RETURNING MATERIALS: IVIESI.J Place in book drop to LIBRARIES remove this checkout from .4--cp--L your record. FINES will be charged if book is returned after the date stamped below. INVESTIGAIIONB INTO THE FATE OF TOXAPHBNB IN NATURAL IITBRB BY Susan Erhardt-Zabik A Dissertation Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry/ Center for Environmental Toxicology 1988 53®ngg ABSTRACT INVESTIGATIONS INTO THE FATE OF TOXAPHBNB IN NATURAL WATERS 3? Susan Brhardt-Zabik This dissertation consists of two separate sections. The first section describes investigations of the fate of toxaphene in natural waters. The second section describes a study of the vapor phase methodology in which the molar response characteristics of three series of chlorinated compounds by negative ion mass spectrometry (NI/MS) and electron capture detection (ECD) is characterized. Toxaphene, a complex chlorinated hydrocarbon pesticide, was exposed to sunlight and UV-light in solutions of natural water and isolated aquatic humic substances. The effects of pH and added iron in addition to the aquatic humic substances on the photolysis of toxaphene were also investigated. Toxaphene in acetonitrile/water solutions with added photosensitizers were also exposed to UV-light and sunlight as a comparison to those solutions containing humic material. Toxaphene produced products indicative of reductive dechlorination in both natural water systems and in aqueous solutions of humic materials. These same products were not observed in the photosensitized solutions indicating that some mechanisms other than a triplet energy transfer from the humics to the toxaphene component was responsible for the observed products. However, overall dechlorination of the toxaphene mixture was more extensive in the presence of photosensitizers than in the presence of humics. A second project investigating the nature of selective structure-related responses of highly chlorinated species by NI/MS was also undertaken. Molar response characteristics of three series of chlorinated compounds were determined by NI/MS and ECD, and compared. For two chlorinated aromatic series, molar response in both types of analysis was found to be directly correlated with degree and position of chlorination. Implications of the phenomenon are discussed in relation to the quantitation of complex chlorinated mixtures such as PCB Alachlors and toxaphene. "Education is experience and the essence of experience is self-reliance." T.H. White once and Future King iv ACKNOWLEDGHENTS When any such undertaking has been completed there are many more people involved than that of the author that have loaned their support and encouragement along the way. These include professors, family and friends made here at M.S.U. I would like to thank Dr. J. T. Watson for his financial and moral support during my research. I know that some of what I accomplished for my research was a little different than what he was used to, but he let me continue none the less. I like to thank Young C. Chung, who showed me some of the inherent beauty in science that I had not seen before and convinced me to continue in the darker days of my career as a graduate student. my mother and father also encouraged me to keep at it, and finally learned not to ask when I would be finished. My thanks go to the graduate students and people at 1R4 at the PRC for their help along the way as well as making life at the PRC enjoyable. There always seemed to be someone to commiserate with over a cup of our gourmet coffee. Included in this group are Lester Geissel, Bob Kohn, Bob Shuetz, Dick Levitt, Glenn Dickmann, Holly Fortnum, Mary Ann Heindorf, Inez Toro-Suarez, Virginia Vega, Gamal Kehdr, Hai Dong Kim and last but by no means least, Dr. Salah Selim. Thanks guys! My thanks go to the entire Zabik family, Mary Ellen, Matt and Mary McMahon for your help getting both of us through the rough times and seeing us to the end. Especially to Matt for keeping the instruments running in times of extreme need and teaching me more about instrumentation than I ever thought I could learn. Thank you just isn’t enough to say sometimes. Finally to my husband Jack Zabik: we knew it wouldn't be easy, but it wasn’t as bad as we thought either! vi TABLE OF CONTENTS LIST OF TABLES O O O O O O O O O O O O O O O O O O O O O x LIST OF FIGURES O O O O O O O O O O O O O O O O O O O O Xii SECTION I: INVESTIGATIONS INTO THE FATE OF TOXAPHENE IN NATURAL ““38 O O O O O O O O O O O O O O O O O O O O O 1 CHAPTER 1: LITERATURE REVIEW OF TOXAPHENE. . . . . . . 2 Introduction and Research Objectives. . . . . . . . 2 Toxaphene Composition and Isolated Components. . . 4 Literature Methods for Toxaphene Analysis and Detection. . . . . . . . . . . . . . . . . . . . .13 Toxicity. . . . . . . . . . . . . . . . . . . . 17 Aquatic Toxicity. . . . . . . . . . . . . . . . . . 19 Environmental Dynamics. . . . . . . . . . . . . . . 21 Global Distribution. . . . . . . . . . . . . . . . .24 CHAPTER 2: HUMIC MATERIALS AND THEIR INTERACTIONS WITH XENOBIOTICS O O O O O O O O O O O O O O O O O O O O O 27 IntrOduCtion O O O O O O O O O O O O O O O O O O 2 7 Structural Characteristics of Humics. . . . . . . . 28 Photophysical Characteristics of Humics. . . . . . 31 Binding of Xenobiotics by Humic Materials. . . . . 32 Photochemical Characteristics of Humics. . . . . . 35 Photo-oxidation. . . . . . . . . . . . . . . . 45 Energy transfer. . . . . . . . . . . . . . . . 48 Attenuation and quenching effects. . . . . . . 51 Photoincorporation of xenobiotics in humics. . 52 CHAPTER 3: MATERIALS AND METHODS FOR EVALUATING THE EFFECT OF WEATHERING ON THE COMPLEX MIXTURE TOXAPHENE. . . . 58 Standards. . . . . . . . . . . . . . . . . 59 Solvents, Chemicals and Resins. . . . . . . . . . . 59 Solvents. . . . . . . . . . . . . . 59 Sensitizers and actinometer chemicals. . . . . 60 Ion-exchange resins. . . . . . . . . . . . . . 60 Collection and Storage of Natural Waters. . . . . . 60 Preparation of MAD-2 Resin for Humic Extraction. . 62 Isolation of Humic Substances. . . . . . . . . . . 63 Sunlight Experiments. . . . . . . . . . . . . . . . 64 Indoor Experiments. . . . . . . . . . . . . . . . . 65 Addition of humics. . . . . . . . . . . . . 66 Addition of photosensitizers. . . . . . . . . . 67 vii Actinometers. . . . . . . . . . . . . . . . . . 67 Sample Extraction and Analysis. . . . . . . . 53 Capillary GC/MS-Negative Ion Mass Spectrometry. . . 70 Statistical Treatment of the Data. . . . . . . . . 71 CHAPTER 4: RESULTS AND DISCUSSION. . . . . . . . . . .' 75 Phase I: Irradiation of Toxaphene in Natural and Distilled Waters. . . . . . . . . . 76 Phase II: The Effect of Simulated Natural Waters . 97 Effect of humic concentration. . . . . . . . . 98 Effect of pH and added humic materials. . . . . 103 Effect of Fe(III) and added humic materials. . 105 Phase III: Effect of Added Photosensitizers on Toxaphene Acetonitrile/Water Solutions. . . . . . 109 Comparison of Residues Isolated from Fish versus Photolyzed Material. . . . . . . . . . . . . . . 136 Implications of the Photolysis of Toxaphene in Natural Waters. . . . . . . . . . . . . . . . . . . . . . 140 CHAPTER 5: SUMMARY AND CONCLUSIONS. . . . . . . . . . 142 CHAPTER 6: FUTURE WORK. . . . . . . . . . . . . . . . 145 LIST OF REFERENCES O O O O O O O O O O O O O O O O O O O l 4 7 SECTION II: SELECTIVE SENSITIVITY OP HIGHLY CHLORINATED SPECIES IN NI/MS: IMPLICATIONS FOR COMPLEX MIXTURE WYBIS O O O O O O O O O O O O O O O O O O O O O O O O 157 CHAPTER 1: LITERATURE REVIEW OF NEGATIVE ION MASS S PE CTROMTRY O O O O O O O O O O O O O O O O O O O O O 1 5 8 Introduction. . . . . . . . . . . . . . . . . . . . 158 Negative Ion Formation. . . . . . . . . . . . . . . 161 The electron capture process. . . . . . . . . . 161 The effect of molecular structure. . . . . . . 164 The effect of moderating/reagent gas. . . . . . 165 The effect of source temperature. . . . . . . . 167 Applications. . . . . . . . . . . . . . . . 168 Chlorinated pesticides. . . . . 169 Polychlorinated and polybrominated biphenyls. . 174 Other xenobiotics. . . . . . . . . . . . . . . 177 NI/MS as a rapid screening technique for environ- mental samples. . . . . . . . . . . . . . . . 180 References . . . . . . . . . . . . . . . . . . . . 184 CHAPTER 2: SELECTIVE SENSITIVITY OF HIGHLY CHLORINATED SPECIES IN NI/MS: IMPLICATIONS FOR COMPLEX MIXTURE ANALYSIS. . . . . . . . . . . . . . . . . . . . . . . . 187 Introduction. . . . . . . . . . . . . . . . . . . . 187 Experimental. . . . . . . . . . . . . . . . . . . . 189 viii Chemicals O O O O O O O O O O O O O O O O O O O 189 ECD anaIYSis O O O O O O O O O O O O O O O O O O 190 Mass spectrometry. . . . . . . . . . . . . . 190 Results. . . . . . . . . . . . . . . . . . . . . . 192 Chlorinated pesticides as a model toxaphen mixture. . . . . . . . . . . . . . . . . . . 192 Chlorinated phenols. . . . . . . . . . . . . . 199 Chlorinated biphenyls. . . . . . . . . . . . . 204 Discussions. . . . . . . . . . . . . . . . . . . . 212 Summary and Conclusions. . . . . . . . . . . . . . 218 Future Work. . . . . . . . . . . . . . . . . . . . 220 References. . . . . . . . . . . . . . . . . . . . 220 APPENDICES O O O O O O O O O O O O O O O O O O O O O O O 2 2 2 ix LIST OF TABLES Section I: Investigations into the Fate of Toxaphene in Table 1: Table 2: Table 3: Table 4: Table 5: Table 6 : Table 7: Natural Waters Concentration of Reactive Oxygen Species in Natural Waters. . . . . . . . . . . . . . . . . 44 pH and Conductivity of Waters Used for Toxaphene PhOtOlYSis O O O O O O O O O O O O O O O O O O O 7 6 Mass Spectral Peaks Used to Integrate the Various Chlorinated Isomers of Toxaphene. . . . . . . . 87 Calculated First Order Rate Constants of lppm Toxaphene in Lake Michigan Water. . . . . . . . 96 Calculated First Order Rate Constants of 4ppm Toxaphene in Distilled water with Added Tahquamenon Humics. . . . . . . . . . . . . . 99 Calculated First Order Rate Constants (kp) (days'l) of 4ppm Toxaphene in Distilled water with added Betsy River Humics at Various pHs. . . . . . . 104 Calculated First Order Rate Constants (hp) (days'l) of 4ppm Toxaphene in Acetonitrile/Water Solution with Photosensitizers. . . . . . . . . . . . . 116 Section II: Selective Sensitivity of Highly Chlorinated Species in Negative Ion Mass Spectrometry: Implications for Table 1: Table 2: Table 3: Complex Mixture Analysis Negative Ion Spectra of Chlorinated Bicyclic Compounds. . . . . . . . . . . . . . . . . . . 198 Negative Ion Spectra of Chlorinated Phenols. . 203 Negative Ion Spectra of PCB Isomers. . . . . . 210 Appendices Table A: Normalized Chromatographic Data for Sensitized Toxaphene Solution Containing Crystal Violet. 223 Table B1: Calculated First Order Rate Constants of lppm Toxaphene in Betsy River Water . . . . . . . 224 Table B2: Calculated First Order Rate Constants of lppm Toxaphene in Distilled Water. . . . . . . . 225 Table B3: Calculated First Order Rate Constants of lppm Toxaphene in Two Hearted River Water. . . . 226 Table C1: Calculated First Order Rate Constants of 4ppm Toxaphene in Distilled Water with added Betsy River Humics O O O O O O O O O O O O O O O O 227 Table C2: Calculated First Order Rate Constants of 4ppm Toxaphene in Distilled Water with added Blind Sucker Flooding Humics. . . . . . . . . . . . 228 Table C3: Calculated First Order Rate Constants of 4ppm Toxaphene in Distilled Water with added Two Hearted River Humics. . . . . . . . . . . . . 229 Table DI: Calculated First Order Rate Constants (kp)(days’1) of 4ppm Toxaphene in Distilled Water at Various pHSO O O O O O O O O O O O O O O O O O O O O O23o Table DZ: Calculated First Order Rate Constants (kp)(days‘1) of 4ppm Toxaphene in Distilled Water with Two Hearted River Humics at various pHs. . . . . .231 LIST OF FIGURES Section I: Investigations into the Fate of Toxaphene in Figure 1. Figure 2. Figure 3. Figure 4. Figure 5. Figure 6. Figure 7. Figure 8. Figure 9. Figure 10. Natural waters Rearrangement of Camphene to Bornane as a Result of Its Chlorination. . . . . . . . . . . 6 Isolated components of toxaphene: Toxicant B, 2,2,5-endo,6-exo,8,9,10-heptachlornorbornane (15), components I, II and III (16), Toxicant A: composed of two isomers (17, 18), 2,5,6-exo,8,8,9,10-heptachlorodihydrocamphene (19, 20), and Toxicant Ac, 2,5-exo,3,6-endo,8,8,9,10, lo-nonachloronorbornane (21). . . . . . . . . . 8 Alternative route for the chlorination of camphene (19’ 20) O O O O O O O O O O O O O O O O O O O 12 Comparison of the emission spectrum of natural sunlight with the absorption spectra of humic acids and propylenethiourea (84) . . . . . . . 37 Various reaction pathways of humic materials which have been shown to influence the rate of degradation of xenobiotics (65). . . . . . . . 39 Oxidation of DMF by singlet oxygen. . . . . . .46 Photolysis products of the herbicide atrazine, identified in aqueous solutions of fulvic acid (111) O O O O O O O O O O O O O O O O O O O O O 54 Collection sites for natural water samples around the state of Michigan. All samples from the Upper Peninsula were high in humic materials. . . . . 61 Representative HPLC chromatogram of the actinometer solution. Conditions are in the body of the text. PNAPap-nitroacetophenone. . . . . 69 A typical toxaphene chromatogram showing the seg- mentation pattern of 16 different segments or zones used to normalize chromatographic data as described in the statistical analysis section.74 xii Figure Figure Figure Figure Figure Figure 11. 12. 13. 14. 15. 16. CC chromatograms of 10 ppm control toxaphene (lower) and 10 ppm toxaphene in Lake Michigan Water (upper) exposed to sunlight for ten days. The arrows designate peaks which either increased or decreased in relative peak area over the course of exposure to sunlight. . . . . . . . 79 Mass spectral data (EI) obtained from the analysis of a component of chromatographic segment #5. Top: Reconstructed total ion chromatogram following identical conditions as those described for NI/MS analysis. Bottom: EI mass spectrum obtained at retention time of 35.4 min. during GLC analysis. . . . .82 Electron impact mass spectrum of reductive de- chlorinated product in chromatographic segment #8. Its empirical formula was determined to be C10H3C13, indicating a species having three degrees of unsaturation. . . . . . . . . . . .34 Histograms of the isomer composition of sunlight exposed solutions of toxaphene in natural and distilled waters at three concentrations. (a)Distilled water, (b)Betsy River Water, (c)Lake Michigan Water, (d)Two Hearted River Water. Note that with the exception of distilled water, an increase in the relative concentration of C17 isomers occurs with decreasing toxaphene concentration. . . . . . . . . . . . . . . . 83 The dendrogram which resulted from the heirach- ical cluster analysis conducted using the statistical analysis program SYSTAT on the normalized chromatographic data of sunlight exposed toxaphene solutions. Two Hearted River (THOTl and THOT2) samples are the least related to control samples (CONTl, 2 and 3) indicating the greatest overall change. . . . . . . . . .92 A plot of ln(C/C.) of chromatographic segment #5 of 1 ppm toxaphene in Lake Mic igan water, exposed to sunlight ( Q ), and UV-light ( O ) . 9S xiii Figure Figure Figure Figure Figure Figure Figure 17. 18. 19. 20. 21. 22. 23. First order plot of ln(C/C.) versus time for chromatographic segment #13 of 4 ppm toxaphene in distilled water with added Tahquamenon humics at 100 ppm ( E] ). 10 ppm ( C) ). and 1 ppm ( ID ) O O O O O O O O O O O O O O O O O O O O O O 00 First order plot of ln(C/C.) versus time for chromatographic segment #5 4 ppm toxaphene in distilled water with added Tahquamenon humics ( O ), Two Hearted humics ( Q ), Blind Sucker Flooding humics ( E] ), and Betsy River humics ( ll ) at 100 ppm. . . . . . . . . . . . . . 101 First order plot of ln (C/Co) chromatographic segment #5 of 4 ppm toxaphene at pH5 in distilled water with added Fe(III)( C) ), and without Fe(III) ( {I ), and Two Hearted Humics with ( A ) and without Fe(III)( A ). Fe(III) was added to a final concentration of 1 x 10' M and humics were added to a final concentration of 100 ppm. . . . . . . . . . . . . . . . . . . 107 Structures and triplet energies of the photo- sensitizers used in acetonitrile/water solution with 4 ppm toxaphene. . . . . . . . . . . . .110 Dendrogram produced by cluster analysis of toxaphene in acetonitrile/water solutions after 10 days exposure to UV-light. Legend to abbreviations is as follows: Fluorene - FLINl, FLINZ Biphenyl - BPIN1, BPIN2 Crystal Violet a CVIN1, CVIN2 Controls - CONTLl, CONTLZ, CONTLB. . . . . . 112 First order plot of ln(C/Co) chromatographic segment #13 crystal violet, and biphenyl on 4 ppm toxaphene over the course of 10 days exposure to UV-light. . . . . . . . . . . . . . . . . . 115 Reconstructed total ion chromatogram (upper) and mass chromatograms representative of C17- isomer composition (middle) and C13- isomer composition (bottom) of toxaphene. The dotted line designates the break between chromatographic segments #8 and #9 . . . . . . . . . . . . . 119 xiv Figure 24. Comparison of the isomer composition of standard toxaphene, 1 ppm toxaphene in distilled water exposed 10 days to UV-light, and 4 ppm toxaphene with fluorene in acetonitrile/water solution after 10 days exposure to UV-light. . . . . .121 Figure 25. Mass chromatograms at m/z=373 representative of [C10H3Clg-Cl] fragment of standard toxaphene (top) and toxaphene in acetonitrile solution with fluorene after 10 days exposure to UV-light (bottom). . . . . . . . . . . . . . . . . . . 123 Figure 26. Photolytic dechlorination and reductive dechlor- ination of Toxicant B (56). . . . . . . . . . 125 Figure 27. Reductive dechlorination and dechlorination by the photolysis of a nonachloro toxaphene component (9' 125) O O O O O O O O O O O O O O O O O O O 127 Figure 28. UV-absorption spectra of biphenyl, fluorene, and crystal violet. . . . . . . . . . . . . . 129 Figure 29. Comparison of unknown spectrum to library spectrum of fluoren-9-one. Top line: unknown mass spectrum Middle line: library mass spectrum of fluoren-9-one. Bottom line: a comparison plot of the two mass spectra. . . . . . . . 131 Figure 30. The photo-oxidation products of fluorene to IIUOI'en'g-One (128) e e e e e e e e e e e e e e 133 Figure 31. Visible absorption spectrum of crystal violet 135 Figure 32. Comparison of the isomer composition of toxaphene residues isolated from Lake Trout bellies and fillets, standard toxaphene and lppm toxaphene in Betsy River water after 10 days exposure to sunlight. . . . . . . . . . . . . . . . . . . 139 XV Section II: Selective Sensitivity of Highly Chlorinated Species in Negative Ion Mass Spectrometry: Implications for Figure 1. Figure 2. Figure 3. Figure 4. Figure 5. Figure 6. Figure 7. Figure 8. Figure 9. Figure 10. Complex Mixture Analysis +CI(a) and NI(b) mass spectra of a cleaned up chicken extract showing the presence of dieldrin (7) O O O O O O O O O O O O O O O O O O O O O O160 Potential energy curves for negative ion fomation (9) O O O O O O O O O O O O O O O O O 163 Negative ion mass spectrum of 50 mg of Lake Ontario lake trout cleaned up by GPC. Only 12C negative mass defect ions are shown for polychlorinated chemicals (32). . . . . . . . 182 Chlorinated bicyclic compounds used to mimic toxaphene constituents: (1)camphene, (2)dichloro- camphene, (3)2,10-dichloronorbornane, (4)aldrin, (5)dieldrin and endrin, (6)heptachlor, (7)hepta- chlorepoxide, (8)cis- and trans- chlordane, (9) nonachlor. . . . . . . . . . . . . . . . . . .194 Molar response curve of chlorinated bicyclic compounds produced by electron impact and negative ion mass spectrometry. Each point represents the average of three replicate injections for each compound. . . . . . . . . . . . . . . . . . . 195 Molar response curve of chlorinated bicyclic compounds produced by capillary GC with Electron capture (EC) detection. . . . . . . . . . . . 197 Molar response curve of chlorinated phenols produced by capillary GC with EC detection. . 201 Molar response curve of chlorinated phenols produced by electron impact and negative ion mass spectrometry. . . . . . . . . . . . . . .202 Molar response curve of chlorinated biphenyls produced by electron impact and negative ion mass spectrometry. . . . . . . . . . . . . . .205 Molar response curve of chlorinated biphenyls produced by capillary GC with EC detection. . 206 xvi Figure 11. Figure 12. Figure 13. Molar response curve of chlorinated biphenyls having single ring substitution; electron impact negative ion mass spectrometry. . . . . . . . 208 Molar response curve of chlorinated biphenyls having symmetrical chlorine substitution on both phenyl rings: electron impact and negative ion mass spectrometry. . . . . . . . . . . . . . 209 A total ion current mass chromatogram and mass chromatogram of the [M-Cl] fragment for aldrin produced by a 1 pl injection of the 10 ppm solution used for NI/MS instrument optimization. This represents the response produced by approximately three nanomoles of each compound 217 xvii Section I INVESTIGATIONS INTO THE FATE OF TOXAPHENE IN NATURAL WATERS Section I Chapter 1 Literature Review of Toxaphene Introduction and Research Objectives Toxaphene is a pesticide composed of a complex chlorinated mixture. It was used in the U.S. from 1946 to 1982 as an insecticide on some 177 different commodities (1). It was produced by Hercules as a broad spectrum insecticide, though its major use was for the control of boll weevil on cotton. Toxaphene was most heavily used in the Southeastern section of the country. Annual use was estimated at 30 to 40 million pounds. Collective use over its tenure as a pesticide, is estimated at 1 billion pounds for the United States alone (2). It was banned in 1982 when studies indicated that toxaphene induced liver carcinomas in rats and mice (3). Restricted use of the pesticide was allowed until the end of 1986 from available supplies. Despite its heavy use, it was considered for many years to be "the most widely used and least understood" pesticide (4). Because it is a mixture, analytical methods development for its detection and confirmation was slow. Residues were not recognized as present in the environment until 1975 (5). 3 Prior to this, toxaphene was considered to be a non-bioaccumlating and non-persistant pesticide (2). As a result of improved analytical procedures, residues identified as toxaphene have been detected world wide in many pristine regions, including the North Seas (6), Bermuda (7) and the Great Lakes (8). However, toxaphene or formulations containing toxaphene had limited use in these regions. These residues are an indication of the ability of toxaphene to spread throughout the environment. However, even at this time little has been learned about its degradation and fate in the environment. The effects of various environmental parameters on toxaphene need to be studied because of the mixture's high toxicity to aquatic organisms and its presence in the Great Lakes basin. Photolytic studies thus far have been conducted only on single toxaphene components in pure solvents under controlled conditions (9, 10). It is important to study the mixture as a whole and its interactions with various environmental compartments. It was the goal of this research to determine if toxaphene exhibits photolytic interactions in natural waters, to characterize and to determine if possible what material present in natural water facilitate these interactions. In addition it was also the intent of these studies to draw some comparison between toxaphene residues weathered in the laboratory versus those extracted from environmental samples. Toxaphene Composition and 4Isolated Components Throughout the following discussion of toxaphene composition, the carbon skeleton of isolated toxaphene components has been referred to as norbornane. This nomenclature, though incorrect by IUPAC standards, was adherred to throughout the discussion because of previous convention by the authors of the original toxaphene papers. These compounds should be referred to as chlorinated bornanes, bornenes and chlorinated bornadienes. The reader is asked to make this correction mentally thoughout the body of the text. Toxaphene is produced by the chlorination of camphene. Technical camphene is dissolved in carbon tetrachloride to which chlorine gas is added to 67 to 69% by weight. The mixture is irradiated with UV light until the yellow chlorine gas is no longer apparent. The resulting mix of indiscriminately chlorinated mono-terpenes is known as ‘toxaphene. It exists at room temperature as an amber waxy solid with a flow point of 82°C (1). When analyzed by capillary GC/MS with electron impact ionization, the mixture Composition was estimated as 76% polychloronorbornanes, 18% Polychloronorbornenes, 2% polychloronornadienes and 4% Chlorinated and non-chlorinated non-identified materials (11) . The total number of individual components have never been actually determined. Estimates range from 177 to 670 5 different possible compounds (12, 13). Holmstead et a1. (12) separated toxaphene into fractions by silica gel chromatography followed by positive chemical ionization (+CI) mass spectrometry to examine the composition of the mixture. Its composition consisted mainly of compounds having the empirical formula of C10H11C17, C10H10C13_ and C10H9C19. The remainder were similarly chlorinated compounds with one or two additional degrees of unsaturation. When the mixture was reduced by the addition of triphenyltin, the major product of the reaction was bornane indicating the majority of the components in the mixture have a norbornane carbon skeleton. This also indicated that the majority of the camphene undergoes a rearrangement reaction during its chlorination to bornane (Figure l). Toxaphene was manufactured for over thirty years by Hercules. Despite that fact, the final product sold as toxaphene appeared to have changed very little from batch to Ibatch. Saleh and Casida analyzed toxaphene batches which had 'been produced over a 26 year period by capillary gas chromatography with electron capture detection (14). The resulting chromatograms were excedingly similar in overall appearance, showing 29 major peaks present in the Chromatogram of which each peak made up some 1 to 8% of the total mixture based on relative peak area. However, if the material produced by Hercules were compared with other Pelychlorocamphene products such as Strobane, a chlorinated Camphene mixture produced overseas, significant pattern + C1' C1 C1 Figure 1. Rearrangment of camphene to bornane as a result of its chlorination 7 changes were observed. The same 29 major peaks were still in evidence, but their relative intensities varied greatly from mixture to mixture. From the original toxaphene mixture only nine compounds have been isolated and characterized. These are shown in Figure 2. The initial intent in isolating these components was to determine the active constituents of the mixture. In other words, the intent was to determine the mechanism of action of toxaphene as a pesticide. 