Heterogeneous photocatalytic treatment of organic dyes in air and aqueous media

Journal of Photochemistry and Photobiology C: Photochemistry Reviews 9 (2008) 171–192 Contents lists available at ScienceDirect Journal of Photochemistry and Photobiology C: Photochemistry Reviews journal homepage: www.elsevier.com/locate/jphotochemrev Review Heterogeneous photocatalytic treatment of organic dyes in air and aqueous media K. Rajeshwar a,∗ , M.E. Osugi b , W. Chanmanee c , C.R. Chenthamarakshan a , M.V.B. Zanoni b , P. Kajitvichyanukul d , R. Krishnan-Ayer a a Center for Renewable Energy Science & Technology, University of Texas at Arlington, Arlington, TX 76019, USA Departamento de Quimica Analitica, Instituto de Quimica de Araraquara, CEP 14801-970, Araraquara, SP, Brazil National Center of Excellence for Environmental and Hazardous Waste Management, Chulalongkorn University, Bangkok, Thailand d Biological Engineering Program, Faculty of Engineering, King Mongkut’s University of Technology Thonburi, Thailand b c a r t i c l e i n f o Article history: Received 24 April 2008 Received in revised form 14 July 2008 Accepted 21 September 2008 Available online 17 October 2008 Keywords: Photoexcitation Free radicals Solar irradiation Electron–hole pairs Langmuir–Hinshelwood mechanism a b s t r a c t This review focuses on the heterogeneous photocatalytic treatment of organic dyes in air and water. Representative studies spanning approximately three decades are included in this review. These studies have mostly used titanium dioxide (TiO2 ) as the inorganic semiconductor photocatalyst of choice for decolorizing and decomposing the organic dye to mineralized products. Other semiconductors such as ZnO, CdS, WO3 , and Fe2 O3 have also been used, albeit to a much smaller extent. The topics covered include historical aspects, dark adsorption of the dye on the semiconductor surface and its role in the subsequent photoreaction, semiconductor preparation details, photoreactor configurations, photooxidation kinetics/mechanisms and comparison with other Advanced Oxidation Processes (e.g., UV/H2 O2 , ozonation, UV/O3 , Fenton and photo-Fenton reactions), visible light-induced dye decomposition by sensitization mechanism, reaction intermediates and toxicity issues, and real-world process scenarios. © 2008 Elsevier B.V. All rights reserved. Contents 1. 2. 3. 4. 5. 6. 7. 8. Introduction and scope of review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Categories of dyes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Evolution of the field of heterogeneous photocatalytic treatment of organic dyes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Dark adsorption of the dye on the semiconductor surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Dark adsorption of the dye and photocatalytic reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Photocatalyst details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1. Titanium dioxide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2. Other semiconductors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Photoreactor configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Photodegradation of organic dyes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.1. Process variants and mechanistic aspects involving light absorption by semiconductor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.2. Comparison with Advanced Oxidation Processes and combination with sonolysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.3. Photosensitized conversion of organic dyes in visible light . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.4. Photoreactions at the solid/air interface or in the gas phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.5. Photoreaction kinetics, variables, and related aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.6. Photoreaction intermediates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.7. Toxicity studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Into the real world: tests with dye wastewaters, process scale-up and economic aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ∗ Corresponding author. Tel.: +1 817 272 3492; fax: +1 817 272 3511. E-mail address: rajeshwar@uta.edu (K. Rajeshwar). 1389-5567/$20.00 © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.jphotochemrev.2008.09.001 172 173 173 173 175 177 177 177 180 180 181 181 183 183 184 184 186 186 189 172 9. K. Rajeshwar et al. / Journal of Photochemistry and Photobiology C: Photochemistry Reviews 9 (2008) 171–192 Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Krishnan Rajeshwar was born in Trivandrum, Kerala State in India. After completing his undergraduate and masters degrees in chemistry at the University College and Indian Institute of Technology respectively, he specialized in solid-state chemistry for his PhD degree at the Indian Institute of Science in 1974. After a post-doctoral stint in Canada (St. Francis Xavier University, Nova Scotia), he was a senior research associate at Colorado State University where he began working on energy and environmental research problems. He began his academic career at the University of Texas at Arlington in 1983 where he is currently a Distinguished Professor of Chemistry & Biochemistry and Associate Dean in the College of Science. His research interests span a broad spectrum in energy, environmental, and materials chemistry problems. Marly Eiko Osugi was born in São Paulo, Brazil in 1980. She obtained her PhD in chemistry from Institute of Chemistry of University of São Paulo State, UNESP, Brazil in 2008 under the supervision of Prof. M. V. B. Zanoni. She further developed her research at The University of Texas at Arlington (2006–2007) under the supervision of Prof. K. Rajeshwar. Her main research interest is on the electrochemistry of dyes and degradation of dyes by photoelectrochemical and electrochemical processes. Wilaiwan Chanmanee was born in Supanburi, Thailand in 1981. She graduated from the National Center of Excellence for Environmental and Hazardous Waste Management, Chulalongkorn University, Thailand with PhD in 2007. Her work was under the joint supervision of Prof. Krishnan Rajeshwar and Assoc. Prof. Puangrat Kajitvichyanukul. Currently, she is doing a post doctoral stint at the Department of Chemistry and Biochemistry, University of Texas at Arlington. Her main research is on the fabrication of titania nano-materials using electrochemical techniques. C.R. Chenthamarakshan was born in Pala, Kerala, India in 1969. He received his MS in chemistry in 1991 from Mahatma Gandhi University, Kerala, India and his PhD in chemistry in 1997 from Regional Research Laboratory, Council of Scientific and Industrial Research (RRL-CSIR), The University of Kerala, Trivandrum, India under the supervision of Dr. A. Ajayaghosh (RRL-CSIR). After a one year postdoctoral research associate appointment at RRLCSIR he was awarded a post doctoral position in 1998 in the group of Prof. Krishnan Rajeshwar, Department of Chemistry and Biochemistry, The University of Texas at Arlington (UT Arlington), Arlington, Texas (USA). Subsequently he was appointed as a research assistant professor at UT Arlington and from 2007-to date he is working as a senior research associate at Corsicana Technologies, Corsicana, TX (USA). His research interests include semiconductor photoelectrochemistry, photocatalysis, conducting polymers, and fatty amine chemistry. Maria Valnice Boldrin Zanoni was born in Tupi Paulistacity, São Paulo, Brazil in 1957. She graduated from Institute of Chemistry of São Paulo University (USP) with PhD in physical chemistry, 1989. Her academic position started at University of São Paulo State (UNESP) as an associate professor in analytical chemistry since 1987. Her main research interests are in organic electrochemistry and environmental electrochemistry, investigating analytical methodologies and new treatment methods for dye effluents. Recently, she expanded her research interests to photoelectrochemistry using nanoporous, nanotubular and nanocomposite semiconductor electrodes for dye degradation. 189 190 190 Puangrat Kajitvichyanukul was born in Chiang Mai, Thailand in 1970. She graduated from Department of Civil and Environmental Engineering of The University of Texas at Arlington with PhD in 2002. Her academic position started at King Mongkut’s University of Technology Thonburi in Bangkok (Thailand) as a lecturer (1996–2003), as an assistant professor (2003–2006), and as an associate professor (2006–present). Her main research interests are on the synthesis and characterization of TiO2 to be used as a photocatalyst for environmental, biological, and health protection applications. She was a recipient of Distinguished Alumni Award of Chiang Mai University, Thailand in 2008. Rebecca Krishnan-Ayer is a senior in high school at the Hockaday School in Dallas, TX, USA. She was involved in this project as a summer intern at the University of Texas at Arlington. She is now looking at Universities starting Fall, 2009 and hopes to complete an undergraduate degree combining writing, art history, and science. 1. Introduction and scope of review Millions of various colored chemical substances have been generated within the last century or so, 10,000 of which are industrially produced [1]. On a global scale, over 0.7 million tons of organic synthetic dyes are manufactured each year mainly for use in the textile, leather goods, industrial painting, food, plastics, cosmetics, and consumer electronic sectors. A sizable fraction of this is lost during the dying process and is released in the effluent water streams from the above industries. Therefore, decolorization and detoxification of organic dye effluents have taken an increasingly important environmental significance in recent years [2,3]. Most conventional methods for the removal of dye pollutants such as adsorption on activated carbon, ultrafiltration, reverse osmosis, etc. are non-destructive and merely transfer pollutants from one phase (for example, aqueous) to the other (for example, adsorbent). Biodegradation is slow and inefficient for many azo dyes and does not work at all for others (for example, Acid Orange 7). Chlorination and ozonation are also relatively inefficient and have high operating costs. Thus, Advanced Oxidation Processes (AOPs), which rely on the generation of hydroxyl and other radical species for environmental remediation, have been successfully deployed for the treatment of organic dye-laden waters [3–6]. This review focuses on one such methodology within the AOP category, namely heterogeneous photocatalysis using an inorganic semiconductor as photocatalyst [3]. Fig. 1 constructed from a sampling of ca. 300 papers in the journal literature over a ∼35 year period, shows an almost exponential growth, attesting to the popularity of this remediation approach. In many instances, as elaborated below, complete mineralization of the dye can be achieved under mild process conditions. The process economics can be further improved if sunlight is used (instead of UV light) for excitation of the semiconductor photocatalyst. Importantly, both oxidative and reductive conversion of the organic dye (into benign and colorless products) is possible using this approach [7,8]. K. Rajeshwar et al. / Journal of Photochemistry and Photobiology C: Photochemistry Reviews 9 (2008) 171–192 Fig. 1. Evolution of the organic dye photocatalysis literature over approximately three decades. Interestingly, the publications on this topic originate from countries all across the globe as Fig. 2 illustrates. Perhaps not surprisingly, China and India dominate in this regard and both economically advanced nations (for example, USA, Japan, Canada) as well as emergent economies such as Brazil and Taiwan feature prominently in Fig. 2. To our knowledge, this is the first review article dealing with the heterogeneous photocatalytic treatment of organic dyes. Previous authors have discussed the kinetics and mechanistic aspects of this process for the specific instance of azo dyes in aqueous solutions [9]. Many review articles also exist on the heterogeneous photocatalytic treatment of organic pollutants in general [8,10–20] and limited discussions on dyestuffs are contained in some of these. We also note the related use of electrochemical treatment of textile dyes and dyehouse effluents [3], a topic not further addressed in detail herein. Because of the plethora of review articles, book chapters, and books on heterogeneous photocatalysis [3–5,8,10–20], familiarity with the underlying principles of this treatment approach will be assumed in what follows. Finally, the literature sampling is representative rather than comprehensive and is almost exclusively limited to journal articles rather than conference proceedings and patents. This choice is deliberate in that we believe refereed journal articles generally contain credible data that have already been vetted by the peer review process. 173 Fig. 3. Types of dyes featured in the studies on organic dye photocatalysis. 2. Background 2.1. Categories of dyes Colorants, or additive substances causing variation in color or visible light absorption, can be divided into two categories: dyes and pigments. The distinct delineation between a dye and a pigment should be noted. Dyes are soluble or partly soluble organic (carbonbased; plant and animal extracted) colored compounds suspended in a medium, and represent one type of colorant [21]. The process of dyeing can be loosely defined as imparting color to textile fiber or leather. Pigments, on the other hand, typically are completely insoluble substances that have no chemical affinity for the substrate to be colored. Dyes can be classified on the basis of structure, function, or both. Table 1 contains a representative listing of dyes grouped according to their chemical structure. Also listed for each example in Table 1 is the color index, which contains reference numbers on the basis of the color and chemical classification [1]. Dyes can also be classified as acid, basic, direct, disperse, reactive, anionic, cationic, etc. and indeed this notation is often simultaneously used with the dye chemical structure type, for example, Basic Blue 41 and Acid Yellow 23 are both monoazo dyes. Of the synthetic dyes manufactured today, azo compounds are dominant (∼50–70%) with anthraquinone dyes being a distant second. The dominance of the azo dyes is reflected in the heterogeneous photocatalysis literature as well, Fig. 3. 