ABSTRACT

Improved Spectrometry Characterization of Organic Compounds Using Analyte Preconcentration and Radio-Frequency Ionization

Abayomi D. Olaitan, Ph.D.

Mentor: Touradj Solouki, Ph.D.

Mass spectrometry (MS) is a powerful and sensitive analytical technique that can be used for analysis of organic compounds in diverse areas of research including biomedical, environmental, and forensic sciences. Detection and identification of volatile organic compounds (VOCs) in real world samples, which are often present at low concentrations and within complex mixtures, require the use of efficient sampling techniques and versatile ionization sources. Presented research projects focus on the use of analyte preconcentration and radio-frequency ionization to improve sensitive characterization of organic compounds with MS.

A home-built post-column cryogenic trap (PCCT) coupled to a commercial gas chromatography (GC)/quadrupole MS system was utilized for VOC analysis. Analytical figures of merit, including limit of detection (LOD) and signal-to-noise ratio, for

GC/PCCT/MS were evaluated and compared to “conventional” GC/MS (i.e., without

PCCT) using direct headspace and solid-phase microextraction techniques. An approximately two orders of magnitude improvement in LOD for the analysis of VOCs using GC/PCCT/MS over the use of conventional GC/MS was observed.

Radio-frequency ionization (RFI), a new and highly efficient ionization method discovered in our lab, coupled with Fourier transform-ion cyclotron resonance (FT-ICR)

MS is an ideal approach for analysis of volatile and semi-volatile organic compounds.

RFI FT-ICR MS was employed for sample fingerprinting and chemical characterization of organic compounds present in bio-oil and commercial gasoline samples. Experimental results, designed to study the effects of amplitude and frequency on ion generation yield, indicate that signal reproducibility is enhanced at higher frequencies

(for RF voltage amplitude between 175 Vp-p and 400 Vp-p). Moreover, experimental results indicate that the ionization threshold RF voltage amplitude in RFI is independent of either the analyte ionization energies or polarizabilities, suggesting that RFI is a universal ionization technique.

The possibility of electron emission in RFI, using both direct charged particle current measurements and indirect electron detection, was also studied. We show that

RF-generated electrons can be captured by high electron affinity molecules (e.g., hexafluorobenzene) in a 9.4 tesla FT-ICR mass . Results from the electron attachment rate constant for hexafluorobenzene, using a post-RF electron trapping FT-

ICR experiment, indicate that RFI process involves RF-generated electrons.

Improved Characterization of Organic Compounds Using Analyte Preconcentration and Radio-Frequency Ionization

by

Abayomi D. Olaitan, B.S.

A Dissertation

Approved by the Department of Chemistry and Biochemistry

Patrick J. Farmer, Ph.D., Chairperson

Submitted to the Graduate Faculty of Baylor University in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

Approved by the Dissertation Committee

Touradj Solouki, Ph.D., Chairperson

C. Kevin Chambliss, Ph.D.

Sascha Usenko, Ph.D.

Darrin Bellert, Ph.D.

Christopher M. Kearney, Ph.D.

Accepted by the Graduate School December 2015

J. Larry Lyon, Ph.D., Dean

Page bearing signatures is kept on file in the Graduate School.

Copyright © 2015 by Abayomi D. Olaitan

All rights reserved

TABLE OF CONTENTS

LIST OF FIGURES………………………………………………………………………ix

LIST OF SCHEMES…………………………………………………………………....xiii

LIST OF TABLES……………………………………………………………………....xiv

LIST OF ABBREVIATIONS……………………………………………………………xv

ACKNOWLEDGMENTS……………………………………………………………..xviii

DEDICATION…………………………………………………………………………...xx

CHAPTER ONE…………………………………………………………………………..1 Introduction………………………………………………………………………..1 1.1. Mass Spectrometry Analysis of Volatile Organic Compounds...... 1 1.2. Sensitivity Improvements in MS Analysis of VOCs..……………...…..2 1.3. Sample Introduction Methods for VOCs.………………………….….3 1.3.1. Membrane Introduction Mass Spectrometry……………….3 1.3.2. Cryogenic Trapping Systems……………………………….4 1.3.2.1. Post-Column Cryogenic Trapping………………..5 1.3.3. Sample Introduction Methods via the Gas Chromatograph……………………………………………………6 1.3.3.1. Static Headspace Analysis………………………..6 1.3.3.2. Dynamic Headspace Analysis………………….....7 1.3.3.3. Solid Phase Microextraction……………..…...…..7 1.4. Analysis of VOCs Using Gas Chromatography…….………………...9 1.4.1. Basic Operating Principle of GC………………………….11 1.5. Mass …………………………………………………13 1.5.1. Quadrupole Mass Spectrometer…………………………..15 1.5.1.1. Operating Principle of the Quadrupole……...….16 1.5.2. Fourier Transform-Ion Cyclotron Resonance Mass Spectrometry……………………………………...... 19 1.5.2.1. Operating Principle of FT-ICR MS……………..21

CHAPTER TWO………………………………………………………………………...25 Ionization Techniques for Volatile Organic Compounds (VOCs)……………....25 2.1. Ionization Sources for VOC Analysis…………….…………………25 2.1.1. ………………………………………...26 2.1.2. ……………………………………….29 2.1.3. Field Ionization……………………………………………31 2.1.4. Photoionization……………………………………………32 v

2.2. Radio-Frequency Ionization…………………...……………………33 2.3. Electron Emission…………………………………………………...36 2.3.1. Thermionic Emission……………………………………...37 2.3.2. Field Emission…………………………………………….37 2.3.3. Secondary Emission……………………………………….38 2.3.4. Photoelectric Emission…………………………………....39 2.4. Goal of the Research Presented in this Dissertation………………..39

CHAPTER THREE……………………………………………………………………...41 Coupling Post-Column Cryogenic Trapping with Conventional GC/MS for the Analysis of Volatile Organic Compounds………………………….…….41 Abstract…………………………………………………………………..41 3.1. Introduction………………………………………………………….42 3.2. Experimental………………………………………………………...44 3.2.1. Materials…………………………………………………..44 3.2.2. Sample Preparation…...…………………………………..45 3.2.2.1. Standard Acetone Samples……………………....45 3.2.2.2. Human Exhaled Breath (HEB) Samples………...47 3.2.3. Instrumentation……………………………………………47 3.2.3.1. Gas Chromatography (GC)……………………..47 3.2.3.2. Mass Spectrometry (MS)………………………...48 3.2.3.3. Post-Column Cryogenic Trap (PCCT) Assembly...... 49 3.2.3.4. GC/PCCT/MS Setup…………………………….50 3.3. Results and Discussion……………………………………………...52 3.3.1. Direct Headspace Analysis of Acetone……………………52 3.3.2. Analytical Sensitivity and Dynamic Range: GC/PCCT/MS vs GC/MS………………………….……………..54 3.3.3. Solid-Phase Microextraction (SPME) Analysis of Acetone at the ppbV Level………………………………….….…59 3.3.4. Solid-Phase Microextraction Analysis of Human Exhaled Breath Sample…………………………………………..61 3.3.5. Cryogenic Trap Coupled to a Catalytic Reaction Unit and MS System………………….…………………………..64 3.4. Conclusion…………………………………………………………..66 Acknowledgments.…………………………………………………….….67

CHAPTER FOUR………………………………………………………………………..68 Analysis of Volatile Organic Compound Mixtures Using Radio-Frequency Ionization/Mass Spectrometry……………...………………...68 Abstract…………………………………………………………………..68 4.1. Introduction………………………………………………………….68 4.2. Experimental………………………………………………………...70 4.2.1. Sample Preparation…...…………………………………..70 4.2.2. Mass Spectrometry (MS)…………………….…………….70 4.2.3. Radio-Frequency Ionization (RFI) Source………………..72

vi

4.2.4. Data Analysis……………………………………………...72 4.3. Results and Discussion……………………………………………...72 4.3.1. RFI/FT-ICR MS Analysis of Bio-Oil Samples……….……73 4.3.2. RFI/FT-ICR MS Analysis of Commercial Gasoline Samples……………………………….………………..77 4.4. Conclusions………………………………………………………….80 Acknowledgments……………….………………………………………..81 References………………………………………………………………..81

CHAPTER FIVE………………………………………………………………………...84 Ionization Yield as Functions of Electric Field Amplitude and Frequency in Radio-Frequency Ionization (RFI)……………………………………………….84 Abstract…………………………………………………………………..84 5.1. Introduction………………………………………………………….85 5.2. Experimental………………………………………………………...87 5.2.1. Samples……………………………………………………87 5.2.2. FT-ICR MS………………….……………………………..88 5.2.3. RFI Setup………………………………………………….88 5.3. Results and Discussion……………………………………………...89 5.3.1. Analyte Ionization Energy and Polarizability versus Threshold RF Voltage Amplitude…...…………….……....89 5.3.2. RFI MS Signal as a Function of Threshold RF Voltage Amplitude and Frequency…………...…………………..91 5.3.3. RFI MS Signal Stability as a Function of Frequency……..94 5.4. Conclusions………………………………………………...... 97 Acknowledgments………….……………………………………………..97

CHAPTER SIX…………………………………………………………………………..98 Evidence for Electron-Based Ion Generation in Radio-Frequency Ionization……………………………………………………...98 Abstract…………………………………………………………………..98 6.1. Introduction……………………………………………………….....99 6.2. Experimental……………………………………………………….101 6.2.1. Samples…………………………………………………..101 6.2.2. FT-ICR MS…………………….…………………………102 6.2.3. RFI Source……………………………………………….103 6.2.4. Current Measurements…………………………………...103 6.3. Results and Discussion…………………………………………….104 6.3.1. Charged Particle Current Measurement during RFI…....104 6.3.2. Indirect Detection of RF-Generated Electrons…………..108 6.3.3. Electron Attachment Reaction Kinetics of Hexafluorobenzene (HFB)……………………………………...113 6.3.4. Potential Mechanism(s) for Electron Generation in RFI……………………………………………...118 6.4. Conclusions………………………………………………………...119 Acknowledgments..…………………………………………….………..120

vii

References………………………………………………………………120

CHAPTER SEVEN…………………………………………………………………….126 Conclusions and Future Directions……………………………………………..126

APPENDIX ...…………………………………………………………………………..135

REFERENCES…………………………………………………………………………140

viii

LIST OF FIGURES

Figure 1.1. Schematic representation of a typical gas chromatography setup……….…..10

Figure 1.2. Schematic representation of the general components of a mass spectrometer...... 14

Figure 1.3. Schematic representation of the ………….……...16

Figure 1.4. Pictorial representation of the Fourier transform-ion cyclotron resonance cell. B: .……………….…………………………22

Figure 1.5. Schematic representation of ion detection in Fourier transform-ion cyclotron resonance mass spectrometry. FT: Fourier transform; m/z: mass-to-charge ratio…………………………………………..…….….23

Figure 2.1. Schematic representation of the electron ionization (EI) source….…………27

Figure 2.2. Schematic representation of the radio-frequency ionization (RFI) source. FTP: Filament trapping plate; ICR: Ion cyclotron resonance; QTP: Quadrupole trapping plate; TTL: Transistor- transistor-logic….…………………………………...…...... 34

Figure 3.1. Overlaid selected ion chromatograms (SICs) of acetone at m/z 43 + (CH3CO ) obtained using GC/PCCT/MS (solid line) and GC/MS (dotted line). A headspace (HS) volume of 5 µL containing 1.0 nmole of acetone was injected into the GC injection port in each instrumental configuration (i.e., GC/PCCT/MS or GC/MS). For comparison purpose, the peak labeled as “GC/MS” is magnified by 14 folds. The two parallel diagonal lines in x-axis denote axis break from 0.7 min to 6.7 min………..53

Figure 3.2. Calibration curves for acetone (between 42 pmoles and 6 nmoles) using GC/PCCT/MS (solid line) and GC/MS (dash lines) based on the (a) total peak areas (sum of the SIC peak areas at m/z 43 Th + + (CH3CO ) and 58 Th (M• )) and (b) peak height for SIC at m/z 43 Th + (CH3CO ). The data points shown with filled square (■) and empty circle (○) symbols correspond to GC/PCCT/MS and GC/MS experiments, respectively. Each data point in the panels (a) and (b) corresponds to an average of eight experimental trials and the absolute errors at the 95 % CL are denoted with the error bars………...... 56

ix

+ Figure 3.3. Selected ion chromatograms (SICs) of acetone at m/z 43 (CH3CO ) obtained from SPME analyses of samples containing (a) 3.8 ppbV and (b) 0.80 ppbV of acetone using GC/MS (dotted line) and GC/PCCT/MS (solid line). The values on top of the peaks are the average S/N values for the SIC peak maxima at m/z 43 from triplicate experiments (the error at 95 % CL is included in parenthesis for each value). In panels (a) and (b), the parallel diagonal lines in x-axes denote the axis break from 0.7 min to 6.7 min…………………………...…60

Figure 3.4. Total ion chromatograms (TICs) (after background correction) obtained from the analysis of a HEB sample using (a) GC/MS and (b) GC/PCCT/MS instrumental setups. In panels (a) and (b), the parallel diagonal lines in x-axes denote the axis break from 2 min to 7 min…………………………………………………………...…..62

Figure 3.5. Temporal plot of normalized ion intensities for NO, N2O, and NO2 with TiO2 P25 catalyst...…………………………………………………….65

Figure 4.1. RFI/FT-ICR mass spectra of VOCs from (a) aqueous and (b) oily phases of bio-oil derived from slow pyrolysis of pine shavings (PS). The two most abundant peaks in each are labeled. For each mass spectrum, 5 FT-ICR MS time-domain transients of 128 k data points were summed prior to Fourier transformation.…………...74

Figure 4.2. Expanded views of m/z range 100.9 to 101.2 for RFI/FT-ICR mass spectra of (a) aqueous and (b) oily phases of PS bio-oil. For each mass spectrum, 100 FT-ICR MS time-domain transients of 512 k data points were summed prior to Fourier transformation…………….…………76

Figure 4.3. RFI/FT-ICR mass spectra of commercial gasoline samples with RONs of (a) 87, (b) 89, and (c) 93. Four most abundant peaks in each mass spectrum are labeled. For each mass spectrum, 5 FT-ICR MS time-domain transients of 128 k data points were summed prior to Fourier transformation…………...... 78

Figure 4.4. (a) Score and (b) loading plots obtained from PCA of RFI/FT-ICR mass spectra of three gasoline grades. Each gasoline sample was analysed in 10 experimental trials. Data points represented by circle (○), triangle (∆), and square (□) symbols correspond to gasoline samples with RONs of 87, 89, and 93, respectively…………….....79

Figure 5.1. Plots of threshold radio-frequency (RF) voltage (Vp-p) at 10.40 MHz versus (a) ionization energy (eV) and (b) polarizability (Å3). Ionization energy and polarizability values were obtained (for the standard samples) from NIST chemistry webbook (webbook.nist.gov/chemistry/). Error bars are reported at the 95 % CL and n= 3……..………………………..90-91

x

Figure 5.2. Plot of threshold radio-frequency (RF) voltage (Vp-p) versus frequency (applied to QIG rods) for acetone performed in three separate experimental trials and error bars are at 95 % CL…………………92

Figure 5.3. Plots of total ion intensity (%) versus radio-frequency (RF) voltage (Vp-p) (applied to QIG rods) for acetone at (a) 10.5 MHz and (b) 5.8 MHz. Error bars are reported at the 95 % CL and n= 3……….…93-94

Figure 5.4. Plots of acquisition time versus positive-ion mode RFI/FT-ICR MS total ion intensity of acetone at RFI frequencies of (a) 11.6 MHz (360 Vp-p) and (b) 5.3 MHz (360 Vp-p) applied to dipole RFI electrodes. Insets in (a) and (b) show RFI/FT-ICR mass spectra of acetone acquired at elapsed time of 100 and 24 minutes, respectively. Relative standard deviation (RSD) for ion generation processes is lower at higher RFI frequency. Errors are reported at the 95 % CL and n= 3.…………………………………………………....96

Figure 6.1. Temporal plots of (a) positive (fifty measurements) and (b) electron/ negative (six) ion currents measured on QTP of the ICR cell at an RF signal of 6.5 MHz (200 Vp-p). Measurement interval for both panels (a) and (b) are 10 s…………………….……………………….105-106

Figure 6.2. Negative-ion mode post-RF FT-ICR mass spectra of hexafluorobenzene after (a) 0 ms, (b) 5 ms, and (c) 13 ms of electron (e-) quenching. The * symbols represent electronic noise….…….111

Figure 6.3. Plot of total ion intensity as a function of post-electron injection quenching time for hexafluorobenzene. Error bars are reported from three experimental trials at the 95 % confidence level……………..………112

●- Figure 6.4. Plot of normalized intensity for C6F6 as a function of reaction time with RF-generated electrons. Average reaction rate constant is reported from a total of four measurements (each measurement was based on a kinetic plot constructed from ~29 data points) and error is reported at the 95 % confidence level…………………………………………..………116

Figure 7.1. (a) RFI FT-ICR mass spectrum of mass spectrometer background vacuum chamber hydrocarbon species obtained at RF signal of 12.2 MHz (400 Vp-p). (b) Picture of electron/ion burn spots on the filament trapping plate of the ICR cell obtained over ~6 months of RFI use…………………………………………………………………..129

Figure 7.2. (a) SIMION captured picture for simulated ion trajectories. Magnetic field, 500 gauss; Repeller, -20 V; RFI electrodes’ DC offset, 10 V; Lens 1, -160 V; Lens 2, -20 V…………………………………..…………130

xi

Figure 7.3. Plots showing simulated ion transmission efficiency as a function of (a) frequency (from 1 to 20 MHz at 200 Vp-p) and (b) RF amplitude/voltage (from 0 to 500 Vp-p at 12.5 MHz) for m/z of 28, 69, 131, 219, and 502…………………………………………………..131

Figure 7.4. (a) AutoCad drawing of RFI source showing the dipole electrodes and the associated ion lenses for improved ion trajectory control and transmission. (b) Picture (side view) of the constructed RFI source…..…..133

Figure A.1. RFI/FT-ICR mass spectra of VOCs from (a) aqueous and (b) oily phases of corn stover (CS) slow pyrolysis bio-oil. The two most abundant peaks in each mass spectrum are labeled. For each mass spectrum, 5 FT-ICR MS time-domain transients of 128 k data points were summed prior to Fourier transformation……………………...……..136

xii

LIST OF SCHEMES

Scheme 3.1. Print screen of the GUI window (“Cryo Controller”) from the home-written software used to control the PCCT unit. The values shown for the “Cryo Controller” parameters are the actual adjusted values used to acquire the GC/PCCT/MS chromatograms in Figures 3.1, 3.3, and 3.4………………...... 50

Scheme 3.2. Schematic representation of the GC/PCCT/MS instrumental setup used for the VOC analyses……………………………………………….....51

Scheme 6.1. Sequence of events for electron (e-) quenching experiments and e- attachment rate constant measurements…………………………………109

xiii

LIST OF TABLES

Table 3.1. Comparisons of the signal-to-noise ratio (S/N) for the direct analysis of acetone using GC/MS and GC/PCCT/MS………………………………...58

Table 4.1. Percentages of oxygenated and non-oxygenated compounds in PS and CS bio-oil aqueous and oily phases and a commercial gasoline sample (RON 87)……………………………………………………………..76

Table A.1. List of the observed and theoretical (exact) m/z values, the mass measurement errors, assigned chemical compositions, and the double bond equivalents (DBEs) for the ionic species observed in the RFI/FT-ICR mass spectra of pine shavings (PS) and corn stover (CS) bio-oil aqueous and oily phases as well as a commercial gasoline sample (RON = 87)…………………………………………………….…...137

xiv

LIST OF ABBREVIATIONS

Abbreviation Description 2-D Two-Dimensional AC Alternating Current APCI Atmospheric Pressure Chemical Ionization APGC Atmospheric Pressure Gas Chromatography AU Arbitrary Unit CI Chemical Ionization CL Confidence Level CS Corn Stover CTE Cryogenic Trap Element DBE Double Bond Equivalent DC Direct Current e- Electron ECD Electron capture detector EI Electron Ionization EM ESI FAB FD FI Field Ionization FID Flame Ionization Detector FT Fourier Transform FT-ICR Fourier Transform-Ion Cyclotron Resonance FTMS Fourier Transform Mass Spectrometry FTP Filament Trapping Plate GC Gas Chromatography GF Geometry Factor GUI Graphical User Interface HS Headspace HEB Human Exhaled Breath

xv

LIST OF ABBREVIATIONS

Abbreviation Description ICR Ion Cyclotron Resonance IE Ionization Energy IM Ion Mobility IS Ionization Sensitivity LC Liquid Chromatography LMCS Longitudinal Modulated Cryogenic System

LN2 Liquid Nitrogen LOD Limit of Detection M Molecule MALDI Matrix-Assisted Laser Desorption Ionization MFE Magnetic Field Effect MIMS Membrane Introduction Mass Spectrometry MPT Microscale Purge and Trap MS Mass Spectrometry MW Molecular Weight m/z Mass-to-Charge Ratio NCI Negative Chemical Ionization PA Affinity PC Preconcentrator PCA Principal Component Analysis PCCT Post-Column Cryogenic Trapping PDMS Polydimethylsiloxane PI Photoionization Ppb Parts-per-Billion PpbV Parts-per-Billion Volume Ppm Parts-per-Million Ppt Parts-per-Trillion PS Pine Shavings QIG Quadrupole Ion Guide QMS Quadrupole Mass Spectrometer

xvi

LIST OF ABBREVIATIONS

Abbreviation Description QTP Quadrupole Trapping Plate RF Radio-Frequency RFI Radio-frequency Ionization RMS Root-Mean-Square RON Research Octane Number RSD Relative Standard Deviation RT Retention Time SIC Selected Ion Chromatogram S/N Signal-to-Noise Ratio SPE Solid-Phase Extraction SPME Solid-Phase Microextraction TCD Thermal Conductivity Detector TIC Total Ion Chromatogram TOF Time-of-Flight TTL Transistor-Transistor Logic UHV Ultrahigh Vacuum UV Ultra-Violet VOC Volatile Organic Compound

xvii

ACKNOWLEDGMENTS

First, I would like to acknowledge the Almighty God for His grace and strength thus far. Without the Lord on my side, I wouldn’t be where I am today.

Next, my profound gratitude goes to my advisor, Dr. Touradj Solouki, for his supervision, support, encouragement, and patience shown to me throughout my graduate program. He was like a father to me. He provided me with guidance and the necessary tools to becoming a mass spectrometrist. He gave me the opportunity to learn and use state-of-the-art instrumentation, to meet with excellent and leading scientists in the field of mass spectrometry, and to improve myself in many areas including but not limited to scientific writing and communication. I also appreciate the privilege to move with him from the University of Maine to Baylor University. He is a great mentor and I would always remain indebted to him.

My appreciation also goes to my graduate committee members, Dr. C. Kevin

Chambliss, Dr. Sascha Usenko, Dr. Darrin Bellert, Dr. Stephen L. Gipson, and Dr.

Christopher Kearney for their assistance, encouragement, time/accessibility, and valuable comments provided to me during my graduate program. I also like to thank them for taking time to review my dissertation and providing me with helpful suggestions and directions.

I am grateful to Dr. Alejandro Ramirez for his assistance whenever I needed specific MS instrument component(s), and training on the mass spectrometers in the MS facility. I am also grateful to Dr. Kevin L. Shuford for his assistance and helpful

xviii discussions on theoretical modeling. My gratitude also goes to all the people to whom I worked with as collaborators including Dr. Patrick J. Farmer (chair of the Chemistry and

Biochemistry Department, Baylor University), Dr. William C. Hockaday (Geology

Department, Baylor University), Dr. Murugaeson R. Kumar, and Birendra Dhungana. I also like to thank Dr. Dennis H. Rabbe for assisting me with the teaching laboratory. And many thanks to the faculty, staff, and graduate students of the Department of Chemistry and Biochemistry, Baylor University, for their direct and indirect support during my graduate program. I also like to appreciate the University of Maine for proving me with financial support during my first year of graduate school.

My thankfulness also extends to the current and past members of my research group including Dr. Behrooz Zekavat, Dr. Mahsan Miladi, Dr. Lu Weigang, Ting Wang,

Brett Harper, Michael Pettit, Jackie Lochridge, Matthew R. Brantley, and Elizabeth

Neumann for providing a friendly and safe laboratory environment to work in. I do appreciate you all for your assistance and encouragement all these years in graduate school. I am particularly grateful to Dr. Behrooz Zekavat for his tremendous help, training, and brotherly support given to me during the majority of my graduate program.

Finally, I’m thankful to every member of my church (RCCG, Living Faith

Chapel, Waco, TX) for their support and prayers, my friends (here and abroad) for their encouragement, and my family and siblings (Mr. Francis Olaitan, Mrs. Fola Enahoro

Bernard, Mr. Dapo Olaitan, and Mrs. Tayo Olaitan-Kolawole) for their moral support, prayers, and unending love.

xix

DEDICATION

To my late parents, Mr. Adetunji Olaitan and Mrs. Adeola Olaitan, for their Godly parental guidance, love, and encouragement

xx

CHAPTER ONE

Introduction

1.1. Mass Spectrometry Analysis of Volatile Organic Compounds

Mass spectrometry (MS) is a valuable analytical tool employed for the analysis of gas-phase molecules or volatile organic compounds (VOCs) present in various sample mixtures.1-4 In addition to using MS for VOC analysis, a large group of non-volatile organic compounds such as proteins,5 lipids,6 carbohydrates,7 nucleic acids,8 and non- biological materials9 can also be detected, identified, and characterized with MS techniques. For the purpose of this dissertation, however, discussions will exclusively focus on MS analysis of VOCs (i.e., organic chemical compounds with low boiling points

(<100 ºC) at standard atmospheric pressure (i.e., 1 atm)).10 Mass spectrometry is employed for VOC analysis in a wide range of areas including, but not limited to, biomedical,2, 11-14 environmental,1, 3, 4, 15 food and flavor,16, 17 forensic,18, 19 and pharmaceutical20, 21 sciences. The majority of the commonly analyzed VOCs are present at low concentrations (e.g., VOCs in human breath, soil, air and water have been reported to be in the parts-per-million (ppm) to parts-per-trillion (ppt) or lower concentration ranges).1, 3, 14, 22-24 Hence, VOC detection, analysis, and characterization often require sensitive analytical techniques.25, 26

1

1.2. Sensitivity Improvements in MS Analysis of VOCs

Mass spectrometry is a sensitive analytical method for the detection and analysis of VOCs.27, 28 Sensitivity can be defined as the ability to quantitatively detect small amounts of analytes within limited amounts of sample or the ability to detect low concentrations of analytes within a large volume of sample.28 One of the important objectives of MS analysis is to detect analytes present at low concentrations within a limited or large sample volume. The factors that limit or determine sensitivity in MS are related to three main areas including: (i) ionization sources, (ii) mass analyzers and detectors, and (iii) sample preparation methods.24, 26, 28

Several research efforts have been devoted to improving sensitivity in MS by developing more efficient ionization sources,29, 30 improved ion transmission components,31, 32 and high performance mass analyzers and detectors.28, 33 The choice of ionization method depends on the characteristics (i.e., polar/non-polar, molecular weight, thermal stability, etc.) and state (i.e., solid, liquid, gas) of the sample to be analyzed. For

VOCs or gas-phase molecules, conventional ionization methods such as electron ionization (EI)34 and chemical ionization (CI)35 are often employed. Experimental parameters that may influence the efficiency of ionization (such as temperature, pressure, and various electrode cleanliness and lens voltages) are carefully controlled and/or optimized to enhance instrumentally achievable analyte sensitivities.

Extracting ions generated in the ionization source requires the use of carefully designed extraction and focusing lenses to minimize ion losses as the ions travel into the mass analyzer.36 Mass analyzers and detectors are constantly being developed and improved to enhance sensitivity, speed, linear dynamic range, mass accuracy, and reliability in order

2 to maximize ion separation and ion detection abilities.33 Sample enrichment methods, such as the use of membrane/sorbent extraction37-39 and analyte preconcentration techniques39, 40 are also being continuously developed to improve the sensitivity for MS analysis of VOCs.

1.3. Sample Introduction Methods for VOCs

Efficient sample introduction into the ion source of a mass spectrometer can improve sensitivity for VOC detection during MS analysis. For most VOC analyses, samples are introduced into the ion source either directly (e.g., using membrane introduction mass spectrometry (MIMS) devices1, 22, 37 or cryogenic trapping systems40-42) or indirectly (e.g., using a gas chromatography (GC) system coupled to the mass spectrometer).2, 43 Commonly employed GC sample injection methods include headspace

(HS) analysis, including static3, 44, 45 and dynamic or purge and trap4, 23 methods as well as stationary phase utilization13, 14, 19 (to concentrate VOC injection).

1.3.1. Membrane Introduction Mass Spectrometry

Membrane introduction involves the use of a silicone polymer for selective introduction of organic compounds into the inlet of a mass spectrometer.22, 37 The first application of a membrane to concentrate and introduce vapors into a mass spectrometer ion source was reported in 1963 by G. Hock and B. Kok.46 In MIMS, the membrane enhances the rapid and simultaneous introduction of the VOCs adsorbed onto the membrane into the vacuum region of the mass spectrometer ion source.37 Extraction, concentration, and injection of the adsorbed VOCs are performed in one single step.37

The main advantages of MIMs are: a relatively low detection limit (in the low parts-per-

3 billion (ppb) levels) for VOCs, rapid response time (~8 s), ease of automation, simplicity, and adaptability to most MS instruments.22, 37 However, the presence of high background signals resulting from thermal decomposition of the membrane material, produced during the heating cycle, and the irreproducibility of MS data limit the sensitivity and application of the technique.22 Qualitative and quantitative analyses of environmental chemical contaminants,47 especially in water3, 48 and air1, 48 samples, are some of the areas where MIMS is presently employed.

1.3.2. Cryogenic Trapping Systems

Cryogenic trapping systems are utilized for preconcentrating VOCs prior to their introduction either into the GC column or directly into the inlet of a mass spectrometer.40,

44, 49, 50 Numerous cryogenic trapping techniques have been coupled directly to GC systems to introduce analytes in a “narrow band” (i.e., quick and concentrated burst of sample) onto the GC column.40, 44, 51 Major advantages associated with concentrating analytes into a small zone prior to introduction onto the analytical GC column include separation enhancement, achievement of narrower chromatographic peak shapes, and sensitivity improvement.41, 44, 49 A commonly used cryogen for analyte preconcentration is liquid nitrogen (LN2) which enables the trapping of VOCs at temperatures below -170

39, 44 ºC. Analyte VOCs are trapped in a metal tube immersed in a LN2 bath for an adjustable period of time (depending on the analytical method) and are subsequently desorbed onto the analytical GC column or inlet of the mass spectrometer. Desorption of the trapped VOCs usually occurs via resistive heating50 (to temperatures in excess of 150

51 ºC) of the metal tube immersed in the LN2 bath.

4

Conventional two-dimensional (2-D) GC systems utilize a form of cryogenic trapping device (i.e., cryogenic modulator) inserted between two GC columns to enhance analytical separation of organic compounds.52, 53 Analytes eluting from the first GC column are trapped in a capillary tube (cooled cryogenically by liquid carbon dioxide

(LCO2) or LN2), focused in a narrow band, and subsequently transferred onto the second

GC column.52-54 The use of a longitudinally modulated cryogenic system (described by P.

Marriott and R. Kinghorn)51 has been interfaced to a detector (i.e., flame ionization detector (FID)) with enhanced sensitivity and lower detection limits for analysis of hydrocarbon molecules.54, 55 An advantage of this technique is improved resolution (i.e., sharper elution profiles through the narrowing of chromatographic bands) for separation of VOCs in complex samples.52-54

Cryogenic systems utilizing a Freon® refrigerator and/or thermoelectric heat pump to concentrate analytes, rather than using cryogens (e.g., LN2 or LCO2), have also been developed.41, 56 These cryogen-free methods of concentrating analytes reduce the cost and inconvenience associated with regular replenishing of the required liquid cryogens. Cryogenic trapping systems have been utilized for VOC monitoring in air,40, 41 water,39 human breath,50 and complex mixtures.57 Detection limits for GC/MS analysis using cryogenic trapping for atmospheric VOCs have been reported to be in the low ppt concentration levels.56

1.3.2.1. Post-column cryogenic trapping. Post-column cryogenic trapping

(PCCT) involves placing a cryogenic trap in-between the GC system and the inlet of the mass spectrometer.50 Volatile organic compounds are trapped in a tube/capillary column

(placed in a cryogenic liquid) and are subsequently desorbed into the mass spectrometer

5 ion source by resistive heating of the tube/capillary column. The design of the PCCT setup utilized in this dissertation was based on the original design of Jacoby et al. where a cryogenic trap was placed between a GC and Fourier transform-ion cyclotron resonance

(FT-ICR) MS system.58 Details of the coupling and instrumental operations of the PCCT system is given in Chapter Three of this dissertation. A similar PCCT setup has been previously demonstrated for use with GC FT-ICR MS systems57 for reducing the pressure load from the GC system into the FT-ICR mass spectrometer. Also, the use of PCCT coupled to a catalysis unit and triple quadrupole mass spectrometer for monitoring nitric oxide (NO) photocatalytic decomposition products was reported recently.59

1.3.3. Sample Introduction Methods via the Gas Chromatograph

One of the most common hyphenated systems coupled to MS for VOC analysis is the gas chromatograph.2, 4, 13 Samples are injected into the GC column and the GC effluents are subsequently introduced into the mass spectrometer ion source through a heated transfer line. To indirectly increase the overall amount of GC-eluting analytes that enter into the mass spectrometer’s ion source, several sampling techniques are commonly utilized, such as static and dynamic HS extraction4, 17, 23, 44 and solid-phase microextraction (SPME).11, 13, 14, 21

1.3.3.1. Static headspace analysis. Headspace (HS) analysis or vapor- phase extraction can be defined as an extraction technique involving the partitioning of analytes between a non-volatile solid or liquid-phase and the vapor-phase above the solid or liquid sample.45 The first explicit mention of vapor analysis above a liquid analyte was reported by R. N. Harger et al., in 1939 for the determination of alcohol content in

6 aqueous solutions.60 Also, static HS was first demonstrated with the GC system in 1958, where sample vapors placed within a gas-tight container were sampled using a gas-tight syringe.61 Static HS sampling is the primary technique for analysis of VOCs in several analytical areas of study, such as clinical,62 environmental,3 forensic,63 flavor and fragrance,64 and pharmaceutical.20 However, static HS sampling is limited in sensitivity and, currently, is only suitable for applications in the high ppb concentration ranges.45, 62-

64

1.3.3.2. Dynamic headspace analysis. Dynamic HS (commonly known as

“purge and trap”) analysis was first shown in 1973 for GC application immediately after the introduction of a commercial adsorbent called Tenax®.65 Dynamic HS sampling involves purging a liquid sample with a carrier gas (e.g., nitrogen (N2) or helium (He)), trapping the volatile analytes present in the purged liquid on a sorbent (e.g., deactivated fused-silica capillary) cryogenically, and finally desorbing the trapped volatiles onto the

GC column.4, 23, 45 The technique is routinely performed for analysis of VOCs, such as chlorinated hydrocarbons present at extremely low (ppb and ppt) concentrations in aqueous matrixes.4, 23, 24 Recent applications for dynamic HS analysis of VOCs are common in biomedical,66 environmental,24, 42 food,17 and human breath2 studies.

1.3.3.3. Solid-phase microextraction. Solid-phase microextraction (SPME) was introduced and developed by J. Pawliszyn and co-workers in the 1990’s.67, 68 SPME is currently one of the most utilized HS sampling techniques and has seen immense growth for use with GC applications.45 Headspace-SPME involves the use of a fused silica fiber coated with polymeric organic liquid/solid material (e.g., polydimethylsiloxane (PDMS)) for trapping and concentrating analytes from a HS volume.68 The extracted VOCs that are

7 concentrated in the coated fiber and transferred into the GC injection port for thermal desorption onto the GC column and subsequent GC/MS analysis.38, 45 The amount of absorbed analyte on the fiber coating is dependent on the overall equilibrium partition of the analyte in the liquid polymeric fiber coating, in the HS (containing the vapor phase of the analyte), and in the aqueous solution (containing the analyte).68

The amount of analyte absorbed by the fiber coating can thus be calculated using equation 1.168: n = C0V1V2K1 / (K1V1 + K2V3 + V2) (Equation 1.1) where n is the mass absorbed by the fiber coating; C0 is the initial concentration of the analyte in the aqueous solution; V1, V2, and V3 are the volumes of the fiber coating, the aqueous solution, and the HS, respectively; K1 is coating/gas partition coefficient; K2 is gas/water partition coefficient. A large K1 value indicates that a significant amount of analyte will be absorbed by the fiber coating even when C0 is low. K2 for most analytes is relatively small, because the volume of the HS is usually much less than the volume of the aqueous solution. Overall, the amount of analyte absorbed by the coating is governed by: (i) the rate at which analytes move from the aqueous phase to the HS and finally to the fiber coating, (ii) the equilibration time (i.e., time at which the mass absorbed by the fiber coating has reached 90 % of its final total mass), and (iii) sampling/extraction time

(determined by diffusion in the vapor phase).68 The chemical structure of the fiber coating is usually modified to enhance the affinity and selectivity for different analytes

(i.e., for analytes with different functional groups).20, 38

8

Solid-phase microextraction is generally applied to samples with concentrations in the low ppm to ppb range for several applications including clinical,11, 14, 38, 69 forensic,18, 19 environmental,70, 71 and pharmaceutical analyses.20, 21 The SPME technique has also been used to sample a wide range of organic compounds present in different sample matrixes such as air,72 water,39, 73 and soil.70 A major drawback in the SPME technique is longer sampling/extraction time (>30 mins) for analytes with low volatility20,

21, 68 and potential contamination of the GC injection port with chemicals from the fiber coating during the desorption process.45

1.4. Analysis of VOCs Using Gas Chromatography

Gas chromatography (GC) technique is one of the oldest and most commonly utilized analytical methods in chemistry for separating and analyzing volatile compounds.

The origin of GC can be linked to the publication on liquid-phase partition chromatography by A. J. P. Martin and R. L. M. Synge in 1941,74 although the contemporary GC technique was invented by A. J. P. Martin and A. T. James in 1952.75

Over the years, several applications have employed GC technique for separation of components in simple and complex mixtures for research, development, and quality control purposes in academia,76-78 biomedical,11, 13, 66, 79 drug and forensic,18, 19, 73 food analysis,16, 17, 64 environmental,4, 23, 71 petrochemical,57, 80, 81 and pharmaceutical20, 21 sciences.

The essential components of a GC system include a gas source (i.e., GC carrier gas), pressure regulator, pressure gauge, injection port (for injecting the sample with a syringe through a self-sealing rubber septum onto the GC column), GC column, an oven

(that houses the GC column), and a detector78 (see Figure 1.1 for illustration). GC

9 injection ports are usually heated to promote rapid volatility of the sample and to prevent zone broadening.78

GC Column

Injection

Port Detector

GC Oven

Pressure Heated Transfer Line Regulator Pressure Gauge He

GC Carrier Gas

Figure 1.1. Schematic representation of a typical gas chromatography setup.

Separations in GC are achieved through the partitioning between a stationary phase and a mobile phase. A stationary phase is composed of a column (packed or open tubular/capillary column) coated with a functional group(s) (e.g., PDMS).78 The nature of the coating functional group(s) determines the polarity of the stationary phase.

Compounds with the same or similar polarity as the stationary phase will spend more time in the GC column than a compound with opposite polarity. The mobile phase in GC is a constantly flowing inert gas (He, hydrogen (H2) or N2) for carrying the separated components through the GC column and to the detector. Commonly used detectors for

GC include flame ionization detector (FID),77 thermal conductivity detector (TCD),76 electron capture detector (ECD),82 and mass spectrometer.83 The choice of detector type depends upon the goal(s) of the analysis. For example, FID is suited for organic

10 compounds (i.e., those with hydrogen and carbon atoms), ECD is selective for halogens and nitrogen-containing compounds, while TCD and mass spectrometers are universal detectors.78, 84

1.4.1. Basic Operating Principle of GC

In GC, separation of analytes occurs as a function of the transport rate of the analyte through the GC column (i.e., time spent in the mobile phase) and retention in the stationary phase (based on the intermolecular interaction with the stationary phase).

