Heterogeneous Photocatalysis in Environmental ... - Wiley Online Library

Dev. Chem. Eng. Mineral Process., 8(5/6),pp.505-550, 2000.

Heterogeneous Photocatalysis in Environmental Remediation Dingwang Chen, M. Sivakumar and Ajay K. Ray* Department of Chemical and Environmental Engineering, The National University of Singapore, 10 Kent Ridge Crescent, SINGAPORE I19260 Semiconductor-based photocatalytic processes have been studied for nearly 1.5 years due to their many intriguing advantages in environmental remediation and other areas. In this paper, underlying principles and mechanism of the photocatalytic degradation process were discussed, followed by thermodynamics, kinetics, mass transfer effects were discussed. The kinetics of photocatalytic degradation was analysed with the aim of determining the favourable conditions to obtain high quantum yield. Primary parameters that injluence the process such as catalyst

loading, dissolved oxygen, pH, temperature, light intensity, crystalline structure, etc. have been discussed in detail. Different types of photocatalytic reactors, process eflciencies and applications in environmental engineering as well as in other areas have also been mentioned. The main barrier to the commercialization of the processes is the low quantum yield. However, it is expected that large-scale applications could be achieved with signijkant progress by improvement of process pelformance.

Introduction Water pollution is a major problem confronting us in the 21" century, particularly in developing countries. People have now realized that water is no longer an endless resource. As the global population grows, so does the need for clean water. With the rapid development of science and technology, many industries, such as chemical, petrochemical, pharmaceutical and mining require large quantities of water. Even

* Author for correspondence. 505

D.Chen, M.Sivakumar and A.K. Ray

ultrapure water is extensively needed in the pharmaceuticals, microelectronics and semiconductor industries. Unfortunately, the discharged water from these industries is often contaminated with toxic organic and inorganic compounds. In the recent two decades, the pollutants, coming from a variety of industrial and agricultural activities, are contaminating water and air to an unacceptable level all over the world. Meanwhile, government regulations and public pressure no longer tolerate untreated discharges. Hence, the need for innovative and effective water treatment processes for both industry and human environment is extremely obvious. There are mainly two purposes in water purification study. One is to reduce the pollution levels in discharged streams to meet the government regulations. The other is to purify water to ultrapure water, mainly for industries such as pharmaceuticals, microelectronics and semiconductor. An ideal water treatment process should completely mineralize all the toxic species present in the waste stream without leaving behind any hazardous residues. It should also be cost-effective. At the current state of development, none of the treatment technologies has reached this ideal situation. There are a number of waste disposal methods currently in practice with varying degrees of success. Figure 1 lists different wastewater treatment technologies either currently available or in varied stage of development. At present, the disposal of the bulk of the industrial wastes is based on the processes developed on phase transfer principles [I], even though none of them is completely satisfactory. Air stripping is widely employed for the removal of volatile organic contaminants in wastewater. But, the process just transfers the pollutants from water phase to air phase rather than destroying them. Granular activated carbon (GAC) adsorption is another commercialized process for water purification. However, the spent carbon, on which pollutants are adsorbed, must then be disposed again. For above reasons attention is being given to alternative destructive oxidation processes for water treatment [2]. In contrast to the above non-destructive technologies, chlorination and ozonation are two destructive technologies used in water industry. But, both use strong oxidants that are seriously hazardous, therefore of undesirable nature. Furthermore, it has now been proven that ozonation generates small levels of bromate ions, a recognized

Heterogeneous Photocatalysis in Environmental Remediation

!I

.m

T

507

D.Chert, M.Sivakumar and A.K. Ray

cancer-causing agent [3]. Biodegradation of industrial wastes is still not common as the degradation rates are usually very slow and at higher concentrations of organic contaminants, it presents difficulties during the operation [4]. Besides, toxic organics at times kill the active microorganisms that are intended to degrade them. Although incineration of organic wastes has been widely used in developing countries, it releases other toxic species into the air, such as dioxin and furan 151. In order to deal with the ever-increasing degenerating environment and to address the problems present in the commercial water treatment technologies, researchers and scientists have been developing innovative treatment processes. A relatively new class of technologies that shows promise for treating water contaminated with organic compounds. These new technologies are known as advanced oxidation processes (AOPs). H202/W, O m , Hz02/03/UV, TiOZ/W and V W processes are major AOPs [61.

In H202/W process, H202 is cleaved into hydroxyl radicals under the illumination of W (mainly at 254 nm). These hydroxyl radicals attack and decompose the organic compounds in aqueous solution by hydrogen abstraction, electrophilic addition and electron transfer. The oxidation rate of contaminants in this process is usually limited by the rate of formation of hydroxyl radicals, due to the rather small absorption cross section of H2OZat 254 nm, in particular, in the cases where organic substrates wili act as inner filter [ 6 ] . Like other hydroxyl radical generating degradation processes.

0 3 / wprocess can also oxidizes a great variety of organic compounds in wastewater. In this process, aqueous systems saturated with ozone are irradiated with W light of 254 nm. Oxidative degradation rates are much higher than those observed in experiments where either W light or ozone has been used separately. The most serious and rather specific problem in the technical development of this process is the low ozone solubility in water and consequent mass transfer limitations in photoreactors [7]. 0 3 / H 2 0 2 / U V is quite similar to 0 3 I L T V process. This process is enhanced by the photochemical generation of hydroxyl radicals by the addition of hydrogen peroxide, due to the dominant production of hydroxyl radicals from

H202.

The vacuum ultraviolet ( V W ) process uses W - C at 190 nm emitted by excimer

508


lamp as light source. V W photolysis of H20is the main and efficient means of the generation of hydroxyl radicals. ... (1)

In general, the V W process is very simple and has the particular advantage that no chemicals need to be added. However, only a few research groups, so far, are active in this field due to the very limited availability of the excimer light sources [ 6 ] . The ultimate goal of AOPs is complete mineralization of organic compounds to carbon dioxide and mineral acids [8]. Among the APOs, heterogeneous photocatalysis (Ti02/UV) has received considerable attention in the last 20 years as an alternative for treating water polluted with toxic organic compounds and some metal ions. This process is based on aqueous phase hydroxyl radical chemistry and couples lower energy UV-A light with semiconductors acting as photocatalyst, and has emerged as a viable alternative for solving environmental problems overcoming many of the drawbacks that exist in traditional water treatment technologies. Photocatalysis research is driven by legislation in industrialized countries that encourage water purification (decontamination, detoxification, decolorization, deodorization) and simultaneous contaminant destruction. Several books [8-171 and review articles [6, 18-22] have been published in the last years reporting studies on photocatalytic

processes both in gas and liquid phases. Heterogeneous photocatalysis differs from the other AOPs as it employs a reusable catalyst and there is no need to add any additional oxidants. Its main advantages over the commercial water treatment technologies include: (i) The oxidation is powerful and indiscriminate leading to the mineralization of almost all-organic pollutants in wastewater. Even carbon tetrachloride which was considered as hydroxyl radical resistant 1231 could also be mineralized [24]. (ii) The process effluents (COz, H20,and mineral acid) are benign to the environment, so it is called a “green technology”. (iii) The oxidant used in the process is atmospheric oxygen, and therefore, in general, there is no need for additional oxidizing agents. (iv) The catalysts are cheap, non-hazardous, stable and reusable. (v) The light required to activate the catalyst is low energy UV-A, and