2.2. Evolution of the field of heterogeneous photocatalytic treatment of organic dyes Fig. 2. Countries of origin of the publications sampled in this review on organic dye photocatalysis. The early photochemical studies on aqueous solutions of organic dyes in the presence of inorganic semiconductors such as silver halides, ZnO, and the like were driven by applications in electroand photography (i.e., dyes were used as visible light sensitizers for electro- or photographic materials, respectively) [22,23]. The first instance for observing dye instability in the presence of an inorganic semiconductor (TiO2 ) and illumination appears to be in 1969 when the photocatalytic reduction of Methylene Blue (a thiazine dye, Table 1) to the leuco form was reported [24,25]. This was subsequently followed by other studies which reported N-dealkylation of dyes such as Rhodamine B and Methylene Blue in aqueous suspensions of CdS [26]. The N-dealkylation was preceded by dye radical cation formation as a result of photoinduced electron transfer from the dye to the semiconductor conduction band. 174 K. Rajeshwar et al. / Journal of Photochemistry and Photobiology C: Photochemistry Reviews 9 (2008) 171–192 Table 1 Representative organic dyes considered in this review. Type of dye Example Color index Azo Reactive Orange 16 C.I. 17757 Xanthene Basic Violet 10 C.I. 45.170 Thiazine Methylene Blue C.I. 52015 Anthraquinone Rective Blue 4 C.I. 61205 Triphenylmethane Basic Violet 4 C.I. 42.600 Phthalocyanine Reactive Blue 15 C.I. 74459 Indigo Indigo Carmine C.I. 73015 Quinoline D&C Yellow 10 C.I. Acid Yellow 3 Phenanthrene D&C Green 8 C.I. Acid Green 9 Deliberate attempts to photochemically destroy the organic dye appear to have been instigated only almost a decade later when photoinduced electron transfer from TiO2 to Methyl Orange (a monoazo dye, Fig. 4) was observed to result in bleaching of the dye absorption (max = 470 nm) and its reductive conversion to a hydrazine derivative [27,28]. Importantly, no dye bleaching was observed in the absence of TiO2 or visible radiation alone Structure ( > 400 nm) indicating that the photocatalysis process involved initial light absorption (max = 310 nm) with the concomitant generation of e− –h+ pairs in the semiconductor. These early studies were accompanied soon by findings in other laboratories which showed that many organic compounds could be decomposed in aqueous media with a combination of TiO2 and near-UV light [3–5,8,10–20]. The photocatalytic oxida- K. Rajeshwar et al. / Journal of Photochemistry and Photobiology C: Photochemistry Reviews 9 (2008) 171–192 175 (Ceq ) according to the Langmuir adsorption isotherm: Nads = Ns KL Ceq 1 + KL Ceq (1) In Eq. (1) Ns is the total number of accessible adsorption sites and KL is the adsorption constant (in M−1 ). From the linearized form of Eq. (1): Ceq Ceq 1 = + Nads KL Nmax Nmax (2) Nmax (the maximum, saturation dye coverage) and KL can be determined from the slope and intercept, respectively. Adsorption isotherms (according to Eqs. (1), (2) or both) appear in Refs. [58–62] for Methylene Blue and in Refs. [32,34,39,44,46,47,49,50,64] for azo dyes. Unfortunately, the results of these analyses are not readily comparable because of the variant units employed for Nads and Ceq and also because of the differing nature of the TiO2 surface in the various studies. However, the KL values are generally high (in the 102 to 105 M−1 range) signaling strong interaction of the dye with the oxide surface. The parameter KL has also been reported to increase with a decrease in the adsorbent particle size because of a higher driving force for adsorption on the finer particles [68]. The shape classification of the adsorption isotherms [69] has been used to deduce, from the L-shape usually observed, that there is no strong competition between the solvent molecules and the dye to occupy the adsorbent surface sites [49,50]. On the other hand, the disparity between Nmax values obtained from the above isotherm data analyses (∼10−5 mol g−1 ) and the number of adsorption sites on the oxide surface (∼10−4 mol g−1 ) has often led to conclusion that the adsorbed dye molecules are surrounded by many anchored water molecules. The value of the dimensionless separation factor (RL ) indicating the shape of the Langmuir isotherm [70]: Fig. 4. Four representative dyes in the photocatalysis literature. See also Table 1. tion of dyes such as Methylene Blue, Rhodamine B, Fluorescein and Methyl Orange was reported thereafter [29,30]; however the field legitimately “took off” only in the late 1980s and early 1990s (Fig. 1). The preceding discussion should have indicated that a multitude of mechanisms exist for dye decomposition involving both oxidative (electron transfer from the dye) or reductive (electron transfer to the dye) routes. The light absorption can occur in the dye or by the semiconductor or both. These mechanistic aspects are addressed later in this review (Section 7.1). RL = 1 1 + KC0 (3) (where C0 is the highest initial dye concentration) has also been used in some studies to indicate the strength of the adsorption process [47]. Values for RL (0 < RL < 1) diagnose the adsorption to be favorable, unfavorable (RL > 1), linear (RL = 1), or irreversible (RL = 0). Generally, values for RL in the 0–1 range have been reported [47]. While the Langmuir isotherm [71] has been overwhelmingly popular in the dye adsorption studies, the applicability of the Freundlich isotherm [72] model has also been considered in several studies [44,62]. This empirical model has been used for non-ideal sorption that includes heterogeneous surface energy systems and is embodied in the equation: 1/n (4) Nads = KF Ceq 3. Dark adsorption of the dye on the semiconductor surface Dye adsorption is studied by equilibrating various concentrations of the dye solution with the powder or colloidal suspensions of the oxide semiconductor particles in the dark for time periods ranging from a few minutes (∼30 min) to several hours depending on the kinetics of the adsorption process. The supernatant solutions are then spectrophotometrically sampled for the amount of dye adsorbed on the semiconductor surface. Tables 2 and 3 summarize the results of such measurements for a variety of dyes. Titanium dioxide was the semiconductor used in the overwhelming majority of these studies with exceptions as noted. Dark adsorption data are processed in the form of a plot of concentration adsorbed (Nads ) versus the equilibrium concentration where KF is the adsorption constant and 1/n is the adsorption intensity. The magnitude of 1/n gives an indication of the favorability of adsorption (similar to the RL parameter above). A third isotherm (Sips adsorption model) combines some of the features of the classical Langmuir and Freundlich models [73]: 1/ns Nads = Nmax Ks Ceq 1/ns 1 + Ks Ceq (5) where Ks is the adsorption constant and 1/ns is the Sips parameter related to the adsorption intensity. For Methylene Blue and Rhodamine B adsorption on TiO2 –SiO2 mesoporous composites, the Langmuir model provided the best fit of the data trends [62]. The Langmuir model fitted the data better for Reactive Yellow 145 while the Freundlich model was 176 K. Rajeshwar et al. / Journal of Photochemistry and Photobiology C: Photochemistry Reviews 9 (2008) 171–192 Table 2 Representative studies on azo dye adsorption on the semiconductor surfacea in the dark. Entry no. Azo dye Comments Reference(s) 1 Solvent Red 1 [31] 2 Acid Orange 7 3 4 Acid Orange 7 Acid Orange 7 5 6 Acid Orange 7 Acid Orange 20 Other aspects of solution spectra (than the peak maximum shift) altered such as fine structure and wavelength range of tail. Both TiO2 and Al2 O3 surfaces studied. Value of adsorption constant (K) from isotherm data analyses reported as 2.48 × 102 M−1 for adsorption on TiO2 . Influence of TiO2 dose on dye adsorption studied. Detailed study using FT-IR, diffuse reflectance and adsorption isotherms revealing adsorption mode of dye on TiO2 and ZnO surfaces. Adsorbed dye exists as hydrazone tautomer. Another FT-IR study of dye adsorption. Adsorption of dye on TiO2 studied by FT-IR and found to be essential for the photocatalytic and photosensitized degradations to be effective. Insignificant (<4%) dye adsorption on TiO2 noted. Similar study as in Ref. [37] by same group. The effect of Au loading on TiO2 surface on dye adsorption studied. Effect of pH on dye adsorption described. K = 2 × 103 M−1 . Increased amount of dye adsorbed on Fe(III)-modified TiO2 surface. Effect of pH noted. Effect of pH noted. Effect of pH noted. Adsorption isotherms presented. Perhaps one of the most comprehensive studies on dye adsorption (see text). Adsorption isotherm studied. Mesoporous TiO2 films used. Adsorption isotherm presented and effect of pH noted. Adsorption isotherm presented. Adsorption followed by FT-IR spectroscopy. 7 8 9 10 11 12 13 14 15 Acid Orange 20 Acid Orange 52b Acid Orange 52b Methyl Red Acid Red 1 Reactive Black 5 Reactive Black 5 Reactive Black 5 Reactive Brilliant Red (X-3B) 16 17 18 19 20 Remazol Black B Basic Blue 41 Direct Black 38 Safira HE XL Amaranth a b [32] [33] [34,35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] TiO2 was used as the semiconductor in these experiments except in Entry 4 when ZnO was also used. Also known as Methyl Orange. better for Reactive Black 5 adsorption on TiO2 [44]. Values of the enthalpy, entropy, and free energy of adsorption are available for Methylene Blue [62], Rhodamine B [62], and Acid Orange 7 [34] on TiO2 and TiO2 –SiO2 composite surfaces. The adsorption kinetics were also studied for the Methylene Blue and Rhodamine B dyes [62]. Oxide surfaces in aqueous media are amphoteric and TiO2 is no exception [74]. The surface groups on titania are involved in acid–base equilibria as follows: (6) Table 3 Representative studies on dye adsorption (non-azo dyes) on the TiO2 surface in the dark. Entry no. Dye category/dye Comments Reference(s) Xanthene 1 Rhodamine B In acidic media, adsorption is inhibited by electrostatic repulsion. Addition of an anionic surfactant induces binding of the cationic dye molecules. In acidic media, adsorption is inhibited by electrostatic repulsion. Addition of an anionic surfactant induces binding of the cationic dye molecules. Adsorption tunable via the positively charged diethylamine group or the negatively charged sulfonate group of the dye molecule. Adsorption of the anionic dye found to be a pre-requisite for its photocatalytic degradation. [52–54] 2 Rhodamine B 3 Sulforhodamine B 4 Eosin Thiazine 5 [55] [56] [57] [58] Methylene Blue Methylene Blue Methylene Blue Methylene Blue Methylene Blue Methylene Blue Probably one of the earliest studies of dye adsorption on TiO2 in a photocatalysis context. The photocatalyst was attached to glass tubing wound in a spiral. Adsorption studied in a batch-recirculated photoreactor with hollow fiber membrane. Both monomeric and dimeric forms of the dye found to adsorb to the same extent. pH effect on dye adsorption noted. As in Entry 2. Adsorption studied as a function of temperature. See also Entry No. 10, Table 2. Anthraquinone 12 13 14 Alizarin S Alizarin Red Acid Blue 25 As in Entry No. 10, Table 2. Electrostatic aspects of dye adsorption discussed. Adsorption found to be important in control of the dye photodegradation rate. [41] [63] [64] Arylmethane 15 Malachite Green In the absence of surfactant, low dye adsorption on TiO2 surface. The role of both anionic and cationic surfactants discussed. [65] Indigo 16 Indigo and Indigo Carmine Adsorption on TiO2 monitored by FT-IR spectroscopy. [66] Cyanine 17 Astrazone Orange G Effect of salt probed. [67] 6 7 8 9 10 11 Methylene Blue [59] [60] [61] [55] [62] [41] 177 K. Rajeshwar et al. / Journal of Photochemistry and Photobiology C: Photochemistry Reviews 9 (2008) 171–192 (7) The point of zero surface charge (pzc) defines the solution pH when the positively charged surface groups on the oxide are exactly counterbalanced by the negatively charged groups. Values quoted for the pHpzc for TiO2 vary from 5.1 to 6.8 (a value of 6.8 appears to be most tenable) and for ZnO it is 9.0 ± 0.3. In view of the oxide surface charge tunability as a function of solution pH, it is hardly surprising that dye adsorption on TiO2 (or ZnO) has been monitored as a function of solution pH in numerous studies (see Tables 2 and 3). In general, cationic dyes (e.g., Methylene Blue, Table 1) bind at pH values greater than pHpzc while anionic dyes (that is, dyes bearing sulfonate groups) show the opposite trend. Other than the characteristics of the photocatalyst surface, solution additives can markedly influence the extent of dye adsorption on the photocatalyst surfaces in the dark. Thus Fe(III) aquo ions enhance azo dye adsorption [42] although the effect appears to be related also to the charge borne by the dye molecule. The adsorption of anionic dyes (for example, Acid Orange 7) is promoted by the presence of FeCl3 . On the other hand, inhibitory effects have also been reported for anions and attributed to their competitive adsorption on the photocatalyst surface [67]. A series of papers discuss the effect of added surfactant on dye adsorption [52–54,56,65]. Thus an anionic surfactant such as sodium dodecylbenzenesulfonate adsorbs strongly on the TiO2 particles and in turn, the bound sulfonate groups electrostatically attract cationic dyes such as Rhodamine B or Malachite Green [52–54,65]. Significantly, in the absence of surfactants, Malachite Green is difficult to degrade in aqueous TiO2 suspensions because of its poor solubility in water and low adsorption on the TiO2 surface. Interestingly, even a cationic surfactant such as hexadecyltrimethylammonium bromide can be deployed at a basic pH (>6.8) when it binds to the negatively charged TiO2 surface (Eq. (7)) [65]. A molecular level understanding of the mode of dye adsorption on the photocatalyst surface is afforded by FT-IR spectroscopy and many examples of application of this spectroscopic tool in dye adsorption studies are available [34,36,37,51]. A particularly detailed probe of the adsorption mode is given for the azo dye, Acid Orange 7, on TiO2 and ZnO surfaces [34]. Based on these FT-IR data, a bi-dentate coordination mode is envisioned for the dye–TiO2 surface complex while only one oxygen atom from the dye sulfonate group is postulated to be involved in binding with the ZnO surface [34]. As discussed later in this review, FT-IR spectroscopy also furnishes useful insights into the course of the photocatalytic process and the chemical nature of any intermediates generated. 4. Dark adsorption of the dye and photocatalytic reactions The foregoing discussion begs the important question: what role, if any, does adsorption of the dye on the photocatalyst surface (in the dark) play in the efficacy of its subsequent photocatalytic degradation? There is much literature on this issue for non-dye substrates including many organic compounds and metal ions [75–83]. At the outset, it may be noted that applicability of a rate equation for the heterogeneous photocatalysis reaction such as Langmuir–Hinshelwood mechanism (see below), can itself be construed as strong evidence for substrate adsorption to be an important element of the overall reaction mechanism. Further, the use of high-surface area, inert adsorbents, such as the candidates discussed earlier and activated carbon [84], silica gel [85] and zeolites [84,86] along with TiO2 for sequestering substrates prior to their subsequent decomposition, contains an implicit recognition of the important role that adsorption can play in the heterogeneous photocatalysis process. Table 4 Representative studies on the use of Degussa P-25 TiO2 for the photocatalytic degradation of various types of organic dyes. Entry no. Dye category Reference(s) 1 2 3 4 5 6 7 8 9 Azo Xanthene Thiazine Anthraquinone Arylmethane Phthalocyanine Indigo Cyanine Phthalein [31,32,34,39,42,43,45,46,47,49,51,88–104] [53,54,56,57,92,98,105–108] [59,60,61,109–112] [63,64,89,108,113–115] [65,108,116] [47,89] [66] [117] [29,118] Given the above, it should not be surprising that many studies indeed confirm and reinforce the notion that adsorption of dye molecules on the photocatalyst (for example, TiO2 ) is a prerequisite for photodegradation [31,41,46,49,58,64,65]. A corollary is that the photoinduced reaction occurs on the catalyst surface rather than in solution bulk. On the other hand, there are also isolated literature data which suggest that adsorption plays a negligible role in the photodegradation of the dye [44,87]. 