Document O3d3bdop5YrkvJaw6LxGowQEQ

CHLORINATED DIBENZOFURANS AND DIBENZO-P-DIOXINS: DETECTION AND QUANTITATION IN ELECTRICAL EQUIPMENT AND THEIR FORMATION DURING THE INCINERATION OF PCBs by Wellington Science Associates Inc. B.G. CHITTIM B.S. CLEGG S.H. SAFE 0. HUTZINGER * 1 ^TKrrr Prepared under contract no: 0SS78-00067 for Fisheries and Environment Canada September, 1979 783369 Fn r-f .i: i :a" c'll EPfnGviiteoc' ion rr rj . ,,'ll n .r o :*ix oe October 19, 1979 ?5 St. Cl ai r Av*. E. Toronto, Ontario M4T 1M2 .OCT 2 _ )ST9' Ou-'t 4405-P75-1 Mr. R.D. Maguire Manager, Engineering Distribution and Specialty Transformer Section Canadian General Electric Co. 940 Lansdowne Avenue Toronto, Ontario Ltd. Dear Mr. Maguire: I, s 7 / ; ... /. a .S _ CC^T- Oni- p / - - (T ^ C *y C ~)f * r f - f t t - i ^ tT 'l'u ^ b r c A ^(f^7c Enclosed is a copy of the final report dealing with the formation of chlorinated dibenzofurans and chlorinated dibenzoparadioxins in operating PCB containing equipment and during low temperature incineration of askarel. The report is being distributed to those who took part in the study as well as to selected individuals for review and comment. We believe the project has been most successful and has resulted in a comprehen sive scientific document. It is the intention of EPS to have this report published for wider distribution after all comments have been received. Should you wish to provide any criticism or suggestions about the contertff^r distribution of the report I would appreciate receiving your n \ewSj D.J. Pascoe A/Manager Environmental Contaminants EPS - Ontario Region DJP:gj GENP 010559 cro-t --y. * it- ;ir u*i.*41it- trl.* .--CO 783370 RR.No 1. Rockwood. Ontario woe 2KO il ^U/eCCinglonScience^Associates*9nc, ENVIRONMENTAL CONSULTANTS October, 1979 Mr. D. Pascoe, Environmental Protection Service, Department of Fisheries and Environment, 135 St. Clair Ave. West, 2nd Floor, Toronto, Ontario Dear Mr. Pascoe, Please find enclosed the final draft of the report entitled ''Chlorinated Dibenzofurans and Dibenzo--dioxins: Detection and Quantitation in Electrical Equipment and Their Formation During the Incineration of PCBs" prepared for the Department of Fisheries and the Environment under contract number: 0SS78-00067. We hope you find it satisfactory. Yours very truly BGC/BSC/dac Enclosure B.G. Chittim B.S. Clegg GENP 01056C 783371 i ABSTRACT In this research, the formation of chlorinated dibenzo--dioxins (PCDDs) and chlorinated dibenzofurans (PCDFs) in PCB based fluids was investigated. The amount of PCDFs in transformer askarels was found to depend upon the length of time the transformer (and askarel) had been in service. PCDFs were also found to form when PCB based fluids were exposed to high temperatures (ca 500*C) with oxygen present. No significant amounts of PCDDs were detected in either atudy. 783372 t a ^ a t a ik inr r\ TABLE OF CONTENTS I PAGE Abstract Letter of Transmittal Table of Contents List of Tables List of Figures Conclusions Recommendations Introduction i ii iii V vi vii viii ix 1. POLYCHLORINATED DIBENZO--DIOXINS AND DIBENZOFURANS: THEIR CHEMISTRY, BIOLOGICAL PROPERTIES AND FORMATION 1.1 Chemical and Physical Properties 1.1.1 Synthesis 1.1.2 General Properties 1 1.1.3 Spectral Properties 1.1.4 Gas Chromatographic Properties jI 1.2 Biological and Toxicological Properties of Chlorinated Dlbenzofurans (PCDFs) and Chlorinated Dibenzo--dioxins (PCDDs) I 1.2.1 Chlorinated Dlbenzofurans(PCDFs) \ 1.2*1.1 Metabolism and Biochemical Toxicology 1.2.1.2 Toxicology j 1.2.2 Chlorinated Dlbenzo--dloxlns (PCDDs) 1.2.2.1 Metabolism and Biochemical Toxicology 1.2.2.2 Toxicology I 1.3 FCDD's and PCDF's in FCB-based Electrical Fluids and Other Commercial Compounds <j 1.3.1 PCDDs and PCDFs In Chlorophenols and Their Derivatives 1.3.2 PCDDs snd PCDFs in 2 (4,5-T Preparations 1.3.3 PCDDs and PCDFs in Fly Ash and Flue Gas 1.3.4 PCDFs in FCB-based Fluids j 1.3.5 Formation of PCDFs from the Pyrolysis of PCBs 1 1 1 3 4 10 11' 11 11 12 13 14 15 18 18 19 20 20 21 2 . CHLORINATED DIBENZOFURANS (PCDFs) AND CHLORINATED DIBENZO-p-DIOXINS (PCDDs) IN PCB-BASED FLUIDS USED IN ELECTRICAL EQUIPMENT 24 2.1 Introduction j 24 2.2 Experimental , 24 22.1 Sampling of PCBs In Electrical Equipment 24 2.2.1.1 Introduction 24 2.2.1.2 Transformer Askarel*Sampling 25 2.2.1.2.1 Sampling Criteria | 25 2.2.1.2.2 Transformer Askarel Sample\Groups 26 2.2.1.2.3 Collection of Askarel Samples j 27 2.2.1.2.4 Determination of Chlorobensene/Polychlorlnated Biphenyl (PCBz/PCB) Ratio 1 j 27 2.2.2 Analysis of Askarel Samples for PCDFs and PCDCs 2.2.2.1 Introduction | 28 2S 2.2.2.2 Isolation of PCDFs and PCDDs from Askarels 29 2.2.2.3 Analysis of PCDF/PCDD Fraction1uslng| Gas Chromatograph/Mass Spectrometry (CC/MS) 2.2.2.4 Analysis of PCDF/PCDD Fraction for pjCDF * * * ` 1 Total FCDF by Rerchlorlnatlon 31 31 31 I iv 2.3 Results 2.3.1 Sampling of Transformer Grade Askareis 2.3.2 Analysis of Askarel Samples for PCDF land PCDD 2.3,2.1 Isolation of PCDF/PCDD Fraction j I 2.3.2.2 GC/MS Analyses of the PCDF/PCDD Fraction 2.3.2.3 Perchlorination of the PCDF/PCDD Fraction to Determine Total Dlbenzofuran 2.3.2.A Quantitation of 2,3,7,8-Tetrachlrodibnzofuran (TCDF) in Askarel 2. A Discussion 3. INCINERATION/TRAPPING EXPERIMENTS USING AROCLORS AND ASKARELS 3.1 Introduction' 3.2 Experimental 3.2.1 PCBs and Askareis Used for Pyrolysis 3.2.2 Pyrolysis of PCBs and Askareis 3.2.2.1 Apparatus 3.2.2.2 Procedure 3.2.3 Analysis of Pyrolysis Products 3.2.3.1 Isolation of PCDF/PCDD Fraction 3.3 Results 3.A Discussion REFERENCES APPENDICES PAGE 33 33 37 37 37 AO AO A6 A8 A8 48 A8 48 A8 A9 A9 A9 52 52 55 59 rQCOIOdNStO 783374 V LIST OF TABLES TABLE PAGE 1 Number of Positional PCDD and PCDF Isomers 1 2 Melting points of some PCDD and PCDF Isomers 4 3 U.V, Spectral Data for some PCDD and PCDF Isomers 6 4 NMR Data for Representative PCDD and PCDF Isomers 8 5 PCDFfs In Aroclor, Clophen and Phenoclor (ppm) 21 6 Transformer Data, Askarel Composition and TCDF Concentration 34 7 Pyrolysis of PCBs and Askarels 51 GENP 010564 783375 vi LIST OF FIGURES FIGURE PAGE 1 Chlorinated Dibenznfurans and Dibenzo--djioxins 1 2 Infrared Spectrum of 2,3,7,8~Tetrachlorodibenzo--dioxin 5 3 El Mass Spectrum of 2,3,7,8-Tetrachlorodibenzo-L-dioxin II 4 El Mass Spectrum of 2,3,7,8-Tetrachlorodibenzofuran 6 7 5 Pyrolysis of 2 t4 t5,2f,4*,5'-Hexachlorobiphenyl 6 Isolation of PCDF/PCDD Fraction from Askarel ! 7 GLC Analysis of Askarel for Chlorobenzenes and PCS 23 30 38 8 GC/MS Analysis of PCDF/PCDD Fraction from Aroclor 1254 39 ir 9 GC/MS SIM Analysis of PCDF/PCDD Fraction from Old Chlorextol 41 10 Ferchlorinatlon Products GLC/ECD 11 GLC/ECD of PCDF/PCDD Fractions 12 Inclneratlon/Trapplng Apparatus for Pyrolysis of PCBs 42 44 50 13 Individual Ion Scans (SIM GC/MS) for Tetrachloro- and Hexachlorodlbenzofurans in PCDF/PCDD Fraction of Pyrolysed Pyranol 53 GENP 010565 783376 vii CONCLUSIONS 1* Significant levels of 2,3,7,8-tetrachlorodilienzofuran (TCDF) (l.e. ranging from ca. 0.005 to 5.0 ppm) were founc in used transformer grade askarels. 2. The amount of chlorinated dioxins (FCDDs), :.f detected, was usually less than 5X of the TCDF concentration. It appears from these initial results that the FCDD levels in FCB Jased fluids are insignificant relative to even individual FCDF concentration. 3. Small amounts, of terphenyls, quaterphenyls and chlorinated terphenyls were detected in some of the askarel samples but they were not quantified. 4. The amounts of TCDF in transformer askarels were found to increase as the time since the transformer was Installed Increased. Time-in- service was the only variable that could be positively attributed to the formation of FCDFs. 5. Although only a limited number of samples were available, the affects of discharging or arcing in a transformer bn the levels of TCDF -sj o0o3 in the 'contained* askarel appears to be negligible. These periods of abnormal operation are moat likely too short to cause the necessary reactions. 6. Laboratory scale pyrolysis of the FCB-jbasfelduids at 500aC resulted in the formation of large amounts of; FCDFs. No PCDDs were formed 7 In this survey the askarel samples ver|e only analyzed for TCDF since it has been shown to be the most toxic (extremely) of all of the FCDF congeners. It should also be pointed out that the analyses were carried out using 'packed* chromatographic columns and It is therefore possible that the peak corresponding to TCDF may be due to more than one GENP 010566 vili RECOMMENDATIONS I. During the sampling portion of this project it was noted that some-units were poorly maintained (eg. corroded, lea icing) and spill preventative measures had not been taken. It is sug gested that a nationil vide inspection of all PCB-filled transformers! be initiated. The purpose i | of this survey would be to locate units that are in need of maintenance i| or spill prevention equipment, and to Instruct the owners of such units on what measures should be taken. 2. Due to the presence of FCDFs in transformer oils, particularly i in the used askarels, the 'total' destruction of FCB based fluids that i are no longer in use is even more imperative. |In this respect, I! |I method to be used for this destruction should .be tested for: the i) its potential to cause the formation of FCDFs from FCBs 11) its destruction efficiency with respect to PCDFs. 3. Recently it was found that FCDFs and, to a lesser extent, PCDDs, 32 were formed by pyrolysis of chlorobenzenes in sealed quartz ampoules. Since askarels are a mixture of chlorobenzenes and F Bs and it has been indicated that chlorobenzenes may be used for retrofilllng askarel filled transformers, the formation of FCDFs from chlorobenzenes and the con tamination of chlorobenzenes by FCDFs should bje investigated. i| 4. Due to recent technical advances it la highly recommended that capillary gas chromatograph methods be developed for the analysis of environmental contaminants. Capillary chromatography would be especially II ' Iapplicable to analysis for PCBs, PCDFs and FCDDs where individual eongeners have been shown to have distinct toxicological properties. CjNP 10567 783378  IX INTRODUCTION The chlorinated dibenzo--dioxins (PCDDs) and dlbenzofurans (PCDFs) have been detected as contaminants In a number of commercial products such as pentachlorophenol and FCBs. A number of Individual PCDD and PCDF congeners have been found to be some of the most toxic compounds known to man* These compounds have also been detected as contaminants *ln various environmental samples. In this report: 1) The chemical, physical and spectral properties of the known PCDD and PCDF Isomers are briefly discussed. 2) The biological and toxicological properties of these two groups of chemicals are summarized and compared. 3) The literature concerning the.prescence and formation of PCDF and PCDD In various commercial products, with emphasis on PCB-based fluids, reviewed and discussed. 4) Samples of askerals were taken from transformers which were operated under varying conditions in such a manner so that an operation or construction variable eould be isolated. These samples were then analyzed for PCDDs and PCDFs, i 5) Samples of PCBs and askerals were subjected to high temperatures (such as those which might be expected during a transformer fire) and the formation of PCDDs and PCDFs investigated. The initial terms of reference for this project are given in Appendix B. GENP 010568 783379 1F 1 POLYCHLORINATED DIBENZO--DIOXINS AND DIBEN20FURANS: THEIR CHEMISTRY|, BIOLOGICAL PROPERTIES AND FORMATION The chlorinated dibenzo-p-dioxlns (PCDDs) and dlbenzofurans (PCDFs) as shown in Figure I, have been found as highly stable contaminants in some widely used industrial chemicals. Their detection as environmental pollutants poses a threat to human health due to their lipophilic and extraordinary toxic nature. 91 /"a a t a t k Trrr\ Figure 1: Chlorinated Dlbenzofurans and Dibenzo-p-dioxins 1.1 CHEMICAL AND PHYSICAL PROPERTIES 1.1.1. Synthesis A total of 75 different PCDD Isomers and 135 different PCDF isomers 'and congeners are possible (see Table 1); since each of these compounds have TABLE 1: Number of positional PCDD and PCDF Isomers Chlorine substitution mono dl - tri tetra - penta hexa hepte octa - Number of chlorines in each carbon ring X1 10 20 11 30 21 A0 31 22 A1 31 A2 33 A3 AA PCDDs Number of 1 somers m PCDFs Number of In each sub-group total in each sul 2 *A 6 2 12 1 8 13 2 12 A 6 2 1 1 ;2 LO , 1A 22 1A 10 2 1 75 A 6 10 A 2A 1 16 21 A -0v3l C0O3 CO 2A O 6 10 A 1 1. I different toxicological properties, their Individual synthesis and ldenclfication in.mixtures is extremely important. Only a small number of papers have been I 123 45 published dealing vith the synthesis of individual FCDD * * and FCDF ' compounds. The most common preparative routes for the PCDDs have been c summarized by Buser. They ire: - the dimerization of chlorophenates <i>7 . C13I0 a T cr^ci (1) 8 - the cydlzation of chlorinated 2-p lenoxyphenols (2) , AT a @ @ c the condensation of catechols vith chlorobenzenes' or chloronitrobenzenes (3)1*2, o* o* C I ^ ^ C I at Cl 505 .. AT - and the direct chlorination of dibenzo-p~dioxln (4) 0 Of j 6 c! -- I -Nl 00 CO CO 00 ci ci Cl A/ TATA J K T ^ A v*i m The major routes possible for the preparation of tie various PCDF isomers have been investigated by Grey and his coworkers. They are: the direct chlorination of dlbenzofuran(5), t o g ) ^ a ' t o ^ C U a'3 t o 9 the diazotizatlon and cydization of substituted o-phenoxy anilines (6), Cl Cl Cl Cl * i & p g r c 1, and, cycllzatlon of substituted o,o'-blphenyols or their derivatives (7). Cl * % Cl Cl Cl Cl Cl (7) More recently, the synthesis of PCDFa has been achieved by the palladium a in chlorinated diphenyl ethers. 