecular dynamics study on aminoacid-based ionic liquids. Journal of Molecular Liquids, 213, 201–212. Hou, X. D., Liu, Q. P., Smith, T. J., Li, N., & Zong, M. H. (2013). Evaluation of toxicity and biodegradability of cholinium amino acids ionic liquids. PloS one, 8(3), e59145. Hunt, P. A., Kirchner, B., & Welton, T. (2006). Characterising the electronic structure of ionic liquids: An examination of the 1-butyl-3-methylimidazolium chloride ion pair. Chemistry-A European Journal, 12(26), 6762–6775. Johnson, E. R., Keinan, S., Mori Sánchez, P., Contreras García, J., Cohen, A. J., & Yang, W. (2010). Revealing non-covalent interactions. Journal of the American Chemical Society, 132(18), 6498–6506.

200 Kurnia, K. A., Neves, C. M. S. S., Freire, M. G., Santos, L. M. N. B. F., & Coutinho, J. A. P. (2015). Comprehensive study on the impact of the cation alkyl side chain length on the solubility of water in ionic liquids. Journal of Molecular Liquids, 210, 264–271. Lin, J., Lü, R., Wu, C., Xiao, Y., Liang, F., & Famakinwa, T. (2017). A density functional theory study on the interactions between dibenzothiophene and tetrafluoroborate-based ionic liquids. Journal of Molecular Modeling, 23(4), 145. Lü, R., Wu, C., Lin, J., Xiao, Y., Wang, F., & Lu, Y. (2017). The study on interactions between 1- ethyl-3-methylimidazolium chloride and benzene/pyridine/pyrrole/thiophene. Journal of Molecular Liquids, 237, 289–294. Mahmood, H., Moniruzzaman, M., Yusup, S., & Welton, T. (2017). Ionic liquids assisted processing of renewable resources for the fabrication of biodegradable composite materials. Green Chemistry, 19(9), 2051–2075. Marium, M., Auni, A., Rahman, M. M., Mollah, M. Y. A., & Susan, M. A. B. H. (2017). Molecular level interactions between 1-ethyl-3-methylimidazolium methanesulphonate and water: Study of physicochemical properties with variation of temperature. Journal of Molecular Liquids, 225, 621–630. Martínez-Magadán, J. M., Oviedo-Roa, R., García, P., & Martínez-Palou, R. (2012). DFT study of the interaction between ethanethiol and Fe-containing ionic liquids for desulfuration of natural gasoline. Fuel Processing Technology, 97, 24-29. Mulliken, R. S. (1955). Electronic population analysis on LCAO–MO molecular wave functions. II. overlap populations, bond orders, and covalent bond energies. The Journal of Chemical Physics, 23(10), 1841–1846. Niedermeyer, H., Ashworth, C., Brandt, A., Welton, T., & Hunt, P. A. (2013). A step towards the a priori design of ionic liquids. Physical Chemistry Chemical Physics, 15(27), 11566–11578. Ortega, J., Vreekamp, R., Marrero, E., & Penco, E. (2007). Thermodynamic properties of 1-butyl- 3-methylpyridinium tetrafluoroborate and its mixtures with water and alkanols. Journal of Chemical and Engineering Data, 52(6), 2269–2276. Perdew, J. P., & Wang, Y. (1992). Accurate and simple analytic representation of the electron-gas correlation energy. Physical Review B, 45(23), 13244.

201 Saielli, G., Bagno, A., & Wang, Y. (2015). Insights on the isotropic-to-smectic a transition in ionic liquid crystals from coarse-grained molecular dynamics simulations: The role of microphase segregation. Journal of Physical Chemistry B, 119(9), 3829–3836. Sasikumar, Y., Adekunle, A. S., Olasunkanmi, L. O., Bahadur, I., Baskar, R., Kabanda, M. M., Obot, I.B., & Ebenso, E. E. (2015). Experimental, quantum chemical and Monte Carlo simulation studies on the corrosion inhibition of some alkyl imidazolium ionic liquids containing tetrafluoroborate anion on mild steel in acidic medium. Journal of Molecular Liquids, 211, 105-118. Shah, S. N., Kallidanthiyil Chellappan, L., Gonfa, G., Mutalib, M. I. A., Pilus, R. B. M., & Bustam, M. A. (2016). Extraction of naphthenic acid from highly acidic oil using phenolate based ionic liquids. Chemical Engineering Journal, 284, 487–493. Shimizu, K., Bernardes, C. E. S., & Canongia Lopes, J. N. (2014). Structure and aggregation in the 1-alkyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide ionic liquid homologous series. Journal of Physical Chemistry B, 118(2), 567–576. Shin, B. S., Kim, E. S., Kwak, S. K., Lim, J. S., Kim, K. S., & Kang, J. W. (2014). Thermodynamic inhibition effects of ionic liquids on the formation of condensed carbon dioxide hydrate. Fluid Phase Equilibria, 382, 270-278. Singh, T., & Kumar, A. (2007). Aggregation behavior of ionic liquids in aqueous solutions: effect of alkyl chain length, cations, and anions. The Journal of Physical Chemistry B, 111(27), 7843- 7851. Sun, H., Jin, Z., Yang, C., Akkermans, R. L. C., Robertson, S. H., Spenley, N. A., Miller, S., & Todd, S. M. (2016). COMPASS II: extended coverage for polymer and drug-like molecule databases. Journal of Molecular Modeling, 22(2), 1–10. Tsuzuki, S., Matsumoto, H., Shinoda, W., & Mikami, M. (2011). Effects of conformational flexibility of alkyl chains of cations on diffusion of ions in ionic liquids. Physical Chemistry Chemical Physics, 13(13), 5987–5993. Ventura, S. P. M., e Silva, F. A., Quental, M. V., Mondal, D., Freire, M. G., & Coutinho, J. A. P. (2017). Ionic-liquid-mediated extraction and separation processes for bioactive compounds: past, present, and future trends. Chemical Reviews, 117(10), 6984–7052.

202 Wang, J., Wang, H., Zhang, S., Zhang, H., & Zhao, Y. (2007). Conductivities, volumes,

fluorescence, and aggregation behavior of ionic liquids [C4mim][BF4] and [Cnmim]Br (n= 4, 6, 8, 10, 12) in aqueous solutions. Journal of Physical Chemistry B, 111(22), 6181–6188. Watanabe, M., Thomas, M. L., Zhang, S., Ueno, K., Yasuda, T., & Dokko, K. (2017). Application of ionic liquids to energy storage and conversion materials and devices. Chemical Reviews, 117(10), 7190–7239. William, H. M. (2016). CRC Handbook of Chemistry and Physics, 97th Edition. Boca Raton: CRC Press. Wu, C., De Visscher, A., & Gates, I. D. (2017). Molecular interactions between 1-butyl-3- methylimidazolium tetrafluoroborate and model naphthenic acids: A DFT study. Journal of Molecular Liquids, 243, 462–471. Wu, C., De Visscher, A., & Gates, I. D. (2018). Interactions of biodegradable ionic liquids with a model naphthenic acid. Scientific Reports, 8(1), 176. Wu, P., Chaudret, R., Hu, X., & Yang, W. (2013). Noncovalent interaction analysis in fluctuating environments. Journal of Chemical Theory and Computation, 9(5), 2226–2234. Yeganegi, S., Sokhanvaran, V., & Soltanabadi, A. (2013). Study of thermodynamic properties of imidazolium-based ionic liquids and investigation of the alkyl chain length effect by molecular dynamics simulation. Molecular Simulation, 39(13), 1070-1078. Zhang, M., Choi, P., & Sundararaj, U. (2003). Molecular dynamics and thermal analysis study of anomalous thermodynamic behavior of poly(ether imide)/polycarbonate blends. Polymer, 44(6), 1979-1986.

203 Chapter Eight: Conclusions and Recommendations

8.1 Conclusions

The objective of the research documented in this thesis was to study the removal of naphthenic acids (NAs) from oil sands process-affected water (OSPW) and crude oil through a theoretical study involving density functional theory (DFT). Ultrasonic irradiation is a potential new technique for the degradation of NAs. Hence, the reactions of benzoic acid (BA) and benzoate

(BZ) with hydroxyl radicals, especially produced by ultrasonic irradiation, were investigated. Ionic liquids (ILs) are potential solvents for the extraction of NAs from crude oil, therefore, the interactions between NAs and ILs were investigated. In particular, the extraction mechanisms of model NAs by 1-butyl-3-methylimidazolium tetrafluoroborate ([BMIM][BF4]) were explored. In addition, biodegradable ILs were studied for their potential to separate NAs, and the physicochemical properties of different ILs were compared through molecular dynamics simulation. The conclusions are listed as follows:

1. All the reactions between hydroxyl radicals and BA, BZ form pre-reactive complexes, which alters reaction energy barriers.

2. The total rate constants for the reactions between BA and OH radicals in the gas and aqueous phase are 2.72×10-11 cm3 molecule-1 s-1, and 1.21×109 M-1 s-1, respectively, whereas it is

9.16×109 M-1 s-1 for the reactions between BZ and hydroxyl radicals.

204 3. The rate constants for the addition reactions of OH adicals to BA and BZ in the aqueous phase are higher than those in the gas phase, whereas the hydrogen abstraction reaction rate constant is lower than in the gas phase.

4. [BMIM][BF4] mainly extracts NAs without long alkyl chain through hydrogen bonding, and separates NAs with long alkyl chain through van der Waals interaction and hydrogen bonding.

F···H hydrogen bonding is the strongest hydrogen bonds, and O···H is the second strongest for all the interactions.

5. [BMIM][BF4] has larger interaction energies with NAs having polycyclic hydrocarbons or multiple carboxylic groups than those with monocyclic hydrocarbons and one carboxylic group.

6. (C-O)-*(C-H), LP(O)-*(C-H), *(C-O)-*(C-H), and (C-H)-*(C-O) interactions can occur between [BMIM][BF4] and model NAs without aromatic ring, whereas π(C-O)-*(C-

H), LP(O)-*(C-H), *(C-O)-*(C-H), and (C-H)-*(C-O) interactions take place for

[BMIM][BF4]-BA.

7. After incorporating a biodegradable group, hydrogen atoms in -NH2, -COOH, and -

- COOCH3 form intramolecular hydrogen bonds with fluorine atoms in [BF4] , whereas hydrogen

- atoms in -OH and -OCH3 do not form hydrogen bonds with [BF4] .

8. [BMIM][BF4] with -COOH has stronger intramolecular hydrogen bonds than other ILs.

9. [BMIM][BF4] with -COOH or -COOCH3 form more hydrogen bonds with NAs, and have higher interaction energy with NAs.

10. Differences between ILs physical properties (density and self-diffusion coefficient) are smaller when there is an aggregation effect in ILs.

205 11. IL anions have greater influence on the density and self-diffusion coefficient than cations.

The intermolecular interaction strength of [BMIM]+-based ILs is stronger than that of [BMPy]+- based ILs.

