Document b5Eod0zNN1Jzzpr35z4LdE11y

Chemoaphere, Vol.13, Vo.9, do 1085 - 1089, 1984 Feinted in Great Britain 0045-6535/84 13.00 + .OO 01984 Peruamon Press Ltd. ANALYSIS OF POLYCHLOROBIPHENYLS FROM USED TRANSFORMERS FOR POLYCHLOROOIBENZOFURANS M R Cull end A J Dobbs frt '! Building Research Establishment Princes Kiaborough Laboratory Princes Riaborough, Aylesbury Buckinghamshire, United Kingdom ABSTRACT The results of analyse* for polychlorodibantofuren (PCDF) content in two samples of polychlorobiphanyl (FCB) from used electrical transformer* are presented and discussed. In neither sample vas there evidence for enhanced PCDF concentrationa even though one of them had bean subjected to overheating while in service. INTRODUCTION Polychlorodibeniofurans (PCDF*) are trace contaminant* of commercial polychlorobiphanyl* (PCBs) eonnonly used in electrical componenta such a* transformer*, where overhesting cen occur. There ie e coneidarebla amount of evidence that PCDFa can ba formed when PCBe sr* heated which i* of concern becauee PCDF* era more toxic than PCB*. This paper present measureeientt of PCDF concentrations from two temple* of PCS taken from used alactrical transformer*, on* of which had bean overheated while in eerviee. EXPERIMENTAL The preciee origin end nature of th* two PCB sample* could not b* established although CC'KS analytic with computer aetching of th* major ion* to the National Bureau of Standards library' gav* either Aroclor 1254 or Aroclor 1260 aa th* nearest equivslents for both aaaplsa, Ths transformers were spparsntly made in 1969 and 1977 and th* newer one had overheated while in service. Th* PCB sample from this Overheated transformer (PCB-A) contained a suspension of fin* black particulate* while the other PCB sample (FCB-B) wee almost clear. About 15 g of PCB-A and 25 g of PCB-B were each diaaolvad in 100 ml atoplas of aromatic-fram hexan* (AFH) . In th* cat* of PCB-A no attempt ws* made to exclude or aepereta out the first black particulataa, but aa hoegeneoua a mixture of oil and particulataa a* possible wet ueed to prepare th* initial solution. In the claan-up procedure 50 ml burattas were plugged with acatona-wathed , gleae-vool, peeked with (bout 22 g of activated basic slumin' (Uoalm Grade B Akt l) under AFH, and topped with cleentd anhydrous sodium sulphate. Th? basic alumina used wet activated by hasting at 440C HONS 1086 overnight and then stored in n oven set at 2)0C, Just prior to use, the required anourt of alumina vaa transferred to an oven kept at llOC containing a bed of ailica gel The anhydrous sodium hulphate n cleaned by Soxhlet extraction vith acetone, hexane and dichlorooechioe prior to use. Allquote (l ml) of each of the PCS eolations -at removed end spiked with 10 u! of e 473 j pi>0 octechlorodibemofuran (OCDF) solution and 200 jl of a 64.& ppm solution of 1,2,3,4 tetrtchlorodibenzp-p-dioxin (10001, These apiked aolutiona were blown dry under Kj, tedissolvtd in AIM and transferred to Che alumina column vith copious amounts of AFH washings. The column* were first eluted with 80 ail of a frathly prepared It CH2CI2/AIH solution and then vith SO mi of a freshly prepared SOX CH2CI2/AFH solution, goth 21 and SOX elutions were separately reduced on Kuderna-Danieh apparatus with final reductiona to dryness under a stream of nitrogen gas in screw-topped glass vials. The residue was redissolved in xylene end etored at -1QC prior to analysis. Anslysis wsa conducted using either a Finnigan 1020 CC-MS with ona single-ion monitoring programme for the PCDF* and the 1,2,3,4-TCDD, and another for the PCDDe, or High Freesure Liquid Chromatography (HPLC) using a Zorbax ODS column and CV detaction. Details of the analytical conditions are given in Table L. The concentrated SOX friction