1.1.2 General Properties The individual PCDD and PCDF congeners are colourless solids at room temperature. The melting points of a few representative congeners are given in Table 2. Like the PCBs, the PCDDs and PCDFs are extremely Insoluble in water and are only sparingly soluble in most organic solvents. At standard temperature and pressure they have no appreciable vapour pressure. 783382 i GENP 010571 * \ Table 2: Melting Points of some PCDD anil PCDF Isomers Dlbenzo-j>-dioxin cM.P. Dlbenzafuran cM.P. unsubstituted 1-chloro2-Chloro2,7-Dichloro2,8-Dichloro2,3-Dichloro1,2,4-Trichloro2,3,7-Trichloro1,2,3,4-Tetrachloro2,3,7,8-Tetrachloro1,3,6,8-Tetrachloro1,2,3,4,7-Pentachloro1,2,3,7,8-Pentachloro1,2,4,7,8-Pentachloro1,2,3,4,7,8-Hexachloro1,2,4,6,7,9-HexachloroOctachloro 122-123 104.5-105.5 88-89 209-210 150.5-151 163-164 128-129 162-163 189 320-325 219-220 195-196 240-241 205-206 275 238-240 330-331 unsubstituted 2-Chl!oro- I 2,3-Dichloro- 2.3.8- TricKloro- 2,4,6-Tricriloro- 2,3,7 *8-Tetra'chloro- 2.3.6.8- Tetrachloro- 2.4.6.8-Tetrachloro- 1.2.4.7.8- Pentachloro- Octachloro 86-87 101.5 125.5-127 189-191 116-117 227-228 202-203 198-200 234-235 256-257 1.3 Spectral Properties The INFRARED SPECTRA (IR) o f the PCDDs and the PCDFs show a characteris1 14 8 tleally strong band due to the stretching of the C-O-C bond * * The location of this absorption band varies vlth the position and numaer of the chlorine substituents^, The PCDD and PCDFs also exhibit bands due to the bending of the aromatic C-H bonds. Mien the hydrogens involved are isolated (the absence of any ortho substituted H-H interactions) the band occurs between 870 and 880 csT \ When these C-H bonds are adjacent the band occurs between 850 and 860 ca'1 . The IR spectrum of 2,3,7,8-tetrachlorodibenzo--dioxin is shown in Figure iJ 2 with the main bands characterised. Many of the jPCDD and PCDF isomers can be distinguished from each other using their IR spectra. j The ULTRAVIOLET SPECTRAL (UV) data for a series of chlorinated dibenzo'-p* dioxins and dlbenzofurans is summarized in Table 3. The UV spectrum of dioxin 7QQ QOQ 7/cmn j w r o * Figure 2. (| Infrared spectrum of 2,3,7,8-tetrachlorodlbenzo^p-dloxin. Aromatic CH stretch: 3122 (uw), 3080(wm), 3028(w). skeletal in-plane stretch: 1569(8), 1493(s sh), 1473(vs), 1464(vs), 1457(s sh),I 1394(b ) . -COC- asymmetric stretch: 1327(s), 1212(s), CCC trigonal aromatic ring in-plane deformation 1173(s). Aromatic ring breathing 115(s). CH out-ofr-plane deformation 876(s sh), 870(vs). Aromatic CC1 stretch 789(m) In chloroform exhibits two at 248 and 293; that of !dlbenzofuran is such more complex (see Table 3). The **** of both compounds are shifted to |i higher wavelengths when chlorines are added to the aromatic ring. The electron-impact (El) HASS SPECTRA of the ?C3Ds and PCDFs both exhibit Intense molecular ions (M+ ) with the expected ion clustering due to the chlorine isotopes. Doubly-charged molecular ions (M^+ ) of some intensity are also ob- served.^ PCDDs mainly fragment to form M+ -C0C1 and M+ -2C0C1 lens 6,11t si- though minor fragmentation occurs by loss of Cl v Cl^ and sometimes HC1 from the molecular and fragment ions. A typical PCDD mass spectrum (El), that of 2 t3 a7 l8-tetrachlorodlbenzo--dloxint is shown in Figure 3. 783384 CiHNF U10i>73 $'CT: 6 Table 3: UV Spectral Data for some PCDD and PCDF Isomers Dibenzo--dioxin UV S jectral Data; CHCl^ Xmax| (c) (nm) Unsubstituted 1- chloro- 2- chloro- 2.7-dichloro- 2.8- dlchloro- 2,3-dichloro- 1.2.4- trichloro- 1.2.3.4- tetrachloro- 2.3.7.8- tetrachloro- 1.3.6.8- tetrachloro- 1,2,3,4,7-pentachloro- 1.2.3.4.7.8- hexachloro 1.2.4.6.7.9-hexachloro octachloro Dlbenzo furan Unsubstituted 2.3.8- trichloro2.3.7.8- tetrachloro2.3.6.8- tetrachloro- 2.4.6.8- tetrachloro- 1.2.4.7.8- pentachloro1.3.4.7.8- pentachloro- 2^8(1020), 293(3680) 248(1320), 294(3190) 248(1140), 299(3700) 2f7(1340), 302(4590) 2f7(1180), 299(4690) 247(1830), 304(3190) 253(5290), 294(2250) 257(6290), 317(2290) 248(2970), 310(5590) 250(5540), 305(3440) 259(5920), 306(2690) 259(5370), 316(3660) 259(4450), 310(1480) 261(13150) ,318(2400) 245(10000) ,250(20000),280(16000), 285(16000) ,295(10000),300(16000) 256(33000) ,302(29000),313(21000) 259(15000) ,306(15000),316(14000) 250(43000) ,260(52500),285(27400), 302(32000) ,34(14100) 257(17000) ,294(18600),310(5800), 323(5800) 256(27000) ,266(41000),297(48000) 263(27000) ,272(40000),297(48000), 320(18000) -0j0 CO CO CcOn GENP 010574 100,% a * CaHja* \s M * - 2 COCI -COC id uXibLUkJL^l 00 100 150 200 -a 250 300 Figure 3 El Mess Spectrum of 2,3,7,8-Tetrachorbdibenzo-p-dioxia $ The PCOFs also show a typical fragmentation with M+ -C0C1 (M+ -63) and M+-C0C1-C12 (M+-133) Ions being formed. As with the ECDDs there is some minor fragmentation due to loss of Cl, Cl^ and HC1. The El mass spectrum (MS) of 2,3,7,8-tetrachlorodibenzofuran is Ghown In Figure A. 100% M+-COO-CI. a ,j.i Jl lTItl| I SO 100 M* I .L 150 M+-coa L M+-0, k -,iim M*-a 200 250 300 Figure A . El Mass Spectrum of 2,3,7,8-Tetrachlorodibenzofuran "1 m/e In most cases the molecular ions of the PCDDs and PCDFs easily distinguish them from other chlorinated aromatic hydrocarbons such as PCBs, There are, however, some compounds which potentially could lnterfer In MS analysis. For example, the Identification of PCDFs may be complicated by the presence of chlorinated dlphenylethers. These compounds yield Intense ions (M^-Clj) with the same exact mass and chlorine number ae a corresponding PCDF. In this case, and In most others, such Interfering compounds have different chromatographic 12 properties and thus can be removed. NUCLEAR MAGNETIC RESONANCE (NMR) spectra, like IR, can be used to distinguish between various PCDD and PCDF Isomers. NHR data for Lome Jf these compounds, synthesized by unambiguous methbds, has previously been reported1.2, A and Is summarized In Table A. 783386 V J JL-fJL U VJ X \J  Table 4: KR Data for Representative PCDD and PCDF Isomers Pibenro-j-dioxin Peak Pattern or Tpe and Numbej of Protons Actual Prtons Involved0 and Chemical Shift (ppm) Unsubstituted 1- chloro- 2- chloro- 2.7- dichloro2.8-dichloro2,3-dichloro- 1,2 ,4-trichloro- 2.3.7- trichloro- 2.3.7.8- tetrachloro1.3.6.8- tetrachloro1,2,3,4-tetrachloro1.2.3.4.7- pentachloro1.2.3.7.8-pentachloro- 1.2.4.7.8- pentachloro- 1.2.3.4.7.8- 1.2.3.6.7.9- hexaehloro- hexachloro- Dibenzofuran Unsubstitutad 2,3,8-trlchloro- 2,3,9-triehloro2,4,6-trichloro- 2,3,6,8-tetraehloro- AA'BB'^ ABCD,4 ABC, 3 A A ' B B 1 ,4 ABC, 3 ABC,6 ABC, 6 AA' BB',4 e,2 m,4 8,1 8*,23 6,4 AB,4 m,4 ABC, 3 s ,1 8,1 88,,21 8,1 88,,22 a, 6 a,2 d,2 8,1 t,l 8,1 a, 3 8,1 8,1 B, 2 d.l d,l B,1 d.l d.l 8,1 8,1 6.81 centred at 6.85 6.78 6.81 6.82 6.80 6.96 Hi and H* 9 7.02 Hc-9 > 6.88 Hj > 7.00 h,h b,h, 6.85 B i and H<, > 6.97 H 6.97 6.90 and 7 .02 H*-9 H,H,Hg H e> 6.96 6.96 6.98 H H,, >* 7.02 7.13 H 1 B B* and and and Hg H B Hg * 7.16 7.19 6.96 7.18 a.%. B- 8 7.03- B] 8 and and H H7 H* H 9>19e 7.82 7.48 7.68 7.84 Hi 9 7.95 I7 and B h* 91 7.257.85 Bi l? and Hg 899 8.5 7.36 As Hi B H7 H R* Hi 9991e99889 7.50 7.78 7.79 7.53 7.77 7.78 7.99 continued. 00 GO CO 00 GENP 010576 9 Table 4 (continued) Dlbenzofuran Peak Pattern or Type and Number of Protons acutal Prgtons Involved0 and Che mlcal Shift (ppm) 2,4,6,8- tetrachloro2,3,7,8-te trachloro- 1 *3,4,7,8-pentachloro- d,2 Hj + H 7 7.52 d,2 * H x + H 9 7.74 s ,2 H i, + He 7.72 3,2 Hi + H, 7.99 3,1 h 2 7.46 8,1 He 7.78 S ,1 H 9 8.30 a s - singlet, d - doublet, t - trlplist, m -- multiplet b where assigned In the literature (see ref 1, 2 and 4) c relative to IMS GENP 010577 783388 m 10 1.1,4 Gas Chromatographic Properties I | Analysis of contaminant residues in environmental samples usually relies on various Isolation techniques (le. solvent extraction, column partitioning) and ultimate quantitation using gas-liquid chromatography (GLC). GLC coupled i with mass spectrometry (MS) has become the primary technique used in structure confirmation. i | 2 12 13 1A GLQ using open-tubular packed columns, separated the FCDDs > * > and the PCDFs 4,12,13,14,15 according to chlorine content but shoved little selectivity towards individual Isomers. In most cases the retention times Increased with ii increasing chlorine content. In the case of the PCDFs, ciilorlne substitution !i ortho to the ring oxygen, that is in the 4 and 6 positions, tends to decrease 14 ' the retention time of the individual isomers. A similar effect has been seen with individual PCDD isomers on glass capillary columns. Generally the reten- tlon times of the PCDDs and PCDFs were in the same range as the PCBs and most `15 of the common chlorinated pesticides. ! j 3 co CO 00 CO Glass capillary columns have been found to successfully resolve Individual PCDD and PCDF isomers^ (eg. 8 of 10 tetrachloro PCDD isomers). As with the open- tubular packed columns, the retention times generally' Increased with increasing chlorine content. :0n the more polar columns (le. Slla'r IOC) , the polarity of the isomers had a pronounced influence leading to a reversal of this elution order. It haa. been suggested that capillary columns are more applicable to PCDD and PCDF analysis because of their ability to separate Isomers. Since only a few of the isomers are of any great Interest due to their extreme toxicity, and I| not all of the possible isomers have been found as contaminants, the packed a o V-- 'L o u: 5 column GLC approach la still widely used. Recent developments In high resolution. | packings (le. the ultrabonded packings), higher detection 1:imite for capillary i columns, and time and cost factors also support the continued use of packed ii | 11 \ 1.2 BIOLOGICAL AND TOXICOLOGICAL PROPERTIES OF CHLORINATED DIBENZOFURANS (PCDFs) AND CHLORINATED DIBENZO--DIOXINS (PCDDs) This discovery that PCDFs and PCDDs existed as contaminants In commercial th icompounds, such as PCBs and PCPs, was due primarily to extreme toxicity of some of their Individual congeners. For example, previous research has shown j| that certain fractions of commercial FCBs were far more toxic to chicks and 38 1 I that these fractions contained PCDFs. In a similar manner, technical penta- chlorophenol, containing PCDDs, was found to give positive results for most toxicological tests whereas the changes effected by 'pure 1 pentachlorophenol 48 were limited to increased liver and kidney weights. A brief discussion of the biological and toxicological properties of PCDFs and PCDDs is given in the following section. 1.2.1 Chlorinated Dibenzofurans (PCDFs) Host biological studies using PCDFs have investigated only the acute toxicity of these compounds. Few research groups have concerned themselves with the actual mechanisms of toxicity or with metabolism! of PCDFs. 1.2.1.1 Metabolism and Biochemical Toxicology M-e--tabolism 0'-0-j CO To our knowledge, only one study has actually identified a FCDF metabolite, eg !o 49 I Zitko and his coworkers ware able to isolate a hydroxylated dichlorodibenzo- furan from fish which were fed 2,8-dichlorodibenzofuran, The actual structure of this metabolite and its mechanism of formation could not be determined. r M O ilMSO Zn what may be considered 'metabolism-related* research, the tissue clearance of a mixture of PCDFs in mice has.been investigated,'50 In this 1 Istudy, it was observed that the PCDFs were primarily concentrated in the liver and fat tissue with a surprisingly high level in the spleen. Gas chromato graphic (GLC) patterns of the PCDFs in the livers indicated that certain Isomers (usually of lower chlorine concent) were not retained as long as others. The authors suggested that this may be due to preferential metabolism 12 and/or excretion of cerealn isomers. The biological half life of the PCDFs used in mice was estimated to be about one week. Enzyme Induction and other Biochemical Effects PCDFs are potent Inducers of enzyme systems, partlcularlly in the liver In this respect TCDF has been found to be JquipoJent with TCDD, 3,4,3t,4*~ tetrachloroazoxybenzene and 3,4,31,4'-tetrachloroazobenzene in inducing aryl hydrocarbon hydroxylase (AHH). These four compounds compete equally with each other for stereospecific binding sites in the hepatic cytosol which are thought to be the receptor sites for the induction of this Jenzyme. Many other biochemical effects were reported phen rats were fed diets containing 1 or lOppm of PCDFs. These Included Increased serum cholesterol levels and cholinesterase activity and decreased triglyceride levels and leucine aminopeptlddse activity. PCDFs also decreased serum glutamic-pyruvic trans aminase activity and testosterone concentrations in the testes and increased serum glutamic oxaloacetic transaminase activity. 1.2.1.2 Toxicology Animals The toxicity of individual PCDFs is extremely dependent upon the degree and position of chlorination and seems to peak at the tetra and pentachlorodlbenzofurans. The most toxic congeners are the 2,3i,7,6-tetra-,l,2,3,7,8- 53 penta- and 2,3,4,7l8-pentachlorodlbenzofuran",<*,while it Is presently proposed 00 CO 00 CO that dlbenzofuran Itself or octachlorodlbenzofuran have little inherent toxicity on an acute basis. 54 In 1961 Bauer et al-'"* observed that a single oral dose of 0.5 mg/kg of a mixture of tri- and tetrachlorodibenzofurans given to rabbits caused severe and often fatal liver necrosis. The single oral Ld |q for TCDF for guinea pigs has been determined to be between 5 and 10ug/kg. The symptoms of toxicity GENP 010580 13 were severe welghc loss and atrophy of the thymus and spleen. Hemorrhages were observed in the adrenals, urinary bladder and single cell necrosis was 55 noted In the liver The oral LD^q for mice for a mixture of PCDFs was found to be approximately 200mg/kg for males, and 400mg/kg for females. In the mice that died, hepatomegaly and atrophy of the thymus were the more common mani I Ifestations; subcutaneous edema and dermal alterations were also observed frequently.