X-RAY SPECTROMETRY X-Ray Spectrom. 2008; 37: 210–214 Published online 26 March 2008 in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/xrs.1038

Observing copper chloride during dioxin formation using dispersive XAFS Takashi Fujimori,1∗ Masaki Takaoka,1 Kazuo Kato,2 Kazuyuki Oshita1 and Nobuo Takeda1 1

Department of Urban and Environmental Engineering, Graduate School of Engineering, Kyoto University, Katsura, Nishikyo-ku, 615-8540, Kyoto, Japan 2 Japan Synchrotron Radiation Research Institute (JASRI), 1-1-1, Kouto, Sayo-cho, Sayo-gun, 679-5198, Hyogo, Japan Received 7 May 2007; Accepted 9 January 2008

The incineration of municipal solid waste generates dioxins. To control dioxins, it is necessary to determine the mechanism of their formation. Because copper chloride (CuCl2 ) is one of the strongest catalysts of dioxin formation in fly ash, it is important to study the chemical changes of this compound. To determine the chemical changes of copper, Cu-K x-ray absorption near edge structure (XANES) spectra were examined during the heating of model fly ash using a time-resolved dispersive x-ray absorption fine structure (DXAFS) technique. The change in the absorption edge of the Cu-K XANES spectrum revealed a change in the valence of copper and the temperature dependence of the reduction and oxidation of CuCl2 . From room temperature to 270 °C, CuCl2 was reduced, and from >270 to 300 °C, copper was reoxidized in the presence of 10% oxygen. At a constant temperature of 210 or 300 °C, copper was dynamically reduced and oxidized. In experiments using different gas streams, HCl gas had no effect on the reaction of copper, while oxygen gas was essential for the reoxidation of copper. Our results suggest that dioxin formation in fly ash occurs via the chlorination of carbon with the reduction of CuCl2 and carbon gasification catalyzed by copper oxides. Copyright  2008 John Wiley & Sons, Ltd.

INTRODUCTION In Japan, gas exhausted from municipal solid waste incinerators in the 1980s formed dioxins. This constituted a serious social and environmental problem, because 80% of municipal solid waste by weight was incinerated. With the recent adoption of various advanced air pollution control devices, dioxins in the gas exhausted from most municipal solid waste incinerators now meet the standards for exhausted gas. The fly ash rising from the combustion furnace is trapped by a bug filter or electrostatic precipitator, and dioxins are prevented from moving into the exhausted gas. However, fly ash contains high concentrations of dioxins.1 If these dioxins are not removed from the fly ash, they will inevitably be discharged into the environment. Dioxins are generated in fly ash at suitable temperatures2,3 (about 300 ° C) in a gas atmosphere that contains some oxygen,4 given sufficient reaction time5 (several hours) with metal catalysts.6,7 This is called de novo synthesis of dioxins from unburned carbon. Copper chloride (CuCl2 ) is a major catalyst in the de novo synthesis of dioxins.6 Recently, some researchers8 – 10 have reported that CuCl2 is reduced or oxidized at temperatures from 200 to 400 ° C at which dioxins are generated. Using conventional x-ray absorption fine structure (XAFS) techniques, Takaoka et al.10 reported that CuCl2 is reduced to Cu2 O, CuCl, or elemental copper at temperatures above Ł Correspondence

to: Takashi Fujimori, Department of Urban and Environmental Engineering, Graduate School of Engineering, Kyoto University, Katsura, Nishikyo-ku, 615-8540, Kyoto, Japan. E-mail: [email protected]

Copyright  2008 John Wiley & Sons, Ltd.

