Environ. Sci. Technol. 2000, 34, 4401-4406
Photoelectrocatalytic Oxidation of Rose Bengal in Aqueous Solution Using a Ti/TiO2 Mesh Electrode X. Z. LI,* H. L. LIU, AND P. T. YUE Department of Civil and Structural Engineering, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong Y. P. SUN Department of Chemical Engineering, Taiyuan University of Technology, Taiyuan, China
To further improve the photooxidation techniques for water and wastewater purification, a new type of photoelectrode, Ti/TiO2 mesh electrode, was successfully prepared by anodizing Ti mesh in 0.5 M H2SO4 solution. The structural and surface morphology of the Ti/TiO2 electrode was examined by Raman spectroscopy and scanning electronic microscopy (SEM), respectively. The examination results indicated that its structure and properties were affected by its growth rate in the anodization process, and anatase TiO2 was dominant in its composition. The photocatalytic (PC) oxidation and photoelectrocatalytic (PEC) oxidation of rose Bengal (RB) in aqueous solution using the Ti/TiO2 electrode were investigated and compared. The experimental results demonstrated that the PEC oxidation by applying an electrical bias between the Ti/ TiO2 electrode and Pt electrode could significantly enhance the degradation rate of RB compared with the PC oxidation. It was found that the best performance of PEC oxidation was achieved by applying an electrical bias of 0.6 V. The efficiency of PEC oxidation was also compared with that of conventional PC oxidation in TiO2 suspension. The kinetic coefficient k for the PEC oxidation was determined to be between that of conventional PC oxidation in 0.3% and 1% TiO2 suspensions.
Introduction Since the first photocatalytic oxidation study for treating organic contaminants in wastewater was carried out by Carey in 1976 (1, 2), many research reports have shown that the semiconductor TiO2 is an excellent photocatalyzer which can break down most kinds of refractory organic pollutants, such as detergents, dyes, pesticides, and herbicides, etc., under UV light irradiation (3-8). However, it has been also known that this type of photooxidation has two typical defects: the difficulty of separating TiO2 particles from aqueous phase, while the photocatalytic oxidation is applied in an aqueous TiO2 suspension, and a very low quantum yield (less than 10%, generally) (9) due to a rapid recombination between active electrons and holes, in which TiO2 particles behave as short-circuited microelectrodes under band gap excitation (10). This high degree of recombination between photogenerated charges carriers wastes radiation energy. * Corresponding author phone: (852) 2766-6016; fax: (852) 23346389; e-mail:
[email protected]. 10.1021/es000939k CCC: $19.00 Published on Web 09/15/2000
2000 American Chemical Society
FIGURE 1. Scanned photo of raw titanium (Ti) mesh which was enlarged with a ratio of 1:2.5. To solve the problem of TiO2 particle separation from wastewater, many efforts and attempts done by researchers were trying to immobilize TiO2 film on a solid carrier such as sand, glass media, or resins by coating, soaking, precipitating, or spinning methods. However, while these immobilized TiO2 photooxidation processes made the TiO2 separation from water phase much easier, they did not achieve any improvement in quantum efficiency. The photoelectrocatalytic (PEC) oxidation using a Pt electrode coated with TiO2 film was first used as a working electrode in the 1982 (11), in which an electrical bias was applied between anode and cathode and the photogenerated electrons on TiO2 were driven away. The process can prevent charge recombination and results in an extension in lifetime of the active holes. Subsequently, a number of studies demonstrated the efficiency of PEC oxidation in organic degradation. Vindogopal et al. (1993) successfully proved that the photocatalytic oxidation efficiency could be enhanced by applying an anodic bias to a TiO2 particulate film electrode on conducting glass plate to degrade 4-chlorophenol, whiles the platinum rod acts as a counter electrode. In the recent years, several studies for degrading formic acid, amino acids, and 4-chlorocatechol were conducted by other researchers (13, 14, 8). In these studies, photoanodes were prepared by coating TiO2 on a conducting glass that was initially covered by an indium tin-oxide. It was found that electron mass transfer between TiO2 film and supporting carriers was not very efficient due to a poor connection between the two materials. In this study, a Ti/TiO2 mesh electrode as a new type of electrode was produced by forming a microporous TiO2 film on Ti mesh in an anodization process, and rose Bengal was used as a dye chemical for photooxidation study under UV irradiation. The objective of this research was to investigate the structural and surface morphology of the Ti/TiO2 mesh electrode and measure its photooxidation efficiency with respect to the degradation of the selected dye chemical in aqueous solution.
