Environ. Sci. Technol. 2010, 44, 1974–1979

Chlorinated Aromatic Compounds in a Thermal Process Promoted by Oxychlorination of Ferric Chloride TAKASHI FUJIMORI,* MASAKI TAKAOKA, AND SHINSUKE MORISAWA Department of Urban and Environmental Engineering, Graduate School of Engineering, Kyoto University, Katsura, Nisikyo-ku, 615-8540, Kyoto, Japan

Received November 3, 2009. Revised manuscript received February 3, 2010. Accepted February 9, 2010.

The relationship between the formation of chlorinated aromatic (aromatic-Cl) compounds and ferric chloride in the solid phase during a thermal process motivated us to study the chemical characteristics of iron in a model solid sample, a mixture of FeCl3 · 6H2O, activated carbon, and boron nitride, with increasing temperature. Fe K-edge extended X-ray absorption fine structure (EXAFS) spectroscopy revealed drastic changes in the chemical form of amorphous iron, consistent with other analytical methods, such as X-ray diffraction using synchrotron radiation (SR-XRD) and Fourier-transform infrared (FT-IR) spectroscopy. Atomic-scale evidence of the chlorination of aromatic carbon was detected by Cl-K X-ray absorption near edge structure (XANES) spectroscopy. These results showed the thermal formation mechanism of aromatic-Cl compounds in the solid phase with ferric chloride. We attribute the formation of aromatic-Cl compounds to the chlorination of carbon, based on the oxychlorination reaction of FeCl3 at temperatures in excess of ca. 300 °C, when the carbon matrix is activated by carbon gasification, catalyzed by Fe2O3, and surface oxygen complexes (SOC) generated by a catalytic cycle of FeCl2 and FeOCl. Chemical changes of trace iron in a thermal process may offer the potential to generate aromatic-Cl compounds in the solid phase.

Introduction Thermal processes are well-known as major anthropogenic sources of chlorinated aromatic (aromatic-Cl) compounds (1, 2), such as polychlorinated dibenzo-p-dioxins (PCDDs) and -furans (PCDFs), biphenyls (PCBs), and chlorobenzenes (CBzs). Aromatic-Cl compounds are known to be emitted from municipal solid waste incinerators (MSWI) (3, 4) and iron ore sintering processes (5, 6). Many researchers have suggested the homo- and heterogeneous formation of aromatic-Cl compounds from macromolecular carbons (7-9). Fly ash collected from the postcombustion zone of such as MSWIs and iron ore sintering plants has the highest concentration of aromatic-Cl compounds, and unburned carbon and chlorine sources in fly ash and surrounding oxygen are known to be essential factors for aromatic-Cl formation (7, 8). Trace metal chlorides in fly ash promote aromatic-Cl formation (10). Although copper chlorides are the most wellknown and studied promoters of aromatic-Cl compounds * Corresponding author e-mail: [email protected]. media.kyoto-u.ac.jp. 1974

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 6, 2010

(11, 12), iron chlorides also have strong potential, and a chlorination mechanism of carbon has been proposed (10, 13-15). Additionally, the amounts of iron in MSWI fly ash have been reported to be much greater than those of copper (15), and, clearly, iron is the main component of the iron ore sintering process (16). Thus, iron is thought to greatly contribute to the formation of aromatic-Cl compounds in fly ash. However, little direct evidence exists regarding the formation mechanism of aromatic-Cl compounds with iron chlorides at the atomic level. X-ray absorption fine structure (XAFS) spectroscopy has recently been used to monitor the redox change of copper (17, 18). The behavior of chlorine and the atomic environment of copper in fly ash were clarified in a previous study using XAFS (18). The chlorination mechanism of carbon by iron chlorides may be better described if the behaviors of iron and chlorine atoms are observed at the same time. In this study, we discuss the behaviors of Fe and Cl in fly ash and, based primarily on XAFS spectroscopy, provide basic information on the chlorination mechanism of carbon by iron chloride that produces aromatic-Cl compounds during thermal processes. Chemical forms of iron at various temperatures were analyzed by Fe K-edge extended X-ray absorption fine structure (EXAFS) spectroscopy. The Fe K-edge EXAFS technique is suitable for amorphous iron characterization, as explained in the present study. We also used Cl-K X-ray absorption near edge structure (XANES) spectroscopy to determine the behavior of chlorine.

