CHARACTERIZATION AND IDENTIFICATION OF CRYSTALLINE STRUCTURES IN CANNABIS SOLVENT EXTRACTS By Otyllia Ruth Abraham A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Forensic Science Master of Science 2021 ABSTRACT CHARACTERIZATION AND IDENTIFICATION OF CRYSTALLINE STRUCTURES IN CANNABIS SOLVENT EXTRACTS By Otyllia Ruth Abraham Cannabis solvent extracts comprise of a variety of products formed through the isolation an d concentration of cannabinoids from either marijuana or hemp using organic solvents. Marijuana and hemp represent two broad classes of Cannabis sativa plants and ar e distinguished based on the concentration of the psychoactive cannabinoid delta - 9 - tetrahyd 9 - THC). A common marijuana solvent extract is butane hash oil, which uses butane to extract and concentrate 9 - THC and its naturally occurring acidic form, delta - 9 - tetrahydrocannabinolic acid 9 - THCA). Hemp solvent extracts, however, typica lly isolate cannabidiol (CBD). This work aimed to comprehensively characterize marijuana and hemp - derived solvent extracts using optical and chemical techniques. Optical analysis via polarized light microscopy (PLM) was performed to characterize crystall ine materials present in both sub sets of extracts and indicated the possibility to differentiate marijuana and hemp extracts based on optical differences. Chemical characterization through infrared spectroscopy and single crystal X - ray diffraction provided the identification of the crysta lline component (THCA for marijuana extracts and CBD for hemp extracts) and supported the PLM findings. Additionally, the derivatization procedure (focusing on reaction temperature, reaction time, and solvent ratio) for THC A using a common silylation reage nt was optimized using full factorial experimental design to allow for the analysis of the solvent extracts by gas chromatography - mass spectrometry. iii A CKNOWLEDGEMENTS First, I would like to thank my advisor, Dr. Ruth Smi th, for her guidance, support, an d high expectations and reassurance, I would not have grown to be the forensic scientist and microscopist I am today. I would like to acknowledge the Michigan State Forensic Science Program for funding this work and for support to present this research. Additionally, I would like to extend a huge especially Sgt. Jim Dunlop for providing case samples used thro ughout this work. My time at Kalamazoo was made worthwhile by the kind words from the crime lab, coffee breaks, and endless talks over donuts. Additionally, I would like to thank Dr. Richard Staples and the Center for Cryst allographic Research for their as sistance with single crystals XRD analysis. Further, I would like to acknowledge my research committee: Dr. Ruth Smith, Dr. Jennifer Cobbina, Dr. Richard Staples, and Sgt. Jim Dunlop for their input, help, and challenging q uestions that have culminated in this immense accomplishment in my life. I would like to extend a big thank you to my forensic science colleagues, as well as Dr. Victoria McGuffin, for their critical insight throughout this work. Your questions and comme nts encouraged me to explore this work more deeply and I have gained a wealth of knowledge from our time together. I would be nowhere without the unending support of my friends and family. Thank you to my wonderful friends, near and afar, for endless pho ne calls akin to therapy, good be er, better gin, and laughter - filled distractions from the stress of grad school. A massive, and well deserved, iv thank you to Niko (and Dante, of course) you gave me peace, kept me sane, and constantly reminded me I can acc omplish anything as long as I tak e it one step at a time. Finally thank you to my family, especially my mother, father, and sister Marci, for their love, support, and inspiration. All glory to God . v TABLE OF CONTENTS LIST OF TABLES ................................ ................................ ................................ ....................... viii LIST OF FIGURES ................................ ................................ ................................ ........................ x 1. INTRODUCTION ................................ ................................ ................................ ...................... 1 1.1 CANNABIS SATIV A MARIJUANA AND HEMP ................................ ......................... 1 1.2 CANNABIS SOLVENT EXTRACTS ................................ ................................ .................. 2 1.3 FORENSIC ANALYSIS OF SEIZED DRUGS ................................ ................................ ... 4 1.4 ADDRESSING THE IDENTIFICATION OF CANNABIS EXTRACTS .......................... 5 1.5 RESEARCH OBJECTIVES ................................ ................................ ................................ . 6 REFERENCES ................................ ................................ ................................ ............................... 9 2. OPTICAL CHARACTERIZATION OF CRYSTALLINE STRUCTURES IN CANNABIS SOLVENT EXTRACTS ................................ ................................ ................................ ............... 12 2.1 POLARIZED LIGHT MICROS COPY ................................ ................................ ............... 14 2.1.1 Observations in Plane - Polarized Light ................................ ................................ ......... 16 2.1.2 Observations in Crossed - Polarized Light ................................ ................................ ..... 18 2.1.3 Conoscopy and Interference Figures ................................ ................................ ............ 21 2.1.4 De termining the Principle Refractive Indices using Biaxial Refractometry ................ 27 2.2 MATERIALS AND METHODS ................................ ................................ ........................ 34 2.2.1 Samples ................................ ................................ ................................ ......................... 34 2.2.2 Sample Preparation Techniques and Macroscopic Observations ................................ . 35 2.2.3 Polarized Light Microscopy ................................ ................................ ......................... 36 2.2.3 Refractive Index Determinations using Biaxial Refractometry ................................ ... 37 2.3 RESULTS OF OPTICAL CHARACTERIZATION BY POLARIZED LIGHT MICROSCOPY ................................ ................................ ................................ ......................... 39 ................................ ............. 39 2.3.2 Skymint Dispensary BHO Samples ................................ ................................ .............. 47 2.3.3 Cannabidiol Life Dispensary Samples ................................ ................................ ......... 53 2.4 DISCUSSION AND COMPARISON OF CRYSTALLINE CHARACTERIZAT ION BY POLARIZED LIGHT MICROSCOPY ................................ ................................ ..................... 62 APPENDIX ................................ ................................ ................................ ................................ ... 65 REFERENCES ................................ ................................ ................................ ............................. 73 3. SPECTROSCOPIC CHAR ACTERIZATION AND IDENTIFICATION OF CANNABIS SOLVENT EXTRACT COMPONENTS ................................ ................................ ..................... 75 3.1 INSTRUMENTAL THEORY ................................ ................................ ............................. 77 3.1.1 Micro - Attenuated Total Reflectance - Fourier Transform Infra - Red Spectroscopy ...... 77 vi 3.1.2 X - Ray Diffraction Theory and Instrumentation ................................ ........................... 80 3.2 MATERIALS AND METHODS ................................ ................................ ........................ 86 3.2.1 Samples ................................ ................................ ................................ ......................... 86 3.2.2 Micro - Attenuated Total Reflectance - Fourier Transform In fra - Red Spectroscopy Sample Preparation ................................ ................................ ................................ ................ 87 3.2.3 Single - Crystal X - Ray Diffraction Sample Preparation ................................ ............. 88 3.3 FTIR SPECTROSCOPIC ANA LYSIS OF CANNABIS SOLVENT EXTRACTS .......... 89 3.3.1 Results of Micro - ATR - FTIR Analysis of Cannabis Solvent Extracts ......................... 89 3.3.1.1 Micro - ATR - FTIR Analysis of Optically Similar Crystals - KCSD Case Samples and Skymint Samples ................................ ................................ ................................ ........ 89 3.3.1.2 Micro - ATR - FTIR Analysis of Cannabidiol Life Dispensary Samples ................ 95 3.3.2 Discussion of Micro - ATR - FTIR Analysis of Cannabis Solvent Extracts .................... 99 3.4 STRUCTURUAL ELUCIDATION OF CANNABIS SOLVENT EXTRACT CRYSTALS ................................ ................................ ................................ ................................ ................. 105 3.4.1 Results of Single - Crystal XRD Analysis for Structural Elucidation .......................... 10 5 3.4.1.1 Single - Crystal XRD Analysis of Optically Similar Samples - KDPS 18 - 9026 and WB THCA Crystals ................................ ................................ ................................ ........ 105 3.4.1.2 Single - Crystal XRD Analysis of CBD Shatter Cry stals ................................ ..... 108 3.4.2 Discussion of Single - Crystal XRD Analysis for Structural Elucidation .................... 111 APPENDIX ................................ ................................ ................................ ................................ . 119 REFERENCES ................................ ................................ ................................ ........................... 130 4. OPTIMIZATION OF THCA DERIVATIZATION USING AN EXPERIMENTAL DESIGN APPROACH ................................ ................................ ................................ ............................... 133 4.1 THEORY ................................ ................................ ................................ ........................... 135 4.1.1 Derivatization Methods ................................ ................................ .............................. 135 4.1.2 Experimental Design ................................ ................................ ................................ . 138 4.1.2.1 Full Factorial Screening Design ................................ ................................ .......... 140 4.1.3 Gas Chromatog raphy - Mass Spectrometry ................................ ................................ .. 143 4.2 MATERIALS AND METHODS ................................ ................................ ...................... 145 4.2.1 Reference Materials and Sample Prepara tion ................................ ............................. 145 4.2.2 Full Factorial Screening Experiments ................................ ................................ ........ 146 4.2.2.1 Sample Preparation - THCA Concentration Study ................................ ............. 148 4.2.2.2 Sample Preparation Cannabis Extract Sample Analysis ................................ .. 149 4.2.3 GC - MS Analysis ................................ ................................ ................................ ......... 149 4.2.4 Data Processing and Analysis ................................ ................................ ..................... 150 4.2.4.1 Full Factorial Analysis ................................ ................................ ....................... 150 4.3 FULL FACTORIAL SCREENING EXPERIMENT RESULTS AND DISCUSSION ... 151 vii 4.3.1 Full Factorial Screening Design Experiments and Optimization ............................... 155 4.3.2 Evaluation of Optimized Method Linearity ................................ ................................ 162 4.3.3 Analysis of Cannabis Solvent Extracts by GC - MS Using the Optimized Derivatization Procedure ................................ ................................ ................................ ............................. 164 4.4 CONCLUSIONS ................................ ................................ ................................ ............... 173 APPENDIX IV ................................ ................................ ................................ ............................ 174 REFERENCES ................................ ................................ ................................ ........................... 185 5. CONCLUSIONS AND FUTURE WORK ................................ ................................ ............. 188 5.1 CONCLUSIONS ................................ ................................ ................................ ............... 188 5.2 FUTURE WORK ................................ ................................ ................................ .............. 192 REFERENCES ................................ ................................ ................................ ........................... 193 viii LIST OF TABLES Table 2.1 Summary of crystal systems and parameters 9 ................................ ............................... 15 Table 2.2 Common interference figures for biaxial refractometry ................................ ............... 32 Table 2.3 Sample identifications, sources, and year obtained ................................ ...................... 35 Table 2.4 KCSD BHO case sample wax consistency and crystalline component summar ies ..... 39 Table 2.5 Summary of Case Sample Optical Properties by PLM ................................ ................. 46 Table 2.6 Skymint BHO macroscopic sample summary ................................ .............................. 47 Table 2.7 Summary of optical characteristics for Skymint dispensary samples ........................... 53 Table 2.8 Cannabidiol Life CBD sample summary ................................ ................................ ...... 54 Table 2.9 Summary of optical characteristics for Cannabidiol Life dispensary samples ............. 61 Table 2.10 Su mmary of optical characteristics from represen tative samples of each subset ....... 62 Table A2.1 Table of refractive index measurements (n ) for KDPS 18 - 9026 .............................. 68 Table A2.2 Table of refr active index measurements (n ) for KDPS 18 - 9026 .............................. 68 Table A2.3 Table of refractive index measurements (n ) for KDPS 18 - 9026 .............................. 69 Table A2.4 Table of refractive index measurements for PPO 14 - 20332 - 10 ................................ 69 Table A2.5 Table of refractive index measurements for KCSD 14 - 10811 - 28964 ....................... 70 Table A2.6 Table of refractive index measurements for KCSD 14 - 10811 - 28967 ....................... 70 Table A2.7 Table of refractive index measurements for KCSD 14 - 10811 - 28960 ....................... 70 Table A2.8 Table of refractive index measurements for WB THCA Crystals ............................. 71 Table A2.9 Table of refractive index measurements for PB THCA Wax ................................ .... 71 Table A2.10 Table of refractive index measurements for CBD Wax ................................ ........... 72 Table A2.11 Table of refractive index measurements for CBD Crystal ................................ ....... 72 ix Table 3.1 Summary of crystal systems with geometric unit cell depictions ................................ . 82 Table 3.2 Sample identifications, sources, and year obtained ................................ ...................... 86 Table 3.3 KCSD and Skymint sample summary ................................ ................................ .......... 90 Table 3.4. Cannabidiol Life CBD - containing sample summary ................................ ................... 96 Table 3.5 Crystal data and structure refinement details of KDPS 1 8 - 9026 and WB THCA Crystal samples ................................ ................................ ................................ ................................ ........ 107 Table 3.6 Crystal data and structure refinement details of CBD Shatter Crystal sample ........... 110 Table 3.7 Comparison of refined data for crystals analyzed by single - crystal XRD ................. 117 Table 4.1 Summar y of Resolution and Confounding Variables ................................ ................. 140 Table 4.2 Example of Experimental Order for Full Factorial Design with Three Factors ......... 141 Table 4.3 Full Factorial Levels for Each Factor ................................ ................................ ......... 147 Table 4.6 Summary of experiment order, levels, averaged THCA - 2TMS abundance, and RSD ................................ ................................ ................................ ................................ ..................... 156 Table 4.7 Optimum derivatization reaction parameters ................................ .............................. 161 Table 4.8 Summary of concentration study normalized THCA - 2TMS abundances and RSDs . 163 T able A4.1 Calculations for the degrees of freedom ................................ ................................ .. 175 Table A4.2 Calculations for the sum of squares for two - way ANOVA 19 ................................ .. 175 Table A4.3 Calculations for the mean squares ................................ ................................ ........... 176 Table. A4.4 Summary of reproducibility for pre - screening hold experiments ........................... 179 Table A4 .5 Inter - and intra - vial RSDs for ethyl acetate and pyridine pre - screening experiments ................................ ................................ ................................ ................................ ..................... 179 Table A4.6 Full ANOVA results from derivatizations using pyridine as support solven t ......... 179 Table A4.7 ANOVA results from derivatizations using ethyl acetate as support solvent .......... 180 x LIST OF FIGURES Figure 2.1 Labeled diagram of a polarized light microscope ................................ ....................... 14 Figure 2.2 Example of relief differences relative to refractive index medium; (a) high relie f, (b) low relief ................................ ................................ ................................ ................................ ....... 17 Figure 2.3 Diagram illustrating the interaction of light with isotropic media (left) and anisotropic media (right) ................................ ................................ ................................ ................................ .. 19 Figure 2.4 Michel - L é vy Birefringence Chart 11 ................................ ................................ ............. 20 Figure 2.5 Example of a uniaxial interference figure with key features labeled .......................... 23 Figure 2.6 Conoscopic images of uniaxial interference figure orientation changes with microscope stage rotation 12 ................................ ................................ ................................ ........... 24 Figure 2.7 Example of a biaxial interference figure ................................ ................................ ..... 25 Figure 2.8 Estimation of biaxial 2V angle based on isogyre curvature in an optic axis interference figure 10 ................................ ................................ ................................ ................................ .......... 26 Figure 2.9 Uniaxial indicatrix with principle refractive index views (a) optic normal and (b) optic axis ................................ ................................ ................................ ................................ ................ 28 Figure 2.10 Biaxial indicatrix with principle refractive index views (a) optic normal and (b) optic axis, and (c) ob tuse bisectrix (Bxo) ................................ ................................ .............................. 29 Figure 2.11 O ptically positive (A) and optically negative (B) biaxial indicatrices ...................... 31 Figure 2.12 (A) Beck e line moving out into higher refractive index medium and (B) into a sample crystal with higher refractive index ................................ ................................ .................. 33 Figure 2.13 Macroscopic view of case sample KDPS 18 - 9026 in glass vial ............................... 40 Figure 2.14 Photomicrograph of KDPS 18 - 9026 at 100X magnification in PPL ........................ 41 Figure 2.15 Photomicrograph of KDPS 18 - 9026 at 100X magni fication in XPL ........................ 42 Figure 2.16 KCSD 18 - 9026 with polarizer at (A) 0° orientation and (B) 90° orientation to demonstrate changes in relief ................................ ................................ ................................ ........ 42 Figure 2.17 Conoscopic view of a KDPS 18 - 9026 crystal displaying an optic axis interference figure ................................ ................................ ................................ ................................ ............. 44 xi Figure 2.18 Macroscopic view of PB THCA wa x (A) and WB THCA crystals (B) in dispe nsary glass containers ................................ ................................ ................................ ............................. 48 Figure 2.19 Stereoscopic view of PB THCA wax (A) and WB THCA crystal s (B) showing the crystal color in reflected light and agglomerated form ................................ ................................ . 49 Figure 2.20 PB THCA wax crystals shown in PPL, note the oily, wax component separate from the crystalline component ................................ ................................ ................................ ............. 49 Figure 2.21 PB THCA wax crystals w ith polarizer at (A) 0° orientation and (B) 90° orientation to demonstrate changes in relief ................................ ................................ ................................ ........ 50 Figure 2.22 Small PB THCA wax crystals in XPL displaying moderate retardation (A) and large cryst als showing high order white retardation (B) ................................ ................................ ........ 51 Figure 2.23 Conoscopic view of a PB THCA wax crystal displaying an optic axis interference figure ................................ ................................ ................................ ................................ ............. 52 Figure 2.24 Macroscopic view o f CBD wax (A) and CBD crystal (B) samples in dispensary containers ................................ ................................ ................................ ................................ ...... 55 Figure 2.25 Stereoscopic view of CBD wax (A) and CBD crystals (B) showing the crystal color in reflected light and the difference in wax presence. ................................ ................................ .. 55 Figure 2.26 CBD wax crystals shown in PPL, note the oily, wax component separate from the crystalline component ................................ ................................ ................................ ................... 56 Figure 2.27 CBD wax crystals with polarizer at (A) 0° orientation and (B) 90° orientation to demonstrate changes in relief. ................................ ................................ ................................ ....... 57 Figur e 2.28 CBD wax crystalline co mponent in viewed in XPL ................................ ................. 58 Figure 2.29 XPL view of CBD Wax crystal with straight edge parallel to the crosshair of the ocular micro meter (red) ................................ ................................ ................................ ................ 59 Figure 2.30 Conoscopic view of the CBD crystal sample displaying a biaxial optic axis interference figure ................................ ................................ ................................ ......................... 60 Figure A2.1 Macroscopic (A) and micros copic views of KCSD case sample PPO 14 - 20332 - 10 in PPL (B) and XPL (C) ................................ ................................ ................................ .................... 66 Figure A2.2 Macroscopic (A) and microscopic views of KCSD case sample KCSD 14 - 10811 - 28964 in PPL (B) and XPL (C) ................................ ................................ ................................ ..... 66 Figure A2.3 Macroscopic (A) and microscopic view s of KCSD case sample KCSD 14 - 10811 - 28960 in PPL (B) and XPL (C) ................................ ................................ ................................ ..... 66 xii Figure A2.4 Macros copic (A) and microscopic views of KCSD case s ample KCSD 14 - 10811 - 28967 in PPL (B) and XPL (C) ................................ ................................ ................................ ..... 67 Figure A2.5 Microscopic views of Skymint dispensary sample WB THCA crystals in PPL (A) and XPL (B) ................................ ................................ ................................ ................................ .. 67 Figure A2.6 Microscopic views of Cannabidiol Life dispensary sample CB D crystals in PPL (A) and XPL (B). ................................ ................................ ................................ ................................ . 67 Figure 3.1 Smit - ATR - FTIR instrumental set - up ................................ ... 79 Figure 3.2 Simplified schematic of an ATR microscope objective and infrared spectrometer commonly used for microscopic infrared analysis ................................ ................................ ....... 80 16 ................................ ..................... 84 Figure 3.4 KDPS 18 - 9026 crystal (top) and wax component (bott om) IR spectrum ................... 91 Figure 3.5 Representative Skymint sample - WB THCA Crystal IR spectrum ............................ 91 Figure 3.6 Stacked FITR spectra for a ll sample crystalline components ................................ ..... 92 Figure 3.7 Stacked FITR spectra for all case s ample wax components ................................ ........ 93 Figure 3.8 Stacked compa rison of THCA CRM (bott om) and case sample KDPS 18 - 9026 crystal (middle) and wax component (top) micro - ATR - FTIR spectra ................................ ..................... 94 Figure 3.9 Stacked comparison of THC CRM (bottom) and case sample KDPS 18 - 9026 crystal (middle) and wax component (top) micro - ATR - FTIR spectra ................................ ..................... 95 Figure 3.10 Micro - ATR - FTIR spectrum for the crystalline component of sample CBD Wax crystal (bottom) and wax (top) co mponents ................................ ................................ ................. 97 Figure 3.11 Micro - ATR - FTIR spectrum for sample CBD Shatter Crystals ................................ . 97 Figure 3.12 Stacked comparison of CBD Shatter C rystal (top) and THCA CRM (bottom) ........ 98 Figure 3.13 Stacked spectra comparing the crystalline components of KDPS 18 - 9026 (bottom), WB THCA Crystals (middle), and CBD Shatter Crystals (top) ................................ ................. 104 Figure 3.14 Crystal structure from case sample KDPS 18 - 9026 determined by single - crystal XRD shown with molecular labelling scheme. ................................ ................................ ........... 106 Figure 3.15 Crystal structure from case sample KDPS 18 - 9026 displaying racemic, dimer crystalline properties ................................ ................................ ................................ ................... 106 xiii Figure 3.1 6 Crystal structure from dispensary sample CBD Shatter Crystal det ermined by single - crystal XRD shown with molecular labelling scheme of chiral atoms. ................................ ...... 109 Figure 3.17 Crystal structure from dispensary sample CBD Shatter Crystal displaying hydrogen bonding and dimer crystalline properties ................................ ................................ .................... 109 Figure 3.18 Comparison of single - crystal XRD refined structures for KDPS 18 - 9026 (A) and CBD Shatter Crystals (B) ................................ ................................ ................................ ............ 117 Figure A3.1 Geometric depictions of Bravais lattices 25 ................................ ............................. 120 Figure A3.2 PPO 14 - 20332 - 10 micro - ATR - FTIR spectrum ................................ ...................... 121 Figure A3.3 KCSD 14 - 10811 28967 micro - ATR - FTIR spectrum ................................ .......... 121 Figure A3.4 KCSD 14 - 10811 28960 micro - ATR - FTIR spectrum ................................ .......... 122 Figure A3.5 KCSD 14 - 10811 28964 micro - ATR - FTIR spectrum ................................ .......... 122 Figure A3.6 Skymint PB THCA Wax crystalline component spectrum ................................ .... 123 Figure A3.7 PPO 14 - 20332 - 10 wax component micro - ATR - FTIR spectrum (baseline corrected) ................................ ................................ ................................ ................................ ..................... 123 Figure A3.8 KCSD 14 - 10811 28967 wax component micro - ATR - FTIR spectru m (baseline corrected) ................................ ................................ ................................ ................................ .... 124 Figure A3.9 KCSD 14 - 10811 28960 wax component micro - ATR - FTIR spectrum (baseline corrected) ................................ ................................ ................................ ................................ .... 124 Figure A3.10 KCSD 14 - 10811 28964 wax component micro - ATR - FTIR spectrum (baseline corrected) ................................ ................................ ................................ ................................ .... 125 Figure A3.11 KDPS 18 - 9026 XRD packing diagram ................................ ................................ 125 Figure A3.12 WB THCA Crystal XRD structure showing chiral centers ................................ .. 126 Figure A3.13 WB THCA Crystal XRD structure showing hydrogen bonding .......................... 126 Figure A 3.14 WB THCA Crystal XRD packing diagram ................................ .......................... 127 Figure A3.15 CBD Shatter Crystal XRD packing diagram ................................ ........................ 128 Figure A3.16. THCA chemical structure ................................ ................................ .................... 128 Figure A3.17. CBD chemical structure ................................ ................................ ....................... 129 xiv Figure 4.1 Example TMS derivatives: (A)hydroxyl, (B) carboxyl, and (C) amide .................... 136 Figure 4.2 Averaged chromatograms of THCA derivatizations using ethyl acetate and pyridine ................................ ................................ ................................ ................................ ..................... 151 Figure 4.3 Mass spectrum and c hemical structure of THCA - 2TMS ................................ .......... 153 Figure 4.4 Mass spectrum and chemica l structure of THC - TMS ................................ ............... 154 Figure 4.5 Overlay of averaged c hromatograms from low (dark blue), high (light blue), and center point (dashed) level pyridine supported derivatizations. ................................ ................. 158 Figure 4.6 Overlay of averaged chromatograms from low (dark green), high (light green), and center point (dashed) level ethyl acetate supported derivatizations ................................ ........... 158 Figure 4.7 THCA - 2TMS abundance changes with changes in reaction temperature ................. 160 Figure 4 .8 Overlay of averaged chromatograms from each concentration study ....................... 162 Figure 4.9 Chromatogram of case sample KDPS 18 - 9026 ................................ ......................... 165 Figure 4.10 Mass spectrum and chemical structure for CBN - TMS ................................ ........... 166 Figure 4.11 Mass spectrum for unidentified peak at 22.2 minutes ................................ ............. 166 Figure 4.12 Mass spectrum and chemical structure for THCA - 2TMS ................................ ....... 167 Figure 4.13 Chromatogram of Skymint THCA Crystal sample ................................ ................. 169 Figure 4.14 Mass spectrum and chemical formula for THCA - 2TMS ................................ ........ 169 Figure 4.15 Chromatogram of Cannabidiol Life CBD Crystal sample ................................ ...... 171 Figure 4.16 Mass spectrum and chemical formula for CBD ................................ ...................... 171 Figure 4.17 Mass spectrum and chemical structure of CBD - 2TMS ................................ ........... 172 Figure A4.1 THC mass spectrum ................................ ................................ ................................ 176 Figure A4.2 NIST library mass spectrum result for THC - TMS ................................ ................. 177 Figure A4.3 Averaged chromatograms comparing refrigerated and 24 hour hold samples of THCA derivatization using ethyl acetate ................................ ................................ .................... 178 Figure A4.4 Averaged chromatograms comparing re frigerated and 24 hour hold samples of THCA derivatization using pyridine ................................ ................................ ........................... 178 xv Figure A4.5 Pareto chart for the pyridine full factorial design ................................ ................... 180 Figure A4.6 Pareto chart for ethyl acetate full factorial design ................................ .................. 181 Figure A4.7 Regression plot for THCA concentration study ................................ ..................... 182 Figure A4.8 Manufacturer information and quantification for Skymint THCA Crystal sample 183 Figure A4.9 NIST library mass spectrum for CBN - TMS ................................ ........................... 183 Figure A4.10 NIST library mass spectrum for CBD ................................ ................................ .. 183 Figure A4.11 NIST library mass spectrum for CBD - 2TMS ................................ ....................... 184 1 1. INTRODUCTION 1.1 CANNABIS S ATIVA MARIJ UANA AND HEMP Cannabis sativa is federally controlled in the Controlled Substances Act (CSA) as a Schedule I substance, indicating that it has no acce pted medical use and a high risk for addiction amongst users. 1 The psychoactive co mponent of Cannabis sat iva responsible for the associated - 9 - 9 - 9 - THC is found in the resinous capitate (globular) trichomes on the leaves and bracts of female Cannabis sativa plants. 2 These globular trichomes also contain over 100 addition al 9 - THC, but these cannabinoids do not demonstrate psyc hoactive effects and are being extensi vely researched for their pharmacological activities. One such cannabinoid of inte rest for the presented research is del ta - 9 - 9 - THCA), as this cannabinoid is 9 - THC thr ough thermal decarboxylation and is of ten present in marijuana products. Additionally, in recent years, the cannabinoid cannabidiol (CBD) has been researched and marketed for its medical benefits. 2 Marijuana and hemp represent two broad classes of Cannab is sativa plants. While both marijuana and hemp originate from the Cannabis sativa species, they differ in their cannabi noid composition and their legality in the United States. The 2018 Unit e d States Agriculture Improvement Act (referred to as the Farm Bi ll ) , which approved the legality of he mp and its products, distinguishes hemp and marijuana on the basis of 9 - THC concentration. 3 As stated in the Farm Bill, any Cannabis sativa plant with less than 0.3% THC (by dry weight) is considered hemp, while any p lant with THC content above that thres hold is still considered marijuana. As such, the 2018 Farm Bill ratifi ed the production and sale of hemp products while maintaining marijuana as a Schedule I drug. 2 1.2 CANNABIS SOLVENT EXTRACTS The isolation and concentration of cannabinoids for the prod uction of cannabis extracts can be performed in a wide range of techni ques. While the production of extracts is more commonly performed for the concentration of THC from marijuana, the recreationa l sale of CBD products has influenced the production of cann abis extracts derived from hemp. Regardless of plant material used, si milar techniques can be used to concentrate cannabinoids. Such techniques include manual separation ( e.g., h ashish), heat and pressure extractions ( e.g., r osin), and solvent extractions ( e.g., b utane h ash o il). Solvent extractions can be produced using a v ariety of solvents, but most commonly hydrocarbons , such as butane or pentane , or supercritical CO 2 ( CO 2 in a liquid state) ar e used . 4 The solvent selected for extraction impacts the ov erall concentration of cannabinoids and terpenes present in the final extracted product. Hydrocarbon extractions have been used historically, due to their efficiency for the extraction of both cannabinoids and terpenes from the plant material . 4 Supercritic al CO 2 extraction is a relatively new technique when applied to cannab process for the selection of specific cannabinoids and terpenes. Tun ing this extraction method by manipulating the pressure and temperature of the extraction provides a variety of products, ranging from pure cannabinoid extracts to mixtures of cannabinoids and terpenes. 5 The extraction of cannabinoids , specifically CBD , fr om hemp is most commonly perf ormed using supercritical CO 2 . Butane hash oil (BHO) is a general cl ass of cannabinoid extract using butane as the extract solvent. Butane hash oil is colloquially referred to as wax, shatter, or crumble each referring to d ifferent textures of product made from similar extraction processes. Similar textures 3 can be achiev ed using supercritical CO 2 extraction. Due to the solubility of the cannabinoids in butane and supercritical CO 2 , their extraction from plant material into the final BHO product increase s 9 - THC or CBD (depending on extraction from mar ijuana or hemp) in products, such that BHO extracts are six to eight times more potent than plant material alone. 6 Prior to recreational legal ization in a number of states, clandestine pr oduction of BHO extracts w ere performed using a simple, yet dangerous , procedure. In clandestine laboratories, BHO was generally produced by packing a glass or steel column with cannabis plant material, fitting a filter to the end, passing butane through t he column, collecting the extracted material in a glass dish, and all owing the solvent to evaporate. As the solvent (butane) passed through the column, the cannabinoids were readily dissolved and extracted from the trichomes. The danger of clandestine oper ations was due to the low vapor pressure of butane which , when evapor ated from the extract , can result in explosions in open systems. While state - wide legalization has allowed for safer, closed - loop manufacturin g systems to be used for large - scale BHO prod uction for recreational sale, clandestine manufacturing of BHO is sti ll performed. The Bureau of Alcohol, Tobacco, Firearms and Explosives (ATF) warns that clandestine BHO production may be greater in states wit h legalized marijuana, as access to larger am ounts of plant material will be easier than in states where recreatio nal sale s remain illegal. 7 S tates with recreational marijuana , including Michigan and California, have combatted this by outlawing the use of butane or other hydrocarbons for the clandest ine production of BHO extracts . 8,9 The popularity of cannabis extrac ts has grown with the federal legality of hemp and the state - wide legalization of recreational ma rijuana. From 2016 to 2019, the Colorado Marijuana Enforcement Division reported a 128% in crease in cannabis extract or concentrate sales. 10 4 Similar growth has been shown internationally with Canadian recreational extract and concentrate sales increas ing 1 54% between 2018 and 2019. 10 Data specifically on the sale of hemp extracts is limited, ho wever economic projections related to CBD products as a whole estimat e the CBD product market will be a $20 billion dollar industry in the United States alone by 2024 . 11 1.3 FORENSIC ANALYSIS OF SEIZED DRUGS The Scientific Working Group for the Analysis of Seized Drugs (SWGDRUG) is a committee working to improve and stan dardize the forensic analysis of illicit drugs. 12 To improve the quality of drug analysis, SWGDRU G has compiled recommendations as a guideline for forensic laboratories to follow as minim um requirements for drug identification. The guidelines follow an ana lytical scheme comprised of techniques that can be categorized (A, B, or C) based on their specif icity and discriminating power. Category C techniques include color tests or melting point measurements are the least discriminatory, providing only general or class characteristics. Category B techniques , such as microchemical tests and microscopic analys is (specifically for cannabis), and various chromatography methods, provide chemical or ph ysical characteristics, making them more discriminatory and selective than Category C techniques. Category A techniques include mass spectrometry, infrared spectrosco py, and X - ray diffraction and are the most selective and discriminatory, providing structu ral information toward identification for drug samples. 12 Minimum rec ommendations for the identification of drug samples include the use of one Category A technique w ith either another Category A, B, or C. When a Category A technique is not available, thre e separate techniques are necessary for identification with at least two Category B techniques. 12 5 1.4 ADDRESSING THE IDENTIFICATION OF CANNABIS EXTRACTS The product ion of cannabis extracts, from both marijuana and hemp, results in a variety of textures o f final extracted products. These textures range from viscous to glas sy, as well as highly crystalline. The current analysis of cannabis extracts utilizes primarily a Duquenois - Levine color test for the presence of cannabinoids followed by gas chromatograp hy - mass spectrometry (GC - MS) for the identification of cannabinoids ( specifically 9 - THC). While this analytical scheme does provide for the identification of the scheduled, federally illegal cannabinoid 9 - THC, little research has been completed characterizing additional cannabinoids or other components of these extracts u sing alternat ive SWGDRUG recommended techniques. While many cannabis solvent extract s appear completely amorphous macroscopically, through microscopic analysis , two distinct components can typically be observed for cannabis solvent extracts. These compon ents include an amorphous wax and well - formed crystalline material. T he analysis of th ese individual components has not been completed using the aforementioned common methods, and a s such the separate components have not been characterized or identified. G iven the macr oscopic similarity of solvent extracts derived from marijuana and hemp, t he presence of crystalline material in both may allow for rapid screening based on the optical properties of the crystals . Differentiating marijuana - based and hemp - based solvent extra cts by a rapid screening method may provide presumptive identification, m ore specific than a Duquenois - Levine color test, prior to identification using a SWGDRUG recommended Category A technique. Additionally, with the passage of the 2018 Farm Bill, the di fferentiation of marijuana and hemp - based products is necessary in a fore nsic context given the legality of most CBD products (derived from hemp) as opposed to the state - specific recreational legality of THC products. 6 Similarly, the state - wid e mandates re garding the identification of THC potency in products differs across the United States. As such, some forensic laboratories are only responsible for the identification and quantification of total THC potency (THC plus THCA) while others need t o identify an d quantify both the cannabinoid acids and neutrals. The extensive use of GC - MS for the quantification of cannabinoids in forensic lab oratorie s readily identifies and quantifies cannabinoid neutrals. The identification of cannabinoid acids, how ever, require s sample derivatization, as decarboxylation of the acids rapidly occurs i n the injection port of the GC - MS. This decarboxylation converts the cannabinoid acids into their neutral forms ( i.e. THCA to THC). In order to combat this issue, researc h has been pe rformed in the forensic community regarding the identification of total T HC potency, 13,14 as well as determining methods through which both the neutral and acidic cannabinoids can be quantified and identified. 1 5 ,1 6 Though methods of cannabinoi d acid deriva tization are available and utilized, this procedure has not been optimize d using single cannabinoids, but rather most utilize a mixture of cannabinoid acids and neutrals in their development. By designing and optimizing the method using a mixt ure of both c annabinoid acids and neutrals, the extent of decarboxylation occurring du e to underivatized acidic product cannot be readily differentiated from the abundance of the neutral cannabinoid itself. As such, optimization using a single cannabinoid acid is neces sary to qualitatively observe the extent of reaction to ensure the comple te derivatization of each cannabinoid acid. 1. 5 RESEARCH OBJECTIVES The proposed research is intended to reduce the knowledge gap present for cannabis extracts by prov id ing a compr ehensive optical and chemical characterization for the two components, with much focus placed on the characterization and identification of the crystalline component. This was accomplished through the analysis of both forensic laboratory case sampl es 7 as we ll as commercially available marijuana solvent extract s and hemp - derived CBD extracts. Optical characterization was completed using polarized light microscopy (PLM), while chemical characterization and identification was performed using micro - atten uated to tal reflectance - Fourier transform infrared spectroscopy (micro - ATR - FTIR) , single crystal X - ray Diffraction (XRD) , and GC - MS. Due to the crystalline material present in the cannabis solvent extracts, extensive optical characterization by PLM w as pe rformed. Using PLM, the morphology, refractive indices, and characteristics in crossed - polarized light and conoscopic view were used to group crystals from each class of cannabis solvent extract (either marijuana or hemp - derived). Next, spectroscopic analy sis of b oth the wax and crystalline components was performed using micro - ATR - FTIR . The chemical characterization by micro - ATR - FTIR spectroscopy allowed for optically similar crystals to be compared further by structural information. Additionally, anal ysis by micro - ATR - FTIR provided preliminary identification of both the crystalline and wax components of each cannabis solvent extract. The confirmatory identification of the crystalline components of select crystalline samples was completed by single - crys tal X RD. Two subsets of crystals from cannabis solvent extracts two from marijuana - derived extracts and one from a hemp - derived extract were identified and spectroscopically characterized by single - crystal XRD to identify the chemical structure as well as t he dimen sions and angles of the unit cell to further associate the optically similar crystals in each solvent extract subset. Additionally, due to the common presence of THCA in marijuana - derived extracts , derivatization studies focused on optimizing the derivati zation reaction (silylation) of THCA was accomplished using an experimental design procedure . Specifically, a published derivatization procedure was optimized in terms of solvent choice, temperature of derivatization, reaction time, 8 and solven t to derivati zing agent ratio. 15 Further, the experimental design followed a full factorial screening analysis to determine significant factors for variability in the abundance of derivatized product when analyzed by GC - MS. The optimized procedure was then used to chem ically characterize a subset of cannabis solvent extracts produced from both marijuana and hemp. While cannabis solvent extracts were solely analyzed in this work, t he optimization of this reaction allows for confident identification and quant ifica tion of THCA in a variety of cannabis products and is not exclusively limited to cannabis solvent extracts. 9 REFERENCES 10 R EFERENCES (1) Drug Enforcement Administration. Drug Scheduling https://www.deadiversion.usdoj.gov/synthetic_drugs/about_ sd.html. (2) Andr e CM, Hausman J - F, Guerriero G. Cannabis sativa: The Plant of the Thousand and One Molecules. Frontiers in Plant Science 2016;7. (3) Agric ulture Improvement Act of 2018 (P.L. 115 - 334), 2018. https://www.congress.gov/115/plaws/publ334/PLAW - 115publ334.pdf (4) Beadle , A. Advances in Cannabis Extraction Techniques. Analy tical Cannabis , 2019. https://www.analyticalcannabis.com/articles/advances - in - cannabis - extraction - techniques - 311772 (5) Rovetto LJ, Aieta NV. Supercr itical carbon dioxide extraction of cannabinoids from Cannabis sativa L. The Journal of Supercritical Flui ds 2017;129:16 27. (6) Elsohly MA, Mehmedic Z, Foster S, Gon C, Chandra S, Church JC. Changes in Cannabis Potency Over the Last 2 Decades (1995 2014): An alysis of Current Data in the United States. Biological Psychiatry 2016;79(7):613 9. (7) Hughes T. ATF warns of danger from 'hash oil' explosions USA Today. 2015; https://www.usatoday.com/story/news/nation/2015/05/23/hash - oil - explosions/27737635/ (8) Michigan Proposal 18 - 1. Ballot M easure. 2018; https://www.michigan.gov/documents/sos/Full_Text_ - _CRMLA_635255_7.pdf (9) California Legislative Information. Code Section 11379.6. http:// http://leginfo.legislature.ca.gov/faces/codes_displaySection.xhtml?lawCode=HSC§ion Num=11379.6 (10) MacIver B. Increasing Popularity of 7/10 Exem pl ifies Rise in C annabis Concentrates . Cannabis Business Times. 2020 ; https://www.cannabisbusinesstimes.com/article/710 - popularity - rises - concentrate - sales / (11) Dance A. As CB D Skyrockets in Popularity, Scientists Scramble to Understand How It's Metabolized. Scientific American. 2019 ; https://www.scientificamerican.com/article/as - cbd - skyrockets - in - popularity - scientists - scramble - to - understand - how - its - metabolized/ 11 (12) S cientific Working Group for the Analysis of Seized Drugs (SWGDRUG) Recommendations, 2019. http://swgdrug.org/Documents/SWGDRUG%20Recommendations%20Version%208_FINA L_ForPosting_092919.pdf . (13) James A. Cannabis Quality Control Testing Using Gas Chromato g r aphy. Cannabis Science and Technology 2018;1(2). (14) Ruppel TD, Kuffel N, PerkinElmer, Inc. Cannabis Analysis: Potency Testing Identification and Quantification of THC and CBD by GC/FID and GC/MS . Application Note. https://www.perkinelmer.com/lab - solutions/resources/docs/app_cannabis - analysis - potency - testing - identifification - and - quantification - 01184 1b _ 01.pdf (15) Rigdon A. Accurate Quantification of Cannabinoid Acids and Neutrals by GC. Restek.com, 2015. (16) Cardenia V, Toschi TG, Scappini S, Rubino RC, Rodriguez - Estrada MT. Development and validation of a Fast gas chromatography/mass spectrometry method fo r the determinati on of cannabinoids in Cannabis sativa L. Journal of Food and Drug Analysis 2018;26(4):1283 92. 12 2. OPTICAL CHARACTERIZATION OF CRYSTALLINE STRUCTURES IN CANNABIS SOLVENT EXTRACTS The use of microscopy as a forensic technique is most root e d in the trace evidence examination of well - formed crystalline materials commonly found during soil analysis, fibers, and amorphous material such as glass. Though seized drugs identification often utilizes instrumentation such as Fourier transform - inf rare d spectroscopy a nd gas chromatography - mass spectrometry, the microscopic characterization of crystalline materials proves beneficial when working with trace amounts of samples. While the use of polarized light microscopy (PLM) has diminished in forensi c ap p lications, rece nt literature highlights the use of PLM as a characterization method for a variety of crystalline samples outside of this field. 1 - 3 Most recent literature shows the use of PLM in the characterization of food and drug crystals, as well a s pl a nt and physiolo gical samples. The application of PLM to a variety of materials for identification and characterization is due to the ability to observe the same identifying characteristics of any sample type, as long as light can be transmitted though the sample. Polowsk y et al. utilized PLM to characterize and identify common surface crystals of smear - ripened cheeses, including ikaite, calcite, and brushite. 1 In this work they specifically utilized PLM for the ability to identify samples based on thei r re f ractive index a nd characteristics in crossed - polarized light (XPL). Further, they compared the identifications made using PLM to those obtained spectroscopically via powder X - ray diffractometry and demonstrated the similarities in identification betwe en t h e two methods. Duncke et al . demonstrated the use of PLM to characterize liquid crystals within Brazilian crude oil samples. 2 This work highlighted the use of PLM to differentiate and characterize lamellar liquid crystals and noted the environment - spe cifi c optical proper ties of such crystals. Such optical characteristics exploited in 13 this work were the morphology and micrometry of sample crystals crystallized in different oil environments and birefringence characteristics in XPL. Nie et al . employed PL M to study the cryst alline solid dispersion rate of salt disproportionation (the conversion from ionized to neutral state) for solid formations of active pharmaceutical ingredients. 3 Finally, PLM has been used extensively in the realm of synthetic and natu ral f iber identifica tion. 4,5 Most recently, Reffner et al . developed a new method by which to differentiate synthetic textile fibers using relative contrast (given by differences in refractive indices) and angle measurements rather than the by standard det ermi n ation of princi ple refractive indices. 6 Similar to aforementioned research regarding crystal characterization, drug analysis utilizes PLM to characterize crystals resulting from microchemical tests often used for presumptive identification of drug sa mple s . 7 Additionally , as many drug samples, such as methamphetamine and cocaine, themselves are crystalline, they too can be characterized using PLM and their properties can be compared against standard materials acquired by laboratories. In the case of t his w ork, crystallin e materials from cannabis solvent extracts of a variety of sources will be characterized and compared. Similar to the literature highlighted, sets of crystals will be characterized to determine similarities in chemical composition based on t he optical prop erties present. Establishing similarity between the optical characteristics of each crystal set will provide precedent that these crystals can be considered comparable in crystal system (unit cell) and chemical structure. 14 2.1 POLARIZED LI GH T MICROSCOPY W hile standard, compound microscopes and stereomicroscopes utilize an unpolarized light source, polarized light microscopes (PLMs) differ in the inclusion of two polarizing films one at the base of the microscope (the polarizer) and one f ol lowing the obj e ctive lens (the analyzer). The use of polarizing films effectively selects the direction of the light rays entering the sample, allowing only the light rays that are oriented in the East - West (E - W) direction to illuminate the sample. This co nfiguration is shown in Figure 2.1 , where the main components of a PLM are highlighted. Figure 2. 1 Labeled diagram of a polarized light microscope Polarized light microscopy is used to observe and record the optical characteristics of anisotropic mat e rials those w ith more than one refractive index. While materials that are isotropic, including glass and cubic crystals such as table salt, display similar optical characteristics regardless of the orientation of light, anisotropic materials have optica l properties th a 15 orienting parallel light rays to the crystallographic axes of a sample material, unique properties can be observed that are unavailable when using stan d ard compound m i croscopes. These unique properties include anisometry, pleochroism, refractive index, birefringence, and extinction characteristics. Utilizing the unique optical properties ascertained by PLM analysis, the crystal system of an unknown samp l e crystalline m aterial can be determined. The ability to establish the crystal system of an unknown sample is especially advantageous in narrowing do wn the list of potential identities of an unknown crystal and allows for comparison of crystallographic pr o perties to othe r techniques, including X - ray crystallography. Crystal systems relate to the shape of the unit cell, including the length and angle be tween crystallographic axes. Such a determination via PLM is made by considering the characteristics of th e sample, specif ically the number of principle refractive indices present and the extinction characteristics. Every crystal is a member of one of the crystals systems: cubic, tetragonal, hexagonal, rhombohedral, orthorhombic, monoclinic, and triclinic. 9 T a ble 2. 1 summari zes the 7 crystal systems and their parameters; a, een the three crystallographic axis a, b and c. A more complete explanation of crystal systems as they rel a te to unit cell s and Bravais lattices is discussed in S ection 3.1.2.1 . Table 2.1 Summary of crystal systems and parameters 9 Crystal System Parameters Cubic Tetragonal Hexagonal Rhombohedral 0 Orthorhombic Monoclinic Triclinic 16 2.1.1 Observations in Plane - Polarized Light Similar to microscopic analysis using standard compound microscopes, simple observations in P L M include morp hology, micrometry, and sample color in transmitted light. Morphology refers to the classification of sample s hape, including acicular (needle like), conchoidal fractures (irregularly curved and striated surfaces, like that of broken glass), cube (like tha t of table salt), and anhedral morphology for particles that lack apparent crystalline structures. Micrometry determines the size of the microscopic sample by utilizing a calibrated ocular micrometer. Sample color is recorded as the observed color in trans mitted, polarized light and may differ from that of the sample in reflected light. When light traveling through the microscope assembly is filtered by the polarizer, the illumination is referred to as plane - polarized light (PPL). As shown i n Figure 2.1, t he polarizer of the PLM sits at the base of the apparatus and filters the randomly - o riented light from the source bulb to an ordered, parallel beam of light traveling in only one direction (E - W). When observing samples in PPL, additional cha r acteristics ca n be determined, including pleochroism, optical relief, and refractive index. Such c haracteristics cannot be identified using standard compound microscopes, because orientation of the light in one direction is necessary. The determination of these addition al characteristics is due to the specific physical phenomena that occur when paralle l light interacts with the crystallographic axes and electron environments of anisotropic materials. 10 Pleochroism is a relatively rare feature displayed by c olored anisotr opic samples and refers to the change in color or brightness of the sample as the vi bration direction of light changes due to the interaction of parallel light with the refractive indices. 10 Optical relief refers to the definition of sample b oundaries and occurs as light is scattered from the uneven surface of the sample. Relief can be th ought of as the contrast between the sample and the 17 mounting medium (generally a Cargille refractive index liquid), and the extent of relief increases as the difference in refractive index between the sample and mounting medium increases ( Figure 2. 2 ). Reli ef is often used to quickly determine the relative refractive index of the sample as compared to the mounting medium. 9 Figure 2. 2 Example of relief differe nc es relative t o refractive index medium; (a) high relief, (b) low relief Refractive i ndex determination is generally considered one of the most powerful techniques performed by PLM. By aligning polarized light with the crystallographic axes of a sample, th e relative re fractive index of the sample can be determined by comparison to the surr ounding mounting medium. Though the measurement of the principle refractive indices occurs in PPL, understanding the crystal morphology and orienting the principle refra ct ive indices r equires a combination of PLM techniques, including extinction characteri stics in XPL and conoscopic observation of the crystal sample. As such, the discussion of these optical characteristics is paramount prior to further explanation of refr ac tometry. More information on (a) ( b ) 18 the use of refractometry for the determination of princi ple refractive indices is provided in S ection 2.1.4. 2.1.2 Observations in Crossed - Polarized Light While light propagating through a PLM sample is subjected to polari za tion due to t he polarizer, a second polarizing film, referred to as the analyzer, can be positioned above the objective lens of the microscope to provide an additional view of light optics through the sample. The analyzer is commonly oriented perpendicul ar to the lower polarizer in the North - South (N - S) position, and as such w hen the analyzer is inserted in the light path the sample is said to be viewed in XPL. Due to the perpendicular orientation of the two polarizers, no light passes through the system an d a darkfield view is observed. 10 In order to observe samples in XPL, interference of the incident light must occur as it passes through the sample. 9,10 As such, isotropic samples which do not split the incident light appear black, or extinct, in XPL. An isotropic sam ples, however, split the incident light beam into two compo nents due to the multiple refractive indices present in the light path. These two - component light rays destructively interfere with one another and produce polarization colors in XPL , more commonly referred to as retardation colors, when the analyzer recom bines components travelling in the same direction and vibrational plane ( Figure 2.3 ). 10 19 Figure 2.3 Diagram illustrating the interaction of light with isotropic media (left) and an iso tropic media (right) The two components of light noted as the ordina ry and extraordinary ray in Figure 2.3 travel orthogonally to one another at different speeds through the sample based on the refractive indices. The difference between the speeds at which the or dinary and extraordinary rays of light is described as bire fringence. Birefringence can be measured quantitatively as the Additionally, birefringence can be observed qualitatively by utilizing a Michel - L é vy birefringence in terference color chart ( Figure 2.4 ). Additionally, the retardation colors of a 20 sample can be characterized by comparing the order of red to those displayed in the Michel - L é vy chart. Fi gure 2.4 Mi chel - L é vy Birefringence Chart 11 Retardation increases linearly with both the thickness and birefringence; thus, the Michel - Lévy chart utilizes both thickness and retardation to determine birefringence of a sample. 10 From Figure 2.4 , retard atio n colors ar e a series of Newtonian colors, which repeat i marked by a red color. Though only showing Newtonian colors through the sixth order, each new order of retardation color becomes less vibrant until approximately the 10 th orde r of red, which 9 Furthermore, the birefringence of a sample can be categorized into three general categories : low (0 0.010), moderate (0.010 0.050), and high (>0.050). 