Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Analysis of polycyclic aromatic hydrocarbons by chemical ionization and ion trap mass spectrometry Mosi, Andrew A. 1998

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

Item Metadata

Download

Media
831-ubc_1998-345939.pdf [ 9.71MB ]
Metadata
JSON: 831-1.0061532.json
JSON-LD: 831-1.0061532-ld.json
RDF/XML (Pretty): 831-1.0061532-rdf.xml
RDF/JSON: 831-1.0061532-rdf.json
Turtle: 831-1.0061532-turtle.txt
N-Triples: 831-1.0061532-rdf-ntriples.txt
Original Record: 831-1.0061532-source.json
Full Text
831-1.0061532-fulltext.txt
Citation
831-1.0061532.ris

Full Text

ANALYSIS OF POLYCYCLIC AROMATIC HYDROCARBONS BY CHEMICAL IONIZATION AND ION TRAP MASS SPECTROMETRY by Andrew A. Mosi B.Sc., University of Victoria, 1992 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in T H E F A C U L T Y OF G R A D U A T E STUDIES Department of Chemistry We accept this thesis as conforming to the required standard: THE UNIVERJSJFY OF BRITISH C O L U M B I A August 1998 © Andrew. A. Mosi, 1998 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of The University of British Columbia Vancouver, Canada DE-6 (2/88) 11 ABSTRACT The aim of this thesis is to develop and explore the use of novel mass spectrometric techniques for the isomer specific analysis of PAHs and the detection of P A H transformation products in environmental samples. The first objective was achieved by using chemical ionization (CI) mass spectrometric techniques. Negative CI with C 0 2 and N 2 0 as reagent gases resulted in the formation of isomer specific adducts, such as [M + G - H]", between P A H molecules (M) and gas phase negative ions (G), enabling the differentiation of several P A H structural and positional isomers. In positive chemical ionization mode, on a linear quadrupole mass spectrometer, isomer differentiation was obtained with dimethyl ether ( C 2 H 6 0 ) as a reagent gas. However, the same results could not be generated using an ion trap mass spectrometer. Consequently, other chemical ionization reagents were investigated for their potential to differentiate PAH isomers in an ion trap. Halogenated hydrocarbons, such as CH 2C1 2 and C H 3 C H F 2 , proved to be the most successful chemical ionization reagents for this task. These substances formed adduct ions of the type [M + R]+ and [M + R - H X ] + (M = P A H molecule, R = reagent ion and X = Cl or F). Isomer identification was possible based on variations in the type and abundance of adducts formed. Application of these chemical ionization techniques to a contaminated environmental sample enabled the differentiation of several P A H isomers, including alkyl substituted ones. Differentiation of P A H isomers was also achieved through the use of ion/molecule reactions with mass-selected fluorocarbons, such as C3F 5 +, generated from the electron ionization of perfluorotributylamine ( F C 4 3 ) . Ul The second objective, the analysis of P A H transformation products, was achieved via GC/MS/MS experiments in an ion trap mass spectrometer. By using optimized collision induced dissociation conditions, several polyaromatic quinones that were undetectable by standard GC/MS analyses, were identified in a contaminated sediment sample collected from an aluminum smelter effluent lagoon. iv TABLE OF CONTENTS ABSTRACT ii LIST OF TABLES vii LIST OF FIGURES ix LIST OF ABBREVIATIONS xiv ACKNOWLEDGMENTS xvii DEDICATION xviii CHAPTER 1: INTRODUCTION 1 1.1 G E N E R A L INTRODUCTION 2 1.2 POLYCYCLIC AROMATIC HYDROCARBONS 4 1.2.1 Formation and Origins 4 1.2.2 Structural Characteristics 5 1.2.3 Chemical and Physical Properties of PAHs 10 1.2.4 Toxicological Properties of PAHs 12 1.3 ANALYSIS OF PAHs IN ENVIRONMENTAL SAMPLES 15 1.4 GAS CHROMATOGRAPHY 19 1.5 MAS S SPECTROMETRY 21 1.5.1 Ion Trap Mass Spectrometry 27 1.6 COMPOUND IDENTIFICATION 31 1.7 OBJECTIVES AND OVERVIEW OF THESIS 33 CHAPTER 2: DIFFERENTIATION OF PAH ISOMERS USING NEGATIVE CHEMICAL IONIZATION MASS SPECTROMETRY 35 2.1 INTRODUCTION 36 2.2 EXPERIMENTAL 41 2.2.1 Materials, Standards and Reagents 41 2.2.2 Environmental Samples 41 2.2.3 Instrumentation 42 2.2.4 Sample Introduction 43 2.3 RESULTS AND DISCUSSION 44 2.3.1 Desorption Negative Chemical Ionization Analysis 44 2.3.1.1 Structural P A H isomers 44 2.3.1.2 Positional P A H isomers 50 2.3.1.3 Limitations of DCI analysis 53 2.3.2 Gas Chromatography Negative Chemical Ionization Analysis 56 2.3.2.1 Using differences in relative response factors for isomer differentiation 61 2.3.2.2 Mass spectral data 66 2.3.2.3 Reactions with gas phase reagent ions 71 2.4 SUMMARY 79 CHAPTER 3: INVESTIGATING THE USE OF DIMETHYL ETHER AS A CHEMICAL IONIZATION REAGENT FOR THE ANALYSIS OF PAHs 80 3.1 INTRODUCTION 81 3.2 EXPERIMENTAL 83 3.2.1 Materials 83 3.2.2 Instrumentation 83 3.2.3 Sample Introduction 90 3.3 RESULTS AND DISCUSSION 91 3.3.1 D M E Chemical Ionization in a Standard Linear Quadrupole MS Ion Source 91 3.3.2 D M E Chemical Ionization in an Ion Trap Mass Spectrometer 100 3.3.3 Investigating the Ion/Molecule Reactions Between D M E Ions and PAHs 5a and 5b 105 3.4 S U M M A R Y 110 CHAPTER 4: DIFFERENTIATION OF PAH ISOMERS BY USING CHEMICAL IONIZATION AND ION/MOLECULE REACTIONS WITH HALOGENATED HYDROCARBONS IN AN ION TRAP MASS SPECTROMETER I l l 4.1 INTRODUCTION 112 4.2 EXPERIMENTAL 116 4.2.1 Standards and Reagents 116 4.2.2 Apparatus 116 4.2.3 Procedures 117 4.2.4 Data Analysis 122 4.3 RESULTS AND DISCUSSION 123 4.3.1 Formation of Reagent Ions 123 4.3.2 Chemical Ionization Reactions 126 4.3.2.1 Halomethane reagents 126 4.3.2.2 Haloethane reagents 132 4.3.3 Mechanism of Adduct Formation 141 4.3.4 Investigating the Mechanism for Isomer Differentiation 143 4.3.5 MS/MS Fragmentation of Adducts 151 4.3.6 Effects of Chemical Ionization Reaction Time on Adduct Formation 154 4.3.7 Effects of Analyte Concentration on Adduct Formation 156 4.3.8 Linear Dynamic Range of CI versus EI Analysis 159 4.4 SUMMARY 162 CHAPTER 5: DIFFERENTIATION OF PAH ISOMERS IN A CONTAMINATED SEDIMENT EXTRACT BY USING 1,1-DD7LUOROETHANE CHEMICAL IONIZATION ION TRAP MASS SPECTROMETRY 163 5.1 INTRODUCTION 164 5.2 EXPERIMENTAL 166 5.2.1 Standards and Reagents 166 5.2.2 Environmental Samples 166 5.2.3 Apparatus and Procedures 167 5.2.4 Data Analysis 167 5.3 RESULTS AND DISCUSSION 168 5.3.1 P A H Homologue Series in Sediment Extract 168 vi 5.3.2 Adduct Formation and Isomer Differentiation 171 5.3.3 Differentiation of Isomers Through the Comparison of Mass Spectral Data 175 5.3.4 Differentiation of Isomers from the Analysis of Mass Chromatograms 177 5.3.5 Mass Chromatogram Transformations 183 5.4 ANALYSIS OF ISOMERS IN REGIONS 2 TO 4 186 5.4.1 Region 2 186 5.4.2 Region 3 191 5.4.3 Region 4 196 5.5 QUANTITATION 200 5.6 SUMMARY 205 CHAPTER 6: ION / MOLECULE REACTIONS OF PERFLUOROTRTOUTYLAMINE CATIONS WITH POLYCYCLIC AROMATIC HYDROCARBONS IN A QUADRUPOLE ION-TRAP 206 6.1 INTRODUCTION 207 6.2 EXPERIMENTAL 208 6.2.1 Standards and Reagents 208 6.2.2 Apparatus 208 6.2.3 Procedure 208 6.3 RESULTS AND DISCUSSION 211 6.3.1 Formation and Isolation of FC43 Reagent Ions 211 6.3.2 Reactivity of Ionized FC43 213 6.3.3 Application to Environmental Analysis 226 6.4 S U M M A R Y 230 CHAPTER 7: ANALYSIS OF PAH OXIDATION PRODUCTS IN A COMPLEX ENVIRONMENTAL MA TREK BY USING GAS CHROMATOGRAPHY ION TRAP TANDEM MASS SPECTROMETRY 231 7.1 INTRODUCTION 232 7.2 EXPERIMENTAL 236 7.2.1 Reagents 236 7.2.2 Instrumentation 236 7.2.3 Analytical Procedures 236 7.3 RESULTS AND DISCUSSION 240 7.3.1 Optimization of CID Conditions 240 7.3.2 Qualitative Analysis of Contaminated Sediment for AQ 245 7.3.2.1 Analysis for AQ 246 7.3.2.2 Analysis for Cl-AQs 248 7.3.2.3 Analysis for C2-AQs 252 7.3.2.4 Analysis of Other PAH Oxidation Products 255 7.3.3 Quantitation 260 7.4 S U M M A R Y 263 CHAPTER 8 CONCLUSIONS 264 REFERENCES 268 Vll L I S T O F T A B L E S Table 1.1(a) Summary of important properties of unsubstituted PAHs 7 Table 1.1(b) Huckel molecular orbital energy values of selected PAHs 8 Table 1.2 Carcinogenic properties of selected PAHs 15 Table 2.1 Major instrumental parameters used during NCI experiments 42 Table 2.2 N 2 0 negative DCI mass spectral data for methylbenz[a]anthracenes (MBAs) 51 Table 2.3 N 2 0 negative DCI mass spectral data for dimethylbenz[a]anthracenes (DMBAs).. 52 Table 2.4 Mass spectral data (% RI) for PAHs analyzed by C 0 2 NCI 67 Table 2.5 Mass spectral data (% RI) for PAHs analyzed by N 2 0 NCI 68 Table 2.6 Mass spectral data (% RI) for PAHs analyzed by N2O/CH4 NCI 69 Table 2.7 Mass spectral data (% RI) for PAHs analyzed by CO2/CH4 NCI 70 Table 2.8 Reagent gas ions observed during NCI 71 Table 2.9 Mass spectral data for the GC/NCI-MS analysis of 16 PAHs using N2O-CH4 with the m/z 17 (OH") ion maximized 77 Table 3.1 Operating parameters for the Delsi-Nermag RIO-10C MS employed during dimethyl ether chemical ionization experiments 83 Table 3.2 Major operating parameters employed during dimethyl ether CI in the Saturn 3D ion trap 84 Table 3.3 Intensities of ions generated from electron ionization of D M E at various ion source pressures 91 Table 3.4 Dimethyl ether positive chemical ionization mass spectra of PAHs 92 Table 3.5 Dimethyl ether positive chemical ionization mass spectra of MBAs and DMBAs . 95 Table 3.6 Dimethyl ether positive chemical ionization mass spectra of PAHs identified in the B-lagoon sediment extract 98 Table 3.7 Mass spectral data extracted from the chromatogram in Figure 3.4 103 Table 3.8 Mass spectral data (% RI) for the ion/molecule reaction products between each of the three mass-selected ions (m/z 45, 47, 61) and phenanthrene and anthracene ... 107 Table 3.9 Ionization energy (IE) and proton affinity (PA) values for the major reaction species investigated 108 Table 4.1 List of major experimental parameters employed during the chemical ionization experiments under ARC conditions 120 Table 4.2 Major experimental parameters employed during the MS/MS CID experiments. 121 Table 4.3 Major reagent ions produced during electron ionization of halocarbons 123 Table 4.4 Major chemical ionization products, m/z (% RI), formed between halomethanes ions (R) and selected P A H isomers (M) 131 Table 4.5 Major products, m/z (% RI), formed between 1,1-dichloroethane ions (R = C H 3 C H C I ) and selected P A H isomers (M) 136 Table 4.6 Major products, m/z (% RI), formed between 1,1-difluoroethane ions (R) and selected P A H isomers (M) 137 1 V l l l Table 4.7 Haloethane chemical ionization [M + R - HX]7[M + R] + % RI ratios 140 Table 4.8 CDC13 CI mass spectral data (m/z and (% RI)) for anthracene (5a), phenanthrene (5b) and their 9,10-dihydro analogues. R = CDC1, 144 Table 4.9(a) Ionization energies (DE) for the formation of R + ions and proton affinities (PA) for halocarbon ions 148 Table 4.9(b) Ionization energies (LE) and proton affinities (PA) for PAHs 149 Table 4.10 Mass spectral data from the CID (60 Volts, 10 ms, Non-resonant) of mass-selected [M+CDC12]+ adducts of phenanthrene (5b) and anthracene (5a) 153 Table 4.11 Mass spectral data from the CID (55 Volts, 10 ms, Non-resonant) of mass-selected [M+R]+ adducts of phenanthrene (5b) and anthracene (5a) 153 Table 4.12 Mass spectral data from the CLD (55 Volts, 10 ms, Non-resonant) of mass-selected [M+R-HF]+ adducts of phenanthrene (5b) and anthracene (5a) 153 Table 4.13 Effect of sample concentration on the percent relative intensities for major ions observed during chemical ionization with 1,1-difluoroethane 157 Table 5.1 1,1-DFE CI mass spectral data for selected PAH isomers from the settling pond extract 170 Table 5.2(a) Effect of methyl substitution on adduct formation under 1,1-DFE CI 172 Table 5.2(b) Ratio of % relative intensity values for [M+R-HF]+ and [M+R]+ adducts....... 173 Table 5.3 Summary of linear regression parameters for data plotted in Figure 5.17" 201 Table 5.4 Relative response factors (RRF), with respect to 5a, for standard PAHs analyzed by 1,1-DFE CI 203 Table 5.5 Concentration, in ppm, (u.g/g) of selected PAHs present in sediment sample determined using 1,1-DFE chemical ionization, together with concentration values determined by an independent analysis employing GC/MS EI 204 Table 6.1 List of major experimental parameters employed during "regular" and "selective" chemical ionization experiments 209 Table 6.2 Ions generated from the EI ionization of FC43 and their MS/MS isolation efficiency 212 Table 6.3 Effects of reaction time between CsF 5 + and PAHs 5a and 5b on the relative intensity of ions produced 218 Table 6.4(a) Mass spectral data (% RI) for ion/molecule reactions between individually mass selected FC43 ions (R+) and anthracene (5a) 219 Table 6.4(b) Mass spectral data (% RI) for ion/molecule reactions between individually mass selected FC43 ions (R+) and phenanthrene (5b) 220 Table 6.5 Mass spectral data (% RI) for the reaction between C3F 5 + (m/z 131) and PAHs 5a to 9e 222 ix LIST OF FIGURES Figure 1.1 Structural formulas for unsubstituted PAHs 6 Figure 1.2 Structural formulas for selected substituted and heteroaromatic polycyclic aromatic compounds 9 Figure 1.3 Localization energy values ((5 units) on benz[a]anthracene (8b) 11 Figure 1.4 General schematic of a mass spectrometer 22 Figure 1.5 General schematic for a mass spectrometer ion source 22 Figure 1.6(a) Time of flight analyzer 25 Figure 1.6(b) Magnetic sector analyzer 25 Figure 1.6(c) Quadrupole analyzer 26 Figure 1.6(d) Quadrupole ion trap analyzer 26 Figure 1.7 Stability diagram for a three dimensional quadrupole ion trap 30 Figure 2.1(a) C 0 2 negative DCI mass spectra of benz[a]anthracene (8b), chrysene (8c), triphenylene (8d) and tetracene (8e) 45 Figure 2.1(b) N 2 0 negative DCI mass spectra of benz[a]anthracene (8b), chrysene (8c), triphenylene (8d) and tetracene (8e) 46 Figure 2.2 Effect of C 0 2 pressure (0.1 mbar and 1.3 mbar) on the negative DCI mass spectra of benz[a]anthracene (8b) 47 Figure 2.3 Experimental blank DCI filament desorption and calculated W O x isotope distribution 47 Figure 2.4 GC/ NCI-MS mass spectra for benz[a]anthracene (8b) and chrysene (8c) using a) C 0 2 and b) N 2 0 reagent gases 50 Figure 2.5 Negative ion N 2 0 DCI analysis of (a) "99+% zone refined" anthracene, (b) dio-anthracene; (c) GC/MS analysis of anthracene 55 Figure 2.6 Relative response factors (RRFs) of 16 standard PAHs analyzed by GC/NCI-MS using different reagent gases. (T = 240°C, eV = 85, Ie = 0.200 mA) 57 Figure 2.7 Effect of ion source temperature on the N 2 0/NCI analysis of the 16 standard PAHs. (a) 175°C, (b) 280"C 58 Figure 2.8 Effect of filament emission current on the NCI/N 2 0 analysis of the 16 standard PAHs. (a) 0.250mA, (b) 0.200mA 60 Figure 2.9 Comparison of a portion of the N 2 0/NCI (top) and positive EI (bottom) total ion chromatograms from the analysis of the B-lagoon sediment extract 62 Figure 2.10 Comparison of a portion of the N 2 0/NCI (top) and positive EI (bottom) total ion chromatograms from the analysis of the B-lagoon sediment extract 64 Figure 2.11 (a) EI, (b) CH4 /PCI, and (c) N 20-CH4/NCI TICs for a P A H contaminated sediment extract 65 Figure 2.12 N 2 0/NCI mass spectral data for three PAH isomer pairs 66 Figure 2.13 Ion chromatograms from the G C / C 0 2 NCI-MS analysis of 16 PAHs: m/z 16 chromatogram (top) and m/z 100-350 ion chromatogram (bottom) 72 Figure 2.14 Ion chromatograms from G C / N 2 0 NCI-MS analysis of 16 PAHs: top to bottom, m/z 30, m/z 16, m/z 32 and m/z 100-350 ion chromatograms 74 Figure 2.15 Expanded view of 14.8 - 16.3 min region of Figure 2.14 75 Figure 2.16 Total ion chromatogram for the GC/NCI-MS analysis of 16 PAHs using N 2 0 -CH4 with the m/z 17 (OH") ion maximized 78 Figure 3.1 Possible mechanisms for the reaction of anthracene (5a) or phenanthrene (5b) with C H 2 O C H 3 ions 82 Figure 3.2(a) Ion trap stability diagram indicating the m/z and (qz, az) values at different points on the d.c. / r.f. = 67 /428 line 88 Figure 3.2(b) Scan functions employed for the isolation of ions using the QISMS trap 89 Figure 3.3 GC/MS D M E CI total ion chromatograms for the analysis of (a) a standard PAH mixture and (b) a P A H contaminated sediment extract 97 Figure 3.4 GC-DME-CI/MS total ion chromatogram from the analysis of a standard 16 PAH mixture by ion trap MS 102 Figure 4.1 GC/EI-MS total ion chromatogram of an aluminum smelter effluent lagoon extract analyzed using an ion trap mass spectrometer 115 Figure 4.2 Standard scan function, with an ARC pre-scan, used for chemical ionization in the Saturn ion trap mass spectrometer 118 Figure 4.3 Schematic of standard MS/MS Scan function employed for mass-isolation and CID experiments 121 Figure 4.4 Effects of (a) reaction time (msec) (b) ionization time (u.sec) and (c) pressure (TIC) on m/z 47, 51 and 65 ions from 1,1-difluoroethane 124 Figure 4.5 Dichloromethane chemical ionization mass spectra of (a) anthracene 5a and (b) phenanthrene 5b 127 Figure 4.6(a) Relative intensity of adduct ions formed by the P A H isomer pair 5a/5b during chemical ionization with halomethanes 128 Figure 4.6(b) Relative intensity of adduct ions formed by the P A H isomer pair 6a/6b during chemical ionization with halomethanes 128 Figure 4.6(c) Relative intensity of adduct ions formed by the PAH isomer group 8b to 8d during chemical ionization with halomethanes 129 Figure 4.6(d) Relative intensity of adduct ions formed by the PAH isomer group 9a to 9e during chemical ionization with halomethanes 129 Figure 4.7 1,1-Difluoroethane chemical ionization mass spectra of (a) anthracene 5a and (b) phenanthrene 5b 132 Figure 4.8(a) Relative intensity of adduct ions formed by the P A H isomer pair 5a and 5b during chemical ionization with CH 3 CHC1 2 and CH3CHF2 134 Figure 4.8(b) Relative intensity of adduct ions formed by the P A H isomer pair 6a and 6b during chemical ionization with CH 3 CHC1 2 and C H 3 C H F 2 134 Figure 4.8(c) Relative intensity of adduct ions formed by the P A H isomer group 8b to 8e during chemical ionization with CH 3 CHC1 2 and C H 3 C H F 2 135 xi Figure 4.8(d) Relative intensity of adduct ions formed by the P A H isomer group 9a to 9e during chemical ionization with CH 3 CHC1 2 and C H 3 C H F 2 135 Figure 4.9 Proposed mechanism for the reaction between 5b and CHC1 2 + 142 Figure 4.10 Effect of P A H ionization energy (IE) on the formation of [M+R-HX]+ elimination products with chloromethane reagent ions 146 Figure 4.11 Effect of P A H ionization energy (IE) on the formation of [M+R-HX]+ elimination products with haloethane reagent ions 146 Figure 4.12 Effect of P A H ionization energy on the formation of Ivf and M H + in the presence of haloethane reagent ions 150 Figure 4.13 Effect of CI reaction time on formation of (a) [M+R]+, (b) [M+R-HF]+ 155 Figure 4.14 Effect of mole fraction composition of phenanthrene / anthracene on m/z 205 [M+ CH 3 CHF-HF] + during difluoroethane chemical ionization 157 Figure 4.15 1,1 -Difluoroethane CI mass spectral data at shoulder and apex of chromatographic elution profile for 5 ng each of phenanthrene (5b) and anthracene (5a) 158 Figure 4.16 Chromatographic integration areas of selected ions of pyrene (6a) and fluoranthene (6b) analyzed by 1,1-difluoroethane chemical ionization under ARC conditions 160 Figure 4.17 Chromatographic integration areas of molecular ions of pyrene (6a) and fluoranthene (6b) analyzed by electron ionization 161 Figure 4.18 TIC integration areas of pyrene (6a) and fluoranthene (6b) analyzed by 1,1-difluoroethane chemical ionization under non-ARC conditions 161 Figure 5.1 Total ion chromatogram of sediment extract obtained using 1,1-difluoroethane chemical ionization showing the four isomer regions of interest. C l = monomethyl PAHs, C2 = dimethyl / ethyl PAHs 169 Figure 5.2 Mass spectra for Cl-5a/5b and C2-5a/5b extracted from chromatogram in Figure 5.1 176 Figure 5.3 Unsubstituted (CO) 5b/5a and 6b/6a ion chromatograms from 1,1-DFE chemical ionization analysis of the sediment extract; bottom: transformation chromatogram. 178 Figure 5.4 Monomethyl (Cl) 5b/5a ion chromatograms and transformation chromatogram from 1,1-DFE chemical ionization analysis of sediment extract 179 Figure 5.5 Dimethyl / ethyl (C2) 5b/5a ion chromatograms and transformation chromatogram from 1,1-DFE chemical ionization analysis of sediment extract 180 Figure 5.6 % RI of [M+27] and [M+47] adducts at selected regions indicated on the chromatograms displayed in Figures 5.4 to 5.6 181 Figure 5.7 Dimethyl / ethyl (C2) 5b/5a sum and m/z 206-m/z 207, (M-[M+1]), transformation chromatogram from 1,1-DFE chemical ionization analysis of sediment extract 185 Figure 5.8 Sum and transformation chromatograms for unsubstituted (CO), monomethyl (Cl) and dimethyl / ethyl (C2) 5a / 5b and 7a/c isomers 189 Figure 5.9 % RI of [M+27] and [M+47] adducts at selected regions indicated on the chromatograms displayed in Figure 5.8 190 X l l Figure 5.10 Ion and transformation chromatograms for unsubstituted (CO) 8a to 8e PAHs (M = 228 u) 192 Figure 5.11 Ion and transformation chromatograms for monomethyl (Cl) 8a to 8e PAHs (M - 242 u) 194 Figure 5.12 Ion and transformation chromatograms for dimethyl / ethyl (C2) 8a to 8e PAHs (M = 256u) 195 Figure 5.13 % RI of [M+27] and [M+47] adducts at selected regions indicated on the chromatograms displayed in Figure 5.10 to 5.12 196 Figure 5.14 Ion and transformation chromatograms for unsubstituted (CO) 9a to 9e PAHs (M= 252 u) 197 Figure 5.15 Ion and transformation chromatograms for monomethyl (Cl) 9a to 9e PAHs (M = 266 u) 198 Figure 5.16 % RI of [M+27] and [M+47] adducts at selected regions indicated on the chromatograms displayed in Figures 5.14 and 5.15 199 Figure 5.17 GC integration area versus injection amount relationship for EI and CI analysis of fluoranthene (6b) 200 Figure 6.1 Scan function employed for mass- isolation and reaction of FC43 fragment ions210 Figure 6.2 FC43 chemical ionization mass spectra of phenanthrene (5b) and anthracene (5a) 213 Figure 6.3 C 3 Fs + ion / molecule reaction spectra of anthracene (5a) and phenanthrene (5b) 214 Figure 6.4 Effect of analyte concentration, along a chromatographic elution band, on the relative intensity of ions produced during ion/molecule reactions with C 3 F 5 + 217 Figure 6.5 Effect of P A H ionization energy (IE) on the formation of (a) [M+C 3F 5] + and (b) M^from ion/molecule reactions with C 3 F 5 + ions 223 Figure 6.6 Effect of benzo interactions (*) on the relative abundance of [M+C 3F 5] + products for isomers 8b, 8c, 8d and 8e 225 Figure 6.7 Analysis of B-lagoon sediment extract with mass-selected C 3 F 5 + ions. Extracted M , [M+49], [M+91], [M+lll] and [M+131] ion chromatograms for phenanthrene (5b) and anthracene (5a) and monomethyl phenanthrenes and anthracenes 227 Figure 6.8 Analysis of B-lagoon sediment extract with mass-selected C 3 F 5 + ions. Transformation ion chromatograms for phenanthrene (5b) and anthracene (5a) and monomethyl phenanthrenes and anthracenes 228 Figure 6.9 Mass spectra from the analysis of B-lagoon sediment extract with mass-selected C 3 F 5 ions 229 Figure 7.1 Polyaromatic ketones and quinones 233 Figure 7.2 Ion trap MS/MS mass-isolation and CED scan function 239 Figure 7.3 Decomposition curves for anthraquinone molecular ion 241 Figure 7.4 Decomposition curves for 2-methylanthraquinone molecular ion 242 Figure 7.5 Possible fragmentation pathways for AQ and 2-MeAQ 243 Figure 7.6 Mass spectral data for AQ and 2-MeAQ standards during non-resonant (NR) and resonant (Res) CID 244 Xlll Figure 7.7 GC/EI-MS total ion chromatogram of B-lagoon sediment extract, with major PAHs and PAQs identified 245 Figure 7.8 GC/MS and GC/MS/MS analysis of sediment extract, (a) TIC and m/z 208, 180, 152 ion chromatograms, (b) background subtracted mass spectra at tR = 19.38 min in (a) 247 Figure 7.9 TIC and m/z 222 (M), 207 (M-CH 3), 194 (M-CO), 179 (M-CO,CH 3 ), 166 (M-2CO), 165 (M-2CO,H), and 151 (M-2CO,CH 3) ion chromatograms for sediment extract in GC/MS and GC/MS/MS non-resonant (NR) and resonant (Res) modes.249 Figure 7.10 Mass spectral data at 20.89 min and 21.55 min for C l - A Q analysis using (a) GC/MS, (b) non-resonant GC/MS/MS and (c) resonant GC/MS/MS 250 Figure 7.11 TIC and m/z 236 (M), 221 (M-CH 3), 208 (M-CO), 207 ( M - C 2 H 5 or M-CO,H), 193 (M-CO,CH 3 ), 180 (M-2CO), 179 (M-2CO,H), 165 (M-2CO,CH 3) ion chromatograms for sediment extract in GC/MS and GC/MS/MS non-resonant analyses 253 Figure 7.12 Mass spectral data at 23.1 min and 23.6 min for analysis of C2-AQs using (a) GC/MS and (b) non-resonant GC/MS/MS 254 Figure 7.13 TIC and m/z 258, 230, 202 ion chromatograms for sediment extract in GC/MS and GC/MS/MS non-resonant modes 256 Figure 7.14 Mass spectral data at 28.5 min and 29.6 min for analysis of M r 258 quinones using (a) GC/MS and (b) non-resonant GC/MS/MS 257 Figure 7.15 (a) TIC and m/z 180, 152 ion chromatograms for sediment extract in GC/MS and GC/MS/MS non-resonant modes, (b) mass spectra at 15.26 min 259 Figure 7.16 Response curves for analysis of AQ by non-resonant (NR) and resonant (R) GC/MS/MS 261 Figure 7.17 Concentration of quinones and their P A H analogues present in the sediment sample analyzed 261 LIST OF ABBREVIATIONS a.c. alternating current b.p. boiling point ASE Accelerated Solvent Extraction CO unsubstituted C l monomethyl C2 dimethyl or ethyl CI Chemical Ionization CFC Chlorofluorocarbon CID Collisionally Induced Dissociation d.c. direct current DCI Desorption Chemical Ionization D C E Dichloroethane D C M Dichloromethane DFE 1,1 -Difluoroethane DMBA Dimethylbenz[a]anthracene D M E Dimethyl ether eth thermal electron E A Electron Affinity EI Electron Ionization EI/MS Electron ionization mass spectrometry FC43 Perfluorotributylamine GC Gas Chromatography GC/CI-MS Gas Chromatography Chemical Ionization Mass Spectrometry GC/EI-MS Gas Chromatography Electron Ionization Mass Spectrometry GC/MS Gas Chromatography / Mass Spectrometry GC/MS/MS Gas chromatography tandem mass spectrometry GC/NEI-MS Gas Chromatography Negative Electron Ionization Mass Spectrometry GC/PEI-MS Gas Chromatography Positive Electron Ionization Mass Spectrometry H M O Huckel Molecular Orbital HOMO Highest Occupied Molecular Orbital IE Ionization Energy IT Ion Trap IT/MS Ion Trap Mass Spectrometry L c Localization energy L C Liquid chromatography L U M O Lowest Unoccupied Molecular Orbital M+ Molecular ion m.p. melting point m/z mass to charge ratio M B A Methylbenz[a]anthracene M H + Pseudomolecular ions M r Molecular mass MS Mass Spectrometry MS/MS Mass Spectrometry / Mass Spectrometry NCI Negative Chemical Ionization NEI Negative Electron Ionization ppb parts per billion ppm parts per million ppt parts per trillion PA Proton Affinity PAC Polycyclic Aromatic Compound P A H Polycyclic Aromatic Hydrocarbon PAQ Polycyclic Aromatic Quinone PAK Polycyclic Aromatic Ketone PEI Positive Electron Ionization PCI Positive Chemical Ionization r correlation coefficient r.f. radio frequency RF Response Factor RA Relative Abundance RI Relative Intensity Reconstructed Ion Chromatogram or Reconstructed Ion Count Relative Response Factor Total Ion Chromatogram or Total Ion Count Atomic mass unit Ultraviolet xvu ACKNOWLEDGMENTS I would like to express my sincere gratitude to my research supervisors Drs. G.K. Eigendorf and W.R. Cullen for their guidance, assistance and insights during the course of my Ph.D. program. I would like to thank three fine mass spectrometrists, Lina Madilao, Chris Edmond and Marshall Lapawa, for their assistance and friendship during the many hours spent at the UBC mass spectrometry laboratories. My gratitude goes to Mr. Rob Christian and Dr. Colin Jennison from Varian for guidance and assistance in solving problems with the Saturn ion trap. Although it may not be clearly reflected in the body of this thesis, some of the most valuable experiences I gained during the last five years have been from the numerous environmental field trips made possible by Dr. Ken Reimer from the Environmental Sciences Group at the Royal Canadian Military College in Kingston, Ontario. I would like to acknowledge all the past and present members of Dr. Cullen's research group for their friendship and assistance during the last five years. In particular I would like to thank Dr. Chris Simpson for his initial guidance in the field of environmental analysis, and the numerous intellectual discussions on polycyclic aromatic hydrocarbons. Last but not least I would like to thank my wife Renee for her patience and help with proofreading my thesis. XV111 / dedicate this thesis to my parents, Delia and Virgilio, for all their support and sacrifices that have enabled me to be where I am today and to Renee for her love and support during our long journey through university. CHAPTER 1 INTRODUCTION 2 1.1 G E N E R A L INTRODUCTION A network of complex interactions between living organisms and chemical substances is found in the natural environment. In air, water, soil and food, living organisms are constantly involved in a dynamic uptake, metabolism and excretion of chemicals. Millions of different chemical substances are present in the environment and during the last century chemists have been able to elucidate the structure and properties of many of these. In the complex network of chemical interactions on our planet many substances can play either a direct or indirect role in sustaining life. For example, oxygen is directly responsible for the minute by minute existence of most organisms, while ozone, an allotrope of oxygen, plays an indirect role in sustaining life on Earth by absorbing harmful ultraviolet radiation in the stratosphere before it can reach our environment. Although many chemical substances are important for sustaining life on the planet, many are also harmful. Ozone, a highly reactive gas, plays an essential role in the stratosphere, but is harmful when present in the troposphere. Most substances present in the environment are of natural origin while a small percentage originate from anthropogenic sources. Many anthropogenic substances, such as polychlorinated biphenyls, have been synthesized deliberately by humans, while others, such as polychlorinated dibenzo-p-dioxins, may be produced inadvertently through human activities. As anthropogenic substances are produced and enter the environment, concerns may arise about their toxicity, or other deleterious environmental effects they may cause. However, because many substances do not always display immediate visible side effects, their long term impact on the environment is often not considered. Humans generally ignore issues that pose only potential risks in the long term, especially if confronting these 3 issues means sacrificing short term gains. The use of chlorofluorocarbons (CFCs) for refrigeration and industrial applications provides a classic example of the difficulty in dealing with long term environmental hazards. Initially, because of their chemical stability and lack of toxicity, CFCs did not appear to pose any hazards. However, through the research of Rowland and Molina, who first unveiled the link between stratospheric ozone depletion and CFCs they were shown to have potential deleterious effects on the planet's stratospheric ozone layer. Therefore, when investigating the long term effects of chemicals in such a complex system as the natural environment, it is important to realize that lack of evidence may be merely a consequence of lack of adequate knowledge of the system. In the past, our understanding of the role many chemicals played in the environment was limited by our ability to detect and identify them at low concentrations in complex mixtures. The development of modern analytical techniques that can separate and detect minute quantities (nanograms or less) of specific substances has enabled a much more detailed investigation and understanding of environmental pollutants. Although humans have introduced a large number of potentially deleterious substances into the environment, there are also many naturally occurring toxic compounds present. For example, aflotoxins, substances found in moldy peanuts and grain, are powerful carcinogens . In some instances, substances with toxic properties may be generated by both natural and anthropogenic processes, as is the case with polycyclic aromatic hydrocarbons (PAHs). This thesis will focus on two main areas of the analysis of PAHs. The main aim will be the development of analytical techniques for the differentiation of isomeric PAH compounds, and the second goal will be the development of an analytical method to detect 4 PAH oxidation products that are obscured by the presence of other more abundant substances. 1.2 POLYCYCLIC AROMATIC HYDROCARBONS 1.2.1 Formation and Origins PAHs can be formed through the diagenesis of organic material, such as during the maturation of petroleum, through combustion and pyrolysis processes, and through biosynthesis 3 . Because PAHs can arise from almost any incomplete combustion process, such as a forest fire, they have probably been present on Earth since its early formation. PAHs have even been detected in space 4 ' 5 and in meteorites originating from Mars 6 . In fact PAHs found in one such meteorite have been attributed to the remnants of past bacterial life on Mars 6 . Anthropogenic sources of PAHs include discharges from the following: (a) industrial processes, such as coke production, catalytic cracking of crude oil, and aluminum smelting, (b) fuel combustion from fossil-fueled power plants, residential space heating or gasoline/diesel powered vehicles and (c) solid waste incineration 3 . The structural features of PAHs formed by various processes can differ significantly. For example, PAHs found in fossil fuels are dominated by alkylated PAHs, while PAHs produced from combustion sources at very high temperatures (2000°C) are generally dominated by unsubstituted PAHs 3 . However, regardless of their origin, PAHs are generally produced as complex mixtures of isomers and congeners. The environmental and analytical chemistry of PAHs has been covered extensively in the literature 3 ' 7 1 1 5 thus only a brief overview with emphasis on matters that are of relevance to this thesis will be presented in this chapter. 1 . 2 . 2 Structural Characteristics Polycyclic aromatic hydrocarbons (PAHs) are molecules consisting of a series of fused aromatic rings, as shown in Figure 1.1. PAHs containing only six-member rings, such as pyrene (6a), are referred to as alternant, while PAHs containing also five-member rings, such as fluoranthene (6b), are referred to as non-alternant. Alternant PAHs can be further sub-classified as pericondensed or orthocondensed. The former comprise non-linear arrangements of benzene rings, such as 6a, while the latter includes all linear arrangements, such as anthracene (5a) and phenanthrene (5b). Names for the PAHs shown in Figure 1.1 are listed in Table 1.1(a) and are those commonly used in the literature. A few examples of substituted PAHs, such as 1-methylanthracene, and of heteronuclear PAHs such as carbazole, are illustrated in Figure 1.2. Heteronuclear PAHs contain at least one atom other than carbon in their ring structure. PAHs and heteronuclear PAHs are sometimes classified together as polycyclic aromatic compounds (PACs). The 5- or 6-membered rings in a PAH can be arranged in a variety of ways, leading to numerous structural isomers (constitutional isomers). The simplest examples of structural isomers are anthracene (5a) and phenanthrene (5b). The addition of a substituent to the basic ring structure can give rise to a class of structural isomers called positional isomers. For example, 1-methylanthracene (Figure 1.2) is only one of three possibl monomethyl anthracene isomers (1-methyl, 2-methyl and 9-methylanthracene). 8 1 6 5 6 5 Figure 1.1 Structural formulas for unsubstituted PAHs. 7 Table 1.1(a) Summary of important properties of unsubstituted PAHs PAH Name M r m.p.a' b.p.a IE EA 0 (u) (°C) (°C) 7 6 0 (eV) (eV) 1 Naphthalene 128 80.5 178 8.14+0.01 0.14+0.1 2 Acenaphthene 154 96.2 278 7.68 n/a 3 Acenaphthylene 152 92 270 8.22±0.04 0.40+0.03 4 Fluorene 166 116s 294 7.9 0.28 5a Anthracene 178 218s 342 7.45+0.03 0.55+0.05 5b Phenanthrene 178 100s 340 7.86+0.03 0.3 6a Pyrene 202 156 404 7.41 0.55+0.05 6b Fluoranthene 202 110 393 7.95+0.04 0.6 7a 1 lH-Benzo[b]fluorene 216 189 402f n/a n/a 7b 1 lH-Benzo[a]fluorene 216 208 407f n/a n/a 7c 7H-Benzo[c]fluorene 216 n/a 406f 8a B enzo [c] phenanthrene 228 68 n/a 7.60 0.545 8b Benz[a]anthracene 228 160s 435 7.43+0.03 0.630 8c Chrysene 228 256s 448 7.59+0.02 0.397 8d Triphenylene 228 199sb 425b 7.84+0.01 0.285 8e Tetracene (naphthacene) 228 341sb 357b 6.97+0.02 0.88 9a Benzo[b]fluoranthene 252 167 481f 7.9d 0.7d 9b B enzo [k] fluoranthene 252 216 481f 7.4d 0.6d 9c Benzo[e]pyrene 252 178 492 7.41 0.534 9d Benzo[a]pyrene 252 177 475 7.12+0.01 0.680 9e Perylene 252 273 350 6.90+0.01 0.7 10a Indeno[l,2,3-cd]pyrene 276 163 n/a 7.5d 0.8d 10b B enzo [ghi] perylene 276 277 n/a 7.15 0.51e 11 Dibenz[a,h]anthracene 278 266sb n/a 7.38+0.04 0.595 n/a not available, s = sublimes (a) Data obtained from the National Research Council of Canada 3 . (b) Data obtained from Merck Index ' 2 . ( c) Data obtained from Lias 1 3 (d) Calculated using a linear regression between EI or E A and H O M O or L U M O P values from Table 1.1(b). IE = 3.176PHOMO +5.946, r = 0.92; E A = 1.508(3LUMO + 1.241, r = 0.88 (e) Data obtained from Lide 1 4 (f) Data obtained from Onuska 1 5 Table 1.1(b) Huckel molecular orbital energy values of selected PAHs PAH Name HOMO a (3) LUMO b (3) 1 Naphthalene 0.618 -0.618 3 Acenaphthylene 0.638 -0.285 5a Anthracene 0.414 -0.414 5b Phenanthrene 0.605 -0.605 6a Pyrene 0.445 -0.445 6b Fluoranthene 0.618 -0.371 8a B enzo [c] phenanthrene 0.568 -0.568 8b B enz [a] anthracene 0.452 -0.452 8c Chrysene 0.520 -0.520 8d Triphenylene 0.684 -0.684 Se Tetracene 0.295 -0.295 9a Benzo[b]fluoranthene 0.602 -0.377 9b Benzo[k]fluoranthene 0.459 -0.400 9c Benzo[e]pyrene 0.497 -0.497 9d Benzo[a]pyrene 0.371 -0.371 9e Perylene 0.347 -0.347 10a Indeno [ 1,2,3 -cd]pyrene 0.475 -0.293 10b B enzo [ghi] perylene 0.439 -0.439 11 Dibenz [a, h] anthracene 0.474 -0.474 (a) Highest Occupied Molecular Orbital. Energy values in arbitrary p units calculated by using "Huckel" Version v. 1.1 software 1 6 (b) Lowest Unoccupied Molecular Orbital. Energy values also calculated with "Huckel" v. 1.1. 9 H 1 -methylanthracene 9,10-dimethylanthracene 3,6-dimethylphenanthrene 6-nitrobenzo[a]pyrene 7-aminobenz[a]anthracene Anthraquinone Figure 1.2 Structural formulas for selected substituted and heteroaromatic polycyclic aromatic compounds. The number of possible isomers increases dramatically with increasing molecular size. For example, there are four structural PAH isomers with the empirical formula C 1 6 H 1 0 and these can have 28 monomethyl positional isomers (C 1 7 H 1 2 ) . Adding one ring to give C 2 o H 1 2 increases the number of structural isomers to 16 and the number of monomethyl positional isomers to 150 1 7 . As for any series of structural isomers, the chemical and physical properties of P A H isomers differ. Thus, in theory, isomers should be separable and identifiable based on these differences. However, in practice, this is often not feasible, especially when only 1 0 small amounts of these compounds are present in a complex mixture such as a contaminated environmental sample. 1.2.3 Chemical and Physical Properties of PAHs The chemistry of PAHs is strongly influenced by their thermodynamic stability associated with the resonance stabilization of the ^-electrons in their conjugated aromatic ring systems. In order for a polycyclic system to be considered aromatic it must be planar 18 and contain (4n+2) Ti-electrons around the periphery of its ring structure . For example, pyrene (6a) has a total of 16 7t-electrons, an anti-aromatic number, but the periphery of its Kekule structure has 14 7t-electrons, an aromatic number. In fact, it has been shown that if the central double bond in pyrene is saturated by the addition of methyl groups, the resulting compound is still aromatic 1 9 . Aromatic systems favor substitution reactions over addition reactions, since in a substitution reaction the aromatic character of the molecule is retained in the products, 18 while in an addition reaction it is at least, partially destroyed . The chemistry of substitutions will tend to favor those reactions in which the disruption of the aromatic system is minimized. For polyaromatic systems it has been shown that the position of 20 substitution will often be determined by the localization energy (Lc) . The localization energy is a measure of the difference in % energy between the initial state and the transition 20 state during a substitution reaction . Consequently, substitutions will be favored in positions with the lowest localization energy. As an example, the L c values for 20 benz[a]anthracene (8b) are indicated in Figure 1.3 . 11 Figure 1.3 Localization energy values (j3 units) on benz[a]anthracene (8b) (position 7 = 1.35, position 12 = 1.44). The 7 and 12 positions on 8b have the lowest L c values, and therefore are the sites where substitution reactions are expected to be favored. In the context of this thesis two other important parameters are ionization energy (IE) and electron affinity (EA). From a thermodynamic prospective the IE of a molecule (M) is defined as the enthalpy change for the removal of an electron, in the gas phase, at absolute zero 1 3 . -> M(g)+ + e AH = I E M The EA of a molecule is defined as the negative of the 0 K enthalpy change for an electron attachment reaction 1 3 . M(g) + e - » M(g) AH = - E A M At a molecular orbital level, the ionization energy of a molecule represents the energy required to remove an electron from its highest occupied molecular orbital (HOMO), while the electron affinity represents the energy released upon addition of an electron to its lowest unoccupied molecular orbital (LUMO). The experimental values for the IE and EA of several PAHs are listed in Table 1.1(a). In cases where experimental IE and EA are not available, theoretical values can be estimated by using the relative energy values of the HOMO or L U M O molecular orbitals listed in Table 1.1 (b). The values in this table were 12 obtained using Huckel Molecular Orbital calculations 1 6 ' 2 1 . Using the relative HOMO and L U M O energy values from Table 1.1(b) and the experimental IE and EA energy values from Table 1.1(a) the following linear correlations were obtained: LE(eV) = 3.176|3HOMO + 5.946 r = 0.92 EA(eV)= 1.508PLUMO+1.241 r = 0.88 The lower correlation coefficient for the EA linear correlation is, in part, a result of the greater uncertainty associated with the E A values. Using these two correlations theoretical values of IE and EA were determined for 9a, 9b and 10a (Table 1.1(a)). 