^ TCDF (6mg/kg) administered subcutaneously to mice produced ^ I^5 similar symptoms but no deaths. Rats fed diets containing 10 ppm of a FCDF mixture over 4 weeks developed chloracne type lesions and showed toxic effects similar to those in m i c e . ^ As yet, there have been no long term or subacute toxicity studies using FCDFs. The reproductive effects of these compounds also remain to be investl- gated. Human Human injestlon of only.PCDFs has not yet been reported. However, in 58 Japan, over 1000 people consumed rice oil contaminated with PC8"'W and developed l lnausea, lethargy, subcutaneous edema of the face, and chloracne. It was cub- aequently demonstrated that the rice oil lnjested was more than twice as toxic 59 * as it should be with the amount of FCB present. In addition It was later oo CO co ^ determined that this PCB contaminated rice oil contained large amounts of 60 FCDFs. Whether these symptoms of the poisoning are due solely to PCBs (l.e. and not in part to FCDFs) is as yet indeterminable. 1.2.2 Chlorinated Dlbenzo-p-dioxins (PCDDs) 2,3,7,8'-Tetrachlorodlbenzo--dloxln (TCDD) is one of the most toxic compounds known to man. PCDDs have received considerably^ more attention than the PCDFs, however, their toxicological properties can. In most cases, be considered to parallel one another. TRCftTfl z T K T T V 14 1.2.2.1 Metabolism and Biochemical Toxicology Metabolism ^C-TCDD fed to rats was found to be rather incompletely adsorbed from the gastrointestinal tract, with the majority ending up in the fesces. Once absorbed in the body, most of the TCDD accumulated in the liver and fat. The level in these tissues was nominally ten times higher than in other tissues. Traces of radioactivity were also detected in the urine and expired air of the rats within the first 10 days; whether this was metabolized or unaltered TCDD was not det,,ermi.ned.. 64,65 TCDD has been considered by most groups to be resistant to metabolism by mammals, in vivo and in vitro. R e c e n t l y , however, Tulp and Hutzinger^ have found that several PCDDs (although not including TCDD) were metabolized by rats to mono- and dihydroxy derivatives. In the case of the monochloro and unsubstituted dibenzo--dioxin, sulfur containing metabolites were also de tected. Primary hydroxylatlon was determined to take place at the 2-, 3-, 7-, or 8-posltlons in the molecule and it was suggested that an intermediate 2,3-epoxlde was involved. Since in TCDD these positions are blocked, this was felt to be a viable reason for this chemical's resistance to metabolism. Enzyme Induction and other Biochemical Effects Some of the PCDDs have been reported to be powerful inducers of the enzymes -aminolevulinic acid synthetase (ALAS) and aryl hydrocarbon hydroxylase (AHH) in the chick e m b r y o . ^ * ^ .The potency of the AHH induction was shown to strongly correlate with general toxic responses. ^ 68 69 For AHH and ALAS induction, the activity of the individual PCDD congeners has been found to be strongly dependent on the position of chlorine substitution; at least three halogens are required In lateral (2-,3-,7- and -- ^ 9*** m 4 * 4 A n t . i j . L . . 1 A . a . -------------------- 1 - J .. 1 783393 GENP 010582 15 The highest potency, shown by TCDD, for inducing AHH was found to parallel 52 other inducers of similar molecular structure (i.e. TCDF) . TCDD is also a 69 strong Inducer of a number of other enzymes including rat j. liver DT-diaphorase and glutathione S-transferase.^ In addition, TCDD has been reported to bind to rat hepatic microsomal 1A 71 macromolecules pretreated with phenobarltol (>40Z of added C-TCDD). Mlcrosomes prepared from nonlnduced rats bound only 7.0Z Df the added *^C-TCDD,* indicating that the binding of this compound is mediated by a mixed function oxidase system. Most biochemical changes noted in serum taken from rats fed TCDD (10mg/kg) were attributable to liver damage. For example, the activits of glutamic- oxaloacetic transaminase, glutamic-pyruvate transaminase, lactic dehydrogenase and hydroxybuturate dehydrogenase were elevated, as were the levels of bill- rubln, cholesterol and urea. In addition, the rats had decreased serum glucose, sodium and protein levels as well as decreased arylesterase and cholinesterase 72 activities. Mice fed various levels of TCDD were reported to have slgnlfl- 73 74 cantly decreased serum protein and Increased liver lipid levels. 1.2.2.2 Toxicology Animal The toxicity of individual PCDD congeners is strikingly dependent upon -si CD CD CD CO the position and number of chlorine substituents. For example, TCDD is re ported to be 1,000 to 10,000 times as toxic as l,2,3,8-teJrachlorodibenzo-- dioxin. The most toxic PCDDs appear to be the 2,3,7,8-tetjra-(TCDD), 1,2,3,7,8- penta-, 1,2,3,6,7,8-and l,2,3,7,8,9-hexachloro-*dloxlns.1 The acute toxicity of TCDD has been tested in several animals. In one study the single oral LD^q ranged from 0.0006 mg/kg in malL guinea pigs to 0.115 mg/kg in rabbits of mixed sex. For male rats the LD value was 50 GENP 010582 { 16 determined to be 0,022 mg/kg whereas female rats were less sensitive giving an U>50 value of 0,045 mg/kg.75 The average mean survival time for these rats was ea. 20 days post administration. For male mice i:he oral was determined to be 0.114 yg/kg with a mean survival t:Lme of, again, approximately 20 days. For female Rhesus monkeys it was given to be <0.070 mg/kg with death 76 occurring, on the average, 30 days after injection, The acute toxicological effects of TCDD common to mo'st of the animals studied, were weight loss (due to a decrease in consumption), increased liver to body weight ratios and decreased thymus to body weight ratios. In mice, TCDD was also found to be porphyrogenlc and caused subcutaneous edema, hepatic | 73 lesions and hemorrhaging in almost all of the organsj. Rhesus monkeys suffered loss of hair and fingernails, gastric ulcers, liver! necrosis, facial alopecia 76 and developed chloracne type eruptions as well as the common effects. TCDD was found to be porphyrogenlc and also created hepatic ana cardiac lesions in 77 rats. Klee fed TCDD on a subacute basis (1,5 and 25 ug/kg) developed similar symptoms to those in the acute experiment. The most severe lesions were observed in the liver and thymus with a 'no-effect1 level reported as a 0.2 73 ug/kg dose. - n| 000303 COOl In a 13-week oral toxicity study using rats fed TCDD at levels, of 0.001, 0.01, 0.1 or 1 ug/kg,5 days per week 77, doses of 1.0 ug/kg TCDD caused some mortality and toxicological symptoms again similar although milder) to those observed when an acute dose was administered. At the 0.1 ug/kg TCDD dosage level, somwhat milder toxicological symptoms were a|aln observed but no deaths. In rats given 0.01 or 0.001 ug/kg, TCDD there was essentially no effect. The results of a 2 year chronic toxicity study of TCDD in rats indicated mcotojnO m 17 that continuous doses of TCDD sufficient to Induce severe toxicity,increased the incidence of some types of tumors while reducing others. These rats | 73 were fed diets supplying 0.1, 0.01 and 0.001 ug of TCDD/kg/day. PCDDs have also been reported to be embryotoxic when administered preatally 79 80 to rats and mice. 1 Three major teratogenic are: - intestinal hemorrhages in rat fetuses * effects il * | i j have been noted; these - an Increased frequency of cleft palate in mouse fetuses l I - kidney abnormalitiesinmouse fetuses. i i Of the PCDDs, TCDD, as expected, is the most active with the hexachloro derivatives shoving some activity. The di- and octachloro Isomers were reI| ported to be nonembryotoxlc. Generally, it has been shown that embryotoxlc activity is also a property of those PCDDs that have a pronounced chloracne r potency. Human ! i Accidental releases- of PCDD into the environment have occurred in the || U.K., U.S.A., Germany, Holland and, most recently, In Seveso, Italy. All of these occurences have been well documented as have the toxicological effects / in exposed humans. I i Ihe most common initial symptom of dioxin poisoning in humans is severe C-vOj CO CO CO O) skin disease (chloracne) Other symptoms which have been observed and attributed to PCDD poisoning are liver cirrhosis, damage to the heart, kidney, spleen, central nervous system and pancreas, and emphysema.j As will as these physiological i| symptoms, psychological effects such as depression and disturbances of memory i '| and concentration have also been observed. Severaljdeaths from TCDD poisoning have been recorded, at least some as a result of liver damage and some many years after exposure. I GENP 010585 8 . !S 18 1.3 PCDDs AND PCDFs IN PCB-BASED ELECTRICAL FLUIDS AND OTHER COMMERCIAL COMPOUNDS 1,3.1 PCDDs and PCDFs In Chlorophenols and their Derivatives Chlorophenols, In particular pentachlorophenol (ICP), and their derivatives are used In large quantities as herbicides, sllmlcldes In paper mills and as wood preservatives. These compounds have been found to contain a variety of !contaminants including other chlorophenols, polychlorinated :phenoxyphenols CPPs), polychlorinated dlphenylethers (PCDPEs), polychlorinated biphenyls (FCBs) 18 19 !I and polychlorinated benzenes (pCBzs). * Of major interest and concern is that they also contain significant quantities of chlorinated dlbenzo--dloxlns and dibenzofurans. A survey of various commercial FCPs and their sodium salts by Buser and 20 Bosshardt revealed that levels of PCDDs and PCDFs varied from 1 to 2 ppm to as high as 200 to 300 ppm. In these samples the major components present were the hepta- and octachloro congeners. Levels of the tetra- and pentachioro congeners were generally less than 0.1 ppm in all the preparations tested. 12 21 Similar results were observed by other researchers * with even higher levels 22 23 I of PCDDs and PCDFs being reported ' Analysis of the lower chlorinated chlorophenols and their sodium salts shoved correspondingly higher levels of tetra-and pentachlorodibenzo--dioxlxis and dlbenzofurans and lover levels of 24 the hepta-and octachloro components, as would be anticipated CD CO CO -CvDl PCDD and PCDF formation in chlorophenol preparations has been studied ]i 3 23 24 I extensively * ' and these compounds*, or their precursors, are formed by thermal dimerization of the chlorophenates. Experimental pyrolysis of commercial chloro phenols and sodlumchlorophenates has been shown to give increased levels of PCDDs and PCDFs only with the c h l o r o p h e n a t e s . T h e formation of the PCDFs is believed to be due to cyclisation of an impurity In the commercial product <jJtJNP010586 i 19 rather Chan the chlorophenate Itself. The polychlorinated diphenyl ethers are presumably the most likely precursors of the PCDFs since they are known contaminants of the chlorophenols and their pyrolysis has been found to yield 9 dibenzofurans. It has also been determined that PCDDs are formed by a unlmolecular cycllzatlon of chlorinated phenoxyphenols (PCPPs). These "predioxins" have 27 28 been found In technical PCP preparations ' and individual PCPPs are known to 19 II 29 cycllze giving dioxins. Photochemical dimerization of chlorinated phenols 30 31 I and cycllzatlon of PCPPs * are also potential sources of PCDDs in chlorophenols 1.3.2 PCDDs and PCDFs in 2,4,5-T Preparations Herbicides derived from 2,4,5-trichlorophenoxy a d d s (2,4,5-T) have re celved considerable attention recently due to their contamination by dioxins and dibenzofurans. PCDDs and PCDFs have been identified in various 2,4,5-T ester f o r m u l a t i o n s ^ s u c h as 'Herbicide Orange', a defoliant used in Vietnam in the late 1960'a32 *34. Analyses carried out in the more recent of these surveys shoved that the phenoxy acids were primarily contaminated by PCDDs, with PCDFs identified in 32 I only a few'of the samples. These results differ somewhat from previous re- 33 search but this is most likely due to the samples chosen for testing. In burning experiments with 2,4,5-T esters, and also its sodium salt, no 00 CO CO CO 00 6 35 2,3,7,8-tetrachlorodibenzo--dioxin was formed. * Although other researchers have suggested that combustion of 2,4,5-T formulation may lead to formation of 33 I small amounts of PCDDs and PCDFs, this route la not an Important source of these contaminants as It was with the ehlorophenates. Slmllarlly the primary 1 photoreactions of chlorinated phenoxy a d d s are dacbllorlnatlon and hydroxyl sub stitution of the chlorines, vith.no formatlor of PCDDs or PCDFs occurring. .L 20 1.3,3 PCDDs and PCDFs In Fly Ash and Flue Gas ! There Is growing public concern over the presence of hazardous organic compounds in the emissions from municipal Incinerators, thermal generating 13 stations and various heating facilities. Olle, and his coworkers first reported the detection of PCDDs and PCDFs as trace components of the fly ash and flue i gases of three incinerators In the Netherlands. They also found that the major i chlorinated compounds In all the fly ash samples were highly chlorinated benzenes, i whereas In flue gas condensates, the most abundantl chlorinated hydrocarbons ii I were the chlorinated phenols. At this time they suggested that these phenols I were most likely the major precursors for the PCDDs and PCDFs detected. Similar surveys were performed In Switzerland using fly ash samples 36,37 collected at a municipal Incinerator and at an industrial heating facility. 