PB87-145785
Character! rat ion of PCB ( Pol ychl or i nat ed
Biphenyl ) Transformer/Capacitor Fluids
-------
• ¦ ¦ t, imptw I
nd
lltr nony 1
: hni a I Rc soup «¦ Inc
¦ ' "•<< I lit $• i , ' ui *• 'H
Roc ¦ ¦ • i ! I • Mi ' ind 1 '
; " I' ¦
B • »r \ »'» ' * ill
• ' 11 if. ¦- I-- hnolort'- 01 > i on
lit . Ef ; l »*«• >* 111 ••>!. R«-ji-• t .•'•~•r «'f-r
11•» • t'r ¦»' ' Oh»¦ » . H
HV'AR S AS: F «*• KR!'.'» If Sf Aft' II U\{WRAfOK*V
>H E 'H -'I ¦ EARCH AM) 011 i l.HPMfM
I. El-,,'Rf'"ME? : PRC T • c ' If»N hi EN' Y
• V1*I HI! • « ' (>-¦
-------
MOT ICE
...
It'-, •imri'ii
th'vlt 5W 1 «"J *
¦ : «¦ l ;
\>*,••!;»' v
yn , - -
i.'-v .ttut ii>i«j
... ,,,
. > * m ,t < • i
* 1 ¦>» J
i |« rojlu' « %
i loi
• r»t*
-------
FOREWORD
Today's rapidly developing and changing technologies and lndtisrfl.il
produces and practices frequently carry with :hem the increased generation of
solid and hazardous wastes. These materials, If improperly dealt with, cui
threaten both public health and the erwiroruaent. Abandoned waste sites ,ird
.icc IdentaI releases of toxic and hazardous substances to the environment also
have important environmental and public health implications. The Hazardous
Waste Engineering Research Labo*atory assists in providing an authoritative .ir.d
defensible engineering basis for assessing and solving these problems Its
products support the policies, programs, and regulations of the Environmental
Protection Agency, the permitting and other responsibilities of State and local
governments and the needs of both large and small business in handling their
wastes responsibly and economically.
This report describes the occurrence of polychlorinated dlbenzo-p-dloxlns
iPCDDs) and polychlofInated d 1 b«?nzo f urans (PCDFs ) in residues following
combustion or pvrolysls of dielectric fluids containing polvchlorinated
b i phenyl s (PCBs). Data from laboratory experiments and accidental tires
involving electrical equipment have been evaluated to determine relative
tendencies for different dielectric fluid compositions to produce PCDOs and
PCDFs under tombustlon/pyrolysis conditions The report will be useful to the
electric utility industry, the owners and Insurers of electrical equipment
containing PCBs, emergency response personnel and organizations responsible tor
cleaning or disposing of contaminated material folloving a fire. For fur:her
information, please contact the Alternative Technologies Division of rhe
H.isa^ous Haste Engineering Research Laboratory.
Thomas R, Hausar, Director
Hazardous Watte Engineering Research Laboratory
111
-------
ABSTRACT
Dielectric fluids tn cvansforme rs and capacitors often contain
poIvchIo r i nated biphenyls i PCBs \ or ch lorobenzenes These substances may
generate po1ychIorinat«d dlhenzofurans (PCDFs) and polychlorinated
d L benzo-p • dioxins (PCDDs} under certain conditions of combusrion pvrolvsl*
W1i«*n #iectrlc«l equipment containing these fluids is Involved in an accicirnt.il
11 r»». the resulting smoke. soot, and residues may be contaminated with P''DDs,
PCDFs. and other chlorinated hydrocarbons.
This report contains a review of several laboratory studies investigating
the sources of PCDDs and PCDFs as well as the conditions under which rhey .ire
formed. In addition, some data ftorn sites of actual fire incidents are
avallabl® and are discussed, Chlorobenienes and PCBs do not form PCDDs and
PCDFs when heated in the absence of oxygen. During fires, the dielectric fluid
of trans formers or capacitors may be leaked or v«ne®d front ruptured casings.
* i th exposure to oxygen, PCBs can produce PCDFs and chloiobenzenes can produce
PCDDs. The particular isomers of PCDDs and PCDFs formed are related to the
number of chlorine substituents in the reacting material.
This report was prepared in fulfillment of Contract No. 68-03¦32 59 by
Technical Resources. Inc., Rockvllle, Maryland, under the sponsorship of the
U.S. Environmental Protection Age ncy, This report covers a period from
September 1983 to April 1986 and was completed as of May 1^86
lv
-------
CONTENTS
Notice . . . .... i i
Foreword . , . . : i i
Abac race iv
Figures . . . ... vi
Tables . >.'ti
LisC of Abbreviations . , . , , ... s
I, Introduce Ion I
2. Summary and Conclusions ...............
) Fire Incidents Involving PCB Transformers and Capacitors
Contaminat i an Cleanup Criteria
•*. PCDFs, PCDDs. and PCBPs from the Pvrolysi* of PCB Fluids
Formation of PCDFs from PCBs ,
Formation of PCDFs from Polvchlorlnated Benzenes j?
Formation of PCDFs from ChlorophenoIs . rA
Formation of PCDDs from PCBs v>
Formation of PCDDs from Polvchlorinattd Benzenes ~><)
Formation of PCDDs from Ch loropheno Is ?>¦«
j. Contamination Situations Involving PCDDs froa PCBs and
Chlorobenzenes ................... 6*
PCDD Formation from i.2,4-Trtch'orobenzene Feedstock
from Lindane Production , ............... ft *
PCDD Formation froa a Pittsbur, v**«» Metal Reclaiming
Facility . hi
PCDD and PCDF Formation from Strandley Metal Recovery
Site, Purdy, UA 63
6- Analysis and Evaluation .............. 7?
Comparison of a PCB Transformer Fire with the
Incineration Situation 77
PCDF and PCDD Levels in PCB Transformers Under Normal
Use Conditions Si-
Electrical Arc it,$ and P'.iD and PCDF Levels .... $$
Correlation of Combustion/Pyrolysts Produces Generated
and Constituents of Transformer Fluids 112
References . !«V*»
v
-------
FIGURES
&inber I'Jfc
1 Relationship Between the Vapor Pressures of the Solid
and Liquid Phases at Temperatures T. the Melting
Point (Tra) , and the Boiling Point (T} , . HO
2 Weight Percentage Yield of PCDFa us a Function
of Oxygen Concentration , . . •
J Weight Percentage Yield of PCDFs as a Function of Temperature "1
-» Concentration of PCDFs Formed as a function of PCB
Concenrratlon ...... '«!
5 Concentration of PCDFs Formed in PCB-Spiked Mineral
Oil by Homo log
6 Concentration of PCDFs Formed as a Function of Temperarure '¦> 3
7 Concentration of TCDD Formed as a Function of Temperature %
8 Concentration of PCDFs Formed as a Function of Temperature 9 7
9 Concentration of PCDFs Formed from Pyroly*is of Aroclor 1254 as
a Function of PCB Concentration . . « :•
10 Concentration of PCDFs Fornwd from Combustion of Aroclor I .'Si
as a Function of PCB Concentration "<•¦1
11 Concentration of PCDFs Formed in PCB-Spiked Mineral Oil
by Howolog *,*
15 Concentration of PCDDs Formed In Chlorobenzene-Spiked
Fluid by Homo log ........... !«»s*
vl
-------
TABLES
%TMBER LiLE
1 Fir* Incidents Involving PC8 Transformers or O.ipac i tors
Since 1978 . . . I
2 Concentrations of PCBs, PCDDs, and PCDFs Found at Binfthamtori
State Office Building Prior to Cleanup
I Levels of PCDFs and PCDDs in Soot from Accidental Burning
of PCB-Containing Electrical Equipment at Blnghamton 1
~ PCBs, PCDFs, and PCDDs in the Askare 1 After the Chicago
Fire . . . I 4
j PCBs. PCDFs, and PCDDs in Wipe Samples from the Interior of
the Transformer Vault from Chicago Fire . . Analysis of PCBs in Air Samples f row School in Cine inn.it i .
Ohio . .
Summary of PCS Wipe Sample Results from School In Cincinnati.
Ohio*
S Levels of PCDFs and °CDDs from the Accidental Burning of PCB-
Containing Electrical Equipment at Boston, Nassachusefts . . . J.
9 PCB Residues from a Transformer Vault Fire In Miami, Florida j
10 Results of Analyses of PCDDs and PCDFs in Sulk Samples of
Residue from Miami Transformer Fire
11 Analysis of Wipe and Soot Samples from Columbus. Ohio Fire .'8
13 Analysis of PCBs in Wipe Samples from the Hill -Murrav School
II Area Concentrations of PCBs and Chlorinated Benzenes at the
Hi 11 -Murray School . . . •<<
1-* PCB and PCDF Levels in Capacitors and Envt ronmental Samples
at Norrralje. Sweden .... ;
15 Concentrations of PCDFs Found in Samples from an Accidental
PCB Fire in Stockholm. Sweden ; <
16 l»eve>is of PCDFs in Wipe Samples fr * e Skovde Fire ' i
1 * PCDFs Found in Ulpe Tests at Surahamm... , Sweden ;•<
18 Levels of PCDFs ind PCDDs in Soot Samples from Imatr«,
Finland Fire I*
19 Levels of PCDFs in Wipe Samples from Ha 11s t ahammar, Sweden
Fire '
. ti
-------
TABLES (Continued)
mmm me
20 Levels of PCDFs. in Railway Locomotive in Sweden 38
21 Levels of PCDFs in Wipe Samples from Kisa, Sweden Explosion )8
22 Levels of PCDFs and PCDDs In Soot Samples from Reims, France-
Explosi on ............ <•>
2 3 PCDf Levels in Other PCB fires .]
2-« C .itamination Cleanup Criteria »2
25 Pvrolvsis of Aroclor 1254 In Quartz Mini-Ampoules ...... o
26 Therraochem cal Conversion of PCBs to PCDFs . .S
2 7 PCDFs Formed i» Combustion Studies 51
28 Relative Amounts of PCDFs Formed from PCB (At oclor 1254)
Pvrolsis and Carbon Filter 'rem a Capacitor Fire >2
29 Concentrations of PCDFs Formed from the Burning of
Askarel ?
10 PCDF Concentration Be fore and After Arcing of Aroclor
12 54/Tricnlorobenzene Mixture In Oil . . . .
11 PCDF Concentration Before and After Arc iof Aroclor 1260
In Oil : .
J2 Identification of PCDF and PCDD Iloners in the Combined
Chlorobenzene Pvroiyzate . v,
13 Format ion of PCDFs from the PvrolvsIs of Chlorobenzenes i*
1* Summary of Formation of PCDFs from the Pvrolysis of
Chlorobenzenes . . ....... '• *
i Levels of PCDFs from Burning of Pentachlorophenol -
Contami nated Waste .
lf> Formation of PCDDs from the Pvrolvs Is of Chlorobenzenes 1
K Summary of Formation of PCDDs from the Pvrolvsis of
Chlorobenzenes . * ;
18 PCDDs Formed in Combustion Studies
)9 Relative Amounts of PCDDs Formed from Analysis of a PCB/
Chlorobenzene Mixture • 1
'.0 PCDD Concentration Before and After Arcing of
trlchlorobenzene in Oil .......
vlii
-------
TABLES (Continued)
mmm
-* l Concent rat ions of 2 . 1. ?. 8 • TCDD Formed fro® thi» Burning
of Askarel T^ ....... .
- 2 Levels of PCDDs from Burning of Pentichlorophenol-
Contaminated Waste . . . . .
<3 Dloxln Isomers from Pyrolyzed HCCH Residues and
Tr lchlorober>zene St i 1 lbottoms . . 's <
-¦* Dloxln and Furan in Dust from Pittsburgh Area Metal
Reclaiming Facility 6«.»
•5 Summary of Sample Results from Strandley Scrap
Metal Site ................. "1
46 Summary of Sample Results from, Strandley Site and
Bur ley Lagoon *»
'* 1 PCDD/PCDF Analysis of Soli Samples from Strandley Site 'i
<8 PCDD/PCDF Analysis of Various Samples from Strandley Site ?«
49 Vapor Pressure Correlations of Dibonzodiox ins . Dibenzofur.ms .
Xanthenes, and Anthracenes ,•««
M) Average Total Congener Class Concentrations of PCDFs in
Four In-Service Dielectric Fluids . "*<>
51 Analysis of Fluids Involved in Transformer Fire Incidents 8"
52 Levels oi PCDFs in Commercial PCBs S •
31 Maximum Weight Percentage Yield of PCDFs as a Function of
Reaction Atmosphere ...... ,
Amount of PCDFs Formed from Pyrolysis of Aroclor 1254 »
•> 5 Percentage of PCDFs Formed from Combustion of Aroclor 1->•¦~ 1 »
ih Correlation of Analytical Data on Transformer Oil and the
Generated Soot from the Blnghamton. KY Incident ;
Ix
-------
LIST or ABBREVIATIONS
r.BR , lATIONS
A .a-II' -- American Conference of Covernmental Industrial Hvgiera-t*,
bp -• boiling point
C •' centigrade
CDD -- chlorodibenzodiox in
CDF - - chlorodibenzofuran
cm -- cein ineter(s)
cm- - squart centimeter(s)
<-! • - day (s)
fH !>D -- dichlorodibenzodiox In
Di DF •• dichlorodibenzofuran
D«M! -- Department of Health
ORE •• destruction and removal ef tie iencv
EPA -- Environmental Protection Agencv
F.PRI •• Electric Power Research Institute
F -- Fahrenheit
ft - foot or feet
g -• gram(s)
gal -• gallon(s)
Cf! HRMS -- gas chromatograph/high resolution mass spec t romei «• r
i MS •* gas chromatographymass spectrometer
Hi 1 H -• hesacf >roc vc lohexar.e
Hp 'BP - heptai lorobiphenylene
HpCDD • • heptac'ilorodibenzodiox In
HoCDF •• hcptachlorod Ibenzofuran
hr - - hour(s)
HRCf/MS -- high resolution gas chromatograph/mass spectrometer
HxCBP -- hexachlorobiphenylene
HxCDO •- htxachlorodlbenzodiox in
HxCDF - - hexachlorodibenzofuran
IAFF -- International Association of Pir«* Fighters
kg •• kilogram(s)
L - - I iteres)
1 (» - • pound
m *• meter(s)
-• square meter(s)
m* * cubic raeter{s)
x
-------
LIST OF ABBF IATIONS (Continued)
•3BREVIATIONS
meg -• mlcrogram(s)
ncI • mlcrolIter(s)
ml * - nii1.liliter(s)
m.p •- melting point
Mono-CDD -- raonochlorodlbenzodloxln
Mow CDF -• monochlorodibenzofuvan
MR* -• Midwest Research Institute
%\F1, • • nonaqueous phase le«chat«
% "J • • nor detected or none detected
ng -* nanogram(s)
N.OSH »• National Institute for Occupational Sat'otv and Health
OCDD octachlorodlbenzodloxIn
OCDF - oetachlorodlbenzofuran
PCi •• polychlor inated biphenyl
PCBP •• polvchiorinated biphenylene
PCDD -- polychlorinatcd dibenzo .loxln
PCDF -- polychlorInated dibenzofuran
OPF. -• polychlerinated diphenyI ether
P •- pentachiorophenol
f PY •• polychlorinated pvrene
tVi DO • pentachlorodibenzodlox in
Pe("0F • • pentachlorodlbenzofuran
pg -- plcogram(s)
PC&E - Pacific Gas 6 Electric Company
pph •• parts per bill ion
ppm -• parts per million
PVC ¦- polyvinyl chloride
TCDD • tetrachlorodibenzodioxin
TCDF -• tetrachlorodlbarTofu-an
Tr i • ("DD *• tr ichlorodibtrizoilioxin
Trt-CDF -• trichlorodlbenzofuran
TSCn - Toxic iubs.: ces Control Act
TV A *¦ t lac - we ight> •. average
wr - - weight
xl
-------
SECTION i
REGULATIONS PERTAINING TO FCBt. PCDDs. AND PCDFs
In August 1982, EPA decided t a permit the continued use of elect t
: r.insfoi'mers containing pol vchlor inated b i phenv 1 s (PCBs) based on t h«> repor: ,«t
low frequency of leaks and spills of PCBs from rhis equipment relative* : <> : in-
High costs of replacing or ? curing these transformers I'l-.U-r
Section 6(e)(2)(B) of the Toxic Subsranees Control Act (TSCA) . r'.PA *
authorize a use of PCBs provided that the use "will not present .in siih.msw.
risk of injury to health or i In- o 11 v 1 rnnrrii-iit " EPA dt» t *> rm i iu*«S iti.i: '
,-ont ituipd use of PCBs 'ottr.imin.it >h! t rails f oi me rs i >0- WO pp:n Pi'BsK two.- i'«
; ransfotmers t< ^0 pp« PCSs* do not present ..nreasonable risks :o public
,md the environment ' ^
A closer evaluation of the f i re * re la ted risks posed bv the continued us,-
of PCB transformers, and the costs and benefits of actions c.es igned to inhiiv
t!mse risks followed the 1982 deterrnln.it Ion EPA issued a Proposed Rnl«- .v.
October U, 1984, concerning PCB transformer fir*s. EPA determined ?»•.., t i ;r.-s
involving transformers containing > 300 ppm PCBs present risks to human h« »; *.•
and the environment,1'-^ The extreme toxicity of materials which can h.« i..r m •:
during fires involving PCB transformers, and the potential for human .tr<;
. n v 1roninent n I exposures to these compounds, contributed lo EPA's proposed ml.
EPA also determined that additional restrictions on the use nt !' B
! tans f orme r s to reduce risk of injury to health and the environment v.-iv
t '.ranted Therefore, EPA proposed additional regulatory controls1- ' «ns *!; t-
¦isf of this equipment, requiring:
• The immediate registration of all PCB t rans formers with apptvpr;
fire department jurisdictions, and the immediate revise rat i.»u
building owners of all PCB transformers located in or near bus Id-,r,•
• The immediate marking of the exterior of the vault door, mach! n«* t->- t .•=«(!•
door. means of egress, or grate (s) accessing a PCB transformer r< K
identification labels
• The Immediate removal of stored combustibles from PCB t r.m^ I «n r< r
I orat ions
I
-------
• The Installation, by Julv 1, 1988, of additional electrical protective
devices on FCB transformers In or near buildings in high secondary
volcag® systems (430/27? vole systems).
• The Isolation, bv JuIv I. 1988, of all PCB transformers in or n«*a *
buildings from building ventilation systems, building ductwork, ,»r.d
openings i n construction to reduce the widespread eonrammar ion of
s c aictui'i's and the environment bv smoke and soot in the event of ,i !*< H
r ran; forme r fire
In order to facilitate monitoring comp I iance, EPA proposed ; h,.: P "H
transformer owners maintain records of their efforts in isolating transfornvr-i
through the completion of PCB Smoke Spread Reduction Plans (PCB-SSRPs) EPA
also proposed to require PCB transformer owners to take expedient measures, :o
contain potential water discharges, and to report all PCB transformer fire
incidents to the National Response Center before Initiating cleanup efforts
On consideration of the extensive comments received during the public
t iiwrt period for the Proposed Rule, EPA modified the Final Rule concerning:
• Evaluation of the use of PCB c vans formers in or near induce r i »;
buildings separately from the use of" PCB 11 ansformers In »•;- .r
commercial buildings
• Relative probabilities of failures and fires in different types "I t'«"B
transformers installations, ^Lacing more stringent controls oi. : hose
transformers which EPA believes pose higher risks of failures and
fires,
• Increased emphasis on the prevention of PCB transformer fI-es through
increased electrical protection, and decrease n has Is on the use of
Isolation measures to minimize the spread c i ready formed and or
released contamina .ts.
On July 9, 1985. FPA promulgated its Final Rule on PCBs in «• )**¦•? i ie.< 1
transformers, the culmination of a long fact-finding and rule - mak i ,tg p; •>« «*.»
that began shortly after the transformer fire at the State Office Bui I.ling in
" inghamton. NY. February 1981 This Final Rule states that PCBs ,»r anv
concentration may be used in transformers (other than ir, railroad Ioccmotives
and self-propelled railroad cars) subject to the following conditions
1) The use of higher secondary voltage (> -»80 vnltsl tie t wo r I i'-"B
transformers in or near commercial buildings after October I I is
proh lbi ted
2
-------
2) The installation of enhanced electrical protection on lower secondary
voltage network PCB transformers and higher secondary voltage radial
PCB transformers in us# in >r near comae rc Lai buildings is required bv
October 1, 1990
3) Further installation of PCB transformers in or near omiiercial
buildings is prohibited after October 1. 198b
4) The registration of all PCB transformers with fire response personnel
and building owners is required by December 1. 1985
5) The narking of the exterior of all PCB transformer locations i s
required by December I, 1985
6) The removal of stored combustibles located near PCB t rans f oi mer.% i •»
required by December 1, 1985.
The Final Rule also requires :
-------
EPA has amended regulations for hazardous waste management under : i,<-
Resource Conservation and Recovery Act iRCRA) . by listing .is arun-lv h,»::a
wastes certain wastes containing particular polvchlor inated d i ben::o - p • <1 i *>\i !•-.
v PCDDs ) . polychlor inated dibenzofurans (PCDFs), and chlorinated phono I. and I» v
spec I fv ing management standards for these wastes . ^ ' The rule, promu 1 r id
January 14, 1985, defines certain waste streams as dioxins-contaminated wastes
and thus, narrows waste treatment options to those approved for dioxins waste.
EPA revoked the regulation (known as the "Vertac Rule") concerning the dispos.»l
of .'.3.7.8- tetrachlorodibenzo-p -dloxln (2,3,7, 8-TCDD) contaminated wastes ur.dc r
TSCA when the regulation under RCRA became effective on July 15, 1985
This regulation proposes a degree of hazard approach for the dioxins
contaminated wastes which prohibits their management at most types of in: «• r iir
status fac 1111les Land disposal of these wastes must be conducted pursuant to
additional special standards implemented during Che course of the permit
proceed! ng. Since all of these wastes are specifically Identified as
candidates for being banned from land disposal In two years under the Hazardous
and Solid Waste Amendments (see RCRA amended Section 3006 (c)) . EPA has
proposed an Interim regulatory regime for their land disposal
The only interim status facilities that are permitted to accept dioxins-
contaminated wastes are:
• Impoundments holding wastewater treatment sludges that, are created in
these impoundments as part of the plant's wastewater treatment svstetr.
