Environ. Sci. Technol. 2003, 37, 5021-5026
High-Performance Ti/BDD Electrodes for Pollutant Oxidation XUEMING CHEN, GUOHUA CHEN,* FURONG GAO, AND PO LOCK YUE Department of Chemical Engineering, The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong, China
Deposition of stable boron-doped diamond (BDD) films on Ti substrates is believed to be very difficult. In the present study, the stability of Ti/BDD electrodes has been significantly improved by using an organic additive, CH2(OCH3)2. The improved electrodes had service lives of 175264 h under accelerated life test conditions, which are 2.33.0 times longer than the service lives of electrodes prepared with the conventional H2 + CH4 mixture. Raman spectroscopy, X-ray diffraction (XRD), and scanning electron microscopy (SEM) examinations demonstrated that the films had well-defined diamond features. The current efficiency (CE) obtained on Ti/BDD was 46.9-78.5% in oxidizing acetic acid, maleic acid, phenol, and dyes, which is 1.6-4.3-fold higher than that obtained on the typical Ti/ Sb2O5-SnO2 electrode. We used a Ti/BDD electrode prepared with H2 + CH4 + CH2(OCH3)2 for over 300 h; its activity remained superior. The successful development of stable and active Ti/BDD electrodes significantly increases the feasibility of industrial applications of anodic oxidation in wastewater treatment.
Introduction In recent years, anodic oxidation has been the focus of much research attention in wastewater treatment because of its high oxidation efficiency, fast reaction rate, and ease of operation. It is well-known that the current efficiency (CE) of an oxidation process depends strongly on the electrode properties. In general, the anodes must have high O2 evolution overpotential. Otherwise, a large amount of the supplied current will be wasted in producing O2, leading to a low CE. Therefore, common electrodes such as graphite, platinum, and dimensionally stable anodes (DSA), which do not have a sufficient O2 evolution overpotential, are not good for use in pollutant oxidation. In fact, there are very few electrodes currently available that have high O2 evolution overpotential. These few are mainly limited to PbO2, SnO2, TiO2, and the recently developed boron-doped diamond (BDD) electrodes (1-4). Nevertheless, the electrochemical stability of the first three types of electrodes is rather poor. BDD electrodes are very effective in oxidizing various compounds such as phenol (5, 6), cyanide (7), carboxylic acids (8), 3-methylpyridine (9), 2-naphthol (10), 4-chlorophenol (11), 4-chlorophenoxyacetic (12, 13), benzoic acid (14), polyacrylates (15), and dyes (16). More importantly, BDD films deposited on Si, Ta, Nb, and W by chemical vapor deposition (CVD) have exhibited good stability. The service life of Nb/BDD, for instance, was shown to be over 850 h in * Corresponding author phone: (852)2358-7138; fax: (852)23580054; e-mail:
[email protected]. 10.1021/es026443f CCC: $25.00 Published on Web 09/27/2003
2003 American Chemical Society
an accelerated life test (5). Despite the good quality of the BDD electrodes using Si, Ta, Nb, and W as substrates, their wide application in wastewater treatment is impossible because of the poor mechanical strength and low conductivity of Si and the unacceptably high costs of Ta, Nb, and W. Ti possesses all the important attributes needed in a substrate material for thin film electrodes: good conductivity, high mechanical strength, and electrochemical inertness. Ti is much cheaper than Ta, Nb, and W. Actually, this metal has been widely used in DSA for over 30 years. In recent years, a few researchers have reported on the preparation and characterization of Ti/BDD electrodes. Fisher and co-workers (17) found that the films deposited on Ti substrates had good diamond quality and that the [Fe(CN)6]4-/[Fe(CN)6]3- redox couple demonstrated a quasi-reversible feature on Ti/BDD electrodes. Beck and co-workers (4, 18) investigated cyclic voltammetric behavior on Ti/BDD electrode and found that the onset potential for O2 evolution on Ti/BDD electrodes was over 2.5 V vs normal hydrogen electrode (NHE) in 1 M H2SO4 solution. It should be noted, however, that deposition of stable BDD films on Ti substrates with CVD from the conventional gas mixture of H2 and CH4 is very difficult. Fryda and co-workers (5) reported that BDD films could be deposited on Ti substrates under proper conditions with sufficient adhesion, but cracks appeared, which led to the delamination of the diamond films under electrochemical attack at high loads. In our previous study, we found that the service life of Ti/BDD electrodes could be improved by selecting proper CVD conditions (19). But the stability of the Ti/BDD electrodes obtained under the optimal conditions, namely, substrate temperature (Tsub) ) 850 °C, filament temperature (Tfil) ) 2120-2150 °C, CH4 ) 0.8%, filamentsubstrate distance (dfil-sub) ) 8 mm, and deposition time (tdep) ) 15 h, is still poor. The service life is only 95 h in the accelerated life test (10 000 A/m2, 3 M H2SO4), much lower than that of Nb/BDD electrodes reported by Frady and coworkers (5). In this paper, we seek to demonstrate significant enhancement of the Ti/BDD stability by using an organic additive, CH2(OCH3)2, to characterize the electrodes, to reveal the stability enhancement mechanisms, and to examine the electrode activity in pollutant oxidation.