1| Both PCDDs and PCDFs were found In these samples, jThe PCDD compounds detected were similar to those formed when a 2,4,6-tri-,3,,6-tetra- and pentachloro- | phenate mixture was pyrolyzed under controlled laboratory conditions. Most of the PCDF congeners detected In the fly ash were also identified in experimental pyrolyses of commercial PCBs. 37 1.3,4 PCDFs in PCB-Based Fluids 783399*i Toa, and his coworkers where the first to Identify PCDFs as toxic impurities In two samples of European commercial PCBs. (Clophen and Phenodor). The toxic effects of the PCBs were found to parallel the levels of the PCDFs present. 38 The Japanese PCB preparation, Kanechlor 400, which jhad been identified as the causative agent of Yusho (a human PCB poisoning incident), was also found to be 39 I contaminated with PCDFs* Bowes and his associates were able to identify PCDFs i in a number of American PCBs (see Table 5)^, in particular the 2,3,7,8-tetra15 1 chloro isomer. They also examined the two European PCBs previously analyzed 38 i I by Vos and found that these PCBs contained at least 10 times more PCDFs than RRcnrn ,mrr % 21 the corresponding Aroclors. Recently Morita et al*1 identified PCDFs in 16 7CB preparations of various origin and also in the Yusho oil (rice bran oil) responsible for the PCB poisoning. This oil was later analyzed by high resolution GC/MS and it vas found 42 that the main FCDF contaminant vas the 2,3,7,8-tetrachloro Isomer. 40 Table 5 PCDFs in Aroclor, Clophen and Phenodor (ppm) PCB Chlorodlbenzofurans Cl, Cl, Cl, TOTAL Aroclor 1248 (1969) 0.5 1.2 Aroclor 1254 (1969) 0.1 0.2 Aroclor 1254 (1970) 0.2 0.4 Aroclor 1260 (1959) 0.1 0.4 Aroclor 1260 (lot.AK3) 0.2 0.3 Aroclor 1016 (1972) ND ND Clophen A-60 1.4 5.0 Phenodor DP-6 0.7 10.0 0.3 1.4 0.9 0.5 0.3 ND 2.2 2.9 2.0 1.7 1.5 1.0 0.8 8.4 13.6 There has been some indication that PCBs also contain minor amounts of penta- and hexachloro dibenzo--dioxins (approximately 32 relative to the I 1corresponding PCDF). The levels of these additional contaminants-, if they are indeed present, are insignificant compared to the levels of the PCDFs. 1.3.5 Formation of PCDFs from the Pyrolysis of PCBs High temperature destruction of PCBs at present appears to be the most efficient and less costly method of disposing of these chemicals. The pyrolysis of PCBs has been tested using various laboratory-scale systems. Vapour injection of PCBs into high temperature air environments (1 second dwell time) resulted in greater than 952 destruction at 740*C, and 99.9952 at 1000'C forming only low molecular weight compounds (as yet unidentified) 43 Atomizing PCBs with oxygen using a burner and Igniting this mixture (giving a 'lame temperature gv 44 2000%C) resulted in >99.999992 destruction. On a larger scale, PCBs mixed with other halogenated hydrocarbons were 783400 TMqrr* b 22 AS burned in a rotary cement Kiln. The effective temperature vas reported as |i 2100*C or higher. In this case the PCBs were destroyed with at least, a 99.982 efficiency with no high molecular weight chlorinated hydrocarbons being emitted. It thus has generally been determined that at temperatures greater than 700*C pyrolysis of PCBs is virtually, complete giving only low molecular weight compounds. At temperatures lower than 700*C, with oxygen present, a significant amount of the PCBs are converted to PCDFs. |41 Morita and his associates found Increased amounts of PCDFs in Aroclor 1248 when it was sealed with air in a glass tube and heated at 300*C for two weeks. They also analyzed a Japanese FCB mixture that had been used in a heat exchange system for long periods of time and discovered chat it too con tained higher levels of PCDFs than in the unused oil. Pyrolysis of Aroclor 1254 at 550-650*C in quartz ampoules in the presence 37 46 of air by Buser et ai ' shoved the formation of up to 60 PCDF components ranging from the mono- to the pentachloro compounds at a combined level of about 1-32. In addition to PCDFs, polychlorinated hydroxyblphenyls were Identified as possible precursors. The pryolysis of Aroclor 1260 showed the formation of PCDFs at a similar level as in the case of Aroclor 1254. The com pounds formed ranged from the tri- to the heptachliro congeners. Buser and hla eovorkers also Investigated the pyrolysis of individual FCB isomers and found that they yielded PCDFs through th formal loss of o r t h o C ^ or HC1. For example, the pyrolysis of 2 t k t 5 t 2 ' ,4*,5'-hexachloroblphenyl yielded three PCDFs as shown in Figure 5. 783401 ViJbMP010590 23 02,AT route 1 route 2 11 1 1 1 i ! l-ro-u-t-e--3-- > i Figure 5: Pyrolysis of 2,4,5,2',4*(5*-Hexachlorobiphenyl The formation of FCDFs from the pyrolysis of PCBs may have environmental complications because a large portion of used PCBs may be disposed of by way of incineration. If the proper conditions are not used this may become a potential source of PCDF contamination. A recent fire in Toronto involving !I a PCB filled transformer Illustrates this risk. Thiis accident released PCBs II 47 and soot containing appreciable levels of PCDFs into the atmosphere. The thermal conversion of PCBs into PCDFs may also occur in askarel filled transformers while in service. Tils possibility ls| investigated as part of this project and will be discussed later. 783402 GENP'010591 I % 24 CHLORINATED DIBENZOFURANS (PCDFs) AND CHLORINATED DIBENZO-P-DIOXINS|(PCDDs) IN PCB-BASED FLUIDS USED IN ELECTRICAL E Q U I P M E N T _________ I I 2.1 INTRODUCTION i Under current federal regulations, PCBs can on|y be used In electrical and some mechanical equipment currently In service. Proposed PCB regulations vlll I i prohibit new PCB containing equipment being imported or manufactured for use in !| Canada. Due to these regulations, PCB filled transformers and capacitors that come out of service will be replaced by equipment using fluids that do not contain PCBs (e.g. silicone based fluids). ! At the present, there are many PCB or Aslcarel filled units in service i and it is anticipated that some of these will stilljbe in use 40 to 50 years 1i from now. During the interim period, PCBs vlll therefore still pose a threat to environmental quality unless properly handled and eventually destroyed. One of the never problems associated with these fluids is their possible contamination by PCDFs and PCDDs. i The purpose of this portion of the project is to investigate the formation !I of PCDFs and PCDDs in PCB baaed fluida that are being used in electrical equipI ment. The ultimate aim of this study is to ascertain those conditions which change the basic composition of these fluids, particularly with respect to the formation of these potentially toxic impurities. 2.2 EXPERIMENTAL 2.2.1 Sampling of PCBa in Electrical Equipment I I 1 783403 2.2.1.1 Introduction At the beginning of this project it vas intendail that both transformers and capacitors would be sampled. Capacitors, though1, were found to be inaDDronrlate for this atudv fearlutA of f h m * n4nr nroK1i.me P 01059: $% 26 \ generally Increased as the kva increases. The sampling was also limited to step-down, power transformers (see Appendix A ) .The available transformers were i thus sampled in groups as described below. { i i 2.2.1.2.2 Transformer Askarel Sample Groups (see Table 6) Group 1 ; Time In Service (1-1, 1-2, etc.) r |The first series of transformers sampled were those where the only variable was the length of time since they had been Installed. The trans formers sampled were all from the same manufacturer and it was therefore assumed i that they contained the same fluid (see Section 2.2.1.2.4). These transformers were rated at 1500 kva and had essentially identical kva/volume ratios. Group 2 : Variable kva/volume Ratio (2-1, 2-2, etc.) i Although as previously mentioned, the kva and fluid volume of a transformer are interdependant, there is some variation in the kva/volume ratio (see Table 6). For each of two fluids, Pyranol and Chlorextol, four 1| transformers were sampled where the only variable was this ratio. Group 3 i Variable Type of Fluid (3-1, 3-2', etc.) As discussed in the Appendix, there are three major askarels (chlorobenzene-PCB mixtures) currently in use in transiiformerIIs in Canada. Zn this sample group, two sets of three transformers were sampled. The only dif!1 ference between the two sets was the kva/volume ratio.1 Essentially, the only variable distinguishing the three transformers from each otlier in each set was the type of fluid they contained (l.e. Inerteen, Pyranol or Chlorextol). i\ Group 4 i (4-1, -2. etc.) j In this group askarels were sampled from transformers which were operated under what may be considered extreme conditions. Samples 4-1 !I OS CO Jos. GENP 010593 $ 27 and 4-2 were taken from two transformers which had been in service since 1947. Sample 4-1 was taken from a 56 kva transformer vitH a fluid volume of 660 Jiters. Sample 4-2 was taken from a 10,000 kva trinsformer with a fluid volume of 6900 liters. The difference in kva betwien 4-j and 4-2 Is quite large when compared to the change in volume. These two samples could there fore be considered to approximate the effect of vatjlable kva. I ISamples 4-3 and 4-4 were taken from transformers which had failed while in service; 4-3 from a transformer which failed and caused a fire (Adelaide St.)f and 4-4 from a transformer which overheated (Valkeriton). Also included in this group are samples of unused Pyranol, Inerteen and Chlorextol (samples 4-5, 4-6 and 4-7 respectively). 2.2.1.2.3 Collection of Askarel Samples Almost all of the askarel samples were taken from the top valve on the transformer. In a few cases the fluid levels were too low and the bottom valve had to be used for sampling. All the samples were collected in prevashed, 250 ml amber glass bottles fitted with teflon-lined screw caps. The askarel samples were stored at ca. 0*C. in the dark until analyzed. 2.2.1.2.4 Determination of Chlorobenzene/Polychlorinated Biphenyl (PCBs/PCB) Ratio CO CO o cn Transformer askarel fluids are mixtures of various chlorobenzenes and PCBs. The PCBx/PCB ratios given by the manufacturer for the three major fluids, Pyranol, Inerteen and Chlorextol (see Appendix A) are only approximate values and may vary from transformer to transformer, The exact PCBz and PCB content of each askarel sample was therefore determined by gas chromatography. i 1 1Since the composition of the Aroclors (1254 and 1260) may yary from batch to batch, the chlorobenzene content was determined and the remainder assumed to x 28 be PCB. The type of Aroclor used In the mixture was determined by comparing the chromatographic pattern of the PCB component to that of the pure Aroclors (1254 and 1260). The actual method used for this determination is as follows: Method : A known weight of each askarel sample was dissolved in petroleum ether to give a concentration |of 10.0 0.01 mg/ml. Standard solutions of each trl-, tetra- and pentachlorobenzene (2.0 0.01 mg/ml) were also prepared in petroleum ether. The PCBz content of each askarel sample was then determined by comparing duplicate 5 y GLC/FID injections of the PCBz solutions to those I I of the askarel solutions. ! GLC/FID was performed using a Hewlett Packard 5710A instru- ment equipped with a flame-ionization detector (FID). The column used was a 4 *1 x V' glass column packed with 32 OV-225 on Gas Chrom Q. The oven was operated isothermally at 125*C for 8 minutes and then temperature programmed to 220*C at 8* per minute. The carrier gas used was helium at 40'ml/mln. and the detector and injection port temperatures were set at 250*C. All chromatographic data was re corded on a H-P 3380S Integrator. j 2.2.2 Analysis of Askarel Samples for PCDFs and PCDDs 2.22.1 Introduction A review of the literature Indicated that PCDFs are the major contaminants of PCBs. If any PCDDs were present in previous analyses. their levels were con sidered relatively insignificant. In this experimental section of the project, the fraction of the askarels containing the PCDFs aniii PCDDsI was first isolated 1` using Florlsil and alumina. This fraction was then analyzed for PCDFs and PCDDs and other contaminants. 