Document 2qz0enkBbbmregw7eNE3819L7

STRUCTURAL IDENTIFICATION OF POLYCHLORINATED BIPHENYLS IN COMMERCIAL MIXTURES BY GLC, NMR AND MASS SPECTROMETRY AUTHORS I- D. SISSONS & D. WELTI ! i i i j SUMMARY 1 The nxijor constituent polychlorinated biphenyls In Aroclor ! 1254 have been characterised hy high resolution Nuclear Magnetic I j . Resonance and Mass Spectronietry following separation by liquid-solid f. j and gas-liquid chroma1tography. The retention indices of these coir x^unds together with those of forty synthesised PCBs have been used to predict the structures of the regaining Aroclor 1254 constituents, together with all those of Aroclor 1242 and 1260. HONS 059835 INITODUCTIOH Itolychiorlnated biphenyls (PCBs) exhibit a high degree of biological and chemical stability. As they are lipid soluble they tend to accumulate In food chains and, since 1966. have been detected in a wide range of marine organisms, sea birds and even human adipose tissue1 -8 . Parti'cularly high levels have been found In rarlne samples . *7 taken from the Baltic and the Pacific coastal waters of USA , Quantitative analytical data are only approximate because of the larce number (210) of PCBs that are theoretlcally possible - up to 10 chlorine atoms may be substituted in the molecule as indicated belowi Quantitative results are estimated by relating the peak areas obtained from extracts of PCBs in an electron-capture (E.C.) GDC chromatogram to the corresponding peak area(s) produced by a known weight of a commercial FCB mixture or an insecticide such as DDE* Certain authors9-11 have shown that the pattern of GDC peaks produced by extracts of marine organisms are very similar to those produced by commercial FCB products containing 5or chlorine. While this is true for species Indigenous to rivers ar.d well defined sea-areas, eg. estuaries, the situation may be quite different in large sea arena such as the North Sea. The PCBs that are present in some species of sea fish have been shown in this laboratory to vary widely in type And quantity and bear little resemblance to any commercial product MONS 059686 - 3- In spite of a growing interest In the toxicology of FCBs, relatively little data has been published on this aspect to date. All available information relates to trials involving commercial products1^ although individual FCBs nay well have differing toxlcltles. Both from the analytical and toxicological viewpoints, therefore, there Is a need to think in terms of individual FCBs rather than complex mixtures. In order to progress to a greater understanding of the problems and significance of FCB contamination. As a first step in this direction we have used a combination of analytical techniques to characterise the component FCBs In Aroclors 1242, 1254, and 1250. (Commercial products manufactured by Monsanto Chemicals Ltd. containing 42$, 54$ and o0$ K/w chlorine respectively). Aroelor 1254 wes fractionated on a coluim of alumina and the major FCBs In the fractions were separated and collected from a gas-liquid chromatograph. The separated FCBs were characterised by trass spectrometry (MS) and high resolution nuclear magnetic resonance spectroscopy (NMR) to determine the number and positions of the chlorine atoms associated with each compound. NMF1 was chosen in preference to infra-red spectroscopy as it is a more definitive technique although the latter has the advantage of requiring a smaller sample size. The NMR 6pectra of chlorinated biphenyls which have been 14.15,16,17 reported pi-eviously were unhelpful in this instance. l8 Mayo and Goldstein have analysed biphenyl itself on 60 and TOO MHz spectrometers. MCN5 059007 FCBs characterised by NMR together with forty Individually synthesised FCBs were subsequently analysed by high resolution OLC to determine their absolute retention values using the Kovats 19 retention index (RI) system. Previous GLC data on FCBs have been restricted 20 to relative retention times . Attempts to derive accurate retention data based on Internal standards were unsuccessful due to variations In conditions during the lengthy GLC runs of up to 16 hours duration, and because of the severe limitations placed on the choice of standards due to the complex Aroclor chromatograms. The structures of the minor PCS constituents of Aroclor 1254 and all PCBs In Aroclors 1242 and 1260 have been predicted by a variation of the Interpolation method 21 suggested by Evans and Smith . EXPERIMENTAL 1. Fractionation of Aroclor 1254 on an alumina colunn. 