CHARACTERIZATION OF SYNTHETIC PHENETHYLAMINES USING LOWRESOLUTION AND HIGH-RESOLUTION MASS SPECTROMETRY By Alexandria Lynn Anstett A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Forensic Science – Master of Science 2017 ABSTRACT CHARACTERIZATION OF SYNTHETIC PHENETHYLAMINES USING LOWRESOLUTION AND HIGH-RESOLUTION MASS SPECTROMETRY By Alexandria Lynn Anstett Definitive identification and differentiation of synthetic designer drugs can be challenging for forensic analysts due to the high structural similarities. The focus in this work was the characterization of synthetic phenethylamines, a common class of designer drugs, using mass spectrometry methods. A set of phenethylamine reference standards was analyzed using both low-resolution and high-resolution mass spectrometry and the mass spectra were probed to identify characteristic and distinguishing features. These features were integrated into a flowchart style characterization scheme for both low-resolution and high-resolution mass spectra. The characterization scheme for low-resolution data utilizes retention index and neutral losses to indicate phenethylamine structural subclass. Further, isotope patterns and characteristic mass spectral features give a preliminary indication of the identity of substituent. This scheme is immediately implementable into forensic practice because it exploits the instrumentation already used for the identification of controlled substances. The high-resolution version of the characterization scheme offers more robust characterization. From high-resolution mass analysis, exact mass and mass accuracy of each ion were determined and mass defect filters were developed. These mass defect filters were successful in characterizing compounds according to structural subclass. Overall, this research provides tools for the characterization of synthetic phenethylamines and highlights the potential for high-resolution mass spectrometry for forensic applications, should this instrumentation become available in forensic laboratories. ACKNOWLEDGMENTS Foremost, I would like to express my deepest appreciation, gratitude, and thanks to my advisor, Dr. Ruth Smith for her guidance and willingness to share her knowledge, expertise, excitement, and laughter throughout my graduate career at Michigan State University. I thank her for challenging me to grow as a scientist and as a person. My gratitude knowns no bounds and without her, this would not be possible. I would also like to thank my committee member Dr. Victoria McGuffin, for her advice throughout this research and for always offering a different perspective and asking questions that have challenged me and helped me to think critically. I would also like to thank Dr. Steve Dow for serving on my committee on shorter notice and agreeing to read my thesis over the Christmas holiday. Further, I would like to thank those who helped facilitate this research, especially Scott Smith and the staff of the MSU Mass Spectrometry and Metabolomics Core Facility, and David Alonso and Joe Binkley from LECO Corp. for help with instrumentation and data collection. I am also grateful to the National Institute of Justice who supported this research via grant number 2015-IJ-CX-K008. Points of view in this thesis are those of the author and do not necessarily represent the official position or policies of the U.S. Department of Justice. Additionally, I would like to thank current and past members of the Forensic Chemistry group for their encouragement, guidance, patience, and support- especially sitting through countless hours of AAFS and thesis defense practices. A special thank you to Fanny Chu, Natasha Eklund, Cindy Kaeser, Amanda Setser, Barb Fallon, and Kristen Reese. An extra special thank you to Trevor Curtis for being my “partner in crime” and making sure I always had a friend to eat with. I would also like to thank my dearest friend and roommate Brianna Bermudez iii for taking Moose and I in, and always offering encouragement, advice, laughs, food, and love. Further, to my “favorite” biologist, I would like to thank Alyssa Badgley for being there for me as a best friend and shoulder to lean on throughout my entire graduate school journey. I honestly couldn’t have done it without you, and wouldn’t have wanted too anyway. Thanks for being the best “trace” partner, tailgater, and overall Spartan enthusiast I could have ever asked for (“GREEEN”). Finally, I would like to thank my friends and family, especially my “moms” for their support from across the country, Tristan Musser for his endless love, support, and patience and my parents, Monica and Paul, for their unwavering, unconditional, encouragement and love in everything I’ve ever done. This one’s for you guys. I am truly grateful to you all and couldn’t have made it this far without you. iv TABLE OF CONTENTS LIST OF TABLES ........................................................................................................................ vii LIST OF FIGURES ..................................................................................................................... viii I. Introduction ................................................................................................................................. 1 1.1 Synthetic Designer Drugs...................................................................................................... 1 1.2 Current Methods of Analysis of Submitted Drug Samples and Limitations ........................ 3 1.3 Current Research of Synthetic Designer Drugs .................................................................... 4 1.4 Research Objectives and Goals ............................................................................................. 8 REFERENCES ............................................................................................................................. 12 II. Theory ...................................................................................................................................... 15 2.1 Chromatography Overview ................................................................................................. 15 2.2 Gas Chromatography Overview .......................................................................................... 15 2.2.1 Retention Index............................................................................................................. 19 2.3 Mass Spectrometry Overview ............................................................................................. 20 2.3.1 Mass Analysis: Single Quadrupole Mass Analyzer ..................................................... 21 2.3.2 Mass Analysis: Time-of-Flight Mass Analyzer............................................................ 24 2.3.3 Comparison of Low-Resolution and High-Resolution Mass Spectra .......................... 26 2.4 Mass Defect ......................................................................................................................... 28 2.4.1 Kendrick Mass Defect .................................................................................................. 29 REFERENCES ............................................................................................................................. 30 III. Materials and Methods ............................................................................................................ 32 3.1 Reference Standards ............................................................................................................ 32 3.2 Gas Chromatography-Mass Spectrometry (GC-MS) Analysis ........................................... 35 3.3 Data Processing ................................................................................................................... 37 3.4 Mass Defect Filters.............................................................................................................. 38 3.4.1 Absolute Mass Defect Filters ....................................................................................... 38 3.4.2 Kendrick Mass Defect Filters ....................................................................................... 39 APPENDIX ................................................................................................................................... 41 IV. Characterization of Synthetic Phenethylamines by Low-Resolution Mass Spectrometry ..... 43 4.1 Retention Index ................................................................................................................... 43 4.2 Electron Ionization Mass Spectra of Synthetic Phenethylamine Subclasses ...................... 45 4.3 Neutral Losses from Molecular Ion to Distinguish 2C- from NBOMe-Phenethylamines.. 49 4.4 Distinction and Identification of Common Substituents for 2C- and NBOMePhenethylamines........................................................................................................................ 53 4.4.1 Halogen Substitutions ................................................................................................... 53 4.4.2 Sulfur and Nitro Substitutions ...................................................................................... 59 4.5 Scheme for Characterization of Synthetic Phenethylamines using Low-Resolution Mass Spectra ....................................................................................................................................... 62 4.6 Summary ............................................................................................................................. 70 v APPENDIX ................................................................................................................................... 71 REFERENCES ............................................................................................................................. 78 V. Characterization of Synthetic Phenethylamines by High-Resolution Mass Spectrometry ...... 80 5.1 Comparison of Low- and High-Resolution Mass Spectra .................................................. 80 5.2 Development of Mass Defect Filters ................................................................................... 84 5.2.1 Absolute Mass Defect Filters for Phenethylamines Based on Molecular Ions ............ 84 5.2.2 Absolute Mass Defect Filter for the APB-Phenethylamine Subclass........................... 87 5.2.3 Absolute Mass Defect Filter for the 2C-Phenethylamine Subclass .............................. 88 5.2.4 Absolute Mass Defect Filter for the NBOMe-Phenethylamine Subclass..................... 90 5.2.5 Kendrick Mass Defect Filters for Phenethylamines Based on Molecular Ions ............ 92 5.2.6 Kendrick Mass Defect Filters of the APB-Phenethylamine Subclass .......................... 93 5.2.7 Kendrick Mass Defect Filters of the 2C-Phenethylamine Subclass ............................. 94 5.2.8 Kendrick Mass Defect Filters of the NBOMe-Phenethylamine Subclass .................... 96 5.2.9 Kendrick Mass Defect Filters for Neutral Losses and Common Fragment Ions .......... 98 5.3 Scheme for Characterization of Synthetic Phenethylamines using High-Resolution Mass Spectra ..................................................................................................................................... 108 5.4 Summary ........................................................................................................................... 115 APPENDICES ............................................................................................................................ 116 APPENDIX A: High-Resolution Mass Spectra ...................................................................... 117 APPENDIX B: Additional High-Resolution Characterization Scheme Examples ................. 125 REFERENCES ........................................................................................................................... 131 VI. Conclusions and Future Work .............................................................................................. 133 6.1 Conclusions ....................................................................................................................... 133 6.2 Future Work ...................................................................................................................... 134 vi LIST OF TABLES Table 2.1 Absolute mass defects of elements commonly used in this research ........................... 28 Table 3.1 Substituents for 2C-phenethylamine compound shown in Figure 3.1 ...........................33 Table 3.2 Substituents for NBOMe-phenethylamine compound shown in Figure 3.2 ..................34 Table A.1 Compound abbreviations and chemical names ............................................................ 42 Table 4.1 Retention index and molecular ion determinations of sample set compounds ............. 44 Table 5.1 Calculation of absolute mass defect molecular ion filter .............................................. 85 Table 5.2 Calculation of APB Kendrick mass defect filter .......................................................... 93 Table 5.3 Calculation of 2C Kendrick mass defect filter.............................................................. 95 Table 5.4 Calculation of NBOMe Kendrick mass defect filter .................................................... 97 Table 5.5 Ion table of 2C-H showing abundant ion elemental composition assignments and mass accuracies ...................................................................................................................................... 99 Table 5.6 Table of remaining ions after common losses of all 2C compounds .......................... 100 Table 5.7 Kendrick mass defect filters associated with ion fragments after common neutral losses ..................................................................................................................................................... 102 Table 5.8 Ion table of 25H-NBOMe with elemental composition assignments and mass accuracies of most abundant fragment ions above m/z 105 ........................................................ 104 vii LIST OF FIGURES Figure 1.1 Phenethylamine ............................................................................................................. 2 Figure 1.2 Phenethylamine analogs. (A) 2,5-dimethoxyphenethylamine (2C-H) (B) 4-ethyl-2,5dimethoxyphenyl-N-(2-methoxybenzyl) ethan-1-amine (25E-NBOMe) (C) aminopropylbenzofuran (4-APB) (D) 1-(3, 5-dimethoxy-4-propoxyphenyl) propan-2-amine (3CP) ..................................................................................................................................................... 3 Figure 2.1 Schematic of gas chromatography (GC) instrument ................................................... 16 Figure 2.2 Example chromatogram of a multi-component gas chromatography separation ........ 19 Figure 2.3 Overall schematic of mass spectrometer (MS) ............................................................ 20 Figure 2.4 Electron ionization (EI) source .................................................................................... 21 Figure 2.5 Quadrupole mass analyzer showing two different ion trajectories occurring simultaneously. The red ion is neutralized as it collides with one of the rods, is pumped away, and not detected, while the blue ion has a stable trajectory through the analyzer and travels to the detector .......................................................................................................................................... 22 Figure 2.6 Time-of-flight mass analyzer showing two different ion trajectories occurring simultaneously. Both ions are accelerated in the pusher region with the same kinetic energy, but the red ion penetrates deeper into the reflectron because it has larger mass, thus reaching the detector after the blue ion ............................................................................................................. 25 Figure 2.7 Spectra and chemical structure of 2, 5- dimethoxyphenethylamine (2C-H) via (A) low-resolution (QMS) and (B) high-resolution (TOFMS) mass spectrometry ............................ 27 Figure 3.1 Structures of the phenethylamines in the reference set (A) 4-(2-aminopropyl) benzofuran (4-APB) (B) 5-(2-aminopropyl) benzofuran (5-APB) (C) 6-(2-aminopropyl) benzofuran (6-APB) (D) 7-(2-aminopropyl) benzofuran and (E) the core structure of 2,5dimethoxyphenethylamine (2C-phenethylamines). The substituents at R1 and R2 and the corresponding 2C compound are given in Table 3.1. ................................................................... 33 Figure 3.2 Structures of more of the phenethylamines in the reference set (A) 3,4,5-trimethoxybenzeneethanamine (mescaline), (B) 4-ethoxy-3,5-dimethoxy-benzeneethanamine (escaline) both 3C-phenethylamines, (C) the core of N-benzyl phenethylamine analogs (NBOMephenethylamines). The substituents at R1 and R2 corresponding NBOMe compound are given in Table 3.2, and (D) 3,4,5-trimethoxy-N-[(2-methoxyphenyl)methyl]-benzeneethanamine (mescaline-NBOMe) ......................................................................................................................34 Figure 3.3 Structures of cathinones in the reference set (A) 4-methylmethcathinone (mephedrone) and (B) 3-methylethcathinone (3-MEC) ............................................................... 35 viii Figure 4.1 Representative spectra of (A) 6-APB, (B) 2C-H, and (C) 25H-NBOMe and proposed structures for the most dominant fragment ions in each spectrum ............................................... 46 Figure 4.2 Mass spectra of (A) 25G-NBOMe and the cannabinoid (B) XLR-115 which both have a molecular ion of m/z 330. NBOMes can be differentiated from cathinones using characteristic peaks at m/z 91, 121, and 150. XLR-11 spectrum obtained from Cayman Chemical .................. 48 Figure 4.3 Mass spectrum of (A) 2C-H and (B) 2C-B showing characteristic 2C neutral losses of 29 and 60 Da and the structures of the fragment ions remaining after each loss ......................... 51 Figure 4.4 Mass spectrum of (A) 25H-NBOMe and (B) 25B-NBOMe showing characteristic NBOMe neutral losses of 31 and 149 Da and the structures of the fragment ions remaining after each loss, as well as common fragment ions (m/z 91, 121, 150) .................................................. 52 Figure 4.5 Characteristic isotope pattern in mass spectra of compounds containing bromine, (A) 2C-B and (B) 25B-NBOMe .......................................................................................................... 54 Figure 4.6 Characteristic isotope pattern in mass spectra of compounds containing chlorine (A) 2C-C and (B) 25C-NBOMe .......................................................................................................... 56 Figure 4.7 Full mass spectrum of (A) 2C-I and (B) expanded section of same spectrum to highlight I+ and HI+ ions ............................................................................................................... 57 Figure 4.8 Full mass spectrum of (A) 25I-NBOMe and (B) expanded section of same spectrum to highlight I+ and HI+ ions ............................................................................................................... 58 Figure 4.9 Mass spectrum of (A) 2C-T and (B) 25T-NBOMe indicating inconsistent sulfur isotope pattern ............................................................................................................................... 60 Figure 4.10 Mass spectrum of (A) 2C-N and (B) 25N-NBOMe indicating M+ with an even mass that suggests an even number of nitrogens present ....................................................................... 61 Figure 4.11 Characterization scheme for low-resolution mass spectra of synthetic phenethylamines to distinguish APB, 2C, and NBOMe subclasses ............................................. 64 Figure 4.12 Characterization scheme for low-resolution mass spectra of synthetic phenethylamines to determine substituents on 2C- or NBOMe-phenethylamines ....................... 65 Figure 4.13 Mass spectrum and structure of cathinone, 3-methylethcathinone (3-MEC) ............ 68 Figure 4.14 Mass spectrum of 3C phenethylamine, mescaline, which would be mischaracterized as a 2C because of its loss of 29 Da (m/z 182) and 60 Da (m/z 151) ............................................ 69 Figure A.1 Low-resolution mass spectra of (A) 4-(2-aminopropyl)benzofuran (4-APB), (B) 5-(2aminopropyl)benzofuran (5-APB), and (C) 7-(2-aminopropyl)benzofuran ................................. 72 ix Figure A.2 Low-resolution mass spectra of (A) 2,5-dimethoxy-4-methylphenethylamine (2C-D), (B) 2,5-dimethoxy-4-ethylphenethylamine (2C-E), (C) 3,4-dimethyl-2,5dimethoxyphenethylamine (2C-G), and (D) 2,5-dimethoxy-4-propylphenethylamine (2C-P) .... 73 Figure A.3 Low-resolution mass spectra of 2,5-dimethoxy-4-ethylthiophenethylamine (2C-T-2) ....................................................................................................................................................... 74 Figure A.4 Low-resolution mass spectra of (A) 2-(2,5-dimethoxy-4-methylphenyl)-N-(2methyoxybenzyl)ethanamine (25D-NBOMe) and (B) 2-(4-ethyl-2,5-dimethoxyphenyl)-N-(2methoxybenzyl)ethanamine (25E-NBOMe) ................................................................................. 75 Figure A.5 Low-resolution mass spectra of (A) 2,5-dimethoxy-N-[(2-methoxyphenyl)methyl]-4[(1-methylethyl)thio]-benzeneethanamine (25T-4-NBOMe), (B) 2,5-dimethoxy-N-[(2methoxyphenyl)methyl]-4-(propylthio)-benzeneethanamine (25T-7-NBOMe), and (C) 3,4,5trimethoxy-N-[(2-methoxyphenyl)methyl]-benzeneethanamine (mescaline-NBOMe) ............... 76 Figure A.6 Low-resolution mass spectra of (A) 4-ethoxy-3,5-dimethoxy-benzeneethanamine (escaline) and (B) 4-methylmethcathinone (mephedrone) ........................................................... 77 Figure 5.1 Comparison of (A) low-resolution and (B) high-resolution mass spectra for 6-APB (left), 2C-H (middle), and 25H-NBOMe (right) ........................................................................... 81 Figure 5.2 Comparison of (A) low-resolution and (B) high-resolution mass spectra for 2C-B. Dominant fragment ions are labeled and in (B) assigned element formulae and mass accuracies are given ........................................................................................................................................ 83 Figure 5.3 Absolute mass defect filter created using a training set of phenethylamines defined in Table 5.1. The absolute mass defect filter was defined at 142.4 ± 54.1 mDa at a 99.9991% confidence level. The horizontal lines represent the average (yellow), and the upper and lower bounds of the mass defect filter (purple) ...................................................................................... 86 Figure 5.4 APB subclass absolute mass defect filter at 99.7 ± 1.6 mDa at a 90% confidence level. The horizontal lines represent the average (black) and the upper and lower bounds of the mass defect filter (red) ........................................................................................................................... 