• Waste piles that meet the requirements of Section 264 250 (c) referred
to as "enclosed waste piles."
• Tanks, containers, and certified Incinerators.
• Certified thermal treatment units subject to regulation under Subpart S*
of Part 265.
EPA mav also issue emergency permits to facilities to store these v.is: %
in situations where there is no other realistically available m.ui,«»;>-»« i«;
e,ipac 1 t v .
4
-------
A waste management plan is required by EPA to ensure, as much as possible,
chat these wastes aie properly managed in » land disposal sicu.it ion The pl-.i.
will be the interim vehicle for assuring individualized consideration that th»-
wastes will be managed safely. The waste management plan should address waste
volume, concentrations of CDDs and CDFs in the waste aerosol/pitt iculace
dispersion, volatilization of the toxicants of concern, soil accertuac ion
properties, waste leaching potential, and anticipated solvent co-disposal
The Fi nal Rule prohibits combustion of diox ins and furans contain in.) ted
wastes in incinerators that have only interim status Interim semis
incinerators can burn these wastes without obtaining a RCRA permit if thev ,ne
certified by the Assistant Administrator for Solid Waste and Emergency Response
as satisfying RCRA standards for incinerators burning these wastes
The hazardous waste regulations impose a more stringent destruction or
removal efficiency (DRE) on incineration of CDDs and CDFs, The new DRE
0 9 9 9 9 9%, is technically feasible and has been measured at numerous
incInerators.
This regulation defines RCRA ecute hazardous wastes and subjects them ra
special standards when land disposed, incinerated, or stored. One part of the
rule is intended to cover * bro:id variety of diox ins contaminated wastes rh.it
are identified as EPA impiemenes its Dioxin Strategy. In the implementation «>t
this strategy, the Agency has established seven categories or t iota ot
investigation. Tier 4 covers: combustion sources such as municipal and
hazardous waste incinerators, PCE transformer/capacitor fires, reactivation
furnaces for spent granular activated carbon, boilers using PCBs and
PCP-treated wood, etc ^ ^ Thus, this rule eventually may cover rex idur
materials generated in a PCBs transformer fire if dloxins-contamln.it ion is
determined, or. In situations where dioxins are found with "PCBs waste " \
case in point is. the problem encountered by a waste management companv us
Alabama. It accepted 40,000 gallons of "PCBs waste" from the Hvde Park
Landfill in Niagara Falls, NY. Since the waste contained > f>0<> ppm .-ft.,,
this waste was legally a PCBs waste. Unfortunately. Alabama State autismi;:«
were not informed via the forwarder's bill of lading that the nonaqueous
leachate (NAPL) contained 20.2 ppm 2.3.7,8 • TCDD along with hundred* ot r
5
-------
halogenated organic arid pesticidal residues and substantiil quant if ws i l't<>
1,000 ppm) of coxic heavy me t a 1s As a result, several of the 41 storage tanks
,»t the waste management company's tank far# became contaminated with dloxin and
1.2 million gallons of PCB«. wastes are now termed "dioxins wastes." with no
short-term solution in sight.
U.S. regulations have focused on PCDDs and PCBs To date, no Federal
regulation has been promulgated specifically for PCDFs However, if PCDDs or
PCDFs are founa In PCBs, disposal options for PCBs could be further restricted
6
-------
SECTION 2
SUMMARY AND CONCLUSIONS
An estimated />'•. 000 tons of PCBs are still in use in U.S. transformers and
capacitors, On July 9, 1985, Che EPA promulgated its Final Rule en PCBs in
electrical transformers.^**^ This rule specifies that PCBs at anv concetti rat Ion
may be used in transformers (other than in railroad locomotives and s» i f-
propelled railroad cars) subject to the following conditions:
1) the use of h ighe r secondary voltage (> 480 volts) network PC P,
transformers in or near commercial buildings after October 1, 1**90, is
prohibited,
2) the installation of enhanced electrical protection on lower secondarv
voltage network PCB transformers and higher secondary voltage radi.il
PCB transformers In use in or near comaercial buildings Is required hv
October I, 1990,
3) further installation of PCB transformers in or near commercial
buildings is prohibited after October I, 1985,
4) the registration of all PCB transformers with f Ire - response per•soni -.«¦•;
and building owners is required by December 1, 1985
5) the markings of the exterior of all PCB transformer location4, is
required by December I, 1985,
6) the removal of stored combustibles located near PCB transformers ts
required by December I, 1985
Tht
-------
The July 9 . 1985 rule provides for a gradual phaseout of some PCB
transformers while recognizing the potential for additional fire accidents in
the interim Moreover, PCB capacitors are not cove red under the July 9, 1985
i ult and, since they have also been Involved in fire incidents, capacitors are
potential release sources for PCHs and combustion by-products into the
envi ronment,
This report presents and evaluates the available literature and published
data on analyses of transformer fluids and soot generated In fires F.ven
though there have been at least 10 reported incidents, a wide varietv of
problems that hinder the analysis and evaluation of the data remain, including:
• Limitation of Analytical Data, V«. y few analytical data have been
generated for each PCB fire incident. Because of the high cost of
Isomer analysis and the large number of Isomers that characterize the
PCBs. PCDFs. PCDDs. and PCBPs, few analyses are actually performed for
any specific Isomer In the aftermath of a PCB fire Incident
• Differences in Saniollag and Analytical Protocols ¦ Sampling and
analysis protocols for contaminants generated in transformer fires are
not yec fully standardized, and thus a wide variety of sampling and
analytical methodologies are often employed Some data are based on
the analysis of sooc and are reported on a weight/weight basis fichu
data are based on the analysis of wipe samples and reported on .i
weight/area basis. Thus, it is very difficult to compare one tit#
Incident with another, or to evaluate the significance of the da ra iis
one Incident relative to that of another.
• Lack of Background Data. Very few background data are available on-
(a) composition of transformer fluids and. (b) composition and 1 eve 1 s
of PCBs, PCDFs, chlorobenzenes, and PCDDs in the environment.
Despite these problems, the following conclusions may be madf from
analvsls of the data from the literature on PCB transformei firesi
1) PCDFs and PCDDs are not formed In transformers containing PCBs under
normal operating conditions. Their formation requires thermal lv
stressful conditions and the presence of oxygen
2) Electrical arcings in transformers do not lead to the formation or
PCDFs and PCDDs
1) k temperature zone between 600°C and 680°C may be regarded as upturn
for the formation of PCDFs
8
-------
4) The amount and the specific PCDF Isomers formed ate rclaced to tiu
concencracion of and type o'~ homologs In the transformer fluid
3) Chlorobenzene diluents in he transformer fluids are required for the
format Ion of PCDDs
The Binghamton, NY transl. - -r f 1 re accident was the first to capture
major media and scientific attent i. » In that fire, both PCDDs and PCDFs were
foand in the generated soot, leadi, to the concern ch.it these compounds were
being formed in situ in PCB cransformers and capacitors under normal operating
conditions. The available evidence does not support this concern Analyses
conducted by Electric Power Research Institute and EPA of samples of die lev:1 Lc
fluids taken from in service transformers at.d capacitors and those involved in
fire accidents showed no appreciable difference in PCDF values from stock
material. No PCDDs were detected Under normal use conditions, it does not
appea. that PCOFs are generated to any significant extent in the transformer
Correspondingly, a second concern was whether PCDDs and PCDFs are formed
during electrical discharges associated with transformer performance To
investigate this issue, experiments involving the arcing of electrical eivrgv
through various transformer fluids were performed There was no appreciable
difference In PCDF levels before and after electrical arcing. The suppos ruv.
is that the level of oxygen is very low in the transformer and, thus. (loos nor
offer the environment for combustion resulting in the formation of PCDFs ,nu!
PCDDs.
The amount and the specific PCDD and PCDF isomers formed in ,i PCB
transformer fire appear to be related to the concentration of and tvpe ol PCB
homologs in the transformer fluids. This supposition is supported bv i» .u-nt
estimations on the boiling points for dioxlns and furans that indicate that : he
boiling temperatures for tri-CDDs and tri-CDFs and the higher chlorinated PCDDr
and PCDFs range from 375°C to 537°C. A first approximation of the th^nro*
dynamic conditions would favor the formation of trl- and higher chlorinated
PCDFs • d PCDDs Combust Ion studies conducted bv EPA and F. PR I also provide
evidence that the concentration of and type of PCB homologs in the r r.ins [ .
-------
6 7 5°C and a residence time of 0,8 second or longer Th«« EPRI stirtv
demonstrated that tetra- and penta-CDF yields are roughly proportional to PCS
concentrations in the starting material, but it indicated that significant
dibenznfuran destruction begins to occur at approximately 5S0°C.(7,8)
A final issue pertains to the question of the use of diluents in
transformer fluids. Of the 30 reported fire incidents involving PCS
transformers and capacitors, unequivocal evidence of PCDDs formation was found
only in the Binghamton, NY fire. The Binghamton, NY transformer contained
chlo robenrenei as a diluent, adding to the evidence that the pyroiysis of
chlorobenz"nt:i leads to the format lea of PCDDs Such evidence has also been
found in chemical manufacturing processes and In metal recovery sites involving
PCS transformers, Chloronsnzenes should be carefullv evaluated for us# a*
transformer fluids or diluents
10
-------
SECTION 3
FIRE INCIDENTS INVOLVING PCS TRANSFORMERS
AND CAPACITORS
PCBs have been used extensively as dielectric fluid in capacitors and
transformers since the 1950s, An estimated ,000 tons of PCBs are still In
use In U.S. transformers and capacI tors ^ ^ At least 30 f i i c* incidents
involving PCB transformers and capacitors have occurred in the past 7 -.-ears
(Table i). Thesc fire incidents can be divided inn three categories: » 1) F* E
fires with mineral oil as the external energy source; (2) PCB fires vith
electrical energy as the external energy source; and (3) explosions in s'CB
filled capacitors The following fires are discussed in more detail i - this
section:
• linghamton. New York
• San Francisco, California
• Chicago, Illinois
• Cincinnati, Ohio
• Boston, Massachusetts
• Miami, Florida
• Columbus. Ohio
• St. Paul, Minnesota
• Norrtalje, Sweden
• Stockholm, Sweden
• Skovde, Sweden
• Surahamroar, Sweden
• Imatra. Finland
• HalIstahammar, Sweden
• Railway Locomotive, Sweden
• Kisa, Sweden
• Reims, France
• Other Fires.
The finding of PCDFs and potential generation of PCODs from ask a r« 1 v umicr
thermal conditions may influence s t ritegles for mitigating and con:rn11; iv
chlorinated dioxins and furans in the nation's environment. Hi** • I'*';,
combustion products have complicated cleanup procedures for buildi.ni'> h,r-::v.
fires involving PCB transformers and capacitors The ana 1vs i s of soot ,
from actual PCB transformer or cap >t fire incidents has indi».irn! • h.
presence of PCDFs, and, in at least two incidents, the presence of P<:!)t)s
11
-------
Table 1. Fire Incldents Involving PCB Transformers
or Capac icors Since 1978(9-l0)
LOCATION
NorrcaLjc. Sweden
September 25, 197S
Cincinnati, Ohio
December 3, 1980
Bi -haniton. New York
February 5, 1981
Stockholm, Sweden
August 25, 1981
Danviken. Sweden
1981
Boston, Massachusetts
January 1982
SkSvde, Sweden
March 19, 1982
Miami, Florida
April 13. 1982
Arvika, Sweden
Nay 1982
Sc. Paul, Minnesota
June 22, 1982
insatra, Finland
August 3, 1982
Helsinki, Finland
August 1982
Surahamnar Sweden
September 23, 1982
Hallstahamaar. Swe en
No\ember 8, 1982
Railway Locomotf ve, Sweden
Winter 1982/83
Kaukopaa, Sweden
1982
Klsa, Sweden
April 25, 1983
San Francisco, California
Hay 15, 1983
Halmstad, Sweden
August 15, 1983
Chicago, Illinois
September 28, 1983
Bofors, Sweden
December 21, 1983
Columbus. Ohio
March 1984
Sodertalje, Sweden
April 27. 1984
Finspang, Sweden
May 24, 1984
Hofors, Sweden
September 13. 1984
Vetlanda, Sweden
October 10, 1984
Reims, France
January 14, 1985
Oslo Lysverker, Norway
January 1985
Sandnes, Norway
February 1985
Raufoss, Norway
February 1985
DATE
12
-------
Singhamcon. Sev York
On February 5, 1981, a fire broke out in r'.te basement mechanics' ioom .
the 18•story State Office Building in Binghamton. New York. The switchgear in
the mechanical room was located adjacent Co two electrical transformers filled
with 1,060 gallons of Pyranol, a nonflammable insulating oil. Electrical
arcing from the switchgear Ignited insulation around one ot th r ranstonners
Several explosions were heard from the machine room and. before the electrical
power could be switched off. It was oovious that the fire had affected cbv
transformer. The heat from the fire cracked a ceramic hushing on one ot the
transformers and approximately 180 gallons of Pyranol c'jbl Aroc lor 125* and i><
chlorinated benzenes) sprayed onto the floor and into the hot switchgear
Immediately following the fire and before the extent and nature of the
contamination was known, more than 300 people reentered the building to help
with the cleanup or to retrieve documents and personal effects Over
$1 billion in lawsuits have been filed tn the names of 23 individuals and
groups against the state, most filed by fir® fighters. Some S00 pec pie .-ho
believe thev were exposed are being monitored.
The New York Office of General Services began an immediate clemup ot ;!u
building. Members of the cleanup crew <" "eloped skin rashes and on Fehi ;.n .•
26. cleanup was ha 1 ted when the presence of PLDFs and PCDDs was discovered.
Analysis of soot samples recovered f rom the building revealed h i »•>,
concentrations of PCBs and the presence of ten different polvchlorinatt-d
dibenzofurans. dibenzo-p dioxins. and biphenylenes at lower concent rat ions
PCBs. 2 . 3 , 7 . 8 • TCDD . and 2.3.7.8 - TC0F were found throughout the i»u t I d i .
Ana I vs Is results are summarized in Tables 7. and 3, The total :* *JF
concentrations in the soot were initially reported to be as high as J. l». J .
!'« DDs 10-20 ppra. and PCBs 100,000-200.000 ppm ( ' - ¦ 12 . I). Ui
Solid waste disposal from the site had reached a total of * , >> 7 #> «••:!»! v
vards of material by rebruai v 1985. Th I s material filled I . ¦*(>•'* barrels vi.ich
were buried in an approved secure landfill site. All ot the furrurm
the building was removed, destroved and buried, and will have to he r e p 1. ic •.!
13
-------
Table 2 Concentrar ions of PCBs, PCDDs, and PCDFs Fo i at Bingham:on
State Office Building Prior to C :eattup' ^ ^*
SAMPLE LOCATION
TYPE OF SAMPLE
CHEMICAL
CONCENTRATION(AVERAGE)
'omposIte (throughout Air
bui lcling)
Composite Air
Compos Ice Air
Composite Soot
Compos It# Soot
«:h and 7th floor Soot
¦**h and 7th floor Soor
Air ducts Soot
Interstitial space
above ceiling levels Soot
Exposed horizontal surfaces
(desk tops, floors,
sills) Soot
Exposed vertical surfaces
(Wilis, desk sides) Soot
Unexposed horizontal
surfaces (inside closed
file cablrecs and desk
drawers) Soot
Unexposed vertical surfaces
(Inside closed cabinets
and drawers) Soot
PCBs
2,3.7,8-TCDD
2 , 3,7,8-TCDF
2.3. 7,8-TCDD
I. 3.7.8-TCOF
2,1,7,8-TCD0
2,3. 7,8-TCOF
PCBs
PCBs
PCBs
PCBs
C-arpge and sn" -basement
floors and surface
areas
Soot (swab)
PCSi
PCBs
PCBs
1.48 ntcg/ii^
0 . 3 p p n-
21 ppm
3. 5 ppm
200 ppm
2•5 ppm
300 ppm
156-1200 mcg/m*
1995 mcg/m*-
162.18 racg/rr-'
6.76 mcg/nr
74.47 rscg/irf
*
4.62 meg, ra~
0.61 mcg/m'
14
-------
Table 3. Levels of PCDFs and PCDDs in Soot from Accidental Burning
of PCB-Contalning Electrical Equipment at Bingharaton^-*-
ISOMERS AMOUNT (ppm)
Total PCDFs 2163
local Tr1-CDFs
Total Tetra-CDF? 28
2.3,7,8-Tetra-CDF 12
Other Tecra-CDF Isomers (6) 16
Total Penta-CDFa 670
I,3,4,7,8-Penta-CDF 65
1,2,4,7,8-Penta-CDF 25
12.4.7 9•Penta-CDF 2?
1 2.3.7.8- Penta- CDF 310
1.2.3.6. 7-Penta-CDF 60
t,2.6,7 8-Penta-CDF 25
J,J,4.7,8-Penta-CDF 48
2 3. •. 6 . 7 ¦ Penta • CDF 12
0:her Penta- CDF I toners (12) 110
Total Hexa-CDFs 965
1.2.3,4,6,8-Hexa-CDF 50
13,4.6,7,8-Hexa-CDF 1 * >
1 2.4.6.7.8-Hexa-CDF 0
1 2.3.4,7.8-Hexa-CDF 510
1.2.3.6.7.8-Hexa-CDF 150
1.2.3.6.8.9-Hexa-CDF 58
2,3,4,6,7,8-Hexa-CDF 10
Other Hexa-CDF Isomers 250
Total Hepta-CDFs 460
12.3.4.6.7.8-Hepta-CDF 230
1.2,3.4.6,7,9-Hepta-CDF 120
12.3.4.6.8.9-Hepta-CDF 55
1 2.3.4,7.8.9-Hepta-CDF 55
Octa-CDF 40
Total PCDDs 20
Total Tri-CDDs
Total Tetra-CDDs I 2
2 3 ' 8-Tetra-CDD 0 6
v her Tetra-CDD Isomers (4) 0 6
15
-------
Table 3, Levels of PCDFs and PCDDs In Soot from Accidental Burning
of P("B-Containing Electrical Equipment at Blnghamton (Continued)
ISOMERS
AMOUNT
Total Penta-CDDs
O
%T\
I
1,2,3,7,8-Penta-CDD
2 . 5
Other Penta-CDD Isomers (7)
? . 5
Total Hexa-CDDs
4 . 7
1,2,3,4,6,8-Hexa-CDD
1.2
1 .2,4,6.8,9- Hexa-CDD
1.2
I,2,3,4,7,8-Hexa-CDD
0.7
1,2,3,6.8.9-Hexa-CDD
0.6
1.2,3,7,8.9-Hexa-CDD
0.4
1,2,3,4,6,7-Hexa-CDD
0.5
Total Hepta-CDDs
7
1.2.3,4.6,7,9-Hepta-CDD
4
1.2 , 3.4,6,7,8-Hepta-CDD
3
Octa-CDD
i
The building is scheduled to be reopened during the .summer of 1986. bv vhich
lime the cost of the cleaning operation Is estimated at close to $22 mill inn.
which exceeds the building's original cost of $20 mill ion. Estimates or the
cost of tearing down the building, safely disposing of the concanu n.ttmt
materials, and erecting a replacement building, reached as high .is $ 1 On
mill Ion.
Reoccupancy of the building is based on an exposure assessment carried nut
by the Bureau of Toxic Substance Assessment of the New York State Dep.trtment «J
Health The contamination cleanup criteria esc blished by the expert advisnrv
panel was lU pg/m^ of 2,3.7.8-TCDD equivalents in the air of the bui Idiiv. uul
1 pg/ra^ of 2,3, 7,8 -TCDD equivalents present on surfaces.
16
-------
San Francisco. CMlllomlM.
On the morning of Sunday, Hay 15. 1983, an electrical fire of unknown origin
occurred in an underground transformer vault located at One Market Plaza in San
Francisco adjacent to a 28-story high-rise The vault, located beneath a
sidewalk manhole grate, contained three Pacific Cas and Electric Company (Pi&£>
transformers and associated network protection equipment Each of the rhtx-
transformers was filled with approximately 515 gal Ions of PCB fluid t 100*
Aroclur 124?).
During the course of the fire, a cooling fin on one of the transformers was
damaged, resulting in the release of approximately 50 gallons of PCB fluid
The fire burned and smoldered for approximateIv 8 hours after discovery
Subsequent sampling of the building and surrounding area indicated that PCBs
and by-products of PCB combust ion had been released through the side- 1 k
grate, through conduits into the building's swi tchgear room, and through
discontinuities in the vault wall into an adjacent basement gar.ir.e . 7lw
ventilation system serving floors B-2 through 6, which draws air from 4 pout:
near the switchgear room, was also shown to have suffice contamination of PCDs
and combustion by-products
Samples taken and analyzed after the incident revealed that the trans*
former oil involved was 1001 Aroclor 1242 and contained no chlorobenrenes
Sampling for PCBs and tetrachlorodibenzofurans (TCDFs) In the vault after the
fire and prior to cleanup revealed the following (h 1 ghe s t recorded)
contamination levels:(15)
Sftiftk E£M IfiftEs
Soot from manhole cover 110.000 ppm
Vlpc sample from floor 910,000 meg/100 cm^
Soot from vault wall -• 15.0 ppm
Analysis revealed 2,3.7,8-TCDD in a concentration approximatelv <> I or
that detected in soot of the Blnghamton State Office Building ^1^
17
-------
Approximately 4,000 air samples were taken in various locations Inside and
outside the high-rise tower to moni tor decontamination efforts The highest
air concentrations were measured In the vault; the second highest concen-
trations were found In the swltchgear room, adjacent to the vault. Prior to
cleanup, air samples in the vault revealed PCB levels up to 1,500 mcg/i»V and
air samples from the switchgear room showed PCB levels up to 98 6 mcg/m* The
highest level outside the vault and switchgear room was found in a sun'.igc
room, where a level of 12 mcg/m* was measured.