60 mg of Aroclor 1254 was washed onto a 25 x 1 cm colur.si of 20 g activity grade I alkaline alumina (M Woelm Eschwege Germany) with flvo 1 ml volumes of hexane. All PCBs were eluted In 900 ml of hexane (flow rate of 1 ml/mln), collected in 25 ml fractions. This fractionation was repeated with a further 60 mg of Aroclor and the corresponding fractions were added together. After suitable dilution, usually 100 or 1000 fold, each fraction was analysed by EC GLC to determine whether further separation of PCB's was required. Analyses were performed on a Pye model 74 Instrument fitted with a nickel 63 E.C. detector and a 7 ft X V1* inch 'pyrex' column of 25I Apiezon 1. + O.tfja Eplkote resin on 80-100 mesh Varaport JO packing materiel. Of a wide variety of stationary phases that were tested, Apiezon 'L1 gave the highest resolution of PCBs In Aroclor 1254, Argon was used as carrier gas at an Inlet pressure of 25 pel and a flow of 80 ml/mln. The column ani Injector were maintained at a temperature of 2J0C and the detector at a temperature of 325C. HONS 059883 -5 - 2. Collection of GLC fractions for NM). Alumina coluim fractions were separated into their component FCB's on a T ft X ^/4 inch OLC coluim similar to that described above but fitted into a Pye 104 chromatograph adapted for fraction collecting. The effluent gas was split in the ratio of 25 parts to the collection outlet and 1 part to the flame-ionisation detector. The temperature was programmed from 220-230C at lC/min. and the carrier gas flow was 40 ml/min of argon measured at 220C. The outlet to the collection trap was maintained at 250C. The traps were interchangeable straight metal tubes loosely packed with glass wool. They were screwed onto a threaded outlet tube and maintained at a temperature of approximately 0C above ft tray of solid carbon dioxide. Trapping efficiencies of nearly 9&j were ootained with 246-2,4,6f hexachlorobiphenyl, but for other FCBs recoveries were often much lower. Trapped FCBs were eluted with light petroleum ether which was evaporated off and replaced with deutero-chloroform or tetrachloroethylene before NMR analyses. NMR. Spectra were obtained on a Varian HR 220 spec trometer, The temperature of the probe varied between ^l^C at the beginning cf a day to ~l?C over the greater part of the day. This temperature difference was shown to be insignificant in the interpretation of spectra as selected spectra obtained at 55C showed only slight shifts in oorne resonance linos. For most gas chromatographic fractions the sample size was snail and several of the spectra were very weak. In these cases, the spcctromctric scanning conditions resulted in the spectra being extremely noisy. HONS 059889 -6- 4, Combined QLC-Mass Spectrometry. Aroclor mixtures were separated using a Perkln-Elmer 50 ft X 0.02 In. l.d. support coated open tubular (SCOT) Apiezon L colunn under the conditions described In the next section. The colum effluent was split between the flame-lonisation detector and an AEI MS12 mass Bpeotrometer. Approximately 10J( of the colum effluent was taken Into the mass spectrometer ion source with no removal of the carrier gas. Peak homogeneity was monitored by repetitively scanning over the parent molecular ion region, observing an oscilloscope screen and recording several spectra during the elution of a peak. The higher chlorinated FCBs in Aroclor 1260, which had very long retention times on the SCOT column, were analysed on the ^ ft X V1* inch o.d. column described previously. The lsrger sample sizes and shorter retention times enabled a higher FOB concentration to be obtained in the carrier-gas stream. The column was connected to the MSI2 through a Watson-Biemann type fritted glass separator. The carrier gas flow rate was 40 ml/min of helium and the colum temperature wa3 maintained at 205C. ICBs fractionated from Aroclor 1251* and all the synthesised PCBs were examined either as effluents i from the '( ft colum or in the mass-spectrometer direct-inlet system. 