v1.5: .94 This is an authorized facsimile printed by microfilm/xerography on acid-free paper in 1983 by UNIVERSITY MICROFILMS INTERNATIONAL Ann Arbor, Michigan, U.S.A. INFORMATION TO USERS This material was produced from a microfilm copy of the original document. Whiie 1he most advanced technological mean s to photograph and reproduce this document have been used, the quality is heavily dependent upon the quality of the original IUbmittad. The following explanation of techniques is provided to help you understand markings or patterns which may appear on this reproduction. 1. The sign or "target" for pages apparently lacking from the document photographed is "Missing Page(s)". If it w• possible to obtain the missint pege(s) or section, they are spliced into the film along with adjacent pages. This may have necessitated cutting thru an image and duplicating adjacent PIIIIS to insure you complete continuity. 2. Whln • image on the film is obliterated with a large round black mark. it Is an indication 11tat the photographer suspected that the copy mey mond durin9 exposure and thus cause a blurred image. You will find a IOOd image of the page in the adjacent frame. have 3.Whln • map, drawing or chart, etc., _. part of the material being photographed the photographer followed · a definite method in ulBClionint" the material. It is customary to begin photoing at the upper left hand comer of a large sheet and to continua photoin;: from left to right in equal sections with a small overlap. If necessary, sectioning is continued again - beginning below the fi,st row and continuing on until oompfele. 4. The majority of users indicate that the taxtual content is of greatest value, however, a somewhat bigher quality repn)dui::tion could be made fronr ..llhotographs,, If essential to the understanding of the dissertation. Silnr prints of "photographs" mey be ordered at additional charge by writing ft· Otder Qepartment. giving the catalog number. title, author· and specific pages you wish repn)duced. &. PLEASE NOTE: Some pages may haw indistinct print. Filmed a rac:eived. Xerox University Microfflrns aGONcMlhZHIIAMI Mlor, MlcllioM 41108 75-22,939 BAUGHMAN, Robert Wfllfam, 1947- lETRACHLORODIBENZO-p -DIOXINS IN THE ENVIRONMENT · HIGH RESOLUTION MASS SPECTROMETRY AT THE PICOGRAK LEVEL. Harvard Uhfversity, Ph.D., 1975 Chemfstry, organic -- -- �--- Xerox University Mlcrofllm1, �-�- -·-· -- - . - . . . , ....•...::_ ..,,oe - _:_·--.�- Ann Arbor, Mlaht;an THIS DISSERTATION HAS BEEN MICROFILMED EXACTLY AS RECEIVED. TETRACHJ,,ORODIBENZO-P-DIOXINS IN THE ENVlRONMENT HIGH RESOLUTION MASS SPECTROMETRY AT THE PICOGRAM LEVEL A thesis presented by Robert w. Baughman to The Department of Chemistry in partial fulfillment of the requirements for the degree of Doctor.of Philosophy in the subject of Chemistry Harvard University Cambridge, Massachusetts December, 1974 .· � .... 1'etrachlorodibenzo-p-dioxins in the Environment. High Resolution Mass Spectrometry at the Picogram Level. 'l'hesis Chairman; Hatthew_Meselson · Summary Robert w. Baughman December, 1974 2 # 3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD) is an ex­ tr,emely toxic compound _(guinea pig single oral dose to50 0.• 6· \lg/kg) that has been distributed in the environm�nt as ·an impurity in-the herbicide 2,4,S-Trichlorophenoxyacetic -ts=-. ·acid (2,4,S-T) and in other chlorophenol related products. A sensitivity of about. 10-129 per gram of sample (1 part per trillion or 1 ppt) is required for environmental mon­ itoring of 'l'CDD. . ... � ·. A time averaged high resolution mass _spectrometric _ -�hod was developed which can detect TCDD at the picogram · (10-lig ) level. _ Procedures were d�eloped for isolating · picogram quantities of 'l'CDD from. tissue samples of · several g�ams, which provided an analytical sensitiv�t:Y on the order -of one ppt. Samples·of fish, crustaceans, and.human milk from ar�s of South Vietnam heavily treated with 2,4,5-T in the ailitary herbicide program were analy?.ed for 'l'CDO. Levels of upto several hundred ppt in fish and up to 40-50 ppt in human milk were.found in samples collected in 1970 shortly ;.- ;; ��?... : A4#,.. .. , after large scale use of 2,4,5-T was ordered di�continued. In samples colle�ted in 1973, TCDD levels appeared to be lower by an order of magnitude or lllOre. Levels of other more stable 2,4,5-T related compounds were found which appear to have increased between 1970 and 1973. The mass spectrometric method.retains the generality of normal MS analysis. It can be employed in situations · where high -sensitivity is required. The use of high resolution . mass measurement of two or more isotopic isomers provided a high degree of specificity about the molecular formula of an observed ion. Information about the origin of an ion was ob-_ tained by monitoring its appearance time and fragmentation pattern. Quantitation was achieved either through t�e addition of a known amount of TCDD or by isotopic ratio measurements relative to 37c1 labelled TCDD, which was prepared for this. purpose. The accuracy of isotopic ratio measurements was enhanced by the use of time averaged multiple ion-detection. A series of confirmation experiments, including photolysis with monochromatic light, were carried out to test the identity . of observed TCDD signals. · Evidence was collected which suggests that chlorinated biphenylenes or acenaphtalenes, derived from polychlorinated biphenyls (PCBs) may be general environmental residues. The chemistry and toxicology of chlorinated dibenzo­ p-dioxins was reviewed. .ACKNOWLEDGEMENTS The valuable guidance arid encouragement of Professor Matthew Meselson in all aspects of this investigation are gratefully and appreciatively acknowledged. The skillful assistance of Maria Gawryl, Kenneth Gross, Lesley Newton and Hal Bakilla in the laboratory was of great aid to this project, particularly in developing the isolation procedures. Appreciation is extended to Dr. John D. Constable, Professors Bui Thi Lang and Pham Hoang Ho, Dr. Robert Cook, and Professor Arthur Westing for collecting and aiding in the identification of the Vietnamese samples. Special thanks go to Dr. David Firest(tne for many stimulating discussions on the dioxin problem and other subjects, to Drs. Charles Hignite and Gerry Dudek. for help in learning the fundamentals of mass spectrometry, and to Dr. David Parrish for assistance in devising the time averaging system on the MS-9. Financial support for this work was provided by �he National Institute of Envirorunental Health Sciences, the American Association for the Advancement of Science and the Ford Foundation. ·�"f' ¥',; p . ,- ¥.¥.fPK- , . •1 feel sure that there are many problems in Che�stry whfch could be solved with far greater ease by this than by any other method. The method is surprisingly sensitive... (itJ requires' an infinitesimal amount ofmatllrial •••• • .,;SirJ.J. Thomson, describing the mass spec­ troscopic method he had recently discovered, in the Preface to his monograph •Rays of Positive Electricity and Their Applicaion to Ch�ical .Analysis,• Longmans, GreeJt ar.d Co., London, 1913, p. v. ., .... ·· TABLE OF CONTENTS I. PAGE INTRODUCTION A. Review of the propl!lrties of TCDD and.other dibenzo-p-dioxins••••••••••••••••1 1. Chemistry ......•..•_ ..•.••..•.•••••..••. 1 a. Synthesis.........- ......•.......••. 1 b. Reactions and derivatives •••••••••13 c. Isotopically labelled 2. d. e. derivatives•••••••..••••••••• .-•.•• 24 Structure and spectra••••• �•••••••24 Natural occurrence•••••••••••••••• 48 Toxicology.•..•••..••••••••••.•••••••• 51 a. b. c. d. e. Chloracne .•.•.••..•.•••••••••••••• 51 Chick edema ..•••.•••••.••••••..•• 69 Acute toxicity._.....•...••.•.••.•• 76-· Chronic toxicity•••••••••••••••••• 79 Teratogenicity, carcinogenicity and mutagenicity•••••••••83 f. Mechanisms of toxicity••••••••••••89 3. Presence in the environment••••••••••108 a. Possible environmental sources of chlorodioxins•••••••••108 b. Bioaccumulation••••••••••••••••••111 8. Review of analytical methods•••••••••••••112 l. Requirements for analysis••••••••••••112 2, Methods other than maas spectrometry.••••.•.......•...•.....•..••... 113 3. Mass Spectrometry••••••••••••••••••••116 C. Aims of the present investigation••••••••119 II. PRELIMINARY EXPERIMENTS A. a. c. III. Gas-liquid chromatography••••••••••••••••120 High ·resolution mass spectrometry with a photoplate detector •••••••••••••••120 High resolution mass spectrometry with an electron multiplier detector•••••121 DEVELOPMEHT OF AN ANALYTICAL PROCEDURE A. Mass spectrometric detection of TCDD at the picogram level•••••••••••••••••••• · 123 l. Repetitive scanning of a 20 m/e unit range..................•..•.•...•123 2. Repetitive scanning of a 0.300 m/e unit range ••••••••••••••••··�···125 3. Time averagedrepetitive scanning of a 0.300 m/e unit range•••••••••••• 127 TABLE OF CONTENTS (con�inued) PAGE III . DEVELOPMENT OF AN ANALYTICAL PROCEDURE (continued) a. C. Quantitation procedures ••••••••.••••••••••••• 135 1. Direct measurement •••••••••••••••••••••• 135 2. Response relative t� a second compound added as an internal . and_ard .••••.••••••••••••••••••••••••• , 13 8 at 3. Response relative to TCDD added as an internal standard•••.•••••••• 138 4. Isotopic dilution ••••••••••••••••••••••• 140 Isolation of TCDD from biological D. E. Separation from major tissue . ..............••••••••••••••• 142 components .. . 2. Separation from other chlorinated residues.................................................... 143 3. Recoveries .......................................................... 148 Errors and reproducibility•••••••••••••••••• 154 Extraction efficiency••••••••••••••••••••••• sample.a •••••..•••-•.....•••.•.•••••.•••••••••• 142 1. IV. RESULTS AND DISCUS.SION A. TCDD levels in environmental samples •••••••• .- •••••••••••••••••••••· •••••••· 1s 7 B'. · ·Confirmation procedures•••••••••••••••••••••162 l. Mass spectrometry•• ·••••••••••••••••••••• 162 2. Separation of 2,3,7,8- from 1,3,6,8-tetrachlorodibenzop-dioxin by preparative glc ••••••••••••• 166 3. Analysis for hexachlorodibenzop-dioxin............................................................... 166 4. Lack of formation of TCDD from 2,4,5-trichlorophenoxyacetic aqid or 2,4,5-trichlorophenol in isolation procedure•••••••••••••••••• 168 5 •. Neutra'i extraction procedure•••••••••••• 169 6.. Treatment with diazomethane ••••••••••••• 169 7. Photolysis with monochromatic ultraviolet light•••••••••••••••••••••••170 8. Partitioning between acetonitrile and hexane.·••.....•...•.......•.••••••.. 172 9. Lack of adsorption of TCDD· on glass during storage ••••••••••••••••••• 172 c. Other 2, 4, 5-T related con,pounds. •••••••••••• · 173 D. Unidentified residues related to polychlorinated biphenylenes•••••••••••••••• 182 E. The mass spectrometric method ••••••••••••••• 185 TABLE OF CONTENTS (continued) V. PAGE EXPERIMENTAL A.· Mass spectrometric procedure•••••••••• ,••••••• 186 1. Photoplate detection•• , •.•••••• , ••••••••••• 186 2. Repetitive scanning of a 20 m/e unit range•••••••••••••••••••••••••l86 3. Repetitive scanning of a · 1s·7 . ..•• 0.300 m/e unit range................ · •• 4. Multichannel averaging with a pulse-height discriminator trigger •..-..•................••....•.•-•••• 188·. S. Multichannel averaging with an internal trigger••••••••••••••••••••••• 189 6. Improvement in signal to .noise ratio with averaging.•••••••••••••••••••••• 190 7. Optimization of scan rate••••••.••• , ••••••• 190 8. Higher sensitivity with higher .filament current••••••••••••. - �•••••••••••• 191 9. Sample. tube preparation••••••••••••••••.••• 192 10. Sample tube loading and direct · insertion probe proce�ure•••••••••••••••• � 193 ·· 11. Effect of sample matrix on . response •••••••••••••••••••••••••••••••••• 193 12. Effects of sol.vent evaporation, adsorption, water on response••••••••••••• 195 13. Linearity of response•••••••••••••••••••••.197 14. Quantitation procedures a. Addition of authentic TCDD•••••••••••• 198 b. Modification of the Varian TAC-1024 analyzer for multiple ion detection••••••••••••••••••• 201 c. Isotopic dilution•••••••••••••••••• · ••• 201 d. Calibration of TCDD stock solutions••• 203 a. Synthesis and reactions•••••••. · ••••••••••••••• , 204 1. Dibenzo-p-dioxin•••••••••••••••••••••••••• 204· · 2. 2,3,7,8-Te}?achlorodibenzop-dioxin-( CllA p•••••••••••••••••••••••••205 3. Synthesis of 2,),7,8-tetra­ chlorcdibenzo-p-dioxin by pyrolysis of 2,4,5-trichlorophenol •••••••• , •••••••••••••••••••••••••••206. 4. 2,3, S, 6-Tetrachloro-4-b,;omoethyl- -. benzene••• � ••••••••••••••••••••••••••••••• 207 s. The sodium salt of 2,4,5-trichloro­ phenoxracetic acid ••••••••••••••• � ••••••• �208. TABLE OF CONTENTS (continued) PAGE V. EXPERIMENTAL (continued) B. 6. Pyrolysis of 2,4,S-tri­ cblorophenoxyacetic acid and of sodium 2,4,5-tri­ cblorophenoxy·acetate••••••••••••••••• 20 8 7. pyrolysis of a 1:1 mixture of n-butyl 2,4,5-trichloro­ phenoxyacetate and n-butyl 2,4-dichlorophenoxyacetate (herbicide formulation Orange) ••••••••••••••••••••••••••••••210 C. Isolation Procedures•••••••••••••••••••••211 1. Isolation Procedure I. Saponification, treatment with sulfuric acid, alumina chromatography and preparative 2. glc ••••••••••••••••••-••••••••.•••••••-.-211 Isolation Procedure II. saponification, treatment with sulfuric acid and alumina chromatography•••••••••••••••214 ·. 3 � Isolation procedure II I. Neutral extraction with methylene chloride and alumina chromatography•••••••••••••••215 4. Isolation .procedure for polychlorinated biphenylenes•• � •••••• 216 D. Apparatus and Reagents•••••••••••••••••••218 E. Confirmation procedures••••••••••••••••••221 1. Bigh·re11olution mass ·aeasurements•••••••••.•••••••••••••••221· 2. Determination o! possible molecular formulas within + 3 IDlllU .. Of TCDD•••••••••••••••••••••.•222 3. Ysotopic isomer, fragmenta-. tion, doubly charged molecular ion and low ionizing voltage ion intensities•••••••••••••• 223 Volatilization time in the mass spectrometer••••••••••••••••••••223 5. Confirmation for routine ,. ,. ., . analyses •••••••••••••••••••••.•••••••• 224- Photolysis with monochromatic ultraviolet light•••••••••••••• 224 Partitioning between acetonitrile and hexane•••••••••••••••••••225 .L¢. TABLE OF CONTENTS. (continued) V. VI. PAGE EXPERIMENTAL (continued) E. 8. Separation of 2,3,7,8from 1,.3, 6, 8-tetrachloro· dibenzo-p-dioxin by pre. · 226 ••••.••••••• • .••.•••••• · •••.••.•• glc · parative .. 9. Analysis for hexachlorod.ibe_nzo-p-dioxin••.••••••••••• • ••••••• • •. • • • 22r 10. Lack of adsorption of 'lCDD on glass in stored vials•••••••••••••••••••••••••••••••••••••••227 11. v Treatntent with diazomethane••••••••••••••••• 229 P. Recoeries ••••..•..••....... � ..•••.•••.•• - •••...••23O 1. overall isolation procedures ••••••••••••••••230 2. Individual isolation steps •••••••••••••••••• 230 3. Extraction efficiency•••••••••••••••••••••••233 . G. ·sample collection and storage •••••••••••.•••••••• 234 APPENDICES A. An Improved Analysis for Tetrachlorodibenzo­ p-dioxins (Environ. Health Pers�c. (5), 27 (1973)) B. An Analytical Method for Detecting TCDD (Dioxin): Levels of TCDD in Samples from Vietnam (Adv. - Chem. -Ser., 120, 92 (1973)) c. Analysis of 'l'wo Hexachlorophene and 'l'wo 2�4,5-Tri­ chlorophenol Samples for Tetrachlorodibenzo-p­ dioxins D. Analysis for TetrachlorodibP.nzo-p-dioxins in a French Talcum Powder-Hexachlorophene Formulation Implicated in the Death of a Number of Infants E. Analysis of Tetrachlorodibenzo-p-dioxins. in some Selected Samples of the Herbicide Formulation •orange•. F. TCDD Production by Pyrolysis of Sodium 2,4,5-T ,i:-2_*§A.¥.,_�- LIST OF TABLES Page TABLE 1. Synt.1eses of dibenzo-p-dioxin and chlorinated dibenzo-p-dioxins••••••••20,21,22 TABLE 2. Isotopically labelled dibenzo_p-�ioxins ••• • ••••.•.••. ••. � •••••.••••••••• 25 TABLE 3. Infrared absorption frequencies of chlorinated dibenzo-p-dioxins••••••••••••30,31 TABLE 4. Ultraviolet absorption of chlorinated dibenzo-p-dioxins •••••••••••••••••••••••• 33 TABLE s. TABLE 6. TABLE 7. Phosphorescence emission of chlorinated dibenzo-p-dioxins•••••••••••••••••• 34 Ultraviolet absorption of chlorinated dibenzo-p-dioxin cation radicals••••••••• 35 Nuclear magnetic resonance of chlorinated dibenzo-p-dioxins••••••••••••37 a. Electron spin resonance of cation radicals of chlorinated dibenzop-dioxins •••.•••• ••••••••••••.••••••••••••3 9 TABLE 9. Mass spectra of chlorinated dibenzo-p-dioxins••••••••••••••••••••••••41 TABLE TABLE 10. Comparison of detailed mass spectro­ metric fragmentation pattern for · three tetrachlorodibenzo-p-dioxins•••••••44 TABLE 11. Acuta toxicity .of chlorinated dibenzo-p-dioxins••••••••••••••••••••••••77 TABLE 12. Teratogenicity of chlorinated dibenzo-p-dioxins •••••••••••••••.•••••••••86 TABLE 13. Structure-activity relationship of substituted dibenzo-p-dioxins for induction of aryl hydrocarbon hydroxylase activity in the chick embryo liver •••••••••••••••••••••••••••••98 LIST OF TABLES (continued) Page TABLE 14. TABLE 15. TABLE.16. TABLE 17. Effect of size of total residue on MS response•••••••••••••••••••••••••••••• 137 Comparison of the isotopic isomer distribution at the.molecular ion of natural!? occuring TCDD and Cl TCuD•••••••••••••••••••.•• •151 synthetic Recoveries of TCDD from isolation procedures•.••••••·• � • ·••••••••••••••••••.•151 Recoveries of TCDD for individual ate�s_in the isolation procedures•••••••• 153 TABLE 18. Recoveries of TCDD at different added levels••••.. �••••••••••••• �•••••••.• 153 TABLE 19.. 'l'CDD levels in fish, crustaceans and human milk collected in South Vietnam in August and September 1970 ..........................................................................159a TABLE 20. TCDD levels in fish, crustaceans and human milk collected in South Vietnam in May 1973••••••••••••••••••••••160a Calculated molecular formulas within 0.0030 mass units of TCDD••••••••••••••16.4 TABLE 21. ! TABLE 22. Separation of 1,3,6,8-TCDD from 2,3,7,8-TCDD by preparative glc••••••••• ,167 TABLE 23. Analysis for hexachlorodioxin••••••••••••167 TABLE 24. High molecular weight peaks observed in a sample of Vietnamese fish collected at site A in 1973•••••••••177 TABLE 25. Approximate levels of the m/e 446 compound in samples of Vietnamese fish•••179 TABLE 26. Possible TCDD "late peak• precursors derived from diphenyl ether not observed in 1973 Vietnamese fish•••••••••••181 TABLE 27. Ratios for various samples of the signals observed at the exact masses (+ 3-4 mmu) of the isotopic isomers oY c12H4c14 ........................................................... 1s1 LIST OF FIGURES Page FIGURE 1 X-Ray crystallographic structure of TCD0•••••••• 28 FIGURE 2 Calculated ff-electron densities of dibenzo-••• p-dioxin FIGURE 3 Calculated ff-electron densities of TCoo••••••••• 47 FIGURE 4 Linearity of 'l'CDD dose�response••••••••••••••••• 81 46 FIGURE 5 Possible mechanisms of toxicity for TCDD•.•••••• 107 FIGURE 6 Glc analysis of chlorodioxinS••••••••••••••••••ll6 Mass spectrum of TCDD and 37c1 'l'CDD••••••••••••l24 FIGURE 7 FIGURE 8 Twenty m/e unit MS scan of 'l'CDD••••••••••••••••i26 FIGURE 9 0.300 m/e unit FIGURE 10 E FIGUR 11 FIGURE 12 FIGURE 13 FIGURE 14 FIGURE 15 FIGURE 16 FIGURE 17 FIGURE 18 MS scan.of TCDD ........... � ••••.128 Improvement in S/N with time averaging•••••••••l29 CAT- MS-9 interfacing••••••••••••••••••••••••••l31 Sample introduction in MS-9••••••••••••••••••••133 Sample tube preparation••••••••••••••••••••••••l34 Signal for 2 pg of TCDD••••••••••••••••••••••••l36 Linearity of response•••••••••••••••••••� ••••••139 Effect of alumina chromatography on residue••••l44 Effect of sample residue on response••�••••••••l45 Resolution of 'l'CDD from DDT, ODE, and PCB••••••l49 FIGURE 19 Effect of two alumina columns on residue••••••• lSO FIGURE 20 Limit of detection and signal from 5 ppt of•••• 155 'l'CDD in human milk FIGURE 21 Samp!e collection sites in South Vietnam ••••••• 158 FIGURE 22 'l'CDD signal in Vietnamese fish•••••••••••••••••l59 FIGURE 24 Photolysis apparatus ••••••••••••••••••••••••••• 17la Apparatus for syntheses of 37c1 Tcoo ••••••••••• 205a FIGURE 23 Appearance time of 'l'CDO in the MS-9 •••••••••••• 174 FIGURE 25 ABBREVIATIONS DDE: 2, 2-Bis (p··chlorophenyl} - 1 , 1:- dichloroethylene 2, 2-Bis (p-chlorophenyl)- 1.,l, 1-trichloroethane ec: Electron capture detector for glc fid: Flame ionization detector for glc glc: Gas-liquid chromatography HMW: High molecular weight m/e Mass to charge ratio of an ion mmu: Millimass unit or 10-J atomic mass unit MS : Mass spectrometry ng or nanogram: · 10-99 pg or picogram: 10 -12g. PCB: Polychlorinated biphenyl PCBene: Polychl«:>rinated biphenylene PFA: Perfluorotributylamine MS m/e reference ppm: Part per million (one part in 10 6 ) ppb: .Part per billion (one part in 10 9 ) ppt: Part per trillion (one part in 1012 ) 2,3,5-T: 2,4,5-'l'richlorophenoxyacetic acid TBB: 2,3,5,6-'l'etrachloro-4-bromoethylbenzene TCDD: 2,3,7,8-Tetrachlorodibenzo-p-dioxin, unless another substitution pattern is indicated 37cl 'l'CDD: 'l'CDD labelled with 37cl 14 c 'l'CDD: TCDD labelled with 14 c . DDT: . I N T R O D U C T I O N A. Review of the Properties of .TCDD and Other Dibenzo-p-Dioxins 1. Chemistry a. Synthesis The first synthesis of a chlori�ated dibenzo-p-dioxi_n was carried out by Merz and Weith just over a century ago.1 From the pyrolysis of either pentachlorophenol or its po­ tassium salt (601 yiel� from the salt) they obtained a white c6 c1 For o. 4 this compound they observed the properties which now have crystalline product with the empirical formula come to be recognized as characteristic of the higher chlor­ inated dioxins: low solubility, thermal stability and chem­ ical inertness. Only by refluxing with concentrated nitric acid were they able to slowly decompose their product. On the basis of its low vapor pressure, Merz and Weith, concluded that the compound must have a molecular formula at least twice the �.mpirical formula. They proposed the structure: They called their compound •pt:rchlorophenylene oxide• and suggested that the condensAtion of chlorinated or brominated phenols to such phenylene oxides might be a general reaction. 1v. Merz and w. Weith, Chem. Ber.,�. 458(1872). 2 Their suggestion turned out to be correct and in fact this reaction is now an important route for the synthesis of chlorinated dibenzo-p-dioxins. In 1889 Huguon enq 2 re ported obtaining a compound iden­ tical to that of Me rz and We ith from the pyrolysis of "hex­ achlorophenol" (he xachloro-1,4-cyclohexadiene.-J-c;,rie) and pro­ r posed the structures: c6c14 ........_ . / 0 . c6c1 ,, 0 L. /· o c6c14 Hugounenz•s reactio� was confirmed hy Zincke andSchaum,3 Barral added the following to the list Barr�1,4· and Biltz.5 of structural candidate a: 0 Cl Cl Cl Cl 0 0 0 2L. Bugounenq, !!!!!• �- �- (3 eserie)!, 805(1889). 3T. Zincke and c. Schaum, Chem.�-!!, 550(1894). · . 4 E. Barral, !!!!!· �- �- (3es e rie)l3, 423(1895). 5 (a) H. ailtz, Chem. Ber., 37, 4003(1904)1 (b) H. Biltz, ibid., 37, 4i>I0(l904). 3 Hugounenq 6 obtained the same compound by heating pentachloro­ anisole with fuming sulfuric acid (the reaction mixture was intensely blue viol e t) and Barral and Jambon 7a reported that it was formed in 501 yield when pentachlorophenol was heated Jambon,7b at 165-170 ° in the presence of c1 and SbC1 • 2 3 in an e xtensive study of salts of pentachlorophenol obtained The structure results similar to those of Merz and Weith. of the "perchlorophenyl e ne oxide" compound was confirmed only recently when Tomita 8 and Denivelle 9 pr e pared an identical ccmipound by perchlorinating dibenzo-p-dioxin(!,) to octa­ chlorodibenzo-p�dioxin{l). ·.o:, 9 ,� .. 10 1 0� 2 oV3 5 ! Cl2 FeCl3 ) or SbCl5/::;. C�O�Cl Cl�O�Cl Cl Cl ! 6i.. Hugounenq, Ann. Chim. Phys. (6 e aerie), 20, 504(1890). 7 ca) E. Barral and L. Jambon, Bull. soc. Chim. France, !!, 822 (1900) 1 (b) L. Jambon, Bull. Soc. Chim. France (l e serie ), 825(1900). ll 8M. Tomita, s. Ueda, and M. Narisada, Yakugaku Zasshi, 1!., 186(1959). 9t. Denivelle� R. Fort., and P. Van Hai, Bull. Soc. Chim. France, 1538 (1960). The first report of a compound actually identified as having a dibenzo-p_-dioxin nucleus was made in 1899 by Hillyer, lO who condensed catechol with picryl chloride to give o: 1,3-dinitrodibenzo-p-dioxin in 851 yield: . . . OH + OH This compound was reduced to give the corresponding diamine.11 At. about the _same time the first report of a compound specifically identified as a halogenated dibenzo-p-dioxin was made by Jackson and Koch at Harvard Unive�sity.12 By condensing lOH. w. llu. w. Hillyer, Arner. Chem. J., ll, 125(1899). Hillyer, Amer. Chem. J., 26, 361 (1901). 12c. L. Jackson and w. Koch, Amer. Chem. J., ll• 10(1901). 5 tetrabromo-o-quinone with tetrabromocatechol they obtained a brominated quinone derivative which could be reduced to a dihydroxy compound: B*; __ ·oa . Br*�.Br. . . o B Br r � Br . ' Na/Hg . OH . + � B . o AcOH, H20 A . o-)yoa · I . Br¥oVoH ·arh ,:,, Br Br Jackson and Mactaurin .later prepared the corresponding chlor­ inated compound.lla This work was repeated by Frejka, et !!_.llb The structure of.Jackson and Koch•s compound was confirmed by Borner and Lingnan 14 who reacted both the hexabromo and hex­ achloroquinones with diazomethane: 13(a) c. L. Jackson and R. D. MacLaurin, Amer. Chem. J., . 37, 7(1907)1 (b) J. Frejka, B. Sefranek and J. Zika, Collect. CZech. Chem. Commun�,!, 238(1937). 14t. Horner and E. Lingnan, Ann. Chem., 573, 30(1951). 6 X X X o o h, � x xVo ·* � o . X o� X* X · I � 0 � I · 0\ CH2 o/ X • Cl, Br Synthesis of unsubstituted dibenzo-p-dioxin (!) was first reported by Ullman and Stein15 in 1906. They cyclized 2 ,2 '-dihydroxy and 2 ,2•-·dimethoxydiphenyl ether with hydro­ bromic acid and red phosphorus: or � o� �Me MeaN JiBr, red P � 185 ° ! In 1909 and 1910 patents were obtained by F. Bayer and Co.16•17•18 for a direct synthesis of dibenzo-p-dioxin by condensation of readily available alkali metal salts of o-chlorophenol: 15F. Ullman and A. Stein, Chem. Ber., 39, 622(1906). 16F. Baye and Co., German patent 223,367 (1909); Chem. Abst •., r !, 2981 (1910). 17F. Bayer and Co., British patent 4,540 (1910); Chem. Abst., f, 2907 {1911). 18R. zaertling a�d H. Friedrich (to F. Bayer and Co.,), U. S. patent 981,348(1910); Chem •. Abstr., _!, 1191(1911). 7 � OH 2 __ 22_0_0__ �cl· M ! M • Na or K OVer the period since these early investigations the chemistry of the dibenzo-p-dioxins has been explored exten­ sively, especially by Tomita, in a series of over forty papers, and Gilman,and their colleagues.19 The basic condensation reactions outlined above remain the most satisfactory means of preparing the dibenzo-p-dioxin nucleus. Symmetrical de- rivatives can be prepared in fair yields by condensation of appropriately substituted Q-chloro or o-bromophenols in the 19ae1evant papers by these authors are listed in reference 37. 8 presence of base and a copper catalyst.8•9• 28 - 28 an example of the Ullman synthesis of aryl ethers This is 29 (not to be confused with the more well known Uliman biaryl synthesis). · er . ,· 2 � . OR Base, Cu X /l, Symmetrical derivatives of! . X • Cl, Br Synthetic Route A 28such condensations are descri.bed by Tomita and his coworkers in papers II, III, IV, XVI, XIX-XXVII, and XXX of their series on Dibenzo-p-dioxin Derivatives. 21N. M. Cyllinane and c. G. Davies., Rec. Trav. Chim., 55, 881(1936). 22a. Gilman and c. G. Stuckwisch, J. Amer. Chem. Soc., 65. 1461 (1943). 23 (1953). M. Julia and M. Baillarge, Bull. Soc. Chim. Fr., 644 24 w. Sandermann, H. Stockmann and R. Casten, Chem. Ber., ,!!, 69 0(1957). 25(a) G. R. Higginbotham, A. Huang, D. Firestone, J. Verrett, J. Ress and A. D. Campbell, Nature, 220, 70 2 (196.8); (b) N. P. Buu-Hoi, G. Saint-Ruf, P. Bigot, M. Mangane, C. R. Acad. Sci., Paris, ser. D, 273, 708(1 9 71). 26A. E. Pohland and G. C. Yang, J. Agr. Food Chem., !Q., 1093(1972). 270. Aniline, Adv. Chem. Ser., 120, 1 26(1973). 28J. Kimmig and K. H. Schulz, Naturwissenschaften, 44, 337(19 57). 29 F. Ullman and P. Sponagel, Chem. Ber., 38, 2211(19 05). 9 Unsymmetrical derivatives can be prepared by the condensa­ · O :o tion of substituted catechols with substituted o-dichloro or o-dibromobenzenes: 26 •30•31•32 O H OH + X � I --->• Symmetrical or unsym­ metrical derivatives o f! X • Cl, Br Synthetic Route B Similar unsymmetrical �ondensations can be carried out with substituted catechols and o-halogenated nitrobenzenes: 10 • 26 •33 30M. Tomita, J. Pharm. Soc. Japan, 52 , 429(1932). 31N. M. Cullinane, H. G. Davey and H. J. H. Padfield, J. Chem. Soc., 716 (1934). 32(a) A. s. Kende and J. J. Wade, Environ. Health Per.­ spectives (5), 49-57(1973); (b) A. s. Kende, J. J. Wade, D. Ridge, A. Poland,!:_ Org. ·chem., 1_!, 931( 1974). 33(a) J. D. Loudon and F. McCapra, J. Chem. Soc., 1 899, (1959}J (b) M. Tovita, J. Pharm. Soc. Japan, 55, 1060(l935). 10 o: N02X:»9'. OH + OH X � I ----�> Symmetrical or unsym­ metrical derivatives of! X • Cl, Br Synthetic Route C At times it may be advantageous to start with one ether link­ �e already formed in which case the second bridge may be closed by an Ullman condcnsation 26 or.by elimination of water lS, 3·4 from a 2,2'-dihydroxy precursor: Synthetic Route D 0: 0 . :c H X .� I Base, Cu Symmetrical or unsymmetrical derivatives of! Synthetic Route E 34ca) I. Keimatsu and E. Yamguchi, J. Pharm. Soc. Japan, 1!,, 680(1936) 1 (b) M. Tomita, J. Pharm. Soc.· Japan,�, 814(1936). 11 These condensation reactions have been used to prepare a number of specific chlorinated isomers of dibenzo-p­ dioxin, 8• 9 • 23 • 25• 26 •27 • 2 8• 32• 3,5 including the highly toxic 2,3,7,8-tetrachlorodibenzo-p-dioxin (l,). 25, 26 , 2 7, 28,3 2 ,35 Cl l O � � � Cl�O�Cl ! The reactions of Hugounenq 2 and Barra1 7a have been developed by Kulka36 into high yield syntheses of octachlorodibenzo-p­ dioxin: Synthetic Route F :::Q::: :)Q::: A Cl Cl Cl Synthetic Route G A ) ! Cl 35(a) E.L. Jones and e. Krizek, J. Invest. Dermatol., 12., 511 (196 2). (b) M.H. Milnes, Nature, 232, 395 (1971). 36 M. . Kulka, Can. J. Chem., l2_, 1973 (1961). 12 In agreement with previous observations concerning the stability of octachlorodibenzo-p-dioxin, Kulka noted that this_ compound could be recrystalized unchanged from hot, con­ centrated sulfuric acid. >_-::':.·:_ 13 b. Reactions and Derivatives Dibenzo-p-dioxin itself readily undergoes electrophilic These reactions have been investigated in detail by Dietrich,37 Gilman andDietrich,3s-4o and Tomita.41 The 2,3,7, and 8 positions are substituted aromatic substitution reactions. first, and for deactivating groups the reaction can be ex­ tended to the 1,4,6, and 9 positions only under forcing conditions. Presumably this is because the 2,3,7, and 8 posi- tions (para to the oxygen atoms) are activated by resonance effects while the 1,4,6, and 8 positions(� to the oxygen atoms}, although also activated by resonance effects, are deactivated by inductive effects.37 (See.also the discussion below of molecular orbital calculations.) synthesis of The first reported 2 ,3,7, 8 -tetrachlorodibenzo-p-dioxin (l) �as achieved by the chlori�ation of 2,7-dichlorodibenzo-p-dioxin (�) It has since in the presence of ferric chloride and iodine.24 been established that the chlorination of dibenzo-p-dioxin precedes stepwise through �he 2-, 2,7-, 2,3,7-, 2,3,7, 8 -, and, eventually, under forcing conditions, the 1,2,3,4,6,7, 8 ,926 2 derivatives: 8 • 5• •37 1957. 37J. J. Dietrich, Ph.D. Thesis, Iowa State College, Ames,Iowa, 38H. Gilman and J. J. Dietrich, J. Amer. Chem. Soc., 1!, 1439 (1957). 39H. Gilman and J. J. Dietrich, J. Org. Chem., ll, 1403(1957). 40u. Gilman and J. J. Dietrich, J. Amer. Chem. Soc., !Q., 366 (1958). 41This work is described by Tomita and his co-workers in papers V, VI,· XI, VII, XXVII-XXIX of their series on Dibenzo-p­ dioxin Derivatives (see Bibliography). 14 ! 1 ! a:,athetic lloute • on the other hand, metalation of dibenzo-p-dioxin with n-butyllithium, 2 2 or phenyllithium,37 a nucleo­ methyllithium, philic reaction, results in substitution in the l and6 (or possibly 9) positions: ! 1) +Li 2) 12 > I Synthetic Route I Cleavage of the ether bridges in dibenzo-p-dioxin is difficult but has been achieved with lithium in ether or in tetrahydrofuran42 • 4 3 or dioxane.44 In tetrahydrofuran: · 4 2 H. Gilman and J. J. Dietrich, J. Org. Chem., 22, 851 (1957). 43H. Gilman and J. J. Dietrich, J. Amer. Chem. Soc., ,!!!, 380 (1958). 44 Y. Watanabe, J. Phar111. Soc. Japan, 75, 313(1955}. 15 ! ! Li, THF, 25 ° 2) CO2 1) 1) Li, .THF, d 2) CO2 0 0::- .� . OH B02C or· .�· :C � OH A similar reaction along with formation of a biphenyl has been observed with sodium in ammonia: 45-so 4Sy . Inubushi, K. Nomura, E. Nishimura, and M. Yamamoto, Yaku9:aku Zasshi, 78, 1189 (1958). 46 Y. Inubushi and K. Nomura, 1!• 838 (1959). nY Inubushi and IC. Nomura, 81, 7(1961). . 48y _ Inubushi and K. Nomura, 49y_ Inubushi and IC. Nomura, �-· �-· �-· 82, 696(1962). .!!?M•. !!, 1341 (1962). soy Sasaki and M. Suzuki, Chem. i>harin. Bull., 18, . 1(1970). 16 �co .OMe 0 OMe Tani 51 Na, NH� q. ·� 0 HO 0 + � OMe OMe With the use of a platinum oxide catalyst Tomita and •52 were able to reduce dibenzo-p-dioxin to dodecahy� drodibenzo-p-dioxin and ci.s-1,2-cyclohexanediol: ! + o: Reversible oxidation of dibenzo�p-dioxins to a radical cation occurs, and in fact this reaction has formed the basis for a diagnostic color test for the dibenzo-p-dioxin nucleus.53154 s111 _ Tomita and c. Tani, J. Chem. .soc. Ja2an, 64, 972 (1943). 52M. Tomita and c. Tani, J. Pharm. Soc. Jaean, 64 242(1944). , 53M Tomita; J. Pharm. Soc. Ja12an, . g, 889 (1932). 54M. Tomita ands. Ueda, Tetrahedron Lett., 1189(1963). ---- -- 17 _ _ _ _ ____________,_ _, In the presence of strong acids such as sulfuric acid, an­ timony pentafluori.de, fuming nitric acid, trichloroacetic acid 54 or trifluoromethanesulfonic acid55 and if necessary an oxidizing agent such as KN03, all known dibenzo-p-dioxins form a deep blue, greenish blue or occasionally violet solu· tion. It is interesting to recall Huguonenq's observation in 1890 t�at the reactie>n mixture in forming octachlorodibenzo­ p-dioxin from pentachloroanisole in fuming sulfuric acid " !,! colore ��violet intense.• 6 �rom electron spin resonance spectroscopy the species formed has been identified as __ a cation radicai. 54•56 · In many cases the parent dibenzo-p-dioxin can be recovered with the addition of water. 5 4 As Hugounenq says, • . . . !'addition d'eau decolore instan_tanement la ligueuer, et il � des flocons cristallises • • • du per­ chlorodioxydiphenylene.• 6 Some dibenzo-p-dioxins also form charge transfer complexes. Such complexes have been observed spectroscopically with iodine, tetrachloro-o-quinone5 7a and 1,2,4,S-tetracyano­ benzene57b and actually have been isolated with tetracyanoethylene,58 55G. c. Yang and A. E. Pohland, J. Phys. Chem., 76, 1504(1972). 56'1'. N. Tozer and L. D. Tuck, J. Chem. Phys., 38, 3035 (1963). 57ca) A. Kuboyama, J. Amer. Chem. Soc., 86, 164(1964). (b) J. E. Bloor, B. R. Gilso�, R. J. Haas, and c. L. Zirkle, J. Med. Chem., 13, 922(1970). 58 s. Teshima, T. Matusuo, and K. Ueno, Asahi Garasu Kogyo Gijutsu Shoreikai ·Kenkyu Hokoku, 13, 339 (1967). ar-nw·v ··; 18 trimesoyl chloride (1:1),59 3,5-dinitrobenzoyl chloride (1:1),59 and 1,3,5-trinitrobenzene (2:1).60 Thermally dibenzo-p-dioxins, in particular the higher chlorinated derivatives, are extremely stable. As Merz and Weith noted for their "perchlorophenylenoxide" (octachloro­ dibenzo-p-dioxin), ·� schmilzt gegen �. sublimirt im langen Nadeln und geh� � uber dem Siedepunkte des Oueck. silbers unverandert iiber."1 Rapid thermal decomposition of 2,3,7,8-tetrachlorodibenzo-p-dioxin occurs only at temperatures above 750 ° .61 Photochemically, in organic solvents, dibenzo-p-dioxin initially undergoes cleavage at one ether linkage to give o­ phenoxyphenol and eventually yields polymeric products with no uv absorption implying complete loss of the aromatic rings.62a. The absoroption maximum near 300 nm is probably important in o this photodecompsition. Chlorinated dibenzo-p-dioxins react similarly, although the rate of decomposition decreases as .the 59G. A. Varvoglis and N. A. I 193.5-195 295 320-325 NR NR 306 305-306 NR NR 305-307 306-307 295-300 300-302 302-304 298-300 195-196 275 230 328-240 Synthetic Route B ff H ·a A A ff A A B+ff B A ff A A C+ff C+ff ** A Yield(I) Ref 40_ 29 44 NR NR 60 NR 39 NR 41 19 15 32 15 1 94 87.6 10.9 32 8• 24 25a 25a 25b 26 27 28 32 32 35 37 C d 26 26 68 26 NR A 1-3 27 NR A NR 2S NR A NR 2S 22 TABLE 1. (cont.) Derivative mp (•c, Octachloro- 320 323 NR NR 325-326 If If " If Synthetic Route )JOO' If NR NR ·. 318-319 326 326 326 326 . 328-331 330 29.8¼0.5 330 325-328 NR If If If If If If If If If If If If If A F F F F *** G A A A F D A+H A G G F G A Yield(I) 60 60 NR NR NR NR 50 NR 35 NR Ref 1 2 3 4 5 6 7a 7b 8 9 9 NR 7_8 9 42 86 24 NR 62 86 83 NR 9 26 27 36 36 97 *Prepared by diazotization of 2,7-dinitrodibenzo-p�dioxin fol:owed by reaction with cuc12• . **Prepared by reaction of pentachloroanisole with fuming sulftlric acid. ***Prepared by reactionof pentachloroanisole with fuming sulfuric acid. •H. J. Vinopal, I. Yamamoto, and J. E. Casida, Adv. Chem. Ser., ill, 7 (1973). b s. Ueo, Bull. Chem. Soc. Japan, 16, 177(1941). cD. A. Elvidge, Analyst, ,!!, 721 (1971). dM. H. Milnes, Nature, fil, 395(1971). NR: Not reported. 23 moist and incubated at 28-30 ° . At the end of one year 50- 701 of .the TCDD remained with little correlation as to soil type or level added. The experiment was repeated with 14 c labelled 2,3,7,8-tetra- and 2,7-dichlorodibenzo-p-dioxin. For the tetrachlorinated compound no labelled metabolites . were observed. A little radioactivity was trapped from the atmosphere which may have been co2 or volatilized TCDD. For the dichlorinated compound at.least one metabolite was observed and a considerably larger amount of radioactivity was trapped from the atmosphere. · A more exhaustive investigation of microbial degrada­ 2,3,7,8tion was carried out by Matsumura ana Benezet.63b . 'tetrochlorodibenzo-p-dioxin was added to cultures of 100 micro­ bial strains selected for their ability to degrade persistent pesticides·� Five strai�s showed a slight ability to degrade the dioxin, and t.he rest were ineffective. Those strains which did achieve some degradation exhibited unusual behavior in that it was not possib1e to increase the rate of degrada­ tion by manipulating the. cultural cor,ditions. These results suggest. that biodegradation of 2,3,7,8tetrachlorodibenzo-p-dioxin by microorganisms probably is not significant. In higher animals this compound also appears to reinain virtually unmetabolized (see Toxicology). 63(b) F. Matsumura a� H. J. Benezet, Environ. Health Perspec. (5) 253(1973). 24 c. Isotopically Labelled Derivatives Several isotopically labelled chlorinated dibenzo-p­ dioxins have been synthesized for use in either toxicological or analytical investigations. These derivatives and their method of synthesis are listed in Table 2. d. Structure and Spectra As mentioned earlier, the "perchlorophenylenoxide" prepared by Merz and Weith in 1872 was not shown to be octaSimilarly, the structure chlorodibenzo-p�ioxin until 1959. of Ullman and Stein's "diphenylene dioxide" was not proven to be a dibenzo-p-dioxin until nearly forty years after· its initial synthesis when Tomita and Tani hydrogenated it to dodecahydro dibenzo-p-dioxin and ois-1,2-cyclohexanediol. Even with the establishment of-the basic dibenzo-p-dioxin skeleton, a question bas remained as mwhether the molecule is planar or bent about 64 the C-0-C bonds. In 1934 Bennet,� observed that the &. dipole moment of dibenzo-p-dioxin in carbon tetrachloride, benzene, and cyclohexane was either very small or zero. this they.concluded that the molecule vith c-o-c bond angles of 120 ° . From was effectively planar It seems that they were correct although other investigators have obtained different results. In 1941 Higasi and Uyeo 65 reported a dipole moment of about 64 G. M. Bennet, D. P. Earp, 1179(1934). s. Gl.asstone, J. Chem. Soc., 65x. Higasi ands. Uyeo, J. Chem. Soc. Japan, g, 396(1941). TABLE 2. Isotopically labelled dibenzo-p-dioxins. Isotope used. Derivative Unsubstituted 3H 14c 2,7-Dichloro2,3,7,8-Tetrachloro- 14c 14c " 3H 36C l " " • Octachloro 37Cl 36C l Specific activity . Percentage labelled isotope of total isoto�e (mliLmmol) 0.085 138 0.91 0.93 1 48 107 NR NR Synthetic route * 0.17 A 0.17 A+H 28 B . 0.13 H NR ** 95.5 H ** NR *Hydrogenalysis of 1,6-dichlorodibenzo-p-dioxin. **Neutron activation of 35c1. aH. J. Vinopal, I. Yamamoto and·J. B. Casida, Adv. Chem. Ser., llQ_, 7(1973). bw. w. Muelder and L. Shadoff, Adv. Chem. Ser., 120, 1(1973) • . csee Experimental Section of this thesis. NR: Not reported. JLGAJtltW@!AA@lk#JM.&Wl!MAiffSMJ.Bl'M.MAlSL®/4#¥1.#J;4@N¥£.\J#it44?�Mt¥&@&.zs. ::et.I£ a ZtNALC Ref a b b 32b a 142c C 142c 26 0.60 Din benzene and cyclohexane and concluded that the aromatic rings of dibenzo-p-dioxin deviated from coplanarity by about 20 ° with a c-o-c bond angle of 117 ° . The same year in an X-ray crystallographic study Wood and Williams 66 con­ cluded thatthe molecule was probably planar, although they could not rule out.other possibilities. In 1969, using highly sensitive microwave absorption spectroscopy, Aroney, � ai.67 reported a zero dipole moment for dibenzo-p-dioxin in cyclohexane. This would appear to confirm that at least on the average the molecule in solution is planar. As a result of the effort to unambiguously identify the toxic •chick edema factor" (see below), Cantrell, Webb, and Mabis 68 determined the structure of 1,2,3,7,8,9-hexachlorodibenzo-p-dioxin by X-ray crystallography. They found that the molecule was very nearly planar, although it was bent about the 0-0 axis at an angle of 4.3 ° . Boer and North; 69 66 · R. G. WOod and G. Will_iams, Phil. .Mag., 31, 115(1941). . 67M. J. Aroney, G. M. aoskins, R. J. w. LeFevre, R. K. Pierens, and D. V. Radford, Anst. J. Chem., 22, 2243(1969). 68J. s. Cantrell, N. c. Webb, A. J. Mabis, Acta Crystal�-, B25, 150 (19.6 9 ). 69F. P. Boer �nd P. P. North, Acta Crystallogr., B28, 1613(1972). i' 27 Boer, Neuman, and Aniline; 70 Boer, et ai., 71 and Neuman, North and Boer, 72respectively investigated the structures of 2,7-dichlorodibenzo-p-dioxin, 2,8-dichlorodibenzo-p-dioxin, 2,3,7,8-tetrachlorodibenzo-p-dioxin, and octachlorodibenzop-dioxin (Figure 1). They found that the 2,7-dichloro, 2,3,7,8-tetrachloro, and octachloro derivatives were pianar, while the by 4.8 . ° 2 ,8-dichloro derivative was bent about the o-o axis They also noted, that the electron density ellipsoids of the bridging oxygen atoms were elongated perpendicular to the plane of the rings, indicating a large component of thermal motion approximately perpendicular to the molecular plane.· This is interpreted as suggesting that, although the molecules are planar on the time average, they are relatively flexible In contrast with respect to bending about the o-o axis.72 the corresponding dithio compound, thianthrene {dibenzo-p- dithiin), is sharply bent about the s-s axis. From X-ray crystallographic studies, the angle between the two aromatic rings in this compound is only 128 ° (a 52 ° deviation from planarity.) 73•74 This probably is because the 3p orbitals 7°F. P. Boer and P. P. North, Acta Crystallogr., 1613(1972). !E!• 71F. P. Boer, F. P. van Remoorter.e, P. P. North, and M. A. Neuman, Acta Crystallogr., fil, 1023(1972). 72M. A. Neuman, P. P. North and F. P. Boer, Acta Crystallogr., 828, 2 313(1972). 73u. Lynton and E. G. Cox, J. Chem. Soc., 4886(1956). 741. Rowe and B. Post, Acta Crystallogr., ll, 372(1958). 28 B H 1.01 Cl 120 120 120.3 122.2 121.2 119.9 122 119 Cl 0 1.388 1.386 Cl 0 0.97 B . B II H 1.01 118 122 Cl 1.383 1.386 119 Cl 121 1.03 B H Figure 1. X-ray crystallographically determined structure of TCDD. Bond angles and distances are given. The TCDD crystal contains two independent molecules situated on the inversion centers at (0,0,0) and(�,�.�) respectively 29 of sulfur are much less able to hybridize and interact with the aromatic.rings than are the 2p orbitals of oxygen. Additional information about the structure of dibenzo­ p-dioxin has been obtained from infrared spectroscopy.75-77 From. 1650'-2000 cm�1, a region where combination and overtone bands ·occur that are characteristic of aromatic substitution patterns,78•79 the absorption observed agrees with that expected for 1,2-disubstituted benzene.79 At 1287 and 1299 cm-l there is very strong absorption from c-o-c stretching (Table 3). I� dipheriyl ether this absorption occurs at 1236 cm-1• The shift to shorter wavelength for dibenzo-p-dioxin partly may be a result of increased double bond charac ter in the c-o bonds.77 With the introduction of groups, such as chlorine or nitro these absorption bonds shift to even shorter wave­ lengths, although the trend is rever�ed for significant sub- stitution in the 1,4,6, or 9 positions. For 2,3,7,8-tetra- chlorodibenzo-p-dioxin this absorption is observed at_ 1312 and 1326 cm·l and for octachlorodibenzo-p-dioxin at 987 and 1002·cm-1 (Table 3). In general pure samples of dibenzo-p -dioxin deriv�tives can easily be distinguished on the basis 75 s. Kimoto, J. Pharm. Soc. Japan, ll, 763(1955). 