783406 vjjudNr V L V j y j $ 29 2.2.2.2 Isolation of PCDFs and PCDDs from Askarels Duplicate samples of each askarel collected in section 2.2.1 were fractionated on Florlsll and alumina columns as described by Roach and 39 ^ Pomerantz. The revised method used in this project is given as a flow chart In Figure 6 All solvents used were pesticide grade. In this procedure, 500 mg of the askarel was first dissolved in petroleum ether and applied to the top of a glass column (25 cm x 1 cm.l:d.) packed with Florisil(PR-60/100; Applied Science) which had been previously washed with i petroleum ether (50 ml). The column was then elutedj, first with 100 ml of I i petroleum ether, then with 50 ml of a 5Z ether/petrotleum ether mixture, and finally with 50 ml of a 25Z ether/petroleum ether mixture. The second eluate was evaporated just to dryness, and the residue remajlning vas then quantitatively transferred, using petroleum ether (ca. 2 ml), to this top of a second column (15 x 0.5 cm l.d.) packed with alumina (neutral, activity 1, 80/200 mesh) which had been prewashed with 2Z methylene chloride/ letroleum ether (15 ml). This column was then eluted first with 2Z methylene chloriLe/petroleum ether I I(15 ml) and then with 50Z methylene chloride/petroleum ether (15 ml). The second eluate from this column was evaporated just to dryness, reconstituted with petroleum ether (5 ml) and then again evaporated to dryness. For gas j |chromatographlc-mass spectrometric (GC/M5) analysis the residue was redis- solved in exactly 1 ml of toluene. For gas chromatographic analysis, using ii 1 an electron capture detector (GLC/ECD), the residue was reconstituted with exactly 1 ml of isooctane. ! "I The recovery of this isolation procedure was tested using A r o d o r 1254, iI previously purified on a Florlsll column, spiked witji 0.01, 0.001 and 0.0001 i mg of 2,3,7,8-tetrachlorodibenzofuran. ' GENP 010596 783407 fron Askarel 783408 GENP 010597 31 2.2.2.3 Analysis of TCDF/PCDD Fraction using Gas Chromatograph/Mass Spectrometry (GC/MS) I I GC/MS was carried out on a Hewlett-Packard 3980 A instrument. The GLC column used was a 6 ft. x 2 mm l.d. glass column filled with 0.2Z Garbowax 20M on Chromosorb W 100-120 mesh (Aue or Ultrabond)j. Helium was used as the carrier gas at a flow rate of ca. 30 ml/mln. The injection temperature was 250*C. and the oven temperature was kept at 170*C for 2 minutes and then temp erature programmed to 240C at 8* per minute. For the interface, a jet separator was used at 350*C. The ion source temperature was 215*C and the temperature of the analyzer was set at 110*C. used with the scan rate set at 160 amu/sec. A 70| ev electron energy was i 1 The FCDF/PCDD fraction dissolved in toluene (1 ml) was injected into the GLC portion of the Instrument. The mass spectra and the :.on current were then monitored as the individual components eluted into the mass spectrometer. Individual ions were monitored by operating the GC/W> in the Single Ion Monitoring (SIM) mode. 2.2.2.4 Analysis of the PCDF/PCDD Fraction for PCDF 1I Due to the results obtained from initial GC/MS [analyses of the PCDF/PCDD i fractions obtained in section 2.2.2.2 (see Results) these fractions were subi sequently only analyzed for PCDFs, in particular 2,3',7,8-tetrachlorodibenzo- furan (TCDF) 2.2.2.4.1 Determination of Total PCDF by Perchlorlnatlon 783409 To analyze the PCDF/PCDD fraction for total PCDF the perchlorlnation method, l 81 I as described by Hutzinger et. al., was used. The sample to be perchlorlnated was transferred using petroleum ether to a glass stoppered test tube. The petroleum ether was then removed by evaporation using a hot water bath and VjJtUNiT W IUJ^C 32 N2 gas flashing. To further insure that all traces of the petroleum ether vere removed, three successive additions and evaporations of methylene chloride (2 ml) were made. The perchlorinatlon reagent (BMC; 1 ml; see paper by 59 I Hutzlnger et.al., ) was then added to the residue, the test tubes loosely capped with foil and the reaction mixture heated at ca. 69*C for 2 hours. 25Z hydrochloric a d d (3 ml) was then added and the resultant solution heated for one hour at 70 to 80*C. The final solution was then allowed to cool to room temperature and was then extracted twice with petroleum ether ( 2 x 10 ml) or another suitable solvent. These extracts were then washed, first with dis- tllled water (5 ml) and then with 10Z sodium bicarbonate (5 ml), and then evaporated just to dryness using gas flashing The residue remaining was dissolved In exactly 1 ml of isooctane and then analyzed by gas chromatography. Gas-liquid chromatography (GLC/ECD) was performed using a Hewlett-Packard 63 I I 5710A chromatograph equipped with a Ni electron capture detector (ECD). The column used was a 4 ft. x 4 mm l.d. glass column packed with 3Z OV-101 coated on Ultrabond (or Aue; 0.2Z Carbowax 20M on ChromosJrb W JoO/120). The carrier gas used was Argon/Methane (5Z) at 60 ml/mln. The detector temperature was set at 300#C. The oven and injection port vere operated isothermally at 240*C* The percent recovery and perchlorinatlon efficiency of tK_' method was tested using 0.01, 0.001 and 0.0001 mg samples of dibenzofuran and then com -043 O J*. paring theoretical and actual amounts of octachlorodlbenzofuran formed. 6 6 C 0 1 0 i-TNJHO 2.2.2.4.2 Analysis of PCDF/PCDD Fraction for 2,3,7,8-Tetrachlorodlbenzofuran (TCDF) The amount of TCDF in the aakarel temples was determined by analyzing the PCDF/PCDD fraction isolated in section 2.2.2.2 using GLC/ECD. The chromato graphic conditions used vere the same as those described earlier in section 33 \ 2.2.2.4,1.except eh? oven and injection port temperatures were operated iso- thermally at 225*C. Quantitation of TCDF vas carried out by comparing thie peak areas of standard solutions of TCDF (courtesy of Dr. C. Rappje) to Chose of the peak occurring at the same retention time In the PCDF/PCDD fractions (duplicate !| 2 ul injections) The amounts of TCDF found were then converted to a ppm value on the basis of 0.50 grams of askarel analyzed. I 2.3 RESULTS 2.3.1 Sampling of Transformer Grade Askarels The data on the transformers sampled and on the askarels they contain is summarized in Table 6. The samples are divided into groups and coded according to the scheme outlined in section 2.2.1.2.2. In the time-in- service study (group 1), three samples vere taken from three separate, but Identical, transformers for comparison purposes. It should be noted at this time that the dielectric strength of each of the askarels sampled Is also Included in this table (most recent measurement available). The dielectric strength, as discussed in Appendix A, is a good Indicator of contaminants or decomposition of the askarel (i.e. the lover the dielectric racing the higher the contaminant*) The chlorobenzene and PCS content of each of the askarels sampled are | also given In Table 6. Generally It vas found that the never askarels had a jI I |higher PCB content (ca. 702) and chose used since 1970 contained Aroclor 1254 rather than Aroclor 1260. The major Individual chlorobenzenes found in all '| these fluids vere the 1,2,4- and 1,2,3-trl- and the 1,2,3,4-tetrachloroI benzenes. Small quantities of the other tri-, ttra^ and pentachlorobenzene congeners vere also detected In most samples. As an'example of the analyses, 783411 u n m - uiuouu  Table 6: 34 Transformer Daca, Askarel Composition and TCDF Concentration u> co 1 X co CS o CO u> c 1 1 -o ro r-- CO ro ro ro ro ii1 oo sj O' cn On o ro ro ro ro C 1 1 l 1 ns X CO ro -- to nO o r- t c- r- rt r- r- r- r- r- c 1 1 1 1 i 1 l 1 1 l 1 v r- rt vO 00 SJ 0 r- O cn X co ro it r- Sample It I--* VO VO O' O' sj SJ H* it OO O' O' sj sj cC nn o r> ps PS M f rt cVO vO VO vO O' O' O' O' cn cn cn cn 1 "1 tn "0 ns ns ns r-- r- r- r-- vO VO VO vO O' O' O' O' sj sj SJ SJ nOon oa Oo PS PS PS PS z rt r- r- rt r- rt it it r- r- PS j: vO SJ X VO sj r-- VO soj vo Soj vO 0 Cn vo 0 CO vO 0 CO vO 0 CO vO 0 r- vO cn SJ nr) nr) t) 11 fj ii t) t) t) ns no ns ns ns n3 ns ns to ns Installation Manufacturer Xx ft H- X CO It U) CO CO CO CO O' O' O' O' CO co CO CO O' O' *--1 co i ro 0D r03o ro ro CoO ro vO co co 00 co 0 co 0 CO 0 Dielectric Strength r- is) sj O Cn oO r- O Cn Oo O O ro ro r-- cn cn cn sj Oo O O O O O O *t r-* H- CD O o cn O' o O' o oo oo rt o 09 oo 00 SJ W SJ Sj SJ ro o cn Ho- Is) r-- O ro O' VO ro CD O' X r- oO' H-< ro sj vo SJ O' o o CO oe ro cn rt o Ho- 0D Cn r- VO O' OO ro H- r- |t ro ro ro o vo O O O' H- r- r- O X sj CoO o SJ O' vO rr rr nn fO IS) sj 00 X ro rr n ro ro O' r o X* rr rr rt rr H r| n H ro ro ro ro 00 sj SJ sj t X O O' rr rt rr z nn nu ro ro ro ro O' O' O' X* o ro X r-- h* r- r--rt r- rt r- c-t r- r-- cn cn cn cn cn cn cn cn cn cn cn O O O O O O O O Oo oO OO OO Oo oO O O r* r- r-- rt r- I-* rt r- ro 1 X X X SJ r-- ro ro ro C OO voO voO VoO O' cn Cn r-- o X cn cn VO 0 r-- r- co 0 M M rt O rt rt H- rt O o 1 o O O 0 co I--r-- M SJ SJ H- r- r-- cn O Sj vO vO ro vO) ft z rt rt rr rr ft ft ft rr ft n a n n n n n n n ft n ro ro ro ro ro ro ro K ro ro ro O t r-- rt ro oe 0 SJ 0 0 oe O' ro ro O X 0 0 X X Power (kva) Volume (litres) lcva/volume 1,3,5-Cl-jBz 1,2,4-Cl Bz -vl ocoo X ro sj 0o Cn o O' oe ro O' O' cn 00 sj 00 O' SJ cn X o ro O' vO 0 O' CD SJ O' 00 SJ O SJ 0 0 t o 0 X 00 ro 0 0 0 1,2,3-C13Bz rr rt ri n rt rt n n rr rt nn rXx O' 0D rt rt ri n Ut U) ro X CO co cn O' ro ro Ox' O' co cc 00 en x O Cn O' ro X cn O X X" OO i-- t-- ro ro O' O' oO r- r- ro Is) O' O oO' o oO rr rr rr rt t n n n rt rt rt rr n r| n n *t rr r-- CO rt r| n O O' rt H- r - (-- CO X X X X X CD oe rt rt ro rt *1 *i H O' ro CO VO co oo ro ro X CO CO X* CO O' cn VO cn o O' X X* O' O' cn O' u> X r- X X" O' O O' X cn X X vO O X cn o O' CO ro cn X cn cn -- vO cn X O X so oo c- H- r- I t ro ro ro ro O' o O' o O' Oo' r- c- 1-- it ro ro ro ro O' O' O' O' o o OO oooo oooo ft n zo ft ft ft ft rt ft rt ft Hn Hn n n Hn H o r- r- M It rt n zo rt n rt rt r-- nn ro X X cn X 0 ro X o ro X z o z a zo z a z a zo ft ro ro co ro n 0 X X CO O raoo ro vo ro ro vO CO sj co cn ro cn O cn co cn 9 O' X O O 0 ro 0 X 0 0 O s. o 0O' sj Sj sj sj 0 X X X X vO r-- r- ro ro r- SJ v 0 vo XOOOX0X0XXO rt *t rt rt C- r- r- r- r- r-- ro ro ro to ro ro ro ro ro ro ro cn cn cn cn cn 0 0 0 0 0 0 X X X X X OOOOOO O O o O O O o O f r- o l f2,3,5- & l,2v4 t5-C 1,2,3,4-Cl^Bz 1,2,3,4,5-C15 Bz PCBz/PCB ratio Aroclor TOQOm /-TKran \ \ 35 X* X' 111 N J O' Ln X* X* 11 X* uj a * o X X- C 1 1 o NJ H- X* u> u> 11 O' cn m Oto 9 c mCm fst--t- nrj n V c o m r- rvO vO nj O' uj cn C> > t-- M VO vO X X NJ NJ nn oo en m r - HVO sO O' O ' Nj nrl 11 OH- ooo Cn O' O' O' VoOO Oo' H- o X o Cn CD r* t zo rr I z zo zo z o oN J NoJ UJ Cn oo KJ X* NJ NO n j N J o NJ NJ cn X' NJ 00 V-* NT VO O UJ 00 o X cn cn NJ NJ O' NJ O' NJ n rr rr n 1n rr rr rl r| r* rr r| ri r- rr *1 Zo O M HX* X* 00 X' r- t--1 X" X* o NJ Z Zo z X* UJ o O' NJ U> O' UoJ UoJ cn 00 O' N J N^ 00 N^ O' O' X* VO VO H- X" 00 N J cn Cn r- O X N J N^ X* X 00 VO OV 00 X* X* 00 vO X* NJ *N* cn cn r- O O' 00 1-- *-* r-- NJ NJ NJ Cn cn cn X* X* X* A r- NNJ NJ O' O' OO t-- HN NJ O' O OO CO u> O' H- rooo- nJ cOn 09 00 uj NJ 00 vO h- o rVO voO nrr rr NJ NJ Cn O o09 N NJ 00 CD rr rr *1 H f* rr *1 *1 r1r rHr UJ UJ N> NJ 00 00 *N *. O' O' NJ NJ NJ N J H- uNJ NJ oO oO Sample Installation Manufacturer _ Dielectric Strength Power (kva) Volume (litres) kva/volume 1 .3 .5 - C13B2 nj CD 1.2,4-Cl.Bz J CO l f2,3-Cl3Bz 1.2.3.5- & 1,2,4,5-C 1,2,3,A-Cl^Bz 1.2.3.4.5- Cl^Bz PCBz/PCB ratio Arodor 700010 rlM ^ n Table 6 : Legend - Samples: Croup 1 (1-1, 1-2, etc.) - variable time-in-service. Group 2 - 2-1 to 2-4 - CCE Transformers; variable kva/volume ratio - 2-5 to 2-6 - FP Transformers; variable kva/volume ratio Group 3 - variable fluid type - 3-1 and 3-2 Pyranol 3-3 and 3-4 Chlorextol 3-5 and 3-6 Inerteen Group 4 - 4-1 and 4-2 - 'very old' transformers; large variation in kva relative to fluid volume. - 4-3 and 4-4 - transformers which had failed in service AD - Adelaide Street (Toronto Hydro) VIA - Walkerton, Ont. (Canada Packers) - 4-5, 4-6 and 4-7 - new Pyranol, Inerteen and Chlorextol. Manufacturer: (fluid type) CGE - Canadlan General Electric FP - Ferranti-Packard W - Westinghouse AD - Brown Boveri WA - Foster Electric (Pyranol) (Chlorextol) (Inerteen) (Pyroclor) (Inerteen) - 1,3,5-ClsBr - 1,3,5-trichlorobenzene content ( 0.05Z w/w) - 1,2,4-CljBz - 1,2,4-trlchlorobenzene content ( 0.05Z w/w) 1,2,3-CljBz - 1,2,3-trlchlorobenzene content ( 0.05Z w/w) - 1,2,3,5- 6 1,2,4,5-CKBz - 1,2,3,5- A 1,2,4,5-tetrachlorobenzene content (i 0.05Z w/w) - 1,2,3,4-CKBz - 1,2,3,4-tetrachlorobenzene content (1 0.05Z w/w) - 1,2,3,4,5-ClsBz - 1,2,3,4,5-pentachlorobenzene content ( 0.05Z w/w) - TCDF (ppm) - concentration of 2,3,7,8-tetrachlorodibenzofuran in askarel (ppm) HPC8Z f090/o Table 6 : Legend u> o 37 the GLC/FID chromatograms of a 'new* Chlorextol and a used (1957) Chlorextol are shown In Figures 7a and 7b respectively. 2.3.2 Analysis of Askarel Samples for FCDF and PCQD 2.3.2.1 Isolation of PCDF/PCDD Fraction A r o d o r 1254, precleaned (i.e. PCDF and PCDD removed) on a florisil column, was spiked with varying amounts of TCDF and then subjected to the fractionation procedure outlined in section 2.2.2.2. In all caseL greater than 802 recovery of the TCDF was obtained using this method. Any interference due to large quantities of PCB was also effectively removed usually after one pass through the alumina and florisil columns. 2.3.2.2 GC/MS Analyses of the PCDF/PCDD Fraction Figure 8 shows the total ion chromatogram of the PCDF/PCDD fraction iso lated from an old Aroclor 1254 sample. The major peaks are assigned according to their mass spectra. It is of some interest that terphenyls and quaterphenyls were detected in this fraction of the PCB mixture. Although not visible as peaks on the total ion chromatogram, penta- and hexachlorodlbenzofuran were also detected in this fraction when the GC/MS system was operated in the SIM mode. Terphenyls and quaterphenyls were identified in most of the PCDF/PCDD fractions analyzed although usually in very small amounts. Tetra-, penta-, and hexachlorodibenzofurans were detected in all of the as carel samples ana lyzed by GC/MS. In some samples analyzed by GC/MS in the SIM mode, traces of PCDDs were detected but their quantities were usually less than 5Z of the individual PCDFs detected. From these initial results it was assumed that these levels of PCDDs were Insignificant relative to those of the PCDFs (i.e. < 0.1 ppb total PCDDs). 783415 010604 % 38 c n o m n risico Figur 7 : GLC Anslysis of Asksrel for Chlorobenzenes nd PCB 783416 39 GENP 010606 Peaks Identified 1 . terphenyl 2 . terphenyl 3. Cl terphenyl 4. CI4 dibenzofuran 5. Cl terphenyl 6 . quaterphenyl 7. Clg terphenyl Figure 8 : GC/MS Analysis of PCDF/PCDD Fraction from Aroclor 1254 783417 40 In the remaining portions of the project, therefore, the askarel samples were only analyzed for PCDFs. As an example of the operation and output of the GC/MS system in the SIM |i mode, that is monitoring for individual ions, the total Ion chromatogram and the individual ion scans for tetra-, penta-, and hexachlorodlbenzofurans for the PCDF/PCDD fraction taken from a used Chlorextol sample are shown in Figure 9. 2.3.2.3 Perchlorination of the PCDF/PCDD Fraction [to Determine Total Dlbenzofuran IThe perchlorination method as described in section 2 .2.2.4.1, was tested for percent recovery using TCDF and dlbenzofuran. In all cases the recovery, Idetermined using standard solutions of octachlorodlbenzofuran (courtesy of Dr. C, Rappe), was quite low. Even when the conditions and extracting solvent were varied, the percent recovery remained < 30Z and was quite variable thus making this method Inadequate for quantitation of total PCDF. This method, however, was used la some initial experiments to confirm qualitatively that i idibenzofurans were present in the PCDF/PCDD fractloJs of some askarel and PCB samples. 