Common intermolecular that exist between analytes and the stationary phase of the

GC include van der Waals forces (dipole-dipole, dipole-induced-dipole, and induced dipole-induced dipole), hydrogen bonding, and electron-donor-acceptor interactions.85

In chromatography, the time an analyte spends in the chromatographic column is called the retention time, which is a combination of the time spent in the stationary phase and time spent in the mobile phase. The time spent in the mobile phase is essentially the same for all analytes, but the time spent in the stationary phase is analyte specific. The distribution constant (K) is an expression of the distribution of an analyte between the stationary and mobile phases given by85:

K = Cs/Cm (Equation 1.2) where Cs is the concentration of an analyte in the stationary phase and Cm is the concentration of an analyte in the mobile phase. Analytes must have sufficiently different

K values to be effectively separated and an analyte with a large K value will spend more time in the stationary phase or, overall, will migrate slower through the column.

Retention times are influenced by column dimensions and GC operational parameters, such as the GC carrier gas flow rate and oven temperature.

11

Analytes can also be separated in the GC system based on their boiling points.

The differences in the boiling points of compounds can be manipulated to cause selective separation by varying the column temperature (a technique called temperature programming).85 Increasing the column temperature decreases the partition coefficients of analytes, hence, causing the retention time (i.e., the time it takes for an analyte to elute through the column at a given experimental condition) to decrease.85

In GC, peak broadening must be reduced to allow for efficient analyte separation.

Peak broadening in GC packed columns was first illustrated by J. J. van Deemter in his presentation on rate theory.86 The van Deemter equation describes the relationship between plate height (H)87 and the average linear velocity of the mobile phase (u)86:

H = A + B/u + Cu (Equation 1.3) where A is the eddy diffusion term (band broadening caused by dispersion or multi- pathway effects), B is the longitudinal diffusion term (i.e., molecular diffusion both with and against the carrier gas flow direction), and C is the mass transfer term (band broadening caused by analyte delay or resistance to mass transfer). As seen in equation

1.3, A is independent of u, B decreases as the u increases, and C increases with increase in u.

For capillary columns, an equation was developed by M. J. E. Golay, which excluded the A term (because capillary columns do not have packed particulates) and included a new term, CM, to account for band broadening caused by diffusion in the mobile phase88:

H = B/u + (CS + CM)u (Equation 1.4)

12

CS and CM represent mass transfer in the stationary and mobile phases, respectively. To achieve a small H (i.e., reduction in band broadening and hence, efficient separation), the

A, B, and C terms must be as low as possible. The A term can be decreased by using small, regular, and uniform column packing material(s); the B term can be reduced by increasing the linear velocity of the carrier gas; and the C term can be minimized by using stationary phases with thin films and low viscosity.

An advantage of GC, besides the separation of components of a complex mixture, is the ability to identify an unknown sample by comparing the retention time of the sample with that of a standard sample (under identical GC experimental conditions). GC is widely utilized because it is efficient, inexpensive, and easily coupled to MS.78 Recent applications of the GC technique are seen in the development of 2-D GC systems (i.e.,

GC × GC)84, 89, 90 and fast GC72, 78, 81 for efficient and robust separation of complex mixtures.

1.5. Mass Spectrometers

A mass spectrometer is an analytical tool that generally consists of an ion source

(for generation of gas-phase ions), a mass analyzer (for separating ions based on mass-to- charge ratio (m/z) using magnetic (B) and/or electric (E) or E×B fields), and a detector

(for counting the separated ions).27, 28 Other important components of a mass spectrometer are the sample introduction source (sample inlet), a vacuum system, and a data acquisition/analysis system.27 A schematic representation showing the general components of a mass spectrometer is illustrated in Figure 1.2.

13

Sample Ionization Mass Data Detector Inlet Source Analyzer System

Vacuum

Figure 1.2. Schematic representation of the general components of a mass spectrometer.

There are several ion sources for converting neutral molecules (that exist in either the gas- or condensed- (liquid/solid) phases) into ions. Examples of ionization techniques for gas-phase molecules include EI,91, 92 CI,35, 92 and field ionization (FI).92, 93 Common ionization techniques for analytes that exist in the condensed-phase are electrospray ionization (ESI),94, 95 atmospheric pressure chemical ionization (APCI),96, 97 matrix- assisted laser desorption ionization (MALDI),98, 99 field desorption (FD),100, 101 and fast atom bombardment (FAB).101 Sample introduction into the mass spectrometer occurs via direct vapor inlet system92 or GC system21 (for gas-phase molecules), direct infusion95

(for liquid-phase samples), and direct insertion probe/stage98, 101 (for low vapor pressure liquids and solids).

Ions formed in the ionization source are accelerated into the mass analyzer in order to separate them according to their m/z.27 The selection of a mass analyzer usually depends upon parameters, such as resolution, mass range, scan rate, and detection limits required for analysis.27, 28 Mass analyzers can be categorized into two major groups: spatial (or beam) and temporal (or trapping) analyzers.27, 28 For beam analyzers, ions leave the ionization source as a beam and pass through the analyzer and to the detector.27

Examples of beam analyzers include transmission quadrupole,102 sector (either magnetic and/or electric),103, 104 and time-of-flight105, 106 mass spectrometers. Trapping analyzers

14 trap ions prior to detection; examples include quadrupole ion trap107, orbitrap108, and FT-

ICR109 mass spectrometers. The separated ions are measured and detected based on the charge or generation and multiplication of secondary electrons.109-111 Important parameters usually considered in choosing detectors include sensitivity, accuracy, dynamic range, and response time.27, 28, 111 Commonly utilized detectors for MS include

Faraday cups,112 electron multipliers,110 and microchannel plates.111

All mass spectrometers require a certain degree of vacuum to minimize ion scattering, ion neutralization, and unwanted ion-molecule collisions.113 Important parameters in MS are resolving power and mass measurement accuracy and these parameters partly depend on the degree of attainable vacuum in the mass spectrometer.114

For example, FT-ICR mass spectrometers (which have high resolving power (e.g.,

109, 115 109, 116 m/Δm50% >300,000) and mass measurement accuracy (e.g., <2 ppm) require pressures on under 1 × 10-9 Torr range; on the other hand, quadrupole mass spectrometers

(which generally have lower mass resolving powers (e.g., ≤1000)117 and mass measurement accuracy (e.g., 100 ppm)117 require pressures in the range of 1 × 10-5 to 1 ×

10-7 Torr for efficient operation.102, 109 The vacuum condition is usually a result of differential pumping (i.e., multiple pump-down regions with progressively lower pressures) using a combination of pumps such as mechanical, diffusion, turbomolecular, and cryogenic pumps.118 A data system is needed to control the experimental parameters and store and process the signal acquired by the mass spectrometer.119

1.5.1. Quadrupole Mass Spectrometer

Quadrupole mass analyzers utilize stable trajectories of ions in an electric field to separate ions according to their mass-to-charge (m/z) ratios. The principle of ion

15 separation in the quadrupole was developed by W. Paul and H. Steinwedel at Bonn

University in 1953.120 Quadrupole mass analyzers are commonly coupled to GC121, 122 and liquid chromatography (LC)123, 124 systems for a number of applications including: residual gas, elemental, and organic compound analyses.117, 121-124 Significant development of the quadrupole mass analyzer (e.g., using a triple quadrupole combination) is seen in tandem MS, or MS/MS, where two or more stages of mass analysis are performed, generally, for structural analyses.125

1.5.1.1. Operating principle of the quadrupole. The quadrupole consists of a set of four electrodes (usually of circular but ideally hyperbolic cross sections) positioned in a radial array (see Figure 1.3 for illustration). A combination of direct current (DC) and alternating current (AC) or radio-frequency (RF) potentials are applied to the quadrupole electrodes to obtain the filtering properties of the quadrupole mass analyzer.102, 126 An ion travelling from the ion source to the detector (in the z-direction) must have stable trajectories in both the x-z and y-z planes.

-(U + V cosωt)

Quadrupole

+ - y z x - Detector + Ion Source

(U + V cosωt)

Figure 1.3. Schematic representation of the quadrupole mass analyzer.

16

The stable trajectory of the ion is determined by the electric field (from the applied RF potential superimposed on the DC potential on the quadrupole electrodes) felt by the ion according to equation 1.5102:

2 2 2 Φ = [U + Vcos (ω)t] x -y /2r0 (Equation 1.5) where Φ represents the combined potential applied to the electrodes, U is the magnitude of the DC potential (in volts), V is the magnitude of the RF potential (volts), ω is the angular frequency (in rad/s = 2πf, and f is the frequency (s-1) of the RF field), t is time (s), x and y are the distances (in meters; m) along the given coordinate axes, and r0 is the distance (in meters) from the center axis (i.e., the z-axis) to the surface of any electrode.

The electric field within the electrode assembly is defined by the potentials applied to the electrodes and is determined by taking partial derivatives of the potential distributions as a function of the distance (x) along the x-y or x-z axes102:

2 Ex = - δΦ/δx = - [U + Vcos (ω)t] x/r0 (Equation 1.6)

2 Ey = - δΦ/δy = [U + Vcos (ω)t] y/r0 (Equation 1.7)

Ez = - δΦ/δy = 0 (Equation 1.8)

The exerted on an ion (as the ion enters the space between the quadrupole rods) can be calculated by multiplying the magnitude of the electric field and the charge

(q = ze) of the ion (where z = electronic charge on the ion and e = charge per electron

(1.60 × 10-19 Coulombs)):

2 Fx = - [U + Vcos (ω)t] zex/r0 (Equation 1.9)

2 Fy = [U + Vcos (ω)t] zey/r0 (Equation 1.10)

Fz = 0 (Equation 1.11)

17

The acceleration (a) (in m/s2) of the ion can be obtained based on Newton’s law of motion, i.e., F = ma:

2 2 2 d x/dt = - zex/mr0 [U + Vcos (ω)t] (Equation 1.12)

2 2 2 d y/dt = + zey/mr0 [U + Vcos (ω)t] (Equation 1.13) d2z/dt2 = 0 (Equation 1.14) where m is the mass (kilograms/particle) of the ion. Solutions to equations 1.12, and 1.13

(note that according to equation 1.14, ion trajectories in the z direction are not influenced by the applied field) give a description of the trajectory of an ion in terms of the position and velocity of the ion at any time as it travels through the quadrupole. In other word, based on equation 1.14, the position and velocity of an ion along the z-axis would not be affected by any potential applied to the quadrupole assembly electrodes. However, if parameters au and qu are defined such that:

2 2 au = 4zeU/mω r0 (Equation 1.15)

2 2 qu = 2zeV/mω r0 (Equation 1.16)

And in combination with variable ξ =ωt/2 (i.e., time in radians of the applied field) and u

= x or y, then equations 1.12 and 1.13 may be transformed into the recognized form of

Mathieu’s differential equation:102, 127

2 2 d u/dξ + [au - 2qu Cos 2ξ]u = 0 (Equation 1.17)

The solutions to Mathieu’s equation show that there is a relationship between the coordinates of an ion (in x and y planes) and time. As long as the position of the ion in the x and y planes from the center of the electrode remain less than r0, the ion will be able to pass the quadrupole without colliding with any of the four electrodes before it reaches the detector. In a given quadrupole, r0 and ω are maintained constant while U and V are

18 varied, hence, an ion can have a stable or unstable trajectory as a function of the applied

U and V. Ions with stable trajectory in the quadrupole are located in stable coordinates in a stability diagram obtained by plotting au (related to DC potential) as a function of qu

(related to RF amplitude).102 In the stability diagram, ions with stable trajectories are those with m/z values defined by the space coordinates at the tip of the diagram (with the quadrupole operated with both the DC and AC potentials).102, 126, 127 On the other hand, when the quadrupole is operated in RF only mode (i.e., no DC potential), then

(theoretically) all ions (i.e., all m/z) should have stable trajectories through the quadrupole.102, 126

It is important to note that the filtering action of the quadrupole is obtained by increasing or decreasing the magnitude of the RF potential and DC potential at a fixed ratio.102, 127 Also, the resolution of quadrupole mass analyzer is established by the ratio of the RF to DC potentials applied on the quadrupole electrodes.102, 127 Quadrupole mass analyzers are inexpensive, simple to operate, robust, easy to maintain, and high- throughput MS instruments.102, 117, 126 Quadrupole mass analyzers do not require a high vacuum (i.e., pressure <1.0 × 10-9 Torr) as other high resolution mass spectrometers (e.g.,

FT-ICR MS instruments)109, and they are ideal for MS/MS analyses.125 However, a major disadvantage of quadrupole mass analyzers is their inherent low mass resolution.117

1.5.2. Fourier Transform-Ion Cyclotron Resonance Mass Spectrometry

The origins of Fourier transform mass spectrometry (FTMS) and ion cyclotron resonance (ICR) could be linked to the initial work of J. A. Hipple et al. (for measurement of cyclotron resonance frequency of )128, H. Sommer and H. A.

+ 129 Thomas (for detection of ICR of H2 ions) , D. Wobschall (for designing a sensitive

19

ICR spectrometer to study low-energy ion-molecule reactions)130, and L. R. Anders et al.

(for using ion cyclotron double resonance to study ion-molecule reactions).131

Subsequently, R. T. McIver introduced a trapped ion analyzer cell for ICR detection of gaseous ions132 and the first FT-ICR mass spectrometer was developed and demonstrated by M. B. Comisarow and A. G. Marshall in 1974.133 In FT-ICR MS, depending on the source of ion generation (either internal or external of the ICR cell)134, all experimental events such as ion quench, ion storage, ion manipulation, ion excitation and detection are performed within the ICR cell and these events are separated in time.135 Vacuum requirements in FT-ICR mass spectrometers are usually higher (<1 × 10-9 Torr) than most other mass spectrometers as the pressure has a major impact on the performance of FT-

ICR MS, especially in achieving high mass resolving power.136

An ICR cell (where ion quenching, trapping, excitation and detection takes place), generally consist of six electrodes (3 opposing pairs of electrically isolated electrodes, in a configuration similar to six faces of a normal cube). In FT-ICR MS, ion quenching is needed to clean the ICR cell of left over ions prior to the start of a new experiment, and ion quenching is performed by applying a DC potential (antisymmetric voltages) to the trap electrodes (generally two trapping electrodes facing each other and orthogonal to the magnetic field lines in the z-direction).135, 136 Ions are stored within the ICR cell before they are manipulated and detected by influences of both electric and magnetic fields.135,

136 The manipulation of trapped ions in ICR cell (usually performed by exciting the cyclotron motion of the ions) is mainly done for two purposes, either to (i) eject unwanted ions from the cell or (ii) promote the interaction of ions with neutral reagents for ion-molecule reactions.135, 136 Once ion manipulation is completed, ion excitation by

20 the application of an excitation waveform to the excitation electrodes (two parallel and opposing electrodes perpendicular to the trapping plates) enhances the coherent cyclotron motion of the ions.135, 136 The excited ions induce image current on the detection electrodes (two additional electrodes), and the image current on the detection electrodes is amplified, digitized, and stored in a computer for discrete Fourier transformation, magnitude calculation, and subsequent frequency-to-m/z conversion.135, 136

1.5.2.1. Operating principle of FT-ICR MS. Although original commercial

(Extrel)137 ICR cells were “close-ended” cubic cells, other configurations (such as “open- ended” cylindrical ICR cells are also common).138, 139 To understand the ion detection principle in FT-ICR MS, consider a cylindrical ICR cell (see Figure 1.4) placed within a

9.4 tesla (T) magnetic field. The ICR cell (utilized in all of the work reported in this dissertation) consists of twelve parallel pairs of capacitively coupled electrodes and two opposite trap electrodes for ion trapping, excitation, and detection purposes. The electrodes include: (i) quadrupole trapping (1 plate), (ii) filament trapping (1 plate), (iii) inner trapping (8 plates), (iv) excitation (2 plates), and (v) detection (2 plates) oxygen- free copper/gold plated/polished plates. An ion entering into an ICR cell has a cyclotron frequency corresponding to its m/z and moves in a circular orbit perpendicular to the magnetic field axis.109 The magnetic field constrains the ion’s motion in the x-y-plane (or it radial motions to a circular orbit), while the z-component or axial motion of the ion is not influenced by the magnet and hence ion is only confined by the applied electric field to the trap electrodes. Ions trapped in an ICR cell undergoes three motions, viz (i) trapping motion (due to the trapping potential in the z-axis), (ii) magnetron motion (~1 to

21

100 Hz; due to electric and magnetic fields), and (iii) cyclotron motion (few KHz to a few MHz; due to the magnetic field in the x, y-axes).136, 140

Figure 1.4. Pictorial representation of the Fourier transform-ion cyclotron resonance cell. B: Magnetic Field.

Based on the Lorentz law, an ion with a charge, q, and mass, m, entering into a fixed magnetic field, B, and moving with a velocity, v, will experience a ,

F109:

F = mass × acceleration = m dv/dt = qv × B (Equation 1.18)

As the ion enters the spatially uniform magnetic field, its motion is bent into a circular orbit (with radius, r) in a plane perpendicular to B (magnetic field lines), and having a natural angular frequency, ω:

ω = v/r (Equation 1.19)

Since angular acceleration, dv/dt = v2/r, then equation 1.18 could be written as: mv2/r = qv × B (Equation 1.20)

Substituting equation 1.19 into equation 1.20 and solving for ω gives:

22

ω = qB/m (Equation 1.21)

Because ω = 2πf (where f is the ion cyclotron frequency); equation 1.21 becomes: f = qB/2πm (Equation 1.22)

Usually B is in T (tesla), f is in hertz (Hz), q is in Coulombs (C), and m is in kilograms/particle. Hence, in FT-ICR MS, if the magnetic field is known, then an ion’s cyclotron frequency can be used to determine its m/z.

In order to detect the ions, their cyclotron radii must be increased and ion motion must be made coherent by applying a linear frequency sweep (“chirp” excitation)141 corresponding to the cyclotron frequencies of the ions orbiting the ICR cell (see Figure

1.5). As the cyclotron radii of the ions increase and as they orbit past the two detection plates, current is induced on the detection plates.134 The difference in the induced current on the two detection plates is amplified, digitized, and stored as a time-domain signal134

(see Figure 1.5). A mathematical procedure known as “Fourier transform” is employed to convert the time-domain signal into a frequency domain signal.140 Performing mass calibration and magnitude calculation (using equation 1.22) on the frequency domain signal generates a mass spectrum.140

Detection Mass Time-Domain Signal Spectrum

Excitation “Chirp” “Chirp” FT

m/z = qB / 2πƒ Amplitude Amplitude m/z Time (s)

Figure 1.5. Schematic representation of ion detection in Fourier transform-ion cyclotron resonance mass spectrometry. FT: Fourier transform; m/z: mass-to-charge ratio. 23

Fourier transform-ion cyclotron resonance MS has inherent benefits over other mass spectrometers in terms of mass resolving power (ability to distinguish two peaks with slightly different m/z values)115, high mass accuracy115, and multichannel or Fellgett advantages (detection of all ionic species simultaneously).134, 140 The achievable high mass resolving power or mass resolution and high mass accuracy enable the accurate determination of elemental composition and identification of unknown compounds.115, 116

Several multistage or MS/MS and ion-molecule experiments can be conducted in FT-ICR to further aid structural elucidation of molecules.116, 142 In addition, several ionization sources can be coupled to FT-ICR mass spectrometer such as ESI95, MALDI99, and EI34; hence, compatibility with multiple ionization sources increases the utility of FT-ICR MS for studying a wide variety of analytes.142 The main disadvantages of FT-ICR MS are the requirements for high-vacuum for efficient operation, high-magnetic fields (which could pose potential safety issues), and high cost of instrument purchase and maintenance.140

24

CHAPTER TWO

Ionization Techniques for Volatile Organic Compounds (VOCs)

2.1. Ionization Sources for VOC Analysis

Common ionization techniques suitable for volatile organic compounds (VOCs) include electron ionization (EI),34, 143 chemical ionization (CI),35, 144 field ionization

(FI),92, 93 and photoionization (PI).145, 146 The choice of ionization technique employed for

VOC analysis depends on the information desired and on the nature of analyte under investigation. For example, EI generates many fragments and is thus suitable for structural characterization28; conversely, CI, FI, and PI generate mostly molecular ions or pseudo-molecular ions and are consequently suitable for analyte identification.93, 147 A molecular ion in an EI process, generally represents a charged species of an intact molecule with an odd number of electrons formed either by loss or gain of an electron.148

A pseudo-molecular (or quasi-molecular) ion is an ion formed by the addition or subtraction of a proton to or from a molecular ion.148 Fragment ions are formed from the decomposition of another (parent) ion.148

In MS, “hard ionization” is referred to an ionization process that primarily produces fragment ions with generation of little-to-no molecular ions whereas “soft ionization” is considered to be a process that primarily generates intact molecular ions with minimal fragmentation.149, 150 Examples of soft ionization techniques include CI,35

FI,93 field desorption,93 FAB,101 ESI,94 PI,147 and MALDI98 and a hard ionization technique is EI.150

25

2.1.1. Electron Ionization

Electron ionization (EI) (formerly known as electron impact ionization) is one of the oldest ionization techniques for gas-phase molecules.91 It was first introduced by A. J.

Dempster in 1916, where he used electrons (generated from a heated cathode) for ionizing hydrogen gas.151 A. J. Dempster also used EI to study the generation of positive ions from aluminum phosphate, potassium chloride, and potassium iodide salts.152

However, it was not until 1940 that the well-known EI source was developed by A. O.

Nier.91 Electron ionization is one of the most utilized ionization techniques for identification and structural analysis of VOCs, and can be coupled to a GC system.28, 143,

150, 153 A conventional EI source consists of a heated filament (i.e., a cathode: where electrons are generated from), a source housing (i.e., ionizing chamber), an electron collector (anode trap), and a set of ion lens (i.e., extracting, focusing, and accelerating lens) (see Figure 2.1). In Figure 2.1, the repeller is used to propel EI-generated ions out of the ion source and toward the mass analyzer. The extraction, focusing, and accelerating lenses are used to extract, focus, and accelerate ions from the ion source into the mass analyzer, respectively.

Ionization occurs in EI by collision of a beam of electrons with neutral gas molecules in the ion source.91, 151, 152 The EI filament is a thin ribbon of metal (e.g., tungsten or rhenium) electrically heated to incandescence, or to a temperature >2000 °C, at which free electrons are emitted.91, 154 Some EI filaments are coated with specialized materials, such as iridium wire coated with thoria (Thorium dioxide (ThO2)) to form thoriated iridium, to reduce the temperature required to emit free electrons.154 The emitted electrons are attracted to the anode trap (or electron collector) placed on the

26

opposite side of the filament (or cathode). This causes the electron beam to transverse

through the central space of the ion source chamber. A magnetic field is superimposed on

the electric field (between the filament slit and the anode trap) to cause the electrons to

move in a tight helical path. This increases the electrons’ path length and overall

interaction of electrons with neutral gas molecules.

Neutral Gas Molecules Source Housing EI Repeller Ion Source Electrode II ICR Cell Filament Chamber Electrode I

d Electron Path Anode Trap EI Filament Extracting Lens Filament Slit Transformer Waveform Focusing Lens TTL Generator Accelerating Lens Amplifier Ions

To Mass Analyzer Figure 2.1. Schematic representation of the electron ionization (EI) source.

It is important to note that the heating current (required for thermal emission of

electrons) through the filament is usually in the order of 3 to 4 amperes for most EI

sources.154 However, not all the emitted electrons are available for ionization, because

most of the electrons are lost to the filament slit and source housing. Only a trap current

(i.e., current available for ionization) in the order of ~100 μA is available for

ionization.154 Hence, recent developments in EI have been aimed at designing EI sources

which maximize the interaction of emitted electrons with the gas-phase neutrals in order

to improve ionization efficiency.29, 30

27

In EI, electron energy is controlled by the potential difference between the filament and the source housing. Most EI experiments are conducted at 70 eV, because, at this energy, the variations in electron energies do not have a significant effect on ion generation. Moreover, at 70 eV, reproducible fragmentation patterns for organic compounds can be obtained and compared with reference mass spectra (also obtained at the “standard” 70 eV electron beam energy) for structural characterization and molecular identification.149, 150 Upon collision of a beam of electrons (at 70 eV) with a neutral molecule (M), the molecule absorbs some of the energy as internal energy, which causes the ejection of an electron (e-) from the molecule.149 Since most organic compounds have ionization energies (i.e., minimum energy to remove an electron from a gas-phase molecule)148 of less than 20 eV, the excess energy from EI (i.e., at 70 eV) leads to the production of fragment ions from the molecular ion.154 The ionization process in EI can be written as154:

M + e-  M•+ + 2e- (Reaction 2.1)

An additional use of an EI source, besides the ionization of VOCs, is for quantitative measurements (e.g., for conducting ionization cross section

155-157 measurements). The total number of ions (or ion current, I+) (in amperes) produced in EI is proportional to the ionizing electron current (Ie) (in amperes), ionizing path length

(d) (in centimeters; cm), ionization cross section of the molecule (Q) (in cm2/molecules), and the concentration or number density of the molecule in the ion source (N)

(molecules/cm3) (related to pressure of the molecule in the ion source, N)156:

I+ = QIedN (Equation 2.1)

Using equation 2.1, the ionization cross section of organic compounds can be obtained.

28

2.1.2. Chemical Ionization

Chemical ionization (CI) is a soft ionization technique originally introduced and developed by M. S. B. Munson and F. H. Field in 1966.158 Chemical ionization is based on the formation of ions through ion-molecule reactions.144, 158 In CI, a gas-phase reagent ion ([R+H]+) collides with a gas-phase molecule (M) to form an adduct ion, usually a protonated species (i.e., [M+H]+) of the intact molecule158, 159:

M + [R+H]+  [M+H]+ + R (Reaction 2.2)

The reagent ion is produced by EI of a selected reagent gas such as ammonia

35, 158 (NH3) or methane (CH4). For CI, high electron ionization energy (in comparison to

EI) is necessary to allow greater penetration of electrons into the ionization chamber which is kept at high pressure (due to high levels of the reagent gas).158 The CI source is similar in design to the EI source except that CI source is constructed with a lower level of gas conductance to allow for a high reagent gas pressure (e.g., >1.5 Torr) in the ionization chamber.158 Chemical ionization sources require higher differential pumping than EI sources in order to minimize the reduction in sensitivity (through poor ion transmission) and resolving power of the MS instrument at high pressure.35, 159 Protonated species of the intact molecule (i.e., [M+H]+) are formed by proton transfer from the reagent ion (i.e., [R+H]+) to the analyte molecule (i.e., M) (reaction 2.2). Since most of the energy transferred to the protonated analyte molecule during the ionization process is lost due to collisional cooling in the high pressure environment, the protonated specie of the analyte molecule (i.e., [M+H]+) is usually stable and with a high abundance in the mass spectrum.35, 158

29

Chemical ionization is similar to atmospheric pressure chemical ionization

(APCI) (developed by E. C. Horning, et al. in 1973) which is used for ionizing analytes

(possessing some degree of volatility) at atmospheric pressure.96 The main disadvantage of CI is the difficulty in reproducing the pressure of the reagent gas in the ion source chamber, which has a negative impact for quantitative analyses.28

The formation of negative ionic species is also possible in CI. Negative chemical ionization (NCI) occurs by ion-molecule reactions involving the interaction of the analyte molecule (M) with an anion (e.g., chloride ion, Cl-) to produce a negatively charge species (reaction 2.3).160 The abstraction of a proton from the analyte molecule upon reaction of the analyte molecule with a basic anion is also possible in NCI (reaction

2.4)161:

M + Cl-  [M + Cl]- (Reaction 2.3)

- - M + OH  [M – H] + H2O (Reaction 2.4)

The possibility of generating low energy electrons or thermal electrons (i.e., electrons with <1.0 eV) in the CI plasma also facilitates the formation of negatively charged species through electron capture (reaction 2.5)160:

M + e-  M•- (Reaction 2.5)

NCI is limited in application due to the fact that the majority of the compounds (such as oxidizing and/or alkylating agents) amenable to NCI are associated with environmental health problems.160

30

2.1.3. Field Ionization

Field ionization (FI) was the first soft ionization technique introduced for MS analysis of organic compounds.162, 163 The invention of FI was initiated by the work of E.

W. Mueller on field ion microscopy in 1953164 and M. G. Ingram and R. Gomer for coupling an FI microscope to a mass spectrometer.162, 165 Additional experiments coupling FI to MS were later conducted by H. D. Beckey, which led to the development of FI for organic compound analysis.93, 166 Field ionization results from the removal of an electron from a molecule (present in the gas-phase) by quantum mechanical tunneling in a high electric field.93, 166

A FI source consists of a field emitter (usually made of a thin tungsten wire) supported on two metal blocks, an extraction plate (counter electrode), and focusing slits/lens (for focusing ions into the mass analyzer).93 The emitter wire (usually coated with micro-needles of pyrolytic carbon or silicon) is positioned a few millimeters away from the extraction plate and held at an acceleration potential. The potential on the extraction plate is usually 8-12 kV lower than the potential on the emitter.93, 167 High field strength of the order of 107-108 V/cm is required for FI.93, 166-168 In FI, the molecular ion

(M•+ in the positive mode) is produced at a higher abundance for nonpolar or slightly polar organic compounds.93

The main advantages of FI include generation of high molecular ion abundances

(for molecular weight (MW) determination) and applicability to a wide range of compounds with MW below 1000 Daltons (Da).93, 166 However, the use of FI is limited partly because of (i) low sensitivity (an order of magnitude lower than EI or CI)93, (ii) expensive and fragile emitters93, 166, and (iii) difficulty in operation.93 In recent reports, FI

31 has been performed using carbon nanotubes169, 170 and whiskered silicon nanowires171 to enhance the applied electric field (e.g., >108 V/cm) at much lower potential (e.g., <10 V); although these techniques have not been utilized with MS.

2.1.4. Photoionization

Photoionization (PI) can be a soft ionization technique that involves the use of photon energies (mostly between 5 and 25 eV) to ionize gaseous molecules.145 The first use of photons for ionization was performed in 1929 by R. W. Ditchburn and F. L. Arnot, where potassium vapor was irradiated with photons and the resulting ions were analyzed in a simple sector mass spectrometer.172 Subsequently, in 1932 A. Terenin and B. Popov utilized PI to study ion pair processes of cadmium, zinc, or aluminum arcs with thallium halides.173 Photoionization has been considered an alternative ionization technique to EI and CI methods for VOC analysis but with limited utility147 because PI suffers from low sensitivity174 and PI requires expensive light sources capable of producing high flux of photons for efficient ionization of VOCs.145, 147 The major advantages of PI include (i) controllable degree of fragmentation (due to the ability to apply low ionization energy

(IE) (e.g., IE < 25 eV))147 and (ii) selective ionization (to ionize specific components in a mixture).145, 146

A PI source consists of a source of light (usually ultraviolet radiation), a monochromator (needed to control the degree of energy spread), and ion lenses (for extracting and focusing ions into the mass analyzer).145 The PI process involves an electron transfer reaction brought about by photo-absorption.145, 147 A molecule (M) absorbs a photon of sufficient energy (hv) to form an ion (M•+) and an electron.146, 147, 174,

175

32

M + hv  M•+ + e- (Reaction 2.6) h = Planck’s constant (6.63 × 10-34 J.s); v = frequency (s-1)

Photoionization is well-suited for selective analysis of organic compounds, such as aromatic and aliphatic compounds and complex organic gases evolved by pyrolysis.146

Photoionization has also been utilized with GC for analyzing reaction products, e.g., aldehydes, nitro compounds, glyoxal, and biacetyl in smog chamber experiments.174

2.2. Radio-Frequency Ionization

A logical approach to improve sensitivity for VOC MS analysis is to develop sensitive ionization sources that can maximize the conversion of gas-phase molecules into charged particles (i.e., ions). Radio-frequency ionization (RFI) is a novel technique that was recently developed in our labs for FT-ICR MS analysis of volatile and semi- volatile organic compounds.176 Radio-frequency ionization FT-ICR MS, in comparison with EI FT-ICR MS, shows higher ionization sensitivity (~6 fold increase) for VOCs analysis.177 It has been shown that RFI is a universal ionization technique, and also based on a limited set of systematic studies, it has been shown that RFI is not significantly influenced by variations of analytes’ ionization energies or polarizabilities.178, 179 Several

VOCs in simple or complex mixtures (e.g., gasoline and bio-oil samples) have been successfully ionized by RFI.180, 181 Although RFI is presently limited to use with FT-ICR mass spectrometer, work is in progress to develop RFI sources that would be used with non-magnet-based MS instruments.182

The existing RFI source consists of (i) RFI electrodes (either quadrupole or dipole rods) placed adjacent to an ICR cell and in the presence of a 9.4 T magnetic field and (ii) a home built RF power supply (for generating RF signals for RFI) (please see Figure 2.2

33 for illustration).176, 181 The first set of RFI electrodes were made up of four aluminum rods (i.e., quadrupole) (length 119.38 cm and outer diameter (O.D.) 0.63 cm) with total capacitance (i.e., capacitance of the rod and BNC cables connected to the rods to deliver

RF signals) of ~230 pF.176, 181 RFI electrodes were later reduced to two sets of aluminum rods (dipole) with length 6.50 cm, (O.D.) 0.63 cm, and total capacitance of ~40 pF.182-184

Radio-frequency ionization efficiencies for the quadrupole or dipole electrodes setups are similar.179, 182

QTP FTP

RF Signal TTL Waveform Transformer Amplifier Switch Generator

Figure 2.2. Schematic representation of the radio-frequency ionization (RFI) source. FTP: Filament trapping plate; ICR: Ion cyclotron resonance; QTP: Quadrupole trapping plate; TTL: Transistor-transistor-logic.

In RFI, radio-frequency signals (from ~3.5 to 12.5 MHz) (with initial RF voltage amplitudes <1.0 Vp-p) generated by a waveform generator are first amplified by an RF amplifier (~20 times amplification) then the amplified RF signal is further amplified by a

176 Ferrite Core Balun type transformer (to give a final ~> 100 Vp-p RF voltage amplitude).

The RF output is split into two separate legs for applying RF signal of opposite RF phases (i.e., positive and negative) onto the RFI electrodes.176 The “on”/“off” state of the

RF signal is controlled by a transistor-transistor logic (TTL) relay switch (signal received

34 from the mass spectrometer – IonSpec Omega software).176, 181 Upon application of the

RF signal to the RFI electrodes (usually for a few milliseconds), RFI-generated ions

(mostly molecular, pseudo-molecular, and fragments ions) are observed.176, 181 The RFI- generated ions are produced as a result of the collision of RF-generated electrons

(obtained via electron emission from RFI electrodes) with VOCs pulsed into the vacuum chamber of the FT-ICR mass spectrometer.183, 184 Extended discussions on RFI are provided in Chapters Four, Five, and Six of this dissertation.

Ionization techniques similar to RFI include spontaneous desorption185, FI93, 163, field desorption (FD)93, 166, and Townsend ionization.186 Spontaneous desorption and FD are surface ionization methods (i.e., they do not ionize molecules in the gas-phase (as in

RFI)).93, 166, 185 Townsend ionization involves the application of high electric field between two parallel plates (cathode and anode) separated by some distance and placed in a chamber filled with gas molecules.186-188 Electrons, which are initially generated by the photoelectric effect, are passed through the two parallel plates held at high potentials, thus leading to the multiplication of the initially generated electrons.186, 187 These electrons collide with the gas molecules in the ionization chamber resulting in ionization.186, 187 Radio-frequency has also been employed in RF mass spectrometry involving the use of RF (e.g., RF signal at 13.58 MHz) to generate high electric field between two electrodes at high pressure.189-191 The high electric field initiates the generation of a plasma within a high pressure region (usually 0.75-11.25

Torr) containing an inert gas, such as Argon.191 The positively charged ions formed within the plasma are in-turn utilized for secondary ionization of analyte species placed

35 on a sample plate within the ionization chamber.189, 191 However, RFI does not require high pressure and hence is suitable to couple to FT-ICR MS.

2.3. Electron Emission

Electron emission involves the liberation of electrons from the surface of a metallic material.192, 193 For electrons to be released from the surface of any metal, sufficient energy must be provided to overcome the surface barrier or work function of the metal.193, 194 The work function is the minimum energy required to emit an electron from the surface a metal.195 The work functions for most metals range from ~2 eV to 6 eV and depend on certain properties of the metals such as their chemical nature, purity, and surface condition (rough or smooth).195 Metals with lower work functions will require lower amount of energy for electron emission.195 The external source of energy could be from heat, electric field, or photons (light energy).196, 197 There are four main processes for electron emission: (i) thermionic emission,196, 198-201 (ii) field emission,192,

194, 202 (iii) secondary emission,203, 204 and (iv) photoelectric emission (i.e., photoemission).197, 205

In some recent reports, electronic excitations from metal surfaces have been shown to occur upon large-amplitude vibrations of reactant molecules (e.g., nitric oxide

(NO)).206-209 These chemically induced electronic excitations have been observed in silver206 and cesiated gold208 surfaces reacted with several gas molecules such as NO and

206-209 nitrogen dioxide (NO2).

36

2.3.1. Thermionic Emission

Thermionic emission is the process of electron emission from a metal surface by heat or thermal energy.196, 198 When a metal is heated to high temperatures (in excess of

2000 ºC), free electrons are able to leave the metal surface.196, 210 Common materials for electron emission include iridium, platinum, tungsten, thoriated tungsten, and metallic oxides of certain metals (e.g., barium and strontium).210 The coated materials enable the use of lower temperatures (e.g., 750 ºC) for electron emission.210 According to the

Richardson-Dushman equation, thermionic emission increases as a function of temperature199, 201:

2 -b/T Js = AT exp (Equation 2.2)

2 where Js is emission current density (amp/m ), T is temperature of the metal or emitter

(K), A is Richardson’s constant depending on the metal or emitter (amp/m2/K2), b is a constant for a metal given by199, 201: b = ϕq/k (Equation 2.3) where ϕ is the work function of the emitter or metal (in Joules; J), q is electronic charge

(1.60 × 10-19 C), and k is the Boltzmann’s constant (1.38 × 10-23 J/K). Increasing the emitter’s temperature or reducing the emitter’s work function increases the overall thermionic emission process.196, 198, 201, 210

2.3.2. Field Emission

Field emission involves the release of electrons from the surface of a metal by application of high electric field.192-194, 202, 211 Several scientists such as R. H. Fowler and

L. Nordheim,202 and J. R. Oppenheimer,199 R. A. Millikan and C. F. Eyring,194 between the mid- and late-1920’s studied the processes of field emission. In field emission,

37 intense electric field in the range of 108 V/m to 109 V/m is required for electrons to leave the surface (by tunneling through the work function energy barrier) of a metal.194, 202, 211,

212 According to the Fowler-Nordheim theory, the electron current density generated in field emission can be calculated using the equation202, 212:

J = q/2πh × μ1/2/(ϕ + μ)ϕ1/2 × E2 × exp(-4/3ĸϕ3/2/E) (Equation 2.4) where J is the current density (amp/m2), q is the electronic charge (1.60 × 10-19 C), h is the Planck’s constant (6.63× 10-34 J.s), ϕ is the metal work function (J), μ is the Fermi energy (J), E is electric field (V/m), and ĸ is a constant = (8π2m/h2)1/2 (m = rest mass of electron in kg). Based on equation 2.4, the electron emission current density depends on both the applied electric field and metal work function. Enhancement of the electric field has been reported for metal electrodes with sharp and micron-sized projections on the surface.211, 213 Hence, the geometry of the metal electrode and the work function of the emitting electrode determine the overall electron emission current density.

