Document wra0dL9QO6GXG02nr7rZRDo4E

!; ARMOUR: PERCHLORINATION OF PCBb 987 Quantitative Perchlorination of Polychlorinated Biphenyls as a Method for Confirmatory Residue Measurement and Identification JUDITH A. ARMOUR Division of Chemical Technology, Food and Drug Administration, Washington, D.C. 0204 The perchlorination procedure for deriva- biphenyl (DCB), that is indicative of the total tization of PCBs described by Berg, Diosady, and Rees has been modified to achieve a micro-scale quantitative conversion (greater than 90%) of commercial PCB preparations (Aroclors) to decachlorobiphenyl. Cleaned up sample extracts containing PCB residues (1-20 pg) are allowed to react with antimony pentachloride in 'the presence of chloroform to form decachlorobiphenyl. This procedure con verts a multicomponent mixture to a single derivative detectable by electron capture GLC, PCB residue and that could be used for quantita tive purposes have previously been reported (7, 8). In this study, the perchlorination procedure described by Berg, Diosady, and Rees (7) has been modified to achieve complete conversion on a micro-scale of all PCB components in a cleaned up sample extract to decachlorobiphenyl, thus providing a qualitative and quantitative con firmatory procedure for PCB determination. A quantitative value obtained by measuring the thus providing an easy method for quanti single GLC peak of a DCB derivative against a ll. tating and identifying PCB residues and at DCB standard of known purity reiuforces the the same time increasing the sensitivity of residue value achieved by the measurement of a IT. 2, detection. The usefulness of the perchlorina tion procedure is demonstrated by comparing multicomponent PCB residue. It also greatly in creases the sensitivity with which PCBs can be results for environmentally contaminated <5. samples quantitated by 2 methods: by meas detected and measured. Measuring the single uring the total area of the electron capture GLC response for DCB will also minimize neces * response for the residue against the Aroclor it sary analytical judgments involved in the physi most closely resembles, and by measuring the cal measurement of the GLC curve of a multi single peak of the decachlorobiphenyl deriv component residue (i.e., position of baseline and ative and expressing the results in terms of method of integration) and in the discrimination the particular Aroclor. S. between PCB and non-PCB components. Analytical methods normally employed for the METHOD ito- 95; extraction, cleanup, and detection of polychlo Reagents rinated biphenyl (PCB) (1-3) residues are gener (a) Antimony pentachloride.--Allied Chemical Re u); ally those used for organochlorine pesticides (4). agent Code 1365. Quantitative values are usually obtained by meas (b) Sodium sulfate.--Anhydrous, granular. It may uring the total area of a PCB residue response be necessary to wash with acetone and ethyl ether to as shown on a gas chromatographic (GLC) tracing remove electron-capturing substances. using an electron capture or microcoulometric de (c) HCl.--61V (1 + 1 dilution of concentrated HC1 tector and comparing it with the total area of a similarly obtained response of the commercial PCB preparation (Aroclor) (5) it most closely resembles (6). This has proved to be a practical procedure for quantitation; however, it is limited because of the lack of sensitivity, the multicom ponent nature of PCB residues, the alterations in in water). (d) Sodium bicarbonate.--10% aqueous solution. (e) Solvents.--Hexane, chloroform, and methanol suitable for use in pesticide residue analysis; distilledin-glass product is satisfactory (available from Bur dick and Jackson Laboratories Inc., 1953 S. Harvey St., Muskegon, Mich. 49442). residue pattern as a result of metabolic and envi ronmental changes, and the need to rely on com mercial Aroclors as reference materials. Procedures to convert all PCB components to a single derivative, e.g., biphenyl or decachloro- Received November 14, 1072. Apparatus (a) Vacuum hydrolysis tube.--150 X 10 mm od with No. 4 Teflon valve (Routes Glass Co. No. K896S60-0004) or Carius Pyrex combustion tube, 200 X 8 mm id, 10 nim od (Corning No. 8640). (b) Silicone oil bath and heating clement.