(2004) KONSTANTINOU e ALBANIS TiO2-Assisted Photocatalytic Degradation of Azo Dyes in Aqueous Solution

Applied Catalysis B: Environmental 49 (2004) 1–14

TiO2-assisted photocatalytic degradation of azo dyes in aqueous solution: kinetic and mechanistic investigatio investigations ns A review Ioannis K. Konstantinou∗, Triantafyllos A. Albanis  Department of Chemistry, Laboratory of Environmental Environmental Technology, Technology, University of Ioannina, I oannina 4511 0, Greece

Received 5 July 2003; received in revised form 24 November 2003; accepted 24 November 2003

Abstract

The photocatalytic degradation of azo dyes containing different functionalities has been reviewed using TiO2  as photocatalyst in aqueous solution under solar and UV irradiation. The mechanism of the photodegradation depends on the radiation used. Charge injection mechanism takes place under visible radiation whereas charge separation occurred under UV light radiation. The process is monitored by following either the decolorization rate and the formation of its end-products. Kinetic analyses indicate that the photodegradation rates of azo dyes can usually be approximated as pseudo-first-order kinetics for both degradation mechanisms, according to the Langmuir–Hinshelwood model. The degradation of dyes depend on several parameters such as pH, catalyst concentration, substrate concentration and the presence of electron acceptors such as hydrogen peroxide and ammonium persulphate besides molecular oxygen. The presence of other substances such as inorganic ions, humic acids and solvents commonly found in textile effluents is also discussed. The photocatalyzed degradation of pesticides does not occur instantaneously to form carbon dioxide, but through the formation of long-lived intermediate species. Thus, the study focuses also on the determination of the nature of the principal organic intermediates and the evolution of the mineralization as well as on the degradation pathways followed during the process. Major identified intermediates are hydroxylated derivatives, aromatic amines, naphthoquinone, phenolic compounds and several organic acids. By-products evaluation and toxicity measurements are the key-actions in order to assess the overall process. © 2003 Elsevier B.V. All rights reserved. Keywords:  Azo dyes; Photocatalytic degradation processes; Operational parameters; Transformation products

1. Introducti Introduction on

Textile dyes and other industrial dyestuffs constitute one of the largest group of organic compounds that represent an increasing environmental danger. About 1–20% of the total world production of dyes is lost during the dyeing process and is released in the textile effluents  [1–4]  [1–4].. The release of  those colored waste waters in the environment is a considerable source of non aesthetic pollution and eutrophication and can originate dangerous byproducts through oxidation, hydrolysis, or other chemical reactions taking place in the wastewater wastewater phase [5–8] phase  [5–8]..

∗

fax:

Corresponding author. Tel.:

+30-26510-98363;

+30-26510-98795.

 E-mail address:   [email protected] (I.K. Konstantinou).

0926-3373/$ – see front matter © 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.apcatb.2003.11.010

Decolorization of dye effluents has therefore received increasing attention. For the removal of dye pollutants, traditional physical techniques (adsorption on activated carbon, ultrafiltrat ultrafiltration, ion, reverse reverse osmosis, osmosis, coagulatio coagulationn by chemical chemical agents, ion exchange on synthetic adsorbent resins, etc.) can generally be used efficiently [9–12] efficiently  [9–12].. Nevertheless, they are non-destructive, since they just transfer organic compounds from water to another phase, thus causing secondary pollution. Consequently, regeneration of the adsorbent materials and post-treat post-treatment ment of solid-was solid-wastes, tes, which are expensiv expensivee operations operations,, are needed needed   [12,13]. [12,13]. Due to the large degree of aromatics present in dye molecules and the stability of  modern modern dyes, dyes, convent conventional ional biological biological treatment treatment methods methods are ineffective for decolorization and degradation  [14–17]  [14–17].. Furthermore, the majority of dyes is only adsorbed on the sludge and is not degraded  [18]  [18]..  Chlorination and ozonation are also being used for the removal of certain dyes but

2

I.K. Konstantinou, T.A. Albanis / Applied Catalysis B: Environmental 49 (2004) 1–14

at slower rates as they have often high operating costs and limited effect on carbon content  [12,19–22]. These are the reasons why advanced oxidation processes (AOPs) have been growing during the last decade since they are able to deal with the problem of dye destruction in aqueous systems. AOPs were based on the generation of very reactive species such as hydroxy radicals ( • OH) that oxidize a broad range of pollutants quickly and non selectively. AOPs such as Fenton and photo-Fenton catalytic reactions [23–27], H2 O2 /UV processes  [28,29]   and TiO2   mediated photo-catalysis [11,30–33] have been studied under a broad range of experimental conditions in order to reduce the color and organic load of dye containing effluent waste waters. Among AOPs, heterogeneous photocatalysis using TiO 2 as photo-catalyst appears as the most emerging destructive technology [4,34–39]. The key advantage of the former is its inherent destructive nature: it does not involve mass transfer; it can be carried out under ambient conditions (atmospheric oxygen is used as oxidant) and may lead to complete mineralization of organic carbon into CO2 . Moreover, TiO2 photocatalyst is largely available, inexpensive, non-toxic and show relatively high chemical stability. Finally, TiO 2  photocatalytic process is receiving increasing attention because of its low cost when using sunlight as the source of irradiation. The utilization of combined photocatalysis and solar technologies may be developed to a useful process for the reduction of water pollution by dying compounds because of the mild conditions required and their efficiency in the mineralization [7,40–44]. The application of photocatalytic procedures for remediation of textile wastewater is rather limited to few investigations [45–49]. There are many studies dealing with the photocatalytic decolorization of specific textile dyes from different chemical categories, and most of them including a detailed examination of the so-called primary process under different working conditions [7,11,32,41,42,50–57]. On the contrary, little information is available on the reaction mechanisms involved in the photocatalytic degradation of  dyes and on the identification of major transient intermediates which have been more recently recognized as very important aspects of these processes, especially in view of  their practical applications  [6,34,35,44,58]. Thus, information about real mineralization of the dye or decreases in toxicity are scarce and therefore our attention has been also focused on the reaction types and mechanisms, based on the identification of the transformation products. Moreover, the effect of common dyebath constituents on the photocatalytic treatment efficiency is also discussed in order to examine the application of the photocatalytic degradation on real wastewater effluents. Of the dyes available on the market today, approximately 50–70% are azo compounds followed by the anthraquinone group. Azo dyes can be divided into monoazo, diazo, triazo classes according to the presence of one or more azo bonds (–N=N–) and are found in various categories, i.e. acid, basic, direct, disperse, azoic and pigments  [3,59]. Some azo

dyes and their dye precursors have been shown to be or are suspected to be human carcinogens as they form toxic aromatic amines [44,60,61]. Therefore azo dyes are pollutants of high environmental impact and were selected as the most relevant group of dyes concerning their degradation using TiO2  assisted photocatalysis. To our knowledge there is not a review dealing with the photocatalytic degradation of dyes although that there are some reviews concerning the photocatalytic degradation of  other pollutants such as pesticides  [37,62–64]. This review intend to assist workers involved in azo dyes photocatalytic treatment using TiO2  by: (a) compiling data on the degree and on the factors influencing dye photodegradation, and (b) Summarizing and discussing data on the mineralization degree, the intermediates and reaction mechanisms followed during the process. The azo dyes were classified in terms of  the characteristic structural groups. 2. Experimental procedures 2.1. Photocatalytic degradation mechanisms 2.1.1. Photocatalytic oxidation

The detailed mechanism of the process has been discussed previously in the literature  [4,6,58,65–68] and will be only briefly summarized here. It is well established that conduction band electrons (e −) and valence band holes (h +) are generated when aqueous TiO 2  suspension is irradiated with light energy greater than its band gap energy ( E g , 3.2 eV). The photogenerated electrons could reduce the dye or react with electron acceptors such as O 2  adsorbed on the Ti(III)-surface or dissolved in water, reducing it to superoxide radical anion O 2 •− . The photogenerated holes can oxidize the organic molecule to form R + , or react with OH− or H2 O oxidizing them into OH • radicals. Together with other highly oxidant species (peroxide radicals) they are reported to be responsible for the heterogeneous TiO 2  photodecomposition of organic substrates as dyes. According to this, the relevant reactions at the semiconductor surface causing the degradation of dyes can be expressed as follows: TiO2 + hv(UV) → TiO2 (eCB − + hVB + )

(1)

TiO2 (hVB + ) + H2 O → TiO2 + H+ + OH•

(2)

TiO2 (hVB + ) + OH− → TiO2 + OH•

(3)

TiO2 (eCB −) + O2

(4)

→

TiO2 + O2 • −

O2 • − + H+ → HO2 •

(5)

Dye + OH• → degradation products

(6)

Dye + hVB +

oxidation products

(7)

Dye + eCB − → reduction products

(8)