509


therefore, solar illumination is possible. (vi) The process is commercially comparable to activated carbon adsorption for intermediate and large capacities [lo, 251. Photoelectrochemistry received a great deal of attention after the discovery by Fujishima and Honda [26] in 1972 that U V irradiated TiOz electrodes immersed in water split water into hydrogen and oxygen gases. As investigation in this field progressed, researchers found that the potential of photocatalytic processes, a subset of photoelectrochemical phenomena, as an alternative to oxidize organic compounds in water to carbon dioxide and mineral acids. Ollis and his co-workers [27-301 were the first to implement photocatalysis as a method for water purification when they conducted experiments for photomineralization of halogenerated hydrocarbon. Since then, numerous studies have shown that a great variety of dissolved organic compounds could be oxidized to CO, by photocatalysis [23, 31-36]. Apart from the oxidation of organic compounds, reduction of some toxic metal ions [37-431 can also be achieved by the heterogeneous photocatalysis. The overall process can be described by the following two equations: C,H,O,

x + ( m+ 5 - 2-2)o2 2 4 2

Semiconduaor, hv

,mCO;?+-n-kz H20+kzH' 2

M"' + n H 2 0 2

Semiconduaor, hv

+zX0Fk

> M o+nH' + L O 2 4

... (2) ... (3)

Basic Principles of Heterogeneous Photocatalysis The term photocatalysis consists of the combination of photochemistry and catalysis and thus implies that light and catalyst are necessary to bring about or to accelerate a chemical transformation [44,451. The catalyst used are inevitably semiconductors, which can act as catalyst due to their specific electronic structure characterized by a filled valence band and an empty conduction band [46]. The energy gap, called bandgap energy, between the conduction band and the valence band is relatively small usually in the order of a few electron-volts. A photon with sufficient energy can activate the catalys by promoting an electron from the filled valence band into the conduction band. The occurrence of this charge separation is one of the first essential

510


steps of a photocatalytic reaction. This process is illustrated in Figure 2 and is represented by Equation (4):

TiO,

hV>E,,#

... (4)

,e, + h i

R'

0 2

Minerals

kox.3

A

OH'ads

RHads

o2

___t

RLJ

Minerals

kox.3

Figure 2. Schematic illustration of generation of electron hole pairs in a spherical semiconductor particle together with some of the consecutive redox reactions that take place.

511

D. Chen, M.Sivakumar and A.K. Ray

As the conduction band is only partially filled, the electron remains free to move

through the semiconductor lattice. The resulting vacancy in the now partially filled valence band is also free to move. This vacancy, or absence of an electron, is usually referred to as hole, designated by h;. The photogenerated electron-hole pairs can subsequently involve in two different reactions: (a) recombination of electrons and holes and subsequent dissipation of the adsorbed energy with the liberation of heat, and (b) reaction with electron donor or acceptor giving rise to oxidation and reduction processes, respectively.

-1.0

I

C

A I

0.0

Ew

1.0

r

w

2.0

3.0

Figure 3. Valence and conductance band positions of semiconductors at pH = 0. Figure 3 illustrates the valence and conductance band positions of various semiconductors, and relevant redox couples. The position of conduction and valence bands of a semiconductor determines the reaction pathway. For example, photooxidation of water in presence of MoS2 (bandgap 1.8 eV) is possible thermodynamically but is not possible in the presence of CdSe (bandgap 1.70 eV) even though their bandgap energies are almost same. Instead, photo-reduction of water would take place more easily in the presence of CdSe, due to the position of its conduction band. Of course, more than one couple can be involved in a photocatalytic

512


process. Electrons and holes can reduce (and/or oxidize) species belonging to different redox couples contemporaneously present in the reacting system. Hence, in order for a semiconductor particle to be photochemically active as a catalyst for both reactions (2) and (3), the redox potential of the photogenerated valence band hole must be sufficiently positive to generate absorbed OH' radical (which is the powerful oxidant for subsequent oxidation of organic pollutants), and the redox potential of the photogenerated conductance band electron must be sufficiently negative to be able to reduce the adsorbed oxygen to superoxide (or reduce the adsorbed metal ions, if there is any). Figure 2 illustrates these reactions on the surface of a semiconductor, the recombination process of the electron-hole pair, and some of the subsequent reactions. The concentration of electron-hole pairs in a semiconductor particle is dependent on the intensity of the incident light, and the material's electronic characteristics that prevent them from recombining and releasing the absorbed energy. Therefore, photocatalytic technology differs from other water treatment methods in that it involves both oxidation and reduction chemistry. Reduction is mainly useful

for removing dissolved toxic metal ions from water whereas the oxidative property is employed to mineralize dissolved toxic organic compounds. It should be emphasized that oxidation and reduction must occur simultaneously to continue the process activity. Organic compounds and water are two potential reductants, while oxygen and metal ion are two potential oxidants in photocatalytic process. Oxidation and reduction kinetics is interdependent, and either process can be rate-controlling 1471. There are a variety of semiconducting materials which are commercially available and investigated in literature as photocatalysts, such as TiOz, ZnO, ZnS,CdS, Fe203, W03, etc. [48-52]. However, only a few of them were found out to be suitable for

efficiently catalyzing reactions in Equations (2) and (3) for a wide range of organic and inorganic compounds. Table 1 lists the semiconductor photocatalysts commonly used, their bandgap energy, and threshold wavelength as reported in literature. The threshold wavelength is calculated using Equation (5):

Abg (nm>=124o/Ebg(eV)

... ( 5 )

However, the bandgap energies listed in Table 1 may be altered when the semiconductor surface is in contact with an electrolyte solution [14, 531. Lower

513

D. Chen, M. Sivakumar and A.K. Ray

bandgap energy semiconductors are preferred when catalyst activation with solar energy is desired. However, low bandgap semiconductors, especially non-oxides such as CdS and CdSe, usually suffer serious stability problem [54]. A desired photocatalyst must posses: (a) high activity; (b) ability to utilize visible and/or nearW light; (c) photostable (durable) and reusable; (d) chemically and biologically inert; and (e) cheap. Of all the different semiconductor photocatalysts tested successfully in laboratory studies. Ti02 appears the most active and is extensively used in both laboratory studies and pilot plants [48, 55, 561. Ti02 is cheap, insoluble under most conditions, photostable, and non-toxic. Moreover, sunlight (about 3% of the solar spectrum contains W - A ) can be used as a possible light source for Ti02 (absorbs light in the UV-A range, h 5388 nm) activation [57-591. Table 1. Bandgap energies (at pH = 0) and corresponding threshold wavelengths of various semiconductors [ I 61.

