Environ. Sci. Technol. 1999, 33, 2234-2240 the concentration of halogen component (12). In catalytic Decomposition of method, the catalysts must be replaced frequently due to the poisoned effect by the reaction of byproducts (10). Dichlorodifluoromethane by Adding Radio-frequency (RF) plasma is a branch of nonequilib- Hydrogen in a Cold Plasma System rium plasma that is often referred to as cold plasma (13). The temperature of the gas molecule in an RF plasma reactor is near room temperature, while the temperature of electrons YA-FEN WANG,² WEN-JHY LEE,*,² will be higher than 104 K. At such high temperature, the energy CHUH-YUNG CHEN,³ AND of an electron is much higher than 200 kJ mol-1.Itistwo LIEN-TE HSIEH² times magnitude higher than the activation energy of a Department of Environmental Engineering and Department conventional chemical reaction (100 kJ mol-1)(14). The of Chemical Engineering, National Cheng Kung University, kinetic energy of electrons and ions is much higher than that No. 1, University Road, Tainan 70101, Taiwan of molecules in the cold plasma system. Therefore, the conventional reaction, which needs to proceed at a very high temperature, can be finished at lower temperature in the - The destruction of chlorofluorocarbons (CFCs) has drawn cold plasma system (15 17). The coexistence of warm gas great attention because of its well-known depletion of and high-temperature active species distinguishes the plasma system from conventional thermal processing (18). Besides, the ozone layer at the stratosphere. Application of RF plasma RF plasma is free from electrode erosion by byproducts such for the decomposition and conversion of dichlorodifluoro- as HCl or HF or spark problem while comparing with DC methane (CFC-12 or CCl2F2) is demonstrated. The decom- plasma or pulsed corona (19, 20). position efficiency (ηCFC-12) and the fraction of total carbon In this study, CFC-12 and H2 mixtures were introduced input converted into CH4 and C2H2 (FCH4+C2H2)inH2 and into the RF plasma reactor with an operational pressure of Ar mixtures have been investigated as a function of input 15 Torr. In contrast with the previous studies (7-10), this power wattage, H2/CFC-12 ratio, operational pressure, work concentrated on not only the CFC-12 decomposition and CFC-12 feeding concentration by using response surface but also on the conversion of the carbon on CFC-12 into methodology and model sensitivity analysis. The results methane (CH4) and acetylene (C2H2). In addition, the fractional factorial design (21) was employed in planning showed that the - is over 94% and the F + is over ηCFC 12 CH4 C2H2 the experiments for studying the effects of experimental 80% under the condition of 100 W of input power wattage, variables on the CFC-12 decomposition efficiency (ηCFC-12) 7.0 of H /CFC-12 ratio, 15 Torr of operational pressure, 2 and the fraction of total carbon input converted into CH4 and 7.6% of CFC-12 feeding concentration. In addition, the and C2H2 (FCH4+C2H2). Furthermore, the results of experiments result of sensitivity analysis indicated that both the were subjected to regression analysis to determine the - + ηCFC-12 and the FCH4+C2H2 are more sensitive to CFC-12 conditions required for a better ηCFC 12 and a higher FCH4 C2H2. feeding concentration and input power wattage. Experimental results showed that CFC-12 could be converted into CH4 Experimental Apparatus and C2H2 up to 80% of conversion in a hydrogen-based RF The experimental apparatus used in this study is schemati- plasma system. cally shown in Figure 1. The CFC-12/H2/Ar mixing gas was metered using Tylan type 1179 mass flow controllers at a total flow rate of 100 standard cm3/min that were entered, Introduction then directed to a mixing vessel, and introduced perpen- dicularly into a 4.14 × 15 cm cylindrical glass reactor. The The occurrence of skin cancer increased approximately 1.5% RF plasma discharge was produced using a plasma generator as the amount of ozone layer decreased 1% in the stratosphere (PFG 600 RF, Fritz Huttinger Elektronik Gmbh) at 13.56 MHz (1). Since 1970, the development of skin cancer increased and with a matching network (Matchbox PFM). RF power from 15% to near 60% (1, 2). Chlorofluorocarbons (CFCs) was delivered through the power meter and the matching were well-known man-made compounds that contributed unit to an outer copper electrode wrapped on the reactor to the depletion of the ozone layer at the stratosphere. The with another electrode earthed. destruction of CFCs has drawn great attention around the Before the experiment, a diffusion pump was used to keep world due to the stepwise forbiddance in accordance with the system pressure lower than 0.001 Torr for the cleanup the 1987 Montreal Protocol and subsequent revision (3). CFCs of contamination. For