Photocatalytic Reduction of NO by CO on Titanium Dioxide under ...

23 downloads 74 Views 203KB Size Report
Sep 12, 2007 - visible light. A reaction mechanism is proposed which includes the following stages: (a) NO photoadsorption along two parallel routes, by e- ...
14440

J. Phys. Chem. C 2007, 111, 14440-14447

Photocatalytic Reduction of NO by CO on Titanium Dioxide under Visible Light Irradiation Andrei A. Lisachenko,† Ruslan V. Mikhailov,† Lev L. Basov,† Boris N. Shelimov,*,‡ and Michel Che§ V.A. Fock Institute of Physics, St.-Petersburg State UniVersity, Ul’yanoVskaya St. 1, 198504 St.-Petersburg, Russia, N.D. Zelinsky Institute of Organic Chemistry, Russian Academy of Sciences, Leninskii Pr. 47, 119991 Moscow, Russia, Laboratoire de Re´ actiVite´ de Surface, UMR 7609-CNRS, UniVersite´ Pierre et Marie Curie-Paris 6, 4, Place Jussieu, 75252 Paris Cedex 05, France; and Institut UniVersitaire de France ReceiVed: March 18, 2007; In Final Form: May 14, 2007

It is shown for the first time that the photoassisted reduction of NO by CO into N2 and N2O can occur on TiO2 catalysts upon visible light irradiation (λ > 380 nm) at room temperature. The selectivity of photoreduction of NO into N2 reaches 90-95%. The CO2 formed which predominantly remains on the surface can be quantitatively desorbed after completion of the photoreaction by heating TiO2 to ∼500 K. The rates of NO consumption and product accumulation remain virtually constant upon successive admissions of the CONO mixture thus indicating a high stability of catalyst activity. It is found that the quantum yield of NO photoreduction by CO is considerably greater for visible light irradiation (λ ) 405 + 436 nm) than for UV irradiation (λ ) 365 nm). Experiments with 18O-enriched NO revealed that, under visible light irradiation, an intense oxygen isotopic exchange between NO and TiO2 develops. The photocatalytic reaction requires the presence in nonstoichiometric TiO2-x of electron-donor centers (Ti3+ ions, F and F+ centers) able to absorb visible light. A reaction mechanism is proposed which includes the following stages: (a) NO photoadsorption along two parallel routes, by e- capture from the electron-donor center by NO to yield adsorbed NO- species and by NO interaction with the hole O- to give adsorbed nitrite NO2-; (b) reduction of NO- into N2O and further into N2; and (c) photoreduction of adsorbed nitrite NO2- into N2O and further into N2 by CO which regenerates the donor centers. It is assumed that the photoinduced oxygen heteroexchange proceeds via adsorbed nitrite complexes.

Introduction Titanium dioxide is widely used as a photocatalyst in environmental catalysis, because it is highly active in degrading many organic and inorganic pollutants, is resistant to photocorrosion, at the same time, is biologically and chemically inert, and is of low cost.1,2 However, the use of TiO2 in photocatalysis requires a source of UV light. The band gap of anatase, the most frequent form of TiO2 used, is relatively large (∼3.2 eV) requiring UV radiation at λ < 385 nm to generate the hole and electron centers active in photocatalysis. This practically rules out the use of sunlight whose UV fraction (λ < 400 nm) of the total solar radiation amounts to less than 4%.3 Sensitization of TiO2 to visible light would allow about 40% of solar energy to be utilized. Different approaches to visible light sensitization of TiO2-based photocatalysts have been developed during the past decades. At present, several techniques are used to introduce active sites in or at the surface of TiO2 able to absorb visible light. These include the doping of TiO2 with noble metal ions (Pt, Ir, Rh, Au, and Pd),4 transition metal ions (Fe, V, and Cr),5-8 and nonmetal ions (C, N, F, and P). The last method, first suggested by Asahi et al.,9 was later exploited by other groups.10-15 It has been shown that N-doping can occur inadvertently during * Corresponding author. E-mail: [email protected]. † St.-Petersburg State University. ‡ Russian Academy of Sciences. § UMR 7609-CNRS.

