1 ION MOBILITY MASS SPECTROMETRY AS A RAPID SEPARATIONS TECHNIQUE FOR ISOMERIC, METABOLOMIC, AND CLINICAL ANALYSIS By CHRISTOPHER RICHARD BEEKMAN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2015
2 Â© 2015 Christopher Richard Beekman
3 To my family without you, I would not have become me
4 A CKNOWLEDGMENTS I would like to extend my sincere gratitude to a number of friends, family, colleagues, and collaborators for making the last four years at the University of Florida a wonderful place to live and work during my graduate career. I would like to first thank my committee as a whole; to have a group of professors that I have the utmost respect for and that truly believed in me was a wonderful source of motivation. I would like to specifically thank my advisor, Dr. Richard A. Yost, for providing a great deal of intellectual freedom when it came to devising experiments and designing new modifications to our instruments. I would also like to thank him for exposing me to the world of mass spectrometry through networking at ASMS. This opportunity ha s truly opened doors to collaborations and friendships that I could have never imagined without being a part of this group. I would like to also thank the members of my committee individually: Dr. Benjamin Smith, for accepting my 3rd letter of recommendat ion from my cousin Dr. Warren Mino (which goes without saying, but thanks , Warren) and giving me the opportunity to become a part of the chemistry program at UF, Dr. Nicolo Omenetto, for his supportive insight on experimental design and my academic work, D r. Nicolas Polfer, for his help with electronics and Labview, and Dr. Timothy Garrett, for his insight I would like to extend a special thanks to all of the Yost group members. I would like to thank Dr. Robert Menger, Dr. Whitney Stutts, and Dr. Chia Wei Tsai for mentoring me since I began work in the Yost group. I want to extend a special thanks to Robin Kemperman for his support as my undergraduate resear cher, and, more importantly, for bringing his Dutch charisma into the group! I also could not have
5 survived graduate school without Christopher Chouinard and our bi Donuts (to see our favorite barista, Evelyn). (For my time here at least 1/8 of the Yost when it came to research and experimental design , and for that I am tremendously thankful. Last, but not least, I would like to thank my Yost g roup partner in crime , Rainey Patterson. We joined the Chemistry Department and the Yost group at the same time and have been through all the ups and downs of graduate school together from our ASMS trip to Vancouver to our annual ski trips to West Virgi nia, we were never I would also like to thank several people who made a world of difference in my graduate career outside of the lab. To start, I would to extend my love and gratitude to my girl friend , Dr. Jessica Towery. She has added a level of balance and who introduced me to slow pitch softball, which has been a wonderful outlet from work, and has broug ht me tremendous enjoyment and many friendships. I also want to thank my parents, Richard Beekman, Barbara Alice Dorothy Cunningham Beekman Powell (the 1st), and my stepdad, Jamie Powell, for all of their support from gas money to priceless advice , it h as all helped me to get to where I am today. I would like to also thank my biggest critics, my siblings, Jonathan and Brittany, and my stepbrothers, Greg, Keith and Brandon, who made me tougher, smarter, and a good cook. Without them, I would be an only child! Ha! But of course I would not have had the time of my life growing up in the zoo that was our household! I would like to thank them for always being there for me and providing me with the best family support system anyone could
6 ask for. Finally, I would like to thank the people who started it all, my grandparents, William and Lorraine Cunningham and John and Barbara Beekman. The humble support and constant encouragement to always try new things, strive to be my best, and to never give up clearly left a lasting impression. I would also like to thank a few people from my undergraduate days who helped me get to where I am today. First, I want to thank my best friend, Ryan De Palma, who ege. He understood the struggles of five classes, three labs, and an extended night of RA duty and campus paired with an extravagant conversation of science and science fiction. I would like to ere two qualities that I I was also lucky enough to have great collaborators throughout my graduate career that provided me with financial support, allowing me to freely do research without additional teaching responsibilities. I would like to thank Agilent Technologies for providing funding to the Yost group, which supported my research assistantship, as well as providing several instruments that provided the foundation fo r my research projects. I would also like to thank Owlstone for supplying the FAIMS instrumentation and equipment as well as support with experimentation.
7 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ .......... 10 LIST OF FIGURES ................................ ................................ ................................ ........ 11 ABSTRACT ................................ ................................ ................................ ................... 14 CHAPTER 1 AN INTRODUCTION TO ION MOBILITY MASS SPECTROMETRY ...................... 16 Research Motivation ................................ ................................ ............................... 16 A Brief Introduction to Ion Mobility Mass Spectrometry ................................ ........... 17 Ion Mobility Spectrometry (IMS) ................................ ................................ ....... 17 Drift Tube Ion Mobility Spectrometry (DTIMS) ................................ .................. 18 High field Asymmetric Wa veform Ion Mobility Spectrometry (FAIMS) .............. 20 Instrumentation ................................ ................................ ................................ ....... 21 Ionization Sources ................................ ................................ ............................ 22 Electrospray ionization ................................ ................................ ............... 22 Atmospheric pressure chemical ionization ................................ ................. 23 Ion Mobility Mass Spectrometry ................................ ................................ .............. 23 Ion Mobility Quadrupole Time of Flight Mass Spectrometer (IM QTOF MS) .... 23 High field Asymmetric Waveform Ion Mobility Spectrometry Triple Quadrupole Mass Spectrometer (FAIMS TQMS) ................................ .......... 25 Scope of the Dissertation ................................ ................................ ........................ 29 2 FUNDAMENTAL STUDIES OF DRIFT TUBE ION MOBILIT Y MASS SPECTROMETRY ................................ ................................ ................................ .. 38 Introduction ................................ ................................ ................................ ............. 38 Model Compounds ................................ ................................ ........................... 39 Experi mental ................................ ................................ ................................ ........... 40 Reagents and Solutions ................................ ................................ ................... 40 Instrument Parameters ................................ ................................ ..................... 41 Dri ft tube ion mobility mass spectrometry (DTIMS MS) analysis ............... 41 Liquid chromatography ion mobility mass spectrometry (LC IMS MS) analysis ................................ ................................ ................................ ... 42 Results and Discussion ................................ ................................ ........................... 43 Ion Mobility Parameters ................................ ................................ .................... 44 Drift Tube Ion Mobility Mass Spectrometry (DTIMS MS) ................................ . 46 Phthalic acid isomers ................................ ................................ ................. 46 Fumaric acid and maleic acid ................................ ................................ ..... 49 Liquid Chromato graphy Ion Mobility Mass Spectrometry (LC IMS MS) ........... 50
8 Tryptophan and d3 tryptophan ................................ ................................ ... 51 Summary ................................ ................................ ................................ ................ 54 3 FUNDAMENTAL STUDIES OF HIGH FIELD ASYMMETRIC WAVEFORM ION MOBILITY MASS SPECTROMETRY ................................ ................................ ..... 65 Introduction ................................ ................................ ................................ ............. 65 High Field Asymmetric Waveform Ion Mobility Mass Spectrometry (FAIMS MS) ................................ ................................ ................................ ............... 65 Model Compounds ................................ ................................ ........................... 69 Experimental ................................ ................................ ................................ ........... 70 Reagents and Solutions ................................ ................................ ................... 70 Instrument Parameters ................................ ................................ ..................... 70 Results and Discussi on ................................ ................................ ........................... 71 Instrumentation Modifications ................................ ................................ ........... 71 Optimization of the CF/DF Scan Range ................................ ........................... 74 Optimization of Gas Flow Rate and Temperature ................................ ............ 74 Modifier Addition: Carbon Dioxide ................................ ................................ .... 77 Modifier Addition: Methanol ................................ ................................ .............. 79 Modifier Addition: Other Modifiers ................................ ................................ .... 81 Summary ................................ ................................ ................................ ................ 81 4 ION MOBIL ITY METHODOLOGIES FOR TARGETED METABOLOMICS ............ 91 Introduction ................................ ................................ ................................ ............. 91 Metabolites of Interest ................................ ................................ ...................... 92 Methylmalonic acid and succinic acid ................................ ........................ 92 Fructose and glucose ................................ ................................ ................. 93 Experimental ................................ ................................ ................................ ........... 94 Reagents and Solutions ................................ ................................ ................... 94 Instrument Parameters Drift Tube Ion Mobility Mass Spectrometry (DTIMS MS) ................................ ................................ ................................ .. 94 Instrument Parameters High Field Asymmetric Waveform Ion Mobility Mass Spectrometry (FAIMS MS) ................................ ................................ .. 95 Results and Discussion ................................ ................................ ........................... 96 Experiments Using Drift Tube Ion Mobility Mass Spectrometry (DTIMS MS) ... 96 Methylmalonic acid and succinic acid ................................ ........................ 98 Fructose and glucose ................................ ................................ ................. 99 Experiments Using High Field Asymmetric Waveform Ion Mobility Mass Spectrometry (FAIMS MS) ................................ ................................ .......... 100 Met hylmalonic acid and succinic acid ................................ ...................... 101 Fructose and glucose ................................ ................................ ............... 104 Summary ................................ ................................ ................................ .............. 106 5 SUMMARY AND FUTURE WORK ................................ ................................ ....... 118 Conclusions ................................ ................................ ................................ .......... 118
9 Future Work ................................ ................................ ................................ .......... 120 LIST OF REFERENCES ................................ ................................ ............................. 122 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 125
10 LIST OF TABLES Table page 1 1 Ion Behavior ................................ ................................ ................................ ....... 34 2 1 LC gradient method. ................................ ................................ ........................... 56 2 2 Calculated CCS values for the phthalic acid isomers. ................................ ........ 58 2 3 Calculated CCS values ................................ ................................ ....................... 60 2 4 Observed drift times and resolving power calculations for each ion of tryptophan. ................................ ................................ ................................ .......... 63 4 1 Important IMS parameters ................................ ................................ ................ 107 4 2 Calculated CCS values for various analytes/ions. ................................ ............ 111
11 LIST OF FIGURES Figure page 1 1 Illustration of a drift tube ion mobility spectrometer. ................................ .......... 31 1 2 Illustration of a high field asymmetric waveform ion mobility spectrometer (FAIMS). ................................ ................................ ................................ ............. 32 1 3 Conceptual representation of different FAIMS mobility trajectories and their reaction to a compensation voltages. ................................ ................................ . 33 1 4 Theoretical differential ion mobility trends. ................................ ......................... 34 1 5 Agilent 6560 ion mobility quadrupole time of flight mass spectrometer (IM QTOF MS). ................................ ................................ ................................ ......... 35 1 6 Cutaway of the Agilent 6560 ion guide (drift tube). ................................ ............. 36 1 7 Diagram of Agilent 6 560 IM QTOF and ion flight path ................................ ........ 36 1 8 FAIMS instrumentation. ................................ ................................ ...................... 37 1 9 Diagram of the Agilent 6460 triple quad rupole mass spec trometer (TQMS). ..... 37 2 1 Phthalic acid isomers molecules ................................ ................................ ......... 56 2 2 Intensity vs drift time plot for the phthalic aci d monomers at 18.6 V/cm. ............ 57 2 3 Intensity vs drift time plot for the phthalic acid dimers at 18.6 V/cm. .................. 57 2 4 I ntensity vs drift time plot for the phthalic acid trimers at 18.6 V/cm. .................. 58 2 5 Fumaric acid and maleic acid. ................................ ................................ ............ 59 2 6 Drift spectrum for the fumaric and maleic acid [M H] ions at m/z 115. ............... 59 2 7 Structure of tryptophan. ................................ ................................ ...................... 60 2 8 LC IMS total ion chromatogram (TIC) of pooled blood plasma spiked with d3 tryptophan (internal standard). ................................ ................................ ............ 61 2 9 IMS MS data of pooled blood plasma extracted from the tryptophan peak ........ 62 2 10 Drift times and m/z values for each tryptophan (or fragment) ion in in Table 2 4. ................................ ................................ ................................ ........................ 64 3 1 Phthalic acid isomeric molecules ................................ ................................ ........ 83
12 3 2 Gas/solvent vapor delivery system for the addition of gas/solvent vapor modifiers to the FAIMS cell. ................................ ................................ ................ 83 3 3 Schematic of the solvent vapor delivery system. ................................ ................ 84 3 4 Schematic of the solvent vapor delivery system incorporated into the heated drying gas for the mass spectro meter. ................................ ............................... 84 3 5 CF/DF plots . ................................ ................................ ................................ ....... 85 3 6 FAIMS CF vs DF intensity plots for ortho, meta, and para phthalic acid [M H] ions at 165 m/z at a 31,000 ppm concentration of methanol in a 100% dry N 2 atmosphere. ................................ ................................ ................................ ........ 86 3 7 FAIMS CF vs DF intensity plots for ortho, meta, and para phthalic acid [M H] ions at 165 m /z under dry CO 2 conditions. ................................ ......................... 87 3 8 FAIMS CF vs DF intensity plots for ortho, meta, and para phthalic acid [M H] ions at 165 m/z under 25% dry CO 2 conditions ................................ .................. 88 3 9 FAIMS CF vs DF intensity plots for ortho, meta, and para phthalic acid [M H] ions at 165 m/z with various concentrations of methanol vapor in a 100% dry N 2 atmosphere. ................................ ................................ ................................ ... 89 3 10 FAIMS CF vs DF intensity plots for ortho, meta, and para phthalic acid [M H] ions at 165 m/z at a 31,000 ppm concentration of methanol in a 100% dry N 2 atmosphere . ................................ ................................ ................................ ........ 90 4 1 Isomers methylmalonic acid (MMA) and succinic acid (SA). ............................ 107 4 2 Isomeric sugars D glucose and D fructose ................................ ....................... 107 4 3 Drift spectrum for methylmalonic and succinic acid. ................................ ......... 108 4 4 Drift spectrum for D glucose, D fructose, L fructose, and mixtures of these sugars. ................................ ................................ ................................ .............. 109 4 5 Drift spectrum of the fragment ions for D glucose, D fructose, L fructose, and mixtures of these sugars. ................................ ................................ ................. 110 4 6 FAIMS CF vs DF intensity plots for MMA and SA, [M H] ion at 117 m/z under dry CO 2 conditions. ................................ ................................ ........................... 112 4 7 FAIMS CF vs DF intensity plots for MMA and SA, [M H] ion at 117 m/z under solvent vapor addition conditions with the head space containing acetonitrile. 113 4 8 FAIMS CF vs DF intensity plots for MMA and SA, [M H] ion at 117 m/z under solvent vapor additio n conditions with the head space containing isopropanol. ................................ ................................ ................................ .... 114
13 4 9 FAIMS CF vs DF intensity plots for MMA and SA, [M H] ion at 117 m/z under solvent vapor addition conditions with the head space containing acetonitrile and water. ................................ ................................ ................................ ......... 115 4 1 0 FAIMS CF vs DF intensity plots for MMA and SA, [M H] ion at 117 m/z under solvent vapor addition conditions with the head space containing acetonitrile and water and an axillary gas addition of carbon dioxide. ................................ 116 4 11 FAIMS CF vs intensity plots of MMA and SA, [M H] ion at 117 m/z under solvent vapor additi on conditions with the head space containing acetonitrile and wa ter with corresponding 2D CF vs DF plot . ................................ ............ 117
14 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partia l Fulfillment of the Requirements for the Degree of Doctor of Philosophy ION MOBILITY MASS SPECTROMETRY AS A RAPID SEPARATIONS TECHNIQUE FOR ISOMERIC, METABOLOMIC, AND CLINICAL ANALYSIS By Christopher Richard Beekman August 2015 Chair: Richard A. Yost Major: Chemistry Ion mobility spectrometry is an analytical technique used to separate ionized molecules based on their differing mobilities (which result from the ion physical properties) through an applied electric field. Despite recent advancements in instrumentation a nd method development in both drift tube ion mobility mass spectrometry ( DT IMS M S) and high field asymmetric waveform ion mobility mass spectrometry (FAIMS MS) , these separation techniques are relatively underutilized. Wit h the growing demand for more rapid and less complex separation techniques, particularly those of clinical relevance, ion mobility methods have the potential to fill this void where traditional LC and/or GC separation methods fall short. The work presented in this dissertation can be divided into three studies. First, a fundamental, proof of concept investigation of drift tube ion mobility mass spectrometry was conducted using an Agilent 6560 drift tube ion mobility quadrupole time of flight mass spectromet er (IMS QTOF). This work explored the optimization of instrumental parameters for improved ion transmission and resolution, as well as the development of DTIMS methodologies for the separation of small molecules. In addition, an LC IMS -
15 MS method was also evaluated, which demonstrated the utility of ion mobility in tandem with traditional separation techniques for an added dimension of separation. Second an investigation of high field asymmetric waveform ion mobility mass spectrometry (ultra FAIMS MS) was conducted using an Owlstone ultra FAIMS A1 paired with an Agilent 6460 triple quadrupole mass spectrometer (QQQ). This work involved an exploration of the fundamentals of separation by FAIMS and included the development of an apparatus for the addition of gas and/or solvent vapors into the FAIMS environment for enhanced separation of small isomeric compounds. The final study involved the development of methods for the separation of clinically relevant isomeric molecules using both DTIMS and FAIMS techniqu es. In summary, the research described in this dissertation introduces two ion mobility mass spectrometry techniques and provides experimental data to support the use of these techniques for the straightforward and rapid separation of small isomeric mole cules.