--'Capable of maintaining temperature at 165-175C. >- DSW 297626 ; v' 'Vv"-** Art*ifcc_.y.. v.; mtsAiziifo STLCOPCB4067539 I". 988 JOURNAL OF THE aoac (Vol. 56, No. 4, 1973) (c) Kudema-Danish evaporative concentrator.-- With Snyder column and collection tube, 500 ml capacity (Kontes Glass Co. No. K5470000, or equiv alent). (d) Micro-Snyder column.--T 19/22, column size 2-19 (Kontes Glass Co. No. K569001, or equivalent). (e) Gas chromatograph.--Equipped with concen tric design electron capture detector operated at a dc voltage to produce M full scale recorder deflection for 1 ng decachlorobiphenyl when full scale deflection is 1 X 10" amp, and 6' X 4 mm id glass column of either: (1) 1% OV-101 on 100-120 mesh Gas-Chrom Q (W. W. Wiencke, 1972), 120 ml nitrogen/min car rier gas, column 220C, injector 230C, detector 200C; (2) 3% Dexsil 300 on 100-120 mesh GasChrom Q (W. W. Wiencke, 1972), 60 ml nitrogen/ min carrier gas, column 240C, injector 250C, de tector 200C. Perchlorination Procedure Extract and clean up sample to be perchlorinated by procedures (such as the FDA multiresidue pesti cide procedure (9)) leaving the residues in solvents that can be selectively evaporated from chloroform (6% ethyl ether-petroleum ether from Florisil or petroleum ether fraction from silicic acid). Transfer 1-2 ml cleaned up extract containing 1-20 Mg PCB residue to vacuum hydrolysis tube (a) (Carius tube may be substituted). Add several drops chloroform and 1-2 20-mesh carborundum chips, and reduce volume to 0.1 nd by carefully heating tube on steam bath. Let tube cool to room temperature, add 2 ml chloroform and 1-2 carborundum chips, and reduce volume to 0.1 ml on steam bath. Repeat addition of chloroform, concentrate to 0.1 ml, and cool to room temperature. Working in well ventilated hood, care fully add 0.2 ml antimony pentachloride to tube. Immediately seal tube tightly to ensure closed sys tem reaction. (Note: It may be necessary to cool Teflon valves in ice before fitting into vacuum hy drolysis tube. Narrowed bottom of valve should fit flush against constriction in glass when sealed. Car ius tubes should be sealed in an oxygen flame.) Im merse tube 2-3* in oil bath heated to 165-175C and let reaction proceed overnight (ca 15 hr). Remove tube from bath and cool to room temperature. In hood, and while pointing the venting side arm of vacuum hydrolysis tube away from face, open by slowly unscrewing valve and pulling it out of tube. (Carius tube should be scored with a file and snapped open in hood at arm's length away from face; use ex treme precaution since there is a rapid release of pressure and sometimes splintered glass.) Slowly add I ml 6.Y HC1 to reaction mixture, tap tube lightly to mix, and quantitatively transfer mixture to 30 ml separatory funnel, rinsing with several small portions of HC1: use a total of 5 ml 6A HC1. (Note: It may be necessary to rock tube back and forth or tap gently in order to drain.) Rinse reaction tube with several small portions of hexane, using a total of 15 ml hex ane, and quantitatively transfer rinsings to the 30 ml separatory funnel containing the HC1. Shake funnel vigorously 1 min; let layers separate. Drain lower HC1 layer into second 30 ml separatory funnel con taining 15 ml hexane and shake 1 min. Drain HC1 layer into third separatory funnel containing another 15 ml hexane and repeat extraction. Discard HC1 layer and combine the 3 hexane extracts in 125 ml sepnratory funnel. Wash hexane extract successively with two 20 ml portions of water, one 20 ml volume of 10% NaHCOa solution, and finally 2 additional 20 ml portions of water. Discard all washes. Dry washed hexane extract by passing through 2* col umn of anhydrous sodium sulfate and rinse column with 100 ml hexane, collecting hexane extract and rinse in Kuderna-Danish concentrator. Add few drops methanol and concentrate on steam bath to ca 3 ml. Separate volumetric receiver from concentra tor, attach micro-Snyder column to receiver, and concentrate extract to final volume <0.5 ml to re move traces of chloroform. Dilute to suitable definite volume for GLC determination. Perform GLC assay