88 Figure 5.5 2C subclass absolute mass defect filter at 133.1 ± 32.2 mDa at a 95% confidence level. The horizontal lines represent the average (light blue) and the upper and lower bounds of the mass defect filters (dark blue) ................................................................................................. 90 Figure 5.6 NBOMe subclass absolute mass defect filter at 179.6 ± 20.5 mDa at a 95% confidence level. The horizontal lines represent the average (light purple) and the upper and lower bounds of the mass defect filter (dark purple) ............................................................................................... 91 Figure 5.7 APB subclass Kendrick mass defect filter at 95.9 ± 1.6 mDa at a 90% confidence level. The horizontal lines represent the average (black) and the upper and lower bounds of the mass defect filter (red) .................................................................................................................. 94 x Figure 5.8 2C subclass Kendrick mass defect filter at 92.2 ± 1.5 mDa at a 95% confidence level. The horizontal lines represent the average (light blue) and the upper and lower bounds of the mass defect filter (dark blue) ........................................................................................................ 96 Figure 5.9 NBOMe subclass Kendrick mass defect filter at 171.5 ± 7.7 mDa at a 99% confidence level. The horizontal lines represent the average (light purple) and the upper and lower bounds of the mass defect filter (dark purple) ............................................................................................... 97 Figure 5.10 Spectrum of 2C-H showing abundant ions................................................................ 99 Figure 5.11 Proposed structures for fragment ions of 2C-H after their neutral losses ............... 100 Figure 5.12 Kendrick mass defect filters developed based on common losses of alkyl-substituted 2C compounds. Points represent KMD of fragment ions remaining after each respective loss. The horizontal lines represent the average (lighter colors) and the upper and lower bounds of each mass defect filter (darker colors) ........................................................................................ 102 Figure 5.13 Selected Kendrick mass defect filters representing losses of CH3N and C2H6NO for all 2C fragments falling within said filters. Fragment shown outside the filter is from 2C-T. The horizontal lines represent the average (light green and purple) and the upper and lower bounds of each mass defect filter (dark green and purple) .......................................................................... 103 Figure 5.14 Spectrum of 25H-NBOMe and most abundant fragment ions above m/z 105 ........ 104 Figure 5.15 Proposed structures for fragment ions of 25H-NBOMe after their neutral losses .. 105 Figure 5.16 Proposed structural elucidation of 2C-N and 25N-NBOMe leading to the same fragment (C9H11NO4) .................................................................................................................. 106 Figure 5.17 Selected Kendrick mass defect filter and corresponding NBOMe fragments falling within the filter. Fragments shown outside the filter are from mescaline-NBOMe and 2C-T. The horizontal lines represent the average (light green) and the upper and lower bounds of the mass defect filter (dark green) ............................................................................................................. 106 Figure 5.18 Selected Kendrick mass defect filters and corresponding 3C fragments falling outside the filters ........................................................................................................................ 107 Figure 5.19 Characterization scheme for high-resolution mass spectral data. M+adj is the mass of the molecular ion adjusted for a halogen/sulfur/nitro substituent ............................................... 109 Figure 5.20 Mass spectrum and structure of cathinone, 3-methylethcathinone (3-MEC) showing loss of C3H8O, which is uncharacteristic of the phenethylamine class....................................... 112 Figure 5.21 Mass spectrum of 3C-phenethylamine, mescaline and fragment ions remaining after neutral losses, the KMD of which can be used to distinguish 2C from 3C-phenethylamines .... 114 xi Figure A.1 High-resolution mass spectra of (A) 4-(2-aminopropyl)benzofuran (4-APB), (B) 5-(2aminopropyl)benzofuran (5-APB), and (C) 7-(2-aminopropyl)benzofuran ............................... 117 Figure A.2 High-resolution mass spectra of (A) 2,5-dimethoxy-4-methylphenethylamine (2C-D), (B) 2,5-dimethoxy-4-ethylphenethylamine (2C-E), and (C) 2,5-dimethoxy-4propylphenethylamine (2C-P)..................................................................................................... 118 Figure A.3 High-resolution mass spectra of (A) 2,5-dimethoxy-4-chlorophenethylamine (2C-C), (B) 2,5-dimethoxy-4-iodophenethylamine (2C-I), and (C) 2,5-dimethoxy-4-nitrophenethylamine (2C-N) ......................................................................................................................................... 119 Figure A.4 High-resolution mass spectra of (A) 2,5 -dimethoxy-4-methylthiophenethylamine (2C-T) and (B) 2,5-dimethoxy-4-ethylthiophenethylamine (2C-T-2) ........................................ 120 Figure A.5 High-resolution mass spectra of (A) 2-(2,5-dimethoxy-4-methylphenyl)-N-(2methyoxybenzyl)ethanamine (25D-NBOMe), (B) 2-(4-ethyl-2,5-dimethoxyphenyl)-N-(2methoxybenzyl)ethanamine (25E-NBOMe) and (C) 2,5-dimethoxy-N-[(2methoxyphenyl)methyl]-3,4-dimethyl-benzeneethanamine (25G-NBOMe) ............................. 121 Figure A.6 High-resolution mass spectra of (A) 4-bromo-2,5-dimethoxy-N-[(2methoxyphenyl)methyl]-benzeneethanamine (25B-NBOMe), (B) 4-chloro-2,5-dimethoxy-N-[(2methoxyphenyl)methyl]-benzeneethanamine (25C-NBOMe), and (C) 4-iodo-2,5-dimethoxy-N[(2-methoxyphenyl)methyl]-benzeneethanamine (25I-NBOMe) ............................................... 122 Figure A.7 High-resolution mass spectra of (A) 2,5-dimethoxy-N-[(2-methoxyphenyl)methyl]-4(methylthio)-benzeneethanamine (25T-NBOMe), (B) 2,5-dimethoxy-N-[(2methoxyphenyl)methyl]-4-[(1-methylethyl)thio]-benzeneethanamine (25T-4-NBOMe), (C) 2,5dimethoxy-N-[(2-methoxyphenyl)methyl]-4-(propylthio)-benzeneethanamine (25T-7-NBOMe), and (D) 3,4,5-trimethoxy-N-[(2-methoxyphenyl)methyl]-benzeneethanamine (mescalineNBOMe) ..................................................................................................................................... 123 Figure A.8 High-resolution mass spectra of (A) 4-ethoxy-3,5-dimethoxy-benzeneethanamine (escaline) and (B) 4-methylmethcathinone (mephedrone) ......................................................... 124 Figure A.9 Mass spectrum of 2C-G and fragment ions remaining after neutral losses .............. 125 Figure A.10 Mass spectrum of 2C-B and fragment ions remaining after neutral losses ............ 128 Figure A.11 Mass spectrum of 25N-NBOMe and fragment ions remaining after neutral losses130 xii I. Introduction 1.1 Synthetic Designer Drugs According to the 2015 National Drug Threat Assessment Summary, the abuse of synthetic designer drugs has remained constant or increased since their popularity skyrocketed in 2008.1 These drugs are typically used by younger individuals, as they are marketed in packages with bright colors and cartoons, and in a variety of fruit or candy flavors.1 By definition, a designer drug is a synthetic version of a controlled substance that is produced with a slightly altered molecular structure to avoid being classified as an already regulated compound.2 Consequently, users may experience the same psychoactive effects as controlled substances without legal ramifications. The 2013 National Drug Threat Assessment Summary defined seven classes of synthetic designer drugs: cannabinoids, phencyclidines or arylcyclohexamines, tryptamines, piperazines, pipradrols, tropane alkaloids, and phenethylamines, which are the focus of this research.3 The Drug Enforcement Administration (DEA) has exercised emergency scheduling authority to temporarily control over 20 synthetic drugs since the 2012 enactment of the Synthetic Drug Abuse Prevention Act. As an amendment of the Controlled Substances Act, the Synthetic Drug Abuse Prevention Act already had regulated 29 synthetic drugs as Schedule I substances. Despite increased efforts in legislation and law enforcement, clandestine chemists are frequently two steps ahead, because as soon as one designer compound is scheduled, a new analog appears on the market. The new analog often differs only slightly in chemical structure or composition from the regulated compound, for example, as an isomeric form or with different substitutions. The core structure of phenethylamine, comprised of a benzene ring and amine side chain, is shown in 1 Figure 1.1, while some of its substituted analogs are shown in Figure 1.2. Within the phenethylamine class, there are subclasses of 2,5-dimethoxyphenethylamines (2C), N-benzyl phenethylamine analogs (NBOMe), aminopropyl benzofuran phenethylamines (APB), and 3,4,5trimethoxyphenethylamines (3C) compounds.4 Further, within each subclass, there are many compounds available with varying substituents around the subclass-core structure. For example, typical substitutions of varying alkyl chain length occur on the popular and well-known 2Cphenethylamine core in the 3’ and 4’ positions, while non-alkyl substituents like halogens, nitro, and sulfur groups occur in the 4’ position. Most compounds in the APB subclass do not have additional substituents around the ring; instead, the location of the furan ring around the benzene ring changes to create different isomers. Compounds in the NBOMe class have substituents in the same locations as the 2C compounds, but also can be altered on the N-benzyl side of the compound by adding substituents or changing the placement of the methoxy group. The combinations of varying substituents and substitution positions to the different subclass core structures is limitless. Thus, new, unscheduled “legal” highs are a challenge for law makers and forensic analysts to identify, despite the legislation that is already in place. NH2 Figure 1.1 Phenethylamine 2 (A) 2C (B) NBOMe O O NH2 3' NH2 H (D) 3C O NH2 N 3' 4' 4' (C) APB O O O O O O Figure 1.2 Phenethylamine analogs. (A) 2,5-dimethoxyphenethylamine (2C-H) (B) 4-ethyl-2,5dimethoxyphenyl-N-(2-methoxybenzyl) ethan-1-amine (25E-NBOMe) (C) aminopropylbenzofuran (4-APB) (D) 1-(3, 5-dimethoxy-4-propoxyphenyl) propan-2-amine (3CP) 1.2 Current Methods of Analysis of Submitted Drug Samples and Limitations The Scientific Working Group for the Analysis of Seized Drugs (SWGDRUG) has recommendations for the analysis of controlled substances, with different analytical techniques placed into categories based upon their maximum potential discriminating power.5 Currently, the method of choice for the analysis of controlled substances in most forensic laboratories is gas chromatography-mass spectrometry (GC-MS). This hyphenated technique separates components in a given sample and the mass spectrum of each separated component is recorded. In these instruments, electron ionization (EI) is used which, as a ‘hard’ ionization method, results in substantial fragmentation of each separated component. As a result, the mass spectral fragmentation patterns contain a significant amount of information from which structural elucidation is possible to determine the identity of the compound. Gas chromatography-mass spectrometry satisfies SWGDRUG’s recommendation that at least two analytical techniques be used for identification. GC-MS is the preferred method for satisfying these recommendations because of its reproducibility, cost effective nature, and the ability for high throughput of samples. 3 However, there are some limitations with using GC-MS for the identification of synthetic designer drugs. Although there is extensive fragmentation of compounds using electron ionization, this is often insufficient for definitive identification of synthetic designer drugs because of the high structural similarity among compounds within the same class. Further, the conventional instruments are equipped with a single quadrupole mass analyzer (GC-QMS), which yields only nominal mass information for each fragment ion. This means that distinction of isomeric compounds is very difficult. Isomeric compounds have the same molecular mass and, therefore, exhibit similar mass spectra, with the same ions appearing only in different ratios. Because new designer drug analogs appear on the market so quickly, an additional obstacle for forensic laboratories to overcome is that oftentimes reference standards are not available. A lack of reference standards is problematic because identification of drugs in forensic labs is based on a visual comparison of the mass spectrum of the reference standard to the mass spectrum of the questioned sample. Therefore, there is a need for improved methods for the identification and characterization of synthetic designer drugs using the conventional GC-QMS instruments available in forensic laboratories. 1.3 Current Research of Synthetic Designer Drugs There has been extensive research of some designer drug compound classes, specifically cathinones and cannabinoids, involving the identification and characterization of designer drugs in street samples. However, the research has been performed primarily using high-resolution mass spectrometry.6 - 10 High-resolution mass spectrometers are capable of acquiring accurate mass data, from which the elemental composition of each ion can be determined with a high degree of confidence. 4 An additional advantage of using high-resolution methods is that from the accurate mass, the mass defect (defined as the difference between the exact and nominal mass) for each ion can also be determined. These mass defects can be used to identify filters that are characteristic of a given compound class and therefore, can be used for characterization of unknown synthetic designer drugs. Grabenauer et al. used mass defect filters to characterize synthetic cannabinoids.11 Compounds analyzed in the study had mass defects ranging from 0.13 and 0.23 Da. A filter was developed that was centered at 0.185 Da with a window of ±50 mDa and this filter encompassed 75% of the known cannabinoids with the indole core structure. However, the study used high-resolution mass spectrometry with liquid chromatography. While liquid chromatography with high-resolution mass spectrometry has advantages in the characterization of synthetic designer drugs (i.e., accurate mass data and mass defect filters), there are limitations. These instruments use electrospray ionization, which is a soft ionization method, resulting in little fragmentation. The outcome is a molecular ion which can allow for molecular mass information, however, also results in less fragmentation and thus little structural information. As structural information is especially important for unknowns, less fragmentation can make definitive compound identification difficult. Another limitation, which is more problematic, is that these instruments are not currently available in forensic laboratories. So, while research is being conducted, the methods developed are neither practical nor implementable for use in a forensic laboratory. Zuba previously described a method for categorizing designer drugs into their compound classes based on mass spectral data collected using the conventional GC-QMS instrument configuration that is available in forensic labs. Zuba also used liquid chromatographyelectrospray ionization with quadrupole time of flight mass spectrometry (LC-ESI/QTOF-MS), a 5 more sophisticated instrument not readily available.4 The characterization of each compound was based on molecular and fragment ions. However, with GC-QMS, only nominal mass data were collected. Thus, only preliminary classification of known “legal highs” into general compound class (e.g., phenethylamine versus cathinone) rather than subclass (e.g., APB versus 2Cphenethylamine) was possible using Zuba’s characterization flow chart. Preliminary classification into general compound class is problematic for large compound classes such as the phenethylamines for which several subclasses exist, as discussed previously. Using the flow chart, these compounds could be identified as phenethylamines, but no further sub-classification would be possible. Because of the wide structural variation among subclasses of the phenethylamine class, a more specific characterization is needed. Additionally, Zuba focused on the investigation of fragmentation of compounds from only the cathinone class. Other characterization methods for compounds in the phenethylamine and other classes have been investigated. For example, Zuba and Sekula characterized the phenethylamine 3,4dimethyl-2,5-dimethoxyphenethylamine (2C-G); however, four different instruments utilizing six analysis methods were needed to complete the characterization.7 These instruments included GC-EI/MS and Fourier transform infrared spectroscopy (FTIR), both of which are commonly available in forensic laboratories, LC-ESI/QTOF-MS, and two types of nuclear magnetic resonance spectroscopy (NMR) which are not commonly used in forensic labs. Overall, using six techniques to characterize one synthetic drug analog is strenuous, time-consuming, and impractical for forensic analysts with large workloads. Similarly, Shevyrin et al. isolated, identified, and characterized several indole-3-carboxylic acid synthetic cannabinoids by similar methods (GC-QTOF-MS, ultra-performance LC-QTOF-MS, NMR, and FTIR)8 while Uchiyama et al. identified fifteen designer drugs in street samples also using UPLC-ESI-MS and GC-EI/MS 6 along with NMR.9 In a different study, Uchiyama et al. characterized several cannabinoids and NBOMe phenethylamines, but again using only one instrumental technique used in forensic laboratories (GC-MS) and two techniques not used in forensic laboratories (LC-MS and NMR).6 The use of so many techniques for characterizing new analogs highlights the lack of a quick, clear, concise, and practical method for characterization of these compounds. Fornal characterized the cathinone compound class by subclass using high performance LC-QTOF-MS.12 Based on structural features such as double-bond equivalency and the characteristics of the amine group, nine different subclasses of cathinones were identified. The fragmentation pathways for each class were proposed, and specific losses common to each class were briefly discussed. However, because ESI was used for ionization, the proposed fragmentation pathways developed will be different than those from the commonly used election ionization method. Therefore, if a forensic analyst received a new analog belonging to one of the nine cathinone compound classes presented, a direct characterization could not be made. A DEA monograph authored by Casale and Hays describes the synthesis, characterization, and differentiation of eleven NBOMe phenethylamines.13 Using GC-MS and FTIR, both instruments used in a typical forensic lab, each NBOMe could be distinguished from their corresponding 2C analog. However, each NBOMe was differentiated from one another by relative ion abundances. If a reference standard is not available for comparison, using relative abundances is problematic because ion abundances can vary from instrument to instrument as well as sample run to run. While differentiation of NBOMes was reported, the method is still somewhat subjective, therefore, an improved method of differentiation is needed. Overall, there is a need for a rapid characterization scheme for synthetic designer drugs that employs the conventional GC-QMS instrument configuration used by the majority of 7 forensic laboratories for controlled substance identification. Current characterization methods for these compounds frequently do not use instrumentation available in forensic laboratories and, therefore, the characterization schemes developed are not directly implementable into labs. 1.4 Research Objectives and Goals This research focused on developing methods for the characterization of synthetic designer drugs according to structural subclass. The focus in this initial work was the synthetic phenethylamine compound class. More specifically, the objectives were: 1. To develop a characterization scheme based on data collected using common GCQMS instruments available in forensic laboratories. 2. To investigate the potential of high-resolution mass spectrometry and mass defect for a more robust characterization of designer drugs. The objectives were accomplished by achieving the following goals: 1. Developing a flow-chart style characterization scheme based on characteristic mass spectral features obtained using GC-EI-QMS. To do this, a range of phenethylamine standards encompassing different subclasses was analyzed by both low- and high-resolution instruments (GC-QMS and GC-TOFMS). Because these instruments use the same electron ionization method, their spectra are comparable. However, the TOF mass analyzer provides accurate mass of each fragment ion. Accurate mass allows for the confirmation of the elemental formula of each ion and the understanding of how these compounds fragment under electron ionization conditions. The fragmentation pathways, mass spectral features, and neutral losses that are characteristic of each phenethylamine subclass were confirmed by the high-resolution spectra and are translatable to the GC-QMS data because the 8 same ionization method is used. Because GC-QMS is the same instrument configuration used in forensic labs, the scheme is immediately implementable in forensic labs. From a confirmed molecular ion, neutral losses were investigated for characterization. A neutral loss is a fragment lost as a neutral molecule during ionization. By investigating what common losses occur from compounds of each subclass, characteristic neutral losses can be identified. Unknown compounds exhibiting those common losses may be able to be characterized into specified subclasses. Additionally, retention index can be calculated and used as another characterization tool.14 In current forensic practices, it is the comparison of chromatographic as well as mass spectral data of reference standards to questioned samples that allows for identification. However, in the event that no reference sample is readily available to analyze on the laboratory instrument, reference chromatograms and spectra from online sources such as the SWGDRUG drug monographs or Cayman Chemical© can be downloaded.15, 16 The retention times from the monographs may differ from that obtained experimentally in the lab because of variations in temperature program, column length and diameter, stationary phase film thickness, or carrier gas velocity and pressure. By calculating and using retention index, these variables are eliminated and retention indices collected on two different instruments can be compared. Further, a range of retention indices for each compound subclass was developed in this work and incorporated into the characterization scheme. For example, a range of retention indices for the 2Cphenethylamine class was determined, and if an unknown were to have a retention index that fell within that range, it would increase the confidence in preliminarily characterizing that compound as a 2C-phenethylamine. 