→

The resulting • OH radical, being a very strong oxidizing agent (standard redox potential +2.8 V) can oxidize most of 

3

 I.K. Konstantinou, T.A. Albanis / Applied Catalysis B: Environmental 49 (2004) 1–14

azo dyes to the mineral end-products. Substrates not reactive toward hydroxyl radicals are degraded employing TiO 2 photocatalysis with rates of decay highly influenced by the semiconductor valence band edge position  [69]. The role of  reductive pathways (Eq. (8)) in heterogeneous photocatalysis has been envisaged also in the degradation of several dyes but in a minor extent than oxidation [58,70]. 2.1.2. Photosensitized oxidation

The mechanism of photosensitized oxidation (called also photo-assisted degradation) by visible radiation ( λ > 420nm) is different from the pathway implicated under UV light radiation. In the former case the mechanism suggests that excitation of the adsorbed dye takes place by visible light to appropriate singlet or triplet states, subsequently followed by electron injection from the excited dye molecule onto the conduction band of the TiO2  particles, whereas the dye is converted to the cationic dye radicals (Dye•+ ) that undergoes degradation to yield products as follows [32,34,50,51,67,68,70–72]: 1

∗

3

∗

Dye + hv(VIS) → Dye or Dye 1

Dye∗ or 3 Dye∗ + TiO2

TiO2 (eCB − ) + O2

→

→

(9)

Dye•+ + TiO2 (eCB − )   (10)

O2 • − + TiO2

 

(11)

Dye• + → degradation products

(12)

The cationic dye radicals readily reacts with hydroxyl ions undergoing oxidation via reactions 13 and 14 or interacts effectively with O2 •− , HO2 • or HO•− species to generate intermediates that ultimately lead to CO 2  (Eqs. (15)–(19)). Dye• + + OH−

→

Dye + HO•

(13)

Dye + 2HO• → H2 O + oxidation p roducts

(14)

O2 • − + H+ → HO2 •

(15)

HO2 • + H++TiO2 (eCB −) → H2 O2 + TiO2 H2 O2 + TiO2 (eCB − ) → HO• + HO− + TiO2 Dye• ++O2 • − → DO2

 

(16)  

(17)

degradation products

(18)

Dye•++HO2 • (orHO• ) → degradationproducts

(19)

→

In experiments that are carried out using sunlight or simulated sunlight (laboratory experiments) it is suggested that both photooxidation or photosensitizing mechanism occurred during the irradiation and both TiO2   and the light source are necessary for the reaction to occur. In the photocatalytic oxidation, TiO2  has to be irradiated and excited in a near-UV energy to induce charge separation. On the other hand, dyes rather TiO2   are excited by visible light followed by electron injection onto TiO 2  conduction band, which leads to photosensitized oxidation. It is difficult to conclude whether the photocatalytic oxidation is superior to the photosensitizing oxidation mechanism, but the photosensitizing mechanism will help to improve the overall

efficiency and make the photobleaching of dyes using solar light more feasible [51]. 2.2. Primary substrate disappearance

Several experimental results indicated that the destruction rates of photocatalytic oxidation of various dyes over illuminated TiO2   fitted the Langmuir–Hinshelwood (L–H) kinetics model [4,9,65,73–75]: r

=

dC dt 

=

kKC 

(20)

1 + KC 

where r   is the oxidation rate of the reactant (mg/l min),  C  the concentration of the reactant (mg/l), t   the illumination time, k   the reaction rate constant (mg/l min), and K   is the adsorption coefficient of the reactant (l/mg). When the chemical concentration  C o  is a millimolar solution (C o  small) the equation can be simplified to an apparent first-order equation [4,9,37]: Ln

  Co C

=

kKt  = kapp. t  or

Ct

=

k Co e− app. t 

 

(21)

A plot of ln C o / C  versus time represents a straight line, the slope of which upon linear regression equals the apparent first-order rate constant  k app . Generally first-order kinetics are appropriate for the entire concentration range up to few ppm and several studies were reasonably well fitted by this kinetic model [8,38,43,44,51,54,55,58,60,76]. The L–H model was established to describe the dependence of the observed reaction rate on the initial solute concentrations. It has been agreed, with minor doubts that the expression for the rate of photomineralization of organic substrates such dyes with irradiated TiO2   follows the Langmuir–Hinshelwood (L–H) law for the four possible situations; (a) the reaction takes place between two adsorbed substances, (b) the reaction occurs between a radical in solution and an adsorbed substrate molecule, (c) the reaction takes place between a radical linked to the surface and a substrate molecule in solution, and (d) the reaction occurs with the both of species being in solution. In all cases, the expression for the rate equation is similar to that derived from the L–H model, which has been useful in modeling the process, although it is not possible to find out whether the process takes place on the surface in the solution or at the interface [6]. It is likely that sorption of the dye is an important parameter in determining photocatalytic degradation rates. All isotherms showed L-shape curves according to the classification of Giles et al. [77] that means there is no strong competition between the water and the dye molecules to occupy the TiO2  surface sites. The adsorption isotherms fit well to Langmuirian type implying a monolayer adsorption model [4,73,75,76]. The color removal of the dye solution was determined usually with the absorbance value at the maximum of the

4

I.K. Konstantinou, T.A. Albanis / Applied Catalysis B: Environmental 49 (2004) 1–14

absorption spectrum for every dye by monitoring UV-Vis spectrum in 200–800 nm zone using a spectrophotometer [41,51,72,78]. Alternatively, the disappearance of dye was monitored by high performance liquid chromatography equipped with a UV diode array detector  [42,58,79,80] or MS detector [3]. 2.3. Factors influencing the photocatalytic degradation 2.3.1. Effect of initial dye concentration

It is important both from a mechanistic and from an application point of view to study the dependence of the photocatalytic reaction rate on the substrate concentration. It is generally noted that the degradation rate increases with the increase in dye concentration to a certain level and a further increase in dye concentration leads to decrease the degradation rate of the dye  [8,42]. The rate of degradation relates to the probability of  • OH radicals formation on the catalyst surface and to the probability of  • OH radicals reacting with dye molecules. As the initial concentrations of the dye increase the probability of reaction between dye molecules and oxidizing species also increases, leading to an enhancement in the decolorization rate. On the contrary, the degradation efficiency of the dye decreases as the dye concentration increases further. The presumed reason is that at high dye concentrations the generation of  • OH radicals on the surface of catalyst is reduced since the active sites are covered by dye ions. Another possible cause for such results is the UV-screening effect of the dye itself. At a high dye concentration, a significant amount of UV may be absorbed by the dye molecules rather than the TiO2 particles and that reduces the efficiency of the catalytic reaction because the concentrations of  • OH and O2 •− decrease [9,52,53,68,74,81,82]. The major portion of degradation occurs in the region near to the irradiated side (termed as reaction zone) where the irradiation intensity is much higher than in the other side [83]. Thus at higher dye concentration, degradation decreases at sufficiently long distances from the light source or the reaction zone due to the retardation in the penetration of light. Hence, it is concluded that as initial concentration of the dye increases, the requirement of catalyst surface needed for the degradation also increases [7]. 2.3.2. Effect of TiO2  loading

Whether in static, slurry or dynamic flow reactors the initial reaction rates were found to be directly proportional to catalyst concentration indicating the heterogeneous regime. However, it was observed that above a certain level of  concentration the reaction rate even decreases and becomes independent of the catalyst concentration. Most of studies reported enhanced degradation rates for catalyst loading up to 400–500mg/l  [8,42,52,53,55,84,85].   Only a slight enhancement or decrease was observed when TiO 2  concentration further increased up to 2000 mg/l. This can be rationalized in terms of availability of active sites on TiO2 surface and the light penetration of photoactivating light into the

suspension. The availability of active sites increases with the suspension of catalyst loading, but the light penetration, and hence, the photoactivated volume of the suspension shrinks. Moreover, the decrease in the percentage of degradation at higher catalyst loading may be due to deactivation of activated molecules by collision with ground state molecules [7].   Agglomeration and sedimentation of the TiO 2   particles were observed elsewhere when 2000 mg/l of TiO 2 was added to the dye solution  [52]. In such a condition, part of  the catalyst surface probably became unavailable for photon absorption and dye adsorption, thus bringing little stimulation to the catalytic reaction. On the contrary, continuous increase of the photocatalytic degradation rate of Reactive Black 5 was found up to 3500 mg/l TiO2  [74]. The crucial concentration depends on the geometry, the working conditions of the photoreactor and the type of UV-lamp (power, wavelength). The optimum amount of TiO 2  has to be added in order to avoid unnecessary excess catalyst and also to ensure total absorption of light photons for efficient photomineralization. This optimum loading of photocatalyst is found to be dependent on the initial solute concentration  [62]. 2.3.3. Effect of pH 

The interpretation of pH effects on the efficiency of dye photodegradation process is a very difficult task because of  its multiple roles. First, is related to the ionization state of  the surface according to the following reactions, TiOH + H+ ⇔ TiOH2 +

