Photodegradation of Benzoic Acid over Metal-Doped TiO2 - Industrial

Apr 4, 2006 - In recent years, attention has been diverted to metal-doped TiO2, ... was added dropwise to the above solution under stirring for 30 min...
0 downloads 0 Views 305KB Size
Ind. Eng. Chem. Res. 2006, 45, 3503-3511

3503

Photodegradation of Benzoic Acid over Metal-Doped TiO2 Jinkai Zhou,† Yuxin Zhang,† X. S. Zhao,† and Ajay K. Ray*,‡ Department of Chemical and Biomolecular Engineering, National UniVersity of Singapore, 10 Kent Ridge Crescent, Singapore 119260, Singapore, and Department of Chemical and Biochemical Engineering, UniVersity of Western Ontario, London, Canada N6A 5B9

A modified sol-gel method was employed to synthesize metal-doped TiO2 with varied dopant concentrations using titanium butoxide and metal nitrate hydrate as the precursors and hydrothermal posttreatment. The prepared samples were characterized by XRD, XPS, TEM, BET, UV-vis/diffuse reflectance, PZC measurement, Raman spectroscopy, photoluminescence, and fluorescence lifetime, while the photocatalytic activities were tested using benzoic acid as the model compound. Ga-doped TiO2 exhibits the highest photoactivity among the prepared metal-doped TiO2 photocatalysts. This is most likely due to the good dispersion of Ga dopant onto the surface of TiO2, adequate surface area, and decrease of recombination center on the surface of Ga-doped TiO2 photocatalyst. The optimum dopant concentration was found to be 0.1 wt % for Ga-doped TiO2. Introduction TiO2 photocatalysis has received a great deal of attention in water purification technology for the past several years due to its optical and electronic properties, low cost, chemical stability, and nontoxicity.1-3 When illuminated by light of energies greater than the band-gap energy (3.2 eV), TiO2 shows strong oxidation abilities, which can oxidize almost all organic compounds to CO2, as well as reduction capabilities, which can reduce toxic metal ions. The mechanisms of photocatalysis have also been discussed in numerous papers.1,4,5 It has been known that TiO2 has three crystal phases: anatase, rutile, and brookite, among which anatase is shown to be the most photocatalytically active one.6 In recent years, attention has been diverted to metal-doped TiO2, which is believed to decrease the band-gap energy. Doping of metal/metal ion in a semiconductor is known to affect both photophysical behavior and photochemical reactivity.1,7 Metal affects the surface properties by generating a Schottky barrier of the metal in contact with the semiconductor surface, which acts as an electron trap. Similarly, doping of transition metal ions to semiconductors improves the trapping of electron and inhibits e--h+ recombination because the relatively low photoactivity of TiO2 is considered to be due mainly to the fast recombination of photogenerated electrons and holes.8 Several transition metals, such as Pt, Ag, and Au, have been shown to promote the separation of photogenerated charge carriers, and thus to improve its photoactivity.9-14 Such an enhancement in the photocatalytic reactivity has been explained in terms of a photoelectrochemical mechanism in which the electrons generated by UV irradiation of the TiO2 semiconductor quickly transfer to the metal particles loaded on the TiO2 surface. These metal particles work to effectively enhance the charge separation of the electrons and holes, resulting in marked improvement in photocatalytic performance. The sol-gel method is an attractive method for lowtemperature synthesis of TiO2, and it is easier to realize metal doping. However, the particle size of TiO2 prepared by the sol* Corresponding author. E-mail: [email protected]. Tel.: +1 (519) 661 2111, ext. 81279. Fax: +1 (519) 661 3498. † National University of Singapore. ‡ University of Western Ontario.

gel method is not uniform and it is difficult to obtain nanosized particles using this method due to calcination posttreatments. The hydrothermal method is a promising method used in the preparation of nanocrystalline TiO2. It has many advantages: (a) A crystalline product can be obtained directly at relatively low temperatures (in general less than 523 K). Hence the sintering process, which results in a transformation from an amorphous phase to the crystal phase, can be avoided. It also favors a decrease in agglomeration between particles. (b) By changing hydrothermal conditions (such as temperature, pH, reactant concentration, and molar ratio), crystalline products with different compositions, structures and morphologies can be formed. (c) The purity of the product prepared under appropriate conditions can be very high owing to recrystallization in hydrothermal solution. (d) The equipment and processing required are simpler, and the control of reaction conditions is easier.15 Recently, a modified sol-gel method,11,16-18 using hydrothermal posttreatment, has been proposed to synthesize nanocrystalline TiO2 of controlled crystal and aggregate sizes. This procedure involved the precipitation of amorphous titanium oxide gels, followed by a controlled crystallization of anatase TiO2 crystals. After hydrothermal treatment, the amorphous gels were transformed into aggregates of TiO2 nanocrystals that retained the original sizes and shapes of the gels. This paper addresses the preparation of metal-doped TiO2 synthesized by a modified sol-gel method, and characterization and catalytic activity of nanocrystalline TiO2 with uniform particle size. The enhanced photocatalytic activities of metaldoped TiO2 have been demonstrated using benzoic acid as the model compound. The prepared undoped TiO2 and metal-doped TiO2 samples were characterized by X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), UV-vis/diffuse reflectance spectroscopy (DRS), photoluminescence, fluorescence lifetime, and Raman spectroscopy. Experimental Section Catalyst Preparation. A modified sol-gel method was employed to prepare metal-doped TiO2 photocatalysts with metal salt hydrates Ni(NO3)2‚6H2O, Ga(NO3)3‚xH2O, Fe(NO3)3‚9H2O, and Cd(NO3)2‚4H2O, and AgNO3 as the precursors. All chemicals were used as received without further treatment. Using the molar ratio Ti(OBu)4 (97%, Aldrich):C2H5OH (99.9%, J.

