UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Development and characterization of two versions of a new single particle mass spectrometer for organic… Simpson, Emily 2009

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2010_spring_simpson_emily.pdf [ 1.71MB ]
Metadata
JSON: 24-1.0061115.json
JSON-LD: 24-1.0061115-ld.json
RDF/XML (Pretty): 24-1.0061115-rdf.xml
RDF/JSON: 24-1.0061115-rdf.json
Turtle: 24-1.0061115-turtle.txt
N-Triples: 24-1.0061115-rdf-ntriples.txt
Original Record: 24-1.0061115-source.json
Full Text
24-1.0061115-fulltext.txt
Citation
24-1.0061115.ris

Full Text

DEVELOPMENT AND CHARACTERIZATION OF TWO VERSIONS OF A NEW SINGLE PARTICLE MASS SPECTROMETER FOR ORGANIC AEROSOL ANALYSIS THAT INCORPORATE A 3D ION TRAP by EMILY ANNE SIMPSON  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Chemistry)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) December 2009  © Emily Anne Simpson, 2009  Abstract Aerosol particles are ubiquitous throughout the atmosphere and play an important role in human health, climate, and the chemistry of the atmosphere. A significant mass fraction of these particles is composed of organic species, which remain poorly characterized due to the number and diversity of species present. This thesis describes the development and characterization of two versions of a new single particle mass spectrometer with a 3D ion trap for organic aerosol studies. Version I combines CO2 laser desorption and electron impact ionization in an ion trap. Mass spectra obtained for four species are comparable to NIST EI spectra. Tandem mass spectrometry studies are also demonstrated. The effects of vaporization energy, ionization delay time, and electron pulse width on the mass spectra and fragmentation patterns are examined. The detection limit of the instrument is found to be ~1x108 molecules (350 nm diameter particle) for 2,4-dihydroxybenzoic acid. Version II integrates CO2 laser desorption and tunable VUV ionization in an ion trap and was used for a detailed study of oleyl alcohol, oleic acid and mixtures thereof. Both the degree of fragmentation in the mass spectra and the translational energy of the vaporized molecules are found to vary as a function of desorption energy in the pure particles and as a function of composition in the mixed particles. These changes can be described by the energy absorbed per particle during desorption. We show that these effects hinder the quantitative response of the instrument and have important implications for other two step laser desorption/ionization systems. The final part of this thesis presents preliminary results from atmospherically relevant particles. Mass spectra of cigarette sidestream smoke, fulvic acid, meat cooking, and ammonium bisulfate aerosols are collected using both versions of the instrument. The two step desorption/ionization process only worked for two types of aerosols, while CO2 only mass spectra were obtained for all four aerosol types. The suitability of CO2 desorption strongly depended on particle composition, which will affect the applicability of the technique to atmospherically realistic aerosols. The results also suggest that CO2 only laser desorption/ionization may be useful for field studies.  ii  Table of Contents Abstract.............................................................................................................................. ii Table of Contents ............................................................................................................. iii List of Tables .................................................................................................................... ix List of Figures.................................................................................................................... x List of Abbreviations ..................................................................................................... xvi Acknowledgements ...................................................................................................... xviii Chapter 1.  Introduction............................................................................................... 1  1.1 Aerosols .................................................................................................................. 1 1.1.1  Motivation for aerosol studies ........................................................................ 1  1.1.2  Aerosol properties and characteristics ............................................................ 2  1.1.2.1  Sources.................................................................................................... 2  1.1.2.2  Sizes ........................................................................................................ 3  1.1.2.3  Composition............................................................................................ 4  1.2 Existing instrumentation for aerosol studies........................................................... 5 1.2.1  Off-line methods ............................................................................................. 5  1.2.2  On-line methods.............................................................................................. 5  Chapter 2.  General Design and Construction of Aerosol Mass Spectrometers for  Aerosol Studies.............................................................................................................. 8 2.1 Inlet systems............................................................................................................ 8 2.1.1  Brief theory of particle motion in an inlet ...................................................... 9  2.2 Particle sizing........................................................................................................ 11 2.3 Particle vaporization and ionization...................................................................... 12 2.3.1  Single step vaporization/ionization............................................................... 12  2.3.2  Two step vaporization/ionization.................................................................. 13  2.4 Mass analysis ........................................................................................................ 13  iii  2.5 Commercial instruments ....................................................................................... 14 2.6 Motivation for this work ....................................................................................... 15 2.7 Thesis Overview ................................................................................................... 17 Chapter 3.  Design and Characterization of the Interface Common to Both  Versions of the Single Particle Ion Trap Mass Spectrometer................................. 19 3.1 Interface overview ................................................................................................ 19 3.2 Particle inlet and focusing region ......................................................................... 20 3.3 Particle sizing region............................................................................................. 21 3.4 Particle sizing region characterization .................................................................. 23 3.4.1  Calibration of sizing region .......................................................................... 24  3.4.1.1  Sizing detection efficiencies ................................................................. 26  3.5 Particle density measurements.............................................................................. 28 3.6 Conclusions........................................................................................................... 29 Chapter 4.  Real-Time Analysis of Single Particles Using Laser Desorption-  Electron Impact Ionization in an Ion Trap Mass Spectrometer (Version I)......... 30 4.1 Introduction........................................................................................................... 30 4.2 Experimental ......................................................................................................... 32 4.2.1  Particle generation ........................................................................................ 33  4.2.2  Particle analysis ............................................................................................ 34  4.2.2.1  Desorption............................................................................................. 35  4.2.2.2  Ionization .............................................................................................. 36  4.2.2.3  Mass analysis ........................................................................................ 37  4.3 Results and discussion .......................................................................................... 39 4.3.1  Mass analysis region characterization .......................................................... 39  4.3.2  Mass spectral identification of aerosols........................................................ 41  4.3.3  MSn performance .......................................................................................... 44  4.3.4  Effect of operational parameters on mass spectra......................................... 45  4.3.4.1  Dependence of total ion signal on laser desorption energy .................. 45  iv  4.3.4.2  Dependence of total ion signal on ionization delay time (time between  firing the CO2 laser and the electron pulse) .................................................. 47 4.3.4.3  Dependence of total ion signal on electron gate pulse width................ 48  4.3.4.4  Dependence of degree of fragmentation on ionization delay time ....... 49  4.3.5  Detection limit .............................................................................................. 50  4.4 Conclusions........................................................................................................... 51 Chapter 5.  One and Two Component Aerosol Studies with Laser Desorption –  Vacuum Ultraviolet Ionization in the Single Particle Ion Trap Mass Spectrometer (Version II) .................................................................................................................. 53 5.1 Introduction........................................................................................................... 53 5.2 Experimental ......................................................................................................... 56 5.2.1  Particle generation ........................................................................................ 56  5.2.2  Overview of version II of the SPITMS......................................................... 57  5.2.2.1  Desorption............................................................................................. 59  5.2.2.2  Ionization .............................................................................................. 61  5.2.2.3  Mass Analysis ....................................................................................... 62  5.2.3  Previous studies by Hanna et al. ................................................................... 63  5.3 Results................................................................................................................... 64 5.3.1  Pure oleic acid particles ................................................................................ 64  5.3.1.1  Mass spectra as a function of CO2 energy ............................................ 64  5.3.1.2  Dependence of degree of fragmentation on CO2 energy ...................... 66  5.3.1.3  Dependence of total ion signal on ionization delay time for a range of  CO2 energies ................................................................................................. 66 5.3.2  Pure oleyl alcohol particles........................................................................... 67  5.3.2.1  Mass spectra as a function of CO2 energy ............................................ 67  5.3.2.2  Dependence of degree of fragmentation on CO2 energy ...................... 69  5.3.2.3  Dependence of total ion signal on ionization delay time for a range of  CO2 energies ................................................................................................. 69 5.3.3  Two component particles of oleic acid and oleyl alcohol............................. 70  5.3.3.1  Mass spectra as a function of composition of mixed particles ............. 71  v  5.3.3.2  Dependence of degree of fragmentation on composition for mixed  particles ......................................................................................................... 73 5.3.3.3  Dependence of total ion signal on ionization delay time for a range of  compositions ................................................................................................. 74 5.4 Discussion ............................................................................................................. 75 5.4.1  Dependence of degree of fragmentation on CO2 energy and composition... 75  5.4.2  Dependence of ionization delay profiles on CO2 laser energy and particle  composition....................................................................................................... 78 5.4.3  Non-linear response of ion signal with particle composition ....................... 80  5.5 Conclusions........................................................................................................... 82 Chapter 6.  VUV Wavelength Scanning Using Version II for Analysis of Two  Component Aerosols................................................................................................... 83 6.1 Introduction........................................................................................................... 83 6.2 Experimental ......................................................................................................... 84 6.2.1  Particle generation ........................................................................................ 85  6.2.2  Particle analysis ............................................................................................ 85  6.2.2.1  Desorption............................................................................................. 85  6.2.2.2  Ionization .............................................................................................. 85  6.2.2.3  Mass Analysis ....................................................................................... 86  6.3 Results and discussion .......................................................................................... 86 6.3.1  Mass spectra for pure 1-octadecene particles ............................................... 86  6.3.2  Mass spectra for mixed oleic acid:1-octadecene particles............................ 88  6.3.3  1-octadecene appearance energy in pure particles........................................ 89  6.3.4  1-octadecene appearance energy in mixed particles..................................... 91  6.4 Conclusions........................................................................................................... 95 Chapter 7.  Studies of Complex and/or Atmospherically Realistic Aerosols Using  Version I and Version II of the SPITMS .................................................................. 97 7.1 Introduction........................................................................................................... 97 7.2 Experimental ......................................................................................................... 99  vi  7.2.1  Particle generation ........................................................................................ 99  7.2.2  Particle analysis .......................................................................................... 100  7.3 Results and discussion ........................................................................................ 101 7.3.1  Ammonium bisulfate .................................................................................. 101  7.3.1.1  CO2 only of ammonium bisulfate particles......................................... 101  7.3.1.2  CO2/EI of ammonium bisulfate particles............................................ 103  7.3.2  Suwannee River fulvic acid (SRFA)........................................................... 104  7.3.2.1  CO2 only of SRFA particles................................................................ 104  7.3.2.2  CO2 /EI and CO2/VUV of SRFA particles.......................................... 108  7.3.3  Cigarette sidestream smoke (SSS) .............................................................. 109  7.3.3.1  CO2 only of SSS particles ................................................................... 109  7.3.3.2  CO2/EI of SSS particles ...................................................................... 110  7.3.3.3  CO2/VUV of SSS particles ................................................................. 111  7.3.4  Meat cooking aerosols ................................................................................ 112  7.3.4.1  CO2 only of meat cooking aerosols .................................................... 112  7.3.4.2  CO2/EI of meat cooking aerosols........................................................ 113  7.4 Conclusions......................................................................................................... 115 Chapter 8.  Concluding Remarks ............................................................................ 117  8.1 Summary of findings........................................................................................... 117 8.1.1  Version I...................................................................................................... 117  8.1.2  Comparison of Version I to existing aerosol mass spectrometers .............. 118  8.1.3  Version II .................................................................................................... 119  8.1.4  Comparison of Version II to existing aerosol mass spectrometers............. 120  8.2 Future directions ................................................................................................. 122 8.2.1  Version I future work.................................................................................. 122  8.2.2  Version II future work ................................................................................ 123  References...................................................................................................................... 125 Appendix I.  Determination of particle density using a DMA coupled to the  SPITMS interface ..................................................................................................... 140  vii  Appendix II.  Determination of oleyl alcohol ionization appearance energy...... 143  Appendix III.  Calculation of energy absorbed per particle during desorption .. 145  viii  List of Tables Table 3.1: Measured PSL size distribution widths with the SPITMS (about 1000 particles per size) compared to the manufacturer’s stated distribution widths........................ 25 Table 3.2: Summary of particle densities determined by simultaneous measurement of mobility diameter (with a DMA) and vacuum aerodynamic diameter (SPITMS) and comparison to literature values. ................................................................................ 28 Table 5.1: Relevant properties of oleic acid and oleyl alcohol......................................... 57 Table 6.1: Relevant properties of oleic acid and 1-octadecene. ....................................... 84  ix  List of Figures Figure 1.1: Average aerosol composition of fine particles (by mass) based on a literature survey by Heintzenberg18 of several field studies....................................................... 4 Figure 1.2: Schematic of aerosol mixing states for (a) an externally mixed system and (b) an internally mixed system. Each circle represents an aerosol particle and each color represents a different chemical species....................................................................... 7 Figure 2.1: General schematic of an aerosol mass spectrometer. ....................................... 8 Figure 2.2: Schematic of inlet designs for particle focusing into vacuum (a) orifice, (b) capillary, and (c) aerodynamic lens system. ............................................................... 9 Figure 2.3: Schematic of particle trajectories through the aerodynamic lens (5 apertures and nozzle) showing the focusing of particles along the centerline. R is the distance from the axis of the lens, X is the distance along the lens axis. Adapted from Zhang et al.68 ........................................................................................................................ 11 Figure 3.1: Schematic of the single particle ion trap mass spectrometer (SPITMS) interface..................................................................................................................... 19 Figure 3.2: Picture of disassembled aerodynamic lens components: (above) aerodynamic lens mount, (below) aerodynamic lens orifices, spacers, and nozzle. ...................... 21 Figure 3.3: PSL calibration curves for particle velocity versus (calculated) vacuum aerodynamic diameter for the one and two laser sizing setups................................. 25 Figure 3.4: Detection efficiency of the SPITMS particle sizing region for PSLs with the one laser and two laser sizing setups. ....................................................................... 26 Figure 4.1: Schematic of SPITMS version I (laser desorption-electron impact ionization). ................................................................................................................................... 33 Figure 4.2: Chemical structures and molecular weights (g/mol) of 2,4-dihydroxybenzoic acid, caffeine, oleic acid, and linoleic acid. .............................................................. 34 Figure 4.3: Timing schematic for CO2 laser firing, electron pulse from the electron gun and the RF voltage scan for the ion trap. Ionization delay times are referenced to the x  CO2 firing. At 0 µs delay, the electron gun gate opens as the CO2 laser is simultaneously discharging. The break in the x-axis scale corresponds to the cooling time used for ions in the trap (10 ms). ...................................................................... 37 Figure 4.4: Spatial profile of MS hit rates for CO2 only ionization of 970 nm DHB aerosols with 22 mJ/pulse (~5x107 W/cm2) at 978 cm-1. By varying the delay between the desorption laser and the aerosol trigger (top X-axis) the aerosols are exposed to different segments of the spatial CO2 power profile, yielding a beam profile (bottom X-Axis). The experimental data could be fitted to a simple Gaussian profile, as expected for a single mode TM00 laser. ................................................... 40 Figure 4.5: Normalized mass spectra of (a) single particle and (b) 200 particle average of 2,4-dihydroxybenzoic acid (879 nm diameter) collected with 14 mJ 944 cm-1 CO2/EI, 1 µs ionization delay and (c) NIST EI standard. ......................................... 41 Figure 4.6: Normalized mass spectra of (a) single particle and (b) 200 particle average of caffeine (887 nm particle) collected with 30 mJ 944 cm-1 CO2/EI, 1 µs ionization delay and (c) NIST EI standard................................................................................. 42 Figure 4.7: Normalized mass spectra of (a) single particle and (b) 200 particle average of oleic acid (1.1 µm particle) collected with 15 mJ 944 cm-1 CO2/EI, 1 µs ionization delay and (c) NIST EI standard................................................................................. 43 Figure 4.8: A sequence of MS/MS studies on caffeine aerosols at 30 mJ CO2/EI and 1 µs ionization delay. (a) caffeine aerosol MS, (b) isolation of parent ion (m/z = 194), (c) fragmentation of parent ion via collision-induced dissociation, (d) subsequent fragmentation of daughter ion (m/z = 193)............................................................... 45 Figure 4.9: Total ion signal as a function of CO2 energy (944 cm-1) for (a) 2,4dihydroxybenzoic acid [m/z 154, 136, 108, 95, 80, 69, 64, 53] and (b) caffeine [m/z 194, 193, 109, 82, 81, 67, 55]. .................................................................................. 46 Figure 4.10: Total ion signal [m/z 154, 136, 108, 80, 69] versus the ionization delay time (time between the desorption laser firing and the electron pulse) (14 mJ 944 cm-1) for 2,4-dihydroxybenzoic acid of two different particle sizes (dve shown)............... 47  xi  Figure 4.11: Effect of EI gate width on ion signal (m/z 136) as a function of ionization delay time for DHB. Also shown is a high resolution ionization delay time profile obtained using CO2/VUV of DHB.166 The VUV profile has been adjusted in time so that maximum signal overlaps with the maximum of the EI profile. ....................... 49 Figure 4.12: Ratio of molecular or fragment ion signal to total ion signal for caffeine as a function of ionization delay time at 35 mJ 944 cm-1 CO2/EI. Error bars represent 1σ. ................................................................................................................................... 50 Figure 5.1: Chemical structures of oleic acid (left) and oleyl alcohol (right)................... 57 Figure 5.2: Instrument schematic for version II (CO2/VUV) of the SPITMS. The inset shows the paths of the particle beam and the IR and VUV lasers through the ion trap.166 ....................................................................................................................... 59 Figure 5.3: Timing schematic for CO2 laser firing, VUV laser firing, and the RF voltage scan for the ion trap. At 0 µs delay, the VUV fires simultaneously with the CO2 laser discharge. .................................................................................................................. 60 Figure 5.4: Schematic of resonance enhanced four wave difference mixing in Xe. ........ 61 Figure 5.5: Normalized mass spectra of oleic acid aerosols (300 shot average) ionized with (top) 7 mJ 944cm-1 CO2 142 nm VUV and (bottom) 10 mJ 944 cm-1 CO2 EI. Adapted from Hanna et al.166 .................................................................................... 64 Figure 5.6: Normalized mass spectra (300 shot average) of oleic acid at 7, 10, 14, 17 and 24 mJ/pulse of CO2 (1056 cm-1) and 142 nm (8.75 eV) VUV.................................. 65 Figure 5.7: Relative intensities of fragment peaks compared to the total ion signal as a function of desorption energy for pure oleic acid particles. ..................................... 66 Figure 5.8: Total ion signal as a function of ionization delay time for oleic acid. Each trace is for a different CO2 pulse energy (at 1056 cm-1) as indicated in the legend (VUV at 142 nm). All traces have been normalized to make comparison easier. .... 67 Figure 5.9: Normalized mass spectra (300 shot average) for oleyl alcohol at 4, 7, 10, 20, and 30 mJ/pulse CO2 (1056 cm-1) and 142 nm (8.75 eV) VUV. .............................. 68  xii  Figure 5.10: Relative intensities of fragment peaks compared to the total ion signal as a function of desorption energy for pure oleyl alcohol particles. ................................ 69 Figure 5.11: Total ion signal as a function of ionization delay time for oleyl alcohol. Each trace is for a different CO2 pulse energy (at 1056 cm-1) as indicated in the legend. All traces have been normalized to make comparison easier. .................................. 70 Figure 5.12: Normalized mass spectra (300 shot average) at the optimum ionization delay time (maximum total ion signal) as a function of particle composition (7 mJ 1056 cm-1 CO2, 142 nm VUV). ......................................................................................... 72 Figure 5.13: Fragmentation ratio of oleic acid (selected peaks relative to m/z 148) as a function of weight percent oleic acid........................................................................ 73 Figure 5.14: Total ion signal as a function of ionization delay time for mixed oleic acid: oleyl alcohol particles. .............................................................................................. 74 Figure 5.15: (a) Ionization delay time profile full width at half maximum (FWHM) versus weight percent oleic acid. (b) Delay time at maximum total ion signal versus weight percent oleic acid. Error bars reflect the resolution of the scan................................ 75 Figure 5.16: Fragmentation ratios for oleic acid as a function of IR energy absorbed during vaporization. Closed symbols show pure oleic acid particle data, open symbols show mixed particle data. ........................................................................... 78 Figure 5.17: (Top) Ionization delay profile FWHM for both pure oleic acid particles and mixed particles as a function of CO2 energy absorbed/particle and (bottom) delay time at maximum total ion signal as a function of CO2 energy absorbed/particle.... 80 Figure 5.18: Relative intensity of oleic acid fragment ion signal to total ion signal for all peaks (oleic and oleyl) as a function of the oleic acid mole ratio............................. 81 Figure 6.1: Chemical structures of oleic acid (top) and 1-octadecene (bottom)............... 84 Figure 6.2: Mass spectra (300 shot average) for 1-octadecene particles at discrete VUV wavelengths (30 mJ 941 cm-1 CO2). ......................................................................... 87 Figure 6.3: Mass spectra (200 shot average) from VUV energy scan of 50:50wt% oleic acid:1-octadecene particles with 6 mJ 944 cm-1 CO2 desorption.............................. 88  xiii  Figure 6.4: Photoionization efficiency curves for 1-octadecene fragments from pure 1octadecene particles (30 mJ 941 cm-1 CO2). The average extrapolated appearance energy is 9.42 ± 0.03 eV. .......................................................................................... 90 Figure 6.5: Photoionization efficiency curves for 1-octadecene fragment at m/z 125: (top) pure 1-octadecene (30 mJ 941 cm-1 CO2), (middle) mixed oleic acid:1-octadecene (6 mJ 944 cm-1 CO2) and (bottom) mixed oleic acid:1-octadecene (16 mJ 944 cm-1 CO2). ......................................................................................................................... 92 Figure 6.6: Photoionization efficiency curves for 1-octadecene fragment at m/z 111: (top) pure 1-octadecene (30 mJ 941 cm-1 CO2), (middle) mixed oleic acid:1-octadecene (6 mJ 944 cm-1 CO2) and (bottom) mixed oleic acid:1-octadecene (16 mJ 944 cm-1 CO2). ......................................................................................................................... 93 Figure 6.7: Photoionization efficiency curves for 1-octadecene fragment at m/z 97: (top) pure 1-octadecene (30 mJ 941 cm-1 CO2), (middle) mixed oleic acid:1-octadecene (6 mJ 944 cm-1 CO2) and (bottom) mixed oleic acid:1-octadecene (16 mJ 944 cm-1 CO2). ......................................................................................................................... 94 Figure 6.8: Photoionization efficiency curves for 1-octadecene fragment at m/z 83: (top) pure 1-octadecene (30 mJ 941 cm-1 CO2), (middle) mixed oleic acid:1-octadecene (6 mJ 944 cm-1 CO2) and (bottom) mixed oleic acid:1-octadecene (16 mJ 944 cm-1 CO2). ......................................................................................................................... 95 Figure 7.1: Mass spectra (400 shot average) of 400 nm ammonium bisulfate particles for 5 - 25 mJ 1037 cm-1 CO2......................................................................................... 102 Figure 7.2: Mass spectrum (400 shot average) of 400 nm ammonium bisulfate particles with 25 mJ 1037 cm-1 CO2...................................................................................... 103 Figure 7.3: IR absorbance spectrum of SRFA Standard I.206 ......................................... 105 Figure 7.4: Proposed average structural model of Suwannee River fulvic acid.209 ........ 105 Figure 7.5: Mass spectra (500 shot average) for Suwannee River fulvic acid aerosols for 6 - 40 mJ 944 cm-1 CO2. The inset in the bottom two panels is shown to provide a comparison of the signal size for peaks between m/z 38-44 relative to the rest of the mass spectrum. ........................................................................................................ 106 xiv  Figure 7.6: Mass spectrum (500 shot average) of 500 nm SRFA aerosols with 40 mJ 944 cm-1 CO2.................................................................................................................. 107 Figure 7.7: Mass spectra (300 shot average) of 700 nm cigarette SSS particles for 10 – 40 mJ/pulse 944 cm-1 CO2. .......................................................................................... 110 Figure 7.8: Mass spectra of 700 nm cigarette smoke particles (300 shot average) with (a) 13 mJ 944 cm-1 CO2, (b) 13 mJ 944 cm-1 CO2/EI, and (c) 13 mJ 944 cm-1 CO2/VUV (142 nm).................................................................................................................. 111 Figure 7.9: Mass spectra (200 shot average) of 950 nm meat cooking particles for 5 - 35 mJ/pulse 944 cm-1 CO2............................................................................................ 113 Figure 7.10: Mass spectra (200 shot average) of 950 nm meat cooking particles with (a) 20 mJ CO2 only and (b) CO2/EI at a desorption energy of 20 mJ. ......................... 114 Figure A 1: Density determination for canola oil particles from measured particle velocity versus mobility diameter......................................................................................... 141 Figure A 2: Density determination for sodium bisulfate particles from measured particle velocity versus mobility diameter. .......................................................................... 142 Figure A 3: Photoionization efficiency curves for oleyl alcohol fragments (15 mJ 944 cm1  CO2, 5 µs ionization delay time, 1 ms cool time). The extrapolated appearance  energy is 8.56 ± 0.05 eV. Data is shown with a 50 point smooth applied to each curve........................................................................................................................ 144  xv  List of Abbreviations a.u.  