5. High Resolution GLC and Retention Indices. The Aroclors and synthesised PCBs were analysed on the SCOT Apiezon 'L* column fitted In a lye model 104 . A flarce-ionlsatlon detector was necessary as the n-alkanes which were added to earn lee for the determination of retention indices produced very small and inverted peaks with an electron capture detector. Helium was used as the carrier gas at a flow of 2.5 ml/mln. The column was operated at 205C which was near the maximum temperature permitted (225C) for this pliAfic. *3110 operatinc temperature was dictated by the need 0 6 8 6 5 0 SNOW -7- to examine all the FCBs under conditions that allowed both the higher chlorinated FCBs in Aroclor 1260 to be eluted quickly and the lower chlorinated species ir. Aroclor 1242 to be resolved adequately. Some column deterioration was noted at 205C and significant variation* in the RIs of FCBs were evident after three weeks. The changes were alnr.nt entirely due to reduced retention volumes of the n-alkancs relative to those of the FCBs. To overcome this effect the RIs of well characterised ICBs were used as sub-standards, as the retention volu''en of FCBs wore constant relative to each other. Colunri efficicnces were constant at between 24,000 and 27,000 theoretical plates, depending on the FCB. Several pairs of peaks eg. 245~2,4I5* and ^,4-2t3,4* changed relative positions as the coluim temperature was increased between 1^0-210C. For this reason isothermal was preferred to temperature programmed gas chromatography Retention Indices of the later FCBs in Aroclor 12oQ were determined on the 7 ft X */4 inch Apiezon L coluim and were correlated graphically with the SCOT column RIs. RK3UI/VS 1. Fractionation on an alumina coluim Only twenty-three distinct peaks were apparent in the gas chromatogram of Aroclor 1254 (Fig.l) analysed on the 7 ft packed Apiezon !I.T column at 2J)0C. Similar GLC analyses of the hexane fractions obtained from the alumina colurm produced chromatograms containing a total of thirty-five peaks. This increase in the number of resolved peaks was due to the combination of the different elution patterns of the two systems. After characterisation of the FCBs by NMR, the order of elution of the substituted phenyl rings, assuming MCNS 059891 -8 - a constant second ring, was shown to be 245-, 25-, 2>45-, 24-, 2^6-, J4-, 2.5-, 2^4-, from the alumina column compared to 25-, 24-, 2>*, 2>5~, }4-, 245-, 2^4-, 2^5- from the GLC column. Three FCBs, corresponding to peaks 39* 44 and 56 In Pig.2, were sufficiently well resolved by alumina column fractionation to be analysed directly by MS and NMR. A further nineteen ICBs, Including all the remaining major peaks In the Aroclor 1254 chromatogram, required further separation by G.L.C. before MS and NMR analyses. 2. NMR and Mass Spectroscopy The structures of the major Aroclor 1254 FCBs that were identified by NMR and Mass spectrometry after their separation and collection from a gas chromatograph are given in Table I. Three further minor peaks are included in Table 2. The NMR chemical shifts for both these FCBs and the synthesised compounds givo a rational gradation of frequencies depending on the substitution patterns of both rings. Full details of the chemical shifts, the line splittings and structural assignments will be published separately22 . The spectra of all the compounds examined from Aroclor 1254 gave near fij'st order spectra. The 2^4- and 23S- substituted phenyls gave AD__ ^AX systems; the 24-, 25- and J4- substituted phenyls gave ABX--^AMX systems; and the 245- and 2.545- substituted phenyls gave two and one singlet(s) respectively. Under the spectrometer conditions used for running these weak spectra, and considering the 220 MHz spectrometer linewidth, the splitting due to para-coupling could not be observed. The majority of the structural assignments in Table I were obtained from NMR data alone - the MS and RI evidence confirmed the results. However, in one or two coses an interplay of all three techniques allowed firm assignments to be made. HONS 059892 -9- The fraction trapped as peak 24 was shown to be a mixture by OLC., The NMR spectrum of the mixture contained resonance lines compatible with the 2>s 24-- and 25- dichlorophenyl groups. A chromatogram of Aroclor 1254 containing added 24-2*4* tetrachlorobiphenyl showed that the retention time of the latter was not coincident with peak 24, although the trapped fraction could have included