10 An additional characteristic observed for anisotropic samples in XPL is extinction. Retardation col ors vary in intensity as the sample is rotated relative to the polarizers. This change 21 in intensity varies cyclically from zero intensity (extinction) to maximum brightness at 45° from extinction and back to ze ro intensity at 90° rotation. At either extinc tion point, one principle refractive index is aligned with the polarizers, indicating that only one vibrational direction is passing through the polarizer. This phenomenon occurs regardless of sa mple rotation f or isotropic samples, as the direction of ligh t exiting the sample is unchanged at any orientation relative to the incident light. The extinction of anisotropic samples can be categorized as complete (zero intensity occurs), incomplete (zero int ensity does not occur regardless of sample rotation), and oblique (zero intensity occurs at an angle not parallel to the boundaries of a well - formed crystalline sample). Incomplete extinction most often occurs with samples that are comprised of stacked cry stal struct ures, due to the inconsistent orientation of p rinciple refractive indices of each crystal. 2.1.3 Conoscopy and Interference Figures When viewing a sample in PPL or XPL, an orthoscopic image is produced due to the sample being viewed perpe ndic ular to the path of incident light. An additional techniq ue utilized during PLM analysis is conoscopic characterization of samples, during which conoscopic light is used to analyze the sample, rather than orthoscopic. Under conoscopic conditions, the l ight coming fro m the sample is viewed at multiple difference angles simultaneously in a cone of converging light. Unlike orthoscopic analysis, which transmits the image to the focal plan though the oculars, conoscopic images manifest on the back focal plan e of the object ive lens. Due to this, the use of a Bertrand l ens is necessary in order to visualize and interpret the interference figures formed by anisotropic samples. 22 In conoscopic view, the converging cone of light from the sample travels in various d irec tions, reta ining the two - component, orthogonal light wave s that originate when polarized light interacts with samples of more than one refractive index. Conoscopic light contains information regarding speed, retardation, and vibration direction from a give n sample. F igure 2.5 provides an example of an interference figure produced by a uniaxial (having two refractive indices) sample. Some features to note include the circles of retardation colors emerging from the center of the interference figure, refer red to as isoch romes. Ea ch circle represents light with the same retardation and birefringence. The dark cross - shape is made up of isogyres, indicating where there is extinction in the sample due to the parallel polarizers and wave vibrations of the sample . Th e wedge - sha ped, smal ler end of the uniaxial isogyre is referred to as the homodrome, and points to the position of the optic axis. The melatope of the interference figure is found where the two isogyres intersect and indicates the position of the optic axi s of the sa mple. 12 23 Figure 2.5 Example of a uniaxial interference figure with key features labeled When rotating the microscope stage, the positioning of the interference figure changes. This is due to the changing shape and symmetry of the isoch romes and isogy res with the changing orientation of the sample. As the stage rotates, the sample orientation moves resulting in changes to the travel time of light rays, retardation, and vibration direction. Figure 2.6 displays how a uniaxial interference figur e may chan ge with stage rotation. An exception to this is when the sample is viewed perpendicular to its optic axis, as shown in Figure 2.5 . 24 Figure 2.6 Conoscopic images of uniaxial interference figure orientation changes with microscope stage rotati on 12 Biaxial (having three principle refractive indices) samples have distinctly different interference fig ures when viewed in conoscopic light. Figure 2.7 provides an example of a biaxial interference figure in a similar optic axis orientation as Figure 2.5 . Similar t o uniaxial interference figures, colorful isochromes emerge from the center of the figure while the black isogyre indicates conoscopic positions of extinction. The melatope can be found at the thinnest point of the isogyre and marks the opti c ax i s for this sample. 25 Figure 2.7 Example of a biaxial interference figure Biaxial interference figures differ in characteristic shape from uniaxial interference figures due to the number and positioning of the optic axes present. Uniaxial samples only have one optic axis, whereas biaxial samples have two at varying positions fro m one another. When one optic axis is centered and viewed conoscopically, the other optic axis is not in the plane of view. The angle between the two optic axes (referred t o as 2 V angle) can be estimated using the curve of the isogyre. Figure 2.8 provides a diagram for comparison. 26 Figure 2.8 Estimation of biaxial 2V angle based on isogyre curvature in an optic axis interference figure 10 When compared to Figure 2.8 , the 2V ang le of th e sample shown in Figure 2.7 is between 70 - 80°. It should be noted th at the curvature of the isogyres decreases until the isogyre forms nearly a straight line at 90°. A 2V angle of 0° indicates a uniaxial sample; as such, it is often difficu lt to d ifferent iate biaxial and uniaxial samples when the 2V angle is below 15°. Whi le estimating the 2V angle using the curvature of isogyres in an optic axis orientation is a common qualitative method, the exact 2V angle can be calculated using an equati on foll owing th e determination of the three principle refractive indices and the opt ic sign biaxial sample. One final characteristic observed in conoscopic light is the optic sign of an anisotropic sample. Anisotropic samples (both uniaxial and biaxial) c an be c ategoriz ed as optically positive (+) or negative ( - ). Optic sign is related t o the refractive indices of anisotropic samples. Uniaxial optic sign is determined by the direction of the principle refractive indices, while the position of the optic axe s relat ive to t he principle refractive indices determines the optic sign of a biaxia l sample. Due to the close relation of optic sign to refractive index, this characteristic is further discussed in S ection 2.1.4 . 27 2.1.4 Determining the Principle Refractiv e Indic es using Biaxial Refractometry Understanding the orientation of crystallographic axes while observing interference figures is necessary to fully characterize a sample. The conoscopic interference figure provides a cal ind icatrix (which does not exist in reality), which is used to predict crystal orientation and understand how refractive indices vary with the direction of a sample. To better illustrate this relationship, an example of a uniaxial indicatrix is provide d in Fi gure 2.9 . For uniaxial indicatri ces, the primary axes are labeled X, Y, and Z, and coincide with the crystallographic axes a, b, and c and the refractive indices labeled by the letter n with the corresponding direction subscripted. Further, the opti c axis coincide s with the c crystallogr aphic axis. 28 Figure 2.9 Uniaxial indicatrix with principle refractive index views (a) optic normal and (b) optic axis The uniaxial indicatrix provides a simplified visualization of crystallographic axes and refra ctive in dex as they coincide with the p rimary axes of a crystalline sample. As shown, uniaxial samples have one optic axis, always parallel to the c crystallographic axis, and two refractive indices. The refractive indices are referred to as ordinary (n ) and extraordinary (n ). The ordinary an d extraordinary refractive indices split the incident parallel light into two perpendicular rays, traveling at different velocities through the sample . The rays are represented by vectors with both direction and ma gnitude, such that the length of the refra ctive index vectors in the indicatrix relate to the velocities of light traveling through the sample and thus is proportional to the refractive index. As such, in Figure 2.9 the velocity of light traveling parallel to Z (the long vector) is greater than th e velocity of light parallel to either Y or X (the short vector). The relationship of velocities can also be expressed as the relationship between 29 refractive indices as n > n . Additionally, the optic axis (denoted as O.A. in Figure 2.10 ) is perpendicular to the n refractive index vector, thus light propagates at the same velocity in all direction (as shown by Figure 2. 9 b). Biaxial indicatrices are more complicated than the previous uniaxial example. This is due to the fact that biaxial samples have thre e different refractive indices and may contained angled crystallographic axes relative to the primary axes. Figure 2. 10 provides an example of a simplified biaxial indica trix with no angled crystallographic axes. Figure 2. 10 Biaxial indicatrix with principle refractive index views (a) optic normal and (b) optic axis, and (c) obtuse bisectrix (Bxo) Unlike uniaxial samples, where the OA coincides with the primary Z axis and C crystallographic axis, the OAs of biaxial samples d o not lie on any primary axis within the crystalline sample. By definition, the OAs are perpendicular to the region of the indicatrix in which light propagates at the same velocity in all directions . In the case of biaxial samples, the OAs are perpendicula r to the n vector direction and lie between n and n . 30 As discussed in Section 2.1.3 , the angle between the two OAs is defined as the 2V angle. Figure 2. 10 displays this characteristic as it relates to the biaxial indicatrix. The relationship between OA location a nd n orientation between n and n gives rise to the optic sign of a biaxial sample, as introduced in Section 2.1.3 . Since n can lie anywhere bet ween n and n , it may lie closer to n or n in a particular biaxial sample. If n is closer to n , the opti c axes are closer to n and the angle between them is bisected by n . In this case, the OA forms an acute angle with n and n is designated as the acute bisectrix (Bxa). Oppositely, n forms an obtuse angle with the OA and is designated as the obtuse bise ctrix (Bxo). In this example, when n is closer to n , the sample is said to be optically positive (+). Figure 2.1 1 A provides an example of a (+) biaxial indicatrix. If, however, n lies closer to n , the OAs are at an acute angle with n and an obtuse an gle with n indicating an optically negative ( - ) sample ( Figure 2.1 1 B ). 31 Figure 2. 11 O ptically positive (A) and optically negative (B) biaxial indicatrices To measure the principle refractive indices of both uniaxial and biaxial samples, observation o f the sample in both orthoscopic and conoscopic light is performed. Conoscopic observation provides a method to orient the cry stallographic axes and refractive indices, while orthoscopic view is used to perform relative refractive index measurements. As su c h, it is necessary to understand the relationship between conoscopic figures and their associated indicatrix to correctly ori ent the refractive indices of anisotropic samples. For the purposes of this work, only the characteristic measuring refractive ind i ces for biaxial samples will be discussed. As discussed in Section 2.1.3 , the interaction of conoscopic light with a biaxia l sample produces characteristic interference figures that can be used in conjunction with the aforementioned biaxial indicatrix t o orient the refractive indices. While the specific methods used to select and orient crystal samples will be discussed in Sec tion 2.2.5 , Table 2. 2 provides a summary of the common interference figures and the associated principle refractive indices. It 32 sh o uld be noted that n can be determined from most interference figures, so long as the optic axial plane can be identified. Ta ble 2.2 Common interference figures for biaxial refractometry Interference Figure Refract ive Index Determined Optic Axis n only Acute Bisectrix (Bxa) & Obtuse Bisectrix (Bxo) n (always) & n or n based on optic sign Pendulum Figures n Optic Normal n and n Following the determination of crystal orientation via refractive index can be char a cterized in PPL using the Becke line method. The Becke line refers to the small halo of bright light that moves at the boundary of a transparent sample when the microscope focus is changed 10 . The formation of the Becke line around transparent samples is d u e to the refraction of light as it interacts with media of different refractive indices. Figure 2.1 2 provides a visual of the Becke line for a sample. 33 Figure 2.12 (A) Becke line moving out into higher refractive index medium and (B) into a sample crys ta l with higher refractive index It should be noted that the Becke line always moves into the medium with a higher refractive index when the microscope is focused above the sample, and as such can be used to compare the relative refractive index of a sam pl e crystal to the refractive inde x of the surrounding medium (most commonly a Cargille refractive index liquid). For example, Figure 2.1 2 A displays a crystal sample that is higher in refractive index than the surrounding medium while Figure 2.1 2 B shows a crystal sample that is lower in refractive index that the surrounding medium. A B 34 2.2 MATERIALS AND METHODS 2.2.1 Samples Five case samples of marijuana extract (categorized as BHO) submitted to the Kalamazoo th roughout this work. These case s amples ranged in consistency, crystal size distribution, and age, which was defined as the time elapsed since the sample was submitted to the laboratory. To identify each sample, the laboratory assigned case number was use d. An additional four samples wer e obtained for further characterization. These samples included two BHO samples (crumble and crystals) procured from a local dispensary (Skymint, Lansing, MI) as well as two hemp derived samples purchased from a dispensary ( Cannabidiol Life, Sanford, Fl) t hat utilizes hemp produced in Colorado. The Skymint dispensary samples were extracted from marijuana using hydrocarbon solvents similar to the KCSD case samples. The Cannabidiol Life samples were extracted from hemp using su percritical CO 2 . Table 2.3 summarizes the source, identity, and age of samples in this work. 35 Table 2.3 Sample identifications, sources, and year obtained Sample Identity Source Year Obtained KDPS 18 - 9026 KCSD 2018 PPO 14 - 20332 - 10 KCSD 2014 KCSD 1 4 - 10811 - 28964 KCSD 2014 KCSD 14 - 10811 - 28967 KCSD 2014 KCSD 14 - 10811 - 28960 KCSD 2014 Sugar/Wax Skymint 2020 crystals Skymint 2020 CBD Shatter Crystals Cannabidiol Life 2020 CBD Crumble/Wax Cannabidiol Life 20 20 2.2.2 Sample Preparation Techniques and Macroscopic Observations Less than 1 mg of each solvent extract sample was placed on a microscope slide and viewed through a stereomicroscope. Samples were first preliminarily screened for the presence of anis ot ropic crys talline structures followed by manual separation of the crystal and wax components. Using a fine tungsten needle, the desired crystal was manipulated out of the wax material, taking care to remove as much of the wax as possible without cleaving t he crystal . Crystals were selected for microscopic analysis within a desired range of sizes (approximately 50 100 µm) as measured using an optical micrometer. Successfully separated crystals were then moved to a separate microscope slide and individual ly placed un der handmade, in - house, 3 mm 36 micro - cover glasses . Handmade micro - cover glasses were made by scoring standard microscope cover glasses ( VWR International, Radnor, PA ) with a diamond scribe. Cargille refractive index liquids (Cargille Laboratorie s, Cedar Gro ve, NJ) were applied dropwise to each crystal sample prior to characterization by PLM. Additionally, consistency and crystal size were noted for the physical, macroscopic description of each sample. Macroscopic photographs of the samples can be f ound in th e appendix ( Figures A 2. 1 - A 2. 9 ). 2.2 .3 Polarized Light Microscopy Each sample was characterized using generally accepted PLM and optical crystallography techniques on an Olympus BX10 PLM (Olympus Corporations of America, Center Valley, PA). Th e following characteristics were determined under PPL: size distribution, color, morphology, pleochroism, and the three principle refractive indices. Size distribution was determined using a calibrated ocular micrometer, taking measurements of the crystall in e componen t in PPL. The three principle refractive indices were determined using biaxial refractometry ( Section 2.2.4 ). Additional characteristics determined using XPL included retardation colors, birefringence, and extinction characteristics. Retardatio n colors wer e compared to the Michel L é vy birefringence chart and categorized by orders of red. Birefringence was determined both quantitatively, following the determination of the three principle refractive indices, as well as qualitatively, using the cat eg ories disc ussed in Section 2.1.3 . Finally, in conoscopic light, the optic sign, 2V angles, and crystal system were determined for each sample. Optic sign was established using a centered optic axis interference figure, with the concave curve of the isogy re facing no rth east (NE), and noting constructive or 37 destructive isochrome interference upon the insertion of a red one waveplate (Olympus Corporations of America, Center Valley, PA). Constructive interference indicates an optically positive crystal, whil e destructiv e interference indicates an optically negative crystal. The 2V angle was estimated by comparing the curve of a well - centered optic axis figure with literature schematics derived from Wright ( Figure 2.7) . 10 The crystal system was deduced using i nf ormation related to the extinction characteristics, number of optic axes, and refractive indices of the sample crystal as discussed in Section 2.1 . 2.2.3 Refractive Index Determinations using Biaxial Refractometry For the determination of the three pr in ciple refractive indices, Cargille® refractive index liquids ranging from refractive index 1.4 1.8 were used (Cargille Laboratories). Biaxial refractometry was used to correctly identify and align the principle re fractive indices of the crystal to the po larizer. Additionally, principle refractive index measurements were determined in triplicate for each refractive index liquid used, and a sodium D filter (Orange21 ) (The Tiffen Company, Hauppauge, NY) was used for a ccurate interpretation of Becke line mo ve ment. Due to the hydrophobic nature of the Cargille refractive index liquids and the high lipid affinity of cannabis products, both the crystal and wax components of BHO dissolved in the Cargille liquids during PLM analysis. The rate at which the crysta ls dissolved was related to the relative strength of the refractive index liquid used. That is, in higher refractive index liquids ( i.e. , n D = 1.68), crystals dissolved more quickly (~ 30 s) than in lower refractive i ndex liquids (~ 15 min in n D = 1.50). W hi le the chemical composition of the Cargille refractive index liquids is proprietary, it is presumed that the liquids are lipid - based and hydrophobic. As such, it is assumed that the concentration of the proprietary hydrophobic mixture increases with refr ac tive 38 index, resulting in a faster rate of dissolution. Due to the different rates at which the crystals dissolved into the refractive index liquid, the optical characterization of the crystals was performed as quick ly as possible while still maintaining pr oper technique. The principle refractive index, n , was determined by the following method. In XPL, a crystal which showed the lowest retardation as the stage rotated was selected for analysis. This display of reta rdation colors in XPL indicates that the optic axis of the crystal is perpendicular (or nea rly so) to the stage ( Figure 2.9B ). In conoscopic view, the interference figure was characterized for uniaxial or biaxial indicatrix and the optic sign was determine d using a red one waveplate (Olympus Corporations of America). Once the orientation for n was confirmed, the microscope was returned to orthoscopic illumination in PPL and the Becke line method was used to compare the refractive index of the crystal to th at of the immersion liquid. To perform the Becke line method, the stage was lowered (raisi ng the objectives), while observing the crystal, and the direction of the Becke line was noted the Becke line always moves into the medium of higher refraction whe n the stage is lowered. This process was repeated, following the movement of the Becke line , until a range of refractive indices was determined. Additionally, pendulum figures were also used to orient n for a given crystal grain. In XPL, a crystal that ex hibited moderate retardation colors (between the highest and lowest observed orders of reta rdation) was selected for analysis. In conoscopic view, n was positioned parallel to the lower polarizer by orienting the pendulum arm of the isogyres in the N - S di rection. To determine n and/or n , with crossed polarizers, crystals with the highest reta rdation colors were selected. Selecting a crystal with the highest possible birefringence increased the likelihood of observing a flash interference figure. This int erference figure occurs when the crystal is oriented with the optic normal perpendicular to the stage and the optic plane horizontal 39 to the stage. In XPL, the crystal vibration direction was determined used a red one compensator in order to orient the fast direction parallel to the lower polarizer (this would orient n for measurement). To orient n , the stage was rotated 90° such that the slow direction of the crystal was parallel to the lower polarizer. Due to the high order retardation colors exhibited b y the crystals in XPL, determining the fast and slow ray direct ions proved difficult using a red one compensator. Additionally, these refractive indices were measured from a Bxa/Bxo figure, however the flash figure method was most commonly used throughout this work. 2.3 RESULTS OF OPTICAL CHARACTERIZATION BY POLARIZE D LIGHT MICROSCOPY Older case samples previously analyzed by KCSD were analyzed via PLM and the optical characteristics of the crystalline components were compared ( Table 2.4 ). Table 2.4 KCSD BHO case sample wax consistency and crystalline component summaries Sample Identity Sample Consistency Crystal Size Range (µm) KDPS 18 - 9026 Wax (Crumble) 10 200 PPO 14 - 20332 - 10 Wax ( Soft Solid/Hard ) 10 150 KCSD 14 - 10811 - 28964 Wax (Soft , Viscous) 10 200 KCSD 14 - 10811 - 28967 Wax/Shatter (Solid, H ard; Glass - like) 10 100 KCSD 14 - 10811 - 28960 Wax (Soft to Glass - like) 10 75 40 Though each sample was fully optically characterized in this work, case sample KDPS 18 - 9026 was used for initial refra ctive index determinations due to the ease of crystal isolation. Case sample KDPS 18 - 9026 was a BHO sample submitted for forensic analysi s in 2018, though the date of manufacture is unknown. Macroscopically, this sample consisted of large, yellow orange, anhedral material ( Figure 2.1 3 ). Microscopically, the well - formed crystals were differentiated from the wax component based on color and a nisotropy. While the wax component of BHO is typically brown - orange in color, the embedded crystals are colorless ( Fig ure 2.14 ), which is readily observed using a standard stereomicroscope or compound microscope. Figure 2.1 4 provides a PPL photomicrograph, showing the difference in color among the particles with colorless crystals (blue circle), dark brown - orange wax (red circle), and light - yellow colored particles where the wax material is lightly stuck to the crystal (green circle). Figure 2.13 Macros copic view of case sample KDPS 18 - 9026 in glass vial 41 Figu re 2.14 Photomicrograph of KDPS 18 - 9026 at 100X magnific at ion in PPL Using a PLM in XPL configuration, the distinction between the wax and crystal component becomes even more visible due to th e very high order retardation colors exhibited by the crystals, while the wax component does not exhibit retardation c ol ors due to its isotropy. Figure 2.1 5 provides a XPL photomicrograph, showing the difference in retardation colors amongst particles with high order white shown by the crystalline material (blue circle) and dark brown - black exhibited by the wax material ( re d circle). After confirming the presence of crystals using PLM in crossed polarized configuration, crystals were manually separated from the wax and mounted in Cargille refractive index liquids for further optical characterization. 42 Figure 2 .15 Photomi cro graph of KDPS 18 - 9026 at 100X magnification in XPL In PPL, crystals from case sample KDPS 18 - 9026 were colorless and displayed no pleochroism. Crystal morphology ranged from thin and platy to anhedral as crystal size increased. Upon the rotation of the mi croscope stage, the crystal relief was observed to range significantly ( Figure 2 .1 6A and B ) indicating that the differ ence between the principle refractive indices is extreme. Figure 2.1 6 KCSD 18 - 9026 with polarizer at (A) 0° orientation and (B) 90° orie ntation to demonstrate changes in relief A B 43 The maximum retardation colors observed for the sample crystals were high order white, even in moderately sized crystals. High order white retardation was exclusively observed in the majority of samples selected fo r optical analysis (~100 µm in size). This indicates retardation of more than six orders of red when compared to the Michel - Lévy chart. In XPL, birefringence was estimated using the Michel - Lévy chart which indicated high order white retardation with sam ple crystals ~50 µm . Further ob servations made in XPL included extinction characteristics. Though the anhedral crystals did not provide a cleavage face by which to measure ex act angles of extinction, noting the degree of rotation on the microscope stage at each exhibited extinction allo wed for approximate measure of extinction angles. Upon rotation of the stage in XPL, crystal grains exhibited highest order retardation at appr oxi mately 45° of stage rotation, and showed no of rotation. Furthermore, the extinction was complete for each crystal, regardless of size or morphology. Biaxial refractometry allowed for co nos copic characterization of the sample as well the determination of the principle refractive indices in PPL. Dur ing biaxal refractometry procedures, analysis of the crystals in conoscopic view indicated that the crystal was biaxial (containing three princ ipl e refractive indices). Biaxial crystals exhibit optic axis interference figures with one curved isogyre that l ook like cat eyes ( Figure 2.17 ). The optic sign was determined to be negative following insertion of a red one waveplate, and destructive inter fer ence in regard to isochrome color was observed. Additionally, the 2V angle, or angle between the two optic axes, was determined to be ~70° when comparing the curvature of the isogyre to literature values 8 . 44 Figure 2.17 Conoscopic view of a KDPS 18 - 9026 cry stal displaying an optic axis interference figure Using biaxial refractometry techniques, the thre e principle refractive indices (n , n , and n ) for KDPS 18 - 9026 were determined. The process of determining each refractive index range was performed in triplicate at each new refractive index. Additionally, new crystals were selected from the bulk BHO ma terial to ensure that the results of the Becke line test were reproducible. As such, to complete the crystal characterization and refractive index determination, over 150 crystals were manually separated from the bulk BHO material, analyzed for characteris tics in PPL and XPL, and subjecte d to biaxial refractometry. T he characterization of the principle refractive indices for sample KCSD 18 - 9026 was reported as ranges, rather than an exact match point. This is a more common approach in optical crystallograph y and was especially necessary du e to any possible change in refractive index from dissolution. Within the range provided, the contrast difference between the crystal and refractive index liquid was extremely low, indicating that the exact match point for the refractive index was very clo se to the current refractive index being measured. The ranges determined for KDPS 18 - 9026 are tabulated in Table 2. 5 . 45 Due to the time - consuming nature of the biaxial refractometry procedure for case sample KDPS 18 - 9026, r elative rather than absolute rang es , refractive indices were determined for crystals of the remaining case samples . Relative refractive index measurements were made in triplicate for each principle refractive index. With that being said, full characterizat ion in PPL, XPL, and conoscopic v iew was completed for each case sample in similar manner to KCSD 18 - 9026. In each case, the crystals were selected from the bulk material, mounted in the refractive index liquid, characterized in PPL and XPL, and the Becke line test was performed following orientation using biaxial refractometry. Though these relative refractive indices reported are not given as ranges, the contrast between the crystal and the respective refractive index liquid was very low, indicating that the principle refractive index wa s very close to the liquid being used for relative determinations. Table 2. 5 summarizes the complete characterization of the five case samples. 46 Table 2 . 5 Summary of Case Sample Optical Properties by PLM Sample KDPS 18 - 9026 PPO 14 - 20332 - 10 KCSD 14 - 10 811 - 28964 KCSD 14 - 10811 - 28967 KCSD 14 - 10811 - 28960 Morphology Anhedral chunks Thin/platy - Anhedral chunks Anhedral chunks Anhedral chunks Thin/platy - Anhedral chunks Size (µm) 10 200 10 150 10 200 10 100 10 75 Color Colorless Colo rless Colorless Colorless Colorless Refractive 1.4920 1.5040 ± 0.0002 Refractive 1.6320 - 1.6330 ± 0.0002 Refractive 1.6850 - 1.6900 ± 0.0002 Birefringen ce High High High High High Extinction characteristics Complete; ~90 Complete; ~90 Complete; ~90 Complete; ~90 Complete; ~90 Optic Sign Biaxial ( - ) Biaxial ( - ) Biaxial ( - ) Biaxial ( - ) Biaxial ( - ) 2V angle ~70° ~70° ~70° ~70° ~70° As summarize d in Table 2.5 , the optical characteristics for each case sample display similarities between crystals in BHO samples irrespective of age or consistency (wax - like versus glass - like). Crystal sizes in each case sample varied widely, indicating that sample c onsistency does not appear to affect the size or presence of crystalline material. It should be noted that although crystals between 50 100 µm were preferentially ana lyzed, a crystal of any size would provide similar optical properties and can be used f or characterization. Properties in PPL, XPL, and conoscopic view were consistent amongst all case samples. Photomicrographs of each case sample in PPL and XPL are provid ed for comparison in Appendix II. 47 Additionally, refractive index measurements were unif orm with the exception of principle refractive index n in the three KCSD 14 - 10811 case samples which was slightly higher than the originally determined range from the a nalysis of case sample KDPS18 - 9026. The Becke line test for the three case samples for which n was slightly higher than the determined range proved difficult due to extremely low contrast between the crystal and refractive index liquid. This very low cont rast indicates that although the refractive index may be greater than the established range, it is very close to n D = 1.6900. Similarly, the relative refractive index measurements for the additional four case samples were all performed in Cargille refracti ve index liquids in which the contrast was very low, indicating that the refractive index of the crystal was very close to that of the liquid . 2.3.2 Skymint Dispensary BHO Samples Samples were purchased from Skymint, a local Lansing, MI dispensary, for an alysis and comparison to the aforementioned samples from KCSD. This subset of crystals was selected to provide characterization of industry - p roduced, BHO samples as a comparison to the KCSD case samples of unknown origin. Additionally, these samples provid e the ability to observe crystal habits in new, relatively young samples, compared to the aged KCSD case samples. Table 2. 6 provides a summar y of the macroscopic sample consistency and microscopic crystal size range. Table 2.6 Skymint BHO macroscopic samp le summary Sample Identity Sample Consistency Crystal Size Range (µm) Sugar/Wax Crystalline/ Wax (oil/wax material) 20 150 THCA crystals Crystalline Material (No wax) 10 150 48 Each dispensary sample was fully characterized using analogous methods to the KCSD case samples. Macroscopically, these samples differed from the KCSD BHO case samples by containing fa sugar/wax sample (referred - brown color and consisted of both crystal and wax material ( Figure 2.1 8A rystal - white color in reflected ligh t and consisted of only agglomerated crystalline material ( Figure 2.1 8B ). Figure 2.18 Macroscopic view of PB THCA wax ( A ) and WB THCA crystals ( B ) i n dispensary glass containers Microscopically, the well - formed crystals in PB THCA wax were readily differentiated from the oily, wax component based on anisotropy. The agglomerated crystals for both samples were colorless, which is readily observed us i ng a standard stereomicroscope or compound microscope ( Figure 2. 