- Email: [email protected]
F formation under stable and transient combustion conditions during MSW incineration
Chemosphere 76 (2009) 767–773
Contents lists available at ScienceDirect
Chemosphere journal homepage: www.elsevier.com/locate/chemosphere
Effects of sulfur on PCDD/F formation under stable and transient combustion conditions during MSW incineration Johanna Aurell *, Jerker Fick, Peter Haglund, Stellan Marklund Department of Chemistry, Umeå University, SE 901 87 Umeå, Sweden
a r t i c l e
i n f o
Article history: Received 15 December 2008 Received in revised form 25 April 2009 Accepted 28 April 2009 Available online 30 May 2009 Keywords: PCDD/F PCDT Waste incineration SO2
a b s t r a c t SO2 levels in the flue gas from a laboratory-scale fluidized bed reactor combusting artificial municipal solid waste (MSW) were varied (resulting in four different SO2:HCl ratios 0, 0.2, 0.7 and 2.7 (by mass)) to study the effects of sulfur on the formation of polychlorinated dibenzo-p-dioxins (PCDDs), polychlorinated dibenzofurans (PCDFs) and polychlorinated dibenzothiophenes (PCDTs). Sampling was performed simultaneously at three fixed points in the post-combustion zone with temperatures of 400, 300 and 200 °C, under normal combustion conditions and both during and after transient combustion conditions. The findings indicate that sulfur has a greater inhibitory effect on PCDF formation than on PCDD formation and that the PCDD/PCDF ratio in the flue gas depends on both the SO2:HCl ratio in the flue gas and memory effects arising from transient combustion conditions. The results also indicate that the relative importance of different pathways shifts in the post-combustion zone; condensation products increasing with reductions in temperature and increases in residence time. However, these changes appear to depend on the SO2:HCl ratio in the flue gas and combustion conditions. Sulfur seems to inhibit the chlorination of PCDFs. A tendency for increased SO2 levels in the flue gas to increase levels of PCDTs was also detected, but the increases were much less significant than the reductions in PCDF levels. Ó 2009 Elsevier Ltd. All rights reserved.
1. Introduction Polychlorinated dibenzo-p-dioxins (PCDDs) and polychlorinated dibenzofurans (PCDFs) are toxic by-products emitted from combustion sources such as municipal solid waste (MSW) incineration plants. The formation of PCDD/Fs is generally believed to occur mainly via heterogeneous reactions (gas phase–solid phase interactions) at temperatures between 400 and 250 °C (Fängmark et al., 1994). The mechanisms of PCDD/F formation are still not fully understood, but two pathways have been shown: de novo synthesis from carbon residuals (Hell et al., 2001) and carbon structures such as polycyclic aromatic hydrocarbons (PAHs) (Iino et al., 1999), and condensation reactions of chlorinated precursors such as chlorophenols (CPh) (Ryu et al., 2005). Notably, formation of PCDDs via CPhs can occur in the gas phase (Sidhu et al., 1995), in reactions catalyzed by copper chloride (CuCl2) (Mulholland and Ryu, 2001), or from CPhs formed from carbon residuals (Hell et al., 2001). It is known that the 2,4,6-TriCPh, 2,3,4,6-TeCPh and PeCPHs correlates to the PCDD congeners: 1,3,6,8- and 1,3,7,9-TeCDD, 1,2,4,6,8/ 1,2,4,7,9- and 1,2,3,6,8-PeCDD and 1,2,3,4,6,8-HxCDD (Sidhu et al., 1995; Mulholland and Ryu, 2001), i.e. these CPhs give rise to a characteristic PCDD congener pattern. Wikström et al. (2003a) showed PCDD and PCDF formation rates to be maximal at 300–400 °C and * Corresponding author. Tel.: +46 907867623; fax: +46 90786755. E-mail address: [email protected] (J. Aurell). 0045-6535/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.chemosphere.2009.04.065
400–500 °C, respectively. This difference in optimal temperatures for PCDD and PCDF formation are presumably indicative of differences in their principal formation pathways. A temperature dependence in chlorophenol precursor abundance with low concentrations at higher temperatures (Hell et al., 2001) may be at least partly responsible for the rates of PCDD formation being greater at lower temperatures than rates of PCDF formation. In addition, chlorine gas (Cl2), produced by CuCl2/CuO-catalyzed transformation of HCl in the so-called Deacon process, has been shown to act as a chlorinating agent (Gullett et al., 1990). Although, ash-bound chlorine has been suggested to be the most important chlorinating source in de novo PCDD/F synthesis during full-scale combustion (Wikström et al., 2003b). In addition, Luijk et al. (1994) suggested that CuCl2 could act both as a catalyst and chlorinating agent. They also found that the relative PCDD and PCDF yields and congener patterns depended on the CuCl2 concentration; low CuCl2 concentrations leading to higher PCDD than PCDF yields, and a PCDD congener pattern characteristic of formation from chlorophenols. Sulfur dioxide has been shown to be an efficient inhibitor of the PCDD/F formation (Ogawa et al., 1996; Raghunathan and Gullett, 1996). However, its effects on PCDD/F formation mechanisms are not fully understood, although recent studies have suggested that the transformation of metal chlorides to metal sulfates may be the most important factor in reducing the PCDD/F concentrations (Ryan et al., 2006; Hunsinger et al., 2007). Ryan et al. (2006)
768
J. Aurell et al. / Chemosphere 76 (2009) 767–773
showed that sulfur reduces PCDF formation more than PCDD formation, and suggested this was due to CuSO4 having effects similar to CuCl2 on biaryl condensation. In addition, Gullett et al. (1992) showed that the transformation of the Cu-based catalyst to CuSO4 did not decrease the rate of the Deacon process, i.e. it did not reduce the production of Cl2, but shifted its optimum temperature from ca. 400 °C to ca. 500 °C. Sulfur has also been suggested to inhibit the Deacon process by converting Cl2 to HCl in the presence of water (Griffin, 1986), although this reaction was recently shown to make minor contributions to the inhibition (Ryan et al., 2006). In addition, the dibenzothiophenes (PCDTs) and polychlorinated thianthrenes (PCTAs) – the sulfur-analogues of PCDFs and PCDDs, respectively – have been suggested to be formed instead of PCDD/Fs (Gullett et al., 1992). The 2,3,7,8-TeCDT congener has been found to have similar toxicity properties (albeit weaker) to 2,3,7,8-TeCDD (Kopponen et al., 1994), warranting further investigation of the formation of PCDTs. Recent studies, both laboratory and full-scale investigations, have shown that not only increases in PCDD/F levels, but also alterations in the PCDD/PCDF ratios and congener patterns, occur during and after transient combustion conditions, i.e. the relative activities of the PCDD/F formation pathways are affected (NeuerEtscheidt et al., 2006; Aurell et al., 2009). However, while we found reductions in the PCDD/PCDF ratio during transient combustion conditions in the cited laboratory study (Aurell et al., 2009), increases in the PCDD/PCDF ratio were observed in the full-scale studies (Gass et al., 2002; Neuer-Etscheidt et al., 2006). A possible reason for this discrepancy was that there were no detectable SO2 levels in the flue gas in our laboratory study, prompting a further investigation to elucidate the effects of sulfur on the formation of PCDD/Fs under different combustion conditions. This may give valuable information in interpreting elevated PCDD/F concentrations emitted from MSW incineration plants and can also give further insight in the mechanism behind the formation of PCDD/F, which can lead to minimization of these toxic by-products. Thus, in the study presented here we varied SO2 levels, collected flue gas samples simultaneously at three locations/ temperatures in the post-combustion zone during normal combustion conditions and both during and after transient combustion conditions, in order to study the formation of PCDD/Fs. The formation of PCDTs was also investigated, although only under normal combustion conditions.
2. Experimental section 2.1. Description of the laboratory-scale incinerator and fuel The 5 kW solid fuel fed laboratory-scale fluidized-bed incinerator schematically illustrated in Fig. 1, which has been fully described elsewhere, was used in the study (Wikström et al., 1998; Aurell et al., 2009). Briefly, the laboratory-scale reactor was constructed to simulate the behavior of a full-scale waste incinerator and consisted mainly of a primary-combustion zone (bed), secondary-combustion zone (freeboard) and a post-combustion zone (convector). The fuel used was an artificial municipal solid waste (MSW) in the form of pellets that mainly consisted of paper, plastic, sawdust, starch, and metals such as copper and iron. The composition and contents of the fuel have been fully described in a previous study (Aurell et al., 2009). In summary, the contents of elements usually considered to be important in PCDD/F formation – Cl, Cu, S and Fe – were 0.7%, 0.007%, 0.07% and 1.0%, respectively. In order to increase the SO2 concentration in the flue gas SO2 gas was injected through I1 (Fig. 1) into the secondary-combustion zone. The SO2 gas was preheated to 250 °C by a small heating jacket before injection. 2.2. Experimental runs Experiments were conducted, with four different SO2 concentrations in the flue gas: 0, 80, 340, and 1495 mg m3. To evaluate the effects of sulfur on PCDD/F formation during stable and transient combustion conditions PCDD/Fs, PCDTs and PAHs were sampled during four combustion periods, as shown by the experimental timeline in Fig. 1 (except that PCDTs were only analyzed in samples collected under normal combustion conditions). The reproducibility of the experimental set-up was examined by conducting five experimental runs with 0 mg m3 SO2 in the flue gas. Three of these five runs only included the normal combustion condition period, i.e. samples were only collected during normal combustion conditions. An additional run with 340 mg m3 SO2 in the flue gas was also conducted, in which samples of PCDD/Fs, PCDTs and PAHs in the flue gases were collected during normal combustion conditions. Simultaneous flue gas sampling of PCDD/Fs, PCDTs and PAHs was performed at sampling points P3, P4 and P7 with flue gas temperatures of 400, 300, and 200 °C, respectively, over periods of 45 min during normal combustion conditions, 20 min during transient com-
Fig. 1. Schematic illustration of the fluidized bed incinerator, not to scale. Timeline for the experiments is shown in the inset panel.