16 CHAPTER 1 AN INTRODUCTION TO ION MOBILITY MASS SPECTROMETRY Research Motivation Over the last decade, the field of ion mobility spectrometry (IMS), and more specifically ion mobility mass spectrometry (IMS MS), has grown in popularity and accepta nce as a useful and advantageous analytical technique for separations. The rapid growth in this field is a result of two important occurrences. First, due to significant advancements in instrumentation design and technology, improved resolution, selectiv ity, and sensitivity comparable to that of traditional separation techniques has been achieved. 1 , 2 Second, t hese advancements have also resulted in the introduction of commercially available ion mobility instruments, featuring traveling wave ion mobility spectrometers (TWIMS), 3 drift tube ion mobility spectrometers (DTIMS), 4 and high field asymmetric waveform ion mobility sp ectrometers (FAIMS), 5 , 6 , 7 which provide a more universal platform for research conducted in the field. With these recent developmen ts in ion mobility, the realm of potential applications has expanded significantly, furthering the acceptance of ion mobility as a fast and selective analytical technique for separations. 1 , 8 In the analytical field of chemistry where hyphenated mass spect rometry (MS) techniques are dominated by liquid chromatography (LC), ion mobility mass spectrometry (IMS MS) coupled with LC offers a new method that would yield more comprehensive data in comparison to that generated by LC MS methods only. 9 , 10 In many cases, ion mobility techniques can provide a degree of isomeric and conformer position data that is not easily attained by LC MS methods alone. It is also possible to use IMS MS as a standalone separation technique; 4 however, its effectiveness in
17 comparison to the current LC MS standard of separation is still under investigation. technique in comparison to LC, where analytical separations may be achieved at least an order of magnitude faster than LC separations and without the use of a mobile phase. With ion mobility, sample throughput can be increased while reducing the amount of waste produced for analytical separations. 11 The research presented in this dissertation is focused on evaluating and demonstrating the utility and novelty of two prototype ion mobility mass spectrometers. The overall goal of this research was to develop a robust foundation and understanding of the ion mobility s pectrometry (IMS) techniques so that it can be applied to the development of IMS MS methods of separation for clinical applications in the growing field of metabolomics. The results of this research indicate that both drift tube ion mobility (DTIMS) and h igh field asymmetric waveform ion mobility (FAIMS) provide rapid separation of small molecules, and can be utilized as tandem analytical techniques with LC (LC IMS MS) or as independent separation techniques (IMS MS). A Brief Introduction to Ion Mobility Mass Spectrometry Ion Mobility Spectrometry (IMS) I on mobility spectrometry (IMS) is reliant on the principle that ions subjected to an electric field in a gas media will flow along the applied field lines with a velocity equal to (eq 1 1) (1 1) where is the drift velocity of the ion, K and E is the electric field strength . If the electric field is held constant, drift velocity is
18 directly proportional to K , and so ions with different K values will have different velocities. The coefficient of mobility ( K les within the drift media. It is from these principles that drift tube ion mobility spectrometry has emerged. Drift Tube Ion Mobility Spectrometry (DTIMS) Drift tube ion mobility spectrometry (DTIMS) is centered on the principle that ions can be effecti arise from their various chemical and physical properties. As illustrated in Figure 1 1, ions enter the hollow drift tube, composed of ring electrodes, from the ionization s ource. The ions are then trapped between the trapping gates to form an ion packet. The ion packet is then quickly released in order to reduce the width (in time) of the ion packet, thus improving resolving power. As the ions pass through the effective dr ift region (the length of the drift tube held at constant field) they are exposed to the drift gas. The collisions sustained with which is dependent on size, shape, and ion charge state, and th us results in a change the detector as a shift in drift time. The calculated mobility is typically expressed as the average rotational collision cros of the Mason Schamp equation (eq 1 2) (1 2)
19 where is the average rotational collision cross secti on (CCS) , ze is the charge state of the analyte ion, k B is the Boltzman constant, T is temperature of the buffer gas (normalized to standard conditions to correct for variation in laboratory conditions), m I is the mass of the ion, m B is the mass of the buf fer gas molecules, t d is the corrected drift time, E is the electric field strength , L is the length of the effective drift tube, P is the pressure of the drift tube ( normalized to standard conditions to correct for variation in laboratory conditions ) , and N is the number density inside the drift tube. It has been demonstrated by comparison with experimental data that this equation accurately measures CCS values under ideal conditions and using helium as a drift gas. However as an economic alternative, it has also been demonstrated that using nitrogen as a drift gas produces good correlation between theoretical and experimental CCS values. 12 Separation utilizing DTIMS is highly effective for ions with significant differences in mobility. If the mobility differences are small, a longer drift tube is required to improve resolution of the various ions . The r es olving power is dependent on the degree of ion packet diffusion, drift tube voltage, drift gas composition, temperature and pressure. 1 , 13 The longer the ion mobility d rift tube the greater the degree of ion packet diffusion, which results in broader drift peaks . In order to improve both resolution and resolving power , advancements in instrumentation, such as the ion funnel, which refocuses ions both in front of the ion trap (front funnel and trap funnel) before the drift tube and after the drift tube (rear funnel) to improve ion transmission, and a dual gate ion trap is used to improve ion packet formation and release into the drift tube, both these have been implemente d in this instrumentation. 14 Various drift gases, which alter the ion neutral collision in the drift tube, may be used, ultimately changing the mobility of ions in order
20 to induce potential separation. 15 However, even with these modifications it is still possible for ions to exhibit similar mobility. As a result, a more non conventional ion mobility technique emerged, which explo its the variation in the ion mobility constant for each individual ion at high and low electric fields. This subdivision of ion mobility is known as differential mobility spectrometry (DMS) or high field asymmetric waveform ion mobility (FAIMS) . High fie ld Asymmetric Waveform Ion Mobility Spectrometry (FAIMS) Like DTIMS, high field asymmetric waveform ion mobility spectrometry (FAIMS) is also centered on the idea that ions may be separated based on their differing velocities, or mobilities, but, in this c ase, the use of two different electric fields greatly increases the potential for variance in mobility between two ions. In FAIMS an asymmetric waveform is applied between two parallel electrodes at atmospheric conditions , which creat es a differential fi eld perpendicular to the direction of ion motion , as shown in Figure 1 2 . 16 , 17 This differential high/low electric field, or dispersion field (DF), induces an ion mobility based on the variance in 17 Under these conditions any two ions with differing net variances, or mobility coefficients, will separate. 16 Under dispersion field conditions alone, only an ion with a net variance of zero will pass through the FAI MS electrodes, while an ion with a positive or negative net variance will drift towards one of the two electrodes and become neutralized. 16 Figure 1 3 illustrates the three types of theoretical ion trajectories: 1) equal high and low field mobilities, 2) larger high field mobilities, and 3) larger low field mobilities. As the difference in ion mobilities increases , an a dditional supplementary field, called the compensation field (CF), may be applied across the DF electrodes . This allows the
21 trajectories of the ions to be redirected so that ions with net variances in mobility other than zero may be detected . 16,17 Based on experimental observations, it has been shown that i ons exposed to varying dispersion field s will demonstrate one of three theoreti cal mobility trends. 17,18 As shown in Figure 1 4 t ype A ions increase in mobility as the strength of the dispersion field increases . Type B ions initially increase in ion mobility , followed by a decrease in mobility as the dispersion field voltage increases. Type C ions decrease in mobility as the dispersion field voltage increases. 17,18 These mobility trends are theorized to be the result of ion interactions with the applied field and/or specific components of the gas phase molecules in the FAIMS a tmosphere. 16 , 19 Increased mobilit y is induced by a greater ion to applied field interaction, whereas decreased mobil ity is the result of greate r ion to atmosphere interaction, 16 thus making it difficult to predict type B ion behavior. Table 1 1 lists theoretically predicted ion behavior base d on the ion polarity, the polarity of the high field, and the applied polarity of the CV need ed to detect the ion under a dry nitrogen atmosphere . These trends will be discussed in further detai l in chapter 3 of this dissertation. Instrumentation The instrumentation used to collect data for this dissertation is described below. All mass spectrometers were manufactured by Agilent Technologies, thus allowing the ionization sources to be used inter changeably between the instruments. Any additional modifications made to the commercial instruments will be further discussed in later chapters of this dissertation.
22 Ioni zation Sources Electrospray ionization E lectrospray ionization (ESI) is a technique used in mass spectrometry that produces charged aerosol particles by applying a high DC voltage to a liquid. This technique applies a high voltage to the ESI needle while the capillary inlet of the mass spectrometer is held at ground, or vice versa. 20 This charge offset produces a Taylor cone 20 at the tip of the ESI needle , thus producing small, highly charged droplets. The droplets are th en expelled away from the tip of the needle by the electric field, assisted by the flow of heated nebulizer gas. As these droplets drift towards the capillary of the mass spectrometer they are heated by a drying gas, which evaporat es the solvent, or desol vat es the analyte ions. As the solvent dissipates, the charges are left behind o n the shrinking droplet as it approaches the Rayleigh limit . 20 At this point , analyte ions can evaporate out of the charged droplet, or as the solvent completely evaporat es , the charge is left attached to the remaining analyte molecule , producin g an ion. The Agilent Jet Stream ESI source has an additio nal heated sheath gas , which is a slight modification from conventional ESI sources. 21 Thi s heated nitrogen sheath gas enhances desolvation, which minimiz es ion/solvent clusters . This is essential for reducing the amount of neutral clusters entering the mass spectrometer and improving ion transmission. This method of ionization m ay also produce multiple charging of analyte compounds as well. There are several advantages of using ESI as opposed to other ionization sources. Possibly the most significant ionization technique it pr oduces very little fragmentation of the compound of interest, so the molecular ion is almost always observed. 20 In addition, ESI is capable of ionizing
23 an array of compounds, forming ions such as charged , nonvolatile small molecules, multiply charged ions, and/or inorganic cations and anions. 20 This capability allows ESI to generate a large field of potential analyte ions that can be s eparat ed and/or observed by ion mobility mass spectrometric methods . Atmospheric pressure chemical ionization A tmospheric pressure c hemical ionization (APCI) is also widely used as an ionization technique in mass spectrometry, where analyte ions are produced via interactions between the analyte and ionized molecules. In APCI , electrons are produced by a corona discharge, 20 which then ionize the reagent gas that occupies the APCI source region . The ionized reagent gas molecules collide with gas phase analyte molecules (provided by a desolvated nebulized spray), inducing chemical i onization of the analyte by transferring the charge on the carrier gas to the analyte molecule. 20 APCI is also considered a soft ionization technique ; however, it does generate ions that differ from those produced by the ESI source , which increases the probability of analyte separation with the availability of two ionizati on techniques. 20 Additionally, the use of APCI may also reduce spectr al noise that results from samples with a high salt content, allowing for the analysis of samples that would otherwise be difficult to investigate with ESI methods. 20 Ion Mobility Mass Spectrometry Ion Mobility Quadrupole Time of Flight Mass Spectrometer (IM QTOF MS) An Agilent 6 560 ion mobility quadrupole time of flight ma ss s pectrometer (IM QTOF MS), shown in Figure 1 5, was used to obtain drift tube ion mobility data for the experiments described in this dissertation. This Agilent system can be equipped with both electrospray and atmospheric pressure chemical ionization sources (described
24 above). Past the ioniza tion source of the instrument are the advanced ion optics and ion mobility drift tube, illustrated in a cutaway view in Figure 1 6. The overall design of the drift tube consists of a series of ring electrodes wit h a potential gradient that directs the ions into the mass spectrometer. This portion of the instrument is divided into four parts: the front funnel, the trapping funnel, the drift cell, and the rear funnel. Ions produced in the ionization source enter the mass spectrometer through the heated glass capillary entering into the front funnel, which provides the first stage of vacuum , lowering the pressure to approximately 20 torr. The vacuum aids in removing excess gas and neutral molecules that have trave lled through the capillary, thus improving sensitivity. The electrodes in the front funnel are tapered to improve ion transmission into the trapping funnel, which further reduces the pressure to the drift tube operational pressure of 4 torr. In order to achieve a constant pressure inside the drift tube, the addition of an auxiliary gas is necessary. This additional gas is the same as the drift gas inside the drift tube and helps reduce the presence of impurities, which could alter the drift conditions an d result in inaccurate drift times. The trapping funnel is tube. These electrodes are necessary in order to acquire a time zero for the drift spectrum. The ion packe t is then pulsed into the drift cell of the instrument. The effective drift region of the drift tube is 78 cm long and is equipped with electrodes, which supply a uniform electric field that directs the flow of ions through the tube where they interact wi th neutral molecules in the drift tube atmosphere. The constant flow of auxiliary drift gas aids in sustaining a constant pressure of 4 torr inside the drift cell (since the
25 vacuum pumps are continuously removing drift gas molecules). As the ions reach th e end of the drift region they enter the rear ion funnel, which is used to focus the ions as they enter the hexapole ion guide and the mass spectrometer. The mass spectrometer is composed of three main components: the quadrupole mass filter, the collision cell, and a time of flight mass analyzer. The quadrupole mass filter is used to isolate and select a mass to charge ( m/z ) range to enter the collision cell in order to generate a multiple stage mass spectrum (MS/MS), or fragmentation spectrum. The collis ion cell may also be used independently of the quadrupole mass difficult to analyze on its own, it can be paired with the ion mobility spectrum ( m/z vs drift time) in order to decipher the fragmentation of all analyte ions. Mass spectra are generated by an orthogonal acceleration time of flight mass spectrometer, where ions are accelerated in the time of flight drift tube in a direction perpendicular to the initial di rection of ion motion. An ion mirror reflects the ions back down the drift tube to the schematic of the IM QTOF MS illustrating the ion flight path is shown in Figure 1 7. High field Asymmetric Waveform Ion Mobility Spectrometry Triple Quadrupole Mass Spectrometer (FAIMS TQMS) All FAIMS studies described in this dissertation were conducted on an Owlstone FA IMS unit as shown in Figure 1 8 . The instrumentation is compo sed of two main components: (1) the FAIMS chip (Figure 1 8 a), which is mounted in front of the sampling capillary (Figure 1 8b ) , and (2) the waveform generator (Figure 1 8c). The micro machined FAIMS chip is composed of multiple vertically stacked micro c hannels, which allow for elevated ion transmission in comparison to a single channel, and is an
26 order of magnitude smaller than conventional FAIMS electrodes, with a gap width of 100Âµm and a thickness of 700Âµm. Due to its size, the FAIMS chip requires muc h lower voltages (in comparison to larger, conventional cells) to achieve the field strength required for separation. A custom Owlstone wave form generator powers the FAIMS chip, providing a waveform dispersion field (DF) range of 0 to 350 Townsends (Td ) ( e lectric field per number density of the gas) and a compensation field (CF) range of 30 to 30 Td (equivalent to a dispersion voltage (DV) of 0 to ~600 V and a compensation voltage (CV) of ~ 50 to +50 V). The lowe r operational voltages permit rapid scanni ng of both the DF and CF, and allows for complete DF/CF scans in less than one minute, as opposed to larger FAIMS electrodes that require higher operational voltages and longer scanning sequences. The operat ional capabilities of the FAIMS cell allow for s ample conservation and minimize analysis time when optimizing the DF and CF settings for individual samples. The OFU hardware was controlled remotely using custom software provid ed by Owlstone/Agilent, which was synchronized with the triple quadrupole mass spectrometer (TQMS) . For FAIMS analysis, the following parameters must be specified: FAIMS cell temperature, cell dimensions, bias voltages, and CF and DF values. The firmware will then calculate the necessary CV and DV to achieve the desired experiment al parameters. T he FAIMS instrumentation may be utilized in one of two modes: primary (P) mode or hop (H) mode. When using the primary mode the CF/DF setting. Using the sweep method, the DF will remain constant while the CF will be scanned over each step between the specified start and end CF. At the conclusion
27 of the CF sweep, the DF will undergo a specified increase and the CF sweep will be repeated. This process will contin ue until the specified DF sweep has concluded. A complete 3 dimensional CF/DF/intensity plot can be constructed from the data obtained from the sweep to visually observe ion mobility behavior under the specific field conditions. Using the 3 D CF/DF/inten sity plot, it is possible to determine specific CF/DF values where analyte separation is observed. With these known values the hop mode, which switches between two or more specific CF/DF values, may be employed. This mode is particularly useful for deter mining the presence or absence of several analytes in a given sample. All exp eriments utilizing tandem FAIMS MS w ere conducted on an Agilent 6460 triple quadrupole mass spectromete r (TQMS) illustrated in Figure 1 9 . 21 The TQMS is capable of single stage mass spectral analysis as well as MS/MS analysis, which can be achieved by tailoring the radio frequency (RF) and direct current (DC) vol tages of the Q1 and Q3 quadrupole mass filters and applying a collision gas (N 2 ) for collision induced dissociation (CID) in Q2. 20 MS/MS analysis is particularly important for the characterization of many isobaric ions, which are identified by their differing fragmentation patterns as a result of CID in the collision cell. 20 The TQMS can be operated i n a number of modes: scan, selected ion monitoring (SIM), product ion, precursor ion, neutral loss, and selected reaction monitoring (SRM). The studies described in this dissertation primarily utilized scan and SIM mode, and us ed the other modes only for specific tracking of known fragmentation pathways. Scan mode operates by applying a scanning RF/DC voltage to Q1 while keeping Q3 in RF only mode (or vice versa) to observe the complete m/z range of the TQMS. This
28 method allows for the detection of any m/z values within the detectable range of a given sample (during a scan duty cycle), and is extremely useful for the detection of multiple ions; however, it is a slower method (approximately 1 2 orders of magnitude) than the the scan mode because only one (or a few) m/z values may be targeted per scan. As in scan mode, SIM mode maintains Q3 in RF only mode, but the RF/DC voltage of Q1 is held constant (rather than scanning the RF/DC voltages). Using these parameters, detecti on of a single m/z value is permitted. This particular mode is useful for rapid analysis of targeted analytes because the MS detection is concentrated on one m/z , optimizing sensitivity and detection speed. For complex analytes, such as structural isomers, it was necessary to utilize one voltage, permitting a single m/z value to pass into Q2 where it will undergo CID. As an example, structural isomers will have the same m/z val ue (held in Q1) and will pass freely into Q2 while all other background m/z values will be removed from the spectrum. After passing through Q2 , the isomers undergo CID, which may result in differing fragmentation patterns based on structural variations. If Q3 is set to scan mode , all fragmented ions will be observed; however, if the fragmentation pathways are known, Q3 can be held at a constant DC and RF voltage for a targeted m/z value, again reducing the scan overhead and MS duty time. SRM mode also ha s preset functional modes: (1) product ion mode where Q1 is held constant, CID is applied in Q2, and Q3 is scanned; (2) precursor ion mode where Q1 is scanned, CID is applied in Q2, and Q3 is held constant; and (3) neutral loss mode where Q1 and Q3 are sca nned simultaneously
29 while CID is applied in Q2. If multiple MS/MS transitions are monitored in succession, the TQMS is sa id to be operated in S RM mode or selected reaction monitoring. Scope of the Dissertation Despite recent advances in instrumentation an d methods in the field of ion mobility mass spectrometry (IMS MS) the technique is still relatively underutilized as a tool for separations, which is likely due to the critical need for a more fundamental understanding of ion behaviors under various condit ions. To address this challenge, two ion mobility techniques were investigated: drift tube ion mobility mass spectrometry (DTIMS MS ) and high field asymmetric waveform ion mobility mass spectrometry (FAIMS MS ) . The studies presented in this dissertation involve the detailed exploration of the fundamentals of each technique and how various conditions and parameters affect ion mobility. It is anticipated that these studies will aid in the understanding of ion mobility behaviors in order to expand the use o f ion mobility spectrometry, particularly in applications of clinical relevance. Chapter 2 investigates the use of drift tube ion mobility mass spectrometry for the separation of small isomeric molecules. Various instrumentation parameters are explored as well as the use of liquid chromatography in tandem with ion mobility (LC IMS MS) for the separation of co eluting compounds. Chapter 3 explores the fundamentals of the new high field asymmetric waveform ion mobility mass spectrometry (ultra FAIMS MS) pl atform as well as the utility of gas/solvent vapor modifiers for enhancing the separation of small isomeric compounds. Chapter 4 assesses the use of both DTIMS MS and ultra FAIMS MS for the separation and analysis of clinically relevant isomeric metabolit es, demonstrating the utility of ion mobility as a rapid separation technique. Finally, Chapter 5 provides a brief summary of
30 the work presented in this dissertation and suggests future areas of study within the field of ion mobility mass spectrometry.
