Formation of polybrominated dibenzofurans duringextrusion of high-impact polystyrene/decabromodiphenylether/antimony(III) oxideCitation for published version (APA):Luijk, R., Govers, H. A. J., & Nelissen, L. N. I. H. (1992). Formation of polybrominated dibenzofurans duringextrusion of high-impact polystyrene/decabromodiphenyl ether/antimony(III) oxide. Environmental Science &Technology, 26(11), 2191-2198. https://doi.org/10.1021/es00035a018

DOI:10.1021/es00035a018

Document status and date:Published: 01/01/1992

Document Version:Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)

Please check the document version of this publication:

• A submitted manuscript is the version of the article upon submission and before peer-review. There can beimportant differences between the submitted version and the official published version of record. Peopleinterested in the research are advised to contact the author for the final version of the publication, or visit theDOI to the publisher's website.• The final author version and the galley proof are versions of the publication after peer review.• The final published version features the final layout of the paper including the volume, issue and pagenumbers.Link to publication

General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright ownersand it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal.

If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, pleasefollow below link for the End User Agreement:

www.tue.nl/taverne

Take down policyIf you believe that this document breaches copyright please contact us at:

openaccess@tue.nl

providing details and we will investigate your claim.

Download date: 02. Feb. 2020

Envlron. Scl. Technol. 1992, 26, 2191-2198

(21) W e n , M. J. Surfactants and Interfacial Phenomenon, 2nd

(22) Huisman, H. F. K. Ned. Akad. Wet., Proc. Ser. B 1964,67,

(23) Anderson, M. A. Agron. Abstr. 1991, 237. (24) Chiou, C. T.; Kile, D. E.; Rutherford, D. W. Enuiron. Sci.

(25) Brusseau, M. L. Enuiron. Sci. Technol. 1991,25,1747-1752.

(26) Anderson, M. A,; Parker, J. C. Water, Air, Soil Pollut. 1990, ed.; John Wiley and Sons: New York, 1989; pp 170-206.

388-395.

1-18.

Received for review April 3,1992. Revised manuscript receiued June 29,1992. Accepted July 8,1992. This research was partially supported by a grant from the Kearney Foundation of Soil Science.

Technol. 1991,25, 660-665.

Formation of Polybrominated Dibenzofurans during Extrusion of High-Impact Polystyrene/Decabromodiphenyl Ether/Antimony ( I I I) Oxide

Ronald LulJk,*gt Harrle A. J. Govers,t and Laurent Nellssent

Department of Environmental and Toxicological Chemistry, University of Amsterdam, Nieuwe Achtergracht 166, 1018 WV Amsterdam, The Netherlands, and Eindhoven University of Technology, Centre for Polymers and Composites, P.O. Box 513, 5600 MB Eindhoven, The Netherlands

Polybrominated dibenzofurans (PBDFs) are formed during extrusion of high-impact polystyrene (HIPS) con- taining the flame retardant system decabromodiphenyl ether/antimony(III) oxide (Br,,,DPO/Sb203). The process of formation of PBDFs was examined at an extrusion temperature of 275 "C. To get a better understanding of processes taking place in the polymer melt during extru- sion) the thermal stability of polystyrene chains was also examined. It was found that the flame retardant system Br&PO/Sb203 shows a stabilizing activity toward ther- mal degradation of polystyrene chains. Temperature de- velopment during extrusion processing of HIPS/ Br1&PO/Sb2O3, simulated in a computer model, supports a mechanism of formation of PBDFs which is initiated by thermal degradation of polystyrene. Polystyrene degra- dation products, i.e., (po1y)styrene radicals, are stabilized by radical scavenging properties of the flame retardant decabromodiphenyl ether (Br&PO). Debromination of Brl&PO is combined with a ring-closure reaction forming PBDFs. It seems inevitable that PBDFs are formed during extrusion of polybrominated diphenyl ethers in addition polymers such as polystyrene. A tentative quantification of the yield of PBDFs in HIPS/BrloDPO/Sb203 by RP- HPLC showed an increase from 1.5 to 9 ppm for hepta- bromodibenzofuran and from 4.5 to 45 ppm for octa- bromodibenzofuran after four extrusion cycles.

Introduction Polybrominated flame retardants are widely used in

synthetic polymers to reduce inflammability. Due to their universality and minor influence on mechanical properties, polybrominated aromatic compounds have a broad ap- plication area. Despite technical advantages in applica- tions, the use of polybrominated compounds in polymers has two disadvantages: First, formation of corrosive HBr during compounding (2' = 190-225 "C) and second, but more important, formation of polybrominated dibenzo-p- dioxins (PBDDs) and -dibenzofurans (PBDFs) during thermal degradation (T = 350-800 "C). The formation of PBDDs and PBDFs has been shown during pyrolysis of pure brominated flame retardants (1-3) and during py- rolysis of polymers with brominated flame retardant ad- ditives (4, 5) . Especially brominated diphenyl ethers (PBDPOs) are reactive precursors in the formation of

University of Amsterdam. 3 Centre for Polymers and Composites.

PBDDs and PBDFs (6, 7). Equation 1 shows the struc- tural similarity of PBDPOs and PBDFs/PBDDs. For- mation of PBDFs consists of a ring-closure reaction with the elimination of a small molecule (Br2, HBr, or H2). An extra oxidation reaction is needed for the formation of PBDDs.