200 ° C after analyzing the Cu-K x-ray absorption near edge structure (XANES) of some fly ashes. In that study, the time required to obtain one Cu-K XANES spectrum was in the order of several minutes. Therefore, there is some uncertainty regarding the temperature when the Cu-K XANES spectrum was obtained, because the temperature rises while measuring a Cu-K XANES spectrum. If the measurement time per spectrum were shortened significantly, we could neglect the effect of any rise in temperature and determine the starting temperature at which dioxin formation occurs. Using dispersive x-ray absorption fine structure (DXAFS), which can record the energy region of Cu-K XANES quickly using a bent crystal (polychromator), Cu-K XANES spectra during heating can be measured in several seconds, and temperature information can be obtained. In this study, we examined the temperature dependence of the reduction and oxidation of CuCl2 in model fly ash with in situ XANES measurements using DXAFS.

EXPERIMENT Model fly ash To determine the mechanism of the chemical change of CuCl2 , we produced a model fly ash that imitates real fly ash from a municipal solid waste incinerator. The model fly ash was a mixture of CuCl2 Ð 2H2 O (Nacalai Tesque, Kyoto, Japan), activated carbon (AC; Takeda Pharmaceutical, Osaka, Japan), which removed any organic compounds by heating at 500 ° C for 60 min under a 100% N2 atmosphere (100 mL/min), and boron nitride (BN; Wako Pure Chemical Industries, Osaka, Japan). This model fly ash contained 1.86

Dioxin formation observed by dispersive XAFS

wt% Cu, 2.0 wt% Cl, and 5.0 wt% AC, and the remainder was almost entirely BN. The model fly ash was ground using a mortar for 10 min and an agate mortar for an additional 10 min, and 70 mg of the model fly ash was then pressed into a disk (1.5 mm thick ð 7.0 mm in diameter). This disk was used for the DXAFS experiment.

In situ cell A U-type in situ cell was constructed, consisting of a glass cell, a mantle heater, and a temperature controller (Fig. 1). The cell consisted of a tubular part 3.0 cm in diameter, and the U measured 28 cm high ð 12 cm wide. A disk sample was placed on a glass stand on a support pillar 1.0 cm in diameter connected to a column 2.5 cm in diameter, and inserted in the cell. The x-rays passed through a window made of Kapton film. The temperature of the sample was increased gradually from room temperature to 400 ° C, following the profile shown in Fig. 2. The gas atmosphere was introduced from the inlet of the U-type cell at 50 mL/min and exhausted from the outlet. A water-cooled tube was coiled outside the Kapton film so that the thermal load did not affect the Kapton film or charged-coupled device (CCD) detector.

Cu-K XAFS was measured with the DXAFS system using the third optical hatch of beam line BL28B2 in SPring-8, a synchrotron radiation facility in Hyogo, Japan. The energy area from 8850 to 9320 eV of Cu-K XAFS could be measured in 6.0 s. Temperature information for DXAFS can essentially be treated as representing the same time when the spectrum was measured, because the Cu-K XAFS spectrum could be measured at 0.5 ° C intervals, even when the temperature was increased at 5.0 ° C/min.

RESULTS AND DISCUSSION Analysis of the edge shift of Cu-K XANES The change in the Cu-K XANES spectra on heating under 10% O2 (and 90% N2 ) gas is shown in Fig. 3. The x-ray absorption edge positions and energy positions of the maximum absorption of the XANES spectra changed. Using the DXAFS system, it is difficult to clearly identify fine changes in the pre-edge position or shoulder shape, because a XANES spectrum measured using the DXAFS system has a lower energy resolution than a conventional XAFS system. For instance, although the pre-edge position of Cu2 O

Dispersive XAFS (DXAFS) Temperature (°C)

400

The DXAFS system consists of a Laue-type polychromator with a bent silicon crystal and a CCD coupled to a scintillating screen and lens system. A detailed explanation is provided elsewhere.11,12 DXAFS allows us to collect continuous energy data from one location on the sample disk because of the bend in the polychromator. When x-rays pass through the sample, the transmittance of the x-rays depends on their energy; consequently, different shades of light are generated at the pre- and postedge of the XAFS. The stripes of light and shade are detected with a CCD camera. Using DXAFS, the energy of a target area can be measured all at once. The

300 200 100 0

0

60

120

180

Time (min)

Figure 2. General time and temperature profile used. Outlet

Inlet

Control valve

Thermocouple Heating device

Cooling tube

Kapton film

Polychromated x-rays

Fly ash disk

Support bed

Figure 1. Schematic of the in situ cell.