2,2,5-endo,6-exo,8,9,10-heptachloronorbornane or toxicant B, was the first compound which was isolated and structurally identified (15). The technical mixture was fractioned by silica gel chromatography. The resulting chromatographic cuts were purified based on their acute toxicity to mice. Toxicant B demonstrated a six fold increase in toxicity over the technical material. Anagnostopoulos et 31., isolated three compounds identified simply as compound I, II and III from toxaphene by Silica gel chromatography (16). Again these were isolated based on their increased acute toxicity to house flies as compared to that of the technical material. Toxicant A was isolated and identified as composed of two different octachloronorbornane isomers: 2:2,5-endo,6-exo,8,8,9,lo-and 2,2,5-endo,6-exo,8,9,9,10- Figure 2. Isolated components of toxaphene: Toxicant B, 2,2,5-endo,6-exo,8,9,10-heptachloronorbornane (15), Compounds I, II and III (16), Toxicant A: composed of two isomers (17, 18), 2,5,6-exo,8,8,9, 10-heptachlorodihydrocamphene (19, 20), and Toxicant Ac, 2,5,-exo,3,6-endo,8,8,9,10,10-nona- chloronorbornane (21) C1 C1 Cl Cl 1 C Toxicant B Cl Compound I Cl 1 Cl C C]. Cl C1c1 Cl C1 C1 Cl C1 Cl 011 CI Cl 1. C1 Compound II c1 Compound III C1 Cl C1 (:1 Cl c1 Cl ‘31 c1 10 Figure 2. (con't.) Cl c1 Cl Cl Cl C1 + Toxicant A Cl Cl C1 Cl C1 C1 C1 C1 Cl C1 C1 C1 C1 Cl A C1 C1 1. Cl A - C1 Cl Cl 2,5,6-exo,8,8,9,10 ’. c1 heptachlorocamphene C1 Toxicant Ac C1 C1 C1 C1 ll octachloronorbornane (17, 18). Its identity as two isomers was confirmed by 1H-NMR, mass spectrometry, and synthesis of the two isomers from 2-exo,lo-dichloronorbornane. The two components which made up toxicant A were not separable by capillary GC. It was further determined by toxicity tests with mice and goldfish, that the majority of toxaphene's toxicity was due to toxicant A. However, toxicant A was estimated to constitute only 3 to 4% of the total mixture by weight. Matsumura (17) suggested that despite the fact that it is a small portion of the total mixture, it makes up some 56% of the toxicity of the mixture. A compound was isolated from toxaphene by Landrum et a1 (19, 20) and was identified as 2,5,6-exo,8,8,9,lO-hepta- chlorodihydrocamphene. This was considered to be an unusual compound as it did not have the characteristic norbornane structure seen in all other compounds previously isolated from technical toxaphene. It was apparent this compound had not undergone the rearrangement to norbornane but retained its original camphene structure during its chlorination. This suggested that the chlorination of camphene to technical toxaphene was likely to proceed through at least two different pathways (Figures 1 and 3). The second pathway proceeds through the chlorination of the double bound without undergoing rearrangement. The toxicity of the dihyrocamphene component was 4 times greater to mice than the technical material. Toxicant Ac was the last compound isolated which 12 I. , 1O .1 Figure 3. Alternative route for the chlorination of camphene (19, 20). 13 contributed to the toxicity of toxaphene (21). The compounds empirical formula was C10H9C19 and its structure was identified as 2,5-exo,3,6-endo,8,8,9,10,10-nona- chloronorbornane. The toxicity of Toxicant Ac to houseflies was identical to toxaphene while its toxicity to minnows was several times greater than that of the technical mixture. Literature Methods for Toxaphene Analysis and Detection A wide variety of analytical methods for the detection of toxaphene have been developed. But most, until the advent of capillary gas chromatography, suffered from a lack of selectivity in that other organochlorides were indistinguishable from toxaphene as well as having limited sensitivity. The analysis of toxaphene offered something of a challenge because of its nature as a complex mixture. It was necessary that the analytical procedures take the unusually complex composition of toxaphene into account during method development. One of the first methods which broke the milligram barrier for the detection limit of toxaphene was developed by Graupner and Dunn (22). Diphenylamine and toxaphene were reacted with ZnCl to produce a toxaphene diphenylamine complex with a maximum absorbance at 640 nm. The technique was linear from 20 to 700 pg of material but suffered from interferences especially from the matrix material. However, other organochlorine pesticides were fairly easily distiguishable from toxaphene even when reacted as a 1:1 l4 mixture with toxaphene. A second spectrophotometric method was developed in which toxaphene was reacted with ethylcyanoacetate to produce a compound having a red to red- violet color (23). This method also suffered from some of the same problems as the first method, such as interferences with the organochloride pesticide lindane. However, the procedure was linear over two orders of magnitude. The use of a reversed phase TLC plate was explored for toxaphene detection (24). However, toxaphene produced a streak of material rather than a single spot. This streak of material was found to overlap the Rf values of those produced by Arochlor mixtures, DDT’s, and Mirex as well. It was not developed further. Packed column gas chromatography (GC) with electron capture detection (ECD) became the preferred method for detection and quantitation of toxaphene (25). However for the analysis to be effective, extensive cleanup of the sample was required to remove other organochloride pesticides as well as matrix interferences. Polychlorinated biphenyls (PCBs), DDTs, chlordane, lindane, aldrin and their metabolites all co-elute with toxaphene when analyzed by GC with ECD. In addition, the chromatograms of toxaphene by packed column consist of a series of poorly resolved peaks many of which are indistinguishable from the interferences (5). As an example, for a number of years because of its co-elution with the PCB fraction, toxaphene was not recognized as a contaminant in the Great Lakes. In fact, 15 toxaphene residues were mistakenly identified as PCB contamination. It was not until a packed GC column was interfaced with a mass spectrometer that the error in identification was recognized. The limits of spectrophotometric detection, in conjunction with TLC or packed column were still in the pg range (22-25). It was, and still is, unusual to find toxaphene residues in samples at these levels. With the development of capillary GC combined with EC detection, the limit of detection for toxaphene residues in samples dropped dramatically as compared to the other methodologies previously employed. In addition, this technique also made it possible to routinely distiguish toxaphene residues from other chlorinated xenobiotics (13, 26). Despite the advantages of capillary chromatography, cleanup and extraction of toxaphene from sample material still remains tedious and time consuming. Even with the ability to distiguish toxaphene from other interferences, those interferences, especially PCBs are, in many instances, present at higher concentrations than that of the toxaphene residues. Ribick et a1. (26), developed a method which was consistently able to isolate and separate toxaphene residues from biological samples, fish, and interfering organochlorides. The procedure requires the use of several column cleanup steps. Initially, the sample is ground with anhydrous Nazso4, and packed as a dry chromatographic column. Methylene chloride is used to simultaneously elute the fat 16 and xenobiotics. from the sample material. The volume of the solution is reduced by roto-evaporation followed by separation of the fat and pesticide fraction by gel permeation chromatography (GPC). The pesticide fraction is further separated by silica gel column chromatography into PCB containing and pesticide containing fractions. A celelite/activated charcoal column is then used to separate chlordane and toxaphene (27). In fish tissue, residues were quantitated by this method at the part per billion (ppb) level. Mass spectrometry has recently been developed as a tool for the quantitation of toxaphene rather than used for confirmation. Originally, with electron impact (EI) available for ionization, the spectra of toxaphene components produced by this ionization process had extemely complex fragmentation patterns (28). In addition, nothing was gained in either selectivity or sensitivity when electron impact mass spectrometry was used to identify residues. Though complex, one fragment at m/z-159 was produced by the majority of the toxaphene components upon fragmentation (ll, 28). A possible structure for the fragment was suggested to be that of the dichloro-tropylium ion (11, 14, 18). The use of this ion for selected ion monitoring (SIM) was never followed up as it was relatively insensitive and non-specific for toxaphene. Positive chemical ionization (+CI) also produced a high degree of fragmentation and offered no increase in sensitivity when compared to E1 (11, 28). 17 Negative ion (NI) mass spectrometry offered both sensitivity and selectivity over the use of either +CI or EI (13, 26). When EI and NI spectra of Toxicant B were compared, an increase in response of 100 to 1 was noted (26). The spectrum produced by NI is extremely simple, consisting of the molecular ion minus one chlorine atom [M-Cl]. In addition to simple spectra, the NI/Ns ionization process is extremely selective for electrophilic compounds such as chlorinated species. Any remaining material from the original matrix such as lipids or phthalates, are not detectable (26). As a result, NI/MS has been recently developed into a selective and sensitive means for quantitation of toxaphene residues (29-31). For example, ppb levels of toxaphene residues have been detected and quantitated in both air and fish samples. Toxicity Toxaphene is considered to act as a neurotoxin (1). 0f its isolated components, Toxicant Aa or 2,2,5-endo,6-exo,8,9,9,lO-octachloronorbornane, has been shown to be the most effective component of the mixture (17, 18). It acts by competitive inhibition with the t-butylbicyclophosphorthionate (TBPS) to the 1-aminobutric acid (GABA)-regulated chloride channel (32-34). This results in central nervous excitation by stimulating neurotransmitter release (33). The toxaphene mechanism of action is similar to that of picrotoxinin and cyclodiene pesticides, i.e., 18 chlordane, which are all potent TBPS inhibitors. Cyclodienes have been shown to inhibit GABA stimulated uptake in coaxal muscle tissue of the american cockroach (35) and directly compete against [3H]-adihydorpictrotoxinin for binding in rat synaptosomes (33). In additions, insect strains which showed resistancy to cyclodiene type insecticides were also resistant to picrotoxinin. It was suggested that the similarities in their mechanisms of action was due to the structural similarities which existed between picrotoxinin and the cyclodiene pesticides. Though toxaphene is not cyclodiene pesticide, it, too, exhibits some structural similarities to picrotoxinin. All three types of compounds have a similar aliphatic bicyclic structure. In addition, toxaphene and its components were found to inhibit TBPS as effectively as the cyclodienes (32). Turner et a1. investigated a variety of toxaphene components to determine structure toxicity relationships (36). He found that their toxicity decreased in the following order, Toxicant Aa> Toxicant Ab>Toxicant B>3~exo-chloro B, S-exo-chloro B, lo-chloro B. This is well correlated with their effectiveness as‘TBPS inhibitors (32). Toxaphene has been shown to have a variety of additional toxic effects. It produces behavioral abnormalities in rats resulting in impaired learning abilities (37). The liver of rats, after exposure to toxaphene, has a reduced ability to metabolize xenobiotics (38). Toxaphene exposure depresses IgG antibody formation and also decreases the phagocytic 19 capabilities of macrophages in mice (39). Toxaphene has also been shown to be mutagenic by the Ames assay (40) and carcinogenic in rats, mice, dogs, and monkeys (3). Specifically, toxaphene induced liver neoplasms in both sexes of animals at both high and low doses. Though an increased incidence of additional types of neoplasms were observed the greatest increase was in the liver in all species. Aquatic Toxicity Toxaphene is extremely toxic to aquatic organism. In fact because of its toxicity, toxaphene was tested for trash fish control in a variety of lakes (41). Plans were abandoned, however, when it was learned that the pesticide persisted and remained toxic to aquatic organisms for up to 5 years after treatment. Acute toxicity Lcso's have been determined for several species of fish. They range from a minimum of 2.5 ppb for Chinook Salmon (Oncorhynchus kisutch) to a maximum of 20.0 ppb for the guppy (Lebistes reticulatus) (42). In comparison, mammalian LDso are in the mg/kg body weight, or ppm range. For marine fish, such as the three spined stickleback (Casterosteus aculeatus), its Lcso is in the same low ppb range. Acute toxicity has been found to vary with age of the organism with embryonic stages and fry exhibiting the greatest sensitivity to toxaphene. Other aquatic organisms such as daphnids or water fleas, 20 had LCso in the ppb range (43). Two species of shrimp at 1.4 ppb and 4.4 ppb are also acutely sensitive to toxaphene. Oysters have been shown to have a 50% decrease in shell deposition in the presence of 16 ppb of toxaphene (44). Temperature has been shown to have an effect on the toxicity of toxaphene to both fish and daphnids (43, 45). With increase in temperature, daphnids and two species of fish, rainbow trout (Salno gairdneri Richardson) and blue gill (Leponis macrocbirus Rafinesque), demonstrated increased susceptibility to the pesticide (45). This may be due to either increased uptake and/or an increase in metabolic activity in the organism. Chronic toxicity tests were conducted using channel catfish (Ictalurus punctatus) and brook trout (Salvelinus fontinalis) (46). The major effect noted was a decrease in the concentration of backbone collagen. A minimum acceptable toxaphene water concentration was determined based on a no effect level for impaired bone development at 49 to 79 ng/l or ppt (46). The levels of collagen were also correlated with a decrease in whole body vitamin C, an important co-factor in the synthesis of collagen (47). In later studies in which fish were exposed to both toxaphene and vitamin C in their diet many of the bone deficiencies were avoided. It was hypothesized that the detoxification of toxaphene in the fish liver required the presence of vitamin C. At high levels of toxaphene, the detoxification process created a whole body vitamin C deficiency resulting in 21 impaired bone development. Toxaphene has been found to bio-accumulate in a most aquatic organisms to varying degrees. In a model aquatic ecosystem, Blue gills (Lepomis macrochirus) accumulated toxaphene to a concentration 6000 times greater than that of the water concentration (47). Snails (Helisona sp.) accumulated 480 times greater concentrations. Longnose killifish (Fundulus slnllls) (44), a marine organism, demonstrated accumulation ranging from 4200 to 60,000 in body tissue with an accumulation of 1000 to 5500 in the ova. The fact that fairly high doses of toxaphene were present in the ova of the fish indicates that embryos of the fish would be exposed to high levels of toxaphene: concentrations at which developmental problems and death have been observed. It is evident that low concentrations of toxaphene are able to induce toxicological changes in aquatic organisms. Estimated open water concentrations for toxaphene in Lake Huron for 1980 and 1982 were 1.5 ng/l (8). These concentrations are significant if the no effect level and bio-accumulating ability of aquatic organisms is taken into account. Environmental Dynamics In the agricultural environment, the major pathway for the attenuation of toxaphene residues is volatilization. Seiber et al. found that 59% of toxaphene sprayed on cotton 22 foliage was removed after a 28 day period (4). In addition, it was determined that out of the mixture the lower chlorinated species disappeared more quickly than the higher chlorinated species. Volatilization was found to be affected by the ambient temperature, the wind speed, and humidity. Of these, humidity showed the greatest influence. With increased humidity over the cotton foliage, a concomitant increase in toxaphene concentration was observed in the air over the crop (49). Any remaining residue after harvest is attenuated by the drying process (50). This was observed in the case of alfalfa treated with toxaphene exposed to sunlight and UV light. It was concluded that the loss of toxaphene was not due to photolytic degradation. On soil, residues have been shown to translocate and attenuate in various manners. High concentrations of toxaphene have been shown to percolate to ground water over an 85 day period (51). If the soil is flooded, up to 95% of the remaining residue may be anaerobically decomposed within a four week period (52). Of other modes of translocation, runoff accounts for only 0.5 to 1% of the total loss of toxaphene. Runoff accounts for only a small portion of the total mass balance (53). However, only very low levels of residue are necessary to elicit toxic side effects in aquatic organisms, thus runoff remains an important mechanism for distribution into the aquatic environment. Toxaphene persistence and elimination have been extensively studied in both model and natural aquatic 23 ecosystems. In lakes treated with toxaphene for trash fish control, the disappearance and viability of the lakes were monitored for up to a 5 year period (41). It was determined that the levels of toxaphene in the water column decreased slowly over the course of a year. One of the influencing factors appeared to be the presence of aquatic flora (54). Aquatic flora accumulated up to 10,000 times the concentration of toxaphene in the water column with the majority located in the root system. This was also the case in an estuary setting, in which case turtle grass (Spartina alterniflora) also accumulated a large portion of the total toxaphene residue (55). The sedimentation cycle perhaps plays a larger role in the aquatic environment for attenuation of toxaphene residues. Veith and Lee demonstrated this in Wisconsin Lakes treated with toxaphene for rough fish control (56). Over a two year period the majority of the residue was removed from the water column and found in the sediment. Further, it was also shown that the material was being vertically transported down into the sediment layer probably due to the action of benthic organisms. Similarly, in the estuarine system, the majority of the toxaphene was found in association with the micro-organic material horizon (55). Concentrations in this band were found to contain up to 50 times the concentration in surrounding layers. Once toxaphene reaches the sediment, it has been shown to undergo both biological and chemical degradation. An iron 24 redox system such as that found in anaerobic estuarine sediments was able to readily degrade toxaphene to a mixture which exhibited compounds with shorter elution times as analyzed by GC with EC detection (57). Biologically, toxaphene is degraded by both anaerobic and aerobic bacteria to a mixture consisting of lower chlorinated components than that of the original mixture (58). A bacterium from the genus Vibrlo was isolated from estuarine sediment and found to readily dechlorinate toxaphene. The extent of the attentuation however was not determined. Toxaphene has been considered to be photolytically inactive. Limited studies have been conducted on single components of toxaphene in pure solvent systems. In these instances, Toxicant 8 exhibited photolytic dechlorination and dehydrohalogenation in either hydrogen donating or non-hydrogen donating systems (9, 59). Only one study has investigated the photochemistry of the entire mixture. It was adsorbed onto silica gel and was shown not to be effected by UV to any detectable extent (59). No studies have been conducted which try to mimic environmental conditions. Global Distribution Toxaphene is a relatively volatile mixture. Sieber et a1., found that toxaphene sprayed on a cotton canopy will dissipate up to 50% of the material over a 24-hour period. Within another 48 hours, 58% of the pesticide in the soil 25 volatilizes (4). As a result, toxaphene and toxaphene-like materials have been found in a variety of matrices world wide. It has been proposed that toxaphene moves through the environment by volatilization, atmospheric transport and finally returns to earth by rain washout of particulate. Its presence in the atmosphere has been confirmed at several locations. It was first detected in the atmosphere over the Mississippi delta region from 1972 to 1974 (10). This represented contamination over a high use area which came as no surprise. Levels of toxaphene residues were found to reflect the spraying schedule for toxaphene usage on agricultural fields. Residues have also been detected over the open Atlantic Ocean and the Island of Bermuda (7). As toxaphene had not been used for over four years prior to the study on Bermuda, the only source for the residues had to be from the Southeastern U.S. Toxaphene residues have also been detected over Sweden, another region in which toxaphene usage. has been very limited or nonexistent (31). Rain washout has been determined to be the main non-point source for toxaphene as well as a variety of other xenobiotics. Levels of toxaphene in the air and rain water were measured over a South Carolina estuary (60). Residues were found to be 34 to 861 times greater in the rain water as compared to the levels in air. This indicated that the majority of toxaphene resides on particulate which is removed during a period of rain activity. Monthly input into the 26 estuary was estimated as 1.2 kg with the assumption being that the majority of the toxaphene input into the estuary was a result of rainfall. As a result of these cycling processes, toxaphene has been detected throughout the globe. In several global base line studies conducted to determine the extent of organochloride contamination, toxaphene as well as a variety of other organochloride pesticides and xenobiotics were found. These were detected not only in heavily industrialized regions but in pristine environments such as the Tyrollian Alps, and Antarctic (6). Toxaphene has also been detected in fish throughout the Great Lakes though the pesticide received limited use in the region (8). On Isle Royale, fish taken from Siskiwit Lake which is land locked and has no water exchange with Lake Superior, show residues of toxaphene at the same or greater level than those from Lake Superior. Toxaphene has spread throughout the trophic levels even to the human population as its residues have been detected in human breast milk (61). 27 Section I Chapter 2 Literature Review of Humic Materials and their Interactions with Xenobiotics Introduction Humic materials are ubiquitous macromolecules occurring in soil and water. They are formed as a result of the decomposition of biological materials, especially lignins, by bacteria (62). The result is a hydrophilic organic molecule with a molecular weight estimated to be from 300 to greater than 300,000. Humic molecules have a polynuclear aromatic core with phenolic, carboxylic and aliphatic functional groups on its exterior. Humics are capable of absorbing sunlight thereby turning waters high in humic materials to a brown color. All natural waters contain humics to various degrees. Concentrations of humics in open lake water range from 5 to 20 mg/liter while bog and estuarine waters have from 60 to 100 mg/liter (63). Though the source of humic materials has not been confirmed, it appears that their presence is a result of leeching from surrounding soils (63). Because of their ubiquitous nature, it would be expected 28 that humic materials will influence the fate of trace organic contaminants. Within the last decade, this has indeed been shown to be the case. For example, numerous instances of increased apparent solubility (64) and altered photochemistry (65, 66) as a result of interactions with humic materials in both soil and water have been discussed in the literature. It is necessary to investigate the interactions of humics with various xenobiotics so that an understanding of the overall fate of xenobiotics in the environment may be reached. structural Characteristics of Humic Materials Humic materials are formed by the bacterial decomposition of lignins, carbohydrates, and proteins (62, 67). Of these three constituents, lignins are considered to be the most important as proteins and carbohydrates decompose more readily. The resulting humic material has been divided into three classes based on their solubility in NaOH. These are: 1. Fulvic Acids-(FA)-Soluble in both dilute basic and acidic solutions. 2. Humic Acids-(HA)- Soluble in dilute base but not in dilute acid. 29 3. Humins- insoluble in both dilute base and acid(62,67). These are functional definitions and are not necessarily an indication of structural differences. The molecular weight of humic materials is variable and dependent upon its source and site of collection. HAs have reported molecular weights ranging from 300 to >200,000. HAS and FAs isolated from marine sediment range from 700 to 200,000 (67). Humic materials isolated from water sources tend to have a lower molecular weight than their soil and sediment counterparts. These range from an average molecular weight of 850 to 1490 g/mole (68). This is to be expected if the source of aquatic humics substances (AHS) is indeed that of soil leechate (63). Lower molecular weight species and therefore more water soluble species of material would be more likely to be found in water than the higher molecular weight species. of aquatic humic substances, FA constitute 90% of the bulk, and tend towards a lower molecular weight than HA. Humic substances contain mainly C, H, N, S and O as the major elements, with C and O predominanting.‘ The composition of soil HA has been determined as: C = 53.8-58.7%, 0 8 32.8-38.3%, H = 3.2-6.2%, N = 0.8-4.3%, and S = 0.1-1.5%. Soil FA composition is as follows: C = 40.7 -50.6%, 0 3 39.7-49.8%, H = 3.8-7.0%, N = 0.9%-3.7 and S = O.l-36%. Water associated humic materials tend to contain less C and N 30 but increased H indicating a lesser degree of condensation as aromatic species (67, 69). The major oxygen-containing functional groups identified in humic substances are carboxyls, hydroxyls, and carbonyls. FAs have a greater total acidity than HAs, as determined by phenols + carboxylic acids. FA also have a greater number of aliphatic hydroxyl groups than HA. Carboxyl functionals for both FA and HA are made up of quinoid and ketone type functional groups (69). The oxidative degradation of humic materials has lent some insight into the structural characteristics of humic materials. Three major types of degradation products have been identified in degradation studies of humic materials. These are aliphatics, substituted benzenes and polycyclic aromatic hydrocarbons (PAHs). Aliphatic products consist of n-alkanes, n-alkanoic acids, fatty acids and alkanedicarboxlic acids. Substituted benzenes include polyphenols, polycarboxylic acids, p-toluenesulfonamides, polyhydroxytoluenes, etc. PAHs isolated from humic degradation mixtures have from 2 to 7 fused rings. These may or may not have substituted side chains. For an illustration of the myriad of compounds isolated from oxidized humics, the reader is directed to table 3 in G. G. Choudhry's book on humic materials (69). In general, humics are considered to have a polynuclear core with aliphatic substituents and oxygen functional groups on the exterior of the molecule. 