On September 28, 1983, an electrical fire in an underground PCB trans-
former vault forced the precautionary evacuation of a Chicago office building
The fire was caused by a 480 volt fault and resulted in the release of 15
gallons of oil from one of four transformers in the vault. The oil was
composed of 651 PCBs (Aroclor 1260) and 35% trichlorobenzene PCB contam-
ination was limited to the vault, an adjacent fan room, underground garage,
vault exhaust, sidewalk grill, and external face of the building, Analvsis of
the oil remaining after the fire revealed the presence of PCDFs and P€DDs», A
square foot (I ft x 1 ft) wipe sample of soot on the vault ceiling contained
PCDFs and PCDDs The results of the analysis of the oil and soot are presented
In Tables 4 and 5,
Cincinnati. Ohio
On December 3, 1980, a capacitor containing PCBs for an electric motor in
a unit heater overheated in a basement storage room of Our Lady of Visitation
Elementary School in Cincinnati, Ohio.The capacitor was used for >
one-half horsepower electric motor in the unit heater. The motor's capacitor
contained a design specification dielectric fluid volume of 22 mi. of which
99 6I by volume (or 21 91 ml) was a biodegradable fluid and 0,4% by volume (or
0 088 ml) was PCBs (Aroclor 1254). On March 18, 1981, the National Institute
for Occupational Safety and Health (NIOSH) was requested by the Hamilton County
Health Department to determine the extent of PCB contamination of the school
Air and surface wipe samples were obtained throughout the school bui Id ion
March 19 and 26 to determine the presence of PCBs.
18
-------
Table 4, PCBs, PCDFs, and PCDDs in the Askarel
After Che Chicago Fire**7*
ISOMERS CONCENTRATION (ppb)
PCB» 650,000,000
PCDFs
2, 3,7, 8-TCDFa 7b
lonal tetta- 230
Penca- 440
Hexa- 980
Hepta- 1,850
Oc ca- 1,600
Total PCDFs 5,100
PCDDs
2,3,7,8-TCDD N D b (2,5)
Total tetra- N.D. (2.5)
Penta- N.D. (2.5)
Hexa- N.D. (3 0)
Hepta- N.D. (2.0)
Octa- 8.0
Total PCDDs 8.0
a2,3,7,8-TCDF and co-elutlng Isomers.
19
-------
Table 5. PCBs. PCDFs, and PCDDs In Wipe Samples from the
Interior of the Transformer Vault from Chicago Fire^7^
ISOMERS AMOUNT (ng/ft2)
PCBs 20,900,000*
PCDFs
2 ,3.7»8-TCDFb 3,800
Total Tetra- 28,000
Penta- 40,000
Hexa- 33,000
Hepta- 11,200
Occa- 1,238
Total PCDFs 113,438
PCDDs
2,3,7.8-TCDD N D c C5)
Total Tetra- N.D. (5)
Penta- N.D. (25)
Hexa- N.0. (55)
Hepta- 314
Octa- 127
Total PCDDs 441
aPCBs from wipe of vault wall; PCDFs/PCDDi from wipe of vault celling
^2.3,7.8-TCDF and co-elutIng Isomers,
cN.d - Not Detected; detection Halts In parentheses
20
-------
Airborne PCB concentrations were reported as micrograms (meg) of PCBs per
cubic meter (ra^) of air sampled. As seen In Table 6, PCBs were not detected in
air samples obtained In the basement storage room nor in another room ces r«-
-------
Table 6, Analysis of PCBs in Air Samples from School In Cincinnati, Oh i o(^1
AIR SAMPLE
SAMPLE LOCATION VOLUME (L) LEVELS
Inside heating unit on floor (Roots 27) 45.2 N D a
Center of room on bookshelf 43 3 N D
Principal's office 34,7 N.D
aN.D or Not Detected means chat the PCBs were not detected at the lowest
level (< 0 05 meg/sample) capable of being measured by the analytical method
equivalent to an airborne concentration of approximately < 1 mcg/m-* for these
samples.
Tabl# 7. Summary of PCBffl Ulpe Sample Results from School In
Cincinnati. Ohlo^®^
PCBs LEVEL (mcg/100 cm2)
SAMPLE LOCATION
NUMBER OF
SAMPLES
MEAN
RANGE
Room 2? (Basement location)^
20
771
<
0.05
?200c
Room 25
4
0 08
<
0 05
0, 14
Room 24
3
0,05
<
0.05
0,06a
Room 26
3
< 0 05
<
0.05'
Hallway between rooms 24-27
4
0.05
<
0 05
0.07
Room 8
4
0,06
<
0 05
0,08
Room 22
6
0.11
<
0 05
0.29
Room 9
2
0.11
<
0 05
0. 16
Room 10
4
0,12
<
0.05
0.20
Room 11
2
0.09
<
0.07
Oil
Room 12
2
0 08
<
0.06
O.Oi
A V room
4
<0.05
<
0,05
PrineIpal* s office
3
0,21
0,05
¦0,45
Mobil# classroom
2
0.06
0,06
Background level (3 locations)
10
0.07
<
0,05
•0 13
aReported as Aroclor 1254.
^Rooms 27, 25, 26, and 24 were all in the basement,
cA < value means PCBs were not detected at detection limit
^Excludes sample taken on capacitor surface showing PCB presence of 4 > iuv.
100 cm^
cExc ludes sample taken on capacitor surface showing PCB presence of 2 J : ci*
100 cm^.
22
-------
Boston. Massachusetts
In January 1982, an electrical fir# Involving a transformer containing
Aroclor 1254 occurred In a government office facility in Boston, Massachusetts
One bulk soot sample was spiked with 200 ng dj2-chrysene and was then Soxblet
extracted for 24 hours with hot toluene This extract was decanted and labeled
fraction I. More hot toluene was added to the sample for an adoltion.il
24 hours. The extract was labeled fraction 2 Both extracts were analyzed
70* of the d^j - chrysene was recovered In fraction 1. The two extracts wele
concentrated. Fraction 1 required multiple HRGC•MS analysis because of its
complex components. PCDFs were confirmed in fraction 1. These results are
shown in Table 8 (
tfijml, Florida
On April 13, 1982, a fire Involving an electrical transformer occurred in
an underground vault In Miami, Florida (20) The International Association of
Fire Fighters (I A F F ) requested NIOSH to perform analyses of surface samples
in order to determine If fire fighters had been exposed to PCBs or ocher trot*
toxic decomposition products such as PCDOs and PCDFs.
in these analyses, NIOSH collected samples using two different methods
Wipe samples were collected using sterile cotton pads saturated with hexane
from sample areas with an approximate size of 100 cm^ Surfaces thought to be
less heavily contaminated were wiped with dry Whatman smear tabs.
Table 9 pre* »nts the results of the surface sample analyses for ?CB con-
tamination. Samples taken within the vault showed the heaviest contamin u ion
Turnout coats and other personal protective equipment were not found to he con -
tamlnated (probably due to the decision to allow the fire to seIf-extinguish•
The only contaminated equipment at the station was found to be the smoke-
ejector fan. Results showed PCB contamination of > 31 meg/100 cm^ l'i ev i .
NIOSH investigations indicate normal background levels for PCBs on non-
contamlnated surfaces should be less than 0,5 meg/I00 cm^ ^20/
23
-------
Table 8. Levels of PCDFs and PCDDs from the Accidental Burning cf
PCB-Containing Electrical Equipment at Boston, Massachusetts*
CONCENTRATION
ISOMERS mcg/g SOOT
Total PCDFs 165
Total Trl-CDFs 50
Total Tetra-CDFs 60
2,3.?,8-Tetra-CDF i
Total Penca-CDFs 35
Total Hexa-CDFs 15
Total Hepta-CDFs 2
Octa-CDF S D.a
Total PCDDs
N.D
Total Tri-CDDs
S . D
Total Te era-CDDs
N.D
2.3,7.8-Tetr?-CDP
N.D
Total Penta-CDDs
N.D
Total Hexa-CDDs
N.D
1.2.3.4.6.7.9-Hepta-CDD
N D
I.2.3,4,6,7,8-Hep c a -CDD
N D
Octa-CDD
N.D
aN D - Not Detected; detection limit at 100 ng/g -01 mcg/g - 0 1 ppo»
24
-------
Tible 9, PCB Residue from a Transformer Vault Fire,
Collected 4/16/82 In Miami. Fl>ridat20)
SAMPLE LOCATION/DESCRIPTION
SAMPLE TYPE
Inside Transformer Vault:
Wall behind removed transformer (soot) wipe
Top of primary cable above fire (soot) wipe
Primary cable support bracket (soot) wipe
Floor near base of Isolating switch (dirt) wipe
Ceiling near fire location (soot) wipe
Secondary bus near vault celling (dirt) wipe
Vail next Co exit ladder (dust) smear cab
Rung of exit ladder (dust) smear tab
Above Transformer V^uit
Sidewalk grating smear cab
Standing water at curb near vault wipe
Fire Fighters' Clothing and Equipment:
Sleeves of turnout coat smear Cab
Top of boot smear tab
Inside helmet smear tab
Front of turnout coac smear cab
Inside helmet face shield smear tab
Outside faceplece smear tab
Smoke ejector fan smear tab
PCBs*
(mcg/LOO cm^)
434b
389
704
2?,400
860
195
2
?9
0. I
0. I
0. i
0.1
0. I
0 2
31
aAs Aroclor 1260 (used as standard for quantitation of samples)
bAs mixture of Aroclor 1254 (231 meg) and Aroclor 1260 (203 meg)
Table 10 presents the results of the analyses • r 6 bulk samples oI \o«t
and other fire residues for the presence of PCDD and I >F isomers No F«:bD*
were detected in these samples, but PCDFs from crichloro* ' > hexachloro-
1st «srs were detected in samples Bl to B6 The samples also exhibited hif;h
levels of PCBs through CIand polychlorInated dlphenylechers (PCDEs) thtouch
Clg. The highly coxic 2.3,7.8- retrachloro Isomer was not detected in ,tnv ot
these samples.
Gas ch roma to graph/mass spectrometer (GC/MS) analyses of soot samp 1 . s
collected from the vault ceiling directly above the fire scene wpto ,
performed. Peaks ident I f led Included penta • , hexa-, and heptachlorobiphfiivl s .
and numerous alkanes. most 1v larger than C^q
25
-------
Table lu Results of Analyses of PCDDs and PCDFs iri Bulk Samples of Residut- from
Miani Transformer Fire^O)
NUMBER CONCENTRATION
ISOMERS POSSIBLE NUMBER DETECTED IN SAMPLE (ng/g or ppb)
Samples:
Bla
B2
83
84
B5
B6
B1
B2
B3
84
B5
Lb
Hono-CDDb
2
3
0
0
0
0
0
N.D c
N.D.
N.D.
N.D.
N.D.
N.D
Di CDD
10
0
0
0
0
0
0
N.D
N.D.
N.D.
N.D.
N.D,
N.D
Tr i - OD
14
0
0
0
0
0
0
N D
N D
N.D.
N.D.
N.D.
N.D
Tetra-CDD
22
0
0
3
0
0
0
N.D,
N.D.
N.D.
N D
N.D.
N.D
Penta CDD
L4
0
0
0
0
0
0
N D
N.D
N.D,
N.D.
N.D.
N.D
Hexa¦CLD
10
0
0
0
0
0
0
N.D,
N.D
N D
N.D.
N.D,
ND
Hepta-CDD
2
0
0
0
0
0
0
N.D.
N.D.
N.D
N.D.
N.D.
N.D
Oct*-CDD
1
0
0
0
0
0
0
N.D.
N.D.
N.D
N.D.
N.D.
N.D
2,3.7,8-TCDD 1 0 0 0 0 0 0 N D N D N O N D N D. N.D
aSa»ple Descript ion/Location: B1 Soot. dusc, and dirt fron top of primary cable support bracket, circuit »2.
B2 Residue from floor under isolating switch near burned transformer,
83 Scrapings fron ceiling above transformer
B4 Scrapings from secondary bus,
85 Vault wall near exit ladder
B6 Scrapings of soot and dirt from snoke ejector fan used to exhaust smoke fro®
vault during fire
''Detection Limit for PCDDs 10 ng/g for samples B1 , BS, and B6; 100 ng/g for s.impl rs B/, 81, and B4 ,
-------
Table 10 Results of Analyses of PCDDs and PCDFs in Bulk Samples of Residue from
Miami Transformer Fire (Continued)
NUMBER CONCENTRATION
ISOMERS POSSIBLE NUMBER DETECTED IN SAMPLE (ng/g or ppb)
Samples:
Bla
B2 B3
B4
85 B6
B1
B2
B3
B5 B6
Mono-CDF1'
U
0
0 0
0
0 0
N.D c
N.D.
N.D.
N.D.
N.D. N.D
Di-CuF
16
0
0 0
0
0 0
N.D,
N.D.
N.D
N.D
N.D. N.D.
Tri CDF
28
6
0 0
0
0 6
180
N.D.
N.D
N.D.
N.D. 110
Tetra-CDF
38
6
0 0
0
0 7
ViO
N.D,
N.D.
N.D.
N.D. 280
Penta-CDF
28
7
0 0
0
0 b
1000
N.D
N.D.
N.D.
N.D. 290
Hexa-CDF
16
3
0 0
0
0 '«
180
N.D.
N.D.
N.D.
N L. 100
Hepta-CDF
4
0
0 0
0
0 0
N D
N.D.
N.D.
N.D.
N.D. N.D.
Octa-CDF
1
0
0 0
0
0 0
N D
N 0,
N.D,
K D
N.D. N.D.
2 ,1,7,8-TCDF
1
0
0 0
0
0 0
N D
N D
N.D.
N.D.
N.D N.D
aSample Description/Location:
B1
Soot,
dust, and
rt
from top of
primary
cable
support bracket, circuit »2
B2
Residue from floor under isolat i
ng
switch
near
burned transformer.
B3
Scrapings from
ceil i
ng abo'- transformer .
B4
Scrapings from
secondary bus .
Bb
Vault
wall near
exi t
ladder,
B6
Scrapings of soot and dirt from
smoke ejector
fan used to
exhaust smoke from
vaul t
during fire
1 i-c i i on Limi t
for PCDFs
10
ng/g for samples
B1 .
B2 . B1), and
B6
; SO
/ lut s.iiap 1 1 K'«
100 ng/g
for
1),
i
-------
Columbus. Ohio
In March 1984, a capacitor fire occurred in a Columbus office building.
The capacitor was filled vi;h Aroclor 126 2 Analysis of soot samples revealed
rhe pvesence of PCDFs and trace amounts of PCDDs (cetra- to octa-isomers),
Chlorubenzene was also detected in trace amounts with a detection limit at ppm
level. Analysis of wipe samples indicated tha presence of PCDFs and
PCDDs (21.22) Ta5ie n presents the results of the analyses of the soot and
wipe samples.
Table 11. Analysis of Wipe and Soot Samples
from Columbus F1 re^^ •
TOTAL
SAMPLE ISOMER CCK-THTRATir*'
Wipe 2,3,7,8-TCDF < 1,000 nt
Total PCDFs < 10,000 n8
2.3,7,8-TCDD < 3.0 ng
Total PCDDs < 44 ng
Soot Total PCDFs 10.9 ppm
Total PCDDs (Cetra to occa) Trace®
Chlorobenzene Trace3
a'. action 1 imi t at ppm level.
Sr. Paul. Hinnesuca
On June 22, 1982, a PCBs fire occurred at the Hill-Murray High School in
St Paul, Minnesota (23) miqsh samples of surface wipes and air were obr.iimd
on June r3 . 1*J82, and again on July 7 The results of the analyses .uv
reported in Tables 12 and 13.
SorrtalLe. Sweden
On September 25, 1978, there vas an intense fire in a capacitor bar tor.-
north of Norrtilje, Sweden. The fire could not be extinguished for
-------
Table 12, Analysis of PCBs (Reported as Aroclor 1260) in Wipe
Samples from the Hill-Murray School(^
PCBs (neg/100 cm*)
SAMPLE LOCATION
6/23/82
7/7/82
Vault Room 100-5000
Vault Door < 5
Corridor out s I tie Vault Room < 5
Roon 048 20
Boiler Room < 5
Cafeteria 0.22-0,24
Corridor leading to the Fieldhouse <0.05
Corridor wall across from Room 109 0 26
Corridor wail adjacent to Music Suite 0 2 7
Corridor wall leading co Office 1.6
Out side Room 106 5,8
Room 106 0.61-0,95
Room 138 1.4-2,1
Gymnasium 0.29
Athletic Building lobby 0.14
Priory < 0.05
2-120
< 0,05
<0,05
< 0.05
< 0.05
O.i
0/2-0 >
29
-------
Table 13. Area Concentrations of PCSs {Reported as Arcolor 1260)
and Chlorinated Benzenes at Hill-Murray Sc hool * ' )
AIRBORNE
CONCENTRATION
(mg/a3)
SAMPLE VICINITY
It DATE
SAMPLE
DURATION
(HOURS)
SAMPLE
VOLUME
(LITERS)
PCBs*
TOTAL
TRICHLORO-
BENZENEb
TOTAL
TETRACHLORO
BENZENEC
Transformer Vault
Transformer Vault
Vault Doors
Room 018
Outside Room 109
Cafeteria
; 54
2.52
2.3?
3.06
2.67
2.98
13?
116
123
165
144
161
<
<
0.05
0.09
0.02
0,004
0 001
0 001
22.6
17.6
10,2
0 73
0. 12
0, 35
U 2
23. *
u ;
1,21
0,26
0.27
Transformer Vault
% ault Doors
Outside Room 018
Outside Room 109
•'.ifece'la
2.82
2.83
2 82
2.92
1. 92
169
170
169
175
175
<
<
<
<
0.007
0.001
0.001
0.001
0,001
0.331
0 048
0,020
0 Oil
0.010
0,6H
o.2,:«
0 . U8'>
0,01'
ti. n.) 2
JNIOSH recommended perrelssLble exposure 1 Imlt tor an 8-hour t Lme • w# Igtu»-r
OSHA have exposure criteria
cSo 8-hour TV/A exposure criteria.
-------
capacitors (100 kVA each) were involved In the fire. When the capacitors we re
denounced. some showed no signs of damage while others had exploded, releasing
their concents. B Jansson and G, Sundstrori, at the Special Analytical
Laboratory . National Swedish Environmental Protection Board located at the
University of Stockholm, were contacted a few weeks after the accident and
asked Co perform sample analysis for PCDF formation, (24) up Co 300 mg samples
together with 100 ng [2 , 3 , 7,8-^^C1\-tetrachloro*(I,4,6,9 *]* dibenzo-p-dioxin
as the internal standard dissolved in 1 ml n-hexane were chromatographed on
aluminum oxide. The first fraction containing most of the PCBs was cluted with
90 ml 51 dichloromethane in n-hexane. The second fraction, which was eluted
with 60 ml dichloromethane, contained all PCDF references, the internal
standard, and at least two PCBs.
Results of the analyses are given in Table 14, An Aroclor 1242 standard
was used as a reference. Samples were taken from intact and exploded
capacitors, from the wire fence surrounding the capacitors (fence was washed
with acetone to obtain sample), and from pine tree needles which were located
about 10 me ters (m) downwind from the fI re
The wire fence was found to contain unexpectedly 1ow PCB concentrations
The extraccable part contained only 0. 3 mg PCBs per g extractables. As .»
result of this low PCB level, the upper limit of the PCDF/PCB rat io for the
fence extract becomes high. While no presence of PCDF was found in the fence
sample, 1f the extremely high detection limit for the dlchlorlnated compounds
Is excluded, the highest possible concentration should be about 300 meg per g
PCBs .
In addition to these environmental samples, blood samples were taken from
six firemen who were exposed to the fire. The samples were not taken until .1
few weeks after the fire incident At chat cime Che PCB levels in the blood
samples were wlchin che same range (23 to 3.6 mg PCBs/g) as those for un-
exposed Swedes (24)
31
-------
Table 14. PCB and PCDF Levels in Capacitors and
Environmental Samples at NorrcalJe . Swtden^ •
PCDP SAMPLE (ncg/g)
SAMPLE PCB** MONO-CDF DI-CDF TRI-CDF TETRA CDF PCDP PCDP/PC1
(1/8) («cg)
Liquid in intact 0 830 86 0,2-0.5 0 2 0.20.3 0.2-0.3 0.9-1.1 1.1-1.3
capacitor
Liquid in exploded 0 93 24 38 11 3 75 81
capacitor
Liquid on exploded 0 '49-0.60 12-21 9 21 3-7 2 27-52 45-107
capacitor
Extract froa wire 0 3x10 3 <0 03 < 0.2 <0 03 <0 04 < 0.3
fence
Pine needles 0 3xlO*6 N D b N.D, N D. NO. NO.
aAroclor 1242.
bN D. - Not Detected.
-------
On August 25, 1981, an explosion (electrical arc) cook place In an 11 kV
capacitor battery in an electrical power station in Stockholm, Sweden The
explosion was caused by an electrical failure in the capacitor battery that was
so violent that it caused the wall to rupture, Wipe samples were taken about
1 ra from the capacitor and were tested for PCDF and PCDO levels. Extraction
and measurement we re accomplished by spiking the wipe tests with 1*5 ng of
labeled COD, The sample was treated with 10 ml of 1M HC1 for 1 hour, The
slurry was filtered and dried, extracted with toluene, and dried with H2 The
residues were 'dissolved in 2 mi of n-hexane, added to a si 1 lea gel col* mr, and
elut#d with ail of n-hexane The hexane was evaporated with U? until 10 ascl
remained. PCDDs and PCDFs were separated from other polychlorInated impurities
using an Alox- column The first fraction, 10 ml n-hexane methylene chloride
(98:2), was discharged. The second fraction, 10 ml n-hexane methylene chloride
(1.1), was collected and dried in N2 This residue was used in the GC/MS
analyses. Individual MS-response factors were used to calculate levels giwn
in Table 15 The total levels of polychlorinatec biphenylenes (PCBPs) were
reported to be 25- 30 mcp,/m^ (9,13,25) Polychlor inated pyrenes (PC FYs 1 were
also detected but not quantified
Skevde. Sweden
On March 19, 1982, a violent fire broke out in a capacitor battery (400 V)
serving a high-frequency oven in a metal treatment factory in Skovde. Sweden
The dielectric fluid In the capacitors consisted of either mineral oil or PCBs,
and both types of capacitors were in use There were no chlorinated benzene
additives in the oil. The fire started in a mineral oil capacitor and 2 in.wrs
elapsed until the fire was extinguished. The smoke spread to an extenr of
1800 m2 (60 m x 30 m) . Copper bars for electricity in the capacitor toon
celling partially melted (m.p 10B0°C) . The capacitor battery contaiiwd .*1
capacitors filled with PCBs (5 kg each). After the fire. 12 of t !•¦«• s.-
capacitors had been opened and 9 remained sealed.