769
J. Aurell et al. / Chemosphere 76 (2009) 767–773
bustion conditions and 30 min after the transient combustion conditions (during memory effect periods). However, in the replicate runs with no SO2 addition (experiment 1B) samples were only collected at P4 (300 °C) and P7 (200 °C), and for 30 min instead of 20 min, during the period with transient combustion conditions. The disturbances simulating transient combustion conditions were created by manually pulsing the fuel feed, i.e. irregular feeding. The residence times from the start of the post-combustion zone (P2) to the sampling points P3, P4 and P7 were approximately 1.4, 2.3, and 4.4 s, respectively (calculated from the air supply, elemental content of the fuel, local temperatures in the post-combustion zone, and volume of the flue duct). Fixed combustion settings in all runs were: fuel feed rate, 0.9 kg h1; primary-combustion zone temperature, 800–850 °C; secondary-combustion zone temperature, 800 °C; primary air flow 110 L min1 and secondary air flow 30 L min1. 2.3. Sampling and analyses Continuous emission measuring (CEM) of O2, H2O, CO2, SO2, NO2, CO, NO, HCl, NH3, N2O, and CH4 was performed during each of the experimental runs (Fig. 1). The PCDD/Fs, PCDTs and PAHs in the flue gas were sampled using the cooled probe polyurethane foam plug (PUFP) sampling technique, according to European standard method EN, 1948:1 (1997). Samples were cleaned-up according to standard methods (Liljelind et al., 2003; Aurell et al., 2009), then PCDD/Fs and PAHs were analyzed as described elsewhere (Aurell et al., 2009). The PCDTs were separated using an Agilent 6890 gas chromatograph equipped with a 60 m J&W Scientific column with an internal diameter of 0.25 mm and a film thickness of 0.25 lm, then detected using a Waters AutoSpec ULTIMA NT 2000D high resolution mass spectrometer. Two of the most intense ions of the molecular ion isotope distribution cluster were monitored for each PCDT homologue. The analysis was time segmented in eight segments, one for each DCDT homologue. For positive identification the CDT congeners had to elute within the expected retention time window (established using a qualitative standard of mono- through octa-CDTs) and have an isotope ratio (of the quantification and confirmation ions) within ±20% of the theoretical. The PCDTs were semiquantified versus 2,3,7,8-tetra-CDT (assuming equal molar response factors) using the internal standard technique (13C12-PCDDs was used as internal standards). All derived concentrations were normalized (norm) to 1 atm, 0 °C, dry gas 11% O2.
3. Results and discussion 3.1. Continuous emission measuring during stable and transient combustion conditions During the stable combustion conditions before and after the transient combustion conditions no CO peaks were detected in any of the experimental runs and the CO levels in the flue gas were low (<5 ppm; Table 1). The SO2 concentrations in the flue gas were below detection limits (3 mg m3) in the runs with no injection of SO2, resulting in a SO2:HCl ratio of virtually zero. In the runs with SO2 gas injection the SO2 concentrations in the flue gas were 80, 340 and 1495 mg m3 (Table 1), resulting in SO2:HCl ratios of 0.2, 0.7 and 2.7 (by mass), respectively. Adding SO2 gas to the combustion zone altered the composition of the flue gases in the postcombustion zone, causing SO2, HCl and N2O levels to be higher, but NO2 levels to be lower than in runs with no SO2 injection (Table 1). The increased HCl levels observed following the addition of sulfur were probably due to the conversion of alkali metal salts (KCl and NaCl) to metal sulfates, with the production of HCl in the presence of water and oxygen according to R1.
1 2MCl þ H2 O þ SO2 þ O2 $ M2 SO4 þ 2HCl ðM ¼ K; NaÞ 2
ðR1Þ
NOx (NO and NO2) and N2O emissions levels have been shown by numerous authors to be affected by the fuel composition. In fluidized bed systems most NO emissions originate from the fuel via oxidation of NH3, and N2O is mainly formed via hydrogen cyanide (HCN), as shown in via R2. The simultaneous reductions and increases in NO2 and N2O emissions, respectively, due to SO2 addition found in the present study are consistent with other studies (DamJohansen et al., 1993), which have suggested that SO2 can decrease the rate of HCN oxidation and thus change the relative amounts of NO2 and N2O emissions.
HCN þ O ! NCO þ H
ðR2Þ
NCO þ NO ! N2 O þ CO
In addition, in experiment 4, with an SO2 level of 1495 mg m3 in the flue gas (SO2:HCl ratio, 2.7), detectable levels of methane (CH4) were found in the flue gas. This is also consistent with earlier studies, that suggested that SO2 inhibits not only the conversion of HCN but also reactions of combustible gases such as CH4 and CO
Table 1 Average values of the continuous emission measurements during the sampling periods (in dry gas).a Experiment
Description
O2 vol %
H2O vol %
CO2 vol %
CO ppm
CH4 ppm
SO2 mg m3
HCl mg m3
NO ppm
NO2 ppm
N2O ppm
NH3 ppm
1 – SO2:HCl 0 1 – SO2:HCl 0 1 – SO2:HCl 0 1 – SO2:HCl 0 1B – SO2:HCl 0 1B – SO2:HCl 0 1B – SO2:HCl 0 2 – SO2:HCl 0.1 2 – SO2:HCl 0.1 2 – SO2:HCl 0.1 2 – SO2:HCl 0.1 3 – SO2:HCl 0.4 3 – SO2:HCl 0.4 3 – SO2:HCl 0.4 3 – SO2:HCl 0.4 4 – SO2:HCl 1.6 4 – SO2:HCl 1.6 4 – SO2:HCl 1.6 4 – SO2:HCl 1.6
Normal combustionb Transient combustion Memory effects 1 Memory effects 2 Transient combustion Memory effects 1 Memory effects 2 Normal combustion Transient combustion Memory effects 1 Memory effects 2 Normal combustionc Transient combustion Memory effects 1 Memory effects 2 Normal combustion Transient combustion Memory effects 1 Memory effects 2
11 10 12 11 8.8 11 10 11 8.7 11 11 11 11 11 10 10 9.4 10 9.4
7.9 7.7 6.5 7.4 9.5 8.0 8.5 8.0 9.3 7.8 8.1 7.6 7.5 7.4 7.9 7.7 8.1 7.7 8.1
9.2 8.9 7.3 8.6 11 9.4 10 9.5 11 9.1 9.5 8.9 8.6 8.5 9.2 9.3 9.5 9.2 9.8
<5 1900 <5 <5 1600 <5 <5 <5 1800 <5 <5 <5 2500 <5 <5 <5 2700 <5 <5
nd 320 nd nd 250 nd nd nd 290 nd nd nd 550 nd nd <5 530 <5 <5
nd 34 nd nd 42 nd nd 81 215 66 52 340 600 365 315 1490 1570 1470 1520
315 240 280 315 315 360 370 400 360 390 430 475 255 460 475 550 400 550 580
470 390 420 460 470 540 530 430 390 430 450 400 330 410 430 410 330 400 410
32 20 31 34 14 33 34 12 3 5 7 10 2 3 5 10 3 4 6
52 28 40 42 29 48 44 69 41 58 64 70 36 61 59 76 41 61 61
nd 34 <5 <5 34 10 <5 nd 39 10 <5 nd 116 10 7 nd 76 9 7
a b c
nd = Not detected, detection limit 1 ppm or 3 mg m3. Mean values of five runs. Mean values of two runs.
770
J. Aurell et al. / Chemosphere 76 (2009) 767–773
Fig. 2. Concentrations of PC1–8DFs (A), PC1–8DDs (B) and PC1–8DTs (C) during normal combustion conditions at three different sampling points P3 (400 °C), P4 (300 °C) and P7 (200 °C) in the post-combustion zone. Cumulative amounts at the respective sampling point, error bars denote 2 SDs.