Experimental Section Electrode Preparation. Hot filament CVD (HFCVD) was used to prepare diamond film electrodes. A schematic diagram of the experimental apparatus is shown in Figure 1. A quartz bell, 300 mm high and 120 mm in diameter, was used as a deposition chamber. Ta coils with 9 turns and 6 mm in diameter, made from φ 0.5 mm Ta wire, was used as the filament. Ti disks, 2 mm thick and 12.7 mm in diameter, were used as the substrates. Prior to diamond deposition, the substrates were sandblasted first, then scratched with diamond paste (1 µm, Kemet International Ltd., UK), cleaned ultrasonically with acetone, and finally rinsed with deionized water. The distance between the filament and the substrate was 5 or 8 mm. The H2 flow was fixed at 100 sccm by a mass flow controller (1179A52CS1BV, MKS). The CH4 flow was adjusted according to the concentration desired by a mass flow controller (1179A12CS1BV, MKS). Both controllers were calibrated using a soap film flow meter (0101-0113, 1-10-100 mL, Hewlett-Packard). B(OCH3)3 was added into the reactive gases as a boron dopant. The concentration of the boron in the reactive gases was about 16 ppm, which was determined by measuring the B(OCH3)3 consumption over a period of 344 h. CH2(OCH3)2 was used as the organic additive; its mass VOL. 37, NO. 21, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5021
FIGURE 2. Accelerated life test comparison between Ti/BDD electrodes deposited with and without addition of CH2(OCH3)2 in 3 M H2SO4 solution under 10 000 A/m2 at 50 °C.
FIGURE 1. Schematic diagram of HFCVD apparatus. 1-gauges; 2-CH4 mass flow controller; 3-H2 mass flow controller; 4-organic additive container; 5-B(OCH3)3 supply; 6-copper pipe; 7-Ta filament; 8-Ti substrate; 9-Nb support; 10-Ta heating element; 11-ceramic support; 12-thermal couple; 13-quartz bell. flow rate was 13.6 mg/h, which was determined by measuring the CH2(OCH3)2 consumption over a period of 60 h. The temperatures of the filament and the substrates were measured with a pyrometer (Model SR35015-0-0-0, Ircon, U.S.A.) and a thermal couple (Model 89000-05, Cole-Parmer, U.S.A.), respectively. Since the thermal couple was located 2.5 mm beneath the substrate surface, the measured temperatures were usually lower than the real values at the substrate surface. To compensate for the temperature differences, calibration was carried out by measuring the melting points of a lead disk and an aluminum disk. More details about the substrate temperature calibration can be found elsewhere (19). For comparison purposes, anodic oxidation of pollutants on Ti/Sb2O5-SnO2, one of the most active oxide-coated electrodes, was also investigated. The Ti/Sb2O5-SnO2 electrode had dimensions of 25 mm × 24 mm × 1.6 mm and was prepared using a standard thermal decomposition method (20). The molar ratio of Sb to Sn in the precursor was 1.5: 98.5. Accelerated Life Tests. Accelerated life tests following the procedure described by Hutchings and co-workers (21) were used to assess the stability of the electrodes. The tests were conducted in 3 M H2SO4 solution at 50 °C. The working electrode was installed on a Ti holder. Pt wire was used as a counter electrode; Ag/AgCl, KCl (sat.) served as a reference electrode. A DC power supply (PD110-5AD, Kenwood, Japan) was used to provide a constant current density of 10 000 A/m2. The potential of the working electrode was periodically monitored. Due to the generation of a large amount of bubbles, use of a Luggin capillary was impossible. However, the reference electrode was placed as close as possible to the working electrode to reduce the ohmic drop from the solution. Characterization. The quality of the diamonds was characterized using Raman spectroscopy (3000, Renishaw, UK). The coating microstructure was analyzed by X-ray diffraction (XRD, PW1830, Philips, The Netherlands). The morphology of the diamond films was examined by scanning electron microscopy (SEM, JSM-6300F, JEOL, Japan). The cyclic voltammetric behavior was investigated using a potentiostat/galvanostat (PGSTAT 100, Atuolab, The Netherlands). 5022
