Effect of Surface Photoreactions on the Photocoloration of a Wide

Feb 25, 2005 - A nonphotocatalytic reaction occurring on the surface of an irradiated wide band gap metal oxide, such as. ZrO2, can affect the process...
1 downloads 0 Views 199KB Size
Download PDF
J. Phys. Chem. B 2005, 109, 5175-5185

5175

Effect of Surface Photoreactions on the Photocoloration of a Wide Band Gap Metal Oxide: Probing Whether Surface Reactions Are Photocatalytic A. V. Emeline,‡,§,† G. V. Kataeva,§,| A. V. Panasuk,§,| V. K. Ryabchuk,§ N. V. Sheremetyeva,§ and N. Serpone*,‡,⊥ Department of Chemistry and Biochemistry, Concordia UniVersity, Montreal, Quebec H4B 1R6, Canada, Department of Photonics, Institute of Physics, St.-Petersburg State UniVersity, St.-Petersburg, Russia, Department of Physics, UlyanoVsk State UniVersity Branch in DimitroVgrad, DimitroVgrad, Russia, and Dipartimento di Chimica Organica, UniVersita di PaVia, Via Taramelli 10, PaVia 27100, Italy ReceiVed: October 20, 2004; In Final Form: December 15, 2004

A nonphotocatalytic reaction occurring on the surface of an irradiated wide band gap metal oxide, such as ZrO2, can affect the process of photoinduced formation of Zr3+, F- and V-type color centers. The effect of such reactions is seen as the influence of photostimulated adsorption on the photocoloration of the metal oxide specimen. In particular, photoadsorption of electron donor molecules leads to an increase of electron color centers, whereas photoadsorption of electron acceptor molecules leads to an increase of hole color centers. Monitoring the photocoloration of a metal oxide during a surface photochemical reaction probes whether the reaction is photocatalytic: accordingly, the influence of simple photoreactions on the photocoloration of ZrO2, reactions that involved the photoreduction of molecular oxygen, the photooxidation of molecular hydrogen, the photooxidation of hydrogen by adsorbed oxygen, and the photoinduced transformation of ammonia and carbon dioxide. Kinetics of the photoprocesses are reported, as well as the photoinduced chesorluminscence (PhICL effect) of ammonia. Thermoprogrammed desorption and mass spectral monitoring of the photoreaction involving NH3 identified hydrazine as an intermediate and molecular nitrogen as the final product. The photoreactions involving NH3 and CO2 are nonphotocatalytic processes, in contrast to the photooxidation of hydrogen which is photocatalytic. Carbon dioxide and carbonate radical anions are formed by interaction of CO2 with Zr3+ centers and hole states (OS-•), respectively. Mechanistic implications are discussed.

Introduction Heterogeneous photochemistry has attracted considerable attention in the last two decades1-6 because of its connection to issues such as the chemical transformation and storage of solar energy, the purification of water and air, the artificial photosynthesis of value-added chemical products, and the production of modern materials. The related field heterogeneous photocatalysis has some potential advantages as it combines photochemistry and heterogeneous catalysis. However, despite the large body of theoretical and experimental results obtained in this field, there is still no clear evidence as to whether a photochemical reaction taking place on the surface of an irradiated metal-oxide specimen in a heterogeneous system is a photocatalytic process. In particular, whether the photocatalyst remains unchanged at the completion of the photocatalytic reaction cycle(s) as required by the definition of catalysis has remained elusive. A major reason that this issue is raised is the photoinduced formation of defects in solids7-12 caused by trapping of photogenerated charge carriers and excitons by intrinsic and extrinsic defects in solids,7,10-12 and by the self-trapping of excitons in regular lattice sites of photosensitive solids.8,9 The * Address all correspondence to this author. E-mail: serpone@ vax2.concordia.ca or [email protected]. ‡ Concordia University. § St.-Petersburg State University. † Present address: Kanagawa Academy of Science & Technology, Kawasaki-shi, Kanagawa, Japan. | Ulyanovsk State University Branch in Dimitrovgrad. ⊥ Universita di Pavia.

corresponding processes are considered photochemical redox reactions in solids that create new defects with higher reduced and/or oxidized states. Examples of such processes are the photoinduced formation of O-• and Ti3+ or Zr3+ states in titanium and zirconium oxides, respectively.10-12 Such processes take place regardless of the advent of surface photochemical reactions.10,12 Accordingly, a solid photocatalyst typically changes its state during photoexcitation. Thermodynamically, this corresponds to the creation of quasi-Fermi levels for the photoinduced defects that generally differ from the quasi-Fermi levels of photogenerated free charge carriers (electrons and holes). Accordingly, with photoinduced formation of new defects, the photocatalyst changes its thermodynamic state, and consequently does not possess the same state as the original state during or after irradiation. This is a typical situation in heterogeneous catalysis when the stationary state(s) of surface structure and composition of the catalyst during the catalytic process differs from the initial state. Rigorously stated, the photocatalyst retains its original state after photoexcitation only in the case of complete relaxation to its original ground state. An example of such relaxation is presented in Scheme 1a, which shows that photoexcitation of the solid specimen results in the generation of free charge carriers (electrons in the conduction band and holes in the valence band) as a result of band-to-band transition. Its relaxation occurs through fast recombination of charge carriers, either through band-to-band or through specific (solid defects) recombination centers (R). As a result of such complete recombination relaxation, the photocatalyst restores its original state at the termination of

10.1021/jp0452047 CCC: $30.25 © 2005 American Chemical Society Published on Web 02/25/2005

5176 J. Phys. Chem. B, Vol. 109, No. 11, 2005

Emeline et al.

SCHEME 1

photoexcitation. However, when relaxation through a recombination pathway is incomplete, the original state of the photocatalyst is not regenerated (Scheme 1b). In this case, the first step of relaxation involves trapping of charge carriers by solid defects (e.g., anion and cation vacancies) in a manner otherwise identical with the first step of recombination through recombination centers (Scheme 1a) leading to formation of Fand V-type defects (i.e. the color centers). The subsequent step of recombination of a trapped charge carrier (e.g., electron) with its free counterpart (e.g., hole) is much less effective (dashed arrows, Scheme 1b). This leads to an accumulation of trapped charge carriers, to incomplete relaxation of the solid, and to a new state of the photocatalyst that differs from the original. The charge conservation law requires that in the case of complete relaxation there are no trapped charge carriers remaining after termination of irradiation; that is, [F] ) 0 and [V] ) 0 (F and V refer to trapped electrons and holes, respectively). For incomplete relaxation, [F] ) [V] * 0, and because the number of preexisting defects in the solid is limited, the kinetics of accumulation of trapped carriers become saturated. The level of saturation is determined by the efficiencies of charge carrier trapping and by the decay of trapped carriers through different pathways including recombination with free charge carriers of the opposite sign. Saturation of photocoloration of the metaloxide specimen occurs when d[F]/dt ) 0 and d[V]/dt ) 0, and [F] and [V] are given by

[F] ) [V] )

ktr(e)[e][Va] (ktr(e)[e] + kr(h)[h]) ktr(h)[h][Vc] (ktr(h)[h] + kr(e)[e])