12. van der Waals interactions, hydrogen bonds, electrostatic interactions, anion-π+, and π+-π+

- stacking widely occur in those ILs, and two strong hydrogen bonds are found in the [HSO4] dimer.

8.2 Recommendations

There are several limitations of the research reported in this thesis. Firstly, the calculated energy levels have an estimated uncertainty of 0.4 kcal mol-1, which causes the expected uncertainty on the calculated rate constants by the order of 2. Secondly, the extraction mechanism study was conducted between one IL molecule and one NAs, whereas different molecules can interact with each other in reality. In addition, the molecular dynamics simulation for self-diffusion coefficient of ILs was only used for comparison purposes. Based on the literature review, results, and limitations of the research, the following recommendations for future research are made to efficiently remove NAs from petroleum-based systems:

1. BA is not necessarily representative of nonaromatic compounds in NAs. Hence, rate constants and mechanisms studies about the reactions between nonaromatic compounds and hydroxyl radicals are necessary. Moreover, an experimental study to analyze the intermediates during reaction is beneficial for investigating the mechanism.

2. ILs are composed of anions and cations. To design suitable ILs for NAs separation, the influence of different anions and cations on the ILs extraction of NAs should be investigated.

206 3. The length of the alkyl chain of the cations impacts the biodegradation and physicochemical properties of ILs. Research to explore the different alkyl chain lengths on NAs isolation by ILs is recommended.

4. ILs with solid support materials are getting more attention recently. A comprehensive study should be conducted to compare the extraction differences of NAs by ILs with solid support materials.

5. Theoretical study alone is not enough for determining the suitable ILs for NAs separation, experimental study is strongly recommended to combine with theoretical study to provide better guidance for the designing of ILs.

6. DFT study is only suitable for studying interactions between countable molecules.

Nevertheless, interactions can happen between a lot of molecules in reality. Combining DFT study and molecular dynamics simulation can help better understand the intermolecular and intramolecular interactions.

7. With the exception of oxygen-containing compounds, crude oil also contains sulfur- and nitrogen-containing compounds. If ILs can extract oxygen-, sulfur-, and nitrogen-containing compounds at the same time, it will make significant contribution to oil refining industry. Hence, experimental and theoretical study to separate oxygen-, sulfur-, and nitrogen-containing compounds simultaneously is suggested. The elemental analysis can be used to investigate the extraction efficiency.

207 APPENDIX A: SUPPLEMENTARY MATERIALS FOR CHAPTER FOUR

(a) o-add (b) o2-add (c) m-add

(d) m2-add (e) p-add (f) H-abs

Fig. S1 Product complexes (Benzoic acid gas phase).

208

(a) o-add (b) o2-add (c) m-add

(d) m2-add (e) p-add (f) H-abs

Fig. S2 Product complexes (Benzoic acid aqueous phase).

209

(a) o-add (b) m-add (c) p-add

Fig. S3 Product complexes (Benzoate aqueous phase).

(a) o-add (b) m-add (c) p-add

210

(d) o2-add (e) m2-add (f) H-abs

Fig. S4 Pre-reactive complexes (Benzoic acid gas phase).

(a) o-add (b) m-add (c) p-add

211

(d) o2-add (e) m2-add (f) H-abs

Fig. S5 Pre-reactive complexes (Benzoic acid aqueous phase).

(a) o-add (b) m-add (e) p-add

Fig. S6 Pre-reactive complexes (Benzoate aqueous phase).

212

(a) o-add (b) o2-add (c) m-add

(d) m2-add (e) p-add (f) H-abs

Fig. S7 Transitions states with one explicit water molecule (Benzoic acid aqueous phase).

213

(a) o-add (b) o2-add (c) m-add

(d) m2-add (e) p-add

Fig. S8 Transitions states with one explicit water molecule (Benzoate aqueous phase).

214

Table S1 Relative energy (Re: kcal/mol) to isolated reactants for benzoic acid reacting with OH radicals in the gas phase. Species Re Species Re o-add pre-reactive complexes -2.85 m2-add pre-reactive complexes -3.04 o-add transition states 3.30 m2-add transition states 2.56 o-add adducts -18.92 m2-add adducts -15.88 o2-add pre-reactive complexes -5.46 p-add pre-reactive complexes -3.15 o2-add transition states 2.75 p-add transition states 2.38 o2-add adducts -20.34 p-add adducts -16.91 m-add pre-reactive complexes -3.17 H-abs pre-reactive complexes -2.87 m-add transition states 2.40 H-abs transition states 3.90 m-add adducts -16.15 H-abs adducts -11.67

Table S2 Relative energy (Re: kcal/mol) to isolated reactants for benzoic acid reacting with OH radicals in the aqueous phase. Species Re Species Re o-add pre-reactive complexes -0.56 m2-add pre-reactive complexes -1.99 o-add transition states 2.59 m2-add transition states 2.11 o-add adducts -17.83 m2-add adducts -14.75 o2-add pre-reactive complexes -2.61 p-add pre-reactive complexes -2.02 o2-add transition states 2.39 p-add transition states 2.46 o2-add adducts -18.30 p-add adducts -18.03 m-add pre-reactive complexes -2.02 H-abs pre-reactive complexes -2.87 m-add transition states 1.97 H-abs transition states 4.27 m-add adducts -14.83 H-abs adducts -2.99

Table S3 Relative energy (Re: kcal/mol) to isolated reactants for benzoate reacting with OH radicals in the aqueous phase. Species Re Species Re o-add pre-reactive complexes -5.80 p-add pre-reactive complexes -2.08 o-add transition states 0.15 p-add transition states 1.04 o-add adducts -18.43 p-add adducts -17.27 m-add pre-reactive complexes -2.20 m-add transition states 1.02 m-add adducts -15.2

215

Table S4 The values for open-shell systems for benzoic acid reacting with OH radicals in the gas phase. Species Species o-add pre-reactive complexes 0.7528 m2-add pre-reactive complexes 0.7532 o-add transition states 0.7866 m2-add transition states 0.7840 o-add adducts 0.7850 m2-add adducts 0.7875 o2-add pre-reactive complexes 0.7527 p-add pre-reactive complexes 0.7532 o2-add transition states 0.7871 p-add transition states 0.7861 o2-add adducts 0.7841 p-add adducts 0.7796 m-add pre-reactive complexes 0.7532 H-abs pre-reactive complexes 0.7528 m-add transition states 0.7839 H-abs transition states 0.7596 m-add adducts 0.7866 H-abs adducts 0.7598

Table S5 The values for open-shell systems for benzoic acid reacting with OH radicals in the aqueous phase. Species Species o-add pre-reactive complexes 0.7527 m2-add pre-reactive complexes 0.7535 o-add transition states 0.7817 m2-add transition states 0.7797 o-add adducts 0.7830 m2-add adducts 0.7877 o2-add pre-reactive complexes 0.7527 p-add pre-reactive complexes 0.7535 o2-add transition states 0.7817 p-add transition states 0.7824 o2-add adducts 0.7827 p-add adducts 0.7783 m-add pre-reactive complexes 0.7535 H-abs pre-reactive complexes 0.7533 m-add transition states 0.7796 H-abs transition states 0.7592 m-add adducts 0.7871 H-abs transition states 0.7597

Table S6 The values for open-shell systems for benzoate reacting with OH radicals in the aqueous phase. Species Species o-add pre-reactive complexes 0.7529 p-add pre-reactive complexes 0.7538 o-add transition states 0.7802 p-add transition states 0.7784 o-add adducts 0.7834 p-add adducts 0.7789 m-add pre-reactive complexes 0.7538 m-add transition states 0.7777 m-add adducts 0.7844

216

Table S7 Reaction rate constants of benzoic acid (BA) and benzoate (BZ) with hydroxyl radical by M06-2X method. Reaction path BA gas phase BA aqueous phase BZ aqueous phase k 훤 k 훤 k 훤 o-add 3.81×10-16 1.343 1.38×106 1.227 2.54×107 1.186 o2-add 1.32×10-15 1.363 9.15×105 1.223 2.54×107 1.186 m-add 2.92×10-15 1.278 3.35×106 1.165 2.10×107 1.122 m2-add 2.23×10-15 1.281 2.00×106 1.166 2.10×107 1.122 p-add 3.27×10-15 1.303 2.00×106 1.212 2.19×107 1.138 H-abs 2.03×10-13 2.992 3.30×104 3.087 Note: Units for gas phase rate constants and aqueous phase rate constants are cm3 molecule-1 s-1, and M-1 s-1, respectively.

Table S8 Reaction rate constants of benzoic acid (BA) and benzoate (BZ) with hydroxyl radical by CCSD(T) correction. Reaction BA gas phase BA aqueous phase BZ aqueous phase path (cm3 molecule-1 s-1) (M-1 s-1) (M-1 s-1) o-add 7.09×10-17 7.56×104 2.59×106 o2-add 1.28×10-16 5.92×104 2.59×106 m-add 2.60×10-16 1.20×105 4.79×105 m2-add 1.87×10-16 6.89×104 4.79×105 p-add 2.76×10-16 9.34×104 4.61×105 H-abs 2.87×10-17 2.44×101

Table S9 Reaction rate constants of BA and BZ with hydroxyl radical with one explicit water molecule by M06-2X method. Reaction path BA aqueous phase BZ aqueous phase k (M-1 s-1) 훤 k (M-1 s-1) 훤 o-add 5.79×105 1.215 5.37×106 1.188 o2-add 1.07×106 1.210 3.66×106 1.181 m-add 4.00×106 1.158 4.80×106 1.124 m2-add 2.03×106 1.160 7.49×106 1.125 p-add 6.56×105 1.199 2.98×106 1.139 H-abs 1.55×103 3.118

217

APPENDIX B: SUPPLEMENTARY MATERIALS FOR CHAPTER FIVE

[BMIM][BF4] HOMO+CHCA LUMO [BMIM][BF4] LUMO+CHCA HOMO Overlap integral = 0.01992 Overlap integral = 0.1314

[BMIM][BF4] HOMO+CPCA LUMO [BMIM][BF4] LUMO+CPCA HOMO Overlap integral = 0.007860 Overlap integral = 0.003406

[BMIM][BF4] HOMO+BA LUMO [BMIM][BF4] LUMO+BA HOMO Overlap integral = 0.001283 Overlap integral = -0.0005741

218

[BMIM][BF4] HOMO+CHPA LUMO [BMIM][BF4] LUMO+CHPA HOMO Overlap integral = -0.01714 Overlap integral = 0.001966

[BMIM][BF4] HOMO+CHDCA LUMO [BMIM][BF4] LUMO+CHDCA HOMO Overlap integral = -0.01607 Overlap integral = 0.004140

[BMIM][BF4] HOMO+DCHA LUMO [BMIM][BF4] LUMO+DCHA HOMO Overlap integral = 0.009666 Overlap integral = -0.1060

Fig. S1 The overlap integral of HOMO of fragment1 and LUMO of fragment2 as well as LUMO of fragment1 and HOMO of fragment2. Note: isosurface is 0.000005 a.u. Unit for integral area is a.u.