1.2.4 Toxicological Properties of PAHs It is generally accepted that environmental or nutritional factors are responsible for 22 as much as 90 % of the cumulative risk of developing cancer in humans . With the exception of U V radiation (e.g. sunlight), the exposure to chemical substances from food, smoking, or the environment is to blame for the majority of these cancers. Genetic factors, viral infections, and nuclear radiation are estimated to be responsible for only 10 22 % of the cumulative cancer risk . Consequently, exposure to chemical carcinogens is by no means a trivial issue. For many carcinogens there is virtually no safe level of exposure since, in theory, even a few molecules could initiate the cellular damage which eventually could lead to production of abnormal cells and cancer. Furthermore, most carcinogenic effects caused by chemicals have a long latency period, often 20 to 30 years, prior to development of tumors, making direct correlations between exposure and cancer very difficult. Consequently, limits of exposure to carcinogens are based upon risk levels. Usually "acceptable" exposure levels are established as a lifetime risk level of one in a 13 million. For example, a 1/106 lifetime risk level is equivalent to one million people being exposed to a substance during their lifetime at a particular "concentration" and only one person developing cancer. Although such risk levels for individual substances may appear to be fairly low, it is important to realize that in their every day existence most living organisms are exposed to a large number of potentially carcinogenic substances. Furthermore, large uncertainties reside in the numerical values for carcinogenic risk levels since, for the most part, they are based on extrapolations from short term animal exposure at high doses. PAHs are considered to be an important class of environmental contaminants 2 23 because many of the compounds have been shown to be carcinogens ' . The mechanism by which PAHs cause cancer is believed to involve the formation of active species which 23 can bind to DNA, or other important cellular structures such as proteins and enzymes . The three main active species believed to be responsible for the carcinogenic activity of 23 24 25-30 31 PAHs are diolepoxide metabolites ' , benzylic esters " , and radical cations Although the majority of research on PAH metabolism has focused on the diol-epoxide 25 30 metabolic pathway, the benzylic ester pathway proposed by Flesher " may explain the high carcinogenicity of some methylated PAHs, such as 7,12-dimethylbenz[a]anthracene. One of the important issues in the toxicology of PAHs is the large difference in the toxicological properties of different isomers. For example, l,12-dimethylbenz[a]anthracene is not carcinogenic while 7,12-dimethylbenz[a]anthracene is one of the most potent PAH 32 carcinogens known . It is interesting to note that the methyl groups on 7,12-dimethylbenz[a]anthracene are situated at the sites of lowest localization energy on benz[a]anthracene (Figure 1.3). Table 1.2 summarizes some of the differences in 14 carcinogenic potency values for a variety of PAH isomers. Because of these differences, it is important to be able to analytically differentiate amongst various isomers in order to obtain toxicologically relevant data. However, the techniques normally employed in the analysis of PAHs are often not good enough to distinguish many of the PAH isomers from each other, particularly when the PAHs are present in a complex mixture, such as in an environmental sample. Thus, it is important for these analytical techniques to be improved or modified so that isomer differentiation can be optimized. In an environmental analysis for polyaromatic compounds, the primary targets are usually PAHs. However, the transformation products of PAHs, such as polyaromatic quinones (PAQs) are often not targeted. Although they may be present at low concentrations, and thus are difficult to detect, some PAH transformation products may exhibit greater toxicity than their precursor compounds. For example, some oxidation 33 34 products of pyrene (6a), a non-mutagenic PAH, have been found to be mutagenic ' . Consequently, the analyses for many PAH transformation products, such as PAQs, need to be improved to facilitate their detection and quantitation at low concentration levels in complex matrixes. The research that will be presented in this thesis has been focused primarily on these two analytical challenges: (1) the ability to resolve and identify PAH isomers, and (2) the detection and identification of PAH transformation products that are obscured by more abundant matrix interferences. As a background to this work some of the most frequently used analytical techniques for extraction, sample clean up, chromatographic separation, detection and identification of PAHs are outlined in Section 1.3. 15 Table 1.2 Carcinogenic properties of selected PAHs Compound Carcinogenicity * anthracene -phenanthrene -9-methyianthracene + 1 -methylphenanthrene + Fluoranthene -Pyrene -B enz [ a] anthracene + 6-methylbenz[a] anthracene ++ 7-methylbenz[a] anthracene +++ 1 -methylbenz[a] anthracene +/-1,12-dimethylbenz[a] anthracene -7,12-dimethylbenz[a] anthracene ++++ Chrysene + Tetracene -Triphenylene +/-Benzo[a]pyrene ++++ B enzo [e] pyrene + Benzo[b]fluoranthene ++ B enzo [k] fluoranthene + B enzo [g, h, i] perylene ++ * The - sign corresponds to no induction of tumor in mice, + corresponds to induction of up to 25 %, ++ up to 50 %, +++ up to 75 %, and ++++ up to 100 %. Data from Mohammad 3 5 and Miller 3 6 . 1.3 ANALYSIS OF PAHs IN ENVIRONMENTAL SAMPLES The first step in most environmental analyses is an extraction procedure to remove the analyte from its matrix. Extraction of PAHs from a solid matrix, for instance a soil or aquatic sediment sample, is usually performed using an organic solvent such as 16 dichloromethane or hexane. The traditional approach has been to use Soxhlet extraction 1 . Although Soxhlet extraction is an effective method for removing PAHs from solid matrixes, and is still widely used in numerous analytical protocols, it suffers from the disadvantage of requiring relatively large amounts of organic solvents for each sample (« 200 mL of solvent per 10 g of sample) and long extraction times (6 - 12 hours). Two relatively new extraction techniques which offer significant advantages in both time and 3 7 solvent reduction over Soxhlet extraction are Supercritical Fluid Extraction and 3 8 Accelerated Solvent Extraction . Both extraction processes operate at elevated pressures and can perform extractions quickly (5 - 15 min) using small quantities of solvents (10 - 30 mL). The former uses a supercritical fluid such as C 0 2 as an extraction medium while the latter employs traditional organic solvents such as dichloromethane. Extraction from an aqueous sample can be performed by liquid/liquid extraction, solid phase extraction or solid phase microextraction. Similarly to Soxhlet extraction, liquid/liquid extraction requires a large volume of solvent (« 500 mL solvent per L of water) and is a time consuming procedure. Solid phase extraction involves passing the sample through a solid adsorbent material which retains the PAHs. Release of the PAHs from the adsorbent is performed by elution with an organic solvent (< 100 mL). Solid phase microextraction involves placing a fiber coated with a hydrophobic polymer, such as polydimethylsiloxane, into an aqueous solution of the analyte 3 9 ' 4 0 . Hydrophobic substances will adsorb onto the fiber and can be analyzed by thermally desorbing the fiber in a GC injector port. Once the desorption process is completed the fiber can be reused numerous times (50-100). 17 PAHs are hydrophobic compounds and one of the major problems in their analysis when present in an aqueous medium is their adsorption onto the surfaces of glass or plastic (polyethylene) containers (sampling or storage). Losses of up to 80 % due to adsorption have been observed over periods of 2 hours 4 1 . This loss can be reduced significantly by addition of a micelle forming surfactant to the aqueous medium 4 1 . For quantitative analysis, internal standards should be added prior to extraction of the analyte. An internal standard should be closely related in chemical and physical properties to the analyte. Thus stable isotope analogues, such as d10-pyrene (C 1 6 D 1 0 ) , are frequently employed 4 2 . If a mass spectrometer is used as a detector, deuterated PAHs can be differentiated from non-deuterated PAHs by differences in molecular mass (Mr) (e.g. M r of pyrene = 202 u, M r of d10-pyrene = 212 u). If Q ; micrograms (u.g) of internal standard are spiked into W grams of sample prior to extraction, the concentration of the target analyte, C a , present in the sample can be determined by using equation 1.1 4 2 , when working within the linear dynamic range of the analysis. R a and R{ represent the detector response measurements (such as the ion chromatogram peak areas obtained from a GC/MS analysis) for the analyte and internal standard respectively. R F a and RF; represent the response factors for the analyte and internal standard, and are determined in a separate analysis using accurate amounts of both the analyte and internal standard (RF= R/m, R = measure response for quantity "m"). F a and F; represent the fractional recovery efficiencies of the analyte and internal standard. Because of the difficulties in determining the value of F a , the ratio of F; / F a is usually [1.1] 18 assumed to be equal to 1. Consequently, one of the main challenges in trying to obtain an accurate value of C a is to attain similar extraction efficiencies for the analyte and the internal standard (i.e. F a « F ;). However, since the internal standards are usually spiked just prior to the extraction process, they will not be bound to the matrix in the same way as the native PAHs. Spiked PAHs have been shown to be extracted more rapidly than native PAHs (F ; > F a) 4 3 . Since it is normally assumed that F, = F a and since F a cannot be readily calculated, the calculated value for C a often underestimates its true value. It has also been found that differences in the extraction rates between native and spiked PAHs are matrix dependent 4 3 . For example, PAHs that are formed together with the matrix, such as in air particles, show the largest differences in extraction rates between native and spiked PAHs 4 3 . The extraction of high M r PAHs (M r > 302) from a matrix can prove to be difficult and requires solvents such as chlorobenzene 4 4 . Consequently, it is unlikely that a standard solvent extract (e.g. CH 2C1 2) will have a profile that exactly represents all the PAHs present in the original matrix. Given the inherent error that can exist in relating spiked and native analyte as well as the fact that this error depends on both the matrix and the nature of the PAH, obtaining true quantitative analytical data remains a difficult challenge. Generally, when investigating any analyte in a complex matrix, it is desirable to remove matrix components that may interfere with the analytical method to be used. For example, when using a non specific detector, such as flame ionization (FID), for gas chromatography (GC) analyses, the presence of aliphatic compounds may obscure the presence of PAHs 1 . Consequently, some pre-separation / pre-purification is often 19 desirable. In the case of P A H analyses, a brief fractionation on a short silica L C column can be used to separate aliphatic compounds from polyaromatic compounds, as well as non-polar aromatic compounds from polar aromatic compounds. If highly polar aromatic compounds, such as phenols, have to be analyzed by GC, it is necessary to precede the GC separation with a suitable derivatization step 4 5 , or to use liquid chromatography instead of G C . The environmental samples that have been analyzed in this study contain a large number of different compounds, including numerous PAH components. However, a detailed work-up scheme, such as multiple chromatographic steps or extractions, has been avoided because of the probable loss of minor components present in the crude extract. With the exception of the pre-clean-up step just mentioned, only gas chromatography and tandem-mass spectrometry have been used in this work as a means of separating mixtures of compounds. 1.4 GAS CHROMATOGRAPHY Capillary GC 4 6 provides very high chromatographic resolution that is particularly suitable for analysis of complex environmental samples. For the analysis of low to medium polar unsubstituted and alkyl substituted PAHs, a low to medium polar GC column, such as a DB-5 (95 % dimethyl-5 % diphenyl-polysiloxane) column should be suitable for the separation process. Although such a column is sufficient for the separation of some unsubstituted structural isomers (such as 5a/5b or 6a/6b), many substituted PAHs, such as methyl PAHs, are not completely separated. Special capillary columns using liquid crystalline phases have been developed which can achieve better chromatographic 20 separation by isolating molecules based upon their molecular shape. By using such columns, the twelve monomethyl isomers of benz[a]anthracene (8b) have been separated 4 7 . However, these columns can only be heated to a maximum of 300°C. Because of their high boiling point, PAHs with M r > 302 cannot be analyzed at these temperatures, or even at the maximum temperatures of a DB-5 column (320°C). High temperature GC columns with cross-linked polymeric or chemically bonded phases have been designed which can be heated to 420°C allowing PAHs up to M r = 374 to be analyzed 4 8 . However, a GC column combining high temperature and molecular geometry selectivity still needs to be developed. In view of these chromatographic limitations, it is desirable to find methods, other than gas chromatography, that would enable differentiation between the various isomers. A number of detectors are currently available for use with GC systems, the most widely used being a flame ionization detector (FID) which, as mentioned earlier, can be categorized as non-specific. When using an FID detector, identification of individual compounds in a chromatogram can only be made by retention time comparisons with suitable standards. Consequently, if standards are not available for an analyte, or if the analyte co-elutes with other compounds, its presence in a sample cannot be confirmed. More specific detectors for GC are the nitrogen-phosphorus, flame photometric, and 49 electron capture detectors, generally used to identify heteroatom-containing compounds . Certainly, the most versatile and, in many cases, the most sensitive detector, is a mass spectrometer. In addition to chromatographic signatures, a mass spectrometer will provide structural and often M r information on the compounds eluting from the GC column. 21 1.5 MASS S P E C T R O M E T R Y Since its introduction in 1907 by J . J . Thompson 5 0 as a technique for separating naturally occurring isotopes, mass spectrometry has evolved into an essential component of many analytical investigations because of its high sensitivity and ability to provide information on a compound's molecular mass and structure. Detailed information on applications of mass spectrometry in chemistry and other disciplines has been published in numerous review papers 5 1 ' 5 2 and monographs 5 3 ' 5 4 . The process of a mass spectrometric analysis can be subdivided into four major stages: sample introduction, ionization, mass separation and detection as outlined in Figure 1.4. For this thesis, sample introduction was achieved either by direct probe desorption or through a GC capillary column. Sample molecules that are introduced into the low pressure environment of an ion source are ionized in the gas phase. This can be achieved by a variety of techniques, the most common being electron ionization (EI). Ion formation in EI mode is achieved via an energy exchange during collisions of energetic electrons with neutral atoms or molecules. As illustrated in Figure 1.5, electrons are emitted from a heated filament F ( generally Tungsten or Rhenium ) and are accelerated by a potential difference V , applied between the filament and plate S, that is normally attached to the ionization chamber I. The magnitude of V will determine the energy of the electron beam that is crossing inside I to collector T and interacting with the sample molecules on this path. Vacuumz IO - 5 mbar Sample Introduction (Inlet) Production of Ions [Source) Mass/Charge Separation of Ions (Analyzer) Ion Defection (Defector) : region under vacuum or at ambient pressure : region always under vacuum Recording of Data (Data System) Figure 1.4 General schematic of a mass spectrometer M Heatable Probe R \ l P N S DE Probe Figure 1.5 General schematic for a mass spectrometer's ion source. I: Ionization chamber, M: Magnet (optional), T: Trap/Collector, F: Filament emitting electrons, R: Repeller, S: electron extraction plate, DE: Direct Exposure probe. Ions exit to mass analyzer. 23 When electrons interact with molecules, they can cause ejection of an electron forming molecular radical cations ( M + ) , often referred to as molecular ions. In general, an electron beam of 70 eV is utilized, often causing fragmentation of the analyte molecules. Negative molecular ions ( M ) may be formed when molecules capture low energy (thermal) electrons. A second and usually milder method of generating ions is chemical ionization (CI). This ionization technique is the most important one in the context of this thesis. In this technique, first introduced by Munson and Field 5 S , ionization of a sample is achieved by gas phase ion/molecule reactions in the ion source with ions generated by electron ionization of a reagent gas. The reagent gas is generally present in large excess (e.g. 10 to 1) compared to the sample molecules. Consequently, the probability of a sample molecule undergoing ionization via EI is very low. When a reagent gas is ionized at appreciable pressures, a variety of ions can be formed by ion-molecule reactions, which in turn collide with the neutral analyte molecules M introduced into the ion source, resulting in a number of product ions. Both positive and negative ions can be generated, depending on the chemical nature of the reagent gas and the analyte molecules. The analyte molecules (M) may undergo charge exchange with positive reagent ions (R +) to form molecular ions (M + ) , or they may form adducts with R + (e.g. [M + R] + ) or undergo ionization by abstracting a proton from R + to yield protonated molecules M H + . Sometimes hydride abstraction takes place to form [M - H ] + ions. Any of these ionization products can undergo further changes, such as elimination reactions, depending on the energy transferred to the molecule during ionization. The ion species that will be produced in the greatest abundance is dependent on the chemical and physical properties of both R 24 and M , such as ionization energy (IE), electron affinity (EA), and proton affinity (PA). When a reagent gas is ionized, a series of different ions may be generated. For example, ionization of methane gas at pressures < 1 mbar will yield C H 4 + , C H 3 + , C H 2 + , as well as products from ion/molecule reactions with C H 4 such as 5 4 ' 5 6 : C H 4 + C H 4 + • - > C H 5 + + C H 3 Many ions can be generated during ionization of a reagent gas, often leading to a complicated reaction chemistry between analyte molecules and reagent ions. Other techniques that can be used to generate ions in a mass spectrometer include 57 58 laser ionization as in MALDI (Matrix Assisted Laser Desorption Ionization) ' , high electric fields (Field Ionization /Desorption) 5 9 ' 6 0 , and ionization through the desolvation of charged droplets (Electrospray ionization)61. Mass separation of the ions produced (parents and fragments) can be performed by using time of flight (TOF) analyzers, magnetic fields (B), linear quadrupoles (Q), and quadrupole ion traps (IT), as outlined in Figure 1.6 (a) to (d). A magnetic sector is often combined with an electric analyzer (E) to form a double focusing MS (E/B or B/E) used for the determination of accurate mass. Time of flight separation is currently the simplest method of mass analysis. Ions are formed in a short time window (e.g. via laser ionization or a pulsed electric beam) and simultaneously accelerated from the ion source by a voltage pulse into a field free drift tube 5 4 . It can be seen from formula 1.2 that heavier ions will take longer to reach the (mL2T detector than lighter ones: t= [1.2] \2zeVj (t = time, L = flight path length, z = ion charge, V = accelerating voltage, e = electron charge). 25 Figure 1.6(b) Magnetic sector analyzer. M = Magnet, F - Flight tube, S = collector slits, D = detector. + ( U + V c o s cot) — Q B © -(3 - ( U + V c o s cot) Figure 1.6(c) Quadrupole analyzer. D = detector. Figure 1.6(d) Quadrupole ion trap analyzer. F = filament, S — electron extraction plate, E — end caps, R = ring electrode, D = detector. 27 Separation of ions by curved E/B sectors is based on the difference in the radius of curvature of the path taken by ions of different m/z values under a given set of electric / magnetic field conditions. Mass separation by linear quadrupoles occurs through the difference in stability of ions in a two dimensional quadrupole electric field 6 2 . Under a given set of r.f. and d.c. voltages, only a narrow range of m/z values will be transmitted to the detector. Ion traps are similar to linear quadrupoles but have a three dimensional quadrupole field 6 2 . Hence ions that are stable under a given set of r.f. / d.c. conditions can be "stored" in the trap. Mass analysis is performed by rendering the ions unstable through changes in the a.c. or d.c. voltages causing a sequential ejection of ions from the trap. Detection of ions in most mass spectrometers is usually performed with an electron multiplier (EMP). An EMP detector can amplify the signal of one ion by a factor of 105 to 106. Because of the importance that ion trap mass spectrometry has played in the research described in this thesis, a more detailed description of this mass spectrometric technique follows. 1.5.1 Ion Trap Mass Spectrometry A good general introduction to the basic principles of ion trap mass spectrometry is outlined in a review paper by March 6 3 . An ion trap (Figure 1.6(d)) consists of three electrodes, two end-caps (E) and a donut-shaped ring electrode (R) with a radius r 0 generally being ^ 1cm. In the EI mode, electrons are accelerated from a filament (F) by applying a voltage (^  150V) pulse (psec-msec) to plate S (gate). The electrons then enter 28 the trap region where they collide with molecules introduced from a sample inlet (I), reagent gas inlet or a gas chromatograph. In the CI mode, a reagent gas is first ionized using a pulse of electrons and then allowed to react with neutral molecules inside the trap for a specified period of time. The ring electrode is supplied with an r.f. voltage ( MHz range) and sometimes also with a d.c. voltage 6 2 . After an ionization pulse, ions fluctuate rapidly in the axial (z) as well as in the radial direction. By increasing the r.f. applied to the ring electrode, ions of increasing m/z values are rendered unstable and are ejected from the trap to reach the detector (D). In addition to this mass scanning mode, one can also isolate ions of certain m/z values and accumulate them inside the trap while a compound is, for instance, eluting from a GC column, resulting in an increase in detection sensitivity. Isolated ions can also be subjected to MS/MS experiments during which the ions are fragmented by ion excitation and collision with the GC mobile phase or other gases followed by mass analysis of the resulting fragments. After an MS/MS experiment of this kind, a product ion could be trapped again and subjected to another CID experiment. This process can be repeated several times giving MS" experiments64. The mass trapping capabilities of an ion trap are also well suited for the study of ion/molecule reactions. By being able to selectively store or eject ions, it is possible to ionize a reagent gas and keep only specific ions inside the trap, enabling ion/molecule reactions to be performed 6 2 . In effect, an ion trap has the capacity to act as an 'electric field test-tube'. The stability of ions in the three dimensional quadrupole field of an ion trap can be illustrated by the trapping parameters and shown in equations 1.3 and 1.4 and the stability diagram in Figure 1.7 6 3 . The diagram in Figure 1.7 represents the values of the a,z and trapping parameters for which ions are stable in a quadrupole ion trap 29 (shaded region). Generally, most ion traps are operated in an r.f. only mode, implying that the value of will always equal zero. In the r.f. only mode of operation, ions become unstable and are ejected from the trap when a qz value of 0.908 is reached. -l6eU SeV m(rl + 2z02)O2 L J m(r20 + 2z2)f22 L ' J U = magnitude of d.c. potential (volts) V = zero-to-peak (0-p) amplitude of r.f. drive potential (volts) Q = angular frequency of r.f. drive = 2jrf (rad s"1), f = drive frequency in Hz r 0 = radius of ring electrode in the horizontal plane (m) 2z0 = separation between endcap electrodes (m) m = mass of the ion (kg) e = charge on ion (C) 30 31 Two other important parameters in ion trap mass spectrometry, in addition to the 2^ and qz trapping parameters, are the axial (z) and radial (r) iso-P lines, (Pz and Pr, Figure 1.7) and the secular frequency of an oscillating ion co. The value of the fundamental frequency co (in Hz) for a trapped ion will depend upon the value of P and the r.f. drive frequency f (in Hz), according to equation 1.5. 0) = ^ - [1.5] 2 For qz < 0.4 the value of Pz can be estimated according to equation 1.6. A*Jf«,+v] [ L 6 ] When chemical ionization (CI) is performed in an ion trap, the reagent gas is usually at a lower pressure (P « 10"5 torr) than in a conventional CI source (P « 0.1 - 1 torr). The lower pressure inside an ion trap results in less collisional stabilization of ion/molecule reactions65 leading to products that may be different from those observed in a conventional CI source. 1.6 COMPOUND IDENTIFICATION The main challenges in the identification of PAHs in environmental samples are: (1) to distinguish them from chemical interferences present in the sample matrix and (2) to differentiate between PAH isomers or isomer groups. If standards are available, and no co-eluting isobaric interferences are present, the identification of PAHs can be made based on a comparison of retention times. However, standards are not available for many isomeric species. Furthermore, when using electron ionization mass spectrometry (EI/MS), the spectral signatures of isomeric PAHs are virtually identical 6 6 . One possible 32 way to achieve isomer differentiation is to exploit differences in the chemical properties of the isomers by way of gas phase ion/molecule reactions under chemical ionization conditions. Under CI conditions, reactions could take place between the reagent gas ions (R + or R") and the neutral sample molecules (M) eluting from the GC column, resulting in the formation of molecular ions ( M + or M"), proton abstraction/donation products ( M H + , [M - H]~), adduct ions ([M + R ] + , [M + R]") or elimination products from adduct ions ([M + R - F ] + or [M + R - F]~, F = neutral fragment). As will be described in later chapters, differences in the type and relative intensities of ions formed can enable PAH isomers to be differentiated. A number of CI techniques have been reported for differentiating some PAH isomers. These CI techniques, listed according to the reagent 67 70 71 72 gas employed include: ethers " , dimethylsulfide methane/argon , carbon dioxide 7 3 ' 7 4 , and oxygen/nitrogen 7 5 , 7 6 . However, only in a few cases have attempts been made to differentiate PAH isomers in environmental samples by using chemical ionization 7 0 ' 7 2 . The other mass spectrometry technique that can be employed to perform isomer differentiation is tandem mass spectrometry (MS/MS). In the MS/MS mode, specific mass-selected ions are fragmented under controlled conditions and the resulting products are then mass analyzed. Although it can be a powerful technique for the differentiation of compounds that undergo extensive fragmentation, it is not as well suited for molecules, such as PAHs, that do not fragment readily. However, research in this field has enabled 77 82 some degree of P A H isomer differentiation to be achieved " . 33 1.7 OBJECTIVES AND OVERVIEW OF THESIS The first objective of this thesis in an investigation of novel CI/MS techniques for the differentiation of PAH isomers. The second objective is the development of MS/MS techniques for the analysis of trace level PAH transformation products in environmental samples. As the focus of this work is to develop techniques suitable for the analysis of environmental samples, authentic environmental samples obtained from Kitimat, British Columbia, are used to test most of the analytical techniques developed and investigated here. The application of negative chemical ionization in a quadrupole mass spectrometer using N 20, C 0 2 , and C H 4 reagent gases for PAH isomer differentiation is discussed in Chapter 2. The use of these gases to generate isomer specific adducts is demonstrated. A comparison of results obtained here with previous work in this field is discussed together with some of the significant variables which affect this technique. Results from positive chemical ionization experiments using dimethylether in both a standard CI ion source and in an ion trap are compared in Chapter 3. The formation of isomer specific adducts and their usefulness in differentiating some PAH isomers is demonstrated using the standard high pressure ion source. Differences in formation of adducts in the ion trap and a standard ion source are described. The discovery of halogenated hydrocarbons, such as CH 2 C1 2 and C H 3 C H F 2 , as a novel set of chemical ionization reagents for the differentiation of PAH isomers, using a quadrupole ion trap, is described in Chapter 4. Due to its novelty and effectiveness at differentiating isomers in environmental extracts, this CI technique is the most important one for the isomer differentiation aspects of this thesis. 34 The application of the CI techniques developed in Chapter 4 to the differentiation of P A H isomers in an environmental sample is described in Chapter 5 . Novel chromatographic analysis and transformation techniques used for the deconvolution of complex chromatographic data wi l l also be presented. The use of mass-isolated perfluorotributylamine ( F C 4 3 ) fragment ions to perform ion/molecule reactions with P A H s , in an ion trap, wi l l be discussed in Chapter 6 . The formation of adduct ions enabled characteristic mass spectra to be generated for the differentiation of P A H isomers. Finally the use of G C / M S / M S with a quadrupole ion trap for the analysis of P A H oxidation products is presented in Chapter 7. Using this technique, it w i l l be demonstrated that polyaromatic quinones that are not detectable by G C / M S can be identified using G C / M S / M S under optimized fragmentation conditions. CHAPTER 2 DIFFERENTIATION OF PAH ISOMERS USING NEGATIVE CHEMICAL IONIZATION MASS SPECTROMETRY 36 2.1 INTRODUCTION Molecules undergoing ionization inside the ion source of a mass spectrometer have the potential to form both positive and negative ions. Organic compounds with a low electron affinity, (EA < 0.5 eV), will generally produce positive ions in greater abundance than negative ions. However, structural features that can increase the electron affinity of a molecule, such as halogen substituents, will increase the yield of negative ions. In order for a molecule to form negative ions it must capture an electron. Under standard EI conditions (e.g. 70 eV, ~ 10"6 mbar source pressure), formation of negative ions is a rather improbable process due to a mismatch between the high translational energies of the electrons and the binding energy of the electron (electron affinity) which must be taken up in the emerging product anion 8 3 . If too much excess energy is present, the negative ion will either fragment or immediately release the electron. One way to decrease the excess energy during electron capture is to decrease the energy of the electrons. This can be accomplished by adding a buffer gas inside the ion source, at typical CI pressures « 1 mbar. The higher gas pressure will reduce the energy of the electrons through inelastic scattering and dissociative ionization processes 5 6 . The resulting lower-energy electrons are often referred to as thermal electrons (eth). The presence of a buffer gas also enables high energy ions to be stabilized by collisional stabilization. Often the buffer gas itself has a low electron affinity so it is not ionized and merely serves to generate the thermal electrons (eth) and to stabilize newly formed negative ions. Inert gases such as argon have been used for this purpose since they yield few negative ions themselves56. The ease of formation of negative ions will also strongly depend on the electron affinity (EA) of the target molecule. Compounds with high E A values will form negative ions more easily due to the larger amount of energy liberated during electron capture, 3 7 as long as the ions can be stabilized. If ionization of analyte molecules via ion/molecule reactions is desired, then buffer gases which can form negative reagent ions need to be used. In such instances the name "reagent gas" is more appropriate than "buffer gas". When an abundant supply of negative ions, G", generated from the reagent gas, is available, analyte molecules (M) may react with G" to form adduct ions such as [M + G]". However, ionization by capturing en, will still be possible as long as the analyte molecules possess a sufficiently high EA. The most common reactions that can occur during negative chemical ionization (NCI) are the following 5 6 : (1) electron capture M + e^ -> [M]~* -» IVT or M + e th -> [MT* -> [M - X]" + X (2) electron transfer M + G -> M" + G ( 3 ) proton transfer M + G"• -> [M - H]" + G H (4) adduct formation M + G" -> [M + G]"* -> [M + G]" or M + G" -> [M + G]"* -> [M + G - X]" + X The presence of neutral species (such as non-ionized G molecules) is also required to collisionally stabilize the energized products (indicated by *) in the reactions shown above. If these products are not stabilized, they may dissociate, releasing a neutral fragment (X). Compounds with a low E A are more likely to undergo ionization via adduct formation or proton transfer reactions, since only small amounts of energy will be liberated by electron capture or charge exchange. 38 The E A values of PAHs can vary significantly as shown previously in Table 1.1(a). Consequently, some P A H isomers should form negative ions more readily than others. The difference in the E A of PAHs was first exploited by Buchanan 8 4 for the differentiation of PAHs via NCI experiments using CH4 as a reagent gas. The PAHs analyzed by this technique were separated into two main groups, those with E A values greater than 0.5 eV, that were able to form molecular ions (M"), and those with E A less than 0.5eV that did not yield significant amounts of molecular ions. When the E A values for a pair of P A H isomers lie on either side of 0.5 eV, then only the isomer with E A > 0.5 eV is expected to form appreciable amounts of negative ions. Hence, isomers such as benzo[a]pyrene (9d) (EA = 0.680 eV) and benzo[e]pyrene (9c) (EA = 0.534 eV) can be distinguished. Although this type of technique can be used to differentiate between isomers, it does not allow characteristic mass spectral signatures to be obtained, since the isomer with the low E A value will not yield a spectrum. However, Buchanan 8 4 observed that PAHs with a saturated five-membered ring, such as fluorene (4) showed the presence of [M + 14]" adduct ions (probably: [M+0-2H]"). The formation of these adducts was attributed to the low electron affinities of these compounds 8 4 . Formation of adduct ions can be a useful analytical tool in the differentiation of isomeric P A H compounds. Isomeric compounds may form different adduct ions or they may form the same adducts but in different proportions. The use of negative adduct ions to 75 76 85 86 distinguish between P A H isomers has been demonstrated by Stemmler and Buchanan ' ' ' . In NCI mode and employing a mixture of N 2 / O 2 as reagent gases, it was shown that a variety of ions could be generated such as [M - H]", [M + O - H]", [M + 02]", [M + O - 2H]" and [M + 20 - 2H]". It has been suggested that the latter two adducts are probably formed via surface catalyzed reactions in the ion source, induced by the formation of rhenium oxides produced from the ion source filament 8 7 . Some P A H structural and positional isomers were 39 differentiated by using N 2 / 0 2 N C I based on differences in the abundances of the adduct ions generated. Although this N C I technique produced interesting results, the application of 0 2 as a CI reagent gas is generally of limited practical use because of its high reactivity with the ion source filament, especially when time consuming analyses are performed as in the case of GC/MS investigation o f environmental extracts. A more practical technique for the differentiation of P A H isomers via adduct formation in N C I has been demonstrated by using C 0 2 as a reagent gas 7 3 ' 7 4 . Through reactions with some of the reagent ions formed during ionization o f C 0 2 (O", 02", and CO3"), adducts of the type [ M + O - H]", [ M + 20 - 2H]" and [ M + 20 - 2H - CO]" were observed. Differences in the abundances o f these adducts enabled five P A H isomers (8a to 8e) to yield significantly different mass spectra 7 4 Another reagent gas that has been used for the analysis of PAHs by N C I is N 2 0 . N C I using a mixture o f N 2 0 and CH4 has been employed in the analysis o f samples containing polar substituted PAHs in an attempt to measure their mutagenic content 8 8 . These experiments enabled a correlation to be developed between the electron capture response and the biological activity (e.g. mutagenicity) o f some compounds 8 8 ' 8 9 . For example, benzo[a]pyrene (9d), a powerful carcinogen, has a greater electron capture response than benzo[e]pyrene (9e), a compound considered to be only weakly carcinogenic 9 0 . The addition o f C H 4 to the N 2 0 reagent gas enables the formation o f OH", according to the following mechanism 9 1 : e t h + N 2 0 -> O"' + N 2 O" + CH4 -> OH" + CH 3 40 The OFT ion can be a useful chemical ionization reagent by acting as a powerful proton acceptor, to generate [M - H]" ions. Based upon some of the research previously referenced, it appears that C 0 2 , and N 2 0 have potential as reagent gases for the NCI analysis of PAHs. Thus, the aim of this chapter was the further investigation of these gases as chemical ionization reagents for the analysis and differentiation of P A H isomers in an attempt to develop a technique that may be suitable for the analysis of environmental samples. 41 2.2 EXPERIMENTAL 2.2.1 Materials, Standards and Reagents A standard mixture of 16 PAHs (2000 u.g mL"1 each) (naphthalene 1, acenaphthene 2, acenaphthylene 3, fluorene 4, phenanthrene 5b, anthracene 5a, fluoranthene 6b, pyrene 6a, benz[a]anthracene 8b, chrysene 8c, benzo[b]fluoranthene 9a, benzo[k]fluoranthene 9b, benzo [a] pyrene 9d, indeno[l,2,3-cd]pyrene 10a, dibenz[a,h]anthracene 11, benzo[g,h,i]perylene 10b) was obtained from Supelco (Oakville, Ontario, Canada) and diluted with toluene to give 50-200 u.g mL"1 solutions. Methyl and dimethylbenz[a]anthracenes were obtained as 50 u.g mL"1 solutions from AccuStandards (New Haven, CT, USA). Individual neat P A H standards (anthracene 5a, phenanthrene 5b, benz[a]anthracene 8b, chrysene 8c, triphenylene 8d, tetracene 8e, carbazole) were obtained from Aldrich (Milwaukee, WI, USA ) and the perdeuterated PAHs dio-anthracene and dio-acenaphthene were obtained from CLL (Woburn, MA, USA). The reagent gases (N 20, C 0 2 and CH4) were obtained from Linde-Union Carbide (Toronto, Ontario). 2.2.2 Environmental Samples The environmental samples analyzed were collected from an effluent settling pond (B-lagoon) at the Alcan aluminum smelter in Kitimat, British Columbia, Canada. Samples were collected in amber glass jars and frozen. In the laboratory the thawed sediments were freeze dried to remove all of the water. A 5 g sample of the dry residue was then placed in a glass thimble, topped with a layer of anhydrous sodium sulfate and Soxhlet extracted with dichloromethane (200 mL) for 8 hours. When required, the internal standards dio-anthracene 42 and dio-pyrene were spiked into the sediments prior to extraction. The dichloromethane extract was concentrated to a volume of approximately 5 mL on a rotary evaporator and then passed through a short silica column (5 g silica), eluting with dichloromethane, to remove polar impurities. The eluted extract was then reduced to a final volume of 1 mL by evaporation under a stream of nitrogen gas. 2.2.3 Instrumentation All experiments were performed using a Delsi-Nermag RIO-IOC single quadrupole mass spectrometer operated in negative ion mode. The major operating parameters are shown in Table 2.1 Table 2.1 Major instrumental parameters used during NCI experiments Parameter Range of values used Electron energy 50 - 100 eV Emission current 0.200 - 0.250 mA Source temperature 180°C to 280°C CI Source pressure 0.1 to 1.3 mbar Analyzer pressure 10"7 mbar Multiplier 1.30 kV Mass scan range 100 - 350 m/z Scan rate 200 ms / scan Mass calibration was performed first in positive ion mode using perfluorotributylamine (FC43, (C4F9)3N) and confirmed in negative ion mode using chloramphenicol (C11H12CI2N2O5). 