some 24-2*4' compound from the tail of the peak. The spectrum also suggested that the other alternative RI structure 5"2I4,6* tetrachlorobiphenyl was not present. Using the total NMR, MS and RI evidence, the most likely structure is 25-2*5* tetrachlorobiphenyl. The fraction trapped as peak 52 gave a very weak spectrum in . tetrachloroethylene which contained lines typical of a 256-trichloropheny] group. Other lines were too weak for certain correlation. However the limited evidence is in agreement with the predicted GLC/RI structure of 25-2,5*6* pentachloroblphenyl. Peaks 42 and 45 were always eluted closely to one another on every chromatogram. Two different fractionations of these peaks gave spectra which Included lines typical of a 245-trichlorophenyl group in addition to those of 25-2*5*4* pentachloroblphenyl. This agrees with a GLC/lU predicted structure of 25-2*4*5' pentachloroblphenyl for Peak 42, The fraction trapped as peak 68 was shown by a subsequent OLC separation to contain a large number of impurities and, in addition, the spectrum was very weak. However, after the GLC/RI values suggested that tho peak might be 245-2*5*4'5' heptachlorobiphenyl it was possible to predict the frequencies of the three singlets in the spectrum from the total FCB NMl data. The three lines were, in fact, present at the predicted frequencier. MONS 059893 - 10 - The mass spectrometric results obtained from the GC/fiS runs are quoted In Tables ], 2, 3 and 4 as the number of chlorine atone associated with each FCB peak* Prediction of PCB structures using K3s Retention Indices are a more reproducible form of presenting logarithmic retention volumes of GI peak maxima and are obtained with reference to the well defined retention values of a series of n-alkanes separated in the same chromatogram. The RI of a compound is directly proportional to its free energy of solution in e stationary phase, which in turn is an approximately additive function of the groups constituting the molecule* Consequently any PCB molecule can be thought of as consisting of two chloro-substituted phenyl groups each with its own (RI) value. All FCBs are composed of twenty such basic groups and the RI of any FCB con therefore be estimated by adding together the i(Rl) values of the two component phenyl groups. The (Kl) values of all substituted phenyl groups containing up to three chlorine atoms and the RIs of the FCBs from which they were derived are shown in table 5* Certain discrepancies are apparent between the (RI)s for the same ring structures calculated from different PCBs. These discrepancies are due to electronic and steric effects which vary according to the substitution patterns in the two rings under consideration. Electronic effects are induced by the introduction of chlorine atoms into the phenyl ring and vary according to the exact position of each substitution. The inclusion of chlorine atoms results in a change in the electronic distribution in each ring and varies the character of the phenyl-phenyl bond. The nett effect is illustrate*! in table 5 by the increased RI (2 to 7 units) for molecule substituted in both rings compared to those substituted in only cne ring. MOWS 059894 -u The steric effect, previously noted in methyl biphenyls, is caused by substitution of chlorine atoms in the ortho (2,2*,6, and 6*) positions in the rings. Increasing substitution in these positions progressively restricts free rotation of the phenyl groups and is accompanied by a reduction in RI. This effect is illustrated in table 5 by the smaller RI for 2b- and 246- rings (-2 and -8 units respectively) derived from molecules containing two ortho chlorine compared to no ortho chlorine in the second ring. Any electronic and stcric effects.must be allowed for when predicting the RIs of FCBs, by using appropriate j(RI) values derived from a similar type of PCB molecule. Using the data in table 5, the experimental and predicted RTs for a further range of PCBs of known structure are compared in table 6. With only one exception, for which no obvious explanation can be given, agreement is to within 4 R1 units Vfhich is consistent with our experimental error of *2 units in determining RIs. Therefore structural predictions of FCBe in Aroclors 1242 and 1254 could be made with confidence. Very few standard