(22)

TiOH + OH−

(23)

⇔

TiO− + H2 O

as well as to that of reactant dyes and products such as acids and amines. pH changes can thus influence the adsorption of  dye molecules onto the TiO2  surfaces, an important step for the photocatalytic oxidation to take place  [86]. Bahnemann et al. [87] have already reviewed that acid-base properties of the metal oxide surfaces can have considerable implications upon their photocatalytic activity. The point of zero charge (pzc) of the TiO2  (Degussa P25) is at pH 6.8  [88]. Thus, the TiO2  surface is positively charged in acidic media (pH  <  6 .8), whereas it is negatively charged under alkaline conditions (pH  >  6.8). Second, hydroxyl radicals can be formed by the reaction between hydroxide ions and positive holes. The positive holes are considered as the major oxidation species at low pH whereas hydroxyl radicals are considered as the predominant species at neutral or high pH levels [78,89]. It was stated that in alkaline solution • OH are easier to be generated by oxidizing more hydroxide ions available on TiO 2 surface, thus the efficiency of the process is logically enhanced [54,55,65,90,91]. Similar results are reported in the photocatalysed degradation of acidic azo dyes and triazine containing azo dyes  [7,9,52,74,92,93],  although it should be noted that in alkaline solution there is a Coulombic repulsion between the negative charged surface of photocatalyst and the hydroxide anions. This fact could prevent the

5

 I.K. Konstantinou, T.A. Albanis / Applied Catalysis B: Environmental 49 (2004) 1–14

formation of • OH and thus decrease the photoxidation. Very high pHs have been found favorable even when anionic azo dyes should hamper adsorption on the negatively charged surface [81]. At low pH, reduction by electrons in conduction band may play a very important role in the degradation of dyes due to the reductive cleavage of azo bonds. Third, the TiO2 particles tend to agglomerate under acidic condition and the surface area available for dye adsorption and photon absorption would be reduced  [86]. Hence, pH plays an important role both in the characteristics of textile waters and in the reaction mechanisms that can contribute to dye degradation, namely, hydroxyl radical attack, direct oxidation by the positive hole and direct reduction by the electron in the conducting band. The degradation rate of some azo dyes increases with decrease in pH as reported elsewhere  [42,67,94,95]. At pH  < 6, a strong adsorption of the dye on the TiO 2  particles is observed as a result of the electrostatic attraction of the positively charged TiO2   with the dye. At pH > 6.8 as dye molecules are negatively charged in alkaline media, their adsorption is also expected to be affected by an increase in the density of TiO − groups on the semiconductor surface. Thus, due to Coulombic repulsion the dyes are scarcely adsorbed [44,65,76]. For the above reasons the photocatalytic activity of anionic dyes (mainly sulphonated dyes) reached a maximum in acidic conditions followed by a decrease in the pH range 7–11  [42,53,58,68,75,76].  Moreover, the higher degradation rate at acid pH is seen also for Vis/TiO 2  experiments due to the efficient electron-transfer process due to strong surface complex bond formation. This effect is less marked in neutral/basic pH solutions  [67]. On the contrary, different optimal pHs (6–7) have been observed for the photocatalytic degradation of other azo dyes, and a decrease of degradation in both acidic and alkaline pH was reported [82,96]. The inhibitory effect seems to be more pronounced in the alkaline range (pH =  11–13). At high pH values the hydroxyl radicals are rapidly scavenged and they do not have the opportunity to react with dyes  [97]. An additional explanation for the pH effects can be related with changes in the specification of the dye. That is, protonation or deprotonation of the dye can change its adsorption characteristics and redox activity  [7]. Since the influence of the pH is dependent on dye type and on properties of TiO 2  surface his effect on the photocatalytic efficiency must be accurately checked before any application. 2.3.4. Effect of light intensity and irradiation time

Ollis et al. [98] reviewed the studies reported for the effect of light intensity on the kinetics of the photocatalysis process and stated that (i) at low light intensities (0–20mW/cm2 ), the rate would increase linearly with increasing light intensity (first order), (ii) at intermediate light intensities beyond a certain value (approximately 25 mW/cm2 ) [62], the rate would depend on the square root of the light intensity (half order), and (iii) at high light intensities the rate is in-

dependent of light intensity. This is likely because at low light intensity reactions involving electron–hole formation are predominant and electron–hole recombination is negligible. However, at increased light intensity electron–hole pair separation competes with recombination, thereby causing lower effect on the reaction rate. In the studies reviewed here, the enhancement of the rate of decolorization as the light intensity increased was also observed  [7,42,52,74,75]. It is evident that the percentage of decolorization and photodegradation increases with increase in irradiation time. The reaction rate decreases with irradiation time since it follows apparent first-order kinetics and additionally a competition for degradation may occur between the reactant and the intermediate products. The slow kinetics of dyes degradation after certain time limit is due to: (a) the difficulty in converting the N-atoms of dye into oxidized nitrogen compounds [99], (b) the slow reaction of short chain aliphatics with • OH radicals [100], and (c) the short life-time of photocatalyst because of active sites deactivation by strong by-products deposition (carbon etc.). 2.3.5. Effect of oxidants

It was observed that H 2 O2  and S2 O8 2− addition was beneficial for the photoxidation of the dyes of different chemical groups included azo dyes  [6,8,43,52]. The reactive radical intermediates (• SO4 − and • OH) formed from these oxidants by reactions with the photogenerated electrons can exert a dual function: as strong oxidant themselves and as electron scavengers, thus inhibiting the electron–hole recombination at the semiconductor surface [101] according to the following equations: H2 O2 + O2 •− → • OH + OH− + O2

 

(24)

H2 O2 + hv → 2• OH

(25)

H2 O2 + eCB −

(26)

→

•

OH + OH−

S2 O8 2− + eCB − → SO4 2− + SO4 • −

(27)

SO4 • − + H2 O → SO4 2− + • OH + H+

(28)

Moreover, the solution phase may at times be oxygen starved, because of either oxygen consumption or slow oxygen mass transfer. Peroxide addition thereby increases the rate towards what it would have been an adequate oxygen supply. The presence of S2 O8 2− positively influences the mineralization rate, despite the decreasing of pH as the oxidant properties of the system probably prevail on the effect of pH reduction. On the contrary, as far as the substrate is concerned, the faster degradation rate can be due to both the decrease of the pH and the oxidant action of S 2 O8 2− [43]. However, H2 O2  can also become a scavenger of valence band holes and • OH, when present at high concentration, [68,102–104]: H2 O2 + 2hVB + → O2 + 2H+

(29)

6

I.K. Konstantinou, T.A. Albanis / Applied Catalysis B: Environmental 49 (2004) 1–14

H2 O2 + • OH → H2 O + HO2 • HO2 • + • OH → H2 O + O2

 

(30)

HCO3 − + • OH → H2 O + CO3 • −

(35)

(31)

SO4 2− + h+ → SO4 • −

(36)

SO4 2− + • OH → SO4 • − + OH−

(37)

As both hVB + and • OH are strong oxidants for dyes, the photocatalytic oxidation will be inhibited when H 2 O2  level gets too high. Furthermore, H2 O2 can be adsorbed onto TiO2 particles to modify their surfaces and subsequently decrease its catalytic activity. Since the influence of the above additives, in particular H2 O2 , has been in some cases controversial and it appeared dependent on the substrate type and on various experimental parameters [105] the usefulness of which must be accurately checked before their application [43]. 2.3.6. Effect of humic acids, natural occurring ions and  solvents

The occurrence of dissolved inorganic ions is rather common in dye-containing industrial wastewater. Often, wastewater contains a mixture of pollutants, organic solvents as well as dissolved organic matter and humic substances, if mixed with other waste streams. These substances may compete for the active sites on the TiO 2   surface or deactivate the photocatalyst and, subsequently, decrease the degradation rate of the target dyes. Alternatively, they may act as light screens, thus reducing the photon receiving efficiency. The Vis/TiO2   photocatalytic degradation of different classes of dyes is reported to be retarded by many commonly used industrial solvents and acids, as well as by many naturally abundant mineral species and dissolved organic matter [99]. The retardation by humic substances may be by the combination effects of light attenuation, competition for active sites and surface deactivation  [106–108]. Finally, various solvents such as acetonitrile and ethanol were found to have a significant retardation effect on the photobleaching of dyes even at low concentrations  [68,106] as it is also stated for phenols and aromatic products  [109]. Of the anionic species studied (HCl, NaCl, NaNO 3 , HNO3 , H3 PO4  and NaHCO3 ), HCl exhibited the strongest inhibition effect followed by H 3 PO4  [68,106]. Inhibition effects of anions can be explained as the reaction of positive holes and hydroxyl radical with anions, that behaved as h + and • OH scavengers (Eqs. (32)–(37)) resulting prolonged color removal. Probably the adsorbed anions compete with dye for the photo-oxidizing species on the surface and preventing the photocatalytic degradation of the dyes [87,93,110]. Formation of inorganic radical anions (e.g. Cl• , NO3 • ) under these circumstances is possible to occur [111]. Cl− + hVB +