19 A and B ). The wax component in the PB THCA wax consisted of an oily, liquid residue that did not readily cling to the crystal component ( Figure 2. 20 ). Using a 49 stereomicroscope, the dispens a ry sample crystals were separated from each agglomerate in order to analyze single cr ystals for analysis. Figure 2.19 Stereoscopic view of PB THCA wax (A) and WB THCA crystals (B) showing the crystal color in reflected light and agglomerated form Fig ur e 2.20 PB THCA wax crystals shown in PPL, note the oily, wax component separ ate from the crystalline component 50 In PPL, crystals from each dispensary sample were colorless and displayed no pleochroism. Crystal morphology ranged from thin and platy (~25 µ m in size) to anhedral as crystal size increased. Crystal relief changed dras tically for each sample when the microscope stage was rotated (represented by PB THCA wax crystals in Figure 2.21 A and B ) indicating that the difference between the principle re f ractive indices is extreme. Figure 2.21 PB THCA wax crystals with polarizer at (A) 0° orientation and (B) 90° orientation to demonstrate changes in relief In XPL, birefringence was estimated using the Michel - Lévy chart. The maximum retardation colors ob gh order white, even in moderately sized crystals. Smaller crystals (~25 µm or less in size) with thin, platy morphology displayed retardation in the 2 nd 3 rd order of colors when compared to the Michel - Lévy char t ( Figure 2.22 A ). High order white retard ation was exclusively observed in the majority of samples selected for optical analysis (those ~50 - 100 µm in size) ( Figure 2.22 B ) . This indicates retardation of more than six orders of red when compared to the M ic hel - Lévy chart and corresponds to high bi refringence. 51 Figure 2.22 Small PB THCA wax crystals in XPL displaying moderate retardation (A) and large crystals showing high order white retardation (B) Further observations made in XPL included extinction cha racteristics. Each dispensary sa mple contained crystals with anhedral morphology, thus no straight cleavage edge by which to measure exact angles of extinction. As such, only approximate extinction angles were noted when rotating the stage by marking th e d egree of rotation at each exhibi ted extinction point for the sample. Upon rotation of the stage in XPL, crystal grains exhibited highest order retardation ap pro ximately every 90° of rotation. Furthermore, the extinction was complete for each crystal, regardless of size or morphology. Biaxial refractometry allowed for conoscopic characterization of the sample as well the determination of the principle refracti ve indices in PPL. During biaxal re fractometry procedures, analysis of the crystals in conoscopic view indicated that the crystals from both dispensary samples were biaxial ( Figure 2.23 ). The optic sign for both dispensary samples was determined to be nega tiv e following insertion of a red one waveplate, and destructive interference in regard 52 to isochrome color was observed. Additionally, the 2V angle, or angle between the two optic axes, was determined to be ~70° when comparing the curvature of the isogyre to literature values 8 . Figur e 2.23 Conoscopic view of a PB THCA wax crystal displaying an optic axis interference figure Using biaxial refractometry techniques, the three principle refractive indices (n , n , and n ) for each dispensary sample were dete rmin ed. The process of determi ning each refractive index range was performed in triplicate at each new refractive index. Complete refractive index ranges were determined for each dispensary sample, rather than match points, due to the rapid dissolution of samp le crystals into the refra ctive index liquids. Within the range provided, however, the contrast difference between the crystal and refractive index liquid was extremely low, indicating that the exact match point for the refractive index was very close to t he current refractive inde x being measured. The ranges determined for both the PB THCA wax and WB THCA crystal samples are tabulated in Table 2. 7 . 53 Table 2.7 Summary of optical characteristics for Skymint dispensary samples Sample WB THCA crystals PB THC A wax Morphology Thin/platy - Anhedral chunks (agglomerated) Thin/platy - Anhedral chunks (agglomerated) Size (µm) 10 150 20 150 Color Colorless Colorless 1.4920 1.5000 ± 0.0002 1.4920 1.5040 ± 0.0002 1.6300 1.6320 ± 0.0002 1.6300 1.6320 ± 0.0002 Birefringence High High Extinction characteristics Complete; ~90 Comp lete; ~90 Optic Sign Biaxial ( - ) Biaxial ( - ) 2V angle ~70° ~70° 2.3.3 Cannabidiol Life Dispensary Samples Hemp - derived solvent extract samples were purchased from Cannabidiol Life, an online - based Florida dispensary, which sources its hemp products from Colorado. This subset of crystals was selected to provide characterizat ion of industry - produced hemp derived samples as a comparison to the BHO samples from KCSD and Skymint dispensary. Additionally, these samples provide the ability to observe crysta l habits of an assumed different cannabinoid, as the samples were derived fr om hemp rather than marijuana. It is the aim of this subset to provide a comprehensive characterization of such crystals in order to propose PLM characterization as a method throug h which extracts can be potentially screened prior to chemical analysis. Tab le 2. 8 provides a summary of the macroscopic sample consistency and microscopic crystal size range. 54 Table 2.8 Cannabidiol Life CBD sample summary Sample Identity Sample Consisten cy Crystal Size Range (µm) CBD Crumble/Wax Crystalline Material Oily Wax 130 - 800 CBD Shatter Crystals Crystalline Material 100 - 1200 Each dispensary sample was fully characterized using analogous methods to the aforementioned KCSD case samples an d Skymint dispensary samples. Macroscopically, these dark orange - light yellow and o ff white in color. Additionall y, the CBD wax sample consisted of both crystal and wax material ( Figure 2. 24A ), while the CBD crystal sample consisted of only well - formed crystalline material ( Figure 2. 24B ). 55 Figure 2.24 Macroscopic view of CBD wax ( A ) a n d CBD crystal ( B ) samples in dispensary containers Microscopically, the well - formed crystals in CBD wax were readily differentiated from the oily, wax component based on anisotropy. The well - formed crystals for both samples were colorless, which is re a dily observed using a standard stereomicroscope or compound microscope ( Figure 2. 25 A and B ). The wax component in the CBD wax sample consisted of an oily, liquid residue that did not readily cling to the crystal component ( Figure 2. 26 ). Using a stereomic r oscope, the dispensary sample crystals were separated from each agglomerate in order to analyze single crystals for analysis. Figure 2.25 Stereoscopic view of CBD wax (A) and CBD crystals (B) showing the crystal color in reflected light and the differe nc e in wax pres ence. 56 Figure 2.26 CBD wax crystals shown in PPL, note the oily, wax component separate from the crystalline component In PPL, crystals from each CBD dispensary sample were colorless and displayed no pleochroism. Both the CBD wax and CBD cry stal samples contained very well - formed crystals that had rod - shaped morphology and slightly angled edges. Crystal size ranged drastically in each sample from ~100 µm to over 1000 µm , each with similar morphologies in the well - formed crystals. Additiona lly , the sample crystals regularly broke apart in irregular patterns, leading to anhedral crystals in the gross sample, as displayed by Figure 2.26 . Crystal relief changed for each sample when the microscope stage was rotated, though not as drastically. Th e d ifferences i n relief is represented by CBD wax crystals in Figure 2.27 A and B and indicates a moderate difference between the principle refractive indices. 57 Figure 2.27 CBD wax crystals with polarizer at (A) 0° orientation and (B) 90° orientation to demo nstrate changes in relief. 58 In XPL, birefringence was estimated using the Michel - Lévy chart. The maximum order white for very large crystals, but moderate retardation (2 nd 3 rd order re d) i n smaller, cleaved crystals ( Figure 2.2 8 ). The high order white retardation in large crystals and moderate retardation in smaller crystals indicates a moderate - high overa ll birefringence of the crystalline component of both CBD dispensary samples. Fi gure 2.28 CBD wax crystalline component in viewed in XPL Further observations made in XPL included extinction characteristics. Each dispensary sample contained crystals with well - formed, rod - shaped morphology with straight cleavage edges by which to measu re ex act angles of extinction. When the crystal edge was parallel to the vertical crosshair of the ocular micrometer, the crystal grain remained visible in XPL ( Figure 2.29 ). This indicated a degree of inclined extinction, which was then measured using the stra ight edge of the crystal, noting the degree of rotation from the straight edge necessary to achieve complete extinction. Crystal extinction occurred regularly at every 9 of stage rotation but was inclined relative to the crystal edge at 44 - 59 Figure 2.29 XPL view of CBD Wax crystal with straight edge parallel to the crosshair of the ocular micrometer (red) During biaxal refractometry procedures, analysis of the cr y stals in conoscopic view indicated that the crystals from both CBD dispensary samples were biaxial ( Figure 2.30 ). The optic sign for both dispensary samples was determined to be positive following insertion of a red one waveplate, and constructive interfe r ence in regard to isochrome color was observed. Additionally, the 2V angle, or angle betwee n the two optic axes, was determined to be ~80 - 90°, as the isochrome curvature was nearly straight. Further, using biaxial refractometry techniques, the ranges for t hree principle refractive indices (n , n , and n ) for each dispensary sample were determin ed ( Table 2.9 ). 60 Figure 2.30 Conoscopic view of the CBD crystal sample displaying a biaxial optic axis interference figure 61 Table 2.9 Summary of optical characte ris tics for Cannabidiol Life dispensary samples Sample CBD crystals CBD wax Morphology Rod - shaped to Anhedral chunks (fragments) Rod - shaped to Anhedral chunks (fragments) Size (µm) 100 - 1200 100 800 Color Colorless Colorless 1 .55 60 1.5600± 0.0002 1.5560 1.5600± 0.0002 1.600 1.6040 ± 0.0002 1.600 1.6040 ± 0.0002 1.6600 1.6700 ± 0.0002 1.6600 1.6700 ± 0.0002 Birefringence High (>0.05) High (>0.05) Extinction characteristi cs Inclined Inclined ~ Optic Sign Biaxial (+) Biaxial (+) 2V angle ~80 - 90° ~80 - 90° As summarized in Table 2.9 , the crystalline components within each CBD dispensary sample were analogous. Additionally, when per for ming biaxial refractometry to determine the principle refractive indices of the crystal samples, it was noted that the high 2V angle directly correlated to the possible numerical value for n . Given that the 2V angle relates to the angle between both op tic axes, and optic sign relates to the bisection of the optic axes by either n or n the assumption could be made by the given characteristics that n would be nearly equal to half the difference between n and n . Further, n would be closer in numerica l v alue to n given the optic sign. This is well represented by the optical results provided in Table 2.9 . 62 2.4 DISCUSSION AND COMPARISON OF CRYSTALLINE CHARACTERIZATION BY POLARIZED LIGHT MICROSCOPY The optical characterization of the KCSD BHO case sample s i ndicated similar classes of crystals both within individual case samples and among the entire subset of BHO samples. Similarly, commonalities in optical properties between the representative KCSD c ase sample (KDPS 18 - 9026) and Skymint dispensary sample s w ere observed by comparing the results summarized in Tables 2.5 and 2.7 . When comparing the optical characteristics of the analogous KCSD and Skymint crystals to that of the Cannabidiol Life, obviou s differences arose. These optical differences directly rel ated to differences in crystal structure and categorization of crystal system for each subset of samples. The optical characteristics for representative samples from each subset are provided by Tab le 2.10 . Table 2.10 Summary of optical characteristics fro m representative samples of each subset Sample KDPS 18 - 9026 WB THCA crystals CBD crystals Morphology Anhedral chunks Thin/platy - Anhedral chunks (agglomerated) Rod - shaped to Anhedral chunks (fragments) Size (µm) 10 200 10 150 100 - 1200 Color Col orless Colorless Colorless 1.4920 1.5040 ± 0.0002 1.4920 1.5000 ± 0.0002 1.5560 1.5600± 0.0002 1.6320 1.6330 ± 0.0002 1.6300 1.6320 ± 0.0002 1.600 1.6040 ± 0.0002 1.6850 1.69 00 ± 0.0002 1.6600 1.6700 ± 0.0002 Birefringence High High High (>0.05) Extinction characteristics Complete; ~90 Complete; ~90 Inclined Optic Sign Biaxial ( - ) Biaxial ( - ) Biaxial (+) 2V angle ~70° ~70° ~80 - 90° 63 B y u tilizing the optical properties recorded for both the KCSD case samples and Skymint dispensary samples the crystal system for this set of crystals can be determined. The characteristics observed in XPL provided insight related to the possible crystal sy ste m of the samples. First, the anisotropy of the crystals, highlighted by the high order retardation and high birefringence, indicated that the sample conta ined more than one refractive index. This refined the possible crystal system to those with unequal cr ystallographic axis lengths ( i.e. the sample, a nd as such can provide insight regarding crystal system. Given that the extinction for the crystals in each KCSD and Skymint sample occurred completely at approximately 90° of rotation, crystal systems could be refined to only those that have mutually perp endicular crystallographic axes ( i.e . ). The characteristics established through biaxial ref ractometry, including conoscopic characterization and refractive index determination, allowed for further crystal system refinement. In conoscopic view, a biaxial interference figure indicated that the sample contained two optic axes and three principle re fractive indices. Noting this, the possible crystal system was c, or triclinic. Taking the XPL characteristics into consideration, specifically extinction charact eri stics, the crystal system for the crystal material from both the KCSD case samples and Skymint dispensary samples was categorized as orthorhombic. Given t he presence of only one analogous crystal type in each case sample, similarities in chemical compos iti on can be inferred between the clandestine and dispensary produced BHO samples. Additionally, the crystal system for the crystalline component of the CBD dispensary samples can be determined in a similar fashion as the KCSD and Skymint samples above. T he 64 anisotropy observed in XPL highlighted by the moderate to high order retardation and high birefringence indicated that the samples contained more than one refractive index. This refined the possible crystal system to those with unequal crystallographic axi s lengths. Given that the extinction for the crystals in the dispensary samples occurred completely at approximately 90° of rotation, but that at least on e principle refractive index aligned with the polarizer at an inclined angle, the crystal systems for both the CBD wax and CBD crystal samples could be refined to only those that have one unequal crystallographic angle ( i.e . ; In conoscopic view, a biaxial interference figure indi cated that the sample contained two optic axes and three principle refractive indices. Noting this, the possible crystal system was identified as one with three unequal crystallograph monoclinic, or triclinic. Taking the XPL c haracteristics into consideration, specifically extinction characteristics, the crystal system for the CBD dispensary samples was categorized as monoclinic. The difference in crystal system ( i.e. optical properties) between the crystals from the KCSD and Skymint subsets and the Cannabidiol Life CBD subset allows for rapid differentiation of the crystals formed in extracts derived from marijuana versus hemp. The readily observed optica l differences between the marijuana - derived and hemp - derived solvent extr act samples provides the opportunity for rapid screening of samples by PLM in forensic laboratories based on their crystalline component. While these samples can be optically differen tiated, identification of the chemical composition using spectroscopic in strumentation is necessary to provide a comprehensive method by which to reliably distinguish the crystal component of marijuana and hemp extracts. The chemical characterization and i dentification of each subset of crystals is provided by Chapter 3 . 65 APPENDIX 66 Figure A2.1 Macroscopic (A) and microscopic views of KCSD case sample PPO 14 - 20332 - 10 in PPL (B) and XPL (C) Figure A2.2 Macroscopic (A) and microscopic views of KCSD case sample KCSD 14 - 10811 - 28964 in PPL (B) and XPL (C) Figure A2.3 Ma croscopic (A) and microscopic views of KCSD case sample KCSD 14 - 10811 - 2896 0 in PPL (B) and XPL (C) 67 Figure A2.4 Macroscopic (A) and microscopic views of KCSD case sample KCSD 14 - 10811 - 2896 7 in PPL (B) and XPL (C) Figure A2.5 Microscopic views of Skymint dispensary sample WB THCA crystals in PPL (A) and XPL (B) Figure A2. 6 Microscopic views of Cannabidiol Life dispensary sample CBD crystals in PPL (A) and XPL (B). 68 Table A2.1 Table of refractive index measurements (n ) for KDPS 18 - 9026 Crystal Number Re lative Refractive Index Crystal Number Relative Refractive Index 1 n < 1.520 20 n 2 n < 1.520 21 n 3 n < 1.520 22 n 4 n 23 n 5 n 24 n 6 n .510 25 n 7 n 26 n 8 n 27 n 9 n > 1.5080 28 n 10 n < 1.5080 29 n 11 n < 1.5080 30 n 12 n < 1.5080 31 n 13 n < 1.5080 32 n 14 n < 1.5080 33 n 1.5 000 15 n 34 n 16 n 35 n 17 n 36 n 18 n 37 n 19 n 38 n Table A2.2 Table of refractive index measurements (n ) for KDPS 18 - 9026 Crystal Number Relat ive Refractive Index Crystal Number R elative Refractive Index 1 n < 1.6500 15 n 1.6330 2 n < 1.6 5 0 0 16 n 3 n < 1.6 5 0 0 17 n 4 n < 1.6400 18 n 5 n < 1.640 0 19 n 6 n < 1.640 0 20 n 25 7 n 21 n 1.6325 8 n 22 n 1.6325 9 n 23 n 10 n 1.6340 24 n 1.6325 11 n 25 n 12 n < 1.6340 26 n 1.632 0 13 n 27 n 1.632 0 14 n > 1.6330 28 n 1.632 0 69 Table A2.3 Table of refr active index measurements (n ) for KD PS 18 - 9026 Crystal Number Relative Refractive Index Crystal Number Relative Refractive Index 1 n > 1.6400 20 n 2 n > 1.640 0 21 n 3 n > 1.640 0 22 n 1.6 844 4 n > 1.6 5 0 0 23 n 1.6 844 5 n > 1.6500 24 n 1.6 844 6 n > 1.650 0 25 n 1.6 850 7 n > 1.6 6 0 0 26 n 1.6 850 8 n > 1.6600 27 n 9 n > 1.6600 28 n 10 n > 1.6 7 0 0 29 n 11 n > 1.6700 30 n 12 n > 1.6700 31 n 90 0 13 n 1.6 8 0 0 32 n 1.6 90 0 14 n 1.6 8 0 0 33 n 1 .6 90 0 15 n > 1.6 8 0 0 34 n 1.69 0 0 16 n 1.6 833 35 n 0 0 17 n 1.6 833 36 n < 1. 700 0 18 n 1.6 833 37 n < 1. 700 0 19 n 1.6 840 38 n < 1. 700 0 Table A2. 4 Table of refractive index measurements for PP O 14 - 20332 - 10 Crystal Number Relative Refractive Index 1 n 1.4960 2 n 3 n 4 n 5 n 6 n 7 n 8 n 6900 9 n 6900 70 Table A2. 5 Table of refractive index measurements for KCSD 1 4 - 10811 - 28964 Crystal Number Relative Refractive Index 1 n 1.4960 2 n 3 n 4 n 5 n 6 n 7 n 1.6900 8 n 1. 6900 9 n 1. 6900 Table A2. 6 Table of refractive index measurements for KCSD 14 - 10811 - 2896 7 Crystal Number Relative R efractive Index 1 n 1.4960 2 n 3 n 4 n 5 n 6 n 7 n 1.6900 8 n 1. 6900 9 n 1. 6900 Table A2. 7 Table of refractive index measurements for KCSD 14 - 10 811 - 2896 0 Crystal Number Relative Ref ractive Index 1 n 1.4960 2 n 3 n 4 n 5 n 6 n 7 n 1.6900 8 n 1. 6900 9 n 1. 6900 71 Table A2. 8 Table of refractive index measurements for WB THCA Cry stals Crystal Number Relative Refrac tive Index Crystal Number Relative Refractive Index n 1 n n 20 n 6320 2 n 1.5040 21 n 6320 3 n 22 n 1. 6320 4 n 1.5040 23 n 1. 6320 5 n 24 n 6320 6 n 1.50 00 25 n 1. 6300 7 n 1.50 00 26 n 1. 6300 8 n 0 0 27 n 1. 6300 9 n 0 0 n 28 n > 1.6850 10 n 1.50 00 29 n > 1.6850 11 n 0 0 30 n > 1.6850 12 n 0 0 31 n > 1.6900 13 n 1. 4960 32 n > 1.6900 14 n 1.4960 33 n > 1.6900 15 n 1.4960 34 n 1. 7000 16 n 1.4960 35 n 17 n 36 n 18 n 19 n Table A2. 9 Table of refractive index measurements for PB THCA Wa x Crystal Number Relative Refractive Index Crystal Number Relative Refractive Index n 1 n n 13 n 6320 2 n 14 n 6320 3 n 15 n 1. 6320 4 n 1. 4960 16 n 1. 6320 5 n 1.4960 17 n 1. 6320 6 n 1.4960 18 n 1. 6300 7 n 1.4960 19 n 1. 6300 8 n 1. 4960 20 n 1. 6300 9 n 1. 4960 n 21 n > 1.6900 10 n 22 n > 1.6900 11 n 23 n > 1.6900 12 n 24 n 1. 7000 25 n 2 6 n 72 Table A2. 10 Table of refractive index measurements for CBD Wax Crystal Number Relative Refractive Index Crystal Number Relative Refractive Index n 1 n n 19 n 6 0 2 n 20 n 3 n 21 n 4 n 1. 5560 22 n 1. 6000 5 n 1. 5560 23 n 1. 6000 6 n 1. 5560 24 n 1. 6000 7 n > 1 .5300 25 n 8 n > 1.5300 26 n 1. 6040 9 n > 1.5300 27 n 1. 6040 n 10 n > 1. 5680 28 n 11 n > 1.5680 n 29 n 6 6 0 0 12 n > 1. 5680 30 n 6 6 0 0 13 n 1. 5800 31 n 1. 6 6 0 0 14 n 1. 5800 32 n 1. 6 6 0 0 15 n 1. 5800 33 n 6 6 0 0 16 n 1. 5900 34 n 1. 670 0 17 n 35 n 1. 670 0 18 n 36 n 1 . 670 0 Table A2. 11 Table of refracti ve index measurements for CBD Crystal Crystal Number Relative Refractive Index Crystal Number Relative Refractive Index n 1 n n 13 n 6 6 0 0 2 n 14 n 6 6 0 0 3 n 15 n 1. 6 6 0 0 4 n 1. 5560 16 n 6 6 0 0 5 n 1. 5560 17 n 1. 670 0 6 n 1. 5560 18 n 1. 670 0 n 7 n 1. 6000 19 n 1. 670 0 8 n 1. 6000 9 n 1. 6000 10 n 11 n 1. 6040 12 n 1. 6040 73 REFERENC ES 74 R EFERENCES (1) Polowsky P, T ansman G, Kindstedt P, Hughes J. Characterization and identification of surface crystals on smear - ripened cheese by polarized light microscopy. Journal of Dairy Science 2018;101(9):7714 23. (2) Duncke ACP, Marinho TO, Barb ato CN, Freitas GB, Oliveira MCKD , Nele M. Liquid Crystal Observations in Emulsion Fractions from Brazilian Crude Oils by Polarized Light Microscopy. Energy & Fuels 2016;30(5):3815 20. (3) Nie H, Xu W, Taylor LS, Marsac PJ, Byrn SR. Crysta lline solid disp ersion - a strategy to slowdown sal t disproportionation in solid state formulations during storage and wet granulation. International Journal of Pharmaceutics 2017;517(1 - 2):203 15. (4) S. Stoeffler, "A Flowchart System for the Identification of Common Synth etic Fibers by Polarized Light Mi croscopy," Journal of Forensic Sciences 41, no. 2 (1996): 297 - 299 (5) Bergfjord C, Holst B. A procedure for identifying textile bast fibres using microscopy: Flax, nettle/ramie, hemp and jute. Ultramicrosco py 2010;110(9):1 192 7. (6) Reffner JA, Kammrath BW, Kaplan S. A More Efficient Method for Synthetic Textile Fiber Analysis Using Polarized Light Microscopy. Journal of Forensic Sciences 2019;65(3):744 50. (7) Brinsko KM, Golemis D, King MB, Laughlin GJ, Sparenga SB. U.S . Department of Justice, 2016; Av ailable from: https://www.ncjrs.gov/pdffiles1/nij/grants/249854.pdf (8) Polarized Light Microscopy [Internet]. Nikon's MicroscopyU. Available from: https://www.microscopyu.com/techniques/polarized - light/pol arized - light - mic roscopy (9) Nesse WD. Introducti on to mineralogy. New York: Oxford Univ. Press, 2012; 10. The Essentials of Polarized Light Microscopy (10) Delly JG. Essentials of polarized light microscopy and ancillary techniques. Westmont, IL: Hooke College of Applie d Sciences, 2019; (11) Michel - Le vy Birefringence Chart [Internet]. Specialized Microscopy Techniques - Michel - Levy Birefringence Chart | Olympus Life Science. Available from: https://www.olympus - lifescience.com/en/microscope - resourc e/primer/techniques/pol arized/michel/ (12) Pichler H., Schmitt - Riegraf C. (1997) Observations under conoscopic light. In: Rock - forming Minerals in Thin Section. Springer, Dordrecht. https://doi.org/10.1007/978 - 94 - 009 - 1443 - 8_3 75 3 . SPECTROSCOPIC CHARACTER IZATION AND IDENTIFICAT ION OF CANNABIS SOLVENT EXTRACT CO MPONENTS Optical characterization, as discussed in Chapter 2 , provides the ability to group and identify samples based on optical characteristics of known crystals. In order to identify the previous ly characterized crysta lline components of each cannabis solvent extract, a variety of methods can be employed. The Scientific Working Group for the Analysis of Seized Drugs (SWGDRUG) provides recommendations for the analysis and identification of drug sam ples, including the can nabis extracts highlighted in this work. The methods by which forensic analysis and identification can be performed are categorized by SWGDRUG, ranking them from most to least discriminating. Based on these recommendations, one of th e most commonly used, s elective forensic analytical metho ds is spectroscopy, including Fourier - transform infrared (FTIR) spectroscopy and X - ray diffraction (XRD). Forensic laboratories often utilize FTIR spectroscopy as a screening or identification tech nique to analyze submit ted samples. As a SWGDRUG Category A method, FTIR spectroscopy can provide structural elucidation and identification, specifically for pure samples. The popularity of FTIR spectroscopy in forensic laboratories is due in part to the p otential for non - destru ctive analysis, depending on the s pecific type of analysis performed as well as the tenacity of the sample. Additionally, such methods can provide quantitative and confirmatory results rapidly when compared to other common analytical methods. 1 Forensic ana lysis using FTIR spectroscopy comm only includes a variety of trace evidence not limited to fibers, paint chips, and adhesives. 2,3 Further, FTIR spectroscopy has demonstrated suitability as both a quantitative and qualitative analytic al method for seized dr ugs analysis, although pure sample s are often necessary for identification. 4 - 6 Moreover, the manufacturing and marketing of FTIR 76 spectrometers specifically for the identification and quantification of cannabinoids has increased recen tly, due to the necessi ty to provide cannabinoid profiles for on - site potency testing. 7 - 9 The cannabinoids of interest for such on - site analyses include 9 - tetrahydrocannabinol ( 9 - THC), tetrahydrocannabinolic acid (THCA), and cannabidiol (CBD). An additional spectroscopic method recommended by SWGRUG as a Category A method for structural elucidation is XRD. Single - crystal XRD is an effective spectroscopi c method through which the exact identity of an unknown can be determined. While FTIR analysis pr ovides spectra to allow for interpretation and identification of functional groups within an organic compound, XRD provides the ability to identify the chemica l formula, structure, and crystalline unit cell of any diffracting material. Moreover, XRD is an excellent identification technique as it allows for the identification of unknown samples without the necessity of a reference standard for comparison. 10 Frequ ently, single - crystal XRD is used to establish the structure of protein crystals, for the purpose of visualizing the interaction between biochemical functions. However, small molecule crystal analysis has been performed, specifically with a forensic applic ation for the characterization and differentiation of designer drugs. 10,11 77 3.1 INSTRUMENTAL TH EORY 3.1.1 Micro - Attenuated Total Reflectance - Fourier Transform Infra - Red Spectroscopy Due to the ease of sample preparation and analysis, attenuated total r eflectance (ATR) - FTIR spectroscopy is commonly employed for IR analysis of forensic evidence. A ttenuated total reflectance analysis is based on the interaction of internally reflected infrared light with a sample that is in close contact with an ATR crys tal. This method can be both non - destructive and destructive, depending on the tenacity of the sa mple analyzed. In the case of this work, ATR - FTIR spectroscopy was a destructive analysis for the cannabis extract crystals. Standard ATR - FTIR analysis limits the size of samples compatible for analysis, as the entirety of the ATR crystal needs to be cover ed (~2mm in diameter) for the collection of representative spectra. While this lower size limit generally does not impact the routine analysis of common forens ic trace evidence materials, it is not well - suited for the analysis of single crystal analysis. A s such, t he spectroscopic analysis of single crystals was performed using a micro - ATR - FTIR due to the ability to collect spectra for crystal samples as small a s 1 0 µm. While the instrumentation between standard ATR - FTIR and micro - ATR - FTIR differs, the gen eral theory behind spectra collection and interpretation remains the same. Infrared light undergoes total internal reflectance between an optically dense ATR c rystal and the sample that is pressed between the crystal and pressure anvil due to the differenc e in refractive index between the two materials. 13 - Moreover, at the point of interaction between the ATR crystal and sample, an evanescent wave of infrared lig ht extends past the ATR crystal, interacts with the sample material, and is absorbed resulting in attenuated total reflection. 13 The penetration of the evanescent wave into the sample media is dictated by the wavelength of light, the angle of incidence be tween the ATR crystal and sample material, and 78 the indices of refraction for the ATR crystal and sample material. 14 Specifically, penetration depth is inversely related to wavelength, such that penetration is greater at lower wavelengths than higher wavele ngths. In general, the depth of penetration for most samples measured by ATR is 1 - 2 µm . Due to th e effect of wavelength on sample penetration, and the fact that an IR light source contains a range of wavelengths, the resulting FTIR spectra produced through ATR - FTIR differs from the standard absorbance spectra measured via transmission of light. 14 Most notably, ATR - the absorbance bands at higher wavenumbers. Due to the diffe rence between the spectra obtained by ATR - FTIR and transmission FTIR, it is ideal for spectral co mparisons to occur between reference and sample spectra obtained on the same instrument. When reference materials are not readily available for comparison, a l ibrary sample collected using ATR - FTIR rather than transmission FTIR can provide a method for pre sumptive identification. The instrumental set - up for micro - ATR - FTIR utilizes a spectrometer and IR light source attached to a fully operational polarized ligh t microscope. Figure 3.1 illustrates this instrumental setup. For analysis by micro - ATR - FTIR, the sample of interest is mounted on a standard microscope slide and can be observed in plane - polarized light and crossed - polarized light as with a standard polar ized light microscope. An ATR diamond crystal objective lens is used to make contact, essentially crushing the crystal sample, in order to produce total internal reflection and the evanescent wave. Additionally, micro - ATR - FTIR fitted microscopes include sp ectrometers with infrared light sources, interferometer, directing optics, and a detector. Figure 3.2 provides a simplified scheme of an ATR objective and spectrometer. 79 Figure 3.1 - ATR - FTIR instrumental set - up 80 Figure 3.2 Si mplified schematic of an ATR microscope objective and infrared spectrometer commonly used for microscopic infrared analysis 3.1.2 X - Ray Diffraction Theory and Instrumentation Single - crystal XRD is a powerful analytical technique through which an unkno w n sample can be characterized and identified. This technique utilizes X - ray radiation as an excitation source to aid in the determination of a crystalline unit cell. Further, to define the crystal structure, X - ray diffraction patterns for a unique crystal ar e solved, providing electron density maps which indicate the likely atom placement. The fundamental concepts related to the determination of the crystalline unit cell and chemical formula include crystal structure, light and X - ray diffraction. Single - c r ys tal XRD is a form of crystallography which refers to the study of the properties, structure, and formation of crystals. Crystal structure is categorized using Bravais lattices, which represents the most basic 81 building block of the crystal. 