Materials and Methods Model Fly Ash. To determine the behavior of Fe and Cl, we prepared a model fly ash (MFA) using a mixture of ferric chloride, activated carbon (AC), and boron nitride (BN). Any organic compounds were removed from the AC by heating at 500 °C for 60 min under a 100% nitrogen stream. Preparation of Ferric Oxychloride (FeOCl). We made FeOCl from ferric chloride hexahydrate, according to Ryan and Altwicker (15). A quartz boat containing FeCl3 · 6H2O (3-4 g) was placed in the center of a quartz tube. The FeCl3 · 6H2O was then heated to 250 °C for about 30 min under a 50mL/min flow of nitrogen. The residue (dark black/red) in the quartz boat was rinsed with deionized water, followed by acetone, on filter paper three times. The product was dried in vacuo and identified as FeOCl by X-ray diffraction (Figure S1) and Fe K-edge EXAFS, by comparison with literature data (19). In Situ Fe EXAFS and Data Analysis. Using in situ Fe K-edge EXAFS spectroscopy, we detected the chemical structure of Fe at the atomic level in the MFA, which contained ferric chloride hexahydrate (FeCl3 · 6H2O), activated carbon (AC), and boron nitride (BN). The composition was 1.5% Fe, 2.9% Cl, and 10% AC, and the remaining 82.7% was almost entirely BN. After the MFA was ground, using a mortar and then an agate mortar for 10 min each, it was pressed into a disk. Fe K-edge EXAFS spectroscopy was performed using beamline BL01B1 at SPring-8 (Hyogo, Japan), with the MFA disk heated in a T-type in situ cell (17, 18) consisting of a glass cell, a mantle heater, and a temperature controller. The temperature of the sample was increased gradually from room temperature to 450 °C at a rate of around 5 °C/min, following the profile in Figure S2. A 10% O2 (90% N2) gas atmosphere was introduced from the inlet of the T-type cell at 50 mL/ min and exhausted from the outlet. The energy area from 6600 to 8700 eV of Fe K-edge EXAFS spectra could be measured in 2 min in quick-scan mode. EXAFS spectra of the MFA disk were collected in transmission mode with a 10.1021/es903337d

 2010 American Chemical Society

Published on Web 02/19/2010

Si(111) monochromator. The spectra of reference materials, FeCl3, FeCl3 · 6H2O, FeCl2 · 4H2O, FeOCl, FeO, Fe2O3, Fe3O4, FeO(OH), Fe3C, and Fe, were measured to compare their spectral shapes and to identify major species, because an EXAFS spectrum can be used as a fingerprint reflecting the local environment of the iron. More reference irons should be studied to characterize iron forms in detail. However, number and species of iron compounds became restricted in case of the MFA. As the MFA and gas phase included five elements (Fe, Cl, O, H, and C), we selected chlorides, oxides, hydroxide, carbide, and metal of iron. These iron compounds express possible chemical forms of iron at each temperature. Species can be distinguished using the least-squares linear combination fit (LSF) technique, in which spectra of known reference species are fitted to the spectrum of an unknown sample. We used the LSF technique on the k3-weighted EXAFS spectrum to determine the major species using the SixPACK software (20) (ver. 0.63). The residual value reducedχ2 )

1 N-P

N

∑ (χ

obs i

- χifit)2

(1)