Arbitrary units  AC  CI  Alternating current Aerosol mass spectrometer (commercially available instrument made by Aerodyne) Aerosol time-of-flight mass spectrometer (commercially available instrument made by TSI Inc.) Chemical ionization  CIMS  Chemical ionization mass spectrometry  CID  Collision-induced dissociation  COA  Constant output atomizer  CPC  Condensation particle counter  cw  Continuous wave  dm  Mobility diameter  dp  Physical diameter  dva  Vacuum aerodynamic diameter  dve  Volume equivalent diameter  DHB  2,4-dihydroxybenzoic acid  DMA  Differential mobility analyzer  EI  Electron impact  FPGA  Field programmable gate array  FWHM  Full width at half maximum  GC/MS  Gas chromatography/mass spectrometry  HULIS  Humic like substances  IE  Ionization energy  IR  Infrared  L2MS  Laser desorption/laser ionization (two step process)  LC/MS  Liquid chromatography/mass spectrometry  LDI  Laser desorption/ionization (one step process)  LPM  Liters per minute  AMS ATOFMS  xvi  m/z  Mass to charge ratio  MALDI  Matrix assisted laser desorption ionization  MS  Mass spectrometer (or mass spectrometry)  MS/MS  Tandem mass spectrometry  OPO  Optical parametric oscillator  PAH  Polycyclic aromatic hydrocarbon  PERCI  Photoelectron resonance chemical ionization  PIE  Photoionization efficiency  PMT  Photomultiplier tube  PSL  Polystyrene latex spheres  QMS  Quadrupole mass spectrometer  REMPI  Resonance enhanced multiphoton ionization  RF  Radio frequency  S/N  Signal to noise ratio  SOA  Secondary organic aerosol  SPI  Single photon ionization  SPITMS  Single particle ion trap mass spectrometer  SPMS  Single particle mass spectrometry  SRFA  Suwannee river fulvic acid  SSS  Sidestream smoke  SWIFT  Stored waveform inverse Fourier transform  TEA  Transversely excited atmospheric pressure  TOF  Time of flight  UV  Ultraviolet  VOAG  Vibrating orifice aerosol generator  VUV  Vacuum ultraviolet  WSOC  Water soluble organic carbon  xvii  Acknowledgements This body of work would not have been possible without the contributions of many people to whom I would like to express my gratitude for all their assistance over the years. First of all, thanks to Dr. Allan Bertram for his mentorship and continual encouragement. Allan, your focus on the overall picture and goals of this research project was invaluable in the context of dealing with the daily successes and failures of the instrument. You have also created a truly collegial environment in your lab that is inspiring to be a part of. Thanks also to Dr. Michael Blades and Dr. John Hepburn, who offered their expertise and support in each of their areas related to this thesis. I was fortunate to share the experience of this project with Dr. Pedro Campuzano Jost and (now Dr.) Sarah Hanna. Pedro, thank you for all the aerosol and atmospheric chemistry expertise you brought to this project, for answering my endless questions, and always remaining calm in the face of seemingly insurmountable instrument difficulties. Sarah, you were in this project from the beginning and completely understand the joys and sorrows of building a new analytical instrument from scratch. I could not have asked for a better co-worker in this enterprise. Thanks for always being willing to lend a helping hand and coaxing the laser system into operating again. A big thank you to the mechanical and electrical engineering shops in the UBC chemistry department without whom this instrument would never have been built. Thanks for all the design advice, emergency repairs, and the endless supply of o-rings. To all the Bertram group members with whom I have been privileged to share an office/lab/supervisor with over the years, thanks for your wonderful camaraderie: Lori, Magda, Matt, Jackson, Mike, Ben, Daniel, Sebastien, Michael, Aidan, Donna, Richard, Song, and Jason. It has been a pleasure to work with you and even more so for all the enjoyable time we’ve spent together outside of work from kayaking to Frisbee to potluck dinners. To Brian Rempel, Sara Van Rooy, and Amanda Richer – you three will always be my “first” lab group. Thanks for being such a great inspiration in the pursuit of both chemistry and the rest of life. To my parents, who have encouraged my education every step of the way and have always supported me, there are not enough thank you’s in a single lifetime for everything you’ve given me. I hope you know how proud and incredibly blessed I am to have you as my parents. Lastly, but certainly not least, to Robyn, thank you for all your support, encouragement, and wholehearted belief in me. By far, the most invaluable part of this entire degree was meeting you, which I wouldn’t trade for the world.  xviii  Chapter 1.  Introduction  1.1 Aerosols 1.1.1 Motivation for aerosol studies Aerosols are defined as a relatively stable suspension of liquid or solid particles in a gas phase medium. Familiar examples include clouds, smoke and fog. Aerosol particles span a diverse range of sizes, number concentrations, and chemical compositions. Likewise, they stem from a variety of anthropogenic (man-made) and natural sources. The physical and chemical properties of an aerosol tend to correlate with the aerosol’s source. Particles are present in all outdoor and indoor atmospheres. In the lower atmosphere (troposphere), particle concentrations range from ~102 to 105 particles/cm3.1 Aerosols play a major role in public health, visibility, climate, and chemistry of the atmosphere. Aerosol effects on human health are well documented. Early studies of severe air pollution episodes in Europe and the U.S. from the 1930's to 1950's correlated exposure to high ambient levels of urban air pollution with increases in daily mortality rates.2 Particles with diameters < 1 µm are able to penetrate deeply into the respiratory passages and produce a range of adverse human health effects, including allergic, infectious, respiratory, and cardiovascular diseases.1, 3-5 Aerosols lead to reduced visibility in the atmosphere by scattering solar radiation. In remote areas, visibility extends up to hundreds of kilometers. In urban environments, the aerosol concentration can reduce visibility by an order of magnitude compared to unpolluted conditions.6 Smaller particles (0.1 - 1 µm) are the most effective at light scattering in the atmosphere since their size is of the same order as the wavelength of solar radiation.1 Aerosols impact the global radiation budget by interactions with both solar and terrestrial radiation. The aerosol direct effect describes the scattering and absorption of radiation by aerosol particles, which is dependent on particle size, structure and chemical composition. Particles can scatter and absorb solar radiation, as well as terrestrial  1  radiation. The alteration of the amount of solar radiation reaching or leaving the earth’s surface significantly impacts global climate.7 Aerosol indirect effects result from a particle’s ability to act as a cloud condensation nuclei or ice nuclei and thus modify the radiative properties of clouds, the conditions required for cloud formation, and cloud lifetimes.1, 7 Furthermore, aerosols have a significant role in the chemistry of the atmosphere. Reactions that do not occur in the gas phase can take place on the surfaces of particles.8 For example, aerosols can impact the oxidation of SO2 to sulfuric acid in the atmosphere through two mechanisms: (1) adsorption to the solid surface of a particle, where SO2 can undergo reactions with hydroxyl radicals present in the atmosphere and (2) absorption of SO2 into an aqueous aerosol (e.g. cloud droplet), where it is oxidized to sulfate.6, 9 In this way, aerosols affect the abundance and partitioning of atmospheric trace gases via heterogeneous chemical reactions and other multiphase processes.1 As a result of these effects, understanding aerosol sources, aging processes, lifetimes, and especially aerosol physical and chemical properties is essential to improving our understanding of the role of aerosols in the earth’s climate and chemistry of the atmosphere as well as their impact on human lives.  1.1.2 Aerosol properties and characteristics 1.1.2.1 Sources Atmospheric particles originate from a diverse collection of anthropogenic and natural sources. Natural sources include biomass burning (forest fires), volcanoes, sea spray, biological sources (e.g. plant fragments, pollen, microorganisms), and dust storms.1, 10 Anthropogenic sources include biomass burning (e.g. field clearing and woodburning stoves); incomplete fossil fuel combustion; industrial dust emissions (fly ash); and traffic-related suspension of road, soil, and mineral dust.1, 10, 11 Additionally, cigarette smoking and meat cooking operations are examples of small scale sources of anthropogenic aerosols.12, 13 Aerosols are formed through one of two processes. Primary particles are emitted directly as liquids or solids into the atmosphere. Secondary particles are formed by gas2  to-particle conversion. This entails new particle formation by nucleation and condensation of gas phase precursors.10 One example would be the oxidation of a gas phase species to produce a highly oxidized, less volatile species, which can then form new particles or condense on existing surfaces.14  1.1.2.2 Sizes Aerosol particles range in size from approximately 0.002 - 100 µm in diameter.1 With regard to atmospheric chemistry and physics, particles in the 0.002 - 10 µm range are the most important due to their relatively long life time, high percentage of the total number of aerosols, and the large amount of surface area they provide for heterogeneous reactions.1,  15  Atmospheric particles can be divided into two classes: coarse particles  (diameters > 2.5 µm) and fine particles (diameters < 2.5 µm). Coarse particles are produced through mechanical processes like erosion, sea spray, wind action (suspension of dust and sand), volcanic eruptions, and industrial processes that produce fly ash and other large particles. This class also includes the majority of biological particles, spores, and pollen. Coarse particles have shorter atmospheric lifetimes (generally on the order of a day) than fine particles due to their size.15 They are removed relatively quickly by gravitational settling or washout, but can at times be transported over long distances by convective processes.1 The fine particle range includes ultrafine particles (< 0.01 µm), the Aitken nuclei range (0.01 – 0.08 µm), and the accumulation range (0.08 - 2 µm).1 The Aitken nuclei mode arises from gas-to-particle conversion and combustion processes where hot, supersaturated vapors are formed and undergo subsequent condensation. Their lifetime is fairly short due to rapid coagulation.1 Particles in the accumulation range arise from low volatility vapor condensation, coagulation of smaller particles in the Aitken nuclei range, or cloud droplet evaporation. The accumulation range has a relatively long lifetime of days to weeks.15  3  1.1.2.3 Composition The chemical composition of aerosol particles, similar to their size, tends to correlate to the source and formation process of the particles.1 Coarse particles, which are formed primarily by mechanical processes, typically contain inorganic species such as sand, sea salt, and soil-related elements. Accumulation mode particles often contain combustion, condensation, and coagulation products like carbon, nitrates, sulfates, and polar organics along with trace metals.1 Once formed, aerosol particles can also undergo various physical and chemical transformations (atmospheric aging) which can alter their size, structure, and chemical composition. Aerosol particles contain both inorganic and organic species, whose distribution will vary with source as shown in Figure 1.1. The inorganic fraction of aerosols is dominated by NH4+, NO3-, and SO42-.1 Organic compounds are also significant components of tropospheric particles. In a review by Kanakidou et al., organic species represented 20 - 60% of the total fine aerosol mass on average and in some cases ranged up to 90% of the total mass.16 In an investigation of particle composition by location by Zhang et al., the organic fraction ranged from 18 - 70% of the particle mass for aerosols sampled at 30 different sites including remote, continental, and urban locations.17  SO 4  2-  NO 3 NH 4  -  +  C (organic) C (elemental) Other 31%  24%  9%  0.3%  5%  11%  11% 19% 18% 4%  8% 6% 28%  Urban  56%  7% 3%  37%  Remote Continental  22%  Remote Marine  Figure 1.1: Average aerosol composition of fine particles (by mass) based on a literature survey by Heintzenberg18 of several field studies. 4  1.2 Existing instrumentation for aerosol studies The variety in chemical composition and size of atmospheric aerosols has led to a plethora of techniques for the study of aerosol properties and interactions. The organic component of aerosols has remained especially challenging to characterize due to the hundreds of species present19 and the range of chemical and physical properties contained therein.20 Techniques for aerosol analysis can be divided into two basic categories: offline and on-line methods. Off-line techniques involve deposition of particles onto a substrate (usually a filter) over a long period of time and later analysis of these particles. On-line techniques involve fast, near real-time or real-time analysis of particles.1, 14, 21  1.2.1 Off-line methods Traditional methods of aerosol collection for off-line analysis include filter collection, inertial impaction, electrostatic or thermal precipitators, and sedimentation collection.1 Particles are subsequently analyzed by a variety of techniques to determine chemical composition including gas chromatography/mass spectrometry (GC/MS), liquid chromatography/mass spectrometry (LC/MS), atomic absorption spectroscopy, scanning electron microscopy, transmission electron microscopy, and laser microprobe mass analysis, to name only a few.1, 21, 22 While off-line particle analysis can provide valuable information, a fundamental limitation of the technique lies in the long time period for sampling and subsequent analysis. Artifacts such as chemical transformation of the collected species or loss of volatile components can occur.21 For example, adsorption and/or reactions of gases can take place on the collection filters or on particles previously collected.23 The particle composition can also vary as a function of temperature, the pressure drop across the collecting medium, or the composition of sampled air, which will result in changes in the gas-particle equilibria and impact the chemical characterization of the particles.24, 25  1.2.2 On-line methods On-line aerosol analysis was developed as a means to minimize chemical modification of particles and provide a high temporal resolution via rapid sampling and 5  analysis of particles. Techniques include both near-real-time (10 - 60 minute cycles) and real-time measurement (immediate) of chemical composition.26 Some examples include absorption spectroscopy measurements of sulfate and nitrate composition,27,  28  identification of biological particles using fluorescence,29 analysis of water soluble particles by ion chromatography,30, 31 and aerosol mass spectrometry. Aerosol mass spectrometry was developed in the 1970’s to provide effective realtime determination of particle composition for a diverse range of particles. The fundamental characteristic of aerosol mass spectrometers is a continuous sample introduction into the ion source region of a mass spectrometer for immediate chemical speciation.32 This technique has become a mainstay of atmospheric aerosol analysis and is used in both laboratory and field studies.1, 14, 21, 33, 34 Depending on the aerosol mass spectrometer design (discussed in Chapter 2), information about individual particles or particle ensembles can be obtained. The advantage of single particle mass spectrometry (SPMS) is it allows for investigation of variations in particle size and composition for a system of particles. Single particle mass spectrometry has the ability to track and identify particles from various sources in the atmosphere, establish correlations between certain chemical species and particle sizes detected, and observe reaction processes occurring on particles.14, 22 SPMS also allows for determination of the mixing state of a system of particles. Externally mixed systems are comprised of an array of different particles where each particle contains a single component, whereas in an internally mixed system each particle contains a mixture of the various components as illustrated in Figure 1.2.1 In a bulk analysis or ensemble particle analysis, externally mixed and internally mixed systems would be indistinguishable which is disadvantageous as the combination of species in a single particle will affect the reactivity, water uptake, and optical properties of the particle.14, 35 These SPMS features provide critical insights into an aerosol’s properties and interactions in the atmosphere, as well as its chemical history.  6  (a)  (b)  Figure 1.2: Schematic of aerosol mixing states for (a) an externally mixed system and (b) an internally mixed system. Each circle represents an aerosol particle and each color represents a different chemical species.  Details regarding the design of these instruments are discussed in Chapter 2, but some of the current applications of aerosol mass spectrometry in both laboratory and field studies are briefly mentioned here. Some examples of laboratory studies with aerosol mass spectrometry include measurement of the uptake coefficients of atmospheric oxidants (e.g. O3) on a variety of pure and mixed particles in studies of heterogeneous reactions,36-40 monitoring density changes in aerosol particles as a result of heterogeneous reactions,41, 42 and secondary organic aerosol (SOA) studies of the reaction of ozone and α-pinene.42-45 Field campaigns have focused on studies such as identification of sources of rural and urban particle emissions,46-49 characterization of traffic and industry emissions,50-56 detection and identification of bio-aerosol particles,57-59 detection of chemical warfare agents,60 and identification of characteristic particles affected by long range transportation.61-65  7  Chapter 2. General Design and Construction of Aerosol Mass Spectrometers for Aerosol Studies The design of most current aerosol mass spectrometers, while somewhat dependent on the types of particles to be analyzed and the information to be extracted, can generally be described as consisting of three components: (1) an inlet from atmosphere to vacuum for particle sampling, (2) a source region for particle vaporization and ionization, and (3) a mass analyzer. A sizing region for the particles may also be included prior to the source region if particle size information is desired. A basic schematic for an aerosol mass spectrometer is shown in Figure 2.1 and the various components are discussed in more detail below.  Aerosols Particle Inlet  Particle Sizing  Particle Vaporization/ Ionization  Mass Analysis  Figure 2.1: General schematic of an aerosol mass spectrometer.  2.1 Inlet systems The particle inlet serves two functions: sampling the aerosols from atmosphere into vacuum (required for operation of the mass spectrometer) and then focusing the particles into a tightly collimated beam with a small divergence angle. Ideally, an inlet provides a high transmission efficiency for a range of particle sizes and shapes without any chemical composition change. Types of inlets historically used in aerosol mass spectrometers have included a simple orifice, a capillary, and an aerodynamic lens system. A schematic of the three inlet types is shown in Figure 2.2. In each case, the gas emerging from the inlet undergoes a supersonic expansion into vacuum wherein the particles are separated from the expanding gas and a particle beam is formed.34 Some brief theory on particle motion through an inlet is included below. 8  (a)  (b)  (c)  Figure 2.2: Schematic of inlet designs for particle focusing into vacuum (a) orifice, (b) capillary, and (c) aerodynamic lens system.  2.1.1 Brief theory of particle motion in an inlet The Stokes number describes particle behavior in a compressible fluid (e.g. air) and can generally be thought of as the ratio of the time it takes a particle to adjust to changes in fluid motion versus the time it actually has to respond to those changes. Particles traveling through a constriction with a Stokes number (St) >> 1 maintain their own trajectories and thus separate from the fluid streamline, either impacting on the walls before entering the constriction or impacting inside the constriction. Particles with St << 1 follow the fluid streamline closely, contracting and expanding with the gas without focusing along the centerline. Particles with St = 1 are tightly focused along the centerline of a gas stream as they travel through a constriction and then maintain that trajectory at the exit of the constriction. The Stokes number can be calculated with Equation 1:  Equation 1  ρ p d p2UC c = St = Do 9ηDo τ oU  where U is the average flow velocity at the orifice, Do is the diameter of the orifice, and τo is the particle relaxation time.34 This equation can be expanded to show the dependence on particle diameter (dp), particle density (ρp), gas viscosity (η) and the Cunningham correction factor (Cc) for non-continuum effects when the particle size approaches the mean free path.  9  A simple orifice can only focus a single particle diameter for given set of inlet conditions (pressure, temperature, and orifice diameter) since only one particle diameter will have a Stokes number = 1. A capillary can effectively focus a small range of particle diameters since it operates at two Stokes numbers: one for the entrance and one for the exit. However, particles outside that range give broadly divergent beams. The aerodynamic lens system was developed to maximize transmission efficiency for an even greater range of particle diameters while maintaining a tightly focused beam with low divergence over relatively long distances.66,  67  The aerodynamic lens system  consists of a series of orifices of decreasing diameter and ends in a capillary, orifice or combination of both as shown in Figure 2.2c. As the particles travel through each contracting and expanding section, they undergo radial acceleration towards the centerline as shown in Figure 2.3 for a range of modeled particle sizes. The pressure drop across each aperture in the lens is negligible; therefore the only variable in the Stokes number of a lens for a given particle size is the aperture diameter (see Equation 1).66, 67 As the aperture size decreases, the lens optimally focuses smaller particles while larger particles that are already entrained on the centerline are unaffected, resulting in an efficient focusing of a large range of particle sizes. At the exit of the lens system, the gas undergoes a supersonic expansion into vacuum where the particles acquire a distribution of terminal velocities dependent on particle diameter.  10  Figure 2.3: Schematic of particle trajectories through the aerodynamic lens (5 apertures and nozzle) showing the focusing of particles along the centerline. R is the distance from the axis of the lens, X is the distance along the lens axis. Adapted from Zhang et al.68  2.2 Particle sizing The function of the particle sizing region is to acquire size data as a characteristic parameter for the aerosols and/or to provide appropriate triggers for pulsed systems of desorption/ionization/mass analysis used to obtain composition information (discussed in section 2.3). Previous methods to determine the size of particles in real-time have included both single laser scattering69,  70  and time-of-flight or velocimetry  measurements.56, 71-77 Light scattering intensity from a single laser-particle interaction has been successfully used to trigger particle ionization and provide a rough determination of particle size;63 however, the applicability of the technique is limited when the particle diameter falls significantly below the wavelength of light. In this region, scattering intensity falls off rapidly (dp-6 dependence).34 Secondly, as particles on the order of the laser wavelength scatter according to Mie theory, the scattering intensity is a complex 11  function of particle size, scattering wavelength, scattering angle, and refractive index. Measuring the intensity at a single angle has been shown to be an unreliable indicator of size,78 and variation in where the particles pass through the beam can result in fluctuations in the scatter intensity.73 Velocimetry measurements include the use of a chopper wheel or velocity analysis via in-flight detection through two timing points. A rotating chopper wheel can be used to select clusters of particles from the particle beam. The time between the opening of the chopper wheel and the detection of ion signal from particles arriving in the ionization region is used to determine the velocity and hence the particle size.79 This method is employed by the commercially available Aerodyne Aerosol Mass Spectrometer (AMS), which is a thermal desorption instrument. The second method of velocity determination involves measuring the transit time of particles as they pass through two continuous-wave laser beams.73, 74, 80 The scattered light is detected by photomultiplier tubes and the particle velocity is calculated based on the measured transit time and known distance of the two scattering events. A trigger can then be generated based on individual particle velocity for the desorption and/or ionization laser to fire upon arrival of the particle in the ionization source region. This method provides precise determination of particle size and correlated single particle mass spectra and composition information.81  2.3 Particle vaporization and ionization 2.3.1 Single step vaporization/ionization The particle vaporization/ionization process is the most widely varying component of the aerosol mass spectrometer design. Depending on the type of aerosol particles, the sampling rate at which they are to be analyzed, and the desired information to be extracted, certain methods of vaporization/ionization will be better suited than others. In many instruments, vaporization and ionization occur in one step using a single pulsed laser, which is referred to as laser desorption/ionization (LDI). Typically a single pulsed ultraviolet (UV) laser is used.63,  74, 82-85  Intense pulses of UV wavelengths are  capable of efficiently desorbing and ionizing salts, metals, and crustal particles which can 12  be advantageous for certain particle studies.86 Conversely, LDI produces massive fragmentation of organic species due to the high laser intensity and absorption of multiple UV photons by organic ions. In general while LDI can work extremely well for detecting multiple components of single particles, it suffers from poor signal reproducibility, matrix effects, and a chemical bias towards certain species.34, 87  2.3.2 Two step vaporization/ionization Two step vaporization/ionization processes were developed as a means to obtain more quantitative and reproducible results. By separating the desorption and ionization steps, optimization of each process can be obtained. Furthermore, matrix effects in single step desorption/ionization that lead to ion-ion recombination reactions and depletion of ion signal are eliminated.88-91 For the desorption step, both a pulsed laser51, 77, 92 and a heated surface93, 94 have been applied. Heated surfaces are commonly used in conjunction with continuous ionization methods while lasers are combined with pulsed ionization methods. Ionization methods can be either continuous or pulsed. Continuous ionization methods include electron impact ionization,79,  95  chemical ionization,40 continuous  vacuum UV photoionization,96 and atmospheric sampling glow discharge ionization.97 Pulsed ionization methods have included resonance enhanced multiphoton ionization (REMPI),56,  77, 92, 98  single photon ionization (SPI),99-101 photoelectron  resonance capture ionization (PERCI),102 and a pulsed form of chemical ionization produced by generating alkali ions via an excimer laser.103  2.4 Mass analysis Types of mass analyzers employed in aerosol mass spectrometers include quadrupole mass filters, time-of-flight (TOF) analyzers, and 3D ion trap mass spectrometers. The quadrupole mass filter was predominantly used in the early days of real-time particle analysis and is primarily associated with continuous ionization sources such as electron impact.47, 71, 79, 95 The quadrupole is also utilized for chemical ionization mass spectrometry (CIMS).40 The quadrupole can be scanned to acquire a full mass  13  spectrum for ensemble averaged particle data; however, for single particle data it is limited to monitoring a single mass-to-charge ratio (m/z) due to the time required to complete a scan in comparison to the timescale of the particle vaporization event.26 The TOF analyzer is capable of both single particle and multiple particle analysis since a full mass spectrum can be acquired for each particle, which allows for investigation of the chemical variability of individual particles. Furthermore the TOF is well-suited to the analysis of pulsed ions produced via laser desorption/ionization.8 The dual TOF analyzer also allows for simultaneous analysis of both positive and negative ions which provide different particle information.104 The TOF analyzer has been implemented in multiple laboratory and field instruments.56, 61, 69, 74, 76, 77, 85, 93, 94, 105 Ion trap mass spectrometers have also been used for single particle analysis although not as widely as TOF analyzers.57, 75, 82, 106-111 The ion trap offers a multitude of desirable properties: relatively low cost, compact size, high duty cycle, high sensitivity, full spectrum collection for single particles, and tandem mass spectrometry capabilities for further speciation. While it does not have as wide a dynamic range as a TOF and can experience space charge issues,21 the ion trap is an extremely versatile mass analyzer that can be coupled to many ion sources.60 The MS/MS capabilities can also provide valuable structural information about particle composition. Successful MS/MS studies have been performed on a variety of aerosols including bacteria,107 uranium oxide particles,112 diesel engine exhaust,52 and chemical warfare species.60  2.5 Commercial instruments Two commercially available, portable aerosol mass spectrometers exist today and are used in numerous field and laboratory studies.41, 45, 46, 80, 113-119 The Aerodyne Aerosol Mass Spectrometer (AMS) uses thermal desorption and electron impact ionization with either a quadrupole mass spectrometer or TOF.47,  61, 79, 85  Non-refractory species,  including ammonium, nitrate, sulfate, and organic carbon, have been successfully studied in a range of particle diameters from ~50 - 1000 nm. The other commercially available aerosol mass spectrometer is the TSI Aerosol Time of Flight Mass Spectrometer (ATOFMS), a single particle instrument that employs  14  single step laser desorption/ionization at 266 nm.76,  81  This instrument uses the  velocimetry sizing technique to determine each particle’s size and provide triggers for the laser desorption/ionization. A dual TOF allows for simultaneous collection of both positive and negative ions, enhancing the amount of chemical information about the particles.  2.6 Motivation for this work As described above, aerosol characterization remains a significant challenge given the variety in size, chemical composition, and properties of atmospheric particles. Given the ubiquity of organic species in atmospheric aerosols and their comparatively poor understanding (only 10-15% of the species in the organic fraction have currently been identified), acquiring accurate information on their properties and behavior is essential for improving our understanding of their effects on human health, atmospheric chemisty, and global climate.14 Current aerosol instruments have both strengths and weaknesses for the analysis of organic species. For example, the Aerodyne instrument (AMS) is capable of determining total organic mass and differentiating oxygenated organic species from hydrocarbon-like organic species.17 However, in terms of chemical speciation, the instrument is better suited to identification of general classes of compounds rather than molecular speciation since electron impact is a high energy ionization process (70 eV) that leads to extensive fragmentation of many organic species. The AMS is not capable of MS/MS studies which can be used for species identification and extracting additional structural information. The TSI instrument is quite powerful for determining composition variability between individual particles, but as the ionization process is single step UV laser desorption/ionization, matrix effects are particularly strong for metal and salt containing particles and organic species undergo extensive fragmentation due to the multiphoton process.86 The TSI instrument is also not capable of MS/MS studies. Continued development of aerosol instrumentation is still necessary to expand the range of particle types and sizes that can be analyzed, as well as increasing the field and laboratory instruments available for future studies.14, 21, 86  15  The overarching goal of this project is to develop and characterize new single particle mass spectrometers for organic aerosol analysis that incorporate a 3D ion trap. These instruments will take advantage of the strengths of an ion trap mass spectrometer and may prove beneficial over existing aerosol instruments in some areas. As discussed earlier, ion traps offer versatility, compact size, a high duty cycle, high sensitivity, full spectrum collection for single particles, a large accessible mass range, and tandem mass spectrometry capabilities for further speciation. Ion/molecule reactions can also be studied with an ion trap. These new instruments may have potential applications in laboratory studies and, in some cases, field studies of organic aerosol particles. Existing ion trap mass spectrometers for aerosol analysis include versions that utilize pulsed laser desorption/ionization,52,  57, 75, 82, 107  thermal vaporization/chemical  ionization (CI),60 or thermal vaporization/EI.60, 109 The pulsed laser desorption system is capable of single particle analysis and provides in-trap ionization, which will enhance the sensitivity of the system.109 However, as laser desorption/ionization is a single step process, issues of charge-transfer matrix effects arise as well as extensive fragmentation of organic species.120 The use of thermal desorption/EI or thermal desorption/CI ameliorates the matrix effect problem, but at the expense of single particle information and sensitivity. Heated surfaces are not generally mounted inside the ion trap due to the disturbances caused in the trapping field, thus ions are created outside the trap in an ion source region and then pulsed inside with an efficiency of 1-5%.109 One other version of an  aerosol  ion  trap  mass  spectrometer  has  employed  two  step  laser  vaporization/ionization for single particle studies which successfully eliminated matrix effects.51 That study also showed a bias towards polycyclic aromatic hydrocarbons since a REMPI ionization process was used.51 This thesis describes the development and characterization of two versions of a novel single particle ion trap mass spectrometer (SPITMS) that provide real-time measurements using a two step desorption/ionization process in an ion trap. Version I of the instrument incorporates a pulsed CO2 laser for desorption and pulsed 70 eV electron impact (EI) ionization which has never been implemented for single particle studies before. Version II of the instrument incorporates a pulsed CO2 laser for desorption and a novel, tunable, laser-based vacuum ultraviolet (VUV) source to provide single photon  16  ionization. Versions I and II of the instrument are examined for strengths and weaknesses in analyzing single organic aerosols and their potential applicability in laboratory and field studies.  