15 The Bravais la ttices are a collection of 14 different groups of points, which can be further categorized into seven crystal systems. Crystal systems were first introduced in Chapter 2 , with their relation to optical properties of a crystal. Nonetheless, a more in - dep t h explanation of crystal systems is necessary in its application to X - ray diffraction. The seven crystal systems are differentiated by the relationship between atomic spacing and angles in the unit cell. Table 3. 1 outlines the crystal systems and provides a ccompanying geometric summaries of each unit cell. The unit cell is the smallest group of atoms that, when repeated, produce the lattice of a crystal system. 15 The length between two points on the corners of the lattice, or crystallographic axis, is deno te d by a, b, or c, while the angle between the crystallographic axes is given by . The unit cells increase in complexity and decrease in symmetry from cubic, in which all three crystallographic lengths a nd angles are equal, to triclinic, in whic h there are no equal lengths or angles. 82 Table 3. 1 Summary of crystal systems with geometric unit cell depictions Crystal System Crystallographic Lengths and Angles Unit Cell Depiction Cubic a = b = c, = 90° Tetragonal = 90° Orthorhombic = 90° Rhombohedral a = b = c , Hexagonal = 120° Monoclinic Triclinic 83 The cubic lattice is the simples t and most symmetrical of all crystal systems with all angles equal to 90 and all axes equal in length. The Bravais lattices associated with the cubic crysta l system are the primitive cubic, body - centered, and face - centered cubic. The body - centered cubic B ravais lattice includes an atom in the center of the cube as well as atoms positioned in the corners of the cubic structure (at the location of the primitive cubic vertices). The face - centered cubic includes atoms at the locations of the primitive cubic l a ttice as well as atoms in the center of each cubes face. An additional form of Bravais lattice is the base - centered lattice, which includes atoms in each pri mitive vertex and atoms positioned on two opposing faces, generally perpendicular to the c axis of the unit cell. 15 Base - centered Bravais lattices are not part of the cubic system, but rather the orthorhombic and monoclinic crystal systems. Figures of the geometric symmetry of crystals by the Bravais lattices are provided in Appendix III. To determine t he crystal system and Bravais lattice of a crystal, X - ray diffraction is often employed. Diffraction occurs based on the interaction of monochromatic light w ith a regular, repeating material. The periodicity of the material increases the magnitude of the d iffraction due to constructive interference of the diffracted light rays. The constructive interference occurs as diffracted light rays combine their amplitu des, resulting in a point of diffraction with an intensity proportional to the square of the light s amplitude. 1 6 The regularly repeating lattice structures of crystals interact with light similarly, producing a diffraction pattern related to the spacing o f atoms in the lattice. In order to produce diffraction on the atomic scale, X - rays are necessary a s their wavelengths are on the same scale as the distance between atoms. The interaction between the crystal lattice and X - w, which states that constructive interference can only occur when Equation 3.1 is satisfied. 16 I n this equation, n is an integer, 84 is the X - ray wavelength, d is the distance between lattice parallel planes, and is the angle of incidence. Figure 3. 3 de monstrates this relationship. Equation 3.1 Figure 3. 3 16 In o r der to completely characterize a crystal lattice using diffraction, the diffraction patterns need to be obtained in many different angles and orientations around the crystal. This is due to the three - dimensional periodicity of crystals, which produce many different diffracting 16 The collection of diffraction reciprocal lattice includes inform atio n regarding the Fourier transform of the spatial wavefunction of the original lattice. 16 It is important to denote the difference between the includes the phy sica l crystal, atoms, and lattice. Reciprocal space is the Fourier transform of real 85 space, resulting from the diffraction pattern of all possible crystal orientations. By collecting diffraction patterns from all possible crystal orientations, a three - dime nsio n al reciprocal lattice can be constructed, thus allowing the real crystal lattice to be solved by performing a Fourier transform. Solving the crystal structure by XRD requires the recombination of the diffracted beams with the correct amplitudes and r elat i ve wave phases. 17 This recombination provides a map of the electron densities of a given plane of the crystal lattice, representing the positioning of atoms in the unit cell. Each diffraction spot (or reflection) includes a different intensity related to t he surrounding spacing of electron density features in each unit cell. 18 The recombination is performed via Fourier transformation, where the amplitude and phase of each diffracted X - ray wavelength act as Fourier coefficients and can be added together by F ourier synthesis to obtain the original periodic function ( i.e. , electron density) represented by the diffraction pattern. 17 In a diffraction experiment, however, only the amplitude of the diffracted beam is known, not the phases. This gives rise to the (less than 20 atoms), solving the phase problem is unnecessary, as the amplitude of a diffracted spot generally provides enough insight related to possible structural features. Rather than solving the phase problem in these small molecules, the Patterson function can be employed, which provides a map of peaks at every interatomic vector position in the structure. However, this function is only useful for small molecules as the number of interatomic vec t ors increases with N atoms, such that the function will contain N(N - 1) vectors, with many overlapped. 18 Additionally, small and intermediate sized crystals can be solved using direct methods, which provide a mathematical method through which phases ar e ch o sen for strong reflections based on knowledge of the possible molecule. The phases for other reflections are then generated based on 86 known interaction relationships between strong reflections and an electron density map is constructed to determine ato mic p ositioning. 18 Structural refinement following the determination of a rough electron density map is performed given the ability to apply chemical knowledge to a structure, improving the electron density and phases together. 3. 2 MATERIALS AND METHODS 3.2. 1 Samples The same nine cannabis extract samples introduced and optically characterized in Chapter 2 were further chemically characterized via micro - ATR - FTIR and single - crystal XRD. Table 3. 2 summarizes the source, identity, and age of the samples. Table 3 . 2 Sample ide ntifications, sources, and year obtained Sample Identity Source Year Obtained KDPS 18 - 9026 KCSD 2018 PPO 14 - 20332 - 10 KCSD 2014 KCSD 14 - 10811 - 28964 KCSD 2014 KCSD 14 - 10811 - 28967 KCSD 2014 KCSD 14 - 10811 - 28960 KCSD 2014 t Sugar/Wax Skymint 2020 THCA crystals Skymint 2020 CBD Shatter Crystals Cannabidiol Life 2020 CBD Crumble/Wax Cannabidiol Life 2020 87 3.2.2 Micro - Attenuated Total Reflectance - Fourier Transform Infra - Red Spectroscopy Sample Preparatio n Characteristic c rystals from each case sample were selected for m icro - ATR - FTIR. Each sample crystal was cleaved from a larger crystal such that both PLM characterization and spectroscopic analysis were performed on equivalent crystal s. Micro - ATR - FTIR an a lysis was performed using an Olympus BX51 PLM (Olympus Corporations of America , Center Valley, PA spectrometer and micro - ATR objective ( Smiths Detection, Edgewood, MD ). Prior to crystal analysis, the FTIR center burst was ali g ned and the strength of the center burst peak was verified to be ~14,500 units. Additionally, between each sample analysis, instrument background scans were collected with the micro - ATR objective placed over, but not in contact with, the sample. Triplicat e spectra were collected for each case sample. Instrument parameters for analysis were as follows: r esolution of 4 cm - 1 , 20 - 1 00 µm aperture (depending on crystal size) , and 16 scans collected. Samples were crushed using the micro - ATR objective until the en t irety of the aperture was covered by the crushed crystal sample, but without pressing so hard that interference from the sup porting glass microscope slide overwhelmed the sample spectrum. Sample spectra were characterized based on functional group frequen c ies tabulated by Pretsch et al . 1 9 and compared to THC and THCA certified reference materials (all Cayman Chemical Co, Ann Ar bor, MI) for preliminary identification. 88 3. 2 . 3 Single - Crystal X - Ray Diffraction Sample Preparation Single crystals from case sa m ple KDPS 18 - 9026 , Skymint THCA crystal sample, and Cannabidiol Life CBD crystal sample were selected for crystallographic id entification via single - crystal XRD . The individual crystal s were mounted on a nylon loop using paratone oil. Single - crystal XRD fo r case sample KDPS 18 - 9026 was performed using a Bruker APEX - II CCD diffractometer (Bruker Analytical X - ray Systems, Madison, WI) at a low temperature T = 173 K with an Oxford Cryosystems low - temperature device (Oxfor d Cryosystems Ltd., Long Hanborough, Ox f ord, UK). In the time bet ween the analysis of the KDPS case sample and the dispensary samples, a new single - crystal XRD was purchased by the Center for Crystallographic Research at Michigan State University. As such, analyses for the Skymint THCA crystal a nd Cannabidiol Life CBD c rystals were performed using a XtaLAB Synergy, Dualflex, HyPix diffractometer ( Rigaku Americas Corporation, The Woodlands, TX ) at a low temperature T=100.01(10) K with an Oxford Cryosystems low - temperature device (Oxfor d Cryosyste m s Ltd.). For all crystal analyses, X - ray intensity data w ere polarization was corrected for the KDPS 18 - 9026 sample using the SAINT software (Bruker Analytical X - ray Systems, V8.38A, after 2013, Madison, WI) , while t he dispensary samples pol arization was corrected using CrysAlisPro software (Agilent Technologies, XRD Products Oxfordshire, UK). For all crystals, the absorption correction was performed in the SADABS - 2016/2 software (Bruker Analytical X - ray Systems V2 . 008/2 2016/2, Madison, WI ). The crystal structure was solved using dual methods using ShelXT (Sheldrick, G.M. (2015). Acta Cryst. A71,3 - 8) . Refinement of the structure was performed using an Olex2 incorporated form of Least Squares (developed based on res e arch by Dolomanov et al . 2 0 ) using version 2014/6 of XL (as developed by Sheldrick 21 ) . All non - hydrogen atoms were refined anisotropically. Hydrogen atom 89 positions were calculated geometrically using the riding model. This excludes hydrogen atoms on non - ca r bon atoms which were foun d by difference Fourier methods and refined isotropically. Crystal dimensions for the KDPS 18 - 9026 case sample crystals were 0.18 × 0.07 × 0.06 mm 3 . Crystal dimensions for the Skymint PB THCA Crystals sample were 0.12 × 0.07 × 0.0 3 mm 3 . Crystals dimensions for the Cannabidiol Life CBD Crystals sample were 0.33 × 0.22 × 0.15 mm 3 . 3. 3 FTIR SPECTROSCOPIC ANALYSIS OF CANNABIS SOLVENT EXT RACTS 3.3.1 Results of Micro - ATR - FTIR Analysis of Cannabis Solvent Extracts 3.3.1.1 Micro - ATR - FTIR A nalysis of Optically Similar Crystals - KCSD Case Samples and Skymint Samples Due to the similarities in optical properties between the KCSD BHO case sampl es and the Skymint BHO crystal and wax dispensary samples outlined in Chapter 2 , it was assumed tha t the crystalline components of these samples shared chemical characteristics. While each of the seven samples between these two groups was characterized by micro - ATR - FTIR, one sample from each subset was selected to represent each group. Ca se sample KDPS 1 8 - 9026 was selected to represent the KCSD case samples, while the Skymint WB THCA Crystal sample represents the Skymint dispensary samples. Additionally, wh en present, the wax components of each BHO sample were also analyzed by micro - ATR - FT IR to determine any chemical similarities between waxes with varying textures. The KCSD and Skymint samples and wax textures are summarized in Table 3.3 . 90 Table 3. 3 KCSD an d Skymint sample summary Sample Source Sample Identity Sample Consistency KCSD Case Samples KDPS 1 8 - 9026 Wax (Crumble) PPO 14 - 20332 - 10 Wax (Soft /Viscous) KCSD 14 - 10811 - 28964 Wax (Soft to Glass - like) KCSD 14 - 10811 - 28967 Wax/Shatter (Solid, hard; Gla ss - like) KCSD 14 - 10811 - 28960 Wax (Soft to Glass - like) Skymint Dispensary Samples PB THCA Wax Cry s talline (Little Oily/Wax Matrix) WB THCA Crystals Crystalline Material (No Wax Material) Crystalline and wax components (when present) of each sample w ere analyzed by micro - ATR - FTIR and compared to a THCA certified reference material. A representativ e spectrum of case sample BHO crystal and wax components is provided by KDPS 18 - 9026 ( Figure 3.4 ) while a representative spectrum of the Skymint dispensary s amples is provided by WB THCA Crystals ( Figure 3.5 ). T hese spectra display a rising baseline in the fingerprint region most likely caused by interference due to the glass microscope slide on which the crystal was mounted for analysis. The stretching freque ncy for the Si - O bond in glass is located from ~1200 - ~800 cm - 1 and produces a very strong, broad p eak. The size constraints of the crystals following cleavage (~25 - 50 µm ) allows for portions of the micro - ATR crystal to contact the supporting glass slide , rather than be entirely covered by the crushed cryst alline sample. Figure 3. 6 provides an overlay of normalized FTIR spectrum of all optically similar crystalline samples for visual comparison, while F igure 3.7 provides the normalized FTIR spectra for th e wax component of each BHO 91 case sample. The separate representative spectra of the additional samp l es are included in Appendix III. Figure 3.4 KDPS 18 - 9026 crystal (top) and wax component (bottom) IR spectrum F igure 3.5 Representative Skymint sample - WB THCA Crystal IR spectrum -0.1 1 650 1150 1650 2150 2650 3150 3650 Normalized Abs. Wavenumbers (cm - 1 ) 0 2.5 650 1150 1650 2150 2650 3150 3650 Normalized Abs. Wavenumber (cm - 1 ) KDPS 18-9026 Wax KDPS 18-9026 Crystal 92 Figure 3.6 Stacked FITR spectra for all sample crystalline componen ts -0.1 8.5 650 1150 1650 2150 2650 3150 3650 Normalized Abs. Wavenumber (cm - 1 ) PB THCA Wax WB THCA Crystals KDPS 18-9026 KCSD 14-10811-28964 KCSD 14-10811-28960 KCSD 14-10811-28967 PPO 14-20332-10 93 Figure 3.7 Stacked FITR spectra for all case sample wax components To presumptively identify the crystalline and wax components of the BHO case sampl es, a representative spectrum from each sample was first subjected to a standard library search and co mpared to the top matches. The most comparable match for each sample crystalline components was a nujol mull preparation o f THCA. Comparatively, the top r esult for the wax component differed between nujol mull preparations of THCA and independent librar y s pectra of 0 6 650 1150 1650 2150 2650 3150 3650 Normalized Abs. Wavenumber (cm - 1 ) KDPS 18-9026 KCSD 14-28960 KCSD 14-28967 KCSD 14-29864 PPO 14-20332-10 94 THC samples. Following the preliminary library identification, the case sample spectra were compared to the FTIR spectra of THCA and THC collect ed using micro - ATR - FTIR ( Figure 3.8 and Figure 3.9 ). Figure 3.8 Stacked comparison of THCA CRM ( bott om) and case sample KDPS 18 - 9026 crystal (middle) and wax component (top) micro - ATR - FTIR spectra -0.1 3 650 1150 1650 2150 2650 3150 3650 Normalized Abs. Wavenumber (cm - 1 ) KDPS 18-9026 wax KDPS 18-9026 crystal THCA CRM 95 Figure 3.9 Stacked c omparison of THC CRM (bottom) an d case sample KDPS 18 - 9026 crystal (middle) and wax component (top) micro - ATR - FTIR spectra 3.3.1. 2 Mic ro - ATR - FTIR Analysis of Cannabidiol Life Dispensary Samples Based on the optical properties discussed in Chapter 2 , the Cannabidiol Life dispensary samp les contained optically different crystal components compared to the KCSD case samples or Skymint s ubset s. As such, micro - ATR - FTIR analysis was performed to determi ne the chemical composition of the crystals for further comparison and differentiation. Tabl e 3.4 provides a summary of the sample identity and consistency . It should be noted that each sampl e inc luded terpene blueberry flavoring, which differs from the KC SD case samples or Skymint dispensary -0.1 3 650 1150 1650 2150 2650 3150 3650 Normalized Abs. Wavenumber (cm - 1 ) KDPS 18-9026 wax KDPS 18-9026 crystal THC CRM 96 samples. Additionally, the wax matrix observed in the CBD Wax sample was oily, as opposed to viscous or sticky. Table 3.4 . Cannabidiol Life CBD - containi ng sa mple summary Sample Identity Sample Consistency CBD Wax Mostly Crystalline Wax (Oily/Wax matrix) CBD Shatter Crystals Crystalline Material (No Wax Matrix) Crystalline components of each sample were analyzed by micro - ATR - FTIR and compared to l ibrar y search results and THCA certified reference material. Figure 3.1 0 provides a spectrum for the Cannabidiol Life CBD Wax sample (both crystalline and oi ly residue), while Figure 3.1 1 displays a spectrum for the CBD Shatter Crystals sample. The noise p resen t in the used for analysis. The lack of residue around the crystals resulted in an overa ll lack of sample for ATR - FTIR analysis, leading to a lower quality s pectr um collected. 97 Figure 3.1 0 Micro - ATR - FTIR spectrum for the crystalline component of sample CBD Wax crystal (bottom) and wax (top) components Figur e 3.1 1 Micro - ATR - FTIR spectrum for sample CBD Shatter Crystals 0 1 650 1150 1650 2150 2650 3150 3650 Normalized Abs. Wavenumber (cm - 1 ) 0 2 650 1150 1650 2150 2650 3150 3650 Normalized Abs. Wavenumber (cm - 1 ) 98 To presumptively identify the crys tallin e and wax components of each Cannabidiol Life sample, each were subjected to a standard library search using the GRAMS IR spectral comparison software. For both crystalline samples, and the wax component of the CBD Wax sample, the top result was to a CBD n ujol mull sample . Additionally, each Cannabidiol Life sample was compared to a THCA CRM. Comparison of the CBD Shatter Crystal sample to the THCA CRM s pectrum ( Figure 3.1 2 ) highligh ted key differences. Figure 3.1 2 Stacked comparison of CBD Shatter Crysta l (top) and THCA CRM (bottom) 650 1150 1650 2150 2650 3150 3650 Normalized Abs. Wavenumber (cm - 1 ) 99 3. 3 . 2 Discussion of Micro - ATR - FTIR Analysis of Cannabis Solvent Extracts As shown by Figures 3.4 and 3.5 , the crystall ine component of KDPS 18 - 9026 samples contains chemical characteristics comparable to that of the S kymint WB THCA Crystal dispensary sample. This is apparent due to both the presence of peaks at similar wavenumbers, as well as the ratio between normalized peak absorbances betwe en the two crystal samples. The peaks present at ~2850 and ~2900 cm - 1 are ind icative of aliphatic C - H stretching. The lack of significant peaks above 3000 cm - 1 suggests that the crystalline structure likely does not include many unsat urated C - C bonds (or t hose with adjacent C - H bonds), hydroxyl, or amine groups. Bands in the region betwee n 1600 and 1650 cm - 1 indicate alkene stretching, including cyclical and conjugated alkenes. In the fingerprint region of the spectrum, medium strength peaks located between ~1350 and ~1450 cm - 1 are commonly due to hydroxyl bending, including bending from a carboxylic acid or phenol functional group. The most intense peak in the spectra at ~1250 cm - 1 is due to C - O stretching, particularly by alkyl aryl e thers. Similarly, the peaks present between ~1120 and ~1190 cm - 1 are caused by C - O stretching from tertiar y hydroxyl groups. Finally, peaks below 900 cm - 1 are generally due to C - H deformations, such as bending, twisting, and rocking. Despite the lack of c haracteristic high fre quency C - H stretching that occurs with unsaturated C - C bonds, strong bands su ggestin g alkene presence are included in the spectra between 1600 and 1650 cm - 1 . This indicates the lack of hydrogens bonded to unsaturated carbons for this structure. Such a patt ern of stretching points to the possibility of a cyclic alkene and multiply s ubstitu ted aromatic ring structure with the addition of a saturated carbon chain, as indicated by the aliphatic C - H stretching and C - H deformations. Addition ally, the stretching f requency in the region significant to alkyl aryl ethers further supports the likelih ood of an 100 aromatic ring as part of the structure. Regardless of the strong stretching bands attributed to C - O bonds and hydroxyl bending, the spectrum lacks the characteris tic broad, high frequency intermolecularly bonded O - H stretching peaks. This may ind icate possible hydrogen bonding present in the crystalline structure, further stabilizing the hydroxyl groups. Comparison of all crystalline sample s pectra indicated simil arities in structural components. Not only were the peaks present between sam ples si milar in location, but the ratios between peaks were also comparable. Most notably, the crystalline component of each crystalline sample lacked O - H st retching in the high f requency region of the spectra but maintained peaks characteristic of C - O str etching and hydroxyl bending in the lower frequency regions of the spectra. Similarly, the presence of peaks indicative of cyclic and conjugated alkenes were found in each case sa mple spectra. These similarities provide confirmation that the crystalline co mponent s in both clandestine BHO samples and dispensary produced BHO samples are chemically comparable, in addition to being optically similar as described i n Chapter 2 . Compariso n of case sample spectrum to a spectrum of THCA CRM ( Figure 3.8 ) highlighted similar ities in organic functional groups. Similarities included peak presence and ratio specifically in the C - H bending (~850 cm - 1 ), C - O alkyl aryl ether st retching (~1250 cm - 1 ), O - H carboxylic acid and phenol bending (~1450 cm - 1 ) and alkene stretching re gion (~ 1600 cm - 1 ). Additionally, comparison of case sample spectra to a THC reference material spectr um ( Figure 3.9 ) indicated stark differences, including a strong O - H stretching peak (~3200 - 3500 cm - 1 ) and defined O - H phenol bending region (~1350 and ~145 0 cm - 1 ) . Due to the similarities between the THCA CRM and each case sample crystal, the crystalline component of BHO extracts was presumptively identified as THCA. 101 The spectrum o thei r cryst alline counterparts, with some minor, but notable, differences. The most readily visible spectral difference between these two components is the weak, broad peak present in the intermolecular O - H stretching region (~3200 - 3500 cm - 1 ) of the wax compon ent spe ctra. This peak is more defined in the individual IR spectra for each case sample (Appendix III). Further, n peaks at similar wav enumbers, the relative ratios between peaks may indicate slight structural di fferenc es between the two components. Most notably, the peak ratios in the O - H bending (~1450 cm - 1 ) regions and C - O stretching (1250 cm - 1 ) differ within the wax samples and between the wax and crystalline component spectra ( Figure 3.4 ). The characterizati on of t he wax component was completed to further identify possible cannabinoids or natural waxes present. The texture and amount of wax component present dif fered between case samples, with textures ranging from soft and viscous to glass - like ( Table 3.3 ). However , despite the differences in texture for the wax component, similar functional groups were identified. The harder, glass - like waxes appear ed to provid e a more defined, high absorbance O - H stretching peak, but the difference was minimal. Comparison o f the w ax component spectra of each case sample to the THC CRM spectrum highlighted similarities in organic functional groups. Similarities inclu ded peak pre sence and ratio specifically in the O - H stretching (~3350 cm - 1 ), C - H bending (~850 cm - 1 ), C - O alkyl aryl e ther stretching (~1250 cm - 1 ), O - H bending (~1450 cm - 1 ) and alkene stretching region (~1600 cm - 1 ). The addition of peaks in the O - H stretch ing region b etter correlated the wax component to THC than THCA, although similarities between the two cannabin oids we re present given their high degree of structural similarity. 102 Given the optical differences between the Cannabidiol Life samples and the c ase samples and dispensary samples discussed previously, structural and chemical differences were expected. Whe n analy zed by micro - ATR - FTIR, the crystalline component of the CBD Wax sample was analogous to the CBD Shatter Crystal sample ( Figure 3.1 0 and 3. 1 1 ). This wa s apparent due to both the presence of peaks as similar wavenumbers, as well as the ratio between n ormaliz ed peak absorbances between the two crystal samples, with the only slight difference in ratios appearing in the fingerprint region of the spectra. The spectrum collected for the wax component of the CBD Wax sample included similar peaks to that of t he crys talline component, indicating related chemical composition. However, some peak intensities and ratios differed between the two CBD Wax sam ple componen t spectra, but this did not lead to differences in library search results or spectral interpretatio n. Wit hin both CBD samples, the peaks present at ~3500 and ~3400 cm - 1 are consistent with intermolecularly bonded O - H stretching. The weak, but reproducible peak at 3071 cm - 1 relates to alkene C - H stretching, though such a stretching frequency is generall y obser ved as a stronger (medium) relative absorbance. The weakness of this peak could indicate that the crystalline structure does not contain a significant amount of C - H bonds adjacent to alkene functional groups. The peaks between ~2850 and ~2900 cm - 1 are ind icative of aliphatic C - H stretching. Bands in the region between 1600 and 1650 cm - 1 indicates alkene (C=C) stretching, including cyclical and conjugat ed alkenes. In the fingerprint region of the spectrum, the medium/strong strength peak located at 1 440 cm - 1 can be attributed to either hydroxyl or methyl bending, as both functional group frequencies occur in this region. The strongest peak in the spectra at ~1210 cm - 1 is due to C - O stretching, possibly from ethers or hydroxyl functional group presence . Addit ionally, the series of sharp peaks present around ~1010 cm - 1 may be due to alkene bending, 103 while the sharp peaks below 900 cm - 1 are genera lly due to C - H deformations, such as bending, twisting, and rocking. As highlighted in Table 3.4 , Cannabidiol L ife CBD Shatter Crystals had only a crystalline component, however CBD Wax contains both a crystalline and an oily, wax component. Given the incl usion of wax components in the THCA - containing samples, chemically characterizing the wax component of the CBD Wax sam ple may provide a more comprehensive analysis of the cannabinoids present in extracts derived from hemp, or other natural oils. Figure 3.1 0 provides a representative spectrum of the oily, wax matrix surrounding the CBD Wax crystalline component. Whe n subje cted to a standard library comparison search, a CBD nujol mull spectrum provided the best visible comparison. Additionally, the two Cannabidiol Life s amples were compared to the THCA CRM material that was spectrally similar to the KCSD case samples and Sky mint dispensary samples ( Figure 3.12 ). Though the Cannabidiol Life and THCA CRM spectra contain similar functional groups based on I R interpretation, the wavenumber and ratios of peaks in each spectra differ drastically. Most notably, the inclusion of O - H stretching and unsaturated C - H stretching peaks provided key discrepancies between the CBD Shatter crystals and THCA CRM. Moreover, the peaks in the C BD Shatter Crystal spectrum are more defined and sharper compared to the THCA CRM spectrum. The inc reased sharpness and definition in the CBD Shatter Crystal spectrum may be due to the well - formed crystalline structure of the sample, comp ared to the dried, oily residue of the THCA CRM. As presented in the previous sections, the two optically different crystal groups, as represented by clandestine and dispensary BHO samples and the Cannabidiol Life dispensary samples, also differ in chemic al characterizatio n. To better showcase the difference in chemical 104 characteristics as determined by micro - ATR - FTIR, F igure 3 .1 3 provides stacked spectra of representative samples from each sample set: KDPS 18 - 9026, WB THCA Crystal, and CBD Shatter Crystal samples. Figure 3.1 3 Stacked spectra comparing the crystalline components of KDPS 18 - 9026 (bottom), WB THCA Crysta ls (mid dle), and CBD Shatter Crystals (top) As displayed, the KDPS 18 - 9026 and WB THCA Crystals samples contain similar functional groups , represented by t he similar peak location and ratios present in the stacked spectra. Oppositely, the CBD Shatter Cry stal sa mple does not spectrally correlate to either of the THCA - containing samples, despite sharing structural functional group similaritie s, such as conjuga ted alkene groups (aromatic ring structures), hydroxyl groups, and alkane moieties. 650 1150 1650 2150 2650 3150 3650 Normalized Abs. Wavenumber (cm - 1 ) KDPS 18-9026 WB THCA Crystals CBD Shatter Crystals 105 Additionally, d espite the inclusion of carboxylic acid and hydroxyl functional groups in the KDPS 18 - 9026 and WB THCA Crystal samples, distinctive O - H str etching is only ob served in the CBD Shatter Crystal sample spectrum. This indicated the likely presence of strong hyd rogen b onding present in the THCA - containing crystal samples, but lack of such strong bonding in the CBD - containing samples. 3.4 STRUCTURUA L ELUCIDATION OF C ANNABIS SOLVENT EXTRACT CRYSTALS 3. 4.1 Results of Single - Crystal XRD Analysis for Structural Elucid ation 3 . 4.1.1 Single - Crystal XRD Analysis of Optically Similar Samples - KDPS 18 - 9026 and WB THCA Crystals The representative samples from each subset of opt ically similar crystals were analyzed by single - crystal XRD for structural elucidation and confirma tory id entification. Following XRD data collection and structural refinement, the crystal component of the BHO case sample extracts, as rep resented by KDPS 1 8 - 9026, was determined to be the cannabinoid THCA. Similarly, the crystals from the Skymint WB THCA Crysta l sample were identified as THCA. The solved crystal structure is displayed in Figure 3.14 . The crystal structure shows dimer proper ties due to hydrog en bonding from carboxylic acid groups ( Figure 3.1 5 ). Details of crystal data and refinement for ea ch samp le are compiled in Table 3.5 . Due to the high degree of similarity in structural results between the two crystals, the figures provi ded here are only a subset of data collected specifically related to the single - crystal XRD analysis of case sample K DPS 18 - 9026; however, similar figures for the WB THCA Crystal sample are shown in Appendix III, along with supplemental figures from each X RD report. 106 Fig ure 3.1 4 Crystal structure from case sample KDPS 18 - 9026 determined by single - crystal XRD shown wit h molecu lar labelling scheme. Figure 3.1 5 Crystal structure from case sample KDPS 18 - 9026 displaying racemic, dimer crystalline properties 107 Table 3. 5 Crystal data and structure refinement details of KDPS 18 - 9026 and WB THCA Crystal samples KDPS 18 - 902 6 WB THC A Crystals Formula C 22 H 30 O 4 C 22 H 30 O 4 D calc. / g cm - 3 1.193 1.213 µ /mm - 1 0.644 0.655 Formula Weight 358.46 358.46 Color colorless colorless Shape chunk block Size/mm 3 0.18×0.07×0.06 0.12×0.07×0.03 T /K 173(1) 100.00(10) Crystal System orthorhombic orthorhombic Flack Parameter 0.2(5) - 0.07(7) Hooft Parameter 0.3(3) - 0.04(6) Space Group P 2 1 2 1 2 1 P 2 1 2 1 2 1 a /Å 11.4189(3) 11.40867(12) b /Å 18.1043(6) 17.98950(19) c /Å 19.3024(7) 19.1297(2) a / ° 90 90 b / ° 9 0 90 g / ° 90 90 V/Å 3 3990.4(2) 3926.11(8) Z 8 8 Z' 2 2 Wavelength/Å 1.54178 1.54184 Radiation type CuK CuK Q min / ° 3.347 3.372 Q max / ° 58.941 77.428 Measured Re flections 14444 27768 5555 801 8 Refl 's with I > 2(I) 3091 7551 R int 0.1207 0.0458 Parameters 482 493 Restraints 0 0 Largest Peak 0.195 0.164 Deepest Hole - 0.218 - 0.191 GooF 0.985 1.038 wR 2 (all data) 0.1390 0.0834 wR 2 0.1098 0.0819 R 1 (all data) 0 . 