Preparation and enhanced visible-light photocatalytic activity of Pancake Rocks-like WO_3−x/C nanocomposite

Abstract

The Pancake Rocks-like WO_3−x/C nanocomposite was synthesized via the in situ solid state thermolysis of dedocylamine–intercalated H₂W₂O₇ (H₂W₂O₇/DDA), leading to the in situ production of interlayered carbon. The Pancake Rocks-like WO_3−x/C nanocomposite has a high conduction band position compared to the pure WO₃, which has an important influence on the photocatalytic process. Furthermore, the interlayered carbon acts as an electron channel that can transfer the photogenerated electrons from WO_3−x. The visible-light response of the Pancake Rocks-like WO_3−x/C nanocomposite was enhanced through the intercalation of carbon.

---

1. Introduction

Semiconductor photocatalysis has attracted considerable interest because of its potential applications in photoelectrochemical cells or photocatalytic.^1,2 In particular, nanostructured semiconductors have aroused great attention compared to the bulk materials, due to their large surface area and size dependent properties such as enhanced charge separation and migration, increased photon absorption, and surface reactions.^3,4 Transition metal oxides, such as TiO₂,⁵ WO₃ (ref. 6) and ZnO,⁷ play an important role in the semiconductor photocatalysis.

Among transition metal oxides, tungsten oxide has been the subject of considerable research over the past few decades due to its outstanding electrochromic, gasochromic, and optoelectronic properties.^8–10 Recently, the study of visible-light-driven photocatalytic or photoelectrochemical properties of WO₃ nanostructures, has attracted much attention due to their small band gap and stable physicochemical properties.^11–14 In particular, the possibility of multi-layered nanostructures with a wide variety of compositions has attracted a great deal of scientific interest.^15,16 For example, the sandwich structured graphene–TiO₂–graphene composite was demonstrated to be a promising photocatalyst, which is several times higher than that of a pure TiO₂ film for the degradation of methylene blue under the same conditions, because the sandwich structured graphene–TiO₂–graphene composite has a two-channel electron conduction path between TiO₂ and graphene and more efficient separation of photoinduced electron–hole pairs can be enhanced greatly at the graphene/TiO₂ interface.¹⁶

Herein, the Pancake Rocks-like WO_3−x/C nanocomposite was prepared from dedocylamine–intercalated H₂W₂O₇ precursor and the overall processes are described in Fig. 1. The structure of Pancake Rocks can be in favor of charge separation and electron transfer because the in situ formation of interlayered carbon can act as an electron channel and promote electron shuttling. In addition, the Pancake Rocks-like WO_3−x/C nanocomposite promotes the electronic interaction and charge equilibration between WO_3−x and interlayered carbon, which leads to a negative shift of the Fermi level and decreases the conduction band potential. The structures, morphologies and properties of the Pancake Rocks-like WO_3−x/C nanocomposite are discussed in detail.

Fig. 1 Schematic of the synthesis process of Pancake Rocks-like WO_3−x/C nanocomposite.

2. Experiment

2.1 Chemicals

All the chemicals were of analytical grade and used without further purification. Tungsten oxide (WO₃), bismuth oxide (Bi₂O₃), dodecylamine (C₁₂H₂₇N), heptane (C₇H₁₆) and sodium sulphate (Na₂SO₄) were purchased from Sinopharm Chemical Reagent Co., Ltd.