2.3.3. Secondary Emission

Secondary emission involves the release of electrons from the surface of a metal by the bombardment of a beam of primary electrons or ions.203, 204 The theory of secondary emission from metals was developed by H. Salow (in 1940),214 E. M. Baroody

(in 1950),215 and H. Bruining (in 1954)216 based on the Sommerfeld free-electron model.217 In general, if the energy of the primary electrons is sufficient or higher than the surface energy barrier (or work function) of metals, then free electrons from the metal surface can be released.204, 215 The intensity of secondary emission depends on the mass and energy of the bombarding particles (i.e., electrons or ions) and the nature of the metal or emitter.203, 204

38

2.3.4. Photoelectric Emission

During photoelectric emission (or photoemission), electron emission occurs from a metal surface by the application of light.197, 205 The concept of photoelectric emission from metals is based on the earlier work on photoelectric effect by H. Hertz (in 1887),218

P. Lenard (in 1902),219 A. Einstein (in 1905),220 and also on the Sommerfeld’s electron- theory of metals.205, 217 Some metals absorb photon energy when a beam of light strikes their surface. If the photon energy is greater than the work function of the metal, then free electrons are released from the surface of the metal. The amount of emitted electrons depends on the intensity and frequency of the incident light on the emitter or metal surface.197, 205

2.4. Goal of the Research Presented in this Dissertation

The focus of this dissertation is to demonstrate the use of a post-column cryogenic trapping (PCCT) system for improving VOC analysis at low VOC concentrations (e.g.,

<3.8 ppb) and to show the application/development of the RFI technique for MS analysis of VOCs using FT-ICR MS. Chapter Three is devoted to showing in detail the coupling and application of a cryogenic trapping system (i.e., PCCT) to a GC/MS system for analysis of standard acetone and VOCs from human exhaled breath samples and for monitoring nitric oxide photocatalytic reaction products (when coupled to a catalytic reaction unit and triple quadrupole mass spectrometer). Chapter Four is devoted to showing the application of RFI FT-ICR MS for VOC analysis in complex samples such as bio-oil and gasoline. Chapter Five evaluates the potential RF parameters (such as RF voltage amplitudes and frequencies) that determine ion yield (in terms of ion intensity and stability) in RFI. Chapter Six discusses the involvement of electrons in RFI using

39 direct electron current measurements and indirect electron quenching and electron attachment rate constant experiments.

40

CHAPTER THREE

Coupling Post-Column Cryogenic Trapping with Conventional GC/MS for the Analysis of Volatile Organic Compounds

Abstract

Details on the coupling of a home-built post-column cryogenic trap (PCCT) to a commercial gas chromatography (GC)/quadrupole mass spectrometry (MS) system for the analysis of volatile organic compounds (VOCs) are presented. The analytical figures of merit (e.g., dynamic range, limit of detection (LOD), and signal-to-noise ratio (S/N)) for the new GC/PCCT/MS were evaluated and compared to the “conventional” GC/MS

(i.e., without post-column cryogenic trapping) using: (i) direct headspace (HS) analysis of the standard acetone samples, (ii) solid-phase microextraction (SPME) analysis of the standard samples containing acetone at 0.80 and 3.8 ppbV, and (iii) SPME analysis of a human exhaled breath (HEB) sample. We show approximately two orders of magnitude improvement in LOD for the analysis of VOCs using GC/PCCT/MS over the use of conventional GC/MS. The combined use of post-column (i.e., cryogenic trapping) and pre-column (i.e., SPME) analyte concentration for the trace VOC analyses is also demonstrated. An application of the GC/PCCT/MS in “real world” sample analysis is demonstrated for the detection of HEB constituents including camphor and benzene

(VOCs observed in HEB which could not be detected without the use of PCCT).

41

3.1. Introduction

The analysis of volatile organic compounds (VOCs) is important in many areas of research such as environmental sciences,4, 221 forensics,222 and human health.223, 224 For instance, air quality monitoring,225, 226 detection of explosives,73 screening for drugs of abuse,19 and detection of disease biomarkers12, 227 require VOC analysis.

The VOCs in complex sample mixtures are normally analyzed using gas chromatography (GC) combined with mass spectrometry (MS) (or GC/MS). The VOC analysis with GC/MS allows for both separation and identification of VOCs in a single experiment. However, biological or environmental VOCs are often present at low concentrations and their comprehensive characterization requires improved analytical sensitivity. Mass spectrometry is one of the most sensitive analytical techniques for complex sample characterization42, 224, 228-231 and further developments in the instrumentation 232 and sample collection techniques225, 233, 234 will undoubtedly enhance the utility of GC/MS. Such improvements are particularly attractive when the sample sizes are limited.235 Our aim was to develop a simple GC/MS approach that would allow us to concentrate the post-GC separated VOCs prior to their detection.

In the past, several approaches have been utilized for analyte preconcentration prior to their injection into the GC column (i.e., pre-column trapping). These approaches include cryogen-free concentration,236 microscale purge and trap (MPT),42, 229, 237 solid- phase microextraction (SPME),238, 239 and solid-phase extraction (SPE).240 The sensitivities in the range of parts-per-million (ppm) to parts-per-trillion (ppt) have been reported, previously, for compounds such as volatile monoaromatics,24 halogenated hydrocarbons,4 and trihalomethanes241 using pre-column trapping GC/MS.

42

Another approach where cryogenic trapping is utilized is in longitudinally modulated cryogenic system (LMCS) developed by Marriot and Kinghorn for manipulating gas chromatographic bands and hence, improving GC separation.51, 52 The

LMCS is employed in two-dimensional (2-D) GC systems where an interface (a thermal modulator) is positioned between two GC columns for analytes’ accumulation and focusing purposes.54, 242 The modulating interface accumulates effluent from the primary column, refocuses the effluent into a narrow chromatographic band, and injects the effluent into the secondary column for improved resolution and overall mass sensitivity.53, 90, 243

It would also be beneficial to “re-concentrate” the analytes after their elution from the GC column and prior to their introduction into the GC detector such as MS (i.e., post- column trapping). Post-column trapping could be combined with other pre-column pre- concentration approaches to yield narrower GC peaks and larger peak heights and therefore, improve the observed signal-to-noise ratio (S/N) and limit of detection (LOD) for VOC analyses.

Previously, Jacoby et al. demonstrated the use of a post-column cryogenic trap

(PCCT) (“cold trap pulsed valve”) interface for a GC/Fourier transform-ion cyclotron resonance (FT-ICR) mass spectrometer.244 Jacoby et al. showed 12 folds improvement in the slope of calibration curves (based on analyte GC peak maximum) for VOCs analyzed using a PCCT interfaced to GC/FT-ICR MS (compared to the experiments without the use of PCCT).244 Based on the earlier work by Jacoby et al.,244 we also demonstrated the advantages of using a home-built PCCT for GC/FT-ICR MS analyses of complex VOC mixtures (e.g., gasoline).57, 245

43

In this manuscript, we report on interfacing a home-built PCCT to a

“conventional” commercial GC/MS instrument, i.e., GC/quadrupole mass spectrometer

(QMS) (model 5972A, former Hewlett Packard - now Agilent Technologies, Inc., CA) for VOC analyses. The design of the PCCT device used in this study was adapted from the original design of a cryogenic trap reported by Jacoby et al.244 and our group57 (for

GC/FT-ICR MS instruments) with the following two modifications. First modification involved the removal of the two pulsed valves used in the original design.57, 244 Second modification involved the addition of a graphical user interface (GUI) computer program to control the flash heating and cryogenic trapping time durations.

The analytical figures of merit (e.g., dynamic range, LOD, and S/N) of the new

GC/PCCT/MS are evaluated and compared to those obtained from the conventional

GC/MS using direct and SPME analyses of standard acetone samples. Moreover, to demonstrate a “real world” application of the GC/PCCT/MS, we show the results from the SPME analysis of VOCs in a human exhaled breath (HEB) sample.

3.2. Experimental

3.2.1. Materials

Commercially available acetone (certified ACS grade, 99.5 % purity (Fisher

Scientific Inc., NJ)) was used (without further purification) as a standard sample to evaluate the analytical figures of merit for GC/PCCT/MS and GC/MS. The 1-L and 5-L

Tedlar gas sampling bags were purchased from Restek (Restek Corp., PA). The 40-mL

EPA glass vials were purchased from Fisher Scientific (Fisher Scientific Inc., NJ). For all the headspace (HS) analyses, we used Hamilton gas-tight syringes (Hamilton Company,

44

NV) with the following volumes (depending on the minimum desired withdrawing volume): 1-L (minimum division marks (MDMs) of 0.01 L), 2-L (MDMs of 0.02

L), 10-μL (MDMs of 0.1 L), 25-μL (MDMs of 0.25 L), and 50-μL (MDMs of 0.5

L). The SPME assemblies (i.e., 85 µm carboxen/polydimethylsiloxane (CAR/PDMS) and 65 µm polydimethylsiloxane/divinyl benzene (PDMS/DVB)) were purchased from

Supelco (Supelco, MO).

3.2.2. Sample Preparation

3.2.2.1. Standard acetone samples. To prepare the HS acetone samples for obtaining the calibration curves, three acetone stock samples (stocks #1 to #3) were used.

The stock samples number 1, 2, and 3 were prepared by injecting 0.1 L, 0.6 L, and 2.0

L of the liquid acetone, respectively, into three clean septum-sealed 40-mL EPA vials.

The selected volumes of 0.1 L, 0.6 L, and 2.0 L for liquid acetone were below the minimum volume required for HS saturation (i.e., ~35 μL, calculated by assuming a density of 0.79 g/mL for the acetone at 25 °C and 1.0 atm.246 in 40-mL volume). The use of the low sample volume ensured complete evaporation of the injected acetone liquid into the EPA vials.

For the injection of 42 pmoles of acetone, 50 L volume of acetone HS was withdrawn from the stock #3 and further diluted in a 40-mL EPA vial. Subsequently, a

HS volume of 50 L (containing 42 pmoles of acetone) was withdrawn from the EPA vial (containing the diluted acetone HS) and injected into the GC injection port. The samples containing 170 and 510 pmoles of acetone were injected into the GC injection port by withdrawing the HS volumes of 5 and 15 L from the stock #1. The samples

45 containing 1 nmole, 2, 3, 4, 5, and 6 nmoles of acetone were injected into the GC injection port by withdrawing the HS volumes of 5, 10, 15, 20, 25, and 30 µL from the stock #2. Based on the multiple sample dilutions, the estimated propagated error for the sample preparation was <2 %. All the direct analyses using GC/PCCT/MS and GC/MS were performed in multiple (at least eight) trials to evaluate the reproducibility of the observed signals.

To prepare 3.8 ppbV and 0.80 ppbV of acetone samples for SPME analyses, an initial volume of 0.1 µL of acetone liquid was injected into a septum-sealed 40-mL EPA vial. Subsequently, separate 5-µL HS volumes (corresponding to 170 pmoles) of acetone were transferred from the 40-mL EPA vial into the helium (He) pre-filled 1-L and 5-L

Tedlar bags to obtain the final acetone concentrations of 3.8 ppbV and 0.80 ppbV, respectively. For GC/MS detections at such low concentrations, it is absolutely necessary to acquire method blanks before and after each analysis. Prior to the usage, we purged each of the Tedlar bags several times with He gas at room temperature (~25 oC) to remove any contaminant present in the bag. The cleanness of the He filled bags was then assessed by SPME sampling and subsequent GC/PCCT/MS and GC/MS analysis of the blank HS volume of the gas sampling bags. The absence of acetone peak in

GC/PCCT/MS and GC/MS analysis of the blank samples confirmed the cleanness of the gas sampling bags and the glassware.

To extract the acetone molecules, a preconditioned and clean SPME fiber was inserted into the Tedlar bag through the septum-sealed input port of the bag. Based on the eight experimental trials performed at different extraction times between 10 min and 80

46 min, the observed optimum extraction times (at 25 °C) for acetone from 1-L and 5-L

Tedlar bags were 40 min and 60 min, respectively.

For the data presented in Figure 3.5, Nitric oxide (NO) (100 ppm balanced with nitrogen (N2)) and ultra-high purity (UHP) oxygen (O2) (purity: 99.993 %) gases were purchased from a commercial source (Praxair, CT, USA). Aeroxide®

P25 titanium dioxide (TiO2) (21 nm, purity: >99.5 %) was purchased from Sigma-

Aldrich (Sigma-Aldrich, MO, USA).

3.2.2.2. Human exhaled breath (HEB) samples. For each analysis, 5 L of breath sample from a “healthy” human volunteer was collected into a clean (pre-purged with He gas) 5-L Tedlar bag. The HEB sample collections were carried out in the laboratory and no clean air was used for the breathing (prior to sample collection). The SPME devices were cleaned at 200 °C for 30 min and SPME blanks were analyzed to avoid potential environmental contaminations. All of the HEB samples for GC/PCCT/MS and GC/MS analyses were collected in the morning (between 9 a.m. to 12 a.m.) and after the subject had brushed teeth with toothpaste (Colgate-Palmolive Company, NY) and fasted for at least 4 hours. The collected HEB samples were analyzed immediately (within the first 5 min) after the collection.

3.2.3. Instrumentation

3.2.3.1. Gas chromatography (GC). Samples were injected into a 60-m (0.28 mm internal diameter (I.D.), 3.0-μm cross bonded 100 % PDMS stationary phase coating)

MTX-1 capillary column (Restek Corp., PA) housed in an SRI GC instrument (model

8610C, SRI Instruments, NV). Helium gas (99.999 % purity, Praxair, CT) was used as 47 the GC carrier gas. The GC injection port was set at 220 °C. For direct HS and SPME analyses of the standard acetone samples, the GC oven temperature and head pressure were set at 120 °C (isothermal) and 20 psi, respectively. For SPME analyses of HEB samples, the GC head pressure was set at 10 psi and the GC oven temperature was programmed as follows: 1 min at 60 °C, ramped at 5 °C / min to 120 °C and held at 120

°C for 12 min. The PeakSimple software (version 2.83, SRI Instrument, NV) running on a Pentium 4 (Microsoft Windows XP 2002) Dell desktop computer (Dell, Inc., TX) was used to control the GC oven temperature. For all of the experiments involving SPME, the analyte desorption time was 2 min at the GC injection port temperature of 220 °C.

3.2.3.2. Mass spectrometry (MS). A bench-top standalone quadrupole mass spectrometer (QMS) (model 5972A, former Hewlett Packard (HP), now Agilent

Technologies, Inc., CA) was used for MS analyses of the GC-eluted compounds. The

QMS parameters, calibration, chromatographic data acquisition, and tuning (“standard spectra autotune”) were controlled using Enhanced ChemStation software (version

B.01.00, HP 1989-1998) running on a Pentium III (Windows NT workstation 4.0) HP desktop computer. The mass calibration and tuning of the QMS were performed daily and prior to running analyte samples using perfluorotributylamine (PFTBA). The QMS m/z scan range was set between m/z of 15 Th and 220 Th (scan rate was 3.58 scans/second).

The electron impact (EI) ionization energy was factory set to 70 eV. The optimized electron multiplier (EM) voltages were used and for all of the autotune experiments the

EM values remained <2400 V. The QMS ion source temperature was set at 150 °C. The

QMS pressure was measured using the direct readout of a Granville-Phillips (series 342) ionization gauge controller (Granville-Phillips Co., CO). The ionization gauge was a mini

48

Bayard-Alpert type ion gauge tube. The QMS background pressures, without the sensitivity247, 248 and geometry factor corrections, at 10 psi and 20 psi of the He carrier gas GC head pressures were 3.6 × 10-5 torr and 8.0 × 10-5 torr, respectively.

The instrument, method, and sample blanks were used to identify the potential background noises. The background chemical noises included the peaks associated with

•+ •+ •+ minor air leaks (at m/z values of 18 Th (H2O ), 16 Th and 32 Th (O2 ), 28 Th (N2 ), 40

•+ •+ Th (Ar ), and 44 Th (CO2 )). All precautions were taken to minimize the air and pumping system (i.e., potential oil contaminant) chemical noises.

3.2.3.3. Post-column cryogenic trap (PCCT) assembly. The cryogenic trap assembly used for the QMS consisted of a capacitive discharge unit for resistive heating of the cryogenic trap element (CTE). The CTE connecting the GC outlet and the MS inlet was an 8-in. piece of sulfinert stainless steel tubing (I.D. 0.021 in., outer diameter (O.D.)

0.029 in., Restek Corp., PA) and was immersed in a liquid nitrogen (LN2) bath. The total resistance of the CTE was 0.72 Ω. The total capacitance of the discharge unit was 0.1 F.

The capacitors within the discharge unit were electrically charged using an Agilent 500-

W direct current (DC) power supply (model 6554A, Agilent Technologies, Inc., CA). A

DC voltage of 40 V (at maximum output current of 4.0 A) was used for charging the capacitors in the discharge unit. The flash heating of the CTE was initiated by discharging the capacitors in the discharge unit. Desorption temperatures in the excess of

200 °C (as measured with careful calibration of the resistance change of the CTE within the liquid nitrogen bath as a function of temperature) could be achieved.

The capacitive discharging event was initiated with a transistor-transistor logic

(TTL) signal. The TTL signal was generated by an in-house designed computer software

49 in Visual Basic (VB version 6.0 for Windows NT). Scheme 3.1 depicts a print screen of the PCCT controller GUI window showing the cryogenic trapping parameters. For instance, as shown in Scheme 3.1, the user-defined inputs could be used to change the cryogenic trapping and heating time durations (in seconds) and optimize these parameters for a particular experiment. We did not use a pulsed valve in the GC/PCCT/MS experiments reported in here. However, the PCCT controller GUI could also be utilized to control a pulsed valve opening/closing time duration (in seconds).

Heating duration of 0.3 seconds

Cycle time (60 seconds)

Three- way pulsed valve CTE heating duration Total number of Cryogenic trapping on/off duration (seconds) (seconds) heating/focusing cycles duration (seconds)

Scheme 3.1. Print screen of the GUI window (“Cryo Controller”) from the home-written software used to control the PCCT unit. The values shown for the “Cryo Controller” parameters are the actual adjusted values used to acquire the GC/PCCT/MS chromatograms in Figures 3.1, 3.3, and 3.4.

For all of the experiments reported herein, we used time durations of 60 s for cryofocusing and 0.3 s for heating. It should be noted that heating time was governed by the time constant () of the RC circuit (used in the discharge unit of the PCCT device) which was ~72.8 ms. In other words, cooling period could be changed but heating occurred during the capacitive discharge event only.

50

3.2.3.4. GC/PCCT/MS setup. Scheme 3.2 shows a pictorial representation of the

GC/PCCT/MS setup used in this study. The CTE was connected to the end of GC column

using a heated transfer line made of an 8-in. long MTX guard column (I.D. 0.01 in., O.D.

0.02 in., Restek Corp., PA). The other end of the CTE was connected to the QMS using a

10-in. long MTX guard column (I.D. 0.01 in., O.D. 0.02 in., Restek Corp., PA). To

connect the two ends of the CTE to either GC or QMS, 1/16-in. stainless steel unions

(Restek Corp., PA) were used. Two brass heating blocks (1.0 in. (length) × 0.7 in. (width)

× 0.5 in. (height)) were used on both ends of the CTE to avoid “cold spots” and GC “peak

tailings”. The entire length of the transfer lines on both sides of the CTE was

continuously heated and kept at 150 °C during the data acquisition. For the experiments

without the cryogenic trapping (i.e., GC/MS), the CTE was used as a transfer line and

heated (at 150 °C) using a heat gun.

Heated Transfer Lines

Quadrupole Mass Heated Metal Blocks Computer Spectrometer Injection Data Port QMS Analysis Teflon Cryogenic GC Oven LN Cup 2 CTrapTE TTL Switch DC Capacitor Supply Bank Cryofocuser Assembly

Scheme 3.2. Schematic representation of the GC/PCCT/MS instrumental setup used for the VOC analyses.

51

3.3. Results and Discussion

In the following sections, we present the results from the analytical performance evaluation of the new GC/PCCT/MS using standards and VOCs in HEB samples249 via a two-step analyte concentration, viz., SPME combined with PCCT.

3.3.1. Direct Headspace Analysis of Acetone

Figure 3.1 shows the overlaid selected ion chromatograms (SICs) of acetone (at

+ m/z 43 (CH3CO )) obtained using GC/PCCT/MS (Figure 3.1, solid line) and the conventional GC/MS approach (Figure 3.1, dotted line). A total amount of 1.0 nmole of acetone HS was injected into the GC injection port in each experiment (i.e.,

GC/PCCT/MS and GC/MS). For better comparison, the peak labeled as “GC/MS” in

Figure 1 is magnified by 14 folds (i.e., the ratio of peak maxima for peaks labeled as

“GC/PCCT/MS” and “GC/MS”).

To acquire the chromatogram labeled as “GC/MS” in Figure 3.1 (dotted line), the

CTE in Scheme 3.2 was not immersed in LN2 (viz., the teflon cup in Scheme 3.2 was empty). The CTE was heated with a heat gun (at 150 °C) and used as a transfer line. The retention time (RT) of the acetone (i.e., the centroid value of the Gaussian-shaped GC peak in Figure 3.1) in the GC/MS was 7.20 (± 0.160) min or 432 (± 9.60) s (errors reported at the 95 % confidence level (CL) for n = 3). The chromatographic peak width of 16 s at the base of the Gaussian profile for acetone in Figure 3.1 (dotted line) was comparable to the peak widths we observed for this particular instrument when GC was directly interfaced to QMS (an indication for the absence of “cold spots” in the exposed regions of the CTE in GC/MS experiments).

52

Direct Headspace (HS) Analysis of 1.0 nmole of Acetone Cryogenic trapping time (~60 s)

1.2x106 GC/MS × 14

5 9.0x10 GC/PCCT/MS 43 (A.U.) 5

m/z 6.0x10

3.0x105

0.0

0.0 0.5 7.0 7.5 8.0 8.5 9.0 Intensity @ Retention Time (min)

Figure 3.1. Overlaid selected ion chromatograms (SICs) of acetone at m/z 43 + (CH3CO ) obtained using GC/PCCT/MS (solid line) and GC/MS (dotted line). A headspace (HS) volume of 5 µL containing 1.0 nmole of acetone was injected into the GC injection port in each instrumental configuration (i.e., GC/PCCT/MS or GC/MS). For comparison purpose, the peak labeled as “GC/MS” is magnified by 14 folds. The two parallel diagonal lines in x-axis denote axis break from 0.7 min to 6.7 min.

To acquire the chromatogram labeled as “GC/PCCT/MS” in Figure 3.1 (solid line), the eluted acetone molecules from GC column were captured in the CTE immersed in LN2 (Scheme 3.2). The cryogenic trapping time was set at 60 s (i.e., the “cycle time” in

PCCT controller GUI in Scheme 3.1 was 60 s and started with the start of the GC/MS data acquisition). The cryogenic trapping duration of 60 s (between 7.04 min or 422.4 s to

8.04 min or 482.4 s) is shown with a horizontal arrow (between the two vertical dash lines) in Figure 3.1.

Once the GC eluting organic molecules (in this case acetone) were trapped, the post-column trapped analyte(s) were desorbed and admitted into the QMS by the resistive heating of the CTE. The shift in the RT (i.e., 1.0 min) of acetone in GC/PCCT/MS

53 experiment (i.e., Figure 3.1, solid line) (compared to GC/MS experiment) is expected and confirms the operator selection of cryogenic trapping time of 60 s. The appearance of a single GC/PCCT/MS peak in Figure 3.1 (solid line) suggests that all of the cryogenically trapped acetone molecules were desorbed in a single flash heating event (viz., no peak was observed in the subsequent flash heating event at 9.04 min in Figure 3.1, solid line).

In order to desorb all of the cryogenic trapped analytes in a single flash heating event, the duration of the cryofocusing and the heating cycles as well as the amount of heating must be sufficient. In other words, it is possible for a single acetone injection to yield more than a single GC peak due to the trapping of the left over acetone molecules in the “back-end” of the acetone GC peak. Our studies under variable cryofocusing and heating durations, He gas flow rates, and discharge timing conditions suggested that the

“back-end” accumulation could yield multiple peaks indicating the highly sensitive nature of the cryofocusing approach.

The comparisons between the two SICs for m/z 43 in Figure 3.1 revealed that the peak width (at the base) of the GC/PCCT/MS peak of acetone (solid line) was decreased by over ten folds (from 16 s to 1.4 s) as compared to GC/MS peak of acetone (dotted data points). Although the peak areas remained comparable (viz., 2.4 × 106 arbitrary unit

(A.U.) for GC/MS and 3.2 × 106 A.U. for GC/PCCT/MS), the peak heights increased by fourteen folds from 6.6 × 104 A.U. for GC/MS to 9.4 × 105 A.U. for GC/PCCT/MS.

3.3.2. Analytical Sensitivity and Dynamic Range: GC/PCCT/MS vs. GC/MS

To examine the dynamic/linearity range of the VOC analyses using

GC/PCCT/MS and GC/MS, we injected eighteen different amounts of acetone HS (i.e.,

42, 170, 510 pmoles, 1.0 nmole, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10, 20, 30, 40, 50,

54 and 100 nmoles) into the GC injection port in each mode (data not shown). For the

GC/MS analysis of acetone, the EM detector response was linear from 170 pmoles to 50 nmoles of the injected acetone and the detector saturation occurred at >50 nmoles of acetone; no signal was observed for 42 pmoles of acetone in GC/MS experiment. For the

GC/PCCT/MS experiment, the EM detector response was linear from 42 pmoles to 6.0 nmoles of the injected acetone and the detector response reached to the saturation level at

>6.0 nmoles of acetone. The observation of detector saturation at lower acetone concentration for GC/PCCT/MS (i.e., >6.0 nmoles) compared to GC/MS (i.e., >50 nmoles) is expected and demonstrates the advantage of using PCCT for enhancing the

S/N ratio in GC/MS analyses (see Table 3.1 and related discussions). The estimated dynamic ranges for acetone analysis using GC/PCCT/MS and GC/MS are 150 (i.e., 6.0 nmoles / 42 pmoles) and 294 (i.e., 50 nmoles / 170 pmoles), respectively. Although the dynamic range for GC/PCCT/MS is smaller than the dynamic range for GC/MS, their combined dynamic range is over three orders of magnitude.

Figures 3.2(a) and 3.2(b) show the calibration curves for acetone based on the total peak area (i.e., sum of the SIC peak areas at m/z 43 Th and 58 Th) and the peak height (for the SIC at m/z 43 Th), respectively. The peak areas and peak heights were corrected for background and used for generating the calibration plots in Figures 3.2(a) and 3.2(b). Each data point for GC/PCCT/MS (denoted with filled squares, ■) and

GC/MS (denoted with empty circles, ○) in Figures 3.2(a) and 3.2(b) corresponds to an average of eight experimental trials and the experimental absolute errors are reported at the 95 % CL. To obtain the best linear fit equations in Figures 3.2(a) and 3.2(b), we utilized the curve fitting function of Origin 7.0 software (Origin Lab Co., MA)

55

(a) Peak Area-Based Calibration Curves (Acetone)

3.2E+07 Y = 3.2 (± 0.1) × 106 X (R² = 0.98) 2.4E+07 ○ GC/MS ■ GC/PCCT/MS

1.6E+07

8.0E+06 Y = 2.2 (± 0.1) × 106 X

(A.U.) Area Peak Total (R² = 0.97) 0.0E+00 0 1 2 3 4 5 6 Amount Injected (nmole)

(b) Peak Height-Based Calibration Curves (Acetone) 7.2E+06 Y = 7.5 (± 0.3) × 105 X

43 43 (R² = 0.96)

5.4E+06 ○ GC/MS m/z ■ GC/PCCT/MS

3.6E+06

(A.U.)

1.8E+06 Y = 4.3 (± 0.2) × 104 X

Peak Height Height @ Peak (R² = 0.96) 0.0E+00 0 1 2 3 4 5 6 Amount Injected (nmole)

Figure 3.2. Calibration curves for acetone (between 42 pmoles and 6 nmoles) using

GC/PCCT/MS (solid line) and GC/MS (dash lines) based on the (a) total peak areas (sum + •+ of the SIC peak areas at m/z 43 Th (CH3CO ) and 58 Th (M )) and (b) peak height for + SIC at m/z 43 Th (CH3CO ). The data points shown with filled square (■) and empty circle (○) symbols correspond to GC/PCCT/MS and GC/MS experiments, respectively.

Each data point in the panels (a) and (b) corresponds to an average of eight experimental trials and the absolute errors at the 95 % CL are denoted with the error bars.

56

The acetone calibration equations obtained based on the acetone total peak area

(Figure 3.2(a)) showed comparable slopes. The comparison of the calibration equations obtained based on the peak heights of the SIC at m/z 43 in Figure 3.2(b) revealed seventeen folds enhancement in the slope of calibration equation for GC/PCCT/MS as compared to GC/MS.

It is important to note that different parameters including vacuum chamber pressure and background contamination levels can significantly affect the observed sensitivity enhancements in GC/PCCT/MS. For example, the observed routine sensitivity enhancement of 17 folds for the detection of acetone using GC/PCCT/MS (in Figure

3.2(b)) was obtained with the vacuum chamber pressure of 7.7 × 10-5 torr to 8.0 × 10-5 torr and a background level of ~2,000 counts/scan. However, with a “cleaner” vacuum chamber (background level of 100 counts/scan), we could obtain up to 40 folds enhancement in the sensitivity for the detection of acetone using GC/PCCT/MS (as compared to GC/MS). The high background level in the MS vacuum chamber of GC/MS instruments results partly from the long history of instrument usage and sample injection.232

Table 3.1 shows the summary of the calculated S/N values of the acetone MS signal (at SIC peak maximum for m/z 43 Th) corresponding to the data points in Figure

3.2. In Table 3.1, columns #1 to #3 correspond to the amount of acetone injected (into the

GC column injection port), the calculated S/N at SIC peak maximum (for m/z 43), and

S/N enhancement (for GC/PCCT/MS as compared to GC/MS), respectively. The S/N values in Table 3.1 were calculated by dividing the peak height of the SIC at m/z 43

(from acetone) by the root-mean-square (RMS) of the background noise (averaged over

57

±30 s time interval from the GC acetone peak centroid). The comparisons of the values in column #2 of Table 3.1 show up to 22 folds enhancement in S/N (at peak maximum for the SIC at m/z 43) for acetone analysis using GC/PCCT/MS (as compared to GC/MS).

The calculated LOD (based on the standard deviation of the sample blank and slope of the calibration curve250) for GC/PCCT/MS and GC/MS were 6.3 × 10-14 moles and 8.3 ×

10-12 moles, respectively.

Table 3.1. Comparisons of the signal-to-noise ratio (S/N) for the direct analysis of acetone using GC/MS and GC/PCCT/MS.

Amount Injected S/N* S/N {nmole(s)} GC/MS GC/PCCT/MS Enhancement

~0.042 ND** ~97 (± 32) -- ~0.17 ~29 (± 7) ~540 (± 271) ~19 ~0.51 ~93 (± 32) ~2085 (± 387) ~22 ~1.0 ~269 (± 33) ~4509 (± 1601) ~17 ~2.0 ~367 (± 38) ~6259 (± 3697) ~17 ~3.0 ~596 (± 43) ~5685 (± 3417) ~10 ~4.0 ~701 (± 124) ~8261 (± 4395) ~12 ~5.0 ~876 (± 51) ~11472 (± 5949) ~13 ~6.0 ~1191 (± 371) ~9435 (± 2610) ~8

+ *The S/N values are the average SIC peak maxima at m/z 43 (CH3CO ) from eight experimental trials. The errors at 95 % CL are included in the parentheses. **ND: Not detected.

We also evaluated the analytical sensitivity of GC/PCCT/MS by the analysis of other VOCs (e.g., benzene, hexane, and toluene dissolved in dichloromethane, data not shown) and observed enhancements both in sensitivity (i.e., 14-22 folds) and LOD (i.e., two orders of magnitude) as compared to GC/MS analyses. Although we consistently observed enhancements in the sensitivity for VOC analyses using GC/PCCT/MS, the chromatographic data acquired using the GC/PCCT/MS were less reproducible than the

GC/MS. We are currently exploring the possibility of using other fast scan mass

58 analyzers (e.g., time-of-flight (TOF)) and modifying the PCCT/MS interface to further enhance the performance characteristics of GC/PCCT/MS.

3.3.3. Solid-Phase Microextraction (SPME) Analysis of Acetone at the ppbV Level

In order to assess the performance of GC/PCCT/MS for VOC analysis at low concentration (i.e., in the ppbV range), we used SPME to preconcentrate the samples containing 3.8 ppbV and 0.80 ppbV of acetone and then inject the SPME-extracted acetone molecules into the GC injection port. All SPME experiments were conducted in triplicate to ensure the reproducibility of the analyses.

Figures 3.3(a) and 3.3(b) show the SICs (at m/z 43) obtained from the analyses of

3.8 ppbV and 0.80 ppbV acetone samples, respectively, using GC/PCCT/MS (solid line) and GC/MS (dotted line). The SPME analysis of acetone at the concentration of 3.8 ppbV using GC/MS (Figure 3.3(a), dotted line) showed a peak at the RT 7.1 (± 0.30) min with poor S/N (viz. S/N of 3.2 (± 0.40) at the peak maximum). However, the SPME analysis of

3.8 ppbV acetone sample using GC/PCCT/MS (with 60 s PCCT cycle) resulted in a sharp and narrow peak with S/N of 29.0 (± 2.00) (Figure 3.3(a), solid line). The comparison of the S/N (for m/z 43) at peak maxima for the two chromatograms in Figure 3.3(a) revealed nine folds enhancement in the S/N for the SPME analysis of 3.8 ppbV acetone sample using GC/PCCT/MS (as compared to GC/MS).

The SIC (at m/z 43) for the SPME analysis of acetone at concentration of 0.80 ppbV using GC/MS (Figure 3.3(b), dotted line) showed no MS signal for acetone. The

SPME analysis of acetone at 0.80 ppbV using GC/PCCT/MS (with 60 s PCCT cycle) resulted in a single sharp peak for acetone (Figure 3.3(b), solid line) with a S/N of 19.0 (±

3.00).

59

(a) SPME Analysis of 3.8 ppbV Acetone Sample

GC/PCCT/MS 5 1.6x10 {S/N ≈ 29.0 (± 2.00)}

▬ GC/PCCT/MS 1.2x105 ○○○

43 (A.U.) GC/MS

4 m/z 8.0x10 GC/MS {S/N ≈ 3.2 (± 0.40)} 4.0x104

0.0 0.5 7.0 7.5 8.0 8.5 9.0 Intensity @

Retention Time (min)

(b) SPME Analysis of 0.80 ppbV Acetone Sample

5 1.0x10 GC/PCCT/MS {S/N ≈ 19.0 (± 3.00)} 7.5x104 ▬ GC/PCCT/MS

43 (A.U.) ○○○ GC/MS

4 m/z 5.0x10 GC/MS 2.5x104 {No signal for acetone}

0.0 0.5 7.0 7.5 8.0 8.5 9.0

Intensity @ Retention Time (min)

+ Figure 3.3. Selected ion chromatograms (SICs) of acetone at m/z 43 (CH3CO ) obtained from SPME analyses of samples containing (a) 3.8 ppbV and (b) 0.80 ppbV of acetone using GC/MS (dotted line) and GC/PCCT/MS (solid line). The values on top of the peaks are the average S/N valuess for the SIC peak maxima at m/z 43 from triplicate experiments (the error at 95 % CL is included in parenthesis for each value). In panels (a) and (b), the parallel diagonal lines in x-axes denote the axis break from 0.7 min to 6.7 min.

60

3.3.4. Solid-Phase Microextraction Analysis of Human Exhaled Breath Sample

To demonstrate the applicability of the GC/PCCT/MS for “real world” sample analyses, we analyzed HEB samples from a “healthy” volunteer. We chose HEB sample because of its potential as a noninvasively collectable biological sample for disease diagnosis and biomarker discovery.12, 227, 249, 251 The previous reports on the GC/MS analysis of HEB samples have shown that HEB contains VOCs at low concentrations in the ppm to ppt ranges.223, 249, 252 However, some of the reported values, in terms of the relative abundance information, are inconsistent. For instance, some of the initial reports indicate the presence of hundreds of VOCs in HEB and none of the reported VOCs are significantly higher (in terms of relative abundance) than the rest.253 However, subsequent reports249, 254 on HEB including our preconcentrator (PC)/GC/FT-ICR MS results255 indicated that acetone and isoprene were at much higher concentrations than most other VOCs in HEB. Although these different results can be explained by considering the experimental variabilities in collecting and analyzing HEB samples, it is desirable to use the most direct approaches for such analyses. We decided to use both

GC/MS and GC/PCCT/MS approaches and identify the most abundant chemicals in

HEB. It should be noted that we did not use clean air256 or exhaustive studies12, 223 for this purpose. In addition, the affinities of different SPMEs were not studied in this report.

However, our preliminary studies demonstrate the advantages of GC/PCCT/MS approach for HEB characterization and VOC analysis (e.g., biological and environmental samples).

Figures 3.4(a) and 3.4(b) show the total ion chromatograms (TICs) (after background correction) obtained from SPME analysis of a HEB sample (from a “healthy” human volunteer) using GC/MS and GC/PCCT/MS, respectively. To acquire the TICs in

61

(a) Human Breath Analysis (GC/MS)

5

8.0x10 Isoprene Acetone 6.4x105 × 20

5 × 20

4.8x10

5

3.2x10

Ethanol Tetradecane

5 Toluene Butanone

1.6x10 -

2

Propanol -

1 0.0 0 8 10 12 14 16 18 20 22 TotalIon Intensity (A.U.) Retention Time (min)

(b) Human Breath Analysis (GC/PCCT/MS)

6

2.5x10

Isoprene × 20 6

2.0x10 Acetone 6 1.5x10

× 20

Butanone

-

6

2

1.0x10

5 Propanol

-

Benzene

Ethanol Toluene

5.0x10 1

Benzene

Camphor Tetradecane 0.0 0 8 10 12 14 16 18 20 22 TotalIon Intensity (A.U.) Retention Time (min) Figure 3.4. Total ion chromatograms (TICs) (after background correction) obtained from the analysis of a HEB sample using (a) GC/MS and (b) GC/PCCT/MS instrumental setups. In panels (a) and (b), the parallel diagonal lines in x-axes denote the axis break from 2 min to 7 min.

Figures 3.4(a) and 3.4(b), two separate 5-L volumes of the breath sample were collected in two separate 5-L Tedlar gas sampling bags. Then the SPME stationary phase was inserted into the Tedlar bag through the septum-sealed inlet port of the bag. The

62 optimized SPME sampling/extraction time was 60 min for both GC/MS and

GC/PCCT/MS experiments. The chemical identities of the chromatographic peaks for each TIC (Figure 3.4) are indicated with the labels on top of each peak in Figures 3.4(a) and 3.4(b).

Figures 3.4(a) and 3.4(b) show the total ion chromatograms (TICs) (after background correction) obtained from SPME analysis of a HEB sample (from a “healthy” human volunteer) using GC/MS and GC/PCCT/MS, respectively. To acquire the TICs in

Figures 3.4(a) and 3.4(b), two separate 5-L volumes of the breath sample were collected in two separate 5-L Tedlar gas sampling bags. Then the SPME stationary phase was inserted into the Tedlar bag through the septum-sealed inlet port of the bag. The optimized SPME sampling/extraction time was 60 min for both GC/MS and

GC/PCCT/MS experiments. The chemical identities of the chromatographic peaks for each TIC (Figure 3.4) are indicated with the labels on top of each peak in Figures 3.4(a) and 3.4(b).

A total of seven VOCs were identified in the analyzed HEB sample using GC/MS

(Figure 3.4(a)). The identified VOCs using GC/MS included (in the increasing order of

RT): ethanol, acetone, isoprene, 1-propanol, 2-butanone, tetradecane, and toluene. The analysis of HEB sample from the same volunteer under the identical experimental conditions (e.g., identical sample collection time duration and temperature) using

GC/PCCT/MS revealed the presence of additional VOCs (Figure 3.4(b)). The additional two VOCs in GC/PCCT/MS experiments were camphor and benzene (Figure 3.4(b)).

Please note that the TIC acquired using GC/PCCT/MS in Figure 3.4(b) shows the presence of benzene as two sharp peaks. The observation of the two peaks for benzene in

63

Figure 3.4(b) could potentially be due to the GC peak broadening or elution of a different compound with the same chemical composition.

Also, for all of the VOCs (in the analyzed HEB sample) detected with both

GC/MS and GC/PCCT/MS techniques, the observed S/N values were larger for the

GC/PCCT/MS analysis. The improved S/N enhances the fragment ion assignments and the matching score for unknown identification using MS library searches.

It is interesting to note that both GC/PCCT/MS and GC/MS results confirmed our initial report255 in that, under our sampling conditions, isoprene and acetone concentrations were higher than the concentrations of other VOCs present in HEB.