the preparation of titanium dioxide depending on experimental conditions. Che and Naccache16 prepared two series of samples by hydrolysis of TiCl4 followed by neutralization by either ammonia or sodium hydroxide to precipitate TiO2. The resulting solids were air treated at 300-400 °C. Only the TiO2 samples prepared with ammonia exhibited complex EPR spectra which were assigned to NO22- species. It was suggested that during the air treatment the Lewis basic ammonia molecules adsorbed during preparation on Lewis acidic Ti4+ ions were transformed into NO. The latter further reacted with basic O2- ions to give the observed NO22- species, explaining the N-doping of TiO2. Other approaches to visible light sensitization of TiO2 worth mentioning include TiO2 pretreatment in plasmas of argon and nitrogen at 300 K17 or of hydrogen at 673 K.18 In the latter case, photo-oxidation of NO to NO3- by an air flow under visible light irradiation (430-570 nm) at 300 K was presumably initiated by light absorption by surface defects of TiO2 (oxygen vacancies filled with e-) whose energy levels are localized within the band gap. As shown earlier,19 TiO2 Degussa P-25 reduced in CO or H2 at 850 K absorbs light in the visible and near IR region (400 < λ < 2500 nm) due to mobile and localized e- (e- trapped in oxygen vacancies and Ti3+ ions). A UPS study20 revealed that the electron donor states in TiO2 fill the band gap up to the Fermi level and that oxide reduction increases the density of surface states. Part of electron centers (up to 75%) of reduced TiO2 rapidly react with gaseous O2, NO, and N2O at 300 K to partially discolor the sample.19 One may assume that this

10.1021/jp072158c CCC: $37.00 © 2007 American Chemical Society Published on Web 09/12/2007

Reduction of NO by CO discoloration is faster and deeper upon visible light irradiation due to photon energy absorbed by electron centers. The present work aims at elucidating whether electron defects of partially reduced TiO2 can initiate the reduction of NO by CO under visible light irradiation. The choice of this test reaction appears logical, because both gases produced in burning processes are present in polluted atmosphere. Upon UV photodecomposition of pure NO on bulk TiO2 (in absence of reducing gas), N2O is predominantly formed with N2 as a minor product21-25 (N2/N2O molecular ratio ∼ 0.38). Hence, TiO2 is not ideal to photocatalytically remove NO from air, because N2O itself is a regulated pollutant. Reducing agents such as NH3 or CO are frequently used to decrease the N2O yield upon photoassisted reduction of NO. Upon UV irradiation of TiO2 in NH3-NO mixtures under static conditions, the N2/N2O molecular ratio is ∼2,22 and under NH3-NO-O2 flow, the selectivity to N2 considerably increases, reaching 96% at 83% NO conversion.26 With CO as reductant, the selectivity to N2 is not so high: under UV irradiation of TiO2 under NO-CO flow, the N2/N2O molecular ratio is 1.86 even at excessive reductant content (CO/NO ) 4).25 To the best of our knowledge, the photocatalytic reduction of NO by CO under visible light irradiation has not been reported so far. Experimental Methods Photocatalysis experiments were performed with Degussa P-25 TiO2 (S ) 50 m2 g-1) and a greaseless vacuum set up with base pressure 380 nm were selected by means of a JS-10 color filter (Figure 1, curve 3) with transmittance 380 nm. The inset shows the kinetics of fast dark CO adsorption on TiO2 after admission of a CO: NO ) 3:1 mixture. The onset of light irradiation is marked by the arrow.

Heating reduced TiO2 sample in oxygen at 750 K for 3 h does not shift the Fermi level, but decreases the intensity of photoelectron spectra by ∼ 40% (Figure 2, curve 1). Thus, even after the oxidative treatment, TiO2 remains nonstoichiometric, oxygen-deficient oxide. Kinetics of Photocatalytic NO + CO Reaction on TiO2 under Visible Light Irradiation. Admission of a CO-NO mixture onto reduced TiO2 is accompanied by fast dark adsorption of CO (Figure 3). At initial CO pressure of 0.45 Torr, approximately half of admitted CO is adsorbed. The kinetics of dark CO adsorption during the first 80 s after admission is shown in the inset of Figure 3. As reported earlier,29 three IR bands appear on exposing TiO2 to CO. The band at 2131 cm-1 remains after evacuation but disappears upon heating to 70 °C. It is assigned to CO adsorbed on Ti3+ ions or in the vicinity of oxygen vacancies. The 2192 and 2209 cm-1 bands are assigned to CO weakly adsorbed on coordinatively unsaturated (cus) Ti4+ ions. Unlike CO, NO does not noticeably adsorb on TiO2: its pressure stays nearly constant in the absence of irradiation. The amount of adsorbed NO is estimated to be ∼10-5 monolayer. Upon visible light irradiation, the pressure of NO drops to zero in 50 min due to photon-induced adsorption on TiO2.

(1)