(1)

(2)

where ktr(e) and ktr(h) are the rate constants of trapping of electrons and holes by anion (Va) and cation (Vc) vacancies, respectively; kr(h) and kr(e) are the rate constants of recombination of free holes and electrons with F- and V-type color centers, respectively. Consequently, the photocoloration of the solid specimen prevents the restoration of the original state of the photocatalyst even when no surface reaction occurs. The ideal photocatalytic cycle is illustrated in Scheme 1c, which shows that the photocatalyst returns to its original ground state in the same manner as in the case of internal charge carrier recombination. This time, however, relaxation takes place through external surface chemical reaction cycles. Generally, it is not necessary for the reaction cycle to be a closed-loop process. It is sufficient that the number of electrons consumed by the electron acceptor A be equal to the number of electrons transferred to the catalyst by the electron donor molecules D, provided that the reaction products are not strongly bonded to the surface of the catalyst so as not to change the chemical composition of the metal-oxide surface. The latter is also true for a photocatalytic process consisting of a closed-loop reaction cycle. In other words, the condition for true photocatalysis is then given by d[A]/dt ) d[D]/dt; that is, the rate of the surface reduction reaction must equal the rate of the oxidation reaction. Otherwise, together with the occurrence of the catalytic process, there would also be a noncatalytic surface chemical side-reaction determined by which half-reaction of the catalytic cycle is the more efficient. In addition, excess charge will accumulate in the solid. In the extreme case of Scheme 1d, when only one half-reaction takes place on the surface, the heterogeneous photochemical reaction is then stoichiometric and not photocatalytic. The simplest example of such a photochemical reaction is the photoinduced chemisorption of molecules on the surface of metal-oxide specimens.10,12 Typically, the fate of charge carriers in a heterogeneous system is given by the equality eq 3. Thus,

[e] + [eR] + [eVa] + [eV] + [eA] ) [h] + [hR] + [hVc] + [hF] + [hD] (3) where [e] and [h] refer to the number of photogenerated free electrons and holes, respectively. Under moderate levels of photoexcitation, these values rapidly become negligible relative to others in eq 3; [eR] and [hR] represent the number of electrons and holes trapped by the recombination centers, R. When fast recombination occurs, [eR] ) [hR] and can thus be eliminated from eq 3; [eVa] and [hVc] denote the number of electrons and holes trapped by the corresponding defect centers (e.g., Va and Vc) yielding F- and V-type color centers; [eV] and [hF] refer to the number of electrons and holes trapped by the corresponding color centers through which complete recombination can occur; and [eA] and [hD] are the number of electrons and holes involved in surface chemical reactions with acceptor and donor molecules, respectively. Equation 3 then simplifies to

{[eVa] - [hF]} + [eA] ) {[hVc] - [eV]} + [hD]

(4)

[F] + [eA] ) [V] + [hD]

(5)

or to

since [eVa] - [hF] ) [F] is the number of photogenerated electron color centers, and [hVc] - [eV] ) [V] is the number of photogenerated hole color centers. Equation 5 thus establishes the correlation between the number of photoinduced color

Photocoloration of a Wide Band Gap Metal Oxide centers and the number of molecules involved in the surface chemical reaction. Clearly, for true photocatalysis (see Scheme 1c) in which [eA] ) [hD], the condition for the photocoloration of the metal oxide sample is the same with or without a surface reaction; i.e., [F] ) [V]. However, if the surface photoreaction is not catalytic, that is if [eA] * [hD], then [F] * [V]. Thus, a nonphotocatalytic surface reaction affects the formation of photoinduced color centers by altering the relationship between electron and hole color centers. The more pronounced the nonphotocatalytic nature of the surface reaction is, the stronger is the deviation of eq 5 from equality. This is observed as the influence of photostimulated chemisorption on photocoloration when photoadsorption of donor molecules increases the number of electron color centers, whereas photoadsorption of acceptor molecules increases the number of hole color centers.10,12 Accordingly, monitoring the photocoloration of the solid during the surface photochemical reaction provides an opportunity to evaluate whether the surface process is photocatalytic and to what extent it is photocatalytic. These two issues are the focus of the studies reported herein in which we examine the influence of simple photoreactions on the photocoloration of the metal oxide ZrO2. In particular we probe such processes as the photostimulated adsorption of oxygen and hydrogen, the photooxidation of hydrogen by adsorbed oxygen, and the photoinduced transformations of adsorbed NH3 and CO2 on the irradiated surface of powdered zirconia. In addition, some conclusions about the mechanisms of the photoreactions are described. Properties of Zirconia. ZrO2 is a direct band gap dielectric with two direct band-to-band transitions at 5.2 and 5.79 eV.13,14 The flatband potential (Efb) of the conduction band is located at +1.0 V vs NHE at pH 0.15 The tetragonal structure of ZrO2 is stable in particles with sizes smaller than 30 nm.16 Zirconia belongs to the class of photoresistant metal oxides, and typical of this class are new defect centers that are formed from trapping of carriers by existing lattice defects. Anion vacancies, Va, constitute the principal intrinsic defects in powdered zirconia.17 UV irradiation of this dielectric gives rise to broad absorptions in the 250-900 nm spectral region assigned to formation of electron (Zr3+, F and F+ species18) and hole (V-like) color centers,19,20 and evidenced by ESR spectroscopy: Zr3+ and F+ centers21 and V-type centers.22 Zirconia exhibits a broad complex emission around 500 nm through intrinsic and extrinsic UV excitation at 290 nm, and also displays X-ray and thermoluminescence in X-ray preirradiated single crystals.23-25 Electron trapping by anion vacancies has been implicated in the mechanism of luminescence.23 Experimental Section High purity grade powdered ZrO2 in the monoclinic crystal form (IREA) was produced from ZrOCl2 and confirmed by XRD methods; the specific B.E.T. (nitrogen) surface area was ca. 7 m2 g-1. Ubiquitous organic impurities and adsorbed molecules on the metal-oxide surface were removed by a thermal pretreatment (T ) 900 K) in an oxygen atmosphere (P ) 100 Pa) and then in vacuo for a few days. Reproduction of the original state of the specimens between experiments was achieved by heating in oxygen for ca. 1 h. Experimental errors in kinetic measurements caused by the reproducibility of the original state of the ZrO2 sample do not exceed ∼10%. Powdered samples were contained in a quartz cell (path length, 5 mm; illuminated area, 6 cm2) connected to a high-vacuum setup equipped with an oil-free pump system. The gas pressure in the reaction cell was ca. 10-7 Pa; the volume of the reactor was 50 cm3.

J. Phys. Chem. B, Vol. 109, No. 11, 2005 5177

Figure 1. Difference DRS spectrum between oxidized and reduced states of zirconia (1), and difference DRS spectrum of surface-active centers for the adsorption of oxygen, centers that were formed by the thermoreduction of zirconia (2).