Semiconductor Ti02 ZnO ZnS CdS FQ.03

wo3

Bandgap (eV) 3.0-3.2 3.2 3.6 2.4 2.3 2.8

Wavelength (nm) 4 13-388 388 335 516 539 443

Titanium dioxide exists primarily in two crystalline forms, anatase and rutile. In many studies, anatase was found more effective than m i l e for the photocatalytic oxidation of acid/acetate mixture [60],phenol [61, 621, cyclohexane and 2-propanol [63]. The anatase form has a reasonably well-defined nature, typically 70 to 30

anatase to rutile mixture, non-porous with BET surface area of 55+15 m2/g and average particle size of 30 nm. It shows substantially higher photocatalytic activity than most other readily available samples of TiO2.

Mechanism of the Photocatalytic Reaction Although Equations (2) and (3) represent the overall reactions, individual steps that drive above reactions are still not well understood. The photocatalytic process can be referred to as “four-phase” system [64]. The fourth phase being the electronic phase (W light source), in addition to the three phases - liquid (water), solid (catalyst) and

514


gas (the oxidant, oxygen). The basic phenomena that take place in a semiconductor particle in aqueous solution irradiated with ultraviolet light of appropriate wavelength are schematically shown in Figure 2. When a semiconductor is illuminated by light of suitable energy, namely higher than its bandgap energy Ebg,electrons are promoted from the valence band to the conduction band, and consequently, the positive holes are created in the valence band, which is represented by Equation (4). The electronhole generation is one of the first essential steps of a photocatalytic reaction. While in contact with a liquid solution, a charge transfer occurs until an electrostatic equilibrium is achieved. The excess of charge in the semiconductor is not confined on the surface as occurs in a metal, but is distributed in a region called the 'space-chargeregion' expanding from the surface to the bulk of the semiconductor. This spacecharge-region plays a significant role in photocatalytic process. In the presence of the surface charge region, the photo-excited electron-hole pair can be separated, and then -

migrate ( kJ# , k ,

h+

) towards the particle surface (Equations 5 and 6) as a result of a

potential gradient that exists between the bulk solid and its external surface. This potential gradient is caused by depletion of conduction band electrons in the spacecharge-region [ 19, 65, 661. Electro-neutrality of the surface requires equal arrival rates of the electrons and holes at steady state [67]. The fate and dynamics of the photogenerated electron-hole pairs are of considerably importance in the subsequent degradation processes. Generally, the mobile electors and holes have two basic destinies. They may recombine (k,)

and dissipate the inputted energy as heat

(Equation 7) during their transport to the particle surface. They can also react with the electron acceptor and electron donor on the particle surface after migration to the catalyst surface, and subsequently, initiate the reduction and oxidation processes, respectively (Equations 8 and 9).

e,

+ hv+b

krec

>heat

... (8)

51.5


A + e,

kred

9 A'-

... (9)

Reaction (7) is one of the main causes for the low quantum efficiencies of photocatalytic processes, and the recombination reaction of electrons and holes occurs mainly in the bulk of the catalytic particle [68]. This recombination possibility can be reduced if these mobile species are separated, and subsequently, trapped by surface adsorbates or other sites. After a photogenerated hole reaches the surface of the semiconductor, it can react with an adsorbed substrate by interfacial electron transfer, assuming that the adsorbate possesses a redox potential appropriate for a thermodynamically allowed reaction. Thus an adsorbed electron donor can be oxidized by transfemng an electron to a hole on the surface, and an adsorbed acceptor can be reduced by accepting an electron from the surface. Hole trapping generates a and electron trapping generates an anion radical, A*-. These radical cation radical Do+, ions can participate in several reaction pathways: (i) they may react chemically with themselves or other adsorbates; (ii) they may recombine by back electron transfer to form an excited state of one of the reactants, or to waste the excitation energy by the non-radiative pathway; (iii) they may diffuse from the semiconductor surface, and participate in chemical reaction in the bulk solution. If the rate of formation of D" is kinetically competitive with the rate of back electron transfer, photoinduced oxidation will occur for any molecule with an oxidation potential less positive than the semiconductor valence band edge, since under these conditions interfacial electron transfer at the illuminated interface is thermodynamically possible. For similar considerations, the photoinduced reduction can occur to any molecule possessing a reduction potential less negative than the conduction band edge [ 191. It is well known that the surface of TiOz is readily hydroxylated when it is in contact with water. Both dissociated and molecular water is bonded to the surface of TiOz. Surface coverage of 7-10 OR/nm2 at room temperature were reported in literature [69, 701. The reactions of the trapped holes react with absorbed solvent molecules (H20 and O H ) have been experimentally observed [32,71-731:

516


kox, 1

G

+ H20ads

hi

+ OHSs

kox, 1

>OH>s

+H+

,OH23

... (11) ... (12)

The trapped holes can also react with absorbed organic substrates RX: h:

+ J?X,

... (13)

*RXzs

Reactions (10) and (11) is more important in oxidative processes due to the high concentration of the absorbed H2O and O H molecules on catalyst particle surface. Experiments conducted in water-free, aerated organic solvents have shown that only partial oxidation can be achieved "731. However, in aqueous solutions complete mineralization of numerous organic compounds has been observed [27-29, 74-76]. Apparently, direct reaction between the organics and the valence band holes (Equation 12) is not significant [77]. The result also implies that the presence of water

or hydroxyl group is necessary for complete destruction of organic compounds. Photogenerated electron must also react to avoid a continuous charge build-up in catalyst particles. The accumulation of photogenerated electrons on Ti02 particles was experimentally observed during the photoassisted oxidation of 1.6M aqueous methanol [77]. At steady state the rate of hole consumption must be equal to the rate

of electron consumption. Therefore, electron scavenger (or acceptor) must be present in the photocatalytic process. Oxygen is the commonly used electron acceptor as it is

available at little or no cost. It reacts with photo-generated electron at the surface of