2.7 Thesis Overview This thesis can be broadly described as consisting of three sections detailing the development and characterization of version I (CO2/EI) and version II (CO2/VUV) of a novel single particle ion trap mass spectrometer (SPITMS), as well as the application of the two versions for studies of atmospherically realistic aerosols. Chapter 4 provides a thorough description of the instrument interface common to both versions of the SPITMS. The particle inlet and sizing region are discussed in detail, including a characterization of the detection efficiency and the calibration of the sizing region. Lastly, the instrument interface is coupled to a differential mobility analyzer (DMA) to allow measurement of particle density for spherical particles. Chapter 5 covers the development and characterization of the laser desorptionelectron impact (CO2/EI) ion trap mass spectrometer (version I). The first results from studies of 2,4-dihydroxybenzoic acid, caffeine, and oleic acid using the SPITMS version I are presented. Tandem mass spectrometry studies up to MS3 of single caffeine particles are also illustrated. The effect of various operational parameters on the mass spectra and fragmentation patterns obtained in these studies are examined. Lastly, the single particle detection limit of the instrument is ascertained for 2,4-dihydroxybenzoic acid. Chapter 6 describes the use of version II of the instrument (CO2/VUV) to study particles of pure oleic acid, pure oleyl alcohol, and mixtures of the two over a range of compositions. The mass spectra, ionization delay profiles, and fragmentation are studied as a function of desorption energy for one component particles and as a function of particle composition for two component particles. In chapter 7 version II is used to acquire mass spectra for two component particles of oleic acid and 1-octadecene as a function of ionization wavelength. We illustrate that we can use the tunable VUV source to simplify the mass spectra and determine the  17  ionization energy for a single component in a mixed particle, both of which could be potentially helpful in identifying an unknown species in a particle. In chapter 8, analysis of a variety of atmospherically realistic aerosols is described using versions I and II of the instrument. Suwannee River fulvic acid particles, cigarette sidestream smoke particles, meat cooking aerosols and ammonium bisulfate are chosen to represent a range of atmospherically realistic particles. Results show that the suitability of both versions for studying these aerosols strongly depends on particle composition. Furthermore, the use of the CO2 laser for a one step desorption/ionization process may be an analysis technique meriting continued research for complex particle analysis, particularly in field instruments. The final chapter summarizes the work completed for this thesis and explores some potential future applications of versions I and II of the single particle ion trap mass spectrometer.  18  Chapter 3. Design and Characterization of the Interface Common to Both Versions of the Single Particle Ion Trap Mass Spectrometer1 3.1 Interface overview Both versions of the single particle ion trap mass spectrometer (SPITMS) for organic aerosol studies utilize the same interface, which includes the particle inlet and particle sizing region as shown below in Figure 3.1. X,YStage Differential Pumping Stage  Light Scattering  Discriminator/ Amplifier PMTs  Aerodynamic lens  -2  10 Torr  10-5 Torr  1-2 Torr  cw Nd:YAG Lasers (532 nm)  Particle Focusing  Particle Sizing and Triggering  Figure 3.1: Schematic of the single particle ion trap mass spectrometer (SPITMS) interface. The aerosol inlet consists of an aerodynamic lens system to sample particles from atmosphere into vacuum and to collimate particles between 0.1 - 1 µm diameter into a 1  A version of this chapter has been published. Simpson, E.A., Campuzano-Jost, P., Hanna, S. J., et al. A laser desorption-electron impact ionization ion trap mass spectrometer for real-time analysis of single atmospheric particles. International Journal of Mass Spectrometry. 2009, 281, 141-149.  19  tightly focused beam. The particle sizing and trigger source region is based in part on the work by Su et al.81 Particle velocity is determined by measuring the transit time of aerosols through two focused continuous wave laser beams and is then used to extract particle size. Custom, real-time data acquisition software records the timing and generates the appropriate triggers for the mass analysis region. These two regions of the interface are discussed in detail below, followed by a rigorous characterization of the particle sizing calibration and sizing detection efficiency.  3.2 Particle inlet and focusing region As mentioned in Chapter 2, an aerodynamic lens consists of a series of apertures of decreasing size where each successive aperture focuses a range of particles sizes with a Stokes number ~1, while particles already traveling along the centerline remain undisturbed. Our aerodynamic lens system was based on the work of Liu et al.66, 67 and machined in house. The aerosols are first sampled through a 100 µm critical orifice which sets the upstream pressure to 1.6 Torr. After a 30 cm long relaxation tube (ID 16 mm), the aerosols enter the aerodynamic lens, which consists of a series of five apertures of 5.0, 4.5, 4.0, 3.75, and 3.5 mm in diameter used to collimate the particle beam. The first aperture (5.0 mm) is a thin capillary 10 mm long. The subsequent apertures are thin plate orifices 0.5 mm thick. Each aperture is separated by a spacer 5 cm in length with a 10 mm ID. The apertures are followed by a nozzle composed of a 10 mm long, 6 mm diameter throat and a 3 mm diameter thin plate exit orifice. Shown in Figure 3.2 are the components of the aerodynamic lens which are stacked into the aerodynamic lens mount. This set of lenses has been shown to provide good focusing and > 90% transmission for particle sizes between roughly 0.1 and 1 µm.  20  Figure 3.2: Picture of disassembled aerodynamic lens components: (above) aerodynamic lens mount, (below) aerodynamic lens orifices, spacers, and nozzle. As the particles exit the nozzle of the lens, they undergo a supersonic expansion into vacuum creating a tight particle beam. Particles achieve terminal velocities between 50 - 120 m/s, which are strongly dependent on their aerodynamic diameter. The aerosols exit the lens into a differential pumping chamber at a pressure of 1.4x10-2 Torr, pumped by a 300 L/s turbomolecular pump (V301, Varian Inc.). A high performance X-Y stage (Thermionics) with 1 µm resolution is used to align the aerosol lens with the mass analysis region (located ~41 cm downstream).  3.3 Particle sizing region The particle sizing region provides single particle information and appropriate triggers for the pulsed vaporization and ionization steps.81 Particles pass through the differential pumping chamber and into the sizing region via a 5 mm orifice. The sizing region, which operates at approximately 7x10-5 Torr, is pumped by a second 300 L/s turbomolecular pump (Varian 300HT). Within the sizing region, particle velocities are determined by measuring the transit time between two continuous wave (cw) 532 nm laser beams located 6 cm apart. The calibration aerosols for the sizing region are NIST traceable polystyrene latex nanospheres (PSLs) of known diameters (Thermofisher, 3000 Series).56, 74, 80, 81  21  Two different configurations for particle sizing were used during the instrument development. In the first sizing configuration of the instrument, a single scattering laser (100 mW, 532 nm cw Nd:YAG; Crystalaser Ltd.) was split into 2 beams and focused via a telescope system to ~80 µm spot sizes for particle detection (referred to as “one laser setup” in the following). In this setup, small, reproducible adjustments in the vertical position of the laser beams were necessary to compensate for the gravitational settling of larger particles that occurs due to the horizontal orientation of the aerodynamic lens. In the second setup (two laser setup), two single mode 100 mW, 532 nm cw Nd:YAGs (Spectra-Physics Excelsior) are focused by single lenses to ~280 µm laser spot sizes. The use of 2 lasers provides higher powers and independent alignment of each beam. The larger laser spot size is a compromise that allows for high detection efficiencies of spherical particles 300 nm and above without any adjustment in the vertical position of the laser beams to compensate for particle settling. The two laser setup was used for all mass spectrometric measurements reported in this thesis. The scattered light is collected by two custom ellipsoidal mirrors to optimize particle detection (f=101 mm, Optiform Ltd).56 Each mirror has four 2 mm apertures drilled at 90o to allow passage of the particle beam and the laser beam. The particle beam and laser beam intersect at one focus of the ellipsoidal mirror and the scattered light is collected and refocused to the second focal point of the mirror behind which the photomultiplier tube (PMT 9001V, Electron Tubes Ltd) is located. The mirrors are mounted at opposing 50° angles relative to the instrument centerline to ensure particle alignment and reduce the number of false triggers. Note that the diagram of the instrument in Figure 3.1 incorrectly suggests that the laser beams enter on the same side for simplicity of drawing, but in reality, the laser beams enter from opposite sides and the two PMTs are on opposite sides (always facing the opening of the mirror) of the interface. The alignment of the ellipsoidal mirrors can be adjusted via custom machined y-z stages. The scattering signals are amplified by a shaping amplifier and discriminated with a single channel pulse analyzer (Ortec 590A), generating TTL pulses with a maximal 50 ns jitter relative to the maxima of the scattering pulses. A field programmable gate array (FPGA) board (PCI-7831R, National Instruments) is used to record the particle velocities, generate the triggers for the  22  desorption and ionization events, and record other particle relevant information such as scattering signal size for the analyzed particle. It also generates the clock for the MS acquisition software. By using a real-time software solution, it is possible to acquire size information independently from the actual triggering so that particle size distributions of sampled aerosols (as opposed to analyzed aerosols) can be accessed at any given time.  3.4 Particle sizing region characterization Polystyrene latex nanosphere standards 200 - 800 nm diameter (Thermofisher, 3000 Series) were used to calibrate the particle sizing region and ascertain sizing detection efficiencies. The standards were diluted in Millipore water (18 MΩ) and aerosolized with a medical nebulizer. The particles were size selected (mobility diameter) with a differential mobility analyzer (DMA) (TSI Inc., Model 3081). The flow was then split into a condensation particle counter (CPC) (TSI Inc., Model W3782) and the single particle ion trap mass spectrometer. In this manner both mobility diameter and aerodynamic diameter could be simultaneously recorded as well as the particle detection efficiency. Some definitions relevant to particle size are listed below. The mobility diameter (dm) of a particle is defined as the diameter of a sphere (of unit charge) with the same migration velocity in a constant electric field as the particle of interest.121 The vacuum aerodynamic diameter (dva) of a particle is defined as the diameter of a sphere with standard density (1 g/cm3) that settles at the same terminal velocity as the particle of interest.122 The volume equivalent diameter (dve) is defined as the diameter of a spherical particle of the same volume as the particle under study.123 For a spherical particle, the volume equivalent diameter is the same as the geometric (physical) diameter. For an aspherical particle or particle with internal voids, the volume equivalent diameter is the diameter a particle would have if it were melted to form a droplet. Additionally for a spherical particle, the mobility diameter and volume equivalent diameter are equal. These terms will be referred to in subsequent sections of this thesis.  23  3.4.1 Calibration of sizing region As mentioned previously in section 2.2, the particles undergo supersonic expansion into vacuum and acquire size-dependent terminal velocities. The terminal velocities can be experimentally correlated to the vacuum aerodynamic diameter (dva) by using calibration aerosols of a known size, shape, and composition and fitting the velocities with the two-parameter function used in Zelenyuk et al.56:  Equation 2  d ρ  v = α ⋅ d vac = α ⋅  ve p   χρ  o    c  v is the particle velocity, α and c are fit parameters, dva is the vacuum aerodynamic diameter, dve is the volume equivalent diameter, ρp is the particle density, ρo is the standard density of 1 g/cm3, and χ is the shape factor (for spherical particles χ = 1, for irregular particles χ is almost always greater than 1).47 It should be noted that in the rest of this work, “aerodynamic diameter” and “vacuum aerodynamic diameter” are used interchangeably. Figure 3.3 shows the particle velocities for the PSL calibration aerosols as a function of vacuum aerodynamic diameter (calculated from known dve values and χ and ρp of 1.0141 and 1.05 g/cm3 (manufacturer specification), respectively) for the two laser and one laser sizing setups described in section 3.3. Note that the different laser setups also had different alignments of the aerosol lens. Despite the measurements being taken 12 months apart with very different sizing detection geometries, the total deviation between both datasets in Figure 3.3 is < 2%. Figure 3.3 also shows the fits correlating velocity to vacuum aerodynamic diameter for the PSL calibrations of the two sizing setups. For the one laser setup, the fit parameters α and c were 0.520 and -0.358, respectively. For the two laser setup, the fit parameters α and c were 0.423 and -0.371, respectively. In both cases r2 is > 99.9%. Based on the PSL calibration fit, the velocity distributions of unknown particle sizes can be converted into aerodynamic size distributions.  24  130  1 Laser setup 2 Laser setup  Particle velocity (m/s)  120  110  100  90  80  200  300  400  500  600  700 800 900  Vacuum aerodynamic diameter d va (nm)  Figure 3.3: PSL calibration curves for particle velocity versus (calculated) vacuum aerodynamic diameter for the one and two laser sizing setups. To illustrate the particle sizing precision, size distribution widths were calculated from the measured PSL velocities and are listed in Table 3.1 as a function of the PSL size (dve) stated by the manufacturer. Measured relative standard deviations varied from 0.9 2.0%, which is only slightly larger than the manufacturer specified relative standard deviations (σPSL/dve).  Table 3.1: Measured PSL size distribution widths with the SPITMS (about 1000 particles per size) compared to the manufacturer’s stated distribution widths. PSL Diameter [nm]  Manufacturer specified width σPSL [nm]  Measured size distribution width σTOF [nm]  Manufacturer sizing relative standard deviation [%]  Measured sizing relative standard deviation [%]  299 404 499 596 707 799  5.1 5.9 6.5 7.7 8.5 8.3  5.7 5.9 8.7 11.6 11.7 7.2  1.7 1.5 1.3 1.3 1.2 1.0  1.9 1.5 1.8 2.0 1.7 0.9  25  3.4.1.1 Sizing detection efficiencies Sizing detection efficiencies were determined by comparing the number of aerosols sized by the instrument with the number of aerosols counted with the condensation particle counter running in parallel. The results for the two sizing arrangements in the instrument are shown in Figure 3.4. The results from the one laser setup were obtained by optimizing the vertical alignment of the laser beams with the particle beam at every size to compensate for vertical settling of larger particles. The results from the two laser setup were obtained at the same alignment position for all sizes.  100 90  Detection efficiency (%)  80 70 60 50 40 30 20  1 laser setup 2 laser setup  10 0 200  300  400  500  600  700  800  PSL geometric diameter (nm)  Figure 3.4: Detection efficiency of the SPITMS particle sizing region for PSLs with the one laser and two laser sizing setups. With our current lens, system pressures, and scattering detection arrangement, sizing efficiencies well over 85% are achievable for particles 400 nm and above and about 1000 particles/s (ie. ~700 particles/cm3) can be accurately sized. Zelenyuk et al.56 reported sizing efficiencies close to 90% at 400 nm, which decreased towards larger sizes (10% at 800 nm). Harris et al.82 reported a 95% detection efficiency at 548 and 740 nm for their vertically mounted ion trap aerosol mass spectrometer, but did not provide data for other size ranges. Our efficiency at 300 nm compares favorably with reported  26  detection efficiencies of ~44% for 300 nm particles by Su et al.81 in an ATOFMS upgrade for enhanced detection of small particle sizes. Zelenyuk et al. have recently redesigned their system for fine and ultrafine particle measurement and detection efficiencies of ~100% for sizes between 200 and 600 nm have been achieved.124 The scattering detection limits of Zelenyuk’s instrument56 and Prather’s instrument81 are 98 nm and 95 nm diameter, respectively. These are both significantly lower than the scattering detection limit of ~225 nm determined for our current system (using a variety of particle types over a large size range for which the smallest size with detectable signal was recorded). As shown in Figure 3.4 the detection limit actually decreased with the two laser arrangement (e.g. not able to detect 200 nm particles) despite a higher laser intensity compared to the previous one laser setup. This is almost exclusively a function of the much larger amount of laser light scattered off the input windows in the new configuration; thus, scattering response is currently background limited rather than signal limited. While the inside of the instrument has already been anodized to reduce stray light, improvements such as the use of higher quality windows, a more precise alignment of the optical detection components, or an input baffle system could substantially improve the detection limit.56, 124 The SPITMS interface performance with other particle types confirmed that regardless of density, phase or chemical composition, the scattering detection limit for the particle sizing region was between 210 and 240 nm diameter, which is the region where scattering intensity for a 532 nm laser declines sharply. This scattering detection limit could be enhanced by going to a shorter wavelength of laser light and further tightening the focus of the beam. The upper limits of our detection range are limited by the vertical displacement experienced over the instrument path length. While particles with aerodynamic diameters up to 1500 nm can be detected with the SPITMS interface, it is currently not practical to sample particles in excess of 1200 nm due to their gravitational settling.  27  3.5 Particle density measurements Particle velocity measurements can also be coupled with particle mobility diameter measurements (using a differential mobility analyzer) to obtain particle density.42,  122, 125-129  Particle density measurements are important for several reasons  including the conversion of measured size distributions to mass loading.130 We tested a variety of solid and liquid particles and were able to determine particle densities within 3% of the literature value for spherical particles. Details of the experiment and theory are included in Appendix I. Table 3.2 summarizes the measured particle densities for a number of liquid and solid aerosols. Liquid compounds should form spherical aerosols, while both NaHSO4 and NH4HSO4 are highly hygroscopic salts expected to persist as droplets rather than completely dry solids.126 For all particles tested, χ was taken as 1 in the fit since the particles are spherical or expected to be spherical based on previous work. Densities determined in this way were within 3% of the reported bulk density. This is comparable to measurements of particle density by Zelenyuk et al. for 10 different compounds determined to be within 2.7% of literature values.126  Table 3.2: Summary of particle densities determined by simultaneous measurement of mobility diameter (with a DMA) and vacuum aerodynamic diameter (SPITMS) and comparison to literature values.  a  Compound  Measured density (g/cm3)  Literature density (g/cm3)  Deviation from literature  Canola oil  0.933  0.915131  2%  Oleic acid  0.906  0.895131  1%  Vacuum pump oil  0.885  0.90132  1.6%  NaHSO4  2.24  2.22130 2.19±0.02126  0.7%  NH4HSO4/ 10wt % NaHSO4  1.82  1.82a  0%  NH4HSO4  1.73  1.78131  2.8%  extrapolated from linear fit between bulk density of pure NaHSO4 and pure NH4HSO4  28  3.6 Conclusions Characterization of the interface common to both versions of the SPITMS shows a comparable performance of the sizing efficiency to existing single particle mass spectrometers that use velocity measurements for size determination. A sizing detection efficiency of 90% or greater is achieved for aerosols ≥ 400 nm diameter and a scattering detection limit of ~225 nm diameter is determined for the sizing region. The particle sizing precision was demonstrated to be within 4.6% at the most conservative estimate. Lastly we demonstrated that we can effectively determine particle densities of spherical particles within 3% of the bulk literature densities. Theses results show that the interface is well-suited to the desired range of particles for analysis (~0.2 – 1 µm), which are of particular interest for atmospheric studies given their long life-times, large surface area to diameter ratio (likely sites for heterogeneous chemistry), and potential health impacts.  29  Chapter 4. Real-Time Analysis of Single Particles Using Laser Desorption-Electron Impact Ionization in an Ion Trap Mass Spectrometer (Version I)2 4.1 Introduction This chapter focuses on the development of a new type of ion trap mass spectrometer for single particle analysis which incorporates a pulsed CO2 laser for in-trap desorption and 70 eV pulsed electron impact (EI) ionization. The instrument uses the interface described in Chapter 3; a pulsed two step desorption/ionization process to improve signal reproducibility, minimize charge-transfer matrix effects, and minimize extensive fragmentation of neutrals; and an ion trap mass spectrometer for tandem mass spectrometry studies. The MS/MS capabilities will allow investigations of the molecular structure of organic aerosol components and further speciation, as well as ion-molecule reaction studies. Thermal vaporization/EI has been successfully used in quadrupole and TOF mass spectrometers for substance class grouping and quantitative measurements of total carbon compounds in aerosols.79, 85 The implementation of thermal vaporization/EI in an ion trap is more difficult due to the geometric constraints of the trap and possible trapping field disturbances. For these reasons, ions are generally created outside of the trap when using this technique in existing ion trap aerosol mass spectrometers.60,  109  Kürten et al.109  estimate a 1 - 5% trapping efficiency for ions created outside the trap, in line with similar efficiencies measured for electrospray ionization outside an ion trap.133 In comparison to existing ion trap instruments that use EI,60,  109  our design is  notably different in that vaporization is achieved with a CO2 laser rather than a heater cartridge. By using a pulsed laser for desorption, we achieve in-trap vaporization. Since both  the vaporization and ionization steps occur inside the trap, we should obtain  improved sensitivity in comparison to existing ion trap instruments utilizing EI which require ions to be generated outside the trap.109 Furthermore, in our instrument the 2  A version of this chapter has been published. Simpson, E.A., Campuzano-Jost, P., Hanna, S. J., et al. A laser desorption-electron impact ionization ion trap mass spectrometer for real-time analysis of single atmospheric particles. International Journal of Mass Spectrometry. 2009, 281, 141-149.  30  desorption and ionization steps are pulsed events triggered by particle velocimetry measurements which makes single particle studies possible. A mid-IR CO2 laser for desorption56, 92, 98, 99, 115, 134 has several advantages over other laser vaporization methods: a low photon energy that allows for high fluencies before ionization takes place,135 minimization of the self-focusing effect for submicron particles which reduces both inhomogeneous heating and preferential surface desorption,100 and a fairly long pulse width that reduces the possibility of particle shattering.136-138 Additionally, many organics exhibit some absorption in the accessible wavelength range of a tunable CO2 laser (9.2 - 10.8 µm). In comparison to a cartridge heater, laser desorption results in superior gas phase background discrimination and does not suffer from particle bounce effects.8 On the other hand, using a pulsed laser requires a single particle triggering scheme that is complex and often leads to lower particle hit rates than heater impaction, especially for smaller particles. As mentioned above, the CO2/EI ion trap mass spectrometer minimizes matrix effects by separating the desorption and ionization steps, similar to two step laser desorption/ionization techniques, but without the added cost and complexity of a second laser. Matrix effects observed previously in one step desorption/ionization processes are attributed to charge-transfer reactions in the ablation plume which affect the detection sensitivity of different analytes and cause variation in the overall measured ion distribution.92, 112, 120 CO2/EI has been employed previously for bulk phase studies,139-144 but to our knowledge no single particle studies have been conducted with this technique. EI provides useful information with linear and reproducible ionization of gas phase molecules, multi-species applicability, quantitative capabilities, and a large library of standard spectra for comparison.47,  79  Additionally, with an ion trap mass analyzer,  tandem mass spectrometry can be performed for structural identification of more robust species.51, 82, 109 In this chapter our new single particle ion trap mass spectrometer, which combines pulsed IR desorption with EI ionization, is described in detail. A preliminary characterization of the CO2/EI technique follows. The instrument provides sizing and compositional information with minimal matrix effects. The mass spectra obtained for aerosols of three different compounds (liquid and solid phase) are comparable to the  31  NIST EI spectra. MS3 studies are successfully performed. We also investigated the influence of vaporization energy and electron gate width on total ion signal and the effects of ionization delay time on the degree of fragmentation. Results from 2,4dihydroxybenzoic acid particles give a detection limit of a 325 nm diameter particle (22 fg).  4.2 Experimental Our novel single particle ion trap mass spectrometer was designed to include many of the advantageous features discussed in Chapter 2. A schematic of the instrument is shown in Figure 4.1. The three basic sections of the instrument are briefly listed below in the direction of particle travel through the instrument: • An aerosol inlet where an aerodynamic lens system is used to sample particles  from atmosphere into vacuum and collimate particles between 0.1 - 1 µm diameter into a tightly focused beam. • A particle sizing and trigger source region based in part on the work by Su et  al.81 Particle velocity is determined by the transit time of aerosols through two, focused, continuous wave laser beams and is used to extract particle size. Custom, real-time data acquisition software records the timing and generates the appropriate triggers for the mass analysis region. • A mass analysis region comprised of a mid-IR CO2 laser for particle  desorption, an electron gun, and the ion trap mass spectrometer.  32  Real Time Control X,YStage  Light Differential Pumping Scattering Stage  MS Control & Acquisition  Discriminator/ Amplifier  PMTs  Aerodynamic Lens  1-2 Torr  Amplifier  Ion Detector 10-6 Torr  10-2 Torr  RF Power Supply  10-5 Torr EI Gun  Tunable CO2 Laser  cw Nd:YAG Lasers (532 nm)  Particle Focusing  Particle Sizing and Triggering  Particle Analysis  Figure 4.1: Schematic of SPITMS version I (laser desorption-electron impact ionization). The aerosol inlet and particle sizing region were characterized in detail in Chapter 3. The specifics of the mass analysis region are included in section 4.2.2.  4.2.1 Particle generation Laboratory-generated aerosols of 2,4-dihydroxybenzoic acid (DHB) (Fluka, ≥98%), caffeine (Aldrich, ≥98.5%), oleic acid (Aldrich, ≥99%), and linoleic acid (Aldrich, ≥99%) were used to test the combined sizing and mass analysis capabilities of the SPITMS version I. The chemical structure and molecular weight of each compound are shown below in Figure 4.2. Chemicals were used without further purification. Particles were produced by aerosolizing solutions with the TSI vibrating orifice aerosol generator (VOAG; TSI Inc., Model 3450) or constant output atomizer (COA; TSI Inc., Model 3076). Solutions used with the VOAG were on the order of 10-5 g/mL concentration and solutions for the COA were on the order of 10-3 g/mL. Oleic and linoleic acid solutions were prepared in 2-propanol (Aldrich, 99.9%). 2,4dihydroxybenzoic acid and caffeine solutions were prepared in Millipore water (18 MΩ). 33  Aerosols were passed through a 85Kr charge neutralizer (TSI Inc., Model 3054) to reduce the high electrostatic charge on the generated aerosols to a bipolar charge distribution thus minimizing coagulation and optimizing effective transport through the tubing connecting the aerosol generation system to the DMA/CPC or instrument. Aerosols traveled through the charge neutralizer with a high flow of air (~30 LPM) and then passed through a 24″ nafion diffusion dryer (Permapure Inc.) before entering the SPITMS to ensure particle dryness and avoid size changes in the aerosol lens.145 O  OH  H N  HN  O  HO  O  OH  2,4-dihydroxybenzoic acid (M.W. 154.12)  N H  N  Caffeine (M.W. 194.19)  HO  HO  O  O  Oleic acid (M.W. 282.46)  Linoleic acid (M.W. 280.45)  Figure 4.2: Chemical structures and molecular weights (g/mol) of 2,4dihydroxybenzoic acid, caffeine, oleic acid, and linoleic acid.  4.2.2 Particle analysis After exiting the sizing region, the particles enter the vacuum chamber housing the mass spectrometer, which is pumped by a third turbomolecular pump (Varian V300HT) and operates at ~6x10-6 Torr. Aerosols enter the ion trap (approximately 41 cm from the exit of the aerodynamic lens) via a 2 mm orifice in the ring electrode. The control software provides TTL triggers to fire the desorption laser upon arrival of the particle in the center of the ion trap. The EI gun is subsequently fired at a programmed ionization delay time and a mass spectrum is acquired.  34  4.2.2.1 Desorption A pulsed, transversely excited atmospheric pressure (TEA) CO2 laser (MTL-3G, Edinburgh Instruments) is used for desorption. The MTL-3G is a single mode, tunable laser with over 60 lines between 1087 and 926 cm-1 (9.2 - 10.8 µm) and a maximum output of 50 mJ/pulse on the strongest lines (~5x107 W/cm2). The CO2 desorption laser enters at a 35.2° angle between the ring and end cap of the trap to intersect with the particle in the center of the trap (see Figure 4.1). A 250 mm ZnSe lens focuses the output of the CO2 laser to a ~1 mm diameter spot inside the ion trap for particle desorption. The main pulse of the CO2 laser is 140 ns wide followed by a ~1 µs broad tail. In this study between 3 - 40 mJ of IR light was used to vaporize the particles. Studies were performed at a desorption wavelength of 944 cm-1 due to its common use in two step laser desorption/ionization experiments.92,  98, 99, 115, 124  The CO2 energy used  depended on the particle composition and size. In all cases, the highest energy possible that did not produce ions from the desorption step alone was chosen. The average CO2 laser power is measured at the exit of the trap with a thermal detector (Ophir Model 3ASH). From the measured absorption cross-sections and a simple Beer’s Law calculation we can determine that all three aerosols (oleic acid, DHB, and caffeine) are optically thin media. A particle is optically thin if the product of the particle’s radius, r, the absorption cross-section, σ, and the concentration, C, is << 1; in this scenario, the particle will be evenly heated.134 The IR absorption cross-sections of 2,4-dihydroxybenzoic acid, caffeine, and oleic acid were measured using a Bruker Equinox 55 FTIR and determined to be 2.38×10−20, 9.00×10−21, and 1.05×10−19 cm2/molec, respectively at 944 cm−1. Based on these numbers rσC << 1 for our experimental conditions and the particles should be uniformly heated by the IR pulse, ruling out front side ablation of the particle as described by Schoolcraft et al.146, 147 Additionally, as the wavelength of the CO2 laser is much larger than the particle radius, there are no self-focusing issues within the particle that can lead to uneven heating and ablation of the shadow side of the particle.57, 148 These arguments suggest the particles should undergo an isotropic expansion after interaction with the CO2 laser.  35  4.2.2.2 Ionization Following desorption, the gas phase species were ionized with a 70 eV electron impact pulse, which is generated by an electron gun located behind the end cap opposite the ion detector. The electron gun consists of a rhenium hairpin filament/cathode held at 0 V and a cathode shield held at -70 V. The ion trap end cap is held at 0 V with a 1 mm entrance hole, which serves to accelerate the electrons when the filament/cathode is pulsed to -70 V. The home built electron gun delivers about 50 - 100 µA total beam current. Based on comparative studies of a gas phase species (toluene) using EI and VUV laser radiation, the effective electron beam current inside the trap is estimated to be on the order of 4 µA. The width of the electron pulse can be varied which will result in enhancement or reduction of the ion signal. Unless otherwise noted an electron pulse width of ~4 µs was used for all experiments. The firing of the EI gun is triggered by particle velocimetry measurements to generate a “pulsed” electron beam that ionizes gas phase molecules produced from the desorption step. The timing scheme for desorption and ionization is detailed in Figure 4.3. The delay time between the CO2 laser firing and the EI gate opening is controlled by the data control/acquisition software and is referred to as “ionization delay time” subsequently in this work. Data can be collected at a fixed ionization delay time or the ionization delay time can be scanned.  36  CO 2 fires  -10  -5  0  5  10  10000 20000 30000 40000 50000 60000  -5  0  5  10  10000 20000 30000 40000 50000 60000  EI gate  open  close  RF Voltage [a.u.]  -10  Cleaning Trapping  Trapping Mass Ejec/Analysis -10  -5  0  5  10  10000 20000 30000 40000 50000 60000  Ionization delay time (time since CO 2 laser fires) (µs)  Figure 4.3: Timing schematic for CO2 laser firing, electron pulse from the electron gun and the RF voltage scan for the ion trap. Ionization delay times are referenced to the CO2 firing. At 0 µs delay, the electron gun gate opens as the CO2 laser is simultaneously discharging. The break in the x-axis scale corresponds to the cooling time used for ions in the trap (10 ms).  4.2.2.3 Mass analysis Once the CO2 laser and EI have fired, the ions are trapped in a custom built 3D ion trap that has been described before in detail.149-151 The ion trap electrodes have an ideal, non-stretched quadrupole geometry (ro = 10.00 mm, zo = 7.07 mm) and are housed in a vacuum manifold specifically designed for optical experiments. The vacuum manifold has four optical ports located in-line with the asymptotes of the ion trap. This allows two direct lines of sight diagonally through the end caps and ring electrodes. The ring electrode has two drilled holes for entry and exit of the particles. A particle beam dump is mounted at the exit hole in the ring electrode to collect excess/unanalyzed particles. The ion trap is operated in mass selective instability mode where an RF voltage is applied to the ring electrode to control the storage and ejection of ions.152 Ions created in  37  the ionization process are trapped by applying a trapping potential from a customized QMS power supply (Extrel Model 011-10; 968 kHz RF) to the ring electrode. Trapped ions are collisionally cooled in ~1 mTorr of He buffer gas inside the trap. Collisions with an inert gas buffer gas (He) reduce the kinetic energy of the ions and collapse the plume of ions into the center of the trap.153,  154  This results in a more efficient detection and  resolution of ion signal as the effectiveness of ion extraction is directly related to the position of ions in the trap prior to ejection.155 After the collisional cooling period (1 - 10 ms), the amplitude of the RF potential is increased causing a sequential destabilization in the axial motion of ions and ions are ejected axially through the holes in the end cap.155 Ions are detected by an electron multiplier (ETP 14138) behind the end cap. The ion signal is amplified by a current amplifier (Keithley 427) and recorded by a 16 bit ADC Card (PCI-MIO-16XE-10, National Instruments) that is also used to create the RF ramp. In mass selective instability mode, a mass range of 10 to 340 m/z is accessible for this instrument. It should be noted that background ions can build up in the trap between particle events, thus the trapping voltage is briefly dropped to zero after every mass scan to empty the trap (see Figure 4.3). Supplemental AC waveforms can be applied to the end caps to perform tandem mass spectrometry and/or conduct mass scans in resonant ejection mode, thereby extending the available mass range. The accessible mass range can currently be extended to about 900 Da in the SPITMS in resonant ejection mode, but this capability was not used in any of the reported experiments. Tandem mass spectrometry was implemented in some experiments by applying supplemental SWIFT (stored waveform inverse Fourier transform) waveforms to the ion trap end caps with in-house Labview software.156-158 Waveforms are generated by an arbitrary waveform generator (PCI-5412, National Instruments) and amplified and inverted by a custom power supply. After ion formation, trapping and cooling, a selected mass can be isolated by applying one or two notched broadband waveforms to the end caps. The isolated ions then undergo collision-induced dissociation with the helium buffer gas. Ions are resonantly excited such that their axial trajectories extend away from the center of the trap to regions of higher potential, which results in increased kinetic energies of the ions. Subsequent collisions with helium atoms  38  induce dissociation of the ions.159 Fragments are then mass analyzed to provide additional insight into molecular structure. For these studies, trapped ions were collisionally cooled for 10 ms in ~1 mTorr of He buffer gas and a mass scan was performed by linearly increasing the RF voltage. Mass scanning rates of 4000 Da/s were used for all the experiments shown here, which led to analysis times of about 48 ms per mass spectrum. Under these conditions, about 5 - 10 particles/s can be analyzed, depending on the aerosol concentration in the sample flow. Mass resolution at 264 Da was typically around 500 (m/∆m). The mass axis was calibrated daily by recording 70 eV EI spectra of perfluorotributylamine (calibration gas). Tandem mass spectrometry was implemented by applying two rounds of notched broadband waveforms for 5 ms and 10 ms with respective widths of 10 Da and 0.5 Da. The isolated ions then underwent collisional-induced dissociation with He buffer gas for 20 ms followed by subsequent analysis. This procedure was repeated to produce further spectral information (MS3).  4.3 Results and discussion 4.3.1 Mass analysis region characterization To evaluate the overall transmission of particles into the ion trap and validate the performance of the triggering system, mass spectrometer (MS) hit rates were recorded. The MS hit rate is defined as the ratio of the number of identifiable aerosol mass spectra to the number of aerosol triggers provided by the particle sizing region. In this work a given mass spectrum was classified as a valid aerosol spectrum if the integrated ion current at a particular mass was above a threshold level (generally 3σ of the background). To validate the detection geometry and timing scheme, the CO2 laser was operated at high energies (978 cm-1) to yield ions by laser ablation of 2,4dihydroxybenzoic acid aerosols. By acquiring MS hit rates as a function of the delay between the second scattering event and the firing of the CO2 laser (referenced as the TOF multiplier), a spatial profile of the CO2 laser beam can be obtained as seen in Figure 4.4. A Gaussian profile is observed as expected for a single mode laser and the measured width (1 mm after correcting for the 35.2º detection geometry) agrees very well with the 39  diameter of laser beam burns taken at the center of the trap (0.9 ± 0.2 mm). The fact that the maximum hit rate is over 90% confirms that both particle focusing and triggering work well. Given the proper CO2 alignment and sufficiently high ion yields, overall detection efficiencies will stay constant over the whole range of accessible sizes. Triggering delay (ms) 2.9250  2.9300  2.9350  2.9400  2.9450  2.9500  2.9550  # Valid MS/ Aerosol Triggers [%]  100  80  60  40  20  0 240.0  240.5  241.0  241.5  242.0  Distance from particle sizer (mm)  Figure 4.4: Spatial profile of MS hit rates for CO2 only ionization of 970 nm DHB aerosols with 22 mJ/pulse (~5x107 W/cm2) at 978 cm-1. By varying the delay between the desorption laser and the aerosol trigger (top X-axis) the aerosols are exposed to different segments of the spatial CO2 power profile, yielding a beam profile (bottom X-Axis). The experimental data could be fitted to a simple Gaussian profile, as expected for a single mode TM00 laser. Vertical beam deflection of larger particle sizes was also observed in the ion trap in addition to the sizing region. In fact, optimum mass spectrometer hit rates for larger particles depend on a slight vertical adjustment of the desorption laser. However, it should be noted that over a wide range of intermediate sizes (350 – 700 nm vacuum aerodynamic diameter) hardly any height adjustment of the desorption laser is necessary.  40  4.3.2 Mass spectral identification of aerosols Figure 4.5 shows normalized single particle and 200 particle average mass spectra of 2,4-dihydroxybenzoic acid vaporized with 14 mJ/pulse 944 cm-1 (10.6 µm) CO2 followed by 70 eV electron impact at the optimal ionization delay time (in this case 1 µs). No ions were obtained from the desorption step alone and no ion signal was observed from electron impact of particles without the desorption laser pulse applied first. The 70 eV NIST spectrum of 2,4-dihydroxybenzoic acid160 is also shown for comparison in Figure 4.5. The single particle mass spectrum (panel a) provides an easily identifiable spectral fingerprint for 2,4-dihydroxybenzoic acid. The oscillation in the baseline in this experiment was due to RF pickup, which has been subsequently eliminated. As can be seen, the CO2/EI molecular and fragment ions in both the single particle and average particle mass spectra match well with the literature fragmentation pattern, however, a higher intensity of low mass fragments are present in comparison to the literature.  1.0  Single particle  (a)  0.5  Ion Signal [a.u.]  0.0  1.0  Average (n=200)  (b)  0.5  0.0 1.0  Literature EI  (c)  0.5  0.0  40  60  80  100  120  140  160  m/z  Figure 4.5: Normalized mass spectra of (a) single particle and (b) 200 particle average of 2,4-dihydroxybenzoic acid (879 nm diameter) collected with 14 mJ 944 cm-1 CO2/EI, 1 µs ionization delay and (c) NIST EI standard.  41  Caffeine particles were also investigated with CO2/EI and demonstrated good correlation to the NIST spectrum160 as shown in Figure 4.6. A higher desorption energy (~30 mJ/pulse 944 cm-1) was applied as the threshold for onset of ions from the CO2 laser alone was greater in comparison to 2,4-dihydroxybenzoic acid (discussed later in the section on vaporization energy effects). Similar to the above observations for 2,4dihydroxybenzoic acid, a higher intensity of low m/z peaks are present in our measurements as compared to the literature spectrum. 1.0  (a)  Single particle  (b)  Average (n=200)  (c)  Literature EI  Ion Signal [a.u.]  0.5  0.0 1.0  0.5  0.0 1.0  0.5  0.0 60  80  100  120  140  160  180  200  m/z  Figure 4.6: Normalized mass spectra of (a) single particle and (b) 200 particle average of caffeine (887 nm particle) collected with 30 mJ 944 cm-1 CO2/EI, 1 µs ionization delay and (c) NIST EI standard. Since the EI spectra of gas phase analytes measured with an ion trap agree well with literature spectra providing that storage time is not prolonged,161 the increased levels of fragmentation observed in both the 2,4-dihydroxybenzoic acid and caffeine spectra suggest that the energy imparted from the laser desorption step is higher than that from the thermal desorption used for NIST standards. Woods et al.136 showed that with increasing CO2 energy, the fragmentation pattern shifts towards a larger percentage of  42  low m/z ions. Other studies have also demonstrated the effects of imparting increased internal energy during the desorption step on fragmentation patterns with both cartridge heaters and a CO2 laser.96, 162-164 Additionally, oleic and linoleic acid aerosols (both liquids) were briefly investigated with CO2/EI. Normalized single particle and 200 particle average spectra at 1 µs ionization delay and the NIST EI spectrum160 are shown in Figure 4.7 for oleic acid. Single particle and averaged spectra for linoleic acid were also collected and closely matched the EI NIST standard (not shown). The geometric particle diameter used for liquids (~1 µm) was larger than for solid particles as similar aerodynamic diameters were selected for comparison. Although oleic and linoleic acid differ only by 2 Da, the observed fragmentation ratios of the pure aerosols reflected unique mass spectral fingerprints that permit the two species to be distinguished. Molecular ions were not observed for either oleic and linoleic acid, which is similar to other studies of aliphatics fragmented by electron impact.26 1.0  (a)  Single particle  (b)  Average (n=200)  (c)  Literature EI  Ion Signal [a.u.]  0.5  0.0 1.0  0.5  0.0 1.0  0.5  0.0 60  80  100  m/z  120  140  160  Figure 4.7: Normalized mass spectra of (a) single particle and (b) 200 particle average of oleic acid (1.1 µm particle) collected with 15 mJ 944 cm-1 CO2/EI, 1 µs ionization delay and (c) NIST EI standard.  43  4.3.3 MSn performance The MSn capabilities of the CO2/EI SPITMS were demonstrated using caffeine particles. Figure 4.8 shows a MS/MS study of caffeine particles with an average of 200 particles displayed for each panel. In Figure 4.8a, the standard mass spectrum of ions produced from 30 mJ CO2 with EI is shown. This EI spectrum is comparable to work by Verenitch et al. with GC-MS (EI ionization) of caffeine.165 Ion fragments identified are listed: m/z = 194 is the parent ion M+, m/z =193 is [M-H]+, m/z = 165 is [M-CO-H]+, m/z = 137 is [M-CH3NCO]+, m/z = 109 is [M-CH3NCO-CO]+, and m/z = 82 is [M-CH3NCOCO-HCN]+. The parent peak at m/z = 194 was subsequently isolated (Figure 4.8b) with an average efficiency of 52%. Collision-induced dissociation (CID) by resonant excitation of m/z = 194 produced predominantly m/z = 193 as shown in Figure 4.8c, along with m/z = 165, 137 and 109. Further isolation and CID of the MS2 peak at m/z 193 produced a range of ions at m/z = 165, 149 [M-H-CO2]+, 134 [M-H-CO2-CH3]+, 122 [M-H-CO2-HCN]+ or [M-CH3NCO-CH3]+, 120 [M-H-CO2-H2CN]+, and 108 (which can stem from 2 fragments) as seen in Figure 4.8d. The overall efficiency for MS3 is approximately 15%. It is worth noting that the MS/MS spectrum collected by Verenitch et al. for caffeine is an amalgamation of the spectra produced in Figure 4.8c and d, indicating their process likely imparted more energy during CID of the parent ion than the waveform used in this study.165  44  1.8 1.2  193  0.6  165  137  0.0 0.8  Ion signal [a.u.]  194  (a)  109  82  (b)  194  (c)  193  (d)  193  0.4 0.0 0.6 0.4 0.2 0.0 0.06  108  0.04  120 122 132  149  0.02 0.00  80  100  120  140  m/z  160  180  200  Figure 4.8: A sequence of MS/MS studies on caffeine aerosols at 30 mJ CO2/EI and 1 µs ionization delay. (a) caffeine aerosol MS, (b) isolation of parent ion (m/z = 194), (c) fragmentation of parent ion via collision-induced dissociation, (d) subsequent fragmentation of daughter ion (m/z = 193).  4.3.4 Effect of operational parameters on mass spectra 4.3.4.1 Dependence of total ion signal on laser desorption energy Mass spectra of pure 2,4-dihydroxybenzoic acid (1 µm diameter) and caffeine (887 nm diameter) particles were collected using CO2/EI over a range of CO2 laser powers. The respective energy ranges used for each species were limited to energies that did not produce ions from the desorption step alone. Figure 4.9 shows the total ion signal (summation of molecular and main fragment ion peaks) for each species integrated over the portion of the ionization delay time profile (scan of total ion signal versus ionization delay time) where S/N is above the detection limit. Each point represents an average of 300 shots. The total ion signal for 2,4-dihydroxybenzoic acid plateaus at 10 - 12 mJ/pulse, while the signal for caffeine is not observed to plateau within the range of energies studied. (It should be noted that a previous study of an 879 nm DHB particle 45  reached a plateau in total ion signal in the same range as the 1 µm particle). The IR absorption cross-sections (see section.4.2.2.1) measured at 944 cm-1 for the two species (DHB = 2.38x10-20 cm2/molec and caffeine = 9.00x10-21 cm2/molec) are consistent with DHB reaching a plateau at lower CO2 energies than caffeine. The DHB plateau of total signal with laser vaporization energy suggests that beyond a certain desorption energy or threshold, the total signal remains constant regardless of CO2 energy and may indicate near-complete or complete vaporization.93,  99  It should be noted that the degree of  fragmentation observed in the mass spectra could also change as a function of CO2 energy as shown in Chapter 5 using Version II of the SPITMS; however, the degree of fragmentation change with CO2 energy will likely be far less in Version I as electron impact ionization is a much more energetic ionization process.  100  (a)  Total Ion Signal [a.u.]  90 80 70 60 50 40 30 20 10 140  Total Ion Signal [a.u.]  130  0  5  10  15  20  25  30  (b)  120 110 100 90 80 70 60 50 15  20  25  30  35  CO 2 energy (mJ/p)  Figure 4.9: Total ion signal as a function of CO2 energy (944 cm-1) for (a) 2,4dihydroxybenzoic acid [m/z 154, 136, 108, 95, 80, 69, 64, 53] and (b) caffeine [m/z 194, 193, 109, 82, 81, 67, 55].  46  4.3.4.2 Dependence of total ion signal on ionization delay time (time between firing the CO2 laser and the electron pulse) The results from scanning the ionization delay time are illustrated in Figure 4.10 for two sizes of 2,4-dihydroxybenzoic acid particles. In this mode, the onset of ion signal is observed as the electron pulse (~4 µs long) overlaps the firing of the CO2 laser, rising to a maximum signal at the optimal ionization delay time (time between the desorption and ionization steps) and then subsequently decaying as the gas phase molecules expand outside of the ionization volume. A signal at a negative ionization delay time merely reflects the ionization delay being defined relative to the CO2 laser firing and the fact that the ionization event is ~4 µs long. The maximum signal generally occurs at a 0 or 1 µs delay with slight variability based on desorption energy. This is the expected behavior for our system based on the electron gate width and the expansion of the vapor plume in the trap. The delay profile observed is roughly comparable to that obtained by Cabalo et al.77 2.5  dve = 737nm dve = 985nm  Total Ion Signal [a.u.]  2.0  1.5  1.0  0.5  0.0 -5  0  5  10  15  20  25  Ionization delay time (µs)  Figure 4.10: Total ion signal [m/z 154, 136, 108, 80, 69] versus the ionization delay time (time between the desorption laser firing and the electron pulse) (14 mJ 944 cm-1) for 2,4-dihydroxybenzoic acid of two different particle sizes (dve shown).  47  4.3.4.3 Dependence of total ion signal on electron gate pulse width Figure 4.11 illustrates the ion signal at 136 m/z for DHB as a function of ionization delay time at three different gate widths, all using 14 mJ CO2 for vaporization. Increasing the gate width enhances the ion signal to a maximal point after which no further effect on ion signal is achieved. Conversely, increasing the gate width reduces the time resolution of the ionization delay time scans such that the temporal characteristics of the vapor expansion profile are obscured. The loss of temporal information would not be problematic for the analysis of atmospheric particles and the use of longer gate width would in fact be more ideal for collection of field data. Although the gate width could be expanded to 10 - 13 µs to increase signal and enhance our detection limit for routine measurements, the 4 µs gate width was used in these experiments to allow us to investigate the profile of the neutral expansion post-desorption and minimize any signal from background residual gases. Also shown in Figure 4.11 is the ionization delay time scan from a CO2/VUV study of 2,4-dihydroxybenzoic acid in the SPITMS version II setup described in Hanna et al.110 In this case, the ionization delay time refers to the time between the CO2 laser firing (15 mJ 944 cm-1 desorption) and the VUV laser firing (144 nm). The VUV laser pulse is 5 ns long which provides excellent time resolution. As can be seen, the full-width half maximum of the VUV profile is ~1 µs shorter than that for the 4 µs EI gate width. This implies that at standard settings the time resolution of the CO2/EI experiment is nearly comparable to the CO2/VUV. In fact, a 4 µs running average of the scaled VUV profile exactly matches the EI profile, further supporting this conclusion.  48  3.8 µs gate width 10.2 µs gate width 21.6 µs gate width VUV delay profile  Ion Signal [a.u.]  1.0  0.5  0.0 -15  -10  -5  0  5  10  15  20  Ionization delay time (µs)  Figure 4.11: Effect of EI gate width on ion signal (m/z 136) as a function of ionization delay time for DHB. Also shown is a high resolution ionization delay time profile obtained using CO2/VUV of DHB.166 The VUV profile has been adjusted in time so that maximum signal overlaps with the maximum of the EI profile.  4.3.4.4 Dependence of degree of fragmentation on ionization delay time Figure 4.12 illustrates the ratio of individual fragment ion signal to total ion signal for caffeine (887 nm diameter particle) as a function of ionization delay time. Each data point is an average of 300 shots. At delay times between -5 and -2 µs, the relative error is significant and S/N values are meaningless due to the low number of valid particle spectra collected, if any are observed at all. For this reason, a restricted range is selected for the integrated delay scan analysis where the S/N ratio is above the detection limit and relative error is low. Similar to previous work, we observe that the degree of fragmentation depends on the ionization delay time.136, 163 The percent of molecular ion (m/z = 194) increased with longer delay times while fragment peaks at m/z 82, 67, and 55 decreased in intensity. The change in intensity was only significant for the m/z 67 and 194. A parallel effect was observed for 2,4-dihydroxybenzoic acid. As Figure 4.12 shows, at longer ionization delay times, fragmentation could be minimized, however the significance of such an effect may  49  depend on particle composition and comes at the expense of the total ion signal. The change in percent abundance of the parent ion is less significant at low vaporization energies for both species studied. Some explanations for the dependence of fragmentation on delay time have been previously discussed by others in significant detail136,  167  and  will not be included here. Another possibility for minimizing fragmentation is adjusting the energy level of the electrons to less than 70 eV, but this would also come at a cost of signal which would in turn limit the detection of single particles.  Ratio of IFragment to Total Ion Signal  194m/z 82m/z  193m/z 67m/z  109m/z 55m/z  0.3  0.2  0.1  0.0 -4  -2  0  2  4  6  8  10  12  14  16  18  20  22  Ionization delay time (µs)  Figure 4.12: Ratio of molecular or fragment ion signal to total ion signal for caffeine as a function of ionization delay time at 35 mJ 944 cm-1 CO2/EI. Error bars represent 1σ.  4.3.5 Detection limit We have acquired CO2/EI spectra from a range of DHB particle sizes down to 350 nm diameter, where MS signal is still distinguishable from background signal (S/N is ~3). Given the S/N ratio, we appear to be virtually at our detection limit for this molecule. This size corresponds to a detection limit of ~1x108 molecules (~26 fg). This number is reasonably consistent with the gas phase detection limit determined for EI in the ion trap in separate experiments using toluene as a calibration gas (8x107 molecules). 50  The Aerodyne aerosol mass spectrometer reported single particle detection limits with a 100% counting efficiency of > ~2x108 (20 fg) for NH4NO3 using thermal desorption/EI in a quadrupole mass analyzer.79 Although the species studied were not the same, the two detection limits are very comparable. A newer version of the Aerodyne instrument produces single particle mass spectra of sufficient quality for identification for particles > 350 nm dva, which depending on the density could give a range of physical particle diameters. For DHB, this would give a limit of ~280 nm physical diameter168 which is slightly smaller than our measured limit of 350 nm physical diameter, despite their much improved electron beam current (see below). The electron gun used for these experiments was originally designed for ionizing the calibration gas (perfluorotributylamine). The electron gun has not been optimized for beam current or beam shape nor has the electron beam produced been thoroughly characterized. In the Aerodyne instrument mentioned above, the total electron beam current is generally 2.5 mA162 in comparison to our estimated 50 - 100 µA, which suggests there is significant room for improvement in the detection limit of Version I with optimization of the electron gun. Additionally, since the electron gun is recessed behind one of the end caps, the RF field undoubtedly distorts the electron beam produced. The ultimate detection limit of this technique could potentially be improved by a superior filament shape (minimizing the emitting area), changing the filament height, better alignment through the trap, increasing the duration of the electron pulse (increasing electron flux by a factor of 2) or enhancing the brilliance of the electron pulse. Varying the anode voltage as well as minimizing the electron beam focus169 may also help optimize the peak current obtained.  4.4 Conclusions In this chapter, the first results for single particle mass spectrometry utilizing laser desorption-electron impact (CO2/EI) in a 3D ion trap are discussed. It was shown that version I of the SPITMS can provide spectra comparable to the NIST spectral library, which may facilitate species identification in the field. Increased levels of fragmentation were observed for 2,4-dihydroxybenzoic acid and caffeine particles in comparison to the  51  NIST spectra. This suggests that the energy imparted in the laser desorption step is higher than that from the thermal desorption used for NIST standards, which could negatively impact analysis of mixed particles. Tandem mass spectrometry up to MS3 was successfully performed for caffeine aerosols, showcasing the ability for identification of a molecular fragment fingerprint. The effect of various parameters such as desorption energy, ionization delay time, and electron pulse gate width on total ion signal obtained were explored. A higher desorption energy resulted in larger total ion signal, presumably as a larger fraction of the aerosol is being vaporized. The scan of total ion signal as a function of ionization delay time allows for a rough examination of the vapor plume dynamics. An ionization delay time of 1 µs provided the maximum total ion signal for the aerosols studied. This optimal delay may vary slightly with aerosol composition. The electron pulse gate width could also be expanded to enhance total ion signal at the expense of resolution in the ionization delay time scan. The extent of fragmentation was also explored as a function of ionization delay time and found to decrease with increasing delay times as observed by other groups, but the overall percent change in molecular ion as a function of ionization delay time is also likely to depend on the stability of the molecule and the type of ionization used. Lastly a detection limit of ~1x108 molecules DHB (350 nm particle) is determined for version I of the SPITMS under current operating conditions. This limit may be lowered by optimization of the electron gun in future experiments. The CO2/EI SPITMS is the first implementation of single particle mass spectrometry with EI in an ion trap mass spectrometer, and the first time CO2 desorption has been used in combination with EI for aerosol studies. The technique is fairly robust, provides spectra comparable to the NIST spectral library, allows for in-trap ionization and increased sensitivity, and reduces the complexity of a pulsed two step laser vaporization/ionization process. This instrument could be complementary to the commercially available ATOFMS and Aerodyne instruments for both laboratory and field studies. The usefulness of this instrument for studying complex or more atmospherically realistic aerosols will be discussed in Chapter 7.  52  Chapter 5. One and Two Component Aerosol Studies with Laser Desorption – Vacuum Ultraviolet Ionization in the Single Particle Ion Trap Mass Spectrometer (Version II)3 5.1 Introduction Recently we have developed a single particle mass spectrometer that incorporates a CO2 laser for vaporization, a tunable laser-based VUV source for ionization, and an ion trap mass spectrometer for mass analysis (version II) with the same interface characterized in Chapter 4. The instrument was initially characterized using single component particles by Hanna et al.110, 166 The introduction of this chapter will briefly review some of the challenges to organic aerosol analysis and the advantages that a soft ionization source, particularly tunable VUV, provides for organic aerosol studies followed by an overview of the work in this chapter. One of the challenges in organic aerosol analysis is the fragility of the molecules under study. One step laser desorption/ionization requires high laser powers which in turn causes extensive fragmentation of organic species through successive absorption of several photons and charge-transfer matrix effects.77 A solution to this is to separate the vaporization and ionization steps51 and use a “soft” ionization source to reduce fragmentation of many organics.26, 162, 166 By using a “soft” ionization source, ions are generated from the vaporized aerosol neutrals with a minimum of excess energy. This reduces the extent of fragmentation and thus simplifies the chemical characterization of the aerosol. Ideally, mass spectra containing only molecular ion peaks would be obtained making molecular identification more straightforward. Given the numerous species present in any atmospheric aerosol, reduction in the complexity of the mass spectra is necessary to provide identification of most organic aerosol species. To date, several research groups have implemented soft ionization sources for aerosol mass spectrometry. Some examples include metal attachment,26 chemical  3  A version of this chapter will be submitted for publication. Simpson, E.A., Campuzano-Jost, P., Hanna, S.J., et al.  53  ionization,38-40, 170 attachment of low-energy photoelectrons (PERCI),50, 102, 171 resonance enhanced multiphoton ionization (REMPI),51,  56, 81, 92, 94, 98, 115, 172  and single photon  ionization.93, 99, 100, 117, 136, 162, 164, 173 A tunable laser-based VUV source is advantageous for several reasons. First, it is a pulsed ionization source which permits single particle studies. Second, single photon ionization (SPI) is suitable for both aliphatic and aromatic compounds whereas resonance enhanced multiphoton ionization (REMPI) is a more selective technique primarily suited for aromatic species since it relies on an intermediate electronic resonance.77,  174, 175  Third, the tunability allows for the photon energy to be selected slightly above the ionization energy of the molecule of interest, which should result in little or no fragmentation in the parent molecule and allow for more accurate identification of an aerosol’s components.176 The tunable source also makes it possible to measure a species’ ionization energy (IE). Additionally, the tunable source can be used to provide selective ionization of a species in multi-component particles by tuning the ionization energy.93, 101 Having both the IE and the SPI mass spectrum of the molecule is a powerful combination for product identification in analytical applications. In the following, we carry out detailed studies of pure oleic acid, pure oleyl alcohol, and mixed oleic acid:oleyl alcohol particles using version II (CO2/VUV) of the single particle ion trap mass spectrometer. This builds on the initial characterization by Hanna et al.110, 166 These investigations are performed to determine the ideal operating parameters of our instrument and any strengths or weaknesses of the two laser approach. Many of the phenomena observed here are a product of the CO2 laser vaporization; thus, our results are relevant to other systems with two step laser desorption/ionization since most of these systems use a CO2 laser for desorption. The three specific questions we are trying to address are detailed below. Question 1: How does the fragmentation vary with CO2 energy and particle composition for one and two component systems? This is critical to interpreting mass spectra from our instrument and also for determining if our system will have a linear response with particle composition. For example, if the fragmentation pattern of one species in a mixed particle changes as function of the particle composition, one could not  54  expect to get a linear response for that species by evaluating the ion signal from a singular molecular or fragment ion. Question 2: How does the ionization delay profile vary with CO2 energy and composition for one and two component systems? The ionization delay profile is a scan of total ion signal as a function of the delay time between the firing of the desorption laser and ionization laser. Understanding the ionization delay profiles is important for optimal instrument performance particularly in a field instrument if only one delay time can be used. Question 3: For a two component system (when the absorption cross-sections are significantly different), does the mass spectrum show a linear response with composition? The results from this study will help determine the usefulness of version II of the instrument for future quantitative and kinetic studies where a quantitative response is required. These results will also apply to similar instruments that use a two step vaporization/ionization process with a CO2 laser for desorption. This study will additionally help identify some of the potential weaknesses of a two laser system for quantitative analysis of multi-component particles. There have not been many studies related to the questions outlined above. Studies of the effects of CO2 power on fragmentation have been previously conducted for single particles of oleic acid,93, 163 ethylene glycol,93, 136 and aniline77 with CO2/VUV in a TOFMS. Hanna et al. investigated the same effect for oleic acid, 2,4-dihydroxybenzoic acid, and caffeine in an ion trap.110, 166 There have been no previous studies of the change in fragmentation as a function of composition for mixed particles. Studies of the ionization delay profile as a function of CO2 energy and/or composition have been done for single component particles of ethylene glycol,136 aniline,77 oleic acid,166 and 2,4-dihydroxybenzoic acid.166 To our knowledge, no one has explored the effect of changing particle composition on ionization delay profiles (or noted such an effect) in a mixed particle system. Lastly, there has only been one study investigating the linear response with particle composition for mixed particles using CO2/VUV. Baer and coworkers explored the quantitative use of CO2/VUV for four varying particle compositions of a three component polycyclic aromatic hydrocarbon (PAH) mixture.99 These aromatic species  55  are reasonably stable under fragmentation and primarily produce molecular ions, which makes the overall analysis more straightforward. As mentioned above, we studied pure oleic acid, pure oleyl alcohol, and mixed oleic acid:oleyl alcohol particles to further characterize version II of our SPITMS and determine the optimal working conditions and the strengths and weaknesses of the technique. These two species are not as robust as PAHs and tend to undergo more fragmentation,163, 164 particularly in the ion trap system where storage time is longer166 (see below for further discussion). Oleic acid is present in the atmosphere at concentrations of ~1 ng/m3 in the particle phase and is known to originate from a variety of sources, including meat cooking operations.12 Oleyl alcohol is structurally comparable to oleic acid with the carboxylic acid group being replaced by an alcohol. The two compounds are liquids and miscible in each other. This should provide a test of the system’s ability to resolve species that fragment similarly. Additionally, as the absorption cross-sections of the 2 species at 1056 cm-1 (the CO2 wavelength utilized) were significantly different (a factor of ~4), the system’s ability to carry out quantitative measurements of mixtures with significantly different CO2 absorption cross-sections is investigated. The two component system was studied at two CO2 desorption energies of 10 mJ/pulse and 7 mJ/pulse (1056 cm-1). At 10 mJ/pulse, both species fragmented extensively and it was impossible to separate the mass spectral peaks of the individual components; hence this chapter focuses on results from the 7 mJ/pulse case. Specifically this chapter shows that the fragmentation and ionization delay profile changes observed for both the one and two component systems studied here can be described by a single key parameter: the energy absorbed per particle in the desorption step. Lastly we demonstrate how the dependence of fragmentation on particle composition affects the linear response of the system.  