Hydrogen gas was obtained by the thermal decomposition of TiH4 and oxygen by the thermal decomposition of KMnO4. A Pirani-type manometer (sensitivity, 20 mV Pa-1 for O2 and 24 mV Pa-1 for H2) measured gas pressures in kinetic studies. Gas composition and identification of thermodesorption products were performed by mass spectral techniques (model MX-7301). Thermodesorption studies were carried out at a rate of 0.25 K min-1 in the linear heating regime. A 120-W high-pressure mercury lamp (DRK-120, MELZ) irradiated the solid specimens: light irradiance below 400 nm was 6 mW cm-2 (water filter and an IOFI thermoelement with sensitivity ca. 1.5 V W-1); the photon flow below 250 nm was ∼1015 photons cm-2 s-1. Diffuse reflectance spectra (DRS; BaSO4 was the reference) were recorded with a Karl Zeiss Spekord M-40 spectrophotometer (Germany) equipped with an integrating sphere assembly and interfaced to a computer. The errors in the DRS spectral measurements are less than (0.001 in ∆R. The movable high-vacuum setup/quartz cell system was moved as a unit between the position of irradiation and of thermal treatment of the specimen to the position for recording spectra; both were fixed with high precision instruments. The oxidized state of the zirconia surface was formed during preheating of the sample in a 100 Pa oxygen atmosphere at 900 K for several hours, after which the sample was cooled to ambient temperatures in an atmosphere of oxygen. Between the series of experiments with the oxidized surface, the sample was treated under otherwise identical conditions for 1 h. The reduced state of the zirconia surface was formed by preheating the oxidized surface in a hydrogen atmosphere for ca. 30 min also at 900 K, followed by heating under dynamic vacuum at the same temperature for 3 h. Between experiments, the reduced sample was heated in oxygen for ca. 15 min followed by heating under dynamic vacuum for 45 min at 900 K. Results and Discussion Photocoloration: Influence of the Reduced or Oxidized State of the Surface. The absorption spectrum 1 of Figure 1 portraying the difference DRS spectra between oxidized and reduced states of zirconia (as ∆R ) Rox - Rred) consists of a broad band around 280 nm corresponding to absorption by Zr3+ defects.10,12,19,20,26 This is in keeping with results from Liu and co-workers,27 who showed that reduction of ZrO2 particles larger than 50 nm generates mostly Zr3+ sites, rather than oxygen vacancies. The low surface area of our ZrO2 sample is in line with a particle size greater than 50 nm. Only a fraction of the Zr3+ states are located on the surface as evidenced from the

5178 J. Phys. Chem. B, Vol. 109, No. 11, 2005

Emeline et al.

Figure 2. Difference DRS spectra of photogenerated color centers formed after irradiation for 1000 s of the oxidized (1) and reduced (2) zirconia in vacuo.

effect of oxygen adsorption on the absorption spectrum of Zr3+ centers whose broad absorption at 240-280 nm is considerably less intense in the presence of oxygen (spectrum 2, Figure 1). The absorption spectrum of irradiated zirconia in vacuo (as ∆R ) Rdark - RhV) consists of three poorly resolved broad bands at 280 nm, 450 nm, and in the 600-650 nm spectral range (Figure 2); they are attributed to absorption by photoinduced color centers. The shape and the intensity of these bands strongly depend on the initial state of the zirconia surface: that is, whether the initial (dark) surface was oxidized or reduced. Clearly, surface reduction results in a smaller level of coloration (i.e., less intense absorption) of zirconia particles affecting mosty the shorter wavelength region (λ ∼ 280 nm) of the absorption spectrum, which corresponds to absorption by Zr3+ centers. Figure 3 illustrates the kinetics of photocoloration at 280 nm (Zr3+ state), 360 nm (mainly V-type hole color centers), and 620 nm (F-type electron color centers) for the reduced (Figure 3a) and oxidized (Figure 3b) ZrO2 surface in vacuo. The kinetics depend sub-linearly on irradiation time approaching a limiting value at the longer times, summarized in Table 1 (column 2). In accordance with Scheme 1b, the kinetics of photocoloration are basically dictated by two opposing processes: (1) formation of color centers (F, V, and Zr3+) by trapping of free charge carriers of the corresponding sign at suitable defect sites (Va, Vc, and Zr4+, respectively), and (2) destruction of color centers through recombination with free charge carriers of the opposite sign. The kinetics of formation of Zr3+ color centers (eqs 6 and 7)

Zr4+ + e f Zr3+

ktr

(6)

Zr3+ + h f Zr4+

kr

(7)

are given by eq 8, whereas the limiting level of photocoloration of ZrO2 based on Zr3+ centers, [Zr3+]lim, is described by eq 9

d[Zr3+] ) ktr[Zr4+][e] - kr[Zr3+][h] dt [Zr3+]lim )

ktr[Zr]o[e] (ktr[e] + kr[h])

(8)

(9)

where [Zr]o is the total number of specific sites on zirconia capable of producing color centers of a given sort. Therefore,

Figure 3. Plots illustrating the kinetics of accumulation of photogenerated color centers formed during irradiation of (a) reduced and (b) oxidized zirconia in vacuo recorded at three different wavelengths.

the kinetics and the limit of photocoloration are dictated by the ratio between the rate constants of charge carrier trapping (ktr) and charge carrier recombination (kr). Moreover, the kinetics of photocoloration depend on the initial concentration of the color centers. Thus for the reduced state of zirconia, the initial concentration of Zr3+ centers is high enough to make the initial rate of defect formation negative (eq 8), whereas for the oxidized state of zirconia the initial concentration of Zr3+ is negligible and the initial rate of photocoloration is therefore positive. As well, from eq 9 and the experimental photocoloration limiting values (∆Rtf∞; Table 1) we conclude that the total concentration of Zr-based defect sites (preexisting plus those photogenerated) that produce Zr3+ centers is greater for the reduced than for the oxidized surface. Concomitantly, the total concentration of vacancies responsible for the formation of other types of color centers (e.g., of the F- and V-type) is somewhat lower for the reduced state, thus leading to the lower limit of photocoloration compared to the oxidized form of the sample.27 Because of the nonuniformity of the intensity of the actinic light within the profile of light propagation caused by light absorption, the (apparent) kinetic data of photocoloration presented in Table 1 can nonetheless be used for a qualitative comparison of the efficiency of attainment of photocoloration in different environments, as the experiments were performed under otherwise identical conditions. Photocoloration: Influence of Photoadsorption of Molecular O2 and H2. The photostimulated adsorptions of hydrogen and oxygen, whereby hydrogen is oxidized to H+ and oxygen is reduced to O2-•, are examples of two simple surface photochemical reactions. Each represents only half of the complete photocatalytic cycle, and the products of photoadsorption are chemisorbed to the corresponding surface-active

Photocoloration of a Wide Band Gap Metal Oxide

J. Phys. Chem. B, Vol. 109, No. 11, 2005 5179

TABLE 1: Kinetic Data for the Biphasic Photocoloration of Oxidized and Reduced Zirconia Surfaces under Various Experimental Conditions (column 1) at Different Wavelengths Corresponding to Zr3+ Centers (280 nm), V-Type Centers (360 nm), and F-Type Electron Centers (620 nm) heterogeneous system

monitored wavelength (nm)