5. Photocatalyst details 5.1. Titanium dioxide On a laboratory scale, the most popular photocatalyst configuration has been in the form of powder suspensions. Commercially available samples of TiO2 have been used in many of these studies and Tables 4 and 5 contain representative listings of these. Table 6 lists the characteristics of these commercial TiO2 samples. Head-to-head comparisons of their photcatalytic activity have also been done in many cases (e.g., Refs. [45,118]). The trends are rather complex and while Degussa P-25 works the best in many cases, there are exceptions, for example, when either the Millenium or the Hombicat samples outperform the benchmark photocatalyst. Photocatalyst parameters such as the surface area and particle size alone are not sufficient indicators of photocatalytic activity and surface chemistry factors may also play a role. Of course, the foregoing comments are not exclusive to the case of organic dyes but apply equally well to the photocatalytic conversion of other organic compounds and metal ions. The influence of crystallographic modification of the TiO2 sample appears to have been only sporadically examined. Generally, the anatase modification is considered to be superior in its photocatalytic activity relative to the rutile counterpart. On the other hand, the mixture of anatase (dominant form) and rutile modifications in Degussa P-25 TiO2 has been touted [126] to play an important Table 5 Representative studies on the use of commercial TiO2 samples (other than Degussa P-25) for the photocatalytic degradation of various types of organic dyes. TiO2 origin Dye category Reference(s) Aldrich Azo, cyanine, thiazine, arylmethane, xanthene Azo Azo Azo Azo Azo Azo Azo ? Azo Thiazine Azo [45,67,116,117,119,120] Millenium (PC 50 and PC 500) Merck Hombikat (UV 100, Pt-UV 100) DuPont Wako Pure Chemical Industries Riedel de Haen Kerr-McGee Fluka Tranox A-K-1 STS-21 (Ishihara Sangyo) Fujititan [45] [121–123] [45] [37] [87] [33] [38] [55] [45] [124] [125] 178 K. Rajeshwar et al. / Journal of Photochemistry and Photobiology C: Photochemistry Reviews 9 (2008) 171–192 Table 6 Specifications and characteristics of commercial TiO2 samples. Entry no. TiO2 sample BET surface area (m2 g−1 ) Average particle size (nm) Composition 1 2 Degussa P-25 Millenium PC 500 PC 50 Aldrich Kerr-McGee Hombikat UV 100 Tranox A-K-1 Mikroanatas IF 9308/18 DuPont R-900 50 287 45 11 90 >250 90 271 NA 21 5–10 75% anatase, 25% rutile Anatase Anatase Anatase NA Anatase Anatase Anatase doped with 0.1% (w/w) Pt Rutile 3 4 5 6 7 8 a NAa 20 5 20 NA NA NA = not available. role in the excellent photocatalytic activity of this material although this proposal must remain speculative pending further scrutiny. A study on the characteristics of the Hombikat UV-100 photocatalyst is also available [127]. Many studies on dye photodegradation by TiO2 involve the use of TiO2 colloidal particles prepared by controlled hydrolysis of a precursor. Thus, titanium(IV) isopropoxide in 2-propanol is injected into acetonitrile and stirred under N2 blanket [128]. This results in a colloidal dispersion of TiO2 particles with an effective hydrodynamic diameter of 0.15 ␮m [128]. This procedure was adopted in many subsequent studies by the same group [129–132]. Other variants of controlled hydrolysis start with titanium(IV) tetraisopropoxide, titanium(IV) butoxide or TiCl4 with details differing in the solution mixing sequence, pH, etc. Non-hydrolytic synthesis approaches have also been deployed for TiO2 . This includes hydrothermal syntheses [133,134] and a solution synthesis method based on reacting TiCl4 and titanium(IV) tetraisopropoxide at 300 ◦ C in a mixture of heptadecane and tri-noctylphosphine oxide (TOPO) as a co-ordinating (capping) agent [135,136]. Nanosized particles (∼7 nm) are “fished out” from the reaction mixture and then re-dispersed in organic solvents. The hydrothermal approach involves a combination of high temperature and pressure in a sealed container such that a constant autogenerated pressure is generated within. Thermal hydrolysis is yet another route to preparing TiO2 particles. Thus, sulfur-doped TiO2 was prepared from titanyl sulfate solution by reacting with ammonia at 85 ◦ C and controlled pH [137]. The product was repeatedly washed with de-ionized water till it was free of sulfate ions and then dried prior to thermal anneal. The photocatalyst thus prepared (along with samples also Pt-modified) were then used for the UV-assisted photooxidation of Acid Orange 7 dye in aqueous media [137]. Another variant involves a “sol-solvothermal” process wherein the polymer-stabilized TiO2 sol (prepared from the tetraisopropoxide precursor as above) is subsequently heated at temperatures in the range, 80–150 ◦ C, for several hours in a stainless-steel autoclave [138]. The advantages accrued from immobilizing or “supporting” TiO2 particles onto a high-surface area support material were identified in a preceding section and this is a lesson learnt from the heterogeneous (thermal) catalysis [139,140] community. Thus even early studies examined the influence of support material on the photocatalytic activity of TiO2 [30,118,141] and a slew of subsequent studies have utilized TiO2 nanoparticles supported on glass beads [118], buoyant polystyrene beads [142], silica [44,143], clay [144], polyoxometallate [145], alumina [141] and composite inert oxide containing either Ru or Ln metal [146,147]. Invariably, solutionbased synthesis approaches (such as the ones outlined in the preceding paragraph), were used to derive supported TiO2 composite photocatalysts. Instead of using TiO2 alone, a “coupled semiconductor” configuration (Fig. 5) [148] improves charge separation in many cases because of “vectorial” electron transfer [149] and reduced carrier recombination as further elaborated below in the discussion on TiO2 photocatalyst films. However, coupled nanoparticles of CdS and TiO2 (denoted as “CdS/TiO2 ”) have been synthesized via a sol–gel/precipitation technique [150]. Other studies report “coupled” semiconductor powders (TiO2 + ZnO, TiO2 + SnO2 , ZnO + SnO2 ) [151,152] but it is not clear whether the two semiconductor particles are truly coupled in these cases. No details are given on how the commercial powder samples were combined in Refs. [151,152]. A logical extension to a supported photocatalyst configuration is to completely immobilize the active material on an inert substrate such as metal, polymer, cloth, etc. The photocatalyst layer is then present as a film on the substrate surface. A major advantage here is that the reaction product(s) and photocatalyst do not have to be separated unlike in the cases with powder or colloidal suspensions of the photocatalyst. The quantum efficiencies are also often low in the latter because of carrier recombination at the photocatalyst surface or in its bulk. A review is available on the strategy of fixing or depositing TiO2 on a variety of polymer-based substrates and the use of these materials in the photodegradation of azo dyes in solution [153]. Perhaps the simplest approach is to make a slurry of TiO2 nanoparticles (for example, in water or alcohol) and then mix it with an organic or polymeric binder (e.g., Nafion). The surface targeted can then be dip-coated with this slurry, the number of dipcoating steps replicated to build up a requisite film thickness. Early studies [29,58] have utilized a similar method to attach Degussa P25 TiO2 nanoparticles as a thin film to the inner wall of borosilicate glass tubing wound in a spiral. Anatase TiO2 supported in fiberglass mesh or glass fiber cloth is available from commercial sources [118]. Other examples of dip-coating Degussa P-25 or anatase TiO2 film on soda-lime glass supports, for applications relevant to dye Fig. 5. Coupled semiconductor photocatalyst configuration. K. Rajeshwar et al. / Journal of Photochemistry and Photobiology C: Photochemistry Reviews 9 (2008) 171–192 photodegradation, are contained in Refs. [154,155]. Photoreactor configurations will be discussed later in this review. There are many examples in the dye photocatalysis literature for colloidal TiO2 (sol–gel) coatings on various substrates [156–159] including SiO2 -coated glass plates [156], glass slides [40,48,157], quartz slides [158–161], titanium foil [158–164], glass reactor tube or ring [165,166], borosilicate Petri dish [167], and indium tin oxide (ITO) or TiO2 glass [168]. Commercial products comprising of colorless TiO2 layers coated on clear glass (for example, Pilkington ActiveTM , Pittsburgh Plate and Glass or PPG) are also available and have been used for the photooxidation of organic compounds including organic dyes such as Acid Blue 9 and Reactive Black 5 [169,170]. The electronically conductive substrates amongst the candidates listed above (i.e., Ti, ITO) offer the crucial advantage of facilitating the imposition of a bias potential across the oxide semiconductor–solution interfere to drive the photocatalytic process. For n-type semiconductor, positive potentials result in the interface being reverse biased, driving the photogenerated holes to the interface and the electrons to the rear contact [171,172]. The result is improved e− –h+ separation and consequently, a lower carrier recombination—a key handicap with the use of TiO2 powder suspensions or TiO2 nanoparticulate films on insulating substrates (see above). This process is termed “photoelectrocatalytic” to distinguish it from its non-biased “photocatalytic” process counterpart. Many examples of photoelectrocatalysis are available in the literature on organic dyes [158–162,168], including studies on coupled semiconductor films involving TiO2 and another semiconductor such as SnO2 [173,174], or even combinations of three semiconductors, TiO2 /CdO–ZnO [175]. Generally, the coupled semiconductor films have been deposited on optically transparent electrodes such as ITO. Electrochemical or electrophoretic deposition [148,176,177] offers a versatile and low-temperature approach to coating TiO2 (or composites) on conductive substrates. Both anodic [178,179] and cathodic [180,181] approaches can be utilized for depositing TiO2 although only the anodic electrosynthesis variant appears to have been deployed for applications related to the photoelectrocatalytic (PEC) degradation of organic dyes. Thus anodization of Ti mesh in aqueous media results in a microporous TiO2 film [182,183] but under carefully tuned conditions can also facilitate the growth and self-assembly of nanotubular arrays on the Ti surface [178,179]. Nanotube arrays facilitate vectorial charge transfer (Fig. 6) relative to the tortuous path undergone by photogenerated electrons in nanoor mesoporous electrode film geometries (c.f., Fig. 6a and b). Thus it is not surprising that such nanotubular TiO2 arrays have afforded 179 faster photoelectrocatalytic degradation rates for three model disperse azo dyes (Disperse Orange I, Disperse Red 1, Disperse Red 13) relative to the nanoporous electrode assemblies prepared via colloidal deposition [184,185]. This brings up the important issues: how do powder suspensions stack up vis-á-vis supported TiO2 photocatalysts in terms of process efficiency? How does the dark electrocatalytic dye oxidation process counterpart perform relative to its photo-variants? Two sets of studies on non-dye substrates have concluded that TiO2 powder suspensions and immobilized configurations perform equally well [186,187]. On the other hand, the photocatalytic activity of TiO2 coated on glass (by the sol–gel method) is found to be intermediate between that of Millenium PC-500 and Degussa P-25 TiO2 powders [41]. Five TiO2 candidates and three different types of organic dyes were included in this particular study [41]. The PEC process was compared with the photocatalysis counterpart for the degradation of Rose Bengal dye [182]. The bias potential served to accelerate the dye photooxidation relative to all the photocatalysis cases where the TiO2 solution dose was varied (from 0.10% to 0.30%) except when the solution dose was 1.0% [182]. Another study on TiO2 –Ti mesh (prepared by anodization, see above) [183] found the PEC variant to be superior to UV photolysis, electrooxidation, and photocatalysis for the conversion of an azo dye. Essentially similar trends have been observed for Remazol Brilliant Orange 3R (azo dye) [158], Reactive Orange 16 (azo dye) [159], Remazol Turquoise Blue 15 (copper phthalocyanine dye) [160,162], and Reactive Blue 4 (anthraquinone dye) [161]. Direct UV photolysis and the PEC and photocatalysis process variants were also compared for SnO2 –TiO2 coupled semiconductor films for Naphthol Blue Black and Acid Orange 7 azo dyes [173,174]. The PEC variant performed the best relative to the other two approaches. The effects of increasing bias potential and of O2 or N2 on the photoelectrode (or photocatalyst) compartment were also investigated in these studies [173,174]. As expected, the more positive the bias potential is, the better the PEC process performs. Unlike in the photocatalyis process where air (or O2 ) is needed to scavenge the photogenerated electrons [8,188,189], the PEC process can be carried out in anaerobic conditions since the (dark) reduction occurs in the counterelectrode compartment of the cell. In electrophoretic deposition, the surface charge on the suspended particles in the solution drives them to the appropriately polarized working electrode surface; this tactic is utilized, for example, for solvent-less painting in the automobile industry. A variant of this is occlusion electrodeposition [176,177] where electrophoretic particle occlusion is combined with continual deposition of a metal “glue” (for example, nickel) which serves to Fig. 6. Schematic diagrams contrasting: (a) nanostructured and (b) nanotubular photocatalyst film configurations and vectorial electron transfer in the latter. 180 K. Rajeshwar et al. / Journal of Photochemistry and Photobiology C: Photochemistry Reviews 9 (2008) 171–192 Table 7 Use of inorganic semiconductors other than TiO2 for the photocatalytic or photoelectrocatalytic degradation of organic dyes. Entry no. Inorganic semiconductor Optical bandgap (eV) Organic dye(s) studied Reference(s) 1 2 3 4 5 6 7 8 9 10 11 12 CdS CdS ZnO ZnO ZnO ZnO ZnO ZnO SnO2 WO3 WO3 WO3 2.4 2.4 3.2–3.4 3.2–3.4 3.2–3.4 3.2–3.4 3.2–3.4 3.2–3.4 3.5 2.5–2.8 2.5–2.8 2.5–2.8 Methyl Orange Reactive Black 5 Rose Bengal Acid Orange 7 Acid Orange 20 Reactive Black 5 Methyl Red Procion Red MX-5B, Amaranth As in Entry No. 8 As in Entry No. 6 Naphthol Blue Black Methylene Blue [27,28] [87] [120] [34] [37] [43] [136] [150] [150] [43] [197,198] [199] immobilize the occluded particles. More modest voltages are used in this process variant relative to the pure electrophoretic approach. Thus Degussa P-25 TiO2 nanoparticles were electrophoretically deposited on stainless-steel substrates and used for the PEC decoloration of Methylene Blue [190]. A similar approach was used in another study along with carbon black to prepare composite films on conductive glass (ITO) [191]. These films were then used for the decoloration and degradation of Acid Orange 7 azo dye [191]. Silver was used as the glue to occlude TiO2 nanoparticles and these composite films were used on Brilliant Green dye [192]. A ␤-PbO2 film was electrodeposited along with the occlusion of TiO2 nanoparticles to yield a ␤-PbO2 –TiO2 composite film [193,194]. The efficacy of these materials was established with Acid Orange 7 azo dye [193,194]. Rounding out the electrosynthetically prepared photocatalysts is a study on anodic alumina (AAO) membranes which were used as a template to house colloidal TiO2 nanoparticles [195]. These immobilized photocatalysts were used for the conversion of Direct Black 168 dye [196]. We end this section on TiO2 photocatalyst with a brief mention of a potpourri of other preparative methods which include magnetron sputtering, chemical vapor deposition or CVD, electron beam evaporation and metal organic CVD or MOCVD. Titania films synthesized by these methods were used for the photodegradation of dyes such as Rhodamine B [196]. corrosion (see Section 7.7). Thus, of the photocatalysts that appear in Table 7, TiO2 , WO3 , and SnO2 are immune from photocorrosion while ZnO and CdS undergo photoanodic corrosion under bandgap illumination [13]. In most of the studies considered in Table 7, commercial samples of the photocatalysts were deployed. WO3 films were prepared by cathodic electrodeposition [197,198] while they were deposited on glass substrates by RF magnetron sputtering in another study [196]. A high PEC activity was observed for the electrodeposited WO3 film electrodes than for TiO2 nanoparticulate films for the degradation of Naphthol Blue Black dye [198]. Finally, anodic nanoporous WO3 films have shown greater photocatalytic activity for Methylene Blue