i)1

was used to evaluate the LSF for the experimental spectra (21). χiobs is the ordinate of the EXAFS spectrum measured from the sample at the ith energy point, χifit is the ordinate of the fitted EXAFS spectrum, N is the number of data points in the fitted wavenumber k range, and P is the number of fitted components. Principal components analysis (PCA) was used to determine the number and type of principal components in k-range 3-12 Å-1, and target transformation was then employed to identify the probable iron species in MFA during increasing temperature for the set of 10 reference compounds. Details of PCA and target transformation and fitting of EXAFS spectra were from Ressler et al. (22) and Manceau et al. (23), respectively. Cl K-Edge XANES. The Cl forms present after the MFA was heated were determined by measuring the Cl K-edge XANES spectra. A MFA contained FeCl3 (2% Fe and 3.8% Cl), AC (5%), and BN (remainder) and was ground using a mortar for 10 min. We then placed the MFA powder on a quartz boat in a quartz tube filled with 10% O2 (90% N2) at 50 mL/min and heated for 30 min in an electric furnace preheated to 200, 300, and 400 °C. After the heating procedure, MFA powder was sealed as quickly as possible and sent for the measurement of Cl K-edge XANES spectra, which was performed using BL-11B in the Photon Factory (Tsukuba, Japan). An in situ cell was not used in the Photon Factory because of physical restrictions of the device. Powdered MFA samples were mounted on carbon tape, and their XANES spectra were collected in total fluorescence yield (TFY) mode in a vacuum. X-ray absorption spectra of Cl in different inorganic and organic reference compounds were collected to assist in the identification of the chemical state of Cl in MFA after heating (Figure S3). Polyvinyl chloride and the reference iron compounds (FeCl3 · 6H2O, FeCl3, FeCl2 · 4H2O, and FeOCl) were measured in total electron yield (TEY) mode in a vacuum. Cl K-edge XANES spectra of chlorobenzenes and chlorophenols were measured under atmospheric pressure by the conversion electron yield (CEY) method at BL-9A at the Photon Factory. Cl bound to inorganic, aromatic, or aliphatic carbon can be distinguished by the features of a Cl XANES spectrum, as reported previously (18, 24). As reflected in these spectral features, analyses were performed by a linear combination fit using reference materials of chlorine and REX 2000 software (ver. 2.5.5; Rigaku, Japan). In Situ Powdery X-ray Diffraction Using Synchrotron Radiation. Trace crystal structures in a MFA were determined by in situ powder X-ray diffraction using synchrotron radiation (SR-XRD). FeCl3 · 6H2O (containing 1% Fe and 2% Cl), 5% AC, and BN (remainder) in a MFA was added to a

FIGURE 1. Contour plot of the k3-weighted Fe K-edge EXAFS spectra (A) and the percentage of iron in different forms, calculated by linear square fitting (LSF) of Fe K-edge EXAFS spectra (B), at each temperature. quartz capillary column (0.5 mm in diameter) using a Pasteur pipet and sealed using a burner under an air atmosphere. We placed the capillary in a Debye-Scherrer camera for measurement of SR-XRD using BL02B2 in SPring-8 (Hyogo, Japan). The capillary was heated in N2 gas from room temperature to 500 °C, and we measured the SR-XRD pattern, from which crystal information was identified using the MDI Jade 6j software (Rigaku, Japan) contained within the International Centre for Diffraction Data powder diffraction file. FT-IR Spectroscopy. The carbon surface of and the chemical form of the iron in a disk of MFA were analyzed by Fourier transform infrared (FT-IR) spectroscopy (FT-IR8400, Shimadzu Co., Ltd.). The MFA contained ferric chloride hexahydrate and AC (1:1 weight) and was heated to 200, 300, and 400 °C for 30 min under 10% oxygen gas stream and 99.5% KBr. Total Organic Carbon. Total organic carbon (TOC) was measured as described previously (18).

Results and Discussion Chemical Form of Iron at Each Temperature. Analyzing the Fe K-edge EXAFS data set, we found evidence of dechlorination and oxidation processes of iron in the MFA upon heating, related to the formation of aromatic-Cl compounds in a thermal process, as discussed below. We selected the EXAFS region to analyze and characterize the chemical form of iron. The amplitude and interval of the oscillation structure of the Fe-EXAFS region showed more different oscillation features at each temperature than the XANES region in the normalized XAFS spectrum (see Figure S4). The k3-weighted Fe K-edge EXAFS spectra changed dramatically upon heating the MFA (Figure 1A), indicating dynamic change in the chemical environment around the Fe atom. We studied the chemical form of the iron by means of LSF analyses of the Fe K-edge EXAFS spectra. Although the Debye-Waller factor is known to affect the amplitude of EXAFS oscillations with temperature rising, peak positions and intervals of EXAFS oscillation which mainly characterize the iron compounds are not affected by the Debye-Waller factor. So, we concluded that analysis by using EXAFS oscillation is a useful method to characterize the iron in case of rising temperature. About six significant components were extracted from PCA of the 35 Fe K-edge EXAFS spectra between 3 and 12 Å-1. We determined the number of components using the minimum value of indicator (IND), defined by Malinowski (25), using SixPACK software (20). VOL. 44, NO. 6, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1975

