<%BANNER%>

Fundamentals and Applications of High-Field Asymmetric Waveform Ion Mobility Spectrometry for the Analysis of Explosives

Permanent Link: http://ufdc.ufl.edu/UFE0024838/00001

Material Information

Title: Fundamentals and Applications of High-Field Asymmetric Waveform Ion Mobility Spectrometry for the Analysis of Explosives
Physical Description: 1 online resource (153 p.)
Language: english
Creator: Wu, Ching-Hong
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: apci, api, dms, dpis, explosives, faims, ims, ion, ionization, mass, mobility, quantitation, reactant, separation, spectrometry
Chemistry -- Dissertations, Academic -- UF
Genre: Chemistry thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Over the past ten years, the world has been stunned and outraged by a series of attacks on civilian targets that used explosive devices. These attacks led to widespread demands for identification of the perpetrators, along with calls for improved security measures to prevent such incidents in the future. Detection techniques such as X-ray scanners, Raman spectroscopy, Terahertz spectroscopy and ion mobility spectrometry are currently in use or under development; however, none of these techniques are appropriate for all necessary applications. High-field asymmetric-waveform ion mobility spectrometry (FAIMS) coupled to a mass spectrometer is an alternative technique that provides improvements to mass spectral signal-to-noise, orthogonal/complementary ion separation to mass spectrometry, enhanced ion and complexation structural analysis, and potential for rapid analyte quantitation. The primary goal of this research is to contribute to the understanding of ionization by an atmospheric pressure ionization (API) source and ion behavior in a FAIMS cell to assist the future development of a portable explosive detector to investigate explosives in the field. In this work, the ionization mechanism of two API sources, atmospheric pressure chemical ionization (APCI) and distributed plasma ionization source (DPIS), are discussed. The spectra of eleven explosives ionized by both sources were collected and characterized. The results show that APCI provides a consistent and simple ionization, while DPIS presents more discrimination by various ion fragments and is more amenable for monitoring a certain classes of explosives in the field. Variation in FAIMS parameters, such as dispersion voltages (DV), compensation voltage (CV) scan rate, curtain gas flow rate, carrier gas composition, and electrode temperature, was explored for their effect on explosive ions. A systematic evaluation of the performance of API-FAIMS-MS demonstrates sensitivity at the picogram level, short detection time (30 seconds), and excellent resolution such that isomers of the same explosive can be successfully resolved. The results from these studies with the laboratory procedure show promise for FAIMS to be used as an explosive detector that presents a sensitive, selective, specific and rapid technique.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Ching-Hong Wu.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Yost, Richard A.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024838:00001

Permanent Link: http://ufdc.ufl.edu/UFE0024838/00001

Material Information

Title: Fundamentals and Applications of High-Field Asymmetric Waveform Ion Mobility Spectrometry for the Analysis of Explosives
Physical Description: 1 online resource (153 p.)
Language: english
Creator: Wu, Ching-Hong
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: apci, api, dms, dpis, explosives, faims, ims, ion, ionization, mass, mobility, quantitation, reactant, separation, spectrometry
Chemistry -- Dissertations, Academic -- UF
Genre: Chemistry thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Over the past ten years, the world has been stunned and outraged by a series of attacks on civilian targets that used explosive devices. These attacks led to widespread demands for identification of the perpetrators, along with calls for improved security measures to prevent such incidents in the future. Detection techniques such as X-ray scanners, Raman spectroscopy, Terahertz spectroscopy and ion mobility spectrometry are currently in use or under development; however, none of these techniques are appropriate for all necessary applications. High-field asymmetric-waveform ion mobility spectrometry (FAIMS) coupled to a mass spectrometer is an alternative technique that provides improvements to mass spectral signal-to-noise, orthogonal/complementary ion separation to mass spectrometry, enhanced ion and complexation structural analysis, and potential for rapid analyte quantitation. The primary goal of this research is to contribute to the understanding of ionization by an atmospheric pressure ionization (API) source and ion behavior in a FAIMS cell to assist the future development of a portable explosive detector to investigate explosives in the field. In this work, the ionization mechanism of two API sources, atmospheric pressure chemical ionization (APCI) and distributed plasma ionization source (DPIS), are discussed. The spectra of eleven explosives ionized by both sources were collected and characterized. The results show that APCI provides a consistent and simple ionization, while DPIS presents more discrimination by various ion fragments and is more amenable for monitoring a certain classes of explosives in the field. Variation in FAIMS parameters, such as dispersion voltages (DV), compensation voltage (CV) scan rate, curtain gas flow rate, carrier gas composition, and electrode temperature, was explored for their effect on explosive ions. A systematic evaluation of the performance of API-FAIMS-MS demonstrates sensitivity at the picogram level, short detection time (30 seconds), and excellent resolution such that isomers of the same explosive can be successfully resolved. The results from these studies with the laboratory procedure show promise for FAIMS to be used as an explosive detector that presents a sensitive, selective, specific and rapid technique.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Ching-Hong Wu.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Yost, Richard A.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024838:00001


This item has the following downloads:


Full Text

PAGE 1

1 FUNDAMENTALS AND APPLICATIONS OF HIGH-FIELD ASYMMETRIC WAVEFORM ION MOBILITY SPECTROMETRY FOR THE ANALYSIS OF EXPLOSIVES By ALEX CHING-HONG WU A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009

PAGE 2

2 2009 Alex Ching-Hong Wu

PAGE 3

3 To my parents and my beloved wife, Rosalind

PAGE 4

4 ACKNOWLEDGMENTS I would like to extend my thanks and apprecia tions to those who have contributed to my achievement, and for their support, encouragemen t and guidance. My deepest thanks and gratitude are addressed to my a dvisor, Dr. Richard Yost, for his consistent support and direction. He has provided me a great example of how to be a successful scientist and a decent person, and he has given me the academic and personal freedom to pursue the projects I enjoyed. I truly appreciated all the insightf ul and thoughtful guidance I received from him. A special thanks to Dr. David Powell for all of his help with the research about DPIS and for sharing his knowledge about mass spectrometry. Dr. Ben Smith is acknowledged as graduate advisor for his advice and guidance, but also as a scientist for answering questions related to my work. Sincere gratitude is extended to the other members of my committee, Dr. Ronald Castellano and Dr. Joseph Delfino, for their in tellectual conversations throughout my research. The members of the Yost groups, past and pres ent, are thanked for th eir help, suggestions, and friendship. I give special th anks to Mike Napolitano, Maril yn Prieto, Erick Molina, Leonard Rorrer, Rich Reich, Dan Magparangalan, and Dr. Dodge Baluya for their nice and warm friendship and critical reading of this dissertation. I would al so like to acknowledge Dr. Jennifer Bryant, Dr. Rachelle Landgraf, Dave Pirman, and Kyle Lunsford fo r their valuable contributions to my research. President Yu-Ih Hou, Dir ector-General Cho-Chi un Wang, and Commissioner Mao-Sui Huang are acknowledged for giving me this opportunity and supporting my study here in the United States. Last, but not least, I would like to thank my parents for their love and support. They have always given me lots of encouragement when I needed it and have been proud of me for whatever I had accomplished. I am grateful to my wife, Rosalind, and my son, Adam, both of whom have sacrificed a great deal of time for me over the past few years. I would especially like

PAGE 5

5 to thank Rosalind for her unfailing support and enc ouragement that made the completion of this dissertation possible. My precious Adam is acknowledged as my motiv ation to succeed.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........8 LIST OF FIGURES................................................................................................................ .........9 CHAPTER 1 INTRODUCTION AND INSTRUMENTATION.................................................................15 Background..................................................................................................................... ........15 High-Field Asymmetric Waveform Ion Mobility Spectroscopy............................................19 Quadrupole Ion Trap Mass Spectrometry...............................................................................22 Quadrupole Ion Trap Theory...........................................................................................22 Ion Motion in the Ion Trap..............................................................................................24 Finnigan LCQ..................................................................................................................25 Atmospheric Pressure Ionization............................................................................................27 Atmospheric Pressure Chemical Ionization (APCI)........................................................27 Distribution Plasma Ioni zation Source (DPIS)................................................................28 FAIMS/MS....................................................................................................................... ......29 Overview of Dissertation....................................................................................................... .29 2 PROPERTIES AND CHARACTERIZATION OF EXPLOSIVES BY ATMOSPHERIC PRESSURE IONIZATION (API)-MASS SPECTROMETRY.............................................39 Introduction................................................................................................................... ..........39 Experimental................................................................................................................... ........40 Atmospheric Pressure Chemical Ionization (APCI)........................................................40 Distributed Plasma Ioni zation Source (DPIS).................................................................41 Results and Discussion......................................................................................................... ..42 Reactant Ions.................................................................................................................. .42 Ionization Chemistry.......................................................................................................45 Nitroaromatic Compounds..............................................................................................46 TNT..........................................................................................................................46 TNB..........................................................................................................................47 Tetryl........................................................................................................................48 DNT..........................................................................................................................48 DNB.........................................................................................................................49 Nitramines..................................................................................................................... ..50 RDX.........................................................................................................................50 HMX.........................................................................................................................51 Nitrate Esters................................................................................................................. ..52 NG............................................................................................................................52 PETN........................................................................................................................53

PAGE 7

7Conclusions.................................................................................................................... .........54 3 FUNDAMENTALS OF HIGH-FIELD ASYMMETRIC WAVEFORM ION MOBILITY SPECTROMETRY (FAIMS) FOR THE ANALYS IS OF EXPLOSIVES.......72 Introduction................................................................................................................... ..........72 Experimental................................................................................................................... ........74 Results and Discussion......................................................................................................... ..76 Effects of CV Scan Rate..................................................................................................76 Effect of Curtain Gas Flow Rate.....................................................................................77 Effects of DV.................................................................................................................. .78 CV value...................................................................................................................79 Signal intensity.........................................................................................................80 Peak width................................................................................................................81 Effects of Carrier Gas Composition................................................................................82 TNT in different carrier gas compositions...............................................................84 TNT in O2 and mixture of N2/O2..............................................................................84 Explosives in mixture of N2/He...............................................................................85 Effects of Electrode Temperature....................................................................................88 Conclusions.................................................................................................................... .........91 4 PERFORMANCE OF APCI-FAIMS-MS FOR ANALYSIS OF EXPLOSIVES...............113 Introduction................................................................................................................... ........113 Experimental................................................................................................................... ......113 Results and Discussion.........................................................................................................114 Repeatability of CV Values...........................................................................................114 Separation..................................................................................................................... .116 Resolving power.....................................................................................................117 Separation and resolution betw een isomeric explosives........................................117 Separation and resolution of explosive mixtures...................................................119 Quantitation...................................................................................................................120 Reproducibility.......................................................................................................120 Limit of detection and linear dynamic range.........................................................121 Conclusion..................................................................................................................... .......124 5 CONCLUSIONS AND FUTURE WORK...........................................................................140 Conclusions.................................................................................................................... .......140 Future Work.................................................................................................................... ......143 Ionization Source...........................................................................................................144 FAIMS.......................................................................................................................... .144 Mass Spectrometer........................................................................................................145 LIST OF REFERENCES.............................................................................................................146 BIOGRAPHICAL SKETCH.......................................................................................................153

PAGE 8

8 LIST OF TABLES Table page 2-1 Gas-phase acidity valu es for reactant ions..............................................................................61 2-2 Mass spectral data of nitroaroma tic compounds analyzed by APCI-MS................................62 2-3 Mass spectral data of nitroaromatic co mpounds analyzed by DPIS-MS in the closed configuration.................................................................................................................. ....63 2-4 Mass spectral data of nitroaromatic compounds analyzed by DPIS-MS in the open configuration.................................................................................................................. ....64 2-5 Mass spectral data of nitr amines analyzed by APCI-MS........................................................65 2-6 Mass spectral data of nitramines analyzed by DPIS-MS in the closed configuration.............66 2-7 Mass spectral data of nitramines analy zed by DPIS-MS in the open configuration...............67 2-8 Mass spectral data of nitrat e esters analyzed by APCI-MS.....................................................68 2-9 Mass spectral data of nitrate esters analy zed by DPIS-MS in the closed configuration.........69 2-10 Mass spectral data of nitrate esters anal yzed by DPIS-MS in the open configuration..........70 2-11 The main ions of explosive com pounds determined by APCI and DPIS..............................71 3-1 The main analytical characteristic s of FAIMS on detecting explosives.................................99 4-1 Repeatability of CV values from five replicate analyzes of explosive compounds..............126 4-2 Resolving power for explosive compounds...........................................................................127 4-3 Resolution between TNT, TNB and DNT isomers...............................................................128 4-4 Resolution of explosive mixtures..........................................................................................132 4-5 Reproducibility of peak areas from five replicate analyzes of explosive compounds..........136 4-6 Linear dynamic range and limits of detecti on for the nitroaromatic explosives collected by full scan and SIM mode..............................................................................................137 4-7 Linear dynamic range and limits of detecti on at the optimum CV fo r transmission of the nitro aromatic explosives fo r varied collection time........................................................137

PAGE 9

9 LIST OF FIGURES Figure page 1-1 Structures of the explosiv es studied in this work....................................................................31 1-2 Hypothetical plots of the dependence of ion mobility on electric field strength for three types of ions.................................................................................................................. .....32 1-3 Ion motion between two parallel plates dur ing the application of an electric field. A simplified asymmetric waveform is applied to the upper plate.........................................32 1-4 Polarities of CV and DV co mbinations required to transm it specific type of ions.................33 1-5 LCQ quadrupole ion trap showing ion trajectory....................................................................34 1-6 Ion motion in a quadrupole ion trap mass spect rometer. For an ideal quadrupole ion trap (r0 2 = 2z0 2) the potential will be purely quadrupolar..........................................................35 1-7 Mathieu stability diagram for an ion trap for the regions of simultaneous stability in both the rand z-directions. The line z=1 intersects the qz axis at 0.908, corresponding to the low mass cut-off (LMCO) of an ion that can be stored in the trap..............................35 1-8 Schematic of the Thermo LCQ ion trap used in these experiments........................................36 1-9 Thermo LCQ APCI source..................................................................................................... .37 1-10 The configuration of distri buted plasma ionization source...................................................37 1-11 Schematic of APCI source, FAIMS cell and heated capillary interface to mass spectrometer. (not to scale)................................................................................................38 2-1 Configuration of DPIS. (A) Schematic, (B) Actual picture....................................................56 2-2 Comparison of reactant ions generate d by DPIS observed with air, methanol, methanol/water, and 10 ppm TNT in (A) ne gative mode and (B) positive mode.............57 2-3 Schematic procedure of reactant ions formation by DPIS......................................................58 2-4 Three different configurations where the DPIS was placed: closed, open and fully open configuration.................................................................................................................. ....58 2-5 Mass spectra of negative ions generated in air by DPIS with fully open, open and closed configuration.................................................................................................................. ....59 2-6 Mass spectra of negative ions generated in air by APCI with fully open, open and closed configuration.................................................................................................................. ....60

PAGE 10

102-7 Comparison of the reactant ions intensity as a functi on of the composition between oxygen and nitrogen...........................................................................................................61 2-8 Negative APCI mass spectra of nitroa romatic compounds: (A) TNT (MW = 227), (B) TNB (MW = 213), (C) Tetryl (MW = 287) (D) 2,4-DNT (MW = 182), (E) 2,6-DNT (MW = 182), (F) 3,4-DNT (MW = 182), (G) 1,3-DNB (MW = 168)...............................62 2-9 Negative DPIS mass spectra of nitroaroma tic compounds in the closed configuration: (A) TNT (MW = 227), (B) TNB (MW = 213), (C) Tetryl (MW = 287), (D) 2,4-DNT (MW = 182), (E) 2,6-DNT (MW = 182), (F) 3,4-DNT (MW = 182), (G) 1,3-DNB (MW = 168)..................................................................................................................... ..63 2-10 Negative DPIS mass spectra of nitroaroma tic compounds in the open configuration: (A) TNT (MW = 227), (B) TNB (MW = 213), (C) Tetryl (MW = 287), (D) 2,4-DNT (MW = 182), (E) 2,6-DNT (MW = 182), (F) 3,4-DNT (MW = 182), (G) 1,3-DNB (MW = 168)..................................................................................................................... ..64 2-11 Negative APCI mass spectra of nitram ines: (A) RDX (MW = 222), (B) RDX + Cl, (C) HMX (MW = 296), (D) HMX + Cl...................................................................................65 2-12 Negative DPIS mass spectra of nitramines in the closed configuration: (A) RDX (MW = 222), (B) RDX + Cl, (C) HMX (MW = 296), (D) HMX + Cl.......................................66 2-13 Negative DPIS mass spectra of nitramines in the open configuration: (A) RDX (MW = 222), (B) RDX + Cl, (C) HMX (MW = 296), (D) HMX + Cl...........................................67 2-14 Negative APCI mass spectra of nitrat e esters: (A) NG (MW = 227), (B) NG + Cl, (C) PETN (MW = 316), (D) PETN + Cl..................................................................................68 2-15 Negative DPIS mass spectra of nitrate esters in the clos ed configuration: (A) NG (MW = 227), (B) NG + Cl, (C) PETN (MW = 316), (D) PETN + Cl.........................................69 2-16 Negative DPIS mass spectra of nitrate esters in the open configuration: (A) NG (MW = 227), (B) NG + Cl, (C) PETN (MW = 316), (D) PETN + Cl............................................70 3-1 The design of the br ass capillary extender..............................................................................93 3-2 The actual picture of the brass capillary extender...................................................................93 3-3 Effect of CV scan rate on CV valu e, peak intensity, and peak width. (DV= 4000V)...........94 3-4 Effect of curtain gas fl ow rate on CV for the ions of tested explosives. (DV= 4000V).......94 3-5 Effect of curtain gas flow rate on peak width for the i ons of tested explosives......................95 3-6 Effect of curtain gas flow rate on signal intensity for the ions of tested explosives...............95 3-7 SI-CV spectra for the [M]ion ( m/z 227) of TNT: variation of the DV..................................96

PAGE 11

113-8 Mass spectra for the TNT: variation of the DV.......................................................................96 3-9 Graph of CV versus DV for th e ions of tested explosives.......................................................97 3-10 Graph of signal intensity versus DV for the ions of tested explosives..................................97 3-11 Graph of peak width versus DV fo r the ions of tested explosives........................................98 3-12 Mass spectra of explosives acquired by APCI-MS and APCI-FAIMS-MS: (A) TNT, (B) TNB, (C) 2,4-DNT, (D) 2,6-DNT, (E) 3,4DNT, (F) Tetryl, (G) RDX, (H) HMX, (I) PETN, (J) NG..............................................................................................................100 3-13 TIC-CV spectra for TNT in different carrier gas composition at DV of -4000V...............101 3-14 SI-CV spectra for the [M-H]ion of TNT ( m/z 226) in oxygen carrier gas at DV from 2500 to 4500 V in 500 V increments........................................................................102 3-15 Graph of (A) CV, (B) peak width, and (C ) signal intensity versus DV for the [M-H]ion of TNT in N2/O2 mixtures from 0% to 50% O2.........................................................103 3-16 Graph of CV versus carrier gas compositi on for the ions of tested explosives in N2/He mixtures....................................................................................................................... .....104 3-17 Graphs of signal intensity versus carri er gas composition for the ions of tested explosives in N2/He mixtures...........................................................................................104 3-18 Graphs of peak width versus carrier gas co mposition for the ions of tested explosives in N2/He mixtures.................................................................................................................105 3-19 Graph of (A) CV, (B) peak width, and (C ) signal intensity versus DV for the [M]ions of TNT in N2/He mixture. Red circle shows that TNT presents an even stronger type C ion behavior in high helium content.............................................................................106 3-20 Graph of (A) CV, (B) peak width, and (C ) signal intensity versus DV for the [M-NO2]ions of tetryl in N2/He mixture. Red circle shows th at Tetryl presents an even stronger type C ion behavior in high helium content.......................................................107 3-21 Calculated electric field as a function of radial distance between cylindrical FAIMS inner/outer cylinders at di fferent temperature and DV....................................................108 3-22 Graph of (A) CV, (B) peak width, and (C) signal intensity versus cell temperature for the [M]ions of TNT and 2,6-DNT.................................................................................109 3-23 Calculated electric field as a function of radial distance between cylindrical FAIMS inner/outer cyli nders at DV of 4500 V. (I: inner electrod e temperature, O: outer electrode temperature, Pl anar: planar FAIMS cell).........................................................110

PAGE 12

123-24 Graph of (A) CV, (B) peak width, and (C ) signal intensity versus inner and outer electrode temperatures ( ) for the [M]ions of TNT.....................................................111 3-25 Graph of (A) CV, (B) peak width, and (C ) signal intensity versus inner and outer electrode temperatures ( ) for the [M]ions of 2,6-DNT..............................................112 4-1 Structures of the isomeric expl osives studied in this research..............................................128 4-2 CV spectra of a solution mixture of 2,4-DNT, 2,6-DNT, and 3,4-DNT at DV of -5000 V and in the carrier gas of 20:80 helium/nitrogen...............................................................129 4-3 CV spectra of a solution mixture of TN T, 2,6-DNT, and 3,4-DNT at DV of -5000 V and in (A) the nitrogen carrie r gas, (B) the carrier ga s of 20:80 helium/nitrogen..................130 4-4 CV spectra of a solution mixture of TNB, 2,4-DNT, and 2,6-DNT at DV of 5000 V and in the carrier gas of 10:90 helium/nitrogen......................................................................131 4-5 CV spectra of a solution mixture of TN T, RDX, and HMX at DV of 4500 V with the carrier gas of 30:70 helium/nitrogen................................................................................133 4-6 CV spectra of a solution mixture of TN T, NG, and PETN at DV of 4500 V with the nitrogen carrier gas..........................................................................................................134 4-7 IS-CV spectrum of nitroaro matic explosives at DV of 4500 V and in the nitrogen carrier gas.................................................................................................................... .....135 4-8 Mass spectra for analytes containing 50 ng/mL explosives collected by APCI-MS and APCI-FAIMS-MS ranging from m/z 50 to 500: (A) TNT, (B) TNB, (C) 2,4-DNT, (D) 2,6-DNT, (E) 3,4-DNT ,(F) 1,3-DNB, (G) Tetryl.....................................................138 4-9 Mass spectra for analytes containing 10 ng/mL explosives collected by APCI-MS and APCI-FAIMS-MS ranging from m/z 150 to 300: (A) TNT, (B) TNB, (C) 2,4-DNT, (D) 2,6-DNT, (E) 3,4-DNT ,(F) 1,3-DNB, (G) Tetryl.....................................................139

PAGE 13

13 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy FUNDAMENTALS AND APPLICATIONS OF HIGH-FIELD ASYMMETRIC WAVEFORM ION MOBILITY SPECTROMETRY FOR THE ANALYSIS OF EXPLOSIVES By Alex Ching-Hong Wu August 2009 Chair: Richard A. Yost Major: Chemistry Over the past ten years, the world has been stunned and outraged by a series of attacks on civilian targets that us ed explosive devices. These attack s led to widespread demands for identification of the perpetrators, along with calls for improved secu rity measures to prevent such incidents in the future. Detection techniques such as X-ray scanners, Raman spectroscopy, Terahertz spectroscopy and ion mobility spectrometry are currently in use or under development; however, none of these techniques are appropriate for all necessary applications. High-field asymmetric-waveform ion mobility spectrometry (F AIMS) coupled to a mass spectrometer is an alternative technique that provides impr ovements to mass spectral signal-to-noise, orthogonal/complementary ion separation to mass spectrometry, enhanced ion and complexation structural analysis, and potential for rapid analyte quantitation. The primary goal of this research is to cont ribute to the understanding of ionization by an atmospheric pressure ionization (API) source and ion behavior in a FAIMS cell to assist the future development of a portable explosive detector to investigate explosives in the field. In this work, the ionization mechanism of two API sour ces, atmospheric pressure chemical ionization (APCI) and distributed plasma ionization source (D PIS), are discussed. The spectra of eleven explosives ionized by both source s were collected and characterized. The results show that APCI

PAGE 14

14 provides a consistent and simple ionization, whil e DPIS presents more discrimination by various ion fragments and is more amenable for monitoring a certain classes of expl osives in the field. Variation in FAIMS parameters, such as disper sion voltages (DV), compensation voltage (CV) scan rate, curtain gas flow rate, carrier gas co mposition, and electrode temp erature, was explored for their effect on explosive ions. A systema tic evaluation of the performance of API-FAIMSMS demonstrates sensitivity at the picogram level, short detection time (30 seconds), and excellent resolution such that isomers of the same explosive can be successfully resolved. The results from these studies with the laboratory proce dure show promise for FAIMS to be used as an explosive detector that presents a sensitive, selective, specific and rapid technique.