10.1021/ie051098z CCC: $33.50 © 2006 American Chemical Society Published on Web 04/04/2006

3504

Ind. Eng. Chem. Res., Vol. 45, No. 10, 2006

T. Baker):H2O (ELGA 18.2 MΩ ultrapure water):HNO3 (69.070.0%, J. T.Baker) ) 1:20:6:0.8, 12.5 mL of Ti(OBu)4 was dissolved in 33.3 mL of absolute alcohol with stirring for 10 min; then 0.25 mL of HNO3 was added dropwise to the above solution under stirring for 30 min. Another solution containing 16.7 mL of absolute alcohol, 1.5 mL of H2O, and metal salt hydrates in the required stoichiometry was slowly added into the above solution. The mixture was hydrolyzed at room temperature for 40 min under vigorous stirring, and the transparent sol was obtained. The gel was prepared by aging the sol for 48 h at room temperature. The derived gel was transferred into a 150-mL, Teflon-lined autoclave vessel for hydrothermal treatment at 493 K for 20 h. Then the suspension was centrifuged for powder separation, washed several times with absolute alcohol (to minimize agglomeration by breaking hydrogen bonding between particles), and dried at room temperature. The samples after hydrothermal treatment were denoted as Me-TiO2-X (Me, doped metal; X, dopant content, wt %). Catalyst Characterization. The structural properties of the prepared photocatalysts were analyzed with a Shimadzu XRD6000 X-ray diffractometer using Cu KR irradiation with a scan rate of 2° min-1. Diffuse reflectance spectra were recorded by a Shimadzu 3100 UV-vis-NIR spectrometer equipped with an integrating sphere ISR-3100, and BaSO4 powder was used as a standard surface. Transmission electron microscopy (TEM) images were taken on a JEOL JEM-2010 TEM instrument. The samples were prepared on a copper mesh substrate covered with an amorphous carbon film. The surface chemical analysis of the samples was made by X-ray photoelectron spectroscopy (XPS, AXIS His; Kratos Analytical) using a Mg KR X-ray source (hν ) 1253.6 eV) and an analyzer pass energy of 40 eV. The physical adsorption of nitrogen was performed on a NOVA 1000 system (Quantachrome) at 77 K. A sample was degassed in a vacuum at 423 K for 4 h before measurement. Surface areas of the samples were calculated based on the BET model. The point of zero charge (PZC) of the prepared samples was determined by measuring the zeta potential, ξ, of each sample at different pH values by a ZetaPlus zeta-potential analyzer (Brookhaven Instruments Corporation). The suspensions were prepared by adding 0.1 g L-1 sample and 0.001 M NaCl in ultrapure water, and then the pH was adjusted. The suspensions were shaken for 24 h in the dark and the ξ-potential was measured after testing the final pH of the suspensions. Raman spectra were detected with a Bruker FRA 106 FT-Raman Accessory using a Nd:YAG laser (λ ) 1064 nm, 350 mW, Bruker Instrument Inc.) at a resolution of 4 cm-1. Photoluminescence spectra were collected on a Perkin-Elmer Luminescence Spectrometer (LS-50B) with a powder holder accessory under room temperature. Fluorescence lifetimes were measured on a GL-3300 Fluorescence Lifetime System (Photon Technology International, Inc.) using a dye laser for excitation. Evaluation of Photocatalytic Activity. A semibatch swirlflow monolithic type photoreactor2 containing 250 mL of aqueous suspension was used, which could provide sufficient mixing among reactant, photon, and catalyst particles. A 125 W high-pressure Hg lamp (Philips, Belgium) and reactor were placed inside a wooden box painted black so that no stray light can enter the reactor. The lamp has a spectral energy distribution with a sharp peak at λ ) 365 nm of 2.1 W with an incident light intensity of 15.05 mW cm-2. The light intensity was measured by a digital radiometer (Model UVX-36; UVP). Oxygen was continuously bubbled into the suspension during the entire experimental process. The amount of catalyst used

Figure 1. Schematic diagram of the experimental setup.

Figure 2. X-ray diffraction patterns of (a) undoped TiO2 and 1 wt % metaldoped TiO2 doped with (b) Ga, (c) Ag, (d) Fe, (e) Cd, and (f) Ni.

for all the experiments was 0.8 g L-1, and the initial concentration of benzoic acid (99.5+%, BDH Chemicals) was 25 mg L-1. The initial pH of the suspension was ca. 3.8, and the temperature inside the photoreactor was kept at 303 K using a water bath. The aqueous suspension was introduced tangentially into the reactor by a peristaltic pump (Watson Marlow 505S) from a beaker with a water jacket and exited from the center of the top plate through an outlet plastic tube of 5 mm diameter. A magnetic stirrer also mixed the suspension in the beaker during the entire reaction period. The detailed schematic view of the experimental setup is illustrated in Figure 1. Before turning on the light, the suspension was circulated for about 30 min, which was chosen on the basis of kinetic experiments, to ensure that the adsorption equilibrium of benzoic acid had been attained on the catalyst surface. Samples of benzoic acid were taken from the suspension every 15 min. Syringe-driven filters (RC membrane, 0.45 µm) were used to remove the catalyst particles before analyzing the samples. The quantitative determination of benzoic acid was performed by measuring its absorption at 226.50 nm with a UV spectrophotometer (UV1601 PC; Shimadzu). Results and Discussion Catalyst Characterization. (a) Structure and Morphology. The XRD patterns of undoped TiO2 and different metal-doped TiO2 with identical dopant concentration of 1 wt % are shown

Ind. Eng. Chem. Res., Vol. 45, No. 10, 2006 3505

Figure 3. TEM image of 0.1 wt % Ga-doped TiO2 (dark particles represent Ga2O3 particles). Table 1. Specific Surface Area (SBET), Anatase Crystal Size (σ), and pH for Point of Zero Charge, PZC (pH) sample