and calculations as described in Discussion. Discussion The micro-scale perchlorination procedure of Berg, Diosady, and Rees was evaluated to deter mine its utility as a routine laboratory procedure for confirmatory PCB quantitation. Parameters for quantitative extraction of DCB from the reaction mixture were not clearly defined by these authors and needed to be established prior to testing the procedure for reproducibility. Even after we developed a method of extraction that would recover greater than 95% of DCB standard from the reaction mixture, perchlorination using the reaction conditions of Berg el al. resulted in 90-100% conversion of Aroclor 1254 to DCB but only 30-70% conversion of Aroclor 1242. Condi tions contributing to inconsistency in behavior between Aroclor 1242 and Aroclor 1254 needed to be determined and modifications needed to be made to achieve conversion of all PCB compo nents in the various Aroclors to DCB.' Perchlorination The perchlorination procedure was designed for use following extraction, cleanup, ancillary (sepa ration), and detection procedures used in multi pesticide residue methodology. In the Berg et al. procedure, cleaned up extracts containing PCBs were evaporated to dryness before reaction with ARMOUR: PERCH; antimony pentac ducible conversic the original proci tion during evap molecular weigh Aroclors with less To reduce the lc while a small ai remains. Howevi vents which can 1 during the perch! solvents with mo exothermically v form a solid bln solvent, howevei favorably at the selective evapora up and separatio Perchlorinatio action tube at ele Increased pressu chloroform vapc dling the Carius Berg ef al. In p' quired sealing i opening, a react specifications, t venting side art Fig. 1). Parameters fo action mixture recover (greate. Traces of chloro be removed beft mination. This < which forms an evaporating to The resulting ex tron capture GI _.In experiment bility of a short as a catalyst. C were reacted an night condition was added after chloride, the tut reacted for 4 : under Recoverie By-products tative derivati; perchlorination Stalling (1972) oSM\| 297627 STLCOPCB4067540 ) ARMOUR: PERCHLORINATION OF PCBb 989 r antimony pentachloride. The observed nonreproi ducible conversion of Aroclor 1242 to DCB with the original procedure is attributed to volatiliza J tion during evaporation to dryness of the lower !l X molecular weight PCB components present in l Aroclors with less than 54% chlorine substitution. a To reduce the loss, evaporation is discontinued while a small amount of solvent (0.1 ml) still a remains. However, this limits the types of sol j vents which can be used since they may interfere y during the perchlorination step. It was found that ie solvents with more than one carbon atom reacted Ell y i- exothermically' with antimony pentachloride to form a solid black product. Chloroform as the tn solvent, however, minimized volatility, reacted id favorably at the given conditions, and permitted w selective evaporation of all solvents used in clean :a up and separation procedures. a Perchlorination is accomplished in a sealed re id action tube at elevated temperature and pressure. e- Increased pressure produced by the presence of te chloroform vapors increased the hazards of han %y dling the Carius tubes recommended for use by Berg ef of. In place of the Carius tube that re quired sealing in an oxygen flame and careful of opening, a reaction tube with the "same volume :r- specifications, tight Teflon sealing valve, and sre venting side arm is recommended for use (see i j ,,3 ! :rs Fig. 1). he Parameters for extraction of DCB from the re FIG. 1--Vacuum hydrolysis tube recommended for sealed system reaction: A, dimensions 10 mm od X 150 .se action mixture were developed to completely mm; B, tight Teflon sealing valve; C, venting side arm. to recover (greater than 95%) DCB standard. en Traces of chloroform present in the extract must product is formed when the antimony penta iat be removed before electron capture GLC deter chloride is contaminated with bromine. Care ird mination. This can be done by adding methanol should be taken to avoid use of contaminated ng which forms an