76M. Narisada, Yakugaku Zasshi, 79, 177(1959). 77J.-Y. Chen. J. Assoc. Off. Anal. Chem., 56, 962(1973). 78 L. J. Bellamy, "The Infrared Spectra of Complex Molecules," John Wiley and Sons, New York, N.Y., 1958, p.67. 79J. R. Dyer, •Application of Absorption Spectroscopy of Organic Compounds," Prentice-Hall, Inc., Englewood Cliffs, N.J., 1965, p.52. TABLE 3. Infrar4!d absorption frequencies of chlorinated.dibenzo-p-dioxins (ref. 77). Derivative C-H·Stretch C-O-C AsymC-O-C Symc-c1 . metric stretch metric stretch stretch Unsubstituted 3058w 3040w 3022w 1625 m 1590 s 1299vs 1287vs 1-Chloro- 3082vw 3056vw 3016vw 1616w 1580111 l297vs 129lvs 930ms 704m 2-Chloro- 3064vw 3044vw 1622w 1585m 1299s 1294s 922ms 718s 1,6-Dichloro- 3088w 3076w 1578s 1570m 1297vs 951vs 710m 7038 2,3-Dichloro- 3086w 3054w 1574m 1314s l305ms 920w 856ms 2,7-Dichloro- 3084w 3064sh 3040vw 1585m 1570w 1302s 892m 804ms 739m 3082w 3044vw 1612wm l582ms 1323m 1307s 942ms 805m 798m 1,2,4-Trichloro- 3092w 3056vw l616wm 1575m 1286vs 968ms 818m 1,2,3,4-Tetrachloro- 3058w 1614w . 1558m l289ms 967ms 839ms 2,8-Dichloro.� \� c-c Aromatic ring skeletal vibration - TABLE 3. (contd.) Derivative C-H Stretch c-c Aromatic ring skeletal vibration c-o-c-Asym111etric stretch C-0-C Symmetric stretch c-c1 stretch 1,3,6,8-Tetrachloro- 3086wm 3072sh 1615m 1675s 1300ms 2,3,7,8-Tetrachloro- 3120vw 3076wm 3026vw 1568s l556sh 1326s 1312s 1,2,3,4,7-Pentachloro- 3108vvw 3070vw 3032vvw .1610w 1558m 129Swm 972m 859m 806m l,2,3,4,7,8-Hexachloro-3108vw 3086vw 3060w 1558m 1110111 .973m 859m 836s l,2,4,6,7,9-Hexachloro-3086m 1568m Octachloro 1552111 1536sh 988s 967ms 1002s 987m 977s 827s 876sh 870s 840ms 854s 845m ... w 32 of their infrared absorption. Data for several chlori- nated derivatives are listed in Table 3. 0V absorption data for a number of substituted di­ benzo•p-dioxins, especially the chlorinated derivatives, have ·. · been tabulated by Pohland and Yang26 and by Kende and_,Wade. 32 TWo maxima are observed, one centered in the vicinity of 250 run and one at about 300 run. For unsubstituted dibenzo-p-dioxin in chloroform the molar extinction coefficients are 1020 As-the degree of chlorine substitution increases the magnitude of these values tends to reverse (in chloroform) until with octachlorodibenzo-p-dioxin they are 13150 and 2400 (Table 4). Thus, to a certain extent, the ratio of the extinction coefficients is diagnostic of the de- gree of substitution. With increased substitution there is also an irregular shift of the absorption maxima to longer wavelengths. Phosphorescence emission spectra have been ob­ tained, for a number of chlorinated dibenzo-p-dioxins 26 (Table 5). As with absorption spectra a bathochromic shift is ob- served with increasing substitution. 'fhe phosphorescence decay time decreases with increasing substitution. Very strong absorption at visible wavelengths is observed for the cation radicals of dibenzo-p-dioxin88 (see below). The position of maximUlll absorption for chlorinated derivatives (Table 6) varies considerably depending on the extent and position of substi­ tution. 88A. E. Pohland, G. C. Yang, and N. Brown Environ. Health , Perspec., (5), 9 (1973). ,, r rz::tce t'':?tZt" < � r»t(ti e 33 TABLE 4. Ultraviolet absorption of chlorinated dibenzo-p-dioxins (ref. 26). a Derivative Unsubstituted 1-Chloro2-Chloro2,7-Dichloro2,8-Dichloro2,3-.Dichloro1,2,4-Trichloro1,2,3,4-Tetrachloro2,3,7,8-Tetrachloro1,3,6,8-Tetrachloro1,2,3,4,7-Pentachloro1,2,3,4,7,8-Hexachloro1;2,4,6,7,9-HexachloroOctachloro4 ). b 1 £ C ).2b 248 248 248 1020 :1.320 1140 1340 1180 1830 5290 6290 2970 5540 5920 5370 4450 13150 293 294 299 302 299 304 247 247 247 253 257 248 250 259 259 259 261 Measured in CHc13• bunits nm. cunits cm. liter. mole-1• 1 294 317 310 305 306 316 310 318 £ 2 C 3680 3190 3700 4590 4690 3190 2250 2290 5590 .3440 2690 3660 1480 2400 34 TABLE 5. Phosphorescence emission of chlorinated dibenzo-p-dioxins (ref. 26) .a Derivative Unsubstituted l-Chloro2-Chloro2,7-Dichloro2,8-Dichloro2,3-Dichloro­ l,2,4-Trichloro1,2,3,4-Tetrachloro-" 2,3,7,8-Tetrachloro '"1,3,6,8-Tetrachloro­ Octachloro8 Excitation, l(nm) 294 296 295 300 304 312 304 304 305 300 310 Emis6ion bands, l(run) 386,395,410,418,426,438 396,402,422,430 400,420 404,412,426,433,458,465 400,404;412,420,428,432,440,450, 408,418 458 412,418,427,439 485 (very broad) 411,418,422,436,440,448,468 410,418,433,438,465 433,458,490 Measured in ethanol-isopentane-ether, 2:5:5 at 77 °K. 35 TABLE 6. Ultraviolet absorption o� ihlorinated dibenzo-p-dioxin _cation radicals (ref. 89). Derivative Unsubstituted l-Chloro2-Chloro2,7-Dichloro2,8-Dichloro2,3-Dichloro1,2,4-Trichloro­ l,3,6,8-Tetrachloro2,3,7,8-Tetrachloro1,2,3,4-Tetrachloro� 1, 2, 3, 4-Pen tachl'oro1, 2, 4, 6., 7, 9-Hexachloro- ' 655 708 674 720 762 742 706 733 845 758 790 750 8 Measured in trifulorosulfonic acid. bsince the concentra,tion of-�ation radical was not known, the moler absorptivity could not be calculated. ,..;;;.,Z...#. v , .. <"), l !Rm.Ill 36 Pohland and Yang 26 recorded the nmr spectra of several chlorinated dibenzo-p-dioxins as well as of the parent com­ pound.BO As anticipated for o-substituted benzenes, the peaks .are centered in the region 66.80-7.20 (relative to te­ tramethylsilane), shifted slightly upfield relative to aromatic systems without oxygen (Table 7). Although 35c1 and 37c1 have non zero magnetic moments, they both have large electric quadrupole moments 81 which rule out nmr studies of these �nuclei in the chlorinated derivatives. As discussed above, 'Tomita 5 3 showed in 1 932 that d;­ benzo-p-dioxins could be oxidized in solution to give an in­ tensely blue species, a species which in 1963 was -identified -by electron spin resonance (esr) spectroscopy as a radical cation.54i56 These cation radicals appear to be extremely stable. Yang and Pohland 55 reported that in a degassed and sealed sample of dibenzo-p-dloxin in trifluoromethanesulfonic acid there was no decrease in cation radical concentration ·after a period of six months. In part because of the well resolved spectrum it produces -- 1e and 13c splittings are readily visible -- the dibenzo-p-dioxin radical has been the 801. c. Calder, R. B. Johns, and J. M. Desmarchelier, Aust. J. Chem., 24, 325(1971). 81J. w. Emsley, J. Feeney, and L. H. Sutcliffe, •nigh Resolution Magnetic Resonance Spectroscopy,• Vol. II, Pergamon Press, Oxford, 1966, p.1095. 37 TABLE 7. Nuclea r magnetic resonance of chlorinated dibenzo-p-dioxins. Derivative Unsubstituted 1-Chloro2-Chloro2.7-Dichloro2,8-Dichloro2,3-Dichloro1,2,4-Trichloro2,3,7,8-Tetrachloro1,3,6,8-Tetrachloro1,2,3,4-Tetrachloro­ l.2;3,4,7-Pentachloro1,2,4,6,7,9-Hexachloro- Chemical shifta 06.81 06.85 06.78 a6.82 a6.80 a6.96 a6.88 a6.97 06.90 06.96 06.96 07.18 (AA'BB') (pattern) (superposition of ABCD.and ABC) (AA' BB'), 06. 81 (ABC) (ABC) (ABC) (AA'BB'), 07.02 {singlet) (broadsinglet), 7.00 (singlet) (singlet) (AB) , o7. 02 (AB) (broad singlet) (ABC) (singlet) &Measured in coc13, rela tive to TMS. rwe·ezzutrznremnt@trctWt'.Cltf"·----,«titirf'it'ititttt:r�t ·rt K trtttitr ·. ··1fi:retiiet?T:s&f'"t rt" :rt�·,.. · 11¥'r 38 subject of a number of esr stu¢lies. 54-56• 82-8 \.ang and Pohland8 9 and Pohland, Yang, and Brown88investigated the esr spectra of the cation radicals of several chlorinated dibenzo-p-dioxins. They found that it was possible to dis- tinguish pure isomers on the basis_ of di.fferences in g-factors · and line widths (Table 8). The ease of oxidat-ion of dibenzo-p-dioxin has also ·provided the basis for electrochemical studies.90-93 82M. Tomita and 83 M. Tomita and s. s. Ueda, Chem. Pharm. Bull., 12, 33(1964). Ueda, Chem. l>harm. Bull., g, 40(1964). 4 & 8• Ueda, Chem. Pharm. Bull., g, 212 (1964). ssJ. P. Halrieu, J. Chim. Ph;k:S., 62, 485(1965). 86B. Lamotte and G. Berthier, J. Chim. Phys., 63, 369(1966). · 87 S. P. Sorenson and W. H •. Bruning, J. Amer. Chem. Soc. , !!, 6352 (1972). 89G. c. Yang and A. E. Pohland, Adv. Chem: Ser., ill• 33(1973). 90c. Barry, G. Cauquis, and M. Maurey, Bull. Soc. Chim. France, 2510 (1966). 91G. cauquis and M. Maurey, c. R. Acad. Sci. Paris, Ser. c., 266, 1021(196 8 ) •. 92w. Schroth, R. Bo:i:sdorf, R. Herzschuh, and J. Seidler, z. Chem., !.Q., 147(1970). 93a. Cauquis and M. Maurey-Mey, Bull. Soc. Chim. France, 3588(1972). ., .%. Jt. _J. %ti>- ,,V#� ,J._A .4U.�¥§ef?'"�.%#-f$.£.lj(.fu(d�-K½Q4f·.4?J.¥?·1'4-�J)J.i .4 ·,:·�- $�41 .. J.1,h:i?-iff;i!ff4!!j:f ... ;�}\,.?)@_ ,\¥.&;,.. %91¾��-i�-tfili_;fl -.�.4!9#?$.#/k..i.:-�.; � aau RIIRiM- .l r ,_, l I 39 TABLE 8. Electron spin resonance of cation radicals of chlorinated dibenzo-p-dioxins (ref. 89). a Derivative g-Factor Unsubstituted t-Chloro2-Chloro:?,3-Dichloro2;7-Dichloro2,8-Dichl.oro1,2,4-Trichloro1,2,l,4-Tetrachloro· 2,l,7,8-Tetrachloro1,3,6,8-Tetra chloro.l,2,l,4,7-Pentachloro·1,2, 3,4, 7 ,8'.'"'Hexachloro1,2,4,6,7,9-HexachloroOctachloro8 2.0038 2.0027 2.0034 2.0018 2.0026 2.0024 2.0029 2.0017 2.0020 2:0025 2.0019 2.0017 2.0024 2.0016 .U(oe) b 0.090 3.26 2.13,4.91 4.34 3.33 2.45 4.34 4.23 3.16 3.79 3.42 3.68 4.86 3.39 1.82 0.67 .3.31 2.17 2.31 1.50 3.47 3.10 2.13 3.18 !.60 2.29 3.76 measured in trifluoromethan�sulfonic Line width Coe) a cid. b.a • H -H where H0 is dibenzo-p-dioxin and a1 is chlorinated dibenzo-p-diOxl81 v0 • 9507.5 GHz. 40 In anhydro�s acetonitrile this compound undergoes a one electron reduction at 1.0 9V versus an Ag/Ag + (l0-2M) electrode. 9 1 The species produced was shown by esr to be dibenzo-p-dioxin cation radical. As another indication of the stability of this cation radical, it was possible to isolate a blue-black perchlorate salt. In solution the cation radical had an intense uv absorption maximum at 666 nm (see above). In the presence of traces of water, decomposition occurred, apparently via a dibenzo-p-dioxin o-quinone.93 The mass spec_trometry of the dibenzo;.. p-dioxins has ·been investigated only recently. Lamotte and Berthier86 and Nounou 94 measured the ionization potential of dibenzo-p -dioxin by mass spectrometry (8.l0eV) and Porter and Baldas 95 reported that this compound lost O and co on _fragmentation, but details of the spectrum were not given. Calder,� !.!_.96 · in 1970 described the complete mass spectrum of dibenzo-p-dioxin. The molecular ion was quite strong and the loss of O, CO, and c2o2 was confirmed (Table 9). In 1971 Buu-Hoi, !_! !.!_.25b reported a spectrum for 2,3,7,8-tetrachlorodibenzo-p-dioxin which dif­ fered significantly from spectra obtained in the present in­ vestigation and in other laboratories (Table 9). An accompanying spectrum of 1,3,6,8-tetrachlorodibenzo�p-dioxin 94P. Nounou, J. Chim. Phys., 63, 9 94(1 966). 950. N. Porter and J. Baldas, "Mass Spectrometry of Heterocyclic Compounds," Wiley-INterscience, New York, 1971, p •. 223. 96 I. c. Calder, R. B. Johns, and J. M. Desmarchelier, Org. Mass• Spectr., !, 121(1970). ,J r. TABLE 9. Mass spectra of chlorinated dibenzo-p-dioxins. Derivative Unsubstittited 1,3-0ichloro .1, 6-Dichloro2,3-Dichloro2,7-Dichloro• 2,3,7-Trichlorol,2,3,4-Tetrachloro1,2,3,8-Tetrachlorol,3,6,8-Tetrachlorol,3,6,8-Tetrachlorol,3,7,8-Tetrachloro2,3,7,8-Tetrachloro- " • • Octachloro- M 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 Relative intensity (l) a M-COCl M-Cl 5 11 48 10 7 10 11 15 5 28 15 11 7 7 6 5(M-CO) 24 38 19 60 46 40 39 29 45 19 40 35 50 35 22 21 14 M-Cl2 10 5 M-c2o2cl2 M2+ Ref. 18(M-c o ) 2 2 14 20 80 32a 97 32a 97 32a 26 41 25 3 20 5 7 12 7 15 14 14 15 5 4 4 3 b 97 26 32a 97 26 32a 101 102 101 C aAll spectr a were obtained at 70 ev ioning voltage, but source temperatures varied. bc.-A. Nilsson, K. Andersson, c. Rappe, and s.-o. Westermark, J. Chromatogr. 96, 137(1974). . .· C L. Shadoff, Dow Chemical co., personal communica tion. .. agreed well with similar spectra obtained here. The 2,3,7,8- isomer used by Buu-Hoi, !!, al. had been prepared by a different synthetic route,· so their synthesis was repeated. The mass spectrum obtained with this product disagreed with that reported by 8uu-Hoi, !!. al. and agreed with spectra obtained previously in this laboratory. Buu-HoI and Saint- Ruf These results were communicated to Buu-HoI. 97 then published a mass spectrometric study of several chlorinated dibenzo-p-dioxins which contained the correct 2,3,7,8-tetrachlorodibenzo-p-dioxin spectrum and attributed the anomalous spectrum to organometallic contamina­ tion in the mass spectrometer source which catalytically de­ composed the 2,3,7 ,8-compound during analysis. Firestone, !!, !!.•,98 Pohland and Yang,26 Jensen and Renberg, 9 9 Rappe and Nilson, 100 Plimmer, !!. !!•,lOl ICende and Wade, 32 and Baughman and Meselson, 1 02 also have published mass spectral data for 97N. P. Buu�HoI, G. Saint-Ruf, and M. Mangane, J. Hetero­ cyclicChem., !, 691(1972). 98 D. Firestone, J. Ress, N.• L. Brown, R. P. Barron and J. N. Damico, J. Assoc. Off. Anal. Chem., 55, 85(1972). 99s. Jensen and L. Renberg, Ambia,!, 1(1972). 10 0c. Rappe and c.-A. Nilsson, .J.Chromatogr., £, 247(1972). lOlJ. R. Plimmer, J. M. Ruth, and E. A. Woolson, J. Agr. . Food Chem., ll, 90(1973). l02R.- Baughman and M. Meselson, Adv. Chem. Ser., 120, 9 2 (1973). 43 .. the chlorinated dibenzo-p-dioxins. marized in Table 9. These results are sum- As is apparent from the Table, frag- mentation is not extensive. In each case the molecular ion predominates and there is a strong signal for the doubly charged molecular ion. the loss of Cl and COCl. The only signific4nt fragmentation is The molecuiar ion constitutes about 201 of the total ion current (including lower molecular weig�t fragments not shown in Table 9). In general the spectra of different isomers are so similar that the isomers cannot readily be distinguished on the basis of their fragmentation patterns. This is illustrated in Table 10 where detailed fragmentatiaipatterns for three tetrachloro isomers are listed. Attempts have been made to determine the properties of dibenzo-p-dioxin by semiempirical molecular orbital calculations. The results although interesting have not always been consistent with measured properties. On the. assumption that dibenzo-p-dioxin is •as expected" nonplanar, and without referring to the structural studies discussed above, Wratten o orbital theory could and Ali103set out to show that mlecular .be used successfully to treat nonplanar as well as planar aromatic systems. They were able to justify their assumption of non- p1anarity and calculated a 20 ° deviation from coplanarity for the two aromatic rings. However, as discussed above, physical measurements indicate that the molecule!! planar. Wratten and Ali were able to produce a reasonably good approximation l03R. J. Wratten and M. A. Ali, Mol. Phys., 13, 233(1967). TABLE 10. Comparison of detailed mass spectrometric fragmentation pattern for three tetrachlorodibenzo-p-dioxins (ref. 26). Ion 320 285 257 250 194 us . 113 110 109 (M) +5 (M- 3 S + + > (M-CO !Cl) (M-35cl )+ CM-c2o2235c12,+ 1,3,6,8 100 15 5 10 12 6 45 40 12 10 40 20 87 18 5 36 20 2 10 71 2 2 86 85 84 1, 75 74 52 51 50 100 15 35 7 99 97 Relative intensity (I) 2,3,7,8 1,2,3,4 45 45 35 20 40. 47 12 10 5 25 22 0 5 68 5 5 50 100 7 40 10 50 5 15 0 7 17 10 15 2 7 4 23 23 25 33 25 55 TtMtH. f ·· r: · ClfH . -,S#ii+tir''liia� 45 of the dibenzo-p-dioxin uv absorption spectrum. Somewhat more refined calculations were carried out by Kamiya 104 using a Pariser-Parr-Pople self consistent field method in­ clu�ing configuration interactions. approximation of the uv spectrum. He also produced a good A value of 8. 4 3 eV was calculated for the ionization potential and an experimental value of 8.10 ev, determined. by electron impact,86 was cited. Using a somewhat more restricted Pariser-Parr-Pople procedure, Bloor, et ai.·57b calculated an ionization potential of 7.85 eV and cited an experimental value 0£ 7.80 eV which was obtained from uv absorption spectra of dibenzo-p-dioxin charge transfer complexes. 57b The ionization potential has not bemmeasured by photoelectron spe·ctroscopy, but presumably this would pro­ vide the most accurate value.105 It is interesting to note that calculated w-electron densities 104 incorrectly predict electrophilic attack at the 1- or a-position of dibenzo-p-dioxin (Figure 2). 104M. Kamiya, Bull. Chem. Soc. Japan, il, 3929(1970 ). 105s. D. Worley, Chem. Rev., 1!., 295(1971). 46 Fi2ure 2 •. Calculated rt-electron densities and bond orders for dibenzo-p-dioxin {ref. 104). A similar situation occurs with thianthrene for which Chandra 106 predicted electrophilic attack at the a-positions, although Gilman and SWayampati l07 had already shown that it occurred at the S-positions. In phenazine {dibenzo-p-diazine) electro.For philic substitution does occur at in the a-positions.1 0 8 anthracene, the parent hydrocarbon of this system, it is not possible to make a direct comparison because initial substitution readily occurs at the s- and 10- positions. After these positions are substituted, however, reaction continues at the S-positions.109 In 4ny case for the heterocyclic systems it is not possible to predict the site of electrophilic 106A. K. Chandra; Tetrahedron, 19, 4 7 1(19 6 3). 107u. Gilman and D. R. Swayampati, J. Org. Chem., ll, 313(1958). 108s. Maffei, s. Pietra, and A. Cattaneo, Gazz. Chim. ll!!:_, 83, 327(1953). 109F. Kaufler and M. Imhoff, Chem. Ber., 37, 4706(1904). 47 attack simply on the basis of calculated ff-electron den­ sities. For 2 ,3,7,8-tetrachlorodibenzo-p-dioxin calculations indicate little change in the ff-electron densities following substitution with chlorine 32 (Figure 3). 1,986 Cl Cl Figure 3. Calculated a-electron densities and bond orders for 2,3,7,8-tetrachlorodibenzo-p-dioxin (ref. 32a). Molecular orbital calculations also have been carried out for the dibenzo-p-dioxin cation radical to attempt to corre­ late spin densities and other properties.85,86,lOl 48· e. Natural Occurrence of Dibenzo-p-dioxins No naturally occurring chlorinated dibenzo-p-dioxins are known. The dibenzo-p-dioxin nucleus, however, is found in bisbenzy�isoquinoline alkaloids isolated from Menisperm­ ac.eaous plantl5. Kondo and Tomita110in 1932 described the main structural features of two of these compounds, trilo­ bine and isotrilobine: OMe Trilobine (R = H) Isotrilobine (R • Me) Since then several other alkaloids from the same or a closely related family of plants have been shown to include the dibenzo-p-dioxin moiety. These include micranthine, menisarine, 110 a. Kondo and M. Tomita, Ann. Chem.,fil, 104( 1 932). / normenisarine, tiliarine, a nd tiliacorine.111-114 _Bow the dibenzo-p-dioxin ring system is formed is an •interesting biosynthetic question. Barton and Cohen115a andBattersby llSb ha ve proposed that coupling is involved. _In fact, an an oxidative.phenolic enzyme has been found in several plant species which oxidizes catechol to dibenzo-p­ dioxin-2,3-diquinone.116a-d Another enzyme which reduces the diquinone back to the c�techol has also been described.116e Although the dioxin containing alkaloids appear to be relatively nontoxic -- tl).e lethal dose for trilob ine in the frog and mouse is about 500-1000 mg/kg -- it may be portentous · _that sev�ral other closely related bisbenzylisoquinoline alka­ loids are noted for their,high toxicity, !.:.S.• tubocurare. 111I. R. c.Bick, J.B. Bremmer, H. M. Leow and P. Wiriyachitra, J. Chem. Soc., 2994 (1972). 112T. Kametani, "The Chemistry of the Isoquinoline Alkaloids,•· Hirokawa Publishing Co., Tokyo, 1968, p.63 •.. 113�. curcwnelli a nd M. Kulka, in nThe Alkaloids,• Vol. IX, R. H. F. Manske., Ed., Academic Press, New York, N.Y., 1967. 114M. Shamma, "The Isoquinoline Alkaloids, n Academic Press, New York, N.Y., 1972. 115 (a) D. H. R.Barton and T. Cohen in "Festschr. A. Stoll,• Birkauser, Basel, 1957. 115(b) A. R. Battersby, in •oxidative coupling of Phenols,• W. ·1. Taylor and A. R. Battersb y, eds., Marce l Dekker, New York, N.Y., 1967i p.119. �(a) w. G. c. Forsyth, v.c. Quensnel and J.B. Roberts, Biochem. and Biophys. Acta, 322 (1960). ll,· a ;mr: we -nrt · wtrt?fffFt:Ytf;' C·--eer irelitii::+·... 5(! ll6 (b) P. M. Nair and L. C. Vining, Arch •. Biochem. Biophys., !ll, 422 (1964). 116 (c) c. Jeandaswarni, P. v. S. Rao, P. M. Nair and c. S. Vaidyanathan, Can. J. Biochem., !1, 375 (1969). 116(d) c. Kandaswarni and C.S. Vaidyanathan, Biochem. J., ill, 30 (1972) • 116( ) P. M. Nair and c. s. Vaidyanathan, Arch. Biochem. e Biophys., 115, S1S (1966). 51 2.. 'l'o;dcoiogy Some of the halogenated dibenzo-p-dioxins are ex- traordinarily toxic. All of the non-halogenated derivatives that have been tested are relatively non-toxic. The toxicity of the halogenated.. -derivatives was discovered accidentally through the appearance of chloracne and other severe toxic response in workers involved in the synthesis of chlorophenols a� their derivatives, and through t he death of large numbers of chickens from pericardial edema and other lesions caused by a toxic fat food supplement containing chlorinated dioxins. a. Chloracne The term chloracne was used first in 1 8 99 by Herxheirner 1l:8 to describe a severe form of acne observed in workers in plants producing chlorine by electrolysis. The disease was characterized by numerous comedones and retention cysts which later were shown to be caused by hyperkeratosis. Herx- heimer attributed the disease to exposure to chlorine gas, but in retrospect it seems likely that the actual cause was • chlorinated aromatic compounds formed by the action of the chlorine on tar used to protect the electrolysis towers.119 Since that time chl.oracne outbreaks have occurred periodically 118x Herxheimer, Munch. Med. Woch., 46.l, 278(1 8 99). . 119x. H. Schulz, Arbeitsmedizin- Sozialmedizin- Arbeits­ hygiene, �. 25(1968). t==>n « tr 52 in chemical industries involved in the synthesis of chlorinated aromatic compounds. In the early part of this century several incidents were reported with chlorinated naphthalenes and biphenyls which were used widely as insulators in the electrical and radio indus�ries. 119, 120 Other pathological symptoms were reported, in particular, damage to the liver. 119•120 Because of the involvement of chlorinated naphthalenes, Wauer and Tel�ky121 •122 •123suggested that the term chloracne be replaced by pernakrankheit or perna disease derived from perchlorinated !!!_Phthalene. In the United States during the second world war chloracne was observed on a large scale when large amounts of chlorinated napthalenes were used to insulate electrical cables which protected ships against mag­ netic mines.11 9•120 Several fatalities involving liver necrosis were reported.119 Shelley and Kligman124 showed that penta- and hexachloronapthalenes were stronger chloracne­ gens than were either the higher or lower chlorinated homologs. 12°K. D. Crow, Trans. of the St. John's Hospital Derm. soc., 56, 79(1970). 121G Wauer, zentralblatt fur Gewerbehygiene, !, 100(1918). . 122L Teleky, Klinische Wochenschrift, 6. 1 , 845(1927). . 123L. Teleky, �, !:.!,, 897(192 7). 124w. Shelley and A. Kligman, Arch. Derm., 12_, 689( 1957). • �vd1C321 i1 ' ' · ·'fl$ t'ftr it 'frtY§ ?r' )l· t --, thtffr · ·:wr· 53 They reported that in conjunction with chloracne in man, •Many cases of fatal hepatic necrosis have been observed.• In 1947 Olafson 125 reported a_ disease in cattle, termed X­ disease, that was characterized by hyperkeratosis, liver degeneration and other symptoms. A rapid decline in.vitamin A levels was observed which it has been suggested may have been the result of the effects on the liver.126a The cause of the disease was traced to animal feeds contaminated with chloronapthalenes. By the 1950's use of chlorinated napthalenes had dropped considerably, but a new and even more potent source of chlor­ acne was encountered in the production and processing of chlorophenols. This new source of chloroacne in fact had made its pre�ence known during the 1930's when chlorophenols were introduced by the Dow Chemical Company under the trade name of Dowicides, as biocides, in particular as wood preservatives.* Patents were obtained in 193 5 and 1936 125P. Olafson, Cornell Vet., 37, 279(1947). 126(a) R. D. Kimbrough, Arch. Environ. Health,�. 125 (1972). In a way the chlorophenols are a geminal product of Dow Chemical Company. The company founder, .Herbert Henry Dow, made his start in the chemical industr in the 1890·s by extracting In 1918 he obtained bromine and chlorine from bri�e. r26e a patent for the preparation of phenol by hydrolysis of 54 for the use o f alkali metal salts of 2,4,5-trichlorophenol as· fungicides.1 26b•1 26cThe Dowicides were described in a However, company prospectus dated 16 December 1936.1 26d even before this prospectus was published, the ,fir�t reports of severe chloracne associated with use o f the chlorophenols had reached the medical literature. Earlier that same year severe cases of chloracne were reported in lumber workers in Mississippi involved in treating wood with a fungicidal chlorophenol formulation, apparently . . Dowic.ide H, which consisted of primarily tetrachlorophenoi.127 In a more detailed report, Stingily 128 indicated that three or four hundred persons were involved. Initially erythema and ulceration of the skin were observed, followed by formation *(continued) bromo-benzene. 126f'l'bis procedure was adapted to hydrolysis of chlorobenzene, and by 1930 the company was operating the largest synthetic phenol plant in the world. l26 g The marriage of the basic chemical products chlorine and phenol soon produced as offspring the chlorophenols and, unwittingly, the chlorinated dibenzo-p-dioxins. 126(b) L. E. Mills (to Dow .Chemical Co.,), U.S. Patent 1,991,329(1935). (c) L. E. Mills (to Dow Chemica1:co.,), 2,039,434(1936). u. s� Patent Dow Chemical Co., Prospectus, 16 December 1936. (e) W. Hayne=, "American Chemical Industry,• !, Vannostrand co., New York, 1949, p.114. (d) (f) (g) H. B. Dow, u. s. Patent, 1,274,394(1918). M. E. Putnam, Plastics and Molded Products, 1, 255(1931) 127oueries and Minor Notes, J. Amer. Med. Assoc.� 106, 2092(1936). 128K. o. Stingily, Southern Med. J., 33, 1268 (1940). -------�.-- _ _ ti' ___ c·-55 of comedones, cysts and pustules, thickening of the skin and urinary disturbances. In some cases leg cramps, thrombosis or highly colored urine were present. The lesions were reported to be remarkably persistent, sometimes Apparently the chickens soon came- home-to roost as in 19 37 Butler 129 • 1 30 lasting several years after the final exposure. reported a similar incident at the Dow Chemical plant in Midland, Michigan involving twenty-oneworkers who had handled 2- (2-chlorophenyl)phenol and tetrachlorophenol. Marked hy- perkeratosis was observed with •enormous numbers of comedones • • • in some cases so numerous as to produce a black discoloration • • • Severe scarring and persistence of the condition were noted. In his paper Butler stated that these chemicals shouid not be used as biocides until the mechanism by which they caused chloracne was understood. The plant was closed temporarily. Butler indicated that experimentation with animals would be undertaken to.attempt to elucidate the mechanism of toxicity, but he was not supported in this effort by the Company.130 129M. G. Butler, Arch. Derm. Syph., 35, 130M. G. Butler, personal communication. 251(1937). 56 If these experiments had been comp�eted, it is possible that the chlorodioxins would have been identified as being highly toxic twenty years before the eventual discovery in 1957. Iri 1941 Adams,� al.131 of Dow Chemical reported a bioassay for chloracnegens which consisted of applying sus­ pected agents to the interior S\lrface of the rabbit ear and. monitoring acne-like changes. They tested five types of chemical products which were acnegenic, including crude chlorophenols. Although they did not discuss ·Which component of each product may have caused the .chloracne, the fact that of the classes of compound they examined the term "crude" was attached only to chlorophenols'suggests that they may have been aware that in this case compounds other than the main chlorophenol products were involved. Following the discovery of the herbicidal properties 2,4,5-T and related phenoxyacetic acids during the second World War,132 production of this compound and its precursor, 2,4,5-trichlorophenol which was prepared by hydrolysis of 1,2,4,5-tetrachlorobenzene, began on a large scale. This production was accompanied by a series of new outbreaks of 131E. M. Adams, D. D. Irish, H. C. Spencer, anci u. K. Rowe, Ind� Med., 10, 1(1941). 132w. B. House, L. H. Goodson, H. M. Gadberry, K. W. Dockter, •Assessment of Ecological Effects of Extensive or Reported Use of Herbicides," Midwest Research Institute, Kansas City, Missouri, 1967, pp.108-110. 57 chloracne. In 194 9 , 228 workers developed chloracne as the result of an accident in Nitro, West Virginia at a 2,4,5-T plant owned by the Monsanto Chemical co.133a,I 33b Other symptoms observed included severe pains in skeletal muscles, shortness of breath, intolerance to cold, palpable -and tender liver, loss of sensation in tle _extremities, demyelination of peripheral nerves, fatigue, nervousness, irritability, insomnia, loss of libido, and vertigo. Por­ phyria was not reported. reported In 1951 Baader and Bauer 134 chloracne in 17 workers in Germany who had participated in an investigation of industrial syntheses of 2,4,5-trichloro­ phenol and pentachlorophenol. Within a few years similar incidents were reported at other 2,4,S-T �nd 2,4,5-trichlorophenol plants.28• 135-l4 0. 133(a) •Report on 2,4,5-T," A report of the Panel on Herbicides of the President's Science Advisory Committee, Office of Science and Technology, March 1971, p.4 8. (b) R. Suskind, "Chloracne and associated problems,• Conference on Dibenzo-p-dioxins and dibenzofurans, Research Triangle Park, N. c., 2 April 1973. 134 E. w. Baader . and - H. J. Bauer, Ind. Med. and Surg., �. 286 ( 1951) �1358• T. Hofmann, Arch. EXE!• Pathol. Pharakd., 232, 228(1957). 136 K . H. Schulz, 137J. Kimmig and 138P. Dugois and . 63, 262 (1956). 139p. Dugois and 1 40P Arch. Klin. EXE!• Derin., lli_, 589(1957). K. H. Schulz, Dermatologica, ill• 540(19S7). L. Colomb, Bull. Soc. Fr. Derin. Syph., L. Colomb, J. Med. Lyon, 38, 899( 1 957). . Dugois, J. Marechal and L. Colomb, Arch. Maladies Prof., 19, 626(1958). I. RP L vii 58 In 1956 Dugois and Colomb 138 described one such case in­ volving 17 workers at a 2,4, 5 -trichlorophenol plant near Grenoble, France. In a unique observation·they reported that in addition to having severe chloracne and internal disturbances, those affected gave off an "intense odeuzo ohZozoe•.• They suggested that the source of the disorde= was some chlorinated cyclic hydrocarbon such as a chloro� napthalene. In 1 957 Hofmann 135described in detail an in­ cident that had happened in November 1 95 3 in Ludwigshafen am Rhein at a 2,4,5-T factory owned by Badischer Anilin& Soda-Fabrik. Hofmann reported that he had found a likely candidate for the toxic agent. He established that the unknown compound was neither 2,4,5-trichiorophenol nor its precursor 1,2,4,5-tetrachlorobenzene and that it must be at least a hundred times more toxic to rabbits than the chloronapthalenes. In fact, the compound was so toxic that con- trol animals in cages next to treated animals developed liver necrosis as did untreate� animals wnich were placed in cages previously housing treated animals. This behavior at first led Hofmann to believe that they were dealing with a virus infection. In retrospect, it appears that cross contamination may have occurred via the excreta. Various chlorinated aromatic ethers and diphenyls were investigated and found not to be particularly toxic. Then a number of chlorinated dibenzofurans were synthesized. -A . ________________________,.., ,...,__...,....,r..c.--r:llil- -•rlliieillilmrliir1111tn111till:11ro:iiirililitttliilf"w,,toilioclilniill- iiiiff'ilil@llil' iilttiiilii=11,·mwlii-""ii11Wiifill-' ... -Ol!"""& 59 Those containing three to five chlorines came close to having the required level of toxicity, both in terms of hepatotoxicity and chloracne. Hofmann proposed that these dibenzo- furans were the source of the problem in the 2,4,5-T and trichlorophenol plants. At the same time in ,�amburg, Schulz and Kimmig 28 •136•137 were carrying out an almost identical investigation. They too found that certain chlorinated dibenzofurans were ex­ tremely toxic. However, as Schulz 136 pointed· out, these compounds did not seem to be present in the technical chlorophenols. most ironic way. The solution to the problem appeared in a A man with an unusually severe case of chloracne, who had no connection at all with the 2,4,5-T or chlorophenol plants, was referred to Schulz. This man was the assistant working with Sandermann (see above) in the attempt to confirm the structure of Merz and Weith's original "perchlorophenyleneoxide" by chlorinating dibenzoBecause of p-dioxin to octachlorodibenzo-p-dioxin.24• 136 decreased solubility and decreased reactivity the reaction had stopped at 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). Kimmig and Schulz readily confirmed the extraordinary chloroacnegenic and, hepatotoxic properties of this compound in tests with rabbits.137 In a self administered skin test, Schulz demonstrated that TCDD is chloracnegenic in 60 humans.• 142 a Schulz suggested that chloracne also might be produced following ing-2stion of TCDD, but this was not tested. At about the same time, a similarly inauspicious episode in which the toxic agent was� recognized, occurred in the United States. As a part of an investigation of the substitution reactions of dibenzo-�dioxin, Dietrich syn­ thesized various halogenated derivatives, including the 2, 3,7,8-compounds. 38 He subsequently became severely ill and was admitted to the cancer ward of the Atomic Energy Commission Medical Center in Chicago. Fortunately, after an .extended period of hospitalization he slowly recovered, al­ though the chloracne remained active for about five years. In the case of Dietrich's misadventure the connection with the compounds he was synthesizing was not made. Only several years later in 1965, at about the time a new series of chloracne incidents struck 2,4,S-T plants in the United States (see below) did he learn the cause of his illness. Dietrich's story provides another ·note of irony. After he had synthesized various chlorinated dioxins, they were submitted to the Army for medicinal evaluation. The only com- pound (a non-halogenated one) which was selected for medicinal use was used in a dermatological salve.141 141J. J. Dietrich, personal communication. -�--------------------------------.. .,- 61 Having shown that TCDD was an extraordinarily potent. chloracnegen, Schulz and his colleagues then demonstrated that this compound was formed during the synthesis of 2,4,Strichlorophenol from 1,2,4,S-tetrachlorobenzene. The con- ditions required to hydrolyze the tetrachlorobenzene to trichlorophenol were found to be sufficient to cause a small fraction of the trichlorophenol to condense to form TCDD according to synthetic Route A of the preceding section. After the nature and identity of the toxic material were established, conditions (unspecified) were found which re­ duced the level of dioxin in the product and which reduced the exposure of the workers to whatever dioxin was present. This resulted in a substantial reduction in the incidence of chloracne. 142a Schulz and his collaborators went on to carry out an extensive investigation of the symptomology and etiology of chloracne. follows: 142a In summary the symptoms were described as Numerous comedones formed, first on the face, especially on the cheeks above the malar bones, forehead, temples, chin and ears, after which folliculitis, pustules, boils and retention cysts occurred as a result of secondary infections. As the disease progressed, these symptoms spread in the majority of patients, especially to the sides of the neck, Lack of the neck, 142(a) H. Bauer, K. H. Schulz, and U. Spiegelberg, Arch. Gewerbepath. Gewerbehyg. 1 ¾_!, 538(1961). 62 upper half of the back, chest, forearms, genitals and thighs. Numerous boils formed, particularly·on the back of the neck and on the back. The efflorescences were generally located so c1o·sely together t;hat scarcely any follicles remained un­ changed. Symptoms also included hyperpigmentation, erythema, swelling of the skin, hirsutism, loss of appetite,· loss of weight, gastritis, severe liver damage, pulmonary emphysema, dyspnea, myocardiac damage, edema, renal damage, plus a .long list of neurological and psychopathological symptoms including mus­ cular distrubances, disturbances in memory and concentration, decrease in initiative and interests, depressions, disturbances in libido and potency and weakness in mental capacity. The dermatological symptoms were reported to be extremely obsti­ nate, in some case� lasting for many years.142a The number of workers involved or their individual symptoms were not given. In referring to this incident, May 142b reported that some fatalities occurred as a result of severe liver damage, and that fifteen years after initial ex­ posure another individual was expected to die soon of this cause. One death from pulmonary carcinoma was rejected as being attributable to dioxin exposure, while one from intestinal sarcoma was accepted. This last observation may be significant in view of the extensive possibly pre-cancerous lesions observed in the intestinal tract of monkeys and other test animals fed a diet containing chlorinated 142(b) G. May, Brit. J. Ind. Med., ll, 276(1973). 63 ·dioxins.142c Recently Goldmann143a upda ted and expanded the de-. scription of the symptoms of the posed to dioxin following an 42 workers who had been ex­ a ccident at the BAS� plant.at Ludwigsh a fen am Rhein in November 1953135 (see. above). Especia lly valuable are deta iled individual case pistories for periods up to eighteen yea rs after exposure. A few of these case hisoq:-ies will be summarized here to supplement the generalized description tha t has been given thus far of the response in humans to 'l'CDD. One case involved a ma n who ha d never entered the He merely sat building in which the accident occurred. next to exposed workers i·n the lunchroom. Chloracne de- No other s�s were developed o n the face and forea rms. observed. In another case chloracne developed which was characterized by hen•s egg-sized abscesses. In this individua l comedones and retention cysts were still occurring eighteen years after exposure. A third case involved a mecha nic who worked for three days in the a rea of the autocla ve used to hydrolyse tetra­ chlorobenzene to trichlorophenol. On the second day pains in the spine and headache occurred. of the skin wa s followed by �hlora cne. Inflammation and swelling Numerous pustuler 142(c) D. H. NOrback a nd J. R. Allen, Environ. Health Perspec., (5), 137(1973). (a) P. J. Goldmann, Arbeitsmed. Socialmed. Arbeitshyg., 1, 12(1972). 64 infections occurred which were not improved by treatments with antibiotics or vaccines. Myocardiac damage, toxic nephrosis, massive bronchitis, sinusitus, tonsillitis, and liver damage were observed. The effects on the liver in­ cluded subacute hepatitis w_ith formation of groups of hya­ line bodies and fat vacuoles and hypertrophy. phlebitis of the left leg developed. .. regained his health. Thrombo­ The patient never Ten years after the original exposure he succumbed to coronary insufficiency and lung embolism with extensive hemorrhaging • . A fourth case involved a 57 year old employee who worked.in one of the autoclaves in 1958, five years after the original accident. Over· this period of time the autoclaves were never used to prepare trichlorophenol. Although he was wearing a complete protective suit, he had removed his mask several times to wipe perspiration from. his face. Within four day.s One headaches, loss of hearing and chloracne were present. JROnth later he was hospitalized with angina pectoris. Six months later acute pankreatitis developed and a large and very painful tumor was found in the upper abdomen. Death attributed to acute pankreatitis occurred shortly th�reafter. Autopsy revealed intestinal ulceration and liver and adipase necrosis. A fifth case involved a young bricklayer who spent two hours in the autoclave room repairing a wall. A persistent, 65 and severe case of chloracne developed. After about a year had passed his temperature became noticeably elevated and a massive X-ray opaque area appeared in the left lung. hemorrhagic pleuritis followed. were negative. slowly. · Severe Animal tests and cultures Eventually his condition began to improve Then about four years 'later he suff.ered acute psychosis with insomnia, loss of ·affect, hearing of voices, suicidal tendencies, physical discomfort and a burning sen­ sation in the back. Within a short time he committed suicide by·hanging. Several general points can be made on the basis of the cases descr�bed by Goldmann. A very short exposure time can be sufficient to produce serious· toxicity. from a single exposure may last many years. Effects Dioxin residues can persist in the chemical plant environment for at least five years. In the present instance the plant was sealed off after the incident in 1958 and later carefully demolished. One of the most prevalent symptoms appears to be decreased resistance to infection. Infections of the respiratory tract such as laryngitis, tracheitis, tonsillitis, and bronchitis were common. Neurological effects ing, smell and taste, including loss of hear- polyneuritua and sensory and motor de­ fects and encephalomyelitis, a disturbance of the central nervous system with paralysis and spasticity on one side of the body, were observed in seven of the forty-two affected · workers. Neurological damage appeared to affect the left 66 side more than the right side. complaint. Drowsiness was a common Hemorrhages in various tissues and gastro- intestinal disturbances, including ulcers, were often present. Many of these effects, and also effects on the heart, liver, and other organ systems, parallel those described · by Ba.uer, !!_ ai. 142a Porphyria was not mentioned. Except for the neurological effects, the symptoms observed in the workers are generally consistant with those observed in toxicity testswith TCDD in various animal species (see Mechanisms of Toxicity). (Few significant neurological effects have been observed in animals.) Perhaps the most stril.cing aspect of the cases described by Goldmann is the tremendous difference in the nature and intensity of the responses from individual to individual. Differences in in- dividual susceptibility also exist in animals, so much so that at times such differences have caused difficulty in determin­ ing LD50 doses for a given species.143b The experiences in Germany described above were re­ peated on a substantial scale in the United States in the mid-1960's. In response to the greatly increased demand for 2,4,5-T which resulted from the u. s. military herbicide pro­ gram in Vietnam, facilities and schedules were put under great pressure in an effort to increase production. Several 143(b) J. B. Greig, G. Jones, w. H. Butler, and J. M. Barnes, Fd. Cosmet. Toxicol., g, 585(1973). 67 outbreaks of chloracne occurred. In 1964, about the time the major phase of the herbicide program was getting under­ way, one such incident involving over seventy workers occurred at a plant operated by the Dow Chemical Company. An investiga- tion was conducted which not surprisingly led to the conclusion Although these that TCDD was the source of the problem.133a results were not published, they were communicated to other manufacturers, including Diamond Alkali, Monsanto and Hercules. In spite of the series of incidents with chloracne in connection with 2,4,S-T production in the United States, the first pub� lication describing such experiences in this country was made by Bleiberg, � !!,.,144 in 1964. They investigated workers at the Newark, New Jersey plant of Diamond. Alkali (now Dimond Shamrock) and found porphyria cutanea tarda as acne and liver dysfunction. earlier report 128 well as chlor­ This brings to mind Stingily's of •highly colored urine• associated with chloracne, with which Bleiberg apparently was not familiar. Darkly colored urine, caused by increased porphyrin excretion Poland, is one of the characteristic symptoms of porphyria. � al., 145 reinvestigated the same chemical plan.t six years later and noted that, although the porphyria was gone, chloracne was still present. 