Wipe tests (tissue 1 dm2) were used to take samples from a) the floor of
the capacitor battery, b) "he floor close to the battery, c) the wall of rhe
33
-------
Table 15, Concentrations of PCDFs Found in Samples from an
Accidental PCI Fire In Stockholm, Sweden^9¦n•25>
PCDP
PCDF CONCENTRATION
(ng/a2)
.3.7,8-Tetra-CDF
I .2.7.8-Tetra-CDF
: . 3.6.8-Tetra-CDF
I ,4.6,9-Tetra-CDF
-W6, 7-Tetra-CDF
1,4,5,7-Tetra-CDF
1,3,6.7- and 1. 3 .6,9-Tetra-CDF
Other Tetra-CDFs
150
150
125
75
37
7.5
300
750
local Tetra-CDFs
1200
J , 3 »4,7,8 -Penta- CDF
I .2,4,7.8-Penta-CDF
1 .2.3,6,7-Penta-CDF
1.3.4,7.8-Penta-CDF
J. 3,4.6,7-Penta-CDF
1.2.4.6.8-Penta-CDF
1 .2.4. 7.8-Penta-f.DF
1.2.3.6.7-Penta-CDF
1.2.4.8.9-Penta-CDF
2.3.4.6.8-Penta-CDF
1.2.3,7,8- and 1.2.3.4.8-Penta-CDF
Other Penta-CDF«
45
38
15
11
15
19
Total Penta-CDF#
Total Hexa-CDFs
175
< 0.5
34
-------
capacitor room, 2 m from the battery, 3 m above che floor, and d) the bench
10 m from the oven (1 floor above the capacitor). Table 16 shows the results
of the analysis. High levels of PCDFs (> 0.10 racg/g) could only be found close
to the fire Although polychlorindeed pyrenes were detected. no PCDDs or PCBPs
could be Identified in the samples from this fire.(9.13.25)
Table 16, Levels of PCDFs in Wipe Samples
from the SkSvde Fire(9-
CONCENTRATION (ng/ra2)
SAMPLE
NUMBER
ISOMER
1
2
3
4
2,3,7
,8-Tetra-CDF
20
100
< I
< I
Total
Tetra-CDFs
100
600
< 1
10®
Total
Penta-CDFs
40
100
< 1
• 1
Total
Hexa-CDFs
(*0
60
< I
< 1
Total
Hepta•CDFs
8
8
< I
< I
Octa -
CDF
5
5
< 1
< 1
aDifferent Isomers than found in samples I and 2.
Sucmhammr, Svedm
On September 23, 1982, a 500-unlt capacitor battery at a steel mill in
Surahammar , Sweden, was ignited by 5 tons of molten steel (1500°C) The
capacitors were filled with PCBs (2 tons) and mineral oil (2 tons). A violent
fire broke out and burned for 2 to 3 hours filling the entire bulIding (140 m x
70 m x 25 n) with smoke O Table 17 presents the results of analyses ot rlu
wipe samples.
35
-------
1antra, Finland
On August 3, 1982, a bank of capacitors in a pa^cr mill outs ide Imacra v.i v
destroyed by explosions and fire. The fire was localized in che capacitor bjnk
behind che 10 kV contactor station. Soot and gas had spread wi ch che pressure
wave through an open doorway to the 10 kV contactor station and the adjacent
3 kV contactor station, and through cable holes to che storage space directly
above che contactor stations, to the floor beneath, and to the Instrument
cro-a-connectic n room.
Table 17. PCDFs Found in Uip« Tests at Surahamraar. Sweden (ng/rr^) * ''
SAMPLE LOCATION
TOTAL
TCDFs
2.3.7,8-
TCDF
CI 5
c H
Cl7
ci8
Capacitor Room (Sample 1)
4000
875
3300
1800
1500
300
Capacitor Room (Sample 2)
1100
365
1250
940
625
U!s
S.E. Corner, 10 m height
1250
300
355
150
65
n
S E Corner, 10 m height
480
120
210
140
60
30
N. E. Corner, floor
100
25
27
15
5
2
S.E. Corner, floor
90
22
25
17
17
4
10 m Outside, downwind
< 250
< 25
< 25
< 60
58a
l?a
300 m Outside, downwind
< 250
< 25
< 25
< 60
< 30
< 12
After Cleaning (Sample 1)
< 20
< 4
< 10
< 12
< 15
2*1
After Cleaning (Sample 2)
< 40
< 8
< 20
< 20
< 30
A*
^Possibly due to use of decachloropheny1 as an additive in casting wax
Each capacitor contained 20 liters of Clophen A 30 (a mixture of PCBs vi ch
an average chlorine content of 42l). Sooc samples were analyzed for PCDFs and
PCDDs. The results are included in Table 18,
36
-------
Table 18. Levels of PCDFs and PCDDs In Soot Samples
from Imatra, Finland Fire^9¦^
CONCENTRATION (ppm)
ISOMERS SAMPLE 1 SAMPLE 2 SAMPLE 3
PCDFs
,ino•COFs
80
>Ji • CDFs
2b0
Trt-CDFs
200
Tetra-CDFs
20
16
1,0
2,3,7,8-Tetra-CDF
1.0
0,0?
Penta-CDFs
1
10
0,2
Hexa-CDF&
0.3
0,04
Hepta-LUFs
0,2
0 02
Occa-CDF
0 1
0,01
PCDDs N D a
aN.D. - Not Detected.
Hallstshaimar. Sweden
On November 8. 1982, a capacitor exploded tn a foundry In Bulcen-Kjmch.il
In Hallstahamnar, Sweden. The capacitor was located close to a melting oven in
the ground floor, and the fumes spread through the entire building (40 m x 13 m
x 5 m). Wipe tests were collected for analysis and the results are presented
In Table 19,
Table 19. Levels of PCDFs In Wipe Samples from
Hal lstahamnar . Sweden FitcO
CONCENTRATION (ng/n2)
SAMPLE TOTAL 2.3,7.8-
TCDFs TCDF PeCDFt HxCDFs HpCDFs OCDF
Cop r bar, 2 da above 1600 540 360 600 800 1 )*<>
Bar, 5 n above 5000 1000 1800 850 550 .-.0
Floor 50 10 12 10 10 20
37
-------
Railway Lo.c
-------
On January 14, 1985, a violent transformer explosion occurred in a 6-s»cory
building In Reims, France. The transformer was filled with an askarcl of 60*
PCBs and 40% tec rachlorobenzene. The dlrcy, highly contaminated soot was
analyzed on February 21. 1985 and April 2, 1985, The results of these analyses
are presented In Table 22.
There have been a hose of other fires associated with PCB trans to risers aiuJ
capacitors. Many In the U.S. probably go unreported. Table 23 lists some
other fire Incidents that have been reported.
-------
Table 22. Levels of PCDFs and PCDDs in Soot Samples from
Reims, France Explosion
ISOMER CONCENTRATION (mcg/m2)
1
2
SAMVLE
3
NUMBER
4
5
6
Tetra-CDFs
>30
>8
>4,5
0.11
0.22
0. 38
Penca-CDFs
960
590
37
0 89
1.5
1 , 4!)
Hexa*CDFs
760
570
36
o
O
12
1 . 1
Hepta-CDFs
530
490
27
12
1.9
3. J
Octa-CDF
290
220
36
N A a
N.A.
N.A,
Tetra-CDDs
N.A.
N.A.
N.A
N.A,
N.A.
N" A.
Penta-CDDs
<1
<1
<0.005
<0 005
<0.005
<0,00J
Hexa CDDs
<1
<1
<0,005
<0 005
<0.005
<0.. on,
Hepta-CDDs
35
32
<0.005
0,11
0, 14
0, 13
Octa-CDD
16
1**
<0,005
0.18
3 45
0.3}
aN A - Not analyzed
Samples I and
transformer
2 were taken on Che
vault (February 21).
ceiling
of the
basement
near the
door of
Sample 3: wipe sample collected externally cn the dcor of the tmnslm ?ci'
vault (February 21).
Sample 4; wipe sample on the second floor, near the dooi of an ap.u'ti-rnr
(April 2).
Sample 5; wipe sample In a bathroom, second floor (April 2).
Sample 6: wipe sample In a kitchen, second floor (April 2)
Samples 4,5, and 6 were collected after cleaning.
40
-------
Table 23
PCDF Levels in Other PCS Fires'27)
SITE PCDF LEVELS PCBP LEVELS
Danvlken. Sweden 1.5 meg/m^ None
Helsinki, Finland
Halmstad, Sweden
Arvika, Sweden
KaukopAA, Sweden None None
CONTAMINATION CLEANIT CRITERIA
There are no Federal guidelines Co define acceptable cleanup levels for
toxic releases from PCB transformer and capaciror fires Current regulative*
state, however, that all spills and leaks wf PCBs diox ins-contaret narod
material should J* cleaned up to preexisting background levels whenever there
Is a threat of contamination of water, food, feed, or huuans. S'ICSH has
detected background levels in urban areas of up to U i meg PCBs/l 'O or- >t
surface area. Following the occurrence of PCBs¦ related fire incident!,. ,s«rvrtl
states and other countries have established contamination cleanup criteria
These criteria are presented tn Table 24,
41
-------
Tabl 24 Coi.taminat ion Cli-anup Criteria'^- ' 3 • 1^.2/)
LOCATION
CONTAMINANT
AIR
SURFACE
Binghaaton. NY Building
PCDDs/PCDFs
10 pg/n3 of 2,3,/,8-TCDD/TCDF
3 pg/nt2 of 2, 3, 7,8-TCDD/TCDF
PCBs
200 ng/«3 of PCBs
60 acg/a? of PCBs
Inside Vault
PCDDs/PCDFs
80 pg/a3 of 2,3,7,8-TCDD/TCDF
24 ng/ra2 of 2,3,7,8-TCDD/TCDF
PCBs
I ocg/«3 of PCBs
I «g/n if PCBs
San Francisco, CA
PCDDs/PCDFs
10 pg/o3 of 2,3,7,8-TCDD/TCDF
3 pk/t»2a
PCBs
200 iig/s3 of PCBs
60 ncg/a^ of PCBs
1nside Vault
PCDDs/PCDFs
80 pg/o3 of 2,3.7.8-TCDD/TCDF
24 ng/a2 of 2,3,7,8-TCDD/TCDF
PCBs
I acg/B3 of PCBs
1 ag/a^ of PCBs
aSum of all PCDD/PCDF isomers Cl^ • CI 7 with CI subst i tut ion in the 2,3,', and 8 positions.
-------
Table 24
LOCATION CONTAMINANT
Sante Fe, NN PCDDs/PCDFs
Finland PCDDs/PCDFs
Sweeten
PCDDs/PCDFs
Contamination Cleanup Criteria (Continued)
AIR SURFACE
1 ng/B2 of 2.3,7,8-TCDD/TCDF
5 ng/m2 of 2,3.7.8-TCDD/TCDF
50 ng/»2 of total TCDF
-------
SECTION 4
PCDFs, PCDDs, AND PCBPi FROM THE PYROLYSIS OF PCD FLUIDS
The formation of PCDFs and PCDDs has beon attributed to the range of
chemical reactions that occur during the pyrolysis of PCBs, the condens.it ion ot
chlorobenzeries, and the condensation or dlmerlzat ion of ehlorophenols. PCBs
and chlorobenzenes are major components of most ?CB transformer fluids. whereas
the ehlorophenols are intermediates formed in the range of 550 to 700°C The
amounts of these compounds formed during the combust lo i process vary with,
temperature and concentration
FORMATION OF PCDFs FROM PCBs
Laboratory studies have indicated that with adequate reaction time and
efficient mixing between air and fuel, PCBs are completely decomposed into H^O,
CO2. and HC1 at a temperature range of 800 -1000°C Pyrolytic conditions,
however, allow the format ion of PCDFs. The pyrolysis of commercial PCBs vields
about 30 major, and more than 30 minor, PCDFs and an indeterminable number ot
PCBPs.
The Importance of combustion as a source of PCDDs and PCDFs (and various
ch1orohydrocarbons) was discovered in 1977, when these toxic materials were
found in fly ash and flue gases of munlclpal lnclnerators.
In 1978, Mor Ita reported that heatlng Aroclor 1248 to 300°C in a sealed
glass ampoule for 2 weeks produced approximately a 4-fold increase In rht-
amount of PCDFs.(28) jn the presence of oxygen. PCDFs formation began .hm
samples were heated for 1 week above 270°C, reaching a maximum of approxitn,>r«*I .
0 2% conversion at 300°C. Dichloro- and tr ichlorodlbenzofurans were .11 so
formed from Che tetrachlorobiphenyls in Aroclor 1248
In a study by Buser ec al., Aroclor 1254 was pyrolvzeu at temperat-av.s « .
550° • 7 50°C In quartz mini - ampoules . (29) CC/MS analysis ihowed PCDI r««
formed In the process, The results are presented In Tjblr 25.
44
-------
Table 25. Pyrolysls of Aroclor 1254 In Quartz Hln .-Aapoules^9)
TEMP.
amt.
DBCOMP.
PCDFs
FORMED*
<%)
<°C)
<®eg)
<*>
MOHO'
DI-
T1I«
TETtA-
PEHTA-
TOTAL
550
100
12
0.25
0.65
0.90
0.75
0,20
2.75
II
10
80
0.02
0.40
0.70
0.70
0.15
1.97
600
100
45
0.10
0,40
0,70
0,60
0, 12
1,92
m
10
90
<0.01
0.10
0.25
0,35
0,05
<0.76
650
100
90
<0,01
0.02
0,18
0.25
0.12
<0,58
m
10
98
0,01
0.12
0.25
0,12
<0 02
<0.52
700
100
>99,9
<0.01
<0.01
<0.01
<0.02
<0 02
<0.07
n
10
>99.9
<0 01
<0,01
<0,01
<0.02
<0,02
<0 07
7 50
100
>99.9
<0.01
<0,01
<0.01
<0,02
<0.02
<0.07
'Combined values of all Isomers of a PCDF.
Mono- Co penta-CDFs were found at temperatures of 550°-650°C at levels
ranging from < 0.01-0,901. The maximum amounts were found at 550°C.
Considering the amount of PCBs recovered after pyrolysls, the total yield of
PCDFs ranged from 3-25%.
^o°c
PCDFs
x • I to 8
> 30 Isomers
1 to 6 (yield 3-2515
> 60 Isomers
45
-------
Buaer and Rappe, In subsequent studies, pyrolyzed 18 Individual PCS
Isomers In the presence of air at 600°C In sealed ampoules.(29,30.31) jt was
found chat the cycllzatlon process which yields PCDFs from PCB* was
intermolecular and followed several competing reaction pathways.^
30.31) In
addition to PCDFs, polychlorlnated bIphenylo1s Identified as possible
precursors, were present at levels of approximately one-fifth ot the amount of
FCDFs present.
From this work, four thermochemlcal reaction mechanisms were proposed
for formation of PCDFs from PCBs,
I. Mechanism l:
Example:
o
o
o
o
b ci
CI
2,4,6,2',4',6'-Hexachloroblphenyl
1,3,7,9-Tetrachlorodlbenzofuran
2. Hgghftnlam 2: Loss of HC1 Involving 2.3-Chlorlne Shift at the Benzene
Hwltui
Example:
o
o
o
o
~ HCI
2,4,6,2',4*,6'-Hexachloroblphenyl
1,3,4,7,9-Pentachlorodlbenzofuran
46
-------
3. Kechanlaa 3; Loss of OethoHCl
Example;
2,3,5,6•TetrachlorobIphenyl i,2,4-Trlchlorodlbenzofuran
<> Hgghan'saJt: Loss of Ortho-M2
Example:
CI
3,4,5,3',4»,5'-Hexachloroblpheny1
2.3,4,6,7,8-Hexachlorodlbenzofuran
Table 26 summarizes the result* ob?erv«d by Buser arid Rappe, Including
each chlorinated blpheny I t * otic r , Its chlorinated dlbenzofuran reaction
products, and the associated theraochemleal reaction mechanism.
EPA, through a contract with Midwest Research Institute, Kansas City,
HO, has recently conducted several studies to evaluate thermal degradation
products from dielectric fluid using a bench scale thermal destruction
system. ( 7 * The purpose of this study was to optimize conditions for PCDF
formation In ordi-r to examine the potential for formation of PCDFs and PCDDs
from combustion of selected PCB- containing dielectric fluid#. The first part
47
-------
Table 26. Thermochealcal Conversion of PCBs Co PCDFs^® ¦ *
PCI COMCENER PCDP REACTION PRODUCTS REACTION MECHANISM
-\ 1 .5-Tecrachloroblphenyl 1.2.3,4-TCDFa 4
2, 3 ^-TrlCDF4'^ 3
i 5,6-Tetrachlorobiphenyl 1,2,4-TriCDFa 3
2.6.." . 6 ' ¦Tecrachloroblphenyl I,4,9-TriCDFa»b 2
1,9-DlCDF® 1
2 . 3 .. 3 . 6 -Pentachlorobiphenyl 1,2.3,4-TCDFa 3
J «.i,2',5'-Pencachloroblphenyi 2.3.8-TriCDF^ I
2, «, 5.2' .5"•Pentachlorobiphenyl 2.3.6.8- and 2,3.4,8-TCDF® 2
. . v . 5'-Pencachloroblphenyi 2.3,6.9-* ami 1,3,4,8-TCDF® 3
2 . .S.25 *-Pencachloroblphenyi 1,3,4.6,9-PeCDF* 4
2. «.5. 3* .4'-Pencachloroblph-nyl 2.3.7,8b and 2,3,6,7-TCDFa 3
2,4.5.3*.4'-Pencachlorobiphe-iyl 1,3,4.7,8- and 1,3.4.6.7-PeCDF* 4
2 < . * . 2 ' . 4 ' - Pencachloroblphenyi 1.3. 7 -TrlCDF^ I
: u.h.r .4*-Pencachloroblphenyi 1.3.6.7- and 1.3.4,7-TCDFa 2
J . «», 2* .4 ' Pencachloroblphenyi 1. 3,7,9-TCDF 3
2 5 5'• Pencachloroblphenyi l,4,8-a*b and 1,2 , 8-TrlCDF® * b 1
1.4.6.8-ab. 1, 2 , 6,8-a. and 2
1.2.4.8-TCDF®
1.2.6.9-* and 1,4,6,9-TCDFa 3
2. J. -.J' .4' .6' -Hexachloro- 1,4,6,9.®. l,2.6.9-a, and I
Hi phenyl 1,2.8,9-TCDF*
1. 2 ,4 , 8 . 9 •b and 1,2.4.6.9•PeCDFa*b 2
1 . >. V .4\V-Hexachloro- 2,3,4,6.7,8 -HxCDF 4
b iphenyl
5 . j 2 ' , i' .5'-Hexachloro- 2.3.7,8-TC0Fb I
blphtnvl 2,3,4,7,8-TCDF 2
1,3,4,7,8-TCDF 3
2 . " .4' .6'-Hexachloro- 1,3,7,9-TCDF I
bi phenyl 1,3.4.7,9-PeCDF 2
'T. nr .it we Isomer 1 dene 1 f icacIon
j(n- Isomer among reaction products.
48
-------
Table 26. Theraochemical Conversion of PCBs co PCDFs (Continued)
PCB CONGENER PCDF REACTION PRODUCTS REACTION MECHANISM
2-4.5.2',4*,6'-Hexa«- ioro- 1.3.7.8-TCDF*•b 1
hi phenyl 2,3.4.7.9-* and 1,3,4,7,8-PeCDF 2
1,3.4,7,9-PeCDF 3
2.3.4,2',3',4*-H«xachloro- 3.4,6.7-TCDF 1
hiphenyl 1.2.3.6.7 -PeCDF 3
2.3.3,2*.3',5-Hexachloro- 2,4,6.8-TCDF 1
btphenyl 1,2,4.6, 8-PeCDF 3
2.3.4.5.2*,3',4'-Heptachloro- 2.3.4,6»7-PeCDF 1
blphenyl 1,2.3,4,6,7-® and I.2.3.6,7.8-HxCDFb 3
1,2,3,4,7,8,9-HpCDF* 4
1.2.3.4.6.7.8- and *
1.2.3.4.6.7.9-HpCDF*
2,3 4.5,2'.4'.5'-Hepeachloro- 2,3,4,7,8-PeCDFb I
btphenyl 2.3.4,6,7,8-HxCDF 2
1,2,3,4,7,8- and 1,3,4,6.7.8-HxCDFa•b 3
1,2,3,4,6,7,9-HpCDF* 4
2 , 3, , 5 , 2 ' , 3 * ,4 ' , 5' -Octa- 2 , 3,4,6 , 7 ,8- and other HxCDFs I
i-hloioblphenyl 1.2,3,4.6.7,8-HpCDFb 3
OCDF 4
^Tentative Isomer identification.
hMajor Isomer anon? reaction product#.
