Ionic Liquid Materials As Gas Chromatography Stationary Phases and Sorbent Coatings

Home , Vinyl group

A Dissertation

entitled

Ionic Liquid Materials as Gas Chromatography Stationary Phases and Sorbent Coatings

in Solid-Phase Microextraction

Qichao Zhao

Submitted to the Graduate Faculty as partial fulfillment of the

requirements for the Doctor of Philosophy Degree in Chemistry

Dr. Jared L. Anderson, Committee Chair

Dr. Jon R. Kirchhoff, Committee Member

Dr. Xiche Hu, Committee Member

Dr. Maria R. Coleman, Committee Member

Dr. Patricia R. Komuniecki, Dean College of Graduate Studies

The University of Toledo December 2011

An Abstract of Ionic Liquid Materials as Gas Chromatography Stationary Phases and Sorbent Coatings in Solid-Phase Microextraction by

Qichao Zhao Submitted to the Graduate Faculty as partial fulfillment of the requirements for the Doctor of Philosophy Degree in Chemistry

The University of Toledo December 2011

Ionic liquids (ILs) are a class of molten salts with melting points below 100 °C.

Their unique properties including high thermal stability, wide viscosity range, negligible vapor pressure under room temperature, and the capability of undergoing multiple solvation interactions make them ideal candidates as gas chromatography

(GC) stationary phases and sorbent coatings in solid-phase microextraction (SPME).

The first part of this dissertation includes an introduction of ionic liquids and their applications as GC stationary phases. The following chapters in this part introduce various example of utilizing ILs as GC stationary phases. Functionalized ILs containing different functional groups and tris(pentafluoroethyl)trifluorophosphate

(FAP) anion have been characterized using the solvation parameter model and applied as selective GC stationary phases. A total of fifteen ILs are examined on the basis of multiple solvation interactions. The effect of cation functional group, cation type, and counter anion has been thoroughly evaluated. The binary mixtures

iii consisting of two PILs have been employed as highly selective GC stationary phases.

The effects of PIL composition on the bleed temperature, system constants, and separation selectivity are examined.

PILs have been proven to be an ideal class of SPME sorbent coatings with promising thermal stability, extraction efficiency, and selectivity. The second part of this dissertation begins with an introduction of the application of ILs and PILs as sorbent coatings in SPME. Two chapters are then presented describing the design, synthesis, and application of two different types of PILs as sorbent coatings in

SPME for the selective extraction of CO2. The presence of functional groups within the PILs results in different mechanism of CO2 capture. The analytical performances of the PIL fibers are evaluated and compared to that of the commercial fibers. The

PIL-fibers and a selected commercial fiber are applied for the selective extraction of

CO2 from simulated flue gas. The effects of humidity and temperature on the performance of different SPME fibers are discussed. The PIL-fiber exhibits superior

CO2/CH4 and CO2/N2 selectivity to that of the commercial fiber. The final chapter introduces the development of a rapid and convenient analytical method for monitoring the enantiomeric purity of chiral molecules in IL-solvents, using headspace SPME coupled to chiral GC. Two commercial fibers as well as a

PIL-based fiber are examined, and the analytical performance of the developed method is thoroughly evaluated. The developed method is successfully applied to the enantiomeric excess determination from selected mixtures of chiral molecules. iv This dissertation is dedicated to my parents: Quansheng Zhao and Minmin Wang for their love, endless support and encouragement

Acknowledgments

First and foremost, I would like to express my sincere gratitude to my advisor,

Dr. Jared L. Anderson, for his guidance, understanding, and patience. He has taught me how good experimental chemistry is done, as well as enlightened how to think logically and execute efficiently. It is my great fortune to be one of his students. For everything you have done for me, Dr. Anderson, thank you.

I would like to acknowledge my committee members, Dr. Kirchhoff, Dr. Hu, and Dr. Coleman for all the helps and suggestions. I would like to also thank all my previous and current labmates, Dr. Yao, Fei, Quinner, Dr. Meng, William, Christa,

Leah, Pamela, Manish, Tianhao, Tien, Honglian, as well as Dr. Li from Dr.

Coleman’s lab. I am grateful to the University of Toledo, Department of Chemistry for giving me the opportunity to pursue my graduate studies here.

Finally, and most importantly, I would like to thank my parents and other family members, as well as all my friends, including that little kitty who stayed with me for a short but memorable time. I am heavily indebted to them for everything they did for me, the time they spent with me and their faith in me. This dissertation would not have been possible without their support, encourage, and love.

vi Contents

Abstract iii

Acknowledgments vi

Contents vii

List of Tables \ xv

List of Figures xix

1 Overview: Application of Ionic Liquids as Gas Chromatography Stationary

Phases 1

1.1 Introduction 1

1.2 Description of Ionic Liquids and Their Advantages When Used as Gas

Chromatography Stationary Phases 2

1.3 General Chemometric and Solvation Models for Gas Chromatography 4

1.3.1 Kovats retention index system 4

1.3.2 Rohrschneider–McReynolds system 5

1.3.3 Abraham’s solvation parameter model 6

vii 1.4 Application of Ionic Liquids as GC Stationary Phases 8

1.4.1 Application of monomeric ionic liquids as GC stationary phases 8

1.4.2 Application of dicationic and tricationic ionic liquids as GC stationary

phases 9

1.4.3 Application of polymeric ionic liquids as GC stationary phases \ 11

1.4.4 Application of ionic liquid mixtures as GC stationary phases 12

1.4.5 Application of functionalized ionic liquids as GC stationary phases 13

1.4.6 Application of ionic liquids as chiral GC stationary phases 15

1.5 Summary 16

2 Using the Solvation Parameter Model to Characterize Functionalized Ionic

Liquids Containing the Tris(pentafluoroethyl)trifluorophosphate (FAP)

Anion 18

Abstract 18

2.1 Introduction 19

2.2 Experimental 22

2.2.1 Materials 22

2.2.2 Methods 24

2.3 Results and Discussion 25

2.3.1 Effects of cation functional group on system constants 31

2.3.2 Effects of cation type on system constants 32 viii 2.3.3 Anion effect on system constants 33

2.3.4 Effect of IL stationary-phase composition on retention behavior for

selected solute molecules 36

2.3.5 Effect of IL stationary-phase composition on separation selectivity for

selected solutes 39

2.4 Conclusions 41

Acknowledgements 44

References 44

3 Evaluating the Solvation Properties of Functionalized Ionic Liquids with

Varied Cation/Anion Composition Using the Solvation Parameter

Model 47

Abstract 47

3.1 Introduction 48

3.2 Experimental 51

3.2.1 Materials 51

3.2.2 Methods 53

3.3 Results and Discussion 54

3.3.1 Effects of cation functional group on system constants 63

3.3.2 Effects of cation type on system constants 64

3.3.3 Effects of anion on system constants 65 ix 3.3.4 Effect of IL stationary phase composition on solute molecule retention

behavior 66

3.3.5 Effect of IL stationary-phase composition on solute pair separation

selectivity 71

3.4 Conclusions 77

Acknowledgements 79

References 80

4 Highly Selective GC Stationary Phases Consisting of Binary Mixtures of

Polymeric Ionic Liquids 82

Abstract 82

4.1 Introduction 83

4.2 Experimental 86

4.2.1 Materials 86

4.2.2 Methods 87

4.3 Results and Discussion 90

4.3.1 Measurement of column bleed temperature 90

4.3.2 System constants of neat and mixed PIL stationary phases 94

4.3.3 Effects of PIL stationary phase composition on retention behavior of

selected solute molecules 99

4.3.4 Effect of PIL stationary-phase composition on separation selectivity for x selected solute pairs 103

4.4 Concluding Remarks 107

Acknowledgements 108

References 109

5 Overview: Application of Ionic Liquids as Sorbent Coatings in Solid-Phase

Microextraction 113

5.1 Introduction 113

5.2 Application of Ionic Liquids as Sorbent Coatings in SPME 114

5.2.1 Physically coated IL-based SPME sorbent coatings 114

5.2.2 Chemically bonded IL-based SPME sorbent coatings 119

5.2.3 PIL-based SPME sorbent coatings 121

5.2.4 Task-specific IL- and PIL-based SPME sorbent coatings 123

5.2.5 IL-mediated SPME sorbent coatings 125

5.3 Summary 126

6 Polymeric Ionic Liquids as CO2 Selective Sorbent Coatings for Solid-Phase

Microextraction 128

Abstract 128

6.1 Introduction 129

6.2 Experimental 132 xi 6.2.1 Materials 132

6.2.2 Methods 133

6.3 Results and Discussion 140

6.3.1 Design of PIL-based sorbent coatings 140

6.3.2 Sorption-time profiles 142

6.3.3 Analytical performance 146

6.3.4 Sorbate storage capacity 150

6.4 Concluding Remarks 152

Acknowledgements 153

7 Selective Extraction of CO2 from Simulated Flue Gas Using Polymeric

Ionic Liquid Sorbent Coatings in Solid-Phase Microextraction 157

Abstract 157

7.1 Introduction 158

7.2 Experimental 162

7.2.1 Materials 162

7.2.2 Methods 163

7.3 Results and Discussion 167

7.3.1 Effects of water on the extraction of CO2 167

7.3.2 Effects of temperature on the extraction of CO2 169

7.3.3 Selectivity of the PIL-based and Carboxen fibers for CH4 and N2 171 xii 7.4 Conclusions 178

Acknowledgements 179

References 179

8 A Rapid Analytical Method for Monitoring the Enantiomeric Purity of

Chiral Molecules Synthesized in Ionic Liquid Solvents 181

Abstract 181

8.1 Introduction 182

8.2 Experimental 186

8.2.1 Chemicals and materials 186

8.2.2 Synthesis of BMIM-NTf2 and poly(VHIM-NTf2) 188

8.2.3 Preparation of PIL-based SPME fiber 189

8.2.4 Extraction of chiral molecules 189

8.2.5 “On-fiber” SPME derivatization 190

8.2.6 Chiral GC separation 191

8.3 Results and Discussion 192

8.3.1 Description of the SPME-GC method 192

8.3.2 Generation of Sorption-Time Profiles 192

8.3.3 Evaluation of Method Performance 194

8.3.4 Determination of enantiomeric excess 201

8.3.5 “On-fiber” SPME derivatization 203 xiii 8.4 Conclusions 206

Acknowledgements 207

References 208

9 Summary 212

References 216

Appendix A. Supplemental Figures and Table Accompanying Chapter 2 234

Appendix B. Supplemental Figures and Table Accompanying Chapter 3 255

Appendix C. Supplemental Figures Accompanying Chapter 4 290

Appendix D. Supplemental Figures Accompanying Chapter 6 295

Appendix E. Supplemental Figures Accompanying Chapter 8 299

xiv List of Tables

- - 2.1 System constants of [FAP] and [NTf2] -based IL stationary phases 27

2.2 Comparison of retention factors for selected solute molecules on four

- functionalised 1-methylpyrrolidinium [FAP] IL stationary phases at 80°C 37

2.3 Comparison of retention factors for selected solute molecules on stationary

phases composed of 2-hydroxyethyl-functionalised ILs at 80°C 38

2.4 Comparison of the selectivity for selected solute molecules on stationary phases

composed of four functionalised 1-methylpyrrolidinium [FAP]- IL stationary

at 80 °C 40

2.5 Effect of 2-hydroxyethyl-functionalised ILs on the selectivity of chosen solute

pairs at 80 °C 42

3.1 System constants of functionalized IL-based stationary phases examined in

this study 57

3.2 Comparison of retention factors for selected solute molecules on stationary

phases composed of ILs containing functionalized pyridinium cations and

FAPˉ anion at 80 °C 67

xv 3.3 Comparison of retention factors for selected solute molecules on stationary

phases composed of ILs containing functionalized imidazolium cations and

FAPˉ anion at 80 °C 69

3.4 Comparison of retention factors for selected solute molecules on stationary

phases composed of ILs containing various cation types with FAPˉ anion

at 80 °C 70

3.5 Comparison of retention factors for selected solute molecules on stationary

phases composed of 1-butyl-1-methylpyrrolidinium cation and various

counter anions at 80 °C 72

3.6 Comparison of selectivity for selected solute molecules on stationary phases

composed of ILs containing functionalized pyridinium cations and FAPˉ anion

at 80 °C 74

3.7 Comparison of selectivity for selected solute molecules on stationary phases

composed of ILs containing functionalized imidazolium cations and FAPˉ anion

at 80 °C 75

3.8 Comparison of selectivity for selected solute molecules on stationary phases

composed of ILs containing various cation types and FAPˉ anion at 80 °C 76

3.9 Comparison of selectivity for selected solute molecules on stationary phases

composed of 1-butyl-1-methylpyrrolidinium cation and various counter anions

at 80 °C 78

4.1 List of probe molecules and their corresponding solute descriptors employed in

xvi the study of the IL stationary phases using the solvation parameter model 91

4.2 On-set bleed temperature of studied neat and binary PIL stationary phases 93

4.3 System constants of neat and mixed PIL stationary phases examined in this

study 95

4.4 Comparison of retention factors for selected analytes on six different PIL

− stationary phases varying in the weight percentage of NTf2 and chloride

anions at 80 ºC 100

4.5 Comparison of retention factors for selected analytes on four different PIL

stationary phases varying in length of cationic side chain at 80 ºC 102

4.6 Effect of PIL stationary phase chloride content on the selectivity of chosen

analyte pairs at 80 ºC 104

4.7 Effect of the length of the imidazolium aliphatic side chain substituent on the

selectivity of chosen analyte pairs at 80 ºC 106

5.1 Application of ILs and PILs in SPME 115

6.1 Reproducibility of four task-specific PIL-based fibers and two commercial

fibers 147

6.2 Figures of merit of calibration curves for two task-specific PIL-based fibers

and one commercial fiber in pure CO2 148

6.3 Figures of merit of calibration curves for two task-specific PIL-based fibers

and one commercial fiber in CO2 spiked with air (70 kPa) 149

7.1 Figures of merit of calibration curves for two task-specific PIL-based fibers xvii and one commercial fiber in dry CO2 and CO2 saturated with water vapor 165

7.2 Figures of merit of calibration curves for two task-specific PIL-based fibers

and one commercial fiber in pure CO2 at various temperatures 172

7.3 Gas pair selectivity for two task-specific PIL-based fibers and one commercial

fiber 177

8.1 Reproducibility of the PDMS, PA, and PIL-based fibers 198

8.2 Figures of merit of the calibration curves and limits of detection by using the

commercial PDMS fiber, PA fiber, and laboratory-made PIL fiber 199

8.3 Determination of enantiomeric excess values (%) of the derivatized methyl-3-

hydroxybutyrate enantiomers using direct injection and SPME methods 202

A.1 Probe molecules and their corresponding solute descriptors employed in the

study of the IL stationary phases using the solvation parameter model 235

B.1 Probe molecules and their corresponding solute descriptors employed in the

study of the IL stationary phases using the solvation parameter model 256

B.2 Comparison of retention factors for selected solute molecules on stationary

phases composed of ILs containing various cation types and NTf2ˉ anion

at 80 °C 258

B.3 Comparison of selectivity for selected solute molecules on stationary phases

composed of ILs containing various cation types and NTf2ˉ anion at 80 °C 259

xviii List of Figures

1-1 Structures of common cations and anions of ILs 2

2-1 Structures and numbering system of ILs that were examined in this study 26

2-2 Scheme demonstrating the acidity/basicity scales of the ten ILs evaluated

in this study 35

3-1 Structures and numbering system of ILs examined in this study 55

4-1 Structures of the PILs used as stationary-phase components in this study 89

4-2 Linear correlation between the stationary phase weight percentage of the

poly(ViHIm-Cl) PIL and the resulting hydrogen bond basicity 98

5-1 Monomeric ILs 1 and 2, and polymeric ILs 3 and 4, were utilized to create

bonded IL-based SPME coatings 120

6-1 Schematic demonstrating the synthesis of the task-specific polymeric ionic

liquids used for the selective capture of CO2 135

6-2 Apparatus of SPME setup used to perform extraction of CO2 138

6-3 Schematic illustrating the reversible reactive capture of CO2 by the

poly(VHIM-taurate) PIL. For simplicity, two imidazolium cations and taurate

anions from the linear polymer backbone are represented 141

xix 6-4 Scanning electron micrographs of a SPME fiber coated with the

poly(VHIM-taurate) PIL 143

6-5 Sorption-time profiles under two different CO2 pressures obtained for four

PIL-based (~10 μm) and two commercial fiber coatings 144

6-6 Comparison of the amount of CO2 sorbate retained in two PIL-based fibers and

the carboxen fiber under various storage conditions 151

7-1 Structures of the PILs used as SPME fiber coatings in this study 161

7-2 Apparatus of SPME setup used in this study to examine humidity effect and

temperature effect 165

7-3 Calibration curves obtained for the poly(VHIM-taurate) fiber coating under

varied temperatures 170

7-4 Scanning electron micrographs of a fiber coated with the poly(VHIM- taurate)

PIL after exposure to CO2, CH4, and N2 175

7-5 Sorption-time profiles of CH4 and N2 obtained for two PIL fibers and one

commercial fiber at room temperature 176

8-1 Structures of studied chiral molecules 187

8-2 Schematic diagram demonstrating the enantiomeric excess determination of

racemic mixture using developed SPME-GC method 193

8-3 Sorption-time profiles obtained for commercial PDMS fiber (100 μm), PA fiber

(85 μm), and laboratory-made PIL fiber (≈ 15 μm) by extracting the studied

analytes at a concentration of 2 mg/g using a constant stir rate of 900 rpm at

xx room temperature (19 ºC) 195

8-4 Schematic diagram demonstrating the enantiomeric excess determination of

a mixture using “on-fiber” derivatization SPME-GC method 204

8-5 Representative chromatograms of extraction of underivatized methyl-3-

hydroxybutyrate enantiomers and extraction of methyl-3-hydroxybutyrate

enantiomers using on-fiber derivatization method 205

1 A-1 H-NMR spectrum of 1-butyl-1-methylpyrrolidinium FAP 237

1 A-2 H-NMR spectrum of 1-(6-aminohexyl)-1-methylpyrrolidinium FAP 238

1 A-3 H-NMR spectrum of 1-ethoxycarbonylmethyl-1-methylpyrrolidinium FAP 239

1 A-4 H-NMR spectrum of 1-(2-hydroxyethyl)-1-methylpyrrolidinium FAP 240

1 A-5 H-NMR spectrum of 1-hexyl-3-methylimidazolium FAP 241

1 A-6 H-NMR spectrum of 1-(2-hydroxyethyl)-3-methylimidazolium FAP 242

1 A-7 H-NMR spectrum of trihexyl(tetradecyl)phosphonium FAP 243

1 A-8 H-NMR spectrum of 1-(2-hydroxyethyl)-1-methylpyrrolidinium NTf2 244

1 A-9 H-NMR spectrum of 1-(2-hydroxyethyl)-3-methylimidazolium NTf2 245

A-10 ESI-MS spectrum of 1-butyl-1-methylpyrrolidinium FAP 246

A-11 ESI-MS spectrum of 1-(6-aminohexyl)-1-methylpyrrolidinium FAP 247

A-12 ESI-MS spectrum of 1-ethoxycarbonylmethyl-1-methylpyrrolidinium FAP 248

A-13 ESI-MS spectrum of 1-(2-hydroxyethyl)-1-methylpyrrolidinium FAP 249

A-14 ESI-MS spectrum of 1-hexyl-3-methylimidazolium FAP 250

A-15 ESI-MS spectrum of 1-(2-hydroxyethyl)-3-methylimidazolium FAP 251 xxi A-16 ESI-MS spectrum of trihexyl(tetradecyl)phosphonium FAP 252

A-17 ESI-MS spectrum of 1-(2-hydroxyethyl)-1-methylpyrrolidinium NTf2 253

A-18 ESI-MS spectrum of 1-(2-hydroxyethyl)-3-methylimidazolium NTf2 254

1 B-1 H-NMR spectrum of N-hexylpyridinium FAP 260

1 B-2 H-NMR spectrum of N-hexyl-4-(N´,N´-dimethylamino)pyridinium FAP 261

1 B-3 H-NMR spectrum of N-hydroxypropylpyridinium FAP 262

1 B-4 H-NMR spectrum of 1-ethyl-3-methylimidazolium FAP 263

1 B-5 H-NMR spectrum of 1-methoxyethyl-3-methylimidazolium FAP 264

1 B-6 H-NMR spectrum of methoxyethyl-dimethyl-ethylammonium FAP 265

1 B-7 H-NMR spectrum of 1-methoxyethyl-1-methylmorpholinium FAP 266

1 B-8 H-NMR spectrum of 1-methoxyethyl-1-methylpiperidinium FAP 267

1 B-9 H-NMR spectrum of 1-methoxypropyl-1-methylpiperidinium FAP 268

1 B-10 H-NMR spectrum of hexyl-trimethylammonium NTf2 269

1 B-11 H-NMR spectrum of 1-propyl-1-methylpiperidinium NTf2 270

1 B-12 H-NMR spectrum of 1-butyl-1-methylpyrrolidinium SCN 271

1 B-13 H-NMR spectrum of 1-butyl-1-methylpyrrolidinium C(CN)3 272

1 B-14 H-NMR spectrum of 1-butyl-1-methylpyrrolidinium B(CN)4 273

1 B-15 H-NMR spectrum of 1-butyl-1-methylpyrrolidinium BOB 274

B-16 ESI-MS spectrum of N-hexylpyridinium FAP 275

B-17 ESI-MS spectrum of N-hexyl-4-(N´,N´-dimethylamino)pyridinium FAP 276

B-18 ESI-MS spectrum of N-hydroxypropylpyridinium FAP 277 xxii B-19 ESI-MS spectrum of 1-ethyl-3-methylimidazolium FAP 278

B-20 ESI-MS spectrum of 1-methoxyethyl-3-methylimidazolium FAP 279

B-21 ESI-MS spectrum of methoxyethyl-dimethyl-ethylammonium FAP 280

B-22 ESI-MS spectrum of 1-methoxyethyl-1-methylmorpholinium FAP 281

B-23 ESI-MS spectrum of 1-methoxyethyl-1-methylpiperidinium FAP 282

B-24 ESI-MS spectrum of 1-methoxypropyl-1-methylpiperidinium FAP 283

B-25 ESI-MS spectrum of hexyl-trimethylammonium NTf2 284

B-26 ESI-MS spectrum of 1-propyl-1-methylpiperidinium NTf2 285

B-27 ESI-MS spectrum of 1-butyl-1-methylpyrrolidinium SCN 286

B-28 ESI-MS spectrum of 1-butyl-1-methylpyrrolidinium C(CN)3 287

B-29 ESI-MS spectrum of 1-butyl-1-methylpyrrolidinium B(CN)4 288

B-30 ESI-MS spectrum of 1-butyl-1-methylpyrrolidinium BOB 289

1 C-1 H-NMR spectra of 1-vinyl-3-hexylimidazolium chloride 291

1 C-2 H-NMR spectra of 1-vinyl-3-hexadecylimidazolium chloride 292

1 C-3 H-NMR spectra of poly(1-vinyl-3-hexylimidazolium chloride) 293

1 C-4 H-NMR spectra of poly(1-vinyl-3-hexadecylimidazolium chloride) 294

1 D-1 H-NMR spectra of 1-vinyl-3-hexylimidazolium bromide \ 296

1 D-2 H-NMR spectra of 1-vinyl-3-hexylimidazolium taurate 297

1 D-3 H-NMR spectra of poly(1-vinyl-3-hexylimidazolium taurate) 298

1 E-1 H-NMR spectra of 1-butyl-3-methylimidazolium NTf2 300

1 E-2 H-NMR spectra of poly(1-vinyl-3-hexylimidazolium NTf2) 301 xxiii Chapter 1

Overview: Application of Ionic Liquids as Gas

Chromatography Stationary Phases

1.1 Introduction

Ionic liquids (ILs) are a class of molten salts with melting points lower than 100 °C

[1]. If an IL has a melting point at or below room temperature (approximately 25 °C), then it is considered a room temperature ionic liquid (RTIL). ILs are usually formed by the combination of nitrogen- or phosphorus-containing organic cations and organic or

− inorganic counter anions, such as halide, tetrafluoroborate (BF4 ), hexafluorophosphate

− − (PF6 ), and bis[(trifluoromethyl)sulfonyl]imide (NTf2 ). The structures of some common cations and anions are shown in Figure 1.1. Compared to traditional organic solvents, ILs possess numerous advantages including their high thermal stability, almost undetectable vapor pressure at room temperature, and the capability of undergoing multiple solvation interactions. In addition, the structure of ILs can be custom-designed to generate task-specific ILs (TSILs), which are capable of interacting with analytes in specific ways

[2].

1 R R R R R1 R2 N 1 4 1 4 N N N P

R R2 R3 R2 R3

Imidazolium Pyridinium Ammonium Phosphonium

F O O F F F F P FC S N S CF B 3 3 F F Cl F F F O O

- - - BF4 PF6 NTf2

Figure 1-1: Structures of common cations and anions of ILs.

2 1.2 Description of Ionic Liquids and Their Advantages When Used as

Gas Chromatography Stationary Phases

ILs have a surprisingly long history that dates back to 1914 when the first RTIL, namely ethyl ammonium nitrate, was reported [3]. However, it exhibited very poor thermal stability. Meanwhile, the first synthesis of imidazolium-based ILs was described by Wilkes and co-workers in 1982 [4]. However, due to the presence of the chloroaluminate anions, the resulting ILs suffered from considerable reactivity to moisture. Wilkes and Zawarotko reported the first development of air- and moisture-stable imidazolium-based ILs in 1992 [5]. Since then, researchers in diverse fields of science and engineering have devoted a significant amount of effort into the study and application of ILs. Currently, numerous ILs have been synthesized and applied in many analytical chemical investigations, including liquid-liquid extraction [6,7], analytical microextraction [8,9], chromatographic separation [10,11], electrochemistry

[12,13], and mass spectrometry [14].

One continually emerging application involving ILs is their employment as gas chromatography (GC) stationary phases [15]. The success of ILs as GC stationary phases is due to the advantages they possess over traditional materials such as the substituted polysiloxanes and polyethylene glycol stationary phases. For example, ILs typically exhibit extremely low volatility and high thermal stability as well as the capability of remaining in the liquid state over a wide temperature range. They can be designed to exhibit low column bleed and longer lifetimes as well as extended operation temperature

3 ranges when used as stationary phases. In addition, ILs are capable of undergoing multiple solvation interactions thereby imparting unique selectivities towards a wide range of molecules with different functional groups. Moreover, the tunability of ILs allows for relatively easy chemical modification of the IL structures, resulting not only in enhanced thermal stability but also tuneable solvation properties and separation selectivity. Currently, the reported modifications include the cation and anion combination [16], introduction of desired functional groups to the IL [17], development of dicationic and tricationic ILs [18,19], as well as polymerization [20].

1.3 General Chemometric and Solvation Models for Gas

Chromatography

1.3.1 Kovats retention index system

The concept of Kovats retention index system was outlined by Kovats in the 1950s during his research into the composition of essential oils [21]. This method converts analyte retention times into system-independent constants and consequently provides accurate retention parameters that can be compared from column to column and lab to lab.

In this method, linear n-alkanes are selected as reference compounds for non-polar stationary phases while fatty acid methyl esters are usually used for polar stationary phases. The calculation of a given analyte retention index involves the logarithmic scale of the retention time or volume of both the analyte and two reference n-alkanes that are used to bracket the analyte, as shown in Equation 1.

4 [Eq. 1]

According to Eq. 1, n is the number of the carbon atoms in the n-alkane eluting before analyte X, n+1 is the number of the carbon atoms in the n-alkane eluting after analyte X, and R is the adjusted retention time or retention volume of the analyte or reference n-alkane. Kovats retention index is an accurate method with high reproducibility and relatively high tolerance to changes in temperature. This system has been applied as a quantitative tool to provide GC data for inter-laboratory substance identification for a wide range of compounds under various temperatures.

1.3.2 Rohrschneider–McReynolds system

The Rohrschneider–McReynolds system, first proposed by Rohrschneider [22] and later revised by McReynolds [23], is one of the most widely used GC stationary phase classification systems. This system is based on the assumption that intermolecular forces are additive and lead to analyte molecule retention collectively. The squalane stationary phase is chosen as the reference stationary phase, and the individual interactions are characterized by the comparison of the retention index of representative probe molecules with that of squalane stationary phase. Five probe molecules, namely benzene, butanol,

2-pentanone, nitropropane, and pyridine, were chosen to represent an individual or a combination of interactions between analyte and stationary phase. Benzene measures dispersive interactions with some weak proton acceptor properties, butanol measures dipolar interactions with both proton donor and proton acceptor capabilities, 2-pentanone measures dipolar properties with proton acceptor properties, nitropropane measures dipolar interactions, and pyridine measures weak dipolar interactions as well as strong

5 proton acceptor capabilities. The illustration of the Rohrschneider–McReynolds system in terms of the five probe molecules is shown in Equation 2:

∆I = aX′ + bY′ + cZ′ + dU′ + eS′ [Eq. 2]

According to Eq. 2, the coefficients a, b, c, d, e are related with the properties of the five probes, namely benzene (a), butanol (b), 2-pentanone (c), nitropropane (d) and pyridine (e). The phase constants, X′, Y′, Z′, U′ and S′, are used to describe the capability of the stationary phase that is involved into specific individual interactions.

However, this system has some weaknesses that limit its applicability in certain aspects. A major weakness is that this system assumes that each solute represents a single dominant intermolecular interaction while ignoring the fact that all the solutes exhibit several interactions simultaneously. Moreover, this system does not consider the contribution of interfacial adsorption as a retention mechanism. In addition, the squalane reference stationary phase has poor thermal and oxidative stability at high temperatures, which limits the applicable temperature range of this system. Lastly, some of the test probes are too volatile and therefore often elute from the column close to or at the dead time, which diminishes the accuracy of the model.

1.3.3 Abraham’s solvation parameter model

Abraham and co-workers developed the solvation parameter model by utilizing a large number of test probes that are capable of undergoing multiple interactions with a stationary phase [24]. The solvation parameter model is a linear free energy relationship that provides considerable advantages over the Rohrschneider–McReynolds classification system. In the solvation parameter model, the total free energy change for the transfer of

6 a given solute X from the gas phase to the stationary phase (solvent) is assumed to be the linear sum of the contributions from different individual free energy. The solvation of solute X can be divided into three steps: (1) a cavity with a suitable size is created in the solvent; (2) the solvent molecules reorganize around the cavity; and (3) various interaction occurs between the solute and solvent which results the introduction of the solute into the cavity. The test probes interact with the solvent through different types of interactions depending on their structural properties. The solvation parameter model, as described in Equation 3, has been successfully employed to evaluate the solvation properties of a wide range of stationary phases.

log k = c + eE + sS + aA + bB + lL [Eq. 3]

According to Eq. 3, log k is the solute retention factor and is calculated by measuring the retention time of the analyte and dead volume of the chromatographic column. Five probe-specific parameters, referred to as solute descriptors (i.e., E, S, A, B,

L), have been determined for a wide range of probe molecules and are used to describe the capability of the solute to interact with the stationary phase [24]. They are defined as:

E, the excess molar refraction calculated from the solute’s refractive index; S, the solute dipolarity/polarizability; A, the solute hydrogen bond acidity; B, the solute hydrogen bond basicity; and L, the solute gas hexadecane partition coefficient determined at 298 K. The system constants (e, s, a, b, l) are used to characterize the strength of each solvation interaction and are defined as: e, the ability of the stationary phase to interact with π and nonbonding electrons of the solute; s, a measure of the dipolarity/polarizability of the stationary phase; a, the IL hydrogen bond basicity; b, a measure of the hydrogen bond acidity of the stationary phase; and l describes stationary phase dispersion forces. The

7 intercept term, c, can be used to determine and verify the phase ratio of the column. The system constants are attained through multiple linear regression analysis of the log k term and the five solute descriptors. The value of each system constant describes the contribution of the particular interaction to the overall solute–solvent retention mechanism.

1.4 Application of Ionic Liquids as GC Stationary Phases

1.4.1 Application of monocationic ionic liquids as GC stationary phases

The first example of utilizing ILs as GC stationary phases was demonstrated by

Armstrong and co-workers in 1999 [15]. Two imidazolium-based ILs, namely

1-butyl-3-methylimidazolium chloride (BMIM-Cl) and 1-butyl-3-methylimidazolium hexafluorophosphate (BMIM-PF6) were coated onto fused silica capillaries and their

Rohrschneider–McReynolds constants were evaluated. It was observed that these two ILs exhibited an interesting “dual nature” phenomenon when used as GC stationary phases, meaning that they were capable of separating non-polar analytes like nonpolar stationary phases and acted as polar stationary phases when separating polar analytes. The

BMIM-Cl stationary phase interacted strongly with proton donor and proton acceptor molecules while BMIM-PF6 was found to be somewhat less polar and interacted more strongly with non-polar molecules.

The Abraham’s solvation parameter model was applied to characterize 17 ILs on the basis of their distinct multiple solvation interactions [16]. The comparison of solvation properties of ILs consisting of the 1-butyl-3-methylimidazolium cation paired with

8 various counter anions revealed that the anion had the dominant effect on the hydrogen bond basicity and dipolarity of the resulting IL. The obtained data provided not only an explanation to why ILs were particularly useful for certain applications such as MALDI matrix but also a model that allowed one to pick the most suitable ILs for a specific organic synthesis, liquid-liquid extraction, or GC stationary phase.

More recently, Armstrong and co-workers utilized phosphonium-based ILs as GC stationary phases [25]. It was observed that the phosphonium IL-based stationary phases exhibited significantly enhanced thermal stability than their imidazolium-based analogues.

The solvation properties of phosphonium ILs were characterized using Abraham’s solvation parameter model, and the results indicated that compared to the imidazolium

ILs, the phosphonium ILs always showed stronger dispersion interaction but had less of an ability to interact with solutes through π and nonbonding electrons. As GC stationary phases, the phosphonium IL columns displayed unique selectivity and good efficiency when they were used for the separation of a complex mixture of 24 flavor and fragrance homologous ester compounds, a homologous mixture of alkanes and alcohols, and the

Grob mixture.

1.4.2 Application of dicationic and tricationic ionic liquids as GC stationary phases

In an attempt to expand the liquid range as well as enhance the thermal stability of

ILs, geminal dicationic ILs were developed [18]. This class of ILs was generated by joining two imidazolium or pyrrolidinium cations with different length hydrocarbon linkage chain. Dicationic ILs exhibited thermal stabilities over 350 °C, which were much higher than those of traditional monocationic ILs. The solvation properties of geminal

9 dicationic ILs were found to be close to those of their monocationic analogues.

Armstrong and co-workers proposed the synthesis of a novel class of dicationic ILs using poly(ethylene glycol) (PEG) as the linkage chain [26]. The PEG-linked geminal dicationic ILs showed thermal stability up to 350 °C, and the column efficiency was found to be the unchanged after conditioning the column to 320 °C. Compared to the hydrocarbon-linked geminal dicationic ILs, the PEG-linked geminal dicationic ILs exhibited slightly higher hydrogen bond basicity and dipolarity/polarizability due to the additional oxygens within the PEG linkers. The separation of complex samples such as a mixture of 24 flavor and fragrance compounds indicated that the PEG-linked dicationic

IL column possessed somewhat better separation selectivities compared to commercial columns such as Innowax and Rtx-5.

More recently, a novel class of trigonal tricationic ILs were synthesized and applied as GC stationary phases [19]. The structure of these ILs was constituted by the attachment of three identical imidazolium or phosphonium moieties attached to a core.

Four different core structures were examined, namely mesitylene core, benzene core, triethylamine core, and tri(2-hexanamido)ethylamine core. The ILs containing the tri(2-hexanamido)ethylamine core exhibited significantly stronger dipolarity/polarizability, hydrogen bond acidity, and hydrogen bond basicity compared to their monocationic or dicationic analogues. This was presumably due to the rigid trigonal geometry as well as the presence of amide group. Separations of the Grob test mixture, a homologous alkane/alcohol mixture, and a flavor and fragrance mixture were successfully performed using the trigonal tricationic IL columns. Results indicated that the IL containing the tri(2-hexanamido)ethylamine core not only exhibited the highest

10 separation efficiency for all the studied mixtures but also was the first IL-based stationary

− phase (containing NTf2 anion) that eliminated peak tailing for separation of alcohols and other hydrogen bonding analytes.

1.4.3 Application of polymeric ionic liquids as GC stationary phases

Due to the fact that most ILs have significantly decreased viscosities at elevated temperature, it is often problematic to maintain the homogeneous and consistent film throughout the capillary column when ILs are used as stationary phases. In order to overcome this limitation, polymeric ionic liquids (PILs) were developed and characterized as GC stationary phases [20]. Partially or highly crosslinked PIL stationary phases were prepared by free radical polymerization. It was observed that by crosslinking the IL monomers, the PIL stationary phase provided efficient separations up to 350 °C when a highly crosslinked PIL was used. The effect of ratio of monomer and crosslinker were studied in order to obtain the optimized separation performance. Four different crosslinkers were mixed for the highly crosslinked PIL stationary phases, and the ratio was optimized to produce a gummy or waxy stationary phase. The solvation thermodynamic and interaction parameter study indicated that the PIL stationary phases retained the dual nature separation selectivity.

More recently, a series of poly(vinylimidazolium)-based PIL stationary phases were developed [27,28]. The effects of various substituents and different anions on the solvation properties of the resulting PILs were thoroughly examined using the solvation parameter model. It was observed that the functional groups (i.e. alkyl group or phenyl group) on the polymerized cation had the potential to provide additional separation

11 selectivity whereas the counter anion had more pronounced effects on the solvation properties, separation selectivity, and thermal stability of the PIL stationary phase. Some of the PIL stationary phases exhibited interesting separation selectivity towards xylene isomers as well as proton donor analytes.

Xing and co-workers proposed the synthesis of IL bonded polysiloxane polymer and its application as a GC stationary phase [29]. Two different counter anions, namely Cl−

− and NTf2 , were paired with the 1-propyl-3-methylimidazolium cationic moieties appended on the polysiloxane backbone. Thermogravimetric analysis data revealed that by attaching the ionic moieties to the polysiloxane backbone, the resulting polymer exhibited thermal stability of 380 °C. The solvation properties of these two columns were characterized using the solvation parameter model, and the results indicated that the polymer containing the Cl− anion possessed a significantly higher hydrogen bond basicity

− compared to the one containing the NTf2 anion. A wide range of analytes including the

Grob mixture, selected aromatic isomers, fatty acid methyl esters, polychlorinated biphenyls, n-alkanes, and polycyclic aromatic hydrocarbons were successfully separated using the IL-based columns.

1.4.4 Application of ionic liquid mixtures as GC stationary phases

Mixed-mode stationary phases have been widely used in GC due to the unique separation selectivities that often cannot be observed when using neat stationary phases.

In the same manner, two monomeric ILs differing in the associated counter anion, namely

BMIM-Cl and BMIM-NTf2, were mixed at various ratios and examined as GC stationary phases [30]. The solvation properties of the IL-mixture columns were characterized using

12 the solvation parameter model, and the result indicated that the hydrogen bond basicity increased linearly as the BMIM-Cl content was increased. It was found that the retention factor of many analytes including alcohols and aromatic compounds can be significantly altered by simply changing the composition of the IL mixture. In addition, many proton donor analytes exhibited a reversal of elution order on the column containing higher weight percentage of BMIM-Cl. More recently, GC stationary phases composed of binary mixtures of two PILs, namely poly(1-vinyl-3-hexylimidazolium)-NTf2

(poly(VHIM-NTf2))/poly(1-vinyl-3-hexylimidazolium)-Cl (poly(VHIM-Cl)) and poly(1-vinyl-3-hexadecylimidazolium)-NTf2 (poly(VHDIM-NTf2))/poly(1-vinyl-3- hexadecylimidazolium)-Cl (poly(VHDIM-Cl)), were evaluated in terms of their on-set bleed temperature and separation selectivity (see Chapter 4) [31]. Compared to their monomeric analogue stationary phases, the stationary phases consisting of the binary mixture of PILs exhibited improved thermal stability. The hydrogen bond basicity of the mixed poly(VHIM-NTf2)/poly(VHIM-Cl) stationary phases was enriched linearly with the increase in the Cl-based PIL content. Results indicated that tuning the composition of the stationary phase allowed for fine control of retention factors and separation selectivity for a wide range of molecules including alcohols, carboxylic acids, as well as selected ketones, aldehydes, and aromatic compounds.

1.4.5 Application of functionalized ionic liquids as GC stationary phases

One of the most unique and exciting properties of ILs is their tunability, meaning that one can introduce functional groups to the cation and/or anion part of IL and impart the resulting IL with desired functionality. Armstrong and co-worker first synthesized ILs

13 containing benzyl and methoxyphenyl groups as GC stationary phases [17]. By introducing the benzyl/methoxyphenyl group to the cationic moiety, the thermal stability of the resulting ILs was significantly enhanced. In addition, these ILs exhibited a stronger capability to interact with solutes via π-π interactions, as suggested by their solvation properties. This resulted in greatly enhanced separation selectivity of polycyclic aromatic hydrocarbons, polychlorinated biphenyl, and aromatic sulfoxides when these ILs were used as stationary phases, in particular, the 1-(4-methoxyphenyl)-3-methyl imidazolium trifluoromethanesulfonate IL.

ILs containing various functional groups, namely primary amine, ester, and hydroxyl groups, were characterized using the solvation parameter model (see Chapter 2) [32]. It was observed that the hydrogen bond acidity, hydrogen bond basicity, and dipolarity of the IL-based stationary phase can be modified through the introduction of the functional group. Particularly, the amine-functionalized IL exhibited significantly increased hydrogen bond basicity and slightly increased dipolarity compared to its unfunctionalized analogue. The presence of the ester and hydroxyl group resulted in enhanced hydrogen bond acidity, largely due to the acidic proton originating from these functional groups.

The chromatographic performance of the functionalized IL stationary phases was evaluated by examining the retention behavior and separation selectivity for selected probes.

Álvarez and co-workers proposed the introduction of 2′-hydroxy-cyclohexyl and

2′-acetyl-cyclohexyl functionality to the cationic moiety of imidazolium-based ILs

− containing BF4 anion [33]. It was observed that with the presence of the

2′-hydroxy-cyclohexyl group, the resulting IL exhibited good thermal stability of 250 °C.

14 The solvation properties of the functionalized ILs were examined using the solvation parameter model. Unique selectivity was found for many analytes, including alkanes, ketones, esters, and aromatic compounds when the functionalized ILs were used as GC stationary phases.

1.4.6 Application of ionic liquids as chiral GC stationary phases

Efforts have been made in order to use IL-based GC stationary phases for chiral separation. Armstrong and co-workers first utilized ILs as solvents to dissolve permethylated-β-cyclodextrin and dimethylated-β-cyclodextrin to prepare stationary phases for chiral GC separation [34]. However, the enantiomeric separation selectivity of the IL-cyclodextrin stationary phases was much lower compared to the commercial columns. It was proposed that the imidazolium ion pair could make an inclusion complex with the cyclodextrin cavity and block it from chiral recognition. More recently, ionic cyclodextrin derivatives were dissolved in ILs and applied to chiral separation [35].

Broader enantioselectivity, significantly enhanced separation efficiency, and improved thermal stability were observed for the IL-ionic cyclodextrin columns compared to the

IL-neutral cyclodextrin columns. When compared to an analogous commercial column, the IL-ionic cyclodextrin columns exhibited improved enantioselectivity for many chiral molecules, better peak shapes for polar analytes, and some complementary enantioseparation.

Chiral ILs are another class of IL-based stationary phases for chiral GC [36]. A total of three different chiral ILs were synthesized and coated onto fused silica capillary columns. The results demonstrated that the studied chiral IL-based stationary phases

15 exhibited enantioselective retention for chiral alcohols (including diols), sulfoxides, acetylated amines, and some chiral epoxides. Interestingly, a reversed enantiomeric elution order was observed for many chiral molecules when different enantiomers of the chiral ILs were used as the stationary phase. This unique property allowed one to modify the structure of the chiral IL to result in a desired enantiomeric elution order.