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 21, 2003
Anodic Oxidation. The Ti/BDD electrode was used as the anode, which was fixed on a Ti holder. A 25 mm × 25 mm × 1.6 mm stainless steel plate was used as the cathode. The electrochemical reactor was 130 mm high and 30 mm in diameter and was operated in the batch mode. The volume of the solution was 30 mL. Na2SO4 (2000 mg/L) was used as the supporting electrolyte. The temperature was controlled at 30 °C using a water bath; the current density was controlled at 100 A/m2 or 200 A/m2 using the potentiostat/galvanostat.
Results and Discussion Stability Enhancement. Diamond films are usually deposited with a mixture of H2 and CH4. To enhance Ti/BDD stability, a third component, CH2(OCH3)2, was added in the present work. Figure 2 compares the potential changes with time in the accelerated life tests for two different Ti/BDD electrodes. For the electrode prepared with the conventional H2 + CH4 mixture, the potential maintained almost unchanged for about 60 h and then increased gradually. A quick potential increase was observed in the last 10 h due to the delamination of the BDD film. The service life of an electrode is defined as the operational time at which the potential increases quickly to 5 V vs NHE. Therefore, the electrode prepared with H2 + CH4 had a service life of 95 h. Obviously, the electrode prepared with H2 + CH4 + CH2(OCH3)2 was much more stable. Its service life reached 264 h. Table 1 compares the service lives of electrodes prepared with different gas mixtures under different conditions. All electrodes prepared with H2 + CH4 + CH2(OCH3)2 were more stable than those prepared with H2 + CH4. Addition of CH2(OCH3)2 to the reactive gases increased the service life of the electrode by a factor of 2.3-3.0 depending on the HFCVD conditions. Table 1 also shows that proper selection of HFCVD conditions is very important because both sorts of electrodes prepared under the conditions called II had longer service lives. The service life difference of the electrodes prepared under the different conditions might result from the difference of film quality, thickness, and compression stress (19). It should be noted that the service lives shown in Table 1 were obtained in the accelerated life tests performed at a current density of 10 000 A/m2. The real service life of a Ti/ BDD electrode will be much longer under normal operational conditions, usually with a current density of less than 500 A/m2. This is because the service life increases significantly when the current density decreases. For example, the service life of Ti/Sb2O5-SnO2 was only 12 h at a current density of 1000 A/m2 (22). But when the current density was reduced to 60-200 A/m2, a Ti/Sb2O5-SnO2 could be used for over
TABLE 1. Service Life Comparison of Ti/BDD Electrodes in Accelerated Life Tests HFCVD conditions
H2 + CH4 (h)
H2 + CH4 + CH2(OCH3)2 (h)
I. Tsub 770 °C, Tfil 2000 °C, CH4 0.5%, dfil-sub 5 mm, tdep 20 h II. Tsub 850 °C, Tfil 2120-2150 °C, CH4 0.8%, dfil-sub 8 mm, tdep 15 h
59 95
175 223, 264
could promote diamond deposition significantly (31-34). In fact, CH3COCH3 and CH2Cl2 were first investigated in the present study. CH2(OCH3)2 was found to give a much longer electrode service life. It is generally believed that •CH3 and •H radicals play very important roles in diamond growth. The promotion of diamond deposition by CH2(OCH3)2 can be explained by the increase in the rate of producing •CH3 radicals and the generation of additional •OH radicals. For the conventional H2 + CH4 mixture, •H and •CH3 radicals are generated through the following reactions:
H2 f 2•H
(1)
CH4 + •H f •CH3 + H2
(2)
The mechanism occurring with the addition of •CH3 to the diamond structure is complex and has not been clearly explained. The GDSB mechanism (35) is now widely believed to be a principal route for dimer opening and carbon insertion:
FIGURE 3. XRD patterns of Ti/BDD electrodes deposited with (a) H2 + CH4 and (b) H2 + CH4 + CH2(OCH3)2. 1000 h without damage (23). Therefore, the Ti/BDD electrodes obtained in the present study are believed to be stable enough for application at low current densities. XRD Analysis. Figure 3 shows the XRD patterns. The film deposited with H2 + CH4 + CH2(OCH3)2 exhibited more intensive diamond crystal peaks than that deposited with H2 + CH4, indicating that CH2(OCH3)2 could promote diamond deposition. Apart from the diamond peaks, TiC peaks were also detected, revealing the existence of a TiC layer between the Ti substrate and the diamond film. This is consistent with the observations by Peng and Clyne (24) and Yan et al. (25). Moreover, addition of CH2(OCH3)2 resulted in a decrease in the TiC peak intensities, indicating that TiC formation was hindered in the presence of CH2(OCH3)2. This is attributed to an increase in the diamond deposition rate. It should be noted that the promotion of the diamond deposition by oxygen-containing substances is not a new finding. It was reported that addition of a small amount of H2O (26) or O2 (27, 28) to the conventional H2 + CH4 mixture or use of oxygen-containing compounds such as CH3OH, C2H5OH, and CH3COCH3 (29) and CO (30) as carbon sources could increase the diamond growth rate significantly. The HFCVD of diamond on Si using the mixture of H2+ CH3COCH3, for instance, had a growth rate of 8-10 µm/h, over 10 times faster than that using the conventional H2 + CH4 mixture (29). In addition, it was reported that use of chlorinated hydrocarbons (CH2Cl2, CHCl3, CCl4, C2H5Cl) or fluorinated hydrocarbons (CF4 and CHF3) as carbon sources
Obviously, the rate of the diamond growth depends on the concentrations of •CH3 and •H radicals available at the depositing surface. The higher the concentrations of •CH3 and •H, the faster the diamond growth rate. When CH2(OCH3)2 is added, in addition to reactions 1 and 2, there are the following possible reactions to generate •CH and •OH radicals: 3
CH2(OCH3)2 f •CH2OCH3 + •OCH3 •
(3)
CH2OCH3 f CH2O + •CH3
(4)
OCH3 + H+ f •CH3 + •OH
(5)
•
The bond dissociation energy of C-O is only 351 kJ/mol (36), much lower than that of CH3-H, 435 kJ/mol (37), indicating easier generation of •CH3 by thermal decomposition of CH2(OCH3)2. Moreover, the bond dissociation energy of HO-H is 498 kJ/mol (37), larger than that of H-H, 435 kJ/mol (37), suggesting that •OH is more effective than •H in abstracting H from the diamond surface. Therefore, addition of CH2(OCH3)2 could promote diamond deposition. Morphology Examination by SEM. Figure 4 displays the typical SEM images of the BDD films deposited with different gas mixtures. Highly faceted diamond crystals can be clearly observed in both films. However, the crystallite size is different. The crystals deposited with H2 + CH4 had an average size of about 2 µm. When CH2(OCH3)2 was added, the average VOL. 37, NO. 21, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5023
FIGURE 4. Typical SEM images of the films deposited with (a) H2 + CH4 and (b) H2 + CH4 + CH2(OCH3)2. crystallite size was reduced to about 1 µm only. The decrease in the diamond crystallite size is probably associated with the increase in the diamond deposition rate and the incorporation of a small amount of sp2 carbon impurities in the diamond crystals. Raman Analysis. Raman spectroscopy is one of the best tools to assess the qualities of the diamond films (38). Figure 5 shows the Raman spectra of films deposited with two different gas mixtures. Strong diamond peaks at 1336 cm-1 for H2 + CH4 and 1334 cm-1 for H2 + CH4 + CH2(OCH3)2 were detected. Apart from the diamond peaks, weak and broad bands were detected, which were from amorphous or graphitic sp2 carbon impurities. Although addition of CH2(OCH3)2 increased the sp2 carbon band intensity significantly, the total sp2 carbon content in the film is essentially quite low. It was reported that