'In all the fractions tested in which TCDF was identified by GLC/ECD, octachlorodlbenzofuran was formed when these samples were perchlorlnated. For example. Figure 10 shows the GLC chromatograms of the products obtained (l.e. octachlorodlbenzofuran) when dlbenzofuran (0.01 mg) and the PCDD/PCDF fraction : I from A r o d o r 1254 were perchlorlnated by this method] 2.3.2.4 Quantitation of 2,3,7,8-Tetrachlorodlbenzofuran (TCDF) in Askarel Samples I 1The amounts of TCDF in ppm detected in the askarel samples are given in Table 6. All the samples were partitioned and analyzed in duplicate and the 7|83418 G E N F uivw o n i N r uiuou 41 I i PCDF/PCDD Fraction from Aroclor 1254 I .01 mg dibenzofuran octachlorodibenzofuran CO CO ji ro o . GENP 010609 Figure 10 : Perchlorination Products GLC/ECD 4 A3 results given are Che average of the values determined. Generally it was found that the fractionation method and subsequent GLC/ECD analyses were quite reproducible. The GLC/ECD chromatograms of the PCDF/PCDD fractions isolated from Inerteen, Pyranol and Chlorextol used in transformers for ca. 30 years (i.e. since 19A7) and the askarel taken from the transformer involved in the Adelaide Street fire are shown in Figure 11. Group 1 : Time-In-Service Although the results are somewhat scattered, the level of TCDF in an askarel was found to Increase as the time since the transformer was installed increases. It was anticipated that this might be a linear relationship but the data does not allow such a fit to be made. The three askarels taken from 'identical' transformers (1-3, 1-A and 1-5) showed a marked variation in TCDF levels. This indicates that time-in-service may not be the only variable in volved for the transformers samples. Group 2 : KVA/Volume For the two sets of transformers (two different manufacturers) where the only variable was the kva/volume ratio, the levels of TCDF showed no observ able trend, in fact, they remained reasonably constant in each series. It is possible that even for the highest kva/volume ratio, the liquid volumes are already so large that any increase in fluid volume does not decrease the operating temperature significantly and thus will have little or no effect on formation of PCDF6. Group 3 ; Manufacturer/Fluld Type In comparing the three major fluid types (Pyranol, Inerteen and Chlorextol), the levels of TCDF in Chlorextol were lower (ca. 25X) than those 783421 vjfcJNj- 0 1 0 6 1 0 (JbNP 0IU6II AA 45 in the other two fluids. Whether this is due to the actual fluid or to the construction of the transformer is not known. The three fluids sampled all have essentially the same composition and the transformers should all be of the same basic construction. The lower levels of PCDFs in the Chlorextol filled units may possibly be due to a 'purer* batch of Aroclor in the original fluid, or to lower operating temperatures of these transformers. Group 4 The two transformer askarels (4-1 and 4-2) sampled from units that had been in service for approximately 30 years, were contaminated with ppm levels of TCDF. The level of TCDF in the askarel from the transformer with the lower kva (kva/volume 0.08) was more than twice that in the higher kva transformer askarel. The levels of TCDF in the two transformers which failed while in service were not extremely high relative to the other askarels tested. Short periods of arcing or 'above normal' temperatures may not be enough to Induce changes in the askarels. The soot from the Adelaide Street fire did contain high levels of PCDFs compared to the transformer oil, but these were more likely formed when the spilled oil was exposed to the high temperatures of the fire. The PCDFs in the oil itself were probably formed during the normal operation of the transformer. The final three samples tested in this group were unused Inerteen, Pyranol and Chlorextol. They contained only low levels of TCDF relative to the used oils and represent a 'zero' level. The TCDF found in these oils is most probably formed from the chlorination of a dlbenzofuran impurity in the original biphenyl. 783423 GENP 010612 46 / 2.4 DISCUSSION l Only one of the variables tested in this portion of the project, that being i time-in-service, vas determined to have an effect on the TCDF levels in askarel; 1 I *the longer the unit vas in service, the higher the concentration of TCDF in the fluid. It vas anticipated that this relationship would be linear, but the final data vas too scattered to be fitted to a line. 1 Variations in the kva/volume ratio and the fluid tyjje had no consistent effect on the TCDF levels. In two 'worst case' samples (4-3 and 4-4), where there vas an extreme increase in the kva relative to the fluid volume in the transformers, there vas a corresponding increase in the TCDF levels. Whether this increase vas only due to the kva or kva/volume ratio could not be resolved ii due to the limited number of samples. i |Analysis of the data also suggests that there may be' a number of unexpected < variables in a transformer and its operation which may all contribute to the ' formation of PCDFs. For example, sample 2-6 (see Table 6) has a different 1 lchlorobenzene (PCBz) content and composition from the other samples in this aeries, and, at the same time, has a higher TCDF concentration. As previously mentioned, PCBzs, when heated at ca, 600*C in quartz ampoules, yield large 82 I amounts of PCDFs and thus the presence of certainPCBz congeners may be responsible for the formation of PCDFs In transformer asklrels. The load under which a transformer Is operated Is also suspected to be a major factor In the formation of PCDFs. Take, for exampli, samples 3-1 and i| 2-1. These transformers are Identical in all respects except that 3-1 is in the Central Utilities Plant and 2-1 is In a residence, both at the University of Guelph. C-riiNP 0 1 0 6 1 3 i 783424 47 Although accurate records are not available, 3-1 is usually operated under a higher, more continuous load than 2-1, and the askarel taken from 3-1 has a higher TCDF concentration. Unfortunately, due to the unavailability of accurate records, transformer load could not be isolated as a single i variable for a sufficiently sized sample group. ! ! The effects of discharging or arcing in a transformer on the levels of FCDFs in the contained askarel appear to be negligible in the two transformers sampled. As previously suggested, these periods of abnormal ii operation are most likely too short in duration toj cause any reactions to I occur in the askarel. j It should be pointed out at this time that, although the retention time of the peak in the PCDF/PCDD fraction corresponds to that of TCDF, ij there may be more than one tetrachlorodibenzofuran|isomer involved in this peak (there are 38 tetrachloro Isomers possible). GC/MS however, does confirm the presence of tetrachlorodlbenzofurans in these fractions. In summary, significant levels of FCDFs are formed in askarel filled transformers during their normal operation. The detection of these com pounds as contaminants in FCB-based fluids currently in use even more necessitates strict controls on the use of FCBs and the fcomplete1 destruction of these fluids when they do come out of service. U1U14 ! I 783425 % 48 3 INCINERATION/TRAPPING EXPERIMENTS USING AROCLORS AND ASKARELS___________________ 3.1 INTRODUCTION Laboratory scale pyrolyses of individual PCB congeners and commercial mixtures, as previously mentioned, has been found to yield significant quan37 41 46 47 titles of PCDFs * * * . To our knowledge there have been no reports as to the products formed when askarels (i.e. chlorobenzene/PCB mixtures) are pyrolyzed. In the following phase of this project, the incineration of PCBs and askarels is simulated using a laboratory scale pyrolysis unit, and the formation of PCDFs and PCDDs investigated. 3.2 EXPERIMENTAL 3.2.1 PCBs and Askarels used for Pyrolysis Two PCB mixtures, Aroclor 1254 and Aroclor 1016, and three askarels, Chlorextol, Pyranol and Inerteen, were obtained during the sampling carried out for Section 2. These mixtures were then purified (i.e. PCDF/PCDD fraction removed) using the method described in Section 3.2.3.1 prior to using them for the pyrolysis experiments. 3.2.2 Pyrolysis of PCBs and Askarels 3.2.2.1 Apparatus The pyrolysis unit consisted basically of two ovens, a movable flash heating oven and a fixed reaction oven, surrounding a quartz tube (20 cm. x 0.4 cm i.d.) as shown in Figure 12. Compressed air (technical), filtered through sulfuric acid on silica gel (20T w/w) and activated charcoal, was fed into the sample at 200 ml/min at a pressure of 6 atmospheres. The pro ducts of incineration were collected at two points; at the end of the quartz 783426 GENPUlUblD I 69 tube, Q, and in a cold trap connected to the end of the tube, C. 3.2.2.2^ Procedure The PCB or askarel, dissolved in hexane, was*deposited on glass vool in the reaction tube (see Figure 12). The weights of PCS or askarel used in the pyrolyses are given in Table 7. The hexane was allowed to evaporate with no air flow prior to switching on the ovens. ^ The fixed and movable ovens were then set to operate at ca. 500*C and ca. 650*C respectively. Once both ovens had reached these temperatures, the air flow was started, dry ice/ethanol cold traps were set in place, and the mov- able oven was positioned around the portion of the quartz tube containing the glass wool and the PCB. After 3 minutes, the ovens were switched off and the cold trap and reaction tube were removed. The products at Q and C were then . f collected for analysis by dissolving them in hexane. 3.2.3 Analysis of Pyrolysis Products 3.2.3.1 Isolation of PCDF/PCDD Fraction .The hexane solutions of the pyrolysis products were first concentrated to ca, 0*5 ml. This concentrate was then quantitatively transferred to the * * ii * top of a 'microcolumn packed with sulfuric a d d on silica gel (20Z by weight) and then eluted with hexane (6 m l ) . This eluate was then again concentrated to ca. 0.5 ml and then applied to the top of a micro alumina column (basic alumina; 1 g.). This column was then eluted, first with 2Z methylene chloride in hexane (10 ml), to remove chlorobenzenes and PCBs, and then with 50Z methylene chloride in hexane (10 ml.), to elute th dioxin and furan pro ducts. The second eluate was evaporated just to dryness and then reconstituted with toluene containing 2,5-dichloro-4'-methylbiphenyl (5 mg/ml) as an internal standard. This solution was then analyzed and quantitated by GC/MS In the SIM 783427 GEJNF UlUOif) movable flash heating oven GENP 010617 783428 Figure 12 ; lnclneratlon/Trapplng Apparatus for Pyrolysis of PCBs m 51 i v \ i Table 7 : Pyrolysis of PCBs and Askarels PCB or Asteare 1 Aroclor 1254 Weight of Tetrachlorodibenzofurans Formed Weight Position Actual Pyrolyzed of Sample Weight (mg) (Ug) Per Gram of Starting Material (VgJ g) Total (Q+C) Average of Both Runs (Vg/ g) 25.0 40.0 Q 2.20 C 1.33 Q 4.90 c 3.03 88. 0 53.\2 122. 5 75. 8 141.2 197.3 169.3 Aroclor 1016 25.0 40.0 Q 0.038 c 0.027 Q 0.072 c 0.026 1 .52 1. 08 1. 80 0, 65 2.60 2.45 2.53 Pyranol 25.0 0.25 Q 0.53 c 0.53 Q 0.048 c 0.074 21. 2 21. 2 191. 7 297. 0 42.4 488.8 KA Inerteen 0.25 0.25 Q 0.030 c 0.030 Q 0.033 C 0.003 16. 3 121. 9 133. 7 12. 2 240.2 145.9 193.1 Chiorextol 0.25 0.25 Q 0.023 C 0.025 Q 0.021 C 0.041 93. 5 100. 2 82 J6 165.7 i ii 193.7 248.5 221.1 ii 8TOOTO -TKTrrrN 783429 52 node. To continually check the response of the GC/MS, repeated Injections were made of a standard solution containing the internal standard, 1,23,4-tetrachlorodlbenzo-g-dloxin and octachlorodibenzo-p-dioxin (all 5 mg/ml in toluene). 3.3 RESULTS The results of the incineration/trapping experiments are summarized In Table 7.. The amounts of tetrachlorodlbenzofuran formed are reported as ug/gm of starting material. Analyses of the 'deaned-up1 starting materials prior to pyrolysis showed no PCDFs or PCDDs. GC/MS analysis of the pyrolysis products revealed that only PCDFs are formed when PCBs or askarels are pyrolyzed. No PCDDs or other com pounds such as terphenyls or quaterphenyls were detected by GC/MS. A typical GC/MS - SIM for the pyrolysis products of an askarel is shown, in Figure 13.. The pyrolysis experiments for only one temperature (^500C) are reported. Preliminary experiments at ^100*C gave no detectable change in the starting material within the .3minute reaction time. 3. A DISCUSSION The formation of PCDFs when PCBs or askarels are Incinerated is consistent with past research. Their formation at temperatures of p . 500*C indicates that for 'safe' Incineration of PCBs, temperatures of at least 700*C are necessary (> 1000*C preferably). The absences of any other contaminants, such as terphenyls and quaterphenyls, !| in the pyrolysis products suggests that their presence in Aroclors or askarels i| is most likely due to contamination of the original biphenyl. They do not Iform under pyrolytic conditions. It is also Interesting to note Chat Aroclor 1016 gave significantly less PCDF than Aroclor 1254 when incinerated. These 783430 TTOATA TkT'nrrx 305.9 3 0 3 . c. Tetrachloro41benzofurans 11 . f 375. E Hexachlorodibenzofurane 373. E l.tt:8Z 07,9010 <USL30 Figure 13 : Individuai Ion Scans (SIM GC/MS) for Tetrachloro- and HexacfrlorodlbenzoPurans In PCDF/PCDD Fractlon oE Pyrolyzed Pyranol * 54 lower levels are to be expected since 1016 is of lower chlorine content and therefore does not contain as much of the higher chlorinated biphenyls (i.e. pentachloro- or hexachloro-) which are necessary precursors of tetrachlorodibenzofurans. In summary, the formation of PCDFs when askarels are exposed to sufficient oxygen and high temperatures demonstrates that fires associated with PCB-filled transformers are potentially hazardous and that any situations where PCBs may be pyrolyzed should be handled with extreme caution. 