31 Photophysical Characteristics of Humic Materials In general, humic materials yield uncharacteristic spectra in the visible (400 to 800 nm) and ultraviolet regions (200 to 400 nm). Solutions of humic materials produce spectra which show a decrease in molar absorptivity with increasing wavelength. The spectra are featureless overall, exhibiting no maxima or minima (67, 70) Light absorption by humic materials appear to increase with an increase in [1] the degree of condensation of the aromatic rings in its structure, [2] the ratio of aromatic to aliphatic carbon, and [3] the total carbon content and molecular weight (68, 71). The absorption of light by a humic molecule also is affected by pH. 'With an increase in pH, the absorbance and also the apparent color of humic solution increases simultaneously. This increase is attributed to the ionization of acidic functional groups present on the molecule. Humic materials are bleached by long term UV (254 nm) irradiation (72). This is accompanied by a decrease in organic carbon in solution, occurring especially in the presence of oxygen. Photo-oxidation then is an important degradative process for humic materials. Non-irradiated humic materials were also found to be more photoreactive and have a more distinctive UV absorption than irradiated humic materials. 32 Binding of zenobiotics by Humic Materials Humic materials can bind hydrophobic xenobiotics and, in doing so, increase the apparent water solubility of the xenobiotic. Shinozuka at al. (64) found that solubility of n-alkanes and PAHs (in solution with humic materials), was governed by two different mechanisms. At low concentrations of humic materials, it is simply a direct absorption of the hydrophobic molecule by the hydrophobic portion of the particle. At higher concentrations of humics, a micellular mechanism appears to account for the observed results (62). The interactions of humic materials and xenbiotics are affected by the pH of the solution. With an increase in pH, the amounts of bound atrazine (73) and DDT (74) were found to decrease. However, eicosane and anthracene complexation increased with increasing pH (64). The effect of pH can be attributed to the degree to which the humic particle is in its ionic form. At high pH, the phenolic and carboxylic acid groups are in their ionic form and therefore would be less likely to interact with hydrophobic molecules (74). In the case of eicosane and anthracene, the concentrations of humics used for these studies (64) were much greater than those used for the DDT study (74). A high pH influences the formation of humic aggregates and thus the increase in eicosane and anthracene solubility is probably due to a micellular mechanism (75). The ability to bind a xenobiotic molecule is also 33 influenced by the solution's ionic strength (74). Though the effect of ionic strength is not as dramatic at that of pH, an increase in ionic strength increases the ability of the humic material to bind a xenobiotic molecule (74). The changes indicate that the humic polymer becomes less hydrophilic as the ionic strength is increased and therefore is more likely to interact with hydrophobic molecules. V The binding of xenobiotics by humic materials can be described by its partition coefficient Kc(74), where, Rc= WW [ Compound in aqueous solution ] In sediment high in humic materials, Kc is directly proportional to the organic carbon content of the solution (74, 76). However as Kc increases with increasing organic carbon in sediment (76), it decreases in aqueous solutions (74, 77-79). It was suggested that the decrease in Kc with increased dissolved organic carbon in solution, was a result of aggregate formation by the humic materials (77, 79). A compound’s K, is directly correlated with its octanol/water partition coefficient or K09, where K0. = [compound in water] In sediments, estimates for Kc were calculated based on Kow (76). At low concentrations of humic materials, Kc was approximately equal to Rev (78-80) and may be used as method for the prediction of the capacity of a molecule to bind with 34 humic materials. The fraction of humic materials most likely to bind organic molecules consists of the high molecular weight components (81). In studies conducted with humics and cholesterol, the majority of the cholesterol was found in conjunction with the high molecular weight portion of the humic material (81). The precise mechanism which accounts for the affinity of humic material for organic molecules has not been fully elucidated. However, there is evidence which suggests that the formation of a charge-transfer complex between the dissolved humics and a organic molecule occurs. Electron donating structures such as hydroquinones, ethers, alcohols, nitrogen-containing groups and phenolic moieties have been identified in humic materials. These functional groups can form electron/donor acceptor complexes (charge-transfer complexes) with electron acceptors such as benzoquinones and many chlorinated compounds in solution. Such complexes have been indirectly observed by UV-difference spectra of the interaction between tetrachlorobenzoquinone, 2,5-dibenzoquinone and humics where a red shift by the maximum absorbance peak of tetrachlorobenzoquinone at 290 nm was observed. The electron affinity of 2,5-dichlorobenzoquinone was less than that of its tetrachloro isomer, and demonstrated a lesser red shift when interacting with humic materials (79). However, with decreased halogen substitution, a molecule’s K," decreases 35 with a concomitant decrease in K¢(76). Similar complexes were observed by infra red (IR) and electron spin resonance (ESR) spectroscopy in solutions containing soil humics and s-triazine (82) and substituted urea herbicides (83). In general, humic materials, because of their organic character, can interact and absorb organic molecules. By this mechanism, the apparent solubility of the molecule and its capacity to react or be degraded, may be altered. Photochemical Characteristics of Humics Aquatic humic substances (AHS) influence photolysis rates of xenobiotics in the aquatic environment through the attenuation of sunlight, and the initiation of indirect photoprocesses through physical or chemical interactions which alter the speciation or microenvironment of pollutants. AHS absorb light across the spectrum of sunlight while many pesticides and xenobiotics absorb in the high UV region. Figure 4 shows the absorption spectra of sunlight, the pesticide propylenethioura (PTU) and humic materials (84). It is because of the overlapping absorption spectra of the humic materials that indirect photoprocesses can occurStudies have indicated that AHS are capable of influencing several different types of direct and indirect photoreactions. The scheme in Figure 5 adopted from Zepp et a1. (65), shows the various pathways which have been demonstrated to occur between humic materials and xenobiotics. 36 Figure 4: Comparison of the absorption spectra of natural sunlight, humic acids and propylenethiourea (84). 37 2:. .35 o. 23.! Run 25:: .I .36 to En l... .o 8625.5 .55 56:22.»; 8.. can con own 8.” com com com SN 8N - p h b p P u b L p cow on. 41/ I. .1 £9.55 E: can ~Eu DE 2.9. 3293 n 1 x3: q d d 11 d d J 4 41d c205 can con 03” Own 80 com own Ova 98 .3 r 3 f3 Ll .. ad f .8 r 06 1 ed r 5.6 1 ad 1 ad r o.— .. 3:352? -l d 22.5% .o Essooom 38 Figure 5: Various reaction pathways of humic materials which have been shown to influence the rate of degradation of xenobiotics (65). 39 as ——> s (2) as. + Q _, s (4) 33' + RH ——> I V Products RH (5) s + 31m SH' + R“ S‘. + R+. s+- + R'- 102 ——> 02 (6) 10"4-1). ""'—’ A02 (7) 40 Light is absorbed by the humic material S, which is excited to its singlet electronic state (eq. 1). The molecule in its singlet state, being a short-lived phenomenon of approximately 10 ns., is unlikely to react with a xenobiotic. By intersystem crossing, some of the population in the singlet electronic state is converted to its triplet excited state, while the remainder of the population in the singlet excited state, reverts to the ground electronic state. The triplet state is a much longer—lived state, on the order of microseconds to seconds, and is therefore much more likely to react with other components in the aquatic system. The triplet state of the humic material can be quenched by a variety of materials that are also present in an aquatic system. Some may revert to the ground state by either phosphorescence or non-radiative decay (eq. 2). The triplet state may interact with oxygen in the solution, resulting in an energy transfer to molecular oxygen thereby producing singlet oxygen (eq. 3). The triplet state of the humic material may be quenched by other components present in natural waters such as polynuclear aromatic hydrocarbons (PAHs) (Q)(eq. 4). Humics, alternatively, may interact with a pollutant (RH) to give reactive intermediates (I) which may either yield products or revert to their ground electronic state. Intermediates might include (a) direct energy transfer from the triplet state of the sensitizer to the triplet state of the molecule, (b) formation of radical 41 intermediates or (c and d) facilitate the oxidation or reduction of the pollutant. Singlet oxygen is a much stronger oxidant than the ground state molecule. Thus, singlet oxygen may be capable of inducing a variety of reactions in its own right (eq. 6 and 7). Singlet oxygen, once formed, may either revert back to its ground state or react with an acceptor (A) to form peroxidized or oxidized species. Phototransients produced by the UV-irradiation of ABS have been detected by direct observation of their lifetimes by laser flash photolysis. Two independent groups utilizing this technique where able to observe two phototransients produced by irradiation of humic-containing solutions (85-88). The first, a long-lived transient with a A max at 475 nm, was attributed to the triplet state of the humics. The second transient was identified as a solvated electron by quenching experiments with nitrous oxide which is known to react with solvated electrons. Its A max was at 720 nm with a lifetime of 10 us (87). The transient attributed to the triplet state of the humic materials had a greater overall excited state absorbence than that of the solvated electron (85, 87). Moreover, this same transient was readily quenched by oxygen, a known triplet quencher, giving further evidence of its triplet character. From further quenching studies using commercially-obtained humic materials, an estimated triplet energy for HS was established of 57 kcals/mole (87). A quantum yield of 0.07 was determined for the solvated 42 electron in natural water from Searsville Lake Water (NW) . A rate of formation of 3.50 x 10'3moles/sec was calculated for this same natural water sample (87). Frimmel et a1. (86) also observed the quenching effect of transition metals on the transients. Ni(II), Co(II), and Cu(II) were all effective quenchers of both transients. The authors speculated that the quenching effect was due to the complexation of the metals by the humic materials. The results of the laser flash photolysis suggest that an energy transfer occurs between the long-lived excited triplet state of AHS to oxygen (85, 87). As a result of the interaction, singlet oxygen is produced. Steady state concentrations of singlet oxygen have been measured in a variety of fresh and salt waters (89, 90). The steady state concentrations as measured under natural sunlight, all fell within the 10‘13 M range. These varied from a minimum of 0.4 x 10'13 M in clear waters with little humics, to a maximum of 28 x 10’13 M in waters with high concentrations of humic materials. The concentration of singlet Oz/mg of ABS was consistent across all waters, indicating a direct interrelationship between humics and the formation of singlet oxygen (102). It was noted that although the surface concentrations of 102 were the greatest in highly colored waters, i.e. high [AHS], the overall concentration at greater depths dropped significantly due to light attenuation by the ABS. Fractionation of the AHS by gel permeation Chromatography followed by subsequent irradiation showed that 43 the lower molecular weight materials appeared to produce the most 102. Steady state concentrations of hydrated electrons (eaq) have been measured in natural waters and solutions of commercial and natural humic concentrates. These were determined to be in the 10"18 M range (87, 91) based on the NO;‘ formation from added nitrate. Further, based on calculations of the diffusion coefficient of the hydrated electron of 4.8 x 10'5 cmz/sec, the average distance travelled by the hydrated electron was estimated at 1.3nm. This suggests that the solvated electron appears to remain confined to the humic particle of its origin and, thus, is unlikely to participate in a direct interaction with micropollutants (91). As a result of these two phototransients, that of the hydrated electron and the triplet state, a variety of reactive oxygen species occur in natural waters at concentrations which are likely to affect micropollutants. These species and their concentrations in natural waters are shown in Table 1 (92). 44 TABLE 1: concentrations of Reactive Oxygen Species in _Nstural_!at2ra_ Species. algcncsntrafign_lhl 'OH 10‘16 -10'17 R00- 10'9 102 10-13 Ro- 10'1‘ cho'3 10‘14 ‘The formation of these reactive species appears to be dependent upon the functional groups, the molecular weight and the degree to which the humic material is either "condensed" or aliphatic. If commercial humics are fractionated by molecular weight, that fraction having the highest absorptivity at 366nm indicating a high degree of aromaticity, also exhibited the greatest rate of reaction with cumene. This fraction also corresponded to the highest molecular weight portion of the humic materials (92, 93). The presence of nitrate in the water also plays a role in the production of free radical species (92, 93). With the addition of nitrate to solutions containing humics, an increase in the concentration of free radicals is observed. Nitrate is present in all natural waters (94) and is therefore a likely candidate for such interactions. However, Haag and Hoigne (95) suggest that at a steady state concentration of less than 10'15 to 10'17 M for -OH is 45 of insufficient quantity to be involved in the attenuation of micropollutants. Reactions indicative of these various reactive species have been observed in solutions containing both probe and xenobiotic molecules Photo-oxidation: Photo-oxidation in a natural water sample solution was first demonstrated to occur by Ross and Crosby (96). Solutions of aldrin in sterilized rice paddy water were exposed to UV and natural sunlight. Within 36 hours, 25% of the total aldrin in solution was oxidized to dieldrin while the corresponding distilled water solutions showed no oxidation products. It was the first study in which the importance of using natural water samples to determine the actual environmental stability of pesticides was demonstrated. Previously, the usual method for evaluating environmental stability had been the irradiation of the xenobiotic in distilled water alone. The molecule 2,5-dimethylfuran (DMF) has been used as a probe to determine the mechanism by which photo-oxidation proceeds (97, 98). Singlet oxygen rapidly reacts with DMF to form cls,1,2-diacetylethylene (97) by the reaction shown (eq. 8) in Figure 6. DMF was found to rapidly photo-oxidize in natural waters or distilled water with added humic material. Solutions of 46 o: o: .Akm.smV somsxc Hosmasm so can so casuavsxo no assess _ no or _ \ / n so o on: «no on: o 47 DMF in distilled water did not produce any detectable change in concentration. To further confirm the importance of singlet oxygen in the reaction, 1,4-diazabicyclo[2,2,2]octane (DABCO) was added to deactivate 102 without reacting with singlet oxygen (98). The addition of DABCO to the solution of DMF in natural water appeared to completely inhibit the photo-oxidation of DMF. By contrast, the addition of D20 to solutions of DMF in natural water, facilitated the photo-oxidation of DMF. This was the expected result as D20 increases the lifetime of singlet oxygen and therefore increases the rate of the photo-oxidation by maintaining a higher steady state concentration of singlet oxygen (98). Similar reactions have been observed for sulfur containing compounds (84, 97, 99, 100) in solutions containing humic materials. Disulfoton, a sulfur containing pesticide, was oxidized to disulfoton sulfoxide in natural water solutions (eq.8)(97) by a singlet oxygen intermediate. 3 S 0 II II II Similarly, coastal sea waters were examined for their capacity to photo-oxidize dimethysulfide (DMS), a major component of the sulfur flux from the ocean to the atmosphere (99). Comparisons were made between seawater, and solutions containing added humics or rose bengal, a known singlet oxygen-producing sensitizer. The amount of oxygen consumed 48 over the course of the experiment in all three types of solutions indicated a 2:1 stoichiometry of DMS to 02, consistent with the reaction: (eq. 9). 2DMS + 101," ——> DMSO (9) Similarly, the oxidation of diethysulfide to diethylsulfoxide was studied in the presence of rose bengal in water. Its rate of oxidation was not enhanced by the addition of the sensitizer indicating that hydrogen bonding with water affects the oxidation reaction (100). Photo-oxidation mediated by AHS has been shown to be of importance not only for the elimination of micropollutants but also in the fixation of oxygen in salt and fresh water (101). Fixation of oxygen was found to increase with an increase in the concentration of AHS. Miles and Brezonik (102) observed a similar phenomenon in fresh water systems high in AHS. They found that the fixation of oxygen was accomplished by the catalytic effect of the ferrous-ferric cycle in water in which humic materials play a major role. Increasing pH, concentration of AHS and iron all increased the rate of oxygen fixation. Energy Transfer: The degradation of xenobiotics has also been shown to proceed through direct energy transfer from either humic 49 materials or singlet oxygen. Both of these energetic species have been observed in natural water systems. The rate of photomineralization of some chlorinated phenols has been shown to increase in the presence of AHS. The isomer, 2,4-dichlorophenol photolyzed at a significantly higher rate in estuarine water than in distilled water, indicating a photosensitized process (103). The mechanism involved in the photosensitized reaction of chlorinated phenols has been identified by Scully and Hoigne (100). Rose bengal was used as a sensitizer for the reaction in aqueous solutions. Rates of reaction were enhanced in the presence of rose bengal and varied with pH. With an increase in pH and, therefore, an increase in the concentration of phenol present in solution as the phenoxy anion, the rate of reaction was enhanced. cls-1,3-pentadiene in the presence of AHS in natural water samples converts to its trans-isomer (eq. 10)(97, 104). The isomerization reaction is known to occur in the presence of triplet sensitizers. K > .— (11) H C ""CH ‘1 3 2 CH Equation 10: The interconversion ofcis—1,3-pentadiene to its trans- isomer. 50 The rate of the conversion reaction was found to increase with an increased in concentration of dissolved humics. The rate of conversion was also affected negatively by the addition of oxygen to the solution. This would confirm that the triplet energy state of the humic material is facilitating the reaction. The steady state concentration of the triplet state of humic materials in natural waters was calculated based on this reaction. It was found to vary from 10’15 to 10'13 M for lake water to bog water but was consistent if based on the amount of AHS in the water. However, the majority of the triplet states appear to interact with oxygen to form singlet oxygen. The herbicides 2,4,5-T and picloram have also been shown to undergo sensitized reactions in water containing humics (105). As with the chlorinated phenols (100, 103), the degree to which picloram and 2,4,5-T were ionized affected the rate of the reaction. The concentration of humic materials also affected the rate of the reaction. It was proposed that the photoreaction induced by humic substances involved an electron transfer from the carboxylate group of the herbicide to the humic material or oxidizing species produced by the humic material. The half lives for DDT and methoxychlor are much reduced in natural water solutions (106) if compared to those in distilled water solutions (107). In distilled water, the half-lives of DDT and methoxychlor were 150 years and 4.5 51 months respectively (106). These were reduced to 2-5 hours in natural water solutions. This was not attributed to a direct energy transfer from the humic material to the pesticide as the triplet energy of methoxychlor was measured at 80 kcals/mole. In comparison, that of humic materials is estimated at 57 kcals/mole (87). Only sensitizers having a triplet energy of greater than or equal to 80 kcals/mole would be capable of sensitizing the reaction. The attack of free radicals generated by the humics appears to be responsible for the decomposition of DDT and methoxychlor. The explosive, 2,4,6-trinitrotoluene (TNT), is photolyzed 10 to 100 times more rapidly in natural water by sunlight than in distilled water (108). The mechanism suggested was that of a triplet intermediate as the same reaction was sensitized by a triplet photosensitizer. Solutions containing the sensitizer were more efficient at the photolysis of TNT, but was a result of the high efficiency of the intersystem crossing from singlet to triplet state for the photosensitizers. Attenuation and Quenching Effects: Because of their capacity to absorb sunlight in the UV-region, AHS also act as quenchers of reactions or reduce the rates of reaction by the preferential absorption of incident light. For example, the apparent rate of photolysis in distilled water for the dimerization of 52 3,4-dichloroaniline is more rapid that the reaction in natural waters containing high concentrations of organic materials (109). Aniline and chloroaniline half lives were determined in estuarine and distilled solutions. Half lives were increased by at least a factor of two in estuarine waters from those in distilled water (110). Quenching of the excited state intermediate by the humic materials was suggested to be the cause. Despite this, photolysis is still considered to be the major pathway for the elimination of aniline from the environment. The photolysis of atrazine in aqueous solutions was slower in presence of fulvic acid than in distilled water alone (111). It was hypothesized that the reduced rate of reaction was due to the binding of the herbicide to the humic materials. However, in the presence of fulvic acids, four products were observed which were not identified in distilled water (Figure 7). Initially, fulvic acids retard the reaction, but in the final analysis, the photolytic decomposition is more extensive in natural waters than in distilled water. Photoincorporation of Xenobiotics in Humic Materials: In many of the above studies, the mass balance of starting material and product were not consistent with the original quantity of starting material. There is some evidence which suggests that some xenobiotics may be removed 53 Figure 7. Photolysis products of the herbicide atrazine, identified in aqueous solutions in fulvic acid (111). 54 elm . 2 m... I% w 1% I \ N Kw... NNIAm m C N IIHA \ IA“... w :A/ \ N I N IAN IV III 55 from the water column by photoincorporation with humic materials (112, 113). A series of polyphenolic and polycarboxylic compounds were used to mimic humic materials. These compounds were selected because they have been detected as a degradation product of humic materials. These were exposed to UV light along with chlorobenzenes in acetonitrile/water solutions. All chlorobenzenes with the exception of hexachlorobenzene yielded products indicative of incorporation products. The reaction proceeds through a photolytically generated polychlorphenyl radical which is then incorporated into the humic monomer (eq. 11 and 12). c“ 'H' P . + G . R1-4 R1-4 Cln-l 1n-l 1-3 0 c ‘H' IS )| + 9 * O Q n-l OH 1’3 C Cl (eq. 12) R - 1-cooa 2-cno 3-ocn3 4-0H There appears to be some specificity associated with the 56 reaction with respect to the orientation of the substitution on the humic monomer. Products were observed which were a result of para- and ortho- orientation with respect to the. polychlorphenyl radical. Summary Humic materials are capable of initiating a number of different indirect photochemical processes. These processes are apparently a result of two phototransients identified as the humic triplet energy state and the hydrated electron. These two transients generate reactive oxygen species such as peroxides and singlet oxygen or directly interact with xenobiotics The photolysis of pesticides and xenobiotics in the solution phase is a major degradation and detoxification pathway. However, most studies have not anticipated the effects that naturally occurring substances might have on a compound of interest. In some cases, the products of the reaction differ from one another depending on whether natural water or distilled water is used for the experiment. Rates of reaction and half-lives for compounds have been observed to increase and decrease in natural waters. Thus it is of extreme importance that these naturally-occurring organic molecules be taken into consideration when evaluating the photolytic reactivity of a specific compound. The photolysis of toxaphene, as mentioned in Chapter 1, 57 has only been studied as a mixture, adsorbed on silica gel. Solution phase photochemistry has been limited to the study of the photolysis of single components of the mixture. Thus, to determine the fate of toxaphene in the aquatic environment, it is necessary to utilize those naturally occurring AHSs in conjunction with aqueous photolytic experiments. In this manner a clearer picture of the fate of toxaphene in the aquatic environment might be reached. 