PAGE 15

15 CHAPTER 1 INTRODUCTION AND INSTRUMENTATION Background With ongoing worldwide terrorist activity, explos ives analysis is becoming an increasingly critical issue of public securit y. Despite long-term research a nd development into explosives analytical techniques, the demand still remains for the development of an explosives detector with properties that include higher sensitivity, sel ectivity, specificity, near-real time analysis and portability. In recent years, a wide variety of te chniques including gas chromatography-electron capture detection (GC-ECD),1-3 gas chromatography-mass spectrometry (GC-MS),4-8 liquid chromatography-mass spectrometry (LC-MS),9, 10 high-performance liquid chromatographytandem mass spectrometry (HPLC-MS/MS), 11, 12 and high-performance liquid chromatographyultraviolet detection (HPLC-UV)13, 14 have been developed and app lied to detect explosives under various conditions.15 These techniques are not, however, void of certain difficulties. Analysis of explosives by GC can be problematic because of their low vapor pressure and thermal instability. Due to its highly polar nature, cyclotrimet hylene trinitramine (RDX) ofte n gives poor peak shapes, and cyclotetramethylene tetranitrami ne (HMX) is often difficult to chromatograph. Further, while ECD is selective for explosives containing n itro groups, it does not conclusively identify separated analytes. HPLC overcomes some of the difficulties associated with the high temperatures required for GC analysis, but su ffers from poorer resolution. Furthermore, UV detection gives little structural information.16 Mass spectrometry is a promising method and has been widely used when coupled with different chromatographic appro aches for explosives analysis. This promise is owed to the

PAGE 16

16 advantages of high sensitivity, fast response, and the additio nal selectivity available from MS/MS and ion/molecule reactions.17 However, the need for sample preparation and dependence on a separation device substantially re stricts the capabilities of mass spectrometry for explosive detection. Therefore, in order to overcome the limitations of MS, the ionization sources recently introduced for explosive research, including electrospray ionization (ESI)10, 18, 19 and atmospheric pressure chemical ionization (APCI),20, 21 have emphasized operation at atmospheric pressure. More and more attention ha s been invested to develop ionization sources with the capacity for the direct ionization of explosives on solid surfaces, such as atmospheric pressure matrix-assisted lase r desorption/ioni zation (MALDI),22 thermal desorption mass spectrometry,23 and secondary ion mass spectrometry (SIMS),24 and to build up an ionization source which can be operated under ambient conditi ons, such as direct an alysis in real time (DART)25, 26 and desorption electrospray ionization (DESI).27, 28 A distributed plasma ionization source (DPIS) was invented according to th e same demand, but consumes less power.29 Without the electrosprayed solvent such as DESI, and co mplex configuration such as DART, the DPIS provides portable characteristics, such as sma ll size and ease of operation, and could potentially be coupled to portabl e mass spectrometers.17 Currently, ion mobility spectrometry (IMS) is used at over 10,000 airports worldwide for screening handcarried articles.30 IMS is an alternative explosiv es separation device that offers the benefit of the favorable gas-phase ionization chemistry for explosives at ambient pressure and the satisfactory selectivity obtained by mobility analysis.31 IMS provides practical advantages of fast, highly sensitive and speci fic detection, instrume ntation simplicity and comparatively low costs of operation.32 However, selectivity of IMS is not enough, resulting in false alarms in some cases.33 Therefore, based on similar principles to IMS, high-field

PAGE 17

17 asymmetric waveform ion mobility spectrometry (FAIMS) was developed as a new technique for atmospheric pressure and room temper ature separation of gas-phase ions.34, 35 With FAIMS, ions are separated based on differences between their mobility in weak and strong electric fields.36 Organic explosives belong to various chemical classes, including nitr ate esters, nitramines and nitroaromatics, and have very different physi cal properties, which make their analysis by a single method difficult. In addition, various amounts of byproducts can be found in an explosive, depending on the way it was manuf actured. Isomers of nitroaromatic compounds, which are typically by-products, ca nnot be easily analyzed on cl assical chromatographic methods because of their similar behavi or on typical stationary phases.37 Current technology attempts to combine several analytical methods to achieve higher selectivity and se nsitivity for explosive detection.21 FAIMS and MS are two se parate orthogonal detection methods, which separate ions depending on their differential ion mobility and ratio of mass-to-charge. FAIMS-MS can be expected to effectively separa te isomers or isobaric compounds38 and provide specific and deterministic detection of the full range of military, commercial, and improvised explosive compounds by matching unknown sample vapors to a known library of mobility and/or mass spectra signatures. Both FAIMS and MS can be downsized to a hand-held, concealed portable detector and used as a field-deployable device for the detection of explosives.39 FAIMS is a valuable separation technique that e xhibits low parts per bi llion (ppb) limits of detection for continuous vapor streams and is su itable as a gas chromatographic detector owing to its fast response and low memory40, 41. It also possesses capabilities of separating isomers and providing additional information orthogonal to MS.38, 42, 43 The interfacing of FAIMS with MS offers potential advantages over the use of mass spectrometry alone or with other chromatographic methods. Such advantages in clude improvements to mass spectral signal-to-

PAGE 18

18 noise, orthogonal/complementary ion separati on to mass spectrometry, enhanced ion and complexation structural analysis, and the potential for rapid analyte quantitation.44 Gaining understanding of how this technique functions will benefit the deve lopment of a highly sensitive, accurate, rapid and on-scene explosives detector, which may be potentially used for the fast detection of bombs or explosives in the anti terrorism field. This system could also be implemented in other areas, such as pharm aceutical analysis, forensic investigation, environmental conservation, and food monitoring. The eleven explosives (Figure 1-1) investigated in this re search are the compounds most widely used for military or terrorist attac k, and can be divided into three categories: nitroaromatic, nitramine, and nitrate esters. The nitroaromatic compounds include 2,4,6trinitrotoluene (TNT), MW=227.13; 1,3,5trinitrobenzene (TNB), MW=213.1; N-Methyl -N,2,4,6-tetranitroaniline (tetryl), MW=287.14; 1,3-Dinitrobenzene (1,3-DNB), MW=168.11; 2,4 -dinitrotoluene (2,4-DNT), MW=182.13; 2,6dinitrotoluene (2,6-DNT), MW =182.13; and 3,4-dinitrotol uene (3,4-DNT), MW=182.13. Among them, TNT is one of the most commonly us ed explosives for military and industrial applications. Tetryl is a sensitive explosive compound used to make detonators and explosive booster charges. TNB and 1,3-DNB are form ed through photodecomposition of TNT from sunlight and are readily detected in explosive contaminated water; and DNTs are the major byproducts from the TNT ma nufacturing process. The nitramine compounds include 1,3,5triazinehexahydro-1,3,5-trinitro (RDX), MW=222.12; and 1,3,5,7-tetranitro-1,3,5,7-tetraazacyclooctane (HMX), MW=296.16. RDX is typically used as a component in mixtures with other explosives such as TNT and as a plastic explosive. HMX is the main byproduct of RDX a nd used exclusively for military application.

PAGE 19

19 The nitrate ester compounds include pentaery thritol tetranitrate (PETN), MW=316.14; and 1,2,3-propanetriol trinitrate (nitr oglycerin, NG), MW=227.09. PETN is primarily used in booster and bursting charges of small caliber ammunition, in upper charges of detonators in some land mines and shells, and as the explosive core of detonation cords. NG, which is widely used in industrial explosives, has been the main component in many dynamites. High-Field Asymmetric Waveform Ion Mobility Spectroscopy High-field asymmetric waveform ion mob ility spectrometry (FAIMS), also commonly referred to as differential mobility spectrometry (DMS), is a new technique used for atmospheric pressure, room temperature separation of gas-phase ions.45 With FAIMS, ions are separated based on differences between the mobilities of the ion in the presence of weak (K0) and strong (Kh) electric fields.36 At low electric fields (e.g., 200 V/cm), the ion drift velocity is proportional to the field strength, and the mobility (K0) which is independent of the applied field is constant. However, at high electric fields (e.g., 10,000 V/ cm) the ion velocity is no longer directly proportional to the applied field, and the mobility (Kh) is dependent upon the applied electric field. The mobility of a given ion under the influe nce of an electric field can be expressed by Kh/K0 = 1 + (E/N)2 + (E/N)4 (1-1) where K0 is the coefficient of ion mob ility at zero electric field and describe the dependence of the ion mobility at a high el ectric field in a particular dr ift gas, and N is the gas number density.46 There are three possible behaviors of changes in ion mobility with elec tric field strength, as illustrated in Figure 1-2: type A behavior, or exponential increas e in mobility proportional to change in field strength, type B behavior, or exponential in crease in mobility followed by exponential decay as field strength increases, a nd type C behavior, which is exponential decay

PAGE 20

20 in mobility as field strength increases. The change in mobility affects the direction the ion travels toward or away from the plates of the FAIMS cell. The change in mobility at high field appears to reflect the size of th e ion, its interaction with the bath gas, and the st ructural rigidity of the ion.30 These designations are not absolute but depend on the buffer gas; ions often shif t toward type C with increasing mass.47 Cations and anions exhibit similar trends, and (in N2 or air) the transition from A to C occurs over the ~100350 Da range. Type B ions are found in that transition region.48 More recent experimental and theoretical work has demonstrated that these differences in ion behavior can be ascribed to interactions of the ion struct ure, collision cross-section and instrumental parameters.49 Most of the explosives studied in this research act as type A or B ions. Figure 1-3 illustrates the ion motion in a FAIMS cell for a positive type A ion. Ions are transmitted past the el ectrode surfaces by a carrier gas that flows between the two parallel electrodes shown in Figure1-3. The waveform, which consists of a high voltage component, Vhigh, lasting for a shorter period of time thigh, and a low voltage component of opposite polarity, Vlow, lasting for a longer period of time, tlow, is applied to the upper plate to produce the required electric fiel d. This waveform is synthesi zed such that the integrated voltage-time being applied to the upper electrode during one complete cycle of the waveform is zero (equation 1-2).50 Vlow tlow + Vhigh thigh = 0 (1-2) If the ion mobility is the same at high and lo w electric fields, the ion will experience zero net displacement towards an electrode and will be transmitted through the FAIMS device. However, the mobilities of most ions depend on electric field strength over the range used in these experiments, resulting in a net displacement of the ions towa rds one of the electrodes. To

PAGE 21

21 select which ions should maintain a trajectory through the electrodes w ithout striking, a direct current (DC) potential, referred to as a compensati on voltage (CV), is applie d to the upper plate. The CV scans are generated by scanning thr ough a range of comp ensation voltages and measuring the ion abundance transmitted through the FAIMS device as a function of compensation voltage.51 The compensation voltage require d depends on the ions ratio of Kh/K0, dispersion voltage (DV), the temperature, the pressure, and the gas flow rates.52 The FAIMS cell used in this research has a cy lindrical geometry, in which an ion focusing region is generated in the annular space between the two concentric cylinders due to the nonuniform electric field in the cell.53 In the FAIMS cell, both the magnitude and the polarity (positive or negative) of the DV have an effect on the CV. The FAIMS instrument works with both positively and negatively charged ions in one of four modes: P1, P2, N1 and N2. The letter portion of the mode indicates the polarity of the ions. The number portion refers to FAIMS instrument conditions. Briefly, P1 and N1 mean that positively and negatively charged ions are optimally separated by a DV with same polarity; P2 and N2 imply positively and negatively charged ions are best separa ted by a DV with opposite polarit y. For example, CV spectra collected for a positive type A ion in P1 mode display the following tendencies with increasingly positive DV: the peak shifts to more negative CV values, the response increases substantially, and the peak widens.52 The waveform (P1 and P2 modes for cations and N1 and N2 modes for anions) in cylindrical FAIMS cell focuses either type A or C ions, but de focuses and eliminates the other type of ions from the gap.50 Ions of type B are normally focused as type C ions; though CV has the same sign as for type A ions. Thus DV and CV have opposite signs for type A (analyzed in the P1 and N1 modes) and same si gns for type C (in P2 a nd N2) ions as Figure 14.54

PAGE 22

22 The fundamentally optimum waveform profile is rectangular, in the sense that the equations relating Kh/K0 to the measured signals are very simple. However, such a waveform requires excessive electronic power consumpti on for typical electrode sizes. Most FAIMS systems, therefore, employ a more practical bisinusoidal waveform. These waveforms are described mathematically by the equation 1-3 VD(t) = f +1 [ f sin(2 wct )+sin(4 wct)] Vmax 2 (1-3) where wc is the frequency, Vmax (dispersion voltage) is the peak amplitude, and f controls the waveform profile. Most FAIMS systems have adopted the optimum f = 2.55 Note that research using the rectangular waveform is underway in our laboratory.56 The advantages of FAIMS as a gas-phase, ionprocessing and separatio n tool include: (1) high sensitivity provided by an ion focusing mech anism; (2) the ability to separate ions at atmospheric pressure and room te mperature; (3) the ability to se parate ions on a continuous basis rather than in discrete pulses; and (4) simplicity in interfacing to a mass spectrometer.47 Quadrupole Ion Trap Mass Spectrometry Quadrupole Ion Trap Theory The quadrupole ion trap mass spectrometer (Q ITMS) was initially described by Paul and Steinwedel in a patent filed in 1953 in Germ any and was awarded a U.S. patent in 1960.57 The quadrupole ion trap is a three-elec trode device that consists of a hyperbolic ring electrode placed between two endcap electrodes (Figure 1-5). Gene rally, the endcap electrodes are held at ground potential and radio frequency (R F) and DC potentials are applie d on the ring electrode. The confinement of the ions within the trap is illustrated by the quadrupolar potential that is represented in Figure 1-6. As can be seen from this figure, ions in the central part of the trap are

PAGE 23

23 confined in the axial z -direction; however, in the radial r -direction ions are accelerated towards the end caps and are not confined. Simultaneous confinement (trapping) of the ion in both directions can be obtained by cha nging the polarity of the field ever y time the ion is approaching the electrodes. The magnitude of the trappi ng potential is described by the equation 1-4. Dz= z/mV/(4z0 22) (1-4) where Dz is the depth of the quadrupolar potential, z/m is the inverse of the mass to charge ratio of the ion, V is the amplit ude of the RF voltage, z0 is the distance from the center to an end cap, and is the frequency of the RF voltage. De pending on the value of the fundamental RF voltage, ions of different m / z are trapped inside the ion trap.58 For an ideal quadrupolar field, the following identity is given as equation 1-5 r0 2 = 2z0 2 (1-5) so that once the magnitude of r0 is given the sizes of all th ree electrodes a nd the electrode spacings are fixed. However, it has been pointed out by Knight59 that the ratio of r0 2 to z0 2 is not necessarily restricted to 2. Re gardless of the value of this ra tio, the size of the ion trap is determined largely by the magnitude of r0, and most commercial ion traps, r0 is either 1.00 or 0.707cm.60 The gaseous ions, positively or negatively char ged, can be stored or confined inside the trapping potential well when appropr iate potentials are applied to the electrodes of the ion trap.61 The ions with varied mass to charge ratios can be measured by changing the electric field within the device when their trajectories become seque ntially unstable. A stable ion will possess a trajectory that allows the ion to be trapped, or contained, within th e specific electric field of the trap; however, an unstable ion will have a trajec tory that increases in magnitude toward the

PAGE 24

24 endcap electrode. Ions that are ejected thr ough the exit endcap electr odes are focused by the conversion dynode accelerating potentia l through the exit lens towards the ion detection system and detected. Ion Motion in the Ion Trap The motion of ions inside the trap can be de scribed mathematically by the solutions to the second-order linear differential equati on described originally by Mathieu.62 Solutions to the differential equation are in term s of two reduced parameters, az and qz, which can be used to calculate whether an ion will have a stable or unstable trajectory in the trap under the defined conditions of the electric field. The values of az and qz depend on the dimensions of the trap and the potentials applied accordi ng to equations 1-6 and 1-7: 2 2 22 16 2 o o r zz r m eU a a (1-6) 2 2 22 8 2 o o r zz r m eV q q (1-7) where the subscripts z and r represent axial and radial moti on between and perpendicular to the endcaps, respectively; U is the DC amplitude applied to the ring electrode (if any), V is the RF potential applied to the ring electrode, e is the charge on an ion, m is the mass of an ion, r0 is the inner radius of the ring electrode, z0 is the axial distance from th e center of the device to the nearest point on one of the endcap electrodes, and is the angular drive frequency. Solutions to these equations in the r-and z-dire ctions are solutions of two kinds which either represent stable or unstable trajectories. The set of solutions can be readily repres ented in the form of a Mathieu stability diagram (Figure 1-6). The coordinates of the stability region in Figure 1-6 are the Mathieu parameters az and qz. According to the Mathieu equati on one can generate a stability

PAGE 25

25 diagram that shows the common region in (az, qz) space for which the radial and axial components of the ion trajectory are stable simulta neously, such that the i on can be confined in the trap. Different point in az and qz coordinates correspond to different values of r and z value which relate to the secular frequency of the ion in the z and r directions, respectively (equation 1-8). u = 0.5 u (1-8) When the value of approaches zero, the ions secular frequency approaches zero, and the ion is not contained. When the value of equals one, the ions secula r frequency equals half the frequency of the RF field and the magnitude of it s oscillation increases so that the ion escapes the trap or collides with one of the endcap surfaces. However, if has a value between zero and unity, the ion can be trapped by the oscillating fields and will oscill ate in a periodic mode at its secular frequencies in z and r direction.63 In Figure 1-7, the position of different ions is depicted in the stability diagram for different RF amp litudes. When the fundamental RF voltage is linearly increased, the ions move toward the boundary of the stability region (qz = 0.908, az = 0). When ions of increasingly m/z reach the qz = 0.908 point, they become unstable in the axial direction and are ejected from the trap. The si mplest way to extend the mass range is to cause the ion to become unstable at a value of qz lower than 0.908. This is achieved by applying an auxiliary RF field across the endcaps with a freq uency matching the oscillation frequency of an ion of particular m/z in the axial direction (qz) while ramping the fundamental RF.64 Finnigan LCQ The mass spectrometer used in this research is a commercial be nchtop Thermo LCQ ion trap, which is a three dimensi onal quadrupole ion trap based inst rument designed for use with

PAGE 26

26 external atmospheric pressure ionization (AP I) sources. Atmospheric pressure chemical ionization (APCI) and electrospray ionization (ESI) source are the two major API sources used with the LCQ. A distributed plasma ionization s ource (DPIS) was also used in this research. The system is easily operated in either positiv e or negative ion mode. In addition, the mass range of this instrument is m/z 150 to 2000 but can be extended to m/z 4000 for some applications, as noted in the previous secti on. The LCQ has a maximum resolution of 10,000 in the zoom scan mode, and 4000 in full scan mode. For APCI and DPIS operation, the sample soluti on is infused by a syringe pump at a flow rate of 20 L/min, and vaporized by the standard LCQ AP CI heated nebulizer. Ions are formed by APCI or DPIS and the ions are guided thr ough heated capillary, wh ich helps desolvate the ions. The ions then pass through a series of le nses, skimmers, and octopole ion guides before making it into the ion trap mass analyzer. After the ions exit the heated capillary, the ions are then gated by the tube lens and passed through a skimmer cone into th e first and second RF octopoles. The skimmer acts as a vacuum baffle between the higher and lower pressure regions. The octopoles act as an ion guides and transm it ions efficiently through the region by focusing the ions into a beam. Figure 1-8 exhibits a schematic of the Thermo LCQ ion trap mass spectrometer used in these experiments. After the ions enter the ion trap mass analyz er through the entrance endcap electrode, they collide with helium buffer gas atoms and are slowed down and maintained near the center of the trap. As described previously, an RF voltage is applied to the hyperbolic ring electrode. The hyperbolic endcap electrodes are held at or near ground. The application of the RF voltage to the ring electrode produces a 3-D quadr upolar field within the mass an alyzer, trapping ions in their stable trajectories. As the ring electrode RF voltage increases, the system produces a mass-

PAGE 27

27 dependent instability to eject ions from the mass analyzer in the axial direction. Once negative ions (as studied here) are ejec ted from the mass analyzer, they are attracted to a conversion dynode, held at +15 kV. Positive ions ejected from the conversion dynode are accelerated into the electron multiplier (held at ~ kV) and then amplified for signal de tection. Data were processed using the instrument software (Xcalibur version 1.3).65 Atmospheric Pressure Ionization Atmospheric Pressure Chemic al Ionization (APCI) Atmospheric pressure chemical ionization (APC I) is a soft gas-phase ionization technique that works at atmospheric pressure. APCI is sim ilar to CI in the type of ionizing reactions that occur, except that it is accomplished in an ioni zation chamber at atmospheric pressure instead of a low pressure environment (~ 1 To rr). In the APCI source, ioni zation is initiated either by lowenergy electrons from a radioactive -emitting or, in our case, by a corona discharge. The APCI technique is mainly applied to polar compounds with moderate molecular weight up to about 1500 Da and generally gives monocharged ions.66 This method was selected because it most closely resembles an ionization source that is am enable in field instruments. Many of the other ionization techniques either requir e vacuum or are large, comple x systems containing lasers or high-voltage; these techniques are less suita ble for a man-portable field instrument.67 The APCI ionization source used for this research (Figur e 1-9) is designed for interfacing to liquid chromatography; therefore, it introduces liquid sa mples. The liquid sample is nebulized by the APCI nozzle into a fine mist of droplets, which are passed pneu matically via nitrogen sheath gas into a heated region where they are vaporized. The nitrogen sheath gas and vaporized solvent molecules then serve as reactant gas as the va por is passed into a corona discharge which produces reactant ions that ionize the anal yte through a series of chemical reactions.63 Analyte ionization can be achieved by primary CI process, with the formation of gas-phase buffer ions,

PAGE 28

28 analyte molecules and solvent molecules, and sec ondary processes, in wh ich electrons from the corona discharge ionize nitrogen or other gases in the APCI source, leading to the eventual CI of the analyte. These ion-molecule reactions incl ude proton transfer, charge exchange, electrophilic (positive ions) or nucleo philic (negative ions) addition, and an ion abstraction (positive ions) or nucleophilic displacement (neg ative ions). Most reactant ions are capable of participating in more than one of the listed reactions. The high io nization efficiency of APCI is due to the short mean free path at 760 Torr and thus the increased number of collisions between the sample molecules and reactant ions. Distribution Plasma Ionization Source (DPIS) A distributed plasma ionization source (DPIS) c onsists of a dielectric between a relatively large electrode and a small electr ode (Fig 1-10). The small elect rode is exposed to the media where ions are to be created. Applying a timevarying (RF) potential between the electrodes produces a glow discharge or plasma. A DC electric field applied between the DPIS and a counter electrode (in this case, the heated cap illary interface to the mass spectrometer) moves ions of the selected polarity away from the DP IS. Reversing the polarity of the potential across the dielectric inhibits the fo rmation of a corona discharge.29 The DPIS source was invented to inhibit cor ona arc discharges caused by point-to-point corona discharge ionization and to minimize dire ct streamers generated by conventional point-toplane corona discharge ionization so urce. It also minimizes corona point erosion and instability. The positive ions that are produced are similar to those generated by 63Ni, 273Am, or a corona discharge. The negative ions produced are similar to those yi elded by point-to-point corona discharge, except that the reac tion region configuration aids in discriminating between the formation of NO3 -, CO3 -, and O2 ions.29, 68 The DPIS used in this research is a bulb design, in

PAGE 29

29 which a neon bulb was used to generate the pl asma in the bulb as a conductive surface and a mesh electrode enclosing the bulb was used as a small electrode. The DPIS has the potential to replace the poi nt-to-plane corona disc harge source for APCI with the advantages that include design configur ation flexibility, dimensional stability, simplicity and ruggedness of design, and exte nded source lifetime. It also has the potential to become a powerful ionization source to detect explosives in the field. FAIMS/MS Experiments were performed employing a FAIM S/MS system, comprising of a cylindrical FAIMS device (Thermo Scientific, San Jose, CA) and a commercial ion trap mass spectrometer (LCQ, Thermo Scientific). Gas-phase explosiv e ions was generated by APCI using a corona discharge needle that is positioned at an angle of 45 and ~1 cm from the opening in the curtain plate of the FAIMS device or by a DPIS source as described in the previous section. The FAIMS is interfaced to the MS with a 9 cm long brass extender (i.d. = 0.76 mm, o.d. = 22 mm). A schematic of the APCI-FAIMS-MS instrument th at was utilized for this work is shown in Figure 1-11. Overview of Dissertation This dissertation presents a de tailed investigation in to the fundamentals and applications of high-field asymmetric waveform ion mobility spec trometry (FAIMS) to explosives analysis. The ultimate goal of the research is to develop an efficient approach to the analysis of explosives through the implementation of FAIMS as a sepa ration device in conjunc tion with a quadrupole ion trap mass spectrometer (QITMS). Chapter 1 has presented an introduction to the fundamentals of FAIMS and the parameters which may affect the separation of explosive compounds by FAIMS. A brief overview of ion trap mass spectrometry was presented becau se a commercial bench-top ion trap mass

PAGE 30

30 spectrometer was used in all applications. Th e introduction of API sources, APCI and DPIS, was also included in this chapter. Chapter 2 compared the performance of differe nt types of API sources. API sources that can be operated under atmospheric pressure and room temperature would be beneficial to developing an on-scene explosives detection system. In addition to APCI, which has been successfully applied for different kinds of explosives DPIS is examined in this research as well. The gas-phase chemistry of the ionization source s for explosive compounds is investigated and presented in this chapter. In Chapter 3, experimental parameters affec ting the ion transmission of explosives in the cylindrical FAIMS analyzer region are explored. The parameters explor ed in this research include dispersion voltage (DV), compensation voltage (CV) scanni ng rate, curtain gas flow rate, carrier gas composition, and elec trode temperature. The eval uation of eleven explosives analyzed by FAIMS is discussed. In Chapter 4, information gained from chap ter 2 and 3 are utilized in the practical application of APCI-FAIMS-MS for explosive analysis. The motivati on for this experiment is to evaluate the analytical perf ormance of the combination of APCI, FAIMS and MS and to understand the limits of detection for explosives by this method. Systematic evaluation of nitroaromatic, nitrate ester, and nitramine e xplosives using APCI/FAIMS/MS is covered. Chapter 5 discusses conclusion and future wo rk. The advantages and disadvantages of utilizing FAIMS with mass spectrometric techni ques for the analysis of explosives and the optimized procedure are presented. The chapte r summarizes the major co nclusions drawn from this work and offers suggestion for future research in this area.

PAGE 31

31 TNT MW=227.13 PETN MW=316.14 HMX MW=296.16 RDX MW=222.12 TNB MW=213.1 NG MW=227.09C H3N+O-O N+O-O N+O-O N+O-O N+O-O N+O-O N N N N+O-O N+O-O N+O-O N N N NN+O-O N+O-O N+O-O N+O-O O O O O N+O-O N+O-O N+O-O N+O-O O O O N+O-O N+O-O N+O-O 1,3-DNB MW=168.11 Tetryl MW=287.14N+O-O N+O-O N+O-O N N+O-O CH3 N+O-O N+O-O 2,4-DNT MW=182.13 3,4-DNT MW=182.13 2,6-DNT MW=182.13CH3N+O-O N+O-O CH3N+O-O N+O-O CH3N+O-ON+O-O Figure 1-1. Structures of the expl osives studied in this work.

PAGE 32

32 A C B Increasing Electric Field StrengthRatio, Kh/K1.0 1.05 0.95 Figure 1-2. Hypothetical plots of the dependence of ion mobility on electric field strength for three types of ions. 52 thightlow DV +4000 V 2000 V 0 V CV Figure 1-3. Ion motion between two parallel plates during the appl ication of an electric field. A simplified asymmetric waveform is applied to the upper plate.

PAGE 33

33 +CV -DV -CV +DV P2P1 N2 N1 Type AType C Type CType A Figure 1-4. Polarities of CV and DV combinations required to transmit specific type of ions.

PAGE 34

34 Figure 1-5. LCQ quadrupole ion tr ap showing ion trajectory.65 r0 z0

PAGE 35

35 Figure 1-6. Ion motion in a quadrupole ion trap ma ss spectrometer. For an ideal quadrupole ion trap (r0 2 = 2z0 2) the potential will be purely quadrupolar.58 Figure 1-7. Mathieu stability diagram for an ion tr ap for the regions of simultaneous stability in both the rand z-di rections. The line z=1 intersects the qz axis at 0.908, corresponding to the low mass cu t-off (LMCO) of an ion th at can be stored in the trap.58

PAGE 36

36 Heated Capillary VACUUM ATMOSPHERE Tube Lens Octopoles Skimmer Ion Trap 15 kv Dynode Electron multiplier Interoctopolelens Figure 1-8. Schematic of the Thermo LCQ ion trap used in these experiments.

PAGE 37

37 Figure 1-9. Thermo LCQ APCI source.65 Large electrode Dielectric material Small electrode Figure 1-10. The configuration of di stributed plasma ionization source.