49
-------
of the experiment was designed to determine the optimum temperature, oxygen,
and residence time conditions for PCDF formation. The feed Into the system v.is
mineral oil spiked with three individual PCB congeners (2 , 3 ,4,6•tetrachlorob1•
phenyl; 3 , 4 , 5 , 3 ' , 4 ' , 5 '-hexachlorob ipheny I, and 2 ,4 , 6 ,2 ' , 4 • , 6 '-hexach loro •
hlphenyl), which form PCDFs by the four reaction mechanisms proposed bv huifr
and Rappe.(29,30,31) Statistical analysis of the results of 33 runs indicates
that both temperature and oxygen have significant effects on the PCDF yield
and that the Interaction between temperature and oxygen Is synergistic. The
results indicate that the optimum conditions for PCDF formation from PCBs are a
temperature of 675°C for > 0,8 second, with an excess oxygen concentration of
8%. The residence time, in the range of 0.3 to 1,5 seconds, did not
significantly affect the yield, although the observed yields of PCDFs were
lower at shorter residence times
in the next part of the study, duplicate test runs were conducted with
mineral oil and silicone oil dielectric fluids containing PCBs (Aroc lor 12 >« >
at concentrations of 0, 5, 50, and 500 ppm An askare 1 fluid containing *«>»
Aroclor 1254/30% tr Ichlorobenzenes, and a non-PCB askarel fluid containing
mostly trichlorobenzenes with some tetrachlorobenzenes, were tested In
duplicate. PCDFs were found In all samples, PCDDs were found in the samples
from the trIchlorobenzene runs and occasionally at low levels in some of th#
other samples. Up to 5,700 ng total PCDFs/ml * spiked fe#d oil or 4|
conversion efficiency (PCBs to PCDFs) was observed tor the mineral oil .ind
silicone oil runs.^Up to 19,000,000 ng total PCDFs/ml feed oil £19 mg/ml)
or 3t conversion efficiency wis observed for the askarel fluid, <7> Statistical
analysis showed a linear relationship for PCDFs formed versus the amount of
PCBs in the feed. Table 2 7 shows the yields of PCDF homo logs from t hi-
different runs. Under the optimal conditions, PCDFs are formed from mim-r,.l
oil or silicone oil contaminated with PCBs at > 5 ppm.
Experiments conducted by Paaslvirta et si . provided evidence to support
Bus»r's results that pyrolys Is of Aroelo-: 1254 produces PCDFs. P.iasivi i r.i
found that the pyrolys Is of Aroclor 1254 at 600°C with a 40 ml/minute ,iir
stream resulted In 0.16% conversion to PCDFs which Is sHghtly lower but of :lu
same order of magnitude as reported by Buser ec al ,(32) The result* <>t
50
-------
Table 27. PCDFs Formed in Combustion Studies *7 ^
ftOVS HoooCDFs DICDFs TrlCDFs TetraCDFs PentaCDFs HexaCDFs HeptaCDFs OctaCDF FCDFs
(ng) (ng) (ng) (ng) (ng) (ng) (ag) (ng) (ng)
Mineral Oil/
.a
-
130
49
NQb
0C
-
180
5 ppm A-1254
-
-
43
23
0
0
-
66
Silicone Oil/
-
.
26
0
90
0
-
116
5 ppm A-1254
.
-
31
9
150
0
-
190
Mineral Oil/
-
-
200
110
39
8.5
-
350
50 ppm A-1254
.
•
140
82
21
2.2
-
250
Silicone Oil/
-
-
290
73
62
0
0
0
420
50 ppm A-1254
-
.
530
640
83
0
0
0
1,300
Mineral Oil/
1,700
0
2,200
690
43
7
0
0
4,700
500 ppm A-1254
.
.
1,300
620
170
13
0
0
2,100
13,000
>19,000
>22,000
5,200
0
0
>90,000
(B)
-
-
81
25
5
0
.
-
110
Lab Blank
0
0
0
0
0
0
.
•
0
a- - Not Analyzed.
**NQ - Not Quantitated.
c0 - Not Detected.
-------
Paaslvi rta ' s GC/KS analysis of the phenolic fraction of Che PCB pyrolvs is
product ar# presented In Table 28
Table 28. Relative Amounts of PCDFs Formed from PCB (Aroclor 125**J
Pyrolys Is and Carbon Filter from a Capacitor F1 '
ISOMER
PTROLYSIS
500°C
PYROLYSIS
600°C
PYROLYSIS
700°C
CARBON
FILTER
Mono-CDF
Dl-CDF
Total PCDFs
Tetra-CDF
Penta-CDF
Hexa-CDF
Tri-CDF
100
0.1
13.7
'8.6
40.0
7.4
0
0,3
12 8
39 3
35.9
11.0
0,8
100
0.1
5.2
29,4
48. 7
15.1
1,5
100
100
51,8
44 , 3
3,9
0
0
0
The major PCDF homologs Identified in the PCB pyrolysis product warp ( in
order of abundance):
• Trichlorodlbenzofuran
• Tetrachlorodibenzofurar
• Dichlorodibenzofuran,
Monochlorodlbenzofuran and hexachlorod ibenzofuran ¦*¦«
-------
Table 29. Concentrations of PCDFs (ppm} Formed from the
Burning of Askare 1 ^
ISOMER A00°C 500°C 600°C 800°C 1000°C
Trl-CDFs N D.1 20 400 15 N D
Tetra-CDFs 5 70 2400 4 NO
Penta-CDFs 4 60 1300 5 N D
'n.O. - Not Detected; detection limit of 0 1 ppm
The second askare1 fluid in Cervason's study, UGILEC T, consisted ot 6u«
UGILEC 141 and 40% triehlorobenzenes (also in the three isomeric fours
described above). UGILEC T is primarily dlchlorobenzene/d ichlorotoluene No
PCDFs were detected from the burning of UCILEC T at temperatures ranging frous
400°C to 1000°C with a detect ion limit of 0 1 ppm.
At the October 1985 PCB Seminar in Seattle, UA, sponsored by the Electric
Power Research Institute (EPRI), Rouse presented data on the formation of
PCDFs from arcing of PCBs .04) j^e study involved the arcing of: (a) a mixrut v
of t r Ichlorobenzene and Aroc lor 1254 ( 1:1 ratio) in oil, (b) Aroc lor I Joi' i n
oil, and (c) trichlorobenzene in oil. The results of the first two afciiu;
studies are presented in Tables 30 and 31. Results of the arcing srudv ot
cr Icwlorobenzene In oil are presented later in Table 40.
The results indicate that although askarels may contain mc.ixur.il>!>¦
amounts of PCDFs, arcing does not increase the concentration of PCDFs in :lu
fluid.
Cooke reported on a simulated transformer fire study using ret tot ill
dielectric Liquids at the EPRI PCB Seminar The purpose of this stui'.v » i>
to determine the anticipated dloxin or furan level in the event of ,t P B
transformer f ire. Cooke used a combustion chamber (20 inches long x 36 inch.-.-,
wide) with an injection post in the combustion region Kerosene with •
silicone was used to set up the nuclei for the catalytic process. The P<'Bn
were Injected at 800°C to cascade through the 350-800°C (optimal condi r i
for PCDF formation) region
53
-------
Table 30. PCDF Concentration Before and After Arcing
of Aroclor 12 54/Tr ichlorobenrrene Mixture
(10,000 ppra, 1:1 ratio) in 0 i I
CONCENTRATION CONCENTRATION
ISOMER BEFORE ARCING AFTER ARCING
(pph) (ppb)
Tetra-CDFs 367 1 ? 3
Penta-CDFs 1,850 L,1U§
Hexa-CDFs 1,790 1,180
HepCa-CDFs 2.960 1,360
Oc ta- CDF 217 171
Total PCDFs 7,184 4,629
Table 31. PCDF Concentration Before and After Arcing
of Aroclor 1260 (10,000 ppm) in O i 1 ^ ^
ISOMER
CONCENTRATION
BEFORE ARCING
'ppb)
CONCENTRATION
AFTER ARCING
(Ppb)
Tetra-CDFs
Penta-CDFs
Hexa-CDFs
Hepta-CDFs
Octa-CDF
Total PCDFs
280
2,000
1, 700
930
170
5,080
2 70
2 . 500
2 ,000
1 ,000
130
5 , 920
Teste were run using Aroclor 1260, 1254, and 1242 at concentrations o
50 ppm. Theoretically, higher chlor 1 nation in the Aroclor should r
higher levels of PCDFs. Cooke found that the PCDF formation is cons i s t «*n: v:;
this theoretical prediction. The conversion factor for PCBs to PCDFs v.i
0 0012 % - 0 . 0 0 3 % , which was low under the combustion conditions. The tot..
TCDF concentration ranged from 100-900 ppm (
54
-------
FORMATION OF PCDF» FROM POLYCHLQRINATED BENZENES
PCDFs can be formed in substantial amounts by pyrolysis of chloro-
benzene® . In 1979, Baser reported the format Ion of PCOFs and PCDDs through
pyroly t ic reactions from chlorobenzenes 00 .'ib) -jable 12 shows the resulting
PCDF and PCDD isomers identified in the chlorobenzene pyrolyzate, In these
experiments, the pyrolysis of tr I - , tetra- , and pentachlorofcenzenes, using
sealed quartz mini -ampoules at 620°C in the presence of air, yielded tetra- to
octa-CDF,
"°°c
x ~ y < 2m
Significant quantities of PCOFs were found in most of the pyrolyzed
samples (see Table 3 3) The likelihood of formation of PCDFs is highly
dependent on the concentration of chlorobenzenes in the reaction system.
As seen In Table 33, significant quantities of PCOFs were formed from the
tri* and tetrachlorobenzenes, and from the combined chlorobenzene sample, In
general, the PCDFs formed had chlorine numbers of 2m-2. 2m-1, and 2m, where m
is the number of chlorine atoms on the reacting chlorobenzene. With the
tr ichloro- benzenes, some higher chlorinated dlbenzo*" rans 'hepta-CDF" were
also observed; presumably, they are formed from higher chlorinated benzenes
produced during pyrolysis,
The optimal temperature for the conversion to PCDFs seef o be 500-
600°C At temperatures exceeding b00•700°C. the degradation of FIDOs and PCOFs
takes place at a faster rate than the rate of formation In audit ion to the
temperature, the retention time is an important parameter. The experimental
results are summarized in .able 34
55
-------
o
r°»
'-
?w
fs-p
w
i ,,
f\j>
,w
r^.
w
W
w
i--
w
4>
W
w
w
n
*
-
*
<*
-
-
*
•
-
-
-
-
53
c-
r
i>*
t
Os
A-
OS
cr*
*»
~i
«
-
•
•
*
*
-
•
-
¦s#
a*
a*
-¦4
Ob
CD
S^l
•si
or*
CD
CD
*sj
-a
CO
OB
OB
m
00
00
<6
sQ
a
X
X
X
X
X
X
X
4
~
»
•
*
t»
ft
9
*
m
IS
m
X
X
X
X
X
X
3
X
X
X
X
X
*
#
9
*
ft
ft
a
ft*
ft
&
&
ft
13
•ts
TJ
*D
«
*
•
*
*
1
•
r?
rt
rf
rf
a
o
i—»
o
o
o
a
a
ft*
ft
ft
fit
o
o
•
o
o
o
o
o
«
»
¦>
«
•n
•n
Ki
•*1
**!
•*1
*n
•v
O
n
O
O
*
O
a
o
O
w
*»1
•n
*n
"*1
£*
&
o
fi
O
O
o
"Si "SJ
CD <0
t*
««l» «¦¦>,
9 *
t> *0
r* r?
ft ft
o n
o o
o o
X
a
X
ft
o
o
•H
*-*
»-*
•—»
#~-
»—
Kd
fO
9S>
K*
«g
K>
fs>
W
w
w
w
*¦
t-
NJ
<7*
C*
x>
K>
w
l>
o
CT
r
VM
w
w
K^p
w«*
w
r-w
£.'-
%**
•-J
l>
£S
4*
as
•^>1
Ofc
|£T*
0D
¦".j
•SJ
•<^J
'•"n!
0^
¦^4
OD
O
CO
ao
•si
00
m
CD
OD
•XJ
"T?
TJ
TJ
•n
"13
13
13
13
»
9
ft
m
»
A
It
m
3
3
3
3
3
3
3
3
3
3
3
r?
n
o
rt
rt
n
r»
a
n
rt
rt
r?
rt
gn
»
3>
Si
ft*
£»
ft*
ft
ft
ft
ft
ft
>
¦
¦
¦
¦
*
*
»
*
*
•
*
O
O
O
o
n
O
O
.
O
o
n
n
n
O
a
<~
o
o
o
o
ru
o
o
o
o
o
•*5
•TJ
*n
•*i
•*3
nr)
w
•»J
**s
•n
*T>
o
00
•u
m
3
n
ft
o
o
»-•
i—»
r-»
•—*
#-»
#-*
K)
fO
K»
to
w
W
u»
Ul
Ul'
V
oo
"si
'si
•sj
0*
a*
¦x>
'si
a»
¦£>
OO
oo
00
tl
TP
u
U
u
9
»
0
f»
#
#
9
D
3
3
9
3
3
3
r»
n
rt
rt
fT
rt
r»
ft
ft
ft
k
ft
ft
ft
O
O
O
O
o
a
n
O
O
o
C3
o
o
o
o
o
o
O
o
o
o
o
*1
*0
sji
¦ m
w
#
f'w
w
w
r-»
•-•
r-*
N*
w
w
W
w
U>
w
'Ni
00
¦si
a»
SI
•sj
00
vC
00
¦SJ
CO
00
H
-4
H
H
HI
H
9
9
s
9
9
©
rt
rf
rt
r*
rt
rt
rt
n
•t
n
n
1
•-I
ft
ft
ft
ft
ft
ft
o
O
n
n
a
O
O
o
o
o
o
o
O
o
o
o
o
o
o
o
o
^1
V)
-------
Table 33. Format Ion of PCDFs from Che PyrolysIs
of Chlorobcnzenes ^ ^0 • ^)
COHPOUND(S)
PCDFs
FORMED
(ng/saraple)
TETRA-
PENTA-
HEXA-
HEPTA-
0CTA
Tr Ichlorobenzenesa
400
1100
550
50
< 5
Tetrachlorobenzene«b
< 2
5
160
450
200
Peruachlorobenzenec
< 2
< 5
< 5
5
30
Combined chlorobenzenesd
80
600
1100
600
60
a200 meg tocal with equal amounts of 1,2,3-, 1,2,4-, arid 1,3,5-crichloro-
benzene.
b200 meg cocal vl ch equal amount# of 1,2,3,4-, 1,2,3,5-, and 1, 2,4 , 5 - tec ra -
chlorobenzene.
c200 meg pencachlorobenzene
^300 meg total with equal amounts of all cr I • , terra- . and pentarhlo r ob.-ru:. i >«•s
1? compounds)
Table 34 Summary of Formation of PCDFs from the Pvro I vs i s
of Ch1orobenzenes ^ ^1
COMPOUND(S) PYR0LYZED PCDFs
YIELD
1.2.3- . 1,2,4-, and 1,3.5 -Tr Ichl orobenzene 1 **1
1,2,3,4-, 1,2,3,5-, and 1,2.4.5 -Tet rach1orobenzene 0 i*b
Pertachlorobenzene 0,02*
ntetr.i - CDFs 0 2% bhtxa-CDFs 0 U
pcnca ¦ CDFs 0 .51 hepta-COFs 0 .!»
hexn•CDFs 0 3% octa-CDF 0 2I
57
-------
Further evidence to support the formation of PCDFs fron the pyrolvs i s of
chlorobenzenes can be found In the results of the Midwest Research Institute
study ^ As mentioned earlier, this study included the pyrolysis •»£ zri-
chlorobenzene dielectric fluid at 675°C for 8 seconds wit i 8% excess oxygen
Up to 110,000 ng total PCDFs/ml feed oil (> 0.004% yield) was observed for the
trichloroben7ene runs. These results, presented in Table 27, support earlier
laboratory data and analytical results of soot material from transformer and
capacitor fires,
Paasivirta «»c a2 . , studied the pyrolysis of a tr Ichlorobenzene/PCB
mixture. The details of this study were described in this section, and the
relative amounts of PCDFs formed from the pyrolysis of the mixture were
presented in Table 28 . (^2)
The study conducted by Cervason, described earlier in this section,
provides evidence in support of the formation of PCDFs from trlchlorobenzenes
This study indicates that the burning of a 60* PCB/40% trIchlorobenrene mixture
results In the formation of PCDFs. C' The results were presented in Table 2)
FORMATION OF PCDFs FROM CHL0R0PHEN0LS
In Buser's experiments, the pyrolysis of chlorobenzenes yielded a series
of chlorinated compounds Including PCDFs, PCDDs, chlorophenols, and in sonse
cases PCBs polychlorinated naphthalenes, and polychlorInated stvrenes ^1
All pyrolyzed samples showed the presence of chlorophenols Baser suggestcd
that these chlorophenols could possibly serve as reaction Intermediate;, in :lu-
fo mat ion of PCDFs from ch lc robenzenes A paction of chlorophenol -irh
unreacted chlorobenzene could lead to the formation of polvchlorinated tliph.-nv!
ethers (PCDPE). which are known to form PCDFs upon pvro1vs i s (^ * 1
Test burns of pentachloropheno I (PCP) waste were performed In a
U.S. industrial boiler. Baghouse ash and bottom ash were analyzed and results
are presented In Table 35. The total level of PCDFs In the baghouse ash wis
2.9 mcg/g (37)
58
-------
PCDPE
PCDF
Table 35 Levels of PCDFs from Burning of Pentachlorophenol-
Contanlnated Uaste^
CONTAMINANT &ACHOUSE ASH BOTTOM ASH
Cppra) (Pl»)
Tetra-CDFs
0 90
N.D
2 . 3. 7,8-Tetra-CDF
0 10
N. D
Penta-CDFi
1.5
N.D
1,2.3,7,8-Penca•CDF
0.05
N.D
2,3.4,7,8-Penta-CDF
0 10
N.D
Hexa-CDFs
0 15
N.D
1.2,3,4,7,8-Hexa- CDF
0 02
N.D
Hepta-CDFs
0 06
N.D
Octa-CDF
0.006
N.D
aN D - Hoc Detected.
FORMATION OF PCDDs FROM PCBs
There Is no experimental evidence for PCDO formation from pvro 1 \m •» «»:
PCIs, and theoretical considerations indicate that this conversion dm- * ~!,>:
occur. The thermolysis of PCBs yields exclusively PCDFs with n«> h !)')>
de tec ted. ()
59
-------
FORMATION OF PCDDs FROM POLYCHLQRINATED BENZENES
Busitr has reported chat PCDDs are formed from the pyrolysis of chloro-
benzenes O0. 36) T»
-------
Table 36, Formation of PCDDs from the PyrolysIs
of Chlorobenzenet
PCDDs FORMED (ng/sample)
COMPOUND(S)
TETRA-
PENTA- HEXA-
HEPTA-
OCTA
Tr Ichlorobenzenes3
30
20 <5
< 5
< •)
Tetrachlorobenzenes^
< 2
5 1«*0
160
10
Pencachlorobenzene0
< 2
A
A
< 5
)
Combined chlorobenzenes^
50
'i 20 220
70
i
a200 meg total with equal
benzt ».
amounts
of 1,2,3-, 1.2.4-
, and 1
,3,5 * tr i
^200 meg total with equal
chlorobenzene.
amounts
of 1,2,3,4-, 1,2,3,
5 -, and
1.2,4.^
c200 meg pentachlorobenzene.
^500 meg total with equal amounts of all tr I - ard tetrachlorobenror,. s
pentachlorobenzfno (7 compounds),
Table 37, S^imary of Formation of PCOOs from the Pyrolysis
of Chlorobenzenes^® ¦
COMPOUND(S) PYROLYZED
1,2,3-, 1,2,4 , and 1,3,5-Trlchlorob*nzene
1,2,3,4-, 1,2,3,5-, and 1,2.4,5-Tetrachlorobenzene
Pentachlorobenzene
PCDD
YIELD
0,02 3*
0 15*
0.002%
61
-------
Table 38, PCOOs Formed in Combustion Studios
RUNS
NonoCDO D1CDD TrlCDD TctraCDO PcnCaCOD H«mCDD HeptaCDD OctaCDD FCDOs
(Dt) (ng) (ng) (ng) (fig) (ng)
-------
transformer arid capacitor fires, which determined chat chlorobenzmes are
required for PCDD formation.
As discussed earlier In this section, Paaslvirca ec j1 , studied tl.c
pyrolysls of a mixture of Aroclor 1254 and ehlorobenzenes. Such mixtures
usually contain anywhere from 13% to 601 t r ichlorabenzene and som* t imes
cetrachlorobenzene. ^2) The results of the CC/MS analysis indicated tli.it PCDDs
were present In the pyrolysls product. Table 39 gives the relative amounts of
PCDDs formed from pyrolysls of this PCB/chlorobenzene mixture.
Table 39. Relative Amounts of PCDDs Formed from Analysis
of a PCB/Chlorobenzene Mixture^"*
PYROLYSIS PYRCLYSIS PYROLYSIS
ISOMER I00°C 600°C 700°C
(%) (%) ft)
Dl-CDD 17.2 0.7 4.6
Trl-CDD 10.0 18 4 13 6
Tetra-CDD 72.9 57.0 66 3
Penta-CDD 0 24.1 17 4
Total PCDDs 100 100 100
The fif.Jor PCDDs detected In the pyrolysls product weie tetra-CDD (highest
concentration) and penta-CDD. Tie pyrolysls of the Aroclor 1254 mixtuiv
resulted in a 0.161 conversion to PCDDs, which Is lower than Buser's esciin.ic«<
but of the same order of magnitude , O?)
PaasIvlrta found high amounts of polychlorophenols and polychlorln.itcd
b I pheny 1 s In the pyrolysls products of the PCB mixture and a capacitor
fire
As discussed earlier. Rouse presented data at the 1985 EPRI PCH Serein,ir in
Seattle, UA. on che formation or PCDDs from arcing of crlchlorohenremt in
The PCDD concentrations resulting from this arcing srudv ,»r,
presented In Table 40
63
-------
Table 40, PCDD Concentration Before and After Arcing
of Trichlorobenzene (10.000 ppm) in 011 )
CONCENTRATION CONCENTRATION
ISOMER BEFORE ARCINC AFTER ARCING
(ppb) (ppb)
Tetra-CDD N D.a SO
Penta-CDD 0.4 N.D.
°N D. - Not Detected.
