Degradation of Bisphenol A in Water by TiO2 Photocatalyst

Apr 27, 2001 - Department of Applied Chemistry, School of Engineering, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan, Department ...
14 downloads 19 Views 73KB Size
Download PDF
Environ. Sci. Technol. 2001, 35, 2365-2368

Degradation of Bisphenol A in Water by TiO2 Photocatalyst Y O S H I H I S A O H K O , †,‡ I S A O A N D O , § C H I S A N I W A , ‡,| T E T S U T A T S U M A , † TSUYOSHI YAMAMURA,§ TETSUTO NAKASHIMA,⊥ Y O S H I N O B U K U B O T A , ‡,| A N D A K I R A F U J I S H I M A * ,†,‡ Department of Applied Chemistry, School of Engineering, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan, Department of Chemistry, School of Science, Science University of Tokyo, 1-3 Kagurazaka, Shinjuku-ku, Tokyo 162-8601, Japan, Department of Urology, School of Medicine, University of Yokohama City, 3-9 Fukuura, Kanazawa-ku, Yokohama 236-0004, Japan, Nagareyama Higashi High School, Nazukari, Nagareyama-shi, Chiba 270-0145, Japan, and Kanagawa Academy of Science and Technology, KSP Building East 412, 3-2-1 Sakado, Takatsu-ku, Kawasaki-shi, Kanagawa 213-0012, Japan

The photocatalytic degradation of bisphenol A (BPA), a representative endocrine disruptor, was carried out in TiO2 aqueous suspension. The main purposes were to confirm the total mineralization of BPA and to evaluate the estrogenic activity in the treated water during the photocatalytic reaction. An initial BPA concentration of 175 µM in water was totally degraded to carbon dioxide by TiO2photocatalyzed reactions under UV irradiation of 10 mW cm-2 for 20 h. Four HPLC peaks indicating intermediate products appeared in chromatograms monitored at 275 nm, but the heights relative to that of the initial BPA were very low, at most 0.04 in the time period 5-10 h after the start of UV irradiation. All of the peaks finally disappeared after 20 h. For the treated water, the transcriptional estrogenic activity in response to human estrogen receptor in a yeast hybrid assay decreased drastically to less than 1% of the initial BPA’s activity within 4 h. On the basis of these results, we conclude that TiO2 photocatalysis could be a useful technology for the purification of water containing BPA without generating any serious secondary pollution.

Introduction Recently, several types of environmental pollutants referred to as endocrine disruptors (EDs) have been suggested to be associated with abnormal sexual development and abnormal feminizing responses of animals in a number of reports (112). EDs exhibit biological behavior that is similar to that of natural and synthetic estrogens with steroid structures and their metabolites. Thus, the development of methods to remove these EDs is needed urgently. However, conventional biological methods to treat organic pollutants in wastewater * Corresponding author phone: +81-3-3812-9276; fax: +81-33812-6227; e-mail: [email protected]. † University of Tokyo. ‡ Kanagawa Academy of Science and Technology. § Science University of Tokyo. | University of Yokohama City. ⊥ Nagareyama Higashi High School. 10.1021/es001757t CCC: $20.00 Published on Web 04/27/2001

 2001 American Chemical Society

require long times, and chemical oxidation methods in general cannot completely eliminate total organic carbon (13-16). Photocatalytic reactions of TiO2 particles and films have been studied extensively for the purification of air and water because most harmful and toxic organic pollutants can be completely mineralized to carbon dioxide with the strong oxidizing power of the photogenerated holes of TiO2 (1727). However, as far as we know, no study has yet appeared in which EDs have been decomposed by means of TiO2 photocatalysis, in particular, as carried out together with an evaluation of changes in estrogenic activity. Bisphenol A [2,2-bis(4-hydroxyphenyl)propane or BPA] is mostly used as a raw material for epoxy and polycarbonate resins and is widely suspected to act as an ED (1-12). In the present study, we have investigated the degradation of BPA to CO2 in water as a result of TiO2 photocatalytic reactions and have sought to identify some of the intermediate products by liquid chromatography/mass spectrometric (LC/MS) analysis. Concurrently, we have monitored the estrogenic activity for the treated water during the photocatalytic degradation of BPA using an estrogen screening assay, which involves recombinant yeast cells with the human estrogen receptor (hER).

