file3.pdf

Chemosphere51 (2003)109–115

www.elsevier.com/locate/chemosphere

Inﬂuence of heavy metals on the formation

and the distribution behavior of PAH

and PCDD/F during simulated ﬁres

M. Wobst, H. Wichmann, M. Bahadir *

Institute of Ecological Chemistry and Waste Analysis, Technical University of Braunschweig, Hagenring 30,

D-38106 Braunschweig, Germany

Received12 April2002;receivedin revisedform 22 October2002;accepted7 November2002

Abstract

Combustionexperimentswereperformedwithanartiﬁcialﬁreload(polystyreneandquartzpowder)inalaboratory

scaleincineratorinthepresenceofgaseousHCltosimulateaccidentalﬁreconditions.Theaimofthisinvestigationwas

totracebackthealterationsoftheformationandthedistributionbehaviorofPAHandPCDD/PCDFtothepresenceof

CuO or a mixture of metal oxides (CdO, CuO, Fe 2 O 3 , PbO, MoO 3 , ZnO). The total amount of the 16 PAH target

compoundswasreducedbythefactorof5–9whenthemixtureofmetaloxideswaspresentratherthanmerelyCuO.PAH

patternsaswellastheir distribution behaviorwere signiﬁcantlyinﬂuencedbytheseoxides.Ingeneral,transportation

insidetheinstallationwasenhancedformostofthe16analyzedPAH.Onlyﬂuoreneanddibenzo[a,h]anthracenewere

transportedtoasmallerextent.IncontrasttoPAH,totalconcentrationsofPCDDwereincreasedbyfactor9andof

PCDFbyfactor10,respectively,whenCuOwaspresent.Addingthemixtureofmetaloxidesresultedinanincreaseof

PCDD by factor 14 and of PCDF by factor 7. CuO and the mixture of metal oxides had a diﬀerent inﬂuence on the

PCDD/Fhomologuepatterns.Forinstance,theHxCDFtoOCDFratioafterincinerationwithoutanymetaloxidewas

1to6,whereasadditionofCuOorthemixtureofthemetaloxidesshiftedtheHxCDFtoOCDFratiostowards1to40or

1 to 17, respectively. Combustion along with CuO increased transportation of higher chlorinated PCDF congeners,

whereasthemixtureofthemetaloxidescausedastrongdecreaseofPCDFdistributionthroughoutthesystem.

2003ElsevierScienceLtd.Allrightsreserved.

Keywords: Heavymetal; PAH;PCDD/PCDF; Combustion;Fireaccident; Laboratory scale incinerator

1. Introduction

mental designs ranged between investigations of the

details applying laboratory scale incinerators and sam-

pling at waste incineration plants, as also frequently

practiced.ButconcerningtheemissionofPAH,PCDD/

PCDF and HM during ﬁre accidents or full scale sim-

ulations under ‘‘uncontrolled combustion conditions’’,

just a few articles can be found citing analytical results

(Meharg and French, 1995; Wichmann et al., 1995;

Ruokoj € aarvi et al., 1999; Wobst et al., 1999; Wichmann

et al., 1999). However, low-volatile organic pollutants

and HM were always separately determined.

The aim of the present investigation was, to con-

ductcombustionexperimentsinsuch awaythat allows

to trace back alterations of the formation and the

Various investigations are published, describing the

releaseofpolycyclicaromatichydrocarbons(PAH)and

polychlorinateddibenzo-p-dioxins(PCDD)anddibenzo-

furans(PCDF)duringcombustionprocesses(Lohmann

and Jones, 1998). In a similar way, the combustion in-

ducedemissionofheavymetals(HM)wasinvestigated,

as well (Wichmann et al., 2000). In both cases, experi-

* Corresponding author. Tel.: +49-531-391-5960; fax: +49-

531-391-5799.

E-mail address: m.bahadir@tu-bs.de (M. Bahadir).

0045-6535/03/$- see frontmatter 2003ElsevierScienceLtd. Allrights reserved.