144J. Bleiberg, J. Wallen, R. Brodkin and I. L. Appelbaum, Arch. Derm., !,!, 793(1964). · · 145A. P. Poland, D. Smith, G. Metter and P. Possick, Arch. Environ. Health, 22, 316(1971). •• 68 Poland,!!_!!.:., did not see the range of psychic disturbances reported by Bauer,!! al.,142a but did observe that chlor­ acne was correlated with a high score on the manic scale of the Minnesota Multiphasic Personality Inventory test. outbreaks of toxicity similar to those described above also have been reported in 2 , 4 ,5-T and 2,4,5-trichlorophenol plants in Italy,14 6 the Netherlands,119 Great Britain 142b •147 and the Sovie.t Union.148 In fact, since the original in- cidents with Dowicides in the 1930's, over a thousand workers in at least fifteen different chemical plants in at least eight different countries have been �fflicted with chloracne and related symptoms in connection with the.production of chlorophe�ols and their derivatives. Th� early history of chloracne was reviewed by Braun.149 Later reviews have been written by Bauer, Schulz and Spiegel­ berg, 14 24 Schulz,119 Crow, 120- Kimbrough 126 and Goldmann. l4la ill• 146M. F. Hoffman and c. L. Meneghini, G. Ital. Derm., 427(1962). 147 N. E. Jensen, Proc. Roy. Soc. Med., 65, 687(197 2 )-. 148K. A. Telegina and L. I. Bikbulatova, Vestr. Dermatol. Venerol., 44(3), 35(1970). 149w. Braun, "Chlorakne," Editio Cantor, Aulendorf, w. Germany, 1955. 69 b. Chick edema factor In 1957 several million chickens died in eastern and mid-western United States of an epidemic disease involving hydropericardium, edema, and liver necrosis. The effort which followed to identify the chick edema factor (CEF) developed into a true� de� of analytical chemistry involving several laboratories.150-15� The origin of the disease was quickly traced to certain fatty acids used as food supplements, but from this point progress was tedious. A number of organic and inorganic pesticides, some of which causes edema-like symptoms, were ruled out, largely because they were not active at sufficiently low levels. CEF was shown to be in the unsaponifiable portion of the fats and a series of chromatographic steps was used to further concentrate the residue.• Three major fractions were obtained1 one nonpolar, one of intermediate polarity and one polar. The CEF was present in the second fraction which ruled out various nonpolar hydrocarbons on one hand and polar 150w. B. Brew, J. B. Dore, J. H. Benedict, G. L. Porter, and E. J. Spias, J. Assoc. Offic. Agr. Chem., 42, 120-128(1959). 151L. Friedman, o. Firestone, W� Horowitz, o. Banes, M. Anstead, and G. Shue, !lli.:., g, 129-140(1959). 152J. c. Wooton and J. c. Alexander, !lli.:., ,!!, 141-148(1959). 153s. R� Ames, w. J. Swanson, .M. I. Ludwig, and G. Y. Brokaw, J. Amer. Oil. Chem., ll, 10-11(1960). 154 R. E. Harman, G. E. Davis, W. H. Ott, N. G. Brink, and F. A. Kuehl., J. Amer. Chem. Soc., 82, 2078-2079(1960). 'Rdlo/itt Crw 70 sterols on the other, both of which were present in significant amounts in the original residue. The second fraction was largely ketonic, apparently mostly dipalmitone. Pure dipalmitone did not produce edema in a chick bioassay and so further purification of the second fraction was undertaken. A 500 tube counter-current distribution between isooctane and methanol, reaction with excess Girard-T reagent, reverse phase chromatography on silane treated Celite, and addition­ al alumina chromatography produced increasingly purer material until suddenly at one stage much of the CEF activity seemed to disappear from the major fractions. The mystery was solved when a residue of 0.5mg in an •empty• flask from a cut between major fractions was found to be the most active material of all. At this stage UV spectra suggested that CEF J!light be a napthalene or _phenanthrene derivative.151•155 One particularly toxic fat was inadverdently discovered by Portman and Andrus at.the Harvard School of Public Health.156 In conducting a nutritional study with a group of Cebus monkeys they used a high quality synthetic triolein as a dietary sup- plement. The results were rather dramatic. died within several months. fat showed high CEF activity. All the monkeys In a chick bioassay this same Using the procedures 155L. Friedman, Feedstuffs, l,i, March 17� 1962. 156A. Yartzoff, o. Firestone, D. Banes, w. Horwitz, L. Friedman, ands. Nesheim, J. Amer. Oil Chem., l!!_, 60(1961). 'NI -r· 1 · wwtm@:ffflt•e · rt" c 71 developed earlier, Yartzoff, � al.156 isolated 2.64 mg of crystalline·CEF from 17.6 kg of fat. Just prior to this Harman,!!:_ ai.154 similarly isolated a small amount of crystalline material which was shown to contain 471 chlorine. The infrared spectrwu in- dicated that carbonyl and hydroxyl groups were not present. The ultraviolet absorption spectrum, with maxima at 244 and 312 nm, suggested the presence of a polynuclear aromatic system. Polychlorinated napthalenes were ruled out by the observation that they were much less toxic and that they had a much shorter glc retention time than did the CEF materiai.156 From microcoulometric and electron capture glc it was established that the final residue consisted of several compounds of varying chlorine content.1·57 In a miscalculation that should serve as a.caveat for others attempting to identify low level enviroruaental residues, Wooton and Courchene,158 in spite of possessing two pure CEF 1570. Firestone, w. Ibrahim and W. Horwitz, J. Assoc. Offic. Agr. Chem.,,!!, 384 (1963). l58J. c. Wootton and w. L. Courchene, J. Agr. Food Chem. 12, 94 (1964). rt: r tztrrt·,@ rn:H:tu@nw-�11 z -,,ww II tn"tt ,. . 72 compounds they had isolated, good infrared and ultraviolet spectral data, complete and accurate low resolution mass spectral data, and a successful hydrogenation experiment, failed to correctly identify the compounds as dibenzo-p-dioxins. Instead they concluded that·they had isolatedderivatives of a partially saturated chlorinated phenanthrene system. They · reasoned that the infrared spectrum ruled out any oxygenated compound. In fact they failed to recognize the·strong, characteristic aromatic ether absorption band that was present '(see above). In the mass spectrum a molecular ion was observed at !!V� 388 that had associated with it an iso­ topic is0111er pattern that could only be derived from six chlorine atoms. This is ·the molecular ion of a hexachloro- dibenzo-p-dioxin, but believing that no oxygen atoms were present, they concluded the molecular formula had to be The second formula was ruled out because such a compound would not give the observed ultraviolet spectrum. 353. atom. The next ion in the spectrum was at !V� This was correctly attributed to loss of one chlorine The next two major peaks corresponded to loss of 63 and 126 mass units. These peaks are derived from loss of COCl and two COCl, fragmentations which are characteristic of chlorinated aromatic ethers. The loss of t�o COCl units • WM n "T®.l"?!fiiilM'll!r'� mt.ti -;;w · r :inn tr _--i· , tttuaz«rrr err ,·' -· · ·n · ·: t· .-, · · t· , « n • c t ""tt 73 is almost unique to chlorinated dibenzo-p-dioxins. However, from their assumption of no oxygen, Wootton and.· Courchene had to conclude that these fragments were c2H Cl and twice 4 c2H c1. A strong double charged molecular i�n (201 of the 4 intensity of the singly charged molecular ion) was observed. Such a strong doubly charged ion is suggestive of a very stable aromatic system. In an attempt to obtain chemical evidence for their, structural assignment.s, Wootton and Couchene chlorinated napthalene, anthracene, phenanthrene, stilben.e and 3,3'-di­ methylbiphenyl by electrophilic substitution until the products �ontained 50-601 chlorine and then catalytically hydrogenated each mixture. Each ring system produced a characteristically different set of hydrogenation products , as determined by glc analysis. Uydrogenation of the CEF compounds gave products with retention times similar to those of phenanthrene, although the product ratios were different. Since from the mass spectral data the unknown compounds had to be partially saturated and since the un­ known compounds might have a chlorine substitution pattern different from that obtained by electrophilic substitution, circumstances which might produce different product ratios, Wootton and Courchene took the result as indicative of a phenanthrene system. structure: _They proposed the following general t tc 74 In retrospect it may be noted that a single high resolution mass spectral measurement of the molecular ion would have ruled out a phenanthrene and probably would have established the correct molecular formula. A short time later the question of the basic structure of the compounds in the CEF mixture was resolved by X-ray crystallography when Cantrell, !!_·a1.68 showed that one of the CEF compounds was l,2,3,7,8,9-hexachlorodibenzo-p-dioxin (see Chemistry section above). Wootton then demonstrated that a synthetic hexachlorodibenzo-p-dioxin prepared by chlorinating dibenzo-p-dioxin, had physical and toxicological properties similar to those of the CEF compounds.159 Higginbotham, !.!:_·a1.25aproduced various chlorinated dioxins by pyrolyzing different chlorophenols and showed that some of these had -glc retention times identical with peaks in CEF residues isolated from toxic fats. Flick,!.!:_ ai.160 159J. c. Wootton, personal communication. 160n. F. Flick, D. Firestone, and G. R. Higginbotham, Poultry Sci., 51, 2026(1972). tr.'·-::,.-;• 7S established that mixtures of other chlorinated dibenzo-p -dioxin homologs fed to chicks also caused edema. The • most toxic mixture tested was a combination of 2,3,7-trichloro- and 2,3,7,8-tetrachlorodibenzo-p-dioxin. Eventually the CEF peaks observed with glc were confirmed as chlorinated dioxins by GC/MS.161 It now appears likely that a substantial fraction of the dioxins in the toxic fats got there through the use of chlorophenols to preserve hides from which ·tallow was extracted. This is discussed in more· detail below in the section on Environmental Sources of ·chlorinated Dioxins. 161J. Ress, G. R. Higginbotham, and D. Firestone, J. Assoc. Offic. Anal. Chem.,!!• 628(1970). 76 c. Acute Toxicity Although chlorodioxins undoubtedly were the cause of the periodic outbreaks of chloracne in industrial chlorophenol plants over the past several decades, the dioxins were not identified and so there was no clue as to their uniquely high toxicity. The first indication of this toxicity came when ltimmig and Schulz 137 observed that 2, 3,7,8-Tetrachlorodibenzo­ l!-dioxin was lethal to rabbits with.single oral doses as low as so µg/kg. Schulz 119 later reported the single oral dose LJ>50 (dose sufficient to cause death in 501 of the animals in a test group) in rabbits .to be 10 µg/kg bodYWeight (10 parts per billion or ppb bodYWeight). The chick edema in- cident confirmed the toxicity o.f th� chlorod"ioxins and in addition brought forth the observation that the toxicity of these compounds was dependent on the degree and position of chlorination.1601162 Acute toxicity data for various chloro- dioxins are sUJD111arized in Table 11. One important.point about dioxin toxicity, which will become more apparent from the discussion to follow, is that "acute" effects may take weeks or possibly even months to become visible. 162 D. P. Flick, D. Firestone, J. Ress and J. R. Allen, Poult;rsci., g, 1637 (1973). Table ll. Derivative Acute toxicity of chlorinated diben:zo-,P-cUoxins. Specie, Route of ad.mini• stration· 'l'ime of death Single oral doae after treat• ·.· Lo 0 (ng/kg or pp·£ bodyweight) inent (days) 2,7-Dichlorooral Rat 2,3,7,8•Tetrachloro- Guinea pig, ule oral 5•34 • Guinea pi9, female ora1 9•42 " Oral Guinea pig 18 (mean) • lnj. ih eft Chick·· embryo Chick Oral • Oral Rabbit II Rabbit Oral • Rabbit Oral 6•39 • Rabbit. Skin 12-22 " Rat, male Oral 9•27 • Rat, female Oral 13•43 Bexachloro• Chick Oral· • Rat Oral OCtachloroChick Oral • Rat Oral • Mouse Oral �.w. >2 X 109 600 2100 <3000 5 X 10 8 . >10 9 >4 X 10 9 :1 Ref. 163 163 163 a .. 25 10 356 11 9 163 163 163 163 163 163 163 163 163 Harris, J.A. Moore, J.G. Vos, B.N. Gu,pta, Environ. Health Pers2ec. (5), 101 (1 9 73).• ... II I . ;- · ·· - ··---- - ·�·- · 78 Also it is important to be aware that individual differences in susceptibility within a species may be so large as to seriously impair determination of an Lo50 •143b,l 6 3 Large strain differences in susceptibility to enzyme induction by 'l'CDD within a species have been observed as we11.164•165 The non-chlorinated dibenzo-p-dioxins are relatively non-toxic. Dibenzo-p-dioxin itself has been tested as an anthelminthic and nematocide, presumably because of its structural resemblence to phenothiazine, but it performed �ta and Watanabe1691170 . rather unsuccessfully.166•1671168 163a. A� Schwetz, J.M. Norris,G. L.·Sparschu, v. K. Rowe. and P. J. Gehring, Adv. Chem. Ser., 55(19 7 3) • 164J. B. Greig and F •. DeMatteis, Envir�n. Health Perspec. . (5), 211(1973). 165A. P. Poland, E. Glover, J. R. Robinson and D. w. Nebert, J. Biol. Chem., in press. 166L. E •. smith and Roy Melvin, J. Econ. Entom. !§., 4 75 (1943). 167w. P. Rogers, J. Cymerman-Craig and G. P. Warwick, Brit. J. Pharmacol., 10, 340(1955). 168N. v. Phiilips, British ?atent 870 ,298(1 961). m, ·· 169M. Tomita and w. Watanabe, J. Pharm. Soc. Japan, !!, 120(1951). 170M. Tomita and w. Watanabe, J. Pharm. soc. Japan, 72, 478(19 52). 79 investigated the antibacterial activity of several substituted dioxins, incl�ding alkyl, alkoxy, amino, keto, amido and other derivatives. The only chlor.inated derivative tested was 2, 7- · dichlorodibenzo-p-dioxin. larly active. None of the compounds were particu­ Qkada and Fuse117 found the dibenzo-p-dioxin containing alkaloids relatively non-toxic. The lethal dose for trilobine in frogs and mice, as mentioned before, was 500 and 1,000 mg/kg respectively. Motor paralysis was observed and death was caused by respiratory failure. d. Chronic toxicity Only a few studies have been done which were designed to measure the effects of chronic exposure to chlorinated dioxins. 'l'he longest by Allen and Carstens171 involved the administration of a CEF toxic fat as a supplemenet to the daily diet of Macaca mulatta monkey over times up to 445 days. The fat used was later shown to contain predominantly tetrachlorodioxin.1 62 Perhaps the most striking result was the remarkably cumulative nature of the dioxin toxicity. This is illustrated in Figure 4 where the mean survival time is plotted against .the reciprocal of the percent of toxic fat present in th.e diet. With the ex- ception of the highest dose, the points conform well to the relation 'J.'=K/0 + K', where 'l' is mean survival time, Dis daily dose, and Kand K' are constants ce>rresponding respectively te> !!, 171J. R. Allen and L. A. Carstens, Amer. J. Vet. Res., ( 1 96 7 ). 1 51 3 400 � t- 300 K K' T• -+ D �� ->o'! �� :::,- 200 100 4 8 RECIPROCAL OF THE PERCENT TOXIC FAT IN DIET Figure,. Mean survival time of monkeys fed toxic fat plotted against the reciprocal of the per cent of toxic fat present in the diet (ref. 189). •... --, . st. t ti ttb" i ·-r&' ·-,c '"'tt'imet&·,·f,�·w·,+e� 'it'!,,,,-, ,·er· i &b'ttnwwr-� t 82 the accumulated lethal dose and to the lag time between the accumulation of this dose and death. This means that at least up to 445 days a dose administered in many small portions over a long time is just as effective as the same dose administered in large portions over a short time. This is supported by the observation that the conditions surrounding the death of animals were the sanie, regardless of the dose level. As Allen and Carstens reported, •the major clinical and pathologic changes were similar and occurred during the terminal 30 days regardless of whether the monkey survived for less than 4 months or longer than l year.• For such behavior to occur the effects of the dioxins must be cumulative over periods of a year or two and possibly longer. The dioxins themselves may or may not be retained for such periods. Since measure- ments of the dioxin residues in the tissues of the experimental animals were not made, there is no evidence from Allen and Carstens• work as to whether the dioxins, as well as their effects, are cumulative. · New chronic toxicity experiments with monkeys, probably the most appropriate test animal for estimating effP.cts in humans ,, are to be under !:liken using pure In thls case dioxin 2,3,7,8-tetrachlorodibenzo-p-dioxin.172 levels in tissue will be monitored. From another point of view Allen and.Carstens' results 172J. R. Allen, personal communication. "t'dtir ,,-·ir · frttitm--Xt · · ·tni-·. 83 can be restated to say that recovery,if any.from dioxin toxicity is slow compared to 445 days. corollary is that chronic exposure does not result in increased tolerance over a period of 445 days. If either of these points did not hold,. a deviation from a cumulative response should have occurred at the longer times. If dioxin toxicity is in fact cumulative, then chronic exposure to levels that are low relative to single dose toxic levels could lead to serious effects over the .long term. This point is of critical importance in deciding what levels of sen­ sitivity are required for an adequate environmental monitoring scheme. The implications of this are discussed in Part B of the Introduction. Some other studies of the effects of repeated exposure to chlorinated dioxins in rats, mice, and guinea pigs have been carried out, but these were of less than about 60 days duration and, although the results are generally consistent . with Allen and Carstens• data, they were not of sufficient length to be a good test of cumulative effects. e. Teratogenicity, carcinogenicity, and mutagenicity The first indication that the chlorinated dioxins were teratogenic in addition to being highly toxic came during the chick edema episode. As a part of a general study of CEF toxicitY, chlorinated dioxions were injected into chicken eggs ·err" o1 - ···-----·�---- ·ltflt#m 84 at an early stage of development. Hatchability was decreased and malformed beak, lack of development of the right mesen­ cephalon, eye defects, growth retardation and leg deformities were observed. 162•173 A further sign of the teratogenicity of chlorodioxins resulted indirectly from the finding in studies at the aionetics Research Laboratory that some 2,4,S-T samples caused birth malformations in rats and mice.174 The 2,4,S•T used was found to contain about 27 ppm 2;3,7,8-tetrachloro­ dibenzo-p-dioxin.175 This prompted new experiments which showed that 2,3,7,8-tetrachlorodibenzo-p-dioxin itself was teratogenic in rats and mice at dose levels of 0.1-1 mg/kg.175•176 Although TCDD probably was the cause of some of the effects observed in the Bionetics study, pure 2,4,S-T also. was found to be teratogenic at higher dose levels.175 A series of ·1730� F. Flick, D. Firestone, and J. P. Marliac, Poultry Sci., 44, 1214(1965). 1740. K. Courtney, D. Gaylor, M. D. Hogan, H. L. Falk, R. R. Bates, and I. Mitchell, Science, ill, 864(1970). 175K. o. Courtney and J. A. Moore, Toxicol. Appl. Pharmacol., .!2,, 396(1971). 176G. L. Sparschu, F. L. Dunn and V. K. Rowe, Toxicol. Appl. Pharmacol., !!, 317(1970). w. u• rw n,,-. 85 additional studies of the teratogenicity of the dioxins followed. THese are summarized in Table 12. One of the most . thorough investigations was carried out by Neubert and his colleagues.177 They noted ,that with an effective single dose ·1f 1-10 a,g/kg, TCDD is the most potent teratoge� known. In the mouse they found that TCDD was quite specific in its effect causing clef palate, kidney abnormalities, and atrophy of the thymus. The coadministration of TCDD was found to potentiate the response of other teratogens, including 2,4,5-T. The aiailarity of some of the toxic respo11ses caused by 1'CDD with those caused by known chemical carcinogens was pointed out by Buu-Hoi, !! !!,.178 They found that 'l'CDI> very greatly activated zoxazolamine hydroxylase, an aryl hydrocarbon hydroxylaae, and that it affected other enzyme systems in a manner similar to that of known carcinogens such as benzo(a] pyrene and p-dimethylamino-azobenzene. 'l'CDD, however, was much a>re potent,. so much .so as to be the most potent inducer of aryl hydrocarbon hydrolyase ever reported. Grieg,179 Greig and DeMatteis,164 Poland and 177 (a) D. Neubert and I. Dillman, Arch. Pharmacol. i fil, 243(1972). (b) D. Neubert, P. Zens, .Rothenwallner and H. J. Merker, Environ. Health Perspec., (5), 67(1973). 178N. P. Buu- Hoi, D.-P. Hien, G. Saint-Ruf, and J. Servoin­ . Sidione, c. R. Acad. Sci •. Paris, Ser. D., fil, 1447 (1971). 179J. B. Greig, Biochem. Pharmacol., !!:, 3196(1972). j I' TABLE 12. Teratogenicity of chlorinated dibenzo-p;.dioxins. Derivative Species 2-Chloro• 2,3-Dichloro• 2,7-Dichloro2,7-Dichloro1,2,3,4-Tetrachloro• 2,3,7,8-Tetrachloro• • • .. • " • • Hexachloro" Octachloro• Lowest dose with signiRoute ficant effect (l-19/kg) >2000: >20()0 .. >200056 > 1.0 > aoo: o.s a 0.2s Rat 0.125 a o.sa . Rat Mouse J.oa 1.oa Mouse 3.oa Mouse s.o b Chick embryo 0.02 b Chick 0.1-1.0° Rat 0.1a Chick 10c Rat >5 X 10 Sa · Chick >5 X 10 50 Teratogenic effect Rat Rat Rat Rat Rat Rat Oral Pre• and postnatal lethality Weight loss Oral Intestinal hemorrhage Kidney abnormality s.c. s.c. Cl�ft palate Kidney abnormality Oral Cleft palate Oral Cleft palate Air s ac Edema Edema ·oral. Delayed ossification of sternabrae Oral Edema Oral -Oral Oral Ref. 187 187 187 163 187 187 176 175 175 177 177 d 163 163 163 163 163 8 Daily dose, days 6-15 of gestation. bsingle dose. cDaily dose •. dJ. Verret, in Effects of 2,4,S-T on Man and the Environment. Hearings before the Subcommittee on Energy, Natural Resources and the Environment of the Committee on Commerce, United States Senate, 91st Congress, April 7 and 15, 1970, p.196� GI 87 Glover,1 8 O•1 8 1and Lucier,� !,!.182 substantially extended Buu-HoI's work using the known carcinogen and aryl hydrocarbon hydroxylese inducer 3-methylcholanthrene as a reference. They found that the response to TCDD paralleled that of 3-methyl­ cholanthrene in every way, except that TCDD was approximately 3 X 10 4 time� as potent an inducer. These effects are de­ scribed in greater detail below in the discussion of mechanisms and toxicity. Greig,� a1.,1 4 3band Gupta,� !!_.¼ 83 pointed out that in livers of rats treated with TCDD in addition to severe de ­ generative lesions, large multinucleated giant hepatocytes were present. The presence of these cells plus pleomorphism of cord cells, and increased numbers of mitotic figures led Gupta,!!:_!!_. to suggest that a long term study be carried out to determine whether neoplasms might eventually develop. back and Allen 1420 expressed a similar concern on the basis Nor­ of the marked hypertrophy of gastric mucosa that they observed in the. intestine of the monkey. lSOA. Poland and E. Glover, Environ. Health Perspec., (5), 245(1973). 181A. Poland and E. Glover, Molec. Pharmacol., 10, 349(1974). 182G. w. Lucier, o. s. McDaniel, G. E. R. Hook, B. Fowler, B. R. Sonawane, and E. Faeder, Environ. Health Perspec., (5), 199(1973). 183B. N. Gupta, J. G. Vos, J. A. Moore, J. G. Zinkl and B. c. Bullock, Environ. Health Perspec., (5) 125(1973). aa:.11 UlfT mrttr···1:;wrm 88 As is discussed below, the immune system appears to be The possible connec- the system most sensitive to TCDD. tion between decreased immune response and the appearance of malignancies, provides another reason to look for long term neQplastic effects with TCDD. A program is presently being carried out at the IIT Research Institute in Chic�go under the auspieces of the . . Ha.tional Cancer Institute to test the carcinogenicity of chlor­ inated dioxins administered dermally and orally 184 to rats and mice. With respect to mutagenicity, Hussain,� !!:.,185 reported that TCDD appeared to be mutagenic in bacteria possibly by intercalation with DNA. Thia latter possibility is sug- gested by the fact that the dioxins closely resemble the acridines, which are known to.intercalate with DNA. Other laboratories have been unable to reproduce these results, although all possible conditions have not been investigated.18 6 In a study of another class of mutagenic activity, Khera and Ruddick 187 conducted a·dominant lethal test with TCDD in rats. 184M. E. King, A. · M. Shefner, and R. R. Bates, Environ. Health Perspec., (5), 163-170(1973). 185 s. Hussain, L. Ehrenberg, G. Lofroth, and T. Gejvall, . �. !, 1(1972). 186s. Ames, personal communication. 187x. S. Khera and J. A. Ruddick, !!9_, 70(1973). Adv. Chem. Ser., 1/abntc··l:ii:,··· 89 Dose levels up to and including lethal levels were tested but reproductive values indicated that no dominant lethal mutations_ occurred during a 35-day post-treatment period, a period sufficient to include the postmeiot.ic stages of spermatogenesis. The incidence of pregnancies, however, was re- duced in the treated groups. f.. Mechani�s _of toxicity At the present the mechanism or mechanisms by which chlorinated dioxins produce toxicity are not known. It is known that they are toxic in very small doses and that an in4uction period o_f two to three weeks or more is needed to reacll the peak of the toxic response. Different mechanisms .-y predominate in different species. For example when a rat or rabbit dies of 'l'CDD poisoning the liver is so seriously damaged that this could be a major contributory cause of death.183 At the time a guinea pig or mouse dies, however, the effects on the liver are relatively mild.183 Effects on the immune system leading to secondary infections appear to be the only likely cause of death in the guinea pig. A wide variety of changes have been observed in organs, tissues, and cellular morphology following dioxin administration. The significance of most of these 'changes is not known at this time, but for the sake of completeness a brief over­ view of these results will be presented in the following 90 discussion. The gross effects on the liver are similar from species to species. although sensitivity varies· considerably. In the rat necrosis, multinucleation, numerous fatty vacuoules, and hypertrophy of parenchymal cells were observed.183•188 Similar l.esions were reported in monkeys,171 chickens,189 rabbits.163 dogs163 and mice.163 Ultrasturctural changes occur in the organization of the rough and smooth endoplasmic reticu- lum in hepatocytes. With chronic exposure to chlorinated dioxins the rough endoplasmic reticulum congregates in char­ acteristic concentric arrays around cell nuclei or fat vacu­ ol.es.142c,l7l Proliferation of the smooth endoplasmic reticu­ 171 •190 mitochondria are also observed.171 lum and enlarged Ufects on the immune system appear to be among the most �ral and most sensitive indications of dioxin toxicity. Allen and carstens171 repc-rted hypoplasia of lymphoid tissue in the spleen and lymph nodes of the monkeys fed toxic fat containing largely TCDD. the white blood cell count. There was a corresponding drop in Norback and Allen142c suggested 188N P. Buu-Hoi, P.-H.-Chanh;, G. Sesque, M. c. Azum-Gelade, . and G. Saint-Ruf, Naturwissenschaften, 59, 174(1972). 189J. R. Allen and L. A. Carstens, Lab. Invest., 15, 970(196 6 ). 190J. G. Vos, J. A. Moore, and J. G. Zinkl, Environ. Health Perspec., (5), 149(1973). 91 that in monkeys and chickens secondary infections as a result of decreased resistance were a major cause of death. Buu-eoi, !! ai. 188 described involution of the thymus in the rat follow­ ing ingestion of TCDD. The first effort to measure directly the effect of TCDD on immune response was made by Vos,!.!:_ !!,.191 They treated rats, mice, and guinea pigs with a wide range of weekly doses of TCDD. In the guinea pig total leucocyte and lymphocyte counts and thymus weights decreased, cell medi­ ated immunity (tuberculin response) was depressed, and at higher doses humeral immunity (tetanus toxin response) also was depressed. .In the rat, thymus and adrenal weights decreased, _but no effect on cell mediated or humeral immunity was detected using tuberculin and tetanus toxin response. Thymus and spleen weight decreased in the mouse and, as indicated by decreased response in a graft �rsus host assay in which spleen cells from treated animals were injected into untreated animals, cell mediated immunity was significantly depressed. At the levels at which these effects on the immune.system occurred, 0.008 �g/kg week for the guinea pig and 5.0µg/kg week for the mouse, other organ systems were unaffected. In contrast to the ex.tensive neurologic.il symptoms that have been observed in workers exposed to TCDD, few effects on the nervous system of animals have been reported. Edema189 and hemorrhaging183 have been observed in the brain of chicks 191J. G. Vos, J. A. Moore, and J. G. Zinkl, Environ. Health Perspec •• (5), 149(1973). 'r I_: &\·· 92 and rats, respectively, but probably are the result of more generalized effects. Bu��HoI, � ai.192 reported finding a five- to sixfold reduction in acetylcholinesterase activity in rats treated with a single intraperitoneal inj!ction of 10 mg/kg of TCDD. As indicated previously, generalized or-localized edema may be observed with dioxin poisoning. one of the most characteristic effects. In some species it is The symptoms of the,•chick edema• disease described earlier include hydro­ pericardium, ascites, edema of the skeletal muscles and pul­ monary edema.193 Similar .symptoms developed in Macaca mulatta monkeys fed a diet i�c-iuding chlorodioxins, mostly TCoo.171 Such effects are much less apparent in rats; mice and guinea 142c,l63•183 although Greig,� !!.:_,143b reported that pi gs, in the rat doses ten times greater than the w5O led to - accumulation of excess peritoneal fluid. The conditions leading to the production of the observed edema are not well under­ stood, but Norback and Allen 142c have proposed that the effects are due to increased permeability of capillaries and, at higher doses, to hepatic dysfunction leading to lowered serum albumin. This in turn is associated with d'ecreased osmolarity of the blood and subsequent extravasation of fluid from the blood 192N. P. Buu-Hoi, P.-H.-Chanh, G. Sesque, M. c. Azum-Gelade, and G. Saint-Ruf, Naturwissenschaften, 59, 173(1972 ). 193J. R. Allen, Amer. J. Vet. Res.,�, 1210(1964). 11. n :amemw > 93 vessels into the surrounding tissue. In their long term feeding study with monkeys Allen and Carstens171 observed serious anemia along with hypoplastic erythroid bone marrow that had been invaded with fatty tissue. In agreement with this Gupta,� !.!,. 183 reported necrosis of megakaryocytes in spleen and bone marrow of rats and guinea pigs treated with TCDD. In contrast, Zinkl,!,! !.!,.194 with rats of the same strain as Gupta,!,! al. saw no sign of anemia or depressed hematopoesis. The difference in results is un­ eq,1ained. Other toxic respQnses to TCDD such as teratogenicity 175 �177 and chloracne 36 •1 1 9• 1 43a• 1 45have been well described clinically, but no underlying mechanisms have been demonstrated. Lucier,� !.!,.182 suggested that the to�ic effects of l!CDD might in part result from disruption of steroid regulation • . Vos,� al.,191 however, observed no changes in the levels of CQrt.isol or corticosteron in the guinea pig even at highly toxic levels. 'l'CDD has been observed to affect cell division. In a potentially very important finding,. with dividing endosperm cells of the African blood lily Jackson 195 reported that mitosis 194 J. G. Zinkl, J. G. Vos, J. A. Moore, and B. N. Gupta, Environ. Health Perspec., (5), 111 (1 973). usw. T. Jackson, J. Cell Sci., 10, 15(1972). was severely impaired by TCDD at c oncentrations The formation of dic entric bridges and c hromatin of 0. 1 ppb. fusion with formation of multinuclei or a·single large nucleus was observed. Impairment of the mitotic apparatus was evidencedby •floating free• in the and Sunner phila. 196 c hromosomes The results of Davring_ c ytoplasm. suggest that a similar phenomenon occurs in Droso- Zn that they show a specific cellular effect at.very low concentrations, these observations may be important in ex- plaining the high toxic fty of TCDD. The spec ulation has been put forth that these effec ts might result from an interaction of TCDD with mic rotubules, the ultrasturctural elements which among other things, organize chromosome movements during mi­ tosis.196 Interference by TCDD with c ell division in animals is suggested by depressed hematopoesis, leucopenia, and sup� pression of spermatogenesis� 42 a, l7 l multinucleate c ells As mentioned earlier, are formed, espec ially in the liver. Zn this organ parenchymal cells with up to thirty nuclei have been obse�ed. 1 43b In some c ases TCDD seems to increase the Hyperplasia o� gastric mucosa,1 7 1 rate of c ell division. hepatoc ytes,17 1 • 1 83 and epithelial cells 142c (in c hlorac ne) has been reported. on the basis of the general emac iation prior to death of animals treated with TCDD, Lucier,� !!,. 1 82 investigated (5), 196R. Baughman and M. Meselson, Environ. Health Perspec., (1973). 27 • t 95 the possibility that bioenergetic pathways might be disrupted. They found that in rats there was no difference in oxidative phosphorylation in treated animals compared to controls. Cunningham and Williams 197 attempted to measure the effect of TCDD on protein and lipid metabolism in the rat liver. · In rats dosed with TCDD the rate of incorporation of 3a-acetate into lipid was unaffected, suggesting no effect on overall lipid synthesis. Total liver lipid increased, however, which might have reflected impaired degradation or transport. ·Although the r.esults were not very reproducible, there seemed to be an increase in the rate of incorporation of 14c-leucine into_protein. Lucier, et ai.182 observed an increase in total rat liver microsomal protein following treatment with TCDD. These effects might result from increased protein synthesis associated with induction of liver enzymes. Buu-Ho1, ·� !,!.178 were the first to- report :that TCDD is an extremely potent. enzyme inducer. By measuring the duration of zoxazolamine paralysis time they found that TCDD was a far stronger inducer of zoxazolamine hydroxylase, an aryl hydroxylase, than the carcinogen and powerful inducer. benzo[a]pyrene,. They noted that for both TCDD and benzo[a}- pyrene there was no further increase in enzyme activity with doses above a certain level and suggested that this might result from saturation of the 96 receptors. Their resu l ts were substantially confirmed and extended by Greig, 1 79 Greig and De Matteis, 1 64 Poland and Glover180 • 181 •198 and Lucier,!.!:_ !!_.1 82 TCDD was found to be over 30, 00 0 times as strong an aryl hydrocarbon hydroxylase inducer as the carcinogen 3-methylcholanthrene.164,lSO, l Sl,lS2, 198 In the chick embryo, a doubling of enzyme activity was· ,obtained with a dose of 1 .55 X 1 0- 12 mole per egg(O.Sng). 180 In every way the 'l'CDD activity seems to parallel that of 3-methylcholan­ thre_ne, although the activity persists longer, probably because of the longer �iological half life of'l'CDD (see be low). The mechanism of enzyme induction of aryl hydrocarbon hydroxyla�e by TCDD has been investigated in considerable c!etail in the chicken embryo and the rat and mouse by 165 In a 1 Poland a�d Glover,18 •199 and Poland, et ai. study of the structure to activity relationship, the enzyme inducing ability of several different halogenated dibenzo­ p-dioxins was measured. Depending on the extent and position of substitution, the activity varied greatly, but in every case this activity closely paralleled the acute toxicity for each compound. For significant toxicity to exist, 198A. Poland and E. Glover, Molec. Pharmaco l ., !, 736 (1973). 199A. P. Poland and E. Glover, manuscript in press. --- - ·- - --------------------------------97 at least three of the 2,3,7, and 8 positions had to be sub­ stituted and at least one hydrogen had to remain on the ring (Table 13 ). · 'l'he simultaneous administration of non-inducing dioxins did not enhance or inhibit the activity of inducing dioxins implying that the non-inducing dioxins are n_either agonists or antagonists for the inducing dioxins. With in- creasing doses, all dioxins which induced eventually led to. the same maximal aryl hydrocarbon hydroxylase response. This maximal response and the slope of the dose-reponse curves were indistinguishable from those obtained with 3-methylcholanthrene. This was taken as suggesting that TCDD and other dioxins.might interact with the same induction receptor as 3-methylcholan- threne, but with a higher binding affinity. �ditional insight was gained from the observation by Greig and De Matt:eis 164 and Poland!.! ai.165 that in certain "non-responsive" inbred strains of mice, which do not undergo aryl.hydrocarbon hy­ droxylase induction even with lethal levels of 3-methylcholan­ threne, high but still non-lethal doses of TCDD are able to produce enzyme induction. As the dose of TCDD was decreased, the enzyme response approached the same maximal level observed For a given response in •non-rein •responsive• mice.165 sponsive• animals a TCDD dose is required that is about ten times that giving the same response i� "responsive" animals.165 Poland and Glover 199 proposed that this effect may be caused by a faulty induction receptor which has a greatly decreased 98 TABLE-13. Structure-activity relationship of substituted dibenzo-p -di�xins for induction of argyl hydrocarbon hydroxylase activity in the chick embryo liver (ref. 32b and 198). Derivative a� Chlorinated 2,3,7,8-Tetrachloro1,2,3,4,7,8-Hexachloro1,3,7,8-Tetrachloro2,3,7-Trichloro­ l,3-Dichloro2,3-Dichloro2, 3-Dichloro2,7-Dichloro2,8-Dichloro1,2,4-Trichloro1,2,3,4-Tetrachloro­ Octachloro- b. Activity/TCDD activity 1.0 0.8 0.2 0.02 0 0 0 0 0 0 0 0 Other derivatives 2,3-Dichloro-7,8-dibromo2,3,7,8-Tetrabromo2,J,7-Tribromo2,3-Difluoro-7,8-dibromo2-Nitro-7,8-dibromo2,3-Dichloro-7,a�dimethyl2,3-Dichloro-7,8-difluoro2,3,7,8-Tetranitro2,3-Dibromo2,3-Difluoro2,7-Diacetamido- 1.1 1.0 0.6 0.2 0.01 0.01 0.01 0 0 0 0 1w·:4M. L. Schafer and J.E. Campbell, Adv. Chem. Ser., !2., 89 (1966) � 205t. _J. Casarett, G. c. Fryer, W. L. Yauger, and H. W. Klemm.er, Arch. Environ. Health, 17, 306(1968). 206M. .· De Vlieger, J. Robinson, M. K. Baldwin, A •. N. Crabtree, and M. c. Dijk, Arch. Environ. Health, 17, 759(1968). 207P. J. Biros and A. c. Walker, J. Agr. Food Chem., ll,, 425(1970). 2080 P. Morgan and c. C. Roan, Arch. Environ. Health, . �- 452 (1970). 103 of DDE and DDT, this difference in adipose to liver distribution, if it is at equilibrium, suggests that some specific, relative­ ly high affinity binding for chlorCldioxins exists in the liver. Indirect support for this view comes from the experiment of Piper,!.!!.!• in which male rats were fed a single dose of 'l'CDD equal to about twice the LD dose. In this case the 50 liver to fat concentration ratio approached one to one,209 which may reflect saturation of specific binding sites and a progression toward a simple lipophilic partitioning. Of particular interest in this regard are Poland and Glover•s 181 results which suggest high affinity binding between TCDD and hydrocarbon hydroxylase •induction receptor.• TCDD apparently is metabolized very little or not at all. Vinopal and Casida 210 treated mice with 130 i,tg/kg of TCDD- 3& in a single intraperitoneal dose and from one to twenty days periodically measured the distribution of radio­ activity in nuclear, mitachondrial, and. microsomal fractions ·isolated from the liver. They reported that a large propor- tion of the label was present in the microsomal fraction. This was taken to indicate selective binding in this fraction. By means of thin layer chromatography and electron capture gas-liquid chromatography all of the radioactivity extracted 209w. N. Piper, J. Q. Rose and P. J. Gehring, Adv. Chem. !!!:,:., 120, 85( 1 973). 210J. H. Vinopal and J. E. Casida, Arch. Environ. Contam. Toxicol., !, 122(1973)� 104 and that collected in feces was shown to be present as un­ transformed Tcoo- 3 u. Gas-liquid chromatography also was used to demonstrate that no trichlorodibenzo-p-dioxin was -formed. About one percent of the label in the liver remained unextracted. This residue may have represented a covalently bound derivative of TCDD or may simply have resulted from incomplete extraction. An,!!! vitro experiment was carried out in which 3 TCD0- H was incubated with liver microsomes and NADPH. eight percent of the Tcoo- 3 a was recovered unchanged. Ninety­ When unsubstituted dibenzo-p-dioxin- H was similarly injected into 3 mice, a large proportion was metabolized to polar compounds. In the in� experiment 72-981 of unsubstituted dibenzo­ p-dioxin-3B was metabolized. Vinopal and Casida's report of localization of TCDD in liver microsomes is supported by Horback and Allen'• observation thatin rats fed octachloro­ dibenzo-p-dioxin- 3 6c1 ninety-five percent of the label present in the liver was located in the microsomal fraction. The in- ability of mice to metabolize TCDD brings to mind the biode­ gradation experiments described earlier which showed that soil microorganisms were similarly unsuccessful at metabolizing TCDD. If TCDD is not metabolized or is metabolized only to a small extent, the question of how it interacts in a molecular sense with living systems becomes even more intriguing. If the toxic effects in fact are due to a metabolite, then this metabolite must clearly have an even more extraordinary toxicity ---�---·· -·------· --- ____ .__, 105 than TCDD itself. Poland and Glover 181 have calculated that at the ED 0 for aryl hydrocarbon hydroxylase induction in 5 the rat liver there are about 2.5 X 10 4 TCDD molecules per cell. If the biologically active molecule is a metabolite of 'l'CDD, formed to the extent of. one percent, then there would be only 2S0 active molecules per cell. The alternatives are that TCDD acts as the neutral molecule, or, that it in some way acts catalytically, with regeneration of the original mole­ cule. As mentioned earlier, in an effort to discover the biologically significant chemical characteristics of 'l'CDD, Poland and Glover,198 carried out a structure-activity study with several different halogenated dioxins. To repeat their results, they found that activity required substitution at three of the 2,3,7- and 8- positions and at least one remain._ lng hydrogen • . Kende and Wade,lla, Jlb using Poland and Glover's chick embryo aryl hydrocarbon hydroxylase bioassay, extended this work, testing derivatives containing bromine, flourine, and methyl substitutents. Their results are summarized in Table Steric effects The order of reactivity is Br>Cl>F>CH3• appear not to be very important, since bromine and methyl 13b. have similar steric properties, although methyl should be much more easily metabolized than the aromatic halogens. Kende and Wade discussed mechanisms by which TCDD might be .."!'�.'1-.+}X 4--- ?.*>se: 106 transformed into highly reactive metabolites. As summarized in Figure 5, these include formation of a benzyne 211 and two different epoxides. nisms has been tested. At the present time none of these mecha­ One approach might be to prepare in­ termediates, such as the quinone, or the corresponding catechol, in mechanism c.32a 211A Streitwieser, personal communication, quoted by Kende . and Wade (Ref. 32a). H+' Cl .A J O :o: Cl � 0 · Cl Cl ?' n :o: . � .1 . -+ . �p, Cl Cl � C ;o: Cl . "..:::: � O O · � � :o c1 1 Cl ·cl 1b: O �. · O . O Ns C I >xx 4' .. Cl Cl ��· I �1,1 Scheme·B 1 � n 0 O�Cl � ?' �· Scheme A c1y<}o��·\.t!• Cl�O�Cl Cl I J · - 0 0 f -... :cc . 4' ?. .N Cl OB . 0Yr��o� 1 O�ClCl�O�l � � Cl O 1 O ·1 � � � � '-'- · o n "..:::: O · · A . O I Cl :(X � Cl Scheme C Figure 5. Possible chemical mechaniems for the toxic action of TCDD. Scheme Aa Formation of a benzyne (ref. 211). Schemes Band c, Formation of an arene oxide (ref, 32a). In each case an intermediate is postulated which reacts irreversibly with cell nucleophiles. 