PAGE 38

38 APCI Probe To MS Brass Extender FAIMS Cell Gas Flow Inner Electrode Outer Electrode Heated Capillary Curtain Plate Corona Discharge Needle Figure 1-11. Schematic of APCI source, FAIM S cell and heated capi llary interface to mass spectrometer. (not to scale)

PAGE 39

39 CHAPTER 2 PROPERTIES AND CHARACTERIZATION OF EXPLOSIVES BY ATMOSPHERIC PRESSURE IONIZATION (API)-MASS SPECTROMETRY Introduction The development of highly sensitive technique s capable of trace expl osives detection and straightforward identification is increasingly desirable in the forensic community. These techniques are also needed to perform field analysis of i nvolatile and thermally unstable explosive compounds with rapi d response times, preferably without complicated sample preparation. Mass spectrometry is a very powerful tool for forens ic analysis, because it offers high sensitivity, high selectivity, and a short detection interval.69 A variety of ionization sources have been explored for use with mass spectrome try for explosives dete ction, including electron ionization (EI),5, 8 chemical ionization (CI),11, 70 photoionization,71 desorption electrospray ionization (DESI),72, 73 direct analysis in real time (DART),25 and API. However, each of these ionizations has characteristics which limit thei r use in a portable explosive detector. For example, EI and CI require reduced pressure to maintain a stable ionization, and photoionization requires the use of an extra power supply and a di scharge lamp and provides selective ionization. DESI and DART generate ions under ambient conditions, allowing for direct detection of samples on surfaces; however, they are still imperfect because the source needs neither electrosprayed solvent to form desorbed ions as for DESI, a device with the complex configuration of DART.74 Among these ionization sources, API has ability to directly sample from the atmosphere and the potential for production of molecula r ions/adducts in high abundance.33 Two API methods, atmospheric pressure ch emical ionization (APCI) and distributed plasma ionization source (DPIS), we re evaluated in this research to investigate the ionization

PAGE 40

40 mechanisms for the detection of eleven expl osive compounds. The APCI source has already been developed and used to detect and analyze e xplosives under various conditions because of its user-friendliness, high sensitivity, reliability, and it s widespread availability, all of which enable the detection in the ambient environment.75 APCI uses a corona discharge at atmospheric pressure and is mainly applied to polar com pounds with molecular weights up to about 1500 Da and generally gives singly-charged ions. Recen tly, the DPIS has been developed to meet the requirements of low detection limits high-throughput, and portability.68 The DPIS is a type of direct ionization technique for mass spectro metry that is based on the production of a nonequilibrium plasma. This plasma is generated around one of the electrodes and is fairly easy to use at atmospheric pressure to generate anal yte ions. These API met hods, with their different ionization mechanisms, were selected because they are potentially amenable to field measurements. Both ionization sources were we ll suited for detecting explosives. Thus, the preference in choosing one ionization source over another is determined by availability, sample medium and convenience of use. Experimental Atmospheric Pressure Chemic al Ionization (APCI) Solutions containing explosives are directly in jected via the syringe pump of the LCQ into the vaporizer at a flow rate of 20 L/min with a maximum ion in jection time of 50 ms for automatic gain control (AGC). Th e discharge current was set at 5 A, the vaporization temperature was held at temper atures of 100, 150, 200, 250, or 300 C, and the flow rate of the sheath gas (N2, unless stated otherwise) was set at 20 (Thermo LCQ arbitrary units). The LCQ software was used to tune the instrument as n eeded throughout the study in order to maximize signal intensity. During sample introduction, these parameters we re changed to optimize the ion intensity of the molecular or major ion of the sa mple. This technique pr oduces ions in air at

PAGE 41

41 atmospheric pressure using a cor ona discharge. If another chem ical ionization (CI) reagent gas is not added, the main components in air serve as the primary CI reagent. A series of ions are generated that undergo a variet y of ion molecule reactions. These reactions include ion formation from the trace species of intere st, allowing their dete ction and measurement.76 In this research, TNT, TNB, tetryl, 2,4-DNT, 2,6-DNT, 3,4-DNT DNB, RDX, HMX, PETN, and NG were selected for analysis by APCI based on their structural classes: nitroaromatic, nitramine, and nitrate ester. The explosive solutions, which were originally prepared in acetonitrile at a concentration from 250 to 2000 g/mL, were diluted in a solvent composed of 65% methanol and 35% deionized wate r. All solutions were further diluted to a concentration of about 10 g/mL. Gas-phase explosive ions were generated by APCI. Negative ion mode was generally chosen fo r detecting the molecular ion [M]-or deprotonated molecule [MH]-. However, addition of an or ganic acid or salt is necessa ry to form adduct ions for nitramine and nitrate ester explosives such as RDX, HMX, PETN, and NG because of their lack of acidic protons. In this research, ap proximately 0.1% carbon tetrachloride (CCl4) was used as an additive in some solutions to form stable adducts ions with nitr amine and nitrate ester explosives. Distributed Plasma Ionization Source (DPIS) The DPIS used for this research was provide d by Implant Science, Inc. (Figure 2-1) A neon bulb can be made to glow by applying direct current between the lead s. This glow comes from the plasma that acts as a conductive surface inside the bulb and serves as an electrode on one side of the glass dielectric surface. A mesh electrode is placed around the bulb to complete the ion source. Argon bulb was also evaluated in this research and generated similar spectra to neon bulb. However, the spectra presented in this chapter are all produced by neon bulb. In this research, a RF voltage (1000-1675 Vp-p at 40 kHz) is applied on lead of the DPIS to create the

PAGE 42

42 plasma in the bulb and an offset DC voltage (-30 to -250 V) is applied to the mesh in order to bias the source and select the polar ity of ions to be produced. Ge nerally, no ions can be observed when the RF voltage was applied under 1000 V, a nd the ion intensity increases gradually as the RF voltage is raised. The neon bulb is positioned 3 mm away from inlet of the heated capillary. All other parameters of the mass spectrometer are the same as APCI. Results and Discussion Reactant Ions The major reactant ions produced by the DPIS in negative mode are m/z 62 (NO3 -), m/z 60 (CO3 -), and m/z 46(NO2 -). These ions occur because the ne gative ions produced are similar to those yielded by point-to-point co rona discharge except the confi guration aids in discriminating between the formation of NO3 -, CO3 -, and O2 ions.68(Figure 2-2-A) The major reactant ions produced by the DPIS in positive mode are m/z 37 [2 H2O + H]+, and m/z 55 [3 H2O + H]+. Spectra with methanol and methanol/w ater show major reactant ions as m/z 33 [MeOH + H]+, m/z 47 [2 MeOH H2O + H]+, and m/z 65 [2 MeOH + H]+. (Figure 2-2-B). The initial negative reactant ions formed by DPIS are primarily Oand O2 from oxygen. Among them, O2 is formed via charge capture (reaction 2-1) and Ois formed by ion-molecule reactions.77 O2 + eO2 (2-1) The reactions produce O2 -, O3 -, and O4 depending on the pressure and the energy of the Oand O2 -ion. Since Oand O2 exist in the upper atmosphe re, there is a great inte rest in the interaction of O2 and Owith O2 in binary and three-body reaction.78 The following reaction occurs very qui ckly to generate a mixture of CO3 -, NO3 -, NO2 -, and HCO3 from air around the DPIS source.79 The CO3 ion is produced via two-body reactions of

PAGE 43

43 O3 (reaction 2-2) and three-body reactions of O(reaction 2-3) with CO2.80 Paulson81 also reported that CO3 may form through the interaction between O2 and CO2 (reaction 2-4). O3 +CO2 CO3 + O2 (2-2) O+ CO2 +CO2 CO3 +CO2 (2-3) O2 +CO2 CO3 + O (2-4) Noted nitrogen monoxide (NO) is produced by DPIS and reacts with O3 and O4 to produce NO2 and NO3 (reaction 2-5, 2-6). Takada, et al.69 have reported that NO is produced by the corona discharge, and is able to react with O2 to produce NO3 -. The NO can also interact with CO3 -, and CO4 to yield NO2 and NO3 -. However, according to the research of Ferguson et al.82, less than 2% of the ground state of NO3 is generated from the reaction between NO and CO4 -. The generation of NO2 can be also inferred by the ioni zation of atmospheric gases through electron capture or charge tran sfer (reaction 2-7) mechanisms owing to the positive electron affinity of NO2 (2.27 eV). The NO2 ion can also react with O3 to form NO3 -. However, Ferguson83 suggested that the NO3 is very non-reactive and might be the terminal ion product of the whole procedure. It appears that NO3 is destroyed by ion recomb ination processes or photo detachment and may solvate with water or other species.84 NO + O3 NO2 + O2 (2-5) NO + O4 NO3 + O2 (2-6) NO2 + O2 (or O-) NO2 + O2 (or O) (2-7) The formation of reactant ions by DPIS is summarized in Figure 2-3. The negative reactant ions ge nerated by DPIS were investig ated in this research under different configurations includi ng fully open, open, and closed environments. (Figure 2-4) The DPIS bulb is placed 3 mm away from inlet of the mass spectrometer. The API source assembly

PAGE 44

44 was removed in the fully open configuration, back ed up 1 cm from the closed position in the open configuration, and attached to the closed position in the closed configuration. Representative spectra from each configura tion are shown in Figure 2-4. The DPIS generated the r eactant ion NO3 and the cluster ion HNO3NO3 in the fully open configuration. In the open and closed configuration, the DPIS produced NO3 -, CO3 -, and NO2 -. In the fully open configuration, NO3 is a very abundant reactant ion ow ing to the larger ionization area surrounds the discharge bulb th an corona discharge needle, which induce more oxygen and nitrogen provided from the open air involving in the reaction. In the closed configuration, the enclosed chamber was filled with carri er gas nitrogen and the formation of NO2 increases. The main reactant ion generated by APCI, as shown in Figure 2-6, is CO3 in all three configurations. This might be due to the smaller ionization area of the corona discharge needle, where insufficient NO and O3 exist to convert CO3 to NO3 or NO2 -. In order to discover how the gas compositi on generates different reactant ions, pure CO2, N2, O2, gas mixtures of N2/CO2, N2/O2, O2/CO2, and air were applied to fill the enclosed chamber as shown in Figure 2-4. (closed configurat ion) No reactant io n was observed with pure CO2 or gas mixtures of N2/CO2, and O2/CO2, due to the deficiency of oxygen, which makes the formation of the initial O2 ion impossible. The results are co nsistent with the reactions reported before.78 It is almost impossible to totally eliminate N2 and O2 from the enclosed chamber; therefore, some weak negative ions can be observed when applying O2 and N2 gas. However, strong NO3 and HNO3NO3 ion peaks were present in air, and a strong NO3 ion signal can be observed in a gas mixture of N2/O2. The results support the a ssumption of reactant ions produced by DPIS only when N2 and O2 are both present.

PAGE 45

45 To study the influence of different compositions of N2 and O2, different mixtures of oxygen in nitrogen were introduced to the chamber as shown in Figure 2-4; the results are shown in Figure 2-7. The high est intensities of NO3 and CO3 ions were generated at oxygen percentages of 5% and 2% in nitrogen, respectively. E xperiments show that the CO3 and NO3 ions intensities decrease with increasi ng amounts of oxygen resulting from less CO2 and N2 existed in the chamber. The CO2 and N2 are the major precursors in the formation of CO3 and NO2 -. The NO2 then quickly converts to NO3 by reacting with O3. When the content of oxygen in nitrogen is less than 2%, the formation of O-, O2 -, and O3 is insufficient to support the reaction with CO2 and NO to generate CO3 and NO2 -. Ionization Chemistry Explosive ions of negative polar ity at atmospheric pressure are formed in two steps: At the first step, reactant ions Rare formed from ionizing radiation; at the second step, explosive ions are formed from ion-molecule reactions of reactant ions with molecules of explosive substances. Ions of explosive substances are formed by reac tions such as electron capture (reaction 2-8), electron transfer (reaction 29), proton abstraction (reacti on 2-10), and adduct formation (reaction 2-11).85 M +eM(2-8) M + RM+ R (2-9) M + R[M-H]+ RH (2-10) M + RMP+ [R-P] (2-11) where M is a molecule of an explosive substance, H is a hydrogen atom, Ris a reactant ion, and P is a part of the reactant ion. Ions of nitroaromatic compounds are formed by reactions 2-8, 2-9, and 2-10. These species are strong gaseous acids because of the electron-withdrawing properties of the NO2 functional

PAGE 46

46 groups on the benzene ring and are responsible fo r the acidic character of the methyl group.85 Electrons from the ionization region or reactant ions can be easily transferred to nitroaromatic compounds due to the high electron affinity of the NO2 functional group, which enables the processes of reaction 2-8 and 29. Electron transfer readily oc curs with a negative corona discharge, where a high density of elect rons is generated that is about 106 times as much as that produced by a 63Ni source.86 In such a high electron density, reaction 2-8 is very efficient and trace amounts of any nitroaromatic compounds re sult in the production of negative ions. The general trend for proton abstraction from molecules of expl osive substances depends on the relative acidity of M and R-. Therefore, the higher the acidity of M compared to R-, the more readily reaction 2-9 proceeds. For both source s used in this resear ch, the ratio between Mand [M H]is quantitatively controlled by the amount of O2 in the nitrogen.87 In DPIS, O2and NO2are generated as reactant ions, and both pos sess strong gas-phase basicity, as shown in Table 2-1. Some nitroaromatic compounds with higher acidity, such as TNT and 2,4-DNT, may be ionized by proton ab straction (reaction 2-10). Nitramines and nitrate esters tend to form i ons by reaction 2-11. Because nitramines and nitrate esters do not have a positiv e electron affinity nor sufficient gas-phase acidity to be ionized by electron transfer or proton ab straction, adduct formation is th e most efficient approach for these compounds.88 Nitroaromatic Compounds TNT The negative-APCI spectrum of TNT (Figur e 2-8A) shows the production of the [M]ion at m/z 227. Spangler and Lawless89 thoroughly studied the ion ch emistry of TNT in air and nitrogen and found that the main ion created in nitroge n at 166 was Mvia electron attachment. It is assumed that an electron capture mechanis m occurs with APCI since TNT has three bulky

PAGE 47

47 electron-withdrawing nitro groups and there is an easily captu red electron produced by the corona discharge. The spectrum also show s two low-intensity fragment ions, [M-NO]at m/z 197 and [M-OH]at m/z 210, which may form either duri ng the ionization or during the desolvation processes in the heated capillary. The negative-DPIS spectra of TNT (Figure 29A, 2-10A) shows the predominant formation of the [M-H]ion. The formation of [M-H]ions may involve proton transfer between the analytes and basic reac tant ions such as NO2 and O2 -. Proton transfer can occur for analytes possessing gas-phase acidity stronger than that of O2 (353 kcal/mol) and NO2 (333.7 kcal/mol), which is true for TNT (315.6 kcal/mol) and 2,4DNT (328 kcal/mol). The comparison of the ratio between [M]and [M-H]reveals that the proton transfer is better in the open configuration because of more reactant ions ge nerated in open air. The DPIS spectra in the open configuration also shows more intense fragment and adduct ions at m/z 197, 260 and 274, corresponding to the ions of [M-NO]-, [M-NO+HNO3]and [M+HNO2]-, respectively, due to more complicated ionization reaction occurring around the DPIS source. TNB The negative-APCI spectrum of the byproduct TNB (Figure 2-8B) shows the major ion at m/z 213, [M]. It also produces a fragment ion at m/z 183, [M-NO]and two adduct ions at m/z 239, [M+CN]-, and m/z 244, [M+CH3O]-. Methanol has been observed to form adducts with other explosives.90 Proton abstraction from meth anol produces a methylate (CH3O-) ion, which is a strong Brnsted base (almost as strong as OH-) that readily reacts with many organic compounds with proton affinities lower than 379 kcal/mol.63 Under APCI, the CNion generated from acetonitrile has also been proven to react with TNB via nucleophilic attack on the benzene ring forming a Meisenheimer complex.91

PAGE 48

48 The major ions that appear in a ne gative-DPIS spectra of TNB are [M]and [M-NO]-. (Figure 2-9B, 2-10B) However, the intensity of [M-NO]in the open configuration is even higher than [M]-, which means more fragmentation occurs when the DPIS source was exposed to open air. The other ions produced include the ion at m/z 259, which is an adduct ion resulting from attachment of NO2 -, and the ions at m/z 239 and 244, which are generated from the adduction of the reactant ions, CNfrom acetonitrile and CH3O-from methanol, respectively. Tetryl The negative-APCI spectrum of tetryl (Fi gure 2-8C) shows a greater abundance of the [M*]ion ( m/z 242) and [M*-H]ion ( m/z 241) of N-methylpicramide. Because a methanol/water solutions were used in this research, no [M]ion was produced in the APCI mass spectrum, but a highly abundant N-methylpicram ide ion wass observed. This is due to the increased hydrolysis effect from the presence of water.92 The [M+CN]ion ( m/z 313) is generated because of acetonitrile. Negative DPIS spectra of tetryl in both the open and closed configur ations include an [MNO2]ion at m/z 241, which is the [M*-H]ion of N-methylpicramide, and the [M-NO]ion at m/z 257. (Figure 2-9C, 2-10C) The major difference between APCI and DPIS is the formation of the [M*]and [M*-H]ions of N-methylpicramide. The most intense ion in DPIS is the [M*-H]ion at m/z 241, while the [M*]ion at m/z 242 for APCI. That is mainly because more basic reagent ions such as O2 and NO2 are generated by the DPIS; then basic reactant ions can induce proton transfer to from N-methylpicramide. DNT Dinitrotoluene (DNT) isomers are byproducts originating from the manufacturing process of TNT, and their combined profile depends on the manufacturing processes (batch or continuous and concentration of acids) as well as the extent of the purification.37 The negative-

PAGE 49

49 APCI spectra of 2,4-DNT; 2,6-DNT; and 3,4DNT (Figure 2-8D-F) all yield an [M]ion as the major ion at m/z 182 and a fragment ion of [M-NO]at m/z 152, as confirmed by Lubman.93 Among the DNT isomers, only 2,6-DNT produces the minor methylate adduct ion [M+ CH3O]at m/z 213. As shown in Figure 2-9D-F, the negative DPIS spectra of DNT isomers in the closed configuration are similar to t hose from APCI. For 3,4-DNT, in addition to the molecular anion [M]at m/z 182, more fragment and dime r ions were observed at m/z 62, 152, and 350, ascribed to [NO3]-, [M-NO2]-, and [2M-CH2]-, respectively. In contrast to the closed c onfiguration, the DPIS spectra in the open configuration show a different ion pattern for these DNT isomers. (210D-F) When the source was exposed to open ai r, the principal ions presented in the DPIS spectra for 2,4-DNT, 2,6-DNT, and 3,4-DNT are [M-H]at m/z 181, [M]at m/z 182, and [MCH2]at m/z 168, respectively. However, all the spectra of the DNT isomers show the same [MNO]ion at m/z 152 and [M-HNO2+NO3]ion at m/z 197. The [M-NO]ion is the most common fragment ion form DNT isomers due to natural losses of NO. The [M-HNO2+NO3]ion is the resultant ion from the reaction between [M-HNO2]and the reactant ion NO3 -, which is especially apparent for 2,4-DNT. This phenomenon also indicates that more NO3 is form with DPIS in the open configuration which may be due to the larger ionization reaction area exposed in open air. DNB The negative-APCI spectra (Figure 2-8G) of 1,3-dinitrobenzene (DNB) shows ions at m/z 168 [M]-, 199 [M+ CH3O]-, and a very weak ion at m/z 138 [M-NO]-. The formation of the [M]ion can be attributed to an electron capture mechanism since DNB possesses a positive electron affinity (1.66eV). The formation of the [M-NO]ion can be interpreted as the result of in-source fragmentation of the [M]ions.88

PAGE 50

50 The negative-DPIS spectra of DNB (Figure 29G, 2-10G) are similar to APCI, with an abundant [M]ion peak at m/z 168 and a weak [M-NO]ion peak at m/z 138. The ratio between [M-NO]and [M]in the open configuration is higher th an the closed configuration, which may be because of a more energetic ionization react ion occurring around the DPIS bulb in open air since more reactant ions were generated. Nitramines RDX and HMX do not have positive electron affinities or sufficient gas-phase acidity to be ionized by electron capture, dissociative electron captu re, or proton transfer.88 However, chloride ion attachment can be a very specifi c and sensitive type of chemical ionization technique for the detection of nitramine and nitrate ester explosives. Caldwell et al.94 demonstrated that the highest sensitivity of halide attachment was for strong acids with Hacid values stronger than 350 kcal/mol. The chloride attachment has been proven as an efficient approach for the detection of RDX, showing th e limit of detection (LOD) in the femtomole range. In this research, 0.1% of carbon tetrachloride was used as an additive for this purpose. RDX RDX is a powerful, highly energetic chemical that is widely used in various military and civilian applications. The negative APCI spect ra (Figure 2-11A) of RDX show a relatively complicated ion pattern. No molecular ion was ob served for RDX. The major ions for RDX are [M+C2H4N3O]at m/z 324, [2M+NO2]at m/z 490 and [M+NO2]at m/z 268. The complicated ion pattern makes it difficult to identify RDX ju st by APCI spectra. The formation of chloride adduct ions [M+Cl]greatly enhances the analysis of RDX. The major ions observed in the spectra (Figure 2-11B) are adduct ions [M+Cl]and cluster ions [2M+Cl]-. The ion of [M + 35Cl]( m/z 257) is characterized by the presence of its isotope [M + 37Cl]( m/z 259), with one third the abundance.

PAGE 51

51 The negative DPIS spectra of RDX in the cl osed and the open conf igurations (Figure 212A and 2-13A) both include ions [M+NO2]at m/z 268, [M+NO3]at m/z 284 and [2M+NO2]at m/z 490. Although DPIS ionization, like APCI, does not generate molecular ions for RDX, the relatively simple ion pattern gives DPIS an advant age to detect RDX in the field where additives might be difficult or impossible to apply. When carbon tetrachloride is used as an additive for DPIS, additional [M+Cl]and [2M+Cl]ions are observed, but the intensity of [M+NO2]and [M+NO3]ions is higher than [M+35Cl]in the open configuration. (Figure 2-12B and 2-13B) This observation can be attributed to more abundant reactant ions NO2 and NO3 generated by the DPIS in open air. The competition between NO2 -, NO3 -, and Clattachment was determined by the concentration of CCl4 in sample solution and the generation of NO2 and NO3 ions. The NO3 adduction in the open configura tion for DPIS is more ready than in the closed configuration due to the evident NO3 ion formation in open air. HMX The negative APCI spectrum (Figure 2-11C) of HMX includes as the base peak the [M-H]ion at m/z 295. Other relevant ions include those at m/z 123, 166, 203 and 342 representing the fragment ions [NO3NO+CH3O]-, [M-C2N2O2-NO2]-, [M-HNO2NO2]-, [M-HNO2NO]-, and the adduct ion [M+NO2]-, respectively. Notwithstanding the ab undance of the de pronated ion of HMX, the numerous ions present in the spectru m might still interfere in the determination of HMX. The use of a chloride additive substantia lly simplifies the spectra. (Figure 2-11D) The main ions observed in the spectrum (Fi gure 2-12D) are only the adduct ion [M+Cl]and the [M+Cl HNO2]fragment ion. The negative DPIS spectra of HMX in both c onfigurations (Figure 2-12C and 2-13C) show the same [M-H]and [M+NO2]ions, except for some weak fragment ions that can be observed in the open configuration. Th e dominant formation of [M+NO2]indicates that HMX prefers to

PAGE 52

52 perform adduct formation rather than proton abst raction by DPIS. In the presence of chloride, DPIS spectra (Figure 2-12D and 2-13D ) show an abundance of the [M+Cl]ion and its fragment ion [M+Cl-HNO2]for both configurations. The major di fference between both configurations is the more intense [M+NO2]ion present in the open confi guration, which is due to more NO2 ions generated by DPIS in open air, which compete with Clfor adduction formation. Nitrate Esters Nitrate esters have a high electron affinit y, which makes them excellent candidates for analysis in the negative-ion mode. However, in the absence of any additives, the mass spectra are usually characterized by various adduct ions formed from the decomposition fragments of the nitrate esters themselves, or the impurities pr esent in the analytical system. The lack of specificity in the mass spectra sometimes makes the unambiguous identification of these nitrate esters difficult.75 In this research, 0.1% carbon tetrachlor ide was added to the solution in order to overcome this problem. NG The negative APCI spectrum (Figure 2-14A) of NG, the active component in dynamite, has NO3 as its most prominent ion at high temperature. Ewing31 proposed that the ionization of NG occurs via loss of NO3 -, which is the only ion observed at high temperature, as shown in reaction 2-12. However, as the temper ature is lowered, the adduct ion [M+NO3]( m/z 289) forms as shown in reaction 2-13. M + eNO3 + [M-NO3] (2-12) NO3 + M [M+NO3](2-13) Besides NO3 and [M+NO3]-, [M+NO2]( m/z 273) is the other intens e ion in the spectrum. Addition of carbon tetrachloride pro duces the base peak ion of [M+35Cl]at m/z 262, with the 37Cl isotope peak at m/z 264, and its dimer ion of [2M-H+35Cl]at m/z 488.(Figure 2-14B) The

PAGE 53

53 fragment ion of NO3 at m/z 62 and the adduct with fragment ion [M+NO3]at m/z 289 can be also observed. In general, the negative DPIS spectra of NG (Figure 2-15A,B and 2-16A,B) show a similar ion pattern with the spectra generated by APCI. However, more abundant [M+NO3]and [M+NO2]ions can be found in the spectra acquire d by DPIS in the open configuration, which suggests that the NO3 and NO2 do not only come from the fr agment of NG, but are also generated by DPIS in open air. PETN Similarly, PETN was fragmented to NO3 and accompanied by adduction to form [M+NO3]in air. As shown in the APCI spectra of PETN (Figure 2-14C), prominent ions at m/z 62, 378, 315, and 362 are attributed to the NO3 -, [M+NO3]-, [M-H]-, and [M+NO2]-, respectively. A number of other weak ions were observed at lower m/z values, which are mainly fragment ions of PETN. Addition of chloride ion gave an improved response for P ETN by generating adduct ions at m/z 351 and 353, corresponding to the 35Cl and 37Cl isotope peaks for [M+Cl](Figure 214D). Clearly, having only an abundant [M+Cl]ion represents a good target ion for monitoring PETN. The negative DPIS spectra of PETN (Fi gure 2-15C and 2-16C) include an intense [M+NO2]ion at m/z 362 and two weaker ions, NO3 at m/z 62 and [M+NO3]at m/z 378, which is less complicated than the spectra generate d by APCI for PETN. Wh en carbon tetrachloride was used as an additive, the primary ion became [M+Cl]ion in the open and the closed configurations. (Figure 2-15D and 2-16D) Sim ilar to NG, the difference between the open and closed configurations by DP IS is the production of [M+NO3] and [M+NO2] ions in the open configuration. In contrast, the spectrum (Figure 2-16D) acqui red by DPIS in the closed configuration is almost identic al to the one produced by APCI (Figure 2-15D) for PETN.