SBET, m2 g-1

σ, nm

PZC (pH)

Degussa P25 undoped TiO2 0.05 wt % Ga-TiO2 0.1 wt % Ga-TiO2 0.5 wt % Ga-TiO2 1.0 wt % Ga-TiO2 0.1 wt % Ag-TiO2 0.1 wt % Fe-TiO2 0.1 wt % Cd-TiO2 0.1 wt % Ni-TiO2

55 ( 15 57 99 121 114 119 89 143 130 142

ca. 30 9.5 11.0 9.6 11.2 12.2 8.6 (for 1 wt % Ag-TiO2) 12.2 (for 1 wt % Fe-TiO2) 9.4 (for 1 wt % Cd-TiO2) 9.2 (for 1 wt % Ni-TiO2)

6.9a 6.7 6.8 6.2 5.9 5.3

a

Dutta et al.41

in Figure 2. From the figure, we can see that the pattern can be indexed to TiO2 in the anatase phase only. The strongest peak at 2θ ) 25.3° is representative for (101) anatase phase reflections. From the full width at half-maximum (fwhm) of the diffraction pattern, we can calculate the particle sizes, which are in the range from 9.2 to 12.2 nm using Scherrer’s equation. The presence of the separate oxide phases was not observed, possibly due to the low dopant concentration. It is reported that the doped TiO2 with metal dopant concentration above 5 wt % can show a small peak attributable to a separate oxide phase in addition to the lines of anatase TiO2.19 Furthermore, if we consider that the ionic radii of the doped metal ions (Ga3+ ) 0.76 Å; Fe3+ ) 0.79 Å; Ni2+ ) 0.83 Å, etc.) are similar to that of Ti4+ (0.75 Å) for a coordination number of 6, these doped metal ions can likely be incorporated into the TiO2 matrix by substituting for the Ti4+ lattice sites.20 The specific surface areas of the prepared samples are listed in Table 1. The surface areas of metal-doped TiO2 are almost twice that of the undoped TiO2 or Degussa P25. For Ga-doped TiO2 samples, the surface areas increase with the increase of Ga dopant concentration up to a maximum of 0.1 wt % and then remain constant with the increase of Ga dopant concentration. When TiO2 is doped with gallium, the well-dispersed Ga2O3 covering the surface of TiO2 (which can be seen from Figure 3) can increase the specific surface area of TiO2. When the amount of Ga2O3 is increased beyond a maximum (covering a monolayer), the Ga2O3 may not only cover the surface of TiO2,

Figure 4. Diffuse reflectance UV-vis spectra of (a) P25, (b) undoped TiO2, and 1 wt % metal-doped TiO2 doped with (c) Ga, (d) Fe, (e) Ni, (f) Cd, and (g) Ag.

but also enter the TiO2 matrix and thus obstruct the pores of the TiO2 support, resulting in a decrease of the specific surface area.13 The TEM image of 0.1 wt % Ga-TiO2 is shown in Figure 3. It can be seen that the prepared sample has agglomerated to larger particles, which have mean particle size about 200 nm. However, the image reveals that the sample is consisted of agglomerates of primary particles with an average diameter of ca. 10 nm, which is in good agreement with the crystallite size calculated from the XRD pattern. It can be clearly seen that Ga dopant is well dispersed onto the surface of the sample of 0.1 wt % Ga-TiO2, and this seems to have some definite relationship with its good performance on the photocatalytic degradation of benzoic acid. It is obvious that the doped Ga exists mostly in the form of Ga2O3 on the surface of TiO2 by surface doping, although substitutional doping may also be possible. However, from the increase of specific surface area after doping Ga into TiO2 and TEM observation, we believe that the good dispersion of Ga2O3 onto the surface of TiO2 may be the main reason for the increase in specific surface area. (b) UV-Vis/DRS. UV-vis/DRS is used to probe the band structure (or molecular energy levels) in the materials since UV-vis light excitation creates photogenerated electrons and holes. The UV-vis absorption band edge is a strong function of titania cluster size for diameters less than 10 nm, which can be attributed to the well-known quantum-size effect for semiconductors.21 The spectra of the metal-doped samples in Figure 4 show a slight red shift in the band-gap transition to longer wavelengths. The absorbance of the metal-doped samples in the visible region is always higher than that of undoped TiO2. In particular, the most pronounced effect occurred in the case of Ag-doped TiO2 sample, which is gray in color. This is due to the absorption of wavelengths longer than 600 nm, which indicates that Ti3+ may exist in the TiO2 lattice.22 Except for the band-gap absorption threshold of titania, a strong wide absorption spectrum dominated the absorption spectrum of Agdoped TiO2, which is significantly different from other metaldoped samples, characteristic of surface plasmon absorption corresponding to silver particles.7,23 The onset of band-gap absorption of Degussa P25 is estimated at 403 nm, which corresponds to 3.08 eV. This may be because Degussa P25 is

3506

Ind. Eng. Chem. Res., Vol. 45, No. 10, 2006

Figure 6. PL spectra of (a) undoped TiO2; 1 wt % metal-doped TiO2 with dopants (b) Ga, (c) Fe, (d) Cd, (e) Ni, and (f) Ag; and (g) P25. Excitation wavelength: 300 nm.