Experimental Section The BPA used in this study was purchased from Tokyo Kasei (GC grade >99%). BPA aqueous solutions were prepared with doubly distilled water. A 40-mL aqueous solution containing 170 µM (40 mg L-1) BPA was put into a Pyrex reaction vessel (100 mL capacity). TiO2 powder (ST-01, Ishihara Sangyo Kaisha, Ltd., 7-nm particle diameter, 320 m2 g-1 surface area, anatase) was added into the solution to give a concentration of 1.0 g L-1. Prior to the photodegradation experiments, the suspension was stirred for more than 12 h in the dark to achieve adsorption equilibrium of BPA on the TiO2 photocatalyst. The TiO2 suspension containing BPA was irradated with a 200-W Hg-Xe lamp (Luminar Ace 210, Hayashi Tokei) together with a 365-nm band-pass filter. The irradiation intensity was 10 mW cm-2, as determined by use of a UV radiometer (UVR-36, Topcon). The amount of BPA in the aqueous solution was measured by use of a high-performance liquid chromatograph (HPLC, Tosoh), equipped with a Tosoh UV-8010 optical detector and an Inertsil ODS-3 column. The elution was monitored at 275 nm. The eluent used was a mixed solvent of acetonitrile and water (6/5, v/v). The TiO2 photocatalyst was removed from the solution by filtration, and the resulting solution was analyzed with HPLC. The amount of CO2 produced as a final product of the BPA degradation was determined by gas chromatography (model GC-8A, Shimadzu), equipped with a Porapak-Q column, a methanizer, and a flame ionization detector with N2 as the carrier gas. LC/MS analysis, used to identify intermediate products, was carried out with a HP1000 series liquid chromatograph (Hewlett-Packard) equipped with a Quattro II mass spectrometer (Micromass). The mass spectrometer was operated in the negative ion mode in the m/z 100-300 range for LC/ MS and 50-300 for LC/MS/MS. The chromatographic separation was carried out with an L-column ODS-3 (Chemicals Evaluation Research Institute, Japan) at 40 °C. The eluate entered the electrospray interface (source temperature, 130 °C; cone voltage, 60-100 V). The elution was carried out at a rate of 0.1 mL min-1 with acetonitrile/10 mM ammonium VOL. 35, NO. 11, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2365

FIGURE 1. Inverse plot of the solution-phase BPA concentration (C, µmol L-1) and amount of adsorbed BPA (M, µmol (m2 of catalyst)-1) on the TiO2 photocatalyst used for the analysis of the Langmuir-type isotherm. acetate aqueous solution (5:5, v/v). For the LC/MS analysis, the whole suspension, after being photoirradiated for 1 h, was subjected to filtration with a 0.2-µm membrane filter, and the filtrate was reduced in volume by a factor of 100 with a rotary evaporator. Transcriptional estrogenic activities in response to a recombinant yeast cell were evaluated for the treated waters as follows (28, 29). A single yeast colony (Sacchromyces cerevisiae BJ3505) (30) was grown overnight at 30 °C on a growth medium (0.67% yeast nitrogen base, 2% dextose) supplemented with lysine and histidine. Then 250 µL of cell solution, 50 µL of 10 mM CuSO4, and 50 µL of BPA sample solution were added to 10 mL of fresh medium. After being treated overnight at 30 °C, a 150-µL aliquot of the yeast cell suspension was added to 250 µL of saline, and then the cells were collected by centrifugation and permeabilized by the addition of 700 µL of Z-buffer (60 mM Na2HPO4, 40 mM NaH2PO4, 10 mM KCl, 1 mM MgSO4, and 35 mM β-mercaptoethanol), 25 µL of 0.1% sodium dodecyl sulfate, and 50 µL of CHCl3 and mixed for 30 min. It was then brought to room temperature with the addition of 200 µL of o-nitrophenyl β-D-galactopyranoside (4 mg mL-1 in Z-buffer) for color reaction. The reactions were terminated by the addition of 1 M Na2CO3, and the absorbance at 420 nm was measured. β-Galactosidase activity was determined by the formula [A420/ (A600 of cells × reaction time (min) × volume of cells (mL))] × 1000. Because the initial BPA concentration used in these experiments (175 µM) was 20 times higher than that at which the maximum estrogenic activity appears in this assay (ca. 10-5 M), the estrogen activities for the treated water samples were evaluated after dilution by a factor of 20.