(Miccio et al., 2001). The reductions in N2O emissions observed during transient combustion conditions compared to stable conditions may be due to the ability of N2O to react with CO, forming CO2 and N2, in the presence of char (a carbonaceous solid product of fuel devolatilization) (Bonn et al., 1995). 3.2. PCDD/F emissions and formation during stable combustion conditions As shown in Fig. 2A and B, with a SO2:HCl ratio of 0 in the flue gas levels of the PCDD/Fs were similar at all three sampling points/ temperatures, i.e. most of the net formation occurred prior to the first sampling point (400 °C). In addition, the SO2:HCl ratio of 0.2 in the flue gas was not sufficient to reduce PCDD/F concentrations (note, data for the sample collected at 200 °C are missing due to losses during its clean-up). In contrast to the results of the experiments with SO2:HCl ratios of 0 and 0.2 in the flue gas, with an SO2:HCl ratio of 2.7 in the flue gas PCDD/F concentrations (especially PCDD concentrations) increased with reductions in temperature and increases in residence time in the post-combustion zone. However, reductions (up to 65%) were observed in both PCDD and PCDF concentrations at the first sampling point P3 (400 °C) with SO2:HCl ratios of 0.7 and 2.7. Although at the last sampling point P7 (200 °C) a significant reduction in PCDF levels was only found with a SO2:HCl ratio of 2.7 in the flue gas. The PCDD/F homologue profiles and congener patterns were found to differ between the experiments with SO2:HCl ratios 60.2 and SO2:HCl ratios P0.7 in the flue gas, implying that adding sulfur to the flue gas results in changes in the formation reactions, particularly for the PCDDs. 3.2.1. PCDDs As shown in Fig. 3, the PCDD homologue profiles at the 400 °C sampling point had lower proportions of the TeCDD to HxCDD homologue groups in the experiments with SO2:HCl ratios P0.7 in the flue gas than in experiments with SO2:HCl ratios 60.2. However, levels of these homologue groups increased with reductions in temperature and increases in residence time with SO2:HCl ratios
P0.7. These changes were found to be largely due to increases in levels of the 1,3,6,8- and 1,3,7,9-TeCDD, 1,2,4,6,8/1,2,4,7,9- and 1,2,3,6,8-PeCDD and 1,2,3,4,6,8-HxCDD congeners, which are known to form from 2,4,6-trichlorophenols, 2,3,4,6-tetrachlorphenols and penta-chlorophenols via condensation reactions (Mulholland and Ryu, 2001). The fraction of the 1,3,6,8-TeCDD congener of the total sum TeCDD isomers (isomer fraction) increased from 13% to 35% from the first sampling point (400 °C) to the last sampling point (200 °C) with the SO2:HCl ratio of 2.7, and from 10% to 18% when no SO2 was added to the flue gas. Formation of these PCDD congeners from PCPhs produced from carbon structures (de novo) has been shown to increase with reductions in temperature from 400 °C to 250 °C (Hell et al., 2001), which may explain the larger amounts of the congeners found at 200 °C than at 400 °C in the present study. In addition, the PCDD levels at 200 °C in experiment 4 with the SO2:HCl ratio of 2.7 in the flue gas may be a result of increased de novo formation due to increased amounts of carbon residuals, as indicated by the inhibited oxidation of CH4, or the presumably lower CuCl2 contents, as shown by Luijk et al. (1994). The results also indicate that the conversion of CuCl2 (and other metal chlorides) to CuSO4 (and other metal sulfates) does not affect the catalysis of condensation reactions. 3.2.2. PCDFs Only slight differences were found in the PCDF homologue profiles between the experiments with SO2:HCl ratios 60.2 and SO2:HCl ratios P0.7 in the flue gas (Fig. 3). The only changes found in the congener patterns were increases in the isomer fraction of 1,2,3,4/2,3,4,9-TeCDF congeners (which co-eluted from the column used in the study) at the last sampling point (200 °C), from 4 ± 3% (2SD) in runs with no added SO2 to 11% in the experimental runs with the SO2:HCl ratio of 2.7 in the flue gas. These congeners have also been shown by Ryu et al. (2005) to be formed from chlorophenols, although they co-eluted with the 2,4,6,7/1,4,6,9-TeCDF congeners from the column used in their study. These congeners may have increased due to the increased CH4 level in the flue gas that appeared with the high SO2:HCl ratio (2.7) in the flue gas, as dis-
J. Aurell et al. / Chemosphere 76 (2009) 767–773
771
Fig. 3. PCDD, PCDT and PCDF homologue profiles in experiments: 1-SO2:HCl 0 (A), 2-SO2:HCl 0.2 (B), 3-SO2:HCl 0.7 (C) and 4-SO2:HCl 2.7 (D) during normal combustion conditions. Error bars denote 1 SD.
cussed above for the PCDDs. The isomer fraction of the 1,2,3,4/ 2,3,4,9-TeCDF congeners was also found to be higher when levels of CuCl2 were low in a study by Luijk et al. (1994). 3.3. PCDTs emissions and formation during stable combustion conditions As shown in Fig. 2C, with an SO2:HCl ratio of 0 in the flue gas PCDT levels were similar at all three sampling points/temperatures, i.e. most of the net formation occurred prior to the first sampling point (400 °C) as for the PCDD/Fs. The PCDT concentrations were significantly higher with an SO2:HCl flue gas ratio of 0.2 than with the SO2:HCl ratio of zero (Fig. 2C), although the SO2:HCl ratio of 0.2 in the flue gas was not sufficient to reduce the PCDD/F concentrations. The observed increase in PCDT levels may have been due to increased sulfonation of the carbon residuals or precursors. Furthermore, as shown in Fig. 2, the PCDT homologue profiles indicated that the degree of chlorination was higher with an SO2:HCl ratio P0.2 than with the ratio of 0, which may also have been due to the amounts of sulfonated backbone structures being lower with no added SO2 in the flue gas. However, the PCDT concentrations were ca. 10-fold lower than the PCDF concentrations, implying that the sulfur analogues were not formed instead of the PCDFs. Furthermore, the PCDT homologue profiles suggest that the PCDTs were formed via similar pathway(s) to the PCDFs, although the increased PCDT levels at 200 °C with an SO2:HCl ratio of 2.7 in the
flue gas suggest that their formation pathway(s) were also similar to those of the PCDDs. 3.4. Effects of SO2 on PCDD/F formation under transient combustion conditions The sum PAH concentrations observed during the sampling periods of the transient combustion phase in experiments 1 and 1B 1–4 (11, 10, 18, 25, and 15 ng PAH m3, respectively) did not correlate with the average CO concentrations in the flue gas in these periods (1900, 1600, 1800, 2500, and 2700 ppm, respectively). A higher CO concentration (2700 ppm), but similar PAH concentration (11 ng PAH m3), was found in the flue gas with the SO2:HCl ratio of 2.7 compared to those obtained with the SO2:HCl ratio of 0 (1900 ppm and 15 ng PAH m3, respectively). The PCDD/F concentrations increased substantially during and after the transient combustion conditions compared to those found during normal combustion conditions (Fig. 4). The trends in the PCDD/F concentrations differed between the replicate runs with the SO2:HCl ratio of zero, possibly due to the longer sampling time (30 min instead of 20 min) in experiment 1B under the transient combustion conditions. However, the PCDD/PCDF ratio was similar in the two runs. As also shown in Fig. 4, with the SO2:HCl ratio of 2.7 in the flue gas there were indications of reduced PCDF levels but increased PCDD levels compared to those observed with the SO2:HCl ratio of 0 in the flue gas under transient combustion con-
772
J. Aurell et al. / Chemosphere 76 (2009) 767–773
Fig. 4. Sum PCDD (A) and sum PCDF (B) concentrations at the sampling points P3 (400 °C), P4 (300 °C) and P7 (200 °C) in the post-combustion zone during the transient combustion period and following memory effect periods (M1 and M2). Cumulative amounts at the respective sampling point.
Table 2 PC4–8DD to PC4–8DF ratio in the collected flue gas samples at 400, 300 and 200 °C under the different combustion periodsa. Combustion period
SO2:HCl 0 400 °C
300 °C
200 °C
400 °C
300 °C
200 °C
400 °C
300 °C
200 °C
400 °C
300 °C
200 °C
Normalb Transient Memory effects 1 Memory effects 2
0.12 0.05 0.09 0.11
0.11 0.06d 0.13d 0.14d
0.11 0.06d 0.13d 0.14d
0.10 0.16 0.28 0.14
0.10 0.21 0.53 0.67
–c 0.20 0.51 0.72
0.11 0.29 0.32 0.14
0.14 0.37 0.83 1.00
0.20 0.41 0.87 1.12
0.17 0.52 0.42 0.21
0.24 0.68 0.90 1.13
0.33 0.77 0.98 1.18
a b c d
SO2:HCl 0.2
SO2:HCl 0.7
SO2:HCl 2.7
PCDD/PCDF in mass ratio. Mean values: of five runs of experiment 1 with SO2:HCl ratio 0 and two runs of experiment 3 with SO2:HCl ratio 0.7. Excluded due to losses during sample clean up. Mean value of two runs.
