Document gD3ZOmVNrz9rv44GbyzNvOg1q

=J jV T Jm21- *" li. (*). nil ' C' ' * . SUISCfOU: j. V '! '."i.JTM LA*"I: ' ""crl,.. " 1. UJCHEL. L. a. LOCK! A A Ifiltii r *w. . i a7*>. * Uil-HmmI Mnwfmii, I. KaMrhlon ud "Ytw OH" fcf |i .Nw>tw>. Mimw Kiiiihm f\ i1 Vohiim Mwi\ I*"1*** CtS'm mt fWavMMW Vi ___ w>pi>*n]r** irusi are widely distributed in the wvitonnut, upteiiUy in the tiinw* of fi*h. wild lives and nan. They are also known as the causal agent of a pass food poisoning which occured among sore than 1,(06 persons in western Japan in 1948. The disease induced by this poisoning is called *Yusho*, namely oil disease, because the patients were proved to have consumed a commercial rice oil contaminated with a large amount of Kanechlor400 (AC-400), a brand of PCS of Japanese make (TSUKAHOTO et al., 19491. feeding experiment using broilers demonstrated, however, that the rica oil taken by patients with Yusho were twice or more toxic than expected from the Mount of KC-400 presented in the rice oil (IKCDA, 1972). This suggested that the "Yusho oil* must contain some more toxic substance or substances besides PCS. VOS et al. (1970), on the other hand, found by the chick embryo assay that toxicity of commercial PC* preparations is greatly affected by theix contaminants, polychlorinated dibenxofurans (PCDf). Bauer et al. (1941) alao demonstrated that structurally related polychlorinated dibensodioxins (PCDD) were extrasmly toxic and acnegenie to man. Approximately 1 ppm of PCDF was estimated to contain in KC-400 (BOACM and POWRAirrz, 1974) . All these facts necessitated an analysis of Xanechlors and the rics oil taken by patisnts with Yusho for their possible content of PCDF end PCDD. EXPEXIMEVTA1. In order to separate PCS, pCOf and PCDD from the *Yuehc oil*, 5 - 10 g of the oil wee taponiictd with ethenol containing an excess amount of sodium hydroxide for 1 hour at I0*C end the n-hexane extract of the reaction mixture wae chromatographed on a column of silica gal (Wakogel S-l, activated by heating at 110*C for 3 hours) with 100 ml of n-hexane. The eluate was concentrated to a small volume. ! G O HONS 017004 Thu cor.cintiitd eluatc and tiaflit of KC-100, 400, S00 and (00 (0.1 - 0.S tit thioMto^nplitd respectivaly. on column of alumina (100 noth, Htko Pure Chaoical Tnd. Ltd., S 9} activated by nesting ac VM*C for 12 hour*, using at alucnta, first 20 al of n-haxane, than 120 al of n-hexnne containing 20 t carbon tetrachloride, 10 al of a-hasana and finally 30 al of n-haxane containing 20 t methylene chloride, tha last aluata was evaporated to drynass, dissolved in n-hexae and subjected to gas chromatography with en electron capture detector (Beckman GC 72-S) and 9aa ehroaatography/aass spectrometry (JKOL 0*1001 for qualitative and quantitative determination of KOF and PCOO. Whan tha separation of PCOF and PCD0 fro* PCI was not adequate, a siailar chromatographic fractionation of tha last aluata was repeatad using a smaller amount of alumina (2 g) and a emallar volume of the same solvent systesw, prior to tha determina tion of PCOf and PCDO. amount of Individual di-, tn-, tetra-, penta-, hexa- and hapta-chlorodibanso- furena was cslculatad from raspective peak Mights compering with those of tears-* pants- end hexa- chlosodibensofuians which war# synthasiied by chlori nation of dibenxofuran. in this determination, the individual peeks were confirmed by GC-MS and their heights were assumed to have the seam sensitivity, the total amount of PCOF was also determined by the perchlorination method. The PCOF from each sample were chlorinated to octachlorodibentofuren (OCOF) with BMC reagents (HUTZIMCEB at el., 1972). The amount of OCDF was determined by ECD-GC comparing with standard OCDF (Analabs, Inc.) and converted to the asuiunt of tetra-, hexa- or hepca-chlorodibentofuran which was the sum component of PCOF in the sample. Possible formation of PCPF from PCB during the saponification process was denied by compering the determined amount of PCOf in RC-400 with that of PCDF in a sample of saponified vegetable oil containing KC-400. The recovery of PCOF in the whole analyti* cal procedure wee examined by analysing 10 9 of vegetable oil containing 1 pg of PCDF, finding more then 90 l. RESULTS AND DISCUSSION Presence of PCOF was demonstrated in ell samples of Kanechlors end *Vusho oils" analysed, as shown in Table 1 ead Figure 1, but PCOO was not detected. As might be expected, kanechlors with higher chlorine contents such as tc-SOO end (00 contained more chlori nated dibensofurans then those with lower chlorine contents. The concentration of total PCOF in RC-400 it ' C a lc u la te d from peak h e ig h ts , b* C a lc u la te d by p e rc h lo rin s tio n m ethod. MOWS 017005 y.XN 1M Figure 1. GC-MS of Tetrachlorodibensofuran A: Kanechlor-400, B; "Yusho oil", Ci Authentic specimen. was astinatad to be 18 ppm, the highest of the lane* chlors tested. GC peak patterns of PCDF present in three saaples of the TYusho oil* were very similar each other consisting of tri- to hexe-chlorodibenso* furans and their concentratione were 5 ppm. The PCS concentration in the "Yusho oil" analysed by us was about 1,000 ppm. Therefora, PCDF concentration in the PCS in the oil was calculated to be 5,000 ppm. In other words. FCOF in the oil seemed to heve been 250 times concentrated as compared with its concent ration in KC-400. VOS et al. (19701 showed in their chick embryo assays that PCDF was far more toxic than PCS. According to ARAKI's (1974) animal experiment, hepatic eniyme-inducing activity of FCOF was 170 times that of KC*4QQ. in view of all these facts, the authors consider that the toxic role of this specific type of contaminants in the `Yusho oil" should not he dismissed in the causetion of Yusho, even if their quantity was small in the "Yusho oil*. I* REFERENCES 1. MlWl, t.s Fukuoka Acta Medica 65, (1 (1974). 2. BAUER, H., K.H. SCHULTE, and U. SPlEGEL&ERki Arch. Gewerbepathol. Gewerbehyg. IS. 538 (19(1). 3. KUTZINGER, 0., S. SAFE, and V. EITRO: Intern. J. Environ. Anal. Chen 2, 95 (1972). 4. 1KE0A, .; J. Food Myg. Soc. Japan 13, 359 (1972). 5. ROACH, J.A.C., and l.H. POMERANI'Z: Bull. Environ. Contan. Toxicol. 12, 336 (1974). 5. TSUKAM0T0, H- at al.: Fukuoka Acta Medica, 60, 496 119*9) . 7. VOS, J.G., J.H. KOEHAN, H.L. VAN OCR MAAS, M.C. TEN NOEVER 0E BRAUN, and R.H. DE VOS: Food Cosnat. Toxicol. 8, 625 (1970). / C&f-ny fOMUfioa or poucHLoamTD oisbizofuiaib u .&) non m mousis or pce Sans Rudolf Bueer and 8oaa~Paul Boosfaardt Swiss Federal tsaeaseft Station CB B820 MUenewil, Switzerland and Ckrtatalfw lafpa Departaent of Offal o Chealstry. Oniveralty of OSes S - 901 ST Oooa, Sweden (Beoetwed la UK 1 Jnwarf 197Sf eaeepte* for pablleatiea 5 faaueiy 1918) awtwtui MuMittiWrt <TCM) M ladoatrlal .naaicala or via* taaittm kM to Bo widely distributed in tbs enwlronaent. The eoMsrcial product* are eoapiex ailture# of at leant SO different substances. Duo to their estreat lability in ocoloflcal zjataaa. the uaa of rao la now restricted to ooalod oyotoaa Ilka capacitors and tnoafomn. 8owm, prior to 19T2t KSa Sm to their food thorns! aad rh--leal otabllity wore used la a variety of other appUeatlsaet la pnrtioular the uao as boat transfer fluids and eaatinf worse vaa aaaoolatoS with tha ftwari stability. toy elaaaly wliiai to the Kfe aro tha polybroal sated biphase (Ptte), ocas of Mich teas boaa asst as flane retardant*. Of spaniel interest la firs Raster W-d or Fire Haster IM, iSM van 8m ta coaolat aalaly of 3<4.9.2lt4t>9,-hoaabwablpMarl.1 lataalcatlM with this n has booo directly aaaoelato4 with a vide-epreed aflenltarsi calastlty u RMla (0U) la WV74.2 10* HONS 017006 Prt or too POo foood la tho oawr at &M tncap*<l opan Application* and cloood acno probably durl&f incineration and burning. > rltacn and Roam 5 f* ,.tu tta t iiaal atebillty 0/ PO*f laboratory aiportaonta and -laotoUcal * ilationo ** tAat at taptnnrtl than 800C PCPo art thorwcdynaolcally unatablo. ITia um d*co> position iwiwta war* fovnd to b* C(a), CO, :02> dCl ud Cly In 1970 Voa ldaati/lod polycblorlnatad dioaoaofurano (PCBfa) aa toalc u*pt*ri*.iaa la iunfiM PCBo at tb* pp^Uwl, and tha imm haa alao boon roportod for Aoarl-v. and *-y--r*~ PCBa,^'6 tha toxie offoeta of tha PCBTa art wry flailat to tbooa roportod for tba polyahlarloatad dlhaaao p dlorln* (PCBDo). la oth oaaaa, tha 23*7.B*tatrachloro c-Tpoupda ara tha aoat toxic looaara, haalaf LS^-valuaa in tha rang* of 1-10 ttj/kf v^uinoa p 1**7* Life* KB8af tha PCXJPo haw boon foaad to loeraaaa eytoehroaa P-430 activity and a nuobor of dro* - **bolislaf aa^raaa. 8.9 Nnlta i| 10 haw roportod loeraaaad aaounta of PCfifa la a aaapla of PCI aaalad with al* la a llaaa taha aad haatad at 300C for two wa <. Za thla payar* a report oa tha formation of PCWi fro* tha pyrolyola of a iinlil K1 (lml 12V) la tha praaaoca of air la a aaala* quarts tuba. Ia oiin a aOa4y taa aaahaalaa of Us raaatlaa, iso hsoa laaluda* ladlclOual K loossooa aa aa4al cooHuli, 2,4,2',f-Tatra- , 2.4,9.2',4',V- oa4 2,4,,2'.4,.<'-liacMobtphahyl son tnfsnS mwlll aa aaaaa aatha4a.U traaloir 12V aa* Plraaaatsr 8P-4 non fna coaaarcial ooureaa. laalfli of tha ta haaahlarahlyfcaapla oa (Isas capillary coluaaa abocaO slnpls Osaka. Iiasi>1 lasilpbiaqrla talag prsnat as laparltloo at local* of 1* la 2,4,9.2'.4',V- as* of O.J0 la 2.4.*.2'.4'.*' tulliwllilsarl. Isao of tho Ia41cl4aal PCI lsoaars or (rsclor 12V 414 aoatala (ataatahlo qaaatltlss (< 0.001-0.01#) of PCSPs, palpchlariaata* banssass, OttaflsOsa oa tlpa--/I si s. itaaisr* aalatloaa of tha ra* mo prop--* * --sotritlaao of 1 as-- 10 sa/al la a boisai or bsasssio. HONS 017007 : alsl-aspoulea wn Mi* fro* quarts tuoing jaing a gac-oxygen v ~ . r^rt* mpoulae war* 2VJ0 m, the internal dlaaetera '.4 m and . . j 0. ;i. i :r,< -eyringa* 10 pi of a standard solution (correeponding t- 10 or l"0 i ** ,U, ..i iAto tb* tip of aa aapoula. i/tar coapiswnporatioo cf the jolvant -*tn the it i j* .*M vacuus, the aspoulee were closed using a v*7 eaall nMi ears aa tn r.^t ;j hsat the tips containing **e coapounde. The volua# of 0.3 ai air in *n i. approvinately constant, the solar ratio# of FCB to oxygonin an aapoule varaabout 1: T5 ind l:7.` for asouste of 10 and 100 ug of PCB, reepectlvely. Iks quarts alniaspoulee were placed in an rleetrleal oven (sax. teM* i000C) preheated to tesperaturea ranging fros 550 to S50C. Yfce tesparatuxo oaa controlled si -: a thetsoceuyle la an aapoule plaead next to the easple*. Heating period# (tlaa betweea placing a easpl* aspeule into the even end zonoving it) oars kept coaatent at 60 sec. The tlaa required IS heat tha mini aapoulee up to a teaparature olth.n 2QC oftha proeat omla* anaaareoad M see, the tine of a easple at actual pyrolyala tespermture thua around 3 eec. After tewtldi aod lasedlate cooling to rocs tespermture, tha Upe of tha Mpeslee aase brakes and 200 *1 of beaaeae added. A 2-til aliquot vae uaed directly for --ljete farther Allatlas ass roqairad for eesplea with Uw iecospoaltlon of PCBa to overcase mrlaail^ effect# of tha GC-fV eystes. Sosa eesplea aero cieaned-up on sa aluslan-aderocolUM, ueing 2!% and Y& Mthyleoachlorida la a-bexane to separately elute PCBa end PCBFs.12 Inti yeas wore carried out os a Flanigan 4000 quadrupole GC-fB lactrunent equipped wits elaotfos Capbdt las aasroa asd s llnnlgwi 6111 data-syetea. 0T-17 glace caaiUary eoiwaa *f 2) asd 50 a length (0.)4 as ZB) sera coupled via a pletlnus interface leading dirsetly lata ,tha les-saoana. ftp eelMi eoadltlosa wore as followat 100 2 sis teotfcexasl. l5/ain ta 145. 5*/sis to 2J0C for tha 25 0T-17 colmbi 100, 2 sin leotharsal. 20/sts to 160, 2/sls to 2J0 for tha 50 a OT-17 eoluMj belies carrier gsa at 0.30 mA 1.40 at, tu^es lively. Isaplea vara aplltleaaly lstrodoeed with tha vaporiser at 2tOC. HONS 0X7008 M m|1h mm directly Mely-- oo the 25 a OT-17 eolena by nptlUff eeaMln^ ' ;opUU MM ejectM 55-4<0 in 1.4 #ee) uric* the d*U eyetM. After J .acquisition, Um MmmM iftctn (a* t 1000) mm Marched for PCBe, PCUTe and other weaible r*c ion prthicts by ramlnc m4 ylottlaf mm chroMtocram. for oytJjnl leaser WfiltUoft com of the Min wort reandjoed on the 50 a QT-17 |UN iimIIItt ml-- M-- lyaelfio deteetio (mm fnpsntngnphy) of PCBTa was carried out by coatiaumialy aoaitarlac aolacular lone at a/e 2C2, 236, 270. y04, ;36 and 37? for en no- to baaabloro e--o--s. Reculta rw^MM or 2.o.5.i-,d-.5'- a* i.a.s.2'.. .& la IhhlM 1 al 2 roport tha naolta or th pyrol/tit at too hotochlonbifao?U at ta^antam rfdlac rna 550 to 450C. Im Indicated, JT* of tfca 2,4,5.2*.4'.5'- ad 22* at tka 2462,t4'6>~boxae!iiorobiphanyl dacoapoad at 550C. Both enapoimdo am hnatml ta > 90* at 450c. coapx.to aoatnotioa oeoumd at 700C tad obo-to. Bo ln--rloatt". at thoaa ti-->1 oroblphaaTt I or daehlorlaatloa to laar ehloalaatod blphanpla tat oboonod to thla thUo 1 * -* o .i'.f-a---~>.<.o-n,i (too ja Tbagntw Deeoapoaltloa <*) FCSfs foned* (jl) totra-CV oente-CDf 5 5T 1.6 MO BJ O.J *58 91 0.2 TOO >99.99 < o.oi 0 >99.99 < 0.01 * 1 tbtn> ad 2 pala Cl# laeaan (taai 0.2 0.1? 0.5V 0.1 1.1 0.1 < 0.01 < 0.01 oaa pataaindlbaaoron-- (tatia- aod pnta-CDPa) rotwd at tapwax at 550-*5d*C. tta lmla aaia 0.2-1.4* ad 0.1-0.5* tat 2.4.5,2',4\V ad 2,4..l*.4*.t'biaaiibotlpaapl. it toaparatma at T00c ad ohm, ooaplata dootnotlaa at thaa KWh MOUTH, a di-, trl' MOWS 017009 Tdftl* 2 fwrolniM at (100 u*) in auM*tj B)n| mr)r iMNntUN Cq) w 600 50 ;oo 950 OaconpoaH ion _ (> 22 ?o >4 > >0.09 (<)PC3Pa f-raad* tatra-CDP 0.9 0.1 0.* 0.55 < 0.01 0.5? < C.CL < 0.01 < 0.01 1 tatm- and 1 panla-CDP iaonar Torsad 7h* foiMtion of panta- and Kexachlorobansena whs obsarvtd with both PC3 1sonars it -hijtaaparaturaa of 690C (G.Ors) and 70OC (J.'-l*). Tk. i-iwar rir.jt*d t-?nanaa way r scapad datact ion (alution from capillary ooluan prior to starting of ?S scar.- -tg). ?u*.h, pant*- and (MiMklorobmtiM warn daeovpoaad at S*OaC, aso 38 ata aoo 3S *c min ttfM* 1 Parti*! mm ff^antntrnM (90 0V-17. / 504 and JJ3) atoovtj* alution of tatm- an* pant* CJWa in a) pyrolyaad 2,4,5.2',4*.S and b) pyrolysad 2,46.2',4*'6'-haa*etilorofciphaftyl. 1 2,J,7,8-tatr-, ? praaMabl/ latra-CDF. HONS 017010 la Vigo* 1 aaaa ftapaao*ma rf Iwtt bajaattontlphanrla pyreljtad at 600 an th* Imtiop of UU^ and pitp-CSFp op tte 50 Cf-17 plaap capillary colupn. Tha r .jin of ,, ,vg,4 5 2 4 teaatelotabipte^* 414 ylold om totn- nd two ponta-CV .^jaora. iy coppafloop with aa aotteatlo itaodard, it could bo ohown that the Cl^-itooer i tho very toxlo *i37>taUn HI* Coppo^ppatly, 00 nmnMnt la involved m th* eyeiitation foaalaf Uda iar< Za tte oppo of tte two poat*-C9Fo, it i ovidoot that a rnmnfpir aupt bo isvolvod. ftp aearopmafod product . houli bo tho 1,;, 1, i-JJF, tioppl wot* ip aooiod to oatpbllafc thp identity of thooo panta-CDfa. 2 isomors ftp pyrolyplo of 2v46t2'(4'*4*-hpxaehioro0iphonyl did yioid on* tttra- and on* p*nta- GW UMPf. ftp d.k4ipaapr pro--hi y ta tha l3t79-totra-CDF but thio cannot oo -onftra*d teo to late of m out*--Up ataaiaat* ftp pmli ffli Ip fotapd in a rtortanfaont and it ten Via iadlpptlac aly iatropniacaUr eyolipotioa proeooooa. tDlalijLOPUP Maalm UM la a imlal ta eaatalalag trl-, tatra-, paaU- and Muoklarokl- paaagla aa tta m)m aittaali. CC-gg aaaljala malag tha 50 a (JT-17 flaaa email1mar eolujp tha ptapaapp of ftp 2,4,},2\4', V- oateUtutod looaar on opo of tte mJpp MOMS 017011 teaeklsnUfkavla. > 2.4.<.2'.4`.<l hnh1imiblhMyl 4>uM KltfnUuw u IMl* } rmlnim at isrelcr 12V < Uni MullII Zsi 550 " MO - 650 * TOO m 50 tat. XmI IX 10 IX 10 IX 10 IX 10 IX Poeoap* JSL 12 ao 49 90 90 9 >99.9 >99.9 >99.9 oaahLnod valaaa of all 0 j __P00F. noM* ( MC- ____ 41- ttfr- ____ StlOe_ 11 0.25 0.02 0.69 0.40 0.90 0.70 0.75 0.20 3 4 0.10 0.40 0.70 0.60 0.12 1 < 0.01 < 0.01 0.01 0.10 0.02 0.12 0.29 0.1s 0.29 0.J5 0.29 0.12 0.09 0.12 < 0.C2 a a e < 0.01 < 0.01 < 0.01 < 0.01 < 0.01 < 0.01 < 0.01 < 0.01 <0.01 < .02 < 0.02 < 0.0a <0. < 0.02 < 0.0a a pear la Tabla J tha rviuiti of tha pyralyala of Aroolor 1294 la tha t^mtan aa^i af 550-d50C and lavala of 10 and 100 uc wa raportad. At 990"690C, tha daataoaalUm at tte IX ac^laval nn|d froa 12)1 to 90)1, at ttaa 10 * laval froa X to 99fL Amlwr U94 waa eoaalataly daattajid at taaparataaaa of 7XC aad aboro. lhaaa japoapaalllaa rmlmm mm hmmt on tba Hawaiiarablphawyla praaant. lha daeoapoaltlon of tha lovor cklorlaatad Mfhaafla aa oaaokat hid** that tha atatad Talaaa, So 000La- or hasMhlowbaasaaa aa dotaotad at taofaaatusa* Afala* loaai efcloTlnatad bonsanoa would 'aa ooeopod dotaatloa. Hoao 1 to paata^CSPa mrm found at 990-650ftC at level* rar-fln* froa 0.1-0. <, Tha HIM awM fTM4 MS* St 550c. ThUa( nto eooalOoratloa tho aaount of PCS racowrad aftoa pfttlnU* tho total yiald of Wfi ruwaod froa 3-29J6* la addltloa to ILWI, pair* okloaiaatad Hfhaayltlt aa paaalbla pnmnon an idoatifiod at lawls of appraxlMtaly oaa fifth of tha aaat of KVa paaaaat. la ?1#bo 2 aaaa fra#watognaa of pprolyaad Aroeior 1294 (IX tif at 600, ola^a**u9 ia>l) oa a 90 a Of-17 |laai capillary coluaa art ahown. Tho olotloa of 09 to )0 aa)ac d aoro than 29 alaor KSPa can bo oboorrod. fhojr raafa froa tho aoao (all 4 tlaoratlaalli poasibio laoaoro) to tha pawta CPFa. Since 2,492* .4* ,VttMMiw>l)hayl la oaa of Ida HONS 017012 MONS 0 1 7 0 1 3 FI#m 2 Mm flHWUm- (50.0V-17, m/. 202. 2*. 270. J04 0 }) wool,* .Jution of t0 MnB. *1-. tri-, totn- wo p.nl.-CL, fro. irrclyioO irooior 1254 (600c). ,no. ^ 4l_, 4 /( tl- wo uuv. I Vi ponts-CBTl 0.5 . NO. 14.atlSie.U0H. 1- 2,),7,-l.tr.-i } 4. p,, ...Wo r, , 2.4.5.2\4\5'-l*U<>roblftwWl, 2- Kw Wore Wtn-CW frr 2.4.6.2-.4-.u'-t..ttchloroL MOttld luU. j mu Htn<in IIM10? I254f it la or intaraat to confirm tha irrnn of 2,3,7,S-tatZfr4V and of tte two panta-Offa in thla coapiai nixtura. Tha 2.* . -atra-c;p la not coaplataly aapamtad on thla coIum free wo thar tatra-CCF u.iar. Tha fnraad fwn 2,4,6,2',4',6*-haxachloroblphanyl waa not f::-.d to > prasant. tn eaaa of tba anjor iaonara, tha dlacrata PCIFa could ia idantlflad by running coaplata II aaau ipactn. All aaaa apoctr* ahowad intanaa acitcular lona with tha trpactad Ion eluataring iua to tha eftlorina xactfpaa v.d *h charictanatle fr^antati n ;`jc i tit fcrmtlon of H*-C0C1 Uld Il'-COCl-Clj. Additional/ wo hava atudiod tha pyrolyala of 2. ' ,V-tatrachlaroblphanyl and ,,i rinaaattr BP-6 (mialy 2,4,5.2,.4'.5,-ho*acraaebipnnyU. Tatrachloroblphany. fomaa at 550C ona 41- and ona trl-CBP iaonar at lavala of 1.6 acid 2.5*, raapoctivaiy. Tha tri-CB? auat ba fomad in a raarranganant. At 700C, tha tatrachlorobiphanyl waa conplttaly laatruyad and no KVa war* found. Flraaaatar BP-6 waa also coaplataly laatroyad at 70QC, hut at 6QQ*C oaw conpouada wazo foraod, ona of which could ba a tatrabrtxodxtanaofuran (It**440, Br^; aajor fra^ant othar fragrant* M*-8r nd fl*-C0flr;. Tn valogy with tha eorrtrpandlac FCB iaonar, it aa-- noat 1Italy that it `a tha 2,5,7,6-tatra-BDF. Tha Mount fomod la about aqual to tha Mount of tatra-CBP fornod fro* tha eorraaponding FCB. Concluaiona Tha raaulta ropo^tad Ln thla payor auggwat that unc.-r.:r^i:d eurr..ng f P7J* can : m laportMt onalronnMtal aourea of tha hasardoua FCBFa. Although PCDFe hava not y*t taan found in tha arrvlronaont, thay fcawo baan .dantlflad in fly Trus rir.ijtp*. .r.cir.arv ra j l induatrlal hooting fonllitioo.^*** Horaovar, Magayam 11 r-*** found highly incnaaol lavola af FCBFa la tha PCS aaaoolatad with tha Yuaho accidant in Japan in l >69, and era of tha aaia oaapaaaata was alaa tha 2,3,7,9-tatra-CDF. 16 In aplta of bolag partly bannod, largo wlunaa of FCBa ara atili u uaa, iapart allj in traaofataars and capaaltara. in Swodan tha anaunt now ln jaa in capaaltora la ahoal 2500 tonat alnilar anaunt appilas for iwittarland according to aurray by tha Swlaa Podatal F*itv.c Haaith. it ia awidant that all diapoaal of FCBa aiwuld ba earafdlly HONS 017014 cotmUtd i ortrt to mil MtUatUl faaitifll of fatMiiout KXft. Oar Unrootijatlaa alto indlootoo tlat opoclftl procosooo itjr bo ooooeiotod -t ;ccupAtloaol hoaltb rltfct dot to foxntlot of PCDts. 'no of thooo proessso* is th wsilir.4 if >sxirjT *yto-- eeetolalflf K8o. Another proeooo it tho nistm* : iron .si:.# ?<*? jontsinin* cuiitf um. Aeeldontol Ooitlm of notorial lrproMto4 *ith 2,4.5.2' .4', V-hozobronoMpiayl ny yisld <imi contointnf tho Tory toxic 2,57,8-'.str^-aDf. Consequently, tho aoo ;* this conpound so o Clone retardent should bo carefully reconsidered. l.foroncoo 1. C. And.r.Aon, A. MorotzOa. C. Rapp*, s. lUoauoon And a. o.ahlln, -My, -fipUtY 1 Won. Sopp. ToA. m, p. 798 (1775). 7. AMhphOUO, Wi ltin hoOmary I, 0- 7 (1?7"-1. 3. L.raolaaoa < I. 8ooaa, Tra)-| Sflpf. 1, 61 (1371). 4. J.C. Too, J.B. IsooMt H.ip, TOO dor Hm*. H.C. toft Hoover do Brauv end K.l. do foo ri. '7--5. TlUWl if (1970; . - . J.A.G. loaoh m4 1.8. rtmxtau. KOI, te n.. Cohtaa. Toxicol.. JJ, 338 (1974). (. G.V. Aoom, 8.J. IklnUll, B.I.T. Slaoaolt. A.L. Burlinoao and a.a. 31ao6rouaX, 8oV. &. 303 (1975). 7. J.O. HoSlooap, I. Chaa. 8.8. Gupta, J.A. Nooro wd J.A. Goldatoln, Tcm-ol, adbI. TllIlSMl &> (97). a. 1.8. RoCoaooU, J.A. Nooro. J.I. Saaoaoa and H.H. aorrio, Tolled. Anal, pvaro*,-:;.. Sir M* (1976). 7. A. Polaa4, 8. Glow. A.8. boAa, J, Uol. Chaa., 791. 4936 (1776). 10. R. Mlo. J* fekapnt K* dklpoao. s. Mm and 8. laono, Boll, apoiroo. Com*a. Tam*- lla (1977). U. 0. rtolo^r, . Wo at T. HUto. T** "nctri *f Mfl 7*C Proaa, Claoolaad. Ohio (1974). 12. htoor. J. Chroaataaa.. 107. 299 (1979). HONS 017015 t. oil*. P.U <M>1> aad o. ftrtalaor, Cfwtm, J, 455 (1977). I. 1. Juaar and I.-f. ftMaftardt, *- '`t`-~',t`,"'<itara. a. ga. 1a vim (197*). J. lipjiM. IL Imlm and T. Handa, laU, a>alron. Contaa. Toaleol.. !>, ) (197t.. C. lappa, A. Can, l.l. Juaar and l.-P. loaahardt, Chaaoaahara. 23) 0977). MONS 017016 COMMENTS OP THE UTILITY SOLID WASTE ACTIVITIES GBCUP, THE EDISON ELECTRIC INSTITUTE, THE AMERICAN PUBLIC POWER ASSOCIATION AND THE NATIONAL RURAL ELECTRIC COOPERATIVE ASSOCIATION IN RESPONSE TO ADVANCE NOTICE OP PROPOSED RULEMAKING CONCERNING RISKS POSED BY FIRES INVOLVING PCB-CONTAINING TRANSFORMERS 49 Fed. Reg. 11070 (March 23, 1984) Docket No. OPTS-62035 of Counsel: WALD, HARJUADER 4 ROSS 1300 Nineteenth Street, N.W. Washington, D.C. 20036 Submitted to The United States Environmental Protection Age June 19, 1964 HONS 017017 TABLE OF CONTENTS INTRODUCTION AND SUMMARY OF POSITION............................................................ .... I. DATA REGARDING USE OF ASKAREL TRANSFORMERS AND FIRE-RELATED INCIDENTS .................................................................................... 5 A. Inventory of Utility-Owned Askarel Transformers . . . . ? 1. Currant Inventory ........................................................................... ^ 2. Voluntary Phase-Out Programs ................................................... 3 B. Frequency and Nature of Fire Incidents ........ 1. Frequency of the Incidents......................................................... 13 2. Exposure Levels........................................................................... ....... 3. Clean-Up Costs....................................................................................... 4. Factors Decreasing Fire Incidents ................................. 12 C. Retrofilling Experience .................................................................. 13 0. Impact of RPC Data on EPA Initial Estimates...................11 1. Frequency of Fire Incidents.....................................................14 2. Clean-up Costs................................................................................ 15 II. REPRESENTATIVES OF THE BINGHAMTON, SAN FRANCISCO AND CHICAGO INCIDENTS ................................................................................ 15 A. Fires Associated with Askarel Transformers Are Rare . le B. Potential Risks Are Overstated .................................................... 17 1. Causes of Fires May Be Preventable.......................................18 2. Levels of Contaminants Formed IsInconsistent . . 13 3. Confounding Sources of PCDFs andPCDDs .......................... 21 4. Formation of Reaction Products .......................................... 22 5. Exposure and Health Effects of PCBs, PCDFs and PCDDs................................................................................. 24 C. Conclusions Regarding the Binghamton, San Francisco and Chicago Incidents ............................................... 25 MOMS 017018 III. MECHANISMS OF PCDF AND PCDD FORMATION..................................... ....... IV. OTBER RELEVANT FACTORS.......................................................................... ....... A. Substitute Fluids for PCBs............................................... j- B. Risk from Contaminated Water................................................... 31 C. The Consequences of Building Fires................................................. CONCLUSION.................................................................................................................... J4 MOWS 017019 The following comments ace submitted on behalf of the Utility Solid Waste Activities Group (USWAG), the Edison Electric Institute (EEI), the American Public Power Association (APPA), and the National Rural Electric Cooperative Association (NRECA) in response to an Advance Notice of Proposed Rulemaking (ANPR) issued by the U.S. Environmental Protection Agency (E?A). 49 Fed. Reg. 11070 (March 23, 1984). Specifically, EPA sought data regarding the risks posed by fires involving electrical transformers that contain PCBs and the mechanisms for eliminating or mitigating these risks. USWAG is an informal consortium of EEI, APPA and NRECA and approximately 65 electric utility operating companies. EEC is the principal national association of investor-owned electric light and power companies. APPA is the national association of publicly-owned electric utilities. NRECA ;s the national association of rural electric cooperatives. Together, USWAG members represent more than 85 percent of the total electric generating capacity of the United States, and service more than 95 percent of the nation's consumers of electricity.i/ */ A list of USWAG members participating in these comments is attached hereto as Exhibit i. HONS 017020 2- . INTRODUCTION AND SUMMARY OF POSITION In August 1992, EPA authorized the use of a population of transformers containing or contaminated with PCBs for the remainder of their useful lives. This authorization was based on EPA's determination that the continued use of such equipment did not pose an unreasonable risk to human health or the environment, underlying this ultimate conclusion were the following findings (see 49 Fed. Reg. at 11070): 1. The resulting reduction in risk, in the absence of such authorization, was insufficient to outweigh the substantial costs (billions of dollars) involved in such an action. 2. Servicing conditions, maintenance activities, ar.d inspection programs reduced the exposure risks associated with the use of PCBs in transformers. 3. Releases of PCBs to the environment and exposure to humans were minimal. * 4. Accelerated phase-out and other risk reduction programs were not reasonable when compared to the potential reduction in release of PCBs achieved. Prompted by a challenge to these findings filed oy EDF/NRDC and reports of three fire incidents involving PCB Transformers, EPA issued the ANPR in March 1934 seeking further information on risks to the public health and environrrr posed by fire-related events. Notwithstanding the fact-finding mission intended by the ANPR, EPA concluded, based only on tne very limited data there presented, that "PCB-Transformer fires pose relatively high risks, occur with unknown frequency and can result in relatively high clean-up costs." Id. at 11071. These comments and the analyses, studies and materials on which they are based demonstrate that EPA's initial conclusions cannot be supported. HONS 017021 -3- Specifically, in response to the ANPR, USWA6 conducted n extensive'survey of electric utility companies regarding the current inventory of high-concentration, PC3 Transformers (referred to hereinafter as 'askarel transformers'); fire incidents that have involved such equipment; ensuing clean-up costs and other matters.*/ These data and other information discussed below regarding the matters on which EPA solicited data, show that: 1. The risks posed to human health and the environment in the event of a fire-related incident involving an askarel transformer, including any effects from exposure to the products of PCB combustion, are de minimis if not non existent. 2. The probability of fire-related events occurring is extremely low. 3. The costs involved in cleaning-up the two fire incidents identified by EPA as "relatively nigh'^V are not representative. 4. The costs associated with the available options for mitigating or eliminating the minimal risks posed by fires involving askarel transformers, are not reasonable when eorparsd to the potential reduction in exposure risk: or put another way, the resulting reduction in risk clearly would not outweigh the substantial costs involved. As EPA is aware, the electric utility industry has been deeply and responsibly involved in the management of PCBs for more than a decade. See USWAG/EEI/NRECA, Comments and Studies on the Use of Polychlorinated Biphenyls in Response to An Order of the United States Court of Appeals for the */ A report on the survey conducted by Resource Planning Corporation is attached as Exhibit 2 ('RPC 1984 Report'). 12/ 49 Fed. Reg. at 11071. HONS 017022 -4- Distriet of Columbia Court, Fob. 12, 1982, Voi. I at 1-2 (*1982 Comments"). Th# industry continues to be supportive of control measures that can be justified by the body of scientific evidence.*/ But however that evidence is weighed, there has been no acknowledgement of the effectiveness of the PCB control measures already in place. Thus there is growing industry concern that this failure to recognize the effect of the controls in substantially eliminating release of PCBs in the environment and exposure of humans to PCBs has distorted the public's view regarding the continued use of PCBs. Current control measures, regardless of whether one now views them to be justified by the scientific evidence, are effective. The need for additional and more restrictive regu.at.zr.i is highly questionable. Even in the case of fire incidents involving askarel transformers, where EPA believes that certain by-products of pyrolysis or combustion -- more so than the PCBs themselves -- pose some sort of risk to human health, the perceived human risk described by EPA appears to be based on the presence of the chemical (whether PCBs, PCDFs or PCCOs) rather than on the actual exposure levels and the effect of that actual exposure. Overall, the existing evidence shows that fires involving askarel transformers are rare, that utilities */ It is USHAG's overall view that this evidence, notwithstanding ~ prior Congressional and regulatory action, does not warrant many of the restrictions in place today. HONS 017023 5- have program* underway to eliminate any perceived potential risks resulting from even these rare incidents, and that further regulation of this industry's equipment is unnecessary. While the focus of these comments is that no further regulation is warranted because of fires involving askarel transformers, USWAG wants to emphasize that the arguments against further regulation of mineral oil transformers are' even stranger. I. DATA REGARDING USE OF ASKAREL TRANSFORMERS AND FIRE-RELATED INCIDENTS The electric utility industry has used equipment containing PCBs for over SO years. Some equipment, because of its locstion or electrical properties, was designed to use PCBs as a dielectric fluid. This equipment included transformers that would be used in locations where fire prevention was a significant factor. Thus, as shown in USWAG's 1981 survey, of the 39,840 askarel transformers then owned by the industry, some 22,489 were located in the utilities' distribution systems. RPC 1984 Report at 6. Other transformers in use on electric systems, numbering about 20 million, customarily used mineral oil as a dielectric fluid. As a result of manufacturing or servicing conditions, some of this equipment was found to contain concentrations of PCBs ever 50 ppm.*/ *'J Approximately 10% of the mineral oil transformers ware found to have PCB concentration levels above 50 ppm, while only 1% were found to have concentration levels above 500 ppm. 1982 Comments, vol. Ill at 9. MONS 017024 In responding to the ANPR, USWAG focused its data gathering efforts only on high concentration, askarel transformers This approach was taken for several reasons. First, the three incidents that appeared to have triggered EPA's concern each involved an askarel transformer. Second, there is a general belief in the industry that the problems associated with mineral oil fires are substantially different than those raised by EPA in the ANPR; i-e., the significant concern of a mineral oil fire is the fire itself and its potential to cause severe injury to people and property. In addition, as discussed in Section III, the potential production of chemical byproducts from conversion of PCBs during a mineral oil fire is believed to be minimal. Finally, as shown in the 1332 RPC report (see, e.g., 1982 Comments, Vol. Ill at ISO-59), and as concluded by EPA in the August 1982 rule (see, e.g., 47 Fed. Reg. 37345-47 (1982)). the cost of the available risk reduction options regarding the use of mineral oil equipment, amounting to billions of dollars, would be totally unreasonable compared to the benefits achieved. The 1984 RPC questionnaire, therefore, principally sought information in three areas: 1. The number and location of askarel transformers owned by the electric utility industry, including the utilities' plans affecting the future use of this equipment; 2. The frequency, timing, location and circumstances of fire incidents that occurred over the last three years and Involved askarel transformers; the cost of clean up following such Incidents, and the populations exposed to such incidents; and 3. The utilities' experience with retrofill options. HONS 017025 7 The comp let* results of the RPC survey ere submitted herewith as Exhibit 2. A summary of the findings follow. A. Inventory of Utility-Owned Askarel Transformers 1. Current Inventory When first surveyed in 1981, the utility industry was estimated to own 39,640 askarel transformers out of a total universe of about 140,000 then in service. Of those 39,<4C, some 571 or 22,469 were located on the utilities' distribution systems. Current data show that the electric utilities now own about 36,000 such transformers of which some 18,291 remain in service on the distribution systems. As concluded by RPC, the magnitude of the decline in the inventory is greater than normally would be expected, primarily due to phase-out programs initiated by some of the companies.V The general location of the askarel transformers on the distribution systems was found to be as follows:**/ Number of Transformers Percent of Total Inside Building Sidewalk vaults Other 6,000 10,919 1,372 32.8 59.7 7.5 Total 18,291 100.0 The overwhelming majority of indoor transformers (4.51) are located in vaults within the building, while 9.39 */ These programs are discussed in the next section. **/ RPC 1984 Report, Table 1 at 6. HONS 017026 8- are located in separate transformer rooms; only 6.It are installed in'open areas. Natural ventilation to the outside predominates even with regard to the indoor equipment. Further, less than half of the indoor transformers are installed in locations where there is direct access to air conditioning or heating system intake vents. In total, of the 6,000 indcor installations, less than 60 transformers are found in vaults or rooms m which such intake vents are located and less than 2,700 are found where such intake vents are in the area immediately adjacent to the transformer vault or room.*/ While other indirect routes to heating and air conditioning systems car he identified, it is considerably less likely that these routes would provide a pathway for significant smoke travel. RPC 1984 Report at 8. 2. Voluntary Phase-Out Programs As indicated in the RPC survey, 766 of the surveyed Utilities have askarel transformer phase-out programs in effect, jd. at 12. Based on these voluntary programs, it is */ RPC 1984 Report at 8. Of the 12,291 transformers " located in outdoor installations, approximately 1,350 are located immediately adjacent to building heating and/or air conditioning systems that could serve as a means of building access for smoke/soot. Id. HONS 017027 -9- estimated that the number of distribution system askarel transformers will be reduced from 18,291 to 12,987 by the end of 1985 ( reduction of 29%), to 6,402 by 1989 (a reduction of 65%), and to 2,927 by 1994. Id., Table 5 at 13. Thus, of the distribution askarel transformers on line today, only 16% of them are expected to remain in use in ten years. Notwithstanding the adoption of these programs by many of the utilities, USWAG believes that the continued use of askarel transformers does not pose an unreasonable risk of injury to human health or the environment. These actions have been undertaken against a background of several concerns. First, there has been media and community expression of perceived risk from any exposure to PCBs. In addition, there have been two instances (Binghamton and San Francisco) in which very high clean-up costs have been incurred, in part because of the absence of any clean-up standards. The uncertainty of the financial consequences that could be imposed in regard to clean up, and from potential litigation, regardless of its merits, has resulted in business judgments to phase out some of the askarel transformers. Me feel the actions the industry is collectively taking deserve some recognition and credit by the Agency in its analysis. At the same time, however, the initiation of those programs should not be viewed as industry agreement that any unreasonable risk exists or that further regulatory action is warranted. MONS 012028 10B. Frequency and Nature of Fire Incidents 1. Frequency of the Incidents As sliown by the RPC data (RPC 1984 Report at 9-H), the annual number of expected fire incidents involving utilityowned askarel transformers is extremely low. Only eight such incidents have occurred over the last three years, indicating an annual estimated rate of 0.019. Of these, none occurred in industrial locations or in utility substation or generating facilities. In addition, despite the one dramatic event at Market Plaza in San Francisco in 1983, the actual amount of smoke travel associated with the incidents was negligible. In one incident no smoke escaped from the transformer. While smoke escaped from the transformer in the other seven incidents, in five cases the smoke either did not enter the building ( cwo outdoor installations) or entered only the area immediately adjacent to the equipment (three indoor installations), in only two incidents involving utility-owned equipment did smoke travel into public areas of a building. These situations involved the Market Plaza/San Francisco and First National Bank/Chicago incidents that are discussed more fully below. 