Document 3NNemY78JBqObgm2K6e3rLka3

i l i . , i I . 1 ; i : .; i . p<-V <Jxa A Method for the Quantitation of Polychlorinated Biphenyl (PCB) Isomers by James W. Rote ami Philip C. Murphy . ' Hopkins Marina Station - Stanford University . Pacific Grove. Calif. 93950 Polychlorinated biphenyls (PCB's) were first de tected ip l$66-67 in European wildlife (1-3). Since that time, workers in several European countries, the U.S., and Canada have reported the presence of PCB's in marine invertebrates, fish, birds, marine mammals, and humans (4-9). Attempts'to quantitate the levels have been complicated by the fact that PCB's are not single compounds, but complex mixtures of isomers having vary ing degrees of chlorination. Monsanto Chemical Company, the sole producer of PCB's in the U.S., markets them under the trade name 'Aroclor*. The products are designated by numbers; the first tSiro digits represent the molecular type, and the last two digits give the weight percent of chlorine. Thus, Aroclor 1254 is a mixture of chlorinated biphenyls having an average chlorine content of 54% (5 Cl atoms). . , C* . I - Due to the complex nature of these mixtures, relat ive amounts of total PCB's have been estimated for tis sue samples and whole organisms. Using electron-capture gas chromatography, the method described here allows for the quantitation of individual PCB isomers. The metnod is based on the fact that detector response is not the same for all PCB components, but depends on the degree of chlorination. METHODS Tissue samples were digested by a perchloric-acetic acid mixture (10) and the fat extracted in hexane. Cleanup methods employed by the Food and Drug Adminis tration (11,12) were used which completely recovered PCB's along with the pesticide residues. After aceto nitrile partitioning and Florisil column cleanup, PCB's were separated from the pesticides by column chromato graphy on silicic acid-Celite (13). . ** f The determinative step was carried out under the following conditions: Gas Chromatograph: Beckman Model GC-4, fitted with helium glow electron-capture detector; column: glass, 6 ft. x 4 mm I.D., packed with a 1:1 mixed bed of 5% DC-200 andj 6% QF-^1 (on Chromosorb W (AW/DMCS) . . ' IT? Bttltaia Environment*! Contamination b Toaicolofy, VoL 6, No. 4, 1971, pabliaHod by Springe*-Ycrlag New York Inc. "gqy"iiw.nrwmm. :ii.uw.^ nium'WOB HUM 0S\N 332566 STLCOPCB4079142 Operating Conditions: Column temperature 200'C., injector temperature 230#C., detector temperature 280C; Carrier gas (He) flow rate, approximately 175 ml/min? Electrometer current 1.6 x 10-9 amps; Recorders Beckman Model 1005 ten inch potentiometric, 1 mV, full scale de flection; chart speed 1 in/min; volume injected 4 ^j.1. Chromatograms of' six PCB standards were obtained for Aroclor 1232, 1242, 1248, 1254, 1260, and 1262. These commercial mixtures correspond approximately to an average of two, three, four, five, six, and seven chlorine atoms per molecule, respectively. The Aroclors were dissolved in hexane to concentrations of 2 ng/p.1. From preliminary injections it was found that a 4 p.1 in jection placed peak heights on the linear portion of the * standard curve for each Aroclor. Duplicate 4 p.1 inject ions were made, providing a total of 16 ng for each Aro clor chromatographed. Total peak area for each Aroclor was determined by the cut and weigh method. The excellent resolution obtained with our operat ing conditions revealed a total of 32 different PCB iso mers in the six Aroclor standards. The peaks were num bered 1 through 32 according to retention time. Indiv idual peaks were identified as to chlorine content from mass spectrometry results of other workers (5,8,9). Thus, peaks 1 and 2 are monochlorobiphenyIs; peaks 3 and 4, dichlorobiphenyIs; peaks 5 through 8, trichlorobiphenyls; peaks 9 through 13, tetrachlorobiphenyIs; peaks 14 through 16, pentachlorobiphenyls; peaks 17 through 22 and 24, hexachlorobiphenyIs; peaks 23 and 25 through 28; heptachlorobiphenyIs; peak 29, octachlorobiphenyl; peaks 30 and 31, nonachlcrobipher.y Is ; and peak 32, decachlorobiphenyl. A standard injection of p,p`DDE was made periodically to provide for a measure of relative retention time and to check detector sensitiv ity. The amount of various chlorinated components in an Aroclor was determined by