31 Figure 1 1. Illustration of a drift tube ion mobility spectrometer. Ions are created in the ion source and enter into the hollow ring electrodes of the drift tube. The ions are then trapped in the trapping gate creating an ion packet. This ion pac k et is then released into the drift tube; the effective drift region of the tube is held at a constant electric field, E, and as a result ions of different mobilities, K, will separate due to differences in velocity. As the ions reach the detector their dr ift times are recorded and plotted against time, resulting in the mobility spectrum.
32 Figure 1 2. Illustration of a high field asymmetric waveform ion mobility spectrometer (FAIMS). Ions are created in the ion source and enter the gap between the FAIMS electrode s , which is composed of two parallel plates. As the ion enters the gap between the electrodes its mobility becomes dependent on the applied field. This field is created by the applied voltage from the asymmetric waveform. The ratio of the time the ion is exposed to the low field t L (illustrated as the blue portion of the waveform) and high field t H (illustrated as the red portion of the waveform) is 2 to 1, with the intensity of the low to high field to be 1 to 2. Ions that have the same net di fference in mobility over this asymmetric field (illustrated as the blue line for low field mobility and the red dotted like for high field mobility) pass through the FAIMS electrode to the mass spectrometer inlet. Any deviations in mobility will result i n a positive or negative shift from zero depending on the change in mobility, and is extrapolated over the length of the FAIMS electrode. Ions with a net difference in mobility greater than the gap distance will be neutralized by colliding with the FAIMS plate electrode. To detect these ions an additional compensation voltage is applied in order to shift the applied waveform to compensate for the ion trajectory.
33 Figure 1 3. Conceptual representation of different FAIMS mobility trajectories and their rea ction to a compensa tion voltages . Zero on the X axis represents the front the FAIMS electrode, and thirty represents the back of the FAIMS cell or the entrance into the mass spectrometer end. The yellow line depicts the detection trajectory of ions that pass through the FAIMS electrode. The solid lines represent ion trajectories under dispersion field (DF) alone, and the black solid line demonstrates a zero difference in mobility, or where the mobility at high field (KH) is equal to the mobility at low fi eld (KL). The zig zag trajectory pattern is representative of the frequency of the applied asymmetric waveform, where if the ion trajectory exist the FAIMS cell around the yellow line it will be detected. The red solid line represents an ion that has a l arger KH and gradually moves towards the high field and away from the detected trajectory. The solid blue line represents an ion with a larger KL and gradually moves towards the low field and away from the detected trajectory. In order to detect ions tha t drift away from the detection trajectory a compensation voltage (CV) is applied. The dotted lines indicate the respective trajectories after a CV is applied to the waveform, resulting in the detection of ions with a larger KL, and equal mobility and the larger KH ions drifting away from the detection trajectory. -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 0 5 10 15 20 25 30 35 Mobility Trajectories Detected KH/KL KH = KL KH > KL KH < KL CV + 1V KH = KL CV + 1V KH > KL CV + 1V KH < KL
34 Figure 1 4 . Theoretical differential ion mobility trends. The solid back line shows equal mobility at high fields (KH) and low fields (KL). The solid red line represents an overall increasin g differential mobility (KH > KL ) and has been referred to as type A ion behavior. The solid blue line shows a decreasing mobility tren d (KH < KL) and is denoted as type C ion behavior. Th e solid green line represents type B ion behavior where the ion mob ility increases at lower fields and then decreases as the field strength increases. The dotted line denotes where type B ion behavior mimics type A and C behavior relative to its own mobility. Table 1 1. Ion Behavior I on Polarity High Field P olarity A pplied CV I on B ehavior + + + type C + + type A + + type A + type C + + type C + type A + type A type C 0 1 2 KH/KL Increasing Electric Field Strength Differential Ion Mobility Trends KH = KL KH > KL KH < KL type B increasing decreasing transition
35 Figure 1 5. Agilent 6560 ion mobility quadrupole time of flight mass spectrometer (IM QTOF MS). 22
36 Figure 1 6. Cutaway of the Agilent 6560 ion guide (drift tube). Photo courtesy of Agilen t Technologies. Figure 1 7. Diagram of Agilent 6560 IM QTOF and ion flight path . 22
37 Figure 1 8. FAIMS instrumentation. A) M icro machined FAIMS chip composed of multiple vertically stacked micro channels . B) FAIMS chip mounted in front of the capillary inlet to the mass spectrometer . Photo courtesy of author. C) FAIMS waveform generat or mounted under the ionization source. D) Agilent 6460 TQMS. Photo courtesy of author . Figure 1 9. Diagram of the Agilent 6460 triple quadrupole mass spectrometer (TQMS) . 21
38 CHAPTER 2 FUNDAMENTAL STUDIES OF DRIFT TUBE ION MOBILITY MASS SPECTROMETRY In troduction As the field of mass spectrometry continues to expand as an importan t analytical technique , particularly with the growing interest in metabolomics and clinical analysis , it is faced with an increasing demand for more rapid, less complex analyse s as well as methods that require smaller sample sizes and minimal amounts of additional chemicals (e.g., solvents) for analysis. To address these challenges, the technique of drift tube ion mobility mass spectrometry (DTIMS MS) was investigated. DTIMS M S is capable of analysis. In addition, many current separation methodologies (such as LC MS and GC MS) of clinical significance require complex and time consuming deriva tization steps prior to analysis, and with DTIMS these steps could be minimized or eliminated altogether. This chapter addres ses the fundamentals of the DTIMS MS technique and its utility as a rapid separation tool. DTIMS is centered on the principle th at ions can be separated, on the millisecond timescale, based on their differing mobilities, which arise from their various chemical and physical properties. Based on the size, shape, and chemical properties of the ions of interest, varying interactions between the analyte ions and gas molecules of the drift tube atmosphere will result in different mobilities for each ion. These differing mobilities translate to slight deviations in drift time for each ion, and even drift times varying just half of a mil lisecond can result in baseline separation of the ions of interest. Additional supporting background information regarding drift tube ion mobility can be found in the previous chapter. The work presented below focuses on 1)
39 instrumentation optimization f or increased ion transmission and improved resolution for mobility data and 2) method development and strategies for separation of small molecules and isomeric compounds. Model Compounds The DTIMS MS technique was evaluated using several sets of isomers. The small isomeric molecules phthalic acid (ortho phthalic acid), isophthalic acid (meta phthalic acid), and terephthalic acid (para phthalic acid) were selected for analysis as they represent relatively noncomplex structural isomers. In addition, the pr esence of the carboxylic acid functional groups made them easy to ionize, which was important for avoiding difficulties with ion transmission. The phthalic acid isomers have also been successfully separated by FAIMS MS techniques (see chapter 3), 19 making them a suitable choice for further ion mobility studies. Another set of small isomeric molecules, fumaric acid and maleic acid, were also examined by DTIMS MS methods. These compounds represent a second type of relatively simple isomers cis /trans stereoisomers. Like the phthalic acid isomers, maleic and fumaric acid also contain the easily ionized carboxylic acid functional group, which minimizes issues with ion intensity related to the chemical structure of the compound. These two isomers also resemble the clinically relevant metabolite succinic acid, which is the saturated form (no double bond) of fumaric acid. The analysis of succinic acid (and methylmalonic acid) by ion mobility methods (DTIMS and FAIMS) is discussed in further detail in chapter 4. Finally, to demonstrate the utility of IMS in tandem with LC techniques (LC IMS MS) a complex sample of pooled blood plasma was analyzed. To reduce complexity, one target compound, tryptophan, was selected for analysis. Tryptophan was a su itable
40 choice as methods for the extraction of tryptophan from blood plasma and analysis by LC were well established, minimizing any issues related to analyte detection. of e collected without hindering LC MS analysis (i.e., IMS did not cause a delay in the introduction of sample into the MS from the LC, which would have resulted in a delayed response from the MS and inaccurate retention times). Deuterated tryptophan (d3 try ptophan) was spiked into the sample of blood plasma as an internal standard for the proof of concept study, and it was also used to monitor analyte fragmentation patterns (by following the deuterated tryptophan fragments) that occurred after the sample exi ted the drift tube. Experimental Reagents and Solutions Phthalic acid (ortho phthalic acid), isophthalic acid (meta phthalic acid), and terephthalic acid (para phthalic acid ) were purchased from Acros Organics. No further purification was necessary , and f resh stock solutions were prepared prior to experiments. Stock solutions were prepared in the following manner: each phthalic acid isomer was weighed in a separate volumetric flask, dissolved in 90/10 methanol/water , and diluted to a final concentration o f 10 ppm. Sample was introduced into the ESI source of the DTIMS MS instrument at a constant flow rate of 10 ÂµL/min from the syringe pump. Fumaric acid and maleic acid were purchased from Fisher Scientific . No further purification was necessary , and fres h stock solutions were prepared prior to experiments. Stock solutions were prepared in the following manner: fumaric acid and maleic acid were weighed in separate volumetric flasks, dissolved in 100% methanol ,
41 and diluted to a final concentration of 1 0 pp m. Sample was introduced into the ESI source of the DTIMS MS instrument at a constant flow rate of 10 ÂµL/min from the syringe pump. Pooled blood p lasma was obtained from the American Red Cross. The blood plasma was purified via a protein precipitation : 100 ÂµL of thawed plasma sample was pipetted into a 2 mL Eppendorf tube. 10 ÂµL of the d3 tryptophan was added, followed by 800 ÂµL of acetonitrile to precipitate the proteins in the solution. The Eppendorf tube was then vortexed for ~15 seconds and cooled for 30 minutes in a 80 C freezer for further precipitation of the proteins. After freezing, the samples were centrifuged for 10 minutes (<10 C) at 20,000 x g, followed by a transfer of 250 ÂµL of the supernatant to a new 2 mL Eppendorf tube. The liquid supernatant was blown t o dryness under N 2 and reconstituted in 150 ÂµL of H 2 O with 0.1% f ormic a cid. The reconstituted sample was vortexed for ~15 seconds, transferred to an LC autosampler vial, and analyzed by LC IMS MS. Instrument Parameters Drift t ube ion m obility mass s pectr ometry (DTIMS MS) analysis An Agilent 6560 drift tube ion mobility quadrupole time of flight mass spectrometer (IMS QTOF) was used for data acquisition. The IMS QTOF was tuned for intensity and resolving power and calibrated for m/z accuracy daily. Additio nally, a collision cross section calibration method was conducted daily in order to normalize calculated CCS values for variations in daily conditions. The CCS calibration method measures the drift time of the m/z 922 calibration ion (with a known CCS valu e) over a range of eight drift tube drift fields (18.6, 17.3, 16.0 14.7, 13.5, 12.2, 10.9, and 9.6 V/cm,
42 synonymous with drift voltages 1450, 1350, 1250, 1150, 1050, 950, 850 , and 750 V). Each spectrum was collected for 30 seconds, resulting in an overall experiment of 4 minutes . The data were then examined in the Agilent IM MS Browser B.06.00, which permits the calculation of the CCS value based on the different drift times of the ion over the varying drift voltages. The CCS calculation was then compared to the known CCS value of the ion to generate a CCS calibration factor, which accounts for daily variations in temperature and pressure in the drift tube . In addition CCS calibration the Agilent IM MS Browser B.06.00 software was used for all subsequent d ata analysis. All direct analysis methods on the IMS QTOF utilized direct injection of the analyte standards in to the ESI source. IMS studies were conducted using nitrogen as the drift tube collision gas with the drift tube pressure held at approximatel y 4 trorr. The drift tube temperature was maintained at approximately 32 C. Both pressure and temperature measurements were recorded for each experiment and used in the calculations for analyte collision cross section (CCS). Standard solutions were directly infused via a syr inge pump at a flow rate of 10 pectrometer and ionization source parameters were as follows : gas flow at 5 L/min, gas temperature at 325 C, nebulizer pressure at 20 psi, sheath gas flow at 8 L/ min, sheath gas temperature at 275 C, capillary voltage at 4000 V, nozzle voltage at 1000 V, a nd fragmentor voltage at 400 V. Liquid chromatography ion mobility mass spectrometry (LC IMS MS) analysis An Agilent 1200 series liquid chromatography system with an Agilent Zorbax SB Aq column, 2.1 x 100 mm, 1.8 Âµm particle size, maintained at 35 C was used for HPLC analysis. Pooled blood plasma samples were analyzed as follows: a flow rate of 0.3
43 mL/min of 100% mobile phase A (0.1% formic acid in water) for 1 minute, followed by a linear ramp to 35% mobile phase A / 65% mobile phase B (0.1% formic aci d in acetonitrile) at 11 minutes and holding at 35% A / 65% B until 13 minutes, a second linear ramp to 5% A / 95% B at 18 minutes and holding at 5% A / 95% B until 20 minutes, and a third linear ramp to 100% A at 21 minutes and holding at 100% A until 22 minutes (Table 2 1). Injection volume was 2 l. The Agilent 6560 drift tube ion mobility quadrupole time of flight mass spectrometer (IMS QTOF) was operated in positive mode in the low mass range. The mass spectrometer and ionization source parameters w ere as follows : gas flow at 10 L/min, gas temperature at 350 C, nebulizer pressure at 30 psi, sheath gas flow at 12 L/ min, sheath gas temperature at 300 C, capillary voltage at 3500 V, nozzle voltage at 2000 V, and fragmentor voltage at 400 V. Results and Discussion In order to apply the DTIMS technique to a wide range of separations applications it was necessary to establish a more fundamental understanding of the instrumentation. Initially, all experiments were conducted using the optimized parameters e stablished for the tune and calibration ions ( m/z 622 and 922); however, the analysis of the smaller molecules ( m/z range of 100 200) described above resulted in a significant decrease in ion transmission and drift time resolution. Due to these observatio ns further optimization over a number of instrument parameters was explored to improve ion transmission and resolution of molecules at the low end of the observable m/z range.