→

NO3 − + hVB +

Cl• orCl− + • OH → ClOH• −

→

(32)

NO3 • or

NO3 − + • OH → NO3 • + OH−

(33)

CO3 2− + • OH → OH− + CO3 • −

(34)

Although the reactivity of these radicals may be considered, they are not as reactive as h + and • OH [112] and thus, the observed retardation effect is still thought to be the strong adsorption of the anions on the TiO 2  surface [110]. The effect of several types of metal ions (Cu 2+, Zn2+, Fe3+ , Al3+ and Cd2+ ) on the photodegradation of non-azo dyes in TiO2  aqueous dispersions under visible light illumination, has been investigated by Chen et al.  [113]. They have concluded that Cu 2+ and Fe3+ ions have a strong suppressing effect on the photodegradation of all three dyes examined, by altering the interfacial electron-transfer pathway under visible light irradiation. They noted that the addition of Cu2+ and Fe3+ decreases the reduction of O 2  by the conduction electrons, subsequently blocks the formation of reactive oxygen species (O 2 •− / • OOH, • OH) and hence suppresses the photodegradation of dyes under visible irradiation. However, other metal ions such as Zn 2+ , Cd2+ and Al3+ affect the photoreaction only slightly through an alteration of the adsorption of dyes. On the basis of hydroxyl radical formation through photocatalytic reactions of Fe3+ ions and the products of their hydrolysis in aqueous solutions  [114] is assumed that, the presence of Fe3+ in the reaction environment, together with TiO2 , should increase the rate of the photocatalytic processes. An increased degradation rate was observed in the photocatalytic degradation of azo dye acid red 1 in TiO 2 suspensions containing Fe(III) aquo ions (10 −5 to 10−4 M) [115]. This beneficial behavior was attributed to the increased amount of dye adsorbed on the iron(III)-modified TiO2  surface and this was further confirmed by the fact that iron species such as Fe 2+ not adsorbed on the semiconductor had no kinetic effects. The beneficial effect of  Fe3+ ions was also found on the photocatalytic degradation of rhodamine B in aqueous TiO2   suspensions [33]. Baran et al. [116] studied the photocatalytic degradation of several anionic and cationic azo dyes in the presence of TiO2  and FeCl3 . They have found that Fe 3+ ions have a catalytic influence on the decolorization of the studied anionic dyes but an inhibiting influence on the decolorization of the cationic dyes. In conclusion, the role of Fe 3+ ions on the photocatalytic degradation of several dyes shows a controversial behavior depending on the physico-chemical properties of dyes. Thus, in the presence of these ions the specific azo dye degradation should be considered in order to determine the treatment efficiency. The photocatalytic decolorization of the triazine azo dye MX-5B was reported to increase slightly in the presence of  1 M of Cu2+ and Ni2+ at pH =  2 .4 [112]. Their reduced forms could trap holes and that explains the decrease of the e− /h+ recombination rate and a higher production of  • OH.

7

 I.K. Konstantinou, T.A. Albanis / Applied Catalysis B: Environmental 49 (2004) 1–14

Excess of Cu2+ and Ni2+ led to short-circuiting reactions, which created a cyclic process without generating active • OH and retarded the reaction. However, at pH 10.8 the photodegradation of MX-5B was completely inhibited by the trace quantities of Cu2+ and Ni2+. The deposition of  NiO2  on the surface of TiO 2  was found to deactivate the photocatalyst [112]. An understanding of the retardation effects not only aids in assessing the feasibility of using photocatalytic oxidations to treat wastewater, but also allows a thoughtful photocatalytic oxidation design. 2.4. Photocatalytic mineralization of dyes 2.4.1. Analysis of the end products

In order to assess the degree of mineralization reached during the photocatalytic treatment the formation of CO 2 and inorganic ions [6,32,34,44,58,91,117], is generally determined. However, in the presence of real wastewaters the monitoring of inorganic ions and CO 2  gives only a global estimation on the well functioning of the treatment, but does not provide information on the real decay of the contaminant. In such cases the determination of total organic carbon (TOC) and/or the measurement of the chemical oxygen demand (COD) or the biological oxygen demand (BOD) of the irradiated solution is generally used for monitoring the mineralization of the dye [7,32,42,52,58,76,80,118]. In general, at low reactant levels or for compounds which do not form important intermediates, complete mineralization and reactant disappearance proceed with similar half lives, but at higher reactant levels where important intermediates occur, mineralization is slower than the degradation of the parent compound. Until now, total mineralization has been observed for the photacatalytic degradation of  most of the azo dyes even at longer irradiation periods [42,44,58,74,76,119]. Only in the case of triazine containing dyes, the mineralization was not complete due to the high stability of triazine nucleus and the stable cyanuric acid was formed, as in the case of s-triazine herbicides [120], which fortunately is not toxic [41,52,80,121]. Usually COD or TOC values decrease with increase in irradiation time whereas the amount of NH 4 + and NO3 − ions increase with increase in irradiation time. However, the formation of Cl− and SO4 2− increases initially and subsequently remains unchanged. COD or TOC curves have an exponential or sigmoidal shape. The sigma-shaped curves indicating that is related to the formation of relative tolerant by-products [44,52,118]. This pattern means that during the first steps of the process where the solution is still colored there is only a small decrease of the parameter measured (TOC or COD or BOD) due to the fact that dye molecules are decomposed to lower molecular weight compounds and the resulting intermediates still contribute to the COD of  the solution. After the decolorization of the solution the COD decreases sharply (the linear segment of the S shaped curve) reaching a plateau that corresponds to the oxidation

of most stable compounds indicating that almost complete mineralization of intermediates has occurred. For chlorinated dye molecules, Cl − ions are easily released in the solution and are the first of the ions appearing during the photocatalytic degradation  [42,43,80]. This could be interesting in a process, where photocatalysis would be associated with a biological treatment which is generally not efficient for chlorinated compounds. Nitrogen is mineralized into NH4 +, NO3 − and N2 . The proportion depends mainly on the initial oxidation degree of nitrogen, the substrate structure and on irradiation time  [122–124]. By comparing the initial rates, NH 4 + appears as the primary product with respect to NO 3 − in the case of amine compounds. The nitrogen atoms in the amino-groups of the dyes can lead to NH4 + ions by successive attacks by hydrogen species R-NH2 + H• → R• + NH3 NH3 + H+

→

NH4 +

 

(38) (39)

The total amount of nitrogen-containing ions present in the solution at the end of the experiments is usually lower than that expected from stoichiometry indicating that N-containing species remain adsorbed in the photocatalyst surface or most probably, that significant quantities of N 2 and/or NH3   have been produced and transferred to the gas-phase.   [42,44,76].   The formation of N2   in azo dyes can be accounted for by the same processes responsible for NH4 + formation: R-N = N-R + H• R-N = N•

→

→

R-N = N• + R H

R• + N ≡ N

(40) (41)

When nitrogen is present in the −3 state as in amino groups or in pyrazoline ring, it spontaneously evolves as NH 4 + cations with the same oxidation degree, before being subsequently and slowly oxidized into nitrate  [58]. In the azo bonds each nitrogen atom is in its +1 oxidation degree. This oxidation degree favors the evolution of gaseous dinitrogen by the two step reduction process expressed previously. N 2 evolution constitutes the ideal case for a decontamination reaction involving totally innocuous nitrogen-containing final product. The dyes containing sulfur atoms are mineralized into sulfate ions [43,80]. In all the studies the formation of SO 4 2− was always observed and in most cases its stoichiometric formation was found in the final steps of the photoreaction when organic intermediates still were present  [43,52,80]. The reported initial slopes are positive indicating that SO 4 2− ions are initial products, directly resulting from the initial attack on the sulfonyl group. Release of sulphate ions upon dye degradation was a little slower than decolorization but much faster than TOC loss. Non-stoichiometric formation of sulphate ions is usually explained by a strong adsorption on the photocatalyst surface [44,125,126]. This strong adsorption could partially inhibit the reaction rate which, however, remains acceptable  [111,127]. Generally, it is found

8

I.K. Konstantinou, T.A. Albanis / Applied Catalysis B: Environmental 49 (2004) 1–14

that nitrate anions have little effect on the kinetics of reaction whereas sulfate, chloride and phosphate ions, especially at concentrations of greater than 10 −3 moldm−3 , can reduce the rate by 20–70% due to the competitive adsorption at the photoactivated reaction sites  [111].  The release of SO 4 2− can be accounted by an initial attack by a photo-induced • OH radical: OH− + p+ → OH• R-SO3 − + OH• SO3 • − + OH−