FIGURE 2. The k3-weighted Fe K-edge spectra at room temperature (rt), ca. 300 °C, and 400 °C, and the spectra of six forms of iron. After PCA, probable iron species were identified by target transformation of 10 reference compounds as described in the Materials and Methods. Six model compounds were found to yield sufficient matches: specifically, four iron chlorides (FeCl3, FeCl3 · 6H2O, FeCl2 · 4H2O, and FeOCl), ferric oxide (Fe2O3), and ferric oxide hydroxide [FeO(OH)]. These iron references were adopted using a combination of SPOIL values, as defined in ref 23, the reducedχ2 value, and similarities in the EXAFS spectra between target and transformation, with help from the report of Slowey et al. (21). When the SPOIL value is <3.0, target transformation is considered “good”. FeO, Fe3O4, and Fe3C had larger SPOIL (>3.0) and reducedχ2 values (Table S1) and showed different EXFAS shapes (Figure S5). Fe had the largest reducedχ2 value, as shown in Table S1. Thus, we deselected Fe, FeO, Fe3O4, and Fe3C. On the other hand, the six selected iron references showed lower SPOIL (<3.0, except FeCl2 · 4H2O) and reducedχ2 values (except Fe2O3; Table S1) and a good fit between target and transformed spectra (Figure S6). Figure 1B shows the change in chemical form of iron with temperature, as determined by LSF of Fe K-edge EXAFS at each temperature. The average and standard deviation of reducedχ2 (n ) 35) was 0.79 ( 0.53. The chemical form of ferric chloride hexahydrate in the MFA at room temperature (rt) was not changed by the prephysical procedure of mixing in a mortar. The spectrum shape at rt was almost the same as that of FeCl3 · 6H2O, as shown in Figure 2. The ratio of FeCl3 · 6H2O decreased linearly from ca. 180 to 380 °C. In contrast, the ratios of ferrous chloride (FeCl2 · 4H2O) and ferric oxychloride (FeOCl) increased. This reflects the damping of the amplitude over ca. 7 Å-1 of Fe K-edge EXAFS, because FeCl2 · 4H2O and FeOCl have smaller amplitudes than ferric chloride (Figure 2). Here, water (H2O) bonded with iron chlorides is thought to vanish with rising temperature. However, LSF results showed the existence of iron chloride hydrates at high temperatures. The Fe K-edge EXAFS spectra of ferric chloride hexahydrate and nonhydrate had quite similar shapes (Figures 2, 5B). It is important to note that Fe K-edge EXAFS spectra are not sensitive to hydration water. The sensitivity of the Fe-EXAFS amplitude derives from surrounding the chlorine atom (Cl). The LSF method provides 1976

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 6, 2010

FIGURE 3. Thermal change of SR-XRD spectra (A) and two-peaks height (B) during heating of the MFA. Normalized height of main two peaks, labeled a and b, was calculated by the peak divided by the strongest peak a at 500 °C.

FIGURE 4. Thermal changes in the FT-IR absorbance. Peaks at ca. 1720 (a) and 1570 cm-1 (b) are surface oxygen complexes (SOCs), such as lactones and diketones. The two peaks at 700-400 cm-1 (c) indicate Fe2O3. fairly accurate information on the valence of iron and the coordination number of Cl atoms in the iron chlorides. Thus, we will not discuss the “hydrate” of iron chlorides. Some FeO(OH) was formed at temperatures over ca. 300 °C (Figure 1B). For example, an EXAFS spectrum at ca. 400 °C had a characteristic vibration pattern between 7 and 10 Å-1 (see asterisks in Figures 1A and 2). This doublet pattern was seen in three iron reference compounds: FeOCl, Fe2O3, and FeO(OH). From the analytical result of LSF [50% Fe(II) chloride, 30% Fe(III) oxychloride, and 20% FeO(OH)], ferric oxide was not detected, because the amplitude of Fe2O3 was larger than that of the two other reference compounds. However, FeO(OH) can be expressed by another chemical formula, Fe2O3 · H2O, indicating the hydrate of the ferric oxide. Thus, the oxidation state of FeO(OH) is not different from that of Fe2O3. As a supporting experiment to examine the chemical form of iron, we also performed SR-XRD. Figure 3A shows the change in diffraction pattern on heating the MFA. Although amorphous iron chlorides were not seen, SR-XRD detected the crystal structure of the Fe2O3 above 250-300

FIGURE 5. Thermal dynamic change in Fourier-transformed Fe K-edge EXAFS spectra (A) and representative spectra at three temperatures [room temperature (rt), ca. 300 °C, and 400 °C], with spectra of six forms of iron (B). Fe-O, Fe-Cl, and Fe-(O)-Fe bonds were derived from the comparison and analysis of six reference spectra.