3.3.5. Cryogenic Trap Coupled to a Catalytic Reaction Unit and MS System

To demonstrate the wide application of the cryofocuser for VOC analysis at low concentrations, we coupled the PCCT to a catalytic reaction unit for preconcentrating nitric oxide (NO) photocatalytic reaction products prior to introduction into a triple quadrupole mass spectrometer. Nitric oxide decomposition reaction in the presence of titanium dioxide (TiO2) catalyst and ultra-violet (UV) radiation has been previously studied using NOx analyzers equipped with either infrared or chemiluminescence

257, 258 detectors. The NOx analyzers based on chemiluminescence measurement systems are limited to indirect detection of NOx gases (i.e., NO and nitrogen dioxide (NO2)) and are not able to detect other potential photocatalytic products such as nitrous oxide

259, 260 (N2O). Coupling the catalytic chamber to MS allows for direct m/z determinations

(i.e., molecular weight determination) of NOx species and NO photocatalytic products.

Figure 3.5 shows the temporal plots of normalized ion intensities for the NO,

N2O, and NO2 species at 80 ppm NO (in simulated air, i.e., 80 % N2 and 20 % O2) and a

64

cryofocusing time of 60 s in the presence/absence of UV light and P25 TiO2 catalyst

(~0.50 grams of P25 TiO2 catalyst was loaded into the reaction chamber). In Figure 3.5,

symbols denoted by empty triangles (Δ), empty circles (○), and solid circles (●) represent

NO, N2O, and NO2 ion intensities, respectively. The normalized ion intensities for the

NO, N2O, and NO2 species were obtained by dividing the total ion intensity of each NO,

N2O, and NO2 species by the sum of the total ion intensities from NO, N2O, and NO2

species at each time (or data) point.

NO Photocatalytic Reaction in the

Presence of TiO2 (P25) Catalyst 1.2 UV “On” 1.0 UV “Off”

0.8 NO

0.6 ○ N2O 0.4 ● NO2

0.2 Normalized Intensity Normalized 0.0 0 10 20 30 40 50 60 70 Elapsed Time (min)

Figure 3.5. Temporal plot of normalized ion intensities for NO, N2O, and NO2 with TiO2 P25 catalyst.

As shown in Figure 3.5, the normalized ion intensities (i.e., peak areas) of NO,

N2O, and NO2, prior to the application of UV radiation (i.e., from elapsed time of 0 to 17

min), were stable. Upon exposure of the P25 TiO2 catalyst to UV radiation (starting at an

elapsed time of 17 min and indicated by the arrow symbol and UV “On”), a significant

(i.e., >5 %) decrease in NO ion intensities and an associated increase in N2O ion

intensities were observed. However, when the exposure of P25 TiO2 photocatalyst to UV

65 radiation was terminated or turned “off” (at 42 min elapsed time and indicated by the arrow symbol and UV “Off”), NO and N2O ion intensities returned to their initial levels after a 5-10 minute delay (i.e., comparable to their respective ion intensities prior to exposure to UV radiation). The observed changes in NO and N2O ion intensities, in response to the catalyst exposure to UV radiation, suggest that N2O was the major photocatalytic product of NO. No significant (i.e., >5 %) changes were observed in NO2 ion intensity in either the presence or absence of P25 TiO2 photocatalyst exposed to UV radiation.

3.4. Conclusions

We successfully coupled a home-built post-column cryogenic trap (PCCT) to a conventional GC/MS instrument for the analysis of VOCs. The performance characteristic of GC/PCCT/MS was evaluated and compared to the conventional GC/MS.

The results from the direct HS analysis of acetone using GC/PCCT/MS revealed approximately two orders of magnitude enhancement in LOD compared to the similar analysis using GC/MS. We also demonstrated the advantage of using PCCT for extending the dynamic range (to lower concentrations) for the VOC analysis with a conventional GC/MS. Moreover, we showed an improved S/N for the detection of acetone at low ppbV (i.e., 0.80 ppbV and 3.8 ppbV) using SPME GC/PCCT/MS as compared to SPME GC/MS.

The performance of GC/PCCT/MS for “real world” sample analysis was also evaluated (and compared to GC/MS) using human exhaled breath (HEB) sample from a

“healthy” human subject. Both the number and S/N (at TIC peak maxima) of detected

VOCs in the analyzed breath samples were higher for the GC/PCCT/MS analysis than the

66 direct GC/MS analysis. We anticipate that these improvements in dynamic range and instrumental sensitivity will contribute to emerging research in biomedical and environmental fields that require enhanced analytical capabilities.

Acknowledgments

Financial support from Baylor University and NSF-EAGER (grant award

#1346596) is greatly acknowledged. We would like to thank Mr. David LaBrecque

(University of Maine, Orono, ME) for constructing the capacitor bank used in these experiments. We would also like to thank Mr. Joe McCulloch for constructing the injection port utilized with the GC system used in this work.

67

CHAPTER FOUR

Analysis of Volatile Organic Compound Mixtures Using Radio-Frequency Ionization/Mass Spectrometry

This chapter published as: Olaitan, A. D.; Zekavat, B.; Dhungana, B.; Hockaday, W. C.; Chambliss, C. K.; Solouki, T. Analysis of volatile organic compound mixtures using radio-frequency ionization/mass spectrometry. Anal. Methods, 2014, 6, 4982-4987.

Abstract

Radio-frequency ionization (RFI) coupled with Fourier transform-ion cyclotron resonance mass spectrometry (MS) was used to analyze volatile compounds from bio-oil and commercial gasoline samples. Fingerprinting of bio-oil and gasoline samples was possible using RFI/MS. Differentiation of three gasoline grades was achieved by principal component analysis of RFI/MS data.

4.1. Introduction

Exact molecular formula determination of volatile organic compounds (VOCs) has many analytical (“fingerprinting”) applications in diverse areas of research including

(but not limited to) disease biomarker detection,1, 2 environmental sciences,3, 4 explosive detection,5 forensics,6 and .7, 8 Mass spectrometry (MS) is one of the most sensitive analytical techniques for VOC analysis.9-11 New developments in instrumentation12 and sample collection techniques13-15 should further enhance analytical utilities of MS. For the past few decades, chemical ionization (CI)16 and electron ionization (EI)17 have been among the most commonly employed ionization methods18 for detection and structural analyses of VOCs. Here, we use VOC mixtures to

68 demonstrate the utility of radio-frequency ionization (RFI) technique for MS analysis of organic compounds.

We recently reported on successful implementation of RFI/ Fourier transform-ion cyclotron resonance (FT-ICR) MS for detection of volatile and semi-volatile organic compounds.19 The observed mass spectral patterns in RFI were similar to those generated by EI at 70 eV and contained both pseudo molecular and fragment ions.19 However, RFI offers several advantages that can minimize the challenges associated with the use of EI for VOC analysis with MS. For instance, issues related to the use of EI with FT-ICR MS include: (a) high background pressure in ICR cell due to outgassing of the heated electrical components surrounding the EI filament,9 (b) presence of chemical noise due to ionization of the outgassed materials, and (c) time-penalties associated with frequent replacement of the fragile EI filaments (which are normally placed within a high magnetic field region for FT-ICR instruments). Because RFI can be operated in pulsed mode,19 surface heating and resultant outgassing are minimal. Moreover, robust experimental setup of RFI is not disposed to filament burning and other damages that are commonly observed in EI experiments.

Here, for the first time, we present the utility of RFI/FT-ICR MS for

“sampleprinting” of VOC mixtures. The analyzed mixtures included room temperature

VOCs of aqueous and oily phases of bio-oil samples derived from slow pyrolysis of pine shavings (PS) and corn stover (CS) biomasses and commercially available gasoline samples (research octane number (RON)20 of 87, 89, and 93). We demonstrate that the use of RFI in combination with high mass measurement accuracy and high mass resolving power of FT-ICR MS enables identification of headspace VOCs. Moreover, we

69 show that VOC mixtures can be differentiated by using (i) information obtained from the degree of analyte oxygenation7 using RFI/FT-ICR MS and (ii) principal component analysis (PCA) of RFI/FT-ICR mass spectral data.

4.2. Experimental

4.2.1. Sample Preparation

PS and CS bio-oil samples were prepared by slow pyrolysis in a custom-built reactor.21 Approximately 1,695 g of biomass (PS or CS) was heated in an air-tight, 20- liter, stainless steel reaction vessel for 400 minutes. Pyrolysis temperature programming was: 30 minutes at 100 ºC, ramp at 6 ºC/minute to 470 ºC, 60 minutes at 470 °C to 490

ºC, and cooling at 4 ºC/minute to ambient temperature. Bio-oil condensation was accomplished by cooling pyrolysis products in a “cold fingerˮ heat exchanger at ambient air temperature of approximately 11 ºC. The representative aqueous and oily portions of

PS and CS bio-oil samples were phase separated and used for MS analyses.

Gasoline samples were collected from a commercial source (a gasoline station in

Waco, TX), stored in a 10-mL eppendorf tube, and analyzed within ~20 minutes of sample collection.

4.2.2. Mass Spectrometry (MS)

Mass spectra were acquired using an IonSpec FT-ICR MS system (former

IonSpec Corp. - now a division of Agilent Technologies, Santa Clara, CA) equipped with a 9.4 tesla superconducting magnet (Cryomagnetics Inc., Oakridge, TN) and a home-built

RFI source.19 A detailed description of the FT-ICR instrument has been reported previously.22 Briefly, RFI-generated ions were trapped in an open-ended ICR cell for 1 s.

70

The trapped ions were excited (for 4 ms) by using dipolar frequency sweep excitation23 and detected in the broadband mode. Fourier transformation of the acquired time-domain signals (128 k or 512 k data points) with one zero fill and Blackman window apodization followed by magnitude calculation and frequency-to-m/z conversion yielded the RFI/FT-

ICR mass spectra shown in this report. Typical ultrahigh vacuum (UHV) base pressures, in the ICR cell region, were below ~3.0 × 10-9 Torr. UHV pressures were measured

(when the FT-ICR chamber was inserted into the bore of the magnet for MS operation) by direct reading of Granville-Phillips dual ion gauge controller and series 274 Bayard-

Alpert type ionization gauge tube. Reported pressures were not corrected for ionization sensitivity, geometry factor, or magnetic field effect. To account for m/z-dependent ICR frequency shifts, we used a “segmented” calibration technique.24 The segmented calibration was performed with 5 segments at segment interval of 30 Da.

Liquid samples (~100 μL of gasoline or bio-oil) were transferred into separate teflon tubes and sealed (at room temperature of ~25 °C). Before MS analyses, samples in the sealed teflon tubes were degassed using conventional freeze-thaw degassing cycles.7,

25 Room temperature VOCs present in the headspace volume of teflon tubes (containing either gasoline or bio-oil samples) were transferred into a heated (~200 °C) expansion reservoir.7, 25 Subsequently, these headspace VOCs were directly introduced into the FT-

ICR vacuum chamber through a heated (~200 °C) transfer line attached to a pulsed valve

(~120 °C). Gasoline and bio-oil samples were introduced into the FT-ICR vacuum chamber for 50 ms and 200 ms, respectively. The selected neutral introduction durations were based on an optimal ICR ion detection pressure of ~3.0 × 10-9 Torr after sample introduction. After each analysis, the expansion reservoir was evacuated using a vacuum

71 pump. To assure that sample expansion reservoir was properly cleaned, prior to each analysis, “blankˮ mass spectra were acquired by pulsing contents of the evacuated reservoir into FT-ICR vacuum chamber for mass analysis. In all cases, “blank” mass spectra showed no detectable VOC signals. Samples were analyzed in triplicate runs to ensure reproducibility.

4.2.3. Radio-Frequency Ionization (RFI) Source

We used a home-built RF power supply26 for ion generation. A detailed description of the RFI source has been reported previously.19 An optimized RF signal at

6.5 MHz (~200 Vp-p) was applied to the quadrupole ion guide (QIG) rods of the FT-ICR instrument for 200 ms.19 A transistor-transistor logic (TTL) relay switch and IonSpec

Omega software (version 8.0) was used to control the “on/off” states of the RF signals on

QIG rods. Under our experimental conditions, RFI/FT-ICR mass spectra of VOCs contained radical cations, protonated species, and fragment ions.19

4.2.4. Data Analysis

MATLAB 7.0 (The MathWorks Inc., Natick, MA, USA) software running on a

Dell (2.13 GHz Intel® Core(TM) i3 CPU with 4.0 GHz RAM) laptop computer (Dell Inc.,

Round Rock, TX, USA) was used for principal component analysis (PCA)27 of gasoline

RFI/FT-ICR mass spectra.

4.3. Results and Discussion

Mass Spectral data presented in here demonstrate the utility of RFI/MS for analysis of VOCs in bio-oil and gasoline samples. We selected bio-oil and gasoline as

72 candidate VOC mixtures because the EI/MS analysis of VOCs in bio-oils28-30 and gasoline7, 31 have been reported previously.

4.3.1. RFI/FT-ICR MS Analysis of Bio-Oil Samples

Complete characterization of bio-oils’ compositions in terms of analysis of volatile and non-volatile organic compounds requires the use of appropriate ionization techniques. For example, electrospray ionization (ESI)28 and atmospheric pressure chemical ionization (APCI)32 are routinely used for analysis of high MW species in bio- oil samples. However, the low MW species can be lost in the ESI and/or APCI due to their low proton affinities (PAs)25, 33 as compared to larger MW compounds. Therefore, use of high sensitivity ionization techniques such as RFI is required for full characterization of low MW species in bio-oils.

Figures 4.1(a) and 4.1(b) show the RFI/FT-ICR mass spectra of VOCs present in the headspace of aqueous and oily phases of PS bio-oil, respectively. RFI of the VOCs in aqueous and oily phases of PS bio-oil generated ions in the m/z ranges of 31 to 131 and

39 to 171, respectively. Total numbers of observed peaks (signal-to-noise ratio (S/N) >3) in RFI/FT-ICR mass spectra of aqueous and oily phases of PS bio-oil were 66 and 64, respectively. Identities of the observed ions in Figures 4.1(a) and 4.1(b) could be assigned as saturated, unsaturated, and heteroatom (nitrogen (N), oxygen (O), and sulfur (S))- containing hydrocarbons (Table A.1).

We also analyzed room temperature headspace VOCs from aqueous and oily phases of corn stover (CS) bio-oil samples (Figure A.1 and Table A.1) using RFI/FT-ICR

MS. Observed VOCs in aqueous and oily phases of PS and CS bio-oil samples were mainly O-containing compounds (Table A.1). Total percentages of oxygenated VOCs in

73

aqueous and oily phases of PS bio-oil were 68 % and 64 %, respectively (Table 4.1,

columns 2 and 3). Similarly, total percentages of O-containing VOCs in aqueous and oily

phases of CS bio-oil samples were 82 % and 65 %, respectively (Table 4.1, columns 4

and 5).

(a) PS Aqueous Phase

+ C2H7O

+ CH5O

50 100 150 m/z

(b) PS Oily Phase

+ C5H9O + C6H9O

50 100 150 m/z

Figure 4.1. RFI/FT-ICR mass spectra of VOCs from (a) aqueous and (b) oily phases of bio-oil derived from slow pyrolysis of pine shavings (PS). The two most abundant peaks in each mass spectrum are labeled. For each mass spectrum, 5 FT-ICR MS time-domain transients of 128 k data points were summed prior to Fourier transformation.

The oxygenated VOC compositions of PS and CS bio-oil samples deduced from

RFI/FT-ICR MS analyses in this work are consistent with previously reported gas

74 chromatography (GC) /EI/MS data.28-30 Previous GC/EI/MS data suggested the presence of various O-containing compounds such as acids, alcohols, esters, furans, and ketones as major classes of compounds in bio-oils.28-30

Comparison between RFI/FT-ICR mass spectra of aqueous (Figure 4.1(a)) and oily (Figure 4.1(b)) phases of PS bio-oil revealed clear qualitative differences. For instance, small mass ions (m/z <70) were present at higher relative abundances in the aqueous phase (Figure 4.1(a)) than in the oily phase (Figure 4.1(b)) of PS bio-oil.

Conversely, the larger mass ions (70

4.1(a)) of PS bio-oil. Similar patterns (in terms of relative abundances of small and large mass ions) were observed in the RFI/FT-ICR mass spectra of aqueous and oily phases of

CS bio-oil samples (Figure A.1).

+ PS aqueous phase contained low mass species such as C2H7O (m/z 47.0491), which were not present in PS oily phase. On the other hand, PS oily phase contained

+ higher carbon number sample-specific species, such as C6H7O (m/z 95.0493) which were not detected in PS aqueous phase. Based on the RFI/FT-ICR MS data presented in

Figures 4.1(a) and 4.1(b) (summarized data in Tables 4.1 and A.1), the aqueous and oily phases of PS bio-oil can be differentiated based on their VOC chemical compositions.

The characteristic m/z spacing (e.g., 0.036 Th for the substitution of CH4 vs. O) for various species present in the two phases of PS bio-oil can be observed in RFI/FT-

ICR mass spectra. For instance, Figure 4.2 shows expanded views of the m/z range 100.9 to 101.2 for RFI/FT-ICR mass spectra of aqueous (panel a) and oily phases (panel b) of

PS bio-oil at mass resolving power (m/∆m50%) of ~20,000. The observed nominally

75

+ isobaric species in Figure 4.2 could potentially be assigned as C4H7NS (at m/z 101.0294

+ with 0.3 ppm error, panel a), C4H5O3 (at m/z 101.0234 with 0.5 ppm error, panel b),

+ + C5H9O2 (at m/z 101.0599 with 2.0 ppm error, panels a and b), and C6H13O (at m/z

101.0960 with 0.6 ppm error, panels a and b). The N- and S-containing compounds were only 13.1 % and 1.5 %, respectively, of all the observed species in RFI/FT-ICR mass spectra of PS bio-oil samples, consistent with previous reports on the presence of negligible amount (<0.5 % by weight) of N- and S-containing compounds in bio-oil samples.34-36

Table 4. 1. Percentages of oxygenated and non-oxygenated compounds in PS and CS bio- oil aqueous and oily phases and a commercial gasoline sample (RON 87). O -Containing x PS (Aqueous) PS (Oily) CS (Aqueous) CS (Oily) Gasoline Classes

O0 32 % 36 % 18 % 35 % 91 % O1 39 % 45 % 51 % 45 % 9 %

O2 27 % 14 % 28 % 20 % 0 %

O3 2 % 2 % 3 % 0 % 0 %

O4 0 % 3 % 0 % 0 % 0 %

(a) C H O + 5 9 2 C H NS+ C H O+ 4 7 6 13

(b) + C5H9O2 + + C4H5O3 C6H13O

100.9 101.0 101.1 101.2 m/z Figure 4.2. Expanded views of m/z range 100.9 to 101.2 for RFI/FT-ICR mass spectra of (a) aqueous and (b) oily phases of PS bio-oil. For each mass spectrum, 100 FT-ICR MS time-domain transients of 512 k data points were summed prior to Fourier transformation.

76

4.3.2. RFI/FT-ICR MS Analysis of Commercial Gasoline Samples

We used RFI/FT-ICR MS to compare the VOC compositions of commercially available gasoline samples and bio-oils. Such comparisons can serve as a potential screening step for selecting bio-oils as renewable energy sources for production of high quality bio-fuels.37, 38 The use of high quality bio-fuels (obtained from bio-oils) improves vehicle engine performance and decreases emissions of carbon monoxide (CO), nitrogen

39, 40 oxides (NOx), and hydrocarbons.

Figure 4.3 shows the representative RFI/FT-ICR mass spectra of headspace VOCs from three commercially available gasoline samples with RONs 87 (panel a), 89 (panel b), and 93 (panel c). Unlike PS and CS bio-oil samples, the analyzed gasoline samples contained a large percentage (e.g., 91 % for gasoline sample with RON 87) of non- oxygenated VOCs (Table 4.1, column 6 and Table A.1). Among the observed oxygenated

+ VOCs in the gasoline samples, C2H7O (m/z 47.0491) (Figure 4.3) had the highest relative abundance.

Gasoline samples with three different RONs could be differentiated based on the observed VOCs in their RFI/FT-ICR mass spectra and using principal component analysis (PCA). Figures 4.4(a) and 4.4(b) show the score and loading plots (PC #1 versus

PC #2), respectively, obtained from PCA of RFI/FT-ICR mass spectra of gasoline samples with three different RONs (e.g., 87, circles symbols (○); 89, triangle symbols

(Δ); 93, square symbols (□)). To obtain the score and loading plots in Figures 4.4(a) and

4.4(b), each of the three gasoline samples was analyzed in multiple (ten) trials and the resulting thirty RFI/FT-ICR mass spectra were subjected to PCA. PCA score plot in

Figure 4.4(a) shows a distinct separation between the three gasoline grades. As shown in

77

Figure 4.4(b) loading plot, the three most important m/z values responsible for the

+ observed differences between the analyzed gasoline samples are 57.0695 (C4H9 ),

+ + 71.0851 (C5H11 ), and 85.1012 (C6H13 ). These three species were present in all three

gasoline grades with varied relative abundances (Figure 4.3).

(a) RON = 87 (b) RON = 89

+ + + + + + C5H11 C H C5H11 C 5H11 C5H11 C 5H11 5 11

+ + + + C2H 7OC 2H7O C4H9 C4H9 + + C2H7OC 2H7O + + + + + + C4H9 C4H9 C2H7OC H OC 6H13 C 6H13 + + 2 7 + + C6H13 C 6H13 C4H9 C4H9 + + C6H13 C 6H13

35 70 105 140 35 70 105 140 35 70 105 140 35 70 105 140 35 70 105 140 35 70 105 140 m/z m/z m/z m/z m/z m/z (c) RON = 93

+ + C H + C5H11 C5H11 5 11

+ + C2H7O C4H9 + C2H7O + + + C4H9 C H O C6H13 + 2 7 + C6H13 C4H9 + C6H13

35 70 105 140 35 70 105 140 35 70 105 140 m/z m/z m/z

Figure 4.3. RFI/FT-ICR mass spectra of commercial gasoline samples with RONs of (a)

87, (b) 89, and (c) 93. Four most abundant peaks in each mass spectrum are labeled. For

each mass spectrum, 5 FT-ICR MS time-domain transients of 128 k data points were

summed prior to Fourier transformation.

PCA results in Figure 4.4 demonstrate the applicability of RF/FT-ICR MS for

VOC sample mixture differentiation.

78

(a) Score Plot 0.1 RON = 89 RON = 87 0.05

0 RON = 93

PC2 (23.75 %) -0.05

-0.1 -0.15 -0.075 0 0.075 0.15 PC1 (58.76 %)

(b) Loading Plot

0.5 m/z 85.1012

0

m/z

57.0695

-0.5 PC2 (23.75 %)

m/z 71.0851 -1 -1 -0.5 0 0.5 PC1 (58.76 %) Figure 4.4. (a) Score and (b) loading plots obtained from PCA of RFI/FT-ICR mass spectra of three gasoline grades. Each gasoline sample was analysed in 10 experimental trials. Data points represented by circle (○), triangle (∆), and square (□) symbols correspond to gasoline samples with RONs of 87, 89, and 93, respectively.

A majority of VOCs observed in the RFI/FT-ICR mass spectra of gasoline samples were similar to the previously observed VOCs in gasoline using GC/EI/FT-ICR

MS.7 The observed minor differences are presumably due to different sources of gasoline samples as well as the different experimental conditions used in RFI/FT-ICR MS (this work) and GC/EI/FT-ICR MS.7 One observed difference that could potentially be sample-related is the detection of different O-containing gasoline additives. In our

79 previous GC/EI/FT-ICRMS analysis of a commercial gasoline sample, methyl tert-butyl ether (MTBE) was detected (evidenced by the presence of the MTBE EI fragment ion at

+ 7 m/z 73.0648 (C4H9O )). However, we did not observe MTBE-related ionic species (e.g.,

MTBE radical cation and/or pseudo molecular ion, or fragment ion at m/z 73.0648) in the

RFI/FT-ICR mass spectrum of the gasoline samples analyzed in this study. Instead, we

+ + observed the presence of C2H7O at m/z 47.0491. The C2H7O ion was not observed in our previous GC/EI/FT-ICR MS study and could potentially be assigned as protonated ethanol (chemical ionization (CI) and self-CI25 product of a gasoline additive). These observations are consistent with the current use of ethanol as a gasoline additive in the state of Texas (e.g., the gasoline sample in this study)41 and the former use of MTBE as an additive in the state of Maine gasoline samples42 (e.g., the gasoline sample in our previous GC/EI/FT-ICR MS study). The use of MTBE has been banned in the state of

Maine since 2007.42

4.4. Conclusions

A newly developed RFI/FT-ICR MS technique was utilized to fingerprint gasoline and bio-oil samples. Our MS results suggest that different classes of VOCs (e.g., saturated, unsaturated, and heteroatom-containing hydrocarbons) can be ionized by RFI.

We showed that aqueous and oily phases of bio-oils could be differentiated by RFI/FT-

ICR MS analyses of headspace VOCs. Furthermore, we presented results from gasoline sample classification by combined use of RFI/FT-ICR MS and PCA. Other applications for using this novel ionization technique can be envisioned in metabolomics, for example, for analysis of VOCs in human exhaled breath samples. In the near future, we plan to report on various aspects of RFI including ionization mechanism(s) and sensitivity.

80

Acknowledgments

Financial support from Baylor University and the U.S. Department of Defense

(grant CDMRP-OC060322) is acknowledged.

References

1. M. Phillips, K. Gleeson, J. M. Hughes, J. Greenberg, R. N. Cataneo, L. Baker and W. P. McVay, Lancet, 1999, 353, 1930-1933.

2. B. Buszewski, M. Kesy, T. Ligor and A. Amann, Biomed. Chromatogr., 2007, 21, 553-566.

3. J. S. Alvarado and C. Rose, Talanta, 2004, 62, 17-23.

4. L. Zoccolillo, L. Amendola, C. Cafaro and S. Insogna, J. Chromatogr. A, 2005, 1077, 181-187.

5. K. G. Furton, L. Wu and J. R. Almirall, J. Forensic Sci., 2000, 45, 857-864.

6. K. G. Furton, J. Wang, Y. L. Hsu, J. Walton and J. R. Almirall, J. Chromatogr. Sci., 2000, 38, 297-306.

7. J. E. Szulejko and T. Solouki, Anal. Chem., 2002, 74, 3434-3442.

8. E. D. Christensen, G. M. Chupka, J. Luecke, T. Smurthwaite, T. L. Alleman, K. Lisa, J. A. Franz, D. C. Elliott and R. L. McCormick, Energy & Fuels, 2011, 25, 5462- 5471.

9. T. Solouki, J. E. Szulejko, J. B. Bennett and L. B. Graham, J. Am. Soc. Mass Spectrom., 2004, 15, 1191-1200.

10. C. Heffner, I. Silwal, J. M. Peckenham, and T. Solouki, Environ. Sci. Technol., 2007, 41, 5419-5425.

11. Z. Luo, C. Heffner and T. Solouki, J. Chromatogr. Sci., 2009, 47, 75-82.

12. A. B. Fialkov, U. Steiner, S. J. Lehotay and A. Amirav, Int. J. Mass spectrom., 2007, 260, 31-48.

13. M. Phillips, Anal. Biochem., 1997, 247, 272-278.

14. T. Gorecki and J. Pawliszyn, Field Anal. Chem. Technol., 1997, 1, 277-284.

81

15. H. E. Daughtrey, K. D. Oliver, J. R. Adams, K. G. Kronmiller, W. A. Lonneman and W. A. McClenny, J. Environ. Monit., 2001, 3, 166-174.

16. B. Munson, Anal. Chem. 1971, 43, 28A-43A.

17. K. R. Jennings, Gas Phase Ion Chemistry, Academic Press: New York, 1979.

18. D. F. Hunt, Int. J. Mass Spectrom. Ion Phys., 1982, 45, 111-123.

19. B. Zekavat and T. Solouki, Angew. Chem. Int. Ed. Engl., 2013, 52, 2426-2429.

20. S. B. Chen, The Determination of Research Octane Number of Gasoline Fuels by Octane Analyzer. Defense Technical Information Center: 1983.

21. C. LeCroy, C. A. Masiello, J. A. Rudgers, W. C. Hockaday and J. J. Silberg, Soil Biol. Biochem., 2013, 58, 248-254.

22. J. E. Szulejko, Z. Luo and T. Solouki, Int. J. Mass Spectrom., 2006, 257, 16-26.

23. M. B. Comisarow and A. G. Marshall, Chem. Phys. Lett., 1974, 6, 489-490.

24. J. J. Savory, N. K. Kaiser, A. M. McKenna, F. Xian, G. T. Blakney, R. P. Rodgers, C. L. Hendrickson and A. G. Marshall, Anal. Chem., 2011, 83, 1732-1736.

25. T. Solouki and J. E. Szulejko, J. Am. Soc. Mass Spectrom., 2007, 18, 2026-2039.

26. B. Zekavat, J. E. Szulejko, D. LaBrecque and T. Solouki, Proceedings of the 57th ASMS Conference on Mass Spectrometry and Allied Topics, Philadelphia, PA, 2009.

27. H. Abdib and L. J. Williams, WIREs Comp. Stat., 2010, 2, 433-459.

28. Y. Liu, Q. Shi, Y. Zhang, Y. He, K. H. Chung, S. Zhao, C. Xu, Energy & Fuels, 2012, 26, 4532-4539. 29. A. S. Pollard, M. R. Rover and R. C. Brown, J. Anal. Appl. Pyrolysis, 2012, 93, 129- 138.

30. N. S. Tessarolo, L. R. Dos Santos, R. S. Silva and D. A. Azevedo, J. Chromatogr. A, 2013, 1279, 68-75.

31. F. Liang, K. Kerpen, A. Kuklya and U. Telgheder, Int. J. Ion Mobility Spectrom., 2012, 15, 169-177.

32. V. K. Venkatakrishnan, J. C. Degenstein, A. D. Smeltz, W. N. Delgass, R. Agrawal and F. H. Ribeiro, Green Chem., 2014, 16, 792-802.

82

33. B. Zekavat, J. E. Szulejko, D. LaBrecque, A. D. Olaitan and T. Solouki, Rapid Commun. Mass Spectrom., 2014, 28, 230-238.

34. M. Garcia-Perez, T. T. Adams, J. W. Goodrum, D. P. Geller and K. C. Das, Energy Fuels, 2007, 21, 2363-2372.

35. C. A. Mullen, A. A. Boateng, N. M. Goldberg, I. M. Lima, D. A. Laird and K. B. Hicks, Biomass Bioenergy, 2010, 34, 67-74.

36. O. D. Mante and F. A. Agblevor, Waste Manage., 2012, 32, 67-76.

37. D. M. Alonso, J. Q. Bond and J. A. Dumesic, Green Chem., 2010, 12, 1493-1513.

38. R. H. Venderbosch, A. R. Ardiyanti, J. Wildschut, A. Oasmaa and H. J. Heeres, J. Chem. Technol. Biotechnol., 2010, 85, 674-686.

39. N. Misron, S. Rizuan, A. Vaithilingam, N. F. Mailah, H. Tsuyoshi, Y. Hiroaki and S. Yoshihito, Energies, 2011, 4, 1937-1949.

40. M. Canakci, A. N. Ozsezen, E. Alptekin and M. Eyidogan, Renew. Energy, 2013, 52, 111-117.

41. Window on State Government, http://www.window.state.tx.us/specialrpt/energy/renewable/ethanol.php, Last accessed March 21, 2013.

42. Maine Department of Environmental Protection, http://www.maine.gov/dep/air/mobile/ethanol.html, Last accessed March 21, 2013.

83

CHAPTER FIVE

Ionization Yield as Functions of Electric Field Amplitude and Frequency in Radio- Frequency Ionization (RFI)

Abstract

Radio-frequency ionization (RFI) is a robust and sensitive ionization technique for mass spectrometry (MS) analysis of volatile organic compounds (VOCs). Recently, we reported experimental evidence suggesting that ion generation mechanism in RFI is primarily electron-based and both positive and negative ions are produced with high yields. In here, we further studied the effects of electric field amplitude and frequency on the ion generation yield in RFI. We show that ionization threshold RF voltage amplitude in RFI does not depend on analytes’ ionization energies and polarizabilities, suggesting that RFI is a universal ionization technique. We also provide experimental evidence indicating that lower RF voltages are accessible for RF signals at higher frequencies. RFI signal at 10.5 MHz has a wider range of accessible RF voltages than RF signal at 5.8

MHz. The total ion intensity for acetone obtained at 11.6 MHz (360 Vp-p) has a better signal reproducibility (with signal relative standard deviation (RSD) of 15 ± 8 %) than at

5.3 MHz (360 Vp-p) (with signal RSD of 155 ± 55 %). The experimental findings suggest that RFI sources would operate efficiently and reproducibly at lower RF voltages and higher frequencies.

84

5.1. Introduction

Mass spectrometry (MS) has become a powerful analytical technique helping scientists from a range of professionals including archaeologists, astronomers, biologists, chemists, geologists, materials scientists, physicist, and physicians for unveiling applied and fundamental questions at the molecular and atomic levels.28, 150, 261 Ionization, as a first step of molecular and atomic analyses using MS, plays a central role for improving the sensitivity of MS techniques.28, 262 Therefore, parallel to various hardware and software innovations, ionization techniques have contributed significantly to advancing

MS technologies.28, 261

Depending on the physico-chemical properties of analytes of interest, specific ionization techniques are utilized to generate ions from neutral analytes. For example, electrospray ionization (ESI)94 and matrix-assisted laser desorption ionization (MALDI)98 are widely used for ionization of semi- and non-volatile chemical compounds from liquid and solid phases, respectively. Atmospheric pressure chemical ionization (APCI)96 and atmospheric pressure photoionization (APPI)263 produce higher ion yields than ESI from non-polar analytes in liquid phase.

Ionization techniques of choice for analysis of volatile organic compounds

(VOCs) are electron ionization (EI),91, 151, 152, 264 chemical ionization (CI),35, 158 field ionization (FI),93, 166, 168 photoionization (PI)147 as well as a recently introduced atmospheric pressure gas chromatography (APGC) ionization.265 Electron ionization is by far the most commonly utilized ionization technique for VOC analysis and involves the interaction of gaseous neutral molecules with an energetic (70 eV) electron beam to generate ionic species. Molecular ions generated in EI (a “hard” ionization technique)

85 have a wide range of internal energies and are subjected to fragmentation.149 Although the molecular ion fragmentation in EI provides valuable structural information for unknown identification, extensive fragmentation obscures molecular weight determination in EI experiments.150 To increase the molecular ion abundances, EI can be operated at lower electron energies; however, low energy electron beams decrease EI ionization yields.149

“Soft” ionization techniques such as CI, FI, and PI produce mostly molecular ions and increase the confidence for molecular weight identification of unknown analytes.150

However, a major challenge in CI technique is the difficulty in MS data reproducibility which depends on experimental conditions (e.g., primary electron energy and current, reagent gas pressure, and ion source temperature) which must be carefully controlled.28, 35

Field ionization involves the utilization of high electric field (~109 to 1010 Vm-1) to ionize a neutral molecule (in the gas phase).166 A major drawback of FI is the requirement of high potential (8-12 kV) for generating high electric field necessary for ionization.166, 167

Photoionization involves the use of photons with specific wavelength to selectively ionize organic molecules.147 Degree of fragmentation of molecular ions in PI can be controlled based on the photon wavelength and/or power density of the laser source utilized for ionization.28, 147 Complexity of the ion source and setup makes PI less common ionization technique for routine analyses.

In an effort towards developing an ionization technique, which brings source design simplicity, robustness, and high sensitivity together for VOC analysis, we recently introduced radio-frequency ionization (RFI).176 We previously showed that RFI has higher ionization yield that EI/CI in both positive- and negative-ion modes.50 This novel

86 ionization technique has been utilized for sample “finger-printing” and analysis of VOCs from complex mixtures such as bio-oil and gasoline samples.181, 266

In a recent study, we showed that RFI is an electron-based ionization technique in which positive and negative ions are generated by electron ejection and electron capture, respectively.267 In here, we present experimental data showing the effects of analyte’s ionization energy (IE) and polarizability (α) on the RFI threshold voltage amplitude for ion generation in positive-ion mode RFI/MS. Correlation between applied RF voltage and ion yield at both low and high RF signals (i.e., frequencies) is also discussed.

5.2. Experimental

5.2.1. Samples

Standard VOC samples were purchased from Sigma Aldrich (St. Louis, MO,

USA) and were used without further purification. Argon (Ar) gas was purchased from

Praxair (Danbury, CT, USA). For all MS experiments, 100 μL of liquid standard VOC sample (at room temperature) was placed in a 1 mL glass vial (VWR Inc., Radnor, PA,

USA) and the headspace volume above the liquid was expanded into a heated (200 ˚C) expansion reservoir. Subsequently, headspace VOC molecules from each standard sample was introduced directly into the FT-ICR MS vacuum chamber (for 2 ms) through a heated (200 ˚C) transfer line attached to a pulse valve (Parker Hannifin Corp.,

Cleveland, OH, USA). Background mass spectra were acquired before and after each sample to confirm that there was no sample contamination in each experiment. All MS analyses were conducted in triplicates to ensure reproducibility.

87

5.2.2. FT-ICR MS

RFI mass spectra were acquired using an IonSpec FT-ICR mass spectrometer

(former IonSpec Corp. – now a division of Agilent Technologies, Santa Clara, CA,

USA), equipped with a 9.4 tesla superconducting magnet (Cryomagnetics Inc., Oakridge,

TN, USA). Detailed description of the FT-ICR instrument has been previously reported268 and are described briefly in here. Following RFI, generated ions were trapped in an open-ended ICR cell for 1 s and excited (for 4 ms) using a dipolar frequency sweep excitation.269 Detection of the RFI-generated ions was done in the broadband mode. The acquired time-domain signals (128 k data points) with one zero fill and Blackman window apodization were subjected to Fourier transformation, magnitude calculation, and frequency-to-m/z conversion134 to produce the RFI/FT-ICR mass spectra shown in this report. Calibration of the RFI mass spectra was done using known m/z values of molecular and fragment ions from standard VOCs. Ultrahigh vacuum (UHV) base pressures in the ICR cell region during MS experiments (with the FT-ICR chamber inserted into the magnet bore) were below ~8.0 x 10-10 torr (pressures were not corrected for ionization sensitivity, geometry factor and magnetic field effects). UHV pressures were measured with a Granville-Phillips dual ion gauge controller and series 274 Bayard-

Alpert type ionization gauge tube.

5.2.3. RFI Setup

RFI setup consisted of either a quadrupole (with total capacitance of ~230 pF)176 or dipole assembly (with total capacitance of ~40 pF).267 The quadrupole RFI setup consisted of 4 cylindrical aluminum rods (outer diameter (O.D.) 0.63 cm and length

119.38 cm) while the dipole RFI set up consisted of 2 cylindrical aluminum rods (O.D.

88

0.63 cm and length 6.50 cm). Several advantages are inherent with utilizing smaller RFI electrodes such as lower capacitance and smaller source size. RFI-generated ions were obtained by applying an RF signal from a home-built RF power supply270 to the QIG rods or dipole rods (for 200 ms) using a transistor-transistor logic (TTL) relay switch controlled by the IonSpec Omega software (version 8.0).

5.3. Results and Discussion

It has been shown that gas sensors based on breakdown discharge current exhibit analyte-specific breakdown voltage.271 As with any ionization technique, it is important to characterize the effects of analyte’s properties on ionization yields. In the following section, we present experimental data from RFI/MS analysis of a series of VOCs with varied (a) ionization energies (IE) and (b) polarizabilities (α). Data presented in this section point to the negligible effects of analytes’ physico-chemical properties (e.g., IE and α) on threshold RF voltage amplitude.

5.3.1. Analyte Ionization Energy and Polarizability versus Threshold RF Voltage Amplitude

Figure 5.1 shows the plots of analyte IE (Figure 5.1(a)) and α (Figure 5.1(b)) versus threshold RF voltage amplitude at an applied RF signal of 10.4 MHz. IE and α values were obtained from the NIST chemistry webbook.272 Threshold RF voltage amplitude is defined as the minimum RF voltage (at a constant frequency) applied to the

RFI electrodes that yielded an RFI/FT-ICR MS signal (with signal-to-noise ratio (S/N)

>3) of an analyte molecule. To obtain the data in Figure 5.1, we introduced (for 2 ms) each analyte (specified on the x-axis of Figure 5.1) into the FT-ICR mass spectrometer vacuum chamber in a separate experiment and recorded the minimum RF voltage

89 amplitude required to observe an RFI/FT-ICR signal of the analyte molecule. Threshold

RF voltage amplitudes in Figure 5.1 are average values from three different measurements and error bars at 95 % confidence level (CL) are included for each value.