the cross-sectional scattering coefficients for diamond and graphite are 9 × 10-7 cm-1/sr and 500 × 10-7 cm-1/sr, respectively (38). By normalizing the band intensities with these coefficients, the sp2 carbon content in the film deposited with H2 + CH4 + CH2(OCH3)2 is estimated to be less than 0.5%, indicating good diamond quality. It was found that addition of CH2(OCH3)2 resulted in a decrease in the diamond peak shift (the natural diamond has a peak around 1332 cm-1). Since the peak shift is a consequence of the compression stress caused by the large difference in the thermal expansion coefficients between a diamond film and a Ti substrate (17), the film deposited with H2 + CH4 + CH2(OCH3)2 had smaller compression stress than that deposited with H2 + CH4. This may be associated with a decrease in the crystallite size. Stabilization Mechanisms. It has been well established that the failure of Ti/BDD electrodes results from the BDD 5024
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 21, 2003
FIGURE 5. Raman spectra of the films deposited with (a) H2 + CH4 and (b) H2 + CH4 + CH2(OCH3)2. film delamination. To know if the delamination was caused by insufficient film adhesion, we wiped new Ti/BDD electrodes with soft paper strenuously. But no film delamination occurred. In addition, we vibrated the electrodes in a strong ultrasonic environment for 2 h. No film delamination occurred. We posited that the BDD films had sufficient adhesion to Ti substrates. To understand the delamination mechanism, the electrolyte was analyzed with Inductively Coupled Plasma (ICP) after an accelerated life test. Over 10 mg/L Ti ions were detected. Since Ti metal is anodically resistive, the Ti ions detected must be a consequence of the electrochemical corrosion of the TiC layer formed during BDD film deposition. This corrosion can undermine the BDD film, leading to the film delamination. The increased stability of Ti/BDD electrodes by addition of CH2(OCH3)2 is first attributed to a decrease in the thickness of the TiC layer and an increase in the thickness of the BDD film as demonstrated in the XRD analysis. As the soluble TiC layer becomes thinner and the protective BDD film becomes thicker, the corrosion is weakened, and, accordingly, the delamination is delayed, leading to an increase in the electrode service life. In addition, the diamond crystallite size is decreased as shown in Figure 4. The penetration rate of the electrolyte across the BDD film decreases. This helps to reduce the corrosion rate and therefore increases the electrode service life. Moreover, the increased stability of the Ti/BDD electrodes by addition of CH2(OCH3)2 is related to a decrease in the BDD film compression stress as demonstrated by the Raman analysis. The decrease in the compression stress makes the film less prone to delaminate, and the electrode service life therefore increases.
that obtained on the Ti/Sb2O5-SnO2 electrode. The high CE of the Ti/BDD electrode for pollutant oxidation is attributed to the difficulty of O2 evolution on Ti/BDD. Actually, anodic oxidation of pollutants involves a series of complicated reactions at the anode surface (S). According to a generalized mechanism for oxidizing organic pollutants in the potential region of O2 evolution, the first step is the split of water to produce adsorbed hydroxyl radicals (40):
S + H2O f S(•OH) + H+ + e-
(7)
The adsorbed hydroxyl radicals can then react with pollutants to produce CO2, H2O, etc.
R + S(•OH) f S + CO2 + H2O + ‚‚‚ FIGURE 6. Cyclic voltammogram development of the stable Ti/BDD electrode at a scan rate of 0.1 V/s in 0.5 M H2SO4 solution.
pollutant acetic acida maleic acida phenolb orange IIa reactive red HE-3Ba a
Ti/Sb2O5-SnO2
initial final charge COD COD CE (Ah/L) (mg/L) (mg/L) (%) 5.53 6.43 4.85 6.25 6.25
Current density ) 200
1090 1230 1175 1120 920
A/m2. b
33 46 39 95 45
64.0 61.7 78.5 54.9 46.9
final COD (mg/L)
CE (%)
756 557 450 814 714
20.2 35.0 50.1 16.4 11.0
Current density )100
A/m2.
