Alkaline reforming of brominated fire-retardant plastics: Fate of bromine and antimony

Chemosphere 74 (2009) 787–796

Contents lists available at ScienceDirect

Chemosphere journal homepage: www.elsevier.com/locate/chemosphere

Alkaline reforming of brominated fire-retardant plastics: Fate of bromine and antimony Jude A. Onwudili, Paul T. Williams * Energy and Resources Research Institute, The University of Leeds, Leeds LS2 9JT, UK

a r t i c l e

i n f o

Article history: Received 16 May 2008 Received in revised form 20 October 2008 Accepted 20 October 2008 Available online 2 December 2008 Keywords: Plastics Supercritical water Reforming Waste

a b s t r a c t High-impact polystyrene (HIPS) flame retarded with decabromodiphenyl ether (DDE), has been reacted in supercritical water from 380 to 450 °C and 21.5 to 31.0 MPa pressure in a batch reactor. Different concentrations of sodium hydroxide additive were used in situ to neutralize the corrosive inorganic bromine species released during the reactions. It appeared that supercritical water conditions lowered the decomposition temperature of both the fire-retardant DDE and HIPS. The reaction products included oils (up to 76 wt%), char (up to 18 wt%) and gas (up to 2.4 wt%) which was mainly methane. The presence of the alkaline water led to up to 97 wt% debromination of the product oil, producing virtually bromine-free oil feedstock. The removal of antimony from the oil product during processing was of the order of 98 wt%. The oil consisted of many single- and multiple-ringed aromatic compounds, many of which had alkyl substituents and/or aliphatic Cn-bridges (n = 1–4). The major single-ringed compounds included toluene, xylenes, ethylbenzene, propylbenzene and a-methylstyrene. Bibenzyl (diphenylethane), stilbene, diphenylmethane, diphenylpropane, diphenylcyclopropane, diphenylpropene, diphenylbutane, diphenylbutene and diphenylbuta-1,3-diene were the major Cn-bridged compounds. Diphenyl ether and acetophenone were the major oxygenated compounds found. The process thus has the potential to produce bromine-free and antimony-free oils from fire-retardant plastics. Ó 2008 Elsevier Ltd. All rights reserved.

1. Introduction There is a growing concern over the environmental consequences of the quantity of polymer products in waste electrical and electronic equipment destined for disposal. The European Union’s Directive on Waste Electrical and Electronic Equipment (WEEE) highlights the importance of the separate collection and end-of-life uses of this category of wastes, with much emphasis on recycling and reuse (European Commission, 2003). High-impact polystyrene (HIPS) and acrylonitrile–butadiene–styrene co-polymer are the main plastic components of WEEE (Tange, 1999; Brebu et al., 2006). These polymers are widely found in TV casings, computers and office equipment and often contain fire-retardants. Particularly of interest are those WEEE plastics retarded with organobromine compounds and antimony trioxide which are toxic. The commonly used organobromine compounds, otherwise known as brominated flame retardants include polybrominated biphenyls, polybrominated diphenyl oxides/ethers, tetrabromo bisphenol-A or polybrominated epoxy resins (Brebu et al., 2006). Antimony oxide synergist is usually added to WEEE plastics to enhance the effectiveness of the brominated flame retardant (Luijk et al.,

* Corresponding author. Tel.: +44 1133432504; fax: +44 1132440572. E-mail address: [email protected] (P.T. Williams). 0045-6535/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.chemosphere.2008.10.029

1991). However, antimony and many of its compounds including antimony trioxide are toxic. Pyrolysis is regarded as a potential route to recycle waste plastics (Bagri and Williams, 2002a,b; Williams and Slaney, 2007), including WEEE plastics (Bhaskar et al., 2004a,b; Hall and Williams, 2006, 2008; Hall et al., 2007), resulting in their decomposition and consequent production of fuel oils and/or chemical feedstock that may be useful for the petrochemical industry. The destruction of the organobromine compounds to recover useful bromine-free chemical feedstocks and fuels via thermochemical means has grown into an important research area (Luijk et al., 1991; Bhaskar et al., 2004b; Brebu et al., 2005; Hall and Williams, 2006; Hall et al., 2007). For example, Luijk et al. (1991) have shown that factors such as pressure, vacuum and type of reaction atmosphere can have a significant effect on the decomposition temperature ranges of the organic components of brominated HIPS (BrHIPS). They showed that the optimum degradation temperature of polystyrene fell by 85 °C from 390 to 305 °C in the presence of air compared to argon. The same authors elucidated the order of the several processes taking place during the pyrolysis of Br-HIPS. This included the debromination of the flame retardant, the bromination of polystyrene, the formation of antimony bromides and antimony oxybromides and lastly the formation of polybrominated dibenzofurans. It was reported that the decomposition of the brominated flame retardant happened independently from that of

788

J.A. Onwudili, P.T. Williams / Chemosphere 74 (2009) 787–796

polystyrene, and also that there was hardly any case of an overlap (Luijk et al., 1991). Hall and Williams (2006) studied the pyrolysis of flame retarded HIPS in a fluidised bed reactor at 440 °C. They reported that the majority of the bromine was found in the oil which constituted about 89 wt% of the reaction products. About 30% of the oil was comprised of compounds like benzene, toluene, ethylbenzene, styrene and cumene. They also reported that antimony was distributed into the oil and char. Brebu et al. (2005) reported that iron and calcium were effective in the removal of bromine and chlorine respectively, when a mixture of polyolefins, polystyrene, PVC, acrylonitrile–butadiene–styrene co-polymer and BrHIPS was pyrolysed in a glass batch reactor at atmospheric pressure. Vasile et al. (2007) demonstrated that catalytic hydrogenation of the liquid product obtained after the pyrolysis of realworld WEEE plastics led to removal of halogen atoms from the oil product. It has also been established that the synergist antimony can react with some of the bromine atoms during pyrolysis to form antimony bromide, which is often distributed between the resultant char and oil (Bhaskar et al., 2004b; Brebu et al., 2005, 2007; Hall et al., 2007). Studies on the use of zeolite catalysts during the pyrolytic degradation of plastics containing different brominated plastics and additives have also been reported with significant results (Blazso et al., 2002; Blazso and Czegeny, 2006; Hall and Williams, 2008). Hall and Williams (2008) reported that zeolite catalysts increased the amount of hydrocarbon gases in the pyrolysis products while the amount of oil was found to decrease. The catalysts also had an appreciable effect on the composition of the product oil. Blazso and Czegeny (2006) found that the molecular size of zeolite catalysts influenced the degree of debromination of tetrabromobisphenol-A to bromophenols and bisphenol-A. Brebu et al. (2006) carried out low temperature treatment of Br-HIPS in the presence of water and aqueous potassium hydroxide solution at 280 °C. They found that water removed about 90% of the bromine and degraded the plastics beyond recovery. KOH solution on the other hand, caused not only debromination, but also left the resultant plastics with a molecular size distribution identical to that of the virgin plastic material. However, the antimony trioxide synergist probably remained in the recovered plastic. The evolved gaseous halogens and hydrogen halides derived from pyrolysis have been trapped by the use of appropriate scrubber chemicals or reagents. These include exit gas alkali scrubber solutions (Hall and Williams, 2006; Hall et al., 2007) and ammonia reagent (Brebu et al., 2007). The thermal treatment of WEEE plastics may become a more acceptable technology if the process not only involves the degradation of the organobromine compounds, but also in situ and safe removal of both the constituent bromine and antimony from the product oil. Supercritical water (Tc P 374 °C, Pc P 22.1 MPa) has physico-chemical parameters comparable to most organic solvents but with added thermal stability (Shibasaki et al., 2004). Some of these properties confer good solvent ability for organic chemical reactions such as depolymerisation, hydrolysis, dehydrogenation, hydrogenation, oxidation and gasification (Savage, 1999; Kozinski and Fang, 2002; Arita et al., 2003; Williams and Onwudili, 2006). The reactions of different plastics have been studied in subcritical and supercritical water conditions. Generally, condensation polymers which have ether, ester and isocyanate linkages such as polycarbonate and polyetherketone resins react under hydrothermal conditions to selectively produce their monomers via hydrolysis (Nagase et al., 1999; Tagaya et al., 2001; Shibasaki et al., 2004). However, under similar conditions, the addition polymers such as polyethylene, polypropylene, etc. undergo pyrolysis to give rise to oils (Fang and Kozinski, 2001; Sato et al., 2006). Hence, degradation of condensation polymers can occur readily under highly acidic subcritical conditions, whereas addition polymers require high temperatures to be degraded. Kwak et al. (2005) found that the