44 Ion Mobility Parameters The m/z 922 ion was used for daily tuning and calibrat ion of the CCS correction factor, with the following normal DTIMS MS operating parameters : fragmentor voltage at 2 00 V , high pressure funnel RF at 150 V, trap funnel RF at 150 V, trap entrance grid delta at 14.5 V, trap exit grid 1 delta at 7 V, trap exit grid 2 delta at 1.5 V, and rear funnel RF at 150 V. These optimized values produced sharp drift time spectra for the calibration ion, resulting in an accurate CCS correction factor and calculated collision cross section value (in comparison to the theoret ical CCS value for the calibration ion). However, for the smaller molecules studied in this work ( m/z range of 100 200) these operational parameters produced broad peaks in the drift time spectra with low ion intensity in comparison to the larger calibrati on ion. As a result, new ion mobility parameters were explored to improve the ion intensity and drift time resolution for the smaller molecules. After manual tuning of the ion mobility parameters it was observed that a higher fragmentor voltage in the ion source improved the ion intensity for lower m/z values. It is likely that dimers, trimers, and adducts are more readily formed after the ionization of compounds of lower m/z . This is theorized to be due to charge stabilization for the smaller molecules. The increased presence of these dimers, trimers, and adducts would require a higher fragmentor voltage in order to break up these species, leaving behind the monomer ion of interest. As a result the fragmentor voltage was increased from 200 V to 400 V f or the molecules of interest in this study. Another important ion mobility parameter that significantly affected the ion intensity of the smaller molecules was the RF voltages used for focusing of the ions in the ion funnels. When the RF voltages were l ower, to 60 V for all three funnels (signal
45 was greater when the voltages for all three funnels were equal), there was an observed increase in the ion intensity for low (50 400) m/z ions. In addition, the lower RF voltages on the ion funnels affected the o bserved ratio in intensity of the two drift peaks that corresponded to the same analyte ion for the lower mass molecules. [An interesting phenomena was observed where two considerably different drift times (approximately 10 milliseconds apart) were record ed for a single analyte ion. This is believed to be the result of fragmentation of a dimer, trimer, or adduct of the analyte ion after exiting the drift tube. Since the larger dimer/trimer/adduct species would exit the drift tube (and then fragment) late r than the monomer species two drift times would be observed for the same ion. This idea is discussed in detail later in the chapter.] The lower RF voltages resulted in an increase in intensity for the true monomer species of the analyte (the faster drif t time). One proposa l to explain this observation is that the lower RF voltages increased pre drift tube fragmentation of the dimer /trimer/adduct species to a more stable monomeric ion, thus increasing the ion intensity of the monomer at the earlier drift time. Another theory is that the lower RF voltages simpl y increas ed transmission for the monomeric ion from the ion source. It was observed that the lower RF voltages increased the intensity of the shorter, monomeric drift time for a number of small mol ecules examined experimentally, and therefore, the lower RF voltage setting was optimized to 60 V. Optimization of other ion mobility parameters addressed drift spectra resolution. The voltages of the ion trap ping gates had a tremendous effect on drift s pectrum resolution. The voltage off sets on these gates are responsible for preventing ions from penetrating the drift tube before the ion packet is released. If the voltages are too low,
46 some ions may not have enough energy to pass through the gates and enter the drift tube, resulting in inaccurate drift times. If the gate voltages are too high the ions can become unstable and collide with the trap, reducing the ion intensity. High voltages can also push the ions further away from the gates, reducing the resolution of the ion pulse when the gates open for ion packet injection into the drift tube. Fine tuning of the ion trap grid voltages improved ion intensity and drift peak resolution , which was thought to be the result of a smoother , more compact io n pulse into the drift tube. The optimized ion trap grid voltages were as follows: trap entrance grid delta at 8.5 V (was 14.5 V), trap exit grid 1 delta at 6 V (was 7 V), and trap exit grid 2 delta at 9 V (was 1.5 V). s also optimized to improve drift peak resolution. It was observed that quicker trap release times and shorter trap fill times improved resolving power; however, for low analyte concentrations decreasing these times drastically reduced ion intensity. Wit h increased trap fill times the signal for the lower concentration analyte ion improved. However, longer trap fill times for high concentration analyte ions induced a space charging effect where a drastic reduction in ion drift spectrum resolution was ob served. Due to the significant effect the analyte concentration has on the optimized ion trap trap time, the trap fill times were adjusted for each experiment. The trap release time was held constant at 150 Âµs, but the trap fill time was adjusted from ru n to run and ranged from 5 20 ms. Drift T ube Ion Mobility Mass Spectrometry (DTIMS MS) Phthalic acid isomers Neat standards of p hthalic acid (ortho phthalic acid), isophthalic acid (meta phthalic acid), and terephthalic acid (para phthalic acid) , shown in Figure 2 1, were
47 analyzed in negative mode by drift tube ion mobility mass spectrometry (DTIMS MS). The most prominent ion observed in the mass spectrum was the [M H] ion at m/z 165 . The ionization source , drift tube, and mass spectrometer were opti mized for the analysis of smaller molecules as described in the previous section. The DTIMS MS data were processed using the IM MS browser to calculate the CCS value for each isomer as described in a previous section. For direct comparison of drift times for each isomeric ion the experimental data were collected on the same day in order to reduce any variation due to day to day temperature and pressure changes. The data were also collected using a single drift tube field setting of 18.6 V/cm. This value resulted in the greatest observed resolution in drift time spectra in comparison to all other drift tube field settings. The drift time spectrum for the phthalic acid isomers is shown in Figure 2 2. The drift time plot shows separation between ortho ph thalic acid and its isomers meta and para phthalic acid, with R s (resolution) = 1.99 between ortho and meta phthalic acid and R s = 1.76 between ortho and para phthalic acid. These differing drift times correlate to the following calculated CCS values f or ortho , meta , and para phthalic acid: 130.3Ã…, 136.3Ã…, and 135.3Ã…, respectively. Separation was not observed between meta and para phthalic acid. After review of the data, additional drift peaks, with longer drift times, were observed for each of the monomeric m/z isomeric ions (as confirmed by the mass spectrum for each peak). Further analysis revealed that the longer monomeric ion drift times matched the drift times for sodiated dimer/double sodiated trimer analyte ion peaks. This suggests that th e secondary monomeric peaks are a result of
48 fragmentation of the dimer and trimer analyte ions (forming the monomer ions) after exiting the drift tube. Since it was clear that dimer and trimer analyte species were present in the drift tube, the drift spe ctra for these species were analyzed. The drift spectra of the dimer [2(M H)+Na] ions, m/z 353 (Figure 2 3), and the trimer [3(M H)+2Na] ions, m/z 542 (Figure 2 4), were particularly interesting and unexpected. The observed separation between the dimer and trimer isomeric peaks was considerably different than the separation observed in the monomeric drift spectrum (Figure 2 2). Ortho phthalic acid was baseline separated from the meta and para phthalic acid ions in the monomeric drift spectrum, but bas eline separation was not quite achieved for any of the ions in the dimer ion drift spectrum (although the small degree of separation between meta and para phthalic acid in the dimer spectrum was not observed in the monomer drift spectrum). Even more inte restingly meta phthalic acid was baseline separated from ortho and para phthalic acid in the trimer drift spectrum. All calculated CCS values corresponding to the monomer, dimer, and trimer species of the three phthalic acid isomers are shown in Table 2 2. The data obtained from these studies suggest that gas phase clustered molecules 1) exist and are stable in the drift tube and 2) have unique drift times and separation trends in comparison to their corresponding monomeric species. The evaluation of th e phthalic acid isomers by DTIMS MS not only demonstrated baseline separation of small isomeric ions, but that separation trends vary significantly between monomer, dimer, and trimer species of the analyte ions. Although it was only possible to achieve se paration between ortho phthalic acid from meta and para -
49 phthalic acid in the monomer drift spectrum, meta phthalic acid could be separated from ortho and para phthalic acid in the trimer drift spectrum, illustrating that the use of higher order ion struc tures could become a useful and versatile strategy for designing DTIMS MS separation methods. This separation strategy also illuminates the possibility of utilizing various ion adducts, such as cation or anion clusters, to induce the formation of higher o rder ion structures. However, further studies regarding the instrumentation and fundamentals of the DTIMS technique are required in order to take advantage of the versatility and complexity that ion adducts may provide for separation purposes. Fumaric ac id and maleic acid Neat standards of fumaric acid and maleic acid (shown in Figure 2 5), along with a mixture of these two standards, were analyzed in negative mode by drift tube ion mobility mass spectrometry (DTIMS MS) . The most prominent ion observed in the mass spectr um was the [M H] ion at m/z 115 . The ionization source , drift tube, and mass spectrometer were optimized for the analysis of smaller molecules as described in the previous section. DTIMS MS data of the maleic and fumaric acid standards were processed using the IM MS browser to calculate the CCS value for each isomer (analogous to the method described for the phthalic acid isomers) . The drift time spectrum for the two standards of the cis/trans isomers, as well as the mixture of the is omers, is plotted in Figure 2 6. The drift time plot shows separation between maleic and fumaric acid (in both the individual standards and the mixture) with R s (resolution) = 1.31. These differing drift times correlate to the following calculated CCS val ues for maleic and fumaric acid: 115.7Ã… and 121.8Ã…, respectively (Table 2 3). Also listed in the table is the CCS value for the monomeric fragment peak of the fumaric acid dimer ion, calculated for the dimer drift time at the monomeric mass. Although thi s
50 mass), this value is reproducible and can be used for monitoring this particular fragment. A CCS value could not be calculated for the maleic acid dimer because of a la ck of intensity for this peak, as shown in Figure 2 6 as the small orange peak overlapping with the fumaric acid dimer peak. Similar to the data obtained from the phthalic acid isomer studies, the data from the fumaric and maleic acid studies provided ad ditional insight into the impact of ion structure on drift time separation. These results indicate that cis and trans isomerism in a relatively small molecule is adequate structural variation for separation by drift tube ion mobility. However, the higher order dimer structures for this particular isomeric pair could not be resolved as they had identical drift times, suggesting that not all ion adducts can be used to improve separation. In addition, these data also illustrate significant differences in io n intensity between the monomer and dimer peaks for each isomer without varying any instrumental parameters. This observation suggests that different isomers may have different gas phase chemical properties that may selectively favor the monomer or dimer structure. It is also possible that the dimer of maleic acid may be fragile and fragment easily, which would alter the observed intensity ratio between the monomer and dimer ion. These behaviors are important factors to consider when processing data and, in the future, and could be used to further improve separation methods by DTIMS. Liquid C hromatography Ion Mobility Mass S pectrometry (LC IMS MS) The work presented in this section focuses on the addition of ion mobility spectrometry (IMS) to existing LC MS methodologies (LC IMS MS). It was of interest to determine whether IMS data could be collected in tandem with LC data without inducing
51 a delayed response from the MS, which could result in a mismatch of a peak at a specific retention time with its mas s spectrum. IMS data collected in tandem with LC data could be particularly advantageous. First, ion mobility adds an additional dimension of separation to chromatography methods. It is possible to separate compounds by ion mobility that could not other wise be separated by LC methods without making any changes to existing LC MS methodologies. Second, in untargeted studies IMS data can be used to confirm the identity of compounds determined by MS analysis. Using drift time data it is possible to determi ne if compounds identified by MS are endogenous in a sample or if they are fragments of larger molecules, thereby In the following experiment a sample of pooled blood plasma was analyzed by LC IMS MS method ologies. The purpose of this investigation was to demonstrate of that is, to show that IMS analysis could be performed in tandem with LC MS methods without compromising the validity or specificity of the LC MS data. For simplification, the experiment focused only on the identification of tryptophan (Figure 2 7) in the blood plasma sample. Deuterated tryptophan was spiked into the plasma as an internal standard. LC IMS MS data were collected for all peaks in the chromatogram, but only t he LC peak at 4.47 5.06 min (tryptophan) was evaluated in the IM MS browser to extract drift time ion mobility data. Tryptophan was confirmed as the identity of the selected LC peak using the LC MS extracted ion chromatogram (EIC) data for the d3 tryptoph an [M+H] + ion at m/z 208. Tryptophan and d3 tryptophan Extracted samples from the pooled blood plasma were prepared as described in the experimental section and auto injected onto the LC column. Samples were
52 analyzed in positive mode by LC IMS MS . The dat a acquired using this technique can be examined using two different methods: 1) analysis of the traditional LC total ion chromatogram (LC TIC), where all eluting compounds are viewed in order of their retention times, or 2) analysis of the ion mobility tot al ion chromatogram (IMS TIC), where all eluting compounds (and fragments of compounds) are viewed in order of their drift times. By combining the LC MS data with the IMS data, a LC retention time / IM drift time / ion intensity plot could be constructed (Figure 2 8), which visually illustrates the additional dimension of separation provided by ion mobility under each LC peak. The blue trace in Figure 2 8 is the LC TIC where the retention time of each peak is plotted vs its intensity. The orange intensit y plot displays the drift time of each compound (or fragment) vs the LC retention time. Although this graphical interpretation is vague in terms of identifying the eluted compounds in the sample, it does illustrate the presence of co eluting compounds hid den under certain LC peaks. The data in this plot can then be further analyzed by extracting the m/z information from a given area on the plot. The green rectangle in Figure 2 8 highlights the tryptophan peak in the LC chromatogram. The ion mobility dat a were extracted from this peak and expressed in Figure 2 9 as the tryptophan drift spectrum vs the corresponding m/z values for each drift peak. This plot illustrates the m/z window where the [M+H] + ions of tryptophan ( m/z 205) and d3 tryptophan ( m/z 208 ) were expected. (The mass spectrum in Figure 2 9 does not show the complete range of m/z values, but dimer and trimer peaks of tryptophan and d3 tryptophan were observed.) The orange traces are the peaks in the mass spectrum ( m/z vs intensity), and the blue intensity plot illustrates the
53 corresponding m/z values for each observed drift peak. The data in Figure 2 9 were simplified further by extracting each individual m/z value and corresponding drift time (Table 2 4) and replotting them to better illust rate several interesting patterns (Figure 2 10). The series of three data points to the far right in Figure 2 10 correspond to the homo dimer, tryptophan tryptophan ( m/z 409), the deuterated homo dimer, d3 tryptophan d3 tryptophan ( m/z 415), and the heter o dimer, tryptophan d3 tryptophan ( m/z 412). As illustrated in the plot two distinct patterns emerge. 1) The vertical association, shown in the blue boxes in Figure 2 10, represents ion lineage. All green dots within the blue box correspond to the same ion, but the ion has several drift times because it arises from fragmentation of a higher order molecule after the drift tube. For example, the longer drift times (~24 and 28 ms) associated with the dimer peaks at m/z 409, 412, and 415 resulted from fragm entation of trimer and tetramer ions, while the earlier drift time (~22 ms) corresponds to the true dimer ions. 2) The horizontal association, shown in the green box in Figure 2 10, represents the fragmentation pattern of a parent drift tube ion. For exa mple, the dimer ions of m/z 409, 412, and 415 have the same drift time as the monomer ions ( m/z 205 and 209) and the / fragment ions ( m/z 188 and 191). This indicates that the monomer ions and the fragment ions at this particular d rift time (~22 ms) are the result of fragmentation of the dimer ions after exiting the drift tube. The collision cell was not active for all data collected in this experiment, indicating that all observed fragmentation arose from normal ion mobility mass spectrometer parameters. However, when the collision cell was active similar fragmentation patterns were observed.
54 The data collected in this experiment demonstrated that ion mobility could be successfully paired with existing LC MS methodologies without affecting the quality of the LC MS data. In addition, these studies illustrated that ion mobility data could offer an additional dimension of separation as shown by the various observed drift peaks under a single LC peak. This is particularly advantageo us for molecules that cannot be separated by traditional LC methods. Finally, some of the most interesting results obtained from these studies revealed that extensive m/z fragmentation and drift time patterns could be deduced by combining the LC, IMS, and MS data in a single plot. With an improved understanding of these fragmentation patterns it may be possible to data alone by determining whether a MS peak is endo genous in a particular sample or if it was the result of fragmentation of a higher order molecule. Summary The work presented in this chapter explored the technique of drift tube ion mobility mass spectrometry (DTIMS MS) and its utility as a rapid separati on technique. Initial investigations demonstrated the need for optimization of several instrumental parameters to improve ion intensity and drift peak resolution for smaller molecules (<200 m/z ). After optimization, the structural isomers, ortho , meta , and para phthalic acid, and the cis/trans stereoisomers, maleic and fumaric acid, were explored by DTIMS MS. The data acquired from the phthalic acid studies not only demonstrated baseline separation of the monomeric phthalic acid ions, but also varying d egrees of separation between the dimer and trimer phthalic acid species. The results of the maleic and fumaric acid study revealed that cis/trans isomerism in relatively small molecules can provide an adequate degree of structural variation to induce sepa ration.
55 Findings in both isomeric separation studies indicated that higher order species of the monomeric analyte ions exist in the drift tube and have unique mobilities/drift times, which, with further fundamental studies, could be exploited for the devel opment of separation methodologies. Finally, the analysis of tryptophan in blood plasma by LC IMS MS demonstrated that ion mobility data could be paired with LC MS without negatively impacting the quality of the LC MS data. In addition, interesting fragm entation patterns could be observed by combining the LC, IMS, and MS data into a single plot. Although additional studies will be required to expand the utility of DTIMS MS as a tool for rapid separations, this work provides a solid foundation for the fun damental understanding of instrumental parameters and basic separation strategies.
56 Table 2 1. LC g radient method . Time (min) Flow (mL/min) %A %B 0 0.3 100 0 1 0.3 100 0 11 0.3 35 65 13 0.3 35 65 18 0.3 5 95 20 0.3 5 95 21 0.3 100 0 22 0.3 100 0 Figure 2 1. Phthalic acid isome rs molecules: phthalic acid (ortho phthalic acid), isophthalic acid (meta phthalic acid), and terephthalic acid (para phthalic acid) .
57 Figure 2 2. I ntensity vs drift time plot for the phthalic acid monomer s at 18. 6 V/cm . Figure 2 3. Intensity vs drift time plot for the phthalic acid dimer s at 18.6 V/cm . 0.0E+00 2.0E+04 4.0E+04 6.0E+04 8.0E+04 1.0E+05 1.2E+05 1.4E+05 1.6E+05 1.8E+05 13 13.5 14 14.5 15 Counts (AU) Drift Time (ms) m/z 165 monomer [M H] OPM0.0E+00 5.0E+03 1.0E+04 1.5E+04 2.0E+04 2.5E+04 3.0E+04 19 19.5 20 20.5 21 21.5 22 Counts (AU) Drift Time (ms) m/z 353 dimer [2M+Na 2H] OPM-
58 Figure 2 4. Intensity vs drift time plot for the phthalic acid trimer s at 18.6 V/cm . Table 2 2. Calculated CCS values for the phthalic acid isomers. Ions m/z CCS (Ã… 2 ) 0.E+00 5.E+02 1.E+03 2.E+03 2.E+03 3.E+03 3.E+03 4.E+03 4.E+03 5.E+03 23 24 25 26 27 28 Counts (AU) Drift Time (ms) m/z 541 trimer [3M+2Na 3H] OPM-
59 Figure 2 5. Fumaric acid and m aleic acid. Figure 2 6. Drift spectrum for the fumaric and mal eic acid [M H] ion s at m/z 115. The peaks to the left (before 14 ms) correspond to the true monomer ions. The peak s to the right at 17.5 ms are the result of fragmentation of the fumaric and maleic acid dimers to the monomer ions after exiting the drift tube. 0.0E+00 2.0E+04 4.0E+04 6.0E+04 8.0E+04 1.0E+05 1.2E+05 12 14 16 18 20 Counts (AU) Drift Time (ms) m/z 115 monomer [M H] Mix Maleic acid [M-H]Fumaric acid [M-H]
60 Table 2 3. Calculated CCS values F igure 2 7. Structure of tryptophan.
61 Figure 2 8. LC IMS total ion chromatogram (TIC) of pooled blood plasma spiked with d3 tryptophan (internal standard) . The blue trace is the LC TIC and the orange intensity plot is the IMS TIC. The green rectangle illustrates the tryptophan peak .
62 Figure 2 9. IMS MS data of pooled blood plasma extracted from the tryptophan peak (green rectangle in Figure 2 8 ) . Th is m/z window show s the mass spectrum produced from tryptophan and d3 tryptophan and the correspondin g drift times for each drift peak .