PCBs containing more than three chlorine atoms per ring were available for analysis but, from the limited data obtained (table 7)* it was evident that electronic and steric effects had a much greater influence on their RIs than on those of the lower chlorinated PCBs. This resulted in increased difficulty In predicting the RIs of some of the hexa, and the hepta, octa and nona-substiluted molecules. However the relation between structure and RI was apparent when msst of the later peaks in the chromatogram of Aroclor 12.60 were studied. After allowance was made for electronic and steric effects, particularly in the case of the 2^40- and 2,550- rings Joined to rings containing two, one or no ^rtho* chlorine atoms, the more obvious correlations for the octa- and nona-substituted compounds could be made, and th'-n MGNS 0 5 9 8 9 5 the remaining peak aasl&iments fitted into place. In this way using the inferred values given in table 8 it was possible to allocate structures to all appropriate peaks in the Aroclor 1260 chromatogram. The values in table 8 are conditional on the presenoe of at least two substi tuted chlorine atoms in the second phenyl ring. This condition was Imposed because it was not clear what effect a greater imbalance in the number of chlorine atoms between the rings would have on RIs. In practice, the MS data giving the number of chlorine atoms present in each PCB enabled potential errors due to the above effects to be greatly reduced. . .Typical chromatograms of Aroclors 1242, 1254 and 1260 obtained from the SCOT colurm are shov/n in figs.2-4. Tables 1,2,5 and 4 list the RIs and chlorine contents of all PCB peaks present In the Aroclors. The tables also Include the alternative PCB structures that have RIs predicted to be within an arbitrary 5 of each peak. The RIs of synthesised FCBs are used where applicable, and those that were derived from the inferred }(RI)values given in table 8 are given in brackets in tables 1-4. PCB molecules having a difference of more than two chlorine atoms between rings arc not included as these would be unlikely to occur in Aroclor mixtures. For the sake of emphasis che major peaks in the chromatograms are presented first, in tables 5 and 4, in order of increasing RI. . Table 8 Inferred Half RI values for four and five substituted phenyl rings as affeoted by 'ortho1 substitution in the second ring MQNS 0 5 9 8 9 6 Substituted phenyl ring 25452>6- ^(RI) values Two ortho One ortho 1353 1363 1245 1250 No ortho 1375 1260 2355- 1219 .1228 1238 25456- 1405 1415 1428 lTie changes In (Rl) values found in this investigation are In line with changes in the .NMR chemical shift data. When considering the protons ir one ring of the molecule* the chemical shifts arc affected by the overall presence of chlorine atoms in either ring, the substitution position of chlorine atoru(s) In the ring being examined* and the number and relative location of tho ortho chlorine atoms causing 6tcric hindrance in the molecule. Allowing for variations in detector response* the SCOT column chromatograms Indicate that the predominant 3j.eciea in Aroclor 1254 are pcrttachloroblphenyls with smaller amounts of tetra and hexachlorobiphenyls. This agrees well with the chlorine content of this Aroclcr which corresponds to a mean value of 5 atoms per molecule, Ihe most abundant chlorine substitutions are those in the 25-* 34-, 224-, 23o-, 2U5- and 2345- positions. When alternative predicted structures are available, therefore, preference should probably be given to those containing these substitution patterns. It is noticeable that certain substitutions, such as 24b-,rarely occur and others,such as 2-, 35- and 23-. occur only infrequently. Aroclor 1242 contains predominately trlchlorobiphenyls with a lesser amount of dichloro and tetrachloro-compounds. The mean chlorine value for this Aroclor is 2.1 r-toms per molecule. The most frequent substitutions occur In the 2-, 24- and 25- positions. The predominant compounds in Aroclor 12o0 were hexachlorobiphenyls with a significant level of heptachlorobiphenyl, The mean chlorine value of this Aroclor lb 6,1 atoms/inolecule. The most frequent substitution patterns were similar to those found in Aroclor 1254, although 235-, 2346- and 235*>- structures were in greater evidence. MGN5 059897 A'lKNOWLEDGEMENTS . The