9 2. Develop mass defect filters from the high-resolution data that can be used to enhance characterization. To do this, the exact mass data obtained were used to develop mass defect filters to use for more confident characterization of phenethylamines into structural subclasses. The exact mass data also highlight the utility of high-resolution mass spectrometry, should it ever become available to forensic labs. Previous preliminary work had identified potential limitations in the development of mass defect filters that this work overcomes.17 The first limitation is a lack of molecular ion. Because molecules are being bombarded with such high energy in electron ionization mode, they often do not remain intact and completely fragment. This means that some compounds do not produce a molecular ion peak in their mass spectrum. This problem can be overcome by determining the molecular ion using chemical ionization. Chemical ionization is a soft ionization technique that most GC-QMS systems can be equipped to perform. A second limitation associated with mass defect filters is determining how wide or narrow the filter should be. If the filter is too wide, then compounds from different subclasses will be included. However, if the filter is too narrow, then compounds from the same subclass may be excluded. These problems are addressed by defining filters based on the Kendrick mass defect. A Kendrick mass defect is an adjusted mass based on the conversion between exact mass of a methylene unit (CH2) and its nominal mass. Kendrick mass is defined as the exact mass multiplied by this conversion and is a way to normalize masses of a similar class of compounds. Thus, members of a homologous series that differ only in the number of methylene groups will have the same Kendrick mass defect. As a result, members of a given subclass will have the same Kendrick mass defects, which will be different from the Kendrick mass defects of another subclass. The use of Kendrick 10 mass defect to overcome problems associated with the width of the filter are utilized in this research. The development of an “easy-to-follow” flow-chart style characterization scheme will be easily and immediately implementable into forensic laboratories because it will have been created using instrumentation and methodology that is already in place. The characterization scheme will be used as an initial screening method to determine if further examination of a submitted controlled substance sample is necessary. Additionally, the scheme will assist in identification of new compounds in a constantly changing drug market, and allow for characterization of unknowns for which no reference standard is available. Further, by investigating the use of mass defect filters for a more robust characterization, this research highlights the potential for high-resolution instrumentation for forensic applications. As molecular ions are not always available for synthetic phenethylamines, mass defect filters are developed for characteristic fragment ions of neutral losses. This, along with the Kendrick mass defect and retention index, will provide sufficient information to distinguish synthetic phenethylamines from different subclasses, and the utility of high-resolution will be highlighted, should those instruments ever be made available. 11 REFERENCES 12 REFERENCES 1. National Drug Threat Assessment Summary. U.S Department of Justice Drug Enforcement Administration, 2015. https://www.dea.gov/docs/2015%20NDTA%20Report.pdf 2. Designer Drug. Merriam-Webster Dictionary. http://www.merriamwebster.com/dictionary/designer%20drug 3. National Drug Threat Assessment. U.S. Department of Justice Drug Enforcement Administration, 2013. http://www.dea.gov/resource-center/DIR-01713%20NDTA%20Summary%20final.pdf 4. Zuba, D. Identification of cathinones and other active components of ‘legal highs’ by mass spectrometric methods. TrAC Trends in Analytical Chemistry. 2012 Feb; 32: 15-30. 5. Scientific Working Group for the Analysis of Seized Drugs (SWGDRUG). Recommendations. http://www.swgdrug.org/Documents/SWGDRUG%20Recommendations%20Version%2 07-0.pdf 6. Uchiyama, N. Shimokawa, Y. Matsuda, S. Kawamura, M. Kikura-Hanajiri, R. Goda, Y. Two new synthetic cannabinoids, AM-2201 benzimidazole analog (FUBIMINA) and (4methylpiperazin-1-yl) (1-pentyl-1H-indol-3-yl) methanone (MEPIRAPIM), and three phenethylamine derivatives, 25H-NBOMe 3, 4, 5-trimethoxybenzyl analog, 25BNBOME, and 2C-N-NBOMe, identified in illegal products. Journal of Forensic Toxicology 2013 Jan; 32(1): 105-15. 7. Zuba, D. Sekula, K. Identification and characterization of 2, 5-dimethoxy-3, 4-dimethylβ-phenethylamine (2C-G) - A new designer drug. Drug Testing and Analysis 2012 Jul; 5(7): 549-59. 8. Shevyrin, V. Melkozerov, V. Nevero, A. Eltsov, O. Shafran, Y. Analytical characterization of some synthetic cannabinoids, derivatives of indole-3-carboxylic acid. Forensic Science International. 2013 Oct; 232(1-3): 1-10. 9. Uchiyama, N. Matsuda, S. Kawamura, M. Shimokawa, Y. Kikura-Hanajiri, R. Aritake, K. et al. Characterization of four new designer drugs, 5-chloro-NNEI, NNEI indazole analog, α-PHPP and α-POP, with 11 newly distributed designer drugs in illegal products. Forensic Science International. 2014 Oct; 243: 1-13. 10. Uchiyama, N. Kikura-Hanajiri, R. Ogata, J. Goda, Y. Chemical analysis of synthetic cannabinoids as designer drugs in herbal products. Forensic Science International. 2012 May; 198(1-3): 31-38. 13 11. M. Grabenauer, W. L. Krol, J. L. Wiley, B. F. Thomas. Analysis of Synthetic Cannabinoids using High-Resolution Mass Spectrometry and Mass Defect Filtering: Implications for Nontargeted Screening of Designer Drugs. Journal of Analytical Chemistry. 2012 June; 84(13): 5574-81. 12. Fornal, E. Study of collision-induced dissociation of electrospray-generated protonated cathinones. Drug Testing and Analysis 2013 Nov; 6(7-8). 705-715. 13. Casale, J. Hays, P. Characterization of Eleven 2,5-Dimethoxy-N-(2methoxybenzyl)phenethylamine (NBOMe) Derivatives and Differentiation from their 3and 4- Methoxybenzyl Analogs – Part 1. U.S. Department of Justice Drug Enforcement Administration Microgram Journal 9(2). 84 – 109. 14. Skoog, D. Holler, F. Crouch, S. Principles of Instrumental Analysis. 6th ed.; Thomas Brooks/Cole: Belmont, CA, 2007. 15. Scientific Working Group for the Analysis of Seized Drugs (SWGDRUG). Micrographs. http://swgdrug.org/monographs.htm 16. Cayman Chemical. Product Search and Drug ID. https://www.caymanchem.com/forensics/search 17. Chu, F. Improving Methods for the Analysis of Controlled Substances. Master’s Thesis, Michigan State University, East Lansing, MI, 2015. 14 II. Theory 2.1 Chromatography Overview Chromatography is an analytical technique used to separate chemical components of a mixture, called analytes, by passing them through two phases in which the individual analytes of the mixture, move at different rates based on their chemical properties. The two phases are called the mobile phase and the stationary phase. The mobile phase is used to move the mixture through a column, while the stationary phase is affixed inside the column. The separation of the different analytes in the mixture is due to the transfer between the two phases. The two phases are selected so that the analytes of the mixture distribute themselves between the mobile and stationary phases to varying degrees.1 Typically, either gas or liquid chromatography can be performed, depending on the matrix and the chemical composition of the analytes of the mixture to be separated. For the purposes of this research, only gas chromatography will be discussed, as it is the prevalent instrumentation used in forensic laboratories for controlled substance identification. 2.2 Gas Chromatography Overview In gas chromatography (GC) the sample to be analyzed is injected into the instrument for separation (Figure 2.1). The sample then flows through the instrument by a carrier gas (the mobile phase) to a capillary column within an oven. The mobile phase is an inert gas (e.g., helium) that does not chemically react with the sample. The inner wall of the column is coated with the stationary phase, which can vary depending on the type of analytes to be separated. At the end of the column is a detector, which generates a chromatogram. Each of these steps will now be discussed in detail. 15 Syringe Injection port Detector Gas cylinder Column Oven Figure 2.1 Schematic of gas chromatography (GC) instrument The mobile phase is an inert gas, typically hydrogen or helium, which will not interact with the sample, rather just flow through the system. The mobile phase is stored in a gas cylinder near the GC instrument and flows into the injection port where samples are first injected. Liquid and solid samples are first prepared by dissolving them in a suitable solvent for GC analysis. A suitable solvent is one that is easily volatilized, non-reactive with the sample, and will cause minimal degradation of the stationary phase inside the column. The sample is then loaded into a syringe and injected into the injection port of the GC instrument. The injection port must be hot enough (e.g., 250 °C) so that the sample is instantly volatilized. The injection port also has mobile phase flowing through it at a set rate (e.g., 1 mL/min). Additionally, inside the injection port there is a valve that can be used to deplete some of the sample to waste. This valve is called the split valve and can be set to bleed off a specific ratio of the injected sample, as defined by the user.1 A typical split ratio can be anywhere from 20:1 to 100:1, and the higher the ratio, the less sample enters the column. A high split ratio would be used for concentrated 16 samples. In this work, a splitless injection was used. A splitless injection is advantageous because all of the sample is introduced onto the column, which is useful for samples in which the analyte is present in low concentrations. In forensic laboratories, many submitted controlled substance samples contain both a cutting agent (e.g., caffeine) and the controlled substance. Because the controlled substance in a submitted sample is at a lower concentration, a splitless injection is typically used to ensure that the controlled substance is detected. After injection, the vaporized sample enters the column. In gas chromatography, the column is typically made of fused silica and can be anywhere from a few meters to 100 m in length, with varying diameters (typically 0.25 mm), and differing stationary phase film thicknesses (typically 0.25 µm). A column with the appropriate dimensions and stationary phase must be chosen for efficient separation of components. The stationary phase is adhered to the inside wall of the column and can vary in chemical composition. The stationary phase is selected based on the chemistry of the analytes to be separated (e.g., polar or non-polar molecules). In controlled substance analysis, typically a nonpolar stationary phase is used, meaning it is hydrophobic, (e.g., 5% diphenyl 95% dimethylpolysiloxane). Analytes partition in and out of the stationary phase, and components that are strongly retained, via strong intermolecular forces, by the stationary phase will move slowly through the column, while those that have weak intermolecular force bonding with the stationary phase travel rapidly. For example, in order to separate a mixture of non-polar compounds, the analyst would choose a non-polar stationary phase because like compounds are attracted to one another, via the intermolecular forces with the stationary phase. As a consequence of these differences in migration rates, the chemical components of the mixture are separated into different bands or peaks.1 17 To aid in separation of similar components, as well as to keep the sample in the gas phase, the column is housed inside an oven that can be temperature programmed. Because the retention of components is also dependent on their boiling points, the rate and how well they are separated from one another can be manipulated by the temperature program of the oven. On a non-polar stationary phase, the different components will interact with that phase based on their polarity and affinity for it. The components will be released from the stationary phase based on their boiling point as the temperature of the oven increases. A fast heating rate will allow a wider range of compounds to be separated rapidly, but the compromise may be lower resolution. Therefore, a compromise is needed to optimize the oven temperature program. Once the various components of the sample have traveled through the column and separated, they reach the detector. Various detectors such as flame ionization or electron-capture devices can be used, depending on the application. In this research, a mass spectrometer (MS) was used that was consistent with standard GC-MS use in forensic laboratories for controlled substance identification. The detector’s primary purpose is to detect the separated components of a mixture, and translate that information to a visual output. It is typically represented as a chromatogram (Figure 2.2), where retention time is on the x-axis and signal abundance is on the y-axis. Each peak represents a different separated component of the mixture. In forensic laboratories, chromatograms are used for controlled substance identification by analyzing both the questioned sample and a reference standard under the same conditions. The retention time and peak shape of the reference standard are then compared to the suspected controlled substance. If the retention times of the sample and reference standard are consistent (within ± 2 sec.), it is contributing evidence toward an identification of the substance.2 18 Figure 2.2 Example chromatogram of a multi-component gas chromatography separation 2.2.1 Retention Index As previously discussed, retention time can be used as contributing information toward the identification of an unknown compound, but it is dependent on many different variants, such as the stationary phase, column length and diameter, and method parameters set by the user. For example, if a sample was analyzed on two different instruments of the same make and model, using the same method and program with the same stationary phase in the column, the retention times may still vary slightly due to minor differences in the injection volume, injection speed, syringe dwell time in the injection port, or flow rate of the carrier gas. However, the retention time can be used to calculate the retention index (IT) of a compound. The IT is independent of variables such as column dimensions, stationary phase thickness, flow rate, and temperature program. By definition, the retention index of a compound is its retention time normalized to the retention times of adjacently eluting normal-alkanes.3 Retention index is advantageous because it allows the comparison of an adjusted retention time for a compound analyzed on different instruments under varying conditions. Retention index is calculated using Equation 2.1 19 IT = 100 × [n + tr(unknown) - tr(n) tr(n+1) - tr(n) ] (2.1) where IT is the retention index, n is the number of carbons in the smaller alkane (the one eluting before the questioned sample), and tr is the retention time. 2.3 Mass Spectrometry Overview For the detection of the samples, mass spectrometry was used in this research. Mass spectrometry can give definitive identification of a compound by ionizing it and separating those ions according to their mass-to-charge (m/z) ratios. This mass information can be used to identify an unknown substance by comparing its information to a library of knowns, a reference standard, or deducing its elemental composition. A mass spectrometer is comprised of three parts: the ionization source, mass analyzer, and detector (Figure 2.3). All three of these components are under vacuum (except the signal processor and readout) to ensure there are no unwanted collisions between ions and to maintain free ion and electron flow. An under-vacuum system ensures reproducible results and no contamination with air molecules. Ionization source Inlet from GC Mass analyzer Detector Output Vacuum pump Figure 2.3 Overall schematic of mass spectrometer (MS) The ionization source is responsible for ionizing the gaseous output from the GC. From the GC, separated components enter the ion source via a heated transfer line (e.g., 280 °C) and are ionized. The transfer line must be sufficiently hot so that the separated components do not condense out of the gas phase. There are many different types of ionization, however in this research, electron ionization (EI) was used. Because samples are bombarded with a significant 20 amount of energy, EI is classified as a “hard” ionization method and results in few intact molecules and extensive fragmentation. Electrons are emitted from a heated tungsten filament in the ionization source and are accelerated by applying 70 eV between the filament and anode (Figure 2.4). As the sample gas travels through the repeller plate, it enters the path of electrons at a 90° angle and through electrostatic repulsion, loses electrons to become positively charged ions. These ions are then directed through the negatively charged focusing lens into the mass analyzer. The mass analyzer is used to separate the newly created ions based on their m/z ratios and send them to the detector. Two types of mass analyzers were used in this research, single quadrupole and time-of-flight mass analyzers. Figure 2.4 Electron ionization (EI) source 2.3.1 Mass Analysis: Single Quadrupole Mass Analyzer A single quadrupole mass analyzer is one of the most common types of analyzers because it is rugged and relatively inexpensive. It consists of four cylindrical rods that are parallel to one another, as shown in Figure 2.5. Ions are filtered through the quadrupole based on the stability of their trajectories as they travel through the oscillating electric fields that are applied to the rods. 21 A radio frequency (RF) voltage and a direct current (DC) offset voltage are applied between each opposing rod pair and only ions of a specific m/z ratio will reach the detector for a given ratio of voltages. Other ions will be unstable, collide with the rods, and be pumped away. The ratio of voltages allows for selection of an ion with a particular m/z value, or allows the user to scan for a range of m/z values by continuously varying the applied voltage while monitoring the RF/DC ratio. Figure 2.5 Quadrupole mass analyzer showing two different ion trajectories occurring simultaneously. The red ion is neutralized as it collides with one of the rods, is pumped away, and not detected, while the blue ion has a stable trajectory through the analyzer and travels to the detector Ions are then attracted to the detector, via its negative electric charge, which converts the ions into an electrical signal that can be processed, stored in the memory of a computer, and displayed.1 The most common detector is a continuous-dynode electron multiplier, which collects, amplifies, and converts positive ions into electrical signal. The electron multiplier is a cornucopia-shaped horn (known as a Woods horn), connected to a power source, with a negatively charge entrance (i.e., -2 kV), and increasingly positively charged gradient walls. As 22 positively charged ions enter the horn, they collide with the wall and are converted to secondary electrons. The secondary electrons are attracted along the positive electrical gradient farther along the Woods horn, and each time they collide with the wall, additional secondary electrons are ejected.1,4 The electrons are then converted to a voltage, the magnitude of which is indicative of the abundance of each ion. From quadrupole mass spectrometry (QMS), nominal mass information about the original molecule is obtained because it is a low-resolution mass analyzer. Nominal mass is defined as the integer mass of the most abundant, naturally occurring stable isotopes of a molecule.4 The ability of the mass analyzer to distinguish between ions of similar m/z value is defined by its resolving power.4 For example, low-resolution mass spectrometers are typically only able to distinguish ions that are 1 mass unit (Da) apart. A limitation of nominal mass resolution is that an elemental formula for each fragment ion cannot be confirmed. For example, many formulae are possible for an ion at m/z 204 (i.e., C10H20O4 and C10H22NO3). Resolution in mass spectrometry is defined by Equation 2.2 using two adjacent peaks in a mass spectrum, R= M ∆m (2.2) where M is the mass of the first peak and Δm is the difference between the masses of the adjacent peak. A larger resolution value indicates better separation of peaks. To measure the minimum peak separation, Δm, and thus, the resolution, the peak width is measured at half of the peak maximum (FWHM).4 In low-resolution mass analysis, resolution is typically on the order of 102. 23 Higher resolution can be desirable as it indicates better discrimination between two adjacent peaks, and allows for accurate mass measurement. Accurate mass measurement is the mass measurement performed to a sufficient number of significant figures to allow for unambiguous determination of an elemental composition, as is obtained with the time-of-flight mass analyzer.4 2.3.2 Mass Analysis: Time-of-Flight Mass Analyzer There are instances, such as within this research, where higher resolution is necessary and can be achieved by high-resolution mass spectrometry. These instruments include mass analyzers such as the time-of-flight (TOF) mass analyzer which yields exact mass to four decimal places. Exact mass is defined as the most abundant naturally occurring stable isotope of an element, also called its monoisotopic mass.4 Instruments like the TOF have resolution on the order of 103 - 105. In TOFMS, the time required for an ion to travel from the ion source to the detector is measured. Because all of the ions from the source are accelerated with the same energy, ions travel at different velocities based on their differing m/z values. As the ions travel through the analyzer, they are separated into different groups according to these velocities. For example, ions of lower m/z value have higher velocity and reach the detector before ions of higher m/z value. To increase resolving power of ions with similar m/z value, a reflectron-TOF was used in this research. An ion mirror, in the form of an electric field with greater and opposite magnitude than the electric field in the acceleration region, is positioned at an angle less than 180° to direct ions toward the detector but not allow them to travel back toward the source.4 Ions with similar m/z values but different energies take longer or shorter flight paths through the reflectron, thus 24 reaching the detector at different times. A schematic of a reflectron-TOF analyzer is shown in Figure 2.6. Detector Pusher From ion source Reflectron Figure 2.6 Time-of-flight mass analyzer showing two different ion trajectories occurring simultaneously. Both ions are accelerated in the pusher region with the same kinetic energy, but the red ion penetrates deeper into the reflectron because it has larger mass, thus reaching the detector after the blue ion After mass analysis, a spectrum is produced with resolution of ions that differ by less than 1 Da and a mass accuracy, given in ppm, is assigned to each. Mass accuracy is a measure of the error of measurement of the mass of the ion (Equation 2.3); therefore, good mass accuracy is represented as a small, positive or negative value (i.e., the closer to zero, the better the mass accuracy).5 Mass accuracy (ppm) = Theoretical exact mass-Measured exact mass Theoretical exact mass × 106 (2.3) Using the accurate mass information, the exact mass of each ion can be assigned an elemental formula, leading to definitive identification of those ions. For example, an ion with the elemental 25 formula assignment of C10H15NO2 would have a theoretical exact mass of 181.110279 Da. If the measured mass of that ion was 181.1104 Da, the associated mass accuracy value would be -0.6 ppm via Equation 2.3. Exact mass from high-resolution mass spectrometry has the potential to overcome limitations of nominal mass resolution data because of the definitive identification that can be obtained. 