1.5 Summary

This chapter provides an overview of the wide application of ILs as GC stationary phases. A brief description about the unique physico-chemical properties of ILs and various advantages of using ILs as GC stationary phases is presented.

The following three chapters of this dissertation provide aspects of characterizing

ILs using the solvation parameter model and applying ILs as GC stationary phases with enhanced selectivity and/or thermal stability. Chapter Two introduces the characterization of functionalized ILs containing tris(pentafluoroethyl)trifluorophosphate (FAP) anion using solvation parameter model. The solvation properties of ILs, including hydrogen bond acidity, hydrogen bond basicity, and dipolarity, can be tuned by introducing functional groups to the cationic moiety. Compared to the NTf2-based analogues, the

FAP-based ILs possessed lower hydrogen bond basicity but higher hydrogen bond acidity.

The FAP anion was found to weakly coordinate the cation and any appended functional groups, promoting properties of the cation which might be masked by stronger interaction with other anion system. The separation performance of the studied IL-based stationary phases was evaluated by examining the retention behavior and separation selectivity for

16 selected analytes.

Chapter Three delineates the evaluation of the solvation properties of functionalized

ILs containing varied cation/anion composition. A total of 16 different ILs were examined using the solvation parameter model. The effect of cation functional groups, cation type, and anion on the system constants was examined. It was observed that the presence of the cation functional group affected the hydrogen bond basicity, hydrogen bond acidity, as well as dispersion interactions of the resulting ILs, while the change of cation type yielded modest influence on the dipolarity. The switch of counter anions in unfunctionalized ILs produced compounds with higher dipolarity and hydrogen bond basicity. Interestingly, the dipolarity and hydrogen bond basicity of ILs possessing cyano-containing anions appeared to be inversely proportional to the cyano content of the anion.

Chapter Four presents the development of highly selective GC stationary phases consisting of binary mixtures of PILs, namely poly(VHIM-NTf2)/poly(VHIM-Cl) and poly(VHDIM-NTf2)/poly(VHDIM-Cl). A total of six neat or binary PIL stationary phases were evaluated in terms of their on-set bleed temperature and solvation properties. The hydrogen bond basicity of the poly(VHIM-NTf2)/poly(VHIM-Cl) increased linearly with the increase of poly(VHIM-Cl) weight percentage. The poly(VHDIM)-based PIL stationary phases exhibited stronger dispersion interaction compared to the poly(VHIM)-based PILs. It was observed that tuning the composition of the PIL stationary phase allowed for fine control of retention factors and separation selectivity for many molecules.

17 Chapter 2

Using the Solvation Parameter Model to Characterize

Functionalized Ionic Liquids Containing the Tris-

(pentafluoroethyl)trifluorophosphate (FAP) Anion

A paper published in Analytical and Bioanalytical Chemistry 1

Qichao Zhao, Jens Eichhorn, William R. Pitner, Jared L. Anderson

Abstract

Ionic liquids (ILs) containing the tris(pentafluoroethyl) trifluorophosphate anion

[FAP]− have attracted increased attention due to their unique properties including ultrahigh hydrophobicity, hydrolytic stability, and wide electrochemical window. In this study, the solvation parameter model is used via gas chromatography to characterize the solvation interactions of seven ILs containing amino, ester, and hydroxyl functional

1 Reprinted from Analytical and Bioanalytical Chemistry, 2009, 395, 225-234. Copyright

18 groups appended to the cation and paired with [FAP]−, as well as three ILs containing the

− bis[(trifluoromethyl)sulfonyl]imide anion [NTf2] . The role of the functional groups, nature of the counter anion, and cation type on the system constants were evaluated. ILs

− containing [FAP] possessed lower hydrogen bond basicity than NTf2-based ILs having the same cationic component; in the case of hydroxylfunctionalized cations, the presence of [FAP]− led to an enhancement of the hydrogen bond acidity, relative to the

− NTf2-analogs. The system constants support the argument that [FAP] weakly coordinates the cation and any appended functional groups, promoting properties of the cation which might be masked by stronger interactions with other anion systems. The chromatographic performance of the IL stationary phases was evaluated by examining the retention behavior and separation selectivity for chosen analytes. The results from this work can be used as a guide for choosing FAP-based ILs capable of exhibiting desired solvation properties while retaining important physical properties including high thermal stability and high hydrophobicity.

2.1 Introduction

Ionic liquids (ILs) refer to a class of nonmolecular ionic compounds that have melting points below 100 °C. This class of unique solvents often results from the combination of substituted nitrogen or phosphorus-containing cations and various counter anions. Compared to traditional organic solvents, ILs have a number of significant advantages due to their unique properties. For example, many ILs possess very low vapor pressure at ambient temperature and exhibit high thermal stability. The physical and

19 chemical properties of ILs, including viscosity and solubility with other solvents, can be tailored by altering the combination of cations and anions as well as by introducing functional groups to either component. Numerous ILs have been designed and applied in many chemical investigations, including organic synthesis [1–4], liquid–liquid extractions [5–7], analytical microextractions (Yao et al., submitted) [8–11], mass spectrometry [12–14], electrochemistry [15, 16], and separation science [17–19].

Perfluoroalkylfluorophosphate (FAP) anions are a promising means of overcoming

− the hydrolytic instability of the hexafluorophosphate anion [PF6] [20, 21]. Recently, the synthesis of novel ILs containing FAP anions was demonstrated by Ignat’ev and coworkers [22]. It was reported that FAP-based ILs possessed numerous interesting properties as well as high hydrolytic stability. Thermal gravimetric analysis of this class of ILs has revealed that imidazolium-based ILs decompose at temperatures above 280 °C.

FAP-based ILs exhibit impressively high hydrophobicity [23], as well as broad electrochemical windows. Endres and coworkers demonstrated the use of FAP-based ILs as a solvent for the electropolymerization of benzene due to the fact that they are more chemically mild, colorless, and stable compared to traditionally employed solvents [24].

Duffy and Bond utilized FAP-based ILs as a supporting electrolyte in toluene [25]. The toluene/FAP-based IL solution was applied to electrochemical techniques in which aromatic solvents are not available, producing well-defined voltammetric measurements.

The FAP-based ILs have also been shown to electrowet a smooth fluoropolymer surface

[26]. In addition, FAP-based ILs have been applied to gas absorption studies [27, 28].

Brennecke and coworkers found that the FAP-based ILs possessed higher CO2 solubility than most other ILs, which is believed to be due to the increased fluoroalkyl chain [27].

20 All of these applications clearly indicate that the FAP anion imparts unique solvation characteristics to the IL. Therefore, it is important to understand how FAP-based ILs differ from ILs containing other anions, and how the solvation properties can be varied by pairing with functionalized cations.

The solvation parameter model [29], developed by Abraham and coworkers, has been successfully employed to evaluate the solvation properties of a wide class of ILs

[30–34]. The solvation parameter model, shown in Eq. 1, is a linear free-energy relationship that describes the contribution of individual solvation interactions of a solvent (e.g., IL) by examining solute/solvent interactions.

log k = c + eE + sS + aA + bB + lL (1)

According to Eq. 1, log k is the solute retention factor and is calculated by measuring the retention time of the analyte and dead volume of the chromatographic column. The solute descriptors (E, S, A, B, L) are probe-specific parameters that have been determined for many molecules [29]. They are defined as: E, the excess molar refraction calculated from the solute’s refractive index; S, the solute dipolarity/polarizability; A, the solute hydrogen bond acidity; B, the solute hydrogen bond basicity; and L, the solute gas hexadecane partition coefficient determined at 298 K. The system constants (e, s, a, b, l) are used to characterize the strength of each solvation interaction and are defined as: e, the ability of the IL to interact with π and nonbonding electrons of the solute; s, a measure of thedipolarity/polarizability of the IL; a, the IL hydrogen bond basicity; b, a measure of the hydrogen bond acidity of the IL; and l describes IL dispersion forces. The system constants are attained through multiple linear regression analysis of the log k term and the five solute descriptors. The intercept term, c,

21 can be used to determine and verify the phase ratio of the column.

In this work, a series of seven FAP-based ILs (four of which contain functionalized cations) was examined using the solvation parameter model. This is the first report to examine the solvation characteristics of FAP-based ILs and also the first report to examine a diverse group of functionalized ILs using the solvation parameter model. A fundamental understanding into the solvation characteristics of these unique solvents will provide additional insight in choosing the appropriate IL(s) for specific applications. To investigate the effect of the counter anion on solvation properties, ILs with the same cations were paired with tris (pentafluoroethyl)trifluorophosphate [FAP]− and

− bis[(trifluoromethyl) sulfonyl]imide [NTf2] . The effect of the nature of the cation (i.e., imidazolium, pyrrolidinium, and phosphonium), functional group substituents, and counter anions on the chromatographic retention factor and separation selectivity for selected solute molecules was explored and will be discussed.

2.2 Experimental

2.2.1 Materials

Seven FAP-based ILs, namely 1-butyl-1-methylpyrrolidinium FAP,

1-(6-aminohexyl)-1-methylpyrrolidinium FAP [35], 1-ethoxycarbonylmethyl

-1-methylpyrrolidinium FAP, 1-(2-hydroxyethyl)-1-methylpyrrolidinium FAP,

1-hexyl-3-methylimidazolium FAP, 1-(2-hydroxyethyl)-3-methylimidazolium FAP, and trihexyl(tetradecyl)phosphonium FAP, as well as two NTf2-based ILs,

1-(2-hydroxyethyl)-1-methylpyrrolidinium NTf2 and 1-(2-hydroxyethyl)-3-

22 methylimidazolium NTf2, were synthesized by Merck KGaA (Darmstadt, Germany) [22,

36]. The identity of each IL was confirmed with 1H-NMR spectroscopy and electrospray ionization mass spectrometry. These spectra are included as Electronic supplementary material (see Figure A-1 to A-18 in Appendix A). Halide content was determined by ion chromatography using a Metrohm system equipped with a model 820 IC separation center, a 819 IC detector, and a Metrosepp A Supp 5 column; the eluent was an aqueous

−3 −1 −3 −1 mixture containing 3.2×10 mol L Na2CO3, 1.0×10 mol L NaHCO3, and 5% acetonitrile. Water content was determined coulometrically using a Metrohm model 831

Karl-Fischer Coulometer with CombiCoulomat fritless Karl-Fischer reagent. All studied

ILs contained less than 100 ppm halide.

A total of 46 probe molecules with varied functional groups were selected for the characterization of the gas chromatographic IL stationary phases using the solvation parameter model. Acetic acid, methyl caproate, naphthalene, and propionic acid were purchased from Supelco (Bellefonte, PA, USA). Bromoethane, butyraldehyde, ethyl acetate, and 2-nitrophenol were purchased from Acros Organics (Morris Plains, NJ,

USA). 1-Butanol, N,N-dimethylformamide, 2-propanol, and toluene were purchased from

Fisher Scientific (Fairlawn, NJ, USA) and p-cresol, m-xylene, o-xylene, and p-xylene were purchased from Fluka (Steinheim, Germany). Cyclohexanol was purchased from J.T.

Baker (Phillipsburg, NJ, USA); ethyl benzene from Eastman Kodak Company (Rochester,

NJ, USA); and acetophenone, aniline, benzaldehyde, benzene, benzonitrile, benzyl alcohol, 1-bromohexane, 1-bromooctane, 2-chloroaniline, 1-chlorobutane,

1-chlorohexane, 1-chlorooctane, cyclohexanone, 1,2-dichlorobenzene, 1,4-dioxane,

1-iodobutane, nitrobenzene, 1-nitropropane, 1-octanol, octylaldehyde, 1-pentanol,

23 2-pentanone, phenetole, phenol, propionitrile, pyridine, pyrrole, and 1-decanol were purchased from Sigma-Aldrich (St. Louis,MO, USA). All probe molecules were used as received. Methylene chloride was purchased from Fisher Scientific. Untreated fused silica capillary tubing (0.25-mm ID) was obtained from Supelco.

2.2.2 Method

Five-meter untreated fused silica capillary columns were coated at 40 °C using the static method. All coating solutions contained 0.45% (w/v) of the studied IL in methylene chloride, except for the 1-butyl-1-methylpyrrolidinium FAP IL, which was prepared at a concentration of 0.25% (w/v) due to the tendency of the evaporation interface to cease movement at higher concentrations. All coated columns were conditioned from 30 to

110 °C at 1 °C/min and held for 1 h, using a constant helium flow at a rate of 1.0 mL/min.

Column efficiency was tested with naphthalene at 100 °C. All coated columns had efficiencies of at least 1,400 plates per meter. Column efficiencies were monitored by recording the retention times of the probe molecules at three temperatures to ensure that the column coating did not change throughout the characterization of the stationary phase.

The 46 probe molecules and their corresponding solute descriptors used in this study are provided as “Electronic supplementary material.” All probe molecules were dissolved in methylene chloride and injected separately at three temperatures (50, 80, and 110 °C).

Some probes with very low boiling points and weak interactions with the IL eluted with the solvent peak, particularly at higher temperatures. On the other hand, other probes exhibited very strong interactions with the stationary phase and were retained in the

24 column for 3 h or longer. Therefore, not all 46 probes could be subjected to multiple linear regression analysis at all temperatures examined. All separations were performed using a Hewlett-Packard 6890 gas chromatograph equipped with flame ionization detector. Helium was used as the carrier gas with a flow rate of 1.0 mL/min. The injector and detector were held at 250 °C. The detector makeup flow of helium was maintained at

23 mL/min, the hydrogen flow at 40 mL/min, and the air flow rate at 450 mL/min.

Methane was used to measure the dead volume of each column at the three different temperatures. Multiple linear regression analysis and statistical calculations were performed using the program Analyze-it (Microsoft, USA).

2.3 Results and Discussion

The ILs in this study were carefully chosen to compare the effect of the cation and anion on the system constants. Figure 2.1 shows the structures of the ILs that were examined in this study, as well as the numbering system that will be used to refer to the

ILs throughout this manuscript.

Seven of the ILs contain functionalized imidazolium, pyrrolidinium, or phosphonium cations paired with [FAP]−. For comparison, three ILs containing the same cation

− component paired with [NTf2] were evaluated. Table 2.1 lists the system constants of the ten ILs evaluated in this study. The system constants for two ILs,

1-hexyl-3-methylimidazolium FAP and 1-butyl-1-methylpyrrolidinium NTf2, were obtained by extrapolation of system constants obtained at 40, 70, and 100 °C. For most of the ILs evaluated, smooth decreases in the system constants are observed with increasing

25 1 2 3 4

5 6 7

8 9 10

Figure 2-1: Structures and numbering system of ILs that were examined in this study. (1) 1-butyl-1-methylpyrrolidinium FAP, (2)

1-(6-aminohexyl)-1-methylpyrrolidinium FAP, (3) 1-ethoxycarbonylmethyl-1-methylpyrrolidinium FAP, (4) 1-(2-hydroxyethyl)

-1-methylpyrrolidinium FAP, (5) 1-hexyl-3-methylimidazolium FAP, (6) 1-(2- hydroxyethyl)-3-methylimidazolium FAP, (7) trihexyl

(tetradecyl)phosphonium FAP, (8) 1-butyl-1-methylpyrrolidinium NTf2, (9) 1-(2-hydroxyethyl)-1-methylpyrrolidinium NTf2, (10)

1-(2-hydroxyethyl)-3-methylimidazolium NTf2.

26 - - Table 2.1 System constants of [FAP] and [NTf2] -based IL stationary phases examined in this study.

IL IL stationary phase/temperature System constantsa No (°C) c e s a b l n R2 F 1 1-butyl-1-methylpyrrolidinium tris(pentafluoroethyl)trifluorophosphate 50 -2.86 0.21 1.61 0.85 0.68 0.54 35 0.99 404 (0.08) (0.11) (0.08) (0.14) (0.02) 80 -3.04 0.21 1.49 0.67 0.66 0.47 34 0.99 403 (0.08) (0.10) (0.07) (0.13) (0.02) 110 -3.22 0.18 1.44 0.56 0.58 0.42 34 0.99 436 (0.07) (0.09) (0.06) (0.12) (0.02)

2 1-(6-aminohexyl)-1-methylpyrrolidinium tris(pentafluoroethyl)trifluorophosphate 50 -2.58 0.26 1.73 2.11 0.48 0.51 29 0.98 234 (0.11) (0.14) (0.21) (0.22) (0.03) 80 -2.57 0.22 1.62 1.61 0.40 0.41 29 0.98 243 (0.09) (0.12) (0.17) (0.19) (0.02) 110 -2.97 0.18 1.76 1.69 0.21 0.36 32 0.98 255 (0.10) (0.12) (0.12) (0.17) (0.03)

3 1-ethoxycarbonylmethyl-1-methylpyrrolidinium tris(pentafluoroethyl)trifluorophosphate 50 -2.86 0 1.99 0.73 1.17 0.53 36 0.99 519 (0.09) (0.08) (0.13) (0.02) 80 -2.86 0 1.78 0.58 0.94 0.43 37 0.99 707 (0.07) (0.06) (0.10) (0.01) 110 -2.82 0.04 1.58 0.45 0.80 0.36 36 0.99 648 (0.05) (0.06) (0.05) (0.09) (0.01)

27 Table 2.1 (continued) IL IL stationary phase/temperature System constantsa No (°C) c e s a b l n R2 F

4 1-(2-hydroxyethyl)-1-methylpyrrolidinium tri(pentafluoroethyl)trifluorophosphate 50 -2.90 0.25 1.73 0.95 2.22 0.49 36 0.99 470 (0.09) (0.11) (0.09) (0.14) (0.02) 80 -3.00 0.23 1.64 0.75 1.81 0.41 36 0.99 575 (0.07) (0.09) (0.07) (0.12) (0.02) 110 -3.10 0.22 1.56 0.60 1.53 0.35 36 0.99 596 (0.06) (0.08) (0.06) (0.10) (0.01)

5 1-Hexyl-3-methylimidazolium tris(pentafluoroethyl)trifluorophosphate b 50 -2.59 0 1.79 0.94 0.72 0.57 43 0.98

80 -2.53 0 1.59 0.67 0.65 0.48 43 0.99

110 -2.47 0 1.40 0.40 0.58 0.39 41 0.98

6 1-(2-hydroxyethyl)-3-methylimidazolium tris(pentafluoroethyl)trifluorophosphate 50 -2.94 0.08 2.01 1.09 2.06 0.46 37 0.99 454 (0.09) (0.11) (0.09) (0.14) (0.02) 80 -2.99 0 1.91 0.83 1.72 0.38 38 0.99 532 (0.09) (0.08) (0.12) (0.02) 110 -2.95 0 1.72 0.65 1.40 0.31 37 0.99 511 (0.07) (0.06) (0.10) (0.01)

28 Table 2.1 (continued) IL IL stationary phase/temperature System constantsa No (°C) c e s a b l n R2 F

7 Trihexyl(tetradecyl)phosphonium tris(pentafluoroethyl)trifluorophosphate 50 -2.59 -0.37 1.56 0.74 0.26 0.73 39 0.99 619 (0.07) (0.08) (0.07) (0.11) (0.02) 80 -2.69 -0.29 1.39 0.49 0.25 0.63 39 0.99 744 (0.06) (0.07) (0.06) (0.09) (0.01) 110 -2.99 -0.32 1.26 0.43 0.29 0.61 39 0.99 528 (0.06) (0.08) (0.07) (0.10) (0.02)

8 1-butyl-1-methylpyrroldinium bis[(trifluoromethyl)sulfonyl]imideb,c 50 -2.79 0 1.64 1.98 0 0.64 34 0.98

80 -2.86 0 1.51 1.72 0 0.54 34 0.98

110 -2.93 0 1.39 1.45 0 0.45 34 0.98

9 1-(2-hydroxyethyl)-1-methylpyrrolidinium bis[(trifluoromethyl)sulfonyl]imide 50 -3.16 0.27 1.96 2.15 1.41 0.50 35 0.99 596 (0.09) (0.11) (0.09) (0.14) (0.02) 80 -3.10 0.24 1.84 1.78 1.10 0.41 39 0.99 939 (0.06) (0.07) (0.06) (0.10) (0.01) 110 -3.25 0.21 1.80 1.58 0.90 0.35 37 0.99 854 (0.05) (0.06) (0.06) (0.09) (0.01)

29 Table 2.1 (continued) IL IL stationary phase/temperature System constantsa No (°C) c e s a b l n R2 F

10 1-(2-hydroxyethyl)-3-methylimidazolium bis[(trifluoromethyl)sulfonyl]imide 50 -3.23 0.25 2.10 2.16 1.39 0.48 35 0.99 902 (0.07) (0.09) (0.08) (0.11) (0.02) 80 -3.30 0.19 2.04 1.93 1.11 0.39 35 0.99 866 (0.06) (0.08) (0.07) (0.10) (0.01) 110 -3.40 0.10 1.97 1.75 0.98 0.36 34 0.98 236 (0.06) (0.10) (0.10) (0.13) (0.02)

n: Number of probe analytes subjected to multiple linear regression, R2: Correlation coefficient, F: Fisher coefficient a System constants: e = nonbonding and π electron interactions, s = IL dipolarity, a = IL hydrogen bond basicity, b = IL hydrogen bond acidity, l = IL dispersion forces bData extrapolated from system constants obtained under 40, 70 and 100 °C. cData extrapolated from ref. [31]

30 temperature. The standard deviation value for each system constant, correlation coefficients of the regression line, and Fisher coefficients indicate that all generated models are statistically sound.

2.3.1 Effects of cation functional group on system constants

In general, the hydrogen bond acidity of unfunctionalized ILs is determined by the cation and modulated by the anion [30, 37]. It was reported previously that simple introductions or alterations of functional groups to the cationic component of the IL are capable of imparting unique solvation properties to the IL, making them particularly useful solvents for retaining enzymatic activity [38]. In particular, ILs containing sulfone-

− − and sulfoxide functionalized cations paired with [NTf2] and [BF4] accentuated specific solvation interactions. In pursuit of identifying functional groups that impart unique solvation characteristics to the IL, the system constants of four FAP based ILs containing substituents and functional group appended to pyrrolidinium cations were determined.

These ILs (1, 2, 3, and 4 in Fig. 2.1) contain an alkyl substituent, amino functionality, an ester group, and a hydroxyl moiety, respectively.

IL 1 possesses a butyl substituent, while IL 2 contains an aliphatic chain with a pendant amine group. At 50 °C, 2.5-fold increase in the hydrogen bond basicity (a term) was observed due to the introduction of the free amine moiety to the cation. This is due to the availability of the lone pair of electrons from the amine group to interact with proton-donating solute molecules. The hydrogen bond acidity of the IL (b term) dropped slightly with the presence of the amine group, while the dipolarity (s term) exhibited a small increase. All other system constants for these two ILs were similar.

31 A comparison of ILs 1 and 3 revealed that the substitution of an ester functional group for the alkyl chain yielded a 1.7-fold increase in the hydrogen bond acidity at 50 °C.

The weakly acidic protons (pKa≈20) [39] originate from the α-carbon adjacent to the ester group (see Fig. 2.1). Therefore, this IL interacted more strongly with hydrogen bond basic solutes, including amines and amides. The introduction of the ester moiety also increased the dipolarity of the IL but provided no significant changes to the hydrogen bond basicity or cohesive interactions.

With the presence of the hydroxyl functionality, IL 4 exhibits a threefold increase in hydrogen bond acidity at 50 °C compared to IL 1. This strong proton-donating ability of the solvent results in increased retention times for basic solute molecules. All other system constants were largely unchanged. The same trend of increasing hydrogen bond acidity was observed when hydroxyl functionality was incorporated into the imidazolium cation. For example, the hydrogen bond acidity of IL 6 increases over 2.8-fold compared to IL 5, which contains an alkyl substituent. Dipolar and hydrogen bond basicity interactions were higher on IL 6, whereas IL 5 was more cohesive.

2.3.2 Effects of cation type on system constants

Ionic liquids 1, 5, and 7 all contain [FAP]− and possess alkyl chain substituents of various lengths on pyrrolidinium, imidazolium, and phosphonium cations, respectively.

ILs 1 and 5 possess similar values for all system constants. A comparison of these two

ILs to IL 7, containing the heavily alkylated phosphonium cation, reveals that the hydrogen bond basicity, hydrogen bond acidity, and cohesive forces are all significantly influenced by the highly hydrophobic cation. Specifically, a slight decrease in hydrogen

32 bond basicity and dipolar interactions was observed whereas the hydrogen bond acidity dropped by over half when examining the phosphonium-based IL. As expected, this IL exhibited the highest cohesion of all the solvents evaluated in this study due to the presence of the long-alkyl-chain substituents: the dispersive forces in IL 7 are at their highest, while the potential for hydrogen bonding are at their lowest.

In another comparison, ILs 4 and 6 possess hydroxyl moieties and are paired with

[FAP]−, whereas ILs 9 and 10 are both hydroxyl-functionalized pyrrolidinium- and

− imidazolium-based ILs containing [NTf2] . A comparison of system constants for these two pairs of ILs reveals that the system constants of the imidazolium and pyrrolidinium-based compounds are nearly identical at all three examined temperatures, regardless of the nature of the counter anion. Interestingly, a comparison of the unfunctionalized [FAP]− ILs (1 and 5) shows that the hydrogen bond acidity is lower for the pyrrolidinium-based IL compared to the imidazolium-based IL, as expected due to the presence of the acidic hydrogen within the imidazolium cation. However, when these cations are functionalized with the hydroxyl group, the pyrrolidinium-based IL 4 attains a higher hydrogen bond acidity than the imidazolium-based IL 6.

2.3.3 Anion effects on system constants

Previous reports have demonstrated that the hydrogen bond basicity is determined by the anion and can be modulated by the cation [30, 37, 38]. To examine the differences

− − in solvation characteristics between [FAP] and [NTf2] , one can compare the system constants of ILs containing the same cation. In the case of ILs 1 and 8, the hydrogen bond

− − basicity at 50 °C decreased 2.5-fold when the anion was switched from [NTf2] to [FAP] ,

33 while the hydrogen bond acidity increased significantly. The same trend was observed for

ILs containing hydroxyl-functionalized cations (i.e., ILs 4 and 9; 6 and 10). Therefore, when comparing ILs with identical cationic components, the FAP-based ILs possess

− lower hydrogen bond basicity relative to their [NTf2] -analogs. A further aspect of

[FAP]− that has not been previously observed with other anions is the tendency to promote the hydrogen bond acidity character of the accompanying cation. This property makes the FAP-based ILs an interesting class of solvents due to their proton donating capability, which can be further enhanced by combination with a hydroxyl-functionalized cation.

From the obtained system constants, the extent to which the cation-appended

− − functional groups are coordinated to [FAP] and [NTf2] can be examined. The purpose of appending functional groups to ILs is to impart additional interactions to the IL. If the anion strongly coordinates to the functional group, this limits the extent to which solutes are able to interact with the functional group(s). To classify the ILs examined in this study, a hydrogen bond acidity/basicity scheme is represented in Fig. 2.2. Two important trends are observed within this scheme. ILs containing [FAP]− are much less basic than

− [NTf2] -based ILs. Anions possessing lower basicity are typically more weakly coordinated compared to anions possessing large basicities (e.g., halides). In addition, the lower coordinating ability of the [FAP]− ILs is evident by noting the higher hydrogen bond acidity values of the hydroxyl-containing cations (4 and 6) compared to the

− analogous ILs containing [NTf2] (9 and 10)

Figure 2-2: Scheme demonstrating the acidity/basicity scales of the ten ILs evaluated in

− − this study. ILs containing [FAP] are much less basic than [NTf2] . The lower coordinating ability of the [FAP]− ILs is evident by noting the higher hydrogen bond acidity values of the hydroxylcontaining cations (4 and 6) compared to the analogous ILs

− containing [NTf2] (9 and 10).

35 2.3.4 Effect of IL stationary-phase composition on retention behavior for selected solute molecules

The vastly different solvation properties exhibited by the functionalized and unfunctionalized FAP-based ILs in this study provide interesting gas chromatographic retention characteristics for various solute molecules. Table 2.2 shows the retention factors for various solutes on four functionalized pyrrolidinium-based stationary phases containing [FAP]−. Alcohols exhibited a significant increase in retention on the stationary phase consisting of IL 2, due to the introduction of the amino group and subsequent increase in the hydrogen bond basicity. IL 3 and IL 4 possessed higher hydrogen bond acidity than IL 1, and therefore strongly retained hydrogen bond basic solutes including dimethylformamide (DMF). The retention factor of acetic acid experienced a 167% increase on the IL 3 stationary phase and a 400% increase on IL 4 compared to the IL 1 stationary phase. For aromatic compounds, such as benzene, xylene, and naphthalene, slight differences in the retention factors were observed on these four stationary phases.

Table 2.3 lists the retention factors for selected solutes on two FAP-based IL (ILs 4 and 6) and two NTf2-based IL (ILs 9 and 10) stationary phases. The ILs contained one of two different functionalized cations: 1-(3-hydroxyethyl)-1-methylpyrrolidinium (ILs 4 and 9) or 1-(3-hydroxyethyl)-3-methylimidazolium. Through the combination of two cations and two anions, the hydrogen bond basicity and acidity of the resulting ILs can be greatly varied. Proton donor solutes, such as acetic acid, benzyl alcohol, and phenol, were retained longer on NTf2-based IL stationary phases due to the inherently higher hydrogen

− bond basicity of [NTf2] . With the exception of acetic acid, all the molecules in this class were retained longer by the pyrrolidinium-based ILs, relative to their respective imidazol-

Table 2.2 Comparison of retention factors for selected solute molecules on four functionalised 1-methylpyrrolidinium [FAP]- IL stationary phases at 80 °C.

Functional Group alkyl 1° amine ester hydroxyl

Solute Molecule IL 1 IL 2 IL 3 IL 4

1-Butanol 0.2 0.8 0.4 0.9

1-Pentanol 0.4 1.6 0.8 1.5

1-Octanol 2.3 7.8 4.0 6.9

1-Decanol 6.7 22.1 11.8 18.7

Cyclohexanol 1.0 3.2 2.0 4.4

Acetophenone 12.7 21.5 25.6 37.6

Benzaldehyde 4.8 N/A 9.1 11.8

Cyclohexanone 3.0 N/A 7.1 12.3

DMF 7.8 14.1 30.1 142.8

Ethyl Acetate 0.2 0.5 0.5 0.7

Acetic Acid 0.3 N/A 0.8 1.5

Propionic Acid 0.6 N/A 1.3 2.3

Benzene 0.2 0.3 0.2 0.2

o-Xylene 0.9 1.5 1.1 1.1

Naphthalene 16.2 29.4 19.7 21.0

37 Table 2.3 Comparison of retention factors for selected solute molecules on stationary phases composed of 2-hydroxyethyl-functionalised ILs at 80 °C.

Anion [FAP]ˉ [NTf2]ˉ [FAP]ˉ [NTf2]ˉ

Cation Pyrrolidinium imidazolium

Solute Molecule IL 4 IL 9 IL 6 IL 10

Acetic Acid 1.5 2.7 1.8 3.0

Benzyl Alcohol 40.2 55.4 33.1 42.7

2-Chloroaniline 27.8 38.7 22.1 30.0

p-Cresol 21.7 66.5 20.7 54.8

Phenol 12.2 41.4 12.4 36.3

Acetophenone 37.6 21.4 33.5 17.8

Benzaldehyde 11.8 8.4 11.0 6.9

Cyclohexanone 12.3 5.4 11.4 4.6

Ethyl Acetate 0.7 0.3 0.7 0.2

2-Pentanone 2.0 0.8 2.0 0.7

Octylaldehyde 4.5 2.2 3.9 1.7

Methyl Caproate 2.8 1.2 2.5 0.9

Ethyl Benzene 0.7 0.6 0.6 0.4

o-Xylene 1.1 0.8 0.9 0.6

Naphthalene 21.0 16.5 15.8 11.8

38 ium analogs, although the effect was not as pronounced as that observed from a change in the anionic component. Interestingly, the interactions of aldehydes and ketones were stronger on the FAP-based IL stationary phases compared to ILs containing the same

− cation and paired with [NTf2] . For example, the retention factors of 2-pentanone, octylaldehyde, and cyclohexanone were doubled when the separation was performed on the FAP-based IL stationary phase. This indicates that FAP-based ILs may be particularly useful in separating lower-molecular-weight aldehydes, ketones, and esters where low column retention often limits the separation performance of the stationary phase.

2.3.5 Effect of IL stationary-phase composition on separation selectivity for selected solutes

The separation selectivity for selected solutes can be calculated by determining the ratio of two solute retention factors. Table 2.4 shows the separation selectivity for various solute pairs at 80 °C on four functionalized pyrrolidiniumbased FAP IL stationary phases.

It can be seen that with specific functional groups, the selectivity can be altered, allowing for increased control over the selectivity of some solutes. For example, the selectivity between cyclohexanol and benzene increased from 1.7 on IL 1 to 3.2 on IL 2. When the hydrogen bond acidity of the stationary phase was increased, the selectivity between

DMF and o-xylene increased from 4.6 on IL 1 to 14.6 on IL 3 and further increased to

68.6 on the IL 4 stationary phase. However, the selectivity between certain aromatic analytes (e.g., naphthalene and benzaldehyde) decreased, whereas the selectivities for other aromatic compounds (e.g., ethyl phenyl ether and ethyl benzene) were largely unchanged. For certain solute pairs, a reversal of elution order was observed. An example

Table 2.4 Comparison of the selectivity for selected solute molecules on stationary

phases composed of four functionalised 1-methylpyrrolidinium [FAP]- IL stationary at

80 °C.

Functional Group alkyl 1° amine ester hydroxyl Solute pair IL 1 IL 2 IL 3 IL 4 1-Octanol/1-bromooctane 1.5 3.1 2.0 3.8 1-Octanol/nitropropane 1.5 2.8 1.4 2.1 DMF/o-xylene 4.6 6.0 14.6 68.6 Methyl caproate/o-xylene 1.2 1.2 1.6 1.8 Acetic acid/benzene 1.1 N/A 1.4 2.0 1-Butanol/benzene 1.0 1.4 1.1 1.5 Ethyl acetate/benzene 1.0 1.1 1.2 1.4 Cyclohexanol/benzene 1.7 3.2 2.4 4.4 Alcohol/Alcohol 1-Pentanol/1-butanol 1.2 1.4 1.2 1.3 1-Octanol/1-butanol 2.8 4.9 3.5 4.2 1-Decanol/1-butanol 6.5 12.8 8.9 10.6 Cyclohexanol/1-butanol 1.7 2.4 2.1 2.9 Aromatic/Aromatic Naphthalene/nitrobenzene 1.5 1.5 0.97 a 0.99 a Naphthalene/acetophenone 1.2 1.4 0.77 a 0.57 a Naphthalene/benzyl alcohol 1.5 N/A 1.0 0.53 a Naphthalene/benzaldehyde 3.0 N/A 2.0 1.7 Naphthalene/o-xylene 9.0 12.0 9.7 10.5 2-Chloroaniline/naphthalene 0.92 a 1.2 1.0 1.3 Benzonitrile/ethyl benzene 4.9 6.3 8.2 10.4 Ethyl phenyl ether/ethyl benzene 2.2 2.7 2.5 2.8 a: By definition, the value of selectivity should not be smaller than unity. However, in

some cases the solute pairs exhibited reversed elution order, which makes it is impossible

to report selectivities greater than one for all IL columns.

40 of this selectivity can be observed for naphthalene and acetophenone.

Interesting separation selectivities were observed for solutes on FAP-based and

NTf2-based stationary phases. As shown in Table 2.5, the selectivities of some solutes were not notably influenced by switching the anion of IL stationary phase. The

FAP-based ILs exhibited increased selectivity for analytes containing a carbonyl group, such as aldehydes, ketones, and esters, over alcohols and aromatic compounds. For instance, the selectivity between cyclohexanone and cyclohexanol decreased from 2.4 on

IL 4 and 2.5 on IL 6, which are two FAP-based IL stationary phases, to 1.2 on the

NTf2-based IL 9 and IL 10 stationary phases. For proton donor solute bond acidity such as acetic acid, their selectivities over aromatic compounds increased when switching from the FAP-based IL to the NTf2-based ILs. For example, the selectivity between p-cresol and naphthalene increased from 1.0 on IL 4 to 3.9 on IL 9 and from 1.3 on IL 6 to 4.4 on IL 10. A reversal of elution order was also achieved for certain solutes by simply switching the counter anion. The selectivity between phenol and naphthalene increased from 0.60 on IL 4 to 2.4 on IL 9 and from 0.80 on IL 6 to 2.9 on IL 10. These observations demonstrate the selectivity advantage of employing functionalized cations in the IL to impart additional solvation interactions.

2.4 Conclusions

The hydrogen bond acidity of most ILs is largely determined by the cation and can be modulated by the anion, while the hydrogen bond basicity is determined by the anion and modulated by the cation. However, this study demonstrates that this is true unless

Table 2.5 Effect of 2-hydroxyethyl-functionalised ILs on the selectivity of chosen solute

pairs at 80 °C.

Anion [FAP]ˉ [NTf2]ˉ [FAP]ˉ [NTf2]ˉ Cation Pyrrolidinium imidazolium Solute pair IL 4 IL 9 IL 6 IL 10 Naphthalene/o-xylene 10.5 9.4 8.9 8.0 1-Octanol/1-butanol 4.2 3.9 3.7 3.2 1-Butanol/o-xylene 0.89 a 1.0 0.94 a 1.1 Naphthalene/benzonitrile 1.2 1.4 1.1 1.2 1-Bromooctane/benzene 1.6 1.7 1.7 1.4 1-Nitropropane/1-pentanol 1.5 1.2 1.5 1.2 Ethyl phenyl ether/o-xylene 2.3 2.0 2.2 1.8

Butyraldehyde/benzene 1.2 1.1 1.3 1.1 Ethyl Acetate/benzene 1.4 1.1 1.4 1.1 Acetophenone/decanol 2.0 1.2 2.4 1.5 Cyclohexanone/cyclohexanol 2.4 1.2 2.5 1.2 Methyl caproate/1-butanol 2.1 1.2 2.0 1.1 Octylaldehyde/1-pentanol 2.2 1.3 2.0 1.2 2-Pentanone/1-pentanol 1.2 0.71 a 1.2 0.75 a

Acetic acid/ethyl benzene 1.4 2.4 1.8 3.0 Benzyl alcohol/naphthalene 1.9 3.2 2.0 3.4 2-Chloroaniline/naphthalene 1.3 2.3 1.4 2.4 p-Cresol/naphthalene 1.0 3.9 1.3 4.4 Phenol/Naphthalene 0.60 a 2.4 0.80 a 2.9 a: By definition, the value of selectivity should not be smaller than unity. However, in

some cases the solute pairs exhibited reversed elution order, which makes it is impossible

to report selectivities greater than one for all IL columns.

42 there is a specific functional group with its own acidic or basic nature (i.e., the amino or ester group). For the first time, the solvation characteristics of functionalized ILs containing [FAP]− have been studied using the solvation parameter model. The hydrogen bond basicity, hydrogen bond acidity, and dipolarity of the resulting ILs can be modified through the introduction of amino, ester, and hydroxyl functional groups to pyrrolidinium, imidazolium, and phosphonium cations. The heavily alkylated phosphonium cation, paired with [FAP]−, exhibited the highest cohesivity and the weakest hydrogen bond interactions. Hydroxyl-functionalized imidazolium-based ILs possessed higher dipolarity interactions than their pyrrolidinium-based analogs. ILs containing [FAP]− possessed lower hydrogen bond basicity than NTf2-based ILs having the same cationic component; in the case of hydroxyl-functionalized cations, the presence of [FAP]− led to an enhancement of the hydrogen bond acidity, relative to the NTf2-analogs. The system constants support the argument that [FAP]− weakly coordinates the cation and any appended functional groups, promoting properties of the cation which might be masked by stronger interactions with other anion systems. This makes [FAP]− a useful component for modulating and tuning bulk solvation properties, particularly for organic synthesis applications. The unique separation selectivity offered by the functionalized

FAP-containing ILs in gas chromatography make them an interesting class of stationary phases. The highly hydrophobic nature of these compounds may make them acceptable stationary phases for analyzing samples with moderate water content without disruption of the stationary phase coating.

43 Acknowledgements

J.L.A. acknowledges funding from the Analytical and Surface Chemistry Program in the Division of Chemistry and the Separation and Purification Processes Program in the Chemical, Environmental, Bioengineering, and Transport Systems Division from the

National Science Foundation for a CAREER grant (CHE-0748612)

References

1. Welton T (1999) Chem Rev 99:2071-2083.

2. Wasserscheid P, Keim W (2000) Angew Chem Int Ed 39:3772-3789.

3. Cole AC, Jensen JL, Ntai I, Tran KR, Weaver KJ, Forbes DC, Davis JH (2002) J Am

Chem Soc 124:5962-5963.

4. Handy ST, Zhang X (2001) Org. Lett. 3:233-236.

5. Dai S, Ju YH, Barnes CE (1999) J Chem Soc Dalton Trans 8:1201-1202.

6. Chun S, Dzyuba SV, Bartsch RA (2001) Anal Chem 73:3737-3741.

7. Visser AE, Swatloski RP, Reichert WM, Mayton R, Sheff S, Wierzbicki A, Davis JH,

Rogers RD (2001) Chem Commun 135-136.

8. Liu JF, Jiang GB, Chi YG, Cai YQ, Zhou QX, Hu JT (2003) Anal Chem 75:5870-5876.

9. Liu J-F, Li N, Jiang G-B, Liu J-M, Jönsson JÅ, Wen M-J (2005) J Chromatogr A

1066:27–32.

10. Hsieh Y-N, Huang P-C, Sun I-W, Whang T-J, Hsu C-Y, Huang H-H, Kuei C-H (2006)

Anal Chim Acta 557:321–328.

44 11. Zhao F, Meng Y, Anderson JL (2008) J Chromatogr A 1208:1-9.

12. Armstrong DW, Zhang LK, He L, Gross ML (2001) Anal Chem 73:3679-3686.

13. Mank M, Stahl B, Boehm G (2004) Anal Chem 76:2938-2950.

14. Li YL, Gross ML (2004) J Am Soc Mass Spectrom 15:1833–1837.

15. Dickinson EV, Williams ME, Hendrickson SM, Masui H, Murray RW (1999) J Am

Chem Soc 121:613-616.

16. Doyle KP, Lang CM, Kim K, Kohl PA (2006) J Electrochem Soc 153:A1353-1357

17. Poole CF (2004) J Chromatogr A 1037:49-82.

18. Anderson JL, Armstrong DW (2003) Anal Chem 75:4851-4858.

19. Stalcup AM, Cabovska B (2004) J Liq Chromatogr Relat Technol 27:1443-1459.

20. Chan SS, Willis CJ (1968) Can J Chem 46:1237-1248.

21. Swatloski RP, Holbrey JD, Rogers RD (2003) Green Chem 5:361-363.

22. Ignat’ev NV, Welz-Biermann U, Kucheryna A, Bissky G, Willner H (2005) J Fluorine

Chem 126:1150-1159.

23. O’Mahony AM, Silvester DS, Aldous L, Hardacre C, Compton RG (2008) J Chem

Eng Data 53:2884-2891.

24. Abedin SZ, Borissenko N, Endres F (2004) Electrochem Commun 6:422-426.

25. Duffy NW, Bond AW (2006) Electrochem Commun 8:892-898.

26. Millefiorini S, Tkaczyk AH, Sedev R, Efthimiadis J, Ralston J (2006) J Am Chem

Soc 128:3098-3101.

27. Muldoon MJ, Aki SN, Anderson JL, Dixon JK, Brennecke JF (2007) J Phys Chem B

111:9001-9009.

28. Chrobok A, Swadzba M, Baj S (2007) Polish J Chem 81:337-344.

45 29. Abraham MH (1993) Chem Soc Rev 22:73-83.

30. Anderson JL, Ding J, Welton T, Armstrong DW (2002) J Am Chem Soc

124:14247-14254.