2. Exposure Levels The only data available regarding the chemical composition in samples taken following fire incidents involving askarel transformers are the data obtained following the Binghamton, Market Plaza and First National Bank incidents. As acknowledged by EPA in the ANFR, no PCDFs and no PCDDs were found in the wipe samples from the bank building. 49 Fed. Reg. at 11073. While contaminant levels were found at the other two locations. USWAG believes that the significance of these findings MONS 017029 -n- en only be determined in conjunction with the degree to which m given populations can be expected to be exposed to the cor.ta.Tir.i.-, and the effect of any such exposure taking into account the expected levels at absorption.^/ USWAG believes that the current information regarding exposure and exposure effects from contaminants chat might be produced when askarel transformers are involved with fires are insufficient to justify any new regulatory programs to restrict the use of such transformers. To provide an analytical tool by which to better evaluate these effects, USWAG is implementing a decision framework model that was developed in a research project by the Electric Power Research Institute t*EPRI") (RP 2S951. The framework will be completed shortly and will be submitted to EPA. We note, though, that preliminary results indicate that even multiple exposure to the contaminants does not lead to unacceptable excess cancer risk. 3. Clean-Up Costs Survey data indicate that the average clean-up cost for an inside building incident, excluding the Market Plaza fire, was about $271,923. Clean-up costs following an incident involving a sidewalk vault averaged a full order of magnitude leas, or about $27,700. Zt thus appears that the clean-up costs associated with Market Plaza are aberrant. . */ For example, as reported in the RPC survey, few utility workers and firefighters were exposed to soot/amoke and even fewer members of the general public were involved. RPC 1984 Report at 10, 11. MONS 017030 -12- 4. Factors Decreasing Fire Incidents As indicated above, phase-out programs in place will substantially reduce the number of askarel transformers that could become involved in a fire incident. That decline, coupled with the overall incident rate, shows conclusively that the risk of a fire incident in the future on utility systems is very low, as shown in the following table: Year 1964 1985 1936 1987 1988 1989 1990 1991 1992 1993 1994 EXPECTED NUMBER OF ASKAREL TRANSFORMER FIRE INCIDENTS IN THE ELECTRIC UTILITY INDUSTRY Estimated * of Askarels** Estimated 4 of Fire Incidents 35,984 30,238 27,601 25,147 23,424 21,885 20,711 19,538 18,364 17,373 16,200 4.6 3.9 3.5 3.2 3.0 2.8 2.7 2.5 2.4 2.2 2.1 RPC 1984 Report, Table 5 at 13, and assumes 20>year average useful remaining life for substaeion/generati.ng facility transformer. HONS 017031 -13- The likelihood of such incidents occurring also is mitigated by the presence of protective devices (fuses, switches or breakers) on many distribution systems. Such devices are installed to isolate the transformer in the event of a fault. Seventy-five percent of the distribution askarel transformers have protective devices on the low voltage side of the transformer, while 19% of the equipment have them installed zr. the high side. RPC 1984 Report at 12. C. Retrofiliing Experience RPC sought data from the utilities concerning their experience with flush/refill and filtering processes to determine the frequency with which these methods have been used, the levels of PCBs achieved over time, and the co3ts of the program. As indicated in the report, the results have been extremely erratic and cannot be correlated with the methods used, the time elapsed or the expenditure made. RPC 1984 Report, Table a at 16. USWAG believes, therefore, that there is no basis to alter the conclusion it reached two years ago that retrofiliing askarel transformers is not, in the overwhelming majority of cases, cost effective. See USKAG/ESI.-NRSCA Reply Comments on Proposed Rules on the Use of Polychlorinated Biphenyls in Electric Equipment, June 24, 1982, at 22. Furthermore, there continues to be too little known regarding the electrical and fire resistant properties and toxicity of the substitute fluids to in any way compel their use. See Section IV. A., HONS 017032 -14- D. Impact of RPC Data on EPA Initial Estimates The 'information provided by RPC demonstrate* that data on which EPA baaed it* preliminary assessment of the risk posed by the continued use of askarel transformers are seriously flawed. Fires involving these transformers are rare, isolated events. In addition, while abnormally high clean-up costs have been incurred in two cases (one of which did not involve a utility-owned transformer), the expected levei of clear.-up costs is substantially less. 1. Frequency of Fire Incidents The RPC data on the number of expected fire incidents contrast significantly with most of the estimates noted by EPA in the AJfPR. If EPA continues to define a catastrophic incident as one in which clean-up costs of S13-20 million can be expeotad, a catastrophic fire incident involving a utility-owned askarel transformer can be expected approximately or.ce every 2.5 years. (This figure is based solely on the number of transformers adjacent to intake systems.)*/ RPC 1934 Report at 19, 20. If the focus is placed on the number of incidents in which ' This frequency rate was derived as follows: RPC identified ~ 2700 Indoor askarel transformers (6000 total indoor transformers (Table 1, p. 6) x 44.7% (the percentage of indoor transformers adjacent to intake vents)) and 1300 outdoor askarel transformers (12,291 total outdoor transformers (Table 1, p. 6, and p. $) x 10.9% (the percentage of outdoor transformers adjacent to Intake vents)) as being potentially involved in what EPA has defined as e catastrophic fire. See generally RPC 1994 Report at 19. These 4000 transformers were multiplied by the reported frequency rate of askarel transformer fires of 0.01% to arrive at 0.4 fires per year, tha equivalent of 1 fire every 2.5 years. If measured against the total population of equipment, the frequency rate would, of course, be dramatically less. mqns 017033 15- smoke reachej} substantial portions of a building, the expectancy rate based on the utility data is the same. Again, it was only in the Market Plaza incident that smoke was carried in any significant way into the public building. 2. Clean-up Costs As discussed above, the RPC data show that the level of clean-up costs averaged about 5279,000 in indoor installations, and $27,700 in sidewalk vaults. These levels, rather than the aberrational $15-20 million figures, should be used in analyzing the regulatory programs relating to the continued use of askarel transformers. II. REPRESENTATIVES CP THE BINGHAMTON, SAN rRANCISCO AND CHICAGO INCIDENTS The issuance of the ANPS was essentially prompted by three fire incidents that involved askarel transformers. These fires occurred in Binghamton, New York, San Francisco, California, and Chicago, Illinois. From its analyses of these fires, EPA appears to have concluded chat PCB Transformer 'fires pose relatively high risks' and 'can result in relatively high clean-up coats.* 49 Fed. Reg. at 11071. As indicated in Section I, these conclusions cannot be supported based on the utility data collected. In addition to seeking new data, EPA solicited cosunents on the representativeness of three fires in terms of the risks posed. This aspect of the ANPR is addressed below. HONS 017034 USWag believes that these incidents, as they are described by EPA, are anomalies for two reasons: (1) they are infrequent occurrences and (2) the potential risks are overstated. While the Agency indicate* these fires are "the three most well- characterized and well-researched incidents," 49 Fed. Re?, at 11071, the knowledge, as presented in the ANPR and rulemaking record, on the causes of the fires, the sources of contamination, the levels and extent of contamination, the exposure risks and the health consequences of that exposure is sparse. In addition, the circumstances and results of the fires are significant!/ different from one another and do not, together, create a picture of significant risk from the continued use of askarsl transformers . V A. Fire* Associated with Askarel Tranformers Are Rare EFA has presented data on the three fires. The data for two of the incidents show the formation and distribution of PCS* and PCB combustion by-products. Based upon USWAG's study of the frequency of significant events, only one utility incident was identified as catastrophic based on EPA's definition that being the San Francisco fire at One Market Plaza, May 15, 1983. While the electric utility industry may own only one- third of the askarel transformers in service, USWAG believes that its data Indicate an infrequent occurrence of this type of event. From USWAG's survey, this indicates that a "catastroph 2/ The three affected transformers on which EPA is focusing contained a high concentration of PCBs (65% or greater). USWAG has previously argued that askarel transformers (containing PCBs) are different from mineral oil transformers that may be contaminated with PCBs. See 1982 Comments, Vol. 1. In this rulemaking, USWAG believes that it is particularly important for the Agency to distinguish between the two types of equipment in order to make a proper evaluation of the risks. It is imperative that the Agency formally acknowledge the difference in the preparation of its proposed rule. MCNS 017035 -17- failure,* as EFA refers to it (49 Fed. Reg. *t 11079), has the r potential to occur approximately once in every 2.5 years (0.4 events per year) for utility-owned askarel transformers. RPC 1984 Report at 19, 20. ERA must consider the frequency with which askarel transformer fires may occur in its risk evaluation. Equally important is the fact that the population of askarel transformers is declining in the utility industry. See Section i. Therefore, the chances that an askarel transformer will become involved in a fire are rapidly diminishing.^/ B. Potential Risks Are Overstated Apart from the infrequency with which a Binghamton or San Francisco fire may potentially occur, the risks involved with such an event and continued use of askarel transformers are overstated as a result of the discovery of PCDFs and PC3Ds at the two buildings. It thus appears, in retrospect, chat the initial finding of concentrations of PCDFs and PCDDs at Binghamton led to an over-reaction and concern for public health protection from the use of askarel transformers. This over-reaction ateas from the mere presence of two substances that are interpreted to be highly toxic. While the distribution 2/ The Agency elso must not lose sight of the fact that the reason askarel fluids were originally used was because of their fire resistant properties. HONS 017036 -18- of these by-products throughout the building in Binghamton and portions of the building in San Francisco is unfortunate, there is still much to be learned about the events and there are many unanswered questions that need be addressed. Sore of the facts surrounding the incidents indicate a far lower hazard than EPA's analysis shows. 1. Causes of Fires May Be Preventable The discussion of the cause of the Binghamton f.re in the ANPR and versar's Report (Exposure Assessment; Fires Involving PCB Transformers) is very sparse. We merely know, based on communications with the New York State Office of Government Services (which owns the equipment) and others that a fire occurred in the switch gear. CSKAG is aware that EPRI is planning to conduct detailed case studies of Binghamton and San Francisco (RP 1263-20, Project Officer: Ralph Komai) that should be completed by October 1984.V The study's results should provide useful information on the actual causes of the events and thereby avoid any mis interpretation based only on current information. We therefore suggest EPA await the results of this report before deterxin.ng whether any changes in the use rules are appropriate. 2/ PCB Technical Report from EPRI, Number 8, attached hereto as Exhibit 3, presents a summary of the many relevant projects now in progress. HONS 017037 19 Siailarly, for the San Francisco and Chicago fires there is little or no information presented on the causes of the events. No conclusion, therefore, can be reached that these events and the occurrence at Binghamton resulted from comparable causes. 2. Levels of Contaminants Formed Is Inconsistent The New York State Health Department was the first group to detect and quantify PCBs, PCDFs, and PCDDs in soot samples taken from the Binghamton fire. USWAG notes the extreme variation in results from the ANPR and the Versar report.*/ These divergent results indicate a great uncertainty about the levels of contaminants to which people potentially may have been exposed. / See ANPR, 49 Fed. Reg. at p. 11072 and Versar ~ Report at 6. HONS 017038 -20- EPA notes the variation in results between samples, but fails to focus on the real import of the problem. That problem occurs because there is a huge disparity in test results for even a single sample. That sample, number 811711965, was tested by three analytical chemists in four chemical analyses. Versar Report at 6. The results are reported for 2,3,7,8-TCDF and 7C3D as follows: Smith 1991a Smith 1991b Races 1931 Stalling 1991 2.3.7.8-TCDF (ug/g) 273, 124 43 12 --------- 2.3.7.8-TCDD (ug/g) 2.8. 2.9 1.2 0.6 3 There is more than a 20-fold difference in result between Rapoe and Smith for 2,3,7,8-TCDF and mors than an 8-fold difference between Rappe and Stalling on 2,3, ',8-TCDD. Notwithstanding these major differences in results, EPA appears to rely on the results of soot samples first taken from Binghamton by the New York State Health Department. These values of 273 ppm and 124 ppm for 2,3,7,8-TCDF are significantly different from the values reported in a separate analysis of ehe sample by New York State, 48 ppm, and Dr. Rappe, 12 ppm. 49 Fed. Reg. at 11072. In light of these disparate results EPA cannot properly view the 273 ppm or 124 ppm levels of contamination as representative. MONS 017039 -21- Evidently, the analytical chemistry of what are the toxic compounds of concern is uncertain. This fact was confirmed in a recent interview with Or. Chnstoffer Rappe of the University of Umea, Sweden,V one of the Binghamton investigators and a leading world expert in the synthesis and analysis of PCTDs and PCDFs. Dr. Rappe indicated that (li there are only three laboratories capable o: performing the analysis c: these compounds, (2) the separation of the 2,3,*,9 isomer from the 2,3,4,8 isomer (the latter, in his assessment, a non-toxic isomer) was critical to proper quantification, ar.d (3) the extraction of PCDOs and PCDFs from soot particles is a rigorous operation requiring strong solvents; because of this problem, he suggested that the question of whether the by-products were actually biologically available to cause harm should be addressed. In light of Or. Rappe's remarks and the results of the 811711965 sample, it appears that there is an inconsistent, if not totally uncertain, knowledge about the contamination at Binghamton. Given the resulting levels and the significance of the level of contamination in evaluating exposure risks, the actual risk posed may be significantly less than assumed by BPA. 3. Confounding Sources of PCDFs and PCDDs A further complicating fact that has not been addressed Z/ Representatives of (JSWAG met with Dr. Rappe in New York City on June 8, 1994, to discuss his scientific opinion on (1) sources of PCDOs and PCDFs, (2) linear by-product formation, (3) by-products of substitute fluids, and (4) human exposure. RONS 017040 -22- at all is the existence of confounding sources of PCDFs and PCDDs. According to Dr. Rappe, it is difficult to distinguish the origin of PCDFs and PCDDs. Conceivable sources are polyvinyl chloride and flame retardant chlorinated hydrocarbons, both commonly used substances in buildings. Although the results of an inquiry into confounding sources could have a substantial bearing on the interpretation of potential risks from PC3 fires, one questions why no investigation has been initiated on this matter.. We also note that PCDDs and PCDFs may be produced in other situations not involving PCBs, such as barbecues, fireplaces and woodstoves in domestic dwellings. Given the economic impact of any further regulation cf askarel transformers, the Agency must clarify whether other sources of PCDFs and PCDDs are present, and, if so, how that fact affected the cost and effectiveness of the clean-up effort. In addition, USWAG believes that it would be relevant to know what expected background levels of PCDFs and PCDDs in buildings might be for purposes of comparison. 4. Formation of Reaction Products USWAG also spoke with Dr. Rappe regarding the formation of PCDFs and PCDDs following fires in which askarel transformers are involved. On the basis of this interview and other informat.o available, USWAG believes that no conclusion can be drawn at this time regarding the linearity of by-product production or the quantitative conversion of PCBs to other contaminants. On the subject of CPA's assumption that PCDFs and PCDDs are linearly formed from PCB mixtures, Dr. Rappe felt that this could not be assessed because at present^no data are HONS 017041 23- available. H stated that the exact nature of the reaction, whether it is monomolecular or bimolecuier, is not known. In feet, more than one mechanism may be operating. Therefore, models used to simulate pyrolysis that show linear relationships, according to Dr. Rappe, are not readily applicable to the fire situations. These models have limitations and do not necessarilv mimic the effects of (1) a clean versus dirty combustion flame, (2! oxygen variation, (3) electrical versus chemical energy, and (4) temperature. Accordingly, without empirical data, the Agency may draw conclusions that misrepresent the severity of PC8 combustion and conversion. In addition, there is r.o adequate information upon which to assess the quantitative conversion of PCS askarel fluids to PCDFs and PCDOs at Binghamton or San Francisco. Only sketchy data have been presented to date. The analytical results we do have indicate the presence of one half to 43 times less 2,3,7,8-TCDF and of 10 times to 50 times less 2,3,7,8-TCDD in soot samples taken from One Market Plaza compared to Binghamton.To OSWAO, this shows that Binghamton and San Francisco are not comparable events, except in the qualitative identification of the compounds. The concentrations are significantly different. EPA is aware that EPRI has contracted with the New York State Health Department to examine the pyrolysis and combustion of PCBs and retrofill fluids (RP 2028-4)./ This V Compare Table 2, p. ( with Table 5, p. 10, of the VersarReport. MONS 017042 SSI Exhibit 3. 24 project aimj to "determine the linearity of PCDF as a pyrolysis or combustion product with decreasing PCB concentration." tx 3 at 1. This study is critical and will have a significant bearing on EPA's findings with regard to askarel transformer use and the purported risks. 5. Exposure and Health Effects of PCBs. pcdf* and Pccgi ^ ------------------------------------ -- ----------- By virtue of the different levels of PCBs, PCDFs and PCDDs found in the Binghamton and San Francisco fires, the potential exposures to firefighters, clean-up workers, office workers, and bystanders is different. OSWAG is attempting to characterize the exposure to and effects of PCDDs and PCDFs from fires. The project, being prepared by Decision Focus Incorporated, will examine dermal contact and inhalation as the primary routes of exposure to PCBs and the by-products. The project, as noted in Section I, is a decision framework model, prepared by EPRI (RP 2595), that allows the assessment of policy options in comparison with estimated health risks. EPRI also is conducting an exposure assessment study for PCB spills and fires (RP 1826-13) that will be completed by December 1984. This project, too, should help EPA in its assessment of exposure risks. In assessing the effects of exposure to PCDDs and PCDFs, it is important to consider the matrix of exposures and the bioavailability of toxins. One literature review study HONS 017043 -25- performed by EPA on the toxicity of the Binghamton soot concludes that "multiple exposures to soot from PCS-Transformer fires have the potential to produce toxicity in the thymus, the hematopoietic system, the salivary gland duct epithelial, and, possibly, the liver (Ref. 6)". 49 Fed. Reg. at 11073. While the effects are not insignificant, the means of exposures to humans during an event and following an event are totally different. Because there is so little information available to simulate the exposure and effects on humans, we suggest the Agency divert resources to this area before it prepares findings on exposure to PCB and PC3 mixture by-products in different media by inhalation and dermal contact. C. Conclusions Pegardino the 9ir.cha.mton, San Francisco and Chicago Incidents EPA must weigh several factors in evaluating the risks actually posed from Binghamton, San Francisco and Chicago incidents and the representativeness of those circumstances. These include: . (1) The frequency of such events; (21 Future preventability and mitigation of the incidents; (3) The actual severity of exposure to fire fighters, clean-up workers, and others; (4) The value of awaiting the completion of EPRI and USWAG studies prior to further action. we have demonstrated through the most recent USWAG survey that the utility industry rarely experiences an askarel HONS 017044 -26- transformer fire of the magnitude of San Francisco, we eatimata 0.27 incident* per year or a frequency of 0.01 per year over the lifetime of the equipment. The industry has reduced it* population of asKarel transformers and this population is rapidly declining. There is a need to further evaluate the causes of these fires, because experiences at Binghamton, San Francisco, and Chicago may help us avoid future incidents and thereby avoid the need for further regulation. The actual severity of the exposure of pcpulaticr.s to the by-products EPA believes are associated with fires involving askarel transformers must be questioned given the (1) infrequency of the events, (2! laclc cf knowledge about exposure to these compounds through the expected primary routes of exposure -- dermal contact and inhalation, (3) the matrices in which FCDFs and PCOOs are bound and their relative biological unavailability, and (4) uncertainty about other sources of dioxins and dibenzofurans and associated fire toxins. Lastly, in addition to voluntary programs for removing askarel transformers, the electric utility industry has shown a significant commitment to properly manage its PCBs through accelerated research programs. These programs, as they relate to fire episodes, are studying under severe time constraints the following matterst HONS 017046 27- (1) Mechanisms of PCDF and PCDD formation from studies of PCB, tri- and cetra-chlorinated benzene pyrolysis and combustion: (2) Case studies of Binghamton and San Francisco; (3) Engineering Options for PCB Askarel Transformer Risk Reduction; (4) Exposure assessment for PCB Spills and Fires; (5) Risk Management of PCB Fires and Spills; (6) Utility Decision Framework for PCB Decisions; (7) Studies on the Use and Toxicology of PCB Substitutes among other important PC3 Research Projects. Many of the projects will not be completed until December 1984 or later into 1985. Yet they could have a significant bearing on the proper direction for EPA to cake in determining the need for further regulation of the use of askarel transformers. As we have shown, there are many uncertainties about the risks associated with PCB fires. Similarly, there are uncertainties about the risks associated with the use of new fluids.*/ Each of these considerations must be oarefully assessed before any action is taken. What appears clear today, however, is that the three fires EPA evaluated as representative of risk do not adequately account for these uncertainties and do not present a sufficient basts on which to conclude that the continued use of askarel transformers pose an unreasonable risk of injury to human health or the environment. MOWS 017046 tf See Section IV.a, infra. -28- III. MECHANISMS OF PCDF AND PCDO FORMATION ' A major issue in this rulemaking is the extent to which PCSs and PC3 mixtures are converted to other compounds under pyrolysis and combustion conditions. As discussed more fully in Section II.B.4., a definitive conclusion on this question cannot be drawn from the currently available dats. while the focus of EPA's question was askarel transformers, the Agency also raised a more general question concerning fires involving mineral oil transformers. In CSWAG's interview with Dr. Rappe, he indicated that an abundance of soot would be formed from the combustion of minersl oil which would probably result in a very complicated matrix. Because cf this complex matrix, it might be very difficult to extract PCDFs and PCDDs and fairly strong solvents would be required. He related experiences he had had in spiking certain sediments with carbon 13-labelled PCDFs ar.d finding it exceedingly difficult to recover the material. He anticipated that similar problems, along with separating PCDCs and PCDFs from other hydrocarbons, could result. In its ANPR, EPA examines laboratory studies and discusses them in terms of fire situations. EPA notes that chemical reactions which occur in a fire with Aroclors is far more complex than laboratory pyrolysis experiments. Yet, the Agency expects the same reactions to occur and chemical products to be formed. Further, EPA suggests that the lower PCB levels HONS 017047 -29- found in mineral oil 'should have a minimal effect on the reaction ratfe or product yield at a given temperature'. 49 Fed. Reg. at 11074. Lower PCB levels mean lower PCDF levels formed, according to the Agency's analysis. From our discussions with Dr. Rappe and our reading of the literature, OSWAG believes that the Agency's first stateme that PCB combustion in fires is complex, is correct. The reaction mechanisms and product yields are affected by many variables other than just temperature. For example, it may be plausible to conclude that diluted PCBs or chlorobenzenes cannot be converted to PCDFs or PCDDs, because mineral oil cr other insulating fluids would consume the oxygen available to nurture the fire. Accordingly, there is no information avails today from which to conclude that PCBs at low concentration levels can be transformed into PCDFs. USWAG believes that research work being performed by CPR1, EPA, and possibly others is pivotal to any finding on the conversion of diluted PCBs and, for that matter, even ?C3 askarel fluids encountering combustion. IV. OTHER RELEVAWT FACTORS HONS 017O4B In assessing the reasonableness of the continued us* of transformers containing or contaminated with PCBs when such equipment may become involved with a fire, a number of other factors must be considered. EPA has sought information on two of these factors, substitute fluids and exposure risk from contaminated water. A third factor, building fires generally. -30- also should b considered. Each of thasa matters is addressed briefly below. A. Substitute fluids for PCBs EFA has requested information on substitutes for PCS fluids. Specifically, it seeks data en the toxicity of substitute fluids in the event of a fire. Reviewing the available materials, USWAG concludes that not much is really known about the properties of these fluids, their toxicity, or the effect on them of pyrolysis or combustion. EPA identifies six major substitute dielectric fluid types: silicones, high-temperature hydrocarbons, chlorinated hydrocarbons, non-PCB askarels, fluorocarbons and mineral oil. Several investigators have reviewed the scientific literature on both PCB capacitor and PCB askarel transformer fluids. ErA itself has performed analyses through SRI International (February 1981) and more recently PEDCO Environmental Incorporated (January 6, 1984). Similarly, EPRI is in the process of completing a study on PCB substitutes. See Exhibit 3. Although a recent article published in Environmental Science and Technology (17:486A, 1983) concludes that some substitutes currently marketed have a low acute toxicity, appear to be more biodegradable than PCSs and lass likely to bioaccumulate appreciably, there Is limited knowledge, especially in comparison with PCBs, on chronic and subchronic health effects. Thus, the normal dayto-day use of the substitutes generally has not been adequately evaluated from a health risk standpoint. HONS 017049 31- Som information is available, however, on the effect of combustion. For example, during arcing of a perchlocoethyler.e transformer, the PEDCO study indicates the breakdown products include hydrogen chloride, carbon, carbon monoxide, and carbcn dioxide. In contrast, initial EPRI literature analysis shows that the predominant resulting product is phosgene, other hazardous decomposition products include hydrogen chloride and carbon monoxide. This disparity indicates that a rigorous analysis of the potential health risks of substitutes and their by-products remains to be made. The current EFRI project should be completed by December 1984. In addition, EFRI will be performing several studies in the area of substitutes: 1. Pyrolysis and Combustion of PCS Contaminated Dielectric Fluids and Retrofill Fluids, such as tetrachloroethylene (RP 2028-4), scheduled for completion by November 1984. 2. Are and Spark By-products (RP 1499-4-5) of Perehloroethylene, scheduled for completion by August 1984. 2. State-of-the-Art Toxicological Review of PCB Substitutes that will focus on pyrolysis and combustion products (RP 2028-12), scheduled for cos^letlon by December 1984. Although these studies may not be definitive, they should help EPA in its deliberations. B. Risk from Contaminated Water HONS 017050 EPA solicits comments on the exposure risk from contaminated water resulting from PCB fires. From the limited Information we have been able to obtain, it appears that both 32- contamination of water and exposure levels are potentially nonexistent. Judging from a conversation with Pacific Gas & Electri Company, there was very little water used to dowse the transformer at One Market Plaza, and the little ultimately used was contained within the vault.V Further, it can be expected in future fires that firefighters will control such types of electrical fires either by removing the source of oxygen or by using carbon dioxide to control the fire. Therefore, USWAG believes there should be little contamination and no spread of contamination through water run-off because water is not required to control such a fire. C. The Ccnsecuer.ces of Building Fires The focus of the ANPR is on askarel transformer fires and the risks from exposure to the associated combustion by products. Given that askarel transformer fires do not burn as raging fires and in fact prevent the spread of fire, a convincing argument can be made that the installation of askarel equipment significantly prevents the direct loss of life from building fires and the damage costs that result frcr those fires. In Binghamton, San Francisco, and Chicago, whether there was a real concern from PCB exposure or not.V V In the case of Binghamton, it is not apparent whether any water was used to control the fire. HONS 017051 -33- there was no loss of Ilfs and psople could be safely evacuated. * USWAG believes that in making any further risk assessment, the measure of safety that askarel transformers provide must be weighed against the consequences of building fires, regardless of the type of transformers that serve them. Every year there are over one million building fires.*/ In 1981, fires were attributed as the cause of 900 deaths in public and work environments at a cost of $3.4 billion. According to a recent article published in Civil Engineering magazine (May 1984, pp. 41-43), about 80% of the deaths that occur in building fires are attributed to the inhalation of toxic gases emitted from certain burning materials. Thus we can estimate that 720 people died in 1981 from exposure to toxic fumes. The toxic fumes result from the decomposition of plastics and other synthetic building and furnishing materials. In a resolution brought before the California Assembly, Assemblywoman Maxine Waters cites a test developed by a University of Pittsburgh toxicologist which "reveals that about S ounces of burning of polyvinyl chloride, one of the commonly used synthetics in buildings, could emit enough hydrogen chloride gas to kill all the people in an averagesite bedroom in 10 minutes.***/ */ In 1981, for example, there were 1,003,77$ building fires. See 1981 Database, Federal Emergency Management Authority, National Fire Incident Report. 11/ Civil Engineerino/ASCE. May 1984, p. 42. HONS 017052 -uCONCLUSION From' the information provided in the RPC 1984 Report and elsewhere, USWAG believes that no unreasonable risk is posed to human health or the environment from the continued use of transformers containing or contaminated with PCBs. Many studies now being conducted will shed further light on many of the questions arising from the involvement of the transformers in fires. As of today, however, there is insufficient reliable information on which to place further restrictions on the use of the transformers under consideration. EPA should, therefore, issue a statement that further rulemaking at this time is unnecessary. HONS 017053 M1n llKtrte taiUnti hUiul tmftl IlKtm CMpmUw tMMUtiM iMrlua NU< hm> UmcIium teuiiu llwuii Nw larrita CorptriUN Appalaefcina Fmf Caaptay CaL**ws * Iwmn Otla Uicuie ^tr-piay aaixu i Rleiiftt IlMtnc cuy iMtwiy l*w Canptuy out am* itimx Atlantis City Klaasria Cl^niy taitlMct SM t C^uy Ciitnl HiaM 6m Utctrit cnatiM Caaml UUmIi UfM ea^ax Cutni IlUMli N*Ur limn Caapaay eaatrai i Intu wait Cirnnun CmuiI Nmt I Uftt Cajuy MU lrriM Ca^aay si OUiIwm iMtAwnttra Umrlc >nr caapaay Nit rasas OtiUtlsa Caapaay Cincinnati Ma a Klaasria Caapaay Ceawa<lta Milan Caapaay Caassll4ta4 KSiaaa Caapaay af Maw far*. :, Dayton Pawar a Liyas Caa^asy Du*a Naif eaapaay Dupjsaas Ut Caaprey PiatiKs Pawns a Uynt Csap any 6i( Itatai DtlUtiss Caapaay Sanataa LiyntLay a Pawar Caapaay tntasia Pawar Caapaay XMlaaapaUa Nwii a Llyat Caapaay iawa*UliAait 6aa a Uaitm Caapaay Zawa Pawar a Ufht Ca^awy Xawa Pvaiic farvica Caapaay Jaraay Caatrsl Pawns a Liyas caapaay Las Aayaiaa Oaparsaaas #1 wasar a Pawar MiMls fautn latviraa, tap. Afkum Pawar % Llt*t Caapaay Louisiana Pawar 6 Liyitt Caapaay Maw orlsam Punlie Samri. tnr. Niiaiaiippl Pawar Liyns Caapaay nananyanala Pawar caapaay saw snyiana Pawar caapaay Maw rats tsasa Blsctris i Oaa Carparasian Niayara Mnftawk Pawar Carparasian Martnarn IMiaas PaPlic Sanlci Caapaay Marsnara Itaaaa Pawar Caapaay Oliia Miaaa Caapaay onia Valtay Klaasria Carparasian Oklanaaa Oaa S Kiaasrie Caapaay Panda Oaa * Klartrls Caapaay Paaaaylvaau Pawns a Liyftt Caapaay miaialpnsa llrnrit Caapaay Paspaaa Klaasria Pawar Caapaay NPU lorries Klsrtflc l Oat Caapaay POPUP Parrita Caapaay af la4iaaa, Inc. laatpam Callioraia Mima Caapaay aanpata Caapaay lirvicti. Saa. Alstons Pawar Caapaay Oaavyia Pawar Caapaay Calf Pawar Caapiay Mississippi Pawar Caapaay Taapa llaasria Caapaay Timi KtlUtlna Oaaarasiay Caapiay TM CiswalaaP Klaasria tllPiaisiaf Caapany Tha pasrais Miaaa Caapaay TPa PaSaaaa Miaaa Caapaay talaOa Miaaa Caapaay Oaiaa Klaasria Caapany viffMii Klaasria i Pawar Caapaay m Paaa Pawwr Caapaay visaaasia Klaasria Pawar caapiay aiaaaaaia Pawar s LiMs Caapaay visaaassa PwpUt farwica Carparasian HONS 017054 PIPCiRT Or TKI STf-jv r- a....... BY ASKAP.IL TRa.VS?Os-'C9 r5s? (/i o RESOURCE PLANNING CORPORATION RPC i22Sl9in$lr*#f.N W Surf:iC Wojnmgion. CC 2CC3a (302) 7C7.tm HONS 017055 0 ,0 RPORT OF THE STUDY OF RISKS = BY ASKARZL TRA-VaFOPO'E3 Prepared For The Edison Electric Institute ud The Utility Solid Waste Activities Group Prepared By: Resource Planning Corporation 1225 19th Street, JJ.W. Suite 650 Washington, D.C. 20036 June 19. 1984 HONS 017056 TABLE OF CONTEXTS 1.0 INTRODUCTION....................................................................... Pag; ., 1.1 PCB Regulatory Background............................................. ; 1.2 Purpose of this Study.....................................................................> 1.3 Study Methodology............................................................... 2 1.4 Organization of this Report .... 4 2.0 SURVEY FINDINGS........................................................................ g 2.1 Survey Response ....................................................................... g 2.2 Askarel Transformer Population.................................................. g 2.3 Location and Environment of Distribution System Askarels 5 2.4 Frequency and Nature of Fire Incidents..........................................3 2.5 Cost of Fire Cleanup............................................. u 2.5 Utility Risk Reduction Measures.................................................12 2.T Retrofill Experience.............................................................. U 3.0 COMPARISON OF SURVEY DATA WITH EPA MATERIALS . 13 3.1 General................................................................................. 13 3.2 Frequency and Nature of Fire Incidents . . . is 3.3 Cleanup Costs..................................................................................... 