2.3.3 Comparison of Low-Resolution and High-Resolution Mass Spectra In mass spectrometry, after ionization, mass analysis, and detection of the ions, an output is given in the form of a mass spectrum (Figure 2.7). The x-axis represents m/z value, while the y-axis represents ion intensity. In both the low-resolution spectra, generated by QMS, and the high-resolution spectra, generated by the TOFMS, the intact, positively charged molecule, called the molecular ion (M+), is present at m/z 181 for the compound 2C-H. In the low-resolution spectrum (Figure 2.7A), nominal mass values are associated with each ion (i.e., m/z 181, 152, 137, etc.). In the high-resolution spectrum (Figure 2.7 B), exact mass values are obtained to the fourth decimal place (i.e., m/z 181.1104, 152.0833, 137.0601, etc.), along with a mass accuracy (i.e., 0.6 ppm, 2.6 ppm, 1.5 ppm, etc.) (Section 2.3.2). From the exact and accurate mass, an elemental formula can be discerned for each ion (i.e., C10H15NO2, C9H12O2, C8H9O2, etc.) and structural arrangement can be proposed with confidence. Each peak in the spectra represents a positively charged fragment of the molecular ion after it has been bombarded by electrons during ionization (Section 2.3). The remaining, uncharged part of the molecule that is pumped away is known as a neutral loss. In a spectrum, the difference in mass between the molecular ion and a fragment ion corresponds to the neutral loss. Elementally, the neutral loss can be identified by taking the chemical formula of the 26 molecular ion and subtracting the chemical formula of the fragment ion. For example, the molecular ion of 2C-H (Figure 2.7) has a chemical formula of C10H15NO2 and the chemical formula of the ion at m/z 152.0833 is C9H12O2, resulting in a neutral loss of CH3N. This research investigates if compounds of similar structural classes fragment similarly and exhibit any common neutral losses. A) 2C-H 152 100 O Abundance (%) NH2 O 137 91 M+ 181 121 0 0 100 200 300 m/z O B) 100 152.0833 137.0601 C 9H12O2 + C C 8H9O2 2.6 ppm 1.5 ppm H O + CH2 O O 121.0645 C 8 H9 O 6.6 ppm O + Abundance (%) CH2 M+ 181.1104 C10H15NO2 0.6 ppm 91.0543 C 7 H7 5.5 ppm + CH2 0 0 100 200 300 m/z Figure 2.7 Spectra and chemical structure of 2, 5- dimethoxyphenethylamine (2C-H) via (A) low-resolution (QMS) and (B) high-resolution (TOFMS) mass spectrometry 27 2.4 Mass Defect Once the exact mass of an ion is obtained using high-resolution mass spectrometry, it can be used to calculate other characteristics of the ions, such as mass defects. Different amounts of energy are released by every elements’ isotope upon binding and stabilizing of its nucleus, called nuclear binding energy. Absolute mass defect, commonly referred to as just mass defect, is the difference in binding energy between every isotope to carbon-12, either positive or negative (Table 2.1). Because each isotope has a different mass defect, each molecule of different elemental composition will have a unique exact mass.6 Absolute mass defect is calculated by Equation 2.4 Absolute mass defect = Exact mass - Nominal mass (2.4) For example, the absolute mass defect of 2C-H would be its exact mass (181.1104 Da) minus its nominal integer mass (181 Da) to yield a defect of 0.1104 Da. Table 2.1 Absolute mass defects of elements commonly used in this research Element Nominal Mass (Da) Most Abundant Isotope Mass (Da) Absolute Mass Defect (Da) C 12 12.00000 0.00000 H 1 1.007825 0.00782 N 14 14.003074 0.00307 O 16 15.994915 -0.00508 Cl 35 34.968853 -0.03115 Br 79 78.918336 -0.08166 F 19 18.998403 -0.00160 I 127 126.904477 -0.09552 S 32 31.972072 -0.02793 28 Mass defect can be a useful tool for the characterization of unknowns. Mass defect ranges (called filters) can be created for classes of compounds in which the mass defect of an unknown could fall within or outside of, indicating potential class characterization. However, the limitation with absolute mass defect is that as mass increases, so too does the absolute mass defect. Therefore, as compounds of increasing mass are added to the filter, it will continue to become wider and less specific. 2.4.1 Kendrick Mass Defect A second type of mass defect, called Kendrick mass defect, can be used for characterizing compounds of the same homologous series. A Kendrick mass is first calculated by normalizing exact mass to the mass of a methylene (CH2) group via Equation 2.5. The difference between the new Kendrick mass and the nominal mass is calculated to obtain the Kendrick mass defect (KMD) of the ion (Equation 2.6).6 Kendrick mass = Exact mass × 14.00000 14.01565 KMD = Nominal mass - Kendrick mass (2.5) (2.6) Theoretically, compounds of the same homologous series, which differ only in the number of CH2 groups, will have the same Kendrick mass defect. For example, 2C-H has a theoretical KMD of 91.95 mDa while 2C-P, which has the 2C-H structure with an additional propyl group, also has a theoretical KMD of 91.95 mDa. KMD filters can be created for different classes of compounds, and have more specificity than absolute mass defect filters for characterization of unknowns. The filters are much narrower and based only on compounds within a homologous series, so the risk of false positive characterization is reduced. 29 REFERENCES 30 REFERENCES 1. Skoog DA, Holler FJ, Crouch SR. Principles of Instrumental Analysis. 6th ed. Belmont, CA: Thomas Brooks/Cole, 2007. 2. Virginia Department of Forensic Science. Controlled Substances Procedure Manual. 2016, 19, 64. 3. IUPAC Gold Book. https://goldbook.iupac.org/R05360.html (Accessed December 13, 2016). 4. Watson JT, Sparkman, OD. Introduction to Mass Spectrometry. 4th ed. Wiley, 2007. 5. Agilent Technologies. Mass Accuracy and Mass Resolution in TOF MS. https://www.researchgate.net/file.PostFileLoader.html?id=567901175cd9e3a6cc8b4571& assetKey=AS%3A309368938008576%401450770710216. (Accessed February 20, 2017). 6. Sleno, L. The use of mass defect in modern mass spectrometry. J. Mass Spectrom. 2012, 47, 226-236. 31 III. Materials and Methods 3.1 Reference Standards Synthetic phenethylamine and cathinone reference standards spanning various structural subclasses were purchased from Cayman Chemical (Ann Arbor, MI). These included four aminopropyl benzofuran phenethylamines (APB) and eleven 2,5-dimethoxyphenethylamines (2C) (Figure 3.1), two 3,4,5-trimethoxyphenethylamines (3C), twelve N-benzyl phenethylamine analogs (NBOMe) (Figure 3.2), and two cathinones (Figure 3.3). Full chemical names for each compound are given in the Appendix for this chapter. Throughout the remainder of this work, all compounds will be referred to by their common abbreviations. All standards were prepared at a concentration of 1 mg/mL of methanol (ACS grade, Sigma-Aldrich, St. Louis, MO). For retention index determination, a mixture of normal (n-) alkanes was prepared using alkanes ranging from C12 – C28, and C30 (Sigma-Aldrich, St. Louis, MO). Each alkane was prepared to an approximate concentration of 13.5 mM in dichloromethane (ACS grade, EMD Millipore, Darmstadt, Germany). Typically, to determine retention index, the n-alkanes are spiked directly into the sample to be analyzed; however, this practice is not practical in a forensic crime laboratory setting. Therefore, the alkane mixture was analyzed independently at the beginning of the sample sequence, after every 10 sample injections, and at the end of the sample sequence. Retention indices were calculated using Equation 2.1. 32 A) 4-APB B) O 5-APB NH2 NH2 O C) 6-APB D) O NH2 O 7-APB NH2 E) O R1 NH2 R2 O Figure 3.1 Structures of the phenethylamines in the reference set (A) 4-(2-aminopropyl) benzofuran (4-APB) (B) 5-(2-aminopropyl) benzofuran (5-APB) (C) 6-(2-aminopropyl) benzofuran (6-APB) (D) 7-(2-aminopropyl) benzofuran and (E) the core structure of 2,5dimethoxyphenethylamine (2C-phenethylamines). The substituents at R1 and R2 and the corresponding 2C compound are given in Table 3.1. Table 3.1 Substituents for 2C-phenethylamine compound shown in Figure 3.1 Compound R1 R2 Compound R1 R2 2C-H -H -H 2C-B -H -Br 2C-D -H -CH3 2C-C -H -Cl 2C-G -CH3 -CH3 2C-I -H -I 2C-E -H -CH2CH3 2C-N -H -NO2 2C-P -H -CH2CH2CH3 2C-T -H -SCH3 2C-T-2 -H -SCH2CH3 33 A) Mescaline B) Escaline O O NH2 O NH2 O O O C) D) Mescaline-NBOMe O R1 O NH O R2 NH O O O O Figure 3.2 Structures of more of the phenethylamines in the reference set (A) 3,4,5-trimethoxybenzeneethanamine (mescaline), (B) 4-ethoxy-3,5-dimethoxy-benzeneethanamine (escaline) both 3C-phenethylamines, (C) the core of N-benzyl phenethylamine analogs (NBOMephenethylamines). The substituents at R1 and R2 corresponding NBOMe compound are given in Table 3.2, and (D) 3,4,5-trimethoxy-N-[(2-methoxyphenyl)methyl]-benzeneethanamine (mescaline-NBOMe) Table 3.2 Substituents for NBOMe-phenethylamine compound shown in Figure 3.2 Compound R1 R2 Compound R1 R2 25H-NBOMe -H -H 25I-NBOMe -H -I 25D-NBOMe -H -CH3 25N-NBOMe -H -NO2 25G-NBOMe -CH3 -CH3 25T-NBOMe -H -SCH3 25E-NBOMe -H -CH2CH3 25T-4-NBOMe -H -SCHCH3CH3 25B-NBOMe -H -Br 25T-7-NBOMe -H -SCH2CH2CH3 25C-NBOMe -H -Cl 34 A) Mephedrone B) 3-MEC O O NH NH Figure 3.3 Structures of cathinones in the reference set (A) 4-methylmethcathinone (mephedrone) and (B) 3-methylethcathinone (3-MEC) 3.2 Gas Chromatography-Mass Spectrometry (GC-MS) Analysis Three different gas chromatography-mass spectrometry (GC-MS) systems were used to analyze the reference set of standards: one low-resolution single quadrupole instrument (GCQMS) and two high-resolution time-of-flight instruments (GC-TOFMS). The GC-QMS consisted of an Agilent 6890N gas chromatograph coupled to an Agilent 5975C mass spectrometer with an Agilent 7683B injector (Agilent Technologies, Santa Clara, CA). The column was coated with a (5% diphenyl)-95% dimethylpolysiloxane (DB-5, Restek, Bellefonte, PA) stationary phase with dimensions 30 m x 0.25 mm x 0.25 µm. The injection temperature was 250 °C and a splitless injection was used. The injection volume was 1 µL. The carrier gas was ultra-high purity helium (Airgas, Radnor Township, PA) at a nominal 1 mL/min flow rate. The oven temperature program was as follows: 40 °C for 1 min, 20 °C/min to 280 °C with a final hold of 7 min. The transfer line temperature was 280 °C. Electron ionization at 70 eV was used, the ion source temperature was 230 °C and the mass analyzer temperature was 150 °C. The mass scan range was 35 – 550 u, with a scan rate of 2.83 scans/s. Retention index data were collected only using this instrument. 35 The GC-TOFMS used to analyze all APB compounds, 2C-B, 2C-D, 2C-E, 2C-G, 2C-H, 2C-P, 2C-T, escaline, mescaline, 3-MEC, and mephedrone was a Waters Micromass GCT Premier (Waters, Milford, MA), which consisted of an Agilent 6890N gas chromatograph coupled to a Waters GCT mass spectrometer with an Agilent 7683B autosampler. The same column dimensions and stationary phase (DB-5) as the GC-QMS analysis were used. The injection temperature was 210 °C and an appropriate split ratio injection was used per sample, ranging from splitless to 100:1. The injection volume was 1 µL. The carrier gas was ultra-high purity helium at a flow rate of 1.3 mL/min. The oven temperature program was as follows: 50 °C for 1 min, 15 °C/min to 280 °C with a final hold of 2 min. The transfer line temperature was 280 °C. Electron ionization at 70 eV was used and the ion source temperature was 180 °C while the mass analyzer was held at 130 °C. The scan range was 35 – 300 u and the rate was 5.00 scans/s. To ensure good mass accuracy, a constant infusion of perfluoro-tertbutylamine (PFTBA), a calibrant, was used during each sample analysis. The resolution of the instrument was 7,000 FWHM. The second GC-TOFMS analysis was used to analyze the remaining sample set compounds (i.e., all of the NBOMe compounds, 2C-C, 2C-I, 2C-N, and 2C-T-2) on a LECO Pegasus GC-HRT (LECO Corp., St. Joseph, MI) which consisted of an Agilent 7890N gas chromatograph coupled to a LECO Pegasus HRT mass spectrometer with a Gerstel MPS2 (GERSTEL, Inc., Linthicum Heights, MD) autosampler. The column was coated with 1,4bis(dimethylsiloxy)phenylene dimethyl polysiloxane (Rxi-5sil ms) stationary phase and dimensions of 20 m x 0.18 mm x 0.18 µm (Restek, Bellefonte, PA). The injection temperature was 250 °C and a 100:1 split injection was used due to the high sensitivity of the instrument. The injection volume was 1 µL. The carrier gas was ultra-high purity helium at a flow rate of 0.85 36 mL/min. The oven temperature program was as follows: 60 °C for 0.5 min, 36 °C/min to 340 °C with a final hold of 4 min. The transfer line temperature was 300 °C. Electron ionization at 70 eV was used and the ion source temperature was 250 °C. The scan range was 35 – 510 u and the rate was 10 scans/s. To ensure good mass accuracy, again a constant infusion of PFTBA was used during each sample analysis. The resolution of this instrument was 50,000 FWHM. Although a different column and GC conditions were used for these samples, it does not affect the mass spectra, which was used for all data processing and analysis. 3.3 Data Processing Low-resolution mass spectra were generated by taking a single scan at the apex of the chromatographic peak in the total ion chromatogram after GC-QMS analysis. All spectra were exported from ChemStation (Agilent Technologies) into Microsoft Excel (Microsoft, Albuquerque, NM). All low-resolution spectra were plotted in Origin (version 8.6, OriginLab Corp., Northampton, MA) to generate spectra of publication quality. The high-resolution spectra obtained from the Waters GCT Premier were generated by taking scans within the peak in the total ion chromatogram and subtracting these from a range of scans in the baseline region immediately before the peak using MassLynx (version 4.1, Waters). The range of scans in the baseline region represented the baseline condition from that sample run and contained background ions at m/z 281, 207, and 73 as well as ions from the calibrant at m/z 218, 131, and 69. The mass accuracy of each ion in the background-subtracted mass spectra was assessed using the elemental composition algorithm in MassLynx. The elemental composition function tabulates the mass accuracies of each ion against a list of potential elemental formulae and assigns them with a given mass accuracy value (in ppm) according to the ion’s exact mass. The potential formulae are restricted by a user-defined tolerance (i.e., 50 ppm mass accuracy) 37 and by user-defined values of the number of each possible element. For this work, the tolerance was defined as 50 ppm mass accuracy and the reference standard’s known elemental formula was used to define the number of carbon, hydrogen, nitrogen, oxygen, bromine, chlorine, iodine, and sulfur elements. Mass spectra containing ions with mass accuracies within ± 20 ppm, were considered acceptable and were exported to Microsoft Excel (version 12.0, Microsoft Corp., Redmond, WA) for further processing. In Microsoft Excel, the ion abundancies were normalized to that of the base peak. The m/z and normalized abundancies were then plotted in Origin. The high-resolution spectra obtained from the LECO Pegasus-HRT were generated from the Peak True data processed files in the ChromaTOF (version 4.2.3.1, LECO Corp.) software. Peak True files include data processing such as background and calibrant ion subtraction. The spectra were generated by taking the scan at the apex of the total ion chromatographic peak in the ChromaTOF software. In a similar manner as previously described, element formulae and mass accuracies were assigned to each ion using the algorithm in the ChromaTOF software. Mass spectra containing ions with mass accuracies ± 10 mDa were considered acceptable and were also exported to Microsoft Excel for further processing. The abundance of each ion was normalized to the relative abundance of the base peak and the spectra were plotted in Origin. 3.4 Mass Defect Filters Only the high-resolution spectra were used to create mass defect filters. The exact masses, nominal masses and mass accuracies for all ions were tabulated in Microsoft Excel. 3.4.1 Absolute Mass Defect Filters The absolute mass defect of each compound was first calculated using Equation 2.4, and expressed in mDa. The absolute mass defect of all the compounds in the training set were then 38 averaged to obtain the centroid of the filter. A confidence interval was then calculated by the following equations: CI = SEM × t CL SEM = σ √n (3.1) (3.2) where CI is the confidence interval, SEM is the standard error of the mean, tCL is the t value for the specified confidence level, σ is the standard deviation, and n is the number of mass defects in the training set. Thus, the filter was represented as the centroid value with a given tolerance, expressed in the form of a confidence interval. All mass defect filters were calculated at commonly used confidence levels of 99.9, 99, 95 or 90%. As confidence levels increase, so does the width of the filter. The wider the filter, the higher likelihood that a compound will incorrectly fall within it, as a false positive. However, a filter that is too narrow will exclude compounds that should fall within it, as a false negative. Therefore, the confidence level is chosen to maximize the specificity of each filter. The mass defects of each test set compound were calculated in the same manner and plotted against the calculated filter to investigate the success of the absolute mass defect for correctly characterizing compounds according to structural subclass. 3.4.2 Kendrick Mass Defect Filters Kendrick mass defect (KMD) filters of molecular ions and fragment ions were calculated in a similar manner as described in Section 3.4.1. The Kendrick mass of each molecular or fragment ion was first calculated using Equation 2.5. The Kendrick mass was then subtracted from the nominal mass and expressed in mDa (Equation 2.6). Using the same method as 39 previously described, the average KMD of the training set was calculated as the centroid of the filter and a confidence level was calculated as the associated tolerance. The KMD filters were developed using the appropriate training set compounds and the test set compounds were used to test the success of the filter in correctly characterizing compounds according to structural subclass. 40 APPENDIX 41 APPENDIX: Compound abbreviations and full chemical names Table A.1 Compound abbreviations and chemical names Compound Abbreviation Full Chemical Name Compound Abbreviation Full Chemical Name 4-APB 4-(2aminopropyl)benzofuran 25H-NBOMe 2-(2,5-dimethoxyphenyl)-N(2methoxybenzyl)ethanamine 5-APB 5-(2aminopropyl)benzofuran 25D-NBOMe 2-(2,5-dimethoxy-4-methylphenyl)-N(2-methoxybenzyl)ethanamine 6-APB 6-(2aminopropyl)benzofuran 25G-NBOMe 2,5-dimethoxy-N-[(2methoxyphenyl)methyl]-3,4-dimethylbenzeneethanamine 7-APB 7-(2aminopropyl)benzofuran 25E-NBOMe 2-(4-ethyl2,5-dimethoxyphenyl)-N-(2methoxybenzyl)ethanamine 2C-H 2,5dimethoxyphenethylamine 25B-NBOMe 4-bromo-2,5-dimethoxy-N-[(2methoxyphenyl)methyl]benzeneethanamine 2C-D 2,5-dimethoxy-4methylphenethylamine 25C-NBOMe 2-(4-chloro-2,5-dimethoxyphenyl)-N(2-methoxybenzyl)ethanamine 2C-G 3,4-dimethyl-2,5dimethoxyphenethylamine 25I-NBOMe 4-iodo-2,5-dimethoxy-N-[(2methoxyphenyl)methyl]benzeneethanamine 2C-E 2,5-dimethoxy-4ethylphenethylamine 25N-NBOMe 2-(2,5-dimethoxy-4-nitrophenyl)-N-(2methoxybenzyl)ethanamine 2C-P 2,5-dimethoxy-4propylphenethylamine 25T-NBOMe 2,5-dimethoxy-N-[(2methoxyphenyl)methyl]-4(methylthio)-benzeneethanamine 2C-B 2,5-dimethoxy-4bromophenethylamine 25T-4-NBOMe 2,5-dimethoxy-N-[(2methoxyphenyl)methyl]-4-[(1methylethyl)thio]-benzeneethanamine 2C-C 2,5-dimethoxy-4chlorophenethylamine 25T-7-NBOMe 2,5-dimethoxy-N-[(2methoxyphenyl)methyl]-4-(propylthio)benzeneethanamine 2C-I 2,5-dimethoxy-4iodophenethylamine MescalineNBOMe 3,4,5-trimethoxy-N-[(2methoxyphenyl)methyl]benzeneethanamine 2C-N 2,5-dimethoxy-4nitrophenethylamine Mescaline 3,4,5-trimethoxy-benzeneethanamine 2C-T 2,5-dimethoxy-4methylthiophenethylamine Escaline 4-ethoxy-3,5-dimethoxybenzeneethanamine 2C-T-2 2,5-dimethoxy-4ethylthiophenethylamine 3-MEC 3-methylethcathinone Mephedrone 4-methylmethcathinone 42 IV. Characterization of Synthetic Phenethylamines by Low-Resolution Mass Spectrometry The fragmentation of various synthetic phenethylamines of different structural subclasses (Section 1.1) is described in this chapter. Characteristic features in the low-resolution mass spectra are identified that can be used for characterization. The two aims are (1) to understand fragmentation of compounds within a structural subclass, and (2) determine which of these fragments can be used to characterize unknowns, or more likely, new analogs of synthetic phenethylamines. Phenethylamines will be analyzed by gas chromatography – quadrupole mass spectrometry (GC-QMS) and chromatograms will be used to determine retention index, while spectra will be probed for characteristic ions to be used to identify phenethylamines. The spectra will be further probed to identify characteristic features of different structural subclasses. Knowing these features, a characterization scheme will be developed and application for characterization of “unknowns” will be demonstrated. 4.1 Retention Index Differentiation of isomeric compounds oftentimes poses a challenge for forensic analysts, particularly in cases where one isomer is controlled and the other is not. Mass spectra do not contribute much to distinguish isomers except relative ion ratios. For example, compounds 2C-G (not controlled) and 2C-E (controlled) are currently differentiated by the intensities of fragment ions at m/z 165 and 180. In the mass spectrum of 2C-G, m/z 165 has a higher abundance than m/z 180 and vice versa for the spectrum of 2C-E. However, distinction based only on ion ratios is challenging because instrument variability can affect the ratios. With GC-MS analysis, the chromatographic retention time can be used for isomer differentiation because isomers have different interactions with the stationary phase inside the GC column. Although not a current practice in forensic laboratories, retention index determination is easily implementable. By 43 calculating the retention index for each compound, there is the potential to distinguish isomers. Using Equation 2.1, the retention indices for each phenethylamine in this study were calculated (Table 4.1). As discussed, 2C-E and 2C-G, have different retention indices at 1706 and 1751, respectively, showing the utility of retention index in isomer differentiation. Additionally, retention index ranges can be determined for each structural subclass. The aminopropyl benzofuran (APB) subclass has a retention index range of 1499 to 1527. The 2,5dimethoxyphenethylamines (2C) subclass has a range of 1590 to 2000, while the N-benzyl phenethylamine analog (NBOMe) subclass has a range of 2475 to 2839. Because these ranges do not overlap, the retention index can be a useful first step in the characterization of a synthetic phenethylamine. Retention index data were not collected for some of the compounds in the sample set, as indicated by “-“ in Table 4.1. Table 4.1 Retention index and molecular ion determinations of sample set compounds Compound Retention Index Molecular Ion Compound CI [M+H]+ EI [M +] 176.12 175.1 25H-NBOMe Retention Index Molecular Ion CI [M+H]+ EI [M +] 2475 302.23 301.1 4-APB 1505 5-APB 1527 176.11 175.1 25D-NBOMe 2519 316.25 315.1 6-APB 1527 176.12 175.1 25G-NBOMe 2619 330.26 329.1 7-APB 1499 176.08 175.1 25E-NBOMe 2569 330.26 329.2 2C-H 1590 182.15 181.1 25B-NBOMe 2746 - 379.0 2C-D 1653 196.19 195.1 25C-NBOMe 2649 336.19 335.1 2C-G 1751 210.18 209.1 25I-NBOMe - - Not detect. 