63 Table 2 4. Observed d rift times and resolving power calculations for each ion of tryptophan.
64 Figure 2 10. D rift times and m/z values for each tryptophan (or fragment) ion in in Table 2 4. The green box shows illustrates fragmentation of higher order molecules to smaller f ragments after exiting the drift tube (resulting in identical drift times). The blue boxes show ion lineage, where identical ion species exhibit multiple drift times . 12 14 16 18 20 22 24 26 28 150 200 250 300 350 400 Drift Time (ms) m/z Tryptophan IMS MS
65 CHAPTER 3 FUNDAMENTAL STUDIES OF HIGH FIELD ASYMMETRIC WAVEFORM ION MOBILITY MASS SP ECTROMETRY Introduction Over the last few years high field asymmetric waveform ion mobility (FAIMS) has found use in a handful of applications, typically for the separation of isomeric compounds. However, d espite its use and recent advances in instrumenta tion the fundamentals of the FAIMS technique are still not well defined or understood , making it difficult to predict separation trends and develop methodologies . In order for FAIMS to become useful in a broader range of applications, studies regarding th e fundamental principles of the separation technique must be conducted. This chapter has aimed to take a detailed look at the various factors that affect FAIMS separation, including auxiliary gas flow rate, temperature, and FAIMS atmosphere composition, i n the hopes of establishing some broad FAIMS separation principles that can aid in method development. The results of the following studies illustrate that the addition of chemical modifiers, in the form of gas and/or solvent vapors, have a significant ef fect on ion mobility behavior, and play an important role in the separation of the phthalic acid isomeric compounds, demonstrating the potential usefulness of the technique for isomeric separations of clinical significance. High Field Asymmetric Waveform Ion Mobility Mass Spectrometry (FAIMS MS) Recent advancements with in the field of ion mobility spectrometry have brought about the development of a new analytical technique called high field asymmetric waveform ion mobility spectrometry (FAIMS). Tradition al drift tube ion mobility methodologies involve the separation of ions at atmospheric or sub atmospheric conditions by applying a counter current drift gas and a uniform electric field through the
66 velocity 23 , as discussed in the previous chapter. In contrast, FAIMS applies an asymmetric waveform between two parallel electrodes at atmospheric conditions creating a differential field perpendicular to the direction of ion motion as described in Chapter 1. 16 , 17 This differential high/low electric field, or dispersion field (DF), induces an ion mobility based on the variance of 17 Under these conditions any two ions with differing net variances, or mobility coefficients, will separate. 16 Ions exposed to a dispersion fi eld (DF) fall into one of three theoretical mobilities as discussed in Chapter 1. Type A ions increase in high field mobility as the dispersion field strength increases, Type B ions initially increase in high field mobility followed by a decrease in ion m obility at higher fields, and Type C ions decrease in high field mobility as the field increases. 17,18 These mobility trends are theorized to be the result of ion interactions with the applied field and/or specifi c components of the gas phase molecules in the FAIMS atmosphere. 16 , 19 Increased mobility is induced by a greater ion to field interaction, whereas decreased mobility is induced by a greater ion to atmosphere interaction. 16 Under dispe rsion field conditions alone, only an ion with a net variance of zero will pass through the FAIMS electrodes, while an ion with a positive or negative net variance will drift towards one of the two electrodes and become neutralized. 16 To detect additional ion mobilities, a supplementary field, called the compensation field (CF), may be applied over the DF electrodes, which allows one to scan ion mobilities with net variances other than zero . 16,17 The FAIMS technique described above can be applied to a number of FAIMS cell designs including (1) a planar FAIMS electrode, composed of t wo parallel plates with an
67 applied waveform over one or both plates, (2) a cylindrical electrode, composed of an inner and outer curved plate with the waveform applied to the inner plate, and (3) a micro machined FAIMS chip (ultra FAIMS), composed of two i nterlocking parallel plates forming multiple planar channels. 24 , 5 , 25 The research presented in this c hapter focuse s on the ultra FAIMS design , which has several advantages over other cell geometries. The ultra FAIM S cell is approximately one order of magnitude smaller than conventional FAIMS electrodes, and is composed of multiple micro channels spaced closely together . T he smaller gaps in the FAIMS electrodes permit lower voltages to be used to achieve the field st rength required for separation. 5 , 25 Due to its size (Âµm scale) and low operational voltages, ultra FAIMS also allows for rapid scanning of the DF and CF for efficient detection of multiple analytes , but most importantly a comprehensive spectrum of the entire mobility field range can be acquired in a short time frame . FAIMS has been employed as a stand alone analytical method for a number of applications ; 25 however, coupling FAIMS with mass spectrometry can make it a more powerful and selective analytical technique. 26 Although FAIMS is capable of separating ions of similar size, shape, and mass to charge ( m/z ) values, it is possible for i ons with differing structures and m/z values to co elut e at the same CF/DF voltages and, therefore, require MS for identification. 16 In addition, mass spectrometry can further enhance analyte resolution/characterization through the use of MS/MS. 26 Although many compounds of clinical and industrial relevance can be resolved by traditional separation methods, FAIMS MS methodologies offer several a dvantages over these techniques by 1) reducing analysis time in comparison to other analytical techniques (e.g., liquid chromatography (LC) and/ or gas chromatography (GC)), 2)
68 5 , 11 , and 3) eliminating the need for derivatization of the analyte(s) prior to separation. 27 Given the growing demand for clinical assays, rapid sample analysis is crucial, and the use of ultra FAIMS MS methodologies could potentially reduce th is analysis time to only a few seconds for a complete CF/DF map (which would improve sample throughput by at least two orders of magnitude) , and also reduce acquisition times to a few m icro seconds for a fixed CF/DF for targeted analysis. 5 Additionally, a complex s ample can be directly injected into the ultra FAIMS ionization source, significantly reducing the amount of generated waste compared to conventional separation methodologies, such as liquid chromatography, which require relatively large volumes of mobile p hase. 11,28 Furthermore, many conventional methods of separation (e.g., LC and/or GC) require time consuming derivatization of the analyte(s) prior to separation. 27 Ultra FAIMS MS has the ability to not only reduce workflow from minutes to seconds as described above, but can also eliminate the need for derivati zation with the addition of FAIMS modifiers, effectively enhancing separation of compounds with only minor structural variations a capability typically not attributed to techniques such as LC or GC. While FAIMS MS does offer its advantages over trad itional separation methods, at the present it has relatively few applications. 5 , 11 Since it is a comparatively new technique to the fiel d of ion mobility spectrometry (approximately 20 years old), there is still much to be understood on a fundamental basis before highly effective, efficient methodologies can be developed. Because ion mobilities are highly dependent on the atmospheric cond itions (e.g., temperature, flow rate , and atmosphere composition ) in which the FAIMS electrodes are held, 16 , 26 , 29 differing atmospheric conditions are
69 capable of inducing considerable differences in ion mobilities. 16 , 17 , 19 , 30 , 31 T he introduction of FAIMS modifiers (gases and/or solvent vapors that are introduced into the FAIMS atmosphere) ha s shown significant promise for the separation of ions, espe cially in clinical applications; 19 , 29 , 31 , 32 , 33 , 34 , 35 however, an in depth understanding of specifically how FAIMS modifiers affect ion mobility is largely unknown. The purpose of the research presented in this chapter was to develop a fundamental understanding of the ultra FAIMS instrumentation as well as to construct a n apparat us that could be used to introduce various gases and solvent vapors into the FAIMS atmosphere. The effects of temperature and auxiliary gas flow rates were investigated. In addition, a number of combinations of gas and/or solvent vapor modifiers were exp lored to potentially identify any trends in ion mobilities or separation based on the composition of the modifier. The information obtained from this study is important for the development of highly effective and efficient ultra FAIMS MS methodologies , an d was applied to the development of methods for the separation of clinically relevant metabolites , as discussed in Chapter 4. Model Compounds The ultra FAIMS technique was evaluated using a common set of isomeric molecules, phthalic acid (ortho phthalic a cid), isophthalic acid (meta phthalic acid), and terephthalic acid (para phthalic acid) , shown in Figure 3 1. These compounds have been separated using the conventional planar FAIMS cell design under multiple conditions. D.A. Barnett et al. 35 separated these molecules using a 5% CO 2 and 95% N 2 atmosphere, wh ile L.C. Rorrer et al. 19 were able to achieve separation in two separate experiments: 1) with the addition of 7,000 ppm water vapor to the N 2 atmosphere and 2) with 15,000 ppm methanol vapor added to the N 2 atmosphere. The
70 use of these isomers allowed u s to directly compare the planar cell design to the micro machined FAIMS chip (ultra FAIMS). Furthermore, they were also an ideal choice as model compounds for illustrating potential mobility trends as they have already been shown to be separated by ion m obility methodologies. Experimental Reagents and Solutions Phthalic acid (ortho phthalic acid), isophthalic acid (meta phthalic acid), and terephthalic acid (para phthalic acid ) were purchased from Acros Organics. No further purification was necessary , and fresh stock solutions were prepared prior to experiments. Stock solutions were prepared in the following manner: each phthalic acid isomer was weighed in a separate volumetric flask, dissolved in 90/10 methanol/water , and diluted to a final concentra tion of 20 ppm. Sample was introduced into the ESI source at a constant flow rate of 15 ÂµL/min from the syringe pump. Instrument Parameters All ultra FAIMS methods utilized direct injection of the standard into the ESI source. All experiments were co nducted on a n Agilent 6460 triple quadrupole mass spectrometer (TQMS), equipped with an Agilent Jet Stream ESI source that was used in tandem with an Owlstone ultra FAIMS chip. The ultra FAIMS chip , composed of multiple micro channels with a gap width of 100 m and a thickness of 700 m, was powered by an Owlstone waveform generator, which generated fields within a range of 0 350 Townsends (Td) over the dispersion field (DF) and 30 30 Td over the compensation field (CF). Standard spectra were acquired u sing scan mode in order to determine ion(s) of interest for evaluation by FAIMS. The most abundant ion for the isomeric phthalic acids
71 was the [M H] ion at m/z 165 . All subsequent data were obtained in SIM mode for targeted analysis of the m/z 165 ion, reducing acquisition time. All standard solutions were directly infused into the ESI source using a syringe pump with a flow rate of 15 parameters were as follows : gas flow at 8 L/min, gas temperature at 150 C, nebulizer pr essure at 12 psi, sheath gas flow at 5 L/min, sheath gas temperature at 250 C, capillary voltage at 2500 V, nozzle voltage at 2000 V, fragmentor voltage at 50 AU, and c ell accelerator voltage at 1 AU. Results and Discussion Instrumentation Modification s N o commercial apparatus for the introduction of auxiliary gases and/or solvent vapors into the FAIMS cell atmosphere is currently available for the Owlstone ultra FAIMS system, so one was constructed in house. The instrument modifications needed to accommo date a wide variety of gases and solvents in order to complete a comprehensive study regarding the effects of the FAIMS atmosphere composition on ion mobility. A survey of the literature revealed that gases such as nitrogen (N 2 ), 26 , 33 , 35 , 11 carbon dioxide (CO 2 ), 26 , 33 , 35 , 11 oxygen (O 2 ), 26 , 33 , 11 sulfur hexafluoride (SF 6 ), 26 , 33 and h elium (He), 29 , 33 and solvents such as isopropanol ( (CH 3 ) 2 CHOH) , 36 acetonitrile (CH 3 CN), 36 ethanol (CH 3 CH 2 OH), 37 and acetone (( CH 3 ) 2 CO) , 30 , 36 were promisi ng candidates for use as FAIMS atmosphere modifiers to induce differential ion mobilities. Therefore, the apparatus was designed keeping in mind compatibility with these particular gases and solvents. To deliver accurate flow rates of each gas (or gase s), mass flow controllers from MKS Instruments (Andover, MA) were used. The mass flow controllers were attached
72 to high purity gas cylinders or, for nitrogen, to a nitrogen gas line supplied by a liquid nitrogen dewar. The flow controllers were calibrate d for their respective gases and operated at flow rates between 0 and 4 L/min (the maximum flow rate obtainable for each flow controller was 4 L/min). Copper and/or stainless steel tubing with Swagelok fittings were used for the gas lines, and the gas plu mbing was designed to connect the gas lines from the mass flow controllers to a common gas manifold, where the gases could mix before entering the FAIMS cell. This common mixing manifold is illustrated on the left in Figure 3 2. In order to regulate the addition of solvent vapor into the FAIMS atmosphere, one gas line from a mass flow controller was redirected into the top of a HPLC bottle. The gas line was connected to a sparger attachment (below the surface of the solvent), which induced aeration (bubb ling) of the solvent and produced a saturated solvent vapor headspace in the HPLC bottle. A second gas line was plumbed out of the top of the HPLC bottle so that the solvent vapor in the head space could be directed into the mixing manifold. The aeration of the solvent was maintained at a set flow rate (into and out of the HPLC bottle) using the mass flow controller, and calibration of solvent flow from the headspace into the FAIMS cell was performed for each set of solvents by measuring the flow rate of the gas and the decrease in mass of the solvent (using a mass balance illustrated on the right in Figure 3 2) over a given time. A schematic representation of the gas/solvent vapor delivery system is shown in Figure 3 3. The exit line from the mixing m anifold, carrying premixed gas and/or solvent vapors, was introduced into the heated drying gas port of the mass spectrometer. The gas and/or solvent vapors were heated to 150 C (temperature details discussed in the following
73 section). Th is heated gas/so lvent vapor was passed over the FAIMS cell, under the capillary spray shield, occupying the atmospher e of the ultra FAIMS cell before entering the gas transfer capillary and mass spectrometer . In order to maintain stable and reproducible conditions in the FAIMS atmosphere there were several other miscellaneous features of the delivery system that were considered. First, all gas lines were made as short as possible in order to reduce the delay of gas delivery into the FAIMS system, which minimized any devi ations in flow rate. Second, smaller volumes of solvent (relative to the volume of the solvent bottle) were used to maximize the headspace within the HPLC bottle, which ensured ample solvent vapor was available to be pumped into the FAIMS atmosphere . Fin ally, to ensure a fixed concentration of solvent was being introduced into the FAIMS cell over the duration of a FAIMS experiment, a mass balance was used to record the loss of solvent mass (inside the HPLC bottle) at a constant gas flow rate over a given period of time. The results of this experiment indicated a linear loss of solvent mass over time, suggesting that a constant concentration of solvent from the headspace of the HPLC bottle was being introduced into the FAIMS cell during the experiment. In order to calculate the concentration of solvent vapor in the FAIMS atmosphere it was assumed that the headspace of the solvent bottle was completely saturated. Using the vapor pressure of the solvent of choice the concentration of the solvent vapor at a g iven solvated gas flow rate and time was calculated and then divided by the total volume of gas entering the cell over the given period of time (dry nitrogen gas was added in the mixing manifold to achieve 8 total L/min of gas flow). This calculation yiel ded the total concentration of solvent vapor in the FAIMS cell. These modifications to the
74 gas/solvent delivery system optimized the quality of the data obtained during FAIMS analysis. Optimization of the CF/DF Scan Range The first step in optimizing a FAIMS method for the separation of the phthalic acid isomers required scanning of the full CF and DF ranges ( 30 to 30 Td and 125 to 300 Td, respectively) in order to ensure the complete FAIMS spectra, which included the most abundant 165 m/z ion, was capt ured during the experiment. After conducting a full CF/DF scan an optimized CF/DF range was selected with a reduced scan window to increase acquisition speed. The optimized CF scan range was set to 2 to 4 Td, with scanning of the range occurring over 3 02 individual steps. The optimized DF scan range was set to 125 to 275 Td, with scanning occurring in 66 increments. Because the number of CF/DF increments had little effect on the resolving power, this method was further refined by decreasing the number of CF steps to 90 and the number of DF steps to 16, which reduc ed the total acquis ition time to approximately 40 60 seconds. Scans of m phthalic acid were obtained using the full 20,000 steps (A) and the shortened 1,440 steps (B) and are displayed in Fig ure 3 5. This figure demonstrates a significant reduction in scan time with minimal loss in resolution. Optimization of Gas Flow Rate and Temperature In order to investigate the effects of the FAIMS atmosphere composition on the mobility and separation of ions it was important to first optimize gas flow rate and study) could be established. Possible flow rates for the mass spectrometer ranged from 5 12 L/min. Under dry N 2 conditions and without the FAIMS cell attached to the mass spectrometer flow rates in the range of 5 6 L/min yielded ion intensities in the range of
75 approximately 100,000 counts more than sufficient for MS analysis. However, after installing t he FAIMS cell (but without turning the voltages on), the ion intensity dropped more than two orders of magnitude (to around 6,000 counts) at flow rates in the 5 6 L/min range. This was due to the decreased efficiency in ion transfer with the addition of the FAIMS chip. This could be observed by buildup of sample on the front face of the FAIMS chip under the spray shield. This suggest s that the stacked parallel plates may cause gas turbulence , resulting in more sample collision s with the FAIMS cell redu c ing ion transfer into the mass spectrometer. (It is likely that the significant decrease in ion transmission is due to a flaw in the instrumental design. If a curved or rounded faced FAIMS chip, instead of a flat faced FAIMS chip, was used it would lik ely reduce gas flow disturbances before entering the parallel plates of the FAIMS cell.) Increasing the N 2 flow rate to 8 L/min (with the FAIMS cell in place) increased the ion intensity counts to around 10,000; flow rates greater than 8 L/min did not pro duce further increases in ion intensity, so 8 L/min was chosen as the optimum flow rate for FAIMS MS analysis. In addition to flow rate, the temperature of the FAIMS cell was also optimized. The possible operating temperatures ranged from 50 150 Â° C. Before the FAIMS chip was installed it was observed that at lower temperatures the ion intensity was slightly lower due to a reduction in desolvation of the ESI droplets. Once the FAIMS chip was installed this effect was observed to have an even greater i mpact on the loss in ion intensity. An increased buildup of sample on the face of the FAIMS chip was also observed. Therefore, to achieve the highest possible signal the temperature for all FAIMS experiments were conducted at 150 C.