Materials Titanium mesh (purity > 99.6%, nominal aperture 0.19 mm, wire diameter 0.23 mm, wires/inch 60 × 60, open area 20%, twill weave) was purchased from Goodfellow Cambridge Limited, and its scanned photo is shown in Figure 1. Titanium dioxide powder (anatase P25) was purchased from BDH laboratory and has an average particle size of 30 VOL. 34, NO. 20, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4401
FIGURE 2. Molecular structure of rose Bengal.
FIGURE 3. Sketch of the photoreactor equipped with two electrodes of a Ti/TiO2 film electrode: a working electrode as anode and a platinum electrode as cathode. nm and surface area of 50 m2 g-1. Rose Bengal (tetraiodnate 4,5,6,7- tetrachlorofluorescein) was purchased from the Aldrich Chemical Company, and its molecular structure is shown in Figure 2. Other chemicals were obtained as analytical grade reagents and used without further purification. Doubly distilled water was used throughout the experiment.
Equipment A bench-scale photoreactor system as shown in Figure 3. consisted of a cylindrical quartz cell with the size of 25 mm in diameter and 50 mm in hight, a 20 W UV lamp (NEC T10 BLACKLIGHT) with a maximum UV irradiation peak at 365 nm, and a potentiostat (ISO-TECH 1PS 1810H). The photoreactor has an effective volume of 10 mL, in which two electrodes of Ti/TiO2 mesh electrode as anode and Pt wire electrode (40 mm in length with a 0.4-mm diameter) as cathode were placed in the center of the reactor in parallel and connected with the potentiostat.
Experimental Section Preparation of Ti/TiO2 Mesh Electrode. A large piece of raw titanium mesh was cut into small rectangle pieces of 25 mm × 10 mm × 0.5 mm as shown in Figure 1, which was then cleaned with alcohol and a subsequent rinse in acetone. The treated Ti mesh and a copper plate having the same size were submerged in a beaker with 1 L of 0.5 M H2SO4 solution (17), and an electrical current was applied by using a laboratory-made DC power supply with a maximum potential of 200 V. The anodization process was conducted in two stages, in which the galvanostatic anodization with a constant current density was first performed until a designated anodeto-cathode voltage (120, 140, 160, and 180 V) was reached. Then the constant voltage was maintained until the end of anodization, while the current was decreasing gradually. Since significant heat was generated during the anodization process, the temperature of H2SO4 solution increased from 22 to 24 °C (room temperature) to 36-40 °C. The freshly generated TiO2 mesh electrode was then rinsed by distilled water and dried in an oven at 105 °C for half an hour. Characterization of Ti/TiO2 Mesh Electrode. Before the Ti/TiO2 mesh electrode was used for the photooxidation study, its structural and surface morphology were examined 4402
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 20, 2000
and studied by using Raman spectroscopy and scanning electron microscopy, respectively. Photooxidation of RB. In the study, five sets of tests were carried out with different objectives. The first set of tests using the different Ti/TiO2 electrodes generated with different working potentials of 120, 140, 160, and 180 V was carried out for 3.5 h under UV irradiation with the intensity of 2.2 mW cm-2 to determine their photooxidation efficiency. The second test using the 160 V Ti/TiO2 electrode to degrade RB in its aqueous solution was carried out with an initial concentration of 20 mg L-1 under the UV irradiation of 1.4 mW cm-2 for 6 h to determine the degree of RB degradation. The third set of tests was performed with different experimental conditions including UV irradiation and an electrical bias to identify the critical conditions. The fourth set of tests was carried out by applying different electrical potentials between the two working electrodes to determine the optimum electrical bias which should be applied. The fifth set of tests utilized conventional PC oxidation in TiO2 suspensions with different concentrations from 0.1% to 1% in order to compare the PEC oxidation with conventional PC oxidation in TiO2 suspensions.
