Document jZYk97gDJNddgYn1kVwz2owy

March 15 1976 Dr. I. Z. Wallen Environmental Protection Agency Office of Toxic Substances 401 'A Street S.W. vra5hinct.cn, DC 204 60 . Dear Dr. Wallen: Monsanto has a concern over the validity of the perchlori nation technique used by EPA. and others to measure and/or confirm polychlorinated biphenyls in environmental materials. In our investigation of the perchlorinaticn method, we have found that such chemicals as biphenyl, alkylated biphenyl, and many other substituted biphenyls, interfere with the perchlorination technique. It also appears that various petroleum components may interfere. If these chemicals were present in environmental materials that were being tested for PCB's using the perchlorination method, erroneously high PCS concentrations would be reported. A recent article in the Journal of the AOAC (Vcl.58, No.3, 1975) points out two other limitations of the perchlorination procedure: 1. High and variable reagent blanks, which cause erroneously high findings. 2. Formation of bromoncnachlorobip'nenyl, which causes- low recoveries. A copy of the article is attached. ' osw 012V* STLCOPCB4002395 Dr. I Wallen 2- - March 15, 19 7 6 H J3 Since these lir.'.itaticns can lead to significant.errors in deterriainc trace levels of rCB's in environmental samples, we suggest that E?A carefully review the val of the perchiorinaticn technique. Results that have obtained using this technique .nay not be valid. If any other information is needed, please let us knew. Sincerely, J. Coleman Weber Manager, Product Acceptedi 11 c- mah cc: Dr. A. C. Trakovski Environmental Protection Agency DSW 012471 STLCOPCB4002396 hOTiCE: This material may be protected by copyright 466 jocrxal or the aoac (Vol. 5S, No. 3, 1973) Limitation on the Use of Antimony Pentachloride for Perchlorination of Polychlorinated Biphenyls WILLIAM J. TROTTER and SUSAN J. V. YOUNG Division oj Chemistry and Physics, Food and Drug Administration, Washington, DC 2020!, Two contaminant] are present in commer cially available antimony pentochloride (SbQj) uied to perehlorinate polychlorinated biphenyl] (PCBs) to decnchlorobiphenyl (DOB). DCB is found in the ShClj perchlorination reaction blank in which no PCBs were added. Bromononachlorobiphenyl (B.NCB) ii found after use of SbCI] to perehlorinate PCBs. Levels of DCB found in the SbClj reaction blanks from vari ous distributors ranged from 8 to 972 ng DCB/ ml SbClj. The relationship of the formation of BNCB to amounts of various PCB Aroclors perchlorinated is examined. Polychlorinated biphenyl (PCBs) residues are extracted, cleaned up, and detected by methods similar to those used for oreanochlorine pesti cides. PCB residues are quantitatively deter mined by comparing the gas-liquid chromato graphic (GLC) response of the multicomponent residue and commercial PCBs (ArcclorS) or a mixture of Aroclors producing a GLC response pattern similar to that of the residue (1). This approach is limited because the multicomponent PCB residue may not have the same propor tional composition as the Aroclor or Aroclors used as the quantitation reference. Residues can bo composed of mixtures of chlorobiphenyl com ponents from more than 1 Aroclor. Metabolic and other environmental factors complicate the description of the PCB residue composition. There has been considerable work to develop methods to convert the multicomponent PCBs to a single derivative on which to base the resi due determination. Procedures have been re ported to catalytically dechlorinate PCBs with hydrogen over palladium or platinum to bi phenyl, cyclohexvlbeniene, .and bicyclohexvl (2, 3). A principal disadvantage with that pro cedure is that the hydrocarbon product is determined with a GLC flame ionization detec tor, resulting iti low sensitivity. Attempts have been made to convert PCBs to the fully chlor inated dccachlorobiphenyl (DCB) (C-o). Ar mour (0) reported optimum conditions Tor perchlorinating PCBs with antimony pentachloride (SbCI,). The method provides a qualitative confirmatory procedure for PCB determination The GLC electron capture detector response Lenhanetd because total FCBs are manifested as single peak for DCB. In measuring tbe siniiv