Environ. Sci. Technol. 2001, 35, 1276-1281
Decomposition of Dichloromethane in a Wire-in-Tube Pulsed Corona Reactor L. HUANG, K. NAKAJO, S. OZAWA, AND H. MATSUDA* Research Center for Advanced Waste and Emission Management, Nagoya University, Nagoya 464-8603, Japan
Decomposition of low concentration dichloromethane in nitrogen-based gas was experimentally investigated by a wirein-tube pulsed corona reactor. Maximum decomposition was found in pure nitrogen, and the decomposition decreased with the increase of oxygen concentration in the gas. The major product detected by FTIR from the decomposition of dichloromethane without oxygen participation was HCl, while CO, CO2, COCl2, and NOx were the main detected products in the presence of oxygen. Aiming at removing the unwanted byproducts from the decomposition reaction, a combination of corona discharge and gas absorption was devised by coating a thin layer of Ca(OH)2 on the inner wall of the corona reactor. It was demonstrated that this kind of combination was capable of scavenging the products of phosgene and nitrogen oxides from the plasma decomposition of dichloromethane.
Introduction Emissions of volatile organic compounds (VOCs) into atmosphere may cause detrimental influences on both human health and global environment. Volatile organic chlorides are a type of compounds that are related to the depletion of stratospheric ozone and have the potential of further formation of more toxic compounds under certain conditions because of the chlorine contained in the compounds (1, 2). Therefore, it is necessary to decompose them before being discharged into the air. However, VOCs such as CFCs are difficult to decompose because of their chemical stability and inertness. In recent years, nonthermal plasma technology was applied to the decomposition of VOCs by different plasma reactors (3-8). The decompositions may experience the process of initiation, propagation, and termination of the concerned reactions. The dissociation and/or excitation of the targeted VOC molecules by direct impact of energetic electrons and the production of active radicals such as O, N, and OH by electron dissociation and ionization of background gas molecules are the initial step. The chain reactions afterward between the produced radicals and VOC molecules will propagate until stable compounds are formed. The ideal products of decomposition of toxic compounds are CO2, H2O, and HCl for organic chlorides. But, since the plasma induced radical reactions are difficult to control, previous investigations showed that some toxic byproducts formed inevitably from the decomposition of some VOCs (6, 8), which may cause additional environmental pollution. Therefore, it is * Corresponding author. Phone: +81-52-789-3382; fax: +81-52789-5619; e-mail:
[email protected]. 1276
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 6, 2001
essential to prevent the production of secondary pollutants for this technology to be applied to practical applications. In this study, we investigated the decomposition of dichloromethane (CH2Cl2) in a cylindrical reactor to which dichloromethane was fed under nonthermal plasma induced by positive pulsed corona discharge. Such a kind of reactor has larger discharge volume and lower pressure drop than the ferroelectric pellet packed bed plasma reactor and, therefore, is more suitable for the treatment of large volume off-gas. In the study, special attention was given to the behavior of chlorine during the decomposition process, and a technique of plasma reaction combined with in-situ absorption was developed to realize simultaneously decomposition of dichloromethane and capture of the unwanted byproducts.
Experimental Section A schematic diagram of the experimental system is shown in Figure 1. The system consists of a sample gas feeding system, a corona reactor, a high pulse voltage generator, and a gas sampling and analysis system. A pressured CH2Cl2 gas cylinder of standard concentration of 500 ppm in N2 was used to make different concentrations of CH2Cl2 by mixing with N2 or O2 gas for experiments. Total gas flow rate into the reactor was 500 cm3/min, and all experiments were implemented at atmospheric pressure and ambient temperature of about 15 °C. A typical wire-tube combination was adopted as the corona reactor for the experiment. It consists of a Pyrex glass tube with an aluminum film attached to the inner wall as the grounding electrode and a coaxial stainless steel wire of 0.5 mm in diameter as the discharge electrode. The inner diameter of the tube is 28 mm, and the aluminum film is 0.3 mm thick and 500 mm long. The high pulse voltage applied to the wire electrode was generated by a rotating spark gap system. It was tested that the width of the pulse was less than 300 ns and that the rising width of the pulse was less than 50 ns. The pulse repetition rate was 50 pulses/s, and the energy injected into the reactor was calculated to be about 620 J/L of treatment gas for a pulse voltage of 20 kV. The analyses of sample gas before and after reaction were carried out using on-line FTIR (SHIMADZU, FTIR-8700) with a sampling cell of 10 cm path length, GC-MS (SHIMAZU, GC-QP5050) with a capillary column(ULBON HR-1 0.25 mm × 25m), and a gas chromatograph with a methanizer and a FID detector for CO2 measurement. The decomposition of CH2Cl2 is defined as
η)
Cin - Cout × 100% Cin
(1)
where Cin and Cout are the inlet concentration of CH2Cl2 and outlet concentration of CH2Cl2, respectively.