--

semiconductor through following equations [60, 61,73,78]:

e,; + 0 2 , udc

H'

*

+ O;;&

2HO;

0:i

... (14)

HO;

... (15)

H 2 0 , + 0,

HO; + O;;ab

-

0, + HO;

HO;+H'-H,O, H,O,

H,+ 0,

H,O,+HO~~OH*+H,O+O,

... (16) ... (17)

... (18) ... (19)

... (20)

517

D.Chen, M.Sivakumar and A.K.Ray

Formation of these species has been experimentally verified by electron spin resonance measurements [71]. Auguliaro et al. [79] also confirmed the formation of

H202 in the degradation of phenol in Ti02 suspensions by aerating the solution with O2 and He respectively. Oxygen-involved species (e.g., HO;,

HO;, 02'-, O2 and

H202) are present either at the interface or in the solution, and can participate in the complex degradation scheme, which leads to the final mineralization of the organic species. It should be noted that it takes three electrons to make one hydroxyl radical through this pathway, but it takes only one hole to produce a hydroxyl radical in the other half-cell reaction, Equations (10) and (1 1). Consequently, most hydroxyl radicals are generated through the hole reaction [61, 801. The hydrogen peroxide formed from reactions (13) to (15) may further decompose to form hydroxyl radicals [61, 79, 811:

... (21)

H20,*20H' H,O,

+ 0;- -OH

+ OH-+ 0,

H,O,+e,--+OH'+OH-

... (22) ... (23)

It has been shown that the addition of hydrogen peroxide considerably improves the photocatalytic process [36, 79, 82-86], most probably through reactions (20) to (22). It has been proposed that the highly reactive hydroxyl radicals formed in above equations are the primary oxidizing species in photocatalytic processes by attacking the organic species and the rate-determining reaction step may be the formation of hydroxyl radicals [32, 65, 731. This is supported by the following facts: (a) Complete mineralization of organic compounds cannot be achieved in water-free organic solvents [73], however it is possible in aqueous solutions; (b) Steady-state OH' concentrations in W-irradiated titanium dioxide aqueous solutions could be as high as lo-' M [87] much higher than those in the processes of ozonation, direct photolysis of H202,and radiolysis (i.e., c 10l2M); (c) During the photocatalytic degradation of aromatic compounds, the detected intermediates typically have hydroxylated structures [32, 61, 621. These intermediates are consistent with those found when similar aromatics are reacted with a known source of hydroxyl radicals [88. 891.

518


Hydroxyl radical is a very reactive species with an unpaired electron. It reacts rapidly and non-selectively in the oxidation of organic compounds. Due to its very high oxidation potential (see Table 2), it is capable of oxidizing almost all organic pollutants in wastewater by hydrogen abstraction. The reactions may occur either in the bulk solution or on the catalyst surface. Based on hydroxyl radical attack ( kox, )

of organic compound in photocatalytic degradation, Turchi and Ollis [32] proposed four different reaction pathways: Reaction occurs while both species are adsorbed.

OHh + ' , , a h

-';.ads

. .. (24)

A non-bound radical reacts with an adsorbed organic species.

OH

+

4.d

-';,ah

... (25)

An adsorbed radical reacts with a free organic species arriving at the catalyst surface.

OH;

+ R, dR;

-

... (26)

Reaction occurs between two free species in the bulk solution.

OH'

+ R,

R;

... (27)

Based on the above different reaction pathways, the four kinetic models developed had similar degradation rate forms and were consistent with reported initial data and temporal degradation data. However, pulse radiolysis and time-resolved diffusion reflectance measurements showed that surface-bound hydroxyl radicals are more stable compared to those in the bulk solution [90].Similar results were obtained by Minero et al. [91] who found that the degradation process occurs at the surface or within a few monolayers around the photocatalytic particles. These results imply that the main reaction of the photocatalytic degradation process takes place on the surface of the catalyst. The organic radicals formed in above reactions can be oxidized ( kox, ) into peroxy radical by addition of molecular oxygen. These intermediates initiate thermal (chain) reactions leading to COz, H 2 0 and mineral acid:

5-19

D.Chen, M. Sivakumar and A.K. Ray

RLfds + 0 2

RO;

++ C02 -i- H20 iMineral Acid

... (28)

Table 2. Oxidation Potentials of Some Oxidants 161. Species fluorine hydroxyl radical atomic oxygen ozone hydrogen peroxide perhydroxyl radical permanganate hypobromous acid clorine dioxide hypochlorous acid hypoiodous acid chlorine bromine iodine

Oxidation Potential (V) 3.03 2.80 2.42 2.07 1.78 1.70 1.68 1.59 1.57 1.49 1.45 1.36 1.09 0.54

Table 3. Reaction rate constants of hydroxyl radical with classes of organic contaminants. Compounds Chlorinated alkenes Phenols N-containing organics Aromatics Ketones Alcohols Alkanes

Reaction rate constant (VmoVs) 109- 10' 109- 10" lo9-lo1' 1oE-1O'O 109-1o10 108-109 lo6- lo9

Table 3 presents a summary of the classes of organic contaminants typically treatable by advanced oxidation process (AOP) technology, as well as their associated reaction rate constants relative to oxidation by hydroxyl radicals. The data indicate that chlorinated alkenes, phenols and nitrogen-containing organics exhibit the highest rate constants principally due to the presence of double bounds within the molecular chain and other characteristics that make theses compounds susceptible to oxidation. During photocatalytic oxidation of organic compounds, oxygen is reduced to superoxide and/or perhydroxyl radicals. These radicals eventually form hydroxyl radicals that enter into the oxidation cycle [3, 19, 201. Besides oxygen, any dissolved metal ions with a reduction potential more positive than the conduction band of the photocatalyst can, in principle, consume electrons and complete the redox cycle. This

520


usually deposits a reduced form of the metal on the catalyst surface, which may or may not be easily removed [92, 931. Figure 4 illustrates the positions of valence and conduction bands of anatase Ti02 photocatalyst in contact with an aqueous electrolyte solution at different pH and compares them with the reduction potentials of environmentally fretful metal ions. It is worth noticing that the positions of both conduction and valence bands are pH dependent. The increase of pH in electrolyte solution makes the positions of the valence and conduction bands to shift to more cathodic potentials by 59mV per pH unit [47]. However, reduction potentials of all metal ions (except for Cr(V1)) illustrated in Figure 4 are independent of pH, thereby resulting in photoreduction of these metal ions more favourable with increasing pH. On the basis of Figure 4, one can note that Cd2+, Fez+ and Cr3+ cannot be photocatalytically reduced because their reduction potentials are more negative than that of electrons. However, photocatalytic reduction is thermodynamically feasible for Au3+, Cr", Hg2+(including HgC12, HgCb'.), Ag', Hg?,