PAGE 54

54 Conclusions The evaluation of the two API sources in th is research are summarized in Table 2-11, which compares the characteristic ions, absolute intensities, and relative intensities between APCI and DPIS for analysis of explosives. Se veral observations should be noted, which help to explain the difference in ionization mechanis m and performance between DPIS and APCI. The first observation is that DPIS typically gives more structural information through increased fragmentation. That is presumably because the reaction region of DPIS, which includes the space around the neon bulb, is far larg er than the area around the corona discharge needle tip where the ions are genera ted by APCI. DPIS creates more O2 -, NO2 -, and NO3 reactant ions, and they enhance proton abstract ion and adduct formation reactions and have more energetic reactions, increasing fragmentation. The generation of reactant ions also explains the lower ionization efficiency of DPIS in the cl osed configuration becau se fewer and different reactant ions are created. Th at is also why the spectra ac quired by DPIS in the closed configuration are more similar to those produced by APCI. The second observation is that the spectra of explosive compounds produced by DPIS are comparable to those formed by APCI; however, th e formation of nitrate and nitrite adduct ions with the explosives is more pronounced with the DPIS source. Typically, APCI produces a complex spectrum of low intensity ions consisting of NO3 -, M-, [M+NO2]-, [M+NO3]-, and other fragment and background ions. The use of DPIS provides re actant ions of NO2 and NO3 -; the spectra appear cleaner even without the addition of chlorine, showing only the NO2 and NO3 reactant ions and the [M+NO2]and [M+NO3]product ions. This will gr eatly enhance sensitivity, selectivity, and remove background interference, a nd will be a benefit for explosive investigation in the field, where additives may not be available for use. In addition, different types of spectra

PAGE 55

55 which either present more information about stru cture or more abundant molecular ions can be obtained from DPIS by adjusti ng the amountt of surrounded air. Lastly, APCI yields a higher ionization effici ency than DPIS for nitroaromatic compounds, and for chlorine adducts of nitr amines and nitrate esters. In fact, only a small portion of ions generated by DPIS are able to be detected in th is research due to the sp atial obstruction of the neon bulb, which is situated betw een the nebulizer and the inlet of mass spectrometer. Further modifications of source geometry can be exp ected to improve the performance of DPIS. In summary, DPIS has been shown in th is research to ionize explosive compounds efficiently, allowing for their identification. The design and low power consumption of DPIS also make it ideal for portable applications Although the geomet ry of DPIS sti ll needs to be modified to obtain better performance, the rich ion patterns and decreased complexity of spectra for nitramines and nitrate esters have provided an alternative choice other than corona discharge for explosive investigation.

PAGE 56

56 mesh RF source voltage DC offset voltage Neon bulb 1cm Figure 2-1. Configuration of DPIS. (A ) Schematic, (B) Actual picture. A B

PAGE 57

57 20 30 40 50 60 70 80 90 1 0 m/z 100 0 50 100 0 50 100 0 50 Relative Abundance 100 0 50 37.0 59.1 54.8 74.9 64.9 33.0 47.1 64.9 33.0 47.1 64.9 33.0 49.9 60.1 42.1 47.1 73.9 58.9 Air Methanol Methanol/Water TNT 10ppm A B Figure 2-2. Comparison of reactant ions genera ted by DPIS observed with air, methanol, methanol/water, and 10 ppm TNT in (A) negative mode and (B) positive mode.

PAGE 58

58 O2 O O2+M O4 -O3 O3 CO3 -CO4 O CO2CO2 NO NO NO2 -NO3 O3 NO NO NO2 NO2ee-O2 y Figure 2-3. Schematic procedure of reactant ions formation by DPIS.77 gas Vaporizer Mass Spectrometer Heated Capillary DPIS Vaporizer Heated Capillary DPIS Heated Capillary DPIS Closed OpenFully open 1 cm Figure 2-4. Three different confi gurations where the DPIS was placed: closed, open and fully open configuration.

PAGE 59

59 20 40 60 80 100 120 140 160 180 200 m/z 0 20 40 60 80 100 0 20 40 60 80 100 Relative Abundance 0 20 40 60 80 100 124.97 62.17 62.17 60.17 124.97 46.07 78.80 62.10 60.04 46.07 92.04Fullyopen NL:4.39 E4 Open NL:4.72 E4 Close NL:1.97E4NO3 -HNO3NO3 -NO2 -CO3 -NO3NO-NO3OH-HNO3NO3 Figure 2-5. Mass spectra of negativ e ions generated in air by DPIS with fully open, open and closed configuration.

PAGE 60

60 20 40 60 80 100 120 140 160 180 200 m/z 0 20 40 60 80 100 0 20 40 60 80 100 Relative Abundance 0 20 40 60 80 100 60.11 62.10 77.14 78.80 46.07122.91 32.1091.97 60.13 61.13 46.13 91.93 32.07 60.13 62.13 122.87 59.13 77.13 46.13 Fully open Open CloseNO3 -CO3 -NO2 -O2 Figure 2-6. Mass spectra of negati ve ions generated in air by APCI with fully open, open and closed configuration.

PAGE 61

61 0.00E+00 5.00E+04 1.00E+05 1.50E+05 2.00E+05 2.50E+05 3.00E+05 3.50E+05 4.00E+05 4.50E+05 0%2%5%10%15%20%25%30%35%40%45%50%Intensity (counts) Gas concentration (O2in N2, v/v)Reactant ion of DPIS vs gas composition 125 [2NO3+H]124 [2NO3]62 [NO3]60 [CO3]46 [NO2]Figure 2-7. Comparison of the reac tant ions intensity as a func tion of the composition between oxygen and nitrogen. Table 2-1. Gas-phase acidity values for reactant ions.95

PAGE 62

62 100 150 200 250 300 350 400 450 500 m/z 0 50 100 0 50 100 0 50 100 0 50 100 Relative Abundance 0 50 100 0 50 100 0 50 100 227.07 213.13 239.07 242.13 182.07 182.07 152.20 182.07 152.33 168.13 199.00 A B C E F D G Figure 2-8. Negative APCI mass spectra of n itroaromatic compounds: (A) TNT (MW = 227), (B) TNB (MW = 213), (C) Tetryl (MW = 287) (D) 2,4-DNT (MW = 182), (E) 2,6-DNT (MW = 182), (F) 3,4-DNT (MW = 182), (G) 1,3-DNB (MW = 168). Table 2-2. Mass spectral data of nitroa romatic compounds analyzed by APCI-MS. Explosive Molcular Weight (g/mole) m/z (Ion Abundance) Opitmal VT ( ) TNT227.13 197(4%)[M-NO]-, 210(3%)[M-OH]-, 227(100%)[M]-100 TNB213.1 183(3%)[M-NO]-, 213(100%)[M]-, 239(23%)[M+CN]-, 244(8%)[M+CH3O]-100 Tetryl287.14 241(68%)[M-NO2]-, 242(100%)[M-NO2+H]-, 313(4%)[M+CN]-130 2,4-DNT182.13 165(3%)[M-OH]-, 182(100%)[M]-130 2,6-DNT182.13 152(7%)[M-NO]-, 182(100%)[M]-, 213(3%)[M+CH3O]-130 3,4-DNT182.13 152(5%)[M-NO]-, 182(100%)[M]-130 1,3-DNB168.11 138(4%)[M-NO]-, 168(100%)[M]-, 199(85%)[M+CH3O]-130

PAGE 63

63 A B C E F D G 100 200 300 400 500 m/z 0 50 100 0 50 100 0 50 100 0 50 100Relative Abundance 0 50 100 0 50 100 0 50 100 226.20 253.00 197.27 213.13 239.07 183.27 258.87 275.87 62.13 241.13 256.93 329.07 228.27 181.07 182.07 166.27 197.07 182.07 152.27 182.07 350.07 152.27 62.13 303.07 197.13 244.00 121.33 361.00 168.13 138.27194.00 Figure 2-9. Negative DPIS mass spectra of nitroaro matic compounds in the closed configuration: (A) TNT (MW = 227), (B) TNB (MW = 213), (C) Tetryl (MW = 287), (D) 2,4-DNT (MW = 182), (E) 2,6-DNT (MW = 182), (F) 3,4-DNT (MW = 182), (G) 1,3-DNB (MW = 168). Table 2-3. Mass spectral data of nitroaromatic compounds analyzed by DPIS-MS in the closed configuration. Explosive Molcular Weight (g/mole) m/z (Ion Abundance) TNT227.13 TNB213.1 Tetryl287.14 2,4-DNT182.13 2,6-DNT182.13 3,4-DNT182.13 1,3-DNB168.11

PAGE 64

64 A B C E F D G 100 200 300 400 500 m/z 0 50 100 0 50 100 0 50 100 0 50 100Relative Abundance 0 50 100 0 50 100 0 50 100 226.20 197.27 273.93 62.13 304.87 183.27 140.93 183.27 258.87 62.13 241.13 256.93 303.80 62.13 227.27 181.07 181.27 226.20 167.20 264.33 62.13 346.00 182.07 152.20 197.27 60.13 168.07 350.07 182.07 303.07 152.27 214.00 168.13 138.27 183.27 62.13 Figure 2-10. Negative DPIS mass spectra of nitr oaromatic compounds in the open configuration: (A) TNT (MW = 227), (B) TNB (MW = 213), (C) Tetryl (MW = 287), (D) 2,4-DNT (MW = 182), (E) 2,6-DNT (MW = 182), (F) 3,4-DNT (MW = 182), (G) 1,3-DNB (MW = 168). Table 2-4. Mass spectral data of nitroaroma tic compounds analyzed by DPIS-MS in the open configuration. Explosive Molcular Weight (g/mole) m/z (Ion Abundance) TNT227.13 TNB213.1 Tetryl287.14 2,4-DNT182.13 2,6-DNT182.13 3,4-DNT182.13 1,3-DNB168.11

PAGE 65

65 A B C D 100 150 200 250 300 350 400 450 500 m/z 0 50 100 0 50 100 0 50 100Relative Abundance 0 50 100 324.00 489.67 268.00 387.87 175.93 219.40 123.27 469.67 102.13 129.13 281.00 344.87 220.93 439.00 189.07 478.80 257.07 259.00 295.00 123.27 203.07 342.07 166.27 219.33 398.07 499.07 322.00 121.20 419.00 129.20 230.00 370.80 331.13 284.07 333.13 286.07 Figure 2-11. Negative APCI mass spectra of n itramines: (A) RDX (MW = 222), (B) RDX + Cl, (C) HMX (MW = 296), (D) HMX + Cl. Table 2-5. Mass spectral data of nitramines analyzed by APCI-MS. Explosive Molcular Weight (g/mole) m/z (Ion Abundance) Opitmal VT ( ) RDX222.12 102(21%)[C2H4N3O]-, 123(36%)[HNO3NO+CH3O]-, 176(42%)[MNO2]-, 268(64%)[M+NO2]-, 324(100%)[M+C2H4N3O]-, 100 RDX+Cl 257(45%)[M+35Cl]-, 259(13%)[M+37Cl]-, 479(100%)[2M+35Cl]-, 481(31%)[2M+37Cl]-100 HMX296.16 123(92%)[NO3NO+CH3O]-, 166(57%)[M-C2N2O2-NO2]-, 203(84%)[M-HNO2NO2]-, 219(39%)[M-HNO2NO], 295(100%)[MH]-, 342(68%)[M+NO2]-130 HMX+Cl 284(40%)[M+35Cl-HNO2]-, 286(13%)[M+37Cl-HNO2]-, 331(100%)[M+35Cl]-, 333(31%)[M+37Cl]-130

PAGE 66

66 A B C D 100 150 200 250 300 350 400 450 500 m/z 0 50 100 0 50 100 0 50 100Relative Abundance 0 50 100 489.67 268.00 284.00 62.13 478.80 257.13 268.00 62.13 342.07 295.00 331.13 333.13 284.07 Figure 2-12. Negative DPIS mass spectra of nitram ines in the closed configuration: (A) RDX (MW = 222), (B) RDX + Cl, (C) HMX (MW = 296), (D) HMX + Cl. Table 2-6. Mass spectral data of nitramines an alyzed by DPIS-MS in th e closed configuration. Explosive Molcular Weight (g/mole) RDX222.12 RDX+Cl HMX296.16 HMX+Cl

PAGE 67

67 A B C D 100 150 200 250 300 350 400 450 500 m/z 0 50 100 0 50 100 0 50 100Relative Abundance 0 50 100 268.00 489.67 284.00 62.13 478.80 284.00 268.00 62.13 342.07 295.00 62.13 356.93 123.27 264.40 331.20 333.13 284.07 Figure 2-13. Negative DPIS mass spectra of nitr amines in the open configuration: (A) RDX (MW = 222), (B) RDX + Cl, (C) HMX (MW = 296), (D) HMX + Cl. Table 2-7. Mass spectral data of nitramines analyzed by DPIS-MS in the open configuration. Explosive Molcular Weight (g/mole) m/z (Ion Abundance) Opitmal VT ( ) RDX222.12 62(27%)[NO3]-, 268(100%)[M+NO2]-, 284(67%)[M+NO3]-, 490(92%)[2M+NO2]-100 RDX+Cl 62(20%)[NO3]-, 257(32%)[M+35Cl]-, 268(41%)[M+NO2]-, 284(54%)[M+NO3]-, 479(100%)[2M+35Cl]-, 490(45%)[2M+NO2]-100 HMX296.16 295(23%)[M-H]-, 342(100%)[M+NO2]-130 HMX+Cl 284(13%)[M+35Cl-HNO2]-, 331(100%)[M+35Cl]-, 333(33%)[M+37Cl]-, 342(19%)[M+NO2]-250

PAGE 68

68 100 150 200 250 300 350 400 450 500 m/z 0 50 100 0 50 100 0 50 100 Relative Abundance 0 50 100 62.13 288.80 272.80 125.00 261.87 62.13 263.80 488.33 62.13 377.93 314.87 361.87 219.40 123.27 166.27 481.73 89.20 350.87 352.80 A B C D Figure 2-14. Negative APCI mass spectra of n itrate esters: (A) NG (MW = 227), (B) NG + Cl, (C) PETN (MW = 316), (D) PETN + Cl. Table 2-8. Mass spectral data of ni trate esters analyzed by APCI-MS. Explosive Molcular Weight (g/mole) m/z (Ion Abundance) NG227.09 62(100%)[NO3]-, 125(7%)[HNO3NO3]-, 273(11%)[M+NO2]-, 289(60%)[M+NO3]NG+Cl 62(54%)[NO3]-, 262(100%)[M+35Cl]-, 264(30%)[M+37Cl]-, 289(15%)[M+NO3]-, 488(21%)[2M-H+35Cl]PETN316.14 62(100%)[NO3]-, 123(18%)[NO3NO+CH3O]-, 315(31%)[M-H]-, 347(14%)[M+CH3O]-, 362(20%)[M+NO2]-, 378(77%)[M+NO3]PETN+Cl 351(100%)[M+35Cl]-, 353(32%)[M+37Cl]

PAGE 69

69 A B C D 100 150 200 250 300 350 400 450 500 m/z 0 50 100 0 50 100 0 50 100Relative Abundance 0 50 100 62.13 288.80 272.80 108.93 228.00 261.87 62.13 263.80 488.53 361.87 62.13 377.93 351.07 350.87 352.87 62.13 Figure 2-15. Negative DPIS mass spectra of nitrat e esters in the closed configuration: (A) NG (MW = 227), (B) NG + Cl, (C) PETN (MW = 316), (D) PETN + Cl. Table 2-9. Mass spectral data of nitrate esters analyzed by DPIS-M S in the closed configuration. Explosive Molcular Weight (g/mole) m/z (Ion Abundance) Opitmal VT ( ) NG227.09 62(100%)[NO3]-, 273(27%)[M+NO2]-, 289(72%)[M+NO3]-100 NG+Cl 62(62%)[NO3]-, 262(100%)[M+35Cl]-, 264(29%)[M+37Cl]-, 289(3%)[M+NO3]-, 488(21%)[M+35Cl]-100 PETN316.14 62(18%)[NO3]-, 315(4%)[M-H]-, 362(100%)[M+NO2]-, 100 PETN+Cl 62(5%)[NO3]-, 351(100%)[M+35Cl]-,353(32%)[M+37Cl]-130

PAGE 70

70 A B C D 100 150 200 250 300 350 400 450 500 m/z 0 50 100 0 50 100 0 50 100Relative Abundance 0 50 100 62.13 288.80 272.87 86.20 109.00 62.13 261.93 288.80 488.47 86.20 361.93 377.93 62.13 350.93 352.87 62.13 377.87 Figure 2-16. Negative DPIS mass spectra of nitr ate esters in the open configuration: (A) NG (MW = 227), (B) NG + Cl, (C) PETN (MW = 316), (D) PETN + Cl. Table 2-10. Mass spectral data of nitrate esters analyzed by DPIS -MS in the open configuration. Explosive Molcular Weight (g/mole) m/z (Ion Abundance) Opitmal VT ( ) NG227.09 62(100%)[NO3]-, 273(16%)[M+NO2]-, 289(91%)[M+NO3]-100 NG+Cl 62(100%)[NO3]-, 262(75%)[M+35Cl]-, 264(30%)[M+37Cl]-, 273(9%)[M+NO2]-, 289(51%)[M+NO3]-, 488(15%)[2M-H+35Cl]-100 PETN316.14 62(21%)[NO3]-, 315(4%)[M-H]-, 362(100%)[M+NO2]-, 100 PETN+Cl 62(12%)[NO3]-, 315(3%)[M-H]-, 351(100%)[M+35Cl]-, 353(33%)[M+37Cl]-, 362(22%)[M+NO2]-, 378(11%)[M+NO3]-130

PAGE 71

71 Table 2-11. The main ions of explosive compounds determined by APCI and DPIS. molecular code MW Major ion ( m/z ) Intensity ( counts ) Relative Percenta g e Major ion ( m/z ) Intensity ( counts ) Relative p ercenta g e Major ion ( m/z ) Intensity ( counts ) Relative p ercenta g e TNT227.13 227[M]-6.25E+06100% 226[M-H]-4.15E+05100% 226[M-H]-2.24E+05100% TNB213.1 213[M]-1.39E+06100% 183[M-NO]-1.78E+05100% 213[M]-5.19E+04100% Tetryl287.14 2 42[M-NO2+H] 1.79E+06100% 241[M-NO2] 8.23E+05100% 241[M-NO2] 1.06E+04100% 24DNT182.13 182[M]-2.30E+06100%181[M-H]-6.87E+04100% 182[M]-2.02E+04100% 26DNT182.13 182[M]-3.57E+06100% 182[M]-1.61E+05100% 182[M]-2.46E+04100% 34DNT182.13 182[M]-1.42E+06100% 168[M-CH2]2.27E+05100%182[M]-1.07E+04100% 13DNB168.11 168[M]-6.24E+05100% 168[M]-8.21E+04100% 168[M]-3.87E+03100% RDX222.12 268[M+NO2]-1.84E+0563% 268[M+NO2]-1.41E+05100% 268[M+NO2]-1.32E+05100% RDX+Cl 257 [M+35Cl]-6.77E+0545% 284 [M+NO3]-1.15E+0554% 257 [M+35Cl]-1.65E+0540% HMX296.16 295[M-H]-1.15E+05100% 342 [M+NO2]-1.46E+05100% 342 [M+NO2]-2.74E+05100% HMX+Cl 331 [M+35Cl]-5.86E+05100% 331 [M+35Cl]-1.13E+05100% 358 [M+NO3]-3.15E+05100% PETN316.14 378[M+NO3]-2.52E+0577% 362[M+NO2]-3.86E+05100% 362[M+NO2]-2.20E+05100% PETN+Cl 351 [M+35Cl]-1.96E+06100% 351 [M+35Cl]-4.54E+05100% 351 [M+35Cl]-9.04E+05100% NG227.09 289[M+NO3]-1.24E+0560% 289[M+NO3]-7.25E+0491% 289[M+NO3]-3.51E+0472% NG+Cl 262[M+35Cl]-4.20E+05100% 262 [M+35Cl]-6.44E+0475% 262 [M+35Cl]-1.93E+05100% APCIDPIS(open)DPIS(closed)

PAGE 72

72 CHAPTER 3 FUNDAMENTALS OF HIGH-FIELD ASY MMETRIC WAVEFORM ION MOBILITY SPECTROMETRY (FAIMS) FOR THE ANALYSIS OF EXPLOSIVES Introduction FAIMS is a sensitive and sele ctive technology for the detecti on and identification of trace constituents in ambient air or liquid samples. This technology separate s gas-phase ions based on certain properties of ions that appear to be independent of both the lo w-field collision cross section and the mass-to-charge ratio.51 The separation of ions by FAIMS is fast and, therefore, may replace slower separation techniques, su ch as gas chromatography (GC), capillary electrophoresis (CE) and high performance liqui d chromatography (HPLC). Furthermore, the signal-to-noise ratio (S/N) can be improved by pa ssing ions that originate from atmospheric pressure sources such as APCI and DPIS through a FAIMS device to select the ion of interest in preference to the background ions. The improvem ent of S/N can lead to increased detection limits, which, in some cases, will allow simplifica tion of sample handling via preconcentration or extraction.36 Therefore, FAIMS can offer an additiona l level of separation to simplify complex mixtures that already provided by chromatography and the m/ z separation by a mass analyzer. The cylindrical geometry of FAIMS used in th is research embodies a unique capability of focusing ions by an electric fi eld at atmospheric pressure.96 This gives higher sensitivity than commercial IMS, in which ion diffusion in th e drift tube causes th e ion cloud to expand, resulting in reduced transfer of ions to the mass spectrometer.53 The appearance of ion focusing is known to depend on ion polarity, -dependence sign (positive sign corresponds to an increase in ion mobility with increasing field st rength), and separation field polarity.97 The use of APCIFAIMS-MS can be expected to eliminate the need for chromatographic separation and allows for very rapid sample processing and sensitive detection.

PAGE 73

73 Investigations of FAIMS for e xplosive detection have been pe rformed by several groups. Buryakov et al. described the detection of explosive vapors in air using FAIMS,98 the qualitative analysis of explosives by FAIMS99, and the analysis of explosiv es with multicapillary-column gas chromatography and FAIMS.100 Eiceman et al.32 examined the separation of ions from explosives in FAIMS by vapor-modified drift gas. Three trends in ion mobility, as a function of electric field, have been reported.51 As electric field strength increases, the mobility of a type A ion increases, a type C ion decreases, and a type B ion initially increases before decr easing. These differences in ion behavior are ascribed to interactions of the ion structure, collisional cross-section, and instrumental parameters including dispersion voltage (DV), ca rrier gas composition, electrode temperature, and others. The ion mobility at high-fiel d strength and, hence the observed compensation voltage, is also affected by the gas composition. This is likely due to long-range, ion-induced dipole attractive interactions betw een the ion and the bath gas. The strength of the interaction depends on the polarizability of the bath gas a nd the size and charge of the ion of interest.47 Changes in temperature are reflected in changes of the thermal energy of the source, which can change the ion-induced dipole in teractions potential well relative to the thermal energy of the source.49 This may cause the compensation voltage (CV) to shift as the temperature changes. The aim of this research is to study the effect of DV, CV scan rate, curtain gas flow rate, carrier gas composition, and elec trode temperature on the separation of negative ions from eleven explosives (TNT, TNB, tetryl, 1,3DNB, 2,4-DNT, 2,6-DNT, 4-DN T, RDX, HMX, NG, and PETN). The other goal is to find optimal parameters of FAIMS separation and to assess their analytical characteristic in the detection of explosives.

PAGE 74

74 Experimental The use of FAIMS requires the op timization of several paramete rs in order to obtain the maximum benefit from the FAIMS device. For th e purpose of this work, five parameters were optimized: DV, CV scan rate, curtain gas flow rate, carrier gas composition, and electrode temperature. The effects that these parameters have on the CV value, the signal intensity, and the peak width were monitored for target explosive compounds. Experiments were performed employing a FAIM S-MS system, comprising of a cylindrical FAIMS device (Thermo Scientific, San Jose, CA ) and a quadrupole ion trap mass spectrometer (LCQ, Thermo Scientific). Gas-phase explosiv e ions were generated by atmospheric pressure chemical ionization (APCI) using a corona discharg e needle positioned at an angle of 45 and ~1 cm from the opening of the curtai n plate of FAIMS device. The cylindrical FAIMS cell consists of two electrodes, an inner and outer electrode. The combinati on of inner electrode, having an outer radius of 6.5 mm, and outer electrode, having an inner radi us of 9.0 mm, makes a gap of 2.5 mm for ion transmission. The asymmetric wave form (750 kHz) and the direct current (DC) CV were both applied to the inner electrode of the FAIMS cell. The DV was V to V. The DV is measured as the magnitude of the high-voltage pulse of the asymmetric waveform. A constant DC bias voltage of 25 V was applied to the outer cylinder of the FAIMS device and to the inlet of the mass spectrome ter. The curtain plate was held at -1000 V to assist negative ions to transport across the desolvation region. In order to connect the Thermo FAIMS cell, designed for Thermo TSQ mass spectrometer, onto the LCQ, a brass capillary extender (i.d. = 0.76 mm, o.d. = 22 mm) (Figure 3-1 and 3-2) was designed to serve as an interface. The setup of the APCI-FAIMS-MS instrument is shown in Figure 1-11. All ions of a given polarity a nd ion type (A or C) can be analyzed by scanning the CV, which results in a CV spectrum, also called th e total ion current-CV spectrum (TIC-CV). An

PAGE 75

75 ion-selective CV spectrum (IS-CV) can be obt ained for each individual ion by plotting the intensity of ions of that specific m/z versus CV. The CV was scanned from -20.0 to 20.0 V at a scan rate between 5.0 and 20.0 V/min. CV values and signal intensity were then taken from the maximum peak height in the IS-CV. The peak width is taken from the full width at half maximum (FWHM) of the peak. The mean of each value for three replicates is reported in these studies. The FAIMS carrier gases were passed through se parate charcoal/molecular sieve filters before being mixed together and introduced in to the region between th e curtain plate and the orifice of the FAIMS analyzer at a flow rate ra nging from 2.0 to 3.5 L/min. In this research, N2, He, CO2, SF6, and the mixture of these gases were used to evaluate their influence on the CV, peak width and signal intensity for the ions of interest. To control the temperature of i nner and outer electrodes, channe ls for the passage of heated gas were drilled into both electr odes. The outer electrode consists of a cylinder with an inner diameter (i.d.) of 18 mm that was bored into a solid block of stainle ss steel. Channels to the left and right of the cylinder carry air in and out of the block. The i nner electrode has a PEEK insert that directs gas to the top of th e electrode and then along the inner surface of the el ectrode to an exhaust port. Under conditions of active heating of the electrode, the in ner and outer electrode temperatures were set between 40-90 For the APCI source, the vaporizer temperatur e was set to 150 C. The heated capillary temperature and voltage were set to 130 C and -25.0 V, respectively. The discharge current was set at 5 A and the tube lens offset was set to 30.0 V. The sheath gas (N2) was set to 20.0 (arbitrary units) and the injection flow ra te of the analyte was maintained at 20.0 L/min.

PAGE 76

76 Eleven explosive compounds (TNT, TNB, tetryl, 1,3-DNB, 2,4DNT, 2,6-DNT, 4-DNT, RDX, HMX, NG, and PETN) were studied. Thes e explosives were pr ovided as acetonitrile solution by Dr. Jehuda Yinon of th e Weizmann Institute of Science, and were obtained from the Analytical Laboratory of the Israeli Police Headqu arters. The explosive solutions were further diluted in a solvent containing 65% methanol and 35% deionized water to a concentration of 10 g/mL. Approximately 0.1% of car bon tetrachloride was used as an additive in some solutions to form stable adduct ions with nitr amine and nitrate ester explosives. Results and Discussion Effects of CV Scan Rate For practical applications, narrow peak width, high transmission, and short detection time are desired for separation and detection techniqu es. In this experiment, CV scan rates were explored to acquire better re solution, transmission, and mi nimized detection time. The effect of the scan rate on CV was characterized. Spectra for TNT ( m/z 227) were collected at various scan rates from 2.5 to 20.0 V/sec; the CV value, peak width, and signal intensity plotted as a function of the scan rate are shown in Figure 3-3. The results show that scans with lower scan rates lead to narrower peak widths, higher intensity, and relatively constant CV value. Higher resolution (narrowe r peaks) and transmission (signal intensity) would be expected when using a lower scan rate because for a given CV range, increased increments of the RF voltage permits more ions to be transm itted at the optimum CV. However, lower scan rates extend the detection time, which is not fa vorable for field applications. A compromised scan rate at 10 V/s was applied throughout the following research, which induced a 30% increase in peak widths, a 10% decrease in signal intens ity, yet a 4-fold increase in detection times, compared to a scan rate of 2.5 V/s.