5.2 Experimental 5.2.1 Particle generation Oleic acid (Fluka, ≥99.0%) and oleyl alcohol (Aldrich, 99%) were used as purchased without further purification. Solutions were prepared in 2-propanol (Aldrich, 56  99.9%) in the concentration range of 10-5 g/mL and particles were generated using a vibrating orifice aerosol generator (TSI Inc., Model 3450). Particles were passed through a 85Kr charge neutralizer (TSI Inc., Model 3054) with ~25 LPM dilution air flow to dry them completely before entering the aerosol mass spectrometer. This prevented additional drying and size changes in the aerosol lens. As the vibrating orifice aerosol generator produces a narrow size distribution, particles were sampled directly into the instrument without being size-selected by a DMA. Particles were ~1.2 µm in diameter. Figure 5.1 shows the chemical structures of oleic acid (left) and oleyl alcohol (right). Relevant properties for each of the species are given in Table 5.1.  HO  OH  O  Figure 5.1: Chemical structures of oleic acid (left) and oleyl alcohol (right). Table 5.1: Relevant properties of oleic acid and oleyl alcohol. Compound  M.W. (g/mol)  Density (g/cm3)  Oleic acid  282.4614  0.895  Oleyl alcohol  268.4778  0.850  a  IE (eV) 8.68164 8.65 ± 0.05166 <9.14a 8.56 ± 0.05b  based on energies given for 2-propen-1-ol and 2-butene-1-ol from NIST Chemistry Webbook177 and trends shown in Adam et al.178 for ionization energy as a function of increasing hydrocarbon chain length. b this work (Appendix II)  5.2.2 Overview of version II of the SPITMS The aerosol mass spectrometer used in these studies is shown schematically in Figure 5.2. The interface and mass spectrometer are the same as described in Chapter 4. The only difference is the VUV source used for ionization. Since many of the components were discussed in detail earlier, they will only be briefly reviewed here. The three basic sections of the instrument are listed below:  57  •  An aerosol inlet where an aerodynamic lens system based on the design of Liu et al.66,  67  is used to sample particles from atmosphere into vacuum and  collimate particles between 0.1 - 1 µm diameter into a tightly focused beam. •  A particle sizing and trigger source region based in part on the work by Su et al.81 Particle velocity is determined by the transit time of aerosols through two focused, continuous wave laser beams and used to extract particle size. Custom, real-time data acquisition software records the timing and generates the appropriate triggers for the desorption/ionization steps and the mass analysis.  •  A mass analysis region comprised of a mid-IR CO2 laser for particle desorption, VUV light for ionization, and the ion trap mass spectrometer.  58  Particle Beam 2.0 mm IR  VUV 36° 0.7mm  Figure 5.2: Instrument schematic for version II (CO2/VUV) of the SPITMS. The inset shows the paths of the particle beam and the IR and VUV lasers through the ion trap.166  5.2.2.1 Desorption In these studies, between 4 - 30 mJ of IR light from a pulsed TEA-CO2 laser (MTL-3G, Edinburgh Instruments) at 1056 cm-1 was used to vaporize the particles. None of the energies used produced ions from the desorption step alone. The average CO2 laser power is measured at the exit of the trap with a thermal detector (Ophir Model 3A-SH). Particles are optically thin with respect to the IR energy in these experiments and are therefore expected to be uniformly heated by the CO2 laser pulse. A particle can be 59  defined as optically thin if the product of the radius, r, the absorption cross-section, σ, and the concentration, C, is << 1.134 The IR absorption cross-sections of oleic acid and oleyl alcohol (solutions prepared in dichloromethane) were measured using a Bruker Equinox 55 FTIR. At 1056 cm-1, the absorption cross-sections were measured to be 2.85x10-20 cm2/molec and 1.03x10-19 cm2/molec, respectively for oleic acid and oleyl alcohol. Based on these numbers, rσC<<1 for the experimental conditions. Additionally, the particle size is small in comparison to the CO2 wavelength making it unlikely that internal focusing of IR light will give rise to a temperature gradient in the particle.57 After the particles were vaporized, neutral molecules in the expanding vapor plume were ionized with vacuum UV light from a custom-built tunable VUV source. A variable ionization delay time of 1 - 30 µs was used. The timing schematic is illustrated in Figure 5.3. Similar to scanning the delay time between the desorption laser and the EI pulse as described in Chapter 4, the delay time between the two laser pulses can be scanned to obtain a profile of the expanding plume (total ion signal versus ionization delay time) from the vaporized aerosol. This profile is referred to as the ionization delay profile and allows us to determine the optimal time for maximum ion signal.  CO 2 fires  -10  -5  0  5  10  10000 20000 30000 40000 50000 60000  VUV fires  -5  RF Voltage [a.u.]  -10  0  5  10  10000 20000 30000 40000 50000 60000  Cleaning Trapping  Trapping Mass Ejec/Analysis -10  -5  0  5  10  10000 20000 30000 40000 50000 60000  Ionization delay time (time since CO 2 laser fires) (µs)  Figure 5.3: Timing schematic for CO2 laser firing, VUV laser firing, and the RF voltage scan for the ion trap. At 0 µs delay, the VUV fires simultaneously with the CO2 laser discharge.  60  5.2.2.2 Ionization The novel, tunable, laser-based VUV source design and characterization were the focus of another thesis,179 thus the source will only be described briefly here. Continuously tunable laser-based vacuum ultraviolet light is produced by resonance enhanced four wave difference mixing in xenon gas. Two of the applied frequencies (ωUV) are set to reach a two-photon resonance in the medium (Xe gas), while the third applied frequency (ωVis/IR) is tunable and allows tunable VUV to be generated as shown in Figure 5.4.180  Xe+ Xe* ωUV  ωVis/IR  ωUV  ωVUV Xe  Figure 5.4: Schematic of resonance enhanced four wave difference mixing in Xe. For this apparatus, an Nd:YAG laser operating at 355 nm pumps both a dye laser and an optical parametric oscillator (OPO) laser as shown in Figure 5.2. The dye laser is used at a fixed frequency (providing the UV wavelength of either 222 nm or 249 nm) and the OPO provides the tunable output of 480 - 1250 nm. Both laser beams are focused into the gas cell to generate tunable, coherent VUV light. VUV light from 10.2 eV (122 nm) to 7.4 eV (168 nm) can be generated with 1010 - 1013 photons/pulse depending on the wavelength.110 A tunable custom monochromator based on a single MgF2 lens (see Figure 5.2) separates the remaining visible/IR and UV light from the desired VUV beam, and maintains a tight and precisely focused beam with high spectral purity over the tunable wavelength range.110 The VUV is refocused into the trap with a parabolic mirror to a slightly vertically elongated spot with an area of ~1 mm2. Data can be acquired in a fixed VUV wavelength mode or a scanning VUV wavelength mode. The fixed wavelength mode uses VUV optimized at a single photon  61  energy. The scanning VUV mode is optimized for a range of wavelengths as the wavelength is scanned continuously over tens of nanometers. This set of experiments utilized the VUV in the fixed wavelength mode where conditions were optimized for generating 142 nm light. A fast phototube (Hamamatsu R1328U-54) was used to measure the single-shot energy of the VUV beam at the exit of the trap. This way, the ion signal could be normalized to fluctuations in the shot-to-shot energy of the VUV beam as recorded by the FPGA board. A linear response to changing VUV intensity has been previously proven in the instrument, which makes it possible to normalize the ion signal to the VUV photon flux.110 The vaporization and ionization laser beams are aligned through the diagonal paths between the ring and end cap electrodes (inset of Figure 5.2). It should be noted that in the current setup the paths of the vaporization and ionization lasers and the path of the particle beam do not actually intersect in the center of the ion trap. The distance the aerosols travel between the two laser beams is on the order of 2 mm.166 This means that no ion signal will be observed if the ionization pulse fires immediately after the CO2 laser (e.g. 0 µs ionization delay time).  5.2.2.3 Mass Analysis Once the CO2 laser and VUV laser fire, trapped ions are collisionally cooled for 10 ms in ~1 mTorr of He buffer gas and a mass scan is performed by linearly increasing the RF voltage (mass selective instability mode) as described in section 4.2.2.3. The ion trap is emptied after every mass scan and laser pulse to avoid the buildup of background ions in the trap. Mass spectra were collected from 50 – 300 m/z for pure component particles and 30 – 300 m/z for mixed particles. Mass scanning rates of 4000 Da/s were used for all the experiments shown here. The mass axis is calibrated daily by recording 70 eV EI spectra of perfluorotributylamine. For the data in this chapter, ionization delay profiles (scan of total ion signal as a function of time between desorption and ionization) were obtained. The mass spectra results shown are an average of several hundred shots at the delay time which gave the  62  maximum total ion signal. For the exploration of effects on fragmentation ratios, the integrated values over the entire delay profile were used.  5.2.3 Previous studies by Hanna et al. Hanna et al. carried out preliminary studies on single component particles of 2,4dihydroxybenzoic acid, caffeine, and oleic acid using the CO2/VUV SPITMS, which are discussed briefly in this section.110, 166 It was determined that using a soft VUV ionization source reduced fragmentation compared to the EI source as demonstrated in Figure 5.5.166 For stable molecules like caffeine, relatively little fragmentation (e.g. the molecular ion was the dominant peak) was observed even at high desorption energies.166 However for less stable molecules like oleic acid, even at low desorption energies and a VUV photon energy just above threshold, significant fragmentation was observed as can be seen in Figure 5.5.166 The degree of fragmentation in the ion trap was observed to be considerably higher than that seen with TOF systems, especially for oleic acid. This was attributed to the long storage interval in the ion trap which allows time for metastable ions to decay.166 For comparison, the half-life for decay of metastable oleic acid ions is on the order of 10 µs, while the storage time (cooling time) in the ion trap is 10 ms and the total storage time and analysis time is on the order of 50 ms.  63  Ion Signal (a.u.)  1  142 nm (8.75 eV)  0 1  Electron Impact  0 50  100  150  200  250  300  m/z  Figure 5.5: Normalized mass spectra of oleic acid aerosols (300 shot average) ionized with (top) 7 mJ 944cm-1 CO2 142 nm VUV and (bottom) 10 mJ 944 cm-1 CO2 EI. Adapted from Hanna et al.166  5.3 Results 5.3.1 Pure oleic acid particles 5.3.1.1 Mass spectra as a function of CO2 energy Particle mass spectra were acquired for pure one component particles of oleic acid using a range of CO2 energies at 1056 cm-1 with 142 nm VUV. Figure 5.6 shows the normalized mass spectra from 7 – 24 mJ/pulse. No ions from the CO2 laser alone were observed over the entire range of energies. As expected, due to the increasing internal energy, the mass spectra show a shift towards lower m/z fragments with increasing vaporization energy.  64  1 55  148 83  7 mJ IR  127 134  246 235  0 1  264  282  10 mJ IR  Ion Signal [a.u.]  0 1  14 mJ IR  0 1  17 mJ IR  0 1  24 mJ IR  0 50  75  100 125 150 175 200 225 250 275 300  m/z Figure 5.6: Normalized mass spectra (300 shot average) of oleic acid at 7, 10, 14, 17 and 24 mJ/pulse of CO2 (1056 cm-1) and 142 nm (8.75 eV) VUV.  65  5.3.1.2 Dependence of degree of fragmentation on CO2 energy Selected peaks of oleic acid were used to assess the degree of fragmentation as a function of CO2 energy. Each fragment’s relative intensity is determined with respect to the total ion signal. As can be seen in Figure 5.7, the general trend reflects the decrease in high mass peaks and increase in lower mass fragments with increasing desorption energy. 56 m/z 134 m/z  Ratio of Fragment to Total Ion Signal  0.25  83 m/z 264 m/z  127 m/z 282 m/z  0.20  0.15  0.10  0.05  0.00 6  8  10  12  14  16  18  20  22  24  CO 2 energy (mJ/pulse)  Figure 5.7: Relative intensities of fragment peaks compared to the total ion signal as a function of desorption energy for pure oleic acid particles.  5.3.1.3 Dependence of total ion signal on ionization delay time for a range of CO2 energies Figure 5.8 shows the ionization delay profiles obtained for oleic acid over a range of desorption energies. The shape of the profile is primarily dependent on the distance between the two laser beams and the translational energy of the vaporized molecules which expand outward from the particle, filling, and then passing beyond the ionization volume. As can be seen, the expansion speed of the desorbed aerosol plume increases with increasing CO2 laser energy, with the maximum total ion signal shifting to shorter delay times.  66  7mJ 17mJ  8mJ 20mJ  10mJ 24mJ  14mJ  Total Ion Signal [a.u.]  1.0  0.5  0.0 0  5  10  15  20  25  30  Ionization delay time (µs)  Figure 5.8: Total ion signal as a function of ionization delay time for oleic acid. Each trace is for a different CO2 pulse energy (at 1056 cm-1) as indicated in the legend (VUV at 142 nm). All traces have been normalized to make comparison easier.  5.3.2 Pure oleyl alcohol particles 5.3.2.1 Mass spectra as a function of CO2 energy Particle mass spectra were acquired for pure one component particles of oleyl alcohol using a range of CO2 energies as shown in Figure 5.9. Increasing the desorption energy resulted in increasing levels of fragmentation. Note that no molecular ion is observed at any desorption energy used and the fragmentation is quite extensive even at the lowest vaporization energy used.  67  1  96 124  4 mJ IR 194  67  250  0 1  7 mJ IR  Ion Signal [a.u.]  0 1  10 mJ IR  0 1  20 mJ IR  0 1  30 mJ IR  0 50  75  100 125 150 175 200 225 250 275 300  m/z Figure 5.9: Normalized mass spectra (300 shot average) for oleyl alcohol at 4, 7, 10, 20, and 30 mJ/pulse CO2 (1056 cm-1) and 142 nm (8.75 eV) VUV.  68  5.3.2.2 Dependence of degree of fragmentation on CO2 energy The relative intensities of selected fragments were compared to the total ion signal as an indication of the degree of fragmentation for oleyl alcohol as a function of desorption energy. Figure 5.10 shows the same trend observed for oleic acid with a decrease in high mass fragments and increase in low mass fragments as desorption energy increases. 250 m/z 194 m/z 124 m/z 96 m/z 67 m/z  Ratio of Fragment to Total Ion Signal  0.20 0.18 0.16 0.14 0.12 0.10 0.08 0.06 0.04 0.02 0.00 5  10  15  20  25  30  CO 2 energy (mJ/pulse)  Figure 5.10: Relative intensities of fragment peaks compared to the total ion signal as a function of desorption energy for pure oleyl alcohol particles.  5.3.2.3 Dependence of total ion signal on ionization delay time for a range of CO2 energies Figure 5.11 shows the ionization delay profiles obtained for oleyl alcohol at a range of desorption energies. The increasing expansion speed of the desorbed aerosol plume can be observed with increasing CO2 laser energy.  69  3mJ 15mJ  4mJ 20mJ  7mJ 10mJ 30mJ  Total Ion Signal [a.u.]  1.0  0.5  0.0 0  5  10  15  20  25  30  Ionization delay time (µs)  Figure 5.11: Total ion signal as a function of ionization delay time for oleyl alcohol. Each trace is for a different CO2 pulse energy (at 1056 cm-1) as indicated in the legend. All traces have been normalized to make comparison easier.  5.3.3 Two component particles of oleic acid and oleyl alcohol Mixed particles of oleic acid and oleyl alcohol were prepared by aerosolizing solutions (using the VOAG) of weight percent 90:10, 75:25, 50:50, and 25:75 oleic acid:oleyl alcohol in isopropanol with the total mass concentration of the solution remaining unchanged (2x10-5 g/mL). Since the molar masses of oleic acid and oleyl alcohol are very similar, the mole ratios vary from the weight ratios by at most 1%. The vibration frequency of the VOAG was adjusted to get nearly the same time of flight (within 1.5%) for each particle composition. Particles of the same aerodynamic diameter were used to provide an optimal laser/particle beam alignment with both the scattering lasers and desorption and ionization lasers that required no adjustment between particle compositions. All results discussed for the two component particles were obtained with 7 mJ/pulse desorption energy, where oleic acid has distinctive high mass peaks (m/z = 264, 246, 148, 134, 127) that do not overlap with oleyl alcohol fragments. Due to the fragmentation of metastable ions observed in the ion trap previously,166 we don’t see  70  significant molecular ions for the analytes chosen in this study (see Figure 5.6 and Figure 5.9). This adds to the complexity of analysis.  5.3.3.1 Mass spectra as a function of composition of mixed particles Normalized mass spectra are shown as a function of particle composition in Figure 5.12. The pure particle spectra for oleic acid and oleyl alcohol are distinctly different at 7 mJ 1056 cm-1 CO2. Oleic acid experiences far less fragmentation and produces fragments primarily above m/z 125, while oleyl alcohol undergoes much more fragmentation and gives fragments largely below m/z 125. The mass spectra of the mixed particles show some interesting features. The 90:10 wt% spectrum appears to most closely resemble a combination of the peak intensities found in each of the 7 mJ pure component spectra. By 75:25 wt%, the dominant peaks for oleic acid (at m/z = 264 for example) are extremely diminished in the spectrum, more so than would be expected based on composition. This could be due to a shift in the fragmentation patterns of oleic acid, perhaps due to experiencing more heating in the vaporization step resulting in more extensive fragmentation, or a preferential vaporization of oleyl alcohol. As the two components are both liquids and miscible, it is not expected that oleyl alcohol partitions to the surface and undergoes preferential desorption at low CO2 energies.  71  1  Oleic Acid  Normalized Ion Signal [a.u.]  0 1  90:10 Oleic:Oleyl  0 1  75:25 Oleic:Oleyl  0 1  50:50 Oleic:Oleyl  0 1  25:75 Oleic:Oleyl  0 1  Oleyl Alcohol  0 50  75 100 125 150 175 200 225 250 275 300  m/z Figure 5.12: Normalized mass spectra (300 shot average) at the optimum ionization delay time (maximum total ion signal) as a function of particle composition (7 mJ 1056 cm-1 CO2, 142 nm VUV).  72  5.3.3.2 Dependence of degree of fragmentation on composition for mixed particles The fragmentation of the mixed particles was examined as a function of composition. Since oleyl alcohol fragments more extensively at 7 mJ than oleic acid, selected high mass peaks in oleic acid were used to monitor the change in fragmentation. Peaks at m/z 127, 134, 235, 246, and 264 were ratioed to m/z 148 to determine the degree of fragmentation of oleic acid. All of the peaks are assigned exclusively to oleic acid. The peak at m/z 148 was chosen for comparison as it belongs exclusively to oleic acid and is present at a range of energies in the pure oleic acid spectra. Figure 5.13 shows the degree of fragmentation as a function of composition for oleic acid in mixed particles. The fragmentation pattern for oleic acid clearly shows that as the percentage composition of oleic acid decreases, oleic acid fragments more extensively until it levels off somewhere between 75 - 50 wt% oleic acid. This behavior may indicate a change in the heating mechanism of the particle or may be a function of the fragments chosen for the analysis, or even a combination of the two. 127 m/z 134 m/z 235 m/z 246 m/z 264 m/z  Ratio of Fragment to 148m/z peak  2.5  2.0  1.5  1.0  0.5  0.0 20  30  40  50  60  70  80  90  100  Weight percent oleic acid  Figure 5.13: Fragmentation ratio of oleic acid (selected peaks relative to m/z 148) as a function of weight percent oleic acid.  73  5.3.3.3 Dependence of total ion signal on ionization delay time for a range of compositions The ionization delay profiles (total ion signal versus ionization delay time) also reflect the effect of composition on vaporization of the particle. In Figure 5.14, it can be seen that the ionization delay profiles (as well as total ion signal) change noticeably with composition until a certain percentage of oleyl alcohol is reached in the particle (~25% or greater). The total ion signal at the peak ionization delay time is also observed to vary as a function of particle composition. The total ion signal reflects the changing amount of particle material vaporized (and thus the number of molecules that can be ionized), due to the dependence of the absorption co-efficient on the composition.  14  Oleyl Alcohol 25:75 Oleic:Oleyl 50:50 Oleic:Oleyl 75:25 Oleic:Oleyl 90:10 Oleic:Oleyl Oleic Acid  12  Total Ion Signal [a.u.]  10 8 6 4 2 0 0  5  10  15  20  25  Ionization delay time (µs)  Figure 5.14: Total ion signal as a function of ionization delay time for mixed oleic acid: oleyl alcohol particles. The full width at half maximum (FWHM) values of the ionization delay profiles and the delay times at maximum total ion signal are plotted in panels (a) and (b), respectively, of Figure 5.15 as a function of particle composition. These parameters provide alternate ways of evaluating the changing delay profiles.  74  Delay Profile FWHM (µ s)  15  (a)  10  5  Delay Time at Maximum Ion Signal (µs)  0 8  (b)  6  4  2 0  20  40  60  80  100  Weight percent oleic acid  Figure 5.15: (a) Ionization delay time profile full width at half maximum (FWHM) versus weight percent oleic acid. (b) Delay time at maximum total ion signal versus weight percent oleic acid. Error bars reflect the resolution of the scan.  5.4 Discussion 5.4.1 Dependence of degree of fragmentation on CO2 energy and composition The degree of fragmentation for a species has been shown to be strongly related to the internal energy of the molecule.93, 101, 134, 136, 162, 163, 166, 176 If the internal energy of the vaporized molecules is high, then increased fragmentation is expected which can complicate the interpretation of mass spectral results.96, 101, 134, 176 For the pure particles, fragmentation escalated with increasing desorption energy, which is consistent with results from other groups that show the degree of fragmentation to be a strong function of  75  particle heating, regardless of whether a heater or laser was used for the vaporization step.93, 101, 134, 136, 162, 163, 166, 176 Additionally, in our studies the extent of fragmentation was observed to depend on particle composition. At an equal desorption energy, pure oleyl alcohol showed more fragmentation than pure oleic acid, likely due in part to the greater absorption cross-section of oleyl alcohol versus oleic acid. The degree of fragmentation of oleic acid in the mixed particles was used as a proxy for the internal energy of the vaporized molecules. It was clearly shown for oleic acid in the two component particle that the internal energy of the molecules produced by laser vaporization (and the degree of fragmentation) was dependent on composition (see Figure 5.13). If the particles experience the same amount of heating during vaporization, the fragmentation patterns of the respective species should not change as a function of the composition. The increasing amount of oleyl alcohol in the particle resulted in increasing fragmentation of oleic acid. We suggest that the presence of oleyl alcohol in the mixed particle enhances the absorption of CO2 laser radiation and leads to an increased internal energy of the vaporized molecules. This observation indicates that the analysis of mixed particles may be far more complicated than anticipated for a two step process with soft vaporization if the degree of fragmentation can change as a function of composition. The mass spectra of the pure component cannot necessarily be used as a fingerprint for that same component in a mixture. The implications for this effect are also important for laboratory studies of particle reactions and subsequent product analysis depending on the extent of composition change. If the extent of fragmentation is simply a function of the internal energy of the vaporized molecules (and thus the heating of the particle), then the fragmentation and delay profile data for oleic acid in both one and two component particles should follow a similar trend when plotted as a function of energy absorbed per particle. The respective absorption cross-sections determined in the IR measurements were used to calculate the amount of energy absorbed by the particle (Eabs) for both the pure oleic acid particles and the mixed particles according to Beer’s law and some simple approximations as described by Equation 3. The derivation of Equation 3 (for a one component system) is given in Appendix III.  76  Equation 3  E abs = ECO2 ⋅  Ap ACO2     4r p ⋅ 1 − exp − σ ⋅ N p ⋅     3         ECO2 is the measured CO2 energy, Ap is the area of the particle, ACO2 is the area of the CO2 beam where it intersects the particle, σ is the absorption cross-section, Np is the number density of the particle, and 4rp/3 is an approximation of the path length the light travels through the particle. Similarly for a two component system, the equation for energy absorbed by the particle during desorption is given below:  Equation 4  E abs = ECO2 ⋅  Ap ACO2   4rp    ⋅ 1 − exp − (σ 1 ⋅ N 1 + σ 2 ⋅ N 2 )  3      where σ1 and σ2 are the respective cross sections of each component and N1 and N2 are the respective number densities of each component in the mixed particle. Figure 5.16 shows the fragmentation ratios for peaks at m/z = 127, 134 and 264 relative to signal at m/z 148 as a function of the calculated CO2 energy absorbed per particle for oleic acid in both the one component and two component particles. The closed data points represent the pure oleic acid particles and the open data points represent the mixed particles. The percent of m/z =264 relative to m/z =148 decreases with increasing energy absorbed. For peaks at m/z 127 and 134, the percent of each peak increases relative to the peak at m/z 148. At higher energies for the pure oleic acid, the error bars are large due to the low signal to noise ratio, thus it is difficult to comment on the continuity of the trend. The data in Figure 5.16 demonstrate that the degree of fragmentation for oleic acid in both one component and two component particles is well described by the energy absorbed per particle in the desorption step.  77  127 m/z 127 m/z  264 m/z 264 m/z  134 m/z 134 m/z  3  Fragmentation ratio (relative to 148 m/z)  2 1 0 3 2 1 0 3 2 1 0 40  60  80  100  120  CO 2 energy absorbed (pJ) Figure 5.16: Fragmentation ratios for oleic acid as a function of IR energy absorbed during vaporization. Closed symbols show pure oleic acid particle data, open symbols show mixed particle data.  5.4.2 Dependence of ionization delay profiles on CO2 laser energy and particle composition The expansion of the aerosol plume can be followed by varying the delay between the CO2 and VUV pulses. The shape of the ionization delay profile as illustrated by the profile FWHM or the delay time at maximum ion signal reflects the translational energy 78  of the vaporized molecules, which is highly dependent on the vaporization energy. The pure component data clearly showed that the translational energy of the desorbed molecules grew with increasing CO2 energy leading to narrower delay profiles and earlier optimal delay times. In the case where a higher translational energy is imparted to the molecules, they will enter and leave the ionization region more quickly, whereas at lower energies the molecules will take longer to fill and then expand beyond the ionization volume. In the mixed particle study, all particles are vaporized at the same CO2 laser energy of 7 mJ, yet the translational energy of the molecules is obviously changing with composition as shown in Figure 5.14. When each particle is identical, as occurs in most laboratory experiments, one can integrate the ion signal over the entire delay profile. In the field however, where each particle is not necessarily identical, it could be problematic to use a fixed ionization delay time. For instance, if the instrument parameters are optimized for the maximum total ion signal with one species, these parameters may be far from ideal for a different species and result in a sensitivity bias. Hanna et al. have shown delay profiles for DHB and oleic acid at 15 mJ vaporization energy that varied widely; collecting data at the optimal delay time for oleic acid would result in almost no signal for DHB as its vaporization profile evolved much more quickly.166 Both one component and two component delay profiles in this study indicate this to be a valid concern. This effect may be diminished at high vaporization energies where there appears to be less of a composition dependence for pure particles, but this will come at the cost of increased fragmentation. Similar to the fragmentation effects, we attempted to describe the vaporization profile changes for the one and two component systems by the energy absorbed per particle during desorption. Figure 5.17 contains the pure and mixed particle data for oleic acid plotted as ionization delay profile FWHM (full width at half maximum) versus absorbed energy in the top panel and optimum delay time versus absorbed energy in the bottom panel. Closed data points represent the pure particle and open data points represent the two component particle. Both panels demonstrate a fairly consistent trend between the pure and mixed particle data for delay profile FWHM and optimum delay time as a function of absorbed energy. Overall, the delay profile changes for both the one and two component systems, which reflect the average translational temperature of the  79  molecules, are reasonably well described by the energy absorbed per particle during desorption.  Pure oleic acid Mixed particle  Delay Profile FWHM (µ s)  20 15 10 5  Delay Time at Maximum Signal (µ s)  0 15  10  5  0 40  60  80  100  120  140  160  180  CO 2 energy absorbed (pJ)  Figure 5.17: (Top) Ionization delay profile FWHM for both pure oleic acid particles and mixed particles as a function of CO2 energy absorbed/particle and (bottom) delay time at maximum total ion signal as a function of CO2 energy absorbed/particle.  5.4.3 Non-linear response of ion signal with particle composition In a previous study Woods et al. showed that a two step process using VUV ionization gave quantitative detection of aromatics (PAHs) in a mixed particle.99 In other words, the signal from a single component in a mixed particle was linear with the amount (mole fraction) of that component in the mixed particle. Because the degree of fragmentation has been shown to change with composition for mixed particles of oleic acid and oleyl alcohol, this is not the case for our studies. This is illustrated in Figure 5.18, which shows a plot of three oleic acid fragments relative to the total signal for all peaks (oleyl and oleic) versus the mole fraction of oleic acid at each composition. This should give a linear relationship even if the amount of material evaporating changes as a  80  function of composition (as was observed in our experiments). The three fragments at m/z = 127, 134 and 264 are used as proxies for the total signal of oleic acid. The relative intensities are normalized to one. The trend is clearly non-linear over the range of compositions, which confirms that the dependence of ion signal on composition is affected by a changing degree of fragmentation with composition.  127 134 264 1:1  Ratio of OleicFragment toTotal Ion Signal (All Peaks) [a.u.]  1.0  0.8  m/z m/z m/z fit line  0.6  0.4  0.2  0.0 0.0  0.2  0.4  0.6  0.8  1.0  Mole ratio oleic acid  Figure 5.18: Relative intensity of oleic acid fragment ion signal to total ion signal for all peaks (oleic and oleyl) as a function of the oleic acid mole ratio.  The above figure shows that a quantitative analysis is difficult (if not impossible) if the energy absorbed per particle during desorption changes significantly with composition. Conditions where quantitative analysis should be successful are (1) where the fragmentation does not change with composition or (2) where the IR absorption cross-sections do not change with composition (assuming heat capacities are similar). An example of the first case would be analytes that do not fragment or produce relatively few fragments, such as mixtures of PAHs as were studied by Woods et al.99 For the second case, species with similar absorption cross-sections and heat capacities are needed or the change in composition restricted to a range where the dependence is not as strong. These conditions also assume liquid phase and miscibility of species.  