Results and Discussion Figure 1 shows the experimental data for the adsorption isotherm in the form of an inverse plot (amount of BPA adsorbed/(m2 of TiO2 photocatalyst)-1, M)-1 (µmol (m2 of catalyst)-1)-1 vs (the solution-phase BPA concentration, C)-1 (µmol L-1)-1. These data were analyzed in terms of a Langmuir-type isotherm, which is described as follows:

M-1 ) µ-1 + (µKC) -1

(1)

where µ is the maximum number of mol of BPA adsorbed/ (m2 of TiO2)-1, and K is the adsorption binding constant (in L µmol-1). The linearity of the plots indicates Langmuir-type adsorption of BPA on the TiO2 photocatalyst. The values of µ and K were ca. 2.0 × 10-2 µmol (m2 of catalyst)-1 and 6.2 × 1010 L µmol-1, respectively. The experimental condition for photodegradation of BPA at 175 µM (40 mg L-1) 2366

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 11, 2001

FIGURE 2. Experimental data for the TiO2 photocatalytic degradation of BPA, including intermediate products and CO2 evolution in terms of moles of carbon (initial BPA concentration, 175 µM; UV intensity, 10 mW cm-2; b, BPA; O, CO2; 2, carbon balance, i.e., products other than generated CO2). corresponded to a coverage of 1 molecule/50 nm2, with only 4% of the total BPA in the solution adsorbed on the TiO2 photocatalyst. UV irradiation was conducted at room temperature after the equilibrium between solution-phase BPA and BPA adsorbed on the TiO2 photocatalyst was achieved. More than 99% of the initial BPA was degraded after 15 h of UV irradiation with first-order kinetics, as shown in Figure 2. In a control experiment, the solution was irradiated without TiO2 photocatalyst, and the disappearance of BPA was negligible. The vertical axis of Figure 2 corresponds to the amounts of each compound expressed as micromoles of carbon. When the photoinduced mineralization of BPA proceeds stoichiometrically using oxygen as an oxidizing agent, the total reaction is given by

C15H16O2 + 18O2 f 15CO2 + 8H2O

(2)

The calculated differences between the amounts of carbon corresponding to the initial amount of BPA and the sum of the amounts corresponding to undecomposed BPA plus generated CO2 were plotted as triangles in the same figure. In the initial stages of the photocatalytic reactions, the amounts of carbon corresponding to generated CO2 were smaller than those corresponding to decomposed BPA. This difference indicates that intermediate products were generated during the photocatalytic reactions, and the corresponding carbon mole percent (with respect to initial BPA) reached at most 20% after 5 h, but these intermediates were completely decomposed to CO2 under UV irradiation for 20 h (31). Figure 3 shows the time dependences of the amounts of several intermediate products formed in the TiO2 suspension during the photocatalytic reactions. Four peaks (labeled a-d in elution order) appeared in the HPLC chromatogram. The vertical axis in Figure 3 corresponds to the heights of the absorption peaks normalized to the height of the initial BPA peak. Maxima were reached after 5 h under UV irradiation for peaks a-c and after 10 h for peak d. However, all of the peak heights were much lower than that of the initial BPA, and all of the peaks disappeared after 20 h of UV irradiation. All four peaks had retention times shorter than BPA, as listed in Table 1, which indicates that they are more polar than BPA. Table 1 lists the main fragments (m/z) and relative abundances (%) obtained for peaks a-d by LC/MS and LC/ MS/MS. The m/z value for each peak corresponds to [M 1]- ions in the negative ion mode of LC/MS. For example, the m/z value of BPA should be 228 but was detected as 227

FIGURE 3. Time course of four peaks appearing in the HPLC trace detected at 275 nm during the TiO2 photocatalytic degradation of BPA in water (0, peak a; 9, peak b; 4, peak c;2, peak d; listed in order of elution). Each peak height was normalized by that of the initial BPA peak. Their detailed characteristics are discussed in the text and listed in Table 1. Experimental conditions were the same as those for Figure 2.