43 2.2.4 Sample Introduction Samples were introduced into the ion source either by desorption off a heated filament probe (DCI) or from a capillary GC column terminating inside the ion source. a) Heated filament probe Desorption chemical ionization (DCI) was performed by placing 1-10 pL solution of a sample (100 to 1000 ng) on a tungsten filament. The solvent was allowed to evaporate at ambient conditions and the probe was inserted into the ion source through a probe lock. Once in the ion source, the current to the tungsten filament was increased to a maximum of 500 mA at a rate of 20 mA s"1 . This current resulted in filament temperatures exceeding 1000°C very rapidly, thus achieving sample temperatures at which the rate of vaporization exceeds that of pyrolysis. 1 2 6 b) Gas Chromatography Sample introduction by GC was performed using a Varian Vista 6000 gas chromatograph equipped with a J&W DB-5 column (30 m x 0.25 mm i.d., 0.25 pm film). A sample of the desired concentration (1.0 - 2.0 pL corresponding to approximately 100 to 200 ng of individual PAHs) was introduced into the GC via a splitless injection. The injector port was maintained at 290 - 310°C. The oven temperature program was: 100°C for 3 min followed by a 10°C min"1 ramp to 300°C, hold at 300°C for 10 min and then ramped to 310°C at 30°C/min. The transfer line to the mass spectrometer, through which the GC column was threaded, was heated to 300°C. The He carrier gas flow rate was 38 cm s"1, corresponding to approximately a lmL min"1 flow rate. 44 2.3 RESULTS AND DISCUSSION 2.3.1 Desorption Negative Chemical Ionization Analysis 2.3. J. J Structural PAH isomers Preliminary NCI experiments were performed by thermally desorbing individual PAHs into the ion source using the DCI probe. A series of four P A H isomers with a molecular mass (Mr) of 228 u was chosen to investigate the effects of NCI using N 2 0 and C 0 2 reagent gases. The DCI mass spectra of PAHs 8b to 8e obtained by using these two reagent gases are reproduced in Figures 2.1(a) and (b). Significant differences can be noted in each set of mass spectra, due to the presence of different ions or due to differences in the intensity of ions. The major ions observed correspond to m/z 227 = [M - H]", 228 = M", 243 = [M + 15]" = [M + O - H]", 258 = [M + 30]" = [M + 20 - 2H]" in accordance with adducts previously reported 7 3" 7 6 ' 8 5 . The small signals observed at m/z 270 and 271 could arise from adducts of the type [M + G - 2H]" (m/z 270), and [M + G - H]" (m/z 271), where G represents either C 0 2 or N 2 0 . Of particular interest is the [M + 20 - 2H]" adduct ion (m/z 258) because its formation has been attributed to surface catalyzed reactions inside the mass spectrometer's ion source75. This ion is most abundant in the spectra of 8b and 8e. A possible explanation for this observation is that these two compounds are the ones which can most easily form stable quinones due to their low localization energy (Lc) at their 7,12 positions (8b) and 5,12 positions (8e) 2 0 . Plausible structures for the [M + 20 - 2H]" adduct ions of 8b and 8e are illustrated in Figure 2.1(a). 45 100-8 0 H ^ 6 0 -to > -55 40H 2 0 H 0-8b 2 5 8 243 228 (A C CO > CD CC 271 2 0 0 2 2 0 2 4 0 2 6 0 2 8 0 3 0 0 100J 8d 80 H ^ 6 0 -> o5 4 0 H 20 H 0 -m/z 2 4 3 2 2 6 _ L L _ 2 5 8 2 7 0 T 2 0 0 2 2 0 2 4 0 2 6 0 2 8 0 3 0 0 100H 8 0 - 6 0 "5 40-2 0 -0-8c 2 4 3 2 3 0 -i—r 2 5 8 2 7 0 "l I" I • "i 2 0 0 2 2 0 2 4 0 2 6 0 2 8 0 3 0 0 cz DC 100-5 so H c CD ^ 6 0 co > 4 0 H 20 H 0 8e m/z 2 4 3 2 5 8 2 0 0 2 2 0 2 4 0 2 6 0 2 8 0 3 0 0 m/z m/z m/z 258 = [M+20-2H]--, M = 8b m/z 258 = [M+20-2H]--, M = 8e Figure 2.1(a) CO2 negative DCI mass spectra of benz[a]anthracene (8b), chrysene (8c), triphenylene (8d) and tetracene (8e). 46 100H •S 80 H c ^ 60 _> 1 40H 20 H 8b 243 228 in c 05 as 258 I i I i I i I 200 220 240 260 280 300 m/z in c CD 03 > _ca CD CC iooH 80 60 40 20 H 0 8d 243 227 -i 1 200 220 m c CD CD CC 2 6 0 2 7 1 240 260 280 300 m/z 100-80-- 60H •35 40H 20 H 8c 243 215 200 227 271 i 1 f 1 1 1 220 240 260 280 300 m/z 100J 8e 80 H - 60 "53 40H 20 258 243 228 1 274 - i 1—1 r ~ — i — " i i 1 1 200 220 240 260 280 300 m/z Figure 2.1(b) N2O negative DCI mass spectra of benz[a]anthracene (8b), chrysene (8c), triphenylene (8d) and tetracene (8e). The mass spectra of benzo [a] anthracene (8b) obtained at two different C O 2 pressures are illustrated in Figure 2.2. At the higher C 0 2 pressure (1.3 mbar) the relative intensity of the m/z 243 adduct [M + 0 - Ffj" decreased. Consequently, at a higher pressure, the m/z 258 adduct [M + 20 - 2H]" is present in a greater abundance relative to the m/z 243 adduct [M + O - H]". An additional difference that can be noted between the two spectra in Figure 2.2 is the presence of two clusters of ions at about m/z 232 and m/z 248, observed only in the 1.3 mbar spectrum. 47 100 H •I 80-c <D * 60-a> > 1 4 0 H cc 20 H 0.1 mbar 258 243 228 -X-271 o-200 220 240 260 m/z c CD CD > JO CD CC 280 300 100H 80 H 60 40 20 H 0 1.3 mbar 258 243 232 248 271 i ' i i i i i 1 i 200 220 240 260 280 300 m/z Figure 2.2 Effect of C02 pressure (0.1 mbar and 1.3 mbar) on the negative DCI mass spectra of benz[a]anthracene (8b). a) e x p e r i m e n t a l 10 c cu CD > CD 60 H = 40-20 H 0 T T b) ca l cu la ted £ CD CD > 0 s 200 220 240 260 m/z — I — i — I 280 300 100-80 H 60-•55 40H 20 0-wo, ~>—r w o . w o c 200 220 T 240 260 m/z — I 1 1 280 300 Figure 2.3 Experimental blank DCI filament desorption and calculated W0X isotope distribution. 48 These two clusters do not correspond to ion/molecule reaction products with the compounds in question. When blank desorption experiments (no sample loaded) were performed, these ion clusters were observed when the filament current exceeded 500 mA, corresponding to filament temperatures exceeding 1000°C (Figure 2.3(a)). These clusters correspond to m/z values calculated for the isotopic distributions of tungsten oxides (WOy) as shown in Figure 2.3(b) (the stable isotopes of W are: 180W(0.12 %), 182W(26.3 %), 183W(14.28 %), 184W(30.7 %),186W(28.6 %)). Apparently, the DCI tungsten filament released WOy" ions upon heating to a high temperature in the C 0 2 atmosphere of the ion source. Higher mass clusters were also observed corresponding to W xO y" ions for values of x ranging from 1 to 3 and y values ranging from 3 to 11. The formation of rhenium oxides (ReCV and ReCV) inside an ion source has been previously mentioned 8 7 as one of the important factors for the occurrence of surface catalysis reactions. However, the large amounts of tungsten oxides observed here from the DCI filament could also be responsible for some catalysis since W O 3 is a known catalyst for the formation of quinones from aromatic compounds 9 2 . The fact that a greater relative intensity of the m/z 258 adduct [M + 20 - 2H]" was observed at 1.3 mbar, when tungsten oxides are generated, provides some evidence that these compounds may also play a role in the formation of this adduct. The mass spectra of 8b and 8c, obtained by GC/MS using NCI with C 0 2 and N 2 0 , are shown in Figure 2.4. Comparison of these two spectra with those obtained by DCI, illustrated in Figures 2.1 (a) and (b), reveals some significant differences in the abundances of the adducts. For example, the m/z 258 adduct formed during the C 0 2 NCI of 8b is more abundant in the DCI spectrum (Figure 2.1(a)) than in the GC/MS spectrum (Figure 2.4). When substances are introduced into the ion source through a GC column there is no opportunity for tungsten oxides to be formed from the DCI filament. The only metal oxides 49 that may form are from the ion source's W/Re filament. Consequently, the lower abundance of the m/z 258 adduct during GC/MS analysis may be partially attributable to a lower abundance of the tungsten oxides. Comparison of the C 0 2 DCI data obtained in this work with that obtained by Elson 7 4 reveals the presence of many of the same type of adducts although with different relative intensities. Differences in source pressure, temperature, reagent gas purity, presence of air/water traces, formation of W or Re oxides, and ion source geometry are all factors which may affect the relative intensity of these ions. Source cleanliness will also be an important factor since accumulation of metal oxides on the surfaces of the ion source may affect any catalysis process that may be taking place. 50 a) NCI with CO. in c CD CP > JO V CC 1 0 0 8 0 6 0 4 0 H 2 0 J 8b 2 4 3 2 2 8 b) NCI with N 2 0 8b 2 4 3 CO C a) CD > JO ce cc 2 7 0 2 5 8 "1 1 1 2 0 0 2 2 0 2 4 0 2 6 0 2 8 0 3 0 0 m/z 2 4 3 cn c CD CD _> V * JO CD CC I O O H 8 0 6 0 -4 0 -2 0 0-8c x " T 2 7 0 cn C as CD > jo CD CC 2 5 8 T " "I 1 1 2 0 0 2 2 0 2 4 0 2 6 0 2 8 0 3 0 0 m/z 1 0 0 -8 0 -6 0 4 0 2 0 0-2 2 7 2 5 8 "1 1 1 1 0 0 8 0 -6 0 -4 0 -2 0 2 0 0 2 2 0 2 4 0 2 6 0 2 8 0 3 0 0 m/z 2 4 3 8c 2 2 6 i—r 271 "I \ f 1 1 1 2 0 0 2 2 0 2 4 0 2 6 0 2 8 0 3 0 0 m/z Figure 2.4 GC/ NCI-MS mass spectra for benz[a]anthracene (8b) and chrysene (8c) using a) C02 and b) N20 reagent gases. 2.3.1.2 Positional PAH isomers Environmental samples often contain a multitude of alkyl and heterogroup substituted P A H compounds in addition to unsubstituted PAHs. As mentioned in Chapter 1, for many of these compounds a large number of positional isomers is possible, with a variety of carcinogenic potencies. To investigate the behavior of positional isomers under NCI conditions, a series of mono and dimethyl benz[a] anthracenes were analyzed using N 2 0 as a 51 reagent gas. The mass spectral results obtained for the monomethylbenz[a]anthracene (MBAs) isomers are summarized in Table 2.2. Table 2.2 N 2 0 negative DCI mass spectral data for methylbenz[a]anthracenes (MBAs) 1 ^ 3 % Relative Intensity (RI) * PAH [M-H] (m/z 241) M (m/z 242) [M+O-H] (m/z 257) [M+20-2H] (m/z 272) [M+20-H-CH3] (m/z 258) 1-MBA 63 25 100 65 12 2-MBA 77 15 100 47 4 3-MBA 65 18 80 100 10 4-MBA 56 30 100 85 12 5-MBA 36 0 56 100 1 6-MBA 34 0 53 100 2 7-MBA 15 25 40 9 100 9-MBA 25 16 57 100 10 10-MBA 28 25 82 100 10 * The range of standard deviations for the % RI of the tabulated ions was 0 to 15 % (three replicate analysis) 52 Although differences can be noted in the mass spectral data for the nine M B A isomers, a variability of up to 15 % in the relative ion abundances was estimated by performing three repeat analyses. Thus, the only M B A that displays a distinguishing spectrum is 7-MBA, since it is the only isomer that forms the m/z 258 ion in significant amounts. This type of adduct has been observed previously using N2/O2 CI and has been attributed to the loss of the methyl group from the oxygen adduct, [M + 20 - H - CH3]~ 7 5 . The most plausible chemical explanation for this adduct is that addition of oxygen must be taking place across the 7 and 12 positions of benz[a]anthracene (8b). These two positions have the lowest localization energy (Figure 1.3) and therefore they are the most reactive positions. Such an oxidation would be expected to give rise to a 7,12-quinone, as illustrated previously in Figure 2.1(a). When the 7 and 12 positions of 5b are both occupied by H, loss of H occurs, and the [M + 20 - 2H]" adduct is observed, but when one of these positions is occupied by a methyl group, the [M + 20 - H - CH3]" adduct can form. Table 2.3 N 2 0 negative DCI mass spectral data for dimethylbenz[a]anthracenes (DMBAs) % Relative Intensity * PAH [M-H] M [M+O-H] [M+20-2H] [M+2O-2CH3] (m/z 255) (m/z 256) (m/z 271) (m/z 286) (m/z 258) 3,9-DMBA 55 36 84 43 5 6,8-DMBA 75 57 100 29 27 7,12-DMBA 0 16 5 0 100 * The range of standard deviations for the % RI of the tabulated ions was 0 to 15 % (three replicate analysis) 53 The results from the analysis of dimethylbenz[a]anthracenes (DMBAs), summarized in Table 2.3, demonstrate that the presence of methyl groups at both the 7 and 12 positions leads to the formation of a [M + 20 - 2CH3]" adduct in greater abundance than the [M + 20 - 2H]" adduct, thus enabling 7,12-DMBA to be distinguished from the other DMBAs. This differentiation is significant because 7,12-DMBA is one of the most potent carcinogens known 3 2 . Similarly, 7-MBA is the most potent carcinogen among the MBA's 9 3 . Based upon these DCI results, the use of NCI with N 2 0 can be a potentially useful technique for differentiating positional isomers that have groups at "active" (low localization energy) positions on a PAH ring. However, because the adduct responsible for this differentiation process appears to be a surface catalyzed species, 7 5 ' 8 7 its formation may be highly dependent upon the catalytic environment, as discussed previously (Section 2.3.1.1). 2.3.1.3 Limitations of DCI analysis The negative chemical ionization experiments discussed to this point have demonstrated that some structural and positional isomers can be differentiated by probe desorption analysis (DCI). Unfortunately, DCI is usually limited to the investigation of pure samples, thus diminishing its usefulness for analyses of complex environmental samples. However, care must also be taken in interpreting results even when "pure samples" are analyzed. For example, when a solution of anthracene (5a), (C14H10, Mr=178 u) prepared from 99+% zone refined anthracene (Aldrich) was analyzed by DCI with N2O, an intense, previously undocumented m/z 166 ion was observed (Figure 2.5(a)). In the mass spectra of PAHs obtained by CI it is unusual to observe intense ions at m/z values below the molecular ion. Furthermore, a similar analysis of dio-anthracene (C14D10, M r =188) (Figure 2.5(b)) did not yield a corresponding ion below M" (m/z 188). A positive EI analysis (via probe 54 desorption) of the same anthracene solution only revealed very small amounts, (<1 % RI), of m/z 166 & 167 ions (data not shown). Analysis of the anthracene sample by N 2 0/NCI via GC introduction provided a chromatogram with one large chromatographic signal corresponding to anthracene, (mass spectrum shown in Figure 2.5(c)) and a small signal with a mass spectrum having a base peak at m/z 166. These data clearly indicate that the intense m/z 166 signal observed in the DCI spectrum of anthracene (Figure 2.5(a)) is due to an impurity. This m/z 166 interference was identified as the [M - H]" ion of carbazole (M r = 167) (the structure of carbazole is illustrated in Figure 1.2). GC/MS analysis by positive EI indicated the presence of approximately 1 % of carbazole, confirming the 99 % purity statement by the supplier. Based upon the relative intensity of anthracene and carbazole ions in the mass spectra in Figure 2.5(a), it is clear that the response of carbazole under negative CI/N 20 conditions far exceeds that of anthracene. This large response of carbazole arises mainly from the formation of m/z 166 ions ([M - H]~) and can be attributed to the presence of an acidic proton on the nitrogen. Consequently, these results illustrate the potential usefulness of N 2 0 as a reagent gas for the analysis of aromatic compounds with acidic hydrogens. However, these results also highlight the problem that impurities can cause during DCI analysis, and indicate that care must be taken in interpreting DCI mass spectral data. For example, the m/z 215 ion in the N 2 0 DCI mass spectrum of chrysene, 8c, in Figure 2.1(b), is most likely an impurity since it is not observed when the sample is analyzed via GC introduction (Figure 2.4(b)). Consequently, although DCI can be a valuable tool in the analysis of individual compounds it is limited by the presence of impurities that need to be separated from the analyte by chromatography. 55 a) cn c CD 0) > J D CD CC 100 80 6 0 -4 0 -20 0 178 b) 166 T 193 cn c CD CD > CD CC 208 - i | i | i 100 80 60 H 40 20 0 202 188 216 160 180 200 220 240 m/z l 1 1 ' 160 180 200 220 240 m/z O co c CD CD > CD CC 100-1 80 60 40 2 0 - | 0 178 193 208 t — i 1 1 160 180 200 220 240 m/z Figure 2.5 Negative ion N2O DCI analysis of (a) "99+% zone refined" anthracene, (b) dio-anthracene; (c) GC/MS analysis of anthracene. 56 2.3.2 Gas Chromatography with Negative Chemical Ionization MS Analysis Use of a GC for sample introduction into the ion source is often an essential part of the analytical procedures for the analysis of complex mixtures, such as environmental samples. Consequently, as a test case, a mixture consisting of 16 standard PAHs was analyzed by using GC/NCI-MS with N 2 0 , N2O/CH4, C 0 2 and C 0 2 / C H 4 as reagents. Each of the total ion chromatograms (TIC) produced from these four experiments displayed only 14 resolvable signals. This was because compounds benzo[b]fluoranthene (9a) and benzo[k]fluoranthene (9b) eluted together as did indeno[l,2,3-cd]pyrene (10a) and benzo[ghi]perylene (10b). The relative response factors (RRFs) for the PAHs present in each of the chromatograms analyzed are summarized in Figure 2.6. The most striking feature notable in this data is the significantly higher RRF values from the higher molecular mass PAHs (9a/b to 10b). Of the lower mass PAHs, only acenaphthylene (3), fluorene (4) and fluoranthene (6b) gave RRF values greater than 0.2, and only with N 2 0 or N 20/CH4. With the exception of 4 which ionizes predominantly via proton transfer to yield [M - H]" ions, the majority of the remaining PAHs displaying large RRF values correspond to compounds with E A > 0.5 eV. Acenaphthylene (3) (EA = 0.40) appears to be the only major exception. PAHs with high E A values produce the largest RRF values because of their ability to form large amounts ofM" ions by electron capture. Comparison of the RRF values of the PAHs, obtained by using different reagent gases, reveals that the largest differences occur in the RRF values for acenaphthylene (3), fluorene (4) and fluoranthene (6b). When using N 2 0 or N2O/CH4 as reagent gases these three PAHs gave rise to significantly greater RRF values that when C 0 2 or C 0 2 / C H 4 were employed. 57 1 -0 .8 -u- 0.6 cc cc 0.4 0.2H 0 T i i i .IWdWlfk I T ~ r . l _ l . X l to XI ( 0 X 5 O T3 ^ 5 CT) 2 • N 2 0 • N 2 0 + CH 4 • c o 2 S C 0 2 + CH 4 Figure 2.6 Relative response factors (RRFs) of 16 standard PAHs analyzed by GC/NCI-MS using different reagent gases. (T = 240°C, eV = 85, Ie = 0.200 mA). 58 a) Intensity 8 5 0 0 0 7 0 0 0 0 A 6 0 0 0 0 5 0 0 0 0 -4 0 0 0 0 3 0 0 0 0 2 0 0 0 0 1 0 0 0 0 b) Intensity 60000 50000 40000 30000 20000 10000i 1 lt4 6b 1 3 , 2 7 4 7 . 5 9 5.66 3 9a+9b I0a+11 2 0 , 0 2 2 5 , 2 4 9d 1 0 . 2 4 5b 5a Ljiinfajll. i|iiin,i 8b 8c 6a V / 1 7 , 1 3 Jlfl . r Mdflluf « 10b 10 15 20 T i m e (min) 25 30 10a+11 25,11 4 7.54 6b 13.. 2 10, 5a I Ml 9a+9b 1 9 . 9 2 8b 8c 6a \ / 16,95 9d 10b 10 15 20 T ime (min) 25 30 Figure 2.7 Effect of ion source temperature on the N2O/NCI analysis of the 16 standard PAHs. (a) 175°C, (b) 280''C. 59 Several ion source parameters were varied to investigate possible effects on the RRFs of the 16 PAHs. The influence of ion source temperature is illustrated in Figures 2.7 (a) and (b). The only significant differences in these two chromatograms is the greater relative response for PAHs 10a and 11 at 280°C. Generally, higher M r PAHs give a lower response at lower ion source temperatures because of their lower volatility. Changes in electron energy between 50 and 100 eV were not observed to have a significant effect on the RRF values of the PAHs analyzed (data not shown). As illustrated in Figure 2.8, a higher emission current (0.250 mA) results in a greater response for the higher molecular mass PAHs, 9a to 10b. These PAHs possess a greater molecular cross section, resulting in a better ability to capture electrons emitted from the filament at 0.250 mA. These results clearly illustrate that it is important to use reproducible instrumental conditions when performing chemical ionization experiments and to verify the results by analysis of suitable standard samples. 60 a) Intensity 190000 150000 T0a+11 100000 50000H b) Intensity 60000 H 50000 40000 30000 20000 10000 Time (min) Figure 2.8 Effect offilament emission current on the NCI/N20 analysis of the 16 standard PAHs. (a) 0.250mA, (b) 0.200mA. 61 2.3.2. 1 Using differences in NCI relative response factors for isomer differentiation by comparison with EI responses Variations in RRF during NCI can be exploited to differentiate some P A H structural isomers. This is highlighted by the bar graph in Figure 2.6, as well as by the chromatograms in Figures 2.7 and 2.8. For example, fluoranthene (6b) has a much higher response than pyrene (6a) when using N 2 0 or N2O/CH4 reagent gases. Although these two isomers are chromatographically separable (see Figure 2.7) and can be identified by using available standards, identification of their methylated analogues is more difficult because of the lack of commercial standards and the fact that many of them co-elute in complex clusters. Since the addition of methyl groups reduces the E A of a compound by only a small amount8 4'9 4, methylfluoranthenes (Cl-6b) are also expected to have a greater response than methylpyrenes (Cl-6a), similar to their unsubstituted analogues (6b and 6a). Analysis of a P A H contaminated sediment extract confirmed this hypothesis. The N2O/NCI and positive EI chromatograms reproduced in Figure 2.9 show the portion of the total ion chromatogram where pyrene (6a), fluoranthene (6b) and their monomethyl analogues (Cl-6a/6b) elute. The signal corresponding to 6a in the NCI chromatogram displays a significantly lower intensity than 6a in the EI chromatogram. This decrease in signal intensity is expected based upon the lower RRF value of 6a with respect to 6b during N 2 0/NCI (Figure 2.6). The series of smaller signals appearing after 6a in the EI chromatogram correspond to the methyl 6a/6b, based upon their EI mass spectral data (i.e. base peak, m/z 216 = IVT ions). Comparison of the EI and NCI chromatograms in this retention region reveals that in the NCI data, some of these signals show a marked decrease in intensity, allowing the retention region of the methyl 6a isomers to be easily identified. 62 Figure 2.9 Comparison of a portion of the N2O/NCI (top) and positive EI (bottom) total ion chromatograms from the analysis of the B-lagoon sediment extract. This data illustrates the elution regions of unsubstituted and methyl (Cl) substituted fluoranthene (6b) and pyrene (6a) isomers. 63 Thus, in environmental samples, methylpyrenes (Cl-6a) can be distinguished from methylfluoranthenes (Cl-6b) by comparing the EI and N 2 0/NCI total ion chromatographic data. This N 2 0/NCI method demonstrates that differentiation of some unsubstituted as well as methylated P A H isomers, in an environmental sample, is possible without using any standards. Another P A H isomer which can be differentiated using this technique is benzo[e]pyrene (9c), as illustrated in Figure 2.10. This figure depicts the elution region of isomers 9a to 9e. Once again, a comparison of EI and NCI chromatographic data reveals that 9c can be distinguished from the other isomers. This differentiation is made possible by the fact that 9c has the lowest E A (0.534 eV) within this set of isomers. This low E A value results in a low response during NCI analysis. However, during positive EI analysis isomers 9a to 9e all have approximately the same response factor. Thus, by comparing the EI and NCI data it is possible to distinguish between isomers with low and high NCI response values. One of the advantages of negative CI over positive CI, for the analysis of compounds with high E A values, is the removal of interferences from substances, such as aliphatics, that have low E A values. This effect can be clearly observed by comparing the positive EI and CI data to the NCI data illustrated in Figure 2.11. In the NCI case, fewer signals are observed, corresponding mainly to high E A compounds. 64 9a+9b Figure 2.10 Comparison of a portion of the N20/NCI (top) and positive EI (bottom) total ion chromatograms from the analysis of the B-lagoon sediment extract. This data illustrates the elution regions of benzofluoranthenes (9a/9b), benzofejpyrene (9c), benzofajpyrene (9d), and perylene (9e). 65 a) 25 30 35 T ima 40 45 SO b) 10 IS 35 40 45 30 c) J—L-10 15 Figure 2.11 (a) EI, (b) CH/PCI, and (c) N2O-CH4/NCI TICs for a PAH contaminated sediment extract. 66 2.3.2.2 Mass Spectral Data In addition to evaluating the GC/NCI-MS results via RRFs, mass spectral data for the 16 PAHs analyzed were also extracted from the GC/MS chromatograms and are summarized in Tables 2.4 to 2.7. A few general trends can be noted in the data shown in these tables. Within each of the three pairs of isomers 5a/5b, 6a/6b and 8b/8c, the isomer with the higher E A yields the greater abundance of M" ions. Each isomer pair also displays notable differences in the relative intensities of adduct ions, enabling them to be differentiated from one another, as illustrated in Figure 2.12. Comparison of the mass spectral data in Tables 2.4 to 2.7 for C 0 2 , N 2 0 , N 20/CFLi and C0 2/CH4 respectively, reveals that these four CI experimental methods generate similar mass spectral data for many of the PAHs analyzed. The major differences observed were the larger amounts of the [M + 43]" and [M + 42]" adducts (see "other m/z" for [M + 42]" values) formed during C0 2 /NCI and CO2-CH4/NCI, and the production of [M - 2]" fragment ions (see "other m/z") during N 2 0/NCI and N 20-CH4/NCI. cc 100 -\ 8 0 A 60 H 4 0 2 0 0 i JD If) CO tL CO . a . co CO J 3 O O o oo • M-• [ M + 1 5] • IM ! 3 0 S IM + 4 3 ] Figure 2.12 N2O/NCI mass spectral data for three PAH isomer pairs. 67 Table 2.4 Mass spectral data (% RI) for PAHs analyzed by C0 2 NCI PAH EA [M-l] [M] [M+15] [M+30] [M+43] other (Mr, u) (eV) m/z (%) 1 (128) 0.14 0 0 100 0 31 3(152) 0.40 0 100 19 0 6 2(154) n/a 13 0 54 0 100 152 (45) 4(166) 0.28 61 13 30 0 14 208 (100) 5b (178) 0.3 0 0 100 0 24 220 (31) 5a (178) 0.55 0 8 100 8 25 220 (25) 6b (202) 0.6 0 100 19 0 4 244 (4) 6a (202) 0.55 0 5 100 0 18 244 (16) 8b (228) 0.630 0 9 100 9 20 270 (21) 8c (228) 0.397 0 0 100 16 21 270(28) 9a+9b (252) 0.7-0.6 0 100 4 0 0 9d (252) 0.68 0 100 21 0 1 267(8) 10a+ll 0.8/0.6 0 100* 1 0 0 (276/278) 10b (276) 0.51 0 100 3 0 0 n/a not available. * only m/z 276 ions (M" for 10a) were observed 68 Table 2.5 Mass spectral data (% RI) for PAHs analyzed by N 20 NCI PAH EA [M-l] [M] [M+15] [M+30] [M+43] other (Mr, u) (eV) m/z (%) 1 (128) 0.14 10 0 100 0 20 126 (20) 3 (152) 0.40 0 100 12 0 0 2(154) n/a 100 12 53 0 0 152 (51) 4(166) 0.28 100 13 12 0 0 164 (25) 5b (178) 0.3 16 4 100 0 11 176 (24) 5a(178) 0.55 22 31 100 6 7 176 (22) 6b (202) 0.6 2 100 11 0 0 200 (2) 6a (202) 0.55 20 17 100 0 6 200 (24) 8b (228) 0.630 30 15 100 8 4 226 (24) 8c (228) 0.397 26 10 100 2 6 226 (33) 9a+9b (252) 0.7-0.6 0 100 9 0 0 9d (252) 0.68 0 100 14 0 0 10a+ll 0.8/0.6 0 100* 1 0 0 (276/278) 10b (276) 0.51 0 100 9 0 0 n/a not available. * only m/z 276 ions (M" for 10a) were observed 69 Table 2.6 Mass spectral data (% RI) for PAHs analyzed by N 20/CH 4 NCI PAH EA [M-l] [M] [M+15] [M+30] [M+43] other (M n U ) (eV) m /z (%) 1 (128) 0.14 6 0 100 0 15 126 (19) 3 (152) 0.40 0 100 9 0 0 2(154) n/a 100 16 63 0 0 152 (87) 4 (166) 0.28 100 13 12 0 0 165 (25) 5b (178) 0.3 15 2 100 1 12 176 (28) 5a (178) 0.55 22 50 100 17 3 176 (26) 6b (202) 0.6 1 100 10 0 0 200 (2) 6a (202) 0.55 15 24 100 0 8 200 (24) 8b (228) 0.630 31 21 100 9 1 226 (23) 8c (228) 0.397 32 6 100 18 4 226 (30) 9a+9b (252) 0.7-0.6 0 100 5 0 0 9d (252) 0.68 0 100 23 0 0 267 (12) 10a+ll 0.8/0.6 0 100* 1 0 0 (276/278) 10b (276) 0.51 n/d n/d n/d n/d n/d n/a not available. * only m/z 276 ions (M" for 10a) were observed, n/d 10b not detected 70 Table 2.7 Mass spectral data (% RI) for PAHs analyzed by C0 2 /CH 4 NCI PAH EA [M-l] [M] [M+15] [M+30] [M+43] other (M„ U) (eV) m/z (%) 1 (128) 0.14 n/d n/d n/d n/d n/d 3(152) 0.40 0 100 21 0 6 2(154) n/a 8 0 61 0 95 152 (66) 4 (166) 0.28 81 8 34 0 14 208 (100) 5b (178) 0.3 0 0 100 1 25 220 (25) 5a (178) 0.55 0 4 79 7 22 220 (17) 6b (202) 0.6 0 100 20 0 4 244 (3) 6a (202) 0.55 1 15 100 0 21 244 (14) 8b (228) 0.630 0 14 100 13 28 270(18) 8c (228) 0.397 0 0 100 31 33 270 (22) 9a+9b (252) 0.7-0.6 0 100 4 0 1 9d (252) 0.68 1 100 7 0 1 10a+ll 0.8/0.6 0 100 1 0 0 (276/278) 10b (276) 0.51 1 100 2 0 0 n/a not available. * only m/z 276 ions (M~ for 10a) were observed, n/d not detected 71 2.3.2.3 Reactions with gas phase reagent ions In order to investigate more closely the nature of the reagent ions and how they interact with PAHs, some GC/MS analyses were performed using a mass scan range low enough to include the m/z values of the reagent ions. These analyses enabled the concentrations of the reagent ions to be monitored during the elution bands of the PAHs. The major ions observed for each of the reagent gases employed are shown in Table 2.8. The addition of CH4 to N 2 0 results in the formation of OH" ions (m/z 17) as described previously by Harrison9 1. Addition of CH4 to CO2 did not yield any new ions. The largest decreases in the signal intensity of the reagent ions investigated was observed for m/z 16, during the C 0 2 analysis, as illustrated in Figure 2.13. Depicted in this figure are the ion chromatograms for the PAHs (m/z 100-350) and O" (m/z 16 ). Table 2.8 Reagent gas ions observed during NCI Reagent gas m/z values of ions observed Identity of ions N 2 0 16, 30 > 32 16 = 0", 30= NO", 32 = 02" N 20/CH4 16, 30>32, 17 17 = OH" co 2 16 » 60 16 = 0", 60 = CO 3" C0 2/CH4 16 » 60 72 m/z 16 T ime (min) Figure 2.13 Ion chromatograms from the GC/CO2 NCI-MS analysis of 16 PAHs: m/z 16 chromatogram (top) and m/z 100-350 ion chromatogram (bottom). 73 The m/z 16 ion chromatogram appears to be almost a mirror image of the P A H chromatogram (m/z 100-350), except in the region where the high molecular mass PAHs elute ( t R = 23 - 30 min). Decreases in the m/z 16 ion chromatogram are most likely a result of a depletion of the m/z 16 ions during ion/molecule reactions (e.g. M + O" -» [M + O - H]" + H) with the eluting PAHs. As a result of their large EA, high molecular mass PAHs can undergo electron capture reactions readily, and consequently this process competes successfully with ion/molecule reactions, as illustrated by the low abundance of adduct ions in their mass spectral data (Tables 2.4 to 2.7). Figure 2.14 illustrates the PAH and m/z 16 ion chromatograms for the N 2 0/NCI analysis. In this instance, the m/z 16 signal is not observed to undergo significant decreases during the elution of the PAHs. The only reagent ion that showed a decrease in intensity during N 2 0/NCI analysis was m/z 32 (02~). This decrease in the m/z 32 signal can be observed more clearly in the expanded view of the 6a / 6b region shown in Figure 2.15. It is interesting to note that the m/z 32 intensity is observed to decrease only during elution of 6b. Perhaps this decrease is due to electron transfer reactions taking place between 02' and 6b, since the major N 2 0/NCI product of 6b is M" (Table 2.5). 74 Figure 2.14 Ion chromatograms from GC/N2O NCI-MS analysis of 16 PAHs: top to bottom, m/z 30, m/z 16, m/z 32 and m/z 100-350 ion chromatograms. Figure 2.15 Expanded view of 14.8 - 16.3 min region of Figure 2.14. 76 During N2O-CH4/NCL hydroxyl ions (OH", m/z 17) are generated (Table 2.8). The hydroxyl ion, being a strong base, is expected to behave as a good proton abstractor. In fact, when a GC/N2O-CH4 NCI-MS analysis was performed under conditions of maximum production of m/z =17, the main product ions generated were [M - H]", M", and small amounts of [M + 43]" (possibly [M + N 2 0 - H]") as summarized in Table 2.9. These results are consistent with OH" being a good proton acceptor. However, only small differences can be noted in the mass spectral data shown in Table 2.9. These differences are mainly in the relative intensities ofM" and [M + 43]" ions. Furthermore, the total ion chromatogram for this analysis, shown in Figure 2.16, demonstrates that under these conditions, the relative response factors of isomeric compounds, such as 6b and 6a, are similar (see also RRF values in Table 2.9). The previous N 2 0 or N2O/CH4NCI analyses performed (with m/z 16, 30 > m/z 17), displayed a significantly higher response factor for 6b than 6a (see Figure 2.7). From these results, it is apparent that the OH' ion is not by itself a useful reagent for isomer differentiation purposes, as it does not yield sufficient differences in the mass spectral data or relative response factors of isomeric compounds. 77 Table 2.9 Mass spectral data for the GC/NCI-MS analysis of 16 PAHs using N 20-CH 4 with the m/z 17 (OH) ion maximized PAH [M-H] M [M+43] RRF 1 n/q n/q n/q n/q 3 100 76 11 1.0 2 100 12 0 0.76 4 100 10 0 0.75 5b 100 14 7 0.45 5a 100 17 3 0.55 6b 100 36 4 0.33 6a 100 16 7 0.31 8b 100 18 1 0.26 8c 100 19 2 0.31 9a+9b 76 100 1 0.49 9c 69 100 1 0.58 10a+ll 20 100 0 0.69 10b 95 100 1 0.49 n/q not quantifiable 78 5b 5a 9a+9b 10a+11 10 15 20 Time (min) Figure 2.16 Total ion chromatogram for the GC/NCI-MS analysis of 16 PAHs using N2O-CH4 with the m/z 17 (OH) ion maximized. 79 2.4 SUMMARY Through differences in relative response factors and mass spectral signatures, chemical ionization in the negative ion mode, using N 2 0 or C 0 2 as reagent gases, has enabled the differentiation of some isomeric PAHs. Of particular interest was the ability to differentiate 7-M B A and 7,12-DMBA from other positional isomers. However, the mass spectral behavior of the PAHs analyzed by using these reagent gases appears to be strongly dependent on the method of sample introduction. For example, when the PAHs were desorbed from a tungsten wire filament (DCI), a greater abundances of adducts of the type [M + 20 - 2H]" was observed for 8b, than when gas chromatography was used for sample introduction. Isomer differentiation was also achieved by exploiting differences in the relative response factors of isomers during GC/MS analysis. Through comparisons between GC/NCI-MS data and GC/PEI-MS data differentiation of isomers present in environmental samples was demonstrated by the analysis of the isomer groups 6a/6b and 9a to 9e. The application of the NCI techniques discussed in this chapter was limited by the sensitivity of the instrument employed. Through the utilization of a GC/MS instrument with higher sensitivity these techniques should provide improved differentiation capabilities in the analysis of PAHs. A better understanding of the surface catalysis process and how to regulate it, in order to maximize the formation of characteristic adducts, may also lead to improvements in the analysis of P A H isomers by negative chemical ionization. 80 CHAPTER 3 INVESTIGATING THE USE OF DIMETHYL ETHER AS A CHEMICAL IONIZATION REAGENT FOR THE ANALYSIS OF PAHs 81 3.1 INTRODUCTION Early work by Keough 6 8 demonstrated that ionization of dimethyl ether (DME, C2H 60) in a conventional mass spectrometer ion source, under positive ion conditions, yielded C 2 H 5 0 + ions (m/z 45) which could form [M + 45]+ and [M + 13]+ adducts with some aromatic compounds. For example, anthracene (5a) formed a [M + 45]+ adduct that was not observed for phenanthrene (5b). The mechanism for the formation of the [M + 45]+ adduct of 5a was postulated to involve a gas phase Diels-Alder type cycloaddition across the 9,10 positions of 5a, as shown in Figure 3.1 6 8 . The [M + 13]+ adducts are believed to be generated as a result of C H 3 O H elimination from [M + 45]+, as illustrated in Figure 3.1 6 8 . Consequently, the lack of a m/z 223 ion in the mass spectrum of 5b suggests that the [M + 45]+ adduct is unstable and undergoes elimination before it can be observed. The lower stability of the [M + 45]+ adduct of 5b has been attributed to the greater loss of resonance energy 5b has to undergo upon addition of the CH2OCH3+ ion 6 8 . Similarly, among the series of compounds 8b to 8e Keough only observed the [M + 45]+ adducts for PAHs 8b and 8e in appreciable amounts. Once again, these two compounds are the isomers that require the smallest loss of resonance energy during cycloaddition. D M E 6 7 , 6 9 and d6-DME 6 9 have also been employed by other researchers for the analysis of PAH isomers. However, none of these studies have reported on the application of DME/CI for the analysis of PAHs in environmental samples. This seemingly good selectivity provided by the C 2 H 5 0 + ion suggested that dimethyl ether (DME) might be a useful CI reagent for the differentiation of P A H isomers in environmental samples. Consequently, a D M E chemical ionization investigation of PAHs both in standard solutions and in extracts from environmental samples, was initiated using a linear quadrupole mass spectrometer. At a later stage of the work an ion trap mass 82 spectrometer became available and the experiments were repeated with this instrument. It was hoped that the significantly higher sensitivity of the ion trap mass spectrometer, over the linear quadrupole instrument previously used, would enable lower limits of detection, thus allowing for the analysis of a greater variety of PAH isomers in environmental samples. Furthermore, an ion trap enables the isolation of individual reagent ions, allowing for selective ion/molecule reaction experiments to be performed. Through such experiments it may be possible to maximize the yields of adduct ions, such [M + 45]+, that enable some P A H isomers to have characteristic mass spectral signatures. The results obtained reveal some important differences in the formation of adducts in the chemical ionization environment of an ion trap versus a standard ion source. 5 b UNSTABLE STABLE m/z 178 m/z 223 m/z 191 [M + 45]+ [M+13] + Figure 3.1 Possible mechanisms for the reaction of anthracene (5a) or phenanthrene (5b) with CH2OCH3+ ions. 83 3.2 EXPERIMENTAL 3.2.1 Materials The sample mixture of 16 standard PAHs, and the Kitimat sediment extract as described in Section 2.2.1, were used. The dimethyl ether reagent, D M E (99+%), was obtained from Aldrich (Milwaukee, WI, USA). 3.2.2 Instrumentation a) Linear Quadrupole Mass Spectrometer Standard CI ion source experiments were performed using a Delsi-Nermag R10-1C single quadrupole mass spectrometer operated in positive ion mode. The major operating parameters are listed in Table 3.1. Mass calibration was performed by using perfluorotributylamine (FC43). The D M E vapours were released at a pressure of 5 psi from a gas cylinder and introduced into the ion source via a needle valve. Table 3.1 Operating parameters for the Delsi-Nermag R10-10C MS employed during dimethyl ether chemical ionization experiments Parameter Values used Electron energy 70 eV Emission current 0.200 mA Source temperature 160°C and 240°C CI Source pressure 0.1 mbar Analyzer pressure 10"7 mbar Multiplier 1.30 kV Mass scan range 100-350 m/z Scan rate 200 ms / scan 84 b) Quadrupole ion traps The ion trap GC/MS experiments were performed using both a Varian Saturn 3D (Varian, Walnut Creek, CA, USA) ion trap and a Varian QISMS ion trap. The former instrument has only an r.f. potential applied to the ring electrode while the latter can also employ the use of an additional d.c. potential to the ring electrode. With either instrument, the D M E was introduced via the CI reagent gas inlet, and the flow rate into the ion trap was regulated via a needle valve. c) Saturn 3D ion trap mass spectrometer The experimental parameters used for the Saturn 3D ion trap experiments are summarized in Table 3.2 Table 3.2 Major operating parameters employed during dimethyl ether CI in the Saturn 3D ion trap Parameter Value(s) used Electron energy 1 70 eV Emission Current 10 pA Maximum Ionization Time (MIT) 2000 ps Reaction Time (RT) 128 ms Automatic Reaction Control (ARC) target 5000 Ionization Storage Level (ISL) 5.0 m/z Reaction storage level (RSL) 20 m/z Mass scan range 100 - 350 m/z Scan rate 600 ms/scan Trap temperature (°C) 240°C Source pressure2 « 10"5 mbar (1) The electron energy cannot be regulated on this instrument. (2) No pressure gauge was available on instrument. Pressure indicated is according to manufacturer's specifications. 85 d) QISMS ion trap When using a Saturn 3D ion trap, only ions with m/z values greater than 60 can be isolated selectively through the application of the MS/MS isolation wave-forms. When using a QISMS system, ions below m/z 60 can be trapped through the application of both r.f. and d.c. potentials to the ring electrode and a high frequency supplemental r.f. potential to the end caps. Application of a positive d.c. ring electrode potential so that the value of the trapping parameter az < - 0.15 enables the storage of only positive ions (ions are stored outside of the negative ion stability region), thus avoiding reactions with any negative ions that might be formed 9 5 . Furthermore, through the application of the correct combination of d.c. and r.f. potentials, a narrow mass range of ions can be isolated. The isolation of m/z 45, m/z 47 and m/z 61 ions, corresponding to the C H 2 O C H 3 + , C2H 7 0 + and C 3 H 9 0 + ions of D M E , was carried out using the scan functions A to C illustrated in Figure 3.2(b). An example of the procedure used to calculate some of the parameters listed in scan function A is illustrated in the following sample calculations. Sample calculations of ion trap scan function parameters: Given a trap ring electrode r.f. voltage frequency = f = 1.05 MHz, the angular frequency O is given by Cl = 2nf= 6.6 x 106 radians / s. For an ion with m/z = 47 inside an ion trap with a ring electrode radius r 0 = 0.0100 m and z 0 = 0.00707 m (one half of the separation distance between the end-cap electrodes), and for applied voltages of 67 volts d.c. (U) and 428 volts* r.f. (V), the values for the trapping parameters az and q z can be determined using equations 1.3 and 1.4, where m represent the mass of a singly charged ion: * Note. The a.c. voltage in the QISMS trap is specified in D A C units. 