azeotrope with chloroform and antimony pentachloride. in evaporating to a small volume before diluting. GLC systems suggested for DCB determina mt The resulting extract is sufficiently clean for elec tion are those originally developed for the detec di- tron capture GLC determination of DCB. tion of polychlorinated terphenyls (W. \V. r ior In experiments designed to determine the feasi Wiencke, 1972). At GLC conditions given in the led bility of a shorter reaction time, iron was tested Method, DCB elutes in 3-4 min on the 1% OV-101 be as a catalyst. One mg samples of Aroclor 1242 column and in 12 min on the 3% Dexsil 300 col !>- were reacted and extracted as described for over umn. These fast eluting systems produce in V night conditions except that 2 mg iron powder creased GLC sensitivity to allow easy detection was added after the addition of antimony penta- and quantitation of as little as 0.2 ng DCB. chloride, the tube was sealed, and the sample was for reacted for 4 and 6 hr. Results are discussed Quanfitation pa- under Recoveries. . Although the perchlorination procedure has lti- By-products that would interfere with quanti practical applications in PCB quantitation, it is al. tative derivatization were not evident in our not meant to completely replace procedures nor :bs perchlorination experiments; however, D. L. mally used for this purpose. The analyst should ith Stalling (1972) has found that a bromiuated by- still be aware of the identity of the contaminating DSW 297628 STLCOPCB4067541 / 990 JOURNAL OF THE AOAC (Vol. 56, No. 4, 1973) Aroclor(s), the presence of other residues which could interfere, and any alteration in Aroclor pat tern which may be due to metabolism and/or environmental exposure. This information may be useful in tracing the source of contamination and may have toxicological significance. Some laboratories may wish to express residues in terms of ppm DCB in order to establish a uniformity in their reporting. However, when the procedure is used for quantitative confirmation, DCB is ex pressed in terms of a specific Aroclor by using a factor for mathematical conversion. The factors used for converting DCB to the .equivalent con centration of a specific Aroclor are derived by assuming the reaction to proceed as follows: Cl Table 1. Factors to mathematically convert decachlorobiphenyl to an equivalent amount of Aroclor Aroclor Av. No. Cl* MW* X* 1221 1232 1242 1016 1248 . 1254 1260 1262 DCB ` 1 2 3 3' 4 S 6 7 10 188.5 223 257.5 257.5 292 326.4 361 395.3 499 0.38 0.45 0.52 0.52 0.59 0.65 0.72 0.79 1 * Average whole number of chlorines calculated from per cent chlorine substitution for a specific Aroclor. * Molecular weight of Aroclor based on the average whole number of chlorines calculated from per cent chlorine substitution.. *X -- molecular wt Aroclor/molecular wt DCB (499). To convert ppm DCB to ppm of a specific Aroclor, multiply ppm DCB by X for the Aroclor. SbCl^ Cl Cl Cl Cl Table Z. Per cent conversion of 1719 PCB to decachlorobiphenyl at overnight conditions Aroclor Trials Range, % Average, % where n = number of chlorines -- approximate whole number of chlorines calculated from per cent chlorine substitution for a specific Aroclor (see Table l). The approximate whole number of chlorines in a specific Aroclor was chosen for calculations, since Aroclors are commercial prepa rations and the actual per cent chlorine may vary from the theoretical for each lot. The factor for each Aroclor (see Table 1) is the ratio of amount of Aroclor necessary to yield one unit of DCB: X = factor for expressing DCB as a specific Aroclor X = molecular wt Aroclor/molecular wt DCB (499) Calculations for measuring PCB residues by GLC based on the DCB derivative formed and converting mathematically to express the residue in terms of a specific Aroclor are: (DCB peak ht sample/DCB peak ht std) X ng std injected = ng DCB in sample' injection ng DCB in sample injection X factor for spe cific Aroclor (X) *= ng Aroclor in sample injection (ng Aroclor in sample injection/wt sample in jected (mg)) = ppm Aroclor in sample This paper was presented at the Symposium on Industrial Chemicals as Food Contaminants at the Both Annual