Gervason's study. previously described In this section, indicated the
formation of PCDDs from Askarel T^ (60% PCB/40% trichlorobenzenes i»ixcui e >
The results are presented In Table 41
Table 41. Concentrations of 2,3.7,8-TCOD Formed from the
Burning of Askarel Tj^ "
TEMPERATURE
(°C)
400
500
600
800
1000
2,3,7.8-TCDD CONCENTRATION
(ppra)
1.5
1.0
15
0.1
0.15
FORMATION OF PCDDs FROM CHL0R0PHEN0LS
While investigating the pyrolysls products of chlorobenzenes Bu*.-r
found that chlorophenols were present In all of the pyrolvzed chloiis.
samples .16) These chlorophenols could possibly serve as i>,n i i ..i,
intermediates In the formation of PCDDs from chlorobenzenes. A reaction a
chlarophenol w 1 r i i nn reacted rhlorobensene coul I 1 c.ul to the form.it h»i.
64
-------
polychlorInated dIphenyl ethers (PCDPEs), which are known to for* PCDOs upon
pyrolysls30,36)
Rappe «c a I . , reported PCDD formation from the burning of chloro-
phenols.^®' Two of the most commonly used chlorophenol formulations on the
Scandinavian market, Servarex Teknlsh (Gulivlks Fabrinks AB, Malmo, Sweden,
lot Npnr 524) and Kyminene KY- 5 (Kyttunene OY. Kuusankoskl, Finland; lot 132/16),
were used. These formulations contain approximately 51 of 2.4,6-crl-, 501 of
2,3,4,6-tetra-, and 10* of pentachiorophenol In their sodium salt form. The
levels of PCDDs found in the two formulations are similar.
m
i-1
PCDD
PCDPE
m
Another route of PCDD formation Involves the dlmerlzatlon of
chlorophenols, although the condensation via PCDPEs nay be preferred.
PCDD
2
Test burns of pentachiorophenol waste were performed in an Industrial
boiler In the US Samples of baghouse ash and bottom ash w»re analyzed and
results are reflected In Table 42. In the baghouse ash the total level of
PCDOs was found to be 5.3 ppo Lower chlorinated PCDDs predominated. In che
bottom ash, che higher chlorinated PCDDs were found In greater amounts •5
65
-------
Table 42. Levels of PCDDs from Burning of Pentachlorophenol-
Concaninaced Vast?-*7*
CONTAMINANT
BAGHOUSE ASH
-------
SECTION 5
CONTAMINATION SITUATIONS INVOLVING
PCDDs FROM PCBs AND CHLOROBENZENES
This section reports some situations of facility contamination by PCDDs.
generated from thermal processes involving chlorobenzenes and/or polybroininated
b I phenyls. As discussed earlier. under the proposed rules for restricting
hazardous waste In landfills. PCDDs -contaminated sites will b<» subject to
substantially more stringent cleanup and dlposal regulations The following
serves to highlight examples of contamination sItuatIons where such regulations
might apply.
PCDD FORMATION FROM 1.2.4-TRICHLOROBENZENE FEEDSTOCK FROM LINDANE PRODUCTION*
A European manufacturing process for trichlorobenzene provides further
evidence of dioxln contaminant formation from chlorobenzene. This process
converts lindane manufacturing wastes ( a ,/9 . & -hexachlorocyclohexane, MCCH) to
predominantly ( 70- 75%) 1.2,4-1r ichlorobenzene. which, after purificatio.i is
transformed to 1 . 2 , 4 . 5 -1 e t rach 1 o robenzene (used as the feedstock tut
1. 4 . 5 -1r ichlorophenol) , The process involved the low temperature 200 • :>
pvrolysis of the lindane wastes with a carbon catalyse using indirect he.it in,',
of a closed reaction mantle. The residues from both the mantle and U,<-
11 Ichlorobenzene still were analyzed by Buser in Switzerland The results o:
this analysis are presented in Table 43.
Extreme ca.e should be taken when pyrolvzing and/or distilling chlorinated
.ii omat ics to exclude air (oxygen) from the syste so that undes i r.th 11>
chlorinated contaminants. includlng PCDDs. will not be formed
PCDDs FORMATION FROM A PITTSBURGH AREA METAL RECLAIMING FACILITY
PCDDs and PCDFs have also been found at active/inact ive scr.ip h. ; .1
reclaiming facilities, whe-e PCB transformer and capacitor units ate proc.•»« J
for their respective copper values. One such example In the Pittsburgh. Pa.
area Is a 1/2 • ac re site In the midst of a residential community, with.
67
-------
1,000 persons within a l/4-mlle radius. This sice Includes a commercial
building containing a snail incinerator chat was useJ not only to combust PCBs,
but also to provide heat to the surrounding area. In August 1982, NIOSH
conducted an Inspection and detected 8 ppm octa-CDD in a sample of soli taken
at a depth of I foot near Che PCB . .quid storage area. In April 1984, EPA
Region III staff removed and analyzed dust samples from inside the building,
the results of which are shown In Table 44.
As a result, an Immediate Removal Action was init iated on Mav 2'J, 1 "'Hi
An extension survey demonstrated that Che dloxln/furan contamination v.is
restricted to the interior of the incinerator building More than 5.U00 : 50 ppm i'CBs , were excavated. loaded, and
transported for disposal to the CECOS facility In Niagara Falls, NY, before the
RCRA Dloxin Listing Rule went Into effect (on July 15, 1985) Scrape samples
from Interior building surfaces showed:
Tetra•CDDs 10-94 ng/g or ppb
Tetra•CDFs 38-615 ng/g or ppb
Sixty transformer casings remain on site today The estimated cost tor
the disposition of the remaining contaminated materials (I.e., PCBs, PCDOs, ami
PCDFs) and security of the site on a temporary or interim basts is $' Hm
million »• for a 1/2 - ac re site.
PCDD AND PCDF FORMATION FROM STRANDLEY METAL RECOVERY SITE, PURDY. WA
Strandlev Scrap Metal occupies 65 acres, Including I to 2 acres of
work areas, along Highway 302 east of Purdy, WA The S t r.md lev .'Hani! i <\,\ i «
contains several buI idings/warehouses . an 8,000 gallon storage t auk ",.f
railroad tank car), transformer casings, a large wire pi le. scrap «•«*?.11 .
forkllfts. and several drums. Transformer salvage operations occurtvd :!.»
site between 1972 md 198 3 Some materials may have been burned u : h«
s i te.^
The St rand ley/Manning Property site is known to contain PCBs conrami nati.m
in the soli and some remaining containers, and there Is evidence of oil
hS
-------
Table 43, Dloxln Isomers from Pyrolyzed HCCH Resldu. ;
and Trichlorobenztne Stlllbotton»#(39)
DIOXIN ISOMER v.ONCt V. > \TI0N (ppa)
Tetra-CDDs 12
2.3.7,8-Tecra-CDD 03
Penta-CDDs
1,2,3.7, 8-Penta-CDD 7
Hexa-CDDs 68i*
1,2, 3,4. 7 . 8 - Hcxa-CDD 4 >
1,2,3,6,7, 8-Hexa-CDD 150
1,2,3,7,8»9-Hexa-CDD 65
1,2,3.4,6,7.9-Hepta-CDD 1.400
1.2,3.4.6,7.8-H«pta-CDD 3.000
Occa-COD 7.600
Table 44, Dloxln and Furan
-------
contaminant migration Dloxlns and furans have also been detected, posslblv as
a result of past burning of oil and transformer wtndlngj
Prior to the sampling for a preliminary sit® assessment, EPA had a I read*,
performed some sanplinj PCBs were found in the storage tank, in soils, and ii,
cream sediments. Very low lewis of PCBs w#td found In the shellfish and
sediments of nearby Burlev Lagoon. Results of swab sampling of ons i ? «•
equipment Indicated some PCBs contamination. Also, 2,3,/,(?¦TCDD was found in a
stove ustfd to burn PCBs- contaminated oil. The presence of several other diosin
and f uran isomers has also been detecte I. Limited excavation of soils with
high levels of PCBs near the old storage tank location resulted in finding
significant PCBs contamination at a depth of 6 feet.
The following analyses were perfomed on the various samples;
• PCBs ifor soil, sediment, drum. and swab samples)
• PCPDs 'PCDFs (for some soil and swab samples)
Total PCDDs
- Total PCDFs
- 2,J.7.8-TCDD
• 2 7,8-TCDP
All cetrc- thro ugh octa-CDDs (total homologs, not isomer•specific*»
All tetra* thrc igh octa-CDFs (total homologs, not Isomer•specifie)
• Halogenated Organics.
A summary of sample results fot PCBs analysI > is presented in Tables
and 4 6.
70
-------
Table 45, Suawary of Sample Results from
Strandley Scrap Metal Slte^*^
SAMPLE MEDIA PCI CONCERTSATION (pptn)*
Soil
136
Soil
27
Sol 1
3?
Soil
1200
Oil Sludge
5600
(wet b?- Is)
Soil
3?
Soil
63
Soil
28
Soil
230
Soil
46
(plus 49 of Aroclor 12~»J i
Soil
62
(plus 10 of Aroclor 12«.'5 ~
Soil
73
Soil
58
(51 duplicate)
Oil
120
(wet basis)
Soil
162
Resldue
4
4
Soil
m
Soil
30
Soil
26
Sediment
88
Sediment
4
. 7
Vate r
2
47 mcg/l
Soli
190
Soli
4
6
Soil
91
(Aroclor 1254)
Soil
71
Empty Sample Bottle
N. D.
Water - Transfer blank
0.069 meg/1 Aroclor 1254
Water - Transport blank
0.42 meg/1 Aroclor l ?t>«
*'PCB concentrar ion Is In uniti of mcg/g (ppm) of Aroclor 1260 on a dr . ;
basis unless otherwise stated.
Li. - Not Detected.
71
-------
Table 46, Sumnary of Sample Resales Iron Strandley Sice and
Bur ley Lagoon^^
SAMPLE MEDIA
DESCIIPTIOH
Tissue
Water
Sediment
Sediment
Sediment
SedIment
Sediment
Sediment
Sediment
Water
Sediment
Sediment
Water
Sediment
Water
Water
Muter
Water
''.iter
Water
Water
Water
Sediment
Sediment
Sediment
Oil
Sedi n.ent
Sediment
Scdluont
Shellfish sample
1 Ine
line
PC# CONCENTRATION
*
Transport blank
Sediment blank
Transport blank
Stream near mouth.
40* up from tree
Stream near mouth
40* up from tree
Stream about 140* up from
tree line
Stream 20' downstream from
culvert below pond
Stream 20* downstream from
culvert be low pond
Sediment from spring
House 1/2 gallon clear
House 1/2 gallon amber
Spring 1/2 gallon clear
Spring 1/2 gallon amber
Transport blank 1/2 gal Ion clear
Transfer blank 1/2 gallon clear
Transport blank 1/2 gallon amber
Transfer blank 1/2 gallon amber
South side access road to Strandley
yard, composite of two namples
North side access road to Strandley
yard, composite of two samples
Auto body yard, composite
of four samples
Lift truck hydraulic reservoir
Oily rag on lift truck
Stream sediment by drain
west of pond
Spring composite below yellow bus
Driveway parking lot by Willow Road
composite of four samples
Sediment blank
0 051 (wet has
1 5 ppb
2 6
15.3
0,43
0.51
0,026
0 072
N.D.b (0 04 ppb)
N.D. (0 001)
N D (0.03 ppb>
67
0.68 ppb
88
102
0.69 ppb
0. 102
N.D, (0.02
ppb)
N.D. (0,02
ppb)
N.D. (0,02
ppb)
N.D. (0.02
ppb)
N.D. (0 02
ppb)
N.D (O.fC
?pb)
N.D (0 ^
ppb)
N.D. (0.02
ppbi
45
178
56
NO. (10)
13
93
6.3
4
N.D
5
(1.0)
Units (ire in ppm (mcg/g; dry weight of Aroc lor 1260 unless n:
spec If led
''N.D. - Not Detected; value In parentheses Is the minimum quant 11 lab Ic » i
72
-------
fable 46, Summary of Sample Results from Strandley Site and
Burley Lagoon (Continued)
SAMPLE MEDIA DESCRIPTION
Sediment
Sediment
Sediment
Sediment
Sediment
Sediment
Sediment
Sediment
Sediment
Station »l
Station »l
Station »2
Scatlon »2
Station »3
Station »4
Station »5
Station i»6
Station ml
Sediment Blank
PCB CONCENTRATION
(ppn)*
0.0100
0.02044
N.D.b (0,01)
NO (0.01)
0.019
N D. (0.010)
0.026
0.36
NO (0 010)
NO (0.010)
Tissue
Raft Culture Pacific Oysters
0 082
(wet baslsl
Tissue
Mid-Bay Touglng Pacific Oysters
0.145
(wet bas:s >
Tissue
West Beach Japanese Little Necks
N O. (0
.016) (wet b,isi>
Tissue
South Beach Japanese Little Necks
0.016
(wet basis)
Tissue
Plant Site Pacific Oysters
0,11
(wet basis)
Tissue
Nortueast Lagoon Macoma Clams
0 048
(wet basis;
Tissue
Northeast Lagoon Ci.osC Shrimp
0,26
(wet bas l s >
Tissue
Northwest Lagoon Macoma Clams
0.053
(wet basis
Tissue
North Lagoon Ghost Shrimp
0. 78
(wet basls»
Tissue
North Lagoon Small Clams
0.49
(wet basis»
Sediment
Station
•8
0 034
Sediment
Station
•9
0 04 7
Sediment
Station
• 10
0.050
Sediment
Station
• 11
0.061
Sediment
Station
• 12
0 035
Sediment
Station
• 13
0 047
Sediment
Stat Ion
• 14
0 078
Sediment
Station
• 15
18 I
Sediment
Station
• 16
40
Sedalent
Station
¦ I?
16 2
Sediment
Station
• 18
3 1
Sediment
Station
• «9
0 25
Sediment
Stat Ion
¦ 50
0 45
Sediment
Station
•21
0 24
Sediment
Station
• 22
0 . 365
Water
Water transfer blank
ID (0
01 ppb)
Water
Water transport blank
N D (0
01 ppb \
Sediment blank
N D. (0
01)
Water
Scat ion
• 15
5. 16 ¦
ppb
Hulls are In ppm (mcg/g) dry weight of Aroclor 1260 unless otlu-t
spec Ifled.
D - Not Detected; value In parentheses Is the minimum quantifiable 1 In-;:
71
-------
California Analytical Laboratories analyzed numerous soli samples from the
Scrandley Sic*. The results of their analyse® are presented in Table 41
Table 47, PCDD/PCDF Analysis of Soli Samples From Strandley SI c e 1 ^ ^
SAMPLE ISOMER CONCENTRATION
Cng/g or ppb)
1 Hepta•CDF 0 96
2 2.3.7,8-Tetra-CDF 0.16
Total Tetra-CDFs 0 23
3 None N.D.l
*lN D. - Not de tec ted
Samples from four different areas of the sice were also analyzed for
PCDDs/PCDFs by California Analytical Laboratories. Table 48 f.iv»tin-
concentration of PCDDs/PCDFs detected In these different samples
Table 46. PCDD/PCDF Analy*It of Various Samples
from Scrandlcy SlC®<42>
SAMPLE DESCRIPTION ISOMER CONCENTRATION
(ppb)
Sample of Buri\ Area Octa-CDD 8 1
from Under Concainer
Sample of Ash on Bark 2,3,7.8 -Tetra- CDF 3 5
of Roadway Total Te t ra -CDFs 13
2.J.4,7,8•Perta-CDF 4 '
Total Penta•CDFs 62
1,2,3,4. 7.8- Hexii - CDF 38
1,2,3,7,8, »H«xa - CDF 14
2.3 4,6.7.8-Hexa- CDF 17
Total Hexa-CDFs 91
74
-------
Table 48 PCDD/PCDF Analysis of Various Samples
from Strandley Site (Continued)
SAMPLE DESCRIPTION ISOKE* CONCENTRATION
(ppb)
1.2.3,4,6.7.8-Hepta•CDF W
Total Hepta-CDFs 90
Octa-CDF 55
Total Tetro-CDDs N D a
Total Penta-CDDs 4 2
Total Hexa-CDDs 1.8
1,2.3.4,6.7,B-Hepta-CDDs 12
Total Hepta-CDDs 25
Octa-CDD 43
Sample of Ash fron 2,3,7.8-Tetra-CDF 160
Stove in Shop Total Tetra•CDFs 545
1 .2.3. 7.8-Per.ta-CDF 81
2,3,4,7,8* Penta-CDF 110
Total Penta-CDFs 2000
1,2,3,4,7,8«Hexa-CDF 7 33
1.2.3.7,8.9-Hexa-CDF 530
2,3.4,6,7,8-Hexa-CDF 38
Total Hexa-CDF# 2300
1.2.3.4.6.7,8-Hepta-CDF 1500
Total Hepta-CDFs 2000
Octa-CDF 320
Total Tetra CDDs 173
1.2.3.7,8-Penta-CDD J I
Total Penta-CDDs 845
1 .2.3.4.7,8- Hexa • CDD •'1
Total Hexa-CDD* 1100
'% D - Not Detected.
I some rs co -e lute .
75
-------
Table 43. PCDD/PCDF Analysts of Various Samples
froffl Strandley Site (Continued)
SAMPLE DESCRIPTION ISOMER CONCENTRATION'
-------
SECTION 6
ANALYSIS AND EVALUATION
In the aftermath of a PCB transformer fire, the usual Immediate actum is
to analyze the soot generated In the fire for toxic combustion products formed
In order to a* ess the immediate health risk that may be present in : lie
building or fac ty. Much has been learned over the last few years from
investigating the PCB transformer fire incidents that have occurred throughout
western Europe and the United States. The issues that the research community
hits focused on appear to be following:
• Are PCDDs and PCDFs formed In PCB transformers under normal ap« «tim*
condltlorj?
• How are the constituents of the transformer fluids re la tori to :(»<• t.p«-
«nd amount of PCL/Ds and PCDFs formed?
• What are the temperature and other environmental conditions that t.ivn,
the formation of the PCB combustion products'
Answers to such questions are necessary to properly assess PCB transf mu-r
f I re incidents and to deve lop appropriate emergency response and longer terr*
remedial measures. This section of the report addresses these questions bv
correlating, ana'/zlng, and evaluating the data presented in the earlier
sec t ions.
COMPARISON OF A PCB TRANSFORMER FIRE UITH THE INCINERATION' SITUATION
The combustion products resulting from PCB fires are .» furction »l ir.iitv
•milables including:
• Carbon-to-hydrogen ratio
• Chlorine content or weight percentage chlorine
• Type of contaminants.
• Reaction (fire) temperature
• Residence time of reactants (air and .ue1)
• Turbulence or mixing efficiency ot fuel and air.
• Atmospheric versus reduced oxygen environment
77
-------
The combust: Ion and molecular decomposition of chemicals in a PCB
transformer fire Is not unlike that of the combustion situation In the
Incineration of chlorinated substances. The general equation for complete
combustion of chlorinated hydrocarbons during Incineration can be written as:
l) CaHbClc + p02 ~ 3 78 pN2 + mH2Q »
3 78 pN2 f (m ~ n/21 H20 + aC02 + (b - n) HC1
~ 1/2(n + c « b) Cl2 «¦ (p - a - n/4) 02 + Qkcal.
The atoms of chlorine will be olstrlbuted between HC1 on one hand, and
stable chlorine molecules on the other hand. This Is consistent with
DEACON tqulllfrrtvffl in Ilia fMEMgf'S CPU?, after proper combustion. The
equilibrium state is reached with a kinetic of each partial reaction,
different from the one of proper combustion reactions.
This equilibrium state can be written as:
I
2) 4 HCl ~ 02 , * 2 H20 *¦ 2 Cl2 • 20.8 kcal (exothermic)
II
If mHjO, nCl2. o02, qHCl, and N are the number of molecules of water,
chlorine, oxygen, and hydrochloric acid, and the total number of
molecules In the combustion gases, respectively, the equilibrium
constant, In the average conditions of a homogeneous furnace, is:
N
-------
co decrease the chlorine's partial pressure In the gases after combustion, the
following conditions have to be fulfilled:
• A high temperature of combustion
• A furnace designed in such a way chat the residence rime of rhe )•,.!•>«>s
at che high temperature Is large enough co enable DEACON's <-quilibrit mi
and complete combusc ion to take place
• A water vapor partial pressure that Is as large as possible, taking
Into account the thermal and water balances of the system
• A very efficient "quench" device to "freeze" the above said
equlllbrlura.
A PCB transformer t Ire, however, Is not a control led combuscIon situation
such r.s that In an Incineration furnace; bur It can be assumed chac at the
point of combustion, the thermal properties of the transformer fluids (vapor
pressure, enthalpies and entropies, etc.) and generated products will deteinir.e
rhe cype of combustion and products generated Unfortunately, because of rhiir
hi^h toxicity, the thermal properties of many PCOFs, PCDDs, and othei cosic
chlorinated combustion products have not been measured.
Rordorf recently developed an approach to esclmace rhe vapor pressure,
boiling point, and other thermal properties of toxic substances such .is
chlorinated dloxlns •/<2) The scheme Is shown In Figure 1
Vapor pressures of liquids of known freezing and boiling points can H»
predicted by equation VII, where K Is from equation V, Equation VII can a! so
be used to estimate boiling points If the melting points and vapor pressures .it
Tm are known for given compounds Equation VII Is first used to pr.ilUt
boiling points of the Invescigaced chlorodloxlns and furans
The enthalpies of fusion need co be known for che projeccion of sol d
vapor pressure curves from (predicted) IIquld vapor pressure data. The i-jbiis
free energies and enthalpies of sub I mac Ion equal the sums of che correspond r. \
funcc Ions of fusion plus melclng (VIII and IX) Equation XI. which is ob: .1 i in* I
from VIII, IX, and X, states chat similar sums hold for rhe ent rop i«-;. Hi.
entropies of fusion at T, equal the enthalpies of fusion divided bv t he ir«-11 n>«-
temperatures (XIII) as the Cibbs free- energies of fusion are zero .it f|n. i,.