2.2 Synthesis of the WO_3−x/C composite

In a typical experiment, the Pancake Rocks-like WO_3−x/C nanocomposite was obtained as follows: 5 g of Bi₂O₃ and 5 g of WO₃ were combined and ground to a fine powder with a mortar and pestle. The powder was then annealed in air at 800 °C for 48 h. The resulting solid (Bi₂W₂O₉) was again ground to a fine powder and 8 g was added to 200 mL of 6 M HCl at room temperature and then stirred seven days to obtain H₂W₂O₇. The solid was centrifuged at 5000 rpm for ten minutes and the acidic supernatant was discarded. The samples were dried in an oven at 60 °C for 12 h. Subsequently, 1 g of H₂W₂O₇, 3.5 g of C₁₂H₂₇N and 50 mL heptane were added to a 100 mL four neck flask stirring for three days. The precipitates in the solution were filtered, washed sequentially with heptane and ethanol, and then dried at 60 °C. Finally, the abovementioned products were placed inside the tube furnace calcined at 600 °C for 2 h in N₂ atmosphere.

2.3 Analysis techniques

X-ray powder diffraction (XRD) measurements were obtained on a Bruker D8 Advance diffractometer operated at 40 kV and 40 mA over the 2θ range from 2° to 70° with Cu-Kα radiation (λ = 0.15406 nm). The particle morphologies of the products were observed by field emission scanning electron microscopy (FE-SEM) (Hitachi SU 8010). Fourier transform infrared spectroscopy (FT-IR) spectra were obtained using the FT-IR-660 spectrometer (Agilent Technologies). Pure KBr pellet was used to analyse the sample. The reflectance spectra of the catalysts were obtained over a range of 1300–4500 cm⁻¹. Raman spectra were obtained on a DXR Raman microscope (Thermo Scientific) with a 532 nm excitation laser (setting 10 s exposure time, 20 accumulations). A Cary 5000 UV-Vis spectrometer (Agilent Technologies) was used to obtain the reflectance spectra of the samples of the reaction solution over a range of 400–800 nm. Electrochemical analysis were carried out with a standard three-electrode system with a Pt sheet as the counter electrode, Hg/Hg₂Cl₂ (saturated with KCl) as a reference electrode, and ITO glass coated with the sample were used as the working electrode. The 0.5 M Na₂SO₄ solution was used as the electrolyte. A 500 W Xe arc lamp was utilized as the light source. EIS experiments were performed at the open-circuit in the frequency range of 0.01 to 10⁵ Hz, which were obtained on a PARSTAT 2273 electrochemical workstation (Ametek). The photocurrent measurements were taken on a CHI660E workstation. The Mott–Schottky experiments were conducted on a CHI660E workstation with amplitude of 50 mV and a frequency of 1000 Hz.

2.4 Photocatalytic degradation tests

The photocatalytic activity of the samples was determined by the degradation of RhB in water under visible light. The concentration of the extracted RhB solution was tested by monitoring the absorbance at the major absorption peak (553 nm). For the photocatalytic degradation measurements, the Xe-JY500 of Rhchem-3 (Niu Bite Ltd in Beijing) was selected as the visible light source. A typical catalytic experiment is described as follows: before illumination, 0.01 g composites and 20 mL of 10 mg L⁻¹ RhB aqueous solution were stirred in the dark for 40 min to ensure the establishment of the adsorption/desorption equilibrium. After illumination, the solution was removed and centrifuged 20 min periodic intervals of irradiation. The solution was analyzed using the Cary 5000 UV-Vis spectrometer to evaluate the adsorption ability. The degradation rate was defined according to the formula as follows: (C₀ − C_i)/C₀ (C_i represents the concentration of the solution obtained at a time of i minutes (i = 20, 40, 60, 80, 100, and 120)).