than their cathodically electrosynthesized counterparts [199]. A crucial advantage with WO3 (relative to TiO2 ) is that its lower optical bandgap (Table 7) results in a much greater utilization of the solar spectrum for solar photocatalysis or solar photoelectrocatalysis applications. In summation of the results on the other semiconductors, comparisons of photocatalytic activity between different semiconductor materials may be confounded by variables related to their morphology. We have seen how important this is even for a single semiconductor (TiO2 ) which shows wide variations in photocatalytic activity for various dyes. 6. Photoreactor configurations 5.2. Other semiconductors After TiO2 , ZnO perhaps is the most studied inorganic semiconductor in the dye photocatalysis community. We have already seen (in Section 2.2) that ZnO actually preceded TiO2 as a semiconductor of choice in the dye sensitization field. Therefore, it is not surprising that early studies on photooxidative degradation of organic compounds via sensitized charge transfer by a dye (a process variant elaborated later) used ZnO as the inorganic semiconductor (e.g., Ref. [120]). Table 7 provides a listing of studies on ZnO and other semiconductors. Head-to-head comparisons of ZnO and TiO2 yield variable results. Thus ZnO was seen to outperform Degussa P-25 TiO2 in the photodegradation of Reactive Black 5 [43]. On the other hand, in terms of CO2 formation as a function of time, ZnO was slower than Degussa P-25 TiO2 [43]. Interestingly, ZnO outperformed Degussa P-25 TiO2 in the UV-assisted photocatalytic conversion of Acid Orange 22 but the trend was reversed in the photosensitization process [37]. For azo dyes, ZnO is reported to be better than TiO2 in some studies (e.g., Ref. [150]). A comparison between TiO2 and CdS photocatalysts also appears for a reactive azo dye [87]. The photocatalytic conversion of Reactive Black 5 obeyed Langmuir–Hinshelwood equation (see below) for CdS but not for TiO2 [87]. The dye wastewater was decolorized in both cases but the toxicity increased for CdS because of photo- The simplest photoreactor design, at a laboratory scale level, is the immersion well type schematically shown in Fig. 7a [45,200–202]. Provisions are made in this set-up for flushing the solution with air or inert gas (for example, N2 ) as needed and also for withdrawing solution aliquots periodically for analyses. In situ measurements within a diode array spectrophotometer can also be performed with a photocell design consisting of an optically transparent electrode substrate for the (transparent) photocatalyst film [203]. The electrical connections for applying bias potentials to the working electrode are incorporated into this design and the illumination light source is placed orthogonal to the light beam path of the spectrophotometer. Diffuse reflectance spectroscopy measurements have been performed [31,141,173,174] with similar cell designs. Fig. 7b illustrates a twin-compartment cell for photoelectrocatalysis experiments [162,184,185]. A single compartment cell design has also been used [182] although this is less desirable because of possible interference from the products generated at the counterelectrode. Batch recirculation is employed in many dye photocatalysis studies and this can be done with either powder photocatalyst suspensions or immobilized photocatalyst configurations. In the former case, a membrane-based particle separation step can also be incorporated for photocatalyst reuse [59,111]. A popular configuration involves the photocatalyst immobilized on a fabric or K. Rajeshwar et al. / Journal of Photochemistry and Photobiology C: Photochemistry Reviews 9 (2008) 171–192 Fig. 7. Schematic diagrams of: (a) an immersion well photoreactor design and (b) twin-compartment electrochemical cell for photoelectrocatalysis. See Ref. [162] for identification of the set-up components in (b). fiberglass mesh wrapped around the inner wall of borosilicate glass tubing. The UV lamp (usually medium-pressure Hg arc) is housed in a quartz inner tube such that the water to be treated flows in the annular space between the lamp housing and the (illuminated) photocatalyst surface. This design can be deployed both for batch recirculation and for continuous flow modes—the latter being amenable for pilot-scale commercial set-ups. Examples (with minor variations in design details) may be found in Refs. [29,58,118,166] in applications related to a variety of organic dyes. The shallow pond or dish photoreactor is another arrangement compatible with both slurry and immobilized photocatalysts; fluid recirculation can be easily built into this photoreactor design. An example for organic dye degradation applications may be found in Ref. [30]. The use of naturally buoyant TiO2 -coated beads is an intriguing approach that was first touted for the solar treatment and remediation of oil spills [205]. A similar design incorporating immobilized TiO2 on polystyrene beads has been combined with a pulsed baffled tube photochemical reactor and used for the photocatalytic mineralization of Methylene Blue [142]. A study incorporating a 1 L plunging tube laboratory photoreactor has been compared with results from field experiments performed at the Solar Platform in Almeria, Spain [41]. Two types of photoreactor designs available at this facility were used [41,204]. One is a cascade falling film photoreactor with the photocatalyst (for example, TiO2 ) supported on non-woven inorganic fibers. 181 Fig. 8. (a) Photooxidation and (b) photosensitized mechanisms for organic dye photocatalysis. Note that light absorption occurs in the semiconductor in (a) and by the dye in (b). The second is a compound parabolic collector photoreactor for experiments and demonstrations with powdered photocatalyst suspensions. A variety of dyes (thiazine, azo, anthraquinone) were included in the study described in Ref. [41]. 7. Photodegradation of organic dyes 7.1. Process variants and mechanistic aspects involving light absorption by semiconductor Consider first, light absorption by the inorganic semiconductor with TiO2 as a prototype (Fig. 8a). When photons with energy (h) equal to or greater than the semiconductor bandgap (Eg ) are incident, electrons (ecb − ) and holes (hvb + ) are generated in the conduction and valence band, respectively [3–5] (Eq. (8)): TiO2 + h → hvb + + ecb − (8) The photogenerated holes that escape direct recombination (Eq. (9)): hvb + + ecb − → heat (9) diffuse to the TiO2 surface (or migrate, if there is an imposed electric field) and react with surface adsorbed hydroxyl groups or water molecules to form trapped holes, Eq. (10): hvb + + ≡ TiIV OH → [≡ TiIV OH• ]+ → ≡ TiIV O• + H+ (10a) 182 K. Rajeshwar et al. / Journal of Photochemistry and Photobiology C: Photochemistry Reviews 9 (2008) 171–192 hvb + + TiO2 (H2 O) → TiO2 (• OH) + H+ (10b) The trapped holes may be regarded as surface-bound hydroxyl radicals. The bound radicals can also diffuse away from the surface toward the solution bulk and exist transiently as free • OH. Either the surface bound or free hydroxyl radicals can react with adsorbed or free dye molecules to oxidatively decompose the latter [3–5]. A preponderance of literature evidence (albeit not much of it on organic dyes, see below) exists for this mechanistic picture based on detection of hydroxylated products and intermediates and also spin-trapping with subsequent EPR detection [206–208]. Complicating the sequence of oxidative steps identified above is direct oxidation of the dye by the photogenerated holes. This pathway is thermodynamically feasible because the redox potential of the dye frequently lies above the semiconductor (e.g., TiO2 ) valence bandedge. The quantum yields for • OH and hole generation have been estimated (7 × 10−5 and 5.7 × 10−2 , respectively) [209] and this suggests that the direct hole transfer pathway should dominate. However, the • OH mediated oxidation route will be important if the dye is adsorbed on the semiconductor surface. Indeed, which of the two competing mechanisms prevail may well depend on the particular organic compound and the experimental conditions. A study on the use of • OH radical or hole scavengers on the TiO2 assisted UV photodegradation of Acid Orange 7 [210] indicated that the direct hole transfer pathway played a major role. We have already seen that in cases where • OH mediated oxidation is dominant, whether the reaction occurs on the semiconductor surface or in the solution bulk, depends crucially on the particular dye. What happens to the photogenerated electrons in Reaction (8) that escape subsequent recombination? They can be trapped at the titanium lattice sites to generate Ti3+ defects (color centers). Recall that photocatalysis usually occurs in an oxygenated aqueous medium so that the following chain of reactions can occur: ∼40–50 times smaller than that for the photoreduction of Methyl Orange by ecb − in CdS suspensions [27,28]. This reduction was promoted by a hole scavenger such as ethylenediamine tetraacetic acid (EDTA) from amongst several electron donors tested [27,28]. The typical UV flux near the terrestrial surface of the earth is ∼20–30 W/m2 in the 300–400 nm wavelength range. Unfortunately a semiconductor such as TiO2 (with an optical bandgap of 3.0–3.23 eV, absorption cut-off: ∼380 nm) harnesses only a small fraction (∼5%) of the entire solar spectrum. For practical considerations, the electrical costs associated with driving a UV lamp (see below) will have to be factored into the overall process economics [212] so that making the process amenable to using sunlight (by using a lower bandgap oxide semiconductor or doping TiO2 ) becomes eminently attractive. Thus it is not surprising that many studies as exemplified by Refs. [37,97,101,109,116,118] (see also the review on azo dyes, Ref. [9]) on the photocatalytic conversion of organic dyes have been performed in natural or simulated sunlight. It is pertinent to mention some cautionary remarks here on the use of light sources and UV cut-off filters (as suggested by a reviewer). Not all the UV component is cut-off with the use of filters and many “visible light” sources (e.g., tungsten halogen lamp) contain a UV component in their output such that the resultant data may lead to erroneous conclusions. It is therefore important to establish the wavelength response of a particular photocatalysis procedure (via photoaction spectroscopy) before conslusions are drawn. As also reviewed in Ref. [9] for azo dyes, solution additives can significantly influence the photooxidation of the organic dye. Thus additives such as H2 O2 and persulfate (S2 O8 2− ) can exert a dual function, first as strong oxidants themselves (see Reactions (13) and (14)) and Reactions (17) and (19): S2 O8 2− + ecb − → SO4 2− + SO4 •− (17) (18) ecb + O2(ads) → O2 •− (11) SO4 •− + H2 O → SO4 2− + • OH + H+ − + ecb + 2H → H2 O2 (12) SO4 O2 •− + H2 O2 → • OH + OH− + O2 (13) H2 O2 + ecb → H2 O2 + h → 2• OH (14) Second, as Reactions (17) and (20) show, these additives can function as electron scavengers minimizing electron–hole recombination at the semiconductor surface and thereby improving the quantum yield for • OH and hvb + production. Consistent with the above mechanistic picture, negligible chemical oxygen demand (COD) abatement was found for the textile dye, Orange 7, using TiO2 in the absence of S2 O8 2− while over 70% COD abatement was obtained with a combination of TiO2 , potassium persulfate, and sunlight [101]. Other studies attest to the beneficial effect of H2 O2 and (NH)4 S2 O8 addition for reactive azo dye decolorization with TiO2 and sunlight [97]. Too high a peroxide level can have an adverse effect on dye photooxidation. This is because H2 O2 is also a scavenger of semiconductor valence band holes and • OH: − O2 •− + We are ignoring back-reactions, for example: ecb − + ≡ TiIV O• + H+ → ≡ TiIV OH (15) in the above scheme. Incidentally, Reaction (15) may be regarded as a trap-mediated recombination of the electron with a surfaceimmobilized hole. Species such as peroxide and the superoxide radical (Reactions (11) and (12)) are important in AOPs and in photosensitized dye degradation as elaborated later. Evidence for H2 O2 generation under UV irradiation of oxide semicondcutors in oxygenated aqueous media was presented even in the early photocatalysis literature, albeit not in the context of organic dyes [211]. The organic dye can also be involved in an irreversible reductive photoreaction pathway involving the photogenerated electrons in the semiconductor conduction band. Thus laser flash photolysis data show evidence for formation of the radical anion of Phenosafranin dye in colloidal TiO2 and CdS suspensions [129] consistent with a one-electron reduction process: ecb − + D → D•− (16) where D is the dye molecule. Bandgap excitation of WO3 and TiO2 suspensions was shown to result in irreversible reduction of two azo dyes, Naphthol Blue Black and Disperse Blue 79 via a trapped electron pathway [132]. The reaction between the dyes and trapped electrons was reported to be diffusion-limited with rate constants of 1.1 × 108 and 4.0 × 107 M−1 s−1 for the above two dyes, respectively [132]. The rate constant for Reaction (11) was found to be •− + D → SO4 − 2− • OH + oxidized products of dye + OH − (20) H2 O2 + 2hvb + → O2 + 2H+ H2 O2 + • OH HO2 + • • OH → H2 O + HO2 → H2 O + O2 (19) (21) • (22) (23) Thus as also pointed out in Ref. [9], the reported influence of oxidant additives such as S2 O8 2− and H2 O2 can be conflicting depending on the dose and the specific experimental conditions. Other chemicals such as Na2 CO3 and NaCl are inherent in the dyeing process itself. For example, Na2 CO3 is added to adjust the pH of the dyeing bath—a crucial variable for fixing the dye onto the fabric and for color fastness. Sodium chloride is mainly used for transferring the dyestuff to the fabric. Therefore, the dye industry K. Rajeshwar et al. / Journal of Photochemistry and Photobiology C: Photochemistry Reviews 9 (2008) 171–192 wastewater has the propensity to contain a considerable amount of CO3 2− and Cl− making an examination of the influence of these species on the photooxidative dye degradation quite relevant. Both species can scavenge • OH and hvb + : CO3 2− + • OH → OH− + CO3 •− HCO3 − − + • OH + Cl + hvb → → H2 O + CO3 •− Cl• Cl• + Cl− → Cl2 •− (24a) (24b) (25a) (25b) Other anions such as SO4 2− and phosphate also have an inhibitory effect [213]. Interestingly, however, no adverse effects were observed in chloride- and sulfate-containing media on the extent of color removal and mineralization of a reactive textile azo dye as long as the pH was carefully controlled in the acidic range [159]. Cations such as Fe3+ can have a beneficial effect because of the possibility of Fenton reactions consuming H2 O2 in the aqueous phase [214]: 183 UV/H2 O2 relies on Reaction (14) with H2 O2 absorbing mainly in the deep UV region (wavelengths shorter than ∼200 nm) [3,218]. Thus while the reaction has been attempted in sunlight because of process energy efficiency considerations (see above), the yield of • OH will be very low. The use of ozone for water treatment is not a new concept, dating back to ∼1907 in the drinking water industry [3]. However, photolytic decomposition of ozone [3]: h O3 −→O(1 D) + O2 (27a) O(1 D) + H2 O → H2 O2 − (27b) + H2 O2 + H2 O ↔ HO2 + H3 O (27c) accompanied by Reactions (14) and (28)–(30): HO2 − + O3 → O3 − + HO2 • (28) O3 − + H2 O → HO3 • + OH− (29) Fe3+ + H2 O2 → Fe2+ + HO2 • + H+ (26a) HO3 + Fe2+ + H2 O2 + H+ → Fe3+ + H2 O + • OH (26b) results in the generation of • OH. Ozone has a strong absorption band centered at 260 nm with a molar extinction coefficient of ∼300 M−1 cm−1 [218]. The combination of UV radiation with H2 O2 and O3 (the so-called Peroxone process) increases the fraction of available ozone and possibly improves the mass transfer rate of ozone in the fluid phase compared to the use of ozone alone [3]. The Fenton reaction (Eqs. (26a) and (26b)) was discussed earlier. The photo-Fenton variant [3] provides for an added source of catalytic Fe2+ species (which are consumed via excess H2 O2 in Reaction (26b)) and • OH as follows: Thus the effect of iron species on the photodegradation of Acid Red 1, an azo dye, was studied in TiO2 aqueous suspensions [42]. The accumulation of H2 O2 was found to be completely suppressed during the photocatalytic process because of Reactions (26a) and (26b) above [42]. However, a confounding decrease in the rate of H2 O2 formation (via Reaction (12)) because of the competition between adsorbed Fe(III) species and O2 also cannot be ruled out here as pointed out by the authors themselves of this study [82]. This example may be taken as illustrative of the complexity of reaction pathways in a heterogeneous photocatalysis system. We close this section with a brief mention of miscellaneous aspects related to the photooxidation of organic dyes. First, attempts have been made to extend the wavelength response of TiO2 toward the visible so as to increase its absorption overlap with the solar spectrum. Thus, precipitation of anatase TiO2 layers on perovskite-type (Sr0.95 La0.05 )TiO3+ı particles yielded a composite that was able to photobleach Methylene Blue at wavelengths longer than ∼420 nm [147]. Another strategy is to dope TiO2 with non-metallic species such as boron, carbon, or nitrogen [215]. Dye photooxidation has been combined with Cr(VI) reduction in a TiO2 slurry reactor [216]. In this scheme, photogenerated electrons are utilized to reduce the toxic Cr(VI) to the environmentally benign and immobile Cr(III) species. This strategy is particularly relevant to practical remediation scenarios involving wastewater effluent streams containing both reducible species (such as metal ions) and oxidizable organic dyes. Yet another value-added innovation to dye photodegradation is combination with photoinduced hydrogen generation as exemplified by a study on the simultaneous degradation of azo dyes and H2 generation using irradiated Pt/TiO2 suspensions [100]. 