1 V = 1.6407 DACs, consequently 428 V correspond to 261 DACs. 86 a = -\6eU = -16(1.602xlO-^C)(67n = _ Q 2 5 2 ™(r2,+2zl)Cl2 g^^^)((0.0100/i i)2+2(0.00707 /»)2)(6.6xlO5i-0 2 8eF 8(1.602x10-'gQ(428F) = Q 8 1 Ht m(r02+2z20)a2 g^l°J^?^(o .010M 2+2(0.00707i«)2X6.6xlOS5-0 2 Because the values of azm and qzm are constant at constant r.f. and d.c. voltages: azm = -0.25 x47u = -12u qzm = 0.81 x47u = 38u the value of m at any particular point (qz, az) on the stability diagram can be calculated by: -12w 38M m or m [3.1a, 3.lbj The mass range that can be stored under these trapping conditions can be determined by calculating the values of m according to equation 3.1a or 3.1 b at each of the intersection points on the d.c./r.f. = 67/428 line (dashed line) shown in Figure 3.2(a). For example, at (qz, az) = (0.67, -0.21), m = 38u/0.67 = 57u m= -12 u/-0.21 = 57 u andat(qz, az) = (1.24, -0.38), m = 38u/1.24 = 31 u m = -12u/-0.38 = 31 u Consequently, the mass range stored under these trapping conditions will be m/z 31 to m/z 57, as indicated in Figure 3.2(a). Within this mass range only the m/z 45 and m/z 47 D M E ions will be stored. In order to isolate the m/z 45 ion by itself, the m/z 47 ion must be ejected. This can be accomplished by applying a supplementary r.f. voltage to the end caps (axial 87 excitation) with a frequency corresponding to the fundamental secular frequency of m/z 47 at (qz, az) = (0.81, -0.25). The axial (z) secular frequency of ions stored at this particular point in the stability region is given by equation 1.5, co = Pzf/2. The pz value at (qz, az) = (0.81, -0.25) is approximately (3 z = 0.25, based upon an interpolation of the pz lines in the ion trap stability diagram (see Figure 1.7) (the (qz, az) = (0.81, -0.25) point lies half-way between the 0.2 and 0.3 pz lines). The value of P z can also be estimated according to equation 1.6: However, the above equation is only valid for q z < 0.4. For the given value of q z = 0.81 the value of pz estimated by the above equation is too high. Therefore, the axial (z) secular frequency of ions stored at pz = 0.25 is: Consequently, if a 131.3 kHz frequency is supplied to the end-caps (axial excitation), at a sufficiently high voltage, the amplitude of motion of the m/z 47 ions will sufficiently increased to cause their ejection. A value of 5 volts was found to be sufficient to achieve ejection. [1.6] co = pzf/2 = (0.25)(1.05 x 106 Hz)/2 = 131.3 x 103 Hz = 131.3 kHz. 88 Figure 3.2(a) Ion trap stability diagram indicating the m/z and (qz, af values at different points on the d.c. /r.f. = 67 /428 line. U = 67 volts d.c. and V =428 volts r.f. at (qz, af = (0.81, -0.25). RF (Volts) 1 428 Isolation Mass Scan DC (Volts) 131.3 kHz I 303 -\ 0 -Reaction Period Time (ms) — H H H 1 0 1 2 5 6 126 (0-1 ms, m/z > 30; 2-5ms, 31 < m/z < 57 - eject 47; 6-126, m/z >30) B RF (Volts) J 468 428 303 DC (Volts) | 6 J -Time (ms) ; / / ' 1 / 1 1 / ' 1 / 1 ! / i i \ Mass Scan \ Reaction Period i i i i i i i i i i i i • - t - H — H — H -0 1 2 5 6 9 10 130 (0-1 ms, m/z > 30; 2-5ms, 31 < m/z < 57; 6-9ms, m/z >46; 10-130ms, m/z > 30) Mass Scan RF (Volts) { 5 1 2 -4 5 4 -428 -Rx. Period 2 Rx. Period 1 DC (Volts) 131.3 kHz 303 — 67_ ._ 5 7 - - -f , I 0 I I I Time (ms) 1—|—| (—| 1 1 0 1 2 5 6 126 2 4 6 (0-1 ms, m/z > 30; 2-5ms, 31 < m/z < 57 - eject 47; 6-126ms, 35 < m/z < 76; 126-246ms, m/z > 50) Figure 3.2(b) Scan functions employedfor the isolation of ions using the QISMS trap. 90 3.2.3 Sample Introduction Gas chromatography was used for sample introduction into the Nermag mass spectrometer as described in Chapter 2. In some cases, a few individual standards were also analyzed by probe desorption, and these data are identified in the experimental discussion. The quantity of individual compounds analyzed per experiment was of the order of 200 ng. Sample introduction into the Saturn ion trap mass spectrometer was always performed by using gas chromatography. A DB5 capillary column (30 m x 0.25 mm I.D., 0.25 pm film thickness) from J & W, (Folsom, CA, USA), was inserted directly into the ion trap through a transfer line maintained at 280-290°C. The samples were introduced via splitless injection at 290°C. The GC column temperature program was as follows: initial temperature of 90°C for 0.1 min, a 6°C min"1 ramp to 280°C, hold at 280°C for 1 min, ramp to 300°C at 20°C min"1, hold for 20 minutes. Helium carrier gas was used at a linear velocity of 32 cm s"1, corresponding to a flow of approximately 1 mL min"1 into the ion trap. The quantity of individual compounds analyzed with the ion traps was approximately 100 ng. 91 3.3 RESULTS AND DISCUSSION 3.3.1 DME Chemical Ionization in a Standard Linear Quadrupole MS Ion Source The major ions produced during electron ionization of D M E (C 2 H 6 0) are listed in Table 3.3, according to their relative intensities at varying source pressures. Protonation of a D M E molecule yields m/z 47 ions (C2F£ 70+), while hydride ion abstraction results in the formation of m/z 45 ions (C 2 H 5 0 +). Ions resulting from the addition of C H 3 + to C 2H60 are also observed at m/z 61 (C 3 H 9 0 + ) . The m/z 91 and m/z 93 ions correspond to gas phase dimer formation between C 2 H 6 0 and either C 2 H 5 0 + or C 2 H 7 0 + . The formation of these ions is in agreement with the results obtained by Keough 6 8 . Table 3.3 Intensities of ions generated from electron ionization of DME at various ion source pressures Intensity Counts x 10 3 Pressure (torr)* m/z 45 (C2H50+) m/z 47 (C 2H 70) m/z 61 (C3H90+) m/z 91 (C 4 H„0 2 + ) m/z 93 (C 4H 1 30 2 ) 0.07 25 75 54 112 290 0.08 651 194 82 133 289 0.10 212 80 57 163 283 0.25 142 47 36 233 268 0.50 45 24 17 172 150 * 1 torr =1.33 mbar, measured on a thermocouple pressure guage. The abundance of the m/z 45 (C 2 H 5 0 +) species shows a maximum intensity at a pressure of 0.08 torr. Because the m/z 45 ion can be used for P A H isomer differentiation 6 8 ' 6 9 , 92 a pressure of 0.08 torr was chosen to carry out the GC/MS experiments. The mass spectral results obtained for the PAHs analyzed at this pressure are listed in Table 3.4. Table 3.4 Dimethyl ether positive chemical ionization mass spectra of PAHs. Ion source conditions: 0.08 torr, and temperature of 160°C or (240°C) % Relative Intensity a (± 3 %) (values in brackets are for data obtained at 240°C) Retention time (min) PAH (Mr, u) M + [MH]+ [M+13]+ [M+45]+ 4.1 1 (128) 3 5 ( 4 3 ) 32 (56) 100 (100) 0 ( 0 ) 9.1 3 " (152 ) 38 (48) 6 8 ( 1 0 0 ) 20 (25) 100(100) 9.7 2(154) 28 (32) 60 (70) 100(100) 1 5 ( 6 ) 11.3 4(166) 27 (30) 50 (70) 100(100) 0 ( 0 ) 14.4 5b (178) 40 (30) 55 (70) 100(100) 0 ( 0 ) 14.6 5a (178) 60 (50) 100 (100) 9 8 ( 1 0 0 ) 67 (40) 18.3 6b (202) 30 (35) 60 (75) 100 (100) 0 ( 0 ) 19.0 6a(202) 58 (50) 86 (90) 100(100) 45 (25) 22.9 8b (228) 62 (42) 100 (100) 86 (85) 47 (12) 23.1 8c(228) 30 (40) 58 (85) 100 (100) 0 ( 0 ) 26.3 9a/9b (252) 46 (44) 100(100) 82 (100) 1 4 ( 1 4 ) 27.4 9d (252) 48 (65) 100 (100) 10 (50) 40 (56) - 11c (278) - (60) - (100) - (82) - ( i o ) (a) Uncertainty in % RI values is (+ 3 %) (b) Other ions produced: m/z 167 = [M+15]+, 20 % (30 %) (c) Compound analyzed only by DCI at 240°C. These data clearly illustrate the isomer selectivity of the [M + 45] adduct. For example, the isomer pairs 5a/5b, 6b/6c, 8b/8c, (9a+9b)/9d all show significant differences in the % RI of [M + 4 5 ] + ions formed. Except for compound 9d, a higher source temperature (% RI values in brackets) results in lower relative intensity for the [M + 4 5 ] + adducts, as was observed by Keough 6 8 . However, if the source temperature is kept too low during GC 93 analysis of PAHs, the chromatographic resolution of the higher M r compounds will suffer. Consequently, it was decided that 240°C would be used for the source temperature during the analysis of the environmental sample extracts. As mentioned in the introduction, (see Figure 3.1), the mechanism for the formation of the two major adduct ions, [M + 45]+ and [M + 13]+, has been speculated to occur via a Diels-Alder cycloaddition of C 2 H 5 0 + (m/z 45) followed by elimination of C H 3 O H 6 8 . Keough's hypothesis of a Diels-Alder mechanisms rests on the observation that compounds with 9,10 anthracene-like units, such as 5a, 8b, and 8e, that have been observed to undergo cycloaddition in solution, yield stable [M + 45]+ adducts 6 8 . However, other structural features may account for the formation of adducts in compounds such as pyrene (6a), that do not possess 9,10 anthracene-like units. For example, formation of the [M + 45]+ adduct ions may take place through an electrophilic attack of the C H 3 O C H 2 + . The hypothesis for the elimination of C H 3 O H from [M + 45]+ to form [M + 13]+ ions received further support from the analysis of dio-anthracene (M r = 188). The observation of m/z 200 ions during DME/CI of dio-anthracene confirmed the elimination of 33 mass units from m/z 233 ([M + 45]+), corresponding to a loss of CH 3 OD. The relative intensities of the [M + 45]+ adducts shown in Table 3.4 are lower than those reported by Burrows 6 9 For example, Burrows reported a 100 % RI value for the [M + 45]+ ion of 5a. These differences can be attributed to variations in experimental conditions. Experiments by Burrows were performed at much higher pressures (6-7 torr) and lower temperatures (80°C). The results in this work were more in line with those of Keough 6 8 , whose experiments were conducted at 0.2 torr and 250°C. Although lower source u temperatures result in greater [M + 45]+ adduct formation 6 8 , analysis of PAHs by GC/MS requires temperatures much higher than 80°C for most PAH species. Despite differences in 94 relative intensities, the relationship between the type of adduct ions formed and the P A H isomers is the same as reported by either Keough or Burrows. The only major difference noted was in the mass spectrum of acenaphthylene (3) obtained by Burrows. Burrows 6 9 reported the appearance of only m/z 243 ions corresponding to a [M + 91]+ adduct. The data in Table 3.4 indicate that 3 formed IVf, M H + , [M + 13]+, [M + 15]+ and [M + 45]+ ions but no [M + 91]+ ions. The higher pressures (6-7 torr) used by Burrows can account for the formation of [M + 91]+ since the m/z 91 reagent ions are more abundant at these high pressures 6 8 . The fact that 3 was observed to form the largest abundance of [M + 45]+ adduct ions out of all the PAHs analyzed (Table 3.4), indicates that this compound undergoes adduct formation more readily than the other PAHs. Analysis of the 1,2-dihydro analogue of 3, (acenaphthene (2)), showed a significantly lower relative intensity of [M + 45]+ ions (Table 3.4). This variation in adduct formation between the two compounds could be due to differences in their structures or ionization energies ( IE(2) = 7.68 eV, IE(3) = 8.22 eV ), or possibly a combination of the two. As has been suggested by Keough 6 8 , the addition of C H 2 O C H 3 + to 5a or 8b to form [M + 45]+ may be occurring across the 9,10 positions of 5a and the 7,12 positions of 8b. If this type of mechanism is important, the presence of methyl groups at these positions may affect [M + 45]+ adduct formation, and possibly enable some positional isomers to be differentiated. In order to investigate this possibility, a series of methyl and dimethylbenz[a]anthracenes (MBAs and DMBAs) were investigated as a large number of positional isomer samples were commercially available. Furthermore, these compounds are of important toxicological relevance. 7-MBA and 7,12-DMBA are very potent carcinogens as discussed in Chapters 1 and 2. The data obtained for the analysis of MBAs and DMBAs by DME/DCI are summarized in Table 3.5. 95 Table 3.5 Dimethyl ether positive chemical ionization mass spectra of MBAs and DMBAs. (DCI sample introduction, 140°C) % Relative Intensity a Compound (M r, u) ivT [M+13]+ [M+45]+ 8b (228) 86 100 88 50 1-MBA (242) 96 100 77 32 2-MBA (242) 100 97 56 21 3-MBA (242) 100 95 60 27 4-MBA (242) 100 87 61 22 5-MBA (242) 100 90 59 21 6-MBA (242) 100 91 54 26 7-MBA (242) 97 100 58 44 9-MBA (242) 100 78 54 22 10-MBA (242) 100 88 63 24 7,12-DMBA (256) 100 86 33 28 3,9-DMBA (256) 100 76 33 23 6,8-DMBA (256) 88 100 60 22 (a) Uncertainty in the % RI values is approximately 3 %. The presence of methyl substituents on 8b appears to result in a slight decrease in the relative intensities of the [M + 13]+ and [M + 45]+ adducts formed. Also, the mass spectra of 7-MBA and 7,12-MBA display small differences from the mass spectra of the other positional isomers. For example, the mass spectrum of 7-MBA shows a slightly greater abundance of [M + 45]+ ions than the other eight MBAs. Similarly, 7,12-MBA displays a slightly greater abundance of [M + 45]+ ions than the other two DMBAs analyzed. However, these differences are close to the uncertainty of the analytical method. If the addition of CH 2 OCH3 + is indeed occurring predominantly across the 7,12 positions, the presence or lack of methyl groups at these positions does not appear to have a pronounced effect. Even if the small 96 differences observed between the mass spectra of different positional isomers are not purely random fluctuations, they are too small to be of practical use from an analytical prospective. As these data reveal, the differentiation of positional isomers remains a much more difficult task than the differentiation of structural isomers. Although D M E CI does not appear to be useful for differentiating P A H positional isomers, its ability to differentiate among structural isomers can still be of practical importance. Figure 3.3 illustrates the GC/MS TIC from the analyses of a standard 16 P A H mixture and a P A H contaminated sediment extract. The mass spectral data for the PAHs identified in the sediment extract are listed in Table 3.6. 97 a) Intensity 5 0 0 0 0 - 2 4 5 0 0 0 4 0 0 0 0 -3 5 0 0 0 3 3 0 0 0 0 -2 5 0 0 0 -1 4 2 0 0 0 0 -1 5 0 0 0 1 0 0 0 0 - i •-5 0 0 0 n |:L . > u-6 5 10 5b \ 6b 6a V I J 5a 9a+9b 8b \ 8c 9d y 1 5 20 2 5 T i m e (min) 3 0 3 5 40 10 1 5 f 2 0 25 16.2 T i m e (min) 3 0 3 5 4 0 Figure 3.3 GC/MS DME CI total ion chromatograms for the analysis of (a) a standard PAH mixture and (b) a PAH contaminated sediment extract. Source temperature was 240°C. 98 Table 3.6 Dimethyl ether positive chemical ionization mass spectra of PAHs identified in the B-lagoon sediment extract (Figure 3.3(b)). Ion source conditions: 0.08 torr, 240°C % Relative Intensity a Retention Time (min) Compound b (Mr) ivT MH+ [M+13]+ [M+45]+ 9.6 2(154) 35 80 100 0 14.4 5b (178) 38 70 100 0 14.6 5a (178) 45 100 100 40 16.0-16.1 Cl-5b (192) 20 85 100 0 16.2 Cl-5a (192) 0 100 15 12 16.3 (190) 40 70 100 0 16.4-16.5 Cl-5b (192) 0 70 100 0 18.2 6b (202) 30 68 100 0 19.0 6a (202) 48 80 100 22 19.8 Cl-6b or 7a/7b (216) 9 90 100 0 20.0 Cl-6b or 7a/7b (216) 32 76 100 0 20.3 Cl-6a (216) 46 90 100 5 21.3 C2-6b (230) ? c 0 44 100 0 21.5 C2-6b (230) ? c 0 84 100 0 22.8 8b (228) 53 100 86 25 23.0 8c(228) 39 80 100 0 23.2 8e d e (228) 46 100 74 46 26.3 9a-9b (252) 46 67 100 22 26.6 ? (252) 5 100 15 50 27.1 9c e (252) 50 100 90 35 27.3 9d (252) 55 100 52 55 27.6 9e e (252) 52 100 0 90 (a) Uncertainty in % RI values ranges from 0 to 10 %. (b) Cl- refers to a monomethyl substitution, (c) Although the [M+45]+ adduct is not present, the low count intensity of these signals makes this identification uncertain, (d) Although this standard was not analyzed, the identification was based upon mass spectral data reported by Keough 6 8 and Burrows 6 9 . (e) Based upon the relative retention times of these compounds. 99 Comparison of the data in Table 3.6 and Table 3.4 (at 240°C) reveals that the mass spectra for many of the unsubstituted PAHs present in the sediment extract are similar to those for the standard PAHs shown in Table 3.4. These similarities allow for some of the structural isomers to be identified based upon not only their retention time, but also their mass spectra. However, differentiation of their methyl substituted analogues, such as methylphenanthrenes (Cl-5b) and methylanthracenes (Cl-5a), is more difficult because many of these compounds elute in poorly-resolved clusters. Nonetheless, the presence of characteristic ions, such as [M + 45 ] +, can allow isomer differentiation of some of these compounds. For example, the compound eluting at 16.2 min cannot be identified as being either a Cl-5b or a Cl-5a based upon its EI spectrum. However, the presence of a [M + 45]+ ion (m/z = 192 + 45 = 237) in the D M E CI spectrum (Table 3.6) enables this compound to be identified as a Cl-5a. Similarly, the presence of a [M + 45]+ ion in the mass spectrum at 20.3 min indicated the presence of a methylpyrene (Cl-6a). Because many of the lower-abundance PAHs present in the sediment extract were below the detection limits ( « 10-50 ng) of the instrument employed, the full potential of D M E as a CI reagent gas for differentiating P A H isomers in environmental samples could not be tested. Consequently, when an ion trap mass spectrometer became available (Varian Saturn 3D GC/MS), some of the D M E CI experiments were repeated. It was hoped that the higher sensitivity of the ion trap, coupled with the selectivity of D M E as a reagent gas, would provide a good combination for the differentiation of P A H isomers in environmental samples. 100 3.3.2 DME Chemical Ionization using an Ion Trap Mass Spectrometer In addition to their high sensitivity, (detection limits < lpg) ion traps allow for selective mass storage. Individual ions, or groups of ions, produced during electron ionization of the reagent vapour can be mass selected and reacted with the neutral molecules entering the ion trap, enabling selective ion/molecule reactions to be performed. In practice, the ability to mass select individual ions on the Varian Saturn 3D ion trap instrument was limited to m/z values greater than 60, hence the m/z 45 ions of D M E could not be selectively isolated. Nevertheless, the abundance of m/z 45 could be optimized and high mass reagent ions (e.g. m/z 91,93) could be removed prior to reaction with the analyte molecules. There are important experimental differences to be recognized when performing CI in a standard ion source versus an ion trap. In standard ion sources, such as those employed in a linear quadrupole system, ionization of the reagent gas and its subsequent reaction with analyte molecules proceed simultaneously because the electron beam is generated constantly. During CI in an ion trap, electrons are briefly pulsed into the ion trap to ionize the reagent gas. High mass analyte ions that may form during this time (e.g. from the ionization of analyte molecules that are entering the ion trap from the GC column) can be ejected from the trap. Once the timed (1 to 2500 ps) ionization period has ended, the ion/molecule reaction period (1-128 ms) begins and it is during this period of time that the reagent ions react with the neutral analyte molecules eluting from the GC column. This approach enables the investigation of analyte interactions with individual reagent ions rather than with a mixture of different ions, as occurs in a standard CI ion sources. Ionization of D M E vapour in the ion trap generated m/z 45, m/z 47, m/z 61 and minor amounts of m/z 91-93 ions. Optimization of the abundance of m/z 45 resulted in the 101 following relative intensities: m/z 45 (100 %), m/z 47 (95 %) and m/z 61 (20 %). The higher mass m/z 91-93 ions were removed by using a reagent ion ejection amplitude of 9.0 volts. The TIC resulting from the analysis of a standard 16 P A H sample mixture by GC-DME-CI/MS in the Saturn 3D ion trap is illustrated in Figure 3.4. The mass spectral data extracted from this chromatogram are listed in Table 3.7. 102 Intensity 1BBX Scan n u m b e r 6 8 8 Time ( m i n ) 5 . 9 9 LL 5b \ 6b r 1288 1 1 . 9 9 1888 1 8 . 8 8 9d L 10a+11 10b 2 4 8 8 2 3 . 9 9 3 8 8 8 2 9 . 9 9 Figure 3.4 GC-DME-CI/MS total ion chromatogram from the analysis of a standard 16 PAH mixture by ion trap MS. 103 Table 3.7 Mass spectral data extracted from the chromatogram in Figure 3.4 % Relative Intensity (± 3 %) Retention PAH (M n u) M+ [MH]+ [M+13]+ [M+45]+ Time (min) 6.2 1 (128) 4 37 100 0 10.6 3a(152) 5 100 58 80 11.2 2(154) 5 60 100 0 12.6 4(166) 5 68 100 0 15.2 5b (178) 4 52 100 0 15.4 5a (178) 8 50 100 2 18.6 6b (202) 3 50 100 0 19.2 6a (202) 6 48 100 0 22.7 8b (228) 6 52 100 2 22.8 8c(228) 4 55 100 0 25.6 9a (252) 3 48 100 0 25.7 9b (252) 5 48 100 0 26.4 9d (252) 25 62 100 3 30.2 10a (276) 10 50 100 0 30.2 11 (278) 10 50 100 0 31.2 10b (276) 15 55 100 0 (a) Other ions: m/z 167 (73 % RI.) The data in Table 3.7 reveal that the adduct formation behavior observed in an ion trap is significantly different from that observed in a standard ion source (Table 3.4). With the exception of acenaphthylene (3), formation of [M + 45]+ is significantly diminished in an ion trap. For example, anthracene (5a) displays significant [M + 45]+ adduct formation in a standard ion source (40-67 % RI, Table 3.4) while producing only barely detectable levels (« 2 % RI) in the ion trap. Pyrene (6a), also yields significant amounts of the [M + 45]+ adduct 104 in a standard ion source but not in an ion trap. Hence the mass spectra of 6a and 6b are practically identical and no isomer differentiation is possible. As mentioned previously, the [M + 13]+ ions are produced as a result of loss of C H 3 O H from the [M + 45]+ adduct. Hence, observation of the m/z 223 ions for both 5a and 5b indicates that the [M + 45 ] + adduct must have formed but was not stable enough to be observed over the time frame of this analysis. During analysis in an ion trap, ions are stored for longer periods of time prior to mass analysis compared to a standard ion source. Reaction times in an ion trap are of the order of 1-128 ms, significantly longer than the 10 u.s interaction times achievable in a conventional CI source 9 1 . Consequently, these differences in source residence times may account for differences in the [M + 45]+ adduct data. However, reducing the reaction times from 128 ms to 1 ms during D M E CI analysis of 5a did not result in an observable increase of the [M + 45]+ adduct, indicating that, if time is a factor, this adduct undergoes C H 3 O H elimination in less than 1 ms. The other important difference between an ion trap and a standard CI ion source is the reagent gas pressure. In an ion trap, reagent gas pressures are of the order of 10"5 mbar, while in an ion source, pressures range from 0.1-1 mbar during CI analysis. The higher pressures in a standard ion source can result in greater collisional stabilization of any adducts that may form during CI. Changes in adduct stabilization have been attributed to differences in the formation of adducts in ion traps versus standard ion sources65. Hence, the lower pressure of D M E in an ion trap may prevent the [M + 45]+ adducts from being stabilized. Consequently, D M E is not a good candidate for PAH isomer differentiation when using an ion trap. Although the reactivity of D M E as a chemical ionization reagent for PAHs has been investigated previously by others 6 8 ' 6 9 , no investigation of the reactivity of individual D M E generated reagent ions with PAHs has been conducted. Thus, upon availability of a 105 Quadrupole Ion Storage Mass Spectrometer (QISMS) ion trap instrument the current study was expanded to investigate this particular aspect. 3.3.3 Investigating the Ion/Molecule Reactions Between DME Ions and PAHs 5a and 5b With a QISMS ion trap isolation of the three D M E ions of interest (m/z 45, 47, 61) was achieved by using the three scan functions described in Figure 3.2(b). These scan functions describe the storage of ions after the ionization period. Consequently, no new ions are formed during these scan functions, except through ion/molecule reactions of the stored ions with PAHs eluting from the GC column. i) Isolation of m/z 45 ions In scan function A, ions above m/z 30 are stored initially (0-1 ms). At 1 ms, the r.f. and d.c. voltages are increased to 428 V and 67 V respectively, resulting in the storage of ions in the range of 31 to 57 m/z. During this period (2-5 ms), a frequency of 131.3 kHz (amplitude = 5 V) is also applied to the end-caps. This frequency causes m/z 47 (C 2 H 7 0 + ) ions to be ejected. The m/z 45 ions are then reacted with the analyte during a 120 ms reaction period (6-126 ms) during which all ions of m/z > 30 are stored in the trap. This is followed by a mass scan (m/z 100-350) to collect the spectrum of the resulting reaction products. Application of this scan function to D M E resulted in almost a complete isolation of the m/z 45 ions. Only a small number of ions at m/z 61 (« 5 % RI) remained. ii) Isolation of m/z 47 ions Scan function B was developed for the isolation of m/z 47 ions. The first step in this function is similar to the one in function A, except that the 131.3 kHz frequency was not 106 applied, since the m/z 47 ions had to be stored. Isolation of m/z 47 from m/z 45 was achieved by increasing the r.f. to 468 V, causing all ions with m/z < 46 to be ejected (6-9 ms). During the reaction period (10-130 ms), the r.f. was returned to 303 V allowing all ions with m/z > 30 to be stored. When using this scan function approximately 83 % of the m/z 45 ions were removed (~ 17 % RI of m/z 45 remained). iii) Isolation of m/z 61 ions Scan function C was developed for generating and reacting m/z 61 ions. Attempts at isolating m/z 61 ions by using a scan function (not shown) during which only the m/z 52 - 95 ions were isolated (r.f. = 726 V, d.c. = 324 V) failed to isolate any m/z 61 ions. However, it was observed that by increasing the D M E flow rate into the ion trap and at the same time storing m/z 45 ions over a period of 120 ms, the m/z 61 ions could be generated in approximately equal proportions to the m/z 45 ions. Scan function C generated m/z 61 ions by first isolating m/z 45 ions as in scan function A (0-5 ms), and then allowing the m/z 45 ions to be stored for 120 ms (6-126 ms, m/z 35-76 stored) to generate m/z 61 ions. During this time, only ions with a m/z value between 35 and 76 remained in the trap, thus removing any high mass ions that may have formed from ion/molecule reactions between analyte molecules and m/z 45 ions. The m/z 45 ions were then removed during the 126-246 ms period, during which only m/z > 50 ions were stored. Hence, during this period of time only the m/z 61 ions were allowed to react. iv) Reaction with DME reagent ions The results obtained for the reaction of phenanthrene 5b and anthracene 5a, with the ions isolated in functions A, B and C, are summarized in Table 3.8. 107 Table 3.8 Mass spectral data (% RI) for the ion/molecule reaction products between each of the three mass-selected ions (m/z 45, 47, 61) and phenanthrene (5b) and anthracene (5a) % RI values for major reaction products Phenanthrene (5b) m/z Anthracene (5a) m/z Ions (m/z) 178 179 191 223 178 179 191 223 C H 2 O C H 3 ( 4 5 ) n.d. 1 100 n.d. n.d. 3 100 4 ( C H 3 ) 2 O H + (47) 9 100 18 n.d. 9 100 18 n.d. ( C H 3 ) 3 0 + (61) n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. m/z 178 = M*~, m/z 179 = [M+l]+, m/z 191 = [M+13]+, m/z 223 = [M+45] n.d. not detectable The data for m/z 45 show predominantly [M + 13]+ ions (m/z 191) corresponding to loss of C H 3 O H from [M + 45]+. A small amount of [M + 45]+ ions (m/z 223) was observed only for anthracene. The data in Table 3.8 also demonstrate that no charge exchange products (m/z 178) are observed with m/z 45, and that only trace amounts of M H + ions (m/z 179) are formed. Overall, these results indicate that the m/z 45 ion reacts almost exclusively by adduct formation followed by rapid elimination of CH 3 OH. Reaction of 5b and 5a with the m/z 47 ions isolated via function B yielded mainly MPT ions (m/z 179). The 18 % relative intensity of the [M + 13]+ ions (m/z 191) in these mass spectra corresponds to the relative intensity of the m/z 45 ions that could not be completely removed (17 % RI) upon application of scan function B. Consequently, these m/z 191 ions are most likely formed from the reaction between the PAHs and the remaining m/z 45 ions. A small amount of m/z 178 was observed indicating that some charge transfer is taking place, however the dominant abundance of m/z 179 (MH +) ions indicates that the m/z 47 ion reacts predominantly by proton transfer with both P A H isomers 5b and 5a. 108 Reaction of 5b and 5a employing scan function C (m/z 61 ions) did not yield any molecular or adduct ions, indicating that the m/z 61 ions are not reacting with either of these two compounds. The recombination energy (RE) and proton affinity (PA) available for the three ions (m/z 45, 47, 61) and 5b and 5a are shown in Table 3.9. Comparison of the R E values for m/z 45 ion and the two PAHs may explain why the m/z 45 ion (CH 2 OCH 3 + ) could not ionize PAHs 5b and 5a by charge exchange. The R E of C H 2 O C H 3 + is lower than the IE of either 5b or 5a, therefore charge exchange with C H 2 O C H 3 + is not an energetically favorable process (i.e. AH > 0) for the formation of m/z 178 ions. For the reaction shown below, if M = 5a, then AH = IE(M) - RE(CH 2 OCH 3 + ) = 7.45 eV - 6.94 eV = 0.51 eV. C H 2 O C H 3 + + M - » C H 2 O C H 3 + M + Table 3.9 Ionization energy (IE) and proton affinity (PA) values for the major reaction species investigated Ion or Molecule M r (u) REa(eV) PA^kJmol1) C H 3 O C H 3 + 46 10.025 804 C H 2 O C H 3 + 45 6.94 n/a (CHa^OH" 47 < 7.45 b n/a (CH 3) 30+ 61 < 7.45 b n/a 5b 178 7.86 832 5a 178 7.45 865 (a) R E values estimated using IE values for the corresponding neutral species from Lias 1 3 . (b) Estimated from the current experimental results based on the lack of charge exchange products with 5a. The lack of any charge exchange products with m/z 47 ((CH 3 ) 2 OH + ) indicates that the RE for this species is probably less that 7.45 eV, the lowest IE value for the two PAHs in Table 3.9. 109 The formation of MET for either 5a or 5b during reaction with (CHs^OtT (m/z 47) can be justified by their higher PA when compared to the PA of C H 3 O C H 3 . Since PA(M) > PA(CH 3 OCH 3 ) , the following reaction will be exothermic: (CH 3 ) 2 OH + + M - » C H 3 O C H 3 + M t T , AH = PA(CH 3 OCH 3 ) - PA(M) The lack of any Ivf, MET/ or adduct ions upon reaction of 5a or 5b with m/z 61 ((CH 3) 30 +) suggests that this ion does not undergo either charge exchange or proton transfer, and that it does not form any stable adduct ions or elimination products. The results from the above experiments have more clearly elucidated the reaction chemistry between D M E ions and PAHs. The m/z 45 ions have been shown to react predominantly by adduct formation, followed by C H 3 O H elimination. The m/z 47 ions appear to react predominantly by proton transfer to form MET ions while the m/z 61 ions do not appear to contribute to any observable reaction products. 110 3.4 S U M M A R Y The chemical ionization of PAHs, performed by using D M E as a reagent gas, enabled the differentiation of several PAH isomers, in accordance with previous results in the literature. Some degree of PAH isomer differentiation was achieved when analyzing a contaminated environmental sample, but these results were limited by the instrument's sensitivity. Application of this CI technique on an ion trap mass spectrometer revealed that the [ M 4- 4 5 ] + adducts, required for isomer differentiation, could not be generated in significant amounts. An increase in the yield of the [ M + 4 5 ] + adduct ions was not achieved by mass-isolating C H 3 O C H 2+ ions (m/z 45) and allowing them to react with PAH molecules such as 5a. Nonetheless, these ion/molecule experiments were able to more clearly elucidate the chemistry taking place during D M E / C I . I l l CHAPTER 4 DIFFERENTIATION OF PAH ISOMERS BY USING CHEMICAL IONIZATION AND ION/MOLECULE REACTIONS WITH HALOGENATED HYDROCARBONS IN AN ION TRAP MASS SPECTROMETER 112 4.1 I N T R O D U C T I O N Environmental concerns about PAHs are predominantly associated with the mutagenic and carcinogenic properties of many of these compounds ' . As mentioned in Chapter 1, their toxicological properties are often isomer specific. For instance, benzo[a]pyrene (9d), is a powerful carcinogen while its structural isomer, benzo[e]pyrene (9c), is non-carcinogenic. This trend is also observed between positional isomers. For example, 7,12-dimethylbenz[a]anthracene is more carcinogenic than 1,12-dimethylbenz[a]anthracene. Consequently, it is important that analyses for PAHs in environmental samples differentiate between isomers as well as their analogues. However, this is particularly difficult in the case of positional isomers where a greater number of species are possible. Although complete isomer differentiation may not be achievable, grouping PAHs into structural isomer classes would at least provide a greater degree of information than is currently available from GC/MS with electron ionization (EI). The EI mass spectra of most PAH isomers are virtually identical 6 6 , thus isomer differentiation can be achieved only by matching GC retention times with those from authentic standards. Consequently, PAHs for which standards are not available, or that are co-eluting with other compounds, cannot be identified. In order to achieve isomer differentiation it is necessary to obtain characteristic mass spectra of the individual isomers. The previous two chapters have described attempts at isomer differentiation by using negative CI as well as positive CI techniques. Although negative CI (Chapter 2) provided some useful results, it did not enable a good systematic investigation of PAH isomers in environmental samples because of the low sensitivity of the mass spectrometer used. Similarly, the use of dimethyl ether (DME) during positive CI (Chapter 3) provided 113 promising results for isomer differentiation with a standard ion source, but was once again limited by the instrument's sensitivity. Unfortunately, application of the D M E CI method on an ion trap mass spectrometer, a more sensitive instrument, did not prove useful for isomer differentiation. The majority of the research published to date on the differentiation of PAHs by 69 73 76 84 85 chemical ionization ' " ' ' has used standard ion source techniques, rather than ion trap methods. However, as mentioned in Chapter 3, the difference in source pressure between these two techniques can result in significant differences in the formation of ion/molecule products. Consequently, chemical ionization techniques that allow isomer differentiation using standard ion source mass spectrometers may not necessarily yield the same results on ion trap instruments. However, since ion traps can be more sensitive and versitile for GC/MS analysis than other types of mass spectrometers, there is a need to develop novel CI techniques for P A H isomer differentiation on these instruments. Consequently, a search was undertaken for chemical ionization reagents that would enable formation of stable adducts with PAHs inside an ion trap, and thereby provide a means to produce isomer specific mass spectra. Benzene has been shown to undergo electrophilic aromatic substitutions in the gas phase similar to those occurring in solution 9 6 . Formation of gas phase adducts between dichloromethane and naphthalene has been demonstrated using a high pressure ion source 97 + . Investigations on the behavior of simple aromatic species with C F 3 ions using an ion-98 102 beam apparatus have also been carried out " . These studies suggested that some aromatic compounds can react with halogenated ions in the gas phase to form stable adduct ions. Although some halogenated hydrocarbons have been used as CI reagent gases, no 114 investigation has been carried out on the use of halocarbon reagents for CI and isomer differentiation of PAHs. Because the studies outlined previously were not performed using ion trap mass spectrometers, two major questions come to mind: (1) will stable adducts be formed with halogenated gaseous ions in the low pressure environment of an ion trap and (2 ) , if adducts are formed, would different PAH isomers produce different adducts and/or different proportions of common adducts? These questions will be explored by using a series of halogenated hydrocarbons (CH 2C1 2, CHC1 3 , CDCI3, CCI4, CH 3 CHC1 2 , C H 2 F 2 , and C H 3 C H F 2 ) for chemical ionization and ion/molecule reactions with PAH isomers. For the current work, the term chemical ionization will imply reactions of an analyte with all of the ions generated from a reagent gas in the ion trap, while ion/molecule reactions will imply reactions with individually mass-selected ions or groups of ions. The major P A H isomer groups of interest in the current work are identified in the chromatogram illustrated in Figure 4.1. This chromatogram represents the results obtained from the GC/EI-MS analysis of a sediment extract collected from an effluent lagoon of an aluminum smelter. 115 5b 6b 5a 5b Phenanthrene 5a Anthracene 6b Fluoranthene 6a Pyrene 8b Benz[a]anthracene 8c/d Chrysene / Triphenylene 9a/b Benzofluoranthenes 9e Benzo[e] pyrene 9d Benzo[a]pyrene 8c 8d 9a/b 14 1 6 18 20 22 24 26 time (min) 28 30 32 34 Figure 4.1 GC/EI-MS total ion chromatogram of an aluminum smelter effluent lagoon extract analyzed using an ion trap mass spectrometer. Major PAHs present in the sample are identified. 116 4.2 EXPERIMENTAL 4.2.1 Standards and Reagents Nine of the PAH isomers of interest (phenanthrene 5b, anthracene 5a, fluoranthene 6b, pyrene 6a, benz[a]anthracene 8b, chrysene 8c, benzo[b]fluoranthene 9a, benzo[k] fluoranthene 9b, benzo[a]pyrene 9d) were present in a standard solution (2000 u.g mL"1 each) purchased from Supelco (Oakville, Ontario, Canada) which was diluted with toluene to give 0.05 to 50 u.g mL"1 solutions. Individual standards for 9,10-dihydroanthracene dh-5a, 9,10-dihydrophenanthrene dh-5b, triphenylene 8d, tetracene 8e, benzo[e]pyrene 9c, and perylene 9e were obtained from Aldrich (Milwaukee, WI, USA). The perdeuterated PAHs d10-anthracene and d10-acenaphthene were obtained from CIL (Woburn, M A , USA). Dichloromethane, chloroform (HPLC grade, Fisher Scientific, Nepean, ON, Canada), chloroform-d3 (CIL, Woburn, M A , USA), carbon tetrachloride (BDH, Toronto, Ont. Canada), and 1,1-dichloroethane (Fisher Scientific, Fair Lawn, New Jersey, USA) were used to generate the corresponding reagent vapours by volatilization into the ion trap via a Negretti needle valve (Fareham, Hants. GB). Difluoromethane (PCR Research Chem. Inc., Gainsville, Florida, USA) and 1,1-difluoroethane (Aldrich, Milwaukee, WI, USA) were introduced into the ion trap from lecture bottles equipped with regulators. 4.2.2 Apparatus All experiments were performed using a Saturn 4D GC/MS/MS system (Varian, Walnut Creek, C A , USA) equipped with a Wave-Board for the generation of user-defined wave functions for application to the ion-trap electrodes. Gas chromatography was 117 performed using a DB5 capillary column (30 m x 0.25 mm I.D., 0.25 pm film thickness) from J & W (Folsom, C A , USA), that was inserted directly into the ion trap through a transfer line maintained at 280-290°C. 4.2.3 Procedures Gas Chromatography: Introduction of 1 pL samples was performed via splitless injection at 290°C. The GC column temperature program consisted of a 90°C initial temperature for 0.1 min followed by a 6°C min 1 ramp to 280°C, holding at 280°C for 1 min and then ramping to 300°C at 20°C min"1 and holding for 20 min. Helium was used as a carrier gas at linear velocity of 32 cm sec"1, corresponding to a flow rate of approximately 1 mL min"1 into the ion trap. Ion trap: The ion trap used in this investigation could only be operated in positive ion mode. Thus, all results reported here refer to positive ions only. The trap temperature was maintained between 220°C and 255°C. No significant changes in the gas phase reactions were observed over this temperature range. The reagent vapours introduced into the trap were ionized with an electron beam of 70 eV. The multiplier voltage was optimized to give a gain of 105. Vacuum conditions were monitored by observing the intensity of the background spectrum (air, water), because the instrument was not equipped with a vacuum gauge. Figure 4.2 displays the scan functions employed for the chemical ionization experiments discussed in this chapter. 