Photocatalytical water decomposition on visible light-driven solid-solution compounds K4Ce2Ta10−xNbxO30 (x=0–10)
Meng-kui Tian,Wen-feng Shangguan
DOI: https://doi.org/10.1016/j.matchemphys.2011.05.008
IF: 4.778
2012-01-01
Materials Chemistry and Physics
Abstract:Highlights ► Visible light driven solid-solution compounds K 4 Ce 2 Ta 10− x Nb x O 30 ( x = 0–10) were synthesized by solid state reaction. ► These solid-solution compounds demonstrated water decomposition activities under visible light irradiation. ► Their electronic structures were studied based on the first principle calculation. ► The proposed band structure has good agreement with their absorption properties and photocatalytic activities. Abstract A series of single phase solid-solution K 4 Ce 2 Ta 10− x Nb x O 30 ( x = 0–10) photocatalysts were synthesized by conventional high temperature solid state reaction. Their UV–vis diffuse reflectance spectra showed their absorbance edges shifted to long wavelength zone consistently with the increase of the amount of Nb for substituting Ta in these compounds, and the onsets of absorbance edges ranging from about 540 nm to 690 nm, corresponding to bandgap energy of 1.8–2.3 eV. These series of photocatalysts possess appropriate band gap (ca. 1.8–2.3 eV) and chemical level to use solar energy to decompose water into H 2 , and the photocatalytical activities under visible light ( λ > 420 nm) demonstrated that the activities decreased correspondingly with the increase of the amount of Nb in these compounds, which is regarded as the result of the differences of their band structures. Furthermore, the photocatalytical activities and the photophysical properties of these visible light-driven photocatalysts K 4 Ce 2 Ta 10− x Nb x O 30 ( x = 0–10) were bridged by the first principle calculation based on Density Functional Theory with General Gradient Approximation and Plane-wave Pseudopotential methods. Keywords K 4 Ce 2 Ta 10− x Nb x O 30 ( x = 0–10) Photocatalyst Photophysical properties First principle calculation Density Functional Theory 1 Introduction The development of hydrogen energy attracts more and more attention for its potential to solve the energy and environmental problems. Among various interesting reactions employed to develop hydrogen, photocatalytic water splitting on semiconductors under solar energy irradiation is one of the most promising ways considering the cost and environmental issues [1] . In the course of photocatalysis process, the most important work is to develop and design photocatalysts possessing appropriate band gap and chemical potential of band edges to make the best use of solar energy and having high quantum efficiency. Up to now, many oxides possessing d 0 electronic configuration such as SrTiO 3 , K 4 Nb 6 O 17 , NaTaO 3 , as well as oxides possessing d 10 electronic configuration such as Zn 2 GeO 4 , SrIn 2 O 4 , NaSbO 3 , have been investigate [2–6] . However, the most of these materials are only activated by ultraviolet light just occupying about 3–5% of solar energy due to their wide band gap. Band engineering methods such as doping with metal cations (Fe 3+ , Cr 3+ , Ni + ) and anions (N, C, S, P, B and halogen family) as well as solid solution of wide and narrow band gap materials were extensively studied and practiced to extend their absorb edges [7–11] . However, these doping levels formed in a forbidden band are generally discrete and also that the mobility of the electron and hole in the doping levels is low, so there lies low quantum efficiency, instability and limited visible light responsive defects in these photocatalysts. Based on these context, the development of series of single phase visible light-driven photocatalysts is urgent [12] . In our previous work, single phase visible light-driven photocatalysts K 4 Ce 2 Ta 10 O 30 and K 4 Ce 2 Nb 10 O 30 were presented. The both photocatalysts have