219

(a) [BMIM][BF4]-CHCA

(b) [BMIM][BF4]-CPCA

220

(d) [BMIM][BF4]-CHPA

221

(e) [BMIM][BF4]-CHDCA

(f) [BMIM][BF4]-DCHA

Fig. S2 The different electron densities of [BMIM][BF4] and model NAs. Note: the purple and green isosurfaces represent the region in which electron density is increased and decreased after interaction between cation and anion of ionic liquids and model NAs, respectively.

222

Table S1 Cartesian co-ordinates for optimized structures of [BMIM][BF4]-CHCA, [BMIM][BF4]-CPCA, [BMIM][BF4]-BA, [BMIM][BF4]-CHPA, [BMIM][BF4]-CHDCA, and [BMIM][BF4]-DCHA.

[BMIM][BF4]-CHCA [BMIM][BF4]-CPCA C 4.72578700 -0.23122600 0.80458200 N -1.99451800 -1.06120000 -1.23569400 C 6.14597300 0.30233000 0.61123000 C -2.69785900 -2.22011400 -0.99322700 C 6.45887500 0.52986700 -0.86887500 C -1.85652500 -3.06579100 -0.34657100 C 5.43070700 1.46532500 -1.50788900 N -0.65547800 -2.40796100 -0.20912800 C 4.00901400 0.93173600 -1.32476000 C -0.76847700 -1.19570400 -0.74110900 C 3.69514500 0.71165200 0.15386300 C 0.50367500 -2.86968500 0.55081100 H 7.46691400 0.93666300 -0.98483000 C -2.53950200 0.18980500 -1.77933200 H 6.24779100 1.25092800 1.15200800 C -3.78485700 0.63211600 -1.01951100 H 6.86545300 -0.39305000 1.05151700 C -4.14472400 2.07530400 -1.36309700 H 4.63949600 -1.21968300 0.33551700 C -5.41474700 2.53001600 -0.64981200 H 4.49605900 -0.35111300 1.86532100 H -3.72733500 -2.33557600 -1.28238800 H 5.51033300 2.45690000 -1.04540200 H -2.00947700 -4.05902500 0.03805300 H 5.64469200 1.59610900 -2.57178000 H 0.01444200 -0.44936500 -0.72879400 H 3.27793700 1.61620800 -1.76265700 H 1.33052900 -2.19436400 0.33958800 H 3.89862300 -0.02266100 -1.85064500 H 0.74999400 -3.88483300 0.24238300 H 3.73083400 1.66679000 0.69167800 H 0.26563900 -2.82466100 1.61232200 H 6.44108800 -0.43393300 -1.39231900 H -2.74256200 0.04457700 -2.84338200 C 2.30941700 0.14702700 0.35619800 H -1.75186500 0.93429000 -1.66531000 O 1.64099200 -0.34913800 -0.52958700 H -3.58583700 0.53860500 0.05052000 O 1.91186000 0.21742700 1.61436900 H -4.62697800 -0.02624300 -1.26239500 H 1.01574700 -0.17622000 1.70774300 H -4.26847300 2.18155600 -2.44723200 N -2.46067000 -0.60026000 -1.49029600 H -3.30825200 2.71825800 -1.07215800 C -3.31266200 -1.67912600 -1.57364700 H -5.65227400 3.56867100 -0.88629900 C -2.61243500 -2.76887100 -1.16995700 H -5.29534000 2.44939500 0.43304700 N -1.34600700 -2.33529400 -0.85042600 H -6.27017300 1.91464500 -0.94072800 C -1.28377600 -1.02161300 -1.03884400 F -1.25309500 1.13459300 0.71143900 C -0.28828500 -3.13084200 -0.23172300 B -1.29309800 0.30111000 1.85303200 C -2.82543400 0.81099000 -1.67495900 F -2.24323700 -0.70873600 1.65061500 C -4.02654900 1.19014100 -0.81606500 F -1.53010200 1.02011400 2.99449700 C -4.19057900 2.70619000 -0.74526400 F 0.00202000 -0.33728100 1.93224400 C -5.40795600 3.10500300 0.08432600 C 5.26064800 -0.49067900 -0.29668100 H -4.33420000 -1.57826400 -1.89470900 C 5.05822300 0.76431900 0.55980000 H -2.90410900 -3.79970600 -1.07041500 C 3.89484700 1.48716100 -0.15374800 H -0.41563400 -0.41586100 -0.81854600 C 4.23998200 1.32199500 -1.63849100 H 0.62971600 -2.54767500 -0.26367400 C 5.00206800 -0.02475700 -1.74833000 H -0.17158600 -4.05960700 -0.78849600 H 6.24955400 -0.93184800 -0.16442600 H -0.55647800 -3.32324000 0.80602200 H 4.52481200 -1.24764600 -0.01188700 H -3.01414900 0.98312400 -2.73760000 H 5.94192800 1.40703900 0.51100400 H -1.95288800 1.38808600 -1.37006900 H 4.84790400 0.55614200 1.60976200 H -3.87494900 0.78191400 0.18562500 H 3.81232800 2.52730600 0.16617300 H -4.93807700 0.74021800 -1.22620400 H 3.34755600 1.34363700 -2.26359200 H -4.27682000 3.12127100 -1.75620000 H 4.88306500 2.15075200 -1.94183800 H -3.28593900 3.13174400 -0.30016000 H 4.41794500 -0.76378500 -2.29708400 H -5.50965400 4.18995200 0.14430500 H 5.94206800 0.11156500 -2.28628900 H -5.32060700 2.71695900 1.10160700 C 2.60091300 0.79430000 0.20591800

223

H -6.32745300 2.70432000 -0.35051200 O 2.02344100 0.00901400 -0.52194000 F -1.51142800 0.92133600 0.96952100 O 2.18522200 1.10142600 1.42083600 B -1.68117800 -0.17213500 1.84887900 H 1.35469600 0.61803500 1.62950400 F -2.74102900 -0.96668200 1.39042900 F -1.86143300 0.24267000 3.14182600 F -0.47675100 -0.96492800 1.74888800 [BMIM][BF4]-BA [BMIM][BF4]-CHDCA N 2.35631500 0.77093600 -1.35335100 N -2.44164300 1.68715400 -0.83545400 C 3.30657900 1.75892500 -1.22088300 C -3.41841300 2.51051700 -0.32059600 C 2.69156200 2.82591900 -0.65166600 C -4.55826100 1.77731200 -0.25786600 N 1.37755400 2.47031600 -0.44936500 N -4.25860900 0.52456600 -0.74160100 C 1.20365500 1.22180900 -0.86910400 C -2.96997900 0.48875500 -1.06738800 C 0.37228500 3.25026000 0.26922500 C -5.15821300 -0.62655600 -0.71315600 C 2.59774700 -0.61515500 -1.77704600 C -1.03040500 2.03309000 -1.05088500 C 3.73606000 -1.24576000 -0.98313100 C -0.32644000 2.43314400 0.23922300 C 3.75703400 -2.76062100 -1.17089700 C 1.13503200 2.78189300 -0.04522600 C 4.92505700 -3.40808800 -0.43248800 C 1.96080200 2.90141300 1.23270400 H 4.32838400 1.61446200 -1.52379500 H -3.21992300 3.53149200 -0.04544700 H 3.07197000 3.78938700 -0.36065500 H -5.54190300 2.02749200 0.09932100 H 0.27487600 0.67507000 -0.78776600 H -2.41275400 -0.37138800 -1.41576800 H -0.59351700 2.77266200 0.11790800 H -4.71670200 -1.42511200 -1.30256200 H 0.36357200 4.26533700 -0.12545800 H -6.11893400 -0.33089700 -1.13211800 H 0.61668800 3.24148700 1.33029300 H -5.25425800 -0.96012000 0.31857800 H 2.80098000 -0.61821100 -2.85089200 H -0.99424200 2.83697200 -1.79133600 H 1.66854700 -1.15260600 -1.58869900 H -0.56441800 1.14177400 -1.47604200 H 3.59766700 -1.00006900 0.07234200 H -0.39902100 1.59927700 0.94151900 H 4.69504500 -0.82152900 -1.30274300 H -0.82641100 3.29188900 0.70022100 H 3.81085500 -3.00423700 -2.23840900 H 1.18887900 3.71439200 -0.61711900 H 2.81260100 -3.16785200 -0.79692100 H 1.56452700 2.00491300 -0.68750600 H 4.92002500 -4.49234300 -0.55592600 H 3.00022800 3.15836300 1.01871400 H 4.87076100 -3.18986800 0.63641200 H 1.95594200 1.95702700 1.78413000 H 5.88148200 -3.03150500 -0.80483100 H 1.55126600 3.67112800 1.89168900 F 1.19386400 -1.03051500 0.77316700 F -1.32280200 -0.43987600 1.00572200 B 1.42772400 -0.13744000 1.84386200 B -2.42236300 -1.21131100 1.44648900 F 2.56621200 0.62768300 1.56115800 F -3.48142000 -0.36124200 1.77910800 F 1.52352800 -0.79795300 3.03994600 F -2.06633300 -2.04483400 2.47428900 F 0.29765600 0.76510600 1.86886100 F -2.84326500 -2.00064700 0.30504100 C -5.97089300 -1.34812500 0.38908300 C 3.93914200 0.00315100 -0.15504400 C -4.68498800 -0.97935000 0.76631100 C 3.21519600 -0.76026800 0.97032200 C -3.89044900 -0.24852900 -0.11641400 C 1.74639900 -0.98171700 0.61676300 C -4.38098200 0.11197100 -1.37025900 C 1.60883100 -1.71042900 -0.73483100 C -5.66640500 -0.25702900 -1.74278300 C 2.32783200 -0.94534000 -1.84276600 C -6.46083800 -0.98739400 -0.86249400 C 3.79946200 -0.72603300 -1.49016400 H -6.59066300 -1.91777500 1.07096000 H 1.23734700 -1.54903400 1.39809400 H -4.28949500 -1.25338700 1.73610000 H 3.70954000 -1.72857600 1.11365100 H -3.74234600 0.67997500 -2.03578100 H 3.30027700 -0.20819500 1.90899100 H -6.05004300 0.02322400 -2.71634100 H 3.49229700 1.00202900 -0.22174800 H -7.46412500 -1.27657600 -1.15323200 H 2.04003200 -2.71182200 -0.62564900 C -2.50565300 0.16792400 0.24691500 H 2.24165300 -1.48390200 -2.78970400 O -1.78664200 0.79231400 -0.51061800 H 1.82615200 0.01802300 -1.98702000 O -2.14391600 -0.19743800 1.46077200 H 4.31755500 -1.68784600 -1.41795900 224