Figure 5. XPS spectra of Ga 3d and Ag 3d peaks.

a mixture of rutile (band-gap energy 3.0 eV) and anatase (bandgap energy 3.2 eV). The estimated band-gap absorption of metaldoped TiO2 samples begins at 399, 405, 412, and 418 nm for Ga-doped TiO2, Fe-doped TiO2, Ni-doped TiO2, and Cd-doped TiO2, respectively, which can be compared with the ca. 393 nm (corresponding to band-gap energy of 3.15 eV) characteristic of the undoped TiO2 sample. The band-gap absorption of undoped TiO2 is lower than that of Degussa P25, which may be due to the pure anatase form in the prepared samples. (c) XPS. X-ray photoelectron spectroscopy (XPS) experiments were performed to elucidate both the titania structure and the chemical state of the doped-metal particles. The peaks at 459.1 and 464.7 eV for the Ti 2p region of 1 wt % metal-doped samples can be attributed to Ti in 4+ state, which is in an octahedral environment (graph not shown here). The O 1s XPS spectra have two peaks, which indicate that two kinds of oxygen species are present in the near-surface region. The first peak at 530.3 eV is assigned to the lattice oxygen atoms of both TiO2 and Ga2O3 (Ti-O bonds and Ga-O bonds), while the second peak at 532.2 eV is due to surface hydroxyl.24 These surface hydroxyl groups can play an important role in the photocatalytic reactions since the photoinduced holes can attack the surface hydroxyl groups and yield surface-bound OH radicals with high oxidation capability.6,24 Figure 5 depicts the XPS spectra of Gadoped TiO2 and Ag-doped TiO2. The Ga 3d5/2 peak shows a binding energy of 20.5 eV, which can be assigned to Ga2O3 (Ga-O bonds). The Ag 3d5/2 peak appears at a binding energy of 368.2 eV, whereas the splitting of the 3d doublet is 6.0 eV.

These values correspond to metallic silver,25 which is in agreement with the result of UV-vis. (d) Point of Zero Charge (PZC). The PZC of an oxide is the value of pH at which the net surface charge is zero. The knowledge of PZC helps to predict whether ion exchange occurs preferentially to a specific component of the oxide system influencing the photoreactivity of the powder. The PZC results of Ga-doped TiO2 are listed in Table 1. It reveals that the PZC values of the prepared samples are lower than that of Degussa P25. The PZC value of the 0.05 wt % Ga-doped TiO2 is not significantly different from that of the undoped TiO2. Furthermore, with the increase of Ga dopant concentration for Gadoped TiO2, the PZC value decreased. This indicates that there is surface enrichment of species with an acidic behavior such as Ga2O3.13 (e) Photoluminescence (PL). Photoluminescence (PL) spectroscopy is an important tool for studying the electronic structure, optical properties, and photoelectric properties in rutile and anatase phases of TiO2 in forms such as single crystals, thin films, and polycrystalline powders. In TiO2 polycrystalline powders, band-gap excitation can result in free or trapped exciton emission due to the recombination of photoinduced electron-hole pairs, or cause broad-band visible emission attributed to luminescence from localized surface states.26-28 Figure 6 depicts the PL spectra of Degussa P25, undoped TiO2, and different metal-doped TiO2 samples with dopant concentration of 1 wt %. The photoluminescence peak position at 393 nm of undoped TiO2 was the same as its onset of UV-vis/ diffuse reflectance spectra, which possibly demonstrates that the PL peak resulted according to the electronic structure of TiO2 from the electron transition from the bottom of the valence band to the top of the conduction band. The first peak around 393 nm on PL spectra of some metal-doped TiO2 is not significant, which may be due to the different characteristics of doped metal ions. Although the PL spectrum of Degussa P25 is broad and much more intense than that of any of the other materials, the Degussa P25 spectrum does not display a distinct peak around 400 nm. However, it can be clearly seen from Figure 7 that the PL spectra show three main peaks for undoped TiO2 and Ga-doped TiO2 with different dopant concentrations at about 393 nm (3.15 eV), 420 nm (2.95 eV), and 484 nm (2.56 eV) using the excitation wavelength of 300 nm. The PL spectra might be closely related to the luminescence of the recombination of photoinduced electron and holes, free excitons, and self-trapped excitons, which possibly resulted from surface

Ind. Eng. Chem. Res., Vol. 45, No. 10, 2006 3507

Figure 7. PL spectra of (a) undoped TiO2 (a) and Ga-doped TiO2 with different dopant concentrations: (b) 0.05, (c) 0.1, (d) 0.5, (e) 1, (f) 2, and (g) 5 wt %. Excitation wavelength: 300 nm.

Figure 8. Fluorescence decay profiles of (a) 0.05 wt % Ga-TiO2, (b) 0.1 wt % Ga-TiO2, and (c) 0.5 wt % Ga-TiO2. Excitation wavelength: 337 nm.

defects in the nanosized TiO2 crystals such as lattice distortion and surface oxygen deficiencies.29 The two peaks at 393 and 420 nm might mainly arise from the self-trapped excitons localized on TiO6 octahedra.27 The peak at 484 nm can likely be attributed to oxygen vacancies.30 (f) Fluorescence Decay Profile. The decay profiles of the prepared samples were investigated by the fluorescence lifetime method. The fluorescence decay profiles of Ga-doped TiO2 with different dopant concentrations and 0.1 wt % Ga-TiO2 under different excitation wavelengths are shown in Figures 8 and 9, respectively. These decay profiles were well traced using the extended exponential function:31 β

[ (τt ) ]

I ) I0tβ-1 exp -

(1)

where τ is the lifetime and β is the shape factor. This function has been widely used for analysis of charge carrier dynamics in amorphous semiconductors and glass.32 Good fitting of the decay profiles of the Ga-doped TiO2 using eq 1 indicates the existence of carrier trapping sites with different energy levels, which lead to a distribution of the carrier transport rates. According to this equation, we can calculate the lifetime of Degussa P25, undoped TiO2, and Ga-doped TiO2 whose results