Copyright  2008 John Wiley & Sons, Ltd.

X-Ray Spectrom. 2008; 37: 210–214 DOI: 10.1002/xrs

211

T. Fujimori et al.

1.1-1.2

3

1.0-1.1 0.9-1.0 Valence of copper

0.8-0.9 0.2-0.8 0.1-0.2 0.0-0.1

28

2

1

80 0

14

0

Absorption (-)

20

0

21

0 23 0

29 Temperature (degree C)

00

3

9030

0

0 2

0

6

4

8

Edge shift (eV)

Figure 4. Relationship between the edge shift and the valence of copper.

31

0

0.05-0.06 Photon energy (eV)

Change in the valence of copper while heating the model fly ash The change in the Cu-K XANES derivative spectra of model fly ash on heating under a stream of 10% O2 gas is shown in Fig. 5. We observed a dynamic change in the edge during heating from room temperature to 400 ° C (Fig. 6). The edge shift at room temperature (28 ° C) was 6.0 eV, which suggests that the copper was bivalent Cu(II). The bivalent copper compound was thought to be CuCl2 , because the only copper compound mixed in the model fly ash at room temperature was CuCl2 Ð 2H2 O, and the edge shift of

Copyright  2008 John Wiley & Sons, Ltd.

0.02-0.03 0.01-0.02 0.00-0.01 9001

8981 0

Photon energy (eV) 0 0 0

34

0

8960 40

30

30

Temperature (degree C)

0

21

differs from that of CuCl, the difference was barely observed using our DXAFS system. Therefore, the chemical form of copper was difficult to identify, although its valence could be determined by analyzing the edge shift13,14 as follows. The position of the absorption edge (simply called the ‘edge’) was taken as the energy point of the first maximum in the derivative of the Cu-K XANES spectrum. The edges of ten different standard copper compounds—CuCl2 Ð 2H2 O, CuO, Cu2
OH3 Cl, CuCO3 , CuPO4 , CuSO4 , CuFe2 O4 , Cu2 O, CuCl, and Cu—were calculated from derivative XANES spectra. The edge of elemental copper (Cu) was taken as 8980.0 eV. Here, we define the edge shift E as the relative edge difference from Cu. The relation between the edge shift and the valence of copper is shown in Fig. 4. The E of univalent Cu(I) compounds was between 1.5 and 6.0 eV (average 3.75 eV), and the E of bivalent Cu(II) compounds ranged from 5.8 to 7.8 eV (average 6.79 eV). A larger edge shift was observed for the higher oxidation state. Using the change in the edge shifts of the Cu-K XANES spectra of the model fly ash on heating, the dynamic change in the valence could be investigated.

0.03-0.04

25

Figure 3. 3D Cu-K edge XANES spectra of cupric chloride model fly ash on heating under a stream of 10% O2 gas.

0.04-0.05

10 0 16 0 21 0

0 40 8959

0.06-0.07

8995

Derivative absorption (-)

0 40

28

37

Figure 5. 3D Cu-K edge XANES derivative spectra of CuCl2 model fly ash while heating under a stream of 10% O2 gas.

8 Cu(II) Edge shift (eV)

212

6

4

Cu(I)

2 Cu(0) 0

0

100

200

300

400

Temperature (°C)

Figure 6. Change in the edge shift on heating the model fly ash under a stream of 10% O2 gas.

standard CuCl2 Ð 2H2 O is 5.8 eV. At temperatures from 50 to 75 ° C, the edge shift was around 5 eV, which is within

X-Ray Spectrom. 2008; 37: 210–214 DOI: 10.1002/xrs

Dioxin formation observed by dispersive XAFS

Copyright  2008 John Wiley & Sons, Ltd.