81  5.5 Conclusions The degree of fragmentation observed in single particle mass spectra of a one component system was demonstrated to be a function of CO2 power. In a two component system, the degree of fragmentation was shown to be a function of the particle composition. The combined one and two component data suggest that the key parameter that determines the extent of fragmentation, which is a proxy for the internal energy of the vaporized molecules, is the energy absorbed by the particle during desorption. The ionization delay profile (and implicitly the translational temperature of the vaporization plume) in a one component system is shown to be strongly dependent on the vaporization energy. In a two component system, the delay profile is shown to strongly depend on the composition of the particle. The combined data suggest that the key parameter that governs the delay profile (and translational temperature of the vaporized molecules) is also the energy absorbed by the particle during desorption. The change in the degree of fragmentation with particle composition resulted in a non-linear dependence of ion signal on composition. This makes any attempt at quantification highly difficult. For species that do not fragment as easily, or have similar absorption cross-sections, quantitative analysis should be possible. These results should be relevant for other two step laser desorption/ionization systems that incorporate a CO2 laser.77,  92, 99, 124  The existing two step laser  desorption/ionization systems with a CO2 laser all use a TOF mass analyzer, which will result in less fragmentation than what is observed with an ion trap.166 Nevertheless the effects observed above will still be of concern for TOF instruments as some level of fragmentation is observed with a TOF analyzer for many types of analytes, particularly long chain hydrocarbons.93, 100, 117, 162, 163  82  Chapter 6. VUV Wavelength Scanning Using Version II for Analysis of Two Component Aerosols 6.1 Introduction Chapter 5 discussed some of the complexities experienced in multi-component organic particle analysis despite the application of soft single photon ionization to reduce the amount of fragmentation observed. In particular, some of the limitations for quantitative analysis of mixed particles with a two laser approach were revealed. However, this type of instrument may still have advantages over other approaches for species identification in mixtures. Using a two step process with VUV ionization results in reduced fragmentation compared to EI or a one laser approach. Additionally, with a tunable VUV system the ionization energies for different components can be determined, which should assist in product identification. Tunable VUV light has been employed in an AMS instrument to distinguish between two C30H62 isomers with ionization energies (IE's) differing by 0.3 eV.101 Heterogeneous reactions of anthracene-coated sodium chloride particles with ozone were also investigated and the tunable VUV capability was used to characterize the products formed.101 These results showcase some of the advantages of a tunable VUV source. It should be noted that Gloaguen et al.101 utilized a synchrotron in their AMS system to provide a continuous VUV source. Furthermore their mixture represented a “best-case scenario”, consisting of anthracene, a PAH that is fairly robust under fragmentation, and sodium chloride, a salt which would provide minimal interfering peaks. No existing studies have utilized a pulsed, tunable VUV source for investigation of mixed particles. This chapter describes the use of version II of the SPITMS to investigate the characteristics of a tunable VUV source for compound identification in a mixed particle system of oleic acid and 1-octadecene. These types of organic species fragment extensively when ionized with EI and have also been observed to have significant fragmentation in our system (because of the ion trap storage time) even when using soft VUV ionization. We demonstrate that by scanning the VUV we can simplify the mass  83  spectrum of the mixed particle and that we can also determine the ionization energy of a single compound in the binary mixture. Oleic acid is a relevant atmospheric species and well-explored in the work of this thesis, thus it was used as one component of the mixed particle. The second component was chosen to have a significantly different ionization energy than oleic acid and also be liquid and miscible. After several tests with various compounds we found 1-octadecene to be a suitable candidate. The appearance energy of 1-octadecene was determined to be 9.42 ± 0.05 eV using the SPITMS and tunable VUV source. In the mixed particle system, the appearance energy of 1-octadecene was determined to match closely with the value obtained in the pure particle study, although the signal to noise was poorer than the pure case.  6.2 Experimental The structures of oleic acid and 1-octadecene are shown in Figure 6.1 along with some relevant properties in Table 6.1.  HO  O  Figure 6.1: Chemical structures of oleic acid (top) and 1-octadecene (bottom). Table 6.1: Relevant properties of oleic acid and 1-octadecene.  a  Compound  M.W. (g/mol)  Density (g/cm3)  Oleic acid  282.4614  0.895  1-octadecene  252.4784  0.789  IE (eV) 8.68164 8.65 ± 0.05166 9.42 ± 0.03a  this study  84  6.2.1 Particle generation Oleic acid (Fluka, ≥99.0%) and 1-octadecene (Fluka, ≥99.5%) were used as purchased without further purification. Solutions were prepared in 2-propanol (Aldrich, 99.9%) in the concentration range of 10-5 g/mL and particles were generated using a vibrating orifice aerosol generator (TSI Inc., Model 3450). Particles were passed through a  85  Kr charge neutralizer (TSI Inc., Model 3054)  with ~25 LPM dilution air flow to completely dry the particles before entering the aerosol mass spectrometer. This prevented additional drying and size changes in the aerosol lens. As the vibrating orifice aerosol generator produces a narrow size distribution, particles were sampled directly into the instrument. Particles were ~1 µm in diameter. This size was selected for both ease of generation and the high sensitivity towards larger particles.  6.2.2 Particle analysis 6.2.2.1 Desorption A mid-IR tunable TEA-CO2 laser (MTL-3G Edinburgh Instruments Ltd) with a single mode output was used for desorption. The focal size of the laser beam was ~1 mm diameter. Pure 1-octadecene particles were desorbed with 30 mJ 941 cm-1. Mixed particles of 50:50 wt% oleic acid:1-octadecene were desorbed with either 6 or 16 mJ/pulse 944 cm-1 CO2. 1-octadecene has no IR features in this region, thus the pure spectra acquired at 941 cm-1 are not expected to be any different than had the data been acquired at 944 cm-1. The lower desorption energies for the mixed particle were used to reduce the fragmentation of both species and prevent ambiguity in peak assignment in the mass spectra. No ions were produced in the desorption step alone in any case.  6.2.2.2 Ionization The tunable laser-based VUV source was described briefly in Chapter 5. The optimal ionization delay time was determined for a set of given experimental conditions (particle type and desorption energy). At the optimal ionization delay time, the VUV source was used in the scanning wavelength mode. Both the VUV wavelength and energy  85  were recorded along with the ion signal for each laser shot. Scan speeds were on the order of 0.005 nm/s (0.0003 eV/s at 140 nm) and spectra were recorded 3 - 4 times per second to give a resolution of ~0.002 nm (0.0001 eV). Any fluctuations in the VUV intensity were compensated for by normalizing the ion signal to the recorded VUV power.  6.2.2.3 Mass Analysis Ions were trapped in the custom built 3D ion trap described in Chapter 4 and cooled for 10 ms in ~1 mTorr He buffer gas. Mass scans were performed by linearly increasing the RF voltage (mass selective instability mode) and collected at a rate of 4000 Da/s. The mass range used was 50 – 300 m/z. The mass axis is calibrated daily by recording 70 eV EI spectra of perfluorotributylamine.  6.3 Results and discussion 6.3.1 Mass spectra for pure 1-octadecene particles A desorption energy of 30 mJ 941 cm-1 was used for the pure particles of 1octadecene. Although this results in significant fragmentation of linear hydrocarbon species, it provides a higher sensitivity as a larger percentage of the particle is vaporized. Averaged mass spectra (normalized to the average VUV power) for 1-octadecene at discrete wavelengths from 127 nm (9.77 eV) to 139 nm (8.93 eV) are shown in Figure 6.2. The majority of fragments appear below m/z 125 with significant fragments at m/z = 83, 97, 111, and 125, likely due to both the high vaporization energy and the storage time in the ion trap.166 Small gas phase background peaks at m/z 186 and 279 are also observed.  86  1  83  127 nm (9.77 eV)  97 111 125  71  0  131 nm (9.47 eV)  1  Ion Signal [a.u.]  0  134 nm (9.26 eV)  1  0  136 nm (9.12 eV)  1  0  139 nm (8.93 eV)  1  0 50  100  150  m/z  200  250  300  Figure 6.2: Mass spectra (300 shot average) for 1-octadecene particles at discrete VUV wavelengths (30 mJ 941 cm-1 CO2).  87  6.3.2 Mass spectra for mixed oleic acid:1-octadecene particles The mass spectra for the mixed particles at discrete VUV wavelengths using 6 mJ 944 cm-1 are displayed in Figure 6.3. At this desorption energy, fragmentation is greatly reduced but conversely a smaller amount of material is evaporated from the particle resulting in a lower overall total ion signal. As can be seen in Figure 6.3, tuning the wavelength towards lower photon energies results in a greatly simplified mass spectrum. The spectrum at 127.5 nm contains approximately 50 peaks as opposed to around 10 peaks at 136.5 nm. Given that many VUV wavelength sources for aerosols function at 118 nm (high energy photons),93,  163, 181-184  the ability to tune to a lower VUV energy  could aid in analysis of multi-component particles.  Normalized Ion Signal [a.u.]  1  127.5nm VUV (9.73 eV)  0 1  130.5nm VUV (9.51 eV)  0 1  131.5nm VUV (9.44 eV)  0 1  136.5nm VUV (9.09eV)  0 80  100  120  140  160  180  200  220  240  260  280  m/z Figure 6.3: Mass spectra (200 shot average) from VUV energy scan of 50:50wt% oleic acid:1-octadecene particles with 6 mJ 944 cm-1 CO2 desorption.  88  6.3.3 1-octadecene appearance energy in pure particles VUV wavelength scans of pure 1-octadecene were collected from 127 – 136 nm (9.8 – 9.1 eV). As the molecular ion signal was not large enough to use for determination of the ionization energy (IE) of pure 1-octadecene, the appearance energies of 4 significant fragments of 1-octadecene (m/z = 83, 97, 111, 125) were compared instead as shown in Figure 6.4. Data from 9.55 - 9.59 eV were removed due to the strong resonance in Xe gas at 129.6 nm (9.57 eV) that prevents VUV generation. The observed ionization threshold for 1-octadecene is obtained by extrapolating the linear portion of the threshold region of the PIE curve to the baseline.110, 166 The appearance energies of peaks at m/z 125, 111, 97, and 83 were 9.39 ± 0.02 eV, 9.41 ± 0.02 eV, 9.42 ± 0.02 eV, and 9.45 ± 0.03 eV, respectively. The uncertainty in these measurements arises mainly from the uncertainty in fitting a line to the linear portion of the PIE curve, which in part stems from the nearby xenon resonance. The fact that all the appearance energies agree within error indicates that the appearance energy of the fragments can be taken as equivalent to the ionization energy of the molecule and seems to indicate a post-ionization fragmentation process. The ionization energy of 1-octadecene was determined to be 9.42 ± 0.03 eV by taking the average appearance energy of the fragments.  89  9.1 6  9.2  9.3  9.4  9.5  9.6  9.7  9.8  9.3  9.4  9.5  9.6  9.7  9.8  m/z 125  3  0 10 m/z 111  Ion Signal [a.u.]  5  0 14  m/z 97  7  0 10 m/z 83  5  0 9.1  9.2  VUV ionization energy (eV)  Figure 6.4: Photoionization efficiency curves for 1-octadecene fragments from pure 1-octadecene particles (30 mJ 941 cm-1 CO2). The average extrapolated appearance energy is 9.42 ± 0.03 eV.  90  6.3.4 1-octadecene appearance energy in mixed particles Mixed particles of 50:50 wt% oleic acid:1-octadecene were analyzed via a VUV wavelength scan from 127 - 139 nm (9.8 – 8.9 eV). This range of photon energies is above the oleic acid ionization energy (8.65 eV) so the oleic acid ion signal should remain relatively constant throughout the scan, while the ion signal for 1-octadecene should disappear at energies below its ionization threshold. The ion signal for oleic acid was observed to remain relatively constant during the scan although is not shown in any of the figures. Figure 6.5 shows the photoionization efficiency (PIE) curves acquired for the peak at m/z 125 for 1-octadecene. The top panel shows pure 1-octadecene for comparison, while the middle and bottom panel show the mixed particle at 6 mJ and 16 mJ 944 cm-1, respectively. Data from 9.55 - 9.59 eV was removed due to the strong resonance in Xe gas at 129.6 nm (9.57 eV) that prevents VUV generation. The observed ionization threshold for 1-octadecene is obtained by extrapolating the linear portion of the threshold region of the PIE curve to the baseline. For the mixed particles, the S/N for the 1-octadecene fragment ion signals is lower due to the smaller amount of 1-octadecene (52% mole ratio of the particle) and the lower vaporization energy applied to reduce fragmentation (less total material evaporated). An appearance energy of 9.41 ± 0.03 eV is determined for 1-octadecene in the mixed particle from the average of linear fits to PIE curves of the four fragments at 6 mJ (middle panels of Figure 6.5, Figure 6.6, Figure 6.7, and Figure 6.8). This value is quite close to the measured appearance energy for the pure 1-octadecene particle. The average appearance energy for fragments collected at 16 mJ (bottom panels of Figure 6.5, Figure 6.6, Figure 6.7, and Figure 6.8) was 9.38 ± 0.05 eV using linear fits to the same portion of the scan used in the 6 mJ data. This also gives good agreement with the measured value for the pure component. Signal to noise is greatly improved at the higher vaporization energy, however, the mass spectra are significantly more complicated due to more extensive fragmentation. While the linear regime of the 16 mJ data is somewhat debatable due to features around 9.45 eV, this feature could arise from a combination of factors such as a low energy tail from an onset of ionization at energies below the IE as well as a poorer normalization of the VUV intensity due to the closeness to the Xe 91  resonance region. It is possible more of the data points should have been removed between 9.5 – 9.55 eV to improve the fit. 6  Ion Signal for m/z 125 [a.u.]  Pure 1-octadecene  0 1  50:50 oleic acid:1-octadecene, 6 mJ IR  0 1.5  50:50 oleic acid:1-octadecene, 16 mJ IR  0.0 9.0  9.1  9.2  9.3  9.4  9.5  9.6  9.7  9.8  VUV ionization energy (eV) Figure 6.5: Photoionization efficiency curves for 1-octadecene fragment at m/z 125: (top) pure 1-octadecene (30 mJ 941 cm-1 CO2), (middle) mixed oleic acid:1octadecene (6 mJ 944 cm-1 CO2) and (bottom) mixed oleic acid:1-octadecene (16 mJ 944 cm-1 CO2).  92  10  Ion Signal for m/z 111 [a.u.]  Pure 1-octadecene  0 1  50:50 oleic acid:1-octadecene, 6 mJ IR  0 3  50:50 oleic acid:1-octadecene, 16 mJ IR  0 9.0  9.1  9.2  9.3  9.4  9.5  9.6  9.7  9.8  VUV ionization energy (eV) Figure 6.6: Photoionization efficiency curves for 1-octadecene fragment at m/z 111: (top) pure 1-octadecene (30 mJ 941 cm-1 CO2), (middle) mixed oleic acid:1octadecene (6 mJ 944 cm-1 CO2) and (bottom) mixed oleic acid:1-octadecene (16 mJ 944 cm-1 CO2).  93  15 Pure 1-octadecene  Ion Signal for m/z 97 [a.u.]  10 5 0 1.0  50:50 oleic acid:1-octadecene, 6 mJ IR  0.5  0.0 4  50:50 oleic acid:1-octadecene, 16 mJ IR  2  0 9.0  9.1  9.2  9.3  9.4  9.5  9.6  9.7  9.8  VUV ionization energy (eV) Figure 6.7: Photoionization efficiency curves for 1-octadecene fragment at m/z 97: (top) pure 1-octadecene (30 mJ 941 cm-1 CO2), (middle) mixed oleic acid:1octadecene (6 mJ 944 cm-1 CO2) and (bottom) mixed oleic acid:1-octadecene (16 mJ 944 cm-1 CO2).  94  10 Pure 1-octadecene  Ion Signal for m/z 83 [a.u.]  5  0 50:50 oleic acid:1-octadecene, 6 mJ IR  0.4  0.2  0.0 50:50 oleic acid:1-octadecene, 16 mJ IR 2  0 9.0  9.1  9.2  9.3  9.4  9.5  9.6  9.7  9.8  VUV ionization energy (eV) Figure 6.8: Photoionization efficiency curves for 1-octadecene fragment at m/z 83: (top) pure 1-octadecene (30 mJ 941 cm-1 CO2), (middle) mixed oleic acid:1octadecene (6 mJ 944 cm-1 CO2) and (bottom) mixed oleic acid:1-octadecene (16 mJ 944 cm-1 CO2).  6.4 Conclusions The ionization energy of pure 1-octadecene was determined to be 9.42 ± 0.03 eV using laser desorption/pulsed tunable VUV ionization in version II of the SPITMS. Mixed particles of two similar aliphatic organic species (oleic acid and 1-octadecene) were analyzed by scanning the VUV wavelength. By tuning the ionization energy, it was possible to greatly simplify the mass spectra of the two component particle. This is one advantage of the tunable VUV ionization method. The mass spectra were comparatively simpler for the lower vaporization energy used (6 mJ) which reduced fragmentation of both species.  95  The appearance energy of 1-octadecene in the mixed particle was determined by analysis of fragments specific to that species assuming that the molecular ion shares the same appearance energy. This is another advantage of the tunable VUV ionization: the potential to determine an analyte’s identity by observing the ion appearance energy. The appearance energy for 1-octadecene in the mixed particle (from the 6 mJ data) was determined to be 9.38 ± 0.03 eV in good agreement with the threshold determined from the pure particle data. A low vaporization energy was used to reduce fragmentation since most species studied in the SPITMS undergo more fragmentation than is observed in a TOFMS due to the long storage time of ions. However, the signal to noise ratio was quite poor for the system picked. While the signal to noise ratio improved at a higher vaporization energy due to the increased amount of particle material evaporated, the degree of fragmentation also increased resulting in more ambiguity in the peak assignment. An improved signal to noise ratio without a concurrent increase in fragmentation is expected if more stable molecules like PAHs are studied. Higher CO2 powers could be used to increase the signal to noise ratio with minimal fragmentation in the mass spectra. For PAHs or other stable molecules, the improved signal to noise ratios and minimal fragmentation should greatly facilitate the identification of analytes in a mixture by tuning the ionization energy.  96  Chapter 7. Studies of Complex and/or Atmospherically Realistic Aerosols Using Version I and Version II of the SPITMS 7.1 Introduction Our previous studies covered micron and submicron, single component, organic aerosols of 2,4-dihydroxybenzoic acid, oleic acid, caffeine, and linoleic acid.110,  111, 166  We also reported results from 2 component mixtures of oleic acid:oleyl alcohol and oleic acid:1-octadecene in Chapter 5 and Chapter 6, some of the very few aerosol mass spectrometry studies of mixed particles with CO2 desorption and VUV ionization. The CO2 laser provided effective desorption of neutrals without ion production until high CO2 energies were reached for the solid particles tested. Ions from CO2 only desorption/ionization were never observed for the liquid organic particles tested. This gave us a wide range of CO2 energies in which to carry out the two step process free of background from ions produced by the CO2 laser alone. Atmospheric aerosols are more complex and can include hundreds of species, including ionic compounds. Due to the variety of species in these aerosols, it is not clear if there will be a range of CO2 energies where sufficient particle vaporization occurs without ion production by the CO2 laser alone or whether the CO2 only ions will comprise a relatively small background signal over which the ionization of gas phase neutrals will be readily apparent. In the following, we study four types of atmospherically relevant and/or complex aerosol particles of both organic and inorganic species. The particles studied include ammonium bisulfate (inorganic salt), Suwannee River fulvic acid (water soluble organic carbon species), cigarette sidestream smoke (complex combustion particle), and meat cooking aerosols (complex organic particle). The rationale behind the study of these species is given in the subsequent paragraphs. Sulfates are one of the dominant species present in atmospheric aerosols.1, 17, 18, 185 In the troposphere, reactions of sulfuric acid aerosols with ammonia lead to the formation of neutralized sulfate species: letovicite ((NH4)3H(SO4)2), ammonium bisulfate and ammonium sulfate. Ammonium to sulfate ion ratios varying from ~1.0 - 2.0 have been  97  observed in tropospheric particles suggesting that ammonium bisulfate aerosols are of atmospheric relevance.186-189 Pure ammonium bisulfate (NH4HSO4) aerosols were studied to determine the system’s ability to characterize a significant inorganic species present in atmospheric particles. HUmic LIke Substances (HULIS) are a diverse collection of organic macromolecular compounds that have been identified in aerosols, fog, and cloud water.190 These humic-like substances are believed to account for a significant fraction of the water-soluble organic carbon (WSOC) in aerosols191-193 and include dialkyl ketols, polyols, polyphenols, alkanedioic acids, hydroxyalkanoic acids, aromatic acids, and polycarboxylic acids.194 Fuzzi et al. suggested that fulvic acid can be used to represent the polycarboxylic acids found in the WSOC fraction of atmospheric aerosols.194 Suwannee River fulvic acid is a certified standard that has been used in several laboratory studies as representative of HULIS found in the atmosphere.195-197 Cigarette smoke is a highly complex and dynamic matrix consisting of thousands of species. Investigations of both gas phase and particle phase smoke constituents, the size and composition characterization of sidestream (from the tip of a lit cigarette) and mainstream smoke (sampled through the filter of the cigarette), and the impacts on indoor air quality and human health have been conducted, to name only a few studies.13, 40, 116, 172, 198-203  Rogge et al. estimated environmental tobacco smoke (comprised of sidestream  and mainstream smoke) aerosols to make up 1.0 - 1.3% of the fine particle mass concentration in the Los Angeles area.13 Tracer compounds identified in these aerosols include nicotine, dicarboxylic acids, phenols, phytosterols, alkanoic acids, alkanols, nalkanes, PAHs, alkaloids, and iso and ante-isoalkanes among numerous other classes.13, 204  Meat cooking aerosols are of interest due to their significant contribution to urban organic aerosol concentrations.205 In a study by Hildemann et al., the contribution of meat cooking operations to organic aerosol emissions was investigated and found to comprise up to 21% of the primary fine organic carbon particle emissions in the Los Angeles Basin.203 Rogge et al. examined particles emitted from meat cooking operations with GC/MS and identified more than 75 organic compounds with prominent species of  98  palmitic acid, stearic acid, oleic acid, nonanal, 2-octadecanal, 2-octadecanol, and cholesterol.12 First we carried out studies with the CO2 laser alone to determine the threshold at which ions were produced from CO2 only desorption/ionization. We then collected mass spectra of the particles using version I (CO2/EI) and/or version II (CO2/VUV) at CO2 energies both above and below the threshold where ions were produced from the CO2 laser only. Preliminary identification of species in the aerosols studied is postulated, but more rigorous studies are required to confirm the species identification. These studies provide the beginnings of a spectral database for complex particle analysis with the single particle ion trap mass spectrometer, as well as illuminating the limitations of a two step desorption/ionization process for some types of atmospherically relevant particles.  7.2 Experimental 7.2.1 Particle generation Pure ammonium bisulfate aerosols were generated using the constant output atomizer (TSI Inc., Model 3076). Ammonium bisulfate (Alfa Aesar, >99.9%) was used without further purification and prepared at 0.1 M concentration in Millipore water (18 MΩ). Solutions were passed through a 0.45 µm filter before entering the constant output atomizer. The generated particles were passed through a  85  Kr charge neutralizer (TSI  Inc., Model 3054) as well as a 24″ nafion diffusion dryer (Permapure Inc) to ensure particle dryness. A TSI differential mobility analyzer (TSI Inc., Model 3081) was then used to select a 400 nm mobility diameter particle which was sampled into the instrument. Fulvic acid particles were also produced using the constant output atomizer. Suwannee River fulvic acid (SRFA) Standard I (International Humic Substances Society, code 1S101F) was prepared in Millipore water (18 MΩ). Solutions were passed through a 0.45 µm filter before entering the constant output atomizer. The same drying procedures as mentioned above were used for the particles generated. Particles with a 500 nm mobility diameter were selected with the DMA.  99  Cigarette sidestream smoke particles were generated by sampling the stream of smoke from the smoldering end of a lit cigarette. Commercially available cigarettes (regular, with filter) with a nicotine and tar content of 1.0 and 13 mg, respectively, per cigarette were used. The cigarette was mounted in a Swagelok fitting through which ~100 cm3/min of air was drawn to keep the cigarette smoldering. An ~20 LPM dilution flow of compressed air was directed upwards underneath the burning tip of the cigarette. The smoke particles were passed through the 85Kr charge neutralizer and the DMA was used to select particles with a 700 nm mobility diameter. Meat cooking aerosols were generated using ground beef (store-bought) with ~25% fat content (representative of fat content of meat used in restaurant/meat-cooking operations) pressed into patties and grilled on a propane BBQ. Particles were sampled into the DMA and a 950 nm mobility diameter was selected. For the cigarette smoke and meat cooking aerosols, the largest diameter possible that could be sampled at a reasonable rate was selected. For fulvic acid and ammonium bisulfate, the particle size selected was smaller due to the detector saturation that occurred at larger sizes and high CO2 energies. To overcome this limitation, one could employ an instrument with a variable dynamic range or selectively trap to exclude major ions that lead to space charge.  7.2.2 Particle analysis The tunable TEA-CO2 laser (MTL-3G, Edinburgh Instruments Ltd.) was used at a desorption wavelength of 944 cm-1 for all particle types studied. This wavelength was chosen due to its common use in aerosol vaporization.92,  98, 99, 115, 124  Additional  wavelengths were also used in some cases to provide desorption at a wavelength onresonance with a species’ absorption band to enhance vaporization of the particle. Specifics for each particle type are discussed in the respective results sections. After a variable ionization delay time, the EI pulse or VUV laser pulse was applied for ionization as described in Chapters 5 and 6. To alternate between the two modes of ionization simply involved switching the ionization trigger from the FPGA board to feed into the external triggering mechanism of the Nd:YAG pump laser for the VUV system or the electron gate trigger for the EI. 100  For electron impact, a gate width of ~4 µs was usually used at a delay time of 0 1 µs where the optimum ion signal is expected.111 For VUV ionization, fixed wavelengths of 142 nm (8.74 eV) and 124 nm (10.0 eV) were used at ionization delay times of 4 µs (determined to be optimal by delay time scans). Hundreds of shots were collected at a single delay time and averaged for the data presented here. Once the CO2 laser and ionization source fired, trapped ions were collisionally cooled for 10 ms in ~1 mTorr of He buffer gas and a mass scan was performed by linearly increasing the RF voltage (mass selective instability mode) as described in section 4.2.2.3. Mass analysis was conducted using a range of m/z appropriate to each particle type at a scan speed of 2000 Da/s to optimize mass resolution and peak shape.  7.3 Results and discussion 7.3.1 Ammonium bisulfate 7.3.1.1 CO2 only of ammonium bisulfate particles Ammonium bisulfate aerosols were first studied over a range of CO2 energies similar to our previous experiments to determine at what threshold level (if any) ions were produced from the CO2 laser alone. Desorption wavelengths of both 944 cm-1 (10.6 µm) and 1037 cm-1 (9.64 µm) were used. NH4HSO4 exhibits stronger absorption at 1037 cm-1 due to an absorption band from the bisulfate molecule. Ions were not observed from the desorption laser alone at any of the energies accessible at 944 cm-1 CO2 (≤ 40 mJ). At the 1037 cm-1 wavelength, CO2 only ions were observed. Figure 7.1 shows the mass spectra obtained from 5 - 25 mJ at 1037 cm-1 CO2. Ions begin appearing at ~10 mJ.  101  0.01  5mJ  Ion Signal [a.u.]  0.00 0.01  10mJ  0.00 0.04  15mJ  0.00 0.08  20mJ  0.00 0.12  25mJ 0.00 20  40  60  80  100  120  140  m/z Figure 7.1: Mass spectra (400 shot average) of 400 nm ammonium bisulfate particles for 5 - 25 mJ 1037 cm-1 CO2.  Figure 7.2 shows the mass spectra acquired at the highest 1037 cm-1 CO2 energy used (25 mJ). Peaks at m/z 18, 80, and 97 due to NH4+, SO3+, and HSO4+ respectively are clearly observed as well as several other peaks, many of which are likely due to alkali or metal cations present in trace amounts. These peaks are tentatively assigned in the mass spectrum. In a study by Gross et al., relative sensitivity factors were determined for a  102  variety of alkali cations and the ammonium cation using the ATOFMS.90 For a 1:1 solution of ammonium and sodium salts, the response factor for NH4+/Na+ was 0.014,90 which could offer some insight into the size of the sodium peak we observe relative to the ammonium peak despite the fact that sodium is present only in trace amounts. Gross et al. also determined the relative sensitivity factor for K+/Na+ to be 5.1,90 which would also support the large signal we see from potassium despite its presence in trace amounts. 0.12 39  K+  0.10  Ion Signal [a.u.]  NH 4+  0.08  0.06  0.04  Na+ FeO+/MnOH+ NH 4OH+  0.02  +  S /O 2+  AlOH+ 41  NO 2+  NaCl+/ 58 CaOH+/ Ni+ Al2O+  K+  CrO 2+ SO + HSO 4+ NiO+ 3 Al2O 2+  0.00 10  20  30  40  50  60  70  80  90  100  m/z  Figure 7.2: Mass spectrum (400 shot average) of 400 nm ammonium bisulfate particles with 25 mJ 1037 cm-1 CO2.  7.3.1.2 CO2/EI of ammonium bisulfate particles Particles were subsequently studied using CO2 desorption in combination with pulsed electron impact ionization. We carried out experiments at 5 mJ 1037 cm-1 CO2, where no CO2 only ions were produced. A ~4 µs long electron pulse was used at a delay time of 0 µs after the desorption step.111 When using EI, there was always a background ion signal from residual gas in the system (due to small leaks in the vacuum system). After the residual gas background signal was subtracted, no ion signal was observed above the background noise in the spectrum. Aerosol spectra were additionally collected for long periods of time (2000 shots) in an attempt to improve the signal to noise but no signal above background was observed. 103  Similar studies using desorption at high energies of 1037 cm-1 CO2 and EI (where CO2 only ions were produced) were also done, but no new peaks nor any quantifiable enhancement of ion signal were observed with the addition of EI. VUV ionization was not pursued as no advantages were expected for analysis of an inorganic particle. These results likely indicate that the vaporization of the particle material is insufficient at low energies of 1037 cm-1 and all energies of 944 cm-1. Any appreciable amount of particle vaporization will automatically result in ions since the species is ionic,99 which can then undergo charge-transfer matrix effects. While the two step process of CO2/EI did not yield characteristic mass spectra for ammonium bisulfate, our results show that ionization via the CO2 laser alone could be successful for determining the presence of inorganic species in aerosol particles. This will most likely depend on the IR absorption of the inorganic species and/or the other components present in a mixed particle.  7.3.2 Suwannee River fulvic acid (SRFA) 7.3.2.1 CO2 only of SRFA particles SRFA particles were studied using two desorption wavelengths of 1079 cm-1 and 944 cm-1, on and off-resonance respectively of a very broad absorption band for SRFA at 1204 cm-1 as shown in Figure 7.3. This absorption band has been ascribed to C-O stretching of the carboxylic groups.191, 206  104  944 cm-1 1079 cm-1  Absorbance [a.u.]  1204 cm-1  Wavenumbers (cm-1)  Figure 7.3: IR absorbance spectrum of SRFA Standard I.206  One of the proposed average structural models of SRFA is shown below in Figure 7.4. The ash content of SRFA Standard I is given as 0.46% w/w for a dry sample of SRFA.207 The ash content is the %(w/w) of inorganic residue in a dry sample of SRFA. Trace inorganic species identified in a study of SRFA 1S101F with content greater than 5 µg/g include silver, aluminum, boron, calcium, iron, molybdenum, sodium, tin, thorium and zirconium.208  Figure 7.4: Proposed average structural model of Suwannee River fulvic acid.209  105  Shown in Figure 7.5 are the mass spectra for SRFA using the CO2 laser only over a range of energies at 944 cm-1 CO2. At approximately 10 mJ and above, ions were generated from the desorption step alone. 0.01  6mJ  0.00 0.01  10mJ  39  Ion Signal [a.u.]  57 71 43  87 94  120 108 134 150  0.00 0.07  20mJ  55 94 69 79  106  120 108 134 146  2 1  158  0 36  38  40  42  44  36  38  40  42  44  0.00 0.21  55  77  65  40mJ 91  115 106 130 138 152  6 4 2 0  167  0.00 50  75 100 125 150 175 200 225 250 275 300  m/z Figure 7.5: Mass spectra (500 shot average) for Suwannee River fulvic acid aerosols for 6 - 40 mJ 944 cm-1 CO2. The inset in the bottom two panels is shown to provide a comparison of the signal size for peaks between m/z 38-44 relative to the rest of the mass spectrum.  106  Figure 7.6 shows the spectrum for a 500 shot average of SRFA particles at 40 mJ 944 cm-1 CO2. The dominant peak is m/z 39, which is ~25x the largest signal observed from the 50 - 300 Da range. Other significant peaks are seen at m/z 40, 41, and 43. The peak at m/z 39 can tentatively be assigned to a combination of C3H3+ and K+, m/z = 40 to Ca+ and C3H4+, and m/z = 41 to C3H5+ and  41  K+. Although potassium was not a major  component of the ash content of SRFA, it may well be present in trace amounts. Species with the greatest ability to undergo charge transfer in the matrix will be detected with the highest sensitivity, like K+ ions.56, 90, 120 The m/z = 43 peak may be attributable to the acetyl moiety (CH3CO+) of a carbonyl-containing compound or the C3H7+ fragment. Smaller clusters of peaks are observed at regular mass intervals (~12-14 units apart) from 50 to 300 m/z, including the ion series CnH2n+1+ and CnH2n-1+. In an HPLC-MS study of the WSOC fraction of aerosols, peaks at m/z 138, 152 and 166 have been identified as nitrophenol and its C1- and C2- homologues, which are also mass peaks observed here that may correspond to these species.210 Some of the higher mass peaks at m/z 179, 189 and 202 may be indicative of the presence of PAHs. 0.25  6 55  77  0.20  39  40mJ  5 4  Ion Signal [a.u.]  3 2  0.15 91 58  65  80  0.10  106  41  1  115 94  40  0 38  117  67  40  43  42  44  46  48  130 100  0.05  138  152  83  167 176 189  202 237  0.00 50  75  100  125  150  175  200  225  250  275  300  m/z  Figure 7.6: Mass spectrum (500 shot average) of 500 nm SRFA aerosols with 40 mJ 944 cm-1 CO2.  107  We also studied SRFA particles at 1079 cm-1 CO2 which falls within the SRFA absorption band at 1204 cm-1 in the IR.206 The CO2 dependence showed a similar trend except CO2 only ions were first observed at a lower energy (5 mJ) as expected based on the higher absorption-cross section.  