Analytical Methods Structural Analysis. Raman spectroscopy was used to determine the phase of TiO2 (anatase and rutile) on the mesh electrode. At room temperature, Raman spectra were excited by a 514.5 nm laser line from a CW argon laser (Coherent Innva 70). The laser power was kept at 250 mW to avoid laser annealing of the samples. A 55 mm f/1.8 lens was used for collecting the scattering light which was dispersed and detected using a double grating monochromator (Spex 1403) equipped with a cooled photomultiplier tube (PMT, Hamamatus R943-2). All spectra were recorded in a small angle scattering geometry. The resolution obtained was expected to be as good as 1 cm-1. Surface Analysis. Scanning electron microscopy (SEM) (Leica Stereoscan 400i Series) was used to study the surface morphology, average pore size, and pore distribution. High tension was selected at 15 kV. Light Intensity Measurement. The intensity of UV irradiation was measured by a Black-Ray UV meter (UVP Inc., model J 221). Chemical Analyses. The concentration of RB was determined by high performance liquid chromatography (HPLC), which includes a ISCO model 2350 High Pump, an ISCO RESTEK Pinnacle Octylamine Column (5 µm 250 × 4.6 mm), and a UV detector with a detecting wavelength of 550 nm. The mobile phase used was aqueous solution with 40% acetonitrile. TOC in wastewater samples was analyzed by a TOC analyzer (SHIMADZU TOC 5000A). The absorbency of RB in its aqueous solution was also analyzed by a UV spectrometer (MILTON ROY Spectronic Genesys 2).
Results and Discussion Preparation of Ti/TiO2 Mesh Electrode in the Anodization Process. In most anodic oxidation processes, metals can dissolve in aqueous acidic solution under enough potential bias or current to form an oxide layer on the metal surface. Theoretically, a higher anodic growth rate is given by a higher current density or activation energy on the electrode. Also, electric field around the electrode is also directly proportional to the current density. In this study, the galvanostatic anodization with a constant current density of 110 mA cm-2 was applied first and characterized by a fast increase of the potential difference (i.e., 1.2 V s-1) across the oxide layer within the first 2 min. This increase then slowed as the effective activation energy of oxide growth reached a near
the cases, while other peaks were broader. Up to 140 V, the Raman spectra showed that the main anatase peaks were increasing while the applied voltage increased. Above 160 V, additional peaks corresponding to the rutile phase of TiO2 were developed. The spectra also indicate that the anatase phase of TiO2 was dominant in all the cases with a voltage growth rate of 1.2 V s-1.
FIGURE 4. Variation of current density and electrical potential applied in the two-stage anodization process. constant level. At this point, the dissolution rate was equal to the oxidation rate. In the second stage, the designated maximum value of a constant voltage (120, 140, 160, and 180 V) was maintained, while the current density gradually decreased from 110 mA cm-2 to 33 mA cm-2. The entire period of anodization lasted for about 10 min. The variation of voltage and current density against time in the anodization process is shown in Figure 4. A linear relationship between voltage and time was found during the initial 1.25 min in all the cases while a constant current density was maintained. The anodic growth rate can be obtained from the curve to be 1.2 V s-1 (up to 1.25 min), when the current density was 110 mA cm-2. However, the slope of voltage curve eventually became more flat due to limited power provided from the DC power supply. The current across the electrodes dropped rapidly since the resistance increased. It was observed that fine gas bubbles were generated from both the anode (Ti mesh) and the cathode (Cu plate), and even a gas film could be seen on the surface of the solution. The generation rate of the gas on the cathode was faster than that on the anode. This was likely due to the fact that most of the oxygen molecules generated at the anode were combined with titanium to form the titanium oxide, while hydrogen gas generated at the cathode was liberated freely. Structural Analysis of Ti/TiO2 Mesh Electrodes. Raw Ti mesh and the Ti/TiO2 mesh electrodes prepared using different voltages were examined by ex-situ Raman spectroscopy. Their Raman spectra are shown in Figure 5, in which the peaks representing the anatase and rutile forms of TiO2 are labeled with A and R, respectively. The phonon peaks of anatase were located at 147, 200, 393, 516, and 640 cm-1, whiles the phonon peaks of rutile were observed at 244, 445, and 612 cm-1. The strongest peak of anatase occurred at around 147 cm-1 with a narrow line shape in all