from the column clean-up proved reluctant to dissolve completely in xylene for CC-MS analysis, so small Amounts of CH2CI] and ecetona were added to encourage complete eolation, Table 1 Analytical conditions Instrument Detector Colvan Oven Ftnnigsn 1020 CC-MS LDC. HPLC. (Milton Ktiy) 70 eV ionizstion potential Manifold temp BoC Separator tamp * ilOC UV detector: 245 nm 30 a x 0.2 ct ID SE54. WCOT on silica quartz 13 cm x 4 d ID Zorbax ODS reverie phzse column. MeOH at 1.0 cml/titi liquid phase Injector temp 100C Detector temp 100C Temp prog: Injection at 90C; beltiscie rise to 21SC'C, temp grsdisnt 15C/min to 300OC for 12 min* Ambient It la difficult to confirm the presence and quantities of the various PCDF isomers vith CC-MS because single isomer standards for these compounds, spirt from OCDF, ware not available at the time of enelysis. The presence of PCDFs we* confirmed both by comparison of retention times with those from an UV-irradiated solution of OCDF in 1:5 methanolt xylene solvent and by euaeuring the ratio between Che intensity of (aeh of the PCDF isomers' main cuss ion (H*) and its daughter fon (M* -61), created by the loss of e COCl fragment. Irradiated solution! were found by CC-MS to contain at leaat two isomers of pentachlorodibenzofuran (PnCDF); three isomers of hexechlocodibensofuritn (HxCDF) end one of hepcachlorodibeniofuran (HpCOF). The retention times of these Isomers relative to 1,2,3,4-TCDD were measured. Injections of the Irradiated aaepla gave HOHS 21*3*7 10 B 7 M*/K* -63 ratio* for PnCDF, HxCDF, HpCDF and OCDF iaoraer* which .ere corjutent with the literature valuta1. In chit atudy all the peek* identified as PCDrs in trt ?C8 aampltt had the expected relative retention tinea and W+/H* -63 ion ratio*. The PCDFa and PCDD* were quantified uaing calibration curvet conatructed from the GC-MS analyait of a aeriea of mixed PCDD/PCPF atendarda, containing 1,2,4,6,7,9-end/or 1,2,1,6,8,9-HxCDD; 1,2,3,4,6,7,9-HpCDD; OCDD and OCDF with 1,2,3,4-TCDD aa the internal etendard dieeolved in xylene. The PCDFa detected in the PCB temple were quantified uaing a aeriaa of atendarda containing 1,2,3,4-TCDD and OCDF. Valuta for the concentration of tttrachlorodibeniofuraa (TCDF) preeent were calculated from a calibration curve derived fron l,2,3,i-TC55 but adjuated for the relative raaponae factor* between 2,3,7,8-TCDF and 1,2,3,4-TCDD found fro= the literature2. The concentration* of PnCDF*, NxCDFa end HpCDFi preaant in the cleaned aarplai -era detarmimd by interpolation between TCDF end OCDF. A particulete-rich temple of 10 g of PCB-A taken from the bottom of the container after attlemenc, vat diaaolved in SO ml AFH and centrifuged to atparetc out the particulate matter. The perticuletee were Soxhlal-extracted for 8 hour* uaing SO ml of a 1:4 fixture of xylena and AFH. The reaulting aolution via concentrated and then blown dry under e steam of clean nitrogen gee. the retidue wee taken up in AFH, column chromatographed and analysed at described above. Mo PCDDs or PCDFa were detected. RESULTS AMD DISCUSSION Recovery efficiencice for tho 1,2,3,4-TCDD end OCDF tpikaa after colucn chromatography wart calculated from HPLC enelyaaa; the rteulct era praeentad in Table 2. Tae clean-up procedure used gavs tha moat coniiatent end highett recovery from the eeveral alter-itive* that wera examined. Table 3 contain* the reault* of the CC-HS analyses expressed as . g PCDF or FCDD per g PCS. Table 2 HPLC-deterained recovery efficienciee for l,2,3,WcDD and OCDF tftar column chroma togreph Chemical 1,2,3,4-TCDD OCDF FCB PCB-A PCB-B PCB-A PCB-B Initial mount of epika added HR 12.9 12.9 9.5 9.5 Poat column chromatograph,/ recovery determined by HPLC vg 11.1 10.7 7.3 7.3 He covery afficiencia* BBT. B3Z 772 772 Sample PCB-A, the