X ' + y ' i S

In this study the thermal behavior of PBDPOs during extrusion (melt compounding) of high-impact polystyrene (HIPS) was examined. HIPS is a blend of polystyrene and polybutadiene (6-12 wt %, dispersed in the polystyrene matrix) of which inflammability is reduced by the flame retardant system decabromodiphenyl ether/antimony(III) oxide (Br,oDPO/Sb203). From pyrolysis studies, it is known that mainly PBDFs are formed from PBDPOs during thermal degradation of HIPS/Br1,DPO/Sb2O3. The formation of PBDFs occurs during polymer degra- dation (2' = 350-400 "C) (8). Recently, formation of PBDFs during compounding of polymers with poly- brominated flame retardants has been shown (9-11). Brenner et al. (12) measured PBDFs in the workplace area during compounding of poly(buty1ene terephthalate) (PBTP)/glass fiber/Br,oDPO/Sb203 polymer blends. They reported total amounts of 34 ng/m3 tetrabromodi- benzofurans (Br,DFs), 143 ng/m3 pentabromodibenzo- furans (Br,DFs), 554 ng/m3 hexabromodibenzofurans (Br,DFs), and about 200 ng/m3 heptabromodibenzofurans (BqDFs). PBDDs were also found but a t much lower concentrations than PBDFs. Around the extruder head the amount of 2,3,7&substituted PBDF congeners was tentatively determined at 1.3 ng/m3 for Br,DFs and 2.6 ng/m3 for Br6DFs. Assuming equal toxicity of bromine- and chlorine-substituted dioxins and furans (13,14), one can estimate the contribution of 2,3,7,84etrachlorodi- benzo-p-dioxin equivalents (toxic equivalents or TEQs) using the toxic equivalence factor method (15). The amount of TEQs in the surroundings of the extruder head during extrusion of PBTP/glass fiber/Br1&PO/Sb2O3 was estimated at 330-910 pg/m3 (12). The tolerable daily

0013-936X/92/0926-2191$03.0010 0 1992 American Chemical Society Environ. Sci. Technol., Vol. 26, No. 11, 1992 2191

intake (TDI), recently established at 10 pg of TEQ kg-l dag ' in The Netherlands (16), can be converted into a proposed occupational exposure limit (OEL) concentration for toxic equivalents in the workplace area:

OEL = - - TDI X body weight breath volume

10 pg kg-' day-l X 70 kg 10 m3/day

- - - 70 pg/m3 (2)

The factors of 70 kg of body weight for an average worker and a l@m3 breath volume in an &h working day are taken from the method of Leung et al. (17). During extrusion of PBTP/glass fiber/Brl$PO/Sbz03, the proposed OEL for toxic equivalents is thus exceeded by at least a factor of 5.

This study focuses in detail on processes which occur during compounding of BrloDPO/Sb2O3 into the HIPS polymer matrix. A better understanding of extrusion processes will give a probable explanation for the formation of PBDFs at relatively low temperatures. During com- pounding HIPS is melted in an extruder with barrel tem- peratures of 190-220 "C. BrloDPO/Sb203 is mixed ho- mogeneously into the polymer matrix. The processes taking place during compounding of HIPS/BrloDPO/ SbzO3 were studied by using the following experiments:

(1) Formation of PBDFs was studied off-line through RP-HPLC analysis as function of the number of extrusion cycles ( n = 1-4) at T = 275 "C.

(2) Temperature development in a standard extruder was modeled using a computer simulation program.

(3) Thermal stability of polystyrene chains was studied by measuring the relative number-average molar mass ( M J , the relative mass-average molar mass (Mw), and the breadth of the molar mass distribution (Mw/Mn or dis- persivity index). After repeated extrusion cycles, these quantities were measured with gel permeation chroma- tography (GPC).

The formation of PBDFs will be described as a logical consequence of polymer degradation processes in HIPS/ Brl,$PO/Sb203 under extrusion conditions. It is the flame retardant activity of decabromodiphenyl ether that will give an explanation for formation of PBDFs at relatively low temperatures.

Experimental Section Materials, These materials were used: HIPS/

BrloDPO/Sb,03 (Dow Chemical Co., Terneuzen, The Netherlands); decabromodiphenyl ether, octabromodi- phenyl ether, and pentabromodiphenyl ether (Broom- chemie B.V., Terneuzen, The Netherlands); homopolymer polystyrene M , N 100 000 (BDH Chemicals Ltd., Poole, England); homopolymer polystyrene (Shell Chemie Ned- erland B.V., Rotterdam, The Netherlands); 13C-labeled 2,3,7,8-Br4DD, 2,3,7,8-Rr4DF, 1,2,3,7,8-Br5DD, and 1,2,3,7,8-BrSDD and I2C-labeled 2,3,7,8-Br4DD, 2,3,7,8- Br4DF, 1,2,3,7,8-Br5DD, 1,2,3,7,8-Br5DD, 1,2,3,4,7,8- 1,2,3,6,7,& 1,2,3,7,8,9-Br6DD, 1,2,3,4,7,8-Rr,DF, Br8DD, and their chlorinated analogues (Cambridge Isotope Labora- tories, Woburn, MA). Reagents which were used in PBDD and PBDF cleanup procedures were similar to those used in PCDD and PCDF cleanup procedures (18).

Extrusion Conditions and Processing Equipment. Repeated extrusion cycles of homopolymer PS and HIPS/BrloDPO/Sb203 (1-4 cycles) were performed on a Bersdorff ZE25 corotating twin-screw extruder (screw diameter 25 mm, screw speed 100 rpm). Temperature settings were T1 = 180 "C, T2 = 200 "C, T3-T7 = 220 "C for a barrel temperature of 220 "C and TI = 180 "C, T2

2192 Environ. Sci. Technol., Vol. 26, No. 11, 1992

"\ c--_- 3 , ____ - 1

BraDPO BrsDPO Br,,DPO 0 00

Br,DF Br,DF 0 0

Br,DF Br,DF 0 0

Br,DD 0

I I l I l 1 1 1 1 1 1 1 1 1 1 1 1 1 I l l

60 70 80 90 i o 0 rnin Flgun 2. RP-HPLC retentlon times of bromlnated dibenzo-pdioxlns, dibenzofurans, and diphenyl ethers.

beled standards available for these compounds at the moment. Consequently, quantitative analysis of these compounds is attended with a set of assumptions in rela- tion to recoveries of extraction and extract cleanup pro- cedures and analytical response factors. Several studies report on the use of electrophilic aromatic bromination of the parent compounds [ 12C]dibenzofuran and [12C]di- benzo-p-dioxin (23,24) for detection and quantification purposes. But even the availability of a set of [12C]PBDFs cannot solve the introduced inaccuracies in quantitative GC/MS analysis of hexa- to octabromodibenzofura. For optimal quantification I3C-labeled internal standards are required.