79
-------
Figure 1. Relationship Between tie Vapor Pr@sst*e$ of tie Sold and liquid Ptiases
at Tempera**®® T, the Meting Point (Tm). and the Boiling Point (Tb>(42)
6hy(Tb )
A8v(T b)
, A^CTb,). **t(Tih). Ah
IB | A8w(Tni), ^SnivTm), AS,
^•K»u»d = "VoUd
1/T (K)
-------
Figure 1, Relationship Between the Vapor Pressures of the Solid and Liquid
Phases at Temperature* T, tha Melting Point (Tm), and the Boiling
Point (Tb) (Continued)
kf the Rolling Point
Ahv(Tb) - Tb Kp (36.61 + R In Tk)
A«v(Tb) - 4hv(Tb) / Tb + R In P®
In Che Liquid Phase
Ahv(T2) - Ahv(T^)
•J
•t2
A gaa-liq.dT
(I)
(II)
(HI)
gives after integration for - Tb and T2 - T:
(IV)
(V)
Ahv(T) - Ahydb) (1 ~ K(l - T/Tb))
with K - - Tb Acp&asll<* / h..(Tw
/ hv(Tb)
h - molar enthalpy
a - aolar entropy
q - aolar glbbs free
energy
Cp - aolar heat capacity
Tb - boiling point (K)
Tb - melting point (K)
R - 8.314 J/no 1 K
P(T) - vapor pressure at
T(K>
Subscript,: subllaatlon
Subscript^,: vaporization
Subscript^: fusion
Setting (IV) Into the Clauslus-Clapeyron equation,
•P°
d In P(T)
(VI)
fTb Tb -
- I Ahydb) (1 ~ K(l - T/Tb)) / R T2 dT
gives after Integration froa the nelt Ing to the boiling point:
In P(Tn) - in P° ¦ (Ahvdb) / R Tb ) <(l ~ K)(Tb/Ta - 1) - K In db/Tffl)) (VI1)
J
b% th« HtIcing Mitt
Fron Ag9 - ~ Agv (VIII),
and 6g - Ah - T As (X) follows
For T - Tm: Agn (Tm) - 0 (XII) and
- Ahs - Ahv
" Ass
As.
(IX)
(XI)
A8rB(Tm) - AhBdB) / Tffl (XIII)
Fron Agv - • R T In P^- Ahv - T Asv
follows for T - T„ Asvdn) - R In P Uq (Tm) ~ Ahv(tffl) / T„
(IV) for T - Tn: Ahyd.) - Ahv(Tb) (1\ K(1 - T./Tb))
In the Solid Phase
Agsd) - • R T In Paolld(T) - Ah,(T) ¦ T A«,(T) gives:
R In P,oltd
-------
• nchalpl.es of evaporation can be estimated by equation IV for the melting
points Expression XV, which is obtained from XIV, permits the calculation of
the entropy of evaporation at Tm In combination with "V and VII
Enthalpies and entropies of fusion can be predicted for Tl9 f r'dip
experimental enthalpies and entropies of sublimation (at Tm) by IV. IX, Ml,
and XV. A correlation of the evaluated enthalpies of fusion with the degr»-e I
chlorine substitution enable the estimation of enthalpies iirnl en rop ieb bv XI I!
of fusion for n-chlorInated dloxlns Equations IK and XI allow the conversinr.
of predicted enthalpies and entropies of evaporation to enthalpies and
entropies of sublimation Solid state vapor pressure curves can now I..-
predicted by equat.^n XVII which is obtained from XVI.
The differences of the heat capacities of the gas and the solid pb,ts»« •,
equal the rotational and the translat lor.a I heat capaclti . the g.is.-s minus
the lattIce-mode heat capacities:
Cp gas - so l id . 1R.C Uttic(
The lattice parts can be estimated by the rule of Dulortg and iv; i:
approaching 6R ac ambient temperatures for crystals composed of mo I «>cu I vi:>-
three nonzero moments of Inertia, Solid to gas heat capacity changes uS •!•• ••
J/fflol K were therefore assumed except for anthracene : • )l . ~ J/i«ol K'¦ ..n>!
J,6-DlCDF (*&5 J/mol K) , The latter values were determined t row tin* «,rroi ,•'.•••
be nt sublimation curves of the two compounds Equation XVI I .-.i c i>«
approximated by XVIII to calculate vapor pressures over a narrow temp.--r i:-u«
range
The boiling point (Tjj) appears In several express ions and was detenr,in«-d
in an Iterative procedure: enthalpies and entropies of sublimation w»-re-
determined by linear regressions of the experimental vapor pressure data over
the indicated temperature intervals (Table 49), Thev hold tor the ir.it!-
temperatures of the investigated temperature ranges defined bv
"""mid " 1 ^ma\ ' 'm In '{^max 1 '"min1
«2
-------
The temperature ranges were recalculated for the melting points bv mr.tns
of the bracket expressions of equation XVII. These equat iors were onrprnt
•ilong with 1, II, IV, V, VII, IX, XI, XIII, XV, and XVII into the spreadsheet
The bol1ing points were then set for correct predict ion of the vapor pressures
.it the melting points. given by:
^llquld^m^ " ^solid^m'
Rordorf's estimates on the heat capacity and the boiling point of
chlorinated dloxlns and furans and the extrapolated and estimated values lor
other thermal propert ies are shown In Table 49
Both the PCDFs and PCDDs exhibit the same ten 'encv of ucronglv decreasing
vapor pressure with increasing chlorine substitution Corresponding!v. the
boiling point temperature also increases with the number of chlorine atom!.
With the boiling point temperature for t r I-CDD and tri-CDF and hir.hii
chlorinated PCDDs and PCDFs ranging between 17 :°C and j i ."°C. a I
approx iraat ion of the thermodvr.anwc conditions would favor the forma' ion «•! tit-
and higher chlorinated PCDFs and I'CDDs in PCS trans former fire incident-. '!
est i ma t ion ol the thermodynamics of a PCB transformer fire tppears : •* •«
supported bv the analvsis of products generated at the incident site is v. I
in laboratory pvro ivs is and combustion studies.
PCOF AND PCDD LEVELS IN PCB TRANSFORMERS UNDER NORMAL USE CONDITIONS
An early concern was whether PCDFs and PCDDs are formed i is P."?.
transformers under normal use :ondlrIons. EPRl conducted i round-rohin :
with three participating laboratories uslt.g four sample*. ,»t » Ic s
taken from In-service transformer and capacitor units ,h Hi.* .ivt-iM,-, ..
congener class concentration found for PCDFs is slmvt, si« Tab !,
Furthermore, EPA obtained f'.uid samples row . transformers that were . '
fire incidents at Miami, FL, Binghamton, NY, and Chicago, IL Tot a I ! .F -
(Table 51) In the vised fluids is within the range of PCPF -a 1 ues prev-i ou- i v
reported for stock material ("table 52) It does not appear that, under n«.nn»l
use conditions, PCDFs are generated to any significant extern in : h«-
t r.ins forme r
H )
-------
T#bl« <•"» Vsftut Ki > >>JI > l ulf r J jl l Ul.v ot bi U-n.-<*li m nit
Dib«nXotur JIM., *1*1 AnllilMnws""'1
S
•a
••at. ranga
1dP(Tb)
K
•»(Tb)
a,(T.)
hv(Ta)
N.«T.>
Tb
(J/w>l)
(J/aol t)
(«*C)
(°C)
(Pa)
UJmol It)
(J/Ml I)
fkj/
lit
(kJ/BOl)
(°C)
ofca.
ob».
oba.
oka.
aatr.
aac.
¦ol)
•ol S)
pr ad
oka.
pre J.
cba.
(»•
*2250
285.2*
W>-*0
122.5*
4.22
111.3
1.24
185.9
228
48*
57
22.4
2J.2*
283.5
24*2
1NCM
*#550
2*1.8
JO-45
1051
1.97
-108.9
1.20
184*
217
74.2
54
21 ,4
21.2*
115.5
2*62
2-MCD0
*7140
292.1
12-75
89*
2.85
-108.9
1.20
184.*
2*1
78.1
51
18.5
21. 1*
314.0
2*8'
2,3-OCDD
1062X.
2*1.oa
11-101
141.5l
5.49
-104.5
1.15
186 9
210
77.8
61
24.7
158.0
J.7-DCDD
105510
279,05
*1-101
209.S1
7.1*
-10*.5
1.1*
187.2
22*
7*.*
55
24.8
17*.5
7 ,8-DCM
109010
291.87
12-90
151'
«.l*
-10* 5
l.t*
187.1
2?7
84.1
55
21.1
we1
182 5
1 ,3,7-TCBO
114200
304.77
17-100
US.5>
l.*7
-100.0
1.09
187.5
2 37
84.4
47
28.4
198.0
l»2,*-TCB0
11(7*0
122.14
J7-10I
128.51
1.15
-100.0
1.09
187.2
218
• 1.9
84
11.9
175.0
1,2,1,*-TC»0
118510
298.15
40-110
189*
*.94
-95.4
1.0*
187.7
211
85.fr
48
11.2
419.0
2,1,7 ,8-TCW)
124001*
287.4®
10- 71*
M5>
8 51
-95.8
1.0*
1M, 1
219
79 9
69
19 9
18.97'
444.5
OCBO
1511)0
114*8
1 * -201
JJ1>
>.87
-77.9
OS*
! «
215
•4 7
132
41.*
510.0
•m* tabic dlaplaya oknr««4 (ohm.), iittipatittl (a«tr.), aattaatad (cat.), and pi< tad valvaa (pr*d. o- ao Indication).
>*.C, PoMand and C.C. Tang, J. Agr. Pood Cha«. 20. 10*1-1099 (1*71).
Kordorf, ». Rtcklar, and A. Cnllro;, napebllaKad rraolta.
® ^CIC Handbook of Data oo Organic CeatpMnda, CSC Praaa. Sot a litm, PL (IM9).
*J. OmxI . >. Ilrckaayr, and D. laarl , Smtmm PatMt tffl. 4021/M, Car. Of fro. ( 1*7*).
*1. Coat ha and C.A. Hachtaalatar» Acta, Ckaa. Scand. )ii»I3 (1*71).
*P. Co*»raot, 1,1. CI rdiiar, and I.P. Wratro, Jr.. J, P*pa. Cbaa. 7*. DW-2MI (1*70).
'p.p. tear, t, P. via baosrdra, and V.V. ftocldar, J. Aaar. Clm. Soc. Hi 1004-1007 (1*71).
*J .*. Scbroy, P.B. ¦ I laaan. a«td S.C. Ckang, CfcaaoapHara 1^ iS73-874 (1*85).
*9,10-Dlhydroantbracana.
Kit tMtad froa ehlerlna corralat Ion an dlomla dat * ¦
'•kjit,) - 11.41 ~ 5.75 • X * 2 kJ/aol
llTk(f) - 283 ~ *2.0 • X - I ."71 * 1 * * ~ 15 (* • ettlortna ajl.it. #).
-------
Table 49, Vapor Fraaiure CorrelatIon* of D1bentodtoiIna,
D(b*n«ofurani t Janlhtnea , and Anlhracrnee < Coal I rated )
h,
(J/«ol)
oka.
•a
(J.'aol I)
oka.
ataa. range
<&C)
oka.
' ¦
< ®C)
oka.
1 nF( T—)
(Fa)
•mt r.
«/•->>
(J7*»l I)
vat .
K
H(Tb>
U/aol K)
•»
1)
efca.
yrad.
Tfe
(°C)
oka
OF*
•5*30
2 78 49
11-70
M.»®
4. *5
•Hi
0.97
185.7
224
65.4
54
19
6
27).0
28 7 ^
J.k-OCW
110870
»0 9
31*101
188.0*
7.03
-78.7
O.i*
lsk.9
276
70 7
75
)4
5
32,4?
357
375*
7 ,4 ,8-TCDF
112)00
299.31
45-ISO
154*
4)5
-74.)
0.S1
187 .4
22)
78.7
7k
)2
*
)92. 5
0C0F
1494)0
I2J.07
100-201
259*
5.0)
-52.1
0.5*
189.0
215
90.0
108
57
5
5)7 .C
liM htM
92450
294.79
12-M
100.55
4.4®
-121.9
1.J5
18k. 0
2)9
74.5
4*
17
1
293.0
311J
1,2,4,5,7,8-HC*
14 7000
1)1.32
•0-174
241.5®
6. 70
-95 4
1.0J
188.2
224
87.1
107
57
4
29.4*
4*).0
3401
Anthracene
91745
274.M
45-190
215.•*
• .50
-79.0
0.17
144.7
216
65.1
59
28
6
140 0
DM-Anthr.*
9,?2l
2(4.9k
31-9*
111®
5.07
-1)0.5
1.45
18*. 0
2)9
74.8
4k
17
k
293.0
3051
pred.
(•red.
c_lc.
pred.
*10
438. )»1
2,J,7,8-Tt0B
121170
284)79
)0- 70
8.71
-95.6
1.04
188 0
208
78.)
*7
18
)• .9'
•flit tab l» d 1 aplaya obarrved (oka.), •>trifel(l«4 (cat r ¦ ), eat latid (cat.), a.»d predlc c rd valnea ( pr#d, or no Indlcac imiI.
Fofcland and C.C. Tang, J. Agr . food Chta. 70,1093-1099 (1971).
^ ?8.F. ¦erdurf, ». Blckl er, and A. Geoffrey, • •pvkllaked raaulla.
v> 'clC Handkook of Data o® Or|*alc Cyptiunda, CtC Freaa, laci ftaton, fL (KM).
*J. Da* 11 , I. Urchmayr, and 9. Hear I , Svlaa Fa (vat Appl, 4021/7), Gar.Of fan . 2 i4l1,65N (1974).
^1. Core ha and C.A. Ifachtae I at er , Acta. Chm. Scand . IliDl) (1973).
*F. Ccw r»ol , * L. Cirdfcar. and I. f. Meet rua, Jr., J, Fkya . CttM. 7* ¦ 1 (1970).
1r.r. Boar, F P. leanertere. and M.tf. Moeldcr. J. Aaar. Chaa. Set. 94.1004-1007 (1972).
®J . H. Schroy, F.D. Hi Iraan, and 9 . C. Cheng, Oweoapfcere 14 i87)-876 ( INi).
*9,10-01 hydroanthracene.
(at taaced frcaa chlorine correlation on dloxln data ¦
!%< T„) « l),*l ~ 5.7$ * I ~ 2 kJ/aol
1 lTk
-------
Table 50, Average Total Congener Class Concentrations of PCDFs
in Four In-Service Dielectric Fluids^
IN-SERVICE SAMPLE (ppm)
CONGENER
CLASS
TCDFs
PeCDFs
HxCDFs
HpCDFs
OCDF
Total PCDFs
ISLAA*
ISL3Ab
0 1
2,6
7 6
16. 3
9«8
36 4
0.5
1.6
1.6
2 0
1.C
6.7
ISL2AC
N. D e
0.01
0.01
N.D.
N.D.
0 02
ISLIOc
N.D.
N.D.
N.D.
N.D.
N.D
N.D.
aISL4A - Askarel fluid with 70% Aroclor 1260/30% trlchlorobei z?nc
from transformer after 31 years of service,
kfSL3A - As above, except 20 vears of service; contained
0.5-2.0 racg/g PCDF.
CISL2A - Askarel containing Aroclor 1242 from capacitor that had
bulged but not ruptured.
dISLI0 - Mineral oil contaminated with 100 racg/g PCB from transformer
that had failed in service by arc ing.
.D - Not Detected.
86
-------
Table 51, Analysis of Fluids Invo lved in
Transformer Fire Incidents*46^
MIAMI. FLl
(May 29, 1984)
(*»cg/g)
BINCHAMTON. NY2
(February 5. 1981)
(«cg/«)
HICACO. IL
(September 28
(mcg/gi
PCDDs
N D a
N.D.®
N. J
PCOFs (Total)
6.0
16 2
3 0
Tetra-CDFs
Penca•CDFs
Hexa-COFs
Hepta-CDFs
Octa>CDF
0 31
0 78
2 2
1.3
1.4
0 48
11
11.0
3.2
0.41
<0.01
0.053
0,47
1.4
1 1
Ch lo robenzenes (Total)
430.000
350.000
160.000
Trlchlorobenzenes
Te c rachlorobenzenes
penta,'hlorobenzene
Hexachlorobenzene
260.000b
I40,000b
29,000b
900d
230.000b
110,000b
13.00nb
45d
ioo.ooov>
51,0n0b
5. i()0b
64b
PCBs (Total)
410,000
580.000
110.000
Monochlorobip.ienyls
Dichloroblphenyls
Trichlorobiphenyls
Tetrachloroblphenyls
Pentachloroblphenyls
Hexa^nlorobiphenyl»
Heptachlorobiphdnyls
OceachlorobIpheny1s
N'onachlok^b I phenyls
Decachlorob 1phenyl
320c
l,100c
2,600c
2,000c
31,000c
160.000c
170.000c
32.000r
14,000
<2,800c
230c
730c
3.500d
88.000
300.000
160,000d
24.000
Fs/PCDDs.
b Results of analysis at a dilution of 0.1 rag of oil/ml of hexane,
c Results of analysis at a dilution of 2,5 mg of oil/ml of hexane
d Results of analysis at a dilution of 0,25 mg of oil/nil of hexane
^ Askarel Type A: 60* PCBs with 60% chlorine (Aroclor 1260) with 40%
trlchlorobenzenes mixture.
^ Askarel Type D: 701 PCBs containing 541 chlorine (Aroclor 1254) and
30% trlchlorobenzenes mixture,
^ Mineral Oil with 25% PCBs (Aroclor 1260).
87
-------
ELECTRICAL ARC INC AND PCDD AND PCDF LEVELS
Electrical arcing has been Implicated as one of the causes ni l'< rf
transformer and capacitor fires, raising concerns as to whether arcing i.iums
the formation of PCDDs and PCDFs In the transformer fluid As di scussvc;
earlier. Rouse ec si., conducted several experiments Involving the arcing of
electrical energy through fluids containing various formulations of transformer
fluids.04) The results of this study were presented in Section 4, Tables ic
and 31. The data show no difference In the PCDF level before and .utir
electrical arcings. The authors suggest that the level of oxvgen is vcrv 1 nw
in the transformer (although It can occur In the cardboard components) m« thus
does not offer the environment for combust Ion and the formation of PCDFs ,twi
PCDDs.
De 11 Inge r ec ai. , has conducted studies to determine the effect of
temperature and oxygen concentration on the rate of format ion and destruction
of PCDFs as well as the thermal stability of the parent PCB. The thermal
degradation of a single PCB Isomer was conducted unde r four rc.u:ion
atmospheres at a constant gap phase residence time of 2.0 seconds. The isomer
selected for study was 2 , 4 , 5» 3 ' »4'-pentachl oroblpheny l (2,4,5.3' , 4' • PeCB) The
oxygen avaliability In the reaction atmosphere Is de sc rIbed using the
equivalence ratio, phi (0) Phi is defined as the ratio of fuel to oxvgen
actually In the reactor divided by the ratio of fuil to oxygen required * or
complete combustion. The values of phi used In this study were 3.0, 1.0. 0.2.
and 0.05, which range from oxygen-rtarved to very oxygen-rich conditions as the
values of phi become progressively smaller. A phi of 1.0 means there is
exactly enough oxygen available for complete combustion; thus, the percent
excess oxygen Is zero. Thermal degradation experiments were conducted a:
various temperatures ranging from 500-1000°C. The results are shown .«».
Table 53 and Flgu> . 2 and 3.
This study shows that the yields of PCDFs forced from the thcim.i I
degradation of 2 ,4 , 5 , 3' , 4' - PeCB are on the order of l to I percent: ..tul t lu-si-
yields Increase as the oxygen concentration increases (0- 3,0 to - 0
The predominant mechanism for their formation is HCl elimination.
88
-------
Table 52. Levels of PCDFs In Commercial PCBs(U)
PCDP LEVEL (pp«)
SAMPLE TRI- TETRA- PENTA- HEXA- HEPTA- TOTAL
Aroc lor 1254 0 10 0.25 0 70 0,81 N .n a 1.3
A oclor 1254
(lot KK602) --- 0.05 0.10 0.02 N D 0
Aroc lor 1260 0 .06 0 .30 1.00 1.10 1.35 3 8
Clorphen T64 0.10 0.30 1.73 2.45 0 82 5.4
Prodelec 3010 0 41 1.08 0.35 0.07 NO. 2.0
Mitsubishi (used) 2.13 4.00 3.30 0.53 N D. 10.0
aN D - Not Detected.
Table 53. Maximum Weight Percentage Yield oC PCDFs
as a Function of Reaction Atmosphere)
WEIGHT % YIELD
0 TEMPERATURE OF
MAXIMUM YIELD (°C) TRI-CDFa TETtLA-CDPa PENTA-CDF« TOTAL PCDFs
0.U5 750 0.60 4.3 2.0 h,"
0.2 800 0.34 1.7 0.56 2
1.0 850 0.13 1.3 0 25 I
3.0 900 0.068 0.71 0.21 0 '
89
-------
a ft m,i^ -- ^
£• wWpm nnvilMpt ¦ ••JQOf rvUri M *
yiffBll \#OflC«III*llWll
ToWPCOft
—•—
Pmrn
-A—
ygen Rich
2.7 3.0
uxygiYi u©f*cwu
$ (Ratio of Fuel to Oxygen in Reactof/Ratio of Fu*l
to Oxygen Required for Complete Combustion)
-------
Trl-COFt
750 800 850 900
Tamptratur* (°C)
Huid = 2, 4. 5, 3'. 4' - PwTtachlor obiph«nyi
-------
CORRELATION OF C0NBUSTION/PYR0LYS1S PRODUCTS GENERATED AND CONSTITUENTS OF
TRANSFORMER FLUIDS
Another issue of interest Is the correlation between the contaminants
found in the soot and the transformer fluid. The ability predict the tvpe
and the quantity of toxic contaminants that may form in a PCB transformer/
capacitor fire is of prime importance in the development of prevention and
control measures, Because of the scarcity and generally poor quality of data
obtained from PCB transform.r fire incidents, pyrolytic studies under
laboratory-controlled conditions have been employed,
EPA, through a contract W th Midwest Research Institute, Kansas City. HO.
has conducted a study to evaluate thermal degradation products using a benc.i
scale thermal destruction system.(^ Details of this study were presented In
Section 4 of this report. The results indicate that both temperature and
oxygen significantly affect PCDF yield. Statistical analysis showed a linear
relationship for PCDFs formed versus the concentration of PCB*. Figurt -
and 5 represent the relationship between the amount of PCDFs for.ned and th#
concentration of PCBs In the fluid.
in a study reported by Cervason. askarel fluid mixtures of 601 PCBs ,-nd
40% trlchlorobenzenes were combusted under varying flame temperatures ( ' The
results of this study were presented in Tables 29 and 41. Figures b and ?.
depicting these results graphically, indicate that the optimal "emperature for
the formation of PCDFs and PCDDs 1$ approximately 600°C This finding Is in
fair agreement with the work done at the Midwest Research Institute, where the
optimal temperature for PCDF formation from pyr&lvsls of PCBs was apprrxl.Mt#lv
675°C (Figure 8) . (7)
Und«r optimal conditions, PCDFs are formed from mineral oil or si 1 icon*
oil contaminated with PCBs at > 5 ppm. PCDFs were also formed from a rr i •
chlorobenzene dielectric ti d that contained no detectable PCBs Tho*.
results supported earlier laboratory work and analytical results of soot
material from transformer and capacitor fires, which determined that
chlorobenzenes are required for PCOD formation.