108 3. Presence in the Environment a. Possible environmental sources of chlorodioxins Since chlorodioxins are not formed naturally, they can be introduced into the environment only in'connection with the distribution of some man made chemical product. Chlorodioxins could be present as.already formed contaminants in a chemical product, or they could be formed via the degra- dation of a chemical product. Or, both routes might be im- portant. As was established by Schultz !l a1.142a and later confirmed by Higginbotham, !lal.,24a chlorodioxins are formed from chlorophenols under conditions which are used to prepare some chlorophenols, such as 2,4,S-trichlorophenol and pent'achlorophenol, by hydrolysis of chlorobenzene pre- cursors. This means that in these chlorophenol products and their derivatives some level of chlorodioxins always will be .present. Any use of such products necessarily in- volves the release of chlorodioxins. Even if chlorophenols and their derivatives could be produced with very low levels of chlorodioxins, the phenols themselves, if exposed later to alkali and heat, always can provide a second source of dioxins. Such reactions, how- ever, require the bimolecular condensation of two chlorophenoxy groups, which means that the liklihood of reaction will drop rapidly as the concentration of precursor is diluted in the 109 environment. Another possibly much more significant source of chlor®ioxins is the monomolecular cyclization of chloro­ phenoxyphenols or their derivatives (Synthetic Route Din the section on Synthesis). The�e compounds were termed "predioxins• by investigators,, 9 •100 who encountered them as an interference in the analysis of c;>ctachlorodioxin in pentachlorophenol. 'l'hey are the intermediate compound in the synthesis of dioxins from chlorophenois ana, under the righ-t conditions .can be converted to dioxins in relatively high yield. Like- the dioxins they are formed during the synthesis of chlorophenols. Since the reaction involved is monomolecular, their potential to form dioxins is unaffected by dilution in the environment. It is possible that other higher oligomers also might decompose to give dioxins. 'l'he chloracne and chick edema incidents and the detection of TCDD in Vietnamese and Thal samples (as discussed below) demonstrate that, whatever the route, chlorodioxins have been introduced into certain areas of the environment. In the case of chloracne the dioxins involved presumably were formed during the synthesis of chlorophenols and their derivatives. case_ of chick edema, Metcalfe 212 and Fireatone 213 In the have proposed that at least one chlorodioxin source was chlorophenols used to preserve animal hides from which tallow was extracted. •toxic fats• were derived from the tallow. 'l'he It has not been established whether they were formed during the high temperature 2 12 L.D. Metcalfe, J. Assoc. Off. Anal. Chem., 56, 542(1972). 213n. Firestone, Environ. Health Perspec. (S),�9(1973). 110 In a group of eight commercial penta­ chlorophenols Firestone, et at.98 found levels of hexachlorodioxin render.tng process. r�ging from 0. 17 to 38 ppm with an average of samples of 2 ,4,5-trichlorophenol to 6."2 ppm with an average of 17 ppm. In six TCDD levels from less than 0.1 1.3 ppm were observed. In the Vietnamese samples, levels of TCDD in various areas appear to be correlated with the extent of use of agent Orange (a 1:1 mixture of the n-butyl esters of 2.4-D) in a particular·area. 2 ,4,S�T and It is not possible to say whether the observed TCDD was that which was originally present in the agent Orange or whether it was formed in the environment. Later shipments of Orange to Vietnam are reported to have contained from less than 0.1 to 47 _ppm of TCDD with a mean of 2-3 ppm. 214 Analyses are not available for the earlier shipments. Residues of TCDD were found in 1973 in soil from areas of Thailand treated with 2,4,5-T during 1964-65. Assuming that no degradation of TCDD occurred, the levels found predict an original level <3 to 50 ppm in the 2,4,5-T. Since TCDD levels rarely were reported to be higher than this in the 2,4,5-T, the results suggest that TCDD may have a half-life for disappearance in soil on the order of several years. 2 14 Crummet and Steh1 215 have reported that with carefully con­ trolled conditions 2,4,5-trichlorophenol and duced with less than 0.1 ppm of TCDD. 2 ,4,5-T can be pro- The level of phenoxyphenols in these products was not reported. 214National Academy of Sciences Committee on the Effects of Herbicides in Vietna.� - Part A," National Academy of Sciences, Washington, o.c. (1974), Section VII. 215W. B. Crummet and R. H. Stehl, Environ. Health Perspec. (5), 15 (197 3). 111 In summary, there _are three general ways in which chloro­ dioxins could be introduced int.o the environment: 1) they may be present as contaminants in chlorophenols or their derivatives, 2) they may be formed from chlorophenols or their monomeric chlorophenoxy derivatives under conditions of use or storage, or 3) they may be formed from phenoxyphenols or other polymeric derivatives under similar conditions. C At the present time, in those cases where TDD residues have been found in environmental samples, it is not known which of these routes was responsible. b. Bioaccumulation The general physical properties of TCDD and other chlori­ nated dioxins are similar to those of DDT and related compounds, for example they are more soluble in lipid than in water by many­ orders of magnitude, and so it is possible that chlorinated di­ oxins WQuld accumulate in food chains in a manner similar to DDT. Matsumura and Benezet6 3b and Isensee 216 attempted to test bioaccumulation of TCDD in model aquatic ecosystems. Under their conditions, which especially for the highest trophic level, fish, were not at equilibrium, TCDD·appeared to bioaccumulate, but to an extent about one tenth that of DDT. The finding of TCDD residues in Vietnamese and other en­ vironmental samples (described later) suggests that under field conditions bioaccumulation of TCDD may occur. 216A. Isensee, cited in R. w. Fullerton, M. B. Carlson, and A. R. Nolting, "Report on 2,4,5-T workshop," Department of Agri­ culture, Office of the General Counsel, Washington, D.C., 8-9 March 1974, p.14. 112 s. Review of Analytical Methods Requirements for Analysis 1. Any analytical method for detecting TCDD in environ­ mental samples should have a sensitivity on the order of part in 10 12 or 1 part per trillion (ppt). for this is as follows. 1 The rationale In the guinea pig, the most sus- ceptible species of those that have been tested and therefore a good choice· for establishing desirable limits of sensitivity; the lethal single oral dose in males for fifty percent of the test animals (Lo 0) is 0.6 µg/kg body weight.163 This 5 means tha.t at lethal doses, if all of the TCDD were retained, the level of TCDD in the whole animal would be.less than 1 part per billion (ppb) O;hat is less than 1 µg/kg). As dis- cussed above, with doses of about 0.01 µg/kg (per week) in the guinea p!.g toxic effects·on the immune system still are observed. This is about a factor of 100 below the LD50• Zf a further factor of 1 0 is provided as a safety factor and to allow for species even more sensitive than the guinea pig, the desired level of sensitivity for environmental mon­ itoring can be calculated to be (10-9) (10-2 ) (10-l) = 10 12 or 1 ppt. This means that for a tection for TCDD �f 1 X 10 is required. 12 1 -g sample a limit of de- g or 1 picogram (pg) or better At such low levels there are not many con- firmation procedures which can be used, and so another desirable characteristic of any analytical method for TCDD is a high 113 degree of specificity. As discussed earlier, other chlorinat.ed dioxins appear to be less toxic than TCDD, but it would be reassuring, as well as convenient, if the method selected for TCDD was suitable for measuring other dioxins as well. Methods other than Mass Spectrometry 2. Probably the most common method for measuring low·levels of organic compounds is gas-liquid chromatography (glc). For chlorinated compounds the most sensitive de- ·tectors, in order of increasing sensitivity, are flame ioni­ . zation (fid), thermionic and electron capture (ec). The ec detector will respond to on the order of l pg of the most favorable chlorinated �ubstrates.217 The limit of d�tection with ec for other chlorinated compounds is higher, but, !. priori, the ec system would deserve to be investigated for . TCOD analysis. In fact, as is discussed in the next section, the limit of detection for TCDD with ec/;9lc is in the range of 50-500 pg! it follows that the less sensitive , fid and.thermionic detectors are of no help. Glc however, does have a substantial degree of spec­ ificity for isomers and homologs in the chlorinated dioxin series. For example with standa.rd analytical columns each group of isomers can readily be separated from other groups of isomers differing from it by one or more chlorines and 114 many isomers actually can be clearly separated (Figure 6).98 Thus glc has considerable potential in terms of specificity if it is usedpreparatively_or is interfaced with some other more powerful detector such as a mass spectromater (see below). For neutral dioxins, which have molar extinction co­ efficients of about 2000-6000 at£!.• 305 nm (Table 4), the limit of detection with uv spectroscopy is on the order of 1 µg or more.26 Since many other aromatic or conjugated mole­ cules absorb at 305 nm or at the other dioxin maximum at£!.• 250 nm, this method is not very specific. The situation is a little better with the dioxin cation radicals (in tri­ fluoromethanesulfonic acid) where the limit of detection is about 100 ng at the maximum at 700-850 nm.$8 This maximum is fairly specific in that none of the other aromatic systems that have been tested, including biphenyls, diphenyl ethers, dibenzofurans, xanthenes and xanthones formed cation ·radicals which means they had no absorption in the visible spectrum. As can be seen in Table 6, cation radica_ls of individual dioxin isomers in some cases can be differentiated on the basis of their visible absorption. Due to its low sensitivity, however, this method is useful only for confirmation of very high levels of dioxins. The dioxin cation radicals can be detected with much greater sensitivity in an esr spectrometer, where the limit R :II B A ...... UI F Fi ure 6. Separation of lower chlorinated dioxins by ec/glc: (R) Aldrin reference, (Af 2-chloro-; (B) 2,7-dichloro-; (C) l,3,6- or 1,3,8-trichloro-; (D) 2,3,7-tri­ chloro-; (E) 1,3,6,8-tetrachloro-; and (F) 2,3,7,8-tetrachlorodibenzo-p-dioxin. Conditions: 2.5\ SE-52 on 60-80 mesh Gas Chrom OJ 5' x 0.25•, column 200 ° c, 50 ml/min N2 carrier g·as, sensitivity lxlo-9 amps full scale (personal communication, o. Firestone). · ¥¥ 116 of detection is a few nanograms.88 Unfortunately this is still about a thousand times above the desired sensitivity. Esr is about as specfic as uv for the dioxin cation radicals. As is indicated in Table 8, many pure isomers and homologs can be differented on the basis of.g factors or line widths. However, mixtures cannot be resolved very well. Esr might be useful for confirming high environmental levels of individual dioxins. If the limit of detection could be lowered, its importance would be much increased. Infrared spectroscopy, with a limit of detection of a few micrograms is much too insensitive for the present purpose.77 With pure samples individual isomers can be dif­ ferentiated with ir (Table 3), but with mixtures or in the presence of other compounds that may be isolated along with the dioxins from environmental samples this method is not likely to be dependable. Similarly, nmr is far too insensitive. Even with Fourier ana.lysis samples of· many micrograms would be required. Almost as great a disadvantage is the lack of spe.cificity. file dioxins give only aromatic signals in the region & 6.8 - 7.2 (Table 7), signals which easily could be masked by many other aromatic compounds. Low temperature phosphorescence emission spectroscopy has a timit of detection of a few nanograms for dioxins.218a 218G� Yang, personal communication. 117 'l'he emission spectra and phosphorescence decay times are fairly specific for different dioxin isomers (Table 5). This method also might be useful for confirming relatively high en­ vironmental levels of individual dioxins. The high degree of b iological activity of TCDD suggests the possibility that sensitive bioassays for TCDD.might be developed. An enzymatic bioassay based on induction of aryl hy- drocarbon hydroylase218b and a radio-immunoassay 218c .are present­ ly being.investigated. An ear bioassay based on the displacement of spin labelled TCDD ha� also been considered. 2 18b 3. Mass Spectrometry Mass spectrometry is a method which has a sensitivity in the range required for analysis of TCDD. Both photographic plate and electron multiplier detectors are available which are capable of responding to very small ion currents in the mass spectro�ter. 2 19 The sensitivity for a particular compound depends on its ionization efficiency and on the composition and structure of the ion formed, but measurements well below the nanogram (picomole) range are often possible. Another advan- tage of a mass spectrometer is that the mass of .the ion measured is known. This provides specificity by restricting the possi- bilities for the molecular composition of the ion. Fragmenta- tion ions and metastable ions, produced by decomposition of the initially formed molecular ion, give additional information 218(b) o. Firestone, personal communication. (c) C. Collier, personal communication. 219w. McFadden, "Techniques of combined gas chromatography/ mass spectrometry," John Wiley & Sons, New York (1973), pp.60-70. \<· ��\fr "",;c9§k4;1#�4,� £ ,§��'5¥1}¥1j,i¼i@3fo##?:•f1*;1�'�4#�,q:��-f�f�?"::;-;i�,8!¥-rFZ;¥4J�?t'ifk\ :�"Mft��¥-i¥¥'*'v-·:•--·· r:nw 118 Measurement of the ratios oY .stable isotopic isomers, such as those derived from 12c- 13c and 35c1-37c1, provides additional information �bout the mole­ about mo.lecuiar structure. cular formula. Certain mass spectrometers, so called •double focussing" spectrometers because they use both magnetic and electrical fo. cussing, can achieve an m/e (mass to· charge ratio, where the charge is normally +l) resolution of 100 ppm or better. With these instruments a.new order of specificity is achieved in that.the molecular formula usually can J>e reduced to a few possibilities. This is because the exact mass for different elements deviates from unit values depending on the nuclear packing fraction for that element. Only a few possible elemental coillhinations will give a particular measured molecular exact mass. One weakness of exact mass measurements, however, is that they do not permit differentiation between.isomeric ions. Such ions might result from the presence of molecular isomers or from fragmentation of other higher molecular weight ions. The ef- feet.of such interferences often can be reduced by the introduc- · · tion of some chromatographic step, such as glc, prior to mass spectrometric analysis. This may be done with a separate preparative glc step, or with a glc directly interfaced with a mass spectrometer. Combined glc-low resolution mass spec- trometry which lacks the specificity of high resolution analysis., e . ·. i :;;;;..,·".. ¼: - .. t·".. t:W .. >w ... ··,... -&... r.. i;,,.,.·,a..,··J:t .. )rrt.. :1:t... Gtiiol"·'··-M-M# ..·w .....}... CM ....... ............. --� c__: __'i!!_:::�:w.a:.?:.?s::r:-· �.-x��ef:.+_.���*'-�½�t.J ·:@_?)§.i'�4zi(;- :;.tJ;· n""v ''t't·,- .. _.1(qi_�¥l-,1J¢{,..�· .. · ....... >- ..., -· .,. ... 119 is now well developed and can be used on a routine basis. Combined glc-high resolution mass spectrometry, however, is The method de­ still in the early stages of development. 2 1 9 scribed here involves high resolution mass spectrometry, with the signal enhanced by multichannel _averaging, with · and without _ a separate preparative glc step. c. Aims of the Present Investigation The aims of the present investigation were 1) to develop a mass spectrometric method suitable for measuring levels of highly toxic TCDD in biological samples at a sensitivity of one part in 10 12 and 2) to use this method to determine whether TCDD was present in environmental samples from South Vietnam which had been heavily exposed to the herbicide 2,4,5-T, a known source of TCDD. It was hoped that the present work might pro- vide a useful groundwork for further studies of the environmental hazard of 'l'CDD and other chlorinated dibenzo-p-dioxins, especially if the results from Vietnam were positive, and that it might provide a highly sensitive mass spectrometric technique of general utility. _-, ... J..''li?fi4,, ,, z 120 PRELIMIHARY EXPERIMENTS A. Gas-Liquid Chromatography The most common method for analyzing residues of chlorinated organic compounds is gas-liquid chromatography (glc) with an ele.ctron capture (ec) detector. This method has been used to measure TCDD levels in tissue, but the limit of sensitivity was reported to be on the order of 50 ppb.220 This limit is a factor of 50,000 higher than that required for adequate environmental monitoring of 'l'CDD (see previous sec­ tion). 'l'o confirm the limit of detection for this method, glc analysis of TCDD was explored at the .beginning of the present investigation. The limit of detection for flame ionization detection was on the order of l ng and for ec about 50 pg, both of which are .well above the desired limit of l pg. Glc analysis of TCDD is aiso susceptible to interference from other residues with similar chromatographic retention times. Zn fact the quoted 50 ppb limit of sensitivity in tissue, with ec/glc largely was determined by such interferences. With the available instrumentation glc clearly was inadequate for part per trillion analysis of TCDD in tissue samples. B. High Resolution Mass Spectrometry with a Photoplate Detector As indicated in the review of analytical methods in the Zntroduction, mass spectrometry is particularly attractive for 120E. A. Woolson, P. D. J. Ensor, w. L. Reichel and A. L. Young, Adv. Chem. Ser., 120, 112(1973). «... UH_i;At. . ·-.. iJi:,!)( .. 9,i.¥.�M�l .. ¥$'$ �-ff..€ -Fil :;ff·,.8Pf,.L 14¢ st . X tl£.aj ..£.,..,.;w.-31g.-&i;S,f'l1$.I/Ql¼4,;tW..$.¥.U-SS4 121 analyses requiring high sensitivity and specificity. The first approach investigated in the present study employed a high resolu.tion (double focusing) mass spectrometer with photop;ate detection. By use of Mattauch-Herzog geometry 219 ·the fields of the electrostatic and magnetic analyzing sectors are constructed. so that ions at all masses simultaneously are focused in. the plane._o.f_a_ silver bromide phot<>l:"late. The ad- vantage of this method is that intensities for all ions are simultaneously integrated on the photoplate which then provides a permanent record of the entire spectrum. With the most sensitive means of plate development the limit of detection for TCDD at resolution 5000 was about 100 pg. Quantitation was made difficult by the nonlinear response curve near the limit of detection and by the rather limited range, a factor of about twenty-five, between the limit of detection and sautration of the photoplate. For these reasons mass spectrometry with an electron multiplier detector was in­ vestigated in the hope that this alternative approach might be more suitable. c. High Resolution Mass Spectrometry with an Electron Multiplier Detector The electron multiplier differs from the photoplate in that it has a much wider linear response range and that it can measure ion intensity at only one m/e value at any given ti�e. For a stationary measurement at a particular value an electron multiplier m/e 123 C DEVELOPMENT OF ANALYTICAL PRO EDURE A. Mass Spectrometric Detection of TCDD at the Picogram Level 1. Repetitive Scanning of a 20 m/e Unit Range One approach to high sensitivity mass spectrometric analysis is to scan repeatedly some restricted m/e region known to contain characteristic ions for .the compound being studied. Because it contains four chlorine atoms, TCDD has a highly ,characteristic set of isotopic isomers associated with its molecular ion. (The complete mass spectrum of TCDD in the Chlorine consists of two naturally occurring stable isotopes, 35c1 and 37c1 range m/e 160-328 is shown in Figure 7A.) which have abundances of 751 and 251 respectively. TCDD therefore has five major isotopic isomers at m/e 319.897,. 321.894, 323.891, 325.888 and 327.885 present in the ratio of 77:100:49: 10:l. There of course are other minor isotopic isomers due 13 2 to a, c, 170, and 180, but these are negligible in the present If the case. region of the TCDD molecular ion is scanned and signals at the above m/e values and intensity rati.o are observed,· this is a strong indication of the presence of TCDD or a structural isomer� Fortunately due to.its· nuclear packing fraction, chlorine also is mass deficient by 31 mmu. This combined with the fact that 'l'CDD contains only four hydrogen atoms (mass heavy by 8mmu} means that TCDD can easily be resolved from most organic residues * The same isotopic isomer pattern also is observed with the photoplate method described earlier, although it is more difficult to measure the ratio of ion intensities with this method. ;- 122 is considerably more sensitive than a photoplate, the limit of detection being less than 10 ions compared with 10 3 or However, .�ith rapid scanning. over more for a photoplate. 2 19 wid.e m/e ranges the sensitivity of the electron multiplier decreases because fewer ions are collected at a�y given m/e value. Since only one m/e value is measured at a time, ion optics of the Nier-Johnson 219 geometry in which only a narrow m/e range is in focus can be used. The electron mul- tiplier·in this case is positioned at the center of focus. The mass spectrometric response for TCDD with an electron multiplier detector was determined under a variety of con­ ditions. As is described in the following section, procedures were found which provided the desired sensitivity for TCDD. ., t@I ts 124 100 80 A. cu,•- 60 40 >!: 20 l: 0 100 -� fi (M•COCI) + (M)I+ l l + (M-2COCI) j B. J 60 40 20 0 ...........__..,....._...,......,...__-y--,--;,....,.�....,.--,---r....,...lll!Mt--6 250 200 H50 m/e. FIGURE 7. Mass spectra of (A) TCDD and (B) 37cl-labeled TCDD. The isotopic purity of the 37ci is 95.5%. The asterisk denotes an impurity. The multiplicity of lines associated with each major molecular species results from the presence of various isotopes of Cl. and c. .. .. -� .,... " ..... �·· . 125 since at m/e 320 they will be mass heavy by 100-300 mmu. (Mass heavy and mass light refer to deviations fron integral masses.) Figure 8 illustrates a scan in the region of the TCDD molecular ion with and without TCDD, The limit of de­ tection for this procedure was on the order of 100 pg. 2. Repetitive Scanning of a 0.300 m/e Unit Range In order to increase the number of ions derived sdely from signals of interest reaching the electron multiplier, scanning was restricted to two narrow m/e regions (each about 0.3 amu or 300 millimass units (mmu) wide) and carried out at resolution 10,000 (minimum between peaks equals 101 of This w�s achieved with the peak switching cir­ peak height). cuit. of the mass spectrometer (see Experimental Section). The two regions selected included a perflurotributylamine (PFA) re ference peak at m/e 313.984 and the TCDD peak at 321.894. 'l'WO scans alternately were made at 314, two at 322, · two at 314, etc. The PFA was bled in from an external reservoir at a constant rate, providing reference peaks that remained at the same height (ion intensity) throughout the analysis. The sample peaks on the other hand rose and fell as the sample volatilized. As is discussed at some length later regarding confirmation of TCDD, the sample peaks form an envelope that is characteristic for the compound being analyzed. The procedure 314 A. Pl"A etaridard B. Pl"A standard plus I I I '1'CJ)J) I 320 322 324 326 328 FIGURE a. Repetitive scanning of a 20 m/e unit range. (m/e 310.-330). r t1 a 127 is illustrated in Figure 9. The limit of detection for TCDD, although lowered to about 3. 20 pg, was still not adequate. Time Averaged Repetitive Scanning of a 0.300 m,k. Unit Range Greater sensitivity can be obtained from the re­ peated narrow scans shown in Figure 9 by combining them into a single time averaged scan. Procedures accanplishing this under low resolution ms conditions had been reported earlier although this was not known at the time that the present averaging system was interfaced with the MS-9. 2 21• 222 A method for averaging the individual scans was devised using a Varian 1024 averaging computer (multichannel analyzer) in conjunction • with the MS-9 mass spectrometer. The resulting increase in sensitivity is illustrated in Figure 10. The signal for a pair of peaks at the limit of detection for a single scan is shown in Figure lOA, and the averaged signal fro:n sixty scans is shown in Figure lOB • 'l'he signal-1D-noise ratio {S/:i) is. . expected to improve approximately as the square root of the 223 With one minute of scanning at one scan number of scans. per second, the observed improvement is approximately that ex­ pected. A variety of experiments were carried out to optimize the conditions for time averaging. These are described in detail in the Experimental Section and can be summarized as 221F. J. Biros, Anal. Chem., Q, 537(1970). 22 2J. R. Plattner ands. P. Markey, Org. Mass Spec.,!, 463 (1971). 223R. R. Ernst, Rev. Sci. Instr., 36, 1689(1965). A. Regions scanned B. Alternating narrow ecane (two half second ecane at 314, two at 322, et�.) (200 pg TCDD) LJ.LJ LJLJLJ 314 322314 322 314 FIGURE 9. Repetitive scanning of a 0.300 m/e unit range .I , ' , ... '.' ... N \0 315.040 314.960 m/• 314.960 llS.040 m/e FIGURE 10. Improvement in sensitivity with time averaging. PFA and a reference peak at m/e 315. Left: one scan. Right: 60 scans (one scan/sec). 130 follows. At scan rates faster than �our scans per second data was inefficiently transferred to the memory of the analyzer and resolution was decreased by damping caused by the time constant of the MS-� circuitry. The fastest suitable scan rate, in this case four scans per second, was used to maximize the S/N improvement. With very short sample volatilization times ( < 10 sec) sensitivity was decreased because fewer scans were made and perhaps in part due to decreased ionization efficiency in the ion source. With volatilizationtimes greater than about 60 sec the drift in peak position over the total course of the analysis was large enough to begin to decrease the resolution observed in the time averaged spectrum. optimum volatilization time was approximately 45 sec. The The interfacing of the analyzer with the MS-9 is illus­ trated in Figure 11. 'l.'he ions in the m/e region of interest, af.ter being focused, pass by a small magnet coil which deflects the bei$11l back and forth over the detector slit. After passing through the slit, the ions strike an electron multiplier, pro­ ducing a signal which is continuously displayed on an oscilloscope on the MS-9. scan.· · memory. This provides a mecU1S of monitoring each Simultaneously, the signal is added to the 1024-channel An oscilloscope on the analyzer continuously displays the total memory content which makes it possible to monitor the o�erall course of the analysis. A potential problem of phasing the beam deflection coil. of the MS-9 with the memory sweep circuit a ... a .AEI MS�9 Double Pocussing Mass Spectrometer ·---- Beam Deflection Coll Electron Multiplier Oscilloscope Source ...... w Memory Circuits ,FIGURE 11. CAT - MS-9 interfacing. XY Recorder .oscUloac:ope Varian 1024 �T &t_.• , ¥0%-•M? ? . . 132 of the analyzer is avoided by using the sweep voltage ramp of the analyzer, via an amplifier and appropriate circuits of the MS-9, to drive the beam deflection coil. The coil is thus necessarily in synchrony with the analyzer. The procedure used for introducing samples into the MS-9 .is shown in Figure 12. It provided reproducible analyses at a high level of sensitivity. The sample tubes were made from one mm id melting point capillaries as shown in Figure 13. A 10-µl syringe was used to introduce residue., ini:.o the sample tube. a 3-4 µl portion of the With a small flame the sample tube was drawn out just above the level of the liquid to produce a capillary constriction about 20 mm long. was then removed at reduced pressure • . by the capillary constriction. with a flame. The solvent· Bumping was prevented The sample tube was then sealed At the,time of analysis the capillary was broken off 2-3 mm above the constriction to give the tube configuration shown in Figure 12. The tubes were introduced into the MS-9 with a wire holder on the tip of a standard MS-9 direct insertion probe. To aid reproducibility all analyses were started at the same time after insertion of the sample tube into the MS-9 source. The temperature o.f the source heating block was adjusted to give a sample volatilization time of approximately 45 sec •. "'· sOW"ce heoting block -· __.,_,1 k · Probe shaft I .�� -� @I · · � >= .I >= Sample tube PIGURB 12. Sample introduction · -.Pf Ctr'�· ,l .. ,fl4$,¢r1Q.48@.§J;t@j.J4ij§ij@;4i¼}@M(PJJ!\ZiJl#fJl!i#J:¥}@£ )4(.@Q)¢fl(; ,fo,.,i44iJ({, :dJfi( ,;4;� f4 8P,.JQf½¢!kl)½¾i?JfoWJ.&f .¢gi(J_.J(R,#£$4.UM $£1!£ UAW Ionizing beam ... w w 1 ·cm . . -V ?at A. I WU kiln?: 134 0 CD >7 . Sample tube preparation I ZEE Experimental Section). 135 i'he analysis of a 0.5 g sample of 'l'CDD is illustrated in Figure 14. An internal standard is provided by a PFA fragmentation peak which is a known distance, 85 mmu, from the TCDD peak • . The limit of detection for TCDD (2.5 x noise) was on the order of 0.5 pg. B. Quantitation Procedures 1. Direc� Measurement One way of qu.-ntitating a 'l'CDD signal in the mass spectrometer is simply to compare its magnitude to that of a knc;,wn amount of authentic compound analyzed independently. To work dependably this procedure assumes that the sample matrix has little effect on.�he response for TCDD. To test this assumption, a series of samples was analyzed containing a fixed amount of 'l'CDD and varying amounts-of squalane, a c30a62 hydrocarbon intended to mimic residues obtained from TCDD isolation procedures. As is indicated in Table 14A, the amount of squalane significantly affected the response for TCDD. This suggests that it would be very difficult to de- termine the amount of TCDD present directly from the size of the TCDD peak alone. The results also suggest that in order not to significantly decrease the response for TCDD, the total sample size should be kept under S µg for a squalane-like matrix. 1) Backround 2) :0.5 pg of TCDD I I 322.000 I 321.900 Pigure 14. Preee.nt limit nf detection for TCDD. Analysis of 0.5 pg of. TCDD at m/e 321.8940 (relative isotopic distribution intensity of 100). Source 230° , electron beam trap current 1.0 mA, electr1Jn multiplier 700, resolution 8000, time average of 165 scans at 4 scans/sec.. · · ...°' w 137 'l'ABLE 14. A. Effect of size of.total residue. Response for 'l'CDD Squalane added (micrograms) 0 s. 25 B. Relative response for 20 pg TCDD 100 75 50 15 Ratio of two componentsa Squalane added (micrograms) 1 'l'CDD/TBB 11 s 1 aRatio of response for 20· pg TCDD to response for 200 pg of 2,3,5,6-tetrachloro-4-bromethylbenzene (TBB). 138 Response Relative to a Second Compound 2. (2,3,4,5-tetrachloro-4-bromoethylbenzene) Added as an Internal Standard Another approach is to add a known amount of a second compound and to measure the response of TCDD relative to this compound. A compound, 2,3,4,S-tetrachloro-4-bromoethyl­ benzen!!(TBB), with a mass (m/e 321.839) close to that of TCDD but well separated from it or any other expected environmenta� residues, was synthesized for this purpose. Thi� procedure requires.that the ratio of the response for TCDD to that for TBB remain constant, independent of the·samplematrix. Table 14B indicates, 3. As this requirement was not met. Response Relative to TCDD Added as an Internal Standard A third alternative is to measure the response for TCDD in a given aliquot of sample relative to the response in a similar aliquot to which a known amount of TCDD has been added. The in­ crease in response caused by the known amount of added TCDD can be used to quantiut.e the response for TCDD in the first aliquot. For the procedur� to be useful, the response to TCDD should be linear in a given sample matrix. Thus addition of authentic TCDD to a second aliquot can be used for quantitation. Figures lSA and 158 illustrate that the response is linear for neat TCDD and for TCDD in the presence of residue corresponding to 700 600 500 Peak height(mm) 400 ... w 'IO 300 200 100 oa;;;.___..,a.____..________.______,j.______,________. 3 5 10 15 TCDO (pg) 20 25 FIGURE lSa. Linearity of response for neat TCDD. The TCDD values are the amounts introduced into individual runs on the MS-9. 0 120 100 80 Peak height (mm) 60 ... w \0 Ill 40 20 40 60 80 100 120 TCDO (pg) Linoarity of reaponoo for Tep:, PIGUU 15b. in the presence of beef liver residue. · The TCDD values are the amounts introduced into individual runs on the :,is-9 140 140· 0.5 9 before cleanup of beef liver. Before the multiple ion detection method described below was developed, this procedure was used routinely to mea�ure levels of TCDD. Details are given in the Experimental section. As described in·Part 3 37 of the next section, recoveries of c1 labelled 'l'CDD also were measured with a variation of this procedure. For determining 37 c1 TCDD was added to a sample recoveries approximately 1 ppb.of before the start of the TCDD isolation procedure. Addition of authentic TCOD has an additional advantage • in that it acts to confirm the a/e value of an observed peak. 4. Isotopic Dilution Another method of quantitation is isotopic dilution. Zf the rati.o of two isotopic isomers- and the · amount of one is known, then the amount of the second can easily be calculated • . 'J.'he method requires an isotopic isomer of the na.turally occurr­ ing compound that can be added to the sample in a known amount and some means of measuring the ratio of the two isomers;mass spectromety provides a direct means of measuring isotopic ratios. Zn the. present case TCOD labelled with 95.51 31c1 was used as the isotopically labelled compound.* c1 TCDD was originally prepared for use as a recovery *The 3 standrad (see Part 3 of the following section). 1•¼- . n( __ , __ , *-Rt ., ,_;@$ U!Qlf,il $6,,.,, 49 ;,- *,f' "lf.' /4¥.J!\%\Jzt; 4 Q-4 s._4.4.f_.1$¼ • ..:; - � .. 1.• W ...fi.!$14'! t ;:qu 141 In using the mass spectrometer to measure isotopic ratios it is desirable to record the intensities of both isotopic isomers during the same analysis. This procedure, which will be referred to as multiple ion detection, minimizes the error introduced by instrumental instability from run to run and at the same time reduces the total number of analyses required. For high sensitivity it is alse>.desirable to be able to time average both signals. As described in the Ex- perimentalsection, the present system was modified to achieve this. The procedure developed also provides an opportunity to confirm the mis value of an observed peak via overlapping peak positions for a particular mass ratio. In· .practice a sample was weighed and a known amount of 37c1 TCDD was added to it. The sample was then taken through the 'l'CDD isolation,procedure, divided into aliquots and analyzed by multiple ion detection. The ratio of isotopic isomers in the final residue was necessarily the same as the ratio before cleanup (no isotope effects were observed). Therefore the amount of naturally occurring TCDD could be determined from this ratio This value divided by and the known e.mount of 37c1 TCDD. the weight of the sample gave the concentration of naturally occurr­ ing 'l'CDD. ° This method provides no information about the percent recovery. As. in the previous procedure recoveries were measured by the addition of 37c1 TCDD to duplicate fractions ! : 142 just prior to MS analysis. A drawback of this method is that since only half as many scans are made at one m/e value, the sensitivity is decreased by a factor of approximately l/V2. After the time averaged multiple ion detection system had been assembled, unless the highest sensitivicywas required, isotopic dilution exclusively was used to determine TCDD levels. c. Isolation of TCDD from Biological. Samples l. Separation from Major Tissue Components Given that the goal of the TCDD isolation procedure was to make possible analysis of l pg of TCDD per gram of tissue in the mass spectrometer, almost total elimination of other tissue components was necessary. As mentioned earlier, no more than about 5-10 µg total residue per fraction was acceptable for high sensitivity MS analysis. This meant that for a fraction corresponding to one gram of tissue, 99.9991 of the other In one procedure treattissue components had to be eliminated. ment with strong base,203 strong acid,203 or both and in a a second procedure extraction into a neutral solvent, methylene chlorid_e, were used as preliminary cleanup steps. case a considerable amount of residue remained. In each This residue was greatly reduced by the use of an alumina chromatography step. The improvement provided by the alumina is illustrated i :: 143 in Figure 16. One procedure was developed which used a preparative glc step after the alumina, and a second pro­ cedure was developed which inste�d employed an additional alumina step. Both of these procedures reduced overall residue levels to such an extent that sampl�s corresponding to up to about 5 g of original sample c�uld be analyzed per MS run. The effect of various amounts of residue from the glc procedure on the response for TCDD is illustrated in Figure 17. The neutral extraction procedure followed by alumina chromatography was less efficient and was suitable only for samples up to about 0.5 g per MS run. The advantage of the neutral extraction was that it avoided strong chemical con­ ditions which might have converted precursors to TCDD. A total sample size of 20 g or more could be handled with the saponification and acid extraction procedure. The neutral extraction was limited to 5-10 g, largely by the capacity of the alumina to hold other tissue components which were co­ extrac.ted. 2. Separation from Other Chlorinated Residues In nearly all environmental samples residues of a variety of man-made chlorinated compounds are present. DDT and its metabolite or degradation product 2,2-bis (p-chlorophenyl)1.1-dichloroethylene (DDE) are often present at the part A, Without alWl!ina cleanup a. Wi�h alWl!ina cleanup ......... Figure i6. Effe� of alumina C'leanup on chicken live r tesidue. Figure 6A before alumina chromatoqraphy and Figure 6B after alumina chromatography. Each·injection corresponds to 100 mg of oriqinal sample. Glc analysis on 5% SE-30 on 60/80 Chromosorb W, 2 m X 2 mm (id) stainless steel, 245° column, 30 ml/min N2' attenuation l X 100. 145 20 1S Peak height (mm) 10 5 l 2 3 4 Beef llver residue (gram orlg. sample/fraction) · FIGUR!l7 Effect of sample residue onTCDD response for two different alumina cleanups. Each fractton contains an . ·tdentlcal amount of TCDD . (20 pg). 146 per million (ppm) level. Polychlorinated biphenyls (PCBs) are often present at or near the ppm level. Other chlorinated residues may be present at levels from a few parts per billion (ppb) to hundreds of ppb. The possibility that such chlorinated organic compounds present in the �nvironment might interfere with or obscure the TCDD peaks in the MS was investigated. Mass spectra were obtained for most of the common organochlorine pesticides ·including lindane, aldrin, dieldrin, mirex, heptachlor, DDD, DDE, and DDT, as well as various polychlorinated biphenyl (PCB) mixtures. In the TCDD mass range DDE from its molecular ion has isotopic isomer peaks at m/e 320, 322, and weakly 324. 'l'he molecular ion of pentachlorobipheny�, a component of some PCB mixtures,· has a peak at m/e 324, and this compound has a weak associated peak at m/e 322. peaks at m/e 320, 322 1c and 324. DDT has weak fragmentation As shown for m/e 322 in Figure 8, all of these compounds can be resolved from TCDD at a resolution of 10,000. The relative input amounts of each compound producing the peaks shown are: DDT, 250; DDE, 25J 'l'CDD, l; PCB ( Al:0chlor 1254), 250, '1'.hus moderate excesses of these compounds will not interfere with TCDD analyses. levels encountered in env:'>.·.: -:mental samples may be_ 10 6 However, times as great as the desired st,:. 1.tivity for TCDD (ppm versus ppt), . which means that a 104 or more reduction in the levels of these 147 compounds relative to TCDD may be required for successful analysis. A procedure was available from the literature for separating chlorinated dioxins from PCBs.224 PCBs were eluted from an alumina column with a 11 solution of methylene. chloride in hexane while the dioxins were eluted with a solution of 201 methylene chloride in hexane. A trial of the pro- cedure established, however, that DDE and the PCB related peak _at m/a 322 eluted with the dioxins. 'l'he PCB related peak be- haved much like a reference sample of a chlorinated biphenylene, and may in fact represent a class of stable environmental residues associated with, but chemically different from PCBs. ('l'his is discussed in greater detail in the Results and Discussion Section.) 'l'O differentiate the unknown peak from PCB it will be hereafter referred to as PCBene. Oxidation with JCMno4 and treatment with Br2 were ineffective in .reducing the amount of DDE. A new solvent system was tried with the alumina column. TWenty percent carbon tetrachloride in hexane was used as the first eluant followed by the usual 201 methylene chloride in hexane. The rationale for this was that DDE and PCBen_e, might be eluted almost as well with carbon tetrachloride as with methylene chloride, while TCDD with its two somewhat polar ether linkages might be less easily eluted with nonpolar 148 carbon tetrachloride and nmain on the column to be eluted with the methylene chloride. As Figure 18 illustrates this procedure did in fact greatly improve the separation of DOE and PCBene from TCDD relative to the previous procedure. For some samples one alumina column was sufficient, for other samples, as shown in Figure 19, a second alumina column was useful in providing additional cleanup. Under the glc conditions used with Isolation Procedure I the retention time of DDE was about one half that of TCDD, ao that the preparative glc step also-provided a substantial cleanup relative to DOE. With the neutral extraction procedure some DDT appeared to survive the cleanup, but not enough to interfere with the determination of TCDD. With the saponification procedure any DDT present is presumably quantitatively dehydrohalogenated to DDE1 no DDT peak was ever seen with this procedure. 3. Recoveries In view of the low level of TCDD being analyzed, and of the delicate balance of parameters in some of the isolation steps, it was desirable to have some way of independently measuring the recovery of TCDD for each individual analysis. Due to the·specificity of the cleanup, particularly if glc is used, the best internal recovery standard would be an ·• 1 • 149 ........ I IIIJ1' I .. -·- .... I I 1aJO PCB C;, t Slondanl I DOT D. .,ca (llrochlor U54) + TCDD I I IIIJ1' DOE I f TCDO PCa ....... r" _________.f'--.__,_____ ..,,. 121.,00 S. DDl'+-+K8+7Clll> -·-- I DO£ iu.aoo I lCDO ..... I I I PCB S- -·-·- FIGURE 18a. Resolution of 'l'CDD from DDE, DDT and PCB with high resolution (10,000) mass spectrometry. Relative input amounts of each compound are: DDT, 250; DDE, 25; TCDD, l; PCB (Arochlor 1254), 250. 149a DOE PCB •• A•. 322,000 :m.aoo FIGURE 18b.· Mass spectra showing reduction of DDE and PCB levels in fish residue by means of alumina chromatography. Following the sulfuric acid cleanup step, the residue in hexane is added to a-column.of activated alumina: (A) Trace from. the mat.erial eluted by ·201 CH2Cl2 in hexane after the column.was first eluted with 201 CCL4 in hexane; . (B) trace·obtained from a similar . 201 CH2c12-in-hexane elution after the column was first eluted with 11 CH2c12 in hexanl!. Elution with CH2Cl2 in hexane was reported to be effective in reducing the amount of PCB residues (5). Elution with 201 cc1 4 is clearly even.more effective and was routinely used in obtaining the results reported here. 150 322.00 I 321.900 I 321.800 •I• FIGURE 19. One versus two alumina columns in the cleanup with Procedure II of tissue from a rat fed TCDD. A. One alumina, 0.24 g per MS run. B. Two aluminas, 0.25 g per MS run. .-, . Wt t··· W ,., ., tttt 151 TABLE 1S. , : Comparison of the isotopic isomer distributions at the molec.ular ion of naturally occurring TCDD and synthetic 37 cl TCDD. 320 Calculated for natural TCDD 0.77 Observed for synthetic 37cl TCDD Calcu! ted for 4.51 gCl and 95.51 37c1 in TCDD. 4.2xl0-6 322 m/e 324 326 328 '.1.00 0.49 0.10 0.0.1 4.Sxl0-4 0.016 0.18 . 