Because of the lack of I3C-labeled internal standards and observed complications in GC/MS analysis of PBDFs, RP-HPLC offers in our opinion an appropriate alternative for GC/MS analysis of hexa- to octabromodibenzofurans. GC/MS analysis of PBDFs in polymer samples is accom- panied with complications regarding their chromatoeaphy. Debromination reactions occur on the column. Peaks are tailing, which is probably caused by the presence of styrene oligomers. Moreover, PBDPOs with the same retention times interfere. In the mass-detection part fragments of PBDPOs (-Br2) with masses similar to PBDFs can in- terfere. In the higher mass range ( m / z 600-800), the sensitivity decreases. In addition, assumptions for as- sessing response factors have to be made. Therefore, an accurate quantification of hexa- to octabromodibenzo- furans cannot be carried out.

Most of these complications do not appear in RP-HPLC analysis. Detection in RP-HPLC with UV absorbance or fluorescence is based upon aromaticity rather than the number of bromine substituents. Therefore, a minor discrimination in response factors for PBDFs is expected. In this study we want to emphasize the main formation processes of PBDFs during extrusion of HIPS/ BrlJ3PO/Sb20,. Therefore, RP-HPLC was used as a qualitative analysis instrument to determine relative in- creases in the yield of hepta- and octabromodibenzofurans. The yield of hexabromodibenzofurans was too low to be measured with semipreparative RP-HPLC. Assuming independency of the response factor on a mole base and the number of bromine substituents, a quantitative analysis of PBDFs with RP-HPLC may be developed.

RP-HPLC analysis was performed with a Beckman HPLC system: pump, Model llOA; eluent, MeOH; flow rate, 2.3 mL/min; detector, Beckman 160 absorbance de- tector (254 nm); precolumn, Chrompack (6 cm X 3.8 mm);

Table I. log k' of Halogenated Dibenzo-p-dioxins and Dibenzofurans

brominated compd log k' chlorinated compd

Dibenzofurans 2,3,7,8-Br4DF 0.442 2,3,7,8-CdDF 1,2,3,7,8-BrSDF 0.766 1,2,3,7&ClbDF 1 ,2,3,4,7,8-BrsDF 0.859 1,2,3,4,7,8-CbDF

Dibenzo-p-dioxins 2,3,7,8-Br4DD 0.574 2,3,7,8-C14DD 1,2,3,7,8-Br5DD 0.711 1,2,3,7,8-C15DD 1,2,3,6,7,8-BrBDD 0.856 1,2,3,6,7,8-C&DD BreDD 1.09 CIBDD

log k'

0.292 0.638 0.804

0.422 0.556 0.697 0.934

column, Zorbax ODS (ClS, 5 pm, 25 cm X 9.4 mm); T = 30 "C. The eluent was collected in three fractions (0-35.5, 35.5-40.0, 40.0-90.0 min) in order to get a confirmation of PBDF assignment with GC/MS. The conditions of GC/MS confirmation analysis were described elsewhere (8).

Characterization Techniques: RP-HPLC Analysis Used for Prediction of Retention Behavior of PBDFs. Assignment of PBDFs in RP-HPLC analysis has been developed by extrapolating simple linear relationships between retention behavior and chemical structure. In prediction models for GC retention behavior of poly- chlorinated dibenzo-p-dioxins and -dibenzofurans based upon chemical structure, the number of chlorine substit- uents is often used as a model parameter (25,26). In a similar way, correlation has been made between RP-HPLC capacity factor k'and the number of halogen substituents (X = Cl, Br). Because there are many more chlorinated compounds available, correlation between brominated and corresponding chlorinated compounds is also of use in predicting retention behavior of unavailable brominated compounds.

The reversed-phase HPLC experiments were performed in triplicate with a Kratos Spectroflow 400 solvent delivery system, a Kratos Spectroflow 757 UV absorbance detector (A = 225 nm), and a Rheodyne Type 7125 injector. The capacity factor k ' of chlorinated and brominated aromatic compounds was measured on a reversed-phase C18 column ( H y p e d ODs, 5 pm, length 20 cm, N = 3100). The eluent was methanol/water (955 v/v). The flow rate was 1 mL/min. The dead volume or to w a ~ measured with uracil.

The logarithm of capacity factors [log k' = log (t, - to/to)] of halogenated dibenzo-p-dioxins and dibenzofurans are listed in Table I.

Linear relationships between retention behavior and the number of halogen substituents are dibenzofurans

log k b , = -0.354 + 0.209nB,

log kbl = -0.702 + 0.256nc1

(3)

(4)

n = 3 R2 = 0.907 SER = 0.094

n = 3 R2 = 0.960 SER = 0.073 dibenzo-p-dioxins

log kbr = 0.066 + 0.129nB,

n = 4 R2 = 0.998 SER = 0.013 log kbl = -0.084 + 0.128nc1

n = 4 R2 = 0.998 SER = 0.011

(5 )

(6)

Linear relationships in RP-HPLC retention character- istics of brominated and chlorinated compounds are

Envlron. Scl. Technol., Vol. 20, No. 11, 1992 2195

dibenzofurans log k i r = 0.207 + 0.833 log kLl (7)

n = 3 R2 = 0.988 SER = 0.034 dibenzo-p-dioxins

log khr = 0.150 + 1.01 log ktCl (8) n = 4 R2 = 0.999 SER = 0.002

Although there are a few standards available, the re- tention behavior of brominated dibenzo-p-dioxins and dibenzofurans showed rather high correlation with the number of halogen substituents (eqs 3-6). The retention behavior of PBDFs in semipreparative RP-HPLC analysis showed a similar linearity with the number of bromine substituents. In addition, a mixture of PBDFs from electrophilic aromatic bromination of dibenzofuran was used as a control (27). Therefore, we have extrapolated eq 3 to assign hepta- and octabromodibenzofuran. Al- though there are four heptabromodibenzofuran isomers, only one peak was detected. In the GC/MS confirmation analysis also, one main peak was detected for hepta- bromodibenzofuran. Linear relationships between the retention behavior of brominated and chlorinated com- pounds, shown in eqs 7 and 8, offer the possibility to predict capacity factors of brominated compounds. This is under further investigation.