H -1.22446000 0.10433900 1.64151000 H 4.30265100 -0.16182500 -2.27889900 H 1.22896200 -0.01744700 0.54598200 C 0.13534000 -1.87789800 -1.00919500 O -0.50987700 -1.15409200 -1.73973200 O -0.38659100 -2.88124100 -0.32192700 H -1.35859200 -2.78298300 -0.28510600 C 5.38666400 0.19726700 0.21846300 O 6.32629400 -0.37069400 -0.26676700 O 5.52776400 1.09680800 1.21697500 H 6.46914600 1.14225200 1.43204100 [BMIM][BF4]-CHPA [BMIM][BF4]-DCHA N -2.73004400 1.73743300 -0.82778900 C 2.54570000 0.28433500 0.75166300 C -3.58734600 2.76552500 -0.50476400 H 2.84294300 0.34616800 1.80420000 C -4.84611900 2.31564900 -0.73946400 C 2.13573200 1.70895700 0.30371800 N -4.73271000 1.02653200 -1.20753100 C 1.91001100 1.86094800 -1.20570200 C -3.44510000 0.69580100 -1.23734300 C 0.90084100 2.20031800 1.07383200 C -5.83461100 0.08074100 -1.35796500 H 2.97102500 2.36627000 0.57996000 C -1.27589800 1.69653900 -0.62001800 C 1.57589900 3.31136800 -1.56441800 C -0.89179000 2.11644600 0.79494900 H 1.09015500 1.20286900 -1.51519100 C 0.49166700 1.57246500 1.14881500 H 2.79548800 1.54081400 -1.76193500 C 0.97504400 2.05661000 2.51162600 C 0.55618000 3.64686700 0.71408200 H -3.23283500 3.70921200 -0.12868400 H 0.04144700 1.55489000 0.84236700 H -5.80328300 2.78661900 -0.59782200 H 1.06973900 2.10289400 2.15081900 H -3.03233800 -0.27280700 -1.49102300 C 0.35216100 3.80597000 -0.79310200 H -5.48169200 -0.77782700 -1.92349900 H 1.41049900 3.40218400 -2.64179000 H -6.65207500 0.56972000 -1.88563200 H 2.43538700 3.94914500 -1.32184200 H -6.14008700 -0.24809300 -0.36532900 H -0.33823600 3.96554600 1.25710600 H -0.80049100 2.32891900 -1.37408600 H 1.37258400 4.30506500 1.03639200 H -0.99204400 0.65948400 -0.79966600 H 0.13501300 4.84874300 -1.04238800 H -1.62661800 1.70647900 1.49162200 H -0.52187700 3.21768300 -1.09899800 H -0.90171400 3.20845500 0.88771700 C 3.70285300 -0.34287600 -0.04582400 H 1.21695500 1.85136900 0.37365700 C 4.94603800 0.55365500 -0.05452300 H 0.42785000 0.47951700 1.14466400 C 4.06047700 -1.72435800 0.52042100 H 1.95169400 1.63473300 2.75706600 H 3.36284700 -0.48578300 -1.07950500 H 0.27360700 1.76043100 3.29503300 C 6.10505700 -0.09413000 -0.81655300 H 1.06480600 3.14584300 2.53292900 H 5.25128600 0.74164100 0.98428800 F -2.20858700 -0.55100500 1.07098700 H 4.71703700 1.52688100 -0.49704900 B -3.54388800 -0.96177600 1.30618400 C 5.21292200 -2.37409700 -0.24649100 F -4.35633300 0.17730400 1.40608500 H 4.34279700 -1.61053500 1.57567600 F -3.63266900 -1.76754700 2.41005700 H 3.18292200 -2.37822700 0.49787500 F -3.96896700 -1.68834900 0.13214500 C 6.44757600 -1.47258500 -0.25080500 C 7.06749600 -0.62226500 0.91730300 H 6.98147500 0.55921600 -0.78929000 C 5.78374400 -1.05023900 0.20221600 H 5.82178600 -0.19965100 -1.87134200 C 4.94889500 0.15382400 -0.24916200 H 5.45138400 -3.34787200 0.18977900 C 5.80458200 1.10033100 -1.09910100 H 4.89724200 -2.55740200 -1.28096100 C 7.08340200 1.53327200 -0.38063500 H 7.25514800 -1.93163000 -0.82768800 C 7.90492200 0.31656100 0.04794400 H 6.81420500 -1.35916300 0.77679800 H 6.04176600 -1.64716200 -0.68407700 C 1.34507300 -0.63474500 0.68953800 H 5.19391000 -1.69569500 0.85855100 O 0.83466900 -0.99544500 -0.35459700 H 6.80309700 -0.10610700 1.84864100 O 0.90524100 -1.01221900 1.87612100 H 7.65096400 -1.50282700 1.19982200 H 0.07969400 -1.53172600 1.77074600 225

H 6.07352800 0.58337300 -2.03046900 N -2.81008500 -0.48936400 -1.66764400 H 5.21239100 1.97580000 -1.38528500 C -3.98114500 -1.06194900 -2.11421200 H 7.67684400 2.18720200 -1.02557900 C -3.84060500 -2.40473100 -1.97887300 H 6.81808100 2.11899500 0.50816500 N -2.58614400 -2.62553000 -1.45630500 H 8.23927100 -0.22403400 -0.84636600 C -1.99114100 -1.45314900 -1.26244100 H 8.80427700 0.63118500 0.58435800 C -2.07931600 -3.89203700 -0.93632700 H 4.63396000 0.70139300 0.65282900 C -2.56770500 0.94216300 -1.43995500 C 3.69056800 -0.26183200 -1.01463500 C -3.65667600 1.54380200 -0.55570000 C 2.73116100 -1.16009900 -0.23580600 C -3.18628200 2.84648400 0.08549600 H 3.98884600 -0.76814100 -1.94233200 C -4.25607900 3.45595700 0.98635900 H 3.15384300 0.64538300 -1.32111900 H -4.80667800 -0.47203200 -2.47184900 C 1.40009400 -1.34825600 -0.95851100 H -4.51955300 -3.21162200 -2.19283500 H 3.18943100 -2.13898100 -0.05957500 H -1.01420500 -1.30986500 -0.81232700 H 2.54671200 -0.72163300 0.75433900 H -1.00690700 -3.79322700 -0.78273400 C 0.44086100 -2.23756100 -0.18032900 H -2.28721300 -4.68327000 -1.65490500 H 0.93675300 -0.37093600 -1.12265300 H -2.55637500 -4.08820400 0.02325900 H 1.56981000 -1.77088300 -1.95436500 H -2.49515400 1.44341700 -2.40841200 H 0.79699600 -3.26963200 -0.11652000 H -1.60067700 1.00152700 -0.93738800 H 0.33530700 -1.89381900 0.85637700 H -3.89649800 0.81953700 0.22601500 C -0.96157800 -2.24872800 -0.72664900 H -4.56546000 1.72067100 -1.14262800 O -1.36888700 -1.49306800 -1.58292500 H -2.89597500 3.56569700 -0.68936000 O -1.71914600 -3.16552200 -0.14026400 H -2.28882200 2.62682500 0.67352700 H -2.65549700 -2.92802200 -0.27414000 H -3.90363700 4.37953300 1.44916300 H -4.52593200 2.76028200 1.78414900 H -5.16186200 3.68873900 0.42001700 F -1.88076600 -0.03770000 1.22793400 B -2.45690800 -1.23327600 1.72479700 F -3.59542400 -1.54180800 0.96514300 F -2.73473200 -1.13924400 3.06207400 F -1.48058800 -2.26803400 1.48803200

226

Table S2 Some donor-acceptor interactions in the structures and their stabilization energies, E(2) (kcal/mol) of [BMIM][BF4]-CHCA, [BMIM][BF4]-CPCA, [BMIM][BF4]-BA, [BMIM][BF4]- CHPA, [BMIM][BF4]-CHDCA, and [BMIM][BF4]-DCHA. Donor Acceptor E(2) [BMIM][BF4]-CHCA From CHCA to [BMIM]+ (C6-C18) *(C26-H34) 0.25 (C18-O19) *(C26-H34) 0.77 (C18-O19) *(C27-H35) 0.43 LP(O19) *(N22-C26) 0.25 LP(O19) *(C26-H34) 5.55 LP(O19) *(C27-H35) 0.78 *(C18-O19) *(C26-H34) 0.18 *(C18-O19) *(C27-H35) 0.06 From [BMIM]+ to CHCA (C26-H34) *(C18-O19) 0.08 (C27-H35) *(C18-O19) 0.08 - From CHCA to [BF4] (O20-H21) *(B48-F51) 0.07 LP(O20) *(B48-F51) 0.10 - From [BF4] to CHCA (B48-F51) *(O20-H21) 0.09 LP(F51) *(O20-H21) 20.18 [BMIM][BF4]-CPCA From [BMIM]+ to CPCA  (C5-H13) *(C45-O46) 0.07  (C6-H14) *(C45-O46) 0.07 From CPCA to [BMIM]+ (C33-C45) *(C5-H13) 0.27 (C45-O46) *(C5-H13) 0.76 (C45-O46) *(C6-H14) 0.40 LP(O46) *(N1-C5) 0.24 LP(O46) *(C5-H13) 5.81 LP(O46) *(C6-H14) 0.69 *(C45-O46) *(C5-H13) 0.15 *(C45-O46) *(C6-H14) 0.06 - From [BF4] to CPCA (B27-F30) *(O47-H48) 0.12 LP(F30) *(O47-H48) 20.62 - From CPCA to [BF4] (O47-H48) *(B27-F30) 0.07 LP(O47) *(B27-F30) 0.09 [BMIM][BF4]-BA

227

From [BMIM]+ to BA (C5-H13) *(C42-O43) 0.08 (C6-H14) *(C42-O43) 0.07 From BA to [BMIM]+ (C33-C42) *(C5-H13) 0.22 (C42-O43) *(C5-H13) 0.84 (C42-O43) *(C6-H14) 0.43 LP(O43) *(N1-C5) 0.25 LP(O43) *(C5-H13) 5.30 LP(O43) *(C6-H14) 0.81 *(C42-O43) *(C5-H13) 0.16 *(C42-O43) *(C6-H14) 0.08 - From [BF4] to BA LP(F30) *(O44-H45) 21.55 - From BA to [BF4] (C42-O44) *(B27) 0.05 (O44-H45) *(B27) 0.06 LP(O44) *(B27) 0.11 [BMIM][BF4]-CHPA From [BMIM]+ to CHPA (C5-H13) *(C60-O61) 0.23 (C7-H18) *(C52-H56) 0.23 (C7-H18) *(C60-O61) 0.08 (C9-H21) *(C52-H56) 0.10 (C9-H22) *(C49-H54) 0.09 (C9-H22) *(C52-H57) 0.07 (C9-H22) *(C55-H59) 0.11 From CHPA to [BMIM]+ (C48-H51) *(C9-H21) 0.09 (C49-H54) *(C9-H22) 0.08 (C52-H56) *(C7-H18) 0.41 (C55-H59) *(C9-H22) 0.28 (C55-C60) *(C5-H13) 0.28 (C60-O61) *(C5-H13) 0.86 (C60-O61) *(N4-C5) 0.10 (C60-O61) *(C7-H18) 0.85 LP(O61) *(N4-C5) 0.09 LP(O61) *(C5-H13) 5.6 LP(O61) *(C7-H18) 1.39 *(C60-O61) *(C5-H13) 0.14 *(C60-O61) *(C7-H18) 0.06 - From [BF4] to CHPA LP(F26) *(C55-H59) 0.06 228