1.00 3.6xl0-4 0.012 0.18 1.00 TABLE 16. Recoveries of TCDD from isolation procedures. Procedure Procedure I: Saponification, acid treatment alumina chromatography, glc Procedure II: Saponification, acid treatment, alumina chromatography Procedure III: Methylene chloride extraction, a lumina chromatography Rec overy (l) a 71±18 (32) b 86±12 (16) c aMean t the standard deviation for the number of independent analyses shown in parentheses. bBased on added 37c1 TCDD. cBased on added 14c TCDD. rzzr· · , t 1:(1 � .,: rri?itn ··ttw- r-et 1t1trw tttme· th t tTrHt'" -ttws- rt · art· ""etti:t ·a I rt Dtttt b 152 TCDD labelled with isotopically labelled form of TCDD. 37 A compQund labelled 95.51 Cl was prepared for this purpose. to a high degree with a stable isotope was selected because levels of such a compound can easily be measured by mass The isotopic isomer distribution of the syn- spectrometry. thesized compound is compared to that of naturally occurring TCDD in Table 15. than 10 3 From the Table it is apparent that more excess of 37c1 'l'CDD can be present without effecting analysis of naturally occurring TCDD at m/e 320 and 322. Thus iri an interesting reversal of the normal situation with tracers, the isotopically labelled compound was used as a carrier for the natural compound. Routinely 37cl TCDD, at about 10 3 times the limit of detection for nat�al TCDD, was added at the be­ ginning of each sample cleanup to act as a recovery standard. As described earlier, the 37c1 TCDD also was used to measure the level of any natural TCDD that was present by means of multiple ion detection. Recoveries for both added natural 37 TCDD.and c1 TCDD were the same, indicating that there was no isotope effect leading to selectivity for one or the other isotopic isomer. Recoveries for each of the three Isolation Procedures are listed in Table 16. Recoveries for the variow; individual steps in the isolation procedures were measured and are listed .in. Table 17. It is apparent that the low recoveries in tf ttir:r·r (-n·· rttai:> · rr iwwmin¼ · Ct · �····-,,-.v ·. ·; ·e- 1 t Tfritt· · -:trtrk½w:. "itY . 153 Table 17. Recoveries of TCDD for individual steps in the cleanup procedures. Cleanup step Recovery (I) Extraction alone Saponification (2 hr) and extraction Treatment with 95-971 H2SO4 Alumina chromatography Microconcentration (slow) Microconcentration (fast) Glc prep 100 + 9 78 + 18 109 + 16 94 + S Table 18. 0 5 20 B. 0 20 100 500 2000 IPS- ,. ±8 Observed ppt (recovery) Human milk (U.S.) HD (LD 1.1 + 0.5) · 3.1 + 0.3 (621) 15 .:!:,-3.5 (751) (5) (3) (2) Fish (U.S.) HD (LD 3.1 + 0.9) 14 + S (70lf 68 + 9 (68\) 380-+ 95 (761) 1400-:!: 380 (701) (6) (8) (8) (7) (9) 1:Pll A&_.,.e:pµ p; : Rif?/1*- .4,!Jf .¢..&,if:,,._A?- --* :'* .. %4[&¥ 31 Recoveries of TCDD at different added levels. Added ppt (no.) A. 1�:} ½� Sf-,::; ®9:¼¥¥1¥¾'"< �"'*''!"'f'"'-?iiii,..,*.""i""1..JJ44.""�y,,... -........ . ti!! ..,_J.',...;:....#""4"'". "'-A,,.,,. ..,.., .*-'"""¥1.. -_µ""· """"MJ ' "".,., ...;, ..,.,..,_;,..,,,.._.,._..,__ z..,.,..,.,..,.u,..,,., __z.....,. ,,.,..,;s _;.,,. ™'"'·"""M""'*' "'" · : ·" 154 Procedure I largely can be attributed to the preparative glc step. Recoveries with Procedures II and III were more satisfactory.- Recoveries with Procedure II of TCDD added over a wide range of levels to human milk and fish are listed in Table 18. Time averaged traces for the limit of detection and for the recovery of 5 ppt of added TCDD in u. human milk are illustrated in Figure 20. D. s. Errors and Reproducibility Overall errors-for the three independent isolation pro­ cedures are listed, expressed as standard deviations, of the mean recovery in Table 16. Expressed as percents the errors are about! 201 for Procedures I and II and about: 101 for Procedure.III. The error for mass spectrometric determinations alone, indicated by error flags in Figure 15, was on the order of ± 101. 'l'hus a substantial part of the observed error can be attributed to the MS measurement itself. Inhomogeneity in the original sample also can produce uncertainty in the final determination. Although some variability possibly attributable to this cause was observed, in general samples which were visibly homogeneous gave results reproducible within the range of other errors in the procedure. Another important source of error can be inaccuracies 155 A.· Human milk 1.23 g TCDD level: ND (LD 1.3 ppt) B .- 1 1 ml!. Buman milk 1.17 g + 5 ppt TCDD TCDD level: 3.8 ppt (LD &.9 ppt) 321!soo FIGURE 20. Limit of detection (LD) and recovery of 5 ppt of added TCDD in a U.S. human iidlk sample. 156 in the amount of 'l'CDD or 37c1 TCDD use as a basis for quan_titation. The concentration of standard solutions must be accurately known. This was checked in two ways in the Ultraviolet absorption was used to measure concentrated stock solutions and 14 c TCDD for which present investigation. the specific activity had been determined by MS was used to calibrate dilute standard solutions of 37c1 TCDD by a combina­ tion of liquid scintillation and MS isotopic dilution measure­ ments (see Experimental). 156a E. Extraction Efficiency It is conceivable that TCDD accumulated in animals may in some way become conjugated or chemically bound so that it cann�t be easily extracted. If this were true, then the recovery of exogenous TCDD added at the beginning of the isolation. procedure would not necessarily reflect the re­ covery of endogenous TCDD. In a test of the extraction efficiency of the isolation procedures used in this investi­ gation, rats.were dosed with S µg/kg (98 mCi/mmole) by intra­ peritoneal injection of 14c labelled TCDD. After 30 days the rats were sacrificed and the level of 14c TCDD in the liver was determined for extraction under strongly basic and acidic conditions (Isolation Procedures I and II) and under neutral conditions (Isolation Procedure III). The total radioactivity in the tissue, which would include any bound or unextract'ed TCDD, was determined by solubilizing the tissue with Protosol and counting directly. The results were as follows: Procedure 14c TCDD level (dpm/g) Strong base and acid (Isolation Procedure II) 1940 ± 270 (7) Neutral extraction (Isolation Procedure III) 1810 ± 230 (lll Protosol (tissue, homogenate) 2060 ! 220 (4) The number in parentheseo ref;;rs to the number of runs. The extraction efficiencies of the base and acid procedure and 156b the neutral extraction procedure both were on the order of 901. Therefore., ·at least over the time-scale of one month, TCDD does not become bound in an unextractable form. In sam­ ples of whole rat (including liver), both Procedures II and III gave similar levels of 14 c TCDD, but the level waA too low to measure with the Protosol treatment which could handle only a few hundred milligrams of tissue. 15? IV. Results and Discussion A. 'fCDD Levels in Environmental Samples Results from . a limited number of i:.nalyses·of fish, . . crustaceans, and human milk collected in South Vietnam in August and September 1970, several months after the cessation of the use of 2,4,5-T are listed in Table 19. The sample col­ lection procedure is described in the Experimental Section.· In Figure 21, the sample collection sites are shown on a map of South Vietnam. The letters correspond to those in Table 19. The stippled areas were treated with 2,4,5-T before its use was directed to be discontinued in April_l970. TCDD levels of up to several hundred ppt were observed in fish from the Dong Nai River (sites A and B) which drains one of the areas most heavily · treated with 2,4,S-T. A TCDD signal from one of these samples·, a carp, is shown in Figure 22. Analysis of mother's milk from Tan lJyen (site A), a village on the Dong Nai River just below the sprayed areas, gave a value of 40-50 ppt.of TCDD. Lower levels were observed in sampies of fish and milk from other sites inland and from can Gio on the seacoaat. No TCDD peaks above a limit of detection of 3 ppt were observed in Cape Cod butterfish used as a control, or above 1 ppt in control mother's milk collected in Boston. Similarly, no peaks were observed in water-blank samples stored in the liquid nitrogen refrigerator throughout the_ sample collection and storage period. Results for analyses of similar samples collected in June 1973, approximately three years after the cessation of 2,4,S-T --V:� t.,. ¢_,4.$£i\Z_?4 _-3i¥Jti/49·,4'?J!®!:4#-.f¼.,.,.r-_.k t4 _ Siuu . .n __ ;;;;,;:,, e H1;.;;,;_*"· flll'� ...•.. 9p:1¢ily4B,.. +. 106· I Cambodia AA. ... - UI South _.,.__ China Sea Gulf � of Thailand IOOKm N•. 106° IOS° FIGURE 21. Map of South Vietnam showing sampling sites. Stippling denotes the principal areas sprayed with herbicides form mid-1965 through 1970 .. The use of Agent Orange was terminated in the spring of 1970. "'-'r'H , 159 A. VIETNAMESE CARP PLUS B ,· VIETNAMESE . CARP C, 322,000 321,900 CAPE Con BUTTERFISH 321.800 FIGURE 22. TCDD signals observed in fish samples: (A) Vietnamese carp plus 60 pg 'l'CDD, (wet weight of fish 0.18g); (B) Viet­ namese carp, (wet weight of fish 0.18g); (C) Cape Cod butterfish, (wet weight of fish 0.16g). TCDD TABLE 3. MaJor polychlortnated molecular tons or ftegments observed 1n the masa spectra of the finalresldues from HCP F-981 and HCP F-996. ( No. of chlorine .atoms Possible molecular formulf Posillble tdentlftcatlon ( F-981 +++' ++ 208 �3 233 3 249 3 252 3 281 4 288 4 315 5 320 4 348 ? f-996 + ++ ++ ++ 38�HC1z Tetrachlorodlbenzo-p•cUoxln ++ ++ ++ ++ +++ + + ++ + 366 s s 381 5? 416-Cl +++ 386 6 He;gschloro:xanthene +++ +++ 407 4? 442-Cl ++ ++ 416 6 Hexachloromethoxyxanthene +++ 431 4? 466-Cl ++ 442 6 ++ + ++ ++ 456 6 ++ + 466 6 ++ ++ 486 6? 351 386-Cl ++ +++ + + + * Molecular ion confirmed for F-996 by high resolution mass spectrometry (resolution 30, 000) + Present at <0.001 to 0.01 ppm in original sample verYapproxlmately (assuming slgnal due to molecular ion) " ++ Present at 0.01 io 0.1 ppm " +++ Present at 0.l to < 1. 0 ppm 1S9a TCDD levels in fish, crustaceans and human milk collected in south Vietnam in August and September 1970 TABLE 19. · Site A B B C C D. D Cape Cod A A ·D D D E E E H C C G G G G F F F Boston noise). Sample Carp (Cyprininae) Catfish (Siluridae) Catfish (Tachysuridae) Catfish (Schilbeidae) River Prawn (Palaemonidae) Croaker (Sciaenidae) Prawn (Peneidae) Butterfish (Stromateidae) Human milk • • • • • • • • • • • • • • • • • • Level (ppt wet weight) 4 540 810 520 70 42 79 18 HD(LD • 3) 55 44 22 11 10 7 HD(LD = 3) 11 5 9 10 HD(LD ND(LD. ND(LD ND(LD ND(LD · ND(LD ND(LD ND(LD = 2) = 3) • 2) = = = = = 6) 4) 15) 7) 1) a ND• None detected; LD = Limit of detection (2.5 x 160 use, are listed in Table 20. Compared to 1970 the TCDD levels appeared to be lower by an order of magnitude or more. In human milk, lov levels of TCDD were observed in samples from · Tan Oyen and Can Gio. As discussed in Section C below, other compounds appar­ ently associated with 2,4,S-T were found at up to ppb levels in the 1973 fish samples. The levels of these compounds appear to be higher in the 1973 than in the 1970 samples. The limits of detection are higher for the 1973 fish samples because a neutral isolation procedure (Isolation Procedure III), which required smaller samples, was used to minimize the level of other 2,4,5-T related compounds in the final residue (see Sec­ tion. C) and to eliminate the possibility of formation of TCDD in the saponification and acid extraction steps. As described in the following section, a series of con­ fizaation experiments was carried out to establish whether the signals observed in the 1970 samples in fact were derived from TCDD. In each case the observed compound behaved in a manner identical to that of TCDD. · 'lhe TCDD levels found in human milk suggest that in reginns .hnediately adjacent to areas treated heavily with 2,4,5-T t:hex:e was human exposure to TCDD. .. However, medical surveys con. ducted in South Vietnam in 1970 and 1973 by Dr. John Constable, in connection with the collection of the samples analyzed in the '*t>.rofessor of Surgery ,. Harvard Medical School, Boston, Massa­ chusetts. 160a TABLE 20. TCDD levels in fish, crustaceans and human milk collected in South Vietnam in May 1973 Site Sample A Carp (Cyprinidae) A A A A A C • Catfish (Siluridae) • Carp (Cyprinidae) • p p River prawn (Palaemonidae) Catfish (Clariidae) · Perch (Anabatida2) D D Human milk A A A p p • • • • • • Level (ppt wet weight) ND(LD ND(LD ND(LD ·. ND(LD ND(LD ND(LD ND(LD ND(LD ND(LD .. 112) 56) = 30) = 76) = 150) = 40) = 23) = 20) • 91) 6 8 NDCLD • 1) 4 ND(LD = 2) ND(LD = l) ND(LD • 1) 161 present investigation, revealed no evidence of severR and wide­ spread i�lness in several villages near sprayed areas or in dis­ cussions with officials of the South Vietnamese Ministry of Health. There was some evidence of increased ,Jlncidence of sti�l- births and certain birth defects, viewed as inconclusive in itself, but as sufficient to justify further investigation of 225 possible connections with herbicide exposure. The Nationa1·.Academy of Sciences Committee on the Effects of Herbicides in Vietnam, on the basis of extensive interviews and comparison of herbicide flight records, concluded :that: •Reports of Highlanders (Montagnards), in comparison with low­ land Vietnamese, on death and illness caused by herbicides are so consistent that despite the lack of medical and toxicological evidence for such effects they cannot be dismissed out of hand and should be followed up as promptly as possible by intensive .226 There is no evidence to indicate whether TCDD .s tudies ••• was or was not involved in these reported incidents. 225American Association for the Advancement of Science Herbicide Assessment Commission, "Preliminary Report and Back­ ground Material,• Congr.,Rec., 3 March 1972, p. 6806. 226cong. Rec., 28 February 1974, p. 52437. 162 B. · Confumation Procedures In routine analyses, to be classified as 'l'CDD the observed compound had to follow ·37c1 TCDD through the isolation pro­ cedure, have an MS signal at !f!. 321.894, and have an isotopic isomer signal with a relative intensity of 0.77 at !!!I!. 319.897. In addition the foilowing additional confirmation tests were . carried out on selected samples. With the use of Isolation Procedure II, which did not in­ clude a preparative glc step, -some samples gave a second signal at the TCDD !f!. values that hal an appearance time in the mass spectrometer about 1.8 times later than that of TCDD. As is discussed at length in the section on other 2,4,5-T related compounds, this second signal apparently was derived frolll a . fragment�.tion ion of some higher molem!lar weight compound asso­ ciated with 2,4�5-=-T• The late signal could be reso:!.ved from TCDD by stopping the analysis with.f;.he disappearance of the 37c1 TCDD peak, by using a preparative glc step, or by using a neutral extraction procedure (Figure 23). The _presumed TCDD signal in all of the following confii:mation experiments had an appearance time in the mass spectr0111eter exactly coincident with that of TCDD. 1. Mass Spectrometry The compound present in Vietnamese fish had relative isotopic intensities at !!Y!. 319.597, 321.894, and 323.891 at 0.77, 1.00, and 0.49, intensities which can be obtained only with a molecular formula containing four chlorine atoms. The ""¥klN'f"\4·""' ""-"'·""*·f""""""' - ·' *""w.'"',....,"'".,,2""*"'""' ... ._._ LJI ..,..,._ ,..m..,, ....,......,......,_N.....n,...,., _,,......,_�-�..,...,.....,..""'"'..," , .¢""""·'*""''"""'" """""it""*""3"".' "'-·-· ""t ,.., _.,...µ_ ,..,,,.,.,, i!. ""'w-· ;...,._,.,.,,�__.,.,...,,,.,: S&"'* ""'· ...,._ "'""· '"'"""""' �-· 163 observed M+l/M ( 13c or other minor isotope) ratio at 319.897 and 320.897 was 0.1 3 . 0.125. The expected M+l/M ratio for TCDD is At a resolution o! 15,000 the!!!/.! value for the base peak in the Vietnamese fish was found to be error of about ;t0.0030 mass units. TCDD is 321.8936. 3 21.8948 with an The theoretical value for By means of.the •t::lcomp• computer program227 all possible molecular fo%111Ul.as within an !!Y!. range of ;t0.0030 mass units containing four chlorine atoms and giving approxim­ ately the observed M+l/M ratio.were calculated. Carbon, hydro­ gen, nitrogen, oxygen, fluorine, silicone, phosphorus, and sul­ fur were included. listed in Table 21. The six resulting possible compounds are Four of the compounds contain fluorine; an unlikely organic substituent in samples from Vietnam. is chemically improbable. that corresponding to TCDD. A fifth The only likely molecular fo:rmula is As is usual with electron impact ionization, the mass spec­ trum of TCDD contains fragmentation ions as well as the molecular ion. For TCDD this fragmentation.is less than usual, but there is a fragmentation ion at!!!/.! 256.933 (actually the !!I.! 258.930 isotopic isomer was used) derive� from loss of COCl which has an intensity about 0.21 times that of the molecular ion and which is characteristic for diphenyl. ethers and dibenzo-p-dioxins. The compound present in.the Vietnamese fish produced the su.e frag­ mentation ion with a relative intensity of 0.22. 227with the assistancA of istry, Massachusetts Institute c. Hignite, Department of Chem­ of Technology. o�-"-:'....,_��---.'"'.e--•.,,-.,;-, -,-,,,. , -.'"'"·"""·"""" L ..... % .,....,,.. ..... P.(,.,f, .......... . 4,_,•.f.""·""'·""4-'"' - ·""*':'*"'"..... · ...,......,....... , .._.................. .4¼""$""!J! .l __... ,..,A,...,. ... .. --··.-�-�-,.,,....,,..? . ,".-s �·""""'" . .. ;: .- " 164 Table 21. Calculated molecular formulas within±. 0.0030 mass units of TCDD (ref. 227). !!/� M-Mc 12H402 Cl 4 3lt.8936 0.0000 319.8912 -0.0024 319.8922 -0.0014 319.8932 -0.0004 319.8948 319.8934 0.0012 -0.0002 Molecular formula C12B4O2c14 C10H3F3Cl 4 C1082N30Cl 4 c 9B6F2SiC14 c9H5O3Fcl 4 c5H 9NF 2SiPC14 165 Upon ionization TCDD al.so gives a significant amount of doubly charged molecular ion. 'rhis presumably results from the stable planar aromatic system of TCDD. Since the ion has a net charge of two, the 321.894 peak is now observed at 160.947. For TCDD the intensity of the doubly charged nolecular ion is about 0.17 times that of the singly charged molecular ion. For the -compound in the Vietnamese fish there was an identical ion with a corresponding intensity of 0.19. �ith electron impact ionization the extent of fragmentation depends on the energy of ionizing electrons. ionizing voltages there is less fragmentation. l'.n general at low l'.f the pres=ed TCDD signal observed from the Vietnamese samples represents a fragment ion of some other higher molecular weight compound, then .the ratio of this signal to the signal of the added 37TCDD should decrease at lower ionizing voltages. A sample of Vietnamese fish which gave a 322/328 ratio of 0.041 (for the particular 37cl TCDD spiking level in that sample) at the normal ionizing voltage of 70 ev, gave a ratio of 0.039 at 25 ev. As samples volatilize after being introduced into the mass spectrometer, the ion intensity associated with a particular com­ pound increases and then decreases with a characteristic time course. This pattern of volatilization can be used to differen­ tiate ions of different origin which occur at the same !Y!. value. As indicated at the beginning of this section, signals 1ttributed to TCDD covolatilized with 37cl TCDD. This was established by means of ion detection at !!!I'� 321.894 and 327.885. 166 2. Separation of 2,3,7,8-Tetrachloro-dibenzo-p-dioxin by Preparative Glc. On the basis of their mass spectra it was not possibl� to distinguish 2,3,7,8- from 1,3,6,8-TCDD. This distinction is important because 1,3,6,8-TCDD is not known to be present in 2,4,5-T (or 2,4-D) although it is present in certain othe� chlor­ ophenol products. If the TCDD present in the Vietnamese samples consisted entirely or.in large part of 1,3,6,8-TCDD, then it would be unlikely_ that the TCDD was derived from use of 2,4,S-T. Another important point is that 1,3,6,8-TCDD is much less toxic than 2,3,7,8-TCDD. Other TCDD isomers have not been·found in 98 known dioxin contai4ing .chemical products • .. On the basis of the observation that 1,3,6,8- and 2,3, 7 ,8TCDD can be resolved successfuly with glc (Figure 6) a.prepara­ tive glc separation procedure was devised. A sample of Vietnamese fish, along with appropriate controls, was put through this pro­ cedure (Table 22). The behavior of the TCDD observed in the fish exactly paralleled that of 2,3,7,8-TCDD. ,.._.,; 3. Analysis for Hexachlorodibenzo-p-dioxin As with 1,3,6,8�TCDD, the presence or absence of hexachlor­ odioxin provides information about the possible origin of any Although hexachlorodioxin has been reported in some 2,4,5-T samples, the levels were lower than for TCDD. 228 TCDD observed. 228E.A. Woolson, R.F-. Thomas, and P.D.J. Ensor, J. Acrr. Food �, .!Q., 351 (1972). 167 Table 22. Separ ation of 1,3,6,8-TCDD from 2,3,7,8-TCDDby preparative glc. (See Experimental Section for conditions.) Percentage ot total TCDD observed Sample l. 75 to 2.30 times· m-terphenyl ta 1.30 to 1.75 times m-terphenyl ta Neat 1,3,6,8-TCI>_Da _ 721 281 Neat 2,3, 7,8-TCDD. . <2.51 >91.51 Butterfish + 500 ppt 1,3,6,8-TCDD 721 281 Butterfish + 500 ppt 2,3,7,8-TCDD <3.21 >96.81 Vietnamese fish (230 ppt total TCDD) <3.81 >96.21 Butterfish blank N. D. (<3.6 ppt) N. D. (<3.1 ppt) �he presence of 1,3,6,8-TCDD signal at the la ter retention time is probably due to tailing, not to the presence of 2,3,7,8-TCDD. (The 1,3,6,8-isomer is >951 pure.) Table 23. Analysis for hexachlorodioxin. Section for conditions.) (See Experiment al Sample Bexachlorodioxin level Butterfish + 1500 ppt hexachlorodioxin 860 ppt (571 recovery) Butterfish + 1170 ppt hexachlorodioxin 880 ppt (751 recoveryl Butterfish bl ank <33 ppt Vietnamese fish (430 ppt. TCDD, uncorrected for recovery) <21 ppt to.·' I fi·. 168 Pentachlorphenol, which is used as a fungicide in the lumber industry, is another source of chlordioxins. In this compound higher chlorina�ed dioxins predominate, levels of hexachloro­ dioxin are much higher than those of TCDo.98 The absence of h!!l(achlorodioxin in the Vietnamese samples would be consistent with the source of the TCDD being 2,4,5-T but it would be in­ consistent with the source being pentachlorophenol. By MS analysis at !!V.! 387.837, the level of hexachlorodioxin in a sample of Vietnamese fish previously measured to contain 759 ppt of TCDD was found to be below the limit of detection of 21 ppt (Table 23). No record could be found of shipment of pentachlorophenol from the United States to South Vietnam. 4. Lack of Formation of TCDO from 2,4,5-Trichlorophen­ oxyacetic acid or 2,4,5-Trichlorophenol in the Isolation Pro­ cedure An experiment was carried out to test whether TCDD might be foJ:1Ded from 2,4,5-T or its degradation product 2,4,5-tri­ chlorophonol during the isolation procedure. 2,4,5-T and 2,4,5-trichlorophenol, both at a level of 100 ppm, were added -to Cape Cod butterfish which previously was shown to have a! 3.0 ppt of TCDD and then carried through Isolation Procedure II. The TCDD level observed,� 4.0 ppt, was not significantly dif­ ferent from the untreated sample. This indicates that TCDD is not formed during the isolation procedure from 2,4,5-T or 2,4,5trichlorophenol when these compounds are initially present at 169 levels up to 100 ppm, that is, less than one part in 107 is formed under these conditions. 5. Neutra l Extraction Procedure. Even though 'l'CDD is not formed from 2,4,5-T or 2,4,5-tri­ chlorophenol in isolation procedures which include saponifica­ tion .and acid treatment, it is·possible that some other related compound might be converted to TCDD under these conditions. In view of the difficulty in identifying and testing every possible precursor, a neutral extraction procedure was developed which comp letely e liminated the strong chemica l conditions of sapon­ ification and acid treatment. The sample was extracted with methy lene _ chloride the residue was chromatographed on alumina and MS ana l ysis was carried out (see the section on Isolation Procedures). For a sample of 1970 Vietnamese fish measured to dure, a of 670 ppt was found with the neutral extraction contain 7SO·ppt of'l'CDD with the saponification and acid proce­ level procedure. This indicates .that the TCDD observed with Isolation Procedures I and II was not fomed in the course of saponifica­ tion or treatment with acid. The neutra l extraction procedure later was used to minimize the TCDD late peak and the levels of HMW compounds in the 1973 Vietnamese fish samples. 6. Treatment with Diazomethane Rappe andNilsson lOO and Jensen and Renberg 99 have reported the cyclization of 2-hydroxynonachlorodiphenyl et.her which they ; \.141¥4'.l'fe:!!'! .. ·A E _ 170 called a •predioxin• to octachlorodioxin duringglc and mass spectrometric analysis. Derivatization-to the methyl ether with diazometha.-ie eliminated the reaction in the glc and mod100 ified that in the mass spectrometer. To test whether the TCDD signal abserved in the Vietnamese fish might have been derived from such a 2-hydroxydiphenyl ether, a portion of residue frOl:\ a 1970 fish sample �as treated with diazomethane and analyzed. The TCDD level found,._ 90 ppt, was the same as that found, 92 ppt, iri an untreated portion of the same residue. Thus the signals observed cannot be attributed to cyclization of a 2hydroxydiphenyl ether. In any case it was reported that the •predioxins• remained adsorbed on alumina under conditions used to elute dioxins and since each isolation procedure used in the present investigation included an alwnina chromatography step, it is unlikely that any hydroxydiphenyl ethers would have sur­ vived. 7. Photolysis_with Monochromatic UV Light The characteristic uv absorption of the diexins provides another approach for determining whether observed TCDD signals are caused by cyclization of a diphenyl ether precursor in the mass spectrometer. The dioxins which absorb in the region 290- 320 nm can be decomposed with uv light ef this wavelength, while the diphenyl derivatives which do not absorb in this region should remain intact. A procedure has been described in which irradiation with a sunlamp is carried out for an extended period 171 to achieve total decomposition of the TCDo.228 This procedure has 'a weakness in that light is available at all wavelengths and by the time •total decomposition• of TCDD is achieved other com­ pounds giving a false signal for TCDD also may have decomposed. A more specific approach is to measure decomposition rela­ tive to an isotopically labelled internal standard and to carry out the photolysis l) with light of very narrow wavelengths and 2) at more than one wavelength. A procedure was devised to achieve this. The apparatw;. is illustrated in Figure 23. Light from a high intensity mercury-xenon lamp was collimated and focussed, with a cylindrical lens, onto the entrance slit of a grating monochromator. Slits were selectea to give a bandwidth of 100 slit. I. The sample t� was positioned directly over the exit Photolyses were carried out in benzene at 314 nm (TCDD extinction coefficient 6000) and 289 nm (TCDD extinction coeffi­ cient 3200). The extent of decomposition was compared to that for 37c1 TCDD, which previously had been added to the sample for quantitation. At 314 nm the level of the compound present in a sample of Vietnamese fish decreased by 701 during 40 minutes, a time sufficient to decompose 691 of ·the 37c1 TCDD. At 289 nm the corresponding reductions after 110 minutes were 841 and 851. These results establish that the compound isolated from the Viet228E.A. Woolson, R.F.' Thomas, and P.D.J. Ensor,· J. Agro? Food �, ll, 351 (1972. 171a @ r�- Sample tube Monochromator FIGURE 23. Cylindrical lens Spherical lens Diaphragm Apparatus for monochromatic uv photolysis. UV lamp 172 riamese fish undergoes photodecomposition in a manner identical to that of TCDD and thereby indicate that the compound is not a cUphenyl ether derivative. 8. Partitioning Between Ac:etonitrile and Hexane Partitioning of Chlorodioxins between acetonitrile and hexane has been repo�ted as a method of confiJ:fflation,228 al­ though the method probably is not very specific. In the pres-. ent case, the r"'sidue from a sample of Vietnamese fish, which showed a positive signal for TCDD, was equilibrated between acetonitrile and hexane. A total of 681 of the· 37c1 TCDD and 611 of the compound in the Vietnamese fish was found in the acetonitrile. Within the error of the experiment this result is consistent with the unknown compound being 9. TCDD. Lack of Adsorption of TCDD on Glass During Storage It has been reported that TCDD is adsorbed onto glass, particularly from aqueous solutions.195 If TCDD were adsorbed onto the walls of the vial during storage of the original sample, the levels measured would be lower than those originally present. To test if this occurred with the Vietnamese samples a sample vial, which had been emptied of the fish it contained, was .ex­ tracted with ethanol followed by methylene chloride. The level of TCDD observed was approximately the same as that expected from the small amount of fish homogenate remaining. able amount of TCDD was adsorbed on the.glass. No measur­ 173 c. other 2,4,5-T Compounds Whenever certain Vietnamese fish samples which had not been put through a preparative ,,glc step were analyzed in the mass spectrometer, a signal appeared at the TCDD !V!. values with an appearance time about 1.8 times later than that of TCDD (Figur�. 24). In terms of exact mass and isotopic isomer distribution the signal was indistinguishable from that of 'l'CDD, which meant that it was an ion identical to or isomeric with TCDD. In the initial analyses of the 1973 fish samples the dif­ ference in the appearance time of this signal f.rom that of TCDD was not appreciated, which led to erroneously high estimates for the levels of TCDD in the 1973 fish in these early runs. Al­ though the late peak had already been discoverea when the multi­ ple ion detection procedure was developed, this procedure pro-. vided a satisfactory way of measuring the appearance time.of any ·observed peaks relative to that of 37c1 TCDD. The I.ate signal was not observed when the preparative glc step was used. Whenever neutral. extraction (Isolation Proce­ dure III) was used instead of saponification and treatment with acid (Isol.ation Procedure II), the signal. either disappeared or decreased by an order of magnitude. By stopping the analysis with the disappearance of the 3lc1 TCDD (as detel:ffli.ned by mul.­ tiple ion detection), the signal could be resol.ved from that of authentic TCDD which vol.atilized exactly with the 37cl. TCDD (Figure 23). The fact that the signal appeared at a time later Late peak seen when acid or base is used !.!!! glc ia not used -===�=:::::!::== o+::::::::t=:=;--._;_________ ___,r--_;,___ Addition of glc step elim­ inates late peak �...;..------,----------=.::::======:. ....GI ; o,;------r----;....-;....____ ! ·c. Neutral cleanup avoids late 'peak even without glc .......... o,+.===�=::::!=::::!:=:=::::...:_________�=�==�= D, 31c1 TCDD reference 0 2.0 1,0 0 Appearance time relative to 37c1 TCDD Figure� Ertecta ot cleanup procedure on TCDD signal. in a a11111ple ot Vietn11111eae fish, Scana A, B, and C are at m/e 321,894 and x50 sensitivity. Scan Dia at m/e 327,885· and xl sensitivity. Appearance · times h.ave · · 't!een normalized to that of 37Cl TCDD. 175 than the signal for 37cl TCDD indicated that it represented a fragmentation ion of sane less volatile compound. The presence of less volatile compounds was confirmed by · MS analysis at higher !!!I! values. As is shown in Table 24 for a sample of 1973 Vietnamese fish from the Dong Nai River, sev­ eral high molecular we_ight (BMW) chlorinated compounds were observed when the saponification and acid treatm�t was used. When the neutral extraction was used, only �one BMW compound was observed at similar levels (Table 24). '.rhe volatilization time of the!!!/� 322 '.rCDD •1ate peak• coinci� with that of the HMW compounds at y� 446 and 496. A 446 signal also was observed with the neutral extraction, but it did not give rise to such a large 322 late peak, and so the 496 signa.l was tentatively identified as the major source of the 322 l.ate peak. tilization times of the 41a, 460, and later than the 446 and 496 peaks. sto The vola­ peaks all were slightly Xt should be pointed out that the BMW s.ignals themselves may be fragmentation peaks. As was mentioned earlier, with the 1973 Vietnamese fish samples the neutral extraction procedure routinely was used.for TCDD anal­ yses to minimize the levels and number of high molecular weight compounds in the final residue. Several experiments were carried out to attempt to charac­ terize the HMW peaks. A rough estimate of their levels was made by using 37c1 TCDD as a reference. Levels for the 1973 fish mentioned above are shown in Table 24. Levels of the 446 compound for neutral cleanups of several different samples are listed in ... Table 24. High molecular weight peaks observed in a sample of Vietnamese fish collected at site A in 1973. Nominal!/! Exact mass (!/!,). 410 446 460 496 S10 409.911 44S.879 4S9.901 495.879 509.887 4 s 2400 Approx. level with saponification and agid extraction (Proc.II) (ppt) 8200 7600 Number of Cl's Probable molecular formula Approx. level with neutral ex�raction (Proc.III) (ppt) 4Based <50 s 6 6 2800 1800 1800 <40 <30 <20 on added 37c1 TCDD. CNRJfA.PJ1·%£lAif4?.#9h94F.Ji!&JQ&IA&AA1J.MRA#.lJl?liK¥MQ!flJJJ.J4W,¥M4&ShAii41All$$}!.Jfifo..1f.¥¥!¥M@ M14iJk¥Mtk¼J;&&UL&t . us. 177 Table 25. :rt appears that the levels of the BMW caupounds may have increased between 1970 and 197'3, but more analyses are required to.confirm this. In any. case the levels.of the BMW compounds in $cine of the 1973 samples appear to be well into the ppb range.· Accurate mass measurements were made on the BMW ions in a ,.·. 1973 fish sample. (Table 24) • The number of chlorine atoms was determine'1 �rom:the isotopic isomer pattern. 416 For the 446 and ions, the M+l/M cl3c) ratios were.measured. From these results the probable molecµlar formulas listed in Table 24 were detemined. The ions appear to be interrelated in a simple man­ ner. The 446 peak corresponds to loss of CH c1 from the 496 3 peak. Similarly the 460 peak corresponds to loss of cs3c1 from the 510 peak. The 446 and 496 peaks correspond to loss of CH2 from the 460 and 510 peaks respectively. ponds to loss of Bel from the 446 peak. The 410 peak corres­ The 446 peak and 496 peak volatilize together as do the 460 and 510 peaks. Treatment with diazomethane prior to MS analysis does not effect the ion intensities, which suggests that the ions are not derived from phenols. From the probable molecular formulas, the 496 and 510 ions must contain three phenyl groups in an open chain configuration. The 446 and 460 ions must contain three phenyl groups and an additional ring or other site of unsatufation. The 410 ion must contain three phenyl groups and two additional rings. Pi ... &,.**� ' 5 rtt:etem 178 TABLE 25. Approximate levels of the m/e 44i compound ain samples of Vietnamese fish (based on added 7cl TCDD). Sample 1970 1973 10 13 19 261 289 294 302 Site TCDD(ppt) m/e 446 (ppt) B 520 140 B 810 260 <6 C A A A C 70 ND(LD ND(LD ND(LD ND(LD --=· 30) 112) 40 ) 23) .>600 96 320 <8 Ratio TCDD/446 3.7 3.1 12.0 0.004 1.2 0.125 a.All samples were treated with Isolation Procedure III. ·, wow·:'i & ·:srtt x tr:& tr t ·--,. ·1c·1zrt f tnt V tm-tt" · iz1tz&c 179 Treatment of the 446 containing residue extracted via the neutra1 procedure with saponification and acid did not produce the other BMW compounds. The HMW peaks and TCDD late peak were not observed in samples of 1973 Vietnamese fish collected from a river,, the Mekong, which did not drain areas sprayed with 2,4,5-T. This suggests that the compounds observed are derived from 2,4,5-T. Additional analyses were carried out to attempt to charac� terize the TCDD late peak. the same y� value As mentioned before this peak had and isotopic isomer distribution as TCDD and h�ce appeared to represent TCDD or an isomer being formed as a fragmentation ion of some higher molecular weight compound. �he signal had the same appearance time as the 446 and 496 BMW ions. The !!V� values for several other possible precursor ions also were monitored, but no signals were observed at any of ,these !!I� values. The ions monitored are listed in Table 26. PLEASE NOTE: Thfs page not included in 111&terfal received from the Graduate Schoo 1. Filmed JS received. UNIVERSITY MICROFIL� 181 TABLE 26. m/111 Possible TCDD "late peak• precursors derived from diphenyl ether not bserved in 1973 Vietnamese fish • 2,2 -Substituents of 4,4',5,S'-tetrachlorodiphenyl ether -o -o.. -o. 336 350 355 356 366 370 516 534 -OH -OMe -OMe -OOC13 -ooc1 3 .o ­ MeO­ Cl� ClMeO­ Cl­ HO­ Cl- TABLE 27. Ratios for various samples of the signals observed at the exact masses (±. 3-4 mmu) of the isotopic isomers of c H c1 • 12 4 4 samele 'theoretical for C12H4Cl4 Authentic tetrachloro PCBene PCB Arochlor 12_54 Fisha Buman liver A b Human liver sb Human liver cb Human liver D b 287.907 289.904 77 100 100 100 100 100 100 100 100 74 71 78 74 "83 79 72 111/, 291.902 293.899 49 49 43 45 46 56 48 48 • Purchased at a local market. b Obtained from Massachusetts General Hospital, Boston, Mass. 10 11 10 11 9 9 10 11 182 D. Polychlorinated Biphenylenes In the mass spectrometric analysis of tissue samples for TCDD, a signal was observed near the !!!I!. �22 peak of TCDD. The signal was measured to be at !!Y!. 321.868, about 26 mmu lower in mass than TCDD, and corresponded to a possible molec­ ular formula of c12H3c15• In a commercial polychlorinated bi­ phenyl (PCB) mixture, Arochlor 1254 (Monsanto Chemical co.), a signal at the same !!Y.! value was.obseLved. The intensity of. the signal was about 11 of the pentachlorobiphenyl base peak at !!!I.! 324. The !!Y.! 321.868 peak, which corresponded to loss of H2.�rom peatachlorobiphenyl, was assumed to be a minor PCB frag­ mentation product. However, whenever an alumina chromatography step designed to eliminate PCB's was �ed, the peak persisted. With the removal of the PCB's it wu possible to ana.l.yze f?r the .higher isotopic isomer lines at !!Y.! 323.865, 325.862, and 327.859 that shoul.d be pr.sent for an ion of the formula c1. 2H c1 • Such 3 5 lines, with intensity ratios appropriate for c12H3ci5 in fact were present. There are three reasonable compounds that could pro­ vide the basic structure for a formula of C12H3Cl5. They are biphenylene, acenaphthalene, and ethynylnaphthalene: Biphenylene Acenaphthalene Ethynylnaphthalene 183 The observed compound is stable to concentrated sulfuric acid which rules out the ethynylnaphthalene. On the basis of the other available data either the biphen­ ylene or acenaphthalene are possible structures. However, on the rationale that the biphenylene structure more likely would be associated with PCB's, this class of compound was investi­ gated in more detail. Chlorinated biphenylenes in general are thermally and chemically stable 229 • 230 and therefore could conceivably form stable environmental residues. chlorinated biphenylenes have been synthesized, A number of 2 31, 232 233 , including 2,3,6,7-tetrachlorobiphenylene, the tetrachlorbi­ phenylene which has the same substitution pattern as 2,3,7,9� tetrachlorodibenzo-e-dioxin. A sample of this compound, which haa the formula c12H4c14, was obtained.234 This compound 229Barton, J.W., in James P. Snyder, Nonbenzoid Aromatics, Vol. I, Academic Press, New York, 1969, pp. 32-62. 230cava, M.P. and M.J. Mitchell, C clobutadiene and Rel­ ated Compounds, Academic Press� New Yer�, l967, pp. 255-316. 231Boulton, A.J. and J.F.W. McOmie, J. Chem. Soc. (1965). 2 549 232Brown, R.F.c., .o.v. Gardner., J.F.W. Mcamie, and R.K. Solly, Chem. Commun., 407 (1966) • 233 Brown, R.F.C., D.V. Gardner, J.F.W. McOmie, and R.K. Sol!Y,. Aust. J. Chem., �. 139 (1967). · U4Courtesy of Prof.. J.F.W. McOmie. 184 separated on alumina just like the c12H3c1 compound observed. 5 earlier in environmental samples. A search was then carried out for a c12u3c14 compound in several ·human l_iver samples, a fish sample, and in PCB Arochlor 1254. As shown in Table 27, in each case signals at the appropriate!!/� values and isotopic isomer ratios were observed. reagent blank. No signals were observed in a With ·the biphenylene as a refere�ce, the approx­ imate level of c12H c1 compound in human liver samples A and 4 4 B was 50 and 30 �b respectively (without correction for losses in re�). 'fhese results suggest that there are environmental resi­ dues of compounds with the general fonnula of chlorinated bi­ phenylenes or acenaphtalenes. Further experiments, possibly in­ cluding catalytic hydrogenation, are necessary to confirm the structure of these compounds. There is no toxicological data available on ·the chlorinated biphenylenes. However, these COJll­ pounds are the next step in the series going from dibenzo-E,­ dioxins to dibenzofurans, and if the biphenylenes retain even a small.. fraction of the toxicity of the dioxins and furans, an4 if ppb levels of these compounds are conunon, additional investi­ gation certainly would be justified. 185 E. The Mass Spectrometric Method The mass spectrometric method described here has sufficient sensitivity to detect on the order of 10-129 of TCDD. The use of high resolution mass measurement for two isotopic isomers provides a high degree of specificity with ·respect to the molecular formula of the observed ion. When the appearance time is taken into account additional speci­ ficity about the origin of the ion is obtained. Accurate quantitation is possible either through the addition of a known amount of TCDD or by isotopic ratio measurements with multiple ion detection relative to 37c1 labelled TCDD. Measurements are reproducible with�n � 10-201. The method retains the generality of normal mass spectrometric analysis, although it is particularly suited to detecting compounds containing mass deficient atoms such as chlorine, bromine, silicone, sulfur, etc. For purposes of perspective, the sensitivity of the present method may be compared to that of radioisotopic analysis. For TCDD labelled with 1001 14c, one pg would pro-· duce 3.7 decays per minute. For 100% 3H labelled TCDD, 580 decays per minute would be produced. For standard liquid scintillation counting techniques this is about an order of magnitude below the practical limit of detection for 14 c and an order of magnitude above that for 3H. Low resolution GC/MS has been used for low level measurements of a wide range of volatile compounds.235, 2 36 185a The limit of detection has generally been in th� range of Recently the sensitivity has been extended to the 1-5 pg range for TCDo,237 a sensitivity similar to that 0.1 - 1.0 ng. of the method described here. The low resolution technique, however, lacks the specificity of high resolution mass spec­ trometry. An attractive approach would be to combine glc and high resolution mass spectrometry. Another recently developed mass spectrometric technique which offers high sensitivity employs·an external ionization source at atmospheric pressure.238 The sample, in nitrogen carrier gas, is exposed to a 63Ni source cs-, 0.067 MeVJ. Ions formed (positive or negative) pass through a 25-� aperture into an ion lens system where they are separated from neutral molecules, focussed, and a _ ccelerated into a standard low resolution quadrapole mass spectrometer. The limit of detection was �eported to be on the order of 5-10 pg for 2.6-dimethyl-�-pyrone. Electron spin resonance spectroscopy of free radi­ cals heretofore has been one of the most sensitive analytical methods available. This technique has been used to detect a 235 Reference 219, Chapter s. ·. 236A. L. Burlingame, R. E. Cox, and P. J. Derrick, Anal. Chem. , 46, 248 R, (1974). 237 w. Crummett, per$onal communication. 238 E.C. Horning, M.G. Horning, D.I. Carroll, I. Dzidic,. and R.N. Stillwell, Anal. Chem., 45, 936 '1973). 185b large number of organic, inorganic, and biochemical free radicals.239•240 The limit of detection under good condi­ tions, however, is on the order of 10 11 radicals.241 Assuming complete conversion to the free radical, this gives a limit of detection for molecular weight 300 of only about 100 pg. At low concentrations of sample, interference from other paramagnetic species becomes important. This .is especially true for organic compounds in that most organic radicals absorb in the same region of the spectrum. Pulsed Fourier transfol;'III esr is presently being investigated and although the work is at a preliminary stage, there is some indication that an increase in sensitivity of about ten­ fold 1c11ay be possible.242 Another analytical method offering high sensitivity ·for certain compounds is laser fluorescence. With this method sens.itivity and specificity is increased substantially relative to conventional fluorescence analysis for several reasons: (l) the laser light source is many times brighter than other available uv sources, (2) sharp temporal, spatial, and energy resolution is available with coherent laser light, 2390.J.E. Ingram, "Biological and Biochemical Applications of Electron Spin Resonance," Plenum Press, New York (1969}. 