92
-------
5000
4000
3000
2000
1000
Total PCDFs
100
200 300
PCB Concentration
(ppm)
400
500
-------
BotowoH JOOd
Bld»H BKOH mU94 wmi
¦ « . . i ¦ mmm
mMm ZZ\
uj
-------
400
$00
Of *CDF» Formed m • FuncikKi
cui
600
700
Temperature CC)
mo
900
Tatr«-CDfs
#_
Pwrts^GDf*
Trt-COF#
1000
-------
96
2, 3, 7. 6-TCOO Concentration
(ppm)
o W 0> «B W
-------
CimmntnMm of PCDFs
mM TviiMritim"'
4000
3000
2000
1000
0
675
676
677
678
679
681
663
665
662
Temperature (°C)
-------
EPRI has also supported a major sCudy of the thermal conversion of various
transformer fluid formulae Ions Co PCDFs and PCDDs.(8,43) Fluids that have
beor. studied include mineral oil, tetrachloroethylene (TCE), and silicone oil.
all spiked with Aroclor 1254. 100 mc1 samples were either pyrolyzed (heated in
an oxygen•deflclent environment) or combusted (injected into a flame or heated
under conditions resulting in self - ignition). Pyrolyses were conduc ted usinp a
simple thermostatically controlled apparatus, capable of accommodating glass or
quartz tubes of diameters up to 6 cm within its 9-cm-long heated region. To
simulate more accurately certain catastrophic Inc i dents, pyrolvses we re
conducted at atmospheric pressure.
Tables 54 and 55, and Figures 9, 10, and 11 summarize the results of the
EPRI study. All data are consistent with the proposition that tetra- and
penta- CDF yields are roughly proportional to PCB concentrations in the starting
material. An Interesting feature of the mineral oll/Aroclor 1254 data is the
clear and reproducible differences between the patterns of tetra-CDFs and
penta-CDFs formed by pyrolysIs of neat Aroclor 1254 versus those formed bv
pyrolysls of the 5,000-ppm mixture. For example, 2,3,7,8-TCDF and co eluters.
comprise 16-211 of the tetra-CDF mixture formed from neat Aroclor, but :h<-v
comprise 4 5-551 of the mlxtuie from mineral oil/Aioclor Furthermore.
combus t ion of bipheny 1 In TCE produced decreas lng net dlbenzofuran as t he
residence time was varied from 18 seconds to 6 seconds and the wall temperature
maintained at 450°C. Ir. contrast, at 550°C, there Is a net increase in
dlbenzofuran yield as residence time decreases. This suggests that at the
higher temperature, significant dlbenzofuran destruction \s occurring. There
are also large effects on dlbenzofuran y'e' ' In the presence of different
solvents. Typically, yields obtained with >ml ustlon In TCE are much higher
than those obtained in silicone or mineral < . Some sharp differences on :ho
effects of particular variable parameters on dlbenzofuran yields have aiso ht-en
noted.
Bo th EPA's and EPRI's pyrolytic and combust Ion studies support : he
proposition that the amount and the specific PCDF Isomers found are related to
Che concentration of and the PCB homologs In the transformer fluid. Additional
supporting data lnclude EPA's data on the transformer oil and the generated
98
-------
Table 54. Aaounc of PCDFs Formed from
Pyrolysis of Aroclor 1254^
PCS CONTAMINANT
TETHA PENTA • V HEPTA
(ng PCDF/g Aroc' 1254)
100* 1254
5000 ppm PC8a in Mineral
Oil
500 ppm In Mineral
Oil
50 ppm In Mineral
Oil
5000 ppn in Silicone
5000 ppm in TCE*
6,900 13,000 6,600 60
7.800 33,600 36,700 3,800
6,200 35,400 34,900 4,400
65,000 70,000
21,400 14.000 1,380 30
580 5,600 2,500 180
lTCE - Tetrachloroethylene or perchloroethylene.
Table 55, Percentage of PCDFs Formed from
Combust '.on of Aroclor 1254^®)
PCI CONTAMINANT
TRI TETRA PENTA HEXA HEPTA
(% PCDF/g Aroclor 1254)
5000 ppn in Mineral
Oil
500 ppn In Mineral
Oil
50 ppn In Mineral
Oil
5000 ppn In TCE®
500 ppm In TCE
50 ppn In TCE
0 . 72
,40
30
.44
,24
,44
0.58
33
,30
1. 50
.70
1. 32
0.17
,084
,056
1.35
.661
1.00
0.33
.23
.38
u
'TCE - Tetrachloroethylene or perchloroethylene
99
-------
v^PW •* %0^KwB^HnKmWKtmmm Wl W W*J^P
OV MUVJU* 1294 ¦¦ ¦ rMHBHSfl OT
,(•»
HwCtfi
#
P«nU-COf«
T«tra-CDF»
+
H«f*a-CDFs
+
1000
2000
3000
PCS Concentration
(PP»n)
4000
5000
-------
(UMfCf)
UOIlBJJU®CHJOO 00,1
ooos ooo» oooe oooe 0001 0
*4QO-*mi
•JOW1
(•)
io • Mi Mil joaaaiv Ml
jp
-------
Figure 11
70.000-
80,000-
50,000-
40.000-
30,000
20.000-
10,000-
T«tra
50 ppm
500 ppm
5000 ppm
-------
soot in che Binghanton, NY fire incident. The data on the transformer oil and
soot from the BingharaCon fire art presented in Table 56.
Table 56, Correlation of Analytical Data on Transformer Oil
and the Generated Soot from the Blnghamton, NY Incident^*^
PCBj
PCDFs
ISOMER/TOTAL PCBs
ISOHER/TOTAL PCDFs
Tetrachloro-
Pentachloro-
Hexachloro-
Heptachloro-
Octachloro-
0,15
0,52
0.28
0,04
0.001
0.013
0.31
0.45
0,21
0.02
Penta-CDF concentration was approximately 7% of the total PCDFs In oil.
Penta CDF concentration increased to 31% in the soot after the fire This
finding corresponds to the 52% of pentachlorobiphenyls in the transformer oil.
Similar observations can be made for the hexa- and hepta-CDF concentrations in
soots relative to the concentrations of hexa- and hepta chlorinated biphenyls
The Binghamton, NY fire data also appear to indicate that the higher
chlorinated biphenyls are more likely to convert to chlorinated dlbenzofurans
and the lower chlorinated PCBs are more likely to decompose in a transformer
The finding of PCDDs in the Binghamton, NY fire incIdent raised concerns
that PCDDs may be generated in PCB transformers. Analyses of subsequent PCS
transformer fire Incidents, however, identifled only one other incident where
PCDDs have been identified. It appears that the presence of chlorobenzene is a
requirement for the format ion of PCDDs, and fires involving transformers
without chlorobenzenes do not generate PCDDs, The research of Erickson.
Cervason, and Buser appears to support this conclusion.(?.30,33.36) The i r
results are shown in Figures 12, 13, 14, and 15, Figure 12 indicates that the
PCDD concentration increases directly with the percentage err * ration of
trichlorobenzenes in fluid. Figure 13 indicates that only the l* robenzine
fluid (with a small amount of tetra isomer) produces significant amounts of
f ire.
103
-------
results are shorn in Figures 12, 13, 14, and 15. Figure 12 indicates Chat Che
PCDD concentration increases directly with the percentage concentration of
trlchlorobenzenes in fluid. Figure 13 Indicates that only the trichlorobenzene
fluid (with a snail amount of tetra-Isomer) produces significant amounts of
PCDD hoaologs -- the highest concentrations being the tri- and tetra-CDDs.
Thi* indicates a correlation between the lower chlorinated chlorobenzenes
Corning the lower chlorinated PCDDs, Figure 14 indicates that
trlchlorobenzenes produce mostly tetra- and penta-CDDs. This again supports
the evidence that lower chlorinated chlorobenzenes form lower chlorinated
PCDDs. Figure 15 indicates that trlchlorobenzenes form significant amounts of
tetra- and penta-CDDs, while tetrachlorobenzenas form hepta- and hexa-CDDs.
The tetrachlorobenzene forms the highest concentration of hepta-CDD. and the
trichlorobenzene forms the highest concentration of tetra-CDD (trl-CDD was not
measured),
104
-------
It.
1500
1200
900
o
000
300
0
Run 2
50
100
of Trlchk>fob«n?ene«
-------
Trt Tvtra Pwita Hexa H«pta Octa
PCDO Homotofl
-------
c*
£01
PCDO Concentration
(ng/ml fluid)
8 S
I
-------
Flour* 15.
wOClCVlVUfllKIfV Uf rVlwl FfWWIVQ Wl _ ^
wMOfOOcnienQ iipiKia rvuM oy 9f09H0§99
PCDD Homotog
~
Trt 4 T«tr« ~ Psnta ~ TCB
Tatrft-TCB
Trt-TCB
PwMa-TCB
-------
REFERENCES
1U S. Federal Register, 1982, 47(165);17342 -3715». August 2S.
JU.S. Federal Register 1984 49{ 198) : 39966 • 3^9 , October II,
^United States of America 1985 Code of Federal Regulations, 40 ¦*»?!,
July 9.
4U,S, Federal Register, 1985. 50(9); 1978-2006 January 14,
\'.S, Federal Register 1986 51(9) 1602•1766. January 14.
''United Stat s Environmental Protection Agencv, 1983. Diosin Str.i;.-
November 3
Erickson, M.D.. Cole, C.J.. flora, J. D. , Corman, P.C. . Hiile, C L . Hinvh.»v.
C.D., Hopkins, F.C.. and Swartion. S E. 1984, Thermal «egr«»d.>: i «»t.
products from dieleetrtcal fluids, EPA-560 3 - 84 - 009 . Noverebo v 1 9
¦a
"Addis, G. 1985, Pyrolysls and combust Ion of PCB as a contaminant in flfctrU'
insulating fluids. Pi e sented at the 3th International Symposium *!"
Chlorinated Dioxins and R»1 ated Compounds . Bayreuch. FRC , September I>
^Rappe. C , Harklund. S , Kjeller.L -0 , Rergqvist, P -A and Hansson. M
1985, Strategies and techniques for sample collection it, I
analysis: experience from the Swedish PCB accidents Envi ron. He,, I *'»
60:279-292,
^nElo, 0., Vuojolahti, P., Janhunen. H. . and Rantanen, J, 1985 Recent P« B
accidents in Finland. En"; ¦ on Health Perspecc 60:31b-SI4'
^Schecter, A. 1982. Contamination of an office bulK'ing in Blnghamton. 7 .
bv PCBs , dioxins, furans. and h phenyl ones .1 f : e 1 .in e lt*c r r ica 1 p.un-! ,r,a
electrical transformer incident, Chcmosohe re 12t,4/5 ) ; - 08O .
*
l-N'ew York State Department of Health l°82 Health questions and .m* • •
related to chemical contamination of the Bingharaton State Office BuiIdn,
^Rappe , C . Marklund, S , Bergqvist, P - A . and Hansson, M 1983 S4.»i v.
chlorinated dioxins. dlbenzofurans. and other polvchlorinated poivrat» i. u
aronacics formed during incineration and PCB fires, Department of w
Chemistry, University of I'mea . S - 901 8? Umea. Sweden
' 'Buser. H.R 1985 Formation, occurrence, an- analysis of polvchior in,
dibenzofurans, dioxins, and related c«rapounds Environ. Health Pct^pn-r
60:259-267
^Lee. A. and Campbell, B, 1984 PCDDs and PCDFs from PCBs r:ransforme 1 .
-------
REFERENCES (Continued)
* 7Hryhorczuk, D.. Orris, P., Burton, U,» Kominsky, JR., and Melius, J 1983
Exposure co PCDFs from a PCB-contain fire: Chicago Presented at the
5th International Symposium on Chlorinated DioxIns and Related Compounds.
Bayreuth, FRG, September 16*19.
'^Kominsky, JR. and Flesch, J.P 1981. Health Hazard Evaluation Report
Our Lady of Vis i tat ion Elementary School. Cincinnati, OH, NIOSH Report
No. HETA 81-237, July.
^Memorandum from G. Choudhary and J C. Posner to K. McManus (Regional Progr.nr
Consultant) 1982 Results of the analysis of chlorinated dioxins 1
''Personal communlcation from J R. Konlnsky (Supervisor Industrial Hygienist,
NIOSH) to Paul E. desRosiers (EPA). September 25, 1984
-^Memorandum from S. Svanson and H. Erickson (Midwest Research Inst itute) co
Dan Heggem (EPA) 1985. Results of analysis of oil and soot samples from
Paul E desRosiers, October 24.
-^Personal communication from JR. Kominsky (Supervisor Industrial Hyglerist.
NIOSH) to D.E. Anderson (Industrial Hygienist Engineer. Minnesota
Department of Health) HETA 82-310, July 26, 1982.
'"Jan.eson, B. and Sundstrom, G. 1982, Formation of polychlorinated dibenzo-
furans (PCDFs) during a fire accident in capacitors containing polychlor•
inated biphenyl (PCB) in: Chlorinated Plains and Rf Utftd Compounds - -
Immt on the Environment (Hutzinger, 0., et al.) , pp. 201 - 207
'^Rappe, C , Marklund, S. . Bergqvist, P.-A,, and Hansson, M. 1982. Polv-
chlorInated dioxins (PCDDs), dibenzofurans (PCDFs), and other polvnucle.ir
aromatics (PCPNAs) formed during PCB fires. Chemlca Scrlota 20 56-61
'"'Rappe, C , Kjeller, L.*0., Marklund. S . Nygren, M. . and Fournie, R 1 ''S j
PCB Accident in France. Proceedings: 1985 EPRI PCB Seminar, Seattle "*A
EPRI CS/EA/EL-4480
- Rappe, C. 19CS European PCB accidents update. Presented «it chi jrh
Internationa Symposium of Chlorinated Dioxir.s and Related Compoui.«N.
Bayre uth, FRG, September 16 19
'^Morita, M . Nakagawa, J., and Rappe, C. 1978 Polychlorinated dibensofm'.m
(PCDF) formation from PCB mixture by heat and oxvgon
Buii £nvtl9,n- Cgntfltn- Toxicol 19:665-670
110
-------
REFERENCES (Continued)
¦'l in r . H R , Bosshardt. HP. and Rappe. C 1978, Formation of poly
chlorinated dlbenzofurans (PCDFs) from the pyrolysls of PCBs. Chemosphere
7(1) : 109-119.
'%appe, C. and Buser, H R. 1979. Formation and degradation of poiychloti-
nated dlbenzo-p-dloxlns (PCDDs) and dlbenzofurans (PCDFs) by thermal
processes. Presented at the 178th National American Chemical Society
Meeting, Washington, DC, September 9-14,
^Buser, H R . and Rappe, C. 19^9. Formation of polychlorinated dibwnzo-
furans (PCDFs) from the pyrolysls of individual PCB isomers Chemosphere
8(3): 157-174.
^Paaslvlrta, J,, Herzschuh. R . Humppl, T . Kantolaht i . E . Knuutlnen, J .
Lahtlpera, H , Laltlnen, R . Solovaara, J . Tarhanen, J . and Virkki. u.
1985. Pyrolys' s products of PCBs. Environ. Health Perspec t
^Personal communication from P. Cervason (ATOCHEM) to Paul E. desRosiers (EPA.
DDAG), March 14, 1984.
"•Route, T , Koch, R E , and CsRoos, F.L. 1985. Formation of PCDFs and PCDDs
from PCBs In electrical discharges. Proceedings: 1985 EPRl PCB Seminar.
Seattle, UA. EPRI CS/EA/EL-4480.
^C'ooke, M , DeRoos. F. L. . and Rls ing, B. 1985. Safety of non-PCB reclas-
sified transformers in fire Incidents, a building scale engineering
study. Proceedings: 1985 EPRI PCB S.ninar, Seattle, WA E PR I
CS/EA/EL-4480
^Iwse r , H . R and Rappe, C. 1979. Formation of polvchlorinated di henno•
furans (PCDFs) and dlbenzo-p-dloxlns (PCDDs* from the pyrolysls ot
chlorobenzenes Chemosohere 6 415-424
"Rappe, C , Marklund, S. . Bergqvlst, P.-A., and Hansson, M. 1983 Po 1 v •
chlorinated dlox ins, dlbenzofurans and other polychlor inated polvnuc le.ir
aromat les forced during incineration and PCBs fires Department of
Organic Chemistry, University of Umea. Sweden.
^Rappe , C. , Marklund, S , Buser, H.R., and Bosshardt. H.P. 19 7 fl Form.it ion
of polychlorinated dlbenzo-p-dloxlns (PCDDs) and dlbenzofurans kPCDFs * hv
burn or heating chlorophenates. Chemosohere 3:269-281,
^Personal comnunication from H Jurgens (DEKONTA Mainz) to Paul E desRosi.is
(EPA, DDaC), September 23. 1985
^personal communication from U. Lee (EPA, Region III) to Paul E. desRos < > is
(EPA, DDAC). March 1985
^PCB/Dioxln Sampling Program for a Preliminary Site Assessment at St rami I ex-
Scrap Metal/Manning Property, Purdy, WA. March 1985
III
-------
REFERENCES (Continued)
**^Rordor£, BP 1985. Thermal properties of dioxlns, furans, and related
compounds. Presented at the 5th International Symposium on Chlorinated
Dioxlns and Related Compounds, Bayreuth, FRC, September 16-19,
•^Gordon, S.H., Miller, M., DeRoos, F. L. , Cooke, M., Cuercln, J., and Addis, C
1985. Measurement of PCDFs/PCDDi in utility equipment. Proceedings
1985 EPRI PCB Seminar, Seattle, UA, EPRI CS/EA/EL-4480.
"^Memorandum from S, Swanson and M. Erlckson (Midwest Research Institute) to
Dan Heggem (EPA), 1985, Results of analysis of oil and soot samples front
Paul E. desRoslers (EPA, DDAG), EPA Contract No, 68-02•3938. June 17
" ^dt-sRoslers . P. E. , Uestfall. B. , Campbell , B . and Lee, A, 1985 PCS Fires:
preliminary correlation of chlorobenr«ne and PCB contents of the fluid
with PCDF and PCDD contents of soot. Proceedings: 1985 EPRI PCB Seminar,
Seattle, UA EPRI CS/EA/EL-4480.
^Delllnger. B,, Rubey, U.. Hall, D., and Mater, S. 1984. Laboratory Investi-
gation of th® high temperature formation and destruction of PCDFs. EPRI
PCB Seminar, Palo Alto, CA.
112
-------
TICMNICAl MK>M DATA
rr»4 /aurrvr rwvu o» Mr T'Ytv »r fan eomf
>UtoAMI AND tOO"IU
Technical Resources, Inc.
3JOJ Monroe street, Suite JOG
Roc kvi11e, MD 208S2
C
10 'MOG**M UlulsV NO
C6wT*Al!T/6rAST,i5'5'
68-03-3269
W AGlNCv N4M1 AND AOOH|||
Hazardous Waste Enqineermq Research Laboratory
USLPA, ORD
26 W. St. CI Air Street
f i rir inn,if i f>H
14 t'ONOMiNG All 1 NC> COOl
EPA/600- •:
»» hO'lt
'l* AtffcAC*
Dielectric fluids in transformers and capacitors often contain polychlcrinated
bi phenyl s (PCBs) or chlorobenzenes . Ihese substances may generate po". vchlen nates
dibenjofurans 1PCDFs ) and po1yc h1 or mated dibenzo-p-dioxins (PCDDs) under certain
conditions of combustion/pvrolysis. When electrical equipment contai'inq tnese
fluids is involved in an accidental fire, the resulting smoke, soot, end residues
may be contaminated with PCODs, PCDFs, and other toxic chlorinated hv rocarbons.
This report contains a review of .everal laboratory studies inve- 11 qati nc; the
sources of PCDDs and PCDFs as well as the conditions under which thev are ferried.
In addition, some data from sites of actual fire incidents are available and are
discussed. Ch]orobenzenes and PCBs do not form PCDDs and PCDFs when heated in the
absence of Oxygen. During fires the dielectric fluid of transformer-, or capacitors
may be leaded or vented from ruptured casings. With exposure to oxygen, PCBs car-
produce PCDFs and chlorobenzenes can produce PCDDs. The particular isomers of
PCDDs and PCDFs formed are related to the number of chlorine substi tuer.ts in the
react inq material.
II" MO'M AMD DCCk/MIXl
» DI»C«ifTP»!
fc iOI»'i»i|»l'OM* INDID W«MI
c cotAti i »id Cio-r
U D'»T»i«wlTlO». r«T|MIX1
Re 1 ease Unlimited
IB HCw«f ClAH (Thaktfm
Unclassified
>1 MO C> **6ii
M tlCwMitv Clam tTkupfri
Uncla^ifipfi
it rmci
|p* im.i (¦.. 4.T7»
------- |