∆Rtf∞

k1 (10-2 s-1)

k2 (10-2 s-1)

vacuum/ZrO2 oxidized

280 360 620 280 360 620 280 360 620 280 360 620 280 360 620 280 360 620 280 360 620 280 360 620 280 360 620

0.052 ( 0.002 0.054 ( 0.002 0.052 ( 0.002 0.004 ( 0.001 0.033 ( 0.001 0.036 ( 0.001 0.066 ( 0.002 0.12 ( 0.03 0.056 ( 0.002 -0.008 ( 0.001 0.036 ( 0.001 0.031 ( 0.001 0.069 ( 0.003 0.086 ( 0.002 0.077 ( 0.003 0.006 ( 0.001 0.032 ( 0.001 0.040 ( 0.002 -0.008 ( 0.001 0.032 ( 0.001 0.034 ( 0.001 0.024 ( 0.001 0.033 ( 0.001 0.041 ( 0.002 0.010 ( 0.003 0.033 ( 0.001 0.043 ( 0.002

5.3 ( 0.5 4.2 ( 0.2 2.7 ( 0.2 5.6 ( 0.3 25 ( 3 13 ( 2 5.6 ( 0.9 2.6 ( 0.1 2.9 ( 0.2 5.9 ( 1.4 24 ( 4 25 ( 6 5.6 ( 1.2 5.0 ( 0.5 7.1 ( 2.0 4.0 ( 0.3 2.0 ( 0.2 11 ( 2 8.3 ( 1.4 20 ( 4 20 ( 4 15 ( 2 31 ( 2 26 ( 1 20 ( 8 24 ( 2 25 ( 6

0.29 ( 0.02 0.29 ( 0.02 0.17 ( 0.01 0.14 ( 0.01 1.3 ( 0.3 0.43 ( 0.05 0.42 ( 0.09 0.030 ( 0.007 0.11 ( 0.02 0.33 ( 0.10 0.91 ( 0.17 0.71 ( 0.10 0.32 ( 0.08 0.17 ( 0.03 0.38 ( 0.10 0.16 ( 0.03 0.67 ( 0.09 0.50 ( 0.13 0.20 ( 0.03 0.67 ( 0.13 0.50 ( 0.08 0.71 ( 0.10 3.3 ( 0.3 0.50 ( 0.06 1.1 ( 0.2 1.3 ( 0.3 0.43 ( 0.16

vacuum/ZrO2 reduced O2/ZrO2 oxidized O2/ZrO2 reduced H2/ZrO2 oxidized H2/ZrO2 reduced (H2+O2)/ZrO2 NH3 ads/ZrO2 CO2 ads/ZrO2

sites. Hence for each photoreaction, eq 5 simplifies to expressions 10 and 11.

[F] + [O2 ads] ) [V]

(10)

[F] ) [V] + [H2 ads]

(11)

Subtracting the condition that describes the photocoloration in vacuo, i.e., [F] ) [V] * 0, yields

∆[F] + [O2 ads] ) ∆[V]

(12)

∆[F] ) ∆[V] + [H2 ads]

(13)

Consequently, in one limiting case, the photostimulated adsorption of oxygen can cause either a decrease of the electron color centers or an increase of hole color centers, whereas in the other limiting case the photoadsorption of molecular hydrogen can lead either to an increase of electron color centers or to the decay of hole color centers. Clearly, some intermediate situation(s) is also possible. {A detailed theoretical analysis of the parameters that determine the direction and the degree to which photostimulated adsorption of these simple molecules influence the photoinduced coloration of metal oxides is currently in progress.} The absorption spectra of photoinduced color centers formed during the photostimulated adsorption of H2 on oxidized and O2 on reduced zirconia surfaces are reported in Figure 4, parts a and b, respectively. The major difference between the spectra of the oxidized and reduced states of the sample is again observed principally in the shorter wavelength spectral region of the Zr3+ centers (as expected; see Figure 4b), since this is the region of the spectra that is influenced mostly by reduction of the surface. In accordance with eqs 12 and 13, the increase in absorption in spectra 2 and 3 relative to spectrum 1 (Figure 4a) corresponds to absorption by hole centers in the case of the photostimulated

Figure 4. Difference DRS spectra of photogenerated color centers formed after irradiation for 1000 s of oxidized (a) and reduced (b) zirconia in vacuo (1), in the presence of oxygen (2), and in the presence of hydrogen (3).

adsorption of oxygen (curve 2) and to electron centers for the photoadsorption of hydrogen (curve 3), respectively. In such a case, the corresponding centers located on the surface of zirconia behave as surface-active centers of post-adsorption of the

5180 J. Phys. Chem. B, Vol. 109, No. 11, 2005

Emeline et al.

Figure 5. Difference DRS spectra of surface photogenerated color centers for the post-adsorption of oxygen (1), hydrogen (2), and ammonia (3) on the preirradiated zirconia surface.

corresponding gases: (i) surface electron centers react with oxygen and (ii) surface hole centers react with hydrogen. As a consequence of post-adsorption, light absorption by the corresponding surface defects will decrease. Absorption spectra of the surface-active centers for the post-adsorption of O2 and H2 are illustrated in Figure 5 (spectra 1 and 2), which also reports the spectrum for the post-adsorption of ammonia (spectrum 3) that will be discussed below. It is evident that the increase in absorption during the photoadsorption process is indeed caused by the corresponding surface-active centers of post-adsorption in keeping with eqs 12 and 13. The kinetics of accumulation of photoinduced color centers in vacuo (1) and in the presence of oxygen (2) and hydrogen (3) are displayed in Figure 6 for the reduced states of ZrO2 at three selected wavelengths: 280, 360, and 620 nm corresponding to absorption by Zr3+ defects, and by V-type and F-type color centers, respectively.10,12,19,20 Also reported are the kinetics of accumulation of photogenerated color centers during the photooxidation of H2 in the presence of molecular oxygen. Despite some differences in the kinetic behavior between oxidized and reduced states of the zirconia sample, in both cases the effects of the photostimulated adsorption of gases on the photocoloration of the metal-oxide specimen are similar and follow the expectations of eqs 12 and 13. For both states, the photoadsorption of hydrogen increases the number of Zr3+ centers (λ ) 280 nm) and electron F-type color centers (λ ) 620 nm): curve 3 of Figure 6a,c relative to those in vacuo (curve 1). Concomitantly, absorption at 360 nm for reduced zirconia, corresponding mostly to hole V-type color centers (curve 3 of Figure 6b), decreases relative to that in vacuo (curve 1). That is, photoadsorption of oxygen leads to increased absorption by hole color centers (curve 2 of Figure 6b) and to diminished absorption by electron color centers (λ ) 280 and 620 nm; curves 2 and 4 of Figure 6a,c). To the extent that photostimulated adsorption is clearly a nonphotocatalytic process, we propose to use the effect of photoadsorption on the photocoloration of ZrO2 in particular, and metal oxides in general, as a reference benchmark to evaluate the catalytic or noncatalytic nature of surface photoreactions. If the spectral and kinetic behaviors of photocoloration during the course of a photoreaction were similar to the effect of photoadsorption of oxygen, then we would infer that the reaction was strongly shifted toward the reduction process. In turn, the similar photocoloration behaviors observed during the photoadsorption of hydrogen would infer an oxidative photoreaction. Accordingly, for a true photocatalytic process the