19 3.4 CommentsRegarding EPA Cost-Effectiveness Analyses................................................................................. 20 APPENDIX A: UST OF UTILITIES SURVEYED APPENDIX B: SURVEY QUESTIONNAIRE ii HONS 01705 7 TABLE l: TABLE 2: TABLE 3: TABLE 4: TABLE 5: TABLE 6: TABLE 7: TABLE 8: LIST OF TABLES Location of Distribution SystemAskarels pace 6 Buildings in Which Transformers areLocated . 7 Location of Transformer Fire Incidents . . 9 Average Cleanup Cost.................................... n Estimated Reductions in Distribution Askarel Transformers Based on Current Phase-Out Programs............................................................................. 13 KVA Rating and Fluid Capacity of Retrofilled Transformers.................................................................... 14 Flush Solvents used in Retrofills 15 Retrofill Results 16 iii HONS 017058 1.0 INTRODUCTION 1.1 PCB Regulator? Background la August, 1982, EPA issued the PCB Electrical Use Rule authorizing the use of certain electrical transformers containing polychlorinated biphenyls (PCBs). Among other provisions, this rule allowed electric utilities to continue to use the approximately 40,000 askarel transformers then in service. In March, 1984, EPA issued an Advance Notice of Proposed Rulemaking (ANPR) stating that the risks posed by fires involving PCB Transformers had not been considered during formulation of the 1982 rule, and that consideration of such risks might result in the necessity for additional regulation. Accordingly, the ANPR solicited data specific to the r.sks of fires posed by PCB Transformers. 1.2 Purpose of this Study In the summer of 1981, Resource Planning Corporation 1RPC), under contract to the Edison Electric Institute and the Utility Solid Waste Activities Group (EEI/USWaG), surveyed the 100 largest utilities in the United States. This survey collected data regarding the use of PCBs by the electric utility industry, and was responsive to information requirements defined by EPA, EEI/USWAG, and the Environmental Defense Fund as part of an order issued by the U.S. Court of Appeals for the District of Columbia Circuit. Although the 1981 survey collected substantial data regarding the use of PCBs and formed the basis for the August, 1982 rule, no data were collected regarding risks posed by PCB Transformer fires. This report describes the results of a 1984 survey of electric utilities specifically designed to obtain data relevant to the fire risk ANPR. The survey was again performed by Resource Planning Corporation under contract to EEI/USWAG. This recent survey provides substantial HONS 017059 empirical data regarding PCB Transformer fire risks, and should he particularly useful in describing the location and environment of askarel transformers and estimating the frequency of PCB Transformer fires - two issues critical to EPA analysis of alternative regulations 1.3 Studr Methodology In order to respond to the fire risk ANPR, three general types of information had to be obtained from utilities: detailed information regarding the location and environment of askarel transformers,1 information regarding transformer fires, and retrofill experience. Four approaches to the collection of these data were considered, including: - Surveying those utilities that reported owning askarel trans formers in 1981 (90 of 100 surveyed). Surveying the 10 or 20 utilities that reported the most askarels in 1981. - Collecting some data (i.e., fire data and retrofill experience) from the 1981 sample, and more detailed data on transformer location and environment frca a small sample of utilities that are known to have such data available. - Surveying those utilities that reported owning askarels in 1981.' and asking them to provide detailed data on a sample of their transformers. The advantages and disadvantages of these various approaches are . relatively straightforward. Using the same utility sample as in 1981 would facilitate data comparisons, i.e., the 1981 transformer inventory could easily be compared with the 1984 inventory. However, it was deemed impractical to expect 90 utilities to promptly provide detailed information regarding the location and environment of each askarel transformer they own. Surveying the 10 or 20 utilities reporting the most askarels in 1981 would provide t somewhat more manageable sample, but the problem facing individual utilities of providing detailed data on hundreds of 1. Although PCBs were sold under a variety of trade names, they are commonly referred to by the generic name "askarel.'' Throughout this report, "askarel transformer" refers to transformers typically containing PCBs at levels of 60% or greater. 2 HONS 017060 askarels would remain. The response rate would likely be low, par ticularly given the time constraints. Collecting some data from the 1981 sample and more detailed data from a small number of utilities (3-6) that are known to possess the data was considered a reasonable approach. The survey would be manage able, and data could be obtained on a substantial number of askarels within a relatively short time frame. The disadvantage with this approach is its representativeness. It would be necessary to generalize to the industry based on a small group of utilities. If the response of all sampled utilities were siniliar, it could be reasoned that such comparability allowed reasonable population projections. However, if responses were dissimiliar projections would be of questionable validity All of the disadvantages associated with the above approaches could be overcome if a random sample of utility industry transformers were selected and data obtained regarding their location and environment. No single utility would be required to provide data on a large number of askarels, thereby alleviating their reporting burden, and reliable projections could be made to ail industry equipment from a relatively small sample of transformers. Such an approach seemed the best methodology for obtaining data required to respond to the ANPR. In refining this methodology, the 1981 survey data were reexamined in light of the ANPR data requirements. Electric utilities have askarel transformers locaced in substations, generating facilities, and distri bution systems. However, the major concern stated by EPA in the proposed rule-making was the risk to the public of askarel trans formers, i.e., distribution system askarels. Given that expressed concern, and in the absence of any reports that askarel trans formers located in utilities' restricted access areas have been involved in fire incidents similar to those that prompted EPA's ANPR,1 2 2. Substations and generating facilities also pose little or no risk. i.e.. there is no public exposure, equipment monitoring and inspection occurs frequently, and utility workers are both sensitive to the hazards posed by electrical equipment and trained in the use of PC3 Nonetheless, the survey did seek information on fire incidents involving askarel transformers wherever the equipment was located 3 HONS 017061 data collection efforts were focused on the distribution system askarsl transformers. In defining our study sample, it was determined that 52 utilities owned 99% of the distribution system askarel transformers reported in 1981. Therefore, it was decided to focus survey activities on these S2 utilities.1 Each of the 52 were asked to provide: General Data, including their current askarel transformer inventor.* and nsk reduction measures taken or planned. ' - Fire Incident Data, including information regarding all fire incidents during tne last three years which involved an askarel transformer, PCB Transformer, or transformers containing askarel substitutes. Askarel Transformer Data, including detailed information regarding Uie location and environment of a sample of their askarel trans formers . A sample of 453 randomly selected askarel transformers was deemed sufficient to make projections to the electric utility industry distri bution system askarel transformer population. Each of the 52 surveyed utilities was given specific instructions regarding the transformers for which data were to be reported. It should be noted that only trans former location and environment data were collected on a sample besis. Ail other survey data covered all askarel transformers owned by respondent utilities. For example, the data on fire incidents covers all tire incidents in the last three years experienced by the sampled utilities. 1.4 Organization of this Report The remainder of this report comprises two major sections: - 2.0 SURVEY FINDINGS describing the data obtained via the survey of 52 electric utilities.3 3. Appendix A provides a list of utilities surveyed, and Appendix B provides t copy of the survey questionnaire. 4 HONS 017062 - 3 0 COMPARISON OF SURVEY DATA WITH EPA MATERIALS analyzing the preliminary EPA regulatory anaivsis4 in light o.' data obtained via the utility survey. ' 4. Putnam, Hayes & Bartlett, Inc., 1984. Preliminary Study and Coat-Effectiveness Analyses of Alternative Regulations tor indoor PCB Transformers. EPA, Office of Pesticides amd Toxic Substances OFiS 6Z035 PCS (Fires), February 1984. 5 HONS 017063 2.0 SURVEY FINDINGS 2.1 Survey Response Of the 52 utilities, 51 provided data. The single non*respondent indicated that insufficient time was available in which to respond. The following results represent projection to the population of utilityowned transformers based on the results of the survey 2.2 Askarel Transformer Population Based on the 1981 utility survey, it was estimated that the electric utilities owned 39,640 askarel transformers, with 22,469 in distribution systems. The 1984 survey indicated that the electric utilities now own 35,984 askarel transformers, and the number of distnbution system askarels has declined by approximately 18.5% since 1981. There are now an estimated 18,291 askarel transformers in service in utility distribution systems.5 This overall decline as well as the decline in distribution system askarels is somewhat greater than would be expected via normal . attrition, because a number of utilities have instituted accelerated phase-out programs. 2.3 Location and Environment of Distribution System Askarels Table 1 indicates the general location of the 18,291 electric utility distribution system askarel transformers. TABLE 1: Location of Distribution System Askarels Inside Building Sidewalk Vault Other TOTAL Number of Transformers "XW 10,919 1,372 18,291 Percent of Total 32.8 59.7 7.5 luO.Q 5. The reported decline in the total number of utility-owned askarel transformers from 1981 to 1984 is somewhat less than the decline in the number of distribution system askarels, i.e., the 1984 survey reported more substation/generating facility askarels than the 1981 survey. It was determined that two utilities underreported their 1981 population of generating facility askarel transformers. The result of this reporting error on 1981 equipment projections was less than 4%. 8 HONS 017064 Approximately 10,919 transformers are located in outdoor sidewalk vaults, and an additional 1,372 are in other outdoor locations. In the case of a fluid loss, 66.6% of all distribution system askarel transformers locations would contain the fluid and prevent release into the environment, 6.5% would contain the fluid in a catchment area. 7.7% of the locations would drain to sewer systems, and 19.2% would drain to surrounding soil, gravel or concrete. For transformer cooling purposes, 94% of all distribution system askarel transformers are ventilated to an outside area, 4.5% vent to an indoor area, and less than 1% are in a sealed room or vault. Less than 20% of these ventilation mechanisms involve forced-air systems. 2.3.1 Indoor Installations Of the 6.0C0 transformers located in buildings, Table 2 provides additional information on the various types of building installations. TABLE 2: Buildings in Which Transformers are Located Building Type Number of Transformers % of Total Residential (Condo, Apt. Hotel) Institution (School,Hospital,Church) Office Industrial/Mfg. Storage (Warehouse) ' Other Public (Store, Mall) Other* 727 318 3,232 180 180 727 636 12.1 5.3 53.3 3.0 30 12.1 10.5 Total 6,000 100% * Primarily includes parking garages, and uninhabited industrial structures. In those instances where transformers are located inside buildings, the average age of the buildings is 29 years, and although the average height of the building is 19 floors, over 50% of the transformers are in buildings of 11 floors or less. 7 HONS 017065 While 51.2% of the indoor transformers are installed in the basements of buildings, only 3.1% of the indoor installations are roof top instal lations; the Pemaining 45.7% are located on other intermediate floors within the building. The overwhelming majority (84.5%) of the indoor transformers are located in vaults within the building, 9.3% are located in separate transformer rooms, and the remaining 6.1% are installed in open areas. Of the transformers that are located in a vault or transformer room, less than 1% have air conditioning or heating system intake vents in the transformer enclosure itself, and less than one-half (44.7%) have air conditioning or heating system intake vents in the area immediately adjacent to the transformer vault or room. Other than these direct routes, approximately 31.2% of the transformer locations are judged to provide some, albeit indirect, access to building heating and air conditioning systems. The responses to the survey indicate these indirect routes may include access to heating and air conditioning through open windows in the building, holes in concrete walls and ceilings, spaces under doors, gaps around electrical conduit, etc. 2.3 2 Outdoor Installations There are approximately 12,291 askarel transformers in sidewalk vaults and other outdoor locations. In the case of a transformer fire. 10.9% of these transformer enclosures are located immediately adjacent to building heating and/or air conditioning systems which could serve as a means of building access for smoke/soot. 2.4 Frequency and Nature of Fire Incidents The 1984 utility survey defined a PCB Transformer fire incident as a fire in which t transformer has been directly exposed to a fire, either from within the transformer or an immediately adjacent source. Using this definition, survey respondents reported that 8 PCB Trans former fire incidents have occurred in the last three years. All of these incidents were reported to involve askarel transformers. 8 HONS 017066 Using these dau and the total number of askarel transformers owned by surveyed utilities, it can be estimated that 0.01 percent of all askarel transformers owned by electric utilities can be expected to be involved in a fire incident each year. Because so few incidents occur, it is difficult to develop reliable incident predictors, i.e , the small size of the sample does not allow the correlation of fire incidents with transformer age, EVA rating, type of building, time of day, or any other measured variable. However, a review of the circumstances surrounding the 8 reported incidents may be useful in considering the likely results of future incidents. Table 3 indicates the locations of utility industry askarel transformers involved in fire incidents during the past three years. TABLE 3: Location of Transformer Fire Incidents Location Outsice - sidewalk vaults Inside - office building Number of Incidents 3 _4 TOTAL 8 Because the risks of exposure posed by outdoor transformers may be significantly different than those of indoor transformers, each is discussed separately below. 2.4.1 Sidewalk Vault Incidents Four of the eight fire incidents involved transformers in outside sidewalk vaults. KVA ratings were reported for three of the four transformers (SOO, 750 and 1,000) and the ages of the four transformer were 8, 12, 15 and 20 years. Two of the fires occurred in the morning, one in the evening, and for one the time was not provided. All but one occurred on a weekday. . In two of the three incidents smoke escaped from the transformer enclosure via the sidewalk grate but did not enter any public buildings In one incident smoke did not escape the transformer vault. In one 9 HONS 017007 incident smoke escaped the vault and entered a nearby building. In one of the three incidents, no fluid was released from the transformer, and in the other three incidents, releases of 5. 25, and 60 gallons occurred. Data regarding exposure of the public, utility personnel, and fire* fighters were reported for two of the four incidents. Neither reported that more than five members of the general public were exposed to smoke/soot from the fire and the average number of firefighters and utility personnel exposed to smoke/soot was nine. Limited data were reported regarding the lapse of time between fire start, de-energization, fire department response, and extinguishment. In one of the incidents, the transformer was de*energi2ed approximately 10 minutes after the fire started, in one incident de-energi2ation * required an hour, and in one incident as much as three hours may have elapsed. For one incident no data were reported regarding de-energization. No data were available regarding the time required for fire department response. For one incident it was reported that two hours elapsed from fire start to extinguishment, and for another total time lapse was reported as three hours. PCB, PCDF, and PCDD test results were generally not available for the four sidewalk vault fires. For one incident PCB fluid levels of 13,000 ppm were reported before cleanup and 12 ppm after. 2.4.2 Inside Building Incidents . The four office buildings in which transformer fires occurred ranged from 1 to 48 floors. One building was 10 years old, one was 13 years old, and the ages of two were not reported. Transformer KVA ratings ranged from 500 to 2,000. The ages of two of the transformers were 10 and 13 years, with two ages unknown. In each of the four incidents occurring inside buildings, the transformer was enclosed in vault, and in each incident smoke/soot escaped the transformer vault. In three of the four incidents the escaping smoke/ 10 HONS 017068 soot only entered immediately adjacent areas of the building. In one incident the smoke escaped the vault and was picked up by the air intake of an adjacent building. , . Fluid was released from the transformer in each of the four incidents, with the quantity released ranging from 4 to 100 gallons and the average being 41 gallons. The extent of exposure of the public, utility personnel and firefighters to smoke/soot was generally not known. However, for one incident no exposure to the public was reported, and for two incidents exposure of 12 and 18 utility personnel and firefighters was reported. For one incident, ?C8 levels were reported in the soot before cleanup of 2*98 micrograms per 100 square centimeters and 1-21 after. Another reported 860-2,000 micrograms of PCB per 100 square centimeters of wipe area in soot before cleanup, with no after cleanup results available. A third incident reported 18 micrograms per 100 square centimeters of PCDF in the transformer vault, with no PCDDs or PCDFs found elsewhere (after extensive testing). 2.5 Cost of Fire Cleanup When calculating the average cleanup cost associated with reported fire incidents, one incident stands substantially apart from all others. Of the 8 incidents reported, cleanup of one is estimated to have cost $15 to 20 million. No other incident reported costs as high as 5425,000. To avoid the distortion of the one $15 to 20 million incident. Table 4 excludes Chat $15 to 20 million incident and indicates the average cast of cleanup associated with all other reported fire incidents. TABLE 4: Average Cleanup Cost Cost Tvp Sidewalk Vault Analysis/Testing Cleanup/Decontamination Disposal Equipment Replacement $ 233 2,467 5,000 20.000 TOTAL $27,700 Inside Building $ 16,625 144,375 8,750 109,175 $278,925 11 HONS 017069 2.6 Utility Hisk Reduction Measures Survey data indicate that utilities have undertaken a variety of measures to reduce the likelihood of PCB fire incidents. More specifically 78% of surveyed utilities have conducted risk-assessmenc inspections of their askarel transformers. 76% of surveyed utilities have askarel transformer phase-out programs completed or currently in effect. 35% of surveyed utilities have notified fire departments in their service area of the location of askarel transformers. 20% of surveyed utilities have posted warning signs for firefighters in the approaches to askarel transformers. - 75% of the distribution system askarels have a protective fuse, switch, or breaker on the low voltage side of the transformer, approximately two percent of which are oil-filled. These pro tective devices are designed to isolate (de-energize) the trans former in the event of a fault in the low voltage line. 19% of the distribution system askarels have a protective fuse, swatch, or breaker on the high voltage side of the transformer, approximately 52% of which are oil-filled. These protective d .-vices are designed to isolate (de-energize) the transformer in t.:e event of a fault in the high voltage line. Askarel phase-out programs are currently in effect in 76% of the surveyed utilities. These programs are primarily directed toward phasing out distribution system askarel transformers, although 13% of these utilities also have reported programs and plans for phasing out substation and generation facility askarels. Based on the phase-out programs voluntarily instituted among the utilities, it is estimated that the number of distribution system askarel transformers will be reduced from 18,291 to 12,987 by the end of 1985. This represents a reduction of 29%. By the end of 1989 the total will have been reduced to 6,102, a reduction of 65%. Table 5 shows the projected reduction in distribution askarels through 1994, given the phase-out programs currently in effect. 12 kONS 017070 TABLE 5: Estimated Reductions in Distribution Askarel Transformers Based on Current Phase-out Proprams Year T35T 1985 1986 1987 1988 1989 199C 1991 1992 1S93 1994 Estimated Number of Distribution Askarels 18,291 12,987 10,792 8,780 7,499 6,402 5,670 4,939 4,207 3,658 2,927 Percent Remaininr --ITXT-* 71 59 48 41 35 31 27 23 20 16 A variety of priority considerations are being taken into account by the utilities in their phase-out proprams. The most common factors considered pertain to the location of askarel transformers within or in relation to the location of buildings. The type and location of the buildings in which these transformers are located are also being considered, with schools, hospitals, commercial and other buildings in congested areas receiving higher priorities. Several utilities are ' prioritizing askarels for phase-out based on the assessed risk to food, feed, and water supplies. Other factors being considered are the age, condition, and voltage of the askarel equipment. Seventy-eight percent of surveyed utilities have conducted risk assessment inspections of their askarel transformers to evaluate exposure risk. These inspections have often been performed in conjunction with phase-out planning. Ten percent of those utilities conducting risk assessments have increased the frequency of their askarel transformer inspections as a fire risk reduction measure. An additional 10% have sealed cable ducts and other holes in transformer areas in order to contain smoke in case of a fire. Other efforts reported to reduce fire risk include installing fire detection or fire suppressant systems within vaults, equipping vaults with a coolant sensor, scaling excess grating area or sidewalk vaults, and assuring that a COj fire response vehicle is available. 13 HONS 017071 2.7 Retrofit Experience Of the 52 utilities surveyed, 11 reported having1 rstrofiiled one or nor* askarel transformers, with a total of 48 askarel transformers retrofilled among these utilities. The average cost of retrofitting reported by those providing cost data was 515,700 per transformer Approximately 45% of the costs reported were 55,000 or less, while the remaining 55% ranged between 510,000 and 541,000. The range in EVA rating and fluid capacity of the askarel trans formers retrofilled is shown in Table 6. TABLE 6: EVA Rating and Fluid Capacity of Retrofilled Transformers Fluid Capacity (in gallons) 5100 101-200 201-300 301-400 401-500 501-600 601-700 701-800 801-900 TOTAL Transformer KVA (Nameplate P.atir.r; 101-500 bQl-looo lTTOT-aOuQ TZ 700 7 1 00 1 14 9 2 25 012 3 023 5 00 1 1 002 2 0 00 0 0 05 5 22 . 12 15 49 None of the retrofilled transformers had nameplate ratings less than 100 EVA or higher than 5,000 EVA. PCB levels reported for these transformers prior to retrofiHing ranged from 400,000 ppm to 750,000 ppm. Over 60% of the transformers had starting PCB levels estimated at 700,000 ppm or above. In 27% of the retrofttls, a flush/refill method was used, and in 35% of the eases s filtering process was used. The remaining 38% were retrofilled using both the flush/refill and filtering methods. In several esses the filtering process was reported to be ineffective. 14 HOMS 017072 whereas in other cases the flush/refili process alone was considered insufficient. The fluid used in retrofflling 4C% of the transformers was silicone. RTemp was used in 29% and oil in 31% of the transformers. A variety of flush solvents were used in these retrofills. Table 7 identifies these solvents along with the frequency of their use. TABLE 7: Flush Solvents used in Retrofills Solvent Number of Retrofills Silicone Trichtore thane Trichlorethane and Silicone Chlorothane Trichlorobeazene Oil Rtemp None No Response 3 1 1 10 7 11 1 3 _4 TOTAL 49 Total does not equal 100% due rounding % of Total 6 3 2 20 14 22 2 15 3 98* In order to assess the effectiveness of the retrofill methods, respon dents were ashed to report the PCB level after their 1st, 2nd. ar.2 3rd refill or filter change. Data on PCB level were provided or. 40 transformers and are summarized in Table 8. 15 HONS 017073 TABLE 8: Retrofill Results Subsequent PC3 Levei (asm) Beginning Type 1st 33 3Fa~ PCB Level** Process* Flush/Filter Flush/Filter Flush/Filter 1,000,000 1,000,000 750.000 750,000 700,000 700,000 700,000 700,000 700.000 700.000 700,000 700,000 700,000 700,000 700,000 700,000 700.000 700,000 700,000 wt ,LVJ 700. COO 700,000 600,000 600,000 600,000 600,000 600,000 600,000 600,000 500,000 500,000 500,000 400,000 400,000 400,000 400,000 400,000 400,000 400,000 400,000 B 703 B 1,251 F F-- B 779 B 760 3 8,673 B 10.103 3 3,744 R 42,000 R 32,000 R 53,000 R 31,000 R 24,000 R 36,000 R 14,200 R 81.0C0 R 1,200 R 3,200 R 1,000 B 57,000 B 47,000 F 1,180 F 578 R 269,016 R 274,690 F 20,000 F 7,075 R 150 F 22 F 19 F 172 B 158,094 B 48,220 B 15,040 B 34,575 B 13,737 B 25,080 B 14,735 B 14.665 43 85 -- 788 130 .. 2,806 2,753 6,443 3,000 2,200 4,600 5,000 1,800 3,200 .. .. .. .. 32,200 3,149 11 25 .. .. 5,000 590 10 7 95 -- 105,840 30,942 10,157 23,087 9,171 16,817 9.849 .9,822 237 316 20 27 472 6,913 6,175 3,416 4,510 4CQ 240 850 2,100 560 300 4,500 37 10 91 Berin-E 3 3 13 3 30 30 6 33 21 *7 33 3 > O 3 5 4 1 2 2 0 2 2 0 2 2 * F * Filter; R * Flush/Refill; B * Combination ** For these transformers whose beginning PCB level was reported as askare a 70% PCB concentration was assumed. 16 HONS 017074 These survey data indicate that 13% of the transformers attained a PCB level of SSOO ppm after the first refill/filter change. Of those . transformers Pefiiled/filtered a second time, 26% attained a PCB level of $500 ppm. After the third refill/filter change, the PCB level in the 13 transformers were lowered to $10,000 ppm and 56% of these 13 attained a PCB level of $500 ppm. However, as is readily apparent, the retrofill results are quite erratic. Results of the first refill/futer change range from 19 ppm to 274,690 ppm, the second refill/filter change from 7 to 105,840 and for the third refill/filter change from 10 to 6,913. There is no apparent correlation with time, beginning level, or process type. In eight cases, the PCB levels increased between refills/filter changes rather than decreased. This was reportedly due to PCB being retained in the paper insulation and windings. Even where PCB levels are apparently decreased substan tially after three refills/filter changes, comments by respondents indicate the possibility of future increases due to continued PCB leaching. Utilities have been spending an average of eleven months between first and third refill/filter change procedures. Thirty-seven percent of those reporting indicated spending two months or less. Twenty-one percent, however, indicated spending over two years. As stated, there appears to be no relationship between the length of time spent for retrofilling and the PCB levels reached. 17 HONS 017075 3.0 COMPARISON OF STUDY DATA WITH EPA MATERIALS 3.1 General' With refard to risks posed by askarel transformer fire incidents, EPA states in the ANPR: If EPA is not provided with adequate data, especially in the areas or the risks posed in the event or a fire, the probability of these fires occurring, and the costs associated with clean-up following these incidents, EPA will sake its regulatory judgments based upon the data set forth in this document. These data indicate that PCB Transformer fires pose relatively high risks, occur with unknown frequency, and can result in relatively high clean-up costs. This report provides substantial empirical data regarding the frequency with which transformer fires occur and the costs of cleanup. A comparison of these data with the estimates developed in the preliminary EPA analysis is presented below. The ANPR also solicited general comments regarding the preliminary EPA cost-effectiveness analyses. As part of our study activities, the EPA analyses were reviewed, and the final section of this report provides our commentary. 3.2 Frequency and Nature of Fire Incidents In the ANPR, EPA presents a variety of estimates of the frequency of PCB Transformer fire incidents, ranging from a low of an expected 8 incidents per year to a high of 1,530 per year. In the preliminary study and cost-effectiveness analyses, the EPA consultant estimates that the number of expected annual catastrophic incidents will be between 0.6 and 11.* 6. The preliminary EPA cost-effectiveness analyses estimates the low probabihty of a catastrophic incident as 0.0005 percent, the high probability as 0.01 percent, and the beginning (1984) transformer population to which these rates apply as 114,469. 18 HONS 0170 76 Developing data comparable to ties* estimated frequencies is difficult because no clear definition of fire incident or catastrop* hic incident has been developed by EPA. In conducting the utility survey, a fire incident was defined as a fire in which a transformer has been directly exposed to a fire, either from within the transformer or an immediately adjacent source. Based on this definition, survey data indicate that one fire incident per every 7,787 askarel transformers can be expected each year. This suggests approximately S fire incidents per year involving utility transformers. If the definition of catastrophic incident is related to cleanup costs as was done in the preliminary EPA cost-effectiveness analyses, only one incident involving a utility-owned transformer in the last three years would be classified as catastrophic. Another way to estimate the likelihood of a catastrophic incident is to determine the number ei askarel transformers located in areas which might result in a Bingham::?, or San Francisco type incident, and estimate the likely frequency oi a fire involving one of these transformers. Data obtained via the survey allowed reasonable estimates of these numbers. Approximately 2,700 transformers are located inside buildings, in areas adjacent to rooms with heating or ventilation system intakes. Another 1,300 are in sidewalk vaults immediately adjacent to building heating and/or air conditioning systems. These 4,000 transformers might be considered as potential catastrophic incidents if the definition relates to the type of incident occurring in Binghamton or San Francisco. Given the reported frequency of askarel transformer fires of 0.01 percent, a catastrophic fire incident of the Binghamton/San Francisco type involving a utility-owned transformer can be expected approximately once every 2.5 years. 3.3 Cleanup Costs Insofar as fire incidents are concerned, the preliminary EPA costeffectiveness analyses focuses solely on catastrophic incidents and assumes t cleanup cost of $20 million. Survey data indicate that the average cleanup cost for s non-cstastrophic transformer fire incident outside a building is $27,700 and inside a building is $278,925. The 19 HONS 017077 single catastrophic incident reported in the survey was estimated to require SIS to 20 million in cleanup costs. 3.1 Comments Regarding EPA Cost-Effectiveness Analyses A preliminary examination of the EPA cost-effectiveness analyses reveals four significant issues: - The estimated frequency of catastrophic incidents may be overstated The methodology used for calculating phase-out costs may be inappropriate. - The assumptions underlying leal-tag* rates used for calculating the quan ciues of PC3s released from askarel transformers may be inappropriate. ' - When calculating PCB exposure, the benefits of spill cleanup are ignored, yet the costs or cleanup are considered. The net effect is to overstate the cost-etfecUveness of additional regulation 3.4.1 Frequency of Catastrophic Incidents The EPA analysis uses a low probability of a catastrophic incident of 0.67 incidents per year and high probability of U incidents per year. As indicated in section 3.2 of this report, survey data indicate a catastrophic fire incident involving a utility-owned transformer is likely to occur approximately once every 2.5 years (0.40 per year). The EPA low probability estimate is reasonable; however, the EPA high probability estimate substantially overstates the apparent risks. 3.4.2 Phase-out Costa The methodology used by EPA for calculating phase-out costs is not clear. The preliminary study indicates the methodology used for the 1982 rule is used with minor modification. We criticized the 1982 methodology in comments submitted to EPA on May 24, 1982. To the degree the 1984 EPA analysis has not incorporated changes resulting from our previous comments, the EPA estimated phase-out costs are understated. 20 HONS 017078 Estimating incremental phase-out casts should contrast normal utility equipment replacement during the phase-out period with replacement . which would be required during the same period as a direct result of the mandated phase-out. The costs associated with the incremental replacement units are the incremental phase-out costs inasmuch as they reflect only those costs associated with the purchase of units phased out prior to the end of their useful life. The 1382 EPA analysis estimated costs that would be incurred by utilities over a 30-year normal replacement period and compared those costs with costs incurred over a 6-year phase-out period. If the 1932 EPA methodology for calculating incremental phase-out costs was used in 1984, the EPA costs are understated. 3.4.3 Askarel Transformer Leakage Rates Again, the 1984 EPA methodology for estimating PCB releases from askarel transformers is not clear, but is apparently the same as that used for the 1982 rule. If the same methodology has been used, it can be summarized as follows: 1. The percentage of askarel transformers expected to fail in a year is estimated. 2. Tbe percentage of failures expected tc result in PCB spills is estimated. 3. Tbe percentage of askarel transformers expected to leak without failing is estimated. 4. Tbe quantity of PCB fluid lost per spill is estimated. 5. The quantity of PCB fluid lost per leak is estimated. 6. Using estimates from Steps 1-5 and an estimate of thenumber of askarel transformers in service, the total annual lossof PCBs is estimated. Conceptually, the approach is fine. However, the methodology whereby estimates are developed by EPA for steps 1-3, i.e., the basis for determining tbe critical factors of failure, spill, and leakage rates, is not clearly defined in either the 1982 or 1984 EPA analysis. These 21 HONS 017079 dau are cited as being based on data obtained from the 1981 EEI.'l'SWAG surrey and industry comments. No precise citations or methodology are provided* The data collected by EEI via the 1981 utility survey do not provide any information on failure rates inasmuch as reliable data on this were not available from the utilities. Similarly, no specific data were collected and reported by EEI regarding the per centage of failed transformers which spill or the percentage of trans formers which leak without failing. Although the EEI data may have been used in conjunction with other information to estimate these factors, the EPA analysis does not describe the methodology used. Because the procedures whereby EPA estimated spill quantities are substantially undocumented, it is difficult to assess whether the benefits calculated are realistic. 3.4.4 Spill Cleanup . A further problem with the EPA cost-effectiveness analyses is the treatment of PCB spill cleanup. Stated simply, although EPA recogr.lres that utilities clean up PCB spills and attributes costs to these activities, they ignore the effect of cleanup when calculating benefits (defined by EPA to be gross pounds of PCBs spilled). In their 1982 analysis, EPA indicated that no empirical data were available which could be used to estimate the degree to which PCB spill cleanup is effective. EPA also indicated that cleanup costs were estimated by the utility industry to range from $100 to $1,000,000. Despite the fact that no empirical cost data were available except this broad range of estimates, EPA made assumptions regarding the cost of cleaning up a PCB spill. These assumptions are apparently also used in the 1984 analysis. Although the precise effects of these estimates on the EPA cost model cannot be determined, estimated phase-out costs are reduced by offsetting cleanup costs that are avoided as a result of phase-out, i.e., industry will avoid costs because equipment cannot spill if it is removed from service. Reducing estimated phase-out costs due to cleanup activities without considering the benefits (pounds spilled) due to cleanup, produces an artifically inflated estimate of the cost-effectiveness of phase-out. 22 HONS 017080 Existing EPA regulations governing the cleanup and disposal of ?C3s are sufficient^ to ensure that spill cleanup efforts by utilities rescue. in excess of 99% of all PCBs spilled from utility-owned equipment.; EPA were to accept the effectiveness of existing regulations, the gross pounds of PCBs estimated to enter the environment without a phase-out would be 99% lower than current EPA estimates (EPA assumes cleanup has no effect on exposure). This would have the effect of substantially reducing the benefits attributable to phase-out, thereby substantially decreasing the cost-effectiveness of phase-out options ?. Assuming a reasonably timely cleanup response, cleanup to any level below 1,000 ppm should conservatively remove 99% of the volume of PCBs spilled. 