peak for DCB the analyst is not faced with analytical judgments such as baseline correction, method of integration, or discrimination between PCBs and non-PCB components. However, it inecessary to be aware that the various Aroclors give rise to different equivalents of DCB and that the nonchlorinated biphenyl (also use-.! as a fungicide) is perchlorinated by SbCI, t DCB. Nonetheless, using the perchlorination derivatization can reinforce the residue vain.' determined by measuring a multicomponent PCB residue. During attempts to apply the perchlorination derivatization in determining low residue levelof PCB and make use of the increased electron capture response to DCB, 2 contaminants we: indicated which led to erratic recoveries of DC3 Experimental Raagants and Apparatus fa) Antimony pentachloride.--Hooker Chemical Niagara Falls, NY 14302 (received in glass hot:'..with lead-lined cap); Matheson Coleman & B-,. (M&B), Norwood, OH 45212 (reticent grad-'' : BifcA (Allied Chemical), Morristown, >"J 077-'-: (reagent grade, 9D"c); Research Organie-Inorgan, Chemical (ROC-RIC), Belleville, NJ . 071.- (99.9977); and J. T. Baker Chemical, Phillipshur.-. NJ 0S5R3 (Baker Analyzed Reagent). (b) Cat chromatograph.--Searlc-ADalytia (D- Plaines, IL 60053) Model 5360 with 6' x I nur. : glass column containing 1% OV-101 on SO-11' mesh Cliromosorb TV (HP). Operating conditions column flow, 60 ml nitrogeo/min; column, 202C detector, 202*C; injector, 225*C; pin-cup (Usic. electron rapture detector with titanium MI tod detector-voltnce (constant dc) adjured t<> canonc-half full scale recorder deflrrtion for 0 7 t. DCB when full scale deflection is I X 10ninp (c) Mms tptclrometer.--Varian MAT (25 I'mi'22, Springfield, NJ 07051) CH5-DK nuiss tpc'-in-ietcr (MS) coupled to Varian Aerograph 2710 ,- DSW 012472 STLCOPCB4002397 TROTTER * YOCVC: PEKCULOr.lSATJON OF PCD. 467 ebronstogniph via sll-gltsj svft'.m uiiag IVatsonBitmian 2-<ugc separator. CLC operating coDdilious: 6'x 4 mm id gJug column containing 37i OV-1 on SO-lOO mh Chroxosorb 'V (HP); column Sow, 60 ml helium xia; column 240*C. MS operating conditions: electron energy, TO ev; emission current, 300 >n; multiplier voltage, 32 kv. Results and DUcucrion A peak identical to that of DCB was found in the reaction blank fur the Armour perchlorinatiou procedure (G) with the described GLC operating condition?. The identification of DCB was confirmed by GLC-M5 of a hexane extract of a hydrolyzed sample of SbCl, which had not been subjected to the perchlorination procedure. Various quantities (0J2-2.0 ml) of SbCl, from the 5 commercial sources were examined to de termine the presence of DCB. SbCl, alone was earned through the perclilorinaiion reaction (G) except that no CHC1, was present with 5bCls iu the reaction vessel. DCB was determined by GLC. Table 1 lists the amounts of DCB found. After perchioi mating PCBs with SbClj, a secondary peak with a GLC retention time rela tive to DCB of 1.31 vras observed similar to that reported by Huckins ft ol. (7). This later eluting peak is seen in Fig. 1, the chromatogram from the 0.2 ml SbC), (Hooker Chemical) perchlorination of 0.50 pc Aroclor 1221. This peak was found when SbO, from each supplier was used. The peak was determined by GLC-MS to be due to broroonouachlorobijihenyl (BXCB). BXCB was assumed to be a competing product with DCB arising from a small amount of SbCI,Br in SbClf, so parameters relating to possible limitations of the perchlorination procedure weTe studied. Various quantities 10.5-10 pp) of Aroclors 1221, 12-12, 125-1, and 1260 in CHCI, were perchlorinated. Recoveries of DCB and estimates of the relative amounts of BXCB,formed are given in Table 2. Calculation of the relative TkU 1. DCB (ng/ml) form4 from vorltvo of Sfcdj Supplier SbCl). ml C.Z 1.0 2.0 Av. Hooker Chemical MC8 Bif ROCRlC .T. B.kcr 17 u 47 (1 IS (2 is U MO 10*: 911 972 iz u 12 12 9 7 71 FIG. I--loctTon ejpluro GLC curvo from th 0.2 ml SbCI (Hooker Chmcol) ptrchJorination of 0.50 ug Aroclor 1221; 0.51 ns equivalent Aroclor 1221 infadatf. rook 1 represents 0.81 n* DCB. Pttk 2 represents 0.28 ng 8NC0. amounts of BXCB product formed was based on comparison of the electron capture GLC peak height of BXCB