Chemosphere 40 (2000) 433±440 Semiconductor-assisted photocatalytic degradation of reactive dyes in aqueous solution Carlos A.K. Gouv^eaa, Fernando Wypycha, Sandra G. Moraesb, Nelson Dur anb, Noemi Nagatab, Patricio Peralta-Zamoraa,* b a Departamento de Quõmica, Universidade Federal do Paran a, P.O. Box 19081, 81531-990, Curitiba-PR CEP, Brazil Biological Chemistry Laboratory, Instituto de Quõmica, Universidade Estadual de Campinas, P.O. Box 6154, 13083-970, Campinas-SPCEP, Brazil Received 14 April 1999; accepted 16 August 1999 Abstract This work reports the semiconductor-assisted photochemical degradation of reactive dyes. In an oxygenated-UVZnO system almost total decolorization of Remazol Brilliant Blue R, Remazol Black B, Reactive Blue 221 and Reactive Blue 222 was observed in reaction times of about 60 min. Extending the photochemical treatment up to 120 min, mineralization higher than 80% for all the dyes was observed. During the same period, the residual acute toxicity was signi®cantly reduced only for Remazol Black B. A systematic optimization study carried out by factorial design showed that for the reactive dyes tested, the ZnO semiconductor exhibits a better eciency than that observed with anatase TiO2 . A synergistic e€ect in the coupled TiO2 ±ZnO system was not observed. Ó 1999 Elsevier Science Ltd. All rights reserved. Keywords: Dye degradation; Heterogeneous photocatalysis; Titanium dioxide; Zinc oxide 1. Introduction The discharge of wastewater that contains high concentrations of reactive dyes is a well-known problem associated with dyestu€ activities. In view of the relatively low ®xation rates (ranging between 60% and 90% (Camp and Sturrock, 1990)) and the low eciency of the biological processes usually used for remediation of the e‚uents (activated sludge systems), about 20% of the un®xed dyes are discharged into the environment (Weber and Stickney, 1993). In view of the carcinogenic or mutagenic character of some reactive dyes, the deleterious e€ect of the color in the receiving waters, and the * Corresponding author. E-mail address: zamora@quimica.ufpr.br (P. Peralta-Zamora). customary resistance of the e‚uents to biological degradation (Chao and Lee, 1994; Knapp et al., 1995), the necessity of investigating new alternatives for the adequate treatment of this kind of residues is evident. In recent years, semiconductor-assisted photocatalysis has been extensively investigated, mainly due to its capacity to degrade a high number of recalcitrant chemicals in gaseous or aqueous systems, through relatively inexpensive procedures. In these reports, anatase titanium dioxide appears as the most extensively investigated semiconductor, mainly due to its characteristics which include: high photochemical reactivity, relatively low-cost, stability in aqueous systems, and low environmental toxicity (Stand€ord et al., 1996; Lisenbigler et al., 1995). Photocatalytic reactions take place when the semiconductor particle absorbs a photon of light more energetic than its bandgap. In this circumstance, the 0045-6535/00/$ - see front matter Ó 1999 Elsevier Science Ltd. All rights reserved. PII: S 0 0 4 5 - 6 5 3 5 ( 9 9 ) 0 0 3 1 3 - 6 434 C.A.K. Gouv^ea et al. / Chemosphere 40 (2000) 433±440 electron is excited from the valence band (VB) to the conduction band (CB), forming an hole±electron pair, able to initiate the oxidation and reduction processes of adsorbed substrates (A) (Fig. 1). In aqueous solutions the holes are scavenged by surface hydroxyl groups to generate the strong oxidizing hydroxyl radical OH†, which can promote the oxidation of organic compounds. The chemical identi®cation of hydroxylated intermediates and the detection of hydroxyl radicals appear to support the hydroxyl radical mechanism. However, several studies have shown that the direct hole oxidation and the participation of several other reactive species play an important role in the TiO2 -assisted photooxidation of organic compounds (Stand€ord et al., 1996; Qu et al., 1998). The photochemical assisted formation of some reactive species is presented in the following reaction sequence: ‡ ZnO ‡ hm ! ZnO eÿ CB ‡ hVB † h‡ ‡ OHÿ ! OH O2 ‡ eÿ ! Oÿ 2 ‡ Oÿ 2 ‡ H ! HO2 HO2 ‡ HO2 ! H2 O2 ‡ O2 H2 O2 ‡ hm ! 