At t ≈ 500 min, the N2/N2O molecular ratio of ∼15 corresponds to a rather high selectivity of NO reduction toward N2. There is a good nitrogen mass balance in the gas phase: -∆P(NO) ) 0.17 Torr compared to 2[∆P(N2O) + ∆P(N2)] ) 0.16 Torr, meaning that N-containing species do not remain adsorbed in appreciable amounts at the end of the reaction. This is also in line with TPD data: after evacuation of the irradiated sample at 300 K, no nitrogen-containing product is found to desorb upon heating to 900 K. During the first 20-25 min of irradiation, the CO pressure slightly increases, probably because of the small rise of sample temperature due to the light beam resulting in partial CO desorption. Upon further irradiation, the CO pressure slowly decreases, but CO2 produced by reaction (1) remains on the surface. CO2 can be completely desorbed upon heating and evacuation of TiO2 at 500 K (Figure 4). There are two peaks at 350 and 420 K in the TPD spectrum, apparently associated with different forms of adsorbed CO2.29 The amount of CO2 desorbed equals the amount of NO consumed in accordance with reaction (1) stoichiometry. It is noteworthy that during the reaction the CO pressure decreases much less than that of NO, i.e., the reaction occurs mainly at the expense of CO adsorbed. The adsorption sites remain occupied by the carbonate species formed, and consequently, the CO pressure decreases with irradiation time only insignificantly. When the photoreaction is carried out at 350 K, CO2 evolves simultaneously with gas-phase N2 and N2O, as shown in Figure 5. A remarkable feature of reduced TiO2 is its very stable photocatalytic activity, illustrated by Figure 6, which shows the kinetics of the NO + CO photoreaction for two successive admissions of reaction mixtures of close compositions. The close similarity of the kinetic curves suggests the absence of noticeable poisoning of catalyst in these conditions.

Reduction of NO by CO

Figure 5. Kinetics of the CO-NO photoreaction on prereduced TiO2 at 350 K. Irradiation with λ > 380 nm.

J. Phys. Chem. C, Vol. 111, No. 39, 2007 14443

Figure 8. Kinetics of the photoreaction upon successive admissions of NO and CO on TiO2. (1) NO admission, irradiation, and evacuation; (2) CO admission, irradiation, and evacuation; (3) NO readmission and irradiation. Light irradiation with λ > 380 nm.

TABLE 1: Efficiency (E) and Quantum Yield (Y) of NO Photoreduction by CO on Reduced TiO2 upon UV and Visible-Light Irradiations

Figure 6. Comparative kinetics of the NO photoreduction by carbon monoxide on prereduced TiO2 for two successive CO + NO admissions and irradiations. After the first irradiation the sample was evacuated at 300 K. Irradiation with λ > 380 nm.

Figure 7. Kinetics of the NO photoreduction by carbon monoxide on (1) oxidized and (2) prereduced TiO2. Irradiation with λ > 380 nm.

The effect of the TiO2 reduction level on the photoreaction is given in Figure 7 with the kinetics of oxidized and reduced samples. The N2 pressure is noticeably higher upon irradiation of reduced TiO2 even at lower initial NO pressure. Accordingly, the amount of CO2 formed by the photoreaction is greater on reduced TiO2: according to TPD, 4.6 × 1017 CO2 molecules are formed on reduced TiO2 vs 2.5 × 1017 molecules on oxidized TiO2 after 230 min of irradiation. These data clearly demonstrate the favorable effect of TiO2 reduction on the photoreaction rate. It should be emphasized that heating TiO2 in oxygen at 760 K

λ, nm

light intensity, photon cm-2 s

E

Y

404 + 436 365

4 × 1015 3.2 × 1015

5 × 10-4 9 × 10-4

6 × 10-3 1 × 10-3

does not eliminate completely the photoactivity in the visible region. This correlates with the stability of electron donor sites of TiO2 under these conditions (see above “Characterization of TiO2 by UV-vis Spectroscopy and UPS”). Upon UV irradiation (λ ) 365 nm) of TiO2 in the intrinsic absorption region, the kinetics of NO photoreduction by CO is similar to that shown in Figure 3. It is important to assess the photoreaction efficiency under UV and visible light irradiations. The photoreaction efficiency (E) is calculated by the following equation: E(λ) ) N(N2)/N(incid), where N(N2) is the number of N2 molecules formed upon incidence of N(incid) photons with wavelength λ on the sample. The quantum yield of the reaction is calculated by the following equation: Y(λ) ) N(N2)/ N(abs), where N(abs) is the number of photons with wavelength λ absorbed by TiO2 with the account of reflection coefficient F determined from diffuse reflectance spectra (Figure 1). The results are summarized in Table 1. The photoreaction efficiency E is higher in the intrinsic absorption region due to practically complete absorption of UV light, whereas the quantum yield (Y) per one photon absorbed by TiO2 is considerably greater in the visible region. Kinetics of the Photoreaction upon Successive Admissions of NO and CO on TiO2. Figure 8 displays the results obtained when NO and CO were successively admitted on TiO2 and irradiated with visible light. The light intensity and its spectral composition were the same for all stages of the run. In stage 1, the irradiation was carried out after NO admission on reduced TiO2. The NO photoadsorption rate is the same as that upon irradiation in the CO-NO mixture at equal initial NO pressures and reduction levels of TiO2. During the first few minutes of irradiation, small amounts of N2 and N2O are evolved (3-4% of the amount of NO photoadsorbed) and then the N2 and N2O pressures level off. After evacuation of the sample at 300 K, CO was admitted onto TiO2. After the adsorption equilibrium was reached, the light was switched on (Figure 8, stage 2). Upon irradiation, the CO pressure decreases; the N2O pressure reaches a maximum

14444 J. Phys. Chem. C, Vol. 111, No. 39, 2007

Lisachenko et al.