Fe3+,Cu' and Cu2+.Among

the above metal ions, Fe3+and Cr6+can only be reduced to Fe2+and Cr3+,rather than iron and chromium metals respectively, because Fez+and Cr3' cannot be reduced further as mentioned above. It is also unlikely for Pb2+and Ni2' to be reduced under most conditions due to the extremely low driving force. One must also note that the reduction potential of a redox couple is concentration dependent. According to Figure 4, one may also expect that oxygen can be reduced preferentially with respect to most metal ions if it is present in solution. That is why the reaction system in most studies was purged with nitrogen or sealed to eliminate oxygen reduction, thereby increasing the photoreduction efficiency of metal ions. However, the photocatalytic reduction rate or to what extent they can be removed for those metal ions that can be reduced thermodynamically at given illumination time is determined by the reduction kinetics. Sometimes, no measurable reduction is observed due to the very low reduction rate even though the reduction reaction is thermodynamically feasible.

Photocatalytic Kinetics Unlike other AOPs that are based on the use of additional chemical reagents, the kinetic behavior for chemical transformations in photocatalysis usually follows a saturation behavior [94]. Therefore, typical applications of photocatalysis are at low

521


level of pollutant concentration. The kinetic modeling of the primary steps is required for any practical application of the process and some kinetic models describing photocatalytic oxidation on illuminated semiconductors have been proposed. 3.2 -1-

2.8 I-

Qz 2.4

1.2

0.8

Potential (V) 0.4

0

-0.4

21

-0.8

0

I

I

I

I

I

I

1

2

3

4

5

6

7

UH

Figure 4. Positions of valence and conduction bands of Ti02 (anatase) and the reduction potentials of metallic ions of interest at different p H .

522


The degradation kmetic rate expressions in literature have focused on the initial disappearance rate of organic compounds [e.g. 78, 961 or the initial formation rate of

C 0 2 [e.g. 33, 54, 971. The initial photocatalytic reaction rate usually follows Langmuir-Hinshelwood equation with respect to initial organic concentration [e.g. 27-33, 62,94-97]:

ro =

kKCo

.. (29)

1+ KC,

The above relationship describes a zero-order rate at high concentration while a first order reaction rate at low organic concentration. Figure 5 illustrates the typical laboratory experimental set-up for photocatalytic kinetic studies. Usually, experiments are carried out in a semi-batch mode, which mainly consists of photoreactor, UV light source, pump, well-mixed reservoir and numerous measuring meters. Adjustments can be easily performed in the reservoir for sampling and suspension chemistry monitoring.

a W lamp

TiOz suspension

lol Magnetic stirrer

Figure 5. Typical experimental setup for photocatalytic kinetic studies in laboratory.

523

D.Chen, M.Sivakumr andA.K. Ray

Mass Transfer in Photocatalytic Processes In heterogeneous catalysis, both internal and external mass transfer of the reactants and products affect the overall rate of degradation. Relationship among the observed degradation rate, kobs,the external and internal mass transfer rates, km,e,, and km,int,and are given by the following expression: the intrinsic kinetic rate, kn,

1

1

km,ext

km,int

1 - 1 ---+-+kobs

krxn

... (30)

Therefore, the apparent rate of photocatalytic reactions can be either be reaction rate controlled or mass transfer controlled. Photocatalytic processes are generally operated in two modes where catalyst is either suspended or immobilized. Although Degussa P25 TiOz is non-porous and its elementary particle size is only about 30 nm, actual catalyst particle size of up to 5 pm in diameter is reported [98] due to the aggregation of the elementary particles when suspended in aqueous solution. Chen and Ray [99] investigated the mass transfer effect in P25 TiOz suspensions using phenol, 4-chlorophenol and 4nitrophenol as model compounds. The results indicated that ratios of surface concentration to the bulk concentration were close to unity, and the minimum effectiveness factors of actual catalyst particles were above 0.9. Therefore, in TiOz suspensions both internal and external mass transfer resistance may be neglected and chemical reaction is the rate determining (controlling) step. However, effect of mass can not be ignored when catalyst is immobilized. Observed photocatalytic degradation rate increases with increasing flowrate of reactants as reported [95, 1001 in both straight and coiled tube photoreactor, where catalyst was coated on the inside surface of the tube. Scalfani et al. [loll found that mass transfer was the main rate-limiting factor in photodegradation of phenol in a continuous annular photoreactor in which Ti02 beads were used. Ray and Beenackers [lo21 investigated the mass transfer in a swirl-flow monolithic-type photoreactor by measuring the dissolution rate of benzoic acid into water flowing at different Reynolds number. The mass transfer coefficient was correlated by: k,

524

= 6.1~10-'

... (31)


When the thickness of catalyst layer (film) is increased, internal mass transfer may also play a significant role in photocatalytic reactions. This influence was determined by Chen et al. [I031 experimentally by conducting dynamic physical adsorption of benzoic acid on P25 Ti02 catalyst film at different catalyst layer thickness. They ' ~ by fitting the experimental reported as effective diffusivity value of ~ . O X I O m2/s

results with a realistic dynamic physical adsorption model.

Kinetic Equations Many researchers have demonstrated that variety of pollutants can be effectively degraded using Ti02 as catalyst. However, the kinetics and mechanism of photomineralization processes are still comparatively ambiguous. Several kinetic models [e.g., 27, 32, 33, 54, 72, 75, 79, 80, 95, 104-1061 for photocatalytic oxidation of organic compounds have been published in literature. Among these reaction models, Langmuir-Hinshelwood type model was the most popular in describing the photocatalytic oxidation, which was further divided into competitive and noncompetitive type adsorption models [32, 1071. The competitive model described that both oxygen and organic compound adsorbed on the same active site of catalyst competitively, while the non-competitive model specify that the organic compound and oxygen adsorbed separately on the different active sites of the catalyst. Turchi and Ollis [32] proposed that the non-competitive model was favored for the photocatalytic oxidation using Ti02 as catalyst. Matthews [33, 75, 95, 1081 studied the photocatalytic oxidation of organic substrates over TiOz, Okamoto [go] investigated the photodecomposition of phenol, Pruden and Ollis [27] studied the photodegradation of trichloroethylene in water, Chen and Chou [ 1051 investigated the photodecolorization of methyl orange, and Mills and Morris [54] studied photooxidation of 4-CP. All of these researchers have pointed out that the Langmuir adsorption model could be applied to their systems as summarized in Table 4. To date, very few photocatalytic kinetics of Ti02 on immobilized system have been published. Sabate et al. [ 1061 investigated the photocatalytic degradation of 3chlorosalicylic acid over Ti02 film supported on glass. In their experiments, they varied the concentrations of organic compound and dissolved oxygen, but no