authors are indebted to Dr. 0. Hutzinger (National Research Council of Canada. Halifax. Nova Scotia). Mr. R. Lid&ett (Monsanto Chemicals Ltd. Ruabon). Dr. D. Osborne (URL Colworth/Welwyn) and Mr. R.II. de Vos (T.N.O. Zeist, The Netherlands) who supplied the synthesised FCBs used throughout this investigation; to Ir, P.E.J, Verwlel and Mr. L.J. Kogendorn of the.Central Laboratory, T.N.O. Delft, The Netherlands, for their cooperation in obtaining the 220 MHz spectra; to Dr. T. Bryce and Dr. W. Kelly for cooperation in obtaining the nass spectra; and to Mr. G.M. Telling for the interest and enthusiasm he has shown in this work. . HONS 059898 - 15 - REFERENCES 1. S. JENSEN. New Sol. 32 (1966) o12, . 2. D.C. HOLMES, J.H. SIWAONS, J.O'O. TATOON. Nature 216 (1967) 227. 3. A.V. HOLDEN and K. MARSDEN. Nature 216 (1967) 1274. k. R.W. RISEBROUGH, P. R1ECHE, D.B. FRAKALL, S.G, HERl'lAN and M.JI. KIHVEN. Nature 220 (1968) 1098. 5. J.H. KOEMAN, M.C. TEN NOEVER DE BRAUW and R.H. DE VOS, Nature 221 (1969) 1126. 6. S. JENSEN, A.O. JOHNEIS, M. OLSSON and 0. OTTERLIND. Nature 224 (1969) 247. T. R.W. RISEBROUGH. Chemical Fallout. First Rochester Conference on . Toxicity 19*59* 8. ?.J. BIROS, A.C. WALKER and A. MEDHURY. Bull. Environ. Centum. and Toxicol. 5 (1970) 317. 9. I. PRESTO, D.J. JEFFERIES and N.W. MOORE. Environ. Pollut 1 (1970) 15. 10. J.G. VOS, J.H. KOEMAN, H.L. VAN DER MAAS, M.C. TEN NOEVER DE BRAl.W and R.H. DE VOS. Food Cosmet. Toxicol. 8 (1970) 625. 11. T.W. DUKE, J.I. LOWE and A.J. VIILSON. Bull. Environ. Contain and Toxicol. 5 ('970) 171. 12. D.J. SISSONS, unpublished work. 13* D-B. FFJVKAtt, and J.I,. LINCER. Bioscience 20 (19J0) 95b, 14. S. BROWNSTEIN. J. Amcr. Chem. Soo. 80 (1958) 2J00. 15- D.M. GRANT, R.C. HIRST and H.S. GUTOWSKY. J. Chem. Phys. 38 (1963) 470. 16. R.J. KURLAND and W.B. WISE. J. Amer. Chem. Soc. 86 (196*1) 1877. 17. Y. NOMURA and Y. TAKEUCHI. J. Chem. Soc. (B) (1970) 956. 18. R.E. MAYO and J.M. GOLDSTEIN. Molecular Riysics 10 (1966) 301. / 19. E. SZ KOVATS. Advances in Chromatography Vol.I, ed. J.C. OIDDINGS ' and R.A. KELLER, publ. Edward Arnold (London)/Marcel Dekker (New York), 19o7, Ch.7. p.229. MONS 059899 20. H. WEINGAKTBN, W.D. BOSS, J.M. SCHIATER and 0. WHEELER. Anal. Chem Acta. 20 (1962) 391. 21. M.B. EVANS and J.P. SMITH. J. Chromatog. 5 (1961) J00. 22. D. SISSONS and D. WELTI. To be published. 2). G.II. EEAVEN, A.T. JAMES and E.A. JOHNSON. Nature 179 (1957) 490. 24. J. HOIOnC H. WIDMEB and T. OAUNANN. J. Chromatog. 11 (19j3) 459. HONS C599G0 FIGURE LEGENDS Fig.l. 2 ng of Aroclor 1254 chromatographed on t. 7 ft Apiezon L packed coluim at 230C. Attenuation setting maintained at 2 x 102 - electron capture detector. Fig.2. 0.2 pi of 5$ Aroclor 125** solution chromatographed on an Apiezon L SCOT column at 205C. Attenuation setting o maintained at 1 x 10 - flame ionisation detector. FigO* 0.2 pi of Aroclor 1242 solution chromatographed on an Apiezon L SCOT coluim at 205C. Initial attenuation setting 2 of 5 * *0 reduced as indicated - flame Ionisation detector. Fig. 4. 0.25 pi 1 Off Aroclor 1260 solution chrona tographed on an Apiezon L SCOT colurrer at 205C. Attenuation setting 2 maintained at 1 X 10 - flame ionisation detector. HONS 059901 Table 1 Tfre Retention Indices, Chlorine numbers, and structures of the major FCB constituents In Aroclor 1251* Neak no: 22 23 2k 29 32 33 36 39 Ul *42 43 I4I4 45 48 50 52 55 56 59 n.i. 1994 2010 2022 2069 2119 2136 2159 2175 2191 2203 2207 2228 2238 Chlorine noi 4 4 4 (5) 5 5 4 5 5 (6) 5 5 5 5 - 2264 2299 2321 2356 2356 2390 6 6 5 5 6 6 (7) NMR determined structure 25 - 2'5' 23 - 2'5' 25 - 2'3,6' 23 - 2'3,6' 25 - yk' . 25 - 2'4'5' 24 - 2'4,5' 23 - 2'4'5i 25 - 2'3,4' 34 - 2'3'6' 230 - 2'4'5' 234 - 2'3,6' 34 - 2'4 '5 ' 34 - 2'3'41 245 - 2'4'5' 234 - 2 'll- '51 Alternative predictions structure R.I. 2 - 2,3,5' 2 - 2'4'5' 24 - 2'5'. 3 - 2'4'6' 24 - 2'4' . 26 - 2'3'6' 26 - 2'3'5I 1996 2012 2010 2021 2027 2017 2092 25 - 2,3'5' 24 - 2'3'5' 35 - 2,4'6' 26 - 2'346' 230 - 2,3'6' 23 - 213'5 2159 2175 2175 (2175) 2172 2189 24 - 2,3'4' 23 - 2'3'4' 25 - 3'4,5' 34 - 3*5' 235 - 2'4'6' 24 - 2,3,4,6' 2226 2240 2240 2238 2240 (2253) 25 - 2'3'4'5I 35 - 2'3'4,6' 23 - 2,3,4'5' 245 - 3*4'5' 235 - 2'.3'.5'.6' (2360) (2357) (2390 2368 (2390) (HI) Inferred prediction using: table 8 (Cl Not) chlorine number of minor constituent not associated with accurate RI, HONS 059902 MENS 0 5 9 9 0 3 >- noi 1 2 3 k 5 6 7 8 9 10 11 1? 