combining all peaks that cor responded to monochlorobiphenyIs, all that were dichlorobiphenyIs, etc. Where there was a problem of peak overlap, the 2 peaks were divided by a straight line drawn perpendicular from the minimum point between the peaks to the baseline. The cut and weigh method, using this technique, approximates automatic integration which operates on the same principle. The s^milogarithmic relationship of detector re sponse (total peak area/ 16 ng) to average chlorine content wis obtained for each Aroclor by the method of least squares. From this relationship, the theoretical response of the detector to each chlorinated biphenyl 1(mono through deca) was calculated (Fig. 1). These* 1 1 1i , (71 I , c i **1 li DSW 332567 r I STLCOPCB4079143 Figure 1. Detector response curve. Response to p,p'~ DDE *= 2.75 in~/ng. (log y = 0.1512x - 0.7jl-28) values were then used to calculate the amount of chlor inated biphenyl represented by each peak in the sample chromatogram. The response curve was also used to de termine the amount of each chlorinated component in the PCB standards. RESULTS When the individual peaks of Aroclor 1232 through 1262 were quantitated by means of the response curve, it was possible to determine the distribution of the 16 ng for each standard. Table I shows the peak area (ng) for each chlorinated specie^ and the percentage of the 16 ng tot^l injection accounted for in each standard. The low percentage for Aroclor 1232 (62.5%) is apparently due to unchlorinated biphenyls in the mix ture. Since this Aroclor has an average of only 2 chlorine atoms per molecule, a certain percentage of the molecules would have no chlorines. The distinct odor of these lower chlorinated Aroclors is in fact due *79 temp* OSNN 332568 I STLCOPCB4079144 to the presence of biphenyl in the mixture (14). These non-electron capturing particles do not contribute to total peak area and therefore the chromatogram does not account for the amount of material injected. Ttie low percentage for Aroclor 1254 (60.4%) may be due to the presence of unchlorinated biphenyls or other non-electron capturing impurities in the preparation. In the synthesis of fVroclors, the degree of chlorination is determined by either measuring the specific gravity or the viscosity of the mixtures. Therefore, it is possible that the average chlorine content of an Aroclor varies from batch to batch. TABLE I Distribution of chlorinated components in Aroclors. Peak area (ng) # Cl 1232 1242 1248 1254 1260 1262 . 1 1.466 .2607 - ' - - - 2 2.472 1.950 .3681 - - - 3 2.732 5.681 3.223 .1896 - -; 4 1.931 4.321 4.559 1.108 - -' 5 1.111 2.541 3.868 3.092 1.139 .7123! 6 .2927 .6039 1.225 4.609 5.590 4.316 ! 7 - - - .6696 5.678 7.034 8 - - - - .7770 2.199 9 - - - - .6700 1.338 10 _ _ _ : 0910 .2850 Total 10.00 15.36 13.24 9.668 13.95 15.88 % 62.5 96.0 82.8 60.4 87.2 99.3 Fig. 2A is the chromatogram of a liver extract from the sea otter Enhydra lutris. The 50.6 g sample, which had about 2% fat content, contained 27 PCB isomers and had a total PCB level of 2.0 p.p.m.(total tissue). The chlorine content of each isomer is indicated below the peaks. A chromatogram of Aroclor 1260 is provided (Fig.2B) for comparing the retention time of each isomer. An no \ i f 05\N 332569 STLCOPCB4079145 9* Ml ^IIJLPSA. !m Wip*1*!! V>|fy Jr*0yi O ai O -H VO *4-1 IN -r4 --< +> c UV o *o r-4 *#4 U wO <<-N* << VD 44 0*0 C I (0 u f*l OO' 4J Xis EV uM0 cu w --a Qa u 0< 4) > *4-1 o-H4 O U B01 H 4-> oc o 00 -44J 0 CV <44 44 O 14 IN 1*1 (0 IN W 0) r-t 0*44 O0M 4J U O 0 -r4 i-4 E *0 U oco 44 -4 M V ,,J C OE -- MO < 44 44 < **4 IN 0 M3 O' ir> DSW STLCOPCB4079146 r arrow marks the retention time of p,p'-DDE relative to the PCB's. Table II gives the amount of chlorinated biphenyl in ng and p.p.b. for each peak of Fig. 2A. As can be . seen in Fig. 2A, there are no monochlorobiphenyls and only one dichlorobiphenyl isomer present in the sample. This may indicate that the lower chlorinated PCB's are less persistent in the environment. A previous study has shown that lower chlorinated components are metabol ized in.birds (5).^ TABLE II Quantitation of PCB components in sample.