As shown in Figure 5.1, threshold RF voltage amplitude (at constant frequency) for ion generation in RFI does not vary significantly as a function of analyte’s IE or α. For example, analytes with IEs ranging from 8.60 to 15.76 eV and α ranging from 6.33 to

3 10.45 Å were successfully ionized using RFI at RF voltage >40 Vp-p. At RF voltage <40

Vp-p (at 10.4 MHz), there was no MS signal for all the analyzed VOCs which could be due to insufficient kinetic energy to get the RF-generated electrons into the ICR cell for the ionization event.267

Although the data in Figure 5.1 are sorted based on analyte’s IE (Figure 5.1(a)) and α (Figure 5.1(b)), the data also show that threshold RF voltage amplitude does not depend on analyte class (e.g., alcohol versus aromatic). Observations in Figure 5.1 suggest that RFI is an unbiased ionization technique and RFI is suitable for ionizing

a) Threshold RF Voltage Amplitude versus Ionization Energy 70

)

p - p 60 V 50

40

30 Argon

Acetone

20 Chloroform

Methanol Isopropylamine

Chlorobenzene 10 Propionate Methyl ( Voltage RF Threshold 0 8.60 9.07 9.70 10.15 10.84 11.37 15.76 Ionization Energy (eV) 90

b) Threshold RF Voltage Amplitude versus Polarizability

70

)

p -

p 60

50

40

30

Heptene

Pentadiene

-

-

1

Benzene Acetone

20 1,4

Diethyl Ether Diethyl

Hexafluorobenzene Chlorobenzene

10 Threshold RF Voltage (V Voltage RF Threshold 0 6.33 8.79 9.19 10.40 11.38 12.14 13.45 Polarizability (Å3)

Figure 5.1. Plots of threshold radio-frequency (RF) voltage (Vp-p) at 10.40 MHz versus (a) ionization energy (eV) and (b) polarizability (Å3). Ionization energy and polarizability values were obtained (for the standard samples) from NIST chemistry webbook (webbook.nist.gov/chemistry/). Error bars are reported at the 95 % CL and n= 3.

different classes of compounds. Also, a wide range of VOCs with different functional groups has been successfully analyzed using RFI-FT-ICR MS both in the positive- and negative-ion modes.273

5.3.2. RFI MS Signal as a Function of Threshold RF Voltage Amplitude and Frequency

In the following section, we present experimental data showing the relationship between the observed RFI signal and the applied RF voltage amplitude and frequency.

Also, a wider range of RF voltages is accessible at RF signal at a higher frequency than at a lower frequency.

Figure 5.2 shows a plot of threshold RF voltage amplitude for RFI of acetone as a function of frequency (in the range of ~3.0MHz to ~12.0 MHz). In Figure 5.2, each data

91

point (denoted with empty circles, ○) correspond to threshold RF voltages with average

values from three separate measurements and error bars reported at the 95 % CL. The

large error bars seen in Figure 5.2 are due to the phase dependency of electrons/ions

during RFI. As shown in Figure 5.2, the threshold RF voltage required to observe an RFI

MS signal (S/N >3) is lower at higher frequencies (e.g., 75 Vp-p at ~12.0 MHz) than at

lower frequencies (e.g., ~325 Vp-p at ~3.0 MHz). Moreover, as we discuss later, our RFI

FT-ICR MS data acquired over a wide range of frequencies suggest that ionization

process in RFI is more reproducible at higher frequencies (as compared with lower

frequencies); hence, it is desirable to use higher RF signals in RFI. In addition, larger

ranges of RF voltages are accessible and can be utilized for RFI at higher frequencies

than at lower frequencies.

Threshold RF Voltage Amplitude versus Frequency 450 Analyte: Acetone

) p

- 400 p 350 300

250 200 150

100

50 Threshold RF Voltage (V Voltage RF Threshold

0 3.0 4.5 6.0 7.5 9.0 10.5 12.0 Frequency (MHz)

Figure 5.2. Plot of threshold radio-frequency (RF) voltage (Vp-p) versus frequency (applied to QIG rods) for acetone performed in three separate experimental trials and error bars are at 95 % CL.

92

Recently, we demonstrated that electrons are involved in the ionization mechanism of RFI.267 The minimum RF voltage required to obtain RFI FT-ICR MS signal could be related to the minimum kinetic energy required by the RFI-generated electrons to get into the ICR cell (placed within a 9.4 T magnetic field) for ionization.

In Figure 5.3, total ion intensity (at m/z values of 43, 58, and 59) for acetone was plotted against RF voltage (Vp-p) at RF signals of 10.50 MHz and 5.8 MHz applied to the

QIG rods. Each data point (denoted with empty circles, ○) in Figure 5.3 correspond to an average of three experimental trials with error bars at the 95 % CL. In Figure 5.3(a), a steady increase in total ion intensity for acetone was observed as RF voltage was increased from ~176 to ~332 Vp-p. The total ion intensity for acetone at RF voltage of

~332 Vp-p is not significantly different (at 95 % CL) from the total ion intensities for acetone at ~372 or ~384 Vp-p indicating that the optimum RF voltage amplitude for maximum (>80 % total ion intensity) ion yield occurred at ~330 Vp-p for RF signal at

10.5 MHz.

a) RFI Signal @10.5 MHz

100 Analyte: Acetone

80

60

40

20 IonIntensity (%) Total 0

150 200 250 300 350 400

RF Voltage (Vp-p) 93

b) RFI Signal @ 5.8 MHz 100 Analyte: Acetone 80

60

40

20 Total IonIntensity (%) Total

0 200 250 300 350 400 450 500 RF Voltage (V ) p-p Figure 5.3. Plots of total ion intensity (%) versus radio-frequency (RF) voltage (V ) p-p (applied to QIG rods) for acetone at (a) 10.5 MHz and (b) 5.8 MHz. Error bars are reported at the 95 % CL and n= 3.

At RF signal of 5.8 MHz (in Figure 5.3(b)), an increase in the total ion intensity for acetone was observed at RF voltage of ~212 to ~308 Vp-p. Total ion intensities at RF voltage of ~308 to ~472 Vp-p are not significantly different (at 95 % CL) from each other.

The potential of obtaining ion yields at lower RF voltages is one of the advantages of performing RFI at higher frequencies than at lower frequencies. Also, at higher frequency

(e.g., at RF signal of 10.5 MHz), the linear range of the applied RF voltage (i.e., from

~176 to ~332 Vp-p) for obtaining RFI MS signal is larger than the linear range of the applied RF voltage (i.e., from ~212 to ~308 Vp-p) at RF signal of 5.8 MHz.

5.3.3. RFI MS Signal Stability as a Function of Frequency

We have recently demonstrated that it is possible to utilize RFI electrodes with shorter dimensions than the original QIG rods of the IonSpec FT-ICR MS system for

94

RFI176, 181 without reduction in ionization sensitivity.274 The new RF electrodes with reduced dimensions (i.e., a short dipole assembly made of two 6.5-cm long (OD = 0.63 cm) aluminum rods) enable us to have access to RF signal at higher frequencies due to the dipole’s reduced capacitance (~40 pF). As we discuss in the following section, our results suggest that ionization process in RFI is more reproducible at higher frequencies

(as compared with lower frequencies); hence, it is desirable to use higher RF signals in

RFI.

Figure 5.4 shows the plots of the total ion intensity (in the positive-ion mode) for acetone as a function of acquisition time at high and low RFI frequencies. In Figure

5.4(a) and Figure 5.4(b), the RFI FT-ICR total ion chromatogram was acquired at an RF signal of 11.6 MHz (~360 Vp-p) and 5.3 MHz (~360 Vp-p) applied to dipole RFI electrodes, respectively. Plot shown in Figure 5.4(a) was reconstructed from a total number of ~900 mass spectra acquired over a period of ~180 mins. For comparison purposes, Figure 5.4(b) shows a plot of total ion intensity for acetone as a function of acquisition time (~65 mins) at an applied RF signal of 5.3 MHz (~360 Vp-p). As shown in

Figure 5.4, RFI MS signal is more reproducible at higher frequencies (i.e., at 11.6 MHz) than at lower frequencies (i.e., 5.3 MHz). For example, the calculated relative standard deviation (RSD) for RFI signal stability is 15 (±8) % at 11.6 MHz and 155 (±55) % at 5.3

MHz. Errors (in parentheses) are at the 95 % CL for three experimental trials (n=3).

The observed high scan-to-scan reproducibility of RFI MS signal has potential advantages for example, in coupling RFI sources to gas chromatography (GC)/MS and ion mobility (IM)/MS instruments where long duration signal stability is desired. Insets in Figures 5.4(a) and (b) show RFI/FT-ICR mass spectra of acetone acquired at elapsed

95

a) RFI Signal @ 11.6 MHz Elapsed time = 100 min + [M - CH3] [M+H]+ Signal RSD: 15 (±5) % ●+ 18 M

15 40 50 60 m/z 12

9 Signal

6

Total Ion Intensity Ion Total 3

0 No Signal 0 30 60 90 120 150 180 Elapsed Time (min) b) RFI Signal @ 5.3 MHz Elapsed time = 24 min + [M - CH3]

[M+H]+ Signal RSD: 155 (±55) % M●+

40 50 60 8 m/z

6

4 Signal

2 Total Ion Intensity Ion Total

0 No Signal 0 10 20 30 40 50 60

Elapsed Time (min)

Figure 5.4. Plots of acquisition time versus positive-ion mode RFI/FT-ICR MS total ion intensity of acetone at RFI frequencies of (a) 11.6 MHz (360 Vp-p) and (b) 5.3 MHz (360 Vp-p) applied to dipole RFI electrodes. Insets in (a) and (b) show RFI/FT-ICR mass spectra of acetone acquired at elapsed time of 100 and 24 minutes, respectively. Relative standard deviation (RSD) for ion generation processes is lower at higher RFI frequency. Errors are reported at the 95 % CL and n= 3. 96 time of 100 and 24 minutes, respectively. RFI generates both molecular (M●+),

+ + 176 pseudomolecular ([M+H] ), and fragment (e.g., [M-CH3] ) ions. The relative higher ion intensity for the [M+H]+ specie than for the M●+ specie is due to self-chemical ionization (CI) within the ICR cell (as a result of time-dependent ion-molecule reactions from long delay time in FT-ICR MS experiments).

5.4. Conclusions

We presented results on the effects of threshold RF voltage amplitude and frequency on the ion generation yield in RFI. RFI/FT-ICR MS results showed that RFI is a universal ionization technique and does not depend on analyte’s physico-chemical properties such as ionization energies and polarizabilities. Furthermore, our RFI/FT-ICR

MS results suggest that RF signal at higher frequencies produce more stable and reproducible RFI signal (and with wider range of accessible RF voltages) than at lower frequencies. In the future, we plan to report on the construction and development of a robust and efficient RFI source for non-magnet based MS instruments that will utilize RF signals at higher frequencies and lower RF voltages.

Acknowledgment

Financial support from Baylor University, the U.S. Department of Defense

(DOD) (grant CDMRP-OC060322), and National Science Foundation-Instrument

Development for Biological Research (NSF-IDBR) (grant award # 1455668) is acknowledged.

97

CHAPTER SIX

Evidence for Electron-Based Ion Generation in Radio-Frequency Ionization

This chapter published as: Olaitan, A. D.; Zekavat, B.; Solouki, T. Evidence for Electron- Based Ion Generation in Radio-Frequency Ionization. J. Mass Spectrom., 2015, DOI 10.1002/jms.3703.

Abstract

Radio-frequency ionization (RFI) is a novel ionization method coupled to mass spectrometry (MS) for analysis of semi-volatile and volatile organic compounds (VOCs).

Despite the demonstrated capabilities of RFI MS for VOC analysis in both positive- and negative-ion modes, mechanism of RFI is not completely understood. Here, we studied the possibility of electron emission in RFI using direct charged particle current measurements and indirect electron detection in a 9.4 tesla Fourier transform-ion cyclotron resonance (FT-ICR) MS. We show that RF-generated electrons can be trapped in the ICR cell and subsequently reacted with neutral hexafluorobenzene (C6F6)

●- ●- molecules to generate C6F6 . Intensity of C6F6 species correlated with the number of trapped electrons and decreased as a function of electron quenching period. We also measured the electron attachment rate constant of hexafluorobenzene using a post-RF electron trapping experiment. Measured electron attachment rate constant of hexafluorobenzene (1.19 (±0.53) × 10-9 cm3 molecule-1 s-1) for post-RF FT-ICR MS agreed with the previously reported value (1.60 (±0.30) × 10-9 cm3 molecule-1 s-1) from low-pressure ICR MS measurements. Experimental evidences from direct and indirect

98 electron detection techniques suggest that RFI process involves RF-generated electrons under ultrahigh vacuum conditions. Better understanding of the ion generation process in

RFI should expand its utility in MS.

6.1. Introduction

Advances in mass spectrometry (MS) are often correlated with innovation and new capabilities to generate ions in the gas phase. Therefore, development of robust, sensitive, and simple ionization techniques has become an important and a growing field in MS.[1-3] Currently, a large set of ionization techniques are available for MS analysis of semi- and non-volatile chemical compounds (e.g., electrospray ionization (ESI),[4, 5] matrix-assisted laser desorption ionization (MALDI),[6] desorption electrospray ionization

(DESI),[7-9] and laser ablation electrospray ionization (LAESI)[10]). On the other hand, conventional ionization techniques for analysis of volatile organic compounds (VOCs) include electron ionization (EI),[11-13] chemical ionization (CI),[14, 15] and field ionization.[16-18] EI as a “hard” ionization technique provides structural information and

CI as a “soft” ionization technique provides molecular mass information.[19] Main disadvantage of EI for VOC analysis is the extensive fragmentation of molecular ions, which makes the identification of parent ion challenging when analyzing complex sample mixtures.[19, 20]

Recently, other ionization techniques such as ultralow-voltage field-ionization[21] and carbon nanotube-based ionization[22] have been introduced for VOC analysis.

Although they have not yet been coupled to MS systems, these two new ionization techniques require low voltages for operation (e.g., 6 V in ultralow-voltage field

99 ionization[21] as opposed to 8 kV to 12 kV in field ionization[18]) and small sizes of their ion sources could be helpful for developing miniature gas sensors.

Most common applications of EI for VOC analysis include positive-ion mode detection; however, negative ions can also be produced via capturing low energy electrons (<1 eV) by electrophilic VOCs in a so called “negative chemical ionization- mass spectrometry (NCI-MS)” technique.[23, 24] Slow electrons in NCI-MS are generally generated by EI of a reagent gas molecule (e.g., methane) at high (70 eV) energy and subsequent cooling of secondary electrons by electron-neutral collisions.[25] The use of a buffer gas for electron energy cooling in NCI-MS can yield negative ion abundance irreproducibility issues that are related to variations in electron energies.[25]

To overcome the challenges associated with electron energy variations in NCI-

MS, the use of trochoidal electron monochromator (TeM) devices has been proposed.[26-

28] TeM devices operate by using crossed electric (E) and magnetic (B) fields for electron energy selection and do not require the use of a buffer gas for electron cooling. Since their introduction by Stamatovic,[27] TeM devices have been coupled to various mass spectrometers (e.g., quadrupoles, time-of-flight analyzers, and magnetic sectors) for analysis of electronegative compounds.[29-31]

Recently, we introduced a novel radio-frequency ionization (RFI) technique coupled to Fourier transform-ion cyclotron resonance (FT-ICR) MS for analysis of volatile organic compounds (VOCs).[32, 33] RFI is a highly efficient ionization method in both positive- and negative-ion modes.[32-34] Although various classes of organic compounds can be successfully ionized using RFI,[33, 34] the exact mechanism(s) of ion generation during the RFI event has not been characterized. Understanding the ion

100 generation mechanism(s) in RFI is important for: (i) optimizations of RFI electrodes and source geometries for constructing highly efficient and versatile ion sources, (ii) further development of the ionization technique for MS analysis of volatile and non-volatile compounds in positive- and negative-ion modes, and (iii) utilization of this new ionization source with a wide range of MS systems.

In here, we present experimental data that suggest RFI involves electrons that are emitted from the ion source electrodes during the ionization process. We discuss results from direct ion (positive and negative) current measurements during the RFI event. MS data from indirect detection of RF-generated electrons by observation of

●- hexafluorobenzene (HFB) negative ions (C6F6 ) produced by electron attachment in the

ICR cell of a 9.4 tesla FT-ICR mass spectrometer are also presented. In addition, experimental data from ICR electron quenching as well as electron attachment reaction rate constant measurements of HFB are discussed.

6.2. Experimental

6.2.1. Samples

Acetone and hexafluorobenzene samples were purchased from Sigma Aldrich (St.

Louis, MO, USA) and used without further purification. For MS experiments, 100 μL of a liquid standard VOC (at room temperature) was placed in a 1 mL glass vial (VWR Inc.,

Radnor, PA, USA) and the headspace volume was expanded into a heated (200 ˚C) expansion reservoir (total volume 300 mL). Acetone or hexafluorobenzene were then introduced directly into the FT-ICR MS vacuum chamber through a heated (200 ˚C) transfer line attached to a pulse valve (maintained at 120 ˚C) (for acetone introduction in charged particle current measurement experiments) or a pulsed-leak valve setup[35] (for

101 controlled introduction of neutral hexafluorobenzene in electron attachment reaction rate measurement experiments).

6.2.2. FT-ICR MS

Positive- and negative-ion mode mass spectra were acquired using an IonSpec

FT-ICR mass spectrometer (former IonSpec Corp. – now a division of Agilent

Technologies, Santa Clara, CA, USA) equipped with a 9.4 tesla superconducting magnet

(Cryomagnetics Inc., Oakridge, TN, USA). RFI-generated ions were trapped in an open end ICR cell. Trapped ions were excited using a dipolar frequency sweep excitation[36]

(sweep rate of 1.6 × 109 Hz/s in the range of ~7.2 MHz to ~720 kHz corresponding to m/z range of 20 to 200 in 9.4 tesla magnetic field) and detected in broadband mode

(bandwidth of 6.5 MHz). Time-domain signals (128 k data points) with one zero fill and

Blackman window apodization were subjected to Fourier transformation, magnitude calculation, and frequency-to-m/z conversion[37] to produce the FT-ICR mass spectra shown in Figure 6.2. RFI mass spectra were calibrated using known m/z values of molecular, pseudo molecular, and fragment ions from acetone and hexafluorobenzene.

The vacuum chamber pressures were measured using a Granville-Phillips dual ion gauge controller and a series 274 Bayard-Alpert type ionization gauge tube (Granville-Phillips

Company - now part of MKS Instruments, Longmont, CO, USA). Direct readouts of the ionization gauge were corrected for ionization sensitivity (IS),[38] geometry factor

(GF),[39] and magnetic field effect (MFE)[40] using a previously reported procedure.[40]

Ionization sensitivity of hexafluorobenzene was calculated based on a previously reported relationship between IS and molecular polarizability (α) (i.e., IS = 0.36 α + 0.30).[38]

Calculated IS for hexafluorobenzene with α = 9.58 Å3[41] was 3.75. Considering GF of

102

0.55,[39] MFE of 1.43, and IS of 3.75 (for hexafluorobenzene), we divided the direct pressure readouts by 2.95 (= GF (0.55) × MFE (1.43) × IS (3.75)) to obtain the corrected hexafluorobenzene pressures for reaction rate constant calculations.

6.2.3. RFI Source

A previously reported RFI source design (consist of a set of quadrupole rods as

RFI electrodes[32]) was used for charged particle current measurements. For all other experiments reported in here, an RFI source with modified RFI electrode configurations and sizes was used. The modified RFI electrode design consisted of a set of two parallel aluminum rods (dipole) (each rod with a length of 6.5 cm and an outer diameter (OD) of

3.0 mm) with a total capacitance (including the capacitance for BNC cable and connections) of ~40 pF. The new RFI dipole electrode design with reduced capacitance

(as compared with the original quadrupole assembly with total capacitance of ~180 pF)[32,

33] allowed access to higher frequency RF signals with peak-to-peak voltages in excess of

400 V (e.g., f = 12.1 MHz with Vp-p = 440). RFI-generated ions were obtained by applying an RF signal from a home-built RF power supply[42, 43] to the dipole RFI electrode assembly (for 200 ms) using a transistor-transistor logic (TTL) relay switch controlled by IonSpec Omega software (version 8.0).

6.2.4. Current Measurements

Charged particle currents were measured using an existing flame ionization detector (FID) amplifier installed on a gas chromatograph (SRI instruments model

8610C, Las Vegas, NV, USA). The SRI gas chromatograph amplifier was set to “High

Gain/Filtered” for all the measurements reported in here. Charged particle currents were

103 recorded using Peak Simple (version 2.83, SRI Instrument, Las Vegas, NV, USA) and converted to ASCII format. Charged particle current data in ASCII format were plotted, as in Figure 6.1, using Origin 7.0 software (Origin Lab Co., MA, USA).

6.3. Results and Discussion

Previous data from RFI/MS analyses of various classes of VOCs showed that RFI was capable of generating both positive and negative ions with higher ion yields than

EI;[32, 44] however, it was not clear whether ion generation mechanism in RFI was a surface phenomenon or a gas-phase event involving electron emission from RFI electrode surfaces.

In the following sections, we present results from direct and indirect detection of electrons generated during the RFI event. Results from electron detection experiments suggest that ion generation mechanism in RFI involves gas-phase phenomenon in which positive and negative ions can be generated by electron impact and electron attachment, respectively.

6.3.1. Charged Particle Current Measurements during RFI

To investigate the possibility of electron generation/emission in RFI, we used the

ICR cell’s quadrupole trapping plate (QTP) as a charge collector electrode and measured charged particle currents during the RFI event. All current measurements were conducted with the ICR cell positioned inside a 9.4 tesla magnet bore and using a flame ionization detector (FID) amplifier installed on a gas chromatography (GC) system (see

Experimental section for more details). ICR cell background pressure was ~1.0 × 10-9

Torr during the ion current measurement experiments.

104

Figure 6.1 shows results from current measurements of RFI-generated charged particles. Panels a and b in Figure 6.1 show results from two separate measurements; for each measurement, charged particle current was plotted as a function of elapsed time.

The current measurements were conducted for durations of 1200 s and 90 s for plots shown in panels a and b (Figure 6.1), respectively. Data in Figure 6.1 were obtained from fifty (panel a) and six (panel b) consecutive measurements with 10-s intervals. Currents in excess of 2.5 nA were recorded for measurements in panel b; and hence, measurements were continued for a shorter period of time (than panel a) to protect the operational amplifier of the FID detector from electrical damages. Charged particle currents in Figure

6.1 (panels a and b) were measured at 50 Hz sampling frequency and with quadrupole direct current (DC) offset voltages (with respect to ground) of 0.0 V (panel a) and -50 V

(panel b). An RF signal of 6.5 MHz (200 Vp-p) for duration of 1 s was applied to the quadrupole rods and charged particle currents were measured with (Figure 6.1(a)) and without (Figure 6.1(b)) acetone pulses into the FT-ICR MS vacuum chamber.

(a) Positive Ion Signals 0.18 20-ms Acetone 5-ms Acetone Pulse Pulse

0.16 A) n

0.14

No Acetone 0.12 Current ( Current Pulse

0.10 0 300 600 900 1200 Elapsed Time (s)

105

(b) Electron + Negative Ion Signals

3.0

2.5

2.0

1.5 1.0 Current (nA) Current 0.5 0.0 20 40 60 80 Elapsed Time (s)

Figure 6.1. Temporal plots of (a) positive (fifty measurements) and (b) electron/negative (six) ion currents measured on QTP of the ICR cell at an RF frequency of 6.5 MHz (200 V p-p). Measurement interval for both panels (a) and (b) are 10 s.

As shown in Figure 6.1(a), measured charged particle currents correlate with the

number density of acetone molecules introduced into the vacuum chamber. For example,

with no acetone pulse, a background current of ~0.15 nA (region labeled as “No Acetone

Pulse” in Figure 6.1(a)) could be detected on QTP when 6.5 MHz (200 Vp-p) RF signal

was applied to the quadrupole. However, simultaneous application of 6.5 MHz (200 Vp-p)

RF signal to the quadrupole and introduction of acetone neutral molecules into the

vacuum chamber increased the measured ion current to values above the background ion

current. For instance, 5 ms acetone pulse into the vacuum chamber increased the ion

current by an average of ~10 pA (region labeled as “5-ms Acetone Pulse” in Figure 6.1(a)

with downward ion current peak shift from ~0.15 nA to ~0.14 nA). Increasing the

duration of neutral acetone pulse into the vacuum chamber (from 5 ms to 20 ms) raised

the measured ion current from ~10 pA (for 5 ms acetone pulse) to an average of ~30 pA

(region labeled as “20 ms Acetone Pulse” in Figure 6.1(a) with downward ion current

106 peak shift from ~0.15 nA to ~0.12 nA). Acetone has a low electron affinity (~1.5 meV))[45] and the observed correlation between the measured currents (magnitude of downward peak shifts in Figure 6.1(a)) with acetone number densities suggest that measured currents in Figure 6.1(a) are from positive ions. Presumably, these positive ions are molecular and fragmented acetone ions generated by the applied RF to quadrupole rods.

To detect electrons/negative ions, a DC offset voltage of -50 V was applied to the quadrupole rods and no acetone neutrals were pulsed into the FT-ICR vacuum chamber.

As expected, application of a -50 V DC voltage to the quadrupole rods (to detect negative ions) changed the direction of current change from downward peaks in Figure 6.1(a) to upward peaks (from ~0 nA to ~2.7 nA). Moreover, the measured currents in Figure 6.1(b) were approximately two orders of magnitude larger than the measured currents in Figure

6.1(a). Observations in Figure 6.1(b) (i.e., change in the direction of charged particle current in the presence of a negative DC voltage and increase in the measured current in the absence of neutral molecule pulse) are consistent with generation/detection of electrons during the RFI process. Note that there might be a minor contribution from negative ions (generated from electron attachment reactions of residual electronegative neutrals) to the total measured currents in Figure 6.1(b); however, additional neutral pulse and electron attachment experiments (as discussed later in this section) confirm that the major portion of the measured currents displayed in Figure 6.1(b) are from RF generated electrons.

107

In the next section, we show that electrons are indeed generated during the RFI event and can be trapped in the ICR cell of a 9.4 tesla FT-ICR MS for indirect detection via attachment to high electron affinity neutral molecules (e.g., hexafluorobenzene).

6.3.2. Indirect Detection of RF-Generated Electrons

Results from direct measurement of charged particles (Figure 6.1) provide evidence for electron generation during RFI. To further support this hypothesis, we attempted to trap RF generated electrons within the ICR cell and use them to generate electron attached negative ions.[46, 47] In this section, we present negative-ion mode post-

RF FT-ICR MS data that confirm the important role of electron emission in RFI. For electron attachment experiments, we selected hexafluorobenzene (C6F6) which has an electron affinity of ~1.8 eV[48] and a high cross section (~1.23 × 10-14 cm2)[49] for capturing low energy (~0.03 eV) thermal electrons.

Scheme 6.1 shows an experimental event sequence for indirect detection of RF- generated electrons. To indirectly probe the generation of electrons during RFI, we first applied appropriate negative and positive voltages to ICR trapping plates for 50 ms to eject any potential background charged species (including electrons and ions of both polarity). After ensuring a successful ICR quench event and obtaining a clean background (by collecting and inspecting blank mass spectra not shown here), we applied an RF signal at 12.0 MHz (360 Vp-p) and duration of 200 ms to the RFI electrodes. The selection of 12.0 MHz for electron attachment experiments was based on our previous findings (unpublished results) that showed RFI signals were more reproducible at 12

MHz than at our previously reported 6.0 MHz experiments. RF generated electrons were allowed to enter into the ICR cell by lowering the potential of trapping plate (near the

108

RFI electrodes) from -30 V to +30 V for 10 ms. Following a 10-ms electron extraction time into the ICR cell, trapping plate potentials for both sides of the ICR cell were decreased to either -0.5 V, -9.0 V or -30 V (in separate experiments). Following a 2-s delay time for electron cooling/relaxation, hexafluorobenzene (HFB) neutral molecules were pulsed into the FT-ICR vacuum chamber for 1 ms (from a reservoir containing HFB at ~10 Torr) by using a miniature solenoid pulse valve (Parker Hannifin Corp.,

Cleveland, OH, USA). Subsequently, HFB neutrals were allowed to react with the trapped electrons for 4 s; trapped negative ions were chirp excited[50] and detected in broadband mode.

- ICR Cell RFI e Quench Time C6F6 Introduction Excitation Quench (50 ms) (200 ms) (variable) (1 ms) Detection

Delay Time Delay Time40

(2 s) (4 s) 20

0 0 0.5 1

Amplitute -20

-40 Time Time

Scheme 6.1. Sequence of events for electron (e-) quenching experiments and e- attachment rate constant measurements.

It has been shown that the dominant energy decay mechanism for electrons in strong magnetic fields is radiative decay (i.e., emission of electromagnetic waves).[46, 51,

52] For instance, the calculated radiative decay time constant[46, 51, 52] for an electron with cyclotron frequency of ~263 GHz in a 9.4 tesla magnetic field is ~29 ms. Therefore, a ~2- s post-electron injection delay time (before introduction of C6F6 neutrals into the FT-ICR vacuum chamber) should be sufficient to “thermalize” electrons for reaction with C6F6. In separate experiments, electron quenching events were also used (Scheme 6.1, event

109 labeled as “e- Quenching Time”) to further confirm the production of electrons during

RFI. Details of these electron quenching experiments are discussed in the following sections.

Representative results from indirect MS-based electron detection experiments are included in Figure 6.2. Figure 6.2(a) shows a negative-ion mode post-RF FT-ICR mass spectrum acquired for C6F6 in the absence of an electron quenching event (i.e., the variable “e- Quench Time” in Scheme 6.1 was set to zero). Mass spectrum in Figure

●- 6.2(a) shows the presence of C6F6 molecular ion (C6F6 ) as the base peak and other

- - minor intensity fragment ions ([C6F6 - CF2] and [C6F6 - C3F3] ). Consistent with direct current measurement in Figure 6.1, observation of negative ions in Figure 6.2(a) suggests that electrons are generated during the RFI event.

To further confirm that the negative ions in Figure 6.2(a) were products of the reaction between RF-generated electrons from RFI electrodes and HFB, an electron quenching event was introduced after 2 s delay time following an RFI pulse of 200 ms

(i.e., the variable “e- Quenching Time” in Scheme 6.1 was set to 5 and 13 ms). For post-

RF electron quenching experiments, ICR cell inner ring and end trapping plates were pulsed to 0.0 V during the quench event. Figures 6.2(b) and 6.2(c) show negative-ion mode post-RF FT-ICR mass spectra of HFB after 5 ms and 13 ms electron quenching time, respectively. Comparison between Figure 6.2(a) and Figure 6.2(b) indicates that

●- there is a decrease in signal-to-noise ratio (S/N) for C6F6 signal when an electron

●- quenching time of 5 ms was used (i.e., S/N values of 146 and 48 for C6F6 in Figures

●- 6.2(a) and 6.2(b), respectively). No C6F6 signal was observed when an electron

110

quenching time of 13 ms (Figure 6.2(c)) was used (an indirect indication that all RF

generated electrons trapped within ICR cell were ejected or “quenched”).

- (a) e Quench time = 0 ms ●- C6F6

-C3F3 -CF2 * * * 50 100 150 200 m/z

(b) e- Quench time = 5 ms ●- C6F6

-CF2 *

50 100 150 200 m/z

(c) e- Quench time = 13 ms

No MS Signal

* 50 100 150 200 m/z

Figure 6.2. Negative-ion mode post-RF FT-ICR mass spectra of hexafluorobenzene after (a) 0 ms, (b) 5 ms, and (c) 13 ms of electron (e-) quenching. The * symbols represent electronic noise.

111

●- Figure 6.3 shows a plot of total C6F6 intensity in post-RF FT-ICR mass spectrum of HFB as a function of electron quenching time (in the range of 1 ms to 13 ms). Each data point in Figure 6.3 corresponds to an average value obtained from three separate experimental trials. In Figure 6.3, error for each data point is reported at the 95

●- % confidence level (n = 3). As is evident in Figure 6.3, C6F6 intensity consistently

●- decreases as a function of electron quenching time. Signal reduction for C6F6 is an indirect indication that increasing the e- quenching time decreases the number of RF generated trapped electrons within the ICR cell.

4 - ●- 3 C6F6 + e  C6F6

2

1

Total Ion Intensity Ion Total 0 0 2 4 6 8 10 12 14 Quenching Time (ms)

Figure 6.3. Plot of total ion intensity as a function of post-electron injection quenching time for hexafluorobenzene. Error bars are reported from three experimental trials at the 95 % confidence level.

We also measured the electron attachment rate constant for HFB by monitoring

●- the appearance of C6F6 species as a function of reaction time between HFB and thermalized RF-generated electrons. Results from electron attachment reaction rate constant measurements are discussed in the following section.

112

6.3.3. Electron Attachment Reaction Kinetics of Hexafluorobenzene (HFB)

For electron attachment reaction kinetics, we studied a well-characterized electron attachment reaction (i.e., reaction of HFB with thermal electrons) to measure HFB’s

●- electron attachment rate constant. If C6F6 species in Figure 6.2 were the products of reaction between RF-generated electrons and HFB neutrals, one would expect to observe

●- generation of C6F6 species (in negative-ion mode FT-ICR MS) with a rate constant that is equal to HFB’s electron attachment rate constant.

Electron attachment reaction kinetics of HFB (reaction 6.1) has been previously studied in both high (300 Torr to 5000 Torr of a buffer gas)[49] and low (<3 × 10-6

Torr)[53] pressure regimes:

C F ●- k 6 6 f * C F + e- {C F ●-} (Reaction 6.1) 6 6 6 6 k b C F ●- + IR radiation 6 6

In reaction 6.1, “kf” and “kb” denote rate constant values for forward (electron attachment) and backward (electron auto-detachment) reactions, respectively. “kc” and

“kr” denote collisional and radiative (infrared (IR)) relaxation/stabilization rate constants

●- ●- * ●- for excited C6F6 , respectively. “{C6F6 } ” denotes a vibrationally excited C6F6 ion.

We assumed that only thermal electrons were involved in reaction 6.1 and therefore, dissociative electron attachment[53] reactions of HFB were ignored; this is a reasonable assumption since we used a sufficiently long electron trapping time (~2 s).

Gant et al. reported a rate constant of ~1.02 × 10-7 cm3 molecule-1 s-1 for electron attachment reaction of HFB using high pressure electron swarm experiments.[49] Woodin

113 et al. measured an electron attachment rate constant of ~1.60 × 10-9 cm3 molecule-1 s-1 for

HFB using low pressure ICR MS experiments.[53] Woodin et al. interpreted the discrepancy between low and high pressure electron attachment reaction rate constants of

●- * HFB in terms of different relaxation/stabilization mechanisms of {C6F6 } (reaction 6.1) in the two pressure regimes.[53]

In other words, in electron swarm experiments, collisional relaxation is the

●- * dominant ion relaxation mechanism for stabilization of {C6F6 } in the presence of high pressure buffer gas (e.g., 300 Torr to 5000 Torr).[49] Note that under high pressure conditions, kc>> kb, kr (reaction 6.1) and contribution of unimolecular electron auto- detachment reaction is negligible. Therefore, the electron attachment rate constant of

[49] HFB at high pressure regime is equal to kf. Conversely, the dominant ion relaxation

●- * mechanism for {C6F6 } (reaction 6.1) in the low pressure regime of ICR MS experiments is through radiation (i.e., radiative relaxation, kr in reaction 6.1) and kc<< kr< kb. Under low pressure electron attachment reactions, electron auto-detachment reaction

(kb in reaction 6.1) becomes non-negligible and electron attachment rate constant of HFB is defined as:[53]

kapp = kf · (kr / kb) (Equation 6.1)

In equation 6.1, “kapp” denotes the apparent electron attachment rate constant (as opposed to the actual electron attachment rate constant, i.e., kf).

Here, we measured the apparent electron attachment rate constant of HFB by

●- following the production of C6F6 as a function of reaction time between neutral HFB molecules and thermalized RF-generated electrons at constant HFB pressure (~3.4 × 10-8

Torr). As described in the following section, a previously reported procedure was used to

114 obtain the electron attachment rate constant of HFB.[46, 53] Experimental sequence of events used for electron attachment rate constant measurements of HFB was similar to the one shown in Scheme 6.1, except no “e- Quench Time” (Scheme 6.1) was used and

“C6F6 Introduction” (Scheme 6.1) duration was varied from 0 ms to 600 ms.

●- To obtain kapp for the reaction between HFB and thermalized electrons, C6F6 ion

●- abundance in post-RF FT-ICR mass spectrum after each reaction time ([C6F6 ]t) was

●- ●- normalized to C6F6 ion abundance at the longest reaction time (i.e., 600 ms, [C6F6 ]600 ms). Normalized intensity (NI) data were then fitted to the following equation using

Origin 7.0 software (Origin Lab Co., MA, USA):

●- ●- NI = [C6F6 ]t / [C6F6 ]600 ms = {1 - exp(- t / τapp)} (Equation 6.2)

In equation 6.2, “t” and “τapp” denote reaction time and apparent time constant, respectively. Subsequently, kapp was calculated by using equation 6.3:

kapp = 1 / (N · τapp) (Equation 6.3)

In equation 6.3, “N” denotes the number density of neutral HFB molecules in the

FT-ICR MS vacuum chamber.

Figure 6.4 shows a typical plot of normalized intensity as a function of reaction

●- -8 time for generation of C6F6 at HFB pressure of ~3.38 × 10 Torr. By fitting the temporal e- attachment kinetic data for HFB (e.g., Figure 6.4) to equation 6.2, we obtained an apparent time constant (τapp) of 770 ms. Substituting this time constant

(obtained from curve fitting) into equation 6.3 and using HFB number density (N) of 1.09

× 109 molecules/cm3 (at pressure of ~3.4 × 10-8 Torr), we obtained an average e-

-9 3 -1 -1 attachment rate constant (kapp) of 1.19 (±0.53) × 10 cm molecule s for HFB (error from four experimental trials at the 95 % confidence level is included in the parenthesis).

115

Our experimentally obtained e- attachment rate constant for HFB using post-RF FT-ICR

MS experiments closely agrees with the previously reported low pressure ICR MS value for HFB’s apparent e- attachment rate constant (1.60 (±0.30) × 10-9 cm3 molecule-1 s-1).[53]

1.0

0.8

0.6

0.4 kapp - ●- C6F6 + e  C6F6 0.2 k = 1.19 (±0.53) × 10-9 cm3 mol-1 s-1

0.0 Normalized Intensity Normalized 0 200 400 600 Reaction Time (ms)

●- Figure 6.4. Plot of normalized intensity for C6F6 as a function of reaction time with RF-generated electrons. Average reaction rate constant is reported from a total of four measurements (each measurement was based on a kinetic plot constructed from ~29 data points) and error is reported at the 95 % confidence level.

●- Our electron attachment rate constant measurements further confirmed that C6F6 species were products of reaction between C6F6 and RF-generated electrons. Therefore, reported experimental electron detection data suggest that the reported RFI is a gas-phase phenomenon (although surface ionization and/or analyte desorption under appropriate experimental settings might also occur and are subjects of our future studies). Note that previously reported high electric field-based ionization techniques such as field

116 desorption/field ionization[54] and spontaneous desorption[55-57] are surface ionization phenomena and do not involve ionization of VOCs in the gas phase.

Although it would have been helpful to measure the kinetic energy ranges of RF- generated (or field-emitted) electrons by using “bracketing” type experiments or appearance energy measurements,[58] the existing instrumental setup limited such options.