Surface Morphology of the Ti/TiO2 Electrodes. From a visual observation, the raw Ti mesh was silver in color, the Ti/TiO2 mesh electrode prepared at 120 V was sky blue, the Ti/TiO2 mesh electrode prepared at 140 V was light gray to blue, the Ti/TiO2 mesh electrode prepared at 160 V was light gray, and the Ti/TiO2 mesh electrode prepared at 180 V was gray. When anodizing titanium, the sparking breakdown voltage that could be achieved was 130 V in 0.5 M H2SO4 solution under a galvanostatic mode anodization, in which the color of the oxide layer changed to gray and surface reflectivity decreased rapidly. The differences in color obtained in oxidized films below the sparking voltage was due to light interference. The morphology of the Ti mesh and Ti/TiO2 mesh electrodes were examined by SEM, and their SEM photos are shown in Figure 6. It was found that the surface of TiO2 film was microporous and rougher than that of raw Ti metal. The micropore size measured by SEM was increased from 17 to 60 nm, while the applied voltage in anodization increased from 120 to 180 V. The earlier work on the titanium anodization indicated that the formation of micropores was attributed to the power dissipation in the barrier oxide layer (18). When anodizing current density was greater than 30 mA cm-2 in sulfuric acid, the power dissipated in the barrier oxide layer increased considerably. This current causes local overheating of the oxide, and the cooling of the oxidized anode by the electrolyte becomes insufficient. This higher temperature leads to a higher ionic current probably along grain boundaries, and the oxide may recrystallize and result in a porous surface. With current densities greater than 100 mA cm-2, a fast dissolution of the oxide is believed to occur with the creation of pores, leading to an even rougher surface. Efficiency of Photooxidation Using Different Ti/TiO2 Electrodes. The first set of tests using the different Ti/TiO2 mesh electrodes anodized at 120, 140, 160, and 180 V to degrade the rose Bengal (RB) in aqueous solution were carried out under UV irradiation at the same density of 2.2 mW cm-2 for 3.5 h, respectively. The water samples before and after each treatment were collected and tested using HPLC for determining RB concentration. The removals of RB achieved in the five tests as shown in Figure 7 demonstrated that the Ti/TiO2 electrode anodized at 160 V achieved the best performance for RB oxidation. Therefore the 160 V Ti/TiO2
FIGURE 5. Raman spectra of raw Ti and Ti/TiO2 electrodes, in which anatase is labeled as A and rutile is labeled as R. VOL. 34, NO. 20, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4403
FIGURE 7. PEC oxidation of RB using different Ti/TiO2 electrodes.
FIGURE 8. UV absorbance of RB (20 mg L-1) in the PEC oxidation against irradiation time in which spectrum 1, 2, 3, 4, 5, 6, and 7 denote the results at the irradiation time of 0, 0.25, 0.5, 1.0, 1.5, 4.0, and 6.0 h, respectively, while light intensity was 1.4 mW cm-2 and electrical bias applied was 0.6 V.
FIGURE 9. Reduction of RB in PC and PEC oxidation, in which 2.2 mW cm-2 of light intensity was used in the tests of PEC1 and PC1 and 1.3 mW cm-2 of light intensity was used in the tests of PEC2 and PC2. The initial concentration of RB was 20 mg L-1, and electrical bias of 0.6 V was applied in the PEC oxidation. FIGURE 6. SEM micrograph of raw Ti mesh and Ti/TiO2 mesh in which A is the raw Ti mesh, B is the 120 V Ti/TiO2 mesh, and C is the 180 V Ti/TiO2 mesh. electrode was applied to the following tests for photodegradation of RB. Degradation of RB during PEC Oxidation. The second test was carried out to treat the RB solution with a light intensity of 1.4 mW cm-2 for 6 h. During the test, several water samples were collected at different time intervals and analyzed for their absorbance of light in the range of 400700 nm. The experimental results are shown in Figure 8, 4404
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 20, 2000