3. Results and discussion

Fig. 2 shows XRD patterns of the prepared samples. According to Fig. 2a, the peak at 9.25° corresponds to the (001) planes of H₂W₂O₇, with an interlayer spacing of 0.96 nm. The (001) peak of H₂W₂O₇/DDA with the strongest intensity centered at about 2.36° corresponds to an interlayer spacing of 3.74 nm (Fig. 2b). The data indicate that dedocylamine had been introduced to the interlayer of H₂W₂O₇.^17,18 Fig. 2c shows the sample can be assigned to sub-stoichiometric WO_3−x. The pattern exhibited two intense diffraction peaks, which corresponded to the [010] and [020] crystal faces of the monoclinic W₁₈O₄₉ (abbreviated as WO_3−x) structure (JCPDS card no. 71-2450).^19–21 All the other diffraction peaks of the sample were weak and broad. The narrow (010) and (020) peaks strongly indicate that the possible crystal growth direction of the WO_3−x is [010]. The XRD patterns of the samples prepared by calcined H₂W₂O₇ and H₂W₂O₇/DDA at 600 °C in N₂ and air atmosphere are shown in Fig. S1a and b,† respectively. The diffraction peaks of both samples can match well with the monoclinic WO₃ phase (PDF no. 83-0950). To study the valence of the W element, XPS was performed (Fig. S2†). For the sample prepared by the H₂W₂O₇ calcined at 600 °C in N₂ atmosphere (Fig. S2a†), the peaks at 35.7 eV and 37.9 eV corresponded to the characteristic W⁶⁺ 4f_7/2 and 4f_5/2 for WO₃.^22,23 For the Pancake Rocks-like WO_3−x/C nanocomposite, a complex energy distribution of W 4f photoelectrons was obtained, as shown in Fig. S2b.† The W 4f core-level spectrum could be fitted to three doublets, associated with three different oxidation states of W atoms. The main peaks, having a W 4f_5/2 at 37.7 eV and a W 4f_7/2 at 35.6 eV, are attributable to the W⁶⁺ oxidation state. The second doublet has binding energies at 34.6 and 36.7 eV, which belong to W 4f_5/2 and W 4f_7/2 core levels of W⁵⁺. Furthermore, the third doublet, observed at 33.7 and 335.8 eV, corresponds to the W⁴⁺ oxidation state. These three oxidation states are the typical oxidation states found in WO_3−x nanomaterials, as reported previously.^21,24,25 The W⁵⁺ oxidation state and W⁴⁺ oxidation state in WO_3−x/C nanocomposite are due to the part of W⁶⁺ reduced in pyrolysis process of DDA.

Fig. 2 XRD patterns of H₂W₂O₇ (a), H₂W₂O₇/DDA (b), and WO_3−x/C (c).

The FT-IR results (Fig. S3†) for H₂W₂O₇/DDA reveal two bands at approximately 2849 cm⁻¹ and 2918 cm⁻¹, which are antisymmetric stretching vibration of methylene and antisymmetric stretching vibration of methylene, respectively. The weak peak appearing at approximately 2108 cm⁻¹ is caused by asymmetric bending vibration and torsional vibration because of the applied force between C₁₂H₂₅–NH₃⁺ and oxygen in H₂W₂O₇,²⁶ which also confirms that dedocylamine has been introduced to the interlayer between H₂W₂O₇ layers. Peaks at around 2900–3000 cm⁻¹ were assigned to the molecular chemisorbed water and hydroxyl groups. The bank at 3431 cm⁻¹ is specified as the stretching vibration of O–H bonding.

Fig. 3 shows of the Raman spectra of the pure WO₃ and WO_3−x/C. Raman vibrations centered at 132, 270, 327, 806 cm⁻¹ characteristic of WO₃ were also detected in the sample of WO_3−x/C composite. These bands are due to the stretching mode O–W–O. The peaks at around 1330 and 1580 cm⁻¹ are the D-band and G-band peak of carbon, respectively. Compared to that of the pure WO₃, the band at 716 cm⁻¹ was broadened and shifted to 697 cm⁻¹ for the WO_3−x/C composite. This means that carbon was in a Pancake Rocks-like structure via a binding interaction rather physically adsorbed on the WO_3−x sheets.²⁷

Fig. 3 Raman spectra of pure WO₃ (a) and WO_3−x/C nanocomposite (b).