PII: S0045-6535(02)00806-8

110

M. Wobst et al. / Chemosphere 51 (2003) 109–115

distributionbehaviorofPAH,PCDDandPCDFtothe

parameter presence of HM in the ﬁre load. For this

purpose, combustion experiments were carried out in a

laboratory scale incinerator. The distribution behavior

oftheheavymetalsCd, Pb,Fe,Cu,Mo,andZninthe

same incinerator was already published (Wobst et al.,

2001). The inﬂuence of several HM on the formation

and the distribution behavior of the organic pollutants

is now presented for the ﬁrst time. It was intended to

simulate accidental ﬁre conditions during fully devel-

opedcombustion.Thiswasachievedbythechoiceofthe

combustion temperature, the compositions of fed gases

andﬁreloads,theheavymetalsknowntobefrequently

analyzed after building or vehicle ﬁres in relevant

amounts,andtheconstructionoftheincinerationdevice.

PAH,16targetcompoundswereanalyzedtogetherwith

benzo[a]pyrene-d 12 as internal standard following US

EPA method 610 (1982). Concerning PCDD/F, 13 C-

labelled standards were used for applying the isotope

dilution method (US EPA method 1613, 1990). Gas

chromatographic separation of the PAH compounds

and PCDD/F homologue groups was conducted using

DB-5capillarycolumn(30m,0.25mmi.d.,0.25 lmﬁlm

thickness, 95%dimethyl-, 5% phenylpolysiloxane). Sep-

arationofthe2,3,7,8-PCDD/Fcongenerswasperformed

using CP-SIL 88 capillary column (50 m, 0.25 lm i.d.,

0.2 lm ﬁlm thickness, cyanopropylpolysiloxane). Re-

ported concentrations of PAH and PCDD/F were ref-

erenced to the amount of polystyrene utilized in each

experiment. To assure the analytical results, the com-

bustion experiments were performed twice.

The following abbreviations are used for the PAH

target compounds:

NAP: naphthalene, ACY: acenaphthylene, ACE:

acenaphthene,FLE:ﬂuorene,PHE:phenanthrene,ANT:

anthracene, FLA: ﬂuoranthene, PYR: pyrene, BaA:

benz[a]anthracene, CHR: chrysene, BbF: benzo[b]ﬂuo-

ranthene, BkF: benzo[k]ﬂuoranthene, BaP: benzo[a]py-

rene, IcdP: indeno[1,2,3,c,d]pyrene, DahA: dibenzo[a,

h]anthracene, BghiP: benzo[g,h,i]perylene.

2. Materials and methods

The combustion experiments were carried out at an

oven temperature of 600 C in synthetic air mixed with

10% HCl gas. The choice of these parameters led to a

combustion atmosphere that can often be met during

real ﬁres. The gas ﬂow was adjusted to 1 l/min and the

thermal treatment was performed for 30 min, respec-

tively. The materials to be thermally stressed, consisted

ofamixtureof50%quartzpowderand50%polystyrene

(PS) powder 500 mg each. The HM added consisted of

CuOorofamixtureofmetaloxides(CdO,CuO,Fe 2 O 3 ,

PbO, MoO 3 , ZnO) 50 mg each calculated as metal.

The incinerator, described in detail in Wobst et al.

(2001), consisted of a furnace reactor with an inserted

quartz tube, a gas supply and a sampling device. After

leaving the quartz tube, the combustion gases passed a

glasswoolplug(GW),aLiebigcooler(CO),aglassﬁber

ﬁlter (GF), and ﬁnally an absorption bottle (AB) ﬁlled

with toluene. Together with the quartz boat (QB), con-

taining the materials to be treated, these were the ﬁve

samplingspotsafterﬁnishingacombustionexperiment.

The residues in the quartz boat and in the glass wool

plug as well as the deposits on the inner surface of the

cooler and on the glass ﬁber ﬁlter were extracted in a

Soxhlet apparatus for 16 h with toluene.

TheanalytesPAHandPCDD/Fwere isolatedusing

appropriatecolumnchromatographicclean-upmethods

as described earlier in Wobst et al. (1999) and Dettmer

et al. (1998). PAH clean-up was done with one column

containing 20 g neutral aluminum oxide deactivated

with 2% water and elution with n-heptane/ethylacetate

(95:5).Aclean-upinthreestepswasappliedforPCDD/

F,usingtwoalkalinealuminiumoxidecolumnsandone

silica gel column with acid and alkaline impregnated

zones and elution with n-heptane.