58 Section I Chapter 3 Materials and Methods for Evaluating the Effect of Weathering on the Complex Mixture Toxaphene The objectives of the research were as follows. First it was the intent to determine if toxaphene was photolytically labile in natural water solutions. Secondly, studies were conducted to determine the effects of various components of natural waters in order to evaluate their effect on toxphene. To gain further insight into photochemical processes, those occuring in natural waters were compared with triplet sensitized solutions of toxaphene. Finally, a comparison between residues isolated from environmental samples to those residues weathered in the laboratory was conducted. Experiments were carried out in several phases. The first phase investigated the effect of natural sunlight and UV-light on toxaphene in three natural water solutions at three concentrations. Secondly, humic materials were isolated from natural water sources and used to mimic natural water systems by their addition to distilled water. The effects of pH and the presence of Fe(III) were investigated 59 as well as the effect of humic concentration in aqueous solutions. The third phase compared natural sensitizers to known chemical sensitizers in order to determine if the effects in theses two systems were similar. These simulated media were exposed to UV-light and natural sunlight. Standards All toxaphene used thoughout these experiments, was from batch #XB 746 and was purchased from City Chemical Corp., New York, NY. The internal standard, 2,4’,5-trichlorobiphenyl was purchased from RFR, Corp., Hope, RI. Humic Acid, sodium salt, was purchased from Aldrich Chemical Co., Milwaukee, WI. NaCl used for the extraction of toxaphene was obtained from Mallinckrodt, Paris, KY. External standards used to calibrate the mass spectrometer for NI/MS were from the following sources. Aldrin and Heptachlor were received as standard from the EPA: trans-chlordane and nonachlor were generously donated by Dr. F. Matsumura. Solvents, Chemicals and Resins Solvents: Hexane used for extraction and sample preparation was of glass distilled grade and obtained from J. T. Baker, 60 Phillipsburg, N.J. Acetonitrile and methanol were also from J. T. Baker. Pyridine was purchased from Mallikrodt, Inc., Paris, KY. All solvents with the exception of methanol were used as received. Methanol was doubly distilled over magnesium metal prior to its use. Sensitizers and Actinometer chemicals: Biphenyl, flourene and crystal violet were purchased from J. T. Baker, Phillipsburg, NJ. 4’-nitroacetophenone, 97% purity, was purchased from Aldrich Chemical Co., Inc., Milwaukee, WI. Ion Exchange Resins: Amberlite MAD-2 and Amberlite IR-45CP were both purchased from Rohm and Haas, Philadelphia, PA. Collection and Storage of Natural Water Samples Waters high in aquatic humic substances and water from Lake Michigan as a comparison, were collected from four different sources. These were the Tahquamenon River, the Two-Hearted River, the Betsy River and Blind Sucker Flooding and Lake Michigan off of Norwood, Michigan (Fig. 8). The headwaters of all these bodies of water, with the exception of Lake Michigan, begin in the bogs of the Upper Pennisula of 61 5 .. 2. mg 2'2“? me: 3:13 3:2 .228 (DE-ma: 22.; C.-o'°; 04:3ch 375.2“;0 comic-:3 00):: =3 3: an :7 ,3 ‘2 033153: Ozmn-I-m #636566 Figure 8. Collection sites for natural water samples around the state of Michigan. All samples from the Upper Pennisula were high in humic materials. 62 Midhigan and flow to Lake Superior. Rivers from such sources have been shown to have from 40 to 60 mg/l of dissolved organic carbon, as aquatic humic substances (63). The waters from these humic containing sources, appeared brown in color. Four to 16 liters of water were collected in brown glass one gallon bottles and stored at 4°C until needed. The pH and conductivity of the waters were measured using a PHH-80 pH/Conductivity meter (Omega Engineering, Inc., Stamford CT). The waters, prior to their use, were filtered through a Whatman GFC glass fiber filter (Whatman Inc., England) to remove particulate and bacteria. Esimated pore size for these filters is 0.4 pm. Preparation of MAD-2 Resin for Humic Extraction Prior to use, the_XAD-2 resin used for the concentration of humics from water was prepared in the following manner. The resin was soaked repeditively in methanol to remove impurites over a 24 hr period. It was then extracted with methanol in soxhlet extractor for 18 hrs and rinsed with clean methanol. A chromatographic column 25 mm x 250 mm of this resin was prepared by plugging the column bottom with glass wool. The resin was introduced into the column and topped with 2 cm of clean sand. The column, equipped with a 500ml reservoir was washed with 3 liters of distilled water to remove any remaining methanol. 63 Isolation of Aquatic Humic Substances from Natural Waters After removal of the particulate as described above, the pH of the water was adjusted to pH 2 or less by the addition of concentrated HCl. If the pH were not adjusted, the humic material would pass through the column without adhering to the resin. The acidified water was introduced onto the prepared resin column. Up to four liters could be passed through the column at one time before the flow was slowed by the accumulation of humic materials on the resin. All color appeared to be removed from the water once it had been passed over the resin bed. The flow rate of the column was maintained between 5 and 10 ml/min. No external pressure was applied. To remove the aquatic humic substances from the column, 0.1 N NaOH was added to the column. The humics eluted as a brown band of material. The brown liquid was collected in 250 ml amber bottles until no further color was detected. The brown solution was again acidified by the addition of concentrated HCl. To remove any other material remaining on the MAD-2 column, the column was stripped with methanol and again rinsed with 3 liters of distilled water. The concentrated solution of humics was introduced onto a second ion exchange column to acid saturate the humics and remove salts. This followed the method of Thurman and Malcolm (114) with the exception of the resin employed. The ion-exchange resin used for the hydrogen saturation process 64 was IR120 HCP (Rohm and Haas, Philadelphia, PA). Initially, 200mls of 0.1N HCl was passed through the column. The column was then rinsed with distilled water until the pH of the eluent was neutral as determined by pH paper. The acidic humic concentrate was introduced onto the column and was eluted with distilled water. The eluent was acidic only if humic materials were present. Once the pH of the eluent was neutral, no further material was collected as this indicated no futher aquatic humic substances were present. The hydrogen-saturated humics were then frozen at -20°C. These frozen concentrates were then freeze dried over a 24 to 48 hr period in a Vitrin Unitrap II freeze drier. The humics at this point were a light brown powdery material. The prepared humic substances were stored in the freezer at -20°C until needed. Sunlight Experiments Samples of various natural waters were prepared as described above. The sample tubes were placed on a rack outdoors at a distance of some 50 yards from the south side of the Pesticide Research Center. The rack was constructed such that the samples hung at approximately 30° to the horizon. Temperature was monitored by a Hygro-Thermograph (Belfort Instrument Co., Baltimore , MD) with the ability to monitor ambient temperature for up to one week. The sunlight intensity was monitored by the actinometer solution described below. Large-volume samples were exposed simultaneously and 65 extracted after ten days. Four milliliter samples were collected after 6 hours, 1 day, 2 days, 6 days and 10 days exposure to sunlight. This sampling protocol is based on that described by Mabey et al. (115). Indoor Irradiation of Aqueous Toxaphene Solutions Three of the natural water samples which had been collected were used for these experiments. These were the Lake Michigan water, Betsy River water, and Two-Hearted River water. The Lake Michigan water was very clear, having little humic content while the other two waters were brown in color due to their high aqueous humic content. Distilled water was used as a control in these experiments. All waters were filtered using Whatman GFC glass fiber filters with an estimated 0.4pm pore size. These were stored at 4°C in brown glass bottles until needed. The waters were spiked with toxaphene by the addition of appropriate volumes of a 10,000 ppm solution of toxaphene in methanol. Concentrations of toxaphene were adjusted to 10, 1.0, and 0.1 ppm. Methanol was added as carrier solvent to a maximum concentration of 1% to ensure solubilization of the pesticide. Four milliliter aliquots of the pesticide mixture were introduced into 13 mm X 100 mm teflon lined screw capped pyrex tubes which had been extensively cleaned prior to their use. The tubes had been boiled once in detergent to remove organics, followed by two eight hour rinses of boiling distilled water to remove any soap residue. These were then 66 baked for 12 hours at 500°C. The sample tubes were placed into a Merry-Go—Round reactor (Rayonet Type RS preparative photochemical reactor, The South New England Ultraviolet, Co., Middle Town Ct.) containing mercury arc lamps with a maximum output at 3000 A. Duplicate samples were removed at 6 hours, 1, 2, 6, and 10 days. This follows the sampling protocol as determined by Mabey et al.(115) with the exception of the final sampling interval. Addition of Humics Humic solutions of Tahquamenon River humics, Two-Hearted River humics, Betsy River humics, and Blind Sucker Flooding humics were made up in distilled water. Three different concentrations of 1 ppm, 10 ppm and 100 ppm were used of each. Toxaphene was spiked by addition of an appropriate volume of the 10,000 ppm solution to get a concentration of 4 ppm in each of these solutions. Four ml samples were prepared, sampled and extracted as described below. The pH was adjusted as needed to pH 5, 7 and 9 by the addition of either 0.1N NaOH or 0.1N HCl. Iron (Fe[III]) was added to a concentration of 1 X 10'5M as FeCla in addition to 100 ppm humics from the Betsy and Two Hearted Rivers. The pH of these solutions were adjusted to 5 by the addition of 0.1N HCl. Addition of Chemical ghotosensitisers Three known triplet sensitizers were used to compare their effect on toxaphene with that of naturally occurring sensitizers. Stock solutions of fluorene, biphenyl, and crystal violet were made in acetonitrile to a concentration of 2000 ppm. A second stock solution of toxaphene of 2000 ppm in acetonitrile was also made. The final concentrations of sensitizer and toxaphene in distilled water were 10 ppm and 4 ppm, respectively. Additional acetonitrile was added to approximately 5% to ensure solubilization of the sensitizers. Their water solubilities were fairly low and it was desirable to maintain all components in solution. Samples were treated in the same manner with respect to sampling time, actinometer solution and extraction as described below. To assure that the same effects seen indoors would occur in sunlight, triplicate samples of each sensitizer plus toxaphene were exposed to sunlight over a 10 day period with control samples kept in the dark. Samples were extracted and analyzed as below. Actinometers Light Intensity was monitored both indoors and outdoos by p-nitrobenzophenone (PNAP)/pyridine actinometer as described by Dulin and Mill (116). Duplicate actinometers were removed at the same time at which samples were removed. These solutions were analyzed by reverse phase HPLC with 68 UV detection at 288nm. The mobile phase was 30% Acetonitrile/70% water with a flow rate of 2 ml/min. An example of a typical chromatogram is shown in Figure 9. The change in the PNAP concentration was monitored with time. The quantum yield of the PNAP reaction and incident photon flux were calculated based on the equation : ¢ = 0.0169 [Pryidine], (Dulin and Mill, 1982 Sample Extraction and Analysis Samples were extracted by the addition of NaCl to a final concentration of 10% (w/v). Once the NaCl dissolved, the 4 ml samples were extracted using three 4 ml portions of hexane. 10 pl of a 10 ppm solution of 2,4,5-trichloro- biphenyl was added to the extract as an internal standard. The extracts were then concentrated to 1 ml under a dry N2 stream. Large volume samples were extracted by liquid-liquid extraction. NaCl was added to a concentration of 10% w/v. The salt solution was then extracted with three 250 ml portions of hexane. These were concentrated by vacuum evaporation and finally to 1 ml under a dry N2 stream. The concentrates were separated by GLC using a 60m X 0.25mm Jaw DB-l column in a Perkin-Elmer 8500 gas chromatograph (GC), equipped with a °3Ni electron capture detector. 69 Figure 9. Representative HPLC chromatogram of the actinometer solution. Conditions are in the body of the text. PNAP-p-nitroacetophenone. 70 Capillary GC/MS - Negative Ion Mass Spectrometry A 60 m by 0.25 mm ID DB-l (J&W Scientific) capillary column was interfaced with a Nermag RIO-10C quadrupole mass spectrometer through a Delsi Di700 GC. A JSW Scientific on-column injector was used for sample introduction. Column head pressure was 20 psi giving approximately a 1 ml/min flow. The oven was temperature programmed from 150°C to 190°C at 30°C per minute with a hold at 190°C for one minute. The oven was then ramped from 190°C to 280°C at a rate of 2°C per minute. The heated interface between the mass spectrometer source and GC was maintained at a constant temperature of 250°C. The mass spectrometric parameters were as follows. The source temperature was held at 100°C as monitored by a thermocouple gauge directly affixed to the source heater block. The filament current varied from 150 to 195 mA depending upon source cleanliness. Methane was used as the moderating gas with a source pressure of 0.035 Torr as measured by a thermocouple gauge adjacent to the source. The electron multiplier voltage was 2.6kV with a dynode voltage of 4.8kV. The quadrupoles were scanned from 70 to 500 u at a rate of 2 scans per second. Optimization of the instrument was accomplished by monitoring the m/zs452 fragment of perfluorotributlyamine. In addition, a mixture of aldrin, heptachlor, 1-chlordane, and trans-nonachlor at a concentration of approximately 10 ppm each was injected daily prior to any chromatographic runs 71 to assess source conditions. If the total ion current as integrated by the SIDAR data system for aldrin fell below 750,000 counts, the source was removed and cleaned. Statistical Treatment of the Data Similarities and differences between laboratory weathered toxaphene, and standard toxaphene were assessed by the method of Morgen et al. (117). The toxaphene chromatogram was divided into 16 time segments within which the areas of the peaks were totaled and divided by the total area of the chromatogram. This was done to reduce the data set to a reasonable size for comparisons between different treatments. Segments represented approximately a 2 min. window of retention time accross the chromatogram. The first segment begins immediately after the elution of the internal standard, and the final segment at approximately 51 min. An example of how the chromatogram was segmented is shown in Figure 10. A typical example of normalized chromatographic data is presented in the appendices. The normalized chromatographic data was first tested by a one way ANOVA test to determine if the treatment accounted for the observed changes within a particular chromatographic segment. The starting material was then compared to the final mixture by cluster analysis using SYSTAT. Cluster analysis is a multivarilate statistical technique which is capable of making comparisons between large data sets. Relative rates of reaction were determined for those 72 segments which exhibited statistical significance at the 95% level. 73 Figure 10. A typical toxaphene chromatogram showing the seg- mentation patten of 16 different segments or zones used to normalize chromatographic data as described in the statitical analysis section. '744 986'99 | SPS'OD IEC'OD PIO'CD GOZ'to _ {‘8". C‘K'lh l'"'l. P zoo ° ‘ BH'SZ fl"! ssz-ei' ['8'82 ace-z: sea-9r 913's: tic-z: '*°"‘ sz-sz SCC'SZ J_4!j; I I35 | l‘i | '15 | I (5 I I ll |617|8i9 l0 1 5 4 3 Clustet No. 55 45 4O 30 35 Retention Time (min) 25 20 75 Section I Chapter 4 Results and Discussion The objectives of the research were as follows. First it was the intent to determine if toxaphene was photolytically labile in natural water solutions. Secondly, studies were conducted to determine the effects of various components of natural waters in order to evaluate their effect on toxaphene. To gain further insight into photochemical processes, results of natural water experiments were compared with triplet sensitized solutions of toxaphene. Finally, a comparison between residues isolated from environmental samples and those residues weathered in the laboratory is discussed. Experiments were carried out in several phases. The first phase investigated the effect of natural sunlight and UV-light on toxaphene in three natural water solutions at three concentrations making a total of nine different solutions. Secondly, humic materials were isolated from natural water sources and used to mimic natural water systems by their addition to distilled water. The effects of pH and 76 the presence of Fe(III) were investigated as well as the effect of humic concentration in aqueous solutions. The third phase compared natural sensitizers to known chemical sensitizers to determine if the effects in these two systems were similar. These simulated media were exposed to UV-light and natural sunlight. Phase I: Irradiation of Toxaphene in Natural and Distilled Waters Aqueous solutions of toxaphene were exposed to sunlight and UV-light over a ten-day period. Three concentrations of toxaphene were used as described in the methods section. The pH and conductivity of the waters were measured and are presented in Table 2. Conductivity was measured in micro Siemens (p8). Table 2 ., ,. .,. . w __. . . .. , 'Noto - 5.91119: RH Comma Lake Michigan 5.65 295 Two Hearted River 4.85 125 Betsy River 4.70 52 Difitill§d_fl§£§r 5:50 4 The waters collected from the Two-Hearted River and the Betsy River were high in humic materials and consequently, yellow 77 in color. The water from Lake Michigan, in comparison was clear and colorless. The waters high in humic materials had lower conductivity measurements than that of Lake Michigan. This is probably due to complexation of dissolved minerals by the humic materials. Humics have been found to be very effective chelating agents for trace metals and other dissolved inorganics (118). Irradiated solutions of natural waters, in either sunlight or UV-light, resulted in the dechlorination of higher chlorinated toxaphene species. Distilled water solutions of toxaphene did not yield these results. A comparison of the control chromatograms of toxaphene and a solution of toxaphene in Lake Michigan water exposed to sunlight is shown in Figure 11. Note the decrease in the relative intensity of the later eluting peaks versus the earlier eluting peaks. As described in Chapter 3 on statistical analysis of the data, each chromatogram was divided into 16 chromatographic segments, corresponding to a retention time window of approximately two minutes within the chromatogram. All calculations and further discussion will be based on theses chromatographic segments. The reader is referred to Figure 10, which shows a typical toxaphene chromatogram and the manner in which the chromatogram is segmented. Two peaks in particular were observed to dramatically Figure 11. 78 10 ppm control toxaphene (lower) and 10 ppm toxaphene in Lake Michigan water (upper) exposed to sunlight for ten days. The arrows designate peaks which either increased or decreased in relative peak area over the course of exposure to sunlight. 79 ‘ “H! ure'as nun-s net's: "£115 ) :ms «rt: ‘ «rt: 8!?“ | "9'36 :e:°:: OIO’I‘ are: ' 3‘ “‘35.: _ “a?“ ":09 054'». “'" :::.:: ifi-z’: esra urn ear» :0: I?! ‘90 ' m .9 ‘34 ‘70 (89°80 -» 028‘s. m urn no earn 953°» __ . are. war - 9995?. t- ,. 'St"v 6'0 , .. 8.6-» ”S N e mu.” ' ’ o m, .56: ,, ‘ 41,, “rum." 9 .“ — '“D'“ 7 -‘-—-—- uncut are: - -- m.“ -‘ ‘ '1 . w' ”3'; ' "ram-u” _ . .._. m.” . -"" a term w m- m . 1%“ 33%" -—9~ ' ' - Wilt» . n It“ "Pt! . “'4‘ * ",.e" o. .‘u " "flu “74 mu: ' ”'°" n“, ur- av?!" ' ' he“: - are: n l are: “I 'f W920“ , ”I“: ‘ . . v are; "34"3: :u'a . c 3 up“ “Vfi . _ en “fir; — 3k 2 3 He's-5 an“ O ’ e- ""‘i::' . . . a: c ‘ In 0‘ g, D 2 ___)umt t . . . O O “at: _ " E g “53:.- o o 0 fi '- "' it“: c E a g 940'“ O o. O O are: 0 :1 a: ‘0 O O o are: | _ _| .. "-9: I Ausuew) ‘flme (min) ‘ 80 increase over the course of the experiments. These were in chromatographic segments #5 and #8 respectively and these two peaks are starred in Figure 11 for clarity. These two species were characterized by capillary GC/MS by electron impact ionization. The peak in segment #5 was identified as a product of reductive dechlorination, with an empirical formula of C10H9C17, and a molecular mass of 374, though the base peak of the isotopic cluster was identified at m/z=376. The first fragment seen in the mass spectrum at m/z=325, with a base peak at m/zs327 for the isotopic cluster, represents a loss of m/z-59 (Figure 12). This would correspond to a loss of a CH2Cl fragment from the molecular ion, which is consistent with the fragmentation scheme shown below (scheme 1). Both the molecular ion isotopic peak cluster starting at m/zs374 and the isotopic peak cluster starting at m/zs325 are consistent with compounds having seven and six chlorine atoms, respectively. Scheme 1 Electron impact fragmentation scheme for reductively dechlorinated toxaphene components. The second peak in segment #8 was also identified as a Figure 12. 81 Mass spectral data (EI) obtained from the analysis of a component of chromatographic segment #5. Top: Reconstructed total ion chromatogram following identical conditions as those described for NI/MS analysis. Bottom: EI mass spectrum obtained at retention time of 35.4min. during GLC analysis. Intensity 82 ‘33 1§29 1333 235'; d 4 f d . I A UK 4 J V f T f I I I fi— 1' I r T —r \U . I— f F s— | r r r-J 3:: 400m . 43H” 30".” Retention 11m ‘1 mova- sscao IT 33-37.; measures-7:2 S i i I I < l i l i i (M-CHZCI) l i , ‘ i 37$ 4 I t. . l g ,7, 34': '. “7 1 i \a N ‘ , - 1 I k N ' . I l" I” m an ass 330 m 4” coo HIS Figure 13. 83 Electron impact mass spectrum of reductive de- chlorinated product in chromatographic segment #8. Its empirical formula was determined to be C10H3C13, indicating a species having three degrees of unsaturation. 84 00M OOH OWN O mmfiqlmovuzkmnquGUm 00? 0mm own Dvm Pb bk-—bN.N-bh-.hr—p-.pbbNL-b.pppLPE—__§ Dun Ova ?« ”Pu m.k«u0?l hm DNM PNP-bbp ”£91.51- KO” ONn 00¢ D u m u O.u w0n0«u INOOq lusuewu 85 product of reductive dechlorination. The mass spectrum of this unknown is shown in Figure 13. It was identified as having an empirical formula of CloHNCls, with a nominal molecular mass of 408 u. The major fragment seen in the spectrum for this product, is a loss of 83 u from the molecular ion; this is consistent with a loss of a CHClz fragment. This is also consistent with the fragmentation scheme of the lower chlorinated species described above, with the exception of the methyl group having one additional chlorine atom. As a confirmation of the identity of these fragments, both isotopic peak clusters starting at m/z=325 and m/z=408 are identical to those obtained from mass spectral analysis of compounds having six and eight chlorine atoms, respectively. Large volume samples at each concentration of toxaphene in each of the natural waters and distilled water were exposed to sunlight for ten days. These were used to evaluate the total change in chlorine isomer composition in the toxaphene mixture. These were analyzed by capillary GC/MS by methane negative ion mass spectrometry. The areas representative of each degree of chlorination (Clg-Cllo) known to be represented in the standard mixture, were integrated by the SIDAR data system. NI/MS produces spectra of toxaphene components which represent a loss of a single chlorine atom from the molecular ion. The masses over which the integration occurred reflect this and are shown in Table 86 3. Table 3 Mass Spectral Peaks Used to Integrate the Various Qhl2rinatsd_lsgmsr§_2f_lsxannens i_9f_§hlerins_Atsms Peaks_lmzzl______ c16 305-317 Cl, 341-351 Cl. 375-335 Cl. 409-421 c11° 443—455 Peak areas were then normalized by using the molar response factors determined for the mass spectrometer calibration mixture as described in the methods section. This mixture was made up of aliphatic bicyclic compounds having from six to nine chlorine atoms. The resulting histograms representing the data are shown in Figure 14. It was observed that with decreasing toxaphene concentrations, the degree of dechlorination increased over the course of the lO-day irradiation. Toxaphene solutions in distilled water did not exhibit these same changes. The relative concentration of C16- through Clg- containing isomers did not Figure 14. 87 Histograms of the isomer composistion of sunlight-exposed solutions of toxaphene in natural and distilled waters at three concentrations. (a)Distilled Water, (b) Betsy River water (c)Lake Michigan Water (d) Two Hearted River water. Note that with the exception of distilled water, an increase in the relative concentration of C17 isomers occurs with decreasing toxaphene concentration. 88 10 9 F. 8 # of Chlorine Atoms per lsomer 7/////////////////////////////////////////fl/////////////// n7 a q d 4 AW nqu m.n-U o 0 “2°86” omcodmom to_o2 oo~=octoz = lOppm Toxaphene 0.1ppm Toxaphene mm 1ppm Toxaphene _ _ V////////////////////////////////////////////////////////////A F Distilled Woter b 10 1 ‘1. G q 4‘- ‘ C 0 w 0 0 a O 6 2 2 2 1 macadmom .632 oo~=octoz m‘mdddo-s new # of Chlorine Atoms per Isomer Betsy River Water 89 10 , ,,/////,////////////////////////////////////..//////////////////l 7/////////////fl///////1////////////A 240-4 1 — q a o 6 2 1| e...- 200 s amcoamom c032 oa~=octoz 0.1ppm Toxaphene IIIID lppm Toxaphene = lOppm Toxaphene m 9 V///////////////flfl/////,7////////////////////////////////////A # of Chlorine Atoms per lsomer Lake Michigan Water d dd-4444J44d‘ddd-5 ‘ 1‘114 m o m m m o 5 m 3 2 4| ‘1‘ omcoamom co_o2 oo~=oEcoz # of Chlorine Atoms per lsomer Two Hearted River Water 90 show any significant change in distilled water while all three natural waters demonstrated dramatic changes in chlorinated isomer composition. Distilled water/toxaphene solutions did show a decrease in the relative concentration of C15- containing isomers with decreasing concentration of added toxaphene. This can probably be accounted for by either selective volatilization or incomplete extraction. It has been observed previously that the earlier eluting, lower-chlorinated species are the first to volatilize from the mixture (4). This would most likely correspond to the Cls-containing isomers. In all three natural waters, the percent composition of the mixture changed dramatically, especially with respect to the C17- and Clg-containing isomers. C17- isomers increased while Clg-isomers decreased consistently with decreasing concentration of toxaphene in solution. The relative concentration of Cllo-isomers also decreased with decreasing concentration but the data are not shown here. The most dramatic change in overall isomer composition was observed in the toxaphene/Two-Hearted River solution. The percent composition of Cly-isomers increased from a minimum of 59% at the lOppm level to a maximum of 89% composition at the 0.1ppm level. This change was confirmed by statistical cluster analysis by a comparison of the chromatograms after 10 days exposure to sunlight. The resulting dendrogram is shown in Figure 15. The Two-Hearted River/toxaphene solutions 91 Figure 15. The dendrogram which resulted from the heirach- ical cluster analysis conducted using the statistical analysis program SYSTAT on normalized chromatographic data of sunlight exposed toxaphene solutions. Two Hearted river (THOTl, and THOT2) samples are the least related to control samples (CONTl, 2, and 3) indicating the greatest overall change. 