31. Mutelet F, Jaubert JN (2006) J Chromatogr A 1102: 256-267.

32. Hsieh YN, Ho WY, Horng RS, Huang PC, Hsu CY, Huang HH, Kuei CH (2007)

Chromatographia 66:607-611.

33. Sprunger LM, Proctor A, Acree WE, Abraham MH (2008) Fluid Phase Equilib

265:104-111.

34. Baltazar QQ, Leininger SK, Anderson JL (2008) J Chromatogr A 1182:119-127.

35. Bates ED, Mayton RD, Ntai I, Davis JH (2002) J Am Chem Soc 124:926-927.

36. Froehlich U (2005) Chirality and Ionic Liquids. Ph.D dissertation. Queen’s University

of Belfast.

37. Crowhurst L, Mawdsley PR, Perez-Arlandis JM, Salter PA, Welton T (2003) Phys

Chem Chem Phys 5:2790-2794.

38. Sharma NK, Tickell MD, Anderson JL, Kaar J, Pino V, Wicker BF, Armstrong DW,

Davis JH, Russell AJ (2006) Chem Commun 646-648.

39. Zhang XM, Bordwell FG, Puy MVD, Fried HE (1993) J Org Chem 58:3060-3066.

46 Chapter 3

Evaluating the Solvation Properties of Functionalized Ionic

Liquids with Varied Cation/Anion Composition Using the

Solvation Parameter Model

A paper published in Journal of Chromatography A 1

Pamela Twu, Qichao Zhao, William R. Pitner, William E. Acree Jr., Gary A. Baker,

Jared L. Anderson

Abstract

Ionic liquids (ILs) are promising gas chromatography (GC) stationary phases due to their high thermal stability, negligible vapor pressure, and ability to solvate a broad range of analytes. The tunability of ILs allows for structure modification in pursuit of enhanced separation selectivity and control of analyte elution order. In this study, the solvation par-

2011 Elsevier

47 ameter model is used to characterize the solvation interactions of fifteen ILs containing various cationic functional groups (i.e., dimethylamino, hydroxyl, and ether) and cation types paired with various counter anions, namely, tris(pentafluoroethyl)trifluorophosphate

− − − (FAP ), bis[(trifluoromethyl)sulfonyl]imide (NTf2 ), thiocyanate (SCN ),

− − tricyanomethide (C(CN)3 ), tetracyanoborate (B(CN)4 ), and bis[oxalate(2-)]borate

(BOB−). The presence of functional groups affected the hydrogen bond basicity, hydrogen bond acidity, as well as dispersion interactions of the resulting ILs, while the change of cation type yielded modest influence on the dipolarity. The switch of counter anions in unfunctionalized ILs produced compounds with higher dipolarity and hydrogen bond basicity. The dipolarity and hydrogen bond basicity of ILs possessing cyano-containing anions appeared to be inversely proportional to the cyano content of the anion. The modification of IL structure resulted in a significant effect on the retention behavior as well as separation selectivity for many solutes, including reversed elution orders of some analytes. This study provides one of the most comprehensive examinations up-to-date on the relation between IL structure and the resulting solvation characteristics and gives tremendous insight into choosing suitable ILs as GC stationary phases for solute specific separations.

3.1 Introduction

Recently, the use of ionic liquids (ILs) has attracted much academic and industrial interest due to their unique physico-chemical properties. ILs are generally defined as non-molecular ionic solvents with melting points below 100 °C [1]. This unique class of

48 solvents is usually composed of a nitrogen or phosphorus containing organic cation and various counter anions. Compared to traditional organic solvents, ILs exhibit significant advantages including nearly undetectable vapor pressures under ambient temperature, wide electrochemical windows, and high thermal stability (often higher than 300 °C). In addition, many physical and chemical properties of ILs, such as viscosity and solubility with other molecules, can be custom designed by simply varying the cation and anion combination or by introducing desired functional groups to either component [2]. One continually emerging application involving ILs is their employment as gas chromatography (GC) stationary phases [3–5].

The success of ILs as GC stationary phases is due to the advantages they possess over traditional materials such as the substituted polysiloxanes and polyethylene glycol stationary phases. For example, ILs typically exhibit extremely low volatility and high thermal stability as well as the capability of remaining in the liquid state over a wide temperature range. They can be designed to exhibit low column bleed, longer lifetimes, as well as extended operation temperature ranges when used as stationary phases. In addition, ILs are capable of undergoing multiple solvation interactions thereby imparting unique selectivities towards a wide range of molecules with different functional groups.

Moreover, the tunability of ILs allows for relatively easy chemical modification of the IL structures, resulting not only in enhanced thermal stability, but tuneable solvation properties and separation selectivity. Currently, the reported modifications include the cation and anion combination [6,7], introduction of desired functional groups to the IL

[8,9], development of dicationic and tricationic ILs [10–12], as well as polymerization

[13–16]. Therefore, IL structural modifications result in unique solvation properties that

49 have instigated the evaluation of new classes of functionalized ILs that consist of more exotic cation and anion combinations.

The solvation parameter model [17], developed by Abraham and co-workers, has been successfully employed to evaluate the solvation properties for a wide class of ILs

[6–16,18]. The solvation parameter model, shown in Eq. (1), is a linear free-energy relationship that describes the contribution of individual solvation interactions of a solvent by examining solute/solvent interactions.

log k = c + eE + sS + aA + bB + lL (1)

According to Eq. (1), log k is the solute retention factor and is determined by measuring the retention time of the analyte and dead volume of the chromatographic column. The solute descriptors (E, S, A, B, L) are probe-specific parameters that have been determined for many molecules [17]. They are defined as follows: E, the excess molar refraction calculated from the solute’s refractive index; S, the solute dipolarity/polarizability; A, the solute hydrogen bond acidity; B, the solute hydrogen bond basicity; and L, the solute gas hexadecane partition coefficient determined at 298 K. The system constants (e, s, a, b, l) are used to characterize the strength of each solvation interaction and are defined as: e, a measure of the IL to interact with π and nonbonding electrons of the solute; s, the dipolarity/polarizability of the IL; a, a measure of the IL hydrogen bond basicity; b, the hydrogen bond acidity of the IL; and l describes the IL dispersion forces. The system constants are attained through multiple linear regression analysis of the log k term and the five solute descriptors. The intercept term, c, can be used to determine and verify the phase ratio of the column.

In this study, a total of fifteen ILs containing different functionalized cations (i.e.,

50 pyridinium, imidazolium, ammonium, morpholinium, piperidinium, and pyrrolidinium) as well as various counter anions, namely tris(pentafluoroethyl)trifluorophosphate [FAP−],

− − bis[(trifluoromethyl)sulfonyl]imide [NTf2 ], thiocyanate [SCN ], tricyanomethide

− − − [C(CN)3 ], tetracyanoborate [B(CN)4 ], and bis[oxalato(2-)]borate [BOB ], were studied for the first time as gas chromatographic stationary phases. It is important to expand the

IL-based stationary phase library with ILs containing different cation functional groups, cation types, or counter anions, since a fundamental understanding of the relationship between the IL structures and the resulting solvation properties will allow for the selection of ideal ILs for specific separations. In this study, the solvation parameter model was used to investigate the change of solvation interactions with the IL stationary phase composition. The effect of the cation/anion composition on the system constants, retention factors, and selectivity for selected solute molecules was evaluated and discussed. This report contains one of the most comprehensive examinations to date of IL solvation interactions consisting of a large number of ILs with varied cation/anion composition.

3.2 Experimental

3.2.1 Materials

A total of fifteen ILs with varied cation and anion composition were employed in this study. Nine FAP-based ILs, namely N-hexylpyridinium FAP,

N-hexyl-4-(N′,N′-dimethylamino)-pyridinium FAP, N-hydroxypropylpyridinium FAP,

1-ethyl-3-methylimidazolium FAP, 1-methoxyethyl-3-methylimidazolium FAP,

51 methoxyethyl-dimethyl-ethylammonium FAP, 1-methoxyethyl-1-methylmorpholinium

FAP, 1-methoxyethyl-1-methylpiperidinium FAP, and 1-methoxypropyl-1- methylpiperidinium FAP, as well as 1-butyl-1-methylpyrrolidinium SCN,

1-butyl-1-methylpyrrolidinium C(CN)3, 1-butyl-1-methylpyrrolidinium B(CN)4 and

1-butyl-1-methylpyrrolidinium BOB, were provided by Merck KGaA (Darmstadt,

Germany). Hexyltrimethylammonium bis[(trifluoromethyl)sulfonyl]imide (NTf2) and

1-propyl-1-methylpiperidinium bis[(trifluoromethyl)sulfonyl] imide were prepared according to previous literature precedence [19]. The identity of each IL was confirmed with 1H-NMR spectroscopy and ESI mass spectrometry. These spectra are included as

Supplementary material.

Forty-six probe molecules with varied functional groups were selected for the characterization of the IL-based GC stationary phases using the solvation parameter model. Acetic acid, methyl caproate, naphthalene, and propionic acid were purchased from Supelco (Bellefonte, PA, USA). Bromoethane, butyraldehyde, and 2-nitrophenol were purchased from Acros Organics (Morris Plains, NJ, USA). 1-Butanol,

N,N-dimethylformamide, ethyl acetate, 2-propanol, and toluene were purchased from

Fisher Scientific, and p-cresol, m-xylene, o-xylene, and p-xylene were purchased from

Fluka (Steinheim, Germany). Cyclohexanol was purchased from J.T. Baker (Phillipsburg,

NJ, USA); ethylbenzene was from Eastman Kodak Company (Rochester, NY, USA); and acetophenone, aniline, benzaldehyde, benzene, benzonitrile, benzyl alcohol,

1-bromohexane, 1-bromooctane, 2-chloroaniline, 1-chlorobutane, 1-chlorohexane,

1-chlorooctane, cyclohexanone, 1,2-dichlorobenzene, 1,4-dioxane, 1-iodobutane, nitrobenzene, 1-nitropropane, 1-octanol, octylaldehyde, 1-pentanol, 2-pentanone,

52 phenetole, phenol, propionitrile, pyridine, pyrrole, and 1-decanol were purchased from

Sigma–Aldrich (St. Louis, MO, USA). All probe molecules were used as received.

Methylene chloride was purchased from Fisher Scientific. Untreated fused silica capillary tubing (0.25 mm I.D.) was obtained from Supelco.

3.2.2 Methods

All ionic liquids were placed under vacuum at 50 °C to remove any excess water.

Seven meter untreated fused silica capillary columns were coated by the static method at

40 °C. All coating solutions contained 0.45% (w/v) of the studied IL in methylene chloride. All coated columns were conditioned from 30 to 110 °C at 1 °C/min and held for 1 h, using a constant helium flow at a rate of 1.0 mL/min. Column efficiency was determined using naphthalene at 100 °C. All coated columns had efficiencies of at least

2000 plates per meter. Column efficiencies were monitored by recording the retention times of the probe molecules at three temperatures to ensure that the stationary phase did not change throughout the characterization experiment.

The forty-six probe molecules and their corresponding solute descriptors used in this study are provided as Supplementary material. All probe molecules were dissolved in methylene chloride and injected separately at three temperatures, namely 50, 80, and

110 °C. It should be noted that probe molecules that interacted weakly with the IL eluted with the solvent peak, especially at higher temperatures. Meanwhile, other probes that exhibited strong interactions with the stationary phase were retained in the column for 3 h or longer. Therefore, not all 46 probes could be subjected to multiple linear regression analysis at all temperatures examined. All separations were performed using a

53 Hewlett-Packard 5890 gas chromatograph equipped with a flame ionization detector.

Helium was used as the carrier gas with a flow rate of 1.0 mL/min. The injector and detector were held at 250 °C. The detector makeup flow of helium was maintained at 23 mL/min, the hydrogen flow at 40 mL/min, and the air flow rate at 450 mL/min. Methane was used to measure the dead volume of each column at the three different temperatures.

Multiple linear regression analysis and statistical calculations were performed using the program Analyze-it (Microsoft, USA).

3.3. Results and Discussion

System constants of the fifteen ILs were determined at three different temperatures

(50, 80, 110 °C) using the solvation parameter model. The ILs were carefully chosen to examine the effect of the cation and anion on the resulting system constants. Fig. 3.1 shows the structures of the ILs evaluated in this study as well as the numbering system used to refer to the ILs throughout this manuscript. Nine of the ILs contain various cations paired with the FAP− anion, while two of the ILs contain an ether functionalized

− cation paired with the NTf2 anion. To further examine the effect of the anion on the system constants, four ILs containing the pyrrolidinium cation paired with various anions,

− − − − namely, C(CN)3 , SCN , BOB , and B(CN)4 , are included in this study.

Table 3.1 lists the system constants of the fifteen ILs evaluated in this study. For most of the examined ILs, nearly all of the system constants decrease smoothly with increasing temperature. In the case of the e system constant, the values tend to drop for most ILs with increasing temperature but do not always appear to follow a consistent trend. All

54 1 2 3 4

5 6 7 8

9 10 11 12

13 14 15 16

Figure 3-1: Structures and numbering system of ILs examined in this study. (1) N-hexylpyridinium FAP, (2)

55 N-hexyl-4-(N′,N′-dimethylamino)pyridinium FAP, (3) Nhydroxypropylpyridinium FAP, (4) 1-ethyl-3-methylimidazolium FAP, (5)

1-methoxyethyl-3-methylimidazolium FAP, (6) methoxyethyl-dimethyl-ethylammonium FAP, (7)

1-methoxyethyl-1-methylmorpholinium FAP, (8) 1-methoxyethyl-1-methylpiperidinium FAP, (9)

1-methoxypropyl-1-methylpiperidinium FAP, (10) hexyltrimethylammonium NTf2, (11) 1-propyl-1-methylpiperidinium NTf2, (12)

1-butyl-1-methylpyrrolidinium SCN, (13) 1-butyl-1-methylpyrrolidinium C(CN)3, (14) 1-butyl-1-methylpyrrolidinium B(CN)4, (15)

1-butyl-1-methylpyrrolidinium BOB, (16) 1-butyl-1-methylpyrrolidinium FAP.

56 Table 3.1 System constants of functionalized IL-based stationary phases examined in this study.

IL IL stationary phase/temperature System constantsa No (°C) c e s a b l n R2 F

1 N-Hexylpyridinium tris(pentafluoroethyl)trifluorophosphate 50 -2.69 0.12 1.71 0.73 0.82 0.60 38 0.98 401 (0.09) (0.11) (0.09) (0.14) (0.02) 80 -2.76 0.24 1.47 0.55 0.79 0.50 39 0.99 513 (0.07) (0.09) (0.08) (0.12) (0.02) 110 -2.85 0.21 1.39 0.47 0.66 0.42 37 0.99 464 (0.06) (0.08) (0.06) (0.11) (0.02)

2 N-Hexyl-4-(N´,N´-dimethylamino)pyridinium tris(pentafluoroethyl)trifluorophosphate 50 -2.62 0.24 1.61 0.86 0.57 0.63 40 0.98 406 (0.09) (0.11) (0.09) (0.13) (0.02) 80 -2.77 0.30 1.44 0.65 0.50 0.54 41 0.99 470 (0.08) (0.09) (0.08) (0.12) (0.02) 110 -2.79 0.26 1.32 0.50 0.41 0.46 40 0.99 455 (0.06) (0.08) (0.07) (0.10) (0.02)

3 N-Hydroxypropylpyridinium tris(pentafluoroethyl)trifluorophosphate 50 -2.89 0.28 1.76 1.01 1.78 0.50 33 0.99 527 (0.08) (0.11) (0.11) (0.14) (0.02) 80 -3.06 0.23 1.71 0.84 1.44 0.43 34 0.99 591 (0.07) (0.09) (0.08) (0.11) (0.02) 110 -3.16 0.13 1.66 0.66 1.20 0.37 33 0.99 463 (0.07) (0.09) (0.08) (0.12) (0.02)

57 Table 3.1 (continued) IL IL stationary phase/temperature System constantsa No (°C) c e s a b l n R2 F

4 1-Ethyl-3-methylimidazolium tris(pentafluoroethyl)trifluorophosphate 50 -2.82 0.17 1.91 0.89 0.89 0.51 39 0.99 469 (0.08) (0.10) (0.09) (0.13) (0.02) 80 -2.85 0.15 1.74 0.70 0.72 0.42 38 0.99 547 (0.06) (0.08) (0.07) (0.10) (0.02) 110 -3.32 0.11 1.74 0.64 0.66 0.41 35 0.99 417 (0.07) (0.09) (0.07) (0.12) (0.02)

5 1-Methoxyethyl-3-methylimidazolium tris(pentafluoroethyl)trifluorophosphate 50 -2.78 0.05 1.96 1.09 0.89 0.53 43 0.98 392 (0.09) (0.11) (0.10) (0.14) (0.02) 80 -2.82 0.06 1.78 0.86 0.73 0.43 42 0.98 401 (0.07) (0.09) (0.08) (0.12) (0.02) 110 -2.87 0.06 1.61 0.68 0.65 0.37 41 0.98 383 (0.07) (0.09) (0.07) (0.11) (0.02)

6 Methoxyethyl-dimethyl-ethylammonium tris(pentafluoroethyl)trifluorophosphate 50 -2.79 0.09 1.97 0.91 0.85 0.53 38 0.99 458 (0.08) (0.10) (0.08) (0.13) (0.02) 80 -2.80 0.13 1.68 0.67 0.76 0.44 39 0.99 453 (0.07) (0.09) (0.08) (0.12) (0.02) 110 -2.91 0.10 1.62 0.62 0.64 0.37 38 0.98 419 (0.06) (0.08) (0.07) (0.11) (0.02)

58 Table 3.1 (continued) IL IL stationary phase/temperature System constantsa No (°C) c e s a b l n R2 F

7 1-Methoxyethyl-1-methylmorpholinium tris(pentafluoroethyl)trifluorophosphate 50 -2.83 0.10 2.05 1.02 0.86 0.51 40 0.98 375 (0.09) (0.11) (0.10) (0.14) (0.02) 80 -2.85 0.10 1.87 0.81 0.70 0.42 39 0.98 432 (0.07) (0.09) (0.08) (0.11) (0.02) 110 -2.93 0.10 1.71 0.68 0.63 0.35 38 0.98 408 (0.07) (0.08) (0.07) (0.10) (0.02)

8 1-Methoxyethyl-1-methylpiperidinium tris(pentafluoroethyl)trifluorophosphate 50 -2.65 0.16 1.82 0.95 0.58 0.56 40 0.98 372 (0.09) (0.11) (0.10) (0.14) (0.02) 80 -2.70 0.16 1.65 0.74 0.46 0.46 38 0.98 356 (0.08) (0.10) (0.08) (0.12) (0.02) 110 -2.81 0.17 1.52 0.58 0.43 0.40 38 0.98 416 (0.06) (0.08) (0.06) (0.10) (0.02)

9 1-Methoxypropyl-1-methylpiperidinium tris(pentafluoroethyl)trifluorophosphate 50 -2.63 0.28 1.77 1.25 0.73 0.53 38 0.98 392 (0.09) (0.11) (0.10) (0.16) (0.02) 80 -2.80 0.23 1.66 0.93 0.67 0.45 38 0.98 406 (0.08) (0.10) (0.08) (0.14) (0.02) 110 -2.79 0.25 1.46 0.74 0.59 0.37 37 0.99 558 (0.06) (0.08) (0.06) (0.10) (0.01)

59 Table 3.1 (continued) IL IL stationary phase/temperature System constantsa No (°C) c e s a b l n R2 F

10 Hexyl-trimethylammonium bis[(trifluoromethyl)sulfonyl]imide 50 -2.77 0 1.90 2.00 0.45 0.59 42 0.99 503 (0.10) (0.09) (0.12) (0.02) 80 -2.78 0 1.73 1.68 0.37 0.48 43 0.99 584 (0.08) (0.07) (0.10) (0.02) 110 -2.93 0 1.66 1.49 0.31 0.41 40 0.99 474 (0.08) (0.06) (0.10) (0.02)

11 1-Propyl-1-methylpiperidinium bis[(trifluoromethyl)sulfonyl]imide 50 -2.76 0.28 1.87 2.11 0.35 0.55 41 0.98 439 (0.09) (0.11) (0.10) (0.14) (0.02) 80 -2.85 0.30 1.77 1.86 0.26 0.45 42 0.98 411 (0.08) (0.10) (0.09) (0.13) (0.02) 110 -2.92 0.27 1.63 1.60 0.21 0.39 40 0.99 488 (0.07) (0.08) (0.07) (0.10) (0.02)

12 1-Butyl-1-methylpyrrolidinium thiocyanate 50 -3.03 0.44 2.21 4.40 0.15 0.54 34 0.99 384 (0.10) (0.12) (0.17) (0.16) (0.03) 80 -3.03 0.41 2.14 4.10 0 0.42 36 0.99 538 (0.09) (0.11) (0.11) (0.02) 110 -3.05 0.51 1.98 3.60 0.10 0.31 34 0.99 418 (0.08) (0.10) (0.11) (0.13) (0.02)

60 Table 3.1 (continued) IL IL stationary phase/temperature System constantsa No (°C) c e s a b l n R2 F

13 1-Butyl-1-methylpyrrolidinium tricyanomethide 50 -2.58 0.18 2.07 3.17 0.16 0.56 34 0.99 439 (0.08) (0.09) (0.09) (0.11) (0.02) 80 -2.87 0.28 1.99 2.81 0.24 0.47 43 0.99 656 (0.07) (0.09) (0.09) (0.12) (0.01) 110 -2.86 0.28 1.82 2.43 0.20 0.39 43 0.99 723 (0.07) (0.08) (0.07) (0.10) (0.02)

14 1-Butyl-1-methylpyrrolidinium tetracyanoborate 50 -2.48 0.11 1.97 2.25 0.29 0.57 36 0.98 385 (0.08) (0.09) (0.10) (0.12) (0.02) 80 -2.62 0.22 1.76 1.95 0.34 0.47 43 0.99 527 (0.08) (0.09) (0.08) (0.12) (0.02) 110 -2.71 0.21 1.67 1.76 0.26 0.40 41 0.99 544 (0.07) (0.07) (0.07) (0.10) (0.02)

15 1-Butyl-1-methylpyrrolidinium bis[oxalato(2-)]borate 50 -3.08 0.11 2.46 2.45 0.16 0.55 37 0.99 491 (0.09) (0.12) (0.09) (0.13) (0.02) 80 -3.06 0.16 2.16 2.04 0.18 0.45 38 0.99 582 (0.07) (0.10) (0.07) (0.11) (0.02) 110 -3.06 0.17 1.97 1.80 0.15 0.38 36 0.99 536 (0.07) (0.09) (0.06) (0.10) (0.02)

61 Table 3.1 (continued) IL IL stationary phase/temperature System constantsa No (°C) c e s a b l n R2 F

16 1-Butyl-1-methylpyrrolidinium tris(pentafluoroethyl)trifluorophosphate† 50 -2.86 0.21 1.61 0.85 0.68 0.54 35 0.99 404 (0.08) (0.11) (0.08) (0.14) (0.02) 80 -3.04 0.21 1.49 0.67 0.66 0.47 34 0.99 403 (0.08) (0.10) (0.07) (0.13) (0.02) 110 -3.22 0.18 1.44 0.56 0.58 0.42 34 0.99 436 (0.07) (0.09) (0.06) (0.12) (0.02)

n: Number of probe analytes subjected to multiple linear regression; R2: Correlation coefficient; F: Fischer coefficient. a System constants: e = non-bonding and π-electron interactions, s = dipolarity, a = hydrogen bond basicity, b = hydrogen bond acidity, l = dispersion forces.

† Data obtained from Ref. [9].

62 generated models were statistically reliable according to the acceptable standard deviation for each system constant, high correlation coefficients of the regression line, and satisfactory Fischer coefficients.

3.3.1 Effect of cation functional group on system constants

In general, the hydrogen bond acidity of unfunctionalized ILs largely depends on the nature of cation and can be regulated by the anion [6,20]. It was reported previously that unique solvation interactions, most importantly, the significantly enhanced hydrogen bond basicity, were observed when sulfone and sulfoxide functionalities were introduced into the cationic moiety of imidazolium-based ILs [21]. In another example, a study of four functionalized FAP-based ILs revealed that the hydrogen bond basicity, hydrogen bond acidity, and dipolarity of ILs can be modified by introducing alkyl, amino, ester, and hydroxyl functional groups to the ILs [9]. In this study, the system constants of five

FAP-based pyridinium and imidazolium ILs (1, 2, 3, 4, and 5 in Fig. 3.1) were determined and compared in order to further evaluate the effect of the cation functional group on the system constants of the resulting ILs.

ILs 1 and 2 contain pyridinium cations and FAP− anions, and are distinguished from each other by the presence of the dimethylamino moiety in the para position of the pyridinium cation (IL 2). The introduction of the dimethylamino group yielded a slight increase in the hydrogen bond basicity (a term) at 50 °C. This may be due to the availability of the lone electron pair from the tertiary amine group to interact with proton-donating solute molecules. Dipolar (s term) and hydrogen bond acidity (b term) interactions were higher on IL 1 while IL 2 was slightly more cohesive.

63 IL 1 contains a hexyl substituent while IL 3 possesses hydroxyl functionality appended to the pyridinium cation. A two-fold increase in hydrogen bond acidity at 50 °C was observed on IL 3 due to the enhanced proton-donating capability resulting from the hydroxyl group. This observation correlated well with a previous study which revealed that hydroxyl-functionalized ILs exhibited increased hydrogen bond acidity compared to their unfunctionalized analogues [9]. The presence of the hydroxyl moiety also yielded a slight increase in the hydrogen bond basicity but provided no significant changes to the overall dipolarity. IL 1 was more cohesive due to the longer alkyl chain substituent.

A comparison between two imidazolium-based ILs, IL 4 and IL 5, revealed that the presence of the ether functionality resulted in a slight increase in the hydrogen bond basicity at 50 °C. This can be explained by the enhanced interaction of the lone pair of electrons from the ether group with proton donating solutes, which results in increased retention of acidic solute molecules including acids and alcohols. All the other system constants remained largely unchanged.

3.3.2 Effect of cation type on system constants

ILs 5, 6, 7, 8, and 9 possess FAP− anions and contain an ether functional group appended to imidazolium, ammonium, morpholinium, and piperidinium cations, respectively. A comparison between these ILs allows for an elucidation into the effect of the cation type on the system constants for the corresponding ILs. The two piperidinium-based ILs, IL 8 and IL 9, showed stronger non-bonding and π-electron interactions than the other three studied ILs at 50 °C. IL 7 exhibited the highest dipolarity while the lowest values were produced by the two piperidinium-based ILs. The hydrogen

64 bond basicities for most of the ILs did not change significantly except for IL 9 which exhibited a hydrogen bond basicity of 1.25 at 50 °C. The hydrogen bond acidities of ILs containing imidazolium, ammonium, and morpholinium cations were similar to each other and higher than those of the piperidinium-based ILs. Dispersion interactions did not appear to be significantly influenced by the cation type.

In another comparison, the difference between the system constants of ammonium and piperidinium-based ILs can be examined by comparing the system constants of ILs

− 10 and 11, which both possess the NTf2 anion. When the cation was switched from ammonium to piperidinium, IL 11 exhibited enhanced nonbonding and π-electron interactions in addition to slightly higher hydrogen bond basicity at 50 °C, while a small drop of the hydrogen bond acidity was observed. IL 10 exhibited higher dispersive interactions due to its longer alkyl chain substituent. The dipolarities of these two ILs were nearly identical.

3.3.3 Effect of anion on system constants

It is widely recognized that the dipolarity and hydrogen bond basicity of unfunctionalized ILs are largely determined by the nature of the counter anion [6]. In this study, a pyrrolidinium-based cation was paired with different counter anions, namely

− − − − SCN , C(CN)3 , B(CN)4 , and BOB , to form ILs 12, 13, 14, and 15, respectively. The system constants of the ILs were examined and compared with that of a corresponding

FAP-containing IL (IL 16 in Fig. 3.1, data obtained from [9]). ILs consisting of cyano-containing anions (i.e., ILs 12, 13, and 14) all exhibited significantly increased dipolarity and hydrogen bond basicity when compared to IL 16. Interestingly, the

65 dipolarity and hydrogen bond basicity appears to be correlated with the number of cyano-moieties contained in the anion. For example, compared to IL 16, the hydrogen

− bond basicity was 2.6-fold higher on IL 14 (containing the B(CN)4 anion), 4.4- fold

− higher on IL 13 (paired with the C(CN)3 anion), and 5.2-fold higher on IL 12 (possessing the SCN− anion) at 50 °C. Correspondingly, IL 16 exhibited higher hydrogen bond acidity than the three comparative ILs. The presence of the SCN− anion resulted in the strongest non-bonding and π-electron interactions at 50 °C, followed by ILs 16 and 13. No significant differences in cohesive forces were observed when comparing these three cyano-containing ILs to each other or to IL 16.

− − ILs 14 and 15 contain the B(CN)4 and BOB anions, respectively. Compared to IL

16, ILs 14 and 15 yielded increased dipolar and hydrogen bond basicity interactions. For example, a 1.5-fold increase in dipolarity and a 2.9-fold increase in hydrogen bond basicity were observed for IL 15 relative to IL 16. Non-bonding and π-electron interactions and hydrogen bond acidity were higher on IL 16 compared to the borate-containing ILs. Dispersion interactions were largely unchanged.

3.3.4 Effect of IL stationary phase composition on solute molecule retention behavior

Structural tunability is an interesting and unique property exhibited by IL-based stationary phases which allows for the modulation of solute retention. Table 3.2 shows the retention factors for selected solute molecules on three FAP-based IL stationary phases containing pyridinium cations at 80 °C. A wide variety of analytes, including alcohols, aromatics, and haloalkanes, exhibited increased retention on IL 2. This is presumably due

Table 3.2 Comparison of retention factors for selected solute molecules on stationary phases composed of ILs containing functionalized pyridinium cations and FAPˉ anion at

80 °C.

Cation functionality Alkyl Tertiary amine Hydroxyl

Probe molecule IL 1 IL 2 IL 3

Benzyl alcohol 27.6 32.7 36.9 p-Cresol 26.6 41.1 29.0

1-Decanol 22.3 31.6 14.5

Benzonitrile 17.8 22.2 14.3

Benzaldehyde 12.7 15.8 10.6

Ethyl phenyl ether 6.7 10.0 3.7

2-Chloroaniline 41.0 66.8 28.1

Octylaldehyde 6.3 7.4 3.6

1-Bromooctane 3.0 5.0 1.3

Pyridine 2.3 3.1 4.6

Toluene 0.9 1.3 0.4

Ethyl benzene 1.5 2.1 0.7 o-Xylene 2.2 3.2 1.1

Naphthalene 48.6 89.5 23.4

67 to the presence of the lone pair electrons from the tertiary amine group as well as the enhanced cohesive interactions of IL 2. Due to the presence of the hydroxyl group, analytes such as benzyl alcohol and pyridine retained longer on IL 3, which can be rationalized by the enhanced hydrogen bond acidity and basicity of IL 3. Table 3.3 compares the retention factors of selected solutes on ILs 4 and 5 which contain imidazolium cations with the FAP− anion. A comparison of retention factors for solute molecules on ILs 4 and 5 revealed that the switch of the alkyl substituent with an ether functionality produced higher retention of proton donating solute molecules. For example, compared to IL 4, the retention factor of propionic acid was enhanced 60% when the separation was performed on IL 5 stationary phase. This is due to the slightly higher hydrogen bond basicity of IL 5 resulting from the presence of the lone pair of electrons from the ether group.

Compared to the cation functional group, the cation type was found to have only a moderate effect on the retention behavior of many solute molecules. As shown in Table

3.4, different analytes, including acids, alcohols, aldehydes, and substituted aromatic compounds, exhibited varied retention factors when subjected to separation on ILs containing different types of cations. Proton donating solutes such as p-cresol and phenol retained longer on IL 9, due to its stronger hydrogen bond basicity. Basic solute molecules, such as N,N-DMF, exhibited higher retention factors on ILs 5, 6, and 7, due to the fact that the imidazolium, ammonium, and morpholinium-based ILs possess higher hydrogen bond acidity than the ILs containing the piperidinium cation. The retention factors for selected solute molecules on ILs 10 and 11 are shown as Supplementary data in Table S-2. Similarly, it was observed that basic solutes exhibited higher retention on IL

Table 3.3 Comparison of retention factors for selected solute molecules on stationary phases composed of ILs containing functionalized imidazolium cations and FAPˉ anion at

80 °C.

Cation functionality Alkyl Ether

Probe molecule IL 4 IL 5

1-Bromooctane 1.2 1.6

Ethyl phenyl ether 3.5 4.3

Cyclohexanol 1.7 2.3

1-Decanol 8.7 12.7

Benzyl alcohol 18.7 24.2

Propionic acid 1.0 1.6

Octylaldehyde 3.0 4.0

Methyl caproate 1.7 2.3

N,N-DMF 19.6 24.5

Pyridine 1.5 2.0

2-Chloroaniline 21.1 26.5

Benzonitrile 10.9 13.8

Table 3.4 Comparison of retention factors for selected solute molecules on stationary

phases composed of ILs containing various cation types with FAPˉ anion at 80 °C.

Cation Imidazolium Ammonium Morpholinium Piperidinium Piperidinium

Probe molecule IL 5 IL 6 IL 7 IL 8 IL 9

1-Bromohexane 0.5 0.5 0.4 0.7 0.7

Propionic acid 1.6 1.3 1.6 1.4 2.2 p-Cresol 22.9 18.9 19.1 26.3 32.5

Phenol 13.0 10.8 11.0 14.5 18.8

Benzyl alcohol 24.2 21.6 24.8 26.3 27.7

1-Octanol 4.4 4.2 3.5 4.9 5.2

Octylaldehyde 4.1 3.9 3.4 4.6 4.3 p-Xylene 0.9 1.0 0.8 1.4 1.3

Ethyl phenyl ether 4.3 4.1 3.6 5.7 5.3

Benzonitrile 13.8 13.0 12.9 15.0 13.6

2-Chloroaniline 24.5 24.7 26.8 36.1 35.0

N,N-DMF 24.5 22.9 26.9 19.4 21.8

70 10 due to its slightly enhanced hydrogen bond acidity, while IL 11 exhibited longer retention for acidic molecules due to its moderately improved hydrogen bond basicity.

The exception occurs for aniline, which exhibited increased retention on IL 11, presumably due to the significantly enhanced non-bonding and π-electron interactions of

IL 11.

The solvation properties of ILs possessing cyano- and borate-containing anions make them exceptionally interesting stationary phases for GC separations. Table 3.5 shows the comparison of retention factors for selected solute molecules on ILs 12, 13, 14, and 15 at

80 °C. For comparison purposes, the retention factors for the same analytes on IL 16 are also included. The higher dipolarity and hydrogen bond basicity of ILs 12, 13, 14, and 15 resulted in greater retention of proton-donating molecules (i.e., alcohols and acids). For instance, the retention factor of propionic acid increased from 4.3 on IL 15 to 98.8 on IL

12. The retention behavior of other molecules was also greatly affected by the counter anion. The retention factor of octylaldehyde experienced a 150% increase on IL 14 and a

92% increase on IL 13 but a 37% decrease on IL 12 when compared to IL 15. Compared to IL 16, a wide variety of solute molecules including acids, alcohols, ketones, and substituted aromatics experienced changes in retention behavior when the separation was performed using ILs possessing cyano- and borate-containing anions due to the significantly enhanced dipolarity and hydrogen bond basicity. For example, the retention factor of 2-chloroaniline increased from 14.8 on IL 16 to 246.4 on IL 12.

3.3.5 Effect of IL stationary phase composition on solute pair separation selectivity

In this study, the separation selectivity of IL-based stationary phases was determined

Table 3.5 Comparison of retention factors for selected solute molecules on stationary phases composed of 1-butyl-1-methylpyrrolidinium cation and various counter anions at

80 °C.

Anion [ SCNˉ] [C(CN)3ˉ] [B(CN)4ˉ] [BOBˉ] [FAPˉ]

Probe molecule IL 12 IL 13 IL 14 IL 15 IL 16

1-Chlorooctane 0.9 1.9 2.4 1.1 0.7

Propionic acid 98.8 23.4 9.1 4.3 0.6

1-Butanol 3.2 2.3 1.7 0.7 0.2

Cyclohexanol 14.5 11.8 8.5 3.3 N/A

2-Pentanone 0.5 1.2 1.4 0.6 0.6

Cyclohexanone 4.2 8.7 10.2 4.2 3.0

Methyl caproate 0.8 2.1 3.0 1.1 1.2

Octylaldehyde 1.9 5.0 6.5 2.6 2.3

Ethyl phenyl ether 4.2 7.7 8.4 3.7 2.6

2-Chloroaniline 246.4 194.7 134.4 80.6 14.8

Nitrobenzene 36.4 58.9 54.3 33.8 10.7

Ethyl benzene 0.8 1.7 1.9 0.7 0.6 o-Xylene 1.3 2.4 2.9 1.1 0.9

Naphthalene 41.7 60.4 60.7 26.9 16.2

72 using the ratio of two solute retention factors. Table 3.6 lists the selectivity for selected molecule pairs on stationary phases composed of FAP-based ILs containing the pyridinium cation at 80 °C, while Table 3.7 lists the selectivity for selected molecule pairs on stationary phases composed of FAP-based ILs containing the imidazolium cation at

80 °C. It was observed that the introduction of a functional group to the IL utilized as stationary phase varied the selectivities of some solutes. For example, the selectivity between 1-decanol and 1-butanol increased from 14.9 on IL 1 to 20.8 on IL 2 due to the presence of the tertiary amine group while it dropped to 9.5 on IL 3 because of the introduction of the hydroxyl functionality. A reversed elution order for some solute pairs can be obtained by simply switching the alkyl chain with an ether moiety. For example, acetic acid eluted quicker than ethyl benzene on IL 4, while it retained longer than ethyl benzene on IL 5.

Table 3.8 summarizes the selectivities for selected solute pairs on five FAP-based ILs containing different cation types. The piperidinium-based ILs exhibited enhanced separation selectivities for solute pairs such as naphthalene and benzonitrile, naphthalene and butyraldehyde, as well as phenol and cyclohexanone. For certain solute pairs, a reversal of elution order was observed. For example, the selectivity for N,N-DMF and p-cresol was 1.1, 1.2, and 1.4 on ILs 5, 6, and 7, respectively. However, when the same separation was performed on ILs 8 and 9, p-cresol eluted quicker than N,N-DMF. The comparison of selectivity for selected solute molecules on two NTf2-based ILs, ILs 10 and 11, is included as Supplementary data in Table S-3. Separation selectivities for solute pairs, for example, N,N-DMF and o-xylene, methyl caproate and benzene, as well as naphthalene and nitrobenzene, can be finely tuned by the switch of cation type. An

Table 3.6 Comparison of selectivity for selected solute molecules on stationary phases composed of ILs containing functionalized pyridinium cations and FAPˉ anion at 80 °C.

Cation functionality Alkyl Tertiary amine Hydroxyl

Solute Pair IL 1 IL 2 IL 3

1-Octanol/1-bromooctane 1.9 1.6 2.7

Methyl caproate/o-xylene 1.3 1.1 1.6

Cyclohexanol/1-butanol 2.6 2.4 2.7

1-Decanol/1-butanol 14.9 20.8 9.5

Naphthalene/nitrobenzene 2.4 1.7 1.2

Naphthalene/acetophenone 1.4 2.1 0.73a

Naphthalene/benzaldehyde 1.5 1.4 2.1

Naphthalene/o-xylene 15.3 21.7 11.7

Benzonitrile/ethyl benzene 7.6 7.6 9.0 a: By definition, the value of the separation factor should be greater than unity. However, some analytes exhibited reversal of elution order making it impossible to report integers greater than unity for all stationary phase compositions.

Table 3.7 Comparison of selectivity for selected solute molecules on stationary phases composed of ILs containing functionalized imidazolium cations and FAPˉ anion at 80 °C.

Cation functionality Alkyl Ether

Solute Pair IL 4 IL 5

Acetic acid/ethyl benzene 0.95a 1.1

Propionic acid/benzene 1.6 2.0

Naphthalene/o-xylene 10.4 11.2

Acetophenone/1-decanol 2.3 2.0

N,N-DMF/o-xylene 9.8 11.2

1-Octanol/1-bromooctane 1.9 2.1

1-Octanol/nitropropane 1.3 1.5

Ethyl phenyl ether/ethyl benzene 2.6 2.8

Benzonitrile/ethyl benzene 6.8 7.9

1-Decanol/2-pentanone 4.9 6.3

Cyclohexanol/benzene 2.1 2.5

Cyclohexanol/1-butanol 2.0 2.2

1-Decanol/1-butanol 7.0 9.4 a: By definition, the value of the separation factor should be greater than unity. However, some analytes exhibited reversal of elution order making it impossible to report integers greater than unity for all stationary phase compositions.

Table 3.8 Comparison of selectivity for selected solute molecules on stationary phases

composed of ILs containing various cation types and FAPˉ anion at 80 °C.

Cation Imidazolium Ammonium Morpholinium Piperidinium Piperidinium

Solute Pair IL 5 IL 6 IL 7 IL 8 IL 9

Acetic acid/benzene 1.5 1.4 1.6 1.3 1.7

N,N-DMF/o-xylene 11.2 10.1 12.9 7.1 8.3

N,N-DMF/p-cresol 1.1 1.2 1.4 0.75a 0.68a

Naphthalene/phenol 1.8 2.3 2.1 2.5 2.0

Naphthalene/benzonitrile 1.7 2.0 1.8 2.4 2.7

Naphthalene/butyraldehyde 17.7 19.3 17.8 27.0 29.4

Phenol/cyclohexanone 1.9 1.6 1.6 2.1 3.3

Acetophenone/1-decanol 2.0 1.9 2.4 1.8 1.6

1-Decanol/1-butanol 9.4 10.2 7.8 11.3 12.2

a: By definition, the value of the separation factor should be greater than unity. However,

some analytes exhibited reversal of elution order making it impossible to report integers

greater than unity for all stationary phase compositions.

76 example of reversed elution order can be observed for acetophenone and 1-decanol.

Table 3.9 shows the selectivity for selected solute molecules on stationary phases composed of the pyrrolidinium cation with various counter anions. ILs possessing the cyano- and borate-containing anions exhibited higher selectivity between proton-donating molecules (i.e., alcohols and acids) and aromatic solutes because of their enhanced dipolarity and hydrogen bond basicity. A comparison of separation selectivity between

1-octanol and 1-bromooctane revealed that the selectivity increased from 1.5 on IL 16 to

2.2 on IL 15, 3.3 on IL 14, 4.8 on IL 13, and further to 8.0 on IL 12. Several reversals of elution order were observed including the separation between acetic acid and ethyl benzene, propionic acid and o-xylene, benzonitrile and acetic acid, benzaldehyde and pyrrole, 2-chloroaniline and naphthalene, benzyl alcohol and naphthalene, as well as naphthalene and 1-decanol. These observations clearly demonstrate that a desirable selectivity can be achieved by choosing a specific cation and anion combination in the make-up and composition of the IL stationary phase.

3.4 Conclusions

ILs have been shown to be very useful classes of GC stationary phases due to their unique and interesting properties. The chemical tunability of ILs is one of their paramount properties that can be exploited in the development of IL-based GC stationary phases. The presence of functional groups, namely dimethylamino, hydroxyl, and ether appended to the cationic moiety, produced varied system constants dependent on the nature of the employed functional group. The cation type also yielded a moderate effect

Table 3.9 Comparison of selectivity for selected solute molecules on stationary phases composed of 1-butyl-1-methylpyrrolidinium cation and various counter anions at 80 °C.