7.2. Comparison with Advanced Oxidation Processes and combination with sonolysis It was mentioned at the outset of this article (Section 1) that heterogeneous photocatalysis fell under the umbrella of technologies called AOPs that rely on the generation and use of free radicals (predominantly • OH). We shall consider here how heterogeneous photocatalysis fares in comparison with their homogeneous process counterparts, namely UV/H2 O2 , UV/O3 , UV/H2 O2 /O3 , and electro-Fenton and photo-Fenton processes for treating organic dyes. A brief outline of each of these process candidates is given first. Further details may be found in Refs. [3–5,217]. • • OH + O2 (30) Fe3+ + h → Fe2+ (31) Fe(OH)2+ + h → Fe2+ + • OH (32) The electro-Fenton scheme involves the electrolytic generation of H2 O2 at the cathode (usually made of carbon) via the 2e− electroreduction of O2 : O2 + 2H+ + 2e− → H2 O2 (33) together with the generation of Fe2+ species at an iron anode (see, for example, Ref. [183], and citations therein). Sonolysis, like photolysis, is a proven method for generating free radicals in aqueous media [219,220]. Propagation of ultrasonic waves, especially at high frequencies (e.g., ∼600 kHz), leads to the formation of cavitation bubbles in the presence of dissolved gas in the medium [219,220]. The collapse of these bubbles spawns extremely high local temperatures (10,000 K) and pressures (10,000 atm). Under these extreme conditions, water molecules dissociate to form hydrogen and hydroxyl radicals: H2 O → H• + • OH (34) While sonolysis can be used in isolation for the oxidative treatment of textile dyes, the beneficial effect of combining sonolysis and photolysis [221] has been recognized in other studies on organic dyes [45,95,222,223]. Examples of organic dyes treated by the combination of UV/TiO2 and sonolysis include Reactive Black 5 [45], Naphthol Blue Black [95], and Remazol Black B [222]. 7.3. Photosensitized conversion of organic dyes in visible light Inspired by early work on the sensitization of wide bandgap oxide semiconductors (Section 2.2), visible light has been used to decolorize and decompose the dye, hopefully to completely mineralized products. Light absorption, unlike in the cases considered 184 K. Rajeshwar et al. / Journal of Photochemistry and Photobiology C: Photochemistry Reviews 9 (2008) 171–192 Table 8 Examples of visible light-induced photosensitized degradation of organic dyesa , b , c . Entry no. Organic dye/dye category Reference(s) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Rose Bengal/xanthene Methylene Blue/thiazine Basic Blue 4/azo Brilliant Red X-3B/azo 15 dyes (various types) Acid Orange 7/azo Acid Orange 7/azo Nile Blue A/oxazine Solvent Red 1/azo Rose Bengal/xanthene Erythrosine/xanthene Rhodamine B/xanthene Malachite Green/arylmethane Eosin/xanthene 7 dyes/various types Alizarin Red/anthraquinone Sulforhodamine B/xanthene 3 dyes/various types [120] [119] [48] [46] [98] [99] [143] [130] [31] [141] [106] [52–54,106] [65] [57] [92] [63,114] [56,107] [108] a See also Ref. [9] for examples of azo dyes. TiO2 used as the oxide semiconductor in all cases except in Entry 1 where ZnO was also used. c See also Ref. [60] for a review of work on Methylene Blue. b in the preceding two sections, now occurs by the dye molecule, D, adsorbed on the semiconductor surface (Fig. 8b): D + h(vis) → 1 D ∗ or3 D∗ (35) 1 (36) D ∗ or3 D ∗ + TiO2 → D•+ + TiO2 (ecb ) D+• → decomposition products +• D − + OH → D + • OH D + • OH → decomposition products (37a) (38) (37b) Reaction (36) is followed by Reaction (11) and the following sequence: O2 •− + H+ → HO2 • +• + O2 +• + HO2 → decomposition products D D •− • → DO2 → decomposition products (39) (37c) (37d) Excited states of the dye are highly reactive (relative to the ground state) and in addition to the oxidative (electron injection) scheme outlined above, the photoexcited dye can also undergo reduction by accepting an electron from the semiconductor conduction band. This is exemplified by the humic acid sensitized photoreduction of oxazine type dyes such as Oxazine 725 and Nile Blue [130] and dyes such as Methylene Blue and Thionine, which are converted to their (colorless) leuco form followed by irreversible decomposition of the dye [24,25,60]. Table 8 lists representative studies on the visible light-induced (photosensitized) degradation of organic dyes using inorganic semiconductor supports such as TiO2 . Reference [9] also provides examples of azo dyes that are converted to cation radicals and decomposition products in this manner. It is worth underlining that close proximity of the dye to the semiconductor surface (whether via adsorption or surfactant-induced binding, see above) is a pre-requisite for this photosensitization degradation mechanism (Fig. 8b). Contrastingly and illustratively, the dye can be stabilized against photodegradation by capping the semiconductor nanoparticles with an inert coating (such as polystyrene sulfonate, Ref. [91]) such that photoinduced electron injection from the excited dye molecule is inhibited. Interestingly, dye sensitization using visible light and an inorganic semiconductor (e.g., TiO2 ) can be used to decompose other organic pollutants (e.g., pesticides and herbicides); this topic has been reviewed [224]. Examples of this environmental remediation approach includes the use of visible light and TiO2 modified with Rose Bengal for the treatment of terbutylazine [225] and with azo and thiazine dyes for treating bromacil [94]. In many of these cases, it is not entirely clear to what extent the oxidized dye is regenerated to sustain the remediation process. This process could well be dye-limited recalling there is only a monolayer of the dye (or at best, a few monolayers) to begin with. In any case, this process is not central to the theme of this review article and is not considered further. Finally, simultaneous and synergistic conversion of organic dyes and metal ions in aqueous TiO2 suspensions under visible light illumination has been described [108]. The presence of Cu2+ or Fe3+ was found to have a deleterious effect on the visible light degradation of Sulforhodamine B, Alizarin Red and Malachite Green while Zn2+ , Cd2+ or Al3+ only had a “slight” influence [108]. These differences were attributed [108] to the decreased reduction of O2 by the semiconductor conduction band electrons in the presence of reducible metal ions and consequently diminished formation of reactive oxygen species (O2 − , • OH) (Reactions (12) and (13)). On the other hand, Zn2+ , Cd2+ or Al3+ are not reduced by the conduction band electrons in TiO2 since their redox levels are not located below the conduction bandedge. 7.4. Photoreactions at the solid/air interface or in the gas phase The oxide semiconductor particles can be stirred with a nonaqueous solution of the dye so as to adsorb the dye onto the surface (Section 3) (for example, Ref. [141]). The stirred particles can be vacuum-treated for evaporating off the solvent and then air dried. Photochemical reactions can then be performed on the dye loaded solid samples in air [31,32,90,141]. For the dye photodegradation to be effective (via either of the mechanisms discussed above), it is imperative that O2 be adsorbed on the semiconductor surface to function as an electron scavenger (see Reaction (11)). Importantly, control experiments performed with Al2 O3 show no photodegradation of the dye signaling that the semiconductor characteristics of the oxide support play a crucial role. Thus organic dyes such as Solvent Red 1 [31], Acid Orange [32,102], Naphthol Blue Black [90], Rose Bengal [141], Acid Blue 9 [102,169], Reactive Blue 19 [102], Reactive Black 5 [102,169], and Indigo [66] have been photodegraded in this manner. Instead of nanosized TiO2 particles, films of the oxide semiconductor on glass (for example, Pilkington ActivTM photocatalytic glass) can also be used (for example, Ref. [169]). Remote bleaching of Methylene Blue by UV-irradiated TiO2 has been reported in the gas phase [124]. A rear illuminated TiO2 -coated glass plate was arranged to face another glass plate coated with the dye with the two surfaces separated by a small gap (12.5–500 ␮m). The authors posit that the dye is not simply bleached to its leuco form but is oxygenated or decomposed by active oxygen species generated on the irradiated TiO2 surface and transported through the gaseous inter-annular space to the Methylene Blue dye layer [124]. 7.5. Photoreaction kinetics, variables, and related aspects In this section, we review data on the kinetics of organic dye photodegradation in UV- or visible light-irradiated semiconductor suspensions and films. Like in all the case discussed above, these studies have mostly featured TiO2 as the semiconductor photocatalyst. The kinetics experiments involve monitoring of either changes in the dye absorbance (at a specific wavelength, usually, max ) or the total organic carbon (TOC) (measured via chromatographic proce- 185 K. Rajeshwar et al. / Journal of Photochemistry and Photobiology C: Photochemistry Reviews 9 (2008) 171–192 dures) as a function of time. The CO2 evolved, from organic carbon decomposition, can also be monitored as a function of time (see below). It must be borne in mind the temporal changes in these three parameters are not identical (see below). The kinetics data are then fit to specific models of which the Langmuir–Hinshelwood (L–H) mechanism [3,5,226] appears to be the most popular: R=− dC k1 KC = 1 + KC dt (40a) where k1 and K are constants. At low dye concentrations and when the product KC ≪ 1, Eq. (40a) collapses to an expression similar to that applicable for first-order kinetics: − dC = k1 KC = k′ C dt (40b) In Eqs. (40a) and (40b), C is the (time-dependent) dye concentration and k′ in Eq. (40b) is a pseudo-first-order rate constant. At the other extreme of high dye concentrations when KC ≫ 1, Eq. (40a) reduces to: − dC = k1 dt (40c) This is the zero-order reaction kinetics limit with a constant reaction rate. This behavior is diagnosed by linear plots of C versus t unlike the exponential decay seen in first-order kinetics processes. First-order or zero-order kinetics behavior has been seen in a number of instances in the organic dye photocatalysis literature. Tables 9 and 10 list studies where the temporal photoconversion of the dye has been monitored. As with their counterparts in Tables 2 and 3, azo dyes dominate the literature (with an extremely diverse group) and they are considered in Table 9 while non-azo dyes are considered in Table 10. Rearrangement and integration of Eq. (40b) yields: ln C0 = k′ t C (40d) In Eq. (40d), C0 is the initial concentration of the dye. Semi-log plots of C0 versus t yield straight-line fits of the kinetics data from which values for k1 (in units of min−1 ) can be extracted from the slopes. Alternately, a linear transform of Equation (40a) yields: 1 1 1 1 = + R k1 k1 K C (40e) where R is the photodegradation (or decolorization) rate embodied by the left-hand term in Eq. (40a). Thus an inverse plot of the rate versus the equilibrium dye concentration (C) would yield values for k′ and K from the intercept and slope, respectively from which corresponding values for k′ can be extracted (k′ = k1 K). Eqs. (40a) and (40e) predict (a Langmuir-type) asymptotic behavior where R increases non-linearly with C ultimately reaching a (surfacelimited) plateau value (given by k′ ) (Eq. (40c)). This can also be seen from Eq. (40e); as C → ∞, R → k1 . Literature values for the pseudo first-order rate constant (k′ ) show a very wide variation for the azo and other dyes, respectively. Thus even if some of the anomalously high values for k′ are discounted, there is a ∼3 orders of magnitude variance in the reported k′ values for the azo dyes and a corresponding variance of ∼500 for the non-azo dyes. Some of these variations can be rationalized on the dependence of k′ on process variables such as the type of photocatalyst used, the illumination details, the initial dye concentration, photon flux, solution pH, the presence of oxidants in the solution such as H2 O2 or Na2 S2 O8 , etc. The effect of inorganic ions on dye adsorption and free radical photochemistry was mentioned in the preceding sections. Their effect on the kinetics of dye photoreactions has been addressed in several studies. Thus Ag+ ions aided in the photodecolorization of Methyl Orange to a greater extent than Cu2+ , Co2+ , Fe3+ , and Ce4+ ions [122]. Similarly, NO3 − was better as a dye counterion than SO4 2− [122]. Transition metal ions such as Cu2+ and Fe3+ were shown to markedly depress the visible light-induced photodegradation of Sulforhodamine B, Alizarian Red, and Malachite Green while other metal ions such as Zn2+ , Cd2+ , and Al3+ had only a slight influence [108]. Sulfate and chloride ions were reported to not significantly affect the reaction rate for Methyl Orange [97]. SO4 2− , Cl− and NO3 − ions decreased the photocatalytic degradation rate Table 9 Representative kinetics studies on the photodegradation of azo dyes in TiO2 suspensions or filmsa . Organic dye Comments Reference(s) Methyl Orange The first azo dye to be studied from a kinetics perspective. Both photooxidation and photoreduction of dye observed. Both the L–H model and other (steady-state approximation derived) models used for analyzing photodegradation kinetics. Pseudo first-order kinetics (Eqs. (40b) and (40d)) observed. Both TOC and dye concentration monitored. Pseudo first-order kinetics observed for both TiO2 suspensions and immobilized photocatalyst. Exponential decay of dye concentration with time seen. Pseudo first-order rate law seen to be obeyed. Pseudo first-order rate law seen to be obeyed. Exponential decay in dye absorbance seen in the photoelectrocatalytic mode. Pseudo first-order kinetics seen. Pseudo first-order kinetics seen. A popular azo dye for kinetics studies. – Degradation