PAGE 77

77 Effect of Curtain Gas Flow Rate In FAIMS, it is essential to introduce cl ean, dry gas into the FAIMS cell to achieve optimum performance. This gas is referred to as the curtain gas and is introduced into the region between the curtain plate and the orifice into the FAIMS analyzer. The majority of the gas exits through the curtain plate to aid in desolvation of ions from the APCI source and to minimize the entrance of droplets and neutral molecules from the solvents into the FAIMS analyzer. The remainder of the gas is drawn into the FAIMS analyzer at a flow rate of ~0.7 L/min, which carries the ions around both sides of the inner cy linder and through the heat ed capillary into the mass spectrometer. In this experiment, nitrogen was used as the curtain gas. The data were acquired with the DV at 4000 V for HMX and PETN, and the DV at 4000V for the rest of explosive compounds. The curtain gas flow rate was varied from 2.0 L/min to 3.5 L/min. Figure 3-4, 3-5, and 3-6 show the effects that curtain gas flow rate has on the CV value, peak width, and signal intensity, respectively. Figure 3-4 illustrates the effect of increasing the gas flow ra te from 2.0 L/min to 3.5 L/min on the CV for explosive compounds. When increasing the curtain gas flow rate, it was observed from the plot that the CV remained relatively constant for all compounds. This result verifies that the gas flow rate has no effect on the CV. Figure 3-5 shows that the narrowest peak width was achieved for nine of the ten compounds at the minimum curtain gas flow rate of 2 L/min. This is unexpected, since an ions resident time in the FAIMS cell is determined by the carrier gas flow rate passing through the cell, and that should be constant at 0.7 L/min, set by the conductan ce of the heated capillary into the vacuum of the mass spectrome ter. In this case, however, the connection between the FAIMS cell and the brass capillary extend er and the extender and the heated capillary were not gas-tight.

PAGE 78

78 At higher flow rates (and thus higher pressures) leakage at these connections will increase, increasing the flow rate through the FAIMS cell. At lower curt ain gas flow rates, the flow through the FAIMS cell will be lower, and resident time for ions will increase. If an ion stays longer in the FAIMS cell, the ion experiences more cycles of the waveform that is used to resolve different compounds. Therefore, lower carri er gas flow rates enab les the FAIMS to more accurately resolve ions with ad jacent CV values, which means higher resolutions or narrower peak widths can be achieved. Operating FAIMS at very low curtain gas flow rates affects the signal intensity by inadequately desolvating ions as they enter the FAIMS cell th rough the orifice in the curtain plate. On the contrary, if the flow rate is t oo high it may cause a decrease in intensity due to a decreased number of ions entering the FAIMS ce ll that have to compete against the high flow rate of gas exiting the orif ice in the curtain plate.101 Figure 3-6 demonstrates that only HMX, PETN, and TNT have an apparent drop in signal in tensity with increasing curtain flow rate while the rest of compounds stay relatively stable. HMX, PETN, and TNT are compounds with higher molecular weights and larger cross sections that may enhance the interac tion between these ions and curtain gas. Consequently, these compounds may have a reduced number of ions entering the FAIMS cell when higher curtai n flow rate is applied. Considering the effect caused by curtain flow rate on the CV value, peak width, and signal intensity, the optimum perf ormance of these ions was achieved at a gas flow rate of 2.0 L/min to 2.5 L/min. Effects of DV The separation of ions in FAIMS is based on th e change in mobility of an ion in strong electric fields. Three trends in ion mobility have been reported that show that as electric field

PAGE 79

79 strength (dispersion voltage) increases the mobility of a type A ion increases, a type C ion decreases, and a type B ion incr eases initially before decreasing. 51 The cylindrical geometry of the FAIMS cell used in this research has been proven to have the ability to focus the ions as they are transmitted.53, 96 The apparently an omalous increase of sensitivity with increasing applied asymmetric wa veform voltage, and the be havior of ions with the change of polarity of the waveform, to th e conclusion that the device was focusing ions.52 Therefore, the magnitude and polarity of the di spersion voltage were bot h evaluated in this research to explore their effects on explosives separation. The wavefo rm with negative DV yields spectra of type N1 for negative ions, wh ereas the reversed polarity waveform yields N2 type spectra for negative ions. In general, low mass ions ( m/z is usually below 300) are type A ions and are detected in N1 mode whereas larger ions are type C ions and are detected in N2 mode. The compounds studied in the rese arch produce ions of all three ions. The DV is the maximum peak of the voltage ap plied and was varied in these studies from V to V in 500 V increments. Figure 3-7 and 3-8 illustrate the effect of increasing DV on CV and mass spectra for TNT. As the magnitude of DV increases, two trends are apparent: first, the peak shifts to more positive CV values; an d second, an increase in signal intensity for the selected TNT ion ( m/z 227). Additionally, the separation of selected ions was improved as the DV increases, generating more sp ecific ion patterns and less background in the mass spectra. Similar phenomenon can be also observ ed for the other explosives, as detailed in the following sections. CV value Table 3-1 and Figure 3-9 present the values and plot of CV as a function of DV. Generally for most ions, the magnitude of the CV increas es as the DV increases in both DV polarities; however, the C type ions, such as the Cladducts of PETN and HMX, show a greater CV shift in

PAGE 80

80 N2 mode. Explosive ions which appear at posit ive CV during application of negative DV can be also seen at negative CV when the polarity of the applied DV is reversed. However, as will be discussed below, the signal intensities for t hose ions under reversed DV decreased since the focusing action within the analyzer region was reversed to a defocusing action. A greater CV shift was observe d at higher fields because th e ions can vary conformation between low and high fields when the field increa ses and, hence, more energy is imparted on the explosive ions.102 The more the conformations vary betw een high and low fields, the greater the change in mobilities, and the greater CV. Ion mobility at high-field strength is a result of the interaction between the ion and the bath gas,36 which is strongly determined by ion size, shap e, rigidity, and properties of the bath gas.103 Lighter explosive ions can be obser ved at higher CV values than heavier ions because ions with smaller size normally have a greater percentage change in cross section than larger ions. However, not all ions possessing same molecular we ight show up at the identical CV value; for instance, the three DNT isomers appear at different CV values because of the variations in how the isomeric ions interact with the bath gas. Signal intensity The spatially inhomogeneous electric field in the cylindrical geom etry FAIMS not only separates ions but also focuses th e ions which are at their correct CV values. Therefore, in most cases, the transmission of ions increases with fi eld strength using cylindrical geometry FAIMS. But focusing is possible only for the ions with a noticeable -dependence in a fairly high field. Only the ions with substantial field dependence of mobility can effectively be focused.97 It has also been reported that increasi ng DV improves the signal intensity for small ions more than for large ones.104

PAGE 81

81 Table 3-1 and Figure 3-10 present signal intens ity as a function of the DV. For type A and most type B ions, including TNT, TNB, Tetryl 1,3-DNB, and DNT isomer s, the largest signal intensities were recorded at maximum DV in N1 mode, with relatively low signal intensity in N2 mode. In contrast, the transmission is maximized with increasing DV in N2 mode for some type B ions and type C ions, including the Cladducts of RDX, HMX, P ETN, and NG. The signal of these ions substantially decreases starting at a DV of -3500 V in N1 mode because the decreased ion mobility of type B ions appears when thos e ions experience higher electric field. This situation mainly depends on a number of factors, namely, an increase in the diffusion coefficient in a strong field, an increase in the amplitude of ion oscillation in th e gap of the separation chamber, and a decrease in the focusing efficiency with a decrease in (E/N).85 The increasing transmission at higher DV for most compounds is due to ion focusing under the effect of the gradient of the alternating field with an unbala nced polarity, which decrea ses ion losses to the walls of the FAIMS analyzer. Peak width The focusing of ions in cylindrical geometry is a principal factor that contributes to the shape of the peaks obtained when sweeping the CV. The fundamental physics responsible for peak shapes has been described in terms of the confining effect of ion focusing between cylindrical FAIMS electrodes and the dispersive effects of diffusion and ion-ion repulsion.104 Guevremont et al.105 also reported that the widths of peaks in FAIMS are a function of the applied DV as well as the radii of the electrode s. The peaks are narrow at low applied DV and largest electrode radii. For a gi ven peak in the CV scan, the lo west CV of ion transmission is characterized by an optimum ion focus point loca ted near the outer electrode. As the CV increases to pull the ion cloud clos er to the inner electrode, this focus point migrates towards the inner electrode. The peak width is determined by the applied CV range, which allows the ion

PAGE 82

82 cloud to migrate through the cell betw een the outer and inner electrodes. Ions are rapidly lost to the walls when the focus location is inside the wa ll of either electrode, less rapidly if the focus location is between the electrodes but is near the wall of either electrode, and minimally if the focus point falls midway between the electrodes. Figure 3-11 shows that type A and most type B ions have broader peaks as DV increases in N1 mode; however, some heav ier ions such as the Cladducts of PETN and HMX have narrower peaks when higher DVs are applied in N2 mode. I ons with larger peak widths due to increased ion focusing are more efficiently transmitte d through the FAIMS device at higher DV values.36 The peak widths increase for type A ions in N1 mode with CV for maximum ion transmission but do not depend on m/z or molecular weight.42 The mass spectra collected from APCI/MS a nd APCI/FAIMS/MS, as seen in Figure 3-12, show that the major ions for both approaches give identical base peaks for most explosives at a concentration of 10 g/mL, except for some differences between the isomers of DNT, which, in turn, may be helpful to discriminate these isomers from each other. Fewer fragment ions, cluster ions, and background ions can be observed in the APCI/FAIMS/MS spectra showing that FAIMS can discriminate against background and ther eby dramatically increase the S/N, reducing or eliminating the need for chromatographic separation. Effects of Carrier Gas Composition A quantitative description of the interaction between ions and the bath gas is a potential well of a given depth in an ener gy diagram at a given temperature.36 The impact of this interaction on the observed CV value depends on the depth of the well in comparison to the thermal energy of the bath gas. The deeper the well relative to the thermal energy, the more mobility will increase with E/N where N is the nu mber density of the bath gas. The mobility increases because of energy gains du ring collisions. If the thermal energy is similar to or greater

PAGE 83

83 than the potential well, the ion mobility can increa se less rapidly or will actually decrease with E/N. This change in ion mobility results from energy losses during collision between the gas and analyte ions.36, 47 Ion mobility at high-field strength is a result of the interaction between the ion and the bath gas.36 To date, most FAIMS work has employed N2 (or air) or a He/N2 mixture;106-108 O2, CO2, N2/CO2, He/SF6, and other compositions were also explored.38, 47, 109-111 Shvartsburg et al.111 reported spectacular non-Blanc effects in mi xture of disparate gases such as He/CO2 or He/SF6 and described the solution as 1/Kmix(E) = xjRj/Kj(E) (3-1) where K mix is the mobility of an ion in a mixture of any number of gases, and each coefficient Rj satisfies RjKj[ ( (m+Mj) wiRi)-1/2] = Kj(E) E Kj(E) xiRiKi(E) (3-2) where E is the strength of electric field, m and Mj are the molecular masses of the ion and the i -th component of gas mixture, and wi terms are given by wi = xi/[(m+Mi)Ki(E)] (3-3) The basis of this effect is related to the wi dely differing molecular masses of these gases, and the large difference in the mobility of an ion in each of the pure gases that compose these mixtures. Resolution and sensitivity of FAIMS us ing binary and ternary mi xtures is often better than that with any individual component because of non-Blanc behavior of ion mobilities at high electric fields.38, 110, 111

PAGE 84

84 In this work, N2, He, CO2, SF6, O2, and the individual gas mixtures were used to evaluate the influence made by different carrier gases. TNT in different carrier gas compositions The effect of different carri er gas composition on compensa tion voltage, peak width, and signal intensity for TNT was studied. The pure N2, O2, He, CO2, or SF6 were evaluated in this research. The mixture of O2, He, CO2, or SF6 with N2 were also studied and their content in N2 carrier gas varied from 0% to 50 %. DV was maintained at -4000V and the gas flow rate was set at 2 L/min. Figure 3-13 shows the TIC-CV spectra of TNT collected in different carrier gas compositions. The results demonstrate TNT in pure N2 has the highest transmission, but poorer resolution. Although the use of pure O2 and the N2/He and N2/O2 mixtures narrows the peak width, it also decreases the signal intensity. TNT e xperiences the greatest change in mobility in the carrier gas mixtures of N2/CO2 and N2/SF6, but wider peak width and lower intensity can be also seen in the spectra. Gas types that may lead to reduced peak wi dth and increasing signal intensity are important considerations for FAIMS; therefore, further st udy of the effect of N2, O2, and mixtures of N2/He and N2/O2 was performed in the following research. TNT in O2 and mixture of N2/O2 In pure O2 or gas that includes O2 content, O2-, which possesses strong gas-phase basicity, is generated as a reactant ion and easily abst racts protons from TNT to produce an [M-H]ion at m/z 226 instead of m/z 227. The [M-H]ion was monitored in the experiment performed in O2 and mixture of N2/O2. IS-CV spectra acquired for TNT in a carrier gas of pure O2 at different DVs are shown in Figure 3-14. At DVs below -4000 V, the signal for TNT was very low due to high diffusional loss to the cylindrical electrodes. However, the observed 10-fold increase in sensitivity at DV of

PAGE 85

85 -4000 V was the result of an ion focusing mechanism inherent to a cylindrical geometry FAIMS cell.53 Comparing to Figure 3-7 acquired for TN T in pure nitrogen, the transmission of TNT observed in oxygen was roughly 60% lower than in nitrogen at DV of -4000 V. The CV value, peak width, and signal inte nsity for TNT are given in Figure 3-15, which shows plots collected at DVs fr om -2500 V to -4500 V in carrier gas compositions from 0% to 50% oxygen in nitrogen. In a mixture of N2/O2, the CV value was reported to comply with Blancs law, which relates the mobility of an ion in a mixture of any number of gases ( Kmix) with abundance xi to its mobilities in individual constituents ( Ki), as shown in equation 3-4.38 1/Kmix = xi/Ki (3-4) In addition, most ions have similar mobilities in N2 and O2, as one might expect from similar molecular mass, size, and other properties of these two gases.111 For each DV, the difference between the maximal and minimal CV value for a mixture of N2/O2 is no greater than 0.5 V. The average difference of CV values is 0.367 V under the influence of all DV values, which indicates the addition of O2 to N2 had little effect on the CV value. However, decreased peak width can be seen as the O2 fraction changes from 0 to 50 %. The signal intensity collected in the gas mixture with the O2 fraction above 20 % starts to drop when the DVs ramps up to 4000 V, which suggests that the [M-H]ion of TNT is on its way to converting into a C-type ion in gas mixtures with higher O2 content. Explosives in mixture of N2/He He atoms are smaller and less polarizable than N2 molecules; therefore, an ion experiences fewer collisions or interactions with th e He atoms compared to the larger N2 molecule. With minimal interaction with the carri er gas, the analyte ion may have more flexibility to alter conformations at high and low field,102 which increases the change in mobility between high and

PAGE 86

86 low field. Addition of He to N2 has been shown to increase re solution (decreased peak width), sensitivity (signal intensity), and peak capacity (higher CV value) of FAIMS measurements.47, 108, 112 Shvartsburg et al.111 also verified that mixtures of grossly disparate gases, such as N2/He, can exhibit large deviations from Bl ancs law that greatly expand the separation space. Because high He content in the FAIMS cell may cause electrica l breakdown in either the mass spectrometer or the FAIMS cell due to the insufficient pumping capacity from the mass spectrometer, the He content in N2 was varied from 0% and capped at 50% DV was varied from V to V in 500 V increments. Figures 3-16, 3-17, and 3-18 show the effects in CV, signal intensit y, and peak width of explosive compounds from increasing the He content from 0% to 50 %. In Figure 3-16, most of the type A and type C ions experience increased CV in N1 and N2 mode, respectively, as the He content is increased. This magnitude of deviatio ns is because of a large difference between the molecular masses of He and N2 and between ion mobilities in these gases. He is lighter, smaller, and less polarizable than N2; therefore, the chance or interacti on with the analyte ion is reduced. With fewer interactions with the bath gas, the analyte ion may have more flexibility to alter conformations at high and low fi eld, resulting in a higher CV to compensate the substantial change. However, for type B ions such as TNT, tetryl, RDX, and NG, the CV decreases in N1 mode and decreases initially before increasing at higher He content in N2 mode. It is notable that a gas composition of greater than 30% He causes a stable or decreasing CV for type A ions in N1 mode. The trends observed in this experi ment are in compliance with the previous report that states that type C ions present even stronger type C ion beha vior in He and, in N2, some ions of type B and A switch to type C ion in N2/He mixtures with a larger He content.107 This may be attributed to low He polarizabil ity, making long-range, attractive interactions (that induce the A-

PAGE 87

87 type behavior) less important than short-range repulsive forces.111, 113 This result may also verify the hypothesis from IMS studies: all ions, except the lightest ones, ar e type C ions in He at any electric field.114, 115 In Figure 3-17, the signal intensity for explos ives generally decreases as the He content increases for some type A ions and type B ions in N1 mode. However, the signal intensity falls off for lighter ions such as DNT and DNB at a gas composition of greater than 30% helium. Higher intensities were seen for type B and t ype C ion when adding more He in N2 mode, showing an even stronger type C i on behavior. It has been reported that there is a boost in signal intensity for type C ions with increasing helium content of up to about 60% helium.111 This phenomenon may be ascribed to fewer interacti ons occurred between analytes and carrier gas with increased He content, which may allow improved focusing.102 In Figure 3-18, the peak width for explosives generally decreases as the He content increases for type A ions and type B ions in N1 mode, and increases for type C ions in N2 mode. This trend shows that the peak s tend to narrow with decreased magnitude of CV and signal intensity. The analyte ions are transmitted over a much narrower range of CV at a given DV in a carrier gas with larger He propor tion, which may be attributed to the less focusing for type A and type B ions in N1 mode. In c ontrast, the reason for the broadened peaks for type B ions in higher He content and type C ion in N2 mode is the enhanced focusing in the field. In order to verify the conversion from type B ion to type C ion caused by He addition, the CV value, peak width, and signal intensity of TNT ([M]-) and tetryl ([M-NO2]-) ions were measured over a range of DV from 2500 to 4000 V and at 0% to 50% He. For both explosive compounds, similar trends were observed as a f unction of both DV and He content: increased CV value and signal intensity, and broadened peak width in N2 mode, which is pointed out by a

PAGE 88

88 red circle in Figure 3-19 and Fi gure 3-20. This phenomenon is typical type C ion behavior, which means these type B ions ar e switching to type C ions at both higher DV and He content. Effects of Electrode Temperature One of the benefits of increas ing the FAIMS cells temperature is the elimination of the residual water and contaminants from the cell, which results in an increase of sensitivity52 Separately controlling the temper ature on the inner and outer electr odes is shown to be extremely useful for manipulating the sele ctivity of FAIMS. The optimal temperature setting depends on the shape of the CV/DV plots; th erefore, the optimal temperature setting needs to be determined experimentally.116 The effect of temperature on ion sepa ration can be described by equations 3-5 to 3-7.116 The mobility at high field for a given ion can be described by: Kh(E/N) = K[1+ f (E/N)] (3-5) where K is the mobility constant, E is the electric field, and N is the gas number density. The number density can be calculated by: N=(n/V) NA (3-6) where NA is Avogadros constant and (n/V) is determined from the ideal gas law: n/V = P/(RT) (3-7) Here, R is 0.082 L atm / mol K, and P and T ar e parameters that are measured during the experiment. Theoretically, the increased temperature can lead to the increased eff ect of electric field, thus increasing ion focusing between the electrodes. In this res earch, different inner and outer

PAGE 89

89 electrode temperatures were mani pulated to investigate the impr ovement in sensitivity and the alteration of ion mobilities for explosive compounds. An increased electric field (E/N) can be e xpected as temperature increases because N decreases with increasing temperature. Figure 3-21 illustrates the calculated electric field in FAIMS cell at different DVs and te mperatures, and indicates that th e electric field generated at a DV of 4500 V and an electr ode temperature of 70 C is greater than it is at a DV of 5000 V and electrode temperature of 30 C. Figure 3-22 presents plots of CV value, peak width, and signal intensity for the ions of TNT and 2,6-DNT as function of el ectrode temperature. As the temperature increases, decreased CV value, peak width, and signal intensity were observed. The situation for both explosives has similarities to what was desc ribed at high applied DV for a t ype B ion, which shows decreased ion mobility at high electric field. Compared to Figure 3-9 to 3-11, th e conversion from type A to type C behavior in this experi ment occurs at a lower electric fiel d than what is calculated from the equations mentioned above. This may be a ttributed to effect of increased temperature isomerization, dissociation,117 or the change in folding structur e other than a simple decrease in number density. Independent control of the inner and outer electr ode temperatures offers extra flexibility to control the CV, peak width, separation, and sensitivity for different analytes.116 In this research, the waveform is only applied to the inner electrode; therefore, the electric field between two concentric cylindrical electrodes is non-uniform. Ions near the inner electrode are subjected to stronger fields than those near th e outer electrode. If the temp erature of either electrode is increased, the applied fields near the heated el ectrode will increase because of decreased number density. Therefore, when a higher temperature is applied to the inne r electrode or a lower

PAGE 90

90 temperature is applied to outer el ectrode, the electric field gradie nt is steepened to enhance ion focusing, as shown in Figure 3-23. In contrast, the electric fiel d generated at a higher outer electrode temperature and a lower inner electrode temperature is less steep and provides less ion focusing. Figure 3-24 and 3-25 demonstrate the effect of temperature gradient on CV value, peak width, and signal intensity. The expected focusi ng was not observed at a higher inner electrode temperature or a lower outer electrode temperature. When the temperature of the inner electrode was increased, the CV for optimal transmission shif ted to a smaller value, and a narrower peak and lower signal intensity were observed. On the other hand, if the temperature of outer electrode was increased, a smaller CV value, a br oader peak, and an incr eased signal intensity were observed. The magnitude of the changing CV value, peak width, and signal intensity for the outer electrode temperature is smaller than that of the inner electrode, which means greater effect was given by the alteration in temperature of the inner electrode than the outer electrode. The results are unexpected because the M]ions of both TNT and 2,6-DNT are type B ions and the differential ion mobility (Kh/K0) starts to decrease when the el ectrode temperature is raised pass the turning point, as was shown in Figure 1-3 for type B ion. Because the electric field near the inner electrode increases with the elevated inner electrode temper ature, it is expected that the Kh/Ko of TNT and 2,4-DNT ions near the inner el ectrode decreases, resulting in the lower focusing strength toward the center of the FAIM S cell and, hence, losing more ions due to diffusion and space-charge repulsion. However, if the temperature of outer electrode is increased, the electric field may focus or defo cus the ion depending on whether the magnitude of electric field reaches the turning point where th e ions possess the largest gap in mobility between low and high electric fields. Th erefore, the plots of peak width and signal intensity may fluctuate

PAGE 91

91 when varied outer electrode temperatures are ap plied. This may also s uggest that changing the outer electrode temperature provides less alteration in CV value, peak width, and signal intensity. Conclusions FAIMS is a promising technology that func tions well as a separation device and is compatible with mass spectrometry. This is the first systematic evaluation of the effect of factors such as DV, CV scan rate, curtain gas flow rate, carrier gas co mposition, and electrode temperature for the analysis of explosive compounds. Experiments showed that a CV scan rate of 10 V/s and curtain gas flow rate from 2.0-2.5 L/min were the optimal conditions for both reso lution and transmission. An increase in CV value and sensitivity was observed at higher a DV for both type A and type B ions in N1 and type C ions in N2 modes. Although the ion focu sing mechanism in the cylindrical cell improves the sensitivity, it also decreases the resolution. Furthermore, th e separation and sensitivity are also influenced in FAIMS by changing the carr ier gas composition. The mixture of helium and nitrogen was shown to provide benefits to achie ve better resolution and sensitivity. For type A ions, the peak width was reduced and the CV wa s shifted to more positive value in N1 mode with increasing helium content in carrier gas, which improved the resolution between those ions. In addition, dramatic increases in sensitivity and peak width for type C ions in N2 mode were obtained when the content of helium in carrier gas was increased; however, a broader peak width and increased signal intensity were also seen fo r type B ions in the same mode, which implied that the type B ion was transformed to a type C ion at higher DV and helium content. Finally, an alteration may also be observed on CV value, peak width, and signal intensity with varied electrode temperatures. Two type B ions, the M]ions of TNT and 2,6-DNT, have decreased

PAGE 92

92 mobilities with increased temperature. That is also the reason that the magnitude of focusing strength decreases when the temperat ure of inner elec trode was raised. Based on the understanding gained here, opt imal separation or transmission can be achieved by controlling different parameters. Th is knowledge will also be beneficial for the further development of devices for explosiv es detection based on the FAIMS technique.

PAGE 93

93 0.8mm 8mm 6mm Heating Capillary Mass Spectrometer 8mm 7mm 7mm 1.4mm9.6mm 5mm 5mm 12mm 8mm 9.6mm 65mm 4mm 2mm 1.6mm ID=0.76 mm OD=1.6 mm 15mm 1mm 5mm Capillary O-ringID= 8mm OD= 11.1mm Wide=1.6mm 3.5mm 3.5mm 19mm FAIMS Figure 3-1. The design of the brass capillary extender. Figure 3-2. The actual picture of the brass capillary extender.

PAGE 94

94 Scan Rate (TNT at -4000 Volts)0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 2.55.010.015.020.0 Scan Rate (Volts/min)Volts0.0 2.0 4.0 6.0 8.0 10.0 12.0Signal Intensity (10 4 counts) CV(Volts) Peak Width (Volts) Intensity of m/z 227(Counts) Figure 3-3. Effect of CV scan rate on CV va lue, peak intensity, and peak width. (DV= 4000V) CV value vs carrier gas flow rate 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 22.533.5 Carrier Gas Flow Rate(L/min)CV(Volts) TNT TNB 2,4-DNT 2,6-DNT 3,4-DNT Tetryl(-NO2) RDX(+Cl) HMX(+Cl) PETN(+Cl) NG(+Cl) Figure 3-4. Effect of curtain ga s flow rate on CV for the ions of tested explosives. (DV= 4000V)

PAGE 95

95 Peak width vs carrier gas flow rate 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 22.533.5 Carrier Gas Flow Rate(L/min)Peak Width(Volts) TNT TNB 2,4-DNT 2,6-DNT 3,4-DNT 1,3-DNB RDX(+Cl) HMX(+Cl) PETN(+Cl) NG(+Cl) Figure 3-5. Effect of curtain gas flow rate on pe ak width for the ions of tested explosives. Intensity vs carrier gas flow rate 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 22.533.5 Carrier Gas Flow Rate(L/min)Signal intensity(10 4 counts) TNT TNB 2,4-DNT 2,6-DNT 3,4-DNT Tetryl(-NO2) RDX(+Cl) HMX(+Cl) PETN(+Cl) NG(+Cl) Figure 3-6. Effect of curtain gas fl ow rate on signal intensity for the ions of tested explosives.