2. Modeling Temperature Development in an Ex- truder. Modeling of the overall temperature profile of a single-screw devolatilization extruder was carried out by Dr. P. Elemans from DSM Research, Geleen, The Neth- erlands. The computer simulation model is based upon extruder geometries and rheological characterization of extruded polymers (28). Flow curves of HIPS/ Br1,$PO/Sb2O3 were measured on a Goffert Rheograph 2002, L/D ratio of the capillary was 30:l. Test tempera- tures were 180, 200, and 220 "C. The shear rate varied between y = 100 and 1000 (l/s). The consistency index vo was measured on a Rheometrics plate/cone viscosimeter.

3. Thermal Stability of HIPS/Br1,DPO/Sb2O3. Characterization Techniques: Gel Permeation Chromatography. Gel permeation chromatography (GPC) was used to determine the relative number-average molar mass (M,,), the mass-average molar mass (Mw), and the breadth of the molar mass distribution (Mw/Mn) of HIPS/BrloDPO/Sb203. After each extrusion cycle (see above) a sample was collected. GPC was performed by using a Waters apparatus composed of a pump, Model 510; injector, WISP 711; UV detector, 254 nm Model 440; column, lo5, lo4, or lo3 p-Styragel(40 "C); columns, cal- ibrated with monodispersive PS standards; and by using tetrahydrofuran (THF) 0.6 mL/min as eluent. The polymer samples were dissolved in THF, filtered, and in- jected on columns. The thermal stability of polystyrene chains in homopolymer PS and in HIPS/BrlJIPO/Sb2O3 was determined by measurement of the breadth of the molar mass distribution (or dispersivity index M,/M,,).

Results and Discussion Formation of PBDFs during Extrusion of

HIPS/BrloDPO/Sb203. Figures 3 and 4 show the rela- tive yield of heptabromodibenzofuran (Br,DF) and octa- bromodibenzofuran (Br8DF) formed during repeated ex- trusion cycles (n = 1-4) of HIPS/Br1,$PO/Sb2O3 at 275 "C. The yields of both Br7DF and Br8DF show a signif- icant increase as a function of the number of extrusion cycles.

The yield of Br7DF increases with factor 6 and that of octabromodibenzofuran with factor 10, from 0-1 to 2-4

1.

0 1 2 3 4

number of extrusion cycles

- Br,DF

Flgwe 5. Formation of BrPF during exbusion of HIPS/Br,pPO/Sb,O, at 275 O C .

0 1 2 3 4

nwnber of extrusion cycles

- Br,DF

Ftgure 4. Formation of Br,DF during extrusion of HIPS/Br,pPO/Sb,O, at 275 O C .

extrusion cycles. Large error bars in the duplicate analysis are explained by inaccuracy of extraction and extract cleanup procedures. The results indicate a ceiling level of PBDFs after repeated extrusion cycles. This may be explained as an equilibrium between formation and de- bromination of PBDFs. Tentative quantification of the yield of PBDFs based upon comparison with RP-HPLC peak area of a known quantity of octabromodibenzo-p- dioxin resulted in concentrations of octabromodibenzo- furan ranging from 4.5 to 45 ppm and heptabromodi- benzofuran ranging from 1.5 to 9 ppm. In this quantifi- cation, it is assumed that overall recoveries for Br7DF and Br8DF in extraction and extract cleanup procedures are 30%. The RP-HPLC chromatogram showed only one heptabromodibenzofuran peak. This was confirmed by GC/MS analysis where again only one hexabromodi- benzofuran was detected. These results indicate that se- lective debromination processes of PBDPOs and PBDFs take place in the polymer melt during extrusion.

The starting material already contained small amounts of Br,DF and Br8DF. Most probably they are formed during industrial compounding of the flame retardant system BrlODPO/Sb2O3 into the matrix of the HIPS polymer blend. Alternatively, the presence of PBDF in the starting material may have been a background pollu- tion of BrloDPO (29, 30).

In the RP-HPLC chromatogram, decabromodiphenyl ether and less brominated diphenyl ethers were present. These less brominated diphenyl ethers are formed by the exchange of bromine and hydrogen in BrIoDPO (31).

2194 Environ. Scl. Technol., Vol. 26, No. 11, 1992

Hydrogen most probably is abstracted from the polymer backbone, The exchange reaction of bromine and hydro- gen in PBDPOs plays a key role in explaining formation of PBDFs at relatively low extrusion temperatures and will be discussed further in sections 2 and 3.

From micropyrolysis of pure polybrominated diphenyl ethers it is known that reactivity toward formation of PBDDs and PBDFs depends on the degree of bromination ( I , @ . Less brominated diphenyl ethers like Br6DP0 give rise to a much higher yield of PBDDs and PBDFs than Brl,$PO does. This phenomenon is explained by ener- getically favorable elimination of HBr in the ring-closure reaction shown in eq 1. During pyrolysis of pure PBDPOs in a gas-phase reaction, the optimum PBDD/PBDF for- mation temperature is 600-700 OC. During macropyrolysis of HIPS/Brl,$PO/Sb203 only PBDFs are formed. The HIPS matrix decreases the optimum PBDF formation temperature from 600-700 to 350-400 O C and enhances the yield of PBDFs by a factor of 7 (8). Molecular pro- cesses taking place during thermal degradation of the HIPS/Br16>PO/Sb203 polymer blend are debromination of the flame retardant Brl,,DPO forming less brominated diphenyl ethers; depolymerization of polystyrene by @- scission forming mainly styrene monomers, dimers, and trimers; bromination of polystyrene; and formation of antimony(II1) bromides and antimony(II1) phenoxy- bromides (31). It has been shown that formation of PBDFs takes place during degradation of the HIPS polymer blend by a condensed-phase mechanism (8).