2C-E 1706 210.19 209.1 25N-NBOMe 2839 347.30 346.1 2C-P 1774 224.21 223.1 25T-NBOMe 2816 348.28 347.0 2C-B 1856 260.08 259.1 25T-4-NBOMe - - 375.1 2C-C 1770 216.13 215.0 25T-7-NBOMe - - 375.2 2C-I 1961 308.08 307.0 MescalineNBOMe - - 331.2 2C-N 2000 277.14 226.1 Mescaline - - 211.1 2C-T 1958 228.14 227.1 Escaline - - 225.2 2C-T-2 - - 241.1 3-MEC - - 191.1 Mephedrone - - 177.1 - Indicates data not collected. A (5% diphenyl)-95%dimethylpolysiloxane (DB-5) stationary phase was used for IT determination. Compound names can be found in Chapter III Appendix. 44 4.2 Electron Ionization Mass Spectra of Synthetic Phenethylamine Subclasses Representative low-resolution mass spectra collected using a single quadrupole mass spectrometer (GC-QMS) for each of the three phenethylamine subclasses of interest in this work, 2C-, APB-, and NBOMe-phenethylamines, are shown in Figure 4.1. Based solely on the lowresolution mass spectra, it is not possible to determine the exact fragmentation mechanism. However, based on known fragmentation of phenethylamines, structures for the dominant ions can be hypothesized.1-4 All phenethylamines are expected to have spectra with some degree of similarity. For example, α-β bond cleavage occurs among all three subclasses, splitting the aromatic ring from the amine chain. In the APB subclass, this results in a base peak of m/z 44 from the amine chain (C2H6N+) (Figure 4.1 A), and in the 2C and NBOMe subclasses, this cleavage results in a methoxy methylbenzene ion (C8H9O+) at m/z 121 (Figure 4.1 B and C). The fragmentation of compounds in the NBOMe subclass further supports the hypothesis of α-β bond cleavage, by the presence of a highly abundant ion at m/z 150, as the amine side of the compound after cleavage (C9H12NO+). Additionally, a common ion among the three subclasses is m/z 77, which corresponds to a positively charged benzene ring (C6H5+), however, this is not a phenethylaminespecific ion and would be present in any aromatic compound spectrum. 45 A) O 44 100 O B) NH2 152 100 O C) NH2 NH 121 100 O O 131 77 137 181 121 O 2C-H Abundance (%) Abundance (%) Abundance (%) 6-APB 150 91 270 175 0 0 0 100 200 300 301 0 0 100 200 m/z 300 0 100 200 m/z 300 m/z O O NH2 O 25H-NBOMe NH2 NH O m/z 175 + O CH2 C + O m/z 77 + CH2 O + H2C NH2 H O + O CH2 m/z 131 m/z 301 O m/z 181 C m/z 44 NH O + + CH2 C + m/z 270 m/z 91 O O O m/z 152 O m/z 137 + CH2 NH m/z 121 H2C + O m/z 150 m/z 121 Figure 4.1 Representative spectra of (A) 6-APB, (B) 2C-H, and (C) 25H-NBOMe and proposed structures for the most dominant fragment ions in each spectrum 46 Despite these similarities, the differences in the spectra are readily apparent due to differences among the structural subclasses. The spectrum of 6-APB (Figure 4.1 A) exhibits a base peak at m/z 44, a molecular ion at m/z 175, and prominent ions at m/z 131 and 77. Besides the ion at m/z 44 previously discussed, α-β bond cleavage also results in an ion at m/z 131 that consists of a benzofuran ring with a methyl group (C9H7O+). Compounds in the APB series are traditionally isomers of 6-APB, differing only in the position of the furan ring around the benzene ring. Therefore, the other APB compounds (4-APB, 5-APB and 7-APB) have very similar spectra although three out of the four can be distinguished from one another based on retention index (Section 4.1). Isomers 5-APB and 6-APB have the same retention indices, however could be distinguished upon further optimization of the GC temperature program, which was outside the focus of this work. The spectrum of 2C-H (Figure 4.1 B) has a base peak at m/z 152, corresponding to a positive radical dimethoxy-methylbenzene ion (C9H12O2+) without the amine chain, cleaved between the α and β carbons. The ion at m/z 137 is a positively charged dimethoxy benzene ring (C8H9O2+). As previously stated, the ion at m/z 121 (C8H9O+) is also present in the spectrum of 25H-NBOMe (Figure 4.1 C) and is the base peak. Other predominant NBOMe ions include m/z 91 (charged methylbenzene, C7H7+) and m/z 150 (C9H12NO+). The three predominant ions at m/z 150, 121, and 91, with m/z 121 as the base peak, are very characteristic of the NBOMe class and can be used to differentiate NBOMes from other compounds of similar mass. For example, 25GNBOMe and the popular cannabinoid XLR-11 have the same nominal mass of 330 Da, but can be differentiated by the presence of the characteristic m/z 91, 121, 150 peaks (Figure 4.2). All NBOMe compounds in the study exhibited these three peaks, with only slight variation in 47 A) 175000 O 121 NH O Abundance Abundance O 25G-NBOMe 150 298 91 180 0 50 100 150 200 250 300 m/z B) F O N XLR-11 Figure 4.2 Mass spectra of (A) 25G-NBOMe and the cannabinoid (B) XLR-115 which both have a molecular ion of m/z 330. NBOMes can be differentiated from cathinones using characteristic peaks at m/z 91, 121, and 150. XLR-11 spectrum obtained from Cayman Chemical 48 abundances between m/z 91 and 150 among the 12 compounds investigated. The molecular ion of 25H-NBOMe is observed in very low abundance (0.2%) at m/z 301. Finally, the ion at m/z 270 is proposed to be the NBOMe molecule without one of its methoxy groups (C17H20NO2+). Overall, the spectra are visually different and through mass spectral interpretation, structural subclass can be determined relatively easily. Although the APBs are readily distinguishable from the 2Cs and NBOMes, the class contains only isomeric compounds, so by low resolution spectra it is difficult to determine exact ring position, and thus differentiate the isomers within the class. Differentiation of 2C and NBOMe compounds is more challenging as they have some common fragments and a whole series of compounds with different substituents. Therefore, further investigation of the mass spectra must be done to distinguish these compounds. 4.3 Neutral Losses from Molecular Ion to Distinguish 2C- from NBOMe-Phenethylamines To distinguish the phenethylamine structural subclasses, particularly the 2C- and NBOMe-phenethylamines, neutral losses from the molecular ion (M+) can be investigated. A neutral loss is a fragment under ionization conditions that is lost as a neutral molecule. To look for neutral losses in a spectrum, the mass of the neutral loss in question is subtracted from M+ (Section 2.3.3). Spectra of 2C-phenethylamines and NBOMe-phenethylamines were assessed for neutral losses characteristic of each subclass. All 2C-phenethylamines in the sample set exhibit losses of 29 and 60 Da from their M+. A loss of 29 Da corresponds to the loss of CH3N, part of the amine side chain, and a loss of 60 Da corresponds to a loss of C2H6NO, part of the amine side chain and one of the methoxy groups. Most of the 2Cs in this study had fragments remaining after a loss of 29 Da as their base 49 peak, otherwise it was a prominent peak. If the neutral loss of 29 Da (CH3N) corresponds to the base peak or a highly prominent peak, this may be supporting evidence that an unknown is a 2Cphenethylamine. Figure 4.3 shows example mass spectra of 2C-H and 2C-B exhibiting these losses and shows how those losses occur structurally. The NBOMe-phenethylamines also exhibit characteristic neutral losses from their molecular ions. A loss of 31 Da corresponds to a loss of CH3O, one of the methoxy groups, and a loss of 149 Da from the molecular ion corresponds to a loss of C9H11NO, the dimethoxy benzene ring side of the structure after α-β cleavage. These losses may indicate an NBOMe compound as preliminary characterization of an unknown. As discussed previously in Section 4.2, further support of preliminary characterization is if the base peak is m/z 121, which is proposed to be a methyl-dimethoxy benzene ring (C8H9O+), and could be formed several different ways, thus causing that ion to be greater in abundance. Figure 4.4 shows the mass spectra of 25H-NBOMe and 25B-NBOMe and ions from the characteristic neutral losses. 50 H 100 A) O -29 Da 152 181 M+ + CH2 2C-H O NH2 O Abundance (%) C 9H12O2 Loss of: CH3N O C10H15NO2 M+ = 181.1 O + CH2 121 C8H9O Loss of: C 2H6NO + -60 Da M 181 0 B) 0 100 200 m/zH 100 O C9H11BrO2 Loss of: CH3N 300 -29 Da 230 259 M+ + 2C-B CH2 O NH2 Br Abundance (%) O Br O C10H14BrNO2 M+ = 259.1 O C8H8BrO Loss of: C 2H6NO + CH2 199 Br -60 Da 259 M+ 0 0 100 200 300 m/z Figure 4.3 Mass spectrum of (A) 2C-H and (B) 2C-B showing characteristic 2C neutral losses of 29 and 60 Da and the structures of the fragment ions remaining after each loss 51 A) 121 100 25H-NBOMe C9H12O2 Loss of: C9H11NO O H O NH + Abundance (%) O O 150 -149 Da 152 O C18H23NO3 M+ = 301.1 CH2 301 M+ O 91 NH -31 Da 270 301+ M O C + C17H20NO2 Loss of: CH3O 0 0 100 200 300 m/z 25B-NBOMe 121 100 B) O C9H11BrO2 Loss of: C9H11NO OH NH + O Br CH2 O Abundance (%) 150 C18H22BrNO3 M+ = 379 Br O -149 Da 230 M+ 379 O NH 91 Br -31 Da 348 379 M+ 0 0 100 200 300 O + C C17H19BrO2 Loss of: CH3O 400 m/z Figure 4.4 Mass spectrum of (A) 25H-NBOMe and (B) 25B-NBOMe showing characteristic NBOMe neutral losses of 31 and 149 Da and the structures of the fragment ions remaining after each loss, as well as common fragment ions (m/z 91, 121, 150) 52 4.4 Distinction and Identification of Common Substituents for 2C- and NBOMePhenethylamines Unlike the APB subclass, the 2C and NBOMe subclasses each contain a series of compounds that differ in substituents, primarily on the aromatic ring. These substitutions often include alkyl chains differing in number of carbons, sulfur or nitro groups, or halogens such as bromine (Figure 4.3), chlorine, and iodine. 4.4.1 Halogen Substitutions The presence of halogen substituents can often be determined by isotope ratios in the mass spectrum. These ratios occur due to the naturally occurring abundance of halogen isotopes. For example, Br has two naturally occurring isotopes: 79Br has 50.5% natural abundance and 81 Br has 49.5% natural abundance, which means either isotope can occur in a molecule with approximately equal probability.6 Spectra of molecules containing Br show characteristic patterns consisting of doublets spaced 2 Da apart (Figure 4.5) in approximately a 1:1 ratio. For example, in the spectrum of 2C-B (Figure 4.5 A), there are doublet peaks at m/z 259.1 and 261.1. These peaks represent the same fragment ion (C10H14BrNO2+) but the ion at m/z 259.1 contains 79 Br whereas the ion at m/z 261.1 contains 81Br. These doublets are observed for all fragment ions that contain Br. In the spectrum of 2C-B doublets are observed at m/z 199, 215, and 230. However, no doublets are observed for lower mass fragments (m/z 77.1, 91.1, or 105) because Br has been cleaved and the remaining ion does not contain it. Similarly, the spectrum of 25BNBOMe (Figure 4.5 B) has bromine-containing doublets at m/z 346, 229, and 198.9 and fragments that do not contain bromine at m/z 150, 121, or 91. Although in different structural 53 subclasses, 2C-B and 25B-NBOMe have similar isotope ratios due to the presence of Br, enabling determination of the substituent by the characteristic isotope pattern. A) 100 2C-B 230 O NH2 Br Abundance (%) O 215 77.1 91.1 105 259.1 261.1 199 0 0 B) 100 121 100 200 300 m/z O NH Abundance (%) O Br 150 O 25B-NBOMe 91 346 348 198.9 229 0 0 100 200 300 400 m/z Figure 4.5 Characteristic isotope pattern in mass spectra of compounds containing bromine, (A) 2C-B and (B) 25B-NBOMe 54 In a similar manner, chlorine can also be identified by its characteristic isotope pattern. Chlorine has two naturally occurring isotopes: 35Cl has 75.7% natural abundance and 37Cl has 24.3% natural abundance, which is approximately a 3:1 ratio.6 Spectra of molecules containing Cl show characteristic patterns consisting of doublets spaced 2 Da apart, in approximately the 3:1 ratio. For example, in the spectrum of 2C-C (Figure 4.6 A), the doublet of peaks at m/z 188 and m/z 186 represent the fragment ion C9H11ClO2+. However, the peak at m/z 186 includes 35Cl, whereas the peak at m/z 188 includes 37Cl. The characteristic 3:1 ratio of these ions, with the intensity of m/z 186 approximately 3 times that of m/z 188, is due to the natural abundance of Cl isotopes observed. Other fragment ions containing Cl can be observed at m/z 215, 171, and 155. Similarly, in the spectrum of 25C-NBOMe (Figure 4.6 B) the peak at m/z 348 has approximately a 3:1 ratio of intensity with the ion at m/z 346. Not all halogen substituents can be identified by isotope ratios. For example, iodine is monoisotopic and, hence, its presence cannot be determined by isotope ratios. However, the iodine in a compound can be identified by ions at m/z 126.9, corresponding to I+, and at m/z 127.9 corresponding to HI+. This may not be true in all cases, depending on the sensitivity of the GC-MS instrument, as these ions are usually observed at relatively low abundances. For example, the I+ and HI+ ions are observed in spectra of 2C-I and 25I-NBOMe (Figure 4.7 and 4.8, respectively), although the intensities of each ion is less than 1% of the base peak. 55 A) 100 O 186 2C-C NH2 Cl Abundance (%) O 188 171 215 155 0 0 B) 100 100 200 m/z 300 25C-NBOMe O NH O Abundance (%) Cl O 302 304 346 348 0 0 100 200 300 m/z Figure 4.6 Characteristic isotope pattern in mass spectra of compounds containing chlorine (A) 2C-C and (B) 25C-NBOMe 56 A) 2C-I 100 O NH2 Abundance (%) I O 0 0 B) 100 200 300 m/z Abundance (%) 10 126.9 127.9 0 120 125 130 135 m/z Figure 4.7 Full mass spectrum of (A) 2C-I and (B) expanded section of same spectrum to highlight I+ and HI+ ions 57 A) 25I-NBOMe 100 O NH O Abundance (%) I O 0 0 B) 100 200 300 400 m/z Abundance (%) 10 126.8 127.9 0 120 125 130 135 m/z Figure 4.8 Full mass spectrum of (A) 25I-NBOMe and (B) expanded section of same spectrum to highlight I+ and HI+ ions 58 4.4.2 Sulfur and Nitro Substitutions Sulfur is also observed as a substituent on synthetic phenethylamines but is problematic to identify in mass spectrometry. Although sulfur is not monoisotopic (32S occurs at 95% and 34S occurs at 4% natural abundance), the isotope ratio is inconsistently observed. An [M+2]+ ion can sometimes be observed at the 95:4 ratio, as is the case with 2C-T (Figure 4.9 A), where the molecular ion is m/z 227 and there is a low-abundant ion at m/z 229. However, this isotope pattern does not occur in all sulfur containing compounds, as seen in the mass spectrum of 25TNBOMe (Figure 4.9 B), where the molecular ion is m/z 347, but there is no corresponding ion at m/z 349. Further, fragments containing 34S are at such low abundance, they may be mistaken as noise, or attributed to isotope peaks from 13C. Additionally, sulfur ions would occur at m/z 32 and 34 which is below the typical scan range for mass spectrometry. Even if the mass scan range was expanded, the sulfur isotopes are not likely to be observed ions in EI-MS by themselves. Unfortunately, using low-resolution mass spectrometry, sulfur is not always identifiable as a substituent. Nitro (NO2) groups are also present as substituents on 2C and NBOMe compounds. There is no specific isotope pattern but the common mass spectrometry “nitrogen rule” can be used to indicate the presence of such a group. If a compound has an odd-mass M+, it contains an odd number of nitrogens. If a compound has an even-mass M+, it contains an even number of nitrogens. Phenethylamines typically contain one nitrogen, from the amine chain, meaning they will have a M+ with an odd mass. For example, M+ in 2C-H is at m/z 181 and M+ for 25HNBOMe is at m/z 301 (Figure 4.1 B and C). However, the spectra of 2C-N and 25N-NBOMe have even M+ of m/z 226 and m/z 346 (Figure 4.10), respectively, indicating an even number of nitrogens on each, due to the NO2 substitution. 59 A) 100 2C-T O Abundance (%) NH2 S O M+ 227 229 0 0 B) 100 100 25T-NBOMe 200 300 m/z O Abundance (%) NH O S O M+ 347 0 0 100 200 300 m/z Figure 4.9 Mass spectrum of (A) 2C-T and (B) 25T-NBOMe indicating inconsistent sulfur isotope pattern 60 A) 100 2C-N O NH2 Abundance (%) O2N O M+ 226 0 0 B) 100 100 25N-NBOMe 200 300 m/z O NH O Abundance (%) O2N O M+ 346 0 0 100 200 300 m/z Figure 4.10 Mass spectrum of (A) 2C-N and (B) 25N-NBOMe indicating M+ with an even mass that suggests an even number of nitrogens present 61 4.5 Scheme for Characterization of Synthetic Phenethylamines using Low-Resolution Mass Spectra From retention index and mass spectra interpretation, the APB structural subclass can be distinguished from 2C-phenethylamines and NBOMe-phenethylamines (Section 4.1 and Section 4.2). Further, distinction of 2C-phenethylamines from NBOMe-phenethylamines is possible based on characteristic neutral losses (Section 4.3). To some extent, identification of substituents (i.e., specific compounds in 2C or NBOMe subclasses) can be determined based on isotope patterns in the mass spectrum as well as the common “nitrogen rule” (Section 4.4). To be more useful in laboratories for unknown identification, a flowchart style characterization scheme was developed based on afore-mentioned features. The scheme is shown in Figure 4.11 and examples follow. The scheme consists of two parts. Part A (Figure 4.11) is designed to (1) distinguish APB from 2C and NBOMe subclasses and (2) distinguish 2C from NBOMe compounds. The second part of the scheme (Part B, Figure 4.12) is designed to identify a likely substituent (halogen, nitro) on 2C or NBOMe compounds. To theoretically determine if the core structure of the unknown is either 2C-H or 25HNBOMe, the mass of the halogen should be subtracted (35, 79, or 126.9 Da for Cl, Br, or I, respectively) and the mass of hydrogen (1 Da) should be added. If the new, adjusted, mass of M+ after subtraction of the substituent and addition of the hydrogen is 181 Da (i.e., M+ for 2C-H), the compound may be a 2C-phenethylamine. If the new mass of M+ is 301 Da (i.e., M+ for 25HNBOMe), the compound may be an NBOMe-phenethylamine. Similar to the halogens, if there is indication of a sulfur substituent, the mass of sulfur (32 Da) should be subtracted and the mass of CH2 (14 Da) should be added. The mass of a methyl group is used instead of hydrogen because of the position of sulfur within an alkyl chain. If the compound has been determined to have an 62 even M+, the mass of a nitro group (46 Da) should be subtracted and the mass of hydrogen should be added. 63 Part A 1 Is retention index available? IT between 1499 – 1527 suggests APB. IT between 1590 – 2000 suggests 2C. IT between 2475 – 2839 suggests NBOMe. APB Other 2 Is there an ion at m/z 131 >10% abundance relative to the base peak? Yes Consistent with an APBphenethylamine Does the spectrum have three dominant peaks at m/z 91, 121, and 150 with the base peak at m/z 121? No No Yes Not consistent with an APBphenethylamine Not consistent with an NBOMephenethylamine. Continue to Step 3. Consistent with an NBOMephenethylamine. Continue to Step 3. 3 Yes 3 Is there a molecular ion? *Can be confirmed by CI data Is there a molecular ion? *Can be confirmed by CI data No 4 Does the compound have common losses of 31 and 149 Da from the molecular ion? Is m/z 121 the base peak? Yes Consistent with an NBOMe-phenethylamine Continue to Part B. Yes 4 See NOTE and Part B Does the compound lose 29 and 60 Da in neutral losses from M+? Does it lose 29 Da from M+ as the base peak? No Not consistent with an NBOMephenethylamine. Yes Consistent with a 2Cphenethylamine Continue to Part B. No See NOTE and Part B No Not consistent with a 2C-phenethylamine. Figure 4.11 Characterization scheme for low-resolution mass spectra of synthetic phenethylamines to distinguish APB, 2C, and NBOMe subclasses 64 Part B 5 Is there a halogen, sulfur, or nitro group present?** Yes No If Br, Cl, or I are present, subtract the mass of the halogen (79, 35, 126.9 Da) from the molecular ion and add the mass of hydrogen (1 Da). If the compound has an even M+ subtract the mass of a nitro group (NO 2) (46 Da) and add the mass of hydrogen (1 Da). If S is present, subtract the mass of sulfur (34 Da) and add the mass of CH2 (14 Da). Is the adjusted mass 181 Da? Yes Is the adjusted mass 301? Yes The unknown is consistent with an NBOMephenethylamine. ** If Br is present, double peaks (doublets) of similar abundance will be present, spaced two mass units apart for higher mass fragments If Cl is present, a 3:1 abundance ratio will be present, spaced two mass units apart for higher mass fragments No The unknown is consistent with a 2Cphenethylamine. Consistent with an alkyl- or sulfursubstituted compound. If I is present, m/z 126.9 and m/z 127.9 should be present (I and HI, respectively) If S is present, a low-abundant ion two mass units higher than M+ should be present No The unknown is not consistent with an APB, 2C, or NBOMe-phenethylamine NOTE: If no molecular ion is confirmed: only halogens can be identified. Cannot replace mass of halogen/sulfur/nitro with mass of hydrogen/methyl Nitrogen rule: If the mass of the molecular ion is even, there is an even number of nitrogens present, or none at all. Ex: The M+ for 2C-H has one nitrogen and its m/z 181 is odd, indicating an odd number of nitrogens, while M+ for 2C-N which has two nitrogens is m/z 226, an even number. Figure 4.12 Characterization scheme for low-resolution mass spectra of synthetic phenethylamines to determine substituents on 2C- or NBOMe-phenethylamines 65 Example 1: 25B-NBOMe (Figure 4.5 B) Part A: 1. Is retention index available? Yes, the retention index is 2746. This retention index is within the retention index range identified for NBOMe-phenethylamines (2475 – 2839). 2. Does the spectrum have three dominant peaks at m/z 91, 121, and 150 with the base peak at m/z 121? Yes. The spectrum has all three prominent peaks (m/z 91, 121, 150) and the base peak is m/z 121. Therefore, the unknown is consistent with an NBOMe-phenethylamine. 3. Is there a molecular ion? Yes, the molecular ion was confirmed to be m/z 379 by EIMS. 4. Does the compound have common losses of 31 and 149 Da from the molecular ion? Is m/z 121 the base peak? Yes, the compound has an ion at m/z 348 (379 – 31 Da) and at m/z 230 (379 – 149 Da). The base peak is at m/z 121. This indicates the unknown is consistent with an NBOMe-phenethylamine. Part B: 5. Is there a halogen, sulfur, or nitro group present? Yes, bromine doublets are present. Doublets of similar intensity indicate the presence of Br. a. Subtracting the mass of Br (79 Da) from the M+ (m/z 379) and adding the mass of H (1 Da) equals a mass of 301 Da. If treated as an unknown, 25B-NBOMe would be correctly characterized as an NBOMephenethylamine with a bromine substituent. 