Cyclic Voltammetric Behavior. Figure 6 displays the cyclic voltammograms (CV) obtained at a Ti/BDD electrode prepared using the gas mixture of H2, CH4, and CH2(OCH3)2 in a 0.5 M H2SO4 solution. Relatively large anodic voltammetric current from an electrochemical “etching” or surface cleaning effect (39) was detected initially. The anodic voltammetric current decreased quickly, however. After about 20 cycles, the shape of CV became identical. There was a wide window ranging from about -0.5 to 2.7 V vs NHE, within which the voltammetric current was very small. The increased currents observed below -0.5 V vs NHE and above 2.7 V vs NHE were attributed to H2 evolution and O2 evolution, respectively. The high onset potential for O2 evolution suggests that Ti/BDD electrodes have high CE for pollutant oxidation. Activity. Good electrodes should not only be stable but also effective in oxidizing various pollutants. In this work, acetic acid, maleic acid, phenol, and two typical types of dyes, i.e. orange and reactive red HE-3B, were employed as model pollutants to examine the activity of the stable Ti/ BBD electrode prepared. For comparison purposes, oxidation with an active oxide-coated electrode, Ti/Sb2O5-SnO2, was also investigated. Table 2 summarizes the experimental results. CE values were calculated according to eq 6
CE )
∆COD × F × 10-3 × 100% 8 × 3600 × q
Meanwhile, the adsorbed hydroxyl radicals can also be further oxidized to generate O2 gas:
S(•OH) f S + 1/2O2 + H+ + e-
TABLE 2. Comparison of Ti/BDD with Ti/Sb2O5-SnO2 for Pollutant Oxidation Ti/BDD
(8)
(6)
where ∆COD is the difference between initial and final COD (mg/L); F is Faraday’s constant (96 487 Coulomb/mole electrons); and q is the charge passed (Ah/L). Obviously, Ti/BDD is much more effective than Ti/Sb2O5SnO2 in oxidizing various pollutants. CE obtained on the Ti/BDD electrode is 46.9-78.5%, 1.6-4.3 times higher than
(9)
Obviously, the more difficult this reaction 9, the higher the CE. The onset potential for O2 evolution on Ti/Sb2O5-SnO2 is 1.9 V in 0.5 M H2SO4 solution (41). In contrast, the Ti/BDD electrode had an onset potential of 2.7 V in 0.5 M H2SO4 solution. That explains the much higher CE of the Ti/BDD electrode than the Ti/Sb2O5-SnO2 electrode for pollutant oxidation. A Ti/BDD electrode has been used for over 300 h; its activity remained superior.
Literature Cited (1) Feng, J.; Johnson, D. C. J. Electrochem. Soc. 1991, 138, 33293337. (2) Kotz, R.; Stucki, S.; Carcer, B. J. Appl. Electrochem. 1991, 21, 14-20. (3) Kesselman, J. M.; Weres, O.; Lewis, N. S.; Hoffmann, M. R. J. Phys. Chem. B 1997, 101, 2637-2643. (4) Beck, F.; Kaiser, B.; Krohn, H. Electrochim. Acta 2000, 45, 46914695. (5) Fryda, M.; Herrmann, D.; Schafer, L.; Klages, C. P.; Perret, A.; Haenni, W.; Comninellis, C.; Gandini, D. New Diamond Front. Carbon Technol. 1999, 9, 229-240. (6) Iniesta, J.; Michaud, P. A.; Panizza, M.; Cerisola, G.; Aldaz, A.; Comninellis, C. Electrochim Acta 2001, 46, 3573-3578. (7) Perret, A.; Haenni, W.; Skinner, N.; Tang, T. M.; Gandini, D.; Comninellis, C.; Correa, B.; Foti, G. Diam. Relat. Mater. 1999, 8, 820-823. (8) Gandini, D.; Mahe, E.; Michaud, P. A.; Haenni, W.; Perret, A.; Comninellis, C. J. Appl. Electrochem. 