LP(F26) *(C60-O61) 1.59 LP(F26) *(C60-O62) 0.14 LP(F30) *(O62-H63) 9.12 - From CHPA to [BF4] (O62-H63) *(B27-F28) 0.05 (O62-H63) *(B27-F30) 0.06 LP(O62) *(F26-B27) 0.07 LP(O62) *(B27-F28) 0.07 LP(O62) *(B27-F30) 0.09 [BMIM][BF4]-CHDCA From [BMIM]+ to CHDCA (C5-H13) *(C47-O48) 0.14 (C7-H18) *(C35-H43) 0.07 (C7-H18) *(C47-O48) 0.12 (C8-H19) *(C33-H46) 0.25 (C9-H22) *(C31-H40) 0.28 (C9-H22) *(C33-H46) 0.14 (C9-H22) *(C35-H43) 0.20 (C10-H23) *(C31-H40) 0.18 (C10-H24) *(C33-H46) 0.13 From CHDCA to [BMIM]+ (C31-H40) *(C9-H22) 0.25 (C31-H40) *(C10-H25) 0.10 (C32-H39) *(C10-H24) 0.07 (C33-H46) *(C8-H19) 0.13 (C33-H46) *(C8-H20) 0.05 (C33-H46) *(C9-H21) 0.08 (C33-H46) *(C9-H22) 0.08 (C34-C47) *(C5-H13) 0.27 (C35-H43) *(C7-H18) 0.10 (C35-H43) *(C9-H22) 0.31 (C47-O48) *(C5-H13) 0.51 (C47-O48) *(N4-C5) 0.08 (C47-O48) *(C7-H18) 0.47 LP(O48) *(N4-C5) 0.11 LP(O48) *(C5-H13) 5.88 LP(O48) *(C7-H18) 1.66 *(C47-O48) *(C5-H13) 0.06 *(C47-O48) *(C7-H18) 0.07 - From [BF4] to CHDCA (B27-F30) *(O49-H50) 0.12 LP(F26) *(C32-C33) 0.20

229

LP(F26) *(C33-H46) 0.14 LP(F26) *(C47-O48) 0.80 LP(F26) *(C47-O49) 0.09 LP(F30) *(O49-H50) 13.03 - From CHDCA to [BF4] (O49-H50) *(B27-F28) 0.06 (O49-H50) *(B27-F30) 0.07 LP(O49) *(F26-B27) 0.09 LP(O49) *(B27-F28) 0.08 LP(O49) *(B27-F29) 0.06 LP(O49) *(B27-F30) 0.08 [BMIM][BF4]-DCHA From [BMIM]+ to DCHA (C45-H53) *(C37-O38) 0.23 (C45-H53) *(C4-H8) 0.05 (C47-H58) *(C5-H11) 0.14 (C47-H58) *(C13-H19) 0.15 (C49-H61) *(C13-H19) 0.20 (C49-H62) *(C5-H11) 0.13 (C49-H62) *(C10-H16) 0.06 (C49-H62) *(C10-H17) 0.06 (C49-H62) *(C13-H19) 0.09 From DCHA to [BMIM]+ (C1-C37) *(C45-H53) 0.45 (C4-H8) *(C47-H58) 0.10 (C5-H11) *(C47-H58) 0.24 (C5-H11) *(C49-H62) 0.08 (C10-H16) *(C49-H62) 0.17 (C13-H19) *(N41-C47) 0.06 (C13-H19) *(C47-H58) 0.12 (C13-H19) *(C49-H61) 0.06 (C37-O38) *(C45-H53) 1.11 (C37-O38) *(C47-H58) 0.06 LP(O38) *(C45-H53) 11.44 *(C37-O38) *(C45-H53) 0.19 - From [BF4] to DCHA (B67-F70) *(O39-H40) 0.23 LP(F66) *(C5-H11) 0.83 LP(F66) *(O39-H40) 0.07 LP(F70) *(O39-H40) 15.95 - From DCHA to [BF4] LP(O39) *(B67-F70) 0.07

230

Table S3 The NBO charge of (a) CHCA, (b) CPCA, (c) BA, (d) CHPA, (e) CHDCA, (f) DCHA, (g) [BMIM][BF4], (h) [BMIM][BF4]-CHCA, (i) [BMIM][BF4]-CPCA, (j) [BMIM][BF4]-BA, (k) [BMIM][BF4]-CHPA, (l) [BMIM][BF4]-CHDCA, and (m) [BMIM][BF4]-DCHA.

a b c d e C1 -0.38356 C1 -0.40833 C1 -0.21113 C1 -0.38846 C1 -0.32234 C2 -0.39053 C2 -0.38511 C2 -0.16761 C2 -0.39181 C2 -0.38252 C3 -0.39248 C3 -0.32799 C3 -0.17660 C3 -0.22378 C3 -0.38252 C4 -0.39049 C4 -0.39559 C4 -0.15855 C4 -0.38563 C4 -0.32234 C5 -0.39213 C5 -0.40503 C5 -0.21046 C5 -0.38922 C5 -0.39059 C6 -0.32258 C6 0.21059 C6 -0.18134 C6 -0.39050 C6 -0.39059 H7 0.20507 H7 0.20566 H7 0.21432 H7 0.19238 H7 0.22348 H8 0.19335 H8 0.20489 H8 0.23152 H8 0.20355 H8 0.20242 H9 0.20751 H9 0.21935 H9 0.23544 H9 0.19127 H9 0.22348 H10 0.19986 H10 0.23817 H10 0.21492 H10 0.20255 H10 0.22795 H11 0.21806 H11 0.22449 H11 0.21247 H11 0.19249 H11 0.22795 H12 0.19140 H12 0.20413 C12 0.81710 H12 0.20335 H12 0.22002 H13 0.20751 H13 0.21715 O13 -0.60595 H13 0.20302 H13 0.21190 H14 0.21466 H14 0.20272 O14 -0.70502 H14 0.19102 H14 0.21190 H15 0.21106 C15 0.83420 H15 0.49089 H15 0.19079 H15 0.22002 H16 0.22502 O16 -0.61724 H16 0.20289 H16 0.20242 H17 0.19367 O17 -0.71054 H17 0.18785 C17 0.83935 C18 0.83994 H18 0.48847 C18 -0.37987 O18 -0.60738 O19 -0.61049 C19 -0.38747 O19 -0.70974 O20 -0.71081 H20 0.19562 H20 0.48743 H21 0.48598 H21 0.19858 C21 0.83935 C22 -0.38962 O22 -0.60738 H23 0.19424 O23 -0.70974 H24 0.19210 H24 0.48743 C25 -0.49186 H26 0.20776 H27 0.20815 H28 0.23032 H29 0.23039 C30 0.83156 O31 -0.60860 O32 -0.70913 H33 0.48608 f g h i j C1 -0.30579 N1 -0.36735 C1 -0.38214 N1 -0.36804 N1 -0.36816 H2 0.22771 C2 -0.02099 C2 -0.39066 C2 -0.01798 C2 -0.01938 C3 -0.21627 C3 -0.01664 C3 -0.39184 C3 -0.01658 C3 -0.01691 C4 -0.40802 N4 -0.36589 C4 -0.39020 N4 -0.36677 N4 -0.36645 C5 -0.39464 C5 0.33952 C5 -0.39057 C5 0.32631 C5 0.32793

231

H6 0.20439 C6 -0.35180 C6 -0.32027 C6 -0.36022 C6 -0.36103 C7 -0.38930 C7 -0.16180 H7 0.20452 C7 -0.16084 C7 -0.16103 H8 0.21800 C8 -0.41464 H8 0.19443 C8 -0.41405 C8 -0.41368 H9 0.20649 C9 -0.39306 H9 0.20657 C9 -0.39176 C9 -0.39200 C10 -0.38931 C10 -0.57904 H10 0.19561 C10 -0.57922 C10 -0.57906 H11 0.19673 H11 0.23633 H11 0.22130 H11 0.23500 H11 0.23479 H12 0.21064 H12 0.23401 H12 0.19220 H12 0.23274 H12 0.23263 C13 -0.39408 H13 0.27431 H13 0.20508 H13 0.29006 H13 0.28888 H14 0.20636 H14 0.22960 H14 0.21172 H14 0.24289 H14 0.24541 H15 0.19056 H15 0.20157 H15 0.20760 H15 0.20130 H15 0.20114 H16 0.20569 H16 0.24416 H16 0.22634 H16 0.23453 H16 0.23433 H17 0.19303 H17 0.20217 H17 0.19172 H17 0.20180 H17 0.20176 H18 0.20438 H18 0.23627 C18 0.86314 H18 0.23423 H18 0.23484 H19 0.19518 H19 0.25298 O19 -0.68748 H19 0.24795 H19 0.24750 C20 -0.22925 H20 0.18745 O20 -0.70697 H20 0.18831 H20 0.18820 C21 -0.39517 H21 0.18570 H21 0.52962 H21 0.18676 H21 0.18677 C22 -0.39188 H22 0.21058 N22 -0.36835 H22 0.20812 H22 0.20841 H23 0.21741 H23 0.20516 C23 -0.01898 H23 0.20511 H23 0.20506 C24 -0.38967 H24 0.20946 C24 -0.01711 H24 0.20806 H24 0.20782 H25 0.19604 H25 0.19266 N25 -0.36655 H25 0.19331 H25 0.19324 H26 0.20741 F26 -0.59453 C26 0.32761 F26 -0.58668 F26 -0.58595 C27 -0.39056 B27 1.34939 C27 -0.36067 B27 1.35486 B27 1.35431 H28 0.19506 F28 -0.58997 C28 -0.16080 F28 -0.57625 F28 -0.57578 H29 0.21059 F29 -0.54547 C29 -0.41441 F29 -0.54386 F29 -0.54345 C30 -0.39350 F30 -0.59011 C30 -0.39178 F30 -0.59857 F30 -0.59714 H31 0.20497 C31 -0.57910 C31 -0.40305 C31 -0.21118 H32 0.19491 H32 0.23480 C32 -0.38238 C32 -0.16376 H33 0.20558 H33 0.23250 C33 -0.32335 C33 -0.17117 H34 0.19669 H34 0.29024 C34 -0.39415 C34 -0.16317 H35 0.20519 H35 0.24505 C35 -0.39969 C35 -0.21347 H36 0.19271 H36 0.20096 H36 0.20854 C36 -0.18368 C37 0.84954 H37 0.23387 H37 0.19826 H37 0.21410 O38 -0.62469 H38 0.20168 H38 0.20271 H38 0.23588 O39 -0.70627 H39 0.23455 H39 0.22108 H39 0.23076 H40 0.48313 H40 0.24832 H40 0.23887 H40 0.21248 H41 0.18809 H41 0.22071 H41 0.21140 H42 0.18657 H42 0.20355 C42 0.83995 H43 0.20856 H43 0.20915 O43 -0.68323 H44 0.20491 H44 0.19945 O44 -0.70190 H45 0.20819 C45 0.85804 H45 0.53400 H46 0.19311 O46 -0.69238 F47 -0.58644 O47 -0.70503 B48 1.35487 H48 0.52912 F49 -0.57660