Figure 9. Fluorescence decay profiles of 0.1 wt % Ga-TiO2 under different excitation wavelengths: (a) 266 and (b) 337 nm. Table 2. Emission Lifetime of Fluorescence Decay Profile sample

measd excitation wavelength, nm

lifetime, ns

Degussa P25 undoped TiO2 0.05 wt % Ga-TiO2 0.1 wt % Ga-TiO2 0.1 wt % Ga-TiO2 0.5 wt % Ga-TiO2 1.0 wt % Ga-TiO2

337 337 337 337 266 337 337

2.14 2.55 2.30 2.15 1.55 2.11 1.79

are listed in Table 2. From Figure 8, it can be seen that the decay profiles of Ga-doped TiO2 were quenched with the increase of the Ga dopant concentration. The lifetimes of Gadoped TiO2 were also decreased with the increase of Ga dopant concentration. These results suggest that, with the increase of Ga dopant concentration, Ga dopants play the role of the recombination center at the TiO2 surface. It can be seen from Figure 9 that the lifetime of 0.1 wt % Ga-TiO2 is dependent on the excitation wavelengths. When the shorter wavelength (λ ) 266 nm) is used to excite 0.1 wt % Ga-TiO2, the lifetime is drastically decreased. This may indicate that the wavelength of 266 nm is not enough to deeply penetrate the bulk of the TiO2 particles. Because the mobility of electrons in TiO2 is slow,33 the decay profile is considered to be chiefly determined by the migration of electrons to the surface of TiO2.31 (g) Raman Spectroscopy. Raman spectroscopy was applied to discriminate the local order characteristics of the samples. The (nondestructive) technique is capable of elucidating the photocatalyst structural complexity as peaks from each material are clearly separated in frequency, and therefore, the phases are easily distinguishable. The Raman spectra for anatase was identified by Hirata and co-workers34 at 144 cm-1 (Eg), 197 cm-1 (Eg), 399 cm-1 (B1g), 513 cm-1 (A1g), 519 cm-1 (B1g), and 639 cm-1 (Eg). However, for our case, a well-resolved TiO2 Raman peak is observed in Figure 10A at 149 ( 1 cm-1 for all samples. This peak is attributed to the main Eg anatase vibration mode, and the peak is intensified with the increase of Ga dopant concentration to a maximum intensity for 0.1 wt % Ga-TiO2 sample. With further increase of Ga dopant concentration, the peak became weaker and weaker. Moreover, vibration peaks at 201 ( 1 cm-1 (Eg, weak), 395 ( 1 cm-1 (B1g), 513 ( 1 cm-1 (A1g), and 637 ( 1 cm-1 (Eg) are present in the spectra of all samples, which indicates that anatase TiO2 nanoparticles are the predominant species. However, it can be seen from Figure 10C that a weak peak at 428 cm-1 belonging to the E1g mode of rutile can be seen for Degussa P25.35

3508

Ind. Eng. Chem. Res., Vol. 45, No. 10, 2006

Figure 11. Photocatalytic degradation of benzoic acid using (9) Degussa P25, (b) undoped TiO2, and 0.1 wt % metal-doped TiO2: (2) Ga, (1) Ag, ([) Cd, (3) Fe, and (4) Ni.

Figure 10. (A) Raman spectra of (a) undoped TiO2, (b) P25, and 0.1 wt % metal-doped TiO2 doped with (c) Ga, (d) Ag, and (e) Fe. (B) Enlarged graph of (A) for undoped TiO2 (i) and P25 (ii). (C) Ga-doped TiO2 with different dopant concentrations: (1) 0.05, (2) 0.1, (3) 0.5, and (4) 1 wt %.

Photocatalytic Activity. (a) Effect of Different MetalDoped TiO2 on Photodegradation of Benzoic Acid. The photocatalytic degradation of benzoic acid over different metaldoped samples with the same doping concentration (0.1 wt %) was evaluated, and the results are shown in Figure 11. It is clear that the photoactivities of Ga-doped and Ag-doped TiO2 are higher than those of Degussa P25 and undoped TiO2. Among all the samples, Ga-doped TiO2 with dopant concentration of 0.1 wt % shows the highest photoactivity. The photocatalytic activity of the crystalline powders depends not only on their electronic properties but also on many other factors such as the particle size distribution, morphology, type of pores present, crystal phase, surface area, d electron structure, differing hydroxylation of the surfaces, acid-base properties, amount of adsorbed reactant species, etc.13 Metal dopants in the TiO2 photocatalytic system have two functions: (a) they can act as both electron traps and hole traps to be photoactive and (b) metal dopants may engage the possibility of charge release and migration to the interface of a solid-liquid reaction system. The reduced activity can be explained as dopants act more as recombination centers than as trap sites for charge transfer at the interface.36 The XRD patterns of different metal-doped TiO2 did not show any significant variations of structure and morphology arising from the metal doping compared with the undoped TiO2 due to