92 :1 RFC”. m. fill abomm HHOZJ NJmFZOQ mmmhzoo Heszoo abomo whomo OCO . H flu. 1H1: NHOZJ HhOIH users ooo.o mwozO NZH>Q ~ZHmm NZHim ahzoo MHZOO NHZOO O 113 the next most effective followed by biphenyl. This was confirmed by statistical cluster analysis of the chromatograms after 10 days exposure to UV-light. The dendrogram comparing the three photosensitizers to controls is shown in Figure 21. The controls (CONTl, 2, 3) are the most closely linked to the biphenyl group (BPINl, 2), indicating the least amount of change. This is followed by crystal violet group (CVIN1, 2) and fluorene group (FLIN1,2), respectively. As described earlier, the longer the arm of the dendrogram, the less related the connection. Thus, the solution of fluorene/toxaphene is the least related to the original toxaphene mixture isolated from non-UV exposed solutions. Rate constants were calculated based on the first 10 to 20% of the reaction for each time segment of the toxaphene chromatogram as described in the statistics section of Chapter 3. In the case of fluorene, the reaction proceeded so rapidly that within six hours of irradiation time, strong deviations from first order kinetics were observed. The ln(C/C.) plots for fluorene are not linear with respect to time after one day of exposure to UV-light (Figure 22). Reactions in the presence of crystal violet and biphenyl, in comparison, demonstrated first order kinetics throughout the time period. The main reason for the lack of linearity observed for the fluorene reaction is that as the reaction proceeds, many back reactions and products may begin to 114 Figure 22. First order plot of ln(C/Co) chromatographic segment #13 crystal violet, and biphenyl on 4 ppm toxaphene over the course of 10 days exposure to UV-light. 115 ocean: olo 6.05 .250 I 3:235 n. or O Amxoov «EC. m e. n — n — h 0 mLmNEmcmmoobsd 0 ’J 5 (00/3) ”I Calculated First Order Rate Constants (hp)(days‘1) of 4ppm 10 11 12 13 14 15 16 -0.32 -0.48 -0.03 -0.43 -0.35 -0.32 NS NS NS 0.35 0.28 1.0 3.7 0.61 1.3 116 Table 7 -0.19 -0.59 -0.29 -0.61 *0.40 -0047 NS NS NS 0.63 0.26 0.533 2.0 1.6 99°H°512~‘_. -4.7 -0.78 -0.56 0.37 0.79 3.8 * H5 3 DQE gjgnjfjggng gt a=.05 117 compete with the original components of the toxaphene mixture. Thus the kinetics of the reaction begin to deviate from that of the original reaction. The estimated rate constants, hp, are shown in Table 7. Note that fluorene, the most efficient of the three, has the greatest overall rate constant values, again confirming the cluster analysis results. Also note that between time segments 8 and 9, that the rate constants change sign indicating a change in the chromatogram from increasing relative peak areas to decreasing relative peak areas. This is also the point at which the composition of the toxaphene chromatogram changes from predominantly isomers having seven chlorine atoms to isomers having eight chlorine atoms. This break in isomer composition is shown in Figure 23 as three reconstructed mass chromatograms. From this then, it can said that isomers having eight or more chlorine atoms are the first in the mixture to be dechlorinated. As a result of this dechlorination of Cla-, Clg-, and C110- containing components, the final mixture, after photolysis, will have a preponderance of lower chlorinated species. Similar results have been observed in the photochemical dehalogenation of polyhalobenzenes (124). In hexane solution photochemistry, it was observed that the greater the degree of halogenation around the benzene ring, the greater the rate of reaction. In a mixture then, the higher chlorinated species would appear to be the first to react as has been 118 Figure 23. Reconstructed total ion chromatogram (upper) and mass chromatograms representative of C17- isomer composition (middle) and Cl.- isomer composition (bottom) of toxaphene. The dotted line designates the break between chromatographic segment #8 and #9. Intensity Total Ion Current CL, isomers Cl, Isomers 120 observed with toxaphene. Further, the estimated rate constants for toxaphene in natural waters, with humics and with the photosensitizers, increased with increasing chromatographic segment number. This is consistent with an increasing degree of chlorination of the eluting toxaphene components. All three sensitizer produced similar chromatographic alterations after UV-irradiation, with identical peaks increasing and decreasing in relative area with UV exposure. Differences between the three were manifested as the degree of chromatographic alteration. As with the natural water experiments, these chromatographic changes manifested themselves as a decrease in overall chlorination of the mixture. This was confirmed by the use of capillary GC NI/MS on a large volume UV-exposed fluorene/toxaphene solution. A comparison of the composition of standard toxaphene, toxaphene exposed to sunlight in distilled water (control) and toxaphene exposed to UV-light with added fluorene is shown in Figure 24. As with the natural water/toxaphene solutions, the percentage of C15- and C17- containing isomers increased while the Cla‘, C19-, and C110- containing isomers decreased after exposure to UV-light for ten days. Unlike the natural water solutions, the two peaks characterized as products of natural water photolysis in phase I, were not observed in photosensitized solutions. Since these two particular peaks did not increase with UV 121 Standard Toxaphene [HID Distilled Water + Toxaphene Flourene + Toxaphene 50- 30- 20- 10- Normolized Molar Response _\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\ W 10 # of Chlorine Atoms per Isomer Figure 24. Comparison of the isomer composition of standard toxaphene, lppm toxaphene in distilled water exposed 10 days to UV-light, and 4ppm toxaphene with fluorene in acetonitrile/water solution after 10 days exposure to UV-light. 122 Figure 25. Mass chromatograms at m/z=373 representative of [C10H3C13-C1] fragment of standard toxaphene (top) and toxaphene in acetonitrile solution with fluorene after 10 days exposure to UV-light (bottom). 1J23 zoom 00”” 000” ODDN Donn 000« GO” . b p lL Pl Ll . OOQN . )3!!! - IF . b Ia Lil .llilfi ii “3.... m"...\.,..m.,:_..__~2...... .5 E, 4. 2:8me >3 $2.0. f Eotgfixoinaoxo... hKMUUuK “nuns“ lflfiaa EL. 2.2338. 23.66 Sana... ovvan axon... I ax. men n .u - .u msuown 124 irradiation, some mechanism other than that of direct energy transfer from the humics to the toxaphene components is involved with the reaction. However, as a result of sensitized photoprocesses, numerous products were observed by NI/MS which were indicative of products produced by reductive dechlorination of toxaphene components. This is shown in Figure 25. Two reconstructed mass chromatograms of mass 373 are shown; the upper is that of standard toxaphene, while the lower is that of toxaphene exposed to UV-light for 10 days in the presence of fluorene. Mass 373 represents the [M-Cl] fragment of isomers having the empirical formula of C10H3Cls, having three degrees of unsaturation. Two would be from the six-membered base ring and the bridging carbon, while the third may be the result of either the formation of a double bound or a bridging bound between carbons 2 and S to form a tricyclic structure. The formation of this bridging bond has been observed with Toxicant B in the presence of the reducing agent triphenyltin (57). The fluorene/toxaphene solution shows an increase in relative peak area of ion current at m/za373, representative of reductively dechlorinated species. Further, it is also obvious from Figure 25, that the number of components with three degrees of unsaturation has increased from that of the original mixture. Similar changes were observed for components having six and seven chlorine atoms as well. The photolytic degradation of several single toxaphene components has been shown to photodegrade through 125 dechlorination and reductive dechlorination. Toxicant B, a major constituent of toxaphene, has been shown to proceed to the reductive dechlorination and dechlorination products shown below in Figure 26 (57). The site of action is the geminal dichloro groups at the 2 position on the ring. Two products, that of the 2-endo and the 2-exo products, were reported. Paarlar et a1. observed similar dehydrohalogenation products of a nona-chloro toxaphene component(9, 125). The toxaphene component underwent both reductive dechlorination and dechlorination at the 3 geminal dichloro position on the six membered ring to give the following products shown in Figure 27. Both reactions were hypothesized as proceeding though a radical intermediate. Because of their structural similarity to one another, the dechlorination and reductive dechlorination of other toxaphene components are likely to proceed through the same reaction pathway.u As described earlier, biphenyl did not have as great an effect on the overall chromatogram as the other two sensitizers. This was unexpected as its triplet energy of 67 kcal/mole is very close to that of fluorene. Upon examination of the UV absorption spectrum of biphenyl, it becomes apparent that 300 nm, the A max of the mercury lamps, is not an optimum wavelength for biphenyl (Figure 28). Crystal violet and fluorene, in comparison, have significant absorption bands at 300 nm. Thus crystal violet, with a much lower triplet energy of 39 kcals/mole, had a greater impact 12(5 o>auoLWMWV mascaouxoa mo cognacauodnooo p cowuocuuoasomo cauxaouonm .om mu m a an 6 S 8 8 8 8 H: S 8 Z 8 8 H0 H0 H0 8 8 8 8 8 A 8 8 8 G 8 fl UGQOfiXOH. 127 Ammm ovucocoeaoo ocogaaxou ouoanoococ a mo mamhaouozd ecu A A muo: pone cofiuocuuodsoop poo couuacfiuoasoop o>auo=pom ww ouom «m S 8 88 G S 8 G a 8 ‘ 8 S 8 ‘ ‘ ‘ Ho Ho 3 8 R A . S ‘ 8 H #0 Ho on toxaphene than biphenyl, with la28trip1et energy of 67 kcals/mole, simply because the absorption spectrum of crystal violet is better matched to that of the light source. Biphenyl would be less likely to have an effect than other PAHs because of its inability to absorb UV-light provided by sunlight. However, as the biphenyl molecule is halogenated to compounds such as polychlorobiphenyls (PCB) and polybromo- biphenyls, a bathochromic shift in the absorption spectrum of the molecule occurs thus matching its absorption spectrum with sunlight (126). PCBs have been shown to behave as triplet sensitizers in the transformation of cis-pentadiene to its trans- isomer (126). They are likely then to behave similarily in the environment. Further, it was observed in our laboratory that a solution of toxaphene in hexane with added nonachlorobiphenyl as an internal standard, was altered over the course of several months as it sat on the bench top under fluorescent lighting. Though this is by no means a confirmation of the sensitizing capabilities of nonachlorobiphenyl, the alterations in the toxaphene chromatogram were similar to those of solutions exposed to UV-light in the presence of sensitizers (127). Fluorene over the course of the irradiation experiment, oxidized to fluoren-9-one. The identity of the oxidized product was confirmed by capillary GC/EI-MS in conjunction I.” .. 129 Biphenyl Assn-- |. 1 u... Fluorene :3 c 3 B '< u- ‘, Crystal Violet L-.. t-.. I.-. t-.. ‘- e ; - r : 4 3 I I I I I I I I I I a i d ‘a‘ i I E 5 3 I 1 nm Figure 28. UV-absorption spectra of biphenyl, fluorene and crystal violet. 130 with a computer assisted library search. The comparison of library and unknown spectra is shown in Figure 29. F1uoren-9-one is also used as a photosensitizer with a triplet energy of 53 kcals/mole. Fluorene has been shown to photo-oxidize in number 2 fuel oil through a hydroperoxide intermediate to f1uoren-9-one by the scheme shown in Figure 30 (128). Thus as fluorene is photo-oxidized to fluoren-9-one, the oxidized product would also sensitize the dechlorination of toxaphene components. The overall dechlorination would be the result of the sensitization by both these molecules with the exception of the beginning or end of the oxidation reaction where either the oxidized or non-oxidized fluorene species predominates. As fluorene and biphenyl are both common environmental contaminants, it was of interest to demonstrate that sunlight would effectively promote the dechlorination and reductive dechlorination reactions observed in the laboratory. To this end, triplicate solutions of each sensitizer in solution with toxaphene were exposed to natural sunlight for ten days. Fluorene was no longer the most effective in the dechlorination of toxaphene, but crystal violet became the more effective sensitizer in sunlight. This was confirmed by the comparison of chromatograms by statistical cluster analysis. This can again be explained by the absorption spectra of these compounds. The most intense absorption band for crystal violet lies in the visible region as shown in Figure 29. 131 Comparison of unknown spectrum to library spectrum of fluoren-9-one. Top line: unknown mass spectrum Middle line: library mass spectrum of fluoren-9-one. Bottom line: a comparison plot of the two mass spectra. 132 00? 00? 00h 00H ODN 00m Ohm 00g on nnlrppppbanbb.Lb—pru—hprLLppu—Pth+b--—--¥—ppbplPIth-pF-bp—pPipnppbl-—b-b-£hb . . . .3. I..- WWW. .. .0 o a” wmml al.012flwno .ocSmluw3 o. looJo..O.Nom'IM. co. IW...HI~..I. .oJ WMW' .loo It .QWWIQNWIVSIII I... 3mN. ”in. I. I 285328 I mm; mm 1 mm o.m wk r «.3 03 . l mmMoSnzcom - l l 1 l l l l one 133 I m9. 3.... mun 13.... man . New mm of l on." mm mm on m... . mm. o a . [I nvmtmomlwnmtquIZduw H.anNNI hm «mkhw l300u anomv ”20 m ZUMDZJk In R? lam 00a l3: sznno >¢¢¢mu4 cahowmw IHZ\¢mU m~40~muJD u>m¢mmn4 0n0awh >00 mxw UZUfiJDJk {out Wkodmhxw DUJDDQ 00l2¢6lfln OuflJka .4 “Jun DDIEHIGA no .10 u H.N D gnumgfiuz 133 326on All . u .Amwav oco-o-couos~m cu ocouooam mo muosvoue couuovaxo-ouo£a one .Om ouswum 134 Figure 31, (320-700 nm) while the most intense absorption bands for fluorene and biphenyl lie below 320 nm in the UV region of the spectrum. The spectral photon flux of natural sunlight is more intense in the visible region than in the UV region of the spectrum. As a result, a sensitizer with a strong visible absorption band will be more effective than that of a sensitizer with a strong UV-absorption based purely on the photon flux of natural sunlight. In general, sunlight is an effective promoter of the dechlorination of toxaphene in the presence of sensitizers. This is of greater importance when the influx of PAHs and PCBs into water bodies is considered. Fluorene, and biphenyl, as was well as many other PAHs, have been detected in fly ash from municipal power plants (129), incinerators (130, 131), coal-fired residential furnaces (132), in diesel exhaust particulate matter (133) and as an emission from power plants (134). Similarly, PCBs have been detected in effluents from coal power plants at the 0.11 pg/m3 level (135). Once in the atmosphere, the majority of the low molecular weight PAHs exist in the vapor phase while those with a molecular weight of greater than approximately 200 are bound to particulate. These gradually return to earth by rain washout of particulate matter (136). By this process, as well as more direct inputs, PAHs are present in the Great Lakes basin. Concentrations of PAHs vary throughout the Great Lakes region. For example, biphenyl concentrations 135 t- A 4 A ‘ 4 ‘ A #4 V f 1 w Crystal Vlolet 2.7. 0 o 2.4. .. 2.1. 0 1o- 0 1% 1.- 0 Abs. Units 1.- q- ..- o I.- 0 a- ., a- .h— 7.4.. 732.- m.- sea.- “.- . “O.- ., Figure 31. Visible absorption spectrum of crystal violet. 136 ranged from a minimum of 0.03 ng/l in Thunder Bay, Ont., to a maximum of 31.9 ng/l at Sault Ste. Marie, Ont. Fluorene varied from 0.1 ng/l at Toronto, to a maximum of 105 ng/l at Sault Ste. Marie, Ont (137). In comparison, toxaphene levels in the Great Lakes have been recorded in the open lakes at 1.4 ng/l (8). With sunlight, the same sort of photolytic dechlorination observed in the laboratory should occur in natural waters. The UV-B region (280-340 nm) of sunlight, has been measured to depths of 9m in the open waters of Lake Michigan (138). Similar depths for UV-B penetration have also been recorded for the open waters of the oceans (139). The presence of PAHs such as biphenyl and fluorene in the aquatic environment, as well as other xenobiotics such as PCBs, will influence the fate of toxaphene in the environment. Comparison of Residues Isolated From Fish Versus Photolyzed material Residues of toxaphene were isolated from Lake Superior Lake Trout bellies and fillets following the procedure of Ribick et al. (26). Pooled residues were analyzed by capillary GC/MS by methane negative ion mass spectrometry for ionization and detection. These residues, however, were separated using a 30 m DB-1 column rather than the 60 m column previously employed (140). The chlorinated isomers of the photolyzed material and fish toxaphene residues were 137 normalized as previously described. The resulting histogram is shown in Figure 32. Fish residues, overall, contained a greater relative concentration of C13- and C19- containing isomers, while the photolyzed material exhibited much greater relative concentrations of C17- containing isomers. This same distribution of chlorinated isomers has also been observed by Swackhamer et al. in residues isolated from fish from Siskiwit Lake on Isle Royale (29, 30) This distribution of isomers in fish residues is to be expected. As has been shown for PCB isomers, the higher chlorinated species are more readily bio-accumulated because of their higher lipophilicity (141). In addition, these higher chlorinated isomers are less likely to be metabolized, while their lower chlorinated counterparts are initially less bio-accumulated and more easily metabolized and eliminated by an organism. Because of selective accumulation by the organism, it is difficult to say whether the residues in fish are a result of uptake of photolyzed material. To better evaluate this, toxaphene residues isolated from lake water should be evaluated to determine if the distribution of chlorinated isomers is similar to the photolyzed residues produced in the laboratory. 138 Figure 32. Comparison of the isomer composition of toxaphene residues isolated from Lake Trout bellies and fillets, standard toxaphene and lppm toxaphene in Betsy River water after 10 days exposure to sunlight. 139 Normalized Molar Response # of Chlorine Atoms per Isomer Standard Toxaphene Lake Superior Lake Trout Residues [EEIJ Lake Trout Belly Residues E lppm Toxaphene in Betsy River Water 140 Implications of the Photolysis of Toxaphene in Natural laters Humic materials, as described in Chapter 2, are ubiquitous organic compounds occurring in natural waters and soils. It is apparent from the results of these experiments that toxapheneis influenced by indirect photoprocesses in the presence of humic materials. As such, toxaphene in the environment will be altered by these interactions to a mixture having a greater relative concentration of lower chlorinated species than the original mixture. These lower chlorinated species are then capable of being readily metabolized by organisms in the environment. Photolysis has also been shown to occur in waters having low concentrations of dissolved organic material, i.e., Lake Michigan water. From these results, it can be shown that photolysis is a major force for the alteration of toxaphene in clear natural water systems as well. However, a mechanism other than those involving humics must be ivolved in these changes. FeOH+2 has been described as a photoreactive species found at relatively high concentrations in natural waters (57). Fe(OH)n*2:F also exists in colloidal form in natural waters as a result of aggregation by FeOH+2 ions. From the distilled water with added iron as Fe(III), implications are that some iron species may be involved in the clear water photolysis of toxaphene. This same sort of behavior has been observed with other pesticides as well 141 (105). In addition, studies conducted in an iron redox system indicated that iron was an effective catalyst for the dechlorination of both toxaphene and DDT (57, 121). . Finally, most natural waters have been contaminated by a variety of xenobiotics such as PCBs and PAHs. These act as effective sensitizers in the dechlorination of toxaphene. These xenobiotics are present in the aquatic environment at the ng/kg level, which is comparable to the concentration of toxaphene in open water. The third element in the fate of toxaphene is that of sunlight which can penetrate into clear waters of Lake Michigan at depths of up to 9 meters (138). Similar results have been reported for the open ocean waters along the Pacific coast (139). Humics and other dissolved organic material are present in many natural waters at greater concentrations than PAHs. These humic materials, on the basis of their higher concentration would be expected to influence the fate of toxaphene to a greater extent. The final result is the essentially the same regardless of the source; that of the overall dechlorination of the toxaphene mixture to a mixture composed of predominately C15- and C17- containing isomers. Thus through the interaction of sunlight, humic materials, or xenobiotics, residual toxaphene will be photolytically altered in the environment. ’ 142 Section I Chapter 5 Summary and Conclusions The complex chlorinated hydrocarbon mixture, toxaphene, was exposed to sunlight and UV-light in solutions of natural and distilled waters. Changes were observed in the overall mixture indicative of dechlorination of the higher (C18, C19, and C110) chlorinated species of the mixture. These produced products having six and seven chlorine atoms many times with an increase in the degrees of unsaturation on the molecule. Two products were characterized by the electron impact mass spectra. The first eluting within chromatographic segment #5 had an empirical formula of C10H9C17. This would indicate that the compound has three degrees of unsaturation, two a result of the bicylic ring structure common to toxaphene components, and the third a result of the formation of a double bond or formation of a tricyclic structure. The second peak eluting in chromatographic segment #8 was characterized as having an empirical formula of C10H3C13, again indicating a species with three degrees of unsaturation. Both were probably produced by the reductive dechlorination of isomers having greater chlorine atom than the final product. The effect of the concentration of aquatic humic 143 material was investigated. The rate of dechlorination was found to increase with increasing humic concentration. pH of the aqueous humic solutions did not appear to affect the rate of reaction. The presence of Fe(III) both with and without added humic materials affected the rate of reaction. There was noted a direct correlation between the rate of dechlorination and the presence of iron as Fe(III). Solutions of toxaphene were exposed to UV-light and sunlight in the presence of photosensitizers. These sensitizers were much more effective at dechlorinating the toxaphene mixture. Species indicative of reductive dechlorination and dechlorination were noted as end products in the mixture. In UV-light, the order of reactivity was determined to be Fluorene > Crystal Violet > Biphenyl. In sunlight, the order of reactivity was altered to Crystal Violet > Fluorene > Biphenyl due to the intense absorption band in the visible light region of crystal violet. In conclusion, toxaphene is not photolytically non-reactive as previously described. Under relevant environmental conditions, toxaphene is dechlorinated in natural water solutions. Therefore toxaphene will be dechlorinated by materials in the environment. The rate at which this occurs will be determined by the organics present, and their concentrations. Transition metals such as iron may be involved as suggested by the data presented here. In addition, other xenobiotics present in natural waters as a result of contamination such as PAHs and PCBs are effective 144 sensitizers in aqueous solutions. From this it can be gathered that the overall composition of the mixture will be dramatically altered as a result of environmental influences. 145 Section I Chapter 6 Future work There are several projects which I can see developing out of the research I have completed on the fate of toxaphene in natural waters. The first project would be an investigatation of whether the toxicity of the pesticide is altered after exposure to light over a period of time. Mosquito larvae and fat head minnow would be the most appropriate organisms to utilize for the study. Mosquito larvae are the traditional species used to test the toxicity of insecticides, while fat head minnows are used to evaluate aquatic toxicity. This study would be of interest as residues isolated from fish out of the Great Lakes demonstrate no change in toxicity. Since the photolyzed material demonstrated a decreased in overall degree of chlorination, this may not be the case. A second project could be instituted to test if toxaphene levels in water are reduced by photoincorporation by humic materials. There is evidence which suggests that photoincorporation may occur between humic materials and xenobiotics. It would be of interest to determine if incorporation actually occurs as a result of photolysis. 146 Solutions containing humic materials and radiolabeled toxaphene could be exposed to UV-light and extracted. Any remaining radioactivity in the aqueous phase of the extract would then be likely a result of photoincorporation. Prior to this study, toxaphene photolyzed with humic "monomers" as described in Chapter 2, could be used to determine if photoincorporation would be likely to occur at all. Other PAHs, such as anthracene and/or chrysene, and chlorinated biphenyls should be investigated for their photosensitizing properties. These are present in most natural waters as a result of anthropogenic contamination and would therefore be likely to affect the dynamics of toxaphene in the aquatic environment. I have shown that fluorene is a potent sensitizer for the dechlorination of toxaphene components and it is likely then that other PAHs may affect toxaphene similarly. Finally a series of binding studies with humics could be initiated. This would examine alternate routes for the removal of toxaphene from the aquatic environment. 147 LIST 0! REFERENCES 10. 11. 12. 13. 14. 15. 16. 