2OH H2 O2 ‡ eÿ ! OHÿ ‡ OH Several studies of photochemical degradation of dyes by using TiO2 have been recently reported (Qu et al., 1998; Lakshmi et al., 1995; Vinodgopal et al., 1996; Tang et al., 1997; Shourong et al., 1997; Naskar et al., 1998). In all cases, reaction times longer than 60 min were necessary to induce almost the total degradation of the dye. However, generally the progress of the reactions is followed only by monitoring the disappearance of the dye at its maximum absorption wavelengths; thus, Fig. 1. Schematic representation of the photochemical activation of a semiconductor and formation of the hydroxyl radical. VB: valence band; CB: conduction band: A: electronic acceptor compound; D: electronic donating compound. information about real mineralization of the dye or decreases in toxicity are, as far as we know, not reported. While most photocatalytic studies have used anatase TiO2 as photocatalyst, numerous studies have been carried out to evaluate the potentiality of other metal oxides. Among others, zinc oxide appears as a very promising photocatalyst for degradation of organic solutes in aqueous systems. In some cases, ZnO was more e€ective than TiO2 (Stand€ord et al., 1996; Villase~ nor and Mansilla, 1996). In this work, a ZnO (and TiO2 )-assisted photocatalytical degradation study of some reactive dyes is reported. The preliminary investigation of a potential synergetic e€ect between both semiconductors was carried out by factorial design. 2. Materials and methods 2.1. Reagents The dyes were purchased from a textile mill located in Americana (S~ ao Paulo, Brazil). The chemical structure of Remazol Brilliant Blue R and Remazol Black B are presented in Fig. 2. Information about the chemical nature of Reactive Blue 221 and Reactive Blue 222 was not available from the current literature, presumably because of commercial con®dentiality. Solutions of the dyes were prepared with distilled water. Titanium dioxide (anatase, Degusa P-25) and zinc oxide (Nuclear, 99%) were used without any pre-treatment. 2.2. Photochemical treatment 100 ml of aqueous dye solution (50 mg lÿ1 ) at a pH of 5.0 (adjusted with diluted aqueous solutionf of NH3 and HNO3 ) and 125 mg of TiO2 (or ZnO) were placed in a 150 ml reactor equipped with water refrigeration and magnetic stirrer. The suspension was irradiated from the top (12 cm from the solution surface) with a 125 W Philips medium pressure mercury lamp without the glass cover. At these conditions the temperature was 20  2 C and the ¯uence rate on the suspension surface was Fig. 2. Chemical structure, colour index and maximum absorption wavelength of the studied dyes. C.A.K. Gouv^ea et al. / Chemosphere 40 (2000) 433±440 15 Jmÿ2 sÿ1 at k > 254 nm (measured with a Cole-Parmer radiometer). The system was bubbled with commercial oxygen through a sintered glass placed in the bottom of the reactor at ¯ows of about 10 ml minÿ1 . Samples of 5 ml were taken at convenient times centrifuged for 15 min at 4000 rpm and ®ltered through a 0.45 lm Millipore ®lter prior to analysis. When the mixture ZnO:TiO2 was used as photocatalyst, the ratio 1:1 was used for both levels of mass (25:25 and 75:75 mg). 3. Analyses The eciency of the processes was evaluated by monitoring the dye decolorization at the maximum absorption wavelength (see Fig. 2), with a Hitachi U-2000 spectrophotometer. The total organic carbon content was measured with a TOC-5000 Shimadzu Total Organic Analyser. Acute toxicity determinations were carried out by measuring the respiration inhibition in Escherichia coli cultures, by using FIA-conductometric methodology (Jardim et al., 1993). The sterilized culture media for E. coli was composed by 7.0 g of K2 HPO4 , 3.0 g of KH2 PO4 , 0.5 g of sodium citrate, 1.0 g of ammonium sulphate, 0.2 g of magnesium sulphate and 20 g of glucose (®nal volume: 1 l with deionized water; pH: 7.2). The same media was used as control, without the addition of the stressing agent. 