Anion [ SCNˉ] [C(CN)3ˉ] [B(CN)4ˉ] [BOBˉ] [FAPˉ]

Solute Pair IL 12 IL 13 IL 14 IL 15 IL 16

Acetic acid/ethyl benzene 40.0 5.9 2.2 2.4 0.84a

Propionic acid/o-xylene 43.2 7.1 2.6 2.5 0.86a

2-Chloroaniline/acetic acid 3.4 12.4 21.3 20.1 11.7

Benzonitrile/acetic acid 0.31a 2.0 5.8 4.8 5.8

Benzaldehyde/pyrrole 0.56a 1.1 1.4 1.3 1.7

2-Chloroaniline/naphthalene 5.8 3.2 2.2 2.9 0.92a

Benzyl alcohol/naphthalene N/A 4.0 2.3 2.2 0.66a

Naphthalene/1-decanol 0.79a 1.0 1.2 1.6 2.2

Nitrobenzene/cyclohexanol 2.6 5.1 5.8 10.5 5.8

1-Octanol/1-bromooctane 8.0 4.8 3.3 2.2 1.5

1-Decanol/1-butanol 12.9 18.5 18.8 10.2 6.5 a: By definition, the value of the separation factor should be greater than unity. However, some analytes exhibited reversal of elution order making it impossible to report integers greater than unity for all stationary phase compositions.

78 on the overall solvation characteristics of the ILs. Four different cyano- and borate-containing anions were paired with a pyrrolidinium cation and the obtained system constants were compared with that of their FAP-based analogue. It was observed that the switch of counter anion produced ILs with enhanced dipolarity and hydrogen bond basicity. The 1-butyl-1-methylpyrrolidinium thiocyanate exhibited the highest hydrogen bond basicity, while the strongest dipolarity was demonstrated by the

1-butyl-1-methylpyrrolidinium bis[oxalato(2-)]borate IL. Correspondingly, all four pyrrolidinium-based ILs exhibited lower hydrogen bond acidity when compared to the analogous FAP-based IL. The dipolarity and hydrogen bond basicity of ILs consisting of cyano-containing anions were inversely proportional to the number of cyano moieties within the anion. The modulation of cation and anion combination allowed for control of solute retention as well as separation selectivity. Reversals of elution order were achieved for many solute pairs by simple IL structure modification of the employed stationary phase. This study demonstrates the utility of tailoring ILs using cation appended

− − − − − functional groups and/or [FAP ], [SCN ], [BOB ], [C(CN)3 ], [B(CN)4 ] anions to produce stationary phases that exhibit unique selectivity often not easily observed with commercially available GC stationary phases.

Acknowledgements

79 National Science Foundation for a CAREER grant (CHE-0748612)

References

1. T. Welton, Chem. Rev. 99 (1999) 2071.

2. J.H. Davis Jr., Chem. Lett. 9 (2004) 1072.

3. D.W. Armstrong, L. He, Y. Liu, Anal. Chem. 71 (1999) 3873.

4. C.F. Poole, S.K. Poole, J. Sep. Sci. 34 (2011) 888.

5. C.F. Poole, K.G. Furton, B.R. Kersten, J. Chromatogr. Sci. 24 (1986) 400.

6. J.L. Anderson, J. Ding, T. Welton, D.W. Armstrong, J. Am. Chem. Soc. 124 (2002)

14247.

7. Q.Q. Baltazar, S.K. Leininger, J.L. Anderson, J. Chromatogr. A 1182 (2008) 119.

8. J.L. Anderson, D.W. Armstrong, Anal. Chem. 75 (2003) 4851.

9. Q. Zhao, J. Eichhorn, W.R. Pitner, J.L. Anderson, Anal. Bioanal. Chem. 395 (2009)

225.

10. J.L. Anderson, R. Ding, A. Ellern, D.W. Armstrong, J. Am. Chem. Soc. 127 (2005)

593.

11. K. Huang, X. Han, X. Zhang, D.W. Armstrong, Anal. Bioanal. Chem. 389 (2007)

2265.

12. T. Payagala, Y. Zhang, E. Wanigasekara, K. Huang, Z.S. Breitbach, P.S. Sharma, L.M.

Sidisky, D.W. Armstrong, Anal. Chem. 81 (2009) 160.

13. J.L. Anderson, D.W. Armstrong, Anal. Chem. 77 (2005) 6453.

80 14. Y. -N. Hsieh, W.-Y. Ho, R.S. Horng, P.-C. Huang, C.-Y. Hsu, H.-H. Huang, C.-H. Kuei,

Chromatographia 66 (2007) 607.

15. Q. Zhao, J.L. Anderson, J. Sep. Sci. 33 (2010) 79.

16. X. Sun, Y. Zhu, P. Wang, J. Li, C. Wu, J. Xing, J. Chromatogr. A 1218 (2011) 833.

17. M.H. Abraham, Chem. Soc. Rev. 22 (1993) 73.

18. S.K. Poole, C.F. Poole, Analyst 120 (1995) 289.

19. L.M. Grubbs, S. Ye, M. Saifullah, W.E. Acree, Jr., P. Twu, J.L. Anderson, G.A. Baker,

M.H. Abraham, J. Solution Chem., in press.

20. L. Crowhurst, P.R. Mawdsley, J.M. Perez-Arlandis, P.A. Salter, T. Welton, Phys.

Chem. Chem. Phys. 5 (2003) 2790.

21. N.K. Sharma, M.D. Tickell, J.L. Anderson, J. Kaar, V. Pino, B.F. Wicker, D.W. Armstrong,

J.H. Davis, A.J. Russell, Chem. Commun. (2006) 646.

81 Chapter 4

Highly Selective GC Stationary Phases Consisting of Binary

Mixtures of Polymeric Ionic Liquids

A paper published in Journal of Separation Science 1

Qichao Zhao, Jared L. Anderson

Abstract

GC stationary phases composed of binary mixtures of two polymeric ionic liquids

(PILs), namely, poly(1-vinyl-3-hexylimidazolium) bis[(trifluoromethyl)sulfonyl]imide poly(ViHIm-NTf2))/poly(1-vinyl-3-hexylimidazolium) chloride (poly(ViHIm-Cl)) and poly(1-vinyl-3-hexadecylimidazolium) bis[(trifluoromethyl)sulfonyl]imide

(poly(ViHDIm-NTf2))/ poly(1-vinyl-3-hexadecylimidazolium) chloride

(poly(ViHDIm-Cl)), were evaluated in terms of their on-set bleed temperature and separation selectivity. A total of six neat or binary PIL stationary phases were characteriz-

WILEY-VCH Verlag GmbH & Co.

82 ed using the solvation parameter model to investigate the effects of the polymeric cation and anion and PIL composition on the system constants of the resulting stationary phases.

The hydrogen bond basicity of the mixed poly(ViHIm-NTf2)/poly(ViHIm-Cl) stationary phases was enriched linearly with the increase in the poly(ViHIm-Cl) content. Results revealed that tuning the composition of the stationary phase allowed for fine control of the retention factors and separation selectivity for alcohols and carboxylic acids as well as selected ketones, aldehydes, and aromatic compounds. A reversal of elution order was observed for particular classes of analytes when the weight percentage of the chloride-based PIL was increased.

4.1 Introduction

Throughout decades, numerous efforts have been devoted toward the development of stationary phases for GC that are capable of providing high selectivity, thermal stability, resolution, and separation efficiency. Ionic liquids (ILs) have recently attracted much attention in the field of separation science. ILs are defined as organic molten salts that have melting points below 100 °C [1]. They possess many interesting and unique properties, such as nearly undetectable vapor pressures under ambient temperature, wide electrochemical windows, and impressive thermal stability. Because of the fact that the different combinations of cation and anion can influence various physical and chemical properties of the resulting ILs, the structures of ILs can be carefully tailored to provide the desired viscosity, solubility, or thermal stability range. Previously reported applications of ILs include organic synthesis [1–5], liquid–liquid extractions [6–8],

83 analytical microextractions [9–13], and mass spectrometry [14–16].

Mixed-mode stationary phases have been widely used in GC due to the unique selectivities that often cannot be observed when using neat stationary phases [17–20].

Acree and coworkers reported the application of binary liquid crystalline solvent mixtures, namely butyl p-(p-ethoxyphenoxycarbonyl) phenyl carbonate and p-(p-ethoxyphenylazo)phenyl undecylenate, as GC-LC stationary phases [17]. A series of alkanes were separated using stationary phases with different compositions (0% butyl p-(p-ethoxyphenoxycarbonyl)phenyl carbonate/100% p-(p-ethoxyphenylazo)phenyl undecylenate, 20/80%, 40/60%, 60/40%, 80/20%, and 100/0%, w/w). The same group then reported the preparation of a mixed-mode GC stationary phase containing hexadecyltrimethylammonium bromide and formamide lyotropic liquid crystals

(50%/50%, w/w), and the separation of selected probes on this stationary phase was performed [18]. Mixed-mode GC stationary phases containing varied ratios of poly(dimethylsiloxane)/poly(ethylene glycol) were characterized through the solvation parameter model by Poole et al., and the relationship between the system constants and temperature and composition were quantitatively explained [19]. More recently, our research group reported that stationary phases composed of a mixture of two different ILs, namely, 1-butyl-3-methylimidazolium chloride (BMIM-Cl) and

1-butyl-3-methylimidazolium bis[(trifluoromethyl)sulfonyl] imide (BMIM-NTf2), can provide enhanced dipole-type and hydrogen bond basicity interactions and enhanced selectivity for proton-donor analytes [20]. However, the operating temperature of stationary phases containing imidazolium-based monomeric ILs is usually lower than

200 °C [21]. One method to increase the thermal stability of ILs is to produce polymeric

84 ionic liquid (PIL)-based materials [22–24]. Previous publications reveal that PIL stationary phases possess enhanced thermal stability allowing for a wider operating temperature range, while largely preserving their dual nature separation selectivity. Our group has also utilized PILs in the development of thermally stable absorbent coatings in solid-phase microextraction (SPME) [12, 13].

The solvation parameter model [25], developed by Abraham, has been extensively used to explore the solvation properties of a wide class of ILs [26–30]. This model is described by the following equation

log k = c + eE + sS + aA + bB + lL (1)

According to Eq. (1), log k is the solute retention factor and is calculated by measuring the retention time of the analyte and dead volume of the chromatographic column. The solute descriptors (E, S, A, B, and L) are probe-specific parameters that have been determined for many molecules [25]. They are defined as: E, the excess molar refraction calculated from the solute’s refractive index; S, the solute dipolarity/polarizability; A, the solute hydrogen bond acidity; B, the solute hydrogen bond basicity; and L, the solute-gas hexadecane partition coefficient determined at 298 K. The system constants (e, s, a, b, and l) are used to characterize the strength of each solvation interaction and are defined as: e, the ability of the IL to interact with π- and nonbonding electrons of the solute; s, a measure of the dipolarity/polarizability of the IL; a, the IL hydrogen bond basicity; b, a measure of the hydrogen bond acidity of the IL; and l describes IL dispersion forces. The system constants are attained through multiple linear regression analysis of the log k term and the five solute descriptors. The intercept term, c, can be used to determine and verify the phase ratio of the column.

85 In this study, binary mixtures of two different PILs, namely, poly(1-vinyl-3-hexylimidazolium) bis[(trifluoromethyl)sulfonyl]imide

[poly(ViHIm-NTf2)]/poly(1-vinyl-3-hexylimidazolium) chloride [poly(ViHIm-Cl)] and poly(1-vinyl-3-hexadecylimidazolium) bis[(trifluoromethyl)sulfonyl]imide

[poly(ViHDIm-NTf2)]/poly(1-vinyl-3-hexadecylimidazolium) chloride

[poly(ViHDIm-Cl)], were used as GC stationary phases. A total of six stationary phases, including two neat stationary phases [poly(ViHIm-NTf2) and poly(ViHDIm-NTf2)], three mixed poly(ViHIm-NTf2)/poly(ViHIm-Cl)-based stationary phases (75% poly(ViHIm-NTf2)/25% poly(ViHIm-Cl), 50%/50%, and 25%/75%, w/w), and one mixed poly(ViHDIm-NTf2)/poly(ViHDIm-Cl)-based stationary phase (50% poly

(ViHDIm-NTf2)/50% poly(ViHDIm-Cl) were characterized using the solvation parameter model. The on-set bleed temperatures of the resulting PIL stationary phases were examined. The effect of the overall cation/anion composition on the retention factors and separation selectivity for selected solute molecules was studied and is discussed.

4.2 Experimental

4.2.1 Materials

The synthesis of all IL monomers and polymers involved the use of the following reagents: vinyl imidazole, 2,2′-azo-bis(isobutyronitrile), 1-chlorohexane, and

1-chlorohexadecane which were purchased from Sigma-Aldrich (St. Louis, MO, USA), lithium bis[(trifluoromethyl)sulfonyl]imide which was obtained from SynQuest Labs

(Alachua, FL, USA), and ethyl acetate and 2-propanol which were purchased from Fisher

86 Scientific (Fairlawn, NJ, USA). A total of 46 probe molecules with varied functional groups were selected for the characterization of the PIL GC stationary phases using the solvation parameter model. Acetic acid, methyl caproate, naphthalene, and propionic acid were purchased from Supelco (Bellefonte, PA, USA). Bromoethane, butyraldehyde, and

2-nitrophenol were purchased from Acros Organics (Morris Plains, NJ, USA). 1-Butanol,

N,N-dimethylformamide, ethyl acetate, 2-propanol, and toluene were purchased from

Fisher Scientific, and p-cresol, m-xylene, o-xylene, and p-xylene were purchased from

Fluka (Steinheim, Germany). Cyclohexanol was purchased from J.T. Baker (Phillipsburg,

NJ, USA); ethylbenzene from Eastman Kodak (Rochester, NY, USA); and acetophenone, aniline, benzaldehyde, benzene, benzonitrile, benzyl alcohol, 1-bromohexane,

1-bromooctane, 2-chloroaniline, 1-chlorobutane, 1-chlorohexane, 1-chlorooctane, cyclohexanone, 1,2-dichlorobenzene, 1,4-dioxane, 1-iodobutane, nitrobenzene,

1-nitropropane, 1-octanol, octylaldehyde, 1-pentanol, 2-pentanone, phenetole, phenol, propionitrile, pyridine, pyrrole, and 1-decanol were purchased from Sigma-Aldrich. All probe molecules were used as received. Methylene chloride was purchased from Fisher

Scientific. Untreated fused silica capillary tubing (0.25mmid) was obtained from Supelco.

4.2.2 Methods

The IL monomers, namely, 1-vinyl-3-hexadecylimidazolium chloride and

1-vinyl-3-hexylimidazolium chloride, and their corresponding polymers were synthesized following previously reported procedures [31]. Briefly, 1-vinylimidazole was dissolved in

2-propanol and reacted with a 10% excess of the corresponding alkyl chloride under reflux and constant stirring conditions followed by extraction with ethyl acetate. After

87 purification, polymerization was performed in the presence of the free radical initiator

2,2′-azo-bis(isobutyronitrile) to generate poly(ViHDIm-Cl) and poly(ViHIm-Cl). The polymerization step was repeated, if necessary, until the peaks belonging to the vinyl group in 1H-NMR disappeared. To perform metathesis anion exchange, PILs containing the chloride anion were dissolved in Milli-Q water and an equimolar amount of LiNTf2 was introduced to the aqueous solution. The resulting precipitate was collected, washed with water to remove any residual halide anion, and dried overnight under vacuum to yield poly(ViHDIm-NTf2) and poly(ViHIm-NTf2). The structures of the PILs evaluated in this study are shown in Fig. 4.1. All IL monomers and polymers were characterized using

1H-NMR and are provided as Supporting Information.

Five-meter untreated fused silica capillary columns were coated at 40 °C using the static method. All coating solutions contained 0.25% (w/v) of the studied neat or mixed

PIL stationary phases at the desired weight percentage of each PIL component in methylene chloride. All coated columns were conditioned from 30 to 110 °C at 1 °C/min and held for 1 h, using a constant helium flow at a rate of 1.0 mL/min. Column efficiency was tested with naphthalene at 100 °C. All coated columns had efficiencies of at least

2000 plates/m, except for the poly(ViHIm-Cl) (75%)/poly(ViHIm-NTf2) (25%) column, which had an efficiency of 1000 plates/m. Column efficiency was monitored by recording the retention times of the probe molecules at three temperatures to ensure that the column coating did not change throughout the characterization of the stationary phase.

Measurements of the onset bleed temperature were performed by ramping the oven temperature from 100 to 400 °C at 3 °C/min. The rising baseline in the chromatogram

(i.e., column bleed) indicated decomposition or volatilization of the PIL.

Figure 4-1: Structures of the PILs used as stationary-phase components in this study.

89 Table 4.1 lists the 46 probe molecules and their respective solute descriptors that were used in this study. All probe molecules were dissolved in methylene chloride and injected separately at three temperatures (50, 80, 110 °C). It should be noted that not all

46 probes could be subjected to multiple linear regression analysis at all examined temperatures due to the fact that some probes with very low boiling points and weak interactions with the IL eluted with the solvent peak, particularly at higher temperature.

Other probes exhibited very strong interactions with the stationary phase and were retained in the column for 3 h or longer. All separations were performed using a

Hewlett–Packard 6890 GC equipped with a flame ionization detector. Helium was used as the carrier gas with a flow rate of 1.0 mL/min. The injector and detector were held at

250 °C. The detector makeup flow of helium was maintained at 23 mL/min, the hydrogen flow at 40 mL/min, and the air flow rate at 450 mL/min. Methane was used to measure the dead volume of each column at the three different temperatures. Multiple linear regression analysis and statistical calculations were performed using the program

Analyze-it (Microsoft, USA).

4.3 Results and Discussion

4.3.1 Measurement of column bleed temperature

The thermal properties of GC stationary phases largely govern the overall lifetime and the maximum operating temperature of the column. In this study, the thermal properties of the PIL stationary phases were investigated by measuring the onset bleed temperature for each composition, as summarized in Table 4.2. The length of the alkyl

Table 4.1 List of probe molecules and their corresponding solute descriptors employed in the study of the IL stationary phases using the solvation parameter model.

Probe molecule E S A B L Acetic acid 0.265 0.65 0.61 0.44 1.750 Acetophenone 0.818 1.01 0.00 0.48 4.501 Aniline 0.955 0.96 0.26 0.41 3.934 Benzaldehyde 0.820 1.00 0.00 0.39 4.008 Benzene 0.610 0.52 0.00 0.14 2.786 Benzonitrile 0.742 1.11 0.00 0.33 4.039 Benzyl alcohol 0.803 0.87 0.33 0.56 4.221 Bromoethane 0.366 0.40 0.00 0.12 2.620 1-Bromooctane 0.339 0.40 0.00 0.12 5.090 1-Butanol 0.224 0.42 0.37 0.48 2.601 Butyraldehyde 0.187 0.65 0.00 0.45 2.270 2-Chloroaniline 1.033 0.92 0.25 0.31 4.674 1-Chlorobutane 0.210 0.40 0.00 0.10 2.722 1-Chlorohexane 0.201 0.40 0.00 0.10 3.777 1-Chlorooctane 0.191 0.40 0.00 0.10 4.772 p-Cresol 0.820 0.87 0.57 0.31 4.312 Cyclohexanol 0.460 0.54 0.32 0.57 3.758 Cyclohexanone 0.403 0.86 0.00 0.56 3.792 1,2-Dichlorobenzene 0.872 0.78 0.00 0.04 4.518 N,N-Dimethylformamide 0.367 1.31 0.00 0.74 3.173 1,4-Dioxane 0.329 0.75 0.00 0.64 2.892 Ethyl Acetate 0.106 0.62 0.00 0.45 2.314 Ethyl benzene 0.613 0.51 0.00 0.15 3.778 1-Iodobutane 0.628 0.40 0.00 0.15 4.130 Methyl Caproate 0.067 0.60 0.00 0.45 3.844 Naphthalene 1.340 0.92 0.00 0.20 5.161 Nitrobenzene 0.871 1.11 0.00 0.28 4.557 1-Nitropropane 0.242 0.95 0.00 0.31 2.894 1-Octanol 0.199 0.42 0.37 0.48 4.619 Octylaldehyde 0.160 0.65 0.00 0.45 4.361 1-Pentanol 0.219 0.42 0.37 0.48 3.106 2-Pentanone 0.143 0.68 0.00 0.51 2.755 Ethyl phenyl ether 0.681 0.70 0.00 0.32 4.242 Phenol 0.805 0.89 0.60 0.30 3.766 Propionitrile 0.162 0.90 0.02 0.36 2.082

91 Table 4.1 (continued) Probe molecule E S A B L Pyridine 0.631 0.84 0.00 0.52 3.022 Pyrrole 0.613 0.73 0.41 0.29 2.865 Toluene 0.601 0.52 0.00 0.14 3.325 m-Xylene 0.623 0.52 0.00 0.16 3.839 o-Xylene 0.663 0.56 0.00 0.16 3.939 p-Xylene 0.613 0.52 0.00 0.16 3.839 2-Propanol 0.212 0.36 0.33 0.56 1.764 2-Nitrophenol 1.015 1.05 0.05 0.37 4.760 1-Bromohexane 0.349 0.40 0.00 0.12 4.130 Propionic acid 0.233 0.65 0.60 0.45 2.290 1-Decanol 0.191 0.42 0.37 0.48 5.628 Data obtained from ref [25]

Table 4.2 On-set bleed temperature of studied neat and binary PIL stationary phases.

Stationary-phase composition Onset bleed temperature (°C)

a Poly(ViHIm-NTf2) 335-345

b Poly(ViHIm-NTf2) (75%)/Poly(ViHIm-Cl) (25%) 235-250

Poly(ViHIm-NTf2) (50%)/Poly(ViHIm-Cl) (50%) 215-230

Poly(ViHIm-NTf2) (25%)/Poly(ViHIm-Cl) (75%) 190-220

c Poly(ViHDIm-NTf2) 325-340

d Poly(ViHDIm-NTf2) (50%)/ Poly(ViHDIm-Cl) (50%) 220-235 a Poly(1-vinyl-3-hexylimidazolium) bis[(trifluoromethyl)sulfonyl]imide b Poly(1-vinyl-3-hexylimidazolium) chloride c Poly(1-vinyl-3-hexadecylimidazolium) bis[(trifluoromethyl)sulfonyl]imide d Poly(1-vinyl-3-hexadecylimidazolium) chloride

93 side chain appended to the imidazolium cation appears to have little effect on the bleed temperature of the PIL. For example, both the neat poly(ViHIm-NTf2) and the neat poly(ViHDIm-NTf2) PIL stationary phases did not exhibit column bleed until 325–335 °C.

On the contrary, the nature of the counter anion has a more dominant effect. For the poly(ViHIm)-based PIL stationary phases, the on-set bleed temperature decreased with an increase in the weight percentage of the chloride-based PIL, as expected. A similar decreasing trend was obtained for the poly(ViHDIm)-based PIL columns due to the introduction of chloride anion. This observation correlated well with previous reports

− which observed that ILs containing the NTf2 anion are more thermally stable compared

− with those with halide anions due to the weak nucleophilic properties of the NTf2 anion

[32, 33]. Compared with the monomeric ILs, for example, BMIM-NTf2, which possesses a bleed temperature of approximately 185 °C, PIL-based stationary phases exhibit considerably higher bleed temperatures.

4.3.2 System constants of neat and mixed PIL stationary phases

The system constants of the six different PIL stationary phases at three temperatures

(50, 80, and 110 °C) are listed in Table 4.3. As the temperature increases, the interactions between the probe molecules and the stationary phase become weaker, resulting in decreased values for the system constants. This is true for most of examined PIL stationary phases in this study. Moreover, the statistical reliability of the generated models was further reinforced by the acceptable standard deviation values for each system constant, high correlation coefficients of the regression line, as well as acceptable values of the Fisher coefficients.

94 Table 4.3 System constants of neat and mixed PIL stationary phases examined in this study.

Stationary phase System constantsa composition/temperature (°C) c r s a b l nb) R2 b) Fb)

Poly(ViHIm-NTf2) (100%) 50 -3.32 -0.43 1.91 1.83 1.27 0.64 40 0.99 576 (0.08) (0.09) (0.08) (0.13) (0.02) 80 -3.48 -0.40 1.87 1.58 0.99 0.55 39 0.99 484 (0.07) (0.09) (0.08) (0.12) (0.02) 110 -3.48 -0.32 1.70 1.27 0.92 0.46 35 0.98 344 (0.07) (0.09) (0.08) (0.11) (0.02)

Poly(ViHIm-NTf2) (75%)/Poly(ViHIm-Cl) (25%) 50 -3.27 -0.36 2.17 3.88 0.10 0.64 34 0.99 420 (0.08) (0.10) (0.12) (0.14) (0.02) 80 -3.22 -0.25 1.94 3.20 0 0.52 37 0.99 636 (0.07) (0.09) (0.07) (0.02) 110 -3.48 -0.24 1.91 2.79 0 0.46 35 0.99 427 (0.08) (0.10) (0.08) (0.02)

Poly(ViHIm-NTf2) (50%)/ Poly(ViHIm-Cl) (50%) 50 -3.27 -0.25 2.04 4.72 0 0.64 32 0.99 474 (0.08) (0.09) (0.13) (0.02) 80 -3.39 -0.19 1.92 4.14 0 0.54 35 0.99 623 (0.07) (0.09) (0.10) (0.02) 110 -3.55 -0.17 1.86 3.66 -0.10 0.47 36 0.99 743 (0.07) (0.08) (0.08) (0.11) (0.02)

95 Table 4.3 (continued) Stationary phase System constantsa composition/temperature (°C) c r s a b l nb) R2 b) Fb) Poly(ViHIm-NTf2) (25%)/ Poly(ViHIm-Cl) (75%) 50 -3.26 -0.14 1.90 5.27 0 0.68 28 0.99 508 (0.07) (0.09) (0.17) (0.02) 80 -3.41 0 1.80 4.72 -0.17 0.57 33 0.99 759 (0.09) (0.10) (0.11) (0.02) 110 -3.50 0 1.71 4.17 -0.17 0.47 31 0.99 878 (0.08) (0.07) (0.11) (0.01) Poly(ViHDIm-NTf2) (100%) 50 -2.98 -0.33 1.25 1.40 0.24 0.75 39 0.99 1224 (0.05) (0.06) (0.05) (0.08) (0.01) 80 -3.06 -0.26 1.12 1.11 0.21 0.64 40 0.99 1025 (0.05) (0.06) (0.05) (0.08) (0.01) 110 -3.15 -0.26 1.11 0.87 0.22 0.56 35 0.99 577 (0.05) (0.06) (0.05) (0.08) (0.01) Poly(ViHDIm-NTf2) (50%)/ Poly(ViHDIm-Cl) (50%) 50 -3.17 -0.22 1.48 4.72 -0.16 0.77 32 0.99 801 (0.06) (0.08) (0.10) (0.10) (0.02) 80 -3.25 -0.16 1.41 4.09 -0.35 0.65 34 0.99 558 (0.08) (0.10) (0.11) (0.14) (0.02) 110 -3.24 -0.16 1.35 3.50 -0.53 0.55 35 0.99 522 (0.08) (0.09) (0.08) (0.13) (0.02) a) System constants: e, nonbonding and π electron interactions, s, PIL dipolarity, a, PIL hydrogen bond basicity, b, PIL hydrogen bond acidity, l, PIL dispersion forces. b) n, Number of probe analytes subjected to multiple linear regression; R2, correlation coefficient; F, Fisher coefficient.

96 To investigate the effect of the percentage of chloride anion within each PIL composition on the system constants, the six studied PIL stationary phases can be divided into two groups, namely, poly(ViHIm-NTf2)/poly(ViHIm-Cl) and poly(ViHDIm-NTf2)/poly(ViHDIm-Cl). For the poly(ViHIm-NTf2)/poly(ViHIm-Cl) group, as the weight percentage of poly(ViHIm-Cl) was increased from 0 to 75%, the hydrogen bond basicity (a-term) was enhanced significantly. The linear correlation between the weight percentage of the poly(ViHIm-Cl) PIL and the corresponding hydrogen bond basicity at three different temperatures is shown in Fig. 4.2. The correlation coefficients reflecting the linearity of the hydrogen bond basicity at 50, 80, and 110 °C were 0.912, 0.951, and 0.946, respectively. Compared with the neat poly(ViHIm-NTf2) stationary phase, the 75% by weight poly(ViHIm-Cl) stationary phase possessed values of the hydrogen bond basicity that were enhanced by 287, 298, and

328% at 50, 80, and 110 °C, respectively. This result supports the previous observations that ILs containing chloride anions exhibit strong hydrogen bond basicity and form hydrogen bonds with proton donor molecules such as alcohols [20, 26]. Correspondingly, as the hydrogen bond basicity of the stationary phase increased with increasing weight percentage of the poly(ViHIm-Cl) PIL, the hydrogen bond acidity decreased. All other system constants were largely unchanged. Similar trends were observed for the poly(ViHDIm-Cl)/poly(ViHDIm-NTf2) compositions. The value of hydrogen bond basicity increased by 337, 368, and 402% when comparing the neat poly(ViHDIm-NTf2) stationary phase to that of 50% by weight poly(ViHDIm-Cl) stationary phases at 50, 80, and 110 °C, respectively. The obtained hydrogen bond acidity decreased correspondingly.

Interestingly, the dipolarity of the poly(ViHDIm-NTf2) (50%)/poly(ViHDIm-Cl) (50%)

Figure 4-2: Linear correlation between the stationary phase weight percentage of the poly(ViHIm-Cl) PIL and the resulting hydrogen bond basicity. (♦) 50 ˚C (R2 = 0.912), (■)

80 ˚C (R2 = 0.951), (▲) 110 ˚C (R2 = 0.946).

98 stationary phase was slightly higher than that of the neat poly(ViHDIm-NTf2). All other system constants for these two PILs were similar.

The length of the alkyl chain appended to the polymerized cation seems to be another factor that affects the system constants of the resulting PIL stationary phases. For example, the poly(ViHDIm)-based PILs were more cohesive than the corresponding poly(ViHIm)-based analogues, whereas all the other system constants of the poly(ViHDIm)-based stationary phases were slightly lower than those of the poly(ViHDIm)-based stationary phases containing the same weight percentage of

Cl-based analogues

4.3.3 Effects of PIL stationary phase composition on retention behavior of selected solute molecules

It has been noted previously that the retention of analytes on IL-based GC stationary phases can be modified by tailoring the structure of the cation and anion that comprise the IL [20, 29]. When two monomeric ILs containing different counter anions (chloride

− and NTf2 ) were mixed and employed as GC stationary phases, interesting chromatographic retention characteristics were obtained by simply altering the ratio of these two components [20]. Similar trends were obtained in this study when binary PIL mixtures were applied as GC stationary phases. The retention factors for selected solutes on PIL stationary phases containing different weight percentages of chloride-based PILs at 80 °C are listed in Table 4.4. Acids and alcohols exhibited increased retention factors on PIL stationary phases containing higher weight percentage of the chloride-based PILs.

For example, compared with the neat poly(ViHIm-NTf2) stationary phase, the retention

Table 4.4 Comparison of retention factors for selected analytes on six different PIL

− stationary phases varying in the weight percentage of NTf2 and chloride anions at 80 ºC.

No. Probe Poly(ViHIm-NTf2)/Poly(ViHIm-Cl) Poly(ViHDIm-NTf2)/ Molecule (% (w/w)) Poly(ViHDIm-Cl) (% (w/w)) 100/0 75/25 50/50 25/75 100/0 50/50 1 Acetic Acid 1.4 9.0 21.6 38.0 0.39 16.6 2 Propionic Acid 2.2 17.8 38.2 67.3 0.8 35.8 3 2-Propanol 0.13 0.30 0.38 0.54 N/A 0.31 4 1-Butanol 0.47 0.89 1.5 2.5 0.30 1.7 5 1-Pentanol 1.0 1.8 3.0 4.8 0.72 3.6 6 1-Octanol 7.4 12.6 21.1 35.7 6.9 37.8 7 1-Decanol 25.3 43.6 75.5 135.0 33.6 190.1 8 Cyclohexanol 2.9 4.4 7.0 10.9 1.8 8.4 9 Benzyl 14.8 43.2 93.2 174.8 7.9 N/A Alcohol 10 Naphthalene 7.1 8.5 8.3 10.4 10.9 13.4 11 p-Xylene 0.33 0.46 0.36 0.42 0.77 0.74 12 Benzonitrile 7.1 6.4 5.6 5.8 4.8 4.4 13 Acetophenone 10.9 9.3 7.6 7.6 7.6 6.6 14 Methyl 1.1 0.87 0.64 0.58 1.2 0.90 Caproate 15 Cyclohexanone 4.1 2.7 1.9 1.5 2.6 1.6 16 2-Pentanone 0.53 0.42 0.27 0.22 0.35 0.20 17 Octylaldehyde 2.8 2.4 1.8 1.7 3.4 2.6

100 factor of propionic acid experienced an 809% increase on the (75%/25%) poly(ViHIm-NTf2)/poly(ViHIm-Cl) stationary phase, a 1736% increase on the (50%/50%) poly(ViHIm-NTf2)/poly(ViHIm-Cl) stationary phase, and a 3059% increase on the

(25%/75%) poly(ViHIm-NTf2)/poly(ViHIm-Cl) stationary phase. When the poly(ViHIm-Cl) weight percentage was increased to 25, 50, and 75%, the retention factors for benzyl alcohol exhibited increases of 292, 630, and 1181%, respectively. On the contrary, aromatic compounds, such as naphthalene and p-xylene, interacted similarly with the PIL stationary phases regardless of the nature of the anion, which resulted in only slight differences in the retention factors. Ketones, aldehydes, and analytes containing nitrile functional groups (e.g., benzonitrile, acetophenone, methyl caproate, cyclohexanone, 2-pentanone, and octylaldehyde) exhibited lower retention factors as the percentage of the poly(ViHIm-Cl) PIL in the stationary phase was increased.

The length of aliphatic side chain appended to the cationic moiety of the PIL also has effects on the retention factors of certain analytes on neat or mixed PIL stationary phases, as summarized in Table 4.5. Analytes possessing longer alkyl chains, for example,

1-chlorooctane, 1-decanol, octylaldehyde, along with several aromatic compounds, tend to retain longer on the poly(ViHDIm)-based PIL stationary phases compared with their poly(ViHIm)-based analogues containing the same percentage of Cl-based PIL. However, as the length of the aliphatic side chain of the polymeric cation was increased, retention factors of analytes containing shorter alkyl chain (e.g., 1-nitropropane, 1-butanol, and butyraldehyde) were decreased.

101

Table 4.5 Comparison of retention factors for selected analytes on four different PIL stationary phases varying in length of cationic side chain at 80 ºC.

No Probe Molecule Poly(ViRIm-NTf2) Poly(ViRIm-NTf2)/Poly(ViRIm-Cl) (% (w/w = 50/50))

R = R = R = Hexyl R = Hexadecyl Hexyl Hexadecyl 1 1-Chlorohexane 0.21 0.56 0.22 0.48 2 1-Chlorooctane 0.82 2.6 0.85 2.4 3 1-Bromohexane 0.37 1.1 0.43 1.0 4 1-Bromooctane 1.4 4.8 1.56 4.4 5 1-Octanol 7.4 6.9 21.1 37.8 6 1-Decanol 25.3 33.6 75.5 190.1 7 Octylaldehyde 2.8 3.4 1.8 2.6 8 1,2-Dichlorobenzene 1.5 2.9 2.0 3.4 9 p-Xylene 0.33 0.77 0.36 0.74 10 Ethyl benzene 0.31 0.68 0.35 0.64 11 Naphthalene 7.1 10.9 8.3 13.4 12 1-Nitropropane 0.94 0.65 0.75 0.54 13 1-Butanol 0.47 0.30 1.5 1.7 14 Butyradehyde 0.18 0.14 0.12 0.09 15 Benzaldehyde 4.5 3.5 3.8 3.4 16 Nitrobenzene 10.5 8.3 9.5 8.2 17 Acetophenone 10.9 7.6 7.6 6.6

102 4.3.4 Effect of PIL stationary-phase composition on separation selectivity for selected solute pairs

The separation selectivity is another chromatographic variable that can be affected by tuning the composition of the PIL stationary phase. The selectivities listed in Tables 4.6 and 4.7 were calculated by determining the ratio of retention factors for given solute pairs.

Table 4.6 lists the separation selectivity for various solute pairs at 80 °C on PIL stationary phases that differ from one another based on the composition of chloride anion within the stationary phase. For poly(ViHIm)-based stationary phases, when the amount of the poly(ViHIm-Cl) PIL was increased, the selectivity between proton-donor solutes and aromatic solutes increased significantly. For example, the selectivity between benzyl alcohol and benzaldehyde increased from 3.2 on the neat poly(ViHIm-NTf2) stationary phase to 10.1 on the 25% by weight poly(ViHIm-Cl) column, 24.4 on the 50% by weight poly(ViHIm-Cl), and further to 41.8 on the 75% by weight poly(ViHIm-Cl) stationary phase. However, the selectivity between aldehydes and ketones and several aromatic compounds decreased as the amount of the poly(ViHIm-Cl) PIL was increased. An example can be observed for acetophenone and naphthalene. On the neat poly(ViHIm-

NTf2) stationary phase, acetophenone eluted before naphthalene but then underwent a reversal of elution order on the 50% poly(ViHIm-NTf2)/50% poly(ViHImCl) stationary phase. Other examples of two analytes that exhibit a reversal of elution order were acetic acid and acetophenone. In this case, the selectivity for acetic acid and acetophenone ranged from 0.12 on the neat poly(ViHIm-NTf2) stationary phase (with acetic acid eluting before acetophenone) to 5.0 on the (25%/75%) poly(ViHIm-NTf2)/poly(ViHImCl) stationary phase (acetophenone eluting before acetic acid). Similar selectivity behavior

103

Table 4.6 Effect of PIL stationary phase chloride content on the selectivity of chosen

analyte pairs at 80 ºC.

No. Solute Pair Poly(ViHIm-NTf2)/Poly(ViHIm-Cl) Poly(ViHDIm-NTf2)/ (% (w/w)) Poly(ViHDIm-Cl) (% (w/w)) 100/0 75/25 50/50 25/75 100/0 50/50

1 Acetic acid/Naphthalene 0.19a 1.0 2.6 3.6 0.04 a 1.2

2 Propionic acid/Naphthalene 0.32 a 2.1 4.6 6.4 0.07 a 2.7

3 Benzyl Alcohol/Benzaldehyde 3.2 10.1 24.4 41.8 2.2 N/A

4 2-propanol/p-Xylene 0.41 a 0.65 a 1.0 1.3 N/A 0.42 a

5 1-Butanol/p-Xylene 1.4 2.0 4.2 5.9 0.39 a 2.3

6 1-Pentanol/Ethyl Benzene 3.3 4.6 8.8 12.0 1.1 5.6

7 Cyclohexanol/Ethyl Benzene 9.1 11.0 20.0 27.6 2.7 13.2

8 1-Octanol/Nibrobenzene 0.71 a 1.1 2.2 3.3 0.84 a 4.6

9 1-Decanol/Nitrobenzene 2.4 4.0 8.0 12.5 4.0 23.2

10 Acetophenone/Naphthalene 1.6 1.1 0.93 a 0.73 a 0.70 a 0.49 a

11 Benzonitrile/1,2-Dichlorobenzene 4.7 3.4 2.8 2.2 1.63 1.31

12 Cyclohexanone/Ethyl benzene 13 6.7 5.4 3.9 3.8 2.6

13 Methyl Caproate/Ethyl benzene 3.4 2.2 1.8 1.5 1.8 1.4

14 2-Pentanone/p-Xylene 1.6 0.92 a 0.74 a 0.52 a 0.45 a 0.27 a

15 Octylaldehyde/p-Xylene 8.5 5.3 5.0 4.0 4.5 3.5

16 Acetic acid/Acetophenone 0.12 a 0.96 a 2.8 5.0 0.05 a 2.5

17 Propionic acid/Acetophenone 0.2 a 1.9 5.0 8.8 0.11 a 5.4

18 Cyclohexanol/Cyclohexanone 0.7 a 1.6 3.7 7.1 0.71 a 5.2

19 1-Butanol/Methyl caproate 0.43 a 1.0 2.4 4.2 0.24 a 1.9

104 Table 4.6 (continued)

No. Solute Pair Poly(ViHIm-NTf2)/Poly(ViHIm-Cl) Poly(ViHDIm-NTf2)/ (% (w/w)) Poly(ViHDIm-Cl) (% (w/w)) 100/0 75/25 50/50 25/75 100/0 50/50

20 1-Pentanol/2-Pentanone 1.9 4.4 11.4 21.3 2.1 18.0

21 1-Octanol/Octylaldehyde 2.7 5.2 11.6 20.9 2.0 14.7

22 Benzyl Alcohol/Benzaldehyde 3.2 10.1 24.4 41.7 2.2 N/A

23 1-Decanol/Benzonitrile 2.6 6.8 13.6 23.1 7.0 43.2

a By definition, the value of selectivity should not be smaller than unity. However, in

some cases, the solute pairs exhibited reversed elution order, which makes it impossible

to report selectivities greater than one for all PIL columns.

105

Table 4.7 Effect of the length of the imidazolium aliphatic side chain substituent on the

selectivity of chosen analyte pairs at 80 ºC.

Solute Pair Poly(ViRIm-NTf2) Poly(ViRIm-NTf2)/Poly(ViRIm-Cl) (% (w/w = 50/50))

R = R = R = Hexyl R = Hexadecyl Hexyl Hexadecyl 1 1-Chlorohexane/Butyraldehyde 1.2 3.9 1.8 5.2 2 1-Bromohexane/1-Nitropropane 0.39 a 1.6 0.57 a 1.8 3 1-Chlorooctane/Butyraldehyde 4.5 18.1 6.9 25.8 4 1-Bromooctane/1-Nitropropane 1.5 7.4 2.1 8.2 5 1-Octanol/Benzaldehyde 1.6 2.0 5.5 11.2 6 Octylaldehyde/1-Nitropropane 3.0 5.3 2.4 4.8 7 1-Decanol/Nitrobenzene 2.4 4.0 7.9 23.2 8 1-Decanol/Acetophenone 2.3 4.4 9.9 28.8 9 Naphthalene/Nitrobenzene 0.67 a 1.3 0.87 a 1.6 10 Naphthalene/Acetophenone 0.64 a 1.4 1.1 2.0 11 Ethyl benzene/Butyraldehyde 1.7 4.8 2.8 6.9 12 1,2-Dichlorobenzene/1-Butanol 3.2 9.9 1.3 2.0 13 p-Xylene/Butyraldehyde 1.8 5.3 3.0 8.0

a By definition, the value of selectivity should not be smaller than unity. However, in

some cases, the solute pairs exhibited reversed elution order, which makes it impossible

to report selectivities greater than one for all PIL columns.

106 was observed when comparing the neat poly(ViHDIm-NTf2) stationary phase to the

(50%/50%) poly(ViHDIm- NTf2)/poly(ViHDIm-Cl).

The aliphatic side chain of the polymeric cation also influenced the separation selectivity of selected solute pairs, as summarized in Table 4.7. It was observed that the selectivity between analytes with long hydrocarbon chains and those with relatively shorter hydrocarbon chain, as well as that between aliphatic and aromatic compounds, was improved on the poly(ViHDIm)-based stationary phases compared with the poly(ViHIm)-based analogues. For instance, the selectivity between 1-decanol and acetophenone on the (50%/50%) poly(ViHIm-NTf2)/poly(ViHIm-Cl) stationary phases was 9.9, and this selectivity increased to 28.8 on the (50%/50%) poly(ViHDIm-NTf2)/poly(ViHDIm-Cl) stationary phase. A reversal of elution order was also achieved for certain solute pairs, such as 1-bromohexane and 1-nitropropane.