ditions. In addition, the PCDD concentrations increased with reductions in temperature in the post-combustion zone during the memory effects periods. These changes were most apparent in the PCDD/PCDF ratios obtained, as shown in Table 2, which revealed that the PCDD/PCDF ratio increased with increases in the SO2:HCl ratio in the flue gas not only during the transient combustion conditions but also in all of the other combustion periods. These findings further support the indications in experiment 4, with the SO2:HCl ratio of 2.7, that sulfur has a stronger inhibitory effect on PCDF reaction pathways than on PCDD reactions. The increases in PCDD/PCDF ratios observed during transient combustion conditions are consistent with findings in full-scale plant investigations (Gass et al., 2002; Neuer-Etscheidt et al., 2006). 3.4.1. PCDDs Irrespective of the SO2:HCl ratio in the flue gas and sampling temperature the PCDD homologue profiles under the transient combustion conditions were similar to those found under normal combustion conditions without SO2 addition, thus the transient combustion conditions had a stronger effect on the homologue profile than the SO2 level in the flue gas. Furthermore, the PCDD congeners patterns were also independent of the SO2:HCl ratio in the flue gas and temperature in the post-combustion zone during transient combustion conditions. In all cases the PCDD congener pat-
terns during the transient combustion periods were different from those found under normal combustion conditions, but similar to those found under normal combustion conditions with an SO2:HCl ratio of 1.6 at 200 °C. The TeCDD and PeCDD congeners most strongly affected by transient combustion conditions were the 1,3,6,8- and 1,3,7,9-TeCDD, 1,2,4,6,8/1,2,4,7,9- and 1,2,3,6,8PeCDD congeners, e.g. the isomer fraction of the 1,3,6,8-TeCDD congener was ca. 40% under transient combustion conditions, and ca. 15% under normal combustion conditions. Again, these congeners are known to form from chlorophenols. However, during the memory effect periods the isomer fraction of these congeners increased with both reductions in temperature in the post-combustion zone and with addition of sulfur to the flue gas, e.g. the 1,3,6,8-TeCDD isomer fraction increased from 18% to 48% with the SO2:HCl ratio of 2.7, and from 15% to 28% with the SO2:HCl ratio of 0, between the 400 °C and 200 °C sampling points. These results imply that the formation of condensation reaction products is not only dependent on the temperature but also on the combustion conditions. 3.4.2. PCDF Under the transient combustion conditions the PCDF homologue profiles differed significantly from those observed under normal conditions, the most abundant homologues being MoCDFs, the proportions of which also increased with increases in the SO2:HCl
J. Aurell et al. / Chemosphere 76 (2009) 767–773
ratio in the flue gas. These findings suggest that sulfur inhibits the chlorination of carbon residuals (de novo synthesis) and/or lightlyor non-chlorinated precursors. The PCDF congener patterns were also affected by the transient combustion conditions, and as for the PCDDs the changes in congener patterns were similar at all sampling points/temperatures during transient combustion conditions. However, the isomer fractions of the TeCDF and PeCDF congeners most affected by the transient conditions (1,2,3,4/2,3,4,9-, 2,4,6,7and 2,3,4,6-TeCDF, 2,3,4,6,8- and 2,3,4,6,7-PeCDF) were 40–50% lower with an addition of SO2 to the flue gas than with no addition of SO2. Of these congeners the 1,2,3,4/2,3,4,9- and 2,4,6,7-TeCDF and 2,3,4,6,8-PeCDF have been found to form from chlorophenols by Ryu et al. (2005) (the TeCDF congeners co-eluted with the 1,4,6,9-TeCDF congener and the PeCDF congener co-eluted with 1,2,3,4,7/1,2,3,5,6-TeCDF congeners from the column used in their study) while the 2,4,6,7- and 2,3,4,6-TeCDF, 2,3,4,6,8- and 2,3,4,6,7-PeCDF congeners were found to be the most abundantly formed congeners in de novo synthesis experiment from the PAH perylene (Iino et al., 1999; Weber et al., 2001). In the memory effect periods the PCDF homologue profiles returned to those found during normal combustion conditions. The 1,2,3,4/2,3,4,9- and 2,4,6,7-TeCDF and 2,3,4,6,8-PeCDF were the only TeCDF and PeCDF congeners that increased with reductions in temperature in the post-combustion zone during the memory effect periods in experiments with and without addition of SO2 to the flue gas, indicating that these congeners are products of condensation reactions from chlorophenols in this case. 4. Conclusions The findings indicate that sulfur has a greater inhibitory effect on PCDF formation than on PCDD formation and that the PCDD/ PCDF ratio in the flue gas depends on both the SO2:HCl ratio in the flue gas and the ash deposits that arise from transient combustion conditions. The results also indicate that the relative importance of different pathways shifts in the post-combustion zone; condensation products increasing with reductions in the temperature and increases in residence time. However, these changes were also found to depend on the SO2:HCl ratio in the flue gas and combustion conditions. The presence of sulfur appeared to inhibit the chlorination of PCDFs. A tendency for increased SO2 levels in the flue gas to increase levels of PCDTs was also detected, but the increases were much less significant than the reductions in PCDF levels. Acknowledgments This study was partly financed by PSO (Public Service Obligations) funds from the Danish Ministry for Transport and Energy, Energinet.dk, whose support is gratefully acknowledged. Special thanks are due to Marie Calla for assistance during experiments and laboratory work and Per Liljelind for help with the GC/MS analyses. References Aurell, J., Fick, J., Marklund, S., 2009. Effects of transient combustion conditions on the formation of PCDD/F, PCBz and PAH during MSW incineration. Environ. Eng. Sci. 26, 509–520.
773
Bonn, B., Pelz, G., Baumann, H., 1995. Formation and decomposition of N2O in fluidized bed boilers. Fuel 74, 165–171. Dam-Johansen, K., Åmand, L.-E., Leckner, B., 1993. Influence of SO2 on the NO/N2O chemistry in fluidized bed combustion 2. Interpretation of full-scale observations based on laboratory experiments. Fuel 72, 565–571. European Committee for Standardization, EN 1948:1–3, 1997. Stationary Source Emissions - Determination of the Mass Concentration of PCDDs/PCDFs. Fängmark, I., Strömberg, B., Berge, N., Rappe, C., 1994. The influence of post combustion temperature profiles on the formation of PCDDs, PCDFs, PCBzs and PCBs in a pilot incinerator. Environ. Sci. Technol. 28, 624–629. Gass, C.H., Luder, K., Wilken, M., 2002. PCDD/F-emissions during cold start-up and shut-down of a municipal waste incinerator. Organohal. Compd. 56, 193– 196. Griffin, R.D., 1986. A new theory of dioxin formation in municipal solid waste combustion. Chemosphere 15, 1987–1990. Gullett, B.K., Bruce, K.R., Beach, L.O., 1990. Formation of chlorinated organics during solid waste combustion. Waste Manage. Res. 8, 203–214. Gullett, B.K., Bruce, K.R., Beach, L.O., 1992. Effect of sulfur dioxide on the formation mechanism of polychlorinated dibenzodioxin and dibenzofuran in municipal waste combustors. Environ. Sci. Technol. 26, 1938–1943. Hell, K., Stieglitz, L., Dinjus, E., 2001. Mechanistic aspects of the de-novo synthesis of PCDD/PCDF on model mixtures and MSWI fly ashes using amorphous 12C- and 13C-labeled carbon. Environ. Sci. Technol. 35, 3892–3898. Hunsinger, H., Seifert, H., Jay, K., 2007. Reduction of PCDD/F formation in MSWI by a process-integrated SO2 cycle. Environ. Eng. Sci. 24, 1145–1159. Iino, F., Imagawa, T., Takeuchi, M., Sadakata, M., 1999. De novo synthesis mechanism of polychlorinated dibenzofurans from polycyclic aromatic hydrocarbons and the characteristic isomers of polychlorinated naphthalenes. Environ. Sci. Technol. 33, 1038–1043. Kopponen, P., Sinkkonen, S., Poso, A., Gynther, J., Karenlampi, S., 1994. Sulfur analogs of polychlorinated dibenzo-p-dioxins, dibenzofurans and diphenyl ethers as inducers of CYP1A1 in mouse heptatoma-cell culture and structure– activity-relationships. Environ. Toxicol. Chem. 13, 1543–1548. Liljelind, P., Söderström, G., Hedman, B., Karlsson, S., Lundin, L., Marklund, S., 2003. Method for multiresidue determination of halogenated aromatics and PAHs in combustion-related samples. Environ. Sci. Technol. 37, 3680–3686. Luijk, R., Akkerman, D.M., Slot, P., Olie, K., Kapteijn, F., 1994. Mechanism of formation of polychlorinated dibenzo-p-dioxins and dibenzofurans in the catalyzed combustion of carbon. Environ. Sci. Technol. 28, 312–321. Miccio, F., Löffler, G., Wargadalam, V.J., Winter, F., 2001. The influence of SO2 level and operating conditions on NOx and N2O emissions during fluidised bed combustion of coals. Fuel 80, 1555–1566. Mulholland, J.A., Ryu, J.-Y., 2001. Formation of polychlorinated dibenzo-p-dioxins by CuCl2-catalyzed condensation of 2,6 chlorinated phenols. Combust. Sci. Technol. 169, 107–126. Neuer-Etscheidt, K., Nordsieck, H.O., Liu, Y., Kettrup, A., Zimmermann, R., 2006. PCDD/F and other micropollutants in MSWI crude gas and ashes during plant start-up and shut-down processes. Environ. Sci. Technol. 40, 342– 349. Ogawa, H., Orita, N., Horaguchi, M., Suzuki, T., Okada, M., Yasuda, S., 1996. Dioxin reduction by sulfur component addition. Chemosphere 32, 151–157. Raghunathan, K., Gullett, B.K., 1996. Role of sulfur in reducing PCDD and PCDF formation. Environ. Sci. Technol. 30, 1827–1834. Ryan, P.S., Li, X.-D., Gullett, B.K., Lee, C.W., Clayton, M., Touati, A., 2006. Experimental study on the effect of SO2 on PCDD/F emissions: determination of the importance of gas-phase versus solid-phase reactions in PCDD/F formation. Environ. Sci. Technol. 40, 7040–7047. Ryu, J.-Y., Mulholland, J.A., Takeuchi, M., Kim, D.H., Hatanaka, T., 2005. CuCl2catalyzed PCDD/F formation and congener patterns from phenols. Chemosphere 61, 1312–1326. Sidhu, S.S., Maqsud, L., Dellinger, B., 1995. The homogeneous, gas-phase formation of chlorinated and brominated dibenzo-p-dioxin from 2,4,6-trichloro- and 2,4,6-tribromophenols. Combust. Flame 100, 11–20. Weber, R., Iino, F., Imagwa, T., Takeuchi, M., Sakurai, T., Sadakata, M., 2001. Formation of PCDF, PCDD, PCB, and PCN in de novo synthesis from PAH: Mechanistic aspects and correlation to fluidized bed incinerators. Chemosphere 44, 1429–1438. Wikström, E., Andersson, P., Marklund, S., 1998. Design of a laboratory scale fluidized bed reactor. Rev. Sci. Instrum. 69, 1850–1859. Wikström, E., Ryan, S., Touati, A., Gullett, B.K., 2003a. Key parameters for de novo formation of polychlorinated dibenzo-p-dioxins and dibenzofurans. Environ. Sci. Technol. 37, 1962–1970. Wikström, E., Ryan, S., Touati, A., Telfer, M., Tabor, D., Gullett, B.K., 2003b. Importance of chlorine speciation on de novo formation of polychlorinated dibenzo-p-dioxins and polychlorinated dibenzofurans. Environ. Sci. Technol. 37, 1108–1113.