525

moUmin.L

Semi-batch Batch Batch Semi-batch

Fixed Slurry Slurry Fixed Slurry

CP

4-CP

Trichloroethylene

3-Chlorosalicylicacid

Chloroform, etc

105

I40

56

mg/min.L

moVm2.s

rn0Vs.L

Batch Batch Batch Batch

Slurry Slurry Slurry Slurry

Methyl orange

4-Chlorophenol'

4-nitrophenol

Phenol

19

80

m0Vmin.L

batch

32

106

21

104

95

33, I5

Ref. 55

Slurry

Kinetic model kKi IS1

Phenol

mmoUmin

m0Vmin.L

pmoUmin.g.c at

mmol/h.L

mglmin.L

Semi-batch

Slurry

Rate unit molh

Benzoic acid, etc*

Operating mode Batch (r0)

Catalyst Fixed

Pollutant Oxalic acid

Table 4. Survey of typical kinetic models proposed in literature

9

a P P

P


influence of other parameters, such as catalyst loading, light intensity and reaction temperature was reported. Ray and Beenackers [lo21 performed a detailed study for degradation of a textile dye on immobilized system. They designed a novel kinetic reactor using which they determined lunetic rate constants as a function of wavelength of light intensity and angle of incidence, catalyst layer thickness, and the effect of absorption of light by liquid film on the overall photocatalytic degradation. Unfortunately, there are two main disadvantages in the reported kinetic equations. Firstly, most of these equations either describe the relation between inirial degradation rate and initial concentration of organic species, or used the kinetic parameters obtained from the initial rate to predict the entire process. Therefore, good agreement can be obtained only for those systems in which intermediates are not formed during photocatalytic degradation such as PCE [31]. However, for most photocatalytic processes, such as the photodegradation of halogenated aromatics, above kinetic equations cannot predict the entire process due to the formation and influence of the intermediates. Furthermore, initial rate data are tedious to obtain, and prone to large errors, and thereby reduces reliability on the results. Secondly, in heterogeneous catalysis, it is customary to report the reaction rates in units of per gram of catalyst. However, heterogeneous photocatalysis is quite different from the ordinary heterogeneous catalytic process due to the presence of W light. Catalyst at different positions in a reactor has different contributions to the reaction rate due to the light distribution within the reactor. In Ti02 aqueous suspensions, the catalyst that is far from the light source has less or even no contribution to the reaction rate. When TiOz is immobilized onto a substrate, not all the catalyst particles take part in the reaction. Obviously, in heterogeneous photocatalytic reaction, it is not reasonable to define the unit of rate as the same as those commensurate with general homogeneous or heterogeneous phase kinetics. However, the definition of reaction rate reported in the literature is either in “moVs” or “moWs” [33, 74, 78, 79, 96, 105, 109, 1101. Since both are dependent on the illuminated catalyst surface area through the reactor window and volume of reaction solution, the reported results are in general not meaningful, and moreover, render the comparison between investigations conducted

in different research groups impossible. In order to avoid above problems, Chen and

527

D.Chen, M.Sivakumar and A. K.Ray

Ray [56, 991 developed new photocatalytic kinetic expressions. Their model correlated as many kinetic parameters as possible. Reaction rate was defined as the mole reduction of pollutant per irradiated reactor window area. So, it was independent of either the illuminated area or the volume of the reaction solution.

Kinetic Parameters Initial Concentration of organic compounds Influence of initial concentration of organic compounds on the photocatalytic degradation rate has been extensively studied and the so-called "saturation behaviour" was observed by many researchers [49,54,94,96, 105, 109, 1111. Langmuir-type rate expression (Equation 28) has been used successfully to describe photocatalytic degradation. However, the mechanism of this effect is still not clear. We believe that three factors might be responsible for the saturation behaviour during the photocatalytic reaction. (a) One of the main steps in the degradation process take place on the surface of the catalyst, and therefore, adsorption of organic compounds onto the catalyst surface affects the reaction, and usually high adsorption capacity favours the reaction. For most organic compounds adsorption capacities on TiOz catalyst are well described by Langmuir-type equation. This has also been confirmed in our laboratory study [59]. It means that at high initial concentration all accessible catalytic sites are occupied. A further increase in pollutant concentration does not increase concentration of pollutant at the catalyst surface. (b) In photocatalytic processes, generation and migration of the photogenerated electron-hole pair, and the reaction between photogenerated hole (or hydroxyl radical) and organic compound are two processes in series. Hence, each step may become rate determining for the entire process. At low organic concentration the latter may dominate the process and, therefore, the degradation rate increases linearly with the concentration. On the contrary, at high organic concentration the former will become the governing step and the degradation rate increases slowly with concentration, and even a constant degradation rate may be observed at higher concentration under a given illuminating light intensity. Gerischer and Heller [81] and Wang et al. [77] provided substantial evidence indicating that the interfacial electron transfer process involving the reduction of oxygen was the rate determining step in the TiOz sensitized 528


photodegradation of organics. (c) Intermediates formed during photocatalytic process also affect the rate constant of their parent compounds. For example, consider two experimental runs with different initial concentrations CIand C2 (Cl > C2). When C, decreases to C2. some intermediates will be formed and subsequently will adsorb competitiveIy on the solid catalyst surface. If the two experimental runs are now compared starting with C1, the adsorption of intermediates will reduce the effective concentration of the parent compound on the catalyst surface for first case over the second. Hence, the observed rate for the first case will be lower.