232

F50 -0.54404 F51 -0.59876 k l m N1 -0.36540 N1 -0.36804 C1 -0.30073 C2 -0.02642 C2 -0.02605 H2 0.22594 C3 -0.01557 C3 -0.01245 C3 -0.21187 N4 -0.36772 N4 -0.36361 C4 -0.40262 C5 0.32069 C5 0.32476 C5 -0.40509 C6 -0.35653 C6 -0.35605 H6 0.20182 C7 -0.18152 C7 -0.17690 C7 -0.38466 C8 -0.41322 C8 -0.41586 H8 0.21001 C9 -0.40377 C9 -0.39886 H9 0.20407 C10 -0.58184 C10 -0.58718 C10 -0.38663 H11 0.23789 H11 0.23412 H11 0.21079 H12 0.23548 H12 0.23440 H12 0.21428 H13 0.29520 H13 0.29124 C13 -0.39308 H14 0.22755 H14 0.22967 H14 0.20096 H15 0.20361 H15 0.20139 H15 0.19046 H16 0.24422 H16 0.24733 H16 0.20660 H17 0.20561 H17 0.20183 H17 0.19282 H18 0.25346 H18 0.25142 H18 0.20221 H19 0.24359 H19 0.24656 H19 0.18654 H20 0.19178 H20 0.19520 C20 -0.22361 H21 0.19206 H21 0.19486 C21 -0.39174 H22 0.20926 H22 0.19388 C22 -0.38970 H23 0.20606 H23 0.21088 H23 0.21039 H24 0.20810 H24 0.20438 C24 -0.38646 H25 0.19623 H25 0.20298 H25 0.19526 F26 -0.59158 F26 -0.59018 H26 0.20505 B27 1.35696 B27 1.35571 C27 -0.38692 F28 -0.57922 F28 -0.57367 H28 0.19609 F29 -0.54427 F29 -0.54457 H29 0.20915 F30 -0.60486 F30 -0.60475 C30 -0.39044 C31 -0.39165 C31 -0.32845 H31 0.20320 C32 -0.39556 C32 -0.38033 H32 0.19136 C33 -0.22813 C33 -0.39018 H33 0.20401 C34 -0.38861 C34 -0.31921 H34 0.19197 C35 -0.39216 C35 -0.38781 H35 0.20283 C36 -0.39354 C36 -0.38776 H36 0.19162 H37 0.19450 H37 0.23957 C37 0.87747 H38 0.20708 H38 0.20542 O38 -0.69721 H39 0.19321 H39 0.22157 O39 -0.69649 H40 0.20464 H40 0.22596 H40 0.52521 H41 0.19408 H41 0.22788 N41 -0.36820

233

H42 0.20311 H42 0.21602 C42 -0.02156 H43 0.20388 H43 0.20166 C43 -0.01694 H44 0.19277 H44 0.21424 N44 -0.36914 H45 0.19262 H45 0.21565 C45 0.32041 H46 0.20430 H46 0.20404 C46 -0.35364 H47 0.19014 C47 0.87825 C47 -0.16816 C48 -0.38339 O48 -0.68707 C48 -0.41226 C49 -0.39385 O49 -0.70280 C49 -0.39997 H50 0.19779 H50 0.52363 C50 -0.57831 H51 0.19641 C51 0.84042 H51 0.23508 C52 -0.39502 O52 -0.60907 H52 0.23231 H53 0.20039 O53 -0.71065 H53 0.29630 H54 0.19654 H54 0.48655 H54 0.23019 C55 -0.49772 H55 0.20273 H56 0.20295 H56 0.23806 H57 0.20722 H57 0.20361 H58 0.23268 H58 0.23665 H59 0.24306 H59 0.24513 C60 0.87055 H60 0.18969 O61 -0.68251 H61 0.18849 O62 -0.70348 H62 0.21250 H63 0.52187 H63 0.20476 H64 0.20891 H65 0.19317 F66 -0.59079 B67 1.35409 F68 -0.57764 F69 -0.54286 F70 -0.59548

234

Table S4 The ,  occupancy and linear combination of some natural atomic orbitals of + [BMIM] in isolated [BMIM][BF4], and [BMIM][BF4]-NAs.

[BMIM][BF4] Bond orbitals Occupancy Linear combination of NAOs (N1-C2) 1.98003 0.7990(sp2.12)N1 + 0.6013(sp2.69)C2 (N1-C5) 1.98321 0.7957(sp1.97)N1 + 0.6059(sp2.23)C5 (C2-C3) 1.97775 0.7069(sp1.66)C2 + 0.7073(sp1.65)C3 (C3-N4) 1.98087 0.6008(sp2.70)C3 + 0.7994(sp2.07)N4 (N4-C5) 1.98371 0.7959(sp1.94)N4 + 0.6054(sp2.25)C5 (C5-H13) 1.98201 0.7993(sp1.60)C5 + 0.6010(s)H13 (C6-H14) 1.98927 0.7837(sp2.88)C6 + 0.6212(s)H14 (C7-H18) 1.98019 0.7867(sp3.07)C7 + 0.6174(s)H18 (N1-C5) 1.89342 0.8456(p)N1 + 0.5337(p)C5 (C2-C3) 1.86074 0.7073(p)C2 + 0.7069(p)C3 [BMIM][BF4]-CHCA (N22-C23) 1.98026 0.7990(sp2.10)N22 + 0.6013(sp2.69)C23 (N22-C26) 1.98294 0.7960(sp1.97)N22 + 0.6054(sp2.26)C26 (C23-C24) 1.97705 0.7070(sp1.66)C23 + 0.7072(sp1.65)C24 (C24-N25) 1.98161 0.6010(sp2.70)C24 + 0.7992(sp2.06)N25 (N25-C26) 1.98348 0.7965(sp1.94)N25 + 0.6046(sp2.33)C26 (C26-H34) 1.98032 0.8063(sp1.53)C26 + 0.5915(s)H34 (C27-H35) 1.98839 0.7891(sp2.79)C27 + 0.6141(s)H35 (C28-H39) 1.97988 0.7860(sp3.09)C28 + 0.6182(s)H39 (N22-C26) 1.89217 0.8485(p)N22 + 0.5292(p)C26 (C23-C24) 1.86109 0.7069(p)C23 + 0.7073(p)C24 [BMIM][BF4]-CPCA (N1-C2) 1.98025 0.7991(sp2.10)N1 + 0.6012(sp2.69)C2 (N1-C5) 1.98291 0.7959(sp1.97)N1 + 0.6054(sp2.26)C5 (C2-C3) 1.97702 0.7070(sp1.66)C2 + 0.7072(sp1.65)C3 (C3-N4) 1.98160 0.6009(sp2.70)C3 + 0.7994(sp2.06)N4 (N4-C5) 1.98371 0.7959(sp1.94)N4 + 0.6054(sp2.33)C5 (C5-H13) 1.98030 0.8063(sp1.53)C5 + 0.5915(s)H13 (C6-H14) 1.98838 0.7885(sp2.80)C6 + 0.6152(s)H14 (C7-H18) 1.97981 0.7859(sp3.09)C7 + 0.6183(s)H18 (N1-C5) 1.89223 0.8483(p)N1 + 0.5296(p)C5 (C2-C3) 1.86037 0.7068(p)C2 + 0.7075(p)C3 [BMIM][BF4]-BA (N1-C2) 1.98023 0.7990(sp2.10)N1 + 0.6013(sp2.69)C2 (N1-C5) 1.98293 0.7960(sp1.97)N1 + 0.6053(sp2.26)C5 (C2-C3) 1.97706 0.7070(sp1.66)C2 + 0.7072(sp1.65)C3 (C3-N4) 1.98156 0.6011(sp2.70)C3 + 0.7992(sp2.06)N4 (N4-C5) 1.98347 0.7964(sp1.94)N4 + 0.6048(sp2.33)C5

235

(C5-H13) 1.98038 0.8058(sp1.53)C5 + 0.5922(s)H13 (C6-H14) 1.98842 0.7892(sp2.79)C6 + 0.6141(s)H14 (C7-H18) 1.97982 0.7861(sp3.09)C7 + 0.6181(s)H18 (N1-C5) 1.89227 0.8483(p)N1 + 0.5295(p)C5 (C2-C3) 1.86104 0.7070(p)C2 + 0.7072(p)C3 [BMIM][BF4]-CHPA (N1-C2) 1.98077 0.7989(sp2.10)N1 + 0.6015(sp2.70)C2 (N1-C5) 1.98295 0.7962(sp1.98)N1 + 0.6051(sp2.31)C5 (C2-C3) 1.97731 0.7073(sp1.66)C2 + 0.7070(sp1.65)C3 (C3-N4) 1.98111 0.6008(sp2.69)C3 + 0.7994(sp2.06)N4 (N4-C5) 1.98357 0.7961(sp1.94)N4 + 0.6051(sp2.29)C5 (C5-H13) 1.97973 0.8081(sp1.53)C5 + 0.5891(s)H13 (C6-H14) 1.98920 0.7829(sp2.90)C6 + 0.6221(s)H14 (C7-H18) 1.97864 0.7933(sp2.92)C7 + 0.6088(s)H18 (N1-C5) 1.89234 0.8465(p)N1 + 0.5323(p)C5 (C2-C3) 1.86034 0.7083(p)C2 + 0.7060(p)C3 [BMIM][BF4]-CHDCA (N1-C2) 1.98069 0.7988(sp2.11)N1 + 0.6015(sp2.70)C2 (N1-C5) 1.98294 0.7965(sp1.97)N1 + 0.6047(sp2.33)C5 (C2-C3) 1.97724 0.7075(sp1.66)C2 + 0.7067(sp1.65)C3 (C3-N4) 1.98104 0.6011(sp2.69)C3 + 0.7992(sp2.08)N4 (N4-C5) 1.98371 0.7959(sp1.93)N4 + 0.6054(sp2.28)C5 (C5-H13) 1.97997 0.8068(sp1.53)C5 + 0.5909(s)H13 (C6-H14) 1.98898 0.7836(sp2.88)C6 + 0.6213(s)H14 (C7-H18) 1.97849 0.7921(sp2.99)C7 + 0.6104(s)H18 (N1-C5) 1.89221 0.8479(p)N1 + 0.5301(p)C5 (C2-C3) 1.86090 0.7095(p)C2 + 0.7049(p)C3 [BMIM][BF4]-DCHA (N41-C42) 1.98043 0.7988(sp2.11)N41 + 0.6016(sp2.69)C42 (N41-C45) 1.98278 0.7962(sp1.97)N1 + 0.6051(sp2.32)C5 (C42-C43) 1.97705 0.7070(sp1.66)C42 + 0.7072(sp1.65)C43 (C43-N44) 1.98144 0.6010(sp2.70)C43 + 0.7992(sp2.06)N44 (N44-C45) 1.98332 0.7965(sp1.94)N44 + 0.6046(sp2.34)C45 (C45-H53) 1.97886 0.8105(sp1.49)C45 + 0.5858(s)H53 (C46-H54) 1.98879 0.7838(sp2.89)C46 + 0.6210(s)H54 (C47-H58) 1.97898 0.7875(sp3.03)C7 + 0.6163(s)H58 (N41-C45) 1.89206 0.8492(p)N41 + 0.5281(p)C45 (C42-C43) 1.86304 0.7073(p)C42 + 0.7069(p)C43

236

APPENDIX C: SUPPLEMENTARY MATERIALS FOR CHAPTER SIX

(a) [C4OHMIM][BF4] (b) [C4NHMIM][BF4]

237

(e) [C4OCMIM][BF4]

Fig. S1 The optimized structures of (a) [C4OHMIM][BF4], (b) [C4NHMIM][BF4], (c) [C4COOHMIM][BF4], (d) [C4COOCMIM][BF4], and (e) [C4OCMIM][BF4].