an appropriate band gap energy ca. 2.3 eV and 1.8 eV respectively, and excellent chemical potential level for utilization of solar energy. The activities for evolving H 2 and O 2 from aqueous solutions under visible light irradiation ( λ > 420 nm) without any co-catalyst were presented, and were greatly enhanced by the incorporation of Pt, RuO 2 and NiO (NiO x ) as co-catalysts on the prepared oxides [13] . In this study, we synthesized a series of solid-solution compounds K 4 Ce 2 Ta 10− x Nb x O 30 ( x = 0–10), and measured their photoactivities for water decomposition. Furthermore, the relationship between their photocatalytical activities and electronic structures were revealed by the first principle calculation based on Density Functional Theory, and the mechanism on visible light-driven photocatalysis for water decomposition was discussed. 2 Experimental 2.1 Preparation and characterization of materials The solid solution compounds of K 4 Ce 2 Ta 10 O 30 and K 4 Ce 2 Nb 10 O 30 phases, denoted as K 4 Ce 2 Ta 10− x Nb x O 30 ( x = 0–10), were synthesized from high purity of K 2 CO 3 ·3H 2 O, Ce(NO 3 ) 3 ·6H 2 O, Ta 2 O 5 and Nb 2 O 5 with corresponding ratio through conventional high temperature solid-state reactions, first at 800 °C for 5 h then 1100 °C for 12 h, with a regrinding process in between. Excess K 2 CO 3 ·3H 2 O (ca. 20%) was used to compensate for the loss caused by volatilization during calcination. After the reaction, the product was washed with distilled water and dried. 1 wt% RuO 2 was loaded by impregnation of RuCl 3 ·3H 2 O and then calcinated at 500 °C for 2 h in atmosphere, as reported [13] . The resulting samples were identified by X-ray diffraction on Bruker – AXS D8 Advance with Cu Kα radiation ( λ = 1.540562 Å). UV–vis diffuse reflectance spectra were measured using TU-1901 spectrophotometer. The reflectivity spectrum was transformed to absorbance intensity through Kubelka–Munk method. SEM images were obtained on field-emission scanning electron microscopy (FE-SEN; S-4700, Hitachi). The BET surface areas of the K 4 Ce 2 Ta 10− x Nb x O 30 ( x = 0–10) powders were determined from N 2 adsorption–desorption isotherm on Quantachrome NOVA 1000-TS. 2.2 Measurement of photocatalytic activity The photocatalytic reaction was carried out in a photo-reactor, which was provided with an entry window of optical flat quartz glass of ca. 64.0 cm 2 . 0.2 g of photocatalyst powder was dispersed and suspended in 100 mL of aqueous solution in the reactor under the vertical irradiation with 300 W Xe lamp. Visible light irradiation condition was carried out through a glass cut-off filter to remove UV light. H 2 and O 2 evolution was measured by gas chromatography (QC-9101) with thermal conductivity detector (TCD) and high purity Ar as a carrier gas. 3 Results and discussion 3.1 Characterization Usually, tantalates and niobates form isostructure compounds for the similar properties of atom Ta and Nb, especially the diameter of their cations (0.68 Å for Ta 5+ and 0.69 Å for Nb 5+ ). There is no exception in current system K 4 Ce 2 M 10 O 30 (M = Ta, Nb), being with the same tetragonal tungsten bronze structured. Their solid solution compounds K 4 Ce 2 Ta 10− x Nb x O 30 , the values of x ranging from 0 to 10, also posed the same tetragonal tungsten bronze [14] , as their XRD patterns shown in Fig. 1 . Their SEM images were shown in Fig. 2 . It can be seen that there were litter difference among them. Since all the solid-solution compounds were prepared by conversional high temperature solid state reaction method and formed iso-structural compounds, consequently and actually, there are no obvious differences among them at morphology and crystal size. Moreover, it is a common feature for solid state reaction that is the high crystalline, agglomeration and large crystal size as well as low surface area. The UV–vis diffuse reflectance spectra of