PAGE 96

96 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2 Time (min) 0 2 4 6 8 10 12 14 16 18 20 22 24 -5000V -4000V -3500V -3000V -2500V -4500VTNT -5 0 5 10 1 5 CV(volts)Signal Intensity ( 105counts) Figure 3-7. SI-CV spectra for the [M]ion ( m/z 227) of TNT: variation of the DV. 100 150 200 250 300 350 400 450 500 m/z 0 50 100 0 50 100 0 50 100 0 50 100 Relative Abundance 0 50 100 0 50 100 227.07 197.27 254.20 462.80 167.13 66.87 227.07 197.27 243.20 167.13 53.80 415.13 227.13 197.13 167.00257.67 73.80 482.93 291.87 351.93 136.13 227.00 197.13 243.00 276.87 167.20 109.53 334.20 408.53 227.00 197.27 291.13 167.00 480.33 100.00 457.07 312.47 361.40 227.07 277.20 197.20 291.13 318.73 244.20 152.13 373.40 420.87 482.40-5000V NL:1.82E4-2500V NL:6.18E3-3000V NL:1.19E4 -3500V NL:1.57E4 -4000V NL:1.80E4 -4500V NL:1.88E4 Figure 3-8. Mass spectra for the TNT: variation of the DV.

PAGE 97

97 CV value of explosives -15.00 -10.00 -5.00 0.00 5.00 10.00 15.00 -4.5-4-3.5-3-2.502.533.544.5 DV(kV)CV(Volts) TNT TNB 2,4-DNT 2,6-DNT 3,4-DNT 1,3-DNB Tetryl(-NO2) RDX(+Cl) HMX(+Cl) PETN(+Cl) NG(+Cl) Figure 3-9. Graph of CV versus DV fo r the ions of tested explosives. Intensity of major ion of explosives 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 -4.5-4-3.5-3-2.502.533.544.5 DV(kV)Signal intensity( 4 counts) TNT TNB 2,4-DNT 2,6-DNT 3,4-DNT 1,3-DNB Tetryl(-NO2) RDX(+Cl) HMX(+Cl) PETN(+Cl) NG(+Cl) Figure 3-10. Graph of signal intensity versus DV for the ions of tested explosives.

PAGE 98

98 Peak width of CV scan for explosives 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 -4.5-4-3.5-3-2.502.533.544.5 DV(kV)Peak Width(Volts ) TNT TNB 2,4-DNT 2,6-DNT 3,4-DNT 1,3-DNB Tetryl(-NO2) RDX(+Cl) HMX(+Cl) PETN(+Cl) NG(+Cl) Figure 3-11. Graph of peak width versus DV for the ions of tested explosives.

PAGE 99

99 Table 3-1. The main analytical characteri stics of FAIMS on detecting explosives. DV(kV)-4.5-4-3.5-3-2.52.533.544.5 CV(Volts)5.74.32.81.00.9-1.0-1.6-2.6-3.6-5.5 Peak Width(Volts)1.31.31.11.01.11.21.41.61.71.8 Intensity of m/z 227(counts ) 1.58E+051.37E+051.08E+056.10E+045.85E+041.18E+041.05E+045.65E+035.13E+033.65E+03 CV(Volts)7.35.53.92.51.3-1.3-2.1-3.4-5.2-6.8 Peak Width(Volts)1.21.21.11.01.01.11.21.41.30.9 Intensity of m/z 213(counts ) 1.60E+051.37E+059.92E+044.09E+041.68E+045.31E+035.09E+032.20E+031.98E+031.44E+03 CV(Volts)9.67.35.33.21.7-1.6-2.5-4.3-6.6-9.2 Peak Width(Volts)2.11.81.41.31.31.82.12.02.31.7 Intensity of m/z 181(counts ) 8.25E+047.64E+045.45E+042.74E+041.80E+045.74E+033.18E+032.08E+031.08E+037.15E+02 CV(Volts)10.37.75.33.31.7-1.8-2.9-4.6-7.2-9.5 Peak Width(Volts)1.91.81.51.21.21.31.42.12.01.9 Intensity of m/z 182(counts ) 1.08E+058.67E+045.34E+042.90E+041.34E+048.21E+035.89E+031.93E+038.35E+028.85E+02 CV(Volts)8.05.83.82.31.3-1.4-1.9-3.6-5.4-7.4 Peak Width(Volts)1.61.51.31.41.41.41.52.11.82.4 Intensity of m/z 182(counts ) 7.41E+045.93E+044.20E+041.96E+041.11E+048.06E+034.35E+032.79E+031.50E+035.67E+02 CV(Volts)10.67.95.63.42.0-1.9-2.8-4.8-7.4-10.0 Peak Width(Volts)1.81.71.41.31.21.01.01.01.31.7 Intensity of m/z 168(counts ) 9.16E+046.88E+044.81E+042.63E+041.41E+045.85E+032.87E+032.44E+031.04E+038.31E+02 CV(Volts)2.72.21.50.90.5-0.5-0.8-1.2-1.8-2.5 Peak Width(Volts)1.31.41.21.21.11.41.41.51.51.4 Intensity of m/z 241(counts ) 1.06E+059.51E+048.25E+044.97E+043.03E+043.58E+043.62E+043.68E+043.56E+044.18E+04 CV(Volts)2.52.22.01.41.00.40.40.40.30.2 Peak Width(Volts)3.23.53.13.33.6 Intensity of m/z 257(counts ) 1.45E+051.46E+051.59E+051.22E+059.12E+043.32E+043.54E+043.89E+045.40E+047.86E+04 CV(Volts)-0.4-0.40.10.70.70.50.71.11.12.3 Peak Width(Volts)3.33.93.13.33.13.12.52.42.12.2 Intensity of m/z 331(counts ) 1.03E+051.35E+051.70E+051.62E+051.41E+051.61E+051.80E+051.92E+052.20E+052.83E+05 CV(Volts)0.00.30.50.50.40.81.11.11.52.0 Peak Width(Volts)3.33.12.92.73.13.13.12.82.82.5 Intensity of m/z 351(counts ) 1.75E+052.04E+052.21E+051.91E+051.48E+056.84E+041.84E+052.43E+052.67E+053.13E+05 CV(Volts)2.21.91.51.40.80.50.10.30.30.3 Peak Width(Volts)3.12.72.72.92.91.71.61.91.71.6 Intensity of m/z 262(counts ) 1.25E+051.75E+052.08E+051.76E+051.46E+052.44E+041.09E+051.33E+051.64E+051.85E+05 TNT TNB 2,4-DNT 2,6-DNT HMX(+Cl) PETN(+Cl) NG(+Cl) 3,4-DNT 1,3-DNB Tetryl(-NO2) RDX(+Cl)

PAGE 100

100 0 50 100 0 50 100 0 50 100 0 50 100 0 50 100 227.07 197.20 260.00 213.13 239.07 183.27 258.80 182.07 165.13 212.93 182.07 212.93 152.20 75.07 182.07 152.27 100 150 200 250 300 350 400 450 50 0 m/z 0 50 100 0 50 100 0 50 100 0 50 100 0 50 100 241.13 313.00 256.93 329.00 225.07 256.87 478.53 258.87 209.87 330.93 291.67 66.80 331.00 332.93 62.00 283.87 357.93 98.47 167.80 350.80 352.80 62.13 306.00 82.07 238.87 261.80 488.40 263.73 62.13 443.40 217.00 398.53 86.13 100 150 200 250 300 350 400 450 50 0 m/z 0 50 100 0 50 100 0 50 100 0 50 100 0 50 100 241.07 256.93 256.78 259.09 330.91 333.08 128.96 314.74 202.88284.22 350.93 352.75 315.09 62.04 355.13 256.85 261.96 263.99 62.18 A G F E C D B J I H APCI-MSAPCI-FAIMS-MS 0 50 100 0 50 100 0 50 100 0 50 100 0 50 100 227.07 210.07 167.00 213.13 183.33 181.20 89.07 182.00 152.13 182.00 168.07 138.27 271.53 212.00 Figure 3-12. Mass spectra of explosives acqui red by APCI-MS and APCI-FAIMS-MS: (A) TNT, (B) TNB, (C) 2,4-DNT, (D) 2,6-DNT, (E) 3,4DNT, (F) Tetryl, (G) RDX, (H) HMX, (I) PETN, (J) NG.

PAGE 101

101 1.0 1.5 2.0 2.5 3.0 3.5 Time (min) 0 50 100 0 50 100 0 50 100 0 50 100 Relative Abundance 0 50 100 0 50 100 N2m/z 227 NL:1.85E5 O2m/z 226 NL:1.17E4 60% N2/ 40% O2m/z 226 NL:3.50E4 60% N2/ 40% He m/z 227 NL:4.02E4 60% N2/ 40% CO2m/z 227 NL:2.32E4 60% N2/ 40% SF6m/z 227 NL:2.21E4 -5 0 5 10 CV (Volts) Figure 3-13. TIC-CV spectra for TNT in diffe rent carrier gas composition at DV of -4000V.

PAGE 102

102 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 Time (min) 0 10 20 30 40 50 60 70 80 90 100 -4500V -3000V -2500V -4000V -3500V TNT in Oxygen -5 0 5 10 15 CV(volts) Figure 3-14. SI-CV spectra for the [M-H]ion of TNT ( m/z 226) in oxygen carrier gas at DV from 2500 to 4500 V in 500 V increments.

PAGE 103

103 Intensity vs DV in N2/O2 mixture (TNT) 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 -2500-3000-3500-4000-4500-5000 DV(Volts)Signal Intensity (10 4 counts) 0% 10% 20% 30% 40% 50% CV vs DV in N2/O2 mixture (TNT) 0 1 2 3 4 5 6 7 -2500-3000-3500-4000-4500-5000 DV (Volts)CV (Volts) 0% 10% 20% 30% 40% 50% Peak Width vs DV in N2/O2 mixture (TNT) 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 -2500-3000-3500-4000-4500-5000 DV(Volts)Peak Width (Volts) 0% 10% 20% 30% 40% 50%C A B% O2% O2% O2 Figure 3-15. Graph of (A) CV, (B ) peak width, and (C) signal in tensity versus DV for the [M-H]ion of TNT in N2/O2 mixtures from 0% to 50% O2.

PAGE 104

104 CV vs Carrier gas composition -6.0 -4.0 -2.0 0.0 2.0 4.0 6.0 8.0 10.0 12.0 50%40%30%20%10%0%0%10%20%30%40%50% %He (V/V) He/N2CV (Volts ) TNT TNB 2,4-DNT 2,6-DNT 3,4-DNT 1,3-DNB Tetryl(-NO2) RDX(+Cl) HMX(+Cl) PETN(+Cl) NG(+Cl)DV -4000VDV 4000V Figure 3-16. Graph of CV versus carrier gas com position for the ions of tested explosives in N2/He mixtures. Intensity vs Carrier gas composition 0.0 20.0 40.0 60.0 80.0 100.0 120.0 140.0 160.0 180.0 50%40%30%20%10%0%0%10%20%30%40%50% %He (V/V) He/N2Signal intensity (10 4 counts) TNT TNB 2,4-DNT 2,6-DNT 3,4-DNT 1,3-DNB Tetryl(-NO2) RDX(+Cl) HMX(+Cl) PETN(+Cl) NG(+Cl)DV -4000VDV 4000V 0.0 5.0 10.0 15.0 20.0 25.0 50%40%30%20%10%0% %He (V/V) He/N2Signal intensity (104 counts ) Figure 3-17. Graphs of signal in tensity versus carrier gas com position for the ions of tested explosives in N2/He mixtures.

PAGE 105

105 %He (V/V) He/N2 DV -4000VDV 4000V Figure 3-18. Graphs of peak width versus ca rrier gas composition for the ions of tested explosives in N2/He mixtures.

PAGE 106

106 Intensity vs DV at varied carrier gas composition TNT(m/z 227 [M]-) 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 -4.5-4-3.5-3-2.502.533.544.5 DV(kV)Signal intensity (10 4 counts ) 0% 10% 20% 30% 40% 50% C A B Figure 3-19. Graph of (A) CV, (B ) peak width, and (C) signal in tensity versus DV for the [M]ions of TNT in N2/He mixture. Red circle shows that TNT presents an even stronger type C ion behavior in high helium content.

PAGE 107

107 Intensity vs DV at varied carrier gas composition Tetryl(m/z 241 [M-NO2]-) 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 -4.5-4-3.5-3-2.502.533.544.5 DV(kV)Signal intensity (10 4 counts) 0% 10% 20% 30% 40% C A B Figure 3-20. Graph of (A) CV, (B ) peak width, and (C) signal in tensity versus DV for the [MNO2]ions of tetryl in N2/He mixture. Red circle shows that Tetryl presents an even stronger type C ion behavior in high helium content.

PAGE 108

108 50 60 70 80 90 100 110 6.577.588.59Electric field (E/N,Td) Electric field vs Radial position in cell 30 -4500V 60 -4500V 70 -4500V 90 -4500V 30 -5000VInner cylinder Outer cylinder Radial distance(mm) Figure 3-21. Calculated electric field as a function of radial distance between cylindrical FAIMS inner/outer cylinders at di fferent temperature and DV.

PAGE 109

109 Signal intensity vs electrode temperature 0.0 2.0 4.0 6.0 8.0 10.0 4050607080 Outer electrode temperature()Signal intensity( 4 counts) TNT 2,6-DNT A C B Figure 3-22. Graph of (A) CV, (B ) peak width, and (C) signal in tensity versus cell temperature for the [M]ions of TNT and 2,6-DNT.

PAGE 110

110 50 60 70 80 90 100 110 6.577.588.59Electric field(E/N, Td)Radial distance(mm) Electric field vs Radial position in cell I40/O40 I40/O90 I90/O40 I90/O90 I40-PlanarInner cylinder Outer cylinder Figure 3-23. Calculated electric field as a function of radial di stance between cylindrical FAIMS inner/outer cyli nders at DV of 4500 V. (I: inner electrod e temperature, O: outer electrode temperature, Pl anar: planar FAIMS cell)

PAGE 111

111 Signal intensity vs electrode temperature 0.0 2.0 4.0 6.0 8.0 10.0 405060708090 Outer electrode temperature()Signal intensity(10 4 counts) 40 50 60 70 80 A C BInner temperature Inner temperature Inner temperature Figure 3-24. Graph of (A) CV, (B ) peak width, and (C) signal in tensity versus inner and outer electrode temperatures ( ) for the [M]ions of TNT.

PAGE 112

112 Signal Intensity vs electrode temperature 0.0 2.0 4.0 6.0 8.0 10.0 405060708090 Outer electrode temperature()Signal intensity(10 4 counts) 40 50 60 70 80 A C BInner temperature Inner temperature Inner temperature Figure 3-25. Graph of (A) CV, (B ) peak width, and (C) signal in tensity versus inner and outer electrode temperatures ( ) for the [M]ions of 2,6-DNT.

PAGE 113

113 CHAPTER 4 PERFORMANCE OF APCI-FAIMS-MS FOR ANALYSIS OF EXPLOSIVES Introduction FAIMS is able to separate analyte ions from chemical background noise and enhance analyte sensitivity.112 The previous chapter described th e characterization of APCI-FAIMS-MS for nitrate ester, nitramine and nitroaromatic compounds. This chapter describes the validation of this method, was carried out in this resear ch, evaluating the repeatability of CV values, resolving power (RP), resolution (RS), linear dynamic range (LDR ), and limit of detection (LOD). Experimental In this research, experiments were perfor med employing a FAIMS-MS system, comprising a cylindrical FAIMS device (Thermo Scientific, San Jose, CA) and a commercial ion trap mass spectrometer (LCQ, Thermo Scientific). Ga s-phase explosive ions were generated by atmospheric pressure chemical ionization (APCI) using a corona discharge needle that is positioned at an angle of 45 and ~1 cm from th e opening in the curtain plate of FAIMS device. The cylindrical FAIMS cell consists of two el ectrodes, inner and outer electrodes. The combination of inner electrode having an outer radius of 6.5 mm and out er electrode having an inner radius of 9.0 mm makes a gap of 2.5mm for ion transmission. The asymmetric waveform (750 kHz) and the DC compensation voltage (CV) we re both applied to the inner electrode of the FAIMS. The dispersion voltage ( DV) was set in the range between 2500 and 4500 V for type A and B ions and at the range between +2500 and +4500 V for type C ions. The CV was scanned between -20.0 to 20.0 V at scan rate of 10.0 V/min. A constant DC bias voltage of 25 V was applied to the outer cylinder of th e FAIMS device and to the inlet of the mass spectrometer. In order to connect the Ther mo FAIMS cell, designed for Thermo TSQ mass

PAGE 114

114 spectrometer, onto the LCQ, a brass capilla ry extender (i.d. = 0.76 mm, o.d. = 22 mm) was designed to serve as an interface. The curtain plate was held at 1000V to assist negative ions to transit across the desolvation region. The nitr ogen carrier gas was in troduced into the region between the curtain plate and the orifice into the FAIMS analyzer at flow rate of 2.0 L/min. The inner and outer electrode temperatures were not heated (left at room temperature). For the APCI source, the vaporizer temperat ure was set to 150C. The heated capillary temperature and voltage were set to 130C and 25.0 V, respectively. The discharge current was set at 5 A and the tube lens offset was set to 30.0 V. The sheath gas was set to 20.0 (arbitrary units) and the injection fl ow rate of the analyt e was maintained at 20.0 L/min. Eleven explosive compounds (TNT, TNB, Tetryl, 1,3-DNB, 2,4-DNT 2,6-DNT, 4-DNT, RDX, HMX, NG and PETN) were studied. Th ese explosives were provided by Dr. Jehuda Yinon of the Weizmann Institute of Science, and were obtained from the Analytical Laboratory of the Israeli Police Headquarters. To build up the calibration curv e, standard solutions of each of the explosive compounds were prepared by serial dilution of the stock solutions (in acetonitrile) with 65:35 metha nol/water. The concentrations ranged between 0.001 and 10 g/mL. Five replicate CV scans were collected for each sample at each concentration. The average peak area and relative standard de viation values were acquired and calculated. Results and Discussion Repeatability of CV Values Repeatability is one of the crucial elements fo r an analytical approach, which describes the consistency of the measurement. The nature of the compound and the composition of the carrier gas dictate the combination of DV and CV th at will permit successful transmission of a particular ion through the FAIMS cell. Simila r to retention time (RT) in chromatography approach, a repeatable CV value can be treate d as a judging index from the FAIMS data to

PAGE 115

115 identify the explosive compounds under investigation. Additionall y, the CV value could be set for a known explosive compound for ra pid identification in screening. The repeatability of CV value was evaluated at different DV values with five repetitive injections of a 10 g/mL standard mixture while employing the same conditions over a period of a few hours. Standard deviations (SDs) and rela tive standard deviations (RSDs) were calculated by analysis of variance. As shown in Table 4-1, type A or type B ions gave the better repeatability at DV values with negative (N1 mode) polarity, and, on the contrary, type C ions are considerably more repeatable at DV values with positive polarity (N 2 mode). The better repeatability was obviously acquire d as higher DV applied. With higher DV, the effect of ion focusing increases the ion transmission of targ et ions, resulting in a more intense and symmetrical CV peak which generates more pr ecise and reproducible CV values. For TNT, TNB, DNT, Tetryl, and NG at a DV of 4500 V, the SDs of CV values are distributed from 0.06 V to 0.2 V and the RSDs of CV values range from 1.0% to 5.3%, and the average of SD and RSD are 0.12 V and 2.2%. For RDX, HMX, and PETN at a DV of 4500 V, the SDs of CV values are from 0.02 V to 0.16 V and the RSD of CV values are from 4.4% to 8.1%, and the average of SD and RSD are 0.09 V and 6.9%. In gene ral, the results indicate a high degree of CV values repeatability while utilizing this method. In addition to the DV, the scan rate is the othe r crucial factor affecti ng the repeatability of CV value. Generally speaking, hi gher scan rates can be expected to decrease the repeatability because the chance of missing the most abundant poi nt of the CV peak increases. However, slower scan rates also suffer from the problem of asymmetrical or zigzag peak shapes which worsen the accuracy of CV values. Therefore, mo derate scan rates need to be applied to obtain better repeatability, which was 10 V/min in this research. The concentration of analytes may

PAGE 116

116 also influence the accuracy and repeatability of CV values. Too high an analyte concentration may lead to saturation of the detector, resulti ng in an inaccurate determination of the optimum CV value, while too low an analyte concentration will lead to poor ion stat istics and a noisy peak, leading to an erroneous CV assignment.52 In addition, for some compounds, a high analyte concentration can lead to multimer forma tion, resulting in multiple CVs observed per compound.49 Separation The ions of different types are separated in FAIMS by ion mobility increments that depend on electric field strength.100 Resolution (RS) is an important characte ristic of any analytical method, and accounts for its capacity to separate specific components of a mixture; resolving power (RP) indicates the ability of the analytical method to produce narrow and well-resolved peaks. The RP for FAIMS was calculated from equation 4-1, where the CV is divided by the peak width at full width half maximum. RP = CV/ W1/2max (4-1) The RS between peaks of explosives was calcula ted by Equation 4-2, used to quantify the degree of a two-component separation. Rs= 2 CV (Wb2+Wb1) (4-2) where CV is the difference in compensation values of maximum intensity of the two peaks and Wb1 and Wb2 are the peak width at 10% height fo r the two species, respectively. In chromatography, a condition of an adequate se paration of two peaks is the equality: RS = 1. For a complete separation RS 1.5, where as an RS 0.5 denotes that sepa ration is unavailable.

PAGE 117

117 Resolving power To assess the impact of changing the electr ic field on resolving power of the FAIMS, varying DVs were applied to acquire the RP for individual expl osive compounds. The RP is higher for narrower peaks at CV values away from DV. A high value of RP corresponding to a good separation of peaks is simila r to the convention used with chromatography separations. The RP values for the eleven explosives obtained wi th the FAIMS system are shown in Table 4-2 and indicate that six of the eight type A and type B explosive ions (TNT, TNB, DNT, DNB, tetryl, and NG) yielded the best RP values at DV of -4500 V. The othe r five explosives (some of type B and the type C explosive ions ) gave a higher value of RP at DV of 4500 V. As described in Chapter 3, broader peaks normally can be observed at higher DV, but this is also accompanied the appearance at higher CV values which benefit to increase the RP. Of significant interest in this table is that higher RP of TNB and 2,4-DNT can be seen at DV of 4500 V than at DV of 4500 V due to a narrower peak produced in CV spect ra. However, the peaks for type A or type B ions acquired at DV with pos itive polarity are not recommende d for neither qualitative nor quantitative analysis because of reduced transmission and lower reproducibility of CV value and peak area. From previous studies, some strate gies can be applied to improve RP, which include adding helium into the carrier gas to decrease peak width for type A ions or to increase the CV shift to greater positive values for type C ions, or using less curvature, li ke a planar FAIMS cell, to decrease the focusing effect to generate a narrower peak. Separation and resolution between isomeric explosives The relative percentage composition within the total nitroaromatic component has been shown to be most useful for the characterizati on of explosive samples in addition to the usual preliminary tests.118 Identification of patterns within the nitroaromatic isomeric explosives can

PAGE 118

118 be use to differentiate explos ives from similar batch types.37 Three nitroaromatic isomeric explosives (Figure 4-1), 2,4-di nitrotoluene (2,4-DNT), 2,6-di nitrotoluene (2,6-DNT) and 3,4dinitrotoluene (3,4-DNT) were i nvestigated by FAIMS to evaluate the capability to separate isomeric and related explosives. The capability to separate the isomeric a nd related nitroaromatic compound by FAIMS is shown in Table 4-3 and Figure 4-2 to 4-4. Among the DNT isomers, only 2,6-DNT and 3,4DNT achieved baseline separation at DV of -5000 V in pure nitr ogen carrier gas; however, the pair 2,4-DNT and 3,4-DNT reached adequate separa tion as the content of helium in carrier gas was increased to 20%, as shown in Figure 4-2 and 4-3(B). Meanwhile, the resolution between 2,4-DNT and 2,6-DNT raised to 0.75, which indicat ed that the resoluti on between these two peaks still remained incomplete. The lower p eak intensity of 2,4-DNT re late to the two other isomers may have been caused by lower ionizati on efficiency as shown in Figure 4-2 and 4-4. This further worsened the identification of the le ss intense isomer peak. Fortunately, separation of DNT isomers in FAIMS is significantly orthogonal to MS dimensions. 2,4-DNT and 2,6DNT can be identified by extracting the mass chromatogram by selecti ng the disparate major ions ( m/z 181 and m/z 182) as mentioned in Chapter 3. It ca n be seen in Figure 4-2 that all three isomers were separated from one another and fr om the background signals when selective ions ( m/z 181 and m/z 182) were monitored. TNT and TNB are two major components wh ich are usually accompanied with DNT isomers. As presented in Figure 4-3 and 4-4 and Table 4-3, TNT and TNB were both resolved completely from these DNT isomers with the conten t of helium in carrier gas up to 10% at DV of 5000 V.

PAGE 119

119 Separation and resolution of explosive mixtures Military explosives, in which several explos ives are mixed to create an anticipated lethality, are often found among terrorist attacks and crime s cenes. The explosives most frequently blended in these explosive mixtures are NG, TNT, RDX, and HMX.119 For example, HMX or RDX powders may be mixed (typically by melt-casting) with TNT to form Octol or Cyclotol. The above compositions may describe the majority of the explosive material, but a practical explosive will often in clude small percentages of othe r materials. To examine the FAIMS method for a mixture sample analysis, se veral solution mixtures of explosives were evaluated in this research. Table 4-4 summarizes the reso lution observed experimentally between explosives, which indicated that the mixture of TNT, RDX, and HMX achieved better resolution at DV of 4500 V with carrier gas of 30:70 helium/nitrogen (Figur e 4-5), and the mixture of TNT, PETN, and NG was well separated at DV of 4500 V with nitrogen carrier gas (Figure 46). From previous experimental experience, optimum resolution fo r explosive compounds can be always expected at higher DV or with helium content in the carrier gas. However, for the mixture of TNT, RDX, and HMX, superior resolution wa s acquired at DV of 4500 V instead of 5000 V. This is results from the CV shifting more rapidly than the p eak width decreasing. For the mixture of TNT, PETN, and NG, improved resolu tion was observed when pure nitr ogen carrier gas was applied instead of a gas mixture including helium. That is because the apparent increase of peak width for type C ions, such as PETN, and the lower CV values for type B ions, such as TNT and NG, both decrease the resolution as the He composition in the carrier gas is increased. In addition, the dramatically reduced transmission of type B i ons with increased helium in the carrier gas also obstructs the identification of NG from PETN. In conclusion, the resolution for explosive mixtures needs to be optimized experimentally for most occasions.