equilibrium conversion of polystyrene was higher in supercritical water (100 wt%) than under subcritical conditions (80 wt%). They found the main degradation products to include toluene, ethylbenzene, isopropylbenzene, triphenyl benzene, styrene monomer, styrene dimer and styrene trimer, with higher selectivity towards the non-styrene compounds in relation to increasing reaction time and temperature. The decomposition of poly(vinylchloride), PVC, under hydrothermal conditions was studied by Takeshita et al. (2004). They observed that the chlorine in PVC dissolved in water to form hydrochloric acid and no harmful chlorinated compounds were observed in the liquid and gas fractions after treatment at 300 °C. The removal of the chlorine atoms left a solid material, polyene between 250 and 350 °C, however, above 350 °C in supercritical water, the polyene decomposed to produce acetone, phenol, benzene derivatives, and aliphatic alkanes and alkenes. Compared to other treatment methodologies for plastics such as pyrolysis, the presence of water during hydrothermal treatment prevents the formation of corrosive gases by dissolving the halogen-containing gases to form their corresponding acids, with a possibility for in situ neutralization. The formation of inorganic water soluble halogen species can also minimize the formation of hazardous organohalogen compounds which often characterize conventional pyrolysis products. In this paper, we report our investigation of the thermal degradation of Br-HIPS in supercritical water in the presence of sodium hydroxide additive. The objective was to produce fuel oils or useful chemical feedstock free from both bromine and antimony.

2. Materials and methods 2.1. Materials The degradation reactions were carried out in a 75 mL Hastelloy-C pressurised, autoclave batch reactor obtained from Parr Instruments CO., IL, USA. It was rated to a maximum temperature of 600 °C and pressure of 40 MPa. Temperature was measured by means of a K-type thermocouple fitted in a thermowell at the bottom-end of the reactor. The reactor was heated with a 1.5 kW ceramic furnace supplied by Carbolite, Sheffield, UK. Organic standards for both qualitative and quantitative analysis were obtained from Sigma–Aldrich, UK. Standard polybrominated diphenyl ether predominant congener mixture was obtained from Cambridge Isotope Laboratories, USA. HIPS flame retarded with decabromodiphenyl ether (DDE) was obtained from Atofina (UK) in pellet form. The elemental composition of samples was determined using a CE Instruments Flash EA 1112 Elemental Analyzer. The bromine content of the HIPS was determined using the standard Oxygen Flask Combustion Method (Association of Official Analytical Chemists, 1990). The Br-HIPS composition (wt%) was found to be carbon, 80.5, hydrogen, 7.3, bromine 7.6, oxygen 0.8 and antimony 3.6. 2.2. Experimental procedure The reactor was pre-weighed and loaded with the Br-HIPS sample and aqueous sodium hydroxide solution. For each experiment, 10 g of Br-HIPS was used and the concentration of alkali solution was varied from zero to 3.4 M. The stoichiometry of the reaction between all the available bromine (as HBr) and sodium hydroxide was considered in choosing the levels of the alkali used. In each case, the reactor was sealed, purged with nitrogen for 10 min and placed in a furnace pre-heated to 600 °C. This procedure helped to reduce the heating-up period such that it took only about 35 min for the reactor to reach 450 °C, the maximum operating temperature used in this work. Hence the average heating reactor

789

J.A. Onwudili, P.T. Williams / Chemosphere 74 (2009) 787–796

rate was approximately 12 °C min1. The temperature range investigated in this paper, from 380 to 450 °C, represented a heating time difference of up to 5.8 min between temperatures. We have previously shown (Williams and Onwudili, 2006) for a batch reactor system that reaction times of up to 30 min did not show a significant difference in product yield. Since the reactor was sealed, heating of the solution in the reactor generated pressure sufficient to pressurize the reactor to supercritical water conditions. The experiments were carried out between 380 and 450 °C, representing pressures between 21.5 and 31.0 MPa. Once the designated temperature was reached, the reactor was rapidly quenched with air-jets. On cooling to ambient conditions, the temperature and pressure of the reactor were noted before gas sampling. All experiments were repeated and the reproducibility was very good in terms of products distribution as shown in Tables 1 and 3. 2.3. Gas sampling and analysis The gaseous effluents were sampled using gas tight plastic syringes via the gas outlet valve situated on the head of the reactor. The permanent gases were analysed using two packed column GC’s, one for permanent gases and the other for hydrocarbon gases. A Varian CP-3380 GC with two packed columns and two thermal conductivity detectors was used to analyse permanent gases. Hydrogen, oxygen and carbon monoxide, methane and nitrogen were analysed on a molecular sieve column and carbon dioxide on a Hysep column. Hydrocarbon gases from C1 to C4 were analysed using a second Varian C-3380 GC/FID. The column used was an 80–100 mesh (150–180 lm) Hysep. The results from the GC were obtained in volume percent, and this was converted to moles and then mass using the general gas equation. 2.4. Liquid product analysis After gas analysis, the reactor was opened to sample the liquid phase. This was quantitatively transferred directly onto a filtration system with ethyl acetate. The mixture was passed through a dried and pre-weighed Whatman filter paper under vacuum. The char held on the filter paper was washed with ethyl acetate until all the oil had been extracted. The filter paper and char were dried to a constant weight to obtain the weight of char. The liquid extract containing both the aqueous and organic phases was separated with ethyl acetate to produce an oil fraction. The ethyl acetate was then carefully evaporated with a gentle stream of nitrogen gas to obtain the mass of product oil in each experiment. The oil was immediately analyzed on a GC/FID with a non-polar BP-5 capillary column. Identification of compounds was via the use of internal standards and the use of relative retention indices from the literature. Identification of compounds was further aided by GC/MS with auto-injection. The GC/MS which used the same temperature programme as with the GC/FID was a Hewlett Packard 5280 GC fitted with a Restek RTX-5MS column, coupled to a