Figure 9. Dynamics of oxygen isotopic exchange of NO with prereduced TiO2 in the presence of the 15N18O-15N16O-14N16O mixture under irradiation with λ > 380 nm.

soon after the onset of irradiation; and the N2 pressure continuously grows. At the end of stage 2, nitrogen becomes a predominant product: after 120 min of irradiation, the P(N2)/ P(N2O) ratio is ∼25. CO2 is not detected in the gas phase. Thus, one can conclude that N2 and N2O are formed in stage 2 by photoreaction of CO with surface species produced by NO photoadsorption. For a long enough irradiation time, the amount of N2 evolved approaches half the amount of NO photoadsorbed. After evacuation at 300 K, the sample was irradiated again in the presence of NO alone (Figure 8, stage 3). The kinetics of NO photoadsorption is identical to that of stage 1 within the accuracy of the measurements, and approximately the same amounts of N2 and N2O are evolved. This implies that the remaining surface CO2 does not block the active sites. Our experiments with 15N- and 18O-enriched NO revealed that, under visible light irradiation, an intense oxygen isotopic exchange between NO and TiO2 develops. The dynamics of this exchange in the presence of a 15N18O-15N16O-14N16O mixture is given in Figure 9. No isotopic exchange occurs in the absence of irradiation: the fraction of gas phase 18O (right Y axis in Figure 9) is constant. Immediately after the beginning of irradiation, the 15N18O pressure and 18O fraction rapidly decrease, whereas the 15N16O pressure goes through a maximum. Thus, NO photoadsorption is accompanied by the following oxygen isotopic exchange:

N18O(g) + 16O(lat) a 15N16O(g) + 18O(lat)

15

(2)

Although oxygen isotopic exchange between N18O and16Olat upon UV irradiation of oxidized and reduced TiO2 has been known for a long time,21 there has been no reaction mechanism proposed. In the next series of experiments, NO and CO were admitted on reduced TiO2 in reverse order (Figure 10). In stage 1, CO was first admitted onto TiO2 which was then irradiated by visible light. The irradiation causes only a slight increase in CO pressure apparently due to a slight temperature rise (∼2 °C) of the sample. After irradiation, the sample was evacuated at 293 K, and the TPD spectrum was recorded. Desorption of CO2 in small amounts (by a factor of 30-50 smaller than that after irradiation in NO-CO mixtures, see Figure 4) was observed upon heating. Hence, photoreduction of reduced TiO2 by CO under visible light is negligible. In stage 2 after evacuation at 293 K, TiO2 was irradiated in the presence of NO. The kinetics of NO photoadsorption is very similar to that shown in Figure 8; that is, preliminary visible

Figure 10. Kinetics of NO photoadsorption on TiO2 after irradiation in the CO atmosphere with λ > 380 nm. (1) CO admission, irradiation, and evacuation; (2) NO admission and irradiation.

light irradiation in CO does not affect the dynamics of NO photoadsorption of the next stage. NO reduction products appear in very small quantities and, similarly to the run of Figure 8, amounts to only 2-3% of NO photoadsorbed. Discussion Catalytic Character of the Photoreduction of NO by CO. The present data show that the photocatalytic activity of TiO2 appearing upon visible light irradiation is associated, at least partially, with the presence of localized electron-donor defects (Ti3+, F+, and F centers) which absorb light in the visible (Figure 1). This conclusion is supported in particular by the larger amounts of N2 (Figure 7) and CO2 (TPD measurements, see “Kinetics of Photocatalytic NO + CO Reaction on TiO2 under Visible Light Irradiation”) produced upon irradiation of reduced TiO2. This is in line with UPS data of Figure 2 which evidence the higher concentration of donor centers in reduced TiO2. The photocatalytic activity of TiO2 cannot be suppressed even after rather severe treatment in oxygen at 723 K for 3 h which does not restore the stoichiometry of TiO2. It is of importance to compare the amount of electron-donor centers initially present in reduced TiO2 and that of N2 and CO2 produced upon irradiation. The concentration of donor centers was derived from the amount of CO2 formed during preliminary UV irradiation in CO of oxidized TiO2 as described in the Experimental Methods. It was accepted that the release of one gas-phase CO2 molecule is accompanied by the incorporation of two electrons into TiO2 lattice and by the formation of one oxygen anion vacancy according to the equation: O2-lat + CO f CO2v + 2e-. It is found that 0.45 µmol of CO2 (or 0.9 µmol of electron centers) per g of TiO2 is produced for 1-2 h of UV irradiation. On the other hand, 4 µmol of N2 and 8 µmol of CO2 per g of TiO2 are produced by visible light irradiation in 1:1 CO-NO mixture for 400 min, i.e., the [N2]/[e-] and [CO2]/ [e-] ratios are ∼4.5 and ∼9, respectively. Thus, NO reduction by CO under visible light irradiation is catalytic in the sense that, under given experimental conditions, each electron-donor center created by preliminary photoreduction of TiO2 participates, at least, 4.5-9 times in the formation of N2 and CO2. Taking into account the high stability of the photocatalyst, much