783432 170010 flhiao < i 55 REFERENCES 1 A.E. Pohland and G.C. Yang, J.Agrie.Food Chem.. 20, 1093 (1972). II 2 A.P. Grey, S.P. Cepa, I.J. Solomon and 0. Aniline, J.Org.Chem., Al_, 2435 (1976). 3 H.R. Buser, J.Chromatogr.. 114, 95 (1975). 4 A.P. Grey, V.H. Dlpinto and I.J. Solomon, J.Org.Chem., 41, 2428 (1976). 5 A.S. Kende, J.J. Wade, M. DeCamp, D. Ridge and aL Pohlainndd,, 167th'Annual Meeting of the ACS, Los Angeles, Calif., 1974, Abstract No.ORGN-130. 6 H.R. Buser, Doctoral Dissertation, University of Ume, 1978. 7 0. Aniline, Adv.Chem.Ser., 120, 126 (1973). 8. M. Harlsada, Yakugaku Zasshi, 79, 183 (1959). 9 A. Norstrom, R. Anders son and C. Rappe, ChemospHere. 5_, 21 (1976). :i 10 G.G. Choudhry, G. Sundstrom, F.W.H. van der Vielen and 0. Hutringer, Cheaosphere. 6., 327 (1977). ! . ii 11 H.R. Buser and C. Rappe, Chemosphere, J7, 199 (1978). 12 H.R. Buser, J. Chromatogr., 107, 295 (1975). ! i 13 R. Olie, P.L, Vermeulen and 0. Hutringer, Chemosphere 6, 455 (1977). 14 D. Firestone, J.Agrie.Food Chem., 25 . 1274 (1977). 15 G.W. Bowes, H.J. Mulvihill, H.R. De Camp and A.S. Kende, J.Agrlc.Food Chem., 23, 1223 (1975). 16 H.R. Buser, Anal.Chem.. 48. 1553 (1976). 17 H.R. Buser, Anal.Chem., 49, 918 (1977). 18 D. Firestone, J. Rees, N.L. Brown, R.P. Barron and J.H. Damico, J.Assoc.Off.Anal.Chem.. 55. 85 (1972). j ~! 19 C.A. Nilsson and L. Renberg, J.Chromatogr.. 89,|325 (1974). 20 H.R. Buser and H.-P. Bosshardt, J.Assoc.Off.Anal.Chemi. 5 9 . 562 (1976). 21 V.V. Blaaer, R.A. Bredeveg, L.A. Shadoff and R.H. Steil, Anal.Chem., 48, 984 (1976). '1 --------- 22 E.C. Villanueva, R.U. Jennings, V.W.'Burse end R.D. Kimbrough, J.Agrlc.Food Chem., 23, 1089 (1975). 783433 . GENP 010622 s ' 57 REFERENCES (continued) 43 D.S. Duvall and W.A. Rubey, Laboratory Evaluation of tigh-femperature Destruction of FCBs and Related Compounds report for EPA under Contract No .R-803540-01-0# 1977. | I 44 R. Komamiya and S. Marisaki, Environ.Sci.Tech.. 12, 1205 (1978). 45 L. P. MacDonald, D.J. Skinner, F.J. Hopton and g Jh . Thomas, Burning Waste Chlorinated Hydrocarbons in a Cement Kiln report for Fisheries and Environment Canada, EPS 4-WP-77-2, 1977. 46 H.R. Buser, H.-P. Bosshardt and C. Rappe, Chemosphere, ]_ % 109 (1978). II 47 Ministry of the Environment, Ontario, Canada, News Release, February 13, 1978. j 48 M. E. King, A.M. Shefner and R.R. Bates, Environjllealth Perspect., _5, 163 (1973). 49 V. Zitko, D.J. Wildish, 0. Hutzinger and P.M.K. Choi, Environ.Health Perspect., .5, 187 (1973). 50 M. Morita and S. Oishi, Bull.Environ.Contam.Toxicol., 18, 61 (1977). 51 ` A. Poland and E. Glover, Mol.Pharmacol., , 736 (1973) 52 A. Poland, E. Glover, A.S. Kende, M. De Camp and C.M. Giandomenico, Science, 194. 627 (1976). i ]53 J.A, Moore, XARC Meeting on FCDDa and PCDFs, Lyon, January 10-11, 1978. 54 H. Bauer, K.H. Schulz and U. Splegelberg, Arch.Geverbepath.Geverbehyg. 18, 538 (1961). 55 J.A. Moore, B.N.* Gupta and J.G. Voa, National Conference on PCBs, Chicago, 1975, pp. 77-80. | 56 M. Nlshlzuml, Toxicol.Appl.Pharmacol.. 45, 209 (1978) 57 S. Oishi, M. Morita and H. Fukuda, Toxicol.AppljPharmacol., 4 3 . 13 (1978) 58 M. Kuratsume, T. Yoshlmura, J. Hatsuzaka and A.Yamaguchi, Environ.Health Perspect.. 1 9 119 (1972). j 59 Y. Ikeda, J .Fd.Hyg.Soc *Japan. 1 3 . 359 (1972). 60 J. Nagayama, M. Kuratsume and Y. Hasuda, Bull.Environ.Contam.Toxicol., 15, 9 (1976). ------ j--------------------- 61 J.H. Vlnopal and J.E. Caslda, Arch.Environ.Contarn.Toxicol., ^1, 122 (1973) 62 G.F. Fries and G.S. Marrow, J.Agric.Food Chem., 23. 265 (1975) 783434 U iN r VI 58 REFERENCES (continued) 63 M.Th.M. Tulp and 0. Hutzinger, Chemosphere, 761 (19711). 64 W.N. Pyser, J.Q. Rose and P. Gehrlng, Environ.Health Perspect., <5, 241 (1973). 65 J.Q`. Rose, J.G. Ramsey, T.H. Uentzler, R.A. Hummel and P.J. Gehrlng, Toxicol.Appl.Pharmacol. 36. 209 (1976). 66 A. Poland and E. Glover, Science, 179. 476 (1973) . 67 A. Poland and E. Glover, Environ.Health Perspect.! _5, 2|3 (1973). .68 A.S. Kende, j.J. Wade, D, Ridger and A. Poland, jlorg.Chem., 39. 931 (1974). II 69 P. Beatty and R.A. Neal, Blochem.Pharmacol., 2 7 , 505 (1978). 70 A.J. Baars, M. Jansen and D.D, Brlemer, Blochem.Pharmacol., 27. 2487 (1978). 71 J.O. Nelson, R.E. Menzer, P.C. Kearney and J.R. Plummer^ Bull.Environ. JContam.Toxicol., 18, 9 (1977). 72 J.G. Zinkl, J.G. Vos, J.A. Moore and B.N. Gupta, Environ.Health Perspect., .5, 111 (1973). 73 J.G. Vos, J.A. Moore and J.G. Zinkl, Toxicol.A p p l Pharmacol., 29 , 229 (1974). 74 G. Jones and J.B. Grelg, Experiencia, 31. 1315 (1975). 75 B.A. Schwetz, J.M. Norris, G.L. Sparschu, V.K. Rowe, P.J. Gehrlng, J.L. Emerson and G.G. Gerblg, Environ.Health Perspect., 87 (1973). 76 E.E. McConnell, J.A. Moore and O.W. Dalgard, Toxicol.Appl.Pharmacol., 43, 175 (1978). " 77R.J. Koclba, P.A. Keeler, C.N. Park and P.J. Gehrlng. Toxicol.Appl. Pharmacol., 35, 553 (1976), 78 R.J. Koclba, D.G. Keyes, J.E. Beyer, R.M. Carreon, C.E. Wade, D.A. Dlttenber, R.R. Kalnlns, L.E. Frauson, C.N. Park, S.D. Barnard, R.A. Hummel and C.G. Humlston, Toxicol.Appl.Pharmacol., 46 . 279 (1978).; 79 D. Neubert, P. Zens, A. Rottenwallner and H.J. Merker, Environ.Health Perspect.. 67 (1973). | j 80 J.A. Moore, B.N. Gupta, J.G, Zinkl and J.G. Vos, Environ.Health Perspect., 5, 81 (1973). j 81 0. Hutzinger, S. Safe and V. Zltko, Intern.J.Environ.Anal.Chem., _2 95 (1972) 62 H.R. Bus er, Chemosphere, 8,, 415 (1979). 783435 GENP 010624 59 APPENDIX A TRANSFORMERS (POWER TYPE) : BASIC THEORY, CONSTRUCTION AND USF. 1 INTRODUCTION The generation, transmission and distribution of electric power 16 achieved by using vhat are called three-phase power systems consisting of three-phase generators, transformers and transmission lines. These systems are more economical chan single-phase systems when Jarge amounts of electri- cal energy have cn be generated and distributed. It Is not very practical to generate electricity at large voltages, but It is practical to transmit power over large distances at high voltages since voltage drops as the distance from the source Increases. Transformers are, therefore, used at generating stations to step up the generated volt ages to transmission levels. A common value for. a generated voltage Is 13,800 volts, while transmission voltages can be as high jas 220,000 volts. Once the electric power reaches the distribution points the voltages are decreased, using step-down transformers, to a levelj suitable for domestic and Industrial use. A typical power system Is shown In Figure 1. Step-up, power trans- formers are primarily located at generating stations where the step- down type are at the substation and distribution Installations. Brief discussions of transformer theory and construction Lre given In the fol lowing sections. For more detailed Information, tJe references given should be consulted. 2 Theory of Transformer Operation 783436 i A transformer Is described as an alectrlc device wit bout continuously I moving parts which, by electromagnetic Induction, transfers energy from one i U M P 010625 & w :; t J' .;L. ' ; . \ GENP 010626 783437 www H v M f f V rsvj= t;*.; 1; 'fb j p v V 'W . ' ..f f r > 783438 G E N P 010628 61 circuit to another. A simple two coil transformer is represented in Figure 2 The two coils, Winding A and Winding B, are inductively coupled through ja ferromagnetic ring called the iron core. Any magnetic flux ($) passing through coil A will also pass through coil B, that is, the two coils have a magnetic circuit common to both. If the current across a, b changes, the Primary Coil Secondary Vp Source Figure 2: Two Coil Transformer mutual flux across both coils changes accordingly and Induces a change in the current in coll B and likewise s change in the voltage across c, d. This induced voltage change in the aecondary coll (B) is called the transformer voltage while the action that creates this change in potential difference is known as transformer action. Host electrical equipment makes use of alternating current (AC) which, in the case of a transformer, would be applied across c, b (see Figure 2). The current in coll A thus changes in magnitude and direction (sinusoidal) which then causes the flux in the core to vary in a similar fashion. The 783440 UJtiiNF U1U629 62 Induced voltage In the secondary coil (B) is thus also a sine wave but lagging the applied voltage by 90*, that is, AC current is [produced in coil B. I| If assumptions are made regarding the various construction features of transformers such as no coil resistance, zero Iron loss and a linear magnet ization curve, the maximum voltage in coll B is determined by the applied voltage across a, b. This voltage induced in coil |b by transformer action is produced by the same flux that crosses coll A. |ln an ideal transformer, the ratio of the induced primary voltage (Ep) in coil A to the secondary induced voltage (Es) in coil B is the same as the ratio of the number of primary turns to the number of secondary turns (Np/Ns). Assuming very little resistance in the windings, the Induced primary voltage (Ep) will equal the applied primary voltage (Vp). across a, b. With the same'assumption, a (Vs) similar relation would exist for the secondary induced (is) and applied voltages and the fallowing relation would exist, Np/Ns" Ep /Es Vp /Vs (1) where "a" is called the transformer ratio. If Vs Is less than Vp (i.e. a>l) the transformer is termed a step-down type. Conversely, if Vs is greater than Vp (i.e. a<l) the unit is called a step-up transformer. 3 Three-Phase Transformers In the pest, when large voltages were to be reduced, a series of three similar transformers were connected together to supply, wihat is termed, three-phase service. There are four standard ways Jo connect three single transformers together to supply this type of servlJe. These four arrange ments are termed Y-Y, A-A, A-Y and Y-A. As an exam pie, Figure 3 is a diagram of a common A-Y (delta-wye) arrangement for a three-phase transformer system (step-down). 783441 GENP 010630 Figure 3 : A-Y Three-Phase Transformer System This type of connection (Figure 3) la useful because the secondary volt ages can be used to either supply three-phase power equipment (208) or lower voltage equipment such as lights (120). \ I IA schematic representation for th A-Y connection Is shown In Figure 4, The common primary and secondary voltages for a power transformer are 13t800 j |and 600 volts respectively. This Is a transformer ratio of 23/1 (i.e. step- down). In the delta (A) arrangement on the primary side the voltage obtained between any two lines Is 13>800 volts (i.e. V, 13,800). The BG Y (wye) arrangement on the secondary side is shown with a ground to the neutral and the line voltages are 600 volts (i.e. - VQF - V " 60) 783442 TCQOTO 64 Primary 13,800 volts 600 volts Figure 4: Schematic of A-Y 3-Phase System Another voltage, 347 volts, can be obtained from the Y side by connecting any one of the three lines to the ground. In the earlier example (Figure 3) voltages of 206 and 120 were obtained this way. Compared to three single-phase transformers (i.e. to make up a threephase system), It la more economical to construct a single three-phase transformer. A three-phase transformer makes use of only one core and thus saves a great deal of iron, copper wire (used In windings) and insulating fluid. These Individual three-phase units may be wired Internally to give the A-A, A-Y, Y-A or Y,Y configurations (Figure 5). 4 CONSTRUCTION 4.1 Iron Core Transformers Host common power transformers use a ferromagnetic or iron core. There are two types of Iron core transformers, core and shell type, which vary basically In the arrangement of the windings o n .the core How the colls are arranged In these two types of transformers Is depicted In Figure 6 This figure also shows how the successive layers of laminations are assembled in 783443 GENP 010632 783444 (a) - (b) Core Type (e) Shell Type 9 0 TO cNi-rr 11 - (e) rt i, r *, : i >- V> p j f !! tt J i ii ) Il II A A Fleure 5 : Throp-Phaoo Tr^nofflt'rurti" l.M I D> t> / 66 kwwwvi ... 1---------------- - . GENP010634 Windings Cors Sr_-- Eleven r\ LO TilW ? lB ra ^ U v a u i % H ( t i g i v jaM .~* ns^im u tuu uimicma. (b) ^ -7 Ldminofions S 1________ %m,A>\L.. /V/>/"*iyr.i/v! - . 1\ W : ' 4'* 7 Secondary Coils /BuffJoinf Figure 6 : Core Type, (a), and Shell Type, ( b ) , Tran*f0m e r Shoving Lamlnaclona 783445 67 both types. In the core type, the primary and secondary colls surround most of the core with each of these coils wrapped around a separate leg of the core. In the shell type transformer, the situation Is reversed with the core sur rounding the coils. 4.2 Windings In three-phase transformers, the primary and secondary windings are situated around the core as shown in Figure 5, which also shows the various connections used. As stated earlier, transformer action exists when a mutual flux is present between the primary and secondary colls. 1The isolation of the individual windings, as depicted In Figure 5, minimizes any flux leakage while at the same time ensuring this mutual flux. This leakage, flux Is further de creased by having the primary and secondary colls Interleaved. This arrange ment also optimizes transformer action and decreases the voltage per coil which In turn extends the life o f the transformer. In the coils of three-phase transformers, the early windings are often paced further apart than the others because under normal conditions the effect of stress voltages are more pronounced in the early part of the coll. 4.3 Insulating Fluids 4.3.1 Composition The main insulating fluids used. In present-day transformers are mineral oils, silicone oils and askarels. 