HP5271 ion trap detector. The ion-mass spectra derived were automatically compared to spectral libraries using HP Chemstation software and similarity indexes of >70% were reliably used to identify compounds. Quantification was carried out by the internal and external standard methods. Analytical recoveries of more than 85% were obtained for the internal standard analytical procedure from extraction through to GC analysis. The organobromine compounds in the oil were analysed using a Varian 3380 GC fitted with an electron capture detector (GC/ECD) (Hall et al., 2007). The elemental bromine content of the oil was determined as bromide using the EPA 5050 Bomb Preparation Method for Solid Waste (EPA, 2007a) followed by Method EPA 9056A; Determination of Inorganic Anions by Ion Chromatography (EPA, 2007b). In this procedure, oil combustion in the calorimeter produces hydrogen bromide gas, which is then scrubbed into an alkali solution. The bromine content of the solution was then determined using a Dionex DX100 ion chromatograph fitted with a Dionex AS4A column. Recovery analyses were carried out using materials with known halogen contents. The antimony content of the oil and char was determined using acid digestion with concentrated sulphuric and nitric acids. The resultant solution was stabilised and analyzed on a Flame Atomic Absorption Spectrophotometer. Method validation was carried out by digesting a known concentration of antimony in polyethylene. 2.5. Char analysis The product chars from the experiments were analysed for their elemental, bromine and antimony contents using the methods described above. Some char samples were further scanned for the presence of inorganic crystal structures using X-ray Diffraction (XRD). The XRD instrument used was a Philips PW1050 Goniometer with a Philips PW1730 generator and a Cu Ka radiation X-ray tube. The sample was ground to <75 mm size and loaded into the 20 mm aperture of an aluminium sample holder. The instrument was equipped with Hiltonbrooks’ HBX data acquisition software to collect the data. GBC Scientific Equipment Ltd.’s TRACES software using the International Centre for Diffraction Data Powder Diffraction Files database was used for phase identification.

3. Results and discussion 3.1. General product distribution The effects of reaction conditions and sodium hydroxide concentration on the product distribution during the alkaline reforming of Br-HIPS were studied from a temperature of 380 to 450 °C and resultant pressures of 21.5 and 31.0 MPa. The results are presented in Table 1. In general, increasing temperature and pressure seemed to only slightly increase the amount of hydrocarbon gases

Table 1 Product distribution from the supercritical water reforming of Br-HIPS in relation to reaction conditions. Temperature (°C) – pressure (MPa)

NaOH (M)

pH

Char (g)

Total gas (g)

Sb in H2O (g)

Br in H2O (g)

Oil (g)

380 400 425 400 425 450 450 450 450 450

0.63 0.63 0.63 2.04 2.04 Nil 0.63 1.23 2.04 3.40

4.56 5.13 5.83 7.62 8.00 1.09 6.05 7.26 8.28 8.85

1.48 ± 0.06 1.49 ± 0.06 1.52 ± 0.01 1.69 ± 0.04 1.73 ± 0.01 2.41 ± 0.01 1.82 ± 0.03 1.85 ± 0.11 1.88 ± 0.01 1.95 ± 0.02

0.15 ± 0.01 0.23 ± 0.01 0.26 ± 0.02 0.19 ± 0.02 0.23 ± 0.01 0.39 ± 0.01 0.30 ± 0.01 0.27 ± 0.01 0.26 ± 0.02 0.24 ± 0.01

0.0011 ± 0.0001 0.0008 ± 0.00006 0.0010 ± 0.0001 0.0018 ± 0.0001 0.0069 ± 0.0002 0.0003 ± 0.00001 0.0017 ± 0.00003 0.0024 ± 0.00006 0.0093 ± 0.0005 0.0230 ± 0.0003

0.62 ± 0.03 0.64 ± 0.01 0.68 ± 0.03 0.65 ± 0.03 0.71 ± 0.02 0.64 ± 0.02 0.66 ± 0.03 0.71 ± 0.03 0.72 ± 0.04 0.74 ± 0.03

7.24 ± 0.12 7.57 ± 0.31 7.60 ± 0.25 7.07 ± 0.15 7.22 ± 0.37 6.62 ± 0.35 7.38 ± 0.16 7.14 ± 0.42 6.98 ± 0.15 6.85 ± 0.33

– – – – – – – – – –

21.5 24.0 27.5 24.0 27.5 31.0 31.0 31.0 31.0 31.0

790

J.A. Onwudili, P.T. Williams / Chemosphere 74 (2009) 787–796

Table 2 Gas composition from the supercritical water reforming of Br-HIPS (mg g1 of Br-HIPS) in relation to reaction conditions. Temperature (°C) – pressure (MPa)

NaOH conc. (M)

CH4

C2H4

380 – 21.5 400 – 24.0 425 – 27.5

0.63 0.63 0.63

6.3 11.8 14.0

0.35 0.28 0.22

400 – 24.0 425 – 27.5

2.04 2.04

8.50 12.5

450 450 450 450 450

Nil 0.63 1.23 2.04 3.40

16.7 16.0 14.9 14.4 13.2

a

– – – – –

31.0 31.0 31.0 31.0 31.0

C2H6

C3H6

C3H8

4.10 6.20 6.80

0.01 0.01 0.02

2.50 3.30 3.30

0.42 0.41

4.40 6.20

0.03 0.03

0.16 0.21 0.25 0.39 0.43

10.1 8.30 4.10 7.70 4.70

0.02 0.02 0.03 0.03 0.03

Butenesa

C4H10

H2

0.01 0.01 0.01

1.50 1.20 0.90

0.30 0.40 0.40

3.30 3.31

0.02 0.01

1.50 1.20

0.60 0.60

9.60 4.70 1.20 2.00 2.00

nd 0.01 0.01 0.01 0.01

1.10 0.70 0.30 0.80 0.70

0.80 0.40 0.60 0.80 0.80

Butene and butadiene.

produced as well as the proportion of antimony and bromine in the water. Recent work (Hall and Williams, 2008) has suggested that hydrocarbon gas production from Br-HIPS is usually low, except in the presence of zeolite catalysts. Moreover, our earlier work (Onwudili and Williams, 2007) with sodium hydroxide showed that its presence does not favour hydrocarbon gas production under supercritical conditions. The compositions of the effluent gases are presented in Table 2. The major hydrocarbon gases formed were in the following order; methane > ethane > propane > butane. The concentration of alkane gases was higher than the alkene gases by more than two orders of magnitude. In general, the alkane gases increased in concentration with increasing temperature and pressure while the alkene gases decreased. However, while increasing sodium hydroxide concentration appeared to decrease the proportion of alkane gases, it seemed to increase the alkenes, albeit, slightly. With regards to other products, the amount of oil product increased with increasing temperature from 380 °C up to 425 °C but fell slightly at a temperature of 450 °C. This was not unexpected, as high reaction temperatures are known to reduce oil product in favour of gas and char production. The thermogravimetric degradation of Br-HIPS under nitrogen atmosphere was studied to determine the degradation of the BrHIPS and showed that DDE started decomposing at about 335 °C, while the polystyrene plastic itself decomposed around 414 °C. However, as noted earlier, the decomposition temperature ranges of DDE and polystyrene could depend on factors such as heating rate, pressure and the reaction atmosphere. The fast heating rate and high pressure of supercritical water used in this work may explain the total degradation of Br-HIPS to products in our experiments even at a temperature of 380 °C and pressure of 21.5 MPa. Low temperature (300–450 °C) hydrothermal degradation of synthetic polymers such as styrene–butadiene rubber (Park et al., 2001), nylon-6 (Iwaya et al., 2006) and other polymers (Goto et al., 2006) have been reported. In addition, it has been reported that debromination of Br-HIPS, and hence degradation of bromi-