each designed experimental condition are, per molecule, 10 000 times more efficient in absorbing [the input power wattage (factor A), the H /CFC-12 ratio (B), infrared radiation than carbon dioxide, and their atmospheric 2 the operational pressure (C). and the CFC-12 feeding lifetimes fall into the range of 40-150 year (4). Furthermore, concentration (D)] were measured more than three times their stability led to rapid accumulation in the global within a 5-min period in order to ensure that steady-state environment and then destroyed the stratospheric ozone conditions were achieved. The reactants and final products layer by invading the Chapman cycle (5). Most investigations were identified by gas chromatography/mass spectrometry focused on how to destroy CFCs, either by thermal technology (Varian Saturn 2000 GC/MS) first and then identified and (6-9) or catalytic method (10, 11). In the incineration method, quantified by an on-line Fourier transform infrared (FTIR) the precursors of PCDD/PCDFs (polychlorodibenzo-p- spectrometer (Bio-Rad, Model FTS-7). dioxin/polychlorinated dibenzofurans) were suspected to be The reactants and products were sampled from the outlet the products of halogenated hydrocarbons while increasing valve downstream from the plasma reactor through a canister sampler (Figure 1). The collected gas was separately injected * Corresponding author telephone: 886-6-275-7575, ext 54531; fax: 886-6-275-2790; e-mail: [email protected]. into the GC/MS and the GC/FID (gas chromatography/flame ² Department of Environmental Engineering. ionization detector) systems by thermal desorption ap- ³ Department of Chemical Engineering. paratus. Chlorofluorocarbons were analyzed by the GC/MS, 2234 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 13, 1999 10.1021/es980757e CCC: $18.00 1999 American Chemical Society Published on Web 05/13/1999 FIGURE 1. Schematic of the RF plasma system. TABLE 1. Factors and Levels for the 24-1 Fractional Factorial TABLE 2. Estimates of the Effects from the 24-1 Fractional Design Factorial Design with the Defining Relation I ) ABCD levels effects estimate (ηCFC-12) estimate (FCH4+C2H2) factors +- *A + BCD 22.95 26.36 B + ACD -15.24 16.73 A: input power wattage (W) 120 50 C + ADB -17.95 -7.33 B: H /CFC-12 ratio 2.0 0.5 2 *D + ABC -36.59 -27.09 C: operational pressure (Torr) 20 10 AB + CD -9.15 2.13 D: CFC-12 feeding concentration (%) 15 8.0 AC + BD -14.49 -13.08 AD + BC -2.73 9.25 while hydrocarbons were analyzed by the GC/FID. Calibra- tion of gaseous reactants and products was made by efficiency (%) withdrawing unreacted gases and by going directly through the sampling line connected to the FTIR. The mass of species ) [(C - C )/(C )] × 100% (1) was calculated by comparing the response factor (absorbance in out in height/concentration) of standard gas at the same IR C : the feeding concentration of CFC-12 (%) C : the effluent wavenumber. To understand the significance of deposition in out concentration of CFC-12 (%) F + (%): the fraction of and condensation that occurred in the sampling and analyz- CH4 C2H2 total carbon input converted into CH and C H (%) ing system, the FTIR quantification data were also checked 4 2 2 through carbon balance. ) [C + /C ] × 100% (2) CH4 C2H2 in Results and Discussion C + : the effluent concentration of CH and C H (%). Fractional Factorial Design. The fractional factorial design CH4 C2H2 4 2 2 method and response surface procedure were established in Estimates of the experimental variable effects were calculated following the procedure recommended by Box et order to identify the key variables in influencing the ηCFC-12 al. (21), and the results are given in Table 2. According to the and the FCH4+C2H2. The design of factors and levels for the first 24-1 fractional factorial experiments are listed in Table 1. The basic assumption of Box et al., higher order interaction effects (above three-factor interaction) being often negligible and ηCFC-12 was higher than 90% when the input power wattage was higher than 100 W; therefore, a higher input power ignored, the results of Table 2 reveal that the input power wattage (factor A), 120 W, was chosen. Soot formation and wattage (A) and CFC-12 feeding concentration (D) were the white deposition led to the decreasing of carbon balance key variables by contrast with other factors and that interaction effects did not matter for the ηCFC-12 or the becoming apparent when the H2/CFC-12 ratio moved away F + . To quantitatively elucidate the effects of all study from 2.0. Accordingly, the high level of factor B was chosen CH4 C2H2 as 2.0. Dimmer plasma was found at the operational pressure variables, the results of fractional factorial design were higher than 20 Torr, which was chosen as the high level in subjected to regression analysis.