TABLE 1. Main Fragment Ions (m/z) and Relative Abundance (%) Obtained for Peaks a-d by LC/MS and LC/MS/MS Spectra

FIGURE 4. Estrogenic activity of aqueous solutions containing various amounts of BPA (a) and time courses of estrogenic activity and BPA concentration during the photocatalytic treatment of a BPA solution (initial concentration ) 175 µM) (b). Estrogenic activities were measured by means of the yeast hybrid assay (see Experimental Section). (O) BPA solutions without photocatalytic treatment; (b) BPA solutions treated photocatalytically for various periods and then diluted 20-fold (see text for further details); (2) BPA solutions treated photocatalytically for 6 h and then diluted 200-20 000-fold. Experimental conditions of photocatalysis were the same as those for Figure 2, and the dashed line indicates the same data for the BPA concentration profile as in Figure 2. as shown in Table 1, which is explained by the characteristic cleavage of the O-H bond in the hydroxyl group. Each peak was identified by structural elucidation by interpretation of the mass spectra obtained, but unfortunately, the main fragments did not exhibit S/N levels good enough to identify the components precisely. The LC/MS/MS spectrum for peak a contained three fragment ions at m/z ) 135, 163, and 207. If only one of the benzene rings of BPA was cleaved at first by a photocatalytic oxidation reaction, the 207 value could be deduced to be 3-(4-hydroxyphenyl)-3-methyl-2-oxobutanoic acid (HPMOBA). The remaining values can be explained by the loss either of -CO2 to give [(MHPMOBA - 1) - 44] ) 163 or -CO-CO2 to give [(MHPMOBA - 1) - 72] ) 135. No explanation is given for peak b at this stage. The mass fragments for peak c resemble those in the lower mass range for the mass spectrum of BPA itself (Table 1). One possibility is that the fragment with m/z 133 could be assigned to 4-vinylphenol (VP, m/z ) 134). However, the fact that the retention time was shorter than that for peak d, which was identified as 4-hydroxyacetophenone (HAP) (see below), cast some doubt on this assignment because the less polar compound (HAP) would be expected to exhibit a longer retention time. In addition, the fragment corresponding to a mass of 40 [(MVP - 1) - 40 ) 93] cannot be explained at this stage. Peak d exhibited a parent peak at m/z 135, with fragment ion peaks at [(MHAP - 1) - 15] and [(MHAP - 1) 43], indicating the loss of -CH3 and -COCH3 groups, respectively. The commercial HAP (Tokyo Kasei) was tested

to confirm the assignment of peak d. The HPLC peak of HAP appeared at the same retention time as that for peak d. Thus, peak d was concluded to be 4-hydroxyacetophenone. In a preliminary experiment, the identification of species was carried out by use of GC/MS and an identification program obtained from NIST (32). A peak with the same molecular weight as that for peak d (m/z 136) was found in the GC/MS chromatogram, which corresponded to HAP, as already deduced by LC/MS. The other peaks (a-c) were not identified clearly in the GC/MS analysis. To determine the detailed pathways by which TiO2 photocatalysts mineralize BPA, the precise identification of these intermediates is now in progress. To assess the additional benefit of the TiO2 photocatalytic degradation of BPA, we have evaluated the transcriptional estrogenic activities in response to hER in a yeast assay system. We were also concerned that some of the intermediate products might have significant estrogenic activity. At first, the activity-concentration relationship (dose-response curve) for BPA was examined (Figure 4a). It is known that the activities of estrogenic chemicals including BPA are not proportional to their concentrations (33, 34). Next, we measured changes in estrogenic activity of a solution containing BPA and intermediate products as well as the BPA concentration during the photocatalytic degradation. In this experiment, the initial BPA concentration was adjusted to 175 µM, which is higher enough than the lower detection limit of the HPLC UV detector for BPA, 10-6 M. The estrogenic activity of the photocatalytically treated solution was assayed VOL. 35, NO. 11, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2367