Meeting <>f the AOAC, Oct. 0-12, 1972, at Washington. D.C. 1242 1016 1248 1254 1260 1262 10 5 5 5 5 3 90-105 87-105 98-102 95-104 90-95 89-103 97 95 99 100 93 95 Recoveries To determine the per cent conversion of spe cific Aroclors to DCB, a series of 1 ng of 6 differ ent Aroclors was allowed to react at the given overnight conditions and was extracted as de scribed. The 1% OV-101 column was used for the gas chromatographic determinations. Results are given in Table 2. The method demonstrated good reproducibility for quantitative conversion of the 6 Aroclors to DCB with results ranging from 87 to 105% and an average recover)' of 97%. In the experiments with shorter reaction times, 10 trials with 4 hr reactions using Aroclor 1242 gave a wide range of-results, although the average conversion to DCB wras 90%. Trials with 6 hr reaction gave results falling in a much narrower range and averaging 102% conversion. Results from these and experiments using the 6 Aroclors at 6 hr conditions are given in Table 3. The per- chlorination is quantitative (99% average con version for 6 Aroclors) and reproducible (range 83-110%) for 6 hr reaction with iron catalyst but offers little advantage over the convenient over night conditions. . ARMOUR: P: In order chlorinatio of fatty sa com oil w< residue me petroleum were fortif Three aliqv equivalent cleaned up perchlorinr age recover results she up extract handle 1-2 Samples environme chicken fa and clean; methodolo separated Table: dec*' Aroclc j, 1242* 1242 1016 1248 1254 1260 1262 4 hr rea; Table 4. cc Sample Chub Chub Chub Chub Sturgeon* Eggs Chicken fa Chicken fa * Expres * Residu * Sam pit study. DSW 297629 STLCOPCB4067542 1973) t ed from clor. average per cent (499). To multiply 6 to ions 7 15 19 n 35 >n of spe>f 6 differthe given ed aa desed for the lesults are rated good sion of the ig from 87 '7%. :tion times, roclor 1242 the average > with 6 hr :h narrower on. Results i 6 Aroclors 3. The perverage coneible (range catalyst but enient over- ARMOUR: PERCHLORINATION OF PCBs 991 In order to determine the behavior of the perchlorination procedure in the presence of extracts of fatty samples, 3 g samples of chicken fat and com oil were extracted by FDA multipesticide residue methodology (9). The 6% ethyl etherpetroleum ether fractions from Florisil cleanup were fortified with Aroclor 1242 at 6.7 ppm. Three aliquots of the cleaned up chicken fat, each equivalent to 0.15 g fat, and 2 samples of the cleaned up com oil, each equivalent to 3 g, were perchlorinated at overnight conditions with aver age recoveries of 107 and 96%, respectively. These results show that the method can tolerate cleaned up extracts of up to 3 g samples of fat and can handle 1-20 pg PCB. Applications Samples of fresh chub, fresh sturgeon, and eggs environmentally contaminated with PCBs, and chicken fat fortified with PCBs were extracted and cleaned up by FDA multipesticide residue methodology (9). Where necessary, PCBs were separated-from pesticides by chromatography on Table 3. Per cent conversion of X PCB to decachlorobiphenyl with Iron catalystand 6 hr reaction Aroclor Trials Range, % Average, % 1242* 1242 1016 1248 1254 1260 1262 ' 10 10 3 3 3 3 3 78-98 96-110 87-95 95-100 98-106 91-108 83-107 90 102 92 98 103 98 98 4 hr reaction. Table 4. PCB residue levels In environmentally contaminated samples obtained by 2 different methods of quantitation Residue, ppm Sample Ref. Aroclor By total area By DCB" - Trial 1 Trial 2 Chub Chub Chub Chub Sturgeon* Eggs Chicken fat1 Chicken fat1 1254 1254 1254 1254 1248 1242 1242 1248 3.5 3.4 4.2 3.4 28.4 7.1 4.6 6.9 4.2. 