said to “approximately” follow first-order kinetics. – – First-order kinetics seen for TiO2 films. – – – – L–H model adherence and pseudo first-order kinetics. Photoelectrocatalytic data shown in terms of dye absorbance vs. time. Both TOC and COD disappearance show initially zero-order kinetics. [27,28,39,40,97,122,123,146,199] Acid Orange 7 Acid Orange 20 Orange G Orange IIb Methyl Red Reactive Red 2 Reactive Orange 16 Reactive Red 120 Reactive Red 141 Reactive Black 5 Congo Red Solvent Red 1 Direct Blue 160 Basic Yellow 15 Basic Blue 41 Disperse Orange 1 Disperse Red 1 Disperse Red 13 Direct Black 38 Safira HEXL Remazol Brilliant Orange 3 R Amaranth a b TiO2 was used as the photocatalyst unless otherwise stated. Also known as Acid Orange 7. [88,93,100,143,166,193,210] [37] [88,96] [92,97] [96] [151] [159] [89] [133] [43–45,87,169] [96,97] [88] [89] [89] [166] [184,185] [184,185] [184,185] [49] [50] [158] [51] 186 K. Rajeshwar et al. / Journal of Photochemistry and Photobiology C: Photochemistry Reviews 9 (2008) 171–192 Table 10 Representative kinetics studies on the photodegradation of non-azo dyes in TiO2 suspensions or films. Organic dye Comments Reference(s) Thiazine Methylene Blue One of the earliest dyes to be studied from a photodegradation kinetics perspective. [29,30,58,59,61,96,98,112,119] A kinetics model incorporating O2 and sacrificial electron donor concentration developed in Ref. [105]. First-order kinetics used in Ref. [30]. Pseudo first-order kinetics seen. Degradation kinetics follows L–H model under visible light irradiation. [30,105,155] Anthraquinone Anthraquinone 2-sulfuric acid Acid Blue 25 Acid Blue 80 Acid Blue 40 Alizarin S Pseudo first-order kinetics seen. Kinetics plots shown using UV–vis spectrophotometry, TOC, and HPLC-diode array detection data. Pseudo first-order kinetics seen. As above but at pH 3 only. – [113] [64] [167] [89] [96] Arylmethane Malachite Green L–H model and its linear transform (Eq. (40e)) used. [65] Indigo Indigo and Indigo Carmine L–H model and its linear transform (Eq. (40e)) used. [66] Phthalocyanine Direct Blue 87 L–H model adherence found at certain pHs but not at others. [89] Cyanine Fluorescein L–H model and pseudo first-order kinetics used in continuous recirculation reactor. [29] Xanthene Rhodamine B Rose Bengal Eosin for Acid Orange 7 [210]. The above discussion makes it clear that these effects (as with the other variables) are very much specific to each dye system and possibly very sensitive to the process conditions employed in a particular study such that generalizations are difficult. Chemometrics [227,228] offers a versatile route to unraveling the effects of these variables on the photocatalysis rate and also for process optimization. Thus a study on Safira HEXL dye [50] shows a response surface (see Fig. 7 in Ref. [50]) of the initial rate as a function of initial dye concentration and solution pH that is curved and complex diagnosing interaction between the two variables. In general, multivariate statistical analyses and central composite experimental designs are powerful tools for screening and optimization of photodegradation rates. Especially, interactions between process variables translates to the fact that traditional experimental designs relying on changing one variable level at a time (while holding other variables constant) will simply not be efficient and may even lead to misleading conclusions. 7.6. Photoreaction intermediates Decolorization of the dye-laden wastewater is but one central objective of heterogeneous photocatalytic treatment. Total mineralization of the dye is an equally crucial end goal. Thus all the original carbon and hydrogen in the organic dye must be converted to CO2 and water. Any nitrogen in the dye (as in the azo compounds) must be similarly converted to N2 , NH4 + and nitrate ions, sulfur to sulfate, phosphorus to phosphate, chlorine to chloride etc. Thus aside from monitoring absorbance changes during the course of the photocatalytic remediation process, other complementary measurements must be carried out to get a total picture of the process efficacy. These include the monitoring of the release of inorganic ions and evolution of CO2 from the dye. Determination of the total organic carbon and/or measurement of the chemical oxygen demand (COD) and biological oxygen demand (BOD) will afford reliable indicators of the extent of mineralization of the dye. In particular, total decolorization of the dye may still be accompanied by significant residual TOC in the solution, signaling that color- [98,120,182] [57,98] less (organic) intermediates are formed during the process. Often these intermediates may be as toxic as the original dye. There are even cases where the original dye is not toxic but the photocatalytic reaction intermediates are (see below). The fate of the original dye has not been tracked in all the studies included in this review. However, many studies do discuss this aspect and Tables 11 and 12 contain examples of these, again for azo compounds (Table 11) and for non-azo dyes (Table 12). Reference [9] contains a good discussion of the nature and evolution of organic intermediates for mono-, di- and triazo dyes and for triazine ring-containing azo dyes. The major degradation pathways are identified for two azo dyes, Acid Orange 7 and Acid Orange 52, respectively [9]. Rather than duplicate the discussion, the reader is instead referred to this prior review article [9] for details. The triazine ring is a recurrent feature in the chemical structure of many azo dyes of the reactive dye family. Triazines are six-membered rings containing three nitrogen atoms in the aromatic ring structure [229]. Instead of terminating in CO2 and other mineralized products (see above), the triazine ring typically yields cyanuric acid via ring hydroxylation [9,229] attesting to its recalcitrant nature. Fortunately, this compound is not toxic as further elaborated in the next section. 7.7. Toxicity studies An early study on the extraction (with methylene chloride) of organic compounds from a dye manufacturing plant wastewater, revealed the presence of several toxic chemicals in it [230]. More recently, a textile azo dye processing plant effluent was identified as a major source of mutagenic activity [231]. Benzidine, a known carcinogen, was additionally found in this effluent [231]. These results underscore the need for proper treatment of the dye wastewater prior to discharging it into the environment. Pertinent to this issue is the question of whether heterogeneous photocatalysis is capable of mitigating the toxicity of the dye effluent. It is therefore surprising that, notwithstanding the existence of hundreds of papers on the treatment of dye-laden water samples using this approach (e.g., Figs. 1 and 2), only a handful of studies 187 K. Rajeshwar et al. / Journal of Photochemistry and Photobiology C: Photochemistry Reviews 9 (2008) 171–192 Table 11 Representative studies probing the formation of reaction intermediates during the heterogeneous photocatalytic conversiona of azo dyes. Dye(s) Technique(s) used Comments Reference(s) Methyl Orange HPLC–MS [97] Acid Orange 7b COD monitored H2 O2 formation via a spectrophotometric assay GC–MS FT-IR spectroscopy Progressively demethylated products found. Hydroxylated products found in laboratory experiments but not observed in field tests. – TOC content only ∼80% after 3 h. C2 aliphatic acids, carbonated and oxygenated sulfur compounds detected. Both qualitative and quantitative detection of intermediates reported. Visible light irradiation leads to decolorization but not complete mineralization. TOC removal found to be pH dependent. UV–vis spectra during reaction indicate quinonic intermediate formation. – Decomposition of the azo group more effective than that of the naphthalene ring structure. Mineralization of the nitrogen more effective than sulfur (from the sulfonate group). Original dye transformed to CO2 and aliphatic (formic, acetic, oxalic) acids which are converted more slowly by trapped holes and • OH. A multiplicity of pathways including demethylation and • OH attack of the aromatic rings or to the azo-linkage bearing carbon. Oxalic acid identified as main degradation product from photoelectrocatalysis, 56–62% TOC reduction observed. Ion evolution found not to track one another for sulfate and nitrate. – Compounds containing aromatic rings as well as some short chain aliphatic acids (butanoic and propanoic) detected. Detailed degradation pathway proposed. [88] [35] ∼70% TOC reduction obtained after ∼3 h of photoelectrocatalysis. Aromatic intermediates detected. [184,185] Complete mineralization reported although color removal slower than that of Amaranth itself diagnosing the intermediate formation of hydroxylated or desulfonated Amaranth. Sulfate evolution rate approaches overall dye degradation rate diagnosing the first step to be C–S bond cleavage. – – [51] Acid Orange 7b Acid Orange 7b Acid Orange 7b Acid Orange 7b Acid Orange 7b COD measurement, GC/MS, FT-IR and ion chromatography H2 O2 determination via colorimetry HPLC, TOC measurement, UV–vis and FT-IR spectroscopy Acid Orange 7b Orange G COD measurement Ion chromatography and TOC measurement Acid Orange 20 FT-IR and TOC Acid Orange 52b , c GC–MS, HPLC and 1 H NMR Reactive Orange 16 HPLC and UV–vis spectrophotometry Reactive Yellow 145 and Reactive Black 5 TOC measurement and ion chromatography Ion chromatography GC–MS Reactive Black 5 Reactive Black 5 Congo Red Disperse Orange 1, Disperse Red 1 and Disperse Red 13 Direct Black 38 Amaranth Ion chromatography and electrospray-MS TOC measurement FT-IR spectroscopy and UV–vis spectroscopy TOC and COD measurements, HPLC and ion chromatography Amaranth Ion chromatography 8 azo dyes Commercial leather dye (not identified) COD measurement NH4 + assay by colorimetry FT-Raman and FT-IR spectroscopy UV–vis spectrophotometry Diffuse reflectance FT-IR spectroscopy Blue textile dye (not identified) Naphthol Blue Black Naphthol Blue Black Procion Red MX-5B a b c HPLC Ion chromatography – A quinonic product speculated as a result of the photosensitized degradation of the dye on TiO2 . As above. Quantity of sulfate and chloride ions found to be sub-stoichiometric. [92] [36] [99] [194] [93] [183] [37] [39] [159] [44] [43] [45] [97] [49] [150] [33] [168] [175] [90] [197] [150] TiO2 photocatalyst was used in all the cases except in Refs. [150,175,197] where ZnO, CdO–ZnO, and WO3 were used, respectively. See also Ref. [9] for further details. Also known as Methyl Orange. have specifically addressed the toxicity issue. A variety of toxicological assays exist based on monitoring the inhibition of growth of bacteria, algae, mammalian cell lines, or aquatic organisms. Two of the more popular candidates are the Microtox® method and inhibition of Escherichia coli growth. In the former assay, luminescence from the bacterium, Vibrio fisheri, is used as an indicator of medium toxicity. Thus Microtox® assay was used in photocataly- sis studies on Reactive Black 5 [87], Remazol Turquoise Blue [162], and in an electrochemical treatment study of textile dyes and dyehouse effluents [232]. Inhibition of E. coli respiration was used as a toxicity indicator in a study on combined photocatalysis and ozonation of a textile effluent [233]. Other types of assays have also been used: inhibition of metabolic degradation of phenol [44], death of Artemia salina cysts [49], and Salmonella/microsome assay 188 K. Rajeshwar et al. / Journal of Photochemistry and Photobiology C: Photochemistry Reviews 9 (2008) 171–192 Table 12 Representative studies probing the formation of reaction intermediates during the heterogeneous photcatalytic conversiona of non-azo dyes. Dye/dye category Technique(s) used Comments Reference(s) Methylene Blue/thiazine CO2 analyses by conductivity measurement. COD and TOC measurements, HPLC, ion chromatography, GC–MS, LC–MS CO2 formation seen to lag considerably behind the rate of dye disappearance. A detailed degradation pathway developed incorporating ring opening and formation of aromatic intermediates. Almost complete mineralization of carbon, nitrogen, sulfur into CO2 , NH4 + , NO3 − and SO4 2− observed. Photobleaching found at low dissolved oxygen concentrations although colored intermediates also found. A comparative study also involving anthraquinone and azo dyes. As above. [58] Methylene Blue/thiazine Methylene Blue/thiazine TOC measurement and UV–vis spectrophotometry. Methylene Blue/thiazine HPLC, ion chromatography, TOC and COD measurements. HPLC, ion chromatography, TOC and COD measurements. COD, 1 H NMR, and EPR. Methylene Blue/thiazine Rhodamine B/xanthene Rhodamine B/xanthene Sulforhodamine B/xanthene Eosin/xanthene Alizarin Red/anthraquinone Alizarin Red/anthraquinone Alizarin Red/anthraquinone Acid Blue 80/anthraquinone Acid Blue 25/anthraquinone Reactive Blue 4/anthraquinone Indigo Carmine/indigo Remazol Turquoise Blue 15/metallo-phthalocyanine Squarylium cyanine dye a H2 O2 evolution by colorimetry and COD measurements. TOC measurements and above techniques combined with 1 H NMR, EPR, and GC–MS. GC-TCD and ion chromatography COD measurements combined with 1 H NMR, UV–vis spectrophotometry, IR spectroscopy, GC–MS and ESR ESR and GC–MS along with MO calculations As in Ref. [61] above CO2 by headspace GC, ion chromatography, and HPLC–MS HPLC-diode array detection. TOC measurement, HPLC–MS TOC measurement, 1 H NMR and IR spectroscopy TOC measurement, CO2 by GC headspace analysis, GC–MS, ion chromatography HPLC-diode array detection, Cu assay by anodic stripping voltammetry 1 H NHR, UV–vis spectrophotometry, ESR and GC–MS De-ethylation and oxidative degradation found to occur in stepwise fashion. Degradation rate correlated with H2 O2 accumulation. Extent of N-de-ethylation found to depend on mode of dye binding to TiO2 surface. [61] [112] [96] [41] [54] [92,108] [56,107] Bromide ion release along with CO2 evolution found concomitantly with the photodecomposition of dye. Peroxides and carbonyl species along with phthalic acid found as first intermediates. No organo-peroxides found. [57] Pathway predicted by frontier electron densities and point charges on all the dye atoms. – Substrate hydroxylation and C–N bond cleavage lead to unstable intermediates. Two different kinetics found for decolorization and dye degradation diagnosing the occurrence of colored reaction intermediates. After 1 h of photoelectrolysis, total color removal observed but only 37% mineralization. Dye decolorization not accompanied by TOC reduction or release of inorganic anions. [63] 100% color removal accompanied by 83% TOC reduction. 