Figure 6. Kinetics of accumulation of the photoinduced formation of (a) F-type color centers, (b) V-type color centers, and (c) Zr3+ color centers during irradiation of reduced zirconia in vacuo (1), in the presence of oxygen (2), in the presence of hydrogen (3), and during the photooxidation of hydrogen in the presence of oxygen (4) recorded at three different wavelengths.

spectral behavior should be similar to the spectrum of color centers after irradiation in vacuo. However, the kinetics of photocoloration need not necessarily correspond to those in vacuo, since such kinetics are essentially dictated by the mechanism(s) of the photocatalytic process. In this regard, it is important to emphasize that a true photocatalytic process begins not only when the surface reaction has reached the stationary state, but also when the photocoloration has reached its limiting value. Photocoloration: Influence of a Photocatalytic Reaction (Photooxidation of H2 in the Presence of O2). Mechanistic details of the photooxidation of hydrogen in the presence of oxygen taking place on zirconia particles were reported in an earlier study,26 which afforded major conclusions on the pathway of the photoreaction. This photoreaction is photocatalytic as confirmed by the experimentally determined turnover numbers: TON ) 6.6 for oxygen and 14 for hydrogen after an

Photocoloration of a Wide Band Gap Metal Oxide irradiation period of 100 min. Moreover, (a) the first limiting step of the photoprocess involved the photostimulated adsorption of oxygen; (b) activation of secondary chemical steps on the particle surface was caused by the dissociative adsorption of hydrogen producing H atoms, which reacted with preadsorbed oxygen; and (c) the most active oxygen species preadsorbed on Zr3+ surface defects that reacted with H atoms did not accumulate on the surface during the photoreaction, unlike the other two oxygen forms (see ref 26). The kinetics of accumulation of color centers during the photoexcitation of reduced zirconia in the presence of the H2/ O2 gas mixture (curves 4), measured at 280, 360, and 620 nm, are shown in Figure 6, together with the kinetics of formation of color centers during irradiation of ZrO2 in vacuo (curve 1), in oxygen (curve 2), and in hydrogen (curve 3) reported as references for comparison purposes. The kinetic time constants for the biphasic increase of color centers (curves 4) are summarized in Table 1: k1 ) 8.3 × 10-2 s-1 and k2 ) 2.0 × 10-3 s-1 (Zr3+ centers, 280 nm); k1 ) 2.0 × 10-1 s-1 and k2 ) 6.7 × 10-3 s-1 (V centers, 360 nm); and k1 ) 2.0 × 10-1 s-1 and k2 ) 5.0 × 10-3 s-1 (F centers, 620 nm). The above kinetic observations reveal that the initial period of irradiation is characterized by a lower rate of formation of electron F-type color centers (Figure 6a) in accordance with the higher rate of accumulation of photoadsorbed oxygen species on the surface of zirconia during the photoreaction.26 Concomitantly, the enhanced rate of photoadsorption of hydrogen at the beginning of irradiation causes a decrease of the rate of formation of hole color centers. It is important to note that the photoinduced coadsorption of hydrogen and oxygen leads to a different behavior of the photocoloration compared to the photoadsorption of each gas separately (see above). To explain this effect we must recognize that most of the additional color centers formed as a result of the photoadsorption of the gases examined are located on the metal oxide surface. Accordingly, coadsorption of another gas results in the disappearance of the absorption of the corresponding surface color centers, and in an increase of the rate of adsorption of the coadsorbate.26 Stated differently, the photoadsorption of oxygen increases the rate of photoadsorption of hydrogen, and the photoadsorption of hydrogen increases the photoadsorption of oxygen. After a certain period of irradiation, however, the role of photoadsorption decreases, and the photocatalytic oxidation of hydrogen in the presence of oxygen reaches the stationary state. Thus, the levels of coloration of zirconia by hole V-type color centers and by electron F-type color centers approach the level of coloration seen in vacuo (see Figure 6a,b). The absorption spectra of color centers produced in the powdered zirconia sample irradiated in vacuo (spectrum 1) and in the presence of a H2/O2 gas mixture (spectrum 2), as well as in the presence of preadsorbed CO2 (spectrum 3) and preadsorbed NH3 (spectrum 4), are presented in Figure 7. Comparison of spectra 1 and 2 shows that the effect in the spectral region of absorption by hole defects (360 nm) is rather small (if any) compared to the spectral region corresponding to absorption by electron F-type color centers (620 nm) in which the effect of irradiation with the H2/O2 gas mixture is similar to the one observed for irradiation of ZrO2 in the presence of oxygen only (Figure 4b, spectrum 2). The most significant effect is observed in the short wavelength spectral region (280 nm), which corresponds to absorption by Zr3+ electron centers. The correlations between the time-evolving rate of consumption of hydrogen and oxygen and the temporal evolution of the difference in coloration levels for hole V centers and electron

J. Phys. Chem. B, Vol. 109, No. 11, 2005 5181

Figure 7. Difference DRS spectra of photogenerated color centers recorded after 1000 s irradiation of zirconia in vacuo (1) (for reference purposes), in a mixture of oxygen and hydrogen (2), and in the presence of preadsorbed carbon dioxide (3) and preadsorbed ammonia (4).

F color centers, respectively, during the photoreactions relative to photocoloration in vacuo are presented in Figure 8, parts a and b. These correlations demonstrate that the kinetic behavior of photocoloration during the photoreactions follows the expectations from eq 5. At shorter times of irradiation, when the rates of consumption of gases are faster, the coloration level tends to decrease, whereas at longer times of irradiation when the rates of consumption of gases decrease and approach the stationary level, the coloration tends to approach the level observed in vacuo. This supports our earlier conclusion26 that at the beginning of the photoreaction, the photostimulated adsorption of gases plays the most significant role in the photooxidation of hydrogen, whereas during the course of the photoreaction this role tends to decrease such that the photocatalytic oxidation becomes the more important process that determines the kinetic behavior of the heterogeneous system. It should also be noted that the level of Zr3+ centers during the photoreaction is much higher than that during irradiation in vacuo, or during the photostimulated adsorption. This also confirms the earlier conclusion26 that the most reactive oxygen species on the zirconia surface are bonded to surface Zr3+ centers. Since these oxygen species do not accumulate on the surface during the photooxidation of hydrogen, they do not occupy the corresponding surface centers, so that enhanced hydrogen adsorption causes an increase in the formation of Zr3+ surface states resulting in the enhanced absorption at 280 nm (see eq 5 and curve 4 in Figure 6c). The experimental results above suggest that during the initial period of irradiation of the heterogeneous system, the photoadsorptions of both oxygen and hydrogen play the most significant role, and lead to the changes observed in the kinetic and spectral behaviors of photocoloration. By contrast, at longer irradiation times the photocatalytic oxidation of hydrogen becomes the dominating process and the level of photocoloration approaches the level observed in vacuo; that is, [F] ) [V]. It is noteworthy that, concomitantly, the reaction rate also approaches the stationary rate of the photocatalytic reaction when [eA] ) [hD], in accordance with eq 5. Consequently, we conclude that the photooxidation of hydrogen becomes truly a photocatalytic reaction when photocoloration of the photocatalyst achieves its limiting value as per the condition [F] ) [V] * 0. Accordingly, this particular state of the photocatalyst must be considered as the working active state of the photocatalytic process. Surface Photooxidation of NH3: Influence on Photocoloration. There is a significant albeit slow saturated adsorption of NH3 on the surface of zirconia in the dark, as is typical of

5182 J. Phys. Chem. B, Vol. 109, No. 11, 2005

Emeline et al.