13 14 15 )5 17 16 19 20 21 25 26 27 28 30 31 > 35 37 38 ko 46 k7 49 0. p.r. 1490 1579 1672 1638 1750 1758 1763 177k 1782 . '833 1849 1863 1879 1893 1922 1935 1043 1952 1963 1966 i960 2040 2051 2058 2072 2097 2115 2146 2152 2l6k 2169 2186 2246 2254 2283 CMerln# no* 0. 1 2 7 2 _ 2 3 2 3 3 3 3 4 3 3 3 3 k k k k . 5 k - (not 6) - - - - - S'.r.ctxt Biphenyl 22-2* 4- . 2-5* 24 - . 2-y 2 - 2*6* 2 - k 2 - 2'5 2 - 2'fc* 2 - 2*3' 3 * 2'6 4 - 2*6* k - 4' 2 - 2,k6* 2 - 3*5' 2 - 2*3*6* 3 - 2'5' - 2'5 3 - a'1*1 3 - 2'y k - 2'k* 3 - 2'3' k - 2*k* 2 - 3*4' 4 - z'yv 23 - 2*4 * 2 - 2'3'k' 23 - 2'3' 23 - 2*3* 26 - 3*4 ` 3 - yyy 2k - 2*4*6* 25 - 3'5* 26 - 2*3'5* 23 - 2*3*6* 23 - 2*4*6* 2 - 3'4,5' 3 - 2'3'k' 3 - 2*3*4 * 2k - 3*U' 23 - 3'k' 2>5 - a'k'6' 23 - 3*k* 236 - 2*3*6' 23 - e's's1 35 - 2*3'6* 3 - 2'4*5* 2k - i'yw 25 - 2*3'4*6* 235 - aW* 4 - 3*4*5.' 23 - 2*3,5,6' 35 - 2'3'5' 245 - 2*4'o' 23 - 2*3*4 *5* 26 - 2'3'k *5* 3k 3k* U91 * 15TT l66o + 166? 175k 1756 1770 1765 1782 ie3i 1 ew 1664 1666 1676 109k 192k * 1932 1932 1932 19kk 19k9 1963 1962 + 1963 1962 + 1975 2041 2041 2047 2055 2055 2071 2097 2097 2094 2092 2120 2112 2lU 2146 2146 2154 2166 2160 2168 2173 2163 2185 2245 (2241) (2247) 2248 2257 (2256) 2259 22*?6 (278) (22?-3) sate * * *ot 51 R.I. 2310 53 2335 5k 2340 57 2372 58 2376 60 2400 61 2413 62 242$ 63 2433 6k 2451 65 2466 66 . 2489 67 2519 68 2543 69 2577 Chlorine rat 5 (7)' 5, (7)4 6 - 6 6 7 6 7 7 7 i 6 7 T Simctv* 25 3*4*5' 35 2*>*4* 236 - 2'5'5'6' - 3'4,5' 236 - 2*3'4'6* 35 - 2'3V6' 235 - 2*4*5f 2k 2> - 2*3*4*5' 2*;'5* > - 2'3,5*6' - 2*;,y6* 2> >* i'i'y :*3,4*6* >5 - 2P3'6* 05 - 2f''4'61 - 2*3'4* 236 * 2 *3*4'5* 245 - 2*3,*6 <*0 - >*?'4'5* - 2*3*5*6' 234 - 2'3*4 *6* 245 - 3'4*5' > - 2'3*1* *51 245 - 2*3*4 '5 1 >5 - 2*3`5'6* 234 * 2?ut5i Predicted R.I 2307 231C 2303 23^6 (2J29) (2335) 2378 2JT4 (2379) 2374 (24C1) 2396 (2412) saw (2439) (2426) fi-jc) (450) (2462) ?.TJ> (541) * R.I. of synthesised ccrpound (RI) Predicted R.i. using Tfcoi# 8 s sceller peeks eluted earlier then mjor certs c: aks 51 and 53 respective r*3L > tv fte'.t:Ion lai'.eM. ChUrLi* *Aib*r *.-i yredirr-d Structures of Areclcr '2k2 Pri< sc;: R.I. Chlorine not Structure Precleted R.l 3 5 3 Q !C 11 12 *3 15 IT >8 19 .---- 20 21 22 __ 2k 31 > 35 ><572 1705 J73i >ej> i8k? 1603 18T9 1696 1935 1951 1961 1966 1=60 1996 2011 202* 2092 2138 211*3 2 2 2 3 3 3 3 2 3 3 3 3 3 k k k 3 u (5) k k 2-2' 2-3* 2 - k* 2 - 2*5' 2 - 2*k' 2 - 2'3* 3 2'j' k - 2'6' k - k* 2 - 3V 3 * 2*5' 3 - 2'k' 3 - 2'3' k - 2k* ? 3 - 2*3' * k - 2V 2 - 3*k* a - a'3'5' 25 - a'5' * 2 ,, a'kV 2k - 2'5' 3 - a'3'S' 3 - a'*'*' k - 3k* 25 - 3*5' 2k - 2k*6* 25 - 2'3'6' * 25 - y.k*. . * a - 3">'5' 3 - a's'k' . 1&9 * T7C J7t2 18*1 i8k8 I8ak l&TO 1376 189k + 193B 1532 I9k9 1963 1902 * 1903 1902 * >9T5 199* 199k * 2012 2010 2029 2021 2068 209k . 2097 2069 2138 21 kk 21k8 MONS 0 5 9 9 0 4 lk90 '580 1750 1758 177k 1923 19kk 2015 2CC8 203T 20k 1 2052 0 Blpftenyl - 2- 2 25 - 2k 3 2 - 2*6' k 2 - 2*k'6' 25 - 2*6' 3 k - 2*5-' k 2 - 2,k,5f k 3 - 2'k'6* 23 - 2'5' 2k - 2*k* 26 - 3'5` k - 2k'6 k ( k - 2'3,S' k ( 23-2'k' k 2 - 2'3'k' 23 - 2'3* lk9J + 1577 ITkk 175o 1765 1?2k 1927 !9kk 2012 2029 2025 2087 * 2028 2033 * 20ki 2^kl 20k7 2055 *s*k not 0 M k2 *3 *5 HI. 206l 2C60 2121 2129 21o7` 217k 2178 2195 2207 22(2 2232 22k2 aao3 2267 C&ierlne net * 3 k k k k - 5 5 5 5 0 - Structure r 3 - 3*kf 23 - yy k . 2k'5 23 - 3*5* k - ah*5* a - 2*3^' . 23 - 3'k' a - a'3'5' 25 - 2*k*5' 26 - a'yh^* 35 - 2*>'5f 35 - 2V6* 2k - 2*k'5* * 23 - 2'k'5` 25 - a^'*' ' 25 -S'}1** * 2k - 2*3'k* 3 - a'3'6' * 2k - 2'3*5'o' as - a'3V<>' 235 - 2'kf6' 2k - 2'yk'6* 216 - 2 *k *51 * Jo . a'3'l>'5' > - 3'k' 93k - 2,k,6' Predicted R.r 2C75 2125 2125 2125 2125 2I6C ' 2)66 2175 2175 (airs) 2183 2175 2191 2206 2210 .J'O ,12} . aaae (J2H) (afT! 22k0 (22S3) 22ok (2283) 2282 2291 * Structures fcuraj in Arcelor 125k by NWt determination*. RIs of synthesised compound or NKR standard. ' (RI) predicted RIs using Table 8. (Cl No.) Chlorine nuaber of minor constituent not associated with accurst* RX. The Retention Indies. Chlorine eusbers ar.g r--11t t-d g- ctures cf Arcelor 129g o** Peak toi 8.1. Chlorine not Structure 31 ?ce6 5 25 - 2*3*6' * 2C8? 