Determination of the PAH and PCDD/F was con-

ducted by means of GC/MS (GC-17A, QP 5050A, Shi-

madzu) in selected ion monitoring mode. In case of

3. Results and discussion

3.1. Polycyclic aromatic hydrocarbon patterns

Total PAH concentrations found in the incinerator

after diﬀerent combustion experiments are given in

Table 1. These are mean values of two similar experi-

Table 1

Total PAH concentrations after combustion experiments (mg/

kg PS)

Without

With

CuO

WithHM

mixture

Naphthalene

5560

1820

803

Acenaphthylene

265

43.5

15.6

Acenaphthene

11.2

< 0.01 < 0.01

Fluorene

580

78.1

499

Phenanthrene

6230

766

298

Anthracene

927

208

90.0

Fluoranthene

686

83.8

45.6

Pyrene

186

12.7

4.78

Benz[a]anthracene

174

9.15

5.49

Chrysene

346

55.7

22.2

Benzo[b þ k]ﬂuoranthene 251

25.1

23.9

Benzo[a]pyrene

30.3

3.68

5.45

Indeno[1,2,3,c,d]pyrene 35.3

3.56

4.95

Dibenzo[a,h]anthracene 117

10.4

7.16

Benzo[g,h,i]perylene

15.4

3.35

4.09

P PAH

15400 3120

1830

M. Wobst et al. / Chemosphere 51 (2003) 109–115

111

ments,respectively,whichisalsotrueofTables3,4,and

6. To demonstrate repeatability, detailed results of two

similar combustion experiments with CuO are exempl-

arily introduced in Table 2.

PAH were formed during all combustion experi-

ments. A strong reduction of the total amount of PAH

couldbeobservedwhenCuOorHMmixturehadbeen

added. Concentrations were reduced by the factor of

5–9,comparedtothecombustionintheabsenceofany

metal oxide. This decrease was stronger in case of the

HMmixturethanincaseofCuOalone,ascanbeseenin

Table 1. The highest concentrations of NAP and PHE

weredetectedintheabsenceofanymetaloxideorwhen

just CuO was applied. The substance patterns were

similartothoseidentiﬁedinsamplestakenfromprivate

residencesafterﬁreaccidents(Wobstetal.,1999).Inboth

cases, comparably high concentrations of FLE, PHE,

ANT, and FLA were detected. In case of the HM mix-

ture applied, the contents of NAP and FLE were the

highest, although being much less than in the former

cases.

TheresultsofthisinvestigationportendthattheHM

oxides actively inﬂuence the combustion process and

complicatethe formationofthe PAH undertheexperi-

mental conditions described. The formation of other

combustionproductsislikelyfavored.GC/MSscreening

analyses gave hint on an increased formation of chlori-

nated compounds.

Table 2

Comparison of PAHamountsresultingfromtwo similarcombustion experimentswith CuOinvolved(mg/kgPS)