Catalyst Dosage In the photocatalytic oxidation of organic compounds in aqueous suspensions of Ti02, the mass concentration of the latter is an important process parameter affecting the observed reaction rates [27, 28, 105, 112, 1131. Most commonly, 0.1-0.2 wt% TiOz slurry are used [36, 50, 75, 1141. In the studies by Ollis and co-workers [36], a 0.1 wt% Ti02 slurry was used to photocatalytically oxidize a variety of halogenated organic compounds in a batch reactor [27-301. Okamoto et al. [61, 801 studied among other variables the effect of Ti02 concentration in the photooxidation of phenol in aqueous suspensions and found that at high light intensities the reaction rate increased with the Ti02 concentration in the solution. Similar studies showed that an optimal concentration range for Ti02 (1-3 g/L)exist for maximum removal of phenol 62. 1151. This optimum concentration range depends on the reactor geometry, intensity of radiation source, and the properties of Ti02 such as particle size, phase composition and impurities. Increasing catalyst concentration beyond the optimum range may result in reduced photon flux caused by a shielding effect of the particles on the light, although the catalyst surface area per unit volume of solution increases [lo, 62, 1151. This shielding effect has been studied in a stirred photoreactor [ 1161. Table 5 lists the reported research results on catalyst dosage used. Obviously, the optimal catalyst dosage reported was in a wide range from 0.15 to 8 g L for different photocatalyzed systems and photoreactors. Even for the same catalyst (Degussa P25), a large difference in optimal catalyst dosage (from 0.15 to 2.5 g/L) was reported. Chen and Ray [99] proposed the following equation that describes the influence of catalyst dosage on the photocatalytic degradation rate:

529

D. Chen, M.Sivakumar and A.K.Ray

... (32)

ri = K[1-exp(+BCCatH)]

where E is the light adsorption coefficient of the reaction system; p is the exponential term of light intensity influence and its value is between 0.5 and 1.O depending on the light intensity; H is solution thickness in light transmission direction. Equation (31) can successfully explain the dependence of degradation rate on the Ti02 dosage at low and high light intensities reported in literature [80].Degradation rate is usually proportional to between p5and I].' at high and low intensities, respectively [ 18, 36, 94, 1021. This indicates that the

p

value is 0.5 at high intensity and 1.0 at low

intensity. Thus, according to Equation (31) the optimal catalyst dosage at high light intensity is higher than that at low light intensity. Experiments conducted under different solution thickness indicated that effective optical penetration length was only a few centimeters in 0.2% Ti02 suspensions [99]. Table 5. List of optimal Ti02catalyst dosage reported in literature. Pollutant(s)

Catalyst

2- and 3-chlorophenol Phenol 4-chlorophenol Nitrophenols Methyl orange 4-chlorophenol Methylene blue Dimethylphenols Malonic acid 2-chlorophenol Cyanides 4-nitrophenol

Degussa P25 Merck Degussa P25 BDH Merck Degussa P25 Degussa P25 Degussa P25 Degussa P25 CERAC Home prepared Degussa P25

Optimal dosage (g/L) 2.5 2.5 2.0 1.o 8 .O 0.5 0.15 1.o 0.8 2.0 2.0 2.0

Ref. 94 115 50 49 105 140 132 112 96 114 113 56

When Ti02 catalyst is immobilized on supports, there exists an optimum thickness for the catalyst film. There are two types of catalyst illumination [102]. In the first

arrangement (SC illumination), UV light is introduced from the catalyst support (substrate) side, while in the second arrangement light is introduced from the liquid side (LC illumination). In SC illumination, catalyst support must be optically transparent, such as quartz and Pyrex glass. The advantage of this arrangement is that UV light doesn't pass through the reaction solution before reaching the catalyst film,

530

HeterogeneousPhotocatalysis in Environmental Remediation

and therefore, light intensity will not be attenuated due to the absorption by reaction medium. Influence of catalyst film thickness, h, on the photocatalytic reaction can be described by following equation [ 1031:

... (33)

and optimum catalyst layer thickness is given by

hop,=

wffpDe 1 k, )

... (34)

ffp--k , De

In the LC illumination case, the effect of catalyst film thickness on the degradation rate is given by [ 1031:

Obviously photocatalytic reaction rate reaches a saturation value as the catalyst layer thickness increases. However, if the film is too thick, the strength of catalyst adhesion is poor and is likely be detached from its supports. Catalyst film with less than 6pm thickness was widely used in laboratory studies [117-1211, because experimental result indicated that 95% of the incident light intensity has been absorbed by the catalyst film with thickness of 4 . 8 of~ P25 Ti02 [ 1031. Dissolved oxygen According to the principles of photocatalytic reaction, the rate of oxidation by holes has to be balanced by the reduction rate of the photogenerated electrons. Therefore, either the oxidation or reduction reaction can be the rate-governing step. Accumulated electrons may also result in increasing rate of electron-hole recombination step, thereby, decreasing in the hole-initiated oxidation of organic molecules. Thus, the excess in photogenerated electron must be removed in order to avoid recombination. Hence, electron scavenger (usually dissolved oxygen) is necessary. It was found [77]

531


that photocatalytic activity nearly completely suppressed in the absence of dissolved oxygen, and the steady-state concentration of dissolved oxygen had a profound effect on the rate of photocatalyzed decomposition of organic compounds. Alberici and Jardim [ 1221 found that decomposition of phenol in non-aerated solutions containing Ti02 was much slower compared with aerated ones. Sabate et al. [lo61 did not observe photocatalytic degradation of 3-chlorosalicylic acid when pure N2 was bubbled through the solution. Hsiao et al. [29] found that the observed rate of disappearance of dichloro- and trichloromethane was much faster when the solution was well purged with oxygen. Sclafani et al. [83] found that the concentration decay of phenol was much faster when oxygen was bubbled in the solution containing

anatase Ti02 instead of He. Similarly, Augugliaro et al. [62, 1161, found that by increasing the partial pressure of oxygen in the He-02 mixture bubbled through a phenol solution containing T i 4 catalyst, the reaction rate constant increased and remained constant for poZ>0.6. It is important, therefore, to provide sufficient oxygen in the organic solution containing the Ti02 catalyst to prevent accumulation of electrons on the TiOz surface. The major role of oxygen in photocatalytic degradation of organic compounds is to act as a scavenger for continuous electron removal from photocatalysts [1231. Gerischer and Heller [81] studied in detail the role of oxygen in the photo destruction of organics on catalyst surfaces. They developed kinetic models to predict the maximum electron uptake by oxygen and found that the latter depends on catalyst particle sizes and oxygen concentration in the solution. Early studies have shown that the adsorption of oxygen on illuminated Ti02 surface depends on the number of hydroxyl groups of the surface [124-1261. The dependence of degradation rate constants of organics on the dissolved oxygen concentration can be well described by the Langmuir-Hinshelwood equation (Table 6 summarizes the

K02

values reported in

literature):

However, in photocatalytic removal of dissolved metal ions, dissolved oxygen is a strong competitor for metal ions to scavenge the photogenerated electrons. So, the

532

Heterogeneous Photocatalysis in EnvironmentalRemediation

presence of dissolved oxygen may significantly inhibit the photoreduction of metal ions [37,47, 127, 1281.