Effects of oxygen and hydrogen chloride gas using DXAFS Cu-K XANES spectra were measured with 1000 ppm HCl gas added to 10% O2 . The change in the XANES spectra upon heating process was not different from that without HCl gas (cf. Figure 3). Therefore, the change in the valence of copper was the same with or without HCl gas. This suggests that the CuCl2 in the fly ash did not react with the HCl gas, which concurs with a report by Wikstrom ¨ et al.19 Oxygen gas is one of the important factors in the de novo synthesis of dioxins.4 The change in the Cu-K XANES spectra on heating was also investigated under 100% nitrogen (no oxygen), as shown in Fig. 7. As the temperature increased, the energy position of the edge decreased, and the maximum absorption of the XANES spectrum weakened. Note that the change in the pattern of the Cu-K XANES spectra was quite different from that under 10% O2 . The change in the Cu-K XANES derivative spectra on heating under 100% N2 gas is shown in Fig. 8. The peak position of the derivative spectrum (i.e., the ‘edge’) shifted from high to low energy on heating, but did not return to high energy, which occurred with 10% O2 (see Fig. 5). The change in E on heating the model fly ash under 100% N2 gas is shown in Fig. 9. The edge shift decreased gradually, and the change in the patterns of the edge shift with and without oxygen was the same from room temperature to 270 ° C. Therefore, the reduction of CuCl2 was thought to be unrelated to the presence of oxygen gas. The edge shift characteristically decreased gradually from 4.7 to 2.7 eV and between 100 and 200 ° C, which implied that the reduction of copper was temperature sensitive. When the temperature was held at 210 ° C for 20 min, the edge shift decreased from 2.7 to 1.7 eV. In comparison, with 10% oxygen, the change in the edge shift over these 20 min was smaller. There are two reasons for this small decrease. First, a majority of the CuCl2 was already reduced before the temperature reached 210 ° C. Second, the edge shifts after 20 min were similar: 1.7 eV under 10% O2 gas vs. 2.3 eV with no O2 gas. It is expected that the reduction of copper at a constant temperature of 210 ° C is limited. From 220 to

1.1-1.2 1.0-1.1 0.9-1.0 0.8-0.9 0.2-0.8 0.1-0.2 0.0-0.1

Absorption (-)