PAGE 120

120 Figure 4-7 illustrates an i on-selected CV (IS-CV) spect rum of seven nitroaromatic explosives. The spectrum was collected by setting the DV to 4500 V and scanning the CV from 0 to 20 V while monitoring the m/z values of the most abundant ion (most of them are [M]or [M-H]ion ) for the analytes. The results sh own in Figure 4-7 demonstrate that all nitroaromatic explosives explored in this rese arch can be identified and separated by APCIFAIMS-MS. Although 2,4-DNT, 2,6-DNT and DN B cannot be separated by FAIMS alone, selected ions collected by mass spectrometer can be used to resolve these target compounds from mixture. This method is operated as a two di mension separation, FAIMS and mass spectrometry, which provides two kinds of orthogonal informa tion to strengthen the power of separation. Quantitation Reproducibility Reproducibility of intensity (peak area) is essent ial for achieving reliable quantification. In this study, reproducibility of selected ion peak area for different compounds under varied DV values was assessed by analyzing FAIMS-MS data sets. The RSD for each compound was calculated from five analyses of a 10 g/mL sample. Since replicate runs used the same amount of explosives from the same sample, lower RSD is favorable and indicates better reproducibility of peak area between replicate runs. Table 4-5 shows that the RSD of peak area be tween replicate runs at different DV values for different explosive compounds range d from 0.9% for 3,4-DNT at DV of 4000 V to 33% for 2,6-DNT at DV of 4500 V. Typically, type A or t ype B ions gave better reproducibility with negative DV (N1 mode) polarity, wher eas type C ions are more reproducible at DV with positive polarity (N2 mode). The RSD range fo r type A or type B ions at DV of 4500 V was from 1.0% to 4.7% and for type C ions at DV of 4500 V wa s from 2.1% to 4.2%, indicating that the peak area for each explosive ion in replicate runs wa s similar when higher DV with adequate polarity

PAGE 121

121 was applied. Table 4-5 also denotes that the h eavier explosives, such as RDX, HMX, PETN and NG, usually defined as type B or type C ions, achieve better reproduc ibility at negative polarities. However, similar conclusions cannot be said for type A ions at N2 mode. The reason is that these type B ions are converting from type A behavior to type C behavior in N1 mode as DV increases; however, type A ions will have an evident decrease in transmission in N2 mode. The optimum RSD for these ions in N1 mode was 3.5% at 3500 V for RDX, 3.0% at 3500 V for HMX, 1.7% at 3500 V for PETN, and 2.2% at 4000 V for NG. Limit of detection and linear dynamic range The ability to quantify a trace element or mol ecule using specific analytical methods is often viewed in terms of the lim it of detection (LOD). The LOD is a value, expressed in units of concentration (or amount), that describes the lowest concentr ation level (or amount) of the element that an analyst can determine statistical ly to be different from an analytical blank.120 In this research, the LOD was taken as three time s the standard deviati on of the blank signal, expressed in concentration. Seven explos ive compounds (TNT, TNB, Tetryl, 1,3-DNB, 2,4DNT, 2,6-DNT, and 4-DNT) were studied. Solu tions at concentrations of 0, 1, 10, 50, 100, 250, 500, 1000, and 10000 ng/mL for each of the compounds were made, and the mass spectrometer was scanned in the full scan and selected ion m onitoring (SIM) mode, using characteristic ion or ions for each compound. The CV was scanned from 0 to 20.0 V at scan rate of 10.0 V/min. The LOD was also collected at varied concentrations by setting to the optimum CV for transmission of the nitro aromatic explos ives for the collection time of 1 minute, 30 and 10 seconds. Linear dynamic range (LDR) was evaluated by five repetitive injections of standard compound at eight different levels of concentr ation ranging from 1 ppb to 10ppm. The lower point of the LDR is equal to the limit of quan titation, namely, the concentr ation yielding a signal

PAGE 122

122 10 times the standard deviation of the blank. The higher point refers to th e highest concentration that shows a linear dependence of the intensity on concentration.121 Results for LOD and LDR for explosives eval uated by this method are given in Table 4-6 and 4-7. Based on the data shown in Table 46, the LOD of explosive compounds collected in full scan MS ranged from 1 to 28 ng/mL in concentration and from 51 to 1113 pg in amount. The correlation coefficient values for the regres sion ranged from 0.9624 to 0.9943, demonstrating a moderate LDR for the FAIMS peak area generati on of these explosive compounds. Notably, for most explosives, the curves begin to flatten at the high concentrations presumably because of saturation of the ion source region, a complication in all uses of gas-phase ion chemistry with sources at ambient pressure, resulting in the decrease on the linearity at higher concentration.32 As expected, the LOD decreased significantly wh en scanning in SIM mode, which ranged from 2 to 7 ng/mL in concentration and from 72 to 276 pg in amount. The linear ity for the calibration curves also improved, which was supported by th e correlation coefficient from 0.9847 to 0.9990. Although the full scan gave higher LOD and decreas ed linearity, it is still indispensable for screening tests because more information about molecular and fragment ions will be provided by this detection mode. The result in Table 4-7 show the LOD a nd LDR collected by setting FAIMS at the optimum CV for transmission for th e nitro aromatic explosives with varied detection time. The average LODs in concentration and the correlat ion coefficients are similar (0.996 and 11 ng/mL, respectively) for the detecti on time of 1minute and 30 seconds. However, the amounts of explosives required to be identified within 30 seconds vary from 32 to 343 pg. The amount required for identification in 10 seconds is 95 pg on average; however the LOD in concentration and the linearity of calibration curve have both deteriorated with only a 10-second acquisition.

PAGE 123

123 Therefore, the preferable dete ction time to monitor specific explosive compounds by setting the CV at a fixed value at which maximum transmission of target ions are achieved is no less than 30 seconds. In this research, the LOD was also restrained by the sensitivity of LCQ mass spectrometer, which was reported down to 5 ppb for each expl osive with a LDR that reaches up to 1000 ppb.63 However, the LOD and LDR for explosive compounds obtained in this research are maintained at the same level to the results reported previously63 despite the loss of 95 % ions in the FAIMS cell and extender capillary, indicating that FAIMS can really improve the sensitivity by filtering out background noise. Although FAIMS generate chromatography-st yle data, its separation is based upon a different principle from chromatographic separati on. In contrast to gas chromatography or liquid chromatography, in which almost all the analyte pa ss the column and can be detected, in FAIMS, only some of the analyte ions w ill pass through the cell and the rest of them will be discharged on the electrodes. Therefore, the quantita tive results acquired by FAIMS without applying internal standards are only relative concentrations (or amount) and can serve for semiquantitative analysis method. Figure 4-8 and Figure 4-9 compare the mass spectra collected by APCI-MS and APCIFAIMS-MS. Figure 4-8 presents the mass spectra for analytes containi ng 50 ng/mL explosives collected by full-scan APCI-M S and APCI-FAIMS-MS from m/z 50 to 500. As seen in Figure 4-8, the filtering capability is evident in the lo w levels of background noise in this system as compared with those collected without FAIMS. Recall that only the ions which are compensated correctly by a specific CV are transmitted through FAIMS cell and detected. Nevertheless, some less abundant ions were still obser ved in the spectra co llected with FAIMS. Some of these are

PAGE 124

124 fragment ions produced from the molecular io ns which have passed the FAIMS cell in the extender or vacuum area of mass spectrometer. In Figure 4-9, the concentration of explosiv e compounds was further reduced to 10 ng/mL and the mass scan range was restricted between m/z 150 and m/z 300 to avoid the fragment or background noise produced in th e low mass region. The select ed CV value mass spectra collected by APCI-FAIMS-MS in Figure 4-9 a ll demonstrate a significant reduction in the background ion signal. The target ion of each explosive compound, which is pointed out by a red square, is still the most abundant peak in e ach spectra and much easier to identify than the target peak in spectra collected without FAIMS. Conclusion In this chapter, the performance of APC I-FAIMS-MS in separation and detection was evaluated. The best repeatability of CV value was obtained at DV of 4500 V for type A or type B ions and at DV of 4500 V for type C ions. Th e SDs and RSDs of CV values were distributed from 0.02 V to 0.2 V and 1.0% to 8.1%, respectively. The ability of FAIMS to separate explosives from mixtures has been demonstr ated. Although higher DV was shown to provide increased CV value and resolving power, it also yielded broader peaks. Addition of helium to the carrier gas improved the separation between isomeric and similar explosives; however, it decreased the transmission of explosive ions for type A and type B ions. The ratio of helium required to resolve explosive peaks needs to be optimized experimentally for different occasions. Two scan modes, full scan and SIM, and three lengths of det ection time, 1minute, 30 seconds, and 10 seconds, were tested for LOD and LDR. The method proved to be sensitive for nitroaromatic explosives down to the average co ncentration of 14 ppb for full scan and 4 ppb for SIM and to extend a linear dynamic range up to 1000 ppb for most nitroaromatic explosive compounds. Although full-scan gave higher LOD a nd decreased LDR, it is still essential for

PAGE 125

125 explosive detection because it provides more info rmation about molecular and fragment ions. The preferred detection time to monitor specific explosive co mpound by setting to the optimum CV is 30 seconds and the amount of explosive compound required for this accumulation time was shown to be as low as 32 pg.

PAGE 126

126 Table 4-1. Repeatability of CV values from fi ve replicate analyzes of explosive compounds. DV(kV)-4.5-4-3.5-3-2.52.533.544.5 CV(Volts)5.74.32.81.00.9-1.0-1.6-2.6-3.6-5.5 SD0.060.060.100.000.060.260.170.260.210.15 RSD % 1.01.33.60.06.226.510.810.25.72.8 CV(Volts)7.35.53.92.51.3-1.3-2.1-3.4-5.2-6.8 SD0.100.120.060.060.060.170.200.210.420.50 RSD % 1.42.11.52.34.613.39.56.28.07.4 CV(Volts)9.67.35.33.21.7-1.6-2.5-4.3-6.6-9.2 SD0.200.100.120.050.100.290.090.100.540.45 RSD % 2.11.32.21.75.918.53.82.48.24.9 CV(Volts)10.37.75.33.31.7-1.8-2.9-4.6-7.2-9.5 SD0.150.210.150.060.100.250.150.210.360.44 RSD % 1.52.72.91.75.914.25.24.55.04.6 CV(Volts)8.05.83.82.31.3-1.4-1.9-3.6-5.4-7.4 SD0.120.120.100.120.200.100.060.060.310.67 RSD % 1.42.02.64.915.47.13.01.65.79.0 CV(Volts)10.67.95.63.42.0-1.9-2.8-4.8-7.4-10.0 SD0.150.200.120.000.120.150.060.360.450.74 RSD % 1.42.52.10.05.97.92.07.56.17.4 CV(Volts)2.72.21.50.90.5-0.5-0.8-1.2-1.8-2.5 SD0.150.130.060.110.040.050.200.050.160.00 RSD % 5.35.94.011.98.710.825.04.38.80.0 CV(Volts)2.21.91.51.40.80.50.10.30.30.3 SD0.070.060.100.180.140.110.050.070.100.05 RSD % 3.23.36.413.217.422.050.322.833.516.7 0.120.120.100.070.100.170.120.170.320.37 2.22.73.24.58.715.113.77.410.16.6 CV(Volts)2.52.22.01.41.00.40.40.40.30.2 SD0.210.250.200.200.240.060.050.050.030.02 RSD % 8.411.710.414.524.813.212.213.410.28.3 CV(Volts)-0.4-0.40.10.70.70.50.71.11.12.3 SD0.110.100.040.160.140.040.110.070.030.10 RSD % 26.724.227.423.321.46.916.07.03.24.3 CV(Volts)0.00.30.50.50.40.81.11.11.52.0 SD0.430.120.140.150.080.130.140.180.080.16 RSD % 18.543.326.828.220.116.213.316.65.38.1 0.250.160.130.170.160.070.100.100.050.09 17.926.421.522.022.112.113.812.46.26.9 Type A or Type B ions Average RSD % Average SD 1,3-DNB Tetryl(-NO2) RDX(+Cl) HMX(+Cl) PETN(+Cl) NG(+Cl) Average SD TNT TNB 2,4-DNT 2,6-DNT 3,4-DNT Average RSD %

PAGE 127

127 Table 4-2. Resolving power for explosive compounds. DV(kV)-4.5-4-3.5-3-2.52.533.544.5 TNT4.463.332.551.030.820.811.141.632.103.09 TNB5.934.453.622.401.231.151.792.414.127.85 2,4-DNT4.524.073.762.441.280.891.192.182.875.43 2,6-DNT5.444.333.552.781.381.332.142.173.554.92 3,4-DNT4.893.772.921.660.950.981.261.702.933.06 1,3-DNB6.014.733.902.681.681.932.834.805.855.76 Tetryl(-NO2) 2.051.591.220.730.450.340.580.801.201.73 RDX(+Cl)0.800.630.640.410.280.220.170.170.130.08 HMX(+Cl)0.120.100.040.200.220.170.260.440.501.03 PETN(+Cl)0.000.090.180.200.130.260.350.380.530.80 NG(+Cl)0.710.700.560.480.280.290.060.160.180.19

PAGE 128

128 2,4-DNT MW=182.13 3,4-DNT MW=182.13 2,6-DNT MW=182.13CH3N+O-O N+O-O CH3N+O-O N+O-O CH3N+O-ON+O-O Figure 4-1. Structures of the isomeric explosives studied in this research. Table 4-3. Resolution between TNT, TNB and DNT isomers. DV(kV)-2.5-3-3.5-4-4.5-5-5-5Carrier gas He/N20/1000/1000/1000/1000/1000/10010/9020/80 2,4-DNTTNT0.361.131.171.141.371.762.593.03 TNB0.200.350.670.720.820.981.671.94 2,6-DNT0.010.080.020.110.200.440.380.75 3,4-DNT0.160.370.640.550.520.590.980.91 2,6-DNTTNT0.391.271.151.281.732.503.053.64 TNB0.220.460.670.871.141.672.142.67 3,4-DNT0.180.450.640.680.791.191.441.68 3,4-DNTTNT0.170.660.490.600.941.482.062.35 TNB0.010.070.030.130.280.470.851.10

PAGE 129

129 RT: 4.00 5.98 SM: 5B 4.0 4.2 4.4 4.6 4.8 5.0 5.2 5.4 5.6 5.8 Time (min) 0 20 40 60 80 100 Relative Abundance 0 20 40 60 80 100 Relative Abundance 0 20 40 60 80 100 Relative Abundance Full Scan m/z 181 [DNT-H]-m/z 182 [DNT]-3,4-DNT 2,6-DNT 2,4-DNT 0 2 4 6 8 10 12 14 16 18 20 CV (Volts) Figure 4-2. CV spectra of a solution mixtur e of 2,4-DNT, 2,6-DNT, and 3,4-DNT at DV of 5000 V and in the carrier gas of 20:80 helium/nitrogen.

PAGE 130

130 RT: 2.00 4.00 SM: 5B 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4 Time (min) 0 20 40 60 80 100 Relative Abundance 0 20 40 60 80 100 Relative Abundance 0 20 40 60 80 100 Relative Abundance Full Scan m/z 227 [TNT]-m/z 182 [DNT]-3,4-DNT 2,6-DNT TNT 0 2 4 6 8 10 12 14 16 18 20 CV (Volts) RT: 2.00 4.00 SM: 5B 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4 Time (min) 0 20 40 60 80 100 Relative Abundance 0 20 40 60 80 100 Relative Abundance 0 20 40 60 80 100 Relative Abundance Full Scan m/z 227 [TNT]-m/z 182 [DNT]-3,4-DNT 2,6-DNT TNT 0 2 4 6 8 10 12 14 16 18 20 CV (Volts) A B Figure 4-3. CV spectra of a so lution mixture of TNT, 2,6-DNT, and 3,4-DNT at DV of -5000 V and in (A) the nitrogen carrier gas, (B) the carrier gas of 20:80 helium/nitrogen.

PAGE 131

131 RT: 2.00 4.00 SM: 5B 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4 Time (min) 0 50 100 0 50 100 0 50 100 0 50 100 Full Scan m/z 181 [DNT-H]-m/z 213 [TNB]-m/z 182 [DNT]-2,4-DNT 2,6-DNT TNB 0 2 4 6 8 10 12 14 16 18 20 CV (Volts) Figure 4-4. CV spectra of a solution mixtur e of TNB, 2,4-DNT, and 2,6-DNT at DV of 5000 V and in the carrier gas of 10:90 helium/nitrogen.

PAGE 132

132 Table 4-4. Resolution of explosive mixtures. DV(kV)2.533.544.54.54.54.55Carrier gas He/N20/1000/1000/1000/1000/10010/9020/8030/7030/70 RDXTNT0.520.640.891.121.571.090.96 HMX0.020.070.160.200.520.960.71 HMXTNT0.420.681.081.432.302.151.57 PETNTNT0.490.700.981.332.061.591.69 NG0.070.250.190.310.490.540.77 NGTNT0.600.670.981.352.020.920.89

PAGE 133

133 Qualifying Examination Ap ril 18, 2008 0.0 0.5 1.0 1.5 2.0 2.5 Time (min) 0 50 100 0 50 100 0 50 100 Relative Abundance 0 50 100 Full Scan m/z 257 [RDX+Cl]-m/z 226 [TNT-H]-m/z 331 [HMX+Cl]RDX HMX TNT -20 -10 0 10 20 CV (Volts) Figure 4-5. CV spectra of a solu tion mixture of TNT, RDX, and HMX at DV of 4500 V with the carrier gas of 30:70 helium/nitrogen.

PAGE 134

134 RT: 0.00 3.00 SM: 3B 0.0 0.5 1.0 1.5 2.0 2.5 3 Time (min) 0 50 100 0 50 100 0 50 100 0 50 100 Qualifying Examination Ap ril 18, 2008 Full Scan m/z 351 [PETN+Cl]-m/z 226 [TNT-H]-m/z 262 [NG+Cl]TNT NG PETN -20 -10 0 10 2 0 CV (Volts) Figure 4-6. CV spectra of a solu tion mixture of TNT, NG, and PETN at DV of 4500 V with the nitrogen carrier gas.

PAGE 135

135 RT: 10.00 12.00 SM: 3B 10.0 10.2 10.4 10.6 10.8 11.0 11.2 11.4 11.6 11.8 1 2 Time (min) 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Relative Abundance Tetryl TNT TNB2,6-DNT 3,4-DNT 2,4-DNT DNB 0 2 4 6 8 10 12 14 16 18 20 CV (Volts) Figure 4-7. IS-CV spectrum of nitr oaromatic explosives at DV of 4500 V and in the nitrogen carrier gas.

PAGE 136

136 Table 4-5. Reproducibility of peak areas from five replicate an alyzes of explosive compounds. DV(kV)-4.5-4-3.5-3-2.52.533.544.5 Peak area468814430507359906305778 22521922662720860923850 5241822165238 RSD % 4.29.85.66.812.78.76.517.911.921.7 Peak area820227703567486363235923 10404872029671523929 92772219791 RSD % 2.35.58.98.52.08.68.622.212.927.6 Peak area13926301122380678174326222 191489100343794804922 83411714313 RSD % 2.42.25.011.911.44.59.87.39.019.6 Peak area17252861248024698343336120 163783128165905065092 11988218977 RSD % 1.04.54.06.518.67.512.114.315.932.5 Peak area1026866805471459985256334 142445119176815296098 43987221790 RSD % 4.70.910.35.416.114.34.314.615.129.8 Peak area1344159990463626242296082 15724981181628403632 91918113408 RSD % 1.45.22.56.314.41.125.020.820.529.2 Peak area635514633901495748300800 19285425608031522030615 1307642403898 RSD % 2.41.01.77.013.421.93.812.17.65.5 Peak area3698229374795939729493374851 27042751649147187102420626 4123367552670661 RSD % 7.44.73.55.95.310.110.95.74.92.1 Peak area3453817408860843699034464514 41438682555905345811944397 5656768816578504 RSD % 8.34.73.05.24.18.710.17.22.44.2 Peak area4637077504501657364544425278 38950872250896569471062944 7064628916666631 RSD % 5.44.41.76.712.816.35.52.41.82.7 Peak area3509916432089547874424212408 35430021954903231753618196 0915296551262850 RSD % 4.52.23.23.45.49.39.411.912.18.2 4.04.14.56.710.610.19.612.410.416.6 TNT TNB 2,4-DNT 2,6-DNT 3,4-DNT 1,3-DNB Tetryl(-NO2) RDX(+Cl) HMX(+Cl) PETN(+Cl) NG(+Cl) Average RSD %

PAGE 137

137 Table 4-6. Linear dynamic range and limits of detection for the nitroaromatic explosives collected by full scan and SIM mode. linear dynamic range (ng/mL) corr coef (R2) concn (ng/mL) amt injected (pg) linear dynamic range (ng/mL) corr coef (R2) concn (ng/mL) amt injected (pg) TNT145-100000.99321663741-100000.99726223 TNB31-100000.994328111322-100000.9978297 2,4-DNT49-10000.98691142735-10000.99455206 2,6-DNT13-10000.991415125-10000.9972292 3,4-DNT47-10000.9711938014-100000.99907276 1,3-DNB22-10000.9828103859-10000.9930287 Tetr y l ( -NO2 ) 45-10000.9624228806-100000.9847272 Average0.9832145530.99484150 limit of detectionlimit of detection Full scanSIM Table 4-7. Linear dynamic range and limits of detection at the optimum CV for transmi ssion of the nitro aromatic explosives for varied collection time. linear dynamic range (ng/mL) corr coef (R2) concn (ng/mL) amt injected (pg) linear dynamic range (ng/mL) corr coef (R2) concn (ng/mL) amt injected (pg) linear dynamic range (ng/mL) corr coef (R2) concn (ng/mL) amt injected (pg) TNT15-100000.997748130-100000.998677249-100000.99712791 TNB19-100000.99631734546-100000.99463434356-100000.995238125 2,4-DNT18-100000.9955612120-10000.998033270-10000.95262065 2,6-DNT22-10000.99641529417-10000.999499017-10000.9962517 3,4-DNT19-10000.9990611935-10000.991866333-10000.99541138 1,3-DNB12-10000.9928510220-10000.997077080-10000.965139130 Tetr y l ( -NO2 ) 29-100000.99492141920-100000.99531515191-100000.975059197 Average0.9961112110.9964121170.98242895 limit of detectionlimit of detectionlimit of detection 1 min 30 s10 s

PAGE 138

138 100 150 200 250 300 350 400 450 50 0 m /z 56 0 20 40 66 0 20 40 46 0 20 93 0 50 R e l a ti ve Ab un d ance 87 0 20 40 60 70 0 20 40 100 0 50 227.07 89.20 75.13 120.93 61.13 93.20 228.07 178.87 135.07 164.93 258.00 89.13 75.13 213.13 120.87 63.00 106.87 178.87 135.07 164.87 214.13 244.07 89.20 75.13 182.07 120.87 61.13 106.80 178.80 133.00 183.07 182.07 89.20 75.13 59.13 120.93 183.07 106.87 178.87 135.00 212.93 89.20 59.13 182.07 75.13 120.93 134.93 178.87 106.80 183.00 89.20 75.13 120.87 199.00 61.13 106.80168.07 178.93 133.00 89.20 75.13 120.93 242.93 61.13 106.73 178.87 226.00 135.07 164.67 243.93 100 150 200 250 300 350 400 450 50 0 m /z 21 0 5 10 15 86 0 20 40 60 51 0 20 100 0 50 R e l a ti ve Ab un d ance 49 0 20 62 0 20 40 58 0 20 40 227.07 197.20 228.13 209.93 192.07 213.07 183.27 115.20 164.93 181.20 152.07 182.00 152.27 184.00 182.00 152.20 168.00 168.13 138.27 241.13 226.13 185.33 213.27 166.00 242.33 432.00 268.67 136.07 482.67 292.87APCI-MSAPCI-FAIMS-MS A B C D E F G Figure 4-8. Mass spectra for analytes containi ng 50 ng/mL explosives collected by APCI-MS and APCI-FAIMS-MS ranging from m/z 50 to 500: (A) TNT, (B) TNB, (C) 2,4-DNT, (D) 2,6-DNT, (E) 3,4-DNT ,(F) 1,3-DNB, (G) Tetryl.

PAGE 139

139 APCI-MSAPCI-FAIMS-MS 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 30 0 m /z 32 0 10 20 47 0 20 62 0 20 40 100 0 50 R e l a ti ve Ab un d ance 78 0 20 40 60 47 0 20 50 0 20 239.00 227.00 165.07 178.87 182.07 177.93 235.07 240.00 212.27 150.93 163.87 167.07 220.27 249.40253.13 227.80 194.73 182.93 275.13 265.07 209.07 291.13 199.13 285.40 297.87 178.87 177.80 213.13 164.87 150.87 177.00 235.13 163.13 239.00 227.07 221.07 182.07 210.80 194.80275.27 249.07 267.33 285.93 293.27 203.27 263.27 296.9 3 182.07 178.87 164.93 239.07 177.93 235.27 150.87 163.53 177.13 227.13 183.00 221.33 249.40 212.20240.07 194.67 265.13 209.07 286.27 259.13 291.07 276.00 297.1 3 182.07 178.93 177.80 164.67 150.73 183.07 212.87235.07 176.87 220.40 227.07249.80 239.00 199.07264.93 193.13 210.27 182.07 178.93 177.93 164.93 150.87 177.07 163.00 235.13 221.13 183.07 239.13 227.13 196.87 249.27 166.87 209.13 275.27 253.93 265.13290.93297.33 285.27 178.80 177.80 164.80 168.13 150.87 198.93 163.87 235.20 239.07 220.40 176.87 227.07 180.93 249.27 212.33 162.33 196.80 291.13 240.07 208.00 265.13275.00286.20 297.07 253.13 178.93 242.13 177.80 150.87 164.80 227.00 158.80 177.13 235.07243.00 220.47 249.80 196.87 181.87 212.27 275.07 207.80 191.00 287.47 265.13 257.20 294.93 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 30 0 m /z 58 0 20 40 70 0 20 40 49 0 20 100 0 50 R e l a ti ve Ab un d ance 23 0 10 24 0 10 95 0 50 226.93 197.20 177.13 228.07 210.00 173.33 167.73 180.27 191.93 151.27 226.40 279.93 197.87 263.00 287.00 239.73293.67 271.07 245.87 213.07 183.27 212.73 155.07 257.67 166.20 215.20 177.93272.73 200.53 247.13 298.33 193.40 260.33 241.93 284.67 291.53 217.80 206.07 223.20 237.27 181.27 182.13 151.20 166.27 159.47 231.60 183.07 179.87 207.93256.27 289.27 279.13 182.00 152.13 182.80 276.20 153.60 179.87 174.47 233.80282.13 238.67 245.93 163.07 182.07 168.00 228.47 273.60 183.00 223.87 234.67 198.00 152.00 243.40 212.13297.80 176.93 165.07 279.53 250.27291.00 270.13 200.93 168.07 169.07 150.40 167.40 237.73 180.53 203.13260.93 241.07 185.33 218.00 217.67 242.00 184.67 218.93 177.07 186.07 201.20225.87 164.93 293.47 209.67 150.40 289.67 248.07 266.80 235.00A B C D E F G Figure 4-9. Mass spectra for analytes containi ng 10 ng/mL explosives collected by APCI-MS and APCI-FAIMS-MS ranging from m/z 150 to 300: (A) TNT, (B) TNB, (C) 2,4DNT, (D) 2,6-DNT, (E) 3,4-DNT ,(F) 1,3-DNB, (G) Tetryl.