23 ONS Q17081 APPEND LX A LIST OF UTILITIES IN SURVEY HONS 017082 Alabama Power Company Appalachian Power Coapany Baltimore Sas & Electric Boston Edison Coirpaay Central Power 4 Light Conpany Coluabua & Southern Ohio Electric Company Cossonwealth Edison Consolidated Edison Co. of New York Delnarva Power & Light Coapany Duke Power Company Duquesne Light Coapany Florida Power 4 Light Co. `Florida Power Corporation Georgia Povsr Ccrpony Idaho Power Coopany Illinois Power Conpany Indiana 4 Michigan Electric Coopany Kansas City Power 4 Light Coapany Kansas Gas 4 Electric Coapany Kentucky Utilities Coapany Los Angeles Departoent of Water 4 Power Louisiana Power 4 Light Coapany Louisville Cos 4 Electric Co.npasy Meaphis Light, Gas 4 Water Divisioa Metropolitan Edison Coapany Minnesota Power 4 Light Coapany New England Power/Massachusetts Electric Co. Niagara Mohawk Power Coopany Northern Indiana Public Service Coapany Northern States Power Coopany Ohio Power Coapany Oklahoaa Gas 4 Electric Coapany Pacific Gas 4 Electric Pacific Power & Light Coopany Pennsylvania Electric Coapany Pennsylvania Power 4 Light Coopany Public Service Coopany of Colorado Public Service Coopany of Indiana, Inc. Public Service Coopany of Oklahoaa Public Service Electric 4 Gas Coapany Sacraoento Municipal Utility District Southern California Edison Southwestern Public Service Coapany Taopa Electric Coopany Tessa Electric Service Coopany ' Tones Power 4 Light Coopany The Toledo Edlaon Coopany Union Electric Coopany Utah Power 4 Light Coapany Virginia Electric 4 Power Coapany Vest Pena Power Coapany Wisconsin Electric Power Coopany Hoars spendonc HONS 017083 APPENDIX 3 SURVEY QUESTIONNAIRE HONS 017084 EDISON ELECTRIC INSTITUTE (EEI) ' and UTILITY SOLID WASTE ACTIVITIES GROUP (USWAG) ASKAREL TRANSFORMER SURVEY MONS 017085 Introduction On March 23, 1984, the U S. Environmental Protection Agency (E?A) published an Advance Notice'of Proposed Rulemaking (ANPR) regarding the use of PCBs in electrical transformers. The ANPR Indicates EPA concerns that inadequate consideration was given to the risks of transformer fires prior to issuance of the August 25, 1982, rule governing the use of electrical transformers containing PCBs. In order to determine whether additional regulation is required. EPA has requested a variety of information regarding transformers and fires. USWAG has retained the services of Resource Planning Corporation (RPC) of Washington, DC., to accumulate the required data and assist in its presentation to EPA. As with the 1981 survey of PCB usage conducted by RPC for USWAG, the study is designed to cover a significant portion of the relevant equipment universe by scientifically sampling a limited number of the largest utilities Since a relatively small number of utilities are being asked to provide data, it is imperative that you fully participate to ensure the validity of the findings ALL RESPONSES PROVIDED IN THIS SURVEY DOCUMENT ARE CONFIDENTIAL THE RESULTS OF THIS STUDY WILL BE REPORTED IN THE AGGREGATE AND UNDER NO CIRCUMSTANCES WILL INDIVIDUAL RESPONSES BE ASSOCIATED WITH AN INDIVIDUAL UTILITY. To facilitate follow up for the purpose of clarifying responses, you are requested to indicate the name of your utility, and the name and phone number of a person to contact if there is some question in processing your responses Name of Utility: ' Person to Contact: Phone Number: *2` HONS 012086 To sect the time requirements imposed by EPA, all responses must be received NOT LATER THAN MAY IS, 1984. Survey forms should b? mailed to: EEI/USWAG Counsel c/o Toni Allen, Esq. Wald, Harkrader k Ross 1300 Nineteenth Street, N.W. Washington, D.C. 20036 All questions or inquiries regarding this survey should be directed to: Rick Bell or Tom Florence Resource Planning Corporation (202) 797-1111 General Instructions This questionnaire is comprised of four major parts: I. GENERAL DATA II. FIRE INCIDENT DATA III. ASHAREL TRANSFORMER DATA IV. RETROFILL DATA Each major questionnaire part has its own instructions. Please read all instruc tions carefully is you complete the questionnaire. To ensure uniformity in the reporting of information, please use the following definitions: Askarel Transformer: Any transformer designed and built with an insulation system using askarel fluid. PCB Transformer: Any mineral oil transformer that contains 500 ppm of PCB or greater. Distribution System Transformer: Any transformer not located in a substation or generating facility. Transformer Fire incident A fire in which a transformer has been directly exposed to a fire, either from within the transformer or sa immedistely sdjaccnt source. If the transformer was not fire damaged, then it should be considered s transformer fire incident only if there were visible signs of fire exposure to the inside of Che transformer compartment or the exterior of che casing. 3- MQNS 017087 I. GENERAL DATA This portion of .the questionnaire is intended to obtain general information necessary to respend to the ANPR- Where information is not readily available, please provide the most complete response possible within the time limits of the study. 1. Indicate the number of askarel transformers owned by your utility as of March 31. 1984: Size--KVA Distribution Generation (name- System Substation Facility late rating)_____ Askarels______Askarels_________ Askarels 553 Wl-TTWfl----------------------------------------------------------------------------------------- >5'.oro~ ~ .. ~ ~ -----mrr.oco------------------------------------------------------------------------------------- Total Askarels 2. If available, indicate the number of transformers containing askarel substitutes (i.e., silicon, RTemp. etc.) owned by your utility as of March 31. 1534: Si2e-KVA (nsme- glate rating) ilCl Distribution System Substitutes Substation Substitutes Generation Facility Substitutes Total Substitutes sroaic-rro-yco.wr >i.ooo " 3. Have you conducted a risk-assessment inspection of your askarel transformers' No Yes If yes, briefly describe actions taken or planned to reduce fire risk (attach additional sheet if necessary) 4. Do you now have an askarel transformer phase-out program in effect? No Yea 1/ yes, briefly describe the phase out methodology, i.e., priority of removal, etc.: _ wnat rear do you anticipate completion of the phase-out program? ~ 5. Have fire departments fa your service area been notified of the location of your askarel transformers? ____No ____ Yes 6. Are there any warning signs for firefighters posted in the approaches to askarel transformers? ____No ____Yes hons owoea -4 13. If known, give the approxucat* time laps* from: a. the beginning of the fire to transformer de-energization hrs b. occurrence to fire department response ~ jj.-s c. occurrence to extinguishment hU'r. s. (indicate tenths~o7 hour if <; 14. Approximately how old was the transformer involved: ___ yrs 15. Did the switch, breaker, or fuse on the low voltage side of the transformer open as designed? ___ no ___ yes ___ no switch/breaker/fuse on the low voltage side 15. Did the switch, breaker, or fuse on the high voltage side of the transformer open as designed? __ no ___ yes ___ no switch/breaker/fuse on the high voltage side Was there any early warning device (alarm or sensor)? no yes If yes, what type? EI3 it function as designed? ___ no yes IS. For incidents occurring inside a building, characterise the extent af smoke travel from hoe fire. ___ smoke was restricted to the transformer enclosure (e.g., vault or rccir. ___ smoke escaped the transformer enclosure and entered immediatelv adjacent areas, but did not enter entire building ' ___ smoke entered the entire building other; describe ______ ____ 19. For incidents occurring outside e building, characterize the extent of smoke travel from the fire: ___ smoke was restricted to the transformer enclosure (e.g., vault or room) ___ smoke escaped the transformer enclosure, but did not enter any public buildings ___ smoke entered nearby public buildings other; describe 20. Excluding fire fighters and utility personnel, approximately how many people were exposed to smoke/soot from the fire? ___' people 21. Approximately how many firefighters and utility personnel were exposed to smoke/toot from the fire? ____ firefighters/utility personnel 22. Indicate the results of any testing: PCB PCDDs Before After Before After Cleanup Cleanup Cleanup Cleanup Trf. fluid Soot _ PCDFi before After Cleanup Cleanup 6- MQNS 017089 II. FIRE INCIDENT DATA This portion of the questionnaire is designed to obtain data on ALL FIRE ixcin?\-- INVOLVING ASKAREL TRANSFORMERS. PCS TRANSFORMERS OTTTRaN5"0RMF-i ' " CONTAINING ASaARki o^mi-TEg 51'RINC THE LaST 3 'Il-aRS tAMirr^ 1531-MARCH 31. 1S3-4). A fire incident is defined as a fire in wtucn a transformer has been directly exposed to a fire, either from wichm the transformer or an immedi ately adjacent source. U the transformer was not fire damaged, then it should only be reported if there were visible signs of fire exposure to the inside of the transformer compartment or the exterior of the transformer casing. Complete ques tions 7*25 for EACH fire incident. Make additional copies of this form as necessary 7. Indicate the si2e and type of transformer involved in the incident: _____ KVA __ askarel _____ PCB (>500 ppm) ____ askarel substitute (high-Fire pc:.-.:: If askarel substitute, specify fluid type:____________________________ 8. Indicate the approximate time, day of the week, and date of the incident time day (Mon.Tue.etc.) month/day/year Location of incident: ___ Inside building ___ Residential (condo, apt., hotel) Institution (school, hospital, church) ' ~ Office ___ Industrial/manufacturing ___ Storage (warehouse) ___ Other public (store, mall) Other, specify __ ___ 3I3ewalk vault ~ Other; specify 10. If the incident occurred inside a building, indicate: a. age of building: ___ years unknown, can't determine b. location within building: ___ basement roof other; specify_ c. no. of floors in the building: ___ floors d. environment of transformer: .11 Cause of fire: ___ within transformer ___ outside the transformer; specify if known_________ vault room open area other; specify 12. Was transformer fluid released and exposed to the fire as s result of damsgs to the transformer, a leak, or rupture? ___ no ___ yes If yes, approximately how much fluid was exposed: ___ gals. (Indicate tenths of gallon if <1 gal.) -5MONS 017090 23. Briefly describe the area cleaned and method of cleanup, i.e.. wall, fleer, ceiling area, HVAC system, etc. (attach additional sheets if necessary) Area Method of Cleanup 24. In the table below indicate approximate cleanup costs. When estimating cleanup costs be sure to consider manpower, equipment, transportation and any relevant overhead charges: Cost Type Cost Estimate Analysis/test costs: _______________________________________ Site deanup/decontamination: Dispos al ____________________________________ Vault/traasforaer/equip. costs: Other (describe): TOTAL: 25. Have there been any lawsuits filed against your company as a result of this incident? ___ no ___ yes If yes, briefly describe nature and amount of damages claimed: -T HONS 017091 III. ASKAREL TRANSFORMER DATA In responding to te EPA data requirements, it is necessary to develop informa tion retarding the location and environment of askarel transformers. UNLIKE THE FIRE INCIDENT PORTION OF THIS SURVEY, THIS PORTION REtf THAT YOU EXAMINE"AND REPORT ON A SMaLL'SCIENTIFICaUV SEI--. SAMPLE OF WR~'PrSTRiaCTIOH SYSTEM aSKaREL TRANSFORMERS l. ASKAREL TRANSFORMERS WOT LOCATED IN~5UaSTATIONS Oft GEVERaTO'G FACILIT155T Based on data provided in the 1981 PCB survey, we have identified the specific askarel transformers on which you are to report. Please follow " these instructions in identifying those units: Prepare a list of all your askarel transformers that are not located in substations or generating facilities. This list can be in any sequence. l.e., by serial number, location code. etc. Beginning with the ___ transformer on the list, select every transformer until yoTThave Identified a coral sample of __ transformers (When you reach the end of the list, go back to the beguining and continue your count until you identify the required number of transformers .) Proper selection of these transformers is critical to the study. If you have questions or problems in performing this task, call Rick Bell or Tom Florence (RPC) at (202) 797-1111. Once you have identified the specific transformers for which data is required, answer questions 25-40 for each transformer selected. Note that if your sample requires data on ter. transformers, you must complete ten sets of questions 26-41 Sufficient copies of this form have been included for your sample. -8- MONS 017092 Answer the following questions far each transformer selected in your sasqle 26. Of the___ transformers in your sample, where in the order was this transformer Selected (1st selected. 2nd selected, etc): Selected 27. Indicate the nameplate rating of the transformer: 28. Indicate the askarel fluid capacity: 29. Specify the transformer location: ___ Inside building ___ Residential (condo, apt., hotel) Institution (school, hospital, church) ___ Office ___ Industrial/manufacturing ___ Storage (warehouse) ___ Other public (store, mall) Other; specify ______________________ ___ 5I3ewalk vault ___ Other; specify ____________________________ KVA gals. 30. if transformer is inside building, indicate: a. age of building: ___ years b. location within building: ___ unknown, can't determine __ basement ___ roof ____ other; specify c. no. of floors in the building: ___ floors " d. environment in which located:___ vault ___ room ___ open area ___ other; specify 31. Check the response which applies to the transformer drainage path: fluid would be contained, i.e., no drain, or drain has been sealed ___ fluid would drain to contained catchment area ___ fluid would drain to sewer system __ fluid would drain to surrounding soil ___ other, specify: 32. Are vault drains required by local building codes? ____ no ____ yes 33. Check the response which applies to the transformer cooling ventilaticn system: ___ no ventilation; transformer in sealed room or vault ___ ventilation through grate in wall/ceiling to outside area ___ ventilation through grate in wall/ceiling to indoor area ___ other; describe:` 34. Is the transformer cooling ventilation system forced or natural? ___natural __ __ forced none If forced, is it designed to atop in the event of a fire? ___no ____yes -9- MONS 017093 35. If this transformer is located inside building, characterize the poten tial spread of^smoke and soot in the event of fire (answer a-d.i: a. Is there an air conditioning or heating system intake inside the transformer enclosure (e.f., vault or room where transformer is Located)' ____no _____ yes _____ transformer not enclosed b. Is there an air conditioning or heating system intake in the area immedi ately adjacent to the transformer enclosure (e g., vault or room)' ____no _____ yes ____ transformer not enclosed c. Other than a yes answer to 35a or 3Sb above, would smoke/soot have other indirect access to the building air conditioning or heating system ___ no ___ yes If yes, describe access: d. Would smoke/soot have direct access to the air conditioning or heating system of any other building? ___ no ____ yes If yes, describe access: 36. If this transformer is located outside a building, characterize the poten tial spread of smoke and soot in the event of are (answer a & b): a. Is there a building air conditioning or heating system intake immediately adjacent to the transformer enclosure' ____no _____ yes ' b. Other than a yea answer to 36a, would smoke/soot have any ocher access to the air conditioning or heating system of nearby buildings' no ___ ye* ' If yes, describe access:________________________________________ 37. Is the transformer protected on the low voltage side by a fuse, switch, or breaker? __ no ........ yes . If yes, is the device oil-filled? ____no _____ yes 38. Is the transformer protected on the high voltage side by a fuse, breaker, or switch et the trensforaer locetion (exclude protection at the substation)' ___ no yes if yes, is the device oil-filled? ____ no _____ yes 39. Is there a smoke or fire alarm or other early warning device or sensor near the transformer? ___ no ___ yes If yes, describe the device 40. Is the transformer protected by a sprinkler system or other fire retardant system? __ no ayetea __ sprinkler system __ other system; describe -10' HONS 01709* IV. RETROFILL DATA This portion of tlve questionnaire is designed to obtain data on each askarel transformer you have attempted to retro/ill. To the extent data are available, pleas* complete questions 41 and 42 for each transformer. Make additional copies of this fora as necessary. 41. Indicate the number of askarel transformers your utility has retrofilied ___ transformers 42. The table below provides space for recording data on two transformers Please record the requested information for each transformer your utility has retrofilied (or is retrofiliing). If space is required for more than two transformers, please make additional copies of this form. When reporting costs please include all associated costs, including labor, materials, disposal, and any additional fire protection costs result ing from use of the new fluid. If additional fire protection was required, please describe and provide specific cost estimate in the comments section. Transformer 1 Transformer 2 a. Transformer KVA (nameplate rating) b. Fluid capacity c. Beginning PC3 level (prior to retrofill) d. Retrofill process (flush/refill or filtering) e. Retrofill fluid f. Flush solvent g. PCB level after 1st refill (or 1st cartridge change if filtering process) h. PCB level after 2nd refill (or 2nd cartridge change if filtering process) 1. PCB level after 3rd refill (or 3rd cartridge change if filtering process) j. Time elapsed from initial draining to last refill (or cartridge change If filtering) (months) k. Approximate costs to date (consider labor, materials, equipment, transports' tion, overhead) Comments: 11MONS 017095 EPRI Technical Newsletter Electrical Systems Division ?C3 TECHNICAL REPORT FROM I?RI NUM2ER 8 J. *C r* : ''NQ.v vt s : ,v ` ? :* s ;j: ** :< c * >^3C3 s3` iii:::: HONS 017096 SOURCE 2028-3 1263-14 2028-4 2029-3 03 13 1263-11 e?r: PDCF PROJECTS Removal of PCBs from Transformer Oil ?C3 Disposal Manual (2nd edition) Pyrolysis and combustion products of PCB, tri-, and tetri-chlorinated benzene CCMPESTICH DATE June 83 September H34 November i?3i Analysis of PCDFs in insulating fluids State-of-the-Art Review: PCDDs and PCDFs m Utility PC3 Fluid ICC 3308) March 1985 Complete December 1983 .mitigation 1263-20 Documentation of building clean-up techniques (Binghamton and San Francisco) 1263-21 Engineering options for ?C3 transformer risk reduction RISK ASSESSMENT AND MANAGEMENT 1926-13 Exposure assessment for PC3 spills and fires 1926-15 Risk mangeoent of PCB fires and spills 2599 Decision framework for PC3 decisions HEALTH EFFECTS 2374 Health effects assessment of PCBs SUBSTITUTE INSULATION 1499-4,-5 Arc and spark byproducts of perchloroethylane 2028-11 Partial and complete combustion of transformer insulating materials 2028-12 (pending) State-of-the-Art Toxicological Reviews of PCS Substitutes October 1594 Octccer 19ii December 1594 December 1394 July 1994 mid year 1985 August 1984 December 1984 December 1984 HONS 017097 i tie: Removal of PC3s from Transformer Oil Contracts: RP2328 - 3 Contractors: Peter Way (-3) Project Manager: Oil Addis Purpose: The objective is to develop and demonstrate on a pile scale a process capable of reducing PCS contamination in transformers below the then EPA acoeptabi1ity level of 33 opm. Scope of Worlt; In an earlier project, RP2023-1, General Electric, after testing in the laboratory chose the solvent extraction process for scale-up. Other processes tested were: sodium reaction, not selected because significant ccmmertial activitiy was under way; electron-beam irradiation; and critical-fluid. The latter two were found to be impractical in me laboratory. A relatively safe and economical solvent extracti; process was successfully scaled up to a 10 gallon per hour pile plant, and plans made to design a full scale plant (RP2025-3). Expected Results; Successful operation of a large-scale sciven extraction plant for removal of PC3s from contaminated mineral oil would provide an alternative, economical means for decor.tam Statusi A 500,000 gal/year plant has been designed and costed. The proposed host utility for this plant has completed site plans. The project is presently on hold to evaluate the epa's recent interpretation requiring a level of 2 ppm PC3 in the treated mineral oil. Completion Date; 1985 / MONS 017098 Title: PC3 Disposal .Manual, second Edition Contract: * RP1263-14 Contractor: SCS Engineers Project Manager: Ralph Komai Purpose: This project will update the reference manual prepared in 1979 as a guide for utilities in evaluating disposal options. Scose of worlc The Contractor will prepare discussions of the available disposal requirements and disposal options, including the chemical treatment technologies which have beeen developed since the first edition. Chemical treatment will include both destruction and retrofill. There will be an update of the estimated ?C3 inventory. Management of PC3s also will be addresss Expected Results: A revised manual will be available to the utility industry to assist in evaluating and planning management options. Status: A draft report has been prepared and is in the process of being revised. The second draft will be distributed to selected industry people for technical review prior to finalisation and publication. Completion date: September 1984 HONS 017099 1 Title: Pyrolysisand Combustion of ?C3 Csrtimiwatto C'elesvt .< ;5 Contract: . P.PZCZS-* Contractor: New 'for* State Cesartmant of health Project .Manager: Sil Addis Purpose: To determine one extant of formation o' PC"/:C" as a ;r;dat: o' pyolysis or comtustion of oot.o i "C'i aslearels ant 3s<s''s a; a-t csnta~.ir.ar.ts in mineral oil or several rstrsfill 'lu'oa. Scope o' he'V: The observation that PC3a -Oder conoitions of oyro'ysis or Comoustion can be converteo to much more toxic ooiycrlorinated oitertofirars (?C"s! has led to heightened interest in the possibility of retrof* 1 "no currently in-vss "C3 trarsfortrs with a Cie'.ect'it vn'i'xa'y to .rte'to such t'ir.sfcr--ations. however, even the nost care'll ret'ofiilir; jrccess i'l lea.a a resio-al amount of ?C3 in the new dielectric fluid. A pr-. tier* is no reason to expect that such low concentrations of ?C5s culo orodjce d'sorooorv:-ately high concentrations of PCCFs on pyrolysis or comtustion. however, since the details of the mechanisms by which this conversion actors ere not t:! established, it is dearly desirable to test this possibility experiments'!/ before launching a major retrofilling program. In a second aspect of this -o'<, the level of formation of PCOFs from mineral oil contaminated with PCHs at a ':w level is of even more importance because of the large number of contamiratei transformers Involved. Expected Results: We expect to determine the linearity of formation of PC" as a pyrolysis or combustfon product with decreasing PC3 concentration in a host HONS 017100 liquid, tithe' itineral oil qr ret.-sfil'. fluid. Status: Pyrolysis end carOustion techniques rave been teveiooed. 3a- 'a elea.n-ua netnods, particularly for pyrolyzed or condusted contaminated -irt-a! Oil. art nearing success. Analytical results will be available in the -.ear future. Co~t'et::n Cate1. .Hcuenber 153a HONS 017101 * -Title: "' . PCDF ar.d PCDD in Askarel and Contaminated * oil Equipment Contracts: Contractors: RP2Q28-5,-10 .Radian (-5), Sattelle <-S), New York State deal: Dept. (-7), General Electric (-81, IITRi (-}), Research Triangle Institute (-10) Project Manage Gil Addis Purpose: to improve tee state of the art of analysis for PCCr To determine under what conditions, if any, PCDFs and related compounds are found in utility equipment. Sccoe of Work: Better analytical separation between 2,3, PCCr and ether less toxic congeners will be made by syr.tr.es :o:n ar.d spiking with added 13c 2,3,',9,X congeners. PCDF content c utility equipment under various operating and contamination scenarios will be determined. Laboratory arcing ar.d corona tests will be made on askarel ar.d contaminated samples. Expected Results: Improved analytical protocol for separating 2,3,7,8,X PCDF and other congeners. Determination of PCDF content of a range of utility equipment. Status: 13c *Pi*ed samples for analysis will be ready around July 1, 1984. A sampling plan for utility equipment has been prepared. Sample collection will start shortly. Completion Date: March 1985 HONS 017102 Title: Contract: Contractor Project Manager: State-of-the-Arr Review: PCCCs and gcCFs in Utility PC3 Fluid RP1263-11 SCS Engineers Ralph Koj.ai Purpose: To examine and sutr.marite the state of the art of knowledge regarding ?C"3, PCTFs ar.d other cyclic hydrocarbons ir. utility PC3 fluids. The study also is to determine what conclusions could be trade about formation conditicns ar.d to define gaps in experimental data. Scene of Work: See Purpose Expected Results: Compilation and critical analysis of information available in the literature. Status: Complete (CC 3308) Completion Date: December 1983 HONS 017X03 i Case Studies of ?C3 Transformer r i:*s: " " San Francisco, California, and Binghamton, * Sew York Contract: Contractor: RP1263-20 IT Ccrporation/W. Corey Trench Project Manager: Ralph Kcitai Purpose: (1) to document remedial ciear.-up techniques which have been successfully and unsuccessfully applied to asxara. fire contar.ination, (2) to generally document the technical and decisio related factors arising from an askarel fire, and (3) to define areas of future research relating to decontamination. Scene of Work: The project may be summarized as collection a review of written documentation, development of an interview- questionnaire, interview of key personnel involved with clean issues at the two fires, data organization and review, and report preparation. Expected Results: Summary case studies will be published wr.i provide utilities with information on what clean-up technique have been successfully used, the levels of allowable residual which were established for the respective buildings, hew they were arrived at, if understood, the basis of the incident, ar. operational changes which have been instituted to reduce the impact of future incidents. Status: Contract has been sent for execution. Completion Date: October 1984. HONS 017104 ; Title Design Manual for Reduction o: Airoo PC9 Contamination Contract: . RP1263-21 Contractor: West and 3ansen/fEI/Fenner Project Manager: Ralph Komai Purpose: This project will develop a manual to ass .s: utility personnel in identifying contamination causes and p :tnwa/s, selection of risx reduction techniques, and prepara non c: isclatic.n designs. Scope of Work: The development of this manual will proceed through a study of historical airberne contaminatio: the pathways cf oentamination, establishment of isolation oriter.ia ccr.3!scant with high voltage operation, design techniques for <iirborr.e isolation, cost factors and cost estimates, sensing systems, and the actual preparation and review of the manual. Expected Results: A manual will be made available :o the utility industry to aid them in planning risx reduo :ion engineerm measures for PCS transformers. Status: Contract has been sent for execution. Completion Date: October 1984. HONS 017105 s Title: Exposure Assessment Methodologies for PCB Spills and Fires Contract: Contractor: RP1826-13 Anderson Nichols West Project Manager: Abe Silvers Purpose: To develop methodologies to estimate exposure levels in soil, air and water of PCB from PCB spills and to estimate exposure levels of fire products from utility transformers containing PCB involved in fires. Scope of Work: A fate model will be developed to estimated exposure levels of PCBs from PCB transformer spills. A corr.partmer.ted model will be developed to estimate PC3 fire products exposure levei3. Expected Results: Code and documentation to estimate exposure levels. Status: In progress. Completion Date: December 1984 for draft of final report. 9 HONS 017106 A Ais* Assess*en; *odel for prs ;,> irz ; Aroducts in Electrical Etuiorert Contract: a?ij'5-:s Contractor: Aesource Alannins Corporation 'reject Manaoer: Abe Silvers BuTdst: 'o estimate tr.e probability of a utility trars'or-er involved In fir* in correrciai, residential and juslij str.ct..'ss. St::e o* o-ic In'orsaticn 111 be developed 'rcn surveys to ss:i-a:e probability :J a transfor-er bein; involved in a fir*. A final resort it.n inftr-atibn on tne ri< o' a Status: In progress. Cfolet'or Tate: Cececber 1934 for draft of final resort. HONS 017107 !0 Title: Teciai.cn traxewcrkj 5cr Indus"--. Level ?C3 Decisions Project h'o. ?.?I595 Contractor: Project Manager: Iecisi.cn Teens, Inc.: 3-sin 3cyd and David Cohan are the principal investigators. * C. Victor Niemeyer Puroove: To develop and test tools that utilities, industry grouts, and other interested parties can use to cam maitnts about ?C3 ris.< management issues. Sccce of Wore An analytical framework is being developed that can trace the health and environ.oer.tal effects steooirg free, alternative ?C3 risk management policies. The framework, irple m the term o: a computer medal, li.-.lts equipment management policies with t.te .mummer of accidental releaaes, the --under c: people m different categories i such as office workers, ut.l.ty personnel, emergency workers) exposed to ?C3s and ccr.tsmmints (such as TCDFs, TCDDs), to estimate the expected number of health effects. It also estimates the costs cf the policy site so that the extra health benefits can be compared tc the extra costs. Parameters that are uncertain can be treated in either sensitivity analysis or with a full probabilistic treatment. Expected P.esults: A complete, documented and tested computer model that can be used in analysis of risk management aiternsti for electric utility PCB-containing equipment. Status; The computer code implementing the framework has been implemented in prototype form. Completion Date: The model will be available in final form later in 1984. The procotype of the model is available .new wit the assistance of the developing contractor. // HONS 017108 Health Effects Assessment cf FCBs and Contaminants Contract: RP 2374 Contractor Clement Associates, Arlington, VA Project Manager: w. Wevie.n Purpose: A comprehensive assessment of the potential health ri3ic associated with occupational exposure to ?C3s and ccnta.ninar.es, based on available information. Sccce cf Work: There are three parts to the study. Part era deals with extracting from the literature data c.n human and animals on health effects associated with exposures. Part two involves characterisation of the utility worlt environment. Part three involves extrapolation of the health ris'xs for persons exposed to ?C5s ar.d contaminants, for selected endpoints. Expected Results: See objective. Status: Part 1 has been completed. Part 2 is delayed but will be completed in about one month. Part 3 depends on information collected in 1 and 2 has not started. Completion Data: Midyear 1985 HONS 012109 TI=lei Contracts: Arc Spark 3y-prcdurts c: CjC^ 351499-4,-5 Contractors: Sprir.gborn Labs, Westinghouse Project Manager: Cil Addis Pursose: To determine arc by-products o percnloroethyiene filled transferrers. Sense of Work: Equipment, procedures, and associated rationale nave been developed a*, west ir.ghcuse for the study of arced prod of C2Clu free t!-.e point of view of toxicity hazards. Spri.-.gbcrn will make use of the equipment and techniques to arc and to analyze the arc products from various combinations of CjC'-i, electrodes and 3ciid insulation material. Measurement of arc energy is provided. Expected P.esults: Levels of potentially toxic arc products generated under varying operating conditions and using a number of materials combinations will be determined. Among the arc products are CI2, HC1, and CCC12Other products, such as carbch tetrachloride, hexaohloroetha.-.e, carbon monoxide and carbon dioxide, will also be included in the analyses. Status: Arcing trials are being run on a regular basis and analyses proceeding accordingly. Most results will be ready by late June 1984. Completion Date: August 1984 HONS 017110 O Hilt: Partial and Corslets Ccrbustipn of Trans'cr-ar , Materials - Contract: . Contractor: RP2023-11 Westlnghouse Project Manaser: Si 1 Add's ' Purpose: To examine the products formed from either corslet or partial combustion of Hpuid dielectrics other than PC3. Scope s* WfX: The contractor wl 11 Initially concoct a theoretital dynamic evaluation of combustion reactions utllltlng available data. ~-iu calculations will tnen be experimentally verified at selected <ey points srp thus provide a continuum of information applicable to all situations. The Investigation will include both theoretical and experimental effects at varies combustion temperatures end varied oxygen levels. The ecrtustion reactions of Insulating materials will be performed In the charter o' a trer-o-grivi-etr'o cell, which will allow monitoring of weight loss, rate of rise of temos-itur, reaction conditions, and epygen content. The combustion products will te tracoe: for subsequent analyses or directed Into a detection device fpr on-stream ara'.,ss. The leboretory studies will use whatever sophisticated analytical technidues a' appropriate (gas chromatography, mass spectroscopy. Infrared speetroscooy, etc.), to determine qualitatively end quantitatively each type of significant corpustipn product. > HONS 017111 jtrgcre- ^asults: Irorsved kncwle-:;* sf :.-e cerc-si mi scrj'.ec* c products of lfculd dielectrics. Stetus: Stsrtlng C:~jlet1:n Sate: February 1925 HONS 017112 n Title: Csntriet: Contractor: State-af-the-Art Taxiac!351 :al review of ~-l -~i: ;.:25 S22223-12 (pending) SCS engineers Project Manager: Si: Addis Purpose; To criti tally review the literature on toxieplogy of PCS substitutes This in'orr.at:an s moil ta avoid, sere tire in t.-.e -"uture, a situation ;i-i t3 trig cur-si: 3C: H a * 1 * Score o* Jo'<: aeratur* search an toxicology of ?C3 substitutes inclucin; pyrolysis an: ::~ ustian ercducts. .<:e:tet substitutes. : A critical review of the toxicology literature a* ?C3 5* * Jj ; C3~t*5Ct being prepared. Ca-t'et'cr Cate: December 1984 HONS 017113 a CAPACITOR FIR IN A PAPER MILL AT IMATRA, FINLAND O.A.Elo and P. Vuojolahti I. INTRODUCTION 1. Early stage* After two consecutive and strong explosions in the electricity room, the control room of the pulp mill storage department sounded an alarm at 12:10 AM on 3 August 1982. The control room became filled with smoke, which caused the shift supervisor, the storage department attendant,, and their assistants to leave the room. ' When the fire brigade arrived, the formation of smoke was abundant. After some thrity minutes of investigation, the explosion was found to have taken place in the capacitor bank situated behind the "10 kv contactor station. The damaged capacitor caused all three board machines and paper machine No. 8 to stop almost immediately. 