Contents lists available at ScienceDirect
Chemosphere journal homepage: www.elsevier.com/locate/chemosphere
Effects of sulfur on PCDD/F formation under stable and transient combustion conditions during MSW incineration Johanna Aurell *, Jerker Fick, Peter Haglund, Stellan Marklund Department of Chemistry, Umeå University, SE 901 87 Umeå, Sweden
a r t i c l e
i n f o
Article history: Received 15 December 2008 Received in revised form 25 April 2009 Accepted 28 April 2009 Available online 30 May 2009 Keywords: PCDD/F PCDT Waste incineration SO2
a b s t r a c t SO2 levels in the flue gas from a laboratory-scale fluidized bed reactor combusting artificial municipal solid waste (MSW) were varied (resulting in four different SO2:HCl ratios 0, 0.2, 0.7 and 2.7 (by mass)) to study the effects of sulfur on the formation of polychlorinated dibenzo-p-dioxins (PCDDs), polychlorinated dibenzofurans (PCDFs) and polychlorinated dibenzothiophenes (PCDTs). Sampling was performed simultaneously at three fixed points in the post-combustion zone with temperatures of 400, 300 and 200 °C, under normal combustion conditions and both during and after transient combustion conditions. The findings indicate that sulfur has a greater inhibitory effect on PCDF formation than on PCDD formation and that the PCDD/PCDF ratio in the flue gas depends on both the SO2:HCl ratio in the flue gas and memory effects arising from transient combustion conditions. The results also indicate that the relative importance of different pathways shifts in the post-combustion zone; condensation products increasing with reductions in temperature and increases in residence time. However, these changes appear to depend on the SO2:HCl ratio in the flue gas and combustion conditions. Sulfur seems to inhibit the chlorination of PCDFs. A tendency for increased SO2 levels in the flue gas to increase levels of PCDTs was also detected, but the increases were much less significant than the reductions in PCDF levels. Ó 2009 Elsevier Ltd. All rights reserved.
1. Introduction Polychlorinated dibenzo-p-dioxins (PCDDs) and polychlorinated dibenzofurans (PCDFs) are toxic by-products emitted from combustion sources such as municipal solid waste (MSW) incineration plants. The formation of PCDD/Fs is generally believed to occur mainly via heterogeneous reactions (gas phase–solid phase interactions) at temperatures between 400 and 250 °C (Fängmark et al., 1994). The mechanisms of PCDD/F formation are still not fully understood, but two pathways have been shown: de novo synthesis from carbon residuals (Hell et al., 2001) and carbon structures such as polycyclic aromatic hydrocarbons (PAHs) (Iino et al., 1999), and condensation reactions of chlorinated precursors such as chlorophenols (CPh) (Ryu et al., 2005). Notably, formation of PCDDs via CPhs can occur in the gas phase (Sidhu et al., 1995), in reactions catalyzed by copper chloride (CuCl2) (Mulholland and Ryu, 2001), or from CPhs formed from carbon residuals (Hell et al., 2001). It is known that the 2,4,6-TriCPh, 2,3,4,6-TeCPh and PeCPHs correlates to the PCDD congeners: 1,3,6,8- and 1,3,7,9-TeCDD, 1,2,4,6,8/ 1,2,4,7,9- and 1,2,3,6,8-PeCDD and 1,2,3,4,6,8-HxCDD (Sidhu et al., 1995; Mulholland and Ryu, 2001), i.e. these CPhs give rise to a characteristic PCDD congener pattern. Wikström et al. (2003a) showed PCDD and PCDF formation rates to be maximal at 300–400 °C and * Corresponding author. Tel.: +46 907867623; fax: +46 90786755. E-mail address: [email protected] (J. Aurell). 0045-6535/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.chemosphere.2009.04.065
400–500 °C, respectively. This difference in optimal temperatures for PCDD and PCDF formation are presumably indicative of differences in their principal formation pathways. A temperature dependence in chlorophenol precursor abundance with low concentrations at higher temperatures (Hell et al., 2001) may be at least partly responsible for the rates of PCDD formation being greater at lower temperatures than rates of PCDF formation. In addition, chlorine gas (Cl2), produced by CuCl2/CuO-catalyzed transformation of HCl in the so-called Deacon process, has been shown to act as a chlorinating agent (Gullett et al., 1990). Although, ash-bound chlorine has been suggested to be the most important chlorinating source in de novo PCDD/F synthesis during full-scale combustion (Wikström et al., 2003b). In addition, Luijk et al. (1994) suggested that CuCl2 could act both as a catalyst and chlorinating agent. They also found that the relative PCDD and PCDF yields and congener patterns depended on the CuCl2 concentration; low CuCl2 concentrations leading to higher PCDD than PCDF yields, and a PCDD congener pattern characteristic of formation from chlorophenols. Sulfur dioxide has been shown to be an efficient inhibitor of the PCDD/F formation (Ogawa et al., 1996; Raghunathan and Gullett, 1996). However, its effects on PCDD/F formation mechanisms are not fully understood, although recent studies have suggested that the transformation of metal chlorides to metal sulfates may be the most important factor in reducing the PCDD/F concentrations (Ryan et al., 2006; Hunsinger et al., 2007). Ryan et al. (2006)
768
J. Aurell et al. / Chemosphere 76 (2009) 767–773
showed that sulfur reduces PCDF formation more than PCDD formation, and suggested this was due to CuSO4 having effects similar to CuCl2 on biaryl condensation. In addition, Gullett et al. (1992) showed that the transformation of the Cu-based catalyst to CuSO4 did not decrease the rate of the Deacon process, i.e. it did not reduce the production of Cl2, but shifted its optimum temperature from ca. 400 °C to ca. 500 °C. Sulfur has also been suggested to inhibit the Deacon process by converting Cl2 to HCl in the presence of water (Griffin, 1986), although this reaction was recently shown to make minor contributions to the inhibition (Ryan et al., 2006). In addition, the dibenzothiophenes (PCDTs) and polychlorinated thianthrenes (PCTAs) – the sulfur-analogues of PCDFs and PCDDs, respectively – have been suggested to be formed instead of PCDD/Fs (Gullett et al., 1992). The 2,3,7,8-TeCDT congener has been found to have similar toxicity properties (albeit weaker) to 2,3,7,8-TeCDD (Kopponen et al., 1994), warranting further investigation of the formation of PCDTs. Recent studies, both laboratory and full-scale investigations, have shown that not only increases in PCDD/F levels, but also alterations in the PCDD/PCDF ratios and congener patterns, occur during and after transient combustion conditions, i.e. the relative activities of the PCDD/F formation pathways are affected (NeuerEtscheidt et al., 2006; Aurell et al., 2009). However, while we found reductions in the PCDD/PCDF ratio during transient combustion conditions in the cited laboratory study (Aurell et al., 2009), increases in the PCDD/PCDF ratio were observed in the full-scale studies (Gass et al., 2002; Neuer-Etscheidt et al., 2006). A possible reason for this discrepancy was that there were no detectable SO2 levels in the flue gas in our laboratory study, prompting a further investigation to elucidate the effects of sulfur on the formation of PCDD/Fs under different combustion conditions. This may give valuable information in interpreting elevated PCDD/F concentrations emitted from MSW incineration plants and can also give further insight in the mechanism behind the formation of PCDD/F, which can lead to minimization of these toxic by-products. Thus, in the study presented here we varied SO2 levels, collected flue gas samples simultaneously at three locations/ temperatures in the post-combustion zone during normal combustion conditions and both during and after transient combustion conditions, in order to study the formation of PCDD/Fs. The formation of PCDTs was also investigated, although only under normal combustion conditions.
2. Experimental section 2.1. Description of the laboratory-scale incinerator and fuel The 5 kW solid fuel fed laboratory-scale fluidized-bed incinerator schematically illustrated in Fig. 1, which has been fully described elsewhere, was used in the study (Wikström et al., 1998; Aurell et al., 2009). Briefly, the laboratory-scale reactor was constructed to simulate the behavior of a full-scale waste incinerator and consisted mainly of a primary-combustion zone (bed), secondary-combustion zone (freeboard) and a post-combustion zone (convector). The fuel used was an artificial municipal solid waste (MSW) in the form of pellets that mainly consisted of paper, plastic, sawdust, starch, and metals such as copper and iron. The composition and contents of the fuel have been fully described in a previous study (Aurell et al., 2009). In summary, the contents of elements usually considered to be important in PCDD/F formation – Cl, Cu, S and Fe – were 0.7%, 0.007%, 0.07% and 1.0%, respectively. In order to increase the SO2 concentration in the flue gas SO2 gas was injected through I1 (Fig. 1) into the secondary-combustion zone. The SO2 gas was preheated to 250 °C by a small heating jacket before injection. 2.2. Experimental runs Experiments were conducted, with four different SO2 concentrations in the flue gas: 0, 80, 340, and 1495 mg m3. To evaluate the effects of sulfur on PCDD/F formation during stable and transient combustion conditions PCDD/Fs, PCDTs and PAHs were sampled during four combustion periods, as shown by the experimental timeline in Fig. 1 (except that PCDTs were only analyzed in samples collected under normal combustion conditions). The reproducibility of the experimental set-up was examined by conducting five experimental runs with 0 mg m3 SO2 in the flue gas. Three of these five runs only included the normal combustion condition period, i.e. samples were only collected during normal combustion conditions. An additional run with 340 mg m3 SO2 in the flue gas was also conducted, in which samples of PCDD/Fs, PCDTs and PAHs in the flue gases were collected during normal combustion conditions. Simultaneous flue gas sampling of PCDD/Fs, PCDTs and PAHs was performed at sampling points P3, P4 and P7 with flue gas temperatures of 400, 300, and 200 °C, respectively, over periods of 45 min during normal combustion conditions, 20 min during transient com-
Fig. 1. Schematic illustration of the fluidized bed incinerator, not to scale. Timeline for the experiments is shown in the inset panel.