1434 0 .0371 R 1 0.0638 0.0341 108 3. 4 . 1. 2 Single - Crystal XRD Analysis of CBD Shatter Crystals Due to the similarity in chemical composition and optical characteristics between the crystalline components of each dispensary sample, a r epresent ative crystal from the CBD Shatter Crystal sample was selected for structural elucidation and identification by single - c rystal XRD. Following XRD data collect ion and structural refinement, the crystal component of CBD - containing dispensary samples, as repr esented by CBD Shatter Crystals, was determined to be the cannabinoid CBD. The solved crystal structure is displayed in Figure 3.16 . The crystal structure sho ws a chiral crystal with dimer properties due to hydrogen bonding from hydroxyl groups ( Fi gure 3.1 7 ). Details of crystal data and refinement for each sample are compiled in Table 3.6. 109 Figure 3.1 6 Crystal structure f rom dispensary sample CBD Shatter Crys tal determined by single - crystal XRD shown with molecular labelling scheme of chiral atoms . Figur e 3.1 7 Crystal structure from dispensary sample CBD Shatter Crystal displaying hydrogen bonding and dimer crystalline pr operties 110 Table 3. 6 Crystal data and s tructure refinement details of CBD Shatter Crystal sample CBD Shatter Crystal Formula C 2 1 H 30 O 2 D calc. / g cm - 3 1.110 µ /mm - 1 0.535 Formula Weight 314.45 Color colorless Shape block Size/mm 3 0.33×0.22×0.15 T /K 100.01(10) Crystal Syste m monoclinic Flack Parameter - 0.02(10) Hooft Parameter 0.01(9) Space Group P 2 1 a /Å 10 .40257(14) b /Å 10.89329(15) c /Å 16.6836(2) a / ° 90 b / ° 95.5081(12) g / ° 90 V/Å 3 1881.82(4) Z 4 Z' 2 Wavelength/Å 1.54184 Radiation type Cu K Q min / ° 2.661 Q max / ° 77.347 Measured Refl's. 24375 Ind't Refl's 7633 Refl's with I > 2(I) 7318 R int 0.0487 Parameters 438 Restraints 1 Largest Peak 0.169 Deepest Hole - 0.147 GooF 1.024 wR 2 (all data) 0.0857 wR 2 0.084 4 R 1 (all data) 0.0363 R 1 0.0348 111 3.4.2 Discussion of Single - Crystal XRD Analysis for Structural Elucidation Case sample KDPS 18 - 9026 and the Skymint dispensary WB THCA crystal sample were both identified as THCA; with analogous crystal structure s identified by single - crystal XRD. As shown in T able 3.5 , the chemical formula determined for the crystal structures match that of THCA ( C 22 H 30 O 4 ) , and the structures between the solved crystalline component and THCA are analogous. A representati ve structu re of THCA is provided in Appendix III. The crystal system for the structure was determined to be orthorho mbic, based on the presence of unequal crystallographic axis lengths (a, b, and c) while all maintaining right angles to one another ( , , ). Additi onally, the space group for the crystal was found to be P2 1 2 1 2 1 , indicating a primitive Bravais lattice (d enoted by P) in the orthorhombic crystal system. The notation 2 1 2 1 2 1 refers to the symmetry related to the screw axes along the x, y, and z coordinate directions. Within the unit cell, eight total THCA molecules are present, denoted by Z in Table 3. 5 . The refers to the number of molecules that make up the asymmetrical unit the smallest portion of the unit cell that can be moved s ymmetrical ly (in this case by screw operations) to produce the full unit cell. A packing diagram depicting the molec ules in the unit cell is provided in Appendix III. Similarities and small differences between the refined data of the two crystalline sample s are pres ent when comparing the results in Table 3. 5 . Most notably, the solved chemical formula, mass, crystal syst em, space group, and crystallographic axis angles are analogous between the two samples. This corresponds to the similarity in their optical propertie s. While the solved structure and identity of the two crystal samples can be considered analogous, there a re numerical differences in the exact results provided by Table 3.5 . During data collection for KDPS 18 - 9026, 14444 total reflections were m easured. O f those measure reflections, 5555 were determined to be independent and not related by symmetry. These dat a 112 were collected to a final completeness of 99.50%, which is desirable for structural refinement. The values for total and independent refle ctions for WB THCA Crystals were larger than that of KDPS 18 - 9026, and the data were collected to a final completene ss of 100%. The difference in total data collected is instrument - dependent and does not reflect an inadequacy or quality issue in regard to the crysta l analysis. The parameters that define the fit of the determined structure to the experimental electron de 2 1 should approach 1 for well - fit models, while wR 2 and R 1 should approach 0. Ideal structural solutions would have R 1 values equal to 0, but due to random error this cannot occur. For case sample KDPS 18 - 9026, the confidence value for GooF, wR 2 (all data) and R 1 are in an acceptable, publishable range. 22 Similarly, for dispensa ry sample WB THCA crystals the confidence values for GooF, wR 2 , and R 1 indicates that the refined experimental struc ture has little error relative to the theoretical solved structure. Comparatively, the WB THCA crystal has less overall error relative to t he theoret ical solved structure than the KDPS 18 - 9026 case sample crystal. The Flack and Hooft parameters provide in sight in regards to the decrease fit of the KDPS 18 - 9026 sample to the theoretical structure. The Flack and Hooft parameters provide data r elated to the absolute structure configuration given by refinement. The Flack parameter is generally found between 0 and 1, with values close to 0 indicating correct structural refinement, while values near 1 suggest that an inverted structure is correct. Additional ly, values around 0.5 indicate a racemic or twinned crystal. 23 The Hooft parameter can be thought of as the Flack parameter determined through Bayesian statistics and is used to define the probability of structural accuracy. 24 For case sample KD PS18 - 9026, the Flack and Hooft parameters indicate a racemic crystal wit h two enantiomers possible in the crystal formation. Conversely, the dispensary WB THCA crystal is chiral, which 113 only one enantiomer possible in the crystal formation. Given the racemi c nature o f the KDPS 18 - 9026 crystal, the fit of the experimental struct ure to the theoretical structure was lower than the chiral WB THCA crystal, resulting in slightly worse goodness of fit values. Due to the racemic nature of the crystals from case sam ple 18 - 902 6 versus the chiral nature of the WB THCA crystal sample, smal l differences exist between the length of crystallographic axes (a, b, and c), with a maximum difference between sample values of ~2.8%. The two enantiomers present in the racemic crys tal would effectively elongate a crystallographic axis, as demonstrated by the numerical data provided in Table 3.5 . Additionally, the chiral nature of the WB THCA crystal could indicate a purer sample, as WB THCA Crystals were purchased from a dispensary utilizing professional, regulated manufacturing techniques compared to t he assumed clandestine production of case sample KDPS 18 - 9026. Additionally, it would be more common for crystalline impurities to be present in a clandestine manufactured product, suc h as KDPS 18 - 9026, which may lead to decreased crystal quality and incre ased error during structure refinement. Such impurities can be incorporated into the crystal during the growing phase of crystallization, especially when this phase occurs in an uncont rolled env ironment, or takes place rapidly. The identification of both the KDPS 18 - crystalline components as THCA confirms the presumptive identification by micro - ATR - FTIR. When comparing the micro - ATR - FTIR spectr a obtained for KDPS 18 - 9026 to the structure provided by single - crystal XRD, complementary functional groups are present. Additionally, the structural elucidation by single - crystal XRD provided insight related to the hydrogen bonding responsible for the la ck of char acteristic O - H stretching peaks in the micro - ATR - FTIR spectra. As shown in Figure 3.1 5 , the dimer properties and hydrogen bonding between the two THCA 114 molecules in the unit cell would reduce the strength of the stretching vibration for the hydrox yl groups present on each molecule. Also included in the refined structu re is an alkyl aryl ether, which correlates to the strong peak at ~1250 cm - 1 as characterized by micro - ATR - FTIR. Furthermore, parallels can be drawn between the optical characterizati on describ ed in Chapter 2 and the single - crystal XRD results presented h ere. Optically, the crystal was characterized as fitting in the orthorhombic crystal system, similar to the XRD determination. While the length of crystallographic axes cannot be measu red by opt ical crystallography methods such as polarized light microscop y (PLM), measuring the principle refractive indices of the crystalline structures within BHO provides insight related to the equality of crystalline axis lengths. Given that each refra ctive inde x had a distinct, different value, the length of the crystallo graphic axes too could be inferred as unequal . This inference can be made due to the fact that the relationship between optical axes (refractive indices) and crystallographic axes is d irect, as demonstrated by optical indicatrices. Additionally, the conosc opic characteristics of the case sample crystalline component indicated that the samples were biaxial (containing two optic axes) which relates to the orthorhombic, monoclinic, and tri clinic cry stal systems. Finally, the extinction characteristics observed for the crystals, with extinction occurring every ~90 rotation indicated that the angles between optic axes, and thus the crystallographic axes, were all equal to 90 , paralleling th e results obtained by single - crystal XRD for the crystalline component o f KDPS 18 - 9026 and WB THCA Crystals. Finally, given th e optical and spectroscopic similarities between each of the case samples and Skymint dispensary samples, the crystalline componen ts of each can be identified as THCA. The Cannabidiol Life CBD Crystal sample was identified as CBD through single - crystal XR D analysis. As shown in T able 3.5 , the chemical formula determined for the crystal structures 115 match that of CBD ( C 2 1 H 30 O 2 ) , and th e structures between the solved crystalline component and CBD are analog ous. A representative structure of CBD is provided in Appendix III. The crystal system for the structure was determined to be monoclinic, based on the presence of unequal crystallograp hic axis lengths (a, b, and c) and the presence of two 90 angles ( , and one inclined crystallographic angle ( ). Addition ally, the space group for the crystal was found to be P2 1 , indicating a primitive Bravais lattice (denoted by P) in the monoclinic crystal system. The notation 2 1 refers to the symmetry r elated to a two - fold screw symmetry along the x coordinate direction. Within the unit cell, a total of four CBD molecules are present, denoted by Z in Table 3. 5 , with two independent molecules symmetrically filling the entire unit cell. A packing diagram d epicting the molecules in the unit cell is provided in Appendix III. The confirmed identification of CBD Shatter Crystal samples as CBD coincides with the pres umptive identification by micro - ATR - FTIR. When comparing the micro - ATR - FTIR spectra obtained for the CBD Shatter Crystal dispensary sample to the structure provided by single - crystal XRD, complementary functional groups are present. Furthermore, parallels can be drawn between the optical characterization described in Chapter 2 and the single - crystal XRD results presented here. Optically, the crystal was characterized as fitting in the monoclinic crystal system, similar to the XRD determination. While the le ngth of crystallographic axes cannot be measured by optical crystallography methods such as PLM, measuring the principle refractive indices of the crystalline struct ures within the CBD - containing dispensary samples provides insight related to the inequalit y of crystalline axis lengths. Given that each refractive index had a distinct, difference value , the length of the crystallographic axes too could be inferred to th e unequal. Additionally, the conoscopic characteristics of the case sample crystalline comp onent indicated that the samples were biaxial (containing two optic axes) which relates to the b iaxial 116 orthorhombic, monoclinic, and triclinic crystal systems. Final ly, the extinction characteristics observed for the crystals, with extinction occurring eve ry ~90 rotation, as well as an extinction observed at an inclined angle to the crystal cleavage boundaries, indicated that the angles between optic axes include both a 90 angle and an inclined angle paralleling the results obtained by single - crystal XRD for th e CBD - containing dispensary samples. The structural differences between the THCA - contain ing samples and CBD - containing sample as determined by single - crystal XRD correlate to discrepancies noted both chemically and optically. Table 3. 7 summarizes k ey res ults related to the refined structures and identities of the crystalline samples represent ing the KCSD case samples, Skymint dispensary samples, and Cannabidiol Life dispensary samples. Additionally, Figure 3.1 8 provides a side - by - side comparison of the re fined crystal structures for KDPS 18 - 9026 ( A ) and the CBD Shatter Crystal samples ( B ). As provided by Table 3. 7 and Figure 3.1 8 , the chemical formula and structure for the KDPS 18 - 9026 and WB THCA Crystal samples was determined to be the cannabinoid THCA, while the CBD Shatter Crystal sample was shown to be the cannabinoid CBD. 117 Table 3. 7 Com parison of refined data for crystals analyzed by single - crystal XRD KDPS 18 - 9026 WB THCA Crystals CBD Shatter Crystals Formula C 22 H 30 O 4 C 22 H 30 O 4 C 2 1 H 30 O 2 D calc. / g cm - 3 1.193 1.213 1.110 µ /mm - 1 0.644 0.655 0.535 Formula Weight 358.46 358.46 314.45 Crystal System orthorhombic orthorhombic monoclinic Flack Parameter 0.2(5) - 0.07(7) - 0.02(10) Hooft Parameter 0.3(3) - 0.04(6) 0.01(9) Space Group P 2 1 2 1 2 1 P 2 1 2 1 2 1 P 2 1 a /Å 11.4189(3) 11.40867(12) 10.40257(14) b /Å 18.1043( 6) 17.98950(19) 10.89329(15) c /Å 19.3024(7) 19.1297(2) 16.6836(2) a / ° 90 90 90 b / ° 90 90 95.5081(12) g / ° 90 90 90 V/Å 3 3990.4(2) 3926.11(8) 1881.82(4) Z 8 8 4 Z' 2 2 2 GooF 0.985 1.038 1.024 wR 2 (all data) 0.139 0 0.0834 0.0857 wR 2 0.1098 0.0819 0.0844 R 1 (all data) 0.1434 0.0371 0.0363 R 1 0.0638 0.0341 0.0348 Figure 3. 18 Comparison of single - crystal XRD refined structures for KDPS 18 - 9026 ( A ) and CBD Shatter Crystals ( B ) 118 The confirmatory id entification of the crystalline components by single - crystal XRD, in addition to the spectroscopic analysis highlighted by micro - ATR - FTIR analysis, provides a compreh ensive chemical characterization of a set of cannabis solvent extracts derived from both m arijuana and hemp. Further, the crystals could be optically differentiated, as discussed in Chapter 2 , due to the crystallographic differences highlighted in Table 3. 7 . Most notably, the difference in crystal systems derived from the crystallographic axes and angles measurements are readily differentiable by optical and X - ray crystallography. The combination of optical and spectroscopic characterization of the crystall i ne components provides a method through which cannabis solvent extracts derived from both marijuana and hemp can be preliminarily screened. Noting the distinct optical differences between the THCA and CBD crystals in each subset, as identified by spectros c opic analysis, optical characterization of cannabis extracts by PLM can provide discrimin ation prior to confirmatory identification by either spectroscopic analysis or additional SWGDRUG recommended analytical techniques. 119 APPENDIX 120 Figure A 3. 1 Geometric depictions of Bravais lattices 25 121 Figure A3.2 PPO 14 - 20332 - 10 micro - ATR - FT IR spectrum Figure A3.3 KCSD 14 - 10811 2896 7 micro - ATR - FTIR spectrum -0.1 1 650 1150 1650 2150 2650 3150 3650 Normalized Abs. Wavenumber (cm - 1 ) -0.1 1 650 1150 1650 2150 2650 3150 3650 Normalized Abs. Wavenumber (cm - 1 ) 122 Figure A3.4 KCSD 14 - 10811 2896 0 micro - ATR - FTIR spectrum Figure A3.5 KCSD 14 - 10811 289 6 4 micro - ATR - FTIR spectrum -0.1 1 650 1150 1650 2150 2650 3150 3650 Normalized Abs. Wavenumber (cm - 1 ) -0.1 1 650 1150 1650 2150 2650 3150 3650 Normalized Abs. Wavenumber (cm - 1 ) 123 Figure A3.6 Skymint PB THCA Wax crystalline component s pectrum Figure A3.7 PPO 14 - 20332 - 10 wax component micro - ATR - FTIR spectrum (baseline corrected) 0 0.5 1 650 1150 1650 2150 2650 3150 3650 Normalized Abs. Wavenumber (cm - 1 ) 0 0.5 1 650 1150 1650 2150 2650 3150 3650 Normalized Abs. Wavenumber (cm - 1 ) 124 Figure A3.8 KCSD 14 - 10811 2896 7 wax component micro - ATR - FTI R spectru m (baseline corrected) Figure A3.9 KCSD 14 - 10811 2896 0 wax component micro - ATR - FTIR spectrum (baseline corrected) 0 0.5 1 650 1150 1650 2150 2650 3150 3650 Normalized Abs. Wavenumber (cm - 1 ) 0 0.5 1 650 1150 1650 2150 2650 3150 3650 Normalized Abs. Wavenumber (cm - 1 ) 125 Figure A3.10 KCSD 14 - 10811 2896 4 wax component micro - ATR - FTIR spectrum (baseline corrected) Figure A3.11 KDPS 18 - 9026 XRD pac king dia gram 0 0.5 1 650 1150 1650 2150 2650 3150 3650 Normalized Abs. Wavenumbers (cm - 1 ) 126 Figure A3.12 WB THCA Crystal XRD structure showing chiral centers Fi gure A3.13 WB THCA Crystal XRD structure showing hydrogen bonding 127 Figure A3.14 WB THCA Crystal XRD packing diagram 128 Figure A3.15 CBD Shatter Crystal XRD packing d iagram Figure A3.16. THCA chemical structure 129 Figure A3.17. CBD chemical structu re 130 REFERENCES 131 R EFERENCES (1) L in, H., Zhang, Y., Wang, Q. et al. (2017). Estimation of the age of human bloodstains under the simulated indoor and ou tdoor cr ime scene conditions by ATR - FTIR spectroscopy. Sci Rep 7, 13254 . https://doi.o rg/10.1038 (2) Pablo Prego Meleiro & Carmen García - Ruiz (2016) Spectroscopic techniques for the forensic analysis of textile fibers, Applied Spectroscopy Reviews, 51:4, 278 - 301 (3) Maxwell, V.M. (2016). Forensic E xamination of Trace Evidence. In Forensic Sci ence (eds E. Katz and J. Halámek). (3) Daéid, N.N. (2019). Systematic Drug Identification . Encyclopedia of Analytical Science, Pages 75 - 80 (4) Hughes J, Ayoko G, Collett S, Go lding G (2013) Rapid Quantification of Methamphetamine: Using Attenuated Total Reflect ance Fourier Transform Infrared Spectroscopy (ATR - FTIR) and Chemometrics. PLOS ONE 8(7): e69609. (5) Pereira, L. S.A., Lisboa, F. L.C., Neto , J.C., Valladão, F. N., Sena, M . M. (20 17). Direct classification of new psychoactive substances in seized blotter pa pers by ATR - FTIR and multivariate discriminant analysis. Microchemical Journal. Volume 133, Pages 96 - 103. (6) González M. Application of Chem ometric Tools on Cannabis Sample s Analyz ed by the FTIR - ATR Method . Brazilian Journal of Forensic Sciences 2020;9. (7) B ruker Corporation . Differentiation of THC and CBD cannabis using FTIR. Application Note. (Date Accessed 11/23/2020) (8) Mainali D. Quick and Real - Time Potency Determination of Canna binoids Using FTIR Spectroscopy. 2020; (9) ssowski, T., Trynda, A., Sikorski, A. Single - crystal X - ray diffraction analysis of designer drugs: Hydrochlorides of metaphedrone and pent edrone. Forensic Science Interna tional. Volume 232, Issues 1 3. 2013. (10) Nycz, J. E ., Pazdziorek, T., Malecki, G., Szala, M. Identification and derivatization of selected cathinones by spectroscopic studies. Forensic Science International. Volume 266. 2016. Pages 416 - 426, (11) Carlton R.A. In frared a nd Raman Microscopy. Pharmaceutical Micro scopy. 2011 . 132 (12) Attenuated total reflectance (ATR) . Anton Paar Wiki. https://wiki.anton - paar.com/en/attenuated - total - reflectance - atr/#:~:text=Attenuated total reflectance or ATR ,where IR light is reflected. (a ccessed Nov 24, 2020). (13) Griffiths P.R., de Haseth J. A. (2007). Fourier Transform Infrared Spectrometry. New York, USA: John Wiley & Sons. (14) Pavan M. V. Barron , R., Barron, A. R., Physical Methods in Chemistry and Nano Scienc e . Chapter 8.1., 2019 (15) Pavan M. V. Barro n , R., Barron, A. R., Physical Methods in Chemistry and Nano Science . Chapter 8.3.4, 2019 (16) Putnis, A. (1992). Introduction to Mineral Sciences. Cambridge, UK: Cambridge University Press. Chapter 3 (pp. 41 - 80). (17) Cowtan, K. Phase Problem in X - ray Cry stallogr aphy, and Its Solution. Encyclopedia of L ife Sciences. 2001. (18) Pretsch, E., Bühlmann, P., Affolter, C. Structure Determination of Organic Compounds. Springer - Verlag Berlin Heidelberg. Edition 3. 2000. (19) Dolomanov, O.V., Bourhis, L.J., Gildea, R.J., H oward, J .A.K. and Puschmann, H. (2009) OLEX2: A C omplete Structure Solution, Refinement and Analysis Program. Journal of Applied Crystallography, 42, 339 - 341. (20) Sheldrick, G.M. (2015) Crystal Structure Refinement with SH ELXL. Acta Crystallographica C, C71, 3 - 8. (21) Th e CIF file, refinement details and valida tion of the structure. CCCW 2011. (22) - 88 1. 1983 (23) Parsons S, Flack HD, Wagner T. Use of intensity qu otients and differences in absolute s tructure refinement. Acta Crystallogr B Struct Sc i Cryst Eng Mater. 2013 Jun;69(Pt 3):249 - 59. doi: 10.1107/S2052519213010014. Epub 2013 May 17. PMID: 23719469; PMCID: PMC3661305. (24) Zhang, Tao & Li, Ling & Yang, Haizhao. (2018). 3D$ Crystal Image Analysis ba sed on F ast Synchrosqueezed Transforms. 133 4 . OPT IMIZATION OF THCA DERIVATIZATION USING AN EXPERIMENTAL DESIGN APPROACH The necessity for forensic laboratories to not only identify cannabinoids but quantify their potency in marijuana and hemp - based produ cts was prompted by the passing of public law num ber 115 - 334 from the Agricultural Improvement Act of 2018 (commonly known as the Farm Bi ll). 1 Essentially, the passing of the Farm Bill require s differentiation of submitted marijuana samples from hemp based on tetr ahydrocannabinol ( THC ) concentration . Thi s requirement ultimately forced the development of validated methods to q uantify THC in samples. Quantification can be performed two ways: through the analysis of both cannabinoid acids and neutrals separate ly, or t Total THC potency analysis combines the concentration of psycho active THC as well as the non - psychoactive tetrahydrocannabinolic acid (THCA). This method of quantification takes into consideration the fact that THCA is converted to THC through de carboxylation , which occurs naturally over time in marijuana but can also be accelerated by heat. While each cannabinoid acid undergoes decarboxylation to the neutral form over time or upon heating , the focu s in thi s work is the decarboxylation of THCA to THC. As the requirements for cannabinoid quantification are set by indiv idual states, it may be necessary to identify the entire profile of cannabinoid acids and neutrals. An analytical scheme commonly used for the analysis of marijuana is gas chromatogra phy - mass spectrometry (GC - MS), due to the fact that it satisfies analysis recommendations provided by the Scientific Working Group for the Analysis of Seized Drugs (SWGDRUG). However, the use of GC - MS for th e analys is of both cannabinoid acids and neutrals poses a challenge directly related to the aforementioned decarboxylation process . As the injection port of the GC - MS is heated at a 134 temperature high enough to volatilize samples prior to analysis, this heat readily decarboxylates the sample, resulting in the conversion of all present cannabinoid acids to their neutral counterp arts ( e.g. , THCA to THC). This conversion occurs with a theoretical maximum efficiency of 87.7% conversion, given the ratio of masses between THCA and THC, but literature indicates 65 % maximum conversion using the heat of the GC - MS injection port. 2 To pro vide accurate identification and quantification of both the acidic and neutral cannabinoids, recent research has been performed focusi ng on in strumentation other than GC - MS to avoid t he potential for decarboxylation ( e.g. , liquid chromatography - mass spectr ometry) . 2 - 4 However, due to the prevalence of GC - MS in forensic laboratories, it is advantageous to develop methods through which the structur al integrity of the cannabinoid acids can be preserved for identification and quantification. As such, the derivat ization of cannabinoid acids prior to GC - MS analysis has been investigated in order to produce a stable product that does not undergo decarbox ylation upon injection. 5 - 7 Derivatizatio n is a chemical process through which a thermally unstable, or low volatility compound can be modified with a variety of functional groups to produce a stable structure suitable for analytical analysis by GC - MS. Rec ent literature has highlighted the use of derivatization for the identification and separation of numerous n ovel psychoactive substances by GC - MS analysis. 8 Additionally, derivatization allows for more sensitive detection and accuracy during GC - MS analysis . 9 While r ecent literature has focused on the method development and validation of various forms of derivati zation specifically for cannabinoid acids , the published methods vary. 5 - 7 Derivatization methods for GC - MS analysis consider multiple facto rs durin g their development, including derivatizi ng agent, support solvents, temperature of reaction, total derivati zation time, 135 and use of a catalyst. 10 Optimization of these parameters, with particular focus on the derivatization of THCA, has not been pe rformed. Due to the unknown interactions between each of the aforementioned derivatization parameters, experimental design procedures can be utilized for method optimization rather than optimizing one parameter at a time. Additionally, recent literature ha s demons trated that experimental design methods e ffectively reduce the overall number of experiments necessary to op timize a set of parameters . Further, these designs provid e statistically relevant data regarding the maxim um yield of the desired derivatiz ed produ ct relative to the interaction of paramet ers. 10 - 13 4.1 THEORY 4 .1.1 Derivatization Methods Derivatization of non - volatile or thermally unstable compounds can be performed using a variety of methods. As derivatization relies on the addition of fu nctional groups to the original compound in order to increase volatility and stability, the derivative functional group selected reagents can be used depending on th e overal l goal of derivatization in the specific analysis ( i.e. , stability, enhanced separation, or ch iral separation). 9 As such, considerations regarding the derivatizing reagent rely both on understanding the structure of the compound of interest as well as the goal for the analysis. Some common functi onal groups used for addition reactions for derivatiz ation include alkyl or aryl, silyl, acyl, and hetero groups. 9 Silylation is a commonly used derivatization method for GC - MS analysis. The addition of si lyl grou ps to a compound increases the stability while reducing the polarity, thus improving the volat ility and stability of the compound during GC - MS analysis. 9 Trimethylsilyl (TMS) 136 reagents produce thermally stable compounds by reacting with active hydro gen atom s. The general chemical reaction for a TM S derivatization is where X methyl groups, and the Sample - OH can be any variety of active hydrogen functional groups (and can be interchanged with Sample - NH). 14 The rate of silylation for unhindered fun ctional groups regardless of silylating reagent is as follows: alcohol > phenols > carboxylic acids > amines > amides. This ca n be fur ther refined relative to alcohol and amin e location on the structure. Example structures of select TMS derivatives are pro vided in Figure 4.1 . Figure 4.1 Example TMS derivatives: (A)hydroxyl, (B) carboxyl, and (C) amide N - Methyl - N - (trimethyls ilyl)trif luoroacetamide (MSTFA) and N,O - Bis(trime thylsilyl)trifluoroacetamide (BSTFA) are two commonly used silylation reagents. B oth ( A ) ( B ) ( C ) 137 reagents react with most alcohols, phenols, and carboxylic acids; however, BSTFA donates the TMS group more readily to hyd roxyl gro ups than MSTFA does. 14 The addition of l ow concentration (1%) trimethylchlorosilane (TMCS) to each silylating reagent imp roves the silylation of hindered functional groups. Derivatization methods vary related to sample size but are generally simp le. Due t o the reactivity of silylating reagents with hydroxyl groups, reactions are often performed in anhydrous environments to prevent competitive reactions and the generation of side products. 13 Additionally, while the silyl reagent may serve as the so lvent, a support solvent may also be necessary fo r complete derivatization. Polar support solvents such as pyridine, dimethylforma mide (DMF), and acetonitrile are often used because they can facilitate the silylation reaction. Further, pyridine is commonly used as a support solvent due to the fact that i t can act as a catalyst for the reaction as a HCl scavenger, as HCl is a byproduc t of the silylation reaction in organosilane reactions. 14 The factors that control the rate and extent of a silylation react ion inclu de silylating reagent, catalyst, use of support solvent, temperature of the reaction, and the time for the reaction. Heat ing may be necessary for some derivatization reactions if the reaction occurs slowly at room temperature. Determining the opti mal param eters necessary for efficient and reprod ucible derivatization can be performed by varying one variable at a time while ma intaining the additional variables at fixed values. While this approach is traditionally used, it can result in a large number of accep table parameters for the derivatization of one compound, due to a reduced number of overall factors taken into considerat ion and the inability to study the influence of multiple factor interactions. Experimental design methods, however, allow for the obser vation of 138 the effect of multiple changin g factors and interactions at the same time, reducing the overall number of exper iments necessary for optimization of the derivatization method. 4.1.2 Experimental Design Experimental design procedures are often employed to determine the statistical s ignificance of factors in a reaction, method, or procedure. Experimental designs are useful in not only identifying significant factors for a procedure, but also can optimize those factors while improving th e overall robustness of the method. 10 The optimiz ation of a method with multiple factors can be achieved more efficiently through experimental design approaches when compared to traditional single - factor analyses . Considerations that are made when perfor ming expe rimental design experiments include the number of factors studied, the range over which the factors will be adjusted (referred to as levels), and the appropriate response variable. An example of an experimental design would be a three - factor, two - level des ign, where pre - determined high and low l evels of each factor are used during experimentation. For ex ample, a derivatization procedure can be investigated by experimental design by comparing the effect of reaction temperature, derivatization time, and deriv atization reagent to solvent ratio at hi gh and low levels on the abundance of the resultant derivati zed product (the response variable). The derivatization experiments can be planned and randomized such that experimenter bias is removed. Additiona lly, if t he experiments cannot all be performed o n the same day, due to length of analysis or other variables , the design can be outlined in the experimental des ign, but also differences between days in the dat a analysis (allowing for temperature and humidity of the lab oratory to be taken into account). 139 Introducing blocks into the derivatization experimental designs inherently introduces confounding of factor an d interac tion impact on the response variable (de rivatized product abundance). Confounding can also occur in designs that are not blocked, based on the number of total experiments performed and the ability to separate the effects of the factors and interac tions on the response variable during data analys is. Confounding prevents complete understanding of the signi ficance of separate factors and interactions. 