Quartz boat

Glasswool plug Liebigcooler

Glassﬁberﬁlter Absorptionbottle

Naphthalene

< 0.01

59.3

298

529

1060

0.101

26.0

248

310

1110

Acenaphthylene

< 0.01

0.612

20.0

11.6

< 0.01

0.697

12.1

12.0

18.4

Acenaphthene

< 0.01

Fluorene

< 0.01

9.20

33.6

41.6

2.29

< 0.01

9.34

29.0

29.5

1.53

Phenanthrene

0.187

232

233

379

14.4

0.290

264

160

237

13.4

Anthracene

< 0.01

5.37

26.8

20.3

151

< 0.01

7.19

26.3

17.3

162

Fluoranthene

0.125

31.8

24.8

30.3

< 0.01

0.143

33.0

21.8

25.6

< 0.01

Pyrene

< 0.01

1.22

9.01

3.88

< 0.01

1.49

7.58

2.26

< 0.01

Benz[a]anthracene

< 0.01

1.57

4.95

3.55

< 0.01

1.63

4.47

2.10

< 0.01

Chrysene

< 0.01

24.6

8.69

20.1

< 0.01

27.6

11.7

18.6

< 0.01

Benzo[b þ k]ﬂuoranthene < 0.01

5.50

9.14

11.8

< 0.01

7.47

7.49

8.82

< 0.01

Benzo[a]pyrene

< 0.01

1.04

0.682

2.02

< 0.01

1.28

0.760

1.58

< 0.01

Indeno[1,2,3,c,d]pyrene

< 0.01

0.774

1.24

1.52

< 0.01

0.934

1.44

1.20

< 0.01

Dibenzo[a,h]anthracene

< 0.01

4.60

4.10

2.73

< 0.01

4.13

3.53

1.80

< 0.01

Benzo[g,h,i]perylene

< 0.01

2.19

0.444

0.681

< 0.01

2.24

0.600

0.538

< 0.01

112

M. Wobst et al. / Chemosphere 51 (2003) 109–115

3.2. Polycyclic aromatic hydrocarbon distribution inside

the incineration device

CuO or HM mixture. In general, the addition of HM

oxidesﬂattenedthedistributioncurveofNAP.Formost

oftheotherPAH,theadditionofHMoxidescausedan

enhancement of volatility. Especially a strong volatil-

izationrespectivetransportationofANTandACYwas

obvious. Seventy ﬁve percent of ANT and signiﬁcant

shares of ACY (27% or 20%) were found in the ab-

sorption solution when CuO or HM mixture were ap-

plied.IncaseofPAHwiththreetosixcondensedrings,

ashiftoftheconcentrationsfromtheglasswooltowards

the cooler was observed. CuO had a stronger inﬂuence

on thevolatilization ofthe3–6-ringPAHthan theHM

mixture.ForBghiPthesmallestandforPYRthebiggest

change of distribution inside the installation was ascer-

tained. In contrast to the other PAH, FLE and DahA

were transported to a smaller extent in the presence of

HM oxides. The concentrations of FLE and DahA on

theglassﬁberﬁlterdeclinedfrom65%and26%without

HM involved, to 48% and 24% with CuO, and to 1%

and 7% after adding the HM mixture.

The addition of CuO or HM mixture did not only

inﬂuencetheamountofPAHformedandthesubstance

pattern but also the distribution behavior inside the in-

cinerator.ThisisexemplarilyshowninTable3forNAP,

PHE, CHR, BaP, and BghiP, representing PAH with

two to six condensed rings.

Combustion without HM resulted in a strong in-

crease of the NAP concentration from quartz boat to

absorptionsolution.Thereasonforthisistherelatively

high vapor pressure of NAP among the investigated

PAH, hence it was transported best through the instal-

lation. The concentrations of ACE also increased from

the quartz boat to the glass ﬁber ﬁlter. Contrary to

NAP,onlyasmallconcentrationofthiscompoundwas

foundintheabsorptionsolution.Theconcentrationsof

the other PAH under investigation were also very low

in the absorption solution. Generally, the distribution

patternofthe3–6-ringPAHwerequitesimilar.Highest

concentrations could be found in the glass wool plug

andontheglassﬁberﬁlter,whereastheywerelowinside

thecooler.Asexpected,increasingconcentrationsinthe

glasswoolplugandreducedconcentrationsontheglass

ﬁber ﬁlter could be observed with rising condensation

degree of the PAH. The percent contents in the glass

woolrosefrom45%for PHEto68%forBghiP. Atthe

sametimetheconcentrationsontheglassﬁberﬁlterwas

reduced from 46% to27%.

There was a varying impact on the distribution be-

havior of the PAH formed caused by the addition of

3.3.Polychlorinateddibenzo-p-dioxin/dibenzofuranhomo-

logue patterns

AmountsofPCDD/Fhomologuesresultingfromthe

combustion experiments are presented in Table 4. Ad-

ditionofHMoxidesobviouslyinﬂuencedtheformation

ofPCDFandPCDD,aswell.AsincaseofPAH(Table

2), detailed results of two similar combustion experi-

ments with CuO are exemplarily introduced in Table 5

in order to demonstrate the repeatability.

Table 3

Distribution of selectedPAHinsidetheincinerationdevice(%)

Quartzboat

Glasswoolplug Liebigcooler

Glass ﬁberﬁlter Absorptionbottle

Naphthalene

0.01

0.07

6.7

23.7

69.5

–

3.05

15.3

27.1

54.5

–

0.53

20.7

9.64

69.1

Phenanthrene

–

45.2

8.93

45.9

0.01

0.02

27.0

27.2

44.1

1.68

–

25.5

47.4

22.8

4.23

Chrysene

–

72.8

4.85

22.3

–

46.0

16.3

37.7

–

70.8

16.8

12.4

–

Benzo[a]pyrene

–

54.0

8.28

37.7

–

27.7

18.2

54.1

–

56.9

9.82

33.3

–

Benzo[g,h,i]perylene – 67.6 5.95 26.5 –

– 66.1 13.4 20.5 –

– 71.7 3.96 24.4 –

Firstline:combustionexperimentswithoutHM;secondline:combustionexperimentswithCuO;thirdline:combustionexperiments

withHM mixture.