For instance, using “bracketing” experiments, energies of electrons generated under the influence of a DC potential (e.g., in electron ionization (EI)) could be determined by utilizing analytes of varied ionization energies (for cations) and electron affinities (for anions).[39] However, such “bracketing” approaches may not provide accurate electron energy measurements in the presence of RF fields (as is the case with RFI). Our electron/ion trajectory simulations (data not shown here) suggest that the velocities and therefore, kinetic energies of electrons are primarily influenced by RF electric fields imposed on the RFI electrodes. For example, our ion trajectory simulations indicate that in the presence of an RF signal (applied to the RFI electrodes), the charged particles

(including electrons) undergo several cycles of acceleration and deceleration before

“escaping” from the RF field. Therefore, the RF-generated electrons (during the application of RF signal to the RFI electrodes) could have kinetic energies ranging from as high as several hundred eV to as low as 0 eV (depending on the RF signal amplitude and frequency). Hence, it is not feasible with our current instrumental setup to use bracketing type experiments to accurately determine the influence of experimental factors on observed electron energies during the RFI processes.

117

6.3.4. Potential Mechanism(s) for Electron Generation in RFI

Experimental data presented in previous sections suggest that (i) electrons are generated by application of an RF signal to RFI electrodes and (ii) RFI is a gas-phase phenomenon in which electrons are involved in the ionization process. The question to address is: what are the mechanisms of electron generation/emission from RFI electrode surfaces?

Electron emission from metal surfaces can occur when surface electrons gain enough energy that exceed metal work functions.[59] For example, in thermionic[60] and photoelectric[61] effects, electrons leave the metal surfaces by thermal excitation and photon absorption, respectively. Electron emission from metal surfaces, known as field emission, can also occur by application of an intense electric field to a metallic (cathode) surface under vacuum.[62] In field emission (or “cold cathode” emission), electrons are generated by electron tunneling through metal work function energy barrier.[62] The quantum mechanical explanation of the field emission has been previously described by

Fowler and Nordheim (F-N) in the context of a classical model.[62] For brevity, we only discuss the most relevant projection model of field emission described by Jimenez et al.[63] in the following section.

In the projection model of field emission, the assumption is that there are sharp and micron-sized projections on the surface of a metallic cathode.[63] These microscopic projections cause local electrical field enhancements of ~100 times larger than the applied field at the projection tips.[63] High electric field enhancements can be achieved in the proximity of metallic electrodes with small curvature where the electric field is locally amplified as compared with planar electrode shapes.[64] Theoretically, the electron

118 emission from each one of these micron-sized sites on the electrode surface follows the

F-N theory.[62] For metal electrodes with large surface areas, typical electric field threshold for electron emission is as low as a few MV/m.[62]

Another mechanism for enhanced electron generation under application of a high electric field to a metal electrode is Townsend ionization/avalanche.[65] In Townsend ionization/avalanche, electron current generated from a cathode electrode surface is amplified by passing the emitted electrons through a gas-filled tube and in the presence of a high DC electric field between two parallel plates (cathode and anode). Acceleration of electrons from cathode to anode and subsequent collisions with gas molecules/atoms results in multiplication of initially generated photoelectric electrons. The electron amplification increases by increasing the distance between the two electrodes at a constant electric field.

Among the mechanisms for electron generation described in the previous paragraphs, field emission is the most probable mechanism for electron generation from

RFI electrodes under ultrahigh vacuum conditions. However, further investigations are necessary to check the validity of field emission phenomenon during RFI.

6.4. Conclusions

Using direct charged particle current measurements and indirect electron detection in a 9.4 tesla FT-ICR MS system, we have shown that electrons play an important role in mechanism of ion generation in RFI. Mechanistic understanding of the

RFI processes opens the door for design/construction of RFI sources with optimized electrode geometries and high ion yields. The high sensitivity advantage of RFI[32, 34] for both positive and negative ion detection combined with the simplicity and scalability of

119 the source design, will enhance functionalities of MS instruments for analyses of complex environmental and biological sample mixtures.

Acknowledgments

Financial support from the U.S. Department of Defense (DOD) (CDMRP-

OC060322) and National Science Foundation-Instrument Development for Biological

Research (NSF-IDBR) (award number: 1455668) is acknowledged.

References

[1] M. E. Monge, G. A. Harris, P. Dwivedi, F. M. Fernandez. Mass spectrometry: recent advances in direct open air surface sampling/ionization. Chem. Rev. 2013, 113, 2269.

[2] S. Trimpin, B. Wang, C. B. Lietz, D. D. Marshall, A. L. Richards, E. D. Inutan. New ionization processes and applications for use in mass spectrometry. Crit. Rev. Biochem. Mol. Biol. 2013, 48, 409.

[3] M. Kandiah, P. L. Urban. Advances in ultrasensitive mass spectrometry of organic molecules. Chem. Soc. Rev. 2013, 42, 5299.

[4] J. B. Fenn, M. Mann, C. K. Meng, S. F. Wong, C. M. Whitehouse. Electrospray ionization for mass spectrometry of large biomolecules. Science 1989, 246, 64.

[5] C. L. Hendrickson, M. R. Emmett. Electrospray ionization Fourier transform ion cyclotron resonance mass spectrometry. Annu. Rev. Phys. Chem. 1999, 50, 517.

[6] F. Hillenkamp, M. Karas, R. C. Beavis, B. T. Chait. Matrix-assisted laser desorption/ionization mass spectrometry of biopolymers. Anal. Chem. 1991, 63, 1193A.

[7] L. S. Eberlin, I. Norton, D. Orringer, I. F. Dunn, X. Liu, J. L. Ide, A. K. Jarmusch, K. L. Ligon, F. A. Jolesz, A. J. Golby, S. Santagata, N. Y. Agar, R. G. Cooks. Ambient mass spectrometry for the intraoperative molecular diagnosis of human brain tumors. Proc. Natl. Acad. Sci. USA 2013, 110, 1611.

[8] D. R. Ifa, N. E. Manicke, A. L. Dill, R. G. Cooks. Latent fingerprint chemical imaging by mass spectrometry. Science 2008, 321, 805.

120

[9] J. M. Wiseman, D. R. Ifa, Y. Zhu, C. B. Kissinger, N. E. Manicke, P. T. Kissinger, R. G. Cooks. Desorption electrospray ionization mass spectrometry: Imaging drugs and metabolites in tissues. Proc. Natl. Acad. Sci. U S A 2008, 105, 18120.

[10] P. Nemes, A. S. Woods, A. Vertes. Simultaneous imaging of small metabolites and lipids in rat brain tissues at atmospheric pressure by laser ablation electrospray ionization mass spectrometry. Anal. Chem. 2010, 82, 982.

[11] A. J. Dempster. The properties of slow canal rays. Phys. Rev. 1916, 8, 651.

[12] A. O. Nier. A mass spectrometer for routine abundance measurements. Rev. Sci. Instrum. 1940, 11, 212.

[13] K. R. Jennings. Gas Phase Ion Chemistry, Academic Press: New York, 1979.

[14] M. S. B. Munson, F. H. Field. Chemical ionization mass spectrometry. J. Am. Chem. Soc. 1966, 88, 2621.

[15] B. Munson. Chemical ionization mass spectrometry. Anal. Chem. 1971, 43, 28A.

[16] E. W. Muller, K. Bahadur. Field ionization of gases at a metal surface and the resolution of the field ion microscope. Phys. Rev. 1956, 102, 624.

[17] H. D. Beckey. Experimental techniques in field ionisation and field desorption mass spectrometry. J. Phys. E: Sci. Instrum. 1979, 12, 72.

[18] R. P. Lattimer, H.-R. Schulten. Field ionization and field desorption mass spectrometry: past, present, and future. Anal. Chem. 1989, 61, 1201A.

[19] T. Kind, O. Fiehn. Advances in structure elucidation of small molecules using mass spectrometry. Bioanal. Rev. 2010, 2, 23.

[20] D. F. Hunt. New ionization techniques in mass spectrometry. Int. J. Mass Spectrom. Ion Phys. 1982, 45, 111.

[21] R. Banan Sadeghian, M. S. Islam. Ultralow-voltage field-ionization discharge on whiskered silicon nanowires for gas-sensing applications. Nat. Mater. 2011, 10, 135.

[22] A. Modi, N. Koratkar, E. Lass, B. Wei, P. M. Ajayan. Miniaturized gas ionization sensors using carbon nanotubes. Nature 2003, 424, 171.

[23] D. F. Hunt, F. W. Crow. Electron capture negative ion chemical ionization mass spectrometry. Anal. Chem. 1978, 50, 1781.

121

[24] R. C. Dougherty. Negative chemical ionization mass spectrometry. Anal. Chem. 1981, 53, 625A.

[25] E. A. Stemmler, R. A. Hites. Methane enhanced negative ion mass spectra of hexachlorocyclopentadiene derivatives. Anal. Chem. 1985, 57, 684.

[26] A. Stamatovic, G. J. Schulz. Trochoidal electron monochromator. Rev. Sci. Instrum. 1968, 39, 1752.

[27] A. Stamatovic, G. J. Schulz. Characteristics of the trochoidal electron monochromator. Rev. Sci. Instrum. 1970, 41, 423.

[28] E. Illenberger, H.-U. Scheunemann, H. Baumgärtel. Negative ion formation in CF2Cl2, CF3Cl and CFCl3 following low energy (0-10 eV) impact with near monoenergetic electrons. Chem. Phys. 1979, 37, 21.

[29] J. A. Laramee, C. A. Kocher, M. L. Deinzer. Application of a trochoidal electron monochromator/mass spectrometer system to the study of environmental chemicals. Anal. Chem. 1992, 64, 2316.

[30] V. G. Voinov, Y. V. Vasil'ev, J. Morre, D. F. Barofsky, M. L. Deinzer, M. Gonin, T. F. Egan, K. Fuhrer. A resonant electron capture time-of-flight MS with trochoidal electron monochromator. Anal. Chem. 2003, 75, 3001.

[31] A. J. Dane, C. D. Havey, K. J. Voorhees. The detection of nitro pesticides in mainstream and sidestream cigarette smoke using electron monochromator-mass spectrometry. Anal. Chem. 2006, 78, 3227.

[32] B. Zekavat, T. Solouki. Radio-frequency ionization of organic compounds for mass spectrometry analysis. Angew. Chem. Int. Ed. 2013, 52, 2426.

[33] A. D. Olaitan, B. Zekavat, B. Dhungana, W. C. Hockaday, C. K. Chambliss, T. Solouki. Analysis of volatile organic compound mixtures using radio-frequency ionization/mass spectrometry. Anal. Methods 2014, 6, 4982.

[34] T. Solouki, B. Zekavat, A. D. Olaitan. A novel radio-frequency ionization method for the analysis of wide classes of volatile organic compounds. Proceedings of the 61st ASMS Conference on Mass Spectrometry and Allied Topics, Minneapolis, MN 2013, 9-13 June.

[35] C. Q. Jiao, D. R. A. Ranatunga, W. E. Vaughn, B. S. Freiser. A pulsed-leak valve for use with ion trapping mass spectrometers. J. Am. Soc. Mass Spectrom. 1996, 7, 118.

[36] M. B. Comisarow, A. G. Marshall. Frequency-sweep Fourier transform ion cyclotron resonance spectroscopy. Chem. Phys. Letters 1974, 6, 489.

122

[37] A. G. Marshall, C. L. Hendrickson. Fourier transform ion cyclotron resonance detection: principles and experimental configurations. Int. J. Mass Spectrom. 2002, 215, 59.

[38] J. E. Bartmess, R. M. Georgiadis. Empirical methods for determination of ionization gauge relative sensitivities for different gases. Vacuum 1983, 33, 149.

[39] T. Solouki, J. E. Szulejko. Bimolecular and unimolecular contributions to the disparate self-chemical ionizations of alpha-pinene and camphene isomers. J. Am. Soc. Mass Spectrom. 2007, 18, 2026.

[40] B. Zekavat, M. Miladi, A. Alfdeilat, A. Somogyi, T. Solouki. Evidence for sequence scrambling and divergent H/D exchange reactions of doubly-charged isobaric b type fragment ions. J. Am. Soc. Mass Spectrom. 2014, 25, 226.

[41] T. M. Miller. Atomic and molecular polarizabilities, 95th ed. In CRC Handbook of Chemistry and , D. R. Lide, H. P. R. Frederikse (Eds). CRC Press Inc.: Baton Rouge, LA, 193.

[42] B. Zekavat, J. E. Szulejko, D. LaBrecque, T. Solouki. Low cost higher frequency RF power supplies for quadrupole ion guides in FT-ICR instruments with external ion sources. Proceedings of the 57th ASMS Conference on Mass Spectrometry and Allied Topics, Philadelphia, PA 2009, 31 May-4 June.

[43] B. Zekavat, J. E. Szulejko, D. LaBrecque, A. D. Olaitan, T. Solouki. Efficient injection of low-mass ions into high magnetic field Fourier transform ion cyclotron resonance mass spectrometers. Rapid Commun. Mass Spectrom. 2014, 28, 230.

[44] A. D. Olaitan, B. Zekavat, B. Dhungana, W. C. Hockaday, C. K. Chambliss, T. Solouki. Application of radio frequency ionization mass spectrometry for volatile organic compound analysis. 69th Southwest Regional Meeting of the American Chemical Society, Waco, TX, 2013, 19 November.

[45] C. Desfranҫois, H. Abdoul-carime, N. Khelifa, J. P. Schermann. From 1/r to 1/r2 potentials: electron exchange between Rydberg atoms and polar molecules. Phys. Rev. Letter 1994, 73, 2436.

[46] G.-Z. Li, H. S. Kim, S. Guan, A. G. Marshall. Radiatively self-cooled penning- trapped electrons: a new way to make gas-phase negative ions from neutrals of low electron affinity. J. Am. Chem. Soc. 1997, 119, 2267.

[47] V. Frankevich, R. Knochenmuss, R. Zenobi. The origin of electrons in MALDI and their use for sympathetic cooling of negative ions in FTICR. Int. J. Mass Spectrom. 2002, 220, 11.

123

[48] C. Lifshitz, T. O. Tiernan, B. M. Hughes. Electron affinities from endothermic negative ion charge transfer reactions. IV. SF6, selected fluorocarbons and other polyatomic molecules. J. Chem. Phys. 1973, 59, 3182.

[49] K. S. Gant, L. G. Christophorou. Attachment of slow electrons to hexafluorobenzene. J. Chem. Phys. 1976, 65, 2977.

[50] A. G. Marshall. Fourier transform ion cyclotron resonance mass spectrometry. Acc. Chem. Res. 1985, 18, 316.

[51] L. S. Brown, G. Gabrielse. Geonium theory: Physics of a single electron or ion in a . Rev. Mod. Phys. 1986, 58, 233.

[52] G.-Z. Li, S. Guan, A. G. Marshall. Sympathetic cooling of trapped negative ions by self-cooled electrons in a Fourier transform ion cyclotron resonance mass spectrometer. J. Am. Soc. Mass Spectrom. 1997, 8, 793.

[53] R. L. Woodin, M. S. Foster, J. L. Beauchamp. Ion cyclotron resonance studies of radiative and dissociative electron attachment processes at low pressures. J. Chem. Phys. 1980, 72, 4223.

[54] H. D. Bekey. Field desorption mass spectrometry: A technique for the study of thermally unstable substances of low volatility. Int. J. Mass Spectrom. Ion Physics 1969, 2, 500.

[55] S. Della-Negra, Y. L. Beyec, P. Htiansson. Spontaneous desorption time-of-flight mass spectrometry (SDMS): Time correlated emission of electrons and negative ions in a constant electric field. Nucl. Instrum. Methods B 1985, 9, 103.

[56] S. Della-Negra, C. Deprun, Y. L. Beyec, F. Rgllgen, K. Standing, B. Monart, G. Bolbach. On the mechanism of ion formation in spontaneous desorption time-of- flight mass spectrometry. Int. J. Mass Spectrom. Ion Processes 1987, 75, 319.

[57] M. Salehpour, J. E. Hunt. Spontaneous desorption mass spectrometry: Applications and mechanisms. Int. J. Mass Spectrom. Ion Processes 1988, 85, 99.

[58] Z. Luo, C. Heffner, T. Solouki. Multidimensional GC-Fourier transform ion cyclotron resonance MS analyses: utilizing gas-phase basicities to characterize multicomponent gasoline samples. J. Chromatogr. Sci. 2009, 47, 75.

[59] H. L. Skriver, N. M. Rosengaard. Surface energy and work function of elemental metals. Phys. Rev. B 1992, 46, 7157.

[60] C. Herring, M. H. Nichols. Thermoionic emission. Rev. Mod. Phys. 1949, 21, 185.

[61] H. Y. Fan. Theory of photoelectric emission from metals. Phys. Rev. 1945, 68, 43.

124

[62] R. H. Fowler, L. Nordheim. Electron emission in intense electric fields. Proc. R. Soc. London. Series A 1928, 119, 173.

[63] M. Jimenez, R. J. Noer, G. Jouve, J. Jodet, B. Bonin. Electron field emission from large-area cathodes: evidence for the projection model. J. Phys. D: Appl. Phys. 1994, 27, 1038.

[64] S. Podenok, M. Sveningsson, K. Hassen, E. E. B. Campbell. Electric field enhancement factors around a metallic, end-capped cylinder. NANO 2006, 1, 87.

[65] J. S. Townsend. The Conductivity produced in gases by the motion of negatively- charged ions. Nature 1900, 62, 340.

125

CHAPTER SEVEN

Conclusions and Future Directions

CHAPTER ONE: In Chapter One, as an introductory section, the importance of utilizing mass spectrometry (MS) for volatile organic compound (VOC) analysis was highlighted. Chapter One also discussed the general approaches for improving sensitivity in MS (e.g., the development of efficient ionization sources, mass analyzers, detectors, and sampling methods). Strengths and weaknesses of different sampling techniques, including membrane introduction MS, cryogenic trapping system, headspace (HS) analysis (static and dynamic), and the use of solid phase micro-extraction (SPME) were discussed. Additionally, operating principles of the techniques utilized to collect the data presented in this dissertation, particularly, gas chromatography (GC), quadrupole MS, and Fourier transform-ion cyclotron resonance (FT-ICR) MS were detailed.

CHAPTER TWO: In Chapter Two, the commonly employed ionization techniques for MS analysis of VOCs were presented. Discussion was focused on the history, methodology, and performance characteristics of electron ionization (EI), chemical ionization (CI), field ionization (FI), and photoionization (PI) sources. The instrumental details and overview of the radio-frequency ionization (RFI) method were also introduced in Chapter Two. Moreover, the four main processes that electron emission can occur in metals (viz., thermionic emission, field emission, secondary emission, and photo emission) were presented in Chapter Two.

126

CHAPTER THREE: Details on improving sensitivity for MS analysis of VOCs using a home-built post-column cryogenic trap (PCCT) coupled to a GC/quadrupole MS system were presented in Chapter Three. Analytical figures of merit, including dynamic range and signal-to-noise ratio (S/N), were evaluated for a conventional GC/MS system with and without the use of PCCT. Compared to the conventional GC/MS detection system (without any preconcentration), the use of PCCT improved the detection limit (for the standard acetone and VOCs in human exhaled breath samples examined in this research) by approximately two orders of magnitude. The use of direct HS analysis and

SPME was combined with the GC/PCCT/MS system to enhance the analysis and detection of standard acetone and VOCs such as ethanol, isoprene, benzene, and toluene present in human exhaled breath samples at the low parts-per-billion (ppb) concentrations.

Future applications of the GC/PCCT/MS system could be employed for the analysis of a wide variety of organic compounds present in trace concentrations (i.e., parts-per-trillion levels) in several different sample matrixes, such as in air and soil. For example, in separate experiments in our lab, analysis of catalytic decomposition products of nitric oxide in the presence of a titanium oxide (TiO2) catalyst and ultra-violet (UV) radiation was achieved through the coupling of a cryogenic trap to a photocatalytic reaction chamber and a triple quadrupole MS system. Further development could be directed towards manufacturing miniaturized cryogenic trapping systems such as inexpensive thermoelectric heat pump that do not rely on liquid cryogens (e.g., liquid nitrogen) to concentrate analytes and could be coupled to portable mass spectrometers for online VOC monitoring in the field.

127

CHAPTER FOUR: The application of RFI for VOC analysis in gasoline and bio- oil samples was presented in Chapter Four. It was demonstrated that RFI was an efficient ionization technique for analysis of VOCs in standard and complex samples. RFI was combined with FT-ICR MS (that offers both ultrahigh resolving power and mass measurement accuracy) to detect the volatile compounds present in gasoline and bio-oil for sample fingerprinting (specifically aqueous and oily phases of pine shaving and corn stover samples). RFI and high resolving power capability of FT-ICR MS were utilized to determine the percent composition of oxygenated and non-oxygenated species in gasoline and bio-oil samples. Principle component analysis was employed to differentiate three gasoline grades (i.e., gasoline with different research octane numbers) based on the VOC compositions of the different gasoline grades as observed in RFI FT-ICR mass spectra of the gasoline samples.

Since RFI is an efficient and sensitive method for ionization of VOCs, an extension of the RFI technique to non-volatile organic compound analysis could be a plausible research to pursue. Preliminary data suggest that desorption of non-volatile compounds from the surface (placed in the path of the RF-generated electrons) is possible with RFI. For example, Figure 7.1(a) shows the RFI FT-ICR mass spectrum of the instrument background vacuum chamber (hydrocarbon species) obtained at RF signal of

12.2 MHz (400 Vp-p). The RFI FT-ICR mass spectrum shown in Figure 7.1(a) was obtained without pulsing neutrals into the vacuum chamber of the FT-ICR mass spectrometer. Figure 7.1(b) shows the picture of the electron/ion burn on the ICR trapping electrode (i.e., filament trapping plate) positioned axially and opposite of the

RFI electrodes. Samples placed on the electron/ion burn spots could be desorbed and

128

subsequently ionized by the RFI-generated electrons. Further studies could verify the

desorption phenomenon in RFI by performing series of experiments that could involve a

combination of desorption using RFI with ionization of the RFI-desorbed species using

internal EI in FT-ICR MS.

a) Mass spectrum of mass spectrometer b) Electron/ion burn spots on the background hydrocarbon species filament trapping plate of the ICR cell

Electron/ion burn spots

50 100 150 200 250 Filament trapping plate m/z Figure 7.1 . (a) RFI FT-ICR mass spectrum of mass spectrometer background vacuum chamber hydrocarbon species obtained at RF signal of 12.2 MHz (400 Vp-p). (b) Picture of electron/ion burn spots on the filament trapping plate of the ICR cell obtained over ~6 months of RFI use.

CHAPTER FIVE: In Chapter Five, potential parameters (including threshold RF

voltage amplitudes and frequencies) that could contribute to ion yield and signal stability

in RFI were discussed. It was shown that analytes with different ionization energies and

polarizabilities are amenable to ionization by RFI, suggesting that RFI is a universal

ionization technique. It was also demonstrated that RF signals at higher frequencies (e.g.,

11.6 MHz) produce a more stable RFI signal than RF signals at lower frequencies (e.g.,

5.3 MHz). For example, an RFI signal at 10.5 MHz has a wider range of accessible RF

voltage amplitudes than an RF signal at 5.8 MHz. For instance, the total ion intensity for

acetone obtained at 11.6 MHz (360 Vp-p) has a better signal reproducibility (with signal

129 relative standard deviation (RSD) of ~15 (±8) % at the 95 % confidence level (CL) for three experimental trials) than at 5.3 MHz (360 Vp-p) (with signal RSD of ~155 (±55) % at the 95 % CL, n=3). Experimental findings presented in Chapter Five suggest that RFI sources would operate more efficiently and reproducibly at lower RF voltages and at higher frequencies.

Future studies might investigate the performance of an RFI source (at lower magnetic field (e.g., <1000 gauss) for non-magnet based mass spectrometers. The RF parameters, such as RF voltage amplitudes and frequencies and signal stability can be verified for RFI sources designed for table-top mass spectrometers. Preliminary data on ion trajectory simulations (shown in Figures 7.2 and 7.3) suggest that efficient RFI sources could be designed and developed for VOC analysis using magnetic field at 500 gauss. Figure 7.2 shows ion trajectories for ions generated within a RFI source placed within a 500 gauss magnetic field. The RFI-generated ions (from 28 Th to 502 Th) were efficiently extracted and focused out of the ion source. A similar RFI source could be designed and experimentally optimized for analyzing VOCs using regular table-top mass spectrometers. The performance characteristics of the designed RFI source could also be studied in the mTorr pressure regime of most conventional EI/CI sources.

Lens 2: Focusing lens Figure 7.2. (a) SIMION captured picture for simulated ion trajectories. Magnetic field, 500 gauss; Repeller, -20 V; RFI electrodes’ DC offset, 10 V; Lens 1, -160 V; Lens 2, - 20 V. 130

Figure 7.3(a) shows the simulated ion transmission efficiency for m/z’s 28, 69,

131, 219, and 502 (representing the molecular and fragment ions of perfluorotributylamine, a commonly used calibrant in most table-top mass spectrometers) against frequency. Ion transmission efficiency for the m/z’s 28 to 502 was greater (>50

%) at higher frequencies than at lower frequencies. In Figure 7.3(b) it is shown that greater (~60 %) ion transmission efficiency (at 12.5 MHz) was achieved at lower RF voltage amplitude (Vp-p) than at higher RF voltage amplitude for m/z’s 28 to 502.

a) Ion Transmission versus Frequency 60

45

30 m/z 502 m/z 219 15 m/z 131 Efficiency (%) Efficiency m/z 69 m/z 28 IonTransmission 0 0 5 10 15 20 Frequency (MHz)

b) Ion Transmission versus RF Amplitude

60

45

30

Efficiency (%) Efficiency

IonTransmission 15

0 100 200 300 400 500 RF Amplitude (Vp-p)

Figure 7.3. Plots showing simulated ion transmission efficiency as a function of (a) frequency (from 1 to 20 MHz at 200 Vp-p) and (b) RF amplitude/voltage (from 0 to 500 Vp-p at 12.5 MHz) for m/z of 28, 69, 131, 219, and 502.

131

CHAPTER SIX: Results presented in Chapter Six demonstrated the involvement of electrons in the ionization process of RFI. Both direct and indirect methods of electron detection were used to explore the possibility of electron emission in RFI. Direct current measurements including electron and ion currents (from ionized acetone molecules and mass spectrometer background hydrocarbon molecules) were detected on the quadrupole trapping plate of the ICR cell upon application of RF signal to RFI electrodes. Similarly, indirect electron detection experiments involving the trapping of RF-generated electrons in the ICR cell of a 9.4 tesla (T) FT-ICR mass spectrometer and subsequent reaction with neutral hexafluorobenzene (C6F6) molecules were described. The rate of electron attachment to C6F6 was determined and compared to the literature reported value. In

●- electron quenching experiments, correlation between the intensity of C6F6 species

(formed from the interaction of C6F6 molecules with the trapped electrons) and electron quenching time was demonstrated and rate constant for electron capture was calculated.

Interpretation of the data shown in Chapter Six suggests that electrons play an important role in the mechanism of ion generation in RFI.

Future studies will focus on better understanding of the ion generation process in

RFI, especially the process of electron emission (via electric field) from the surface of

RFI electrodes. Different RFI electrodes with differing work functions could be utilized to determine the effects of metal work function on RFI ion yield. Based on the Fowler-

Nordheim theory presented in chapter two of this dissertation, a lower metal function should yield higher current density. Similarly, the geometry of RFI electrodes in terms of shape (planar, round, or pointed) and surface properties (rough or smooth) could be investigated for field enhancement in RFI. Selecting appropriate and optimized RFI

132 electrodes that maximize electron emission in RFI will enhance the design/construction of RFI sources with higher ion yields.

In Chapter Six, utilizing dipole RFI electrodes with reduced capacitance (i.e., ~40 pF) enabled higher frequencies to be accessible. An important future study could be the development of miniaturized RFI sources with micron-sized electrodes for coupling to portable mass spectrometers, which would expand MS utility of RFI. Preliminary results from the construction of an RFI source with smaller electrode sizes and reduced capacitance is illustrated in Figure 7.4.

a) AutoCad Drawing of RFI Source b) Constructed RFI Source

Focusing Focusing lens

electrodesRFI

Aluminum housing

RFI electrodes Extractor lens Teflon Extractor lens Repeller lens 1 cm Inlet for neutrals support Focusing Aluminum housing lens Teflon support RFI Electrodes’ Outer Diameter: 0.3 cm Inlet for neutrals RFI Electrodes’ Length: 2.0 cm 1 cm Total Capacitance: ~36 pF

Figure 7.4. (a) AutoCad drawing of RFI source showing the dipole electrodes and the associated ion lenses for improved ion trajectory control and transmission. (b) Picture (side view) of the constructed RFI source.

In Figure 7.4(a)), an AutoCad drawing of an RFI source consisting of two RFI electrodes (made from aluminum rod), ion lenses (repeller, extractor, and focusing), inlet for neutrals, aluminum housing, and teflon/ceramic supports is shown. The constructed

RFI source (shown in Figure 7.4(b)) comprising of RFI electrodes with length of 2.0 cm,

133 outer diameter of 0.3 cm, and capacitance of ~36 pF have been used to obtain reproducible and stable RFI signals when placed within a 9.4 T FT-ICR mass spectrometer.

134

APPENDIX

135

APPENDIX

Analysis of Volatile Organic Compound Mixtures Using Radio-Frequency Ionization/Mass Spectrometry

(a) CS Aqueous Phase + C3H7O

+ C4H9O

50 100 150

m/z (b) CS Oily Phase + C5H7O

+ C6H9O

50 100 150 m/z Figure A.1. RFI/FT-ICR mass spectra of VOCs from (a) aqueous and (b) oily phases of corn stover (CS) slow pyrolysis bio-oil. The two most abundant peaks in each mass spectrum are labeled. For each mass spectrum, 5 FT-ICR MS time-domain transients of 128 k data points were summed prior to Fourier transformation.

136

Table A.1. List of the observed and theoretical (exact) m/z values, the mass measurement errors, assigned chemical compositions, and the double bond equivalents (DBEs) for the ionic species observed in the RFI/FT-ICR mass spectra of pine shavings (PS) and corn stover (CS) bio-oil aqueous and oily phases as well as a commercial gasoline sample (RON = 87).

Error Chemical PS PS CS CS Obs. m/z Exact m/z DBE* Gasoline (ppm) Composition (Aqueous) (Oily) (Aqueous) (Oily) + 29.03845 29.03858 -4.48 C2H5 0.5 x x x x + 31.01788 31.01784 1.40 CH3O 0.5 x x + 32.02585 32.02567 5.62 CH4O 0.0 x x x x + 33.03342 33.03349 -2.12 CH5O -0.5 x x x + 39.02291 39.02293 -0.56 C3H3 2.5 + 41.03858 41.03858 0.00 C3H5 1.5 x x x + 42.03388 42.03383 1.07 C2H4N 1.5 x x x + 42.04622 42.04640 -4.28 C3H6 1.0 x x x x + 43.01783 43.01784 -0.23 C2H3O 1.5 + 43.05420 43.05423 -0.70 C3H7 0.5 x x x x + 45.03356 45.03349 1.48 C2H5O 0.5 x x + 47.04911 47.04914 -0.57 C2H7O -0.5 x x + 53.03858 53.03858 0.00 C4H5 2.5 x + 55.01807 55.01784 4.18 C3H3O 2.5 x x x + 55.05438 55.05423 2.79 C4H7 1.5 x x + 56.04961 56.04948 2.32 C3H6N 1.5 x x x x + 56.06207 56.06205 0.36 C4H8 1.0 x x x x + 57.03347 57.03349 -0.35 C3H5O 1.5 x + 57.06953 57.06988 -6.13 C4H9 0.5 x x x x + 58.04103 58.04132 -4.99 C3H6O 1.0 x x x x + 59.04925 59.04914 1.86 C3H7O 0.5 + 60.05553 60.05562 -1.58 CH6N3 0.5 x x x + 61.02839 61.02841 -0.33 C2H5O2 0.5 x x x + 65.03835 65.03858 -3.54 C5H5 3.5 x x x + 67.05431 67.05423 1.12 C5H7 2.5 x + 69.03354 69.03349 0.65 C4H5O 2.5 x + 69.07000 69.06988 1.66 C5H9 1.5 x + 70.06476 70.06513 -5.28 C4H8N 1.5 x x x x + 70.07781 70.07770 1.57 C5H10 1.0 x x x x + 71.04921 71.04914 1.02 C4H7O 1.5 x + 71.08507 71.08553 -6.47 C5H11 0.5 x x x x + 72.05640 72.05697 -7.91 C4H8O 1.0 x x x x + 73.02817 73.02841 -3.29 C3H5O2 1.5 x x x x + 73.06480 73.06479 0.17 C4H9O 0.5 x + 75.04416 75.04406 1.37 C3H7O2 0.5 x + 76.05194 76.05188 0.79 C3H8O2 0.0 x x x x + 77.03954 77.03858 12.46 C6H5 4.5 x x x x + 78.04685 78.04640 5.70 C6H6 4.0 x x x + 79.05421 79.05423 -0.25 C6H7 3.5 x x x x + 80.05020 80.04948 8.99 C5H6N 3.5 x x x x + 81.03311 81.03349 -4.72 C5H5O 3.5 x + 81.07004 81.06988 2.01 C6H9 2.5 x x + 82.04153 82.04132 2.56 C5H6O 3.0 x x + 82.07691 82.07770 -9.63 C6H10 2.0 x x x x + 83.04922 83.04914 0.96 C5H7O 2.5 x + 83.08574 83.08553 2.57 C6H11 1.5 x x + 84.05565 84.05562 0.31 C3H6N3 2.5 x x + 84.09315 84.09335 -2.38 C6H12 1.0 x x x x + 85.06482 85.06479 0.38 C5H9O 1.5 x + 85.10117 85.10118 -0.12 C6H13 0.5 x x x x + 87.04378 87.04406 -3.27 C4H7O2 1.5 x x x + 87.08049 87.08044 0.52 C5H11O 0.5 x

137

+ 89.05975 89.05971 0.48 C4H9O2 0.5 x + 91.01752 91.01784 -3.52 C6H3O 5.5 x x x x + 91.05442 91.05423 2.11 C7H7 4.5 x + 92.06177 92.06205 -3.08 C7H8 4.0 x x + 93.06992 93.06988 0.43 C7H9 3.5 x x x + 94.07770 94.07770 0.00 C7H10 3.0 x x x x + 95.01248 95.01276 -2.95 C5H3O2 4.5 x x x x + 95.04933 95.04914 1.95 C6H7O 3.5 x x x + 95.08561 95.08553 0.88 C7H11 2.5 x x + 96.02062 96.02058 0.36 C5H4O2 4.0 x x x + 96.05745 96.05697 4.96 C6H8O 3.0 x x + 96.09339 96.09335 0.42 C7H12 2.0 x x x x + 97.02821 97.02841 -2.04 C5H5O2 3.5 x + 97.06481 97.06479 0.23 C6H9O 2.5 x + 97.10136 97.10118 1.82 C7H13 1.5 x x + 98.03574 98.03623 -4.99 C5H6O2 3.0 x x x + 98.07276 98.07262 1.43 C6H10O 2.0 x x x x + 98.10910 98.10900 1.02 C7H14 1.0 x x x x + 99.08055 99.08044 1.06 C6H11O 1.5 x + 99.11701 99.11683 1.82 C7H15 0.5 x x x x + 101.02337 101.02332 0.49 C4H5O3 2.5 x x x x + 101.02940 101.02937 0.29 C4H7NS 2.0 x x x x + 101.05989 101.05971 1.78 C5H9O2 1.5 x + 101.09603 101.09609 -0.59 C6H13O 0.5 x + 102.04530 102.04640 -10.78 C8H6 6.0 x x x x + 103.05416 103.05420 -0.68 C8H7 5.5 x x x x + 103.07570 103.07536 3.29 C5H11O2 0.5 x + 105.06473 105.06585 -10.66 C3H9N2O2 0.5 x x x x + 105.06952 105.06988 -3.43 C8H9 4.5 x x + 106.07799 106.07770 2.87 C8H10 4.0 x x + 107.08609 107.08553 5.23 C8H11 3.5 x x x + 109.06458 109.06479 -1.93 C7H9O 3.5 x x x + 109.10101 109.10118 -1.56 C8H13 2.5 x x + 110.03700 110.03623 6.99 C6H6O2 4.0 x x x x + 110.07303 110.07262 3.73 C7H10O 3.0 x x x + 110.10908 110.10900 0.73 C8H14 2.0 x x x x + 111.03171 111.03148 2.07 C5H5NO2 4.0 x x x x + 111.04376 111.04406 -2.70 C6H7O2 3.5 x x + 111.08029 111.08044 -1.35 C7H11O 2.5 x + 111.11730 111.11683 4.23 C8H15 1.5 x x x + 112.08590 112.08692 -9.10 C5H10N3 2.5 x x x x + 112.12526 112.12465 5.44 C8H16 1.0 x x x x + 113.09500 113.09475 2.21 C5H11N3 2.0 x x x x + 113.09606 113.09609 0.27 C7H13O 1.5 x x + 113.13210 113.13248 -3.36 C8H17 0.5 x x x x + 115.07575 115.07536 3.39 C6H11O2 1.5 x + 117.05500 117.05462 3.25 C5H9O3 1.5 x x x + 117.06976 117.06988 -1.03 C9H9 5.5 x x x x + 117.09071 117.09101 -2.56 C6H13O2 0.5 x x x + 119.08556 119.08553 0.25 C9H11 4.5 x x + 120.09302 120.09335 -2.75 C9H12 4.0 x x x + 121.10200 121.10118 6.77 C9H13 3.5 x x + 123.07989 123.08044 -4.47 C8H11O 3.5 x x x + 123.11756 123.11683 5.93 C9H15 2.5 x x x x + 124.12455 124.12465 -0.81 C9H16 2.0 x x x x + 125.05876 125.05836 3.20 C5H7N3O 4.0 x x x x + 125.09587 125.09609 -1.76 C8H13O 2.5 x + 125.13245 125.13248 -0.24 C9H17 1.5 x x x x + 128.06166 128.06205 -3.05 C10H8 7.0 x + 129.07010 129.06988 1.70 C10H9 6.5 x x x + 129.08906 129.08966 -4.65 C5H11N3O 2.0 x x x x + 130.07504 130.07368 10.46 C5H10N2O2 2.0 x x x x 138

+ 131.07266 131.07295 -2.21 C9H9N 6.0 x x x x + 131.08543 131.08553 -0.76 C10H11 5.5 x x x x + 132.04204 132.04171 2.49 C5H8O4 2.0 x x x x + 133.04911 133.04954 -3.23 C5H9O4 1.5 x x x x + 133.10139 133.10118 1.58 C10H13 4.5 x x x x + 134.10900 134.10900 0.00 C10H14 4.0 x x x x + 135.11672 135.11683 -0.81 C10H15 3.5 x x x x + 137.13272 137.13248 1.75 C10H17 2.5 x x x + 138.10314 138.10392 -5.65 C9H14O 3.0 x x x x + 139.11227 139.11174 3.81 C9H15O 2.5 x x x x + 143.05134 143.05117 1.19 C5H9N3S 3.0 x x x x + 155.10667 155.10666 0.06 C9H15O2 2.5 x x x x + 163.14786 163.14813 -1.65 C12H19 3.5 x x x x + 171.13787 171.13796 -0.53 C10H19O2 1.5 x x x

* Double Bond Equivalent (DBE) = (Number of Carbon Atoms) - (Number of Hydrogen Atoms / 2) + (Number of Nitrogen Atoms / 2) + 1. : Present x: Not Present

139

REFERENCES

Chapters One, Two, Three, and Five

1. Cisper, M. E.; Gill, C. G.; Townsend, L. E.; Hemberger, P. H., On-line detection of volatile organic compounds in air at parts-per-trillion levels by membrane introduction mass spectrometry. Anal. Chem. 1995, 67, 1413-1417.

2. Ochiai, N.; Takino, M.; Daishima, S.; Cardin, D. B., Analysis of volatile sulphur compounds in breath by gas chromatography-mass spectrometry using a three- stage cryogenic trapping preconcentration system. J. Chromatogr. B 2001, 762, 67-75.

3. Alvarado, J. S.; Rose, C., Static headspace analysis of volatile organic compounds in soil and vegetation samples for site characterization. Talanta 2004, 62, 17-23.