after a 20-fold dilution because the estrogenic activity of BPA is maximum at around 10-5 M (33). We found that the estrogenic activity was reduced to less than 10% of the initial activity for BPA after 6 h of UV illumination, even though 35% of the initial amount of BPA remained in the solution with some intermediate products, as shown in Figure 4b. The activity of the treated solution was replotted as a function of the concentration of BPA contained in the solution (after 20-fold dilution) in Figure 4a. The curve obtained is in good agreement with the curve obtained for BPA solutions before photocatalytic treatment (Figure 4a). This means that no significant increase in the estrogenic activity was caused by the photocatalytic treatment; the estrogenic activity of the intermediate products can be negligible for those treated solutions. In addition, we examined the solution treated photocatalytically for 6 h after further dilution over a range of 3 orders of magnitude (Figure 4a, triangles) because some estrogenic compounds exhibit activity at rather low concentrations. As a result, no increase in estrogenic activity was observed. These results show that there is not likely to be any secondary risk to activating hER as a result of the TiO2 photocatalytic degradation of BPA in water under these experimental conditions. In summary, we confirmed for the first time the total mineralization of BPA in water and the dramatic reduction of the estrogenic activity in the treated water as a result of photocatalytic reactions. We have made some progress in clarifying the BPA degradation mechanism. In addition, we are now applying TiO2 photocatalysis in a similar way to the degradation of other EDs, for example, 17β-estradiol (35). TiO2 photocatalysis could be applied to sewage treatment works as a new developing methodology for reducing levels of EDs without generating secondary pollution.

Acknowledgments We are grateful to Dr. Kenji Kusumoto (Chemicals Evaluation and Research Institute, Japan) for running the LC/MS analyses. We thank Prof. D. A. Tryk for carefully reading the manuscript. This work was supported in part by a Grantin-Aid for Scientific Research from the Ministry of Education, Science, Sports and Culture of Japan and the Joint Research Projects for Regional Intensive of Kanagawa Prefecture.

Literature Cited (1) Fukazawa, Y.; Iguchi, T. Zool. Sci. 1999, 16, 153. (2) Hilliard, C. A.; Armstrong, M. J.; Bradt, C. I.; Hill, R. B.; Greenwood, S. K.; Galloway, S. M. Environ. Mol. Mutagen. 1998, 31, 316. (3) Pfeiffer, E.; Rosenberg, B.; Deuschel, S.; Metzler, M. Mutat. Res. 1997, 390, 21. (4) Ramasabaphathy, R.; Tom, M.; Post, C. Biochem. Pharmacol. 1997, 53, 1425. (5) Tsutsui, T.; Tamura, Y.; Yagi, E.; Hasegawa, K.; Takahashi, M.; Maizumi, N.; Yamaguchi, F.; Barrett, J. C. Int. J. Cancer 1998, 75, 290. (6) Michelangeli, F.; Tovey, S.; Lower, D. A.; Tien, R. F.; Mezna, M.; Mclellan, H.; Hughes, P. Biochem. Soc. Trans. 1996, 24, 293S. (7) Nagel, S. C.; vom-Saal, F. S.; Thayer, K. A.; Dhar, M. G.; Boechler, M.; Welshons, W. V. Environ. Health Perspect. 1997, 105, 70. (8) Reel, J.; George, J.; Lawton, A.; Meyers, C. Environ. Health Perspect. 1997, 105, 273. (9) Saunders: P. T.; Majdic, G.; Parte, P.; Millar, M. R.; Fisher, J. S.; Turner, K. J.; Sharpe, R. M. Adv. Exp. Med. Biol. 1997, 424, 99. (10) vom-Saal, F. S.; Cooke, P. S.; Buchanan, D. L.; Planza, P.; Thayer, K. A.; Nagel, S. C.; Parmigiani, S.; Welshona, W. V. Toxicol. Ind. Health 1998, 14, 239.