4.4. Concluding Remarks

ILs have been shown to be a unique class of GC stationary phases that exhibit attractive separation selectivity and thermal stability. The on-set bleed temperature of ILs can be further increased by examining PILs. In this study, mixtures containing two different PILs were employed as GC stationary phases for the first time. Compared with the GC stationary phases that are composed of monomeric IL, the bleed temperatures of stationary phases composed of binary PIL mixtures were significantly improved, although the increase of chloride anion content resulted in a decrease of the on-set bleed temperature of resulting PIL mixture. To investigate the effect of the polymerized cation

107 and counter anion on the system constants of resulting stationary phase, two different types of polymerized cations, namely, poly(ViHIm)-based PIL and poly(ViHDIm)-based

PIL, as well as two different types of anions, chloride and NTf2, were involved in this study. As expected, stationary phases containing higher percentage of chloride anion possess significantly enriched hydrogen bond basicity interactions, whereas the poly(ViHDIm)-based PIL show higher dispersion interactions. This modulation in the system constants allows for the fine tuning of analyte retention and the selectivity of chosen analytes. The enrichment of the PIL mixture with chloride anion resulted in improved separation selectivity of analytes capable of undergoing hydrogen-bonding interactions with the stationary phase as well as ketones, aldehydes, and aromatic compounds. With the lengthening of the cationic side chain, the resulting stationary phases tend to selectively retain probes via dispersive-type interactions. Reversal of analyte elution order was observed for chosen analytes by tuning the composition of the mixed PIL stationary phase. The results demonstrate that mixtures of PILs can be used as promising GC stationary phases that can provide unique chromatographic selectivity.

Acknowledgements

National Science Foundation for a CAREER grant (CHE-0748612)

108 References

1. Welton, T., Chem. Rev. 1999, 99, 2071-2083.

2. Wasserscheid, P., Keim, W., Angew. Chem. Int. Ed. 2000, 39, 3772-3789.

3. Cole, A. C., Jensen, J. L., Ntai, I., Tran, K. L. T., Weaver, K. J., Forbes, D. C., Davis, J.

H. Jr., J. Am. Chem. Soc. 2002, 124, 5962-5963.

4. Sheldon, R. A., Lau, R. M., Sorgedrager, M. J., van Rantwijk, F., Seddon, K. R., Green.

Chem. 2002, 4, 147-151.

5. Morris, R. E., Chem. Commun. 2009, 21, 2990-2998.

6. Dai, S., Ju, Y. H., Barnes, C. E., J. Chem. Soc. Dalton. Trans. 1999, 8, 1201-1202.

7. Chun, S., Dzyuba, S. V., Bartsch, R. A., Anal. Chem. 2001, 73, 3737-3741.

8. Visser, A. E., Swatloski, R. P., Reichert, W. M., Mayton, R., Sheff, S., Wierzbicki, A.,

Davis, J. H. Jr., Rogers, R. D., Chem. Commun. 2001, 135-136.

9. Liu, J. F., Jiang, G. B., Chi, Y. G., Cai, Y. Q., Zhou, Q. X., Hu, J. T., Anal. Chem. 2003,

75, 5870-5876.

10. Yao, C., Pitner, W. R., Anderson, J. L., Anal. Chem. 2009, 81, 5054-5063.

11. Liu, J. F., Li, N., Jiang, G. B., Liu, J. M., Jönsson, J. Å., Wen, M. J., J. Chromatogr. A

2005, 1066, 27-32.

12. Zhao, F., Meng, Y., Anderson, J. L., J. Chromatogr. A 2008, 1208, 1-9.

13. Meng, Y., Pino, V., Anderson, J. L. Anal. Chem. 2009, 81, 7107-7112.

14. Armstrong, D. W., Zhang, L. K., He, L., Gross, M. L., Anal. Chem. 2001, 73,

3679-3686.

15. Mank, M., Stahl, B., Boehm, G., Anal. Chem. 2004, 76, 2938-2950.

109 16. Tholey, A., Heinzle, E., Anal. Bioanal. Chem. 2006, 386, 24-37.

17. Rouse, C. A., Acree, W. E. Jr., J. Chromatogr. 1986, 357(1), 33-37.

18. Byrd, J. M., Acree, W. E. Jr., Mol. Cryst. Liq. Cryst., Letter Sections.1988, 5(5),

163-169.

19. Poole, C. F., Kiridena, W., Nawas, M. I., Koziol, W. W., J. Sep. Sci. 2002, 25,

749-759.

20. Baltazar, Q. Q., Leininger, S. K., Anderson, J. L., J. Chromatogr. A 2008, 1182,

119-127.

21. Yao, C., Anderson, J. L., J. Chromatogr. A 2009, 1216, 1658-1712.

22. Anderson, J. L., Armstrong, D. W., Anal. Chem. 2005, 77, 6453-6462.

23. Hsieh, Y.-N., Ho, W.-Y., Horng, R.S., Huang, P.-C., Hsu, C.-Y., Huang, H.-H., Kuei,

C.-H., Chromatographia. 2007, 66, 607-611.

24. Hsieh, Y.-N., Horng, R.-S., Ho, W.-Y., Huang, P.-C., Hsu, C.-Y., Whang, T.-J., Kuei,

C.-H., Chromatographia. 2008, 67, 413-420.

25. Abraham, M. H., Chem. Soc. Rev. 1993, 22, 73-83.

26. Anderson, J. L., Ding, J., Welton, T., Armstrong, D. W., J. Am. Chem. Soc. 2002 124,

14247-14254.

27. Mutelet, F., Jaubert, J. N., J. Chromatogr. A 2006, 1102, 256-267.

28. Sprunger, L. M. , Proctor, A., Acree, W. E., Abraham, M. H., Fluid. Phase. Equilib.

2008, 265, 104-111.

29. Zhao, Q., Eichhorn, J., Pitner, W. R., Anderson, J. L., Anal. Bioanal. Chem. 2009, 395,

225-234.

30. Sharma, N. K., Tickell, M. D., Anderson, J. L., Kaar, J., Pino, V., Wicker, B. F.,

110 Armstrong, D. W., Davis, J. H. Jr., Russell, A. J., Chem. Commun. 2006, 646-648.

31. Marcilla, R., Blazquez, J. A., Rodriguez, J., Pomposo, J. A., Mecerreyes, D., J. Polym.

Sci. A: Polym. Chem. 2004, 42, 208-212.

32. Anderson, J. L., Armstrong, D. W., Anal. Chem. 2003, 75, 4851-4858.

33. Ngo, H. L., LeCompte, K., Hargens, L., McEwen, A. B., Thermochim. Acta. 2000,

357-358, 97-102.

111 Chapter 5

Overview: Application of Ionic Liquids as Sorbent Coatings in

Solid-Phase Microextraction

5.1 Introduction

SPME is a sample preparation technique introduced by Pawliszyn and co-workers in the early 1990s [1]. Since its introduction, SPME has obtained increasing attention due to the fact that it possesses numerous advantages over traditional sample preparation methods, including its speed, simplicity, and ease of operation and automation. Moreover,

SPME combines sampling and sample preparation into one step and can be easily coupled with various separation techniques such as gas chromatography (GC). SPME consists of a thin, fused-silica fiber coated or bonded with a polymer sorbent material.

During exposure to sample matrix via headspace or direct-immersion mode, the analytes partition to the sorbent coating resulting in an equilibration between the analytes and the coating material. When coupled with GC, the extracted analytes are desorbed from the sorbent coating by thermal desorption in the injection port.

112 ILs have proven to be ideal candidates for the development of novel SPME sorbent coatings and often exhibit high selectivity, sensitivity, thermal stability, and long lifetimes.

This is largely due to the unique properties of ILs, such as their high thermal stability, low vapor pressure, capability of interacting with analytes through a multitude of solvation interactions, as well as their structural tunability. Currently, a wide range of ILs and their polymeric analogs have been successfully applied as SPME sorbent coating for various target analytes. The application of ILs and PILs in SPME is summarized in Table

5.1.

5.2 Application of Ionic Liquids as Sorbent Coatings in SPME

5.2.1 Physically coated IL-based SPME sorbent coatings

Jiang and co-workers [2] first reported the development of a disposable SPME fiber using the 1-octyl-3-methylimidazolium hexafluorophosphate (OMIM-PF6) IL. The coating solution was prepared by mixing the IL with dichloromethane at a ratio of 9:1

(v/v). A pretreated stainless steel wire or fused silica fiber was then dipped into this solution followed by the evaporation of the dichloromethane. After optimizing the coating procedure, desorption time and temperature, extraction temperature, extraction time, and salt effect, the IL-coated fiber was used to extract benzene, toluene, ethylbenzene, and xylenes (BTEX) from paints. More recently, Rouhollahi and co-workers used a similar approach to generate disposable IL-based SPME coating using four imidazolium-based ILs containing various types of counter anions [3]. The resulting fibers were used for the determination of methyl tert-butyl ether (MTBE) in a gasoline

113 Table 5.1 Application of ILs and PILs in SPME.

Applied ILs Fiber coating composition Application 2 OMIM-PF6 Pure IL coating BTEX in paints

BMIM-BF4; OMIM-BF4; OMIM-PF6; Pure IL coating MTBE in a gasoline sample EMIM-ethylsulphate3 OMIM-TfO; Nafion membranes-supported PAHs in water sample 1-Benzyl-3-methylimidazolium TfO; IL coating 1-Phenylpropyl-3-methylimidazolium TfO4 5 BMIM-PF6; OMIM-PF6 IL coated on etched PAHs released from the fused-silica fibers burning of mosquito coil incense 6 1-Ethoxylethyl-3-methylimidazolium NTf2 IL impregnated in a Methamphetamine and cross-linked silicone amphetamine in human urine elastomer samples 7 BMIM-BF4 Polyaniline-IL composite Benzene derivatives in water film coated on Pt wires samples Bis-hydroxyethyl imidazolium trioxyethylene ILs and PILs bonded to silica Short-chain alcohol and NTf2; particles, and then glued onto amines Bis-hydroxyethyl imidazolium trioxyethylene TfO; fused silica fibers Poly[1,1’-(1,6-hexanediyl) bis-styrene imidazolium-NTf2]; Poly[1,1’-(1,6-hexanediyl) bis-styrene imidazolium-TfO]8 1-Allyl-3-methylimidazolium PF6; ILs bonded to fused silica Phenolic environmental 9 1-Allyl-3-methylimidazolium NTf2 fiber by sol-gel technology estrogens and aromatic amines from water samples 1-Methyl-3-(3-trimethoxysilylpropyl) imidazolium IL bonded to fused silica MTBE in a gasoline sample 10 NTf2 fiber

114 Table 5.1 (continued) Applied ILs Fiber coating composition Application Poly(1-vinyl-3-hexylimidazolium-NTf2); PIL Esters and FAMEs from Poly(1-vinyl-3-dodecylimidazolium-NTf2); synthetic wine 11 Poly(1-vinyl-3-hexadecylimidazolium-NTf2) 12 Poly(1-vinyl-3-hexadecylimidazolium-NTf2) PIL Low-Volatility aliphatic hydrocarbons and FAME 13 Poly(1-vinyl-3-hexadecylimidazolium-NTf2) PIL Direct-immersion extraction of PAHs and substituted phenols from water samples Poly[(1-4-vinylbenzyl)-3-hexadecylimidazolium-N PIL with aromatic Direct-immersion extraction 14 Tf2] functionality of PAHs from aqueous samples Poly(1-vinyl-3-hexylimidazolium-chloride)15 PIL with Cl− anion Volatile fatty acids (VFAs) and alcohols from aqueous or heptane samples Poly(1-vinyl-3-hexylimidazolium-NTf2); Binary mixture containing Selected analytes from water Poly(1-vinyl-3-hexadecylimidazolium-NTf2); various weight ratio of PILs samples Poly(1-vinyl-3-hexylimidazolium-Cl); Poly(1-vinyl-3-hexadecylimidazolium-Cl)16 Poly(1-vinyl-3-hexylimidazolium-NTf2); Task-specific PILs CO2 from simulated flue gas Poly(1-vinyl-3-hexylimidazolium-taurinate)17,18 1-Allyl- 1-allyl-3-(6’-oxo-benzo-15-crown- 5 Task-specific ILs bonded to Alcohols, phthalate esters, 19 hexyl) imidazolium PF6 fused silica fiber by sol-gel phenolic environmental technology estrogens, fatty acids, and aromatic amines. OMIM-Cl; IL-mediated sol-gel coatings Polar, non-polar, and Trihexyltetradecylphosphonium BF4; for in-tube SPME moderately polar analytes 20,21 N-butyl-4-methylpyridinium BF4 from aqueous matrices

115 sample. While the relatively high viscosity of ILs allowed for successful coating on the fused silica fiber, loss of the IL sorbent phase was observed during the thermal desorption process due to the decreased viscosity of the IL at elevated temperature. This resulted in the IL flowing off the fiber and into the injection port requiring a washing and subsequent recoating step after each extraction.

Various methods have been proposed in order to obtain thicker IL-based SPME sorbent coatings. Nafion was utilized as a support medium to aid in the formation of a thicker and more even IL coating layer on the SPME fiber, as demonstrated by Kuei and co-workers [4]. ILs containing the trifluoromethanesulfonate (TfO−) anion and imidazolium cations with various substituents were chosen as sorbent coatings for the microextraction of PAHs in water samples. A comparison of the extraction efficiencies between the IL-based fibers coated with and without Nafion-supported membrane for the tested PAHs revealed that significantly larger amounts of analytes could be extracted by the Nafion-supported IL fiber due to the increased coating thickness. Liu and co-workers

[5] developed a simple and inexpensive SPME device using a plastic syringe and etched fused silica fiber to pursue higher fiber surface area and coating thicknesses. By immersing of the fused silica fiber into a saturated solution of ammonium hydrogen difluoride in methanol followed by heating at high temperature, a uniform and porous fiber surface was obtained. An increased amount of the BMIM-PF6 IL was coated on the etched fiber. The etched IL-based fiber exhibited highest extraction efficiency for PAHs compared with the IL-based fused silica fiber without etching or the Nafion-membrane supported fiber. However, fibers developed using these procedures still required washing

116 and recoating after every extraction/desorption step.

Efforts have been devoted to overcome the tendency of ILs to flow at elevated temperatures typically employed in GC injection ports. He and co-workers [6] demonstrated the feasibility of generating reusable IL-based SPME fibers through the cross-linkage of IL impregnated silicone elastomers on a fused silica support. The silicone elastomer was mixed with the 1-ethoxylethyl-3-methylimidazolium NTf2 IL to allow for the complete cross-linking of the polymer. A film thickness of approximately 50

μm was achieved by this approach. Extractions of methamphetamine and amphetamine were carried out using the IL-based SPME fiber as well as commercial polydimethylsiloxane (PDMS) fibers. No change in the extraction efficiency was observed after more than 100 extractions. Another approach involves the electrochemical polymerization of polyaniline-BMIM-BF4 composite film coated platinum wire

(PANI-IL/Pt) for the extraction of benzene derivatives [7]. After pretreatment, the Pt wire was immersed into 0.1 M aniline + 0.1 M HNO3 + 0.3% (V/V) BMIM-BF4 solution, and the potential scan between -0.1 V and 0.9 V for 200 times at scan rate of 50 mV/s was cycled for 200 times. The coated Pt wire was then immobilized on the laboratory-made

SPME device using high-temperature epoxy resin. A scanning electron microscopic study revealed that the proposed electrochemical polymerization method produced an even and porous sorbent coating with a film thickness of approximately 80 μm. It was observed that the PANI-IL/Pt fiber exhibited much higher extraction efficiency for the studied benzene derivatives when compared to the commercial PDMS fiber with a film thickness of 100 μm. The obtained sorbent coating was found to be thermally stable up to 320 °C, however, no data regarding the fiber lifetime was reported.

117 5.2.2 Chemically bonded IL-based SPME sorbent coatings

In order to enhance the thermal and chemical stability of IL-based SPME sorbent coating materials, Armstrong and co-workers developed silica particles bonded with two monomeric and two polymeric dicationic ILs, and the obtained silica particles were then glued on fused silica fiber to form a 50-μm highly porous coating layer [8]. Figure 5.1 shows the structures of ILs that were used to form the bonded IL-based sorbent coatings.

The resulting fibers were applied to the extraction of short-chain alcohols via both headspace and direct-immersion modes. The ethanol content of several alcoholic beverages and a National Institute of Standards and Technology reference material were examined for validation.

Sol-gel technology has been applied as an alternative to conventional coating preparation methods to generate more thermally stable as well as pH and solvent resistant fiber coatings. Feng and co-workers utilized two allyl-functionalized ILs, namely

1-allyl-3-methylimidazolium PF6 and 1-allyl-3-methylimidazolium NTf2, to prepare bonded IL-based sorbent coatings using sol-gel technology and free radical cross-linking method [9]. The bonded IL-based sorbent coatings were found to have highly porous surface, thermal stability of up to 380 °C, good solvent resistance, and a wide range of pH stability (from 0 to 14). Extraction of phenolic environmental estrogens (PEEs) and aromatic amines from water samples revealed that the developed sorbent coatings exhibited good reproducibility, extraction efficiency, and selectivity towards these highly polar analytes.

118 OEt O H SiO2 Si N O O O OH O O 1 O

NTf2 NTf2

OEt O H SiO2 Si N O O O OH O O

2 O TfO TfO

NTf2 NTf2 NC O n

NTf2 NTf2 NH

O Si O SiO NTf2 2 n NTf2 O Si O n 3 NH NTf2 NTf2

O n NC

TfO TfO NC O n

TfO TfO NH

O Si O SiO2 n TfO TfO O Si O n 4 NH TfO TfO

O n NC

Figure 5-1: Monomeric ILs 1 and 2, and polymeric ILs 3 and 4, were utilized to create

bonded IL-based SPME coatings.

119 Rouhollahi and co-workers proposed an easy method of chemically modifying the surface of fused silica fiber by using 1-methyl-3-(3-trimethoxysilylpropyl) imidazolium

NTf2 [10]. After removal of the polyimide polymer from the surface, the pretreated fused silica fiber was immersed into a methylene chloride solution comprising of 5%

1-methyl-3-(3-trimethoxysilylpropyl) imidazolium NTf2 and allowed to react under N2 atmosphere and reflux for 24 h. Compared to fibers that were physically coated with the same IL, the chemically bonded IL-based sorbent coatings exhibited enhanced thermal stability. Extraction of MTBE from a gasoline sample was successfully performed using the developed fiber and satisfactory recoveries were obtained.

5.2.3 PIL-based SPME sorbent coatings

As described in previous chapters, one of the approaches to improve the thermal stability of ILs involves the synthesis of PILs. PILs were first applied as SPME sorbent coatings by our group, and provided enhanced uniformity, thermal stability, and extended fiber lifetime [11]. Three different types of PILs were synthesized through the free radical polymerization of imidazolium-based IL monomers containing varied alky chain length

- substituents followed by exchanging the counter anion to NTf2 through a metathesis

- reaction. The NTf2 anion imparted the PIL enhanced thermal stability and increased the hydrophobic nature of the PIL. The polymerized cation produced a higher viscosity material that prevented the IL coating from dripping off the fiber at high temperature.

Moreover, this stable and reusable coating allows for longer fiber lifetimes, exceptional extraction-to-extraction reproducibility, and eliminates the need for compromising between desorption time and temperature within the injection port. Scanning electron

120 micrographs illustrated that a smooth, even, and homogeneous PIL-based coating can be formed on the surface of fused silica supports. When used in the selective extraction of esters and fatty acid methyl esters (FAMEs) from a synthetic wine matrix, comparable or superior sensitivities and detection limits were obtained for the PIL-coated fibers relative to the commercial PDMS coatings with a similar film thickness. PIL-based SPME fibers were also applied to the determination of analytes with high boiling points and low vapor pressures using high-temperature static headspace extraction [12]. Three distinct ILs were carefully designed, tailored, and utilized as thermally stable solvents for high-temperature headspace extraction, thermally stable and selective sorbent coatings for headspace

SPME, as well as a highly selective and low-bleed GC stationary phase.

PIL-based coatings have also been used in direct-immersion SPME [13]. A highly hydrophobic PIL, poly(1-vinyl-3-hexadecylimidazolium-NTf2), was applied in the direct-immersion extraction of PAHs and substituted phenols from deionized (DI) water as well as real water samples. The performance of the PIL fiber was compared with commercial SPME fibers including PDMS and polyacrylate (PA). The sensitivities of the

PIL, PDMS, and PA coatings were normalized by their corresponding film thicknesses since it is well known that in SPME thicker sorbent coatings are beneficial for obtaining higher extraction efficiencies under equilibrium conditions. It was observed that the PIL coating possessed enhanced affinity towards all the analytes studied compared to the

PDMS coating. In terms of the fiber lifetime, the performance of the PIL fiber was still within acceptable range after 50 direct-immersion extractions.

More recently, several examples of improving the performance of PIL-based sorbent coating by tailoring the PIL structures were demonstrated by our group. A new generation

121 of PILs, namely poly[(1-4-vinylbenzyl)-3-hexadecylimidazolium-NTf2], was synthesized and applied as SPME sorbent coating for the extraction of PAHs from aqueous samples using the direct-immersion mode [14]. It was observed that with the introduction of aromatic moieties to the PIL, the developed sorbent coating exhibited impressive selectivity towards PAHs due to enhanced π-π interactions. In another example, poly(1-vinyl-3-hexylimidazolium chloride) was applied as a selective SPME coating for the extraction of polar compounds including volatile fatty acids (VFAs) and alcohols [15].

Compared to the poly(1-vinyl-3-hexylimidazolium NTf2) fiber, the chloride-based PIL sorbent coatings provided higher selectivity towards more polar compounds due to the higher hydrogen bond basicity provided by the chloride anion. In a similar fashion, four

PIL-based coatings consisting of two different cations, namely poly(1-vinyl-3-hexylimidazolium) and poly(1-vinyl-3-hexadecylimidazolium), coupled with NTf2 and Cl anions, were mixed at various weight ratios and applied as SPME coatings for the extraction of selected analytes including alcohols [16]. It was found that the extraction efficiency of alcohols increased with the increase of the weight percentage of the chloride-based sorbent coating. These studies clearly indicated the possibility of tuning the interaction between the analytes of interest and PIL-based coatings by simply tailoring the structure of the applied PILs resulting in improved extraction performance of the developed method.

5.2.4 Task-specific ILs and PILs as SPME sorbent coatings

As described previously, one of the most interesting characteristics of ILs is their structural tunability, which allows for the modulation of physical and chemical properties

122 by using different cation/anion combinations and/or by introducing desired functional groups to either component. The resulting ILs are defined as task-specific ILs and are capable of interacting with analytes or substrates in specific ways due to the presence of the given functionality. Our group first reported the design, synthesis, and application of task-specific PILs for the selective extraction of CO2 using SPME (see Chapter 6, 7)

[17,18]. Two different PIL-based sorbent coatings, namely poly(1-vinyl-3-hexylimidazolium NTf2) and poly(1-vinyl-3-hexylimidazolium taurinate), exhibited varied selectivities towards CO2 due to the functional groups within the PIL that imparted varied mechanisms of CO2 capture. The NTf2-based PIL, which is capable of extracting CO2 more by physical sorption, demonstrated comparable extraction of CO2 with that of the commercial Carboxen-PDMS fiber despite the fact that the surface area of the PIL-based coating was much smaller than that of the commercial fiber. Moreover, the NTf2-based PIL exhibited superior CO2/CH4 and CO2/N2 selectivity compared to the commercial fiber. A microscopic study illustrated the morphology of the fiber coating changed drastically under CO2 uptake. Due to the fact that the amine groups in the taurinate anion can reversibly and selectively react with CO2 and form a carbamate salt, the taurinate-based PIL exhibited enhanced storage capacity and improved resistance to water vapor. The abilities of the PIL-based coatings to extract CO2 from simulated flue gas were evaluated by examining the effect of humidity and temperature on CO2 extraction and selectivity. This novel class of sorbent coatings has opened up the possibility of designing task-specific PIL materials for the selective and sensitive determination of target analytes using SPME.

123 A crown ether functionalized IL, namely 1-allyl-3-(6’-oxo-benzo-15-crown-5 hexyl) imidazolium PF6 was synthesized by Liu and co-workers, and chemically bonded on a fused silica fiber surface using sol-gel method [19]. The formation of the IL-based copolymer coating was confirmed by FT-IR spectroscopy. Compared to the previously reported sol-gel derived 1-allyl-3-methylimidazolium PF6-based coating, this task-specific IL-based sorbent coating exhibited enhanced extraction efficiency of polar analytes including alcohols, phthalate esters, phenolic environ

mental estrogens, fatty acids, and aromatic amines.

5.2.5 IL-mediated SPME sorbent coatings

While ILs have been successfully applied as sorbent coating materials for SPME, they have also been exploited in in-tube SPME studies. In-tube SPME, also known as capillary microextraction, consists of a capillary in which the sorbent coating is placed on the inner wall. To increase the thermal and chemical stability of the sorbent coating in in-tube SPME, hybrid organic–inorganic coatings have been bonded on the fused silica capillary inner wall surface by sol-gel methods. Due to the enhanced stability of reactants in ILs, Malik and co-workers utilized ILs as the co-solvent for the synthesis of hybrid organic–inorganic sol-gel coatings for in-tube SPME [20,21]. It should be noted that the presence of ILs in the sol-gel synthesis played no role in the extraction as the capillaries were thoroughly rinsed and dried prior to use to ensure removal of undesired debris as well as the residual ILs on the surface of the sol-gel coatings. The ILs contributed as porogens in the sol-gel synthesis; it was observed that the cation and anion portions of the

ILs have unique effects on the pore structure and distribution, which make them

124 advantageous porogens compared with organic molecules. A scanning electron microscopic study revealed that the IL-mediated sol-gel coatings were significantly more porous than their counterparts prepared without ILs. In some of the sol-gel coatings studied, the increase in the surface area of sol-gel coatings resulted in enhanced extraction efficiencies for selected polar, nonpolar, and moderately polar analytes. The application of ILs in the preparation of sol-gel coatings shows great potential for producing stable and sensitive sorbents for in-tube SPME as well as fiber SPME.

5.3 Summary

This chapter provides an overview of the wide application of ILs and PILs in SPME.

A brief description about SPME and various advantages of using ILs as SPME sorbent coating is presented. The current literature relative to the application of ILs and PILs in

SPME are extensively and critically reviewed.

The following three chapters of this dissertation provide aspects in the ILs as SPME sorbent coatings for various applications. Chapter six introduces the design, synthesis, and application of two different PILs as SPME sorbent coatings for the selective capture of CO2. The two PIL coatings exhibited different mechanisms of CO2 capture, namely physical sorption by the poly(1-vinyl-3-hexylimidazolium NTf2) coatings and complex formation by the poly(1-vinyl-3-hexylimidazolium taurinate) coatings. Compared to commercial SPME fibers such as PDMS and Carboxen-PDMS fibers, the poly(1-vinyl-3-hexylimidazolium NTf2) fiber exhibited comparable CO2 extraction efficiency at CO2 pressure of 112 kPa, despite the fact that the carboxen fiber possessed a

125 much larger coating film thickness. A sorbate storage capacity study revealed that the poly(1-vinyl-3-hexylimidazolium taurinate) coatings exhibited superior capability in retaining the extracted CO2 on the fiber coating, mostly due to the formation of the carbamate complex.

Chapter seven presents the application of PIL-based SPME sorbent coatings for the selective extraction of CO2 from simulated flue gas. For comparison purpose, the commercial Carboxen-PDMS fiber was also included in this study. A study into the effect of humidity revealed that the poly(1-vinyl-3-hexylimidazolium taurinate) fiber exhibited enhanced resistance to humidity, presumably due to the unique mechanism of CO2 capture. The poly(1-vinyl-3-hexylimidazolium NTf2) fiber, on the other hand, showed superior CO2/CH4 and CO2/N2 selectivity compared to the Carboxen fiber. The effect of temperature on the performance of the PIL-based fibers as well as the Carboxen-PDMS fiber was evaluated.

Chapter eight introduces the application of SPME coupled to chiral GC as a rapid and sensitive sampling method for the “on-line” analysis of chiral molecules from IL solvents. Two commercial SPME fibers, namely PDMS and PA fibers, as well as a

PIL-based fiber, poly(1-vinyl-3-hexylimidazolium NTf2) fiber, were used and compared.

The analytical performance of the developed method was thoroughly evaluated. The

SPME method was successfully applied to determine the enantiomeric excess from selected mixtures of chiral molecules. A preliminary study using an “on-fiber” derivatization approach was performed, revealing that the stereoisomers extracted by the

SPME fiber can be efficiently derivatized using a short “on-fiber” derivatization step.

126 Chapter 6

Polymeric Ionic Liquids as CO2 Selective Sorbent Coatings for

Solid-Phase Microextraction

A paper published in Analytical Chemistry 1

Qichao Zhao, Jonathan C. Wajert, Jared L. Anderson

Abstract

Two polymeric ionic liquids (PIL) were synthesized and employed as sorbent coatings in solid-phase microextraction (SPME) for the selective extraction of CO2. The two coatings, poly(1-vinyl-3-hexylimidazolium) bis[(trifluoromethyl) sulfonyl]imide

[poly(VHIM-NTf2)] and poly(1-vinyl-3-hexylimidazolium) taurate [poly(VHIM-taurate)], exhibited varied selectivity toward CO2 due to functional groups within the PIL that imparted different mechanisms of CO2 capture. Extraction efficiencies were compared to those of two commercial SPME fibers [poly(dimethyl siloxane) (PDMS) and Carboxen-

Chemical Society

127 PDMS]. The poly(VHIM-NTf2) PIL fiber exhibited comparable extraction efficiency at high CO2 pressure compared to the Carboxen-PDMS fiber, even though the PIL-based fibers possessed much smaller film thicknesses. Calibration curves generated in pure CO2 showed that the sensitivity of the poly(VHIM-NTf2) coating was comparable to that of the Carboxen-PDMS fiber with both PIL-based fibers exhibiting larger linear ranges and higher extraction-to-extraction reproducibility. The storage ability for selected fibers was examined and revealed that the PIL-based coatings exhibited superior capability in retaining the CO2 sorbate on the fiber under different storage conditions, particularly for the poly(VHIM-taurate) PIL which reversibly captures CO2 as a carbamate salt.

6.1 Introduction

An intense area of research today lies in the development of new methods and materials for the sequestration of CO2. Global warming, due in large part to greenhouse gas emission from fossil fuel combustion, has attracted concerns from the scientific community, industry, and the general public. According to a report from the Energy

Information Administration in 2005, approximately 28.2 billion metric tons of CO2 has been released to the atmosphere due to the consumption and flaring of fossil fuels [1].

However, in the foreseeable future, fossil fuels will continue to be one of the primary global energy feedstock sources. Hence, it is no surprise that efforts have been devoted to developing technologies and materials that can effectively monitor and stabilize atmospheric CO2 levels. Sensitive, selective, convenient, and reproducible methods are desired in order to perform sampling of CO2. Moreover, materials with high affinity and

128 selectivity for CO2 need to be designed and coupled with the available sampling methods.

Due to its speed, simplicity, and ease of operation, solid-phase microextraction

(SPME) has become a widely used extraction technique since its introduction by

Pawliszyn and co-workers in the early 1990s [2-4]. SPME possesses many advantages over conventional sample preparation methods as it combines sampling and sample preparation into one step and can be readily automated and coupled with separation techniques such as gas chromatography (GC). SPME consists of a fiber support coated with a sorbent material that is exposed to the sample matrix by either headspace or direct-immersion sampling. Equilibration is subsequently established between the analytes and the fiber coating. When coupled with GC, the analytes are desorbed from the sorbent coating by thermal desorption in the injection port.

SPME has been demonstrated as a fast, accurate, and reproducible sampling method for airborne compounds. It has been reported that SPME can be utilized for the sampling of volatile chlorinated hydrocarbons [5], volatile organic compounds (VOCs) [6], formaldehyde [7], particulate matter [8,9], hydrocarbons [10,11], trimethylamine [12], and hydrogen cyanide [13]. SPME can also be applied as a time-weighted average (TWA) sampler for gas-phase analytes by retracting the coated fiber a known distance into the needle housing during the sampling period [14]. Gas-phase compounds such as alkanes, primary alcohols, and methyl esters containing up to 22 carbons can be quantified by

SPME during equilibration [15]. Special interfaces have been developed for SPME to allow for high-speed GC analysis of benzene, toluene, ethyl benzene, and xylene (BTEX)

[16], sampling from inhaler-administered drug, spray insect repellent, and tailpipe diesel exhaust [17], as well as nonequilibrium sampling and determination of airborne VOCs

129 [18]. A major thrust of current SPME research involves the development of new sorbent coatings that are capable of expanding the range of analytes that can be selectively extracted. Our group is particularly interested in developing “task-specific” sorbent coatings based on polymeric ionic liquids (PILs) that exhibit high selectivity and high extraction efficiency for targeted analytes. With the growing interest in detecting and quantifying CO2 in gas streams, our efforts have been focused on developing sorbent coatings suitable for its selective and sensitive detection.

Ionic liquids (ILs) are a class of nonmolecular ionic compounds that have melting points below 100 °C [19]. ILs have a number of significant advantages over traditional organic solvents; for example, many ILs possess very low vapor pressure at ambient temperature and exhibit high thermal stability. The physical and chemical properties of

ILs, including viscosity and solubility with other solvents, can be tailored by altering the combination of cations and anions. In addition, functional groups can be introduced to either component to form task-specific ionic liquids (TSILs) [20], which are capable of interacting with analytes or substrates in specific ways. Due to their unique properties and versatility, ILs have been examined as extraction media for analytical microextractions

[21-26]. Liu et al. first used IL-coated disposable SPME fibers for the analysis of BTEX in paints [24]. Alternative methods have been proposed to pursue smooth, thick, and stable IL-based coatings on SPME fibers using a Nafion membrane-supported IL coating

[25], IL coated on etched fused-silica fibers [26], and IL impregnated in a cross-linked silicone elastomer [27]. Our group first utilized PILs as SPME sorbent coatings [28,29], which resulted in enhanced thermal stability, high extraction-to-extraction reproducibility, and extended fiber lifetime. A logical progression of our work is to design sorbent

130 coatings based on task-specific PILs that exhibit high extraction selectivities and efficiency for desired analyte(s) in addition to producing highly reproducible and long lifetime devices.

In this work, PIL-based sorbent coatings were used for the determination of CO2 using SPME. The structures of these PILs, namely, poly(1-vinyl-3-hexylimidazolium) bis[(trifluoromethyl)sulfonyl]imide [poly(VHIM-NTf2)] and poly(1-vinyl-3-hexylimidazolium) taurate [poly(VHIM-taurate)], were carefully tailored to enhance CO2 solubility. SPME fibers were coated with neat PILs as well as mixtures with desired weight percentage of these two PILs. For comparison, two commercially available SPME fibers, namely, poly(dimethylsiloxane) (PDMS) (film thickness of 7 μm) and Carboxen-PDMS (film thickness of 75 μm), were included in this study. The sensitivity, linearity, and linear range of these sorbent coatings were determined from calibration curves generated in pure CO2 and CO2 spiked with a known amount of air.

The storage capability under different storage conditions was examined for selected fibers, revealing that PIL-based sorbent coatings provided superior abilities in retaining CO2 compared to the commercial carboxen fiber. Although SPME is widely used for the determination of airborne organic contaminants, no report to our knowledge has demonstrated the application of SPME in the direct analysis of CO2.

6.2 Experimental

6.2.1 Materials

The synthesis of all IL monomers and polymers involved the use of vinyl imidazole,

131 1-bromohexane, 2,2′-azobis(isobutyronitrile) (AIBN), and taurine, which were purchased from Sigma-Aldrich (St. Louis, MO). Lithium bis[(trifluoromethyl)sulfonyl]imide was obtained from SynQuest Laboratories (Alachua, FL). Deuterated chloroform and dimethyl sulfoxide were obtained from Cambridge Isotope Laboratories (Andover, MA).

Deionized water (18.2 MΩ/cm) was obtained from a Milli-Q waterpurification system

(Millipore, Bedford, MA). Ethyl acetate, chloroform, 2-propanol, hexane, acetone, methanol, methylene chloride, and sodium hydroxide were obtained from Fisher

Scientific (Fairlawn, NJ). Propane and microflame brazing torches were purchased from

Sigma-Aldrich. Amberlite IRA-400(OH) anion-exchange resin was also obtained from

Sigma-Aldrich.

All laboratory-made SPME devices were constructed using a 5 μL syringe purchased from Hamilton (Reno, NV) and 0.05 mm i.d. fused-silica capillary obtained from Supelco (Bellefonte, PA). Commercial SPME fibers of PDMS (film thickness of 7

μm) and Carboxen-PDMS (film thickness of 75 μm) were obtained from Supelco. A fiber holder purchased from the same manufacturer was used for manual injection of the commercial fibers. All commercial fibers were conditioned in the inlet according to the instructions provided by the manufacturer. Gas sampling bulbs (250 mL) with thermogreen LB-1 cylindrical septa were obtained from Supelco and were used in all extraction studies. A manometer pressure/vacuum gauge (0 to ±30 psi), obtained from

Fisher Scientific, was used to record the pressure of the system.

6.2.2 Methods

The IL monomers were synthesized using a slight modification of a previously

132 published procedure [28], as shown in Figure 6-1. Briefly, 1-vinyl-3-hexylimidazolium bromide (VHIM-Br) was produced by mixing 1-vinylimidazole with an equimolar amount of 1-bromohexane in 2-propanol. The mixture was then allowed to react at 60 °C under constant stirring for 24 h. After removal of 2-propanol under vacuum, the product was dissolved in a small amount of Milli-Q water and then extracted with ethyl acetate five times to remove any unreacted starting materials. Ethyl acetate was then removed, and the product was collected and dried in a vacuum oven. The purity of the VHIM-Br was confirmed by 1H NMR before polymerization or metathesis anion exchange.

To obtain poly(VHIM-NTf2), polymerization of VHIM-Br was carried out by free radical polymerization following the procedures described elsewhere [30]. Briefly, 5.0 g of purified VHIM-Br was dissolved in 30 mL of chloroform. Then, 0.1 g (~2%) of the free radical initiator AIBN was introduced, and the solution refluxed for 3 h under N2 protection. Chloroform was then removed under vacuum, and the product was dried in a vacuum oven. The polymerization step was proven to be complete by the disappearance of the peaks that represent the vinyl group in the 1H NMR. The polymerization was repeated when necessary. The obtained poly(1-vinyl-3-hexylimidazolium) bromide was dissolved in Milli-Q water, and an equimolar amount of lithium bis[(trifluoromethyl)sulfonyl]imide was introduced to perform metathesis anion exchange.

This solution was stirred overnight, and the resulting PIL precipitate, poly(VHIMNTf2), was collected and washed with three aliquots of water and then dried under vacuum at

70 °C for 2 days.

The synthesis of poly(VHIM-taurate) involved exchanging the counteranion of

VHIM-Br to hydroxide by passing the monomer through a column packed with anion-

133

Figure 6-1: Schematic demonstrating the synthesis of the task-specific polymeric ionic liquids used for the selective capture of CO2.

134 exchange resin in the hydroxide ion form, following a method by Fukumoto et al. that was used to synthesize amino acid-based ILs [31]. Specifically, 100 mL of the regenerated anion-exchange resin was packed into a 50 cm × 2 cm column followed by flushing an excess amount of 5 M NaOH to ensure the resin was entirely in the hydroxide ion form. This was verified by adding silver nitrate to a collected fraction of the eluent.

White silver bromide precipitate formed when bromide ions persisted in the solution.

Nitric acid was used to avoid the potential interference of silver oxide, which is a dark precipitate and can be dissolved by introducing nitric acid. After regenerating the resin,

VHIM-Br was dissolved in water and passed through the anion-exchange column at an appropriate flow rate. The IL eluent was kept in aqueous solution due to its limited stability. An acid-base titration was used to determine the concentration of

1-vinyl-3-hexylimidazolium hydroxide (VHIM-OH) in the aqueous solution. The final step was a neutralization reaction between VHIM-OH and taurine. An equimolar amount of taurine was dissolved in water and added into an aqueous solution of VHIM-OH dropwise to avoid intense reaction. The reaction was stirred and allowed to stand overnight. Water was then removed by rotary evaporation, and the product dried under vacuum for 48 h. Polymerization was performed using the aforementioned conditions to yield poly(VHIM-taurate). 1H NMR spectra for VHIM-Br, VHIM-taurate, and poly(VHIM-taurate) are included in the Supporting Information.

Laboratory-made SPME devices were constructed following the previously published procedure by our group [28]. A microflame torch was used to seal the end of the capillary followed by removing the polyimide polymer from the last 1.0 cm segment of the fiber by a high-temperature flame. The fiber was then washed with methanol,

135 hexane, acetone, and dichloromethane, followed by a 10 min conditioning step in the GC injection port at 250 °C. Coating solutions were prepared by mixing the PIL in chloroform at a ratio of 9:1 (v/v). Binary PIL mixtures containing poly(VHIM-NTf2) and poly(VHIM-taurate) were mixed in chloroform at the desired weight percentage of each component. The coating solution was shaken for 5 min to ensure that the two PILs were homogeneously mixed. The pretreated SPME fiber was dipped into the PIL coating solution, held for approximately 20 s, and then removed from the coating solution. The coated fibers were allowed to dry in air for 10 min and then conditioned in the GC injection port for 10 min at 250 °C for the poly(VHIM-NTf2) coating and 180 °C for all other remaining PIL coatings. The film thickness of the PIL-based coatings was estimated by scanning electron microscopy.

All extractions were performed in a 250 mL gas sampling bulb at 14 °C. A schematic describing the extraction apparatus used in this study is shown in Figure 6-2.

With the regulator closed, valve 1 and valve 2 were opened, and the entire system evacuated via vacuum pump until the pressure reading from the pressure gauge was constant. Valve 1 was closed to isolate the system from the atmosphere and the initial pressure was recorded from the pressure gauge. The regulator was opened to introduce a desired amount of CO2 into the sample bulb. The reading from the pressure gauge was recorded as the final pressure when the pressure reached a constant value. Valve 2 was closed and the SPME syringe injected into the sample bulb through the septum, and the fiber exposed to the gas sample for a desired length of time. The fiber was withdrawn into the syringe, and the syringe was removed from the sampling bulb. The captured CO2 was released from the SPME fiber by high-temperature desorption in the GC injection port.

136

Figure 6-2: Apparatus of SPME setup used to perform extraction of CO2.

137 The obtained CO2 peak area was normalized by ΔP (final pressure - initial pressure).

Calibration curves were generated under two different conditions to perform a quantitative study of CO2 extraction efficiency. This involved exposing the SPME fiber to the sample bulb containing pure CO2 at a given pressure, as well as performing extractions in the sample bulb containing a fixed amount of air with varied concentrations of CO2. For the latter condition, vacuum was first applied to the entire extraction system, followed by introducing air into the sample bulb up to a pressure of 70 kPa. CO2 was subsequently injected into the sampling bulb to increase the total pressure (i.e., 75, 80, 85 kPa, etc.). In each case, the partial pressure of CO2 was known (i.e., 5, 10, 15 kPa, etc.).

All separations were conducted using an Agilent Technologies 6890N gas chromatograph (Agilent Technologies, Palo Alto, CA). The GC was equipped with thermal conductivity and flame ionization detectors coupled in series. All separations were performed using a Carboxen 1010 PLOT capillary column (30 m × 0.32 mm i.d.) purchased from Supelco. The following temperature program was used for the separation of CO2: initial temperature of 35 °C held for 10 min and then increased to 225 °C employing a ramp of 12 °C/min. Helium was used as the carrier gas with a flow rate of 1 mL/min. The inlet temperature was maintained at 250 °C for PDMS, Carboxen-PDMS, and poly(VHIM-NTf2) PIL fibers and 180 °C for the remaining fibers. An inlet desorption time of 2 min was used for all fibers. Splitless injection was used, and a purge flow to split vent of 20.0 mL/min at 0.10 min was applied. The thermal conductivity detector was held at 250 °C using a reference flow of 20.0 mL/min and a makeup flow of helium at 7.0 mL/min. Agilent Chemstation software was used for data acquisition.