with that of a DCB reference. The amount of DCB determined in the reac tion blank was directly proportional to the amount of SbCl3 used (Table 1). This indicates SbClj was the source of the DCB and that contamination from other possible sources dur ing the perchlorination was negligible. The pro cedure for percblorinating PCBs specifies the use of OB ml SbCl,. SbCl, producing S-972 ng SbCl,/ral in the reaction blank would add 0.5 65 ppb, based on a 3 g sample. DCB produced in the reaction blank was as sumed to come from PCB contamination of SbCl.. In an effort to locate the origin of this contamination, SbCIs bottle closures were investi gated. GLC analysis of hexane, in which the plastic caps were soaked for 4 day.*, did not reveal PCBs. Hooker Chemical, the sole do mestic source of SbClj, supplied SbCl, in glass bottles with lead-lined caps. This bulk supplier of SbClj indicated that the production of chlo rine in carbon anode half-cells with linseed oil or other organic binders forms certain organic compounds: however, the destructive oxidative environment in the electrolytic cells would make the production of PCB unlikely as a result of this pathway. On the other hand, antimony metal is commonly obtained ns a metallurgical by-prod uct by carlxxn reduction of its oxide; therefore. DSW 012^73 -f *- * ---4., 468 JOURNAL or THE AOAC (Vol. 55. No. 3, UI7j' TabU 2. DCS and BNCB from parehloHnatlon of various AroClpo with 0.2 mi SbClj* of lower amounts of PCBs the relative yield e BNCB to DCB is greater because the brorni: Aroelor Ami, P8 DCB BNC8* DCB 4- BNCB* ating agent is the limiting quantity in contar-. rcd, combined %% rC, % inated SbClj. (2) Bromination occurs to a targ degree for a given quantity of the less chlori 1260 1254 1242 1221 10 10 10 10 M0 M0 IS 4 67 ' 16 nated PCBs such as Aroclors 1221 and 12. 84 92 rather than for 1254 and 1260. This likely is .!. 13 to a greater number of reactive sites and !..- 1260 4 11 0 Sl steric hindrance. (3) In the range of PCBs per 1254 4 SO 0 so chlorinated (0.5-10 ng) in the above study, it : 1242 1221 4 4 7S S 70 18 86 8 likely that with lower amounts of PCBs am! ; 1260 0.5 S9 1254 0.5 78 2 6 91 or less chlorinated Aroclors the decrease in DC. 84 recovery is principally due to the increase 1242 0.5 72 10 82 BNCB formed. 1221 0.5 60 IS 79 One of the major advantages of perchloric.', Hoohtr Chemical SbClj. 1 Quantity calculated by comparison of electron cap ture QIC response to 8NC8 vs. response to OCB refer* enee standard. tion in determining minute quantities of PC:' is the inherent increase in effective GLC doc tor response. Contaminated SbClj, as descrir here, would preclude its use in many of (lit cases. it is conceivable that PCBs could be associated with the antimony metal employed in the SbCl, References process. No heat transfer systems containing 0) Official Method* of Analytn (1075) 10th E-' PCBs -are used in either the chlorine or SbClj AOAC, Washington, DC, sees. 29.001-29.C(,_ production facilities, and SbCl5 does not come (2) Asai, R., Gunther, R., Westlake. W,, Iw.v into contact with plastics in the manufacturing Y. (1071) J. Agr. Food Chem. 19, 396-39S operation or in shipping containers (Hooker (3) Berg, O. W.. Diosady, P. I,., Rees, G. A. Chemical and Plastics Corp., 1974. private com (1972) Bull. Environ. Contam. Toxicol. 33S-347 munication). (4) Hutzinger, O. \V., Safe, S.. & Zitko, V. (107 \ Two parameters (various quantities and vari Ini. J. Environ. Anal. Chem. 2, 95-106 ous Aroclors) were studied in relationship to (5) Hutzinger, O. W, Jamieson, D., Safe, S., the production of BNCB ns a competing prod Zitko, V. (1973) JAOAC 56. 9S2-9S6 uct of DCB during the perchlorination of PCBs. BNCB was calculated by comparison of the elec tron enpture C>I.C response to BNCB vs. the (6) Armour, J. A. (1973) JAOAC 56. 9S7-99J (7) Huckins, J. X., Swanson. J. E.. Stalling, I L. (1974) JAOAC 57, 416-417 response to DCB. Several factors are considered: (1) In this reaction bromination i= kineticallv RwfivH Au{ft.*1 71, 1974. Ihi* papvr wnj i>rv?nti at the SVh Annual Mrrtir.x ) favored over chlorination. With perchlorination tht AOAC, Oct. 11-17, 1974. At Wuhtntton. DC. * DSW 012474 STLCOPCB4002399