76 While it was deter mined that higher temperatures increase ion intensity, it is also important to note that temperature is a critical component of the FAIMS atmosphere and could potentially have a strong influence on mobility behavior. The temperature parameter was reevalua ted after separation of the phthalic acids isomers was achieved by solvent vapor addition (discussed in further detail later in the chapter). An example of the potential importance of utilizing temperature as a modifying parameter is illustrated in Figur e 3 6, where a change in temperature at an already recognized mobility separation setting for methanol vapor (with regards to the phthalic acid isomers) demonstrates a completely different ion mobility behavior at a lower temperature (50 C. However, the ion intensity loss from the decrease in temperature reduces signal an additional order of magnitude, which introduces problems with signal stability, thereby hindering the usefulness of this parameter. Additional modifications to the instrumentation are necessary to further the utility of the temperature parameter in order to explore ion mobility behavior at low temperatures. Using the baseline parameters established above (dry N 2 atmosphere, 8 L/min gas flow rate, and FAIMS cell temperature of 150 Â° C), the model compounds ortho, meta, and para p h thalic acid s were analyzed by FAIMS MS. Analysis by mass spectrometry was performed using SIM mode, monitoring the m/z value of 165. The data from these experiments were extracted and plotted in a three dimensi onal DF and CF vs intensity plot for each individual isomer. These individual plots were overlaid into a comprehensive CF vs DF plot, shown in Figure 3 7A, which illustrates minimal separation between the ortho, meta, and para phthalic acid isomers. It s hould be noted that two ion traces for each isomer are observed. The traces on the left are the result of
77 the monomer species of the phthalic acid isomers and the traces on the right belong to fragmented dimer and trimer species (these species were confirm ed by mass spectrometry). Data obtained for the dimer and trimer phthalic acid species were not used in this work; additional studies regarding the formation and mobility of these species are necessary in order to exploit them for separation purposes. Alt hough the isomers could not be separated under dry N 2 conditions (and the parameters specified above), which was thought to be due to a lack of selective ion molecule interactions between the inert N 2 gas molecules and the phthalic acid isomer ions, the me thod was able to demonstrate the successful and rapid acquisition of FAIMS MS data using the CF/DF scan function. In addition, the data provided an excellent reference point for future experimentation, where data obtained utilizing the gas/solvent vapor m odifier Modifier Addition: Carbon Dioxide Using the baseline parameters established above a number of gas/solvent vapor modifiers were explored with the hopes of improving the separation between ortho, meta, and para p h thalic acid s. Initial experiments investigated the use of carbon dioxide as an auxiliary gas modifier. Carbon dioxide was expected to add a greater degree of chemical interaction between the isomeric ions a nd the CO 2 molecules in the FAIMS atmosphere because of the increased polarizability of CO 2 (in comparison to N 2 ) under applied electric field, which would, therefore, more likely induce mobility separation. The isomeric phthalic acid standards were analy zed by FAIMS MS using various ratios of carbon dioxide/nitrogen, with carbon dioxide percentages in the FAIMS atmosphere ranging from 0 to 65 percent (greater percentages of CO 2 were not explored because a minimum N 2 gas flow was required by the mass spect rometer in
78 order to operate, and this was maintained to avoid re plumbing the gas lines for the mass spectrometer). The percentage of carbon dioxide gas was adjusted using the mass flow controllers on the gas/solvent vapor delivery system, and each set of FAIMS atmospheric conditions were allowed to normalize for one minute in order to avoid collecting data based on an unstable atmospheric environment. As described above, t he mass spectrometer was operated in SIM mode m onitoring the m/z 165 [M H] ion. T he data collected for each individual isomer were plotted in a 3D CF and DV vs intensity plot and then overlaid to generate the CF vs DF plots displayed in Figure 3.7. In contrast to the lack of differential ion mobility for each phthalic acid isomer unde r standard conditions (dry N 2 only), shown in Figure 3.7A, the subsequent plots (Figure 3.7B D) illustrate a significant effect on ion mobility with the addition of CO 2 gas to the FAIMS atmosphere. As shown in plots B D, a stronger shift to greater CF val ues as the DF field increases is observed in the ion intensity traces as the percentage of CO 2 in the FAIMS atmosphere increases. Furthermore, the addition of 25% CO 2 to the FAIMS atmosphere led to differential ion mobilities that resulted in partial sepa ration of the phthalic acid isomers, as shown in Figure 3 8. This plot illustrates the conditions at which the best separation could be achieved for all CO 2 percentages and CF/DF values. The addition of carbon dioxide to the FAIMS cell atmosphere success fully altered the ion mobility behavior of the three isomeric ions; however, baseline separation using CO 2 alone could not be achieved. The major contributing factor to the lack of separation is thought to be due to the relatively short length of the FAIM S cell. The reduces the number of oscillations experienced by the ions. This limits the overall
79 differential mobility between different ions, resulting in minimal difference s in CF values and insufficient separation. Since the length of the FAIMS cell could not be altered, various solvent vapor modifiers were applied to the FAIMS cell atmosphere in order to potentially induce more extreme differences in ion mobilities, there by improving separation. Modifier Addition: Methanol According to the literature the addition of solvent vapor to the FAIMS atmosphere should result in larger shifts in CF values 19 (in comparison to auxiliary gas alone) for each isomeric compound, prod ucing the most significant separation between the isomers. Under a 100% N 2 gas atmosphere, a number of solvent vapors (in various combinations) were investigated by adding 100 mL of the solvent in a 1 L HPLC bottle and using the mass flow controllers and manifold gas lines to introduce the solvent into the FAIMS atmosphere. Each isomeric standard was analyzed by FAIMS MS using various concentrations of solvent vapors ranging from 0 to 80,000 ppm in the N 2 atmosphere. The various solvent vapor concentratio ns were achieved by changing the flow rate of N 2 gas bubbled through the solvent a s well as the flow rate of dry N 2 pumped into the FAIMS cell atmosphere . The FAIMS atmospheric conditions were allowed to normalize for one minute before data acquisition , a nd data was acquired in SIM mode monitoring the m/z 165 [M H] ion. Of the various solvents investigated the most significant changes in ion mobilities were observed using methanol. CF vs DF plots of increasing methanol concentrations (Figure 3 9A D) disp lay drastically different mobility behaviors than what was observed under CO 2 /N 2 atmospheric conditions. With increasing methanol, the ion intensity traces shift to more negative CF values as the DF field increases. Even more
80 interestingly, the o phthali c acid ion intensity trace has the most negative CF value of the isomers at low DF values, but as the DF field increases the mobility of the o phthalic acid ion shifts to more positive CF values in comparison to m and p phthalic acids. Although these res ults support the idea of chemical selectivity based on ion/atmosphere interactions, the specific cause of the observed mobility shifts is not well understood. Figure 3 10 illustrates the conditions (31,000 ppm methanol) where near baseline separation was achieved for the phthalic acid isomers. It should be noted that separation was also achieved at all concentrations of methanol above 31,000 ppm (as shown in plots c and d in Figure 3 9 the separation was observed at more negative CF values in plot d); h owever, as the concentration of methanol increased the intensity of the ion traces decreased, to the point that the traces nearly disappeared at the point of separation. The conditions of 31,000 ppm methanol and 100% dry N 2 atmosphere resulted in the most ideal combination of separation and ion intensity. Complete baseline separation of these isomers under similar conditions was expressed in literature, but using a conventional FAIMS cell (larger gaps between the parallel plates and greater cell length). Again, as described above, it is speculated that the slightly lesser degree of separation using the ultra FAIMS cell is due to the shorter length of the cell in comparison to the conventional cell. Another possible contributing factor may be the accelera ted gas flow of the ultra FAIMS MS system compared to conventional FAIMS MS systems. This accelerated gas flow of 8 L/min (in comparison to ~2 L/min) reduces the time the ion is exposed to the FAIMS waveform, resulting in smaller differences in ion mobili ty.
81 Modifier Addition: Other Modifiers In addition to methanol, other gas/solvent modifier combinations were investigated including the use of acetone, acetonitrile, ethanol, isopropanol, and water in a dry N 2 atmosphere, as well as combinations of modifi ers such as carbon dioxide/methanol and carbon dioxide/water (discussed in chapter 4) . These modifiers and combinations of modifiers resulted in significantly different ion mobilites from one modifier to the next, as well as varying degrees of separation between the isomers (but none that achieved baseline separation). These studies demonstrate the utility of the gas/solvent vapor delivery system and the substantial and meaningful effects the FAIMS atmosphere composition has on ion behavior. However, obt aining a full understanding of ion behaviors under various conditions remains an extremely complex task. Summary T he work presented in this chapter explored the utility of the new Owlstone ultra FAIMS MS instrumentation. Baseline parameters of gas flo w rate and temperature were established before investigating the effects of various gas and solvent vapor modifiers on ion mobility behavior. The isomeric compounds ortho , meta , and para phthalic acid, which have been previously studied using a convent ional FAIMS cell, were used as model compounds to demonstrate the effects of various FAIMS conditions on isomeric separation. Using the home built gas/solvent vapor delivery system it was shown that partial separation of the isomers could be achieved with a FAIMS atmosphere composition of 25% CO 2 (75% N 2 ), and near baseline separation was observed with 31,000 ppm methanol (in N 2 ).
82 In addition to the instrument modifications described above, it was also shown that the smaller ultra FAIMS MS system (in com parison to conventional FAIMS cells) could scan a complete CF/DF spectrum in under one minute without a substantial loss of resolution. This scanning mode is tremendously useful for complete CF/DF screening under various modifier conditions, which should u ltimately aid in the understanding of FAIMS ion behaviors. Although the small ultra FAIMS chip does offer rapid screening in comparison to conventional cells, its size is also one of its limitations, as it reduces the retention of the analyte in the FAIMS cell. This reduces the amplification of the differential mobilities between two different ions, minimizing separation by the ultra FAIMS MS instrumentation . In order to overcome this limitation of the ultra FAIMS cell it is necessary to find gas/solvent vapor modifier combinations that considerably enhance differential ion mobilities. Although it is clear from the studies described above that various FAIMS atmosphere conditions impacted ion behavior in different ways, it was difficult to extract any info rmation regarding mobility trends from the data, making it challenging to predict experimental outcomes based on various FAIMS modifier additions. Indeed, the studies above show the usefulness of FAIMS MS as a powerful tool for separations, but, at the pr esent, method development for this technique remains a challenging, trial and error process.
83 Figure 3 1. P hthalic acid isomeric molecules : phthalic acid (ortho phthalic acid), isophthalic acid (meta phthalic acid), and terephthalic acid (para phthalic a cid) . Figure 3 2. G as/solvent vapor delivery system for the addition of gas/solvent vapor modifiers to the FAIMS cell . A) Multi port gas manifold. Photo courtesy of author. B) HPLC solvent degassing bottle . Solvent evaporates into the headspace of the bottle and the balance is used to record the loss in solvent mass. Photo courtesy of author. A B
84 Figure 3 3. Schematic of the solvent vapor delivery system. MKS digital mass flow controllers regulate the flow of dry nitrogen and dry carbon dioxide to t he cell, as well as provid e nitrogen and carbon dioxide for bubbling through the solvent in the HPLC solvent degassing bottle . Figure 3 4. Schematic of the solvent vapor delivery system incorporated into the heated drying gas f or the mass spectrometer. T he heated drying gas is used to provide the FAIMS The yellow cylinder represent s w h ere the ultra FAIMS chip would be positioned on the capillary of the mass spectrometer.
85 Figure 3 5. CF/DF plots. A) Comprehensive CF/DF scan of the m phth alic acid [M H] ion taken in SIM mode wit h 20,000 steps in 11 minutes. B ) Rapid CF/DF scan of the same ion taken in SIM mode with 1440 steps in 48 seconds.
86 Figure 3 6 . FAIMS CF vs DF intensity plots for ortho, meta, and para phthalic acid [M H] ion s at 165 m/z at a 31,000 ppm concentration of methanol in a 100% dry N 2 atmosphere . Ortho phthalic acid is the red intensity trace, meta phthalic acid is the blue intensity trace and para phthalic acid is the green intensity trace. The plot to the left show s the ion intensity traces at 150 C and the plot to the right shows the ion behavior at 50 C.
87 Figure 3 7 . FAIMS CF vs DF intensity plots for ortho, meta, and para phthalic acid [M H] ion s at 165 m/z under dry CO 2 conditions. Ortho phthalic acid i s the red intensity trace, meta phthalic acid is the blue intensity trace , and para phthalic acid is the green intensity trace. The monomer intensity traces are on the left in each plot and the dimer traces hover over 0 to 1 Td in the CF range on the righ t . Each sequential plot going from left to right is an increasing concentration of CO 2 in the FAIMS atmosphere, going from 0 to 50 percent. A B C D
88 Figure 3 8 . FAIMS CF vs DF intensity plots for ortho, meta, and para phthalic acid [M H] ion s at 165 m/z u nder 25% dry CO 2 conditions (left plot) . Ortho phthalic acid is the red intensity trace, meta phthalic acid is the blue intensity trace and para phthalic acid is the green intensity trace. The monomer intensity traces are on the left in each plot and the dimer traces hover over 0 to 1 Td in the CF range on the right . The plot to the right is the 2D CF vs intensity plot at 255 Td DF field , which shows partial separation of the isomeric ions.
89 Figure 3 9 . FAIMS CF vs DF intensity plots for ortho, m eta, and para phthalic acid [M H] ion s at 165 m/z with various concentrations of methanol vapor in a 100% dry N 2 atmosphere . Ortho phthalic acid is the red intensity trace, meta phthalic acid is the blue intensity trace and para phthalic acid is the gree n intensity trace. The monomer intensity traces are on the left in each plot and the dimer traces hover over 0 to 1 Td in the CF range on the right . Each sequential plot going from left to right is an increasing concentration of methanol vapor in the FAIM S atmosphere, going from 0 to 60,000 ppm. A B C D
90 Figure 3 10 . FAIMS CF vs DF intensity plots for ortho, meta, and para phthalic acid [M H] ion s at 165 m/z at a 31,000 ppm concentration of methanol in a 100% dry N 2 atmosphere (left plot) . Ortho phthalic acid is the red intensity trace, meta phthalic acid is the blue intensity trace and para phthalic acid is the green intensity trace. The plot to the right is the 2D CF vs intensity plot at 255 Td DF field , which shows near baseline separation of the isomeric i ons.
91 CHAPTER 4 ION MOBILITY METHODOLOGIES FOR TARGETED METABOLOMICS Introduction The results shown in previous chapters demonstrate the utility of drift tube ion mobility mass spectrometry (DTIMS) and high field asymmetric waveform ion mobility mass spect rometry (FAIMS), both of which can be used as stand alone separation techniques, or in tandem with conventional separation approaches for overall improvement of existing analytical methodologies for separation. The work presented in this chapter is a culmi nation of the fundamentals and principles established in the earlier chapters of DTIMS and FAIMS applied to the development of ion mobility methodologies for the separation and analysis of compounds of clinical relevance. Current methods for clinical analy sis are burdened by extensive sample preparation, time demanding LC MS MS methods, and isomeric or isobaric interferences. Although many of these methods are dependable and robust, the growing demand for analysis of clinical samples requires an increase i these existing methodologies to accommodate this growing demand is not an economical or practical solution. However, improvements to the existing analytical methods may provide the necessary solution. The em ployment of ion mobility instrumentation for rapid acquisition of mobility data, with an additional separation dimension, may not only provide the necessary increase in sample throughput, but also offer additional ion mobility information not currently acq uired through conventional methodologies. The following studies demonstrate the use of ion mobility techniques that, in comparison to current analytical methodologies for targeted metabolomics, significantly
92 reduce sample preparation and/or analysis time. In addition, this work also establishes the applicability of DTIMS and FAIMS, facilitating the development of future ion mobility methodologies for clinically, as well as industrially, relevant separations. Metabolites of Interest This investigation f ocus ed on two pairs of common isomeric metabolites: (1) methylmalonic acid (MMA) and succinic acid (SA), relevant to the detection of a vitamin B 12 deficiency, and (2) fructose and glucose , compounds related to the onset of obesity. These molecules were s elected for study based on clinical demand, complexity of current conventional methods of separation, and clinical importance . Improving detection/separation methodologies for common isomeric compounds will greatly expand the field and study of metabolomi cs. It is anticipated that ion mobility will fill a void among current analytical techniques, not only in terms of reducing sample preparation and improving throughput, but by expanding upon the information obtained from experimental data, which will aid in the understanding of certain metabolic pathways. Methylmalonic a cid and s uccinic a cid The first set of isomers, methylmalonic acid (MMA) and succinic acid (SA) shown in Figur e 4 1 , are part of the metabolic pathway for cobalamin (vitamin B 12 ), wherein MMA is converted into SA through a reaction that requires vitamin B 12 . 38 In the absence o f vitamin B 12, the MMA concentration will become elevated indicating a vitamin B 12 deficiency and methymalonic acidemia. 38 , 39 , 40 Both condi tions may result in mild to severe symptoms of neuropathy, memory impairment, depression, and dementia, all of which may become permanent if the conditions are not diagnosed and treated. 38 Conventional methods for MMA de tection involve extensive LC MS MS or GC MS MS
93 assays to differentiate between SA and MMA. These methods may be carried out in o ne of two ways: 1) verifying MS MS fragmentation pathways of an MMA d 3 standard and comparing to the fragmentation pathways of an unknown sample , this method suffers from contaminating fragmentation peaks from SA, or 2) the t butyldimethylsilyl derivat ives of both MMA and SA can be separated on an LC MS method in approximately 10 minutes . 38 , 39 Both methods involve extensive sample preparation and/or analysis time, resulting in low throughput and costly met hodologies. 38 Fructose and g lucose The second s et of clinical isomers studied in this work is glucose and fructose , shown in Figure 4 2. These isomers also cu rrently require extensive LC MS or GC MS separation methodologies that are costly and time consuming (over 5 min LC methods) . 41 Due to the growing concern that fructose may be a contributing factor in the onset of obesity, lipogenesis, insulin resistance and metabolic dyslipidemia, 42 , 43 , 44 it has become increasingly necessary to look for an alternative analytical method for the separation of these compounds. Although fructose is a natur al source of sugar found in many fruits and vegetables, additional intake of high fructose c orn syrup and sucrose (glucose/ fructose ) disaccharide , added to processed foods and beverages, may have harmful effects on many metabolic pathways. 42 , 43 , 44 Although fructose rapidly metabolizes in the liver, it does not stimulate insulin (an energy regul ator), as would 42 The detection of fructose in the absence of glucose interference in blood sugar would be important for the regulation and diagnoses of fructose induced conditions.