138 6.3 Results and Discussion

6.3.1 Design of PIL-based sorbent coatings

In this study, two different PIL-based SPME fiber coatings were designed, synthesized, and evaluated for the purpose of selective capture of CO2. The polymeric imidazolium cation was chosen due to its high thermal stability as well as the tendency of these materials to form smooth and homogeneous coatings [28]. It has been shown previously that for unfunctionalized ILs, the nature of the counteranion exhibits a more dominant effect on overall CO2 solubility [32]. The poly(VHIM-NTf2) PIL was chosen as

− it contains the bis[(trifluoromethyl)sulfonyl]imide (NTf2 ) anion and ILs consisting of this anion often exhibit high thermal stability and higher CO2 solubility. Another PIL, poly(VHIM-taurate), was chosen as it contains a primary amine in its counteranion. The mechanism of CO2 capture by the poly-(VHIM-taurate) PIL coating is believed to be largely dominated by the reversible formation of a carbamate salt. Task-specific ILs of this class were first introduced by Davis and co-workers [33,34]. As shown in Figure 6-3,

CO2 can be extracted and sequestered in the sorbent coating through the reaction of CO2 with two amine groups of the PIL, followed by thermally desorbing CO2 in the GC injection port resulting in regeneration of the PIL. By carefully tailoring the structures of the PILs as well as incorporating various functional groups to the cation or anion, the selectivity of sorbent coatings for CO2 can be optimized.

Initial SPME studies utilizing these sorbent coatings revealed impressive extraction efficiencies of CO2 at various pressures. To investigate the extraction performance of these PILs, the morphology of the poly(VHIM-taurate) coating was examined by scann-

139

Figure 6-3: Schematic illustrating the reversible reactive capture of CO2 by the poly(VHIM-taurate) PIL. For simplicity, two imidazolium cations and taurate anions from the linear polymer backbone are represented.

140 ing electron microscopy (SEM). As shown in Figure 6-4A, the poly(VHIM-taurate) PIL produced an even and smooth surface when coated on the fused-silica support. After exposure of this PIL-coated fiber to CO2, the coating underwent a distinct morphology change presumably due to the dramatic increase in viscosity of the PIL (see Figure 6-4B).

Previous studies have also observed a pronounced increase in viscosity, and occasionally solidification, for similar classes of ILs [34,35]. When the complexed CO2 was desorbed from the fiber in the GC injector, the morphology of the sorbent coating returned to its original state, as shown in Figure 6-4C.

6.3.2 Sorption-time profiles

The effect of the extraction time on the amount of CO2 extracted by the fiber coating was investigated by plotting the CO2 peak area versus the extraction time for four different PIL-based fibers including: neat poly(VHIM-taurate), neat poly(VHIM-NTf2),

50% poly(VHIM-taurate)/50% poly(VHIM-NTf2) (w/w), and 15% poly(VHIMtaurate)/85% poly(VHIM-NTf2) (w/w). Sorption-time profiles were generated under two different CO2 pressures (i.e., 12 and 112 kPa), which represent two different

CO2 concentration levels. Triplicate extractions were performed at selected time intervals to examine the reproducibility of the fiber coatings.

As shown in Figure 6-5, parts A and B, sorbent coatings consisting of neat poly(VHIM-NTf2) and PIL binary mixtures reached equilibration in 20 min under both

CO2 pressures. However, the neat poly(VHIM-taurate) sorbent coating did not achieve equilibration until 30 min under a CO2 pressure of 112 kPa. Under low CO2 pressure (12 kPa), the poly(VHIM-NTf2) fiber exhibited higher extraction efficiency compared to the

141

Figure 6-4: Scanning electron micrographs of a SPME fiber coated with the poly(VHIM-taurate) PIL: (A) before exposure to CO2, (B) after exposure to CO2, and (C) after thermal desorption of CO2.

142

Figure 6-5: Sorption-time profiles under two different CO2 pressures, (A) CO2 Pressure =

12 kPa and (B) CO2 Pressure = 112 kPa, obtained for four PIL-based (~10 μm) and two commercial fiber coatings: (▲) carboxen (75 μm), (■) poly(VHIM-NTf2), (♦) poly(VHIM-NTf2) (85%)/poly(VHIM-taurate) (15%), (□) poly(VHIM-NTf2)

(50%)/poly(VHIM-taurate) (50%), (X) poly(VHIM-taurate), (+) PDMS (7 μm).

143 poly(VHIM-taurate) fiber and two binary PIL-coated fibers, which all exhibited similar extraction efficiencies. However, when the CO2 pressure was increased to 112 kPa, the extraction efficiency of the neat poly(VHIM-NTf2) PIL was distinctly higher than that of the 15% poly(VHIM-taurate)/85% poly(VHIM-NTf2) fiber, followed by the 50% poly(VHIM-taurate)/50% poly(VHIM-NTf2) coating and the neat poly(VHIM-taurate) coating.

For comparison purposes, sorption profiles of two commercial SPME sorbent coatings, namely, PDMS and Carboxen-PDMS (carboxen), were also generated. It can be observed that equilibration was reached within approximately 5 min for both coatings under the two different CO2 pressures. The PDMS fiber provided the lowest extraction efficiency among all of the evaluated coatings in this study. The carboxen fiber, which possessed a film thickness of 75 μm and a highly porous surface, exhibited the highest extraction efficiency under both CO2 pressures. In comparison to the carboxen fiber, the poly(VHIM-NTf2) fiber, with an approximate film thickness of 10 μm, produced an extraction efficiency approximately 45% of that of the carboxen fiber under a CO2 pressure of 12 kPa, whereas a comparable extraction efficiency was attained under a CO2 pressure of 112 kPa. In the case of the poly(VHIM-taurate) fiber, the extraction efficiency was nearly 22% of that of the carboxen fiber at 12 kPa and increased to nearly 45% when the CO2 pressure was increased to 112 kPa. It should be emphasized that both of the neat

PIL-coated fibers possessed film thicknesses of approximately 10 μm, resulting in a much smaller volume of the sorbent phase compared to that of the carboxen fiber. These results clearly indicate that comparable, if not higher, extraction efficiencies could be attained with these sorbent coatings if similar film thicknesses were employed.

144 6.3.3 Analytical performance

The sorbent coatings and the extraction method were evaluated by examining the analytical performance in terms of reproducibility and figures of merit of calibration curves including sensitivity, standard deviation of the regression, linear range, and correlation coefficients. The reproducibility was examined by studying six different coatings at their equilibration times under CO2 pressures of 12 and 112 kPa. As shown in

Table 6.1, relative standard deviation (RSD) values of the PIL-based fibers ranged from

3.6% to 7.2% and were significantly lower than those of the commercial fibers which ranged from 7.4% to 10.0%.

Calibration curves were obtained for the carboxen, poly(VHIMNTf2), and poly(VHIM-taurate) fibers at their equilibration times. Table 6.2 includes the figures of merit of the calibration curves generated in a sampling bulb containing pure CO2 with a pressure ranging from 1.5 to 125 kPa. The poly(VHIM-NTf2) fiber exhibited similar sensitivity to the carboxen fiber, despite the fact that the film thickness of the poly(VHIM-NTf2) coating is approximately 15% that of the carboxen coating. Both PIL fibers exhibited lower standard deviations of the regression line and wider linear ranges compared to the carboxen fiber.

To evaluate the capability of the sorbent coatings to selectively extract CO2 in a mixed gas matrix, a preliminary study was performed by carrying out extractions in CO2 spiked with 70 kPa of air at the equilibration time of the sorbent coatings. With a fixed pressure of air, an increased amount of CO2 was introduced into the sample bulb resulting in an increased total pressure. The mole fraction of CO2 ranged from 0.067 to 0.417. The figures of merit from the obtained calibration curves are listed in Table 6.3. As expected,

145

Table 6.1 Reproducibility of four task-specific PIL-based fibers and two commercial fibers.

Sorbent coating % RSD at low % RSD at high CO2 pressure CO2 pressure (12 kPa)a (112 kPa)a Poly(VHIM-taurate) 6.8 5.9

Poly(VHIM-NTf2) 5.3 6.5

Poly(VHIM-NTf2)(85%)/poly(VHIM-taurate)(15%) 3.6 4.5

Poly(VHIM-NTf2)(50%)/poly(VHIM-taurate)(50%) 7.2 4.7

Carboxen-PDMS 9.0 7.4

PDMS 8.7 10.0

a: Relative standard deviation was determined after performing eight (8) successive extractions at the equilibration time of the sorbent coatings.

146

Table 6.2 Figures of merit of calibration curves for two task-specific PIL-based fibers and

a one commercial fiber in pure CO2.

b Sorbent coating Slope ± error Syx Linear range (kPa) R

Carboxen 28.71 ± 1.7 119.3 1.5 - 75 0.990

Poly(VHIM-NTf2) 23.28 ± 0.4 54.63 1.5 - 125 0.999

Poly(VHIM-taurate) 12.30 ± 0.4 49.62 1.5 - 125 0.996

a : Extraction time: 5 min for the carboxen fiber, 20 min for the poly(VHIM-NTf2) fiber, and 30 min for the poly(VHIM-taurate) fiber. b: Standard deviation of the regression.

147

Table 6.3 Figures of merit of calibration curves for two task-specific PIL-based fibers and

a one commercial fiber in CO2 spiked with air (70 kPa).

b Sorbent coating Slope ± error Syx Linear range (kPa CO2) R

Carboxen 13.31 ± 0.3 15.70 5 - 50 0.997

Poly(VHIM-NTf2) 5.25 ± 0.2 6.87 15 - 50 0.995

Poly(VHIM-taurate) 4.76 ± 0.2 6.46 15 - 50 0.995

a : Extraction time: 5 min for the carboxen fiber, 20 min for the poly(VHIM-NTf2) fiber,

and 30 min for the poly(VHIM-taurate) fiber.

b: Standard deviation of the regression.

148 the sensitivity of all fibers for CO2 decreased in the mixed gas matrix. The poly(VHIM-taurate) fiber provided comparable sensitivity to the poly(VHIM-NTf2) fiber, whereas the carboxen fiber exhibited higher sensitivity and larger linear range than the

PIL-based fibers. The standard deviations of the regression and correlation coefficients were comparable for all fibers. Again, due to the fact that the PIL-based fibers possess film thicknesses of approximately 10 μm, it is expected that enhanced sensitivity would be achieved if a thicker PIL sorbent coating was used.

6.3.4 Sorbate storage capacity

The capability of sorbent coatings to retain extracted analyte(s) during storage is an important parameter to consider when designing new coating materials. To determine the storage capacity of the task-specific PIL-based coatings, extractions were carried out at the equilibration time for each fiber at room temperature. Three storage conditions were examined in this study. Following extraction, the SPME fiber was retracted into the syringe and left uncapped for 1 min. In addition, the syringe was capped with a thermored septum for 1 and 10 min following extraction. When the storage time had elapsed, the fiber was immediately desorbed in the GC injection port to determine the amount of CO2 that was retained in the coating.

As shown in Figure 6-6, a significant amount of CO2 was lost for all fibers when a sampled SPME fiber was withdrawn into the syringe needle and left uncapped at room temperature for 1 min. In comparison to direct desorption, an average of 54.7% of the captured CO2 was lost from the carboxen fiber under this storage condition. The task-specific PIL-based fibers exhibited an enhanced ability to retain the extracted CO2 as

149

Figure 6-6: Comparison of the amount of CO2 sorbate retained in two PIL-based fibers and the carboxen fiber under various storage conditions.

150 25.9% of the captured CO2 was lost from the poly(VHIM-NTf2) fiber and 30.6% of the

CO2 was lost from the poly(VHIM-taurate)fiber. Using a septum to seal the tip of the syringe needle for 1 min prior to desorption appears to have imparted improved stability of CO2 within the sorbent coating. Approximately 21.2% of the captured CO2 was lost from the carboxen fiber after capping for 1 min, whereas only 12.4% and 2.2% of the captured CO2 was lost under the same condition for the poly(VHIM-NTf2) and poly(VHIM-taurate) fibers, respectively. When a sampled carboxen fiber was sealed for

10 min after CO2 extraction, 43.1% of the captured CO2 escaped from the fiber coating.

For the PIL-based sorbent coatings, 34.7% of captured CO2 was lost from the poly(VHIM-NTf2) fiber, whereas 33.0% was lost from the poly(VHIM-taurate) fiber.

These results indicate that the task-specific PIL coatings exhibit distinct advantages in their storage capabilities compared to the carboxen coating. Despite its lower extraction efficiency compared to the poly(VHIM-NTf2) fiber, the poly(VHIM-taurate) fiber was capable of retaining more CO2 after short periods of storage. This is likely due to the fact that the CO2 is chemically captured by the sorbent coating, whereas the poly(VHIM-NTf2) and carboxen coatings retain the CO2 via physical sorption.

6.4 Conclusions

Polymeric ionic liquids have proven to be a useful class of sorbent coatings for

SPME due to their unique physicochemical properties. For the first time, the structures of two PILs were carefully tailored to incorporate specific functional groups to serve as task-specific PIL sorbent coatings that exhibit high extraction efficiencies and

151 selectivities for CO2. The PIL sorbent coatings are capable of undergoing two different types of mechanisms responsible for capturing the CO2 sorbate, namely, physical sorption by the poly(VHIM-NTf2) coating and carbamate formation by the poly(VHIM-taurate) coating. The PIL-based coatings exhibited impressive extraction efficiencies of CO2 as well as exceptional extraction-to-extraction reproducibility. In comparison to the commercial carboxen SPME fiber, similar extraction efficiency of CO2 was achieved using the poly(VHIM-NTf2) PIL fiber at high CO2 pressure, despite the fact that the carboxen fiber possessed a much larger film thickness. The poly(VHIM-taurate) sorbent coating exhibited enhanced storage capacity of CO2 on the fiber compared to the poly(VHIM-NTf2) and carboxen coatings, most likely due to the ability of the task-specific PIL-based sorbent coating to chemically react and sequester CO2 within the coating.

Future studies involving these sorbent coatings will focus on examining the effect of humidity and temperature on the extraction efficiency of CO2. To further investigate the advantages of employing task-specific PILs that capture CO2 by two different mechanisms, the performance of the coatings will be examined in a mixed gas matrix.

The results from this work indicate the promise that PIL-based SPME coatings possess in developing task-specific microextractions that result in high extraction efficiencies, high selectivities, and low detection limits.

Acknowledgements

J.L.A. acknowledges funding from the Analytical and Surface Chemistry Program

152 in the Division of Chemistry and the Separation and Purification Processes Program in the Chemical, Environmental, Bioengineering, and Transport Systems Division from the

National Science Foundation for a CAREER grant (CHE-0748612)

References

1. DOE/EIA. International Energy Annual 2005; Energy Information Administration:

Washington, DC, 2005.

2. Arthur, C. L.; Pawliszyn, J. Anal. Chem. 1990, 62, 2145-2148.

3. Arthur, C. L.; Killam, L. M.; Motlagh, S.; Lim, M.; Potter, D. W.; Pawliszyn, J.

Environ. Sci. Technol. 1992, 26, 979-983.

4. Arthur, C. L.; Killam, L. M.; Buchholz, K. D.; Pawliszyn, J.; Berg, J. R. Anal. Chem.

1992, 64, 1960-1966.

5. Chai, M.; Arthur, C. L.; Pawliszyn, J. Analyst 1993, 118, 1501-1505.

6. Chai, M.; Pawliszyn, J. Environ. Sci. Techno. 1995, 29, 693-701.

7. Martos, P. A.; Pawliszyn, J. Anal. Chem. 1998, 70, 2311-2320.

8. Koziel, J.; Jia, M.; Khaled, A.; Noah, J.; Pawliszyn, J. Anal. Chim. Acta. 1999, 400,

153-162.

9. Odziemkowski, M.; Koziel, J. A.; Irish, D. E.; Pawliszyn, J. Anal. Chem. 2001, 73,

3131-3139.

10. Martos, P. A.; Pawliszyn, J. Anal. Chem. 1997, 69, 206-215.

11. Martos, P. A.; Saraullo, A., Pawliszyn J. Anal. Chem. 1997, 69, 402-408.

12. Chien, Y.-C.; Uang, S.-N.; Kuo, C.-T.; Shih, T.-S.; Jen, J.-F. Anal. Chim. Acta. 2000,

153 419, 73-79.

13. Smith, P. A.; Sheely, M. V.; Kluchinsky Jr., T. A. J. Sep. Sci. 2002, 25, 917-921.

14. Martos, P. A.; Pawliszyn, J. Anal. Chem. 1999, 71, 1513-1520.

15. Bartelt, R. J.; Zilkowski, B. W. Anal. Chem. 1999, 71, 92-101.

16. Gdreckit, T.; Pawliszyn, J. Anal. Chem. 1995, 67, 3265-3274.

17. Koziel, J.A.; Odziemkowski, M.; Pawliszyn, J. Anal. Chem. 2001, 73, 47-54.

18. Augusto, F.; Koziel, J., Pawliszyn, J. Anal. Chem. 2001, 73, 481-486.

19. Wasserscheid, P.; Welton, T. Ionic Liquids in Synthesis; Wiley-VCH: New York,

2003.

20. Davis, J. H., Jr. Chem. Lett. 2004, 9, 1072.

21. Liu, J. F.; Jiang, G. B.; Chi, Y. G.; Cai, Y. Q.; Zhou, Q. X.; Hu, J. T. Anal. Chem.

2003, 75, 5870-5876.

22. Yao, C.; Pitner, W. R.; Anderson, J. L. Anal. Chem. 2009, 81, 5054-5063.

23. Yao, C.; Anderson, J. L. Anal. Bioanal. Chem. 2009, 395, 1491-1502.

24. Liu, J.-F.; Li, N; Jiang, G.-B.; Liu, J.-M.; Jönsson, J. Å.; Wen, M.-J. J. Chromatogr. A

2005, 1066, 27-32.

25. Hsieh, Y.-N.; Huang, P.-C.; Sun, I.-W.; Whang, T.-J.; Hsu, C.-Y.; Huang, H.-H.;

Kuei, C.-H. Anal. Chim. Acta. 2006, 557, 321-328.

26. Huang, K.-P.; Wang, G.-R.; Huang, B.-Y.; Liu, C.-Y. Anal. Chim. Acta. 2009, 645,

42-47.

27. He, Y.; Pohl, J.; Engel, R.; Rothman, L.; Thomas, M. J. Chromatogr. A 2009, 1216,

4824-4830.

28. Zhao, F.; Meng, Y.; Anderson, J. L. J. Chromatogr. A 2008, 1208, 1-9.

154 29. Meng, Y., Pino, V., Anderson, J. L. Anal. Chem. 2009, 81, 7107-7112.

30. Marcilla, R.; Blazquez, J. A.; Rodriguez, J.; Pomposo, J. A.; Mecerreyes, D. J. Polym.

Sci. Part A: Polym. Chem. 2004, 42, 208-212.

31. Fukumoto, K.; Yoshizawa, M.; Ohno, H. J. Am. Chem. Soc. 2005, 127, 2398-2399.

32. Cadena, C.; Anthony, J. L.; Shah, J. K.; Morrow, T. I.; Brennecke, J. F.; Maginn, E. J.

J. Am. Chem. Soc. 2004, 126, 5300-5308.

33. Bates, E. D.; Mayton, R. D.; Ntai, I.; Davis, J. H., Jr. J. Am. Chem. Soc. 2002, 124,

926-927.

34. Soutullo, M. D.; Odom, C. I.; Wicker, B. F.; Henderson, C. N.; Stenson, A. C.; Davis,

J. H, Jr. Chem. Mater. 2007, 19, 3581-3583.

35. Gutowski, K. E.; Maginn, E. J. J. Am. Chem. Soc. 2008, 130, 14690-14704.

155 Chapter 7

Selective Extraction of CO2 from Simulated Flue Gas Using

Polymeric Ionic Liquid Sorbent Coatings in Solid-Phase

Microextraction

A paper published in Journal of Chromatography A 1

Qichao Zhao, Jared L. Anderson

Abstract

The CO2 selectivity of two polymeric task-specific ionic liquid sorbent coatings, poly(1-vinyl-3-hexylimidazolium) bis[(trifluoromethyl)sulfonyl]imide [poly(VHIM-

NTf2)] and poly(1-vinyl-3-hexylimidazolium) taurate [poly(VHIM-taurate)], was examined using solid-phase microextraction (SPME) for the determination of CO2 in simulated flue gas. For comparison purposes, a commercial SPME fiber, CarboxenTM-

2010 Elsevier

156 PDMS, was also studied. A study into the effect of humidity revealed that the poly(VHIM-taurate) fiber exhibited enhanced resistance to water, presumably due to the unique mechanism of CO2 capture. The effect of temperature on the performance of the

PIL-based and Carboxen fibers was examined by generating calibration curves under various temperatures. The sensitivity, linearity, and linear range of the three fibers were evaluated. The extraction of CH4 and N2 was performed and the selectivities of the

PIL-based and Carboxen fibers were compared. The poly(VHIM-NTf2) fiber was found to possess superior CO2/CH4 and CO2/N2 selectivities compared to the Carboxen fiber, despite the smaller film thicknesses of the PIL-based fibers. A scanning electron microscopy study suggests that the amine group of the poly(VHIM-taurate) is capable of selectively reacting with CO2 but not CH4 or N2, resulting in a significant surface morphology change of the sorbent coating.

7.1 Introduction

Solid-phase microextraction (SPME) is a widely used extraction technique that was introduced by Pawliszyn and co-workers [1–3] in the early 1990s. Compared to conventional sample preparation methods, SPME provides many advantages including its speed, simplicity, and ease of operation. In addition, SPME combines sampling and sample preparation into one step and can be readily automated and coupled with separation techniques such as gas chromatography (GC). SPME consists of a fiber support coated with a sorbent material that is exposed to the sample matrix by either headspace or direct-immersion sampling. Equilibration is subsequently established

157 between the analytes and the fiber coating. When coupled with GC, desorption is accomplished by exposing the sampled sorbent coating to the injection port, resulting in thermal desorption of the analytes from the sorbent coating.

The nature of the sorbent coating can be considered the heart of SPME. Recently, new SPME sorbent coating materials based on ionic liquids (ILs) and polymeric ionic liquids (PILs) have attracted increased attention [4–11]. ILs are a class of non-molecular ionic compounds that possess melting points below 100 ºC [12]. Compared to traditional organic solvents, ILs have a number of significant advantages including their low vapor pressure at ambient temperature and high thermal stability. The combination of cations and anions can be altered to custom design the physical and chemical properties of the resulting IL. In addition, functional groups can be introduced into either component to form task-specific ionic liquids (TSILs), which are capable of interacting with analytes or substrates in specific ways [13–15].

Our group first reported the utilization of PILs as SPME sorbent coatings. These coatings exhibit enhanced thermal stability, high reproducibility, and extended fiber lifetimes [9,10]. More recently, we reported the first example of SPME sorbent coatings based on a task-specific PIL capable of selectively extracting CO2 [11]. By tailoring the structures of the PILs, the resulting sorbent coatings exhibited different mechanisms of

CO2 capture and sequestration. Compared to the commercial SPME fibers with much larger film thicknesses, the PIL-based fibers exhibited comparable extraction efficiency, wider calibration range, and enhanced sorbate storage capacity.

It is well-known that the extraction conditions affect the analyte extraction efficiency and often the equilibration time in SPME [16]. Previously, Martos and Pawliszyn [17]

158 utilized the polydimethylsiloxane (PDMS) fiber for the extraction of hydrocarbons in the gas phase, and the effect of relative humidity, temperature, gas velocities, and gas concentrations on the extraction efficiency was studied. The encouraging CO2 selectivity exhibited by the PIL-based sorbent coatings may make them particularly advantageous in the analysis of multicomponent gas streams. Typical flue gas produced by coal-fired power plants consists of a number of gas species including N2, CO2, H2O, and O2 with varied concentration, depending on the location of sampling [18]. The temperature of flue gas often varies depending on the requirements of the scrubbing system employed. Hence, it is important to identify the sorbent coatings that exhibit high CO2/N2 selectivity in the presence of water vapor while maintaining performance at various temperatures.

In this work, two different task-specific PIL-based coatings, namely poly(1-vinyl-3-hexylimidazolium) bis[(trifluoromethyl)sulfonyl]imide [poly(VHIM-

NTf2)] and poly(1-vinyl-3-hexylimidazolium) taurate [poly(VHIM-taurate)], were employed for the extraction of CO2. The structures of the studied PILs are shown in Fig.

7-1. For comparison purposes, a commercial CarboxenTM-PDMS (Carboxen) SPME coating was included in this study. Calibration curves of CO2 saturated with water as well as under various temperatures were generated to study the effect of humidity and temperature on the performance of these fibers. The studied fibers were also used to extract CH4 and N2, and the CO2/CH4 and CO2/N2 selectivities were determined and compared. To further investigate the surface morphology change of the poly(VHIM-taurate) coating, scanning electron microscopy (SEM) was employed to examine the coating morphology after exposure to CH4 and N2.

159

Figure 7-1: Structures of the PILs used as SPME fiber coatings in this study.

160

7.2 Experimental

7.2.1 Materials

Vinyl imidazole, 1-bromohexane, 2,2′-azo-bis(isobutyronitrile) (AIBN) and taurine were purchased from Sigma–Aldrich (St. Louis, MO, USA). Lithium bis[(trifluoromethyl)sulfonyl]imide was obtained from SynQuest Labs (Alachua, FL,

USA). Deuterated chloroform and dimethylsulfoxide were obtained from Cambridge

Isotope Laboratories (Andover, MA, USA). Deionized water (18.2MΩ/cm) was obtained from a Milli-Q water-purification system (Millipore, Bedford, MA, USA). Ethyl acetate, chloroform, 2-propanol, hexane, acetone, methanol, methylene chloride, and sodium hydroxide were obtained from Fisher Scientific (Fairlawn, NJ, USA). Propane and microflame brazing torches were purchased from Sigma–Aldrich. Amberlite

IRA-400(OH) ion-exchange resin was obtained from Sigma–Aldrich.

All laboratory-made SPME devices were constructed using a 5-μL syringe purchased from Hamilton (Reno, NV, USA) and 0.05-mm I.D. fused silica capillary obtained from Supelco (Bellefonte, PA, USA). Commercial SPME fibers of

CarboxenTM-PDMS (film thickness of 75 μm) were obtained from Supelco. A fiber holder purchased from the same manufacturer was used for manual injection of the commercial fibers. Gas sampling bulbs (250 mL) with thermogreenTM LB-1 cylindrical septa were obtained from Supelco and used to perform CO2 extraction. A pressure gauge

(0 psi to ±30 psi), obtained from Fisher Scientific, was used to record the pressure.

161

7.2.2 Methods

The monomeric and polymeric ILs were synthesized following previously reported procedures [11]. Briefly, 1-vinylimidazole was dissolved in 2-propanol and reacted with an equimolar amount of 1-bromohexane under reflux and constant stirring conditions to generate 1-vinyl-3-hexylimidazolium bromide (VHIM-Br). After extraction with ethyl acetate, polymerization was performed in the presence of the free radical initiator AIBN to generate poly(1-vinyl-3-hexylimidazolium) bromide. The polymerization step was repeated, if necessary, until the peaks belonging to the vinyl group in 1H NMR disappeared. To perform metathesis anion-exchange, poly(1-vinyl-3-hexylimidazolium) bromide was dissolved in Milli-Q water and an equimolar amount of lithium bis[(trifluoromethyl)sulfonyl]imide was introduced to the aqueous solution. The resulting precipitate was collected, washed with water to remove any residual halide anion, and dried overnight under vacuum to yield poly(VHIM-NTf2). To generate poly(VHIM-taurate), purified VHIM-Br was dissolved in water and passed through an anion-exchange column at an appropriate flow rate to exchange the counter anion. The IL eluent, 1-vinyl-3-hexylimidazolium hydroxide (VHIM-OH), was kept in the aqueous solution due to its limited stability. An equimolar amount of taurine was dissolved in water and added to an aqueous solution of VHIM-OH dropwise to avoid intense reaction.

The reaction was stirred and allowed to stand overnight. Water was then removed by rotary evaporation and the product dried under vacuum for 48 h. Polymerization was performed using the aforementioned conditions to yield poly(VHIM-taurate). The purity of VHIM-taurate and poly(VHIM-taurate) were characterized using 1H NMR and are reported in our previous study [11].

162

Laboratory-made SPME devices were constructed following the previously published procedure by our group [9]. Briefly, the end of the capillary was sealed using a microflame torch and the polyimide polymer from the last 1.0 cm segment of the fiber was removed by a high temperature flame. The fiber was then washed with methanol, hexane, acetone, and dichloromethane, followed by a 10-min conditioning step in the GC injection port at 250 ºC. Coating solutions were prepared by mixing the PIL in chloroform at a ratio of 9:1 (v/v). The pre-treated SPME fiber was dipped into the PIL coating solution, held for approximately 20 s, and then removed from the coating solution. The coated fibers were allowed to dry in air for 10 min followed by a 10-min condition in the

GC injection port at 250 ºC for the poly(VHIMNTf2) coating and 180 ºC for the poly(VHIM-taurate) coatings. The film thickness of the PIL-based coatings was estimated by SEM.

Schematics describing the extraction apparatus used in this study are shown in Fig.

7-2. The regular operating process was described elsewhere [11]. To study the effect of humidity on the performance of the PIL-based and commercial fibers, approximately 50 mL of water was introduced to the trap (a 500 mL Erlenmeyer flask), as shown in Fig.

7-2A. With both valves open and the gas cylinder regulator closed, the entire system was evacuated via the vacuum pump. When the pressure reading from the pressure gauge was constant, valve 1 was closed to isolate the system from the atmosphere and the entire system was then saturated with water vapor. The initial pressure was recorded when the pressure reading from the pressure gauge was constant. A desired amount of CO2 was injected into the sample bulb and the SPME fiber exposed to the sample bulb containing

CO2 saturated with water vapor at room temperature (15 ºC).

163

Figure 7-2: Apparatus of SPME setup used in this study to examine (A) humidity effect and (B) temperature effect.

164

To study the effect of temperature on the performance of the PIL-based and commercial fibers, a slight modification was made to the gas sampling bulb, as illustrated in Fig. 7-2B. The sampling bulb was modified so that the valves were higher than the body of the bulb by lengthening the connecting glass tubing. The modified sampling bulb was then placed in a water bath allowing both valves and the septum inlet to lie above the water. An ice water and thermostatted warm water bath were used to achieve and maintain the desired temperatures (i.e., 0 ºC, 30 ºC, 40 ºC, and 65 ºC). Calibration curves were then generated by exposing the SPME fiber in the modified sample bulb containing dry CO2 at varied temperatures.

All separations were conducted using an Agilent Technologies 6890N gas chromatograph (Agilent Technologies, Palo Alto, CA, USA). The GC was equipped with thermal conductivity and flame ionization detectors coupled in series. All separations were performed using a CarboxenTM 1010 PLOT capillary column (30m×0.32mm I.D.) purchased from Supelco. The following temperature program was used for the separation of CO2: initial temperature of 35 ºC held for 10 min and then increased to 225 ºC employing a ramp of 12 ºC/min. Helium was used as the carrier gas with a flow rate of 1 mL/min. The inlet temperature was maintained at 250 ºC for the Carboxen and poly(VHIM-NTf2) PIL fibers and 180 ºC for the poly(HVIM-taurate) fiber. An inlet desorption time of 2 min was used for all fibers. Splitless injection was used and a purge flow to split vent of 20.0 mL/min at 0.10 min was applied. The thermal conductivity detector was held at 250 ºC using a reference flow of 20.0 mL/min and a make-up flow of helium at 7.0 mL/min. Agilent Chemstation software was used for data acquisition.

165

7.3 Results and Discussion

7.3.1 Effects of water on the extraction of CO2

Humidity is known to have notable effect on the performance of SPME fiber coatings. Previously, an approximate 10% decrease in the extraction efficiency was observed when PDMS fibers were used to extract hydrocarbons and volatile organic compounds (VOCs) under humid conditions [17,19]. It was rationalized that the water vapor partitioned into the SPME fiber coating and interfered with the mass uptake of the analytes.

In this study, calibration curves were generated for the Carboxen, poly(VHIM-NTf2) and poly(HVIM-taurate) fibers by extracting CO2 saturated with water vapor. The obtained figures of merit for these calibration curves are listed in Table 7.1. For comparison purposes, the figures of merit that were recently reported by our group for the calibration curves generated in dry CO2 are also included. It can be seen that the presence of water vapor significantly decreased the extraction efficiency of CO2 for all fibers. The sensitivity of the Carboxen fiber experienced a dramatic 75% decrease (from 28.71 to

7.05) due to the presence of water vapor. For the PIL-based coatings, a 40% decrease

(from 23.28 to 14.11) in sensitivity was observed when the poly(VHIM-NTf2) fiber was used to extract CO2 saturated with water vapor. The sensitivity of the poly(HVIM-taurate) experienced a 28% drop (from 12.30 to 8.90) when extracting CO2 saturated with water vapor. Both PIL-based coatings exhibited similar linearity to the Carboxen fiber. These results clearly indicate that the PIL fibers, particularly the poly(HVIM-taurate) fiber, showed enhanced resistance to the reduction of CO2 extraction sensitivity under humid

166

Table 7.1. Figures of merit of calibration curves for two task-specific PIL-based fibers and one commercial fiber in dry CO2 and CO2 saturated with water vapor.

a Condition Sorbent coating Slope ± Syx Linear range R

error (kPa)

b,c Dry CO2 Carboxen 28.71 ± 1.7 119.3 1.5 - 75 0.990

(75 μm)

Poly(VHIM-NTf2) 23.28 ± 0.4 54.63 1.5 - 125 0.999

(10 μm)

Poly(VHIM-taurate) 12.30 ± 0.4 49.62 1.5 - 125 0.996

(10 μm)

CO2 saturated Carboxen 7.05 ± 0.2 21.10 1.5 - 87.5 0.996 with waterc (75 μm)

Poly(VHIM-NTf2) 14.11 ± 0.5 65.05 1.5 - 125 0.995

(10 μm)

Poly(VHIM-taurate) 8.90 ± 0.2 21.76 1.5 - 125 0.998

(10 μm)

a: Standard deviation of the regression. b: Data obtained from reference [11] c : Extraction time: 5 min for the Carboxen fiber, 20 min for the poly(VHIM-NTf2) fiber, and 30 min for the poly(VHIM-taurate) fiber.

167 conditions compared to the Carboxen fiber. It is expected that the PIL fibers would provide further improved performance if the coating film thickness was increased.

7.3.2. Effects of temperature on the extraction of CO2

It is well documented that the change of temperature can affect both the analyte equilibration time and the extraction efficiency in SPME. Increased temperature enhances the mass transfer process and consequently shortens the equilibration time. In headspace

SPME, heating the sample to an elevated temperature results in an increased equilibrium concentration of the analyte(s) in the headspace. On the other hand, the sorption of analytes by the fiber coating decreases with increasing temperature due to a decrease in the analyte–sorbent coating partition coefficient. It has been reported that the amount of

VOCs extracted by the PDMS fiber increases approximately 20% when the temperature is dropped by 10 ºC [17].

In this study, the effect of temperature on the extraction performance of the

PIL-based and commercial Carboxen fibers was examined by generating the calibration curves at varied temperatures, namely 0 ºC, 30 ºC, 40 ºC, and 65 ºC. It was found that the

PIL fibers required considerably longer time to reach equilibrium when the extractions were performed at 0 ºC. All extractions performed at 15 ºC, 30 ºC, 40 ºC, and 65 ºC utilized the same extraction time for each fiber. As shown in Fig. 7-3, the amount of dry

CO2 extracted by the poly(HVIM-taurate) coating decreased with increasing temperature.

The calibration curves of the poly(HVIM-taurate) coating exhibited great linearity under all studied temperatures when the CO2 pressure was varied 1.5–125 kPa. Triplicate extractions were performed at selected CO2 pressures to examine the reproducibility of

168

Figure 7-3: Calibration curves obtained for the poly(VHIM-taurate) fiber coating under varied temperatures: (◊) 0 ºC, (■) 15 ºC, (Δ) 30 ºC, (♦) 40 ºC, (х) 65 ºC. Extraction time:

30 min.

169 the fiber coatings and generally produced relative standard deviations ranging from 1.4% to 4.4%.

Table 7.2 lists the figures of merit for calibration curves generated at 0 ºC, 15 ºC

(data obtained from Ref. [11]), 30 ºC, 40 ºC, and 65 ºC. Overall, the sensitivities of all three fibers decreased smoothly with increasing temperature. Interestingly, the exceptions occur for the poly(VHIM-NTf2) and Carboxen fibers which exhibited slightly increased sensitivities as the temperature was varied from 0 ºC to 15 ºC. Under all the examined temperatures, the 75-μm Carboxen fiber provided the highest sensitivity, while both

PIL-based fibers exhibited wider linear ranges compared to the Carboxen fiber. It should be emphasized that the PIL-based fibers possess film thicknesses of approximately 10 μm, which is much smaller compared to that of the Carboxen fiber. It can be expected that higher sensitivities would be obtained if a larger volume of the PIL-based sorbent coating was used.

7.3.3 Selectivity of the PIL-based and Carboxen fibers for CH4 and N2

In our previous study, it was shown that the poly(HVIM-taurate) sorbent coating undergoes a distinct change in morphology when exposed to CO2. The smooth and even surface of the coating can be recovered by exposing the sampled poly(HVIM-taurate) fiber to the high temperature of GC injection port. It was proposed that the amine groups within the taurate anion allowed for the reversible capture of CO2 within the sorbent coating, thereby significantly increasing the viscosity of the coating and imparting a morphological change to the coating surface.

170

Table 7.2. Figures of merit of calibration curves for two task-specific PIL-based fibers and one commercial fiber in pure CO2 at various temperatures.

a Extraction Sorbent coating Slope ± Syx Linear range R

Temperature error (kPa)

0 ºCb Carboxen 25.30 ± 2.1 186.1 1 - 87.5 0.976

(75 μm)

Poly(VHIM-NTf2) 19.58 ± 0.8 109.5 1 - 125 0.992

(10 μm)

Poly(VHIM-taurate) 12.40 ± 0.5 75.58 1 - 125 0.991

(10 μm)

15 ºCc,d Carboxen 28.71 ± 1.7 119.3 1.5 - 75 0.990

Poly(VHIM-NTf2) 23.28 ± 0.4 54.63 1.5 - 125 0.999

Poly(VHIM-taurate) 12.30 ± 0.4 49.62 1.5 - 125 0.996

30 ºCd Carboxen 22.36 ± 1.0 94.72 1.5 – 87.5 0.992

Poly(VHIM-NTf2) 13.67 ± 0.6 88.91 1.5 - 125 0.990

Poly(VHIM-taurate) 8.34 ± 0.3 42.56 1.5 - 125 0.994

40 ºCd Carboxen 19.75 ± 0.5 44.70 1.5 - 87.5 0.998

Poly(VHIM-NTf2) 10.35 ± 0.3 21.92 1.5 - 125 0.996

Poly(VHIM-taurate) 6.31 ± 0.2 13.70 1.5 - 125 0.995

171

Table 7.2 (continued)

a Extraction Sorbent coating Slope ± Syx Linear range R

Temperature error (kPa)

65 ºCd Carboxen 14.60 ± 0.3 30.44 1.5 - 100 0.998

Poly(VHIM-NTf2) 7.16 ± 0.3 21.92 1.5 - 125 0.990

Poly(VHIM-taurate) 4.85 ± 0.2 25.69 1.5 - 125 0.993

a: Standard deviation of the regression. b : Extraction time: 5 min for the Carboxen fiber, 50 min for the poly(VHIM-NTf2) and poly(VHIM-taurate) fibers. c: Data obtained from reference [11] d : Extraction time: 5 min for the Carboxen fiber, 20 min for the poly(VHIM-NTf2) fiber, and 30 min for the poly(VHIM-taurate) fiber.

172

To further explore the unique selectivity offered by the poly(HVIM-taurate) coating, three fused silica fibers were coated with poly(HVIM-taurate) and used to sample CO2,

CH4 and N2. The morphology of the coating surface was then studied by SEM for each gas/fiber pair. As shown in our previous work [11] and in Fig. 7-4A, the exposure of the poly(HVIM-taurate) coating to CO2 results in a wrinkled surface. However, no noticeable morphology change is observed when the poly(HVIM-taurate) coating was exposed to both CH4 and N2 (see Fig. 7-4B and C). In fact, these two surfaces are nearly identical.

These observations further support the argument that the amine groups of the poly(HVIM-taurate) PIL are largely responsible in imparting the CO2 selectivity to this sorbent coating.

To evaluate the selectivity of the PIL-based fibers and the Carboxen fiber for the extraction of CH4 and N2, extractions were performed at 112 kPa of pure gas. The obtained peak areas were plotted versus the extraction time, as shown in Fig. 7-5A and B.

The reproducibility was examined by triplicate extractions at selected time intervals. In the case of CH4 and N2, equilibration was reached for the Carboxen fiber within 5 min while the poly(HVIM-taurate) and poly(VHIM-NTf2) fibers required approximately 10 min and 15 min, respectively. The Carboxen fiber exhibited the highest extraction efficiency for CH4, which was approximately twice that of the PIL-based fibers. The extraction efficiencies for N2 were similar for all three examined fibers.

To further discern the three fibers evaluated in this study, the CO2/CH4 and CO2/N2 selectivities were determined. The selectivity was calculated using the peak area at the equilibration time for a particular gas and its relative thermal response [20]. Table 7.3 includes the selectivity of CO2 versus CH4 and N2 for the PIL-based and Carboxen

173

Figure 7-4: Scanning electron micrographs of a fiber coated with the poly(VHIM- taurate) PIL after exposure to (A) CO2, (B) CH4, and (C) N2.

174

Figure 7-5: Sorption-time profiles of (A) CH4 (pressure = 112 kPa) and (B) N2 (pressure

= 112 kPa), obtained for two PIL fibers and one commercial fiber at room temperature

(15 ºC): (♦) Carboxen (75 μm), (▲) poly(VHIM-NTf2) (10 μm), (■) poly(VHIM-taurate)

(10 μm).

175

Table 7.3. Gas pair selectivity for two task-specific PIL-based fibers and one commercial fiber.

Sorbent coating CO2/CH4 CO2/N2

Carboxen (75 μm) 3.27 7.53

Poly(VHIM-NTf2) (10 μm) 6.44 8.03

Poly(VHIM-taurate) (10 μm) 2.95 3.34

176 fibers. It can be seen that the three fibers tended to selectively extract CO2 relative to CH4 and N2. The poly(VHIM-NTf2) fiber provided the highest CO2/CH4 (twice that of the

Carboxen fiber) and CO2/N2 selectivities. The poly(HVIM-taurate) fiber, on the other hand, exhibited comparable CO2/CH4 selectivity to the Carboxen fiber. However, the

CO2/N2 selectivity of the poly(HVIM-taurate) fiber was lower than that of the Carboxen fiber.

7.4 Conclusions

Polymeric ionic liquids (PIL) are a promising class of sorbent coatings for SPME, particularly in gas analysis. In this study, two different PIL-based SPME sorbent coatings and commercial Carboxen fiber were examined to understand the effect of humidity and temperature on CO2 extraction. It was observed that the poly(VHIM-taurate) coating exhibited the lowest sensitivity drop in the presence of water vapor, while the sensitivity of the Carboxen fiber dropped substantially. As expected, all three fibers exhibited lower

CO2 extraction efficiencies with increasing extraction temperature. The Carboxen fiber provided the highest sensitivity under all studied temperatures while the PIL-based fibers were superior in terms of calibration curve linearity and linear range. The poly(VHIM-NTf2) sorbent coating was found to possess higher CO2/CH4 and CO2/N2 selectivities compared to the Carboxen fiber. Scanning electron micrographs of the poly(VHIM-taurate) coating revealed that the sorbent coating only undergoes a distinct change in morphology upon exposure to CO2. This thorough study provides an examination into the performance of these sorbent coatings under real-world sampling

177 conditions, particularly for a simulated flue gas matrix.

Acknowledgements

The authors wish to thank Dr. Pannee Burckel for her assistance in using the SEM, and Mr. Steven Moder for his assistance in producing the modified gas sampling bulb.

J.L.A. acknowledges funding from the Analytical and Surface Chemistry Program in the

Division of Chemistry and the Separation and Purification Processes Program in the

Chemical, Environmental, Bioengineering, and Transport Systems Division from the

National Science Foundation for a CAREER grant (CHE-0748612)

References

1. C.L. Arthur, J. Pawliszyn, Anal. Chem. 62 (1990) 2145.

2. C.L. Arthur, L.M. Killam, S. Motlagh, M. Lim, D.W. Potter, J. Pawliszyn, Environ. Sci.

Technol. 26 (1992) 979.