69% of copper in the macrocycle is released as free metal ions. H2 O2 generated during visible light photo-degradation but no organo-peroxides found. Cleavage of the cyanine C C bond dominated yielding 1-sulfopropyl-3,3dimethyl-5-bromoindolenium-2-one. This intermediate not excited by visible light and photo-degradation stops. [160,162] [114] [41,96] [115] [64] [161] [66] [117] TiO2 was used as the photocatalyst. [231]. The above studies were all performed on azo dye-laden water samples. How effective is heterogeneous photocatalysis in reducing the toxicity of dye-laden water? Toxicity reduction was seen for Reactive Black 5 but after a treatment period of several hours, residual toxicity was still found [45,87]. Toxicity data are often quantified in terms of the parameter, EC50 , which is the effective concentration of a toxin causing 50% reduction in luminescent light emission in the assay. Thus the 15 min EC50 values for Procion Red MX-5B and Reactive Brilliant Red K-2G were found to be 23.57% and 21.68%, respectively at an initial dye concentration of 50 mg/L [234]. Complete detoxification (but not mineralization) was obtained in these cases diagnosing that the final product, cyanuric acid, is not toxic. Effective detoxification was also noted for an unspecified leather dye after UV/TiO2 treatment [49]. Very recently, we have studied a novel class of mutagenic compounds, derived from dinitrophenyl azo dyes assigned as: C.I. Disperse 373, C.I. Disperse violet 93 and C. I. Disperse Orange 37, respectively [235]. These dyes were identified as contaminants in rivers in Brazil, accounting, in some cases, for at least 50% of the mutagenic activity detected in the water bodies. The degradation of these mutagenic dyes was instigated by photoelectrocatalytic oxidation. All photoelectrochemical experiments were conducted using Ti/TiO2 thin film photoelectrodes prepared by the sol–gel method and their mutagenic responses were compared and analyzed by using the strains YG1041 and YG1042 of Salmonella typhimurium. Using optimized conditions for photoelectrocatalytic oxidation carried out on Ti/TiO2 anode and UV illumination in 0.25 M NaCl, 100% of decoloration and 60% of total organic carbon reduction were obtained after 120 min of photoelectrocatalysis [235]. In addition, the method pro- K. Rajeshwar et al. / Journal of Photochemistry and Photobiology C: Photochemistry Reviews 9 (2008) 171–192 moted an amelioration of the mutagenic properties of these dyes (83% revertants/L), A variety of confounding factors not related to the organic dye itself can influence toxicity results. The use of ZnO as photocatalyst leads to a very toxic medium for the bacterium, V. fisheri, as a result of ZnO photodissolution, which releases Zn2+ ions into the aqueous medium [236]. Similarly, CdS photocatalysis increased the medium toxicity because of CdS photocorrosion and release of Cd2+ ions [87]. Release of copper ions from the phthalocyanine macrocyclic assembly because of photodegradation caused an increase in the acute toxicity for V. fisheri [162]. Finally, the use of EmulsogenTM surfactant to solubilize Disperse Orange 1, Disperse Red 1 and Disperse Red 13 azo dyes resulted in acute toxicity for 293F cell lines [184,185]. 8. Into the real world: tests with dye wastewaters, process scale-up and economic aspects Real world dye wastewaters often contain high levels of suspended solids as well as other additives including salt (see above). Relatively few studies exist where the efficacy of heterogeneous photocatalysis has been assessed to treat wastewater effluents from dyeing industries. This void is exacerbated by the fact that many of the titles of papers in the relevant literature are misleading in that they contain the word “wastewater” in them, although the test samples used in these studies were actually pristine water to which dye had been added (e.g., Refs. [196,237]). Further, even in some cases where real-world dye wastewater samples were used, their source was not fully described nor was their characteristics (TDS, pH, TOC, COD, BOD, etc.) identified. A study on dye waste containing a mixture of reactive azo dyes compared the efficacy of UV/TiO2 with the photo-Fenton process; UV/TiO2 was found to be less effective for bleaching and degradation of the dye waste [237]. On the other hand, another study on municipal wastewater contaminated with textile dyes found that TiO2 -based photocatalysis was a viable method for decolorizing and oxidizing the organics in the dye wastewater [104]. Wastes from dye rinse baths were compared and blue-colored rinse was found to degrade faster than pink, orange, or yellow colored wastewater [238]. Yet another study reports a lower degree of effectiveness and this study also compares raw, coagulated and biologically pretreated textile effluent [239]. UV/TiO2 (or even solar/TiO2 ) could prove to be a versatile polishing step for water streams to be posttreated by other (e.g., biodegradation) methods. Interestingly, TiO2 -based photocatalysis can also improve the biodegradability of dyeing wastewater. The increase in the BOD value observed on photocatalysis can be taken to signal this possibility. This was confirmed in a study on anthraquinone 2-sulfonic acid (sodium salt) [113]. Another study on Reactive Yellow 145 [44] subjected the products of photocatalysis to bacterial degradation after 6 h of UV/TiO2 pre-treatment. A 47% TOC reduction was noted after 4 h of biological post-treatment while the original (unirradiated) sample showed only a 4% reduction [44]. Switching focus to process scale-up efforts, the section on photoreactor designs (Section 6) already alluded to studies going beyond the laboratory scale [41,204]. Water throughput volume is always an important design criterion in a practical water treatment system and in this regard, photoreactor designs incorporating immobilized photocatalyst films are likely to be more practical. However, films are also prone to fouling and residue accumulation on their surface. The final aspect considered in this penultimate section on the more pragmatic aspects of organic dye photocatalysis, concerns process economics. It may be argued that costing exercises may 189 be a little premature at this juncture considering the lack of many pilot-scale studies (see above). However, examining how UV/TiO2 would fare in an energy-efficient and economic sense against other AOPs is still relevant given that a common thread here is the cost of electrically driving the UV lamps. A useful parameter here is EE/O, which is defined as the electrical energy in kilowatt hours (kWh) required to instigate the degradation of a pollutant by one order of magnitude in 1000 L of contaminated water or air [212]. For the purpose of this review, we can adapt EE/O to be the electrical energy required to remove the initial color of the organic dye by one order of magnitude. In one study, EE/O was compared for two inorganic semiconductor candidates (TiO2 and ZnO) [240] while in two other studies, this parameter was estimated and compared for UV/TiO2 against other AOP variants [241,242]. Thus EE/O was three times higher for TiO2 than for ZnO in the case of Reactive Blue 19 dye [240]. The use of UV alone resulted in very high EE/O values for Acid Orange 7 and the values were ordered thus for other process variants: UV/TiO2 > UV/TiO2 /H2 O2 > UV/TiO2 /IO4 − with the specific values being dependent on the oxidant (H2 O2 and IO4 − ) concentrations [241]. In another study, the following ordering was reported for the EE/O values for Reactive Orange 4 and Reactive Yellow 14 azo dyes: UV/H2 O2 > UV/TiO2 > Fe2+ /H2 O2 /UV [242]. The EE/O calculations, of course, are necessary but not sufficient, and the costs of chemicals and capital outlays must be factored in to gain an overall comparative picture of the process economics. 9. Concluding remarks This review has shown that heterogeneous photocatalytic treatment of organic dyes is a mature science that has percolated to all parts of the globe. On the scientific front, while much has been learnt about the process (in all its interesting variations) and mechanistic aspects, research on new generations of photocatalyst materials (moving beyond TiO2 ), reaction intermediates/pathways, and toxicity issues must continue. Especially much can be learnt from corresponding advances in the discovery of new oxide photocatalyst materials for solar-driven hydrogen generation from water [215]. While effluent streams from dyeing processes are fairly nontoxic in many cases (relative to trends with other organic materials such as pesticides), the potential exists for generating toxic by-products and intermediates. Photocatalysis appears to be unsurpassed in its capability to remove color quickly (unlike biological treatment) and it appears to be energy-efficient relative to AOP competitors such as UV/H2 O2 . Other AOP candidates such as ozonation or UV/O3 (and variants thereof), which are also very effective for dye decolorization and degradation, have accompanying handicaps. For example, the chemical instability of ozone necessitates its generation on site and its hazardous nature requires proper venting to be built into the water treatment system. On the other hand, the visible light-induced photosensitization approach based on TiO2 (and other oxide semiconductors) is particularly amenable to use in developing country scenarios and in tropical or desert locales where solar insolation is plentiful. On the technological front, the scientific advances summarized in this review article, must be translated to large-scale systems capable of continuously and reliably treating large volumes of dyeing wastewaters. Adaptation of the photocatalytic process to treat poorly soluble dyes and dye waste sludge/cake must also be developed. In this regard, very few pilot and/or field studies exist at present to assess the large-scale feasibility of the photocatalysis approach. Indeed the field is now poised to move from the 190 K. Rajeshwar et al. / Journal of Photochemistry and Photobiology C: Photochemistry Reviews 9 (2008) 171–192 chemistry research laboratories to the chemical engineering/civil engineering/environmental engineering realm for further breakthroughs and advances. Finally, process economics comparisons pitting UV/TiO2 and other process variations (discussed above) against other more conventional methods such as biological treatment, adsorption, chemical coagulation, etc. for treatment and remediation of dye wastewaters, are lacking as are many studies aimed at evaluating the use of heterogeneous photocatalysis to supplement these existing methods as a pre- or post-treatment priming or polishing step, respectively in a hybrid scheme. Acknowledgments One of us (K.R.) thanks the University of Texas System for a Science and Technology Acquisition and Retention (STAR) grant and the U.S. Department of Energy (Basic Energy Sciences) for partial funding support. We also thank Rohini Krishnan, Reena Krishnan and Asha S. Nair for assistance in manuscript preparation. Profs. Akira Fujishima, Prashant Kamat, Leonardo Palmisano, and Nick Serpone are thanked for comments on an initial draft of the manuscript. Finally, the two anonymous reviewers are thanked for constructive criticisms of an earlier manuscript version. References [1] H. Zollinger, Color Chemistry, VCH, Weinheim, 1991. [2] D. Brown, A. Mark, S.C. De Vito, Crit. Rev. Environ. Sci. Technol. 23 (1993) 249. [3] K. Rajeshwar, J. Ibanez, Environmental Electrochemistry, Academic Press, San Diego, 1997. [4] O. Legrini, E. Oliveros, A.M. Braun, Chem. Rev 93 (1993) 671. [5] G.R. Helz, R.G. Zepp, D.G. Crosby (Eds.), Aquatic and Surface Photochemistry, Lewis Publishers, Boca Raton, FL, 1994. [6] J.T. Spadaro, L. Isabelle, V. Renganathan, Environ. Sci. Technol. 28 (1994) 1389. [7] K. Rajeshwar, C.R. Chenthamarakshan, S. Goeringer, M. Djukic, Pure Appl. Chem. 73 (2001 1849). [8] K. Rajeshwar, J.G. Ibanez, J. Chem. Educ. 72 (1995 1044). [9] I.K. Konstantinou, T.A. Albanis, Appl. Catal. B: Environ. 49 (2004) 1. [10] A. Mills, R.H. Davies, D. Worsley, Chem. Soc. Rev 22 (1993) 417. [11] P.V. Kamat, Chem. Rev 93 (1993) 267. [12] M.R. Hoffmann, S.T. Martin, W. Choi, D.W. Bahnemann, Chem. Rev. 95 (1995) 69. [13] K. Rajeshwar, J. Appl. Electrochem. 25 (1995 1067). [14] A.L. Linsebigler, G. Lu, J.T. Yates Jr., Chem. Rev. 95 (1995) 735. [15] A. Mills, S. LeHunte, J. Photochem. Photobiol. A: Chem. 108 (1997) 1. [16] M.I. Litter, Appl. Catal. B: Environ. 23 (1999) 89. [17] D.A. Tryk, A. Fujishima, K. Honda, Electrochim. Acta 45 (2000) 2363. [18] A. Fujishima, T.N. Rao, D.A. Tryk, J. Photochem. Photobiol. C: Photochem. Rev. 1 (2000) 1. [19] D.M. Blake, Bibliography of Work on the Heterogeneous Photocatalytic Removal of Hazardous Compounds from Water and Air, NREL/TP-570-26797, update No. 4 (to October 2001), National Renewable Energy Laboratory, Golden, CO, 2001. [20] K. Rajeshwar, Guest Ed., J. Appl. Electrochem., vol. 35 (2005) (Special Issue on Heterogeneous Photocatalysis). [21] F. Delamare, B. Guineau, Colors-The Story of Dyes and Pigments, Harry N. Abrams Inc., New York, 2000 (Engl. Translation by S. Hawkes). [22] H. Gerischer, F. Willig, Top. Curr. Chem. 61 (1976) 31. [23] R. Memming, Prog. Surf. Sci. 17 (1984) 7. [24] A.V. Pamfilov, Ya.S. Mazurkevich, E.P. Pakhomova, Kinet. Catal. (USSR) 10 (1969) 915. [25] H. Yoneyama, Y. Toyoguchi, H. Tamura, J. Phys. Chem. 76 (1972) 3460. [26] T. Takizawa, T. Watanabe, K. Honda, J. Phys. Chem. 84 (1980) 51 (see also references therein). [27] G.T. Brown, J.R. Darwent, J. Chem. Soc., Faraday Trans. I 80 (1984) 1631. [28] G.T. Brown, J.R. Darwent, J. Phys. Chem. 88 (1984) 4955. [29] R.W. Matthews, J. Phys. Chem. 91 (1987) 3328. [30] R.W. Matthews, Water Res. 25 (1991) 1169. [31] M.S. Dieckmann, K.A. Gray, R.G. Zepp, Chemosphere 28 (1994) 1021. [32] K. Vinodgopal, D.E. Wynkoop, P.V. Kamat, Environ. Sci. Technol. 30 (1996) 1680. [33] M.S.T. Gonçalves, A.M.F. Oliveira-Campos, E.M.M.S. Pinto, P.M.S. Plasência, M.J.R.P. Queiroz, Chemosphere 39 (1999) 781. [34] C. Bauer, P. Jacques, A. Kalt, Chem. Phys. Lett. 307 (1999) 397. [35] C. Bauer, P. Jacques, A. Kalt, J. Photochem. Photobiol. A: Chem. 140 (2001) 87. [36] M. Stylidi, D.I. Kondarides, X.E. Verykios, Appl. Catal. B: Environ. 40 (2003) 271. [37] C. Galindo, P. Jaques, A. Kalt, Chemosphere 45 (2001) 997. [38] T. Papadam, N.P. Xekoukoulotakis, I. Poulios, D. Mantzavinos, J. Photochem. Photobiol. A: Chem. 186 (2007) 308. [39] C. Galindo, P. Jaques, A. Kalt, J. Photochem. Photobiol. A: Chem. 130 (2000) 35. [40] I.M. Arabatzis, T. Stergiopoulos, D. Andreeva, S. Kitova, S.G. Neophytides, P. Falaras, J. Catal. 220 (2003) 127. [41] C. Guillard, H. Lachheb, A. Houas, M. Ksibi, E. Elaloui, J.M. Herrmann, J. Photochem. Photobiol. A: Chem. 158 (2003) 27. [42] M. Mrowetz, E. Selli, J. Photochem. Photobiol. A: Chem. 162 (2004) 89. [43] I. Poulios, I. Tsachpinis, J. Chem. Technol. Biotechnol. 74 (1999) 349. [44] A. Aguedach, S. Brosillon, J. Morvan, E.K. Lhadi, Appl. Catal. B: Environ. 57 (2005) 55. [45] D.E. Kritikos, X. Xekoukoulotakis, E. Psillakis, D. Mantzavinos, Water Res. 41 (2007) 2236. [46] Y. Xu, C.H. Langford, Langmuir 17 (2001) 897. [47] I. Arslan, I.A. Balcioğlu, Dyes Pigments 43 (1999) 95. [48] E. Stathatos, T. Petrova, P. Lianos, Langmuir 17 (2001) 5025. [49] R.F.P.M. Moreira, T.P. Sauer, L. Casaril, E. Humeres, J. Appl. Electrochem. 35 (2005) 821. [50] T. Sauer, G.C. Neto, H.J. José, R.F.P.M. Moreira, J. Photochem. Photobiol. A: Chem. 149 (2002) 147. [51] M. Karkmaz, E. Puzenat, C. Guillard, J.M. Hermann, Appl. Catal. B: Environ. 51 (2004) 183. [52] P. Qu, J. Zhao, L. Zang, T. Shen, H. Hidaka, Colloid Surf. A: Physicochem. Eng. Aspects 138 (1998) 39. [53] J. Zhao, T. Wu, K. Wu, K. Oikawa, H. Hidaka, N. Serpone, Environ. Sci. Technol. 32 (1998) 2394. [54] T. Wu, G. Liu, J. Zhao, H. Hidaka, N. Serpone, J. Phys. Chem. B 102 (1998) 5845. [55] D. Chatterjee, A. Mahata, J. Photochem. Photobiol. A: Chem. 153 (2002) 199. [56] G. Liu, X. Li, J. Zhao, H. Hidaka, N. Serpone, Environ. Sci. Technol. 34 (2000) 3982. [57] F. Zhang, J. Zhao, T. Shen, H. Hidaka, E. Pelizzetti, N. Serpone, Appl. Catal. B: Environ. 15 (1998) 147. [58] R.W. Matthews, J. Chem. Soc., Faraday Trans. I 85 (1989) 1291. [59] K. Sopajaree, S.A. Qasim, S. Basak, K. Rajeshwar, J. Appl. Electrochem. 29 (1999) 533. [60] A. Mills, J. Wang, J. Photochem. Photobiol. A: Chem. 127 (1999) 123. [61] A. Houas, H. Lachheb, M. Ksibi, E. Elaloui, C. Guillard, J.M. Herrmann, Appl. Catal. B: Environ. 31 (2001) 145. [62] P.V. Messina, P.C. Schulz, J. Colloid Interface Sci. 299 (2006) 305. [63] G. Liu, X. Li, J. Zhao, S. Horikoshi, H. Hidaka, J. Mol. Catal. A: Chem. 153 (2000) 221. [64] I. Bouzaida, C. Ferronato, J.M. Chovelon, M.E. Rammah, J.M. Herrmann, J. Photochem. Photobiol. A: Chem. 168 (2004) 23. [65] J. Zhao, K. Wu, T. Wu, H. Hidaka, N. Serpone, J. Chem. Soc., Faraday Trans. 94 (1998) 673. [66] M. Vautier, C. Guillard, J.M. Herrmann, J. Catal. 201 (2001) 46. [67] M. Sökmen, A. Özkan, J. Photochem. Photobiol. A: Chem. 147 (2002) 77. [68] H.Z. Zhang, R. LeePenn, R.J. Hamers, J. Phys. Chem. B 103 (1999) 4656. [69] C.H. Giles, A.P. D’Silva, I.A. Easton, J. Colloid Interface Sci. 47 (1974) 766. [70] R.H. Perry, D. Green (Eds.), Perry’s Chemical Engineers’ Handbook, McGrawHill, New York, 1984. [71] I. Langmuir, J. Am. Chem. Soc. 38 (1916) 2221. [72] H. Freundlich, Z. Phys. Chem. A 57 (1906) 228. [73] M.F. Sawalha, J.R. Peralta-Videa, J. Romero-González, J.L. Gardea-Torresdey, J. Colloid Interface Sci. 300 (2006) 100. [74] G.D. Parfitt, Prog. Surf. Sci. 11 (1976) 181. [75] J. Cunningham, G. Al-Sayyed, J. Chem. Soc., Faraday Trans. 