240E.G. Jansen, Anal. Chem., 46, 478 R (1974). 241c.P. Poole, "Electron Spin Resonance: A Comprehensive Treatise on Experimental Techniques," Interscience . Publishers, New York (1967), p. 539. 242 s. Weissman, personal communication. 185c which minimizes phosphorescence and scattered light and which makes possible accurate measurement of the flourescence decay half-life, (3) half-life measurements combined with wave­ length measurements provide considerable specificity and make possible analysis of mixtures of different fluorescent compounds.243 The great potential of the method was demonstrated by the detection of 10-lBg of fluorescent species.244 Detec­ tion of 200 pg of aflatoxin dire�tly on chromatographic plates by means of laser fluorescence was described recently.243 The major obstacle to further improving the sensitivity was reported to be background interference. For liquid phase analysis this technique appears to offer the possibility of sensitivity well below the picogram level, if suitable sample and solvent cleanup can be achieved. In that there is con­ sider.able overlap between the fluorescence emission spectra of different organic molecules, the method may be susceptible to interference from trace amounts of impurities and therefore may be less specific than high resolution mass spectrometry. Thus, although other promising methods are being developed, the picogram level mass spectrometric method des­ cribed here represents one of.the most sensitive and specific analytical techniques presently available. At the moment, it is the method of choice for analysis of 'J!CDD. 243 M.R. Berman and R.N. �are, Anal. Chem., in press. 244R.N. Zare and P.J. Dagdigian, Science, ill, 739(1974). 186 EXPERIMENTAL MASS SPECTROMETRIC PROCEDURE Photoplate Detection. Initial experiments were carried out with a CEC-110 double focusing photoplate mass. spectrometer. were: Conditions ° source 200 , with the is.ample on a heated direct insertion probe that was programmed from 25-400° over 2-3 min, accelerat­ ing voltage 8 kV, ionizing voltage 70eV,. trap current 300 µA• resolution approximately 50�0 (10% valley)1evaporated silver bromide photoplate (Ionomet Co., Waban, Massachu.setts 02168). The limit of detection for TCDD under these conditions was about 100 pg. Quantitation was made difficult by the non- linear response curve near the limit of detection and by -the rather limited range, a factor of about twenty-five, between the limit of detection and saturation of the silver bromide. Analyses were also carried out using the electron multiplier of this instrument as the detector. Narrow scans were made over a m/e range of about 0.3 ainu centered on the m/e 321. 8935 peak of TCDD. In another experiment the de­ tector was fixed at m/e· 321.8935 and was used as a specific ion detector for TCDD. In both cases the limit of detection for TCDD again was on the order of 100 pg. Repetitive Scanning of a 20 m/e Unit Range. This and all following experiments were carried out with the MS-9 mass spectrometer. A sample of TCDD in a 187 capillary tube (Figure 2) was introduced into the mass spec­ trometer via the direct insertion probe and volatilized over 30-60 sec (source 200 ° ). The m/e range 319-330 was repeatedly scanned manually at a rate of approximately 5 sec/scan. -scans were visualized with an oscillographic recorder. The A mass reference was obtained by the introduction at' a constant rate of PFA, whi.ch has characteristic fragmentation.peaks at m/e 314 and 326. The limit of detection of TCDD (limit at which the m/e 320, 322 and 324 peaks could be detected) was about 100 pg (Figure 8). At resolution 10,000 (10% min)uver the coorse of a 4S sec analysis, the detector spends a total of about 0.072 sec in the region of the TCDD peak (taken to be 0.032 amu wide}. Repetitive Scanning of a 0.300 m/e Unit Range. As in the preceding experiment, a sample of TCDD was introduced into the MS-9 and volatilized over 30-60 sec. (200 ° ). .At a rate of 1.3 sec/scan the m/e ranges 313.8 - 314.l and 321.7 - 322.0 were scanned repetitively, two scans at the lower mass range, two at the higher, etc. (Figure 9). The scans were visualized with an oscillographic recorder. PFA was. introduced at a constant rate to provide a reference peak of m/e 314. At resolution 10,000 over the course of a 45 sec analysis, the detector spends a total of about 2.4 sec in.the region of the TCDD peak (taken to be 0.032 amu wide). 188 The l.inlit of detection for 'l'CDD with this procedure was about 20 pg. Multichannel Averaging with a Pulse Height Discriminator Trigger. The separate scans over a narrow mass range described in·the previous paragraph were combined with the aid of a Varian 10 24 time averaging computer (which is effectively a multichannel analyzer) to produce a single averaged trace.· In the first con- figurationinvestigated, the PFA peak at m/e 314 in one scan was used to trigger the analyzer to accumulate the following �can at m/e 32 2. Then m/e 314 was scanned again for the next trigger pulse, and so forth. The analyzer required a 2 volt pulse to trigger, but the PFA peak height in the MS-9 output corresponded to a few hundred mill.ivolts. A pulse-height discriminator was used to pn>duce a 3-volt trigger signal when it received a SO mv or greater signal in the MS-9 output at m/e 314. The input to the. computer was taken directly from the output o.f the MS-9 electron multiplier amplifier. During a 4S-sec analysis at 1.3 s�c/scan, the built in scan rate of the MS-9, a total of 17 scans could be carried out at m/e 322 (45-sec analysis x SOI of scans at m/e 32 2 x l scan/ 1.3 sec : 17 scans). Time averaging this number of scans theoretical- ly would provide a fouriold increase in the S/N ratio ( 17 '¥ 4) �223 .,. •Jilt ,e 189 Multichannel Averaging with an Internal Trigger. The Varian 1025 analyzer produces a sawtooth voltage ramp of Oto +25 V simultaneously with its memory sweep. This signal was amplified to Oto +100 V with a de amplifier (Krohn­ Hite Model DCA-lOR) and placed in series with a +45 V dry cell to provide an overall sweep voltage of +45 to 145 v. This saw- tooth signal was used successfully to replace the +SO to +150 V triangular wave ramp of the MS-9 circuitry which drives -both the horizontal sweep of the MS-9 oscilloscope and the small mag­ net coil used to deflect the ion beam.· back and forth over the O.jOOamu range being scanned. The +SO to +150.-V ramp of the MS-9 is produced at the anode of a vacuum tube (Vl) in a Miller integrator.stage of the peak switching unit. Thi,s tube was removed, interrupting the existing circuits, and the +45 to The +145 V analyzer ramp was fed into the anode output lead. ramp of the MS-9 is a triangular wave and there is a relay (RLA2 of the peak switching unit) which shuts off the oscillo- scope trace during the negative-going portion of the wave. When the analyzer sawtooth ramp (which has a nearly instantaneous negative-going flyback) was substituted,. the result was that every other trace on the oscilloscope was cut off. dition was alleviated by shortening RLA2. This con­ It was possible to repetitively scan only the •/• 322 region -- no scans at ml• 314 beyond initial tuning were necessary �- with the various rates available on the Varian computer. During these scans the sweep of the ion beam • ?¥04, !!:'li(\ZH&)J!L)k-i ,;,t¥Plff$ .. if!k -!l \.jlk _,.¥¥:P.Yh .. .$-¼ %".¼-::+¥ e . . u,,y,.;; 1,.¢ _..Z.1Pl§.91Ni4&.?U) ·, ..... A:W4__ ,.,� .&_.,z. Li.� ,.+1-<- .. ¥¾,!.J.(Jb�- *'- i90 in the MS-9 was necessary in synchronism with the analyzer memory. he sweep of Based on experiments described below, a scan rate of 4 scans/sec was selected as optimum. 45-sec analysis this provides 180 scans at m/e 322. Fora ·This in turn makes possible an increase in the S/N ratio on the order of tenfold. Improvement in Signal to Noise Ratio with Time Averaging. The increase in signal to noise ratio resulting from time averaging was measured for a PFA reference peak (m/e 315) under normal conditions of analysis. The signal to noise.ratio was defined as the signal (mm peak height) divided by the root­ mean-square noise (here taken to be 0.7 times the average peak­ to-peak value).223 Five runs each were made of one 25-sec and twenty-five averaged 1/sec scans and the improvement with time averaging was calculated. The increase in S/N ratio for twenty-five 1-sec scans relative to one 25-sec scan was 4.0±1.2. The theoretical inc�ease for 25 scans is 5.o.223 Optimization of Scan Rate. The response on the TAC-1025 analyzer for a constant scan period of thirty seconds was compared for scan rates of 1,2,4, The N2 peak at m/e as a standard reference peak. 10,20 and 40 scans/sec. 28.0061 was used .. ··.:m=::=. . 191 At scan rates of 10, 20 and 40 scans/sec the response was more than an order of magnitude lower than that at the lower scan This effect may have been caused by inefficient trans- rates. fer of information to the analyzer memory. To gain the maxi- · mum benefit from time averaging it is necessary to obtain the maximum possible number of scans, as long as the information from each faster scan rate is efficiently incorporated into the In the present case a rate ofJscans/sec was used. memory. Higher Sensitivity with Higher Source Filament Current. The maximum current no%111ally available nt the trap (or collector) of the ion source of the MS-9 is 0.5 mA. electron impact In an source more ions can be produced, thereby increasing sensitivity, by increasing the- current of the electron beam at least until the space charge associated with the beam becomes large enough to interfere with the focusing of the ions formed. In the MS-9 this current was increased io 1.0 mA and 2.0 mA by decreasing the values of. resistors in series with the source filament. The response for a PFA reference peak was measured under conditions of optimal tuning for each current. The PFA was leaked into the source· at a constant rate throughout the experiment. The mean value of our measurements of the response £or each current were calculated. Although a twofold increase in response was obtained in going from 0.5 mA to 1.0 mA, the increase in going from 1.0 mA to·2.o mA was not significant. Throughout the present investigation . ...,.,.-:· ·' ,_mzsr .,_ 192 a filament current of 1.0 ma routinely was used. The fila- ments used were rhenium wire 0.20 mm in diameter (No. 51804, Vacumetrics Corp., Waltham, Massachusetts). Sample Tube Preparation. Sample tubes were prepared from 1.6 - 1.8 mm x 100 mm borosil.icate glass open ended melting point capillaries (Kimble No. 34500) (Figure 13A}. The tubes were fused about 40 Jlllil from one end with a small flame to ·give the configuration shown in Figure 13B. (Fj,gure llC). The shorter side was then cut to 10 mm The tubes were stored in this form until use. To carry out an analysis the tube was loaded with the sample dissolved in 2-8 µl (depending on the type of sample, amount of calibration solution added, etc.) of benz�ne (Figure 13D). 'l'he tube.was then drawn out with a flame 5-6 mm above the level of the liquid (Figure llE and the solvent.was pumped off through the capillary constriction (to prevent bumping) at reduced pressure. After removal of the solvent, the capillary was sealed with a flame (Figure llF). Such sealed tubes could be stored several weeks or'more before analysis. Variations may occur during the constriction of the tubes prior to pumping off solvent. 'l'Wo extremes are charring of the solvent around the constriction and an inadequate constriction. Conditions were selected to minimize these effects. .i! f!if�4:.�\��2?:���¥:f::f.�:¥-¥ffffe--¥%�,.,. r'._#i�-�:t��-4:?%#1:ji1J.¥�t-��¾f'¾1t������rt4tq·r��¾':�#ffefu£%:f%i-: ::!.���P$lf&.A.*#·�*K�,'r 4 193 MS-9 Direct Insertion Probe Procedure. Each sample was placed with a 10 µl syringe into a tube made from a 1.6-1.8 mm i.d. melting point capillary (Kimax No. 34500) as described above. The tube was introduced into the MS-9 with a wire holder placed on the tip of a standard MS-9 .direct insertion p�obe. The probe was lowered into the source until the tip of the sample tube blocked the ionizing electron beam, as indicated by a sharp drop in the total ion current monitor, and pulled out until the ion current monitor just recovered� This configm.:-ation is illustrated in Figure 12. "l'O further aid reproducibility, all analyses were started at the same time {l min) after the probe was inserted into the source housing. The temperature of the source heating block was adjusted to give a TCDD volatilization time of 30-60 sec and ranged from 180-240 ° depending on the sample matrix. Effect of Sample Matrix on Response. Several experiments were conducted to test the effect · of the nature and amount of sample residue present along with TCDD during mass spectrometric analysis. As a preliminary ex­ periment O.l, 5 and 25 µg portions of squalane dissolved in ben­ zene were added to a number of Salllc)le tubes containing a fixed amount, 20 pg, of TCDD and the TCOD response was measured as a -function· of the amount of squalene. The results are sum­ marized in Table 14A. In going from Oto 25 µg the response :+std'"'tffiH('j]M ftfL�it:rrti,pt_L�)'·i:..�_--p'<� "!Jt'_ ' _t___ ' .?&! 11W"fr YrWTftti:t·mr · ... 'tfft_�_JM1fiflf'fi1tri t1· 1-,.·1rt ·• 1,�f" ,., ¥&"M'f 1 ·, t· f mntrttttl '-., 194 dropped by 851. Simi.larly the ratio of the response for 20 pg of. TCDD to the response for 200_pg of 2,3,5,6-tetrachloro-4bromethylbenzene was measured in the presence of 1,3, and 9 µg of squalane. The results are listed in Table 14B. The ratio 'was not constant and dropped from 11:1 at 1 µg to 1:1 at 9 µg. When a preparative glc step was included in the cleanup pro-· cedure, the effect of the presence. of 0,5,10 and 201 of the re­ sidue from a blank glc prep on the response for 20 pg of TCDD was measured. Up �o 201 of the residue from the blank glc prep had no significant effect on the response for TCDD. The higher amounts of residue did significantly increase the vola­ tilization time of the TCDD -- about 20 sec for 01 residue and 60 sec for 201 residu�. With the preparative glc procedure the effect of various amounts of beef liver residue on the response for TCDD was measured. To test different alumina chromotgraphy procedures, two experiments actually were carried out. In one experiment the alumina step involved elution with 11 dichloromethane in hexane followed by 201 dichloromethane in hexane. In the second experiment the alumina step involved elution with 201 carbon tetrachloride i� hexane followed by 201 dich1oromethane in hexane. Figure .17 illustrates the effect of various amounts of beef liver residue from these two cleanup procedures on the response for 20 pg of TCDD. With the 11 dichloromethane pro- cedure, _above a sanple size corresponiin;J to about lg of residue per analysis, ,.,.VP, : 2 I . tr' SIT t Hti ., 195 decrease in response cancels the effect of increasing the sample size. For the 201 carbon tetrachloride procedure, which was selected for actual use, the residue from as much as 4 g of beef liver could be included in an analysis without seriously depressing the response for TCDD. When the prepara- tive glc step was replaced by a second alumina chromatography step, similar experiments were carried out. The response for 100 pg of TCDD was measured in the presence of no residue and in the presence of 101 (the fraction used in routine analysis) of the residue froa a blank run of the final alumina chroma.tography step. The alumina chromtography residue had no sig- nificant effect on the response. The response for 40 pg of TCDD was then measured in the presence of the residue from 0,0.6,2.0, and 6.0 g of whole fish homogenate. Above 0.6 g the response decreased with increasing sample size. The slope of this decrease, however, was only about 1:3, so that up to at least 6.0 g and possibly beyond increased sensitivity could be gained by increasing the sample size. Effects of Solvent Evaporation, Adsorption and Water on TCDD Response. The following experiments were carried out to measure sources of error that might have an effect during sample handling. Two 18-1,.11 benzene solutions of 120 pg of TCDD-{Ct 37 )4 were prepared and each was fractionated and loaded into six 3-µl 196 sample tubes. .In both cases the sixth 3- µl fraction was either negligible or nonexistent because of evaporation of the solvent during preparation of the earlier fractions. However, no increase in the concentration of. the later fractions which might be anticipated as the result of evaporation of solvent, was observed. Under some circumstances TCDD reportedly is adsorbed To test if this might from solution onto glass surfaces.195 be important under the present conditions of sample tube load­ ing, a 20 pg/µl benzene solution of TCDD was prepared and di­ vided intc;, two portions, one of which remained in the original vial while the other was exposed to three additional vials for 15 min each. The concentration measured for the solution exposed to three additional vials, 20 ! 2 pg/JJl, was not dis­ tinguishable from that of the first solution. Similarly, over a six month period the concentration of a 1 ng/ 1 c1 37 TCDD benzene solution changed.less than± 101. In some cases solu- tions used frequently for spiking became somewhat more concen­ trated after several weeks of use. This effect probably was· caused by loss of solvent through evaporation. For example a TCDD:solution with an original concentration of 20 pg/µl after six weeks of regular use (251 of initial volume left) had a concentration of 34 pg/µl relative to a freshly prepared solution. The effect of such a concentration change is to make 197 measured levels of TCDD in samples lower than the true values. 'l'o test whether water, which might accidently condense in the flask during the final microconcentration, has any effect on TCDD response, the response was measured for 20 pg samples of TCDD prepared from dry and water saturated benzene solutions. Within± 101 no difference was observed. Linearity of the Response for.TCDD During Mass Spectrometric Analysis. Response curves for TCDD were obtained both with and without_the presence of residue from a tissue sample. the first case sample tubes containing and 27 pg of TCDD were prepared. o, In. 0.5, 1.5, 3.0, 9.0 'l'o minimize the effect of any slow change in the response of the mass spectrometer the samples were run in the order 9.0, 3.0, 1.5, 0.5, o, o, o, 0.5, 1.5, 3.0, 9.0, 27, 27, 0.5, 1.5, etc. They were in the se- quence given so that a tube containing a low level was never run· immediately after a tube containing .a high level, thereby minimizing the effects of "memory" or residual sample, in the spectrometer. The results plotted in FigurelSAindicate that the response is linear. In the second case an identical ex- •• periment was performed with sample tubes containing 5, 15, 45 and 135 pg of TCDD, except that each sample tube also contained an amount of residue from beef liver corresponding to 0,5 g of 1,\-' .�_.,A ,:.:t ,n/$..,4 ., .p,-41 .;.:,5_ ;a 4!-2.. g ;, .. 1¾$,Z.J,!#4.-.*.¥ ,, ,;. (� t.. FWA:U,¥#L4k? t., ¥ fllif ._._mp 8m� 7'1'�,,���c:-�-���-T��� 198 original sample. Once again (Figure 133) the respO"'\�e was . linear. Quantitation by Addition of Authentic TCDD. Fractions for 35c1 TCDD measurement normally each con­ tained 101 of the total residue while fractions for 37c1 TCDD recnvery measurement each contained only 0.6% (see Isolation Procedures), and since the larger fractions introduced back­ .ground into the mass spectrometer which temporarily inter­ 37 fered with measuring c1 TCDD recoveries with the smaller 37 samples. cl TCDD recoveries always were measured before 35c1 TCDD levels were determined. Recoveries of 37c1 TCDD were measured at m/e 327.885. For each sample three fractions were run to which no addition­ a1·37c1 TCDD was added and three fractions were run to which a known amount of 37c1 TCDD (normally 90 pg) was added. The mean peak height (less bcakground) was calculated in mm for t�e unspiked and spiked fractions. The difference in the mean peak heights was determined and this difference was di­ vided into the amount of TCDD added to the spiked fr.actions. The result is a calibration factor in pg/mm for the response caused by the known amount of added TCDD. Since the response to TCDD is linear (if the sample matrix is held constant), the peak height for the unspiked fractions was multiplied by this calibration factor to give the amount of TCDD in that caused the response in the unspiked fractions. < _,Jhaliiii!#l�Wai.W-ati.i�1111,: -.tat{¢lt'r\£ti'Q . .,�dtliiii'a·i-UW··�·,Rewi:Nt-i t:i{tifwkif ¼tufit "t�t:»r#tttWft ·--wt}t@ t r(J(· . , . 199 amount of 37c1 TCDO in pg per fraction was then divided by the number of pg per fraction expected for 1001 recovery (con­ centration of 37c1 TCDD added to sample before cleanup x grams of sample/number of fractions) to give the recovery of 37c1 TCDD. For each 35c1 TCDD analysis at m/e 321.894 a similar procedure was followed, except that the known amount of 35cl TCDD added for calibration was either 3-4 times the amount of · ·TCDD expected from the sample residue or 40 pg, whichever was larger. To obtain the concentration of 35c1 TCDD in ppt in the originai sample, the calculated number of pg of 35c1 TCDD per unspiked fraction was divided by the number of grams of original. sample represented by each fraction. To confirm a 35c1 TCDD peak observed at 321.894, an additional fraction Any authen­ was run at an isotopic isomer peak at m/e 319.897. tic 35c1 TCDD peak observed at 321.894 will have associated with it an isotopic isomer peak at m/e 319.897 with an inten­ sity ratio of o. 77. Procedures were developed to take into account inter­ .ferences which might have affected the measurement of 37c1 TCDD recovery at m/e 327.885. One 37c1 TCDD fraction was also run at m/e 325.888. If the peak a t m/e 327.885 was caused by only 37c1 TCDO, then, since the 37c1 TCDD actua�ly contained 95.51 37c1, the ratio of the peak at m/e 325.888 to the peak at m/e 327.885 was 0.17. . . 6'#¾fr"f ,-,::r«M¥fftdttrit:. · 1 e ·tr ··:., If pentachlorobiphenyl (which could 0 200 not be resolved from TCDD at resolution 10,000) or 35c1 TCDD contributed to the signal at 327.885, this ratio was larger. If the 325.888 to 327.885 ratio was greater than 0.22, a correction was applied to the peak height attributed to 37c1 TCOD at 327.885. For the case in which only pentachlorobi- phenyl is interfering , the correction is given by the follow­ ing equation based on the isotopic isomer distributions of pentachlorobiphenyl and 37c1 TCDD: Peak height at ril/e 327.885 caused by 37c1 TCDD • 0.73 {(l.54)(observed peak height at m/e 327.885) observed peak height at m/e 325.888)} This correction was only occasionally applied in the analysis of the Vietnamese samples. For the even rarer case in which 35 both c1 TCDD and pentachlorobiphenyl are interfering, the correction is given by the following equation based on the · isotopic isomer distribution of 35c1 TCDD, pentachlorobiphenyl and 37cl TCDD: Peak height at m/11 327.885 caused by 37c1 TCDD = f(observed peak height at m/e 327.885) (0.65)(observed peak height at m/e 325.888) + (0.055)(observed peak height at m/e 321.894)1 /0.889. Since no interference of the sort described above was 201 was ever observed during analysis of 35c1 TCDD, no corrections were made in determination of 35c1 TCDD levels except for subtracting occasional low level background. Modification of the Varian TAC-1024 Analyzer for Multiple Ion Detection. The Varian TAC-1924 analyzer contained circuits for ac­ cumulating data in two different sets of.channels (0-512 and 513-1025), but there was no way to automatically switch the in­ put back and forth between the two sets in synchrony with the peak switching circuit of the MS-9. To achieve such automatic switching, a relay that was activated by the MS.-9 peak switching relays was introduced into the TAC-1024 circuitry. In this manner signals at two different m/e values could simultaneously be averaged in separate sets of channels on the TAC-1024. This system was used tQ carry out isotopic dilution measurements of TCDD levels (see below). Quantitation by Isotopic Dilution. At the start of the isolation procedure th� sample was weighed and a known amount, usually 10-15 ng, of 37c1 TCDD was added. The sample was then cleaned up as usual. During mass spectral analysis the TCDD peaks at 321.894 and 327.885 were scanned alternately with the peak switching (multiple ion detection) circuit of the MS-9 and stored in two separate, sets of channels in the TAC.-1024 analyzer. ··-\v/--·-. t' 202 • 'l'he concentration of natural TCDD present was calcu­ lated from the following equation: Ratural TCDD level (pg/g or ppt) = {x/y)(2.00) c 37c1 TCDD original spike (pg)/original sample weight (g)) -where x = Peak height at 321.894 contributed by natural TCDD y • Peak height at 327.885 contributed by 37cl 'l'CDD. Given a relative 35c1; 37c1 abundance of 0.7 57 to 0.243 for natural '1."CDD and 0.045 to 0.955 for 37c1 labelled 'l'CDD, X - a - b/2778 y• b - a/117.6 where a• Total peak height at 321.894 b • TOtal peak height at 327.885 :In practice the a/b ratio is usually such that x ":! a and y ':! b. :In this case peak heights a and b can be substituted direct ly in the equation for the level of natural TCDD. The factor of 2.00 appears in the_ectuationbecause the intensity for a given amount of 37c1 TCDD.at 327.885 is twice that.for the same amount This results from the fact that of natural 'J.'CDD at 321.894. 35 37 c1; c1 abundance of 0.757 to 0.243 in natural TCDD with a the intensity is more widely dis�ributed among the possible isotopic isomers. If necessaey, correction factors for PCB interference can be-introduced as 4escribed in the previous section. J-- ' :_-­ . . -•.....-� ""'"!!" 3 f"'!,)l!!!' -- 4""".· ·•""' '.""""· '"",>t · -� "".. !"lL""3:!'"l'�f4"'lJ-"'l- A?!\'>:5""-·* _,,,._;:;;..,,#�- .\"1'>:�" -1"'?��"11:t��-..¥!!" _ '-1"-'!-!i }'""--?�,, .-"1'1<'.lrl!"'_', "1!-.':")?"'-'""!¥""'".•""·f""t.""'3t!"".'"! ""· 4J\I"!,. ......4!!'1'..*!!'!-...,,+f . .u"".'!'!JG"" ____..,.p,..�:.."'-""""""""·-- """"-·.l!!0\ ... .w.c ,... - '"'f"""'ff¥'"''� · � ¥ ,% "" , ..,t,.,m)4"' ¥ . "''.�"'",j."",;1 -¥.4""(""*:�T""'""-il*"'·-�"' "'¥,i#\""') -- ""->"'�.,.,,¼-- 203 Calibration of TCDD Stock Solutions Used for Ouantitation. Levels of TCDD in samples were determined on the basis of addition of aliquots of known volume of TCDD or 37c1 TCDD stock solutions. in two ways. Concentrations of the stock solqtions were determined Concentrated storage solutions (10-100 �g/ml} were measured on a tJV spectrometer and concentrations were determined from known extinction coefficients.26 More dilute solutions, used for actual quantitation. of samples, were cali­ brated relative to a known amount of 14c TCDD. '.the s�cific activity of the 14c TCDD was calculated on·the basis of its isotopic isomer distribution as determined by MS. '.the pattern observed revealed that the 14c TCDD solution con­ unlabelled TCDD and 461 14c labelled TCDD. The labelled material contained 39. 9.1 14c and had a specific sisted of 541 activity of 214 mci/mmole. The overall mixture contained 18.41 14c and had a specific activity of 98 mci/mmole. ( The literature value for this material was 148 mCi/11U11ole.32b} The concentration of the 14c TCDD stock soluti�n was determined accurately by liquid scintillation counting and the specific activity calculated above. Analysis of isotopic ratios in a mixture of known volumes of 37c1 and 14c TCDD solutions by multiple ion detection then provided the concentration of the 37c1 solution. 204 SYNTHESES AND REACTIONS Dibenzo-p-dioxin A literature procedure 38 was followed with some modifications. ; ·A··well mixed slurry of 2-chlorophenol (5.0 g, 39 mmole) (Aldrich 98+1, as rec.), Na2co3 (4.1 g, 39 mmole) and Cu powder (0.25 g, 3.9 mmole) in a 100-ml, three-necked, 14/20 RB flask with a reflux condensor was heated in an oil bath at 175 ° . gas. After From 1- 3 hr. there was visible evolutio� of 7.5 hr. the oil bath was removed and when the dark residue had cooled, 50 ml of ether and 25 ml of 10% aq. KOH were added. The phases were separated and the aqueous phase was extracted with 20 ml of ether which was then combined with the original ether solution. The ether was extracted with three 15-ml portions of 10 aq. KOH, dried over 5 g anhydrous Mgso , and filtered to give .. a clear yellow solution. The 4 solution was concentrateq to dryness at reduced pressure. The residue was dissolved in a minimum amount of benzene and The chromatographed on Al2o3 (25 gin a 15 mm id column). column was eluted with 75 ml of benzene (at which time no more residue was observed in evaporating the eluant solution). The benzene was removed at reduced pressure and the pale yellow residue was sublimed (20 mm, 140 ° bath) to give 0.5 g (14%, lit.38 10-20%) of colorless crystals, mp. 120-122 ° ). The infra- red spectrum was identical with that reported for dibenzo-p-dio�in.77 �le analysis indicated a purity of 9 8 +%. ! · · c,·· iit:'t ts ·r "T rY: · < rm e- r-· 205 2, 3 ,7,8-Tetrachlorodibenzo-p-dioxin-( 37 c1)4 The chlorination of dibenzo-p-dioxin with carried out in the apparatus shown in Figure 25�, c 37c1)2 was Dibenzo-p­ dioxin (1.2 mg, 6.5 µmole), iron powder (9. 3 mg, S µ11ram atom) (Baker); and iodine (0. 3 mg, 1.2 µmole) in 100µ1 nitrobenzene were placed in the predried reaction tube. A connecting tube packed with granular anhydrous Na2 so4 was attac�ed to the reaction tube. Fuming H2 so 4 ( 3 01 so3 , 450 µl) premixed in an ice bath with 301 a2o2 (150 µ1) was placed in one arm of the Y-tube. · Sodium chloride- 3 7c1 (5.0 mg, 8.6 1,1mole) (95.981 3 7c1, 4.021 35c1, from Oak Ridge National Lab.) was placed in the other arm. The tip of the reaction tube was cooled in liquid ni_trogen and the Y-tube was rotated to permit the a2 so4 /a 2 o2 solution to flow into the arm with the Nac1- 3 7c1. After After 15 min •. visible evolution of c12 gas had stopped. 30 min. the Y-tube was removed and replaced with a connection to a water aspirator. The reaction tube was then sealed (20 mm pressure with the tip of the reaction tube was then sealed (20 at all times) and suspended over refluxing methanol for 20 hrs. The tube was then opened and 200 1,11 of benzene were added. Glc analysis of the solution indicated dibenzo-p-dioxins in the following ratio: 37 unreacted starting material : 10 monochloro: 25 dichloro : 1 trichloro, with, no visible tetrachloro. 'l'Wenty- five percent of the benzene�nitrobenzene solution was transferred _£ .11UEUA4M 205a Y-tube. ---- Dtbenzo-p-dtoxtn, iron powder,---­ lodlne in nttrobenzene · · . 37 Figure 25. Mtcrochlorinatton apparatus used to prepare TCDD-(Cl · )4 .• lft&A §.4. �� W{§tsfQg:;;z. . QWWW�& ,.£ MJLt:Z-BLi JJJ44Ji@�£¥J_t;;.W#.6¥'Jk .§ib2"'!:v,..1. U d Q�;afu. �t-RfflJ•Btif�r.gg.;;:..$$4,;.-¢.A.,.!J!,.,.!¥4.J¼ .41�, - ·11:2: 206 to a new reaction tube, tresolvent was removed at reduced pressure, and the reaction was repeated as above. This time glc analysis indicated di­ tri- : tetra- : pentachloro- dibenzo-p-dio�in in the ratio 1 1: 100 : 25 with no other Mass spectroscopic analysis indicated dioxins visible. TCDD-( 37 c1)4 was present in 95.51 isotopic purity. Isotope peaks were observed at m/e 327. 8840 (theoretical 327. 8853) and at 325. 8870 (theoretical 3 25. 8832). + A. major fragmenta­ tion peak was present at m/e 263 (M - COCl ) at 191 of the base peak (211 in the naturally occurring compound) a nd a doubly charged ion was present at m/e 164 (M2+) at 91 of the base peak _(101 in the naturally occurring compound. The .benzene/nltrobenzene solution (500 µ1) was extracted with lN NaOH (100 µ1), filtered through powdered Na2co3 (100 mg) in a disposable pipette, and pumped off (through a trap cooled in liquid nitrogen) to give a film of light-brown crystals. The crys.tals were dissolved in benzene (800 µ1) and glc analysis indicated a concentration (assuming an FlD response identical to that of the naturally occurring compound) of 48 ng/µl for a total of 38 µg (81 yield). Synthesis of 2,3,7,8-Tetrachlorodibenzo-p-dioxin by Pyrolysis of 2,4,S-Trichlorophenol. A mass spectrum reported for 2, 3 ,7,8-TCDD by this procedure 25b prepared differed significantly in the extent of 207 fragmentation from the mass spectrum that we had obtained from TCDD prepared by chlorination of dibenzo-p-dioxin.102 'l'he synthesis by pyrolysis therefore was repeated. In a borosilicate glass tube (3 mm id) were placed 2,4,S�trichloro ,;. phenol (2.0 mg, 10 ·µmole) (Eastman Organic Chemicals, as rec.), K2co3 (1 mg) and copper powder _(O.3 mg). The tip of the tube containing the reagents was heated at 240-250 ° for 1 hr. Mass spectral analysis of the product gave a 2,3,7;8-TCDD spectrum that agreed with the 2,3,7,8-TCDD spectrum we had obtained previously, not with the sp�ctrum reported in the literature for this procedure. The yield, estimated by mass spectrometry,. was about 11 (£!, reported yield of 60i 25b). The anomalous spectrum was later attributed to the presence of contaminants in the mass spectrometer which catalyzed the fragmentation of 'l'CDD,97 2,3,S,6-Tetrachloro-4�bromethylbenzene. In a 100-ml, three necked, 14/20 RB flask equipped with a reflux condenser were placed 4-bromoethylbenzene (18.4 g, 0.10 mole) (Chemical Samples Co., 991, as rec.), antimony pentachtoride (l,Sg, 5 µmole) and 40 ml of carbon tetrachloride. The flask was cooled. in a water-ice bath and chlorine gas was slowly bubbled through the solution. After 30 min. the ice bath was removed and the reaction was continued at room temperature 208 for another 5 hr. On completion of the reaction the solution was extracted with three 20-ml portions of lN NaOH followed by three 20-ml portions of lN HCl and filtered through Most of the carbon MgSO4 to give a clear yellow solution. tetrachloride was distilled off and on cooling white crystals separated. The crystals were recrystall:iz�d twice from hexane to give 4.9 g (151 yield) of white crystals, mp 61-62 ° . Mass spectrometric analysis confirmed the structure of the product. A molecular ion was present at m/e 305 (base peak). The Sodium Salt of 2,4.5-Trichlorophenoxyacetic Acid. A solution of 1.0 g (3.9 µmole) of 2,.4,S-trichloro­ phenoxyacetic acid (Eastman Organic Chemicals, practica.l, as .received) in 20 ml ofethanol 2.0 g of NaOB in SO mlethanol was poured into a solutdon of The white sodium salt of 2,4,S-trichlorophenoxyacetic acid immediately preciptated. The salt was filtered, washed with three 25-ml portions of ethanol followed by two 25-ml portions of acetone, and dried to give o.eo g (2.9 IJ.IDOle, 741 of a white solid (dee. 275-290 ° ). Pyrolysis of 2,4,5-Trichlorophenoxyacetic Acid and of Sodium 2 1 4,5-Trichlorophenoxyacetate. 2,4,5-Trichlorophenoxyacetic acid (15 mg, 60 µmole) (Eastman Organic Chemicals, practical, as received) was placed in an open 7 mm x 250 mm borosilicate glass tube with a glass 209 wool plug and heated in a molten lead bath at 450 ° for 30 min. Only the lower 20 mm of the tube were placed in the bath. Charring occurred in the bottom of the tube and white crystals condensed on the glass above the bath. After the tube had cooled, one ml of be�zene was added to the tube, refluxed for 2-3 min on a steambath, filtered through 30 mm of Na2 co 3 in a 7mm x 150 mm disposable pipette, and analyzed by glc. No '.rCDD peak was observed which indicated a yield of TCDD < 0.011. Sodium 2,4,5-trichlorophenoxyacetate (.15 mg, 54 1,1mole) (see above) was pyrolyzed in an identical manner. Glc analysis indicated the presence of TCDD in a yield of about 0.11. The peak was confirmed by coinjection with authentic TCDD. - Glc analysis of a benzene extract of unpyrolyzed sodium 2,4,5-trichlorophenoxyacetate showed no indication of TCDD (<0.011). Mass spectrometric analysis confirmed the presence of TCDD and also revealed the presence of pentachloro­ dibenzo-p-dioxin and tetra- and pentachlorodibenzofuran. A large trichlorophenol peak and a weak dichlorophenol peak were also obs_erved. Pyrolyses of. sodium 2,4, 5-trichlorophenoxy- acetate were carried out under a number of other conditions in which both time and temperature were varied. TCDD was in- variably produced, but the yield never differed much from 0.11 210 Pyrolysis of a 1:1 Mixture of n-Butyl 2,4,S-Trichlorophenox­ .acetate and n-Butyl 2, 4-Dichlorophenoxyacetate (Herbicide Formulation Orange). A l�O g portion of formulation Orange was heated at reflux for 20 min. in a 7 inm i.d. x 200 mm borosilicate glass tube. Considerable darkening occurred. Ten ml of ethanol was added and the solution was transferred to a 125-ml Erlenmeyer flask. Twenty ml of 401 aqueous KOH (to saponify the esters) and 6 ml of benzene were added. After 5 min,60 ml of ,,ater was added with stirring and the aqueous phase was pipetted off. Sixty ml of 101 aqueous KOH was added with After 2-3 g of powdered Na2 co3 was added the benzene was analyzed by glc. No TCDD was det�ted stirring and pipetted off. ( <0.011). A 250 mg portion of formulation Orange in a similar tube was heated with an open flame until about half the material had charred to a black solid. After the tube cooled 4 ml of benzene was added, heated to reflux on a steam bath, and Twenty ml of transferred to a 125-ml Erlenmeyer flask. ethanol, 10 ml of 401 aqueous KOH, and 60 ml of water were added with stirring. After the aqueous phase was pipetted ; off, the benzene was dried over 2-3 g of Na2 co 3 and analyzed No TCDD was observed ( 0.011). by glc. Te··· "t rs -. 211 ISOLATION PROCEDURES Isolation Procedure I. Saponification, Treatment with Sulfuric Acid, Alumina Chromatography, and Preparative Glc. 1. Weigh the sample and homogenize it with 1.0-1.2 parts EtOH. ·(One part will refer throughout to the original wet weight of the sample.} 2 •. Transfer the homogenate to a round bottomed (RB) flask equipped with a reflux condenser (Teflon tape Spike the is used on the ground glass joint). 37 cl TCDD (more if large sample with 10 ng of ·amounts of PCB are known to be present), add 2 ·parts 3. c·• 401 aq. KOH, and reflux for 2 hrs. Let the solution partially cool and add l part· he}!:ane. . Transfer the solution to .a separatory funnel, separate phases and extract with three more identical portions of hexane. s. Collect the hexane in the original. RB.flask. Transfer the hexane to the separatory funnel, rinse the RB flask twice with a few ml of EtOH and then twice with a few ml hexane, refl.uxing solvent each time, and extract the haxane with l part l.0N 6. KOH. Extract the hexane four times (or· until acid phase is colorless) with 2 parts 95-971 e2so4• Break up emulsions with a few drops of saturated Na2co3 solution. ' �::; __ , ..._.: '· 212 7. Extract the hexane with l part water and add 8. several grams of Na2co3 to the hexane. Filter the hexane through a column of Na2co3 (SO.mm in 10 mm id column for 100 ml hexane). 9. Prewash Na2co3 with several ml of hexane. Concentrate the hexane (Snyder column) to 3-4 ml. Allow sample to cool. ·10. Chromatograph the hexane residue on a 50 mm column of Fisher A-540 alumina (activated at 130 ° for 24 h r) in a 7 mm x 150 mm disposable pipette. Prepare the column dry. U. Do not prewash. Elute wit� 12 ml of 201 cc14 in hexane, then 1 ml of hexane, and finally 4 ml of 201 CH2c12 in hexane. Concentrate the 201 CH2c12 fraction carefully to about 50 µ1, add 100-200 µl benzene and concentrate again to 20 µl. 12. Add a few µg of m-terphenyl in benzene to the residue and preparatively chromatograph by glc (51 SE-30 on 60/80 Ch�omosorb w, 2 m x 2 mm id stainless steel column; injector 280 ° , column 225 ° , detector 280 ° , Be carrier 30 ml/min, thermal conductivity detector). The retention time of m-terphenyl relative to that of 'l'CDD should be determined before hand and used to make certain that the TCDD collection is carried out .. :---.o. . . ..._. - ., . . ... .. --.�� -- 213 at the right retention time. The glc trap (see Apparatus and Reagents) should be wetted with hexane a� cooledin dry ice immediately before injection of the sample. 13.. Elute the glc trap with 60 µl followed by 10 µl of benzene. 14. Measure the total amount of residue and calculate the fraction size in µl for the planned number of fractions (typically nine fractions for 35c1 TCDD analysis and one fraction divided into 17 subfractions for measur­ ing 37c1 TCDD recovery). 15. Prepare the fractions for 'l'CDD analysis in the sample tubes described previously (see above). 16. Analyze the fractions with the MS-9. ditions are: Typical con- ° source 220 C, resolution 10,000 (based on a 101 valley betwee·.1 peaks), trap current 1.0 mA (rhenium filament), el�ctron multiplier 3.1 kV, ionizint voltage 70 ev, time averaging at four scans per second. 17. Calculate the recovery of 37c1 TCDD and the level of 35c1 TCDD (see Quantitation Procedures). 214 Isolation Procedure II. Saponification, Treatment with Sul­ furic Acid, and Alumina Chromatography. Carry out steps 1-9 of Isolation Procedure I and continue as follows: 10� Chromatograph hexane residue on a 40 mm column of Woelnl neutral alumina, activity grade 1, in a 7 mm x 150 mm disposable pipette. To prepare the column, submerge the lower part of the pipette in hexane .in a 10-ml graduated cylinder and add the alumina throu�h a funnel. Rinse RB flask with one ml of hexane and add this to the column. of carbon tetrachloride. discarded. 11. Elute with 8 ml This .fraction may be Then elute with 4 ml of dichloromethane. Add 4 ml of hexan.e to the dichloromethane fraction and concentrate this solution to about 200 µl in a microconcentration flask with a 10 cm 14/20 Rinse vial which held dichloro- Vigreaux column. methane fraction with 1 ml of hexane. Add this hexane to the microconcentration flask and concentrate again to about 200 µl. 12. Allow sample to cool. Chromatograph the residue on 50 mm of alumina in a Rinse micro- 1.5 mm id x 250 mm column. Load dry. concentration flask with 60 l of hexane and add this to the column. tetrachloride. methane. Elute with 500 µl of carbon Then elute with 350 µl of dichloro- Collect the dichloromethane fraction in a 215 clean microconcentration flask. 13. Add 350 µ1 of benzene to the dichloromethane fraction and concentrate (Vigreaux column) to 60 µ1. Add 200 µ1 of· benzene and again concentrate to 60 _µ1. The Vigreaux column may be removed in the latter stages of concentration. Continue with steps 14-17 of Isolation Procedure I. Isolation Procedure I�I. Neutral Extraction with Methylene chloride and Alumina Chromatography. l. Weigh the sample (up to approximately 10 g) and combine it with 1.0-1�5 parts Na2so4 and 4.0 parts ca2c12 (One part will refer in a homogenizing flask. throughout to the original wet.weight of the sample. 2. Homogenize the sample for 5 min (setting 40 on a Virtis 45 Homogenizer) and transfer the ca2c12 Repeat the (pipette or syringe) to a clean flask. 3. extraction four times with fresh ca2c12• Add o.s-1.0 g Celite to the combined ca2c1 2 extract and filter it with suction (water aspirator) through a sintered glass funnel. 4. Elute the CH2c12 through a column (lS x 80 mm for Elute 80 ml of ce2c12) of WOelm neutral alumina. Before use the an additional 2 parts of CH2c12• column should be prewashed with 4 parts of fresh ce2c12• 216 S. Concentrate the ca2c12 to about S ml (Snyder column)i add 30 ml of cc14 and concentrate to S ml, again and 30 ml of cc1 and concentrate to 5 ml. 4 Chromatograph the residue on 15 x 200 mm column 6. 7. of Woeimneutral alumina (made up in ccl ). 4 Elute with 70 ml of cc1 , SO ml of 101 cn2c12 in 4 . cc14., and 75 ml of ca2c12• Concentrate the ca c1 fraction (Snyder colwnn) 2 2 Transfer to a microconcentrator, add to 2-3 ml. 2 ml benzene and concentrate to about 60 µl. Add 200-300 µl benzene and concentrate to 60 µl. Continue with _steps 14-17 of Isolation Procedure l. Isolation Procedure for Polychlorinated Biphenylenes in Tissue Samples. carry out steps 1-9 of Cleanup Procedure for TCDD and continue as follows: 10. Chromatograph the hexane resldue on a 35 mm column of Fisher A-540 alumina {activated at 130 ° for 24 hr.) in a 7 mm x 150 mm disposable pipette. the column dry. Do not prewash. Prepare Elute with 6 ml of 201 dichloromethane in hexane. 