25 - 2'*,6' 2Obi 37 2168 6 236 - 2*3*o* 2172 38 217k 5 2k - 2'3*5* 25 - 2'V5* * 2175 2175 * *3 2239 6 >5 - 2'Wa' 2k - 2*3'5'6* 2175 (22*1) 235 - 2''*,6' 22*0 *5 5 2k - i'j'lt'o1 {a*)) 236 - 2'k`5' * 22ok * *7 2296 *8 2>X> 6 T 23k a'j'Q* 2J6 - 2299 (2303) 52 2358 6 2*5 - 2,3*5*6' 25 - 2*3'**5* (2295) (2360) 35 - 2*>'**6* (2357) . 2k5 - 2rk>5* 2356 * 55 2390 6 23 - 2*3'**5* (2391) 25k 2*4'5* 239 * >5 - a'k'S' 2388 57 2kt 1 <T) 7 235 - 2,3V6* 235 - 2*3'*,5' (2390) (2*12) al5 - a'j's'S' (2*06) 58 2*28 6 25k - 2`3'k* * 2*25 + 7 2*5 * 2,3'k,o (akaB) 59 2k32 7 2*5 - 2'3,k6* (2*26) 60 2**5 8 7 2*6 - Z'j'kV 2556 - a'j'j'i' 23* - 2'yyz* (2*3) 2*38 * (2**0) 66 25k2 8 2575 7 7 236 - 2k5 - a'J'k'S' 2>* - 2*3** *5 * (2V59) (asu) (2575) >5 - 2'j'*'6' (25TO) 76 272k 8 23*5 - 2*3** *5* (27a6) Is Aroelor 125* chrosatogra* were also present In Aroclor 1260 22 199k . * 23 2009 k 2013 2022 ) 202k ) - 2030 - 20 20*0 29 2050 JO ?os8 J2 21 >7 * * * 2 - 2*3*5 * 25 - 2'5' a - a'k's' 2k - 2*5* 2 - 2'*V 2k - 2'5* 3 - 2***6* 23 - 2'5* 2* - 2**' 3 - 3*5* > - a'3'6' * - a'k'S* 2k - 2** 26 * 3*5' k - 2*3'6# 23 - 2'** 2 - 2*3k 23 - 2*3* 23 - 2*3* 3 - 2'-*'5* 23 - 2'3'6' 1996 199* * 2012 2010 2012 2010 coei 2025 2027 2033 2029 20 * 2027 2026 20*1 20*: 0*7 2055 2C^5 ?t ik 21 :k Peak net 33 > 35 36 39 *0 *2 . k6 *9 50 51 53 5k 55 61 62 3 6* 65 67 69 70 71 72 . 73 7k T5 a *.I. 2*25 2135 21*o 2158 2192 SC5 ) 2209 ) 2229 225* 2263 2321 2320 23*0 2772 2379 2*02 2*60 2*6* 2*80 2518 2522 550 2578 2587 2597 2605 2616 - 2733 770 ~ Chlorine m> _ * * 5 - 5 - 5 7 6 - - 6 7 8 8 6 5 T 6 8 8 8 9 8 9 9 10 Structure * - 2***5* 23 - >'5* 25 - 3*** * 2 - 3***5* 3 - 2,3*** 25 - 2*3*5* 23 - 2*3*5' 2k - 2**V i> - 2,*'5* 25 - 2*3*** * 2* - 2*3*k* > - 2*3*6* 23 - 2*>*5'6* *5 - 2*k`6* 23 - 2*3'fc*6* a6 - a'7'V 3* * 3**' 2* - y*'5* 3* - 2***5* 236 - 2'3,*,5' ako - a'j'k'o' 35 - 2*3'5*6* 235 - 2*k *5* ak-a'j'k'S1 23* - 2'3'5* 24 - 2*3'*,5* 3* - 2*3V6* 23* - 2*3*5* >. - 2*3*k'6* 23* - 2*3***6* 23*5 - 2*3*5*6* 36 - 2*3'**5*6* 2*6 - 2*3'*,5,6I 2?*6 - 2*3'%'6* 2`5 - )'k'5' > - 2*3**v * 235 - 2,3*k*5, 3*5 - 2*3*5*6* 235 - 2*3***5'6' 23*5 - z'yyv 23*5 - 2*3*5 '6* 2*5 - 2*3'*'5*6' 23*5 - 2,3**,6' 2756 - 23* - 2'3**,5*6' 23*6 - 2,3'*'5,6' >5 - 2'3'*'5' 3*5 - 2,3**'5,6' 23*5 - 2*3***5*6* a>5o - a'j'k's'S' Predicted*. 2125 >25 2138 21** 2199 22>0 222 9 2226 (i25i) 225o (22-5) (2293) 2*C2 2323 231? (2') (i-71 ) (2335; (2376) 23'k (2779) 237* (2*01) (2*6d} (2*6*) (2*89) (2*31) 2*90 * 2*88 (2516) (2525) (25*6) (2577) (2581) (2581) (2595) (2oC <26%) (2627) (265C) (2685) (2738) (2798) *SiO Structure* found la Aroelor 125* by mi dctervtasUooa (RI) Predicted Rls uslug T*ol* % (Cl.Jto.) Chlorine aueber of slnor constituent oot associated witb accurst* III * Peaks oteerved ca ati spectrometer oscilloscope, cut net an G.L.r. Half HI values of phenyl rinRs containing uo to threw chlorine a tons Rlrw structure i(Rl) value Derivation H.I. phenyl- 747 biphenyl 1494 2> 42> 24- 834 829 935 930 947 940 1028 1025 1013 1009 2-2' 2- 3- 3' 3- 4-4' 4- . 23 - 2'4V 23 - 24 - 2'4' 24 - 1669 157T 1871 1678 1894 1687 2206 1772 2026 1756 25- 997 25 - 2'5' 1994 26- 930 26 - 2'6' i860 932 26 - 1679 34- ll4l 34 - 3'4' 2282 1134 34 - 1881 55- 234235- 1097 1095 1213 U62 35 - 3*51 35 - 234 - 2'4V 235 - 2 '4 '51 2194 1842 2391 2340 236245246- 1086 1178 1076 1086 1090 1084 236 - 2'4'5' 245 - 2'4'5' 246 - 2'4'6' 246 - 4' 246 - 2' 246 - 2264 2356 2151 2033 1924 1831 345- 1310 345 - 2'4'5' 2488 HONS 059906 i TABLE 6 Comparison of experln*ental and predicted RI? of FCBa Structure 2 - 1*' 245 - 24 - 4' 25 - 2'3,6I 25 - 5'4' 25 - `S'L-'S' 24 - 2,4'5' 25 - 2'>'4' 34 - 2'3'6' 234 - 2'3,6' 34 - 2'4'5' 235 - 2 '4 '51 34 - 2'3,4' 234 - 2,3,4i HI Experimental 1782 1922 i960 2089 2136 2175 2191 2207 2228 2299 2221 2339 2356 2425 RI Predicted 1781 1925 1956 2083 2138 2174 2191 2210 2227 229Q 2319 2340 2354 2424 MGNS 059907 TABLE 7 Experimental half RI values of phenyl rings containing 4 and 5 chlorine atoms Rinp? Structure 2>5- i(HI) value 136U 078 1380 Derivation 2345 - 2,3'4l 2345 3*4 ' 2345 - R.I. 25T7 2519 2127 2 346- 1245 2346 - 2,3,4,6' 2490 2356- 1219 1273 2356 - 2 * 315'6' 2356 - 2438 2020 231*56- 1405 23456 - 2'3'4'5'6' 2810 MOMS 059908 6 0 6 6 S 0 SNOW 15 mins 5 OT 6 6 5 0 SNOW ts3* *** FIG. 3 MCNS 0 5 9 9 1 1 ' '.f/G.f 7 hr 6 hr W % 9hr .' ' ,n llhr 14 hf . 77 73 ' 72 8hr 78 71 70 . 10 hr ., 13 hr. 78 15 hr MOMS 059912