→

→

(42) R-OH + SO3 •−

SO4 2− + H•

(43) (44)

The attack of sulfonate groups would be favored if the molecule is adsorbed with its SO3 − group orientated to the surface [76]. The hydrogen atom generated can react with other radicals or with a neutral functional group such as an amino group. 2.4.2. Nature and evolution of organic intermediates

Cost effective treatment to complete pesticide mineralization is usually not feasible and by-products generation appears to be unavoidable with photocatalytic degradation process. Kinetics of formation and decomposition of  the intermediates are needed and identification of these byproducts needs to be established in order to (1) determine which specific compounds will appear in the effluent (2) increase our knowledge on the degradation pathways in order to reveal which step is crucial for the global reaction of the process. Identification of by-products is one of the keys to maximizing the overall process efficiency. Since hydroxyl radicals react non-selectively, various by-products are formed at low concentration levels. Various analytical techniques such as high performance liquid chromatography (HPLC) [65,67,128], gas chromatography–mass spectrometry (GC–MS) [34,44,65,68], liquid chromatography–mass spectrometry (LC–MS) [4,6,128],  1 H NMR [34,65,72], diffuse reflectance FT-IR [44,72,79,80,129] and electron spin resonance (ESR) [33,34] were used for the determination of organic intermediates. The azo dyes that have been already studied for their TiO2 -mediated photocatalytic degradation are summarized in   Table 1.  Generally, the sites near the azo bond (C–N=N–bond) is the attacked area in the photocatalytic degradation process, whilst the TiO 2  photocatalytic destruction of the C–N=  bond and –N–N– bonds leads to fading of the dyes   [32].  Aromatic intermediates were identified for most dyes. They are either aromatic amine or phenolic compounds. The formation of aminobenzenesulfonate suggests the reductive cleavage of the azo group prior to the opening of the aromatic ring   [58,96].  The formation of aromatic amines has also been reported in the natural aerobic reduction of azo dye   [130].   On the other hand, the formation of phenolic compounds as intermediates is commonly observed in the photocatalytic degradation of  other aromatic compounds  [37,58]. Several organic acids were found as aliphatic intermediates. Main products were

formic and acetic acids. Other organic acids detected were oxalic, glycolic, glyoxylic and malonic acids. The formation of these acids could correspond to the opening of aromatic and naphthalene rings followed by a sequence of oxidation steps which leads to progressively lower molecular weight acids and the evolution of CO 2 . Formation of CO2  takes place via decarboxylation of carboxylic acids according to the “photo-Kolbe” reaction [126]: R-COO− + h+

→

R + CO2

 

(45)

For most of the intermediate compounds the maximum concentration of formic acid was larger than that of acetic acid, as formic acid is more degradable by photocatalytic processes than acetic acid   [44,58]. Typical bell-shaped curves were obtained in some studies for the aromatic and naphthalene intermediates and the resulted acids [44,58] as in the case of other pollutants as pesticides  [37]. The possibility of generating molecular fragments during the photocatalyzed degradation that can be more toxic than the parent compound  [131] make also toxicity measurements obligatory part of the experiments. Moreover, if only partial degradation is envisaged, toxicity assessment of treated water becomes necessary. Toxicity analyses of the phototreated solutions of pesticides indicate the toxicity of the sum of compounds formed and not only those that have been identified. This parameter is very important for the treatment operation. The toxicity of all transient intermediates was compared to that observed for the parent compound. A variety of toxicity measurement systems exist, including those based on bacteria and algae, animal cells, swell mammals fish fly and zooplankton [132,133]. Two of them, Microtox method and the inhibition of   Escherichia coli  bacterial respiration, were usually applied in studies dealing with photocatalytic degradation of dyes. With Microtox method, the comparison was carried out by monitoring the bioluminescence of the bacteria   Vibrio fisheri   as a function of illumination time  [74,134]. It has been used extensively to assess the toxicity of a wide range of aquatic and terrestrial pollutants. This assay appears to be the best available method with high sensitivity and reliability for field monitoring and toxicity screening of industrial effluents. Alternatively the inhibition in bacterial respiration (e.g.   Escherichia coli) was followed [46,119,135]. This test has been reported to be well adapted to toxicity screening of effluents, being both easy to use and capable of providing reproducible and repeatable results [132]. The resulted EC50 values after the decolorization of the solution from the triazine dyes MX-5B and K-2G indicating that the intermediates produced from photocatalytic degradation were not toxic [52,80]. On the contrary, the presence of residual toxicity together with the permanence of about 20% of TOC content during the photocatalytic degradation of Remazol Black B (Reactive Black 5)  [119] and other azo dyes [117] indicate that the molecular fragments produced

9

 I.K. Konstantinou, T.A. Albanis / Applied Catalysis B: Environmental 49 (2004) 1–14

Table 1 Summary of TiO2 -mediated photocatalytic degradation studies of azo dyes Azo dyes

Light irradiation

Transformation products

UV Vis Vis Simulated solar Solar

−

Acid orange 20 Acid orange 52

References

 Monoazo dyes

Acid orange 7

−

         

[50,55,58,70,79,136,139,140] [71,137] [50,116] [44,67] [43]

UV-Vis

+

 

[12]

UV UV-Vis Simulated solar Solar

−

       

[55,76] [65] [128] [43]

UV UV UV UV/Solar Vis Vis Vis UV Vis UV Solar/UV UV-Vis Vis Vis UV UV-Vis Solar/UV UV

−

     

 

[58,139] [58] [58] [76,93] [116] [51] [51] [68] [115] [9] [7] [112,118] [116,142] [116] [54,56] [50] [76,93] [66,141]

Reactive red 120 Direct fast scarlet 4BS

UV UV

− −

 

[9] [56]

Reactive black 5

UV-Vis UV

−

   

[74] [85,119]

Reactive blue 221 Acid brown 14 Direct blue 1

UV Solar UV

−

     

[105] [42] [70]

Congo red

UV Solar

−

   

[58,73] [93]

UV-Vis UV

−

   

[53] [58]

Vis UV UV Vis UV

+

   

−

   

[129] [60] [9] [57] [75]

UV UV UV UV UV-Vis Vis Solar/UV

−

 

Tartazine, Acid yellow 17 New coccine Orange G Basic orange 66 Acid red 27, acid red 33, Allura red AC Acid red 14 Acid red 1 Basic yellow 15 Reactive yellow 17 Cationic blue X-GRL Basic blue 41 Acid yellow 23 Acid red 3B Red acid G Methyl red 30 H/K azo dyes

+ − +

+ + −

− − − / − − − − − −

           

− − − − − − −

         

− / − −

 

 Di- and triazo dyes

Acid black 1 Naphthol blue black Solophenyl green BLE, Direct blue 160 Lanasol blue Safira HEXL

−

− −

−

−

− − −

Triazine-containing azo dyes

Reactive brilliant red K-2G Procion red MX-5B Reactive yellow KD-3G Reactive red 15, reactive red 24 Reactive brilliant red X3B Reactive red 45 Reactive blue 4

+ / − − − − − −

         

[80] [52,80,112] [118] [118] [73] [116] [7]

10

I.K. Konstantinou, T.A. Albanis / Applied Catalysis B: Environmental 49 (2004) 1–14

at higher photochemical reaction times are toxic. In the case of anthraquinone dye Remazol brilliant Blue R the toxicity is completely removed at the first reaction times, but almost totally recovered with the progress of the photochemical process [119]. Acid orange 7 (AO 7) is the most studied compound among the azo dyes as far as its photocatalytic degradation under several experimental conditions. The degradation pathways and the formation of by-products is also fully described  [44,50,55,58,67,70,71,79,136–140] thus, AO 7 could be used as a model compound for oxidative degradation studies of azo dyes. Oxidative attack of an azo dye from the phenyl azonaphthol family as AO 7 leads to benzene sulfonate and naphthoquinone as primary degradation products. Vinodgopal et al.  [71] reported the formation of four by-products (benzene sulphonic acid, sulphoanilic acid, 1,4-naphthoquinone and phthalic acid) and Bauer et al.   [137]   have identified in addition quinone and 4-hydroxybenzene sulphonic acid during the first steps of Vis/TiO 2   photosensitized degradation of AO 7. The former products were also identified by Stylidi et al. [44], which studied the complete degradation of AO 7 under solar light irradiation. Twenty-two transformation products were identified in total, including 2-naphthol, 2-hydroxy-1,4-naphthoquinone, smaller aromatic intermediates such as pthalic acid and phtalimide and aliphatic acids such as fumaric, succinic, maleic and malonic acids. The lowest molecular weight compounds detected in that study, are oxalic, acetic and formic acids. The photoxidation of AO 7 para isomer (acid orange 20) was also studied under Vis/TiO2  [12,137]. Based on the previous identified by-products the major degradation pathways for AO 7 are shown in Fig. 1. From the aminoazobenzene sub-category of monoazo dyes, acid orange 52 (AO 52, namely also methyl orange) is studied in details by various advanced oxidation processes UV/H2 O2 , UV/TiO2 , Vis/TiO2 , solar light/TiO2 [43,55,65,76,128].  Up to 18 intermediates were identified including aniline, N , N -dimethyl aniline, hydroxy anilines, hydroxy analogues of AO 52, phenols, quinone, benzene sulphonic acid, demethylated analogues of AO 52, and various aliphatic and carboxylic acids. Spadaro et al.   [117] proposed that oxidation of aminoazobenzene dyes proceeds by the addition of an hydroxy radical to the carbon atom bearing the azo bond, followed by the breaking of the resulting adduct. The products such as benzenesulfonic acid,  N , N -dimethylaniline and 4-hydroxy- N , N -dimethylaniline could arise from such reactions. The electron withdrawing sulphonate group inhibits reactivity towards • OH of the ring that carried it, thus the ring with the amino group is the first target for the hydroxy radicals  [65]. The addition of  • OH on the carbon atom bearing the sulphonate group and the subsequent elimination of SO 3   is an non probable pathway due to the electron withdrawing effect of sulphonate group and steric hindrance. On the contrary,