118 RF Voltage 4 D ARC prescan Analytical scan I • Time A: ARC reagent gas ionization period (100 ps) B: ARC Reaction period (128 ms) C: Ejection of low mass ions (i.e. CI reagent ions) D: Ejection of all ions for total ion count measurement (E) F: Ionization time as determined by ARC pre-scan (10 to 2500 ps) G: Reaction period as determined by ARC pre-scan (1 to 128 ms) H: Ejection of reagent gas ions prior to mass scan I: Mass scan of ions present in trap Figure 4.2 Standard scan function, with an ARC pre-scan, used for chemical ionization in the Saturn ion trap mass spectrometer. (Figure adapted from Varian Saturn 3D GC/MS/MS instruction manual). The primary variable parameters for chemical ionization in the Saturn 4D ion trap are indicated in Figure 4.2. When the Automatic Reaction Control (ARC) scan function is employed, the reagent gas ionization time (F), and the reaction time (G) between sample molecules and reagent gas ions is determined automatically according to the TIC measurements (E) obtained during the ARC pre-scan. The ARC function calculates optimal ionization time and reaction time values in order to achieve a pre-specified ARC target value of 10 to 24999 (arbitrary units). This controls the number of ions generated, preventing space charge effects from occurring. When required, experiments were also 119 performed in manual mode, with the ARC function disabled. During manual operation, the ionization time (IT) is preset and the reaction time (RT) is determined according to the formula 4.1: MRT and MIT correspond to the maximum reaction time and maximum ionization time values specified in the mass spectrometer method. If more than one type of reagent ion is formed during the ionization of a specific reagent gas, the values of IT and RT may affect the abundance ratios of these ions. Consequently, IT and RT may have to be optimized manually to obtain a desired distribution of reagent ions. The abundance ratios of reagent ions are also influenced by the amount of reagent vapour introduced into the trap. Other important parameters are the ionization storage level (ISL) and reaction storage level (RSL). The ISL corresponds to the r.f. voltage set during the ionization period (A or F). In the Saturn instrument this value is specified, in m/z units, as the lowest m/z value that is stored during the ionization period. The RSL is a parameter similar to the ISL, but corresponds to the lowest m/z value that is stored during the reaction period (B or G). Collision-induced dissociation (CID) experiments on mass-selected ions were performed by applying a MS/MS scan function (Figure 4.3) after chemical ionization. A one m/z window was used to select the precursor ions. Fragmentation was achieved via CID with the carrier gas using a non-resonant excitation waveform at 55 - 60 volts for a duration of 10 ms. A more detailed description of non-resonant excitation will be provided in Chapter 7. [4.1] 120 Unless otherwise specified, experiments were conducted using all the reagent gas ions generated. When required, isolation of specific mass reagent ions, for m/z > 60, was achieved by incorporating a MS/MS scan function, such as that illustrated in Figure 4.3, into the CI scan function. When these experiments were performed a CID voltage value of 0 volts was used to avoid fragmentation of the mass-selected reagent ions. Removal of high mass ions prior to the reaction period was achieved by using the Selected Ejection Chemical Ionization (SECI) function. Mass ejection with SECI results from the application of a low frequency square wave to the end caps of the trap. The list of the major experimental parameters employed during CI and MS/MS experiments are summarized in Tables 4.1 and 4.2 respectively. For the MS/MS CID experiments, the excitation storage level corresponds to the r.f. storage level in m/z units when the CID waveform is applied during the CID period (T c i d). Table 4.1 List of major experimental parameters employed during the chemical ionization experiments under ARC conditions. * Experimental Parameter Value (s) Used ARC Target 10000 ARC pre-scan ionization time 100 (J.s Maximum ionization time (MIT) 2500 u.s Maximum reaction time (MRT) 128 ms Ionization storage level (ISL) 25 m/z Reaction storage level (RSL) 25 m/z Emission current 10 u.A Ejection amplitude 0 - 6 V Unless otherwise specified 121 RF Voltage • Time For isolation and CID experiments on mass "m" A: Ejection of m/z < m B: Ejection of m/z > m C: CID of m during time period (Tdd). A supplementary r.f. voltage is applied to the end caps. D: Mass scan of resulting CID products Figure 4.3 Schematic of standard MS/MS Scan function employed for mass-isolation and CID experiments Mass-selected ions were allowed to react with analyte molecules during time period C with the r.f supplementary voltage set to a value of 0 volts. Table 4.2 Major experimental parameters employed during the MS/MS CID experiments Experimental Parameter Value (s) Used Isolation window 1 u Waveform type Non-resonant CID r.f. voltage 55 - 60 V CID time (Tcid) 10 ms Excitation storage level 48.0 m/z 122 4.2.4 Data Analysis Unless otherwise indicated, the mass spectral data presented in this chapter consists of background subtracted average values obtained across the chromatographic peak of an eluting compound. The data for most of the mass spectra and chromatograms shown were replotted with GraFit software (1994 Version 3.0, London, U.K.) by using the original data extracted from the Saturn data files (*.ms files). Conversion of the binary data in the Saturn *.ms files to ASCII text files for use in Grafit was performed by using MS2TXT sofware (Unitech, Cincinnati, USA). 123 4.3 RESULTS AND DISCUSSION 4.3.1 Formation of Reagent Ions The major reagent ions produced during ionization of the halocarbons are listed in Table 4.3. Characteristic isotopic ratios were observed for the chlorinated reagents. Table 4.3 Major reagent ions produced during electron ionization of halocarbons Reagent Major Reagent Ions R+ (m/z) (> refers to relative intensity) CH2CI2 CH2Cl+(49/51) > CHCl2+(83/85/87) CHC13 CHCl2+(83/85/87) CDC13 CDCl2+(84/86/88) CCL CC1 3 + (117/119/121/123) CH3CHC12 CH3CHCl+(63/65) > CHCl2+(83/85/87) CH 2F 2 CH 3CHF 2 C H 2 F + (33) C H 3 C H F + (47) > C H 3 C F 2 + (65) > CFTF2 + (51) When ionization of a halocarbon vapour produces more than one reagent ion, (e.g. C H 3 C H F 2 -> CH 3 CHF + & C H 3 C F 2 + & C H F 2 + ) , the relative amounts of each type of ion will depend on reaction time (RT), ionization time (IT) and the gas pressure, as demonstrated for C H 3 C H F 2 in Figure 4.4(a) to (c). 124 a) 0 20 40 60 80 100 120 140 Reaction Time (msec) b) 0 200 400 600 800 1000 Ionization Time (usee) Figure 4.4 Effects of (a) reaction time (msec) (b) ionization time (jusec) and (c) pressure (TIC) on m/z 47, 51 and 65 ions from 1,1-difluoroethane. 125 c) 0 2e+006 4e+006 TIC Figure 4.4 (continued). For example, when the reaction time (RT) was increased from 1 to 128 msec the relative inensity of the m/z 47 C H 3 C H F + ions increased, while the relative intensity of m/z 51 and 65 ions decreased (Figure 4.4(a)). Increasing the ionization time for C H 3 C H F 2 from 10 to 1000 usee gave a linear increase in the ion count for m/z 47, 51 and 65 (Figure 4.4(b)). When the amount of 1,1-difluoroethane gas introduced into the trap was increased, as measured by an increase in the total ion count (TIC), the concentration of the m/z 47 ion increased while the m/z 51 and 65 ions decreased (Figure 4.4(c)). The predominance of m/z 47 at higher concentrations and reaction times suggests that the loss of F" from the neutral C H 3 C H F 2 molecule is probably an important mechanism for the formation of C H 3 C H F + . 126 In order to employ all three of the major C H 3 C H F 2 reagent ions (m/z 47, m/z 65, m/z 51), during chemical ionization, the flow rate was adjusted to produce a relative intensity ratio of 8 : 2 : 1 (m/z 47: m/z 65: m/z 51). Unless otherwise specified, chemical ionization experiments were performed using all the reagent ions listed in Table 4.3 simultaneously. 4.3.2 Chemical Ionization Reactions Chemical ionization of the PAHs resulted in formation of adduct ions for most of the PAHs analyzed. Several intense adducts were observed for some PAHs, sometimes even approaching base peak intensity. The adducts observed can be grouped into two main classes, namely: direct addition products of the type [M -I- R ] + and elimination products of the type [M + R - HX] + , where M represents the PAH molecule, R the reagent gas ion and X = Cl or F. The halocarbons investigated can be classified into two groups, halomethanes and haloethanes. Although both groups of halocarbons produced adducts, some significant differences in their chemical ionization behavior were observed. 4.3.2.1 Halomethane reagents The major products resulting from the chemical ionization of PAHs with halomethanes can be expressed as [M + R - H X ] + (M = PAH molecule, R = reagent ion, X = Cl or F). Formation of this type of product, after reaction with dichloromethane ions, is illustrated in Figure 4.5 for the structural isomer pair anthracene (5a) and phenanthrene (5b). The two isomers are able two yield distinctly different mass spectra. Ions are observed at m/z 178 (M + ) and m/z 191 ([M + CH 2C1 -HC1]+) for both PAHs. 127 However, m/z 191 is more prominent in the mass spectra of 5b. In addition, for isomer 5b, ions are also observed at m/z 225/227 which corresponds to [M + CHC1 2 - HC1] + . The mass spectrum for phenanthrene 5b also displays small amounts of m/z 239/241, possibly due to a reaction product of the m/z 97/99/101 ions (probably C 2 H 3 C 1 2 + ions which were observed in small quantities during ionization of CH 2C1 2) followed by loss of HC1. In general, no significant amounts of the [M + R ] + adduct were observed in the mass spectra of any of the halomethanes investigated here. The difference in the amounts of adducts formed by the four isomer pairs investigated during chemical ionization with halomethanes, are illustrated in Figures 4.6(a) to (d). The numerical values for the data shown in these figures are summarized in Table 4.4. a) in c cu CD > J O CD CC 1 0 0 -8 0 -6 0 4 0 H 2 0 ' 0 1 7 8 c CD CD _> JS CD CC 191 1 1 I 1 I i I i I i I i I ' i 1 6 0 1 8 0 2 0 0 2 2 0 2 4 0 2 6 0 2 8 0 3 0 0 m/z 1 0 0 -8 0 H 6 0 4 0 -2 0 -1 7 8 191 2 2 5 T—rn—I |' i—r~"i—T—i—i—i—I i I 1 6 0 1 8 0 2 0 0 2 2 0 2 4 0 2 6 0 2 8 0 3 0 0 m/z Figure 4.5 Dichloromethane chemical ionization mass spectra of (a) anthracene 5a and (h) phenanthrene 5b. (m/z 178 = M+, m/z 179 = 13C of M+, m/z 191 = [M + CH2Cl -HCl]+, m/z 225 = [M + CHCl2 - HCIJ).+ 128 or 4 0 H 20 H 0 C H 2 C l 2 I o X I o r x o + o X I r o X o + C D C L O X I o Q O + C C L O X I C o o + C H 2 F 2 X o + Figure 4.6(a) Relative intensity of adduct ions formed by the PAH isomer pair 5a/5b during chemical ionization with halomethanes. (Ivf = 100 %). on C H 2 C I 2 4 0 H 2 0 H I o X o o X o + o X 1 o o X o + C D C L X o X I o o Q o + C C L O X I c o o + C H 2 F 2 X I Li_ r x o + Figure 4.6(b) Relative intensity of adduct ions formed by the PAH isomer pair 6a/6b during chemical ionization with halomethanes. (]vf = 100 %). 129 CH 2CI 2 CDCL I J o o o X X 1 X 1 o CM o CM O X X Q o o O + + + 2 2 2 CCL o X I CO o o + C H 2 F 2 x o + • 8b • 8c • 8d B 8e Figure 4.6(c) Relative abundance of adduct ions formed by the PAH isomer group 8b to 8d during chemical ionization with halomethanes. (M* = 100%). 40 30 DC 20 10 0 CH 2 CI 2 CDCL JLii o X 6" X o + o X CJ o X o + o X I CJ O Q O + CCI 4 o X I CO o O + C H 2 F 2 JZL. X I LL CM X o + Figure 4.6(d) Relative abundance of adduct ions formed by the PAH isomer group 9a to 9e during chemical ionization with halomethanes. (Af1" = 100%). 130 This set of figures clearly illustrates the differences in ion intensities between many of the isomeric compounds investigated. These differences enable the isomers to be differentiated from one another. It is important to note that although the reaction of PAHs with each of the four halomethane reagents yields different amounts of adducts, the trends in reactivity are similar. For example, all the halomethanes produced higher yields of [M + R - H X ] + upon reaction with 5b than upon reaction with 5a. Similar trends in reactivity can also be noted with the other isomers groups investigated. However, there are some important differences. For example, the greater the number of chlorines in the reagent gas, the lower the yield of adducts formed. Variations can also be noted in the relative intensity of [M + CHC1 2 - HC1] + and [M + CDC1 2 - HC1] + adduct ions (R = C H C 1 2 + and R = C D C 1 2 + data in Table 4.4). Since the C H C 1 2 + and C D C 1 2 + ions are chemically similar they should be expected to yield equal amounts of adduct ions. This was indeed the case when the C H C 1 2 + ions were mass-selected from the C H 2 C 1 + during CH 2 C1 2 chemical ionization. In an experiment, in which ion/molecule reaction reactions were performed between mass-selected m/z 83 (CHC1 2 +) ions and molecules of 5b, the relative intensity of [M + CHC1 2 - HC1] + adduct ions was found to be to 36 %. This value is comparable to the 30 % relative intensity of the [M + CDC1 2 -HC1] + adduct ion obtained during CDC1 3 chemical ionization of 5b (Table 4.4). The lower value for the [M + CHC1 2 - HC1] + obtained during regular CH 2 C1 2 chemical ionization (R = C H C 1 2 + data, Table 4.4) can be attributed to the C H C 1 2 + ions having to compete with the C H 2 C 1 + ions. 131 Table 4.4 Major chemical ionization products, m/z (% RI), formed between halomethanes ions (R) and selected PAH isomers (M) CH2C12 CDC13 CC14 CH 2 F 2 R=CH2C1+ R=CHC12+ R=CDC12+ R=CC13+ R=CH 2r (M+13) (M+47) (M+48) (M+81) (M+13) PAH M + a [M+R-HC1]+ [M+R-HCI]+ [M+R-HC1]+ [M+R-HC1]+ [M+R-HF]+ 5a 178 (100) 191(13) 225 (0) 226 (1) 259 (0) 191 (4) 5b 178 (100) 191(44) 225 (14) 226 (30) 259(14) 191 (28) 6a 202 (100) 215(13) 249 (1) 250 (3) 283 (0) 215 (5) 6b 202 (100) 215(47) 249 (14) 250(31) 283 (18) 215 (28) 8b 228 (100) 241(12) 275 (1) 276 (2) 309 (0) 241 (5) 8c 228 (100) 241(25) 275 (3) 276 (20) 309 (5) 241(12) 8d 228(100) 241 (24) 275 (5) 276 (12) 8e 228 (100) 241 (8) 275 (0) 276 (0) 309 (0) 9a 252 (100) 265 (34) 299 (8) 300 (24) 333 (10) 265 (21) 9b 252 (100) 265 (18) 299 (4) 300 (9) 333 (0) 265 (9) 9c 252 (100) 265 (12) 299 (0) 300 (2) 333 (0) 9d 252 (100) 265 (5) 299 (0) 300 (0) 333 (0) 265 (1) 9e 252 (100) 265 (3) 299 (0) 300 (1) 333 (0) Notes. General: 3 7C1 isotope values observed but not indicated in table. Small amounts of [M + R ] + adducts are produced with some PAHs, but the data are not shown since they represents values < 1 % RI. % RI values were obtained by averaging mass spectral data over the entire chromatographic elution peak. The uncertainty in these averaged % RI values is approximately 10 % of the %RI value (i.e. 30 %+3 %). The error bars in Figures 4.6 and 4.8 are based upon this error level, (a) M + is the base peak for all the substances shown in this table. 100-•5 80 H c CO £ 60-cu ,> 1 40H CC 20 H 0 178 225 205 I 2 4 3 b) c cu 100-80" 60 > u 40H cu cc —I—r—|—i—r~i—| i | i | — i — | — i — | 160 1 80 200 220 240 260 280 300 m/z 20 H 132 179 205 223 243 l i I i i i r~~i i > I 1 i 1 l 160 1 80 200 220 240 260 280 300 m/z Figure 4.7 1,1-Difluoroethane chemical ionization mass spectra of (a) anthracene 5a and (b) phenanthrene 5b. (m/z 178 = M+, m/z 179 = MH+ and 1 3 C of M+, m/z 205 = [M + CH3CHF - HFJ+, m/z 223 = [M + CH3CF2 - HF]+, m/z 225 = [M + CH3CHFJ+, m/z 243 = [M + CH3CF2J+). 4.3.2.2 Haloethane reagents The major adduct ions formed between haloethane reagent ions (R) and PAH molecules (M) were [M + R - H X ] + (X = Cl or F) and [M + R ] + . The most significant difference observed between the chemical ionization behavior of haloethanes and halomethanes was that haloethanes formed significant amounts of the [M + R ] + adduct ions, while halomethanes did not. For example, as shown in Figure 4.7, 1,1-difluoroethane and anthracene 5a form [M + R ] + adducts with both C H 3 C H F + and C H 3 C F 2 + , yielding ions at m/z 225 and 243 respectively. In the mass spectrum of the 5a isomer, the [M + C H 3 C H F - H F ] + elimination product (m/z 205) was observed, but no ions for the [M + C H 3 C F 2 - H F ] + elimination product (m/z 223) were detected. However, phenanthrene 5b yielded both elimination products [M + C H 3 C H 2 F - HF] + , (m/z 205) and [M + C H 3 C H F 2 - H F ] + , (m/z 223), as well as small amounts of the [M + 133 C H F 2 - H F ] + (m/z 209) elimination product. On the other hand, the mass spectrum of 5b shows the presence of only one addition product: [M + C H 3 C F 2 ] + (m/z 243). Furthermore, comparison of the two mass spectra reveals that the 5a isomer yields a significantly lower intensity of the m/z 205 elimination product. It is clear from the comparison of the mass spectra of 5a and 5b shown in Figure 4.7 that these two isomers can be distinguished from each other when using 1,1-difluoroethane chemical ionization. In fact, 1,1-difluoroethane is a better chemical ionization reagent that any of the halomethanes tested since it generates a greater abundance of distinguishing mass spectral features (e.g. compare Figure 4.5 and Figure 4.7). A summary of chemical ionization products formed by all four groups of PAH structural isomers when analyzed with 1,1-difluoroethane and 1,1-dichloroethane is illustrated in Figures 4.8 (a) to (d). These data are also summarized in Tables 4.5 and 4.6. 134 rr C H 3 C H C I 2 C H , C H F . ZZ + o zz I o zz CJ zz cj + zz CJ ZZ o + zz + + + EL + EL ZZ EL I LL I 1 ZZ CJ LL om CO zz CJ zz" CJ zz o + CJ + + 2 CJ CT ZZ CJ + Figure 4.8(a) Relative intensity of adduct ions formed by the PAH isomer pair 5a and 5b during chemical ionization with CH3CHCI2 and CH3CHF2. CH,CHF. DC + CJ zz 2 zz u " o + + s X + + + •HF] + EL •HF] EL X LL X 1 CN X CJ LL O M CO X O CO x" CJ X 0 + 0 + + CJ Figure 4.8(b) Relative intensity of adduct ions formed by the PAH isomer pair 6a and 6b during chemical ionization with CH3CHCI2 and CH3CHF2. 135 CH 3CHCI 9 x + tkaCL, CJ X I CJ X CJ o X CJ + CJ X CJ CO X CJ + 111 CH 3 CHF 2 I + x + F-HF] + EL7 HF] F-HF] X I CM X CJ LL CJ CO CO X O CO X CJ X o + o + + 2 2 i o CO X o + • 8b • 8c • 8d • 8e Figure 4.8(c) Relative abundance of adduct ions formed by the PAH isomer group 8b to 8e during chemical ionization with CH3CHCI2 and CH3CHF2. CH,CHCL 100 80 1 S 60 c# 40 20 1 0 J I I I 11 1 U J 11 CH„CHF, Ll 1 • 9a • 9b • 9c i n 9d m 9e X + CJ X I 0 X 0 CO X CJ + CJ X o CO X CJ + X + X CJ CO X CJ + X 0 CO X CJ + LL. X CJ CO X CJ + 0 CO X CJ + Figure 4.8(d) Relative abundance of adduct ions formed by the PAH isomer group 9a to 9e during chemical ionization with CH3CHCI2 and CH3CHF2. 136 Table 4.5 Major products, m/z (% RI), formed between 1,1-dichloroethane ions (R = CH3CHC1+) and selected PAH isomers (M) (M) (M+l) (M+27) (M+63) PAH M + [M+H]+ [M+R-HC1]+ [M+R]+ 5a 178(100) 179 (32) 205 (11) 241 (3) 5b 178 (55) 179 (68) 205 (100) 241 (2) 6a 202 (100) 203 (33) 229 (20) 265 (4) 6b 202 (64) 203 (62) 229 (100) 265 (1) 8b 228(100) 229 (45) 255 (15) 291 (7) 8c 228 (100) 229 (56) 255 (58) 291 (2) 8d 228(100) 229 (77) 255 (74) 291 (1) 8e 228(100) 229 (34) 255 (7) 291 (3) 9a 252 (68) 253 (95) 279(100) 315 (10) 9b 252 (100) 253 (40) 279 (35) 315 (2) 9c 252 (22) 253(100) 279 (42) 315 (35) 9d 252 (100) 253 (30) 279 (2) 315 (3) 9e 252 (100) 253 (27) 279 (0) 315 (1) Note: C l isotope values observed but not given in table. 137 Table 4.6 Major products, m/z (% RI), formed between 1,1-difluoroethane ions (R) and selected PAH isomers (M) R = CH3CHF* R = CH 3 CF 2 + (M) (M+l) (M+27) (M+47) (M+45) (M+65) PAH M+ [M+H]+ [M+R-HF]+ [M+R]+ [M+R-HF]+ [M+R]+ 5a 178 (100) 179 (42) 205 (10) 225 (18) 223 (0) 243 (13) 5b 178 (66) 179 (100) 205 (88) 225 (0) 223 (16) 243 (7) 6a 202 (100) 203 (46) 229 (20) 249 (20) 247 (0) 267 (20) 6b 202 (55) 203 (100) 229 (86) 249 (9) 247 (18) 267 (22) 8b 228(100) 229 (52) 255 (13) 275 (30) 273 (1) 293 (12) 8c 228 (98) 229(100) 255 (50) 275 (11) 273 (6) 293 (16) 8d 228 (42) 229 (100) 255 (16) 275 (10) 273 (2) 293 (13) 8e 228(100) 229 (33) 255 (15) 275 (20) 273 (8) 293 (7) 9a 252 (2) 253(100) 279 (60) 299 (25) 297(11) 317 (30) 9b 252(100) 253 (64) 279 (25) 299 (28) 297 (3) 317(22) 9c 252 (2) 253(100) 279 (20) 299 (86) 297 (4) 317 (32) 9d 252(100) 253 (28) 279 (1) 299 (19) 297 (0) 317 (7) 9e 252(100) 253 (26) 279 (1) 299 (6) 297 (0) 317 (1) Note: C l isotope values observed but not given in table. 138 The relative intensity distribution graphs shown in Figures 4.8 (a) to (d) clearly illustrate the isomer differentiation potential of the two haloethane reagent gases investigated. Chemical ionization with either 1,1-dichloroethane or 1,1-difluoroethane results in the formation of a greater abundance of ions than chemical ionization with the halomethane reagents discussed previously. This greater abundance of ions enables a greater choice of comparisons to be made for isomer differentiation purposes. Unlike the halomethanes, the haloethane reagents also produce protonated molecules ( M H + ) , whose intensities vary among the different isomers. Although the molecular ion (M + ) is the base peak in the mass spectra of many of the isomers analyzed by haloethane chemical ionization, occasionally adduct ions, or M H + ions, can also reach base peak intensity, as illustrated for the 5b and 6b data in Figures 4.8 (a) and (b). Although many of the product ions listed in Figures 4.8 (a) to (d) show striking differences in relative intensity between isomers, this is not the case for all ions. For example, the [M + C H 3 C F 2 ] + adduct ions do not offer significantly different relative intensity values to enable clear isomer differentiation. It is often necessary to compare several ions to obtain isomer differentiation, particularly for isomer groups containing more than two isomers, such as those shown in Figures 4.8 (c) and (d). For example, in the C H 3 C H F 2 data shown in Figure 4.8(c), the relative intensities of the [M + C H 3 C H F ] + ions of the 8c and 8d isomers are of a similar magnitude, while the intensity of the [M + C H 3 C H F - H F ] + ions for these same isomers are significantly different. Comparison of the CH 3 CHC1 2 and C H 3 C H F 2 chemical ionization results reveals that for similar adducts, (i.e. [M + R - H X ] + formed by the R = C H 3 C H C 1 + or 139 C H 3 C H F + ions), both halocarbons exhibit similar trends in the relative intensities of product ions formed with different isomers. For example, both halogens yield a greater amount of [M + R - H X ] + ions for 5b than for 5a (Figure 4.8(a)). To further highlight the isomer differentiation capabilities of the haloethane gases, the [M + R - H X ] + / [M + R ] + ion ratios are tabulated in Table 4.7. Comparison of the data for the two gases reveals that in most cases the value of the [M + R - H X ] + / [M + R ] + ion ratios are lower when C H 3 C H F 2 is used as a reagent gas, indicating that chemical ionization with C H 3 C H F 2 yields a greater relative intensity of [M+R] + ions. It is also interesting to note that the lowest [M + R - H X ] + / [M + R ] + ratios are for R = C H 3 C F 2 + , indicating the highest relative intensity of the [M + R ] + ions. These results suggest that the [M + R ] + ions are more stable when R contains F atoms rather than Cl , and that two F atoms (R = C H 3 C F 2 + ) yield a more stable [M + R ] + ion than one F atom (R = C H 3 C H F + ) . These results are consistent with Cl" being a better leaving group than F" 1 8 . Table 4.7 Haloethane chemical ionization [M + R - HX]+/[M + R]+ % RI ratios CH3CHCI2 CH 3CHF 2 R = CH3CHC1+ R = CHaCHF" R = CH 3 CF 2 + PAH (M+27)/(M+63) (M+27)/(M+47) (M+45)/(M+65) 5a 4 0.6 0.04 5b 50 200 2 6a 5 1 0.02 6b 100 10 0.82 8b 2 0.43 0.08 8c 29 4.5 0.4 8d 74 1.6 0.2 8e 2 0.75 1 9a 10 2.4 0.37 9b 18 0.89 0.1 9c 1 0.23 0.12 9d 0.7 0.05 0.07 9e 0.5 0.2 0.5 [ M + R - H X ] + ions = (M+27), (M+45) [ M + R ] + ions = (M+63), (M+47), (M+65) 141 4.3.3 Mechanism of Adduct Formation The reaction between PAH molecules and halocarbon ions probably proceeds through an electrophilic aromatic substitution mechanism ( E A R ) . In previous chemical 97 ionization studies of benzene and naphthalene that used CH 2 C1 2 as a reagent gas , the formation of [M + 13]+ ions was attributed to [M + CH 2C1 - HC1] + and the presence of [M + 47]+ was attributed to the elimination of H 2 from [M + CH2C1] + . However, the [M + 47]+ ion could also have been formed through loss of HC1 from [M + CHC1 2 ] + . Results obtained from the reaction of mass-selected C H C 1 2 + and C H 2 C 1 + ions with phenanthrene (5b) support the latter mechanism involving elimination of HC1 from [M + CHC1 2 ] + . When C H C 1 2 + was mass-selected and reacted with 5b only the [M + 13] = [M + CHC1 2 - HC1] + product was observed, and when CH 2C1 was reacted only [M + CH 2C1 - HC1] + was formed. The lack of formation of any [M + 47]+ ions from the reaction with C H 2 C 1 + demonstrates that H 2 is not eliminated from [M + CH 2 C1] + under ion trap conditions. Thus, formation of [M + 47]+ ions during CH 2 C1 2 chemical ionization in an ion trap must result solely from the loss of HC1 from [M + CHC12] + rather than the loss of H 2 from [M + CH 2 C1] + . A mechanism for the reaction of 5b with C H C 1 2 + is proposed in Figure 4.9. It is important to note that substitution at the 2 position on 5b illustrated in this mechanism is not meant to imply a favored site for adduct formation. In fact, as it will be discussed later there does not appear to be a favored site of attack. Although a o-bonded complex 1 0 3 is shown in Figure 4.9, a 7t-complex is also a possible intermediate during electrophilic aromatic substitution 1 0 4 ' . 142 ci i M + C H C 1 2 + [M + CHC1 2 ] + [ M + C H C 1 2 - HC1] + Figure 4.9 Proposed mechanism for the reaction between 5b and CHClf' • Further confirmation of the HC1 elimination mechanism was obtained through the use of C D C I 3 to generate C D C 1 2 + ions. Reaction of phenanthrene 5b with C D C 1 2 + ions resulted in the production of ions at m/z 226 / 228 corresponding to [M + CDC1 2 - HC1] + (M = 5b, 178 u). The loss of HC1 rather than DCI is a consequence of elimination of hydrogen from the P A H aromatic ring rather than from the reactant ion. In a complementary approach d10-acenaphthene (d10-2) was reacted with C H C 1 2 + leading to the formation of ions at m/z 210/212, corresponding to [M + CHC1 2 - DC1] + (M = d 1 0-2, M r = 164 u). This elimination product corresponds to a loss of D from the aromatic ring of d 1 0-2, as is expected in an E ^ mechanism (d10-2 was used since d1 0-5b was not available and this compound undergoes adduct formation readily). In order for the [M + R ] + adducts to be observed they must be stabilized either through collisions with neutral species (collisional stabilization) or by emission of a photon (radiative association). The former is usually the dominant process at ordinary chemical ionization pressures (0.1-1 Torr), but in the lower pressure environment of an ion trap («10~ 5 Torr), the latter mechanism could also be an important stabilization process, especially for larger molecules with 50 or more internal degrees of freedom 1 0 5 . The data in Table 1.1(a) and Tables 4.4 to 4.6, indicate that in most cases for PAHs with similar 143 ionization energies, those of higher molecular mass tend to form larger amounts of [M + R ] + adducts. For example, 9c (M r = 252, IE = 7.41 eV) forms approximately four times the amount of [M + C H 3 C H F ] + adduct as 6a (M r = 202, IE = 7.41 eV). This difference may be a result of the larger 9c molecule undergoing a greater degree of stabilization through radiative association. 4.3.4 Investigating the Mechanism for Isomer Differentiation The difference in reactivity between the structural isomers investigated was initially believed to be associated with the presence or absence of reactive sites, such as the 9,10 20 double bond in 5b. This is the site with the lowest localization energy (Lc) . To test this hypothesis, 9,10-dihydrophenanthrene (dh-5b) was reacted with C D C 1 2 + . If the reactivity of the C D C 1 2 + ion is selective for the 9,10 double bond then no adduct should be formed at this position on dh-5b, hence resulting in a decrease in adduct formation. However, as shown in Table 4.8, (dh-5b) formed similar amounts of [M + CDC1 2 - HC1] + as 5b, indicating that the 9,10 double bond is not essential for the formation of this ion. 9,10-Dihydroanthracene (dh-5a) was also analyzed to determine the effects of saturating the reactive positions on 5a. Surprisingly, the reaction of C D C 1 2 + with dh-5a resulted in large amounts of [M + CDC1 2 - HC1] + , exceeding those of both phenanthrene 5b and dh-5b. Furthermore, the mass spectrum of dh-5a also displayed large amounts of a secondary HC1 elimination product, [M + CDC1 2 - 2HC1] +, at m/z 192. 144 Table 4.8 CDC13 CI mass spectral data (m/z and (% RI)) for anthracene (5a), phenanthrene (5b) and their 9,10-dihydro analogues. R = CDC12+ Compound M (M+l) (M+48) (M+84) other ions M + [M+H]+ [M+R-HC1]+ [M+R] + 5a 178 (100) 179 (15) 226 (1) 262 (0) dh-5a 180 (30) 181 (5) 228 (100) 264 (0) 179 (32), 192 (44) 5b 178 (100) 179 (15) 226 (32) 262 (0) dh-5b 180 (100) 181 (14) 228 (29) 264 (0.5) The results from these latter two experiments seem to indicate that adduct formation is not selective towards the more reactive positions in an aromatic system. Consequently, other factors must play a role in the adduct formation reactions observed. A comparison between the mass spectral data (Tables 4.4 to 4.6) and IE data (Table 1.1(a)) revealed that the isomers forming the largest amounts of [M + R - H X ] + ions have IE values greater than 7.5 eV while those forming the smallest amounts of these ions have IE values below 7.5 eV. This difference is clearly exemplified by the two isomer pairs 5a / 5b and 6a / 6b. PAHs 5b and 6b, both have IE values greater than 7.5 eV, and yield large amounts of [M + R - H X ] + ions. PAHs 5a and 6a have IE values below 7.5 eV and only yield small amounts of [M + R - H X ] + ions. When the relative amounts of [M + R - H X ] + products from chloromethane ions for the four PAHs isomer groups (5a to 9e) are plotted as a function of their ionization energy, a positive trend with increasing PAH ionization energy is observed (Figure 4.10). A positive correlation between [M + R - H X ] + formation and PAH ionization energy was also obtained with haloethane reagent ions (Figure 4.11). 145 Based on the correlation between [M + R - H X ] + formation and ionization energy 9,10-dihydrophenanthrene (dh-5b) would be expected to possess a similar IE to 5b as it yielded a similar intensity of [M + R - H X ] + ions to 5b (Table 4.8). However, the actual IE value for dh-5b is 7.55 eV 1 3 , significantly less than the 7.86 eV value of 5b, suggesting that factors, other than ionization energy, may be influencing adduct formation. Similarly, 9,10-dihydroanthracene (dh-5a) would be expected to possess a higher IE value than 5b, since it yielded a greater abundance of [M + R - H X ] + ions than 5b (Table 4.8). Unfortunately, no experimental IE values are available for dh-5a to permit an evaluation of this latter hypothesis. 146 c o o T3 CD N ro E O c « 0.8 0.6 0.4 0.2 0 o R = cn.2cr • R = C H C I 2+ • R = C C I 3+ J H5H t=CN H-QJ-CH 7.2 " i — i r~ 7.4 7.6 IE (eV) H H 7.8 8 Figure 4.10 Effect of PAH ionization energy (IE) on the formation of [M+R-HX]+ elimination products with chloromethane reagent ions. +-* c o o T3 CD N O C 0.8 0.6 E 0.4 H 0.2H o R = CH 3 CHCI + R = C H 3 C H F+ • R = C H 3 C F 2 h#=l ^ 0 # - ^ ^ i—Q—i I — | — i — | — i — | i | i r 7.2 7.4 7.6 7.8 8 IP (eV) Figure 4.11 Effect of PAH ionization energy (IE) on the formation of [M+R-HX]+ elimination products with haloethane reagent ions. The normalized count data represents counts/TIC values obtained at the apex of a chromatographic peak, that have been normalized with respect to the highest counts/TIC value. Errors estimates in Figures 4.10 to 4.12: IE ± 0.04 eV and normalized counts: ± 10 % of count/TIC values (based upon average relative deviation of counts/TIC for mass spectral data obtained across chromatographic peak, at a 90 % confidence level). 147 The fact that PAHs with a low ionization energies yield lower amounts of the [M + R - H X ] + ions suggests that these compounds may undergo charge exchange more readily than adduct formation. The ionization energies for the formation of the reactant ions studied are listed in Table 4.9(a) 1 3 . When the IE for the formation of a reagent ion (R +) is greater than the ionization energy for the formation of a PAH ion (M + ) an exothermic (AH<0) charge exchange reaction can occur between M and R + as shown in the scheme below: * M -> M + + e" AHj = IE(M) e + R + -> R A H 2 = -IE(R) M + R + -> M + + R A H 3 = AHj + A H 2 - IE(M) - IE(R) A H 3 < 0 when IE(R) > IE(M) (* At temperatures > 0 K, AH « IE(M) 13) The IE values for the reagents R investigated are generally greater than the IE of the PAHs (IE(M) < 8 eV), allowing for charge exchange to occur with A H 3 values of approximately -0.1 to-2 eV. Analysis of the [M + R ] + ion data (Table 4.6) does not reveal a clear trend with PAH IE values. For the isomer pairs 5a/5b and 6a/6b the % RI values of the [M + C H 3 C H F ] + ions are higher for the lower IE isomer (i.e. 5a and 6a), an opposite trend to the [M + C H 3 C H F - H F ] + ions. However, the data for the other two isomer groups does not enable a clear correlation to be established between [M + R ] + adduct ion intensity and PAH IE values. 148 Table 4.9(a) Ionization energies (IE) for the formation of R+ ions and proton affinities (PA) for halocarbon ions Ion (R+) IE(R) (eV)a R-H PA of (R-H) (kcal mol1)b CH 2 Cf 8.6 CHC1 208 CHC12+ 8.1 CC12 193 c c V 7.8 N/A N/A CPEX 9.05 CHF 193 CHsCHF* 7.93 CH 3CF 175 CH 3 CF 2 + 7.92 CH 2CF 2 174 CH3CHC1+ n/a CH3CC1 172 Notes: n/a not available, N / A not applicable, (a) From IP values of neutral R from n . (b) Calculated by using enthalpies of formation from Lias 1 3 . Sample calculation: PA(CC1 2) = - A H for CC1 2 + H4" ->• CHC1 2 + . Therefore, PA(CC1 2) = - [ A H ^ C H C l ^ - A H ^ C C ^ 4 ) - A H ^ C H C l ^ ] = - (212-39-365.7) kcal mol"1 = -193 kcal mol"1. Protonation of PAHs can also take place if their proton affinity (PA) (Table 4.9(b)) exceeds that of the deprotonated reagent ions [R-H] (Table 4.9(a)). None of the halomethanes tested here yielded any significant amounts of PAH protonation (the M + l 13 signals were not significantly larger than the expected C contribution), while the haloethanes resulted in formation of M H + ions (Tables 4.5 - 4.6). The formation of M H + ions with haloethanes can be explained by the lower PA of the (R-H) species (172 - 175 kcal mol"1) in comparison to the PA of PAHs (196 - 207 kcal mol"1). Halomethane [R-H] species have PA values approximating those of PAHs (193 -208 kcal mol"1), explaining the lack of M H + formation with these reagents. Interestingly, some of the PAHs with the highest PA values, but low IE values, (Table 4.9(b)), such as 5a, 6a and 8e, displayed the smallest amounts of M H + formation. Consequently, PAHs with low IE values seem to 149 undergo charge exchange to M + preferentially over M H + formation as illustrated in Figure 4.12. Table 4.9(b) Ionization energies (IE) and proton affinities (PA) for PAHs PAH IE(eV)a PA (kcal mol *) a 5a 7.45 207 5b 7.86 199 6a 7.41 206 6b 7.95 199 8b 7.43 198 b 8c 7.59 202 8d 7.84 198.5 8e 6.97 218 9a 7.9 c 196 b 9b 7.4 c 198 b 9c 7.41 197 b 9d 7.12 201.5 b 9e 6.90 197 (a) Experimental values obtained from Lias 1 3 . (b) Theoretical values (PM3 calculations) obtained from trouk-Pointet1 (c) Calculated using H M O values as discussed in Table 1.1(a). 150 03 c D O o T3 CD N " r o E o c 1 H w 0.8 0.6 0.4H 0.2 OH KD-I H | H r McJjjjl HCH 9c hCH ^ i i i " i i r 6.8 7 7.2 7.4 7.6 IE (eV) h-CH i—i—r 7.8 8 o M * + (1,1-DFE) • M * + (1,1-DCE) ~ 0 .8H c D O o CD N " r o o c 0.6 0.4 H 0.2H 0 9c o M H + (1,1-DFE) • M H + (1,1-DCE) — i | i | i | i | i | i | i 6.8 7 7.2 7.4 7.6 7.8 8 IE (eV) Figure 4.12 Effect of PAH ionization energy on the formation of Af and MIT in the presence of haloethane reagent ions (1,1-DFE = 1,1-difluoroethane; 1,1-DCE = 1,1-dichloroethane). , 151 4.3.5 M S / M S Fragmentation of Adducts The lack of observable adducts for some PAHs during chemical ionization with a particular reagent gas may be a result of the adduct decomposing prior to detection, or it may be a result of the PAHs undergoing charge exchange preferentially to adduct formation and thus yielding mainly molecular ions. To investigate possible differences in the nature of adducts formed by 5a and 5b, a series of collision-induced dissociation (CID) experiments was performed on mass-selected adduct ions following chemical ionization. As illustrated in Table 4.4, [M + R - H X ] + ions represent the major adducts produced during chemical ionization of 5a or 5b with halomethanes. However, small amounts (less than 1 % RI) of [M + R ] + ions are were observed in some experiments, enabling them to be isolated for MS/MS experiments. The data summarized in Table 4.10 summarize the % RI values from the CID mass spectra of [M + CDC1 2 ] + adduct ions. When mass-selected [M + CDC1 2 ] + adduct ions (m/z 262) of 5a and 5b were fragmented via CID, [M + CDC1 2 - HC1] + daughter ions (m/z 226) were observed for both adducts, but only the 5a adduct yielded M + ' ions (m/z 178) (Table 4.10). These results indicate that the M = 5a, [M + CDC1 2 ] + adduct is less stable than the M = 5b adduct suggesting that the two adducts may be different in nature. CID of the mass-selected [M + CDC1 2 - HC1] + elimination product of 5b (m/z 226) produced predominantly [M + CDC1 2 - 2HC1]+ ions (m/z 190) from the loss of a second HC1, and did not yield any m/z 178 ions (data not shown). No CID experiment could be performed on the [M + CDC1 2 - HC1] + elimination product of 5a because this ion could not be detected. 152 Using 1,1-difluoroethane, CID experiments were performed on the mass-selected [M + R ] + and [M + R - H X ] + adduct ions of 5a and 5b, for R= C H F 2 + , C H 3 C H F + and C H 3 C F 2 + (Tables 4.11 and 4.12). The CID experiments on the [M + R ] + adducts indicate that the M = 5b adduct is the most stable species, since it produced a lower abundance of fragment ions, particularly when R = C H F 2 + . Similarly, the CID of the [M + R - H X ] + adducts also revealed that 5b formed the most stable species, based upon its lower yield of fragment ions. 153 Table 4.10 Mass spectral data from the CID (60 Volts, 10 ms, Non-resonant) of mass-selected [M+CDCI2]+ adducts of phenanthrene (5b) and anthracene (5a) R+ M [M+CDCl2]+ [M+ CDC12-HC1]+ M+ (m/z 262) (m/z 226) (m/z 178) CDC12+ 5b 100 30 0 5a 100 82 76 Table 4.11 Mass spectral data from the CTD (55 Volts, 10 ms, Non-resonant) of mass-selected [M+R]+ adducts of phenanthrene (5b) and anthracene (5a) R+ M [M+R]+ [M+R-HF]+ [M+R-2HF]+ M+ CHF 2 + 5b 100 0 0 5 5a 100 0 0 45 CHaCHF^ 5b n/q n/q N/A n/q 5a 4 100 N/A 20 CH 3 CF 2 + 5b 100 40 25 32 5a 28 100 32 68 n/q not quantifiable, N / A Not Applicable (only one F atom available) Table 4.12 Mass spectral data from the CTD (55 Volts, 10 ms, Non-resonant) of mass-selected [M+R-HF]+ adducts of phenanthrene (5b) and anthracene (5a) R+ M [M+R-HF]+ [M+R-2HF]+ M+ CHF 2 + 5b 100 6 4 5a 100 0 62 CHaCHr 5b 100 N/A 2 5a 100 N/A 17 CH 3 CF 2 + 5b 100 60 1 5a 100 50 30 N / A Not Applicable (only one F atom available) 154 4.3.6 Effects of Chemical Ionization Reaction Time on Adduct Formation When CDCI3 was used as a reagent gas, no significant changes in the relative amount of adduct formation (i.e. as a percent of TIC) were noted over a series of CI reaction times ranging from 10 to 128 ms. However, during chemical ionization with 1,1-D F E , increasing the CI reaction time resulted in higher amounts of the adducts formed with R = C H 3 C H F + , (m/z 205 and 225) as illustrated in Figures 4.13 (a) and (b). This result was entirely consistent with the previous observations (Figure 4.4(a)) that increasing reaction time yields larger amounts of C H 3 C H F + ions (m/z 47). The data for the adducts formed with R = C H 3 C F 2 + (m/z 243 and 223) shows a slight initial increase in intensity followed by a decrease at longer reaction times. The decrease observed at longer reaction times is most likely a result of the decrease in the abundance of R = C H 3 C F 2 + (m/z 65) (see Figure 4.4(a)). These results indicate.that reaction time can be an important factor in the relative amounts of adducts formed during chemical ionization when more than one reagent ion is present. a) 0 20 40 60 80 100 120 140 Reaction Time (ms) b) O H 0 20 40 60 80 100 120 140 Reaction Time (ms) Figure 4.13 Effect of CI reaction time on formation of (a) [M+R]+, (b) [M+R-HFJ 225 and 205, R = CH3CHF*; m/z 243 and 223 R = CH3CF2+). 156 4.3.7 Effects of Analyte Concentration on Adduct Formation It has been found (Table 4.13) that under ARC conditions changes in concentration over two or three orders of magnitude do not significantly affect the relative intensities of the major adduct ions discussed in this work. For example, the mass spectral data for 50 to 57000 pg each of phenanthrene (5b) and anthracene (5a), are summarized in Table 4.13. It was also found that by using mathematical correlations of the different m/z values it is possible to determine the relative amounts of each member of an isomer group. This is illustrated in Figure 4.14 showing the relationship between the integrated area of the m/z 205 ([M + C H 3 C H F - HF] + ) chromatogram and the changing mole fraction of 5b in a series of standard mixtures of 5b and 5a. Phenanthrene yields a larger abundance of m/z 205 ions which results in an increase in the total m/z 205 count as its mole fraction increases. Observation of the effects of changes in concentration on the mass spectrum of a compound can also be performed by comparing the mass spectral data obtained at the shoulder and apex of a chromatographic elution profile. Such a comparison is depicted in Figure 4.15 for the chromatographic elution profiles of 5 ng each of 5a and 5b. No significant changes can be noted between the mass spectra obtained at the apex and shoulder of the chromatographic peaks. 157 Table 4.13 Effect of sample concentration on the percent relative intensities for major ions observed during chemical ionization with 1,1-difluoroethane. Average mass spectral % RI values obtained across chromatographic peaks. Amount Phenanthrene (5b) Anthracene (5a) (pg) 178 179 205 223 225 243 178 179 205 223 225 243 57000 95 100 65 16 <1 <1 100 32 8 1 7 8 29000 80 100 65 16 <1 3 100 35 9 2 8 8 10000 55 100 72 13 <1 3 100 40 12 2 9 8 5000 48 100 75 12 <1 3 100 40 13 <1 10 7 572 35 100 72 8 5 2 100 40 16 1 8 4 286 35 100 75 7 5 1 100 38 13 1 8 4 100 38 100 80 5 <1 2 100 38 13 <1 8 5 50 40 100 79 8 <1 <1 100 40 16 <1 10 4 80000 - i 0 0.2 0.4 0.6 0.8 Mole Fraction Phenanthrene Figure 4.14 Effect of mole fraction composition of phenanthrene / anthracene on m/z 205 [M+ CH3CHF-HFf+ during difluoroethane chemical ionization (correlation coefficient (r) = 0.98 ). 158 Intensity l 1 1 ' 1 1 1 1 1 1 i T | i i • i . 1 0 5 0 1 0 6 Q 1 0 7 0 1 0 8 8 1 8 9 8 1 5 . 9 9 1 6 . 1 5 1 6 . 3 1 1 6 . 4 7 1 6 . 6 4 Figure 4.15 1,1-Difluoroethane CI mass spectral data at shoulder and apex of chromatographic elution profile for 5 ng each of phenanthrene (5b) and anthracene (5a). 