86 (1990) 3935. [76] J. Cunningham, S. Srijaranai, J. Photochem. Photobiol. A: Chem. 58 (1991) 361. [77] J. Cunningham, G. Al-Sayyed, S. Srijaranai in Ref. [5], Ch. 22, pp. 317–348. [78] C.R. Chenthamarakshan, K. Rajeshwar, Electrochem. Commun. 2 (2000) 527. [79] C.R. Chenthamarakshan, H. Yang, Y. Ming, K. Rajeshwar, J. Electroanal. Chem. 494 (2000) 79. [80] P. Kajitvichyanukul, C.R. Chenthamarakshan, K. Rajeshwar, S.R. Qasim, J. Electroanal. Chem. 519 (2002) 25. [81] Y. Ming, C.R. Chenthamarakshan, K. Rajeshwar, J. Photochem. Photobiol. A: Chem. 147 (2002) 199. [82] P. Kajitvichyanukul, C.R. Chenthamarakshan, S.R. Qasim, K. Rajeshwar, Langmuir 17 (2001) 3792. [83] P. Kajitvichyanukul, C.R. Chenthamarakshan, K. Rajeshwar, S.R. Qasim, Adsorp. Sci. Technol. 21 (2003) 217. [84] T. Torimoto, S. Ito, S. Kuwabata, H. Yoneyama, Environ. Sci. Technol. 30 (1996) 1275. [85] H. Tada, M. Akazawa, Y. Kubo, S. Ito, J. Phys. Chem. B 102 (1998) 6360. [86] Y. Xu, C.H. Langford, J. Phys. Chem. B 101 (1997) 3115. [87] L.B. Reutergárdh, M. Iangphasuk, Chemosphere 35 (1997) 585. [88] K. Hustert, R.G. Zepp, Chemosphere 24 (1992) 335. [89] W.Z. Tang, H. An, Chemosphere 31 (1995) 4157. [90] C. Nasr, K. Vinodgopal, L. Fisher, S. Hotchandani, A.K. Chattopadhyay, P.V. Kamat, J. Phys. Chem. 100 (1996) 8436. [91] L. Ziolkowski, K. Vinodgopal, P.V. Kamat, Langmuir 13 (1997) 3124. [92] T. Wu, G. Liu, J. Zhao, H. Hidaka, N. Serpone, J. Phys. Chem. B 103 (1994) 4862. K. Rajeshwar et al. / Journal of Photochemistry and Photobiology C: Photochemistry Reviews 9 (2008) 171–192 [93] F. Kiriakidou, D.I. Kondravides, X.E. Verykios, Catal. Today 54 (1999) 119. [94] L. Feigelson, L. Muszkat, L. Bir, K.A. Muszkat, Water Sci. Technol. 42 (2000) 275. [95] N.L. Stock, J. Peller, K. Vinodgopal, P.V. Kamat, Environ. Sci. Technol. 34 (2000 1747). [96] H. Lachheb, E. Puzenat, A. Houas, M. Ksibi, E. Elaloui, C. Guillard, J.M. Herrmann, Appl. Catal. B: Environ. 39 (2002) 75. [97] V. Augugliaro, C. Baiocchi, A.B. Prevot, E. García-López, V. Loddo, S. Malato, G. Marci, L. Palmisano, M. Pazzi, E. Pramauro, Chemosphere 49 (2002) 1223. [98] G.A. Epling, C. Lin, Chemosphere 46 (2002) 561. [99] M. Stylidi, D.I. Kondarides, X.E. Verykios, Appl. Catal. B: Environ. 47 (2004) 189. [100] A. Patsoura, D.I. Kondarides, X.E. Verykios, Appl. Catal. B: Environ. 64 (2006) 171. [101] S.F. Villanueva, S.S. Martinez, Solar Energy Mater. Solar Cells 91 (2007) 1492. [102] A.J. Julson, D.F. Ollis, Appl. Catal. B: Environ. 65 (2006) 315. [103] Y. Zhiyang, M. Bensimon, V. Sarria, I. Stolitchnov, W. Jardim, D. Laub, E. Mielczarski, J. Mielczarski, L. Kiwi-Minsker, J. Kiwi, Appl. Catal. B: Environ. 76 (2007) 185. [104] R.J. Davis, J.L. Gainer, G. O’Neal, I.W. Wu, Water Environ. Res. 66 (1994) 50. [105] A. Mills, A. Belghazi, R.H. Davis, D. Worsley, S. Morris, J. Photochem. Photobiol. A: Chem. 79 (1994) 131. [106] F. Zhang, J. Zhao, L. Zang, T. Shen, H. Hidaka, E. Pelizzetti, N. Serpone, J. Mol. Catal. A: Chem. 120 (1997) 173. [107] G. Liu, J. Zhao, New J. Chem. 24 (2000) 411. [108] C. Chen, X. Li, W. Ma, J. Zhao, H. Hidaka, N. Serpone, J. Phys. Chem. B 106 (2002) 318. [109] R.F.P. Nogeira, W.F. Jardim, J. Chem. Educ. 70 (1993) 861. [110] S. Lekshmi, R. Renganathan, S. Fujita, J. Photochem. Photobiol. A: Chem. 88 (1995) 163. [111] K. Sopajaree, S.A. Qasim, S. Basak, K. Rajeshwar, J. Appl. Electrochem. 29 (1999) 1111. [112] T. Zhang, T. Oyama, A. Aoshima, H. Hidaka, J. Zhao, N. Serpone, J. Photochem. Photobiol. A: Chem. 140 (2001) 163. [113] J. Kiwi, C. Pulgarin, M. Grätzel, New J. Chem. 17 (1993) 487. [114] G. Liu, T. Wu, J. Zhao, H. Hidaka, N. Serpone, Environ. Sci. Technol. 33 (1999 2081). [115] A.B. Prevot, C. Baiocchi, M.C. Brussino, E. Pramauro, P. Savarino, V. Augugliaro, G. Marci, L. Palmisano, Environ. Sci. Technol. 35 (2001) 971. [116] K.D. Giglio, D.B. Green, B. Hutchinson, J. Chem. Educ. 72 (1995) 352. [117] T. Wu, T. Lin, J. Zhao, H. Hidaka, N. Serpone, Environ. Sci. Technol. 33 (1999) 1379. [118] P. Reeves, R. Ohlhausen, D. Sloan, K. Pamplin, T. Scoggins, C. Clark, B. Hutchinson, D. Green, Solar Energy 48 (1992) 413. [119] K. Tennakone, S. Senadeera, A. Priyadharshana, Solar Energy Mater. Solar Cells 29 (1993) 109. [120] K. Tennakone, W.C.B. Kiridena, S. Punchihewa, J. Photochem. Photobiol. A: Chem. 68 (1992) 389. [121] L.C. Chen, T.C. Chou, J. Mol. Catal. 85 (1993) 201. [122] L.C. Chen, T.C. Chou, Ind. Eng. Chem. Res. 33 (1994) 1436. [123] L.C. Chen, T.C. Chou, Ind. Eng. Chem. Res. 32 (1993) 1520. [124] T. Tatsuma, S.I. Tachibana, T. Miwa, D.A. Tryk, A. Fujishima, J. Phys. Chem. B 103 (1999) 8033. [125] K. Tanaka, K. Padermpole, T. Hisanaga, Water Res. 34 (2000) 327. [126] R.I. Bickley, T.G. Carreno, J.S. Lees, L. Palmisano, R.J.D. Tilley, J. Solid State Chem. 92 (1991) 178. [127] M. Lindner, D.W. Bahnemann, B. Hirthe, W.D. Griebler, J. Solar Energy Eng. 119 (1997) 120. [128] P.V. Kamat, M.A. Fox, Chem. Phys. Lett. 102 (1983) 379. [129] K.R. Gopidas, P.V. Kamat, Langmuir 5 (1989) 22. [130] K. Vinodgopal, Res. Chem. Intermed. 20 (1994) 825. [131] K. Vinodgopal, I. Bedja, S. Hotchandani, P.V. Kamat, Langmuir 10 (1994 1767). [132] C. Nasr, K. Vinodgopal, S. Hotchandani, A.K. Chattopadhyay, P.V. Kamat, Res. Chem. Intermed. 23 (1997) 219. [133] F. Sayilkan, S. Erdemoğlu, M. Asiltůrk, M. Akarsu, Ş. Şener, H. Sayilkan, M. Erdemoğlu, E. Arpac, Mater. Res. Bull. 41 (2006) 2276. [134] M. Nag, P. Basak, S.V. Manorama, Mater. Res. Bull. 42 (2007) 1691. [135] M.L. Curri, R. Comparelli, P.D. Cozzoli, G. Mascolo, A. Agostiano, Mater. Sci. Eng. C 23 (2003) 285. [136] J.E.B. Katari, V.L. Colvin, A.P. Alivisatos, J. Phys. Chem. 98 (1994) 4109. [137] G.N. Kryukova, G.A. Zenkovets, A.A. Shutilov, M. Wilde, K. Günther, D. Fassler, K. Richter, Appl. Catal. B: Environ. 71 (2007) 169. [138] L. Wu, J.C. Yu, X. Wang, L. Zhang, J. Yu, J. Solid State Chem. 178 (2005) 321. [139] G.A. Somorjai, Introduction to Surface Chemistry and Catalysis, J. Wiley & Sons, New York, 1994. [140] A.W. Adamson, A.P. Gast, Physical Chemistry of Surfaces, J. Wiley & Sons, New York, 1997. [141] K.R. Gopidas, P.V. Kamat, J. Phys. Chem. 93 (1989) 6428. [142] M.E. Fabiyi, R.L. Skelton, J. Photochem. Photobiol. A: Chem. 132 (2000) 121. [143] Y. Chen, K. Wang, L. Lou, J. Photochem. Photobiol. A: Chem. 163 (2004) 281. [144] J. Feng, R.S.K. Wong, X. Hu, P.L. Yue, Catal. Today 98 (2004) 441. [145] D. Li, Y. Guo, C. Hu, C. Jiang, E. Wang, J. Mol. Catal. A: Chem. 207 (2004) 181. [146] A.O. Ibhadon, G.M. Greenway, Y. Yue, Catal. Commun. 9 (2008) 153. 191 [147] S. Otsuka-Yao-Matsuo, M. Ueda, J. Photochem. Photobiol. A: Chem. 168 (2004) 1. [148] N.R. de Tacconi, R.O. Lezna, K. Rajeshwar, Chem. Mater. 15 (2003) 3046. [149] E.S. Smotkin, A.J. Bard, A. Campion, M.A. Fox, T. Mallouk, S.E. Webber, J.M. White, J. Phys. Chem. 90 (1986) 4604. [150] Y. Bessekhouad, N. Chaoui, M. Trzpit, N. Ghazzal, D. Robert, J.V. Weber, J. Photochem. Photobiol. A: Chem. 183 (2006) 218. [151] C.H. Wu, Chemosphere 57 (2004) 601. [152] C.H. Wu, C.L. Chang, J. Hazard. Mater. B 128 (2006) 265. [153] Y. Zhiyong, E. Mielczarski, J. Mielczarski, D. Laub, Ph. Buffat, U. Klehm, P. Albers, K. Lee, A. Kulik, L. Kiwi-Minsker, A. Renken, J. Kiwi, Water Res. 41 (2007) 862 (see also references therein). [154] H. Tada, M. Tanaka, Langmuir 13 (1997) 360. [155] Y. Ma, J. Yao, Chemosphere 38 (1999) 2407. [156] M. Miyauchi, A. Nkajima, A. Fujishima, K. Hashimoto, T. Watanabe, Chem. Mater. 12 (2000) 3. [157] S. Sitkiewitz, A. Heller, New J. Chem. 20 (1996) 233. [158] M.V.B. Zanoni, J.J. Sene, M.A. Anderson, J. Photochem. Photobiol. A: Chem. 157 (2003) 55. [159] P.A. Carneiro, M.E. Osugi, J.J. Sene, M.A. Anderson, M.V.B. Zanoni, Electrochim. Acta 49 (2004) 3807. [160] M.E. Osugi, G.A. Umbuzeiro, M.A. Anderson, M.V.B. Zanoni, Electrochim. Acta 50 (2005) 5261. [161] P.A. Carneiro, M.E. Osugi, C.S. Fugivara, N. Boralle, M. Furlan, M.V.B. Zanoni, Chemosphere 59 (2005) 431. [162] M.E. Osugi, G.A. Umbuzeiro, F.J.V. De Castro, M.V.B. Zanoni, J. Hazard. Mater. B 137 (2006) 871. [163] M.A. Anderson, M.J. Gieselmann, Q. Xu, J. Membr. Sci. 39 (1988) 243. [164] D.H. Kim, M.A. Anderson, Environ. Sci. Technol. 28 (1994) 479. [165] T. Watanabe, A. Nakajima, R. Wang, M. Minabe, S. Koizumi, A. Fujishima, K. Hashimoto, Thin Solid Films 351 (1999) 260. [166] P. Bouras, P. Lianos in Ref. [20], p. 831. [167] C.H. Ao, M.K.H. Leung, R.C.W. Lam, D.Y.C. Leung, L.L.P. Vrijmoed, W.C. Yam, S.P. Ng, Chem. Eng. J. 129 (2007) 153. [168] L.C. Macedo, D.A.M. Zaia, G.J. Moore, H. de Santana, J. Photochem. Photobiol. A: Chem. 185 (2007) 86. [169] P. Chin, D.F. Ollis, Catal. Today 123 (2007) 177. [170] A. Mills, A. Lepre, N. Elliott, S. Bhopal, I.P. Parkin, S.A. O’Neill, J. Photochem. Photobiol. A: Chem. 160 (2003) 213. [171] K. Rajeshwar, in: V. Balzani (Ed.), Electron Transfer in Chemistry, Wiley–VCH, Weinheim, 2001. [172] K. Rajeshwar, in: S. Licht (Ed.), Encyclopedia of Electrochemistry, Wiley–VCH, Weinheim, 2001, pp. 3–53, Ch. 1. [173] K. Vinodgopal, I. Bedja, P.V. Kamat, Chem. Mater. 8 (1996) 2180. [174] K. Vinodgopal, P.V. Kamat, Environ. Sci. Technol. 29 (1995) 841. [175] R. Suárez-Parra, I. Hernández-Pérez, M.E. Rincón, S. López-Ayala, M.C. RoldánAhumada, Solar Energy Mater. Solar Cells 76 (2003) 189. [176] K. Rajeshwar, N.R. de Tacconi, in: A. Wieckowski (Ed.), Interfacial Electrochemistry, Marcel Dekker, New York, 1999, pp. 721–736, Ch. 40. [177] K. Rajeshwar, N.R. de Tacconi, C.R. Chenthamarakshan, Curr. Opin. Solid State Mater. Sci. 8 (2005) 173. [178] N.R. de Tacconi, C.R. Chenthmarakshan, G. Yogeeswaran, A. Watcherenwong, R.S. de Zoysa, N.A. Basit, K. Rajeshwar, J. Phys. Chem. B 110 (2006) 25347. [179] C.A. Grimes, O.K. Varghese, S. Ranjan, Light, Water, Hydrogen, Springer, New York, 2008 (ch. 5), pp. 257–369. [180] G.H.A. Therese, P.V. Kamath, Chem. Mater. 12 (2000) 1195. [181] C. Natarajan, G. Nogami, J. Electrochem. Soc. 143 (1996) 1547. [182] X.Z. Li, H.L. Liu, P.T. Yue, Y.P. Sun, Environ. Sci. Technol. 34 (2000) 4401. [183] X.B. Xie, X.Z. Li, Mater. Chem. Phys. 95 (2006) 39. [184] M.E. Osugi, C.R. Chenthamarakshan, N.R. de Tacconi, G.A. Woldermariam, S.S. Mandal, M.V.B. Zanoni, K. Rajeshwar, Proceedings of the International Society of Electrochemistry (ISE) Meeting Foz da Iguazu, Brazil, 2008. [185] M.E. Osugi, C.R. Chenthamarakshan, N.R. de Tacconi, G.A. Woldermariam, S.S. Mandal, M.V.B. Zanoni, K. Rajeshwar, J. Adv. Oxid. Technol 11 (2008) 425. [186] M.F.J. Dijktra, A. Michorius, H. Buwalda, H.J. Panneman, J.G.M. Winkelman, A.A.C.M. Beenacker, Catal. Today 66 (2001) 487. [187] S. Parra, S.E. Stanca, I. Guasaquillo, K.R. Thampi, Appl. Catal. B: Environ. 51 (2004) 73. [188] H. Gerischer, A. Heller, J. Phys. Chem. 95 (1991) 5261. [189] H. Gerischer, A. Heller, J. Electrochem. Soc. 139 (1992) 113. [190] E. Vlatka, Ž. Kulešius, J. Appl. Electrochem. 37 (2007) 415. [191] J.M. Peralta-Hernández, J. Manríquez, Y. Meas-Vong, F.J. Rodriguez, T.W. Chapman, M.I. Maldonado, L.A. Godínez, J. Hazard. Mater. 147 (2007) 588. [192] K. Ui, T. Fujita, N. Koura, F. Yamaguchi, J. Electrochem. Soc. 153 (2006) C449. [193] G. Li, J. Qu, X. Zhang, J. Ge, Water Res. 40 (2006) 213. [194] G. Li, J. Qu, X. Zhang, H. Liu, H. Liu, J. Mol. Catal. A: Chem. 259 (2006) 238. [195] H. Zhang, X. Quan, S. Chen, H. Zhao, Y. Zhao, Sep. Purif. Technol. 50 (2006) 147. [196] T. Wang, H. Wang, P. Xu, X. Zhao, Y. Liu, S. Chao, Thin Solid Films 334 (1998) 103. [197] M. Hepel, J. Luo, Electrochim. Acta 47 (2001) 729. [198] J. Luo, M. Hepel, Electrochim. Acta 46 (2001) 2913. [199] A. Watcherenwong, W. Chanmanee, N.R. de Tacconi, C.R. Chenthamarakshan, P. Kajitvichyanukul, K. Rajeshwar, J. Electroanal. Chem. 612 (2008) 112. [200] W.Y. Lin, C. Wei, K. Rajeshwar, J. Electrochem. Soc. 140 (1993) 2477. [201] W.Y. Lin, K. Rajeshwar, J. Electrochem. Soc. 144 (1997 2751). 192 K. Rajeshwar et al. / Journal of Photochemistry and Photobiology C: Photochemistry Reviews 9 (2008) 171–192 [202] C.R. Chenthamarakshan, K. Rajeshwar, E.J. Wolfrum, Langmuir 16 (2000 2715). [203] N.R. de Tacconi, J. Carmona, K. Rajeshwar, J. Electrochem. Soc 144 (1997) 2486. [204] J.M. Herrmann, Catal. Today 53 (1999) 115. [205] N.B. Jackson, C.M. Wang, Z. Luo, J. Schwitzgebel, J.G. Eckerdt, J.R. Brock, A. Heller, J. Electrochem. Soc. 138 (1991) 3660. [206] M.A. Grela, M.E.J. Coronel, A.J. Colussi, J. Phys. Chem. 100 (1996) 16940. [207] L.Z. Sun, J.R. Bolton, J. Phys. Chem. 100 (1996) 4127. [208] Y. Nosaka, S. Komori, K. Yawat, T. Hirakawa, A.Y. Nosaka, Phys. Chem. Chem. Phys. 5 (2003) 4731. [209] K.I. Ishibashi, A. Fujishima, T. Watanabe, K. Hashimoto, J. Photochem. Photobiol. A: Chem. 134 (2000) 139. [210] Y. Chen, S. Yang, K. Wang, L. Lou, J. Photochem. Photobiol. A: Chem. 172 (2005) 47. [211] V. Rao, K. Rajeshwar, V. Pai Verneker, J. Dubow, J. Phys. Chem. 84 (1980 1987). [212] J.R. Bolton, K.G. Bircher, W. Tuma, C.A. Tolman, J. Adv. Oxid. Technol. 1 (1996) 13. [213] M. Abdullah, G.K.C. Low, R.W. Matthews, J. Phys. Chem. 94 (1990) 6820. [214] C. Walling, Acc. Chem. Res. 8 (1975) 125. [215] K. Rajeshwar, J. Appl. Electrochem. 37 (2007) 765. [216] S.G. Schrank, H.J. José, R.F.P.M. Moreira, J. Photochem. Photobiol. A: Chem. 147 (2002) 71. [217] W.W. Eckenfelder, A.R. Bowers, J.A. Roth (Eds.), Chemical Oxidation: Technologies for the Nineties, vol. 2, Technomic Publishing, Lancaster and Basel, 1992; W.W. Eckenfelder, A.R. Bowers, J.A. Roth (Eds.), Chemical Oxidation: Technologies for the Nineties, vol.3, Technomic Publishing, Lancaster and Basel, 1993. [218] J. R. Bolton and S. R. Cater in Ref. [5], Ch. 33, pp. 467–490. [219] K.S. Suslick, Ulrasound: Its Chemical, Physical and Biological Effects, VCH, New York, 1988. [220] K.S. Suslick, Science 247 (1990) 1439. [221] M. Stevenson, K. Bullock, W.Y. Lin, K. Rajeshwar, Res. Chem. Intermed. 23 (1997) 311. [222] K. Vinodgopal, J. Peller, O. Makogon, P.V. Kamat, Water Res. 32 (1998) 3646. [223] T. Velegraki, I. Poulios, M. Charalabaki, N. Kalogerakis, P. Samaras, D. Mantzavinos, Appl. Catal. B: Environ. 62 (2006) 159. [224] D. Chatterjee, S. Dasgupta, J. Photochem. Photobiol. C: Photochem. Rev. 6 (2005) 186. [225] H. Ross, J. Bendig, S. Hecht, Solar Energy Mater. Solar Cells 33 (1994) 475. [226] D.F. Ollis, J. Phys. Chem. B 109 (2005) 2439 (see also references therein). [227] S.N. Demming, S.L. Morgan, Experimental Design: A Chemometric Approach, Elsevier, Amsterdam, 1987. [228] G. Box, W.G. Hunter, Statistics for Experimenters: An Introduction to Design, Data Analysis, and Model Building, Wiley, New York, 1987. [229] N. Watanabe, S. Horikoshi, H. Hidaka, N. Serpone, J. Photochem. Photobiol. A: Chem. 174 (2005) 229. [230] L.M. Games, R.A. Hites, Anal. Chem. 49 (1977) 1433. [231] R.O. Alves de Lima, A.P. Bazo, D.M.F. Salvadori, C.M. Rech, D. de Palma Oliveira, G. de Aragão Umbuzeiro, Mutat. Res. 626 (2007) 53. [232] E. Chatzisymeon, N.P. Xekoukoulotakis, A. Coz, N. Kalogerakis, D. Mantzavinos, J. Hazard. Mater. B 137 (2006) 998. [233] S. Gomes de Moraes, R.S. Freire, N. Duran, Chemosphere 40 (2000) 369. [234] C. Hu, J.C. Yu, Z. Hao, P.K. Wong, Appl. Catal. B: Environ. 42 (2003) 47. [235] M.V.B. Zanoni, P.A. Carneiro, G.A. Umbuzeiro, M.A. Anderson, in press. [236] M. Kositzi, I. Poulios, K. Samara, E. Tsatsaroni, E. Darakas, J. Hazard. Mater. 146 (2007) 680. [237] G. Ruppert, R. Bauer, G. Heisler, Chemosphere 28 (1994) 1447. [238] A.N. Gopalakrishnan, V. Mohan, Ind. J. Environ. Protect. 17 (1997) 268. [239] I.A. Balcioglu, I. Arslan, J. Adv. Oxid. Technol. 4 (1999) 189. [240] C. Lizama, J. Freer, J. Baeza, H.D. Mansilla, Catal. Today 76 (2002) 235. [241] W.A. Sadik, J. Photochem. Photobiol. A: Chem., in press. [242] M. Muruganandham, K. Selvam, M. Swaminathan, J. Hazard. Mater. 144 (2007) 316.