15 When two variables are confounded, such as the reaction temperature factor and the interaction betw een react ion temperature and reaction time, the e ffect they have on the derivatized product abundance cannot differentiated. In this case of confounding, an increase in derivatized product abundance cannot be attributed to only the reaction temperature nor the inte raction between temperature and time, bu t to both the factor and the interaction, effectively reduci ng the ability to fully characterize the reaction. As such, confounding limits the overall usefulness of the experimental design data. 16 The extend of confo unding in an experimental design can be described by the design resolution. Resolution of an experim ental design is generally portrayed by considering the lowest order of interactions (using letters to represent the factors) in the defining relati on of the design. 16 Defining relations are create d in fractional factorial or blocked full factorial designs to show the combination constant factors capable of defining the experimental design. The defining relation can be used to determine which factors are confo unded by multiplying the factor or inter action of interest by the defining relation. In order to sep arate the main factor effects from each other (necessary for any useful experimental design), the defining relationship must be represented as I = ABC. This defining relation indicates a resolutio n III design, related to the minimum number of factors in th e defining relation. This method for determining resolutions for experimental designs gives rise to three general resolutions - III, IV, and V. Whi le resolu tion II is technically 140 possible, this wo uld indicate that the main factors effects are confounded wi th each other (I = AB) and would not be useful in characterizing the procedure. As resolution number increases, the confounding between factors an d interac tions decreases such that Resolution III provides the least design characterization and Resolution V provides the most. Table 4.1 summarizes the confounding of a design present at each resolution level. Table 4.1 Summary of Resolution and Confoun ding Vari ables Confounding Resolution III Resolut ion IV Resolution V Main effects with each other - - - Main effects are with two - factor interaction X - - Main effects with three - factor interactions N/A X - Two - factor interactions with each other N/A X - Two - f actor interactions with three - factor int eractions N/A N/A X 4.1.2.1 Full Factorial Screening Design Screening designs are often utilized as the first round of an experimental design procedure and provide the opportunity to determine significan t factors in a method or procedure. There are a v ariety of screening designs available, includi ng full factorial, fractional factorial, and Plackett - Burman designs. 17 Full factorial designs provide the most insight related to parameter and interaction effe cts on th e response variable with no confounding present (unless 141 blocked experiments are perfor med), while each of the latter designs confound the parameters and interactions to some degree. Full factorial screening designs limit the number of experiments being pe rformed based on Equation 4.1: E=K N (4. 1) In Equation 4.1, E represents the total num ber of experiments, K is the number of level, and N represents the number of factors. For example, the two - level, three - factor experimental design to optimize a derivatiz ation procedure requires eight experimen ts to complete the full factorial. The design of such a full factorial experiment would follow Table 4.2 , where +1 represents the high value for a level and - 1 represents the low value for a level. Table 4 .2 Exampl e of Experimental Order for Full Factori al Design with Three Factors Experiment Order Factor A Factor B Factor C 1 +1 +1 +1 2 +1 +1 - 1 3 +1 - 1 +1 4 - 1 +1 +1 5 +1 - 1 - 1 6 - 1 +1 - 1 7 - 1 - 1 +1 8 - 1 - 1 - 1 The positive and negative level v alues in Table 4 .2 represent pre - determined high and low experimental values based on knowledge of the reaction or method being studied. For 142 example, in the context of this work, factor A represents the time of a reaction, with +1 equal to 60 minutes, whil e - 1 is e qual to 10 minutes. Factor B represents the temperature of a reaction with a high value of 100 ° C and a low value of 30 ° C. Finally, factor C is the ratio of a solution with a high value of 90:10 and a low value of 50:50. Following the experimenta l order i n Table 4 .2 , experiment four would be pe rformed for 10 minutes at 100 ° C with a solution ratio of 90:10. Following the c ompletion of the full factorial screening design experiments, linear regression mathematical models are used to determine the sources o f significant variation in the response variable. When analyzing the response of one variable against multiple factors, a nalysis of variance (ANOVA) can be utilized to determine statistically significant factors and interactions. Statistical analy sis by AN OVA determines significance of factors b y comparing variation of the response variable due to random error against the va riation caused by changing an experimental factor. ) is calculated between each experimenta l factor mean and the overall mean. Additionally, the SS is calculated for the er ror and for the design as a whole. Next, the total degrees of freedom are calculated as well as degrees of freedom for each fa ctor and error. Following this, the mean sum of s quares (MS) are calculated to measure the variation explained by the factor s and model assuming that all other factors are in the model. Finally, an F - value is calculated for each test, which is further us ed to cal culate the p - value used to determine sta tistical significance. If the p - value for the model is less than the signific ance level set for analysis, then the model explains the variation in the response variable. Similarly, if the p - value for a parti cular fac tor studied in the design is less than t he significance level it causes significant variability in the response varia ble. The equations used for ANOVA calculations are provided in Appendix IV. 143 4.1.3 Gas Chromatography - Mass Spectrometry GC - MS i s a hyphe nated, two - part technique which combines chromatographic separation (GC) with spectrometric analysis (MS) for the separation, characterization, and identification of volatile samples. The general components of a GC include a heated injectio n port, a capill ary chromatography column, and an oven w hich houses the column. Following injection via the GC injection port and volatilization of the sample, an inert carrier gas (such as He, H 2 , or N 2 ) moves the samples through the GC column at a pre - de termine d flow ra te ( e . g. , 1 mL/min). As sample mixture m ove through the column, the individual analytes are separated by both their boiling points and chemical interaction with the stationary phase, resulting in more volatile analytes travelling more quick ly thro ugh the c olumn than less volatile analytes. The e nd of the GC column is connected to a detector (in this case a MS) which distinguishes the components of the mixture from each other. The total time taken for the analyte to travel through the column to the detector is the retention time. 18 A mass spectrom eter functions as the detector in GC - MS. The MS and GC are connected by a heated transfer line such that the volatile compounds leaving the GC do not condense prior to MS analysis. Mass spectrometers general ly consis t of three components: an ion source, a mass analyzer, and an ion detector. Though there are many variations of mass spectrometers on the market, i n benchtop instruments that are common in forensic lab oratorie s, an electron ionization sourc e, sing le - quadru pole mass analyzer, and electron multipl ier detector are used . As mass spectrometers identify and separate charged fragments, ionization must occur prior to detection. In EI, a heated filament emits high - energy electrons (commonly 70 eV) th at bomb ard the g as - phase analyte exiting the GC column. The use of 70 eV for bombardment provides sufficient energy to break bonds in organic molecules and ensures reproducible compound fragmentation. 144 The interaction between the electrons in the EI source and the analyte causes fragmentation while also inducing a net positive charge due to the loss of electrons during fragmentation. From this, positively charged fragments continue through the mass spectrometer, while neutral loss fragments are not detected. Next, the charg ed fragments are separated using a quadr upole mass analyzer, which consists of four parallel molybdenum rods with oscillating radio frequency (RF) and DC voltage. The oscillation of voltages between the metal rods provides the ability to se parate ions acco rding to their mass over charge ( m/z ) ra tios. 18 The path of a charged fragment through the quadrupole is determine by the sweeping of RF and DC voltages. Specific RF and DC voltages are used to separate the fragment ions because ions have s table o scillatio ns only within particular values of volt ages applied. For example, if the voltages currently applied to the quadrupole allow an ion with an m/ z equal to 55 to oscillate through to the detector, an ion with an m/z of 155 would not be stable and its oscillat ing path would cause the ion to be neutr alized by the quadrupole and pumped away by the vacuum system. The ions separated by the mass analyzer are then detected by an electron multiplier. Ions are detected by th e electron multiplier through a sec ondary em ission process, in which the detected io n produces a cascade of secondary electrons that are collected by a metal anode and converted to a computable signal. The separation of m/z values and their detection by an electron multiplier allows for the creation of a mass spectrum providing the m/z va lues and their associated intensities. 18 The resultant mass spectrum can then be analyzed for abundant fragments and compared to commercially available library spectra such as the National Institute of Stand ards and Technology (NIST) EI mass spectral libra ry. Further, more confident 145 identification of the analyte is performed by comparing the resultant mass spectrum to that of a certified reference material analyzed under sim ilar conditions as recommended by S WGDRUG. 4.2 MATERIALS AND METHODS 4.2.1 Referen ce Materials and Sample Preparation Pre - screening and full factorial screening design experiments were performed using THCA certified reference material (CRM) from Cayman Chemical ( Ann Arbor, MI). Solvents used thro ughout the pre - screening and full factor ial screening design experiments included ethyl acetate (99.7% pure, HPLC grade, Millipore Sigma, Burlington, MA ), pyridine (>99.9% pure, ACS grade, Sigma - Aldrich, St. Loui s, MO ), and BSTFA - 1% TMCS (Supelco Analytica l, Bellefonte, PA). The THCA CRM was pre pared at 0.1 mg/mL in methanol (ACS grade, VWR Chemicals BDH®, Radnor, PA - Aldrich) which was prepared at 0.9 mg/mL in both ethyl acetate and pyr idine sol utions and added to the samples followin g derivatization at a final concentration of 0.3 mg/mL. All THCA CRM samples were prepared for derivatization by pipetting 150 µL of the 0.1mg/mL THCA into a GC - MS vial fi t with a 250 µL glass insert (Agile nt Techno logies, Santa Clara, CA). The THCA CRM s olution was dried down under house nitrogen to constant mass to create an anhydrous environment for the derivatization reaction. A general derivatization method (provid ed by Restek) 5 was utilized and acted as the f ramework by which the full factorial exp eriment derivatizations were modified. The dried THCA CRM was reconstituted with 50 µL ethyl acetate (Millipore Sigma) and 50 µL BSTFA - 1% TMCS (Supelco Analytical). The s ample was heated for 30 min at 70 °C, then all owed to 146 cool to room temperature prior t o GC - MS analysis. Progesterone i nternal standard (0.9mg/mL) was added to the sample before GC - MS analysis in an additional 50 µL aliquot of ethyl acetate. Initial experiments were performed prior to the ful l factori al screening design using the Restek der ivatization method. 5 These experiments investigated appropriate support solvents for the derivatization as w ell as derivatized sample hold conditions prior to GC - MS analysis. A summary of the pre - screening e xperiment s is provided in the appendix. From thes e experiments, pyridine was selected as an additional support solvent based on the abundance of derivatized THCA (THCA - 2TMS) and reproducibility between and within derivatizations. Further, refrigeration was selected as a suitable sample hold environment b ased on reproducibility and feasibility. Moving forward, both ethyl acetate and pyridine were considered as support solvent through the experimental design. 4 . 2 . 2 Full Factorial Screening Experiments Focu sing only on the derivatization of a 0.1mg/mL THC A CRM sample, the silylation reaction was characterized for significant factors. Specifically, the factors i nvestigated for optimization included temperature of reaction, total reaction time at temperature, and the r atio between BSTFA - TMCS reagent and the secondary solvent (either pyridine or ethyl acetate). Full factorial analysis was selected as both main fact or and interactions between factors could be investigated for their significance on the overall abu ndance of the THCA - 2TMS product. The high and lo w levels for the full factorial parameters were selected as the extreme values at which the derivatization w ould move forward. The levels for the temperature of reaction were chosen considering the necessity of heat for the derivatization of THCA (based on 147 prior experimentation), as well as the thermal instability of THCA. As such, the lowest temperature of 30 ° C was selected to provide the lowest, controllable temperature comparable to room temperature. The highest t emperature was selected as 100 ° C to min imize sample decarboxylation. 2 The total derivatization time was selected based on previous experimentation and with regard to the thermal instability of THCA over time. Prior experimentation using the Reste k method indicated that immediate derivatization (less than 5 minutes total reaction time prior to sample injection) produced little to no derivatized sample and with low reproducibility. Noting that, 10 minutes was used as the shortest reaction time to pr ovide rep roducible reactions using both pyridine and ethyl acetate as support solvents, while 60 minutes ensured little decarboxylation, even in combination with the high temperature level (100 ° C). Finally, the levels used for solvent ratio were selected with the knowledge that excess BSTFA - TMCS derivat izing reagent is necessary for full, reproducible derivatizations, as well as the necessity for a support so lvent for the formation of the THCA - 2TMS product. Table 4.3 summarizes the high, low, and center v alues for these factors. Table 4.3 Full Factorial Levels for Each Factor Time (min) 10 60 35 Temperature ( ) 30 100 65 BSTFA : Solvent 50 : 50 90 : 10 70 : 30 The experimental order was designed and randomized using Minitab (Minitab, LLC, State College, Pennsylvania). A total of 24 experiments were designed in which the support solvent was pyridine, allow ing for the entire full factorial design to be repeated tw ice. These experiments were broken up into four blocks of six experiments each comprising of four different factorial experiments and two center point experiments per block. Blocks were 148 analyzed on sequential days. This design resulted in a resolution IV e xperimental design, capable of providing sufficient characterization information for the experimental protocol. A full table of design order and factor values can be found in Section 4.4.1 . An addi tional full factorial screening design was performed in wh ich ethyl acetate was used as the support solvent. This was completed in 12 total experiments using the same levels and factors as outlined in Table 4.3 . The experiment was broken into two blocks of 6 experiments each comprising of four different factorial experiments and two center point experiments per block. The experiments performed mimic blocks one and two of the pyridine support solvent full factorial analysis. Following full factorial analysis , the derivatization method using pyridine as the support solvent was optimized to provide the following method: The dried THCA CRM was reconstituted with 50 µL pyridine ( Sigma - Aldrich ) and 50 µL BSTFA - 1% TMCS (Supelco Analytical). The sample was heated fo r 10 min at 65 °C, then allowed to cool in the refrigerato r prior to GC - MS analysis. Progesterone i nternal standard (0.9mg/mL) was added to the sample before GC - MS analysis in an additional 50 µL aliquot of pyridine . 4.2.2.1 Sample Preparation - THCA Con centration Study A range of THCA concentrations was deriv atized using the optimized method in order to demonstrate method linearity. The concentrations included in this study were: 10 µ g/mL, 50 µ g/mL, 0.1 mg/mL, 0.25 mg/mL, and 0.5 mg/mL. Samples at each concentration level were derivatized using the optimized m ethod as determined through the full factorial screening 149 experiments in triplicate vial to assess the reproducibility of the method both between and within sample vials in terms of RSD. 4.2.2.2 Sa mple Preparation Cannabis Extract Sample Analysis A sub set of three cannabis solvent extract samples was subjected to derivatization by the optimized method and analysis by GC - MS. These samples included two THCA - containing samples, case sample KDPS 18 - 9 026 and Skymint sample WB THCA crystals. Additionally, the CBD Crystal sample was derivatized using the optimized method to show the suitability of this derivatization method for cannabinoids other than THCA. For each solvent extract sample, a 1mg/mL solut ion of crystalline material was prepared in methanol and t he sample was diluted by one - third for derivatization (50 µL evaporated to dryness as opposed to 150 µL ). Each sample was derivatized in triplicate to evaluate the reproducibility of the derivatizat ion given any matrix effects present in the samples (additional cannabinoids or wax component). 4.2.3 GC - MS Analysis All derivatized samples were analyzed b y GC - MS using an Agilent Technologies 7890A GC and 5975C MS equipped with an autosampler (all A gilent Technologies). The GC injection port was set at 280 °C and a 1 µL aliquot of sample was injected with a 25:1 split ratio. Ultra - high purity helium (Air Ga s, Radnor, PA) was used as the carrier gas at a nominal flow rate of /min. The GC was equ ipped with a 5% diphenyl - 95% dimethyl polysiloxane column (30 m x 0.25 mm x 0.25 µm, Agilent Technologies ). n ramped to 280 °C at 12 °C / 150 was operated 7 50 amu, and a scan rate o f 2.00 scans/sec. The ion source and quadrupole were heated to 230 °C and 150 °C, respectively . Each sample was analyzed in triplicate. 4.2. 4 Data Processing and Analysis The chromatographic abundances for each derivatized sample were collected and exported from ChemStation ( version E.01.00.237, Agilent Technologies) to Microsoft Excel (version 16.0, Microsoft Corporation, Redmond, WA). The chromatographic abundan ces were normalized to the abundance of the progesterone internal standard. Additionally, averaged chromatograms were produced by averaging the normalized chroma tographic abundances to accurately represent chromatographic abundances of derivatized products within one derivatized sample as well as between triplicate derivatizations. The reproducibility of peaks was evaluated based on the relative standard deviation (RSD) of the peak abundance of the derivatized product (THCA - 2TMS or THC - TMS) within and betwe en derivatizations. 4.2. 4 .1 Full Factorial Analysis All full factorial analysis was performed using Minitab software ( Minitab® , version 19.2020.1 (64 - bit) , State College, PA ) The averaged, normalized abundance of the THCA - 2TMS chromatographic peak fo r each experiment was input relative to the experiment number for each screening design. The full factorial analysis function was utilized, with a two - sided confidence level set to 95. The Minitab software provides multiple tables and plots for additional analysis, including the coded coefficients table, ANOVA table, Pareto chart of standardized effects, and 151 multiple residual plots. Significan t factors were determined based on the ANOVA output, while the additional table and plots were used for supporting i nformation. 4. 3 FULL FACTORIAL SCREENING EXPERIMENT RESULTS AND DISCUSSION Using the derivatization method outlined in Section 4.2.1. 3, the THCA CRM was derivatized, using both ethyl acetate and pyridine as support solvents, and characterized by GC - MS. Figure 4.2 provides an overlay of averaged chromatograms from both the ethyl acetate and pyridine derivatizations. Derivatizations were performed in triplicate using ethyl acetate and pyridine as support solvents. Figure 4. 2 Averaged chromatograms o f THCA derivatizations using ethyl acetate and pyridine The peaks of interest regarding the derivatization of THCA are between retention times ( t R ) 21 and 24 min. Derivatizations performed using pyridine or ethyl acetate as the support solvent displayed s imilar chromatographic peaks ( t R = 21.6 min and t R = 23.0 min), however 152 ethyl acetate - supported derivatizations also include a t hird peak ( t R = 22.4 min). The peak present at t R = 21.6 min is the THCA - 2TMS product while the peak at t R = 22.4 min is consis t ent with a THC - TMS product (derivatized THC, not THCA). Finally, the progesterone internal standard was present at t R = 23.0 min. The early eluting peaks in the chromatogram (between t R 4.5 9.5 min) are BSTFA - TMCS solvent fragments, including single TMS groups and those bound to alkyl groups not associated with the THCA - 2TMS ( t R = 21.6 min), as well as various siloxane peaks from the GC - MS inlet and stationary phase. Peaks at the listed retention times were identified using their associated mass spectrum . The mass spectra and chemical structure for THCA - 2TMS ( t R = 21.6 min) is provided in Figure 4. 3 . Though mass spectra l library searches are generally used to presumptively identify an unknown analyte, the mass spectrum of THCA - 2TMS was not included in the NIST mass spectral library used in this work. As such, identification of the THCA - 2TMS product peak was accomplished through structural elucidation based on fragment ions present in the mass spectrum. The molecular ion peak at m/z 502.3 for THCA - 2TMS is p r esent at low abundance in the mass spectrum. The accompanying base peak at m/z 487.4 represents the loss of one methy l group from the silylated compound. The THCA - 2TMS mass spectrum does not display many high abundance fragment ions, which is expected due to the stability of the compound following derivatization. The next highest abundance ion is m/z 73.1 which is repres entative of a single TMS group (molecular weight = 73.2). The presence of this peak is indicative of a derivatized sample but cannot be us e d toward the identification of the cannabinoid structure, as it is not . 153 Figure 4. 3 Mass spectrum and chemical structure of THCA - 2TMS The mass spectrum of THC - TMS ( t R = 22.4 min) and chemical structure are provided by Figure 4. 4 . Presumptive identification was performed using a mass spectral library search. The experiment al mass spectrum for THC - TMS includes high abundance background ions in the range m/z 150 - 200, which are not present in the NIST library spectrum (Appendix IV). However, these background ions are instrument - rather than compound - specific and correspond t o polysiloxanes commonly found in the column stationary phase or injection port septum. In spite of the high abundance background, identif i cation of the peak as THC - TMS was possible due to ions characteristic of the derivatized product, including the high abundance molecular ion ( m/z 386.3). Additionally, the high abundance fragment at m/z 371.3 represents the loss of a methyl group from th e derivatized structure. Finally, the characteristic single TMS fragment peak is present at m/z 73.1, indicating a derivatized sample. 154 Figure 4. 4 Mass spectrum and chemical structure of THC - TMS Each of these high abundance peaks a re present in the li br ary mass spectrum; however, lower abundance peaks from the library mass spectrum are not present in the experimental spectrum. The lack of lower abundance ions as well as the inclusion of high abundance background ions correspond to t he low chromatograph ic abundance of the THC - TMS peak. Despite these features, this peak can be presumptively identified as THC - TMS. A representative mass spectrum for THC (underivatized) is also provided in Appendix IV. The presence of underivatized THCA in the form of THC - T MS in the ethyl acetate derivatizations indicates an inefficient derivatization method. This product may be produced by the decarboxylation of remaining THCA in the injection port of the GC - MS and immediate derivatization with remaining BSTFA - TMCS, though th is was not specifically studied in this work. Additionally, the formation of this derivative may provide insight related to preferred derivatization of the THCA molecule, given that BSTFA - TMCS is known to react with hydroxyl groups mo re readily than carb on yl groups. This additional peak is not present in any 155 derivatization using pyridine as the secondary solvent, possibly indicating full derivatization of the THCA during the reaction time in the oven. 4.3.1 Full Factorial Screening Design Experiments a nd Optimization A duplicated full factorial screening design was used to characterize the significant parameters of THCA derivatization using pyridine as the support solvent. This resulted in a total of 24 experiments, including center point and factorial ex periments. Ethyl acetate supported derivatizations were also characterized using an unduplicated full factorial design, consisting of 12 experiments which followed the order of the first two blocks of the pyridine study. Following eac h derivatization, th e samples were analyzed by GC - MS and the peak abundance for the THCA - 2TMS peak was averaged across triplicate instrument injection for each derivatized sample. The full list of experiments and averaged THCA - 2TMS normalized peak abundanc es and RSDs are summ ar ized in Table 4.6 . 156 Table 4.6 Summary of experiment order, levels, averaged THCA - 2TMS abundance, and RSD Factors Pyridine Ethyl Acetate Order Block Time (min.) Temp. (°C) BSTFA: Solvent Abundance (normalized) % RSD Abundance (normalized) % RSD 1 1 60 30 90 :10 0.6844 5.3 0.2374 4.4 2 1 10 100 90 :10 0.7403 4.8 0.5229 5.6 3 * 1 35 65 70 :30 0.8219 2.5 0.4428 7.3 4 * 1 35 65 70 :30 0.6883 2.9 0.4400 5.2 5 1 60 100 50 :50 0.7394 1.8 0.7328 2.5 6 1 10 30 50 :50 0.2849 2.5 0.1527 4.5 7 4 10 100 50 :50 0 .6 906 2.8 0.6846 1.3 8 * 4 35 65 70 :30 0.5802 5.7 0.6807 3.2 9 4 10 30 90 :10 0.5275 7.1 0.2732 16.1 10 4 60 30 50 :50 0.5977 2.6 0.5154 9.4 11 * 4 35 65 70 :30 0.6698 4.3 0.5758 5.7 12 4 60 100 90 :10 0.7986 7.9 0.8035 4.5 13 2 60 30 50 :50 0.5993 1.0 - - 1 4 2 60 100 90 :10 0.8053 2.0 - - 15 * 2 35 65 70 :30 0.7568 1.3 - - 16 * 2 35 65 70 :30 0.6893 7.0 - - 17 2 10 30 90 :10 0.6296 10.1 - - 18 2 10 100 50 :50 0.8598 6.9 - - 19 3 10 30 50 :50 0.6580 5.9 - - 20 * 3 35 65 70 :30 0.7902 5.4 - - 21 3 10 100 90 :10 0. 8008 6.1 - - 22 3 60 100 50 :50 0.8114 6.8 - - 23 * 3 35 65 70 :30 0.6434 3.6 - - 24 3 60 30 90 :10 0.7138 1.2 - - * denotes a center point analysis Both the pyridine and ethyl acetate supported derivatizations provided reproducible THCA - 2TMS production , with nearly all derivatizations resulting in THCA - 2TMS peak abundances under 10% RSD. The outliers from this trend included experiment number 17 for t he pyridine supported derivatizations and experiment number 9 for ethyl acetate. Both of these experimen ts were performed under similar conditions, with a 10 - minute reaction time, at 30 ° C, and with a 90:10 reagent:solvent ratio. Despite the similar reacti on conditions, the overall 157 abundance of THCA - 2TMS in the pyridine experiment was significantly higher th an the THCA - 2TMS abundance in the ethyl acetate experiment ( = 0.05, p = 0.003). The increased THCA - 2TMS abundance for pyridine supported derivatizatio ns as compared to ethyl acetate holds true for a majority of the experiments, with the exception of expe ri ment numbers 8 and 12. Statistical analysis of means ( = 0.05), however, concluded that the average THCA - 2TMS abundances from either experiment were not statistically different ( p = 0.103, p = 0.950). Figure 4.5 provides an overlay of the representativ e chromatograms from derivatizations using the lowest, highest, and center point levels for each of the three factors considered for the pyridine design [ e.g. , low indicates 10 - min reaction time, 30 ° C reaction temp erature, and a 50:50 (BSTFA : solvent) ra ti o]. These chromatograms were averaged between the two replicate factorial analyses, while the center point chromatogram is the average of all center point analyses. Similarly, Figure 4.6 displays an overlay of the chromatograms at low, high, and center p oi nt levels from the ethyl acetate supported derivatizations. Refer to Table 4.3 for the values of each factor at a given level. 158 Figure 4.5 Overlay of averaged chromatograms from low (dark blue), high (light blue), and center point (dashed) level pyridine s upported derivatizations. Figure 4.6 Overlay of averaged chromatograms from low (dark green), high (light green), and center point (dashed) level ethyl acetate supported derivatizations 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 21 21.5 22 22.5 23 23.5 24 Normalized Abundance Ret. Time (min.) 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 21 21.5 22 22.5 23 23.5 24 Normalized Abundance Ret. Time (min.) 159 As shown from Figure 4 .5 and 4.6 , the significant peaks present i n chromatograms from derivatizations using either support solvent reflect that of the initial chromatograms provided in Figure 4.2 . Regardless of factor level for each experiment, the derivatization using ethyl ac etate continued to include underivatized TH CA, represented in the chromatogram as THC - TMS ( t R = 22.4 min). Conversely, the derivatizations using pyridine contained only THCA - 2TMS. For each full factorial analysis, ANOVA calculations were performed at the 95% confidence level ( = 0.05). For the derivatizations using pyridine as a support solvent, derivatization temperature was the single significant factor ( p = 0.001) of the three factors studied. Conversely, for the ethyl acetate supported derivatizations both temperature of r eaction ( p = 0.000) and time at temperature ( p = 0.003) were significant to the variability of the THCA - 2TMS product abundance. ANOVA t ables for both full factorial experiments, as well as Pareto charts of standardized effects are provided in Appendix IV. Given the reduced number of significant variables determined for derivatizations using pyridine as the support solvent, as well as the full derivatization of THCA, final derivatization optimization experiments were performed using pyridine rather than eth yl acetate as the support solvent. When performing optimization experiments, time and solvent ratio were held constant, while the tempe rature of reaction was varied between the low, center point, and high levels used for the full factorial analysis. A reac tion time of 10 min was selected, as increasing the total time of reaction was not significant to THCA - 2TMS production, nor did a reduc ed reaction time significantly impact the reproducibility of THCA - 2TMS abundance in the screening design experiments. A 5 0:50 ratio of BSTFA:pyridine was used during optimization, as it provided the most support solvent to act as a derivatization catalyst while maintaining an excess of BSTFA, and offered lower RSD values when compared to the 90:10 ratio experiments at each t ime and 160 temperature level. Temperature was varied between 30 °C, 65 °C, and 100 °C to reflect the conditions of the screening design. A dditionally, these experiments were completed using both a THCA concentration of 0.1 mg/mL and 0.2 mg/mL to determine the effect of analyte concentration on the overall efficiency of the derivatization. Efficiency was evaluated based on derivatized product abundance as well as any excess underivatized product present following the reaction. Figure 4.7 summarizes the optimiza tion studies with averaged THCA - 2TMS abundances for each concentration at each temperature. Figure 4.7 THCA - 2TMS abundance changes wi th changes in reaction temperature Similar to the full factorial screening design results, complete THCA derivatization was observed at each temperature, regardless of concentration. This was indicated by the absence of THC - TMS in the resultant chromatogr - tests of the mean THCA - 2TMS abundance at each temperature indicated that a t 0. 