2000, 30, 1345-1350. (9) Iniesta, J.; Michaud, P. A.; Panizza, M.; Comninellis, C. Electrochem. Commun. 2001, 3, 346-351. (10) Panizza, M.; Michaud, P. A.; Cerisola, G.; Comninellis, C. J. Electroanal. Chem. 2001, 507, 206-214. (11) Rodrigo, M. A.; Michaud, P. A.; Duo, I.; Panizza, M.; Cerisola, G.; Comninellis, C. J. Electrochem. Soc. 2001, 148, D60-D64. (12) Boye, B.; Michaud, P. A.; Marselli, B.; Dieng, M. M.; Brillas, E.; Comninellis, C. New Diamond Front. Carbon Technol. 2002, 12, 63-72. (13) Boye, B.; Michaud, P. A.; Marselli, B.; Dieng, M. M.; Brillas, E.; Comninellis, C. New Diamond Front. Carbon Technol. 2002, 12, 73-81. (14) Montilla, F.; Michaud, P. A.; Morallon, E.; Vazquez, J. L.; Comninellis, C. Electrochim. Acta 2002, 47, 3509-3513. (15) Bellagamba, R.; Michaud, P. A.; Comninellis, C.; Vatistas, N. Electrochem. Commun. 2002, 4, 171-176. (16) Chen, X. M.; Chen, G. H.; Yue, P. L. Chem. Eng. Sci. 2003, 58, 995-1001. (17) Fisher, V.; Gandini, D.; Laufer, S.; Blank, E.; Comninellis, C. Electrochim. Acta 1998, 44, 521-524. (18) Beck, F.; Krohn, H.; Kaiser, W.; Fryda, M.; Klages, C. P.; Schafer, L. Electrochim. Acta 1998, 44, 525-532. (19) Chen, X. M.; Chen, G. H.; Yue, P. L. J. Electrochem. Soc. Submitted for publication. (20) Chen, X. M.; Chen, G. H.; Yue, P. L. J. Phys. Chem. B 2001, 105, 4623-4628. VOL. 37, NO. 21, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5025
(21) Hutchings, R.; Muller, K.; Kotz, F.; Stucki, S. J. Mater. Sci. 1984, 19, 3987-3994. (22) Correa, B.; Comninellis, C.; Battisti, A. D. J. Appl. Electrochem. 1997, 27, 970-974. (23) Lipp, L.; Pletcher, D. Electrochim. Acta 1997, 42, 1091-1099. (24) Peng, X. L.; Clyne, T. W. Thin Solid Films 1997, 93, 261-269. (25) Yan, B.; Loh, N. L.; Fu, Y. J. Mater. Sci. Lett. 2000, 19, 1139-1141. (26) Saito, Y.; Sato, K.; Tanaka, K.; Fujita, K.; Matuda, S. J. Mater. Sci. 1988, 23, 824-846. (27) Mucha, J. A.; Flamm, D. L.; Ibbotson, D. E. J. Appl. Phys. 1989, 65, 3448-3452. (28) Chen, C. F.; Huang, Y. C.; Hosomi, S.; Yoshida, I. Mater. Res. Bull. 1989, 24, 87-94. (29) Hirose, Y.; Terasawa, Y. Jpn. J. Appl. Phys. 1986, 25, L519-L521. (30) Saito, Y.; Sato, K.; Gomi, K.; Miyadera, H. J. Mater. Sci. 1990, 25, 1246-1250. (31) Hong, F. C. N.; Hsieh, J. C.; Wu, J. J.; Liang, G. T.; Hwang, J. H. Diam. Relat. Mater. 1993, 2, 365-372. (32) Hentschel, F.; Schmidt, I.; Benndorf, C. Thin Solid Films 1996, 290-291, 196-199.
5026
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 21, 2003
(33) Schmidt, I.; Hentschel, F.; Benndorf, C. Solid State Ionics 1997, 101-103, 97-101. (34) Schmidt, I.; Benndorf, C. Diam. Relat. Mater. 1998, 7, 266-271. (35) Garrison, B. J.; Dawnkaski, E. J.; Srivastava, D.; Brenner, D. W. Science 1992, 255, 835-838. (36) Lazar, M.; Rychly, J.; Klimo, V.; Pelikan, P.; Valko, L. Free radicals in chemistry and biology; CRC Press: FL, 1989; p 9. (37) Wade, L. G. Organic chemistry; Prentice Hall: NJ, 1991; p 133. (38) Knight, D. S.; White, W. B. J. Mater. Res. 1989, 4, 385-393. (39) Swain, G. M. J. Electrochem. Soc. 1994, 141, 3382-3392. (40) Comninellis, C. Electrochim. Acta 1994, 39, 1857-1862. (41) Correa, B.; Comninellis, C.; Battisti, A. D. J. Appl. Electrochem. 1996, 26, 683-688.
Received for review December 19, 2002. Revised manuscript received July 16, 2003. Accepted August 21, 2003. ES026443F