* Peak area (ng) Peak # 3 ng. .2686 D.P.b. 41 Peak # 20 ng 1.423 P-P-b210 6 .2036 30 7 .2598 39 21 22 . .2285 2.013 35 300 8 .2212 33 23 .6775 100 9 .1116 17 24 .2964 45 10 .3099 47 11 .1537 23 12 .1537 23 14 .4358 67 15 .8995 140 16 .4463 68 17 .6397 97 18 .1964 30 19 .7990 120 25 26 27 28 29 30 31 32 Total .2738 .3679 1.419 .5632 .5195 .1821 .2642 .0822 13.41 41 56 220 85 79 27 39 12 2024 * Sea Otter from Monterey, California. STLCOPCB4079147 PI8CUSSI0H In quantitating levels in marine birds and fish, Risebrough (15) assumed that each PCB compound produced tne same peak height with the electron-capture detector as the same amount by weight of p,p'-DDE. After sum ming the contributions of the individual peaks, -the total was multiplied by a factor derived from measure ments of . standard solutions. Our results (Fig. 1), show that detector response is not the same for all isomers, but is dependent on the degree of chlorination. The principle of an electron-capture detector is based on the electron affinities of compounds. A polychlorinat ed compound can have 10^ times the relative electron absorption coefficient as a monochloro compound and 10 times that of an unchlorinated compound (16). Koeman et al. (5) semiquantitatively measured PCB residues by using one peak of a commercial mixture as a standard. Reynolds (17) used a similar method but based his quantitation'on an average of two or more peaks. He reported his results as Aroclor 1254 or 1260 depending on the overall pattern of the chromatogram profile. Both of these methods assume that the peak(s) used as a standard give the same detector response as the rest of the peaks in the sample. Several workers have compared the PCB components in field samples to Aroclor 1254 because the sample chromatograms were most similar to this mixture (8,18, 19). If Aroclor 1254 had been used as a standard for measuring levels in the sea otter liver, a value of 4.2 p.p.m. would have been obtained. This figure is over twice the 2.0 p.p.m. value that we calculated with the detector response curve. As stated before, there may be impurities in the Aroclor being used as a standard and any quantitation based on that one Aroclor would be err oneous. By using all six Aroclor standards, the errors due to impurities and unchlorinated biphenyls are min 3 imized. If the standard injection of p,p'-DDE shows that detector sensitivity has changed from when the response curve was calculated, a simple factor can be applied to the entire curve and all isomers quantitated as before. We feel that this method should be used by all workers when reporting PCB levels found in field sam ples. A- more accurate quantitation will be achieved when individual PCB components are synthesized for use as standards. M3 *' I - STLCOPCB4079148 ACKNOWLEDGEMENTS We wish to thank Judson E,. Vandevere for supplying the tissue samples and the Monsanto Chemical Company for providing the Aroclor standards. 1. 2. 3. 4. )5' 6. 7, 8. 9. 10, 11, 12, 13, 14, 15, 16 17 18 19 REFERENCES JENSEN, S., New Sci. 32,612 (1966) W1DMARK, G., J. Assoc. Offic. Anal. Chem. 50,1069 (1967) HOLMES, D.C., SIMMONS, J.H. and TATTON, J., Nature 216,227 (1967) RISEBROUGH, R.W., REICHE, P., PEAKALL, D.B., HERMAN S.G. and KIRVEN, M.N., Nature 220,1098 (1968) KOEMAN, J.H., TEN NOEVER DE BRAUW, M.C. and DE VOS, R.H., Nature 221,1126 (1969) JENSEN, S., JOHNELS, A.G., OLSSON, M. and OTTERLIND G., Nature 224,247 (1969) DUKE, T.W., LOWE, J.I. and WILSON, A.J., Bull. Environ. Contam. Toxicol. 5,171 (1970) BAGLEY, G!E., REICHEL, W.L. and CROMARTIE, E., J. Assoc. Offic. Anal. Chem., 53,251 (1970) BIROS, F.J., WALKER, A.C. and MEDBERY, A., Bull. Environ. Contam. Toxicol. 5,317 (1970) STANLEY, R.L. and LEFAVOURE, H.T., J. Assoc. Offic. Agri. Chem. 48,666 (1965) Pesticide Analytical Manual Vol. I, Food and Drug Admin. (1968) Pesticide Analytical Manual Vol. Ill, Food and Drug Admin. (1970) ARMOUR, J.A. and BURKE, J.A. , J. Assoc. Offic.: Anal. Chem. 53,761 (1970) PENNING, C.H., Ind. Eng. Chem. 22,1180 (1930) RISEBROUGH, R.W. pp. 5-23. Ins Chemical Fallout, BERG, G.C. and MILLER, M.W. (eds.), Charles C. Thomas, Springfield, 111. (1969) GREGORY, N.L. and LOVELOCK, J.E., Anal. Chem. 33.45A (1961) REYNOLDS, L.M. Residue Rev. 1970 (in press) REYNOLDS, L.M., Bull. Environ. Contam. Toxicol. 4,128 (1969) RISEBROUGH, R.W., REICHE, P. and OLCOTT, H.S., Bull. Environ. Contam. Toxicol. 4,192 (1969) Mr i h ( ' H iT iilfr u iiiTun jewmpwrwi i. wwtw m \ STLCOPCB4079149