66 Example 2: 3-methylethcathinone (3-MEC) (Figure 4.13) Part A: 1. Is retention index available? No, the retention index of 3-MEC was not available. 2. Does the spectrum have three dominant peaks at m/z 91, 121, and 150 with the base peak at m/z 121? No, the prominent peaks in the spectrum are at m/z 44.1, 72.1, and 91.1. 3. Is there a molecular ion? Yes, a molecular ion was confirmed to be m/z 191.1 by EIMS. 4. Does the compound lose 29 and 60 Da as neutral losses from M+? Does it lose 29 Da from M+ as the base peak? No. No ion was observed at m/z 162 (191 – 29 Da) and therefore, it was also not the base peak. There was a low abundant ion at m/z 131 (191 – 60 Da). Part B: 5. Is there a halogen, sulfur, or nitro group present? No evidence of halogens, sulfur, or nitro groups was observed. If treated as an unknown, 3-MEC would not be characterized as an APB or NBOMe. It cannot be determined if it would be characterized as a 2C compound. The fragment after the loss of 60 Da cannot be confirmed to be from a loss of C2H6NO (Section 4. 3) using the current instrumentation. High-resolution mass spectrometry would allow elemental formula assignment for this fragment, along with an accurate mass measure of the confidence in that elemental assignment. Using the current low-resolution flowchart, some subjectivity still remains because it would be at the analysts’ discretion whether or not to 67 preliminarily characterize 3-MEC as a 2C-phenethylamine, as one of the two characteristic 2C neutral losses is present. O 225000 NH 72.1 Abundance 3-MEC 44.1 91.1 119.1 0 0 50 100 150 200 m/z Figure 4.13 Mass spectrum and structure of cathinone, 3-methylethcathinone (3-MEC) Example 3: Mescaline (Figure 4.14): Part A: 1. Is retention index available? No, the retention index of mescaline was not available. 2. Does the spectrum have three dominant peaks at m/z 91, 121, and 150 with the base peak at m/z 121? No, these ions were not prominent in the mass spectrum. 3. Is there a molecular ion? Yes. The molecular ion was confirmed to be m/z 211.1 by EI-MS. 68 4. Does the compound lose 29 Da and 60 Da in neutral losses from M+? Does it lose 29 Da from M+ as the base peak? The compound does lose both 29 and 60 Da, m/z 182 and 151, respectively, with the loss at 29 Da as the base peak. This indicates the unknown is consistent with a 2C-phenethylamine. Part B: 5. Is there a halogen, sulfur, or nitro group present? No evidence of halogens or nitro groups was observed. If treated as an unknown, mescaline would be incorrectly characterized as a 2Cphenethylamine. O 500000 NH2 182 O O Abundance Mescaline 211 151 0 0 100 200 m/z Figure 4.14 Mass spectrum of 3C phenethylamine, mescaline, which would be mischaracterized as a 2C because of its loss of 29 Da (m/z 182) and 60 Da (m/z 151) 69 4.6 Summary A characterization scheme has been designed to be immediately implementable into forensic laboratories as a “quick and easy” guide for preliminary characterization of unknowns. Through retention index determination, mass spectral investigation, and neutral loss determination, three phenethylamine structural subclasses can be differentiated. Additionally, some substituent identification and isomer differentiation is possible. However, some limitations have been highlighted using the current instrumentation. Without definitive identification of the fragment element compositions, 2C- and 3C-phenethylamines cannot be differentiated, and some subjectivity remains in differentiating cathinones from phenethylamine compounds. 70 APPENDIX 71 APPENDIX: Low- Resolution Mass Spectra A) B) 100 100 4-APB O 5-APB NH2 Abundance (%) Abundance (%) NH2 O 0 0 0 100 200 300 0 100 m/z 200 300 m/z C) O 7-APB 100 Abundance (%) NH2 0 0 100 200 300 m/z Figure A.1 Low-resolution mass spectra of (A) 4-(2-aminopropyl)benzofuran (4-APB), (B) 5-(2aminopropyl)benzofuran (5-APB), and (C) 7-(2-aminopropyl)benzofuran 72 A) 2C-D 100 B) 2C-E 100 O O NH2 Abundance (%) Abundance (%) NH2 O O 0 0 0 100 200 300 0 100 m/z C) 2C-G 100 200 D) 2C-P 100 O O NH2 NH2 Abundance (%) Abundance (%) 300 m/z O 0 O 0 0 100 200 300 m/z 0 100 200 300 m/z Figure A.2 Low-resolution mass spectra of (A) 2,5-dimethoxy-4-methylphenethylamine (2C-D), (B) 2,5-dimethoxy-4-ethylphenethylamine (2C-E), (C) 3,4-dimethyl-2,5dimethoxyphenethylamine (2C-G), and (D) 2,5-dimethoxy-4-propylphenethylamine (2C-P) 73 2C-T-2 100 O NH2 Abundance (%) S O 0 0 100 200 300 m/z Figure A.3 Low-resolution mass spectra of 2,5-dimethoxy-4-ethylthiophenethylamine (2C-T-2) 74 A) 25D-NBOMe 100 O NH Abundance (%) O O 0 0 100 200 300 m/z B) 25E-NBOMe 100 O NH Abundance (%) O O 0 0 100 200 300 m/z Figure A.4 Low-resolution mass spectra of (A) 2-(2,5-dimethoxy-4-methylphenyl)-N-(2methyoxybenzyl)ethanamine (25D-NBOMe) and (B) 2-(4-ethyl-2,5-dimethoxyphenyl)-N-(2methoxybenzyl)ethanamine (25E-NBOMe) 75 A) 25T-4-NBOMe B) O 100 25T-7-NBOMe O 100 NH NH O S O Abundance (%) O Abundance (%) O S 0 0 0 100 200 300 400 0 100 200 m/z 300 400 m/z C) Mescaline-NBOMe 100 O NH O Abundance (%) O O 0 0 100 200 300 m/z Figure A.5 Low-resolution mass spectra of (A) 2,5-dimethoxy-N-[(2-methoxyphenyl)methyl]-4[(1-methylethyl)thio]-benzeneethanamine (25T-4-NBOMe), (B) 2,5-dimethoxy-N-[(2methoxyphenyl)methyl]-4-(propylthio)-benzeneethanamine (25T-7-NBOMe), and (C) 3,4,5trimethoxy-N-[(2-methoxyphenyl)methyl]-benzeneethanamine (mescaline-NBOMe) 76 A) Escaline 100 O NH2 O Abundance (%) O 0 0 100 200 300 m/z B) Mephedrone 100 O Abundance (%) NH 0 0 50 100 150 200 m/z Figure A.6 Low-resolution mass spectra of (A) 4-ethoxy-3,5-dimethoxy-benzeneethanamine (escaline) and (B) 4-methylmethcathinone (mephedrone) 77 REFERENCES 78 REFERENCES 1. Chu, F. Improving Methods for the Analysis of Controlled Substances. Masters Thesis, Michigan State University, East Lansing, 2015. 2. Zuba, D.; Sekula, K. Identification and characterization of 2,5-dimethoxy-3,4-dimethylβ-phenethylamine (2C-G) – A new designer drug. Drug Test. Analysis. 2013, 5, 549-559. 3. Chen, B. et. al. A general approach to the screening and confirmation of tryptamines and phenethylamines by mass spectral fragmentation. Talanta. 2008, 74, 512-517. 4. Awad, T; DeRuiter, J.; Clark, C. R. GC-MS Analysis of Ring and Side Chain Regioisomers of Ethoxyphenethylamines. J. Chromatogr. Science. 2008, 46, 675-679. 5. XLR-11. Cayman Chemical. https://www.caymanchem.com/product/11565 (accessed December 1, 2016). 6. Reusch, William. Michigan State University. https://www2.chemistry.msu.edu/faculty/reusch/virttxtjml/spectrpy/massspec/masspec1.h tm (accessed October 5, 2016). 79 V. Characterization of Synthetic Phenethylamines by High-Resolution Mass Spectrometry Limitations of the characterization scheme for low-resolution data were highlighted at the end of Chapter IV, such as the inability to distinguish 2C- from 3C-phenethylamines (Section 1.1), and the inconclusive characterization of cathinones. Additionally, the elemental composition of each fragment ion remaining after neutral losses could not be determined with a high degree of certainty. Overall, nominal mass data were not sufficient for definitive identification of structurally similar compounds, therefore a new approach is necessary. Highresolution mass spectrometry measures the accurate mass of each ion, from which elemental formulae can be assigned with a high degree of confidence. This leads to a better understanding of the fragmentation of the phenethylamine compounds. High-resolution mass spectrometry also enables the exploitation of the mass defect that can be investigated as a tool for characterizing new analogs. In this chapter, the comparison of low-resolution and high-resolution spectra will first be discussed, followed by the development of mass defect filters, and a discussion of their implementation into a high-resolution version of the characterization scheme. 5.1 Comparison of Low- and High-Resolution Mass Spectra The low- and high- resolution spectra of 6-APB, 2C-H, and 25H-NBOMe were compared to ensure consistency in electron ionization (EI) between the ionization sources of the two instruments (Figure 5.1). Although the high-resolution spectra have more peaks because they were generated on a more sensitive instrument, both the low-resolution (Figure 5.1 A) and highresolution (Figure 5.1 B) spectra display the same peak patterns, molecular ions, and base peaks for each compound. The same principles apply for mass spectral interpretation as discussed in Chapter IV, such as characteristic NBOMe peaks at m/z 91, 121, and 150, and substituent identification. However, with high-resolution mass spectrometry, the elemental formula for 80 25H-NBOMe (IT = 12475) 2C-H (IT = 1590) 6-APB (IT = 1527) O A) 44 100 NH2 O 152 100 O NH2 NH 121 100 O 131 137 77 91 175 0 181 121 O Abundance (%) Abundance (%) Abundance (%) O 100 200 300 100 175.0986 C11H13NO 6.3 ppm 0 100 200 m/z 100 300 300 121.0649 C 8 H9 O 0.68 ppm 100 181.1104 C 10H15NO2 0.6 ppm 91.0543 C 7 H7 5.5 ppm 200 m/z 0 0 0 91.0543 C 7 H7 1.29 ppm Abundance (%) 131.0506 C 9 H7 O 6.9 ppm 121.0645 C 8 H9 O 6.6 ppm Abundance (%) Abundance (%) 300 152.0833 C9H12O2 2.6 ppm 137.0601 C 8 H9 O 2 1.5 ppm 100 77.0382 C 6 H5 11.7 ppm 200 m/z 44.0488 C 2 H6 N 27.2 ppm 100 301 0 0 m/z B) 91 270 0 0 150 150.0916 C9H12NO 1.79 ppm 270.1495 C17H20NO2 2.55 ppm 0 0 100 200 m/z 300 0 100 200 300 m/z Figure 5.1 Comparison of (A) low-resolution and (B) high-resolution mass spectra for 6-APB (left), 2C-H (middle), and 25H-NBOMe (right) 81 every ion can be determined, leading to more confidence of the identity of each fragment ion. For example, 25H-NBOMe has the same three characteristic ions (m/z 91, 121 and 150, Figure 5.1 A) using high resolution mass spectrometry, at m/z 91.0453, 121.0649, and 150.0916 (Figure 5.1 B) but now the elemental formulae of each can be assigned as C7H7+, C8H9O+, and C9H12NO+, respectively, with high degrees of accuracy at 1.29, 0.68, and 1.79 ppm, respectively. The formulae for these ions confirm the structural fragment elucidation proposed in Chapter IV (Figure 4.1). The spectra of 2C-B (Figure 5.2) further highlight the similarities between low- and high-resolution spectra, showing consistent fragmentation and doublets due to the presence of bromine. The assigned elemental composition in the high-resolution spectrum confirmed the presence of bromine. It should also be noted that the retention index (IT) for these compounds is the same for high- and low-resolution instruments (Figure 5.1), as expected. 82 2C-B (IT = 1856) A) 80000 O 230 NH2 Br Abundance O 215 77.1 259.1 199 0 0 50 100 150 200 300 229.9938 C9H11BrO2 1.7 ppm m/z B) 250 100 Abundance (%) 214.9696 C8H8BrO2 9.7 ppm 77.0413 C 6 H5 28.6 ppm 198.9772 C 8H8BrO 6.5 ppm 259.0183 C10H14BrNO 2 9.7 ppm 0 0 50 100 150 200 250 300 m/z Figure 5.2 Comparison of (A) low-resolution and (B) high-resolution mass spectra for 2C-B. Dominant fragment ions are labeled and in (B) assigned element formulae and mass accuracies are given 83 5.2 Development of Mass Defect Filters Accurate mass data can be used not only to assign elemental formulae, but also to calculate the mass defect of each ion. Compounds in the same structural class should theoretically have similar mass defects because the core structure is consistent among analogs. Because the mass defect of the core structure has a larger contribution to the overall mass defect, the addition of various substituents should not change the overall mass defect substantially. Therefore, mass defect was used as a tool to characterize compounds according to structural class. To do this, mass defects of the molecular ions were calculated for phenethylamines and a filter was developed as the mean mass defect ± a given tolerance. The efficacy of this filter to characterize compounds as phenethylamines was then tested. Mass defects based on molecular ions were calculated for a test set of compounds and tested to determine if the mass defect was within the previously defined filter. 5.2.1 Absolute Mass Defect Filters for Phenethylamines Based on Molecular Ions A training set contained 16 phenethylamines that were randomly selected from the full sample set (Section 3.4.1). These included APB, 2C, and NBOMe compounds and the mass defects of their molecular ions (M+) were calculated. Table 5.1 shows the exact masses, mass accuracies, and mass defects for the M+ of all compounds in the training set. The mean mass defect represented by the training set was 142.4 mDa and the tolerance was calculated as a confidence interval at the 99.9991% confidence level. This confidence level was necessary to encompass the range of mass defects in the training set. Thus, the filter was defined as 142.4 ± 54.1 mDa and is shown graphically in Figure 5.3, where the yellow line represents the average 84 mass defect and the purple lines represent the upper and lower bounds of the filter. All the mass defects of the training set compounds fell within this filter. Table 5.1 Calculation of absolute mass defect molecular ion filter Training or Test Set Training Test Compound 2C-D  2C-E  2C-H  2C-P  2C-N  2C-T-2  4-APB  25C-NBOMe  25D-NBOMe  25E-NBOMe  25G-NBOMe  25H-NBOMe  25N-NBOMe  MescalineNBOMe  Escaline  Mescaline  * 99.9991% CL 2C-G  2C-B  2C-C  2C-I  2C-T  5-APB  6-APB  7-APB  25B-NBOMe  25T7-NBOMe  3-MEC  Mephedrone  Nominal Mass Exact mass mass defect (Da) (Da) (mDa) 195.1243 195 124.3 209.1404 209 140.4 181.1104 181 110.4 223.1573 223 157.3 226.0956 226 95.6 241.1134 241 113.4 175.0999 175 99.9 335.1188 335 118.8 315.1813 315 181.3 329.1962 329 196.2 329.1945 329 194.5 301.1654 301 165.4 346.1493 346 149.3 Mass Filter accuracy (mDa)* (ppm) 8.2 5.7 0.6 0.4 3.53 1.1 1.1 28.25 5.01 142.4 54.1 7.21 12.38 5.96 8.76 331.1755 331 175.5 6.84 225.1351 211.1207 225 211 135.1 120.7 6.2 0.5 209.1421 259.0203 215.0710 307.0067 227.0988 175.1005 175.0986 175.0993 379.0602 375.1807 191.1310 177.1150 209 259 215 307 227 175 175 175 379 375 191 177 142.1 20.3 71.0 6.7 98.8 100.5 98.6 99.3 60.2 180.7 131.0 115.0 2.4 0.37 1.06 0.89 3.50 4.6 6.3 2.3 46.33 14.80 0.0 2.3  Compounds analyzed on Waters system that measures mass accuracy to one decimal place Compounds analyzed on LECO system that measures mass accuracy to two decimal places (Section 3.3)  85 Absolute Mass Defect (mDa) 250 200 150 100 2C-C 50 25B-NBOMe 2C-B 0 165 215 265 2C-I 315 365 m/z Training Set Phenethylamine Test Set Cathinone Test Set Figure 5.3 Absolute mass defect filter created using a training set of phenethylamines defined in Table 5.1. The absolute mass defect filter was defined at 142.4 ± 54.1 mDa at a 99.9991% confidence level. The horizontal lines represent the average (yellow), and the upper and lower bounds of the mass defect filter (purple) The filter was tested using the remaining compounds in the sample set, with the exception of 25I-NBOMe, 25T-NBOMe, and 25T-7-NBOMe, which did not exhibit molecular ions. The mass defect of 2C-G was 142.1 mDa and, hence, falls inside the filter. However, the halogenated compounds in the test set pose problems because of the large mass defect associated with the halogen (Section 2.4). For example, the mass defect associated with bromine is -81.6 mDa and, thus, has a significant impact on the mass defect of any fragment ion containing bromine. As a result, compounds with halogens have smaller mass defects than similar compounds that do not contain a halogen. For example, the mass defect of 2C-B is 20.3 mDa compared to 110.4 mDa for 2C-H, where the only difference is the presence of Br. Therefore, 2C- and NBOMe-phenethylamines with halogens are not correctly characterized using this filter. 86 More problematic, the absolute mass defects of the cathinone M+ from the test set also fall within the filter (Figure 5.3). For example, 3-methylethcathinone (3-MEC) and mephedrone have mass defects of 131.0 and 115.0 mDa, respectively. Although the cathinones are structurally similar, differing from the core structure of phenethylamine by an addition of a carbonyl group, there is a need to distinguish them from phenethylamines for a robust characterization scheme. The filter based on absolute mass defect of molecular ions shows potential, but the tolerance is too wide, resulting in a filter that is not sufficiently specific to distinguish phenethylamines from cathinones. Further, despite a large tolerance, phenethylamines containing halogens are not successfully characterized due to the large mass defect contribution from the halogen. In an effort to improve specificity of the filter, separate mass defect filters were developed for the three different phenethylamine structural subclasses. 5.2.2 Absolute Mass Defect Filter for the APB-Phenethylamine Subclass The training set for the APB mass defect filter contained 4-APB, 5-APB, and 6-APB. From the data in Table 5.1, the APB filter based on absolute mass defect of molecular ions was defined as 99.7 ± 1.6 mDa, at the 90% confidence level (Figure 5.4). This filter is very narrow because it was defined with a set of isomeric compounds. Theoretically, the exact masses and mass defects of isomers should all be the same, but because of the instrument variation, the experimentally collected exact masses vary slightly, as represented by the mass accuracies. The test set contained 7-APB, as well as the remaining 2C, NBOMe, 3C, and cathinone compounds from the sample set. None of these fall inside the filter, indicating correct characterization. 87 Absolute Mass Defect (mDa) 103 7-APB 99 95 165 175 185 195 205 215 225 235 245 m/z APB Training Set Test Set Figure 5.4 APB subclass absolute mass defect filter at 99.7 ± 1.6 mDa at a 90% confidence level. The horizontal lines represent the average (black) and the upper and lower bounds of the mass defect filter (red) 5.2.3 Absolute Mass Defect Filter for the 2C-Phenethylamine Subclass Because of large negative mass defect contribution of halogens to the mass defect of a compound, only 2C-phenethylamines with alkyl side chains were used in the training set (i.e., 2C-H, 2C-D, 2C-E, 2C-P). The 2C absolute mass defect filter was defined as 133.1 ± 32.2 mDa at the 95% confidence level. This tolerance is narrower compared to the full set of phenethylamines (Section 5.2.1), allowing for a more specific filter of the 2C subclass. The test set contained the remaining 2C-phenethylamines and all APB, NBOMe, 3C, and cathinone compounds. The compounds with a halogen, sulfur, or nitro group will fall outside the filter due to the mass defect contribution from the substituent (Section 2.4). However, like the method described in Section 4.4, halogens, sulfur, and nitro groups can be identified by isotope patterns and mass spectral features in the mass spectrum. Further, the mass defect of the suspected 88 halogen and nitro group can be replaced with that of hydrogen and then tested against the filter (Section 4.5). Although a sulfur substituent may not be able to be discerned from the mass spectral features (Section 4.4.2), high-resolution mass spectrometry offers the advantage of including sulfur during elemental formulae assignment for each fragment ion, and thus a sulfur substituent can be identified and further replaced with a CH2 group (Section 4.5). For halogen, nitro, and sulfur group replacement, the following exact masses are used to adjust the mass of the molecular ion: 34.9689 Da for Cl, 78.9183 Da for Br, 126.9045 Da for I, 45.9929 Da for NO2, 31.9721 Da for S, 1.0078 Da for H, and 14.0157 Da for CH2.1 This adjusted mass, when applicable, is used to calculate all mass defects. After halogen/sulfur/nitro group replacement, all the mass defects of the 2Cphenethylamines in the test set correctly fall within the 2C filter (Figure 5.5). However, the two cathinone and two 3C-phenethylamine test compounds also fall within the 2C filter, highlighting a lack of specificity despite the narrower tolerance associated with this filter. Furthermore, because there is no limit to the m/z range the 2C filter extends, it encompasses the APB filter, and significantly overlaps and encompasses many NBOMe compounds. 89 Absolute Mass Defect (mDa) 225 195 165 135 105 75 150 200 250 300 350 400 m/z Training Set 2C Test Set NBOMe Phenethylamines Cathinone & 3C Test Set APB Phenethylamines Figure 5.5 2C subclass absolute mass defect filter at 133.1 ± 32.2 mDa at a 95% confidence level. The horizontal lines represent the average (light blue) and the upper and lower bounds of the mass defect filters (dark blue) 5.2.4 Absolute Mass Defect Filter for the NBOMe-Phenethylamine Subclass Similar to the 2C absolute mass defect filter (Section 5.2.3), only NBOMephenethylamines with alkyl side chains were used in the training set (i.e., 25H-NBOMe, 25DNBOMe, 25E-NBOMe, and mescaline-NBOMe) to develop the filter. The NBOMe absolute mass defect filter was defined as 179.6 ± 20.5 mDa at the 95% confidence level. The test set contained the remaining NBOMe-phenethylamines, and all of the APB, 2C, 3C, and cathinone compounds. After halogen/nitro/sulfur group replacement, all the mass defects of the test NBOMe-phenethylamines correctly fell within the NBOMe filter except 25B-NBOMe and 25T7-NBOMe (Figure 5.6). The mass accuracy of the molecular ion of 25B-NBOMe was poor at 46.33 ppm, causing the mass defect (149.7 mDa) to fall outside the filter, despite substituting the halogen with hydrogen. Absolute mass defect has a positive correlation with mass, so 90 compounds of higher mass will have higher mass defects, as is the case with 25T-7-NBOMe, causing it to fall outside the filter, again despite replacing sulfur with CH2. As stated previously, a limitation of this filter is that it overlaps with the 2C filter and lacks specificity. Additionally, the theoretical mass defects of six of the most popular synthetic cannabinoids (JWH-018, JWH073, CP 47,497, AM-2201, UR-144, and XLR-11) were used to further test the specificity of the NBOMe filter because cannabinoids have similar molecular masses as many NBOMe- Absolute Mass Defect (mDa) phenethylamines. Three of the six cannabinoids would fall within the filter. 235 25T-7-NBOMe 195 155 25B-NBOMe 115 75 150 200 250 300 350 400 m/z NBOMe Training Set NBOMe Test Set 2C Phenethylamines APB Phenethylamines Cathinone & 3C Test Set Theoretical Cannabinoids Figure 5.6 NBOMe subclass absolute mass defect filter at 179.6 ± 20.5 mDa at a 95% confidence level. The horizontal lines represent the average (light purple) and the upper and lower bounds of the mass defect filter (dark purple) Although a good starting point for differentiation, the absolute mass defect filters based on the molecular ion have some limitations. The first is that based on mass defect alone, the filters overlap if the m/z ranges have no limit. However, the use of retention index can be used to overcome this limitation. Because the retention index ranges of APB, 2C, and NBOMe 91 subclasses are distinctly different, this information can be used to determine which filter to test the compound against. Second, mass defects of all the 3C-phenethylamines and cathinones fall within the 2C filter and many of the cannabinoid mass defects fall within the NBOMe filter. This highlights a lack of specificity when using the absolute mass defects of a molecular ion. To further investigate specificity, absolute mass defects of fragment ions and neutral losses common to each subclass were also investigated; however, these filters were still not sufficiently specific and the m/z and mass defect ranges overlapped. Because absolute mass defect filters were non-specific for distinguishing the structural subclasses, Kendrick mass defect filters were investigated. 