148 Section I: References Anonymous, in e s v Aspects, Technical Report Prepared by Hercules Inc, 1970. J. E. Casida, R. L. Holmstead, S. Khalifa, J. R. Knox, T Oshawa, K. L. Palmer, R. Y. Wong, Sci; 1974 181 520-521. M- D- Reuber. With 1979 5 729-748- J. N. Seiber, S. C. Madden, M. M. McChesney, and W. L. Winterlin. Miami 1979 2.4 284-290. G. D. Veith, D. W. Kuehl, F. A. Puglisi, G. E. Glass, and J- 6- Eaton. WW 1977 i 487-499- M. Zell, and K. Ballschmiter, Ere§L_zL_Ana;&_§ngm; 1980 100 387-402. T. F. Bidleman, and C. E. Olney, Nature 1975 251 475-477. C- P. Rice. and M. 8. Evans. in WW Great_Lakg§ Wiley-Interscience Publications, 1984, pp. 163-194. H. Parlar, D. Kotzias, and F. Korte, Abstracts Sixth International Cong. Pest. Chem., IUPAC, 1986, pg. 6A-16. R. D. Arthur, J. C. Cain, and B. F. Barrentine, Bull. Enxrigni_fign&§mi_lgxi 1976 15 129-134- H- A- Saleh, 1;_AQIIQI_EQQQ_Qh§mi 1933 11 743-751- R. L. Holmstead, S. Khalifa, and J. E. Casida, J&_Ag:i§& dn§_EQQQ_£h§ni 1974 22 939-944- B. Jansson. and U- Wideqvist. W gum 1983 3; 309-321. M. A. Saleh, and J. E. Casida, J&_Agzig‘_figgd_ghem& 1977 25,63-67. S. Khalifa, T. R. Man, J. L. Engel, and J. E. Casida, 11.59Ii£1_£29§_£h§m1 1974 22 653-557- M. L. Anagnostopoulos, H. Palar, and F. Korte, Smasher: 1974 2 65-70. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 149 F. Matsumura, R. W. Howard, and J. 0. Nelson, ghomooohoro 1975 5 271-276. W. V. Turner, 8. Khalifa, and J. E. Casida, 11.59I191 £221.9h2n1 1975 5 991-994. P. F. Landrum, G. A. Pollock, J. N. Seiber, H. Hope, and K. L. Swanson, Qhomooohoro 1976 2 63-69. J. N. Seiber, P. F. Landrum, S. C. Madden, K. D. Nugent, and W. L. Winterlin, 11.90I9n1 1975 111 361-368. P. S. Chandurkar, F. Matsumura, and T. Ikeda, ghomooohozo 1978 2 123-130. A. J. Graupner, and C. L. Dunn, Jo_Ag;ioo_£ooo_Qhomo 1960 8 286-289. R. Liebmann, D. Hempel, and E. Heinisch, Aroho_2flonoonoo 1971 1 131-150. R. J. Ismail, and F. L. Bonner, Jo_ofifo_5nolo_§homo 1974 51 1026-1032. R. A. Hughes, G. D. Veith, and G. F. Lee, E§L£I_B§§i 1970 A 547-558. M. A. Ribick, G. R. Dubay, J. D. Petty, D. L. Stalling and C. J. Schmitt, Enyirono_§oio_1oono 1982 1o 310-318. J- C- Underwood. 831W 1978 2.0 445-446. J. E. Casida, and J. E. Saleh, onioolooy, EPA report, EPA-600/1-78-060, 1978. D. L. Swackhamer, M. J. Charles, and R. A. Hites, Extended Abstract, ASMS 1986 Cincinnati, OH, pp. 238-239. D. L. Swackhamer, M. J. Charles and R. A. Hites, Anal. Qhomo 1987 52 913-917. T. F. Bidleman, U. Wideqvist, B. Jannson, and R. Sénderlund, ALEQ§1_EDXIIQDL 1987 21 641-654. L. J. Lawrence, and J. E. Casida, L¢1o_§o1o 1984 3; 171-178. F. Matsumura, and S. M. Ghiasuddin, 1o_£nyirono_§oio Hgglgh 1983 318 1-14. 150 34. I. M. Abalis, M. E. Eldefrawi, and A. T. Eldefrawi, Resti_Bioohsmi_£hxsi 1985 21 95-102. 35. S. M. Ghiasuddin, and F. Matsumura, 99nn1_2122h£m1_2h¥§1 1982 zoo 141-144. 36. W. V. Turner, J. L. Engel, and J. E. Casida, Jo_Agzioo Eoog_ghomo 1977 25 1394-1401. 37. K. L. Olson, F. Matsumura, and G. M. Bousch, 5:281 Enxiroai_gontami_Toxi 1977 25 1394-1401. 38. H. M. Mehendale, £oo§_§oomooo_1oxo 1978 1o 19-25. 39. A. L. Allen, L. D. Koller, and G. A. Pollock, qo_1oxo Enxirgni_flealsh 1983 11 61-69- 40. N. K. Hooper, B. N. Ames, M. A. Saleh, and J. E. Casida, $21; 1979 295 591-593. 41. D. L. Henegar, in Publication #3m 1966, pp 1-16. 42. M. Katz, I:§D§&-Afl&-Ei§h&-§QQ& 1951 29 264-253- 43- H- 0- Sanders. and 0- B- Cope. Transi_Amsri_Ei§hi_§oci 1966 25 165-169. 44. S. C. Schimmel, J. M. Patrick, Jr., and J. Forrester, Arshi_Enxironi_Qontami 1977 5 353- 367- 45. K. J. Macek, C. Hutchinson, and O. B. Cope, Boll; Enxironi_£ontami_loxi 1969 4 174-183- 46. F. L. Mayer, P. M. Mehrle, and W. P. Dwyer, in onoohono; a! on 0 0 2 9-2! a. ll°ws ._,._ 1-11 EPA Ecol. Res. Series, EPA-600/3/75- 013, 51 pgs, 1977. 47. F. L. Mayer, P. M. Mehrle, and P. L. Crutcher, IIQn§1_AE1 Ei§h1_§221 1978 107 326-333. 48. A. R. Isensee, G. E. Jones, J. A. McCann, and F. G. Pitcher. £1.89Ii§1_£290_§hsm1 1979.21 1041- 1046- 49. G. H. Willis, L. L. McDowell, S. Smith, L. M. Southwick, and E. R. Lemon, AQIQDLAlL 1980 12 627-631. 50. T. E. Archer, lo_D§1I!_§91o 1971 51 1180-1183. 51. K. S. LaFleur, G. A. Wojick, and W. R. McCaskill, 11 Eaxiroai_gaal- 1973 2 515-518- 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 151 J. F. Paar, and S. Smith, §911_§Qi. 1976 121 52-57. L. L. McDowell, G. H. Willis, C. E. Murphree, L. M. Southwick, and S. Smith, I._Enxirsn._Qual. 1981 120-125. L. C. Terriere, U. Kiigemagi, A. R. Gerlach, and R. L. Borovioka. LAW. 1966 14 66-69- J. L. Gallagher, 8. E. Robinson, W. J. Pfeiffer and D. M. Seliskar, Hygrooioo 1979 o; 3-9. G- D- Veith. and G. F- Lee. W 1971 s 230-234. R. R. Williams, and T. F. Bidleman, l._AinQ._EQQ§ Qhomo 1978 go 280-282. F. H. Weber. and F- A. Rosenberg. W m 1980 2; 85-89. M. A. Saleh, and J. E. Casida, £._Asris._Essd_£hsn. 1978 2§ 583-590. H. W. Harder, E. C. Christensen, J. R. Matthews and T. F. Bidleman, Eatoorioo 1980 1 142-147. R. Vaz, and R. Blomkvist, Qhomooonoro 1985 11 223-231. F. J. Stevenson, in Wis. QQERQEILLQRLLBQQQLIQDS John Wiley and Sons, New York, NY 1982. E. H. Thurman. in W. Martinus Nijhoff/Dr. W. Junk Publishers, Dordrecht, Netherlands, 1984. N. Shinozuka, C. Lee and S. Haynoi, £211_IQ§1_EDYIIQD. 1987 §Z 311-314. R. G. Zepp, P. F. Shlotzhauer, M. S. Simmons, G. C. Miller, G. L. Baughman, and N. L. Wolfe, Exooo_zo_5nolo Chem. 1984 112 119-125. 0. C. Zafiriou, J. Jaoussot-Dubien, R. G. Zepp, and R. G. Zika. W... 1984 12 358A-371A- G. G. Choudhry, BQ§L_B§1o 1984 22 59-112 R. Beckett, z. Jue and J. C. Giddings, W. Ififihs 1987 21 289-295. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 152 G- a. Choudhry. in nanis_Substansss.__s;rus§nral. '9' 09 '10 0 9 m_ - 10:: ‘ 21° Intsra_ti9ns_xith_En21r2nasntal__hsmisals Gordon and Breach Science Publishers, New York, NY, 1984, 57-89. A. P. Black, and R. F. Christman, Amozo_flotor_florko Afifioooolo 1963 55 753-707. D. Kotzias, M. Hermann, A. Zsolnay, H. Russi and F. Korte. Natnrsxisssn. 1986 13.35-36- F. H. Frimmel, and H. Bauer, §£1._IQL._EDYIIQD. 1987 o; 139-148. D. S. Gamble, M. I Haniff, and R. H. zienius, Analo_ghomo 1986 58 727-731. C. W. Carter and I. H. Suffet, Enyirono_§oio;Iooho 1982 1o 735-740. E. W. Perdue, in 1983 Ann Arbor Science, Ann Arbor, MI Edited by R. F. Christman and E. T. Gjessing, 441-461. S. W. Karickhoff, D. S. Brown and T. A. Scott, flooor Booo 1979 11 241-248. C. N. Haas and B. M. Kaplan, Enyirono_§oio_moono 1985 12 643-645. P. F. Landrum, S. R. Nilhart, B. J. Eadle and W. S. Gardner. Enxirsn._§si._Tssh. 1984 18 187-192- M. E. Melcer, M. S. Zalewski, M. A. Brisk, and J. P. Hassett, gnomooonogo 1987 1o 1115-1121. C. W. Carter and H. Suffett, in Enxizonmont, 1983 ACS Symposium Series 225, American Chemical Society, Washington D.C. J. P. Hassett and M. A. Anderson, Eoyirono_§oio_1ooho 1979 11 1526-1527. N. Senesi, and C. Testini, goooozno 1982 28 129-146. N. Senesi and C. Testini, £§§§._§§i. 1983 19 79-89. U. Jensen-Korte, C. Anderson, and M. Spiteller, Ssi._12t. 32111201 1987 fig 335-340. 153 85. A. M. Fischer, D. S. Kliger, J. S. Winterland and T. H111, Qhomofiohozo 1985 11 1299-1306. 86. F. H. Frimmel, H. Bauer, J. Putzien, P. Murasecco and A. M. Braun. Enxiron_Ssi._Tssh. 1987 21 541-545- 87. A. M. Fischer, J. S. Winterland and T. Mill, in Symposium series #327,1986,141-166. 88. F. H. Frimmel, and H. Bauer, So11_1oto_znyizon1 1987 o2 139-148. 89. W. R. Haag and J. Holgne, Eny1ron1_§o11_1ooh1 1986 29 341-348. 90. R. G. Zepp, N. L. Wolfe, G. L. Baughman, and R. G. Hollis, Nooozo 1977 2oz 421-423. 91. P. Breugem, P. van Noort, S. Velberg, E. Wondergen, and J. Zijlstra, gnomooonozo 1986 15 717-724. 92. D. Kotzias, K. Hustert, and A. Weiser, ghomoooozo 1987 1o 505-511. 93. D. Kotzias, M. Herrmann, A. Zsolnay, H. Russi, and F. Korte. Natnrnisssnshaffsn 1986 13 35-36- 94. J. D. Hem, in Stadx_and_IntsIaIstatisn_sf_ths_9hsmisal Qharsstsristiss_af_Natural_flatsr 1975 Geological Survey Water-Supply Paper 1473, U.S. Government Printing Office Washington, D.C. 95. W. R. Haag and J. Hoigné, gnomooohoro 1985 11:12 1659-1671. 96. R. D. Ross and D. G. Crosby, gnomooonoro 1975 5 277-282. 97. R. G. Zepp, B. L. Baughman and P. F. Schlotzhauer, gnomooooro 1981 19 109-117. 98. R. M. Baxter and J. H. Carey, EISSDW._5191. 1982 12 285-292. 99. P. Brimblecombe, and D. Shooter, Mo;1_gnom1 1986 12 343-353. 100. F. E. Scully, and J. Hoigné, gnomooonoro 1987 16 691-694. 154 101. R. W. P. M. Laane, W. W. C. Gieskes, G. W. Kraay, and 3- EVEISdijk. Hfithi_l._§£§_3§§. 1985 12 125-128- 102. C. J. Miles, and P. L. Brezonik, Enyiron1_fio11_1oon1 1981 15 1089-1095. 103. H.-W. HHwang, R. E. Hodson, and R. F. Lee in ACS Symposium Series 327, American Chemical Society Washington D. C., 27-43. 104. R. G. Zepp, P. F. Schlotzhauer, and R. M. Sink, Eoyiron1 5911.159h. 1985 12 74-81- 105. Y. I. Skurlatov, R. G. Zepp, and G. L. Baughman, go AQIi§._EQQ§_§b£m. 1933 21 1065-1071- 106. R. G. Zepp, N. L. Wolfe, J. A. Gordan, and R. C. Fincher. 1._Asris._food_§hsn. 1976.24 727-733. 107. R. G. Zepp, N. L. Wolfe, L. V. Azarraga, R. H. Cox, and G- W- Pepe. Arsh..£nxir9n._£9ntan.;lsx. 1977 6 305-314- 108. W. R. Mabey, D. Tse, A. Baraze, and T. Mill, gnomooonoro 1983 12 3-16. 109. G. C. Miller, R. zisook, and R. G. Zepp, J1_Agrioo_£ogd thm1 1980 25 1053-1056. 110. H. -M. Hwang, R. E. Hodson, and R. F. Lee, Wotor_3oo1 1987 21 309-316. 111. s. U. Khan, and M. Schnitzer, 11_zn21ron._soi._nsalth 1978 211 299-310. 112. G. C. Choudhry, J. A. van den Broecke, G. R. B. Webster, and O. Hutzinger, Qhomooohogo 1987 2-3 495-502. 113. G. C. Choudhry, J. A. van den Broecke, G. R. B. Webster, and O. Hutzinger, Enyigon1_1oxo_ghom1 1986 5 625-635. 114. E. M. Thurman, and R. L. Malcolm, Enyiron1_fioi1_logn1 1981 15 463-466. 115. Mabey, W. R., T. Mill and D. G. QHendy. in 1 s - . oco T .- . . ' Qhsmis91s_in_ths_Air_and_flatsr by T- M111. 0- C- Bomberger, T. -W. Chou, D. G. Hendry and J. H. Smith (1982) EPA-600/3-82-022 49-103. 155 116. D. Dulin and T. Mill, Engirgn1_§g11;1ggn‘ 1982 16 815-820. 117. S. L. Morgan, M. D. Walla and J. C. Rogers, shrgmasgsrannx 1987 2. 12-20- 118. F. J. Stevenson in ° , 1982, Wiley Interscience Publications, New York, pp 337-352. 119. R. G. Zepp, G. L. Baughman, and P. F. Schlotzhauer, Chemeenhers 1981 10 119-126- 120. R. G. Zepp, N. L. Wolfe, L. V. Azarrage, R. H. Cox, and C- W- Papa. Ar9h1_Enxir2n1_§2n§am1_Iex1 1977 6 305-314- 121- B- L- Glass. £1.59ris1_£226_£hem1 1972 22523211 122. B. A. Goodman, and M. V. Chesire, Sgi1_1g;1_3311;9n& 1987 fig 229-240. 123. J. Sedlacek, E. Gjessing, and J. P. Rambach, Sgi;_19§& Engixgn1 1987 62 275-279. 124. M. Nakada, S. Fukuhsi, H. Hishiyama, K. 0kubo, K. Kume, M- Hirota. and T. Ishii. BullL_Qnem1_§291_lnn1 1983 56 2447-2451. 125. H. Parlar, S. Gab, S. Nitz, and F. Korte, 0 re 1976 5 333-338. 126- L- 0- Rn20. thesis, Ine_2n9L9snem1§1r2_and_zin§tigs_ei zgng‘ 1972. 127. S. Erhardt-Zabik, unpublished data. 128. Larson, 1977, Engi;gn1_§giL_Iggn‘ 1977 11 492-496. 129. R. E. Clement, A. C. Viau, and F. W. Karasek, Intern141; Enxir2n1_bnal1_snem1 1984 11 257-266- . 130. H. Y. Tong, D. L. Shore, F. W. Karasek, P. Helland, and E. Jellum, 1‘_§nzgm& 1984 16 430-448. 131. R. E. Clement, A. C. Viau, and F. W. Karasek, anL_JL Qn§m1 1984 52 2629-2633. 132. G. Grimmer, J. Jacob, G. Dettbarn, and J. W. Naujack, EIEEL_Z1_An§11_§h§m1 1985.122 595-502- 156 133. H. Y. Tong, and F. W. Karasek, 53311_Qngm1 1984 56 2129-2134. 134. E. Hasénen, V. Pohjola, M. Hahkala, R. Zilliacus, and K. Wichstrém. W 1986 5.4. 29-51- 135. C. T. Farmer, and T. L. Wade, flgtg: A, S, 2, 1986 22 439-452. 136. J. J. Richard, and G. A. Junk, EngingnL_§911_I§gn1 1981 15 1095-1100. 137. D. J. Williams, E. R. Nestmann, G. L. LeBel, F. M. Benoit, and R. Otson, gngmggpnggg 1982 11 263-276, 138. W. R. Gala, and J. P. Geisey, 1988, J1_§;gg;_Lgkg§_Bg§1 submitted for Publication. 139. R. C. Smith, and K. S. Baker, Rhgggghgm1_£n9392191 1979 22 310-318. 140 J- w. Gooch MW 0 .13.. o 0‘ :0”. . _o :1; ‘ o_ _‘ 1986 Thesis 141. A. J. Nimi, and B. G. Oliver, . s . 1988 A5 222-227. 157 Section II SELECTIVE SENSITIVITY OP HIGHLY CELORINATED SPECIES IN NI/MS: IMPLICATIONS EOE COMPLEX MIXTURE ANELYSIS 158 Section II Chapter 1 Literature Review of Negative Ion Mass Spectrometry Introduction Electron capture negative ion mass spectrometry (NI/MS) has been developed into a powerful tool for the selective detection and quantitation of electrophilic molecules. Much of the original work on negative ion formation was conducted by electron swarm techniques with the goal being the determination of basic molecular parameters such as electron affinity (EA) or ionization potential (1, 2). It was not until high source pressure ionization experiments were conducted which demonstrated an increase in the production of negative ion current by a molecule, that the potential of NI/MS as an analytical technique was recognized (3,4). Since these first experiments, the field and applications of NI/MS have grown extensively, becoming a major force for detection and analysis in many scientific disciplines. NI/MS has several advantages over the traditional form 159 of mass spectrometry which used electron impact for the ionization of molecules. First, the ionization process is not indiscriminent as is electron impact, but is selective for electrophilic molecules (5). Secondly, the spectra produced by the capture of a low energy electron are usually simple, consisting of only one or two major peaks. In addition to their simplicity, the spectra for some analytes are more intense than those produced by electron impact mass spectrometry (6). These two characteristics make the use of NI/MS ideally suited for selected ion monitoring (SIM). Finally because of the selective nature of the electron capture process, many problems associated with matrix effects with environmental or biological samples are avoided. Lipids which might be present in environmental samples as a result of the extraction procedure, would be invisible to the electron capture process (7)(Figure 1). Thus, only the electrophilic xenobiotics in the matrix would be detected with limited background interferences. Because of their electrophilicity, limits of detection for many xenobiotics using NI/MS for quantitation are comparable and in some instances when SIM is employed, lower than those of the more traditional methodology of gas chromatography with electron capture detection. The use of negative ions for detection is becoming more common as instruments easily capable of preforming these analyses become more available. As a result, the interest in 160 :5 2: 13 4:) 1 4331114.... ‘5 80 I00 I20 I40 ISO ISO 200 220 240 260 280 300 320 340 360 380 400 3 ac» (b) ‘5 GOP ‘15 40> C: firfrvrt rrTTTL r‘hfrr'rf—rttrrr4LF-Fh—1 IZO I40 ISO I80 200 220 240 250 280 300 320 340 360 380 400 420 m/z Figure 1. (3) +01 and (b) NI mass spectra of a cleaned up chicken extract showing the presence of dieldrin (7). 161 the use of negative ions as a viable analytical technique for the analysis of xenobiotics in the environment increases as well. In a recent study (8), it has been demonstrated that if similar instrumental conditions were used by different laboratories the spectra produced by each independent laboratory were similar to one another. Until this interlaboratory study had been conducted, one of the greatest perceived disadvantages of NI/MS was thought to be its irreproducibility between laboratories. With this last problem eliminated, the use of NI/MS in environmental chemistry may be extended further. Negative Ion Formation The Electron Capture Process The formation of a negative ion by electron capture proceeds through the formation of a transient negative parent ion. After this point, the transient negative ion may undergo a number of different reactions which are dependent upon electron affinity of the original analyte and the initial energy of the electron. Negative ions are produced as a result of electron/molecule interactions by three general processes depicted for the molecule Mk: 1. Ion-Pair Formation: e'+MX -------- >M*+X’+e’ 162 2. Electron Attachment e' + Mx -------- > MK“ 3. Dissociative Electron Attachment: e' + MX -------- > M + X‘ These electron molecule interactions are better understood with reference to the potential energy diagrams shown in Figure 2. In ion-pair formation, Figure 2a, the electron is energetic enough that the initial electron/molecule interaction results in an excited state species which then dissociated to a positive and negative ion. The threshold energy required for ion pair formation is in the 10 to 20 eV range (9). Dissociative electron capture occurs if the capture of an electron by a molecule leads to a vertical Franck-Condon transition producing a repulsive state of MX' (Figure 2b). The repulsive intermediate would then dissociate to M° and X' (9). For electron attachment to occur, there are two likely pathways, which a dependent upon the inherent electron affinity (EA) of the molecule. If EA is less than 0, a Franck-Condon vertical transition leads to an unstable [MX‘J' molecular anion (Figure 2c). This unstable entity may either disappear by autodetachment of the electron from the molecule or by dissociation to M' and X'. Lifetimes for such intermediates are very short and are therefor incapable of 163 (a) ( b) MX" M+X N - M“ M+X" LV L MX’ Mx M+X MX W. M“ M+X' , M+x' (c) (d ) Figure 2. Potential energy curves for negative ion formation (9). 164 being stabilized by collisional deactivation. If the electron affinity is greater than 0, the potential energy curve for MX'may or may not intersect with a Franck-Condon vertical transition region. If such an intersection does occur, the ensuing reactions are similar to those described for the same situation if EA<0, shown in Figure 2c. If not, the attachment reaction is different in that the anionic molecular ion is a longer lived species as a result of interactions with thermal energy electrons. As such, any internal energy as result of the interaction may be redistributed throughout the vibrational levels of the molecule and a stable molecular anion may then be produced by collisional stabilization (Figure 2d). All electron capture processes are resonant processes as no electron produced is capable of removing any excess energy. These reactions are observed at electron energies ranging from 0 eV (thermal electrons) to <15 eV, with low energy interactions producing electron attachment reactions and higher energies producing dissociative electron interactions. The Effect of Molecular Structure For the transient negative ion to form a stable molecular ion without autodetachment of the electron or fragmentation into neutrals and ions, the original molecule’s electron affinity must be positive (5). The electron 165 affinity of a molecule increases with increasing halogenation until it becomes positive. This has been documented in the case of a series of fluorinated benzenes. In addition to . increasing EA with increasing degrees of fluorination, a concomitant increase in the lifetime of the molecular negative ion was observed (10). These transient negative molecular ions must be stabilized by inelastic collisions in the source region to ensure detection. Overall molecular structure will also affect the position of the stable negative ion resonance states (NIRs) of a molecule. The addition of a double bond to a system increases the EA, and a second double bond increases it further (5). Addition of an aliphatic group decreases the electron affinity of a molecule, by increasing the molecule’s aliphatic character. As mentioned above, a halogen in place of a hydrogen atom increases the EA (10). Thus the type and number of substituents around a basic molecular system appear to influence the EA of the overall molecule. Effect of Moderating/Reagent Gas The moderating/reagent gas is introduced into the source of the instrument to serve two major functions. First, the interaction of the electron beam and the moderating gas serves as a ready source of low energy electrons. Methane, a commonly used moderating gas, is ionized by the primary 166 electrons produced by the source filament, thus producing low energy electrons, as well as ions and radicals as shown by equations 4 and 5 (8). CH4 + eprlmgpy """" > CH4 + eeecondary + eprimary (4) CH4 + eprlmary """" >CH4 + esecondary+H'+eprimary (5) The secondary electrons undergo further inelastic and elastic collisions in the source reducing the energy spread of the secondary electrons. The final energy spread of the secondary electrons is dependent upon the ion source pressure, the moderating gas pressure and the electron fields in the source (11). The second major function of the moderating gas is the enhancement of the analyte anions by collisional stabilization. The source pressure must be operated at pressures of approximately 1 Torr for significant collisional stabilization to occur. Under these conditions, the lifetime of the intermediate anionic species must be > 1 psec for collisional deactivation to have a significant effect (9). Though methane is perhaps the most commonly utilized moderating gas, a variety of other monoatomic and polyatomic gases have been investigated for their applicability to NI/MS (12). Such studies have reveals that molecular size and shape of the moderating gas are important parameters for both the formation of secondary electrons and affect the capacity 167 of the gas to collisionally stabilize anionic species. Monoatomic gas species exhibited a direct linear correlation between the degree of electron attachment observed for an analyte and increasing atomic number, which corresponds directly with atomic radius. Polyatomic molecules demonstrated an increase in overall attachment rate with the available degrees of freedom of the moderating gas molecule. The relative effectiveness of the polyatomic gases to stabilized negative ions with excess internal energy would be expected to increase then, with the availability of vibrational and rotational modes in the moderating gas. Polyatomic gases would be expected, on this basis, to be more effective for collisional stabilization than monoatomic species of corresponding size. In general, source pressure for any moderating gas should be such, that the formation of secondary electrons are maximized but not so dense that the ion beam produced by the analyte is scattered by the diffusion of the moderating gas in the ion optics region of the instrument (12). The Effect of Source Temperature Source temperature has been found the affect the degree to which a molecule will fragment in NI/MS. High pressure studies of the positive ion spectra of benzene demonstrated dramatic changes in the spectral composition with changing 168 temperature (13). With increasing temperature, an increase in the relative abundance of fragments occurred and at the same time, the formation of a benzene adduct decreased . Similar results have been observed for a variety of electrophilic compounds analyzed by NI/MS. Cyclopentadiene derivative pesticides (14), chlorinated phenol derivatives (15), PBB isomers (16) and polychlorinated ethers (17) all demonstrated a decrease in the relative abundance of the molecular anion with increasing temperature. This phenomenon was found to be more pronounced for higher chlorinated species and species which were capable of nondissociative electron capture. The observed increase in fragmentation at elevated temperatures has been explained that at these elevated source temperatures, the initial electron/molecule interaction may not necessarily occur at the ground vibrational level of the molecule (9). The extra thermal energy is great enough to enhance the degree of fragmentation which occurs and thus decrease the relative abundance of the molecular anion species. Because of this effect, the source would remain as cool as possible which conducting an NI/MS experiment. For reproducibility between laboratories, the source temperature becomes an important parameter (8). Applications 169 Molecules which readily undergo formation of stable negative ions must have a high electrophilic character as discussed in the previous section. Many compounds of environmental concern exhibit such characteristics. As a result, the use of negative ion mass spectrometry has been extensively studied for the low level detection and quantitation of a variety of xenobiotics in a variety of environmental matrices. Chlorinated Pesticides Some of the first investigations of negative ion spectra of chlorinated pesticides were undertaken by R. C. Dougherty et a1. (18) at Florida State University. The first negative ion spectra of a variety of hexachlorocyclopentadiene derivative pesticides were reported with methane as the moderating gas. These included such compounds as a- and T-chlordane, trans-nonachlor, dieldrin and aldrin, among others. These compounds are composed of two or more fused aliphatic ring systems with varying degrees of chlorine substitution around the molecule. The analytes were introduced into the source by direct probe at high concentrations (10 to 100 pg). As a result of both high analyte concentration and impure methane, the spectra showed numerous adducts as well as fragments indicative of electron capture. These adducts included the following sorts of 170 fragments: (M+ClO)‘, (M+OH)', and (M+O-Cl)'. Ions indicative of the electron capture process were (M)', (M-Cl)‘, and (M-HCl)’. Though a stable molecular ion was observed for all compounds except hydroxychlordene, the author noted that there was no intensity enhancement of the molecular ion if compared to its intensity by electron impact. However, overall spectra were much simplified from those produced by electron impact. The formation of the (M+Cl)' adduct for these compounds has been recently show to be a function of analyte concentration in the source (14). With increasing source concentration, there was a simultaneous increase in (M+Cl)' formation observed. Adducts of oxygen, such as those described above, were not observed in the more recent study (14) demonstrating the importance of the use of high grade methane for the moderating gas. Lower concentrations of these compounds were introduced into the source by GC/MS simulating those concentrations more indicative of concentrations found in environmental matrices. The spectra of these same compounds had their base peaks at (M+H-Cl)’ ions with subsequent losses of Cl. Negative ion behavior of the family of DDT pesticides and metabolites has also been explored. Dougherty and Roberts (19) observed that the NI spectra of these compounds were dominated by (M+Cl)' adducts. These adducts were attributed to the formation of the chloride ion by the 171 chlorinated analyte itself with subsequent reactions of the Cl‘ with the non-ionized analyte in the source. The only other significant ions observed for this class of compounds were the (M-H)‘ for p,p'-DDT and (M+H)‘ for o,p'-DDT. Further fragmentation was not observed in most instances. At higher analyte concentrations, the formation of DDT dimers were seen. It was pointed out that in most environmental samples such high concentrations of DDT would be highly unlikely and therefor dimer formation should not be a problem. In a more recent examination of NI/MS spectra of DDT and its analogs, Stemmler and Hites (20) noted that the spectra of these compounds were dominated by the molecular ion of the analyte and (M-Cl)‘. It was pointed out that the concentrations used by Stemmler and Hites (20) were several orders of magnitude lower than the original study conducted by