4. Results and discussion Preliminary optimization of some experimental parameters was carried out by the factorial designs 435 presented in Tables 1 and 2 Basically, these two experimental designs were performed to examine the e€ect of the semiconductor type in the eciency of the decolorization process. Due to the close connection between the semiconductor nature, its mass and the presence of electron scavengers, the latter two variables were also included in this study. The ®rst study shows that the three evaluated variables play an important role in the photochemical decolorization process. The main e€ect of the semiconductor type (S: 30) indicates that the ef®ciency of the process was enhanced nearly 30% points when the semiconductor was changed from the low (TiO2 ) to the high level (ZnO). The observed higher ef®ciency for the ZnO semiconductor can be explained on account of its high adsorption capacity (See Fig. 4). For ZnO, approximately 10% of the decolorization ratio are due to adsorption of the substrate, while for anataseTiO2 the adsorption capacity are considerably lower (near of 4%). Considering that the principal step of the photochemical degradation process involves hydroxyl radical attack on previously adsorbed substrate molecules, the importance of the adsorption capacity of the semiconductor is quite evident. This higher decolorization capacity is not surprising, because better eciency of ZnO photocatalysts has been also reported in degradation studies involving recalcitrant molecules such as lignin (Stand€ord et al., 1996; Villase~ nor and Mansilla, 1996). A similar tendency was observed for the nature of the electron scavenger agent. When the system was changed from the aerated to the oxygenated mode, the decolorization process showed an increment of about 17% points (A: 17). In additional experiments executed in the absence of an oxygen source, decolorization eciency was decreased strongly. This result Table 1 Preliminary optimization of the photochemical process by a 23 complete factorial design. Dye: Remazol Brilliant Blue R (50 mg lÿ1 ); volume: 100 ml; Photochemical reaction time: 30 min; pH: 5.5 Variable Semiconductor Atmosphere Mass (mg) Level ())aps TiO2 Air 25 Level (+) ZnO Oxygen 75 Runa Semiconductor (S) Atmosphere (A) Mass (M) Decolorizationb (%) 1 2 3 4 5 6 7 8 ) + ) + ) + ) + ) ) + + ) ) + + ) ) ) ) + + + + 14 27 24 30 14 67 44 93 Main e€ects: S : 30  2; A : 17  2; M : 31  2 Combined e€ects: S  M : 21  2; A  M : 11  2 a Experiments performed in random order. b Mean of two determinations (Typical Relative Standard Deviation of about 2%). 436 C.A.K. Gouv^ea et al. / Chemosphere 40 (2000) 433±440 Table 2 Preliminary optimization of the photochemical process by a 23 complete factorial design. Dye: Remazol Brilliant Blue R (50 mg lÿ1 ); volume: 100 ml; Photochemical reaction time: 30 min; pH: 5.5 Variable Semiconductor Atmosphere Mass (mg) Level ()) ZnO Air 25 Level (+) ZnO:TiO2 Oxygen 75 Runa Semiconductor (S) Atmosphere (A) Mass (M) Decolorizationb (%) 1 2 3 4 5 6 7 8 ) + ) + ) + ) + ) ) + + ) ) + + ) ) ) ) + + + + 27 20 30 24 67 52 93 57 Main e€ects: S: ÿ16  2; A: 10  2; M: 42  2 Combined e€ects: S  A : ÿ5  2; S  M : 9  2; A  M : 6  2; S  A  M : ÿ6  2 a Experiments performed in random order. b Mean of two determinations (Typical Relative Standard Deviation of about 2%). con®rms the importance of oxygen in avoiding the hole± electron pair recombination process, by scavenging the photogenerated electrons. The e€ect of the photocatalyst mass was also evident (M: 31). When the variable was changed from 25 to 75 mg, a mean increase of about 30% points was observed in the decolorization process. In view of these preliminary results, it is possible to conclude that the best conditions for the decolorization process are represented by the use of ZnO, an oxygenated system and a higher mass (run 8: 93% of decolorization). Since the existence of substantial combined e€ects (S  M: 21 and A  M: 11) necessarily implicates a close correlation between the variables, the application of a conventional univariate optimization system would be completely inadequate. From the geometrical representation given in Fig. 