FIGURE 6. Magnitudes of changes in Fe-Cl and Fe-(O)-Fe in Fourier-transformed Fe K-edge EXAFS. °C. We also measured SR-XRD patterns of only powdery AC at room temperature and 300 °C. Then, we confirmed the broadenings of the diffraction peaks from 8° to 20° (peak at about 15°). There was not a difference in two SR-XRD patterns at room temperature and 300 °C. The same broadenings are also shown in the SR-XRD patterns in case of the MFA (Figure 3A) at each temperature. Therefore, these broadenings were thought to be derived from amorphous of carbon matrix in the AC. Two major acute peaks, labeled “a” and “b” in Figure 3A, increased in intensity with increasing temperature (Figure 3B). Thus, the ratio of Fe2O3 to total iron apparently increased on heating. Bands of Fe2O3 at 700-400 cm-1 (26) were also detected by FT-IR spectroscopy in the temperature range of 200-300 °C, and its absorbance increased with temperature (see Figure 4). These results, from EXAFS, SR-XRD, and FTIR, suggest that oxidation of iron, such as Fe2O3, in MFA started at ca. 300 °C. Bonding with iron atoms was examined by Fouriertransform EXAFS spectra, as shown in Figure 5. The linear decline at ca. 180-380 °C of Fe(III) chloride represents dechlorination of the iron atom, with drastic changes in the Fe-Cl bond. The average value of the Fourier-transformed magnitude derived from the Fe-Cl bond during R + ∆R ) 1.72 -1.90 Å (∆R is the differential of the radial coordinate) decreased linearly from ca. 60 to 400 °C (Figure 6). The transformed EXAFS spectra included a doublet peak structure at 7-10 Å-1 of EXAFS oscillation, indicating the environment surrounding the Fe atom (see Figures 2 and 5B). The Fouriertransformed magnitude of the Fe-(O)-Fe bond (R+∆R: 2.61-2.79 Å) gradually increased above ca. 300 °C, at which temperature the growth of iron oxides, such as Fe2O3, started. Carbon Chlorinated by Oxychlorination of Ferric Chloride. Although chlorine bonded only with iron atoms at room temperature in the MFA, the interaction of chlorine atoms with the carbon matrix after the dechlorination process can be examined using the Cl-K XANES technique. Figure 7 shows the Cl-K XANES spectra at each temperature. The specific acute peak at ca. 2821 eV of chlorine connected with

FIGURE 7. Cl-K edge XANES spectra at room temperature (rt), 200 °C, 300 °C, and 400 °C, indicated by the bold solid line. Dotted circles indicate calculated spectra from combined Cl references (thin solid line). Red and blue thin lines indicate aromatic- and aliphatic-Cl, respectively. aromatic carbon (18, 24) was confirmed visually from the Cl-XANES shape at 400 °C. A linear combination fit of Cl-K XANES showed a rise in the aromatic-Cl ratio, with a maximum at 400 °C (aromatic-Cl, ca. 30%). The aromatic-Cl ratio increased with temperature over 200-300 °C, when the chlorine-source ferric chloride decreased its ratio of iron and dechlorination progressed (Figures 1B and 6). Moreover, ferric oxide appeared and increased its ratio of iron at temperatures over ca. 300 °C (Figures 1B, 3, and 4). Thus, oxychlorination of ferric chloride primarily occurred to generate chlorinated aromatic compounds over ca. 300 °C, by the following proposed reaction 4FeCl3 + 3O2 + 6A-H f 2Fe2O3 + 6A-Cl + 6HCl (R1) where A-H and A-Cl are aromatic and chlorinated aromatic compounds, respectively. Ryan and Altwicker reported that concentrations of PCDDs and PCDFs in model samples (such as our MFA sample) increased from ca. 300 to 400 °C, and the PCDD/F yield at 400 °C was three times larger than that at 300 °C (15). Measuring the amount of PCBs and CBzs in MFA at 300 and 400 °C, we found the same tendency, the yield at 400 °C was greater than that at 300 °C. The ratio of aromatic-Cl showed the same increasing pattern, as did those of the following chlorinated aromatic compounds: PCDDs, PCDFs, PCBs, and CBzs. Thus, we suggest that the oxychlorination reaction R1 of ferric chloride is a key mechanism contributing to the formation of chlorinated aromatics in the solid phase of a thermal process. Catalytic Cycle of Iron Chlorides Promotes Surface Oxygen Complexes. A decrease in ferric chloride and an increase of ferrous chloride and ferric oxychloride in the temperature range of ca. 180-300 °C (Figure 1B) indicate the following two reactions: VOL. 44, NO. 6, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1977