769
J. Aurell et al. / Chemosphere 76 (2009) 767–773
bustion conditions and 30 min after the transient combustion conditions (during memory effect periods). However, in the replicate runs with no SO2 addition (experiment 1B) samples were only collected at P4 (300 °C) and P7 (200 °C), and for 30 min instead of 20 min, during the period with transient combustion conditions. The disturbances simulating transient combustion conditions were created by manually pulsing the fuel feed, i.e. irregular feeding. The residence times from the start of the post-combustion zone (P2) to the sampling points P3, P4 and P7 were approximately 1.4, 2.3, and 4.4 s, respectively (calculated from the air supply, elemental content of the fuel, local temperatures in the post-combustion zone, and volume of the flue duct). Fixed combustion settings in all runs were: fuel feed rate, 0.9 kg h1; primary-combustion zone temperature, 800–850 °C; secondary-combustion zone temperature, 800 °C; primary air flow 110 L min1 and secondary air flow 30 L min1. 2.3. Sampling and analyses Continuous emission measuring (CEM) of O2, H2O, CO2, SO2, NO2, CO, NO, HCl, NH3, N2O, and CH4 was performed during each of the experimental runs (Fig. 1). The PCDD/Fs, PCDTs and PAHs in the flue gas were sampled using the cooled probe polyurethane foam plug (PUFP) sampling technique, according to European standard method EN, 1948:1 (1997). Samples were cleaned-up according to standard methods (Liljelind et al., 2003; Aurell et al., 2009), then PCDD/Fs and PAHs were analyzed as described elsewhere (Aurell et al., 2009). The PCDTs were separated using an Agilent 6890 gas chromatograph equipped with a 60 m J&W Scientific column with an internal diameter of 0.25 mm and a film thickness of 0.25 lm, then detected using a Waters AutoSpec ULTIMA NT 2000D high resolution mass spectrometer. Two of the most intense ions of the molecular ion isotope distribution cluster were monitored for each PCDT homologue. The analysis was time segmented in eight segments, one for each DCDT homologue. For positive identification the CDT congeners had to elute within the expected retention time window (established using a qualitative standard of mono- through octa-CDTs) and have an isotope ratio (of the quantification and confirmation ions) within ±20% of the theoretical. The PCDTs were semiquantified versus 2,3,7,8-tetra-CDT (assuming equal molar response factors) using the internal standard technique (13C12-PCDDs was used as internal standards). All derived concentrations were normalized (norm) to 1 atm, 0 °C, dry gas 11% O2.
3. Results and discussion 3.1. Continuous emission measuring during stable and transient combustion conditions During the stable combustion conditions before and after the transient combustion conditions no CO peaks were detected in any of the experimental runs and the CO levels in the flue gas were low (<5 ppm; Table 1). The SO2 concentrations in the flue gas were below detection limits (3 mg m3) in the runs with no injection of SO2, resulting in a SO2:HCl ratio of virtually zero. In the runs with SO2 gas injection the SO2 concentrations in the flue gas were 80, 340 and 1495 mg m3 (Table 1), resulting in SO2:HCl ratios of 0.2, 0.7 and 2.7 (by mass), respectively. Adding SO2 gas to the combustion zone altered the composition of the flue gases in the postcombustion zone, causing SO2, HCl and N2O levels to be higher, but NO2 levels to be lower than in runs with no SO2 injection (Table 1). The increased HCl levels observed following the addition of sulfur were probably due to the conversion of alkali metal salts (KCl and NaCl) to metal sulfates, with the production of HCl in the presence of water and oxygen according to R1.
1 2MCl þ H2 O þ SO2 þ O2 $ M2 SO4 þ 2HCl ðM ¼ K; NaÞ 2
ðR1Þ
NOx (NO and NO2) and N2O emissions levels have been shown by numerous authors to be affected by the fuel composition. In fluidized bed systems most NO emissions originate from the fuel via oxidation of NH3, and N2O is mainly formed via hydrogen cyanide (HCN), as shown in via R2. The simultaneous reductions and increases in NO2 and N2O emissions, respectively, due to SO2 addition found in the present study are consistent with other studies (DamJohansen et al., 1993), which have suggested that SO2 can decrease the rate of HCN oxidation and thus change the relative amounts of NO2 and N2O emissions.
HCN þ O ! NCO þ H
ðR2Þ
NCO þ NO ! N2 O þ CO
In addition, in experiment 4, with an SO2 level of 1495 mg m3 in the flue gas (SO2:HCl ratio, 2.7), detectable levels of methane (CH4) were found in the flue gas. This is also consistent with earlier studies, that suggested that SO2 inhibits not only the conversion of HCN but also reactions of combustible gases such as CH4 and CO
Table 1 Average values of the continuous emission measurements during the sampling periods (in dry gas).a Experiment
Description
O2 vol %
H2O vol %
CO2 vol %
CO ppm
CH4 ppm
SO2 mg m3
HCl mg m3
NO ppm
NO2 ppm
N2O ppm
NH3 ppm
1 – SO2:HCl 0 1 – SO2:HCl 0 1 – SO2:HCl 0 1 – SO2:HCl 0 1B – SO2:HCl 0 1B – SO2:HCl 0 1B – SO2:HCl 0 2 – SO2:HCl 0.1 2 – SO2:HCl 0.1 2 – SO2:HCl 0.1 2 – SO2:HCl 0.1 3 – SO2:HCl 0.4 3 – SO2:HCl 0.4 3 – SO2:HCl 0.4 3 – SO2:HCl 0.4 4 – SO2:HCl 1.6 4 – SO2:HCl 1.6 4 – SO2:HCl 1.6 4 – SO2:HCl 1.6
Normal combustionb Transient combustion Memory effects 1 Memory effects 2 Transient combustion Memory effects 1 Memory effects 2 Normal combustion Transient combustion Memory effects 1 Memory effects 2 Normal combustionc Transient combustion Memory effects 1 Memory effects 2 Normal combustion Transient combustion Memory effects 1 Memory effects 2
11 10 12 11 8.8 11 10 11 8.7 11 11 11 11 11 10 10 9.4 10 9.4
7.9 7.7 6.5 7.4 9.5 8.0 8.5 8.0 9.3 7.8 8.1 7.6 7.5 7.4 7.9 7.7 8.1 7.7 8.1
9.2 8.9 7.3 8.6 11 9.4 10 9.5 11 9.1 9.5 8.9 8.6 8.5 9.2 9.3 9.5 9.2 9.8
<5 1900 <5 <5 1600 <5 <5 <5 1800 <5 <5 <5 2500 <5 <5 <5 2700 <5 <5
nd 320 nd nd 250 nd nd nd 290 nd nd nd 550 nd nd <5 530 <5 <5
nd 34 nd nd 42 nd nd 81 215 66 52 340 600 365 315 1490 1570 1470 1520
315 240 280 315 315 360 370 400 360 390 430 475 255 460 475 550 400 550 580
470 390 420 460 470 540 530 430 390 430 450 400 330 410 430 410 330 400 410
32 20 31 34 14 33 34 12 3 5 7 10 2 3 5 10 3 4 6
52 28 40 42 29 48 44 69 41 58 64 70 36 61 59 76 41 61 61
nd 34 <5 <5 34 10 <5 nd 39 10 <5 nd 116 10 7 nd 76 9 7
a b c
nd = Not detected, detection limit 1 ppm or 3 mg m3. Mean values of five runs. Mean values of two runs.
770
J. Aurell et al. / Chemosphere 76 (2009) 767–773
Fig. 2. Concentrations of PC1–8DFs (A), PC1–8DDs (B) and PC1–8DTs (C) during normal combustion conditions at three different sampling points P3 (400 °C), P4 (300 °C) and P7 (200 °C) in the post-combustion zone. Cumulative amounts at the respective sampling point, error bars denote 2 SDs.