3. C.L. Arthur, L.M. Killam, K.D. Buchholz, J. Pawliszyn, J.R. Berg, Anal. Chem. 64

(1992) 1960.

4. J.-F. Liu, N. Li, G.-B. Jiang, J.-M. Liu, J. Å. Jo¨nsson, M.-J. Wen, J. Chromatogr. A

1066 (2005) 27.

5. Y.-N. Hsieh, P.-C. Huang, I.-W. Sun, T.-J. Whang, C.-Y. Hsu, H.-H. Huang, C.-H.

178

Kuei, Anal. Chim. Acta. 557 (2006) 321.

6. K.-P. Huang, G.-R. Wang, B.-Y. Huang, C.-Y. Liu, Anal. Chim. Acta. 645 (2009) 42.

7. Y. He, J. Pohl, R. Engel, L. Rothman, M. Thomas, J. Chromatogr. A 1216 (2009)

4824.

8. E. Wanigasekara, S. Perera, J.A. Crank, L. Sidisky, R. Shirey, A. Berthod, D.W.

Armstrong, Anal. Bioanal. Chem. 396 (2010) 511.

9. F. Zhao, Y. Meng, J.L. Anderson, J. Chromatogr. A 1208 (2008) 1.

10. Y. Meng, V. Pino, J.L. Anderson, Anal. Chem. 81 (2009) 7107.

11. Q. Zhao, J.C. Wajert, J.L. Anderson, Anal. Chem. 82 (2010) 707.

12. T. Welton, Chem. Rev. 99 (1999) 2071.

13. J.H. Davis Jr, Chem. Lett. 9 (2004) 1072.

14. E.D. Bates, R.D. Mayton, I. Ntai, J.H. Davis Jr, J. Am. Chem. Soc. 124 (2002) 926.

15. M.D. Soutullo, C.I. Odom, B.F. Wicker, C.N. Henderson, A.C. Stenson, J.H. Davis Jr.

Chem. Mater. 19 (2007) 3581.

16. J. Pawliszyn, Solid Phase Microextraction: Theory and Practice, Wiley-VCH, New

York, 1997.

17. P.A. Martos, J. Pawliszyn, Anal. Chem. 69 (1997) 206.

18. D. Aaron, C. Tsouris, Sep. Sci. Technol. 40 (2005) 321.

19. M. Chai, J. Pawliszyn, Environ. Sci. Technol. 29 (1995) 693.

20. A.E. Messner, D.M. Rosie, P.A. Argabright, Anal. Chem. 31 (1959) 230.

179

Chapter 8

A Rapid Analytical Method for Monitoring the Enantiomeric

Purity of Chiral Molecules Synthesized in Ionic Liquid

Solvents

A paper accepted by Chirality 1

Qichao Zhao, Pamela Twu, Jared L. Anderson

Abstract

Ionic liquids (ILs) have been widely used as reaction solvents in asymmetric synthesis due to their interesting physical and chemical properties. However, monitoring reactant-to-product conversion and the enantiopurity of formed stereoisomers often involves a tedious extraction step prior to chromatographic analysis. In this study, a rapid and sensitive sampling method using headspace solid-phase microextraction (SPME)

& Co.

180 coupled to chiral gas chromatography (GC) was developed for the “on-line” analysis of chiral molecules in the IL solvent. Three different SPME sorbent coatings, namely polydimethylsiloxane (PDMS), polyacrylate (PA), and a polymeric ionic liquid

(PIL)-based fiber, were examined in this study. The analytical performance of the developed method was evaluated in terms of reproducibility, slope of calibration curve, linear range, calibration linearity, and the determination of detection limits. The SPME method was successfully applied in the determination of enantiomeric excess from selected mixtures of chiral molecules. A preliminary study was performed using an

“on-fiber” derivatization approach revealing that the stereoisomers extracted by the

SPME fiber can be efficiently derivatized using a short “on-fiber” derivatization step. The developed SPME method eliminates the need of sequestering the reaction, separating the compounds of interest from the IL solvent, and the addition of a derivatizing reagent.

8.1 Introduction

Recently, increasing interest from a wide range of scientific fields has centered on the application of ionic liquids (ILs). ILs are typically defined as a class of molten salts which possess melting points below 100 ˚C [1,2]. This unique class of solvents is usually composed of substituted nitrogen or phosphorus-containing cations and various counter anions. Compared to traditional organic solvents, ILs possess a number of unique properties, including their extremely low vapor pressure at ambient temperature as well as high thermal stability. In addition, a wide variety of cations and anions can be used to generate ILs permitting the customization of physical and chemical properties of the

181 resulting ILs to produce solvents with varied viscosity and miscibility/immiscibility with a number of solvents. Thus far, ILs have been utilized in liquid-liquid extraction [3-5], analytical microextraction [6-8], electrochemistry [9,10], chromatographic separation

[11-13], and synthesis including asymmetric synthesis [1,2,14-16].

ILs have been successfully applied as reaction media for asymmetric synthesis since the first example was proposed by Chauvin and co-workers [17]. ILs can be employed in place of volatile organic solvents and provide the opportunity to recycle catalysts. It has been found that organometallic and biocatalysts can exhibit enhanced stability in

IL-based reaction media [15,16]. Moreover, one can easily introduce chirality into the structure of ILs, resulting in chiral ILs (CILs) with desired functionalities and properties

[18]. It was reported that CILs can induce chirality when used as the asymmetric synthesis reaction solvent, and yield appreciable enantoselectivities for some transformations [19]. CILs have also been applied as organocatalysts in Michael reactions, and excellent enantioselectivities were observed [20].

While ILs are considered as emerging solvents for asymmetric reactions, their non-volatile nature precludes them from being easily removed from the products under vacuum. Therefore, it is often difficult to monitor the progression of a given reaction as well as to determine the final enantiomeric excess (ee) of the chiral products without sequestering the reaction and subjecting the solvent to numerous separation methods to recover the product(s). Liquid-liquid extraction is a commonly employed method used to separate and recover products from the IL. This method can involve a large amount of organic solvents and only works if the chiral products partition into the IL-immiscible organic phase. Alternative methods have been developed to avoid the large consumption

182 of organic solvents. Brennecke and co-workers reported the use of supercritical CO2 to efficiently extract organic compounds from the IL-phase [21]. However, the supercritical fluid extraction method is expensive and is not suitable for polar compounds unless an organic modifier is used [22]. Tanaka and co-workers recovered products from the IL by distillation after reaction [23]. However, this method requires relatively expensive apparatuses and is limited to thermally stable and volatile compounds. It was also reported that some compounds can be precipitated from the IL-phase by adding a proper antisolvent, even though the mechanism of the precipitation process is complex and not well understood [24]. It is desired to carry out the final work-up step (e.g., extraction, distillation, precipitation) once a proper method indicates completeness of the reaction or good enantioselectivity. Therefore, there is a great need for a rapid, high-throughput

“on-line” sampling technique that can be readily coupled with chiral chromatography to provide information regarding reactant-to-product conversion and the enantiomeric purity of any chiral products produced in the reaction.

Solid-phase microextraction (SPME) is a sample preparation method that was developed by Pawliszyn and co-workers in the early 1990s [25]. Typically SPME consists of a thin, fused-silica fiber coated or bonded with a polymer sorbent material. When exposed to the sample matrix via headspace or direct-immersion mode, the compounds of interest are extracted by the sorbent coating. The extracted compounds can be easily thermally desorbed from the sorbent coating in the injection port of gas chromatograph

(GC). SPME has become a widely used technique due to the fact that it possesses significant advantages over traditional sample preparation methods, including its ease of operation as well as its amenability to automation with various chromatographic

183 techniques. In addition, the sampling and sample preparation are consolidated making it fast and simple. As a non-exhaustive sampling technique, only a small portion of the total compounds are extracted by the sorbent coating. SPME sampling devices are commercially available and permit the experimentalist to choose between three general categories of sorbent coating polarities: non-polar, semi-polar, and polar. To achieve superior extraction results, the polarity of the sorbent coatings should be similar to that of the compounds being extracted.

In this study, SPME was applied for the first time as a rapid, convenient, and sensitive sampling method in examining the enantiomeric excess of various chiral molecules dissolved in the 1-butyl-3-methylimidazolium bis[(trifluoromethyl)sulfonyl]imide (BMIM-NTf2) IL. The compounds examined in this study were previously examined via enantioselective synthesis in which ILs were used as the reaction solvent [26-28]. Two commercial SPME fibers possessing polydimethylsiloxane (PDMS) and polyacrylate (PA) sorbent coatings were used. A laboratory-made fiber possessing the polymeric ionic liquid (PIL)-based sorbent coating, namely poly(vinylhexylimidazolium bis[(trifluoromethyl)sulfonyl]imide)

[poly(VHIM-NTf2)] fiber, was also included in this study in order to evaluate the performance of this new class of sorbent coating and compare its performance to the commercially available sorbent coatings. The analytical performance of the developed method was thoroughly evaluated. In addition, the enantiomeric excess of mixtures containing various volume ratios of two pure enantiomers was determined in order to demonstrate the speed, precision, and accuracy of the developed method for the analysis of reaction mixtures. A preliminary study was performed using an “on-fiber”

184 derivatization approach to further expand the range of chiral molecules that can be studied using this method thereby eliminating tedious bulk-scale derivatization that is often required prior to chiral chromatographic separation.

8.2 Experimental

8.2.1 Chemicals and materials

It is important to note that in this study the chiral molecules were not synthesized using a chiral IL solvent. Instead, racemic mixtures and pure enantiomers of the chiral compounds were purchased and spiked at desired levels in an IL to develop the method.

The compounds studied in this work were purchased from Sigma-Aldrich (St. Louis, MO,

USA): methyl (R)-3-hydroxybutyrate, 1-phenylethanol,

α-(trimethylsilyloxyl)phenylacetonitrile, 1-methyl-3-phenylpropyl acetate, and α-ionone.

Methyl (S)-3-hydroxybutyrate was obtained from Fluka (Steimheim, Germany). The structures of these compounds are shown in Figure 8-1. Acetic anhydride was purchased from Sigma-Aldrich. The reagents involved in the synthesis of the IL solvent and the

PIL-based extraction phase, namely methylimidazole, vinylimidazole, 1-bromobutane,

1-bromohexane, and 2,2´-azo-bis(isobutyronitrile) (AIBN), were purchased from

Sigma-Aldrich. Lithium bis[(trifluoromethyl)sulfonyl]imide was obtained from SynQuest

Labs (Alachua, FL, USA). Deionized water (18.2 MΩ/cm) was obtained from a Milli-Q water-purification system (Millipore, Bedford, MA, USA). Ethyl acetate, chloroform,

2-propanol, hexane, acetone, methanol, methylene chloride, and sodium bicarbonate

(NaHCO3) were obtained from Fisher Scientific (Fairlawn, NJ, USA). Propane and

185

OH O OH O

 O  O

Methyl (S)-3-hydroxybutyrate Methyl (R)-3-hydroxybutyrate

OH CN

Si   O

1-Phenylethanol α-(Trimethylsilyloxyl)phenylacetonitrile

O O

  O

1-Methyl-3-phenylpropyl acetate α-Ionone

Figure 8-1: Structures of studied chiral molecules.

186 microflame brazing torches were purchased from Sigma-Aldrich.

All laboratory-made SPME devices were constructed using a 5-μL syringe purchased from Hamilton (Reno, NV, USA) and 0.05 mm I.D. fused silica capillary obtained from

Supelco (Bellefonte, PA, USA). Commercial SPME fibers of polydimethylsiloxane

(PDMS, film thickness of 100 μm) and polyacrylate (PA, film thickness of 85 μm) were obtained from Supelco. A fiber holder purchased from the same manufacturer was used for manual injection of the commercial fibers.

8.2.2 Synthesis of BMIM-NTf2 and poly(VHIM-NTf2)

The ILs and PIL used in this studied were synthesized following previously reported procedures [29-31]. The synthesis of the BMIM-NTf2 IL involved dissolving

1-methylimidazole in 2-propanol and reacting with an equimolar amount of

1-bromobutane under reflux and constant stirring conditions to generate

1-butyl-3-methylimidazolium bromide (BMIM-Br). After extraction with ethyl acetate, metathesis anion exchange was performed by dissolving BMIM-Br in Milli-Q water followed by the addition of an equimolar amount of lithium bis[(trifluoromethyl)sulfonyl]imide. The resulting precipitate was collected, washed with water to remove any residual halide anion, and dried overnight under vacuum. The purity

1 of BMIM-NTf2 was confirmed by H-NMR, which is included in the supplementary material.

The synthesis of the poly(VHIM-NTf2) PIL involved producing

1-vinyl-3-hexylimidazolium bromide (VHIM-Br) by reacting 1-vinylimidazole with an equimolar amount of 1-bromohexane in 2-propanol under reflux with constant stirring.

187

The VHIM-Br monomer was purified using ethyl acetate. Polymerization was performed in the presence of the free radical initiator AIBN to generate poly(1-vinyl-3-hexylimidazolium bromide) [poly(VHIM-Br)]. The polymerization step was repeated, if necessary, until the peaks belonging to the vinyl group in 1H-NMR disappeared. The obtained poly(VHIM-Br) product was dissolved in Milli-Q water, and an equimolar amount of lithium bis[(trifluoromethyl)sulfonyl]imide was introduced to the aqueous solution to perform metathesis anion exchange. The precipitated organic phase was collected, washed with water to remove residual halide impurities, and dried

1 overnight under vacuum. The purity of poly(VHIM-NTf2) was confirmed by H-NMR and is included in the supplementary material.

8.2.3 Preparation of PIL-based SPME fiber

A laboratory-made SPME device was constructed following a previously published procedure [30,31]. Briefly, a 1.0 cm segment of the polyimide polymer was removed from the end of the fiber using a high temperature flame followed by sealing the end of the capillary using a microflame torch. The fiber was then washed with methylene chloride, acetone, methanol, and hexane followed by 10 min of conditioning in the GC inlet at 250 °C. A coating solution was prepared by mixing the poly(VHIM-NTf2) PIL in chloroform at a ratio of 9:1 (v/v). The conditioned fiber was coated by dipping it into the

PIL coating solution, held for approximately 30 sec, and then removed from the coating solution. The coated fiber was allowed to dry in air for 15 min, followed by a 10 min conditioning step in the GC injection port at 250 ºC to remove any residual solvent.

188

8.2.4 Extraction of chiral molecules

Separate stock solutions were prepared by dissolving 60 mg of each compound into

3 g of the BMIM-NTf2 IL. The stock solutions were stored at 4 ºC and were used to prepare the working solution by mixing desired amounts of each stock solution with 120 mg of IL. All extractions were performed by exposing the SPME fiber into the headspace of clear crimp seal glass vials (2.0 mL) capped with an 11 mm magnetic crimp cap with

PTFE septa containing 300 mg of the IL working solution. All extractions were performed using a constant stir rate of 750 rpm at room temperature (approximately 19 ºC) maintained using a water bath.

In order to achieve satisfactory enantioresolution of some enantiomeric compounds using chiral gas chromatography, derivatization of the chiral compound is often necessary.

In this study, methyl-3-hydroxybutyrate and 1-phenylethanol were derivatized before preparing the stock solution, unless otherwise specified. To perform derivatization, 100

μL of the chiral alcohol was mixed with 120 μL of acetic anhydride and this mixture heated overnight at 90 ºC under constant stirring. A saturated NaHCO3 solution was introduced to the reaction mixture in order to hydrolyze any excess acetic anhydride as well as neutralize the generated acetic acid. The obtained organic phase was extracted with methylene chloride and washed with water. The complete removal of acetic acid was confirmed by testing the pH of the aqueous phase. Finally, the organic phase was collected and dried at 40 ºC overnight prior to use. The pure derivatized standards were used for the subsequent SPME study as well as the determination of enantiomeric excess by direct injection chiral GC analysis, unless otherwise specified.

189

8.2.5 “On-fiber” SPME derivatization

The working solution for the on-fiber SPME derivatization study was prepared by dissolving 10 μg of underivatized methyl (R)-3-hydroxybutyrate and 2 μg of underivatized methyl (S)-3-hydroxybutyrate in 2 g of BMIM-NTf2. One g of the working solution was transferred into a 2.0 mL extraction vial and a 45 min extraction was performed using a commercial PA fiber under a constant stir rate of 750 rpm at room temperature (approximately 19 ºC). Afterwards, the sampled fiber was removed from the extraction vial and exposed for 10 seconds to the headspace of acetic anhydride (5

μL)that was previously sealed in a 20 mL vial. The fiber was then transferred to the GC injection port permitting the derivatized analytes to be thermally desorbed from the

SPME fiber.

8.2.6 Chiral GC separation

All analyses were carried out using an Agilent 6850N gas chromatograph (Agilent

Technologies, Palo Alto, CA, USA) equipped with an Agilent 7683B autoinjector and a flame ionization detection (FID) system. All separations were performed using an Astec

CHIRALDEX G-TA (2,6-di-O-pentyl-3-trifluoroacetyl-γ-cyclodextrin) capillary column

(30 m × 0.25 mm I.D., 0.12 μm film thickness) provided by Supelco. The following optimized temperature program was used for the separation of the chiral compounds examined in this study: initial temperature of 80 ºC held for 5 min, then increased to 130

ºC at a ramp of 10 ºC/min and held for 5 min, and then raised to 170 ºC at a ramp of 2

ºC/min. Helium was used as the carrier gas with a flow rate of 1 mL/min. The inlet temperature was maintained at 250 ºC with an inlet desorption time of 5 min for all fibers.

190

Splitless injection was used with an applied purge flow to split vent of 20.0 mL/min at

0.10 min. The detector was held at 250 ºC. The detector make-up flow of helium was maintained at 45 mL/min, the hydrogen flow at 40 mL/min, and the air flow at 450 mL/min.

8.3. Results and Discussion

8.3.1 Description of the SPME-GC method

The main goal of this study is to develop and demonstrate a high-throughput sampling method based on SPME-GC for the analysis of chiral molecules (particularly the enantiomeric excess) from an ionic liquid. This “on-line” method, represented in

Figure 8-2, permits rapid sampling of the chiral compounds without disturbing the reaction. The extraction is performed by exposing the SPME fiber to the headspace of the reaction mixture allowing the chiral compounds, products, etc. to partition between the reaction mixture, headspace, and the sorbent coating on the fiber. Following the brief extraction step, the sampled fiber is removed from the reaction flask and thermally desorbed in the GC injection port, thereby subjecting the extracted molecules to the chiral

GC column for separation. This method circumvents the need to utilize organic solvents which are often implemented to pre-concentrate or phase separate the product from the IL.

In addition, the technique does not disturb the reaction/sample and can be very advantageous for monitoring the progress of the chemical reaction (rate of reactant conversion to products) as well as the reaction enantioselectivity.

191

Figure 8-2: Schematic diagram demonstrating the enantiomeric excess determination of racemic mixture using developed SPME-GC method

192

8.3.2 Generation of Sorption-Time Profiles

In SPME, the time in which the polymer coated fiber is exposed to the sample (i.e., sampling time) is an important parameter since the extraction efficiency of the extracted compounds increases with increasing time until equilibrium is established. To investigate the equilibration time for the studied SPME fibers, sorption-time profiles were generated by plotting the chromatographic peak area of each compound versus the extraction time.

As shown in Figure 8-3A for the PDMS fiber, 1-phenylethanol and methyl-3-hydroxybutyrate reached equilibrium in 60 min while the remaining compounds reached equilibrium after 90 min. For the PA fiber, none of the studied compounds reached equilibrium after an extraction time of 120 min (see Figure 8-3B), presumably due to the slower diffusion into the solid polymeric PA coating [32]. With respect to the

PIL fiber, equilibrium for all of the compounds was achieved at 60 min, as shown in

Figure 8-3C. The optimum extraction times, namely 90 min for the PDMS fiber, 120 min for the PA fiber, and 60 min for the PIL fiber, were chosen to study the reproducibility of the method as well as for calibration studies.

Sorption-time profiles also allow one to compare the extraction efficiencies between different polymer coated SPME fibers. It can be observed that at the optimum extraction time, the PDMS fiber exhibited the highest extraction efficiencies for all of the studied compounds. The extraction efficiency of the PIL fiber was considerably lower than the two commercial fibers largely due to its much smaller film thickness (15 μm for the PIL fiber versus 100 μm for PDMS fiber and 85 μm for PA fiber). A higher extraction efficiency for the PIL fiber would be expected if a fiber employing a thicker film was used.

193

Polydimethylsiloxane (PDMS) (A)

2000

1500

Peak area1000

500

0 0 30 60 90 120 Time (min)

(B) Polyacrylate (PA)

1000

Peak area Peak 500

0 0 30 60 90 120 Time (min)

400 Peak area Peak

200

0 0 30 60 90 120 Time (min)

Figure 8-3: Sorption-time profiles obtained for (A) commercial PDMS fiber (100 μm), (B)

194

PA fiber (85 μm), and (C) laboratory-made PIL fiber (≈ 15 μm) by extracting the studied analytes at a concentration of 2 mg/g using a constant stir rate of 900 rpm at room temperature (19 ºC): (♦) derivatized methyl-3-hydroxybutyrate, (▲) derivatized

1-phenylethanol, (X) α-(trimethylsilyloxyl)phenylacetonitrile, (●)

1-methyl-3-phenylpropyl acetate, (■) α-ionone.

195

8.3.3 Evaluation of Method Performance

When developing new sampling methods, it is important to examine its overall performance. The analytical performance of the method including reproducibility, slope of calibration curve, standard deviation of the regression line, linear range, correlation coefficient (R), and detection limits, was evaluated using the three SPME fibers. The reproducibilities of the examined SPME fibers in the extraction of the 5 compounds from the IL sample were determined by four duplicate extractions at their optimum extraction time. As shown in Table 8.1, the obtained relative standard deviation (RSD) values ranged from 3.6% to 8.4% for the PDMS fiber, 2.4% to 14.2% for the PA fiber, and 5.8% to 7.5% for the PIL fiber, respectively. The precision of the developed method employing all three SPME fibers is quite exceptional.

In order to demonstrate the applicability of this method towards quantitative analysis, calibration curves of each compound were obtained using the PDMS, PA, and PIL fibers at their respective optimum extraction time. The figures of merit of the calibration curves are listed in Tables 8.2. The obtained correlation coefficients (R) of the calibration curve, which is a measure of linear relationship among the points of the calibration curve, varied from 0.991 to 0.997 for the PDMS fiber, from 0.997 to 0.999 for the PA fiber, and from

0.994 to 0.997 for the PIL fiber. The detection limit of the extraction method, generally defined as the concentration of a studied compound that gives a signal 3 times the standard deviation of the background signal (noise), was calculated based on three times the standard deviation of the peak area obtained at the lowest concentration level of the calibration curve divided by the slope of the calibration curve. The obtained detection limits ranged from 2.3 to 21.7 μg/g (mass of compound/mass of IL) for PDMS fiber, 22.3

196

Table 8.1 Reproducibility of the PDMS, PA, and PIL-based fibers.

Analyte RSD (%)a

PDMS (100 PA PIL

μm) (85 μm) (≈ 15 μm)

Methyl-3-hydroxybutyrate 5.5 2.4 7.5

1-Phenylethanol 7.4 4.1 6.4

α-(Trimethylsilyloxyl)phenylacetonitrile 3.6 14.2 6.2

1-Methyl-3-phenylpropyl acetate 6.2 3.2 6.2

α-Ionone 8.4 5.0 5.8

a: Relative standard deviation was determined after performing four (4) successive extractions at the optimum extraction time of the respective sorbent coating.

197

Table 8.2 Figures of merit of the calibration curves and limits of detection by using the commercial PDMS fiber, PA fiber, and laboratory-made PIL fiber.a

b Sorbent Compounds Slope ± error Syx Linear range R Detection

coating ( x103) (μg g-1) limit

(μg g-1)

PDMS Methyl-3-hydroxybutyrate 559 ± 15 48.3 50 - 3000 0.997 6.7

1-Phenylethanol 874 ± 25 78.1 50 - 3000 0.997 4.6

α-(Trimethylsilyloxyl)phenylacetonitrile 108 ± 3 10.5 50 - 3000 0.996 21.7

1-Methyl-3-phenylpropyl acetate 446 ± 13 41.9 50 - 3000 0.997 6.6

α-Ionone 905 ± 45 140.9 50 - 3000 0.991 2.3

PA Methyl-3-hydroxybutyrate 436 ± 7 20.7 50 - 3000 0.999 23.4

1-Phenylethanol 629 ± 8 23.4 50 - 3000 0.999 22.5

α-(Trimethylsilyloxyl)phenylacetonitrile 66 ±2 6.0 50 - 3000 0.997 22.3

1-Methyl-3-phenylpropyl acetate 367 ± 6 18.1 50 - 3000 0.999 33.1

α-Ionone 411 ± 9 27.8 50 - 3000 0.998 23.9

198

Table 8.2 (continued)

b Sorbent Compounds Slope ± error Syx Linear range R Detection

coating ( x103) (μg g-1) limit

(μg g-1)

PIL Methyl-3-hydroxybutyrate 133 ± 5 17.0 50 - 3000 0.994 32.0

1-Phenylethanol 170 ± 5 15.1 50 - 3000 0.997 34.4

α-(Trimethylsilyloxyl)phenylacetonitrile 18 ± 0.5 1.5 50 - 3000 0.997 9.8

1-Methyl-3-phenylpropyl acetate 146 ± 5 17.0 50 - 3000 0.995 5.2

α-Ionone 331 ± 12 37.1 50 - 3000 0.995 10.3

a: Conditions: Sample weight, 300 mg; extraction time, 90 min for PDMS fiber, 120 min for PA fiber, and 60 min for PIL fiber; stir rate, 750 rpm; room temperature. b: Standard deviation of the regression.

199 to 33.1 μg/g for PA fiber, and 5.2 to 34.4 μg/g for PIL fiber. It should be noted that these detection limits are significantly lower than the concentration levels observed previously when the compounds were produced asymmetrically in IL solvents (i.e., 3.0×104 μg/g for

1-methyl-3-phenylpropyl acetate, 2.8×105 μg/g for

α-(trimethylsilyloxyl)phenylaceto-nitrile, and 4.5×104 μg/g for methyl-3-hydroxybutyrate

[26-28]). Therefore, this method can be easily detect organic molecules at concentration levels typically employed in organic reactions.

8.3.4 Determination of enantiomeric excess

The objective of this work is not only to develop a SPME method that can extract chiral molecules from the IL solvent but to also demonstrate the applicability towards the rapid and sensitive “on-line” determination of enantiomeric excess. To examine this approach, acetylated pure enantiomers of R and S methyl-3-hydroxybutyrate were chosen as representative compounds and mixed at desired volume ratios (i.e., methyl

(R)-3-hydroxybutyrate/methyl (S)-3-hydroxybutyrate = 3 μL/1.5 μL, 5 μL/1 μL, 50 μL/1

μL, 1.5 μL/3 μL, 1 μL/5 μL, and 1μL/50 μL) in 2 mL of BMIM-NTf2. The determination of enantiomeric purity was performed by headspace extraction using the PDMS and PIL fibers. For comparison purposes, standard mixtures using the same volume ratios were prepared in methylene chloride and examined by direct injection chiral GC analysis. It should be emphasized that it is only necessary to derivatize the chiral alcohols in order to achieve suitable enantioresolution during chromatographic separation and accurately determine the peak area of each eluting enantiomer.

Table 8.3 shows the enantiomeric excess values (%) of the R and S enantiomer

200

Table 8.3 Determination of enantiomeric excess values (%) of the derivatized methyl-3-hydroxybutyrate enantiomers using direct injection and SPME methodsa

Direct Injection PDMS fiber PIL fiber

R/S = 2/1 34.4 ± 0.2 33.0 ± 0.2 33.8 ± 0.1

R/S = 5/1 67.7 ± 0.2 68.1 ± 0.2 68.8 ± 0.3

R/S = 50/1 95.7 ± 0.0 96.0 ± 0.1 95.8 ± 0.1

S/R = 2/1 32.2 ± 0.6 31.1 ± 0.2 32.3 ± 0.6

S/R = 5/1 66.9 ± 0.3 66.8 ± 0.2 66.7 ± 0.6

S/R = 50/1 93.9 ± 0.1 93.6 ± 0.0 93.2 ± 0.0

a: Conditions: Direct injection volume, 0.5 μL; SPME sample weight, 300 mg; extraction time, 1 min for PDMS fiber and 5 min for PIL fiber; stir rate, 750 rpm; room temperature.

201 mixtures of methyl-3-hydroxybutyrate obtained using both direct injection and the SPME method. It was found that most of the determined enantiomeric excess values obtained by the SPME method were nearly identical to that of direct injection with the largest observed deviation of only 1.4% (for the sample with an R/S ratio of 2/1 in which 33.0%

± 0.2 was obtained for the SPME method using PDMS fiber versus 34.4% ± 0.2 for direct injection). Moreover, it should be noted that the SPME extraction time was shortened to only 1 min for the PDMS fiber and 5 min for the PIL fiber, making it a rapid method for evaluating reactant to product conversion as well as ascertaining reaction enantioselectivity. These extraction times were selected based on the equilibration time as well as extraction efficiency of the examined fibers. These results reveal the remarkable accuracy of the developed extraction method while underscoring the fact that the method can be used in a high-throughput fashion.

8.3.5 “On-fiber” SPME derivatization

The previous section clearly indicates the precision of the SPME approach for determining the enantiomeric excess of chiral compounds from IL solvents. However, if the chiral purity of the synthesized compound is desired, any derivatization reaction required for chromatographic analysis should be performed after the extraction step to eliminate adding derivatizing reagents to the reaction medium. In order to further demonstrate the rapid and accurate nature of enantiomeric excess determination exhibited by this approach, a study was performed to examine the feasibility of performing

“on-fiber” SPME derivatization using a commercial PA fiber. As shown in Figure 8-4, following the extraction of methyl-3-hydroxybutyrate, the SPME fiber was then exposed

202

Figure 8-4: Schematic diagram demonstrating the enantiomeric excess determination of a mixture using “on-fiber” derivatization

SPME-GC method.

203

(A)

(B)

Figure 8-5: Representative chromatograms of (A) extraction of underivatized

methyl-3-hydroxybutyrate enantiomers and (B) extraction of methyl-3-hydroxybutyrate

enantiomers using on-fiber derivatization method.

204 to the headspace of acetic anhydride allowing for the fiber loaded methyl-3-hydroxybutyrate to be derivatized “on-fiber”. The comparison between the extraction of underivatized methyl-3-hydroxybutyrate enantiomers and the one involving on-fiber derivatization is shown in Figure 8-5. As stated earlier, the underivatized methyl-3-hydroxybutyrate enantiomers can be successfully extracted by the PIL-fiber.

However, the peak shape and enantioresolution are very poor (see Figure 8-5A). One way to improve the separation is to derivatize the chiral products before thermally desorbing them into the GC inlet (as performed in all earlier experiments). As shown in Figure 8.5B, the extracted methyl-3-hydroxybutyrate enantiomers in the sorbent coating were successfully derivatized by the 10 second post-extraction derivatization resulting in significantly improved enantioresolution permitting enantiomeric excess determination.

In addition, the enantiomeric excess values obtained from the on-fiber derivatization method were in excellent agreement with the theoretical value. Therefore, one can easily expose the fiber to the headspace of a reaction mixture allowing for rapid extraction followed by quickly derivatizing the extracted compounds “on-fiber”, and then analysis of the conversion and enantiomeric excess of the chiral compounds. This approach eliminates the need to separate the IL solvent from the compounds of interest to simply monitor the reaction while also allowing for derivatization of any compounds prior to chromatographic separation.

8.4 Conclusions

ILs have gained increasing popularity in a variety of research fields, including

205 organic synthesis, due to the unique advantages they provide over traditional organic solvents. For the first time, a commercially available sampling technique, solid-phase microextraction (SPME), was coupled to chiral GC and applied as a high-throughput and sensitive sampling method for the determination of chiral compounds dissolved in ILs.

Fibers coated with two commercial SPME sorbent coatings as well as a laboratory-made

PIL sorbent coating were examined in this study. All three studied fibers exhibited exceptional reproducibility and low detection limits. The developed method was successfully applied toward the enantiomeric excess determination of mixtures containing various ratios of methyl (R)-3-hydroxybutyrate and methyl

(S)-3-hydroxybutyrate. It was observed that the enantiomeric excess results obtained from the SPME method correlated well with that from direct injection. In addition, the precise and accurate results were obtained using brief extraction times (1-5 minutes). An

“on-fiber” SPME derivatization approach further demonstrated the applicability of the developed method for the rapid and convenient enantiomeric excess determination of chiral molecules that cannot be well resolved chromatographically without derivatization.

This SPME approach uses cheap commercially-available fibers ($300 USD) and can be coupled with any gas chromatograph. This method is capable of quickly monitoring the progression of organic reactions as well as determination of reaction enantioselectivity without sequestering the reaction prior to work-up.

Acknowledgements

J.L.A. acknowledges funding from the Analytical and Surface Chemistry Program in

206 the Division of Chemistry and the Separation and Purification Processes Program in the

Chemical, Environmental, Bioengineering, and Transport Systems Division from the

National Science Foundation for a CAREER grant (CHE-0748612)

References

1. Welton T. Room-temperature ionic liquids. Solvents for synthesis and catalysis. Chem

Rev 1999;99:2071-2083.

2. Hallett JP, Welton T. Room-temperature ionic liquids: Solvents for synthesis and

catalysis 2. Chem Rev 2011;111:3508-3576.

3. Poole CF, Poole SK. Extraction of organic compounds with room temperature ionic

liquids. J Chromatogr A 2010;1217:2268-2286.

4. Martinis EM, Berton P, Monasterio RP, Wuilloud RG. Emerging ionic liquid-based

techniques for total-metal and metal-speciation analysis. TrAC-Trend Anal Chem

2010;29:1184-1201.

5. Francisco M, Arce A, Soto Ana. Ionic liquids on desulfurization of fuel oils. Fluid

Phase Equilibr 2010;294:39-48.

6. Liu R, Liu J-F, Yin Y-G, Hu X-L, Jiang G-B. Ionic liquids in sample preparation. Anal

Bioanal Chem 2009;393:871-883.

7. Aguilera-Herrador E, Lucena R, Cardenas Soledad, Valcarcel M. The roles of ionic

liquids in sorptive microextraction techniques. TrAC-Trend Anal Chem

2010;29:602-616.

8. Zhao Q, Anderson JL. Task-specific microextraction using ionic liquids. Anal Bioanal

207

Chem 2011;400:1613-1618.

9. Shiddiky MJA, Torriero AAJ. Application of ionic liquids in electrochemical sensing

systems. Biosens Bioelectron 2011;26:1775-1787.

10. Walsh DA, Lovelock KRJ, Licence P. Ultramicroelectrode voltammetry and scanning

electrochemical microscopy in room-temperature ionic liquid electrolytes. Chem Soc

Rev 2010;39:4185-4194.

11. Yao C, Anderson JL. Retention characteristics of organic compounds on molten salt

and ionic liquid-based gas chromatography stationary phases. J Chromatogr A

2009;1216:1658-1712.

12. Han X, Armstrong DW. Ionic liquids in separations. Acc Chem Res

2007;40:1079–1086.

13. Stalcup AM, Cabovska B. Ionic liquids in chromatography and capillary

electrophoresis. J Liq Chromatogr 2004;27:1443-1459.

14. Ma Z, Yu J-H, Dai S. Preparation of inorganic materials using ionic liquids. Adv

Mater 2010;22:261-285.

15. Baudequin C, Baudoux J, Levillain J, Cahard D, Gaumont A-C, Plaquevent J-C. Ionic

liquids and chirality: Opportunities and challenges. Tetrahydron: Asymmetry

2003;14:3081-3093.

16. Toma S, Mecirova M, Sebesta R. Are ionic liquids suitable media for organocatalytic

reactions? Eur J Org Chem 2009;321-327.

17. Chauvin Y, Mussmann L, Olivier H. A novel class of versatile solvents for two-phase

catalysis: Hydrogenation, isomerization, and hydroformylation of alkenes catalyzed

by rhodium complexes in liquid 1,3-dialkylimidazolium salt. Angew Chem Int Ed

208

Engl 1995;34:2698–2700.

18. Earle MJ, McCormac PB, Seddon KR. Diels–Alder reactions in ionic liquids. Green

Chem. 1999;1:23 – 25.

19. Gausepohl R, Buskens P, Kleinen J, Bruckmann A, Lehmann CW, Klankermayer J,

Leitner W. Highly Enantioselective Aza-Baylis–Hillman Reaction in a Chiral

Reaction Medium. Angew. Chem. Int. Ed. 2006;45:3689-3692.

20. Luo S, Mi X, Zhang L, Liu S, Xu H, Cheng J-P. Functionalized chiral ionic liquids as

highly efficient asymmetric organocatalysts for michael addition to nitroolefins.

Angew. Chem. Int. Ed. 2006;45:3093-3097.

21. Blanchard LA, Brennecke JF. Recovery of organic products from ionic liquids using

supercritical carbon dioxide. Ind Eng Chem Res 2001;40:287-292.

22. Hawthorne SB. Analytical-scale supercritical fluid extraction. Anal Chem

1990;62:633A–642A.

23. Mizushima E, Hayashi T, Tanaka M. Palladium-catalysed carbonylation of aryl

halides in ionic liquid media: High catalyst stability and significant rate-enhancement

in alkoxycarbonylation. Green Chem 2001;3:76-79.

24. Sanders MW, Wright L, Tate L, Fairless G, Chrowhurst L, Bruce NC, Walker AJ,

Hembury GA, Shimizu S. Unexpected preferential dehydration of artemisinin in ionic

liquids. J Phys Chem A 2009;113:10143-10145.

25. Arthur CL, Pawliszyn J. Solid phase microextraction with thermal desorption using

fused silica optical fibers. Anal Chem 1990;62:2145–2148.

26. Kim K-W, Song B, Choi M-Y, Kim M-J. Biocatalysis in ionic liquids: Markedly

enhanced enantioselectivity of lipase. Org Lett 2001;3:1507-1509.

209

27. Baleizão C, Gigante B, Garcia H, Corma A. Ionic liquids as green solvents for the

asymmetric synthesis of cyanohydrins catalysed by VO(salen) complexes. Green

Chem 2002;4:272–274.

28. Ngo HL, Hu A, Lin W. Highly enantioselective catalytic asymmetric hydrogenation

of β-keto esters in room temperature ionic liquids. Chem Commun 2003;1912–1913.

29. Anderson JL, Ding J, Welton T, Armstrong DW. Characterizing ionic liquids on the

basis of multiple solvation interactions. J. Am. Chem. Soc. 2002;124:14247-14254.

30. Zhao F, Meng Y, Anderson JL. Polymeric ionic liquids as selective coatings for the

extraction of esters using solid-phase micorextraction. J Chromatogr A

2008;1208:1-9.

31. Zhao Q, Wajert JC, Anderson JL. Polymeric ionic liquids as CO2 selective sorbent

coatings for solid-phase microextraction. Anal Chem 2010;82:707-713.

32. Louch D, Motlagh S, Pawliszyn J. Dynamics of organic compound extraction from

water using liquid-coated fused silica fibers. Anal Chem 1992;64:1187-1199.

210

Chapter 9

Summary

The first part of this dissertation describes the applications of ionic liquids (ILs) and binary mixture containing polymeric ionic liquids (PILs) as GC stationary phases.

Functionalized ILs containing tris(pentafluoroethyl)-trifluorophosphate (FAP)-based ILs were characterized using the solvation parameter model. The role of cation functional groups, cation type, and nature of counter anion on the system constants were evaluated.

With the presence of functional groups in the cationic moieties, namely amino, ester, and hydroxyl groups, the dipolarity, hydrogen bond basicity, and hydrogen bond acidity of resulting ILs can be modified. The IL containing alkylated phosphonium cation, paired with FAP− anion, exhibited the strongest cohesivity and weakest hydrogen bond interactions. The FAP-based ILs was found to possess lower hydrogen bond basicity than the NTf2-based ILs having the identical cationic component. The unique and interesting separation selectivity offered by the functionalized FAP-based ILs in gas chromatography make them a promising class of stationary phases.

The solvation parameters of fifteen ILs containing various functional groups (i.e., dimethylamino, hydroxyl, and ester) and cation types paired with various counter anions,

− − − − − − namely FAP , NTf2 , SCN , C(CN)3 , B(CN)4 , and BOB were characterized using the

211 solvation parameter model. The presence of cation functional groups yielded influence on the hydrogen bond basicity, hydrogen bond acidity, as well as dispersion interaction of the resulting ILs, while the change of cation type affected dipolarity modestly. It was observed that the switch of counter anion produced ILs with enhanced dipolarity and hydrogen bond basicity. The modulation of cation and anion combination allowed for fine control of solute retention as well as separation selectivity.

Mixtures containing two different PILs were applied as GC stationary phases and characterized in terms of thermal stabilities, solvation properties, and separation selectivities. Compared to the GC stationary phases that were composed of monomeric IL mixtures, the bleed temperatures of stationary phases composed of binary PIL mixtures were significantly improved. The stationary phase containing higher weight percentage of chloride anion exhibited enhanced hydrogen bond basicity interaction, whereas the PIL with longer alky chain appended to the cationic moiety showed stronger dispersion interaction. This modulation in the system constants allows for the fine tuning of analyte retention and the separation selectivity of chosen analytes, including proton-donating analytes as well as ketones, aldehydes, and aromatic compounds.

The second part of this dissertation presented the application of PILs as sorbent coatings in SPME. The structures of two PILs were carefully tailored to incorporate specific functionalities to serve as task-specific PIL sorbent coatings that exhibited high extraction efficiency and selectivities for CO2. The PIL sorbent coatings were capable of undergoing two different types of mechanisms responsible for capturing the CO2 sorbate, namely physical sorption by the poly(VHIM-NTf2) coating and carbamate formation by the poly(VHIM-taurinate) coating. Extraction efficiencies were compared to those of two

212 commercial SPME fibers, namely PDMS and carboxen fibers. It was observed that the poly(VHIM-NTf2) fiber exhibited comparable extraction efficiency and sensitivity to that of carboxen fiber, while the poly(VHIM-taurinate) fiber showed superior capability in retaining CO2 sorbate on the fiber under various conditions. The extraction performance of the PIL fibers was expected to be further improved if a thicker sorbent coating can be used.

Two PIL-based SPME fibers, namely poly(VHIM-NTf2) and poly(VHIM-taurinate) fibers, were used for the selective extraction of CO2 from simulated flue gas. For comparison purposes, a commercial carboxen fiber was also studied. It was observed that the poly(VHIM-taurinate) fiber exhibited the lowest sensitivity drop in the presence of water vapor, while the sensitivity of the carboxen fiber dropped substantially. The effect of temperature on the analytical performance of the PIL-based fibers was examined. The extraction of CH4 and N2 was performed and the selectivities of the PIL-based and carboxen fibers were compared. A scanning electron microscopy study suggested that the amino group of the poly(VHIM-taurinate) coating was capable of selectively reacting with CO2 but not CH4 or N2, resulting in a significant surface morphology change of the sorbent coating.

A rapid and sensitive sampling method using headspace SPME coupled to chiral GC was developed for “on-line” monitoring the reactant-to-product conversion and enantiopurity of stereoisomers formed in the IL solvent. Two commercial SPME fibers, namely PDMS and PA fiber, as well as a PIL-based fiber, namely poly(VHIM-NTf2), were utilized and compared. The analytical performance of the developed method was thoroughly evaluated in terms of reproducibility, sensitivity, linear range, calibration

213 linearity, and the determination of detection limit. The SPME method was successfully applied towards the enantiomeric excess determination of mixtures containing various ratios of selected chiral molecules, and the results correlated well with that from direct injection. An “on-fiber” derivatization approach was demonstrated, further indicating the applicability of the developed method for the rapid and convenient enantiomeric excess determination of chiral molecules that can not be well resolved chromatographically without derivatization.