2FeCl3 f 2FeCl2 + Cl2

(R2)

2FeCl3 + O2 f 2FeOCl + 2Cl2

(R3)

Reaction R2 indicates the dechlorination of ferric chloride, which has a negative Gibbs free energy change (∆G) over 150 °C at thermodynamic equilibrium, as calculated by FactSage software. Oxychlorination of ferric chloride is described in R3. Although chlorine is generated from ferric chloride via reactions R2 and R3, it has been reported that few aromaticCl compounds, such as PCDDs and PCDFs, are generated under ca. 300 °C (15). Regarding surface oxygen complexes (SOCs), we attempt to explain this conflict in which chlorine was hardly consumed by carbon. Figure 4 shows the FT-IR spectrum of MFA at each temperature. The absorbance of SOC bands (27) at 1720 and 1570 cm-1 grew from 200 to 300 °C. SOC formation is simply described as C + 1/2O2 f C-Os

(R4)

where C-Os indicates the SOC. No SOC band appeared in cases of AC alone. Thus, iron chlorides are related to SOC growth. Between 200 and 300 °C, reactions R2 and R3 progressed and Cl2 was generated. Here, we suggest that SOC formation R4 was catalyzed by the following reactions of iron chlorides, as spillover effects (27): 1

1

FeCl2 + /2O2 f FeOCl + /2Cl2

(R5)

FeOCl + 1/2Cl2 + C f FeCl2 + C-Os

(R6)

This catalytic cycle, via R5 and R6, is important. First, FeCl2 and FeOCl show the same mole amount in each reaction. This is supported by the Fe-EXAFS analyses, where the ratio of FeCl2 was almost the same as the ratio of FeOCl over the 200-300 °C range (Figure 1B). Second, Cl2 derived from FeCl3 via R2 or R3 is used in the catalytic cycle, which explains the absence of the chlorination of carbon. Furthermore, absorption bands of SOC increased from 300 to 400 °C, as shown in Figure 4. The ratio of FeCl2 and FeOCl did not change greatly in this temperature range (Figure 1B). Another reason for the acceleration of SOC formation is thought to relate to the oxidation process of iron, starting at ca. 300 °C (Figures 1B, 3, and 4). Metal oxides, including Fe2O3, catalyze carbon gasification, as reported by McKee (28). When ferric chloride hexahydrate was added to AC (1:1 weight), organic carbon began to be consumed, starting at 300 °C, according to a TOC analyzer. We conclude that this consumption of carbon was catalyzed by Fe2O3, which is generated at temperatures in excess of 300 °C. Then, reaction R4 may proceed, because of the loss of C-Os on the right side of the R4 reaction. Alternatively, increasingly strong carbon reactivity at the edge of the carbon matrix, such as a graphite sheet, by carbon gasification, may lead to high levels of SOCs in MFA at 400 °C. Aromatic-Cl Compounds Derived from Iron Chlorides. Here we present a summary of the formation mechanism of aromatic-Cl compounds by ferric chloride. In the solid phase, FeCl3 changes its chemical form to FeCl2, FeOCl, and Fe2O3 with increasing temperature. These iron chlorides and oxychloride interact with the solid-phase carbon matrix and oxygen in the gas stream to derive chlorine. Especially, at temperatures over 300 to 400 °C, at which aromatic-Cl compound generation exponentially increases, (i) chlorination of the carbon matrix by oxychlorination (R1) of ferric chloride, (ii) carbon activated from its gasification, catalyzed by ferric oxide, and (iii) SOC formation, by the catalytic cycle of FeCl2 in (R5) and FeOCl in (R6), occur together and strongly affect the formation of chlorinated aromatic compounds in the solid phase. The relationship between Fe2O3 generation 1978