Reduction of NO by CO

J. Phys. Chem. C, Vol. 111, No. 39, 2007 14445

higher [N2]/[e-] and [CO2]/[e-] ratios can be obtained upon successive irradiations of TiO2 in several fresh portions of the reaction mixture (Figure 6). It is of primary importance however that, even for a single cycle, the above ratios considerably exceed unity. Mechanism of the Photocatalytic Reduction of NO by CO. The establishment of the mechanism has to adequately account for the following data: NO photoadsorption on TiO2, formation of reaction products (N2O, N2, and adsorbed CO2), oxygen isotopic exchange between N18O and the TiO2 lattice under irradiation, production of N2O and N2 from photoadsorbed NO under irradiation in CO, as well as the shape of the kinetic curves. In what follows, the successive stages of the proposed mechanism are considered. Photoadsorption of NO. Visible light irradiation of TiO2 in NO alone gives strongly adsorbed species which cannot be removed by evacuation at room temperature. Photoadsorption in the absence of CO is accompanied by release of only small amounts of N2O and N2 (Figure 8). It should be stressed here that the number of photoadsorbed NO molecules considerably exceeds the number of electron donor sites accumulated during preliminary UV activation of TiO2. For example, in the run of Figure 8, an amount of 5.3 µmol of NO is photoadsorbed on 1 g of TiO2, which is greater by a factor of 6 than the number of accumulated donor species. In order to explain this fact, one has to assume two parallel processes, one of which consumes electron donor species, the other which proceeds without participation of these species. NO photoadsorption, with consumption of donor species, involves the following sequence of reactions:

where O2- is an oxide ion of TiO2 and NO2- denotes surface linear (I) or bidentate (II) nitrites:

Reaction 5 can proceed upon absorption of a photon in the 2.7 e Ehν e 3.2 eV visible region. Two intense lines emitted by the Hg lamp (at 436 and 405 nm) fall within this interval (Figure 1). As seen from the energy diagram of Figure 2, the density of states is high enough both for oxidized and reduced TiO2. The formation of various types of nitrites and nitrates (NO3-) on TiO2 in presence of NO was detected by IR and reported in several papers.31-33,35 The intensity of IR absorption bands assigned to these species gradually increases with time in the absence of irradiation after admission of NO at 293 K, whereas at 110 K, nitrites or nitrates were not observed at all.34 The rate of NO2- and NO3- accumulation considerably increases under UV irradiation suggesting that formation of these species is activated and that the activation barrier can be overcome thanks to photon energy. We suggest that, under visible light irradiation, nitrites are reduced by the reverse reaction in equilibrium (7). An alternative mechanism for NO photoadsorption without consumption of electron-donor centers can be described by the following sequence of reactions: hν

NO(g) a NO(a) hν

NO(a) + Ti3+(e-) {\} NO-(a) + Ti4+

(3) (4)

where e- stands for F+ and F centers, i.e., oxygen anion vacancies filled with one or two e-, respectively. NO adsorption resulting in nitrosyl complexes NO(a) can occur on Ti4+cus ions and/or F+ and F centers. The formation of nitrosyl complexes upon NO admission on TiO2 at room and lower (110 K) temperatures was detected by IR and reported in several studies.31-34 As a result of visible light absorption by a donor center, an electron transfer to NO(a) can occur (reaction 4). To some extent, this reaction can proceed without irradiation as evidenced by the decrease of donor center absorption bands after NO admission on reduced TiO2.19 NO- and N2O2 species were detected by IR (absorption bands at 1170 and 1135 cm-1, respectively).35 The formation of NO- is also proved by EPR data:36,37 upon NO admission on UV irradiated TiO2 and Ti/ ZSM-5, the intensity of the EPR signal of electron donor centers Ti3+(e-) decreases due to oxidation of the centers by NO (reaction 4). NO photoadsorption, accompanied by formation of electron centers, occurs as follows: hν

Ti4+ + O2- {\} Ti3+ + O-

(5)

O- + NO(a) f NO2-(a)

(6)

which can be grouped into the overall reaction: hν

} NO2-(a) + Ti3+ Ti4+ + O2- + NO(a) {\ hν

F+ + O2- 98 F + O-

(8)

F + NO(a) a F+ + NO-(a)

(9)

O- + NO(a) f NO2-(a)

(6)

with the overall process written as hν

2 NO(a) + O2- {\} NO-(a) + NO2-(a)

(10)