11. Make up fractions from the 20% dichloromethane in hexane solution containing an amount of polychloro­ biphenylene suitable for mass spectrometric analysis 217 (fractions corresponding to 0.015 g each of original sample were used for domestic -liver samples). o. S. human Add an appropriate amount of polychlorobiphenylene (e.g. 2,3,6,7-tetrachlorobi­ phenylene} to some fractions for quantitation. Analyze by mass spectrometry in a manner analogous to that desribed for 'l'CDD • . 4P 218 APPARATUS AND REAGENTS. 1. Associated Electrical Industires model MS-9 double focusing mass spectrometer. 2. Varian TAC-1024 multichannel analyzer • . 3� Kroh n-Hite model DCA-lOR wide band amplifier. 4. Hewlett-Packard model 7035-13 X-Y recorder. 5. Virtis model 45 homogenizer. 6. Bendix model 2200 gas liquid chromatograph equipped with thermal conductivity and flame ionization detectors. 7. Leeds and Northrup model XL-610 strip chart recorder. 8� Gas chromatographic columns: 9. 10. w, a. 51 SE-30 on 60/80 Chromosorb stainless steel. 2 m x 2 mm i.d. b. 31 poly-m-pbenyl ether (6-ring) on 50/60 Anakrom AB; l m x 2 mm i.d. stainless steel. Trap for preparative gas chromatography: 150 mm x 1.5 mm i.d. borosilicate glass tube. packed with 30 mm of glass wool. Alumina: a. Fisher A-540, activated at 130 ° for 24 hr. b. Woelm, neutral, activity grade I. 11. Linde LR-30 liquid nitrogen refrigerator. 12. Kimble No. 34500 1.6-1.8 x 100 mm melting point capillaries ·(for sample tubes). 13. Chase Instruments P5180-2 7.0-7.5 x 150 mm disposable pipettes (for alumina columns). 14. Perfluorotributylamine (PFA): Peninsular Chemresearch (internal 111ass reference for mass spectrometric analysis). 15. 2,3,7,8 - Tetrachlorodibenzo-p-dioxin: :�· ":�fy."j .&§.�::�·':' ·?,? ¥--�>'*M?™ieft�%-#f?:}¥.9�h_,·:4!#?���*W'df-��%�¥f*-::- ,� ·"'k¾"�"';\5:�:,c;/_�·?. =v�t'PS 1 0 ! ;:..;-c­ _:�!¥!"!:��:*'�· · •. - i; � .J,k�'."·?':. 'hf� .fy������: : 219 a. From condensation of 2,4,5-trichlorophenol (courtesy of D. Firestone and A. Pohland, Division of Chemistry and Physics, Bureau of Foods, Food and Drug Administra­ tion, Washington, D.C.). b. From chlorination of dibenzo-p-dioxin (courtesy of s. Slapikoff, Department of Biology, Tufts University, leifoxd, Mass.) • c. From condensation of 4,5-dichlorocathechol with 1,2,4,5-tetrachlorobenzene (courtesy of A. Poland, Department of Pharmacology, University of Rochester, Rochester, N.Y.). 16 •. 2, 3,7,8-Tetrachlorodibenzo-p-dioxin -[ 37c1] 4 (see .synthesis above). 17. 2, 3 ,7,8-Tetrachlorodibenzo-p-dioxin A. Poland, address above). 18. 1,3,6,8-Tetrachlorodibenzo-p-dioxin: From condensation of 2,4,6-trichlorophenol (courtesy of D. Firestone and A. Pohland, address above). 19. Mixtures of mono-, di-, tri-, and tetrachlorodibenzo-p­ .dioxins and hexa-, hepta-, and octachlor00ioenzo-p-dioxins (courtesy of D. Firestone and A. Pohland, address above). 20. DDE (l,l;..dichioro-2,2-bis (4-chlorophenyl) ethylene): Reagent grade (Analabs). 21. DDT (1,l,l-trichloro�2, 2-bis (4 chlorophenyl) ethylene): Practical grade (Analabs). _[ 14c] (courtesy of 22. · PCB (Arochlor 1254, a mixture of tri-, tetra, penta- and hexachlorobiphenyls): Technial grade (Monsanto). w. 23. 2,3,6,7-Tetrachlorobiphenylene (courtesy of J. F. University of Bristol ., , Bristol, England). 24. Octachlorobiphenylene (courtesy of R.F.C. Brown, Monash University, Clayton, Victoria, Australia). 25. m-Terphenyl: 26. Ethanol: 27. Water: McOmie, Reagent grade (Eastman). Pesticide grade (Matheson, Coleman and Bell). Glass distilled. 220 28. Hexane: Pesticide grade (Fisher). 29. Dichloromethane:· 30. Carbon tetrachloride: 31-. Benzene: Pesticide grade (Fisher). 32. Acetone: Reagent grade (Mallinckrodt). 33. Potassium hydroxide:. 34. Sulfuric acid, 95-971: 35. Sodium carbonate (powdered): 36. SOdium sulfate: 37. Celite 545 (Fisher). 38. Bis-(N-methyl-N-nitroso)terephthalamide diazomethane precursor (E. I. duPont). 39. Protosol tissue solubilizer (New England Nuclear). 40. Aquasol scintillation fluid for aqueous samples (New England Nuclear). 41. Omnifluor scintillation fluid (New England Nuclear). 42. Beckman LS-250 liquid scintillation counter. Pesticide grade (Fisher). Pesticide grade (Fisher). Reagent grade (Merck). Reagent grade (Baker). Reagent grade (Mallinckrodt). REagent grade (Baker). 221 CONFIRMATION PROCEDURES High Resolution Mass Measurements. High resolution mass measurements were carried out with the standard peak switching circuits of the MS-9. A peak of known mass was selected by tuning the magnetic deflection field. The peak then was scanned by means of a small magnetic sweep coil positioned just before the magnetic analyzer and displayed on an oscilloscope. The peak switching circuit was used to select a second mass region which alternately also was displayed on the oscilloscope. The accelerating and electrostatic analyzer voltages determining the position of this region were varied by adjusting the value of a precision resistor decade box in the circuit. In practice the position of the second region was adjusted so as to include the unknown peak to be measured, and the decade box was tuned until the unknown peak directly superimposed with �e known peak on the oscilloscope. The value on the decade box then was used to calculate the m/e of the unknown peak relative to that of the known peak. Cor­ rections were made when necessary by measuring the m/e value of some other known peak close to the unknown peak and correcting the decade box value by the appropriate amount. For routine analysis the mass spectrometer was tuned to a resolut.ion of a/'00-10,000. For confirmation the resolution was. increased ., At 15,000 resolution with the peak matching method to 1s,000. .. ; described above masses could be measured to within about 10 ppm or± 0.0033 mass units nt m/• 300. 222 The mass of the 'l'CDD peak in Vietnamese fish at m/e 320 was measured to be 319.8988 and at m/e 322,321.8948. The cor- responding theoretical values for TCDD are 319.8969 and 321.8939. Determination of Possible Molecular Formulas Within¼ 3 mmu of TCDO This calculation was carried out with the aid of the computer program for determining possible molecular formulas from high resolution mass spectrometric data. From the isotopic isomer pattern of the peaks present at m/e 320, 322 and 324, the compound observed in Vietnamese fish must contain four chlorine atoms. attributed:t o four chlorine atoms Thu.s the weight can be subtracted from the observed mass at 3i9.897 to give a residual mass of 180.021. All molecular fo.rmulas within ± 3 mmu of this value were de­ termined for compounds containing c, H, N, o, F, Si, P, ands. The results, with those formulas which would not give the ob­ served 321/320 (M+l) +/M+ ratio (0 •. 13) eliminated, are listed The signal at M+l is determined by the presence of minor stable isotopes, such as 13c, 29si, or 33s. one mass in Table 21. unit higher than the major isotope. 223 Isotopic Isomer, Fragmentation, Doubly Charged Molecular Ion, and Low Ionizing Voltage Ion Intensities. Ion intensities for isotopic isomers of the molecular ion, a fragmentation ion, the doubly charged molecular ion, and JllOlecular ion at low ionizing voltage were measure for the compound observed in Vietnamese fish and compared to those observed for TCDD. All measurements wex:e carried out at high resolution (10,000). selected were m/e . 323.891 (. 37Cl2). The molecular ion isotopic isomers 319.897, 320.900 c 13 c), 321.894 ( 37 c1), and The fragmentation ion at �/e 258,930 cor. responding to the 37c1 isotopic isomer of �e (M-COCl) + ion (the peak at m/e 256.933 was obscur.ed bya·PFA. reference peak) was measured, as was the doubly charged molecular ion at m/e 159.948. Intensity at the molecular ion for an ionizing voltage of 25eV was compared to that observed with the normal 75 ev. By means of multiple ion detection, ion intensities in each . case were simultaneously obtained for 37Cl TCDD which was added to the sample prior to analysis, thus providing an internal reference for the behavior of TCDD. listed in the Discussion section. The results are Volatilization Time in the Mass Spectrometer. Ion int'!nsity at m/e 321.894 (TCDD) as a function of time was compared to that at 327.885 c 37 c1 TCDD). This was achieved by using the peak switching circuit of the MS-9 to 224 altemately scan each region while the signal was continuously recorded on an oscillographic recorder. In order for a signal to be classified as TCDD, it had to produce an envelope of intensities as it volatilized that was dixectly s�perimposable Representative volatilization with that of the 37cl TCDD. patterns for.different sample types are illustrated in Figure 23. confirmation for Routine Analyses. For.a sample to be classified as co�taining TCDD under conditions of routine analysis, the following criteria had to be met: (1) signals at both m/e 319.897 and 321.894 had to be present, (2) these signals had to h11ve a 319.897/321.894 ratio in the range 0.70-0.85 (the theoretical value for TCDD is 0.77), and (3) the signals had to appear with the sam� time �ourse as 37c1 TCDD at m/e 327.885. Photolysis with Monochromatic UV Light. A portion of the final residue from certain samples which were positive for TCDD was dissolved in 10-20,-.l of hexane. placed in a 1.3-1.5 mm id quartz tube, and positioned to completely cover the exit slit of a monochromator. for these photolyses is shown in Figure 24. The apparatus used A 100-watt mercury lamp (0.25 mm arc length) was used as a light source. The 225 li9ht was collimated and then focused by means of a cylindrical lens on the entrance slit of an f 3.6 grating monochromator·. (Jarrell-Ash model 82-410). Slits were selected so that 95\ of the passed light was contained in a 100 X band. were carried out at peaks at 2894 (E=3200) and 3130 in the mercury emission spectrum. l Photolyses (E=6000) Under the conditions described, the t� for destruction of TCDD at 3130 i was about 20-25 min. and at 2894 A about 60-80 min. The presence of sample residue tended to shorten the t� relative to that for neat TCDD. Half lives in benzene were about 1.5 times longer than those in hexane. In an effort to obtain a third point, photolysis was attempted at 3220 K, even though there is·no mercury peak in this. region. Ho decomposition had occurred at·the end of 60 min, indicating that under the present conditions, reaction at this wavelength was too slow to be useful. In each case another portion of the sample being photolyzed was stored ini thedark during the photolysis. This pro- vid� the control from which the relative decomposition of the various sample components was calculated. Results of such photolyses are described in the Results and Discussion section. Partitioning Between Acetonitrile and Hexane. A sample of Vietnamese fish, which prevlously had been 226 found to contain 540 ppt of TCDD, was treated with Isolation Following the alumina step, one portion of the Procedure I. eluant was chromatographed preparatively by glc as usual. A second portion of the eluant was equilibrated with two parts of acetonitrile. The top (hexane) layer was concentrated and From MS analysis 391 of the· 35cl and 321 . glc prepped. of the 37c1 TCDD were ;ecovered relative to the untreated portion of .the sample. Seeration of 2,3,7,8- from 1,3,6,8-Tetrachlorodibenzo-p-dioxin . by Preparative Glc. By ·glc analysis it is possible to resolve 1,3,6,8and 2,3,7,8-TCDD (Figure 6). With the .51 SE-30 column (see Apparatus and Reagents) at 230 ° , with a 30 ml/min He carrier flow rate, the retention time of 1,3,6,8-TCDD is 1.45 times the retention time of m-terphenyl, an internal reference compound. 2,3,7,8-TCDD appears at 1.95 times the m-terphenyl Four hundred pg of 1,3,6,8-TCDD were glc retention time. prepped and two fractions were collected, one from 1.30-1.75 . times and the other from 1.75-2.30 times the m-terphenyl re­ tention time. A similar glc prep of 400 pg of 2,3,7,8-TCDD was carried out. As shown in Table 22, most of the 1,3,6,8- TCDD appears in the first fraction, while essentially all of the 2,3,7,8-TCDD appears in the second fraction. The presence.of a 'l"CDD signal from the 1,3,6,8-TCDD injection collected at 227 the later retention time was probably due to tailing, not to the presence of 2;3,7,8-TCDD. isomer was 951. The purity of the 1,3,6,8- Samples of domestic u. s. fish with <3-4 ppt of either TCDD isomer were spiked separately with 500 ppt of each isomer and carried through the preparative glcpro­ Separation paralleled that observed for the neat cedure. isomers. A sample of Vietnamese fish known to contain �30 ppt of TCDD was treated similarly. The TCDD in this sample behaved in a manner identical with that of 2,3,7,8-TCDD. Analysis for Hexachlorodioxin in Vietnamese Fish. o •. s. The recovery of hexachiorodioxin in a spiked domestic fish was measured withthe cleanup procedure using Woelm alumina (Table 23) • All aspects of the'procedure were identi�al to those used for TCDD, except that MS analysis was carried out at m/e 388 (387.837). Recoveries were adequate and the procedure was repeated with a blank domestic u. s. fish sample and with a sample of Vietnamese fish known to contain 750 ppt TCDD. As is shown in Table 23, the level of hexachlorodioxin in the Vietnamese fish was <21 ppt and <33 ppt in the domestic o. s. fish. Lack of Absorption of TCDD on Glass in Storage Vials. 'l'Wo 20-ml glass bottles used for storing sanples·were used in this experiment. One previously had contained a sample 228 of domestic U. s. fish with 3 ppt of TCDD, and the other had contained a sample of Vietnamese fish with 540 ppt of TCDD. Both bottles were emptied in the normal Jnanner with a spatula which leaves a residue of about 1 g of fish in the bottle. The bottles were extracted with two 5-ml portions of. ethanol followed by three S-ml portions of dichloromethane, heating to reflux each time. Twenty ml of hexane was added and the combined organic phat.e was extracted once with 40 ml of water and four times with 40 ml of 95-971 sulfuric acid. The organic phase was concentrated to 3-4 ml, 10 ml of hexane was added, and the solution was again concentrated_to 3-4 ml. From this point the cleanup was carried on from step 11 of Isolation Procedure II. bottle a total of For the domestic u. s. fish 35 pg was calculated to be present. no 'l'CDD is absorbed on the glass, this gives a in the one gram of fish residue in the bottle. If level of <35 ppt For.the Viet- namese fish bottle a total of 610 pg was calculated to be present. If no TCDD is absorbed on the glass, this gives a level of 610 ppt in the one gram of fish residue in the bottle. Since the level previously observed for this fish was 540 ppt, it seems likely that in fact little or no TCDD was adsorbed on the glass. 229 Treatment with Diazomethane. A sample of 1970 Vietnamese fish was treated. according to Isolation Procedure III (neutral extraction). Following elution from the 15 X 200 mm alumina column, the eluant was divided into two equal portions. One was concentrated to 8 ml, treated with 12 ml of 0.5 ! diazomethane in ether for 30 minutes,· concentrated to dryness, dissolved in cc1 4 and chromatographed on alumina according to step 10 of Isolation Procedure II. The second portion was concentrated and directly chromatographed by the same procedure. The TCDD level observed without diazomethane treatment was.92 ppt. treatment it was 90 ppt. With diazomethane This indicated that the TCDD peak observed in normal analyses is not a fragmentation peak.of a phenoxyphenol •predioxin" in the mass spectroine.ter. 230 RECOVERIES Recoveries for overall Isolation Procedures. Three differen.t isolation procedures for TCDD were in­ vestigated (see above). The first of these involved saponi­ fication, acid extraction, alumina chromatography and gas-liquid chromatog:..·aphy. In the second the glc step was replaced with a second alumina step. In the third saponification and acid extraction were eliminated and only neutral extraction into an organic solvent and alumina chromatography.were used. coveries for each procedure are listed in Table 16. Re­ In each case a known amount of TCDD was added to the sample immediately before cleanup and the recovery was measured at the end of the ·procedure. Recoveries for Individual Steps of the Cleanup Procedure. Recoveries of TCDD were measured individually for each step of the cleanup procedure. typical for each case. The following experiments were For the saponification recoveries were measured after two hours. In each run 50 1,11 of a one ng/1 benzene solution of 37c1 TCDD was combined with 2 ml of ethanol, 4 ml of 401 aqueous KOH and l.5 ml of water. Following saponification this solution was extracted with three 2.ml portions of hexane. The combined hexane phase was filtered through a prewashed column of 2 g of Na2co3 in a disposable.pipette. Fractions containing 0.11 of the total 231 residue were prepared and analyzed. Saponification of most tissues seemed to be complete by 2 hours and so this time was used routinely. Four ml of a 20 pg/1 benzene solution of 37cl.TCDD were extracted four times with 2 ml of concentrated (95-971) .sulfuric acid, rinsed once with 3 ml of water and dried over N•2co3• Fractions containing 0.151 of the. original solution were prepared and analyzed. As is indicated in Table 17, the recovery was essentially quantitative. A solution of 5 ng of 37c1 TCDD in l.ml of hexane ·. was chromatographed ori alumina. Twenty percent of the TCDD fraction was concentrated to 50 µl and fractions containing 21 of the original 37cl TCDD were prepared and analyzed. The recovery (Table 17) was again nearly quantitative. A solution of 1 ng of 37c1 TCDD in l ml of 201 dichloro­ methane in hexane {corresponding to the effluent from the alumina chromatographyin method 1) was concentrated to 50 µl in.the microconcentration apparatus. Fractions containing 21 . . ii of·the or g nal 37Cl TCDD were prepared and analyzed. The results (Table 17) sug.gest that· this step is quantitative. Occasionally losses .due to bumping occurred during micro­ concentration. During the period when a preparative glc step was in­ cluded in the cleanup procedure, recoveries were measured. Twenty µl of a 30. pg/1 solution of 37cl TCDD were glc prepped. ,.,.4. ¥ t 232 The trapped effluent was fractionated' and analyzed. Recoveries as shown in Table 17, were lower f9r this step than for any other. To attempt to discover the fate of the unrecovered TCDD, a run.was made in which the regions 50 sec prior to and 100 sec after the no:rmal preparative region were collected in addition to the normal region. A 61 recovery of TCDD was obtained in the preceding 50 sec fraction and a 2.51 recovery was a1so obtained in the following 100 sec fraction. A second rinse of the glc trap gave only 41 additional recovery. A la.1:'9e portion (30-401) of the TCDD still was unaccounted for. Possibly it was tightly adsorbed on surfaces in the glc and slowly eluted over several hours • . J?Al .4!Q Pk - "' RI. . *..+'ff.¼?. P:¥<·½.S . 0¢ '· ,_AfZ4 .¼.ti?- 233 Extraction Efficiency TWO male and two female rats (Sprague-Dawley) were injected intraperitoneally with 5 µg/kg of 14c TCDD (98 mCi/ . mmole). One female rat died on the day of injection and-one male rat died three days after injection. After 30 days the remaining rats were sacrificed. The rats· were dissected into three portions: liver, sub­ cutaneous fat, and all other tissue, including skin and hair. The liver was homogenized with one part water. Al�quots of the homogenate were treated with Isolation Procedure III, Isolation Procedure II (up to the end of the first alumina chromatography), or digested_ directly with Protosol (New England Nuclear). The.·r�sidues from Isolation Procedures II and III were divided into two equal portions, one of which was spiked with a known amount of 14c TCDD for quantitation, dilu.ted with Omnifluor (New England Nuclear) , and counted by liquid scintillation. The residue from the Protosol digestion· was decolorized with 301 aqueous e2o2, divided into two portions as above, diluted with Aquasol (New England Rllclear), and counted. sion Section. The results are listed in the Discus­ 234 SAMPLE COLLECTION AND STORAGE Samples from Vietnam were collected fran the sites shown in Figure 20.245Fish and shell fish samples w� bought directly from local fisherman, homogenized with a hand grinder, which was cleaned between samples, placed in borosilicate glass storage vials with Teflon or alwninum foil lined caps. Mother's milk samples were expressed into a glass storage vial. The samples were immediately frozen on dry ice and later, the same day, transferred to a Linde LR-30 liquid nitrogen freezer. The freezer was returned by air i:o the United States and then used for permanent storage of the samples. Enough liquid notrogen was placed in the freezer before shipment from the u.s. last throughout the collection period and return trip. after storage of some samples at -ss0 to Later, was found to be satis­ factory, all of the samples were transferred to a Harris LTF 6.5 freezer and stored at -ss0 • 245 The samples were collected by J. Constable, P.H. Ho; B.T. Lang, M. Meselson, R. Cook, and A. Westing. Identification of fish and crustacean samples was provided largely by B.T. Lang and P.H. Ho. ·· -·c 't ctr 10 a.printed from ADVANCES IN CHEMISTRY SERIES. Number 120 "Ollorodioxin1-0rigin and Fate" Copyright 1973 by the American Chemical Society Reprinted by permission of the copyright owner An Improved Analysis for Tettachlorodibenzo:P-dioxins .• JIOBERT. BAUGHMAN 1111d MAITHEW MESELSON Deputment of Chem&tJy and Department of Biochemistry 1111d Molec,alar Biolo&Y, Huvard Univenfty, Cambridge, Mus. 02138 A � � of t1ie enoironmental lewu of 1,,.1,1,8-lfflachlorodlbenzo..p-diorin (TCDD), an mroordi­ -U, fr)zlc compound prmnt a, 0.1 unputity in t1ie hnbi­ .t:IM J,4,,S.T and In .t0me commercial cel«opheno&, can be aode bv eoaluating representatioe iamplu with• auffo­ cfnilg lffllihoe analgtical method. The fflllitwity required ,. eeU be,ond l1tat ooailable with curmd methoh. We re­ ,-, 4 � UMng time OOef'Oged high molution fflO.f,f � tDith o '6Mtitlity (IO·IZ gram) auUoble for ,ucla 411 � Interference from pentochlotolnphenvl ill aitlaln. mo1erio1, from the entlironment pnmatlv limita '4mlng full Hfllflit>itll of the method although toOl'king toward • ,uolution of ,,. problem. onlv ot­ on, 0-themteiest in the chJorodioxin problem stems from our work with Herbicide Assessment Commission of the American Association i:r the AdVIIQCellleDt of f Science which was organized in 1970 to initiate a study of the. e!fects o herbicide use .in Vietnam. As one part of that iwestigation we are analyzing various sample'$ from Vietnam for TCDD, abown impurity in U.5-T (I, J, 3, 4). This herbicide in a one-to-one mflltw:e with 2,4-D is a component of agent Orange, the herbicide that used most widely in Vietnam. Our aim has been to determine wlae.ther TCDD has accumulated in food chains to any significant extent. We were surprised to &nd that no method existed that was sensitive IDOUgh to detect TCDD in animal tissues even after administration in - species· of lethal doses. An example is the guinea pig, the � f sasceptible species o the few that have been tested, and therefore a good dloice- for establishing desirable limits of detection. The lethal single -' dose ( W..) in males of this species is 0.6 ,.g/kg body weight was 81 t3 (S). This means that if all of the TCDD were retllined, the level of TCDD would be less than 1. part per billion ( ppb) In the whole animal The lowest reported limit of detection for TCDD in whole tissue ls 50 ppb (6). Thus, a guinea pig could be killed with TCDD, and it "'®Id lie impossible to establish this fact with the analytical procedurea In eurrent use. Such analytical procedures· are clearly of little value in D)ODitoring food chains for the buildup of TCDD. This is even more apparent lf one eonsfders the possibility of sub-lethal toxic effects and allows a margin for a safety factor. If we provide a factor of about 100 for noo-lethal toxicity (6) and a further factor of 10 for a safety margin and to allow for the possible existence of species even more sensitive than the guinea pig. we would require a level of detection of ( 10"') (}(ti) ( 10-1) - 10·11 or I ppt (I part fn 1011 ) for environmental monJtorfng. For a. 1 gram sample this would require a limit of detection of I pg ( 10·11 gram). The limit of detection of TCDD for the electron capture detector, the key­ stone of current analytfcal procedures, is not much less than l ng ( 10"' gram). Jn addition to high sensitivity, a requirement for any acceptable analytical method Is high speculcity because at very. low levels few coo8rmatory procedures can be used to establish the identity of a particular compound. A method which uniquely combines high Jemitivi.ty with high speculcity 1s high resolution mass spectrc)metry. we have \ISed this method as the basis for an approach which we believe will male possible a� assessment of TCDD levels in the enviroDID!!Dt. Figure i shows WIit the mass spectrum of TCDD is relatively simple. (All the work reported here was done with an IJSOC:iated Elcctrical lnuo �-----------------,---, .. .....,.. .. a.laHw­ - _,I . T'"'"_�--�-----1--,.-..-..l-.--1 - •+-,fl,-.,.;...._.... uo - .,_,,_,,_.of 1111•- .... .,_,,_lmpurii,. � l. !l,3,1,8-tmoc� (TCDD). Tu inolecular Ion (M•J ,..al m/r, 320. lon&l.,g� Ti) eV, -1so•c AllerW: IN CRIOIM - AND ,PAT& cltllbies MS-9 double fOCUJing mass spectrometer.) The base peak is the molecular ion at m/e 320. As a result of the various possible combina­ tions of the naturally occuning UC! and 11Cl isotopes in a tetrachloro compound, the signal for the molecuJar ion Is a pentuplet with peaks at ,n/e320,322, 324,326, and 328 with intensities in the ratio 77:100:49:10:1. In addition, the four chlorine atoms and the limited number of hydro�"" atoms make the compound slgnilkantly n:ass deficient ( the m/e 320 peak Is actually 319.8956) and, therefore, relatively easily resolved from most other organic residues (for which ml• 320 would be 320.1-320.2). First, we tried scanning the region m/e 310-330 (Figure .2). As the sample was introduced into the mass _'spectrometer, signals appeared at m/e 320, 3.2.2, and 324 and then, ai the sample became exhausted, disap­ peared. Under these conditions the limit of sensitivity was on the order af 100 pg. We next reduced the scanning interval to about one third of a mais unit. This allows the detector to spend more time in the region of interest, considerably increasing the signal At a resolution of 10,000 a series of scans was made, alternately two at 3.22, two at 314 perfluoro­ tributylamine (PFA) reference peak, two at 3.2.2, etc. (Figure 3). The PFA wu bled in from an external reservoir at a constant rate. providing nterence peaks that remain at the same height throughout the. analysis while the sample peaks rise and then fall as the sample volatilizes. This procedure with a sensitivity of about .20 pg was still riot adequate. It is possible· to obtain greater sensitivity from the repeated narrow scam in Figure 3 by combining them to produce a single time averaged scan. Procedures accomplishing this under low resolution con­ ditions have been reported previously (7, Bj. Under the present condi­ tions a system was devised for doing 'this using a Varian 10.24 averaging eomputer (CAT) in conjunction with the MS--9. The result is shown m Figure 4. The signal for a pair of peaks at the limit of detection for a nagle scan is shown In Figure 4A, and the averaged signal from sixty scam is shown in Figure 4B. The signaJ-to-nbise ratio is expected to Improve approximately as the square root af the number of scans (9). With 1 min of scanning at a rate of one scan per second, the observed Improvement Is approximately that expected'. At very fast scan rates data Is lnefliciendy transferred to the memory of the CAT, and resolution fl decreased by damping caused by the time. constant of the MS-9 cir­ cuitry. In the present system this limits the maximum scan rate to four scans per second. With very short volatilization times ( < 10 sec) sensi­ . tMty is decreased, perhaps In part because of decrffased ionization effi­ deilcy. With volatilization times longer than a001,1t 60 sec the drift in peak position from scan to scan Is large enough to decrease siP.iJlcantly · the. resolution 9bserved In the time averaged specttum. '"'be optimum volatilization time Is from 30 to 60 aec. mown q.t. .,, .b ,\4( __ f._,_4.i •.. ,½4•!1Ji*'.. OJ.I J.K4%§!.�l.#,.. I}" 10, IIAVCIIMAH AND· MUSI.ION 95 The interfacing of the CAT with the MS-9 fs _mustrated in Figure 5. The ions in the m/e region of interest, a�er being focussed, pass by a small magnet coil which deflects the beam back and forth over the detector slit. After passing through the slit, the ions strike an electron 314 i 326 328 Ffgu" J. Repetitwe aamning of m/e 310-330 (5 lll!C/6COR). Standard eoi&ditlona for t1"' and all following figures: Ionizing tJOltQf!.6 70 eV, tic­ «lnolmg oolta,e 8 kV, Crop CUtfflll 300 ,.A. ,m,/dplin600, source lSO-C. • u fflCffA> L..JI.....JL.JWW fflHUlUJJJU Flgvre3. Altemalingllllrl'OIOICtlll6 l'op:,..w«:tfflfflldoneacAIMl...,t1i'IC8fl. B_,.,,......,.,,,._,_(ltl/0 lwlf---' ,_,, a, ."114, IIDo Ill 3a2, ne.) (200 P& TCDD). multiplier, producing a signal which is continuously displayed on an CIICilloscope on the MS-9. This provides a means of monitoring each acan. Simultaneously, the signal is added to the IOU.Channel memory of the CAT. An oscilloscope on the �T continuously displays the total memory content which makes it possible to monitor the overall cow:se of the analysis. A potential problem of phasing the beam deflection coil with the memory sweep circuit of the CAT is avoided by using the aweep voltage ramp of the CAT, ala an amplifier and appropriate drcuib of the MS-9, to drive the beam deJlectfon coil. The coil is thus necessarily in synchrony with the CAT. The procedure we have adopted for introducfng samples Into the MS-9 is shown in Figure 6. It provides reproducible analyses at a high level of sensitivfty. The sample tubes are made from 1 mm id melting point capillaries. A Hamilton 10-,.J syringe is used to introduce a 3-4 ,.J portion of the residue into the sample tube. With a small Bame the ample tube Is draMa out just above the level. of the liquid to pro­ duce a car'.:lary constriction about 20 mm long. The solvent is then ., ... 10, IIAVCIIMAX AND ..,..._ An lmprooetl Anolr* removed at reduced pressure. Bumping b prevented by the capillary COllltriction. The sample tube Is th«m sealed with a flame. At the time of analysll the capillary II broken off� mm above the constriction to give the tube configuration shown in Figure 6. The tubes are introduced into the MS--9 with a wire holder on the tip of a standard MS-9 direct fmertion probe. To aid reproducibility all analyses are started at the tame time ,after insertion of the sample tube into the MS--9 source. The temperature of the source heating block Is adjusted to give a sample wlatilization time of 30 to 60 seconds. The result of combining these various components in the anal)'IIII of a 2-pg sample of TCDD is illustrated in Figure 7. An internal standard b given by a PFA fragmentation peak which II a known distance, 85 mmu (1 milllma.os unit or mmu - 10-• atomic mus unit), &om the TCDD peak. In Its present form the MS-9-CAT system hu a limit of detection for TCDD of about 1 pg. The procedure, we haVf' described retains the generality of normal 1111111 spectral anal)'IIII. It. is particularly suited, however, to compounds aw.HO 31$,CMO :114,MO �- :I !J,CMO F� 4. lmprotJflfflfflt In lffllitMl-,1 wUla th. CAT, PFA and o "'•,_ TM obHrt,ed Im� In llgnal-10-noue rollo ruulu from th. longer total ,canning llme and oho tlu! faet that manr, _.,,. ore ..,,. tlurinB thll llme. Tiu! oonall lmpl'OOIJfflt/flt In ngnal-lo-no(n ratio tJa. ,i,ndl on tu dlitalllld potD1Jr �ctrum of tlu! nol# (9). &aolutlon 12,000 . • and for all . iUowin& tlnw --,tnl .,,.c1,11 ,-a Id m/e 315. ·w,-- lllpe:BO_(_..,.,_,. 18 containing atoms with signilcant mass defects, such u heavy metal or Ol'glU10Chlorin compounds, which are easily resolved from other residues, Before the procedure is applied to tissue or other samples from the environment, some potential complications must be taken into account. One is the possibility that other chlorinated organic compounds present. fn the environment might interfere with or obscure the TCDD peab. To test this, - obtained mass spectra on the MS-9 for most of the common organochlorine pesticides including lindane, · aldrin, dieldrin, mirex. heptachb, DDD, DDE, and DDT, as well as various polychlo­ rinated biphenyl (PCB) mixtures. In the TCDD mass range DOE from its molecular ion bas isotopic isomer peaks at m/e 320; 322, and weakly 324. The molecular ion of pentachlorobiphenyl, a component of some .PCB mixtures, has a peak at ml• 324, and this compound has a weak &agmentation peak at ml• 322. DDT has weak fragmentation peaks f at m/e 320, 322, and 324. As shown for m/e 322 in Figure 8, all o these compounds can . be resolved from TCDD at our normal resolu· tioD of 12,0l'O (!7 mmu at m/e 322). The relative input amounts of. each compound producing the peaks shown are: DDT, 250; DDE, 25; TCDD, l; PCB · ( iuochlor 1254 }. 250. Even though moderately �e excesses · of these intecfamces caa be tolerated, it is necessary to use highly l •.. ..�·, ···rrt-tftf»r fsr 1e· 1--IIIOlinfbloclt ______t........k @ >== Probe lhclft tcm ___$ ...�..-->- :120.000 .,. n,.,oo ,.,...1. l.Wtofat«uoa Top: I PC (t X JO-- ..-JTCDD. Boaoin, � · - irattnW ·,·ii, 100 elBdent cleanup procedures to any out ana1)'III b TCDD at very JowleYets. Another complicating characteristic of -terials &om the environ­ � is that the size and nature of the residue to be analyzed in the mass apeetiometer will change from sample to sample. To determine if this might have an el£ect on the observed TCDD signal, we analyzed identical umples of TCDD with di.lfering amounts of squalane, a saturated hydro­ carbon selected as a model for residues obtained from standud extraction wl cleanup procedures. As is indicated in Table I (Part A), there was C r.,...........--..--�.- --- • • _.,.________ I t I ,-, ........... .,,......_.,._,.. .. .t..:.. _______ _,._._, -- ---- --air• ............... ,,. a - F,._8. lwolulioaofTCDDt- ,,.,.,.,_ A: DDr + TCDD. B: DDE + TCDD. C: l'CB iAndlor 1254} + TCDD. D: DDT + DDE + l'CB. E: DDT + DDE + lCB + TCDD. s".w·--· \'-:.#'ttik½v;> · t):"r0'K"1fWrtt· _·:__f ·. W.r.it __it ·. Hi __>• - - :t'ti'fir:iQ·_ :lli_ --- )f� 1?C1#'fftvjaff# #a rfit tt'tr«wtrzr'� ··· ;; trtflttfXttt 1t@ltr:Wt n< rw :I: 101 Table L r Elf.« of Siu of Total lletid1111 A. llMpo,iM/ur TCDD Bgualan. added (microgram,) 0 1 6 26 .RelatiN rc,pon,efurMJ pg TCDD 100 76 so · 16 -� B. Ratio of two eomponmta • · TCDD/TBB Spalan. added (microgram,) 11 1 6 3 1 9 • Ratio or reoponae for 20 pg TCl>D to reopona, for 200 Pl of 2,3,$,11-tetl'lldlloro+ llromoethylbenzene (TBB). a ligniflcant effect on the respPnSe to TCDD. This eliminates any hope of determining the amount of TCDD present directly from the area of the TCDD peak. In an attempt to resolve this diiliculty, we added an · internal standard which does not coincide with TCDD or with peaks from any of the potential chlorinated hydrocarbon impurities and mea­ sured the ratio between TCDD and the star .ard in the presence of vari­ ous amounts of squalane. A compound w. ·• a suitable mass, .2,3,S,6tetrachloro-4-bromoethylbiinzene (TBB), Wlb synthesized ·for this pur­ pose. However, as is apparent from Table I (Part B), the ratio of its signal to that of TCDD was far from constant, and this procedure was ruled out. Table· I also suggests that in order not to cause a signiiicant loss of sensitivity, the total sample size should be kept under 5 ,.g. Some alternative method had to be devised to quantify the TCDD -ements. The problem was solved · with the observation, illus­ tiated in Figure 9, that the response to TCDD is linear over a wide con­ centration range as long as the size and nature of the sample matrix remain the same. Thus, it is possible to divide a sample into two equal portions, run orie, then add an appropriate known amount of TCDD to the other, run- it, a11d by simply notirig the increase in area caused by the added TCDD to calculate the l!lllOUllt of TCDD presesit in the first l portion. Figure 9 illustrates the reproducibi ity of the system. Each point was obtained from four or five independent analyses with an (root mean square) of 5-10%, as indicated by the error Bags. •lpch is acceptable for the present purposes. To satisfy the requirement of having a_ �tal residue of cioly a few mfcrognms, the sample cleanup must be very thorough (10). The pro­ eeclure which we. have used to accomplish this is the following. A 10. error ¥ . d tt ti cmwm.• --ran IOI • IO 15 TCIIO C'91 20 ��O:. _,. ..... F,,_. 9. Uaearitr, .of tta,,Dn# and�- Tu error -.0 illdicm ,_ roo, _ _,_ error ·rJw_-.,, pam 58lllple of human milk was combined with 10 ml oE ethanol ( all mlvelits were of pesticide grade) and 20 ml of aqueous 40'J, KOH an:I le8wrecl b 6 hours. This solution was then exttacted v.ith four. &-ml pClltiom of 5% benzene in hexane. The organic phase was � with four 8-ml portions of 85% a.so•. filtered thrQugh a 10 mm id column of IO.grams of powdered Na2C01, t"Oncentrated carefully to about 10 ;J. aacl peparaJiwly c:hromatographed by gas-liquid chromatography oo a­ l m X 1/.r' column of 3% OV-101 methyl silirone polymer liqujd phase on. S0/60 mesh Anahom AB solid support (200 ° C column. 30 ml/min · Be). The CU: trap consisted of a 150 mm X 1.5 mm id borosilicate glass tul;e pacud with 30 mm of 100/120 mesh glass beads retained with glass wool plugs ( 11, lJ ). ·Toe trap was wetted with.hexane and cooled in dry ice. The CU: cleanup is carried out through a thermal car:ductnity detector. A smaD amount of an internal standard, m-terphenyl with a ·bown retention time relative to TCDD was added to make certain that -the TCDD collection was carried out at the right retention � 1he total residue &om the CU: cleanup when divided into twelve frac:tions · povjded a suitable sample size. Figure 10 illustrates the results of a typical analysis oE a 10-gram l8Dlple of human mUk to which 0.1 ppb TCDD bas. been added. Blank 1tF 10. N- .._ -- ua.- All lm,-cf � ..,. ,21.to0 10., ,21.to0 Plgtf,w JO. TCDD � In IO g,mn huff!Oft milk to which 10-- gmm (0.1 ppb) TCDD waa tldded. Each trace repre,enr. 8'1& of tla. tot4l ruidue. . 7op: Nllt1fl6 (1159' -,,,}. """-1-"'- + /M1 Pl TCDD. samples showed no indication of TCDD at this Jevet Each trace repre· seats SCA, of the total sample and ·since the area under the unknown TCDD peaJi. calculated &om the nm with TCDD added, corresponds to approximately 2.0 pg. the recovery is about 25%. The other sjgnals observed in the TCDD mass region are probably from DDE and penta­ ehJorobiphenyL At � present time the major impediment to mending the proc:edure to the desired level of 1 ppt Is interference from penta­ chlon,biphenyL We are investigating changes in the cleanup procedure which show excellent p,nmdse of reducing the level of this and other fnCerferences in the &naJ residue. are also developing a procedure f'm directly measuring the recovery of TCDD in each cleanup by adding �p1a] labelled TCDD � the sample before cleanup. We �,..., We would liJce to thank David Firestone of the Division of Chemistry and Physics, FDA for providing us with samples of TCDD, IClaus Bie­ mana and Charles Hignite of the Department of Chemistry, Massachu­ setts Institute of Technology for assistance in the early stages of this work, David- Parrish of the Department of Chemistry, Harvard Uni. Wdity for assistance in developing the MS-9-CAT system, and William Doering of the Department of Chemistry, Harvard University for the use of laboratory facilities. This war!-: _ _,pported by the Herbicide Assess- ICM ment Commission of the Ameribln Allociation for the Advancement of Sdeoce and by the Ford Foundation. l.lln-.,_.., CII•� l. Brenner, JC. S., M!ilter, JC., Sattel, P., /. Chromalogy. (1972) 84, 39. I. EMdge, D. A., Analyn (1971) 98, 721. 3. John,on, J. E., Hearings Before the Subcommittee on Energy, Natural lle­ llOUJ'Cel, and die Environment of the Committee on Commerce, United States Senate on "Effects or 2,4,5-T on Man and the Environment,• April 7 and 15, 1970. 4. Woolson, E. A., Thoma,, R. F., Enlor, P. D. J., J. A,r. Food Chem. (1972) IO, 351. S. "Report on 214,5-T," A Feport of the Panel on Herblddes of the President's Science Aavilory Committee, Executive Oftlce of the President, OlBce of Science and Technolo1CY (March 1971). Woolson. E. A., Reichel, W. L., Youna, A. L., ADv.ur. Caw. Su. (1973} HO, 112. 1. Blroa, F. ,., Anal. Chem. ( 1970) 4i, 537. 8. Plattner, .R., Markey, S. P., Org. Mau Spec•. (1971) S, 463.. 9. Ernst,R.R., &o. Sci. ln,t. (1965) 38, 1689. 10, Rea, J., Higginbotham, G. R., Fireltone, D., J. A.la. Oflic. Anal. Chem. , (19'70) 53; 628. 11. Bierl, B. A., Beroza. and Ruth, J. M., J. Goa Chromatogr. (1968) 8, 286. l!. Munay, JC. E., Shipton, J., Robertson, A. V., Smyth, M. P., CMm. Ind. (1971) 401• e. . Rscava February 8, 1972. .. .• An Analytical Method for Detecting TCDD (Dioxin): Levels of TCDD in Samples from Vietnam b la•ert Baaghman· and Matthew Meserson· 2,8.7,8-Tetrachlorodibenzo..p.dioxin (TC- of detection for TCDD, about 10.:10g, is in­ . DD) is· an extraordinarily toxic substance adequate. This method is _also susceptible to that is produced as an unwanted side p1·oduct interference from other compounds and so · in the industrial synthesis of 2,4,6-tri­ is not very specific. chlorophenoJ. an intermediate in the manu­ Masa spectrometry offers better pouibili­ facture of the herbicide 2,4,6-trichlorophen­ ties. It is high sensitive an_d · in the high oxyacet,ic acid (2.4,5-T) (1, .t). Because of ita resolution. mode of operation it is highly chemical stability and its lipophilic nature, · 1tpeciftc. We have previously described a tiR)e • the possibility exists that TCDD released averaged mass spectroscopic method with into the environment could accumulate in an adequate limit of detection (4). However; food cha.ins. � direct test of the possibility full sensitivicy could not be realized in most of biologically significant accumulation in sample types because of interference from animal tissues requires an analytical method DDE (a major degradation product of able to detect TCDD at levels well below DDT) and polychlorinated biphenyls thc:,se known to be toxic. The lowest value (PCBs). In this paper we describe a clean­ known for the lethal dose of TCDD is that up procedure that overcomes this difficulty. Homogenized samples are saponified in al­ observed in the guinea pig, for which the aing)e oral dose LD.. is 600 parts per trillion coholic potassium hydroxide and extracted (ppt) body weight (8). Allowing for sub­ with hexane. The extract is shaken with sul­ furic acid and chromatographed on alumina. lethal toxic effects and providing for a servative margin of safety, it seems desir­ Elution with carbon tetrachloride-hexane able to have an analytical sensitivity of at removes most of the DDE and PCBs•. Chlori­ least 1 ppt. For a 1-g sample this means the nated dioxins are then eluted with dichloro­ method must have a sensitivity· of. about methan&-hexane. The TCDD containing frac­ tion is further purified by preparative·gas­ 10·11g or 1 picogram (pg). . The DIG8t common method for analyzing liquid chromatograph! and analyzed by mass chlorinated organic compounds in tiuue spectroscopy by use of a multichannel analy­ zer to average successive scans. aamplea is gas-liquid chromatography (t:'l.C) We also report the levels of TCDD found with an electron capture detector. Ita limit in a limited number of samples of fish and crustace!lns from locations in South_ Vietnam near areas heavily expused to 2,4,6-T. con­ Septeaber 1978 27 of EtOH and then twice with a few milliliters Experimental of hexane: the solvent was refluxed each Rn,ents and Apparatua time; and the hexane was extracted with 1 Hexane (pesticide grade, Fisher Scienti­ part 1.0N NaOH. ftc). dichloromethane (reagent grade, · East­ (6) The hexane was extracted four times man). carbon tetrachloride (reagent grade, (or until acid phase was colorless} with 2 Jhrck), 96-97% sulfuric acid (reagent parts 96-97% H,SO•• Emulsions were broken grade. Dupont), sodium carbonate (pow­ with a few drops of saturated Na,co. dered) (reagent grade, Mallinckrodt), and solution. ethanol (pesticide grade, Matheson, Coleman (7) The hexane was extracted with 1 part and Bell) were used. water, and several grams of Na,CO. were Activated alumina waa Fisher A-1540, added to the hexane. activated at 180 ° C for 24 hr. (8) The hexane was filtered through a The gas chromatograph was a Bendix Mo-, column of Na,CO. (100 mm x 10 mm id for del 2200 equipped with a thermal conductiv­ 300 ml hexane), the Na,CO, first being pre­ ity detector. The column was 6% SE--30 on washed with several milliliters of hexane. 