the Cu(I) region. Therefore, some of the bivalent copper was reduced to univalent copper. When the temperature was held at 80 ° C for about 5 min, the edge shift decreased to 4.5 eV, which implies that the time it is held at a stable temperature influences the reduction of copper. E remained at 4.5 eV as the temperature was increased to 200 ° C. On holding the temperature at 210 ° C for around 20 min, the edge shift decreased dynamically from 4.5 to 2.3 eV, which is smaller than the average edge shift of Cu(I) (E D 3.75 eV). Most of the copper was thought to have been reduced to Cu(I) during the 20 min at 210 ° C. At temperatures above 230 ° C, the edge shift was less than that of Cu(0) (E D 1.5 eV). At 260 and 270 ° C, the edge shift reached a minimum of 0.5 eV. Therefore, a significant reduction of Cu(II) or Cu(I) to Cu(0) was thought to have occurred. It has been suggested that CuCl2 acts as a chlorine source for carbon,1 because the chlorine atom moves away from the copper atom during this reduction process (particularly at temperatures from 210 to 270 ° C). Conversely, from 270 to 290 ° C, the edge shift increased from 0.7 to 1.4 eV and then increased markedly from 1.4 to 6.4 eV when a constant temperature of 300 ° C was maintained for 20 min. This suggests that the copper was oxidized, because the valence of the copper increased with the edge shift. The copper compounds that formed were thought to be copper oxides (Cu2 O or CuO), because 10% O2 gas surrounded the model fly ash and served to oxidize the copper. This was determined by comparing the effects of 10% O2 and 100% N2 gas. At temperatures from 270 to 300 ° C, copper oxides have reported to catalyze carbon gasification,15,16 destroying or breaking up the carbon matrix in the model fly ash.17,18 From 310 to 400 ° C, the edge shifts were almost 6.4 eV, which is in the bivalent Cu(II) region. Between 310 and 400 ° C, the copper may exist as CuO, which would catalyze carbon gasification continuously. The amount of unburned carbon in the CuCl2 model fly ash was measured at room temperature and after heating to 200, 300, or 400 ° C for 30 min under 10% O2 using a total organic carbon analyzer (TOC-V, Shimadzu, Japan). The content of unburned carbon (equal to the total carbon) at 300 and 400 ° C decreased to 81 and 44 wt%, respectively. By contrast, when only activated carbon was heated, the content of unburned carbon did not change from room temperature to 400 ° C. The chemical form of the copper at 300 and 400 ° C was thought to be an oxidative product based on the DXAFS experiment. According to these results, carbon gasification catalyzed by the copper occurred via the formation of copper oxides at temperatures between 200 and 300–400 ° C. The combination of the reduction of copper at temperatures up to 270 ° C and its oxidation from 270 to 400 ° C drove the chlorination and gasification of the carbon matrix. The probability of dioxin formation is thought to be the highest at 300 ° C, which was the temperature at which the reported concentrations of dioxins in fly ash were maximal with de novo synthesis,2 because almost all of the copper was reduced at temperatures up to 210 ° C and then oxidized to form copper oxides at 300 ° C. It is believed that destruction of the carbon structure of dioxins is facilitated when there is no chlorination of the carbon; for this reason, dioxin formation decreases at temperatures over 300 ° C, as reported by Vogg et al.2

28 90 140 190 230 Temperature 280 (degree C) 320

9030

8995

370 8959

Photon energy (eV)

Figure 7. 3D Cu-K edge XANES spectra of CuCl2 model fly ash during heating under a stream of 100% N2 gas.

X-Ray Spectrom. 2008; 37: 210–214 DOI: 10.1002/xrs

213

T. Fujimori et al.

0.06-0.07 0.05-0.06 0.04-0.05 0.03-0.04 0.02-0.03 Derivative absorption (-)

0.01-0.02 0.00-0.01 9001

28

8981

0 0

0 0

8960

Figure 8. 3D Cu-K edge XANES derivative spectra of CuCl2 model fly ash during heating under 100% N2 gas.

6

4

Cu(I)

The temperature dependence of the CuCl2 in the model fly ash during the de novo synthesis of dioxin was revealed using DXAFS. The change in the absorption edge of the CuK XANES spectrum in the model fly ash corresponded to the change in the valence of the copper. Under a stream of 10% O2 gas, copper was reduced to Cu(I) or Cu(0) from room temperature to 270 ° C. When the temperature was held at 210 ° C for 20 min, dynamic reduction of the copper occurred. When the temperature exceeded 270 ° C, copper was reoxidized by the oxygen gas because there was no reoxidation of copper under 100% N2 . When held at 300 ° C under the oxidative atmosphere, the oxidation of the copper was dynamic. This suggested that copper oxides did not change the oxidation state at temperatures from 300 to 400 ° C for 20 min. When HCl gas was mixed with the 10% O2 , there was no difference in the change of the Cu-K XANES spectrum, which suggests that the CuCl2 in the fly ash did not react with the HCl gas. Therefore, the chlorination and gasification of carbon are thought to be important for dioxin formation in fly ash containing CuCl2 .

Acknowledgements The synchrotron radiation experiments were performed at the SPring-8 with approval of the Japan Synchrotron Radiation Research Institute (JASRI) (Proposal No.2004B0512-Nxa-np). We thank Dr. Kentaro Teramura (Kyoto University) for performing the experiments using DXAFS.