238

(a) [C4OHMIM][BF4]-CHCA (b) [C4NHMIM][BF4]-CHCA

(e) [C4OCMIM][BF4]-CHCA

Fig. S2 The electrostatic potential (a.u.) of (a)[C4OHMIM][BF4]-CHCA,(b) [C4NHMIM][BF4]- CHCA, (c) [C4COOHMIM][BF4]-CHCA, (d) [C4COOCMIM][BF4]-CHCA, and (e) [C4OCMIM][BF4]-CHCA.

239

(a) [C4OHMIM][BF4]-CHCA

(b) [C4NHMIM][BF4]-CHCA

240

(d) [C4COOCMIM][BF4]-CHCA

241

(e) [C4OCMIM][BF4]-CHCA

Fig. S3 The different electron densities of five types of ionic liquids and model NAs. Note: the purple and green isosurfaces (0.001 a.u.) represent the region inwhich electron density is increased and decreased after interaction between cation and anion of ionic liquids and model NAs, respectively.

242

Table S1 Bond length (Å) of (a) [C4OHMIM][BF4], (b) [C4NHMIM][BF4], (c) [C4COOHMIM][BF4], (d) [C4COOCMIM][BF4], and (e) [C4OCMIM][BF4].

(a) (b) (c) (d) (e) F23···H13 2.346 F24···H13 2.340 F29···H16 2.222 F32···H13 2.615 F23···H13 2.346 F23···H18 2.379 F24···H18 2.447 F31···H13 2.710 F32···H14 2.526 F23···H18 2.378 F23···H19 2.441 F24···H19 2.462 F31···H14 2.465 F32···H31 2.197 F23···H19 2.447 F25···H13 2.295 F24···H22 2.398 F31···H28 1.994 F34···H16 2.215 F25···H13 2.297 F25···H14 2.606 F24···H31 2.526 F33···H19 2.275 F35···H19 2.252 F25···H14 2.610 F25···H16 2.531 F26···H13 2.335 F33···H24 2.676 F35···H24 2.316 F25···H16 2.526 F26···H19 2.413 F26···H14 2.631 F33···C5 2.816 F35···C5 2.783 F26···H19 2.412 F26···C5 2.728 F26···H16 2.546 O26···H13 2.164 O26···H13 2.185 F26···C5 2.731 O30···H20 2.528 F27···H19 2.405 O26···H18 2.307 O26···H18 2.279 O30···H21 2.614 O30···H21 2.601 F27···C5 2.712 O30···H28 2.063 O30···H28 2.078 N30···H19 2.417 O30···H29 2.065 O30···H29 2.081 N30···H22 2.728 O30···H32 2.073 N30···H23 2.080 O30···H33 2.022 N30···H29 2.154 O30···H34 2.072

243

Table S2 The NBO charge of (a) [C4OHMIM][BF4], (b) [C4NHMIM][BF4], (c) [C4COOHMIM][BF4], (d) [C4COOCMIM][BF4], (e) [C4OCMIM][BF4], (f) [C4OHMIM][BF4]- CHCA, (g) [C4NHMIM][BF4]-CHCA, (h) [C4COOHMIM][BF4]-CHCA, (i) [C4COOCMIM][BF4]-CHCA, and (j) [C4OCMIM][BF4]-CHCA. a b c d e N1 -0.36618 N1 -0.36619 N1 -0.36939 N1 -0.36814 N1 -0.36613 C2 -0.01966 C2 -0.01957 C2 -0.02524 C2 -0.02423 C2 -0.01971 C3 -0.01712 C3 -0.01607 C3 -0.00032 C3 -0.00204 C3 -0.01696 N4 -0.36617 N4 -0.36591 N4 -0.36389 N4 -0.36476 N4 -0.36614 C5 0.33827 C5 0.33649 C5 0.33543 C5 0.33689 C5 0.33834 C6 -0.35163 C6 -0.35259 C6 -0.36221 C6 -0.36021 C6 -0.35160 C7 -0.16071 C7 -0.16030 C7 -0.17057 C7 -0.17487 C7 -0.16048 C8 -0.42475 C8 -0.43210 C8 -0.42669 C8 -0.42102 C8 -0.42334 C9 -0.41055 C9 -0.39859 C9 -0.38720 C9 -0.38602 C9 -0.40831 C10 -0.02038 C10 -0.17614 C10 -0.49482 C10 -0.49692 C10 -0.01317 H11 0.23856 H11 0.23751 H11 0.23423 H11 0.23382 H11 0.23842 H12 0.23383 H12 0.23402 H12 0.24028 H12 0.23929 H12 0.23383 H13 0.27261 H13 0.27133 H13 0.27291 H13 0.27246 H13 0.27260 H14 0.22873 H14 0.22805 H14 0.23254 H14 0.22603 H14 0.22857 H15 0.20170 H15 0.20208 H15 0.19981 H15 0.19915 H15 0.20169 H16 0.24386 H16 0.24537 H16 0.25270 H16 0.25691 H16 0.24400 H17 0.20408 H17 0.20360 H17 0.20422 H17 0.20371 H17 0.20396 H18 0.23303 H18 0.22833 H18 0.24325 H18 0.24682 H18 0.23334 H19 0.25036 H19 0.26956 H19 0.24643 H19 0.24572 H19 0.25073 H20 0.20987 H20 0.19353 H20 0.20490 H20 0.20291 H20 0.20846 H21 0.19725 H21 0.19488 H21 0.20985 H21 0.21047 H21 0.19771 H22 0.22228 H22 0.21398 H22 0.20488 H22 0.20241 H22 0.22219 F23 -0.59525 H23 0.17359 H23 0.23472 H23 0.23132 F23 -0.59526 B24 1.34972 F24 -0.59883 H24 0.24881 H24 0.26938 B24 1.34975 F25 -0.58979 B25 1.34946 C25 0.86864 C25 0.87465 F25 -0.58991 F26 -0.58936 F26 -0.58635 O26 -0.68976 O26 -0.68947 F26 -0.58936 F27 -0.54646 F27 -0.58400 O27 -0.69256 O27 -0.54920 F27 -0.54632 H28 0.16043 F28 -0.54981 H28 0.51843 C28 -0.21706 H28 0.16456 H29 0.16545 H29 0.16391 F29 -0.57595 H29 0.17449 H29 0.16848 O30 -0.75493 N30 -0.87253 B30 1.35752 H30 0.18917 O30 -0.61501 H31 0.46291 H31 0.37678 F31 -0.60678 H31 0.21076 C31 -0.19682 H32 0.35651 F32 -0.54798 F32 -0.60000 H32 0.15763 F33 -0.59618 B33 1.35662 H33 0.18335 F34 -0.57777 H34 0.16090 F35 -0.60047 F36 -0.55079 f g h i j

244

N1 -0.36773 N1 -0.36797 N1 -0.36985 N1 -0.36765 N1 -0.36771 C2 -0.01865 C2 -0.01834 C2 -0.03167 C2 -0.02799 C2 -0.01822 C3 -0.01794 C3 -0.01680 C3 0.00247 C3 -0.01725 C3 -0.01786 N4 -0.36678 N4 -0.36727 N4 -0.36327 N4 -0.36765 N4 -0.36679 C5 0.32694 C5 0.32549 C5 0.32613 C5 0.32286 C5 0.32644 C6 -0.36075 C6 -0.36169 C6 -0.36785 C6 -0.35555 C6 -0.36035 C7 -0.15982 C7 -0.15973 C7 -0.17381 C7 -0.17815 C7 -0.15944 C8 -0.42409 C8 -0.43214 C8 -0.42354 C8 -0.41656 C8 -0.42254 C9 -0.40904 C9 -0.39805 C9 -0.38726 C9 -0.39117 C9 -0.40668 C10 -0.02024 C10 -0.17676 C10 -0.49584 C10 -0.47706 C10 -0.01302 H11 0.23688 H11 0.23612 H11 0.23316 H11 0.23345 H11 0.23669 H12 0.23231 H12 0.23252 H12 0.23977 H12 0.23235 H12 0.23243 H13 0.28877 H13 0.28788 H13 0.27967 H13 0.29666 H13 0.28892 H14 0.24554 H14 0.24458 H14 0.24807 H14 0.22810 H14 0.24516 H15 0.20088 H15 0.20115 H15 0.19832 H15 0.20216 H15 0.20102 H16 0.23358 H16 0.23498 H16 0.24555 H16 0.24317 H16 0.23343 H17 0.20328 H17 0.20316 H17 0.20388 H17 0.20095 H17 0.20319 H18 0.23161 H18 0.22697 H18 0.24633 H18 0.27129 H18 0.23135 H19 0.24605 H19 0.26581 H19 0.24384 H19 0.23804 H19 0.24658 H20 0.21053 H20 0.19420 H20 0.20358 H20 0.19449 H20 0.20917 H21 0.19844 H21 0.19488 H21 0.21048 H21 0.20777 H21 0.19888 H22 0.21790 H22 0.21364 H22 0.20404 H22 0.21312 H22 0.21735 F23 -0.58755 H23 0.17429 H23 0.23328 H23 0.23170 F23 -0.58761 B24 1.35394 F24 -0.59202 H24 0.25656 H24 0.22815 B24 1.35382 F25 -0.59611 B25 1.35519 C25 0.86910 C25 0.86345 F25 -0.59596 F26 -0.57575 F26 -0.59165 O26 -0.69364 O26 -0.65796 F26 -0.57554 F27 -0.54467 F27 -0.57132 O27 -0.70004 O27 -0.55834 F27 -0.54449 F28 0.16410 F28 -0.54913 H28 0.51321 C28 -0.21354 H28 0.16722 H29 0.16009 H29 0.16362 F29 -0.57094 H29 0.17424 H29 0.16414 O30 -0.75420 N30 -0.87252 B30 1.36634 H30 0.18712 O30 -0.61427 H31 0.46306 H31 0.37475 F31 -0.61088 H31 0.21504 C31 -0.19686 C32 -0.39055 H32 0.35636 F32 -0.54588 F32 -0.59752 H32 0.15751 C33 -0.39001 C33 -0.38236 F33 -0.59280 B33 1.35477 H33 0.18386 C34 -0.39190 C34 -0.39062 C34 -0.38105 F34 -0.57331 H34 0.16078 C35 -0.39050 C35 -0.39191 C35 -0.38966 F35 -0.59022 C35 -0.39057 C36 -0.38206 C36 -0.39020 C36 -0.39173 F36 -0.54804 C36 -0.39010 C37 -0.32010 C37 -0.39039 C37 -0.38910 C37 -0.38142 C37 -0.39174 H38 0.20430 C38 -0.32037 C38 -0.39153 C38 -0.38894 C38 -0.39048 H39 0.19160 H39 0.20455 C39 -0.32227 C39 -0.39109 C39 -0.38202 H40 0.20506 H40 0.19413 H40 0.20450 C40 -0.38852 C40 -0.32013 H41 0.20803 H41 0.20672 H41 0.19391 C41 -0.39293 H41 0.20420 H42 0.21081 H42 0.19609 H42 0.20732 C42 -0.32317 H42 0.19169 H43 0.19375 H43 0.22170 H43 0.20065 H43 0.20192 H43 0.20501 H44 0.20697 H44 0.19198 H44 0.21648 H44 0.19240 H44 0.20789