159 4.3.8 Linear Dynamic Range of CI versus EI Analysis In order to verify the applicability of this technique for quantitative purposes, a series of analyses were performed by injecting amounts ranging from 5 to 57000 pg under both 1,1-DFE CI and standard EI conditions. The results from these analyses are shown in Figures 4.16 and 4.17 for the isomer pair pyrene (6a) and fluoranthene (6b). For the EI data, only the molecular ion integration values are shown since this is the major ion formed. As these experiments demonstrate, the CI values produce a linear response (log-log linear) and are comparable to the EI data. However, slightly lower detection limits are possible by using EI (injections of 5 pg are not detectable by CI, but are detectable by EI). It is important to note that these CI experiments were performed under A R C conditions. When non-ARC conditions were employed some deviation from linearity was observed at higher concentrations as shown in Figure 4.18. Consequently, the use of A R C is of particular importance at higher concentrations as it prevents excessive ionization which leads to space charge effects and non linearity. However, at low concentrations, A R C becomes less important since ionization times are usually near their maximum value (2500 Us). 160 a) 1 e + 0 0 6 - | 1 e + 0 0 5 - | 1 0 0 0 0 ^ 1 0 0 0 ^ 1 0 0 - | 1 0 -6a O O • o • • o • • • • i A e • A A A • I o i n—i i 11111| 1—i i 111ii| i—i i 11111| 1—rm 10 100 1000 Amount (pg) 10000 o m/z 202 • m/z 203 • m/z 229 • m/z 249 A m/z 247 A m/z 267 1e+005 10000 -J 1 0 0 0 ^ 1 0 0 ^ 10 6b • o • o 0 a o o a • 5 A O m/z 202 8 A A A A • m/z 203 A A • • • m/z 229 A • • • m/z 249 A m/z 247 A m/z 267 10 "i i i 11111| i i i 111ii| i i i 11111| i r 100 1000 10000 Amount (pg) ~n Figure 4.16 Chromatographic integration areas of selected ions of pyrene (6a) and fluoranthene (6b) analyzed by 1,1-difluoroethane chemical ionization under ARC conditions. 161 co CD 1e+007-| 1e+006 l 1e+005-= 10000^ 1000-100 O 6a m/z 202 * 6b m/z 202 ® ® 9 9 —i i 111iii|—i i 11 ni i |— i i 111ni|—i i 1111ii| i i 111 1 10 100 1000 10000 Amount (pg) Figure 4.17 Chromatographic integration areas of molecular ions of pyrene (6a) and fluoranthene (6b) analyzed by electron ionization. 1e+005d co CD 10000H 1000-J • 6a TIC o 6b TIC 8 8' • oo o o —i—i i 11111| 1—i i 111ii| 1—i i 111111 i n 10 100 1000 10000 Amount (pg) Figure 4.18 TIC integration areas of pyrene (6a) and fluoranthene (6b) analyzed by 1,1-difluoroethane chemical ionization under non-ARC conditions. 162 4.4 S U M M A R Y Reactions of halogenated hydrocarbon ions (R +) with PAHs in an ion trap result in the formation of adduct ions [M + R ] + and elimination products [M + R - H X ] + . Differences in the intensity values of these ions can be used to differentiate between structural isomers. Use of haloethane gases results in the greatest number of adduct and elimination ions and thus offers the best isomer differentiation capability. Of the haloethanes tested, 1,1-difluoroethane was found to yield the most characteristic mass spectra. Although the exact mechanism for the isomer differentiation process is not completely understood it appears to be at least partially correlated to the IE of a PAH, rather than to the reactivity of specific molecular regions. The ability to generate isomer characteristic mass spectra, by using a sensitive instrument such as an ion trap, offers the potential for identification and isomer differentiation of PAHs in complex mixtures. A detailed account of the application of this technique to the analysis of environmental samples containing PAHs will be reported in Chapter 5 by using 1,1-difluoroethane as a reagent gas. 163 CHAPTER 5 DIFFERENTIATION OF PAH ISOMERS IN A CONTAMINATED SEDIMENT EXTRACT BY USING 1,1-DIFLUOROETHANE CHEMICAL IONIZATION ION TRAP MASS SPECTROMETRY 164 5.1 INTRODUCTION The use of halocarbons, C n H m X p (n=l- 2, m= 1- 4, p=2 - 3, X = Cl, F), as chemical ionization (CI) reagents in an ion trap was shown in Chapter 4 to result in formation of reactive ions R + that yielded adducts of the type [M + R] + and / or [M + R - H X ] + with PAH molecules (M) 1 0 7 . Structural isomers of PAHs were distinguished based on differences in their adduct-forming behavior. When using haloethanes both of the above adduct types could be generated, but when halomethanes were employed only [M + R - H X ] + adducts were observed in significant amounts. The formation of two different types of adduct ions during CI with haloethanes makes these reagents more useful for isomer differentiation purposes. Among the haloethane reagents examined, 1,1-difluoroethane (1,1-DFE) provided the largest number of adduct-forming reactive ions R (R = C H 3 C H F + , C H 3 C F 2 + and CHF 2 + ) . Furthermore, 1,1-DFE yielded the greatest relative intensity of [M + R] + adducts. Consequently, this compound was selected as a reagent for developing a method for the differentiation of P A H isomers in environmental samples. Analysis of environmental samples was also investigated with the other reagents described in Chapter 4. However, only the results with 1,1-difluoroethane will be reported here since the CI methodology for the other reagents was similar. Isomeric differentiation is often not possible when using standard GC/MS methodologies with electron ionization (EI). Co-eluting isomers, or isomers for which standards are not readily available, cannot be identified and are generally grouped together according to their molecular masses 1 0 8 . For example, methylfluoranthenes, methylpyrenes and benzofluorenes all give rise to identical mass spectra (base peak = M*' = m/z 216). 165 The GC/MS method reported here using 1,1 -difluoroethane (1,1-DFE) as a chemical ionization reagent generates isomer-characteristic mass spectral signatures, thus enabling isomer differentiation. The formation of isomer characteristic ions or ion intensity ratios is particularly important when isomers are closely eluting or co-eluting. Software extraction of individual ion chromatograms from the total ion chromatogram (TIC), corresponding to the characteristic ions of individual isomers, enables the identification and differentiation of isomers in complex mixtures. However, when mass spectral differences between isomers are based only on variations in relative ion intensities, the analysis becomes more complicated. However, as outlined in Figure 4.14, the proportion of an individual isomer (e.g. 5b) in an isomer mixture (e.g. 5a/5b) can be determined from the intensity of a particular ion (e.g. m/z 205). The aim of the work summarized in this chapter is to investigate the feasibility of differentiating P A H isomers present in an environmental matrix by using 1,1-difluoroethane chemical ionization. 166 5.2 EXPERIMENTAL 5.2.1 Standards and Reagents Pure standards of 1-methylanthracene, 2-methylanthracene, 2-methylphenanthrene, 3,6-dimethylphenanfhrene, 9,10-dimethylanfhracene and 1 lH-benzo[b]fluorene were purchased from Aldrich (Milwaukee, WI, USA) while 1-methylphenanthrene, 2-methylfluoranthene, 2-,4-,7-methylbenz[a]anthracene, 3,9-6,8- and 7,12-dimethylbenz[a]anthracenes, 2-,3-,4-, and 5-methylbenzo[c]phenanthrenes, 1,12-dimethylbenzo[c]phenanthrene, 8-,9- and 10-methylbenzo[a]pyrene were purchased from AccuStandards (New Haven, CT, USA). All solvents used were HPLC grade (Fisher Scientific, Nepean, ON, Canada). The perdeuterated PAH d10-anthracene was obtained from CIL (Woburn, M A , USA). 1,1-Difluoroethane (Aldrich, Milwaukee, WI, USA ) was introduced into the ion source via the chemical ionization gas inlet from a lecture bottle equipped with a regulator. 5.2.2 Environmental Samples The contaminated sediment was obtained from an effluent settling pond at the Alcan aluminum smelter in Kitimat, British Columbia, Canada. The sediment was freeze dried and a 5 g sample was spiked on its surface with 4.5 pg of d10-anthracene and extracted for 15 minutes with 40 mL of dichloromethane / acetone (5:1) using an Accelerated Solvent Extraction System (Dionex, Sunnyvale, CA, USA) operated at 125°C and 2000 psi. A 1 mL aliquot, containing 10 mg of crude extract, was flushed through a column containing 5 g of silica gel. A 10 pl (« 100 pg of crude extract) portion of the extract was diluted to 1 mL with toluene and used for the GC/MS analysis. 167 5.2.3 Apparatus and Procedures All experiments were performed using a Saturn 4D GC/MS ion trap system (Varian, Walnut Creek, CA, USA) as described in Chapter 4, using Automated Reaction Control (ARC) conditions. 5.2.4 Data Analysis The mathematical transformations discussed in this work were performed on mass spectral files that were first converted into text files using a binary to ASCII converter (MS2TXT file converter, Unitech, Cincinnati, USA). ASCII text files were then imported into a Excel spreadsheet (Microsoft) for performing the required calculations and plotted by using GraFit sofware (1994 Version 3.0, London, U.K.). Prior to performing each mathematical transformation, the data set for each ion chromatogram in a particular retention time segment was normalized to the largest count value. 168 5.3 RESULTS AND DISCUSSION 5.3.1 PAH Homologue Series in Sediment Extract The total ion chromatogram (TIC) that was obtained from the 1,1-DFE chemical ionization analysis of the sediment extract is reproduced in Figure 5.1 (regions 1 to 4). These four chromatograms represent the four retention time windows (Rw) in which the four main groups of structural isomers (5a to 9e) elute. The major PAHs eluting in these four retention time (tR) slots are identified in Table 5.1, together with the relative intensity values of their adducts (M + 27 = [M + CH 3 CHF - HF] + , M + 47 - [M + CH 3 CHF] + , M + 45 = [M + C H 3 C F 2 - HF] + , M+65 = [M + CH 3 CF 2 ] + ) . From these data it appears that the m/z 47 (CH 3 CHF + ) ion gives rise to the highest relative intensity for most of the adducts, making it the most useful for isomer differentiation. Although the m/z 65 ion (CH 3 CF 2 + ) results in the formation of a lower adduct abundance with some of the PAHs, it can still be useful, especially in cases where isobaric interferences in the environmental matrix may obscure the m/z 47 adducts. Presence of m/z 65 is not detrimental to the analysis and therefore this ion was not removed from the reagent ion group. The homologue series described in each region (Table 5.1), as CO, C l and C2, refer to the carbon substitution pattern on the P A H (CO = unsubstituted parent compound, C l = monomethyl substitution, C2 = dimethyl or monoethyl substitution). With increasing substitution pattern (CO —> C2) the number of possible isomers will increase, leading to a larger number of co-eluting species and thus complicating the chromatographic data. The relative concentrations of PAHs decreases along the homologous series from CO to C2; this is expected since the PAHs in this sample were formed at high temperatures 1 0 9 Alkyl PAHs above C2 will not be discussed here since their concentrations were near or below the detection limits of the ion trap. 169 Region 1 Region 2 1740 2015 2315 2450 2693 2922 16 17 18 19 20 21 21 22 23 24 25 26 27 Region 3 Region 4 i 2922 3164 3393 3512 3721 26 27 28 29 30 31 31 32 33 34 Scan number Time (min) Figure 5.1 Total ion chromatogram of sediment extract obtained using 1,1-difluoroethane chemical ionization showing the four isomer regions of interest. Cl = monomethyl PAHs, C2 = dimethyl / ethyl PAHs. 170 Table 5.1 1,1-DFE CI mass spectral data for selected PAH isomers from the settling pond extract, t R refers to chromatogram shown in Figure 5.1 Rw PAH M r Scan tR % Relative Intensity a No. (min) M M+l M+27 M+47 M+45 M+65 1 CO 5b 178 1740 16.1 38 100 68 <1 10 3 CO 5a 178 1763 16.3 100 45 13 12 <1 6 Cl 5b 192 1993 18.2 53 100 47 3 10 6 Cl 5a 192 2015 18.4 100 42 7 12 <1 7 C2 5b 206 2169 19.7 70 100 50 4 7 4 C2 5a 206 2227 20.2 100 54 10 13 2 5 2 CO 6b 202 2315 20.9 42 100 83 4 11 7 CO 6a 202 2423 21.8 100 45 15 17 4 14 7a,7b, 216 2630 100 84 36 23 7 17 7c Cl 6b 216 2543 22.8 63 100 58 7 8 8 Cl 6a 216 2681 23.9 100 41 10 20 1 9 C2 6b 230 2780 24.8 26 100 46 7 6 6 C2 6a 230 2841 25.3 100 95 24 22 4 9 3 CO 8a 228 2910 25.9 95 100 50 4 10 10 CO 8b 228 3015 26.7 100 56 15 30 1 11 CO 8cb 228 3030 26.8 98 100 50 11 6 16 CO 8db 228 3030 26.8 42 100 16 10 2 13 CO 228 3025- 26.8- 90 100 39 11 6 16 8c/8d 3040 26.9 CO 8e 228 3060 27.1 100 30 4 17 3 7 Cl 242 3100- 27.2- * * * * * * 8a/8e 3360 29.3 C2 256 3200- 28.3- * * * * * * 8a/8e 3500 31.6 4 CO 9a 252 3490- 30.8- 2 100 60 25 11 30 3500 31.7 CO 9b 252 3490- 30.8- 100 64 25 28 3 22 3500 31.7 CO 9c 252 3615 31.7 2 100 20 86 4 32 CO 9d 252 3640 31.9 100 28 1 19 0 7 CO 9e 252 3673 32.2 100 26 1 6 0 1 Cl 266 3600- 32.0- * * * * * * 9a-4e 4000 34.5 ( a ) M = = M+, M + l = = Mrf", M+27 = [M+CH 3 CHF-HF] + , M+47 = [M+CH 3 CHF] + , M+45 = [M+CH 3 CF 2 -HF]+ , M+65 = PVI+CH 3CF 2] + (b) Data for standard 8c and 8d compounds shown. * Complex mass spectral data containing many isomeric compounds. 171 5.3.2 Adduct Formation and Isomer Differentiation Co-eluting substances that possess identical mass spectra cannot be differentiated from each other. In order to differentiate co-eluting substances by using mass spectrometry they must either yield unique characteristic ions or, if they yield ions at identical m/z values, the relative intensities of these ions must be significantly different. The process of isomer differentiation by chemical ionization relies on the formation of adducts in the gas phase leading to characteristic mass spectral fingerprints. As demonstrated in Chapter 4, certain P A H structural isomers, such as 5a and 5b, will react differently, leading to characteristic mass spectra under 1,1-DFE CI conditions. Although PAHs 5a and 5b are chromatographically separable, and therefore can be identified by a combination of their retention time and standard mass spectral information, the presence of characteristic mass spectral signatures for these compounds will still be useful for resolving them in samples containing co-eluting isobaric interferences. However, the method of isomer differentiation discussed.here will be particularly useful for the alkyl substituted analogues. There are many structural isomers in the C l and C2 alkyl P A H homologue series indicated in Figure 5.1 which cannot be differentiated from each other because of either the lack of analytical standards, or more importantly, because they co-elute or elute closely. Standards for many alkyl PAHs are not available, thus their anticipated adduct-forming behavior has to be extrapolated from that of their unsubstituted parent analogues. To validate this extrapolation, a series of available methyl and dimethyl P A H standards were analyzed by using 1,1-DFE/CI and the results were compared to those for the corresponding unsubstituted PAHs (Table 5.2). 172 Table 5.2(a) Effect of methyl substitution on adduct formation under 1,1-DFE CI conditions % Relative Intensity a Substance IE b (eV) M M+l M+27 M+47 M+45 M+65 5b 7.86 38 100 68 <1 10 3 l&2-Methyl-5b 7.7 40 100 50 <1 8 5 3,6-Dimethyl-5b 7.6-7.8 c 98 100 46 <1 6 2 5a 7.45 100 45 13 12 <1 6 l-Methyl-5a n/a 100 47 12 12 <1 5 2-Methyl-5a 7.37 100 43 8 13 <1 6 9,10-Dimethyl-5a 7.22 d 100 25 1 7 <1 2 6b 7.95 42 100 83 4 11 7 2-Methyl-6b n/a 52 100 62 7 2 9 6a 7.41 100 45 15 17 4 14 8b 7.43 100 56 15 30 1 11 2-Methyl-8b 7.30 100 52 10 21 <1 7 4-Methyl-8b 7.30 100 47 7 30 1 8 7-Methyl-8b 7.24 100 48 4 23 <1 5 3,9&6,8-Dimethyl-8b 7.20 6 100 40 6 15 <1 4 7,12-Dimethyl-8b 7.10 100 29 1 8 <1 4 (a) M = M 4 ", M + l = M r C , M+27 = [M+CH 3 CHF-HF] + , M+47 = [M+CH 3 CHF] + , M+45 = [M+CH 3 CF 2 -HF] + , M+65 = [M+CH 3 CF 2 ] + (b) A l l values from Lias 1 3 . (c) EI data for 2,7; 4,5; and 9,10 - 5b isomers. (d) EI data for 9-methyl-5a. (e) EI data for 3,9-dimethyl-8b. 173 Table 5.2(b) Ratio of % relative intensity values for [M+R-HF]+ and [M+R]+ adducts Substance [M+R-HF1+ b [M+R1+ M+27 M+45 M+47 M+65 5b lxlO 2 3 l&2-Methyl-5b lxlO 2 2 3,6-DimethyI-5b lxlO 2 3 5a 1.1 0.08 l-Methyl-5a 1.0 0.1 2-Methyl-5a 0.6 0.08 9,10-Dimethyl-5a 0.1 0.3 6b 2xlOx 2 2-Methyl-6b 9 0.2 6a 0.88 0.3 8b 0.50 0.09 2-Methyl-8b 0.48 0.07 4-Methyl-8b 0.2 0.1 7-Methyl-8b 0.2 0.1 3,9&6,8-Dimethyl-8b 0.4 0.1 7,12-Dimethyl-8b 0.1 0.1 (a) M+27 = [M+CH 3 CHF-HF + , M+47 = [M+CH3CH] (b) % RI is <1 were rounded to a value of 0.5 for the purpose of these calculations, n/a not available. The % RI values shown in Table 5.2(a) appear to indicate that in some cases the presence of methyl substituents results in a decrease of the relative amounts of both the protonated species MFC and of the elimination products [M + R - HF] + (i.e. [M + 27] and [M + 45]). The decrease in the relative intensities of M H + and [M + R - HF] + ions may be attributable to the slightly lower ionization energy of the methyl isomers. As was demonstrated in Chapter 4 (Section 4.3.4), decreases in IE generally result in a decrease in the relative intensities of both MET and [M + R - FfX]+ ions. Although the addition of one or two methyl substituents does slightly affect the amounts of adducts formed within one isomer group (e.g. 5a and its methyl isomers) it does not affect the relative adduct forming behavior between structural isomers (e.g. between methyl-5a and methyl-5b). For example, the [M + 27] / [M + 47] relative intensity ratio of 5b and its C l and C2 analogues, shown in Table 5.2(b), is two orders of 174 magnitude larger than for 5a and its C l and C2 analogues. It is these differences in relative intensity that are important for isomer differentiation purposes. Thus, it appears that the adduct forming behavior of methyl substituted PAHs can be extrapolated from that of their unsubstituted analogues. This extrapolation is important as it enables the prediction of the mass spectral behavior of isomers for which standards are not available. The data in Table 5.2(b) also show that the [M + CH 3 CHF-HF] + ions (M+27), and [M + CH 3 CHF] + ions (M+47) offer the best differentiation, since their ratio displays the greatest difference between isomers. Although the differentiation between different positional isomers would also be desirable, it appears from the data shown in Table 5.2(a) that the mass spectral variations between positional isomers are not sufficient to permit such a process. 175 5.3.3 Differentiation of Isomers Through the Comparison of Mass Spectral Data The simplest way to search for the presence of different structural P A H isomers in a complex chromatographic profile is to examine the mass spectral data extracted from the TIC. For example, in Figure 5.2 the mass spectra from three different scans along the Cl-5a/5b homologue series (scans 1981, 2015, 2037) and three positions along the C2-5a/5b homologue series (scans 2169, 2227, 2253) are shown. Mass spectra at scans 1981 and 2015 clearly indicate the presence of a methylphenanthrene (Cl-5b) and a methylanthracene (Cl-5a), respectively, as indicated by the large amounts of m/z 219 (M+27) in the former and the presence of m/z 239 (M+47) in the latter. The mass spectrum at scan 2037 contains lower amounts of m/z 219 than one would expect for a Cl-5b isomer but more than is expected for a Cl-5a isomer. Furthermore, this mass spectrum displays both m/z 239 and m/z 237 ions indicating the presence of both Cl-5a and Cl-5b species co-eluting. The higher % RI of m/z 239 over m/z 237 suggests a greater abundance of the Cl-5a isomer eluting at scan 2037. Similarly, scans 2169 and 2227 are indicative of the presence of C2-5b and-5a respectively, while scan 2253 shows the presence of both isomers. Although comparison of mass spectral data along different regions of the TIC can confirm or exclude the presence of a particular target compound, it does not allow a clear understanding of the distribution of isomers eluting along the chromatogram. To obtain a better understanding of the elution profile of the PAHs present in the sample, ion chromatograms corresponding to the molecular ion and the major adduct ions should be extracted from the TIC. S c a n 1 9 8 1 1 0.4 0.2 Cl-5b 1 9 3 150 180 2 1 9 2 3 7 210 240 m/z 2 5 7 , I, S c a n 2 0 1 5 1 « 0.6H CC 270 300 0.2 Cl-5a 1 9 2 S c a n 2 0 3 7 -i 0.8-• in c - £ 0.6--t Relative 0.4-- ' ' I f , T.T.T, I 0.2-0-210 240 m/z 270 300 176 Cl-5a/5b 1 9 2 -<—r-f 180 2 1 9 2 3 7 | 2 | 7 210 1 i i I i i I 240 270 300 S c a n 2 1 6 9 1 " I C2-5b 2 0 7 2 3 3 2 5 1 2 7 1 S c a n 2 2 2 7 0.6-0.4-150 180 210 240 270 300 m/z C2-5a 2 3 3 2 5 3 . . 2 7 1 0.8' 150 180 210 240 270 300 m/z C2-5a/5b --2 3 3 2 5 3 2 7 1 1 1 1 1 1 I . . 2511 , 180 210 240 270 300 m/z Figure 5.2 Mass spectra for Cl-5a/5b and C2-5a/5b extracted from chromatogram in Figure 5.1. 177 5.3.4 Differentiation of Isomers from the Analysis of Mass Chromatograms Analysis of the data from the sediment extract TIC was carried out by first plotting the M , M+l, M+27 and M+47 mass chromatograms for the TIC regions of interest. These chromatograms represent the dominant ions in the mass spectra of the PAHs, and will generally display the largest mass spectral differences between isomers. In some instances these differences are evident, while in other cases, particularly for the methylated analogues, the differences are more difficult to observe. For example, comparison of the M and M+47 ion chromatograms in Figure 5.3 clearly reveals major differences between 5a and 5b, and between 6a and 6b. These differences are particularly notable in the M+47 ion chromatograms, which is dominated by the 5a and 6a isomer signals. Similarly, the M+27 chromatogram is dominated by the 5b and 6b isomers. It was found that a simple transformation of the mass chromatograms, such as the subtraction of two normalized ion chromatograms from each other, for example [(M+47) -(M+27)], can enhance differences in the ion chromatograms, thus facilitating the identification of the different isomers. To illustrate this method the standard mass chromatograms and "transformation chromatogram" (TCH) for the unsubstituted (CO) isomer pairs 5a/5b and 6a/6b are shown in Figure 5.3. This transformation enables the 5a and 6a signals to plot positively and the 5b and 6b signals to plot negatively with respect to the baseline. 178 i — i — i — i — i — i — i — i — i — i — i | i i i i i i i i i i i i i i i i i i i i i i i i i i scan number 1 7 2 0 1 7 6 0 1 8 0 0 2320 2400 2480 t ime (min) 15.93 16.27 16.60 20.94 21.59 22.27 Figure 5.3 Unsubstituted (CO) 5b/5a and 6b/6a ion chromatograms from 1,1-DFE chemical ionization analysis of the sediment extract; bottom: transformation chromatogram. The chromatographic transformation approach becomes particularly useful when applied to the retention time regions where the methylated PAHs elute, as these regions contain an even greater abundance of isomers. Figures 5.4 and 5.5 display the chromatographic data for the monomethyl (Cl) and dimethyl/ethyl (C2) isomer series for 5a and 5b. Figure 5.4 illustrates the m/z 192 (M), m/z 193 (M+H), m/z 219 (M+27) and m/z 239 (M+47) mass chromatograms for the Cl-5a/5b isomer series, while Figure 5.5 shows the m/z 206 (M), m/z 207 (M+H), m/z 233 (M+27) and m/z 253 (M+47) mass chromatograms for the C2-5a/5b isomer series. scan number i960 1980 2000 2020 2040 2060 t ime (min) 17.94 18.27 18.60 Figure 5.4 Monomethyl (Cl) 5b/5a ion chromatograms and transformation chromatogram from 1,1-DFE chemical ionization analysis of sediment extract. 180 9 C2-5b i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i scan number 2160 2200 2240 2280 t ime (min) 19.60 19.94 20.27 20.61 Figure 5.5 Dimethyl / ethyl (C2) 5b/5a ion chromatograms and transformation chromatogram from 1,1-DFE chemical ionization analysis of sediment extract. 60 5b 40 H 20 0 5a co CD CD C1 5a/5b c m I I I *— *— CD o oo *— *— CNJ CO LO LO CD 03 CO 00 00 00 00 C2 5a/5b 7 is 8 m 00 CD 00 CD i - CO CJ) o o T - CN CM • [M + 27] • [M + 47] Time (min) Figure 5.6 % RI of [M+27] and [M+47] adducts at selected regions indicated on chromatograms displayed in Figures 5.3 to 5.5. 182 The ion chromatograms displayed in these figures represent the most characteristic ions for these two groups of isomers, and certain differences can be observed between the first four ion chromatograms in each figure. These differences are of a nature similar to those observed between the chromatograms in Figure 5.3. For example, the m/z 219 (M+27) chromatogram in Figure 5.4 shows mainly the presence of methylphenanthrenes (Cl-5b) since 5b produces much larger amounts of this ion than 5a, while the m/z 239 (M+47) chromatogram emphasizes the elution profile for the methylanthracenes (Cl-5a) since 5a produces a larger amount of this ion than 5b. In theory, there are three possible monomethyl isomers for anthracene (1-, 2- and 9-methyl) and five for phenanthrene (1-, 2-, 3-, 4- and 9-methyl). Retention time information from available standards and literature values 1 1 0 indicate that the 1- and 2-methylanthracene co-elute with 1-, 4- and 9-methylphenanthrenes (scans 2010 to 2050). The isomer 9-methylanthracene elutes at a later retention time and is not present in Figure 5.4. Under EI GC/MS conditions it would not be possible to verify the presence of both the 1- and 2-methylanthracene isomers because of the lack of any characteristic ions that would distinguish them from the methylphenanthrenes. However, by generating a characteristic ion (e.g. m/z 239) through reaction with C H 3 C H F + ions the presence of 1- and 2-methylanthracene in the sediment extract can be clearly confirmed in the m/z 239 chromatogram (2-methylanthracene at position 3, 1-methylanathracene at position 5). The % RI of the [M + 27]+ and [M + 47]+ adduct ions observed in the mass spectra at the numbered positions indicated on the chromatograms in Figures 5.4 (1 to 6) and 5.5 (7 to 9) are displayed in the bar graph in Figure 5.6. By comparing the % RI values of the two adduct ions eluting at regions 1 to 7, to those for 5b and 5a, it becomes clear that the isomers eluting at positions 3 and 8 are predominantly 5a derivatives and those eluting at positions 1, 183 2, 4, 6 and 7 are predominantly 5b derivatives. At positions 5 and 9 the relative intensity of [M + 27] is greater than expected for pure 5a methyl isomers, indicating the presence of both 5a and 5b methyl isomers. 5.3.5 Mass Chromatogram Transformations As mentioned earlier, transformations can be performed to improve the comparison between mass chromatograms; for this purpose the chromatograms were normalized prior to subtraction. The results obtained from taking the difference between the normalized m/z 239 and m/z 219 ion chromatograms for the Cl-5a/5b data are illustrated in Figure 5.4 (bottom chromatogram). This [(M+47)-(M+27)], (m/z 239-m/z 219), chromatogram indicates the regions where the Cl-5a and Cl-5b isomers are dominant. The region of dominant Cl-5a abundance is identified by the chromatogram plotting positive, while the region of dominant Cl-5b abundance is identified by the chromatogram plotting negative. The usefulness of subtracting normalized ion chromatograms for highlighting differences in the elution profile of isomers is further exemplified in the C2-5a/5b homologue series (Figure 5.5) by the (m/z 253-m/z 233) transformation chromatogram. C2 homologue series are generally more complex because of the larger number of isomers present. Even if a complete set of standards were available, the number of co-eluting species would make differentiation between C2-5a and C2-5b isomers virtually impossible by using GC/EI-MS. However, by using the (M+47)-(M+27) transformation mass chromatograms shown in Figure 5.5 the presence of both C2-5a and C2-5b structural isomers groups can be confirmed. While there are some differences in the elution profiles for the C2-5b and the C2-5a in the m/z 233 (M+27) and m/z 253 (M+47) ion chromatograms, the normalized transformation chromatogram [(M+47) - (M+27)], (m/z 253 - 233), clearly shows the tR locations for the C2-5b (negative plot) and C2-5a (positive 184 plot). Analysis of the mass spectra in Figure 5.2 (scans 2169, 2227, 2253) and of the % RI data in Figure 5.6 (signals #7 to #9) further confirms the presence of both C2-5b and C2-5a isomers. The chromatogram transformation technique can also be applied to mass chromatograms, such as the M and M+l chromatograms, that display only minor differences. In fact, it is in these instances when the transformation technique becomes most useful. For example, the results of subtracting the normalized m/z 207 (M+l) chromatogram from the normalized m/z 206 (M) chromatogram are shown in Figure 5.7 (negative plot: C2-5b, positive plot: C2-5a). Also plotted are the normalized sum of the mass chromatograms of the six major ions listed in Table 5.2(a) for C2 5b/5a. Although small differences can be observed in the M and M+l chromatograms in Figure 5.5, the transformation chromatograms shown in Figure 5.7 highlight these differences more clearly. Comparison between the transformation chromatograms in Figure 5.5 (m/z 253-m/z 233) and 5.7 (m/z 206-m/z 207) reveals that these chromatograms are similar, although the degree of positive and negative enhancements varies slightly. 185 | 1 1 1 1 1 1 1 1 1 1 1 1 1 1 scan number 2160 2200 2240 2280 time (min) 19.60 19.94 20.27 20.61 Figure 5.7 Dimethyl / ethyl (C2) 5b/5a sum and m/z 206-m/z 207, (M-[M+JJ), transformation chromatogram from 1,1-DFE chemical ionization analysis of sediment extract. A second type of chromatographic transformation that enabled enhancement of ion chromatographic differences in regions 2 to 4 (Figure 5.1), was [(A-B) x C] where A, B and C represent individual normalized ion chromatograms, for example: [(M+47) - (M+27)] x 186 (M+47). Multiplying the difference chromatogram [(M+47) - (M+27)] by (M+47) allows a stronger enhancement of the isomers possessing a large abundance of [M + 47]+ ions. So far chromatograms for region 1 (Figure 5.1) have been analyzed in detail. To simplify the presentation of the data for regions 2 to 4, where deemed appropriate, only the normalized sum for the six major ions (M, M+l, M+27, M+45, M+47 and M+65) together with the respective transformation chromatograms are shown. 5.4 ANALYSIS O F ISOMERS IN REGIONS 2 T O 4 5.4.1 Region 2 As has been shown previously, fluoranthene (6b) and pyrene (6a) compounds eluting in region 2 have adduct forming behavior similar to phenanthrene (5b) and anthracene (5a) respectively (Tables 4.6 and 5.1). Once again it is the alkyl homologue series C l and C2 that are of interest because 6a and 6b are themselves chromatographically well separated. The number of possible monomethyl substitutions on 6a and 6b are three and five respectively. However, three other compounds that could potentially co-elute with the C l 6a and 6b isomers are the benzofluorenes (7a, 7b and 7c). The latter three compounds have a molecular mass of 216 u and exhibit EI spectra virtually identical with the Cl-6a and Cl-6b species. Fortunately, the DFE-CI mass spectra of these three groups of compounds (Cl-6a, Cl-6b and 7a-7c) are significantly different, as illustrated by the % RI values of the [M + 27]+ and [M + 47]+ ions in Table 5.1. The sum of the normalized extracted ion chromatograms for the six most important ions for the C l series (m/z 216, 217, 243, 261, 263 and 281) is shown in Figure 5.8. Identifications have been made based on mass spectral data and by comparing the [(M+47) -(M+27)] transformation ion chromatogram of the unsubstituted isomers 6b /6a (m/z 249 -187 m/z 229) to that of the Cl -6b/6a (m/z 263-m/z 243). The 6b and 6a isomers plot negative and positive respectively in the CO chromatogram. Based upon this trend the region where the Cl -6a compounds are eluting is identified as the positive regions of the chromatogram (signals #5 to #7). In contrast, the region where the Cl -6b compounds elute are identified by the chromatogram going negative (signals # 1 and #2). The remaining regions (# 3 and #4) where the chromatographic signals are of low intensity correspond to regions where the 7a-7c benzofluorene isomers elute. The presence of benzofluorenes was confirmed by comparison of the mass spectral data between scans 2600 - 2640 with that for a standard of 7a. The small peak that occurs between signals 5 and 7 in the transformation chromatogram is not due to a Cl -6a isomer. The mass spectral data at this scan revealed the presence of a compound with a molecular mass of 218 u. This compound is able to form a M+45 adduct at m/z 263 that interferes with the M+47 signal of the Cl-6a/6b isomers (M = 216). Also illustrated in Figure 5.8 are the summed mass chromatogram (m/z 230, 231, 257, 275, 277 and 295), and the transformation chromatogram for the C2 series. The transformation chromatogram clearly shows the predominance of C2-6b isomers at position 8, as indicated by the transformation chromatogram plotting negative. The other regions in the transformation chromatogram display mainly weak positive enhancements, indicating that a mixture of C2-6a and C2-6b isomers is probably co-eluting, with perhaps a greater abundance of the C2-6a isomers. It is also important to realize that the possibility exists for C l isomers of benzofluorenes (7a/c) to be present since they possess the same molecular mass as C2-6a/6b isomers. To further clarify the elution profile of the isomers shown in Figure 5.8, the relative intensities of the [M + 27]+ and [M + 47]+ adduct ions observed in the mass spectra at the numbered positions in Figure 5.8 are summarized in Figure 5.9. This clearly confirms the 188 identification of Cl-6a isomers at positions 5, 6 and 7, based upon the close match of their abundance ratios with the one for 6a. Similarly, positions 1 and 2 closely match the identity of 6b and can therefore be considered Cl-6b type isomers. Positions 3 and 4 correspond to the mass spectral values of the 7a/c isomers. In the C2 isomer region (positions 8 to 11) position 8 clearly shows the predominance of a C2-6b isomer, and position 10 of a C2-6a isomer. The abundance ratios at positions 9 and 11 appear to suggest a mixture of isomers (C2-6a, C2-6b and possibly Cl-7a/c). 189 scan number 2320 2400 2 4 8 0 2560 2640 2720 2 8 0 0 2880 t ime (min) 20.94 21.61 22.27 22.94 23.61 24.27 24.94 25.61 Figure 5.8 Sum and transformation chromatograms for unsubstituted (CO), monomethyl (Cl) and dimethyl / ethyl (C2) 5a / 5b and 7a/c isomers. For the (M+47)-(M+27) chromatogram M = 202 for the CO data, M = 216 for the Cl data and M = 230 for the C2 data. 190 [M + 27] [M + 47] LO CN o o o ^ c N c o ^ t c o r ^ ^ r ^ c o oo r^ - oo T - CN LO LO cn o o CM O r - CM 00 00 00 00 00 <^  ^ LO LO LO CM CN CM CM CM CM CM CM CM CM CM CM CM Time (min) Figure 5.9 % RI of [M+27] and [M+47] adducts at selected regions indicated on the chromatograms displayed in Figure 5.8. 191 5.4.2 Region 3 Isomer analysis becomes more complicated in region 3 because of the presence of five structural isomers, benzo[c]phenanthrene (8a), benz[a]anthracene (8b), chrysene (8c), triphenylene (8d) and naphthacene (8e). Identification of these five compounds in the mass chromatogram is shown in Figure 5.10. A standard was not available for 8a but its retention time relative to the other four isomers is consistent with literature values 1 1 0 . The only co-elution problem in the CO series is with compounds 8c and 8d. The major difference between the mass spectra (Tables 5.1) of these two compounds is the larger amounts of m/z 255 (M+27) formed by 8c. Another difference is the greater abundance of m/z 229 (M+l), with respect to m/z 228 (M) for isomer 8d (intensity ratio of m/z 229/228 ~ 1 for 8c and ~ 2 for 8d). Based upon these relative differences, the two compounds were partially de-convoluted through the [(M+l) - (M+27)] x (M+l) transformation shown in Figure 5.10. By enhancing the signal for the characteristic ion of 8d (M+l) and reducing the signal for the characteristic ion of 8c (M+27) this transformation chromatogram produces a positive plot in the region where 8d is dominant and a negative plot in the region where 8c is dominant. From this chromatogram it is possible to observe that 8d begins to elute just prior to 8c. Comparison of the mass spectral data in Table 5.1 at scans 3025-3040 with the mass spectral data of 8c and 8d standards confirms that both of these isomers are present. A [(M+27) - (M+47)] x (M+27) transformation was also performed. This transformation enabled the 8a and 8c signal to plot positively and 8b negatively. 192 m/z 228 m/z 229 M+27 M+47 [(M+27)-(M+47)] x (M+27) [(M+1)-(M+27)]x(M+1) i — i — | — i — i — i — | — i — i — i — | — i — i — i — | — i — i — i — i — i — i scan number 2920 2960 3000 3040 3080 t ime (min) 25.94 26.27 26.66 26.94 27.77 Figure 5.10 Ion and transformation chromatograms for unsubstituted (CO) 8a to 8e PAHs (M = 228 u). The m/z 242, 243, 269 and 289 ion chromatograms for the Cl-8a to Cl-8d isomers are illustrated in the upper part of Figure 5.11. Two transformation chromatograms, of the same type as in Figure 5.10, are displayed in the lower part of Figure 5.11. Analysis of these chromatograms reveals the presence of a strong positive signal at scan 3236 (signal number 5, 28.53 min) in the [(M+27) - (M+47)] x (M+27) transformation and a corresponding negative signal in the transformation [(M+l) - (M+27)] x (M+l). These signals indicate the presence of Cl-8c in accordance with the behavior of 8c in Figure 5.10. Furthermore, the adduct ion 193 intensity ratios shown in Figure 5.13 for signal 5 also indicate a significant contribution from a Cl -8c isomer. The other positive signals in the [(M+27) - (M+47)] x (M+27) chromatogram (numbers 2, 3 and 6) are expected to be either Cl -8a or Cl -8c as suggested by the data in Figures 5.10 and 5.13. Although 8a and 8c exhibit similar mass spectra, 8a elutes earlier. If this relative elution order can be extrapolated to their C l analogues (Cl-8a and Cl -8c) then signals 2 and 3 are probably Cl -8a isomers and signal number 6 represents Cl -8c isomers. The remaining labeled signals, numbers 1 and 4, have ion ratios (Figure 5.13) resembling those of 8e and 8b respectively. The analysis of the C2 data is shown in Figure 5.12 (chromatograms) and Figure 5.13 (mass spectral data). From the transformation chromatograms it appears that C2-8b compounds dominate this homologue series. This is based upon the negative peaks in [(M+27) - (M+47)] x (M+27) and corresponding positive peaks in [(M+l) - (M+27)] x (M+l), in accordance with the earlier discussed behavior of 8b in Figure 5.10. The negative signal at position number 12 in [(M+l)-(M+27)] x (M+l) chromatogram of Figure 5.12 indicates the possible presence of a C2-8c isomer, in accordance with the behaviour of 8c (Figure 5.10). Figure 5.11 Ion and transformation chromatograms for monomethyl (Cl) 8a to 8e PAHs (M = 242 u). 195 12 Sum [(M + 27HM + 47)] x (M + 27) [(M + 1)-(M + 27)] x (M + 1) scan number 3200 3280 3360 3440 time (min) 28.27 28.94 29.60 30.27 Figure 5.12 Ion and transformation chromatograms for dimethyl / ethyl (C2) 8a to 8e PAHs (M = 256 u). 196 60 H E 4 0 20 0 8a 8d8c + r^ i 8c 8b 8d 8e 1 *N C1 8a/8e 3 C2 8a/8e 8 12 0 1 1 CD CO 00 00 oc i ts oo 0 -a T -LT5 CD cd 0 0 0 0 CNI CN CN CN 00 CO O CO CD in O <- in oo rv T - o LO CD co LO o CN co T -N 00 OO 00 00 00 CN CN CN CN CN CN Time (min) co co co co o o CN CN CN CM 00 CO • [M + 27] • [M+47] Figure 5.13 % RI of [M+27] and [M+47] adducts at selected regions indicated on the chromatograms displayed in Figure 5.10 to 5.12. 5.4.3 Region 4 The major CO structural isomers observed in region 4 are benzo[b]fluoranthene (9a) benzo[k]fluoranthene (9b), benzo[e]pyrene (9c), benzo[a]pyrene (9d) and perylene (9e). This group of isomers is of both toxicological and environmental interest. For example, 9d is a potent carcinogen while 9c is not 9 0 . Compound 9e is a natural occurring P A H which can be used as a marker for natural and anthropogenic inputs m . Their elution order in the sediment sample is illustrated in Figure 5.14. With the exception of 9a and 9b these compounds are well resolved. A compound that belongs to this set of isomers (M r = 252 u), but whose identity is unknown has been labeled with a question mark (?). The mass spectra for these six unsubstituted isomers are distinctly different, as noted in Figure 5.16 by the % RI values of the [M + 27] and [M + 47] adducts. These isomers also exhibit distinctly different mass chromatogram profiles, as illustrated in Figure 5.14. For example, the m/z 252 (M) 197 chromatogram is dominated by isomers 9d and 9e, while the m/z 279 (M+27) chromatogram is dominated by isomers 9a and 9b. The transformation chromatograms that were found to highlight the most important differences are shown in Figure 5.14. Transformation [(M+47) -(M+27)] x (M+47) enables 9a and 9b to plot negative while enhancing 9c over 9d and 9e. In contrast, the transformation [M- (M+27)] x M was found to strongly enhance the 9d signal. M M+1 M+27 M+47 9a/9b [(M+47)-(M+27)] x (M+47) [ M - (M+27) ] x M m/z 252 m/z 253 m/z 279 m/z 299 i i i i i i i scan number 3520 time (min) 30.94 i l i i i l i i i I i i i I 3600 3680 31.60 32.27 Figure 5.14 Ion and transformation chromatograms for unsubstituted (CO) 9a to 9e PAHs (M= 252 u). Although the CO isomers are chromatographically well separated the same is not true for the C l isomers as shown in the normalized summed (m/z 266, 267, 293, 311, 313, 331) 198 chromatogram in Figure 5.15. Fortunately, these compounds can be distinguished based upon the adduct forming behavior of their unsubstituted analogues. When the same set of transformations used for the CO isomers was applied to the C l data a remarkable deconvolution resulted, as illustrated in Figure 5.15. Sum [(M+47)-(M+27)] x (M+47) [ M - (M+27) ] x M C1-9a/9b scan number 3600 time (min) 31.60 3800 33.27 4000 34.93 Figure 5.15 Ion and transformation chromatograms for monomethyl (Cl) 9a to 9e PAHs (M = 266 u). The [(M+47) - (M+27)] x (M+47) transformation clearly illustrates the elution region of the C l 9a and Cl-9b isomers (negative plots) and the Cl-9c isomers (enhanced positive plots). 199 The second transformation, [M - (M+27)] x M , highlights the elution region of the Cl-9d isomers. Further confirmation of the elution of the Cl-9d isomers in this region was obtained from three available standards (8, 9, and 10 monomethyl 9d). The last eluting compounds are probably Cl-9e isomers in accordance with the behavior of the 9e in Figure 5.14. The % RI values in Figure 5.16 are consistent with the chromatographic data. The % RI values of signal 1 correspond to the values for 9a or 9b, signal 2 most closely resembles the 9c isomer, and signal 3 the 9d isomer. Signal 4 has a % RI intensity slightly closer to 9e than 9d, although the [M + 27] intensity is greater than expected for either isomer. 