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 6, 2010

and aromatic-Cl compound formation at temperatures in excess of 300 °C (as described above) suggests the importance of the increase in the activated edge of the carbon matrix, such as in a graphite sheet, by carbon gasification, catalyzed by Fe2O3. In the thermal-solid phase, the mixture of carbon and trace iron might have a serious potential to generate harmful aromatic-Cl compounds in the environment. We need to rethink and assess trace iron in solid phases, such as fly ash, in a thermal process. Because the analytical methodology of Fe K-edge EXAFS spectroscopy is well-suited to the characterization of amorphous iron compounds, this technique will allow for future in depth scientific studies of iron in environmental and artificial processes.

Acknowledgments We thank N. Takeda, K. Oshita, K. Shiota, T. Yamamoto, and Y. Tanino for supporting this study; H. Tanida and T. Uruga (BL01B1) (Proposal No. 2007A1798) and K. Kato (BL02B2) (Nos. 2006A1093 and 2007B1687) for helping with Fe XAFS and SR-XRD measurement at SPring-8; and Y. Kitajima (BL11B) and Y. Inada (BL-9A) for helping with Cl XANES measurement at Photon Factory (Proposal No. 2007G069).

Supporting Information Available One supporting table and six supporting figures are available. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Bumb, R. R.; Crummett, W. B.; Cutie, S. S.; Gledhill, J. R.; Hummel, R. H.; Kagel, R. O.; Lamparski, L. L.; Luoma, E. V.; Miller, D. L.; Nestrick, T. J.; Shadoff, L. A.; Stehl, R. H.; Woods, J. S. Trace chemistries of fire: a source of chlorinated dioxins. Science 1980, 210, 385–390. (2) Czuzuwa, J. M.; Hites, R. A. Environmental fate of combustiongenerated polychlorinated dioxins and furans. Environ. Sci. Technol. 1984, 18, 444–450. (3) Olie, K.; Vermeulen, P. L.; Hutzinger, O. Chlorodibenzo-p-dioxins and chlorodibenzofurans are trace components of fly ash and flue gas of some municipal incinerators in The Netherlands. Chemosphere 1977, 6, 455–459. (4) Ballschmiter, K.; Braunmiller, I.; Niemczyk, R.; Swerev, M. Reaction pathways for the formation of polychloro-dibenzodioxins (PCDD) and sdibenzofurans (PCDF) in combustion processes: II. Chlorobenzenes and chlorophenols as precursors in the formation of polychloro-dibenzodioxins and sdibenzofurans in flame chemistry. Chemosphere 1988, 17, 995–1005. (5) Anderson, D.; Fisher, R. Sources of dioxins in the United Kingdom: the steel industry and other sources. Chemosphere 2002, 46, 371–381. (6) Cieplik, M. K.; Carbonell, J. P.; Mun ˜ oz, C.; Baker, S.; Kru ¨ ger, S.; Liljelind, P.; Marklund, S.; Louw, R. On dioxin formation in iron ore sintering. Environ. Sci. Technol. 2003, 37, 3323–3331. (7) Addink, R.; Olie, K. Mechanisms of formation and destruction of polychlorinated dibenzo-p-dioxins and dibenzofurans in heterogeneous systems. Environ. Sci. Technol. 1995, 29, 1425– 1435. (8) Tuppurainen, K.; Halonen, I.; Ruokoja¨rvi, P.; Tarhanen, J.; Ruuskanen, J. Formation of PCDDs and PCDFs in municipal waste incineration and its inhibition mechanisms: A review. Chemosphere 1998, 36, 1493–1511. (9) Huang, H.; Buekens, A. On the mechanisms of dioxin formation in combustion processes. Chemosphere 1995, 31, 4099–4117. (10) Fujimori, T.; Takaoka, M.; Takeda, N. Influence of Cu, Fe, Pb and Zn chlorides and oxides on formation of chlorinated aromatic compounds in MSWI fly ash. Environ. Sci. Technol. 2009, 43, 8053–8059. (11) Luijk, R.; Akkerman, D. M.; Slot, P.; Olie, K.; Kapteijn, F. Mechanism of formation of polychlorinated dibenzo-p-dioxins and dibenzofurans in the catalyzed combustion of carbon. Environ. Sci. Technol. 1994, 28, 312–321. (12) Weber, P.; Dinjus, E.; Stieglitz, L. The role of copper(II) chloride in the formation of organic chlorine in fly ash. Chemosphere 2001, 42, 579–582.