2. Persons immediately exposed The capacitors contain a liquid known as polychlorinated biphenyls (PCB). FBC is used both as an insulating and a cooling agent. gassify and to decompose. One substance given off was hydrochloric acid, which causes acute irritation. Immediately after the explosion, it was not known that the damaged parts contained PCB. Thus respirators were not used. Only after some IS people had symptoms of irritation did it become obvious that toxic gases were spreading as a consequence of the explosion. The most common symptoms reported by this group were headache, smarting and redness of the eyes, nausea. Irritation of the throat and respiratory tract, and itching and smarting of the HONS 017114 2 kin. Due to these symptoms, ten people were taken to the out patient department of the Imatra District Hospital some two hours after the explosion. The symptoms resolved slowly in fresh air. About five hours after the explosion, the ten most heavily exposed were interviewed at the company health service center in Vuoksenniska. Blood and urine samples were taken in the laboratory the next morning. Within 24 to 48 hours all ten reported that the symptosis had slowly disappeared. As Investigations proved that not only PCB but also dibenzofurans had been given off by the explosion, it was decided that the exposed persons should be carefully examined and followed. The test program arranged for this purpose is based on the findings of both foreign and Finnish specialists; also considered were the .views of the physicians who participated in the work. 3. Persons exposed later Since at the initial stages there was no knowledge of the presence or the effects of PCB or its disinitegration products, many members of the mill staff moved about in the polluted areas without the proper protective equipment or clothing. Some of the people who thus became exposed had come to restart the machines and operations. Also, despite the clear smell of gas in the area, people moved about and inspected the affected area without protective clothing. Some of the exposed were merely people who out of curiosity wanted to get as close to the explosion area as possible even though posters prohibiting entry had been erected at an early stage. The exposed also included one person who was there to take samples for indus trial hygienic measurements and others who were cleaning up. Al together some ISO people were thus slighly exposed. All of them have been interviewed and records made of the places and tiswa at which they were in the areas. In addition, blood samples have been coiy^^sd from all of them for later determination of the levels of ne. Similar tests will be carried out simultaneously with a control group of about 200 people in order to determine the present level of PCB in the blood of a normal Finnish population. HONS 017115 3 4. Examination and treatment of the exposed In this study those immediately and more heavily exposed form group A, and those later and less heavily exposed comprise group B. Group A was initially examined at intervals of a few days and later at about Monthly intervals. Group B was interviewed and examined during the early stages. No clinical follow-up tests have been carried out for group B. In future, tests to determine the levels of PCB in the blood will be carried out: decisions about further measures depend on their results. There is no treatment recommended for this sort of exposure. Animal tests have indicated that large doses of carbon tablets during the acute stages could help clear the substances from the body. Thus every exposed person was given 50 g of carbon daily for an average of one week. 5. what is thought to have happened Harmonics, the exceptional connection positions of the network, end possibly also the hot weather - which had lasted for a long period blew a number of fuses in the capacitors, the imbalance relay was not operational, and the over-current relay, due to its high setting, does not yet operate in a case like this: therefore one of the units 'exploded". The explosion caused the impedances of the network to move closer to resonance condition, and the harmonics caused the voltage to increase. This situation led to the destruction of other units when the cap'acltors went into resonance condition. A total of eleven units were destroyed by pressure, two units were destroyed by external electric arc, and two more units were destroyed mechanically. It is estimated that altogether about 100 litres of PCB leaked out from the destroyed units. The over-current relays .tripped the capacitors from the network two seconds after the electric arc. HONS 017116 4 6. Initiation of investigations Soma three to four hours after the explosion, the Labor Protection Manager was asked to come to the mill. He called an industrial hygienic expert from the Lappeenranta District Occupational Health Center to test the concentrations of gas. MD P. Vuojolahti was alerted at about 4:00 AM to organize the required medical measures. At 7100 AM a conference was held at the XaukopSA mill in which he explained the toxicity of PCB and the possibility of the presence of the even more dangerous pyrosynthetic derivatives. He decided ' to call in specialists to carry out investigations, first from the industrial Insurance Company and a representative of Oy Nokia ^b. The first samples of soot dust were flown to the Department of Chemistry of the University of Helsinki in the after noon of 3 August 1982. That same evening we were informed that the samples analyzed.contained furan. Thus,' that day the Industrial Insurance Company immediately contacted specialist institutions in Europe to collate the information available from reports on similar accidents or investigations on PCB derivatives, products. On 4 August 1982 Prof. Lindberg from the University of Helsinki arrived at the scene of the accident. At his instructions, the systematic collection of floor samples was started in the mill departments thought to have been polluted. On 5 August Dr. Reggiani from Hoffman La Roche in Basel, Switzerland arrived at the mill. He was known to have extensive experience in the investigation of PCB deriva tiyes and .the clearing up of polluted areas. The Institute of Occupational Health participated in the investiga tions during the initial stages by measuring the concentrations of PCB in the air and, later, in dust samples. The University of Helsinki concentrated on the collection of samples of PCB disintegraderlvatives and the analysis of the results of soot dust tests. HONS 017117 5 7. Effect* of the explosion The explosion took place in the board mill capacitor bank behind the 10 kV contactor station. The pressure wave, accompanied by clouds of soot and gas. went mostly through an open doorway into the 10 kV contactor station and further into the adjacent 3 kV contactor station. The filter and control rooms of the storage department are located directly above capacitor room and the neighboring room. The cloud of soot and gas entered these rooms through openings in the floor; the openings include three uncovered cable holes, one of which led underneath the floor level of the instrument cross-connection room situated under the control room. The pollued area at the floor level of the storage department was limited to the surroundings of the control room and the end of the floor. . Some traces of PCB were also discovered on the story below the floor level of the storage department; these were, however, considerably lower than the. concentrations detected at the floor level. The gases and soot were not able to penetrate into the board and paper machine halls. Once the eleietric power had been cut off, it was no longer possible to run the three board machines and two paper machines, the outer pulper station, and the Kamyr pick-up machine. After the pulp storage tanks had been filled, the pulp production line also had to be stopped. HONS 017118  8. Cleaning and other follow-up measures 8.1. General Depending on the concentration of PCB, the cleaners used different kinds of protective clothing: VSS suits and compressed air respira tors; disposable protective clothing with filter mask; or disposable clothing that included gloves and head protection. The vacuum cleaners used were all fitted with microfilters. Once collected all the waste was packed in metal containers or specially made plywood cases. 8.2. Control room and cross-connection rooms The control room and cross-connection rooms were vacuum cleaned several times with microfilter-equipped vacuum cleaners and wiped several times with methylene chloride and methyl alcohol. 8.3. Switching station and adjacent rooms The switching station and the adjacent rooms were vacuum cleaned as above, wiped with methyl alcohol, and high-pressure washed with warm water. The walls and the ceiling were painted twice over with epoxide pitch,and the floors received a coating of epoxide. 8.4. Room of the explosion The floor, an area of approx. 30 n', was pick dressed and the plaster was removed from the walls. .It was vacuum cleaned several times (as in 8.1.) and wiped several times with methyl alcohol. The floor and walls were ground. The floor underwent the following treatment; moisture Insulation, new concrete, and epoxide paint. The walls and ceiling received double coatings of epoxide pitch. HONS 017119 7 8.5. Mill area The storage department was vacuum cleaned and washed with water two of three times. The Xa 4 beater department was vacuum cleaned, washed with water, and partly wiped with methyl alcohol. The room adjacent to the control room was vacuum cleaned and washed with water several times. The surroundings of the bobbin machine were vacuum cleaned and partly wet-wiped. 8.6. Electrical and other appliances in the various rooms These were vacuum cleaned with microfilter-equipped vacuum cleaners several times and wet-wiped with methylene chloride and methyl alcohol several times. 8.7. Special equipment and supplies for these operations The following list of items was used: 1,300 respirators 4,900 overalls of non-woven material 1,200 protective hoods 4,100 shoe covers 2,800 disposable gloves 40 ordinary overalls 45 pairs of rubber boots 35 raincoats and hats 600 glass jars 120 kg of seasUng foam 4,000 kg of ABSOL absorption agent 93 VSS suits, in circulation'(350 washes) compressed air i I 1 > 200 BAR: 1,087 units at 3.8 - 12.8 1,195 units at 13.8 - 18.1 ' c. 2,000 litres of paint . HONS 017120 8 8.8. Cleaning waste All cleaning waste, a total of about 100 tons, was first packed in closed casks or boxes which were stored in the vicinity of the nill, either in bundles or in containers. The waste materials were later sent to be destroyed in England. 9. Results of the measurements at Kaukop88 Mills 9.1. Concentrations of PCB measured in air samples The first measurements for the levels of PCB in the air were carried out about five and a half hours after the explosion,at a place close to pillar B 2 in the storage department. An open cable channel leads to this site from the condenser room. The concentration of PCB in the air at this site was 16,000 pg/m* . A similar test was done in the capacitor room itself; the result was 8,000 pg/m'. (This latter figure can hardly be a correct reading.) A sample was taken in the control room of the storage department only at 4;00 PM on the day of the explosion; the finding was 77 pg/m1. By this time the level of PCB in the air close to pillar B 2 had gone down to 80 pg/m'. Further tests were carried out in the capacitor room only on 12 August.1982, and the reading was still high, 1,600 pg/m*. The electricity department was later fitted with special ventilation, for which suction fans equipped with activated carbon filters were used. By 6 August`1982 the level of PCB was below the critical value of 10 pg/m' in areas other than the electric power areas (i.e., the capacitor room, the 10 kv and 3 kv contactor stations, and at point B 4 on the second floor). By 18 August 1982 the concentration in the 3 kv contactor station had dropped below the critical level. On 29 September 1982 the concentrations-recorded in the capacitor room and the 10 kV contactor station-still exceeded the critical level. Only on 12 November 1982 was the concentration in the capacitor room below the critical level; at that time PCB levels of 4.9-6 pg/m* were measured. HONS 017121 9 Mechanical ventilation had been arranged for the electric power areas by the use of temporary pipe constructions. The ventilation systesi was fitted with activated carbon filters; these reduced the level of PCM by 80 - 90 I. 9.2. Determination of PCB and pyrosynthetie products on surface samples At the time of the accident it was not possible to measure the con centrations of PCB on dust samples in Finland. The Department of Chemistry of the University of Helsinki carried out tests to deter mine the contents of furan from 3 August 1982 onwards. Samples taken from the 10 kv contactor station on the day of the accident con tained 10,000 ng/m* of furan. On 4 August 1982 samples were collected from the storage department; the lowest findings were about 10 ng/m', the highest 1,800 - 3,600 ng/m'. In the electric power area, in the 3 kV contactor station, the levels detected were SOB - 630 ng/s^ . Tests were carried out in the capacitor room on 11 August, when the concentration oaths floor varied from 11,600 to 25,900 ng/m* and on the walls from 209 to 576 ng/m' . - Dust samples were also sent for analysis to Sweden, to Prof. Rabbe of the Department of Organic Chemistry of the University of Unei. These analyses showed concentrations of furans at varying degrees of chlorination, biphenyls and concentrations of tetra- and pentachlordibenzodioxins about tenfold less than the levels of furans. . - ~ After the explosion at KaukopiK, the Institute of Occupational Bealth in Helsinki began to develop methods to determine the levels of PCB and its pyrosynthetics in dust and soot samples. However, HONS 017122 10 auch analyses were done on duat samples taken on 24 August only, by which tine e fair amount of cleaning had already taken place. Analysis of soot samples taken in the capacitor room after the explosion showed that some dibenzofurans at various degrees of chlorination had been created in the explosion. Sample No. 1 con tained about 20 pg/g of tetrachloride benzofurans, and isomers 2, 3, 7, 8 2 pg/g. ' .Soot sample No. 2 Con tained about 90 ug/g tetrachloride benzofurans, and 6,5 pg/g - Isomers 2,3,7,8.. Traces of polychlorinated pyrenes and polychlorinated biphenyls were detected in the soot samples. No dioxin was found in the samples; in this respect the results of the analyses carried out by the Institute of Occupational Health differed from those reported by Prof. Rabbe. 9.3. Concentrations in mill products The concentrations of PCB in mill products were also measured. The tests proved that the content of PCB in paperboards did not exceed the criticsl value. The levels of PCB in clean washing water and used washing water were also recorded.II. II. EXAMINATIONS OF PERSONS EXPOSED TO PCB AND ITS PYROSYHTHETICS 1. Composition of the group and data on their symptoms This report' deals with the examinations of the people who cosiprise group A, i.e., those with the heaviest exposure. The group contains the storage department attendants and supervisors on duty st the time of the accident, members of the fire brigade, and labor pro tection personnel,_a total of 16 persons thought on the basis of HONS 017123 11 of the. interviews to have been the most heavily exposed. Th# group contains only men. Of the acute symptoms, 63 t had headache, 63 t experienced smarting of the eyes, and 50 I reported Irritation of the throat and res piratory tract. Nausea, dizziness, increased sweating, and irrita tion of the skin were markedly less common. The acute symptoms of all the exposed persons disappeared within 24 to 4S hours. Table 4 presents the results of seven months of follow-up with regard to diagnosed dlkeaaes or changes In health status among the exposed group. The more unusual conditions observed include ridges in the nails of two persons and reduced immunity to infections among two persons. The acne of an exposed person worsened visibly and required a dermatologist's treatment. Nine had.lesser skin symptoms such as flaking of .the face skin,, redness, itching, and pimples. One of them also had increased pigmentation of the neck skin. Immediately after the explosion one person had an allergic-type, itchy rash on the upper part of his body that lasted for five days and disappeared by itself; this does not fit the picture of chlorine- induced acne. There was not one case of chlorine-induced acne. The association of these mild skin symptoms with PCB is questionable. The exposed were requested to check for any possible skin changes, and the reporting of milder changes can merely be the result of close, active observation. ' HOMS 01?lj>4 12 2. Measurements of PCB levels 2.1. Blood and its subtractions Some of the blood samples taken on the day of the explosion were sent for analysis to Sweden, to Prof. Rabbe of the Department of Organic Chemistry of the University of Umel. On IS august BB2 we were informed by telephone that no PCDF could be detected or that the concentrations were below the detection limit of 0.15 ppt 0.15 pg/g. Ho dioxins were found, either. 2.2. adipose tissue Biopsies of adipose tissue have been taken from some of the most heavily exposed. These samples also awsit analysis by the Institute of Occupational Health. 2.3. Urine Urine samples have been collected from the group with the heaviest exposure. The semples ere now being stored at the Institute of Occupational Health in the event of future analyses. 3. Measurements to reveal effects 3.1. Hematologic measurements ~ Samples have been taken to analyse the levels of enzyme activities in the red blood cells. These determinations have not yet been done. HONS 017125 13 Typical blood teata (for La, Hb, Hkr, MCHC, and leuk) and differential cunts of the leukocytes were followed for a period of aix montha. Patient No. 3 hat continually had elevated La levela that range between 26 and 37; the reaaon ia rheumatic arthritia. Patient No. 6 has alto had moderately elevated La levels (between 12 and 22), with no clear reaaon for the elevation. Other have had mild fluctuationa above the La level considered normal (i.e., 10), but these increases do not appear to be a result of exposure. For the differential cunts of the leukocytes, an unusual finding was that the proportion of lymphocytes exceeded 40 t in 43 of the 8< analyzed samples, i.e., in half the samples. Fifteen of the sixteen studied had at least'one elevated level of lymphocytes. A similar finding of elevated lymphocyte levels has been reported in Japan for patients with Yusho disease. No clear anemic trend was observed. A slightly increased percentual distribution of eoslnophiles was found in 18 of the 86 samples; these increases are probably not connected with the exposure. The blood samples were analyzed at the company health service center in Vuoksenniska. 3.2. Biochemical analyses Liver functions were followed by analyses of enzymes (Gamma-Gt, AFOS, ALAT, and ASAT) over a period of six months. The determina tions were made by the Institute of Occupational Health in Helsinki. The reference values were; 40 U/l for Gamma-Gt; 35 U/l for S-ASAT and S-ALAT; and 60 - 270 U/l for AFOS. The following values in excess of the reference values were found: Gamma-GT, 35 of 95 values; ASAT, 7 of 95 values; ALAT, 22 of 95 values; and AFOS, 10 of 95 values. Pathological enzyme levels were detected at least once for twelve of tha exposed, and four persona had completely nonul enzyme levels at each examination. Figures 1-4 show the changes in enzymes activities for the individuals as a function of time. HONS 017126 14 As wen in the figure, about two weeks after the explosion, there was a slighlty rising trend in the activities of Gamna-Gt, ASAT, and ALAT enzymes at group level. N6 similar trend was detected for the entire group at later stages; rather, variations probably had other causes. In some cases alcohol may have been a factor con founding the initial group trend of increased liver enzyme activities. Art ASAT/ALAT ratio greaterthan 1 is highly indicative of increased liver enzyme activity'caused by alcohol. Of the eleven in the group for whom saa^les were analyzed for enzyme activities after two weeks of follow-up, seven had an ASAT/ALAT ratio under L and four had a ratio over 1. The latter had consumedslight amounts of alcohol a couple of days before the samples werecollected; the liver enzyme activities of two of them did-not exceed the normal limits. Thus the slight increase in liver enzyme activities detected at group level dan at least in part be interpreted as an effect of exposure to PCS and its pyrosynthetics. No essential change was noted in the levels of AFOS. Serum proteins and their electrophoreisis were also followed. The reference values for proteins were 65 - SO g/1. The determina tions were done by United (Khtyneet) Laboratories. Protein levels above the reference values were found in 20 of 98 samples. The increases were only slight and most likely due to slight dehydra tion caused by fasting. Ten of 98 eleetrophoreisic values deviated from the reference values. No specific changes were dis covered. Hypoalbuminevia was not detected at all. Thus the levels of exposure in question did not bring about noticeable effects in serum proteins. HOMS 017127 15 To follow the fatty"metabolism the levels of cholesterol, trigylcerides, and the electrophoreisls of lipcproteina were measured frem the sera. X clear Increasing tendency was noted for triglicerldes in the serum. For many of the exposed, the peat was reached a week after exposure; the peak was reached two weeks after exposure for the remainder. Four weeks after exposure the baseline values were attained again. The baseline value for the ten exposed persons frost whom all four samples were collected was 1.39 mmol/1. For this group the swan peak level of triglycerides was 2.14 nssol/l; their mean level at four weeks was 1.13 mmol/1. At group level a 54 increase was found (Figure 5). A clear decreasing tendency in the level of cholesterol was observed for the same group. The lowest levels were attained, again, at either one or two weeks after exposure. The mean baseline value was 6.2S mmol/1, the mean of the lowest values was 4.62 mstol/1; and four weeks after exposure it was 5.69 mmol/1. Compared with its baseline value, a mean drop of 26 t occurred in the level of cholesterol (Figure 6). The electrophoreisls of the lipoproteins was followed. Electrophoreisic measurements of the samples collected the first few days after the explosion could not be done because the samples had been stored in deep freeze. The first electrophoreisic analyses were carried out for the samples taken two weeks after tha explosion, at a time when the peak values for trigylcerides had already passed. In electrophoreisls triglicerldes are transported in pre-^-fractions. HONS 017126 16 Increased levels of pre-j/i-fractions were detected as follows; 2 weeks 4-5 weeks 12 weeks 3/12 ' 1/15 3/16 18-19 weeks 4/1$ 26 weeks 5/16 Otherwise the electrophoreisis of the lipoproteins was normal. The results for these sanples do not allow any conclusions to be drawn. Electrophorelsic analyses at earlier stages after exposure would probably have provided more information. at least sofce of the changes detected in the fatty metabolism were an effect of the exposure. Alcohol may be a confounding factor that accounts for some of the changes. 3.3 Cytogenic determinations It is known that chlorinated dibenzodioxins and dibenzofurans are mutagenic. Therefore chromosomal aberrations and sister chromatid exchanges were studied in the lymphocytes of ten persons in the group with the. heaviest exposure. A reference group was composed that contained a group of workers with similar smoking habits from the Tainionkoeki factories. Chromosomal aberrations reflect damage that can lead to the rise of somatic chromosomal mutations. The analysis of sister chroewtid exchanges reveals changes in the organization of the chromosomes within the DNA strands. Many substances known to be mutagenic cause increased frequencies of sister chromatid exchanges in cell cultures. The results should be examined at the group level only, as the meaning of intraindi vidual variations is not yet clear. Compared with the reference group, the exposed did not have statis tically significantly more chromosomal aberrations. Smokers in HONS 017129 these groups did not have more chromosomal aberrations than non smokers. The frequencies of sister chromatid exchanges among the exposed were not higher than the frequencies among the referents. The frequencies were somewhat higher among the smokers, but the difference from the nonsmokers was not statistically significant. The findings indicate that the exposure to mutagenic agents was not heavy enough to induce an increase in chromosomal damage. Appendix 8 presents the data in detail. 3.4. Immunologic tests 3.4.1. Immunoglobins and compleawnts The levels of immunoelectroplioreisls and complements |C4, Cj, and CI-INH (Hane method fI were followed in the exposed. The determinations were made at the Clinical lanunologlcal laboratories. The samples were taken one month (ten persons) and six Mnths (sixteen persons) after exposure. The lmmunoeledtrophoteialc values were normal except far one sample taken six months after exposure; in this sample a polyclonal increase was detected in ICM. Five of the samples contained amounts of complements that deviated from the . normal values. The deviations were usually values above the normal range. No sign of deficiencies was detected in the complements. It can thus be concluded that the exposure did not alter either the complement system or the immunoglobins. 3.4.2. Cell immunologic tests These were done at the Transplantation Laboratory of the University of Helsinki, by Ors. E. Taskinen (D. Med. Sc.) and R. Renkonen (D. Mad. Sc.). The mononuclear cells were isolated by the method of Ficoll-Isopaque (TIP). In addition to overall and differentialcell cants, the staphylococcus technique and monoclonal antibodies were used to study the following subclasses of T-lymphocytes T cells (OKT-ll)j T-heiper cells (ORT-4); and T-suppressor cells (OKT-8). Surface and/or intracytoplasmic immunoglobinpotitive tests were used to study B cells. Furthermore, the follow ing mitogens were used for stimulation tests of the lymphoeyt s: HONS 017110 18 phytohemagglutin (PDA); concanavsll A (ConA); and pokeweed mitogen (PWM). Sample! were taken from ten persons five weeks after the explosion. Since the results contained many pathologic findings, control samples were taken from six persons twelve weeks after the explosion and, again, from five persons six months after the explosion. Results The FIP isolation method yielded a range in the number of cells froai 9 x 10 to SB x 10) the means were 33.8 x 10 five weeks after the explosion and 24 x 10 twelve weeks after the explosion. The differential cell count revealed that three of the ten samples had more than 10 t large granular lymphocytes (LGL cells). In all of the samples studied (8/8), the level of T cells was low (< 35 ) five weeks after the explosion. Four of the samples taken seven weeks later had normal T cell levels. The ratio of T-helper cells to T-suppressor cells was low (< 1.4) in six of seven samples taken five weeks after the explosion. The ratio was still low in two of the samples taken six months after the explosion. No decisive conclusions could be made with respect to the percentages of 8 cells. The results of the stimulation tests varied according to the mitogen and the time of the test (see the table below). MOWS 017131 19 TABLE. Frequencies of reduced mitogenic responses of the mononuclear cells in the blood (< 66 4 of normal control values) among persons exposed to PCB. Duration after the explosion PHA ConA pm five weeks twelve weeks six months 4/10 1/6 0/3 3/10 5/6 1/5 5/9 2/4 0/5 Thus the responses that reflect the functions of T cells were reduced for nearly half of the exposed, and, except for FRA, the reduction was still observed twelve weeks after the exposure. Only one of the exposed had reduced ConA responses six months after the explosion. Summary Reduced percentages of T cells, a reduced ratio of T-helper/T-suppressor cells, and reduced responses of T cell functions in stimulation tests were found in the blood samples taken from persons exposed to PCB five weeks after exposure. Seven weeks later these changes ' had levelled off to some extent, and the percentages of T cells had become normal. Six months after exposure only two of the ten had a reduced ratio of T-helper/T-suppreasor cells.The results of the stimulation tests indicated that T cell functions had normalised. The findings suggest that some temporary damage occurred in the number and functions of T cells. 3.3. Neurophysiological examinations . The nerve conduction velocities were measured for the first time on 1 October 1982~and for the second time on 9 February 1983. The motor conduction velocities (MCV) were measured from the median, the dinar, and the peroneal nerves. The distal sensory conduction velocities IdSCV) were measured from the median and the ulnar nerves, aid the sensory conduction HONS 01713*! 20 velocities (SCV) vmre measured frcm the sural nerve In the area of the calf, lhe *! below presents the group means and standard deviations (SO). Nerve Median, MCV, m/s Median, dSCV, m/s Dinar, MCV, n/s Ulnar, dSCV, m/s Peroneal, MCV, m/s Sural, SCV, m/s Under 40 years old Mean SO 57.0 47.3 59.6 43.8 47.2 36.0 0.8 3.1 4.0 4.8 5.2 3.3 Over 40 years old Mean SO 58.2 45.8 57.7 41.0 50.9 38.2 4.8 3.2 5.5 4.8 6.1 4.0 The means were somewhat lower than the mean conduction velocities of referents. At group level significantly slower sensory conduction velocities were measured for the ulnar and sural nerves, and sig nificantly slower motor conduction velocities were measured for the peroneal nerve. Theta findings are indicative of damage to nerve axons. The measurments were repeated about six months after the explosion fox the 16 exposed. The following table shows the results of these latter measurements. Nerve Median, MCV, a/s Median, dSCV, m/s Ulnar, MCV, m/s Ulnar, dSCV, m/a Peronaal, MCV, m/s Sural, SCV, m/s Under 40 years old Mean SD 60.1 54.2 58.9 45.9 -51.0 39.3 5.3 9.4 3.4 3.0 6.1 2.6 Over 40 years old Mean SD 57. 50.0 59.7 44.3 54.5 40.1 4.9 4.9 4.8 4.2 5.8 3.3 HONS 017133 21 Th* mans for the letter measurements were higher than those for the preyious measurements and were now on the same level with the means for referents. The mean amplitudes of the.sensory nerve potentials were also higher than for the previous measurments. Furthermore, all of the conduction velocity measurements fell in the normal range 6f variation. The measurements were taken by Docent SeppSllinen from the Institute of Occupational Health, in In whose opinion the findings indicated temporary, mild neuropathy of the sensory distal nerves in particular. HONS 017134 HONS 017135 TABLE 1. Diagnosed diseases and findings during the seven-month follow-up (3 August 1982-3 March 1983) Patient's nunber Diseases, signs, symptoms t 2 3' 4 5 6 7. 8 9 10 11 12 13 14 15 16 Barber's Itch. Sone plnples on the face. Aching Joints. Acne worsened. Repeated sinus infections (treated four tines with antibiotics). One respiratory infection. Ridges in th^ nails. One respiratory infection. Gastroenteritis once. Redness in the nasal area two weeks after exposure. One respiratory infection. Ridges in the nails. Erysipelas. One respiratory infection. One respiratory infection, flaking, itching rash on the face. Itching rash on the arna two nonths after exposure. One respiratory infection. Pimples on theface,temporarily. Nonspecific urethritis, One respiratory infection. Slightly increased pigsentation of the skin on the neck. Pharyngitis once. One respiratory infection. One attack of gout. Four or five aphthae in the mouth. Dry skin. Itching, allergic-type rash on the upper body between 4 August and 9 August 1982. One respiratory infection. Cold sores twice. -. HONS 0 1 7 1 3 7 I .Figuras 1-6 J-------- I------~3-------3----S------- S' mo 1 S-ganma GT HONS 017138 TABLE 3 THE RELATION BETWEEN PCB-LEVELS AND OTHER FINDINGS PCB GAfVUGT't AFOSt 3.8. 6.8. ALATt 1 3.5 2 4,1 30 4 11.2 5 5.2 6 19.9 7 48.3 8 33.4 90 10 10.0 11 0 12 0 13 36.9 14 6.9 15 4.5 16 19.8 0 20,3 20,6 25,5 19,0 31,9 .13,4 11,2 0 7,7 0 0 7.1 0 5.0 6,3 .+ + ,+ +- 0 0 0 + 0 - -+ - .+ -++ -+ ++ +-+ 00 +- 00 00 ++ 00 -. + - +H ASATt KoLEST-t . - ' +- + +(-) + +- 0 0 0 +H 0 - + + + + + 0 + 0 0 + +- TriglyT Cell. Nerve Immun.Tesi Conduct i or \ Velocity! -0+ +-+ + .- . - + -' - -+- ++- ++ -+ +? 0 +0 00 + + - + 00+++ +0 - --- f - f t f U I O SHOW TECHNICAL PAPER The Degradation of Polychlorinated Benzenes in Electrical Equipment IEEE/PES Conference New York City February 1*5,1M2 Dr. Ian Webber Engineering Manager RTE Corporation RTE CORPORATION 1900 E. NORTH ST. WAUKESHA, W1 531M HONS 0171*0 THE DEGRADATION OF POLYCHLORINATED BENZENES IN ELECTRICAL EQUIPMENT IEEE/PES CONFERENCE, NEW YORK CITY, 1982 The currant legislation governing polychlorinated biphenyls (PCBs) was precipitated by the Yusho incident in southwestern Japan in 1961. Approximately 1300 people Ingested an average of about 2q each of PCB fluid from a leaking heat exchanger. The first contaminant of the rice oil to be identified was PCB, and the symptoms of the poisoning were therefore blamed upon PCBs as the causative agent. Subsequent analysis^) of the Yusho oil has revealed that the apparent toxicity of the aakarel is likely to have been due to the presence of PCB degradation products such as polychlorinated dibenzo furans (PCDFs). In 1970 Vos at' al^2' showed a correlation between the toxic effects of European PCBs and the concentration levels of PCDFs. The major PCDF components contained in Yusho oil were the highly toxic 2,3,7,8-tetra-CDF and 2,3,4,7,8-penta-CDF. The relative concentra tions of the PCDF isomers present in Yusho oil and in two samnles . of used heat exchanger PCBs (Kanechlor KC 400 and Mitsubishi ' Monsanto T 1248) were found to be strikingly similar.'3' This fact underscores the findings of other workers in this field that there is a connection between the toxicity of degraded PCB fluids, in cluding tri-/tetra-chlorinated benzene/PCB blends, and the con centration of PCDFs. The overall toxicity of the fluid may then be attributable, as Vos indicates, to the two particular components mentioned earlier. (1) II. R. Buser, C. Rappe and A. Gart; "Polychlorinated Dibento Furans Found in Yusho Oil and in Used Japanese PCB"; Cheansphere 7, 439 1197*). (2) J. G. Vos, J. H. Koeman, H. L. Van der Maas, M. C. Ten Hoever de Braun and R. H. de Vos; "Identification and Toxicological Evaluation of Chlorinated Dibenzofuran and Chlorinated Naphthalene in Two Commercial Polychlorinated Biphenyls"; Fd. Cosmet. Toxicol. 8, 425 (1970). (3) M. Kuratsune, Y. Masuda and J. Naqayuma; Proceedings of the National Conference on Polychlorinated Biphenyls, Chicago, November, 1973; EPA-560/6-75-004. HONS 017141 Buaer has shown that the pyrolysis reactions of chlorobenzenes in the presence of air yield tetra- to octa- CDFs and tetra- to octaCDDs. <') The experimental conditions used in his laboratory study were intended to provide easily determined concentrations of degradation products {approximately II) within a short period of time. The fluid temperature inside a faulting transformer is not likely to reach 600C for a sufficient length of time to cause the formation of PCDFs, but there may be hot-spot temperatures of 300C for significant periods. It is of concern that heating PCBe in air at 300OC for only one week is sufficient to generate PCDFs (5) j.n the ppm concentration range. Chlorobenzenes with a higher deqree of chlorination than the compounds used for pyrolysis were observed in each reaction mixture. These must have been formed by a chlorination process from components of the mixture containing less chlorine. However, the formation of chlorobenzenes with less chlorine than the compounds used for pyrolysis was not observed. This may indicate that once the chloro benzene molecule has become reactive by the removal of a chlorine atom the product formed will depend upon the concentration of inter active molecules in the system. In any case, the partially dechlorinated chlorobenzene does react, probably with oxygen, and transanites to a stable form which is not a chlorobenzene. For example, the reaction mixtures were found to '-contain ehlorophenols in addition to PCDFs and PCDDs. It was therefore suggested that the reaction of chlorophenol with chlorobenzene could lead to polychlorinated diphenyl ethers (PCDPEs) which are known to form PCDFs upon pyrolysis. This mechanism does not seem probable, however, because PCDPEs were not detected. PCDFs and PCDDs can be formed from ehlorophenols in three ways. (1) The dimerization of chlorophenates. (2) The cyclization of PCDPEs. (3) The cyclization of polychlorinated phenoxy phenols termed `pre-dioxins*. The dimerization of chlorophenates is likely to be bimolecular and therefore the products formed should be highly dependent upon the chlorophenate concentrations. The PCDD isomers fonsed in the system, and their quantities, will depend upon the relative kinetics . of the alternative reaction routes. It would be expected that the relative importance of the benoleculsr mechanism would decrease with the concentration of chlorophenates.4 5 6 (4) H. X. Buser; `Formation of PCDFs and PCDDs from the Pyrolysis of Chlorobenzenes'; Chemosphere, S, 415 (1979). (5) H. X. Buser and C. Xappe; `Formation of Polychlorinated Dibenzofurans (PCDFs) from the Pyrolysis of Individual PCS Isomers*; Chemosphere 157 (1979). (6) C. Xappe, S. Marklund,~H. X. Buser, H. -P. Bosshardt; `Formation of PCDDs and PCDFs by Burning or Heating Chlorophenoles*; Chemosphere 7, 269 (1978). HONS 017142 The pyrolysis of PCDPEs follows two competitive reaction pathways vi*., reductive dechlorination cr ring closure to dibenzofurans. >7>* 9 This is not likely to be an important route in the laboratory pyrolyses because of the apparent absence of PCDPEs. The cyclization of pre-dioxins is a molecules reaction caused by heating. The final concentration of dioxins in the heated chloro benzene system should ultimately be dependent upon the concentration of polychlorinated phenols which are formed. The presence of chlorobenzenes and chlorophenols in the product mixture which contain a higher degree of chlorination than the chlorobenzenes in the starting mixture indicates the probability of free radical reactions. Louw, Rothuizen and Wegman'*) have interpreted the pyrolysis of chlorobenzene as a radical chain reaction involving CjHrCl, -Cl and H* as carriers._ The authors discuss a radical reaction sequence which explains the observed product pattern, including the formation Of PCBs. The pyrolysis of PCBs occurs intramolecularly by four alternative reaction routes to yield different isomeric PCDF products. () The extent to which the free radical reactions yield PCBs and the extent to which the eventual cyclization of the PCBs contributes to the overall yield of PCOFs is probably small. Indeed, the quantities of PCBs detected were small. The formation of polychlorinated naphthalenes in Buser's experiments {Ref. 4) can be explained by either invoking the formation of benzyne intermediates or the re-arrangement of intermediates formed between an ortho-chloro phenyl radical with a chlorobenzene. The overall effect of radical reactions on the product distribution in a pyrolysis reaction will be effected by both temperature and the availability of oxygen. No work has so far been done to examine the effects of these parameters on the total quantity of PCOFs produced. (7) A. Nor strom, K. Andersson and C. Rappe; 'Studies on the PoraMtion of Chlorodibenzofurans by Irradiation or Pyrolysis of Chlorinated Diphenyl Ethers*; Chemosphere, 241 (1977). (I) R. Louw, J. W. Rothuizen and R. C. C. Wegman, 'Vapour Phase Chemistry of Arenes. Part II. Thermolysis of Chlorobenzene and Reactions with Aryl Radical* and Chlorine and Hydrogen Atoms at 500*; J. Cheat. Soc. Perkin Trans., 2, 1635 (1973). (9) H. R. Buser and C. Rappe; 'Formation of PCDFs from the Pyrolysis of Individual PCB Isomers*' Chemosphere, 8. 