100 -\ 80 E 60 40 H 20 0 9a 9b 9c 9d 9e C1 9a/9e 2 i l l Time (min) LO 00 LO o CO CO LO OO CO * — 1 ^ co CM t — LO o o o , — , — , — CM CM CO CO • S t CO CO CO CO CO CO CO CO CO CO I I [M + 27] • [M+47] Figure 5.16 % RI of [M+27] and [M+47] adducts at selected regions indicated on the chromatograms displayed in Figures 5.14 and 5.15. 200 5.5 QUANTITATION As demonstrated previously in Chapter 4, (Table 4.13), under automatic reaction control (ARC) conditions the effect of concentration does not alter the adduct forming behavior of isomers over a concentration range of at least three orders of magnitude. The results obtained from the analysis of 5 to 57000 pg of 5b by both 1,1-DFE/CI and EI analyses are summarized in Figure 5.17 and Table 5.3. The EI data (m/z 202 ion) display a longer linear range and a higher slope signifying a slightly better calibration sensitivity. For the CI analysis, the most reliable ion for quantitation purposes appears to be the molecular ion (m/z 202), however, the m/z 229 (M+27) and m/z 249 (M+47) adduct ions also display reasonably good linear fits making them also analytically reliable. • 6b m/z 202 El • 6b m/z 202 Cl o 6b m/z 229 Cl A 6b m/z 249 Cl log pg Figure 5.17 GC integration area versus injection amount relationship for EI and CI analysis of fluoranthene (6b). 201 Table 5.3 Summary of linear regression parameters for data plotted in Figure 5.17 a Ionization method Quantitation ion (m/z) slope (log A/ log Q) intercept r EI 202 1.17 1.48 1.000 DFE/CI 202 1.05 0.70 0.999 DFE/CI 229 0.866 1.44 0.998 DFE/CI 249 1.08 -0.93 0.994 A = integration area (arbitrary units), Q = quantity in pg The limits of detection (LOD) for PAHs in the sediment extract were calculated using 112 equation 5.1 LOD = m [5.1] where "m" represents the slope of the calibration curve (A vs. Q, on a linear scale) at the low concentration limit, and A,, represents the integrated area of the noise above the baseline. LOD values of 170 pg in the sediment (A„ = 500, m - 9 pg"1) extract and 50 pg (A„ = 150, m = 9 pg"1) in a standard solution were obtained using the m/z 202 CI data for compound 6b. The presence of multiple adduct ions for most isomers in this study provides a wide choice for quantitation. Generally one selects the most intense ion, but when co-eluting compounds have to be quantified the best choice will be ions that are characteristic of each compound. For example, for a C l 5a/5b mixture, using the m/z 239 and m/z 219 ions should offer the best results. As reported in Figure 4.14, the relative compound proportions for a pair of co-eluting isomers can be determined from the measurement of a characteristic ion. Quantitation of PAHs present in the sediment was performed for a series of compounds for which standards were available in order to verify the reliability of this chemical 202 ionization technique in comparison to standard E I analysis. Table 5.4 lists the relative response factors (RRF) of several PAHs, with respect to anthracene (5a). The concentration of the PAHs present in 5.0 g of sediment was estimated according to equation 5.2. Q = Areax ^ QdX0_5a [ 5 2 ] Aread]0_5a RRF^ d\-ia Q x represents the quantity (in u.g) of a given PAH "x", and Q dio-5a represents the 4.5 pg of dio-anthracene internal standard spiked into the sediment prior to extraction. Areax and Areaaio-5a represent the integration areas using a particular ion chromatogram, (e.g. IVf) for P A H "x" and dio-anthracene respectively. RRF represents the ratio of the RF values for PAH "x" and dio-5a (RRF = RF(x) / RF(di0-5a). 203 Table 5.4 Relative response factors (RRF), with respect to 5a, for standard PAHs analyzed by 1,1-DFE CI PAH M M+l M+27 M+47 5b 0.28 1.45 3.29 0.62 5a 1.00 1.00 1.00 1.00 6b 0.24 1.12 6.20 0.49 6a 0.71 0.78 2.10 0.73 8b 0.59 0.81 1.34 0.65 8c 0.42 1.24 3.66 0.66 9a+9b 0.13 0.61 1.21 0.27 9d 0.52 0.36 0.06 0.47 RRF = RF(PAH)/RF(5a), RF = A/Q, A = integration area, Q = quantity (2 ng) The results obtained from this analysis are listed in Table 5.5 along with an independent analysis of the same sediment (different sub-sample) performed by a commercial analytical lab employing GC/EI-MS analysis using a quadrupole instrument. Comparison of the CI and EI results indicates that the molecular ion (M+) serves as the best quantitation ion for CI analysis. A larger uncertainty in the CI data results from the use of dio-anthracene as the only internal standard. Ideally, the use of a larger selection of perdeuterated internal standards would offer a more accurate estimate of the true concentrations present in the sediment. Unfortunately, from a practical perspective the use of many perdeuterated PAH internal standards is limited by their high purchase price. 204 Table 5.5 Concentration, in ppm, (u.g/g) of selected PAHs present in sediment sample determined using 1,1-DFE chemical ionization, together with concentration values determined by an independent analysis employing GC/MS EI CI EI a PAH Quantitation Ion M M+l M+27 M+47 Average StDev. 5b 1600 640 250 1700 1000 720 1500 5a 470 340 140 160 300 160 510 6b 1300 670 130 270 600 540 1300 6a 800 560 110 220 420 310 1100 8b 530 400 90 160 300 210 280 8c+d 500 420 80 130 280 210 370 9a+b 220 900 350 330 450 310 580 9c 19 250 36 350 160 160 190 9d 520 440 n/d 170 380 180 270 9e 160 130 n/d 18 100 75 67 n/d = not detectable, (a) Quadrupole GC/MS analysis performed by an independent analytical laboratory on a sub-sample of the sediment analyzed in the current work. This analysis was performed using perdeuterated internal standards for most of the target PAHs analyzed. 205 5.6 SUMMARY Analysis of a contaminated sediment extract by using GC/1,1-DFE-CI has enabled the identification of several P A H isomers which cannot be identified by GC/EI-MS. Through the use of ion chromatogram analysis and transformations techniques a number of P A H isomer groups eluting as complex clusters were deconvoluted. Of particular interest was the ability to distinguish between monomethyl isomers of benzo[e]pyrene (9c) and benzo[a]pyrene (9d), since 9d is a more powerful carcinogen than 9e. Quantitation with this CI technique was possible by using relative response factors (RRFs) calculated with respect to di0-anthracene. However, a high relative standard deviation (50 - 100 %) resulted between the values obtained using different quantitation ions. Improved quantitation would be possible if in each case the RRFs were obtained with respect to the perdeuterated analogue of the compound quantified. The reasonably low limits of detection (170 pg in sediment) coupled with a linear dynamic range comparable to the one obtained by EI-GC/MS, suggest that this CI technique has a good potential to be exploited for a large variety of P A H environmental analyses. 206 CHAPTER 6 ION / MOLECULE REACTIONS OF PERFLUOROTRIBUTYLAMINE CATIONS WITH POLYCYCLIC AROMATIC HYDROCARBONS IN A QUADRUPOLE ION-TRAP 207 6.1 INTRODUCTION The results outlined in Chapters 4 and 5 have demonstrated that halogenated hydrocarbons, such as 1,1-difluoroethane, are good CI reagents for generating isomer specific mass spectra of PAHs. These reagents are able to form adducts of the type [M + R ] + and [M + R - H X ] + (M = PAH, R = reagent gas ion, X = Cl or F). Differences in the relative intensity of these adducts with respect to M + and M H + , form the basis for the differentiation of PAH isomers. The results from these experiments suggest that other halocarbon positive ions might also form adducts with PAHs. A potentially useful reagent that had not been investigated previously is perfluorotributylamine ((C 4F 9) 3N). Perfluorotributylamine, commonly referred to as FC43 is frequently used as a mass calibration compound in mass spectrometry. Upon electron ionization FC43 yields numerous fragment ions of the type C n F m + and C n F m N + (n = 1-12, m = 1-24) which are used for mass calibration. In consideration of the results obtained in Chapters 4 and 5 it was thought that these ions could provide a series of useful reagents for adduct formations with PAHs. However, to study the reactions of the individual FC43 ions, they have to be mass-isolated, a task that can be readily performed on an ion trap. The work outlined in this chapter reports on the study of gas phase reactions between neutral PAHs and a number of FC43 ions inside a quadrupole ion trap. 208 6.2 EXPERIMENTAL 6.2.1 Standards and Reagents The source of the PAHs employed in this study has been outlined in previous chapters. FC43 (Varian, Walnut Creek, C A , USA) was introduced into the ion trap through a Negretti needle valve (Fareham, Hants. GB). 6.2.2 Apparatus All experiments were performed using the Saturn 4D GC/MS/MS system described in Chapter 4. 6.2.3 Procedure Gas chromatography: Introduction of solutions containing 50 ng of each of the PAHs was performed by gas chromatography by using the same conditions reported in Chapter 4. Ion trap: The most important CI parameters employed are listed in Table 6.1. The initial experiments with FC43 were performed under "standard" CI conditions using all of the FC43 ions generated. "Selective" CI refers to ion/molecule reaction experiments performed by using individually mass-selected ions. Isolation of individual ions was achieved by using the standard MS/MS isolation function shown in Figure 6.1. Mass-isolation was performed with a 1 m/z unit window, and the CID energy was maintained at 0 volts. The standard CI reaction time value was kept at 1 ms (minimum value) since the reaction time between isolated ions and analyte molecules depended upon the CID time (T c i d) value in the MS/MS function. Unless otherwise specified, a T c i d of 100 ms was 209 used, providing a 100 ms reaction time between the isolated FC43 fragment ions and the neutral PAH molecules entering the ion trap. Table 6.1 List of major experimental parameters employed during "regular" and "selective" chemical ionization experiments Experimental Parameter Value(s) Used Emission current 10 uA CI reaction storage level 35 m/z CI ionization storage level 35 m/z CI reaction time (RT) (regular CI) 128 ms CI reaction time (RT) (selective CI) 1 ms CI background mass 35 m/z Ionization time (IT) (regular CI) 500 us Ionization time (selective CI) 1000-2500 ps Mass isolation window 1 u Isolation time 1 ms CID time (T c i d) 100 ms CID supplementary r.f. voltage 0 V CID storage level 35 m/z The automatic reaction control (ARC) function used for constantly optimizing the RT and IT values in the CI scan function was not employed for any of the experiments outlined in this chapter. Consequently, all experiments were carried out under constant IT and RT conditions. 210 RF Voltage • Time A: Ejection of m/z < m B: Ejection of m/z > m (with a 1 ms isolation time) C: Reaction of ion "m" for a time period = T d d CID Excitation voltage = 0 volts. D: Mass scan of resulting ion/molecule reaction products Figure 6.1 Scan function employedfor mass- isolation and reaction of FC43 fragment ions. 211 6.3 R E S U L T S A N D DISCUSSION 6.3.1 Formation and Isolation of FC43 Reagent Ions The major fragment ions generated from the electron ionization of FC43 in an ion trap are listed in Table 6.2, together with their relative intensity values. To be analytically useful a reagent ion should generally have a m/z value lower than the m/z of the molecular ion of the smallest analyte molecule, so as to avoid interferences. The lowest molecular mass compounds analyzed in this study were 5a and 5b, (m/z 178), making C 3 F 7 + (m/z 169) the largest useful ion for ion/molecule experiments. However, for the purpose of studying the interactions between PAHs and fluorocarbons, ions with m/z greater than 178 were also investigated. The mass-isolation efficiencies of the FC43 fragment ions are also listed in Table 6.2. Comparing the EI counts with the counts after mass-isolation for the ions that were able to be isolated, reveals that mass-isolation efficiency is in the range of 0.1 to 1.4 %. Consequently, only ions with sufficiently high ionization counts were usefully mass-isolated. 212 Table 6.2 Ions generated from the EI ionization of FC43 and their MS/MS isolation efficiency m/z Assigned Formula EI Ionization Counts (% RI) Mass-Isolation Counts (% RI) Mass-isolation % efficiency a 69 C F 3 + 291000(100) 4000(100) 1.4 93 C 3 F 3 + 6900 (2) 100 (3) 1.4 100 C 2 F 4 + 59000(20) 600 (15) 1.0 114 C 2 F 4 N + 22000 (7) 200 (5) 0.9 119 C 2 F 5 + 34000 (12) 40(1) 0.1 131 C 3 F 5 + 243000(84) 2000 (50) 0.8 150 C 3 F 6 + 7000 (2) 60 (2) 0.8 164 Q F X 5700 (2) 50(1) 0.9 169 C 3 F 7 + 11000 (4) 10(0.3) 0.1 181 C4V 18000 (6) 100 (3) 0.6 219 49000 (17) 0(0) 0 264 CsFioN* 58000(20) 100 (3) 0.2 314 CsFialST 3900 (1) 0(0) 0 364 C T F M N " 4700 (2) 0(0) 0 414 CgFigN" 26000(9) 20(1) 0.1 426 CgFjsN" 5000 (2) 10 (0.3) 0.2 464 10000 (3) 0(0) 0 502 C 9 F 2 0 N + 8000 (3) 0(0) 0 614 9500 (3) 0(0) 0 (a) % efficiency = (mass-isolation counts / ionization counts) x 100 213 6.3.2 Reactivity of Ionized FC43 Initially FC43 was tested as a CI reagent gas by allowing all of its ions (Table 6.2) simultaneously to react with anthracene (5a) and phenanthrene (5b). The resulting mass spectra are shown in Figure 6.2. The two spectra look very similar. Formation of M H + (m/z 179) by proton transfer, and IVf (m/z 178) by charge exchange, appear to be the two dominant ionization processes, together with some fragmentation products ( [M-2H]+ and [M-C 2H 2] +). 5a 5b in c 0} 0 ) > JO V CC 1 0 0 -8 0 -6 0 -40 20 0 179 152 m c QJ JO V oc - , , , •, , , , 2 , 6 4 , , , " i—|—i—^—i—|—i—r-1—|—i—i 100 8 0 -6 0 -4 0 -20 0 179 152 227 264 160 200 240 m/z 280 - i —| — i — r " n — | — i — n - i — |— i — i 160 200 240 280 m/z Figure 6.2 FC43 chemical ionization mass spectra of phenanthrene (5b) and anthracene (5a). No ions of significant intensity were observed above m/z 300. The absence of H atoms in FC43 suggests that MFT must be formed via proton transfer from other M + ions, or via collisions with residual H 2 0 molecules. The lack of significant amounts of adduct ions, a requirement for isomer differentiation, suggests that FC43 is not a useful reagent under standard CI conditions. To investigate whether individual FC43 ions could form adducts with PAHs it was necessary to perform selective ion/molecule reactions, that is, to mass-isolate individual FC43 214 ions and allow them to react with the PAHs eluting from the GC column into the ion trap. The results obtained were dramatically different from the standard CI experiments. For example, when the m/z 131 (C3F5+) ion was isolated and allowed to react with PAHs 5a and 5b, a number of intense adducts were obtained for both compounds, as shown in Figure 6.3. 5a 100H •3 80-c ^ 60-> 1 40-or 20 H 5b 178 151 160 cn c 03 > JO 309 100H 80 H 60 40 H 20 H 269 i l i i i l i i i i i i l | i | l | | | I 309 289 178 1.51 227 200 240 280 320 360 400 m/z l i i i l i i i l i i i l i i i l i i i l i i i l 160 200 240 280 320 360 400 m/z Figure 6.3 C ? F j + ion / molecule reaction spectra of anthracene (5a) and phenanthrene (5b). The larger number of adducts was formed with phenanthrene (5b), (m/z 227 = [M + 49] = [M + C F 2 - H], m/z 289 = [M + 111] = [M + C3F4 - H], and m/z 309 = [M + 131] = [M + C3F5] ). Anthracene (5a) reacted differently with C3F 5 + : like 5b it also formed m/z 309 ions but at a significantly lower relative intensity, but it did not yield any m/z 289 or m/z 227 ions but did form small amounts of m/z 269 ([M + 91] = [M + C3F3 - 2H]). Comparison of the two spectra reveals that these two structural isomers can be clearly distinguished from each other by using C 3 F 5 + as a reagent ion. The [M + 111] and [M + 91] adducts correspond to successive losses of HF from the [M + 131] adduct. Loss of HC1 and HF from a halocarbon/PAH adduct was demonstrated in Chapter 4, and was proven to involve loss of H 215 originating from the PAH. The m/z 227 ion corresponds to a loss of 82 u from the m/z 309 adduct, which could arise from elimination of a C 2 F 3 H moiety from the [M + C 3 F 5 ] + adduct. Compound 5b also displays evidence of MFF formation, since the relative intensity of the [M + 1]+ ion exceeds the expected 1 3 C molecular ion relative intensity. The effect of analyte concentration, on the formation and distribution of product ions during ion/molecule reactions with C 3 F 5 + , was investigated by analyzing the mass spectra generated along the GC elution profile of compounds 5a and 5b. A relative measure of the increase in analyte concentration is provided by the RIC (Reconstructed Ion Count) values, corresponding to the sum of all ions counts. The results for the main product ions of interest are summarized in Figure 6.4. Large RIC values correspond to high concentration regions at the apex of the GC elution profile. For both 5a and 5b, no significant changes are noted in the relative ion counts (counts/RIC) at higher concentrations (i.e. RIC > 2000). It is only in the lower concentration region (RIC < 2000) that variations can be noted. For 5a, only Ts/f (m/z 178) was formed at low concentrations, possibly indicating a limiting ratio of analyte vs. reagent ions for adduct formation to occur. In contrast, 5b displayed significant adduct formation (m/z 289, m/z 309) at low concentrations with no JvT formation, indicating a preferential reactivity of this compound with C 3 F 5 + . Although there is some variability in the relative intensities of the ions as a function of concentration (Figure 6.4), the ratios of ion counts between different product ions remains relatively constant. For example, for 5b, the average intensity ratio for m/z 309/289 over the concentration range shown in Figure 6.4 is 1.8 ± 0.2 (at a 90 % confidence level). Similarly, for 5a the average intensity ratio for m/z 178/309 ratio is 5.0 ± 0.6 (90 % confidence level). Consequently, isomers can be clearly distinguished based upon their relative ion intensity ratios even when their % RI values vary as a consequence of changes in the concentration of 216 the analyte. It is important to emphasize that these experiments were performed under constant ionization time (IT) and reaction time (RT) conditions. Under these conditions the number of ions produced will vary significantly as the concentration of the analyte changes. In most of the experiments performed in Chapters 4 and 5, the IT and RT were optimized through the use of the ARC function, in order to maintain an optimum ion population in the trap. Under these conditions the changes in the relative intensity of ions along a chromatographic profile were less pronounced. However, because the ARC function is not designed to control Tc;d (the ion/molecule reaction time), it could not be used for the ion/molecule reaction experiments. 217 1 H 0.8 0.6 H o CC co I 0.4 H o 0.2 0 oo meam Anthracene (5a) o -SI • • o m/z 178 a m/z 179 • m/z 269 A m/z 309 i I 1 I " I 1 I 1 0 2000 4000 6000 8000 RIC (all ions) o CC co c 3 O o 0.6H • 0.4 0.2 • 0 Phenanthrene (5b) O • ** * Cj • O A A • • 9 9 I 2000 4000 6000 RIC (all ions) • o I 8000 o m/z 178 A m/z 179 * m/z 227 • m/z 289 A m/z 309 Figure 6.4 Effect of analyte concentration, along a chromatographic elution hand, on the relative abundance of ions produced during ion/molecule reactions with C3F/ (low RIC values represent the lower ends of the chromatographic band while high RIC values represent the upper portion of the chromatographic band. 218 In order to investigate the effects of reaction time between C 3 F 5 + and PAHs 5a and 5b the value of T c i d was varied between 1 and 100 ms. The results from these experiments are summarized in Table 6.3. Increasing the reaction time resulted in a decrease in the relative amounts of m/z 178 (IVT) for both 5a and 5b, although this effect was much more pronounced for the latter. The largest increase in relative abundance was observed for the m/z 309 [M + 131] adducts for both 5a and 5b. Table 6.3 Effects of reaction time between C 3 F 5 + and PAHs 5a and 5b on the relative abundance of ions produced 5a (counts / E counts)a 5b (counts / £ counts) a T c i d b (ms) m/z 178 m/z 269 m/z 309 m/z 178 m/z 227 m/z 289 m/z 309 1 0.90 0.06 0.05 0.40 0.10 0.18 0.14 10 0.83 0.06 0.10 0.25 0.13 0.23 0.34 100 0.79 0.06 0.15 0.14 0.13 0.29 0.41 (a) 2 counts represents the sum of all the product ions shown in the table. (b) T o i d represents the ion/molecule reaction time. The results from the ion/molecule reactions between other mass-isolated FC43 fragment ions and 5a/5b are summarized in Tables 6.4(a) and 6.4(b). The ions undergoing significant adduct formation are m/z 69 (CF 3 +), 93 (C 3 F 3 + ), 114 (C 2 F 4 N + ), 131 (C 3 F 5 + ) and 264 (C 5 F 1 0 N + ) . The m/z 100 and 119 and 414 ions display only ]Vf and M H + formation. For several of these reagent ions distinct differences can be seen in the adduct formations with compounds 5a and 5b. These differences can be used to differentiate these two compounds. The reaction behavior of the m/z 69 and 93 ions is similar to that observed for C H 3 C H F + and other halocarbon ions previously investigated (Chapter 4). That is, 5b reacts to form predominantly the elimination product [M + R addition product. 219 - HF] + and proportionally less of the [M + R] + Table 6.4(a) Mass spectral data (% RT) for ion/molecule reactions between individually mass selected FC43 ions (R+) and anthracene (5a) R+ (m/z) RICa M+ [M+l]+ [M+R]+ [M+R-HF]+ [M+R-2HF]+ cry (69) 1600 100 15 5 5 0 C 3 F 3 + (93) 300 100 20 10 20 0 C 2 F 4 + (100) 2000 100 15 0 0 0 c 2F 4isr f (114) 400 100 10 50 0 0 C 2 F 5 + (119) 400 100 20 0 0 0 C 3 F 5 + (131) 3000 100 15 15 0 10 C 3 F 6 + (150) 20 100 0 0 0 0 c 3 F 6 isr (164) 0 0 0 0 0 0 C 3 F 7 + (169) 0 0 0 0 0 0 (181) 10 100 0 0 0 0 C 5 F 1 0 N + (264) 200 100 15 20 0 0 C 8 F 1 6 N + (414) 30 100 30 0 0 0 (a) Sum off all ion counts for ion/molecule reaction products. 220 Table 6.4(b) Mass spectral data (% RI) for ion/molecule reactions between individually mass selected FC43 ions (R+) and phenanthrene (5b) R+ (m/z) RIC a [M]+ [M+l]+ [M+R]+ [M+R-HF]+ [M+R-2HF]+ C F 3+ (69) 1500 100 30 10 80 10 C 3 F 3+ (93) 400 70 10 0 100 0 C 2 F 4+ (100) 1500 100 10 0 0 0 C 2 F 4 N+ (114) 300 5 0 0 0 0 C 2 F 5+ (119) 450 100 15 5 1 0 C 3 F 5+ (131) 3000 50 15 100 60 0 C 3 F 6+ (150) 20 100 0 0 0 0 CaFgN" (164) 0 0 0 0 0 0 C 3 F 7+ (169) 0 0 0 0 0 0 C ^ (181) 10 100 0 0 0 0 CsFjoN" (264) 100 100 20 0 0 0 CgF^N" (414) 20 100 0 0 0 0 (a) Sum off all ion counts for ion/molecule reaction products. In contrast, 5a, yields significantly less of the [M + R - FTF]+ product but is able to form [M + R] + in greater abundance than 5b. However, when 5b was reacted with C 3 F 5 + (m/z 131) it yielded a greater abundance of [M + C 3 F 5 ] + ions than [M + C 3 F 5 - FTF]+ ions. The experiments described in Chapter 4 have demonstrated, that within isomers groups, the PAHs with higher IE values (e.g. 5b) generally produce greater amounts of the [M + R - H X ] + elimination products (X=C1, F). As previously discussed in Section 4.3.3, PAHs possessing higher IE values undergo adduct formation more readily since their charge exchange reactions are less exothermic. These earlier observations, that the higher IE PAH produce a greater abundance of [M + R -HX] + ions than [M + R] + , were also attributed to the [M + R] + ions undergoing ready elimination. Consequently, the high yield of [M + C 3 F 5 ] + ions observed for 5b is 221 probably a result of the greater stability of this adduct ion. This is consistent with previous results obtained during C H 3 C H F 2 / C I experiments (Chapter 4) when it was observed that a greater relative abundance of [ M + R ] + , with respect to the [ M + R - F£F]+ elimination product, resulted with R = C H 3 C F 2 + than with R = C H 3 C H F + , suggesting a greater stability for [ M + R ] + when R contains a larger number of F atoms. Also, it is interesting to note that the C F 3 + ion produced some [ M + C F 3 ] + adducts, while all the other halomethane ions studied previously (R = CF£ 2F +, CF£ 2C1+, C H C 1 2 + , CC1 3 + ) did not yield any appreciable amounts of [ M + R ] + adducts, further testifying to the greater stability of [ M + R ] + ions with highly fluorinated R groups. The reactivity of the C 3 F 5 + ion was investigated in more detail by performing ion/molecule reaction experiments with P A H s 5a to 9e. The data from these experiments, summarized in Table 6 .5 , demonstrates that isomers in the four P A H groups display significant variations in their mass spectra, thus allowing differentiation from each other. These data also shows that except for 8e and 9e, all other P A H s were able to form [ M + C 3 F 5 ] + adducts ( [ M + 131]). The lack of adducts for 8e and 9e can be attributed to their low IE values. Because of their low IE values these two compounds can readily undergo charge exchange reactions. 222 Table 6.5 Mass spectral data (% RI) for the reaction between C 3 F 5 + (m/z 131) and PAHs 5a to 9e PAH IK (eV) RIC: [M] [M+l] [M+49] [M+91]b [M+lll] b [M+131] 5a 7.45 6500 100 20 0 10 0 15 dio-5a 7.45 2200 100 15 0 7 2 15 5b 7.86 6100 30 20 30 0 50 100 6a 7.41 5900 100 20 0 0 0 25 dio-6a 7.41 5700 100 15 0 0 0 35 6b 7.95 6100 25 5 15 4 7 100 8b 7.43 4400 100 30 0 0 0 40 8c 7.59 4100 80 25 5 0 4 100 8d 7.84 4700 20 10 0 0 0 100 8e 6.97 3000 100 20 0 0 0 0 9a 7.9 1300 40 15 5 0 0 100 9b 7.4 1300 90 35 0 0 0 100 9c 7.41 2400 100 25 0 0 0 60 9d 7.12 2300 100 35 0 0 0 20 9e 6.90 800 100 35 0 0 0 0 (a) Sum of all ion counts at the mass scan where the mass spectrum was selected. (b) For the two perdeuterated PAHs (di0-5a, d]0-6a) HF losses correspond to DF (21 u) Figure 6.5 illustrates the normalized counts for the [M + C 3 F 5 ] + and M^ ions of isomer groups 5a/b to 9a/e plotted as a function of PAH IE. The [M + C 3 F 5 ] + data clearly shows an increase in the relative abundance of [M + C 3 F 5 ] + with the increasing IE of the PAH. This trend is a similar to that demonstrated for the [M + R - HX] + ions (R+ = C i H m X n , 1=1-2, m=0-5, n = 1-3, X=C1, F) in Chapter 4. The major difference is that [M + C 3 F 5 ] + ions are [M + R]+ type ions, not [M + R - HX]+. None of the halocarbon ions tested previously yielded clear correlations between IE and [M + R]+ adduct formation. As discussed earlier, the [M + C 3 F 5 ] + ions appear to be more stable towards elimination than the previously observed [M + R]+ type ions (R+ = CiHmXn, 1=1-2, m=0-5, n = 1-3, X=C1, F). This enhanced stability may account for the observable trend with IE. 223 a) 0.6 H cc 0 .4 H 10 +J C 3 O « 0 .2 0 9e 8e l-A-H-AH 9Ch^H H§H8b 9d 6a -d 1 0 hSH 6a 9a hp 8d 5b 6b — i | l | l | I | I | l | I 6.8 7 7.2 7.4 7.6 7.8 8 b) 0.8 - i o CC co <-« c 3 O CJ 0.6 0 .4 - \ 0.2 0 8e IE (eV) 5a 8b 6a-d 1 0 6a • 9b 8c 9a ^ 5b 8d HCDH 6b hCH n | i | r -i 1 1 1 r 6.8 7 7.2 7.4 7.6 7.8 8 IE (eV) Figure 6.5 Effect of PAH ionization energy (IE) on the formation of (a) [M+C3F5]* and (b) AT from ion/molecule reactions with C V 7 / ions. 224 In contrast to the [M + C3F 5 ] + data, the relative abundance of ions increases with decreasing PAH IE, as illustrated in Figure 6.5(b). A greater abundance of IVF ions for lower IE PAHs is expected as charge exchange becomes a more favored process. A good linear correlation of [M + 131] adduct formation was also obtained when the data for the four MW 228 PAHs 8b to 8e were plotted as a function of benzo interactions (Figure 6.6). The number of benzo interactions represents the number of angular regions present in the compound, as outlined in Figure 6.6. These interactions have been described previously as a mass spectral correlation parameter for PAHs 7 8 . This observed trend suggests that structural features, not just physical parameters such as TE, may also be important factors in the formation of adducts with the C 3F 5+ ions. 225 Figure 6.6 Effect of benzo interactions (*) on the relative abundance of [M+C3F5J products for isomers 8b, 8c, 8d and 8e. 226 6.3.3 Application to Environmental Analysis In order to test the viability of using mass-selected FC43 fragment ions for performing isomer differentiation in environmental samples, a P A H contaminated sediment extract was analyzed by using C3F 5 + (m/z 131) as a reagent ion. Figure 6.7 illustrates the M , [M + 49], [M + 91], [M + 111] and [M + 131] extracted ion chromatograms for the PAHs 5a and 5b along with their monomethyl analogues. Comparison of the ion chromatograms clearly reveals differences that can be used to differentiate these compounds. For example, the monomethyl-5b isomers, indicated by * in the [M + 49] chromatogram can be distinguished from the methyl-5a isomers (remaining peaks). These differences can be enhanced by performing transformations on some of the ion chromatograms, as described in Chapter 5. For example, taking the difference between the normalized [M + 131] and M ion chromatograms results in the phenanthrenes (5b and methyl-5b) plotting positive and the anthracenes (5a and methyl-5a)plotting negative (Figure 6.8). The mass spectra extracted at positions 1 to 4 in Figures 6.7 and 6.8 are illustrated in Figure 6.9. The mass spectra at positions 1 and 2 closely resemble the standard mass spectra of 5b and 5a respectively (Figure 6.3). The mass spectra extracted at positions 3 and 4 contain the same type of adducts as the 5b and 5a spectra respectively, except that the m/z values are mass shifted by 14 u. Consequently, these mass spectra clearly confirm the presence of Cl -5a and Cl -5b in accordance with the results outlined in Figures 6.7 and 6.8. 227 5a m/z 1 7 8 5 b m/z 2 2 7 m/z 2 6 9 m/z 2 8 9 m/z 3 0 9 A 1 2 methy l 5b & 5a M [ M + 4 9 ] [M + 9 1 ] [M + 1 1 1 ] [ M + 1 3 1 ] 3 4 m/z 1 9 2 m/z 2 4 1 m/z 2 8 3 m/z 3 0 3 m/z 3 2 3 i — • — i — i — i — i — i i — i — i — i — i — i — i — i — i — i — i 960 980 1000 1020 1060 1080 1100 1120 1140 1160 mass scan number mass scan number Figure 6.7 Analysis of B-lagoon sediment extract with mass-selected C i F / ions. Extracted M, [M+49], [M+91], [M+lll] and [M+131] ion chromatograms for (left) phenanthrene (5b) and anthracene (5a) and (right) monomethylphenanthrenes and anthracenes. 228 methyl 5b j i methyl 5a 960 1000 1040 1080 1120 1160 mass scan number Figure 6.8 Analysis of B-lagoon sediment extract with mass-selected C3F/ ions. Transformation ion chromatograms for (left) phenanthrene (5b) and anthracene (5a) and (right) monomethyl phenanthrenes and anthracenes. 229 Position 1 (5b) 100- i 8 0 -60-40-20-309 289 178 227 269 -| 1 1 1 1 1 I i I r 160 200 240 280 320 m/z Position 2 (5a) 100n 1 7 8 80 H 60 H 40 H 20 i . ll 309 269 1 1 1 1 1 r - 1 I r 160 200 240 280 320 m/z Position 3 (C1-5b) 100-i 8 0 -60 H 40 H 20-192 241 323 303 1 1 1 1 1 1 1 r—i"—r 160 200 240 280 320 m/z Position 4 (C1-5a) 1 0 0 - 1 9 2 80 H 60-40-20 -\ 323 0 - I | r " r I | i | i r r 160 200 240 280 320 m/z Figure 6.9 Mass spectra from the analysis of B-lagoon sediment extract with mass-selected C J F J + ions. These mass spectra represent data from the mass scans at positions 1 to 4 in the chromatograms displayed in Figures 6.7 and 6.8. 230 6.4 SUMMARY Individual FC43 fragment ions, mass-selected by using an ion trap, can be used as reagents for the differentiation of PAH isomers. The four FC43 ions that proved to be the most suitable for adduct formation with PAHs were C F 3 + , C 3 F 3 + , C2F4N4" and C 3 F 5 + . Ion/molecule reactions between C 3 Fs + and a series of fifteen PAHs demonstrated that a correlation exists between P A H IE and the abundance of M1" and [M + C 3 F 5 ] + ions formed. As demonstrated by the analysis of a contaminated sediment extract, these reagent ions are not suitable to just differentiate unsubstituted P A H isomers standards but can also be applied to the analysis of substituted P A H isomers such as alkyl-PAHs. Furthermore, the results from the experiments performed in this chapter suggest that other fluorocarbons, such as perfluorokerosene (PFK), used frequently as mass calibrants, may provide useful reagent ions for the differentiation of P A H isomers. 231 CHAPTER 7 ANALYSIS OF PAH OXIDATION PRODUCTS IN A COMPLEX ENVIRONMENTAL MATRIX BY USING GAS CHROMATOGRAPHY ION TRAP TANDEM MASS SPECTROMETRY 232 7.1 INTRODUCTION Because many PAHs are potent mutagens or carcinogens 2 2 3 numerous analytical techniques have been developed for determining their presence in environmental samples. However, the toxicity of a P A H contaminated sample does not exclusively depend upon the parent compounds. Air particulates contaminated with PAHs have been found to be more carcinogenic than can be accounted for by their P A H content alone 1 1 3 . This increased toxicity is potentially attributable to PAH oxidation products 3 3 ' 3 4 - 1 1 4 such as the polyaromatic quinones (PAQs) and polyaromatic ketones (PAKs) illustrated in Figure 7.1. Recently, the polyaromatic ketone, 3-nitrobenzanthrone (3-NBK), was discovered in the exhaust fumes of diesel engines 1 1 5 . According to the Ames mutagenicity test 3-nitrobenzanthrone is one of the most mutagenic (and thus potentially highly carcinogenic) compounds ever tested 1 1 5 . Because the toxicological and environmental effects of a P A H can be significantly altered as a result of a chemical transformation, such as oxidation, a full account of the environmental impact of PAHs should also include their transformation products. However, because PAHs are often associated with complex matrixes, such as soils or aquatic sediments, the detection and quantification of trace levels of transformation products is often a difficult task, usually requiring extensive fractionation by liquid chromatography 3 4 This in turn will generally lead to loss of analyte. In order to routinely screen for trace levels of PAH transformation products together with their precursors, new methods that do not require extensive sample manipulation need be developed. Such methods should involve rapid solvent extractions followed by a short chromatographic step to remove highly polar impurities, if gas chromatography is to be used. Anthraquinone 1-Methvlanthraquinone 2-Methylanthraquinone (AQ) (1-MeAQ) (2-MeAQ) Benzanthrone 3-Nitrobenzanthrone Pyrene-2,7-dione ( B K ) (3-BK) (PyQ) Figure 7.1 Polyaromatic ketones and quinones. Capillary GC/MS is usually the analytical technique of choice when analyzing complex mixtures of PAHs, as it combines efficient chromatographic separation with mass spectral information. Selectivity and sensitivity can be achieved by software extraction of the molecular or fragment ion chromatograms for a particular analyte, as long as co-eluting matrix 234 compounds do not interfere at the specific m/z values of the analyte. However, when analyzing for trace level substances, such as the transformation products of PAHs, their molecular or characteristic fragment ions are often obscured throughout the ion chromatogram by abundant matrix ions. Also, standards are often not available for many P A H analogues so that a confirmatory spectrum has to be obtained, a difficult task when other co-eluting interferences are present. To differentiate analytes obscured by co-eluting interferences a GC tandem mass spectrometry method (GC/MS/MS) is required. In the first MS step, molecular ions (or fragment ions) of the appropriate m/z value can be pre-selected, eliminating interfering ions with m/z values different from the analyte. The pre-selected ions can then be fragmented via collision-induced dissociation (CTD) leading to characteristic fragments. If co-eluting interferences of the same pre-selected m/z are also present they will likely lead to different fragment ions. Consequently, identification and quantitation of the target analyte can be performed using its characteristic CTD product ions, either from a single CID experiment (MS/MS) or from multiple CID processes (MSn). Although GC/MS/MS can be performed using multi-sector 1 1 6 , 1 1 7 or triple quadrupole instruments 1 1 8 ' 1 1 9 5 the use of an ion trap affords potentially better sensitivity since it eliminates losses of ions during transfer between different analyzers 1 2 0'. The recent availability of reliable commercial GC Ion Trap MS (GC/ITMS) systems has enabled GC/MS/MS experiments to be readily performed at pg to fg levels 1 2 1 1 2 2 . In addition to improved sensitivity, other advantages of an ion trap include the possibility of performing MS" experiments as well as analyses in the field because of the compact nature of the instrument. Furthermore, the cost of an ion trap is significantly lower than the cost of multi-sector instruments. The work presented in this chapter demonstrates a relatively simple yet powerful technique for the analysis of trace level PAH oxidation products in a complex matrix by 235 performing GC/MS/MS in an ion trap. The ion trap used in this study has two modes of CID excitation, resonant excitation (RCID) and non-resonant excitation (NRCID). Resonant excitation involves the application of a high frequency r.f. potential to the end-caps, corresponding to the oscillation frequency of the selected ion. Non-resonant excitation uses a low frequency dipole square wave which causes simultaneous excitation of all ions in the trap. The amount of energy imparted onto the excited ion will depend upon the amplitude of the CID waveform, its duration, the r.f. storage level, and the nature and pressure of the collision gas. The helium mobile phase from the GC generally serves as the collision gas. The CID amplitude and the r.f. storage level were the two parameters optimized to yield the maximum amounts of the required fragment ions. It is important to note that CID in an ion trap leads to the formation of products via the lowest energy dissociation pathways 6 4 , resulting in non-standard MS/MS spectra, when compared with the higher energy CID MS/MS spectra obtained in triple quadrupole or multi sector instruments. Thus, the MS/MS spectra from ion trap experiments will not necessarily be identical to published reference electron ionization (EI) mass spectra, and the extent of fragmentation will depend upon the CID conditions. Nevertheless, most of the major fragments present in EI spectra should be observed under varying CID conditions via sequential fragmentation on the ion-trap timescale. Anthraquinone (AQ), 2-methylanthraquinone (2-MeAQ), benz[a]anthracene-7,12-dione (B[a]A-Q) and 9-fluoranone (9FK) were selected as the primary target compounds for optimization of the CID method. These compounds may be produced from the oxidation of their respective P A H precursors, substances that are present in high concentration in the contaminated sediment analyzed 8 . 236 7.2 E X P E R I M E N T A L 7.2.1 Reagents Anthraquinone, 2-methylanthraquinone, 9-fluoranone, and benz[a]anthracene-7,12-dione were obtained from Aldrich (Milwaukee, WI, USA). The internal standard di 0-anthracene was obtained from CTL (Woburn, MA, USA). All solvents used were of HPLC grade (Fisher Scientific, Nepean, ON, Canada). 7.2.2 Instrumentation All experiments were performed using a Saturn 4D GC/MS/MS system (Varian, Walnut Creek, CA, USA) equipped with a Wave-Board for the generation of user-defined wave forms applied to the ion-trap electrodes. The software used for general operation of the instrument was Saturn version 5.2. The "Toolkit" version 1.0 software was used for performing sequential MS/MS experiments during the optimization procedure. 7.2.3 Analytical Procedures a) Sample preparation and quantitation The contaminated sediment analyzed was the B-lagoon sample described in previous chapters. A portion of this sample was freeze dried and extracted as described in Chapter 5. A 1 mL aliquot, containing 10 mg of crude extract, was placed on a column containing 5 g of silica, and flushed with acetone / dichloromethane (1:1) to collect all the non-polar (aliphatics, PAHs) and semi polar compounds (heteroaromatics). This solvent extract was used to perform all the analyses discussed in this chapter. 237 A fractionation of the aliphatics, PAHs and heteroaromatics was not attempted in this work in order to determine the viability of performing analyses on a single complex extract. The amounts of anthracene, anthraquinone, 2-methylanthracene and 2-methylanthraquinone in the sediment extract were determined by using the dio-anthracene internal standard and appropriate relative response factors obtained by analyzing a mixture containing these four compounds together with dio-anfhracene under non-resonant CID conditions. Calibration curves were obtained for AQ, anthracene and dio-anthracene to determine the dynamic range of the method. A series of control samples of uncontaminated sediment (as determined by prior analysis) spiked with 20 pg of anthracene or AQ were also extracted to determine the extraction efficiency of AQ and to monitor for any oxidation of anthracene to AQ during the extraction or cleanup processes. b) Gas chromatography GC experimental conditions were the same as described in Chapter 4. c) Ion trap analysis All experiments were performed under automatic gain control (AGC) with a target value of 30,000 for GC/MS and 10,000 for GC/MS/MS. The filament emission current was 40 pAmps and the multiplier voltage was set to give a gain of 105. The r.f. was ramped to produce a scan rate of 5,600 u s"1 over a range of m/z 100 - 300. The MS/MS parameters used are indicated in Figure 7.2. For all CID experiments the storage r.f. (SRF) was maintained to store ions above m/z 80. For the AQ and 2-MeAQ CID experiments the optimum excitation voltage yielding maximum production of fragment ions was determined by a series of multistep experiments in which the CID voltage (resonant or non-resonant) was increased in ten sequential steps. With 238 a scan rate of 0.2 seconds per step a ten step experiment took 2 seconds allowing five experiments to be performed during the 10 second wide chromatographic band containing 50 ng of compound. The data obtained for each ion of interest were normalized as a percent of the total ion current (TIC). Fine Isolation Waveform 4 $ Eject > m CID Waveform ^ 7 — > CID 8 Axial Modulation Waveform Eject < m Scan 1_ 1 Lowest mass ions stored during ionization: 48 m/z 2 Ejection amplitude: 20 volts 3 Parent ion selection (1 u window): - m/z 208 (AQ), 222 (Cl-AQ), 236 (C2-AQ), 258 (B[a]A-Q), 180 (9FK) 4 Broadband amplitude: 30 volts 5 Dwell time prior to CID: 5 ms 6 CID Storage RF (SRF) low mass cutoff: 80 m/z 7 CID time: 20 ms 8 CID supplemental RF: - Non-Resonant: 60 volts (AQ, C l - A Q and C2-AQ) - Resonant: 1.1 volts (AQ), 1.5 volts (Cl -AQ and C2-AQ) 9 Mass scan: 100 - 300 m/z Figure 7.2 Ion trap MS/MS mass-isolation and CID scan function. 