3, it is possible to observe that the correlation between the decolorization eciency and the mass of catalyst depend largely on the nature of the semiconductor. Thus, the modi®cation of the mass in this studied interval causes an enhancement of 63% points for ZnO and only 20.4 for TiO2 . To verify the existence of a synergistic e€ect between both semiconductors a second factorial design was carried out (Table 2). Again, the presence of oxygen and the increase of the semiconductor mass were very important in enhancing the decolorization eciency of the photochemical process. The e€ect of the semiconductor nature was also signi®cant (S:)16). However, the ()) signal implicate that, when the ZnO was replaced by the ZnO:TiO2 mixture, the decolorization was decreased by nearly 16% points. This result con®rms the nonexistence of any synergistic effect between these semiconductors. Many studies that report a signi®cant enhancement of the photochemical activity by coupling semiconductors have been recently published (Jardim et al., 1993; Hotchandani and Kamat, 1992; Ranjit and Viswanathan, 1997; Lin and Yu, 1998). However, this enhancement, which is the result of an ecient charge separation process and a subsequent higher disposability of the reactive electron±hole pair, was not observed in this study. In view of the important e€ect of the semiconductor mass on the degradation eciency, this variable was studied in more detail for the ZnO-oxygenated system. The results, Fig. 4, showed that the decolorization eciency increases almost linearly with the increment of the Fig. 3. Geometrical interpretation of the combined e€ect of mass and semiconductor type in an oxygenated system. C.A.K. Gouv^ea et al. / Chemosphere 40 (2000) 433±440 437 Fig. 4. E€ect of the ZnO mass on the decolorization eciency. Dye: Remazol brilliant blue R (50 mg lÿ1 ); volume: 100 ml; Photochemical reaction time: 30 min; pH: 5.5. semiconductor mass. However, although this improvement was observed up to 225 mg, a mass of 125 mg was used for subsequent experiments, because above this amount the subsequent photocatalyst separation procedure became too laborious. By using the ZnO as photocatalyst and the previously optimized experimental conditions, the e€ect of the pH on the decolorization eciency was evaluated. A signi®cant enhancement on the decolorization process of Remazol Brilliant blue R at pH between 7 and 11 was observed (Fig. 5). The interpretation of pH e€ects on the eciency of the photochemical process is a very dicult task, because three possible reaction mechanisms can contribute to dye degradation, namely: hydroxyl radical attack, direct oxidation by the positive hole, and direct reduction by the electron in the conducting band. The importance of each one depends on the substrate nature and pH (Tang et al., 1997). In our particular case we can presume that the main reaction is represented by the hydroxyl radical attack, which can be highly favored by the high concentration of adsorbed hydroxyl groups at high pH values. An additional explanation for the pH e€ects can be related with changes in the speciation of the dye. That is, deprotonation of the dye can change its adsorption characteristics and redox reactivity. However, to avoid dissolution of ZnO, which is very signi®cant at high pH values, a pH of 5 for subsequent experiments was selected. Fig. 5. E€ect of the pH on the decolorization eciency. Dye: Remazol Brilliant Blue R (50 mg lÿ1 ); volume: 100 ml; Photochemical reaction time: 30 min; photocatalyst: ZnO (125 mg). 