nated flame retardant occurred under hydrothermal conditions at temperatures as low as 280 °C (Brebu et al., 2006). The range of analytical procedures adopted in this work ensured quantitative recovery of the char formed during these reactions. Char formation increased with both reaction temperature and sodium hydroxide concentration. Our experiments without sodium hydroxide alkali yielded more char than those with the alkali. The higher proportion of antimony and bromine in the char obtained without alkali compared to others indicates that the formation of some antimony bromide must have occurred. The ability of antimony trioxide to enhance the formation of highly cross-linked carbonaceous char under thermochemical processes is also well reported (Hall et al., 2007). The lower proportion of char with increasing alkali concentration at 450 °C could be explained by the increasing partitioning of antimony into the aqueous phase, thereby depleting the antimony in the char. The results of the elemental analysis of the chars obtained are shown in Table 3. The percentage closure was around 90% but not up to 100% and the difference could be attributed to oxygen, which was not analysed for in our procedures. The results show that carbon and antimony were the predominant elements in the char. Carbon content generally increased slightly with increasing temperature and alkali concentration. Antimony content on the other hand, seemed to decrease with both increasing temperature and sodium hydroxide concentration. Bromine was found more in the char obtained at 380 °C and 21.5 MPa (0.63 M NaOH) and at 450 °C and 31.0 MPa without the alkali. This means that at constant alkali concentration, bromine removal from char was due to increasing temperature and pressure, whereas at constant temperature and pressure it was due to increasing alkali concentration. 3.2. Bromine distribution The effect of reaction temperature on the distribution of bromine atoms between the oil, aqueous phase and char during the

Table 3 Elemental analysis of chars derived from the supercritical water reforming of Br-HIPS in relation to reaction conditions. Temperature (°C) – pressure (MPa)

NaOH (M)

380 400 425 400 425 450 450 450 450 450

0.63 0.63 0.63 2.04 2.04 Nil 0.63 1.23 2.04 3.4

– – – – – – – – – –

21.5 24.0 27.5 24.0 27.5 31.0 31.0 31.0 31.0 31.0

Elemental analysis (wt%) Carbon

Hydrogen

Br

Sb

64.6 ± 1.34 64.8 ± 2.51 66.4 ± 2.12 67.4 ± 1.52 69.2 ± 1.33 71.2 ± 1.65 67.1 ± 1.32 67.6 ± 0.83 71.6 ± 1.81 73.1 ± 2.33

2.5 ± 0.22 2.4 ± 0.13 2.3 ± 0.05 2.5 ± 0.15 2.5 ± 0.07 2.9 ± 0.01 2.3 ± 0.15 2.3 ± 0.02 2.3 ± 0.13 2.5 ± 0.11

4.2 ± 0.10 3.8 ± 0.15 3.1 ± 0.08 2.9 ± 0.11 1.8 ± 0.13 6.0 ± 0.31 3.2 ± 0.22 2.6 ± 0.14 1.1 ± 0.01 <0.2

23.5 ± 1.3 23.1 ± 1.8 22.5 ± 1.1 20.5 ± 0.8 19.9 ± 1.5 14.6 ± 1.1 21.5 ± 0.3 21.0 ± 0.1 18.2 ± 1.0 16.8 ± 0.5

791

J.A. Onwudili, P.T. Williams / Chemosphere 74 (2009) 787–796

Oil

Aqueous Phase

Wt% Bromine

100.00 80.00

a

80.0

60.00

60.0

40.00

40.0

20.00

20.0

0.00

0.00 M0.63 M1.23 M2.04 M3.40 M 100.0

100.0

Wt% Antimony

b

0.0 380400425450

80.0

Char

100.0

c

80.0

60.0

60.0

40.0

40.0

20.0

20.0

0.0

d

0.0

380

400

425

450

Reaction Temperature (°C)

0.00 M

0.63 M

1.23 M

2.04 M

3.40 M

Sodium Hydroxide Concentration (M)

Fig. 1. Effect of reaction conditions on bromine and antimony distribution in products derived from the supercritical water reforming of Br-HIPS: (a) Br distribution related to temperature (with 0.63 M NaOH at 380 °C – 21.5 MPa, 400 °C – 24.0 MPa, 425 °C – 27.5 MPa and 450 °C – 31.0 MPa); (b) Br distribution related to sodium hydroxide concentration (at 450 °C and 31.0 MPa); (c) Sb distribution related to temperature (with 0.63 M NaOH at 380 °C – 21.5 MPa, 400 °C – 24.0 MPa, 425 °C – 27.5 MPa and 450 °C – 31.0 MPa); (d) Sb distribution related to sodium hydroxide concentration (at 450 °C and 31.0 MPa).

alkaline supercritical water treatment of Br-HIPS is shown in Fig. 1. Our results show that at a constant sodium hydroxide concentration of 0.63 M, the partition of bromine atoms into both water and char increased gradually but decreased in the oil product, with increasing reaction temperature. The bulk of the bromine atoms were found in the aqueous phase, followed by char and oil in very low concentrations respectively. Only at 380 °C and 21.5 MPa, did we measure 2.86 wt% of bromine in the oil; between 400 and 450 °C (24.0 and 31.0 MPa), the bromine content of the oils was less that 1 wt%. In general up to 97 wt% of the bromine content was found in both the aqueous phase and char. Increasing temperature and pressure led to appreciably small increases in the bromine content of the aqueous phase while that of the char decreased. This may suggest that increasing temperature and pressure, and hence increasing supercritical character of the water, enhances the removal of char-bromine into the aqueous phase. The presence of sodium hydroxide may also have been beneficial to this effect; such that higher temperature favoured the removal of antimony bromide from the char via the possibility of reacting sodium hydroxide to form sodium antimonate. Brebu et al. (2006) reported that with water, up to 90% debromination of Br-HIPS was achieved under hydrothermal conditions at 280 °C and 7 MPa, suggesting that under hydrothermal conditions the decomposition of the DDE occurred at a much lower temperature range. The decomposition of DDE involves the release of Br2, HBr and other brominecontaining species (Luijk et al., 1991). During conventional pyrolysis, the formation of HBr is favoured in the absence of antimony trioxide (more gas product) otherwise its presence would lead to the formation of antimony bromide (more oil product) instead (Hall et al., 2007). In this work, no appreciable brominated gas formation was observed, as is found with the degradation of HIPS as reported in the literature. Fig. 1 also shows the effect of sodium hydroxide concentration on the distribution of bromine atoms during reactions at 450 °C and 31.0 MPa. In the absence of sodium hydroxide, the aqueous

phase was greenish with a pungent smell and a strongly acidic pH of 1.09. In this experiment more than 93 wt% debromination was achieved with only supercritical water, with about 74 wt% in the aqueous phase and close to 19 wt% in the char. This agrees with the work of Brebu et al. (2006) cited earlier. The only difference being that their work, at low temperature, did not give rise to char formation. The supercritical water more or less affected the partition of bromine atoms into water as hydrobromic acid. The formation of the acids most probably resulted from: (a) the dissolution of hydrogen bromide gas in water (Eq. (1)); (b) the reaction of bromine gas with water could only result in species like hydrobromic acid and hypobromous acid (Eq. (2)); and (c) the decomposition of antimony bromide in the presence of water (Eq. (3))

HBrðgÞ ðþwaterÞ ! HBrðlÞ

ð1Þ

Br2ðgÞ þ H2 O ! HBr þ HOBr

ð2Þ

2SbBr3 þ 3H2 O ! Sb2 O3 þ 6HBr

ð3Þ

Eq. (3) obviously affected the fate of antimony which is discussed in Section 3.3. These acids can be neutralized by sodium hydroxide to form the corresponding sodium salts. The colour of the cloudy liquid effluents was typically whitish to greyish. Thus, sodium hydroxide neutralized the acids in situ which also prevented reactor corrosion. In addition, the alkali served as a bromine removal agent, preventing the formation of antimony bromide. This may explain why the weight percents of oil products in this work were lower that those reported under pyrolysis, where antimony bromide is a significant constituent. The pH of the resulting aqueous effluents ranged from slightly acidic to neutral and finally to basic with increasing alkali concentrations. At the highest sodium hydroxide concentration of 3.4 M, 97 wt% bromine was found in the aqueous phase, with only 0.05 and 0.13 wt% in the oil and char, respectively. This indicates that highly alkaline supercritical water can totally recover the bromine atoms released from Br-HIPS at 450 °C, thus producing bromine-free oils.