4. Zoccolillo, L.; Amendola, L.; Cafaro, C.; Insogna, S., Improved analysis of volatile halogenated hydrocarbons in water by purge-and-trap with gas chromatography and mass spectrometric detection. J. Chromatogr. A 2005, 1077, 181-187.

5. Domon, B.; Aebersold, R., Mass spectrometry and protein analysis. Science 2006, 312, 212-217.

6. Schiller, J.; Sub, R.; Arnhold, J.; Fuchs, B.; LeBig, J.; Muller, M.; Petkovic, M.; Spalteholz, H.; Zschornig, O.; Arnold, K., Matrix-assisted laser desorption and ionization time-of-flight (MALDI-TOF) mass spectrometry in lipid and phospholipid research. Prog. Lipid Res. 2004, 43, 449-488.

7. Harvey, D. J., Matrix-assisted laser desorption/ionization mass spectrometry of carbohydrates. Mass Spectrom. Rev. 1999, 18, 349-451.

8. Nordhoff, E.; Kirpekar, F.; Roepstorff, P., Mass spectrometry of nucleic acids. Mass Spectrom. Rev. 1996, 15, 67-138.

9. Saf, R.; Mirtl, C.; Hummel, K., Electrospray ionization mass spectrometry as an analytical tool for non-biological monomers, oligomers and polymers. Acta Polym. 1997, 48, 513-526.

10. Art, H. W., Volatile organic compounds. In Art, H.W., ed. 1993, A dictionary of ecology and environmental science: New York, Henry Holt and Company, p. 581.

140

11. Zhang, Z.-M.; Cai, J.-J.; Ruan, G.-H.; Li, G.-K., The study of fingerprint characteristics of the emanations from human arm skin using the original sampling system by SPME-GC/MS. J. Chromatogr. B 2005, 822, 244-252.

12. Buszewski, B.; Kesy, M.; Ligor, T.; Amann, A., Human exhaled air analytics: biomarkers of diseases. Biomed. Chromatogr. 2007, 21, 553-566.

13. Chik, Z.; Mustafa, A. M.; Mohamed, Z.; Lee, T. C., Analysis of captopril in human plasma using gas chromatography-mass spectrometry (GCMS) with solid- phase extraction (SPE). Curr. Anal. Chem. 2010, 6, 329-333.

14. Ulanowska, A.; Kowalkowski, T.; Hrynkiewicz, K.; Jackowski, M.; Buszewski, B., Determination of volatile organic compounds in human breath for Helicobacter pylori detection by SPME-GC/MS. Biomed. Chromatogr. 2011, 25, 391-397.

15. Richardson, S. D., Mass spectrometry in environmental sciences. Chem. Rev. 2001, 101, 211-254.

16. Imhof, R.; Bosset, J. O., Quantitative GC-MS analysis of volatile flavour compounds in pasteurized milk and fermented milk products applying a standard addition method. LWT-Food Sci. Technol. 1994, 27, 265-269.

17. Heikes, D. L.; Jensen, S. R.; Fleming-Jones, M. E., Purge and trap extraction with GC-MS determination of volatile organic compounds in table-ready foods. J. Agric. Food Chem. 1995, 43, 2869-2875.

18. Furton, K. G.; Wang, J.; Hsu, Y. L.; Walton, J.; Almirall, J. R., The use of solid- phase microextraction-gas chromatography in forensic analysis. J. Chromatogr. Sci. 2000, 38, 297-306.

19. Furton, K. G.; Hong, Y. C.; Hsu, Y. L.; Luo, T.; Rose, S.; Walton, J., Identification of odor signature chemicals in cocaine using solid-phase microextraction-gas chromatography and detector-dog response to isolated compounds spiked on U.S. paper currency. J. Chromatogr. Sci. 2002, 40, 147- 155.

20. Camarasu, C. C.; Mezei-Szuts, M.; Varga, G. B., Residual solvents determination in pharmaceutical products by GC-HS and GC-MS-SPME. J. Pharm. Biomed. Anal. 1998, 18, 623-638.

21. Camarasu, C. C., Headspace SPME method development for the analysis of volatile polar residual solvents by GC-MS. J. Pharm. Biomed. Anal. 2000, 23, 197-210.

141

22. Mendes, M. A.; Pimpim, R. S.; Kotiaho, T.; Eberlin, M. N., A cryotrap membrane introduction mass spectrometry system for analysis of volatile organic compounds in water at the low parts-per-trillion level. Anal. Chem. 1996, 68, 3502-3506.

23. Chen, T. C.; Her, G. R., On-line monitoring of trihalomethanes in drinking water using continuous-flow purge and cryofocusing gas chromatography-mass spectrometry. J. Chromatogr. A 2001, 927, 229-235.

24. Jochmann, M. A.; Blessing, M.; Haderlein, S. B.; Schmidt, T. C., A new approach to determine method detection limits for compound-specific isotope analysis of volatile organic compounds. Rapid Commun. Mass Spectrom. 2006, 20, 3639- 3648.

25. Daughtrey, J., H. E. ; Oliver, K. D.; Adams, J. R.; Kronmiller, K. G.; Lonneman, W. A.; McClenny, W. A., A comparison of sampling and analysis methods for low-ppbC levels of volatile organic compounds in ambient air. J. Environ. Monit. 2001, 3, 166-174.

26. Kim, Y.-H.; Kim, K.-H., Ultimate detectability of volatile organic compounds: how much further can we reduce their ambient air sample volumes for analysis? Anal. Chem. 2012, 84, 8284-8293.

27. Glish, G. L.; Vachet, R. W., The basics of mass spectrometry in the twenty-first century. Nature 2003, 2, 140-150.

28. Kandiah, M.; Urban, P. L., Advances in ultrasensitive mass spectrometry of organic molecules. Chem. Soc. Rev. 2013, 42, 5299-5322.

29. Zavilopulo, A. N.; Mironets, E. A.; Agafonova, A. S., An upgraded ion source for a mass spectrometer. Instrum. Exp. Tech+. 2012, 55, 65-71.

30. Vitcher, S.; Charvy, C.; Dudragne, L.; Tabet, J. C., Characterization of an electron ionization source trap operating in the presence of a magnetic field through computer simulation. J. Am. Soc. Mass Spectrom. 2013, 24, 1130-1136.

31. Chen, Y.; Leach III, F. E.; Kaiser, N. K.; Dang, X.; Ibrahim, Y. M.; Norheim, R. V.; Anderson, G. A.; Smith, R. D.; Marshall, A. G., Improved ion optics for introduction of ions into a 9.4-T Fourier transform ion cyclotron resonance mass spectrometer. J. Mass Spectrom. 2015, 50, 280-284.

32. Krutchinsky, A. N.; Padovan, J. C.; Cohen, H.; Chait, B. T., Maximizing ion transmission from atmospheric pressure into the vacuum of mass spectrometers with a novel electrospray interface. J. Am. Soc. Mass Spectrom. 2015, 26, 649- 658.

142

33. McLuckey, S. A.; Wells, J. M., Mass analysis at the advent of the 21st century. Chem. Rev. 2001, 101, 571-606.

34. Fu, J.; Purcell, J. M.; Quinn, J. P.; Schaub, T. M.; Hendrickson, C. L.; Rodgers, R. P.; Marshall, A. G., External electron ionization 7T Fourier transform ion cyclotron resonance mass spectrometer for resolution and identification of volatile organic mixtures. Rev. Sci. Instrum. 2006, 77, 025102(1)-025102(9).

35. Munson, B., Chemical ionization mass spectrometry. Anal. Chem. 1971, 43, 28A- 43A.

36. Kane, D. B.; Oktem, B.; Johnston, M. V., An electrostatic lens for focusing charged particles in a mass spectrometer. Aerosol Sci. Technol. 2001, 35, 990- 997.

37. Kotiaho, T.; Lauritsen, F. R.; Choudhury, T. K.; Cooks, R. G.; Tsao, G. T., Membrane introduction mass spectrometry. Anal. Chem. 1991, 63, 875A-883A.

38. Grote, C.; Pawliszyn, J., Solid-phase microextraction for the analysis of human breath. Anal. Chem. 1997, 69, 587-596.

39. Demeestere, K.; Dewulf, J.; Witte, B. D.; Langenhove, H. V., Sample preparation for the analysis of volatile organic compounds in air and water matrices. J. Chromatogr. A 2007, 1153, 130-144.

40. McClenny, W. A.; Pleil, J. D.; Holdren, M. W.; Smith, R. N., Automated cryogenic preconcentration and gas chromatographic determination of volatile organic compounds in air. Anal. Chem. 1984, 56, 2947-2951.

41. Bertman, S. B.; Buhr, M. P.; Roberts, J. M., Automated cryogenic trapping technique for capillary GC analysis of atmospheric trace compounds requiring no expendable cryogens: application to the measurement of organic nitrates. Anal. Chem. 1993, 65, 2944-2946.

42. Solouki, T.; Szulejko, J. E.; Bennett, J. B.; Graham, L. B., A preconcentrator coupled to a GC/FTMS: advantages of self-chemical ionization, mass measurement accuracy, and high mass resolving power for GC applications. J. Am. Soc. Mass Spectrom. 2004, 15, 1191-1200.

43. Fialkov, A. B.; Steiner, U.; Jones, L.; Amirav, A., A new type of GC-MS with advanced capabilities. Int. J. Mass Spectrom. 2006, 251, 47-58.

44. Kolb, B., Headspace sampling with capillary columns. J. Chromatogr. A 1999, 842, 163-205.

143

45. Snow, N. H.; Slack, G. C., Head-space analysis in modern gas chomatography. Trends Anal. Chem. 2002, 21, 608-617.

46. Hoch, G.; Kok, B., A mass spectrometer inlet system for sampling gases dissolved in liquid phases. Arch. Biochem. Biophys. 1963, 101, 160-170.

47. Kotiaho, T.; Hayward, M. J.; Cooks, R. G., Direct determination of chlorinated products of organic amines using membrane introduction mass spectrometry. Anal. Chem. 1991, 63, 1794-1801.

48. Allen, T. M.; Falconer, T. M.; Cisper, M. E.; Borgerding, A. J.; Wilkerson, J. C. W., Real-time analysis of methanol in air and water by membrane introduction mass spectrometry. Anal. Chem. 2001, 73, 4830-4835.

49. Klemp, M. A.; Akard, M. L.; Sacks, R. D., Cryofocusing inlet with reverse flow sample collection for gas chromatography. Anal. Chem. 1993, 65, 2516-2521.

50. Olaitan, A. D.; Zekavat, B.; LaBrecque, D.; Solouki, T., Coupling post-column cryogenic trapping with conventional GC/MS for the analysis of volatile organic compounds. Proceedings of the 61st ASMS Conference on Mass Spectrometry and Allied Topics 2013, Minneapolis, MN.

51. Marriott, P.; Kinghorn, R., Cryogenic solute manipulation in gas chromatography - the longitudinal modulation approach. Trends Anal. Chem. 1999, 18, 114-125.

52. Marriott, P. J.; Kinghorn, R. M., Modulation and manipulation of gas chromatographic bands by using novel thermal means. Anal. Sci. 1998, 14, 651- 659.

53. Phillips, J. B.; Beens, J., Comprehensive two-dimensional gas chromatography: a hyphenated method with strong coupling between the two dimensions. J. Chromatogr. A 1999, 856, 331-347.

54. Edwards, M.; Mostafa, A.; Gorecki, T., Modulation in comprehensive two- dimensional gas chromatography: 20 years of innovation. Anal. Bioanal. Chem. 2011, 401, 2335-2349.

55. Lanning, L. A.; Sacks, R. D.; Mouradian, R. F.; Levine, S. P.; Foulke, J. A., Electrically heated cold trap inlet system for computer-controlled high-speed gas chromatography. Anal. Chem. 1988, 60, 1994-1996.

56. Sive, B. C.; Zhou, Y.; Troop, D.; Wang, Y.; Little, W. C.; Wingenter, O. W.; Russo, R. S.; Varner, R. K.; Talbot, R., Development of a cryogen-free concentration system for measurements of volatile organic compounds. Anal. Chem. 2005, 77, 6989-6998.

144

57. Luo, Z.; Heffner, C.; Solouki, T., Multidimensional GC-Fourier transform ion cyclotron resonance MS analyses: utilizing gas-phase basicities to characterize multicomponent gasoline samples. J. Chromatogr. Sci. 2009, 47, 75-82.

58. Jacoby, C. B.; Remple, D. L.; Gross, M. L., A cold trap/pulsed valve GC/FTMS interface: ultra-trace analysis. Proceedings of the 38th ASMS Conference on Mass Spectrometry and Allied Topics 1990, Lincoln, NE.

59. Lu, W.; Zekavat, B.; Olaitan, A. D.; Brantley, M. R.; Erdogan, D. A.; Ozensoy, E.; Solouki, T., A cryofocuser/quadrupole mass spectrometer coupled to a catalysis unit for detection of nitrogen oxide catalysis products. Proceedings of the 63rd ASMS Conference on Mass Spectrometry and Allied Topics 2015, St. Louis, MO.

60. Harger, R. N.; Bridwell, E. G.; Raney, B. B., An aerometric method for the rapid determination of alcohol in water and body fluids. Proc. Am. Soc. Biol. Chem., J. Biol. Chem. 1939, 128, 38.

61. Bovijn, L.; Pirotte, J.; Berger, A., in D. H. Desty (editor). Gas Chromatography 1958 (Amsterdam Symposium) 1958, Butterworths, London, pp. 310-320.

62. Jones, A. W.; Holmgren, P., Uncertainty in estimating blood ethanol concentrations by analysis of vitreous humour. J. Clin. Pathol. 2001, 54, 699-702.

63. Sharp, M.-E. E.; Dautbegovic, T., Ether: Stability in preserved blood samples and a case of ether-assisted suicide. J. Anal. Toxicol. 2001, 25, 628-630.

64. Panosyan, A. G.; Mamikonyan, G. V.; Torosyan, M.; Gabrielyan, E. S.; Mkhitaryan, S. A.; Tirakyan, M. R.; Ovanesyan, A., Determination of the composition of volatiles in cognac (brandy) by headspace gas chromatography- mass spectrometry. J. Anal. Chem. 2001, 56, 945-952.

65. Zlatkis, A.; Lichtenstein, H. A.; Tishbee, A., Concentration and analysis of trace volatile organics in gases and biological fluids with a new solid adsorbent. Chromatographia 1973, 6, 67-70.

66. Szulejko, J. E.; Hall, S. K.; Jackson, M.; Solouki, T., Differentiation between pure cultures of streptococcus pyogenes and pseudomonas aeruginosa by FT-ICR-MS volatile analysis. Open Spectrosc. J. 2009, 3, 9-13.

67. Arthur, C. L.; Pawliszyn, J., Solid phase microextraction with thermal desorption using fused silica optical fibers. Anal. Chem. 1990, 62, 2145-2148.

68. Zhang, Z.; Pawliszyn, J., Headspace solid-phase microextraction. Anal. Chem. 1993, 65, 1843-1852.

145

69. Djozan, D.; Baheri, T.; Farshbaf, R.; Azhari, S., Investigation of solid-phase microextraction efficiency using pencil lead fiber for in vitro and in vivo sampling of defensive volatiles from insect's scent gland. Anal. Chim. Acta 2005, 554, 197- 201.

70. Ghiasvand, A. R.; Hosseinzadeh, S.; Pawliszyn, J., New cold-fiber headspace solid-phase microextraction device for quantitative extraction of polycyclic aromatic hydrocarbons in sediment. J. Chromatogr. A 2006, 1124, 35-42.

71. Heffner, C.; Silwal, I.; Peckenham, J. M.; Solouki, T., Emerging technologies for identification of disinfection byproducts: GC/FT-ICR MS characterization of solvent artifacts. Environ. Sci. Technol. 2007, 41, 5419-5425.

72. Gorecki, T.; Pawliszyn, J., Field-portable solid-phase microextraction/fast GC system for trace analysis. Field Anal. Chem. Technol. 1997, 1, 277-284.

73. Furton, K. G.; Wu, L.; Almirall, J. R., Optimization of solid-phase microextraction (SPME) for the recovery of explosives from aqueous and post- explosion debris followed by gas and liquid chromatographic analysis. J. Forensic Sci. 2000, 45, 857-864.

74. Martin, A. J. P.; Synge, R. L. M., A new form of chromatogram employing two liquid phases 1. A theory of chromatography 2. Application to the micro- determination of the higher monoamino-acids in proteins. Biochem. J. 1941, 35, 1358-1368.

75. James, A. T.; Martin, A. J. P., Gas-liquid partition chromatography: the separation and micro-estimation of volatile fatty acids from formic acid to dodecanoic acid. Biochem. J. 1952, 50, 679-690.

76. Ray, N. H., Gas chromatography. I. The separation and estimation of volatile organic compounds by gas-liquid partition chromatography. J. Appl. Chem. 1954, 4, 21-25.

77. Harley, J.; Nel, W.; Pretorius, V., Flame ionization detector for gas chromatography. Nature 1958, 181, 177-178.

78. Bartle, K. D.; Myers, P., History of gas chromatography. Trends Anal. Chem. 2002, 21, 547-557.

79. Mendis, S.; Sobotka, P. A.; Euler, D. E., Pentane and isoprene in expired air from humans: gas-chromatographic analysis of single breath. Clin. Chem. 1994, 40, 1485-1488.

146

80. Szulejko, J. E.; Solouki, T., Potential analytical applications of interfacing a GC to an FT-ICR MS: fingerprinting complex sample matrixes. Anal. Chem. 2002, 74, 3434-3442.

81. Pavon, J. L. P.; Pena, A. G.; Pinto, C. G.; Cordero, B. M., Differentiation of types of crude oils in polluted soil samples by headspace-fast gas chromatography-mass spectrometry. J. Chromatogr. A 2006, 1137, 101-109.

82. Lovelock, J. E.; Lipsky, S. R., Electron affinity spectroscopy-a new method for the identification of functional groups in chemical compounds separated by gas chromatography. J. Amer. Chem. Soc. 1960, 82, 431-433.

83. Gohlke, R. S., Time-of-flight mass spectrometry and gas-liquid partition chromatography. Anal. Chem. 1959, 31, 535-541.

84. Dorman, F. L.; Whiting, J. J.; Cochran, J. W.; Gardea-Torresday, J., Gas chromatography. Anal. Chem. 2010, 82, 4775-4785.

85. Miller, J. M., Chromatography concepts and contrasts. John Wiley & Sons, Inc.: Hoboken, New Jersey, 2005; p 35-182.

86. van Deemter, J. J.; Zuiderweg, F. J.; Klinkenberg, A., Longitudinal diffusion and resistance to mass transfer as causes of nonideality in chromatography. Chem. Eng. Sci. 1956, 5, 271-289.

87. James, A. T.; Martin, A. J. P.; Smith, G. H., Gas-liquid partition chromatography: the separation and micro-estimation of ammonia and the methylamines. Biochem. 1952, 52, 238-242.

88. Golay, M. J. E., Theory of chromatography in open and coated tubular columns with round and rectangular cross-sections, in: Desty D. H. (editor). Gas Chromatography 1958 (Amsterdam Symposium) 1958, Butterworths, London, pp. 139-143.

89. Beens, J.; Blomberg, J.; Schoenmakers, P. J., Proper tuning of comprehensive two-dimensional gas chromatography (GC x GC) to optimize the separation of complex oil fractions. J. High Resol. Chromatogr. 2000, 23, 182-188.

90. Meinert, C.; Meierhenrich, U. J., A new dimension in separation science: comprehensive two-dimensional gas chromatography. Angew. Chem. Int. Ed. 2012, 51, 10460-10470.

91. Nier, A. O., A mass spectrometer for routine isotope abundance measurements. Rev. Sci. Instrum. 1940, 11, 212-216.

147

92. Gierlich, H. H.; Heindrichs, A.; Beckey, H. D., A combined electron impact, chemical ionization, field ionization and field desorption mass spectrometer source. Rev. Sci. Instrum. 1974, 45, 1208-1211.

93. Lattimer, R. P.; Schulten, H.-R., Field ionization and field desorption mass spectrometry: past, present, and future. Anal. Chem. 1989, 61, 1201A-1215A.

94. Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S. F.; Whitehouse, C. M., Electrospray ionization for mass spectrometry of large biomolecules. Science 1989, 246, 64-71.

95. Hendrickson, C. L.; Emmett, M. R., Electrospray ionization Fourier transform ion cyclotron resonance mass spectrometry. Annu. Rev. Phys. Chem. 1999, 50, 517- 536.

96. Horning, E. C.; Horning, M. G.; Carroll, D. I.; Dzidic, I.; Stillwell, R. N., New picogram detection system based on a mass spectrometer with an external ionization source at atmospheric pressure. Anal. Chem. 1973, 45, 936-943.

97. Herrera, L. C.; Grossert, J. S.; Ramaley, L., Quantitative aspects of and ionization mechanisms in positive-ion atmospheric pressure chemical ionization mass spectrometry. J. Am. Soc. Mass Spectrom. 2008, 19, 1926-1941.

98. Hillenkamp, F.; Karas, M.; Beavis, R. C.; Chait, B. T., Matrix-assisted laser desorption/ionization mass spectrometry of biopolymers. Anal. Chem. 1991, 63, 1193A-1203A.

99. Solouki, T.; Marto, J. A.; White, F. M.; Guan, S.; Marshall, A. G., Attomole biomolecule mass analysis by matrix-assisted laser desorption/ionization Fourier transform ion cyclotron resonance. Anal. Chem. 1995, 67, 4139-4144.

100. Beckey, H. D., Field desorption mass spectrometry: a technique for the study of thermally unstable substances of low volatility. J. Mass Spectrom. Ion Phys. 1969, 2, 495-500.

101. Komori, T.; Kawasaki, T.; Schulten, H.-R., Field desorption and fast atom bombardment mass spectrometry of biologically active natural oligoglycosides. Mass Spectrom. Rev. 1985, 4, 255-293.

102. Miller, P. E.; Denton, M. B., The quadrupole mass filter: basic operating concepts. J. Chem. Educ. 1986, 63, 617-623.

103. Cody, R. B.; Tamura, J.; Musselman, B. D., Electrospray ionization/magnetic sector mass spectrometry: calibration, resolution, and accurate mass measurements. Anal. Chem. 1992, 64, 1561-1570.

148

104. Jakubowski, N.; Moens, L.; Vanhaecke, F., Sector field mass spectrometers in ICP-MS. Spectrochim. Acta, Part B 1998, 53, 1739-1763.

105. Wolff, M. M.; Stephens, W. E., A pulsed mass spectrometer with time dispersion. Rev. Sci. Instrum. 1953, 24, 616-617.

106. Guilhaus, M., Principles and instrumentation in time-of-flight mass spectrometry. J. Mass Spectrom. 1995, 30, 1519-1532.

107. March, R. E., An introduction to quadrupole mass spectrometry. J. Mass Spectrom. 1997, 32, 351-369.

108. Perry, R. H.; Cooks, R. G.; Noll, R. J., mass spectrometry: instrumentation, ion motion and applications. Mass Spectrom. Rev. 2008, 27, 661- 699.

109. Marshall, A. G.; Hendrickson, C. L.; Jackson, G. S., Fourier transform ion cyclotron resonance mass spectrometry: a primer. Mass Spectrom. Rev. 1998, 17, 1-35.

110. Dietz, L. A., Basic properties of electron multiplier ion detection and pulse counting methods in mass spectrometry. Rev. Sci. Instrum. 1965, 36, 1763-1770.

111. Liu, R.; Li, Q.; Smith, L. M., Detection of large ions in time-of-flight mass spectrometry: effects of ion mass and acceleration voltage on microchannel plate detector response. J. Am. Soc. Mass Spectrom. 2014, 25, 1374-1383.

112. Scheidemann, A. A.; Darling, R. B.; Schumacher, F. J.; Isakharov, A., detector array with electronic multiplexing for multichannel mass spectrometer. J. Vac. Sci. Technol., A 2002, 20, 597-604.

113. Watson, J. T.; Sparkman, O. D., Introduction to mass spectrometry: instrumentation, applications and strategies for data interpretation. John Wiley & Sons Ltd.: 2007; Vol. Chapter Two, p 147-162.

114. Xian, F.; Hendrickson, C. L.; Marshall, A. G., High resolution mass spectrometry. Anal. Chem. 2012, 84, 708-719.

115. Scigelova, M.; Hornshaw, M.; Giannakopulos, A.; Makarov, A., Fourier transform mass spectrometry. Mol. Cell. Proteomics 2011, 10, 1-19.

116. Barrow, M. P.; Burkitt, W. I.; Derrick, P. J., Principles of Fourier transform ion cyclotron resonance mass spectrometry and its application in structural biology. Analyst 2005, 130, 18-28.

149

117. Dawson, P. H., Quadrupole mass analyzers: performance, design and some recent applications. Mass Spectrom. Rev. 1986, 5, 1-37.

118. Busch, K. L., High-vacuum pumps in mass spectrometers. Spectroscopy 2001, 16, 14-18.

119. Sturm, M.; Bertsch, A.; Gropl, C.; Hildebrandt, A.; Hussong, R.; Lange, E.; Pfeifer, N.; Schulz-Trieglaff, O.; Zerck, A.; Reinert, K.; Kohlbacher, O., OpenMS - An open-source software framework for mass spectrometry. BMC Bioinformatics 2008, 9, 1-11.

120. Paul, W.; Steinwedel, H., Ein neues massenspektrometer ohne magnetfeld. Z. Naturforsch. 1953, A8, 448-450.

121. Fiehn, O.; Kopka, J.; Trethewey, R. N.; Willmitzer, L., Identification of uncommon plant metabolites based on calculation of elemental compositions using gas chromatography and quadrupole mass spectrometry. Anal. Chem. 2000, 72, 3573-3580.

122. Adahchour, M.; Brandt, M.; Baier, H.-U.; Vreuls, R., J. J.; Batenburg, A. M.; Brinkman, U. A. T., Comprehensive two-dimensional gas chromatography coupled to a rapid-scanning quadrupole mass spectrometer: principles and applications. J. Chromatogr. A 2005, 1067, 245-254.

123. Wang, C.; Xie, S.; Yang, J.; Yang, Q.; Xu, G., Structural identification of human blood phospholipids using liquid chromatography/quadrupole-linear ion trap mass spectrometry. Anal. Chim. Acta 2004, 525, 1-10.

124. Diaz-Cruz, M. S.; Garcia-Galan, M. J.; Barcelo, D., Highly sensitive simultaneous determination of sulfonamide antibiotics and one metabolite in environmental waters by liquid chromatography-quadrupole linear ion trap-mass spectrometry. J. Chromatogr. A 2008, 1193, 50-59.

125. de Hoffmann, E., : a primer. J. Mass Spectrom. 1996, 31, 129-137.

126. Batey, J. H., The physics and technology of quadrupole mass spectrometers. Vacuum 2014, 101, 410-415.

127. Paul, W., Electromagnetic traps for charged and neutral particles. Rev. Mod. Phys. 1990, 62, 531-540.

128. Hipple, J. A.; Sommer, H.; Thomas, H. A., A precise method of determining the Faraday by magnetic resonance. Phys. Rev. 1949, 76, 1877-1878.

150

129. Sommer, H.; Thomas, H. A., Detection of magnetic resonance by ion resonance absorption. Phys. Rev. 1950, 78, 806.

130. Wobschall, D., Ion cyclotron resonance spectrometer. Rev. Sci. Instrum. 1965, 36, 466-475.

131. Anders, L. R.; Beauchamp, J. L.; Dunbar, R. C.; Baldeschwieler, J. D., Ion- cyclotron double resonance. J. Chem. Phys. 1966, 45, 1062-1063.

132. McIver Jr., R. T., A trapped ion analyzer cell for ion cyclotron resonance spectroscopy. Rev. Sci. Instrum. 1970, 41, 555-558.

133. Comisarow, M. B.; Marshall, A. G., Fourier transform ion cyclotron resonance spectroscopy. Chem. Phys. Lett. 1974, 25, 282-283.

134. Marshall, A. G.; Hendrickson, C. L., Fourier transform ion cyclotron resonance detection: principles and experimental configurations. Int. J. Mass Spectrom. 2002, 215, 59-75.

135. Gross, M. L.; Rempel, D. L., Fourier transform mass spectrometry. Science 1984, 226, 261-268.

136. Amster, I. J., Fourier transform mass spectrometry. J. Mass Spectrom. 1996, 31, 1325-1337.

137. Marshall, A. G.; Schweikhard, L., Fourier transform ion cyclotron resonance mass spectrometry: technique developments. Int. J. Mass Spectrom. Ion. Processes 1992, 118/119, 37-70.

138. Beu, S. C.; Laude, J. D. A., Elimination of axial ejection during excitation with a capacitively coupled open trapped-ion cell for Fourier transform ion cyclotron resonance mass spectrometry. Anal. Chem. 1992, 64, 177-180.

139. Solouki, T.; Russell, D. H., Laser desorption studies of high mass biomolecules in Fourier-transform ion cyclotron resonance mass spectrometry. Proc. Natl. Acad. Sci. USA 1992, 89, 5701-5704.

140. Marshall, A. G.; Grosshans, P. B., Fourier transform ion cyclotron resonance mass spectrometry: the teenage years. Anal. Chem. 1991, 63, 215A-229A.

141. Comisarow, M. B.; Marshall, A. G., Frequency-sweep Fourier transform ion cyclotron resonance spectroscopy. Chem. Phys. Lett. 1974, 6, 489-490.

142. Marshall, A. G., Fourier transform ion cyclotron resonance mass spectrometry. Acc. Chem. Res. 1985, 18, 316-322.

151

143. Samokhin, A. S.; Revel'skii, A. I.; Revel'shii, I. A., Comparison of electron ionization mass spectra of some organic compounds (MW<200 Da) recorded on mass spectrometers of various types. J. Anal. Chem. 2011, 66, 1432-1436.

144. Munson, B., Chemical ionization mass spectrometry. Anal. Chem. 1977, 49, 772A-775A.

145. Reid, N. W., Photoionization in mass spectrometry. Int. J. Mass Spectrom. Ion Phys. 1971, 6, 1-31.

146. Arii, T.; Otake, S.; Takata, Y.; Matsuura, S., Evolved gas analysis using photoionization mass spectrometry. J. Mass Spectrom. Soc. Jpn. 2006, 54, 243- 249.

147. Hunt, D. F., New ionization techniques in mass spectrometry. Int. J. Mass Spectrom. Ion Phys. 1982, 45, 111-123.

148. Murray, K. K.; Boyd, R. K.; Eberlin, M. N.; Lanhley, G. J.; Li, L.; Naito, Y., Definitions of terms relating to mass spectrometry (IUPAC recommendations 2013)*. Pure Appl. Chem. 2013, 85, 1515-1609.

149. Vekey, K., Internal energy effects in mass spectrometry. J. Mass Spectrom. 1996, 31, 445-463.

150. Kind, T.; Fiehn, O., Advances in structure elucidation of small molecules using mass spectrometry. Bioanal. Rev. 2010, 2, 23-60.

151. Dempster, A. J., The properties of slow canal rays. Phys. Rev. 1916, 8, 651-662.

152. Dempster, A. J., A new method of positive ray analysis. Phys. Rev. 1918, 11, 316- 325.

153. Ma, Y.; Hites, R. A., Electron impact, electron capture negative ionization and positive chemical ionization mass spectra of organophosphorus flame retardants and plasticizers. J. Mass Spectrom. 2013, 48, 931-936.

154. Busch, K. L., Electron ionization sources: the basics. Spectroscopy 2006, 21, 14- 18.

155. Otvos, J. W.; Stevenson, D. P., Cross-sections of molecules for ionization by electrons. J. Am. Chem. Soc. 1956, 78, 546-551.

156. Harrison, A. G.; Jones, E. G.; Gupta, S. K.; Nagy, G. P., Total cross sections for ionization by electron impact. Can. J. Chem. 1966, 44, 1967-1973.

152

157. Crawford, C. K., Mass spectrometer ion source for ionization cross-section measurement. J. Vac. Sci. Technol. 1968, 5, 131-140.

158. Munson, M. S. B.; Field, F. H., Chemical ionization mass spectrometry. J. Am. Chem. Soc. 1966, 88, 2621-2630.

159. Munson, B., Development of chemical ionization mass spectrometry. Int. J. Mass Spectrom. 2000, 200, 243-251.

160. Dougherty, R. C., Negative chemical ionization mass spectrometry. Anal. Chem. 1981, 53, 625A-636A.

161. Stemmler, E. A.; Yoshida, E.; Pacheco, J.; Brunton, J.; Woodbury, E.; Solouki, T., Reactions of O*- with methyl benzoate: a negative ion chemical ionization and Fourier transform ion cyclotron resonance study. J. Am. Soc. Mass Spectrom. 2001, 12, 694-706.

162. Inghram, M. G.; Gomer, R., Mass spectrometric analysis of ions from the field microscope. J. Chem. Phys. 1954, 22, 1279-1280.

163. Anbar, M.; Aberth, W. H., Field ionization mass spectrometry: a new tool for the analytical chemist. Anal. Chem. 1974, 46, 59A-64A.

164. Mueller, E. W., Feldemission. Ergeb. Exakt. Naturwiss. 1953, 27, 290-360.

165. Inghram, M. G.; Gomer, R., Mass spectrometric investigations of the field emission of positive ions. Z. Naturforsch. 1955, 10A, 863-872.

166. Beckey, H. D., Experimental techniques in field ionisation and field desorption mass spectrometry. J. Phys. E: Sci. Instrum. 1979, 12, 72-83.

167. Tsong, T. T.; Muller, E. W., Measurement of the energy distribution in field ionization. J. Chem. Phys. 1964, 41, 3279-3284.

168. Muller, E. W.; Bahadur, K., Field ionization of gases at a metal surface and the resolution of the field ion microscope. Phys. Rev. 1956, 102, 624-631.

169. Riley, D. J.; Mann, M.; MacLaren, D. A.; Dastoor, P. C.; Allison, W.; Teo, K. B. K.; Amaratunga, G. A. J.; Milne, W., Helium detection via field ionization from carbon nanotubes. Nano Lett. 2003, 3, 1455-1458.

170. Luo, J.; Mark, L. P.; Giannakopulos, A. E.; Colburn, A. W.; Macpherson, J. V.; Drewello, T.; Derrick, P. J.; Teh, A. S.; Teo, K. B. K.; Milne, W. I., Field ionization using densely spaced arrays of nickel-tipped carbon nanotubes. Chem. Phys. Lett. 2011, 505, 126-129.

153

171. Sadeghian, R. B.; Islam, M. S., Ultralow-voltage field-ionization discharge on whiskered silicon nanowires for gas-sensing applications. Nat. Mater. 2011, 10, 135-140.

172. Ditchburn, R. W.; Arnot, F. L., The ionization of potassium vapour. Proc. Roy. Soc. A 1929, 123, 516-536.

173. Terenin, A.; Popov, B., Photodissociation of salt molecules into ions. Phys. Z. Sowjetunion 1932, 2, 299-318.

174. Washida, N.; Akimoto, H.; Takagi, H.; Okuda, M., Gas chromatography/photoionization mass spectrometry. Anal. Chem. 1978, 50, 910- 915.

175. Syage, J. A., Mechanism of [M + H]+ formation in photoionization mass spectrometry. J. Am. Soc. Mass Spectrom. 2004, 15, 1521-1533.

176. Zekavat, B.; Solouki, T., Radio-frequency ionization of organic compounds for mass spectrometry analysis. Angew. Chem. Int. Ed. 2013, 52, 2426-2429.

177. Olaitan, A. D.; Zekavat, B.; Dhungana, B.; Hockaday, W. C.; Chambliss, C. K.; Solouki, T., Application of radio frequency ionization mass spectrometry for volatile organic compound analysis. 69th Southwest Regional Meeting of the American Chemical Society 2013, Waco, TX.

178. Solouki, T.; Zekavat, B.; Olaitan, A. D., A novel radio-frequency ionization method for the analysis of wide classes of volatile organic compounds. Proceedings of the 61st ASMS Conference on Mass Spectrometry and Allied Topics 2013, Minneapolis, MN.

179. Olaitan, A. D.; Zekavat, B.; Solouki, T., Ionization mechanism in radio-frequency ionization (RFI). Proceedings of the 62nd ASMS Conference on Mass Spectrometry and Allied Topics 2014, Baltimore, MD.

180. Zekavat, B.; Olaitan, A. D.; Rash, C.; Chambliss, C. K.; Solouki, T., Analysis of volatile organic compounds from gasoline samples using radio-frequency ionization/mass spectrometry. Proceedings of the 61st ASMS Conference on Mass Spectrometry and Allied Topics 2013, Minneapolis, MN.

181. Olaitan, A. D.; Zekavat, B.; Dhungana, B.; Hockaday, W. C.; Chambliss, C. K.; Solouki, T., Analysis of volatile organic compound mixtures using radio- frequency ionization /mass spectrometry. Anal. Methods 2014, 6, 4982-4987.

154

182. Olaitan, A. D.; Zekavat, B.; Brantley, M. R.; Solouki, T., Performance characteristics of a unique radio-frequency ionization source. Proceedings of the 63rd ASMS Conference on Mass Spectrometry and Allied Topics 2015, St. Louis, MO.

183. Olaitan, A. D.; Zekavat, B.; Solouki, T., Electron generation in radio-frequency ionization (RFI). Proceedings of the 2015 Pittsburgh Annual Conference 2015, New Orleans, LA.

184. Olaitan, A. D.; Zekavat, B.; Solouki, T., Evidence for electron-based ion generation in radio-frequency ionization. J. Mass Spectrom. 2015, DOI: 10.1002/jms.3703.

185. McAfee, C. D.; Da Silveira, E. F.; Schweikert, E. A., Spontaneous desorption: filed assisted ion induced desorption mass spectrometry. Int. J. Mass Spectrom. Ion Phys. 1990, 97, 311-324.

186. Townsend, J. S., The conductivity produced in gases by the motion of negatively- charged ions. Nature 1900, 62, 340-341.

187. Sanders, F. H., The value of the Townsend coefficient for ionization by collision at large plate distances and near atmospheric pressure. Phys. Rev. 1932, 41, 667- 677.

188. Scheibel, R. C.; Tanner, O. P.; Wood, K. V., Townsend discharge ionization probe for a mass spectrometer. Anal. Chem. 1981, 53, 550-552.

189. Duckworth, D. C.; Marcus, R. K., Radio frequency powered glow discharge atomization/ionization source for solids mass spectrometry. Anal. Chem. 1989, 61, 1879-1886.

190. Marcus, R. K., Operation principles and design considerations for radiofrequency powered glow discharge devices. J. Anal. At. Spectrom. 1993, 8, 935-943.

191. Winchester, M. R.; Payling, R., Radio-frequency glow discharge spectrometry: a critical review. Spectrochim. Acta, Part B 2004, 59, 607-666.

192. Bonin, B., Field emission and surface conditioning. Vacuum 1995, 46, 907-912.

193. Ghosh, A.; Chakrabarty, S., Fowler-Nordheim electron cold emission formalism in presence of strong magnetic field. Mon. Not. R. Astron. Soc. 2011, 000, 1-9.

194. Millikan, R. A.; Eyring, C. F., Laws governing the pulling of electrons out of metals by intense electric fields. Phys. Rev. 1926, 27, 51-67.

155

195. Skriver, H. L.; Rosengaard, N. M., Surface energy and work function of elemental metals. Phys. Rev. B 1992, 46, 7157-7168.

196. Murphy, E. L.; Good, J. R. H., Thermionic emission, field emission, and the transition region. Phys. Rev. 1956, 102, 1464-1473.

197. Braun, J.; Donath, M., Theory of photoemission from surfaces. J. Phys.: Condens. Matter 2004, 16, S2539-S2556.

198. Richardson, O. W., The emission of electrons from Tungsten at high temperatures: an experimental proof that the electric current in metals is carried by electrons. Science 1913, 38, 57-61.

199. Oppenheimer, J. R., On the quantum theory of the autoelectric field currents. Proc. Natl. Acad. Sci. USA 1928, 14, 363-365.

200. Dushman, S., Thermionic emission. Rev. Mod. Phys. 1930, 2, 381-476.

201. Crowell, C. R., The Richardson constant for thermionic emission in Schottky barrier diodes. Solid-State Electron. 1965, 8, 395-399.

202. Fowler, R. H.; Nordheim, L., Electron emission in intense electric fields. Proc. R. Soc. London. Series A 1928, 119, 173-181.