94 Experimental Reagents and Solutions Methylmalonic acid and succinic acid were purchased from Acros Organics and Fisher Scientific, respectively . No further purificati on was necessary , and fresh stock solutions were prepared prior to experiments. Stock solutions were prepared in the following manner: MMA and SA were weighed and dissolved in 100% methanol and diluted to a final concentration of 1 ppm. D fructose , D gl ucose , and L fructose were purchased from Fisher Scientific . No further purification was necessary , and fresh stock solutions were prepared prior to experiments. Stock solutions were prepared in the following manner: D fructose, D glucose, and L fructose were weighed and dissolved in 60:40 methanol:water and diluted to a final concentration of 10 ppm. Mixtures of these standards were also prepared: (1) D fructose and D glucose, 2) D fructose and L fructose, and 3) D glucose and L fructose from the individ ual standard solutions at concentrations of 10 ppm. For FAIMS experimentation, stock solution of D fructose and D glucose standards were weighed out and dissolved in 60:40 methanol :water , and diluted to a final concentration of 1 0 ppm solutions with 1mM of ammonium nitrate . Instrument Parameters Drift Tube Ion Mobility Mass Spectrometry (DTIMS MS) An Agilent 6560 IMS QTOF instrument was use for all data acquisition, and all data were processed using Agilent IM MS Browser B.06.00. All IMS QTOF methods util ized direct injection of the analyte standards in to the ESI source. IMS studies were conducted using nitrogen as the drift tube gas with the drift tube pressure held at approximately 4 torr. Drift tube temperature was maintained at approximately 32 C. Both pressure and temperature of the drift tube were recorded for each experiment and
95 used in the calculations for analyte collision cross section (CCS). The CCS measurements were calibrated daily by evaluating the varying drift time measurements of known calibration ions over a series of different drift tube voltages, which resulted in a calibration factor. The measured CCS values are adjusted using this calibration factor, which takes into account the day to day temperature and pressure changes in the d rift tube. For a direct drift time comparison of two ions, one experimental run is conducted at a single drift tube field setting, normally 18.6 V/cm (or 1450 V as a drift tube voltage) to reduce the likelihood of peak shifting due to instrumentation condi tion (temperature and pressure) instability. All drift time data presented in this chapter are of same day experiments, and at a drift tube field of 18.6 V/cm, to reduce any drift variance due to day to day changes in temperature and pressure. All standar d solutions were directly infused into the ESI source using a syr inge pump at a flow rate of 10 and ionization source parameters were as follows : gas flow at 5 L/min, gas temperature at 325 C, nebulizer pressure at 20 psi, s heath gas flow at 8 L/ min, sheath gas temperature at 275 C, capillary voltage at 4000 V, nozzle voltage at 1000 V, and fragmentor voltage at 400 V. Instrument Parameters High Field Asymmetric Waveform Ion Mobility Mass Spectrometry (FAIMS MS) All ultra FAIMS MS methods utilized direct injection of the compound standard in to the ESI source. All experiments were conducted using an Agilent 6460 triple quadrupole mass spectrometer (TQMS), equipped with an Agilent Jet Stream ESI source , that was used in tand em with an Owlstone ultra FAIMS chip. Standard spectra were acquired using scan mode in order to determine ions of interest for evaluation by FAIMS. Once the ions of interest were determined all data were obtained in SIM mode
96 for targeted analysis, reduc ing acquisition time. The full CF and DF ranges ( 30 to 30 Td and 125 to 300 Td, respectively) were evaluated for each FAIMS gas/solvent vapor additive to ensure the complete FAIMS spectra was captured during the experiment. All standard solutions were d irectly infused into the ESI source using a syringe pump at as follows : gas flow at 5 L/min, gas temperature at 150 C, nebulizer pressure at 12 psi, sheath gas flow at 5 L/min, sheath gas te mperature at 250 C, capillary voltage at 2500 V, nozzle voltage at 2000 V, fragmentor voltage at 50 AU, and cell accelerator voltage at 1 AU . Results and Discussion The isomeric pairs methylmalonic acid (MMA) / succinic acid (SA) and fructose / glucose wer e analyzed by both DTIMS MS and FAIMS MS. The fundamental ideas and principles established in the previous chapters were taken into consideration when developing separation methods for the isomeric compounds. The current methods, described below, demonst rate FAIMS MS separation of MMA and SA, while separation of glucose and fructose is possible utilizing DTIMS MS methods. At the present, separation of MMA and SA by DTIMS MS has not been achieved, nor the separation of glucose and fructose by FAIMS MS. A lthough further research into the fundamentals of each technique may result in respective methods of separation for each set of isomers, the current work illustrates the uniqueness and applicability of each analytical technique. Experiments Using Drift T ube Ion Mobility Mass Spectrometry (DTIMS MS) Important IMS QTOF parameters considered in these experiments are displayed in Table 4 1. These values were established based on data collected from experiments
97 described in chapter 2, where it was determined that changing various voltage settings can improve intensity of the monomer ions as well as the dimer ions, with little dimer ion fragmentation. Currently, our work is concerned with the separation of monomeric ions only; dimer separation, although promis ing, is still in an exploratory phase. Additionally, because small molecules have typically been difficult to tune for on the IMS QTOF, extensive method optimization was necessary to tune the drift tube for ion transmission. It was determined that higher fragmentor voltages yield greater ion intensity for lower m/z values. In addition to tuning of the fragmentor voltage, fine tuning of the RF voltages in the ion funnels and ion trap gates played an important role in signal intensity, and more specifically , selected ion transmission and drift spectrum resolution. Low RF voltages (~60 V) were found to increase intensity for the low mass monomeric ion at the monomeric drift time, which suggests increased pre drift tube fragmentation of the dimer and a more s table monomeric ion transition, or simply increased transmission for the monomeric ion at the low RF voltage. Fine tuning of the ion trap grid voltages (shown in Table 4 1) resulted in improved monomeric ion intensity, as well as improved drift peak resol ution. This was thought to be the result of a smoother ion pulse into the drift tube, where ions are ejected from the trap in a more compact ion packet, resulting in less diffusion and improving the drift spectra. CCS measurements were obtained by collect ing drift time spectra for each ion of interest over a range of eight drift tube drift fields (18.6, 17.3, 16.0 14.7, 13.5, 12.2, 10.9, and 9.6 V/cm, synonymous with drift voltages 1450, 1350, 1250, 1150, 1050, 950, 850, and 750 V). Each spectrum was coll ected for 30 seconds, resulting in an overall experiment of 4 minutes. Drift time results collected over this field range yielded a large
98 data set, and precise CCS measurements for the ions of interest were obtained. However, to reduce data acquisition ti me, the same experiment can be conducted with as little as three different drift fields, with collection times as short as 10 seconds for each spectra, with little reduction in CCS precision. This shortened acquisition time is comparable to a typical LC pe ak width, which would allow drift tube ion mobility data to be collected in tandem with chromatographic separation data. Methylmalonic acid and succinic acid The MMA and SA standards, along with a solution of a mixture of these analytes, were analyzed in negative mode of the mass spectrometer. The most prominent ion observed in the mass spectrum was the [M H] ion at m/z 117 . The ionization source , drift tube, and mass spectrometer parameters were then optimize d for signal intensity for the [M H] ion a t the monomeric ion drift time.. After optimization for the [M H] ion at m/z 117, all standards (individual analyte and mixed analyte solutions) were analyzed by drift tube ion mobility to calculate the CCS value for each isomer. The drift time spectrum for each ion was also extracted at the 18.6 V/cm drift field and plotted in Figure 4 3. The [M H] ions for both isomers have similar drift times with peaks appearing at 13.64 ms for MMA and 13.88 ms for SA. If we compare the CCS values for both the MMA and SA ions (116.5 Ã… 2 and 118.1 Ã… 2 , respectively), it is clear there is no significant difference in the collision cross sections, which results in no separation of the ions as shown by the gray trace in Figure 4 3. This is believed to be due to the lack of rigidity in the small molecules , causing free rotation of the compounds as they travel through the drift tube, which results in similar drift times of the isomeric compounds. (In Chapter 2 it was shown that the cis/trans isomers fumaric and maleic acid could be baseline resolved, demonstrating that molecular
99 rigidity may play an important role in separation by drift tube ion mobility.) Although separation could not be achieved for the monomeric [M H] ions, it is possible that other ion adducts of thes e isomers may have potential for separation by inducing different rigid ion structures. However, it will first be necessary to gain a better understanding of the behavior of these ion adducts before the development of future separation methods. Fructos e and glucose The D fructose, L fructose, and D glucose standards, along with solutions of mixtures of these sugars, were analyzed in positive mode of the mass spectrometer. The mos t prominent ion observed in the mass spectrum was the [M+Na ] + ion at 203 m/ z , with a small secondary [2M+Na]+ peak at 383 m/z . It should be noted that sodium was not added to these, but was expected to have come from the glassware in which the standards were stored. The ionization source , drift tube, and mass spectrometer paramet ers were then optimize d for signal intensity for the most prominent [ M+Na] + ion at the monomeric ion drift time. After optimization for the [ M+Na] + ion at m/z 203, all standards (individual analyte and mixed analyte solutions) were analyzed by drift tube ion mobility to calculate the CCS value for each isomer. The drift time spectrum for each ion (individual analyte standards and mixtures) was also extracted at the 18.6 V/cm drift field and plotted in Figure 4 4. Near baseline separation was achieved fo r D fructose and D glucose ( R s = 0.88) and L fructose and D glucose (R s = 0.86). The [M+Na] + ions for all isomers , D fructose, L fructose, and D glucose (individual analyte standards), ha d different drift times with peaks appearing at 17.35 ms , 17.47 ms, and 17.94 ms, respectively. However, there was no significant difference in the CCS values (which take in to account minor fluctuations in instrument conditions) for D and L fructose, indicating
100 these compounds could not be separated , which was expected . The CCS values for D/L fructose and D glucose were 141.7 Ã… 2 and 147.0 Ã… 2 , respectively. Peaks in the 22 25 ms range are fragment ions of the [2M+Na] + ion ([2M+Na] + [M+Na] + ). Fragmentation occurs after the dimer ion has drifted through the drift tube, resulting in the longer observed drift times. As shown in Figure 4 5 the homo dimer peaks for fructose and glucose (from the individual analyte solutions of each) are baseline separated (R s = 1.52); however, in the fructose and glucose mixture a drift pea k appears between the two homo d imer peaks, which indicates the presence of a hetero dimer at the intermediate drift time. All CCS values for the individual analyte standards and mixed standards were calculated from the 203 m/z drift peaks and listed in Table 4 2. The results indicate that the isomeric sugars D glucose and D/L fructose can be separated by drift tube ion mobility methods. In addition, based on the results obtained in these experiments significant potential also exists for the separation of isomeric dimers and hybrid dimers, although additional research is needed in this area in order to refine this approach. Experiments Using High Field Asymmetric Waveform Ion Mobility Mass Spectrometry ( FAIMS MS) All FAIMS experiments were conducted on an Agilent 6460 triple quadruple mass spectrometer (TQMS). Standards were initially analyzed in scan mode of the mass spectrometer in order to select specific ions to be investigated by FAIMS. Once an ion was selected, FAIMS MS data were collected in SI M mode so that FAIMS CF/DF scans could be completed in less than one minute. The instrumental design of the FAIMS system included a gas/solvent vapor modification apparatus (discussed in the previous
101 chapter), allowing for exploration of various FAIMS cell environments that could promote separation of the isomers. M ethylmalon ic acid and succinic acid The MMA and SA standards, along with a solution of a mixture of these analytes, were analyzed in negative mode of the mass spectrometer. The most prominent i on observed in the mass spectrum was the [M H] ion at m/z 117 . The ionization source and mass spectrometer parameters were then optimized for signal intensity of the [M H] ion. An ultra FAIMS dispersion field (DF) and compensation field (CF) method was used to obtain CF/DF vs. intensity plots, which illustrate separation trends. The FAIMS method was developed by optimizing a CF/DF scan range where the ions of interest are expected to be transmitted. (The full range of the CF/DF fields are 30 30 Towns end (Td) and 125 275 Td, respectively.) After conducting a full CF/DF scan, an optimized CF/DF range is selected with a reduced scan window to increase acquisition speed. The optimized CF scan range was set to 5 to 1 Td, with scanning of the range occurr ing over 90 individual steps. The optimized DF scan range was set to 125 to 275 Td, with scanning occurring in 16 steps. These parameters allowed for a 60 second scan time while still covering the entire mobility range of the two isomers. Both MMA and SA were analyzed under standard conditions (pure N 2 atmosphere, 8 L/min gas flow rate, and 150 Â°C FAIMS cell temperature) ; however, no separation between the two isomers was observed, which was similar to the observed ion behavior described in chapter 3 for the phthalic acid isomers . Preceding experiments utilized the custom FAIMS gas/solvent vapor modifier apparatus (described in chapter 3) in order to modify the environmental conditions inside the FAIMS cell. Modifier experiments combined gas and/or solve nt vapor in a common bottle head
102 space, which was then pumped into the FAIMS cell in order to modify the FAIMS atmosphere. Flow of the gas and/or solvent vapor into the FAIMS cell is controlled by an additive gas mass flow controller, where any flow from 0 to 4 L/min of head space gas/solvent can be mixed in with dry nitrogen to achieve the operational flow parameter of 8 total L/min. During method development, head space flow was increased in a slow, step wise manner, where the increased flow rate was al lowed to equilibrate before data acquisition. The first gas modifier introduced into the FAIMS cell was dry CO 2 . Based on the behaviors of the phthalic acid isomers, it was anticipated that CO 2 , may induce shifts in the mobilities of MMA and SA that would result in separation of the isomers; however, as shown in Figure 4 6 separation could not be achieved under these conditions. Although the mobility peaks of each isomer drift to more negative CF values as the DF increases, very little separation is obser ved. It is important to note, however, that under dry CO 2 conditions both compounds had strong ion transmission intensities at high DF values, which was an improvement over standard (dry N 2 ) conditions. Further experimentation with FAIMS modifiers includ ed the use of solvent vapor, mixed solvent vapors, and the continuous addition of CO 2 since it was demonstrated that increased ion transmission at high DF values could be achieved under dry CO 2 conditions. The first solvent modifier introduced into the FA IMS cell was methanol, with the expectation, based on the behavior of the phthalic acid isomers, that it would induce large mobility changes in MMA and SA ions. However , the addition of methanol actually induced a significant loss in total ion transmissio n and no relevant variations in the ion mobilities . The addition of water as a solvent modifier resulted in similar ion behavior.
103 Acetonitrile was selected next and showed interesting intensity behaviors shown in Figure 4 7. As illustrated in the FAIMS CF/DF vs intensity plots there was an overall decrease in ion intensity as the acetonitrile concentration increased in the FAIMS cell, but more interestingly the MMA peak did not exist past 180 Td in the DF range. This same phenomenon was also observed whe n isopropanol was used as a solvent vapor, as shown in Figure 4 8. Although the cause of this behavior is not fully understood, it is thought to be the result of drastic differences in the interactions of the isomeric ions with the FAIMS atmosphere. Alth ough the isomers have the same chemical formula, the variation in their structures is enough that interactions of each ion with the FAIMS atmosphere could induce different mobility behaviors, and, in this case, cause one of the isomeric ions to become unst able and drift into the FAIMS electrode, becoming neutralized. A number of mixed solvents were also used as modifiers, with the most favorable results unexpectedly being obtained from the combination of acetonitrile and water, as shown in Figure 4 9. S ince the MMA isomer was neutralized above 180 Td in the DF range when using 100% acetonitrile as the modifier, it was interesting that the largest partial separation observed thus far was achieved with the acetonitrile/water mixture. In hopes of improving the separation of these isomers further, CO 2 was added as an additional modifier. In previous experiments, CO 2 improved the MMA ion intensity at higher DF values, and, based on previous experimental trends, it was expected that larger ion mobility differ ences would be observed at higher DF values. However, while the addition of CO 2 did increase the intensity of the MMA ion at higher DF values, it also
104 negatively impacted the ion mobility trends observed with only acetonitrile/water as the modifier and re duced the separation between MMA and SA (Figure 4 10). Since the addition of CO 2 to the acetonitrile/water modifier did not have a positive impact on the separation of MMA and SA, it was decided that another approach utilizing ld be used to isolate each ion. Using the partial separation achieved with acetonitrile/water FAIMS was used as a filter, where only a single DF/CF value, where each isomer exists separately, was selected at a time. Figure 4 11 illustrates how FAIMS can be used to isolate these isomers. As shown in the top plot in Figure 4 11 the SA isomer ion can be transmitted at DF and CF values of 205 Td and 2.2 Td, respectively, while the MMA ion is not transmitted at all. Unfortunately, MMA is the more significan t molecule of the isomeric pair because it is indicative of a vitamin B 12 deficiency. However, it is possible to detect both isomers independent of each other. Using the hop function the FAIMS system can jump back and forth between two DF/CF values. In this case MMA is transmitted at a DF value of 175 Td and a CF value of 1.4 Td, while SA is neutralized. On the other hand, SA is transmitted at a DF value of 175 Td and a CF value of 2.2 Td, while MMA is neutralized (Figure 4 11). Using this hop functio n to jump back and forth between CF values of 1.4 Td and 2.2 Td allows the detection of these isomers to be exclusive for each FAIMS DF/CF setting. Fructose and glucose D fructose and D glucose standards were analyzed in both positive and negative modes of the mass spectrometer. (The L fructose standard was not available at the time of these experiments.) The most prominent ion observed in the mass spectrum was the [M+NH 4 ] + at 198 m/z in positive mode and [M+NO 3 ] at 242 m/z in negative
105 mode . The sour ce and mass spectrometer parameters were optimized for signal intensity in both positive and negative ionization modes for signal intensity of the respective ions of interest . As was done in the optimization for MMA and SA, a preliminary full CF/DF scan w as conducted for the D fructose and D glucose standards in order to select a reduced scan window while still capturing both isomers. The optimized CF scan range was set to 5 to 1 Td, with scanning of the range occurring over 90 individual steps. The opt imized DF scan range was set to 125 to 275 Td, with scanning occurring in 16 steps. These parameters allowed for a 60 second scan time while still covering the entire mobility range of the two isomers. Both D fructose and D glucose were initially analyzed under standard conditions (pure N 2 atmosphere, 8 L/min gas flow rate, and 150Â°C FAIMS cell temperature) in both positive and negative mode , and, similar to the phthalic acids and MMA and SA, there was no separation observed under these conditions. The ad dition of CO 2 gas as a modifier also resulted in overlapping peaks for the isomeric sugars. Methanol was also used as a solvent vapor modifier and again, no separation was observed for the [M+NH 4 ] + isomer ion in positive mode or the [M+NO 3 ] in negative m ode. Water vapor was also investigated as a modifier, but after a few injections of the standards under these conditions, and again seeing no separation, a total loss in ion intensity was observed. A close inspection of the FAIMS cell revealed the presen ce of a thick caramelization layer caused by the build up of sugar after interacting with the superheated ESI Jetstream gas. In order to maintain the integrity of the FAIMS chip for future experiments the investigation of isomeric sugars was halted.
106 Su mmary The work presented in this chapter describes two examples of the separation of isomeric metabolites using ion mobility methodologies. Both drift tube ion mobility (DTIMS) and high field asymmetric waveform ion mobility (FAIMS) techniques can provide separation on the millisecond time scale. Although MMA and SA could not be separated by DTIMS, the sodium adduct ions of D/L fructose and D glucose could be nearly baseline separated, demonstrating a rapid ion mobility method for the differentiation of is omeric sugars, and the potential for the redevelopment of current analytical methods for sugar analysis. FAIMS, on the other hand, could not be used to separate the sugars, but was able to achieve separation of MMA and SA by using a isomer was observed. Although the sugar analysis by FAIMS was halted before completion, this study could be potentially be resumed with an adjustment in concentration in the standard solutions. Although these studies demonstrate separation by ion mobility there is still a large body of work to be completed regarding the fundamentals of each technique and how those fundamentals impact ion behavior. However, this research was intended to offer a glimpse into the future of metabolomics and clinical separations. Based on this work ion mobility techniques offer a promising approach to address common difficulties in cl inical analyses, as well as provide a wealth of knowledge regarding ion mobility behavior that cannot be obtained using traditional separation techniques.
107 Figure 4 1. Isomer s methylmalonic acid (MMA) and succinic a cid (SA). Figure 4 2. Isomeric sugars D glucose and D fructose Table 4 1. Important IMS parameters Category Name Value (V) Source Fragmentor 400 IM FrontFunnel High Pressure Funnel RF 60 IM FrontFunnel Trap Funnel RF 60 IM Trap Trap Entrance Grid Delta 14.5 IM Trap Trap Exit Grid 1 Delta 6 IM Trap Trap Exit Grid 2 Delta 1.5 IM RearFunnel Rear Funnel RF 60 .