(Miccio et al., 2001). The reductions in N2O emissions observed during transient combustion conditions compared to stable conditions may be due to the ability of N2O to react with CO, forming CO2 and N2, in the presence of char (a carbonaceous solid product of fuel devolatilization) (Bonn et al., 1995). 3.2. PCDD/F emissions and formation during stable combustion conditions As shown in Fig. 2A and B, with a SO2:HCl ratio of 0 in the flue gas levels of the PCDD/Fs were similar at all three sampling points/ temperatures, i.e. most of the net formation occurred prior to the first sampling point (400 °C). In addition, the SO2:HCl ratio of 0.2 in the flue gas was not sufficient to reduce PCDD/F concentrations (note, data for the sample collected at 200 °C are missing due to losses during its clean-up). In contrast to the results of the experiments with SO2:HCl ratios of 0 and 0.2 in the flue gas, with an SO2:HCl ratio of 2.7 in the flue gas PCDD/F concentrations (especially PCDD concentrations) increased with reductions in temperature and increases in residence time in the post-combustion zone. However, reductions (up to 65%) were observed in both PCDD and PCDF concentrations at the first sampling point P3 (400 °C) with SO2:HCl ratios of 0.7 and 2.7. Although at the last sampling point P7 (200 °C) a significant reduction in PCDF levels was only found with a SO2:HCl ratio of 2.7 in the flue gas. The PCDD/F homologue profiles and congener patterns were found to differ between the experiments with SO2:HCl ratios 60.2 and SO2:HCl ratios P0.7 in the flue gas, implying that adding sulfur to the flue gas results in changes in the formation reactions, particularly for the PCDDs. 3.2.1. PCDDs As shown in Fig. 3, the PCDD homologue profiles at the 400 °C sampling point had lower proportions of the TeCDD to HxCDD homologue groups in the experiments with SO2:HCl ratios P0.7 in the flue gas than in experiments with SO2:HCl ratios 60.2. However, levels of these homologue groups increased with reductions in temperature and increases in residence time with SO2:HCl ratios
P0.7. These changes were found to be largely due to increases in levels of the 1,3,6,8- and 1,3,7,9-TeCDD, 1,2,4,6,8/1,2,4,7,9- and 1,2,3,6,8-PeCDD and 1,2,3,4,6,8-HxCDD congeners, which are known to form from 2,4,6-trichlorophenols, 2,3,4,6-tetrachlorphenols and penta-chlorophenols via condensation reactions (Mulholland and Ryu, 2001). The fraction of the 1,3,6,8-TeCDD congener of the total sum TeCDD isomers (isomer fraction) increased from 13% to 35% from the first sampling point (400 °C) to the last sampling point (200 °C) with the SO2:HCl ratio of 2.7, and from 10% to 18% when no SO2 was added to the flue gas. Formation of these PCDD congeners from PCPhs produced from carbon structures (de novo) has been shown to increase with reductions in temperature from 400 °C to 250 °C (Hell et al., 2001), which may explain the larger amounts of the congeners found at 200 °C than at 400 °C in the present study. In addition, the PCDD levels at 200 °C in experiment 4 with the SO2:HCl ratio of 2.7 in the flue gas may be a result of increased de novo formation due to increased amounts of carbon residuals, as indicated by the inhibited oxidation of CH4, or the presumably lower CuCl2 contents, as shown by Luijk et al. (1994). The results also indicate that the conversion of CuCl2 (and other metal chlorides) to CuSO4 (and other metal sulfates) does not affect the catalysis of condensation reactions. 3.2.2. PCDFs Only slight differences were found in the PCDF homologue profiles between the experiments with SO2:HCl ratios 60.2 and SO2:HCl ratios P0.7 in the flue gas (Fig. 3). The only changes found in the congener patterns were increases in the isomer fraction of 1,2,3,4/2,3,4,9-TeCDF congeners (which co-eluted from the column used in the study) at the last sampling point (200 °C), from 4 ± 3% (2SD) in runs with no added SO2 to 11% in the experimental runs with the SO2:HCl ratio of 2.7 in the flue gas. These congeners have also been shown by Ryu et al. (2005) to be formed from chlorophenols, although they co-eluted with the 2,4,6,7/1,4,6,9-TeCDF congeners from the column used in their study. These congeners may have increased due to the increased CH4 level in the flue gas that appeared with the high SO2:HCl ratio (2.7) in the flue gas, as dis-
J. Aurell et al. / Chemosphere 76 (2009) 767–773
771
Fig. 3. PCDD, PCDT and PCDF homologue profiles in experiments: 1-SO2:HCl 0 (A), 2-SO2:HCl 0.2 (B), 3-SO2:HCl 0.7 (C) and 4-SO2:HCl 2.7 (D) during normal combustion conditions. Error bars denote 1 SD.
cussed above for the PCDDs. The isomer fraction of the 1,2,3,4/ 2,3,4,9-TeCDF congeners was also found to be higher when levels of CuCl2 were low in a study by Luijk et al. (1994). 3.3. PCDTs emissions and formation during stable combustion conditions As shown in Fig. 2C, with an SO2:HCl ratio of 0 in the flue gas PCDT levels were similar at all three sampling points/temperatures, i.e. most of the net formation occurred prior to the first sampling point (400 °C) as for the PCDD/Fs. The PCDT concentrations were significantly higher with an SO2:HCl flue gas ratio of 0.2 than with the SO2:HCl ratio of zero (Fig. 2C), although the SO2:HCl ratio of 0.2 in the flue gas was not sufficient to reduce the PCDD/F concentrations. The observed increase in PCDT levels may have been due to increased sulfonation of the carbon residuals or precursors. Furthermore, as shown in Fig. 2, the PCDT homologue profiles indicated that the degree of chlorination was higher with an SO2:HCl ratio P0.2 than with the ratio of 0, which may also have been due to the amounts of sulfonated backbone structures being lower with no added SO2 in the flue gas. However, the PCDT concentrations were ca. 10-fold lower than the PCDF concentrations, implying that the sulfur analogues were not formed instead of the PCDFs. Furthermore, the PCDT homologue profiles suggest that the PCDTs were formed via similar pathway(s) to the PCDFs, although the increased PCDT levels at 200 °C with an SO2:HCl ratio of 2.7 in the
flue gas suggest that their formation pathway(s) were also similar to those of the PCDDs. 3.4. Effects of SO2 on PCDD/F formation under transient combustion conditions The sum PAH concentrations observed during the sampling periods of the transient combustion phase in experiments 1 and 1B 1–4 (11, 10, 18, 25, and 15 ng PAH m3, respectively) did not correlate with the average CO concentrations in the flue gas in these periods (1900, 1600, 1800, 2500, and 2700 ppm, respectively). A higher CO concentration (2700 ppm), but similar PAH concentration (11 ng PAH m3), was found in the flue gas with the SO2:HCl ratio of 2.7 compared to those obtained with the SO2:HCl ratio of 0 (1900 ppm and 15 ng PAH m3, respectively). The PCDD/F concentrations increased substantially during and after the transient combustion conditions compared to those found during normal combustion conditions (Fig. 4). The trends in the PCDD/F concentrations differed between the replicate runs with the SO2:HCl ratio of zero, possibly due to the longer sampling time (30 min instead of 20 min) in experiment 1B under the transient combustion conditions. However, the PCDD/PCDF ratio was similar in the two runs. As also shown in Fig. 4, with the SO2:HCl ratio of 2.7 in the flue gas there were indications of reduced PCDF levels but increased PCDD levels compared to those observed with the SO2:HCl ratio of 0 in the flue gas under transient combustion con-
772
J. Aurell et al. / Chemosphere 76 (2009) 767–773
Fig. 4. Sum PCDD (A) and sum PCDF (B) concentrations at the sampling points P3 (400 °C), P4 (300 °C) and P7 (200 °C) in the post-combustion zone during the transient combustion period and following memory effect periods (M1 and M2). Cumulative amounts at the respective sampling point.
Table 2 PC4–8DD to PC4–8DF ratio in the collected flue gas samples at 400, 300 and 200 °C under the different combustion periodsa. Combustion period
SO2:HCl 0 400 °C
300 °C
200 °C
400 °C
300 °C
200 °C
400 °C
300 °C
200 °C
400 °C
300 °C
200 °C
Normalb Transient Memory effects 1 Memory effects 2
0.12 0.05 0.09 0.11
0.11 0.06d 0.13d 0.14d
0.11 0.06d 0.13d 0.14d
0.10 0.16 0.28 0.14
0.10 0.21 0.53 0.67
–c 0.20 0.51 0.72
0.11 0.29 0.32 0.14
0.14 0.37 0.83 1.00
0.20 0.41 0.87 1.12
0.17 0.52 0.42 0.21
0.24 0.68 0.90 1.13
0.33 0.77 0.98 1.18
a b c d
SO2:HCl 0.2
SO2:HCl 0.7
SO2:HCl 2.7
PCDD/PCDF in mass ratio. Mean values: of five runs of experiment 1 with SO2:HCl ratio 0 and two runs of experiment 3 with SO2:HCl ratio 0.7. Excluded due to losses during sample clean up. Mean value of two runs.