these solid solution compounds were shown in Fig. 3 . It is demonstrated the absorption edges for K 4 Ce 2 Ta 10− x Nb x O 30 are about 540 nm, 590 nm, 640 nm, 670 nm and 690 nm corresponding to x = 0, 2, 5, 8 and 10 respectively. This indicates that the absorption edges shift to long wavelength zone correspondingly with the increase amount of Nb for substituting Ta, and got the utmost when there comes the x = 10 phase, which is consistent with their appearance colors, from gold yellow for x = 0 phase to buff, fawn, then turn to brown for x = 10 phase. As a comparison, the UV–vis spectrum of physical mixture of powder samples of K 4 Ce 2 Ta 10 O 30 and K 4 Ce 2 Nb 10 O 30 with 1:1 ratio is also shown in Fig. 3 . The physical mixture displayed obvious two-step absorption edges belonging to both pure phases. The distinction in absorption feature further confirmed the formation of single phase solid solution compounds K 4 Ce 2 Ta 10− x Nb x O 30 ( x = 0–10). 3.2 Photocatalytic activities for water decomposition under visible light The time dependent photocatalytic activities for water decomposition on the compounds without and with loading RuO 2 under visible light irradiation ( λ > 420 nm) in Na 2 SO 3 aqueous solution were shown in Fig. 4 . On these without any loading, their activities for H 2 evolution decreased with the increase of x in K 4 Ce 2 Ta 10− x Nb x O 30 ( x = 0–10), and the activities of these solid solution compounds are in the between of K 4 Ce 2 Ta 10 O 30 and K 4 Ce 2 Nb 10 O 30 . The same trend appears on these loading with RuO 2 . The remarkable promotion effect of RuO 2 -loading on photocatalytic activity is attributed to the electron migration from the conduction band of K 4 Ce 2 Ta 10− x Nb x O 30 to the RuO 2 nanoparticles, which function as H 2 production sites on the surface of host bulk catalyst. The facilitation for the transfer of photogenerated electrons from the conduction bands to the co-catalyst RuO 2 particle anchoring on the surface in diameter of 10–30 nano-sized particles reduced the recombination ratio of photoexcited electrons and holes pairs. Consequently, more electrons effectively taking part in the reducing water into H 2 reaction [15] . On the other hand, when it comes to the activities for H 2 evolution under λ > 510 nm versus experimental band gaps on K 4 Ce 2 Ta 10− x Nb x O 30 ( x = 0–10) with loading 1 wt% RuO 2 , it can be seen that the activities increased with the decrease of their band gaps in order, which is in inverse trend compared with that under λ > 420 nm condition, as shown in Fig. 5 . That is to say, the activities get higher with the more effective photons taking part in the photocatalytical reaction, as their absorb edges shift to long wavelength zone with the increase of x in solid solution compounds K 4 Ce 2 Ta 10− x Nb x O 30 ( x = 0–10) correspondingly. Combined with the above analysis, it can be proposed that the photocatalytical activity of a photocatalyst is not only dominated by the chemical level of its conduction band, but also the amount of photons effective in the reduction reaction. In order to confirm whether the photogenerated holes on these photocatalysts possesses positive chemistry level enough to oxidize H 2 O into O 2 or not, the photocatalytical activities for O 2 evolution on these photocatalysts were investigated in the presence of the electron acceptor AgNO 3 aqueous solution. The results shown in Table 1 indicated that these series of compounds had the ability to evolve H 2 and O 2 in the presence of electron donor and acceptor respectively. The activities for H 2 evolution decreased with increasing x in K 4 Ce 2 Ta 10− x Nb x O 30 , while the solid solution compounds with x = 2, 5 and 8 have higher activity for O 2 evolution, compared to both of the single phase compounds with x = 0 and 10, and their difference in photocatalytic activities seemed to have less