Table 6. Adsorption constant of dissolved oxygen on TiOl catalyst. Pollutant

Catalyst

KO,(atm-')

Ref.

4-CP

Degussa P25

10.5

93

4-CP"

Degussa P25

4.4

140

Phenol

Wako Chem.

11.1

61

Methyl orange

Merck

4.2

105

4-CP

Degussa P25

17.6

99

4-NP

Degussa P25

9.98

99

Phenol

Degussa P25

12.7

99

UV Light Intensity The incident light intensity determines the photogenerated electrodhole pair concentration, and thus the hydroxyl radical formation rate. Therefore, at sufficiently low levels of illumination (catalyst dependent), degradation rate is of first-order in intensity. At high intensity level, on the other hand, the reaction rate increases with the square root of intensity because of the increasing electrodhole pair recombination (Equation 7) during their migration to the particle surface and recombination of hydroxyl radicals resulting from their high concentrations.

OH' +OH' + H202

. .. (37)

As a consequence, this is obviously detrimental to the photocatalytic process as it

results in the decrease in quantum efficiency. The optimal light power utilisation should be in the domain where degradation rate is proportional to the incident light intensity. Several previous studies [35, 50, 941 have demonstrated above conclusion. The transition points between these regimes, however, will vary with the photosystem. In the degradation of 3-CP in TiOz aqueous suspensions, the initial degradation rate was reported to be proportional to radiant flux for values smaller than 20 mW/cm2, while above this value it was proportional to its square root [94]. The initial degradation rate of 4-CP against radiant flux was also investigated in the range

533


of 2-50 mW/cm2 [50], where similar rate shift from first-order to half-order in

intensity was observed.

Table 7. Photocatalytic activation energies (E) reported in literature. Reactant 4-CP 4-CP 4-CP 4-CP

Methyl orange Malonic acid phenol phenol Salicylic acid 2-propanol Oxalic acid Formic acid xylenols

Catalyst Fixed Slurry Slurry Slurry Slurry Slurry Slurry Slurry Slurry Slurry Slurry Slurry

Slurry

4-Np

Slurry

Benzoic acid

Slurry Slurry Fixed

SBS

Methylene blue

E (kJ/mol) 20.6 16 5.5 13.7 18 9.99 10 11.8 11 31 13 16.7 8.8 1.42 9.13 13.4 60.6

T (“1

Ref.

10-60 15-55

119 18 50 99 105 96 80 99 95 67 55 150 112 99 59 110 118

10-45

15-50 27-46 21-51 20-50 15-50 25-45 4-40 5-70 6-60 15-50 15-50 25-38 16-52

Reaction Temperature In photocatalysis, irradiation is the primary source of electron-hole pair generation at ambient temperature as the band-gap energy is too high to be overcome by thermal activation. The true activation energy should be very small or even zero. Thus, the overall photocatalytic reaction (Equation 2) is usually found to be temperature insensitive. Increasing the reaction temperature may increase the oxidation rate of organic compounds at the interface. The dependence of degradation rate on temperature is reflected by the magnitude of activation energy. Table 7 lists the apparent activation energies of some organic compounds in reaction photocatalyzed by TiOz. All the reported activation energies [less than 21 kT/mol] are much lower

than those of ordinary thermal reactions except the degradation of methylene blue reported by Naskar et al. [118]. These values are quite close to that for a hydroxyl radical reaction [95],suggesting that the photodegradation of most organic pollutants

534


be governed by hydroxyl radical reaction. This indicates that the temperature influence is quite weak. This is because that the low thermal energy (kT = 0.026 eV at room temperature) has almost no contribution to the activation of Ti02 due to its high bandgap energy (3.2 eV). Pichat and Herrmann [129] reported that, at ‘high’ temperature (T > 343K),

even a negative apparent activation energy was found in photocatalytic degradation.

A possible explanation is that the adsorption of the reactants may have a significant effect on the degradation rate because adsorption capacity is usually affected inversely by increasing temperature. As a consequence, contrary to thermal reactions there is no need to heat the system. This absence of heating is very attractive for photocatalytic processes carried out in aqueous media, especially for photocatalytic water purification, because there is no need of additional energy in heating large volume of water, which has a high value of heat capacity. p H in Solution and Anions The rate of photocatalytic degradation of an organic compound not only depends on the parameters discussed above, but also on pH and presence of interfering adsorbing species. Theoretically, pH value of the solution has strong influence on all oxide semiconductors, including the surface charge on the solid catalyst particles, the size of the aggregates formed and the band-gap energies of the conductance and valence bands. Besides, pH also influences the adsorption properties of organic compounds and their dissociating state in solution. In literature, influence of pH on the photocatalytic degradation rate is diversified. Higher reaction rates for various photocatalytic processes have been reported at both low and high pH [18-20] and no general conclusions have been obtained till now. Typically, reaction rate varied by less than one order of magnitude from one end of the pH range to the other. Maximum degradation rate of 4-NP was reported [I301 at pH 4.5, but the degradation rate increased insignificantly with the increase of pH value beyond 10. Augugliaro et a]. [ 1313 investigated the effect of pH in the range of 3 to 11 on the degradation of 4NP in Ti02 suspension and found that the degradation rate decreased slightly with increasing pH value. Herrmann et al. [ 5 5 ] reported that both low and high pH were not favorable to the photocatalytic oxidation of oxalic acid in the TiO, aqueous

535


suspensions. Kim and Anderson [ 1321 found the best pH value for both photocatalytic and photo-electrocatalytic degradation of formic acid over immobilized Ti02 film is around its pK, value, i.e. pH = 3.75. They explained the result in terms of the influence of pH on the adsorption of formate ion (HCOO') on Ti02 surface. Chen and Ray [56] also found that both low and high pH were not suitable for the degradation of 4-nitrophenol in P25 Ti02 suspensions. The high O H ion content of the system enhances the electrodhole separation through reaction (Equation 11). However, in alkaline system the photogenerated COz will be trapped in the solution, and bicarbonate and carbonate are formed. According to the carbonate equilibrium: H 2 C 0 3 ( C 0 2 -I- H 2 0 ) HCOT

PK1= 10.2 \

PKl = 6.3 \

-Cot-

+ H'

-

HCOF+H+

... ... (39)

Between pH 6.3 and 10.2, the inorganic carbon is mainly present in the form HC03while pH >10.2 as CO:-.

At pH < 6.3, molecular C02 dominates, which evaporates

easily in an open system. Hence, the reason for retardation of the photo-degradation at pH > 6.3 may be an inhibiting effect of HCO