To explain the formation of PBDFs during extrusion of HIPS/Br1$PO/Sb2O3 at relatively low temperatures, more insight in temperature development in the extruder and in the thermal stability of polymer chains during ex- trusion processes is needed. In the next section temper- ature development in a model-type extruder will be dis- cussed.

2. Modeling Temperature Development in an Ex- truder. The temperature (T) development in model-type extruders can be calculated with a computer simulation program which is based mainly on extruder geometrics and rheological characterization of extruded polymer systems. This computer model is developed by Meijer and Elemans (28). The temperature development of HIPS/ Brl0DPO/Sb2O3 is calculated in a single-screw extruder. It is supposed that this extruder is a good model system to simulate conditions during industrial compounding of the flame retardant system Brl&PO/Sb203 into an HIPS polymer matrix.

Rheological parameters like the power law index n, the consistency index qo, and the temperature index b (Ar- rhenius constant) were used in the computer model to calculate the power consumption and the temperature development in the pumping zone of the extruder, i.e., the part filled with the melt. The extruder throughput was set on 75 kg/h with a screw speed of 200 rpm. The tem- perature and shear rate dependent viscosity q (Pa s) is described using a power law

where n is the power law index, qo is the consistency index (Pa s), b is the temperature index (Arrhenius constant) ( O W ) , To is the reference temperature ("C), and y is the shear rate (s-'). The flow curve in Figure 5 shows the viscosity (log q) of HIPS/Br1&PO/Sb2O3 as a function of shear rate (log y). In the range 100 < y < 1000 it follows readily that the power law index n is given by

(10) For HIPS/Br1&PO/Sb2O3 the power law index n is 0.26.

q = qo,-b(T-To)yn-l (9)

n = d(1og q)/d(log y) + 1

1.00€+04

16

1.00E+04 1.00Ec03 1 .OOE+O 1 1.00€+02

Shear rate L 7/51 1 - Flgue 5. HIPS/Br,&FQ/Sb,03 viscosrty q as a function of shear rate 7.

2301 280 I I

I I / 200 v 0 2 4 6 8 10 12 I 4

Length (LID1

Flgure 6. Simulation of temperature and pressure development in a single-screw extruder during compounding of HIPS/BrloDPOISb,O, under given condltlons (for detailed technical data see Table 11).

From the flow curves at different temperatures (T = 200, 220, and 240 "C), the temperature dependency of the viscosity, expressed in the Arrhenius constant b, can be calculated. The shear rate is kept constant, and the vis- cosity is measured as a function of temperature. For HIPS/Brl,,DPO/Sb20, the Arrhenius constant b is 0.02 O C - l . The last important rheological parameter is the consistency index lo, i.e., the viscosity a t shear rate log y = 0 or y = 1. For HIPS/Br1,$PO/Sb2O3 the consistency index v0 is 3.64 X lo4 Pa s. The values of these rheological parameters were used in the computer simulation program for temperature development in a model-type extruder. The results of the computer calculations are shown in Table I1 and in Figure 6.

If the barrel temperature is set on 200 "C, which is a commonly used extrusion temperature for HIPS/ Brl,,DPO/Sb203, the temperature in the main channel rises to 260 OC. At the extruder flights, especially in high shear pumping zones, the temperature can rise to above 400 "C. It is likely that the material passes these local hot spots several times with an estimated mean residence time of a few seconds (32). Temperature development in high shear pumping zones rather depends on the viscosity of the melt. At higher barrel temperatures, viscosity will be lower, and consequently, the temperature difference be- tween main channel and flights decreases. In general the temperature a t the flights can rise to approximately 100

Environ. Sci. Technoi., Vol. 26, No. 11, 1992 2195

6 r I

Kdp L I I

0 100 200 300

Temperature ["C] F W e 7. Varlatlon of polymerization rate K,[M] and depolymerization rate K, with temperature for styrene.

"C above the temperature of the bulk material in the main channel (33).

Now that we have gained more insight into temperature development in a model-type extruder, it is interesting to evaluate the two processes occurring during extrusion of HIPS/Brl$PO/Sb20,: (i) debromination of Brl$PO to less brominated diphenyl ethers and (ii) formation of PBDFs. The first process, debromination of BrloDPO forming less brominated diphenyl ethers, can possibly be explained by the calculated temperature development a t kneading flights in a model single-screw extruder. Strie- bich et al. (34) reported that homogeneous decomposition of pure BrloDPO already starts a t 400-425 "C. In heter- ogeneous systems such as compounding of Brl$PO into polymer matrices at relatively low temperatures of 190-220 "C, corrosive HBr can be formed. This indicates debro- mination of BrloDPO. Another heterogeneous process concerns thermal instability of high brominated organic compounds in GC analysis. Debromination of poly- brominated dibenzo-p-dioxins, dibenzofurans, and di- phenyl ethers was observed in the GC/MS confirmation analysis at temperatures above 300 "C on a relatively inert chemically bonded DB-5 column. The second process, formation of PBDFs, cannot be explained entirely by temperature development in the extruder. The difference in temperature between optimum PBDF formation in an homogeneous gas-phase reaction (7' = 600-700 "C) and local temperatures a t extruder flights (T = 400 "C) is more than 200 "C. Although it has been shown that ring-closure reactions of PBDPOs in an HIPS polymer blend take place during thermal degradation of the HIPS matrix at 350-400 "C (8), more information about the thermal stability of the polymer matrix under extrusion conditions is needed. Examining thermal (in)stability, e.g., decomposition re- actions of polymer chains during extrusion, will give a probable explanation for formation of PBDFs. In the next section the thermal stability of HIPS/Brl$PO/Sb203 will be compared with that of homopolymer PS after repeated extrusion cycles.