2368

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 11, 2001

(11) Welshons, W. V.; Nagel, S. C.; Thayer, K. A.; Judy, B. M.; vomSaal, F. S. Toxicol. Ind. Health 1999, 15, 12. (12) Krishnam, A. V.; Starhis, P.; Permush, S. F.; Tokes, L.; Ferdman, D. Endocrinology 1993, 132, 2279. (13) Dorn, P. B.; Chou, C.; Gentempo, J. J. Chemosphere 1987, 16, 1501. (14) Spivack, J.; Leib, T. K.; Lobos, J. H. J. Biol. Chem. 1994, 269, 7323. (15) Hoigue, J.; Bader, H. Water Res. 1983, 17, 173. (16) Hoigue, J.; Bader, H.; Haag, W. R.; Taehein, J. Water Res. 1985, 19, 993. (17) Photocatalysis: Fundamentals and Applications; Serpone, N., Pellizzetti, E., Eds.; Wiley-Interscience: Amsterdam, 1989. (18) Photocatalytic Purification and Treatment of Water and Air; Ollis, D. E., Al-Ekabi, H., Eds.; Elsevier: Amsterdam, 1993. (19) Aguado, M. A.; Anderson, M. A.; Hill, C. G. J. Mol. Catal. 1995, 165, 89. (20) Schwiztgebel, J.; Ekerdt, J. G.; Gerischer, H.; Heller, A. J. Phys. Chem. 1995, 99, 5633. (21) Theron, P.; Pichat, P.; Guillard, C.; Petrier, C.; Chopin, T. Phys. Chem. Chem. Phys. 1999, 1, 4663. (22) Sopyan, I.; Murasawa, S.; Hashimoto, K.; Fujishima, A. Chem. Lett. 1994, 723. (23) Sopyan, I.; Watanabe, M.; Murasawa, S.; Hashimoto, K.; Fujishima, A. Chem. Lett. 1996, 69. (24) Ohko, Y.; Hashimoto, K.; Fujishima, A. J. Phys. Chem. A 1997, 101, 8057. (25) Ohko, Y.; Tryk, D. A.; Hashimoto, K.; Fujishima, A. J. Phys. Chem. B 1998, 102, 1724. (26) Kikuchi, Y.; Sunada, K.; Iyoda, T.; Hashimoto, K.; Fujishima, A. J. Photochem. Photobiol. A 1997, 106, 51. (27) Sunada, K.; Kikuchi. Y.; Hashimoto, K.; Fujishima, A. Environ. Sci. Technol, 1998, 32, 726. (28) Coldham, N. G.; Dave, M.; Sivapathasundaram, S. Environ. Health Perspect. 1997, 105, 734. (29) Collins, B. M.; McLachlan, J. A.; Arnold, S. F. Steroids 1997, 62, 365. (30) The recombinant yeast of S. cerevisiae BJ3505 was expressed as [Mat R, pep4::HIS3, prb-∆1, 6R, his3-∆200, lys2-801, trpl-∆101, ura3-52(can1)] with an expression plasmid containing the complementary DNA of hER and a reporter plasmid containing estrogen response elements linked to the lacZ gene. (31) Since 40 mL of 170 µM BPA was used in the present experiments, CO2 production of 105 µmol is expected when the complete mineralization is achieved. (32) GC mass spectrometric analyses were performed by GC/MS (Shimadzu, QP-5050A), equipped with a fused silica caplillary column (NEUTRA BOND-5, GL Sciences Inc., 30 m long, 0.25 mm i.d., 0.4 µm film thickness). A split-splitless injection port was used in the splitless mode at high pressure (350 kPa). The column temperature was programmed as follows: 2 min at 50 °C, 20 °C/min to 130 °C, 10 °C/min to 300 °C, and 8 min at 300 °C. The helium gas flow rate was 1.63 mL min-1 (at 50 °C). Electron impact was used for ionization of samples for the GC/ MS analysis. The identification of some of the intermediates was made by use of an identification program of the U.S. National Institute of Standards and Technology (NIST) library (Shimadzu). For the GC/MS analysis, the water in the aqueous sample solution was largely substituted by methanol followed by N2. This procedure was repeated a total of three times. (33) Gaido, K. W.; Leonard, L. S.; Lovell, S.; Gould, J. C.; Babai, D.; Portier, C. J.; McDonnell, D. P. Toxicol. Appl. Pharmacol. 1997, 143, 205. (34) Arnold, S. F.; Klotz, D. M.; Collins, B. M.; Vonier, P. M.; Guillette, L. J., Jr.; McLachlan, J. A. Science 1996, 272, 1489. (35) Kubota, Y.; Niwa, C.; Iguchi, T.; Ohko, Y.; Nakajima, T.; Tatsuma, T.; Fujishima, A. unpublished results (Tokyo, 2001).

Received for review October 11, 2000. Revised manuscript received February 8, 2001. Accepted February 23, 2001. ES001757T