157 (1979). HONS 017143  } Norite(10) and Nagayamatm hava determined the concentrations of PCDFs in "Yusho oil" end in new European and Japanese PCB samples. The combined concentration of the different isomers identified in the unused PCB samples was from 1-24 ppm. while in Yusho oil, the concentration was reported as 2700 ppm in Ref. (10) and 5000 ppm in Ref. (11). The concentrations of PCDFs in new and used PCBs is shown in Table (1). The PCDFs in the unused PCBs were probably formed by the chlorination of dibenzofuran as an impurity in the original biphenyl. Chittim, Clegg, Safe and Hutzinger have analyzed unused North American askarels (Interteen, Pyranol and Chlorextol) and foundU2) that they contained less than 0.05 ppm. As stated earlier, Vos et al have foundthat the toxic effects of PCB based fluids could be correlated with the levels of PCDFs contained in them. It is therefore of considerable concern that, in Ref. (12), it was determined that the level of tetra-CDF in an askarel was found to increase as the time since the transformer was Installed increases, see rig.(1). The small number of samples tested were insufficient to obtain a definite, correlation between kVA and oil volume, manufacturer and fluid type or TCDP concentration and transformer loading. Nhile the results indicated that transformer load is probably a major factor in the formation of PCDFs, the effects of discharging or arcing appeared to be negligible. Chittim et al have concluded that the PCDFs found in the oil were probably produced during the normal operation of the transformer and that a number of different variables all contribute to the foneatlon of PCDFs. In particular, it was observed that the samples which contained the most polychlorinated benzenes also had the hiaher TCDF concentrations. Thus, the presence of certain polychlorinated benzenes may be the major contributor to the formation of PCDFs in transfonaer askarels. The results of different studies on the formation of PCDFs in chlorobentene/PCB fluids are shown in Table II to IV. A list of the reactions involved in the formation of toxic products is given in Table V.10 11 12 (10) (11) (12) M. Merits, J. Nakegawa, K. Akiyama, S. Mimurz and N. Ieonoi 'Detailed Examination of PCDFs in PCB Preparations and Kanemi Yusho Oil"; Bull. Environ. Contam. Toxicol.; 18, 67 (1877). J. Nagayama, M. Kuratsune and Y. Masusda; 'Determination of PCDFs in Kanechlors and Yusho Oil"; Bull. Environ. Contam. Toxicol. 15, 9 (1976). B. C. Chittim, B. S. Clegg, S. H. Safe and O. Hutzinger; 'Chlorinated Dlbenzofurans and Dibenzo-p-Dioxins: Detection and Quantitation in Electrical Equipment and their Fonaation During the Incineration of PCBs.*; Report prepared for Fisheries and Environment Canada, under Contract No. 05578 00067; September 1979. HONS 017144 TABLE : PCB NEM: Aroclora 1254, 1260*11* Clophan A760O-9) Phanoclor DP-6(13) T 1242<*) T 1248 T 1254<> T 1260 O) Kanachlor 300(** Kanachlor 400^9* Kanachlor 500(*) Kanachlor 600l9* USED: T 1248<9> Aroclor 1254(ll) Aroclor 1260(H) Yusho Oil PCB (KC400) - ' Polychloro Dibanzo Purans Total Cl4 5 Cl6 1.4 0.7 2.3 0.5 0.1 0.8 6.7 12.2 1.7 0.2 ` 5.0 10.0 2.2 2.3 3.6 0^9 1.6 10.4 1.1 0.5 2.2 2.9 - 1.9 O.S ' '0.9 3.1 0.4 ' <0.05 8.4 13.6 4.5 2.8 5.6 2.2 8.3 23.8 6.1 1.1 S. 8 5.6 0.7 12.4 0.55 1.1 520 1330 810 2680 (11) G. M. Bovaa, H. J. Nulvihill, 8. R. T. Siaonait, Natura, 256, 305 (1875). A. L. Burlinqaaa and R. M. Rlsabrouqh, S M M O SNOW Chlorobanssita Studiad Mixtura of tri-/tetrachlorobanzene: pentachlorobanzans Honochlorobenzsns Monochlorobanzsna w TABLE TT Chlorobenzene Pyrolysis Process Pyrolysis in air ProdMct(s) PCDO, PCDP, PCB, higher chlorinated benzenes. Cl phenols Yield 2,000 ppai Pyrolysis in 2 HOno- and di-Cl biphenyl Pyrolysis di-Cl biphenyl 50 ppm 10 Ref 4 8 14 lHONS 0 1 1 *4 * (14) C. P. Cullis and J. E. Hanton; Trans. Faraday Soc. 54. 381 (1958) Chlorophanols Studiaa Various tri-*, tetra- and pantachlorophanola Trichlorophanol Tecrachlorophanol 2,4,5-Trichlorophanol 2,4,(-Trichlorophanol 2,3,4,6-Tatrachlorophenol TABLE III Chlorophenol Pyrolysis Process Product(s) Pyrolysis Various PCDOs Pilot scale coabustion on mod chips Pyrolysis Mainly TCDDs -* Mainly HCDDs 2,3,7,8-TCDO 2,4,7,9-TCDD 1,3,4,(,8,9-HCDD Yield ' 3-10% Ref IS 230 ppa 16 350 ppn 1% up to 151 30% 17 {.*rl T O SNOW (15) UC> (17) H. R. IttMti * SpectroMtry*; SjP*ehtl0n. ,nd J* CnroMt09r 114* 95 of PCODs by Gaa Chroautoaraphy - Maas (1975)# I. Jansson, C. Sundstron and # AKling; 3ci. Total Environ# 10, 209 (19701, H. C. Linger, T. P. Brad/ and P. R, Brlqqsr Environ# Haalth Firspaot. 5, 3 (1973). Chlorobiphonyl Studied 18 different chlorobiphenyla containing 4,5,6,7 and 8 Cl atoms 2, 21, 6, 61 - TCB 2, 21, 4, 4l. 5, 51 - HCB 2, 21, 4. 41, 6, 61 - HCB Jtrochlors 1248 and 1260 Archlor 1248 2 C 2 Ln o o*> Process Pyrolysis in air Pyrolysis in air Product(a) PCDP , PCDP ^2,3,7,8-TCDP Pyrolyiii in air in O2 in M2 PCDP Yield 0.1-1% Ref 5 2t 18 2t 300 ppm 1100 ppm 2 ppm 19 (It) H. I. Buser, H. P. Bosshardt, C. Kappa and R. Lindahl; Chamoaphare, 7, 419 (1978). (19) N. Morita, J. Nakagava and C. Kappa; Bull. Environ. Contarn. Toxicol.; 19, 665 (1977). TABLE V I (1) Formation of chlorophenols: Cl, OH CU (2) Formation of polychlorinated dipehnyl ethart/PCDF*: oH O. Clj ----------------- (3) Dimerization of ehlorophenataa: PCDF CL HONS 017149 TABLE V (cont'd) HONS 017150 / HONS 01?151 I nyears Service RECOMMENDATIONS FOR THE PREVENTION OF PCDF CONTAMINATION FROM FIRES INVOLVING ELECTRICAL EQUIPMENT by Conaerelal Chenicala Branch Environmental Protection Programs Dlractorata Environmental Frotaction Sarvlca Environment Canada Tbla report baa not ondargoao detailed technical review by the Environmental Frotaction Service and the content doee not neceaaarlly reflect the vlewa and polldea of tavlroooent Canada. Mention of trade nanea or coumerclal producta doea not cooetltate endoraaaent for nae. Thle unedited version la undergoing a United distribution to transfer the laforuatloo to people working In related studies. This distribution la not Intended to signify publication and, If the report la referenced, the aether should cite It as an unpublished report of the Branch Indicated below. Any coaaents concerning Its content should be directed to: . Coamerelala Cheelcals Branch Environmental Protection Progress Directorate Envlronnental Protection Service Environment Canada Ottawa, Ontario K1A ICE CO-B February, IMA MOMS 017152 TABLE OP CONTENTS 1 SUMMARY 1 INTRODUCTION 1.1 Background 1.2 Objectives 2 POLYCHLORINATED D1BENZOFURANS (PCDFs) 2.1 Introduction 2.2 PCDFs Their Properties 2.3 PCDFs - Thermal/Pyrolytic Generation 2.S General Observations and Conclusions 3 ELECTRICAL EQUIPMENT 3.1 General 3.2 Transformers 3.3 Capacitors 3.a Electromagnets a CASE HISTORIES .l General Observations *.2 North American Fires b.3 Scandinavian Fires a.4 Conclusions 3 RECOMMENDATIONS 3.1 General 3J Discussion of Recommendations Preventive Measures JJ Discussion of Recommendations Firefighting Measures J.S Discussion of Recommendations Clean-up Measures APPENDIX A CASE HISTORIES APPENDIX B REFERENCES Page iii 1 1 I 3 20 2t 31 37 31 HONS 017153 ii UST OF TABLES Tible J NUMBER OF POSITIONALPCDF ISOMERS 2 PREVENTIVE MEASURES J FIREFIGHTING MEASURES CLEAN-UP MEASURES UST OF FIGURES Figure I BASIC STRUCTURES Page i 25 30 3) j HONS 017154 SUMMARY Ill , The purpose of this publication is to enhance awareness through recommendations (or the prevention of contamination by polychlorinated dibenzofurars (PCDFs) from fires involving electrical equipment containing polychlorinated biphenyls (PCBs). A study was carried out by M. M. Dillon Limited, Consulting Engineers md Planners, In association with Veilington Environmental Consultants at the request oi the Commercial Chemicals Branch, Environmental Protection Service, Environment Cwiada, following a recommendation made by the Interdepartmental Committee on Toxic Chemicals (ICTC) in August 1912. The properties of PCDFs and the conditions under which they are formed we reviewed in Section 2 of this report. A computer survey of literature covering the period from 19(7 to 19J3 was carried out. Human exposure to only PCDFs has not occurred. PCDFs are usually accompanied by PCBs and sometimes by other contaminants, such as polychlorinated dlbenso-p-dioxlns (PCDDs). Although the effects observed in humans cannot be solely attributed to PCDFs, it Is reasonable to assume that these compound like PCDDs ere extremely toxic, both in -the short and the long term. The toxicity ol various PCDF congeners depends upon the position and degree ol chlorination; the tetra- and pcntachlorodibenzofurans being the most toxic. It was found that PCDFs are generated when PCBs and/or chlorinated benzenes (PCBZs) are heated in the presence of oxygen at temperatures In the range of 250*C 700*C. The optimum temperature for the formation of PCDF* appears to be around 300*C. The formation of PCDFs is catalyzed by the presence of certain metals (e.g. iron and copper) or their salts. In addition to PCDFs, PCDDs are formed from PCBZs, but not from PCBs. The types of electrical equipment containing PCBs which are used inside or in a clot* proximity to buildings, are described in Section 3. These are; transformers, capacitors and electromagnets, the latter being quantitatively Insignificant a* compered to the first two. By far the largest quantity ol PCBs still in use are Inside transformers, which on the average contain 1 SCO kg of PCBs per unit. Compared to that, the per unit content of PCBs in capacitors is very small, particularly in those capacitors that are encased In lamp ballasts used for fluorescent and high intensity dbcharge lamps. However, unlike transformers and large capacitors Installed usually In Isolated locations. HONS 017X55 iv the capacitors in lamp ballasts are installed throughout the buildings, often in large quantities as part of each fluorescent lighting fixture. Cast histories of ten fires which involved PCB-containing electrical equipment and occurred during 1977-1982 period, were reviewed. Two of those fires occurred in Canada, one in the United States of America, live in Sweden and two in Finland- A review of the ten fires is presented in Section 4 of the report and all available data on these fires are Included in Appendix A. It is concluded that fires Involving PCB-containing electrical equipment are usually caused by electrical faults or malfunction of electrical equipment. Frequently, these fires start in electrical equipment containing mineral oil and then spread into PCBcontalning equipment. All fires involving PCBs are smokey and yield large amounts of Mack oily soot. This soot is what Is contaminated with PCBs, PCDFs and, if chlorobenzenes are present, PCDDs. PCBs and PCDFs are non-volatile from soot. Soot analysis shows that approximately one per cent of the PCBs are converted to PCDFs. No PCDDs were found in case of fires involving PCB capacitors. Ten times more PCDFs are formed than PCDDs in fires involving PCB transformers. The recommendations contained in Section 3 of the report are presented in three partsi PREVENTIVE MEASURES FIREFIGHTING MEASURES CLEAN-UP MEASURES - Table 2 - Table 3 - Table4 The course of action recommended in Tables 2, 3, 4 fails into four categories one of a general nature and three that address specific areas of concern related tot - Individual Safety and Health - Equipment and Facilities, and - Environmental Concerns It is hoped that this first attempt to produce a comprehensive set of recommendations concerned specifically with the prevention of PCDF contamination resulting from fires of electrical equipment, will form a useful basis lor the development of effective measures to deal with this serious problem. HONS 0X7156 1 INTRODUCTION 1 1.1 Backgroiatd \ Polychlorinated biphenyls (PCBs) are known to produce, under certain condi tions, polychlorinated dibenzofurans (PCDFs) which are more harmful than PCBs. It is estimated that approximately 13 million kg of PCBs are still in use throughout Canada, as an integral part of electrical equipment, mostly in transformers and capacitors. 1 The majority of this PCB-containing equipment Is installed inside or adjacent to buildings. It was found that fires involving PCB-containing electrical equipment croato physio-chemical conditions that are conducive to the generation of PCDFs from PCBs, in significant quantities.2,3 A working group formed by the interdepartmental Committee on Toxic Chemicals (ICTC), entrusted with the development of a management strategy for dioxins and dibenzofurans, prepared an interim report (ICTC Report No. 1) in August 1912. One of the recommendations contained in this report reads) "That guidelines for the prevention of severe PCDF contamination from accidental fires In PCB-containing electrical equipment be developed". The responsibility for the development of the above guidelines was placed with the Commercial Chemicals Branch, Environmental Protection Service, Environment Canada. M. M. DUlon Limited, Consulting Engineers and Planners, in association with Wellington Environmental Consultants, were retained to conduct a study, under the direction of the Commercial Chemicals Branch, that would address the above-mentioned recommendation of the ICTC Working Croup. 1.2 Objectives The principal purpose of this study is to develop a set of recommendations for the prevention of PCDF contamination from fires involving electrical equipment. More specifically, the recommendations are to relate to the conditions: before the fire - during the fire, and - after the fire. HONS 017157 2 The recommendations must be concerned with the aspects of: > heaith and individual safety * equipment and facilities, and environmental control. The recommendations deal with the PCDF contamination only. PCDDs and other contaminants that are known to form during some of the fires of electrical equipment, are outside the scope of this report. Where references to PCDDs appear in this report, they are incidental. HONS 017158 3 2 POLYCHLORINATED OIBENZOFURANS (PCDFs) 2.1 , Introduction Polychlorinated dlbenzofurans (PCDFs) have received considerable attention recently due to their severe toxicological properties and their presence as contaminants h widely used Industrial chemicals such as the chlorinated phenols (CPs) and In (ly ash from municipal Incinerators. In addition, they can be thermally and pyrolitlcally generated from CPs, polychlorinated biphenyls (PCBs) and chlorobenzenes (PCBZs). In Canada, there are considerable quantities of PCBs and PCBZs that are still In service In electrical equipment (see Section 3) or In storage awaiting disposal. Other fluids that are being used as substitutes for PCBs In such equipment may also be contaminated with PCBs (and PCBZs). Fires involving electrical equipment containing PCBs and PCBZs can lead to the formation of PCDFs. It Is also possible that failures in this equipment (e.g. electrical arcing) may also produce PCDFs. The recommendations presented as part of this report were developed so that the formation of PCDFs in electrical equipment can be limited and human exposure to these chemicals minimized in fire situations involving electrical equipment. A thorough knowledge of the properties of the PCDFs and the conditions under which they are formed was essential for the development of these guidelines. For this purpose a computer survey was performed using the Lockheed DIALOG" data system. The literature was surveyed for the period from 19(7 to 1983. Those articles relevant to this project were collected and the Information contained In them Is discussed in the following sections. For more detailed discussions of PCDFs, their chemistry, biological/toxicological effects and their presence as contami nants In industrial chemicals, the reader is referred to the following publications: - Jones, P.A., Chlorophenols and Their Impurities in the Canadian Environment, Environmental Protection Service, Report EPS 3-EC-S1-2, March, 1981. - Chlttlm, B., B.S. Clegg, S. Safe and O. Hutzinger, PCDFs and PCDOs: Detection and Quantification in Electrical Equipment and Their Formation During the Incine ration of PCBs, Wellington Science Associates, Prepared for Environment Canada, Contract No. 03378-000(7 (1979). - National Research Council (NRC) oi Canada (a document on PCDFs Is to be published in early 1988). hO*S 017159 4 U PCDFs - Their Properties U.1 Physical/Chemlcal Properties. The PCDFs are a series of chlorinated tricyclic aromatic compounds having the same basic structure* As can be seen in Figure 1, they are very similar in structure to another series ol compounds, the chlorinated dlbenzo-g- dioxlns (PCDDs)j it has been shown that the properties of these two groups olten, if not always, parallel one another. As shown in Table 1, there are 133 possible PCDF congeners ranging from mono- to octachloro. They are all crystalline solids at room temperature, the individual congeners melting at temperatures ranging from ca. 100 to 250*C. At standard temperature and pressure they have no appreciable vapour pressure.*,3 Like the PCBs and PCDDs, the PCDFs are extremely insoluble in water and only sparingly soluble in moat organic solvents. The solubility of the individual congeners generally decreases with increasing chlorine content.* The PCDFs behave in the environment like most halogenated aromatic compounds; they are highly stable, tend to persist and bioaccumulate In fatty tissues. Although there have been few reports on their ecokinetic properties, it is generally assumed that movement of PCDFs in aquatic systems and in air would largely occur via their adsorption on particulate.* 2JJ Biological/ToxicologicaJ Properties. The discovery ol PCDF contamination ol PCBs was due in large part to the toxicity of tome of the individual PCDF isomers. For example, previous research has shown that certain fractions ol PCBs were tar more toxic to chicks than others and that these fractions contained PCDFs.* The toxicity of individual PCDF congeners is extremely dependent upon the position and degree of chlorination. It would appear that the tetra- and penta- chJorodibenzofurans are the most toxic, in particular, the 2,3,7,S-tetra, 1,2,3,74-penta and 2,3,*,7,S-pentachiorodlbenzofurans.7 The oral LD)o fr 2,3,7,8-TCDF in moat animal species has been reported to be of the order of micrograms per kilogram.*!* Human exposure to only PCDFs has not occurred although related incidents have been reported. in 3apan, over 1,000 people consumed rice oil contaminated with PCBs.10 It teas subsequently postulated that the toxic symptoms observed were due in part to PCDF contamination of the PCBs. In addition, exposure to PCDDs and PCDFs has been HONS 0171*0 5 Polychlorinated dibenzofurans (PCDFs) Polychlorinated dibenzo-p-dioxins (PCDDs> Polychlorinated biphenyls (PCBsh 'V'V/ FIGURE I BASIC STRUCTURES HONS 017161 TABLE 1 6 NUMBER OF POSITIONAL PCDF ISOMERS Chlorine Substitution Mono- ` Number of Cl Atoms in Each Carbon Ring *X 10 PCDFs Number of Isomers in Each Sub-Group a Dl- 20 6 1l 10 Trl- 30 4 2I 2a Tetra- 0 31 22 l 16 21 Ptnta- a1 32 a 2a Hex*- a2 33 6 10 HeptaOcti- a3 aa a l Total a 16 2S 38 28 16 a ___ i. 133 HONS 017162 7 discussed In relation to malignant tumors among spraymen exposed to phenoxy acid herbicides in the 193(Ps and 1960's in Sweden.! 1 Although the effects observed in humans cannot be solely attributed to PCDFs, it is reasonable to assume that these compounds like the PCDDs are extremely toxic, both in the short and the long term. 2.1 PCDFs - Thermal/Pyrolytic Generation 2.3.1 Background. It is well known that PCDFs are present as contaminants in chemicals such as CPs and phenoxy acid herbicides. Of major importance to this study is that PCDFs have also been detected as contaminants in PCBs and askarels and that they exist in new fluids. The total levels of PCDFs found in these unused products have been reported to range from 0.1 to 33 ppm.12 It has also been determined that the levels of PCDFs are higher In used PCB based fluids. For example, in PCBs used as a heat transfer medium, levels of PCDFs up to 300 ppm have been reported.12 The levels of 2,3,7,8-TCDF In askarel filled transformers has also been found to be higher in those units that have been in service for long periods of time.1 * It is therefore most likely that PCDFs already exist in the electrical equipment in service containing PCBs (and PCBZs). Under severe conditions, such as fires Involving this equipment, much higher levels of PCDFs can be formed. Several incidents of this have occurred and have been well documented; these are discussed in Section * of the report. In addition, laboratory research has been carried out to further delineate the conditions under which PCDFs are formed thermally and pyrolytically from PCBs and PCBZs. 2JJ Pyrolysis of Chlorobenzenes. Transformer grade askarels, most widely used, may contain from 30 to *0 per cent (w/w) chlorobenzenes; the balance being PCBs (either Arodor 1239 or 1260), The chlorobenzenes used are mixtures of trichlorobenzenes, primarily the 1,2,9- and 1,2,3- Isomers, and tetrachlorobenzenes, primarily the 1,2,3,9- Isomer. Studies involving the pyrolysis of PCBZs have shown that in the presence of oxygen, 'oxygen containing* aromatic compounds (e.g. PCDDs and PCDFs) are formed while in its absence (e.g. under nitrogen) oxygen free aromatic products are produced.1*,1* , Buser,1* in 1979, reported the formation of PCDFs (and PCDDs) through pyrolytic reactions of chlorobenzenes. In this research, individual as well as combined HONS 017163 8 samples of tri-, tetra- and pentachlorobenzenes were pyrolyzed at 620*C in quartz ampoules with air present. The major components of the linal mixture were chlorobenzene, with decomposition of the lower chlorinated benzenes occurring to a greater extent than the higher chlorinated benzenes (l.e. more than 95* for trichlorobenzenes, approxi mately 90* for tetrachlorobenzenes and approximately 50% lor pentachlorobenzenes in the combined sample). Other products of the pyrolyses included PCDFs, PCDDs, PCBs, CPs, chlorinated naphthalenes (PCNs) and chlorinated styrenes (PCSs). Significant quantities (l.e. up to 10% yields) of PCDFs were formed In the pyrolyses. Generally the PCDFs formed had chlorine numbers of 2m-2, 2m-1 and 2m where m is the chlorine number of the chlorobenzene used (e.g. pyrolysis of the trichtorobenzenes gave tetra-, penta- and hexachlorodibenzolurans). According to Buser the thermochemical formation of the PCDFs (and PCDDs) from the PCBZs is bimolecular and so their formation is dependent on the PC8Z concentration. Buser also proposed that the CPs (present In all pyrolyzed samples) were Intermediates in the formation of the PCDFs: Reaction of the CPs with PCBZs could lead to chlorinated diphenylethers (PCDPEs) which upon pyrolysis are known to yield PCDFs. 2J.1 Heating/Pyrolysis of PCBs. Investigations into the presence of PCDFs in PCBs lead to the now well-known fact that heating of PCBs at sufficiently high temperatures results in the formation of PCDFs. Buser and his co-workers1* found increased levels of PCDFS On the ppm range) In a commercial PCB mixture heated at 300*C for a week in the presence of air and in a PCB used in a heat exchange system. Several groups have since studied the pyrolytic/thermal formation of PCDFs from PCBs In the laboratory to determine the mechanisms Involved, the additional reactants necessary and any catalysts' that may enhance the formation of these compounds. The results of these studies are discussed in the following sections. MOhS 017164 9 2.3.3.1 Mechanisms. In 1979, Buser and Rappe16 studied the pyrolyses of IS PCB congeners and determined that the formation of PCDFs in these reactions resulted from intramolecular cycllzatlon processes. They determined that the formation of the PCDFs from PCBs follows one or more of four general reaction routes: a) Loss of ortho-Cl? Cl* ci ci _i!i c,y Cl CL b) Loss of HC1 with or without a 2-3 chlorine shift Q-Qf ciV ^ci cTh * y c) Loss of ortho-HCI K V/ A CL* -----c^i h ' CLy HCI 9. vy .w ci HONS 017165 d) Loss of Qftho-H? 10 A further mechanism for the formation of PCDFs from pyrolysis of PCBs has been proposed as part of an investigation of trace organic compounds in fly ash from municipal and Industrial incinerators. Choudhry and Hutzinger1? have proposed that the reaction proceeds via the mechanism shown below, that is: i) chlorinated diphenyl radicals are first formed from PCBs by either abstraction of ortho-hydrogen by free radical species (e.g. R*l, alkyl or aryl, Cl* orH*) or by scission of an ortho C-Cl bond, li) reaction of these diphenyl radicals with O2 to form peroxide radicals which In turn form ortho-hydroxychlorinatedbiphenyls, and finally. Hi) cydization of these chlorinated biphenyls to form PCDFs. HONS 017166 II The intermediation oi the biphenylols is supported by their detection in PCS pyrolysates.19 UJJ Temperature. PCB pyrotyses have been carried out in the laboratory at temperatures ranting from 200`C1* to *50*C.l9 There is some discrepancy, however, as to what effect temperature has on the amounts of PCDFs formed. The optimum temperatures for PCDF formation in two separate studies were reported to be 300`C20 and JJO'C1* while in yet another study, the authors found that at 330*C the PCDFs formed began to decompose.21 The minimum temperature reported lor PCDF formation is 270*C by Morita and his co-workers.2* in their article they suggested that the PCDF levels formed are governed by the transient equilibrium of thermal formation and decomposition. Buser et ai.19 investigated the pyrolysis of two hexachlorobiphenyls at temperatures ranging from 550*C to S50*C. PCDFs were formed at temperatures of 330 to 650*C. However, at and above 700*C complete destruction of these PCDFs occurred. 2JJ.3 Time. Only one study has been reported on the effect of time on the pyrolytic formation of PCDFs from PCBs.21 in this the authors found that at 300*C the amounts of PCDFs formed increased to a maximum (7 days with O2, 14 days with air) and then gradually decreased. 2JJ.I Atmosphere. It is obvious when one examines the structures of PCBt and PCDFs that a source of oxygen is needed for this conversion. This was confirmed in one report where PCBs were pyrolyzed under oxygen, air and nitrogen.21 PCDFs were formed in good yields when oxygen was used; lesser amoimts were formed (ca. one tenth that of Oy) in the presence of air and very little, if any, were formed under nitrogen. (The PCDFs formed were attributed to either oxygen, water or tome other oxygen containing contaminants in the nitrogen.) 2JJ.) Catalysis. Addition of Fe,1 FeClj20 or Cu-Fe powder20 has been found to Increase the amounts of PCDFs formed during the pyrolysis of PCBs. It is likely that the reactions involved are similar to Ullman type catalyzed reactions12 in which cate many other metals may also catalyze the reaction. Ud Pyrolysis of Askareb. As discussed earlier, askarels are mixtures of PCBZs and PCBs. Pyrolysis of these mixtures is therefore likely to yield PCDFs and PCDDs (the latter from the PCBZs). To our knowledge only one study has reported on the in-lab pyrolysis of askarels.'5 In this samples oi Aroclors* 1234 and 1016 and the askarels Pyranol*, HONS 017162 12 Inerteen* and Chlorextol* were pyrolyzed under air at 300*C. The pyrojysates were analyzed for PCDFs and the tetrachlorodibenzofurans (TCDFs) quantitated. The askarels all yielded significant quantities of TCDFs (ca. *0 to JOO micrograms TCDF per gram of starting material) as did Aroclor 12JS. Arocior 1016, on the other hand, yielded much lower quantities of TCDFs. This is most likely due to the lower chlorine content of the PCBs In this mixture (i.e. very little perns- and hexachloroblphenyls). It is interesting to note that PCDDs were not detected. It is possible that they were formed but In such low quantities that they were not detected or were masked by the higher levels of PCDFs. 2.3.3 Pyrolysis of PCB Contaminated Mineral OU. To date there have been no reports of experiments involving the heating or burning of PCB contaminated oils (of any type) and analysis of the products formed. Mineral dielectric insulating oils (also called naphthenic or white oils) are mixtures of paraffinic (ca. 70%) and aromatic (ca. 30%) hydrocarbons.22 These oils are frequently contaminated with PCBs in concentrations that often exceed 50 ppm (see Section 3). It is difficult to predict what quantities of PCDFs will be formed in fire situations involving such units. If a 1% PCB to PCDF conversion is assumed, a 1 000 L mineral oil filled transformer containing 100 ppm PCB would yield 1 ppm PCDFs or a total of 1 gram of PCDFs. The conversion of PCBs to PCDFs could, however, be higher or lower. For example, the temperatures Involved in a mineral oil fire could be high enough so that all of the PCBs and PCDFs are destroyed. On the other hand, the mineral oil could act as a fuel continuing to burn over a longer period of time (unlike PCBs or askarels which would quench a fire) and potentially produce more PCDFs. _ There is also the possibility that the mineral oil components could take part chemically in the conversion of PCBs to PCDFs as in the reaction mechanism proposed by Choudhry and Hutzinger17 (see 2.3.3.1 above). This is one area where much more research is required. M General Observations and Conclusions From this laboratory research several conclusions can be made with respect to the thermal/pyrolytic generation of PCDFst HONS 017166 13 PCDFs are formed when PCBs and/or PCBZs are heated in the presence of oxygen at temperatures greater than ca. 230*C but less than ca. 700*C. The optimum temperature for the formation of PCDFs is ca. 500*C, however lower temperatures have been reported. - Oxygen (pure 02, air or possibly other oxygen containing compounds) is required. The yields of PCDFs are dependent on the amount of oxygen present. - The formation of PCDFs is catalyzed by the presence of certain metals (e.g. iron or copper) or their salts. - Percentage yields of PCDFs (i.e. 1% or greater) have been reported when PCBs, PCBZs and askarels have been pyrotyzed. - In addition to PCDFs, PCDDs are also formed when PCBZs are pyrolyzed. To accurately predict what will be formed in a fire situation Involving electrical equipment is, however, complicated by the number of variables. This will become more evident when reading the "case histories" in Section a of this report. In conclusion it must be assumed that PCDFs are formed in such situations. HONS 017109 14 3 ELECTRICAL EQUIPMENT 3.1 Genera] Fire resistant dielectric liquids, known under the genericname of askarels, have been used for electrical equipmentin Canada sincethe early1930's.Transformer askarels consist of polychlorinated biphenyls (PCBs) usually blended with chlorinated benzenes (PCBZs) in varying ratios.23 Transformer askarels most widely used In Canada contain (0 or 70 per cent of PCBs and 40 or 30 per cent of PCBZs. Capacitor askarels (A5TM Types 2233A, B and D) contain no PCBZs only PCBs.2* Askarel Type 2233C is a mixture of 73 per cent PCBs and 23 per cent trichlorobenzene. Most of the capacitors used in Canada contain no PCBZs. The following electrical equipment containing askarels have been used for indoor applications] - transformers - capacitors - electromagnets. The main reason for using askarel immersed equipment indoors, in preference to oil immersed equipment, is the superior fire retardant property of askarel compared to mineral Oil. The main reasons for using askarel immersed equipment indoors, in preference to dry type, air cooled equipment, are: - the superior dielectric strengthof askarel - suitability for use in areaswhere the ambient atmosphere is badly contaminated with dust, dirt, or corrosive fumes - where the ambient atmosphere is extremely damp - where the hazard of lightning surge is unduly high. L2 Transformers The transformers which present a potential PCDF contamination hazard, fall Into two categories] - those that were designed to use askarel as the dielectric andcooling medium, and - those that were designed to use mineral oil as thedielectric and cooling mecfium, but contain PCBs due to adventitious contamination of the transformer oiL Askarel transformers were designed as sealed tank type units complete with pressure relief vents. Relief vents are usually designed to operate at a tank pressure In HONS 017170 15 the range of S to 10 psi. The tanks could bulge when tank pressure exceeded 10 psi, and tank rupture could occur with pressures In excess of 25 psi. The typical range of askarel immersed transformers is 300 kVA to 5 000 kVA, with an average transformer containing approximately 1 *00 kg of PCBs. Askarel will not support or sustain combustion until it reaches its boiling point at approximately 205*C. ft has a high dielectric strength, exceeding the dielectric strength of mineral oil. Internal arcing faults may create conditions in which PCDF will be generated. With respect to PCDF formation potential, two types of fault in the transformer windings have been considered: - short circuit within the transformer, and escalating high impedance fault, accompanied by arcing. In the first case, the tank should remain intact, being disconnected by the primary circuit breaker or fuse protection assembly. The speed at which the transformer will be disconnected depends on the magnitude of the fault current, and the speed at which the protection mechanism operates. The second case is more serious. This type of arcing fault may cause the current to increase over an extended period of time, increase the askarel temperature, and increase the tank pressure to a dangerous level. This will result in the relief vent operating, causing askarel and gases to be expelled from the tank. 3.3 Capacitors Capacitors containing askarel are used for the following applications: power factor correction and voltage regulation of high-voltage lines and transformer stations - power factor correction for indoor power distribution systems - fluorescent and high intensity discharge (HID) lamp ballasts. Capacitors containing PCBs have been manufactured since the 1930`S up until 1979 when askarel was no longer available. All askarel capacitors are sealed in metal enclosures. It would be necessary for the enclosure to rupture to spill askarel. Dielectric breakdown within a capacitor could cause hot spots within the capacitor, and the formation of PCDF, without rupturing the capacitor enclosure. HONS 017171 16 High voltage power factor correction and voltage regulation capacitors are normally located outdoors on poles or structures, and at outdoor transformer stations. Indoor power distribution systems may have power factor correction capaci tors located at the main secondary switchboard, motor control centres, or at individual electric motors. The degree of hazard is similar to that of small askarel transformers. Capacitors used for power factor correction are installed either outdoors or in Isolated indoor locations. Capacitors used in lamp ballasts, however, are installed throughout the building as part of fluorescent or HID lighting systems. In case of a large building, the lighting system may include many thousands of PCB capacitors encased in the lamp ballast. Therefore these lamp ballasts present a higher degree of hazard than power factor correction capacitors, in spite of the fact that a single lamp ballast unit contains relatively small amount of PCBs.2* 3.* Electromagnets . Electromagnets may be located in industrial plants, scrap metal yards, and mineral conveyor systems. The number of units and total amount of PCB contained in electromagnets is considered to be small compared to transformers and capacitors. The degree of hazard Is similar to that of small askarel transformers. HONS 017172 * CASE HISTORIES 17 Cue histories were collected through a combination of personal contact, telephone conversations, literature search and newspaper information services. They are given in detail in Appendix A. All occurrences, actions, etc., are discussed with reference to the situation before the fire, what happened during the fire and how clean-up was achieved after the fire. 1.1 General Observations Electrical fires in transformers and capacitors have historically been very common occurrences. Toronto Hydro has documented an average of one transformer fire per year. Only one of those fires involved a transformer containing PCB. There have been few documented incidents of PCB fires. Several hypotheses are advanced for this. The first and most common is that PCBs do not burn easily, therefore they do not become involved in fires: secondly awareness that PCBs and their combustion products act as extreme risks if involved in electrical fires did not exist until alter the Binghamton, New York fire in 1911; a speculated third reason for minimal documented Incidents in North America particularly, Is that companies that have small fires involving PCBs do not report their Involvement in order to avoid expensive cleanups and Intense public concern that coincides with PCB Incidents. PCB fires are a normal occurrence in the U.S. but all except Binghamton have been quite small with no in-depth study or analysis made into them. The majority of PCB fires go by without PCDFs being tested for, due to the lade of awareness of the hazards ol PCDFs.** All of the North American incidents documented involved fires in transformers due to electrical malfunction causing fire. In ail of these cases the characteristics were very consistent. They involved very small amounts of flame giving oft large amounts of heat and thick black smoke leaving residues of oily black soot. All of the Scandinavian incidents documented involved capacitors containing PCB. All but one of these fires were caused by electrical malfunction in the capacitors. In all of these fires PCDFs and no PCDDs were found in the resulting soot. The lack of PCDD formation in the Scandinavian fires has been explained by the fact that chloroben zenes were not used in the capacitors Involved in these fires. HONS 017173 IS M North American Fires Case histories of three North American fires involving PCB-containing electri cal equipment were reviewed. These ares Toronto Hydro Fire Binghamton State Office Building Fire University of Manitoba Fire J Scandinavian Fires 9 December 1977 J February 1981 29 March 1982 Seven PCB fires or explosions have occurred in Sweden and Finland since 1977. These are, In chronological order: * Norrtalje, Sweden Danviken, Sweden Skovde, Sweden * Imatra, Finland * Helsinki, Finland Surahammar, Sweden * HaJ1stshammar, Sweden 3une 1977 August 1981 March 1982 August 1982 August 1982 September 1982 November 1982 All of these fires involved capacitors containing PCB only (no chlorobenzenes) and, with the exception of the Surahammar fire, were caused by electrical malfunction within a capacitor. * Specific details are not shown. .