relationship with their crystal and surface properties, which showed little difference from SEM images as shown in Fig. 2 and BET surface area among them. 3.3 Electronic structures study of K 4 Ce 2 Ta 10− x Nb x O 30 ( x = 0–10) based on DFT Electronic structure is an important factor dominating the properties of a materials and the study of it is necessary to bridge their relationship between structure and properties [16] . First principle calculation based on Density Functional Theory (DFT) has been extensively used to study the electronic structures of a semiconductor photocatalyst [17,18] . In general, efficient photocatalytic materials contain either transition-metal cations with a d 0 electronic configuration locating in d block (e.g., Ti 4+ , Nb 5+ and Ta 5+ ) or typical metal cations with d 10 electronic configuration locating in p block (e.g., Ge 4+ , In 3+ and Sn 4+ ) as principal cation components, the empty d or sp orbitals of which form the bottom of conduction bands. While the tops of the valence bands usually consist of O 2p orbitals, which are positioned at about +3 eV or higher versus NHE and, as such, produce a band gap too wide to absorb visible light [19] . In order to investigate the electronic structures of these series of solid solution K 4 Ce 2 Ta 10− x Nb x O 30 ( x = 0–10), the first principle calculation based on DFT with Plane-wave Pseudopotential method and Generelized Gradient Approximation were carried out on K 4 Ce 2 Ta 10 O 30 , K 4 Ce 2 Nb 10 O 30 and K 4 Ce 2 Ta 5 Nb 5 O 30 , as a representative for x = 5 of these series of solid solution K 4 Ce 2 Ta 10− x Nb x O 30 ( x = 0–10). The core orbitals were replaced by the ultrasoft core potentials, and the O 2s 2 2p 4 , K 3s 2 3p 6 4s 1 , Ce 4f 1 5s 2 5p 6 5d 1 6 s 2 , Ta 5d 3 6 s 2 and Nb 4s 2 4p 6 4d 1 5s 1 electrons were treated explicitly [20] . The kinetic energy cutoff was set at 330 eV. The partial density of states for K 4 Ce 2 Ta 10 O 30 , K 4 Ce 2 Nb 10 O 30 was shown in Fig. 6 and that for K 4 Ce 2 Ta 5 Nb 5 O 30 was in Fig. 7 respectively. The obvious results obtained from Figs. 6 and 7 are the theoretical band gaps for these photocatalysts. The theoretical band gaps for K 4 Ce 2 Ta 10 O 30 , K 4 Ce 2 Nb 10 O 30 are 1.6 eV and 1.1 eV respectively. The band gap calculated by DFT was smaller than that obtained experimentally, 2.3 eV and 1.8 eV for K 4 Ce 2 Ta 10 O 30 , K 4 Ce 2 Nb 10 O 30 respectively, which is frequently pointed out as a common feature of DFT calculations for the discontinuity in the exchange-correlation potential is not taken into account within the framework of Density Functional Theory [11] . The experimental band gap for K 4 Ce 2 Ta 10− x Nb x O 30 ( x = 5) is about 1.9 eV, deducing from the absorb edge at about 640 nm, while the calculated band gap of it is about 2.2 eV. The higher calculated band gap than that of experimental is proposed as the result the crystal distortion and the electron repulsion between Ta 5d–Nb 4d resulting from the substitution of Ta by Nb. When it comes to the configuration of their band structure and the adscription for every electron orbital, it depends on the close investigation of density of states distribution and energy level for different electron orbital. As for conventional tantalates and niobates, a typical semiconductor photocatalysts locating in d block possessing d 0 electronic configuration, usually their conduction band and valence bands are mainly consisting of Ta 5d (or Nb 4d) and O 2p orbitals respectively, such as Sr 2 M 2 O 7 (M = Ta, Nb) [21] . When it comes to K 4 Ce 2 M 10 O 3 (M = Ta, Nb), based on their DOS distribution shown in Fig. 6 , it was proposed that their band structures and the ascription for every atom electrons orbitals are as following: the top of their valence bands of K 4 Ce 2 M 10 O 3 (M = Ta, Nb) are made up of the hybridization of O 2p, Ta 5d (or Nb 4d) and occupied Ce 