Results and Discussion The decomposition of CH2Cl2 as a function of O2 concentration is shown in Figure 2. The pulse voltage (peak voltage) applied to the reactor was 20 kV (620 J/L). The experimental results showed that the maximum decompositions of CH2Cl2 were found in the absence of O2 for all tested concentrations of 100, 200, and 300 ppm and decreased with the increase of O2 concentration in the sample gas. The decrease of decomposition of CH2Cl2 by the presence of O2 was also verified by Penetrante et al. (5) and very recently by 10.1021/es0011414 CCC: $20.00
2001 American Chemical Society Published on Web 02/16/2001
FIGURE 1. Schematic diagram of experimental system.
FIGURE 2. Destruction of CH2Cl2 as a function of O2 concentration. Pulse voltage, 20 kV. Fitzsimmons et al. (9). But we did not find a peak decomposition that was found by Fitzsimmons et al. (9) by a dielectric packed-bed plasma reactor when the carrier gas contains 1-3% O2. This difference may be ascribed to the difference in plasma generation methods. The mechanism for the deterioration of CH2Cl2 decomposition by O2 coexisted in the gas was thought that the electronegative gas of O2 captured a certain number of energrtic electrons that might be contributed to the destruction of CH2Cl2. That is to say, the initiation step for CH2Cl2 decomposition, direct dissociation or excitation of CH2Cl2 and production of active radicals by energetic electron impact might be restrained to some degree. Furthermore, although O2 is easy to dissociate under corona discharge to produce active O radicals due to low dissociation energy, the reaction rate of O radical with CH2Cl2 is much lower as compared with the reaction rate of N radical with CH2Cl2 (5). Additionally, the production of NOx also consumed a number of effective active radicals.
With the increase of O2 concentration to 10% in the sample gas, we detected the production of ozone from the off-gas, which has been regarded having slow reaction rates with VOCs (10). The FTIR spectra of the sample gas before and after reactions are shown in Figure 3. A production of HCl was detected while the CH2Cl2 destruction was performed in N2 atmosphere as shown in Figure 3b, indicating that the Cl in CH2Cl2 might exist in a chemically stable form of HCl after destruction. It was found that over 80% of total Cl content in CH2Cl2 to form HCl after the decomposition of 100 ppm CH2Cl2. Another possible product was thought to be HCN as it was the main product detected by Fitzsimmons et al. (9). But Penetrante et al. (5) thought the production of HCN was not appreciable by pulsed corona processing of dichloromethane in N2. After a long time running, a carbon-like particulate layer was found to be deposited on both the wire surface and the inner wall of the tube, which was regarded VOL. 35, NO. 6, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1277
FIGURE 3. FTIR spectrum patterns before and after decomposition reaction. (a) CH2Cl2, 300 ppm; O2, 0%; no discharge. (b) CH2Cl2, 300 ppm; O2, 0%; pulse voltage, 20 kV. (c) CH2Cl2, 300 ppm; O2, 20%; pulse voltage, 20 kV. as the solid products produced from recombined of carbon element or molecular fragments from CH2Cl2 decomposition reactions. The CO peak in the FTIR pattern might be due to the traces of oxygen contained in the sample gas. With the coexisting of O2, the ideal final products from CH2Cl2 destruction are expected to be CO2 and HCl. But from the FTIR pattern in Figure 3c, almost no HCl was detected in the case of 20% O2 contained in the sample gas. On the contrary, many product peaks appeared in the FTIR pattern. 1278
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 6, 2001
Further investigation identified these products mainly as COCl2, NOx, CO, and CO2. It seems that oxygenation reactions dominated the whole decomposition process. But the complete oxidation of CH2Cl2 to CO2 seemed to be prevented by side reactions such as the reaction CH2Cl2 + O f COCl2 + H2. Also, the radical reaction between O and molecular fragments from CH2Cl2 might form RO compounds, where R represents molecular fragments from CH2Cl2. But it is difficult to identify these compounds in the FTIR pattern. In
FIGURE 4. Influence of reaction temperature on CH2Cl2 decomposition. CH2Cl2, 100 ppm.
FIGURE 5. Influence of O2 concentration on HCl production. the destruction process, nitrogen oxides were the unavoidable byproducts produced by the reactions between various radicals from dissociative excited N2 and O2 (11). The influences of pulse voltage and temperature on CH2Cl2 decomposition were shown in Figure 4. It was confirmed that with the increase of the pulse voltage applied to inner electrode resulted in the increase of decomposition efficiency, because more energetic electrons were generated with higher pulse voltage. High temperature at 100 °C will be more favorable for the decomposition as compared with room temperature. It indicated although the direct electron impact dissociation is almost independent of temperature, the radical reactions were enhanced at high temperature (12). The HCl yield from CH2Cl2 decomposition is shown in Figure 5, suggesting that the HCl yield decreased sensitively with the increase of O2 concentration in the sample gas. But at a higher temperature of 100 °C, there was still a certain amount of HCl produced even with high O2 concentration, indicating
that the oxidation destruction of CH2Cl2 processed more completely at high temperature. As a demonstration, it was found that more CO2 was produced when the reaction took place at 100 °C as shown in Figure 6. To prevent the production of unfavorable byproducts such as phosgene from CH2Cl2 decomposition, a thin Ca(OH)2 layer of about 1 mm thick was coated on the aluminum surface, the grounding electrode of the corona reactor. The function of Ca(OH)2 is as a dry absorbent for in-situ capture of the chlorine produced from the plasma reaction, leading the reaction to avoid the production of COCl2 or other chlorinous compounds. Figure 7 shows the FTIR spectra of the decomposition of CH2Cl2 in the corona reactor combined with Ca(OH)2 absorbent. Comparing Figure 7a with Figure 3b, it can be seen that the peak of the product HCl disappeared from FTIR pattern under the reaction condition without supplying O2 to the reactor. The CO peak in Figure 3b produced by oxidizing VOL. 35, NO. 6, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1279
FIGURE 6. CO2 yield from CH2Cl2 decomposition. CH2Cl2, 300 ppm.