792

J.A. Onwudili, P.T. Williams / Chemosphere 74 (2009) 787–796

3.3. Antimony distribution The fate of antimony trioxide synergist during the alkaline treatment of Br-HIPS was also studied simultaneously with that of bromine above. Considering that different species of antimony compounds could be formed during the reactions, it became imperative to monitor the partition of antimony metal into the different reaction products. Fig. 1 shows the distribution of antimony with respect to reaction temperature at a constant alkali concentration of 0.63 M. In most pyrolysis work regarding brominated plastics, antimony is usually found in the oil and char (Hall and Williams, 2006, 2008). The results of this work showed that antimony was located mainly in the char. In conventional pyrolysis, the antimony trioxide usually reacts with HBr from the flame retardant to form antimony oxybromide in the condensed state thus facilitating the breaking of C–Br bonds (Luijk et al., 1991). Antimony bromide (SbBr3) which is the only known chemical combination of antimony and bromine atoms, is then formed when the oxybromide is subjected to continuous heating or high temperatures in the gaseous state. In this work, water and alkali appeared to have stronger affinities towards bromine species and as such, suppressed the synergistic reactions of antimony trioxide. In particular, water limited the availability of gaseous hydrogen bromide. Antimony bromide, even if formed readily decomposes in water to form hydrobromic acid and antimony trioxide (Eq. (3)). Antimony oxybromide is also unstable at about 300 °C, decomposing by releasing antimony bromide and forming antimony trioxide. In essence, only a minute percent of the bromine atoms could be held by antimony in the presence of water, indicating that much of the antimony in the char will be in the form of antimony trioxide. XRD analysis of the char derived from the reaction of the Br-HIPS under supercritical water conditions showed the crystalline components of the char samples. These included two crystalline forms of antimony trioxide, and antimony oxybromide, sodium bromide, sodium bromide hydrate, nickel antimonide and metallic antimony. It is thus possible to recover antimony trioxide synergist from the char in its original form. Additionally, inorganic compounds are highly insoluble in supercritical water (Armellini and Tester, 1994). All these could explain the predominant location of antimony in the char. Moreover, antimony compounds/salts are sparsely soluble in ambient water, so that antimony would preferentially partition into the char or oil when the reaction reached ambient conditions. The level of antimony remained very low in the oil products, hence most of the metal was found in the char, with levels of over 95 wt%. In the presence of higher concentrations of alkali (Fig. 1), there was an increasing presence of antimony in the aqueous phase. For instance while there was little or no antimony in the aqueous phase obtained in the absence of the alkali, at 3.4 M alkali concentration, the antimony level was up to 6 wt% in the aqueous phase. The predominant levels of antimony still remained in the char with very low levels in the oil. The reason for the increasing partitioning of antimony into the aqueous phase may be the possible ionic reaction of any antimony bromide with sodium hydroxide to precipitate antimony hydroxide (Sb(OH)3). This reaction is very well known in material science (Pentia et al., 2001). The antimony hydroxide may sorb to the char, however in the presence of excess sodium hydroxide, antimony hydroxide will react to form sodium antimonate. Thus excess hydroxide ions may facilitate the formation of antimonate ion SbðOHÞ 6 which would become soluble in the aqueous phase and is possibly stabilised in solution by sodium ions. From both the limited available data of the speciation of antimony in water and thermodynamic predictions, the most favoured species of antimony in water is the antimonate ion, also known as the pentavalent oxoanion, SbðOHÞ 6 (Mohammad et al., 1990; Cotton and Wilkinson, 1999). In addition, an earlier study on the prep-

aration of sodium antimonate (Stewart and Knop, 1970) reported that in alkaline solutions, the ionic form of antimony in water is SbðOHÞ 6 . To this extent, we can infer that under alkaline conditions, water soluble antimony existed as NaSb(OH)6, this may explain the enhanced level of antimony in the more alkaline effluents. However, under acidic conditions the level of antimony in the aqueous effluent was minimal due to the poor solubility of the likely antimony species such as antimony trioxide, antimony bromide, etc. The equations below attempt to describe the possible reactions involving antimony species under the alkaline supercritical water conditions employed in this work

SbBr3 þ 3NaOH ! SbðOHÞ3 þ 3NaBr

ð4Þ

In alkaline solutions, antimony is oxidized from its usual Sb(III) oxidation state to a Sb(V) oxidation state (Stewart and Knop, 1970), thus

SbðOHÞ3 þ 3OH ! SbðOHÞ6 þ 2e

ð5Þ

The antimonate ion thus becomes soluble in aqueous environments and is stabilised by sodium ions,

SbðOHÞ6 þ Naþ ! NaSbðOHÞ6

ð6Þ

The product in Eq. (6) can readily lose water to form the commercial sodium antimonate, NaSbO3 thus,

NaSbðOHÞ6 ! NaSbO3 þ 3H2 O

ð7Þ

During the acid digestion of samples for antimony determination, chars obtained under highly alkaline conditions contained white precipitates. When the resultant suspension was decanted and the clear portion analysed, it gave antimony levels about three times lower than when the precipitate was treated with concentrated hydrochloric acid. This indicated that antimony was significantly present in the precipitate. However, we could not determine to forms of antimony in the precipitate. XRD analysis has already confirmed the presence of antimony trioxide, antimony oxybromide and elemental antimony in the char and available data from the literature suggests that in the presence of concentrated sulphuric and nitric acids (digestion reagents), antimony sulphate (Sb2(SO4)3) and antimony nitrate (Sb(NO3)3) would form. The latter is easily and completely hydrolysed in water to antimony pentoxide (Moody, 1991). Thus, these are the likely compounds in the precipitate, which dissolved on treatment with concentrated hydrochloric acid. Acid digestion is a widely known technique for extracting metals from ores, and may thus be used to recover antimony compounds from the char. 3.4. Analysis of oil products Oil was the major product obtained in all the experimental conditions studied. More than 70 wt% conversion of the Br-HIPS to oil was achieved in most experiments. The oils were mostly dark brown in colour, but the colour improved to lighter shades with increasing reaction temperature. A range of analytical techniques were used to characterize the organic composition of the oils. These included GC/MS, GC/FID, GC/ECD and FT-IR. The GC/MS was used to identify the compounds while GC/FID was used to determine their concentrations. For this paper, we present details of the composition of some of the oils, noting that temperature could be the major factor affecting the composition of the oils. In the work of Brebu et al. (2006) alkaline subcritical water was able to produce a debrominated material with molecular weight in the same range as polystyrene. The results of the GC analysis of the oils are presented in Table 4. A typical GC/MS chromatogram of the product oils is shown in Fig. 2. The chromatograms showed over

793

J.A. Onwudili, P.T. Williams / Chemosphere 74 (2009) 787–796

two hundred peaks but we have simplified this in Fig. 2 by focussing on the 50 largest peaks which make up more than 85% of each of the oils. The oils comprised a rich cocktail of aromatic compounds; no aliphatic compounds were identified. Some of the major compounds common to all the oils included toluene, ethylbenzene, styrene, cumene, a-methylstyrene, bibenzyl,

diphenylpropane, diphenyl ether, diphenylmethane, naphthalene and substituted naphthalenes, anthracene and triphenylbenzenes. Generally the concentrations of the lighter aromatic compounds increased with increasing temperature possibly due to thermal cracking of the heavier compounds, which constituted the bulk of the oils at lower temperature. At 380 °C and 21.5 MPa, diphenyl-