(13) Hoffman, R. V.; Eiceman, G. A.; Long, Y.-T.; Collins, M. C.; Lu, M.-Q. Mechanism of chlorination of aromatic compounds adsorbed on the surface on fly ash from municipal incinerators. Environ. Sci. Technol. 1990, 24, 1635–1641. (14) Stieglitz, L.; Vogg, H.; Zwick, G.; Beck, J.; Bautz, H. On formation conditions of organohalogen compounds from particulate carbon of fly ash. Chemosphere 1991, 23, 1255–1264. (15) Ryan, S. P.; Altwicker, E. R. Understanding the role of iron chlorides in the de novo synthesis of polychlorinated dibenzep-dioxins/dibenzofurans. Environ. Sci. Technol. 2004, 38, 1708– 1717. (16) Xhrouet, C.; Pirard, C.; De Pauw, E. De novo synthesis of polychlorinated dibenzo-p-dioxins and dibenzofurans an fly ash from a sintering process. Environ. Sci. Technol. 2001, 35, 1616–1623. (17) Takaoka, M.; Shiono, A.; Nishimura, K.; Yamamoto, T.; Uruga, T.; Takeda, N.; Tanaka, T.; Oshita, K.; Matsumoto, T.; Harada, H. Dynamic change of copper in fly ash during de novo synthesis of dioxins. Environ. Sci. Technol. 2005, 39, 5878– 5884. (18) Fujimori, T.; Takaoka, M. Direct chlorination of carbon by copper chloride in a thermal process. Environ. Sci. Technol. 2009, 43, 2241–2246. (19) Kauzlarich, S. M.; Teo, B. K.; Averill, B. A. Extended x-ray absorption fine-structure and powder x-ray-diffraction study of feocl intercalated with tetrathiafulvalene and related molecules. Inorg. Chem. 1986, 25, 1209–1215. (20) Webb, S. M. SIXpack: a graphical user interface for XAS analysis using IFEFFIT. Phys. Scr. 2005, T115, 1011–1014.

(21) Slowey, A. J.; Johnson, S. B.; Rytuba, J. J.; Brown, G. E. Role of organic acids in promoting colloidal transport of mercury from mine tailings. Environ. Sci. Technol. 2005, 39, 7869–7874. (22) Ressler, T.; Wong, J.; Roos, J.; Smith, I. L. Quantitative speciation of Mn-bearing particulates emitted from autos burning (methylcyclopentadienyl)manganese tricarbonyl-added gasolines using XANES spectroscopy. Environ. Sci. Technol. 2000, 34, 950– 958. (23) Manceau, A.; Marcus, M. A.; Tamura, N. Quantitative speciation of heavy metals in soils and sediments by synchrotron X-ray techniques. Applications of Synchrotron Radiation in LowTemperature. Geochem. Environ. Sci. 2002, 49, 341–428. (24) Fujimori, T.; Tanino, Y.; Takaoka, M.; Morisawa, S. Chlorination mechanism of carbon during dioxins formation by using Cl-K near edge X-ray absorption fine structure. Bunseki Kagaku 2009, 58, 221–229, in Japanese. (25) Malinowski, E. R. Determination of the number of factors and the experimental error in a data matrix. Anal. Chem. 1977, 49, 612–617. (26) Mendelovici, E.; Villalba, R.; Sagarzazu, A. Processings of iron oxides at room temperature. I. Solid state conversion of pure alpha-FeOOH into distinctive alpha-Fe2O3. Mater. Res. Bull. 1982, 17, 241–249. (27) Mul, G.; Kapteijn, F.; Moulijn, J. A. Catalytic oxidation of model soot by metal chlorides. Appl. Catal. B: Environ. 1997, 12, 33– 47. (28) McKee, D. W. Metal oxides as catalysts for the oxidation of graphite. Carbon 1970, 8, 623–635.

ES903337D

VOL. 44, NO. 6, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1979