ditions. In addition, the PCDD concentrations increased with reductions in temperature in the post-combustion zone during the memory effects periods. These changes were most apparent in the PCDD/PCDF ratios obtained, as shown in Table 2, which revealed that the PCDD/PCDF ratio increased with increases in the SO2:HCl ratio in the flue gas not only during the transient combustion conditions but also in all of the other combustion periods. These findings further support the indications in experiment 4, with the SO2:HCl ratio of 2.7, that sulfur has a stronger inhibitory effect on PCDF reaction pathways than on PCDD reactions. The increases in PCDD/PCDF ratios observed during transient combustion conditions are consistent with findings in full-scale plant investigations (Gass et al., 2002; Neuer-Etscheidt et al., 2006). 3.4.1. PCDDs Irrespective of the SO2:HCl ratio in the flue gas and sampling temperature the PCDD homologue profiles under the transient combustion conditions were similar to those found under normal combustion conditions without SO2 addition, thus the transient combustion conditions had a stronger effect on the homologue profile than the SO2 level in the flue gas. Furthermore, the PCDD congeners patterns were also independent of the SO2:HCl ratio in the flue gas and temperature in the post-combustion zone during transient combustion conditions. In all cases the PCDD congener pat-
terns during the transient combustion periods were different from those found under normal combustion conditions, but similar to those found under normal combustion conditions with an SO2:HCl ratio of 1.6 at 200 °C. The TeCDD and PeCDD congeners most strongly affected by transient combustion conditions were the 1,3,6,8- and 1,3,7,9-TeCDD, 1,2,4,6,8/1,2,4,7,9- and 1,2,3,6,8PeCDD congeners, e.g. the isomer fraction of the 1,3,6,8-TeCDD congener was ca. 40% under transient combustion conditions, and ca. 15% under normal combustion conditions. Again, these congeners are known to form from chlorophenols. However, during the memory effect periods the isomer fraction of these congeners increased with both reductions in temperature in the post-combustion zone and with addition of sulfur to the flue gas, e.g. the 1,3,6,8-TeCDD isomer fraction increased from 18% to 48% with the SO2:HCl ratio of 2.7, and from 15% to 28% with the SO2:HCl ratio of 0, between the 400 °C and 200 °C sampling points. These results imply that the formation of condensation reaction products is not only dependent on the temperature but also on the combustion conditions. 3.4.2. PCDF Under the transient combustion conditions the PCDF homologue profiles differed significantly from those observed under normal conditions, the most abundant homologues being MoCDFs, the proportions of which also increased with increases in the SO2:HCl
J. Aurell et al. / Chemosphere 76 (2009) 767–773
ratio in the flue gas. These findings suggest that sulfur inhibits the chlorination of carbon residuals (de novo synthesis) and/or lightlyor non-chlorinated precursors. The PCDF congener patterns were also affected by the transient combustion conditions, and as for the PCDDs the changes in congener patterns were similar at all sampling points/temperatures during transient combustion conditions. However, the isomer fractions of the TeCDF and PeCDF congeners most affected by the transient conditions (1,2,3,4/2,3,4,9-, 2,4,6,7and 2,3,4,6-TeCDF, 2,3,4,6,8- and 2,3,4,6,7-PeCDF) were 40–50% lower with an addition of SO2 to the flue gas than with no addition of SO2. Of these congeners the 1,2,3,4/2,3,4,9- and 2,4,6,7-TeCDF and 2,3,4,6,8-PeCDF have been found to form from chlorophenols by Ryu et al. (2005) (the TeCDF congeners co-eluted with the 1,4,6,9-TeCDF congener and the PeCDF congener co-eluted with 1,2,3,4,7/1,2,3,5,6-TeCDF congeners from the column used in their study) while the 2,4,6,7- and 2,3,4,6-TeCDF, 2,3,4,6,8- and 2,3,4,6,7-PeCDF congeners were found to be the most abundantly formed congeners in de novo synthesis experiment from the PAH perylene (Iino et al., 1999; Weber et al., 2001). In the memory effect periods the PCDF homologue profiles returned to those found during normal combustion conditions. The 1,2,3,4/2,3,4,9- and 2,4,6,7-TeCDF and 2,3,4,6,8-PeCDF were the only TeCDF and PeCDF congeners that increased with reductions in temperature in the post-combustion zone during the memory effect periods in experiments with and without addition of SO2 to the flue gas, indicating that these congeners are products of condensation reactions from chlorophenols in this case. 4. Conclusions The findings indicate that sulfur has a greater inhibitory effect on PCDF formation than on PCDD formation and that the PCDD/ PCDF ratio in the flue gas depends on both the SO2:HCl ratio in the flue gas and the ash deposits that arise from transient combustion conditions. The results also indicate that the relative importance of different pathways shifts in the post-combustion zone; condensation products increasing with reductions in the temperature and increases in residence time. However, these changes were also found to depend on the SO2:HCl ratio in the flue gas and combustion conditions. The presence of sulfur appeared to inhibit the chlorination of PCDFs. A tendency for increased SO2 levels in the flue gas to increase levels of PCDTs was also detected, but the increases were much less significant than the reductions in PCDF levels. Acknowledgments This study was partly financed by PSO (Public Service Obligations) funds from the Danish Ministry for Transport and Energy, Energinet.dk, whose support is gratefully acknowledged. Special thanks are due to Marie Calla for assistance during experiments and laboratory work and Per Liljelind for help with the GC/MS analyses. References Aurell, J., Fick, J., Marklund, S., 2009. Effects of transient combustion conditions on the formation of PCDD/F, PCBz and PAH during MSW incineration. Environ. Eng. Sci. 26, 509–520.
773
Bonn, B., Pelz, G., Baumann, H., 1995. Formation and decomposition of N2O in fluidized bed boilers. Fuel 74, 165–171. Dam-Johansen, K., Åmand, L.-E., Leckner, B., 1993. Influence of SO2 on the NO/N2O chemistry in fluidized bed combustion 2. Interpretation of full-scale observations based on laboratory experiments. Fuel 72, 565–571. European Committee for Standardization, EN 1948:1–3, 1997. Stationary Source Emissions - Determination of the Mass Concentration of PCDDs/PCDFs. Fängmark, I., Strömberg, B., Berge, N., Rappe, C., 1994. The influence of post combustion temperature profiles on the formation of PCDDs, PCDFs, PCBzs and PCBs in a pilot incinerator. Environ. Sci. Technol. 28, 624–629. Gass, C.H., Luder, K., Wilken, M., 2002. PCDD/F-emissions during cold start-up and shut-down of a municipal waste incinerator. Organohal. Compd. 56, 193– 196. Griffin, R.D., 1986. A new theory of dioxin formation in municipal solid waste combustion. Chemosphere 15, 1987–1990. Gullett, B.K., Bruce, K.R., Beach, L.O., 1990. Formation of chlorinated organics during solid waste combustion. Waste Manage. Res. 8, 203–214. Gullett, B.K., Bruce, K.R., Beach, L.O., 1992. Effect of sulfur dioxide on the formation mechanism of polychlorinated dibenzodioxin and dibenzofuran in municipal waste combustors. Environ. Sci. Technol. 26, 1938–1943. Hell, K., Stieglitz, L., Dinjus, E., 2001. Mechanistic aspects of the de-novo synthesis of PCDD/PCDF on model mixtures and MSWI fly ashes using amorphous 12C- and 13C-labeled carbon. Environ. Sci. Technol. 35, 3892–3898. Hunsinger, H., Seifert, H., Jay, K., 2007. Reduction of PCDD/F formation in MSWI by a process-integrated SO2 cycle. Environ. Eng. Sci. 24, 1145–1159. Iino, F., Imagawa, T., Takeuchi, M., Sadakata, M., 1999. De novo synthesis mechanism of polychlorinated dibenzofurans from polycyclic aromatic hydrocarbons and the characteristic isomers of polychlorinated naphthalenes. Environ. Sci. Technol. 33, 1038–1043. Kopponen, P., Sinkkonen, S., Poso, A., Gynther, J., Karenlampi, S., 1994. Sulfur analogs of polychlorinated dibenzo-p-dioxins, dibenzofurans and diphenyl ethers as inducers of CYP1A1 in mouse heptatoma-cell culture and structure– activity-relationships. Environ. Toxicol. Chem. 13, 1543–1548. Liljelind, P., Söderström, G., Hedman, B., Karlsson, S., Lundin, L., Marklund, S., 2003. Method for multiresidue determination of halogenated aromatics and PAHs in combustion-related samples. Environ. Sci. Technol. 37, 3680–3686. Luijk, R., Akkerman, D.M., Slot, P., Olie, K., Kapteijn, F., 1994. Mechanism of formation of polychlorinated dibenzo-p-dioxins and dibenzofurans in the catalyzed combustion of carbon. Environ. Sci. Technol. 28, 312–321. Miccio, F., Löffler, G., Wargadalam, V.J., Winter, F., 2001. The influence of SO2 level and operating conditions on NOx and N2O emissions during fluidised bed combustion of coals. Fuel 80, 1555–1566. Mulholland, J.A., Ryu, J.-Y., 2001. Formation of polychlorinated dibenzo-p-dioxins by CuCl2-catalyzed condensation of 2,6 chlorinated phenols. Combust. Sci. Technol. 169, 107–126. Neuer-Etscheidt, K., Nordsieck, H.O., Liu, Y., Kettrup, A., Zimmermann, R., 2006. PCDD/F and other micropollutants in MSWI crude gas and ashes during plant start-up and shut-down processes. Environ. Sci. Technol. 40, 342– 349. Ogawa, H., Orita, N., Horaguchi, M., Suzuki, T., Okada, M., Yasuda, S., 1996. Dioxin reduction by sulfur component addition. Chemosphere 32, 151–157. Raghunathan, K., Gullett, B.K., 1996. Role of sulfur in reducing PCDD and PCDF formation. Environ. Sci. Technol. 30, 1827–1834. Ryan, P.S., Li, X.-D., Gullett, B.K., Lee, C.W., Clayton, M., Touati, A., 2006. Experimental study on the effect of SO2 on PCDD/F emissions: determination of the importance of gas-phase versus solid-phase reactions in PCDD/F formation. Environ. Sci. Technol. 40, 7040–7047. Ryu, J.-Y., Mulholland, J.A., Takeuchi, M., Kim, D.H., Hatanaka, T., 2005. CuCl2catalyzed PCDD/F formation and congener patterns from phenols. Chemosphere 61, 1312–1326. Sidhu, S.S., Maqsud, L., Dellinger, B., 1995. The homogeneous, gas-phase formation of chlorinated and brominated dibenzo-p-dioxin from 2,4,6-trichloro- and 2,4,6-tribromophenols. Combust. Flame 100, 11–20. Weber, R., Iino, F., Imagwa, T., Takeuchi, M., Sakurai, T., Sadakata, M., 2001. Formation of PCDF, PCDD, PCB, and PCN in de novo synthesis from PAH: Mechanistic aspects and correlation to fluidized bed incinerators. Chemosphere 44, 1429–1438. Wikström, E., Andersson, P., Marklund, S., 1998. Design of a laboratory scale fluidized bed reactor. Rev. Sci. Instrum. 69, 1850–1859. Wikström, E., Ryan, S., Touati, A., Gullett, B.K., 2003a. Key parameters for de novo formation of polychlorinated dibenzo-p-dioxins and dibenzofurans. Environ. Sci. Technol. 37, 1962–1970. Wikström, E., Ryan, S., Touati, A., Telfer, M., Tabor, D., Gullett, B.K., 2003b. Importance of chlorine speciation on de novo formation of polychlorinated dibenzo-p-dioxins and polychlorinated dibenzofurans. Environ. Sci. Technol. 37, 1108–1113.