* Condusiom The case histories discussed in this report are typical of electrical fires involving PCB containing transformers or capacitors. In fires Involving PCB equipment consistent observations can be reached. The fires are rarely, If ever, caused by the PCB itself. The PCB fluid normally acts as a by stander to a tire caused by an electrical malfunction or a mineral oil fire. Fire Involving PCBs are smokey and yield large amounts of black oily soot. This soot Is what is contaminated with PCBs, PCDFs and. If chlorobenzenes are present, PCDDs. Testing has Indicated that PCBs and PCDFs are non-volatile from soot. In all of these case history studies testing was conducted to determine the amount of PCDFs termed during the fire In relation to the amount of PCBs available. It was found that approximately one per cent of the PCBs are converted to PCDFs and 10 times more PCDFs than PCDDs are formed. HONS 017174 19 J RECOMMENDATIONS M . General A review of current hazardous material regulations shows that the PCDF issue has not been specifically addressed in most cases. There is a considerable body of work on PCB fires, however, and it is this that forms most of the basis for the handling of specific aspects of PCDF incidents. The bulk of hazardous material legislation which is generally applicable to PCDFs must be applied to these materials with caution due to their extreme environmental and toxic threat. The development of recommendations can best be pursued by subdividing the issue into 3 partsi Preventive Measures (Table 2) Procedures to reduce the probability of fire may eliminate the threat at the source. Planning for a fire, in terms of the measures to extinguish it and contain its products, can limit the PCDF threat to a tolerable level. Firefighting Measures (Table 3) Firefighting methods that result in quick control of the fire with the minimum damage to health, safety or the environment must be employed. Procedures should consider the clean-up requirements after the fire. Clean-up Measures (Table ) Techniques for hazardous material spill clean-up have been well developed, and In many cases apply directly to PCB fires. Refinements to these techniques based on PCDP's special problems and the high level of cleaning necessary are, however, required. These three phases of a PCDF contamination control program can each be addressed in terms oft - general - individual health and safety facilities and equipment environmental concerns HONS 017175 20 U Dlculon of Recommendations - Preventive Measures GENERAL ' 1. Identify Equipment Each piece of electrical equipment containing PCDF precursors should be identified, properly lebelled, and its location recorded on a schematic lay-out of the facility. This will prevent exposure to those taking uninformed actions in case of fire, and instill caution. In addition, areas to be avoided or restricted to authorized personnel can be more eesily defined. The emergency response crews (fire department, environmental protection departments) should also be informed on the identity end location of such equipment in order to, in the event of an emergency, enable them to use appropriate material and Implement adequate countermeasures. 2. biform Personnel Personnel should be informed on the procedures to follow in case of fire. Fire wardens should be familiarized with the additional hazards posed by PCDFs, end with any special countermeasures necessary in the event of a fire involving electrical equipment and possibly PCDFs. X Set Up Contingency Plan This plan should contain dear courses of action and various options for dealing with emergency situations. The plan should also consider evacuation needs and actions to be taken prior to arrival of emergency response crews. This contingency plan should also be provided to and discussed with emergency response crews. HONS 017176 INDIVIDUAL HEALTH * SAFETY 21 These measures are primarily aimed at the benefit of personnel; they will also facilitate action by the emergency response crews. L Ensure Training of Personnel This will prevent exposure to those taking uninformed actions in case of fire, and Instill caution. Practice, rehearsal and training are necessary to ensure that personnel can be evacuated or take appropriate precautions and countermeasures, pending arrival of emergency response crews. 2. Ensure identification of Equipment . 3. Enstre Availability of Adequate Personal Protective Material HONS 01717? EQUIPMENT tc FACILITIES 22 These meesures ere primarily aimed at protecting equipment and facilities; they will also protect people and lower the costs of repairs. 1. Implement and Rigorously Follow a Preventive Maintenance and taspectien Program Repair of equipment before it becomes hazardous Is particularly desirable In this case since the benefits of preventing fire incidents outweigh the costs. Rebuilding of transformers is not recommended. 2. Equipment Should Not Be Operated Above Design Ratings Undersized (overloaded) equipment is more prone to deterioration and failure, thereby increasing the risks for fire. 3. Install Alarms and/or Shut-down Systems on Equipment Shutting the power off reduces both the -arcing and the production of PCB/PCBZ degradation products; it also cuts the energy input and reduces the intensity of fires resulting from it. In addition, firefighting is always delayed pending equipment shut-off. During that delay, more damage to equipment can occur as well as more generation of smoke and fumes. For transformers, a two-stage gas alarm relay installation should be conside red; the first stage will trigger an alarm when the tank pressure reaches a certain level (*! 6 PSIG or ca 10 kPa) and the second-stage will de-energize the transformer (and possibly shut down Electrical Room ventilation, including dampers) when the tank pressure reaches a higher level (e.g. > PSIG or ca. }) kPa). *. Install Automatic Fire Extinguishing Systems A properly designed fire extinguishing system can begin the fire control procedure even before emergency response crews are aware of the fire. & fcstall an Auto-dose Mechanism on Ventilation Systems . The ventilation system is a major route of contamination. Smoke and fumes may be carried throughout an entire building by convective effect, resulting in massive widespread contamination from a relatively minor localized occurrence. HONS 017178 23 (. Do Not Mix PCB Equipment With Non-PCB Equipment holate PCB Equipment , PCB transformeri and capacitors should be physically separated from non-PCB equipment so that in case of a fire in a mineral oil-filled transformer, the fire will not Involve PCB equipment. A fire starting in PCB equipment may not be self-sustaining when the power is off, whereas a lire in an oil-filled transformer will continue to burn. If PCBs are involved In a self-sustaining fire the difficulty of control of contaminant dispersion is compounded; not only must contaminants be contained, but the containment must take place during the control of another and seemingly more urgent emergency. In such a situation it is difficult both to assign environmental priorities and to carry out tasks effectively. 7. Remove PCBs From Equipment and Retrofit! With Non-PCB Fluids Where possible, this course permits the reduction of hazard without incurring the major capital expense of replacing the equipment. There are several caveats with this course: the flammability with a retrofitted appliance is greater than with a PCB appliance; draining and retrofitting of equipment may not be practical; the replacement fluid may contain more than 30 ppm PCBs due to Incomplete draining. HONS 017179 24 ENVIRONMENTAL CONCERNS ment. These measures are primarily aimed at minimizing dispersion into the environ 1. Emure Containment of Fluid PCB equipment must be within a containment system capable of holding more than the equipment volume of fluids. Berms, dams, curbs or other barriers around the equipment should be capable of holding the equipment contents, plus any other fluid likely to be involved. Drains in the floor must be outside the containment area. A drain sealing system must be in place to prevent contaminant from escaping to the sewer system. 2. Emure Containment of Contaminants All surfaces (walls, ceilings and floors) should be painted with impermeable (to PCBs and PCDFs) material. This precaution will facilitate clean-up and may also avoid the necessity to remove building materials during the decontamination process. HONS 0l?l80 TABLE 2 PREVENTIVE MEASURES General 1. Identify equipment 2. hferm persannel X citlBfwcr fatdlvlduel Health end Safety I. Ensure traMaf *f personnel 2- Gsem identification mi equipment 3. Ensure availability of personal protective material Equipment and FaciHties Environmental Concerns I* Implement and rigorously follow a preventive mafattensnee and faiqMction program 1. Ensure containment of fluid 2. Equipment PwH not be operated above rated capacities 2. Enure containment of contaminants 3. fatstaO alarms adfar Putdosn systems on equipment Install automatic Or* eetfaifiriahing systems 3. btstal an auto cleat mechanism on nntilMkai systems f. 0* not mis PC&>centaMnf equip ment widi rwn-PCB-contairtinf equipment halite PCIkvrtahbq equipment /. Replace PCBs/PCBZs with substitutes HONS 0 1 7 1 8 1 26 J.) Discussion oi Recommendations - Firefighting Measures GENERAL 1. Worm Emergency Response Crews Warn emergency response crews about the identity and location of equipment Involved in the lire. This will enable the use of proper extinguishing techniques. Including personal protective material and countermeasure precautions. 2. Consider Evacuation of Neighbouring Buildings All ventilation openings (inlets and outlets) in the building should be dosed automatically in case of a fire. Otherwise evacuation of adjacent buildings must be considered whenever there is any possibility that contamination could have reached other buildings. Considering the hazardous nature of PCDFs a conservative course must be followed in order to protect safety and health. 3. Access to Area Should be Restricted to Essential Personnel HONS 017182 INDIVIDUAL HEALTH * SAFETY 27 1. .Ensure That Emergency Response Crews Are Aware at the identity end Location oi Equipment 2. Protect Against All Anticipated Hazards Protection against PCDFs should be based on a limit of 1.5 ng/m3 (2b hour average)!. This means use of full protective suit and self-contained breathing apparatus. Note that other hazards are likely to be associated with an electrical fire aa well - heat, asphyxiants and corrosive atmospheres. Protective clothing should guard against all of these. For PCDFs, as for any unknown or highly toxic material, response personnel requires - self-contained breathing apparatus fully enclosed chemical-resistant suit - protective headgear, gloves, boots Where personnel has been exposed to gases from a fire, decontamination of both protective material and skin is required. Medical testing is also required. X Arrange for Emergency Medical AM on call. Due to the extremely toxic effects involved, emergency medical aid should be . * The Ontario Ministry of Labour, in its "Chlorinated Dioxins and Chlorinated Dibenzofurans Ambient Air Guidelines (Dec. 19*2)" has recommended an annual exposure limit for TCDD of 50 pg/m3. PCDFs have not been specifically addressed, but the limit for mixtures of TCDD and PCDFs Implies that TCDD is more toxic by a factor of 50. This would suggest an annual average limit of 1 500 pg/m3, Le. 1.5 ng/m3. HONS 017183 EQUIPMENT * FACILITIES 2S 1. Ensure That Power b Off A power-on situation complicates firefighting considerably, requiring extra precautions to avoid potentially live sources, both for personal safety and fire control. 2. Us* Correct Firefighting Materials - Do NOT USE WATER Water provides a cooling effect in firefighting. It should never be used on an electrical fire. Firefighting with water has resulted in transformer explosions. Water b generally unsuitable for use with fires involving organic liquids as it causes them to spatter and to flow over a wider area without effectively cooling the fuel material. Carbon dioxide (CO2) and HALON* act by segregating the oxidant (O2, Air) from the fuel. Although they do not have as significant a cooling effect as water does, they do not have the drawbacks that water has. 3. Ensure That Air Inlcts/Outlets Are Closed If one of the three necessities for fire b absent - in thb case the oxidant (02, Air) - then control and extinguishment may be achieved at much less cost and with much less damage. Once a fire has been extinguished in this way CAUTION b required to avoid re-ignition when air b re-admitted, if ignition sources still remain. HONS 017184 ENVIRONMENTAL CONCERNS 29 1. Ensure That Access to Building* Ventilation b Closed ` M this is not done, what should have been a localized conlined contamination can become a major environmental problem, contaminating ail areas accessible to the building's ventilation system. HONS 017185 TABLE 3 FIREFIGHTING MEASURES Cmml 1. hitnMMrgiKy mpam crm beUdbgs X Acci- R area dtould te rwticirt esasntial Hrwww1 DO NOT USE WATER Mhridual Health and Safety Equipment and Facilities 1. Ewiv tfitt imr|iiicy respwsse Cim m wm el At Identity end lecatton ( equipment 1. Ensure that power h off 2. AoWCt ifiiMl an tntklpiUd hazards 2. Use correct flrefigHtinf materials DO NOT USE WATER 3* Airifl|t far nnttincy medical aid 3. Ensure that air inists/outlets are closed BirlrenmstiUl Concerns 1. Ensure that access to butMine* ventilation is closed HONS o 31 M Discussion of Recommendations - Clean-up Measures GENERAL 1. Restrict Area to Essential Personnel 2. Monitor Contamination ci the Area Determination of the extent of contamination will: - indicate the degree of hazard, both personal and environmental provide boundaries for the clean-up and how much is needed - make prompt treatment of affected personnel easier - provide valuable information lor the establishment of clean-up measures and lor the need of protective equipment - Indicate when clean-up is satisfactory ' 3. Set up Decontamination Procedures Decontamination procedures must ensure that no hazardous material remain on the clean-up crew members, and that clothing and equipment has also been cleaned or disposed of. HONS 017187 32 INDIVIDUAL HEALTH & SAFETY 1. Worm Personnel Personnel must be informed and trained in the occupational health concerns of PC6s, PCDDSt PCDFs and the cleaning materials used. Training of clean-up crews in recognizing the hazards of the task is a necessary adjunct to the provision of protective equipment and health and safety facilities. 2. Year Proper Protective Material Although fire-related hazards are not present in the clean-up phase, protection against PCDFs must be maintained at a high level. Generally, protective equipment must feature provision for decontamination or discard as well because of the nature of clean-up work (active, hands-on scouring activities). 3. If Clean-up b in an Enclosed Area, Set Up an Air Cleaning Vacuum System Reduction in air contamination will reduce the inhalation hazard and permit less stringent breathing apparatus requirements. Skin contamination from airborne particulates and recontamination of cleaned surfaces will also be reduced. a. Ensure That Medical Back-up b Available Prompt medical attention b essential for any case of exposure to the substances produced in a fire involving electrical equipment. Personnel feeling ill or who feel they have absorbed or contacted contaminant mutt tee the physician immediately. Response crew members should undergo medical examination if there has been any possibility of exposure. HONS 017188 EQUIPMENT & FACILITIES 33 1. Use a Curtain Vail System to Segregate Cleaned Areas Irom Undewed ones The moving curtain wall technique is effective in permitting the intensive cleaning of small sections of an area without recontaminating the area with debris or mobile contaminant from uncleaned sections. Although installation and moving of curtain walls Is expensive in terms of material and time. It minimizes the need to reclean areas. 2. Clean Up Soot Immediately The PCDFs produced in a fire involving PCB-equipment Is largely present in the carbonaceous residues. The vapour pressure of PCDF adsorbed on soot Is very low so the handling and protection techniques suitable are those for fine particulate solids. PCDFs are quite strongly adsorbed, so that removing the soot will effect removal of much of the hazard. The Site Should be Cleaned with a High Efficiency Vacuum dewier. Followed by a Wash with Organic Solvent and Detergent. Washing Should be Hepeased Until the Surface Contamination is Reduced to 10 ng/m2 PCDFs. The major location of PCDF contamination is expected to be In soot deposited on surfaces. Removal of loose soot is best accomplished by vacuuming, with high efficiency filtration for the exhausted air. Remaining soot adhering too well to be removed by vacuum, is removed by washing with organic solvents. Although many solvents have sufficient solvent power, those with low vapour pressures, posing no flammability or health problems, are the best examples. Water and detergent systems, with mechanical scouring, may also be effective, particularly as a final cleaning step. All cleaning materials and residues must be contained for treatment or disposal. X bvpect Ventilation Inlets of Contaminated Areas and Adjacent Buildings Ventilation inlets of all buildings in the vicinity of the fire should be closely examined. Should any soot be found at these inlets, it must be analyzed immediately and appropriate clean-up procedures must be put into effect with respect to all contaminated areas. HONS 017189 34 ENVIRONMENTAL CONCERNS 1. Maintain the Contaminated Area Under Negative Pressure Prevention of contamination spread is achievable by maintaining the contami nated area under reduced pressure. This technique is best used with high-efficiency filtration of exhausted air. 2. Exhaust Air Through High Efficiency Filter To maximize the benefits of decontamination, containment must be during the Clean-up phase. PCDF contamination is predominantly absorbed to carbon particles. This carbon must be removed from air before it is exhausted, to prevent spread of contamina tion. Contamination by PCDF vapour has not been detected so that particle removal Is an effective way to reduce PCDF dispersion. J. Decontaminate Workers and Clean-up Equipment on Site This practice keeps contamination localized. In addition to decontaminating personnel, the decontamination material must be monitored and maintained contaminantfree. 4. Store Contaminated Materials In Labelled, Numbered Containers Labelling containers is necessary for hazard identification. Numbering them reduces the possibility of loss and permits better control of wastes prior to storage or disposal. All clean-up materials must be considered contaminated and must be stored until treated or disposed of. This requirement includes clothing and tools; once these are decontaminated, the trash liquids must be treated as contaminated and stored. It is advantageous to store material in numbered, identified containers, to enable quick recognition of any losses. 5. Dispose of Contaminated Material in an Approved Hazardous Waste Disposal Faculty Contaminated materials should be collected in labelled, numbered drums and moved to the hazardous waste faculty. MONS 017190 TABLE* CLEAN-UP MEASURES General I. toWci ar-- id parwwl btdMdual HmMi and Safety 1. Mena personnel 2. MerHot canlambniton of theme 3. Sata, dacwinlnntfan fraciAaii 2. Wear proper protective material J. If clean op is in an enclosed area, set op an air cleaning vacuum system %. Enure diet medteal beck-4* is avaiahk equipment and Facilities 1. Use a curtain wall system to segregate cleaned areas Iron* wcleaned ones 2. Clean up soot immediately 3. bnpect ventilation Inlets of contaminated areas and adja cent buildings Emironmenal Concerns i. Maintain the contaminated area under negative pressure 2. Exhaust air through high effi ciency filters X Decentaminate markers and material on site *. Stare contaminated material in labelled, numbered containers 5. Pli,m of contaminaltd material in an abated huardeta waate diipoaai facility 37 APPENDIX A CASE HISTORIES HONS 017192 39 TORONTO HYDRO FIRE - 9 December 1977 , An electrical fire in an askarel transformer located in an underground transformer vault under the sidewalk across from (0 Adelaide Street East, Toronto. Before A transformer vault designed for two oil insulated transformers. In 1965 one of the oil insulated transformers had to be replaced with a 1 500 kVA askarel transformer containing 600 gallons (I MO L) of askarel. The askarel transformer was equipped with a 2 500 amp low voltage circuit breaker (called a network protector by Toronto Hydro). On 9 December 1977 the transformer vault contained one oil insulated transformer and one askarel transformer. At the start of the morning rush hour on Friday 9 December 1977 a fault . (crack) developed in the network protector (circuit breaker). The crack was at a position on the protector that could not be cleared immediately by the protective relays. During Arcing took place across the crack starting a small electrical fire. The crack spread to the porcelain bushing on the askarel transformer causing one hundred gallons of askarel to leak out onto the floor, some of the askarel vapourized across the electrical arc. Protective relays in the substation opened the primary feeder. Low voltage fuses blew disconnecting the askarel transformer lrom the oil transformer. THE OIL INSULATED TRANSFORMER WAS NOT INVOLVED IN THE FIRE 22 firemen from four fire halls were at the scene. THEY DID NOT AT ANY TIME ENTER THE VAULT. Firemen filled the vault with COj (standard practice) through the ventilation gratings and extinguished It. Firefighters domed breathing apparatus (standard smoke breathing apparatus) because of volumes of black smoke coming from the vault. THEY DID NOT KNOW AT THIS POINT THAT PCBs WERE INVOLVED. HONS 017193 *0 The 100 gallons oi askarel that spilled were contained in a pit around the transformer. Soinq of the askarel that vapourized escaped through ventilation gratings in the sidewalk contaminating the front of an adjacent building and the surrounding area. The buildings adjacent to the fire were not evacuated. After Primary feeder supplying the oil filled transformer was opened to completely de-energize the spot network. spilled. Toronto Hydro personnel entered the vault and determined that askarel had ONLY AFTER THE FIRE WAS IT KNOWN THAT PCBs WERE DEFINITELY INVOLVED. . Hydro personnel isolated the askarel unit so that power could be restored to an adjacent building. An electrical maintenance company, experienced In dealing with askarel, pumped the unspllled askarel out of the transformer and removed the transformer. Some Hydro employees complained of illness after the fire, symptoms included headache, eye discharges and irritation, nausea, vomiting and skin rashes similar to cold sores about the face and hands. The 22 firemen and 23 Hydro employees who went into the lire without proper protective equipment were tested tor PCB levels one week after the fire occurred! no high PCB levels were found. Firemen underwent dose medical scrutiny for ( months alter the fire: men showed no ill effects. Much controversy stemmed from the fact that firemen. Hydro employees and other personnel at the site were not told of the PCB danger until five days after the fire. Clean-uo The area was enclosed by a sand berm and barricade. The vault was decontaminated by spraying first with No. 2 Fuel Oil and then with detergent by an experienced contractor. All contaminated dothlng and equipment from the dean-up contractor, Toronto Hydro personnel and firemen who had been on-site, was collected. HONS 017194 *1 Any fixtures in the vault that could not be decontaminated were removed tor disposal. A second contractor was hired to wash down the wall of an adjacent building that had been contaminated. Cleaning agents and water used for cleaning were collected lor disposal. Contaminated water, clothing, equipment, fixtures etc. that could net be cleaned were disposed of at a PCB secure landfill in Niagara Falls, New York. PCB in air levels were tested in the vault after initial clean-up and were fouid to be above MOE standards of 30 ug/ml. Vault was re-scrubbed and cleaning agents collected as in initial clean-up: H reduced PCB levels sufficiently. Soot containing 10 000 ppm PCB covered a 700 ft^ area adjacent to the fire site. PCDFs were found in concentrations of 3 ppm in the soot. Dioxin was detected in the test samples of air and soot taken at the site in late December 1977 but firefighters and Hydro personnel were not told until February 197S. Cost $100 000 Cost Includes: Decontamination of building; vault and surrounding area - Disposal of contaminated material - Value of transformer destroyed and labour for its replacement. HONS 017195 *2 BINGHAMTON STATE OFFICE BUILDING FIRE - 5:30 ajn. Thtnday. 5 February IM1 An extremely hot electrical lire in a switching mechanism. The heat caused bushings on an askarel transformer to crack and askarel to leak out. PCB contaminated soot was transported throughout the 18-storey building. Before Occurred at Binghamton State Office Building Central tower within a governmental building complex building completed Spring 1973 and rises 13 storeys (260*) from street level with two subsurface levels 700 employees of 33 state agencies normally occupy building. Building occupied by two people (a security guard and a stationary engineer) at the time of the fire. Purine An electrical failure in the building's main switching equipment within a basement mechanical room, possibly caused by a loose connection or a piece of dirt. The electrical failure touched off an intensely hot fire. The heat caused the porcelain bushings on a nearby transformer, containing 1 0(0 gallons of Pyranol ((lb Aroclor 125% and 35% chlorobenzenes), to crack and leak 180 gallons onto the floor. The Pyranol vapourized due to the Intense heat and contaminated soot and ash. The fire burned for 45 minutes until the power was shut off; during the *5 minutes temperatures at the centre of the fire reached two thousand degrees F. Firemen waited for power to be shut off before they tackled fire. Fire took 35-40 minutes (total exposure time for firemen) to extinguish. Contaminated soot was spread throughout all 13 storeys of the building. The drastic temperature change from the warm basement to the roof (-4*F) propelled the soot Into the fire stairwell and into an airshaft that ran through every floor. State fire investigator determined that soot rose from the basement through a door that was left ajar to prevent pipe freezing and then distributed through the rest of the building with the aid of a steady stream of air that was escaping through two roof-top HONS 017196 *3 scape hatches, above the {ire stairs, that automatically opened when the lire alarm sounded. ` Fire was quite small but very hot. After the power was shut down and firemen were informed that they were dealing with Pyranol, the fire was extinguished by ventilating the area and dousing flames with water. . Donned Scott Airpaks* and standard full enclosure protective gear (poly/canvat jackets and pants, rubber boots and gauntlet gloves) did utmost to ensue that no skin was exposed. The water used was not contained. All equipment and clothing were disposed of afterwards. All firemen took a shower immediately after fire (standard lire department procedure). One fireman developed a recurrent rash on hit lace. Two other developed "rashy bums" when the super hot fluid penetrated their gear. After Immediately after the fire it was known that ISO gal of PCB had spilled. Wipe tests of the toot were immediately taken and tested for other chemicals. A 2*-hour security system was maintained to control access to the buikfing. A trailer that served as an entry/exit module was set up at the basement loading entrance. The trailer contained entry facilities, locker, areas, showers, rest rooms and security offices. All personnel entering the building were required to wear a full-face respirator and protective clothing that comprised socks, underwear, sneakers and rubber boots, coveralls, an outer "Tyvek" protective suit, and both cotton and rubber gloves. Personnel were required to remove all protective equipment and shower on exiting the building through the module. Respirators were cleaned and filters replaced. The respirator weighed t lb and featured both activated carbon and high efficiency particulate filters. MONS 017197 Outer suits, gloves end respirator filters were disposed of after each use. A week after the fire, test results indicated that the soot was also contamina ted with PCDFs, PCODs, biphenylenes and naphthalenes with concentrations as high as 2.9 ppm of 2,3,7,8-TCDD and 12* ppm of 2,3,7,8-TCDF contained in the soot found between tlte floors. The horizontal surfaces of the building were covered with soot contaminated with 1(2 vg/m2 PCB. Ail surfaces contaminated by soot and smoke particles were cleaned initially by a high-elf Iciency vacuum with vacuum bags approved for use with asbestos and washed with water and an anionic detergent (Triton X-100). soot. Clean-up rags were disposed of as soon as they appeared slightly blackened by Exposed areas not easily cleaned (papers, carpets, drapes etc.) were put in plastic bags and stored in the sub-basement of the building. All water discharged from the building during cleaning was deposited into JJ- gallon barrels for treatment and disposal until a carbon filter treatment system, using three plastic above-ground swimming pools, was set up in the basement of the building. The contaminated water was recirculated through the carbon filtration system until it was sufficiently decontaminated (less than 1 ug/L) to send to the city's sanitary sewage treatment plant. An air pollution control filtration system was installed to filter the air In the building such that It was clean before discharge from the building. The system created a negative pressure throughout the building to ensure that air flowed from outside, through the building and linally through the filters on the roof. Preliminary cleaning has reduced the levels to I ug of removable PCB/m2 on glass and painted surfaces. Cleaning procedures were not successful decontaminating porous ceramic and vinyl floor surfaces, therefore they must be removed tor disposal, many people suspect that the floors could have been cleaned if the oily soot had not been left to penetrate Into the porous surfaces for upwards of ( months (during which clean-up and safety procedtres were being developed). HONS 017198 5 Clean-up personnel have been continually monitored with blood tests. No exposure symptoms other than chloracne and "rashes" have been docu mented. ` Blood tests were given on a voluntary basis to anyone who thought that they were exposed during and after the fire. Results of these tests were inconclusive and the tests are presently being redone. Preliminary clean-up required I* months for completion. The criterion for completion of preliminary clean-up was to remove all loose PCDF and PCDD contaminated soot from the building to allow opening of the building to normal ventilation) normal ventilation is required in order to use organic solvent based cleansers which would further decontaminate the building. All material that could not be decontaminated was drummed and landfilled at the SECOS secure landfill at Niagara Falls, New York. Final clean-up Is currently pending a decision by scientific authorities on how dean (i.e. how low do the levels of PCDF and PCDD contamination have to be brought) the building must be before it can be reopened. Final clean-up will probably include replacement of all vinyl and ceramic floor tiles as they could not be sufficiently cleaned by detergents and repainting the walls with chemical resistant paint. On 31 January 19S3 Governor Mario Cuomo of the State of New York stated that $L( million had been spent on the Binghamton clean-up thus far and that a further $9 million will be allocated this year to complete the final clean-up. HONS 017199 *6 UNIVERSITY OF MANITOBA FIRE - 29 March I9S2 Transformer lire in a transformer vault located In the basement of the electrical engineering laboratory of the University of Manitoba at Winnipeg. Transformers were filled with mineral oil contaminated with 250 ppm PCB. Bgfore A transformer vault contained six mineral oil transformers. Transformers were used for both experimental and operational purposes. No one was at the fire site at the time of the fire (early morning). During An electrical failure occurred in one of the transformers causing arcing and a fire. The mineral oil in the transformer caught fire and spread to some of the others (all six transformers were damaged by fire). Firemen attempted to extinguish the fire with water. Electricity must not have been shut off because the transformer exploded spraying oily soot all around the room. Fire was extinguished with water which was not contained. Firefighters store standard firefighting equipment with no specialized chemi cal gear as the transformers were supposed to contain only mineral oil. The entire laboratory was contaminated with PCB-containing soot. Alter The Manitoba Department of the Environment responded to the fire and tested the oil that was in the damaged transformers and sprayed on the walls to ensure that It was not contaminated with PCB. Test results indicated that the soot that had sprayed on the walls eras contaminated with 250 ppm PCBe. They tested the soot for PCDFs but found no detectable levels. Blood tests for PCB-exposed personnel were taken. Precautions for PCB contamination were taken in clean-up. HONS 017200 47 Two clean-op contractors were hired to dean the laboratory, one cleaned up the high voltage equipment with Keysolve tlie rest of the lab was scrubbed down with Varsol. ` A vacuum truck exhaust system was set up to remove Varsol fumes. Clean-up crew were protected with air-line respirators (because of Varsol), rubber gloves and boots, and disposable chemical impervious suits. NOTEs requirements lor protective equipment would have been more stringent if PCDFs had been Involved. All solid and liquid contaminated material that could not be decontaminated was drummed in metal drums and sent to Kinetic Contaminants for storage. HONS 017201 DANVKEN, SWEDEN - August I9SI A transformer station IS PC8 capacitors *8 fiSElQK Electrical malfunction caused a fire that burned IS capacitors. Firemen wore no special protective equipment against PCBs. After 1 - 3 vt/m? PCDFs in soot no PCDDs detected chlorinated pyrenes formed Clean-upi high efficiency vacuuming followed by detergent washing. Exposed personnel were tested by blood samples: no elevated PCB or PCDF levels were found. No incidents of exposure symptoms were reported. All contaminated wastes were drummed and stored, pending decision of the Swedish government on disposal. HONS 017202 *9 SKOVDE, SWEDEN March 19S2 A Volvo metal treatment factory. Fire broke out in a capacitor room containing mineral oil capacitors and 21 PCB capacitors. Purina An electrical malfunction in a mineral oil capacitor serving a high frequency oven in a casting line. A very hot fire ensued and spread to the capacitors: 12 were broken open, 9 remained sealed. The fire burned for two hours until it was completely extinguished. Temperatures were high enough to melt a copper pipe (1 100'C). Firemen were aware of the presence of PCBs and wore protective equipment accordingly which was disposed of afterwards. Fire extinguishing method unknown. Alter up to 900 ug/m2 PCDFs no PCDDs detected Clean-up: high efficiency vacuuming followed by detergent washing. Exposed personnel were tested by blood samples: no elevated PCB or PCDF level* were found. No incidents of exposure symptoms reported. PCDF contamination levels were lowered to <10 ug/m? after the ciean-up. All contaminated wastes were drummed and stored, pending decision by the Swedish government on disposal. HONS 017203 50 SURAHAMMAR, SWEDEN - September 23,19*2 A steel mill 500 capacitors involved (300 mineral oil, 200 PCB) An explosion in a steel kiln caused 10 tonnes ol molten steel (1 500*0 to spread through the steel mill. The molten steel melted a metal door sealing oil the capacitor room. A lire ensued in the 500 capacitors, 200 ol which contained PCB (ca. 2 000 kg) Firemen attempted to extinguish the lire resulting from the molten steel with water. Explosions resulted. They then decided to let the lire burn itseil out. The building was occupied at the time ol the lire but everyone was evacuated. The heat was very intense and large amounts ol HCt were given oil by the burning PCB. Alter up to * M g/m2 TCDF in soot (condenser room) no PCODs detected Clean-upt high etliciency vacuuming followed by detergent washing. The vacuum-up soot was treated by carbon filtration and the contaminated carbon was stored in steel drums for future disposal. A movable curtain wall was erected In the steel mill to prevent re-contamination ol areas already cleaned. Clean-up personnel wore disposable protective suits, rubber boots and gloves. They wore air stream, helmets (air was filtered by glass fibre filter before breathing). Exposed personnel were tested by blood samples: no elevated PCB or PCDF levels were found. Two incidents of skin problems resulting from exposure to HCl were reported. PCDF contamination levels were lowered to <10 ug/m2 PCDFs after clean up. All contaminated wastes were drummed and stored, pending decision by the Swedish government on disposal. HONS 017204 51 APPENDIX B REFERENCES HONS 017205 52 REFERENCES J. Environment Canada, Preliminary Socio-Economic Impact Analysis for Chloro- biphenyl Draft Regulations. Inventory and Attrition Rate, p. 5, M. M. Dilion Limited, April, 1981. 