4.2 5.1 4.5 28.6 6.8 5.0 6.6 28.4 7.1 * Expressed as reference Aroclor. * Residue is reported on a lat basis. * Samples fortified with Aroclor for an interlaboratory study. silicic acid. The PCB residues were quantitated on electron capture GLC with a 10% DC-200 column (9,10) by comparing the total area of the residue with total area of the Aroclor reference with most similar GLC pattern. Aliquots of the 6% ethyl ether-petroleum ether Florisil eluate equivalent to 3 g chicken fat, 0.2 g sturgeon fat, and 0.2 g egg were perchlorinated for qualitative and quantitative confirmation. The residue level reported for each sample after perchlorination was determined by measuring the DCB deriva tive (1% OV-101 column) and mathematically converting it to the equivalent amount of the Aroclor used as reference in the total area quanti tation. Comparison of the residue levels obtained by the 2 determinative procedures is presented in Table 4. Results obtained by the 2 methods of quantitation are in close agreement, demon strating the practical application of the per chlorination procedure for quantitative confirma tion. Gas chromatograms of an egg sample (Aroclor 1242) quantitated Tor PCBs as a multi component residue before perchlorination and as the DCB derivative after perchlorination are shown in Fig. 2. In the case of the chub samples, the extracts containing residues of both PCB and DDT analogs were separated on silicic acid prior to quantitation of the multicomponent PCB resi dues. Aliquots equivalent to 0.3 g chub from the silicic acid petroleum ether (PCB) fraction were perchlorinated and results are compared (see Table 4). Gas chromatograms of a chub sample quanti tated for PCBs (Aroclor 1254) as a multicompo nent residue before perchlorination and as the DCB derivative after perchlorination are shown in Fig. 3. Figures 2 and 3 also illustrate the in crease in sensitivity obtained by perchlorination. For approximately equivalent peak heights on GLC only about xg as much of the egg sample (Aroclor 1242) and xg as much of the chub sample (Aroclor 1254) were injected after perchlorination as was needed before perchlorination. Organochlorine pesticides, if present in a sam ple, should be separated from PCBs prior to derivatizing the residue to DCB. Pesticides such as DDT may be chlorinated to derivatives other than DCB. Preliminary studies show that per chlorination of DDT itself results in the forma tion of several derivatives which are separated from DCB on the 3% Dexsil 300 column at the specified conditions and do not interfere. t, i i: '. ! d! IK t ! f :l > I 1 ( i DSW 297630 STLCOPCB4067543 992 s+~--- . tft-aamiwfu'iiT ~i -- i jivca**'-'-'--' JOURNAL OF THE AOAC (Vol. 56, No. 4, 1973) ARMOUR: PERCHI FIG. 3--GLC curv cuiated as 4.2 PC .eluate from silici' Aroclor 1254), pc -------------- iy MINUTES ns FIG. 2---GLC curves of egg samples containing 7.1 ppm PCBs (as Aroclor 1242) before and after perchlorination: A, 1.11 mg sample (7.9 ng Aroclor 1242). 6% Florisil eluate before perchlorination, 10% DC-200 column (9,10); B, 0.07 mg sample (0.5 ng equivalent Aroclor 1242), 6% Florisil eluate after perchlorination, 1% OV-101 column. GLC conditions given in Method. (1) Peakall, D Science 20, (2) Reynolds, (3) Risebroug! (1969) Bid 201 (4) Armour, J S3, 761-7t (5) Aroclor O/PL306, DSW 297631 reuTOitiemiH ;wi|i'.hv j ti f;i ywyi STLCOPCB4067544 i Veil r ARMOUR: PERCHLORINATION OF PCBs iMmrrmrrri'fi'iri^mmdrnaiiitKUXi T Tt^-r *] I i. i. 993 : 1 r }. * t FIG. 3--GLC curvM of chub samplo calculated as 3.4 ppm PCBs (as Aroclor 1254) before perchlorination and cal culated as 4.2 PCBs (as Aroclor 1254) after perchlorination: A, 1.5 mg sample (5.1 ng Aroclor 1254), petroleum ether eluate from silicic acid before perchlorination, 10% DC-200 column (9,10); B, 0.15 mg sample (0.63 ng equivalent Aroclor 1254), petroleum ether eluate from silicic acid after perchlorination, 1% OV-101 column. GLC conditions given in Method. References (1) Peakall, D. B., & Lincer, J. L. (1970) Bio Science 20, 958-964 (2) Reynolds, L. M. (1970) Residue Rev. 34, 25-57 (6) Armour, J. A. (1972) J. Chromatogr. 72,275-282 (7) Berg, O. W., Diosady, P. L., & Rees, G. A. (1972) Bull. Environ. Conlam. Toxicol. 7, 338 (3) Risebrough, R. W., Reiche, P., &01cott, H. S. (1969) Bull. Environ. Contain. Toxicol. 4, 192 201 (4) Armour, J. A., & Burke, J. A. (1970) JAOAC 53, 761-76S (5) Aroclor Plasticizers, Technical Bulletin 347 (8) Hutzinger, O., Safe, S., & Zitko, V. (1972) Ini. J. Environ. Anal. Chem. 2, 95-106 (9) Pesticide Analytical Manual (1972) Vol. I, Food and Drug Administration, Washington, D.C. (10) Official Methods of Analysis (1970) 11th Ed., O/PL306, Monsanto Co., St. Louis, Mo. AOAC, Washington, D.C. .1 i DSW 297632 STLCOPCB4067545