PAGE 140

140 CHAPTER 5 CONCLUSIONS AND FUTURE WORK Conclusions As demonstrated in this dissertation, FAIMS is a promising technique for separating gasphase explosive ions at atmospheric pressure, ba sed on changes in their mobility at high electric fields relative to low electric field. FAIMS can behave as an ion filter, capable of transmitting selected compounds in a mixture on to a mass spect rometer. Mass spectrom etry is by far the most widely used technique for explosive iden tification and provides information orthogonal to that provided by FAIMS. In addition, both FA IMS and mass spectrometry instrumentation can be manufactured in a small scale, offering th e potential for API-FAIMS-MS instruments to be portable. An instrument of this kind may be able to replace conventional ion mobility spectrometers (IMS) in the field for explosive detection. In this study, the combination of a mass spectrometer with the FAIMS cell, in which only selected ions are transmitted through the cell, has been proved to greatly simplify mass spectr a over those acquired us ing a conventional mass spectrometer, significantly improving the sensi tivity and selectivity of this method. Two API sources, APCI and DPIS, were i nvestigated and were observed to produce different characteristic ions, and relative intens ities for analysis of explosives. Typically, the DPIS gave more structural information over AP CI through increased frag mentation, presumably due to more abundant O2 -, NO2 and NO3 -. In addition, spectra which present either more information about structure or more abundant molecular ion can be obtained from DPIS by adjusting the components in th e surrounding air. Overall, th e mass spectra of explosive compounds produced by DPIS are comparable to those formed by APCI although the formation of nitrate and nitrite adduct i ons with the explosives is more pronounced with the DPIS source. This phenomenon will benefit explosive investiga tion especially in the field, where additives

PAGE 141

141 may not be available for use. Although the signal intensity generated by DPIS is somewhat lower than APCI due to the spatial obs truction of the neon bulb, the merits presented by DPIS, such as rich ion patterns and decreased complexity of sp ectra for nitramines and nitrate esters, makes DPIS an attractive alte rnative to APCI for e xplosive investigation. The use of FAIMS as a potential method for characterizing explosive compounds has been evaluated. A thorough understanding of ion behavior influenced by experimental parameters was obtained in order to gain the optimum separation or transmission of ions. This knowledge will also be beneficial for the further development of devices for explosives detection based on the FAIMS technique. In this work, the CV scan rate, curtain gas flow rate, dispersion voltage (DV), carrier gas composition and electrode temperature were optimized. The effect of these parameters on the signal intensity (sensitivity), peak width (re solution) and compensation voltage (peak capacity) was studied for explosive compounds of interest. Experiments showed that a CV scan rate of 10 V/s and curtain gas flow rate from 2.0-2.5 L/min were optimal for both resolution and tr ansmission. An increase in CV value and sensitivity was observed at higher DV values for bot h type A and type B ions in N1 and type C ions in N2 modes. Although the ion focusing mechanism in the cylindrical cell improves the sensitivity, it also decreases the resolution due to peak broadening. Furthermore, the separation and sensitivity can also be controlled in FA IMS by changing the carrier gas composition. A mixture of helium and nitrogen was demonstrated to improve resolution and sensitivity. For type A ions, the peak width was reduced and the CV wa s shifted to more positive value in N1 mode with increasing helium content in carrier gas, improving the resolution between those ions. In addition, the sensitivity and peak width were increas ed dramatically for type C ions in N2 mode with increased content of helium in carrier ga s. A broader peak wi dth and increased signal intensity was also seen for type B ions in the same mode at higher DV, indicating that a type B

PAGE 142

142 ion was transformed to a type C ion at higher el ectric field as the helium content in carrier gas increased. Finally, variations were also observed in CV value, peak width, and signal intensity with varied electrode temperatur es. Two type B ions, the [M]ions of TNT and 2,6-DNT, have decreasing mobilities at lower elect ric field with increased temperature, which implies that the raised temperature on electrodes not only reduced the number density of the curtain gas in the FAIMS cell but also affected the inter action of these ions with the gas. The performance of APCI-FAIMS-MS in separa tion and detection was evaluated in this research. The best repeatability of CV value was obtained at DV of 4500 V for type A or type B ions and at DV of 4500 V for type C ions. Th e SDs and RSDs of CV values were distributed from 0.02 V to 0.2 V and 1.0% to 8.1%, respectively. The ability of FAIMS to separate explosives from mixture has been demonstrat ed. Although higher DV wa s proved to increase CV values and resolving power, it also leads to peak broadening. Addition of helium to the carrier gas improved the separation between simi lar explosives and isomers; however, it also decreased the transmission of explos ive ions for type A and type B ions. In this research, most explosives can be resolved by the combination of CV spect rum from FAIMS separation and mass select-ion from the mass spectrometry. In br ief, the optimum condition of each parameter required to resolve explosive peaks needs to be di scovered experimentally fo r different analytical situations. Although quantitation of these compounds was not the purpose of this research, calibration curves were constructed in order to test linear ity and sensitivity of the APCI-FAIMS-MS method. Two scan modes, full scan and SIM, and three lengths of det ection time, 1minute, 30 seconds, and 10 seconds, were tested for LOD and LDR. The quantitative study showed that FAIMS was sensitive for nitroaromatic explosives down to the average con centration of 14 ppb for full scan and 4 ppb for SIM, and provided linear calibration for at least 3 orders of magnitude. Although

PAGE 143

143 the full scan gave higher LOD and decreased LDR, it is still essential for explosive detection because of more information about molecular and fragment ions provided by this detection mode. The preferred detection time to monitor specific explosive co mpound by setting to the optimum CV is 30 seconds and the amount of explosive compound required for this accumulation time can be down to 32 pg. The ultimate goal of this research was to evaluate the feasibility of API-FAIMS-MS to detect explosive compounds, and to ascertain me thodologies to eventually do so in a field environment. According to the experimental da ta shown in this resear ch, the integration of FAIMS with mass spectrometry for the analysis of explosive compounds was very fruitful, permitting a sensitive, selective, and rapid analysis of explosives. The method is able to perform a fast separation of explosive ions and a selective and se nsitive detection of different classes of explosives, pointing to the develo pment of a powerful portable de vice for monitoring explosives in field. The development of fieldable explos ive device based on this concept could make a contribution toward the protection of first respon ders and emergency personnel, diagnosis of the nature of the attack, and gathering of forensic data, or even for the prevention of terrorist activity. Future Work Preliminary results have shown that API-FAIMS -MS is a viable method for the analysis of explosive compounds, but further studies are re quired to improve the sensitivity, resolution, portability, and reliabilit y of the method to be employed in the field. This work would likely include further research in several modifications on each component (ion source, FAIMS, and mass spectrometer) to allow the instrument to serve its purposes.

PAGE 144

144 Ionization Source The DPIS has presented the superiority over APCI on the explosives detection by generating diversified and selectab le ion pattern on mass spectra. It may be potentially used for the fast detection of explosives in the field with the advantages that include design configuration flexibility, dimensional stability, simplicity and ruggedness of design, and extended source lifetime. However, in this research, th e ion intensity produced by DPIS is only 10 to 50 of ion intensity by APCI due to th e inefficiency on the transmissi on of the ions from ionization source to detector, which is primary caused by th e spatial obstacle of the DPIS neon bulb itself. A possible resolution is to reshape the DPIS neon bulb into a cylinder of tube, which may increase the efficiency of both ionization and i on transport by directing the analytes through the ionization area inside the cylinder. The other important issue for the desirable io nization source employing in the field is the capability to ionize explosives directly from various matrices. A design based on similar mechanism termed dielectric ba rrier discharge ionization (DBDI)17, 74 has been demonstrated to permit desorption and ionization of the explosives from solid surfaces. The uses of discharge gases to assist desorption and ioni zation may provide an alternative to integrate with DPIS in the future design. FAIMS How to downsize the FAIMS cell without comp romising resolution and sensitivity will be a concern when developing a portable explosive detector. Ongoing efforts in our group are aimed at developing FAIMS cells of varied geometries (planar, cylindrical, hemispherical and spherical cells), and varied dimensions in order to fabricate a field-porta ble explosive detector for providing early and timely de tection of different classes of explosives. Current research

PAGE 145

145 demonstrates that the planar geometry provide s better resolution and the spherical geometry gives higher transmission. The resolution and se nsitivity are also influenced by the dimensions (gap width, height and length) of the cell. Therefore, further st udies are necessary to determine the ideal FAIMS cell design which may combine several geometry and optimum dimensions, which will provide the best combination of resolution and sensitivity. Additional experiments using FAIMS could include the evaluation of different waveform types and frequencies, curtain gas combinations, and designs for introduction of analyte ions into the FAIMS cell. It would be advantageous to improve our understanding of how ions behave under the influence of different parameters of FAIMS in order to apply the most optimum condition when more sensitivity is required for the application of interest. Mass Spectrometer As suggested in the previous chapter, the limit of detection of the API-FAIMS-MS method was limited primarily by the mass spectrome ter. For this reason, the use of more sensitive mass spectrometer, such as triple quadrupole, can be expect ed to increase the sensitivity of this method. In addition, modification of the commercial FAIMS cell or the mass spectrometer so the FAIMS cell could be attached directly onto the inlet of mass spectrometer could eliminate the capillary extender used in th is research that caused a 75% decrease on signal intensity.

PAGE 146

146 LIST OF REFERENCES (1) Zhang, B.; Pan, X.; Cobb, G. P.; Anderson, T. A. J Chromatogr B Analyt Technol Biomed Life Sci 2005 824 277-282. (2) Calderara, S.; Gardebas, D.; Martinez, F. Forensic Sci Int 2003 137 6-12. (3) Walsh, M. E. Talanta 2001 54 427-438. (4) Sigman, M. E.; Clark, C. D.; Fidler R.; Geiger, C. L.; Clausen, C. A. Rapid Commun Mass Spectrom 2006 20 2851-2857. (5) Calderara, S.; Gardebas, D.; Martinez, F.; Khong, S. P. J Forensic Sci 2004 49 10051008. (6) Sharma, S. P.; Lahiri, S. C. J Energetic Mater 2005 23 239-264. (7) Bradley, K. S. J Forensic Sci 2005 50 96-103. (8) Sigman, M. E.; Ma, C. Y. J Forensic Sci 2001 46 6-11. (9) Pan, X.; Zhang, B.; Tian, K.; Jones, L. E.; Liu, J.; Anderson, T. A.; Wang, J. S.; Cobb, G. P. Rapid Commun Mass Spectrom 2006 20 2222-2226. (10) Mathis, J. A.; McCord, B. R. Rapid Commun Mass Spectrom 2005 19 99-104. (11) Xu, X.; van de Craats, A. M.; Kok, E. M.; de Bruyn, P. C. J Forensic Sci 2004 49 12301236. (12) Chen, Y.; Kim, S. C.; Zhao, Y. Anal Chem 2005 77 8179-8184. (13) Marple, R. L.; Lacourse, W. R. Talanta 2005 66 581-590. (14) Schmidt, A. C.; Niehus, B.; Matysik, F. M.; Engewald, W. Chromatographia 2006 63 111. (15) Yinon, J. Forensic Applications of Mass Spectrometry ; CRC Press: Boca Raton, FL, 1995 (16) McAvoy, Y.; Dost, K.; Jones, D. C.; Cole, M. D.; George, M. W.; Davidson, G. Forensic Sci Int 1999 99 123-141. (17) Na, N.; Zhang, C.; Zhao, M.; Zhang, S.; Yang, C.; Fang, X.; Zhang, X. J Mass Spectrom 2007 42 1079-1085. (18) Pan, X.; Zhang, B.; Cox, S. B.; Anderson, T. A.; Cobb, G. P. J Chromatogr A 2006 1107 2-8. (19) Gapeev, A.; Yinon, J. J Forensic Sci 2004 49 227-237.

PAGE 147

147 (20) Fu, X. F.; Zhang, Y.; Shi, S. H.; Gao, F. ; Wen, D. W.; Li, W.; Liao, Y. P.; Liu, H. W. Rapid Commun Mass Spectrom 2006 20 2906-2914. (21) Tachon, R.; Pichon, V.; Le Borgne, M. B.; Minet, J. J. J Chromatogr A 2007 1154 174181. (22) Zhang, M.; Shi, Z.; Bai, Y.; Gao, Y.; Hu, R.; Zhao, F. J Am Soc Mass Spectrom 2006 17 189-193. (23) Popov, I. A.; Chen, H.; Kharybin, O. N.; Nikolaev, E. N.; Cooks, R. G. Chem Commun (Camb) 2005 1953-1955. (24) Gillen, G.; Mahoney, C.; Wight, S.; Lareau, R. Rapid Commun Mass Spectrom 2006 20 1949-1953. (25) Cody, R. B.; Laramee, J. A.; Durst, H. D. Anal Chem 2005 77 2297-2302. (26) Kpegba, K.; Spadaro, T.; Cody, R. B.; Nesnas, N.; Olson, J. A. Anal Chem 2007 79 54795483. (27) Justes, D. R.; Talaty, N.; Cotte-Rodriguez, I.; Cooks, R. G. Chem Commun (Camb) 2007 2142-2144. (28) Song, Y.; Cooks, R. G. Rapid Commun Mass Spectrom 2006 20 3130-3138. (29) Blanchard, W. C. In U.S. Pat. Appl. Publ. : US, 2006 pp 11. (30) Eiceman, G. A. K., Z. Ion Mobility Spectrometry ; CRC Press: Boca Raton, FL, 1993 (31) Ewing, R. G.; Atkinson, D. A. ; Eiceman, G. A.; Ewing, G. J. Talanta 2001 54 515-529. (32) Eiceman, G. A.; Krylov, E. V.; Krylova N. S.; Nazarov, E. G.; Miller, R. A. Anal Chem 2004 76 4937-4944. (33) Evans, C. S.; Sleeman, R.; Luke, J.; Keely, B. J. Rapid Commun Mass Spectrom 2002 16 1883-1891. (34) Guevremont, R.; Purves, R. W. J Am Soc Mass Spectrom 1999 10 492-501. (35) Barnett, D. A.; Ding, L.; Ells, B.; Purves, R. W.; Guevremont, R. Rapid Commun Mass Spectrom 2002 16 676-680. (36) Guevremont, R. J Chromatogr A 2004 1058 3-19. (37) Zhao, X.; Yinon, J. J Chromatogr A 2002 946 125-132. (38) Barnett, D. A.; Purves, R. W.; Ells, B.; Guevremont, R. J Mass Spectrom 2000 35 976980.

PAGE 148

148 (39) Gao, L.; Song, Q.; Patterson, G. E.; Cooks, R. G.; Ouyang, Z. Anal Chem 2006 78 59946002. (40) Eiceman, G. A.; Nazarov, E. G.; Mill er, R. A.; Krylov, E. V.; Zapata, A. M. Analyst 2002 127 466-471. (41) Eiceman, G. A.; Tadjikov, B.; Krylov, E. ; Nazarov, E. G.; Miller, R. A.; Westbrook, J.; Funk, P. J Chromatogr A 2001 917 205-217. (42) Robinson, E. W.; Sellon, R. E.; Williams, E. R. Int J Mass Spectrom 2007 259 87-95. (43) Tang, K.; Li, F.; Shvartsburg, A. A. ; Strittmatter, E. F.; Smith, R. D. Anal Chem 2005 77 6381-6388. (44) Levin, D. S.; Miller, R. A.; Nazarov, E. G.; Vouros, P. Anal Chem 2006 78 5443-5452. (45) Guevremont, R.; Purves, R. W. J Am Soc Mass Spectrom 1999 10 492-501. (46) Buryakov, I. K., E.; Nazarov, E.; Rasulev, U. Int. J. Mass Spectrom. 1993 128 143-148. (47) Barnett, D. A.; Ells, B.; Guevremont, R.; Purves, R. W.; Viehland, L. A. J Am Soc Mass Spectrom 2000 11 1125-1133. (48) Shvartsburg, A. A.; Bryskiewicz, T.; Pu rves, R. W.; Tang, K.; Guevremont, R.; Smith, R. D. J Phys Chem B 2006 110 21966-21980. (49) Kolakowski, B. M.; Mester, Z. Analyst 2007 132 842-864. (50) Guevremont, R.; Purves, R. W. Rev Sci Instrum 1999 70 1370-1383. (51) Purves, R. W.; Barnett, D. A.; Guevremont, R. Int J Mass Spectrom 2000, 197 163-177. (52) Purves, R. W.; Guevremont, R.; Day, S.; Pipich, C. W.; Matyjaszczyk, M. S. Rev Sci Instrum 1998 69 4094-4105. (53) Guevremont, R.; Purves, R. W. Rev Sci Instrum 1999 70 1370-1383. (54) Shvartsburg, A. A.; Bryskiewicz, T.; Pu rves, R. W.; Tang, K. Q.; Guevremont, R.; Smith, R. D. J Phys Chem B 2006 110 21966-21980. (55) Shvartsburg, A. A.; Tang, K.; Smith, R. D. J Am Soc Mass Spectrom 2005 16 2-12. (56) Tsai, C.-W. P., M.; Kaminsky, I.; Ferran, R.; Boumsellek, S.; Yost, R.A. In 57th ASMS Conf. Mass Spectrometry and Allied Topics : Philadelphia, PA, 2009 (57) Paul, W. S., H. In U.S. Pat. Appl. Publ. : US, 1960 ; Vol. 2939952. (58) Fernandez, L. E. M. Carbohydrate Polymers 2007 68 797-807.

PAGE 149

149 (59) Knight, R. D. Int J Mass Spectrom Ion Proce 1983 51 127-131. (60) March, R. E. J Mass Spectrom 1997 32 351-369. (61) March, R. E.; Splendore, M.; Reiner, E. J.; Mercer, R. S.; Plomley, J. B.; Waddell, D. S.; MacPherson, K. A. Int J Mass Spectrom 2000 194 235-246. (62) Mathieu, E. J. Math. Pure Appl 1868 13 137. (63) Reich, R. Fundamentals and Applications of Atmos pheric Pressure Chemical Ionization Quadrupole Ion Trap Mass Spectrometry for the Analysis of Explosives ; MS thesis, University of Florida., 2001 (64) Kaiser, R. E.; Cooks, R. G.; Stafford, G. C.; Syka, J. E. P.; Hemberger, P. H. Int J Mass Spectrom Ion Proce 1991 106 79-115. (65) Finnigan LCQ Series APPI/APCI Combinati on Probe Operators Manual, Rev. A 2004 (66) Hoffmann, E.; Stroobant, V. Mass Spectrometry: Principle and Application 2nd ed.; John Wiley: Chichester, 2002 (67) Boock, J. J. Characterization of Commonly Encount ered Explosives using High-Field Asymmetric Waveform Ion Mobility Spectro metry Coupled with Mass Spectrometry ; MS thesis, University of Florida, 2007 (68) Waltman, M. J.; Dwivedi, P.; Hill, H. H.; Blanchard, W. C.; Ewing, R. G. Talanta 2008 77 249-255. (69) Takada, Y.; Nagano, H.; Suga, M.; Hashim oto, Y.; Yamada, M.; Sakairi, M.; Kusumoto, K.; Ota, T.; Nakamura, J. Propellants Explos Pyrotech 2002 27 224-228. (70) Sigman, M. E.; Ma, C. Y.; Ilgner, R. H. Anal Chem 2001 73 792-798. (71) Sielemann, S.; Baumbach, J. I.; Schmidt, H.; Pilzecker, P. Anal Chim Acta 2001, 431 293301. (72) Cotte-Rodriguez, I.; Takats, Z.; Talaty, N.; Chen, H.; Cooks, R. G. Anal Chem 2005 77 6755-6764. (73) Talaty, N.; Mulligan, C. C.; Justes, D. R.; Jackson, A. U.; Noll, R. J.; Cooks, R. G. Analyst 2008 133 1532-1540. (74) Na, N.; Zhao, M.; Zhang, S.; Yang, C.; Zhang, X. J Am Soc Mass Spectrom 2007 18 1859-1862. (75) Zhao, X.; Yinon, J. J Chromatogr A 2002 977 59-68. (76) Caldwell, T. E.; Foster, K. L.; Benter, T.; Langer, S.; Hemminger, J. C.; Finlayson-Pitts, B. J. J Phys Chem A 1999 103 8231-8238.

PAGE 150

150 (77) Dillard, J. G. Chem Rev 1973 73 589-643. (78) Moruzzi, J. L.; Phelps, A. V. J Chem Phys 1966 45 4617-4627. (79) Thomson, B. A.; Davidson, W. R.; Lovett, A. M. Environmental Health Perspectives 1980 36 77-84. (80) Fehsenfe.Fc; Schmelte.Al; Schiff, H. I.; Ferguson, E. E. Planet Space Sci 1967 15 373. (81) Paulson, J. F. J Chem Phys 1970 52 963-964. (82) Fehsenfe.Fc; Ferguson, E. E. J Chem Phys 1974 61 3181-3193. (83) Ferguson, E. E. Rev Geophys Space Phys 1971 9 997. (84) Fehsenfe.Fc; Ferguson, E. E.; Bohme, D. K. Planet Space Sci 1969 17 1759. (85) Buryakov, I. A.; Kolomiets, Y. N.; Luppu, B. V. J Anal Chem 2001 56 336-340. (86) Tabrizchi, M.; Khayamian, T.; Taj, N. Rev Sci Instrum 2000 71 2321-2328. (87) Ewing, R. G.; Atkinson, D. A. ; Eiceman, G. A.; Ewing, G. J. Talanta 2001 54 515-529. (88) Song, L. G.; Wellman, A. D.; Yao, H. F.; Bartmess, J. E. J Am Soc Mass Spectrom 2007 18 1789-1798. (89) Spangler, G. E.; Lawless, P. A. Anal Chem 1978 50 884-892. (90) Harrison, A. Chemical Ionization Mass Spectrometry ; CRC Press: New York, NY, 1983 (91) Song, Y. S.; Cooks, R. G. Rapid Commun Mass Spectrom 2006 20 3130-3138. (92) Yinon, J.; McClellan, J. E.; Yost, R. A. Rapid Commun Mass Spectrom 1997 11 19611970. (93) Huang, S. D.; Kolaitis, L.; Lubman, D. M. Applied Spectroscopy 1987 41 1371-1376. (94) Caldwell, G. W.; Masucc i, J. A.; Ikonomou, M. G. Org Mass Spectrom 1989 24 8-14. (95) Linstrom, P. J.; Mallard, W.G. NIST Standard Database Number 69 ; National Institute of Standards and Technology: Gaithersburg, MD, 2005 (96) Krylov, E. V. Tech Phys 1999 44 113-116. (97) Krylov, E. V. Int J Mass Spectrom 2003 225 39-51. (98) Buryakov, I. A.; Kolomiets, Y. N.; Luppu, B. V. J Anal Chem 2001 56 336-340. (99) Buryakov, I. A. Talanta 2003 61 369-375.

PAGE 151

151 (100) Buryakov, I. A. J Chromatogr B Analyt T echnol Biomed Life Sci 2004 800 75-82. (101) Otero Santos, S. M. Allelochemicals that Modify Mos quito Most-seeking Behavior: Ggas Chromatography/Mass Spectroetery and High-fiel d Asymmetric Waveform Ion Mobility Spectrometery PhD dissertation, University of Florida, 2007 (102) Bryant, J., A,G Separation of Tryptic Peptide Ions by High-Field Asymmetric Waveform Ion Mobility Spectrometry ; PhD Dissertation, University of Florida, 2008 (103) Keller, T.; Keller, A.; Tu tsch-Bauer, E.; Monticelli, F. Forensic Sci Int 2006 161 130140. (104) Shvartsburg, A. A.; Tang, K.; Smith, R. D. J Am Soc Mass Spectrom 2004 15 1487-1498. (105) Guevremont, R.; Purves, R. J Am Soc Mass Spectrom 2005 16 349-362. (106) McCooeye, M. A.; Ells, B.; Barnett, D. A.; Purves, R. W.; Guevremont, R. J Anal Toxicol 2001 25 81-87. (107) McCooeye, M.; Ding, L.; Gardner, G. J.; Fraser C. A.; Lam, J.; Sturgeon, R. E.; Mester, Z. Anal Chem 2003 75 2538-2542. (108) Purves, R. W.; Barnett, D. A.; Ells, B.; Guevremont, R. J Am Soc Mass Spectrom 2001 12 894-901. (109) Ells, B.; Barnett, D. A.; Purves, R. W.; Guevremont, R. Anal Chem 2000 72 4555-4559. (110) Cui, M.; Ding, L.; Mester, Z. Anal Chem 2003 75 5847-5853. (111) Shvartsburg, A. A.; Tang, K.; Smith, R. D. Anal Chem 2004 76 7366-7374. (112) McCooeye, M. A.; Mester, Z.; Ells, B.; Ba rnett, D. A.; Purves, R. W.; Guevremont, R. Anal Chem 2002 74 3071-3075. (113) Mitchell Roberts, A. C. Advanced Techniques in Mass Spectrometry for Forensic Applications ; PhD Dissertation, University of Florida, 2005 (114) Thackston, M. G.; Ellis, H. W. ; Pai, R. Y.; McDaniel, E. W. J Chem Phys 1976 65 20372038. (115) Thackston, M. G.; Eisele, F. L. ; Ellis, H. W.; McDaniel, E. W. J Chem Phys 1977 67 1276-1277. (116) Barnett, D. A.; Belford, M. ; Dunyach, J. J.; Purves, R. W. J Am Soc Mass Spectrom 2007 18 1653-1663. (117) Robinson, E. W.; Shvartsburg, A. A.; Tang, K.; Smith, R. D. Anal Chem 2008 80 75087515.

PAGE 152

152 (118) Burns, D. T.; Lewis, R. J. Anal Chim Acta 1995 309 189-195. (119) Yinon, J. Forensic and Environmental Detection of Explosives ; John Wiley: Chichester ; New York, 1999. (120) Long, G. L.; Winefordner, J. D. Anal Chem 1983 55 712A-724A. (121) Kolakowski, B. M.; D'Agostino, P. A.; Chenier, C.; Mester, Z. Anal Chem 2007 79 82578265.

PAGE 153

153 BIOGRAPHICAL SKETCH Alex Ching-Hong Wu was born in 1972, in Kaohsiung County, Taiwan. He attended Central Police University where he received a ba chelors degree in forensic science in 1994 and started undergraduate research on glass evidence analys is by ICP-AES under the supervision of Dr. Chien-Min Hsu. After three years of worki ng experience in the fore nsic science field, he came back to Central Police University to purs e his masters degree in 1997 and focused his research on amphetamines analysis by SPE a nd GC/MS under the direction of Dr. Sheng-Meng Wang. After graduating from Central Police University, he worked for the Forensic Science Center of the Criminal Investigation Bureau in Taiwan, which is responsi ble for the investigation of major crimes nationwide. With the Bureau, he served as a forensic expert and police officer while performing crime scene investigations and fo rensic evidence analyses. In addition to his laboratory experience, he also sp ent a few years working with pract ical crime case investigation. In 2004, he married Rosalind Yi-Chun Lin in Taipei During this time, he was selected for the students studying abroad by the Taiwanese government, sponsored by the Ministry of Education, to study for his Ph.D. in the United St ates. In fall 2006, he chose to pursue his Ph.D. degree in analytical chemistry at the University of Florida and joined the Yost Lab. He received his PhD degree in August 2009 under the supervision of Dr. Richard A. Yost.