which demonstrated that about 65% of RB in the aqueous solution were reduced rapidly during the first 0.5 h and about 85% of RB were removed after 6 h. At the same time, a slight shift of its maximum adsorption peak from 550 to 540 nm during the photodegradation was found. Photooxidation of RB under Different Experimental Conditions. The third set of tests was carried out under different conditions as shown in Figure 9. Test 1 (dark) was conducted in a dark condition without applying any electrical bias between the electrodes. The experiment demonstrated a significant decrease of RB during the first few minutes due to adsorption of RB by the Ti/TiO2 electrode and a recovery of RB in the solution due to desorption of the electrode at
FIGURE 10. Degradation of RB and TOC against irradiation time, in which initial concentration of RB ) 20 mg L-1, light intensity ) 2.2 mW cm-2, and electrical bias ) 0.6 V. the later stage. Tests 2 and 3 were carried out with a light intensity of 1.4 mW cm-2, in which an electrical bias of 0.6 V was applied in test 2 (PEC1) representing the PEC oxidation, while no electrical bias was applied in test 3 (PC1) representing the PC oxidation. Tests 4 (PEC2) and 5 (PC2) were then carried out with a higher light intensity of 2.2 mW cm-2 with and without applying electrical bias, respectively. The experimental results indicated that using the higher intensity of UV irradiation achieved the faster photodegradation rate of RB. In both pairs of the tests, the PEC oxidation demonstrated a significant priority over that of PC oxidation, especially during the first 2 h. The experimental results as shown in Figure 9 indicate that the Ti/TiO2 electrode demonstrated a strong ability to adsorb RB in its aqueous solution. Since the PEC oxidation took place on the catalyst surface, not in the bulk of the solution,17 any accumulation of the adsorbate on the surface of the electrode may result in a decline of photooxidation rate. Therefore, the photooxidation rate on the surface of electrode would be a critical factor affecting the overall reaction rate in the PEC oxidation process. To investigate the mineralization of RB, the TOC concentration was also monitored against reaction time and compared with the variation of RB concentration as shown in Figure 10. During the first 2 h, RB concentration was reduced faster than the TOC concentration significantly, which may indicate the reaction in this stage has only achieved an incomplete mineralization of RB. However, after 6 h reaction, the TOC removal up to 90% was almost the same as the RB removal. This seems to indicate that a complete mineralization of RB can be achieved within 6 h under this experimental condition. While the exact mechanism of the PEC process has not been totally clarified, the adsorption behavior of the TiO2 electrode and the electron mass transfer rate between the Ti electrode and TiO2 film must play very important roles in this system. Photodegradation of RB Affected by Applying Different Electrical Bias. To study the reaction rate of RB in the PEC oxidation, the fourth set of tests was performed to treat the RB solution, including three tests with the same light intensity of 2.2 mW cm-2, but with different electrical biases of 0.3, 0.6, and 1.0 V. The tests lasted for 3.5 h, and the experimental results are shown in Figure 11. The results indicated that 0.6 V may be the best voltage for the overall reaction, but only a slight difference appeared mainly in the first few hours. This behavior is similar to the results of others, which indicated that the effect of applying an electrical bias between the electrodes is more important than the difference of the voltage applied. Comparison of the PEC Oxidation Using the Ti/TiO2 Mesh Electrode with the PC Oxidation in TiO2 Suspension. Although the PEC oxidation may follow a very different reaction mechanism from the PC oxidation, it is important
FIGURE 11. Degradation of RB in the PEC oxidation affected by applying different electrical bias between the electrodes, in which the initial concentration of RB ) 20 mg L-1, and light intensity ) 2.2 mW cm-2.
FIGURE 12. PEC oxidation compared with conventional PC oxidation in TiO2 suspension, in which the initial concentration of RB ) 20 mg L-1, light intensity ) 2.2 mV cm-2, and electrical bias ) 0.6 V.