REFERENCES 1. 2. 3. 4. 5.

2 Cu(0) 0

-2

CONCLUSIONS

37

32

28

0

23

19

14

0

90

Photon energy (eV)

Temperature (degree C)

Edge shift (eV)

214

6.

0

100

200

300

400

Temperature (°C)

Figure 9. Change in the edge shift when the model fly ash was heated under a stream of 100% N2 gas.

7. 8. 9. 10.

270 ° C, the edge shift decreased from 1.4 to 0.3 eV, which was very similar to the change with 10% O2 . Over 270 ° C, the edge shift continued to decrease under 100% N2 , while the edge shift increased again under 10% O2 gas. There was no reoxidation at temperatures over 270 ° C under 100% N2 . At 280 ° C, the edge shift was 0.0 eV, i.e., that of Cu(0), and from 300 ° C to a constant temperature at 400 ° C, the edge shift decreased
0.8 eV. Consequently, most of the copper in the model fly ash was thought to be elemental copper. Therefore, oxygen gas is necessary in the surrounding atmosphere for the reoxidation of the copper. Few dioxins formed in the fly ash under 100% N2 . Conversely, the presence of oxygen gas stimulates dioxin formation.4 Our results suggest that the reoxidation of the copper at temperatures >270 ° C plays an important role in dioxin formation.

Copyright  2008 John Wiley & Sons, Ltd.

11. 12. 13. 14.

15. 16. 17. 18. 19.

Addink R, Olie K. Environ. Sci. Technol. 1995; 29: 1425. Vogg H, Metzer M, Stieglitz L. Waste Manage. Res. 1987; 5: 285. Milligan MS, Altwicker E. Environ. Sci. Technol. 1993; 27: 1595. Addink R, Olie K. Environ. Sci. Technol. 1995; 29: 1586. Stieglitz L, Zwick G, Beck J, Bautz H, Roth W. Chemosphere 1989; 19: 283. Stieglitz L, Zwick G, Beck J, Roth W, Vogg H. Chemosphere 1989; 18: 1219. Stieglitz L, Vogg H, Zwick G, Beck J, Bautz H. Chemosphere 1991; 23: 1255. Weber P, Dinjus E, Stieglitz L. Chemosphere 2001; 42: 579. Kuzuhara S, Sato H, Kasai E, Nakamura T. Environ. Sci. Technol. 2003; 37: 2431. Takaoka M, Shiono A, Nishimura K, Yamamoto T, Uruga T, Takeda N, Tanaka T, Oshita K, Matsumoto T, Harada H. Environ. Sci. Technol. 2005; 39: 5878. Iwasawa Y. J. Catal. 2003; 216: 165. Spring-8 Website, 2007; http://www.spring8.or.jp/ [26 April 2007]. Brown NMD, McMonagle JB. J. Chem. Soc., Faraday Trans. 1 1984; 80: 589. Wong J, Rek ZU, Rowen M, Tanaka T, Sch¨afers F, Muller B, ¨ George GN, Pickering IJ, Via G, DeVries B, Brown GE Jr, Froba ¨ M. Physica B 1995; 208&209: 220. Mackee DW. Carbon 1970; 8: 623. Kyotani T, Yamada H, Yamashita H, Tomita A, Radovic LR. Energy Fuels 1992; 6: 865. Wilhelm J, Stieglitz L, Dinjus E, Zwick G. Organohalogen Compd. 1999; 41: 83. Olie K, Schoonenboom MH, Addink R. Organohalogen Compd. 1995; 23: 329. Wikstrom ¨ E, Ryan S, Touati A, Telfer M, Tabor D, Gullett BK. Environ. Sci. Technol. 2003; 37: 1108.

X-Ray Spectrom. 2008; 37: 210–214 DOI: 10.1002/xrs