245

H45 0.22313 H45 0.20519 H45 0.19208 H45 0.20418 H45 0.21087 H46 0.19699 H46 0.21146 H46 0.20501 H46 0.19571 H46 0.19379 H47 0.22468 H47 0.20747 H47 0.21939 H47 0.21785 H47 0.20690 H48 0.19184 H48 0.22599 H48 0.20787 H48 0.19173 H48 0.22362 C49 0.86303 H49 0.19192 H49 0.22096 H49 0.20210 H49 0.19684 O50 -0.68785 C50 0.86380 H50 0.19278 H50 0.22626 H50 0.22470 O51 -0.70700 O51 -0.68769 C51 0.88428 H51 0.20217 H51 0.19167 H52 0.52920 O52 -0.70729 O52 -0.66031 H52 0.23184 C52 0.86344 H53 0.52962 O53 -0.75502 H53 0.19089 O53 -0.68803 H54 0.53893 C54 0.88817 O54 -0.70687 O55 -0.68602 H 55 0.52873 O56 -0.71675 H57 0.52272

246

Table S3 Electron densities () and Laplacian of electron density (2) of BCPs in (a) [C4OHMIM][BF4], (b) [C4NHMIM][BF4], (c) [C4COOHMIM][BF4], (d) [C4COOCMIM][BF4], (e) [C4OCMIM][BF4], (f) [C4OHMIM][BF4]-CHCA, (g) [C4NHMIM][BF4]-CHCA, (h) [C4COOHMIM][BF4]-CHCA, (i) [C4COOCMIM][BF4]-CHCA, and, (j) [C4OCMIM][BF4]-CHCA.

CP label X···Y  (a.u.) 2(a.u.)

(a)[C4OHMIM][BF4] 53 F26···C5 0.1387E-01 0.5513E-01 56 F23···H18 0.1209E-01 0.4838E-01 58 F23···H13 0.1401E-01 0.5828E-01 59 F25···H13 0.1498E-01 0.6191E-01 64 F25···C6 0.1088E-01 0.4746E-01 66 F26···H18 0.9306E-02 0.3376E-01 67 F23···H19 0.1076E-01 0.4173E-01

(b)[C4NHMIM][BF4] 54 F27···H19 0.9309E-02 0.3415E-01 58 F24···H19 0.1149E-01 0.4625E-01 61 F24···H18 0.1122E-01 0.4582E-01 65 F27···C5 0.1441E-01 0.5706E-01 66 F24···H22 0.1066E-01 0.4160E-01 71 F24···H13 0.1420E-01 0.5899E-01 75 F24···H31 0.6897E-02 0.2748E-01 78 F26···H13 0.1348E-01 0.5550E-01 83 F28···H31 0.5676E-02 0.2362E-01 85 F26···C6 0.1034E-01 0.4442E-01

(c) [C4COOHMIM][BF4] 49 O26···H18 0.1534E-01 0.5598E-01 54 O26···H13 0.1710E-01 0.7209E-01 60 F33···N1 0.1208E-01 0.4952E-01 61 F29···N4 0.1094E-01 0.4570E-01 63 F33···N4 0.1018E-01 0.4177E-01 66 F33···H19 0.1171E-01 0.4686E-01 69 F31···H14 0.1118E-01 0.4730E-01 71 F29···H16 0.1482E-01 0.5883E-01 76 F33···C25 0.1151E-01 0.4832E-01 85 F31···H28 0.2072E-01 0.8397E-01

(d) [C4COOCMIM][BF4] 48 O26···H18 0.1559E-01 0.5611E-01

247

CP label X···Y  (a.u.) 2(a.u.) 61 O26···H13 0.1576E-01 0.6438E-01 62 F35···H19 0.1255E-01 0.5107E-01 63 F35···N1 0.1303E-01 0.5448E-01 75 F32···C5 0.1212E-01 0.4962E-01 76 F35···H24 0.1271E-01 0.4885E-01 78 F34···N4 0.9173E-02 0.3944E-01 84 F32···C6 0.1167E-01 0.5220E-01 87 F34···H16 0.1459E-01 0.5671E-01 91 F32···H31 0.1427E-01 0.5684E-01

(e) [C4OCMIM][BF4] 56 F26···C5 0.1381E-01 0.5488E-01 58 F23···H18 0.1211E-01 0.4842E-01 61 F23···H13 0.1401E-01 0.5829E-01 62 F25···H13 0.1497E-01 0.6191E-01 67 F25···C6 0.1093E-01 0.4762E-01 71 F26···H19 0.9325E-02 0.3382E-01 72 F23···H19 0.1063E-01 0.4128E-01

(f) [C4OHMIM][BF4]-CHCA 58 O50···H13 0.9534E-02 0.4119E-01 62 O50···H14 0.9571E-02 0.3142E-01 66 F25···C5 0.1003E-01 0.4191E-01 68 F26···C5 0.1181E-01 0.4856E-01 71 F25···H52 0.3653E-01 0.1526E+00 82 F25···H16 0.1903E-01 0.7802E-01 83 F26···H19 0.9821E-02 0.3560E-01 84 F23···C5 0.1248E-01 0.5121E-01 92 F23···H19 0.9386E-02 0.3661E-01 97 F23···H18 0.1116E-01 0.4501E-01

(g) [C4NHMIM][BF4]-CHCA 59 F28···H31 0.7535E-02 0.2965E-01 65 F24···H31 0.4673E-02 0.2078E-01 72 F24···H22 0.1019E-01 0.3909E-01 82 F26···H35 0.3647E-01 0.1528E+00 92 F24···H19 0.1013E-01 0.4128E-01 97 F24···H18 0.1042E-01 0.4337E-01 101 F27···H19 0.1048E-01 0.3847E-01 103 F24··C5 0.1316E-01 0.5365E-01

248

CP label X···Y  (a.u.) 2(a.u.) 111 F26···C5 0.7603E-02 0.3396E-01 113 O51···H13 0.1914E-01 0.7901E-01 114 F27···C5 0.1227E-01 0.5022E-01 115 F26···H16 0.8502E-02 0.3557E-01 120 O51···H14 0.9461E-02 0.3129E-01

(h) [C4COOHMIM][BF4]-CHCA 56 F29···H16 0.1152E-01 0.4446E-01 59 F31···C6 0.1151E-01 0.5138E-01 64 F31···H54 0.3689E-01 0.1606E+00 68 O52···H14 0.1049E-01 0.3529E-01 70 F29···N4 0.1033E-01 0.4182E-01 82 O53···H28 0.2841E-01 0.1097E+00 85 O52···H13 0.1609E-01 0.6472E-01 95 F33···C5 0.1203E-01 0.4882E-01 98 O26···C51 0.1134E-01 0.4283E-01 103 F25···C25 0.1038E-01 0.4432E-01 104 O26···H13 0.1041E-01 0.3934E-01 110 F33···H24 0.9388E-02 0.3893E-01 115 F33···H19 0.1152E-01 0.4591E-01 122 O26···H47 0.6241E-02 0.2231E-01 130 O26···H18 0.1594E-01 0.5824E-01

(i) [C4COOCMIM][BF4]-CHCA 61 F32···C6 0.7957E-02 0.3508E-01 64 F34···H16 0.9748E-02 0.3976E-01 67 F34···H16 0.3089E-01 0.1306E+00 71 F32···H13 0.1266E-01 0.5116E-01 74 F34···H16 0.1239E-01 0.5000E-01 79 O55···H13 0.2361E-01 0.9804E-01 94 F35···O55 0.4746E-02 0.1969E-01 95 F35···C5 0.1147E-01 0.4610E-01 97 O56···H31 0.8688E-02 0.2873E-01 101 F36···H31 0.7013E-02 0.2711E-01 108 O55···H18 0.9560E-02 0.3203E-01 111 F35···H31 0.9752E-02 0.3707E-01 113 O26···C54 0.8920E-02 0.3741E-01 117 F35···H18 0.9016E-02 0.3910E-01 119 O26···H52 0.9163E-02 0.3550E-01 127 F35···O26 0.9903E-02 0.4312E-01

249

CP label X···Y  (a.u.) 2(a.u.) 130 F35···H19 0.1101E-01 0.4219E-01 131 O26···H18 0.1026E-01 0.3443E-01 132 O26···H50 0.8367E-02 0.3068E-01

(j) [C4OCMIM][BF4]-CHCA 85 F23···H18 0.1097E-01 0.4431E-01 86 F23···H19 0.9410E-02 0.3665E-01 101 F25···H55 0.3678E-01 0.1534E+00 103 F26···H19 0.9870E-02 0.3582E-01 104 F23···C5 0.1261E-01 0.5168E-01 109 O53···H13 0.1936E-01 0.7974E-01 114 F25···C5 0.1001E-01 0.4173E-01 115 F26···C5 0.1171E-01 0.4836E-01 118 O53···H14 0.8928E-02 0.2894E-01 122 F25···H16 0.9718E-02 0.4225E-01

250

APPENDIX D: COPYRIGHT PERMISSIONS

251

252

253

254

255

256

257

258