60/80 Chromosorb W, 2 m x 2 mm (id) (9) The hexane was concentrated to 3-4 stainless ateel. The trap for preparative gas chromatography was a 160 mm X 1.6 mm ml (Snyder column). (10) The hexane residue was chromato­ (id) glaaa tube packed with SO mm of rlau graphed on a column of activated Al,O, (50 wool. mm In a 6 mm disposable· pipet). The column An Associated Electrical Industries MS-9 should not be prewashed. Elution w:iis with- 12 double focusing mass spectrometer and a ml of 20% cct. In hexane, then.1 ml of hexaVarian 1024 time-averaging computer Inter­ 0 faced with the MS-9 as described earlier . ane, and finally .4 ml of 2 % CU.Cl, in· hexane. were uaed. (11) The 20% CH,CI, fraction .was con­ Cleuap Procedure for the Analysis of centrated carefully to about 60 ,J, l�0 TCDD bl Tiaaue Santplea ,.J benzene added, and concentration re·· peated to 20 "'· (1) The aariiple was wefrhed and homo­ (12) A few micrograms or' m-terphenyl senned with U--1.2 parts EtOH. (2) Thia hon.ogenate was transferred to in . benzene were added · to the residue and a round.bottomed flask equipped with a re­ the mixture subjected to preparative chro­ flux condenser (Tefton tape used on the m•toirraphy. The retention time of tn-ter­ phenyl relative to that of TCDD was deter­ ground glass joint). The sample was spiked 1 mined beforehand and used to make certain with appn)Ximatety 1000 ppt 'CI TCDD; 2 that the TCDD collection was carried out at parts 40% aqueous KOH were added, and thia mmure was reftuxed for 2 hr. One the right retention time. part always. ref'en to the original samples. (13) The GLC trap containing TCDD (3) The aulution was partially cooled and was eluted with 601'I followed by 10,,.1 of benzene. The total amount of eluant collected l put hexane added. <•> Theiolution was transferred to a sep.. was measured, and the fraction size for the araiory funnel, ·and the phases were sepa­ planned number of fractions (typically ten) rated. The aqueous phase was extracted with calculated. (14) The fractions for TCDD analysis three more identical portions of hexane; the hexane extracts were combined and collected were prepared in. the sample tubes described previously ("). A known amount of TCDD in the oririnal round-bottomed ftask. (6) The hexane phase was transferred to was added to three or more fractions for the separatory funnel, the round-bottomed quantitatlon o: any TCDD obs�rvcd. The flask was rinsed twice with a few milllliten amount of TCDD added per fra�tion for m. Z8 BnYlronmental Health Perspectives quantitation should be approximately three or four times the amount expected to be present. (15) The fractions were analyzed with the MS-9 instrument. Typical conditions were: aource 220 ° c, resolution 10,000 (based on i. 10% valley between peaks), trap cur­ rent 1.0 mA (rhenium filament), electron multiplier 700, ionizing voltage 70 eV, time averaging at four scans per second. (16) Peak heights were measured at m/e 821.894. The quantity of TCDD (picagrams), present in the fractions to which TCDD haa not been added was computed from the ratio of their mean peak heights to the mean peak heights found with added TCDD. (17) Steps (14}-(16) were repeated, but "'Cl TCDD was added and peak heights were measured at m/e 827.885 in order to compute the amount of "Cl TCDD recovered. The recovery through the complete clean•tp pro­ cedure waa then calculated based on the amount of "Cl TCDD added to the sample at the beginning of the cleanup•. (18) The quantity of TCDD com'Juted in atep (15} was corrected by th� recovery . factor obtained in atep (16) to give' the final result. Sample Collection Freshly caught ftsh and crustaceans were coll.?Cted in South Vietnam. in August and September 1970 from local &Immen. The samples were homogenized with a meat vlnder, placed in acetone-rinsed glass bot­ tles with aluminum foil-lined caps. and im­ mediately frozen in solid co•. Later on the aame day, samples were placed in a Linde LR-85 liquid nitrogen refrigerator where they · remained · until analysis. Water blanks were present in the liquid nitrogen refrigera­ tor throughout the storage period and were analyzed with the samples. Fresh Cape Cod butterfish (Poronotua tricanthus, family Stromffttidae) were obtained from a focal market, homogenized, and kept· at -20° C until analysis. Domestic beef livers were ob­ tained and treated similarly. September · 1973 Results Methodolou The masa spectra of natural and "Cl TCDD are shown in Figure 1. The most intense signal for natural TCDD occurs at m/e 821.894 (nominal m/e 822), correspond­ ing to the isotopic isomer with one atom of "Cl and three atoms of aac1. The natural abundances of the Cl isotopes are 75.58 and 24.47%, respectively. The observed spectrum for the synthetic "Cl TCDD corresponds to an isotopic purity of 95.5% "Cl, the same as the value claimed by the manufacturer (Oak Ridge National Laboratory) of the NaCl used in the synthesis of the labeled TCDD. Tha synthetic 11Cl TCDD contributes only 0.042% as much to the peak at m/e 822 as it contrib­ utes to its most intense signal at m/e 828. The contribution at m/e 820 is even lower, by a factor of nearly 100. This allows an exceaa of "Cl TCDD to be added to each sample before cleanup without interfering IOO A. ll> • - ' -r ,.. 40 t: z ! i IO ..... 0 IOO IIO IIO -· � .. 1 w- .....,_. .oa•:;. 40 IO 0 IIO J I IOO - IOO I IOO FIGt:u 1. lllaaa apeetra of (A) TCDD and (B) "Cl-labeled TCDD. The Isotopic purity of the "Cl is 95,69<. The aaterlllk denote• an impurity, Tbe multiplicity ot lines aaaoclatecl with each ma­ jor molecular apeclea resulta from the preaence of Yarioua laowpea of Cl and C, fQ .. L Jl'l,IIOO' 521-900 ""- ·m.a ._ I. llaA 1peetra lhowlnir reduction ef DDE ... PCB levela ln.ftlh reaidue by means of ahmsma .-tography. Following the aulturie Kiel. , eliuap atep, the naldue In henne Is a4ded tD a mlama of activated alumina: (.i) True tr--. tile · aaterial eluted by 20% CH,Ct. in huue after 0. ealama wu ftnt eluted with 20% CCl. in baane; (B) trace obtained from a limilar 20� CH.Cir iD-Mune elution after the column wu ftnt e11i1al widl 1" CB.Cl. in hexane. Elutioa wit!, 1� CUA in hexane waa reported to be e�ecm. ia ·nchlcinir the amount of PCB reliduea (S). Emioa .witlt 20% CCI. la clearly even more e,ftctt,,e ud - routinely ued In obtlinlnir tile naia1ta -. .... here. with analyaia of natural TCDD at •I• S2Z and 320. The addition of 17Cl TCDD provides a carrier and makes possible the calcula­ tion of absolute recoveries. An alumina chromatography step hu been developed · which, when combined with the cleanup atepa described previoual7, (-0 • makea poaaible the measurement of pico­ pam quantities of TCDD in samples initially containing more than a millionfold excess of DDE and PCBs. Figure 2 shows the ef­ fectiveness of this procedure. The calculation of TCDD levels described iJl steps (14)-(16) of the experimental sec­ tion assumes a iinear relationship between · peak height and amount of TCDD present in any given sample. Figure 3 demonstrates that the response is indeed linear over the · full ·range of TCDD amounts introduced into the MS:.S in the course of the analyses re­ ported here• The reproducibility and overall recovery of the complete analytical procedure is illu­ strated in Table 1. A sample of beef liver was homogenized and divided into three portions each of which was then spiked with 20 ppt TCDD and 1000 ppt "Cl TCDD. The three samples were independently put through the cleanup procedure up to the GLC step. Each sample wu then split into three portions before preparative CLC and mass spectromet­ ric analysis, giving rise to a · total of nine separate -values for the recovery of both TCDD and "Cl TCDD. The average re­ covery waa 84 :I: 7% for TCDD and 27 :!: '" f�r ITCJ TCDD. When the slight back­ ground ai,rllal at ml• 822 in an l!!.lpilced 40 eo eo 1COO '"' IOO 1111 MO :navu a. Llnaarft,, of napo1IM for TCDD In th• preance of beet liver realdue. The TCDD Yaluea an the amounts introduced Into Individual runa ea the KS-I. En'ff:ronmen1al Health Perapectivea i. Talle I•. lleeeftriea at TCDD (..Wed at II t,pt) u4 "O 'l'CDD (added at 1000 ppt) from beef U-... Recover;,, "' TCDD "Cl TCDD Sample A. 14 GLC l 88 80 GLC2 88 GLC8 Sample B 81 28 GLC 1 81 GLC 2 14 IO GLC8 Sample C ff GLCl 81 GLC2 40 8'7 11 GLC8 ff :t LO 84 :t '7.2 forA,.B.ani C ,., .... _..,. • • • aample of the $&me liver is taken into ac­ count, the calculated recoveries from the spiked samples become even · more nearly· equal. Experiment.a perfoffl!C!d separately with each individual deanup step established that the step with lowest recovery is prepara. tive pa-liquid chromatography. ·We conclude fro:n these and other controls that the present . analytical method pro­ . 'rides the aenaitivity and reproducibility· re­ quired f1)r biologically meaningful analyses of animal tissue samples. The method makes poasibie investigations of such samples at levela approximately 10-- times those report� edheretofote (6). ...... es.,, Observed TCDD l..eftla Signals at m/e 820 and 822 were con­ spicuously present in each of the ftsh and crustacean samples from Vietnam. The cal­ �ted levels of TCDD, summarized in Table 2. range from 18 ppt to 814 ppt, based on total wet body weight. No peak was observed at m/e S20 or Si?2 with Cape Cod butterfish. The background aigna) corresponded . to a level of · S ppt of TCDD. No peaks were observed in water blank samples present.in the iiquld nitrogen refrigerator throughout the sample collec­ tion and storage period. Confirmation that peaks observed at m/e 820 and S22 are in fact produced by TCDD ia routinely provided · by the criteria out­ lined in part A of Table S. All three of these criteria are met by the mass spectra from each of the Vietnamese samples. The additional confirmatory procedures listed in part B of Table S were carried out on a sample of Vietnamese fish. This sample. carp from .the Dong Nai River, exhibited a . mean TCDD level of 540 ppt. The mass spectrum in the region mle 822 is shown in Figure 4. The compound observed in this fish behaved identically to TCDD in each of the three additional confirmatory tests. · We consider it extraordinarily unlikely that this compound is anything other than a tet­ rachlorodibenzo-p-dioxin. In contrast to the . aipiftcant amount.a of ZS,7,8-tetrachlorodi- - Lnel, ppt total wet bod7 wefaht• m lien II I 820 610 640 GO 810 1020 610 Do..- Nal JUwr (interior) (CJprilliue) DIIBg Nai Ri•er (interior) Catft•h (Silaridae) O.W Nai Rlwr (interior) Catftah (Tadi:,suridae) no no B Bai Goll Ri•er (interior) '70 u 89 Catdah (Schilbeldu) C Sal Con JU.er (interior) 42 Rhv Prawn (Palaemonidae) 84 C 49 Cu Glo Village (aeacoa1t) 49 Croaker (Sclunldae) 110 D 18 Cape Gio VUJap (-,t) lll 14 Prawn (Pneidae) D Cape Cod, xa..aclimetta Buttedah (Stromateldu) �8 • Letlien refer to sites on map in Figure 5. • ._.. numerals refer to independent eleanupa of di!rerent portion• of the - umple. All •al- ue � for reeo-,_er;r. A B .,. ,, f A. Boutlne L P'ollowa •a TCDD tmoach � apeei& eleanup I. Bu apeeted mau (:tW -u) at •I• 820 andBH &. Hu expected ratio of laotople i-ra at fll/e 820 aud 822 B. AddltlonaL• L lt··COCI tra1111entatlon puk haa apeded IIIU8 and laotople laomer rado I. Percent �very after partial photol,tle d• eompoeltlon equals that ot "Cl TCDD (1, 8) &. Partition coefficient betwwn dichlorometlaane1*!ane and aeetonltrile equala tllat ot ·"'Q TCDD (r). • Stepa t and 8 ot the additional procec:haffl! _.,. earried out on the dichlorom.,thane-hexane eluam from the alwnlna ehromatosrapb1' prior to preparative GLC. _/\·--.. A·-�- ____ m.11111 Jl1,911J m.a l'IGvllll 4. TCDD alpal1 o'baernd In ftah umplea: (A) V-ietnan:eae carp plus 60 pg Teno;· (wet 'ftlpt of ftah G.18 g); (B) Vietnameae carp, (wet -18'ht of ftsh 0.18 I'); (C) Cape Cod butterllsh (wet weight of ftsh o.is s>benzo.p-dioxin known to have been: dissemi-­ nated aa a contaminant of 2.4.5-T (Z), we know of no likely route by which-other isom­ ers of TCDD might have been introduced into the Vietnamese environment. The locations frc:n which the Vietnamese samples were obtained are designated in Fig­ ure 5. The letters correspond · to those in Table 2. Areas heavily ti:eated with 2,4,5-T before its use was ordered discontinued in az FtGVD 6. Kap ahowlns umpllq 1ltes In relation to rivers and principal a�-.ayed areas. Sites A and B are loeaW on the Doq Nai River, aite C i1 ea the Sai Goll River, and site D i• on the coaat at Can Gio. .Sprayed areu are depicted only wltJiln the reaioa bowulecl by tile .daahed lines �-� April 1970 are shown as stip�. The number of samples is not adequate to permit relia,ble conclusions concerning the differences be­ tween various locationa and species, although this certainl7 should be a subject of future studies. Discussion Considering the limited number of. um­ pies we.have analyzed and the fact that they wen collected 2¾ yr ago, it does not seem appropriate to attempt any detailed evalua,­ tion of the possible toxicological significance of our results. Such discussion is zna,de even more difficult by the complexity and in­ completeness of the existing toxicological data. However, in order to provide persp�­ tive for such discussion, a tabulation of some of the principal toxicity data on TCDD is presented in Table 4. It may be noted that gui�q pigs consuming their we.ight of food contaminated with TCDD at a level of 600 ppt would have ingested a quantity cor­ respon�ng to the lethal dose. In contrut, a far greater quantity of TCDD is required to rea,ch the LD,. cited for rats•. The table shows that teratogenesis in the rat occurs at doses substantially lower than those re­ quired to kill Blniroamental Bealtla Penpeetivea ·--·�·· 2· .. _.fl ·i"'7f"'·- rm ,iiti'"' t'· ., ...... ·'·m·· Letlaaliiy Female nt, mgJe oral doN LI>. (obeinatiOM terminated at « da,a) Kale nt, 1inp, oral dON LO. (obllf'rvatiom terminated at « da,a) Kale guinea Pie, single oral doee LD. · (obaervatiom terminated at IIO da,a) Tentoceniclty Cleft palate la &O� NMRI mice, dally oral. doN, cla7a6-15 lnteatinal hemorrbap and 111beutaneou1 edema ln II091, Sprasu•Dawlq rata, dally oral doae, da,.. 6-16 �•ma and death in chicken eml,170, lincle ·Injection E-,- Induction Doublinc of f.aminoleYUlinlc acid IIJ'llthetue ·1n clllcua embl')'O. aincl• Injection Kllcltic arnat � endoaperm. ambient coneeatratloa Feeding studies in monkeys show that dioxin poisoning is cumulative (18). Various levels of a toxic fat known to contain chloro­ dioxins were incorporated into the daily diet . of .IICacaeG mulatta monkeys. As pointed out by the investigators, the mean survival time depended inversely on the daily dose. A plot of their data (Fig. 6.) conforms rather well to the relation T •KlP .. + K', where T is mean survival time, D is -daily dose, and K . I( •IC' ·.r-•, 100 ' • I 1£CIPROCAL OF Tl£ PERCEflT TOXIC FAT IN DIET J'I01JU I. Mean aurrival time of monkeJ'I fed toJdc fat plotted apiMt the reciprocal of the per cent of toxic fat preaent ill the diet (ll). September 1973. TCDD to obtain effect, ppt body weight Reference 41,000 (I) 11,000 (I) eoo (I) 1,000 UMOO (I) (I) IO (IO) IO (It) nta of which spindle fibers are constructed and which are ubi­ quitous in their structural roles in cell ex­ tension and cell movement. Acknowledgement This research was initiated by the Herbi­ cide Assessment Commission of the Ameri­ can Association for the Advancement of Science. The work has been supported by funds from the AAAS, the Ford Foundation, and the National Institute of Environmental Health Sciences (NIH grant no. 1 ROI ES00851-01 TOX). We thank Professor Bui Th! Lang, Mr. Robert Cook, Professor Arthur Westing, and Dr. John Constable for aiding in the collec­ tion of samples, and Professor Larig for helping to identify the specimens collected. We also thank Kenneth Gross and Lesley Newton for their expert. assistance in the laboratory. REFERENCES l. Report on 2,4,&-T. A Report of the Panel on Herbicide& of the Prealdent's Science Advisory Committee, Executive Office of the Pruident, Offla. at Science and Technology, March 1971. I. EA'eeta ot 2,4,&-T on Man and the Environment. Hearlnp before the Subcommitt.ee on Energy, Summary Natural Reaourcea, and the Environment of ti!• A procedure has been developed for the Committee on Commerce, United Statea Senate, e reliable detection of TCDD in animal tissus 911t Consr-1, April 'l and 111, 19'10. down .to levels approaching 1 ppt. It makes IJ. Spanchu, G. L., Dunn, F. L., and Rowe, V. K. use of chemieal cleanup, preparative gas. Study of the Teratogenicity of 2,3,7,8-Tetra­ i chlorodlbenzo-p-dioxln in the Rat. Food Comiet. lquid chromatography, and analysis by time­ ToxieoL 9: 406 (1971). averaged high resolution mass SJ)eCtrcscopy. C. Baughman, R.; and Meaelilon, M. An Improved ·. A limited number of fish and crustacean � analysis for 2,3,7,8-tetrachlo_rodlbenzo-p-cl!oxin. samples was collected in South Vietnam in In: Advances In Chemistry Ser. JZO, E. Blair, Ed., American Chemical Socletr, Wa1hlnston, 1970 near areas heavily exposed to the herbi• DC., 1978, p. 91!. cide 2,4,5-T. TCDD was detected in these S. Porter, M. L., and Burke, J. A. Separation of aamples at levels ranging from 18 to 810 thro chlorodibenzo-p-dioxlna from aome poly­ ppt. TCDD was not detected hi a sample of chlorinated biphenyla by chromatography on an Cape Cod butterftsh used as a control. aluminum oxide column, 1971. J. Aaaoc. Oftlc. Anal. Chem. IIC: 1'26 (1971). These results suggest that TCDD may have 8. Woolson, E. A., Reichel, W. I., and Young, A. L. accumulated to biologically significant levels Dioxin residues in lakeland und and eagle in food chains in some areas· of South Viet­ aample1. In Advances in Chemistry Ser. UO, nam exposed to herbicide spraying. E. Blair, Ed., American Chemical Society, Waahlncton, D�C., In preu. Note added In proof: Overall recoveries have 7. Woolson, E. A., Thomas, R. F., and Ensor, P. been increased to 60-80� by replacing the D. J. Survey ot poiychlorotlibenso-p-dloxln con­ GLC step with an additional AlaO, column tent In 1elected pe1ticide1. J. Agr. Food Chem. step. Details of this procedure will be de­ IO: 8111 (1972). scribed ·fn a future publication. & Croaby, D. G., Wong, A. S., Plim�r, J. It., and Environmental Health Perspectives . .. .. Woo!aon, E. A. Photodecompoaltlon of ehtori­ aated dibemo-p.dloxbia. Science 171: 748 (1971). t. Neubert, D,, and Dll!inan, I, Embeyot,oxic eft'ecta In mice treated with 2,4,5-T and tetrachloro­ dibemo-p-dloxln. Naun,n-Schmlaclberp Arch. Pbannacol. 272: .243 (1972). 10. Verret, J. In: Eft'eeta of 2,4,5-T on Kan and the EnYilonment. Hearinga before the Sub­ committee on Enern-, Natural Re.ourcea, and 'the Environment of the Committee on Com­ --. United States Senate, 91st ConlftU, April f and 15, 1970. 11, Poland, A., and Glonr, E. 2,ll,7,S.Tetrachloro­ dlbe-p.dloxln: a potent Inducer of ,.amtno­ lnulinlc add qnthetue. Sdnee ·171: 2'8 (1978). 11. Jacbon, W. T. Reculatlon of mltosfa: lll. Cyto­ Joaical d'ecta of 2,4,5-T and of cllozfn eontam- September 197i 18. 14, 15. 18. lnanta In 1,4,5-T t-ulatiou, J, Ctll 8d. 11: 15 (1972). Allen, J. R., and Cantene, L. A. Llpt and electron microacoplc oblervatlona . in M­ ..zott.. monke,a fed toxic fat. Am. J. Vet. llaa. 18: 1&11 (1967). Herbicide Alaeaanent Conunlulon. Background Material Relevant to Preeeiltatfona at the 1970 Annual MeetlJI&' of the AAAS and Preliminary Report of the Herbicide Aueumenl Commla­ alon; both reprinted !n ConrreuiOllal Record 118(12): 83226-1288 March a, 1972. Baa-Ho!, N, P., et al. Orran• u ·larireta of dioxin intoxleatfon, Naturwlu. 59: 174 (1972). D..,r!q, L., and . Bunner, M. C,tosenetlc ef­ tecta of 2,4,5-tril'hlorophenox,acetlc acid on 01119nesia and eul7 embryo�neala In Dronplila -'4M,ulff. Hereditu 18: 115 (1971), .. Analysis of Two Hexachlorophene and Two 2, 4, 5Trichlorophenol Samples for Tetrachlorodibenzo-p-dioxins Robert W. Baughman Lesley Newton Department of Chemistry Harvard University Cambridge, Massachusetts 02138 l June 1972 INTRODUCTION The samples analyzed in the present investigation included two samples of 2.4.5-trichlorophenol, designated phenol F-980 and phenol F-995, and two samples of hexachlorophene (HCP), designated HCP F-981 and HCP F-996. The samples were received from Dr. David Firestone, Division of Chemistry and Physics, Food and Drug Administration. The analyses were carried out with the method (MS/CAT analysis) reported earlier. I METHOD Reagents and Apparatus (a) Hexane: Pesticide grade, dist1lled in glass (Fisher Scientific). (b) Dlchloromethane: Reagent grade (Eastman). (c) 85% H SO4 solution: Prepared from reagent grade concentrated 2 H2so4 (Dupont, 95-97%). (d) Na2CO3 (rowdered): Reagent grade {M:-:.llinckrodt). (e) Ethanol: Pesticide grade (Matheson, Coleman and Bell). (f) Activated AI o3: Fisher No. A-540, activated at 130 ° for 24:hr. 2 (g) Gas chromatograph: A Bendix Model 2200 gas chromatograph equipped With dual flame ionization and dual thermal conductivity detectors was used in the present investigation. (h) Gas chromatographic column: 5% SE-30 on 60/80 Chromosorb W, 2 m x 2 mm (id) stainless steel. (i) Trap for preparative gas chromatography: 150 mm x 1.5 mm (id) glass tube · packed with 30 mm of 100/120 mesh glass beads (glass wool plugs). 6) Mass spectrometer: Associated Electrical Industries MS-9 double focussing mass spectrometer equipped with a Varian 1024 time averaging computer. PRO CEDURE l) Weigh out 5. 00 g of phenol or HCP in a 250-ml Erlenmeyer flask. 2) Add 100 nil 1.0 !{ NaOH and warm solution on steam bath to dissolve sample. 3) Cool and extract with four 35-ml portions of hexane in a 250-ml separatory -2- '-' funnel. Collect hexane fractions in original 250-ml Erlenmeyer flask. 4) Transfer hexane to 2 50-ml seperatory funnel (rinse Erlenmeyer with two S-ml portions of ethanol followed by two 5-ml portions of hexane, heating on steam bath each time) and extract with two 20-ml portions of 1,0 N NaOH followed by one 20-ml portion of water. ,5) Extract hexane solution with three 30-ml portions of 85% H2SO4• 6) Extract hexane with one 20-ml portion of water and add S.O g of powdered Na2CO3 to hexane with swirling. 7) Filter hexane solution through 60 mm of powdered Na2 co 3 in an 11 mm 1d column. Prewash column with 30-ml clean hexane. Rinse separatory .funnel with 20 ml of clean hexane and filter this through the Na2 co3 column. 8) Con ,:a:ntrate hexane (Snyder column) to 1.5 ml. 9) Chromatograph hexane residue on activated AI2o3 (35 mm) in a Dispo­ plpette. Do not prewash column. After eluting hexane, elute with 6 ml of 1% dlchloromethane in hexane. Then elute with 6 ml of 20% dlchloromethane in hexane. · 10) Concentrate the 20% dichloromethane fraction carefully to about 25 µi. ll) Add 100-200 µI benzene, depending on the amount of solid residue. Dissolve any solid that may be present and store the benzene solution in a screw cap vial with a metal foll cap liner. l2) Analyze 1-4 µI of the residue on a suitable gas chromatograph. The glc column described previously, operated at 250 ° C and 30 mVmin carrier flow, was used for the present analyses. 13) Obtain a mass spectrum of the residue (the solid obtained by evaporating 1-2 µI of the benzene solution on. the tip of a direct insertion probe should provide an adequate sample) in the range m/e 0-600. Resolution of 1,000-3,000 is sufficient. 14) Preparatively chromatograph approximately 10% of the total residue by glc. Collect the eluant in the region of the TCDD retention time. See Reagents anc;I Apparatus for a description of a suitable trap. The trap should be wetted with hexane and cooled in dry ice prior to collection. Warning: It is necessary to collect all other regions as well since toxic compounds other than TCDD may be present. For these regions an uncooled, empty 200 mm x 1.5 mm id Pyrex tube should be adequate. The -3'-" most convenient way to carry out the glc prep is through a thermal conductivity detector. Then a small amount of an internal standard, such as !!l-terphenyl; whose retention time relative to TCDD is known, can be added to make certain that the TCDD collection is carried out at the right retention time. 15) Elute the glc trdp containing TCDD with 60 ml of benzene. Measure the total amount of eluant collected and calculate the fract1011 size in µl for the planned number of fractions. In the present study 12 or 15 fractions ·of 3-4 µl typic:;lly were prepared. 16) Prepare the fractions in the sample tubes previously described .1 Add a known amount of TCDD to three or more ,of the fractions for quantification of the observed TCDD. The amount of TCDD added per fraction should be approximately the sam� as the amount expected to be present in the residue. 17) Analyze the fractions by the time averaged, high resolution mass spectral method reported earlier. l All analyses reported here were carried out at the m/e 322 isotope peak of the molecular ion of TCDD. Resolution of approximately 10,000 ·was used, based on a ten per cent v.alley between resolved peaks. RESULTS AND DISCUSSION MS/CAT Analyses for TCDD The results of the analyses for TCDD are summarized in Table l. In phenol F-995 the level of TCDD was high enough to be determined by flame ionization (FID) glc analysis. The level observed in this phenol was approximately six times higher than tho level of 70 ppb reported by other workers. 2 Since such high level residues were not of primary interest-in the present study, recoveries using glc analysis were not determined. Possible the glc peak on which the quantitation was based contained other compounds in addition to TCDD. It is true, however, that recoveries for the procedure which presumably was used in the previous analyses are somewhat low, on the order of 30%. 3 . Phenol F-980 contained a much lower level of TCDD, 1, 6"!:: 0.8 ppb, than · did phenol F-995, HCI' F-'996, prepared from phenol F-995, contained 3.8±1.9 ppb -4'-' TCDD. For HCP F-981, prepared from phenol F-980, a TCDD level of O.54 ± 0. 2 7 ppb was observed, but since this level is close to that observed in the reagent blanks, 0 ;13 ! 0. 06 ppb, it is safer to say that in HCP F-981 the TCDD level gid not exceed 0.54:!:0.27 ppb. Before it was apparent that the levels in HCP F-981 and the reagent blanks would be so high, samples of HCP F-981 spiked with 0.2 ppb of TCDD were worked up to provide a test for the ncovery of TCDD. Although the value obtained, 0 � 76 ! o·. 38 ppb, is in good agreement with the TCDD level_ in unspiked HCP F-981, a higher spiked level had to be used to measure recovery. With 20 ppb of TCDD added to HCP F-981, the observed level was 11.2 ppb, which represents a recovery of 56%. Throughout the analyses background of the s_ort that was present in the reagent blanks and interference, caused by a nearby fragmentation ion, were relatively high. 'l'he hexachlorophene samples contained a significant amount (much greater than 1 ppm) of a neutral impurity that, from its exact mass and isotope pattern (discussed below), appeared to be one or more isomers of hexachloroxanthenc (I). Either the hexachloroxanthene or some other compound in the hexachloraphene _ Clm·�-� . ? - �o...V-- Cln n+m=6 1 residue produced ions with exact masses close to those of the molecular ion -of TCDD. The level of the hexachloroxantherie was so high that it became a constant component of the glc eluant. Analysis of preparative glc blanks indicated that approximately half of the background and interference in the reagent blanks was introduced in the glc cleanup step. Wide scan, low resolution (resolution 3,000) mass :;pectral analysis indicated that non-negligible residues of hexachloroxanthe.ne, about 1-10 ppb, were present in the reagent blanks prior to the glc cleanup. These residues most likely resulted from cross contamin-3tion from glassware. (The same glassware ... was used for the phenol, HCP, and reagent blank cleanups.) Representative �races from the MS/CAT analyses of the various samples are presented in Figure 1-5. · ·,. y· r , · tt S? -�r '-zttic �.'1 · ·-,, r · ·�� -s'-' In the present analysis each fraction analyzed by MS/CAT analysis corresponded to about o.os g of original sample (10% of final residue glc prepped, divided into ten or m ore fractions. with an original sample of S. 0 g). The minimum detectable amount of TCDD was about 2 pg. This provided a limit of detection of approximately 0.04 ppb. For other types of samples, such as milk, liver, or fish, background and interference are less severe, and fractions corresponding to about 3 g of original sample can be used. With a minimum detectable amount of about one pg, this provides a limit of detection on the order of one ppt. If greater sensitivity were required for other phenol or HCP samples, methods could be devised to reduce the levels of the interferences to provide a similar limit of detection. Wide Scan, Low Resolution Mass Spectral Analyses In Tables 2 and 3 the major molecular and fragmentation ions for the phenol and HCP samples are tabulated. In phenol F-995 there was a very strong signal for the molecular ion of TCDD (Cl4 isotope pattern at m/e 320; base peak 11,/e 322; observed exact mass 321.8932, theoretical 321.8940), which suggested that the glc peak observ�d (see below) at-the :rcDD glc retention time was in fact due to TCDD. In phenol F-98o·a glc peak ata sllghtly later retention time than TCDD did not give rise to any si�ificant signals from polychlorinated ions over the range m/e 500-250. At m/e 204 both phenols containe(f a relatively large amount of a dlchloro compound which did not correspond to any of the expected methoxy or dlmethoxy derivatives. Two molecular.formulas which correpond to a molecular weight.of 204 are CsH202Clz at m/e 203.9752 and C9H10oc12 at m/e 204.0122. The observed value was 203. 9790, which is in much closer agreement with the first fonnula. A possible structure for this fonnula is dichlorodibenzodioxane (V} • - II Eliminating the second formula rules out such compounds as dichlorophenylpropyl ether and ethyl- or dimethyldichloroanisole. It is unlikely that the ion at m/e 204 1 ft#" '; .· "'·ffl ") .. ,.,., ,, . ¥f" tr� -6- '- 1s .a cyclized fragmentation product formed by the loss of H2 from a dichloro-1, 2dimethoxybenzene, since such fragmentation is not observed for unsubstituted 1,2-dimethoxybenzene.4 In. view of the similarity of V to TCDD, it may be of interest to investigate further the structure of this impurity. Both of the HCP samples e::ontained relatively high levels of a hexachloro compound with a molecular ion at m/e 386. At 388, the second isotope peak and also the molecular ion for hexachlorodibenzo-p- 1. O ppm ++ + * Molecular fonnula confinned for F-995 by high resolution mass spectrometry (resolution 30,000) ++ Present at O • 01 to O .1 ppm ++ N II 1' ( ( B C Flee·1. A. MS/CAT analy•i• o� phenol F-9Sq cleanup A, range 2 17 • B. MS/CAT analyeie ot phenol F-980, cleanup B, range 2 '·. C. MS/CAT aqalysis ot' phenol r-98·0, cleanup B, plus 20 pg TCDD, range 21s. Conditions: Source 200°, resolution 10,000, 0 • .5 mA t'ilament current, electron multiplier· �oo, ionizing voltage 70 eV · (same conditions t'or all t'ollowi�g analyses). .. .,aj,, w­ . , .. �·�1 ,..,��-•••"-· w-ar t'rf:Mt"Vt·;y 'Jlfft:\.,._ �-..... ,.; , �--�...... Jitf·Wil'iti'tt' ·.ali!i.oa--ii..i nu:ti'.{ �liili!i-liliili� ini'M'\ TfHi#........ ,7 t � . ""'. :ca) IQ rft· • z. � IO. � IQ C ...=• • Cl, :, 0 -0 °'I e \C)rO\Clt .... IO . c i ... Do. .�... �·!ct tlQ ...••� ...... •• 1 0 \C) ·= (ft ... ED. f -� � "••... ....• ;,,. '14 C ( ( .J . ··.·� Figure l• A. MS/CAT analyeh or HCP F-9814 ele-,iup A, range 2 1 ). B. MS/CAT analyeis or HCP F-981, cleanup B, range 21 • 'l'he unusually high noise was cauaed by problem• in the electron 111ul.Uplier t)ircuitry or the HS•9, It doe• not effect significantly the measurement or TCDD levels. ( ( .A B ·1 ... Fi«ure 4. A. MS/CAT analyeie o� HCP F-981 epiked with 0.2 ppb TCDD, cleanup A, range 215. B. MS/CAT analyeie o� HCP 1"�981 epi�ed with 0.2 ppb TCDD, cleanup B, range 21s. · .r nm: QJt8ilT!'fRttffff W tYtMTTMR nrnsres··-,·trn1ttffart&ttfftfttrrtWWf'f'#trtfafttitt±hr:riilttti± ·,a, :s....... .. ( ( 1, Iii,'. it A. !- ; I 1· l B !1_· Fi re • A. MS/CAT ana1yaia ot a reagent blank, cleanup A, range 2 1 ,. B. MS CAT analy•i• ot a reagent blank, ole•nup B, range 213. See Figure 3 tor •n expianation ot the high noiae level. I ( •. MC,.- - •• " ,v TV T"ll: 11111:NTHilfftllt ''..: ,. • ...... •wu,PQ.a-111100. . . . , ..a; j ... ",\$-,h, ,·. u,;;, .&.)iPUJ •• .."·' f� ' 1 , .. 11111 ,.. . .. ... .., .· -· . ' .... . ... , , " '" , I . ,,,,,,,,,, .. , ··:,, ,,. ·,,r." ,., '''Lj ,-,1 ( ... .-., ' . ,.: ,f ,.,;• : trer r:rr::r:mmr err rn :n • Wtft'&:Wtrrt: :&¥':t@ttr'tt'frr: r w:: (:iibt : 111Wriii%ft � - 1 I ( ( j ·�� i :, f 2 I it r • 1 I A B I l j ·, ·1· . l"i re 7• A. Gas chromatogram or (1-) 10 ng !!!-terphenyl internal standard and (2r 100 ng TCDD. B. Gas chromatogram o� a reagent blank •. FID glc conditions a si SE-)0 on 60/80 mesh Chromosorb V, 2 111 X 2.2 mm id stainless steel column; column temperature 21'5 ° , helium carrier· f'low rate )0 ml/min, signal attenuation 20X, injection or one JAl of' concentrat_ed residue or 2� CH2Clz fraction from alumina chromatography. r ·i r- ... a • .. .., ; 0 I .,• 0 C • --c·• IQ t ...-• § 0 0 co 0\ ...�g J D, It 0 :IQ It . -c !: 0"'4 • .! c1o .. • 0\ -C I f .zs: a:, 0 C), .... fM .. 0 • • ... ...•0 E ....a0 0 •• ' Ill SQ C) • -• IIQ r ...-• � . I • :�� 0 ...�I It\ 0\ 0\ • t � ...t,...:. 0 I OIIQ It 0 C: • ·- i: Uf"'f 0 C)"' 0\ •O\ � �l i i GI .E: Cl, ... '4 ... 0 l ( ·, A Figure 10. A. ·aa. chromatogram or phenol P-980 (cleanup A) plue 20 ng TCDD. B. Gas chroma�ogram ot phenol F-995 (cleauup A) plus 20 ng TCDD. • tiara 11. Gas chromatograi or H015 r9981 (cleanup A). Ittitf -s: .·.,, � 0 0 .,..... 0 ......!•• . ...• •i ..c • -� • � 0 • .. .Q,ez°.A!J -¥.;s,s,p.t ..!3;i�Li!i5,F-':$ f., .. 34. tE.A,1;;;_4)¥# . I ,� -• IQ ...-•• A, :, s: � •.. 0 ., 0\ I ."1! • •• I • .,• 0 - I 0 0 i ,,..A, '7¥ ·*··'·· i.J,Y:· 4_hffe •.hi dfyS .{M J), )? . _.,¥fett,;; . _¥)1,H.4, . __ ._k_?iifA!j_µ )f \c'll.\'t:;:,:;; ( ( it '� r.! · t.'\iii, " 1)I , I Figure 14. Gae chro.. togr.. or HCP r-996 (cl.eanup A). ·�A!hlft!,jfi)k.Yff.�:�:·�.�Wf,F t '*.!�<½.£?¥ l 4?.+£?.W.�1;�¥"',.R�.*---f-{AUA hfif(!(.i -A .1:PA41#f L�_..)¢41,f .Ji ,1:M.4-fi*i kf9, i,M�_.�P*9-� AW¥'!49l. ¥.41 _@,Alf%,. M,A•:?-�W k-4:¥4,S¾.P)!¢%)4i,. f!9,.--M:st{_·f¼#}A\9fJ4J.G!H#)P<,At44Aiit ,,. u;x;: .z a, ' %Ell tittliit:t rt ' • IQ ...-• 0 .. \0 0\ 0\ I Dt i "'0 I ..I &-., 0 0 JI0 •• 0 >< F i y' tr :e&k ANALYSIS FOR TETRACHLORODIBENZO-P•DIOXINS IN A FRENCH TALCUM POWDER-HEXACHLOROPHENE FORMULATION IMPLICATED IN THE DEATH OF A NUMBER OF INFANTS. Robert Baughman and Lesley Newton Department of Chemistry Harvard University October 1972 IB�DUCTION Recently in France a talcum powder foz:mulation reportedly containing six percent hexachlorophene (HCP) was implicated in the deaths of a nwnber of infants. Since chlorodioxins, especially tetrachlorodioxins (TCDD) may be associated with HCP,1 it is of interest to know the level of 'l'CDD p�esent in this talcum powder. We report here the results of analyses for TCDD performed on such a sample received from Dr. David Firestone of the Division of Chemistry and Physics, Food and Drug Administration. METHOD Reagents and Apparatus Bell). 1) Ether: Pesticide grade (Matheson, Coleman and -2-· 2) Hexane, Pesticide grade, distilled in glass (Fisher Scientific). 3) 4) 5) 130 ° Dichloromethane: 95-971 H2so4: Reagent grade (Dupont). Na2co3 (powdered): Activated Al2o3: for 24 hr. 6) 7) Reagent grade (Eastman). Reagent grade (Mallinckrodt). Fisher No. A-540, activated at Gas chromatograph: Bendix Model 2200 gas chroma- · · tograph equipped with dual flame ionization and dual thermal conductivity detectors. 8) Chromosorb 9) Gas �tographic column: w, 2 m x 2 11111\ 51 SE- 3 0 on 60/80 (id) stainless steel. Trap for preparative gas chromatography: 150 mm x 1.5 mm (id) glass tub.t packed with 30 mm of 100/120 mesh glass beads (glass wool plugs). 10) Mass spectrometer: Associated Electrical Industries MS-9 double focussing mass spectrometer equipped with a Varian 1024 time averaging computer. Procedure In a 10-ml Erlenmeyer flask 0.40 g of talc spiked with·lB ng '.rCDD�(c1 37 >4 (prepared by chlorinating dibenzo-p­ dioxin with (c1 37 )2) 2 was extracted twice with 2�ml of ether. The ether-talc slurry was heated for 10 min on a steam bath. The ether was pipetted off, and filtered through a glass -3- wool plug in a disposable pipette. The combined ether por­ tion was e�aporated to dryness and the resulting pale yellow. residue was redissolv�. in 5 ml of hexane. After being ext�acted three times with 5 ml 1.0 ! aqueous NaOH, 5 ml of water, two times with s ml of 95-97' s2so4, 5 ml of water and filtered through 30 mm of Na2co3 in a disposable pipette, the hexane was chromatographed on 50 mm of Al 203 in a dis­ posable pipette. The A12o3 was eluted with 5 ml of 11 CB2c12 in hexane followed by S ml of 201 cs2c12 in hexane. The 201 CH2c12 fraction was concentrated to about 20 ml, chromatographed preparatively. by gas liquid chromatography (glc), and analyzed for TCDD with the mass spectroscopic technique described previously.3 The level of TCDD was measured for the talc as a whole and for the HCP, assuming the talc contained 6 I HCP. A level for TCDD . in the talc and HCP corrected for losses in the cleanup was calculated based on the recovery of added TCDD-(c1 37 ) • 4 Before the French talc was investigated, duplicate samples of domestic talc spiked with 61 HCP and 8 ng TCDD­ ·(Cl 37) 4 were analyzed to confirm that · 'l'CDD could be recovered with the procedure described above. RESULTS The table below conta:l.ns a summary of the results of the analyses. Correc:ted for losses in the cleanup, in the . or11 ·:::mt·rrrr: -4l'rench talc the level of 'l'CDD was about 20 ppt and in the HCP, assuming the HCP was 61 of the talc formulation, about 0.30 ppb. In a sample of "clean• commercial HCP prepared in the United States a TCDD level of 0.50 ppb was observed,1 so the HCP in the French talc was not unusually highly con- taminated with TCDD. Level of TCDD in rrenc:h talc:. Cleanup A. o.S0g us talc:+ 61 HCP + Sng TCDD-(c137)4 B. • A. o.40g French ta}� + l6n9 _TCDD-(Cl )4 B, • Observed Corrected level of level of TCDD in 'l'CDD in talc (ppt) talc (ppt) Observed Corrected level of level of 'l'CDD in 'l'CDD in 61 HCP 6' HCP (ppb) )ppb) (Corresponds to about 0.5 ppm 'l'CDDc137 )4 in HCP and 2.7 ppb in talc) ,.,� 3.62 Recovery of TCDD-)Cl37)4 241 351 19.8 0.113 0.333 341 21.3 0.060 0.350 171 I .:J1'!.!Bl.iMR BtflMIG'"ijflif'ldrlifT'i s,1dg�-- --- �--ii�tniifr'·:'ftt'ljr(fedf'ftwi�'tt1H: v-,. ·et -·rm 7 X'iWMtt:if#tteite1ttffit:ftt ({f??!tfrfifti'r''IMNW-H· "ft{ W(1f1siti tw:t&Hi "� -6- REFERENCES 1) a. Baughman and L. Newton, unpublished results. 2) R. Baughman, unpublished results. 3) R. Baughman and M. Meselson, Adv. in Chem., 119, (in t,rei:s) • "" -.;. ...1 1$_ l , J MALYSIS·OF TETRACHLOROD!BENZO-P-DIOXINS IN SOME SELECTED.SAMPLES OF THE HERBICIDE FORMULATION "ORANGE." Robert Baughman and Lesley Newton Department of .Chemistry Harvard University November 1972 INTRODUCTION Three samples of the military herbicide formula� tion "Orange,• a one-t.o-one by weight llli.xture of the n­ bUtyl esters of 2,4-D and 2,4,5-T were analyzed for tetra­ chlo:rodibenzo-p..:dioxins (TCDD's). The samples were obtained from (1) Ft. Detrick via Dr. Fred Tschirley, (2) Pt. Detrick via Dr. Arthur Westing and (3) Dr. Bert Pfeiffer via Dr. Jacqueline Verrett. The intent of the experiment was to develop a mass spectroscopic analytical procedure for chlorodioxins in ester formulations of chlorophenoxy acid herbicides� This procedure cOJilplementa existing methods based primarily on electron capture glc analysis.1, 2 , 3 ,4, 5 1-- -2- .. METHOD · Reagents and Apparatus 1) Hexane: (Fisher Scientific). Pesticide grade, distilled in glass 2) 95-971 H so : 2 4 Reagent grade (Dupont). 3) �a2co3 (powdered): 4) Ethanol: Reagent grade (Mallinckrodt). Pesticide grade (Matheson, Coleman and Bell). 5) _Mass spectrometer: Associated Electrical Industries MS-9 double focussing mass spectrometer equipped with a Varian 1024 time averaging computer. Procedure In a 250 ml separatory funnel 0.50 g of orange plus 200 ng of TCDD-(c137 )4 (prepared by chlorinating dibenzo-p-dioxin with (c137)2) 6 was dissolved in 30 ml of ethanol. 'l'Wo ml of aqueous 401 (by wt) KOH was added. white precipitate appeared immediately. The slurry was A allowed to sit at. room temperat�re for 30 min with occa­ sional swirling. One hundred ml of water was added, which dissolved the precipitate, and the aqueous phase was extracted with three SO-ml portions of hexane. The hexane was extracted with three 50-ml portions of 1.0 ! NaOH, 40-ml of Yster, 40-ml of 95-971 a2 so4, 40-ml of water, -3- filtered through a 60 cm column (15 mm id) of powdered Na2co3, and concentrated (with a Snyder co;umn> to about_ 101 of its original volume. Fractions corresponding to 0.011 of the original sample, or 50 µg Orange _per fraction, were then analyzed by the mass spectroscopic technique described earlier.7 The level of TCDD with the naturally occurring c1 351c1 37 isotope ratio was measured and corrected based on the recovery of the·c1 37 l3.belled material added at the beginning of the cleanup. RESULTS The results are summarized in the table below. In all three samples the TCDD level, corrected for losses in the cleanup, was about 0.1 ppm. 501. The recoveries were about From other similar experiments, the error in the reported values is probably less than a factor of two. of TCDD in some samples of the herbicide !!!!!!· Level formulation Orange. Sample Observed level Corrected level 'l'CDD {ppm) 'l'CDD (ppm) Recovery of TCDD-(Cl 37 ) -I 1) Pfeiffer 0.0375 O.O66 57' 2) Ft. Detrick (Tschirley) O.O52 O.O91 57' 3) Ft. Detrick (Westing) O.060 O.146 41' -4- REFERENCES .1) K. s. Brenner, K. Muller, P. Sattel, J. Chroma­ �-,!!, 39-48 (1972). 2) D. A. Elridge, Analyst,!!, 721-727 (19'71). 3) D. Firestone, J. R:ass, N. L. Brown, R. P. Barron, and J. N. Damico, J.A.O.A.C., �, 85-91 (1972). 4) s. Jensen and L. Renberg, �, !,, 1-4 (1972). 5) B. A. Woolson, R. F. Thomas, and P. D. J. Ensor, J. Agr. Food Chem.,�, 351-354 (1972). 6) 7) in press. R. Baughman, unpublished results. :a. Baughman and M. Mesel.son, Advances in Chemistry, MEMORANDUM ON TCDD PRODUCTION BY PYROLYSIS OP SODIUM 2,4,S-'l' In view of the possible importance of 'l'CDD production from 2,4,5-T and related compounds by pyrolysis, we wish to communicate the following results. Pyrolysis of 15 mg Na salt of 2,4,5-T in an open tube (5 mm i.d. x 150 mm, pyrex). Heating was at temperatures form 300 to 450 ° c by means of an oil or metal bath applied to the bottom of the tube, times ranging from 30 minutes to 12 hours. The tubes were rinsed . with benzene which was then extracted with 2N aqueous NaOH� The organic phase·,··was analysed by glc and also by mass spectrometry. Yields of 'l'CDD ranged from 0.1 to 0.3 percent (1,000 to 3,000 ppm). Amass spectrum of the pyrolysate of Na 2,4,S-T-is attached. 'ftlese results have previously been desribed at the NIEHS Conference on Dibenzofurans and Dibenzodioxins held at Research T�iangle Park, Borth carolina on 2-3 April 1973 and at a meeting on the pyrolysis problem at the EPA on 23 July 1973. ll.W. Baughlllan M. Meselson Department of Chemistry y Department of Biochemistr and Molecular . Biology Harvard University Cambridge, Massachusetts 02138 30 July 1973 . I' '' PRODUCTS OF 'l'HE PYROf,YSIS 0.F 'l'IIE NA S6L'l' OF 2 111.,j;:! i- .. ,·.,.·· \.· ;� Tetrachlorodibenzo-,a-d,ioxin Pentachlorodibenzoturan Tetraohlorodibenzoturan I 300 I 310 Pentaohlorodibenzo-.t-4ioxin I..V'""-�, 320 . 330 m/e �JJL,JJ�� 340 350 360 Mass spectrum in the region m'e 300-360 ot the benzene-soluble residue trom the pyrolysis (450° . 30 min) ot 15·mg ot the Na salt ot 2, ,5-trichlorophenox;yacetic acid. A large trichlorophenol peak (m/e 196) and a weak dichlorophenol peak (m/e 162) were also observed. AEI MS-9 spectrometer, source 200° , ionizing voltage 70 eV, accelerating voltage 8 kV. By glc the yield ot tetrachlorodibenzo-E,-