5.2.5 Kendrick Mass Defect Filters for Phenethylamines Based on Molecular Ions To overcome the limitations of non-specific, overlapping, absolute mass defect filters, Kendrick mass defect (KMD) filters were developed, again based on molecular ions. Only alkylsubstituted phenethylamines were used in the subclass training sets to define the filters. Because Kendrick mass defects are used to identify members of a homologous series, differing only in the number of methyl (CH2) groups, compounds containing halogens, nitro groups, or sulfur are not members of this homologous series, and therefore were not used to create the filters. All compounds containing halogens or nitro groups had the masses of these substituents replaced with hydrogen, while compounds containing sulfur had the masses replaced with CH2 similar to Section 5.2.3, before calculating their KMD and being used to test the filter (Section 3.4.2). 92 5.2.6 Kendrick Mass Defect Filters of the APB-Phenethylamine Subclass Table 5.2 shows the KMD and the associated filter for the APB-phenethylamine subclass (Section 3.4.2).2 The APB KMD filter was determined to be 95.9 ± 1.6 mDa at the 90% confidence level. This tolerance is less than those used in defining absolute mass defect filters because the training set compounds should, theoretically, all have the same KMD. Thus, the filter should be significantly more narrow, and further, more specific. The test set contained the remaining APB-phenethylamines, all 2C, NBOMe, 3C, and cathinone compounds. The APB test set compound, 7-APB, had a KMD that fell within the filter (96.2 mDa), indicating correct characterization (Figure 5.7). The remaining test set compounds had KMD that did not fall within the APB KMD filter, also indicating correct characterization. Further, the APB and 2C KMD filters do not overlap, overcoming a limitation of the absolute mass defect filters discussed in Section 5.2.4. Table 5.2 Calculation of APB Kendrick mass defect filter Compound Nominal Mass (Da) Kendrick Mass (Da) Kendrick Mass Defect (mDa) 4-APB 175 174.9044 95.6 5-APB 175 174.9045 95.0 6-APB 175 174.9031 96.9 *90% CL 93 KMD Filter 2 (mDa)* 95.9 1.6 Kendrick Mass Defect (mDa) 100 7-APB 95 90 165 175 185 195 205 215 225 m/z Training Set 2C Phenethylamines APB Test Set 3C & Cathinone Test Set Figure 5.7 APB subclass Kendrick mass defect filter at 95.9 ± 1.6 mDa at a 90% confidence level. The horizontal lines represent the average (black) and the upper and lower bounds of the mass defect filter (red) 5.2.7 Kendrick Mass Defect Filters of the 2C-Phenethylamine Subclass Table 5.3 shows the KMD and the associated filter for the 2C-phenethylamine subclass. The 2C KMD filter was determined to be 92.2 ± 1.5 mDa at the 95% confidence level. The test set contained the remaining 2C-phenethylamines, all APB, NBOMe, 3C, and cathinone compounds. After halogen/nitro group substitution, all the 2C-phenethylamines had KMD that fell within the filter. The sulfur-containing compounds, 2C-T and 2C-T-2 have KMD around 155 mDa due to the contribution of sulfur to the KMD, which would cause these compounds to fall outside the filter. However, because sulfur can be identified by including it in element composition selection, the mass of sulfur is replaced with the mass of a methylene group 94 (14.01565 Da), and compounds 2C-T and 2C-T-2 then correctly fall within the filter (Figure 5.8) as members of the homologous series. The 3C compounds (mescaline and escaline) have KMD that fall outside and above the 2C filter around 115 mDa, indicating correct characterization. Both 3C compounds have KMD that fall near one another because they are members of their own homologous series, and thus have similar KMD. The cathinone compounds (3-MEC and mephedrone) have KMD that fall outside and below the 2C filter, around 82 mDa, and would also be correctly characterized. The KMD of the APB- and NBOMe-phenethylamines did not fall within the 2C filter, further indicating correct characterization. Table 5.3 Calculation of 2C Kendrick mass defect filter Compound Nominal Mass (Da) Kendrick Mass (Da) Kendrick Mass Defect (mDa) 2C-H 181 180.9082 91.8 2C-D 195 194.9064 93.6 2C-G 209 208.9086 91.4 2C-P 223 222.9082 91.9 KMD Filter 2 (mDa)* 92.2 *95% CL 95 1.5 Kendrick Mass Defect (mDa) 120 3C-phenethylamines 100 2C-T-2 2C-T cathinones 80 170 180 190 200 210 220 230 240 250 m/z Training Set 2C Test Set APB Phenethylamines 3C & Cathinone Test Set Figure 5.8 2C subclass Kendrick mass defect filter at 92.2 ± 1.5 mDa at a 95% confidence level. The horizontal lines represent the average (light blue) and the upper and lower bounds of the mass defect filter (dark blue) 5.2.8 Kendrick Mass Defect Filters of the NBOMe-Phenethylamine Subclass The NBOMe KMD filter was defined using a training set of only alkyl-substituted NBOMe compounds (Table 5.4). The NBOMe KMD filter was determined to be 171.5 ± 7.7 mDa at the 99% confidence level. The test set contained the remaining NBOMe phenethylamines, all 2C, APB, 3C, and cathinone compounds, as well as the theoretical KMD of six cannabinoids (Section 5.2.4). The cannabinoids are tested against the NBOMe KMD filter because some were incorrectly characterized within the NBOMe filter when their absolute mass defects were tested. All the NBOMe test set compounds had KMD that fell inside the KMD filter with the exception of 25B-NBOMe and mescaline-NBOMe (Figure 5.9). As discussed in Section 5.2.4, 96 Table 5.4 Calculation of NBOMe Kendrick mass defect filter Compound Nominal Mass (Da) Kendrick Mass (Da) Kendrick Mass Defect (mDa) 25H-NBOMe 301 300.8291 170.9 25D-NBOMe 315 314.8294 170.6 25G-NBOMe 329 328.8269 173.1 KMD Filter 2 (mDa)* 171.5 7.7 Kendrick Mass Defect (mDa) *99% CL 225 195 25B-NBOMe Mescaline-NBOMe 165 135 105 2C-T 75 165 2C-T-2 215 265 315 m/z Training Set 2C Phenethylamines 3C & Cathinone Test Set NBOMe Test Set APB Phenethylamines Theoretical Cannabinoids Figure 5.9 NBOMe subclass Kendrick mass defect filter at 171.5 ± 7.7 mDa at a 99% confidence level. The horizontal lines represent the average (light purple) and the upper and lower bounds of the mass defect filter (dark purple) the mass accuracy of the molecular ion of 25B-NBOMe is poor (46.33 ppm), causing the KMD to fall outside the filter. Mescaline-NBOMe has a KMD that falls outside and above the NBOMe KMD filter at 194.3 mDa because it is not a member of the same homologous series as the other 97 NBOMe compounds, much like its 3C counterparts’ relation to the 2C compounds. All the KMD of the APB- and 2C-phenethylamine compounds fall outside the filter, indicating correct characterization. The six theoretical KMD of the cannabinoids also fall outside the NBOMe filter, illustrating KMD as a more specific and robust filter than the NBOMe absolute mass defect filter. Overall, KMD has the most specificity to differentiate and characterize unknown compounds and give a preliminary indication of subclass. Further investigation of fragment ions, neutral losses and common fragments was also performed to enhance the confidence of KMD characterization as well as provide evidence toward characterization in the event that no molecular ion is present or cannot be confirmed by chemical ionization. 5.2.9 Kendrick Mass Defect Filters for Neutral Losses and Common Fragment Ions One of the limitations of using fragment ions in the low-resolution scheme was the lack of elemental formulae assignment after a neutral loss. Using high-resolution mass spectrometry this can be overcome by using exact mass for formulae assignment for each ion. Further, KMD filters can be developed for the fragments remaining after common neutral losses. To investigate KMD filters based on fragment ions, first the high-resolution spectra were probed and tables were created for each 2C compound detailing the most prominent fragment ions, their mass accuracies, and elemental compositions as shown in Figure 5.10 and Table 5.5 for 2C-H. Knowing the elemental composition, the neutral losses from the molecular ion were then determined, as shown in Figure 5.11 for 2C-H. Neutral losses were compiled for each 2Cphenethylamine to identify common losses that may be characteristic of this subclass (Table 5.6). 98 2C-H 152.0833 100 O Abundance (%) NH2 O 137.0601 121.0645 181.1104 0 0 50 100 150 200 250 300 m/z Figure 5.10 Spectrum of 2C-H showing abundant ions Table 5.5 Ion table of 2C-H showing abundant ion elemental composition assignments and mass accuracies m/z Mass accuracy (ppm) Elemental composition m/z Mass accuracy (ppm) Elemental composition 181.1104 0.6 C10H15NO2 121.0645 6.6 C8H9O 152.0833 2.6 C9H12O2 109.0643 9.2 C7H9O 137.0601 1.5 C8H9O2 105.0342 1.9 C7H5O 99 2C-H H O + CH2 Molecular Formula: C9H12O2 Monoisotopic Mass: 152.083181 Da v Loss of: CH3N O O O NH2 C Molecular Formula: C8H9O2 Monoisotopic Mass: 137.059706 Da + O Molecular Formula: C10H15NO2 Monoisotopic Mass: 181.110279 Da Loss of: C2H6N O O + CH2 Molecular Formula: C8H9O Monoisotopic Mass: 121.064791 Da Loss of: C2H6NO O + H Molecular Formula: C7H9O Monoisotopic Mass: 109.064791 Da Loss of: C3H6NO Figure 5.11 Proposed structures for fragment ions of 2C-H after their neutral losses Table 5.6 Table of remaining ions after common losses of all 2C compounds LOSS 2C-H 2C-D 2C-G 2C-E 2C-P 2C-B 2C-C 2C-I 2C-N Molecular Formula C10H15NO2 C11H17NO2 C12H19NO2 C12H19NO2 C13H21NO2 C10H14NO2Br C10H14NO2Cl C10H14NO2I C10H14N2O4 2C-T 2C-T-2 CH2N CH3N CH4N C2H6N C2H6NO C9H13O2 C10H15O2 C11H17O2 C11H17O2 C12H19O2 C9H11O2 C10H13O2 C11H15O2 C11H15O2 C12H17O2 C9H10O2Br C9H10O2Cl C8H9O2 C9H11O2 C10H13O2 C10H13O2 C11H15O2 C8H8O2Br C8H8O2Cl C8H8O2I C8H9O C9H11O C10H13O C10H13O C11H15O C8H8OBr C9H12O2Cl C9H12O2I C9H12NO4 C9H12O2 C10H14O2 C11H16O2 C11H16O2 C12H18O2 C9H11O2Br C9H11O2Cl C9H11O2I C9H11NO4 C11H17NO2S C10H15O2S C10H14O2S C10H13O2S C9H11O2S C9H11OS C12H19NO2S C11H17O2S C11H16O2S C11H16O2S C10H13O2S C10H13OS Ion present, but lower than 5% relative abundance 100 C8H8OI Ion not present The common losses were CH2N, CH3N, CH4N, C2H6N, and C2H6NO. From the five alkyl-substituted 2C compounds, the KMDs of ion fragments remaining after each of these losses were used to calculate Kendrick mass defect filters. The filter for each loss and their respective confidence levels can be seen in Table 5.7. The five filters are shown in Figure 5.12 with the KMDs of the alkyl-substituted 2C fragment ions used to define them. Only one filter is distinctly separated from the rest: KMD of fragments resulting from a loss of C2H6NO. This filter was selected as one to use in the characterization scheme. The remaining filters (blue, orange, green and black) all had some degree of overlap because they all were representing members of the same homologous series, where only the number of carbons and hydrogens were different. Based on the structural elucidations of fragments, and commonality of the loss among all 2C compounds, the filter representing the loss of CH3N was also chosen to use as part of the characterization scheme. The KMDs of the fragments of nonalkyl substituted 2Cs resulting from these losses were then calculated, replacing halogens and nitro groups with hydrogen and sulfur groups with CH2 when appropriate (Section 5.2.7), and plotted against the selected filters (Figure 5.13). After replacement, the KMD of fragments of 2C-B, 2C-C, 2C-I, 2C-N, and 2C-T-2 showing a loss of CH3N correctly characterized within the CH3N loss filter (dark green). The KMD of fragments after a loss of C2H6NO of 2C-B, 2C-I, 2CN, 2C-T, and 2C-T-2 correctly fall within that respective filter (dark pink). Compound 2C-T also showed a loss of CH3N, however because the remaining fragment ion had a poor mass accuracy of 23.2 ppm, the KMD of the ion falls outside the corresponding filter. The incorrect characterization of 2C-T highlights the importance of having good mass accuracy of fragment ions. 101 Table 5.7 Kendrick mass defect filters associated with ion fragments after common neutral losses Neutral Loss Filter (mDa) Confidence Level CH2N 83.6 1.93 99% CH3N 86.0 0.72 95% CH4N 91.7 5.3 99% C2H6N 92.8 0.16 99% C2H6NO 69.6 5.16 99% Kendrick Mass Defect (mDa) 100 95 90 85 80 75 70 65 60 115 135 155 175 195 215 m/z Loss CH2N Loss CH3N Loss CH4N Loss C2H6N Loss C2H6NO Figure 5.12 Kendrick mass defect filters developed based on common losses of alkyl-substituted 2C compounds. Points represent KMD of fragment ions remaining after each respective loss. The horizontal lines represent the average (lighter colors) and the upper and lower bounds of each mass defect filter (darker colors) 102 Kendrick Mass Defect (mDa) [M-CH3N]+ 87 82 2C-T 77 [M-C2H6NO]+ 72 67 62 115 135 155 175 195 215 m/z Fragments after loss of CH3N Fragments after loss of C2H6NO Figure 5.13 Selected Kendrick mass defect filters representing losses of CH3N and C2H6NO for all 2C fragments falling within said filters. Fragment shown outside the filter is from 2C-T. The horizontal lines represent the average (light green and purple) and the upper and lower bounds of each mass defect filter (dark green and purple) Common losses for NBOMes were investigated in the same way as the 2C compounds. The high-resolution spectra were examined and ion tables were created (Figure 5.14, Table 5.8), which facilitated structural elucidation of some of the fragments (Figure 5.15). In comparing common neutral losses, it was observed that all NBOMe compounds lost one methoxy group (CH3O) and exhibited a loss of C9H11NO, which is proposed to be a loss of the amine and methoxy-phenyl chain. It was also observed that all compounds exhibited the same fragment of m/z 121 as the base peak, m/z 150 and m/z 91 (methyl-benzene, not pictured). 103 25H-NBOMe 121.0649 100 O NH O Abundance (%) O 91.0543 150.0916 270.1495 0 0 100 200 300 m/z Figure 5.14 Spectrum of 25H-NBOMe and most abundant fragment ions above m/z 105 Table 5.8 Ion table of 25H-NBOMe with elemental composition assignments and mass accuracies of most abundant fragment ions above m/z 105 m/z Mass accuracy (ppm) Elemental composition m/z Mass accuracy (ppm) Elemental composition 301.1654 1.8 C18H23NO3 150.0916 1.79 C9H12NO 270.1495 0.69 C17H20NO2 122.0684 34.28 C8H10O 152.0835 2.36 C9H12O2 121.0649 0.68 C8H9O 104 O NH 25H-NBOMe C + Loss of: CH3O O Molecular Formula: C17H20NO2 Monoisotopic Mass: 270.148855 Da H O O NH + CH2 NH Loss of: C9H11NO v O O O Molecular Formula: C18H23NO3 Monoisotopic Mass: 301.167794 Da Molecular Formula: C9H12O2 Monoisotopic Mass: 152.083181 Da H2C + O Molecular Formula: C9H12NO Monoisotopic Mass: 150.09134 Da Base peak O + CH2 H2C OH + Molecular Formula: C8H10O Monoisotopic Mass: 122.072616 Da Molecular Formula: C8H9O Monoisotopic Mass: 121.064791 Da Figure 5.15 Proposed structures for fragment ions of 25H-NBOMe after their neutral losses The NBOMe fragment remaining after the C9H11NO loss was observed to be the same as the fragment remaining after a loss of CH3N in the 2C compounds (Figure 5.11). This fragmentation is shown in Figure 5.16, using 2C-N and 25N-NBOMe as examples. Therefore, the filter previously developed is applicable here – although it corresponds to a different neutral loss, the same ion is remaining. The KMD of these NBOMe fragment ions was calculated (halogens/nitro/sulfur groups replaced when applicable) and plotted against the 2C CH3N loss filter. Except that from mescaline-NBOMe, all the fragments correctly characterized within the filter (Figure 5.17) as shown in yellow. This helps to illustrate how NBOMe compounds fragment, and their relationship with 2C compounds. No NBOMe-specific KMD neutral loss filters were developed because NBOMes could be definitively identified by their characteristic mass spectral features, presence of characteristic neutral losses, and the applicability of the 2C CH3N loss filter. 105 O NH2 Loss of CH3N O2N HO + CH2 O 2C-N O 2N O O NH Loss of C9H11NO O O2N O 25N-NBOMe Figure 5.16 Proposed structural elucidation of 2C-N and 25N-NBOMe leading to the same fragment (C9H11NO4) Kendrick Mass Defect (mDa) Mescaline-NBOMe 105 100 95 90 [M-CH3N]+ 85 2C-T 80 145 155 165 175 185 195 205 215 m/z 2C fragments after loss of CH3N NBOMe fragments after loss of C9H11NO Figure 5.17 Selected Kendrick mass defect filter and corresponding NBOMe fragments falling within the filter. Fragments shown outside the filter are from mescaline-NBOMe and 2C-T. The horizontal lines represent the average (light green) and the upper and lower bounds of the mass defect filter (dark green) 106 To test the KMD fragment ion filters developed, the two 3C-phenethylamines and two cathinone compounds were analyzed for fragment ions after common neutral losses. Neither mephedrone nor 3-MEC exhibited losses of CH3N or C2H6NO. Mescaline and escaline did exhibit losses of both CH3N and C2H6NO; however, when the KMD values of each of the four remaining fragment ions were calculated and plotted, all four correctly characterized as being Kendrick Mass Defect (mDa) outside both 2C fragment filters (Figure 5.18). Mescaline-NBOMe 110 Escaline Mescaline Mescaline Escaline [M-CH3N]+ 85 2C-T [M-C2H6NO]+ 60 115 135 155 175 195 215 m/z 2C fragment ions after loss of CH3N 2C fragments after loss of C2H6NO NBOMe fragment ions 3C fragment ions Figure 5.18 Selected Kendrick mass defect filters and corresponding 3C fragments falling outside the filters Developing KMD filters on common fragment ions is not possible because the substituent on each compound causes different m/z in a spectrum, leading to a lack of ions in common across a subclass. Further, some ions that are common across a subclass are not necessarily characteristic, e.g., m/z 77, which is present in all spectra for aromatic compounds. Developing the filters related to common neutral losses are more successful because members of the 2C subclass have the same neutral losses, despite having different substitutions (Section 4.3). 107 5.3 Scheme for Characterization of Synthetic Phenethylamines using High-Resolution Mass Spectra With the addition of Kendrick mass defect filters based on molecular ions and Kendrick mass defect filters based on fragment ions after neutral losses, the characterization scheme for low-resolution data can be modified to create a characterization scheme based on high-resolution data. Many of the same components of the low-resolution data scheme are retained, including retention index determination, molecular ion confirmation, and substituent identification, and where applicable, halogen/nitro/sulfur group replacement with the exact masses of hydrogen or CH2 (Section 5.2.3). The order of the revised characterization scheme is slightly different such that the substituent must be accounted for before the Kendrick mass defect filters can be applied. The characterization scheme is presented in Figure 5.19 and two examples demonstrating application of the scheme follow. 108 1 Is retention index available? IT between 1499 – 1527 suggests APB. IT between 1590 – 2000 suggests 2C. IT between 2475 – 2839 suggests NBOMe. Other APB 2a Is there an ion at m/z 131 (C H O+) >10% 9 7 abundance relative to the base peak? Yes No Not consistent with an APBphenethylamine Consistent with an APBphenethylamine 3 Is there a molecular ion? *Can be confirmed by CI data Yes 4 Yes Consistent with an APBphenethylamine Yes No Consistent with an NBOMephenethylamine. Continue to Step 3. 3 Is there a molecular ion? *Can be confirmed by CI data No Does the Kendrick mass defect** of the M+adj. fall in the APB filter between 95.9 ± 1.6 mDa (94.2 – 97.5 mDa)? 2b Does the spectrum have three predominant peaks at m/z 91 (C 7H7+), 121 (C8H9O+) and 150 (C9H12NO+) with the base peak at m/z 121? See NOTE No Not consistent with an APBphenethylamine Yes No Is there a halogen, sulfur, or nitro group present? ***(next page) Yes See NOTE No Continue to 3a. ** Kendrick Mass defect is calculated by: Exact mass * (14/14.01565) = Kendrick mass (Nominal mass - Kendrick Mass) * 1000 = Kendrick Mass defect in mDa Not consistent with an NBOMe-phenethylamine. Continue to Step 3. 3 Is there a molecular ion? *Can be confirmed by CI data Yes Is there a halogen, sulfur, or nitro group present? ***(next page) Yes Continue to 3b. No Continue to 3c. See NOTE No Continue to 3d. NOTE: If no molecular ion is confirmed: only halogens can be identified. Figure 5.19 Characterization scheme for high-resolution mass spectral data. M+adj is the mass of the molecular ion adjusted for a halogen/sulfur/nitro substituent 109 Figure 5.19 (con’t) 3a 3b If Br, Cl, or I are present, subtract the mass of the halogen (78.9183, 34.9689, 126.9045 Da) from the molecular ion and add the mass of hydrogen (1.0078 Da). If the compound has an even M+ subtract the mass of a nitro group (NO2) (45.9929 Da) from the molecular ion and add the mass of hydrogen (1.0078 Da). If S is present, subtract the mass of sulfur (31.9721 Da) from the molecular ion and add the mass of CH2 (14.0157 Da). 3c Consistent with an alkylor sulfursubstituted compound. Continue to Step 4 Yes 5 If the compound has an even M+ subtract the mass of a nitro group (NO2) (45.9929 Da) from the molecular ion and add the mass of hydrogen (1.0078 Da). If S is present, subtract the mass of sulfur (31.9721 Da) from the molecular ion and add the mass of CH2 (14.0157 Da). 4 Does the Kendrick mass defect of the M+adj. fall in the NBOMe filter between 171.5 ± 7.7 mDa (163.8 – 179.2 mDa)? Does the Kendrick mass defect of the M+adj. fall in the 2C filter between 92.2 ± 1.5 mDa (90.7 – 93.7 mDa)? No Consistent with an NBOMephenethylamine. Continue to Step 5. Yes Yes No Not consistent with a 2C-phenethylamine Consistent with a 2Cphenethylamine. Continue to Step 5. Not consistent with an NBOMe-phenethylamine 5 Does the compound lose CH3N (approx. 29 Da) and C 2H6NO (approx. 60 Da) in neutral losses from M+? Does it lose CH3N (approx. 29 Da) from M+ as the base peak? Does the compound have common loses of CH3O (approx. 31 Da) and C9H11NO (approx. 149 Da) from the molecular ion? Does the fragment remaining after loss of C 9H11NO fall within the CH3N KMD filter?* Consistent with an NBOMe-phenethylamine. Consistent with an alkylor sulfursubstituted compound. Continue to Step 4 This new adjusted M+ should be approximately 181 Da. Use it for Step 4. This new adjusted M+ should be approximately 301 Da. Use it for Step 4. 4 3d If Br, Cl, or I are present, subtract the mass of the halogen (78.9183, 34.9689, 126.9045 Da) from the molecular ion and add the mass of hydrogen (1.0078 Da). No Yes Not consistent with an NBOMe-phenethylamine ***If Br is present, double peaks (doublets) of similar abundance will be present, spaced two mass units apart for higher mass fragments If Cl is present, doublets in a 3:1 abundance ratio will be present, spaced two mass units apart for higher mass fragments If I is present, m/z 126.9 and m/z 127.9 should be present (I and HI, respectively) Nitrogen rule: If the mass of the molecular ion is even, there is an even number of nitrogens present, or none at all. Ex: The M+ for 2C-H has one nitrogen and its m/z 181 is odd, indicating an odd number of nitrogens, while M+ for 2C-N which has two nitrogens is m/z 226, an even number. If S is present, there may be an [M+2] + ion of low abundance No Do the fragments remaining after the losses of CH3N and C 2H6NO fall within the KMD fragment filters?