FIGURE 7. FTIR patterns of the products from the plasma decomposition combined with in-situ Ca(OH)2 absorption. (top) CH2Cl2, 300 ppm; O2, 0%; pulse voltage, 20 kV. (bottom) CH2Cl2, 300 ppm; O2, 20%; pulse voltage, 20 kV. reaction due to trace O2 in the reactor was replaced by CO2 peak in Figure 7a, which means the decomposition of CH2Cl2 proceeded more completely in this reactor. It was also found from Figure 7b that the major products were CO2 and H2O under the condition of 20% O2 contained in the sample gas. Compared with Figure 3b, the byproduct of phosgene was not detected. It was therefore considered that the 1280
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 6, 2001
Ca(OH)2 layer played an important role in preventing unfavorable plasma reactions by in-situ capturing of the chlorine produced by plasma dissociation of CH2Cl2 molecule, therefore promoting the decomposition reaction to the expected direction. The existence of captured chlorine by Ca(OH)2 was ascertained as the form of CaCl2 by using an ion chromatograph to analyze the water solution of the Ca-
(OH)2 layer after the decomposition reaction. It was also found that the products of nitrogen oxides were captured by the Ca(OH)2 layer to form Ca(NO3)2. On the basis of the present experimental results, it can be summarized that decomposition of CH2Cl2 was successfully achieved in a wire-in-tube pulsed corona reactor. The decomposition efficiency of CH2Cl2 decreased with the increase of O2 concentration in the sample gas. The major product from CH2Cl2 decomposition was HCl without O2 participation. On the other hand, the HCl yield decreased dramatically with the increase of O2 concentration in the sample gas. Instead, many oxygenous products of phosgene and nitrogen oxides were formed in the presence of O2. Raising the reaction temperature and the pulse voltage resulted in an increase of dichloromethane decomposition and CO2 yield. It was demonstrated that the production of the unwanted intermediates was suppressed by a combination of plasma decomposition reaction and in-situ absorption. The Cl derived from the rupture of the C-Cl bond was found to be absorbed by the Ca(OH)2 layer, which was coated on the inner wall of the corona reactor, thereby inhibiting the reaction to form COCl2.
Literature Cited (1) Rowland, F. S. Environ. Sci. Technol. 1991, 25, 622.
(2) Wendell, L. D.; Corwin, J. B.; Nancy, B. T. Environ. Sci. Technol. 1974, 10, 351. (3) Yamamoto, T.; Ramanathan, K.; Lawless, P. A.; Ensor, D. S.; Newsome, J. R.; Plaks, N.; Ramsey, G. H. IEEE Trans. Ind. Appl. 1992, 28 (3), 528. (4) Chang, M. B.; Lee, C. C. Environ. Sci. Technol. 1995, 28, 181. (5) Penetrante, B. M.; Hsiao, M. C.; Bardsley, J. N.; Merritt, B. T.; Vogtlin, G. E.; Kuthi, A.; Burkhart, C. P.; Bayless, J. R. Phys. Lett. A 1997, 235, 76. (6) Kohno, H.; Berezin, A. A.; Chang, J.-S.; Tamura, M.; Yamamoto, T.; Shibuya, A.; Honda, S. IEEE Trans. Ind. Appl. 1998, 34 (5), 953. (7) Sano, N.; Tamon, H.; Okazaki, M. Ind. Eng. Chem. Res. 1998, 37, 1428. (8) Yamamoto, T.; Jang, B. W.-L. IEEE Trans. Ind. Appl. 1999, 35 (4), 736. (9) Fitzsimmons, C.; Ismail, F.; Whitehead, J. C.; Wilman, J. J. J. Phys. Chem. A 2000, 104, 6032. (10) Falkenstein, Z. J. Appl. Phys. 1999, 85 (1), 525. (11) Eliasson, B.; Kogelschatz, U. IEEE Trans. Plasma Sci. 1991, 19, 1063. (12) Evans, D.; Rosocha, L. A.; Anderson, G. K.; Coogan, J. J.; Kushner, M. J. J. Appl. Phys. 1993, 11 (1), 5378.
Received for review March 29, 2000. Revised manuscript received October 27, 2000. Accepted January 8, 2001. ES0011414
VOL. 35, NO. 6, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1281