4f orbitals, while the conduction bands are mainly consisted of the Ta 5d (Nb 4d), still little degree of O 2p. Although the unoccupied Ce 4f orbitals appear at the bottom of conduction band, their highly localized nature in the conduction band make them less effective in photoconductivity or photocatalysis process. As for K 4 Ce 2 Ta 10− x Nb x O 30 ( x = 5), the sum and partial density of states was shown in Fig. 7 . Corresponding to the energy region from −55 eV to near the valence band, there positions Nb 4d, Ta 5d 6s, Ce 6s, K 4s, O 2s, Ce 5p in energy order. The valence bands are consisting with O 2p, Nb 4d, Ta 5d. On the other hand, the conduction band region appears unoccupied Ce 4f, Ce 5p5d, Ta 5d, Nb 4p4d and a part contribution from O 2p. It is worthy to noting that Ta 5d and Nb 4d are divided into three part: nonbonding part locating far low from the valence band edge, Ta–O and Nb–O bonding part in the valence band and antibonding part in the conduction band, which is a result of the crystal field splitting in the octahedral TaO 6 and NbO 6 environments. The nonbonding properties were deduced from their high concentration DOS distribution properties. The bonding and antibonding of Ta–O and Nb–O have great effluence on their photophysical properties of K 4 Ce 2 Ta 10− x Nb x O 30 ( x = 5). When it comes to the Ce 4f adscription in the conduction band, it is still proposed that it has little contribution to the conduction band configuration for its high localized nature, so do its contribution to the photoexcited electron transfer and photocatalytical activities for water decomposition. The additional contributions from Ce 5p5d and Nb 4p to the conduction band compared with those K 4 Ce 2 Ta 10 O 30 , K 4 Ce 2 Nb 10 O 30 single phases, extended the conduction band width of K 4 Ce 2 Ta 10− x Nb x O 30 ( x = 5) in some degree, which is helpful to the electrons transfer from bulk to the surface. Therefore, there showed a less gap on photocatalytical activities between K 4 Ce 2 Ta 10− x Nb x O 30 ( x = 5) and K 4 Ce 2 Ta 10 O 30 compared with the obvious difference in their light absorbance edges shown in diffuse reflectance spectra. Based on the electronic structures study on these photocatalysts K 4 Ce 2 Ta 10− x Nb x O 30 ( x = 0–10) and their UV–vis diffuse reflectance spectra, it can be proposed a schematic band structures and the chemical levels for band edges of these series solid solution compounds K 4 Ce 2 Ta 10− x Nb x O 30 ( x = 0–10), as shown in Fig. 8 . Since these series of solid-solution compounds have the abilities to reduce and oxidize water into H 2 and O 2 , their conduction bands edges must be locating in the positions negative than that of H + /H 2 ( E 0 = 0 V), and with the increase of x value in K 4 Ce 2 Ta 10− x Nb x O 30 ( x = 0–10), the band edges becomes be close to H + /H 2 . At the same time, their valence bands edges all locate positive than the chemical potential needed to oxidize water into O 2 ( E = 1.23 V versus NHE), and because of the great amount of contribution from O 2p, there lie little difference between K 4 Ce 2 Ta 10 O 30 and K 4 Ce 2 Nb 10 O 30 . Furthermore, the top of the valence band for solid-solution compounds ( x = 2, 5, 8) should show small lower than that of K 4 Ce 2 Ta 10 O 30 and K 4 Ce 2 Nb 10 O 30 for little contribution from the occupied Ce 4f to the top of their valence bands compared with that for K 4 Ce 2 Ta 10 O 30 and K 4 Ce 2 Nb 10 O 30 . The rationality of the schematic band structure is good consistent with their photoactivities for H 2 and O 2 evolution. Since the conduction band edges becomes low and be close to H + /H 2 potential with the increase value of x in K 4 Ce 2 Ta 10− x Nb x O 30 ( x = 0–10), resulting the lower ability to reduce water into H 2 , so there showed the H 2 evolution activities decreased with the increase amount of Nb in K 4 Ce 2 Ta 10− x Nb x O 30 ( x = 