203. Lye, R. G.; Dekker, A. J., Theory of secondary emission. Phys. Rev. 1957, 107, 977-981.

204. Scholtz, J. J.; Dijkkamp, D.; Schmitz, R. W. A., Secondary electron emission properties. Philips. J. Res. 1996, 50, 375-389.

205. Fan, H. Y., Theory of photoelectric emission from metals. Phys. Rev. 1945, 68, 43-52.

206. Gergen, B.; Nienhaus, H.; Weinberg, W. H.; McFarland, E. W., Chemically induced electronic excitations at metal surfaces. Science 2001, 294, 2521-2523.

207. White, J. D.; Chen, J.; Matsiev, D.; Auerbach, D. J.; Wodtke, A. M., Conversion of large-amplitude vibration to electron excitation at a metal surface. Nature 2005, 433, 503-505.

208. LaRue, J. L.; Schafer, T.; Matsiev, D.; Velarde, L.; Nahler, N. H.; Auerbach, D. J.; Wodtke, A. M., Electron kinetic energies from vibrationally promoted surface exoemission: evidence for a vibrational autodetachment mechanism. J. Phys. Chem. A 2011, 115, 14306-14314.

156

209. Larue, J.; Schafer, T.; Matsiev, D.; Velarde, L.; Nahler, N. H.; Auerbach, D. J.; Wodtke, A. M., Vibrationally promoted electron emission at a metal surface: electron kinetic energy distributions. Phys. Chem. Chem. Phys. 2011, 13, 97-99.

210. Tuck, R. A., The use of platinum metals in modern thermionic emitters. Platinum Metals Rev. 1982, 26, 167-173.

211. Jimenez, M.; Noer, R. J.; Jouve, G.; Jodet, J.; Bonin, B., Electron field emission from large-area cathodes: evidence for the projection model. J. Phys. D: Appl. Phys. 1994, 27, 1038-1045.

212. Stern, T. E.; Gossling, B. S.; Fowler, R. H., Further studies in the emission of electrons from cold metals. Proc. R. Soc. London. Series A 1929, 124, 699-723.

213. Podenok, S.; Sveningsson, M.; Hassen, K.; Campbell, E. E. B., Electric field enhancement factors around a metallic, end-capped cylinder. NANO 2006, 1, 87- 93.

214. Salow, H., Sekundare elektronen emission. Physik. Z. 1940, 41, 434-442.

215. Baroody, E. M., A theory of secondary electron emission from metals. Phys. Rev. 1950, 78, 780-787.

216. Bruining, H., Physics an application of secondary electron emission. 1954, London:Pergamon, Chapter 6.

217. Hall, E. H., Sommerfeld's electron-theory of metals. Proc. Natl. Acad. Sci. USA 1928, 14, 370-377.

218. Hertz, H., Ueber einen Einfluss des ultravioletten Lichtes auf die elektrische Entladung. Ann. Physik 1887, 31, 983-1000.

219. Lenard, P., Ueber die lichtelektrische Wirkung. Ann. Physik 1902, 8, 149-198.

220. Einstein, A., Uber einen Erzeugung und Verwandlung des Lichtes betreffenden heuristischen Gesichtspunkt. Ann. Physik 1905, 17, 132-148.

221. Alvarado, J. S.; Rose, C., Static headspace analysis of volatile organic compounds in soil and vegetation samples for site characterization. Talanta 2004, 62, 17-23.

222. Furton, K. G.; Wang, J.; Hsu, Y. L.; Walton, J.; Almirall, J. R., The use of solid- phase microextraction--gas chromatography in forensic analysis. J. Chromatogr. Sci. 2000, 38, 297-306.

157

223. Ulanowska, A.; Kowalkowski, T.; Hrynkiewicz, K.; Jackowski, M.; Buszewski, B., Determination of volatile organic compounds in human breath for Helicobacter pylori detection by SPME-GC/MS. Biomed. Chromatogr. 2011, 25, 391-397.

224. Watanabe, K.; Fujita, H.; Hasegawa, K.; Gonmori, K.; Suzuki, O., GC/MS with post-column switching for large volume injection of headspace samples: sensitive determination of volatile organic compounds in human whole blood and urine. Anal. Chem. 2011, 83, 1475-1479.

225. Hunter E. Daughtrey, J.; Oliver, K. D.; Adams, J. R.; Kronmiller, K. G.; Lonneman, W. A.; McClenny, W. A., A comparison of sampling and analysis methods for low-ppbC levels of volatile organic compounds in ambient air. J. Environ. Monit. 2001, 3, 166-174.

226. McClenny, W. A.; Pleil, J. D.; Holdren, M. W.; Smith, R. N., Automated cryogenic preconcentration and gas chromatographic determination of volatile organic compounds in air. Anal. Chem. 1984, 56, 2947-2951.

227. Phillips, M.; Gleeson, K.; Hughes, J. M.; Greenberg, J.; Cataneo, R. N.; Baker, L.; McVay, W. P., Volatile organic compounds in breath as markers of lung cancer: a cross-sectional study. Lancet 1999, 353, 1930-1933.

228. Mackintosh, S. A.; Perez-Fuentetaja, A.; Zimmerman, L. R.; Pacepavicius, G.; Clapsadl, M.; Alaee, M.; Aga, D. S., Analytical performance of a triple quadrupole mass spectrometer compared to a high resolution mass spectrometer for the analysis of polybrominated diphenyl ethers in fish. Anal. Chim. Acta 2012, 747, 67-75.

229. Kim, Y.-H.; Kim, K.-H., Ultimate detectability of volatile organic compounds: how much further can we reduce their ambient air sample volumes for analysis? Anal. Chem. 2012, 84, 8284-8293.

230. Solouki, T.; Marto, J. A.; White, F. M.; Guan, S.; Marshall, A. G., Attomole biomolecule mass analysis by matrix-assisted laser desorption/ionization Fourier transform ion cyclotron resonance. Anal. Chem. 1995, 67, 4139-4144.

231. Strobel, F. H.; Solouki, T.; White, M. A.; Russell, D. H., Detection of femtomole and sub-femtomole levels of peptides by tandem magnetic sector/reflectron time- of-flight mass spectrometry and matrix-assisted laser desorption ionization. J. Am. Soc. Mass Spectrom. 1991, 2, 91-94.

232. Fialkov, A. B.; Steiner, U.; Lehotay, S. J.; Amirav, A., Sensitivity and noise in GC-MS: Achieving low limits of detection for difficult analytes. Int. J. Mass spectrom. 2007, 260, 31-48.

158

233. Gorecki, T.; Pawliszyn, J., Field-portable solid-phase microextraction/fast GC system for trace analysis. Field Anal. Chem. Technol. 1997, 1, 277-284.

234. Phillips, M., Method for the collection and assay of volatile organic compounds in breath. Anal. Biochem. 1997, 247, 272-278.

235. Mieth, M.; Schubert, J. K.; Groger, T.; Sabel, B.; Kischkel, S.; Fuchs, P.; Hein, D.; Zimmermann, R.; Miekisch, W., Automated needle trap heart-cut GC/MS and needle trap comprehensive two-dimensional GC/TOF-MS for breath gas analysis in the clinical environment. Anal. Chem. 2010, 82, 2541-2551.

236. Sive, B. C.; Zhou, Y.; Troop, D.; Wang, Y.; Little, W. C.; Wingenter, O. W.; Russo, R. S.; Varner, R. K.; Talbot, R., Development of a cryogen-free concentration system for measurements of volatile organic compounds. Anal. Chem. 2005, 77, 6989-6998.

237. Ochiai, N.; Takino, M.; Daishima, S.; Cardin, D. B., Analysis of volatile sulphur compounds in breath by gas chromatography-mass spectrometry using a three- stage cryogenic trapping preconcentration system. J. Chromatogr. B 2001, 762, 67-75.

238. Campo, L.; Mercadante, R.; Rossella, F.; Fustinoni, S., Quantification of 13 priority polycyclic aromatic hydrocarbons in human urine by headspace solid- phase microextraction gas chromatography-isotope dilution mass spectrometry. Anal. Chim. Acta 2009, 631, 196-205.

239. Hara, K.; Kashiwagi, M.; Kageura, M.; Matsusue, A.; Kubo, S., Solid-phase microextraction for amphetamines in solid tissues: washing the homogenates with ethyl ether enables their measurement by GC-MS after heptafluorobutyryl derivatization. Forensic Toxicol. 2009, 27, 52-53.

240. Chik, Z.; Mustafa, A. M.; Mohamed, Z.; Lee, T. C., Analysis of captopril in human plasma using gas chromatography-mass spectrometry (GCMS) with solid- phase extraction (SPE). Curr. Anal. Chem. 2010, 6, 329-333.

241. Chen, T. C.; Her, G. R., On-line monitoring of trihalomethanes in drinking water using continuous-flow purge and cryofocusing gas chromatography-mass spectrometry. J. Chromatogr. A 2001, 927, 229-235.

242. de Geus, H.-J.; de Boer, J.; Brinkman, U. A. T., Development of a thermal desorption modulator for gas chromatography. J. Chromatogr. A 1997, 767, 137- 151.

159

243. Panic, O.; Gorecki, T.; McNeish, C.; Goldstein, A. L.; Williams, B. J.; Worton, D. R.; Hering, S. V.; Kreisberg, N. M., Development of a new consumable-free thermal modulator for comprehensive two-dimensional gas chromatography. J. Chromatogr. A 2011, 1218, 3070-3079.

244. Jacoby, C. B.; Remple, D. L.; Gross, M. L., A cold trap/pulsed valve GC/FTMS interface: ultra-trace analysis. Proceedings of the 38th ASMS Conference on Mass Spectrometry and Allied Topics 1990, Lincoln, NE.

245. Solouki, T.; Szulejko, J. E.; Luo, Z., Multidimensional GC Fourier transform ion cyclotron resonance mass spectrometry: resolving complex mixtures. Proceedings of the 52nd ASMS Conference on Mass Spectrometry and Allied Topics 2004, Nashville, TN.

246. Handbook of chemistry and physics. CRC Press, Inc.: Boca Raton, FL, 1983.

247. Bartmess, J. E.; Georgiadis, R. M., Empirical methods for determination of ionization gauge relative sensitivities for different gases. Vacuum 1983, 33, 149- 153.

248. Nakao, F., Determination of the ionization gauge sensitivity using the relative ionization cross-section. Vacuum 1975, 25, 431-435.

249. Szulejko, J. E.; McCulloch, M.; Jackson, J.; McKee, D. L.; Walker, J. C.; Solouki, T., Evidence for cancer biomarkers in exhaled breath. IEEE Sens. J. 2010, 10, 185-210.

250. Sanagi, M. M.; Ling, S. L.; Nasir, Z.; Hermawan, D.; Ibrahim, W. A.; Abu Naim, A., Comparison of signal-to-noise, blank determination, and linear regression methods for the estimation of detection and quantification limits for volatile organic compounds by gas chromatography. J. AOAC Int. 2009, 92, 1833-1838.

251. Phillips, M.; Cataneo, R. N.; Ditkoff, B. A.; Fisher, P.; Greenberg, J.; Gunawardena, R; Kwon, C. S.; Tietje, O.; Wong, C., Prediction of breast cancer using volatile biomarkers in the breath. Breast Cancer Res. Treat. 2006, 99, 19- 21.

252. Libardoni, M.; Stevens, P. T.; Waite, J. H.; Sacks, R., Analysis of human breath samples with a multi-bed sorption trap and comprehensive two-dimensional gas chromatography (GCxGC). J. Chromatogr. B 2006, 842, 13-21.

253. Phillips, M., Breath tests in medicine. Sci. Am. 1992, 267, 74-79.

254. Phillips, M.; Herrera, J.; Krishnan, S.; Zain, M.; Greenberg, J.; Cataneo, R. N., Variation in volatile organic compounds in the breath of normal humans. J. Chromatogr. B 1999, 729, 75-88.

160

255. Szulejko, J. E.; Bishop, K.; Solouki, T., High resolution biomarker identification in human exhaled breath of smokers and non-smokers, and cigarette smoke: connectivities to chronic obstructive pulmonary disease (COPD). Proceedings of the 50th ASMS Conference on Mass Spectrometry and Allied Topics 2002, Orlando, FL.

256. Gordon, S. M.; Wallace, L. A.; Pelllzzari, E. D.; O'Neill, H. J., Human breath measurements in a clean-air chamber to determine half-lives for volatile organic compounds. Atmos. Environ. 1967, 22, 2165-2170.

257. Zhang, J.; Ayusawa, T.; Minagawa, M.; Kinugawa, K.; Yamashita, H.; Matsuoka, M.; Anpo, M., Investigations of TiO2 photocatalysts for the decomposition of NO in the flow system - the role of pretreatment and reaction conditions in the photocatalytic efficiency. J. Catal. 2001, 198, 1-8.

258. Soylu, A. M.; Polat, M.; Erdogan, D. A.; Say, Z.; Yildirim, C.; Birer, O.; Ozensoy, E., TiO2-Al2O3 binary mixed oxide surfaces for photocatalytic NOx abatement. Appl. Surf. Sci. 2014, 318, 142-149.

259. Rusu, C. N.; Yates, J. J. T., Photochemistry of NO chemisorbed on TiO2(110) and TiO2 powders. J. Phys. Chem. B 2000, 104, 1729-1737.

260. Henderson, M. A.; Szanyi, J.; Peden, C. H. F., Conversion of N2O to N2 on TiO2 (110). Catal. Today 2003, 85, 251-266.

261. Trimpin, S.; Wang, B.; Lietz, C. B.; Marshall, D. D.; Richards, A. L.; Inutan, E. D., New ionization processes and applications for use in mass spectrometry. Crit. Rev. Biochem. Mol. Biol. 2013, 48, 409-429.

262. Monge, M. E.; Harris, G. A.; Dwivedi, P.; Fernandez, F. M., Mass spectrometry: recent advances in direct open air surface sampling/ionization. Chem. Rev. 2013, 113, 2269-2308.

263. Purcell, J. M.; Hendrickson, C. L.; Rodgers, R. P.; Marshall, A. G., Atmospheric pressure photoionization Fourier transform ion cyclotron resonance mass spectrometry for complex mixture analysis. Anal. Chem. 2006, 78, 5906-5912.

264. Jennings, K. R., Gas phase ion chemistry. Academic Press: New York, 1979.

265. Pintado-Herrera, M. G.; Gonzalez-Mazo, E.; Lara-Martin, P. A., Atmospheric pressure gas chromatography-time-of-flight-mass spectrometry (APGC-ToF-MS) for the determination of regulated and emerging contaminants in aqueous samples after stir bar sorptive extraction (SBSE). Anal. Chim. Acta 2014, 851, 1-13.

161

266. B. Zekavat; A. D. Olaitan; C. Rash; C. K. Chambliss; Solouki, T., Analysis of volatile organic compounds from gasoline samples using radio-frequency ionization/mass spectrometry. Proceedings of the 61st ASMS Conference on Mass Spectrometry and Allied Topics 2013, Minneapolis, MN.

267. Olaitan, A. D.; Zekavat, B.; Solouki, T., Electron Generation in Radio-Frequency Ionization (RFI). Proceedings of the Pittsburgh Annual Conference 2015, New Orleans, LA.

268. Szulejko, J. E.; Luo, Z.; Solouki, T., Simultaneous determination of analyte concentrations, gas-phase basicities, and proton transfer kinetics using gas chromatography/Fourier transform ion cyclotron resonance mass spectrometry (GC/FT-ICR MS). Int. J. Mass Spectrom. 2006, 257, 16-26.

269. Comisarow, M. B.; Marshall, A. G., Frequency-sweep Fourier transform ion cyclotron resonance spectroscopy. Chem. Phys. Letters 1974, 6, 489-490.

270. Zekavat, B.; Szulejko, J. E.; LaBrecque, D.; Solouki, T., Low cost higher frequency RF power supplies for quadrupole ion guides in FT-ICR instruments with external ion sources. Proceedings of the 57th ASMS Conference on Mass Spectrometry and Allied Topics 2009, Philadelphia, PA.

271. Modi, A.; Koratkar, N.; Lass, E.; Wei, B.; Ajayan, P. M., Miniaturized gas ionization sensors using carbon nanotubes. Nature 2003, 424, 171-174.

272. WebBook., N. C. Available: http://webbook.nist.gov/chemistry/.

273. Solouki, T.; Zekavat, B.; Olaitan, A. D., A novel radio-frequency ionization method for the analysis of wide classes of volatile organic compounds. Proceedings of the 61st ASMS Conference on Mass Spectrometry and Allied Topics 2013, Minneapolis, MN.

274. Olaitan, A. D.; Zekavat, B.; Solouki, T., Ionization Mechanism in Radio- Frequency Ionization (RFI). Proceedings of the 62nd ASMS Conference on Mass Spectrometry and Allied Topics 2014, Baltimore, MA.

Chapter Four

1. Phillips, M.; Gleeson, K.; Hughes, J. M.; Greenberg, J.; Cataneo, R. N.; Baker, L.; McVay, W. P., Volatile organic compounds in breath as markers of lung cancer: a cross-sectional study. Lancet 1999, 353, 1930-1933.

2. Buszewski, B.; Kesy, M.; Ligor, T.; Amann, A., Human exhaled air analytics: biomarkers of diseases. Biomed. Chromatogr. 2007, 21, 553-566.

162

3. Alvarado, J. S.; Rose, C., Static headspace analysis of volatile organic compounds in soil and vegetation samples for site characterization. Talanta 2004, 62, 17-23.

4. Zoccolillo, L.; Amendola, L.; Cafaro, C.; Insogna, S., Improved analysis of volatile halogenated hydrocarbons in water by purge-and-trap with gas chromatography and mass spectrometric detection. J. Chromatogr. A 2005, 1077, 181-187.

5. Furton, K. G.; Wu, L.; Almirall, J. R., Optimization of solid-phase microextraction (SPME) for the recovery of explosives from aqueous and post- explosion debris followed by gas and liquid chromatographic analysis. J. Forensic Sci. 2000, 45, 857-864.

6. Furton, K. G.; Wang, J.; Hsu, Y. L.; Walton, J.; Almirall, J. R., The use of solid- phase microextraction-gas chromatography in forensic analysis. J. Chromatogr. Sci. 2000, 38, 297-306.

7. Szulejko, J. E.; Solouki, T., Potential analytical applications of interfacing a GC to an FT-ICR MS: fingerprinting complex sample matrixes. Anal. Chem. 2002, 74, 3434-3442.

8. Christensen, E. D.; Chupka, G. M.; Luecke, J.; Smurthwaite, T.; Alleman, T. L.; Lisa, K.; Franz, J. A.; Elliott, D. C.; McCormick, R. L., Analysis of oxygenated compounds in hydrotreated biomass fast pyrolysis oil distillate fractions. Energy & Fuels 2011, 25, 5462-5471.

9. Solouki, T.; Szulejko, J. E.; Bennett, J. B.; Graham, L. B., A preconcentrator coupled to a GC/FTMS: advantages of self-chemical ionization, mass measurement accuracy, and high mass resolving power for GC applications. J. Am. Soc. Mass Spectrom. 2004, 15, 1191-1200.

10. Heffner, C.; Silwal, I.; Peckenham, J. M.; Solouki, T., Emerging technologies for identification of disinfection byproducts: GC/FT-ICR MS characterization of solvent artifacts. Environ. Sci. Technol. 2007, 41, 5419-5425.

11. Luo, Z.; Heffner, C.; Solouki, T., Multidimensional GC-Fourier transform ion cyclotron resonance MS analyses: utilizing gas-phase basicities to characterize multicomponent gasoline samples. J. Chromatogr. Sci. 2009, 47, 75-82.

12. Fialkov, A. B.; Steiner, U.; Lehotay, S. J.; Amirav, A., Sensitivity and noise in GC-MS: Achieving low limits of detection for difficult analytes. Int. J. Mass spectrom. 2007, 260, 31-48.

13. Phillips, M., Method for the collection and assay of volatile organic compounds in breath. Anal. Biochem. 1997, 247, 272-278.

163

14. Gorecki, T.; Pawliszyn, J., Field-portable solid-phase microextraction/fast GC system for trace analysis. Field Anal. Chem. Technol. 1997, 1, 277-284.

15. Daughtrey, J., H. E. ; Oliver, K. D.; Adams, J. R.; Kronmiller, K. G.; Lonneman, W. A.; McClenny, W. A., A comparison of sampling and analysis methods for low-ppbC levels of volatile organic compounds in ambient air. J. Environ. Monit. 2001, 3, 166-174.

16. Munson, B., Chemical ionization mass spectrometry. Anal. Chem. 1971, 43, 28A- 43A.

17. Jennings, K. R., Gas phase ion chemistry. Academic Press: New York, 1979.

18. Hunt, D. F., New ionization techniques in mass spectrometry. Int. J. Mass Spectrom. Ion Phys. 1982, 45, 111-123.

19. Zekavat, B.; Solouki, T., Radio-frequency ionization of organic compounds for mass spectrometry analysis. Angew. Chem. Int. Ed. 2013, 52, 2426-2429.

20. Chen, S. B., The determination of research octane number of gasoline fuels by octane analyzer. Defense Technical Information Center: 1983.

21. LeCroy, C.; Masiello, C. A.; A., R. J.; Hockaday, W. C.; Silberg, J. J., Nitrogen, biochar, and mycorrhizae: alteration of the symbiosis and oxidation of the char surface. Soil Biol. Biochem. 2013, 58, 248-254.

22. Szulejko, J. E.; Luo, Z.; Solouki, T., Simultaneous determination of analyte concentrations, gas-phase basicities, and proton transfer kinetics using gas chromatography/Fourier transform ion cyclotron resonance mass spectrometry (GC/FT-ICR MS). Int. J. Mass Spectrom. 2006, 257, 16-26.

23. Comisarow, M. B.; Marshall, A. G., Frequency-sweep Fourier transform ion cyclotron resonance spectroscopy. Chem. Phys. Lett. 1974, 6, 489-490.

24. Savory, J. J.; Kaiser, N. K.; McKenna, A. M.; Xian, F.; Blakney, G. T.; Rodgers, R. P.; Hendrickson, C. L.; Marshall, A. G., Parts-per-billion Fourier transform ion cyclotron resonance mass measurement accuracy with a "walking" calibration equation. Anal. Chem. 2011, 83, 1732-1736.

25. Solouki, T.; Szulejko, J. E., Bimolecular and unimolecular contributions to the disparate self-chemical ionizations of alpha-pinene and camphene isomers. J. Am. Soc. Mass Spectrom. 2007, 18, 2026-2039.

164

26. Zekavat, B.; Szulejko, J. E.; LaBrecque, D.; Solouki, T., Low cost higher frequency RF power supplies for quadrupole ion guides in FT-ICR instruments with external ion sources. Proceedings of the 57th ASMS Conference on Mass Spectrometry and Allied Topics 2009, Philadelphia, PA.

27. Abdi, H.; Williams, L. J., Principal component analysis. WIREs Comp. Stat. 2010, 2, 433-459.

28. Liu, Y.; Shi, Q.; Zhang, Y.; He, Y.; Chung, K. H.; Zhao, S.; Xu, C., Characterization of red pine pyrolysis bio-oil by gas chromatography−mass spectrometry and negative-ion electrospray ionization Fourier transform ion cyclotron resonance mass spectrometry. Energy & Fuels 2012, 26, 4532-4539. 29. Pollard, A. S.; Rover, M. R.; Brown, R. C., Characterization of bio-oil recovered as stage fractions with unique chemical and physical properties. J. Anal. Appl. Pyrolysis 2012, 93, 129-138.

30. Tessarolo, N. S.; Dos Santos, L. R.; Silva, R. S.; Azevedo, D. A., Chemical characterization of bio-oils using comprehensive two-dimensional gas chromatography with time-of-flight mass spectrometry. J. Chromatogr. A 2013, 1279, 68-75.

31. Liang, F.; Kerpen, K.; Kuklya, A.; Telgheder, U., Fingerprint identification of volatile organic compounds in gasoline contaminated groundwater using gas chromatography differential ion mobility spectrometry. Int. J. Ion Mobility Spectrom. 2012, 15, 169-177.

32. Venkatakrishnan, V. K.; Degenstein, J. C.; Smeltz, A. D.; Delgass, W. N.; Agrawal, R.; Ribeiro, F. H., High-pressure fast-pyrolysis, fast-hydropyrolysis and catalytic hydrodeoxygenation of cellulose: production of liquid fuel from biomass. Green Chem. 2014, 16, 792-802.

33. Zekavat, B.; Szulejko, J. E.; LaBrecque, D.; Olaitan, A. D.; Solouki, T., Efficient injection of low-mass ions into high magnetic field Fourier transform ion cyclotron resonance mass spectrometers. Rapid Commun. Mass Spectrom. 2014, 28, 230-238.

34. Garcia-Perez, M.; Adams, T. T.; Goodrum, J. W.; Geller, D. P.; Das, K. C., Production and fuel properties of pine chip bio-oil/biodiesel blends. Energy Fuels. 2007, 21, 2363-2372.

35. Mullen, C. A.; Boateng, A. A.; Goldberg, N. M.; Lima, I. M.; Laird, D. A.; Hicks, K. B., Bio-oil and bio-char production from corn cobs and stover by fast pyrolysis. Biomass Bioenergy. 2010, 34, 67-74.

36. Mante, O. D.; Agblevor, F. A., Storage stability of biocrude oils from fast pyrolysis of poultry litter. Waste Manage. 2012, 32, 67-76.

165

37. Alonso, D. M.; Bond, J. Q.; Dumesic, J. A., Catalytic conversion of biomass to biofuels. Green Chem. 2010, 12, 1493-1513.

38. Venderbosch, R. H.; Ardiyanti, A. R.; Wildschut, J.; Oasmaa, A.; Heeres, H. J., Stabilization of biomass-derived pyrolysis oils. J. Chem. Technol. Biotechnol. 2010, 85, 674-686.

39. Misron, N.; Rizuan, S.; Vaithilingam, A.; Mailah, N. F.; Tsuyoshi, H.; Hiroaki, Y.; Yoshihito, S., Performance improvement of a portable electric generator using an optimized bio-fuel ratio in a single cylinder two-stroke engine. Energies 2011, 4, 1937-1949.

40. Canakci, M.; Ozsezen, A. N.; Alptekin, E.; Eyidogan, M., Impact of alcohol- gasoline fuel blends on the exhaust emission of an SI engine. Renew. Energy 2013, 52, 111-117.

41. Window on State Government, http://www.window.state.tx.us/specialrpt/energy/renewable/ethanol.php, Last accessed March 21, 2013.

42. Maine Department of Environmental Protection, http://www.maine.gov/dep/air/mobile/ethanol.html, Last accessed March 21, 2013.

Chapter Six

1. Monge, M. E.; Harris, G. A.; Dwivedi, P.; Fernandez, F. M., Mass spectrometry: recent advances in direct open air surface sampling/ionization. Chem. Rev. 2013, 113, 2269-2308.

2. Trimpin, S.; Wang, B.; Lietz, C. B.; Marshall, D. D.; Richards, A. L.; Inutan, E. D., New ionization processes and applications for use in mass spectrometry. Crit. Rev. Biochem. Mol. Biol. 2013, 48, 409-429.

3. Kandiah, M.; Urban, P. L., Advances in ultrasensitive mass spectrometry of organic molecules. Chem. Soc. Rev. 2013, 42, 5299-5322.

4. Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S. F.; Whitehouse, C. M., Electrospray ionization for mass spectrometry of large biomolecules. Science 1989, 246, 64-71.

5. Hendrickson, C. L.; Emmett, M. R., Electrospray ionization Fourier transform ion cyclotron resonance mass spectrometry. Annu. Rev. Phys. Chem. 1999, 50, 517- 536.

166

6. Hillenkamp, F.; Karas, M.; Beavis, R. C.; Chait, B. T., Matrix-assisted laser desorption/ionization mass spectrometry of biopolymers. Anal. Chem. 1991, 63, 1193A-1203A.

7. Eberlin, L. S.; Norton, I.; Orringer, D.; Dunn, I. F.; Liu, X.; Ide, J. L.; Jarmusch, A. K.; Ligon, K. L.; Jolesz, F. A.; Golby, A. J.; Santagata, S.; Agar, N. Y.; Cooks, R. G., Ambient mass spectrometry for the intraoperative molecular diagnosis of human brain tumors. Proc. Natl. Acad. Sci. USA 2013, 110, 1611-1616.

8. Ifa, D. R.; Manicke, N. E.; Dill, A. L.; Cooks, R. G., Latent fingerprint chemical imaging by mass spectrometry. Science 2008, 321, 805.

9. Wiseman, J. M.; Ifa, D. R.; Zhu, Y.; Kissinger, C. B.; Manicke, N. E.; Kissinger, P. T.; Cooks, R. G., Desorption electrospray ionization mass spectrometry: Imaging drugs and metabolites in tissues. Proc. Natl. Acad. Sci. U S A 2008, 105, 18120-18125.

10. Nemes, P.; Woods, A. S.; Vertes, A., Simultaneous imaging of small metabolites and lipids in rat brain tissues at atmospheric pressure by laser ablation electrospray ionization mass spectrometry. Anal. Chem. 2010, 82, 982-988.

11. Dempster, A. J., The properties of slow canal rays. Phys. Rev. 1916, 8, 651-662.

12. Nier, A. O., A mass spectrometer for routine isotope abundance measurements. Rev. Sci. Instrum. 1940, 11, 212-216.

13. Jennings, K. R., Gas phase ion chemistry. Academic Press: New York, 1979.

14. Munson, M. S. B.; Field, F. H., Chemical ionization mass spectrometry. J. Am. Chem. Soc. 1966, 88, 2621-2630.

15. Munson, B., Chemical ionization mass spectrometry. Anal. Chem. 1971, 43, 28A- 43A.

16. Muller, E. W.; Bahadur, K., Field ionization of gases at a metal surface and the resolution of the field ion microscope. Phys. Rev. 1956, 102, 624-631.

17. Beckey, H. D., Experimental techniques in field ionisation and field desorption mass spectrometry. J. Phys. E: Sci. Instrum. 1979, 12, 72-83.

18. Lattimer, R. P.; Schulten, H.-R., Field ionization and field desorption mass spectrometry: past, present, and future. Anal. Chem. 1989, 61, 1201A-1215A.

19. Kind, T.; Fiehn, O., Advances in structure elucidation of small molecules using mass spectrometry. Bioanal. Rev. 2010, 2, 23-60.

167

20. Hunt, D. F., New ionization techniques in mass spectrometry. Int. J. Mass Spectrom. Ion Phys. 1982, 45, 111-123.

21. Banan Sadeghian, R.; Islam, M. S., Ultralow-voltage field-ionization discharge on whiskered silicon nanowires for gas-sensing applications. Nat. Mater. 2011, 10, 135-140.

22. Modi, A.; Koratkar, N.; Lass, E.; Wei, B.; Ajayan, P. M., Miniaturized gas ionization sensors using carbon nanotubes. Nature 2003, 424, 171-174.

23. Hunt, D. F.; Crow, F. W., Electron capture negative ion chemical ionization mass spectrometry. Anal. Chem. 1978, 50, 1781-1784.

24. Dougherty, R. C., Negative chemical ionization mass spectrometry. Anal. Chem. 1981, 53, 625A-636A.

25. Stemmler, E. A.; Hites, R. A., Methane enhanced negative ion mass spectra of hexachlorocyclopentadiene derivatives. Anal. Chem. 1985, 57, 684-692.

26. Stamatovic, A.; Schulz, G. J., Trochoidal electron monochromator. Rev. Sci. Instrum. 1968, 39, 1752-1753.

27. Stamatovic, A.; Schulz, G. J., Characteristics of the trochoidal electron monochromator. Rev. Sci. Instrum. 1970, 41, 423-427.

28. Illenberger, E.; Scheunemann, H.-U.; Baumgärtel, H., Negative ion formation in CF2Cl2, CF3Cl and CFCl3 following low energy (0-10 eV) impact with near monoenergetic electrons. Chem. Phys. 1979, 37, 21-31.

29. Laramee, J. A.; Kocher, C. A.; Deinzer, M. L., Application of a trochoidal electron monochromator/mass spectrometer system to the study of environmental chemicals. Anal. Chem. 1992, 64, 2316-2322.

30. Voinov, V. G.; Vasil'ev, Y. V.; Morre, J.; Barofsky, D. F.; Deinzer, M. L.; Gonin, M.; Egan, T. F.; Fuhrer, K., A resonant electron capture time-of-flight MS with trochoidal electron monochromator. Anal. Chem. 2003, 75, 3001-3009.

31. Dane, A. J.; Havey, C. D.; Voorhees, K. J., The detection of nitro pesticides in mainstream and sidestream cigarette smoke using electron monochromator-mass spectrometry. Anal. Chem. 2006, 78, 3227-3233.

32. Zekavat, B.; Solouki, T., Radio-frequency ionization of organic compounds for mass spectrometry analysis. Angew. Chem. Int. Ed. 2013, 52, 2426-2429.

168

33. Olaitan, A. D.; Zekavat, B.; Dhungana, B.; Hockaday, W. C.; Chambliss, C. K.; Solouki, T., Analysis of volatile organic compound mixtures using radio- frequency ionization /mass spectrometry. Anal. Methods 2014, 6, 4982-4987.

34. Solouki, T.; Zekavat, B.; Olaitan, A. D., A novel radio-frequency ionization method for the analysis of wide classes of volatile organic compounds. Proceedings of the 61st ASMS Conference on Mass Spectrometry and Allied Topics 2013, Minneapolis, MN.

35. Jiao, C. Q.; Ranatunga, D. R. A.; Vaughn, W. E.; Freiser, B. S., A pulsed-leak valve for use with ion trapping mass spectrometers. J. Am. Soc. Mass Spectrom. 1996, 7, 118-122.

36. Comisarow, M. B.; Marshall, A. G., Frequency-sweep Fourier transform ion cyclotron resonance spectroscopy. Chem. Phys. Letters 1974, 6, 489-490.

37. Marshall, A. G.; Hendrickson, C. L., Fourier transform ion cyclotron resonance detection: principles and experimental configurations. Int. J. Mass Spectrom. 2002, 215, 59-75.

38. Bartmess, J. E.; Georgiadis, R. M., Empirical methods for determination of ionization gauge relative sensitivities for different gases. Vacuum 1983, 33, 149- 153.

39. Solouki, T.; Szulejko, J. E., Bimolecular and unimolecular contributions to the disparate self-chemical ionizations of alpha-pinene and camphene isomers. J. Am. Soc. Mass Spectrom. 2007, 18, 2026-2039.

40. Zekavat, B.; Miladi, M.; Alfdeilat, A.; Somogyi, A.; Solouki, T., Evidence for sequence scrambling and divergent H/D exchange reactions of doubly-charged isobaric b type fragment ions. J. Am. Soc. Mass Spectrom. 2014, 25, 226-236.

41. Miller, T. M., Atomic and Molecular Polarizabilities. In CRC Handbook of Chemistry and Physics; Lide, D. R.; Frederikse, H. P. R., Eds. CRC Press Inc.: Baton Rouge, LA, 95th Edition; Vol. 10, pp 193-202.

42. Zekavat, B.; Szulejko, J. E.; LaBrecque, D.; Solouki, T., Low cost higher frequency RF power supplies for quadrupole ion guides in FT-ICR instruments with external ion sources. Proceedings of the 57th ASMS Conference on Mass Spectrometry and Allied Topics 2009, Philadelphia, PA.

43. Zekavat, B.; Szulejko, J. E.; LaBrecque, D.; Olaitan, A. D.; Solouki, T., Efficient injection of low-mass ions into high magnetic field Fourier transform ion cyclotron resonance mass spectrometers. Rapid Commun. Mass Spectrom. 2014, 28, 230-238.

169

44. Olaitan, A. D. Zekavat, B. Dhungana, B. Hockaday, W. C. Chambliss, C. K. Solouki, T., Application of radio frequency ionization mass spectrometry for volatile organic compound analysis. 69th Southwest Regional Meeting of the American Chemical Society 2013, Waco, TX.

45. Desfranҫois, C.; Abdoul-carime, H.; Khelifa, N.; Schermann, J. P., From 1/r to 1/r2 potentials: electron exchange between Rydberg atoms and polar molecules. Phys. Rev. Letter 1994, 73, 2436-2439.

46. Li, G.-Z.; Kim, H. S.; Guan, S.; Marshall, A. G., Radiatively self-cooled penning- trapped electrons: a new way to make gas-phase negative ions from neutrals of low electron affinity. J. Am. Chem. Soc. 1997, 119, 2267-2272.

47. Frankevich, V.; Knochenmuss, R.; Zenobi, R., The origin of electrons in MALDI and their use for sympathetic cooling of negative ions in FTICR. Int. J. Mass Spectrom. 2002, 220, 11-19.

48. Lifshitz, C.; Tiernan, T. O.; Hughes, B. M., Electron affinities from endothermic negative ion charge transfer reactions. IV. SF6, selected fluorocarbons and other polyatomic molecules. J. Chem. Phys. 1973, 59, 3182-3192.

49. Gant, K. S.; Christophorou, L. G., Attachment of slow electrons to hexafluorobenzene. J. Chem. Phys. 1976, 65, 2977-2981.

50. Marshall, A. G., Fourier transform ion cyclotron resonance mass spectrometry. Acc. Chem. Res. 1985, 18, 316-322.

51. Brown, L. S.; Gabrielse, G., Geonium theory: Physics of a single electron or ion in a Penning trap. Rev. Mod. Phys. 1986, 58, 233-311.

52. Li, G.-Z.; Guan, S.; Marshall, A. G., Sympathetic cooling of trapped negative ions by self-cooled electrons in a Fourier transform ion cyclotron resonance mass spectrometer. J. Am. Soc. Mass Spectrom. 1997, 8, 793-800.

53. Woodin, R. L.; Foster, M. S.; Beauchamp, J. L., Ion cyclotron resonance studies of radiative and dissociative electron attachment processes at low pressures. J. Chem. Phys. 1980, 72, 4223-4227.

54. Bekey, H. D., Field desorption mass spectrometry: A technique for the study of thermally unstable substances of low volatility. Int. J. Mass Spectrom. Ion Physics 1969, 2, 500-502.

55. Della-Negra, S.; Beyec, Y. L.; Htiansson, P., Spontaneous desorption time-of- flight mass spectrometry (SDMS): Time correlated emission of electrons and negative ions in a constant electric field. Nucl. Instrum. Methods B 1985, 9, 103- 106.

170

56. Della-Negra, S.; Deprun, C.; Beyec, Y. L.; Rgllgen, F.; Standing, K.; Monart, B.; Bolbach, G., On the mechanism of ion formation in spontaneous desorption time- of-flight mass spectrometry. Int. J. Mass Spectrom. Ion Processes 1987, 75, 319- 327.

57. Salehpour, M.; Hunt, J. E., Spontaneous desorption mass spectrometry: Applications and mechanisms. Int. J. Mass Spectrom. Ion Processes 1988, 85, 99- 115.

58. Luo, Z.; Heffner, C.; Solouki, T., Multidimensional GC-Fourier transform ion cyclotron resonance MS analyses: utilizing gas-phase basicities to characterize multicomponent gasoline samples. J. Chromatogr. Sci. 2009, 47, 75-82.

59. Skriver, H. L.; Rosengaard, N. M., Surface energy and work function of elemental metals. Phys. Rev. B 1992, 46, 7157-7168.

60. Herring, C.; Nichols, M. H., Thermoionic emission. Rev. Mod. Phys. 1949, 21, 185-270.

61. Fan, H. Y., Theory of photoelectric emission from metals. Phys. Rev. 1945, 68, 43-52.

62. Fowler, R. H.; Nordheim, L., Electron emission in intense electric fields. Proc. R. Soc. London. Series A 1928, 119, 173-181.

63. Jimenez, M.; Noer, R. J.; Jouve, G.; Jodet, J.; Bonin, B., Electron field emission from large-area cathodes: evidence for the projection model. J. Phys. D: Appl. Phys. 1994, 27, 1038-1045.

64. Podenok, S.; Sveningsson, M.; Hassen, K.; Campbell, E. E. B., Electric field enhancement factors around a metallic, end-capped cylinder. NANO 2006, 1, 87- 93.

65. Townsend, J. S., The Conductivity produced in gases by the motion of negatively- charged ions. Nature 1900, 62, 340-341.

171