Reaction 8 first suggested by Crawford was found to occur in Al2O3.38 Later, it was also observed in other wide band gap oxides.39 It was shown38 that the formation of O- hole centers by reaction 8 is possible when the lowest empty orbital of an excited F+ center lies below the top of the valence band. Since photoadsorption of NO by reactions 7 and 10 is not accompanied by diminution of the number of donor centers accumulated during preliminary photoreduction, the number of NO photoadsorbed may substantially exceed the number of these centers. Thus, actually NO photoadsorption is a disproportionation reaction of NO:NO(a) is reduced to NO-(a) by donor centers via reaction 4, whereas NO(a) is oxidized into NO2-(a) by O2- ions via reaction 7. Formation of N2O and N2 upon Photoadsorption of NO. Figure 8 shows that at the early stage of NO photoadsorption small amounts of gaseous N2O and N2 are released in the absence of CO. It appears that N2O is produced via NO-(a)

NO-(a) + NO f N2O + O-

(11)

with the O- species rapidly reacting with NO molecules

(7)

O- + NO(a) f NO2-(a)

(6)

14446 J. Phys. Chem. C, Vol. 111, No. 39, 2007

Lisachenko et al.

The overall reaction of N2O production (reactions 4, 6, and 11) can be written as

SCHEME 1: Photocatalysis Reaction Cycle

Ti3+(e-) + 3 NO(a) f Ti4+ + NO2-(a) + N2O (12) Most likely, N2 is formed by N2O reduction by donor centers, i.e., N2O is an intermediate in the reduction of NO

N2O + Ti3+(e-) f N2 + O- + Ti4+

(13)

The overall reaction can be written as the sum of reactions 6, 12, and 13

2 Ti3+(2e-) + 4 NO(a) f 2 Ti4+ + 2 NO2-(a) + N2

(14)

Following the stoichiometry of reactions 12 and 14, reduction of one photoadsorbed NO mole can generate 1/4 N2 mole or 1/3 N2O mole. However, the observed yields of N2 and N2O are very small and in absence of CO are only ∼2.5 and ∼0.8%, respectively, from the amount of NO photoadsorbed (Figure 8). Hence, the major route for NO photoadsorption is reaction 7 or 10 which do not produce N2 and N2O and do not consume electron-donor centers. Therefore, the amount of NO photoadsorbed largely exceeds the initial amount of available donor centers in TiO2. A reaction scheme for UV-assisted decomposition of NO on TiO2 was proposed earlier, in which a key step is the interaction of NO-(a) with hole centers h+(O-)

e- + NO f NO-(a)

(15)

UV

TiO2 98 h+(O-) + e-

(16)

NO-(a) + O- f N(a) + O(a) + O2-

(17)

Pichat21

Reaction 17 proposed by Courbon and was later supported by Cant and Cole,22 who suggested N2 and N2O to be formed via adsorbed atomic nitrogen

N(a) + NO f N2O

(18)

2 N(a) f N2

(19)

This situation can hardly be applied to the case of visible light irradiation of TiO2 in view of the following considerations: (a) nitrite complexes NO2-(a) are not considered although they play a key role in NO photoadsorption and were detected by IR; (b) mass spectrometric analysis does not reveal O2 and NO2 in the gas phase. The formation of atomic oxygen by reaction 17 is thus doubtful [It should be noted that the formation of O2 and NO2 upon UV irradiation of TiO2 in NO has not been reported in the literature so far.20-24]; (c) N2O and N2 are assumed to be formed in parallel by reactions 18 and 19; however, this is not consistent with the shapes of kinetic curves of these products characteristic of a consecutive reaction (Figures 3 and 5-7). Oxygen Isotopic Exchange between N18O and 16O of TiO2. It is suggested that oxygen isotopic exchange of the N18O-Ti16O2 system (Figure 9) occurs via NO2-(a) species by the following reaction: hν

} (N18O16O)-(a) + Ti3+(e-) {\} N18O(a) + Ti4+ + 16O2- {\ hν N16O(a) + Ti4+ + 18O2- (20) This reaction considered above in the discussion of the mechanism of NO photoadsorption (reaction 4) assumes a

reversible electron transfer from the excited electron-donor center to NO2-(a). It is worth noting that a slow isotopic exchange in the N18O-Ti16O2 system takes place in the absence of irradiation both on oxidized and UV-activated TiO2, apparently due to e- localized in shallow traps.21 Under visible light irradiation, the exchange rate dramatically increases, probably because additional e- localized in deeper traps are involved in the process. Reaction Scheme upon Visible Light Irradiation of TiO2 in CO-NO Mixtures. Upon visible light irradiation of TiO2 in CO-NO mixture, the N2 and N2O yields are much higher than those in the absence of CO, the major part of N2 being released into the gas phase after completion of NO photoadsorption (Figures 3 and 5-7). A similar behavior is observed when TiO2 is irradiated in pure CO after preliminary NO photoadsorption (Figure 8, stage 2). Since, as shown above, the main photoadsorption product is NO2-(a) (reaction 7), one may assume that N2 and N2O are formed under irradiation in CO via reduction of NO2-(a) by the following sequence of reactions: hν