'Askarel' Is the generic name for a series of synthetic liquids which have been used In transformers and capacitors for over forty years. These liquids provided a non-flammable, non-oxldlzlng, inaula ting and heat transfer medium for use in this equipment operated at power frequencies. There arc two general grades of askarJl fluids; transformer 783446 GENP 010635 II 68 askarel (Arodor 1242, 1254 or 1260 and chlorinated benzenes) and capacitor askarel (Aroclor 1016). Transformer askarels were first introduced in 1932 by the General Electric Company and marketed under their trade name Pyranol, which was essentially a blend of Aroclor 1260 (60Z) and trichlorobenzenes (40Z). Westinghouse and I Ferranti Packard also later Introduced their own askarels,, Inerteen and Chlorextol, which were of the same composition. In the early 1970's, Aroclor 1254 was substituted for Aroclor 1260 In most askarel mixtures. Also around this time, the compositions of Inerteen and Chlorextol were changed to a 70Z PCB : 301 chlorobenzene mixture while Pyranol retained the same basic com position. In the late 1940's most transformer manufacturing companLesbegan adding what are termed scavengers to their askarels. The first scavenger used was tetraphenyl tin which was added at a concentration of ca. 0.125Z. However, at very low temperatures it had a tendency to crystallize out of solution and, | float on top of the askarel. Because of this solubility problem, Vestlnghouse later Introduced phenoxy-propene oxide as a substitute for tetraphenyl tin. These scavengers were added to transformer aslcarels to trap gases that are given off when arcing occurs in a transformer.| The major gas formed during arcing is HC1 gas, which is very corrosive to the internal components of a transformer. The scavengers react with this gas to neutralize it and pre- I vent any corrosion Inside the transformer. ] 4.3.2 Dielectric Strength I The moat important characteristic of any askarel, from the stand-point of transformer operation, is its dielectric strength. lie dielectric strength i is a direct measure of the askarel*s ability to withstand electrical stress GENP01063i 783447 69 without failure. New askarel has a minimum dielectric strength of 35 KV and a maximum water content of 30 ppm. Dissolved water has a significant effect on the insulating capacity of askarel and other transformer oils. The graph in Figure 7 shows the effect of water content, in askarel and mineral oil, on the dielectric strength. Figure 7 : Dielectric Strength (KV) vs. Veter Content (H^O, ppm). Transformer manufacturers advise that the water level in an askarel should always be lower than 70 ppm to Insure optimum insulation. They also advise that If the dielectric strength drops below 22 KV, and water Is pre sent, It should be removed. The dielectric strength Is also s good Indicator of contamination by particulate matter, or dissolved contaminants such as decomposition products resulting from arcing or abnormal operating conditions. For example, a dark askarel having a dielectric strength less than 22 KV moat likely contains a 783448 /cooin riKi^rn I 70 large amount of particulate carbon due to repeated! arcing 4.4 Transformer Tank The winding and core of a transformer are usually enclosed in a rein- forced steel tank that is permanently sealed except for the service and main tensnee ports. These tanks are very heavily constructed1 to minimize vibration ii I and noise. Host power transformers are constructed in such a way that external cooling radiators can be added. These radiators are usually forced-air cooled i| by fans and when they are installed on self-cooled transformers, they allow i i such a unit to be operated at 115 to 1302 of the self-cooled kva rating (see Figure 8). i i 5 OPERATING TEMPERATURE | I Transformers possess self Induction as veil as mutual Induction between individual windings on the core. These characteristics, along with the fact || that any electric current flowing in a coll of wire Is subject to resistance, will cause a transformer to produce heat when operating] Transformers are usually constructed to operate safely at a maximal tempirature rise of 65*C above ambient temperature. The ambient temperature Itself should not exceed 40*C so that the maximum operating temperature of|any transformer Is usually less than 105*C. 783449 s I o POoN oo GENF 010639 e ri References | i X. V. Del Toro, Principles of Electrical Engineering, Prentice Hall Inc., i New Jersey, 1965. | 2, V, Gourlshankar and D.H. Kelly, Electromechanical Energy Conversion, Intext Educational Publishers, Hew York, 1i973. I 3. C.S. Siskind, Electrical Machines - Direct and Alternating Current, McGraw-Hill, Toronto, 1959. I 4. R.L. Shrader, Electrical Fundamentals for Technicians, McGraw-Hill, Toronto, 1969. 5. R. Stein and V.T. Hunt, Static Electromagnetic Devices, Allyn and Bacon, Boston, 1963. 6. G. Wilcox and C.A. Husselbirth, Electricity for Engineering Technology; Allyn and Bacon, Boston, 1970. 7. H.tf. Jackson, Introduction to Electric Circuit's, PrejnticeHall, Hew Jersey, 1959. nATr> 783451 73 APPENDIX B: TERMS OF REFERENCE CHLORINATED DIBENZOFURANS AND DIBENZO-P-DIOXINS: DETECTION AND QUANTITATION IN ELECTRICAL EQUIPMENT AND THEIR FORMATION DURING THE INCINErL t ION OF PCBs, PROJECT PROPOSAL This project is a combination of two related projects concerning the detection and quantitation of the potentially toxic polychlorinated dibenzofurans (PCDF) and dibenzo--dioxins (PCDD). The tvo primary topics will focus on; 1) the detection and quantitation of residual PCDF and PCDD in new and used PCB-based transformer oils and 2) the detection and quantitation of PCDF (and PCDD) formed during the incineration of PCB. The approaches to the tvo related projects will be considered together in an Integrated fashion in phases. Phase 1 The first phase will involve a detailed background description of PCDF and PCDD with emphasis on the following proposal-related topics. 1) A brief discussion of the chemical piysical and spectral properties of the known PCDF and PCDD isomers. 2) A brief susmary of the biological and toxicological properties with an emphasis on those l6omer(s) which are toxic. 3) A review of the literature concerning the presence and formation of PCDF and PCDD in commercial PCB-containing electrical fluids and their formation during the Incineration of PCB, chlorinated organics and waste material. 783452 TWOTO4KTirn b 74 Phase II In discussions with Ontario Hydro personnel, arrangements will be made to sample PCB-containing electrical flu:.ds (transformers and cap acitors). The study will attempt to select fluids with respect to a number of possible variables which include; 1) the type of PCB-containing fluid (i.e. PCB/chlorinated benzene ratio), 2) the length of time In service (i.e. 5, 10, 15 and 25 years), 3) the power rating, I 4) the load, 5) the .length of time in service since maintenance has been carried out 6) fluids from transformers which have arced or discharged while in service 7) the effects of transformer KVA/Askarel Fluid volume. The major aim of this study is to ascertain those conditions which I |change the basic composition of the PCB-containing fluid particularly with respect to the formation of potentially toxic PCDF impurities. This data will be evaluated along with the physical properties of the fluids (i.e. dielectric strength, moisture content, etc.). The sampling process will attempt to obtain and analyse a sufficient numbir of PCB fluids so that the potential differences between the seven variables can be ascertained. All of the above samples will be split and stored in the dark at 0* prior to use. 783453 rwn torTM'qrn 75 Phase III The Isolation and quantification of PCDF and FCDD as minor constituents in PCB fluids involves a number of important analytical procedures which will be carried out as described. Since the PCDF are i| the most likely contaminants present the methods|vill primarily refer to their analysis. 1) PCDF Isolation and Clean-up. The Isolation of PCDF from the PCB- icontaining electrical fluid will be carried out by the methods 12 described in the literature. * The PCB sample (Ig) will be chromatographed on activated Florlsll (20 g, 80/100 mesh) and the PCB eluted with hexane (200 ml), hexane:acetone (95:5, 50 ml) and finally acetone (100 al). The latter fraction contains 90Z of the PCDF free from PCB impurities. i 2) Identification and Quantitation of PCDF, The Isolated PCDF II samples will be analyzed by GC and GC-MS to determine the range of molecular species present in the diverse electrical fluids. GC vili be carried out on a Hewlett-Packard 5710, chromatograph using electron capture detection with a 6* glass column picked with Ultrabond (RFR Corp) or Aue packing. This technique has been Ii shown to give excellent separation of Isomers Iiand rIesults in I minimum column bleed on to the mass spectrometer. A VG Micromass I 7070 mass spectrometer and data system locateti at the University of Guelph will be used on a contract basis. Since there are 75 possible PCDF isomers, most of which cannot or have not been made, the task of quantitation of the mixture 783454 (jrJtliNJK U 1U 64 % 76 I i 2I becomes prohibitive . Moreover quantitation of the highly toxic 2,3,7,8-tetrachlorodibenzofuran (TCDF) is also difficult since there are 2 2 possible tetrachlorodibenzofuran isomers, howevjer, TCDF quantit ; Iation will be made by the GC method. Quantitation will also be attempted by perchlorinating the mixture to give octachlorodibenzofuran as the sole reaction product. The procedure will be carried out as described II 3 by Hutzinger, Safe and Zitko. The method uses the BMC reagent (sulfuryl chloride ( 1 ), sulfur monochloride (5 g), and aluminum chloride (2.5 g)) which on heating converts the mixture into the single octachloro product. Preliminary experiments will be carried out to determine the optimum conditions for this analysis. In addition, the presence of other possible products present in PCB-contalning fluids (e.g. polychlorinated terphenyls, organotln preservatives) will also be investigated using GC and GC-MS. Phase IV The incineration and accidental combustion of PCB are known to produce PCDF by-products which depend on the incineration conditions. 4 II Recent work by Suser and colleagues have examined the variable temper ature pyrolysis of PCB isomers and commercial PCB (Aridor 1254). The results Indicate that below 700# the pyrolysis producia contain measurable quantities of dibensofuran products whereas above 700j the PCB was destroyed (>99.9X decomposition) with the levels of PCDF Isomers being aon-datactable. i <TKraWOrO rr\ 783455 77 The experiments which will be carried out will attempt to simulate conditions which would occur during a transformer fire. Several (3-4) i representative PCB-containing askarel fluids (Including fluid from an arced transformer) which have been analysed during Phase II of this project will be subjected to Incineration at a range of temperatures which might be expected during a transformer fire (200*-500#C). In addition the Incineration will be carried out using low oxygen conditions and ambient air levels* The off-gases will be trapped In a glycol solvent and the particulate collected in a glass fibre Alter. The incineration unit is currently available in Dr. Hutzinger's laboratory and the PCDF analysis will be carried out as described In Phase III. The organic extracts of the soot can be obtained by repeated soxhlet extraction with methylene chloride. Extraction of the glycol solvent with hexane will remove the PCDF/FCB gaseous material. Thus the simulated environmental burning will evaluate the formation of the potentially toxic PCDF. Capability | Wellington Science Associates has the analytical and chemical expertise to carry out these projects. Our personnel have access to I PCDF and PCDD isomeric standards, and are well versed and experienced in the analytical methodology which will be required. In addition Dr. Hutzlnger has direct experience In the Incineration/trapping/ analysis experiments. Listed below are research and development projects on PCDF and PCDD which have been carried out by Wellington Science personnel* GENP 01064i> 783456 4 s- 78 1 Chlorodibenzo--dioxins And chlorodibenzofurans are trace components of fly ash and flue gas of some municipal incinerators in the Netherlands. Chemos. 455-460 (1977) 2. Hass and Ion Kinetic Energy Spectra of Some Chlorinated Dlbenzo*--dioxins. I Anal. Chem., 47, 327-329 (1975). 3. Exhaustive Chlorination as a Technique in the Analysis of Aromatic Hydrocarbons. J. Assoc. Offlc. Anal. Chem. 36, 982-986 (1973). 4. Photochemical Degradation of Di- and Octachlorodlbenzofuran. Environ. Health Persp., 5 , 267-271 (1973). | 5. Analysis of Chlorinated Aromatic Hydrocarbons by Exhaustive Chlorination. Quantitative and Structural Aspects of Perchloroderivatives of Biphenyl, Naphthalene, Terphenyl, Dlbenzofuran, Dibenzodloxin and DDE. Environ. Anal. Chem., 2 t 95-106 (1972). 6. i An Assessment of the Effects of Enzyme Inducers on Aryl Hydro carbon Hydroxylase Activity. Bas. Commun. Chem. Pathol, and Pharmacol., 18, 59-66 (1977). I I GENF 01Ub4 I I I 783457 I  79 References 1. J.A.C. Roach and I.H. Pomerantz, Bull. Environ. Concam. Toxicol. 12, 33d (19743. 2. M. Merita, J. fiakagawa, K. Akiyama, S. Mimuraand, N. Isono, Bull. Environ. Contam. Toxicol. 67 (197d). 3. 0. Hutzinger, S. Safe and V. Zitko, Environ. Anal. Chem. _2, 95 (1972). 4. H.R. Buser, H.P. Bosshardt and C. Rappe, Chemos. 1, 109 (1978). 5. A. Poland, E. Glover, A.S. Rende, M. DeCamp and C.M. Giandomenico, Science 194, 627 (1976). UEMP 010647 783458