the low dopant concentration. The UV-vis/diffuse reflectance spectra of the metal-doped TiO2 reveal a slight red shift in the band-gap transition, which can be explained by the introduction of energy levels of the doped metal into the band gap of TiO2.10 The energy levels below the conduction band edge and above the valence band edge influence the photoreactivity of TiO2 since the doped metal ions can act as electron (or hole) traps, which can alter the electron-hole pair recombination rate. The UV-vis/DRS results revealed the existence of Ti3+ on the surface of Ag-doped TiO2, and it is regarded that Ti3+ states on the surface of nanosized TiO2 are disadvantageous to photocatalytic oxidation reactions because they can easily become recombination centers of photoinduced electrons and holes.37 However, metallic silver in a low amount can play a favorable role by attracting electrons, thus helping electronhole pair separation and preventing electron-hole recombination.38 On the other hand, the ionic radius of Ga3+ (0.76 Å) is very close to that of Ti4+ (0.75 Å), which indicates that Ga3+ might be much easier than other metal ions whose radii are larger than that of Ti4+ to enter the TiO2 lattice to replace Ti4+. As a result, it may decrease the coverage of gallium dioxide on the surface of TiO2, which decreases the recombination center. For other metal-doped TiO2, many researchers have reported their detrimental function in the photocatalytic oxidation reaction. However, the role of Fe3+ dopant in TiO2 is controversial. Some reports revealed that Fe3+ behaves as an electron-hole recombination center, which decreases its photoactivity. Serpone et al.39 reported that addition of transition-metal dopants, such as Fe, Cr, or V, to the anatase TiO2 lattice may tend to decrease the photooxidative activity of TiO2 for the photooxidation reaction of oxalic acid. The dopants behave as an electronhole recombination center. This is supported by the observation that increasing the dopant content decreases the photocatalytic activity and thus increases the recombination center. Other authors have postulated that the role of dopant would favor electron-hole separation, which enhances the photoactivity. Choi et al.10 reported that Fe3+ can act as both electron trapping and hole trapping. Trapping either an electron or a hole alone is ineffective because the immobilized charge species quickly recombines with its mobile counterpart. Zhang et al.11 introduced a new concept, shallow trapping sites, which can be used to describe the role of Fe3+ dopant in TiO2. Shallow trapping means that the trapped charge carrier can be thermally excited from the trapping site and become mobile again, preferably before the generation of the next e--h+ pair inside the same

Ind. Eng. Chem. Res., Vol. 45, No. 10, 2006 3509

Figure 12. Concentration-time fitting curve based on LangmuirHinshelwood kinetic model: (9) Degussa P25, (b) undoped TiO2, and 0.1 wt % metal-doped TiO2 doped with (2) Ga and (1) Fe. Table 3. Concentration-Time Fitting Results Based on Langmuir-Hinshelwood Model catalyst

rate const, mg L-1 min-1

R2

Degussa P25 undoped TiO2 0.1 wt % Ga-TiO2 0.1 wt % Ag-TiO2 0.1 wt % Cd-TiO2 0.1 wt % Fe-TiO2

0.152 0.254 5.544 0.334 0.033 0.085

6.16 3.36 14.84 10.28 3.25 1.97

particle. Photocatalytic decomposition efficiencies can be increased when using Fe3+-doped TiO2, in which Fe3+ ions serve as shallow trapping sites for the charge carriers, thereby separating the arrival time of e- and h+ at the surface. A better and more quantitative way of presenting the activities of a series of metal-doped TiO2 photocatalysts is the use of the rate constant k, which is independent of the concentrations used. In photocatalytic reactions, k is almost independent of temperature because of the photoactivation process and depends only on the radiant flux and on the UV spectrum of the lamp. The photocatalytic degradation of organic pollutants in water generally follows a Langmuir-Hinshelwood mechanism:7

r)

kKc 1 + Kc

(2)

where k is the true rate constant which is a function of various parameters such as the mass of catalyst, the flux of efficient photons, and the coverage in oxygen. K is the adsorption constant. Fitting the experimental data into the above kinetic model (Figure 12), the rate constants were obtained and are listed in Table 3. The experiment data agree closely with the fitting curve. It can also be seen from Table 3 that 0.1 wt % Ga-doped TiO2 shows the highest rate constant. (b) Effect of Ga Dopant Concentration on Photodegradation of Benzoic Acid. Figure 13 shows the results of the photocatalytic degradation of benzoic acid for 60 min irradiation over Degussa P25, undoped TiO2 catalyst, and Ga-doped TiO2 photocatalysts with different doping concentrations. It is obvious that the undoped TiO2 shows higher photoactivity than P25. For the samples doping with different concentrations of gallium in TiO2, 0.1 wt % Ga-TiO2 catalyst exhibits the highest photoactivity.

Figure 13. Photocatalytic degradation of benzoic acid using (9) Degussa P25, (b) undoped TiO2, and Ga-doped TiO2 with different dopant concentrations: (2) 0.05, (1) 0.1, ([) 0.5, and (4) 1 wt %.

From XPS results, it was observed that gallium exists on the surface mainly in the form of Ga2O3. Therefore, it is reasonable to believe that the possible reason for this result is that the excess amount of gallium oxide covering the surface of TiO2 increases the number of recombination centers and results in low photoactivity. Xu et al.16 reported that an excess amount of rare earth oxide would cover the surface of TiO2, which results in the decrease of photocatalytic activity since the existence of an excess amount of rare earth oxide would increase the recombination centers and this increase will be detrimental to photocatalytic reactions. TEM imaging revealed that the gallium sesquioxide is dispersed very well over the surface of 0.1 wt % Ga-TiO2. This may benefit its enhancement on photodegradation of benzoic acid. The PZC values of Ga-doped TiO2 decreased with the increase of dopant concentration. It seems that no direct relationship between the PZC values and photoactivity of Gadoped TiO2 can be made. However, it can be explained by taking into account the nature of the reacting molecules and the acidbase and electronic properties of the metal-doped TiO2 photocatalysts.40 Co-doped, Cu-doped, and Fe-doped TiO2 with high PZC values show high photoactivity over the degradation of methanoic acid. The author attributed this to the relatively basic PZC values of these samples since methanoic acid is more acidic and a stronger interaction with a quite basic catalyst surface probably occurs. However, due to the good performance of 0.1 wt % Ga-TiO2 on the photodegradation of benzoic acid, we might consider that benzoic acid is more likely to be degraded under a relatively weak acidic PZC value since it is a weak acid. Fluorescence decay profiles have revealed that, with the increase of Ga dopant concentration, the lifetime of Ga-doped TiO2 decreased. This is attributed to the increase of the recombination centers on the surface of TiO2, which lead to the decrease of photoactivity due to the increase of dopant concentration. It can also be seen from Table 1 that 0.1 wt % Ga-TiO2 has the smallest particle size and the largest specific surface area, and shows the highest photocatalytic activity. The reduction in particle size, which leads to larger surface area, will be beneficial to the increase of the available surface active sites. It is wellknown that the successful separation of photoinduced electronhole pairs is crucial to photocatalytic reactions. With the decrease of particle size, the photoinduced electrons (or holes) when illuminated are likely to show easier transit onto the