* Loss CH3N KMD filter = 86.0 ± 0.7 mDa (85.2 – 86.7 mDa) Loss C 2H6NO KMD filter = 69.6 ± 5.2 mDa (64.5 – 74.8 mDa) Yes Consistent with a 2Cphenethylamine Not consistent with a 2Cphenethylamine No Not consistent with a 2Cphenethylamine. If KMD falls between 95 – 110 mDa, it is consistent with a 3C-phenethylamine *Replace halogens/sulfur/nitro group when appropriate 110 Example 1: 3-methylethcathinone (3-MEC) (Figure 5.20) 1. Is retention index available? No, the retention index of 3-MEC was not available. 2a. Is there an ion at m/z 131 (C9H7O+) >10% abundance relative to the base peak? No, there is an ion at m/z 131.0748 but the abundance is 1.1% relative to the base peak. 2b. Does the spectrum have three predominant peaks at m/z 91 (C7H7+), 121 (C8H9O+) and 150 (C9H12NO+) with the base peak at m/z 121? No, these peaks are not present, therefore this compound is not consistent with an NBOMe-phenethylamine. 3. Is there a molecular ion? Yes, the molecular ion was confirmed to be m/z 191.1310 with a mass accuracy of 0.0 ppm. a. Is there a halogen, sulfur, or nitro group present? No, no halogens, sulfur, or nitro groups were determined to be present. 4. Does the Kendrick mass defect of M+adj fall in the 2C-phenethylamine filter between 92.2 ± 1.5 mDa (90.7 – 93.7 mDa)? No, the KMD of M+ (82.4 mDa) does not fall within the 2C-phenethylamine KMD filter. This compound is not consistent with a 2Cphenethylamine. If treated as an unknown, 3-MEC would be not be characterized as an APB, NBOMe, or 2C-phenethylamine. 111 3-MEC 100 O Abundance (%) NH 50 Loss of C3H8O 131.0748 9.9 ppm C9H9N M+ 191.1310 0.0 ppm C12H17NO 0 0 50 100 150 200 m/z Figure 5.20 Mass spectrum and structure of cathinone, 3-methylethcathinone (3-MEC) showing loss of C3H8O, which is uncharacteristic of the phenethylamine class In the characterization scheme for low-resolution data, 3-MEC exhibited a loss of 60 Da, which is the nominal mass of a loss of C2H6NO, a common neutral loss for 2C compounds. However, with the advantage of high-resolution mass spectrometry, leading to elemental formulae assignment, the loss of 60 Da from M+ of 3-MEC corresponds to a fragment at m/z 131.0748 and a formula assignment of C9H9N with a mass accuracy of 9.9 ppm. This would equate to a loss of C3H8O, which also gives a nominal mass of 60 Da. Through the characterization scheme for low-resolution data, it was determined that 3-MEC would not be characterized as an APB or NBOMe, but it could not be discerned whether or not it was a 2Cphenethylamine. However, using high resolution, 3-MEC would be correctly characterized as being inconsistent with an APB-, NBOMe- and 2C-phenethylamine. 112 Example 2: Mescaline (Figure 5.21) 1. Is the retention index available? No, the retention index of mescaline was not available. 2a. Is there an ion at m/z 131 (C9H7O+) >10% abundance relative to the base peak? No, there is no ion at m/z 131. 2b. Does the spectrum have three predominant peaks at m/z 91 (C7H7+), 121 (C8H9O+) and 150 (C9H12NO+) with the base peak at m/z 121? No, these peaks are not present, therefore this compound is not consistent with an NBOMe-phenethylamine. 3. Is there a molecular ion? Yes, the molecular ion was confirmed to be m/z 211.1207 with a mass accuracy of 0.5 ppm. a. Is there a halogen, sulfur, or nitro group present? No, no halogens, sulfur, or nitro groups were determined to be present. 4. Does the Kendrick mass defect of M+adj fall in the 2C-phenethylamine filter between 92.2 ± 1.5 mDa (90.7 – 93.7 mDa)? No, the KMD of M+ (115.0 mDa) does not fall within the 2C-phenethylamine KMD filter. This compound is not consistent with a 2Cphenethylamine. 5. Does the compound lose CH3N (approx. 29 Da) and C2H6NO (approx. 60 Da) in neutral losses from M+? Does it lose CH3N from M+ as the base peak? Yes, the spectrum has ions at m/z 182.0967 (211.1207 – 29.0240 = 182.0967) and m/z 151.0725 (211.1207 – 60.0482 = 151.0725). The ion at m/z 182.0967 has an elemental formula of C10H14O3 with a mass accuracy of 13.2 ppm, corresponding to a neutral loss of CH3N (C11H17NO3 – CH3N = C10H14O3). The ion at m/z 151.0725 has an elemental formula of C9H11O2 with a mass accuracy of 22.5 ppm, corresponding to a neutral loss of C2H6NO 113 (C11H17NO3 – C2H6NO = C9H11O2). The loss of CH3N from M+ (m/z 182.0967) is the base peak. HO + CH2 100 Loss of CH3N M+ 182.0967 211.1207 0.5 ppm 13.2 ppm C10H14O3 C11H17NO3 Mescaline O O O NH2 Abundance (%) O Loss of + C2H6NO M 211.1207 + 151.0725 CH2 0.5 ppm 22.5 ppm C9H11O2 C11H17NO3 50 O O O 0 0 50 100 150 200 250 m/z Figure 5.21 Mass spectrum of 3C-phenethylamine, mescaline and fragment ions remaining after neutral losses, the KMD of which can be used to distinguish 2C from 3C-phenethylamines a. Do the fragments remaining after the losses of CH3N and C2H6NO have KMD that fall within the KMD fragment filters? [M-CH3N]+ KMD filter = 86.0 ± 0.7 mDa (85.2 – 86.7 mDa). [M-C2H6NO]+ KMD filter = 69.6 ± 5.2 mDa (64.5 – 74.8 mDa). No, the fragment remaining after a loss of CH3N (m/z 182.0967) has a KMD of 106.6 mDa and does not fall within the CH3N KMD filter. The fragment remaining after a loss of C2H6NO (m/z 151.0725) has a KMD of 96.2 mDa and does not fall within the C2H6NO KMD filter. 114 b. Are the KMD of the fragments remaining after the neutral losses of CH3N and C2H6NO between 95 – 110 mDa? Yes, the KMD of the fragments remaining are 106.6 and 96.2 mDa. This is indicative of a 3C-phenethylamine. If treated as an unknown, mescaline would be correctly characterized as a 3Cphenethylamine. In the characterization scheme for low-resolution data, mescaline was incorrectly characterized as a 2C-phenethylamine. However, with the addition of mass defects filters, 2C- and 3Cphenethylamines can be easily differentiated, and mescaline is correctly characterized. 5.4 Summary High-resolution mass spectrometry can overcome the limitations of low-resolution mass spectrometry by giving definitive identification of ions through elemental assignment and mass accuracy measurements. Using exact mass measurements, mass defects can be explored for use in characterization of unknown compounds to a specific designer drug class or subclass. This allows for a more accurate preliminary characterization and a more detailed, descriptive characterization scheme. 115 APPENDICES 116 APPENDIX A: High-Resolution Mass Spectra A) B) 100 100 4-APB NH2 NH2 Abundance (%) Abundance (%) O 5-APB 50 O 50 0 0 0 50 100 150 200 250 300 0 50 100 150 m/z 200 250 300 m/z C) O 7-APB 100 Abundance (%) NH2 50 0 0 50 100 150 200 250 300 m/z Figure A.1 High-resolution mass spectra of (A) 4-(2-aminopropyl)benzofuran (4-APB), (B) 5-(2aminopropyl)benzofuran (5-APB), and (C) 7-(2-aminopropyl)benzofuran 117 A) 100 2C-D B) 2C-E 100 O O NH2 Abundance (%) Abundance (%) NH2 O 0 O 0 0 50 100 150 200 250 300 0 50 100 m/z 150 200 250 300 m/z C) 2C-P 100 O Abundance (%) NH2 O 0 0 50 100 150 200 250 300 m/z Figure A.2 High-resolution mass spectra of (A) 2,5-dimethoxy-4-methylphenethylamine (2C-D), (B) 2,5-dimethoxy-4-ethylphenethylamine (2C-E), and (C) 2,5-dimethoxy-4propylphenethylamine (2C-P) 118 A) B) 100 O 2C-C 100 2C-I O NH2 I Abundance (%) Abundance (%) NH2 Cl O O 0 0 0 50 100 150 200 250 300 0 50 100 m/z 150 200 250 300 350 m/z C) 100 2C-N O Abundance (%) NH2 O2N O 0 0 50 100 150 200 250 300 m/z Figure A.3 High-resolution mass spectra of (A) 2,5-dimethoxy-4-chlorophenethylamine (2C-C), (B) 2,5-dimethoxy-4-iodophenethylamine (2C-I), and (C) 2,5-dimethoxy-4-nitrophenethylamine (2C-N) 119 A) 2C-T 100 O NH2 Abundance (%) S O 0 0 50 100 150 200 250 300 200 250 300 m/z B) 2C-T-2 100 O NH2 Abundance (%) S O 0 0 50 100 150 m/z Figure A.4 High-resolution mass spectra of (A) 2,5 -dimethoxy-4-methylthiophenethylamine (2C-T) and (B) 2,5-dimethoxy-4-ethylthiophenethylamine (2C-T-2) 120 A) B) 25D-NBOMe 25E-NBOMe O O NH NH O 100 O 100 O Abundance (%) Abundance (%) O 0 0 0 100 200 300 0 100 200 m/z 300 m/z C) 25G-NBOMe 100 O NH Abundance (%) O O 50 0 0 100 200 300 m/z Figure A.5 High-resolution mass spectra of (A) 2-(2,5-dimethoxy-4-methylphenyl)-N-(2methyoxybenzyl)ethanamine (25D-NBOMe), (B) 2-(4-ethyl-2,5-dimethoxyphenyl)-N-(2methoxybenzyl)ethanamine (25E-NBOMe) and (C) 2,5-dimethoxy-N-[(2methoxyphenyl)methyl]-3,4-dimethyl-benzeneethanamine (25G-NBOMe) 121 A) 25B-NBOMe B) 25C-NBOMe O O NH 100 NH O Br 100 O Cl Abundance (%) O Abundance (%) O 0 0 0 100 200 300 400 0 100 200 m/z 300 400 m/z C) 25I-NBOMe 100 O NH O Abundance (%) I O 0 0 100 200 300 400 m/z Figure A.6 High-resolution mass spectra of (A) 4-bromo-2,5-dimethoxy-N-[(2methoxyphenyl)methyl]-benzeneethanamine (25B-NBOMe), (B) 4-chloro-2,5-dimethoxy-N-[(2methoxyphenyl)methyl]-benzeneethanamine (25C-NBOMe), and (C) 4-iodo-2,5-dimethoxy-N[(2-methoxyphenyl)methyl]-benzeneethanamine (25I-NBOMe) 122 A) 25T-NBOMe 100 B) 25T-4-NBOMe 100 O O NH NH O O S Abundance (%) Abundance (%) S O O 0 0 0 100 200 300 400 0 100 200 m/z C) 300 400 m/z 25T-7-NBOMe D) 100 Mescaline-NBOMe 100 O NH O O NH O O Abundance (%) S Abundance (%) O 0 O 0 0 100 200 300 400 m/z 0 100 200 300 400 m/z Figure A.7 High-resolution mass spectra of (A) 2,5-dimethoxy-N-[(2-methoxyphenyl)methyl]-4(methylthio)-benzeneethanamine (25T-NBOMe), (B) 2,5-dimethoxy-N-[(2methoxyphenyl)methyl]-4-[(1-methylethyl)thio]-benzeneethanamine (25T-4-NBOMe), (C) 2,5dimethoxy-N-[(2-methoxyphenyl)methyl]-4-(propylthio)-benzeneethanamine (25T-7-NBOMe), and (D) 3,4,5-trimethoxy-N-[(2-methoxyphenyl)methyl]-benzeneethanamine (mescalineNBOMe) 123 A) Escaline 100 O NH2 O Abundance (%) O 50 0 0 100 200 300 m/z B) Mephedrone O 100 Abundance (%) NH 50 0 0 50 100 150 200 m/z Figure A.8 High-resolution mass spectra of (A) 4-ethoxy-3,5-dimethoxy-benzeneethanamine (escaline) and (B) 4-methylmethcathinone (mephedrone) 124 APPENDIX B: Additional High-Resolution Characterization Scheme Examples Example 1: 2C-G (Figure A.9) 1. Is retention index available? Yes, the retention index is 1751. This retention index falls within the retention index range identified for 2C-phenethylamines (1590 – 2000). 2b. Does the spectrum have three predominant peaks at m/z 91 (C7H7+), 121 (C8H9O+) and 150 (C9H12NO+) with the base peak at m/z 121? No, these peaks are not present, therefore this compound is not consistent with an NBOMe-phenethylamine. 3. Is there a molecular ion? Yes, the molecular ion was confirmed to be m/z 209.1421 with a mass accuracy of 2.4 ppm. a. Is there a halogen, sulfur, or nitro group present? No, no halogens, sulfur, or nitro groups were observed. 100 O + Loss of CH3N H 180.1152 1.1 ppm C11H16O2 CH2 Abundance (%) O 209.1421 2.4 ppm C12H19NO2 2C-G O NH2 + CH2 O 149.0985 Loss of 209.1421 12.7 ppm C2H6NO 2.4 ppm C10H13O C12H19NO2 O 0 0 50 100 150 200 250 300 m/z Figure A.9 Mass spectrum of 2C-G and fragment ions remaining after neutral losses 125 4. Does the Kendrick mass defect of M+ fall in the 2C-phenethylamine filter between 92.2 ± 1.5 mDa (90.7 – 93.7 mDa)? Yes, the KMD of M+ (91.4 mDa) does fall within the 2C-phenethylamine KMD filter. This compound is consistent with a 2Cphenethylamine. 5. Does the compound lose CH3N (approx. 29 Da) and C2H6NO (approx. 60 Da) in neutral losses from M+? Does it lose CH3N from M+ as the base peak? Yes, the spectrum has ions at m/z 180.1152 (209.1421 – 29.0269 = 180.1152) and m/z 149.0985 (209.1421 – 60.0436 = 149.0985). The ion at m/z 180.1152 has an elemental formula of C11H16O2 and a mass accuracy of 1.1 ppm, corresponding to a neutral loss of CH3N (C12H19NO2 – CH3N = C11H16O2). The ion at m/z 149.0985 has an elemental formula of C10H13O with a mass accuracy of 12.7 ppm, corresponding to a loss of C2H6NO (C12H19NO2 – C2H6NO = C10H13O). The loss from CH3N from M+ (m/z 180.1152) is not the base peak, but is a highly abundant ion. a. Do the fragments remaining after the losses of CH3N and C2H6NO have KMD that fall within the KMD fragment filters? [M-CH3N]+ KMD filter = 86.0 ± 0.7 mDa (85.2 – 86.7 mDa). [M-C2H6NO]+ KMD filter = 69.6 ± 5.2 mDa (64.5 – 74.8 mDa). Yes, the fragment remaining after a loss of CH3N (m/z 180.1152) has a KMD of 85.9 mDa and does fall within the CH3N KMD filter. The fragment remaining after a loss of C2H6NO (m/z 149.0985) has a KMD of 68.0 mDa and falls within the C2H6NO KMD filter. If treated as an unknown, 2C-G would be correctly characterized as a 2C-phenethylamine. 126 Example 2: 2C-B (Figure A.10) 1. Is retention index available? Yes, the retention index is 1856. This retention index falls within the retention index range identified for 2C compounds (1590 – 2000). 2. Does the spectrum have three predominant peaks at m/z 91 (C7H7+), 121 (C8H9O+) and 150 (C9H12NO+) with the base peak at m/z 121? No, these peaks are not present, therefore this compound is not consistent with an NBOMe-phenethylamine. 3. Is there a molecular ion? Yes, the molecular ion was confirmed to be m/z 259.0203 with a mass accuracy of 0.37 ppm. a. Is there a halogen, sulfur, or nitro group present? Yes, bromine doublets are present. Doublets of similar intensity indicate the presence of Br. i. Subtracting the mass of Br (78.9182 Da) from the M+ (m/z 259.0203) and adding the mass of H (1.0078 Da) yields an adjusted molecular ion (M+adj) of m/z 181.1098. 4. Does the Kendrick mass defect of M+adj fall in the 2C-phenethylamine filter between 92.2 ± 1.5 mDa (90.7 – 93.7 mDa)? Yes, the KMD of M+adj (92.4 mDa) does fall within the 2C-phenethylamine KMD filter. This is consistent with a 2C-phenethylamine. 5. Does the compound lose CH3N (approx. 29 Da) and C2H6NO (approx. 60 Da) in neutral losses from M+? Does it lose CH3N from M+ as the base peak? Yes, the spectrum has ions at m/z 229.9938 (259.0203 – 29.0265 = 229.9938) and m/z 198.9772 (259.0203 – 60.0431 = 198.9772). The ion at m/z 229.9938 has an elemental formula of C9H11O2Br and a mass accuracy of 1.7 ppm, corresponding to a neutral loss of CH3N (C10H14NO2Br – CH3N = C9H11O2Br). The ion at m/z 198.9772 has an elemental formula of C8H8OBr with a mass accuracy of 6.7 ppm, corresponding to a loss of C2H6NO 127 (C10H14NO2Br – C2H6NO = C8H9OBr). The loss from CH3N from M+ (m/z 229.9938) is the base peak. a. Do the fragments remaining after the losses of CH3N and C2H6NO have KMD that fall within the KMD fragment filters? [M-CH3N]+ KMD filter = 86.0 ± 0.7 mDa (85.2 – 86.7 mDa). [M-C2H6NO]+ KMD filter = 69.6 ± 5.2 mDa (64.5 – 74.8 mDa). Yes, after replacing the Br with a H on each fragment, the fragment remaining after a loss of CH3N (m/z 152.0833) has a KMD of 86.5 mDa and does fall within the CH3N KMD filter. The fragment remaining after a loss of C2H6NO (m/z 121.0667) has a KMD of 68.5 mDa and falls within the C2H6NO KMD filter. If treated as an unknown, 2C-B would be correctly characterized as a 2C-phenethylamine with a bromine substituent. H O + CH2 100 Br Loss of CH3N 229.9938 259.0203 1.7 ppm 0.37 ppm C9H11O2Br C10H14BrNO2 O O 2C-B NH2 Loss of C2H6NO Abundance (%) O + CH2 Br 198.9772 6.7 ppm C8H8OBr Br 259.0203 0.37 ppm C10H14BrNO2 O 0 0 50 100 150 200 250 300 m/z Figure A.10 Mass spectrum of 2C-B and fragment ions remaining after neutral losses 128 Example 3: 25N-NBOMe (Figure A.11) 1. Is retention index available? Yes, the retention index is 2839. This retention index falls within the retention index range identified for NBOMe compounds (2475 – 2839). 2b. Does the spectrum have three predominant peaks at m/z 91 (C7H7+), 121 (C8H9O+) and 150 (C9H12NO+) with the base peak at m/z 121? Yes, the spectrum has prominent peaks at m/z 91.0543, 121.0649, and 150.0915. These three peaks are consistent with the NBOMe phenethylamine subclass. 3. Is there a molecular ion? Yes, the molecular ion was confirmed to be m/z 346.1493 with a mass accuracy of -8.76 ppm. a. Is there a halogen, sulfur, or nitro group present? Yes, an even-massed molecular ion indicates the presence of more than one nitrogen. A nitro group is present. i. Subtracting the mass of NO2 (45.9929 Da) from the M+ (m/z 346.1493) and adding the mass of H (1.0078 Da) yields an adjusted molecular ion (M+adj) of m/z 301.1642 4. Does the Kendrick mass defect of M+adj fall in the NBOMe-phenethylamine filter between 171.7 ± 7.7 mDa (163.8 – 179.2 mDa)? Yes, the KMD of M+adj (172.1 mDa) does fall within the NBOMe-phenethylamine KMD filter. This is consistent with an NBOMe-phenethylamine. 5. Does the compound lose CH3O (approx. 31 Da) and C9H11NO (approx. 149 Da) in neutral losses from M+? Yes, the spectrum has ions at m/z 315.1284 (346.1493 – 31.0209 = 315.1284) and m/z 197.0686 (346.1493 – 149.0807 = 197.0686). The ion at m/z 315.1284 has an elemental formula of C17H19N2O4 with a mass accuracy of -17.53 129 ppm, corresponding to a neutral loss of CH3O (C18H22N2O5 – CH3O = C17H19N2O4). The ion at m/z 197.0686 has an elemental formula of C9H11NO4 with a mass accuracy of 1.74 ppm, corresponding to a neutral loss of C9H11NO (C18H22N2O5 – C9H11NO = C9H11NO4). a. Does the KMD of the fragment remaining after the losses of C9H11NO fall within the CH3N KMD filter (after replacement of the nitro group)? [MCH3N]+ KMD filter = 86.0 ± 0.7 mDa (85.2 – 86.7 mDa). Yes, after replacing the NO2 with a H, the fragment remaining after a loss of C9H11NO (m/z 152.0835) has a KMD of 86.3 mDa. This KMD does fall within the [M-CH3N]+ KMD filter. If treated as an unknown, 25N-NBOMe would be correctly characterized as a NBOMephenethylamine with a nitro group substituent. 25N-NBOMe O NH 121.0649 100 HO + O O2N CH2 O O2N O Abundance (%) 197.0686 1.74ppm C9H11NO4 O Loss of C9H11NO 346.1493 -8.76 ppm C18H22N2O5 Loss of CH3O 315.1284 346.1493 -17.53 ppm -8.76 ppm C17H19N2O4 C18H22N2O5 NH O2N O C + 0 0 100 200 300 m/z Figure A.11 Mass spectrum of 25N-NBOMe and fragment ions remaining after neutral losses 130 REFERENCES 131 REFERENCES 1. CRC Handbook of Chemistry and Physics, 89th ed.; Lide, D.R., Ed.; CRC Press: Boca Raton, FL, 2008; Section 3, No. 339. 2. Chu, F. Improving Methods for the Analysis of Controlled Substances. MS Thesis, Michigan State University. 2015 132 VI. Conclusions and Future Work 6.1 Conclusions A sample set of designer drugs characteristic of the phenethylamine compound class was analyzed by gas chromatography and both low- and high-resolution mass spectrometry. The chromatographic data were used to develop characteristic retention index ranges for each structural subclass. The spectral data were probed to identify characteristic spectral features to identify compounds of similar subclasses. These features included the investigation of common fragment ions, characteristic neutral losses, and substituent identification. The characteristic subclass features were, in turn, used to develop a characterization scheme in the format of a flow chart which crime laboratories can use as an initial screening method to determine if further examination of a submitted controlled substance sample is necessary. This low-resolution characterization scheme is immediately implementable in forensic laboratories because it has been created using the gas chromatography-mass spectrometry (GC-MS) instrumentation already in place and conventionally used for the identification of controlled substances. The characterization scheme was successful in characterizing all APB-, 2C-, and NBOMephenethylamines into their respective subclasses. However, some of the 3C-phenethylamines and cathinone compounds used to test the scheme were mischaracterized or not characterized at all. The lack of correct characterization means that while the low-resolution scheme is most applicable in a forensic laboratory, there are some limitations, such as a lack of elemental formulae assignment, and thus definitive identification, of the fragment ions in the mass spectra. A high-resolution version of the same characterization scheme was developed for increased confidence of a characterization and to overcome the limitations of the characterization scheme for low-resolution data. This scheme exploits the use of accurate mass and mass defect 133 obtained from high-resolution mass spectrometry, with definitive identification of fragment ions. Absolute and Kendrick mass defect filters were developed but only Kendrick mass defect filters were implemented into the characterization scheme for structural subclass characterization due to the greater specificity afforded. The characterization scheme for high-resolution data was successful in characterizing all the phenethylamine and cathinone compounds, including those mischaracterized and uncharacterized by the scheme for low-resolution data. Kendrick mass defect filters offer a more specific characterization into structural subclass, and overcame many limitations of mischaracterization using absolute mass defect. Overall, the utility of highresolution mass spectrometry for robust characterization of synthetic designer drugs was highlighted, should that instrumentation ever be made available to forensic laboratories. 6.2 Future Work Further investigation of the electron ionization-mass spectral features of sulfur and other substituents should be performed. The presence of sulfur could not always be identified in the mass spectra because it inconsistently exhibited characteristic features such as distinguishing isotope patterns. Identifying other substituents such as fluorine, multiple nitrogens in a fragment ion, or having several, differing substituents in a compound is an additional aim that could be pursued. Another area of future direction would be the optimization of a GC-MS temperature program for the differentiation of all retention indices of phenethylamine isomers. Some of the isomers of the APB-phenethylamine subclass had the same calculated retention indices and thus could not be distinguished from one another. However, if the gas chromatography temperature program was further refined, this limitation could be overcome. Additionally, more research should be conducted on how the substituent, such as a halogen, of a compound affects the retention index. A third direction of future work should build on this work for the 134 characterization of compounds when a molecular ion is not available. Although the molecular ion can be confirmed by chemical ionization, the ability to perform CI analysis may not be possible. Therefore, more work is needed to identify characteristic features of the structural subclasses based on fragment ions alone. Because only reference standards were analyzed in this work, further experimentation would be to first test the characterization schemes against a set of hypothetical unknowns, in which the analyst does not know what they are. Following this, street samples would then be tested, containing cutting agents and lower concentrations of unknown controlled substances. Additionally, synthetic designer drugs of other compound classes, such as tryptamines, could be tested against the current flow charts, or could be used to expand and refine the flow chart for characterization of other compound classes. While there are certain directions for future work and expansion, this research developed two characterization schemes that will be able to assist in identification of compounds in a constantly changing drug market, as well as allow characterization of unknowns for which no reference standard is available. 135