438 C.A.K. Gouv^ea et al. / Chemosphere 40 (2000) 433±440 Fig. 6. Kinetics of the photochemical decolorization process of Remazol Brilliant Blue R. Applying the photochemical process in the optimized experimental conditions a kinetic study of the dye decolorization process was carried out. For Remazol Brilliant Blue R (Fig. 6) the color removal process (monitoring at the maximum absorption wavelength, 592 nm) was very fast, reaching almost total decolorization at reaction times of about 30 min. During the ®rst 30 min the degradation kinetics can be quantitatively described as ®rst order with respect to the dye concentration (k ˆ 0.053 minÿ1 ). Under the same conditions, the observed decolorization by adsorption on the photocatalyst surface or by application of UV light was negligible (less than 10% for a reaction time of 30 min). For the next three dyes, Remazol Black B (Fig. 7A), Reactive Blue 221 (Fig. 7B) and Reactive Blue 222 (Fig. 7C), the spectrophotometric study demonstrated that the degradation process not only induces a rapid decolorization of the dye, but also signi®cant degradation of the aromatic structure. To verify the photochemical mineralization of the dyes, the reduction of the total organic carbon (TOC) content was evaluated after 30 min of photochemical reaction with the oxygenated-ZnO system. The results (Fig. 8), demonstrated that in these conditions the degradation process proceeds very rapidly, reaching a TOC reduction higher than 50% at 30 min. After stabilizing at around 60 min, the photochemical mineralization attained values over 75% for all dyes. Almost all the recent studies related to photochemical degradation of dyes were carried out with the UV-TiO2 system. In a large part of these works, the degradation process was monitored by quantitative determination of the dye (generally by HPLC) or by evaluation of color and COD reduction. Thus, information about the real mineralization of the dyes was not Fig. 7. Kinetics of the photochemical decolorization process of Remazol Black B (A), Reactive Blue 221 (B) and Reactive Blue 222 (C). Dyes concentration: 50 mg lÿ1 ; volume: 100 ml; pH: 5.0; photocatalyst: ZnO (125 mg). provided. Nevertheless, comparing the degradation ef®ciency reported in these recent works with the real degradation capability shown by the process proposed here, it is possible to observe that the ZnO photocatalyst presents an interesting potential, especially for recalcitrant commercial dyes. Due to the incompleteness of the mineralization process (around 80%) and the real possibility of generating molecular fragments during the photochemical degradation process that can be more toxic than the 439 C.A.K. Gouv^ea et al. / Chemosphere 40 (2000) 433±440 Table 3 Photochemical reduction in color, TOC and acute toxicity. Dyes concentration: 50 mg lÿ1 ; volume: 100 ml; pH: 5.0; photocatalyst: ZnO (125 mg) a Reaction time (min) Remazol Black B Color (%) TOC (%) Toxicitya (%) Color (%) Remazol Brilliant Blue R TOC (%) Toxicitya (%) 0 30 60 120 100 90 0 0 100 30 22 20 100 (30) 73 (22) 53 (16) 53 (16) 100 88 0 0 100 45 20 10 100 (28) 0 (0) 21 (6) 79 (22) The absolute toxicity is show in bracket. parent compound (Jardim et al., 1997), an evaluation of the residual toxicity is an evident necessity. The acute toxicity was evaluated by monitoring the respiration inhibition in Escherichia coli cultures submitted to the stress agents. While for Remazol Black B an almost linear relationship between reaction time and toxicity reduction was observed (Table 3), with a maximum reduction of about 50% for a reaction time of 120 min, the Remazol Brilliant Blue R showed a behavior that typically characterize the formation of toxic transient species. With a photochemical reaction time of 30 min the toxicity was totally removed. From this time, the toxicity was successively increased until an almost complete recovery of the original value at a time of 120 min. That is, the toxicity is completely removed at the ®rst reaction times, but almost totally recovered with the progress of the photochemical process. In both cases, the presence of residual toxicity together with the permanence of about 20% of the original TOC content, indicate that the molecular fragments produced at higher photochemical reaction times are toxic. To solve this problem, the photochemical treatment must be extended to complete TOC removal. According to our observations, the reactive dyes can be completely mineralized with reaction times above 240 min. 5. Conclusions The semiconductor-assisted photochemical process permits almost total decolorization of the reactive dyes with reaction times around 60 min, when carried out in the oxygenated-UV-ZnO system. The