Valence State and Catalytic Role of Cobalt Ions in Cobalt TiO2

ARTICLE pubs.acs.org/JPCC

Valence State and Catalytic Role of Cobalt Ions in Cobalt TiO2 Nanoparticle Photocatalysts for Acetaldehyde Degradation under Visible Light Dambar B. Hamal and Kenneth J. Klabunde* Department of Chemistry, Kansas State University, Manhattan, Kansas 66506, United States

bS Supporting Information ABSTRACT: Cobalt carbon sulfur-codoped anatase photocatalysts (size 1% Co(II)/(C, S) TiO2 > 2% Co(II) TiO2 > 2% Co(III) TiO2 > (C, S) TiO2 under visible light and 2% Co(II)/(C, S) TiO2 > 1% Co(II)/(C, S) TiO2 > 2% Co(II) TiO2 ≈ (C, S) TiO2 > TiO2 P25 under UV light. It was concluded that the doped cobalt ions play a vital role and act as active sites in the photodegradation reaction. A reaction mechanism is proposed that may explain the synergistic effect of the codopants in visiblelight-induced photodegradation of acetaldehyde. This new photocatalyst system, Co/(C, S) TiO2, can have other potential applications that only need visible light as energy input.

’ INTRODUCTION Doping1 6 and codoping7 10 of nonmetals into ultravioletlight-active titanium dioxide photocatalysts have been the most attractive strategies to design and develop efficient visible light active TiO2 nanoparticle photocatalysts for decontamination of toxic organic compounds in polluted air and water. Particularly, doping and codoping of nonmetals allow researchers to reduce the band gap of TiO2 ( 400 nm) to generate electron hole pairs, which can subsequently engage in surface redox reactions, contingent upon their lifetimes and recombination rates. Furthermore, it has been proven that codoping TiO2 with two nonmetals rather than doping with a single nonmetal induces synergistic effects in accelerating photomineralization of toxic organic pollutants primarily by suppressing both surface and bulk charge carrier recombination processes. As a result, a marked improvement has been observed in the activity of the codoped TiO2 photocatalyst under visible light irradiation compared with that of pure and single nonmetal doped TiO2 photocatalysts.9 r 2011 American Chemical Society

Besides doping a single metal or nonmetal, further improvements in visible light activities of TiO2 photocatalysts have been achieved by codoping a metal and a nonmetal. On the basis of this design principle, visible-light-responsive new photocatalytic materials, such as (Sr, N)/TiO2,11 (Ni, B)/TiO2,12 (La, N)/TiO2,13 (La, S)/TiO2,14 (Fe, C)/TiO2,15 (V, B)/TiO2,16 and (V, C)/ TiO2,17 have been developed and investigated especially for environmental remediation purposes. Moreover, these works have shown that new photocatalyst systems of TiO2 derived by codoping a metal and a nonmetal displayed higher visible light activities for degradation of highly toxic organic pollutants than that of the single metal, nonmetal doped, and pristine TiO2. From the mechanistic viewpoint, it turns out that in metal and nonmetalcodoped photocatalyst systems the codoped nonmetal actually reduces the band gap and induces visible light absorption to generate electron hole pairs while the codoped metal suppresses the charge-carrier (electron hole) recombination process.13,14 Thus, codoping TiO2 with a metal and a nonmetal has been considered a good strategy for developing visible-light-driven new photocatalytic materials.12 14,16,17 Received: January 14, 2011 Revised: July 22, 2011 Published: July 27, 2011 17359

dx.doi.org/10.1021/jp200405y | J. Phys. Chem. C 2011, 115, 17359–17367

The Journal of Physical Chemistry C In recent work, we have shown that visible light activities of TiO2 photocatalysts could be markedly improved by codoping TiO2 with a noble metal (Ag) and two nonmetals (C and S).18 This work actually motivated us with further interest to design visible-light-driven TiO2 photocatalyst codoped with a transition metal and two nonmetals. In this context, several authors have reported TiO2 photocatalysts doped with 3d-transition metals— such as V,19 Cr,20 Mn,21 Fe,22 Co,23 Ni,24 Cu,25 and Zn26—for decontamination of hazardous organic pollutants. To our knowledge, no research work has been carried out to study the effect of codoping a 3d-transition metal, cobalt, with two nonmetals (C and S) on the textural, optical, and photocatalytic properties of the TiO2 photocatalyst. As compared with V and Cr, cobalt is less toxic, and cobalt-doped photocatalysts have shown commendable activities for degradation of acetaldehyde27,28 and phenol29 and production of hydrogen30 from aqueous ethanol solution. In the present study, we synthesized and characterized a cobalt carbon sulfur-codoped TiO2 and developed a new visible-light-driven photocatalyst with enhanced photoactivity for acetaldehyde degradation. Moreover, the role of codopants— cobalt, carbon, and sulfur—in visible-light-induced photocatalysis was investigated, and a photocatalytic reaction mechanism is proposed.

’ EXPERIMENTAL SECTION Photocatalyst Preparation. All chemical reagents—Ti[OCH(CH3)2]4 (97% Sigma-Aldrich), NH4SCN (97.5% Alfa Aesar), C2H5OH (200 Proof, Aaper Alcohol and Chemical Co.), and Co(NO3)2 (99.9+%, Alfa Aesar)—were used without further purification. Typically, Co/(C, S) TiO2 photocatalyst preparation involves 0.031 mols (8.5 g) of Ti[OCH(CH3)2]4 and 0.124 mols (9.44 g) of NH4SCN dissolved in 200 mL C2H5OH and stirred vigorously followed by dropwise addition of 0.125 mols (2.25 g) of deionized water containing the desired amount of Co(NO3)2 (0 2 mol %). The contents of the reaction were stirred for 5 min for complete hydrolysis, and the solvent was then removed by using a rotavap apparatus. After keeping them in a drying cabinet overnight, the samples were annealed in a Chamber Furnace (Carbolite, CWF-1100) to 500 °C and were held for 2 h. The final product was ground well into fine particles. The same method was used to prepare TiO2, (C, S) TiO2, and Co/TiO2 as controls. Characterization. After annealing at 500 °C for 2 h in air, the samples were characterized by various spectroscopic techniques. The powder X-ray diffraction (XRD) patterns of samples were recorded on a Scintag XDS 2000 D8 diffractometer with Cu KR radiation of wavelength 1.5406 Å and were analyzed from 15° to 75° (2θ) with a step size of 0.05° and a step time of 3 s to determine the crystalline phase. The specific surface areas were determined from nitrogen adsorption desorption isotherms recorded at liquid nitrogen temperature (77 K) on a Quantachrome Instrument (NOVA 1000 Series) by using the Brunauer Emmett Teller (BET) method. The pore size distributions were derived from the BJH desorption isotherms on the basis of the Barrett Joyner Halenda (BJH) method. UV vis optical absorption spectra were collected on a Cary 500 scan UV vis near-infrared (NIR) spectrometer from 200 to 800 nm by using poly(tetrafluorethylene) (PTFE) as a reference. Energy-dispersive X-ray (EDX) measurement was performed on scanning electron microscope-S-3500N and absorbed electron detectorS-6542 (Hitachi Science Systems, Ltd.), EDXA (Inca Energy, Oxford Instruments Microanalysis Ltd.), to determine the surface

ARTICLE

composition of the samples under the specified conditions of 20 keV and 15 mm working distance. Transmission electron microscopy (TEM) images were taken with a Philips CM100 operating at 100 kV. The TEM samples were prepared on a carbon-coated Formvar copper grid and were allowed to dry under ambient conditions for 5 h. The photoluminescence (PL) emission spectra for the hydroxyl radical generation and the electron hole separation were measured with a Fluoro Max-2 instrument from HORIBA Jobin Yvon Company at room temperature. To determine the amount of the hydroxyl radicals, 56 mg of photocatalyst sample was suspended in 80 mL of 3 mM terephthalic acid prepared in 0.01 M NaOH aqueous solution. The suspension was sonicated for 5 min and was stirred in the dark for 10 min, and a 5 mL aliquot of the dispersion was extracted and centrifuged. Then, 3 mL of the supernatant was taken for PL measurements before and after visible light illumination. In principle, hydroxyl radicals, if produced, readily react with terephthalic acid to form 2-hydroxyterephthalic acid that gives a PL emission peak at 425 nm upon excitation by 315 nm light. Thus, the PL emission intensity of the 2-hydroxyterephthalic acid of the supernatant solution directly gives the amount of the hydroxyl radicals. To study the electron hole separation, 100 mg of photocatalyst sample was pelletized and was used for the PL measurement at 420 nm excitation. Photoactivity Test. Photoactivity tests were performed in a 305 mL cylindrical air-filled static glass reactor with a quartz window at room temperature (Figure S1 of the Supporting Information). The 0.103 g photocatalyst samples were placed in a circular glass dish sample holder. Then, 100 μL liquid CH3CHO (0.08 g) was added into the bottom of the reactor and was constantly stirred for 40 min in order to allow acetaldehyde to vaporize and equilibrate. After 40 min stirring in dark conditions, 35 μL acetaldehyde air mixtures were periodically extracted (every 10 min) and were injected into the gas Gas Chromatography-Mass Spectrometry (GC-MS) port (Shimadzu GCMS-QP 5000). The temperatures of the column, injector, and detector were maintained at 40, 200, and 280 °C, respectively. The dark sampling was performed five times in order to examine the dark activity of the sample. For visible light photocatalysis, the sample was irradiated with a 1000 W high-pressure mercury lamp (Oriel Corp.) at a distance of 20 cm from the top by using combined filters that contained one vis NIR long pass filter (400 nm) and another colored glass filter (>420 nm) (see Figure S1 of the Supporting Information). After visible light was turned on, nine injections (35 μL each time) from the reactor were made at every 10 min interval to examine the photoactivity of the sample. The UV light photocatalysis was performed exactly in the same way by using two cutoff filters that transmit light of wavelength 320 400 nm. To compare photoactivity of photocatalyst samples, initial rates of acetaldehyde degradation to carbon dioxide were determined from the slopes of the straight lines of the concentration versus time plots with standard deviation less than 5% and the coefficient of linearity R2 = 0.995 0.999. For this, one dark point (40 min) to four light on points (50, 60, 70, and 80 min) in the concentration versus time plot of acetaldehyde and carbon dioxide were selected and were linearly fitted to get the initial rate (mmol/min) of acetaldehyde degradation and carbon dioxide production.

’ RESULTS AND DISCUSSION Characterization. Energy-dispersive X-ray spectral analysis showed the presence of codoped elements in various samples 17360

dx.doi.org/10.1021/jp200405y |J. Phys. Chem. C 2011, 115, 17359–17367

The Journal of Physical Chemistry C annealed at 500 °C in air (Figure S2 of the Supporting Information). During hydrolysis of titanium(IV) isopropoxide, the presence of ammonium thiocyanate as nonmetal precursor resulted in codoping of both carbon and sulfur into the TiO2 photocatalyst (Figure S2a, c, and d of the Supporting Information), whereas its absence resulted in only Co TiO2 without C and S (Figure S2b of the Supporting Information). From the EDX spectrum, it can be seen that the doped nitrogen was absent. Table S1 of the Supporting Information shows an average amount of codoped carbon and sulfur present on the surface of the annealed TiO2 samples with 2% Co-ion loading. Powder X-ray diffraction analysis of the annealed samples confirmed the formation of only anatase TiO2 (Figure 1). Further analysis of the XRD profile of cobalt-doped TiO2 with or without codoped carbon and sulfur ruled out the formation of cobalt oxides and sulfides as separate phases. This indicates that codoped Co, C, and S are uniformly dispersed or are coordinated in such a way that the X-rays are insensitive to detect them (that is, no crystallites of graphite or sulfur species are present). This can be explained on the basis of the crystallite size of materials under study. The smaller the crystallite size, the higher is the dispersion of codoped elements. The crystallite size of each sample determined by using the Debye Scherrer equation (crystallite size = 0.89λ/β cos θ,

Figure 1. Powder X-ray diffraction patterns of the undoped and codoped TiO2 samples.

ARTICLE

where λ = wavelength of X-ray, β = full width at half-maxima, and θ = diffraction angle) for the most intense (101) peak at around 2θ = 25.25° is shown in the respective XRD pattern (Figure 1). Obviously, the crystallite size of the anatase TiO2 codoped with C and S is smaller (8.4 nm) that that of Co(II) or Co(III) TiO2 (15 and 17.5 nm) and undoped TiO2 samples (20.4 and 21 nm) implying that codoping of C and S resists substantially the aggregation and coalescence of the anatase TiO2 nanocrystals during the annealing step at 500 °C in air as compared with doping Co only. This antisintering effect of the codoped nonmetals could be seen in the crystallite sizes of 1% Co(II)/(C, S) TiO2 (8.5 nm), 2% Co(II)/(C, S) TiO2 (7.4 nm), and 2% Co(III)/(C, S) TiO2 (6.6 nm). To confirm the existence of the nanosized domains, we obtained the TEM images of 2% Co(II)/(C, S) TiO2 sample. Now, it is clear that this sample actually constitutes very small nanocrystals (Figure 2a) and that the mean size by this method is 5.7 ( 2 nm (Figure 2b). The formation of very small TiO2 nanoparticles ( 420 nm) because the onset of its band edge is shifted to the visible region at about λ = 548.7 nm (Figure 3). Now, it is interesting to note how fast acetaldehyde degrades over various photocatalysts. From Table 2, the photodegradation of acetaldehyde proceeds 2.3, 4.6, and 5.8 times faster over 2% Co(II) TiO2, 1% Co(II)/(C, S) TiO2, and 2% Co(II)/(C, S) TiO2 systems as compared with the (C, S) TiO2 photocatalyst under visible light irradiation resulting in 5.0, 8.9, and 15.4 times production of carbon dioxide, respectively. Moreover, the 2% Co(II)/(C, S) TiO2 sample degrades acetaldehyde 2.5 and 1.9 times faster than the 2% Co(II) TiO2 along with 2.6 and 2.0 fold concomitant increase in carbon dioxide production under visible light and UV light, respectively. Also, the 2% Co(II)/(C, S) TiO2 sample has the ability to degrade acetaldehyde 7.0 and 22.5 times faster than the TiO2 P25 and undoped TiO2 photocatalysts with almost 2-fold increase in carbon dioxide production under UV light. These results proved that the doped cobalt has a significant effect on the photoactivities of the (C, S) TiO2 system indicating that the codoping of cobalt, carbon, and sulfur is essential to achieve high photoactivities for both visible and UV light irradiations. Furthermore, the enhancement in visible light activity could be attributed to the further red shift of the onset of the absorption edge in visible region at about λ = 756.4 762.5 nm (Figure 3). Consequently, this additional red shift enables Co/(C, S) TiO2 photocatalysts to absorb more visible light photons as compared with (C, S) TiO2 and undoped TiO2 systems. Yet, of the various attributes of ideal photocatalysts, the stability and the reusability of the photocatalyst are important for long-term applications. For this purpose, we performed successive long-run experiments with the 2% Co(II)/(C, S) TiO2 photocatalyst under visible light irradiation (Figure 4). In each experiment, acetaldehyde was completely degraded over five hours, and carbon dioxide production was slowed. Note that 2 mmol aldehyde was degraded yielding about 0.4 mmol CO2. Clearly, other products, such as acetic acid and formic acid, are also formed and remain in the reaction chamber under the analysis conditions employed. At any rate, these results show that the 2% Co/(C, S) TiO2 system is stable and reusable without any significant loss in photoactivity as indicated by the small error bars in Figure 4a and b. Thus, we conclude that cobalt carbon sulfur-codoped TiO2 nanoparticle photocatalysts have promise for harvesting abundant solar energy for environmental remediation. Recognition of Active Sites. Now, it seems important to know and understand the active sites for enhanced visible-light-induced photodecomposition of gas-phase acetaldehyde over 2% Co/ (C, S) TiO2 samples. To elucidate the active sites, we examined

Table 2. Initial Rates of Acetaldehyde Degradation and Carbon Dioxide Production over the Undoped and Codoped Titania Photocatalysts under Visible and UV Light Irradiationsa photocatalysts TiO2 P25

a

r[CH3CHO]VIS nil

TiO2

nil

(C, S) TiO2

4.6  10

1% Co(II)/(C, S) TiO2 2% Co(II)/(C, S) TiO2 2% Co(II)/TiO2

r[CO2]VIS

r[CH3CHO]UV

nil nil

r[CO2]UV

5.1  10

3

4.0  10

3

4.2  10

3

1.6  10

3

2.8  10

4

17  10

3

3.7  10

3

21  10

3

2.5  10

3

27  10

3

7.0  10

3

27  10 11  10

3

4.3  10 1.6  10

3

36  10 19  10

3

8.0  10 3.8  10

3

3

3

3

3

3

Linear fit initial rate, r, is expressed in mmol/min. 17362

dx.doi.org/10.1021/jp200405y |J. Phys. Chem. C 2011, 115, 17359–17367

The Journal of Physical Chemistry C

ARTICLE

Table 3. Effect of Initial Valence State of Cobalt Ions on the Initial Rate of Acetaldehyde Degradation and Carbon Dioxide Production under Visible Light Irradiationa cobalt

designated

precursors

sample

r[CH3CHO]VIS

r[CO2]VIS

2% Co(II) TiO2

11  10

3

1.6  10

3

2% Co(III) TiO2

5.2  10

3

1.0  10

3

Co(NO3)2 + NH4SCN

2% Co(II/

27  10

3

4.3  10

3

Co(acac)3 + NH4SCN

(C, S) TiO2 2% Co(III)/

28  10

3

4.0  10

3

Co(NO3)2 Co(acac)3

(C, S) TiO2 a

Figure 4. Long-run kinetic plot for (a) CH3CHO degradation and (b) CO2 evolution over the 2% Co(II)/(C, S) TiO2 sample under visible light irradiation.

the effect of valence states of cobalt precursors on the initial rate of visible-light-induced photodegradation of acetaldehyde with or without codoped carbon and sulfur precursor (Table 3). Here, we found that the 2% Co(II) TiO2 sample degrades acetaldehyde 2.0 times faster than the 2% Co(III) TiO2 sample under visible light. This agrees with the rate of carbon dioxide formation (Table 3) and the UV vis absorption spectra of the samples (Figure 3). However, in the presence of nonmetal precursors, both 2% Co(II)/(C, S) TiO2 and 2% Co(III)/(C, S) TiO2 samples exhibited remarkable activities for CH3CHO decomposition and CO2 evolution under visible light irradiation regardless of the initial valence state of cobalt in its precursors (Table 3 and Figure S4 of the Supporting Information). The results proved that the doped cobalt ions, which could be Co(II), Co(III), or both, act as active sites for enhanced visible light photocatalysis. Both samples, 2% Co(II)/(C, S) TiO2 and 2% Co(III)/(C, S) TiO2, have equal BET surface areas (89 m2/g and 88 m2/g) with the same amount of the doped cobalt ions. To determine the chemical valence state of Ti, O, C, S, and Co, we measured the X-ray photon spectroscopy (XPS) spectra of all samples and determined the core level binding energy (BE) for each element. For simplicity, a typical XPS spectrum of 2% Co(II)/(C, S) TiO2 is shown in Figure 5, and binding energy values of all constituent elements are shown in Table 4. The Ti2p XPS spectrum (Figure 5a) and binding energy values for Ti2p3/ 2 and Ti2p1/2 spin orbit doublet confirmed that titanium is present as Ti4+.4 6 The O1s XPS spectrum with a shoulder (Figure 5b) and BE values showed that oxygen is present as O2 in the O Ti lattice and as OH M on the surface.4 6 The C1s XPS

Initial rate, r, is expressed in mmol/min.

spectrum (Figure 5c) and BE values indicated that carbon is mostly present as graphitic carbon.7 The S2p3/2 XPS spectrum (Figure 5d) and BE values showed that sulfur is present as sulfate species.8,9 The bulk elemental analysis of Co(II)/(C, S) TiO2 sample showed that cobalt is 1.2 at. (atomic percent), that carbon is less than 0.5 at.%, and that sulfur is 2 at.%. These results suggest that the graphite-like carbon is mostly distributed on the surface whereas the doped cobalt ions and the doped sulfur as sulfate are distributed on the surface and throughout the bulk of the TiO2 nanocrystals. For cobalt, it is difficult to differentiate Co(II) and Co(III) species from the measured Co2p3/2 core level binding energy values (Table 4). However, some clear features in the Co2p core level spectra will help us to differentiate Co(II) and Co(III) oxidation states. If the Co(II) species is present, the Co2p XPS spectra exhibit strong shakeup satellite peaks at higher binding energies. Figure 6a shows the Co2p XPS spectrum of 2% Co(II)/ (C, S) TiO2 sample before visible light photocatalysis. The shakeup satellite features obtained in the present study are similar to the reported ones31 confirming the presence of Co(II) species. Figure 6b shows the Co2p core level XPS spectrum of 2% Co(III)/(C, S) TIO2 sample before visible light photocatalysis. Since the shakeup satellite peaks of Co(III) are weaker than that of Co(II), the Co2p features in Figure 6b confirm the presence of Co(III) species. However, after visible light photocatalysis, the shakeup satellite features in Figure 6a became very weak indicating the possible oxidation of Co(II) to Co(III). Interestingly, after visible light photocatalysis, no distinct shakeup satellite peaks appeared in Co2p XPS spectrum (Figure 6d). This ruled out the thermodynamically feasible reduction of Co(III) to Co(II). From these results, we conclude that in addition to C and S dopants, the Co(III) species act as active sites for the photocatalyzed degradation of acetaldehyde upon visible light excitation. This is further supported with the measured binding energy value of Co in 2% Co(II) or Co(III)/(C, S) TiO2 samples before and after photocatalysis with visible light (Table 4). Overall, the XPS results confirmed that the Co(III) species are predominantly present in Co/(C, S) TiO2 systems and act as active sites for acetaldehyde degradation under visible light irradiation. If the Co(III) species were not the active sites in the presence of codoped C and S into TiO2, the long-run kinetic results obtained for acetaldehyde degradation over 2% Co(II)/(C, S) TiO2 system under visible light irradiation could not have been as reproducible (Figure 4). Role of Codoped Cobalt, Carbon, and Sulfur in VisibleLight-Induced Photocatalysis. It is known that heterogeneous photocatalysis is a complex process and that its outcome is affected by several factors, such as crystalline structure, surface area, band gap, reactive oxygen species, and electron hole separation. It is 17363

dx.doi.org/10.1021/jp200405y |J. Phys. Chem. C 2011, 115, 17359–17367

The Journal of Physical Chemistry C

ARTICLE

Figure 5. XPS spectra for (a) Ti, (b) O, (c) C, and (d) S of the 2% Co/(C, S) TiO2 sample.

Table 4. Core Level Binding Energy (eV) for Ti, O, C, S, and Co sample

Ti2p3/2

O1s

C1s

S2p3/2

(C, S) TiO2 2% Co(II)TiO2

458.6 459.5

529.9 530.8

284.7

168.9

2% Co(II)/(C, S) TiO2 before reaction

459.6

530.9

285.9

170.1

782.1

2% Co(II)/(C, S) TiO2 after reaction

459.6

531.0

285.5

170.1

782.0

2% Co(III)/(C, S) TiO2 before reaction

459.6

530.9

285.2

169.9

782.3

2% Co(III)/(C, S) TiO2 after reaction

459.6

530.9

285.3

169.9

782.6

important to know how codopants influence some characteristic properties of the titania photocatalyst. BET Surface Area Increase. Specific surface areas of pure TiO2 nanomaterials depend upon the synthesis routes and calcination temperatures. Even though the present synthesis route resulted in very low surface areas of pure TiO2 (5.4 m2/g) sample, the codoped samples, Co/(C, S) TiO2 and (C, S) TiO2, have high BET surface areas in the range of 69 89 m2/g (Table 1). This indicates that the codoped cobalt, carbon, and sulfur could prevent the sintering of TiO2 nanoparticles during the annealing step possibly because of the evolution of burned-out gases and chelating effect of the doped metal ions. Thus, cobalt-, carbon-, and sulfur-codoping improved the specific surface areas of the codoped TiO2 nanoparticles providing more absorption/desorption sites for photocatalytic reactions, which is an added benefit for enhancing activity. Band Gap Reduction. One purpose of doping or codoping TiO2 with metal/nonmetals is to reduce its wide band gap (3.2 eV). Clearly, the codoped nonmetals, carbon and sulfur, reduced the band gap of the TiO2 from 3.2 to 2.26 eV (Table 5) indicating visible light utilization. In the absence of codoped

Co2p3/2

782.6

carbon and sulfur, the doped cobalt ions also reduced the band gap of TiO2 from 3.2 to 2.32 eV and 2.69 eV with an additional interband state of 1.6 eV (possibly because of O2 f Co(III)).32 Moreover, the doped cobalt ions introduced two optical band edges, one 420 nm) at room temperature can be used to investigate the photogenerated charge-carrier separation and to understand why codoping TiO2 with cobalt, carbon, and sulfur led to enhanced visible light activity for CH3CHO degradation. Basically, photoluminescence is a process that involves emission of light photons when excited conduction band electrons recombine with valence band holes of semiconductors.35 A high PL signal indicates a high rate of electron hole pair recombination under light irradiation (λ g 420 nm); a low PL signal indicates the separation of 17365

dx.doi.org/10.1021/jp200405y |J. Phys. Chem. C 2011, 115, 17359–17367

The Journal of Physical Chemistry C

ARTICLE

Figure 9. Proposed reaction pathways for the generation and separation of electron hole pairs on Co/(C, S) TiO2 photocatalyst under visible light irradiation. Figure 8. Photoluminescence (PL) emission spectra of the codoped TiO2 samples obtained at room temperature at 420 nm excitation wavelength.

photogenerated electron hole pairs. As the probability of the electron hole pair separation increases, the photocatalytic activity will supposedly increase. Figure 8 shows the PL spectra of the codoped TiO2 samples. We assume that the prominent peaks at around 464 465 nm and 483 485 nm are possibly ascribed to the band gap recombination of electron holes pairs and to the electron transition mediated by intrinsic oxygen vacancies, respectively. Moreover, the results from Figure 8 prove that the doped cobalt ions, with or without C and S, are very effective in electron hole pair separation. This is possible because the Co(III) ions can scavenge or trap conduction electrons effectively because of its high reduction potential (Co(III) + e = Co(II), E0 = +1.81 V). Also, the doped cobalt ions could induce p-type conductivity onto the TiO2 samples. This improves the mobility and migration of the valence band holes toward the catalyst surface. This could be one reason that the 1% and 2% Co(II)/(C, S) TiO2 photocatalysts demonstrated higher activities for CH3CHO degradation than the (C, S) TiO2 photocatalyst under visible light (Table 2). Proposed Reaction Mechanism for Visible-Light-Induced Acetaldehyde Degradation. Photocatalysis processes involve both reductive and oxidative reaction pathways of the photoinduced charge carriers (electron hole pairs) for degrading organic pollutants. On the basis of the above characterization and activity results, possible reaction pathways for generation, separation, and involvement of electron hole pairs during acetaldehyde degradation over the Co/(C, S) TiO2 nanoparticle sample are shown in Figure 9. In the presence of the codoped carbon and sulfur, visible light photon excites an electron from the ground state to the excited state of the doped graphite-like carbon.36 The excited electron is transferred into the conduction band of TiO2. The conduction band electron is then subsequently transferred to oxygen molecules adsorbed on the catalyst surface directly or via the doped sulfur as sulfates (Lewis acid sites) forming reactive superoxide anions (O2 + eCB f O2 ) that degrade acetaldehyde (O2 + CH3CHO f CO2 + H2O) on the (C, S) TiO2 catalyst surface in a series of reaction events. When cobalt is present along with the doped carbon and sulfur, the scenario is, however, different in that the doped cobalt ions act as more efficient electron traps (Co(III) + eCB f Co(II)) than the adsorbed oxygen and sulfate species. This results in even more efficient migration and separation of conduction band electrons over the 2% Co/(C, S) TiO2 photocatalyst than (C, S) TiO2 producing more mobile holes in the hybridized

valence band (O2p + Co3d)37 of the codoped TiO2 photocatalyst at the same photoexcitation event. The valence band holes (E0 = +2.96 V) that are more powerful oxidants than hydroxyl radicals (E0 = +2.76 V) and superoxide anions (E0 = +0.56 V) are transferred to the catalyst surface because of the p-type doping effect of cobalt ions. These holes are responsible for further mineralization of acetaldehyde molecules (hVB+ + CH3CHO f CO2 + H2O) on the catalyst surface. This is supported with the photoactivity and PL results (Table 2, Figures 7 and 8). To rejuvenate the photocatalytic cycle, valence band holes oxidize conduction band electron reduced Co(II) to Co(III) ions (hVB+ + Co(II) f Co(III)). This catalytic cycle of the doped cobalt ions, irrespective of their initial valence states in precursor materials, continues as long as visible light is irradiated over the 2% Co/(C, S) TiO2 catalyst. This proves that the doped cobalt ions improve the charge separation and enhance the oxidation ability of photogenerated holes, no matter free or trapped, to degrade acetaldehyde over 2% Co/(C, S) TiO2 photocatalyst under visible light. Also, other factors—such as surface area, anatase crystallinity, introduction of interband state, and OH radical generation—partly contributed to the optimum photoactivity of the 2% Co/(C, S) TiO2 system for acetaldehyde degradation under visible light (λ > 420 nm). Zhu et al.38 reported that noncompensated nature of n p codoping of n- and p-type dopants with unequal charge states ensures the creation of tunable intermediate bands that effectively narrows the band gap and results in efficient electron hole separation in the visible light region. This could be one reason for enhanced visible light photoactivity of the 2% Co/(C, S) TiO2 photocatalyst prepared by simultaneous incorporation of C, S, and Co with n- and p-type nature and unequal charge states. The same authors also found that the presence of N breaks up clustering of Cr ions resulting in a more efficient photocatalyst, Cr N TiO2. Therefore, we assume that the presence of C, S improves surface area and the dispersion of Co ions that behave as active sites for visible light absorption and catalysis. Using electron paramagnetic resonance (EPR) spectroscopy, Li et al.39 concluded that the presence of both highly dispersed CuO clusters and substitutional Cu2+ sites (Ti O Cu linkages) account for the improved photooxidative activity of the 0.1 wt % CuO TiO2 nanocomposite. Hurum et al.40 also concluded that the high photoactivity of a mixed-phase titania photocatalyst is due to fast electron transfer from the small-sized rutile crystallites to anatase that makes catalytic hot spots at the rutile/anatase interface. However, the present photocatalyst system, Co/(C, S) TiO2, does not have mixed-phase titania particles other than the codoped 17366

dx.doi.org/10.1021/jp200405y |J. Phys. Chem. C 2011, 115, 17359–17367

The Journal of Physical Chemistry C cobalt, carbon, and sulfur. Thus, to gain more insights on the role of Co, C, and S codopants in visible light absorption and electrontransfer mechanism, we hope to study the Co/(C, S) TiO2 photocatalyst system with EPR technique in the future.

’ CONCLUSIONS Cobalt carbon sulfur-codoped anatase TiO2 nanoparticles were prepared in one step and were developed as a new visible light active photocatalyst. This new photocatalyst system, Co/ (C, S) TiO2, demonstrated high photoactivity for acetaldehyde degradation as compared with the Co TiO2 and (C, S) TiO2 photocatalysts under visible light irradiation. Enhancement in photoactivity of the Co/(C, S) TiO2 system is basically attributed to the key role of the codoped cobalt ions that increase surface area, extend visible light absorption up to 762 nm, generate hydroxyl radicals, and improve electron hole pair separation. Moreover, the 2% Co/(C, S) TiO2 photocatalysts of the same specific surface areas exhibited almost the same photoactivities for acetaldehyde photodegradation irrespective of the initial valence state of cobalt ion in precursor materials. This study will provide some useful strategies to prepare new visible-light-driven photocatalysts codoped with other types of transition-metal ions and nonmetals. ’ ASSOCIATED CONTENT

bS

Supporting Information. EDX Table S1 and Figures S1 S4. This material is free of charge via the Internet at http:// pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected].

’ ACKNOWLEDGMENT The authors thank NanoScale Corporation and the U.S. Army Research Office for financial support to complete the research. Also, Dr. Dan Boyle, Division of Biology; Myles Ikenberry and Professor Keith Hohn, Department of Chemical Engineering, Kansas State University, are acknowledged with gratitude. ’ REFERENCES (1) Nagaveni, K.; Hegde, M. S.; Ravishankar, N.; Subbanna, G. N.; Madras, G. Langmuir 2004, 20, 2900–2907. (2) Asahi, R.; Morikawa, T.; Ohwaki, T.; Aoki, K.; Taga, Y. Science 2001, 293, 269–271. (3) Lin, L.; Lin, W; Zhu, Y.; Zhao, B.; Xie, Y. Chem. Lett. 2005, 34, 284–285. (4) Ohno, T.; Mitsui, T.; Matsumura, M. Chem. Lett. 2003, 32, 364–365. (5) Li, D.; Haneda, H.; Labhsetwar, N. K.; Hishita, S.; Ohashi, N. Chem. Phys. Lett. 2005, 401, 579–584. (6) Hong, X.; Wang, Z.; Cai, W.; Lu, F.; Zhang, J.; Yang, Y.; Ma, N.; Liu, Y. Chem. Mater. 2005, 17, 1548–1552. (7) Sun, H.; Bai, Y.; Cheng, Y.; Jin, W.; Xu, N. Ind. Eng. Chem. Res. 2006, 45, 4971–4976. (8) Cong, Y.; Chen, F.; Zhang, J.; Anpo, M. Chem. Lett. 2006, 35, 800–801. (9) Li, D.; Haneda, H.; Hishita, S.; Ohashi, N. Chem. Mater. 2005, 17, 2588–2595. (10) Ozaki, H.; Iwamoto, S.; Inoue, M. Chem. Lett. 2005, 34, 1082–1083.

ARTICLE

(11) Sakatani, Y.; Ando, H.; Okusako, K.; Koike, H.; Nunoshige, J.; Takata, T.; Kondo, J. N.; Hara, M.; Domen, K. J. Mater. Res. 2004, 19, 2100–2108. (12) Zhao, W.; Ma, W.; Chen, C.; Zhao, J.; Shuai, Z. J. Am. Chem. Soc. 2004, 126, 4782–4783. (13) Sakatani, Y.; Nunoshige, J.; Ando, H.; Okusako, K.; Koike, H.; Takata, T.; Kondo, J. N.; Hara, M.; Domen, K. Chem. Lett. 2003, 32, 1156–1157. (14) Liu, H.; Gao, L. Chem. Lett. 2004, 33, 730–731. (15) Tryba, B.; Morawski, A. W.; Inagaki, M.; Toyoda, M. J. Photochem. Photobiol., A: Chem. 2006, 179, 224–228. (16) Bettinelli, M.; Dallacasa, V.; Falcomer, D.; Fornasiero, P.; Gombac, V.; Montini, T.; Romano, L.; Speghini, A. J. Hazard. Mater. 2007, 146, 529–534. (17) Yang, X.; Cao, C.; Hohn, K.; Erickson, L.; Maghirang, R.; Hamal, D.; Klabunde, K. J. Catal. 2007, 252, 296–302. (18) Hamal, D. B.; Klabunde, K. J. J. Colloid Interface Sci. 2007, 311, 514–522. (19) Klosek, S.; Raftery, D. J. Phys. Chem. B 2001, 105, 2815–2819. (20) Shi, J. Y.; Leng, W. H.; Zhu, W. C.; Zhang, J. Q.; Cao, C. N. Chem. Eng. Technol. 2006, 29, 146–154. (21) Othman, I.; Mohamed, R. M.; Ibrahem, F. M. J. Photochem. Photobiol., A: Chem. 2007, 189, 80–85. (22) Zhu, J.; Chen, F.; Zhang, J.; Chen, H.; Anpo, M. J. Photochem. Photobiol., A: Chem. 2006, 180, 196–204. (23) Iwasaki, M.; Hara, M.; Kawada, H.; Tada, H.; Ito, S. J. Colloid Interface Sci. 2000, 224, 202–204. (24) Chen, S.; Zhang, S.; Liu, W; Zhao, W. J. Hazard. Mater. 2008, 155, 320–326. (25) Colon, G.; Maicu, M.; Hidalgo, M. C.; Navio, J. A. Appl. Catal., B: Environ. 2006, 67, 41–51. (26) Jing, L.; Xin, B.; Yuan, F.; Xue, L.; Wang, B.; Fu, H. J. Phys. Chem. B 2006, 110, 17860–17865. (27) Martyanov, I. N.; Uma, S.; Rodrigues, S.; Klabunde, K. J. Langmuir 2005, 21, 2273–2280. (28) Kapoor, P. N.; Uma, S.; Rodriguez, S.; Klabunde, K. J. J. Mol. Catal. A: Chem. 2005, 229, 145–150. (29) Long, M.; Cai, W.; Cai, J.; Zhou, B.; Chai, X.; Wu, Y. J. Phys. Chem. B 2006, 110, 20211–20216. (30) Wu, Y.; Lu, G.; Li, S. J. Photochem. Photobiol., A: Chem. 2006, 181, 263–267. (31) Li, J. G.; Buchel, R.; Isobe, M.; Mori, T.; Ishigaki, T. J. Phys. Chem. C 2009, 113, 8009–8015. (32) Zhang, Y.; Chen, Y.; Wang, T.; Zhou, J.; Zhao, Y. Microporous Mesoporous Mater. 2008, 114, 257–261. (33) Hirakawa, T.; Nosaka, Y. Langmuir 2002, 18, 3247–3254. (34) Yu, J.; Wang, W.; Cheng, B.; Su, B.-L. J. Phys. Chem. C 2009, 113, 6743–6750. (35) Yang, X.; Cao, C.; Erickson, L.; Hohn, K.; Maghirang, R.; Klabunde, K. J. Catal. 2008, 260, 128–133. (36) Dong, F.; Zhao, W.; Wu, Z. Nanotechnology 2008, 36, 365607/ 1–365607/10. (37) Yin, J.; Zou, Z.; Ye, J. J. Phys. Chem. B 2003, 107, 4936–4941. (38) Zhu, W.; Qiu, X.; Lancu, V.; Chen, X. Q.; Pan, H.; Wang, W.; Dimitrijevic, N. M.; Rajh, T.; Meyer, H. M.; Paranthaman, M.; Stocks, G. M.; Weitering, H. H.; Gu, B.; Eres, G.; Zhang, Z. Phys. Rev. Lett. 2009, 103, 226401/1–226401/4. (39) Li, G.; Dimitrijevic, N. M.; Chen, L.; Rajh, T.; Gray, K. A. J. Phys. Chem. C 2008, 112, 19040–19044. (40) Hurum, D. C.; Agrios, A. G.; Gray, K. A.; Rajh, T.; Thurnauer, M. C. J. Phys. Chem. B 2003, 107, 4545–4549.

17367

dx.doi.org/10.1021/jp200405y |J. Phys. Chem. C 2011, 115, 17359–17367