3. Thermal Stability of HIPS/BrloDPO/Sb203. The mechanism of polystyrene degradation is a radical depolymerization. The depolymerization process can be imagined as the unzipping of polymer chains initiated by radicals. A theoretical consideration of polymerization versus depolymerization results in a thermodynamically defined ceiling temperature (TJ at which the polymeri- zation rate Kp(M] equals the depolymerization rate Kdp (35). At thermodynamic equilibrium (AG = 0):

T, = AH/(R In [MI, + AS) (12) Optimum polymerization temperature Tp. and ceiling

temperature T, of polystyrene are shown in Figure 7. While the ceiling temperature is 300 "C, polystyrene al- ready starts to depolymerize at 200 "C. Under given ex- trusion conditions, the temperature in HIPS/Br,&PO/ Sb203, a8 discussed in the previous section, locally rises to 400 "C. Therefore, radical (de)polymerization processes occur during extrusion of polystyrene blends.

To get a better understanding of the thermal stability of polystyrene, the homopolymer PS and HIPS/ Brl$PO/Sb203 have been repeatedly extruded (n = 1-4) under normal temperature conditions ( Th, = 220 "C) and above practical temperature conditions = 275 "C). The thermal stability of polystyrene chains in homo- polymer PS and in HIPS/Brl$PO/Sb203 was compared using the breadth of the molar mass distribution (or dis- persivity index M,/M,) (36). The value of the dispersivity index is a measure for the chain length distribution. In contrast to the relative number-average molar mass M,, where all polymer chains are of equal weight, the relative mass-average molar mass M , outweighs the contribution of longer polymer chains. Therefore, a change in disper- sivity index indicates formation or degradation of polymer chains.

Table I11 shows that the polymer chain length distri- bution in homopolymer PS and in HIPS/Brl$PO/Sb20, remain unchanged during extrusion at 220 and 275 "C. After repeated extrusion of the homopolymer PS at 220 "C, M, and M, show a slight decrease of 5%. After re- peated extrusion of homopolymer PS at 275 "C, both M , and M, show a significant decrease of 20%. This indicates that polystyrene depolymerizes during extrusion at 275 OC and that styrene oligomers are formed. HIPS/ Brl$PO/S~03 shows a relatively smaller decrease in M, and M, of 5 4 % after extrusion at 275 "C. Depolymeri- zation of polystyrene is stabilized in HIPS/Brl,,DPO/ Sb203. An explanation for the stabilizing activity of BrloDPO/Sb20, is found in its flame retardant activity based on scavenging the H' and OH' propagation radicals in a real fire situation. During extrusion a nearly identical radical stabilization of HIPS degradation products is ex- pected to occur. Radicals which start the unzipping of polystyrene chains are scavenged by Brl$PO/Sb20s. The radical stabilizing activity of Br,$PO is accompanied by debromination of Brl$PO and formation of more stable Br' radicals. This mechanism forming less brominated diphenyl ethers is in good agreement with the observation of Brauman and Chen (37), who found that degradation of the HIPS polymer matrix initiates the autocatalytic debromination of Brl$PO. Debromination of Br,&PO, initialized by degrading polystyrene chains, is the fmt step to PBDF formation. Therefore, incorporation of Brl$PO into a thermally degrading polymer matrix is necessary for the low-temperature formation of PBDFs.

Conclusions and Recommendations After repeated extrusion cycles of HIPS/Br,,,DPO/

Sb203, a significant increase in the yield of PBDFs was shown. A RP-HPLC analysis method was used to describe the relative increase of hepta- and odabromodibenzofuran. After four extrusion cycles a t 275 "C, the yield of octa- bromodibenzofuran increases by a factor of 10 (tentatively quantified as an increase from 4.5 to 45 ppm). The yield of heptabromodibenzofuran increases by a factor of 6 (tentatively quantified as an increase from 1.5 to 9.0 ppm). The presence of PBDFs in the starting material most probably is explained by the formation of PBDFs during

2186 Envlron. Sci. Technol., Vol. 26, No. 11, 1992

Table 11. Modeling of Single-Screw Devolatilization Extruder

material HIPS/Br$PO/Sb203 melt index (dg/min) 4.50 screw diameter (mm) 60.0 screw flight clearance cold (mm) 0.20 kneading flight clearance cold (mm) 0.50 melt throughput (kg/h) 75.0 screw speed (rpm) 200 barrel temperature ("C) 200

temperature index (l/"C) 0.020

drive power consumption (kW) 5.7

total cooling power (kW) -1.9 total pumping power (kW) -0.1

consistency index (Pa s) 36400

power law index 0.26 screw power consumption (kW) 5.1

specific energy drive (kWh/kg) 0.076

die pressure (MPa) 0.4 die melt temperature ("C) 260.0

com- feeding pressing pumping

length (mm) 240 240 360 pitch angle (deg) 17.7 17.7 17.7 channel number 1 1 1 channel depth (mm) 10.0 7.5 5.0 channel width (mm) 51.2 51.2 51.2 screw flight width (mm) 6.0 6.0 6.0 kneading flight width (mm) 0.0 0.0 0.0 channel power (kW) 0.9 0.8 1.9

kneading flight power (kW) 0.0 0.0 0.0

pumping power (kW) 0.0 0.0 -0.1

screw flight power (kW) 0.7 0.4 0.5

cooling power (kW) -0.1 -0.4 -1.4

filling factor 0.30 0.39 0.74 pressure (MPa) 0 0 0.4 melt temperature channel ("C) 212.4 219.5 225.8 melt temperature flight ("C) 267.1 284.1 408.1 melt temperature out ("C) 227.1 243.6 260.0

Table 111. Dispersivity Indexes of HIPS/BrloDPO/Sb208 and Homopolymer PS after Repeated Extrusion Cycles

cycle T M, M, polymer no. ("C) ()(lo3) (x103) M,/M,

HIPS/Brl&PO/Sb203 1 2 3 4

PS 1 2 3 4

PS 1 2 3 4

275 275 275 275 220 220 220 220 275 275 275 275

88.8 81.8 83.3 81.8

100 100 94 95 88 90 75 68

183 180 181 173 281 281 274 267 256 250 220 202

2.1 2.2 2.2 2.1 2.8 2.8 2.9 2.8 2.9 2.7 2.9 3.0

industrial compounding of decabromodiphenyl ether into the HIPS matrix. Another explanation may be contam- ination of decabromodiphenyl ether.