240 7.3 R E S U L T S AND DISCUSSION 7.3.1 Optimization of CTJ) Conditions The resonant and non-resonant CTD decomposition curves for the molecular ions of AQ and 2-MeAQ are shown as a function of CTD voltage in Figures 7.3 and 7.4 respectively. The fragmentation pathways for these two compounds are illustrated in Figure 7.5. From the CID decomposition curves optimum voltages were selected to provide maximum mass spectral information concerning characteristic fragmentation of these compounds. Under non-resonant excitation, the CTD voltage selected for both AQ and 2-Me AQ was 60 volts. At this voltage the M-CO (m/z 180), and M-2CO (m/z 152) fragment ions were present in approximately equal proportions. For 2-MeAQ, the M - C H 3 (m/z 207), M-(CH 3 ,CO) (m/z 179), and M-(CH 3,2CO) (m/z 151) ions were present in significantly lower amounts compared to the M - C O (m/z 194), M-2CO (m/z 166) and M-(2CO,H) (m/z 165) ions. For resonant excitation the CID voltages chosen were 1.1 volts for AQ and 1.5 volts for 2-MeAQ. The latter voltage was the smallest value that yielded detectable amounts of the m/z 151 fragment. For all experiments, the CID excitation period was kept at 20 ms; variations from 10 - 40 ms were not found to significantly affect the fragmentation behavior of either AQ or 2-MeAQ. The mass spectra of AQ and 2-MeAQ obtained at the selected CID voltages for both resonant and non-resonant excitation are shown in Figure 7.6. The most striking difference in these spectra is the greater complexity of the fragment ions in the non-resonant spectra (Figures 7.6 (a) and (c)), which is expected since during non-resonant excitation the fragment ions can themselves undergo further fragmentation. Although standards are available for AQ and 2-MeAQ, allowing their identification in unknown samples via mass spectral data and retention times, other alkylated AQs isomers for 241 which standards were not available (i.e. 1-MeAQ and C2-AQs) can only be identified by their mass spectral patterns. Consequently, it is necessary to obtain a good characteristic spectrum with enough fragmentation data to identify the compound. (b) 20 40 60 80 RF Supplemental Voltage 0.8 h I 0.6 •B 0.4 h 0.2 100 - A - m/z 207-208 -O- m/z 180 - A - m/z 152 0 0.4 0.8 1.2 1.6 2 RF Supplemental Voltage Figure 7.3 Decomposition curves for anthraquinone molecular ion using (a) non-resonant excitation: parent mass = m/z 208, CID time 20 ms, CID energy 10 to 100 volts; and (b) resonant excitation: parent mass = m/z 208, CID time 20 ms, CID energy 0.1 to 1.9 volts. Relative intensity values correspond to normalized data (ion count / TIC). 242 (a) 0 20 40 60 80 100 - • - m/z 194 RF Supplemental Voltage m/z 179 m/z 165-166 m/z 151 U i I i I i I i I i I i I i l_ 0 0.4 0.8 1.2 1.6 2 2.4 2.8 RF Supplemental Voltage Figure 7.4 Decomposition curves for 2-methylanthraquinone molecular ion using (a) non-resonant excitation: parent mass = m/z 222, CID time 20 ms, CID energy 0 to 90 volts; and (b) resonant excitation: parent mass = m/z 222, CID time 20 ms, CID energy 0 to 2.7 volts. Relative intensity values correspond to normalized data (ion count / TIC). AQ 2-MeAQ Figure 7.5 Possible fragmentation pathways for AQ and 2-MeAQ. a) AQ, NR, 60 Volts b) AQ, Res, 1.1 Volts 244 100 80-60-40-20-0 180 152 207 100 80 E 60 40 20 i — i — i — i — i — i — i — i — i — r ~ i — i — i — i — i 100 120 140 160 1 80 200 220 240 m/z c) 2-MeAQ, NR, 60 Volts 180 152 o ~ i — i — i — i — i — i — I ' I T i i i | — i r~> 100 1 20 140 1 60 1 80 200 220 240 m/z d) 2-MeAQ, Res, 1.5 Volts 100-80 60-40-20-0 194 166 151 — i i-179 221 207 I — i — i — i— i— i— r - 1 — i — 1 100 120 140 160 180 200 220 240 m/z 100-80 H 60 H 40-20-0 165 151 194 179 [207 1—1—1—1—1—1—1—1—1 1 1—1—p—1—1 100 120 140 160 1 80 200 220 240 m/z Figure 7.6 Mass spectral data for AQ and 2-MeAQ standards during non-resonant (NR) and resonant (Res) CID obtained at the indicated voltages, (a) AQ, NR, 60 Volts, (b) AQ, Res, 1.1 Volts, (c) 2-MeAQ, NR, 60 Volts, (d) 2-MeAQ, Res, 1.5 Volts. 245 5b AQ = Anthraquinone 1-MeAQ, 2-MeAQ = Methylanthraquinones C2-AQ = dimethyl or ethylanthraquinones B[a]A-Q = Benz[a]anthracene-7,12-dione 6b >a 2-MeAQ 9e 14 1 6 18 20 22 24 26 28 30 32 34 t ime (min) Figure 7.7 GC/EI-MS total ion chromatogram of B-lagoon sediment extract, with major PAHs and PAQs identified. 7.3.2 Qualitative Analysis of Contaminated Sediment for AQ The GC/EI-MS total ion chromatogram for the B-lagoon sediment extract is shown in Figure 7.7. The major unsubstituted PAHs present in this sample have been labeled according to the number labels used in this thesis. Some of the more abundant alkyl homologue groups have also been indicated using the C l (monomethyl) and C2 (dimethyl or ethyl) annotations (e.g. Cl-5a,b for monomethylanthracenes/phenanthrenes). Also indicated in this figure are the retention time regions where some of the polyaromatic quinones (PAQs) present in this sample have been identified using GC/MS/MS. It is important to note that the arrows for the PAQs in Figure 7.7 merely indicate the retention time region where these compounds have been identified. They do not imply that the PAQs correspond to a specific signal in this region. In fact, the PAQ signals cannot be observed in the TIC, as their signals are obscured by other more abundant compounds present in the extract. 246 7.3.2. J Analysis for AQ A portion (tR = 18.9 min to 19.8 min ) of the TIC, and three mass chromatograms (m/z = 208, 180 and 152) from the GC/MS analysis of the sediment extract are illustrated in Figure 7.8(a). The retention time window shown corresponds to the elution region of AQ. The various signals observed in the chromatogram during this retention period originate from intact molecular ions and/or fragment ions of different compounds present in the sediment extract. The mass spectrum at the retention time of AQ (19.38 min) is reproduced in Figure 7.6(b). Although ions characteristic for AQ (208, 180, 152) are observable, the spectrum is dominated by a co-eluting interference with m/z 204. By employing GC/MS/MS in non-resonant or resonant modes the m/z 180 and 152 mass chromatograms are "cleaned up" significantly resulting in clear confirmatory spectra for AQ. 247 a) GC/MS m/z TIC m/z 208 M+ m/z 180 [M-CO]+ m/z 152 [M-2CO]+ Res GC/MS / MS i— i— i— i— i— i— i— i— i— i 11—i— i— i— i— i— i— i— i i | i i m i i 1 i r 19 ia2 19.4 196 m 19 ia2 19.4 19.6 19. 19 ia2 1S4 19.6 i a time (rrin) trrelnti) b) cc MS 100-80 60 40 20 H 204 152 180 NR MS/MS 100-cc 208 80 60 40- | 20 -i 0 180 152 M-2C0 i I i I M-CO 207 M-H # Res MS/MS 100-8 0 -60- 1 M-2CO 4 0 -2 0 -o- 1 I 1 I 180 M-CO i i ' i i i ' 100 120 140 160 180 200 220 240 100 120 140 160 180 200 220 240 100 120 140 160 180 200 220 240 m/z m/z m/z Figure 7.8 GC/MS and GC/MS/MS analysis of sediment extract, (a) TIC and m/z 208, 180, 152 ion chromatograms, (b) background subtracted mass spectra at tR = 19.38 min in (a). 248 7.3.2.2 Analysis for Cl-AQs The real advantage of the MS/MS technique is demonstrated in the search for MeAQs (H. = 222) and C2-AQs (M r = 236) in the sediment extract. As illustrated in Figures 7.5 and 7.6, 2-MeAQ yields, in addition to the molecular ion (m/z 222), fragments at m/z 207, 194, 179, 166, 165 and 151. The TIC and mass chromatograms corresponding to these ions, extracted from the GC/MS and GC/MS/MS non-resonant and resonant analyses, are illustrated in Figure 7.9. The mass spectra at tR = 20.89 min and tR = 21.55 min, corresponding to the two major peaks observed in the m/z 166, 165 and 151 GC/MS/MS chromatograms, are reproduced in Figures 7.10(a) to (c). The compound eluting at tR = 21.55 was identified as 2-methylanthraquinone (2-MeAQ) on the basis of the MS/MS spectrum and the retention time match with the 2-MeAQ standard. Because AQ appears to be the only dominant M r = 208 quinone detected in the sediment extract, it is unlikely that other polyaromatic quinones with a mass of 222 u are not AQ analogues. The standard library reference spectra of 1,4-anthraquinone (NIST #48970), another possible anthraquinone isomer, displays an intense M-82 ion due to loss of C O C 2 H 2 C O which is not observed in any of the spectra obtained in this work. The retention time of 9,10-phenanthrenequinone, another structural isomer of AQ, was found to be over three minutes longer than that of AQ. No detectable levels of 9,10-phenanthrenequinone were found in the environmental sample analyzed. Given that only two possible monomethyl isomers of AQ exist, (1-MeAQ and 2-MeAQ), the chromatographic signal at tR = 20.89 min must correspond to 1-methyl AQ (1-MeAQ). A standard for this compound is not available, so it is necessary to examine the full mass spectral data for structural confirmation. GC/MS m/z TIC (\ _ A NR GC/MS/MS _ 7 V v A ^ v A Res GC/MS / MS m/z 222 ^h^V^y^j^ m/z 207 'M|l/V^L^jL m/z 194 Ajy^K^\]A m/z 179 r J l / W _ ^ v _ J V V / j | ^ m/z166JLJn IbiAjV m/z 165 N V N A / ^ A A A m/z 151 A \ _ I I ill I 1-MeAQ l\ 2-MeAQ A 1-MeAQ 1 2-MeAQ i i i i i i i i i i i i i 20.8 21 21.2 21.4 21.6 21.8 t i me (mi n) I 1 I 1 I 1 I 1 I 1 I 1 I 20.8 21 21.2 21.4 21.6 21.8 t i m e ( m i n ) 1 1 1 1 1 1 1 1 1 1 1 1 1 20.8 21 21.2 21.4 21.6 21.8 t i me ( mi n ) Figure 7.9 TIC and m/z 222 (M), 207 (M-CH3), 194 (M-CO), 179 (M-CO,CH3), 166 (M-2CO), 165 (M-2CO,H), and 151 (M-2CO,CH3) ion chromatograms for sediment extract in GC/MS and GC/MS/MS non-resonant (NR) and resonant (Res) modes. 250 a) 20.89 min 21.55 min 1 o o 8 0 H 6 0 H 4 0 2 0 2 0 2 1 7 4 1 0 0 - i 8 0 6 0 -\ 4 0 2 0 165 2 0 2 i I i I — 1 4 0 1 6 0 1 8 0 2 0 0 2 2 0 2 4 0 20.89 min 1 0 0 - i 8 0 6 0 -4 0 -2 0 0 i/z 1 94 2 2 2 1 - M e A Q 1 66 151 1 77 1 I'- LL. 1 9 4 i - n— T 2 2 2 1 4 0 1 6 0 1 8 0 2 0 0 2 2 0 2 4 0 m/z 1 0 0 8 0 -\ 6 0 H 4 0 H 2 0 - i 1 ' 1 4 0 1 6 0 1 8 0 2 0 0 2 2 0 2 4 0 m Iz 20.89 min 21.55 min 2 - M e A Q 1 66 1 94 151 1 79 2 2 2 2 0 7 -t 1 1 1 1 1 ' 1 ' 1 > 1 4 0 1 6 0 1 8 0 2 0 0 2 2 0 2 4 0 m Iz 21.55 min 1 o o -8 0 -6 0 4 0 H 2 0 165 151 1 - M e A Q 1 76 1 9 4 1 1 I 1 I '" I 1 I 1 1 4 0 1 6 0 1 8 0 2 0 0 2 2 0 2 4 0 m/z 1 0 0 -8 0 -6 0 4 0 2 0 H 0 165 2 - M e A Q 151 I 1 9 4 1 79 2 0 7 1 1 1 1 1— ' 1 ' 1— 1 4 0 1 6 0 1 8 0 2 0 0 2 2 0 2 4 0 m/z Figure 7. JO Mass spectral data at 20.89 min and 21.55 min for Cl-AQ analysis using (a) GC/MS, (b) non-resonant GC/MS/MS and (c) resonant GC/MS/MS. 251 Upon examining Figure 7.10(b) it is noticed that ions m/z 207 (M-CH 3) and 179 (M-CO,CH 3 ) in the mass spectrum at t R= 21.55 min (2-MeAQ) are absent in the mass spectrum obtained at tR = 20.89 min (1-MeAQ). This absence of C H 3 fragmentation products in the mass spectrum of 1-MeAQ may be due to a stabilizing interaction between the carbonyl oxygen and the adjacent methyl protons. The mass spectrum of 1-MeAQ also displays m/z 177 ions (M-45) not present in the 2-MeAQ spectrum. A number of small signals at m/z 202, 210 and 211 are observed in the NR mass spectra of 1-MeAQ (Figure 7.10(b)). These ions do not correspond to any logical losses from m/z 222, and they are most likely due to co-eluting interferences. For example, the m/z 202 ion probably corresponds to pyrene (6a, M r 202), since this substance, present in concentrations two orders of magnitude higher than C l - A Q , is found to elute in this region (tR = 20.86 min). It is also important to recognize that during the CTD process, (step 7, Figure 7.2), eluents from the GC column are constantly entering the trap and can interact with the ions which have been isolated. Given that the ionization potential of anthraquinone (9.3 eV) 1 3 , is greater than that of pyrene (7.41 eV) 1 3 , the m/z 202 signal may be M 4 ions of 6a formed via charge transfer reactions with the stored m/z 222 ions. The m/z 210 and 211 ions could arise from ion/molecule reactions between matrix molecules and stored ions. Although the isolation of m/z 222 for the CID experiments was performed with a 1 u window, these data demonstrate that the presence of relatively large amounts of co-eluting substances can still generate some minor mass spectral interferences. Comparison of the mass spectra in Figures 7.6 and 7.10, reveals that non-resonant CID provides more intense characteristic fragment ions. Therefore, the non-resonant CID mass spectra enable a better identification of the molecule undergoing fragmentation. 252 Consequently, for the C2-AQs analyses only data obtained in non-resonant mode are presented. 7.3.2.3 Analysis for C2-AQs The search for C2-AQs (dimethyl or monoethyl) used m/z 236 as a precursor ion. Fragments at m/z 221 (M-CF£3), 208 (M-CO), 207 (M-C 2 H 5 or M-HCO), 193 (M-CO, CH 3), 180 (M-2CO), 179 (M-2CO, H) and 165 (M-2CO, CH 3 ) can be expected based on the fragmentation behaviour of 2-MeAQ. The ion chromatograms corresponding to these ions are illustrated in Figure 7.11. The signals present in the ion chromatograms from the GC/MS data merely correspond to baseline noise level that has been normalized to the scale shown. On the other hand, the m/z 165 ion chromatogram from the GC/MS/MS analysis reveals the presence of two clear signals above the baseline noise. Mass spectra extracted at positions 1 (tR = 23.1 min) and 2 (tR = 23.6 min) are shown in Figure 7.12. These spectra identify the presence of C2-AQ in the sediment extract. Ions of m/z 207 at position 1 could be due to loss of C2H5 from the molecular ion or loss of CO+H because of a close proximity of the carbonyl oxygen(s) and an alkyl group. The three experiments outlined above, designated to detect anthraquinone and its C l and C2 alkylated analogues in the sediment extract, were first carried out individually (i.e. selecting for only one precursor mass for the entire chromatographic run) to determine the appropriate elution windows, and were then followed by a single combined experiment in which three time windows were defined for the three different CID precursor masses (m/z 208, 222, 236). As expected, there were no detectable difference in the results obtained. The ability to perform such multiple CID experiments during a single GC run, either for method 253 development or for analytical purposes, is advantageous when small amounts of sample are available, or if a large number of samples are to be analyzed. G C / M S m/z TIC NR GC / MS / MS m/z 207 m/z 193 m/z 180 m/z 179 m/z 165 1 | I | I | I 1 1 1 I I 1 1 1 1 1 I | I | I | I | 2 2 . 8 23 2 3 . 2 2 3 . 4 2 3 . 6 2 3 . 8 2 4 2 2 . 8 2 3 2 3 . 2 2 3 . 4 2 3 . 6 2 3 . 8 2 4 t i m e ( m i n ) t i m e ( m i n ) Figure 7.11 TIC and m/z 236 (M), 221 (M-CH3), 208 (M-CO), 207 (M-C2H5 or M-CO,H), 193 (M-CO,CH3), 180 (M-2CO), 179 (M-2CO,H), 165 (M-2CO,CH3) ion chromatograms for sediment extract in GC/MS and GC/MS/MS non-resonant analyses. a) 23.1 min 23.6 min 254 100 80 60 40 20 O 215 189 "i r JllL i-llil 100n 80-60 H 40 H 20 0 216 189 " l — i 140 160 180 200 220 240 m/z b) 23.1 min (# 1) 100-80-60-40-20-0 2 3 1 236 1 0 ° -80 A 140 160 180 200 220 240 m/z 23.6 min (#2) 165 208 165 207 180 193 - i — r ^ - i — T — i i i — T — — i 140 160 180 200 220 240 m/z 60-40-20 0-193 221 236 i — | — i — | — i — i — i — r i i— i 140 160 180 200 220 240 m/z Figure 7.12 Mass spectral data at 23.1 min and 23.6 min for analysis of C2-AQs using (a) GC/MS and (b) non-resonant GC/MS/MS. 255 7.3.2.4 Analysis of Other PAH Oxidation Products GC/MS and NRCID GC/MS/MS analyses were also performed for benz[a]anthracene-7,12-dione ([B[a]A-Q, M r 258) and 9-fluoranone (9FK, M r 180). Standards were available for these compounds to establish retention times and develop optimum CTD conditions. 9FK only offers one major fragment for identification (m/z 152, M-CO) while B[a]A-Q offers two fragments (m/z 230, M - C O and m/z 202, M-2CO). The non-resonant CTD voltages required to yield maximum amounts of m/z 202 for B[a]A-Q and m/z 152 for 9FK were 60 and 90 volts respectively. Applying these CTD conditions, during the appropriate retention time windows, these two compounds were targeted in the sediment extract. a) Benz[a]anihracene-7,12-dione analysis The m/z 258, (JVF), 230, ([M - C O f ) and 202 ([M - 2CO]+) ion chromatograms observed within the retention time region of B[a]A-Q are shown in Figure 7.13 for both GC/MS and GC/MS/MS analysis. The GC/MS/MS data enable the detection of two compounds (tR = 28.5 min and tR = 29.6 min) that are almost undetectable by GC/MS. The elution time of the latter compound corresponds with that of standard B[a]A-Q. The mass spectra obtained at these two retention times are shown in Figure 7.14. The GC/MS/MS mass spectrum at tR = 29.6, clearly confirms the presence of B[a]A-Q, by removing the dominant m/z 241 interference observed in the GC/MS mass spectrum. The MS/MS mass spectrum at tR = 28.5 min also displays the presence of m/z 202 (M-2CO), 230 (M-CO) and 257 (M-H) ions (marked with a * in Figure 7.14(b)) corresponding to a M r 258 quinone. However, in this mass spectrum some interferences are observed at m/z 228, 242, and 243. 256 G C / M S m/z TIC m/z 258 M + m/z 230 [M-CO]+ m/z 202 [M-2CO]+ i i i i i i i i i i i i i i i i i i i i i 2 8 . 4 2 8 . 8 2 9 . 2 2 9 . 6 3 0 time (min) G C / M S / M S i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i 3 0 . 4 3 0 . 8 2 8 . 4 2 8 . 8 2 9 . 2 2 9 . 6 3 0 3 0 . 4 3 0 . time (min) Figure 7.13 TIC and m/z 258, 230, 202 ion chromatograms for sediment extract in GC/MS and GC/MS/MS non-resonant modes. CC 0 s 257 a) 28.5 min 100-8 0 -6 0 -40 20 29.6 min 0" 242 231 I II 255 271 100-i 8 0 -6 0 -4 0 -20-1 1 ™ - ! 0 241 202 230 X .L_L T r 258 X ~r—I 180 200 220 240 260 280 180 200 220 240 260 280 m/z m/z b) 28.5 min 29.6 min 100 80 60 H 40 20 H 0 * 202 242 228 J 257 i 1 1 100-80-6 0 -4 0 -20 0 202 230 ' r 258 "i 1 180 200 220 240 260 280 180 200 220 240 260 280 m/z m/z Figure 7.14 Mass spectral data at 28.5 min and 29.6 min for analysis ofMr 258 quinones using (a) GC/MS and (b) non-resonant GC/MS/MS. The source of these interferences is most likely a dimethyl or ethyl substituted P A H with a mass of 256 u, such as a C2-8b isomer. Based upon previous analysis (Chapter 5) it was found that these compounds elute between 29 - 30 min. As discussed previously (analysis of Cl-AQs, Section 7.3.2.2), these interferences may arise from ion/molecule or charge exchange reactions. 258 b) 9-Fluoranone analysis The TIC, m/z 152 and m/z 180 ion chromatograms observed within the retention time region of 9-fluoranone (9FK) are shown in Figure 7.15(a) for both GC/MS and GC/MS/MS analysis. The lack of any significant signals in the m/z 180 chromatogram for the GC/MS/MS non-resonant analysis is a result of the m/z 180 precursor ions undergoing fragmentation to m/z 152. The m/z 152 chromatograms observed in the GC/MS/MS data contains several signals. The chromatographic peak eluting at tR = 15.3 min in the GC/MS/MS analysis was identified as 9FK, based upon the retention time of the 9FK standard. The MS/MS spectrum at this retention time (Figure 7.15(b)) confirms the identity of this compound based upon a single loss of 28 u (m/z 152 = m/z 180 - 28). Furthermore, the same MS/MS spectrum was obtained when pure 9FK was analyzed. The MS/MS spectra for compounds number 1 and 2 in the GC/MS/MS chromatogram (Figure 7.15(a)) do not correspond to 9FK as they contain ion at m/z 176, 163, 165 and 176 (spectra not shown). The mass spectrum obtained at tR = 15.3 min during the GC/MS analysis does display the m/z 152 and m/z 180 ions corresponding to 9FK, but in addition, it contains numerous other ions from co-eluting substances. Although the GC/MS/MS analysis results in a strong enhancement of the 9FK (tR = 15.3 min) signal, when compared to the GC/MS data, it did not eliminate other compounds from appearing in the m/z 152 ion chromatogram to the same extent as GC/MS/MS analysis of compounds containing two keto groups (i.e. AQs, B[a]A-Qs). 259 a) GC / MS G C / M S / M S m/z TIC m/z 180 M + m/z 152 [M-CO] + i 1 i 1 i i i i i i i — i — i — i — i — i — i — i — i — i 15 1 5 . 2 1 5 . 4 1 5 . 6 1 5 . 8 15 1 5 . 2 1 5 . 4 1 5 . 6 1 5 . 8 Time (mini Time (min) b) GC/MS 100-1 8 0 -6 0 -tr 4 0 -2 0 -0 115 1_L 195 180 152 i | r 165 j III.' . i. 1 r GC/MS/MS i 1 i 100n 80-60-40 20 0 152 4 -"i—i i—r i ' i i i i i 100 120 140 160 180 200 220 100 120 140 160 180 200 220 m/z m/z Figure 7.15 (a) TIC and m/z 180, 152 ion chromatograms for sediment extract in GC/MS and GC/MS/MS non-resonant modes, (b) mass spectra at 15.26 min. 260 7.3.3 Quantitation A large number of fragment ions is desirable for compound identification. However, for quantification purposes it is important to remove as many co-eluting interferences as possible. Consequently, it is necessary to optimize the yields of the fragment ions that offer the best selectivity over matix interferences. As clearly demonstrated by the ion chromatograms in Figures 7.8, 7.9 and 7.11, secondary fragmentation products provide the best selectivity over matrix interferences (e.g. the m/z 152 [M - 2CO] ion of AQ, the m/z 151 [M - 2CO, CH 3] and m/z 166 [M - 2CO] ions of MeAQs and the m/z 180 [M - 2CO], and m/z 165 [M - 2CO, CH 3] ions of C2-MeAQs). Therefore, for quantitation purposes the CID conditions should be optimized to yield maximum amounts of these ions. For example, in the case of AQ and 2-MeAQ, a non-resonant CID energy of 80 volts (see Figures 7.3 and 7.4) provides the maximum yield of secondary fragments. Calibration curves for AQ, obtained using the m/z 152 ions were linear (r = 0.99) over a range of 1 pg to 14000 pg for both resonant and non-resonant CID analyses (Figure 7.16). The lowest detectable amount of AQ by GC/MS/MS was approximately 0.5 - 1.5 pg using the m/z 152 ion for quantitation. The extraction efficiency of AQ determined using a spiked uncontaminated sediment was approximately (70-130 %). This value is higher than the recovery of anthracene or dio-anthracene (50-70 %). This large difference can be partially explained by the fact that acetone was employed in the extraction process, thus facilitating the extraction of the more polar components. Furthermore, it was found that approximately 6 % of the anthracene in a spiked sample was converted to AQ. Injections of pure anthracene revealed that oxidation to AQ did 261 not occur in the injector. Consequently, some oxidation must have taken place during the extraction or cleanup procedures. <0 CU 4 0 0 0 I i | i i i | i i i 8 0 0 0 1 2 0 0 0 pg AQ Figure 7.16 Response curves for analysis of AQ by non-resonant (NR) and resonant (R) GC/MS/MS. 800 .1 600 A T3 CD W >-•a c 400 H E £ 200 0 (0 LO a < LO CM a < a> E I CM < a • < JO m Substance Figure 7.17 Concentration of quinones and their PAH analogues present in the sediment sample analyzed. 262 These results indicate the importance of monitoring for transformation products as part of an analytical protocol to determine if low extraction efficiencies are due to adherence to the matrix or to actual loss of analyte through chemical transformation. The importance of the transformation of PAHs during extraction has been discussed by Soheila 1 2 3 who identified the presence of bianthracene as a polymerization product of anthracene (5a). The concentrations of 5a, 2-Me-5a, AQ and 2-MeAQ and B[a]A-Q in the freeze dried sediment were determined by using the internal standard anthracene-dio m/z 188 molecular ion chromatogram and the M-2CO ion chromatograms for the quinones, together with the appropriate relative response factors. For the anthraquinone and 2-methylanthraquinone, a correction factor corresponding to 6 % of the non-oxidized compound, (anthracene or 2-methylanthracene), was employed to account for potential oxidation during the sample preparation. The results for the concentrations of polyaromatic quinones, together with their P A H analogues, are summarized in Figure 7.17. For all three quinone groups the concentrations of the non-oxidized P A H analogue exceeded the concentrations of quinone by approximately one order of magnitude. 263 7.4 S U M M A R Y The data obtained in this study demonstrate the usefulness of ion trap GC/MS/MS experiments for detecting and quantifying trace amounts of polyaromatic quinones which would ordinarily be obscured by matrix interferences when using standard GC/MS techniques. Improvements are especially notable when secondary fragmentation products, such as M -2CO, are monitored. Non-resonant CID is preferred when characteristic spectra are required for confirmation, because it provides intense fragment ions together with molecular ion information. CHAPTER 8 CONCLUSIONS 265 The differentiation of P A H isomers in environmental sample extracts is an important but not trivial analytical task. The research summarized in this thesis has demonstrated that differentiation of PAHs within several isomer groups is possible by using gas chromatography coupled with chemical ionization mass spectrometry. The reagents employed for chemical ionization yield ions that can react with P A H molecules to form characteristic adducts. Through the formation of different adduct ions, or varying abundance of the same adduct ions, isomeric PAHs can yield distinct mass spectral fingerprints. The use of CO2 and N 2 0 as reagent gases under negative chemical ionization conditions enabled differentiation of several PAH isomers, including some alkylated positional isomers of benz[a]anthracene (8b). Of particular interest is the ability to differentiate 7,12-dimethylbenz[a] anthracene, one of the most potent P A H carcinogens, from other dimethylbenz[a] anthracenes by using N2O/NCI. However, the yields of adduct ions display significant variability, and depend in particular upon the method of sample introduction. Isomer differentiation of unsubstituted and methylated PAHs present in an environmental sample extract was achieved by comparing GC/MS chromatograms obtained under negative CI conditions to those obtained under positive EI conditions. Positive CI using dimethyl ether (DME) as a reagent gas in a linear quadrupole ion source enabled the differentiation of several P A H isomers, through the formation of [M+C 2H 50] + adducts. However, when using an ion trap mass spectrometer the same experiments did not yield [M+C 2H 50] + adducts. These experiments clearly demonstrated that the use of a particular chemical ionization technique may yield significantly different results depending on the type of mass spectrometer employed. The use of halogenated hydrocarbons (halomethanes and haloethanes) as chemical ionization reagent gases has enabled the development of novel techniques for the 266 differentiation of P A H isomers using an ion trap mass spectrometer 1 0 7 ' 1 2 4 Several P A H isomers could be differentiated through the formation of adducts of the type [M+R]+ and [M+R-HX]+ (M = P A H molecule, R = Reagent gas ion, X = Cl or F). Furthermore, with mass chromatogram data extraction and transformation methods, it was possible to differentiate several groups of P A H isomers present in complex environmental mixtures. Of particular interest was the ability to differentiate between methylbenzo[a]pyrenes and methylbenzo[e]pyrenes using C H 3 C H F 2 CI, since the former PAHs are more potent carcinogens than the latter. FC43, a fluorocarbon that yields numerous fragment ions upon ionization, can also be used to achieve P A H isomer differentiation, but only when individual fragment ions are mass-selected and allowed to undergo ion/molecule reactions with PAH molecules. Although the analysis for standard PAH compounds (unsubstituted and alkyl substituted) is a necessary first step in an environmental investigation, it is also important to analyze for P A H transformation products, such as polyaromatic quinones (PAQs) . Such compounds may serve as indicators for transformation reactions of PAHs during their residence in the environment. GC/MS/MS using an ion trap has been demonstrated to be a powerful analytical technique for the detection and identification of PAQs 1 2 5 . By monitoring secondary fragment ions (e.g. M-2CO) of a target analyte (M), it was possible to almost completely filter out the chemical noise from the matrix. This enabled the detection of PAQs that were previously undetectable via standard GC/MS methods. The results obtained from the research conducted for this thesis expand the array of mass spectrometric techniques available for the analysis of PAHs. Although this thesis has stressed the toxicological risk posed by PAHs as a motivation for their analysis, it is by no means the only reason to study these compounds. A detailed analysis of the P A H distribution 267 in a sample could also be of potential use for detailed source fingerprinting, such as determining the origin of an oil spill. In addition, as PAHs are byproducts of incomplete combustion, a detailed P A H fingerprint analysis could potentially be of use for arson investigations. In addition, the ability to detect certain PAH transformation products readily, by using GC/MS/MS, demonstrates that it is possible to monitor for minor components in a mixture. Such analyses may be of importance in monitoring the degradation of PAHs, or other environmental pollutants, in contaminated sites. 268 REFERENCES 1. Molina, M . J. and Rowland, F. S. Nature 1974, 249, 810. 2. Gelboin, H . V. and Tso, P. O. Polycyclic Hydrocarbons and Cancer; Academic Press: London, 1981; Vol. 3. 3. Polycyclic Aromatic Hydrocarbons in the Aquatic Environment: Formation, Sources, Fate and Effects on Aquatic Systems; NRCC: Ottawa, 1983; Vol. Publication No. NRCC 18981. 4. Bohme, D. K. Chem. Rev. 1992, 92, 1497. 5. Szczepanski, J. and Vala, M . Nature 1993, 363, 699. 6. McKay, D. S., Gibson Jr., E . K . , Thomas-Keptra, K. L . , Vali, H . , Romanek, C. S., Clemett, S. J . , Chiller, X. D. F . , Maechling, C. R. and Zare, R. N. Science 1996, 273, 924. 7. Simpson, C. D . , Cullen, W. R., Quinlan, K. B. and Reimer, K. J. Chemosphere 1995, 31, 4143. 8. Simpson, C. D. , Mosi, A. A . , Cullen, W. R. and Reimer, K. J. Science Tot. Environ. 1996, 181, 265. 9. Cooke, M . , Loening, K. and Merritt, J. (Eds.) Polynuclear Aromatic Hydrocarbons: Measurements, Means and Metabolism; Battelle Press: Gaithersburg, Md. , 1991. 10. Bjorseth, A. (Ed.) Handbook of Polycyclic Aromatic Hydrocarbons; Marcel Dekker: New York, 1983. 11. Vo-Dinh, T. (Ed.) Chemical Analysis of Polycyclic Aromatic Compounds; John Wiley & Sons: New York, 1989; Vol. 101. 12. Budavarvi, S. (Ed.) The Merck Index; 12th ed.; Merck & Co.: Whitehouse Station, New Jersey, 1996. 13. Lias, S. G . , Bartmey, J. E . , Liebman, J. F . , Holmes, J. L . , Levin, R. D. and Mallard, W. G. J. Phys. Chem. Ref. Data Suppl. 1988, 17. 14. Lide, D. R. (Ed.) CRC Handbook of Chemistry and Physics; 73rd ed.; CRC Press: Boca Raton, Florida, 1992. 15. Onuska, F. I. Analysis of Polycyclic Aromatic Hydrocarbons in Environmental Samples; CRC Press: Boca Raton, Florida, 1989. 269 16. Houser, J. J. Huckel Version 1.1; Department of Chemistry, University of Akron, Akron, Ohio, 1991. 17. Lee, M . L . , Novotny, M . V. and Bartle, K. D. Analytical Chemistry of Polycyclic Aromatic Hydrocarbons; Academic Press: New York, 1981. 18. Solomons, T. W. G. Organic Chemistry; 4th ed.; John Wiley & Sons: New York, 1988. 19. Boekelheide, P. J. Am. Chem. Soc. 1967, 89, 1695. 20. Dewar, M . J. S. J. Am. Chem. Soc. 1952, 74, 3357. 21. Isaacs, N. S. Physical Organic Chemistry; John Wiley & Sons: New York, 1987. 22. Berkow, R. and Fletcher, A. J. The Merck Manual of Diagnosis and Therapy; 16th ed.; Merck and Co.: Rahway, N.J, 1992. 23. Harvey, R. G. Polycyclic Aromatic Hydrocarbons: Chemistry and Cacinogenesis; Cambridge University Press: Cambridge, England, 1991. 24. Conney, A. H . Cancer Res. 1982, 42, 4875. 25. Flesher, J. W. , Stansbury, K. H . and Sydnor, K. L . Cancer Lett. 1982, 16, 91. 26. Flesher, J. W. , Myers, S. R. and Lovelace, G. E. Bio-Transformation of Benzofajpyrene to 6-methylbenzofaJpyrene and Related Metabolites in Rat Liver and Lung; 8th International Symposium of Polynuclear Aromatic Hydrocarbons; Battelle Press: Columbus, Ohio, 1983. 27. Flesher, J. W. , Myers, S. R. and Blake, J. W. Bioalkylation of Polynuclear Aromatic Hydrocarbons: A Predictor of Carcinogenic Activity; 9th International Symposium of Polynuclear Aromatic Hydrocarbons; Battelle Press: Columbus, Ohio, 1984. 28. Flesher, J. W. , Myers, S. R. and Blake, J. W. Bioaklyation of Polynuclear Aromatic Hydrocarbons in vivo: A Predictor of Carcinogenic Activity; 10th International Symposium of Polynuclear Aromatic Hydrocarbons; Battelle Press: Columbus, Ohio, 1985. 29. Flesher, J. W. , Myers, S. R. and Blake, J. W. Cancer Lett. 1984, 24, 335. 30. Flesher, J. W. and Myers, S. R. (Eds.) Polynuclear Aromatic Hydrocarbons: Measurements, Means, and Metabolism; Battelle Press: Gaithersburg, Md. , 1987, pp 271. 31. Cavalieri, E . and Rogan, E. EHP Environ. Health Perspect. 1985, 64, 69. 32. Richard, A. M . and Woo, Y. A. Mutation Res. 1990, 242, 285. 270 33. Salamone, M . F . , Heddle, J. A. and Katz, M . Environ. Int. 1979, 2, 32. 34. Ramdahl, T. Environ. Sci. Technol. 1983, 17, 666. 35. Mohammad, S. N. Metabolic Activation and Carcinogenecity of PAHs: a New Quantum Mechanical Theory; 9th International Symposium of Polynuclear Aromatic Hydrocarbons; Battelle Press: Columbus Ohio, 1984. 36. Miller, E . C. Cancer Res. 1978, 38, 1479. 37. Kristen, J. H . , Hansen, B. N . , Cravens, E . and Sievers, R. E. Anal. Chem. 1995, 67, 3541. 38. Richer, B. E . , Jones, B. A . , Ezzell, J. L . , Porter, N. L . , Avdalovic, N. and Pohl, C. Anal. Chem. 1996, 68, 1033. 39. Chai, M . and Pawliszyn, J. Environ. Sci. Technol. 1995, 29, 693. 40. Zhouyao, Z. and Pawliszyn, J. Anal. Chem. 1993, 65, 1843. 41. Garcia, A. L . , Gonzales, E . B. , Alonso, G. J. I. and Sanz-Mendel, A. Anal. Chim. Acta 1992, 264, 241. 42. Quilliam, M . A . , Hardstaff, W. R., Anacleto, J. F . , LeBlanc, M . D. , Stergiopoulos, V . , Dick, K. L . , Bowser, M . T . , Curtis, J. M . , Embree, D. J . , Sim, P. G. and Boyd, R. K. Fresenius J. Anal. Chem. 1994, 350, 109. 43. Burford, M . D. , Hawthorne, S. B. and Miller, D. J. Anal. Chem. 1993, 65, 1497. 44. Pace, C. M . and Betowski, L . D. J. Am. Soc. Mass Spectrom. 1995, 6, 597. 45. Blau, K. and King, G. (Eds.) Handbook of Derivatives for Chromatography; Heyden: London, 1978. 46. Grob, K. Classical Split and Split less Injection in Capillary GC; 2nd ed.; Huethig Verlag: Heidelberg, 1988. 47. Lee, M . L . , Kong, R. C , Woolley, C. L . and Bradshaw, J. S. J.Chromatogr. Sci. 1984, 22, 136. 48. Wright, B. W. , Paenden, P. A . , Lee, M . L . and Stark, T. J. J. Chromatogr. 1981, 213, 211. 49. Hyver, K. J. High Resolution Gas Chromatography; Hewlett-Packard Co.; 1989. 50. Thompson, J. J. Phil. Mag. 1907, 13, 561. 271 51. Burlingame, A. L . , Millington, D. S., Norwood, D. L . and Russell, D. H . Anal. Chem. 1990, 62, 268R. 52. Mosi, A. A. and Eigendorf, G. K. Curr. Org. Chem. 1998, 2, 15. 53. McDowell, C. A. Mass Spectrometry; McGraw-Hill: New York, 1963. 54. Roboz, J. Introduction to Mass Spectrometry; Interscience: New York, 1968. 55. Munson, M . S. B. and Field, F. H . J. Am. Chem. Soc. 1966, 88, 2621. 56. Harrison, A. G. Chemical Ionization Mass Spectrometry; 2nd ed.; CRC Press: London, 1992. 57. Karas, M . , Bachmann, D. and Hillenkamp, F. Anal. Chem. 1985, 10, 995. 58. Karas, M . , Bahr, U. and Hillenkamp, F. Int. J. Mass Spectrom. Ion Processes 1989, 60, 11 A. 59. Beckey, H. D. Z. Naturforsch. 1959, 14a, 712. 60. Beckey, H . D. Z. Naturforsch. 1962, 17a, 1103. 61. Whitehouse, C. M . , Dreyer, R. N . , Yamashita, M . and Fenn, J. B. Anal. Chem. 1985, 57, 675. 62. March, R. E . and Hughes, R. J. Quadrupole Storage Mass Spectrometry; John Wiley & Sons: New York, 1989; Vol. 102. 63. March, R. E . / . Mass Spectrom. 1997, 32, 351. 64. McLuckey, S. A . , Glish, G. L . and Van Berkler, G. J. Int. J. Mass Spectrom. Ion Processes 1991, 106, 213. 65. Brodbelt, J . , Liou, C. C. and Donovan, T. Anal. Chem. 1991, 63, 1205. 66. Sharp, T. R., Lee, H . , Ferguson, A. and Marsh, K. N. (Eds.) Electron Impact Mass Spectra of Polycyclic Aromatic Hydrocarbons: A Reference Collection; ASMS Conference Proceedings: San Francisco, C A , 1988, pp 278. 67. Bazdikian, J . , March, R. E . and Mendez, A. (Eds.) Positive Ion CI of PAHs and Related Compounds; ASMS Conference Proceedings: San Francisco, California, 1993, pp 989a. 68. Keough, T. Anal. Chem. 1982, 54, 2540. 272 69. Burrows, E . P. J. Mass Spectrom. 1995, 30, 312. 70. Pointet, K . , Milliet, A . , Renou-Gonnord, M . F. and Fleurat-Lessard, P. Eur. Mass Spectrom. 1997, 3, 281. 71. Milliet, A . , Renou-Gonnord, M . F. and Pointet, K. Anal. Environ. Chem. 1996, 322, 347. 72. Simonsick Jr, W. J. and Hites, R. A. Anal. Chem. 1984, 56, 2749. 73. Sim, P. G. and Elson, C. M . Rap. Comm. Mass Spectrom. 1988, 2, 137. 74. Elson, C. M . and Sim, P. G. Rap. Comm. Mass Spectrom. 1990, 4, 37. 75. Stemmler, E . A. and Buchanan, M . V. Org. Mass Spectrom. 1989, 24, 705. 76. Stemmler, E . A. and Buchanan, M . V. Org. Mass Spectrom. 1989, 24, 94. 77. Shushan, B. , Safe, S. H. and Boyd, R. K. Anal. Chem. 1979, 51, 156. 78. Shushan, B. and Boyd, R. K. Org. Mass Spectrom. 1980, 15, 445. 79. Dass, C. Org. Mass Spectrom. 1993, 28, 470. 80. Wang, X . , Becker, H . , Hopkinson, A. C , March, R. E . , Scott, L . T. and Bohme, D. K. Int. J. Mass Spectrom. Ion Processes 1997, 161, 69. 81. Sim, P. G . , Jamieson, W. D. and Boyd, R. K. Org. Mass Spectrom. 1989, 24, 327. 82. Pyle, S. M . , Betowski, L . D. , Marcus, A. B. , Winnik, W. and Brittain, R. D. / . Am. Soc. Mass Spectrom. 1996, 8, 183. 83. Futrell, J. H . (Ed.) Gaseous Ion Chemistry and Mass Spectrometry; John Wiley & Sons: Toronto, 1986, pp 61-93. 84. Buchanan, M . V. and Olerich, G. Org. Mass Spectrom. 1984, 19, 487. 85. Stemmler, E . A. and Buchanan, M . V. Rap. Comm. Mass Spectrom. 1988, 2, 184. 86. Stemmler, E . A. and Buchanan, M . V. / . Mass Spectrom. 1993, 28, 953. 87. Stemmler, E . A. Int. J. Mass Spectrom. Ion Processes 1995, 142, 111. 88. Oehme, M . Chemosphere 1985, 14, 1285. 89. Oehme, M . Anal. Chem. 1983, 55, 2290. 273 90. Ayrton, A. D . , McFarlane, M . , Walker, R., Neville, S., Coombs, M . M . and Ioannides, C. Toxicology 1990, 60, 173. 91. Harrison, A. G. / . Chem. Phys. 1993, 90, 1411. 92. Pyatnitski, Y. I. Russ. Chem. Rev. (Engl. Transl.) 1976, 45, 762. 93. Wislocki, P. G . , Fiorentini, K. M . , Fu, P. P., Yang, S. K. and Lu, A. Y. H . Carcinogenesis 1982, 3, 215. 94. Streitwiesser, A. Molecular Orbital Theory for Organic Chemists; John Wiley and Sons: New York, 1961. 95. Catinella, S., Traldi, P., Jiang, X . , Londry, F. A . , Morrison, R. J. S., March, R. E . , Gregoire, S., Mathurin, J. and Tabet, J. Rap. Comm. Mass Spectrom. 1995, 9, 1302. 96. Miller, D. L . , Lay, J. O. and Gross, M . L . J. Chem. Soc. Chem. Commun. 1982, 970. 97. Lane, D. C. and McGuire, M . Org. Mass Spectrom. 1983, 18, 494. 98. Tsuji, M . , Aizawa, M . and Nishimura, Y. Chem. Lett. 1995, 211. 99. Tsuji, M . , Aizawa, M . , Ujita, H . and Nishimura, Y. Bull. Chem. Soc. Jap. 1995, 68, 2385. 100. Tsuji, M . , Aizawa, M . and Nishimura, Y. J. Mass Spectrom. Soc. Jpn. 1995, 43, 109. 101. Tsuji, M . , Aizawa, M . and Nishimura, Y. Bull. Chem. Soc. Jpn. 1996, 69, 147. 102. Aizawa, M . , Tsuji, M . and Nishimura, Y. Bull. Chem. Soc. Jpn. 1996, 69, 1363. 103. Olah, G. A. and Mo, Y. K. J. Am. Chem. Soc. 1972, 94, 9241. 104. Olah, G. A . , Kobayashi, S. and Tashiro, M . Chem. Rev. 1970, 70, 295. 105. Cheng, Y. W. and Dunbar, R. C. / . Phys. Chem. 1995, 99, 10802. 106. Trouk-Pointet, K . , Milliet, A. and Renou-Gonnard, M . F. J. Mass Spectrom. 1995, 30, 1495. 107. Mosi, A. A . , Cullen, W. R. and Eigendorf, G. K. J. Mass Spectrom. 1997, 32, 864. 108. Roussel, R., Allaire, M . and Friar, R. S. J. Air Waste Manage. Assoc. 1992, 42, 1609. 274 109. Yunker, M . B., MacDonald, R. W., Cretney, W. J . , Fowler, B. R. and McLaughlin, F. A. Geochim. Cosmochim. Acta 1993, 57, 3041. 110. Grimmer, G . , Jacob, J . , Dettbarn, G. and Naujack, K. W. Fresenius Z. Anal. Chem. 1985, 322, 595. 111. Venkatesan, M . I. Marine Chemistry 1988, 25, 1. 112. Riley, C. M . and Rosanske, T. W. (Eds.) Development and Validation of Analytical Methods; Elsevier: New York, 1996; Vol. 3. 113. Pitts, N. J . , Grosjean, D. , Mischke, T. M . , Simmon, V. F. and Poole, D. Toxicol. Lett. 1977, 1, 65. 114. Pierce, R. C. and Katz, M . Environ. Sci. Technol. 1976, 10, 45. 115. Enya, T. E . , Suzuki, H . , Watanabe, T . , Hirayama, T. and Hisamatsu, Y. Environ. Sci. Technol. 1997, 31, 2772. 116. Slayback, J. R. B. and Taylor, P. A. Spectra 1983, 9, 18. 117. D'Agostino, P. A. and Provost, L . R. / . Chromatogr. A 1995, 695, 65. 118. Lindstrom, K. and Shubert, R. J. High Res. Chrom. Comm. 1984, 7, 68. 119. Hofmann, U . , Fromm, M . F . , Johnson, S. and Mikus, G. J. Chromatogr. B 1995, 663, 59. 120. Johnson, J. V . , Yost, R. A . , Kelley, P. E . and Bradford, D. C. Anal. Chem. 1990, 62, 2162. 121. Schachterle, S., Brittain, R. D. and Mills, J. D. J. Chromatogr. A 1994, 683, 185. 122. Plomley, J. B. , Koester, C. J. and March, R. E . Anal. Chem. 1994, 66, 4437. 123. Soheila, K. L . , Pickard, M . A. and Gray, M . R. Environ. Sci. Technol. 1996, 30, 1145. 124. Mosi, A. A . , Cullen, W. R. and Eigendorf, G. K. J. Mass Spectrom. 1998, 33, 250. 125. Mosi, A. A . , Reimer, K. J. and Eigendorf, G. K. Talanta 1997, 44, 985. 126. Daves, Acc. Chem. Res. 1979, 12, 359. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0061532/manifest

Comment

Related Items