Dougherty and Roberts (19) and therefore more indicative of concentrations found in environmental samples. It was noted as with the hexachlorocyclopentadiene derivative pesticides, that the concentration of the analyte affected the spectrum of the compounds. At higher concentrations, in the pg range, the adduct (M+Cl)‘ appeared while at environmentally significant levels (1 to 10 ng) these adducts were not detected. The spectra for the DDTs exhibited a greater degree of fragmentation overall than previously reported (19). More importantly, however, the fragmentation was found to be indicative of the aliphatic character 172 associated with the molecule. For example, a stable molecular ion was not detected for DDT isomers, but rather the loss of HCl was observed. This led to the formation of a stable diphenyl ethylene negative ion. Further, for the various DDTs there was exhibited some isomer specificity in that one isomer was distiguishable from another by their spectra. The use of negative ion mass spectrometry has been extended from the detection of single compounds to the detection, confirmation and quantitation of complex mixtures. Such was the case with toxaphene, a complex chlorinated pesticide mixture consisting of some 670 components. The mixture contains a variety of isomers and congeners of camphene and bornane with degree of chlorination varying from six to ten chlorine substituents. The use of negative ion mass spectrometry for toxaphene was first described by Ribick et a1. (6) for detection of toxaphene at the part per billion (ppb) level. NI/MS was found to be ideally suited for the detection of toxaphene components. Rather than the complex spectra and fragmentation patterns produced by electron impact and positive chemical ionization, spectra were simple, consisting of two fragments; an (M-Cl)‘ and (M-HC12)'. This was observed for a single toxaphene component, Toxicant B, which has a bornane structure substituted with seven chlorine atoms. In addition, the ratio of the total ion current for NI/EI was greater than 100. Interferences which would 173 normally compound the difficulty of the analysis by EI, such as lipids, hydrocarbons and phthalates were not detectable by NI/MS. The toxaphene components produced unique stable ions which could easily be distiguished from other similar chlorinated compounds such as chlordanes. In actual sample analysis (21) by capillary GC/MS, chlordanes were found to significantly interfere with the quantitation of toxaphene mixture. This was circumvented by adjusting a series of retention time windows to monitor chlordanes and subtract any interferences produced by chlordane components from the toxaphene mixture. The use of a modified SIM program has been further developed for the quantitation of toxaphene (22, 23). The mass spectrometer was set to monitor certain a.m.u. windows at certain times during the chromatographic run. These ions were representative of the various chlorinated isomers of the toxaphene mixture. Additional ions were monitored for interferences such as chlordanes within specific retention time windows. The ion current representative of the interferences would be subtracted from the TIC after the completion of the run, with the final adjusted TIC being compared to a standard curve for toxaphene. Using this series of windows, concentrations of toxaphene were determined in fish from Lake Huron and Siskiwit Lake, Isle Royale, Michigan at the high ppb level. Through the use of NI/MS, the detection and quantitation of toxaphene has been 174 accomplished in an environmental matrix. This was achieved because of the simplicity of the spectra produced by the electron capture negative ion formation process. Polychlorinated and Polybroninated Biphenyls Polychlorinated biphenyls (PCBs) and polybrominated biphenyls (PBBs) are complex mixtures made up of various halogenated congeners and isomers. PCBs in the form of Arochlors were originally used in transformers and capacitors because of their insulating properties. PBBs in the form of Fimemaster, were used as fire retardants in clothing. Both of these halogenated species have spread throughout the environment. In particular, PCBs are commonly found in most aquatic biota at the parts per million level (ppm). Because of their halogenation and aromaticity, these compounds are amenable to analysis by NI/MS. Crow et al. used various combinations of reagent gases to test their selectivity for the analysis of PCBs by NI/MS (24). These were methane, for electron capture negative ion formation, methane/methylene chloride, to investigate chloride ion attachment and methane/oxygen, to investigate 0‘ adduct formation. Of these three combinations, the formation of the oxygen adduct was the most sensitive for chlorinated biphenyls. The formation of molecular anions and dissociative electron capture were not favored. In 175 particular, the lower chlorinated species produced little or no detectable ion current. Those isomers which were detectable, produced spectra which were made up of the Cl'. It was also observed that negative ion formation, whether as an adduct or by electron capture, was more favored by the higher chlorinated species in the Arochlor mixtures. The formation of halogen anions by PBBs, PCBs and bromobenzenes was investigated by electron capture NI/MS (16). Methane and N; were used as moderating gases. The PBB and PCB spectra consisted of two isotopic clusters, representing the Br‘ or Cl‘ and [M'] for the specific isomer. The moderating gas was not found to affect the pattern of the spectra but was found to enhance the formation of the halogen anion. N2 produced spectra with a greater relative abundance of either halogen anion. This was explained as a consequence of the size of the moderating gas. The larger methane molecule was more efficient at removing excess energy by collisional deactivation than the smaller N2 molecule. With either moderating gas, the ratio of halogen anion to molecular anion decreased logarithmically with increasing halogenation around a molecular system. This logarithmic decrease was also found to be dependent upon the position of the halogens on the molecular system. Since all PBB isomers were found to produce Br' in abundance by dissociative electron capture, it was use as a means to monitor the presence of PBBs in biological samples 176 (25). Three ions were monitored: m/z=79, 81 and 498 representing the Br‘ isotopes and decachlorobiphenyl which was added as an internal standard. As Br' is not a commonly occurring atom in biological samples, it was assumed that the only source would be from brominated xenobiotics present as trace contaminants. The identity of specific isomers relied on the isomers' retention time. Using this method, the limit of detection was reduced to 2 picograms (pg) of Firemaster B-6 in serum. This was equivalent to 1.4 pg of the hexabromo isomer present in the PBB mixture at the higher concentration. This represented a 20 fold increase in detection limits over what had previously been accomplished by conventional GC with EC detection. A variety of gas mixtures were used to determine the best method for residue level detection of PCBs in marine sediment (26). The mixtures used were CH4/02, Ar/Oz, CH4/H20 and CH4 alone. By using these combinations of reagent and moderating gases, both the electron capture ability and the adduct reactivity of PCBs were investigated. In all instances, the highly chlorinated species of the PCBs mixtures produced a greater abundance of ions than their lower chlorinated counter parts. The molecular ion for the chlorinated isomers was detected only when methane alone was used as the moderating gas for electron capture NI/MS. In all other gas mixtures, oxygenated species representative of the isomers were present in the greatest abundance in the 177 spectra. PCBs and PBBs when analyzed by NI/MS produce spectra which consist of essentially two major ions; the corresponding halogen anion and molecular anion. In addition, the formation of ions by electron capture is selective towards those isomers having a greater degree of halogen substitution. Even so, the overall sensitivity and selectivity gained through the use of NI/MS by the elimination of matrix interferences compensates and produces an overall enhancement in sensitivity over traditional NI/MS. other Xenobiotics There are a wide variety of additional xenobiotics which fall neither into the halogenated biphenyl category or that of the chlorinated pesticide category. These include polychlorinated dibenzo-p-dioxins and dibenzofurans, chlorinated phenols and related ethers, and halogenated benzenes. All of these compounds have a high electrophilic character and have been analyzed by NI/MS. Of this group, perhaps the most widely studied class of compounds are the dioxins and dibenzofurans (27, 28). Dioxins and dibenzofurans have been shown to be a by-product of the production of the herbicide 2,4,5-T, and chlorinated phenols. In addition, it has been demonstrated that dioxins are produced by the incomplete combustion and are therefore 178 ubiquitous in the environment. Because these compounds are halogenated and aromatic, they have been shown to be amenable to the use of NI/MS for their detection and quantitation. Wood treated with chlorinated phenols and cow liver and fat were analyzed for chlorinated dioxins (27). NI/MS and positive CI were compared for dioxin detection with methane used in both instances as the reagent gas. Electron capture NI/MS produced the greatest overall response. Limits of detection of 2,3,7.8-TCDD and octachlorodibenzo-p-dioxin were found to be 0.2 and 0.5 pg respectively. Levels in the tissue and wood were quantitated in the low part per trillion (ppt) level. Derivatized chlorinated phenols have been analyzed by NI/MS (15, 29). The tri-, tetra-, and pentachlorphenol isomers were derivatized to their pentafluorobenzylethers, pentafluorobenzoesters, heptafluorobenzylesters, trifluoroacetates, acetates and 2,4-dinitrophenyl ether derivatives (29). The acetoxy derivatives of the tri-, tetra-, and pentachlorophenol were examined for their sensitivity to NI/MS using methane as the moderating gas. As with the other chlorinated aromatics, it was observed that the greater the chlorination, the greater the negative ion response produced by an individual molecule. These spectra were consistent with those produced by electron capture having a single major ion detected: that of the molecular ion or the molecular ion less one hydrogen atom. In contrast, 179 Trainor and Vouros (15) found that the electron capture NI/MS spectra of derivatized chlorinated phenols were not representative of the original molecule. However those which were representative of the original molecule, were found to be strongly affected bye the original degree and position of halogen substitution Polychlorinated diphenyl ethers are contaminants in technical chlorophenol preparations at the 1 to 5% level. OH C1 C1 C1 General structure of polychlorobiphenyl ethers These unusual compounds are of environmental concern because they have been shown to undergo ring closure by either photochemical or thermal processes to produce chlorinated dioxins. These species have been analyzed by both methane (30) and methane-oxygen enhanced (17) negative ion mass spectrometry. Methane enhanced spectra were found to be dominated by either the M' or [M-C1]' ion of each isomer. Methane-oxygen enhanced spectra of polychlorinated diphenyl ethers showed species indicative of both oxygen adducts and electron capture. Higher chlorinated species were found to 180 produce a greater abundance of adduct ions than the lower chlorinated isomers. The limits of detection and sensitivity however were not determined. Chlorinated and bromobenzenes are additional xenobiotics which have the electrophilic characteristics necessary for NI/MS (16, 24). When compared with electron impact ionization, their spectra are more intense and dependent upon the degree and position of the halogenation. The lower halogenated species produced spectra with halogen anions as the base peak. The molecular anion did not become the most abundant ion until five or six halogen atoms were substituted around the benzene system (24). With bromobenzenes, (l6)it was also observed that the ratio of [Br‘J/[M'] decreased logarithmically with increasing bromine substitution. In general, the electron capture spectra of chlorinated xenobiotics are dominated by their halogen and molecular anion. In some instances, losses of Cl‘ and Br’ are also observed. However, the structural information obtainable from such spectra is limited. The major advantage of NI/MS lies in its intense and simple spectra making the process amenable to SIM. NI/MS as a Rapid Screening Technique for Environmental Samples Since many xenobiotics are amenable to either electron 181 capture or the production of negative ions as adduct species, rapid screening methods for their simultaneous extraction and analysis have been developed (7, 11, 31, 32). The screening method is composed of a simple extraction procedure, followed by direct probe introduction of the extract into the source (7, 31, 32). Though the use of direct probe for sample introduction has many disadvantages (31), it was noted that certain compounds which were not detected by GC/MS were detected if direct probe introduction were used on the same sample. A variety of samples were extracted using continuous steam distilled liquid-liquid extraction (7, 31, 32). Sample sizes were many times as small as 250 mg of material, with recoveries in the 90% range for PCBs, PCDDs, and polycyclic pesticides. Higher molecular weight and higher chlorinated species were not found to be as efficiently extracted. For extracts with greater lipid content, a separation step utilizing gel permeation chromatography was required (31). Source conditions were adjusted such that either electron capture or adducts representative of additional Cl'or 0; would be formed depending upon the compounds present in the samples. Polychloronaphthalenes, PCDDs, and PCDFs were detected in carp taken from the Tittabawassee River downstream of Dow Chemical, Midland, MI. Further, an extract of lake trout from Lake Ontario showed that various chlorinated phenols, hexachlorobenzene (HCB) and a variety of chlorinated pesticides were present (Figure 3) (32). In INIENSITY RELATNE 0' IOO P Q )- Cl 80 _ 5 mil 263 60 r- h— 0' 40 r- " 04 0 250 Figure 3. CI4 m/Z 305 Q Cl6 182 ICI C-IC \ CI 0 CI3 ml 323 0' as 0 mil 339 0' / (‘zl /C' O‘ gave-=6». a - a 0 04 mm tall 357 ca, CLORDANE CI (“/2 407 «VI 44! ”I m/Z 3736 NONACHLOR m/Z 475 MEI I I ,. (“in .. I”? A. 350 400 450 500 MASS NUMBER Negative ion mass spectrum of 50 mg of Lake Ontario Lake Trout cleaned up by GPC. Only 12C negative mass defect ions are shown for polychlorinated chemicals (32). 183 human seminal fluid, a variety of polychlorinated compounds indicative of chlorinated phenols, and PCBs as well as several unidentified C17 containing compounds were identified. A 250 mg sample of human adipose tissue was found to be contaminated with chlordane, heptachlor epoxide, nonachlor, PCBs, HCB and chlorinated phenols (7) 10. 11. 12. 13. 14. 184 References J. J. Decorpo, and J. L. Franklin, J P . 1971 65 1885-1888. R. E. Fox, and R. K. Curran, 61_§hgm1_£ny§6 1961 26’1595-1601. R. C. Dougherty, and C. R. Weisenberger, 62_Am1_§h6m1 29;; 1968 22 6570-6571. D. F. Hunt, G. C. Stafford Jr., F. W. Crow, and J. W. Russell, Aggl1_gngm6 1976 62 2098-2105. L- G- Christophorou, Enxironl_nealth_£er§neel 1980 16 3-32. M. A. Ribick, G. R. Dubay, J. D. Petty, D. L. Stalling, and c. J. Schmitt, Engizgn1_§611_zggh6 1982 16 310-318. D. W. Kuehl, M. J. Whitaker, and R. C. Dougherty, 53616 Q2236 1980 62 935-940. E. A. Stemmler, R. A. Hites, B. Arbogast, W. L. Budde, M. L. Deinzer, R. C. Dougherty, J. W. Eichelberger, R. L. Foltz, C. Grimm, E. P. Grimsrud, C. Sakashita, and L. J. Sears, 53611_thm6 1988 66 781-787. A- G- Harrison. in Qhemi2al_I6nizatien_ua§s_§nestrgmetrx 1983 , CRC Press, Inc., Boca Raton, FL. J. R. Frazier, L. G. Christophorou, J. G. Carter and H. C. Schweinler, g, Chgm, Rhys, 1978 62 3807-3818. R. C. Daugherty, An§16_§D§E; 1981 22 625A-636A. I- K- Gregor. and M. Guilhaus. Inti_ll_hass_§nesl_lon_ thsl 1984 64 167-176. F. H. Field, P. Hamlet, and W. F. Libbey, J, Am. Chgm. $09, 1969 21 2839-2842. E. A. Stemmler, and R. A. Hites, An611_gngm6 1985 61 684-692. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 185 T. M. Trainor, and P. Vouros, 83611.2hfim1 1987 62 684-710. J. Greaves, J. G. Bekesi, and J. Roboz, 216m661_M6§§ §p§§1 1982 2 406-410. K. L. Busch, A. Norstrém, M. M. Bursey, J. R. Haas, C. -A. Nilsson, 216E661_M§§§_§pg§1 1970 6 157-161. R. C. Dougherty, J. Dalton, and F. J. Biros, ng6_ng§§ 5.9.9.9... 1972 6 1171-1181. R. C. Dougherty, J. D. Roberts, and F. J. Biros, A6616 CHEM; 1975‘11 54-59. E. A. Stemmler, and R. A. Hites, 53611_§n6m1 1988 66 787-792. R- Jansson. and U- Wideqvist, W 911611.. 1983 1.1 309-321. D. L. Swackhamer, J. M. Charles, and R. A. Hites, 34th Annual Conference on Mass Spectrometry and Applied Topics, 1986, Cincinnati, OH. D. L. Swackhamer, J. M. Charles, and R. A. Hites, 53616 Qh§m1 1987 62 912-917. F. W. Crow, A. Bjorseth, K. T. Knapp, and R. Bennett, Ans11_§n§m1 1931 21 619-625- J. Roboz, J. Greaves, J. F. Holland and J. G. Bekesi, An§16_§ngm1 1982 21 1104-1108. E. Lewis, and W- D- Jamieson. W123 m 1983 4.8 303-306. J. R. Haas, M. D. Friesen, D. J. Harvan, and C. E. Parker, Angl1_ghgm1 1978 66 1474-1479. C. Rappe, H. R. Buser, D. L. Stalling, L. M. Smith, and R. C. Dougherty, HQEBIQ 1981 222 836-838. S. Shang-Zhi, and A. M. Duffield, 16_§hrgm1 1984 266 157-165. K. L. Busch, A. Norstrém, C. -A. Nilsson, M. M. Bursey and J- R- Haas. W 1980 3.6 125-132- R- C- Daugherty. 91mm 1981 6 283-292. D. W. Kuehl, and R. C. Dougherty, AQYI_MQ§§_§E§§1 1980 186 QB 1451-1459. 187 Section II Chapter 2 SELECTIVE SENSITIVITY OP HIGHLY CHLORINLTED SPECIES IN NEGATIVE ION MASS SPECTROMETRY (NI/MS): IMPLICATIONS EOR CONPLEX MIXTURE ANALYSIS Introduction Negative ion mass spectrometry (NI/MS) has been used for the quantitation of a variety of compounds of environmental concern. These include PCBs(1), dibenzodioxins and dibenzofurans(2,3), chlorinated phenols (4,5), hexachlorocyclopentadiene derivative pesticides such as chlordane (6), DDTs (7), polynuclear aromatic hydrocarbons (PAHs) (8), and toxaphene (9,10,11). With exception of the PARS, all compounds listed above are chlorinated species. Because of the electronegative chlorine atom, chlorinated compounds have a higher cross section for the NI/MS process than their non-chlorinated analogs. Dioxins, for example, have been easily detected by NI/MS in the low part per trillion level when using selected ion monitoring (SIM) (3). Toxaphene, a complex chlorinated pesticide mixture, has been quantitated using a modified SIM method at the low part per billion level (9). However, despite the advantages of NI/MS, there are some difficulties with the technique especially when quantitating complex mixtures such as toxaphene. In working with toxaphene residues, it was noted that significantly different responses were recorded when 188 analyzing the same residue by EI/MS and then by NI/MS. The later-eluting or higher chlorinated compounds exhibited more of a preferential response to NI/MS than to EI/MS. When low levels of toxaphene were analyzed by the two techniques, the differences in detection of higher and lesser chlorinated species became more exaggerated. For example, a total ion chromatogram of the complete mixture by NI/MS indicated that only compounds having a high degree of chlorination were detected. These observations prompted a systematic study involving three series of chlorinated compounds. The first consisted of compounds which were similar in structure to toxaphene components; the other series consisted of isomers and congeners of chlorinated phenols and chlorinated biphenyls. The goal was to characterize the response of chlorinated species as a function of degree of chlorination. A secondary objective was to determine optimum conditions for analysis of complex residues containing low levels of chlorinated hydrocarbons. 189 Experimental Chemicals: PCB isomers were purchased from Foxboro Analabs, North Haven, CT and Accustandard, New Haven, CT. The chlorinated phenols and camphene were purchased from Aldrich Chemicals, Milwaukee, WI. cis- and trans-chlordane and trans-nonachlor were generously donated by Dr. F. Matsumura. Heptachlor, heptachlor epoxide, aldrin, dieldrin, and endrin were received as standards from the EPA. All compounds, with the exception of camphene, were used as received with no further purification. Technical camphene was recrystallized three times in ethanol to 95% purity as determined by capillary GLC with FID. Dichlorocamphene and 2,10-dichloronorbornane were synthesized from camphene using the method of Jennings and Herschbach (12). The two-component mixture was purified by vacuum distillation followed by recrystallization in ethanol. The mixture was determined to be 2:1 dichlorocamphene to 2,10-dichloronorbornane as quantitated using GC/MS by electron impact ionization. Solutions of PCB isomers and chlorinated pesticides were made up in hexane to a concentration of 200 ppm. The chlorinated phenols were dissolved in methanol to a concentration of 200 to 250 ppm. All solvents were glass 190 distilled. These stock solutions were then diluted as necessary for capillary GC/ECD and NI/MS analysis. ECD Analysis: Response factors were determined using a capillary GLC with electron capture detection. A DB-l J & W 60 M x .25mm ID column in a Perkin Elmer GLC equipped with a 63Ni detector was used for separation and detection. Column head pressure was 25 psi producing a flow rate of approximately 1 ml/min. The GLC oven was temperature programmed from 190° to 270° at 2°C per minute with a two-minute hold at 190° prior to ramping and an lS-minute hold at 270°. Total run time was one hour. The injector, at 220°, was operated in the split mode with a 3 to 1 split. Aliquots of 1 p1 of the 10 ppm standard solutions were injected for the determination of response. Molar response was calculated using the area integrated by the Perkin Elmer data system divided by the total number of nanomoles of injected compound. Mass Spectrometry: Solutions of 10 ppm of each standard were used for GC/MS. One microliter of each solution was injected into a J & W DB-1 30 M column in a Delsi Di700 gas chromatograph which was directly interfaced to the source of a Nermag R10-1OC 191 quadrupole mass spectrometer. The GLC oven was temperature programmed from 100° to 270° C holding at 100°C for 1 minute followed by 30° C/min to 190° then ramped from 190° to 270° at 5° C/min. A J 8 W on-column injector was used for sample introduction. The column head pressure was set at 15 psi. The heated interface between the gas chromatograph and mass spectrometer source was maintained at 250°C throughout the course of the experiments. For EI/MS analysis, the source temperature was maintained at 200°C. An electron beam of 200 mA at 70 eV was used for ionization. The mass scan ranges were from 35 u to 450 u for the chlorinated pesticides and chlorinated biphenyls, and from 35 u to 300 u for chlorinated phenols. The quadrupoles were scanned at a rate of 2 scans per second. The multiplier was set at -2.48 kV with the conversion dynode at -4.83 kV. NI/MS conditions varied for each series as described below. Methane was used exclusively as the moderating gas. Methane pressure was measured adjacent to the source by a thermocouple gauge. For the model toxaphene system, source pressure was 0.04 Torr with an ionizing current of 200mA at 70 eV. Pressure for the PCB isomers was 0.1 Torr with 159 mA to 165 mA ionizing current with 70 eV electron energy. Chlorinated phenol NI/MS conditions were 0.085 Torr moderating gas pressure with 130 mA ionizing current at 70 eV electron energy. The source temperature was maintained in 192 all cases at 100° C. The electron multiplier was set at -2.59 kV with the conversion dynode at +4.87 kV. Mass scan ranges for each of the three series remained as those described above for EI/MS. Molar response was calculated using the area of each chromatographic peak as integrated by the SIDAR data system of the Nermag R10-10C. As with the ECD analysis, total ion current area was divided by the total number of nanomoles of injected compound. Results The calculated molar response factors for three series of chlorinated compounds are presented below. Results show that the degree to which a molecule is chlorine substituted, and in some instances, the positions of the chlorine atoms on the molecule, determine the molar response for a compound. Tabulated spectral data for all compounds are also presented. Chlorinated Pesticides as a Model Toxaphene Mixture Toxaphene is a complex mixture consisting of mainly chlorinated norbornanes and camphenes. The majority of components comprising the mixture have 6 to 9 chlorine atoms. Since individual toxaphene components are not easily obtainable, a group of readily available compounds was 193 selected which were similar in structure to toxaphene components. Aldrin, dieldrin, endrin, heptachlor, heptachlor epoxide, cis and trans-chlordane, and trans-nonachlor were selected as representative of the toxaphene mixture (Figure 4). These compounds exhibit the bicyclic non-aromatic structures common to components of toxaphene. All have the bridged cyclohexane ring system with varying degrees of chlorine substitution. Camphene, dichlorocamphene, and 2,10-dichloronorbornane were added to determine the degree of chlorination necessary to elicit an NI and ECD response. When analyzed using EI/MS, the molar responses calculated for these compounds were within one order of magnitude of one another (Figure 5). The molar response factors varied from 3.9 x10 7to 12.0 x 107 TIC/nanomole. When the same mixture was analyzed using NI/MS, the molar response varied from non-detectable for camphene, dichlorocamphene, and 2,10-dich1oronorbornane with 0, 2, and 2 chlorine atoms, respectively, to a maximum of 123 x 107 arbitrary units/nanomole for trans-nonachlor, with 9 chlorine atoms. As shown by Figure 5, the response in NI/MS is logarithmic with respect to the number of chlorine atoms in the molecule. It is of interest to note that those compounds in the mixture which also were substituted with oxygen, i.e., dieldrin, heptachlorepoxide and endrin, showed no difference in NI/Ms molar response from their non-oxygenated analogs, aldrin and heptachlor. 194 Figure 4. Chlorinated bicyclic compounds used to mimic toxaphene constituents: (1) camphene, (2) dichlorocamphene, (3) 2,10-dichloronorbornane, (4)aldrin, (5) dieldrin and endrin, (6) heptachlor, (7) heptachlorepoxide, (8) cis- and trans- chlordane, (9) nonachlor. 195 .pcsomeoo some you mcofiHOOncu oumofiamou moan» mo oomuo>o on» musmmounou unwon zoom .huuosouuoonm mama cow 0>wuooos can Hosea“ souuooHo an couscoum upssomaoo ofiao>0an mousswuoHnO no o>uso omsoauou undo: mEB< octoio Co 1% m h m 0 ¢ 0 _F .m musmfim N _ _ _ _ _ g :0. 96.092 o “009.5 coboofi o woofio; Imo+mo.N Imo+mo.