HO +Na O 3 S

N

HO 3 S

N

2.4.2.1. Monoazo dyes.

N

OH

H NH 2

OH OH + N2

HO 3 S

O OH O HO 3 S

O

OH

O

2-

SO 4

+

O

O

O

COOH COOH

O O

Benzoic acid derivatives

CO 2 + H 2 0

Aliphatic + Carboxylic acids

Fig. 1. Major photocatalytic pathways of acid orange 7, a representative azo dye for phenyl-azonaphthol chemical group, based on the identification of by-products from previous reported degradation studies [12,44,67,71,137].

hydroxy-analogue derivatives of AO 52 are identified  [65]. Similar reaction was observed also during the degradation of aminoazobenzene acid orange 5 (AO 5) and the hydroxyazo dye AO 7   [138]   but is presumed not to be the major oxidation way. The –N(CH 3 )2  substituent group is also an important site of attack  [65,128]  and thus the demethylated analogues could be formed by such reactions. The full characterization of the resulted degradation products is given in Baiocchi et al.  [128]. According to the previous identified by-products of AO 52 the major degradation pathways for aminoazobenzene dyes are shown in Fig. 2. Bezenesulfonic acid and phenols were identified also as intermediates for the photooxidation of other monoazo dyes such as Tartazine, Acid yellow 17, New Coccine and Orange G [58,76,139]. Zhang and Tian [66,141] studied the effect of substituents and the spectral characterization of excited states on the TiO 2  photocatalytic degradation of 30 H/K acid based azo dyes. They have concluded that the substituent dependent intramolecular resonance energy in the conjugated molecules of  o -arylazonaphthols determines the light fastness of the dyes and that the triplet state of the dyes

11

 I.K. Konstantinou, T.A. Albanis / Applied Catalysis B: Environmental 49 (2004) 1–14

e ring hydroxylated dye derivatives

e

OH

.

OH R1

+Na O3S c

d

.

N N a1

c

ring hydroxylated dye derivatives

a1

b

a2

R2 b

Dealkylated Dye derivatives

a2 R1 N

+ N2 +

HO3S

ring hydroxylated dye derivatives

d

N

R2

OH

R1

.

HO

N R2

a2

OH

.

a1

H HO3S

OH

N R2

b

a1 NH2

c OH

.

HO

NH2

OH

.

OH 2SO4 +

HO

O

OH

O

+ + NH4 + NO3 Carboxylic + Aliphatic Acids

CO2 + H2O

Fig. 2. Major photocatalytic degradation pathways for aminoazobenzene dyes based on the identification of by-products from previous reported degradation studies of acid orange 52 [65,128].

does not react with • OH and O2 •− . Finally, several studies have been reported for the TiO 2  photocatalytic degradation of other monoazo dyes (Table 1) compiling data only on the decolorization of the solution and the monitoring of the end products (TOC or COD and inorganic ions) under various operational conditions. Similar results were obtained for some di- and triazo dyes. Phenol, and 4-nitro-2-hydroxy phenol were identified as intermediate products for the TiO2   photocatalytic degradation of Congo Red and Acid Black 1, respectively   [53,58,75,93]. 1,2 Naphthaquinone were identified for Naphthol Blue Black azo dye   [129]. Much work has been done on the operational conditions and the parameters influencing the photocatalytic degradation of di and triazo dyes be several researchers ( Table 1) [9,42,56,57,70,74,75,82,119]. In some studies comparative photocatalytic degradation rates between azo dyes from different groups are reported. Monoazo dyes are easier oxidized than diazo dyes which may in turn be easier than triazo dyes as long as auxochrome groups are similar   [9,58].   This trend is also observed in ozonation of azo dyes   [143].   Dyes of  the -naphthol type degrade faster than dyes derived from 2.4.2.2. Di- and triazo dyes.

 N , N -dimethylaniline and dyes containing the sulphonic acid

group degrade faster than those with the carboxylic group [55]. Several studies reported the photocatalytic oxidation of triazine azo dyes in various operational conditions [7,52,73,80,118] but only one of them reported the photocatalytic pathways and the formation of transformation products [80]. The photocatalytic degradation of triazine-containing azo dyes, Procion Red MX-5B and Reactive Brilliant Red K-2G was investigated in aqueous TiO2  dispersions [80]. The whole photocatalytic degradation proceeds in three steps. In the first step, the more active bonds were hydroxylated. These included the C–N bond linked to the benzene ring or the naphthalene ring and the C–S bond of sulfonate group linked to the naphthalene ring or the benzene ring, to form organic acids with or without hydroxyl group and the related ions (SO 4 2− and NH4 + ). For the second step, the groups linked to the triazine ring were replaced by hydroxyl group to yield the well-known cyanuric acid, as in the case of s-triazines herbicides, and the related ions (NO 3 − and Cl−). At the same time the aromatic acids produced from the first step were subsequently hydroxylated and led to the cleavage of aromatic ring to form 2.4.2.3. Triazine containing azo dyes.

12

I.K. Konstantinou, T.A. Albanis / Applied Catalysis B: Environmental 49 (2004) 1–14

aliphatic acids. The third step involved the further oxidation of the aliphatic acids to produce CO 2  and water [80]. 3. Conclusions

Effective destruction of azo dyes belonging to different chemical groups is possible by photocatalysis in the presence of TiO2  suspensions and UV, Vis or solar light. The kinetics of the photocatalytic oxidation follows a Langmuir–Hinshelwood model and depends on several factors such as, dye concentration, mass of catalyst, wavelength, radiant flux and the addition of oxidants or the presence of natural occurring substances (humic substances and/or inorganic ions). Since the influence of the above factors has been in some cases controversial and it appeared dependent on substrate type and on various experimental parameters their impact on the dye degradation must be accurately checked before their application. The great part of the studies on the photocatalytic degradation of azo dyes relies only on the monitoring of solution decolorization, TOC or COD and inorganic ions. Monitoring the disappearance rate of the target dyes is not the most appropriate parameter to classify the efficiency of this process. Only few studies reported thorough mechanisms with detailed reaction steps of the different pathways leading to several photoproducts. Lot of work has to be done also in the quantification of these intermediates. Quantitative data about the relative importance of the different routes of degradation is lacking and will help further in the mechanistic studies of the photocatalytic reactions. Cost-effective treatments to complete compound mineralization are usually not practicable, therefore, by product evaluation is one of the keys to optimize each treatment and to maximize the overall process. Toxicity tests of the treated water will gather also useful information about the practical application of the photocatalytic process. The better understanding of the photocatalytic process and the operative conditions could give great opportunities for its application for the destruction of  environmental organic contaminants.

References [1] H. Zollinger (Ed.), Color Chemistry: Synthesis, Properties and Applications of Organic Dyes and Pigments, second revised ed., VCH, 1991. [2] J. Weber, V.C. Stickney, Wat. Res. 27 (1993) 63. [3] C. Ràfols, D. Barceló, J. Chromatogr. A 777 (1997) 177. [4] A. Houas, H. Lachheb, M. Ksibi, E. Elaloui, C. Guillard, J.M. Hermann, Appl. Catal. B: Environ. 31 (2001) 145. [5] U. Pagga, D. Bruan, Chemosphere 15 (1986) 479. [6] A. Bianco-Prevot, C. Baiocchi, M.C. Brussino, E. Pramauro, P. Savarino, V. Augugliaro, G. Marci, L. Palmisano, Environ. Sci. Technol. 35 (2001) 971. [7] B. Neppolian, H.C. Choi, S. Sakthivel, B. Arabindoo, V. Murugesan, Chemosphere 46 (2002) 1173. [8] M. Saquib, M. Muneer, Dyes Pigments 56 (2003) 37–49.