7.3.2.2 CO2 /EI and CO2/VUV of SRFA particles At 10 mJ 944 cm-1 CO2, where only a small amount of ion signal was observed, an EI pulse ~4 µs long at a 1 µs ionization delay was used.111 After correcting for EI residual gas background, the total ion signal for CO2/EI showed no quantifiable enhancement over the CO2 only ion signal. No new peaks in the mass spectrum were observed either. Similarly at a high CO2 energy of 40 mJ/pulse 944 cm-1 with EI, neither enhancement of ion signal nor any change in the mass spectrum was observed. Additionally, at 3 mJ 1079 cm-1 (where no CO2 only ions were observed) an EI pulse 13 µs long was used and produced no change in the mass spectrum. The longer EI pulse was applied to attempt enhancing the ion signal and reduce the possibility of selecting a nonoptimal ionization delay time for the experimental conditions. SRFA particles were then studied using CO2 energies both on and off-resonance where no CO2 only ions were produced with VUV ionization at a 4 µs ionization delay time to determine if any effect or advantage was gained in switching to a soft ionization process.166 Two VUV wavelengths, 142 nm (8.74 eV) and 124 nm (10 eV), were used for comparison. Neither VUV wavelength produced a quantifiable change in the ion signal nor new mass peaks. This seems to suggest that at CO2 energies below the CO2 only ionization threshold, the vaporization step does not produce a sufficient number of gasphase molecules that can be detected upon ionization with the VUV. To date, only one group that we are aware of has measured fulvic acid in an aerosol mass spectrometer45, 211 in order to compare the mass spectral signatures of FA to organic particulates measured in rural and remote sites. This study was conducted with thermal desorption (600oC) and electron impact in a time-of-flight mass spectrometer with ensemble averaged results. The mass spectra were characterized by fragments below m/z 55.45  108  While our results perhaps surprisingly showed that the species in a predominantly organic particle are not being sufficiently vaporized at these low CO2 energies to produce ion signal with either EI or VUV, we can observe well-resolved mass peaks up to m/z = 300 in our CO2 only spectra. This suggests that CO2 laser desorption/ionization may be a useful tool for single particle measurements of these types of aerosols and this single step desorption/ionization method may be complimentary to UV desorption/ionization, which is often used for field measurements.  7.3.3 Cigarette sidestream smoke (SSS) 7.3.3.1 CO2 only of SSS particles Studies of SSS particles were conducted at 944 cm-1. Figure 7.7 shows the mass spectra collected over a range of CO2 energies. Ions were produced from CO2 only desorption/ionization once an energy of ~15 mJ/pulse was reached. Morrical et al. also observed significant ion production from the CO2 laser alone in their mass spectrometry studies of cigarette smoke.92 Peaks observed in the CO2 only spectra are tentatively assigned based on previous compounds identified in other cigarette smoke studies. The peak at m/z 163 may be the protonated molecular ion for nicotine (M+H)+, m/z = 84 could be a nicotine fragment, m/z = 80 could be pyrazine, and m/z = 69 could be pyrroline. Other peaks assignments include phenol or 2-vinylfuran for m/z = 94; xylene, benzaldehyde or ethylbenzene for m/z = 106; indole for m/z =117; methylindene for m/z = 130; isopropyltoluene for m/z =134; 5-quinolineamine for m/z = 144; nicotyrine for m/z = 158; and 3,5-dimethyl-1phenylpyrazole for m/z= 172.202 Several higher mass peaks are present but have not been assigned.  109  0.02  10mJ  0.00  Ion Signal [a.u.]  0.02  163  0.00 0.06  15mJ  163  20mJ  173 187  134 147  0.00 0.1  80 158163 173 30mJ 106 213 134 144 84 186 69 94 197 224 123 55 238  0.0 0.3  80  84 106  55  77  94  158  117  130 144  40mJ  172  163  186  213  0.0 50  75  100 125 150 175 200 225 250 275 300  m/z Figure 7.7: Mass spectra (300 shot average) of 700 nm cigarette SSS particles for 10 – 40 mJ/pulse 944 cm-1 CO2.  7.3.3.2 CO2/EI of SSS particles Particles were then studied using CO2 desorption with EI. At 13 mJ 944 cm-1 (no CO2 only ions produced as seen in Figure 7.8a), an EI pulse ~13 µs long was applied for ionization and produced observable new peaks in the mass spectrum as shown in Figure 7.8b. Most peaks are clustered between m/z 50 – 100 as expected for a higher energy ionization process. The EI spectrum does not show the nicotine parent ion at m/z 162, however there is a large peak at m/z 84, likely from the nicotine fragment. 110  Ion Signal [a.u.]  0.08  (a)  0.00 0.08 57 55  (b)  81 91 67 69 84 95 79 93 107 119 105  133  0.00 110  0.20 0.08  81 58  (c)  84 95  124 136  161 150  62  162 176  192  0.00 50  75  100  125  150  175  200  225  250  275  300  m/z  Figure 7.8: Mass spectra of 700 nm cigarette smoke particles (300 shot average) with (a) 13 mJ 944 cm-1 CO2, (b) 13 mJ 944 cm-1 CO2/EI, and (c) 13 mJ 944 cm-1 CO2/VUV (142 nm).  A high desorption energy of 40 mJ 944 cm-1 was also used with EI. In this case, the application of EI after a high desorption energy did not result in a significant enhancement of ion signal, but did slightly increase the abundance of low mass peaks relative to high mass peaks.  7.3.3.3 CO2/VUV of SSS particles Cigarette SSS particles were then studied using 13 mJ 944 cm-1 with 142 nm VUV (8.73 eV) at a 4 µs ionization delay, which produced the spectrum shown in Figure 7.8c. The parent peak for the nicotine molecular ion at m/z 162 is observed in the spectrum as well as the expected peak at m/z 84 for the nicotine fragment. The peak at m/z 110 is tentatively assigned to the molecular ion of any of the dihydroxybenzenes.13,  111  172  The IE for a dihydoxybenzene is 7.94 eV so it should be possible to ionize at 142 nm  (8.73eV). Probable peak assignments include: m/z = 136 to benzeneacetic acid or limonene,13 m/z = 124 to 3 or 4-methyl catecol,13, 162 m/z = 95 to a hydroxypyridine,13 m/z = 81 to methylpyrrole, and m/z = 69 to pyrroline.202 The peak at m/z 176 could be due to cotinine.13, 201 The peak at m/z 178 could be either phenanthrene or anthracene. The spectrum is markedly different from the CO2/EI spectrum with the softer ionization method producing higher m/z peaks as expected, similar to the work by Northway et al. comparing thermal desorption with EI and VUV of cigarette smoke.162 Conversely, while their results showed the nicotine fragment at m/z = 84 as the dominant peak in the thermal desorption/VUV mass spectrum, the dominant peak in our spectrum is at m/z 110. This could be due to the difference in the energy imparted during the desorption step or the difference in the photoionization wavelength used (142 nm in our case and a mixture of 118 and 123.6 nm in their case).  7.3.4 Meat cooking aerosols 7.3.4.1 CO2 only of meat cooking aerosols Mass spectra were collected using a desorption wavelength of 944 cm-1 over a range of energies from 5 - 35 mJ/pulse as shown in Figure 7.9. Ion signal is first observed at ~10 mJ/pulse. The peak at m/z 178 could be either phenanthrene or anthracene. None of the other peaks are assigned at this time.  112  0.04 5 mJ  0.00 0.04  85  10 mJ  113  87  115  Ion Signal [a.u.]  0.00 0.04 157  85 120 132  87  15 mJ  168 178  0.00 0.05  113 85  132  178  157 144  120  20 mJ  168  87 104  0.00 0.05  25 mJ  157 104 120 132 85 95 113  55 57  213  178  0.00 0.25 157 113  55 57  85 95  120 132  213 178  0.00 50  75  35 mJ  100  125  150  175  200  225  m/z Figure 7.9: Mass spectra (200 shot average) of 950 nm meat cooking particles for 5 35 mJ/pulse 944 cm-1 CO2.  7.3.4.2 CO2/EI of meat cooking aerosols Data were collected at 5, 10, 15, and 20 mJ/pulse desorption energy. Particles were ionized with a ~4 µs long EI pulse 1 µs after the desorption laser fired. The 20  113  mJ/pulse data is shown below in Figure 7.10 as it provided the most significant enhancement of ion signal upon application of EI (no enhancement observed at 5 or 10 mJ/pulse, and very little at 15 mJ/pulse). The EI ionization pulse results in a nearly 20 fold increase in the signal of the most intense peak of the spectrum comparing panels (a) and (b) of Figure 7.10. The spectrum is very similar to a fatty hamburger spectrum recorded with thermal desorption/EI in the Aerodyne AMS with clusters of hydrocarbon peaks centered around m/z 55, 69, and 83.205 Possible species (based on previously identified species in the particle phase using charbroiled regular ground beef) for the peak at m/z =157 may be N,N-dibutyl formamide, [M]+, or 5-pentyldihydro-2(3H)-furanone, [M+H]+.12, 212 0.8  (a)  Ion Signal [a.u.]  0.4  0.0 50  0.8  75  100  125  150  175  55  200  (b)  57 69  0.4  67  71  43  81 83 95  79  97  157  109 111  0.0 50  75  100  125  150  175  200  225  m/z  Figure 7.10: Mass spectra (200 shot average) of 950 nm meat cooking particles with (a) 20 mJ CO2 only and (b) CO2/EI at a desorption energy of 20 mJ.  The major organic species identified in meat charbroiling aerosols are palmitic acid, stearic acid, and oleic acid.12,  212  Other significant species include long chain (9  carbons or longer) alkanoic acids, alkenoic acids, ketones, and aliphatic aldeyhydes, as  114  well as cholesterol.12, 212 Under analysis with EI, these long chain hydrocarbon species will form short chain fragments as seen here and in other studies.111, 205 Studies that have utilized softer ionization sources such as CI have observed the molecular ions of the pure components named above as well as the molecular ions of most of these species when analyzing meat cooking aerosols.37,  40, 102, 176  It is expected that the use of laser  desorption/VUV ionization would allow us to obtain reduced fragmentation and larger m/z peaks, if not more molecular ions.  7.4 Conclusions This chapter presents the first results obtained from a variety of atmospherically relevant and/or complex aerosols using version I and version II of the single particle ion trap mass spectrometer. Laboratory generated aerosols of ammonium bisulfate, Suwannee River fulvic acid, cigarette sidestream smoke, and meat cooking aerosols were analyzed in real-time on a single particle basis. Ideally a particle should have good absorbance at the IR wavelength used without undergoing ionization during the desorption step. For all systems tested there was only a small range of CO2 energies where no ions were produced from the CO2 laser alone. However there were two scenarios that provided for a successful two step desorption/ionization experiment: 1) sufficient neutral desorption without production of ions by the CO2 laser only or 2) production of ions during the desorption step but significant enhancement of ion signal upon application of the ionization step, suggesting that the majority of desorbed molecules were neutral species rather than ionic species. Combining CO2 desorption with EI or VUV was successful for cigarette smoke particles and meat cooking aerosols. Below 15 mJ, cigarette smoke particles were vaporized (without ionization) using the CO2 laser and the gas phase neutrals were then ionized with EI or VUV. The two step process using VUV allowed us to reduce fragmentation in comparison to the EI spectra with characteristic fragments observed at masses greater than m/z 100 and clear identification of the nicotine ion. For meat cooking aerosols, at ~20 mJ desorption energy, the use of EI resulted in significantly enhanced ion  115  signal with characteristic hydrocarbon peaks identified, despite ion production during the desorption step. The two step process did not work for ammonium bisulfate or Suwannee River fulvic acid particles. For these systems, at low CO2 energies, not enough material was vaporized to give a signal above background. At higher CO2 energies, ions were produced by the CO2 laser alone and the application of an ionization source did not result in any significant enhancement of ion signal. These results show that the suitability of CO2 desorption for atmospherically relevant particles will strongly depend on particle composition. However, the two step vaporization/ionization process can work well for certain systems and could be applied for laboratory or focused field studies. Our data also potentially suggests that a CO2 only instrument may be more useful to develop for field studies than CO2/EI. Other representative types of atmospheric aerosols would need to be studied first to confirm this; however, we were able to acquire mass spectra for all four types of particles over a range of CO2 energies using only the CO2 laser. The spectra furthermore contained an interesting range of well-resolved high mass fragments for the fulvic acid and cigarette smoke particles. This indicates that using a CO2 laser for single step vaporization/ionization may be a feasible ionization technique for atmospheric aerosol studies and provide complimentary information to the UV laser desorption/ionization technique. It has been demonstrated that pulsed IR matrix assisted laser desorption ionization (MALDI) (at 3 and 10 µm) is useful for studying particularly large and/or labile compounds in comparison to UV-MALDI. IR MALDI is generally regarded as a softer desorption method, leading to a low degree of metastable ion fragmentation.213,  214  Similar to the effects experienced in MALDI, possible benefits to using a CO2 laser over a UV laser for single step desorption/ionization may include less fragmentation in the spectra which would aid in identification of species in multicomponent atmospheric aerosols. Further studies would need to be done to verify the potential advantages single step CO2 laser desorption/ionization might have over UV laser desorption/ionization.  116  Chapter 8.  Concluding Remarks  8.1 Summary of findings This thesis focused on the development and characterization of two versions of a new single particle mass spectrometer for organic aerosol studies that incorporate a 3D ion trap for mass analysis. Version I of the SPITMS integrated a CO2 laser for desorption and electron impact ionization. Version II of the SPITMS incorporated a CO2 laser for desorption and a novel, tunable laser-based VUV ionization source developed in our laboratory.  8.1.1 Version I Version I of the instrument worked well in preliminary experiments on simple organic aerosols, providing comparable mass spectra of 2,4-dihydroxybenzoic acid, caffeine, oleic acid, and linoleic acid to the NIST standard EI spectra. High quality single particle spectra were obtained representing the first reported single particle mass spectra collected using EI in an ion trap aerosol mass spectrometer. Tandem mass spectrometry studies up to MS3 of single caffeine particles were also demonstrated with an acceptable overall efficiency (~15%). These results show that for the right types of organic species (e.g. robust aromatic molecules), this version could be successfully used for MSn studies. The detection limit of the ionization region determined for this technique was ~1x108 molecules (350 nm particle) for 2,4-dihydroxybenzoic acid. Version I was then used to study a selection of atmospherically relevant particles. Mass spectra of cigarette smoke particles and meat cooking aerosols showed significant total ion signal enhancement upon application of EI. The spectra were comparable in quality to previous studies with well-established aerosol instruments.92,  162, 205  For  ammonium bisulfate and Suwannee River fulvic acid aerosols, application of EI post desorption did not lead to a significant enhancement in ion signal. We suggest this is due to either insufficient vaporization of particle material or a predominant amount of ionic species formed in the desorption step. Whether this scenario would occur with a wider  117  range of atmospherically realistic aerosols is unknown and further work would need to be done to ascertain the applicability of the technique for a broader range of aerosols (see section 8.3).  8.1.2 Comparison of Version I to existing aerosol mass spectrometers It is useful to make some comparisons of Version I to existing aerosol instruments utilizing EI. Kürten et al.109 and Harris et al.60 used ion trap mass spectrometers with thermal vaporization/EI and generated ions outside of the ion trap. Externally generated ions are trapped with an efficiency of ~1 - 5%.133 We generate our ions directly in the center of the trap by CO2 laser vaporization which should give an enhanced sensitivity in our instrument based on trapping efficiencies. Kürten et al. and Harris et al.’s instruments are also not capable of single particle studies and thus cannot investigate changes in composition on a particle to particle basis, while Version I can. The most comparable instrument to Version I is the Aerodyne aerosol mass spectrometer, which utilizes thermal desorption/EI and a time-of-flight mass spectrometer. The Aerodyne is able to provide single particle and particle ensemble information on composition and quantitative results for particle composition. The ability of Version I of the SPITMS to provide a quantitative response has been initially tested, but the results were inconclusive. More work would need to be done to prove the quantitative capabilities. We did however achieve a single particle detection limit comparable to what has been demonstrated with the Aerodyne instrument despite the fact that our electron gun generates far less total beam current. One advantage of Version I compared to instruments that use thermal desorption is that we have the ability to study refractory materials (such as sea salt) by use of CO2 laser vaporization. Conversely, a CO2 laser requires more maintenance and optimization than a heater and can give varying amounts of vaporized particle material depending on the absorption cross-sections and heat capacities of the compounds within the particle. One major drawback of the Aerodyne that has been noted is the particle bounce loss of ~ 50% upon impaction on the heater.168 It is possible that the percentage of particles not volatilized by the CO2 laser will be less than the particle bounce loss, but this cannot be stated for certain without a side-by-side comparison of the instruments. The other benefit 118  of Version I is a faster acquisition and data saving time for single particle information. The Aerodyne instrument has a very low duty cycle for single particle measurements: to acquire and save one single particle mass spectrum takes ~800 ms.168 Version I of the SPITMS offers much faster acquisition times (~60 ms) and could thus acquire more single particle results in the same time period. Overall, Version I of the SPITMS appears to offer some promising improvements over existing EI instruments as long as its limitations are kept clearly in mind. Chemical identification in multi-component aerosols will remain challenging due to the harsh form of ionization employed, similar to other EI instruments, and will be limited to class groupings of molecules. If there are species resistant to fragmentation, the ability to do MS/MS with the ion trap will provide additional advantages over existing instrumentation for molecular identification.  8.1.3 Version II Version II of the instrument, which incorporates CO2 laser desorption and VUV ionization, was used to study pure component oleic acid and oleyl alcohol particles as well as mixtures of the two over a range of compositions. One of the main rationales for a two-step desorption/ionization approach in aerosol mass spectrometry has been that by avoiding matrix effects such instruments are superior for the analysis of mixed particles in that quantitative results can be obtained. However, this is one of the few in-depth studies of mixed particles with such an instrument, particularly of long chain hydrocarbon species that showed otherwise. As the percent of oleyl alcohol increased in the mixed particles, an increase in the degree of fragmentation (internal energy) and faster ionization delay profiles (translational energy) were observed up until a certain composition. The changes in fragmentation and delay profiles for both one and two component systems were described by a key parameter: the energy absorbed per particle during the desorption step. Due to the change in the degree of fragmentation with composition observed, the ion signal responded non-linearly with particle composition. These effects are inherent (and should be kept in mind) in any aerosol mass spectrometer where laser desorption is used as the composition of the particle, the energy of the laser, and where the particle passes through the profile of the desorption beam will 119  result in variability in energy absorbed during desorption. Operating at a fixed desorption temperature (cartridge heater) has advantages in terms of being able to create or reference mass spectral libraries in that a fixed internal energy is imparted to the vaporized molecules which will in turn provide a fixed degree of fragmentation during ionization. However, Northway et al. concluded that the optimal desorption temperatures for soft ionization techniques to further reduce fragmentation will vary with the aerosol of interest.162 Thus depending on the composition of the aerosol, optimizing the temperature for minimal fragmentation for all compounds is not possible and/or would lead to partial desorption in many cases.162 This again makes the quantitative analysis of mixtures very difficult. To determine what advantages could be gained from the tunability of the VUV source in such a scenario as above with a mixture of long chain hydrocarbon species, two component particles of oleic acid and 1-octadecene (50:50 weight percent) were studied. We illustrated that we can use the tunable VUV source to simplify the mass spectra and determine the ionization energy for a single component in a mixed particle. Both of these capabilities are advantageous for species identification, as well as the ability to isolate the total ion contribution from one species which should aid in the ability for quantification. Version II of the instrument was also used for preliminary studies of some atmospherically relevant aerosols. Cigarette smoke aerosols produced significant enhancement of ion signal with the two step process, while the fulvic acid aerosols showed no detectable enhancement of signal or change in the mass spectra with VUV ionization post CO2 desorption. Ammonium bisulfate aerosols were not studied with this version as no ions were produced with CO2/EI and the soft photoionization technique is not particularly advantageous for inorganic aerosols.  8.1.4 Comparison of Version II to existing aerosol mass spectrometers The most relevant studies for comparison to Version II are: Baer et al’s work37, 93, 99, 100, 136, 164  which uses thermal or CO2 laser desorption, VUV ionization, and a TOFMS;  Zelenyuk et al.’s work with SPLAT56, 126, 127, 215 which uses CO2 laser desorption, UV ionization, and TOFMS; and Petrucci et al.’s work with PERCI50,  102, 171, 216  which  combines thermal desorption, low energy electron ionization, and a TOFMS. These 120  instrumental configurations all use two step vaporization/ionization processes and soft ionization techniques. Baer’s group has used both fixed wavelength VUV generation37, 93, 99, 100, 136, 163, 217 and synchrotrons164 to produce VUV radiation. In comparison to a fixed wavelength source, the tunability of our VUV source allows for the minimization of fragmentation through near threshold ionization.163 It also allows for the differentiation of species by appearance energies and the measurement of ionization energies for additional identification of unknown species.166 A linear response to particle mass and a detection limit of 8x105 molecules of caffeine (75 nm particle) was determined for Version II in previous work; however, the optical detection of the instrument’s sizing region limits us to particles >250 nm diameter.110 This is still an improvement to the single particle detection limit of 300 nm estimated by Baer et al.’s group.86 In comparison to synchrotron generated VUV, the primary advantage of our source is the lab availability rather than the limited access and travel requirements for a synchrotron source. The synchrotron is also a continuous ionization source paired with thermal vaporization, which does not lend itself to single particle measurements. In comparison to SPLAT, the use of UV ionization has been shown to result in more fragmentation than single photon ionization.166 Single photon ionization is also more general than multiphoton ionization (UV),99 which will be beneficial in heterogeneous reaction studies (discussed in section 8.2.2) and would allow for the possibility of observing non-aromatic products. Furthermore, single photon ionization is preferred for quantitative work as the single photon ionization cross-sections vary much less from compound to compound than REMPI cross sections.99 SPLAT does achieve a lower sizing and detection limit (~50 nm)56 than Version II of the SPITMS due to the superior configuration of the particle sizing region. If the optical components for the particle sizing region of the SPITMS could be improved, the overall detection limit in the ionization region, ~75 nm, would be very comparable with SPLAT. PERCI (photoelectron resonance capture ionization) is applicable to a broader range of compound classes than REMPI and has been demonstrated to provide greatly reduced fragmentation levels and an improved ability to identify mixtures over other instruments utilizing VUV. Conversely, particle phase detection limits are determined to  121  be on the order of 10-8 to 10-9 g total deposited mass with PERCI,50 which is roughly 4-5 orders of magnitude worse than our detection limits (compared to the optical detection limit). The low sensitivity of PERCI will hinder its ability to study multi-component particles of 1 µm or smaller. Additionally, the technique is not velocimetry triggered and thus does not provide single particle information. In all three instruments compared to above, a TOFMS is used and positive and negative ion detection is obtained. The combination of positive and negative ion information provides better identification of species within the aerosols. Furthermore, instruments with a TOFMS will provide spectra with less fragmentation than ion traps due to the longer storage time of ions before detection in traps.166 We can still target Version II towards focused studies where the fragmentation will be less of an issue (cyclic structures etc.), where the MSn capabilities will provide additional insight, and where a tunable source will be beneficial. In further development of the instrument, a negative ion detector could easily be implemented in place of the electron gun to provide us with the ability to detect both positive and negative ions.  8.2 Future directions The application of the two versions of this new instrument in future studies could follow a number of different routes. The studies of two component and multi-component aerosols have revealed a number of unanticipated results in our studies, but still provided some valuable insights into aerosol analysis techniques and show promise for specific applications.  8.2.1 Version I future work Version I is a conceivable predecessor to a single particle field instrument due to the compact size of the ion trap mass analyzer, the ruggedness of the ionization source, and the successful use CO2 lasers for vaporization in the field. Some essential work in the laboratory that should be completed prior to field implementation includes optimization of the EI source to improve sensitivity of the instrument, demonstration of quantitative  122  capabilities, and verification that refractory particles such as sea salt or mineral dust could be effectively studied. An enhanced understanding of the CO2 desorption process should be undertaken before field experiments particularly since some of the atmospherically realistic particles studied did not produce CO2/EI spectra. This could be done in conjunction with the expansion of the spectral library to include more atmospherically relevant aerosols like biomass  burning  particles  (e.g.  wood  fire),  diesel  and  gasoline  exhaust,  bacteria/biological particles, soil samples, and marine aerosols. Marker peaks for different types of atmospherically relevant particles could then be identified to allow for improved identification of aerosol sources. Further optimization of the instrument could include studies where if ions are formed during the desorption step, as observed in Chapter 7, a method to extract them from the ion trap pre-ionization could be developed. Additionally, pursuing the application of the ion trap mass spectrometer for MS/MS studies to better resolve components in a complex particle or in particle reaction studies would be a worthwhile investigation. For instance, some of the high mass peaks detected in cigarette smoke particles could be isolated and fragmented using collision-induced dissociation to determine their structure. With regards to field instruments, our preliminary results from Chapter 7 also suggest the potential usefulness of a one-step CO2 laser desorption/ionization for a simplified laboratory or field instrument. Obviously a one step desorption/ionization process will be easier to implement and maintain in the field. If the comparison to IR MALDI is valid, it may provide less fragmentation than UV ablation/ionization which would improve the ability to identify chemical species. Testing this configuration with more aerosol types would be an interesting direction to pursue. In particular, mixed aerosols of inorganic and organic components (ionic and neutral components), such as oleic acid and sodium chloride would be a good first test system.  8.2.2 Version II future work Version II of the instrument is suited for laboratory stories and would be useful in studies of heterogeneous reactions (kinetic rates and product studies) relevant to atmospheric chemistry. As the heterogeneous reactions of organic particles with gas 123  phase species, such as ozone, hydroxyl radicals and nitrate radicals, can result in oxidation of organic species and alter the particle’s hygroscopicity, viscosity, morphology and overall reactivity, it is essential to better characterize the reaction rates and products of these heterogeneous interactions.37, 38, 40, 171, 218-222 Field measurements suggest that heterogeneous chemistry may change the composition of atmospheric aerosols.223 This could affect the use of certain “molecular markers” to identify a particular aerosol source, e.g. oleic acid for meat cooking processes. Furthermore, some studies suggest that these reactions can lead to rapid volatilization of organic particulate matter224, 225 (resulting in volatile organic carbon production in the atmosphere226, 227) and may serve as a significant sink for organic particles.225 Due to the increased fragmentation experienced in our system with the ion trap, reactions with PAHs or other atmospherically important aromatic species (e.g. pinene or levoglucosan) would be the most logical heterogeneous studies to pursue. PAHs are produced during incomplete combustion processes such as diesel and gasoline engines and biomass or coal burning.1 The number of studies on condensed phase PAHs and gas phase oxidation with NO3 are fairly limited at this point.228-233 Currently in our lab group, the oxidation of organic surfaces via nitrate radicals has been studied using a flow tube reactor.228, 231, 234-236 The loss of gas phase reactant is monitored by chemical ionization mass spectrometry (CIMS) and the heterogeneous reaction rates can be determined, however there is no ability to identify the products formed. The single particle mass spectrometer would be useful for monitoring the products formed in the particle phase when the gas phase species reacts with the aerosol particles as the tandem MS capabilities would provide additional structural identification of products in comparison to other aerosol mass spectrometers. The particle density could also be monitored for evidence of the extent of reaction. Comparing the same heterogeneous reaction using both the CIMS and the single particle mass spectrometer would allow for quantitative determination of uptake coefficients as well as elucidating reaction pathways via product studies.  124  References  (1)  Finlayson-Pitts, B. J.; Pitts Jr., J. N. Chemistry of the Upper and Lower Atmosphere: Theory, Experiments, and Applications; Academic Press: San Diego, 2000.  (2)  U.S. EPA, Air Quality Criteria for Particulate Matter, U.S. Environmental Protection Agency: Washington D.C., 1996, EPA 600/P-95/001aF.  (3)  Dockery, D. W.; Pope, C. A.; Xu, X. P.; Spengler, J. D.; Ware, J. H.; Fay, M. E.; Ferris, B. G.; Speizer, F. E. New Engl J Med 1993, 329, 1753-1759.  (4)  Schwartz, J.; Dockery, D. W.; Neas, L. M. J Air Waste Manage 1996, 46, 927939.  (5)  Metcalfe, S.; Derwent, D. Atmospheric Pollution and Environmental Change; Hodder Arnold: London, 2005.  (6)  Jacob, D. J. Introduction to Atmospheric Chemistry; Princeton University Press: Princeton, 1999.  (7)  Solomon, S.; Qin, D.; Manning, M.; Alley, R. B.; Berntsen, T.; Bindoff, N. L.; Chen, Z.; Chidthaisong, A.; Gregory, J. M.; Hegerl, G. C.; Heimann, M.; Hewiton, B.; Hoskins, B. J.; Joos, F.; Jouzel, J.; Kattsov, V.; Lohmann, U.; Matsuno, T.; Molina, M. J.; Nicholls, N.; Overpeck, J.; Raga, G.; Ramaswamy, V.; Ren, J.; Rusticucci, M.; Somerville, R.; Stocker, T. F.; Whetton, P.; Wood, R. A.; Wratt, D. In Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change; Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K. B., Tignore, M., Miller, H. L., Eds.; Cambridge University Press: Cambridge, 2007, pp 19-91.  (8)  Murphy, D. M. Mass Spectrom. Rev. 2007, 26, 150-165.  (9)  Herring, J. A.; Ferek, R. J.; Hobbs, P. V. J. Geophys. Res. [Atmos.] 1996, 101, 14451-14463.  (10)  Pöschel, U. Angew. Chem. Int. Edit. 2005, 44, 7520-7540.  (11)  Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. Environ. Sci. Technol. 1993, 27, 1892-1904.  (12)  Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simonelt, B. R. T. Environ. Sci. Technol. 1991, 25, 1112-1125.  125  (13)  Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R. Environ. Sci. Technol. 1994, 28, 1375-1388.  (14)  Prather, K. A.; Hatch, C. D.; Grassian, V. H. Annu. Rev. Anal. Chem. 2008, 1, 485-514.  (15)  Wallace, J. M.; Hobbs, P. V. Atmospheric Science: An Introductory Survey, Second ed.; Elsevier Inc.: Amsterdam, 2006.  (16)  Kanakidou, M.; Seinfeld, J. H.; Pandis, S. N.; Barnes, I.; Dentener, F. J.; Facchini, M. C.; Van Dingenen, R.; Ervens, B.; Nenes, A.; Nielsen, C. J.; Swietlicki, E.; Putaud, J. P.; Balkanski, Y.; Fuzzi, S.; Horth, J.; Moortgat, G. K.; Winterhalter, R.; Myhre, C. E. L.; Tsigaridis, K.; Vignati, E.; Stephanou, E. G.; Wilson, J. Atmos. Chem. Phys. 2005, 5, 1053-1123.  (17)  Zhang, Q.; Jimenez, J. L.; Canagaratna, M. R.; Allan, J. D.; Coe, H.; Ulbrich, I.; Alfarra, M. R.; Takami, A.; Middlebrook, A. M.; Sun, Y. L.; Dzepina, K.; Dunlea, E.; Docherty, K.; DeCarlo, P. F.; Salcedo, D.; Onasch, T.; Jayne, J. T.; Miyoshi, T.; Shimono, A.; Hatakeyama, S.; Takegawa, N.; Kondo, Y.; Schneider, J.; Drewnick, F.; Borrmann, S.; Weimer, S.; Demerjian, K.; Williams, P.; Bower, K.; Bahreini, R.; Cottrell, L.; Griffin, R. J.; Rautiainen, J.; Sun, J. Y.; Zhang, Y. M.; Worsnop, D. R. Geophys. Res. Lett. 2007, 34, L13801, doi:10.1029/2007GL029979.  (18)  Heintzenberg, J. Tellus, Ser. B. Chem. Phys. Meteorol. 1989, 41, 149-160.  (19)  Saxena, P.; Hildemann, L. M. J Atmos Chem 1996, 24, 57-109.  (20)  Jacobson, M. C.; Hansson, H. C.; Noone, K. J.; Charlson, R. J. Rev. Geophys. 2000, 38, 267-294.  (21)  Hinz, K. P.; Spengler, B. J. Mass Spectrom. 2007, 42, 843-860.  (22)  Sipin, M. F.; Guazzotti, S. A.; Prather, K. A. Anal. Chem. 2003, 75, 2929-2940.  (23)  Eatough, D. J.; Lewis, L. J.; Eatough, M.; Lewis, E. A. Environ. Sci. Technol. 1995, 29, 787-791.  (24)  Kaupp, H.; Umlauf, G. Atmospheric Environment Part A - General Topics 1992, 26, 2259-2267.  (25)  Zhang, X. Q.; Mcmurry, P. H. Environ. Sci. Technol. 1991, 25, 456-459.  (26)  Canagaratna, M. R.; Jayne, J. T.; Jimenez, J. L.; Allan, J. D.; Alfarra, M. R.; Zhang, Q.; Onasch, T. B.; Drewnick, F.; Coe, H.; Middlebrook, A.; Delia, A.; Williams, L. R.; Trimborn, A. M.; Northway, M. J.; DeCarlo, P. F.; Kolb, C. E.; Davidovits, P.; Worsnop, D. R. Mass Spectrom. Rev. 2007, 26, 185-222.  126  (27)  Hansen, A. D. A.; Novakov, T. Aerosol Sci. Tech. 1990, 12, 194-199.  (28)  Stolzenburg, M. R.; Hering, S. V. Environ. Sci. Technol. 2000, 34, 907-914.  (29)  Pinnick, R. G.; Hill, S. C.; Nachman, P.; Videen, G.; Chen, G.; Chang, R. K. Aerosol Sci. Tech. 1998, 28, 95-104.  (30)  Boring, C. B.; Al-Horr, R.; Genfa, Z.; Dasgupta, P. K.; Martin, M. W.; Smith, W. F. Anal. Chem. 2002, 74, 1256-1268.  (31)  Loflund, M.; Kasper-Giebl, A.; Tscherwenka, W.; Schmid, M.; Giebl, H.; Hitzenberger, R.; Reischl, G.; Puxbaum, H. Atmos. Environ. 2001, 35, 2861-2869.  (32)  Noble, C. A.; Prather, K. A. Mass Spectrom. Rev. 2000, 19, 248-274.  (33)  Wexler, A. S.; Johnston, M. V. In Aerosol Measurements: Principles, Techniques, and Applications; Baron, P. A., Willeke, K., Eds.; John Wiley: New York, 2001, pp 365-385.  (34)  Johnston, M. V. J. Mass Spectrom. 2000, 35, 585-595.  (35)  Thomson, D. S.; Murphy, D. M. Chemtech 1994, 24, 30-35.  (36)  Katrib, Y.; Martin, S. T.; Hung, H. M.; Rudich, Y.; Zhang, H. Z.; Slowik, J. G.; Davidovits, P.; Jayne, J. T.; Worsnop, D. R. J. Phys. Chem. A 2004, 108, 66866695.  (37)  Smith, G. D.; Woods, E.; DeForest, C. L.; Baer, T.; Miller, R. E. J. Phys. Chem. A 2002, 106, 8085-8095.  (38)  Hearn, J. D.; Smith, G. D. J. Phys. Chem. A 2004, 108, 10019-10029.  (39)  Hearn, J. D.; Lovett, A. J.; Smith, G. D. Phys. Chem. Chem. Phys. 2005, 7, 501511.  (40)  Hearn, J. D.; Smith, G. D. Int. J. Mass Spectrom. 2006, 258, 95-103.  (41)  Katrib, Y.; Martin, S. T.; Rudich, Y.; Davidovits, P.; Jayne, J. T.; Worsnop, D. R. Atmos. Chem. Phys. 2005, 5, 275-291.  (42)  Zelenyuk, A.; Yang, J.; Song, C.; Zaveri, R. A.; Imre, D. Environ. Sci. Technol. 2008, 42, 8033-8038.  (43)  Tolocka, M. P.; Jang, M.; Ginter, J. M.; Cox, F. J.; Kamens, R. M.; Johnston, M. V. Environ. Sci. Technol. 2004, 38, 1428-1434.  (44)  Tolocka, M. P.; Heaton, K. J.; Dreyfus, M. A.; Wang, S. Y.; Zordan, C. A.; Saul, T. D.; Johnston, M. V. Environ. Sci. Technol. 2006, 40, 1843-1848.  127  (45)  Alfarra, M. R.; Paulsen, D.; Gysel, M.; Garforth, A. A.; Dommen, J.; Prevot, A. S. H.; Worsnop, D. R.; Baltensperger, U.; Coe, H. Atmos. Chem. Phys. 2006, 6, 5279-5293.  (46)  Alfarra, M. R.; Coe, H.; Allan, J. D.; Bower, K. N.; Boudries, H.; Canagaratna, M. R.; Jimenez, J. L.; Jayne, J. T.; Garforth, A. A.; Li, S. M.; Worsnop, D. R. Atmos. Environ. 2004, 38, 5745-5758.  (47)  Jimenez, J. L.; Jayne, J. T.; Shi, Q.; Kolb, C. E.; Worsnop, D. R.; Yourshaw, I.; Seinfeld, J. H.; Flagan, R. C.; Zhang, X. F.; Smith, K. A.; Morris, J. W.; Davidovits, P. J. Geophys. Res. [Atmos.] 2003, 108, 8425, doi:10.1029/2001JD001213.  (48)  Zhang, Q.; Canagaratna, M. R.; Jayne, J. T.; Worsnop, D. R.; Jimenez, J. L. J. Geophys. Res. [Atmos.] 2005, 110, D07S09, doi:10.1029/2004JD004649.  (49)  Buzorius, G.; Zelenyuk, A.; Brechtel, F.; Imre, D. Geophys. Res. Lett. 2002, 29, 1974, doi:10.1029/2001GL014221.  (50)  LaFranchi, B. W.; Petrucci, G. A. Int. J. Mass Spectrom. 2006, 258, 120-133.  (51)  Lazar, A.; Reilly, P. T. A.; Whitten, W. B.; Ramsey, J. M. Environ. Sci. Technol. 1999, 33, 3993-4001.  (52)  Reilly, P. T. A.; Gieray, R. A.; Whitten, W. B.; Ramsey, J. M. Environ. Sci. Technol. 1998, 32, 2672-2679.  (53)  Slowik, J. G.; Stainken, K.; Davidovits, P.; Williams, L. R.; Jayne, J. T.; Kolb, C. E.; Worsnop, D. R.; Rudich, Y.; DeCarlo, P. F.; Jimenez, J. L. Aerosol Sci. Tech. 2004, 38, 1206-1222.  (54)  Spencer, M. T.; Shields, L. G.; Prather, K. A. Environ. Sci. Technol. 2007, 41, 1303-1309.  (55)  Gross, D. S.; Galli, M. E.; Silva, P. J.; Wood, S. H.; Liu, D. Y.; Prather, K. A. Aerosol Sci. Tech. 2000, 32, 152-163.  (56)  Zelenyuk, A.; Imre, D. Aerosol Sci. Tech. 2005, 39, 554-568.  (57)  Harris, W. A.; Reilly, P. T. A.; Whitten, W. B. Int. J. Mass Spectrom. 2006, 258, 113-119.  (58)  Stowers, M. A.; van Wuijckhuijse, A. L.; Marijnissen, J. C. M.; Scarlett, B.; van Baar, B. L. M.; Kientz, C. E. Rapid Commun. Mass Spectrom. 2000, 14, 829-833.  (59)  Mansoori, B. A.; Johnston, M. V.; Wexler, A. S. Anal. Chem. 1996, 68, 35953601.  128  (60)  Harris, W. A.; Reilly, P. T. A.; Whitten, W. B. Anal. Chem. 2007, 79, 2354-2358.  (61)  DeCarlo, P. F.; Kimmel, J. R.; Trimborn, A.; Northway, M. J.; Jayne, J. T.; Aiken, A. C.; Gonin, M.; Fuhrer, K.; Horvath, T.; Docherty, K. S.; Worsnop, D. R.; Jimenez, J. L. Anal. Chem. 2006, 78, 8281-8289.  (62)  Schneider, J.; Hings, S. S.; Hock, B. N.; Weimer, S.; Borrmann, S.; Fiebig, M.; Petzold, A.; Busen, R.; Karcher, B. J. Aerosol Sci. 2006, 37, 839-857.  (63)  Murphy, D. M.; Cziczo, D. J.; Froyd, K. D.; Hudson, P. K.; Matthew, B. M.; Middlebrook, A. M.; Peltier, R. E.; Sullivan, A.; Thomson, D. S.; Weber, R. J. J. Geophys. Res. [Atmos.] 2006, 111, D23S32, doi:10.1029/2006JD007340.  (64)  Murphy, D. M.; Hudson, P. K.; Thomson, D. S.; Sheridan, P. J.; Wilson, J. C. Environ. Sci. Technol. 2006, 40, 3163-3167.  (65)  Tolocka, M. P.; Lake, D. A.; Johnston, M. V.; Wexler, A. S. Atmos. Environ. 2004, 38, 3263-3273.  (66)  Liu, P.; Ziemann, P. J.; Kittelson, D. B.; Mcmurry, P. H. Aerosol Sci. Tech. 1995, 22, 293-313.  (67)  Liu, P.; Ziemann, P. J.; Kittelson, D. B.; Mcmurry, P. H. Aerosol Sci. Tech. 1995, 22, 314-324.  (68)  Zhang, X. F.; Smith, K. A.; Worsnop, D. R.; Jimenez, J. L.; Jayne, J. T.; Kolb, C. E.; Morris, J.; Davidovits, P. Aerosol Sci. Tech. 2004, 38, 619-638.  (69)  Mckeown, P. J.; Johnston, M. V.; Murphy, D. M. Anal. Chem. 1991, 63, 20692073.  (70)  Thomson, D. S.; Murphy, D. M. Appl. Optics 1993, 32, 6818-6826.  (71)  Sinha, M. P. Rev. Sci. Instrum. 1984, 55, 886-891.  (72)  Sinha, M. P.; Friedlander, S. K. Anal. Chem. 1985, 57, 1880-1883.  (73)  Nordmeyer, T.; Prather, K. A. Anal. Chem. 1994, 66, 3540-3542.  (74)  Prather, K. A.; Nordmeyer, T.; Salt, K. Anal. Chem. 1994, 66, 1403-1407.  (75)  Yang, M.; Reilly, P. T. A.; Boraas, K. B.; Whitten, W. B.; Ramsey, J. M. Rapid Commun. Mass Spectrom. 1996, 10, 347-351.  (76)  Gard, E.; Mayer, J. E.; Morrical, B. D.; Dienes, T.; Fergenson, D. P.; Prather, K. A. Anal. Chem. 1997, 69, 4083-4091.  (77)  Cabalo, J.; Zelenyuk, A.; Baer, T.; Miller, R. E. Aerosol Sci. Tech. 2000, 33, 3-19.  129  (78)  Salt, K.; Noble, C. A.; Prather, K. A. Anal. Chem. 1996, 68, 230-234.  (79)  Jayne, J. T.; Leard, D. C.; Zhang, X. F.; Davidovits, P.; Smith, K. A.; Kolb, C. E.; Worsnop, D. R. Aerosol Sci. Tech. 2000, 33, 49-70.  (80)  Noble, C. A.; Prather, K. A. Environ. Sci. Technol. 1996, 30, 2667-2680.  (81)  Su, Y. X.; Sipin, M. F.; Furutani, H.; Prather, K. A. Anal. Chem. 2004, 76, 712719.  (82)  Harris, W. A.; Reilly, P. T. A.; Whitten, W. B.; Ramsey, J. M. Rev. Sci. Instrum. 2005, 76, -.  (83)  Carson, P. G.; Johnston, M. V.; Wexler, A. S. Rapid Commun. Mass Spectrom. 1997, 11, 993-996.  (84)  Cziczo, D. J.; Thomson, D. S.; Thompson, T. L.; DeMott, P. J.; Murphy, D. M. Int. J. Mass Spectrom. 2006, 258, 21-29.  (85)  Drewnick, F.; Hings, S. S.; DeCarlo, P.; Jayne, J. T.; Gonin, M.; Fuhrer, K.; Weimer, S.; Jimenez, J. L.; Demerjian, K. L.; Borrmann, S.; Worsnop, D. R. Aerosol Sci. Tech. 2005, 39, 637-658.  (86)  Nash, D. G.; Baer, T.; Johnston, M. V. Int. J. Mass Spectrom. 2006, 258, 2-12.  (87)  Mansoori, B. A.; Johnston, M. V.; Wexler, A. S. Anal. Chem. 1994, 66, 36813687.  (88)  Dessiaterik, Y.; Nguyen, T.; Baer, T.; Miller, R. E. J. Phys. Chem. A 2003, 107, 11245-11252.  (89)  Dessiaterik, Y.; Baer, T.; Miller, R. E. J. Phys. Chem. A 2006, 110, 1500-1505.  (90)  Gross, D. S.; Galli, M. E.; Silva, P. J.; Prather, K. A. Anal. Chem. 2000, 72, 416422.  (91)  Kane, D. B.; Johnston, M. V. Environ. Sci. Technol. 2000, 34, 4887-4893.  (92)  Morrical, B. D.; Fergenson, D. P.; Prather, K. A. J. Am. Soc. Mass Spectrom. 1998, 9, 1068-1073.  (93)  Sykes, D. C.; Woods, E.; Smith, G. D.; Baer, T.; Miller, R. E. Anal. Chem. 2002, 74, 2048-2052.  (94)  Bente, M.; Sklorz, M.; Streibel, T.; Zimmermann, R. Anal. Chem. 2009, 81, 25252536.  (95)  Tobias, H. J.; Kooiman, P. M.; Docherty, K. S.; Ziemann, P. J. Aerosol Sci. Tech. 2000, 33, 170-190. 130  (96)  Wilson, K. R.; Jimenez-Cruz, M.; Nicolas, C.; Belau, L.; Leonet, S. R.; Ahmed, M. J. Phys. Chem. A 2006, 110, 2106-2113.  (97)  Dalton, C. N.; Jaoui, M.; Kamens, R. M.; Glish, G. L. Anal. Chem. 2005, 77, 3156-3163.  (98)  Zelenyuk, A.; Cabalo, J.; Baer, T.; Miller, R. E. Anal. Chem. 1999, 71, 18021808.  (99)  Woods, E.; Smith, G. D.; Dessiaterik, Y.; Baer, T.; Miller, R. E. Anal. Chem. 2001, 73, 2317-2322.  (100) Woods, E.; Smith, G. D.; Miller, R. E.; Baer, T. Anal. Chem. 2002, 74, 16421649. (101) Gloaguen, E.; Mysak, E. R.; Leone, S. R.; Ahmed, M.; Wilson, K. R. Int. J. Mass Spectrom. 2006, 258, 74-85. (102) LaFranchi, B. W.; Zahardis, J.; Petrucci, G. A. Rapid Commun. Mass Spectrom. 2004, 18, 2517-2521. (103) Lazar, A. C.; Reilly, P. T. A.; Whitten, W. B.; Ramsey, J. M. Anal. Chem. 2000, 72, 2142-2147. (104) Hinz, K. P.; Kaufmann, R.; Spengler, B. Aerosol Sci. Tech. 1996, 24, 233-242. (105) Hinz, K. P.; Kaufmann, R.; Spengler, B. Anal. Chem. 1994, 66, 2071-2076. (106) Yang, M.; Dale, J. M.; Whitten, W. B.; Ramsey, J. M. Anal. Chem. 1995, 67, 1021-1025. (107) Reilly, P. T. A.; Gieray, R. A.; Yang, M.; Whitten, W. B.; Ramsey, J. M. Anal. Chem. 1997, 69, 36-39. (108) Hoffmann, T.; Bandur, R.; Marggraf, U.; Linscheid, M. J. Geophys. Res. [Atmos.] 1998, 103, 25569-25578. (109) Kürten, A.; Curtius, J.; Helleis, F.; Lovejoy, E. R.; Borrmann, S. Int. J. Mass Spectrom. 2007, 265, 30-39. (110) Hanna, S. J.; Campuzano-Jost, P.; Simpson, E. A.; Robb, D. B.; Burak, I.; Blades, M. W.; Hepburn, J. W.; Bertram, A. K. Int. J. Mass Spectrom. 2009, 279, 134146. (111) Simpson, E. A.; Campuzano-Jost, P.; Hanna, S. J.; Robb, D. B.; Hepburn, J. W.; Blades, M. W.; Bertram, A. K. Int. J. Mass Spectrom. 2009, 281, 140-149. (112) Gieray, R. A.; Reilly, P. T. A.; Yang, M.; Whitten, W. B.; Ramsey, J. M. Anal. Chem. 1998, 70, 117-120. 131  (113) Dzepina, K.; Arey, J.; Marr, L. C.; Worsnop, D. R.; Salcedo, D.; Zhang, Q.; Onasch, T. B.; Molina, L. T.; Molina, M. J.; Jimenez, J. L. Int. J. Mass Spectrom. 2007, 263, 152-170. (114) Zhang, Q.; Stanier, C. O.; Canagaratna, M. R.; Jayne, J. T.; Worsnop, D. R.; Pandis, S. N.; Jimenez, J. L. Environ. Sci. Technol. 2004, 38, 4797-4809. (115) Bente, M.; Adam, T.; Ferge, T.; Gallavardin, S.; Sklorz, M.; Streibel, T.; Zimmermann, R. Int. J. Mass Spectrom. 2006, 258, 86-94. (116) Dall'Osto, M.; Harrison, R. M.; Charpantidou, E.; Loupa, G.; Rapsomanikis, S. Sci. Total Environ. 2007, 384, 120-133. (117) Shu, J. N.; Gao, S. K.; Li, Y. Aerosol Sci. Tech. 2008, 42, 110-113. (118) Angelino, S.; Suess, D. T.; Prather, K. A. Environ. Sci. Technol. 2001, 35, 31303138. (119) Yadav, R.; Saoud, K.; Rasouli, F.; Hajaligol, M.; Fenner, R. Journal of Analytical and Applied Pyrolysis 2004, 72, 17-25. (120) Reilly, P. T. A.; Lazar, A. C.; Gieray, R. A.; Whitten, W. B.; Ramsey, J. M. Aerosol Sci. Tech. 2000, 33, 135-152. (121) Flagan, R. C. In Aerosol Measurement: Principles, Techniques, and Applications; Baron, P. A., Willeke, K., Eds.; John Wiley: New York, 2001, pp 537-568. (122) DeCarlo, P. F.; Slowik, J. G.; Worsnop, D. R.; Davidovits, P.; Jimenez, J. L. Aerosol Sci. Tech. 2004, 38, 1185-1205. (123) Baron, P. A.; Sorensen, C. M.; Brockmann, J. E. In Aerosol Measurement: Principles, Techniques, and Applications; Baron, P. A., Willeke, K., Eds.; John Wiley: Newy York, 2001, pp 705-749. (124) Zelenyuk, A.; Yang, J.; Imre, D. Int. J. Mass Spectrom. 2009, 282, 6-12. (125) Zelenyuk, A.; Cai, Y.; Imre, D. Aerosol Sci. Tech. 2006, 40, 197-217. (126) Zelenyuk, A.; Cai, Y.; Chieffo, L.; Imre, D. Aerosol Sci. Tech. 2005, 39, 972-986. (127) Zelenyuk, A.; Imre, D.; Han, J. H.; Oatis, S. Anal. Chem. 2008, 80, 1401-1407. (128) McMurry, P. H.; Wang, X.; Park, K.; Ehara, K. Aerosol Sci. Tech. 2002, 36, 227238. (129) Hand, J. L.; Kreidenweis, S. M. Aerosol Sci. Tech. 2002, 36, 1012-1026. (130) Tang, I. N.; Munkelwitz, H. R. J. Geophys. Res. [Atmos.] 1994, 99, 18801-18808.  132  (131) Lide, D. R., Ed. CRC Handbook of Chemistry and Physics, 89 ed.; CRC Press: Boca Raton, 2008-2009. (132) Kurt J. Lesker Company. Fluid Properties Cross-references. http://www.lesker.com/newweb/fluids/fluids_selectionguide_properties.cfm?pgid =0 (accessed June 2009). (133) Quarmby, S. T.; Yost, R. A. Int. J. Mass Spectrom. 1999, 191, 81-102. (134) Ferge, T.; Mühlberger, F.; Zimmermann, R. Anal. Chem. 2005, 77, 4528-4538. (135) Focsa, C.; Mihesan, C.; Ziskind, M.; Chazallon, B.; Therssen, E.; Desgroux, P.; Destombes, J. L. J. Phys.: Condens. Matter 2006, 18, S1357-S1387. (136) Woods, E.; Miller, R. E.; Baer, T. J. Phys. Chem. A 2003, 107, 2119-2125. (137) Parra, E.; Alexeev, I.; Fan, J. Y.; Kim, K. Y.; McNaught, S. J.; Milchberg, H. M. J. Opt. Soc. Am. B: Opt. Phys 2003, 20, 118-124. (138) Menzel, C.; Dreisewerd, K.; Berkenkamp, S.; Hillenkamp, F. J. Am. Soc. Mass Spectrom. 2002, 13, 975-984. (139) Lincoln, K. A. Int. J. Mass Spectrom. Ion Phys. 1969, 2, 75-83. (140) Lincoln, K. A. Int. J. Mass Spectrom. Ion Phys. 1974, 13, 45-53. (141) Vanbreemen, R. B.; Snow, M.; Cotter, R. J. International Journal of Mass Spectrometry and Ion Processes 1983, 49, 35-50. (142) Fenselau, C.; Emary, W. B.; Cotter, R. J. Org. Mass Spectrom. 1989, 24, 694-698. (143) Yuan, Z. M.; Fenselau, C.; Dulik, D. M.; Martin, W.; Emary, W. B.; Brundrett, R. B.; Colvin, O. M.; Cotter, R. J. Anal. Chem. 1990, 62, 868-870. (144) Drewnick, F.; Wieser, P. H. Rev. Sci. Instrum. 2002, 73, 3003-3006. (145) Zelenyuk, A.; Imre, D.; Cuadra-Rodriguez, L. A. Anal. Chem. 2006, 78, 69426947. (146) Schoolcraft, T. A.; Constable, G. S.; Zhigilei, L. V.; Garrison, B. J. Anal. Chem. 2000, 72, 5143-5150. (147) Schoolcraft, T. A.; Constable, G. S.; Jackson, B.; Zhigilei, L. V.; Garrison, B. J. Nucl Instrum Meth B 2001, 180, 245-250. (148) Wood, C. F.; Leach, D. H.; Zhang, J. Z.; Chang, R. K.; Barber, P. W. Appl. Optics 1988, 27, 2279-2286. (149) Robb, D.; Blades, M. W. J. Am. Soc. Mass Spectrom. 1997, 8, 1203-1205. 133  (150) Specht, A. A.; Blades, M. W. J. Am. Soc. Mass Spectrom. 2003, 14, 562-570. (151) Rolland, D.; Specht, A. A.; Blades, M. W.; Hepburn, J. W. Chem. Phys. Lett. 2003, 373, 292-298. (152) Stafford, G. C.; Kelley, P. E.; Syka, J. E. P.; Reynolds, W. E.; Todd, J. F. J. International Journal of Mass Spectrometry and Ion Processes 1984, 60, 85-98. (153) Wu, H.-F.; Brodbelt, J. S. International Journal of Mass Spectrometry and Ion Processes 1992, 115, 67-81. (154) March, R. E. International Journal of Mass Spectrometry and Ion Processes 1992, 118, 71-135. (155) March, R. E. Rapid Commun. Mass Spectrom. 1998, 12, 1543-1554. (156) Julian, R. K.; Cooks, R. G. Anal. Chem. 1993, 65, 1827-1833. (157) Doroshenko, V. M.; Cotter, R. J. Rapid Commun. Mass Spectrom. 1996, 10, 6573. (158) Soni, M. H.; Cooks, R. G. Anal. Chem. 1994, 66, 2488-2496. (159) March, R. E. J. Mass Spectrom. 1997, 32, 351-369. (160) Stein, S. E. In NIST Chemistry WebBook, NIST Standard Reference Database Number 69; Linstrom, P. J., Mallard, W. G., Eds.; National Institute of Standards and Technology: Gaithersburg MD, June 2005. (161) March, R. E. In Practical Aspects of Ion Trap Mass Spectrometry: Chemical, Environmental, and Biomedical Applications; March, R. E., Todd, J. F. J., Eds.; CRC Press, Inc.: Boca Raton, 1995; Vol. III, pp 34-59. (162) Northway, M. J.; Jayne, J. T.; Toohey, D. W.; Canagaratna, M. R.; Trimborn, A.; Akiyama, K. I.; Shimono, A.; Jimenez, J. L.; DeCarlo, P. F.; Wilson, K. R.; Worsnop, D. R. Aerosol Sci. Tech. 2007, 41, 828-839. (163) Nash, D. G.; Liu, X. F.; Mysak, E. R.; Baer, T. Int. J. Mass Spectrom. 2005, 241, 89-97. (164) Mysak, E. R.; Wilson, K. R.; Jimenez-Cruz, M.; Ahmed, M.; Baer, T. Anal. Chem. 2005, 77, 5953-5960. (165) Verenitch, S. S.; Mazumder, A. Anal. Bioanal. Chem. 2008, 391, 2635-2646. (166) Hanna, S. J.; Campuzano-Jost, P.; Simpson, E. A.; Burak, I.; Blades, M. W.; Hepburn, J. H.; Bertram, A. K. Phys. Chem. Chem. Phys. 2009, 11, 7963-7975, DOI: 10.1039/b904748d.  134  (167) Woods, E.; Dessiaterik, Y.; Miller, R. E.; Baer, T. J. Phys. Chem. A 2001, 105, 8273-8280. (168) Cross, E. S.; Onasch, T. B.; Canagaratna, M.; Jayne, J. T.; Kimmel, J.; Yu, X. Y.; Alexander, M. L.; Worsnop, D. R.; Davidovits, P. Atmos. Chem. Phys. 2009, 9, 7769-7793. (169) Ragheb, M. S.; Zakhary, S. G. Radiat. Phys. Chem. 2000, 58, 1-8. (170) Voisin, D.; Smith, J. N.; Sakurai, H.; McMurry, P. H.; Eisele, F. L. Aerosol Sci. Tech. 2003, 37, 471-475. (171) Zahardis, J.; LaFranchi, B. W.; Petrucci, G. A. Atmos. Environ. 2006, 40, 16611670. (172) Morrical, B. D.; Zenobi, R. Atmos. Environ. 2002, 36, 801-811. (173) Shu, J. N.; Wilson, K. R.; Ahmed, M.; Leone, S. R. Rev. Sci. Instrum. 2006, 77, 043106. (174) Vanbramer, S. E.; Johnston, M. V. Appl. Spectrosc. 1992, 46, 255-261. (175) Vanbramer, S. E.; Ross, P. L.; Johnston, M. V. J. Am. Soc. Mass Spectrom. 1993, 4, 65-72. (176) Oktem, B.; Tolocka, M. P.; Johnston, M. V. Anal. Chem. 2004, 76, 253-261. (177) National Institute of Standards and Technology. http://webbook.nist.gov/chemistry/. (accessed May 27 2009). (178) Adam, T.; Zimmermann, R. Anal. Bioanal. Chem. 2007, 389, 1941-1951. (179) Hanna, S. J. Development and characterization of a single particle mass spectrometer with soft photoionization for organic aerosol studies. Ph.D. Thesis, University of British Columbia, Vancouver, BC, July 2009. (180) Lipson, R. H. In An Introduction to Laser Spectroscopy; Andrews, D. L., Demidov, A. A., Eds.; Kluwer Academic: New York, 2002, pp 257-309. (181) Butcher, D. J.; Goeringer, D. E.; Hurst, G. B. Anal. Chem. 1999, 71, 489-496. (182) Vanbramer, S. E.; Johnston, M. V. J. Am. Soc. Mass Spectrom. 1990, 1, 419-426. (183) Shi, Y. J.; Lipson, R. H. Can. J. Chem. 2005, 83, 1891-1902. (184) Mühlberger, F.; Zimmermann, R.; Kettrup, A. Anal. Chem. 2001, 73, 3590-3604. (185) Sheridan, P. J.; Brock, C. A.; Wilson, J. C. Geophys. Res. Lett. 1994, 21, 25872590. 135  (186) Turpin, B. J.; Saxena, P.; Allen, G.; Koutrakis, P.; McMurry, P.; Hildemann, L. J Air Waste Manage 1997, 47, 344-356. (187) Kline, J.; Huebert, B.; Howell, S.; Blomquist, B.; Zhuang, J.; Bertram, T.; Carrillo, J. J. Geophys. Res. [Atmos.] 2004, 109, D19S08, doi:10.1029/2004JD004540. (188) Ma, Y.; Weber, R. J.; Maxwell-Meier, K.; Orsini, D. A.; Lee, Y. N.; Huebert, B. J.; Howell, S. G.; Bertram, T.; Talbot, R. W.; Dibb, J. E.; Scheuer, E. J. Geophys. Res. [Atmos.] 2004, 109, D15S06, doi:10.1029/2003JD003673. (189) Bahreini, R.; Jimenez, J. L.; Wang, J.; Flagan, R. C.; Seinfeld, J. H.; Jayne, J. T.; Worsnop, D. R. J. Geophys. Res. [Atmos.] 2003, 108, 8645, doi:10.1029/2002JD003226. (190) Graber, E. R.; Rudich, Y. Atmos. Chem. Phys. 2006, 6, 729-753. (191) Havers, N.; Burba, P.; Lambert, J.; Klockow, D. J Atmos Chem 1998, 29, 45-54. (192) Mukai, H.; Ambe, Y. Atmos. Environ. 1986, 20, 813-819. (193) Rogge, W. F.; Mazurek, M. A.; Hildemann, L. M.; Cass, G. R.; Simoneit, B. R. T. Atmospheric Environment Part A - General Topics 1993, 27, 1309-1330. (194) Fuzzi, S.; Decesari, S.; Facchini, M. C.; Matta, E.; Mircea, M.; Tagliavini, E. Geophys. Res. Lett. 2001, 28, 4079-4082. (195) Dinar, E.; Mentel, T. F.; Rudich, Y. Atmos. Chem. Phys. 2006, 6, 5213-5224. (196) Facchini, M. C.; Fuzzi, S.; Zappoli, S.; Andracchio, A.; Gelencser, A.; Kiss, G.; Krivacsy, Z.; Meszaros, E.; Hansson, H. C.; Alsberg, T.; Zebuhr, Y. J. Geophys. Res. [Atmos.] 1999, 104, 26821-26832. (197) Decesari, S.; Facchini, M. C.; Fuzzi, S.; Tagliavini, E. J. Geophys. Res. [Atmos.] 2000, 105, 1481-1489. (198) Morawska, L.; Hofmann, W.; Hitchins-Loveday, J.; Swanson, C.; Mengersen, K. J. Aerosol Sci. 2005, 36, 939-957. (199) Adam, T.; Mitschke, S.; Streibel, T.; Baker, R. R.; Zimmermann, R. Anal. Chim. Acta 2006, 572, 219-229. (200) Adam, T.; Baker, R. R.; Zimmermann, R. Anal. Bioanal. Chem. 2007, 387, 575584. (201) Carre, V.; Aubriet, F.; Muller, J. F. Anal. Chim. Acta 2005, 540, 257-268. (202) Mitschke, S.; Adam, T.; Streibel, T.; Baker, R. R.; Zimmermann, R. Anal. Chem. 2005, 77, 2288-2296. 136  (203) Hildemann, L. M.; Markowski, G. R.; Cass, G. R. Environ. Sci. Technol. 1991, 25, 744-759. (204) Stedman, R. L. Chem. Rev. 1968, 68, 153–207. (205) Mohr, C.; Huffman, J. A.; Cubison, M. J.; Aiken, A. C.; Docherty, K. S.; Kimmel, J. R.; Ulbricht, I. M.; Hannigan, M.; Jimenez, J. L. Environ. Sci. Technol. 2009, 43, 2443-2449. (206) International Humic Substances Society. Spectroscopic Properties of IHSS Samples. http://ihss.gatech.edu/ihss2/spectra.html (accessed May 2009). (207) International Humic Substances Society. Elemental Compositions and Stable Isotopic Ratios of IHSS Samples. http://ihss.gatech.edu/ihss2/elements.html (accessed May 2009). (208) Taylor, H. E.; Garbarino, J. R. In Humic substances in the Suwannee River, Georgia: interactions, properties, and proposed structures; Averett, R. C., Leenheer, J. A., McKnight, D. M., Thorn, K. A., Eds.; U.S. Geological Survey: Denver, 1994, pp 45-55. (209) Leenheer, J. A.; McKnight, D. M.; Thurman, E. M.; MacCarthy, P. In Humic Substances in the Suwannee River, Georgia: Interactions, Properties, and Proprosed Structures; Averett, R. C., Leenheer, J. A., McKnight, D. M., Thorn, K. A., Eds.; U.S. GPO: Denver, 1994, pp 195-211. (210) Krivacsy, Z.; Gelencser, A.; Kiss, G.; Meszaros, E.; Molnar, A.; Hoffer, A.; Meszaros, T.; Sarvari, Z.; Temesi, D.; Varga, B.; Baltensperger, U.; Nyeki, S.; Weingartner, E. J Atmos Chem 2001, 39, 235-259. (211) Zhang, Q.; Alfarra, M. R.; Worsnop, D. R.; Allan, J. D.; Coe, H.; Canagaratna, M. R.; Jimenez, J. L. Environ. Sci. Technol. 2005, 39, 4938-4952. (212) Schauer, J. J.; Kleeman, M. J.; Cass, G. R.; Simoneit, B. R. T. Environ. Sci. Technol. 1999, 33, 1566-1577. (213) Dreisewerd, K.; Berkenkamp, S.; Leisner, A.; Rohlfing, A.; Menzel, C. Int. J. Mass Spectrom. 2003, 226, 189-209. (214) Menzel, C.; Berkenkamp, S.; Hillenkamp, F. Rapid Commun. Mass Spectrom. 1999, 13, 26-32. (215) Zelenyuk, A.; Yang, J.; Choi, E.; Imre, D. Aerosol Sci. Tech. 2009, 43, 411-424. (216) Zahardis, J.; Geddes, S.; Petrucci, G. A. Int. J. Environ. Anal. Chem. 2008, 88, 177-184. (217) Rosen, E. P.; Garland, E. R.; Baer, T. J. Phys. Chem. A 2008, 112, 10315-10324.  137  (218) Broekhuizen, K. E.; Thornberry, T.; Kumar, P. P.; Abbatt, J. P. D. J. Geophys. Res. [Atmos.] 2004, 109, D24206, doi:10.1029/2004JD005298. (219) de Gouw, J. A.; Lovejoy, E. R. Geophys. Res. Lett. 1998, 25, 931-934. (220) George, I. J.; Vlasenko, A.; Slowik, J. G.; Broekhuizen, K.; Abbatt, J. P. D. Atmos. Chem. Phys. 2007, 7, 4187-4201. (221) McNeill, V. F.; Wolfe, G. M.; Thornton, J. A. J. Phys. Chem. A 2007, 111, 10731083. (222) Morris, J. W.; Davidovits, P.; Jayne, J. T.; Jimenez, J. L.; Shi, Q.; Kolb, C. E.; Worsnop, D. R.; Barney, W. S.; Cass, G. Geophys. Res. Lett. 2002, 29, 1357, doi:10.1029/2002GL014692. (223) Robinson, A. L.; Donahue, N. M.; Rogge, W. F. J. Geophys. Res. [Atmos.] 2006, 111, D03302, doi:10.1029/2005JD006265. (224) McNeill, V. F.; Yatavelli, R. L. N.; Thornton, J. A.; Stipe, C. B.; Landgrebe, O. Atmos. Chem. Phys. 2008, 8, 5465-5476. (225) Molina, M. J.; Ivanov, A. V.; Trakhtenberg, S.; Molina, L. T. Geophys. Res. Lett. 2004, 31, L22104, doi:10.1029/2004GL020910. (226) Ellison, G. B.; Tuck, A. F.; Vaida, V. J. Geophys. Res. [Atmos.] 1999, 104, 11633-11641. (227) Kwan, A. J.; Crounse, J. D.; Clarke, A. D.; Shinozuka, Y.; Anderson, B. E.; Crawford, J. H.; Avery, M. A.; McNaughton, C. S.; Brune, W. H.; Singh, H. B.; Wennberg, P. O. Geophys. Res. Lett. 2006, 33, L15815, doi:10.1029/2006GL026144. (228) Gross, S.; Bertram, A. K. J. Phys. Chem. A 2008, 112, 3104-3113. (229) Gross, S.; Bertram, A. K. J. Geophys. Res. [Atmos.] 2009, 114, D02307, doi:10.1029/2008JD010987. (230) Gross, S.; Knopf, D. A.; Mak, J.; Iannone, R.; Bertram, A. K. Geochim Cosmochim Ac 2009, 73, A470-A470. (231) Mak, J.; Gross, S.; Bertram, A. K. Geophys. Res. Lett. 2007, 34, L10804, doi:10.1029/2007GL029756. (232) Pitts, J. N.; Atkinson, R.; Sweetman, J. A.; Zielinska, B. Atmos. Environ. 1985, 19, 701-705. (233) Pitts, J. N.; Sweetman, J. A.; Zielinska, B.; Atkinson, R.; Winer, A. M.; Harger, W. P. Environ. Sci. Technol. 1985, 19, 1115-1121.  138  (234) Knopf, D. A.; Mak, J.; Gross, S.; Bertram, A. K. Geophys. Res. Lett. 2006, 33, L17816, doi:10.1029/2006GL026884. (235) Cosman, L. M.; Knopf, D. A.; Bertram, A. K. J. Phys. Chem. A 2008, 112, 23862396. (236) Cosman, L. M.; Bertram, A. K. J. Phys. Chem. A 2008, 112, 4625-4635.  139  Appendix I. Determination of particle density using a DMA coupled to the SPITMS interface As briefly described in section 3.5, the velocity measurements with the SPITMS interface can be used in tandem with a DMA to determine particle density for spherical particles. Additional details on these measurements are included below. Chemicals used include canola oil (store-bought), oleic acid (Aldrich, ≥99%), pump oil (Kurt J. Lesker Vacuum Pump Fluid, TKO 19 Ultra), sodium bisulfate (MCB reagents, technical grade), ammonium bisulfate (Aldrich, ≥99%), and a 10% w/w solution of sodium bisulfate/ammonium bisulfate (same as previous compounds). Particles were generated by aerosolizing solutions of the above compounds using a constant output atomizer (TSI Inc., Model 3076) or a vibrating orifice aerosol generator (TSI Inc., Model 3450). Aerosols were then passed through a 85Kr charge neutralizer (TSI Inc., Model 3054) with a large dilution flow of air as well as a 24″ nafion diffusion dryer (Permapure Inc) to ensure particles are thoroughly dried before they reach the DMA. This will avoid size changes in the aerosol lens and hence inaccuracies in the measured densities.145 After a monodisperse sized aerosol is selected through the DMA, the majority of the aerosol flow is sent to the condensation particle counter (TSI Inc. model W3782) to monitor the particle number concentration, and a small fraction is drawn into the SPITMS for aerodynamic sizing. Particle time of flight distributions are recorded with the realtime software and then converted to velocity by using the known separation of the two scattering lasers. The TSI differential mobility analyzer (DMA) is used to select particles by mobility diameter. The mobility diameter (dm) of a particle is defined as the diameter of a sphere (of unit charge) with the same migration velocity in a constant electric field as the particle under study.121 For spherical particles with no internal voids, the mobility diameter, particle diameter, and volume equivalent diameter are all equal. The density for a spherical particle can then be determined as follows: Equation A 1  ρ p = ρo  d va dm 140  where ρp is the particle density, ρo is the standard density of 1.0 g/cm3, dva is the vacuum aerodynamic diameter, and dm is the mobility diameter.126 The velocimetry data is then fit with Equation A 2 below, which was obtained by replacing the volume equivalent diameter with the mobility diameter in Equation 2, which are equal for a spherical particle.  dm ρ p v = α ⋅ d vac = α ⋅   χρ o   Equation A 2       c  ν is the particle velocity, α and c are fixed parameters, the shape factor χ is set equal to one, the mobility diameter is known from the size-selection, and the density term (ρp) is allowed to vary to find the best fit. Figure A 1 shows the velocimetry measurements of a range of size-selected canola oil particles. The error bars in the particle velocity reflect the distribution width of velocities obtained rather than the error of the determination of the most probable velocity. The α and c parameters are determined previously from the PSL calibration curve (see section 3.4.1). The density determined was 0.933 g/cm3. The r2 value for the fit is > 0.99, indicating the behavior is well-described by the two parameter fit. 140 135  R^2  = 0.99871  130  α ρ χ  0.52037 0.93316 1 ±0 -0.35795  Particle velocity (m/s)  125 120  c  ±0 ±0.00442 ±0  115 110 105 100 95 90 85 80 75 150 200 250 300 350 400 450 500 550 600 650 700 750 800  Mobility diameter (nm)  Figure A 1: Density determination for canola oil particles from measured particle velocity versus mobility diameter.  141  Figure A 2 shows the fit applied to a range of size-selected sodium bisulfate particles and their measured velocities in the SPITMS. The density determined was 2.25 g/cm3. The r2 value for the fit is > 0.99, indicating the behavior is well-described by the two parameter fit. Particle densities can be determined in this way for solid or liquid spherical particles using simultaneous measurements of the vacuum aerodynamic diameter with the aerosol ion trap mass spectrometer and the mobility diameter with a differential mobility analyzer. Results are summarized in section 3.5 and Table 3.2 of the main body of the thesis. 100  95  R^2  = 0.99406  α ρ χ  0.52037 ±0 2.25166 ±0.014 1 ±0 -0.35795 ±0  Particle velocity (m/s)  c  90  85  80  75  70 200  250  300  350  400  450  Mobility diameter (nm)  Figure A 2: Density determination for sodium bisulfate particles from measured particle velocity versus mobility diameter.  142  Appendix II. Determination of oleyl alcohol ionization appearance energy The PIE curves were obtained by continuously scanning the VUV wavelength and recording both the ion signal and VUV power for each laser shot. This way, ion signal could be normalized to the VUV power on a shot to shot basis. Particles of ~1 µm diameter were used. A desorption energy of 15 mJ at 944 cm-1 was used. Figure A 3 shows the photoionization efficiency (PIE) curves obtained for four significant oleyl alcohol fragments. The water loss fragment (m/z = 250) and three of the larger fragments (m/z = 138, 110, and 82) are shown for comparison in Figure A 3 since the molecular ion was not observed. The gap between 8.42 to 8.48 eV occurs in the scan because of the strong resonance line in xenon which prevents VUV generation in this region. The observed ionization threshold is obtained by extrapolating the linear portion of the threshold region of the PIE curve to the baseline. In general, a low energy tail is also observed that does not fall into the linear portion of the threshold and indicates an onset of ionization at energies below the literature value for the IE.110, 166 All fragments appeared at similar ionization energies. For the four fragments shown the extrapolated IEs were 8.58, 8.55, 8.56, and 8.57 eV. Although the molecular ion was not observed in this case, results from our previous work showed that both oleic acid fragment and parent ions appeared at the same ionization energy indicating a postionization fragmentation process166. A post-ionization fragmentation process means that vaporization of the particle produces the neutral molecule which is then ionized by the VUV light and subsequently decays to give fragments. Assuming this process also holds true in this case, which seems reasonable given the similar appearance energies for fragments of widely varying mass to charge ratio, the ionization energy for oleyl alcohol was determined based on an average of these four fragments to be 8.56 ± 0.05 eV. The uncertainty in these measurements arises mainly from the uncertainty in fitting a line to the linear portion of the PIE curve.  143  200  m/z m/z m/z m/z  180 160  Ion Signal [a.u.]  140  250 138 110 82  120 100 80 60 40 20 0 8.2  8.3  8.4  8.5  8.6  8.7  8.8  8.9  VUV ionization energy (eV)  Figure A 3: Photoionization efficiency curves for oleyl alcohol fragments (15 mJ 944 cm-1 CO2, 5 µs ionization delay time, 1 ms cool time). The extrapolated appearance energy is 8.56 ± 0.05 eV. Data is shown with a 50 point smooth applied to each curve.  144  Appendix III. Calculation of energy absorbed per particle during desorption Beer’s law, which describes the absorption of light by a sample is given below, where I is the intensity of light measured after the sample, Io is the initial intensity of light, σ is the absorption cross-section, c is the concentration of the species, and l is the path length the light travels:  Equation A 3  I  − ln   = σ ⋅ c ⋅ l  Io   In this case we are interested in the CO2 energy absorbed by the particle (Eabs). The energy measured after the particle is equal to the the difference of the initial CO2 energy (Ei) and the energy absorbed by the particle (Eabs), thus the equation can be rewritten:  Equation A 4   E − E abs   = σ ⋅ c ⋅ l − ln  i E   i  Solving for the absorbed CO2 energy gives: Equation A 5  E abs = E i (1 − exp(− σ ⋅ c ⋅ l ))  The initial CO2 energy the particle experiences is defined as the measured CO2 energy times the fraction of the CO2 beam intercepted by the particle: Equation A 6  Ei = ECO2 ⋅  Ap ACO2  where ECO2 is the measured CO2 energy, Ap is the area of the particle and ACO2 is the area of the CO2 beam where it intersects the particle. The fraction of Ap over ACO2 is included to take into account the geometric fraction of a flat beam profile intersecting the particle. We know the CO2 profile is Gaussian, so we could be off be as much as a factor of 2 in  145  the absolute energy absorbed per particle, but the relative comparison between the pure and mixed particles will not be affected. Equation A 6 can be rewritten as follows:  Equation A 7  E abs = ECO ⋅ 2  Ap (1 − exp(− σ ⋅ c ⋅ l )) ACO 2  The absorption cross-section (σ) was measured with an FTIR for pure components of oleyl alcohol or oleic acid and is given in units of cm2/molec. The concentration c is given as the number density (Np) in the particle (molec/cm3). The path length l (units of cm) is approximated by calculating the height of a cylinder with the same volume and radius (rp) as the particle studied which yields a path length of 4rp/3. There is some uncertainty in the absorption cross-section term as absorption cross-sections can vary with solvents and vary with temperature during the heating process. Also, the absorption cross-sections determined for bulk solutions do not necessarily apply directly to aerosol particles. Thus Eabs can be determined as follows:  Equation A 8  E abs = ECO2 ⋅  Ap ACO2     4r p ⋅ 1 − exp − σ ⋅ N p ⋅     3         For the two component particle, Equation A 8 can be rewritten as follows where the contribution from each component is taken into account:  Equation A 9  E abs = ECO2 ⋅  4rp   A p   1 − exp − (σ 1 N 1 + σ 2 N 2 )  3   ACO2    σ1 and σ2 are the respective cross-sections of each of the pure components and N1 and N2 are the molecular densities of each species in the mixed particle.  146  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0061115/manifest

Comment

Related Items