Figure 10. Formation and decay of the photoinduced chesorluminescence (PhICL) emission (see text) produced by the post-adsorption of ammonia on zirconia preirradiated for 1000 s.

resulting from the interaction of NH3 with surface hole states (such as OS-•) as evident from experiments of the post-

NH3 + OS-• f OH- + NH2•

(14)

adsorption of NH3 on the zirconia surface that had been preirradiated in vacuo. In this regard, the post-adsorption of NH3 causes a decrease of absorption by surface hole centers in the DRS spectra, in the same manner as hydrogen does (compare, for example, spectra 2 and 3 of Figure 5 with spectrum 1 in vacuo of Figure 4a), reaction 15. It is significant to note that after preadsorption of hydrogen on ZrO2, the post-adsorption Figure 8. (a) Time evolution of the rate of hydrogen consumption during the photooxidation of hydrogen in the presence of oxygen (1), and kinetics of the difference between accumulation of color centers (2) during irradiation of ZrO2 in the presence of hydrogen and in a mixture of oxygen and hydrogen recorded at 360 nm; (b) time evolution of the rate of oxygen consumption during the photooxidation of hydrogen in a H2/O2 gas mixture (1), and kinetics of difference between accumulation of color centers (2) during irradiation of ZrO2 in the presence of oxygen and in a mixture of oxygen and hydrogen recorded at 620 nm.

Figure 9. Time dependence of the rate of evolution of molecular nitrogen during irradiation of zirconia with preadsorbed ammonia.

most metal oxides that possess surface Lewis and Bronsted acid sites.28,29 Irradiation of zirconia with preadsorbed NH3 leads to the evolution of molecular nitrogen into the gas phase (Figure 9) and to the formation of hydrazine on the particle surface, as evidenced from the mass spectral monitoring of products of thermoprogrammed desorption. The kinetics of evolution of molecular nitrogen into the gas phase (k = 6 × 10-4 s-1) suggest that the process is a multistep reaction resulting in the loss of hydrogen during the formation of molecular nitrogen. The first step of the reaction is formation of NH2• radicals (reaction 14)

H2 + OS-• f OH- + H•

(15)

of NH3 causes no changes in the DRS spectra. The same is true for the reverse situation. That is, when NH3 is preadsorbed before the post-adsorption of hydrogen, no changes are seen in the DRS spectra. This infers that both gases interact with the same type of photoinduced surface-active centers, namely surface hole V-type centers. Additional support for this inference comes from experiments designed to observe the photoinduced chesorluminescence (PhICL) emission.30 The post-adsorption of NH3 on the preirradiated zirconia surface yields a PhICL emission whose temporal evolution is illustrated in Figure 10, after which introduction of hydrogen gas causes no further PhICL emission, as expected if the photoinduced surface hole centers are located on the surface of the metal oxide specimen. In turn, the post-adsorption of hydrogen on the preirradiated ZrO2 surface displays a PhICL emission,30 whereas subsequent addition of NH3 yields no additional detectable emission. Unfortunately, the temporal evolution of the PhICL emission and the corresponding kinetics from the post-adsorption of NH3 are rather badly distorted by the strong adsorption of NH3 on surface acid sites, thus precluding a detailed mechanistic analysis of this PhICL effect, as was reported earlier for hydrogen.30 The product from the post-adsorption of NH3 detected in the TPD spectra is hydrazine (mass spectra; m/z 32) produced from a radical recombination process, reaction 16.

NH2• + NH2• f N2H4

(16)

It is clear from reactions 14 and 16 that to transform hydrazine into the final product, viz., molecular nitrogen, hydrazine must first lose four H atoms in the hydrazine oxidation process. This will necessitate four additional hole states. Accordingly, the reaction will be very consuming with respect to the number of photogenerated holes and will thus strongly affect the photo-

Photocoloration of a Wide Band Gap Metal Oxide

J. Phys. Chem. B, Vol. 109, No. 11, 2005 5183

Figure 12. Difference DRS spectra of surface-active centers for the post-adsorption of carbon dioxide on zirconia preirradiated for 1000 s in vacuo (1), in the presence of oxygen (2), and in the presence of hydrogen (3).

Figure 11. Kinetics of accumulation of photogenerated (a) F-type color centers and (b) V-type color centers during irradiation of zirconia in vacuo (1), in the presence of oxygen (2), in the presence of hydrogen (3), and with preadsorbed ammonia (4) recorded at two different wavelengths.

coloration of zirconia in a manner similar to the effect caused by hydrogen. The absorption spectra of color centers in zirconia irradiated in vacuo (spectrum 1) and in the presence of preadsorbed ammonia (spectrum 4) were displayed earlier in Figure 7. The kinetics of accumulation of electron F-type color centers (620 nm; Figure 11a) show that the presence of preadsorbed NH3 increases the saturation limit of absorption by the color centers (relative to vacuum, curve 1), as is also typical of other donor gases such as hydrogen (curve 3). Both the photoadsorption of hydrogen and the photoreaction of NH3 display a similar influence on the decreased accumulation of hole V-type color centers (360 nm; curves 3 and 4; Figure 11b), and on the photocoloration of zirconia. The greater degree of consumption of photoholes in the presence of NH3 results in a faster approach of the kinetics to the saturation limit (curve 4 of Figure 11b). Accordingly, the kinetics of coloration confirm the notion that the surface photoreaction involving preadsorbed NH3 is due to the trapping of photogenerated holes. This results in a lower limit of saturation for hole color centers and in a higher limit of electron F-type color centers (eq 5), in a manner similar to what happens during the photooxidation of hydrogen molecules. As a result, we conclude that the indirect photolysis of NH3, which yields hydrazine as an intermediate product and molecular nitrogen as the final product, is initiated by interaction of preadsorbed molecules with photogenerated surface hole states only and does not involve photogenerated electrons. The kinetic and spectral behaviors of the photocoloration of the metal-oxide specimen in the presence of ammonia show that the condition [F] ) [V] * 0 is never fulfilled in this