[9] W.Z. Tang, H. An, Chemosphere 31 (1995) 4157. [10] V. Meshko, L. Markovska, M. Mincheva, A.E. Rodrigues, Wat. Res. 35 (2001) 3357. [11] W.S. Kuo, P.H. Ho, Chemosphere 45 (2001) 77. [12] C. Galindo, P. Jacques, A. Kalt, Chemosphere 45 (2001) 997. [13] P. Cooper, J. Soc. Dyers. Colour. 109 (1993) 97. [14] S.S. Patil, V.M. Shinde, Environ. Sci. Technol. 22 (1988) 1160. [15] A.T. More, A. Vira, S. Fogel, Environ. Sci. Technol. 23 (1989) 403. [16] M.C. Venceslau, S. Tom, J.J. Simon, Environ. Technol. 15 (1994) 917. [17] I. Arslan, I.A. Balcioglou, Dyes Pigments 43 (1999) 95. [18] U. Pagga, K. Taeger, Wat. Res. 28 (1994) 1051. [19] H.L. Sheng, M.L. Chi, Wat. Res. 27 (12) (1993) 1743. [20] S.H. Lin, W.Y. Liu, J. Environ. Eng. 120 (1994) 437. [21] F. Strickland, S. Perkins, Text. Chem. Color. 5 (1995) 11. [22] Y.M. Slokar, A.M. Le Marechal, Dyes Pigments 37 (1998) 335. [23] W.G. Kuo, Wat. Res. 26 (1992) 881. [24] E. Balanosky, J. Fernadez, J. Kiwi, A. Lopez, Wat. Sci. Technol. 40 (1999) 417. [25] W. Feng, D. Nansheng, Z. Yuegang, Chemosphere 39 (1999) 2079. [26] C. Morrison, J. Bandara, J. Kiwi, J. Adv. Oxid. Technol. 1 (1996) 160. [27] S.F. Kang, C.H. Liao, S.T. Po, Chemosphere 41 (2000) 1287. [28] I. Arslan, T. Akmehmet, T. Tuhkamen, Environ. Technol. 20 (1999) 921. [29] N.H. Ince, D.T. Gonenc, Environ. Technol. 18 (1997) 179. [30] K. Vinodgopal, P. Kamat Chemtech. 4 (1996) 18. [31] C. Lizama, M.C. Yeber, J. Freer, J. Baeza, H.D. Mansilla, Wat. Sci. Technol. 44 (2001) 197. [32] F. Zhang, J. Zhao, T. Shen, H. Hidaka, E. Pelizzetti, N. Serpone, Appl. Catal. B: Environ. 15 (1998) 147. [33] P. Qu, J. Zhao, T. Shen, H. Hidaka, J. Mol. Catal. A: Chem. 129 (1998) 257. [34] G. Liu, T. Wu, J. Zhao, H. Hidaka, N. Serpone, Environ. Sci. Technol. 33 (1999) 2081. [35] J. Zhao, T. Wu, K. Wu, K. Oikawa, H. Hidaka, N. Serpone, Environ. Sci. Technol. 32 (1998) 2394. [36] M.R. Hoffmann, S.T. Martin, W. Choi, D.W. Bahnemann, Chem. Rev. 95 (1995) 69. [37] I.K. Konstantinou, T.A. Albanis, Appl. Catal. B: Environ. 42 (2003) 319. [38] A.L. Linsebigler, L. Guangquan, J.T. Yates, Chem. Rev. 95 (1995) 735. [39] P. Reeves, R. Ohlhausen, D. Sloan, K. Pamplin, T. Scoggins, Solar Energy 48 (1992) 413. [40] X.Z. Li, M. Zang, Wat. Sci. Technol. 34 (9) (1996) 49. [41] Y. Wang, Wat. Res. 34 (2000) 990. [42] S. Sakthivel, B. Neppolian, M.V. Shankar, B. Arabindoo, M. Palanichamy, V. Murugesan, Solar Energy Mater. Solar Cells 77 (2003) 65. [43] V. Augugliaro, C. Baiocchi, A. Bianco-Prevot, E. Garcia-Lopez, V. Loddo, S. Malato, G. Marci, L. Palmisano, M. Pazzi, E. Pramauro, Chemosphere 49 (2002) 1223. [44] M. Stylidi, D.I. Kondarides, X.E. Verykios, Appl. Catal. B: Environ. 40 (2003) 271. [45] I.A. Balcioglou, I. Arslan, Environ. Pollut. 103 (1998) 261. [46] S. Gomes de Moraes, R. Sanches Freire, N. Duran, Chemosphere 40 (2000) 369. [47] I.A. Balcioglou, I. Arslan, Environ. Technol. 18 (1997) 1053. [48] P. Peralta-Zamora, S. Gomes de Moraes, R. Pelegrini, M. Freire Jr., J. Reyes, H. Mansilla, N. Duran, Chemosphere 36 (1998) 2119. [49] I.A. Alaton, I.A. Balcioglou, D.W. Bahnemann, Wat. Res. 36 (2002) 1143. [50] F. Chen, Y. Xie, J. Zhao, G. Lu, Chemosphere 44 (2001) 1159. [51] G.A. Epling, C. Lin, Chemosphere 46 (2002) 561. [52] C.M. So, M.Y. Cheng, J.C. Yu, P.K. Wong, Chemosphere 46 (2002) 905.

 I.K. Konstantinou, T.A. Albanis / Applied Catalysis B: Environmental 49 (2004) 1–14

[53] J. Grzechulska, A.W. Morawski, Appl. Catal. B: Environ. 36 (2002) 45. [54] Z. Shourong, H. Qingguo, Z. Jun, W. Bingkun, J. Photochem. Photobiol. A: Chem. 108 (1997) 235. [55] M.S.T. Conçalves, A.M.F. Oliveira-Campos, M.M.S. Pinto, P.M.S. Plasencia, M.J.R.P. Queiroz, Chemosphere 39 (1999) 781. [56] C. Zhu, L. Wang, L. Kong, X. Yang, L. Wang, S. Zheng, F. Chen, F. MaiZhi, H. Zong Chemosphere 41 (2000) 303. [57] R. Suárez-Parra, I. Hernadez-Perez, M.E. Rincon, S. Lopez-Ayala, M.C. Roldan-Ahumada, Solar Energy Mat. Solar Cells 76 (2003) 189. [58] K. Tanaka, K. Padermpole, T. Hisanaga, Wat. Res. 34 (2000) 327. [59] P.C. Vandevivere, R. Bianchi, W. Verstraete, J. Chem. Technol. Biotechnol. 72 (1998) 289. [60] C. Gomes da Silva, J.L. Faria, J. Photochem. Photobiol. A: Chem. 155 (2003) 133. [61] M.A. Brown, S.C. De Vito, Crit. Rev. Environ. Sci. Technol. 23 (1993) 249. [62] J.M. Hermann, Catal. Today 53 (1999) 115. [63] H.D. Burrows, M. Canle, J.A. Santaballa, S. Steenken, J. Photyochem. Photobiol. B: Biol. 67 (2002) 71. [64] S. Chiron, A. Fernández-Alba, A. Rodriguez, E.C. Calvo, Wat. Res. 34 (2000) 366. [65] C. Galindo, P. Jacques, A. Kalt, J. Photochem Photobiol. A: Chem. 130 (2000) 35. [66] H. Zhan, H. Tian, Dyes Pigments 37 (1998) 231. [67] J. Bandara, J.A. Mielczarski, J. Kiwi, Langmuir 15 (1999) 7680. [68] N. Daneshvar, D. Salari, A.R. Khataee, J. Photochem. Photobiol. A: Chem. 157 (2003) 111. [69] V. Maurino, C. Minero, E. Pelizzetti, M. Vincenti, Coll. Surf. A 151 (1999) 329. [70] K. Vinodgopal, I. Bedja, S. Hotchandani, P.V. Kamat, Langmuir 10 (1994) 1767. [71] K. Vinodgopal, D. Wynkoop, P. Kamat, Environ. Sci. Technol. 30 (1996) 1660. [72] C. Galindo, P. Jacques, A. Kalt, J. Photochem. Photobiol. A: Chem. 141 (2001) 47. [73] Y. Xu, C.H. Langford, Langmuir 17 (2001) 897. [74] L.B. Reutergarth, M. Iangphasuk, Chemosphere 35 (1997) 585. [75] T. Sauer, G.C. Neto, H.J. Jose, R.F.P.M. Moreira, J. Photochem. Photobiol. A: Chem. 149 (2002) 147. [76] H. Lachheb, E. Puzenat, A. Houas, M. Ksibi, E. Elaoui, G. Guillard, J.M. Hermann, Appl. Catal. B: Environ. 39 (2002) 75. [77] C.H. Giles, A.P. Da Silva, I.A. Easton, J. Colloid. Interf. Sci. 47 (1974) 766. [78] W.Z. Tang, C.P. Huang, Wat. Res. 29 (1995) 745. [79] L. Lucarelli, V. Nadtochenko, J. Kiwi, Langmuir 16 (2000) 1102. [80] C. Hu, J.C. Yu, Z. Hao, P.K. Wong, Appl. Catal. B: Environ. 42 (2003) 47. [81] A. Mills, R.H. Davis, D. Worsley, Chem. Soc. Rev. 22 (1993) 417. [82] I. Poulios, I. Aetopoulou, Environ. Technol. 20 (1999) 479. [83] L. Zhang, C.Y. Liu, X.M. Ren, J. Photochem. Photobiol. A 85 (1995) 239. [84] R.J. Davis, J.L. Gainer, G. O’Neal, I. Wu, Wat. Environ. Res. 66 (1994) 50. [85] B. Zielinska, J. Grzechulska, B. Grzmil, A.W. Morawski, Appl. Catal. B: Environ. 35 (2001) L1–L7. [86] M.A. Fox, M.T. Dulay, Chem. Rev. 93 (1993) 341. [87] D.W. Bahnemann, J. Cunningham, M.A. Fox, E. Pelizzetti, P. Pichat, N. Serpone, in: R.G. Zepp, G.R. Heltz, D.G. Crosby (Eds.), Aquatic Surface Photochemistry, Lewis Publishers, Boca Raton, 1994, p. 261. [88] I. Poulios, I. Tsachpinis, J. Chem. Technol. Biotechnol. 71 (1999) 349. [89] S. Tunesi, M. Anderson, J. Phys. Chem. 95 (1991) 3399. [90] A. Sharma, P. Rao, R.P. Mathur, S.C. Ametha, J. Photochem. Photobiol. A 86 (1995) 97.