fast degradation process is possible by virtue of the great ability of the photochemically formed hydroxyl radical to react with the chromophore group of the dyes. However, due to the complex reaction sequence, the mineralization process was incomplete even after reaction times of 120 min. After the same time, residual acute toxicity was observed, mainly on account of the presence of residual organic fragments that represent 20% of the initial content. The systematic optimization study showed that, for these kinds of substrates, the ZnO semiconductor presents a better eciency than that exhibited by anatase TiO2 . A synergistic e€ect in the coupled TiO2 ±ZnO system was not observed. Acknowledgements The authors thank Prof. Caroll Collins for critically reading the manuscript. References Fig. 8. Photochemical reduction of the total organic carbon content. Dyes concentration: 50 mg lÿ1 ; volume: 100 ml; pH: 5.0; photocatalyst: ZnO (125 mg). Camp, R., Sturrock, P.E., 1990. The identi®cation of the derivatives of C.I. Reactive Blue 19 in textile wastewater. Wat. Res. 24, 1275±1278. Chao, W.L., Lee, S.L., 1994. Discoloration of azo dyes by three white-rot fungi: in¯uence of carbon source. World J. Microbiol. Biotechnol. 10, 556±559. Hotchandani, S., Kamat, P.V., 1992. Charge-transfer processes in coupled semiconductor systems. Photochemistry and photoelectrochemistry of the colloidal CdS±ZnS system. J. Phys. Chem. 96, 6834±6839. Jardim, W.F., Gimenez, S.M.N., Canela, M.C., de Moraes, S.G., 1993. Acute toxicity of Hg and Hg2‡ ions to Escherichia coli. Chem. Spec. Bioavail. 5, 97±100. Jardim, W.F., Moraes, S.G., Takiyama, M.M.K., 1997. Photocatalytic degradation of aromatic chlorinated compounds 440 C.A.K. Gouv^ea et al. / Chemosphere 40 (2000) 433±440 using TiO2 : Toxicity of intermediates. Wat. Res. 31, 1728± 1732. Knapp, J.S., Newby, P.S., Reece, L.P., 1995. Discoloration of dyes by wood-rotting basidiomycete fungi. Enzyme Microb. Technol. 17, 664±668. Lakshmi, S., Renganathan, R., Fujita, S., 1995. Study on TiO2 mediated photocatalytic degradation of methylene blue. J. Photochem. Photobiol. A: Chem. 88, 163±167. Lin, J., Yu, J.C., 1998. An investigation on photocatalytic activities of mixed TiO2 -rare earth oxides for the oxidation of acetone in air. J. Photochem. Photobiol. A: Chem. 116, 63±67. Lisenbigler, A.L., Lu, G., Yates, J.T., 1995. Photocatalysis on TiO2 surfaces: principles, mechanism, and selected results. Chem. Rev. 95, 735±758. Naskar, S., Pillay, S.A., Chanda, M., 1998. Photocatalytic degradation of organic dyes in aqueous solutions with TiO2 nanoparticles immobilized on foamed polyethylene sheet. J. Photochem. Photobiol. A: Chem. 113, 257±264. Qu, P., Zhao, J., Shen, T., Hidaka, H., 1998. TiO2 -assisted photodegradation of dyes: a study of two competitive primary processes in the degradation of RB in an aqueous TiO2 colloidal solution. J. Mol. Catal. A: Chem. 129, 257± 268. Ranjit, K.T., Viswanathan, B., 1997. Synthesis, charcaterization and photocatalytic properties of iron-doped TiO2 catalysts. J. Photochem. Photobiol. A: Chem. 108, 79± 84. Shourong, Z., Quingguo, H., Jun, Z., Bingkun, W., 1997. A study on dye photoremoval in TiO2 suspension solution. J. Photochem. Photobiol. A: Chem. 108, 235±238. Stand€ord, U., Gray, K.A., Kamat, P.V., 1996. Photocatalytic degradation of organic contaminants:halophenols and related model compounds. HCR Advanced-Education Rev., Het. Chem. Rev. 3, 77±104. Tang, W.Z., Zhang, Z., An, H., Quintana, M.O., Torres, D.F., 1997. TiO2 /UV photodegradation of azo dyes in aqueous solutions. Environ. Technol. 18, 1±12. Villase~ nor, J., Mansilla, H., 1996. E€ect of temperature on kraft black liquor degradation by ZnO-photoassisted catalysis. J. Photochem. Photobiol. A: Chem. 93, 205±209. Vinodgopal, K., Wynkoop, D.E., Kamat, P.V., 1996. Environmental photochemistry on semiconductor surfaces: Photosensitized degradation of a textile azo dye acid orange 7 on TiO2 particles using visible light. Environ. Sci. Technol. 30, 1660±1666. Weber, J., Stickney, V.C., 1993. Hydrolysis kinetics of reactive blue 19-vinyl sulfone. Wat. Res. 27, 63±67.