2. B. Sansson and C. Sundstroem, "Formation of Polychlorinated Dibenzoiurans (PCDF) During a Fire Accident in Capacitors Containing Polychlorinated Biphenyls (PCBP, Vol. St Chlorinated Dioxins and Related Compounds, p. 201, Pergamon Press (1982). 3. Malarek, Victor, "Workers Not Told Second Chemical Released by Fire". The Globe and Mail (3 February, 1978). -------------- 9. Jones, P.A., Environment Canada, Chlorophenols and Their impurities in the Canadian Environment. Report EPS 3-EC-81-2, March, 1981. Buaer, H-R., PCDDs and PCDFst Formation. Occurrence and Analysis of Environ mentally Hazardous Compounds. Doctoral Dissertation. University of Umea-Sweden U"7e)i 6. Vos, 3.G., J.H. Koeman, H.L. Van der Maas, M.C. Noever de Brauw, and R. H. de Vos, Food Cosmet. Toxic.. I, 623 (1970). 7. Moore, 3.A., 1ARC Meetins on PCDDs and PCDFs. Lyon. January 10-11, 1978. 8. Hryhorczuk, D.D., W.A. Withrow, C.S. Hesse and V.R. Beasley. Arch. Environ. Health. 36(5), 228 (1981). 9. Bauer, H., R.H. Schulz and U. Spiegeberg, Arch. Gewerbeoath. Gewerbehve.. 18.338 10. Kuratsune, M., T. Yoshimura, 3. Matsuzaka and A. Yamaguchl, Environ. Health Perscect.. 1, 119 (1972). 11. Buser. H-R.. H-P. Bosshardt. C. Raooe and R. Lindahl. Chemosohere. 7(3). 819 (1978). " 12. Reference 2, pp. 201-8. 13. Chlttlm, B., B. Clegg. S. Sale and O. Hutzinger, PCDFs and PCDDsi Detection and Quantitation in Electrical Equipment and their Formation Purina the Incinera tion of fCBs. Report to Environment Canada. DSS File No. OaSSJCEl09-8-6377 mwT 18. Lustenhouwer, 3.W.A., K. Olie and O. Hutzinger, Chemosohere. 9(7/8), 301 (1980). 13. Buser. H-R.. Chemosohere. 8(6). 913 (1979). 16. Buser, H-R., and C. Rappe, Chemosohere. 8(3), 137 (1979). 17. Choudhry, G.G. and O. Hutzinger, Toxicol. Environ, Chem,. 3, (3-8), 277 (1982). HONS 017206 53 II. Nagayama, 3., M. Kuratsune and Y. Masuda, Fukuoka ttaku Zasshi. 72(9), 136(1911) Chemical Abstracts 95(1 l)979S9h (1911). ~ 19. Buser, H-R., H-P. Bosshardt and C. Rappe, Chemosphere. 7(1), 109 (1978). 20. Nakagawa, 3., M. Morita, K. Akiyama, Y. Higuchi and S. Mimura, Tokyo-TorHsu Eisei Kenkyusho Vempo, 21(1), 260(1977). Chem. Abstracts 91(5)t39223d. 21. Morita, M., 3. Nakagawa and C. Rappe, Bull. Environ. Contain. Toxicol.. 19(6). 665 (1971). ----------------------------------------------- 22. Weber, 1., RTE Corporation, Wisconsin, Personal Communication, March, 1913. 23. American National Standard ANSI/ASTM D22I3-75, Standard Specification for Chlorinated Aromatic Hydrocarbons (Askarels) for Transformers. 29. American National Standard ANSI/ASTM D2233-79, Standard Specification for Chlorinated Aromatic Hydrocarbons (Askarels) for Capacitors. 23. Report on Lamp Ballasts Containing PCB. M.M. Dillon Limited, 31 March, 1913. 26. Personal Communication - Thomas Tieman, Brehm Laboratories, Wright State University (1* February, 1913). HONS 017207 6560-50 ENVIRONMENTAL PROTECTION AGENCY 140 CFR PART 7611 (OPTS 62035;TSH FRL ] POLYCHLORINATED BIPHENYLS (PCBs) MANUFACTURE, PROCESSING, DISTRIBUTION IN COMMERCE AND USE PROHIBITIONS: USE IN ELECTRICAL TRANSFORMERS REFERENCE DOCUMENT VII. (1) MOWS 017208 DSEAP, OPTS. BED. Exposure OPTS 62035 PCB (Fires) 0 File Reports DRAFT REPORT EXPOSURE ASSESSMENT: FIRES INVOLVING PCB TRANSFORMERS EPA Contract No. (8-01-627! Task No. IS - Prepared for: U.S. Environmental Protection Agency Exposure Evaluation Division Office of Toxic Substances ' 401 M Street, S.M. Washington, D.C. 20460 Prepared by: Versar Inc. (8SO Versar Center P.O. Box 1646 Springfield, Virginia 221S1 December 16, 19(3 Revised January 13, 19(4 HONS 017209 DUcllliaer This document Is i preliminary drift. It hat not been released forwlly by thi Offlet of Toxic Subttancct, Offle* of Pesticides and Toxic Substancac, U.S. Environmental Protaction Agency, and should not at this stage be construed to represent Agency policy. It Is being circulated for coamnts on Its technical earlt and policy Valuations MOMS 017210 TABLE Of COWTEHTS PQ 1. INTRODUCTION................................................................................................. 1 R. PYROLYSIS OP PCBS BY FIRE........................................................................ 2 3. SUHNARY Of FIRESINVOLVING PCS TRANSFORNERS................................. 3 . DISCUSSION..................................................................................................... 17 S. REFERENCES..................................... !.................. ..........................If J HONS 017211 i. irmooucTiON Electrical transforMrs containing KB-based collant liquids were Mnufactured In th Unltud SUtti from the early 1930's until 1978. tore than 100,000 such units My still ba In sarvlca. Although transforMrs filled with KBs wara nor* expensive than transforMrs filled with mineral oil, tha PCB units wara usad In thosa locations wharo tha flaxeaablllty of mineral oil would prasant a significant flra hazard to adjacent equIpMnt or structuras. In Mny applications, such as in underground alactrlcal vaults, tha decision as to whether tha transforMrs should contain KBs was ..a Mtter of policy which varied significantly aMng tha various alactrlcal utilities In the United States. Although nonaally operating KB transforMrs are totally enclosed units, there have bean a few docixeented Incidents wham physical damage to a transformer or alactrlcal failure within tha unit has resulted In tha loss of KBs with significant exposure to people or tha envIronMnt. Tha MSt serious consequence of physical daMge My have bean tha transforMr leak that occurred whan a transformer being stored for eventual must by fierce Backing Corporation In Billings, Montana, was struck by a truck and leaked Into a water recovery simp In a Mat packing plant (Anonymous 1979). Tha resulting contamination of anlMl feed resulted In a FDA recall of food products In 17 states, and led directly to the EPA ban on the use of KB transforMrs in locations where food or feed could be contMlnated (EPA 1982). KB transforMrs can also lose PCBs to the envIronMnt by venting through the safety valve. The transforMrs am equipped with pressure relief valves set at about IS psl which are designed to prevent rupture of the casing If the transforMr becoMS pressurized due to the generation of gat or vapors due to Internal arcing. However, the release of hot pressurized KB vapors can entrain considerable quantities of 1 HONS 017212 liquid PCBs at a fin* mist. Documented safety valv* releases fr*m transformers In Washington, O.C. (Lots and Irtyss* 1IS3) and Frankfurt, Bermany (Folx 19(3) damonstrat* that such Incidents can roltaso hundreds of pounds of PCBs and, depending on the wind conditions, distribute the aerosol for several hundred feet. However, the arcing conditions are usually of fairly short duration, and the resulting pyrolysis of PCBs does not result In the forawtlon of more than a few parts per million of pyrolysis products In the oil remaining In the transformer (Lees and Breysse 1((3, pp. 2D-3S.). Breater pyrolysis of PCBs has occurred In several recent Incidents where PCB coolant fluid has leaked from transformers Into an externally fueled fire. The formation of polychlorinated blphenylenes (PCBPs), dlbenaofurans (PCDFs), and dlbemo-p-dloxlns (PCDDs) and their wide dispersion In the smoke generated by 'the fires may present a ' significantly greater risk to human health and the environment than results from the previously Identified transformer failure mechanisms. 2. PYROLYSIS OF PCBS BY FIRE Laboratory experiments have demonstrated that PCBs are converted into polychlorinated dlbentofurans when heated for two weeks In the presence of oxygen at temperatures between 2?0*C and 330*C, but that the furans decompose at temperatures above 330*C (Horlta et al. 197B). Experimental pyrolysis of PCBs held at elevated temperatures for less than on* minute In the presence of air demonstrated that PCBs were completely destroyed at temperatures above 100*C, but that measurable amounts of PCDFs wore formed by exposure to temperatures as high as (S0*C (Buser et al. 197B). A comparison of the pyrolysis products from two capacitor fires In Sweden suggests that exposure of PCBs to high temperatures in the presence of oxygen can form PCDFs, but that exposure of PCBs to these temperatures in the absence of oxygen can lead to the formation of PCBPs and other chlorinated polynuclear aromatic chemicals (Rapp* etal. 1(82). PCBs 2 HONS 017213 apparently do not degrade to PCDOs, but PCDOs can be formed by the pyrolysis of the trl- and tetrachlorobemenes that are mixed with the PCBs In aany transforaers (Webber 1982). The chemical reactions leading to the generation of PCDDs and PCOFs are discussed In detail In a recent review article by Choudhry and Hutzlnger (1982). The forawtlon of planar chlorinated coapounds, such as PCDFs, PCDDs. and PCBPs, Is of concern following the pyrolysis of electrical fluids as soae of the congeners of these coapounds are auch more toxic than PCBs (Vucita 1983). An assessaent of the data from recent fires In the United States that Involved the loss of PCBs from transforaers Indicates that the foraatlon of these coapounds by pyrolysis of PCBs Is a codiaon occurrence that does not depend on a peculiar set of conditions, and that the noraal air convection caused by a fire aay spread these chemicals far beyond the confines of a transformer vault. ' ]. SUMMARY OF FIRES INVOLVING PCB TRANSFORMERS State Office Building. Blnohaaton. New Tort At 5:30 a.a. on February S, 1981, a fire occurred In the switchgear adjacent to a PCB transformer In the basement Mechanical room of the Binghamton State Office Building. This building Is an.18-story office tower that was completed In 1973 (OGS 1992). Power to the building was supplied by two transforaers which contained a coolant fluid consisting of BSX Aroclor 1254 and 355 Mixed trl- and tetrachlorlnated benzenes (Schecter 1983). Although the city fire department responded within alnutes, they did not enter the mechanical room until the power was disconnected to the transformers approximately fifty minutes after the fire started (Gelaann 1983). During this period, there was repeated electrical arcing and reports of loud explosions occurring In the mechanical room. Saoke generated by the fire was distributed by convection throughout the 3 HONS 017214 building through on open vortical chase that started In the Mechanical room and ran to the penthouse. This concrete block chase contained the sheet natal duct for thd exhaust air from the Men's restrooms on all the floors. The chase was not air tight, and allowed smoke to escape Into the space between the structural celling and the suspended celling on each floor of the building. As a result, the entire Inside of the building was coated with black soot. Air and wipe samples were taken one to two weeks after the fire occurred to define the levels of PCBs prior to cleaning. The results are simaaarlzed In Table 1. The probable source of the soot was combustion of the Materials in the switch gear. The heat of the fire apparently caused a ceramic bushing to crack on one of the transformers, allowing approximately 1M gallons of Insulating liquid to drain onto the floor In the vicinity of the fire. Photographs taken shortlyiafter the fire was extinguished showed that the switchgear was completely destroyed, but that there was little daawge to the transferors. The Mechanical roam was heavily coated with soot, and puddles of liquid remained on the floor. The heat from the fire was sufficient to slightly damage a structural steel beam that was directly over the twitch gear. However, the fire apparently did not spread to any Mterlal other than the switchgear Itself. Elevated levels of PCOfs In the soot were reported'within a week after the fire. Initial analyses of soot samples were performed by the Hew York State Department of Health, U.S. Department of fish and Hlldllfe, and the University of Umea. Each of these labs analysed for those constituents for which they had standards available. The results of these analyses are sunaarlted In Table 2. The Hew York State Department of Health subsequently did a more thorough Investigation ef the PCI and PCOF levels In soot taken from the top of the suspended celling In each of the floors of the building and of a large composite ; sample of soot that was collected by vacuum cleaner from the third and * e 4 HONS 017215 Ml* I. PCO (avals la Ml* 31*M**ta* SUM Offlo* Pal1*1a, Mm Ctamlif t)M* of M*l* Dal* of aaplinp lecatioa-floor lyp* of *arfo* * Air Dry *ip*** (ry irtpa** Dry *ip*** Ory nip*** Dry vipa** Dry *i|MH* *ry *ip* Dry vip*** Dry ip** 2/11-2/13, 1301 2/10,2/12,2/13 2/10,2/13 2/10.2/11,2/13 2/n 2/12,2/13 2/12 2/11-2/13 2/11-2/12 2/12 1,2.4,1,10.13,13,10 0,1.2 1.2 3,3.1.11 u 3.10.13 10 3,10,13 10,11 (.14 0p* arwa, korizoatal aurfaem Mai ar**s, mrtlcal aurfaem flpM m, korimUl mrfaoM 4pa* aram. mrtlcal aarfao* bcM 1 horlniUl MrfMH CncM rut,*** wrttal arfacti OovittfMMctHInf, tarimUl nrfon /bow MpaoM otlllof, wrilql sarficM HV roaaB, leaaara.'an* faa bladm 14 3 3 4 1 4 3 4 3 0 Ml air h*Ni *r* Mk1; all alpm ar* voto* at Araclar la*. Mi#* M*l*l (aka* ky vlkinf a** liaura *a(*r *i(k Pry lllt*r papar. **Iaat*a ctoaaO M an* Ml* cakiaat *wra. *P atama* to k* 1/2 of MtoctM IMt. lama: MM 1). Man,** Bnatric Wiii* 0.21-10.1 <.W148 0.3H4 1.4D 2.4S**** 4.14 SH3D 142.18 1.8 1.8 If.OMM 14.41 .42*14 4.42 400*14,000 IMS 1.3*3.4 2.42 3.4.020 131.83 HONS 017216 PblychlortuMd SIpMnylt WycMtrlMM DtMonfarm 2.3.7.1-Utri Tout Utr lout ponu 1oU1 Mu ToMI MpU Tout OCU PolycMortuUd 1IMM0 dtuliw 2.3.7.1-Utn 1oU1 Wtr* ToUl poou ToUl Mu Tout MpU Tout ecu PolycMortuUd MpMuylout 2,1,4,7-totrt ToUl Utr* Hatm libit 2. Capofttlon of Soot Savin Fm> tM ilnfhanton SUM Offlco tulldlnt - vV9 Mtt* Wli Mth IMIb lipp* 1M1 SUUInt 1W 273 124 41 S97 12 21 S. 470 MS SIS S12 101 ' 30 2.1 2.1 U2 l.t 0.4 1.2 S 4.7 7 2 5 M"PM till IMSnot wed 1* anlaal Indies itedlot 6 SO HONS 017217 fourth floor of the building and subsequently used for anlnal foodlng studios. The results of those analysts art summarized In Table 3. The variation In the concentration of PCDfs from floor to floor Is thought to be a function of the extent to which the various soot samples were contaailnated (and thereby diluted) with noraial building dirt that had accumulated on the ceilings, Animal feeding studies indicated that the homogenized Material from Floors 3 and 4 had an acute toxicity In guinea pigs equal to that expected for an Inert Material containing SB ppm 2.3.7.I-TCOD (Eadon at al. 1982). The building has been extensively cleaned by vacuunlng and washing to remove the soot. The effectiveness of the cleaning has been Monitored by wipe saMples taken both with dry filter paper and with hexane saturated game pads. The results of a study to coupare the effect of the coupling procedure on the Measured levels of PCBs on surfaces after completion of the prellMlnary cleanup are siamaarlzed in Table 4. a nunber of dry wipe sauples froM vinyl walls were analyzed for PCOOs and PCDfs. The results of these analyses are presented In Table S. Analysis of air sauples (Sulth et al. 1913a. 1983b), wipe sauples frou vinyl walls, and bulk sanples of sprayed-on celling Insulation (Versar New York Inc. 1993a, 1983b) have deMonstrated that the ratio of the various toxic constituents Is not constant frou one Matrix to another. The relative concentrations of PCDfs having various degrees of chlorination In different Media are summarized In figure 1. These data suggest that vaporization and redeposition processes are causing a gradual redistribution of the Material within the building. The Blnghanton State Office Building Is still closed to norual use awaiting the establlshnent of a reoccupancy criteria which will specify the residual levels of PCBs, PCDfs, and PCDOs that will be considered to be safe. The Most recent draft risk assessnent for re-entry to the Binghamton State Office Building would require that the air contain no 7 HONS 017218 HW* 1- CMcntratlon (apfp) if MytMoriaated of MyMrhaM Olkaamfanai la Soot Sa*ta* Idwn Fra tka Otatfiwtea State OFFtea Oaitdiap Floor S1 2 3 4 4 1 0 0 10 11 12 13 14 IS IS 17 . tst h* Kfc* 8 SKI* Vfctra CBF 0.2 taate OF <0.21 mu af <0.3* Hapta OF 4.H OcteOF <0.3* Total KOF - Natto Total PCPF cone. FCP ooac. 1300 M0 220 2000 M00 1000 niwff MOO 21000 11000 3400 930 10000 4200 0500 7200 5000 <** <0 <11* <12* <I4> n 220 2i m 13 100 1.2 02 <1.2 21 SI MO 23 144 10 X 100 14 8 32 3 3.1 14 o.s - 3 74 410 200 320 140 2SO 440 in 300 240 04 240 too O.S 35 2" 31 1200 400 1000 tf d d d d - .Ml .11.027 .042 .052 .042 .040 d 32 300 4 220 d 31 3M no* 200 d X 01* 140 d 5.0 n 8* 32 d 2.0 II 11h 15 - 07 no on 14S M0 145 120 X 70 31 20 11 4 430 310 .094 .015 .103 .044 .042 HONS 0 1 7 2 1 9 'QiatIUM Aroclor 12S4; data uncorracted for racmary. *Can<actiaa far raraary lacorpcratad. caot iapM bacaaaa aF dlnlallar oailtap eoaatractiaa. ^Oita not aiad kacaaw F poor raconry eF lateral 'Cl-Tatra OF standard (<I0I). ^Nte laat Oacaasa of taatraaatel aalFaactlaa; tar parpoaai of calculatloa of tote) FCOF, It I* ainad that kaaa COF l< 1/2 tka coasaatratloa of paate. . oaI lacted ky how claaaar aad oaa4 la aaloal taalooloiy ataatoa. Nfetacttan I lain ara par eaapaaar, not par cfclariaatioa Mtar. Vila aat eanactei For racovary. lastraaaat ailFaaclIaa caasad Ion oF date oa * 'Cl-Tatra OF raeoaary. loam: fadoa Ml. MN . KB (Mil la Ilia (la^Maa Mata orfte# krtlalaf ATtar Pralkairy Cltav Tjpa af wapla Bata of upHay locMioa-floor Dry alpa* MM alpa** My alpa MM alpa My alp* MM alpa 5/5/02 5/5/02 S/V02 5/5/02 5/5/02 5/5/02 9 9 * 4 9 Tjpa of arfia Viayl floor Vinyl floor olayl aall Vlayl aall Plotter Hill Plaatar aall n Onpa Oaaaatrlc data ao/M* a*** 12 3.2*10.0 0.45 12 214 12 *.*12.4 3.C4*** 12 3.0*240 44.3 3 1.4*1 aft I.S 3 1.2*10.4 4.2 My alpa MM alpa 5/5/02 5/5/02 9 9 Qaarry tlla floor Quarry tlla floor 3 14*21.4 3 WHIM 20.1 0 My alpa MM alpa 5/5/02 5/5/02 9 Mtadoa flats ItindM flan c <0.0*4.0 ft <0.0*2.7 1.1*** HONS 0 1 1 2 2 0 Or* nipt . 0.25 a1 arai alpa* altk try /IItar papar. *NM alpa - 0.25 a* araa alpa* alt* a cattaa pant pa4 aattad altk 0 at at kaiaat. "00 aaauaa* to ka 0.4. " laarca: Manor Maa <Nrk 1402. Ckaaical Mto 2. CMteBiaatlaa af Otay! Milt te a* telM ___ HQ floar State afftoa MMtef after IMIaf Dry Mpa S*lai Lacatloa lit* now Nmr SU. Oavlatlaa m ran 2.3.T.9-tetra Total tatra Total teU Total Mnu Total H*ta total Octa W*1 2.4 1.0 1.2 1.3 2.0 1.2 0.1 1.0 1.4 1.12 0.0 0.29 0.29 f.25 4.20 4.29 4.22 2.20 4.00 2.10 2.20 2.01 1.00 1.19 0.92 0.02 o.so 0.29 0.13 0.22 0.13 0.10 0.13 0.09 9.22 0.03 0.22 0.15 4.0D 0.03 9.00 2.15 3.22 3.22 4.20 2.10 0.20 3.4T 2.20 0.02 0.20 9.T2 0.95 0.90 0.29 0.10 0.39 0.10 2.01 3.19 2.10 0.22 0.22 0.10 0.22 1.04 1.23 1.04 0.29 0.10 ! 2.3.T.0-tatra Total tatra Total faate Total Hom Total k^ta Total octa O^ 10 op NO NO NO 0.90 NO ND ND NO NO ND 0.14 ND 0.59 ND ND ND NO ND ND m ND ND ND" 0.23 ND m ND ND ND ND 0.141 ND 0.49 m ND ND ND 0.10 0.02 * * * ** ** ** T722T0 SNOW teO - awanfi Oatactkm Iteit - 0.l *%* teD - awifi Oatacttoa IMt 0.1 agfe* 1901 190(1 99 CN190UMTI9N AM 0v mmt <tj ij %**?) 1 ,J i 8 1__ _ III 9(9911 OF CNL90HIATI99 Ctnni0 MS9UT19N COM! 99 PI190MATIM KO0tf or CNL90MAT19M Figure 1. Relative Abundance of Chlorodlbenzofuran Congeners In Various Media In the Blnghanton State Office Building Source: Cadon 1962; Smith 1983a, 1983b; Versar New York Inc. 1983a, 1983b. 11 HONS 017222 or* thin 0.4 pg/m* 2,3,7.8-TCOF ind tint surface contamination bt no or* than 2.5 ng/m* 2,3,7,8-TCOF (KlM 1983). 7h1s criteria assumes th*t th* ratio of 2,3,7,8-TCOF to other toxic chemicals ruealns th* sum t In th* toot and that th* 1*v*1 of contamination ranilns constant for 30 ytart following th* reopening of th* building. Th* draft risk *ittssn*nt It b*1ng reviewed by a panel of expert! and th* calculated re-entry criteria are subject to revision. On* Market Plait. Sen freneltco. California Shortly after 11 a.a. on Nay 15, 1883, a fir* started In th* sub-baseewnt transformer vault of th* On* Market Plaza office building at th* corner of Steuart end Mission streets In San Francisco. Th* "vault contained three transformers that wort filled with Aroclor 1242 coolant liquid. Heavy black saok* Issued fro^ th* sidewalk grating adjacent to the building for about three hours until the high voltage power to the transformer tras Interrupted by th* utility (Burks 1883). During this three-hour period, there ms considerable vibration and loud noises occurring In th* vault that Mr* described by reporters as explosions (Smith and Solffer 1883). No Information It available on th* size of th* transformers, but It ms reported that only on* transformer leaked, and that after th* fir* there ms 50 to 60 gallons of liquid In th* vault, with th* floor and Mils of th* vault caked with black soot and liquids dripping from th* vault colling (Bartlett and Solffer 1883). Th* liquid remaining In th* transformer contained 127 ng/g TCOFt and no detectable TCDOs (detection limit *7.4 ng/g) (Anonymous 1883a). Contributing to th* seriousness of th* fir* ms th* failure of a protector In th* vault (Htlnlek 1883). Smoke and soot from th* transformer vault contaminated th* adjacent switch gear room through th* bus duct, and small areas of th* adjacent parking garage and work shop areas through smII cracks In th* concrete block vault Mils. Some of th* heavy smoke Issuing from th* sldeMlk 12 HONS 017223 (rating was apparently pulled In through the street level air makeup louvers on Steuart Street end Into the ventilating fans that supply air through the sub-basement, basement, plaza level, and floors 2 through i of the Steuart Street office tower. Tests later Indicated that the contamination was limited to the second basement, the air handling system through floor S, and the exterior of the building. The upper floors of the Steuart Street Tower receive air from ventilating fans mounted In a penthouse above the 28th floor. Air samples taken In the vault starting a week after the fire measured 320 to 1SOO pg/m* PCSs, depending on the rate of ventilation through the vault. The switch gear room adjacent to the vault had air concentrations as high as 08 vg/m* PCSs before ventilation was started. Air levels In the sub-basement fan room (outside the fan housings) found a peak level of 2.7 vg/m* PCBs : ; (Anonymous 1B83b). The San Francisco Department of Public Health restricted access to any area having a level of PCBs In the air In excess of 1 vg/m* or surface contamination In excess of 100 vg PCBs/m* when sampled by wiping with a cloth wetted with octane. Although no samples collected In the offices on floors 2 through S were contaminated above these levels, the city requested that the ventilation system not be operated because of the possibility of spreading contamination to these clean areas; without ventilation, the floors could not be occupied. Analysis of the soot collected In the transformer vault and In the sub-basement adjacent to the wall of the transferor vault showed the pretence of PCBs, TCOFt, and TCOOt at shown in Table B. Only one sample was completely characterized for polychlorinated dlbenzofurans. This was a wipe sample taken August t from the floor of the HVAC fan room In the second basement that had the following amounts of CDFs (Anonymous igaib): . 13 MONS 017224 II --1 lW.of eaecle Soot toot table (. KO*. KOFf, and KMX In tablet frw One Harltt Plea* Pefore dewlap location Ootod Panel* PCSl wp/p total ICOFt up/p 2,3,73tCOf Up/p total tCOOl up/p 2,3,73tcoo up/p odjoeont to vmH letlde of vault -- 11,000 -- n.p 13.1 3.3 7.3 .324 .031 .400 - PpAa? * <*/* np/W? np/e? Wipe** Wipe** wipe** Wipe** wipe** wipe** fen reoo witch pear roe* - fan -- - witch pear roan - floor ZPth floor whH Hoot 7/17 free row top of fen 7.3 .001 - '- .016 , 1.03 1 7.3 <0.03* 2.3 7.1 .0046 .13 2.3 .066 2.4 UP/W? np/o3 ep/a3 .001* .017 - np/*3 -.017 - " - Air letlde vwlt -- no P.2 1.13 <.003* - Mr Hit floor - .1 .006 <.001* - - Mr vwlt Mr vwlt i/u 490 .w .15 -- - 1/17 MO 2.1 1.0 ... - Air snitch |iar 1/20-71 2.1 .07 .022 -- -- roao Air fee row /It1 .14 .011 .004 - - bone detected at detectlee ltarit Indicated. ' Wipe uncle collected vtlnp Rlcnilpe pacer netted with octane. Source: anonym 1103b. .. HONS 017225 Chemical Total TCOFs 2,3,7,8-TCOF Total ponta COFs Total hexa CDFs Total hopta COFs Total octa COFs Amount 13 ng/ng* 29 ng/m* 32 ng/m* 4 ng/m* 2.3 Itg/IM .3 ng/m* AfUttv abundance 1 .40 .30 .06 .03 <01 The avallable reports on the analytical results from One market Plata do not specify tho analytical procedures nor. In most cases, the detection limits. However, It Is apparent that the conversion of PCts to PCOFs occurred with about the same efficiency In both Blnghaarton and San Francisco. Reoccupancy of One Market Plaza depends on whether the cleaning efforts achieve the criteria established by the State of California (firavltz 1383) and the City.of San Francisco (Silverman 1983). The criteria established by the City are summarized in Table 1. first Rational Bank Building. Chicago. Illinois On September 38, 1983, a fire occurred In a bus bar between a transformer and the switchgear In a transformer vault under the plaza on the seam block as the First Rational Bank Building. Although power to the transformer was cut after about 10 minutes, smoke continued to Issue from the side walk grating for about 46 minutes. Inspection of the transformers after the fire was extinguished found that one of the four transformers In the vault had leaked approximately 16 gallons of coolant consisting of 65X Aroclor and 35X chlorinated benzenes (Simon 1983a, 1983b). Tho source of tho leak was a samll hole In the transformer casing that was perhaps caused by electrical arcing. Significant PCB contamination was limited to tho vault, the exhaust air ducts from the vault, and to the exterior surfaces of a small building that was adjacent to the transformer vent. The highest readings outside the vault and Its exhaust system were on a window near the vault (119,000 ug/m*) and on an Intake air well, grille (181,000 ug/m*) and window (96,000 ug/m*). The transformer vault was 15 HONS 01722b Tabla ?. Maoccupancy Criteria for On* Harlrat Plan anuM. oPFict sna All pcootmsfi* Kb WfKK PCOOa/PCCPt* PC*I IMNSFONCR VAULT All KOOa/KOFt* Kb hrftai KOOi/PCOFf* PCOS 10 Pf/e* 1 M/ta . 3 np/la1 100 ppM* 0 Pf/P* 1 pf/a* 24 itp/ta* 1000 pp/b* *0*fln*4 at Ik* tun of all PC00 and PCOP conpnri chlorinated at to* 2,1,1, and I petitIona. ieurco: lllmrea* no. ***- 16 HONS 017227 adjacent to on underground porting gorogc ond shored o vontllotlng system with the gorogo ond the smell adjacent building; none of the adjacent roes were heovlly contaminated. The toll First Notlonel Bonk Building wos obout one holf block froa the vault ond wos not contaalnated. Although one fiber gloss filter on on elr Inteke vent on the fourth floor of the First Notlonel Bonk wos * { found to contoln About >0 ug/g PCBs, the adjacent filters did not show f high levels. None of the aeosureaents of PCBs token In the building Indlceted unsefe levels of contealnatlon. However, the building wos evecueted when saoke wos teen cooing froa e window vent; the source of the saoke wos lotor Identified os the oxhoust froa on auxiliary diesel generator that started when power wos shut off at the transforaers. No Analyses were reported for PCOFs or PCDOs In the soot. lA;nuaber of wipe saaplas were collected in the First Notional Bonk building and froa the air supply intake near the transforaer, but no PCDFs or PCDOs i were found at a detection Halt of ZSO ng/A ft* (673 ng/a') for total PCDFs and PCODs by full spectrua NS scanning, and a detection Halt of froa 0.B to IS ng/ft* (8.6 to 160 ng/a*) 2,3,7,8-TCOO. The First National Bonk building was evacuated at the tlae of the fire, but reoccupied the following day. Cleaning was Halted to the transforaer vault, Its exhaust systea, and the exterior of the saoll building next to the vault. The criteria of cleanliness was established by Coasoonwealth Edison at 400 pg PCBs/ft* (4300 vg/a*) when i saapled by wiping with a filter paper wetted with hexane. 4. DISCUSSION The data that has been developed In the laboratory and froa the analysis of samples of coabustlon products froa various fires Involving PCBs Indicates that the formation of polychlorinated dlbentofurant can be expected In cases where PCBs are exposed to air and elevated temperatures. A fire involving a PCB transforaer can generate HONS 017228 considerable smoke and the damage to surrounding property depends on where the smoke goes. Elevated levels of PCBs have been found several hundred feet from the source of the fire; If the smoke gets Into a building through ventilating ducts, the elevated levels of PCBs and pyrolysis products will be found In the building. A review of transformer Incidents suggests several precautions that would limit the threat to human health and the environment resulting from the continued use of PCB transformers. First, the possibility of fire should be minimized by preventive maintenance of the various electrical components used In conjunction with the transformers (switchgear, breakers, cables, etc.) end the area around the transformers should be kept clear of flammable materials. Second, there should be a method or procedure for quickly disconnecting the power to the transformers If a fire should occur In the vault. The installation of circuit breakers on the high voltage side, perhaps activated by temperature. Infrared, or smoke alarms In the vault, would do much to reduce the duration of the fire and the generation of smoke and the installation of network protector on the secondary side would prevent current from being fed back Into a failed transformer. The apparent advantage of the protective equipment should, however. Be carefully balanced against the Increased likelihood of failure of a more complex Installation, bearing In mind that the failure of a protector contributed to the severity of the Incident In San Francisco and the failure of a primary switch filled with PCBs resulted In the contamination of an office building recently in Syracuse, New York (Heath 1BB3). Third, design features to seal off the vault from the adjacent buildings, and perhaps to seal off the normal vault ventilating ducts In the event of a fire, would decrease the dispersion of smoke and the resulting contamination. Fourth, since continued operation of the building HVAC system can contribute to the spread of smoko, tha HVAC controls could be tied to a fire or smoke sensor In the vault to allow for automatic shut down In the case of a fire. ' 18 HONS 01?i29 REFERENCES Anonyms. 1*7*. Cast study of contamination: OCR's in ftod. Scltnco Nows, Vol. 1H, No. 14 (Oct. R. lilt), p. 228. Anonymous. 1913a. Summary of OCR air sampling results - On* Market Plaza. August 15, 1983. Anonymous. 1983b. Summary of air sampling results for OCR's and TCDF's - On* Market Plaza. September 12, 1983. Bartlett R, Solffer 8. 1983. Cleanup after S.F. explosion. San Fransclsco Chronicle, Hay 17, 1983, p. 1. Burks J. 1983. Downtown dioxin danger. Firehouse, November 1983, pp. 71-72, 114. Ruser NR, Bosshardt H-O, Rapp* C. 1978. Foremtlon of polychlorinated dlbonzofurans (OCDFs) from tho pyrolysis of OCBs. Chemosphere, No. 1, pp. 109-119. ( Choudhry 88, Hutzlnger 0. 1982. Mechanistic aspects of tho thermal formation of halogonatad organic compounds Including polychlorinated dlbonzo-p-dloxlns. Toxicology and Environmental Chemistry, Vol. 5, No. 1. pp. 57-93. DOM. 1981. Background. (Attachment to agenda sent to the members ef the Expert Panel for tho Binghamton State Office Building prior to a meeting of the panel in Mow York City on April 3, 1981.) Albany, my: New York State Oepartawnt of Health. ladon 8, Aldous K, Frenkel 8, at *1. 1982. Comparisons of chemical and biological data on soot samples from the Binghamton State Office Building. Albany, NY: New York State Oepartment of Health. March 1982. EM. 1982. Polychlorinated biphenyls (PCBs) manufacturing, processing, distribution In commerce and use prohibitions; us* In electrical equipment. Federal Register, vol. 47, No. 185 (August 25. 1982), pp. 37342-37380. Folx R. 1983. Hochglftlges Clophen sprltzte aus den Trafos. Frankfurter Rundschau, Hay g, 1983, pp. 13-14. 6a1mann S. 1983. City crisis began after fir* ended. (Series of articles on tha fire In tho Binghamton State Office Building). Binghamton (N.V.) Sunday Press. Binghamton (N.Y.) Evening Press, Jan. 30 Fab. 5. 1983. 19 HONS 017230 $. REFERENCES (continued) Bravltz N, Fan A, Noutro RR. 1183. Intori* guideline! for acceptable exposure level! In office settings contaminated with PCB and KB combustion products. Berkeley, CA: California Oepartawnt of Health Servlcas. Novomber 1, 1*83. Heath T. 1BB3. Tests show KB danger unlikely at Chimes Building. The Post Standard (Syracuse, NT), Vol. 155. No. 18 (December 33, 11B3), p. A-1. Jansson B, SundstrBm B. 1182. Formation of polychlorinated dlboniofurans (PCOf) during a fire accident In capacitors containing polychlorinated biphenyls (KBs). In: Chlorinated dioxins and related compounds - lupact on the environment. 0. Hutzfnger, ed. New York: Pergamon Presi, pp. 201-308. Kin Wt, Hawley J. 1183. Revised risk assessment Binghamton state office Building. Draft Report submitted to the expert panel on the Blnglwmton State Office Building. Albany, NY: New York State Department of Health. Juno 7, 1183. loos PSJ, Broysso PN. 1183. Final report on industrial hygiene - assessment: KB exposure of 6SA switchgear operators. Report submitted to Accident and Fire Branch, Oeneral Services Administration. Baltimore. NO: Center for Occupational and Environmental Health, The Johns Hopkins University. Jan. 20, 1103. Helnlck N. 1103. 100 KBs leaks reported In downtown S.F. San Francisco Examiner, September 21, 1183, pp. Al, All. Horlta N, Nakagawa J, Rappe C. 1178. Polychlorinated dlbemofuran (KBF) formetlon from KB mixture by heat and oxygen. Bull. Environ*. Contam. Toxicol., vol. 11, pp. BBS-870. OBS. 1183. The Binghamton State Office Building clean-up: a progress report update. Albany, NY: New York State Office of Boneral Services. January 1183. Rappe. 1181. Rappe C, Herklund S, Bergqvlst P-A, et al. 1182. Polychlorinated dioxins (KOOs), dlbemofuran! (KOfs) and other polynuclear aromatics (KPNAs) formed during PCB fires. Chemlca Script*. Vol. 20, pp. SB-B1. 20 H HONS 017231 S. REFERENCES (continued) Nappe C, Harklund S, Oergqvlst P-A, t. H. 1J83. Polychlorinated dioxins, dlbenzofurans and ether polychlorinated polynutltar aroaatlcs foraed during InclMratlon and polychlorinated biphenyl Fires. In: Chlorinated dioxins and dlbentofurans In the total envlronaent. Boston, HA: Butterworth Publishers, pp. 99-124. Schecter A. 1IS3. Contaalnatlon of an office building In Blnghaaton, New York by PCBs, dioxins, furans and blphenylenes after an electrical panel and transferaar fire. Cheaosphere, Vol. 12, No. 4/S, pp. MS-MO. Sllveraan Nf. 1083. Letter to Nr. Nynn Oliver, Tlshaan Hast Building Nanagoaent Corporation (sic) frea Nervyn F. Sllveraan, Director of Health, City and County of San Francisco. Noveaber 11, 1003. . Slaon S. 1003a. Suaaary of October 3, 1003, aeetlng on clean-up ef PCB transforaer fire at First National Bank Building. ... Slaon S. 1003b. untitled ea aeae dated 12/13/03, with suaaary of stapling results frea First National Dank, Chicago. ^ Salth 0, Selffer B. 1003. Big PB t E explosion - S.F. htghrlsd shut. San Francisco Chronicle, Hay 10, 1003, p. 1. Salth ON, O'Keefe PM, Hllker OL. et al. 1001a. Analysis ef 2,3,7,0 tatrachlerobeniofuran and 2,3,1,0 tetrachlerodlbenio-p-dlextn In a seat staple frea a transforaer explosion In Blnghaapton [sic]. New York Albany, NY: New York State Departaent ef Health. Feb. 20, 1001. Salth OH, Hllker OL, O'Keefe PM, et. al. 1001b. Analysis ef a Blnghaaton soot staple for tetrachlorodlbenrofurans and tatrachlorodlbonzo-p-dloxlns. Albany, NY: New York State Bepartaent of Health. Oct. 1. 1001. Salth RH, O'Keefe PM, Hllker ON, et. al. 1002a. Analysis for 2,3,1,0-tetrachlorodlbentofuran and 2,3,7,0-tetrachleredlbenzo-p-dlexln In a soot saapla frea a transforaer explosion In Blnghaaton, New York. Cheaosphere. Vol. 11, No. 0. pp. IIS-120. Salth M. Hllker OB. O'Keefe PM. et. al. 1002b. Oeteralnatlao of polychlorinated dlbentofurans In soot staples froa a contaalnated office building. Albany, NY: New York State Departaent ef Health. Harch 1002. I Salth RH, Hllker 0, O'Keefe P, et al. 1003a. Oeteralnatlon of TCBfs and TCODs In air staples froa the sixteenth fleer of the Blnghaaton State Office Building. Albany, NY: New York State Oepartaent ef Health. Harch 10. 1003. 21 I HONS 017232 S. REFEREHCES (continued) . Smith RH. Xllkir D. O'Keefe P. at al. 1183b. Determination of tetra-hexa COfs and tetra-CODs In air samples from th* 11, 14, H and 11th fleers of the Blnghaarton State Office Building. Alban*, NY: New York State Oepartaent of Health. Hay 16, 1163. Stalling. 1181. * "1?, *I Vertar Hew Verb Inc. 1182. Investigation ef the contamination remaining In the llnghamten State Office Building following completion of the prellmlnery cleanup. Heport submitted to Hew York Office ef leneral Services. Springfield, VA: Verier Hew York Inc., October 20, 1182. Verier Hew York Inc. 1163a. Determination of toxic compounds in sprayed-on celling Insulation: sampling and analysis (PCBs, PCOOs, and PCOFs). Report submitted to Hew York State Office of Senerel Services. Springfield, VA: Versar Haw York Inc., July 18, 1183. Verier Hew York Inc. 1163b. Determl|iat1on of the ratios of various toxic chemicals on vinyl wall surfaces In the Binghamton State'Office Building. Report submitted to Hew York State Office of Senerel Services. Springfield, VA: Versar Hew York Inc., September 26, 1183. Vacate J, Harsh JR, Kennedy S, at el. 1183. State-of-the-art review: PCOBs and PCBFs In utility fluid. Palo Alto, CA: Electric Power Research Institute. Webber I. 1182. The degradation of polychlorinated benxenes In electrical equipment. Paper delivered at the IEEE/PES Conference, Hew York, HY. Feb. 1-S, 1162. - Negars B. 1183. Herkars returning to PCS skyscraper, sen Francisco Chronicle, Hay 26. 1183. p. 4. 22 HONS 017233