¢ T Imo+wo.o ' [mo+mo.m I Imo+wo. — I Imo+uw. P punodwog 1o e|owouoN/Q|l 196 Other compounds were remarkably insensitive to NI/MS. For example, 2,10-dichloronorbornane and dichlorocamphene produced no response in NI/MS at the 10 ng level. Even when greater amounts of these compounds were injected (100 ng), there was no detectable signal by NI/MS resulting from their ionization, though a response was produced in ECD. When the same solution containing lOOng/ul of dichlorocamphene and 2,10-dichloronorbornane was analyzed by capillary GC with ECD detection, they produced a nanomolar response of 89.7 and 96.7, respectively. It seemed that some threshold of electrophilicity must be reached by the molecule before it is capable of producing a signal under NI/MS conditions. Once reached, there was a dramatic increase in response with increasing degree of chlorination on the molecule. However, since an ECD response was produced by the dichlorinated compounds, it would appear that this threshold may not be as critical in ECD or is lower than it is for NI/MS. All other compounds with the exception of the dichloro isomers and camphene, when analyzed by ECD produced molar responses within the same order of magnitude despite their differing degrees of chlorination (Figure 6). This would explain why the similarities between EI and ECD toxaphene chromatograms occurred while the same similarities were not observed when EI results were compared to those from NI/MS. The NI/MS spectrum of each of the chlorinated bicyclic compounds making up the model toxaphene system varied with 197 .cofiuomumo Aomc 0090680 souuooam an“: 0O annanwmoo >3 omoscoum upcsomaoo Dadowofin mouosHHoHno mo 0>uso omsonmou Hoaoz .m ousowm mEB< 05.6.50 .6 4% o v r — P — b -00 hCfl CD punodwog 1o e|owouoN/esuodsea (133 198 Table 1 Negative Ion Spectra of Chlorinated Bicyclic Compounds Compound [MJ' [M-C1]‘ [M-ClZJ' [HC12]' [ClJ' Camphene N/D1 2,10-dichloronorbornane N/D dichlorocamphene N/D Aldrin T2 5 --- --- 100 Dieldrin 4 5 --- --- 100 Endrin 5 6 --- --- 100 Heptachlor 5 4 67 7 100 Heptachlor 10 2 T --- 100 Epoxide cis-Chlordane 100 4 --- 3 25 trans-ChlordanelOO 6 --- 14 85 Nonachlor 100 8 --- 34 35 degree of chlorination. (Table 1). The relative abundance of 199 the molecular ion increased with increasing chlorine substitution until at the higher degrees of chlorine substitution, the M' ion became the base peak. Concomitantly, there was also observed a decrease in the relative abundance of the Cl‘ ion with increasing halogen substitution. Chlorinated Phenols Chlorinated phenols exhibit similar molar responses in ECD and NI/MS (Figures 7 and 8). In both instances, an‘ increase in response factor was observed until three to four chlorine atoms were substituted on the benzene ring. Molar response factors for NI/MS varied from 0.26 x 107 to 44 x 107 TIC/nanomole for m-chlorophenol to 2,3,4,6-tetrachlorophenol. When fully halogenated the phenol system, as represented by pentachlorophenol, exhibited a sharp drop in response to NI/MS. ECD response/nanomole varied from 70 to 12,000 for p-chlorophenol and 2,3,5,6-tetrach1orophenol, respectively (Fig. 7). On the other hand, as indicated in Figure 8, all of the chlorinated phenols had approximately the same molar response when analyzed by E1. The NI/MS spectra of most of the chlorinated phenols were dominated by the Cl’ ion as the base peak (Table 2). The mono- and dichlorinated species produced an (M-l)’ ion while those having more than two chlorine atoms produce an M' 200 ion. Other fragments were detected which represented the loss of a single chlorine atom from the parent molecule. The relative abundance of the molecular or pseudomolecular ion increased with increased halogen substitution. This was at a maximum with the trichloro-isomers and dropped significantly with the tetra- and penta- isomers. The chloride ion in these spectra became the most abundant ion for the higher chlorinated isomers. 201 .sofiuoouoo um bags 00 >h0a~wnso >2 couscous muososm pouocwuoaso no 0>hso oncommou undo: .h ousoflm 030202 too 2.92 05.620 Co 1% m 4 n a F o P _ _ — 0.0 no 88 T: r O $.88 % d . nu nu 8.88 a / . N . w :o 88 o . w m. i.e.-mo. F a . t0 I¢o+mm. — 202 no maococn mousswuoHno mo 0>bso mmsonmmu undo: 8an. coboom e :0. 250qu o .>uuoaouuooam mama sow m>flummoc oco nouns“ souuooHo an couscoum 0300.02 ton mEob< 05.620 Co u. n .v n N WWII P .m ousmflm c U mo+mo. F Imo+mo. F Imo+moN Imo+uo.n Ioo+mo..v punodwoo 1o alowouoN/oll 203 Table 2 Negative Ion Spectra of Chlorinated Phenols 8.2161134611066119; Isomer [M] ' [M-l] ' [M-HCl] ' [c1] ' p-chloro --- 100 4 6 2,3-dichloro 100 72 65 45 2,4-dichloro 71 67 27 100 3,5-dichloro 18 30 25 100 3,4-dich1oro 42 100 42 89 2,3,4-trichloro 100 --— 10 15 2,3,6-trichloro 95 30 20 100 2,4,5-trich1oro 100 --- 18 50 3,4,5-trichloro 100 --- 5 60 2,3,4,5-tetrachloro 37 14 17 100 2,3,4,6-tetrachloro 4 --- 5 100 2,3,5,6-tetrachloro 8 11 13 100 pentachloro --- 2 6 100 204 Chlorinated Biphenyls The chlorinated biphenyls when analyzed by EI/MS produced similar molar responses. These varied only from 9 x 106 to 3 x 107 TIC/nanomole for 2,4,4'-trichlorobiphenyl and 2,2’4,6,6'-pentachlorobiphenyl, respectively (Figure 9). In comparison, ECD and NI/MS did not demonstrate molar responses of similar magnitude across all degrees of chlorination. The ECD response per nanomole of compound demonstrated an essentially linear increase with increasing chlorination as shown in Figure 10. It varied from 14 to 4500 for 3-chlorobiphenyl to 2,3,4-trichlorbiphenyl, respectively. In comparison those isomers having from one to three chlorine atoms present in the molecule had similar low responses in NI/MS as shown in Figures 11 and 12, however these responses increased significantly when five to eight chlorine atoms were present in the molecule. Overall, the NI/MS nanomolar response varied from a minimum of 1.4 x 106 for 3,4-dichlorobiphenyl to a maximum of 2100 x 106 TIC/nanomole for 2,2’,3,3',4,4’,5,5'-octach1orobipheny1. An increase in molar response for NI/MS was not observed until at least three chlorine atoms were substituted on one ring of the biphenyl ring system. When the PCB isomers were divided into two subgroups of single ring substituted (Figure 11) and 205 OF .ahumaouuommm 0005 :Ow 0>wuomms 6:0 #0093“ souuooa0 an pmospoum m~>s0nmwn o0umcfiuoHso no 0>hso 00somw0u undo: .0 0usofim mEob< 05.6.50 Co 1% m m e N no. 05302 o Lucas: coboom e o Imo+mo. _. Imo+mo.N wmo+mo€ “mo+mo.o .Imo+mo.m Hmo+mo. F H.mo+wN. _. fih.mo+m._w. F .Imo+mo. F wmo+mm. — .Imo+mo.N punodwoo ,lo a|ouJouoN/3LL 206 .co«uomumo Om saw: 00 >unadfinmo >3 couscoum maasmnmwn U0uosfiuoaso mo 0>Mso mmsommmu undo: .OH 0usmwm 0:62 05.620 .6 % punodwog 1o elowouoN/esuodsea (333 207 symmetrically (Figure 12) substituted isomers, this molar response characteristic became more apparent. An increase in response was not observed in the symmetrically substituted species until three chlorine atoms were substituted on one of the phenyl rings (Figure 12). As with the chlorinated phenols, the chlorine-saturated molecule decachlorobiphenyl showed a drop in response from those of the nona- and octa- chlorinated isomers,in both ECD and NI/MS. The nature of the NI/MS spectra of the PCB isomers varied with the degree of chlorination (Table 3). The abundance of the molecular ion was dependent upon the number and position of the chlorine atoms. For example, the relative abundance of the molecular ion of 2,3,4,5,6-pentachlorobiphenyl was greater by an order of magnitude than that for the 2,2',4,6,6'-pentachloro isomer. This indicates that the fully substituted phenyl ring system has a greater inherent stability than those having lesser halogenation. The molecular anion was the base peak for isomers having five or more chlorine atoms per molecule. Once the molecule was substituted with nine or ten chlorine atoms, a drop in molecular ion abundance was observed. When the system was close to halogen saturation, as with the chlorinated phenols, the molecular anion became less stable than the fragment resulting from the loss of a single chlorine atom. 208 .auu0souuomau 0005 so“ 0>wu000s oco pooaafl couuo0a0 «sawusuwuunsm 0swu 0Hmcfi0 mew>sc 0H>G0AQHQ cmumcwuoHcO mo 0>hso mmsommmu undo: .HH 0usmwm 0E3... 056.50 .6 .0. n .v n N o I? .0 . 0 Ie O O .6. 026002 o goods... 6.3.00.0 e 00+wo.— Imo+mo. _. Imo+m0N I00+wo.n punodwoo 4o elowouoN/QLL 209 .Nuu0aouu00nm 000a QOF 0>wu000s 0:0 nouns“ souuo0H0 «mmcFu Hmc0sm suon so sofiusuwumnsm 0swuoHno Hoowpumaahm mcw>0s mamcmnmfin p0uscwuoaso mo 0>uso 0msommmu uoaoz .NH 0usmFm 0:62. 05620 00 .m 0F 0 0 e. N .JQ XXV O 0 “009.5 6.60.0 e .6. 036002 o o I00+mo. F Imo+mo..v Imo+mo.m Imo+wN. F Imo+m0. F Imo+mo.N punodwog 1o elowouoN/Qu 210 Table 3 N6gati26_I6n.6266tra_9f_268_162mers. Belatixe_AbunQanse i§l__I§2mer fH‘l fuzll' 16:211' 1911’ 1 2-chloro --- 100 7 7 3-chloro --- 61 1 --- 2 2,5-dichloro --- 100 8 20 2,4’-dichloro --- 100 17 11 2,4-dichloro 3 48 7 100 3,3’-dichloro --- 100 4 74 3,4-dichloro --- 52 --- 100 3,5-dichloro --- 46 3 100 2,3-dichloro --- 16 1 I100 4,4’-dichloro --- 100 10 70 2,2'-dichloro --- 100 6 3 2,6-dich1oro --- 100 3 55 3 2,2’,5-trichloro 10 100 16 14 2,3',5-trichloro 10 100 28 18 2,3,4-trich1oro 2 8 3 100 2,3,6-trichloro --- 9 4 100 2,4,4’-trichloro --- 10 3 100 2,4,5-trichloro --- 7 3 100 2,4’,5-trichloro --- 100 26 26 211 Table 3 (con’t.) 4 2,3’,4’,5-tetrachloro 32 100 32 20 2,2’,3,5’-tetrachloro 12 100 4 62 5 2,2',4,6,6'-pentachloro 7 25 7 100 2,3,4,5,6-pentachloro 100 9 1 8 6 2,2',4,4',6,6'- 28 100 40 70 hexachloro 7 2,2',3,4,4',5,6- 100 35 1 8 heptachloro 8 2,2',3,3',4,4',5,5'- 100 9 4 --- octachloro 9 2,2',3,3’,4,4',5,5',6- 54 100 6 .5 nonachloro 10 Decachlorobiphenyl 75 --- 100 7 212 Discussion The analysis of toxaphene and the study of individual components of toxaphene is inherently difficult because of its nature as a complex mixture. Only a very few toxaphene components have been isolated and with great difficulty (13-19). Toxicant B, a major component of the mixture, has been previously synthesized (20). However, the purification steps are tedious and time consuming with very low yields. In order to conduct a systematic study of important variables involving degrees of chlorination within the toxaphene mixture, a group of chlorinated pesticides with similar structural characteristics was selected as described in the results section. In this manner, it was possible to gain some insight into the NI/MS behavior of individual toxaphene components. It is of interest to note that changes in selectivity are consistent with a concomitant increase in chlorine substitution around a carbon skeleton. A similar phenomenon has been described for PBBs analyzed using NI/MS (21). With an increase in bromine substitution around the biphenyl system, it was noted that the abundance of the Br‘ dropped: this was accompanied by an increase in the molecular ion abundance. The ratio of Br' to M' decreased logarithmically with the increase in bromine substitution. The same observation was noted when using nitrogen or methane as the 213 moderating gas. The physical basis for these two phenomena are the same. The formation of a stable negative ion is dependent upon three molecular parameters.(22) These are the electron attachment cross section (0a) (e) of a molecule, the number of degrees of freedom (N) inherent in the molecule, and the electron affinity (EA) of the molecule. As halogen substitution on a molecule increases, the EA of the molecule increases as well. At low degrees of halogen substitution on a molecular system, the molecule has a negative EA. The most stable ion produced by capture of an electron is its substituted halogen anion. The molecular ion becomes the most stable ion with higher degrees of halogenation on a molecular system. Thus, as EA increases with increasing halogenation and the relative abundance of the molecular anion within the mass spectrum becomes greater than that of the halogen anion. In addition, the (aa) (a) of a molecular system increases in magnitude with an increase in halogenation. This parameter describes the probability of a favorable interaction between an electron and a molecule. The N of a molecule must be high initially, to better facilitate the distribution of excess energy acquired as a result of the original electron/molecule interaction throughout the vibrational levels of the molecule. Otherwise, autodetachment of the electron is likely to occur. Once the molecule is fully halogenated, even with the 214 increased EA and (0a)(e), N decreases due to steric hindrances caused by neighboring halogen atoms. A drop or leveling off of molar response for the halogen-saturated species would be expected. This was observed in both case of the PCBs and chlorinated phenols (Figures 8 and 9). How might this affect quantitation of compounds when using NI/MS? When quantitating a single compound there will be little effect: only that of the limit of detection for a specific compound being increased or decreased. The major impact of this phenomenon will be observed in the quantitation of complex mixtures which contain species of widely varying degrees of chlorination. Complex mixtures such as toxaphene are altered as they move through the environment (23). In the case of toxaphene, there is evidence which suggests that this alteration is mainly a result of dechlorination (23). The chromatogram of toxaphene residues isolated from an environmental matrix when compared with that of the toxaphene standard used for quantitation, differ greatly. In general, a shift is observed in the relative concentrations of components eluting at longer and shorter retention times: this can be interpreted as a shift in the predominance of compounds having higher to those having lower degrees of chlorination. Thus, when comparing toxaphene from environmental residues with standard material, quantitation using NI/MS may produce artificially low results. This could be due to a shift to lower chlorinated 215 species which give a much lower response under NI/MS analysis. This phenomenon was noted in a recent article in which the authors were quantitating toxaphene or polychlorinated camphenes in the air over Sweden (11). The author stated that concentrations of toxaphene when quantitated by capillary GC with ECD were two to four times greater than quantitation by NI/MS. Though they attributed this observation to co-eluting pesticides in ECD, it probably was due to the lower response produced by the bulk of the toxaphene residue as a result of the "weathering" of the toxaphene mixture. It is likely that the same phenomenon would occur if mixtures of chlorinated phenols and Arochlor residues were quantitated using ECD and NI/MS. In several of the proposed quantitation methods for toxaphene which use NI/MS for analysis, a single compound was used to judge the performance of the mass spectrometer before analyzing the toxaphene residues and standards. This may lead to some problems, especially if the standard used is one having a higher degree of chlorine substitution on the molecule, such as chlordane or dechlorane(9,10). The instrumental response would be skewed favoring detection of higher chlorinated species. As a result, the lower chlorinated species may not be detected. To circumvent the problem, we recommend the following. To better judge the performance of the instrument, a mixture which mimics the chlorine substitution likely to be present 216 in the standard toxaphene mixture should be used. As a test mixture, aldrin, heptachlor, trans-chlordane, and nonachlor is recommended. A total ion chromatogram of the simple mixture and the mass chromatogram of the [M-Cl]' fragment of aldrin are shown in Figure 13. A 1 pl aliquot of the solution containing lOppm of each of the above compounds was used to obtain the TIC profile shown in Figure 13. Even though aldrin appears to be non-detectable at this level, if the [M-ClJ‘ fragment for aldrin is monitored, the compound produces an appreciable response. This represents the response produced by three nanomoles of each compound. Since aldrin is the least sensitive to the NI/MS process, a minimum acceptable total ion current should be determined for aldrin alone. Using this technique, the source pressure, filament current, and electron energy are optimized for this simple mixture. These parameters produced the highest TIC at full scan for aldrin of 900,000 counts as integrated by the SIDAR data system. Once the parameters are optimized, prior to beginning the chromatographic analysis of samples, the simple mixture is injected to determine the condition of the source. In this manner, the detection of components in toxaphene residue having six chlorine atoms or greater is assured. The other three compounds in the test mixture are used to determine approximate response factors for congeners having higher degrees of chlorination in the complex mixture. The advantages of this are several. First, these 217 .ossomsoo 2000 no 00Hoaocws 00ucu >H0ucawxouams >0 amoscoum mmsommmu 0:9 musmmmummu mace .cofiumuwaaumo usmssuumsw m:\Hz you 600: mcoFHSHom ammoa 0:» mo coauomnsfl Ana 0 >8 U0osooua canvac you unmawcuu Facts. 0:» mo Ecumoucaouno 0003 6:0 acnoouosouso muss usmuuso :ofi Fancy a .MH musmfim zoom . . Oh won an? OD? own OOH onfl DON 90F 00F PI . 6 . P 7 . 1E 1.5.x... :élactg £11....) E: can-Q: 39.54.33,...3é2393L. 565396.!) 14“) Eudszg>§§ \ so.goocezleses. OBI‘bO-8OIIBUL. .0 .0 mOmONNnF INOOF .0 .0 Mgsuagul I 218 compounds are readily available to most laboratories which are conducting environmental analyses. In addition, if a temperature program similar to that described above is used, the mixture will elute in less than 20 minutes. Thus, the sensitivity of the instrument to the mixture can be quickly determined by its response to a simple mixture prior to the actual analysis. As a result of using this optimization procedure, levels of toxaphene could be determined accurately. Summary and Conclusions The molar reaponse characteristics produced by electron capture detection, electron impact mass spectrometry, and negative ion mass spectrometry were determined for chlorinated aliphatic pesticides, chlorinated phenols and PCB isomers. Electron impact ionization produced similar molar responses across all three chlorinated species at all degrees of halogenation. ECD and negative ion molar responses were not comparable. Aliphatic chlorinated pesticides were found to exhibit different molar responses by ECD and negative ion mass spectrometry. In ECD, all chlorinated compounds examined produced a dectable signal. Further, those compounds having six, seven, eight, and nine chlorine atoms exhibited molar. responses within the same order of magnitude. This was not 219 the case in NI/MS. The two dichlorocompounds produced no detectable signal, while thouse having six through nine chlorine atoms demonstrated a logirithmic increase with increasing chlorination. The molar response characteristics of chlorinated phenols were similar whether analyzed by ECD or NI/MS. The molar response increased linearly up to the substitution of three to four chlorine atoms on the phenol system. A drop or leveling off of the molar response was observed with compounds having greater chlorine substitution. PCBs exhibitied a linear response as analyzed by ECD. However, when analyzed by NI/MS, the molar response of the system did not increase until at least four chlorine atoms were substituted on one of the phenyl rings of the biphenyl system. Molar response, after this point, increased logrithmically as analyzed by NI/MS. Molar response for the biphenyl system was determined to be influenced the greatest by substitution on a single ring of the biphenyl system rather then the total number of chlorine atoms in the system. The degree of halogen substitution on a molecular system determines the molar response of a molecule. This variation in molar response of chlorinated species needs to be taken into consideration when analyzing complex chlorinated mixtures, such as toxaphene and PCB Alachlors, by NI/MS. A method for determining the molar response characteristics for toxaphene is presented, as a means of recognizing these 220 differences prior to the actual sample analysis. Future Work The application of molar response characteristics to complex mixture analysis should be extended to other halogenated xenobiotic series. These might include compounds such as chlorinated benzenes, PBBs, halonaphthalenes and halowaxes. These should be analyzed to determine molar response characteristics by NI/MS. These compounds have all had methodologies developed which utilized NI/MS for their detection and analysis. Differing molar response characteristics will affect the final quantitation of these 'types of species BIBLIOGRAPHY 1. E. Lewis and W. D. Jamieson, $3 on . 1983 52 303-306. 2. C. Rappe, H. R. Buser, D. L. Stalling, L. M. Smith and R C. Dougherty, £66222 1981 222 524-526. 3. J. R. Hass, M. D. Friesen, D. J.Harvan and C. E. Parker, Aggl6_gn§m1 1978 52 1474-1479. 4. S. Shang-Zhi and A. M. Duffield, J. QDIQQ, 1984 284 157-165. 5. T. M. Trainor and P. Vouros, 53611_thm1 1987 62 601-610. 6. E. A. Stemmler and R. A. Hites, Ang11_§ngm. 1985_61 684-692. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 221 R. C. Daugherty, J. D. Roberts and F. J. Biros, 56611 §n2m1 1975 61 54-59. M. Oehme, 53211_Qh2m1 1983 22 2290-2295. D. L. Swackhamer, M. J. Charles and R. A. Hites, 8261. thm1 1987 22 912-917. U. Wideqvist, B. Jansson, L. Reutergardh and G. Sundstroem,_§h§m61 1984 12 367-379. T. F. Bidleman, U. Wideqvist, B. Jansson and R. Soederlund, A§m1_£ny1ggn1 1987 21, 641-654. B. H. Jennings and G. B. Herschbach, 11_Q;g1_§ngm6 1965 22 3902-3908 . J. N. Seiber, P. F. Landrum, S. C. Madden, K. D. Nugent and W. L. Winterlin, 11_Qh26m1 1975 116 361-368. W. V. Turner, 8. Khalifa and J. E. Casida, Q1_Agr1g1_£666 EDEN. 1975 22 991-994. P. S. Chandrukar, F. Matsumura and T. Ikeda, QDQEQ. 1978 2 123-130. F. Matsumura, R. W. Howard and J. O.Nelson, thmg. 1975 6 271-276. M. L. Anagnostopoulos, H. Parlar and F. Korte, 966mg; 1974 2 65-70. P. F. Landrum, G. A. Pollock, J. N. Seiber, H. Hope, H. and K. L. Swanson, thmg1 1976 21 63-69. S. Khalifa, T. R. Mon., J. L. Engel and J. E. Casida, l. AQI1£._EQQQ_§h§m. 1974 22 653-657- W. V. Turner, J. L. Engel and J. E. Casida, 11_Ag:161 EQQ§_Qh2m1 1977 25 1394-1401. J. Greaves, J. G. Bekesiand J. Roboz, 2166661_Mg§§_§pgg1 1982 2 406-410 . L. G. Christophorou, D. L. McCorkle and A. A. Christodoulides in Ele6tr6n:Hglesule_Interagtigns_gnd vol. 1, ed L. G. Christophorou, p. 478. Academic Press, Orlando (1984). M. A. Saleh and J. E. Casida, l._AQI1§._EQQQ_Qn§m. 1978 26, 583-590 . 222 APPENDICES 223 Table A1 MIWW 16362116119 WW Sample ID i chl cv6hl cv6h2 cvll cle cv2l cv22 cvSl cv52 clel cleZ l .25 .30 .41 .56 .63 .72 .21 .81 1.20 1.93 2.38 2 .16 .ll .21 .21 .ll .13 .34 .30 .62 .81 .73 3 1.99 2.52 2.09 2.36 1.97 2.30 3.04 2.81 3.82 7.50 5.81 4 3.30 3.45 3.64 4.03 2.85 4.16 3.46 4.80 6.72 5.16 7.88 5 4.93 3.24 5.74 5.89 4.66 7.21 6.98 7.20 9.07 8.74 10.9 6 6.62 7.32 7.34 7.38 6.76 8.22 7.87 8.10 7.78 8.26 9.56 7 14.4 16.5 16.1 15.6 15.1 17.7 16.5 16.5 18.6 16.4 18.3 8 8.13 8.40 8.13 8.09 7.88 8.33 8.41 7.86 9.17 5.63 7.97 9 18.5 18.9 16.7 18.2 18.6 17.6 18.0 14.7 12.4 17.7 16.8 10 19.0 19.3 18.7 18.8 19.6 18.0 17.7 16.4 17.9 9 97 14.3 11 10.7 8.90 9.37 9.21 10.3 7.56 7.00 7.15 5.85 2.80 1.95 12 6.15 6.02 5.51 5.00 5.87 4.41 4.33 4.86 5.27 3.54 3.15 13 4.16 4.06 4.36 3.64 4.27 3.01 2.96 3.03 1.22 .49 .23 14 .SO .31 .83 .28 .48 .20 .27 .23 .23 .07 .07 15 .37 .24 .26 .26 .30 .13 .10 .16 0.0 .01 0.0 16 .82 .48 .59 .46 .55 .26 .29 .11 .08 .02 0.0 224 Table 81 Calculatud First Order Rate Constants of 1 ppm toxapheno in _ W W. 15121:“ 1)_.921;.§29_r_$_£p1§m’ 11.__ 1 NS -0.05 3.02 2 NS NS 3 -0.14 t 0.05 -0.05 t 0.02 4 -0.12 t 0.02 -0.06 2 0.02 5 NS NS 6 NS -0.07 t 0.01 7 NS -0.02 t 0.01 8 0.06 t 0.02 -0.04 t 0.01 9 0.04 t 0.01 NS 10 NS NS 11 0.21 t 0.03 0.60 t 0.01 12 0.12 t 0.05 NS 13 0.37 t 0.1 0.07 t .02 14 NS 0.30 1 .04 15 NS NS 16 NS 0.57 z 0.08 225 Table 32 Calculated First Order Rate Constants of 1 ppm toxaphene 1n .121§51111§;!§§23 QJH§EQI 1 Ingggzg h, ngxfi'll QQEQQQI§_£pLQ§¥§-ll—————— 1 NS NS' 2 -o.22 1 .08 NS 3 NS NS 4 -o.13 1 0.05 NS 5 -o.1o 1 0.04 NS 6 -o.oe t 0.02 NS 7 NS -o.os 1 0.02 8 NS 0.04 t 0.01 9 NS 0.03 1 0.01 10 NS NS 11 -o.09 : 0.02 NS 12 NS NS 13 NS NS 14 NS 0.16 1 .01 15 NS Ns 16 NS NS 226 Table 33 calculated First Order Rate Constants or 1 ppm toxaphene 1n 1:2.S:1:1:0.31221.1311:11 anatgr_£___lndggra_£1 19:25?11___Qntdggra_11ldaxa’11...... 1 -0.11 1 .02 -1.3 1 0.1 2 NS NS 3 -0.10 1 0.03 NS 4 -0.03 1 0.02 NS 5 -0.06 1 0.02 NS 6 NS NS 7 -0.01 1 0.004 NS 8 NS 0.09 1 0.03 9 NS 0.03 1 0.01 10 NS NS 11 0.04 1 0.01 0.40 1 0.01 12 NS 0.31 1 0.01 13 0.07 1 0.02 0.03 1 0.01 14 0.19 1 0.052 NS 15 0.12 1 0.05 NS 16 0.12 1 0.04 NS 227 Table C1 Calculated rirst Order Rate Constants of 4ppm toxaphene 1n n111111gd_z1sgr_11sh_s9339.31111_8122r_32n121__ anaggr1i___199_nnm11 10 nnm1 1dppm____ 1 NS NS NS 2 -0.08610.01 -0.104 1 0.021 NS 3 NS NS NS 4 -0.0sa10.010 -0.030 1 0.016 NS 5 -0.02110.003 -0.022 1 0.009 NS 6 NS NS NS 7 NS NS NS 8 NS NS NS 9 0.02010.005 NS Ns 10 0.01610.025 NS NS 11 NS 0.019 1 0.005 NS 12 NS NS NS 13 0.02610.03 NS NS 14 NS NS NS 15 0.1321o.03 NS NS 16 NS NS NS * NS = n91.51gnitlsans_at_g=1951 228 Tabll 02 Calculated First Order Rate Constants ct 4ppm toxaphene 1n .3 f _. ‘1, y , ;.. . ; _,. .u.‘. , ... ,. :11 .. £12512r_1___199_nnn 10 ppm 1 ppm________ 1 NS NS NS 2 -0.08310.02 NS NS 3 -0.04510.02 NS NS 4 NS -0.0257 1 .007 NS 5 NS NS NS 6 -0.01610.007 NS NS 7 NS NS NS 3 -0.02010.009 NS NS 9 0.01610.005 NS NS 10 NS NS NS 11 0.01110.005 NS NS 12 NS NS NS 13 NS NS NS 14 NS NS NS 15 0.10810.05 NS NS 16 0.15310.05 NS NS 229 Table C3 Calculated first Order Rate Constants of 4 pp. toxaphene in NW 919§§£I_1___199_22m 10 gym 1 ppm 1 -0.043:0.009 -0.041:0.003 NS 2 NS NS NS 3 NS NS NS 4 NS NS NS 5 NS NS NS 6 -0.052:0.019 NS NS 7 NS NS NS 8 -0.018:0.005 NS NS 9 NS NS NS 10 NS NS NS 11 NS NS NS 12 NS NS NS 13 0.03310.013 NS NS 14 0.136:0.024 0.079:0.03 NS 15 0.11510.043 NS NS 16 NS NS NS 230 Table D1 Calculated rirst Order Rate Constants (h,)(dsys'1) of 4ppm toxaphene in distilled water at varigns_2n§11 £195121_1____nn=5 pfi=7 0359 1 NS -0.087 1 0.02 -0.072 1 0.020 2 NS -0.196 1 0.061 NS 3 NS NS NS 4 NS -0.045 1 0.078 NS 5 NS Ns -0.03s 1 0.010 6 -0.01 10.002 NS -0.012 1 0.042 7 NS NS NS 8 -0.012 1 0.026 0.066 1 0.002 NS 9 -0.059 1 0.023 0.097 1 0.004 NS 10 NS 0.008 1 0.002 NS 11 0.012 1 0.005 Ns NS 12 0.023 1 0.011 0.073 1 0.022 0.026 1 0.007 13 NS NS 0.013 1 0.043 14 NS 0.057 1 0.014 0.070 1 0.032 15 NS NS 0.112 1 0.05 16 NS NS NS z_u§_:_n2&_fiignliigan§_at_a=.05 231 lel. 02 . calculated First Order Rate Constants (h,)(days'1) or 4ppm Toxaphene in distilled water at with Two Hearted River Eunics VM Wfi pH=7 0_H=9 1 NS -0.118 1 0.040 -0.100 1 0.020 2 -0.074 1 0.030 -0.133 1 0.040 NS 3 -0.029 1 0.012 -0.320 1 0.10 NS 4 -0.021 1 0.009 -0.067 1 0.029 -0.019 1 0.008 5 0.053 1 0.017 -0.053 1 0.017 -0.037 1 0.008 6 NS NS NS 7 NS NS -0.012 1 0.002 8 0.008 1 0.004 -0.014 1 0.005 NS 9 NS NS NS 10 NS NS NS 11 NS 0.011 1 0.004 0.028 1 0.005 12 0.026 1 0.008 0.022 1 0.005 0.020 1 0.008 13 NS 0.023 1 0.006 0.024 1 0.005 14 NS NS NS 15 NS 0.191 1 .061 0.085 1 0.018 16 NS NS NS 010100r