heterogeneous system. Consequently, the surface photoreaction of ammonia on zirconia particles is not photocatalytic. It is simply a stoichiometric surface photochemical oxidation of ammonia. Photoreactions of CO2 on the Zirconia Surface: Influence on Photocoloration. Introduction of carbon dioxide into the reactor containing zirconia particles leads to a rapid, strong, and practically unsaturated adsorption of CO2 on the metal oxide surface producing surface carbonate groups as one of the products of adsorption.29,31-34 The changes in absorption at 280 nm infer that CO2 also adsorbs, albeit partially, on Zr3+surface centers to yield carbon dioxide radical anions (CO2-•), a process that is typical of adsorption of carbon dioxide on electron centers.35,36 However, post-adsorption of carbon dioxide on the zirconia surface that was preirradiated in vacuo results in changes in the DRS spectrum (1 of Figure 12) that is similar to the complex spectrum of both electron and hole color centers. To resolve this spectrum of surface-active centers for the postadsorption of CO2, zirconia was also preirradiated in the presence of oxygen and in the presence of hydrogen. Irradiation in oxygen removes surface electron centers and promotes the accumulation of surface-active hole centers. The resulting spectrum of the surface-active centers of CO2 post-adsorption after preirradiation in oxygen is also illustrated in Figure 12 (spectrum 2). Addition of hydrogen after the post-adsorption of CO2 caused no changes in this absorption spectrum 2. As well, prior post-adsorption of hydrogen after preirradiation of ZrO2 in oxygen also resulted in no further changes in spectrum 2 after the post-adsorption of CO2 (e.g., compare spectrum 2 of Figure 5 with spectrum 2 of Figure 12). Evidently, both CO2 and H2 molecules interact with the same photogenerated surface-active hole centers. The result of this interaction between adsorbed CO2 and the photoactivated zirconia surface (i.e., with trapped holes) is formation of CO3-• anion radicals (reaction 17). However, if preirradiation of zirconia was carried out in the presence of hydrogen, which causes a diminution of surface hole centers and promotes accumulation of surface

CO2 + OS-• {hS,tr} f CO3-•

(17)

electron centers, the post-adsorption of CO2 then results in changes in the absorption spectrum that are typical of postadsorption of oxygen on surface electron centers (compare spectrum 1 of Figure 5 with spectrum 3 of Figure 12). Addition of oxygen after the post-adsorption of CO2 caused no further changes in spectrum 3 of Figure 12. As noted earlier, prior post-

5184 J. Phys. Chem. B, Vol. 109, No. 11, 2005

Emeline et al.

Figure 13. Graph showing the kinetics of photodesorption of CO2 from the zirconia surface.

adsorption of hydrogen after preirradiation of zirconia in the presence of oxygen resulted in no changes in the DRS spectrum after the post-adsorption of CO2. Consequently, in this case also we infer that CO2 adsorbs on surface electron centers (reaction 18). Clearly, then, carbon dioxide interacts with both surface electron centers and surface hole centers (reaction 17).

CO2 + eS,tr f CO2-•

(18)

Irradiation of the zirconia surface with preadsorbed CO2 (in the dark) leads to the photodesorption of carbon dioxide through biphasic kinetics (Figure 13): k1 ) 4.0 × 10-2 s-1 and k2 ) 1.3 × 10-3 s-1. The spectrum of color centers recorded after 1000 s of irradiation of ZrO2 in the presence of preadsorbed CO2 is reported in Figure 7 (spectrum 3). The kinetic behavior of the photocoloration of this ZrO2/CO2 system at three wavelengths is illustrated in Figure 14 {for comparison purposes, the corresponding kinetics from irradiation of ZrO2 in vacuo, in the presence of oxygen, and in the presence of hydrogen are also given}. From the spectral results and the kinetic data (Table 1), the altered behavior caused by preadsorbed carbon dioxide is similar to that observed from the photostimulated adsorption of hydrogen donor molecules. This suggests that preadsorbed CO2 interacts principally with photogenerated holes. However, the behavior in the spectral region around 280 nm (Zr3+ states) differs from the results of the influence of hydrogen on photocoloration. In the presence of preadsorbed CO2, the number of Zr3+ centers is increased significantly (e.g., compare curve 4 with curve 3 in Figure 14c). This increase in Zr3+ absorption correlates with the kinetics of photodesorption of carbon dioxide. Hence, photodesorption of carbon dioxide originates from CO2 molecules that were preadsorbed on the Zr3+ surface centers. Accordingly, photodesorption of CO2 and the corresponding formation of Zr3+ centers take place as proposed in reactions 19 and 20, respectively.

{Zr3+-CO2} + h f Zr4+ + CO2v

(19)

Zr4+ + e f Zr3+

(20)

The process of photodesorption of carbon dioxide consumes photoholes. This results in an increased absorption by electron F-type centers (λ ) 620 nm) and in a decreased absorption (λ ) 360 nm) by V-type hole centers (curves 4 in Figure 14, parts a and b, respectively). Thus, photodesorption of CO2 is one of the processes that affects the photocoloration of zirconia, and preadsorbed CO2 behaves as an electron donor species partici-

Figure 14. Kinetics of accumulation of the photoinduced formation of (a) F-type color centers, (b) V-type color centers, and (c) Zr3+ color centers during the irradiation of zirconia in vacuo (1), in the presence of oxygen (2), in the presence of hydrogen (3), and with preadsorbed carbon dioxide (4) recorded at three different wavelengths.

pating in surface oxidative reactions. However, photodesorption is not the sole surface photoreaction that takes place in this heterogeneous system. As noted earlier from the spectra of surface-active centers for the post-adsorption of carbon dioxide, CO2 can also react with both surface holes and surface electrons (reactions 17 and 18). At the same time, the data on photocoloration in the presence of preadsorbed CO2 show that the interaction with surface hole centers is the dominant process, leading to formation of carbonate radical anions CO3-• (reaction 17) as a result of hole trapping by surface carbonates produced by preadsorption of carbon dioxide (in the dark) on lowcoordinated oxide ions on the ZrO2 particle surface.33,34 Experiments with post-adsorption of CO2 on preirradiated zirconia have shown that CO3-• species are formed through adsorption on the surface hole states OS-•. Prior to irradiation, these OS-• species were originally the low-coordinated lattice oxide ions, O2-, which are the stronger Lewis base centers for adsorption of CO2 on metal oxides. Consequently, adsorption complexes formed during irradiation by interaction of CO2 with these base sites that acted as effective surface hole traps.