NO2-(a) + Ti3+ {\ } NO(a) + Ti4+ + O2hν hν

(7)

Ti4+ + O2- {\} Ti3+ + O-

(5)

Ti4+ + O- + CO f Ti3+(e-) + CO2

(21)



NO(a) + Ti3+(e-) {\} NO-(a) + Ti4+

(4)

NO-(a) + NO f N2O + O-

(11)

N2O + Ti3+(e-) f N2 + O- + Ti4+

(13)

O- + NO(a) f NO2-(a)

(6)

The above sequence of reactions can also be presented by the reaction cycle shown in Scheme 1. The overall process according to this reaction sequence can be written as hν

NO + CO 98 CO2 + 1/2 N2

(1)

i.e., as the selective photocatalytic reduction of NO by CO into N2. At the final stages of the runs shown in Figures 3 and 5-7,

Reduction of NO by CO NO2-(a) is practically completely reduced into N2 with a small admixture of N2O. This is evidenced by the nitrogen mass balance in the gas phase (see section “Kinetics of Photocatalytic NO + CO Reaction on TiO2 under Visible Light Irradiation”) and TPD data. This conclusion is consistent with IR data.31 The above scheme suggests the formation of adsorbed CO2(a) or carbonate species. It has been found29 that carbonate species (bridged, bidentate and monodentate) on TiO2 surface are rather thermally stable: according to IR data, these species are destroyed only partially upon heating to 340 K. This may explain why gaseous CO2 is not detected upon irradiation at room temperature and only appears upon irradiation at 340 K (Figure 5). Conclusions We report here for the first time on photocatalytic reduction of NO by CO into N2 and N2O on TiO2 Degussa P-25 under visible light irradiation (λ > 380 nm) at room temperature under static conditions. The remarkable features of the photoreaction are the high selectivity of photoreduction of NO to N2 (9095%) and very stable activity of the TiO2 photocatalyst. Similar kinetic curves are obtained with NO-CO mixtures successively admitted onto the photocatalyst indicating no noticeable poisoning in these experiments. The efficiencies of NO photoreduction by UV and visible light are compared. The photoreduction efficiency per one incident photon is higher in the intrinsic absorption region due to practically complete absorption of UV light, whereas the quantum yield per one photon absorbed by TiO2 is much greater in the visible region. It is concluded that the photocatalytic activity of nonstoichiometric TiO2-x in the visible region is associated with the presence of localized electron-donor centers (Ti3+ ions, F and F+ centers). The amount of such centers in the initial TiO2-x sample is much smaller than the amount of N2 and CO2 produced by the photoreaction even in case of single admission of NO-CO mixture; that is, each electron-donor center can participate up to approximately 4.5 and 9 times in the formation of N2 and CO2, respectively. A reaction mechanism is proposed which includes the following stages: (a) NO photoadsorption along two parallel routes: by e- capture from the electron-donor center by NO to yield adsorbed NO- species and by NO interaction with the hole Oto give adsorbed nitrite NO2-; (b) reduction of NO- into N2O and further into N2; and (c) photoreduction of adsorbed nitrite NO2- into N2O and further into N2 by CO which regenerates the donor centers. Finally, it should be underlined that the reaction scheme proposed in this work allows, in our opinion, to completely and self-consistently describe all the experimental findings observed. The scheme is also in accord with rather abundant literature data on IR and EPR study of the NO-TiO2 system. Acknowledgment. This work was supported by INTAS under Grant 03-51-6088 and by the Russian Foundation for Basic Research under Grant 05-03-32490. References and Notes (1) Linsebigler, A. L.; Lu, G.; Yates, J. T., Jr. Chem. ReV. 1995, 95, 735.