108 Figure 4 3. D rift spectrum for methylmalonic and succinic acid. S eparation of the isomeric monomer ions was not achieved even with optimized IM parameters. 0.00E+00 1.00E+05 2.00E+05 3.00E+05 4.00E+05 5.00E+05 6.00E+05 7.00E+05 0.00E+00 5.00E+03 1.00E+04 1.50E+04 2.00E+04 2.50E+04 3.00E+04 3.50E+04 4.00E+04 4.50E+04 12 12.5 13 13.5 14 14.5 15 Counts (AU) Drift Time (ms) m/z 117 Monomer [M H] MMA and SA MMA SA
109 Fig ure 4 4. Drift spectrum for D g lucose, D f ructose, L f ructose, and mixtures of these sugars. Separation of these sugars was achieved as their sodiated ions. The monomer ions are the peaks with drift times between 16 19 ms , and the dimer fragment ions are between 22 25 ms. The small D glucose peak at 17.11 ms can be attributed to in source contamina nt of remaining fructose as a result of the order of experimentation. 0.00E+00 5.00E+05 1.00E+06 1.50E+06 2.00E+06 2.50E+06 3.00E+06 3.50E+06 4.00E+06 4.50E+06 5.00E+06 16 17 18 19 20 21 22 23 24 25 Counts (AU) Drift time (ms) m/z 203 Monomer [M+Na] + D-Glucose D-Fructose L-Fructose D-Fructose + D-Glucose Mix L-Fructose + D-Glucose Mix L-Fructose + D-Fructose Mix
110 Figure 4 5. Drift spectrum of the fragment ions for D g lucose, D f ructose, L f ructose, a nd mixtures of these sugars. D imer fragment ions were partially separated; however , the hetero dimer ion drifts between the two homo dimer pairs. 0.00E+00 2.00E+05 4.00E+05 6.00E+05 8.00E+05 1.00E+06 22 22.5 23 23.5 24 24.5 25 Counts (AU) Drift Time (ms) 203 m/z Fragment Ion [2M+Na] + [M+Na] + D-Glucose D-Fructose L-Fructose D-Fructose D-Glucose L-Fructose D-Glucose L-Fructose D-Fructose
111 Table 4 2. Calculated CCS values for various analytes/ions.
112 Figure 4 6. FAIMS CF vs DF i ntensity plot s for MMA and SA , [M H] ion at 117 m/z under dry CO 2 conditions. MMA is the blue /purple intensity trace and SA is the red intensity trace . T he left peak in each plot is the monomer trace , and the right peak hovering over 0 to 1 Td in the CF range is the dimer fragment ion peak. The artifact in the 25 ppm CO 2 plot at 190 Td DF and 2 Td CF is a total loss ion count for SA. This was a glitch i n the system that disrupted the FAIMS waveform causing no ions to be transmitted. 0 ppm CO 2 10 ppm CO 2 14 ppm CO 2 18 ppm CO 2 25 ppm CO 2 CF (Td) CF (Td) CF (Td) CF (Td) CF (Td)
113 Figure 4 7. FAIMS CF vs DF i ntensity plot s for MMA and SA , [M H] ion at 117 m/z under solvent vapor addition conditions with the head space containing acetonitrile. MMA i s the blue /purple intensity trace and SA is the red intensity trace . T he left peak in each plot is the monomer trace , and the right peak hovering over 0 to 1 Td in the CF range is the dimer fragment ion peak. 0 ppm ACN 5,500 ppm ACN 7,100 ppm ACN 9,300 ppm ACN 13,000 ppm ACN CF (Td) CF (Td) CF (Td) CF (Td) CF (Td)
114 Figure 4 8. FAIMS CF vs DF i ntensity plot s for MMA and SA , [M H] ion at 117 m/z under solvent vapor addition conditions with the head space containing isopropanol. MMA is the blue /purple intensity trace and SA is the red intensity trace . T he left peak in each plot is the monomer trace , and the r ight peak hovering over 0 to 1 Td in the CF range is the dimer fragment ion peak. 0 ppm ISO 3,300 ppm ISO 4,300 ppm ISO 5,600 ppm ISO 7,900 ppm ISO CF (Td) CF (Td) CF (Td) CF (Td) CF (Td)
115 Figure 4 9. FAIMS CF vs DF i ntensity plot s for MMA and SA , [M H] ion at 117 m/z under solvent vapor addition conditions with the head space containing acetonitrile and wa ter. MMA is the blue /purple intensity trace and SA is the red intensity trace . T he left peak in each plot is the monomer trace , and the right peak hovering over 0 to 1 Td in the CF range is the dimer fragment ion peak. The artifact in the 800 ppm ACN an d 1,600 ppm H 2 O plot at 200 Td DF and 1 Td CF is a total loss ion count for SA. This was a glitch in the system that disrupted the FAIMS waveform causing no ions to be transmitted. 0 ppm ACN 0 ppm H 2 O 800 ppm ACN 1,600 ppm H 2 O 1,000 ppm ACN 2,000 ppm H 2 O 1,400 ppm ACN 2,600 ppm H 2 O 1,600 ppm ACN 3,000 ppm H 2 O CF (Td) CF (Td) CF (Td) CF (Td) CF (Td)
11 6 Figure 4 10. FAIMS CF vs DF i ntensity plot s for MMA and SA , [M H] io n at 117 m/z under solvent vapor addition conditions with the head space containing acetonitrile and water and an axillary gas addition of carbon dioxide. MMA is the blue /purple intensity trace and SA is the red intensity trace . T he left peak in each plot is the monomer trace , and the right peak hovering over 0 to 1 Td in the CF range is the dimer fragment ion peak. The artifact in the 1,600 ppm AC N, 3,000 ppm H 2 O, and 10 ppm CO 2 plot at 175 Td DF and 2 and 0 Td CF is a total loss in ion count for MMA . This was a glitch in the system that disrupted the FAIMS waveform causing no ions to be transmitted. 0 ppm ACN 0 ppm H 2 O 0 ppm CO 2 800 ppm ACN 1,600 ppm H 2 O 10 ppm CO 2 1,000 ppm ACN 2,600 ppm H 2 O 10 ppm CO 2 1,400 ppm ACN 2,600 ppm H 2 O 10 ppm C O 2 1,600 ppm ACN 3,000 ppm H 2 O 10 ppm CO 2 CF (Td) CF (Td) CF (Td) CF (Td) CF (Td)
117 Figure 4 11. FAIMS CF vs intensity plot s of MMA and SA , [M H] ion at 117 m/z under solvent vapor addition conditions with the head space con taining acetonitrile and water (left) with corresponding 2D CF vs DF plot ( right ) . MMA is shown as the blue intensity trace and SA is the red intensity trace . T he left peak s in the traces are the monomer peaks and the right peaks ( hovering over 0 to 1 T d in the CF range ) are dimer fragment s . The green line in the 2D plot (corresponding to the top left plot) indicates FAIMS conditions under which only SA is present . The black line in the 2D plot (corresponding to the bottom left plot) shows partial separ ation of MMA and SA under FAIMS modifier conditions of 1,000 ppm acetonitrile and 2,000 ppm H 2 O at DF 175 Td . MMA can be isolated by the use of the FAIMS hop function at a CF value of 1.4 Td (blue dotted line) , and SA can be isolated by the use of the FAIMS hop function at a CF value of 2.2 Td (red dotted line). 1,000 ppm ACN 2,000 ppm H 2 O CF (Td)
118 CHAPTER 5 SUMMARY AND FUTURE WORK Conclusions The research presented in this dissertation focuses on the emerging field of ion mobility spectrometry (IMS) and its use, in tandem with mass spect rometry (IMS MS), for the rapid separation and detection of small isomeric compounds and metabolites. Despite the recent advancements in instrumentation, ion mobility applications are still somewhat limited due to a lack of a fundamental understanding of how certain conditions affect ion mobility behaviors. The studies described above took a detailed look at the techniques of drift tube ion mobility mass spectrometry (DTIMS MS) and high field asymmetric waveform ion mobility mass spectrometry (FAIMS MS) i n the hopes that the insight gained from these investigations would aid in the development of new and innovative methodologies for small molecule separations by ion mobility. First, an investigation of drift tube ion mobility mass spectrometry (DTIMS MS) l ed to several key findings: 1) ion transmission could be enhanced with higher fragmentor voltages and lower RF voltages for the ion funnels and 2) improved drift spectrum resolution could be achieved by fine tuning of the ion trap grid voltages and decreas ing the ion trap fill and release times. Additionally, further experimentation not only revealed that small isomeric monomer species could be baseline separated by DTIMS MS, but that dimer and trimer ion species of the molecules of interest exist in the io n mobility drift tube, demonstrating the potential that dimer and trimer species could be examined for separation when the monomeric analyte species cannot be resolved. Finally, ion mobility mass spectrometry (IMS MS) methods were paired with LC (LC IMS M
119 analytes of interest that cannot be resolved in complex samples can be separated based on their drift times by the additional IMS dimension. Following the investigation of DTIMS M S methodologies, the technique of high field asymmetric waveform ion mobilit y mass spectrometry ( FAIMS MS ) was explored. In comparison to conventional FAIMS cells it was shown that the new ultra FAIMS instrumentation could complete a full CF/DF scan in sec onds as opposed to minutes. In addition, it was demonstrated that the incorporation of gas and/or solvent vapor modifiers in the FAIMS atmosphere had a significant impact on ion mobility behavior, resulting in the separation of the phthalic acid isomeric species. However, identifying trends in mobility behavior based on the chemical properties of the FAIMS atmosphere remains a significant challenge. Finally, the fundamental understanding attained from the studies of the DTIMS MS and FAIMS MS techniques was applied to the development of methods for the separation of small molecules and metabolites of clinical significance. Standards of the isomeric pairs methylmalonic acid (MMA) and succinic acid (SA), and D/L fructose and D glucose were analyzed by DTIM S MS and FAIMS MS. Although DTIMS MS methods were unsuccessful in separating MMA and SA , it was possible to resolve sod ium adduct ions of D/L fructose from D glucose with nearly baseline separation. Additionally, FAIMS MS methods were unsuccessful in sep arating the isomeric sugars, but could between two different CF/DF values where only one of the analyte ions is present at each value.
120 The central theme of the work presented in t his dissertation focuses on the potential of ion mobility spectrometry techniques for the rapid, separation of small molecules, particularly those of clinical significance. Both DTIMS MS and FAIMS MS offer their own distinct advantages for the separation of small isomeric molecules. MS is relatively understood (i.e., molecules are separated based on their drift times, which are influenced by their chemical and physical properties), it is possible to predict the mo bility of ions based on their theoretical CCS (collision cross section) values. Ion mobilities based on FAIMS MS analysis, however, are much more difficult to predict since the effects of ion atmospheric molecule interactions on mobility are not well unde rstood. Nevertheless, the studies presented in this dissertation demonstrate the separation of a number of isomeric molecules by FAIMS atmosphere with various gas and solvent vapor modifiers. This is a di stinct advantage of the FAIMS MS technique over DTIMS MS methodologies the FAIMS MS platform permits altering the chemical composition of the FAIMS cell atmosphere in order to induce differential ion mobilities; at this time the DTIMS MS platform does no t have this capability the separation of ions can only be minimally influenced by the fine tuning of the ion trap grid voltages. These observations in ion mobility suggest that mobility behaviors may have a greater dependence on chemical interactions ra ther than the physical structure of the molecule. Future Work Future studies will continue to focus on broadening the use of ion mobility spectrometry as a rapid tool for separations, particularly for clinical applications. Further investigations should i nclude an exploration of the dimer and trimer analyte species and
121 their potential for expanding ion mobility separation strategies when monomeric analyte species cannot be resolved. Additionally, the use of cations (e.g., sodium, potassium, cesium, lithiu m, etc.) for analyte adduct formation should also be investigated as another potential separation strategy, as it would require very little sample preparation in comparison to LC derivatization methodologies. Finally, if instrument modifications are possi ble, changing the drift gas inside the drift tube of the DTIMS could potentially influence ion mobility behaviors, similar to that of the various FAIMS modifiers, enhancing the separation capabilities of the DTIMS MS platform. Overall, achieving a more in depth understanding of the fundamentals of DTIMS MS / FAIMS MS and ion mobility behaviors is vital to expanding the applications of ion mobility spectrometry.
122 LIST OF REFERENCES 1 . Borsdorf, H.; Eiceman, G. A. Applied Spectroscop y Reviews 2006 , 41 , 323 3 75. 2 . May, J. C.; McLean, J. A. Analytical Chemistry 2015 , 87 , 1422 1436. 3 . Shvartsburg, A. A.; Smith, R. D. Analytical Chemistry 2008 , 80 , 9689 9699. 4 . May, J. C.; Goodwin, C. R.; Lareau, N. M.; Leaptrot, K. L.; Morris, C. B .; Kurulugama, R. T.; Mordehai, A.; Klein, C.; Barry, W.; Darland, E.; Overney, G.; Imatani, K.; Stafford, G. C.; Fjeldsted, J. C.; McLean, J. A. Analytical Chemistry 2014 , 86 , 2107 2116. 5 . Shvartsburg, A. A.; Smith, R. D.; Wilks, A.; Koehl, A.; Ruiz Alo nso, D.; Boyle, B. Analytical Chemistry 2009 , 81 , 6489 6495. 6 . Schneider, B. B.; Covey, T. R.; Coy, S. L.; Krylov, E. V.; Nazarov, E. G. Analytical Chemistry 2010 , 82 , 1867 1880. 7 . Barnett, D. A.; Belford, M.; Dunyach, J. J.; Purves, R. W. Journal of t he American Society for Mass Spectrometry 2007 , 18 , 1653 1663. 8 . Kanu, A. B.; Dwivedi, P.; Tam, M.; Matz, L.; Hill, H. H. Journal of Mass S pectrometry 2008 , 43 , 1 22. 9 . T a raszka, J. A.; Kurulugama, R.; Sowell, R. A.; Valentine, S. J.; Koeniger, S. L.; Arnold, R. J.; Miller, D. F.; Kaufman, T. C.; Clemmer, D. E. Journal of Proteome Research 2005 , 4 , 1223 1237. 10 . Baker, E. S.; Livesay, E. A.; Orton, D. J.; Moore, R. J.; Danielson, W. F.; Prior, D. C.; Ibrahim, Y. M.; LaMarche, B. L.; Mayampurath, A. M. ; Schepmoes, A. A.; Hopkins, D. F.; Tang, K.; Smith, R. D.; Belov, M. E. Journal of Proteome Research 2010 , 9 , 997 1006. 11 . Tsai, C. W.; Yost, R. A.; Garrett, T. J. Bioanalysis 2012 , 4 , 1363 1375. 12 . Campuzano, I.; Bush, M. F.; Robinson, C. V.; Beaumon t, C.; Richardson, K.; Kim, H.; Kim, H. I. Analytical Chemistry 2012 , 84 , 1026 1033. 13 . Loeb, L. B. Basic processes of gaseous electronics ; Univ of California Press, 1955 . 14 . Baker, E.; Clowers, B.; Li, F.; Tang, K.; Tolmachev, A.; Prior, D.; Belov, M. ; Smith, R. Journal of the American Society for Mass Spectrometry 2007 , 18 , 1176 1187. 15 . Asbury, G. R.; Hill, H. H. Analytical Chemistry 2000 , 72 , 580 584.
123 16 . Buryakov, I. A.; Krylov, E. V.; Nazarov, E. G.; Rasulev, U. K. International Journal of Mass Spectrometry and Ion Processes 1993 , 128 , 143 148. 17 . Punles, R. W.; Purves, R. W.; Guevremont, R.; Day, S.; Pipich, C. W.; Matyjaszczyk, M. S. Review of Scientific Instruments 1998 , 69 , 4094. 18 . Shvartsburg, A. A.; Tang, K.; Smith, R. D. J ournal of t he Am erican Soc iety Mass Spectrom etry. 2005 , 16 , 2 12. 19 . Rorrer Iii, L. C.; Yost, R. A. International Journal of Mass Spectrometry 2011 , 300 , 173 181. 20 . Watson, J. T.; Sparkman, O. D. Introduction to mass spectrometry: instrumentation, applications, and strategies for data interpretation ; Wiley. com, 2007 . 21 . Agilent. 6400 Series Triple Quad LC/MS Concepts Guide . 22 . Agilent 6500 Series Triple Quad LC/MS Concepts Guide. 23 . Kanu, A. B., Dwivedi, P., Tam, M., Matz, L. and Hill, H. H. ( 2008 ), Ion mo bility mass spectrometry. J ournal of . Mass Spectrom etry ., 43: 1 22. doi: 10.1002/jms.1383. 24 . Beata, M. K.; ZoltÃ¡n, M. Analyst 2007 , 132 , 842 864. 25 . Wilks, A.; Hart, M.; Koehl, A.; Somerville, J.; Boyle, B.; Ruiz Alonso, D. International Journal for Ion Mobility Spectrometry . 2012 , 15 , 199 222. 26 . Guevremont, R. Journal of Chromatography A 2004 , 1058 , 3 19. 27 . Cowan, D. A.; Kicman, A. T.; Kubli Garfias, C.; Welchman, H. J. Steroids 2008 , 73 , 621 628. 28 . Beach, D. G.; Gabryelski, W. Analytical C hemistry 2011 , 83 , 9107 9113. 29 . Shvartsburg, A. A.; Danielson, W. F.; Smith, R. D. Analytical Chemistry 2010 , 82 , 2456 2462. 30 . Ross, S. K.; McDonald, G.; Marchant, S. Analyst 2008 , 133 , 602 607. 31 . Buryakov, I. A. Tech nical Phys ics Lett ers . 2007 , 3 3 , 861 864. 32 . Buryakov, I. A.; Kolomietz, Y. N.; Bolotov, A. V.; Vasin, A. I.; Vlasov, Y. N. Journal of Analytical Chemistry 2002 , 57 , 606 611. 33 . Shvartsburg, A. A.; Keqi, T.; Smith, R. D. Analytical Chemistry 2004 , 76 , 7366 7374.
124 34 . Xia, Y. Q.; Je mal, M. Analytical Chemistry 2009 , 81 , 7839 7843. 3 5 . Barnett, D. A.; Purves, R. W.; Ells, B.; Guevremont, R. Journal of Mass S pectrometry 2000 , 35 , 976 980. 36 . Schneider, B.; Nazarov, E.; Covey, T. Int ernational J ournal for Ion Mobil ity Spec trometry . 2 012 , 15 , 141 150. 37 . Buryakov, I. A. Journal of Analytical Chemistry 2011 , 66 , 1095 1102. 38 . Magera, M. J.; Helgeson, J. K.; Matern, D.; Rinaldo, P. Clinica l C hemistry 2000 , 46 , 1804 1810. 39 . Marcell, P. D.; Stabler, S. P.; Podell, E. R.; Allen, R. H . Analytical Biochemistry 1985 , 150 , 58 66. 40 . Norman, E. J.; Martelo, O.; Denton, M. D. Blood 1982 , 59 , 1128 1131. 41 . Kawasaki, T.; Akanuma, H.; Yamanouchi, T. Diabetes Care 2002 , 25 , 353 357. 42 . Basciano, H.; Federico, L.; Adeli, K. Nutrition & Met abolism 2005 , 2, 5. 43. Elliott, S. S.; Keim, N. L.; Stern, J. S.; T eff, K.; Havel, P. J. Th e American Journal of Clinical N utrition 2002 , 76 , 911 922. 44. Stanhope, K. L.; Havel, P. J. The American Journal of Clinical N utrition 2008 , 88 , 1733S 1737S.
125 BIOGRAPHICAL SKETCH Christopher Richard Beekman was born in Westwood, New Jersey , i n 1988, to Richard and Barbara Beekman. In 1991 his brother, Jonathan, was born and in 1993 his sister, Brittany, was born. Christopher spent his childhood playing basebal l and soccer on traveling teams as well as looking after his younger siblings. His interest in science came at a young age of 6, after watching a NASA special on the Discovery Channel with h is grandfather, William Cunningham. Christopher attended Westwood Senior Barre, PA. focusing on his double major in chemistry and biology. In addition to taking t he maximum course credit load every semester , Christopher also found time to work as a dir ection of Dr. Ronald Supkowski (chemistry) and Dr. Brian Mangan (biology). His research involved the analysis of mercury concentrations in Susquehanna River organisms, as well as developing a new mercury standard for thermal decomposition analysis. After , Christopher began his graduate career at the University of Florida under the direction of Dr. Richard Yost. After graduation he looks forward to a fulfilling research and development career in industry or government.