Table 4 GC/MS & GC/FID analyses of oils derived from the supercritical water gasification of Br-HIPS (mg g1 of Br-HIPS) in relation to reaction conditions. Name of compound

Toluene Ethylbenzene p-Xylene m-Xylene o-Xylene Styrene Cumene 3-Ethyltoluene 4-Ethyltoluene Propylbenzene Trimethylbenzene a-Methylstyrene Propyl toluene p-Methylstyrene 1,3-Diethylbenzene 1,4-Diethylbenzene 1,2-Diethylbenzene Acetophenone 1-Methyl-Indane Naphthalene 2-Methylnaphthalene 1-Methylnaphthalene Diphenyl ether Diphenylmethane Diphenyl ethene (Stilbene) Diphenyl ethane (Bibenzyl) Diphenylmethylethene Diphenylmethylethane Diphenyl-m-ethylmethane Diphenyl-p-ethylmethane Diphenyltetramethylethane Diphenyl-o-ethylmethane Diphenylcyclopropane Diphenylpropane Diphenylmethylpropene Phenylmethyl-2,3-dihydro-1H-Indene Diphenyl-2,3-dihydro-1H-Indene Diphenylmethylpropane 1,2-Diphenylmethylcyclopropene Cyclopropylidene(phenyl)methylbenzene Diphenylbutane Phenyltetrahydronaphthylmethane 9-Methylanthracene 9,10-Dimethylanthracene Anthracene Diphenylmethylbutene Diphenylmethylcyclopropene Diphenyldimethylbutene 1,5-Diphenyl-1,3-butadiene 4-Phenyl-1,2-Dihydronaphthalene 1-Phenylnaphthalene Diphenyl-2-Methyl-1,3-Pentadiene 1,5-Diphenyldimethyl-1,3-Pentadiene 3-Methyl-1-phenyl-1H-Indene Tetrahydro-1-phenyl-1,2,3-methanonaphthalene 2-Phenylnaphthalene 9-Phenyl-5H-benzocycloheptene Phenylnaphthylmethane 11-H-Benzo(b)fluorene Diphenyldimethylpropylbutane 1,5-Diphenyl-3-phenylpropylpentane 1,3,5-Triphenylbenzene a b

0.63 M NaOH. No NaOH additive.

Temperature (°C) – pressure (MPa) 380 – 21.5a

425 – 27.5a

450 – 31.0a

450 – 31.0b

11.2 55.9 0.45 0.46 0.45 69.4 57.6 0.74 0.25 5.93 0.08 6.13 2.95 Nd 0.39 0.52 1.22 2.28 1.47 0.99 Nd 1.16 2.34 2.29 2.58 14.8 5.46 1.11 2.65 Nd 8.75 Nd Nd 169 1.46 127 Nd 44.8 29.1 29.9 9.90 3.38 6.35 29.2 Nd 11.0 5.91 1.44 0.69 1.66 Nd 72.1 1.09 16.6 2.05 7.37 19.8 7.33 Nd 1.40 Nd 2.09

13.6 84.2 0.60 0.62 0.61 15.5 77.0 2.37 7.41 19.3 2.47 17.7 7.02 5.60 1.57 2.01 2.92 3.75 5.23 2.79 2.05 2.92 4.30 7.20 7.01 26.7 9.94 3.49 6.31 3.11 5.77 2.39 120 9.15 92.5 43.7 Nd 14.6 48.9 Nd 13.7 Nd Nd 17.0 Nd 11.0 12.3 Nd 6.35 5.48 60.8 41.6 Nd 40.4 3.06 21.7 32.8 19.3 1.93 Nd 2.08 3.21

34.0 116 0.83 0.81 0.82 14.8 80.7 3.78 1.24 30.5 0.35 20.6 7.64 6.00 1.51 1.71 2.99 4.22 5.36 4.81 3.05 4.38 4.45 9.84 7.74 27.3 6.38 4.52 5.71 4.52 2.85 3.34 91.5 6.84 Nd 62.2 33.0 7.89 33.1 Nd 12.8 Nd 14.6 13.5 5.46 7.80 13.9 Nd 4.79 6.62 66.2 43.0 Nd 44.3 3.27 36.1 31.5 2.57 4.23 Nd Nd 5.32

26.0 347 2.13 2.84 2.61 168 92.7 11.5 3.76 92.7 1.07 10.9 5.18 0.94 2.88 3.20 5.60 Nd 12.0 5.92 2.52 2.48 2.10 4.54 4.40 10.1 1.74 Nd 2.78 1.30 Nd Nd 24.0 1.93 0.50 8.51 17.3 Nd Nd Nd Nd Nd Nd Nd 2.24 Nd Nd Nd Nd 3.92 34.1 Nd Nd 5.24 1.47 17.7 3.54 5.42 2.13 Nd Nd 3.66

794

J.A. Onwudili, P.T. Williams / Chemosphere 74 (2009) 787–796

Intensity (10 6)

24.41 5.51

6.46

23.82 24.75

75

25.04 28.13

50

4.22 30.54

7.30 3.20 25 7.87

32.52

20.31 22.52

9.04 9.53 17.34 18.10 9.83 0

0

5

10

15

20

25

30

38.16

37.51 40.04

45.80

35

45

40

50.21 53.82 56.65 50

55

60

78.85 63.86 68.12 71.59 65 70 75 80

Time (min) Fig. 2. Typical GC/MS chromatogram of oil derived from the supercritical water reforming of Br-HIPS (with 0.63 M NaOH).

propane, phenylmethyl-2,3-dihydro-1H-Indene and 1,5-diphenyl2-methyl-1,3-butadiene were the dominant compounds in the oil. Of importance was the range of diphenyl compounds in the oil. They occupy the middle portion of the chromatographs but are characterized by a sequential elongation of the carbon chain between the benzene rings in the diphenyl compounds observed. Compounds such as diphenylmethane, diphenylethane, diphenylpropane, diphenylbutane and up to diphenylpentane were reliably identified along with some of their unsaturated, cyclic and alkylated homologues. HIPS consists of a polystyrene phase and a dispersed polybutadiene phase (Jakab et al., 2003; Hall et al., 2007). Hence aliphatic radicals existed in the reaction system as decomposition products from the polybutadiene phase. The formation of the diphenyl compounds may be related to them. There were probably too many aromatic radicals to allow chain termination reactions between two aliphatic radicals or between aliphatic and hydrogen radicals, thus resulting in insignificant levels of aliphatic compounds in the oil. For example combination of a phenyl radical and a phenylethyl radical or styrene radical would yield diphenylethane. More importantly, the elongation of the carbon chain between the benzene rings was possibly a precursor to the formation of condensed polyaromatic hydrocarbons such as anthracene, 11-H-Benzo(b)fluorene, diphenyl-2,3-dihydro-1H-Indene, Diphenyl-1,3-butadiene, Phenyl-1,2-dihydronaphthalene, phenylnaphthalene, etc. It is well known that the possibility of cyclization increases with increasing carbon chain; hence the absence of bridging aliphatic chains with more than five carbon atoms. The compositional analyses of the oils produced at a temperature of 450 °C and pressure of 31.0 MPa, with and without sodium hydroxide illustrate the effect of the alkali on the distribution of compounds in the oil between simple light and complex multiringed aromatic compounds. For instance, in the presence of 0.63 M NaOH, ethylbenzene constituted about 12% of the oil, whereas in the absence of the alkali, it was about 36% of the oil. In addition, with alkali, increasing temperature and pressure seemed to slightly increase the percentage of styrene in the oil such that it was 0.81% at 380 °C (21.5 MPa), 1.6% at 425 °C (27.5 MPa) and 1.5% at 450 °C (31.0 MPa), but without the alkali, it increased to about 17.5% at 450 °C (31.0 MPa). The same trend was observed for cumene and propylbenzene. On the contrary it was clear that temperature resulted in a decreasing percentage of heavier compounds such as diphenylpropane, diphenylcyclopro-