1 mg/mL , the average abundances were statistically comparable, and ANOVA analysis confirmed that there was no 0 0.5 1 1.5 2 30 65 100 THCA - 2TMS Normalized Abundance Temperature ( C) 0.1 mg/mL 0.2 mg/mL 161 statistically significant difference in abundance of THCA - 2TMS as a function of temperature ( p = 0.617 ). However, at 0.2 mg/mL, the mean abundance of THCA - 2TMS at 65 ° C was statistically higher than the abundance s at 30 ° C and 100 ° C - test ( p = 0.03 5 and p = 0.004, respectively) and ANOVA confirmed the significance of temperature for this set of optimization experime n ts ( p = 0.001 ). The optimum temperature was determined based on the abundance and reproducibility of the THCA - 2TMS peak. Derivatizatio n performed at 65 ° C provided the highest abundance of THCA - 2TMS product most reproducibly (RSD = 0.5%). Accordingly, th e optimized method for the derivatization of THCA by BSTFA - 1%TMCS using pyridine as a support solvent is summarized by Table 4.7 . Table 4.7 Optimum derivatization reaction parameters Parameter Optimized Value Reaction Time 10 min Reaction Temperature 65 ° C BSTFA : Pyridine 50:50 The optimized method shares some similarity with the published Restek method. 5 While pyridine was selected as a more suitable support solvent, based on reproducibility and full derivatization of THCA, the reagent:support solv e nt ratio remains at 50:50 to accommodate the necessity for excess BSTFA - TMCS as well as support solvent volu me needed to act as a reaction catalyst. The optimized temperature of 65 ° C is comparable to the published temperature of 70 ° C, with each providi n g ample heat to expedite the derivatization without risking THCA decarboxylation. Finally, the reaction time of the optimized method was reduced from that of the 162 published method. The reduction from 30 min to 10 min reflects the efficiency of the derivati z ation using pyridine as opposed to ethyl acetate. 4. 3 . 2 Evaluation of Optimized Method Linearity Using t he optimized method as outlined in Section 4.3.1, a concentration study was performed to assess the linearity and upper and lower limits of detecti o n for the derivatization of THCA. These reactions were performed in triplicate using concentrations ranging from 10 µg/mL to 0.5 mg/mL. Figure 4.8 displays an overlay of the chromatograms for each concentration analyzed. At each concentration, full deriva t ization of THCA to THCA - 2TMS is observed.. The inset chromatogram provides a closer look at the THCA - 2TMS pe ak abundances at each concentration. Figure 4.8 Overlay of averaged chromatograms from each concentration study With the lack of significant pea k s at the retention times associated with THC or THC - TMS at the highest concentration (0.5 mg/mL), the optimu m derivatization parameters provide 163 full THCA derivatization that is reliable even at high concentrations. Moreover, the calibration curve demonstr a ted linearity ( R 2 = 0.995) over the range of concentrations investigated. Table 4.8 provides the average nor malized THCA - 2TMS abundances for each concentration as well as the associated RSD value. A calibration plot is provided in the appendix. Table 4.8 Summary of concentration study normalized THCA - 2TMS abundances and RSDs THCA Concentration (mg/mL) THCA - 2TMS Norm. Abundance % RSD 0.01 0.013 8 21.1 0.0 5 0.166 15.1 0.1 0 0.568 8.4 0. 25 2.06 3.2 0. 50 4.49 4.3 As summarized by Table 4.8 , the THCA - 2TM S peak is present and discernable from background noise in concentrations as low as 0.01 mg /m L. Derivatizations of low concentrations (0.01 and 0.05 mg/mL) demonstrated high RSD values, which are expected due to the fact that small changes in abundances of that magnitude result in large relative standard deviations. As concentration increases, how ever, the RSDs for triplicate derivatizations are minimized, with acceptable experimental reproducibility (< 10%) starting at 0.1 mg/mL and higher concentrations. T his range of concentrations encompasses most recreational product THCA concentrations, with the highest concentration marijuana solvent extracts available with nearly 99% purity capable of being diluted and characterized. 164 4.3.3 Analysis of Cannabis Sol v ent Extracts by GC - MS Using the Optimized Derivatization Procedure To further validate the optimized derivatization method, cannabis solvent extract samples were derivatized and analyzed by GC - MS. This provided the ability to observe any matrix effects o n the derivatization reaction efficiency given the fact that additional cannabinoids may have been present in the cannabis solvent extract. Further, derivatization and GC - MS analysis of the cannabis solvent extracts served as a method to complete the compr e hensive chemical characterization and identification of a subset of extracts for inclusion w ith the previously discussed optical and spectroscopic characterization. The first cannabis solvent extract derivatized and analyzed represented the subset of KC S D case samples. A representative chromatogram of derivatized material from case sample KDPS 18 - 9026 is provided by Figure 4.9 . Similar to the chromatograms for the derivatized THCA reference material, the peaks of interest in this sample occur between t R 1 9 and 24 min, while other peaks present in the chromatogram are due to remaining BSTFA - TMS r eagent or siloxanes. 165 Figure 4.9 Chromatogram of case sample KDPS 18 - 9026 The inset chromatogram highlights lower abundance peaks eluting at t R = 20.2 min and at t R = 22.2 min. The mass spectrum of the first peak at t R = 20.2 min ( Figure 4.10 ) indicates a cannabinol derivative (CBN - TMS) . However, the peak at t R = 22.2 min ( Figure 4.11 ) was not readily identified due to the low number of ions present in the spectru m . The peaks at t R = 21.6 minutes and t R = 23 minutes are THCA - 2TMS ( Figure 4.12 ) and progesterone, respectively. A library spectrum for CBN - TMS is provided in the appendix. 166 Figure 4.10 Mass spectrum and chemical structure for CBN - TMS Figure 4.11 Mass spectrum for uni dentified peak at 22.2 minutes 167 Figure 4.12 Mass spectrum and chemical structure for THCA - 2TMS The mass spectrum for the peak present at t R = 22.2 min is not spectrally similar to THC - TMS and has not been observed in previous ex p erim ents for this work. Given the lack of total ions in the mass spectrum and no comparable library spectrum, this peak cannot be confidently identified. Some feature s of the mass spectrum are comparable to the THCA - 2TMS mass spectrum, including the base p eak at m/z 483.3 and the TMS fragment ion at m/z 73. The expected base peak for THCA - 2TMS, however, occurs at m/z 487.4. Noting this, the peak can be assumed to be a derivatized product, but the cannabinoid cannot be distinguished due to the low abundan c e an d low quality of the mass spectrum. The presence of derivatized CBN indicates that additional cannabinoids other than THCA are present in the KDPS 18 - 9026 samp le. This sample consists of two separate components, crystalline and wax, and was analyze d in bulk, with no separation of the components prior to 168 GC - MS analysis. The crystalline material has been previously identified through this work as THCA (Chapter 3). Noting that, the presence of CBN in this case sample may indicate the low concentration o f CB N and THC present in the wax component of the solvent extract. The small abundance of CBN may be due to the natural decomposition of THC and THCA over time, which would be expected in a sample approximately three years of age (estimated from sample su b miss ion). Additionally, the CBN - TMS product could be due to the rapid oxidation of THC to CBN due to the heat of the injection port, followed by immediate derivatizat ion. This process was not studied in the context of this work, and as such the definitive iden tification of the additional cannabinoid in the wax component is not possible. The THCA present in KDPS sample was quantified using the calibration curve. Given that this sample was clandestinely manufactured, it was not subjected to quantification p r ior to sale, nor was it quantified during forensic analysis. Using the regression equation available in the appendix (Figure A4.7), the case sample KDPS 18 - 90 - 26 was determined to contain 48.65% THCA by mass. The WB THCA Crystal sample, representing the S kymi nt dispensary solvent extracts, was derivatized using the optimum method and analyzed by GC - MS ( Figure 4.1 3 ). Similar to the KDPS 18 - 9026 sample chromatogram, the profile was the THCA - 2TMS product peak a t t R = 21.6 min. This peak was identified using its mass spectrum ( Figure 4.1 4 ). Additionally, low abundance peaks were present at the retention times associated with the CBN - TMS and THC - TMS peaks but could not be confidently identified by their mass spec t ra d ue to the low number of ions present in the spectra and the overlap of background m/z ions in each spectra. 169 Figure 4.1 3 Chromatogram of Skymint THCA Crystal sam ple Figure 4.14 Mass spectrum and chemical formula for THCA - 2TMS 0 2.5 4.5 7 9.5 12 14.5 17 19.5 22 24.5 27 Normalized Abundance Ret. Time (min.) 170 The lack of additiona l can nabinoids in this sample is representative of the purity achieved by the regulated production of solvent extractions for recreational use. The overall purity of t his sample was independently tested prior to sale and was listed as 76.66% THC ( ± 10%) on the p roduct label (provided in Appendix IV). Comparatively, the concentration of THCA in this sample determined in this work was 83.69%. Given that this determined val ue falls within the range reported from the independent testing facility, the accuracy of the THCA derivatization method can be displayed, although more samples would need to be quantified to further confirm the functional reliability of the derivatization method for the use of THCA quantification. In order to compare the THCA concentration to the reported THC concentration, the decarboxylation conversion rate (87.7%) was applied. Following this, the total maximum THC potency was found to be 73.4%. Finally , the Cannabidiol Life CBD crystal sample, representing solvent extracts derived from hem p, wa s derivatized and analyzed by GC - MS. The CBD Crystal was analyzed with and without derivatization ( Figure 4.1 5 ). The retention time of CBD was 19.9 min, while CBD - 2TMS was t R = 18.8 min. The chromatographic peaks were identified using their mass spec tra ( Figure 4.16 and 4.17 ), and spectral comparison was employed using the NIST spectra library. The library mass spectra are provided in the appendix. It should be no ted that no remaining CBD was present in the derivatized CBD chromatogram, indicating ful l der ivatization of CBD at the given concentration (~0.33 mg/mL) As the sample was previously identified as CBD (Chapter 3), as opposed to CBDA, derivatization of the sample prior to GC - MS analysis would not be necessary for differentiation and identificat ion p urposes, but does display the robustness of the derivatization method. 171 Figure 4.1 5 Chromatogram of Cannabidiol Life CBD Crystal sample Figure 4.16 Mass spectrum and chemical formula for CBD 0 7 4.5 7 9.5 12 14.5 17 19.5 22 24.5 27 Normalized Abundance Ret. Time (min.) CBD CBD-2TMS 172 Figure 4.17 Mass spectrum and chemical structure of CBD - 2TMS 173 4. 4 CONCLUSIONS Through this work, a method for the silylation of THCA throu gh a reaction with BSTFA - 1%TMCS was optimized using experimental design protocols to increase the overall abundance and reproducibility of the derivatized product, TH C A - 2T MS. Using an existing, published derivatization method as a starting point, the facto rs for derivatization evaluated included support solvent, reaction temperature, total time of reaction, and the ratio between the derivatizing reagent (BSTFA - 1%TMCS) a nd t he support solvent. Throughout the pre - screening and screening designs, pyridine perf ormed more efficiently and more reproducibly than ethyl acetate as a support solvent. Further, for reactions performed using pyridine only, derivatization temperature was found to be a significant variable for THCA - 2TMS abundance. Through ANOVA analysis an d optimization experiments a method with optimized time, temperature, and reagent:support solvent ratio was proposed. The optimized method was used to analyze and qua n tify marijuana - derived solvent extracts, and hemp - derived CBD extracts were additionally analyzed. The optimized method displayed excellent linearity within the concentration range studied, and efficiently derivatized samples with multiple different canna b inoi ds. 174 APPENDIX 175 ANOVA calculations: Table A4.1 Calculations for the degrees of freedom Source of Variation Degrees of Freedom Factor A = a - 1 Factor B = a - 1 Interaction AB = (a 1)(b 1) Error = Total Where: a and b levels of Factor A and B Table A4.2 Calculations for the sum of squares for two - way ANOVA 19 Source of Variation Sum of Squares Factor A Factor B Interaction AB Error Total Where: a and b number of levels in factor A or B n number of trials y mean of the i th factor level of factor A y mean of the j t h factor level of factor B - overall mean of all observations y ij mean of the observations at the i th level of factor A and the j th level of factor B 176 Table A4.3 Calculations for the mean squares Source of Variation Mean Squares Factor A Factor B Interaction AB Error Figure A4.1 THC mass spectrum 177 Figure A4.2 NIST library mass spectrum result for THC - T MS Pre - Screening Experiment Data: To compare the hold conditions of both solvents, reactions were performed under equivalent conditions, but on separate days given the length of the GC - MS analysis . Figure A 4.3 provides an overlay of averaged chromatograms f rom the ethyl acetate derivatization with refrigerated and 24 hour hold conditions. Similarly, Figure A 4.4 provides an overlay for the chromatograms from the refrigerated and 24 hour hold experiments using py ridine as a support solvent. 178 Figure A 4.3 Ave ra ged chromatograms comparing refrigerated and 24 hour hold samples of THCA derivatization using ethyl acetate Figure A 4.4 Averaged chromatograms comparing refrigerated and 24 hour hold samples of THCA deriv atization using pyridine 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 21 21.5 22 22.5 23 23.5 24 Normalized Abundance Ret. Time (min.) Refirgerated 24 hour hold 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 21 21.5 22 22.5 23 23.5 24 Normalized Abundance Ret. Time (min.) Refrigerated 24 hour hold 179 Table. A 4. 4 Summary of reproducibility for pre - screening hold experiments Ethyl Acetate Pyridine Refrigerated 24 Hr. Hold Refrigerated 24 Hr. Hold THCA - 2TMS 14.0% 12.5% 7.2% 9.6% Progesterone 17.5% 11.0% 5.1% 9.1% Table A4. 5 Inter - and intra - vial RSDs for ethyl acetat e and pyridine pre - screening experiments Ethyl Acetate Pyridine Vial Refrigerated 24 - hour hold Refrigerated 24 - hour hold 1 6.2% 4.5% 1.6% 0.64% 2 7.0% 5.6% 1.0% 6.2% 3 8.4% 3.2% 2.4% 11.1% Total 14.0% 12.5% 7.2% 9.6% Table A4. 6 Full ANOVA results f ro m derivatizations using pyridine as support solvent Source DF Adj SS Adj MS F - Value P - Value Model 10 0.243462 0.024346 3.12 0.029 Blocks 3 0.037604 0.012535 1.61 0.236 Linear 3 0.183026 0.061009 7.81 0.003 Time 1 0.019493 0.019493 2.5 0.138 Temperature 1 0.150358 0.150358 19.26 0.001 Ratio 1 0.013175 0.013175 1.69 0.217 2 - Way Interactions 3 0.02045 0.006817 0.87 0.48 Time*Temperature 1 0.011664 0.011664 1.49 0.243 Time*Ratio 1 0.000152 0.000152 0.02 0.891 Temperatur e* Ratio 1 0.008634 0.008634 1.11 0.312 Curvature 1 0.002381 0.002381 0.31 0.59 Error 13 0.101503 0.007808 Lack - of - Fit 9 0.075507 0.00839 1.29 0.432 Pure Error 4 0.025997 0.006499 Total 23 0.344965 180 Table A4. 7 ANOVA results fro m derivatizations using ethyl acetate as support solvent Source DF Adj SS Adj MS F - Value P - Value Model 8 0.461012 0.057627 28.47 0.010 Blocks 1 0.084092 0.084092 41.55 0.008 Linear 3 0.367635 0.122545 60.55 0.003 Time 1 0.05373 0.05373 26.55 0. 01 4 Temperature 1 0.306188 0.306188 151.29 0.001 Ratio 1 0.007717 0.007717 3.81 0.146 2 - Way Interactions 3 0.004005 0.001335 0.66 0.63 Time*Temperature 1 0 0 0 0.989 Time*Ratio 1 0.003452 0.003452 1.71 0.283 Temperature*Ratio 1 0.0 00 553 0.000553 0.27 0.637 Curvature 1 0.005281 0.005281 2.61 0.205 Error 3 0.006072 0.002024 Lack - of - Fit 1 0.000566 0.000566 0.21 0.695 Pure Error 2 0.005506 0.002753 Total 11 0.467084 Figure A4. 5 Pareto chart for the pyridin e f ull factorial design 181 Figure A4. 6 Pareto chart for ethyl acetate full factorial design ANOVA Discussion: During Minitab full factorial analysis, linear regression models are formed and fit using the provided data. In addition to the ANOVA output, a mod el fit parameter is provided, incl uding an R 2 value for the model and predicted R 2 for any new data added to the model. In the case of the pyridine full factorial experiments, the R 2 value for the model was low, at 70.58% compared to the ethyl acetate full factorial fit with an R 2 value of 98.70%. Additionally, the predicted R 2 value for the pyridine model could not be estimated, while the predicted R 2 value for the ethyl acetate model was 73.91%. The difference in fits between both models may be due to the fact that for the pyridine experiment, only one factor was found to be significant (temperature), while both temperature and time were significant f or the ethyl acetate experiment. Given that only one factor was significant, and the variance between t he g roups of data for each temperature was not large, the model was not fit as accurately to a regression model. Conversely, 182 the ethyl acetate experiment provided more significant variation between responses for both the time and temperature factors, allow ing for a more well - fit model to be produced. As the ANOVA results were primarily used to determine the significant factors for each experiment, however, the model was found to describe the variation in the pyridine experiment (p=0.029) and ethyl acetate e xper iment (p=0.010) regardless of regression model fit as described by the R 2 values. Figure A4. 7 Regression plot for THCA concentration study T he re gression equation for the fitted line (R 2 = 0.995) is: (Eq. A1) 0.00 1.00 2.00 3.00 4.00 5.00 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 Normalized Abundance THCA concentration (mg/mL) 183 Figure A4. 8 Manufacturer information and quantification for Skymint THCA Crystal sample Figure A4. 9 NIST library mass spectrum for CBN - TMS Figure A4. 10 NIST library mas s spectrum for CBD 184 Figure A4. 11 NIST library mass sp ec trum for CBD - 2TMS 185 REFERENCES 186 R EFERENCES (1) A griculture Improvement Act of 2018 (P.L. 115 - 334), 2018 ; https://www.congress.gov/115/plaws/publ334/PLAW - 115publ334.pdf (2) Wang M, Wang YH, Avula B, et al. Decarboxylation Study of Acidic Cannabinoids : A Novel Approach Using Ultra - High - Performance Supercritical Fluid Chromatography/Photodiode Array - Mass Spectrome try. Cannabis Cannabinoid Res. 2016 Dec 1;1(1):262 - 271 (3) Aizpurua - Olaizola, O., Omar, J., Navarro, P. et al. Identification and quantification o f cannabinoids in Cannabis sativa L. plants by high performance liquid chromatography - mass spectrometry. Anal Bio anal Chem 2014; 406, 7549 7560 . (4) Backer BD, Debrus B, Lebrun P, Theunis L, Dubois N, Decock L, et al. Innovative development and validation o f an HPLC/DAD method for the qualitative and quantitative determination of major cannabinoids in cannabis plant mat erial. Journal of Chromatography B 2009;877(32):4115 24. (5) Politi M, Peschel W, Wilson N, Zloh M, Prieto JM, Heinrich M. Direct NMR analysis o f ca nnabis water extracts and tinctures and semi - quantitative data on delta9 - THC and delta9 - THC - acid. Phytochemistr y. 2008 Jan;69(2):562 70. (6) Rigdon, A. Accurate Quantification of Cannabinoid Acids and Neutrals by GC. Restek.com, 2015. (7) Cardenia V, Toschi TG , Scappini S, Rubino RC, Rodriguez - Estrada MT. Development and validation of a Fast gas chromatography/mass spect rometry method for the determination of cannabinoids in Cannabis sativa L. Journal of Food and Drug Analysis 2018;26(4):1283 92. (8) Milman G, B ar ne s AJ, Lowe RH, Huestis MA. Simultaneous quantification of cannabinoids and metabolites in oral fluid by two - dim ensional gas chromatography mass spectrometry. Journal of Chromatography A 2010;1217(9):1513 21. (9) Kranenburg RF, Verduin J, Stuyver LI, Ridd er R D, Beek AV, Colmsee E, et al. Benefits of derivatization in GC MS - based identification of new psychoactive subs tances. Forensic Chemistry 2020;20:100273. (10) S.C. Moldoveanu, V. David, Derivatization methods in GC and GC/MS Gas Chromatography - Derivatiza ti on , Sample Preparation, Application, IntechOpen 2019; (11) Poole CF. Alkylsilyl derivatives for gas chromatography. J ournal of Chromatography A 2013;1296:2 14. 187 (12) Prata VDM, Emídio ES, Dorea HS. New catalytic ultrasound method for derivatization of 11 - nor - - t et ra hydrocannabinol - 9 - carboxylic acid in urine, with analysis by GC - MS/MS. Analytical and Bioanalytical Chemistry 2 012;403(2):625 32. (13) Jurado - Sánchez B, Ballesteros E, Gallego M. Determination of carboxylic acids in water by gas chromatography mass spectro me tr y after continuous extraction and derivatisation. Talanta 2012;93:224 32. (14) Sigma - Aldrich Co. BSTFA Product Specification. 1997; https://www.sigmaaldrich.com/Graphics/Supelco/objects/4800/4746.pdf (15) Sigma - Aldrich Co. Derivatization Reagents for Selective R es ponse and Detection in Complex Matrices. 2011; https://www.sigmaaldrich.com /content/dam/sigma - aldrich/migrationresource4/Derivatization%20Rgts%20brochure.pdf (16) Araujo PW, Brereton RG. Experimental Design I. Screening. Trends Anal Chem 1996; 15:26 - 31. (17) S tu fk en J, Dean A, Morris M, Bingham D, Dean. Handbook of Design and Analysis of Experiments. CRC Press, 2015; (18) Brereton RG. Chemometrics: data driven extraction for science. Chichester: Wiley Blackwell, 2018; (19) Watson, J. T.; Sparkman, O. D. Introduction to M as s Spectrometry: Instrumentation, Applications and Strategies for Data Inter pretation; Wiley: Chichester, 2011; 188 5. CONCLUSIONS AND FUTURE WORK 5 .1 CONCLUSIONS Throughout this work, sets of cannabis solvent extracts, derived from either marijuana or he mp , were comprehensively characterized. The characterization of both the crys talline and wax components of the cannabis solvent extracts included optical characterization by polarized light microscopy (PLM), spectroscopic characterization by infra - red (I R) s pectroscopy and single crystal X - ray diffraction (XRD), and finally spectro metric identification by gas chromatography - mass spectrometry (GC - MS). These characterizations were performed in order to further research the components of cannabis solvent ext ra ct s using a variety of instrumentation available to forensic scientists, as w ell as provide methods by which solvent extracts derived from marijuana or hemp could be differentiated. Analysis of the crystalline component of each cannabis solvent extract b y polarized light microscopy allowed for subsets of extracts to be compared a nd grouped based on crystal optics. As demonstrated, the KCSD case samples each contained analogous crystals regardless of sample age or extract texture. Similarly, the crystall in e material from the Skymint dispensary samples were optically similar within the Skymint subset and were comparable to the KCSD case sample crystals. As both subsets were derived from marijuana (either clandestinely or through regulated manufacturing), t hi s substantiated that a preferred crystalline material forms in marijuana solv ent extracts regardless of manufacturing technique or product texture. Additionally, the samples derived from hemp from the Cannabidiol Life dispensary were optically characteri ze d and were found to contain analogous crystals within their subset but did no t share optical characteristics with the marijuana extract subset. The difference in crystal optics between the marijuana and hemp - derived solvent extracts provides the opportun it y to distinguish 189 the two optically based on their crystal presence, allowing for a rapid screening for macroscopically similar samples. Spectroscopic analysis of the sample subsets provided chemical characterization not offered by PLM analysis. Micro - at te nuated total reflectance - Fourier transform infrared spectroscopy (micro - ATR - FTIR) analysis was used to chemically characterize both the crystalline and wax components (when present) of each sample and provided presumptive identification for each compon en t. The marijuana solvent extract subset, comprised of KCSD case samples and S kymint dispensary samples, had comparable IR spectra for both the crystalline and wax components of each sample. Each of the crystalline sample components shared spectral charac te ri stics analogous to tetrahydrocannbinoic acid (THCA), while the wax componen t shared spectral characteristics with both THCA and tetrahydrocannabinol (THC). Conversely, the Cannabidiol Life dispensary sample crystalline and wax components were spectrall y si milar to cannabidiol (CBD). Further analysis of two crystalline samples fro m the marijuana solvent extract subset and one crystalline sample from the hemp extract subset was performed using single crystal XRD, which confirmed the presumptive identifica ti on for each sample via micro - ATR - FTIR. The spectroscopic characterization and identification of each subset further strengthened the hypothesis that cannabis solvent extracts derived from marijuana and hemp would produce products differentiable by forens ic t echniques. Not only were the samples grouped successfully based solely on o ptical characteristics, but their optical similarities (within subsets) and differences (between subsets) can be chemically explained based on the prominent cannabinoid present in t he final extracted product. 190 Given the wide - spread use of GC - MS for seize d drug analysis in forensic crime labs, it was necessary to perform spectrometric analysis of samples from each subset in order to provide a comprehensive characterization of ca nn ab is solvent extracts. Due to the identification of the marijuana - derived sub GC - MS analysis was necessary to avoid decarboxylation to THC. As such, a previously published silylatio n pr ocedure was optimized for the derivation of THCA. 1 A combination of pre - screening experiments, replicated full factorial screening design experiments, and optimization experiments were performed to establish the significant parameters of derivatizatio ns u sing N,O - Bis(trimethylsilyl)trifluoroacetamide (BSTF A) - 1% trimethylchlorosilane (TMCS) and either pyridine or ethyl acetate as support solvents. The three factors studied during the full factorial screening design included temperature of reaction, ti me o f reaction, and the ratio between BSTFA and the supp ort solvent. From this design, temperature was found to be the only significant parameter for derivatization using pyridine as the support solvent, while both temperature and time were significant for e th yl acetate - supported derivatizations. The preferred solvent, as determined by both the pre - screening and full factorial screening design experiments, was pyridine, rather than ethyl acetate due to increased reproducibility and overall abundance of der iv at ized product THCA (2) trimethylsilyl (THCA - 2TMS). The optimization of the pyridine - supported derivatization was performed by adjusting the temperature of reaction while maintaining the time and solvent ratio. The optimum parameters for the derivatiza ti on of THCA were determined to be a 10 - minute reaction at 65 ° C using a ratio of 50:50. This optimized reaction was validated using a concentration study that 191 showed excellent linearity and full derivatization of THCA to THCA - 2TMS within the concentration r an ge of 10 µg /mL to 0.5 mg/mL (R 2 = 0.995). Following optimization, a selection of cannabis extract samples two crystalline samples from the marijuana solvent extract subset and one crystalline sample from the hemp extract subset were derivatized and a na lyzed by GC - MS. The samples from the marijuana extra ct subset contained analogous prominent peaks, with THCA - 2TMS making up most of the chemical composition of the sample. Additionally, the hemp - derived extract sample was analyzed with and without deri va ti zation, providing spectrometric identification of th e main component as CBD and CBD - 2TMS respectively. This confirmed the robustness and versatility of the optimized derivatization parameters for the derivatization of additional cannabinoids other than T HC A. This work showcases the possible methods of scree ning and analysis for cannabis solvent extracts derived from either marijuana or hemp. Given the optical and chemical differences between marijuana and hemp - derived solvent extracts, the research pres en te d provides the opportunity to apply optical and spec troscopic methods in forensic laboratories for the purpose of screening macroscopically similar samples prior to confirmatory analysis. The readily observed optical properties between crystalline comp on en ts from the marijuana and hemp subsets offers rapid screening, with more chemically specific information than what is generally determined by a Duquenois - Levine color test, while additionally reducing sample destruction. Further, the optimized derivati za ti on protocol can be readily implemented in forensic a nd independent laboratories where complete cannabinoid potency determination is necessary by GC - MS analysis. 192 5 .2 FUTURE WORK Portions of this work can be expanded to provide more thorough character iz at ion of cannabis solvent extracts. For example, the m arijuana - derived extracts used in this work were hydrocarbon solvent extracts, most commonly extracted using butane, while the hemp - derived extracts were extracted using supercritical CO 2 . Moving forw ar d, research in this field could compare the crystal ha bits between marijuana - derived extracts using extraction solvents other than hydrocarbons. It is expected that the crystalline components would exhibit similar optical and chemical characteristics, gi ve n the cannabinoid profile for such samples. Addition ally, during this work it was concluded by single - crystal XRD that the crystalline THCA from a clandestine case sample was a racemic crystal, while the crystalline THCA from the Skymint dispensary sam pl e was chiral. Given the illegal nature of clandestine solvent extract manufacturing using hydrocarbons such as butane, a case study comparing the single - crystal XRD results of clandestine - manufactured solvent extracts versus regulated dispensary solvent ex tr acts may provide the opportunity to differentiate th e two based on sample chirality. This may allow for the effective differentiation of legally (regulated) and illegally produced (clandestine) marijuana solvent extracts. Finally, further research ca n be performed using the optimized derivatization method for the analysis of multi - cannabinoid samples and samples of different matrices. In this work, a standard THCA certified reference material was used for all of the optimization and linearity validati on e xperiments. As such, this method was optimized for t he derivatization of THCA only, and not THCA in combination with other common cannabinoids (including THC). It would be beneficial to further optimize this procedure for each common cannabinoid indivi du al ly, and in a mixture, in order to provide the most c onfident and comprehensive derivatization protocol. 193 REFERENCES 194 R EFERENCES (1) Rigdon, A. Accurate Quantification of Cannabinoid Acids and Neutrals by GC. Restek.com, 2015.