TABLE 1. Comparison of Coefficient k Values Calculated Based on the Assumption That the Kinetics of Both PEC Oxidation Using the Ti/TiO2 Electrode and PC Oxidation in TiO2 Suspension Follow the First-Order Reaction TiO2 (%)
k (min-1)
correlation coefficient R
0.1 0.2 0.3 1.0 PEC
0.0141 0.0183 0.0216 0.0566 0.0240
0.9968 0.9922 0.9933 0.9781 0.9722
to compare the PEC and conventional PC (TiO2 suspension) oxidation rates. This comparison may provide valuable information leading to a better understanding of the PEC oxidation performance. In the fifth set of tests, TiO2 suspensions were prepared by dispersing TiO2 powder in an aqueous solution of RB. Four TiO2 suspensions with the same RB concentration of 20 mg L-1 but having different TiO2 concentrations of 0.1%, 0.2%, 0.3%, and 1% were treated under the same UV irradiation as PC oxidation (Table 1). Several studies (16, 19) indicated that the rate of photocatalytic degradation of various organic contaminants over illuminated TiO2 could be fit with the Langmuir-Hinshlwood kinetics expression. When the initial concentration of organic contaminants was low, the Langmuir-Hinshlwood rate form reduced to an apparent first-order kinetics form. If we suppose that both PEC and PC oxidation reactions simply VOL. 34, NO. 20, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4405
follow the first-order kinetics under these experimental conditions, the kinetic coefficient k in the tests can be calculated and are listed in Table 1, and four degradation curves against time are shown in Figure 12. The experimental results demonstrated that the reaction rate of RB degradation in the PEC oxidation was faster than that in the PC oxidation with TiO2 concentration below 0.3%, but not as fast as that in the PC oxidation when TiO2 concentrations reached levels of 1%. Although a higher concentration such as 1% used in TiO2 suspension can achieve a higher reaction rate, the separation difficulty would be a factor to consider as possibly limiting. Actually, the concentration of TiO2 suspension used for most PC oxidation is in the range of 0.1-0.3%, which means that the PEC oxidation achieved a higher reaction rate than most PC oxidation processes. In this experiment, the PEC oxidation demonstrated a very efficient reduction of RB during the first hour, which may result from the strong adsorption on the electrode surface. However, to maintain a substantial rate of RB reduction, one must also maintain a significant PEC oxidation rate on the surface of the counter electrode. If oxidation rate is faster than adsorption rate, there would be no significant accumulation of adsorbate on the surface of the electrode and a sustainable degradation can be continued. Otherwise, the accumulation of adsorbate residue on the electrode surface may reduce the efficiency of photooxidation.
Literature Cited (1) Carey, J. H. Bull. Environ. Contam. Toxicol. 1976, 16 (6), 663. (2) Aurian, V. Toxicol. Environ. Chem. 1988, 16, 89. (3) Ollis, D. F.; Pelizzetti, E.; Serpone, N. Environ. Sci. Technol. 1991, 25, 1523. (4) Ollis, D. F.; Al-Ekabi, H. Elsevier Amsterdam 1993, 39. (5) Hidaka, H.; Zhao, J.; Pelizzetti, E.; Serpone, N. J. Phys. Chem. 1992, 96, 2226. (6) Ross, H.; Bendig, J.; Hecht, S. Sol. Energy Mater. Sol. Cells 1994, 33, 475. (7) Ma, Y.; Yao, J. N. J. Photochem. Photobiol. A: Chem. 1998, 116, 167. (8) Hidaka, H.; Kazuhiko, T. S.; Zhao, J.; Serpone, N. J. Photochem. Photobiol. A: Chem. 1997, 109, 165. (9) Choi, A.; Termin, W.; Hoffmann, M. R. J. Phys. Chem. 1994, 98, 13669. (10) Kormann, C.; Bahnemann, D. W.; Hoffmann, M. R. Environ. Sci. Technol. 1991, 25, 494. (11) Ward, M. D.; Bard, A. J.; Bard, A. J. J. Phys. Chem. 1982, 86, 3599. (12) Vinodgopal, K.; Hotchandani, S.; Kamat, P. V. J. Phys. Chem. 1993, 97, 9044. (13) Kesselman, J. M.; Lewis, N. S.; Hoffman, M. R. Environ. Sci. Technol. 1997, 31, 2298. (14) Kim, D. H.; Anderson, M. A. Environ. Sci. Technol. 1994, 28, 479. (15) Mikula, M.; Blecha, J.; Ceppan, M. J. Electrochem. Soc. 1992, 139, 3470. (16) Pruden, A. L.; Ollis, D. F. J. Catal. 1983, 82, 404. (17) Leitner K.; Schultze, J. W. Electrochem. Sci. Technol. 1986, 133 (8), 1561. (18) Matthews, R. W. Water Res. 1990, 24, 653. (19) Ollis, D. F.; Hsiao, C. Y.; Budiman, L.; Lee, C. L. J. Catal. 1984, 88, 89.
Acknowledgments The authors thank the Hong Kong Government Research Grant Committee for financial support to this work under the RGC Grant Q-228/98.
4406
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 20, 2000
Received for review January 31, 2000. Revised manuscript received June 22, 2000. Accepted July 5, 2000. ES000939K