214

References

Chapter 1

1. Hallett, J. P.; Welton, T. Chem. Rev. 2011, 111, 3508.

2. Davis Jr., J. H. Chem. Lett. 2004, 9, 1072.

3. Walden, P. Bull. Acad. Sci. St. Petersburg 1914, 405.

4. Wilkes, J. S.; Levisky, J. A.; Wilson, R. A.; Hussey, C. L. Inorg. Chem. 1982, 21,

1263.

5. Wilkes, J. S.; Zaworotko, M. J. Chem. Commun. 1992, 965.

6. Poole, C. F.; Poole, S. K. J. Chromatogr. A 2010, 1217, 2268.

7. Martinis, E. M.; Berton, P.; Monasterio, R. P.; Wuilloud, R. G. TrAC-Trend Anal Chem

2010, 29, 1184.

8. Aguilera-Herrador, E.; Lucena, R.; Cardenas, S.; Valcarcel, M. TrAC-Trend Anal Chem

2010, 29, 602.

9. Zhao, Q.; Anderson, J. L. Anal. Bioanal. Chem. 2011, 400, 1613.

10. Yao, C.; Anderson, J. L. J. Chromatogr. A 2009, 1216, 1658.

11. Stalcup, A. M.; Cabovska, B. J. Liq. Chromatogr. Relat. Technol. 2004, 27, 1443.

12. Shiddiky, M. J. A.; Torriero, A. A. J. Biosens. Bioelectron. 2011, 26, 1775.

13. Walsh, D. A.; Lovelock, K. R. J.; Licence, P. Chem. Soc. Rev. 2010, 39, 4185.

215 14. Tholey, A.; Heinzle, E.; Anal. Bioanal. Chem. 2006, 386, 24.

15. Armstrong, D.W.; He, L.; Liu, Y. Anal. Chem. 1999, 71, 3873.

16. Anderson, J. L.; Ding, J.; Welton, T.; Armstrong, D. W. J. Am. Chem. Soc. 2002, 124,

14247.

17. Anderson, J. L.; Armstrong, D. W. Anal. Chem. 2003, 75, 4851.

18. Anderson, J. L.; Ding, R.; Ellern, A.; Armstrong, D. W. J. Am. Chem. Soc. 2005, 127,

593.

19. Payagala, T.; Zhang, Y.; Wanigasekara, E.; Huang, K.; Breitbach, Z. S.; Sharma, P. S.;

Sidisky, L. M.; Armstrong, D. W. Anal. Chem. 2009, 81, 160.

20. Anderson, J. L.; Armstrong, D. W. Anal. Chem. 2005, 77, 6453.

21. Kovats, E. Helv. Chim. Acta 1958, 41, 1915.

22. Rohrschneider, L. J. Chromatogr. 1966, 22, 6.

23. McReynolds, W. O. J. Chromatogr. Sci. 1970, 8, 685.

24. Abraham, M. H. Chem. Soc. Rev. 1993, 22, 73.

25. Breitbach, Z. S.; Armstrong, D. W. Anal. Bioanal. Chem. 2008, 390, 1605.

26. Huang, K.; Han, X.; Zhang, X.; Armstrong, D. W. Anal. Bioanal. Chem. 2007, 389,

2265.

27. Hsieh, Y.-N.; Ho, W.-Y.; Horng, R. S.; Huang, P.-C.; Hsu, C.-Y.; Huang, H.-H.; Kuei,

C.-H. Chromatographia 2007, 66, 607.

28. Hsieh, Y.-N.; Horng, R. S.; Ho, W.-Y.; Huang, P.-C.; Hsu, C.-Y.; Whang, T.-J.; Kuei,

C.-H. Chromatographia 2008, 67, 413.

29. Sun, X.; Zhu, Y.; Wang, P.; Li, J.; Wu, C.; Xing, J. J. Chromatogr. A 2011, 1218,

833.

216 30. Baltazar, Q. Q.; Leininger, S. K.; Anderson, J. L. J. Chromatogr. A 2008, 1182, 119.

31. Zhao, Q.; Anderson, J. L. J. Sep. Sci. 2010, 33, 79.

32. Zhao, Q.; Eichhorn, J.; Pitner, W. R.; Anderson, J. L. Anal. Bioanal. Chem. 2009, 395,

225.

33. Álvarez, J. G.; Gomis, D. B.; Abrodo, P. A.; Llorente, D. D.; Busto, E.; Lombardía, N.

R.; Fernández, V. G.; Álvarez, M. D. G. Anal. Bioanal. Chem. 2011, 400, 1209.

34. Berthod, A.; He, L.; Armstrong, D. W. Chromatographia 2001, 53, 63.

35. Huang, K.; Zhang, X.; Armstrong, D. W. J. Chromatogr. A 2010, 1217, 5261.

36. Ding, J.; Welton, T.; Armstrong, D. W. Anal. Chem. 2004, 76, 6819.

Chapter 2

1. Welton T (1999) Chem Rev 99:2071-2083.

2. Wasserscheid P, Keim W (2000) Angew Chem Int Ed 39:3772-3789.

3. Cole AC, Jensen JL, Ntai I, Tran KR, Weaver KJ, Forbes DC, Davis JH (2002) J Am

Chem Soc 124:5962-5963.

4. Handy ST, Zhang X (2001) Org. Lett. 3:233-236.

5. Dai S, Ju YH, Barnes CE (1999) J Chem Soc Dalton Trans 8:1201-1202.

6. Chun S, Dzyuba SV, Bartsch RA (2001) Anal Chem 73:3737-3741.

7. Visser AE, Swatloski RP, Reichert WM, Mayton R, Sheff S, Wierzbicki A, Davis JH,

Rogers RD (2001) Chem Commun 135-136.

8. Liu JF, Jiang GB, Chi YG, Cai YQ, Zhou QX, Hu JT (2003) Anal Chem

75:5870-5876.

217

9. Liu J-F, Li N, Jiang G-B, Liu J-M, Jönsson JÅ, Wen M-J (2005) J Chromatogr A

1066:27–32.

10. Hsieh Y-N, Huang P-C, Sun I-W, Whang T-J, Hsu C-Y, Huang H-H, Kuei C-H (2006)

Anal Chim Acta 557:321–328.

11. Zhao F, Meng Y, Anderson JL (2008) J Chromatogr A 1208:1-9.

12. Armstrong DW, Zhang LK, He L, Gross ML (2001) Anal Chem 73:3679-3686.

13. Mank M, Stahl B, Boehm G (2004) Anal Chem 76:2938-2950.

14. Li YL, Gross ML (2004) J Am Soc Mass Spectrom 15:1833–1837.

15. Dickinson EV, Williams ME, Hendrickson SM, Masui H, Murray RW (1999) J Am

Chem Soc 121:613-616.

16. Doyle KP, Lang CM, Kim K, Kohl PA (2006) J Electrochem Soc 153:A1353-1357

17. Poole CF (2004) J Chromatogr A 1037:49-82.

18. Anderson JL, Armstrong DW (2003) Anal Chem 75:4851-4858.

19. Stalcup AM, Cabovska B (2004) J Liq Chromatogr Relat Technol 27:1443-1459.

20. Chan SS, Willis CJ (1968) Can J Chem 46:1237-1248.

21. Swatloski RP, Holbrey JD, Rogers RD (2003) Green Chem 5:361-363.

22. Ignat’ev NV, Welz-Biermann U, Kucheryna A, Bissky G, Willner H (2005) J Fluorine

Chem 126:1150-1159.

23. O’Mahony AM, Silvester DS, Aldous L, Hardacre C, Compton RG (2008) J Chem

Eng Data 53:2884-2891.

24. Abedin SZ, Borissenko N, Endres F (2004) Electrochem Commun 6:422-426.

25. Duffy NW, Bond AW (2006) Electrochem Commun 8:892-898.

218

26. Millefiorini S, Tkaczyk AH, Sedev R, Efthimiadis J, Ralston J (2006) J Am Chem

Soc 128:3098-3101.

27. Muldoon MJ, Aki SN, Anderson JL, Dixon JK, Brennecke JF (2007) J Phys Chem B

111:9001-9009.

28. Chrobok A, Swadzba M, Baj S (2007) Polish J Chem 81:337-344.

29. Abraham MH (1993) Chem Soc Rev 22:73-83.

30. Anderson JL, Ding J, Welton T, Armstrong DW (2002) J Am Chem Soc

124:14247-14254.

31. Mutelet F, Jaubert JN (2006) J Chromatogr A 1102: 256-267.

32. Hsieh YN, Ho WY, Horng RS, Huang PC, Hsu CY, Huang HH, Kuei CH (2007)

Chromatographia 66:607-611.

33. Sprunger LM, Proctor A, Acree WE, Abraham MH (2008) Fluid Phase Equilib

265:104-111.

34. Baltazar QQ, Leininger SK, Anderson JL (2008) J Chromatogr A 1182:119-127.

35. Bates ED, Mayton RD, Ntai I, Davis JH (2002) J Am Chem Soc 124:926-927.

36. Froehlich U (2005) Chirality and Ionic Liquids. Ph.D dissertation. Queen’s University

of Belfast.

37. Crowhurst L, Mawdsley PR, Perez-Arlandis JM, Salter PA, Welton T (2003) Phys

Chem Chem Phys 5:2790-2794.

38. Sharma NK, Tickell MD, Anderson JL, Kaar J, Pino V, Wicker BF, Armstrong DW,

Davis JH, Russell AJ (2006) Chem Commun 646-648.

39. Zhang XM, Bordwell FG, Puy MVD, Fried HE (1993) J Org Chem 58:3060-3066.

219

Chapter 3

1. T. Welton, Chem. Rev. 99 (1999) 2071.

2. J.H. Davis Jr., Chem. Lett. 9 (2004) 1072.

3. D.W. Armstrong, L. He, Y. Liu, Anal. Chem. 71 (1999) 3873.

4. C.F. Poole, S.K. Poole, J. Sep. Sci. 34 (2011) 888.

5. C.F. Poole, K.G. Furton, B.R. Kersten, J. Chromatogr. Sci. 24 (1986) 400.

6. J.L. Anderson, J. Ding, T. Welton, D.W. Armstrong, J. Am. Chem. Soc. 124 (2002)

14247.

7. Q.Q. Baltazar, S.K. Leininger, J.L. Anderson, J. Chromatogr. A 1182 (2008) 119.

8. J.L. Anderson, D.W. Armstrong, Anal. Chem. 75 (2003) 4851.

9. Q. Zhao, J. Eichhorn, W.R. Pitner, J.L. Anderson, Anal. Bioanal. Chem. 395 (2009)

225.

10. J.L. Anderson, R. Ding, A. Ellern, D.W. Armstrong, J. Am. Chem. Soc. 127 (2005)

593.

11. K. Huang, X. Han, X. Zhang, D.W. Armstrong, Anal. Bioanal. Chem. 389 (2007)

2265.

12. T. Payagala, Y. Zhang, E. Wanigasekara, K. Huang, Z.S. Breitbach, P.S. Sharma, L.M.

Sidisky, D.W. Armstrong, Anal. Chem. 81 (2009) 160.

13. J.L. Anderson, D.W. Armstrong, Anal. Chem. 77 (2005) 6453.

14. Y. -N. Hsieh, W.-Y. Ho, R.S. Horng, P.-C. Huang, C.-Y. Hsu, H.-H. Huang, C.-H. Kuei,

Chromatographia 66 (2007) 607.

220

15. Q. Zhao, J.L. Anderson, J. Sep. Sci. 33 (2010) 79.

16. X. Sun, Y. Zhu, P. Wang, J. Li, C. Wu, J. Xing, J. Chromatogr. A 1218 (2011) 833.

17. M.H. Abraham, Chem. Soc. Rev. 22 (1993) 73.

18. S.K. Poole, C.F. Poole, Analyst 120 (1995) 289.

19. L.M. Grubbs, S. Ye, M. Saifullah, W.E. Acree, Jr., P. Twu, J.L. Anderson, G.A. Baker,

M.H. Abraham, J. Solution Chem., in press.

20. L. Crowhurst, P.R. Mawdsley, J.M. Perez-Arlandis, P.A. Salter, T. Welton, Phys.

Chem. Chem. Phys. 5 (2003) 2790.

21. N.K. Sharma, M.D. Tickell, J.L. Anderson, J. Kaar, V. Pino, B.F. Wicker, D.W.

Armstrong, J.H. Davis, A.J. Russell, Chem. Commun. (2006) 646.

Chapter 4

1. Welton, T., Chem. Rev. 1999, 99, 2071-2083.

2. Wasserscheid, P., Keim, W., Angew. Chem. Int. Ed. 2000, 39, 3772-3789.

3. Cole, A. C., Jensen, J. L., Ntai, I., Tran, K. L. T., Weaver, K. J., Forbes, D. C., Davis, J.

H. Jr., J. Am. Chem. Soc. 2002, 124, 5962-5963.

4. Sheldon, R. A., Lau, R. M., Sorgedrager, M. J., van Rantwijk, F., Seddon, K. R., Green.

Chem. 2002, 4, 147-151.

5. Morris, R. E., Chem. Commun. 2009, 21, 2990-2998.

6. Dai, S., Ju, Y. H., Barnes, C. E., J. Chem. Soc. Dalton. Trans. 1999, 8, 1201-1202.

7. Chun, S., Dzyuba, S. V., Bartsch, R. A., Anal. Chem. 2001, 73, 3737-3741.

221

8. Visser, A. E., Swatloski, R. P., Reichert, W. M., Mayton, R., Sheff, S., Wierzbicki, A.,

Davis, J. H. Jr., Rogers, R. D., Chem. Commun. 2001, 135-136.

9. Liu, J. F., Jiang, G. B., Chi, Y. G., Cai, Y. Q., Zhou, Q. X., Hu, J. T., Anal. Chem. 2003,

75, 5870-5876.

10. Yao, C., Pitner, W. R., Anderson, J. L., Anal. Chem. 2009, 81, 5054-5063.

11. Liu, J. F., Li, N., Jiang, G. B., Liu, J. M., Jönsson, J. Å., Wen, M. J., J. Chromatogr. A

2005, 1066, 27-32.

12. Zhao, F., Meng, Y., Anderson, J. L., J. Chromatogr. A 2008, 1208, 1-9.

13. Meng, Y., Pino, V., Anderson, J. L., Anal. Chem. 2009, 81, 7107-7112.

14. Armstrong, D. W., Zhang, L. K., He, L., Gross, M. L., Anal. Chem. 2001, 73,

3679-3686.

15. Mank, M., Stahl, B., Boehm, G., Anal. Chem. 2004, 76, 2938-2950.

16. Tholey, A., Heinzle, E., Anal. Bioanal. Chem. 2006, 386, 24-37.

17. Rouse, C. A., Acree, W. E. Jr., J. Chromatogr. 1986, 357(1), 33-37.

18. Byrd, J. M., Acree, W. E. Jr., Mol. Cryst. Liq. Cryst., Letter Sections.1988, 5(5),

163-169.

19. Poole, C. F., Kiridena, W., Nawas, M. I., Koziol, W. W., J. Sep. Sci. 2002, 25,

749-759.

20. Baltazar, Q. Q., Leininger, S. K., Anderson, J. L., J. Chromatogr. A 2008, 1182,

119-127.

21. Yao, C., Anderson, J. L., J. Chromatogr. A 2009, 1216, 1658-1712.

22. Anderson, J. L., Armstrong, D. W., Anal. Chem. 2005, 77, 6453-6462.

222

23. Hsieh, Y.-N., Ho, W.-Y., Horng, R.S., Huang, P.-C., Hsu, C.-Y., Huang, H.-H., Kuei,

C.-H., Chromatographia. 2007, 66, 607-611.

24. Hsieh, Y.-N., Horng, R.-S., Ho, W.-Y., Huang, P.-C., Hsu, C.-Y., Whang, T.-J., Kuei,

C.-H., Chromatographia. 2008, 67, 413-420.

25. Abraham, M. H., Chem. Soc. Rev. 1993, 22, 73-83.

26. Anderson, J. L., Ding, J., Welton, T., Armstrong, D. W., J. Am. Chem. Soc. 2002 124,

14247-14254.

27. Mutelet, F., Jaubert, J. N., J. Chromatogr. A 2006, 1102, 256-267.

28. Sprunger, L. M. , Proctor, A., Acree, W. E., Abraham, M. H., Fluid. Phase. Equilib.

2008, 265, 104-111.

29. Zhao, Q., Eichhorn, J., Pitner, W. R., Anderson, J. L., Anal. Bioanal. Chem. 2009, 395,

225-234.

30. Sharma, N. K., Tickell, M. D., Anderson, J. L., Kaar, J., Pino, V., Wicker, B. F.,

Armstrong, D. W., Davis, J. H. Jr., Russell, A. J., Chem. Commun. 2006, 646-648.

31. Marcilla, R., Blazquez, J. A., Rodriguez, J., Pomposo, J. A., Mecerreyes, D., J. Polym.

Sci. A: Polym. Chem. 2004, 42, 208-212.

32. Anderson, J. L., Armstrong, D. W., Anal. Chem. 2003, 75, 4851-4858.

33. Ngo, H. L., LeCompte, K., Hargens, L., McEwen, A. B., Thermochim. Acta. 2000,

357-358, 97-102.

Chapter 5

1. Arthur, C. L.; Pawliszyn, J. Anal. Chem. 1990, 62, 2145.

223

2. Liu, J.-F.; Li, N.; Jiang, G.-B.; Liu, J.-M.; Jösson, J. Å.; Wen, M.-J. J. Chromatogr. A

2005, 1066, 27.

3. Amini, R.; Rouhollahi, A.; Adibi, M.; Mehdinia, A. Talanta 2011, 84, 1.

4. Hsieh, Y.-N.; Huang, P.-C.; Sun, I.-W.; Whang, T.-J.; Hsu, C.-Y. ; Huang, H.-H.; Kuei,

C.-H. Anal. Chim. Acta 2006, 557, 321.

5. Huang, K.-P.; Wang, G.-R.; Huang, B.-Y. ; Liu, C.-Y. Anal. Chim. Acta 2009, 645, 42.

6. He, Y.; Pohl, J.; Engel, R.; Rothman, L.; Thomas, M. J. Chromatogr. A 2009, 1216,

4824.

7. Zhao, F.; Wang, M.; Ma, Y.; Zeng, B. J. Chromatogr. A 2011, 1218, 387.

8. Wanigasekara, E.; Perera, S.; Crank, J. A.; Sidisky, L.; Shirey, R.; Berthod, A.;

Armstrong, D. W. Anal. Bioanal. Chem. 2010, 396, 511.

9. Liu, M.; Zhou, X.; Chen, Y.; Liu, H.; Feng, X.; Qiu, G.; Liu, F.; Zeng, Z. Anal. Chim.

Acta 2010, 683, 96.

10. Amini, R.; Rouhollahi, A.; Adibi, M.; Mehdinia, A. J. Chromatogr. A 2011, 1218,

130.

11. Zhao, F.; Meng, Y.; Anderson, J. L. J. Chromatogr. A 2008, 1208, 1.

12. Meng, Y.; Pino, V.; Anderson, J. L. Anal. Chem. 2009, 81, 7107.

13. López-Darias, J.; Pino, V.; Anderson, J. L.; Graham, C. M.; Afonso, A. M. J.

Chromatogr. A 2010, 1217, 1236.

14. Meng, Y.; Anderson, J. L. J. Chromatogr. A 2010, 1217, 6143.

15. Meng, Y.; Pino, V.; Anderson, J. L. Anal. Chim. Acta 2011, 687, 141.

16. Graham, C. M.; Meng, Y.; Ho, T.; Anderson, J. L. J. Sep. Sci. 2011, 34, 340.

17. Zhao, Q.; Wajert, J. C.; Anderson, J. L. Anal. Chem. 2010, 82, 707.

224

18. Zhao, Q.; Anderson, J. L. J. Chromatogr. A 2010, 1217, 4517.

19. Zhou, X.; Xie, P.-F.; Wang, J.; Zhang, B.-B.; Liu, M.-M.; Liu, H.-L.; Feng, X.-H. J.

Chromatogr. A 2011, 1218, 3571.

20. Shearrow, A. M.; Harris, G. A.; Fang, L.; Sekhar, P. K.; Nguyen, L.; Turner, E. B.;

Bhansali, S.; Malik, A. J. Chromatogr. A 2009, 1216, 5449.

21. Shearrow, A. M.; Bhansali, S.; Malik, A. J. Chromatogr. A 2009, 1216, 6349.

Chapter 6

1. DOE/EIA. International Energy Annual 2005; Energy Information Administration:

Washington, DC, 2005.

2. Arthur, C. L.; Pawliszyn, J. Anal. Chem. 1990, 62, 2145-2148.

3. Arthur, C. L.; Killam, L. M.; Motlagh, S.; Lim, M.; Potter, D. W.; Pawliszyn, J.

Environ. Sci. Technol. 1992, 26, 979-983.

4. Arthur, C. L.; Killam, L. M.; Buchholz, K. D.; Pawliszyn, J.; Berg, J. R. Anal. Chem.

1992, 64, 1960-1966.

5. Chai, M.; Arthur, C. L.; Pawliszyn, J. Analyst 1993, 118, 1501-1505.

6. Chai, M.; Pawliszyn, J. Environ. Sci. Techno. 1995, 29, 693-701.

7. Martos, P. A.; Pawliszyn, J. Anal. Chem. 1998, 70, 2311-2320.

8. Koziel, J.; Jia, M.; Khaled, A.; Noah, J.; Pawliszyn, J. Anal. Chim. Acta. 1999, 400,

153-162.

9. Odziemkowski, M.; Koziel, J. A.; Irish, D. E.; Pawliszyn, J. Anal. Chem. 2001, 73,

3131-3139.

225

10. Martos, P. A.; Pawliszyn, J. Anal. Chem. 1997, 69, 206-215.

11. Martos, P. A.; Saraullo, A., Pawliszyn J. Anal. Chem. 1997, 69, 402-408.

12. Chien, Y.-C.; Uang, S.-N.; Kuo, C.-T.; Shih, T.-S.; Jen, J.-F. Anal. Chim. Acta. 2000,

419, 73-79.

13. Smith, P. A.; Sheely, M. V.; Kluchinsky Jr., T. A. J. Sep. Sci. 2002, 25, 917-921.

14. Martos, P. A.; Pawliszyn, J. Anal. Chem. 1999, 71, 1513-1520.

15. Bartelt, R. J.; Zilkowski, B. W. Anal. Chem. 1999, 71, 92-101.

16. Gdreckit, T.; Pawliszyn, J. Anal. Chem. 1995, 67, 3265-3274.

17. Koziel, J.A.; Odziemkowski, M.; Pawliszyn, J. Anal. Chem. 2001, 73, 47-54.

18. Augusto, F.; Koziel, J., Pawliszyn, J. Anal. Chem. 2001, 73, 481-486.

19. Wasserscheid, P.; Welton, T. Ionic Liquids in Synthesis; Wiley-VCH: New York, 2003.

20. Davis, J. H., Jr. Chem. Lett. 2004, 9, 1072.

21. Liu, J. F.; Jiang, G. B.; Chi, Y. G.; Cai, Y. Q.; Zhou, Q. X.; Hu, J. T. Anal. Chem. 2003,

75, 5870-5876.

22. Yao, C.; Pitner, W. R.; Anderson, J. L. Anal. Chem. 2009, 81, 5054-5063.

23. Yao, C.; Anderson, J. L. Anal. Bioanal. Chem. 2009, 395, 1491-1502.

24. Liu, J.-F.; Li, N; Jiang, G.-B.; Liu, J.-M.; Jönsson, J. Å.; Wen, M.-J. J. Chromatogr. A

2005, 1066, 27-32.

25. Hsieh, Y.-N.; Huang, P.-C.; Sun, I.-W.; Whang, T.-J.; Hsu, C.-Y.; Huang, H.-H.; Kuei,

C.-H. Anal. Chim. Acta. 2006, 557, 321-328.

26. Huang, K.-P.; Wang, G.-R.; Huang, B.-Y.; Liu, C.-Y. Anal. Chim. Acta. 2009, 645,

42-47.

226

27. He, Y.; Pohl, J.; Engel, R.; Rothman, L.; Thomas, M. J. Chromatogr. A 2009, 1216,

4824-4830.

28. Zhao, F.; Meng, Y.; Anderson, J. L. J. Chromatogr. A 2008, 1208, 1-9.

29. Meng, Y., Pino, V., Anderson, J. L. Anal. Chem. 2009, 81, 7107-7112.

30. Marcilla, R.; Blazquez, J. A.; Rodriguez, J.; Pomposo, J. A.; Mecerreyes, D. J. Polym.

Sci. Part A: Polym. Chem. 2004, 42, 208-212.

31. Fukumoto, K.; Yoshizawa, M.; Ohno, H. J. Am. Chem. Soc. 2005, 127, 2398-2399.

32. Cadena, C.; Anthony, J. L.; Shah, J. K.; Morrow, T. I.; Brennecke, J. F.; Maginn, E. J.

J. Am. Chem. Soc. 2004, 126, 5300-5308.

33. Bates, E. D.; Mayton, R. D.; Ntai, I.; Davis, J. H., Jr. J. Am. Chem. Soc. 2002, 124,

926-927.

34. Soutullo, M. D.; Odom, C. I.; Wicker, B. F.; Henderson, C. N.; Stenson, A. C.; Davis,

J. H, Jr. Chem. Mater. 2007, 19, 3581-3583.

35. Gutowski, K. E.; Maginn, E. J. J. Am. Chem. Soc. 2008, 130, 14690-14704.

Chapter 7

1. C.L. Arthur, J. Pawliszyn, Anal. Chem. 62 (1990) 2145.

2. C.L. Arthur, L.M. Killam, S. Motlagh, M. Lim, D.W. Potter, J. Pawliszyn, Environ. Sci.

Technol. 26 (1992) 979.

3. C.L. Arthur, L.M. Killam, K.D. Buchholz, J. Pawliszyn, J.R. Berg, Anal. Chem. 64

(1992) 1960.

227

4. J.-F. Liu, N. Li, G.-B. Jiang, J.-M. Liu, J. Å. Jo¨nsson, M.-J. Wen, J. Chromatogr. A

1066 (2005) 27.

5. Y. -N. Hsieh, P.-C. Huang, I.-W. Sun, T.-J. Whang, C.-Y. Hsu, H.-H. Huang, C.-H. Kuei,

Anal. Chim. Acta. 557 (2006) 321.

6. K.-P. Huang, G.-R. Wang, B.-Y. Huang, C.-Y. Liu, Anal. Chim. Acta. 645 (2009) 42.

7. Y. He, J. Pohl, R. Engel, L. Rothman, M. Thomas, J. Chromatogr. A 1216 (2009) 4824.

8. E. Wanigasekara, S. Perera, J.A. Crank, L. Sidisky, R. Shirey, A. Berthod, D.W.

Armstrong, Anal. Bioanal. Chem. 396 (2010) 511.

9. F. Zhao, Y. Meng, J.L. Anderson, J. Chromatogr. A 1208 (2008) 1.

10. Y. Meng, V. Pino, J.L. Anderson, Anal. Chem. 81 (2009) 7107.

11. Q. Zhao, J.C. Wajert, J.L. Anderson, Anal. Chem. 82 (2010) 707.

12. T. Welton, Chem. Rev. 99 (1999) 2071.

13. J.H. Davis Jr, Chem. Lett. 9 (2004) 1072.

14. E.D. Bates, R.D. Mayton, I. Ntai, J.H. Davis Jr, J. Am. Chem. Soc. 124 (2002) 926.

15. M.D. Soutullo, C.I. Odom, B.F. Wicker, C.N. Henderson, A.C. Stenson, J.H. Davis Jr.

Chem. Mater. 19 (2007) 3581.

16. J. Pawliszyn, Solid Phase Microextraction: Theory and Practice, Wiley-VCH, New

York, 1997.

17. P.A. Martos, J. Pawliszyn, Anal. Chem. 69 (1997) 206.

18. D. Aaron, C. Tsouris, Sep. Sci. Technol. 40 (2005) 321.

19. M. Chai, J. Pawliszyn, Environ. Sci. Technol. 29 (1995) 693.

20. A.E. Messner, D.M. Rosie, P.A. Argabright, Anal. Chem. 31 (1959) 230.

228

Chapter 8

1. Welton T. Room-temperature ionic liquids. Solvents for synthesis and catalysis. Chem

Rev 1999;99:2071-2083.

2. Hallett JP, Welton T. Room-temperature ionic liquids: Solvents for synthesis and

catalysis 2. Chem Rev 2011;111:3508-3576.

3. Poole CF, Poole SK. Extraction of organic compounds with room temperature ionic

liquids. J Chromatogr A 2010;1217:2268-2286.

4. Martinis EM, Berton P, Monasterio RP, Wuilloud RG. Emerging ionic liquid-based

techniques for total-metal and metal-speciation analysis. TrAC-Trend Anal Chem

2010;29:1184-1201.

5. Francisco M, Arce A, Soto Ana. Ionic liquids on desulfurization of fuel oils. Fluid

Phase Equilibr 2010;294:39-48.

6. Liu R, Liu J-F, Yin Y-G, Hu X-L, Jiang G-B. Ionic liquids in sample preparation. Anal

Bioanal Chem 2009;393:871-883.

7. Aguilera-Herrador E, Lucena R, Cardenas Soledad, Valcarcel M. The roles of ionic

liquids in sorptive microextraction techniques. TrAC-Trend Anal Chem

2010;29:602-616.

8. Zhao Q, Anderson JL. Task-specific microextraction using ionic liquids. Anal Bioanal

Chem 2011;400:1613-1618.

9. Shiddiky MJA, Torriero AAJ. Application of ionic liquids in electrochemical sensing

systems. Biosens Bioelectron 2011;26:1775-1787.

229

10. Walsh DA, Lovelock KRJ, Licence P. Ultramicroelectrode voltammetry and scanning

electrochemical microscopy in room-temperature ionic liquid electrolytes. Chem Soc

Rev 2010;39:4185-4194.

11. Yao C, Anderson JL. Retention characteristics of organic compounds on molten salt

and ionic liquid-based gas chromatography stationary phases. J Chromatogr A

2009;1216:1658-1712.

12. Han X, Armstrong DW. Ionic liquids in separations. Acc Chem Res

2007;40:1079–1086.

13. Stalcup AM, Cabovska B. Ionic liquids in chromatography and capillary

electrophoresis. J Liq Chromatogr 2004;27:1443-1459.

14. Ma Z, Yu J-H, Dai S. Preparation of inorganic materials using ionic liquids. Adv

Mater 2010;22:261-285.

15. Baudequin C, Baudoux J, Levillain J, Cahard D, Gaumont A-C, Plaquevent J-C. Ionic

liquids and chirality: Opportunities and challenges. Tetrahydron: Asymmetry

2003;14:3081-3093.

16. Toma S, Mecirova M, Sebesta R. Are ionic liquids suitable media for organocatalytic

reactions? Eur J Org Chem 2009;321-327.

17. Chauvin Y, Mussmann L, Olivier H. A novel class of versatile solvents for two-phase

catalysis: Hydrogenation, isomerization, and hydroformylation of alkenes catalyzed

by rhodium complexes in liquid 1,3-dialkylimidazolium salt. Angew Chem Int Ed

Engl 1995;34:2698–2700.

18. Earle MJ, McCormac PB, Seddon KR. Diels–Alder reactions in ionic liquids. Green

Chem. 1999;1:23 – 25.

230

19. Gausepohl R, Buskens P, Kleinen J, Bruckmann A, Lehmann CW, Klankermayer J,

Leitner W. Highly Enantioselective Aza-Baylis–Hillman Reaction in a Chiral Reaction

Medium. Angew. Chem. Int. Ed. 2006;45:3689-3692.

20. Luo S, Mi X, Zhang L, Liu S, Xu H, Cheng J-P. Functionalized chiral ionic liquids as

highly efficient asymmetric organocatalysts for michael addition to nitroolefins.

Angew. Chem. Int. Ed. 2006;45:3093-3097.

21. Blanchard LA, Brennecke JF. Recovery of organic products from ionic liquids using

supercritical carbon dioxide. Ind Eng Chem Res 2001;40:287-292.

22. Hawthorne SB. Analytical-scale supercritical fluid extraction. Anal Chem

1990;62:633A–642A.

23. Mizushima E, Hayashi T, Tanaka M. Palladium-catalysed carbonylation of aryl

halides in ionic liquid media: High catalyst stability and significant rate-enhancement

in alkoxycarbonylation. Green Chem 2001;3:76-79.

24. Sanders MW, Wright L, Tate L, Fairless G, Chrowhurst L, Bruce NC, Walker AJ,

Hembury GA, Shimizu S. Unexpected preferential dehydration of artemisinin in ionic

liquids. J Phys Chem A 2009;113:10143-10145.

25. Arthur CL, Pawliszyn J. Solid phase microextraction with thermal desorption using

fused silica optical fibers. Anal Chem 1990;62:2145–2148.

26. Kim K-W, Song B, Choi M-Y, Kim M-J. Biocatalysis in ionic liquids: Markedly

enhanced enantioselectivity of lipase. Org Lett 2001;3:1507-1509.

27. Baleizão C, Gigante B, Garcia H, Corma A. Ionic liquids as green solvents for the

asymmetric synthesis of cyanohydrins catalysed by VO(salen) complexes. Green

Chem 2002;4:272–274.

231

28. Ngo HL, Hu A, Lin W. Highly enantioselective catalytic asymmetric hydrogenation

of β-keto esters in room temperature ionic liquids. Chem Commun 2003;1912–1913.

29. Anderson JL, Ding J, Welton T, Armstrong DW. Characterizing ionic liquids on the

basis of multiple solvation interactions. J. Am. Chem. Soc. 2002;124:14247-14254.

30. Zhao F, Meng Y, Anderson JL. Polymeric ionic liquids as selective coatings for the

extraction of esters using solid-phase micorextraction. J Chromatogr A 2008;1208:1-9.

31. Zhao Q, Wajert JC, Anderson JL. Polymeric ionic liquids as CO2 selective sorbent

coatings for solid-phase microextraction. Anal Chem 2010;82:707-713.

32. Louch D, Motlagh S, Pawliszyn J. Dynamics of organic compound extraction from

water using liquid-coated fused silica fibers. Anal Chem 1992;64:1187-1199.

232

Appendix A

Supplemental Figures and Tables Accompanying

Chapter 2

233 Table A.1: Probe molecules and their corresponding solute descriptors employed in the study of the IL stationary phases using the solvation parameter model.

234 Table A.1 (continued) Probe molecule E S A B L Pyridine 0.631 0.84 0.00 0.52 3.022 Pyrrole 0.613 0.73 0.41 0.29 2.865 Toluene 0.601 0.52 0.00 0.14 3.325 m-Xylene 0.623 0.52 0.00 0.16 3.839 o-Xylene 0.663 0.56 0.00 0.16 3.939 p-Xylene 0.613 0.52 0.00 0.16 3.839 2-Propanol 0.212 0.36 0.33 0.56 1.764 2-Nitrophenol 1.015 1.05 0.05 0.37 4.760 1-Bromohexane 0.349 0.40 0.00 0.12 4.130 Propionic acid 0.233 0.65 0.60 0.45 2.290 1-Decanol 0.191 0.42 0.37 0.48 5.628 Data obtained from ref [30].

235

Figure A-1: 1H-NMR spectrum of 1-butyl-1-methylpyrrolidinium FAP

236

Figure A-2: 1H-NMR spectrum of 1-(6-aminohexyl)-1-methylpyrrolidinium FAP

237

Figure A-3: 1H-NMR spectrum of 1-ethoxycarbonylmethyl-1-methylpyrrolidinium FAP

238

Figure A-4: 1H-NMR spectrum of 1-(2-hydroxyethyl)-1-methylpyrrolidinium FAP

239

Figure A-5: 1H-NMR spectrum of 1-hexyl-3-methylimidazolium FAP

240

Figure A-6: 1H-NMR spectrum of 1-(2-hydroxyethyl)-3-methylimidazolium FAP

241

Figure A-7: 1H-NMR spectrum of trihexyl(tetradecyl)phosphonium FAP

242

1 Figure A-8: H-NMR spectrum of 1-(2-hydroxyethyl)-1-methylpyrrolidinium NTf2

243

1 Figure A-9: H-NMR spectrum of 1-(2-hydroxyethyl)-3-methylimidazolium NTf2

244

Figure A-10: ESI-MS spectrum of 1-butyl-1-methylpyrrolidinium FAP

245

Figure A-11: ESI-MS spectrum of 1-(6-aminohexyl)-1-methylpyrrolidinium FAP

246

Figure A-12: ESI-MS spectrum of 1-ethoxycarbonylmethyl-1-methylpyrrolidinium FAP

247

Figure A-13: ESI-MS spectrum of 1-(2-hydroxyethyl)-1-methylpyrrolidinium FAP

248

Figure A-14: ESI-MS spectrum of 1-hexyl-3-methylimidazolium FAP

249

Figure A-15: ESI-MS spectrum of 1-(2-hydroxyethyl)-3-methylimidazolium FAP

250

Figure A-16: ESI-MS spectrum of trihexyl(tetradecyl)phosphonium FAP

251

Figure A-17: ESI-MS spectrum of 1-(2-hydroxyethyl)-1-methylpyrrolidinium NTf2

252

Figure A-18: ESI-MS spectrum of 1-(2-hydroxyethyl)-3-methylimidazolium NTf2

253

Appendix B

Supplemental Figures and Table Accompanying

Chapter 3

254 Table B.1: Probe molecules and their corresponding solute descriptors employed in the study of the IL stationary phases using the solvation parameter model.

255 Table B.1 (continued) Probe molecule E S A B L Pyridine 0.631 0.84 0.00 0.52 3.022 Pyrrole 0.613 0.73 0.41 0.29 2.865 Toluene 0.601 0.52 0.00 0.14 3.325 m-Xylene 0.623 0.52 0.00 0.16 3.839 o-Xylene 0.663 0.56 0.00 0.16 3.939 p-Xylene 0.613 0.52 0.00 0.16 3.839 2-Propanol 0.212 0.36 0.33 0.56 1.764 2-Nitrophenol 1.015 1.05 0.05 0.37 4.760 1-Bromohexane 0.349 0.40 0.00 0.12 4.130 Propionic acid 0.233 0.65 0.60 0.45 2.290 1-Decanol 0.191 0.42 0.37 0.48 5.628 Data obtained from ref [30].

256

Table B.2. Comparison of retention factors for selected solute molecules on stationary phases composed of ILs containing various cation types and NTf2ˉ anion at 80 °C.

Cation Ammonium Piperidinium

Probe molecule IL 10 IL 11

1-Chlorooctane 1.5 0.8

Phenol 48.2 70.8

p-Cresol 87.3 121.9

2-Pentanone 1.1 0.8

Cyclohexanone 6.2 4.8

Ethyl acetate 0.4 0.3

Methyl caproate 2.2 1.6

Octylaldehyde 4.5 3.3

N,N-DMF 15.7 11.6

Aniline 30.6 44.5

Benzene 0.3 0.4

Naphthalene 26.1 35.7

257

Table B.3. Comparison of selectivity for selected solute molecules on stationary phases composed of ILs containing various cation types and NTf2ˉ anion at 80 °C.

Cation Ammonium Piperidinium

Solute Pair IL 10 IL 11

Methyl caproate/benzene 2.4 1.9

Acetophenone/1-decanol 0.9a 1.1

N,N-DMF/o-xylene 6.6 4.8

1-Decanol/1-bromooctane 8.5 7.7

Benzyl alcohol/naphthalene 1.7 1.5

Naphthalene/o-xylene 10.7 14.0

Naphthalene/nitrobenzene 1.0 1.4

Cyclohexanol/1-butanol 2.7 2.1

1-Octanol/1-butanol 5.5 4.9

1-Decanol/1-butanol 16.1 13.5 a: By definition, the value of the separation factor should be greater than unity. However, some analytes exhibited reversal of elution order making it impossible to report integers greater than unity for all stationary phase compositions.

258

Figure B-1: 1H-NMR spectrum of N-hexylpyridinium FAP

259

Figure B-2: 1H-NMR spectrum of N-hexyl-4-(N´,N´-dimethylamino)pyridinium FAP

260

Figure B-3: 1H-NMR spectrum of N-hydroxypropylpyridinium FAP

261

Figure B-4: 1H-NMR spectrum of 1-ethyl-3-methylimidazolium FAP

262

Figure B-5: 1H-NMR spectrum of 1-methoxyethyl-3-methylimidazolium FAP

263

Figure B-6: 1H-NMR spectrum of methoxyethyl-dimethyl-ethylammonium FAP

264

Figure B-7: 1H-NMR spectrum of 1-methoxyethyl-1-methylmorpholinium FAP

265

Figure B-8: 1H-NMR spectrum of 1-methoxyethyl-1-methylpiperidinium FAP

266

Figure B-9: 1H-NMR spectrum of 1-methoxypropyl-1-methylpiperidinium FAP

267

1 Figure B-10: H-NMR spectrum of hexyl-trimethylammonium NTf2

268

1 Figure B-11: H-NMR spectrum of 1-propyl-1-methylpiperidinium NTf2

269

Figure B-12: 1H-NMR spectrum of 1-butyl-1-methylpyrrolidinium SCN

270

1 Figure B-13: H-NMR spectrum of 1-butyl-1-methylpyrrolidinium C(CN)3

271

1 Figure B-14: H-NMR spectrum of 1-butyl-1-methylpyrrolidinium B(CN)4

272

Figure B-15: 1H-NMR spectrum of 1-butyl-1-methylpyrrolidinium BOB

273

Figure B-16: ESI-MS spectrum of N-hexylpyridinium FAP

274

Figure B-17: ESI-MS spectrum of N-hexyl-4-(N´,N´-dimethylamino)pyridinium FAP

275

Figure B-18: ESI-MS spectrum of N-hydroxypropylpyridinium FAP

276

Figure B-19: ESI-MS spectrum of 1-ethyl-3-methylimidazolium FAP

277

Figure B-20: ESI-MS spectrum of 1-methoxyethyl-3-methylimidazolium FAP

278

Figure B-21: ESI-MS spectrum of methoxyethyl-dimethyl-ethylammonium FAP

279

Figure B-22: ESI-MS spectrum of 1-methoxyethyl-1-methylmorpholinium FAP

280

Figure B-23: ESI-MS spectrum of 1-methoxyethyl-1-methylpiperidinium FAP

281

Figure B-24: ESI-MS spectrum of 1-methoxypropyl-1-methylpiperidinium FAP

282

Figure B-25: ESI-MS spectrum of hexyl-trimethylammonium NTf2

283

Figure B-26: ESI-MS spectrum of 1-propyl-1-methylpiperidinium NTf2

284

Figure B-27: ESI-MS spectrum of 1-butyl-1-methylpyrrolidinium SCN

285

Figure B-28: ESI-MS spectrum of 1-butyl-1-methylpyrrolidinium C(CN)3

286

Figure B-29: ESI-MS spectrum of 1-butyl-1-methylpyrrolidinium B(CN)4

287

Figure B-30: ESI-MS spectrum of 1-butyl-1-methylpyrrolidinium BOB

288

Appendix C

Supplemental Figures Accompanying Chapter 4

289

Figure C-1: 1H-NMR spectra of 1-vinyl-3-hexylimidazolium chloride

290

Figure C-2: 1H-NMR spectra of 1-vinyl-3-hexadecylimidazolium chloride

291

Figure C-3: 1H-NMR spectra of poly(1-vinyl-3-hexylimidazolium chloride)

292

Figure C-4: 1H-NMR spectra of poly(1-vinyl-3-hexadecylimidazolium chloride)

293

Appendix D

Supplemental Figures Accompanying Chapter 6

294

Figure D-1: 1H-NMR spectra of 1-vinyl-3-hexylimidazolium bromide

295

Figure D-2: 1H-NMR spectra of 1-vinyl-3-hexylimidazolium taurate

296

Figure D-3: 1H-NMR spectra of poly(1-vinyl-3-hexylimidazolium taurate)

297

Appendix E

Supplemental Figures Accompanying Chapter 8

298

1 Figure E-1: H-NMR spectra of 1-butyl-3-methylimidazolium NTf2

299

1 Figure E-2: H-NMR spectra of poly(1-vinyl-3-hexylimidazolium NTf2)

300