The formation of PBDFs from decabromodiphenyl ether during extrusion of HIPS could be explained by the in- trinsic features of the polymer blend; i.e., radical degra- dation of polystyrene chains is stabilized by the radical scavenging activity of decabromodiphenyl ether. Reactive radicals are transformed into leas reactive bromine radicals. However, the reactivity of decabromodiphenyl ether in the formation of PBDFs is increased because debromination is followed by a ring-closure reaction to dibenzofurans. The formation of PBDFs is initiated by radical degradation of polystyrene. Therefore, optimum temperature of for- mation of PBDFs in polymer blends containing PBDPOs

is determined by the thermal stability of the polymer matrix.

When the results of this model study are translated to industrial processes such as homogeneous mixing of de- cabromodiphenyl ether into polymer matrices or shaping a final product, it seems inevitable that PBDFs are formed and partially released into the surrounding area. It is of great importance to get an impression, by monitoring programs, of the possible concentrations and exposure levels in the near surroundings of extruders. The toxicity of PBDFs makes an occupational exposure limit necessary. Exceeding this limit, three possible actions can be taken to reduce occupational exposure. First, extrusion tem- perature can be reduced to a minimum value to reduce formation of PBDFs. However, the results of this study indicate that formation of PBDFs in extrusion processing cannot be avoided entirely due to intrinsic features of addition polymers containing PBDPOs. Second, local exhaust ventilation can be used to control hazardous air contaminants, and third, the use of PBDPOs can be lim- ited. Further research is required in order to examine whether these measures are sufficiently effective.

The extract cleanup procedure may be improved by the use of a preparative RP-HPLC step instead of a carbon column. Because of the low solubility of decabromodi- phenyl ether in organic solvents, a large injection volume is necessary. Great differences in polarity give rise to a complete separation of hexa- to octabromodibenzofurans from PBDPOs. It is worthwhile to apply a nondestructive detection method to determine recoveries of hepta- and octadibenzofuran in the RP-HPLC step. As long as there are no I3C-labeled internal PBDF standards available, RP-HPLC is preferable to GC/MS analysis.

Acknowledgments We sincerely thank Dr. P. H. M. Elemans, DSM Re-

search BV, Geleen, The Netherlands, for his support on the rheological characterization of HIPS/Brl,-,DPO/Sb203 and the calculations of temperature development in a model-type extruder.

Registry No. BrloDPO, 1163-19-5; Sb209, 1309-64-4; Br7DF, 62994-32-5; Br8DF, 103582-29-2; polystyrene, 9003-53-6; poly- butadiene, 9003-17-2.

Literature Cited (1) Buser, H. R. Environ. Sci. Technol. 1986, 20, 404-408. (2) Thoma, H.; Hutzinger, 0. Chemosphere 1987, 16,

(3) Thoma, H.; Rist, S.; Hauschulz, G.; Hutzinger, 0. Che- mosphere 1986, 15, 649-652.

(4) Dumler, R.; Thoma, H.; Lenoir, D.; Hutzinger, 0. Chemo- sphere 1989,19, 2023-2031.

( 5 ) Dumler, R.; Lenoir, D.; Thoma, H.; Hutzinger, 0. J. Anal. Appl. Pyrolysis 1989, 16, 153-158.

(6) Thoma, H.; Hauschulz, G.; Knorr, E.; Hutzinger, 0. Che- mosphere 1987, 16, 277-285.

(7) Dumler, R.; Lenoir, D.; Thoma, H.; Hutzinger, 0. Chemo- sphere 1990,20, 1867-1873.

(8) Luijk, R.; Wever, H.; Olie, K.; Govers, H. A. J.; Boon, J. J. Chemosphere 1991,23, 1173-1183.

(9) Donnelly, J. R.; Grange, A. H.; Nunn, N. J.; Sovocool, G. W.; Brumley, W. C.; Mitchum, R. K. Biomed. Environ. Mass Spectrom. 1989,18, 884-896.

(10) McAllister, D. L.; Mazac, C. J.; Gorsich, R.; Freiberg, M. Chemosphere 1990,20, 1537-1541.

(11) Thies, J.; Neupert, M.; Pump, W. Chemosphere 1990,20,

(12) Brenner, K. S.; Knies, H. Dioxin 90, 10th International Meeting, Short Papers Organohalogen Compounds; Ecoinforma Press: Bayreuth, Germany, 1990; Vol. 2, pp

1353-1360.

1921-1928.

319-324.

Environ. Sci. Technol., Vol. 26, No. 11, 1992 2187

Environ. Sci. Technol. 1992, 26, 2198-2206

(13) Nagao, T.; Neubert, D.; Loser, E. Chemosphere 1990,20,

(14) Schulze-Schalge, T.; Koch, E.; Schwind, K.; Hutzinger, 0.; Neubert, D. Chemosphere 1991,23, 1925-1931.

(15) van Zorge, J. A.; van Wijnen, J. H.; Theelen, R. M. C.; Olie, K.; van den Berg, M. Chemosphere 1989,19, 1881-1895.

(16) Dioxins and planar PCBs in food: Concentrations in food products and intake by the Dutch population: RIVM 730501.034; National Institute for Public Health and En- vironmental Hygiene: The Netherlands, 1991.

(17) Leung, H.; Murray, F. J.; Paustenbach, J. Am. Ind. Hyg.

(18) van Berkel, 0. M.; Olie, K.; van den Berg, M. Int. J . En- viron. Anal. Chem. 1988, 34, 51-67. Billmeyer, F. W. Textbook of Polymer Science, 3rd ed.; J.

1189-1192.

ASSOC. J . 1988, 49, 466-474.