overheatad PCB, ha* marginally higher PCDF concent,ations then sample PCB-B but *11 the PCDF* era within the range quoted for 10 Aroclor end Clophen PCS temple* in the literature^ Ces* Tebl* 4) end eince PCB-A end B hed different origins the difference* in PCDF concentration* cannot be coneidared aignificant. HONS 214348 1088 Table 3 Results of GC-MS *nalyie Chemical Retention cine (relative to 1,2.3,4-TCDD) PCB-A (l.g g_1 PCS) PCB-B (p* g-t PCB) TCDF TCDD PrtCDF PnCDD HxCDP HxCDD HpCDF HpCDD OCDr OCDD 0.94 0.96 0.9B 1.0 1.14 - 1.39 1.47 - 1.19 - 2.45 - ' 0.06 0.05 0.09 pike 0.46 NO 0.53 0.59 m spike ND pike ND NP 0.04 0-07 spike 0.15 HD 0.13 0.16 ND tpike ND spike ND Hoc**: 1 The analytical error eeeoclated with there reeulce, attimatad free repeat daterminertone ia 46 par cant for the TCDF and TCDD end 62 per cent for the tixCDF and KxCDD value*. 2 ND " not detected. The detection limit for TCDD and TCDF vaa 0-04 pg/g FCB; for OCDF . 0-06 pg/g FCB; and for OCDD 0.05 pg/g PCS. 3 'Spike' indicatea that only the apiked chemical vat detected, the OCDF contained trace* of KpCDF. , Table a Measured concentration* of PCDFt Total FCDT (pg g'1) Range of literature concentrations (mean) PCB-A PCB-B TCDF PnCDF HxCDF 0.04-1.4 (0.34) 0.2 0.1 0.1-5.0 (t.l) 0.5 0.2 0.02-: (1.0) i.i 0.3 Thua there 1* no evidence for tlgnifLeant PCOF production in the overheettd transformer. The pretence of black particle*, assumed to be carton, in the overheated sample indicate! thet during overheating, condition# ineida tha trentforner v*r* reducing tether than oxidising. Thi. ie coneietent vlth the observation of no significant PCDP formation from PCBt because tuch HONS 2X4349 1069 raction* would requir* oxidizing condition*. Neither PCS sample contained detectable concentration* of PCDD* and it is noticeable that even in fire*, where oogen ia presumably more freely available, PCDF production predoninetee over PCDD formation4. Retulte of PCB analyata from an i.ntact capacitor involved i.n a fira<1 have been reported giving comparable PCDP concentration# to thoae in Table 4 wheteea residual fluid fron a ruptured ctpecitor involved in the fame fire hed markedly enhanced PCDP concentration!*. Thit support a the epeculetion above and implies that provided the PCB-containing vessel does not contain air end remains intact end mealnd during overheating, formation of PCDF will be nagllgibte. CONCLUSIONS The PCDF concentre!Iona in <vo samplaa of PCB from used tranaforset*. ona of which had ov*rheated, ere within the range of values previously published for the indue trial product. Theta results indicate that in thit case there it no evidanca for PCDF formation during overheating. This i* the expected result if the overheating occurs in tha abstnea of oxygen end impliet that if electrical components remain intact and taaled while in use end during overheating, conversion of PCB to PCDF should ba negligible. REFERENCES 1 Buttr H R and Boaahardt HP. J Alt. Off, Analytical Cham. 59 (3) 562-569 (1974), 2 Taylor M L, Tisrnan T 0, Garrett J H, Van Net* C F end Solch J C. In `Chlorinated Dioxin* and Dibanaofurana in tha Total Environment' - Ed* Choudry C C, Kieth L H, Rapp* C. Buttarworth* 1983 p 123-164 and 'PCDD* end PCDFtt Formation. Occurrence end Anelyaie of Environmentally Hettrdou* Compounds' - Buter H R, (1978). 3 Repp* C, Nygren H, Buaer N R end Ktuppinen T in 'Chlorinated Dioxin* end Related Compounds*. Ed Hutciager 0, Frei R W, Herien E, end Pocchitri F, Ftrgaaon Pres* 1982. p 493-314. 4 Sehecter A. Chemotpher* 12, 669 (1983). 3 Jtntton B end Sundetrom G in 'Chlorinated Dioxins end Related Compound*'. Ed Huttingtr 0, Frei R V, Harlan E and Pocch(*ri F, Ftrgamon Fr** 1982. p 201-208. (Recalved In UK 3 June 1984) * * ' i , 1 ' j , 1 HONS 2X4350