0–10), as shown in Fig. 4 . On the other hand, the valence band edges for K 4 Ce 2 Ta 10− x Nb x O 30 ( x = 2, 5, 8) are located lower than that of K 4 Ce 2 Ta 10 O 30 and K 4 Ce 2 Nb 10 O 30 , resulting in higher oxidize ability of photogenerated holes, consequently, it is no wonder that K 4 Ce 2 Ta 10− x Nb x O 30 ( x = 2, 5, 8) show higher O 2 evolution activities than that of K 4 Ce 2 Ta 10 O 30 and K 4 Ce 2 Nb 10 O 30 . The reason for this phenomenon may be ascribed to their differences in the band structures. Based on the analysis in our previous work [13] , the higher photocatalytic activities of K 4 Ce 2 Ta 10 O 30 than that of K 4 Ce 2 Nb 10 O 30 is due to the higher chemical potential of Ta 5d than that of Nb 4d, which is proposed to be the main configuration of their conduction bands respectively. The higher the conduction band is, the higher the ability to reduce H + to H 2 it has. UV–vis diffuse reflectance spectra of these solid solution compounds indicate their absorption edges shift to long wavelength with increasing x in K 4 Ce 2 Ta 10− x Nb x O 30 , correspondingly resulting in lower conduction band and lower ability to reduce water into H 2 . Thus, the activity decreases with the increase of x . 4 Conclusions A series of solid solution compounds K 4 Ce 2 Ta 10− x Nb x O 30 ( x = 0–10) were prepared by high temperature solid state reaction, they are isostructures of tetragonal tungsten bronze, confirmed by their XRD patterns. UV–vis diffuse reflectance spectra showed that with the increase amount of Nb in these compounds, their absorb edges shift to long wavelength zone correspondingly and their band gap ranging from 1.8–2.3 eV. The abilities to decompose water into H 2 and O 2 in the presence of Na 2 SO 3 and AgNO 3 aqueous solution under visible light ( λ > 420 nm) indicate that these photocatalysts had appropriate chemical levels for band edges using solar energy to oxidize and reduce water. Their differences in photocatalytical activities for water decomposition have little relation with their crystalline and surface properties, but result from their differences in electronic structures and photophysical properties. Acknowledgements This work was financially supported by the National Key Basic Research and Development Program ( 2009CB220004 ); Key Project for Society Development, Guizhou Province ( SZ[2008]3008 ); Key Project of Science and Technology, Guiyang City ( [2008]014 ) and Program for Talent Induced, Guizhou University ( 017[2007] ). References [1] X.B. Chen S.H. Shen L.J. Guo Chem. Rev. 110 2010 6503 [2] M.G. Walter E.L. Warren J.R. Mckone Chem. Rev. 110 2010 6446 [3] B.D. Alexander P.J. Kulesza I. RutKowska J. Mater. Chem. 18 2008 2298 [4] J. Sato H. Kobayashi K.K. Isarashi J. Phys. Chem. B 108 14 2004 4369 [5] A. Kudo Y. Miseki Chem. Soc. Rev. 38 2009 253 [6] H. Kadowaki J. Sato H. Kobayashi J. Phys. Chem. B 109 48 2005 22995 [7] R. Niishiro H. Kato A. Kudo Phys. Chem. Chem. Phys. 7 10 2005 2241 [8] K. Maeda K. Domen Chem. Mater. 22 2010 612 [9] K. Maeda K. Teramura D.L. Lu T. Takata N. Saito Y. Inoue K. Domen Nature 440 7082 2006 295 [10] I. Tsuji H. Kato H. Kobayashi J. Am. Chem. Soc. 126 41 2004 13406 [11] J. Sato N. Saito Y. Yamada J. Am. Chem. Soc. 127 12 2005 4150 [12] H.G. Kim D.W. Hwang J.S. Lee J. Am. Chem. Soc. 126 29 2004 8912 [13] M.K. Tian W.F. Shangguan J. Yuan M.X. Chen J.W. Shi Appl. Catal. A: Gen. 309 2006 76 [14] F. Brik R. Enjalbert C. Roucaju J. Solid State Chem. 122 1996 122 [15] H. Kato K. Asakura A. Kudo J. Am. Chem. Soc. 125 2003 3082 [16] J. Hafner Acta. Mater. 48 2000 71 [17] R. Asahi T. Morikawa T. Ohwaki Science 293 2001 269 [18] Z.G. Zou H. Arakawa J. Photochem. Photobiol. A 158 2003 145 [19] A. Ishikawa T. Takata J.N. Konda J. Am. Chem. Soc. 124 2002 13547 [20] M.D. Segall P.J.D. Lindan M.J. Probert J. Phys. Condens. Matter 14 2002 2717 [21] A. Kudo M. Sekizama Chem. Commun. 2000 1371