Photocoloration of a Wide Band Gap Metal Oxide To the extent that coverage by preadsorbed carbon dioxide (∆PCO2 ) 200 Pa; about 10-1 of a monolayer) is much greater than the concentration of surface-active centers of postadsorption (about 10-5 of a monolayer37), hole trapping by surface carbonates becomes the dominating photoprocess. Note that these CO3-• radical anions play the role of so-called longlived excited states of adsorbed CO2 in the photoreaction with hydrogen.36 As inferred earlier in the case of NH3, the kinetic and spectral behaviors of photocoloration of the ZrO2 specimen in the presence of preadsorbed carbon dioxide also show that the condition [F] ) [V] * 0 is also never fulfilled in this heterogeneous system. Accordingly, the surface photoreaction involving CO2 is not photocatalytic, and represents merely a surface photooxidation reaction: that is, photooxidation of the CO2-• radical followed by desorption of CO2 (reaction 19). Conclusions This study was aimed at examining the effect(s) of surface photochemical reactions involving H2, O2, NH3, and CO2 on the photocoloration (i.e. formation of F- and V-type color centers) of ZrO2 in particular, and metal oxides in general. The results demonstrate that for a true photocatalytic process, two conditions must be met: (i) the rates of consumption of charge carriers in the reduction and oxidation half-reactions must be equal, and (ii) photocoloration of the metal oxide must achieve the limits of accumulation of hole and electron color centers when the number of electrons trapped by the defects in the solid equals the number of trapped holes. This last condition corresponds to the requirement that at the completion of the reaction cycle the photocatalyst retains its original state. Any deviation from these two conditions will point to a noncatalytic nature of the overall surface photochemical process. In this regard, the effect of nonphotocatalytic reactions, such as the photostimulated adsorption of molecular oxygen and molecular hydrogen on the coloration of the metal-oxide sample, can be used as benchmark references to estimate the extent to which a particular surface photoreaction deviates from being a true photocatalytic process. The spectral and kinetic results from the examination of the effect of photoreactions involving NH3 and CO2 on the photocoloration of zirconia have demonstrated that these photoreactions are not photocatalytic processes. By contrast, the photooxidation of hydrogen in the presence of oxygen is indeed a photocatalytic reaction.26 Acknowledgment. We are grateful to the North Atlantic Treaty Organization, Brussels, for a Collaborative Grant (No. PDD(CP)-PST CLG-979700) between our respective laboratories at Concordia University in Montreal and at the University of St.-Petersburg (Russia). Studies carried out at the Universita di Pavia are supported from a grant from the Ministero dell’Istruzione, dell’Universita e della Ricerca (MIUR, Roma, Italy, to N.S.). References and Notes (1) Serpone, N.; Pelizzetti, E., Eds. Photocatalysis. Fundamentals and Applications; John Wiley & Sons: New York, 1989.

J. Phys. Chem. B, Vol. 109, No. 11, 2005 5185 (2) Zamaraev, K. I.; Parmon, V. N., Eds. Photocatalytic Transformation of Solar Energy. Heterogeneous, Homogeneous and Organized Molecular Structures; Nauka: Novosibirsk, Soviet Union, 1991. (3) Fox, M. A.; Dulay, M. T. Chem. ReV. 1993, 93, 341. (4) Hoffman, M. R.; Martin, S. T.; Choi, W.; Bahnemann, D. W. Chem. ReV. 1995, 95, 69. (5) Linsebigler, A. L.; Lu, G.; Yates, J. T. Chem. ReV. 1995, 95, 735. (6) Fujishima, A.; Rao, T. N.; Tryk, D. A. J. Photochem. Photobiol. C: Photochem. ReV. 2000, 1, 1. (7) Pankove, J. I. Optical Processes in Semiconductors; Dover Publications: New York, 1971. (8) Luschik, C. B.; Luschik, A. C. Decay of Electronic Excitation with Formation of Defects in Solids; Nauka: Novosibirsk, Soviet Union, 1989. (9) Special Issue: Defects and Impurity Centers in Ionic Crystals: Optical and Magnetic Properties, Part 1. In J. Phys. Chem. Solids 1990, 51, No. 7. (10) Ryabchuk, V. K.; Burukina, G. V. SoV. J. Phys. Chem. 1991, 65, 1621. (11) Serpone, N.; Khairutdinov. R. F. In Semiconductor NanoclustersPhysical, Chemical, and Catalytic Aspects; Kamat, P. V., Meisel, D., Eds. Studies Surf. Sci. Catal. 1997, 103, 417. (12) Emeline A. V.; Kataeva G. V.; Litke A. S.; Rudakova A. V.; Ryabchuk V. K.; Serpone N. Langmuir 1998, 14, 5011. (13) Bluvshtein, I. M.; Nizhnikova, G. P.; Farberovich, O. V. SoV. Phys. Solid State 1990, 32, 548. (14) Kwok, C. K.; Aita, C. R. J. Appl. Phys. 1989, 66, 2756. (15) Sayama, K.; Arakawa, H. J. Photochem. Photobiol. A: Chem. 1996, 94, 67. (16) Gravie, R. C. J. Phys. Chem. 1965, 69, 1238. (17) Strekalovsky, V. N.; Polejaev, Yu. M.; Palguev, S. V. Oxides with Extrinsic Disorder; Nauka: Moscow, Soviet Union, 1987. (18) Harrison, O. E.; Melamed, N. T.; Subbarao, E. G. J. Electrochem. Soc. 1963, 110, 23. (19) Mikhailov, M. M.; Kuznetsov, N. Y. Inorg. Mater. 1988, 24, 656. (20) Mikhailov, M. M.; Dvoretsky, M. I.; Kuznetsov, N. Y. Inorg. Mater. 1984, 20, 449. (21) Zhao, Q.; Wang, X.; Cai, T. Appl. Surf. Sci. 2004, 225, 7. (22) Burukina, G. V. Ph.D. Thesis, Leningrad State University; Leningrad, Soviet Union, 1990. (23) Llopis, J. Phys. Status Solidi A 1990, 119, 661. (24) Arsenev, P. A.; Bagdasarov, Kh. S.; Niklas, A.; Ryazantsev, A. D. Phys. Status Solidi A 1980, 62, 395. (25) Bettinali, C.; Ferraresso, G.; Manconi, J. W. J. Chem. Phys. 1969, 50, 3957. (26) Emeline, A. V.; Panasuk, A. V.; Sheremetyeva, N. V.; Serpone, N. J. Phys. Chem. B. 2005, 109, 2785. (27) Liu, H.; Feng, L.; Zhang, X.; Xue, Q. J. Phys. Chem. 1995, 99, 332. (28) Hertl, W. Langmuir 1989, 5, 96. (29) Cutrufello, M. G.; Ferino, I.; Monaci, R.; Rombi, E.; Solinas, V. Top. Catal. 2002, 19, 225. (30) Andreev, N. S.; Emeline, A. V.; Khudnev, V. A.; Polikhova, S. A.; Ryabchuk V. K.; Serpone, N. Chem. Phys. Lett. 2000, 325, 288. (31) Pokrovski, K.; Jung, K. T.; Bell, A. T. Langmuir 2001, 17, 4297. (32) He, M.-J.; Ekerdt, J. G. J. Catal. 1984, 87, 238. (33) Bachiller-Baeza, B.; Rodriguez-Ramos, I.; Guerrero-Ruiz, A. Langmuir 1998, 14, 3556. (34) Tsuji, H.; Okamura-Yoshida, A.; Shishido, T.; Hattori, H. Langmuir 2003, 19, 8793. (35) (a) Volodin, A. M.; Cherkashin, A. E. Kinet. Katal. 1981, 22, 1227. (b) Volodin, A. M.; Cherkashin, A. E. Kinet. Katal. 1981, 22, 979. (c) Volodin, A. M.; Cherkashin, A. E.; Zakharenko, V. S. React. Kinet. Catal. Lett. 1979, 11, 221. (36) Kohno, Y.; Tanaka, T.; Funabiki, T.; Yoshida, S. Phys. Chem. Chem. Phys. 2000, 2, 2635. (37) Polikhova, S. A.; Andreev, N. S.; Emeline, A. V.; Ryabchuk, V. K.; Serpone, N. J. Phys. Chem. B 2004, 108, 2354.