13

[91] S. Sakthivel, B. Neppolian, M. Palanichamy, B. Arabindoo, V. Murugesan, Indian J. Chem. Tech. 6 (1999) 161. [92] K. Hustert, R.G. Zepp, Chemosphere 24 (1992) 335. [93] C. Guillard, H. Lachheb, A. Houas, M. Ksibi, E. Elaloui, J.M. Hermann, J. Photochem. Photobiol. A: Chem. 158 (2003) 27. [94] S. Naskar, S. Arumugam Pillay, M. Chanda, J. Photochem. Photobiol. A 113 (1998) 254. [95] B. Neppolian, S. Sakthivel, M. Palanichamy, B. Arabindoo, V. Murugesan, Stud. Sur. Sci. Cat. 113 (1998) 329. [96] W.Z. Tang, Z. Zhang, H. An, M.O. Quintana, D.F. Torres, Environ. Technol. 18 (1997) 1. [97] A.P. Davis, C.P. Huang, Wat. Sci. Technol. 21 (1990) 455. [98] D.F. Ollis, E. Pelizzetti, N. Serpone, Environ. Sci. Technol. 25 (1991) 1523. [99] J. Bandara, V. Nadtochenko, J. Kiwi, C. Pulgarin, Wat. Sci. Technol. 35 (1997) 87. [100] C. Walling, Acc. Chem. Res. 8 (1975) 125. [101] C.K. Grätzel, M. Jirousek, M. Grätzel, J. Mol. Catal. 60 (1990) 375. [102] O. Legrini, E. Oliveros, A.M. Braun, Chem. Rev. 93 (1993) 671. [103] M.W. Peterson, J.A. Tuner, A.J. Nozik, J. Phys. Chem. 95 (1991) 221. [104] S. Malato, J. Blanco, C. Richter, B. Braun, M.I. Maldonado, Appl. Catal. B: Environ. 17 (1998) 347. [105] A.B. Isil, I. Yuksel, J. Environ. Sci. Health A 31 (1996) 123. [106] G.A. Epling, C. Lin, Chemosphere 46 (2002) 937. [107] J. Wiszniowski, D. Robert, J. Surmacz-Gorska, K. Miksch, J.V. Weber, J. Photochem. Photobiol. A: Chem. 152 (2002) 267. [108] M. Bekbolet, A.S. Suphandag, C.S. Uyguner, J. Photochem. Photobiol. A: Chem. 148 (2002) 121. [109] C. Renzi, C. Guillard, J.M. Hermann, P. Pichat, G. Baldi, Chemosphere 35 (1997) 819. [110] M. Sökmen, A. Özkan, J. Photochem. Photobiol. A: Chem. 147 (2002) 77. [111] M. Abdullah, G.K.C. Low, R.W. Matthews, J. Phys. Chem. 94 (1990) 6820. [112] C. Hu, J.C. Yu, Z. Hao, P.K. Wong, Appl. Catal. B: Environ. 46 (2003) 35. [113] C. Chen, X. Li, W. Ma, J. Zhao, H. Hidaka, N. Serpone, J. Phys. Chem. B 106 (2002) 318. [114] W. Feng, D. Nansheng, Chemosphere 41 (2000) 1137. [115] M. Mrowetz, E. Selli, J. Photochem. Photobiol. A: Chem., in press. [116] W. Baran, A. Makowski, W. Wardas, Chemosphere 53 (2003) 87. [117] J.T. Spadaro, L. Isabelle, V. Renganathan, Environ. Sci. Technol. 28 (1994) 1389. [118] H. Chun, W. Yizhong, Chemosphere 39 (1999) 2107. [119] C.A.K. Gouvea, F. Wypych, S.G. Moraes, N. Duran, N. Nagata, P. Peralta-Zamora, Chemosphere 40 (2000) 433. [120] C. Minero, V. Maurino, E. Pelizzetti, Res. Chem. Interm. 23 (1997) 291. [121] C. Hu, Y. Wang, Chemosphere 39 (1999) 2107. [122] E. Puzenat, H. Lacheb, M. Karkmaz, A. Houas, C. Guillard, J.M. Hermann, Int. J. Photoenergy 5 (2003) 51. [123] K. Nohara, H. Hidaka, E. Pelizzetti, N. Serpone, J. Photochem. Photobiol. A: Chem. 102 (1997) 265. [124] V. Maurino, C. Minero, E. Pelizzetti, P. Piccinini, N. Serpone, H. Hidaka, J. Photochem. Photobiol. A: Chem. 109 (1997) 171. [125] J.M. Hermann, C. Guillard, P. Pichat, Catal. Today 17 (1993) 7. [126] M. Vauthier, C. Guillard, J.M. Hermann, J. Catal. 201 (2001) 46. [127] M. Kerzhentsev, C. Guillard, J.M. Herrmann, P. Pichat, Catal. Today 27 (1996) 215. [128] C. Baiocchi, M.C. Brussino, E. Pramauro, A. Bianco-Prevot, L. Palmisano, G. Marci, Int. J. Mass Spectrom. 214 (2002) 247. [129] C. Nasr, K. Vinodgopal, L. Fisher, S. Hotchandani, A.K. Chattopadhyay, P.V. Kamat, J. Phys. Chem. 100 (1996) 8436. [130] E.J. Weber, R.L. Adams, Environ. Sci. Technol. 29 (1995) 1163. [131] W.F. Jardim, S.G. Moraes, M.M.K. Takiyama, Wat. Res. 31 (1997) 1728.

14

I.K. Konstantinou, T.A. Albanis / Applied Catalysis B: Environmental 49 (2004) 1–14

[132] I.E. Tothill, A.P.F. Turner, Tr. Anal. Chem. 15 (1996) 178. [133] V.B. Manilal, A. Haridas, R. Alexander, G.D. Surender, Wat. Res. 26 (1992) 1035. [134] S. Ledakowicz, M. Solecka, R. Zylla, J. Biotech. 89 (2001) 175. [135] I. Poulios, E. Micropoulou, R. Panou, E. Kostopoulou, Appl. Catal. B: Environ. 41 (2003) 345. [136] K. Vinodgopal, P. Kamat, Environ. Sci. Technol. 29 (1995) 841. [137] C. Bauer, P. Jacques, A. Kalt, J. Photochem. Photobiol. A: Chem. 140 (2001) 87.

[138] C. Galindo, P. Jacques, A. Kalt, J. Adv. Oxid. Technol. 4 (1999) 400. [139] K.V. Subba Rao, B. Lavedrine, P. Boule, J. Environ. Photochem. Photobiol. A: Chem. 154 (2003) 189. [140] J. Fernandez, J. Kiwi, C. Lizama, J. Freer, J. Baeza, H.D. Mansilla, J. Photochem. Photobiol. A: Chem. 151 (2002) 213. [141] H. Zhan, H. Tian, Dyes Pigments 37 (1998) 249. [142] E. Stathatos, T. Petrova, P. Lianos, Langmuir 17 (2001) 5025. [143] H.Y. Shu, C.R. Huang, Chemosphere 29 (1995) 2597.