3510

Ind. Eng. Chem. Res., Vol. 45, No. 10, 2006

surface of metal-doped TiO2 to react with the metal ions or adsorbed organic compounds rather than to recombine with each other to give off heat. Such an efficient migration of electrons or holes to the surface of metal-doped TiO2 can appreciably improve the charge separation process and inhibit the recombination of electron-hole pairs. Consequently, the photocatalytic activity of metal-doped TiO2 can be increased. According to the above-mentioned mechanism, it is not difficult to understand that 0.1 wt % Ga-TiO2, with the smallest particle size and the largest surface area compared with other Ga-doped TiO2 samples, shows the highest photocatalytic activity. Conclusions A modified sol-gel method was utilized to synthesize different metal-doped TiO2 with different dopant concentrations. Among these prepared samples, 0.1 wt % Ga-TiO2 exhibited the highest photocatalytic activity over the degradation of benzoic acid when 365 nm UV light was used for irradiation. The enhancement of the photoactivity can be attributed to its smallest particle size and largest surface area, good dispersion of gallium sesquioxide over the surface of TiO2, and suitable dopant concentration, which might have decreased the recombination center. Consequently, it could be considered that the photactivity of the photocatalysts depends on many factors, which can affect the photoactivity of the photocatalysts comprehensively. Future investigations will be focused on the acidbase properties and charge separation dynamics of the prepared photocatalysts. Literature Cited (1) Hoffmann, M. R.; Martin, S. T.; Choi, W.; Bahnemann, D. W. Environmental applications of semiconductor photocatalysis. Chem. ReV. 1995, 95, 69. (2) Chen, D.; Ray, A. K. Photocatalytic kinetics of phenol and its derivatives over UV irradiated TiO2. Appl. Catal., B 1999, 23, 143. (3) Sivalingam, G.; Nagaveni, K.; Hegde, M. S.; Madras, G. Photocatalytic degradation of various dyes by combustion synthesized nano anatase TiO2. Appl. Catal., B 2003, 45, 23. (4) Mills, A.; Davies, R. H.; Worsley, D. Water purification by semiconductor photocatalysis. Chem. Soc. ReV. 1993, 22, 417. (5) Fox, M. A.; Dulay, M. T. Heterogeneous photocatalysis. Chem. ReV. 1993, 93, 341. (6) Litter, M. I. Heterogeneous photocatalysis transition metal ions in photocatalytic systems. Appl. Catal., B 1999, 23, 89. (7) Herrmann, J. M.; Tahiri, H.; Ait-Ichou, Y.; Lassaletta, G.; GonzalezElipe, A. R.; Fernandez, A. Characterization and photocatalytic activity in aqueous medium of TiO2 and Ag-TiO2 coatings on quartz. Appl. Catal., B 1997, 13, 219. (8) Kumar, A.; Jain, A. K. Photophysics and photocatalytic properties of Ag+-activated sandwich Q-CdS-TiO2. J. Photochem. Photobiol., A 2003, 156, 207. (9) Butler, E. C.; Davis, A. P. Photocatalytic oxidation in aqueous titanium dioxide suspensions: the influence of dissolved transition metals. J. Photochem. Photobiol., A 1993, 70, 273. (10) Choi, W.; Termin, A.; Hoffmann, M. R. The role of metal ion dopants in quantum-sized TiO2: correlation between photoreactivity and charge carrier recombination dynamics. J. Phys. Chem. 1994, 98, 13669. (11) Zhang, Z.; Wang, C. C.; Zakaria, R.; Ying, J. Y. Role of particle size in nanocrystalline TiO2-based photocatalysts. J. Phys. Chem. B 1998, 102, 10871. (12) Furube, A.; Asahi, T.; Masuhara, H.; Yamashita, H.; Anpo, M. Charge carrier dynamics of standard TiO2 catalysts revealed by femtosecond diffuse reflectance spectroscopy. J. Phys. Chem. B 1999, 103, 3120. (13) Di Paola, A.; Marci, G.; Palmisano, L.; Schiavello, M.; Uosaki, K.; Ikeda, S.; Ohtani, B. Preparation of polycrystalline TiO2 photocatalysts impregnated with various transition metal ions: characterization and photocatalytic activity for the degradation of 4-nitrophenol. J. Phys. Chem. B 2002, 106, 637.

(14) Arabatzis, I. M.; Stergiopoulos, T.; Andreeva, D.; Kitova, S.; Neophytides, S. G.; Falaras, P. Characterization and photocatalytic activity of Au/TiO2 thin films for azo-dye degradation. J. Catal. 2003, 220, 127. (15) Cheng, H.; Ma, J.; Zhao, Z.; Qi, L. Hydrothermal preparation of uniform nanosize rutile and anatase particles. Chem. Mater. 1995, 7, 663. (16) Xu, A. W.; Gao, Y.; Liu, H. Q. The preparation, characterization and their photocatalytic activities of rare-earth-doped TiO2 nanoparticles. J. Catal. 2002, 207, 151. (17) Wang, C. C.; Zhang, Z.; Ying, J. Y.; Photocatalytic decomposition of halogenated organics over nanocrystalline titania. Nanostruct. Mater. 1997, 9, 583. (18) Yeung, K. L.; Yau, S. T.; Maira, A. J.; Coronado, J. M.; Soria, J.; Yue, P. L. The influence of surface properties on the photocatalytic activity of nanostructured TiO2. J. Catal. 2003, 219, 107. (19) Martin, C.; Solana, G.; Rives, V.; Marci, G.; Palmisano, L.; Sclafani, A. Physico-chemical properties of WO3/TiO2 systems employed for 4-nitrophenol photodegradation in aqueous medium. Catal. Lett. 1997, 49, 235. (20) Zhu, J.; Zheng, W.; He, B.; Zhang, J.; Anpo, M. Characterization of Fe-TiO2 photocatalysts synthesized by hydrothermal method and their photocatalytic reactivity for photodegradation of XRG dye diluted in water. J. Mol. Catal., A 2004, 216, 35. (21) Anpo, M.; Shima, T.; Kodama, S.; Kubokawa, Y. Photocatalytic hydrogenation of CH3CCH with H2O on small-particle TiO2: size quantization effects and reaction intermediates. J. Phys. Chem. 1987, 91, 4305. (22) Sakatani, Y.; Ando, H.; Okusako, K.; Koike, H. Metal ion and N co-doped TiO2 as a visible-light photocatalyst. J. Mater. Res. 2004, 19, 2100. (23) Stathatos, E.; Lianos, P.; Falaras, P.; Siokou, A. Photocatalytically deposited silver nanoparticles on mesoporous TiO2 films. Langmuir 2000, 16, 2398. (24) Xu, Z.; Shang, J.; Liu, C.; Kang, C.; Guo, H.; Du, Y. The preparation and characterization of TiO2 ultrafine particles. Mater. Sci. Eng., B 1999, 56, 211. (25) Practical Surface Analysis; Briggs D., Seah, M. P., Eds.; John Wiley and Sons: New York, 1983. (26) Rahman, N. M.; Krishna, K. M.; Soga, T.; Jimbo, T.; Umeno, M. Optical properties and X-ray photoelectron spectroscopic study of pure and Pb-doped TiO2 thin films. J. Phys. Chem. Solids 1999, 60, 201. (27) Li, X. Z.; Li, F. B.; Yang, C. L.; Ge, W. K. Photocatalytic activity of WOx-TiO2 under visible light irradiation. J. Photochem. Photobiol., A 2001, 141, 209. (28) Zhang, W. F.; Zhang, M. S.; Yin, Z.; Chen, Q. Photoluminescence in anatase titanium dioxide nanocrystals. Appl. Phys. B 2000, 70, 261. (29) Zhang, L.; Mo, C. M. Luminescence in nanostructured materials. Nanostruct. Mater. 1995, 6, 831. (30) Serpone, N.; Lawless, D.; Khairutdinov, R. Size effects on the photophysical properties of colloidal anatase TiO2 particles: size quantization or direct transitions in this indirect semiconductor? J. Phys. Chem. 1995, 99, 16646. (31) Fujihara, K.; Izumi, S.; Ohno, T.; Matsumura, M. Time-resolved photoluminescence of particulate TiO2 photocatalysts suspended in aqueous solutions. J. Photochem. Photobiol., A 2000, 132, 99. (32) Suzuki, K.; Bley, G.; Neukirch, U.; Gutowski, J.; Takojima, N.; Sawada, T.; Imai, K. Stretched-exponential decay of the luminescence in ZnSe-ZnTe superlattices. Solid State Commun. 1998, 105, 571. (33) Enright, B.; Fitzmaurice, D. Spectroscopic determination of electron and hole effective masses in a nanocrystalline semiconductor film. J. Phys. Chem. 1996, 100, 1027. (34) Hirata, T.; Kitajima, M.; Nakamura, K. G.; Asari, E. Infrared and Raman spectra of solid solutions Ti1-xZrxO2 (xe0.1). J. Phys. Chem. Solids 1994, 55, 349. (35) Falaras, P.; Hugot-Le Goff, A.; Bernard, M. C.; Xagas, A. Characterization by resonance Raman spectroscopy of sol-gel TiO2 films sensitized by the Ru(PPh3)2 (dcbipy)Cl2 complex for solar cells application. Sol. Energy Mater. Sol. Cells 2000, 64, 167. (36) Navio, J. A.; Colon, G.; Litter, M. I.; Bianco, G. N. Synthesis, characterization and photocatalytic properties of iron-doped titania semiconductors prepared from TiO2 and iron (III) acetylacetonate. J. Mol. Catal., A 1996, 106, 267. (37) Jing, L.; Sun, X.; Cai, W.; Xu, Z.; Du, Y.; Fu, H.; The preparation and characterization of nanoparticle TiO2/Ti films and their photocatalytic activity. J. Phys. Chem. Solids 2003, 64, 615. (38) Sclafani, A.; Mozzanega, M.-N.; Pichat, P. Effect of silver deposits on the photocatalytic activity of titanium dioxide samples for the dehydrogenation or oxidation of 2-propanol. J. Photochem. Photobiol., A 1991, 59, 181.

Ind. Eng. Chem. Res., Vol. 45, No. 10, 2006 3511 (39) Serpone, N.; Lawless, D.; Disdier, J.; Herrmann, J.-M. Spectroscopic, photoconductivity and photocatalytic studies of TiO2 colloids: naked and with the lattice doped with Cr3+, Fe3+ and V5+ cations. Langmuir 1994, 10, 643. (40) Di Paola, A.; Garcia-Lopez, E.; Ikeda, S.; Marci, G.; Ohtani, B.; Palmisano, L.; Photocatalytic degradation of organic compounds in aqueous systems by transition metal doped polycrystalline TiO2. Catal. Today 2002, 75, 87.

(41) Dutta, P. K.; Ray, A. K.; Sharma, V. K.; Millero, F. J. Adsorption of arsenate and arsenite on titanium dioxide suspensions. J. Colloid Interface Sci. 2004, 278, 270.

ReceiVed for reView September 30, 2005 ReVised manuscript receiVed March 6, 2006 Accepted March 10, 2006 IE051098Z