pane, phenylmethyl-2,3-dihydro-1H-Indene, etc. in the oils. For instance, diphenylpropane was over 19% of the oil at 380 °C and 21.5 MPa and less than 1% at higher temperatures. In addition, it was the case that at higher temperatures, diphenylcyclopropane and diphenylmethylpropene were more prominent than diphenylpropane. However, their concentrations decreased with increasing temperature. Comparison between the two oils produced at a temperature of 450 °C and 31.0 MPa, with and without alkali, showed that oil from the experiment with the alkali contained a higher percentage of heavier compounds than the oil obtained without sodium hydroxide. This may suggest that the alkali influenced the formation of oligomers from the single-ringed compounds. The oils obtained at 380 °C (21.5 MPa), 425 °C (27.5 MPa) and 450 °C (31.0 MPa) with 0.63 M NaOH were analysed to identify brominated organic compounds using GC/ECD. The chromatograms and compound identification showed that the GC results corroborated other results with respect to the organobromine content of the oils. It can be seen from the chromatograms in Fig. 3 that the numbers of and responses to organobromine compounds declined with increasing temperature. Considering that the same (50,000 lg g1 of oil dissolved in ethyl acetate) high concentrations of oil samples were injected into the GC, the detector signals are an indication of the low concentration of organobromine compounds in the oils. Each chromatogram is divided into three segments with respect to retention times. Segment one ranges from 0.7 to 30 min; this segment contains aliphatic-bromine, mono-bromine and di-bromine compounds such as bromobutane, bromobenzene, (1-bromoethyl)benzene, (2-bromoethyl)benzene and 1-bromo-2ethylbenzene. Apart from 2-bromoethylbenzene, the concentrations of the other compounds remained low, although they were generally higher at 380 °C than at 425 °C and 450 °C, respectively. The same trend of decreasing concentration with increasing temperature was observed for the higher organobromine compounds in the second segment (30–50 min). These compounds included bromonaphthalene, dibromodiphenylethane, and bromodiphenyl ethers particularly in the range of mono-bromodiphenylether to pentabromodiphenyl ether. The cluster of peaks indicating these compounds was much denser at 380 °C than at higher temperatures. Generally, these peaks appeared roughly around the retention time range of the diphenyl hydrocarbon compounds identified on the GC/MS-FID. The poor separation of these compound-clusters as well as the relatively lower concentrations of the organobromine compounds may well be the reason the latter

795

J.A. Onwudili, P.T. Williams / Chemosphere 74 (2009) 787–796

400

a

300 200 100 0 100

b

75 50 25 0

75

c

50 25 0

10

20

30

40

50

60

70

Time (min) Fig. 3. GC/ECD chromatograms of oils derived from the supercritical water reforming of Br-HIPS (a) 380 °C – 21.5 MPa, (b) 425 °C – 27.5 MPa (c) 450 °C – 31.0 MPa, in the presence of 0.63 M NaOH.

were not detected on the GC/MS or GC/FID. It is equally interesting to note that more late-eluting compounds (51–76 min) such as higher polybrominated diphenylethers and others, e.g. bromotriphenylmethane (56.3 min) were detected at 425 and 450 °C. Luijk et al. (1991) had proposed that higher temperatures could lead to secondary reaction products such as polybrominated dioxins and furans; however these are not likely to form at 450 °C (31.0 MPa). The oils obtained by using higher concentrations of sodium hydroxide did not contain any quantifiable organobromine compounds. This corroborates the results presented in Table 1, indicating that more bromine atoms were removed as sodium bromide with increasing alkali concentration. 4. Conclusions The degradation of Br-HIPS, a representative polymer for WEEE plastics has been carried out under alkaline supercritical water conditions. Supercritical water lowered the decomposition temperature range of both DDE and the polystyrene components of the plastic. Hence, at a temperature as low as 380 °C and 21.5 MPa, HIPS was completely degraded into oil and char. Water also played a prominent role in the removal of bromine species in the reaction medium forming hydrobromic acid. The presence of sodium hydroxide provided a system to neutralize the formed acids in situ and thus prevented reactor corrosion. Sodium hydroxide also influenced the fate of antimony trioxide synergist ensuring the metal was predominantly located in the char. Both water and sodium hydroxide prevented the usual predominance of antimony bromide and ensured that majority of the antimony in the char were in the form of antimony trioxide, antimony hydroxide and possibly sodium antimonate. Experiments without sodium hydroxide suggested that the alkali also played a role in the

composition of the oils. Results showed that the preference for lower molecular weight compounds like ethylbenzene, cumene, styrene and propylbenzene was higher in the absence of sodium hydroxide. This indicated that the alkali might have enhanced the oligomerization of these compounds to form the kind of complex compounds such as the variety of diphenyl compounds, triphenylbenzenes and polyaromatic hydrocarbons found in the oils. The intervention of the alkali ensured the non-availability of bromine species for organic reactions. The process thus demonstrates the potential to produce bromine-free and antimony-free oils from WEEE plastics. Acknowledgements This work was supported by the UK Engineering and Physical Sciences Research Council under EPSRC Grant EP/D053110/1. References Arita, T., Nagahara, K., Nagami, K., Kajimoto, O., 2003. Hydrogen generation from ethanol in supercritical water without catalyst. Tetrahedron Lett. 44, 1083– 1086. Armellini, F.J., Tester, J.W., 1994. Precipitation of sodium chloride and sodium nitrate in water from sub- to supercritical conditions: 150–550 °C, 100–300 bar. J. Supercrit. Fluid 7, 147–158. Association of Official Analytical Chemists, 1990. AOAC Method 974.36. Microchemical Determination of Bromine, Chlorine, or Iodine by Oxygen Flask Combustion Method, in Association of Official Analytical Chemists. Official Methods of Analysis. Association of Analytical Chemists, Washington DC. Bagri, R., Williams, P.T., 2002a. Fluidized bed catalytic pyrolysis of polystyrene. J. Inst. Energy 75, 117–123. Bagri, R., Williams, P.T., 2002b. Catalytic pyrolysis of polyethylene. J. Anal. Appl. Pyrol. 63, 29–41. Bhaskar, T., Kaneko, J., Muto, A., Sakata, Y., Uddin, M.A., Jakab, E., 2004a. Effect of poly(ethylene terephthalate) on the pyrolysis of brominated flame retardant