Wiley & Sons: New York, 1986.- Wegman, R. C. C.; Freudenthal, J.; de Korte, G. H. L.; Groenemeijer, G. S.; Japenga, J. Chemosphere 1986, 15,

Donnelly, J. R.; Munslow, W. D.; Vonnahme, T. L.; Nunn, N. J.; Hedin, C. M.; Sovocool, G. W.; Mitchum, R. K. Biomed. Environ. Mass Spectrom. 1987, 14, 465-472. Unpublished results. Sovocool, G. W.; Munslow, W. D.; Donnelly, J. R.; Mitchum, R. K. Chemosphere 1987, 16, 221-224. Munslow, W. D.; Sovocool, G. W.; Donnelly, J. R.; Mitchum, R. K. Chemosphere 1987,16, 1661-1666.

1107-1 112.

(25) Donnelly, J. R.; Munslow, W. D.; Mitchum, R. KO; Sovocool, G. W. J. Chromatogr. 1987, 392, 51-63.

(26) Hale, M. D.; Hileman, F. D.; Mazer, T.; Shell, T. L.; Noble, R. W.; Brooks, J. J. Anal. Chem. 1985,57, 640-648.

(27) Unpublished results. (28) Meijer, H. E. H.; Elemans, P. H. M. Polym. Eng. Sci. 1988,

(29) Timmons, L. L.; Brown, R. D. Chemosphere 1988, 17,

(30) Hileman, F.; Wehler, J.; Wendling, J.; Orth, R.; Ritchie, C.;

(31) Luijk, R.; Govers, H. A. J.; Eijkel, G. B.; Boon, J. J. J. Anal.

(32) Manas-Zloczower, I.; T a b o r , Z. Adv. Polym. Technol. 1983,

(33) Winter, H. H. Polym. Eng. Sci. 1976, 20, 406-412. (34) Striebich, R. C.; Rubey, W. A.; Tirey, D. A,; Dellinger, B.

(35) Odian, G. Principles of Polymerization, 2nd ed.; John Wiley

(36) Young, R. J. Introduction to Polymers; Chapman and Halk

(37) Brauman, S. K.; Chen, I. J. J. Fire Retard. Chem. 1981,8,

28, 275-290.

217-233.

McKenzie, D. Chemosphere 1989,18, 217-224.

Appl. Pyrolysis 1991, 20, 303-319.

3, 213-221.

Chemosphere 1991,23, 1197-1204.

& Sons: New York, 1981; p 268.

London, 1981; p 8.

28-36.

Received for review June 11, 1992. Accepted July 7, 1992.

Transformation of Carbon Tetrachloride in the Presence of Sulfide, Biotite, and Vermiculite

Michelle R. Kriegman-Klng and Martin Reinhard"

Department of Civll Engineering, Stanford University, Stanford, California 94305-4020

I Carbon tetrachloride is transformed in aqueous solutions containing dissolved hydrogen sulfide more rapidly in the presence of the minerals biotite and vermiculite than in homogeneous systems. Approximately 8045% of the CC4 was transformed to COP via the measured intermediate, CS2. Chloroform comprised 5-15% of the products. The remaining 5% of the products were an unidentified non- volatile compound and CO. At 25 OC, the half-life of CCl, with 1 mM HS- was calculated to be 2600,160, and 50 days for the homogeneous, vermiculite (114 m2/L), and biotite (55.8 m2/L) systems, respectively. The CCl, transforma- tion rate was found to be dependent on the type and quantity of the solid and the temperature, but was inde- pendent of pH and HS- concentration above a critical HS- concentration. The activation energies (f95% confidence intervals) were determined to be 122 f 32,91.3 f 8.4, and 59.9 f 13.3 kJ/mol for the homogeneous, vermiculite, and biotite systems, respectively. The CCl, transformation rate exhibited first-order behavior with respect to biotite sur- face area concentration (SCbiotite) below 55.8 m2/L. The rate of CCll transformation was independent of HS- con- centration when [HS-] = 0.5-4 mM and SCbiotit, = 55.8 m2/L. Below [HS-] = 0.5 mM, the rate law was dependent on HS- concentration.

Introduction Abiotic transformations, such as reductive dehalogena-

tion and nucleophilic substitution, can influence the fate of halogenated aliphatic compounds in aqueous environ- ments. Sulfide, commonly found in hypoxic environments such as landfill leachate, hazardous waste plumes, and salt marshes at levels ranging from 0.2 pM to 5 mM, can act

as electron donor (1-3) or as a nucleophile (4-6) to promote transformation of halogenated organics. In subsurface environments, the transformation rates of halogenated organic compounds in aqueous solution may be influenced by mineral surfaces ( I , 2, 7, 8).

We investigated the effect of mineral surfaces in the presence of dissolved sulfide on the transformation rate of carbon tetrachloride. CCl, was chosen as the compound of interest because it is a contaminant which is frequently found in groundwater, it is a suspected human carcinogen (9), it causes ozone depletion, and it is a greenhouse gas (10). Laboratory studies were conducted to identify and quantify the environmental parameters that govern the transformation rate of CCl, in a heterogeneous aqueous environment containing sulfide and sheet silicates (biotite and vermiculite). The parameters studied were temper- ature, pH, mineral surface area, and sulfide concentration. The reaction rate was hypothesized to be a function of surface area because the reactions have been shown to be faster in the presence of mineral surfaces (2). Sulfide was also expected to play a role in the kinetics because it could act as either an electron donor or a nucleophile. Data are presented in terms of [HS-] instead of total sulfide because HS- is a much stronger nucleophile than H2S (5,11) and HS- is a stronger reductant than H2S (12). Lastly, pH dependence was studied because it affects the surface properties of the minerals, the aqueous speciation of sulfide, and probably the surface speciation of sulfide.

In these systems, sheet silicates and sulfide were hy- pothesized to influence the mechanism of CCl, transfor- mation in the following three ways (2): (1) CCl, can un- dergo electron transfer with ferrous iron in the sheet sil-

2198 Environ. Sci. Technol., Vol. 26, No. 11, 1992 0013-936X/92/0926-2198$03.00/0 0 1992 American Chemical Society