J. Phys. Chem. C, Vol. 111, No. 39, 2007 14447 (2) Fujishima, A.; Zhang, X. C.R. Chimie 2006, 9, 750. (3) Zhao, J.; Chen, C. Ma, W. Top. Catal. 2005, 35, 269. (4) Zang, L.; Macyk, W.; Lange, C.; Maier, W. F., Antonius, C.; Meissner, D.; Kisch, H. Chem. Eur. J. 2000, 6, 379. (5) Serpone, N.; Lawless, D.; Disdier, J.; Herrmann, J.-M. Langmuir 1994, 10, 643. (6) Takeuchi, M.; Yamashita, H.; Matsuoka, M.; Anpo, M.; Hirao, T.; Itoh, N.; Iwamoto, N. Catal. Lett. 2000, 67, 135. (7) Yamashita, H.; Harada, M.; Misaka, J.; Takeuchi, M.; Neppolian, B.; Anpo, M. Catal. Today 2003, 84, 191. (8) Klosek, S.; Raftery, D. J. Phys. Chem. B 2001, 105, 2815. (9) Asahi, R.; Morikawa, T.; Ohwaki, T.; Aoki, K.; Taga, Y. Science 2001, 293, 269. (10) Kumar, S.; Fedorov, A. G.; Gole, J. L. Appl. Catal. B 2005, 57, 93. (11) Kobayawa, K.; Murukami, Y.; Sato, Y. J. Photochem. Photobiol. A 2005, 170, 177. (12) Sato, S.; Nakamura, R.; Abe, S. Appl. Catal. A 2005, 284, 131. (13) Belver, C.; Bellod, R.; Stewart, S. J.; Requejo, F. G.; FernandezGarcia, M. Appl. Catal. B 2006, 65, 309. (14) Morawski, A. W.; Janus, M.; Tryba, B.; Inagaki, M.; Kalucki, K., C. R. Chim. 2006, 9, 800. (15) Yin, S.; Ihara, K.; Aita, Y.; Komatsu, M.; Sato, T. J. Photochem. Photobiol. A 2006, 179, 105. (16) Che, M.; Naccache, C. Chem. Phys. Lett. 1971, 8, 45. (17) Yamada, K.; Nakamura, H.; Matsushima, S.; Yamane, ; Haishi, T.; Ohira, K.; Kumada, K. C. R. Chim. 2006, 9, 788. (18) Nakamura, I.; Negishi, N.; Kutsuna, S.; Ihara, T.; Sugihara, S.; Takeuchi, K. J. Mol. Catal. A 2000, 161, 205. (19) Lisachenko, A. A.; Kuznetsov, V. N.; Zakharov, M. N.; Mikhailov, R. V. Kinet. Catal. 2004, 45, 189. (20) Lisachenko, A. A.; Mikhailov, R. V. Tech. Phys. Lett. 2005, 31, 21. (21) Courbon, H.; Pichat, P. J. Chem. Soc. Faraday Trans. 1 1984, 80, 3175. (22) Cant, N. W.; Cole, J. R. J. Catal. 1992, 134, 317. (23) Lim, T. H.; Jeong, S. M.; Kim, S. D.; Gyenis, J. J. Photochem. Photobiol. A 2000, 134, 209. (24) Zhang, J.; Ayusawa, T.; Minagawa, M.; Kinugawa, K.; Yamashita, H.; Matsuoka, M.; Anpo, M. J. Catal. 2001, 198, 1. (25) Bowering, N.; Walker, G. S.; Harrison, P. G. Appl. Catal. B 2006, 62, 208. (26) Teramura, K.; Tanaka, T.; Yamazoe, S.; Arakaki, K.; Funabiki, T. Appl. Catal. B 2004, 53, 29. (27) Lisachenko, A. A.; Chihachev, K. S.; Zakharov, M. N.; Basov, L. L.; Shelimov, B. N.; Subbotina, I. R.; Che, M.; Coluccia, S. Top. Catal. 2002, 20, 119. (28) Lisachenko, A. A.; Aprelev, A. M. Tech. Phys. Lett. (Pis’ma V ZhTF) 1995, 21, 9. (29) Liao, L.-F.; Lien, C.-F.; Shieh, D.-L.; Chen, M.-T.; Lin, J.-L. J. Phys. Chem. B 2002, 106, 11240. (30) Emanuel, N. M.; Knorre, D. G. Textbook of Chemical Kinetics, Izd. Vysshaya Shkola; Higher Education Publishing House: Moscow, 1984. Emanuel, N. M.; Knorre, D. G. Homogeneous Reactions. Chemical Kinetics; Wiley: New York, 1973. (31) Alekseev, A. V.; Filimonov, V. N. Russ. J. Phys. Chem. 1984, 58, 2023. (32) Ramis, G.; Busca, G.; Lorenzelli, V.; Forzatti, P. Appl. Catal. 1990, 64, 243. (33) Dines, T. J.; Rochester, C. H.; Ward, A. M. J. Chem. Soc. Faraday Trans. 1 1991, 87, 643. (34) Rusu, C. N.; Yates, J. T., Jr. J. Phys. Chem. B 2000, 104, 1729. (35) Hadjiivanov, K.; Kno¨zinger, H. Phys. Chem. Chem. Phys. 2000, 2, 2803. (36) Primet, M.; Che, M.; Naccache, C.; Mathieu, M. V.; Imelik, B. J. Chim. Phys.-Chim. Biol. 1970, 67, 1629. (37) Zhang, S.; Fujii, N.; Nosaka, Y. J. Mol. Catal. A 1998, 129, 219. (38) Crawford, J. H. Semicond. Insulat. 1983, 5, 599. (39) Kuznetsov, V. N.; Lisachenko, A. A. Rus. J. Phys. Chem. 1991, 65, 1328.