796

J.A. Onwudili, P.T. Williams / Chemosphere 74 (2009) 787–796

containing high impact polystyrene and catalytic debromination of liquid products. J. Anal. Appl. Pyrol. 71, 765–777. Bhaskar, T., Kaneko, J., Muto, A., Sakata, Y., Jakab, E., Matsui, T., Uddin, M.A., 2004b. Pyrolysis studies of PP/PE/PS/PVC/HIPS-Br mixed with PET plastics and dehalogenation (Br, Cl) of liquid products. J. Anal. Appl. Pyrol. 72, 27–33. Blazso, M., Czegeny, Z., 2006. Catalytic destruction of brominated aromatic compounds studied in a catalyst microbed coupled to gas chromatography/ mass spectrometry. J. Chromatogr. A 1130, 91–96. Blazso, M., Czegeny, Z., Csoma, C., 2002. Pyrolysis and debromination of flame retarded polymers of electronic scrap studied by analytical pyrolysis. J. Anal. Appl. Pyrol. 64, 249–261. Brebu, M., Bhaskar, T., Murai, K., Muto, A., Sakata, Y., Uddin, M.A., 2005. Removal of nitrogen, bromine, and chlorine from PP/PE/PS/PVC/ABS–Br pyrolysis liquid products using Fe- and Ca-based catalysts. Polym. Degrad. Stabil. 87, 225–230. Brebu, M., Bhaskar, T., Muto, A., Sakata, Y., 2006. Alkaline hydrothermal treatment of brominated high impact polystyrene (HIPS-Br) for bromine and bromine-free plastic recovery. Chemosphere 64, 1021–1025. Brebu, M., Jakab, E., Sakata, Y., 2007. Effect of flame retardants and Sb2 O3 synergist on the thermal decomposition of high-impact polystyrene and on its debromination by ammonia treatment. J. Anal. Appl. Pyrol. 79, 346–352. Cotton, F.A., Wilkinson, G., 1999. Advanced Inorganic Chemistry, sixth ed. John Wiley & Sons, Berlin. EPA, 2007a. Method 5050. Bomb Preparation Method for Solid Waste. Environmental Protection Agency, Washington, DC. EPA, 2007b. Method 9056A. Determination of Inorganic Anions by Ion Chromatography, Environmental Protection Agency, Washington, DC. European Commission, 2003. Directive 2002/96/EC of the European Parliament and of the Council on Waste Electrical and Electronic Equipment. Official Journal of the European Commission. Official Journal of the European Commission, L37/ 24, Brussels. Fang, Z., Kozinski, J.A., 2001. A comparative study of polystyrene decomposition in supercritical water and air environments using diamond anvil cell. J. Appl. Polym. Sci. 81, 3565–3577. Goto, M., Sasaki, M., Hirose, T., 2006. Reactions of polymers in supercritical fluids for chemical recycling of plastic wastes. J. Mater. Sci. 41, 1509–1515. Hall, W.J., Williams, P.T., 2006. Pyrolysis of brominated feedstock plastic in a fluidized bed reactor. J. Anal. Appl. Pyrol. 77, 75–82. Hall, W.J., Williams, P.T., 2008. Removal of organobromine compounds from the pyrolysis oils of flame retarded plastics using zeolite catalysts. J. Anal. Appl. Pyrol. 81, 139–147. Hall, W.J., Mitan, N.M.M., Bhaskar, T., Muto, A., Sakata, Y., Williams, P.T., 2007. The co-pyrolysis of flame retarded high impact polystyrene and polyolefins. J. Anal. Appl. Pyrol. 80, 406–415. Iwaya, T., Sasaki, M., Goto, M., 2006. Kinetic analysis for hydrothermal depolymerization of nylon 6. Polym. Degrad. Stabil. 91, 1989–1995. Jakab, E., Uddin, M.A., Bhaskar, T., Sakata, Y., 2003. Thermal decomposition of flameretarded high-impact polystyrene. J. Anal. Appl. Pyrol. 68, 83–99. Kozinski, A.J., Fang, Z., 2002. A study of rubber liquefaction in supercritical water using DAC-stereomicroscopy and FT-IR. Fuel 81, 935–945. Kwak, H., Shin, H.-Y., Bae, S.-Y., Kumazawa, H., 2005. Characteristics and kinetics of degradation of polystyrene in supercritical water. J. Appl. Polym. Sci. 101, 695– 700.

Luijk, R., Govers, H.A.J., Eijkel, G.B., Boon, J.J., 1991. Thermal degradation characteristics of high impact polystyrene/decabromodiphenylether/antimony oxide studied by derivative thermogravimetry and temperature resolved pyrolysis-mass spectrometry: formation of polybrominated dibenzofurans, antimony (oxy)bromides and brominated styrene oligomers. J. Anal. Appl. Pyrol. 20, 303–319. Mohammad, B., Ure, A.M., Reglinski, J., Littlejohn, D., 1990. Speciation of antimony in natural waters: the determination of antimony (III) and antimony (V) by continuous flow hydride generation–atomic absorption spectrometry. Chem. Spec. Bioavailab. 2, 117–122. Moody, B., 1991. Comparative Inorganic Chemistry, third ed., vol. 9. Edward Arnold, a division of Hodder and Stoughton, London. Nagase, Y., Yamagata, M., Fukuzato, R., 1999. Development of a chemical recycling process for waste plastics using supercritical water. Kolbeco Technol. Rev. 22, 11–14. Onwudili, J.A., Williams, P.T., 2007. Hydrothermal catalytic gasification of municipal solid waste. Energy Fuel 26, 3676–3683. Park, Y., Hool, J.N., Curtis, C.W., Roberts, C.B., 2001. Depolymerization of styrene– butadiene copolymer in near-supercritical water. Ind. Eng. Chem. Res. 40, 756– 767. Pentia, E., Pintilie, L., Trivarus, C., Pintilie, I., Botila, T., 2001. Influence of Sb3+ ions on photoconductive properties of chemically deposited PbS films. Mater. Sci. Eng. B 80, 23–26. Sato, O., Konno, M., Wakul, Y., Matsunaga, H., Shirai, M., 2006. Supercritical water decomposition of polyethylene samples for the determination of their trace cadmium content. Anal. Sci. 22, 1461. Savage, P.E., 1999. Organic chemical reactions in supercritical water. Chem. Rev. 99, 603–621. Shibasaki, Y., Kamimori, T., Kadokawa, J., Hatano, B., Tagaya, H., 2004. Decomposition reactions of plastic model compounds in sub- and supercritical water. Polym. Degrad. Stabil. 83, 481–485. Stewart, D.J., Knop, O., 1970. Pyrochlore. VI. Preparative chemistry of sodium and silver antimonates and related compounds. Can. J. Chem. 48, 1323– 1332. Tagaya, H., Suzuki, Y., Komuro, N., Kadokawa, J., 2001. Reactions of model compounds of phenol resin in sub- and supercritical water under an argon atmosphere. J. Mater. Cycles Waste Manag. 3, 32–37. Takeshita, Y., Kato, K., Takahashi, K., Sato, Y., Nishi, S., 2004. Basic study on treatment of waste polyvinyl chloride plastics by hydrothermal decomposition in subcritical and supercritical water regions. J. Supercrit. Fluid 31, 185– 193. Tange, L., 1999. BSEF Technical Workshop Report on Sustainable Management of Plastics with Bromine in Tokyo, Tokyo. Vasile, C., Brebu, M.A., Karayildirim, T., Yanik, J., Darie, H., 2007. Feedstock recycling from plastics and thermosets fractions of used computers II. Pyrolysis oil upgrading. Fuel 86, 477–485. Williams, P.T., Onwudili, J.A., 2006. Composition of products from the supercritical water gasification of glucose: A biomass model compound. Ind. Eng. Chem. Res. 44, 8739–8749. Williams, P.T., Slaney, E., 2007. Analysis of products from the pyrolysis and liquefaction of single plastics and waste plastic mixtures. Conserv. Recycling 51, 754–769.