M. Wobst et al. / Chemosphere 51 (2003) 109–115

113

Table 4

Amounts of PCDD/F homologue groups after combustion

experiments(lg/kgPS)

WithoutHM With CuO WithHM

mixture

PCDD/Fconcentrationssigniﬁcantlyincreasedwiththe

chlorinationdegree.IncontrasttoPAH,theadditionof

HM oxides led to an increased formation of PCDD/F.

In case of CuO, Hx-OCDD/F concentrations were en-

hanced. This was especially evident for OCDD (from

94.7to1440 lg/kgPS)andOCDF(from2250to33000

lg/kg PS). Compared with PCDD/F amounts resulting

from incineration without any HM the concentrations

were enhanced by factor 9 for PCDD and by factor 10

forPCDFwhenCuOwaspresent.AddingHMmixture

resultedinanconcentrationincreaseofPCDDbyfactor

14and ofPCDF by factor 7.

AcomparisonoftheresultswithCuOandwithHM

mixture indicated the following diﬀerences: Strongest

diﬀerence was identiﬁed for TeCDD. In this case the

concentrationwaselevatedfrom12.2 lg/kgPS(CuO)to

134 lg/kg PS (HM mixture). Regarding the analytical

resultsforPCDFitwasstrikingthatTeCDFformation

was elevenfold higher in the presence of HM mixture

thanwithoutHMoxides.Incontrasttothat,noTeCDF

enhancement was caused by CuO. Nevertheless, con-

cerningOCDF,theadditionofCuOresultedinahigher

concentration (33000 lg/kg PS) than the addition of

HM mixture (18900 lg/kg PS). The concentrations of

HxCDF and HpCDF were quite similar after both ex-

periments.TheHxCDFtoOCDFratioaftercombustion

P TeCDD 14.6

12.2

134

P PeCDD 15.6

12.7

16.3

P HxCDD 13.5

23.3

96.3

P HpCDD 43.6

184

310

OCDD

94.7

1440

1990

P PCDD

182

1670

2550

P TeCDF 189

192

2130

P PeCDF 170

280

741

P HxCDF 356

832

1120

P HpCDF 864

3530

3370

OCDF

2250

33000

18900

P PCDF

3830

37800

26300

As expected, PCDF were predominantly formed,

when polystyrene was combusted in an atmosphere of

synthetic air and HCl. Concentrations of single PCDF

homologuegroupswerehigherbyafactorof10–46than

those of the PCDD. Total concentrations of 182 lg

PCDD/kg PS and 3830 lg PCDF/kg PS were detected

inside the combustion apparatus without HM. The

Table 5

Comparison of PCDD/Famountsresultingfrom two similarcombustion experimentswith CuOinvolved(lg/kgPS)

Quartzboat

Glasswool plug

Liebigcooler

Glass ﬁberﬁlter

Absorption bottle

P TeCDD

< 0.2

7.96

1.13

11.3

< 0.2

3.75

< 0.2

P PeCDD

< 0.2

12.6

< 0.2

6.68

< 0.2

5.77

< 0.2

P HxCDD

< 0.2

16.5

< 0.2

8.83

< 0.2

15.5

< 0.2

5.60

< 0.2

P HpCDD

< 0.2

102

4.78

58.8

< 0.2

170

3.85

28.2

< 0.2

OCDD

< 0.2

865

25.5

402

< 0.2

1240

13.0

325

< 0.2

P TeCDF

< 0.2

102

13.6

91.7

< 0.2

96.9

9.54

69.1

< 0.2

P PeCDF

< 0.2

215

6.69

77.1

< 0.2

190

7.74

63.2

< 0.2

P HxCDF

< 0.2

588

18.4

215

< 0.2

651

22.6

169

< 0.2

P HpCDF

0.398

2780

61.1

1050

0.170

0.557

2410

36.6

706

0.242

OCDF

5.92

28800

564

6470

0.334

6.30

22900

287

6910

4.64

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika: