- Email: [email protected]
Fabrication and comparison of highly efficient Cu incorporated TiO2 photocatalyst for hydrogen generation from water
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 5 2 5 4 e5 2 6 1
Available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/he
Fabrication and comparison of highly efficient Cu incorporated TiO2 photocatalyst for hydrogen generation from water Shiping Xu, Jiawei Ng, Xiwang Zhang, Hongwei Bai, Darren Delai Sun* School of Civil and Environmental Engineering, Nanyang Technological University, 50 Nanyang Avenue, Singapore 639798, Singapore
article info
abstract
Article history:
Efficient Cu incorporated TiO2 (CueTiO2) photocatalysts for hydrogen generation were
Received 8 December 2009
fabricated by four methods: in situ solegel, wet impregnation, chemical reduction of Cu salt,
Received in revised form
and in situ photo-deposition. All prepared samples are characterized by good dispersion of
15 February 2010
Cu components, and excellent light absorption ability. Depending on the preparation
Accepted 26 February 2010
process, hydrogen generation rates of the as-prepared CueTiO2 were recorded in the range
Available online 31 March 2010
of 9e20 mmol h1 g1 catalyst, which were even more superior to some noble metal (Pt/Au)
Keywords:
as different distribution ratio of Cu between surface and bulk phases of the photocatalyst.
CueTiO2
Both factors have been proven to influence photocatalytic hydrogen generation. In addition,
Hydrogen
the Cu content in the photocatalyst played a significant role in hydrogen generation. Among
loaded TiO2. The various fabrication methods led to different chemical states of Cu, as well
Fabrication methods
the four photocatalysts, the sample that was synthesized by in situ solegel method exhibited the highest stability. High efficiency, low cost, good stability are some of the merits that underline the promising potential of CueTiO2 in photocatalytic hydrogen generation. Crown Copyright ª 2010 Published by Elsevier Ltd on behalf of Professor T. Nejat Veziroglu. All rights reserved.
1.
Introduction
Massive global utilization of fossil fuels has caused critical and irreversible environmental problems throughout the world. Since fossil fuels are non-renewable, alternative energy sources are urgently needed. Among available fuels, hydrogen has been regarded as an excellent energy carrier, due to its high energy content, zero emission of greenhouse gas, and ease of storage and distribution. Currently, the majority of hydrogen demands are sourced from steam reforming of natural gas; however, this process does not mitigate the dependence on fossil fuels, and emits environmentally unfriendly greenhouse gases. Since the discovery of photoelectrochemical water splitting at a TiO2 electrode by Fujishima and Honda [1], there has
been much emphasis on leveraging solar energy to photo-split water molecules. However, hydrogen generation efficiency over bare TiO2 is low, largely due to the fast recombination of electron/hole pairs [2]. Noble metal modified TiO2 has been proven to be very effective in overcoming this drawback [3e7], since the lower Fermi level of noble metals facilitates electron transfer from TiO2 to the loaded noble metal [8], resulting in an efficient separation of charge carriers, thus enhancing the potential of hydrogen generation. However, noble metals are expensive and rare; hence identification of common metals with similar enhancement to photocatalytic activity increasingly draws research attention. Cu is much less costly as compared to noble metals, and CueTiO2 has been studied extensively in the treatment of gaseous pollutants [9,10]. Interestingly, some researchers found that CueTiO2
* Corresponding author. Tel.: þ65 67906273; fax: þ65 67910676. E-mail address: [email protected] (D.D. Sun). 0360-3199/$ e see front matter Crown Copyright ª 2010 Published by Elsevier Ltd on behalf of Professor T. Nejat Veziroglu. All rights reserved. doi:10.1016/j.ijhydene.2010.02.129
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 5 2 5 4 e5 2 6 1
possessed satisfactory photocatalytic hydrogen generation ability. Dhanalakshmi et al. [11] and Sreethawong and Yoshikawa [12] reported that, in comparison to Pt, Au, and Pd metals, Cu showed only slightly weaker enhancement of TiO2 activity for hydrogen generation. In our previous study [13], we deposited CuO on TiO2, with an optimum Cu content and a high dispersion of CuO. Our experimental results exhibited excellent hydrogen generation activity, even higher than some noble metal loaded TiO2. Furthermore, several Cu species (CuO/Cu2O/metallic Cu) modified TiO2 have been proven to be active photocatalysts in reduction of water under sacrificial conditions [14e18]. This further affirms the huge potential of CueTiO2 in photocatalytic hydrogen generation. Hydrogen generation ability of photocatalysts depended on the synthesis method to a large extent. AueTiO2 as fabricated by photo-deposition showed greater activity in hydrogen generation than samples prepared by depositioneprecipitation and wet impregnation methods [19]. Likewise for CueTiO2 photocatalysts, fabrication methods would also affect photocatalytic activity, as well as characteristics of the end-products. Boccuzzi et al. [20,21] compared properties and activity of CueTiO2 prepared by wet impregnation and chemisorptionhydrolysis methods, and found that samples with the same chemical composition exhibited a marked difference of up to 100 times in the hydrogenation of 1,3-cyclooctadiene. Wu and Lee [14] reported that Cu doping within the TiO2 lattice had a negative effect on photocatalytic hydrogen generation as opposed to Cu deposition. Different hydrogen generation activity of CueTiO2 prepared via varying preparation methods had also been mentioned by Kang’s group [17,18]. However, majority of the reported CueTiO2 photocatalysts employed for hydrogen generation were fabricated by the wet impregnation method [13,16,17,22]. To date, no systematic study about the effect of CueTiO2 preparation methods on hydrogen generation rates has been reported. In this paper, we compared several fabrication methods, and this may pave way for fabrication of highly productive CueTiO2 photocatalysts for generation of hydrogen. This paper highlights four different methods that are widely used in the fabrication of metaleTiO2, namely, in situ solegel (SG), wet impregnation (WI), chemical reduction of Cu salt by NaBH4 (NR) [23], and in situ photo-deposition (PD) [24]. These four methods were employed to fabricate CueTiO2 photocatalysts, and their hydrogen generation activities were subsequently compared. It was discovered that the fabrication methods determined the chemical state of Cu, distribution ratio of Cu within the photocatalyst, BET surface area of the catalyst, and crystal structure of the TiO2 support. The influence on hydrogen generation activity will also be discussed in this paper.
2.
Experimental
2.1.
Preparation of CueTiO2 photocatalysts
With reference to our previous study [13], Cu content in all samples was controlled at 8 atom% Cu/Ti, with Cu(NO3)2$3H2O (Merck, GR) as the copper precursor.
5255
In the solegel process, Ti(OBu)4 (Merck) was employed as the titanium precursor. TiO2 was synthesized in a 1:20:6:0.8 molar ratio of Ti(OBu)4:C2H5OH (Merck, GR):H2O (Milli-Q, ultrapure):HNO3 (Merck, AR) [25]. 17 ml Ti(OBu)4 was dissolved in 38 ml absolute ethanol with stirring for 10 min; 2.5 ml HNO3 was then added dropwise into the above solution and vigorously stirred for 30 min. Another solution containing 20 ml absolute alcohol, 4.3 ml H2O, and required amount of Cu(NO3)2 was slowly added into the above solution, to produce transparent sol under rigorous stirring. After 7 days of aging, the obtained gel was evaporated at 378 K, and calcined at 723 K for 4 h to obtain the final product. SGeTiO2 was produced via identical procedures, with the exception of omitting Cu salt addition. Degussa P25 (Germany) was employed as the TiO2 precursor in the following three processes. For WI sample, P25 powder was added into Cu(NO3)2 solution at a ratio of 1 g/3 ml, followed by ultrasonication and thorough mixing to obtain a homogeneous slurry. Concentration of Cu(NO3)2 solution was adjusted to control Cu content in the product. After wet impregnation, the sample was dried at 378 K overnight, and finally calcined in air for 4 h at 723 K. For NR sample, solid NaBH4 (Merck, GR) was added into a vigorously stirred, homogeneous suspension containing Cu (NO3)2 and P25 in an ice water bath. The amount of NaBH4 and Cu(NO3)2 were decided by complete reduction of a required amount of Cu on TiO2. During the reduction process, the originally white TiO2 was turned into a chocolate-brown coloured precipitate. After continuous mixing for 2 h, the colour turned blackish green. Final product was recovered and washed with large amount of ultrapure water, and dried at 378 K overnight. In situ photo-deposition was employed to prepare PD sample. P25 suspension was first mixed and sonicated to break up loosely attached aggregates. Required amounts of Cu (NO3)2 and methanol (Fisher, HPLC) were then added, and made up to the final volume in view of photocatalytic hydrogen generation, which would be described later.
2.2.
Characterization of CueTiO2 photocatalysts
X-ray diffraction (XRD) patterns were obtained using a Bruker D8 Advance X-ray diffractometer with monochromated highintensity Cu Ka radiation (l ¼ 1.5418 A). X-ray photoelectron spectroscopy (XPS) analyses were carried out in an ultrahigh vacuum chamber with a base pressure below 2.66 107 Pa at room temperature. Photoemission spectra were recorded by a Kratos Axis Ultra spectrometer equipped with a standard monochromatic Al Ka excitation source (hn ¼ 1486.71 eV). All binding energies were referenced to C 1s at 284.8 eV. Sample morphologies were observed by transmission electron microscope (TEM, Jeol JEM-2010) and scanning electron microscope (SEM, Zeiss Evo 50); surface element composition was detected using the energy dispersive X-ray spectrometer (EDS) attached to the SEM. BET surface area and pore size distribution were determined at liquid nitrogen temperature (77 K) using a Micromeritics ASAP 2010 system. Absorption spectra of CueTiO2 photocatalysts were recorded by Thermo Scientific Evolution 300 UVevisible spectrophotometer
5256
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 5 2 5 4 e5 2 6 1
equipped with an integration sphere. Inductively coupled plasma emission spectroscopy (ICP, Perkin Elmer Optima 2000DV) was employed to determine bulk Cu content in the CueTiO2 catalysts, and quantify dissolved Cu in the solution during hydrogen generation.
2.3.
Photocatalytic hydrogen generation
The photocatalytic reaction was performed in an innerirradiation type Pyrex reactor (volume: 1.35 l) with a 400 W high pressure Hg lamp (Riko, UVL-400HA) as the light source. To maintain a constant reactor temperature of 298 K, a quartz water jacket, which was cooled by recycled water, was utilized to cover the lamp. Powdered CueTiO2 was suspended in 10 volume% methanol water mixture at a concentration of 1 g/l. A magnetic stirrer was placed at the bottom of the reactor to homogeneously churn the suspension during the reaction. Prior to irradiation, the photocatalyst suspension was deaerated thoroughly for 30 min by nitrogen gas purging. Gas produced via the photocatalytic reaction was analyzed using a TCD-type gas chromatography (HP 5890A, Carboxen 1010 plot). pH value of the solution was determined by a pH meter (Horiba F-53).
3.
Results and discussion
3.1.
Characterization of CueTiO2
Fig. 1 shows the XRD patterns of the as-prepared CueTiO2 photocatalysts, with P25 as comparison. In the XRD patterns of WI and NR samples, signals of Cu components were clearly observed. In the case of WI sample, diffraction peaks at 2q ¼ 35.5 and 38.7 were ascribed to the respective (1 1) and
(1 1 1) planes of CuO (JCPDS 48-1548); while diffraction peaks at 2q ¼ 36.4 and 42.3 were respectively identified as (1 1 1) and (2 0 0) planes of Cu2O (JCPDS 05-0667) for the NR sample. However, no obvious phases of Cu components were detected in SG and PD samples. As results of EDS and ICP confirmed the existence of Cu in all photocatalysts, it is believed that the dimensions of Cu components in SG and PD samples were below detection limit of the XRD, thus evidencing high copper dispersion in the samples [26]. Taking into account the fabrication process, Cu component in PD sample might be in ground metal state, since in situ photocatalytic reduction of Cu salt favored formation of metallic Cu [27]. The chemical state of Cu in SG sample was identified by XPS analysis. With regard to TiO2, the XRD pattern of SG sample exhibited characteristic features of anatase phase (JCPDS 21-1272), with an average crystal size of 8.5 nm, as calculated by Scherrer’s equation from the full width at half maximum of anatase (1 0 1), after correction for instrumental broadening. For P25-based photocatalysts (WI, NR, PD), no significant changes of P25 diffraction patterns were observed, indicating that the TiO2 support had scarcely undergone changes in particle size and phase transformations during the synthesis processes. Quantitative XPS analysis of SG sample was performed, and the survey spectrum and high-resolution scans are shown in Fig. 2. From the XPS spectra, Cu, Ti and O photoelectron lines from SG sample were detected along with C peaks. The XPS observations were consistent with EDS results that apart from Cu, Ti and O, no other elements were detected from the samples. In Fig. 2(b), Cu 2p3/2 and Cu 2p1/2 spin orbital splitting photoelectrons were located at binding energies of 933.6 and 953.4 eV respectively, and were assigned to the compound of CuO [28]. The respective binding energies of Ti 2p3/2 and Ti 2p1/2 were found to be 458.4 and 463.9 eV. These bands were assigned to typical Ti4þ [17], which was consonant with the
Fig. 1 e XRD patterns of CueTiO2 photocatalysts.
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 5 2 5 4 e5 2 6 1
C 1s
Ti 2p
15000
Cu 2p
O 1s
20000
10000
Ti 3p
Intensity
O KLL
25000
Cu 2s
a
5000
0
0
200
400
600
800
1000
1200
Binding energy (ev)
Intensity (a.u.)
b
2000
Cu 2p3/2 1800
Cu 2p1/2
1600 1400
925
930
935
940
945
950
955
960
965
970
Binding energy (ev)
Intensity (a.u.)
c
3000
Ti 2p3/2
2000
Ti 2p1/2 1000 0 454
456
458
460
462
464
466
468
Binding energy (ev)
Fig. 2 e (a) XPS survey spectrum of SG photocatalyst; (b) Highresolution Cu 2p spectrum; (c) High-resolution Ti 2p spectrum.
XRD analysis. The XPS spectra confirmed the chemical composition of SG sample to be CuO and TiO2. From the TEM image shown in Fig. 3(a), it is apparent that the WI sample consisted of two types of particles coalesced
5257
together, with diameters of ca. 10 nm and 25 nm. With reference to the XRD patterns, particles of 10 nm were most likely CuO component, and the larger particles of 25 nm were postulated as TiO2 support. The agglomeration of small Cu component particles around TiO2 endorsed efficient charge transfer and high photocatalytic activity. A similar result was obtained for the NR sample. It is noteworthy that for both SG and PD samples, only particles assigned to the TiO2 support could be distinguished from the TEM images. According to the XRD analysis, Cu components within SG and PD samples were highly dispersed, and this may result in poor identification from the TEM images. Table 1 summarizes surface atomic composition, Cu content, and BET specific surface area of each prepared photocatalyst. All prepared CueTiO2 possessed similar Cu content and the ratios of Cu/Ti in all bulk samples were about 8 atom%. Nevertheless, from the EDS results, Cu content on the surface was relatively higher, especially for P25-based photocatalysts (WI, NR, PD). Difference of Cu/Ti ratio in the surface and bulk phases, gives a good indication of whether the Cu is deposited on the surface of TiO2 or dissolved into the crystal lattices of TiO2. During the preparation process of WI, NR and PD samples, interactions between Cu and TiO2 support were confined to the TiO2 surface, and the existing TiO2 crystal structure limited dissolution of Cu into its lattice, leading to relatively greater accumulation of Cu on the surface. Contrary to the P25-based samples, Cu in the SG sample was allowed to homogenize with TiO2 in the molecular scale during the solegel process [14]. This resulted in a relatively uniform dispersion of Cu on both the surface and bulk of the sample. From Table 1, it can be noted that all CueTiO2 photocatalysts possessed a moderate BET surface area, thus facilitating absorption of reactants on the catalyst surface. Among all the samples, SG possessed the largest surface area, mainly due to composition of 8.5 nm TiO2 crystals; while, particles in P25-based samples (WI, NR, PD) were ca. 25 nm. Moreover, P25-based photocatalysts tended to agglomerate into larger secondary particles during the synthesis processes, which led to lower surface areas. SEM images (Fig. 4) and pore size distribution results (Fig. 5) further manifested the formation of secondary particles.
Fig. 3 e TEM images of CueTiO2 photocatalysts (a: WI; b: SG).
5258
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 5 2 5 4 e5 2 6 1
Table 1 e Summary of surface element composition, Cu content and BET specific surface area of as-prepared CueTiO2 photocatalysts. Catalyst Surface element Cu/Ti Cu/Ti BET composition (surface, %) (bulk, %) surface (atom %) area (m2 g1) Cu Ti O SG WI NR PD
4.1 3.5 4.2 4.3
43.0 30.7 34.7 30.8
52.9 65.8 61.1 64.9
9.4 11.3 12.2 13.8
8.1 8.5 8.6 8.2
87.3 36.1 44.8 46.8
In the SEM image of the SG sample, dense cakes instead of particulates were observed, owing to coagulated particles of small dimensions. Under the same conditions, nano-scale particles of P25 and uniformly large agglomerates in P25-based samples (pointed in squares) were apparent in the SEM images. The pore size distribution of CueTiO2 photocatalysts in Fig. 5 supported the existence of larger secondary particles formed in all P25-based samples. It is clear that the curves of P25-based samples had at least two components with peak values of 2e3 nm and 30 nm. When compared against P25, intensity of the first component decreased in the order of PD, NR, WI; whereas, intensity of the second component exhibited the reverse sequence. Incorporated with the BET data and SEM images, it is reasonable to hypothesize that the first component corresponded to the pore volume distribution of primary nanoparticles (ca. 25 nm); the second component correlated to pores formed within the larger secondary particles [29], and intensity of the second peak corresponded to the amount of secondary particles present in the sample. In accordance to this hypothesis, the dominant composite of P25 was primary nanoparticles; whereas secondary particles were prevalent gradually in P25-based samples, in the sequence of PD, NR, and WI. This sequence was associated with the preparation processes, where the in situ PD method allowed preservation of P25 structure to a big extent, and the calcination procedure used to prepare WI sample further promoted formation of secondary particles. For the SG sample, there was only one pore size distribution peak that was centered at 10 nm, and apparently, the low position of this peak corresponded to large surface area of the catalyst. Fig. 6 shows the UVevisible absorption spectra of CueTiO2 photocatalysts. In addition to the strong absorption bands at around 350 nm which were associated with excitation of O 2p
electron to the Ti 3d level, large absorption bands between 400 and 800 nm as attributed to the presence of Cu species were distinct. These visible light absorption bands might be due to metallic Cu absorption (225e590 nm) [30], 3-D Cu1þ clusters (400e500 nm) [31], excitation of CuO electron from the valance band to the exciton level (<730 nm), and ded transition of Cu2þ (600e800 nm) [30]. Improvement of visible light absorption of CueTiO2 photocatalysts led to a reduction of its inherent band gap. For example, band gap of SG sample decreased from 3.1 eV of bare SGeTiO2 to 2.4 eV, as calculated with application of the Kubelka-Munk function. Cu species (CuO/Cu2O/metallic Cu) with smaller band gap and higher work function than bare TiO2 facilitated light harvest and charge carrier separation in CueTiO2 samples, and were highly regarded to improve photocatalytic reaction efficiencies. In addition, the differential reflectance spectra of WI, NR and PD samples showed two maxima: one was in the UV region, and the other was under the visible light range. It indicated that there were two-phase materials or two different types of electron transitions in those samples [32], which could be respectively attributed to TiO2 and Cu species. For the SG sample, only one peak in the visible light region was observed, which suggested relatively high dispersion of Cu component in the sample, and good contact between the Cu component and TiO2 [33].
3.2.
Photocatalytic hydrogen generation over CueTiO2
Fig. 7 shows time courses of hydrogen evolution over fabricated CueTiO2 photocatalysts, with bare TiO2 supports (P25, SGeTiO2) as comparison. Changes of hydrogen generation rates as calculated on basis of time courses are summarized in the inset of Fig. 7. It can be seen that the addition of Cu components significantly promoted hydrogen generation activity over bare TiO2. Hydrogen evolution rate was a mere 0.05 ml min1 (about 90 mmol h1 g1 catalyst) for pure TiO2, and this poor hydrogen production efficacy had been reported by Choi and Kang [17] and Puangpetch et al. [34] under similar reaction conditions. A much higher hydrogen generation rate was achieved in the range of 5e11 ml min1 (9e20 mmol h1 g1 catalyst) over CueTiO2 photocatalysts, as dependant on the preparation process. The great improvement rooted primarily from efficient separation of photo-induced electrons and holes in TiO2. In SG, WI and NR samples, conduction band of CuxO (CuO/Cu2O) is lower than TiO2 [22], hence excited electrons in TiO2 were inclined to transfer to CuxO, instead of staying in TiO2 to recombine with holes. In PD sample, electron migrated from the semiconductor
Fig. 4 e SEM images of CueTiO2 photocatalysts (a: SG; b: NR; c: P25).
5259
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 5 2 5 4 e5 2 6 1
0.0008
0.008
Pore volume, dV/dD (cc/Å/g)
0.0007
SG WI NR PD P25
0.0006 0.0005
0.007 0.006 0.005
0.0004
0.004
0.0003
0.003
0.0002
0.002
0.0001
0.001
0.0000
0.000 10
10 0
10 00
Pore diameter, D (Å) Fig. 5 e Pore size distribution of CueTiO2 photocatalysts.
to Cu metal owing to the higher work function that Cu possesses [8]. Formation of Schottky barrier at the metalesemiconductor interface served as an efficient electron trap in preventing electronehole recombination. Accumulation of electrons enabled the Cu component to act as water reduction sites in enhancing production of hydrogen. In addition, all fabrication methods employed in this study, equipped CueTiO2 photocatalysts with good dispersion of Cu components, moderate BET surface area, and excellent light absorption ability over the entire UVevisible range. These characteristics are advantageous in efficient photocatalytic hydrogen generation. Thus, under similar reaction conditions, hydrogen generation rates of the as-synthesized CueTiO2 photocatalysts were even higher than some noble metal (Pt/Au) loaded TiO2 [12,35]. From Fig. 7, it can be found that the chemical state of Cu and Cu content in the photocatalyst, were two major factors that would affect hydrogen generation of CueTiO2. Effect of Cu chemical state on hydrogen generation activity can be elucidated by comparing initial hydrogen generation
rates of all samples. NR with Cu2O showed the highest activity, followed by WI and SG samples with CuO, while PD with metallic Cu was the most inferior out of the four samples. The diverse performances of Cu components were due to different competencies of Cu species (CuO/Cu2O/metallic Cu) in the capture of electrons. As compared to metallic Cu, Cu oxides were more inclined to receive electrons from TiO2; while under the same conditions, the redox potential of Cuþ/Cu is higher than Cu2þ/Cu, inferring that electrons would preferentially attack Cu2O than CuO. The increase in efficiency of electronehole separation resulted in a higher hydrogen generation rate. In this study, it was found that high Cu content in photocatalyst, within the range of Cu < 8 atom% Cu/Ti, significantly enhanced hydrogen generation activity. Take PD for example, during Cu2þ reduction and concurrent in situ formation of the sample (first 30 min reaction), Cu was gradually deposited on TiO2, thereby increasing Cu content of the sample. An increase
1.0
Absorbance
0.8
(a)
0.6
(b) (c) (d)
0.4
(e) (f) 0.2
0.0 200
300
400
500
600
700
800
Wavelength (nm)
Fig. 6 e UVevisible absorption spectra of CueTiO2 photocatalysts (a: WI; b: PD; c: NR; d: SG; e: SGeTiO2; f: P25).
Fig. 7 e Time courses of hydrogen evolution over CueTiO2. Inset: Changes in hydrogen evolution rates across the 5 h reaction.
5260
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 5 2 5 4 e5 2 6 1
in the subsequent hydrogen generation rate was correspondingly recorded. During hydrogen generation over mature PD sample (after 70 min), a slight reduction in Cu content was observed, since pH of reaction solution declined gradually due to formation of CO2 and HCOOH byproducts, leading to Cu dissolution into the liquid phase (Fig. 8). As a result, efficiency of hydrogen generation was observed to abate slightly with time. Decline of hydrogen generation rate corresponding with decreased Cu content in the photocatalyst was also observed during hydrogen generation over SG, WI and NR samples. It gives a strong indication that higher Cu content in photocatalyst would yield higher hydrogen generation efficiency. According to our previous study [13], optimum Cu content in maximizing active surface sites and UV penetration to TiO2 was 10 atom% Cu/Ti for WI sample. In this study, Cu content in all four photocatalysts (SG, WI, NR, and PD) was less than 10 atom% Cu/Ti; moreover, a positive effect of Cu content on hydrogen generation activity was observed. Therefore it is predicted that the optimum Cu content for all four preparation methods should be higher than 8 atom% Cu/Ti. Optimum Cu content may be primarily determined by nature of Cu and TiO2, and less affected by the preparation methods. Further investigation is still required to clarify the relationship between optimum Cu content and fabrication methods. During hydrogen generation, Cu components underwent physicochemical and kinetic transformations. As the reaction progressed, the discrepancy in photocatalytic activities due to different chemical states of Cu became less apparent. This indicates that chemical state of Cu in the photocatalysts tended to transform and homologize during photocatalytic reactions. In addition, according to the results of EDS and ICP, there was no significant change in Cu content present in the spent photocatalysts. However, after the process of hydrogen generation, there were no XRD diffraction peaks of Cu species in all CueTiO2 photocatalysts. Disappearance of diffraction peaks implied that all Cu components were well dispersed after the photocatalytic reactions. We postulate that the kinetic transformations were redox in nature, and were related to the dissolution and deposition of Cu components in aqueous solution and TiO2 support. Cu oxides were first
Fig. 8 e Cu leaching from CueTiO2 into the aqueous phase corresponding to pH decline across the 5 h reaction. (For PD sample, Cu2D reduction occurred in the first 30 min of reaction.)
dissolved in acid solutions to form Cu2þ, and the Cu2þ ions then captured excited electrons to form metallic Cu, which was in situ photo-deposited on TiO2. This process produced high dispersion of small dimensional metallic Cu on the photocatalyst. Finally, metallic Cu on TiO2 was oxidized back to Cu oxide by photo-generated holes, which closed the Cu mechanistic cycle. Considering the initial difference of hydrogen generation rates between SG and WI samples, distribution of Cu in the photocatalyst might also affect photocatalytic activities, since Cu content and chemical states for both samples were similar. In the WI sample, more Cu was accumulated on the TiO2 surface than the SG sample. This improved hydrogen generation due to an increased probability for protons in the aqueous phase to come into contact with reduction sites. More Cu loaded on the surface of photocatalysts would assist in electron transfer and consumption. It is noted that the SG method produced sample with the highest stability. In SG synthesis, the Cu component was allowed to dissolve into the lattice of TiO2, which restricted Cu component from leaching into the liquid phase during photocatalytic reactions, hence resulting in stable and consistent hydrogen generation throughout the 5 h reaction. After 225 min of reaction, hydrogen generation rate of the SG sample was highest among all four CueTiO2 samples. Excellent stability endowed SG sample to be a potential candidate in performing intensive photocatalytic reactions. Characteristics of the TiO2 support were found to be noncritical in influencing hydrogen generation. Pure anatase phase in SG samples showed similar trends as anatase/rutile mixture in the WI sample. Difference in BET surface area of the TiO2 supports also had insignificant influence on hydrogen generation activity. Therefore, it can be concluded that an active crystal phase with a moderate BET surface area were the main requirements for TiO2 support in CueTiO2 photocatalysts.
4.
Conclusion
Efficient CueTiO2 photocatalysts for hydrogen generation were successfully fabricated by SG, WI, NR and PD methods. All products were equipped with good dispersion of Cu component and excellent light absorption ability. Hydrogen generation rates over the as-prepared CueTiO2 photocatalysts were recorded in the range of 9e20 mmol h1 g1 catalyst. The various fabrication methods led to different Cu chemical states, as well as different Cu distribution ratios in the surface and bulk phases of photocatalysts. Both factors were proven to influence photocatalytic activities. Cu2O in the NR sample, exhibited the greatest enhancement in terms of initial hydrogen generation, followed by CuO in WI and SG samples, while effect of metallic Cu in PD sample was relatively inferior among the four samples. It was also postulated that Cu components distributed on the surface of TiO2 facilitated charge transfer more efficiently than that dissolved in TiO2 lattice, thus leading to a higher hydrogen generation rate. It can be concluded that within the range of Cu < 8 atom% Cu/Ti, increasing Cu content of photocatalysts would generally induce better photocatalytic performance. Among the four photocatalysts, the SG sample demonstrated greatest potential in engineering applications. It
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 5 2 5 4 e5 2 6 1
achieved the highest rate of hydrogen generation over an extended period of time owning to excellent stability, and this stability is largely attributed to the restriction of Cu leaching by its crystal configuration.
references
[1] Fujishima A, Honda K. Electrochemical photolysis of water at a semiconductor electrode. Nature 1972;238:37e8. [2] Ni M, Leung MKH, Leung DYC, Sumathy K. A review and recent developments in photocatalytic water-splitting using TiO2 for hydrogen production. Renew Sust Energ Rev 2007;11:401e25. [3] Sayama K, Arakawa H. Effect of carbonate salt addition on the photocatalytic decomposition of liquid water over PteTiO2 catalyst. J Chem Soc, Faraday Trans 1997;93: 1647e54. [4] Galinska A, Walendziewski J. Photocatalytic water splitting over PteTiO2 in the presence of sacrificial reagents. Energy Fuels 2005;19:1143e7. [5] Park JW, Kang M. Synthesis and characterization of AgxO, and hydrogen production from methanol photodecomposition over the mixture of AgxO and TiO2. Int J Hydrog Energy 2007;32:4840e6. [6] Fu XL, Long JL, Wang XX, Leung DYC, Ding ZX, Wu L, et al. Photocatalytic reforming of biomass: a systematic study of hydrogen evolution from glucose solution. Int J Hydrog Energy 2008;33:6484e91. [7] Sakthivel S, Shankar M, Palanichamy M, Arabindoo B, Bahnemann D, Murugesan V. Enhancement of photocatalytic activity by metal deposition: characterisation and photonic efficiency of Pt, Au and Pd deposited on TiO2 catalyst. Water Res 2004;38:3001e8. [8] Linsebigler AL, Lu GQ, Yates JT. Photocatalysis on TiO2 surfaces: principles, mechanisms, and selected results. Chem Rev 1995;95:735e58. [9] Boccuzzi F, Guglielminotti E, Martra G, Cerrato G. Nitric oxide reduction by CO on Cu/TiO2 catalysts. J Catal 1994;146:449e59. [10] Larsson P-O, Andersson A. Complete oxidation of CO, ethanol, and ethyl acetate over copper oxide supported on titania and ceria modified titania. J Catal 1998;179:72e89. [11] Dhanalakshmi KB, Latha S, Anandan S, Maruthamuthu P. Dye sensitized hydrogen evolution from water. Int J Hydrog Energy 2001;26:669e74. [12] Sreethawong T, Yoshikawa S. Comparative investigation on photocatalytic hydrogen evolution over Cu-, Pd-, and Auloaded mesoporous TiO2 photocatalysts. Catal Commun 2005;6:661e8. [13] Xu SP, Sun DD. Significant improvement of photocatalytic hydrogen generation rate over TiO2 with deposited CuO. Int J Hydrog Energy 2009;34:6096e104. [14] Wu NL, Lee MS. Enhanced TiO2 photocatalysis by Cu in hydrogen production from aqueous methanol solution. Int J Hydrog Energy 2004;29:1601e5. [15] Sakata Y, Yamamoto T, Okazaki T, Imamura H, Tsuchiya S. Generation of visible light response on the photocatalyst of a copper ion containing TiO2. Chem Lett; 1998:1253e4. [16] Bandara J, Udawatta CPK, Rajapakse CSK. Highly stable CuO incorporated TiO2 catalyst for photocatalytic hydrogen production from H2O. Photochem Photobiol Sci 2005;4:857e61. [17] Choi HJ, Kang M. Hydrogen production from methanol/water decomposition in a liquid photosystem using the anatase structure of Cu loaded TiO2. Int J Hydrog Energy 2007;32:3841e8. [18] Jeon MK, Park JW, Kang M. Hydrogen production from methanol/water decomposition in a liquid photosystem
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
5261
using the anatase and rutile forms of CueTiO2. J Ind Eng Chem 2007;13:84e91. Bamwenda GR, Tsubota S, Nakamura T, Haruta M. Photoassisted hydrogen production from a watereethanol solution: a comparison of activities of AueTiO2 and PteTiO2. J Photochem Photobiol, A 1995;89:177e89. Boccuzzi F, Chiorino A, Gargano M, Ravasio N. Preparation, characterization, and activity of Cu/TiO2 catalysts. 2. Effect of the catalyst morphology on the hydrogenation of 1,3cyclooctadiene and the COeNO reaction on Cu/TiO2 catalysts. J Catal 1997;165:140e9. Boccuzzi F, Chiorino A, Martra G, Gargano M, Ravasio N, Carrozzini B. Preparation, characterization, and activity of Cu/TiO2 catalysts. 1. Influence of the preparation method on the dispersion of copper in Cu/TiO2. J Catal 1997;165:129e39. Jin ZL, Zhang XJ, Li YX, Li SB, Lu GX. 5.1% Apparent quantum efficiency for stable hydrogen generation over eosinsensitized CuO/TiO2 photocatalyst under visible light irradiation. Catal Commun 2007;8:1267e73. Foster NS, Noble RD, Koval CA. Reversible photoreductive deposition and oxidative dissolution of copper ions in titanium dioxide aqueous suspensions. Environ Sci Technol 1993;27:350e6. Herrmann JM, Disdier J, Pichat P. Photoassisted platinum deposition on TiO2 powder using various platinum complexes. J Phys Chem 1986;90:6028e34. Zhou JK, Zhang YX, Zhao XS, Ray AK. Photodegradation of benzoic acid over metal-doped TiO2. Ind Eng Chem Res 2006; 45:3503e11. Xu B, Dong L, Chen Y. Influence of CuO loading on dispersion and reduction behavior of CuO/TiO2 (anatase) system. J Chem Soc, Faraday Trans 1998;94:1905e9. Wang YY, Wan CC. Products and nucleation analysis of photoelectrochemical reduction of cupric ions in the CuSO4emethanoleTiO2 suspension system. J Photochem Photobiol, A 1994;79:203e11. Tseng IH, Wu JCS, Chou HY. Effects of solegel procedures on the photocatalysis of Cu/TiO2 in CO2 photoreduction. J Catal 2004;221:432e40. Bavykin DV, Parmon VN, Lapkin AA, Walsh FC. The effect of hydrothermal conditions on the mesoporous structure of TiO2 nanotubes. J Mater Chem 2004;14:3370e7. Praliaud H, Mikhailenko S, Chajar Z, Primet M. Surface and bulk properties of CueZSM-5 and Cu/Al2O3 solids during redox treatments. Correlation with the selective reduction of nitric oxide by hydrocarbons. Appl Catal B 1998;16:359e74. Colon G, Maicu M, Hidalgo MC, Navio JA. Cu-doped TiO2 systems with improved photocatalytic activity. Appl Catal, B 2006;67:41e51. Michalow KA, Logvinovich D, Weidenkaff A, Amberg M, Fortunato G, Heel A, et al. Synthesis, characterization and electronic structure of nitrogen-doped TiO2 nanopowder. Catal Today 2009;144:7e12. Bokhimi X, Morales A, Novaro O, Lopez T, Chimal O, Asomoza M, et al. Effect of copper precursor on the stabilization of titania phases, and the optical properties of Cu/TiO2 prepared with the solegel technique. Chem Mater 1997;9:2616e20. Puangpetch T, Sreethawong T, Yoshikawa S, Chavadej S. Hydrogen production from photocatalytic water splitting over mesoporous-assembled SrTiO3 nanocrystal-based photocatalysts. J Mol Catal A: Chem 2009;312:97e106. Sreethawong T, Puangpetch T, Chavadej S, Yoshikawa S. Quantifying influence of operational parameters on photocatalytic H2 evolution over Pt-loaded nanocrystalline mesoporous TiO2 prepared by single-step solegel process with surfactant template. J Power Sources 2007;165:861e9.
Available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/he
Fabrication and comparison of highly efficient Cu incorporated TiO2 photocatalyst for hydrogen generation from water Shiping Xu, Jiawei Ng, Xiwang Zhang, Hongwei Bai, Darren Delai Sun* School of Civil and Environmental Engineering, Nanyang Technological University, 50 Nanyang Avenue, Singapore 639798, Singapore
article info
abstract
Article history:
Efficient Cu incorporated TiO2 (CueTiO2) photocatalysts for hydrogen generation were
Received 8 December 2009
fabricated by four methods: in situ solegel, wet impregnation, chemical reduction of Cu salt,
Received in revised form
and in situ photo-deposition. All prepared samples are characterized by good dispersion of
15 February 2010
Cu components, and excellent light absorption ability. Depending on the preparation
Accepted 26 February 2010
process, hydrogen generation rates of the as-prepared CueTiO2 were recorded in the range
Available online 31 March 2010
of 9e20 mmol h1 g1 catalyst, which were even more superior to some noble metal (Pt/Au)
Keywords:
as different distribution ratio of Cu between surface and bulk phases of the photocatalyst.
CueTiO2
Both factors have been proven to influence photocatalytic hydrogen generation. In addition,
Hydrogen
the Cu content in the photocatalyst played a significant role in hydrogen generation. Among
loaded TiO2. The various fabrication methods led to different chemical states of Cu, as well
Fabrication methods
the four photocatalysts, the sample that was synthesized by in situ solegel method exhibited the highest stability. High efficiency, low cost, good stability are some of the merits that underline the promising potential of CueTiO2 in photocatalytic hydrogen generation. Crown Copyright ª 2010 Published by Elsevier Ltd on behalf of Professor T. Nejat Veziroglu. All rights reserved.
1.
Introduction
Massive global utilization of fossil fuels has caused critical and irreversible environmental problems throughout the world. Since fossil fuels are non-renewable, alternative energy sources are urgently needed. Among available fuels, hydrogen has been regarded as an excellent energy carrier, due to its high energy content, zero emission of greenhouse gas, and ease of storage and distribution. Currently, the majority of hydrogen demands are sourced from steam reforming of natural gas; however, this process does not mitigate the dependence on fossil fuels, and emits environmentally unfriendly greenhouse gases. Since the discovery of photoelectrochemical water splitting at a TiO2 electrode by Fujishima and Honda [1], there has
been much emphasis on leveraging solar energy to photo-split water molecules. However, hydrogen generation efficiency over bare TiO2 is low, largely due to the fast recombination of electron/hole pairs [2]. Noble metal modified TiO2 has been proven to be very effective in overcoming this drawback [3e7], since the lower Fermi level of noble metals facilitates electron transfer from TiO2 to the loaded noble metal [8], resulting in an efficient separation of charge carriers, thus enhancing the potential of hydrogen generation. However, noble metals are expensive and rare; hence identification of common metals with similar enhancement to photocatalytic activity increasingly draws research attention. Cu is much less costly as compared to noble metals, and CueTiO2 has been studied extensively in the treatment of gaseous pollutants [9,10]. Interestingly, some researchers found that CueTiO2
* Corresponding author. Tel.: þ65 67906273; fax: þ65 67910676. E-mail address: [email protected] (D.D. Sun). 0360-3199/$ e see front matter Crown Copyright ª 2010 Published by Elsevier Ltd on behalf of Professor T. Nejat Veziroglu. All rights reserved. doi:10.1016/j.ijhydene.2010.02.129
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 5 2 5 4 e5 2 6 1
possessed satisfactory photocatalytic hydrogen generation ability. Dhanalakshmi et al. [11] and Sreethawong and Yoshikawa [12] reported that, in comparison to Pt, Au, and Pd metals, Cu showed only slightly weaker enhancement of TiO2 activity for hydrogen generation. In our previous study [13], we deposited CuO on TiO2, with an optimum Cu content and a high dispersion of CuO. Our experimental results exhibited excellent hydrogen generation activity, even higher than some noble metal loaded TiO2. Furthermore, several Cu species (CuO/Cu2O/metallic Cu) modified TiO2 have been proven to be active photocatalysts in reduction of water under sacrificial conditions [14e18]. This further affirms the huge potential of CueTiO2 in photocatalytic hydrogen generation. Hydrogen generation ability of photocatalysts depended on the synthesis method to a large extent. AueTiO2 as fabricated by photo-deposition showed greater activity in hydrogen generation than samples prepared by depositioneprecipitation and wet impregnation methods [19]. Likewise for CueTiO2 photocatalysts, fabrication methods would also affect photocatalytic activity, as well as characteristics of the end-products. Boccuzzi et al. [20,21] compared properties and activity of CueTiO2 prepared by wet impregnation and chemisorptionhydrolysis methods, and found that samples with the same chemical composition exhibited a marked difference of up to 100 times in the hydrogenation of 1,3-cyclooctadiene. Wu and Lee [14] reported that Cu doping within the TiO2 lattice had a negative effect on photocatalytic hydrogen generation as opposed to Cu deposition. Different hydrogen generation activity of CueTiO2 prepared via varying preparation methods had also been mentioned by Kang’s group [17,18]. However, majority of the reported CueTiO2 photocatalysts employed for hydrogen generation were fabricated by the wet impregnation method [13,16,17,22]. To date, no systematic study about the effect of CueTiO2 preparation methods on hydrogen generation rates has been reported. In this paper, we compared several fabrication methods, and this may pave way for fabrication of highly productive CueTiO2 photocatalysts for generation of hydrogen. This paper highlights four different methods that are widely used in the fabrication of metaleTiO2, namely, in situ solegel (SG), wet impregnation (WI), chemical reduction of Cu salt by NaBH4 (NR) [23], and in situ photo-deposition (PD) [24]. These four methods were employed to fabricate CueTiO2 photocatalysts, and their hydrogen generation activities were subsequently compared. It was discovered that the fabrication methods determined the chemical state of Cu, distribution ratio of Cu within the photocatalyst, BET surface area of the catalyst, and crystal structure of the TiO2 support. The influence on hydrogen generation activity will also be discussed in this paper.
2.
Experimental
2.1.
Preparation of CueTiO2 photocatalysts
With reference to our previous study [13], Cu content in all samples was controlled at 8 atom% Cu/Ti, with Cu(NO3)2$3H2O (Merck, GR) as the copper precursor.
5255
In the solegel process, Ti(OBu)4 (Merck) was employed as the titanium precursor. TiO2 was synthesized in a 1:20:6:0.8 molar ratio of Ti(OBu)4:C2H5OH (Merck, GR):H2O (Milli-Q, ultrapure):HNO3 (Merck, AR) [25]. 17 ml Ti(OBu)4 was dissolved in 38 ml absolute ethanol with stirring for 10 min; 2.5 ml HNO3 was then added dropwise into the above solution and vigorously stirred for 30 min. Another solution containing 20 ml absolute alcohol, 4.3 ml H2O, and required amount of Cu(NO3)2 was slowly added into the above solution, to produce transparent sol under rigorous stirring. After 7 days of aging, the obtained gel was evaporated at 378 K, and calcined at 723 K for 4 h to obtain the final product. SGeTiO2 was produced via identical procedures, with the exception of omitting Cu salt addition. Degussa P25 (Germany) was employed as the TiO2 precursor in the following three processes. For WI sample, P25 powder was added into Cu(NO3)2 solution at a ratio of 1 g/3 ml, followed by ultrasonication and thorough mixing to obtain a homogeneous slurry. Concentration of Cu(NO3)2 solution was adjusted to control Cu content in the product. After wet impregnation, the sample was dried at 378 K overnight, and finally calcined in air for 4 h at 723 K. For NR sample, solid NaBH4 (Merck, GR) was added into a vigorously stirred, homogeneous suspension containing Cu (NO3)2 and P25 in an ice water bath. The amount of NaBH4 and Cu(NO3)2 were decided by complete reduction of a required amount of Cu on TiO2. During the reduction process, the originally white TiO2 was turned into a chocolate-brown coloured precipitate. After continuous mixing for 2 h, the colour turned blackish green. Final product was recovered and washed with large amount of ultrapure water, and dried at 378 K overnight. In situ photo-deposition was employed to prepare PD sample. P25 suspension was first mixed and sonicated to break up loosely attached aggregates. Required amounts of Cu (NO3)2 and methanol (Fisher, HPLC) were then added, and made up to the final volume in view of photocatalytic hydrogen generation, which would be described later.
2.2.
Characterization of CueTiO2 photocatalysts
X-ray diffraction (XRD) patterns were obtained using a Bruker D8 Advance X-ray diffractometer with monochromated highintensity Cu Ka radiation (l ¼ 1.5418 A). X-ray photoelectron spectroscopy (XPS) analyses were carried out in an ultrahigh vacuum chamber with a base pressure below 2.66 107 Pa at room temperature. Photoemission spectra were recorded by a Kratos Axis Ultra spectrometer equipped with a standard monochromatic Al Ka excitation source (hn ¼ 1486.71 eV). All binding energies were referenced to C 1s at 284.8 eV. Sample morphologies were observed by transmission electron microscope (TEM, Jeol JEM-2010) and scanning electron microscope (SEM, Zeiss Evo 50); surface element composition was detected using the energy dispersive X-ray spectrometer (EDS) attached to the SEM. BET surface area and pore size distribution were determined at liquid nitrogen temperature (77 K) using a Micromeritics ASAP 2010 system. Absorption spectra of CueTiO2 photocatalysts were recorded by Thermo Scientific Evolution 300 UVevisible spectrophotometer
5256
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 5 2 5 4 e5 2 6 1
equipped with an integration sphere. Inductively coupled plasma emission spectroscopy (ICP, Perkin Elmer Optima 2000DV) was employed to determine bulk Cu content in the CueTiO2 catalysts, and quantify dissolved Cu in the solution during hydrogen generation.
2.3.
Photocatalytic hydrogen generation
The photocatalytic reaction was performed in an innerirradiation type Pyrex reactor (volume: 1.35 l) with a 400 W high pressure Hg lamp (Riko, UVL-400HA) as the light source. To maintain a constant reactor temperature of 298 K, a quartz water jacket, which was cooled by recycled water, was utilized to cover the lamp. Powdered CueTiO2 was suspended in 10 volume% methanol water mixture at a concentration of 1 g/l. A magnetic stirrer was placed at the bottom of the reactor to homogeneously churn the suspension during the reaction. Prior to irradiation, the photocatalyst suspension was deaerated thoroughly for 30 min by nitrogen gas purging. Gas produced via the photocatalytic reaction was analyzed using a TCD-type gas chromatography (HP 5890A, Carboxen 1010 plot). pH value of the solution was determined by a pH meter (Horiba F-53).
3.
Results and discussion
3.1.
Characterization of CueTiO2
Fig. 1 shows the XRD patterns of the as-prepared CueTiO2 photocatalysts, with P25 as comparison. In the XRD patterns of WI and NR samples, signals of Cu components were clearly observed. In the case of WI sample, diffraction peaks at 2q ¼ 35.5 and 38.7 were ascribed to the respective (1 1) and
(1 1 1) planes of CuO (JCPDS 48-1548); while diffraction peaks at 2q ¼ 36.4 and 42.3 were respectively identified as (1 1 1) and (2 0 0) planes of Cu2O (JCPDS 05-0667) for the NR sample. However, no obvious phases of Cu components were detected in SG and PD samples. As results of EDS and ICP confirmed the existence of Cu in all photocatalysts, it is believed that the dimensions of Cu components in SG and PD samples were below detection limit of the XRD, thus evidencing high copper dispersion in the samples [26]. Taking into account the fabrication process, Cu component in PD sample might be in ground metal state, since in situ photocatalytic reduction of Cu salt favored formation of metallic Cu [27]. The chemical state of Cu in SG sample was identified by XPS analysis. With regard to TiO2, the XRD pattern of SG sample exhibited characteristic features of anatase phase (JCPDS 21-1272), with an average crystal size of 8.5 nm, as calculated by Scherrer’s equation from the full width at half maximum of anatase (1 0 1), after correction for instrumental broadening. For P25-based photocatalysts (WI, NR, PD), no significant changes of P25 diffraction patterns were observed, indicating that the TiO2 support had scarcely undergone changes in particle size and phase transformations during the synthesis processes. Quantitative XPS analysis of SG sample was performed, and the survey spectrum and high-resolution scans are shown in Fig. 2. From the XPS spectra, Cu, Ti and O photoelectron lines from SG sample were detected along with C peaks. The XPS observations were consistent with EDS results that apart from Cu, Ti and O, no other elements were detected from the samples. In Fig. 2(b), Cu 2p3/2 and Cu 2p1/2 spin orbital splitting photoelectrons were located at binding energies of 933.6 and 953.4 eV respectively, and were assigned to the compound of CuO [28]. The respective binding energies of Ti 2p3/2 and Ti 2p1/2 were found to be 458.4 and 463.9 eV. These bands were assigned to typical Ti4þ [17], which was consonant with the
Fig. 1 e XRD patterns of CueTiO2 photocatalysts.
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 5 2 5 4 e5 2 6 1
C 1s
Ti 2p
15000
Cu 2p
O 1s
20000
10000
Ti 3p
Intensity
O KLL
25000
Cu 2s
a
5000
0
0
200
400
600
800
1000
1200
Binding energy (ev)
Intensity (a.u.)
b
2000
Cu 2p3/2 1800
Cu 2p1/2
1600 1400
925
930
935
940
945
950
955
960
965
970
Binding energy (ev)
Intensity (a.u.)
c
3000
Ti 2p3/2
2000
Ti 2p1/2 1000 0 454
456
458
460
462
464
466
468
Binding energy (ev)
Fig. 2 e (a) XPS survey spectrum of SG photocatalyst; (b) Highresolution Cu 2p spectrum; (c) High-resolution Ti 2p spectrum.
XRD analysis. The XPS spectra confirmed the chemical composition of SG sample to be CuO and TiO2. From the TEM image shown in Fig. 3(a), it is apparent that the WI sample consisted of two types of particles coalesced
5257
together, with diameters of ca. 10 nm and 25 nm. With reference to the XRD patterns, particles of 10 nm were most likely CuO component, and the larger particles of 25 nm were postulated as TiO2 support. The agglomeration of small Cu component particles around TiO2 endorsed efficient charge transfer and high photocatalytic activity. A similar result was obtained for the NR sample. It is noteworthy that for both SG and PD samples, only particles assigned to the TiO2 support could be distinguished from the TEM images. According to the XRD analysis, Cu components within SG and PD samples were highly dispersed, and this may result in poor identification from the TEM images. Table 1 summarizes surface atomic composition, Cu content, and BET specific surface area of each prepared photocatalyst. All prepared CueTiO2 possessed similar Cu content and the ratios of Cu/Ti in all bulk samples were about 8 atom%. Nevertheless, from the EDS results, Cu content on the surface was relatively higher, especially for P25-based photocatalysts (WI, NR, PD). Difference of Cu/Ti ratio in the surface and bulk phases, gives a good indication of whether the Cu is deposited on the surface of TiO2 or dissolved into the crystal lattices of TiO2. During the preparation process of WI, NR and PD samples, interactions between Cu and TiO2 support were confined to the TiO2 surface, and the existing TiO2 crystal structure limited dissolution of Cu into its lattice, leading to relatively greater accumulation of Cu on the surface. Contrary to the P25-based samples, Cu in the SG sample was allowed to homogenize with TiO2 in the molecular scale during the solegel process [14]. This resulted in a relatively uniform dispersion of Cu on both the surface and bulk of the sample. From Table 1, it can be noted that all CueTiO2 photocatalysts possessed a moderate BET surface area, thus facilitating absorption of reactants on the catalyst surface. Among all the samples, SG possessed the largest surface area, mainly due to composition of 8.5 nm TiO2 crystals; while, particles in P25-based samples (WI, NR, PD) were ca. 25 nm. Moreover, P25-based photocatalysts tended to agglomerate into larger secondary particles during the synthesis processes, which led to lower surface areas. SEM images (Fig. 4) and pore size distribution results (Fig. 5) further manifested the formation of secondary particles.
Fig. 3 e TEM images of CueTiO2 photocatalysts (a: WI; b: SG).
5258
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 5 2 5 4 e5 2 6 1
Table 1 e Summary of surface element composition, Cu content and BET specific surface area of as-prepared CueTiO2 photocatalysts. Catalyst Surface element Cu/Ti Cu/Ti BET composition (surface, %) (bulk, %) surface (atom %) area (m2 g1) Cu Ti O SG WI NR PD
4.1 3.5 4.2 4.3
43.0 30.7 34.7 30.8
52.9 65.8 61.1 64.9
9.4 11.3 12.2 13.8
8.1 8.5 8.6 8.2
87.3 36.1 44.8 46.8
In the SEM image of the SG sample, dense cakes instead of particulates were observed, owing to coagulated particles of small dimensions. Under the same conditions, nano-scale particles of P25 and uniformly large agglomerates in P25-based samples (pointed in squares) were apparent in the SEM images. The pore size distribution of CueTiO2 photocatalysts in Fig. 5 supported the existence of larger secondary particles formed in all P25-based samples. It is clear that the curves of P25-based samples had at least two components with peak values of 2e3 nm and 30 nm. When compared against P25, intensity of the first component decreased in the order of PD, NR, WI; whereas, intensity of the second component exhibited the reverse sequence. Incorporated with the BET data and SEM images, it is reasonable to hypothesize that the first component corresponded to the pore volume distribution of primary nanoparticles (ca. 25 nm); the second component correlated to pores formed within the larger secondary particles [29], and intensity of the second peak corresponded to the amount of secondary particles present in the sample. In accordance to this hypothesis, the dominant composite of P25 was primary nanoparticles; whereas secondary particles were prevalent gradually in P25-based samples, in the sequence of PD, NR, and WI. This sequence was associated with the preparation processes, where the in situ PD method allowed preservation of P25 structure to a big extent, and the calcination procedure used to prepare WI sample further promoted formation of secondary particles. For the SG sample, there was only one pore size distribution peak that was centered at 10 nm, and apparently, the low position of this peak corresponded to large surface area of the catalyst. Fig. 6 shows the UVevisible absorption spectra of CueTiO2 photocatalysts. In addition to the strong absorption bands at around 350 nm which were associated with excitation of O 2p
electron to the Ti 3d level, large absorption bands between 400 and 800 nm as attributed to the presence of Cu species were distinct. These visible light absorption bands might be due to metallic Cu absorption (225e590 nm) [30], 3-D Cu1þ clusters (400e500 nm) [31], excitation of CuO electron from the valance band to the exciton level (<730 nm), and ded transition of Cu2þ (600e800 nm) [30]. Improvement of visible light absorption of CueTiO2 photocatalysts led to a reduction of its inherent band gap. For example, band gap of SG sample decreased from 3.1 eV of bare SGeTiO2 to 2.4 eV, as calculated with application of the Kubelka-Munk function. Cu species (CuO/Cu2O/metallic Cu) with smaller band gap and higher work function than bare TiO2 facilitated light harvest and charge carrier separation in CueTiO2 samples, and were highly regarded to improve photocatalytic reaction efficiencies. In addition, the differential reflectance spectra of WI, NR and PD samples showed two maxima: one was in the UV region, and the other was under the visible light range. It indicated that there were two-phase materials or two different types of electron transitions in those samples [32], which could be respectively attributed to TiO2 and Cu species. For the SG sample, only one peak in the visible light region was observed, which suggested relatively high dispersion of Cu component in the sample, and good contact between the Cu component and TiO2 [33].
3.2.
Photocatalytic hydrogen generation over CueTiO2
Fig. 7 shows time courses of hydrogen evolution over fabricated CueTiO2 photocatalysts, with bare TiO2 supports (P25, SGeTiO2) as comparison. Changes of hydrogen generation rates as calculated on basis of time courses are summarized in the inset of Fig. 7. It can be seen that the addition of Cu components significantly promoted hydrogen generation activity over bare TiO2. Hydrogen evolution rate was a mere 0.05 ml min1 (about 90 mmol h1 g1 catalyst) for pure TiO2, and this poor hydrogen production efficacy had been reported by Choi and Kang [17] and Puangpetch et al. [34] under similar reaction conditions. A much higher hydrogen generation rate was achieved in the range of 5e11 ml min1 (9e20 mmol h1 g1 catalyst) over CueTiO2 photocatalysts, as dependant on the preparation process. The great improvement rooted primarily from efficient separation of photo-induced electrons and holes in TiO2. In SG, WI and NR samples, conduction band of CuxO (CuO/Cu2O) is lower than TiO2 [22], hence excited electrons in TiO2 were inclined to transfer to CuxO, instead of staying in TiO2 to recombine with holes. In PD sample, electron migrated from the semiconductor
Fig. 4 e SEM images of CueTiO2 photocatalysts (a: SG; b: NR; c: P25).
5259
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 5 2 5 4 e5 2 6 1
0.0008
0.008
Pore volume, dV/dD (cc/Å/g)
0.0007
SG WI NR PD P25
0.0006 0.0005
0.007 0.006 0.005
0.0004
0.004
0.0003
0.003
0.0002
0.002
0.0001
0.001
0.0000
0.000 10
10 0
10 00
Pore diameter, D (Å) Fig. 5 e Pore size distribution of CueTiO2 photocatalysts.
to Cu metal owing to the higher work function that Cu possesses [8]. Formation of Schottky barrier at the metalesemiconductor interface served as an efficient electron trap in preventing electronehole recombination. Accumulation of electrons enabled the Cu component to act as water reduction sites in enhancing production of hydrogen. In addition, all fabrication methods employed in this study, equipped CueTiO2 photocatalysts with good dispersion of Cu components, moderate BET surface area, and excellent light absorption ability over the entire UVevisible range. These characteristics are advantageous in efficient photocatalytic hydrogen generation. Thus, under similar reaction conditions, hydrogen generation rates of the as-synthesized CueTiO2 photocatalysts were even higher than some noble metal (Pt/Au) loaded TiO2 [12,35]. From Fig. 7, it can be found that the chemical state of Cu and Cu content in the photocatalyst, were two major factors that would affect hydrogen generation of CueTiO2. Effect of Cu chemical state on hydrogen generation activity can be elucidated by comparing initial hydrogen generation
rates of all samples. NR with Cu2O showed the highest activity, followed by WI and SG samples with CuO, while PD with metallic Cu was the most inferior out of the four samples. The diverse performances of Cu components were due to different competencies of Cu species (CuO/Cu2O/metallic Cu) in the capture of electrons. As compared to metallic Cu, Cu oxides were more inclined to receive electrons from TiO2; while under the same conditions, the redox potential of Cuþ/Cu is higher than Cu2þ/Cu, inferring that electrons would preferentially attack Cu2O than CuO. The increase in efficiency of electronehole separation resulted in a higher hydrogen generation rate. In this study, it was found that high Cu content in photocatalyst, within the range of Cu < 8 atom% Cu/Ti, significantly enhanced hydrogen generation activity. Take PD for example, during Cu2þ reduction and concurrent in situ formation of the sample (first 30 min reaction), Cu was gradually deposited on TiO2, thereby increasing Cu content of the sample. An increase
1.0
Absorbance
0.8
(a)
0.6
(b) (c) (d)
0.4
(e) (f) 0.2
0.0 200
300
400
500
600
700
800
Wavelength (nm)
Fig. 6 e UVevisible absorption spectra of CueTiO2 photocatalysts (a: WI; b: PD; c: NR; d: SG; e: SGeTiO2; f: P25).
Fig. 7 e Time courses of hydrogen evolution over CueTiO2. Inset: Changes in hydrogen evolution rates across the 5 h reaction.
5260
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 5 2 5 4 e5 2 6 1
in the subsequent hydrogen generation rate was correspondingly recorded. During hydrogen generation over mature PD sample (after 70 min), a slight reduction in Cu content was observed, since pH of reaction solution declined gradually due to formation of CO2 and HCOOH byproducts, leading to Cu dissolution into the liquid phase (Fig. 8). As a result, efficiency of hydrogen generation was observed to abate slightly with time. Decline of hydrogen generation rate corresponding with decreased Cu content in the photocatalyst was also observed during hydrogen generation over SG, WI and NR samples. It gives a strong indication that higher Cu content in photocatalyst would yield higher hydrogen generation efficiency. According to our previous study [13], optimum Cu content in maximizing active surface sites and UV penetration to TiO2 was 10 atom% Cu/Ti for WI sample. In this study, Cu content in all four photocatalysts (SG, WI, NR, and PD) was less than 10 atom% Cu/Ti; moreover, a positive effect of Cu content on hydrogen generation activity was observed. Therefore it is predicted that the optimum Cu content for all four preparation methods should be higher than 8 atom% Cu/Ti. Optimum Cu content may be primarily determined by nature of Cu and TiO2, and less affected by the preparation methods. Further investigation is still required to clarify the relationship between optimum Cu content and fabrication methods. During hydrogen generation, Cu components underwent physicochemical and kinetic transformations. As the reaction progressed, the discrepancy in photocatalytic activities due to different chemical states of Cu became less apparent. This indicates that chemical state of Cu in the photocatalysts tended to transform and homologize during photocatalytic reactions. In addition, according to the results of EDS and ICP, there was no significant change in Cu content present in the spent photocatalysts. However, after the process of hydrogen generation, there were no XRD diffraction peaks of Cu species in all CueTiO2 photocatalysts. Disappearance of diffraction peaks implied that all Cu components were well dispersed after the photocatalytic reactions. We postulate that the kinetic transformations were redox in nature, and were related to the dissolution and deposition of Cu components in aqueous solution and TiO2 support. Cu oxides were first
Fig. 8 e Cu leaching from CueTiO2 into the aqueous phase corresponding to pH decline across the 5 h reaction. (For PD sample, Cu2D reduction occurred in the first 30 min of reaction.)
dissolved in acid solutions to form Cu2þ, and the Cu2þ ions then captured excited electrons to form metallic Cu, which was in situ photo-deposited on TiO2. This process produced high dispersion of small dimensional metallic Cu on the photocatalyst. Finally, metallic Cu on TiO2 was oxidized back to Cu oxide by photo-generated holes, which closed the Cu mechanistic cycle. Considering the initial difference of hydrogen generation rates between SG and WI samples, distribution of Cu in the photocatalyst might also affect photocatalytic activities, since Cu content and chemical states for both samples were similar. In the WI sample, more Cu was accumulated on the TiO2 surface than the SG sample. This improved hydrogen generation due to an increased probability for protons in the aqueous phase to come into contact with reduction sites. More Cu loaded on the surface of photocatalysts would assist in electron transfer and consumption. It is noted that the SG method produced sample with the highest stability. In SG synthesis, the Cu component was allowed to dissolve into the lattice of TiO2, which restricted Cu component from leaching into the liquid phase during photocatalytic reactions, hence resulting in stable and consistent hydrogen generation throughout the 5 h reaction. After 225 min of reaction, hydrogen generation rate of the SG sample was highest among all four CueTiO2 samples. Excellent stability endowed SG sample to be a potential candidate in performing intensive photocatalytic reactions. Characteristics of the TiO2 support were found to be noncritical in influencing hydrogen generation. Pure anatase phase in SG samples showed similar trends as anatase/rutile mixture in the WI sample. Difference in BET surface area of the TiO2 supports also had insignificant influence on hydrogen generation activity. Therefore, it can be concluded that an active crystal phase with a moderate BET surface area were the main requirements for TiO2 support in CueTiO2 photocatalysts.
4.
Conclusion
Efficient CueTiO2 photocatalysts for hydrogen generation were successfully fabricated by SG, WI, NR and PD methods. All products were equipped with good dispersion of Cu component and excellent light absorption ability. Hydrogen generation rates over the as-prepared CueTiO2 photocatalysts were recorded in the range of 9e20 mmol h1 g1 catalyst. The various fabrication methods led to different Cu chemical states, as well as different Cu distribution ratios in the surface and bulk phases of photocatalysts. Both factors were proven to influence photocatalytic activities. Cu2O in the NR sample, exhibited the greatest enhancement in terms of initial hydrogen generation, followed by CuO in WI and SG samples, while effect of metallic Cu in PD sample was relatively inferior among the four samples. It was also postulated that Cu components distributed on the surface of TiO2 facilitated charge transfer more efficiently than that dissolved in TiO2 lattice, thus leading to a higher hydrogen generation rate. It can be concluded that within the range of Cu < 8 atom% Cu/Ti, increasing Cu content of photocatalysts would generally induce better photocatalytic performance. Among the four photocatalysts, the SG sample demonstrated greatest potential in engineering applications. It
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 5 2 5 4 e5 2 6 1
achieved the highest rate of hydrogen generation over an extended period of time owning to excellent stability, and this stability is largely attributed to the restriction of Cu leaching by its crystal configuration.
references
[1] Fujishima A, Honda K. Electrochemical photolysis of water at a semiconductor electrode. Nature 1972;238:37e8. [2] Ni M, Leung MKH, Leung DYC, Sumathy K. A review and recent developments in photocatalytic water-splitting using TiO2 for hydrogen production. Renew Sust Energ Rev 2007;11:401e25. [3] Sayama K, Arakawa H. Effect of carbonate salt addition on the photocatalytic decomposition of liquid water over PteTiO2 catalyst. J Chem Soc, Faraday Trans 1997;93: 1647e54. [4] Galinska A, Walendziewski J. Photocatalytic water splitting over PteTiO2 in the presence of sacrificial reagents. Energy Fuels 2005;19:1143e7. [5] Park JW, Kang M. Synthesis and characterization of AgxO, and hydrogen production from methanol photodecomposition over the mixture of AgxO and TiO2. Int J Hydrog Energy 2007;32:4840e6. [6] Fu XL, Long JL, Wang XX, Leung DYC, Ding ZX, Wu L, et al. Photocatalytic reforming of biomass: a systematic study of hydrogen evolution from glucose solution. Int J Hydrog Energy 2008;33:6484e91. [7] Sakthivel S, Shankar M, Palanichamy M, Arabindoo B, Bahnemann D, Murugesan V. Enhancement of photocatalytic activity by metal deposition: characterisation and photonic efficiency of Pt, Au and Pd deposited on TiO2 catalyst. Water Res 2004;38:3001e8. [8] Linsebigler AL, Lu GQ, Yates JT. Photocatalysis on TiO2 surfaces: principles, mechanisms, and selected results. Chem Rev 1995;95:735e58. [9] Boccuzzi F, Guglielminotti E, Martra G, Cerrato G. Nitric oxide reduction by CO on Cu/TiO2 catalysts. J Catal 1994;146:449e59. [10] Larsson P-O, Andersson A. Complete oxidation of CO, ethanol, and ethyl acetate over copper oxide supported on titania and ceria modified titania. J Catal 1998;179:72e89. [11] Dhanalakshmi KB, Latha S, Anandan S, Maruthamuthu P. Dye sensitized hydrogen evolution from water. Int J Hydrog Energy 2001;26:669e74. [12] Sreethawong T, Yoshikawa S. Comparative investigation on photocatalytic hydrogen evolution over Cu-, Pd-, and Auloaded mesoporous TiO2 photocatalysts. Catal Commun 2005;6:661e8. [13] Xu SP, Sun DD. Significant improvement of photocatalytic hydrogen generation rate over TiO2 with deposited CuO. Int J Hydrog Energy 2009;34:6096e104. [14] Wu NL, Lee MS. Enhanced TiO2 photocatalysis by Cu in hydrogen production from aqueous methanol solution. Int J Hydrog Energy 2004;29:1601e5. [15] Sakata Y, Yamamoto T, Okazaki T, Imamura H, Tsuchiya S. Generation of visible light response on the photocatalyst of a copper ion containing TiO2. Chem Lett; 1998:1253e4. [16] Bandara J, Udawatta CPK, Rajapakse CSK. Highly stable CuO incorporated TiO2 catalyst for photocatalytic hydrogen production from H2O. Photochem Photobiol Sci 2005;4:857e61. [17] Choi HJ, Kang M. Hydrogen production from methanol/water decomposition in a liquid photosystem using the anatase structure of Cu loaded TiO2. Int J Hydrog Energy 2007;32:3841e8. [18] Jeon MK, Park JW, Kang M. Hydrogen production from methanol/water decomposition in a liquid photosystem
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
5261
using the anatase and rutile forms of CueTiO2. J Ind Eng Chem 2007;13:84e91. Bamwenda GR, Tsubota S, Nakamura T, Haruta M. Photoassisted hydrogen production from a watereethanol solution: a comparison of activities of AueTiO2 and PteTiO2. J Photochem Photobiol, A 1995;89:177e89. Boccuzzi F, Chiorino A, Gargano M, Ravasio N. Preparation, characterization, and activity of Cu/TiO2 catalysts. 2. Effect of the catalyst morphology on the hydrogenation of 1,3cyclooctadiene and the COeNO reaction on Cu/TiO2 catalysts. J Catal 1997;165:140e9. Boccuzzi F, Chiorino A, Martra G, Gargano M, Ravasio N, Carrozzini B. Preparation, characterization, and activity of Cu/TiO2 catalysts. 1. Influence of the preparation method on the dispersion of copper in Cu/TiO2. J Catal 1997;165:129e39. Jin ZL, Zhang XJ, Li YX, Li SB, Lu GX. 5.1% Apparent quantum efficiency for stable hydrogen generation over eosinsensitized CuO/TiO2 photocatalyst under visible light irradiation. Catal Commun 2007;8:1267e73. Foster NS, Noble RD, Koval CA. Reversible photoreductive deposition and oxidative dissolution of copper ions in titanium dioxide aqueous suspensions. Environ Sci Technol 1993;27:350e6. Herrmann JM, Disdier J, Pichat P. Photoassisted platinum deposition on TiO2 powder using various platinum complexes. J Phys Chem 1986;90:6028e34. Zhou JK, Zhang YX, Zhao XS, Ray AK. Photodegradation of benzoic acid over metal-doped TiO2. Ind Eng Chem Res 2006; 45:3503e11. Xu B, Dong L, Chen Y. Influence of CuO loading on dispersion and reduction behavior of CuO/TiO2 (anatase) system. J Chem Soc, Faraday Trans 1998;94:1905e9. Wang YY, Wan CC. Products and nucleation analysis of photoelectrochemical reduction of cupric ions in the CuSO4emethanoleTiO2 suspension system. J Photochem Photobiol, A 1994;79:203e11. Tseng IH, Wu JCS, Chou HY. Effects of solegel procedures on the photocatalysis of Cu/TiO2 in CO2 photoreduction. J Catal 2004;221:432e40. Bavykin DV, Parmon VN, Lapkin AA, Walsh FC. The effect of hydrothermal conditions on the mesoporous structure of TiO2 nanotubes. J Mater Chem 2004;14:3370e7. Praliaud H, Mikhailenko S, Chajar Z, Primet M. Surface and bulk properties of CueZSM-5 and Cu/Al2O3 solids during redox treatments. Correlation with the selective reduction of nitric oxide by hydrocarbons. Appl Catal B 1998;16:359e74. Colon G, Maicu M, Hidalgo MC, Navio JA. Cu-doped TiO2 systems with improved photocatalytic activity. Appl Catal, B 2006;67:41e51. Michalow KA, Logvinovich D, Weidenkaff A, Amberg M, Fortunato G, Heel A, et al. Synthesis, characterization and electronic structure of nitrogen-doped TiO2 nanopowder. Catal Today 2009;144:7e12. Bokhimi X, Morales A, Novaro O, Lopez T, Chimal O, Asomoza M, et al. Effect of copper precursor on the stabilization of titania phases, and the optical properties of Cu/TiO2 prepared with the solegel technique. Chem Mater 1997;9:2616e20. Puangpetch T, Sreethawong T, Yoshikawa S, Chavadej S. Hydrogen production from photocatalytic water splitting over mesoporous-assembled SrTiO3 nanocrystal-based photocatalysts. J Mol Catal A: Chem 2009;312:97e106. Sreethawong T, Puangpetch T, Chavadej S, Yoshikawa S. Quantifying influence of operational parameters on photocatalytic H2 evolution over Pt-loaded nanocrystalline mesoporous TiO2 prepared by single-step solegel process with surfactant template. J Power Sources 2007;165:861e9.