The morphology of samples is characterized by SEM, as shown in Fig. 4. Fig. 4a shows a SEM image of H₂W₂O₇. The sample surface displayed a shape like a well-regulated layer-by-layer structure throughout its cross-section. Compared with pure WO₃ (Fig. S4†), the WO_3−x/C nanocomposite consisted of large nanometer-sized scaled sheets, which are tightly stacked together, similar to natural Pancake Rocks. These thin pieces can provide more active sites and adsorb more reactive species, which could have been the reason for its enhanced photocatalytic activity.²⁸

Fig. 4 SEM images of H₂W₂O₇ (a), WO_3−x/C nanocomposite (b).

Fig. 5a shows the representative electrochemical impedance spectra, which are presented as a sum of real impedance (Z′) and imaginary impedance (Z′′) in the form of a Nyquist diagram for WO₃ and WO_3−x/C nanocomposite. In the Nyquist plot, the semicircle portion is characteristic of the high frequencies and represents the electron-transfer-limited process, whereas the linear part, which was observed at low frequency range, corresponds to the diffusion-limited electron-transfer processes. The impedance in the plot of the WO_3−x/C nanocomposite is smaller than that of bare WO₃ and significantly decreased under light irradiation. The constructive effect of the interlayered carbon in promoting electron shuttling and suppressing the recombination of charge carriers, wherein carbon acted as an electron transfer channel, was confirmed. Therefore, the WO_3−x/C nanocomposite can inhibit the recombination of charge carriers and enhance the photocatalytic efficiency. Fig. S5† shows the I–t curves for the film electrodes with several on–off cycles of intermittent visible-light irradiation. The photocurrent curves have an anodic photocurrent spike at the initial time of irradiation. The initial current is due to the plasmon-induced electron–hole pairs at the semiconductor/electrolyte interface: holes are trapped or captured by reduced species in the electrolyte, while the electrons are transported to the back contact.^29,30 After the spike current has been attained, a continuous decrease in the photocurrent with time can be observed until a constant current is reached. During decay, the holes competitively recombine with electrons from the conduction band, instead of being captured by reduced species in the electrolyte. Under visible light irradiation, the stable photocurrent value of the WO_3−x/C nanocomposite electrode is higher than that of the WO₃ electrode. This obvious enhancement of photocurrent indicates an efficient photoinduced electrons and holes separation, which in turn results in enhanced photocatalytic performance.

Fig. 5 Impedance curve (Nyquist plots) of WO₃ and WO_3−x/C nanocomposites (a), Mott–Schottky (MS) plots of WO₃ and WO_3−x/C electrodes in the dark in 0.5 M Na₂SO₄ (b).

Fig. 5b shows the Mott–Schottky plots generated from the capacitance values. Both of these samples show positive slopes in the Mott–Schottky plots, as expected for n-type semiconductors. The flat band potential was measured by the onset potential for the photocurrent. Herein, the flat band potential of electrodes was determined by the Mott–Schottky relation given as follows:

where C is the space charge layer capacitance, e is the electron charge, ε is the dielectric constant, ε₀ is the permittivity of a vacuum, N_d is the electron donor density, V_a is the applied potential, k is the Boltzmann's constant, T is the temperature of operation, and V_fb is the flat band potential. The flat band potential (V_fb) was determined by taking the x intercept of a linear fit to the Mott–Schottky plot, C⁻², as a function of applied potential (V_a). The V_fb of pure WO₃ and Pancake Rocks-like WO_3−x/C nanocomposite were found to be 0.11 V and −0.74 V vs. SCE (equivalent to 0.36 V and −0.49 V vs. NHE), respectively. The V_fb of WO_3−x/C nanocomposite shows a large negative shift compared to pure WO₃. It is generally known that the conduction band potential (E_CB) of a n-type semiconductor is very close to (0–0.2 V more negative) the V_fb and is dependent on the carrier concentration and the electron effective mass.³¹ Herein, the voltage difference between the conduction band and the flat potential was set to 0.1 V. Therefore, the negative shift in the Fermi level of WO_3−x/C causes the E_CB shift of WO_3−x/C from 0.26 V to −0.59 V vs. NHE. The standard redox potential of O₂/×O₂⁻ (0.28 V vs. NHE) is more positive than the E_CB of the WO_3−x/C nanocomposite, which indicates that the photogenerated electron could theoretically react with the adsorbed ×O₂⁻. The WO_3−x/C nanocomposite shows substantially smaller slopes compared to the pure WO₃ sample, which suggests significantly increased donor densities based on the Mott–Schottky equation given as follows:

The donor densities of WO₃ and WO_3−x/C were calculated to be 4.5 × 10¹⁹ cm⁻³ and 1.58 × 10²⁰ cm⁻³, respectively.

To further demonstrate the charge separation and electron transfer of the Pancake Rocks-like WO_3−x/C nanocomposite, the experiment on photocatalysis degradation of rhodamine B was chosen as an example. Fig. 6a shows the degradation rate of RhB irradiation for 2 h in the presence of WO₃ and the WO_3−x/C nanocomposite. The WO_3−x/C nanocomposite exhibited a 61.0% degradation rate of RhB after adsorption/desorption equilibrium and a 92.75% degradation rate in visible light irradiation for 60 min. Furthermore, the RhB was degraded almost completely after 120 min, which is apparently higher than pure WO₃. The quantum yield (Φ) was estimated to examine the photodegradation efficiency.^32–34 The quantum yield of WO_3−x/C for this photocatalytic reaction was (2.0 ± 0.21) × 10⁻² (see ESI† for detail). The results indicate that Rocks-like WO_3−x/C nanocomposite has a higher efficiency for the decomposition of RhB than the pure WO₃. The reasons for these improvements can be attributed to its unique Pancake Rocks-like structure. Lin et al.³⁵ also reported that carbon could extend the lifetime of photoinduced electron–hole pair, showing a significant photoresponse in the visible region. The interlayered carbon can act as an electron transfer channel, transferring the photogenerated electrons from WO_3−x. In addition, the negative shift in the Fermi level of the Rocks-like WO_3−x/C nanocomposite and the high migration efficiency of photoinduced electrons may suppress the charge recombination effectively.

Fig. 6 Photodegradation of WO₃ and Pancake Rocks-like WO_3−x/C nanocomposite (a), photodegradation of RhB over ​WO_3−x/C photocatalyst in the presence of different scavengers (b), energy band diagram and photocatalytic scheme of WO₃ and WO_3−x/C nanocomposite (c) and recyclability of WO_3−x/C was carried out for 4 cycles used (d).

To quantitatively investigate the photodegradation rate of WO_3−x/C under visible light, the experimental data were analyzed by using the pseudo-first-order equation as follows:

where k is the rate constant (min⁻¹). Linear relationships were obtained as shown in Fig. S6.† The rate constant of WO_3−x/C is calculated to be 0.035 ± 0.00374 min⁻¹.

Photoinduced active species (˙OH, h⁺, and ˙O₂⁻) are very important for the mineralization of organic contaminants. To reveal the photocatalytic mechanism in the photocatalytic process of RhB, the trapping experiments were performed to examine the roles of active species. The sacrificial agents, including ammonium oxalate (AO), isopropanol (IPA) and benzoquinone (BQ), were added as the scavengers for holes (h⁺), hydroxyl radicals (˙OH) and superoxide radicals (˙O₂⁻), respectively.^36,37 Fig. 6b shows the RhB degradation over the ​WO_3−x/C nanocomposite in the presence of different quenchers. The degradation efficiency of RhB had almost no effect in the presence of IPA, indicating that very few ˙OH species were involved in the degradation process. On the contrary, the degradation rate for RhB decreases significantly with the addition of AO or BQ, which implies h⁺ and ˙O₂⁻ play a major role in the degradation process. The mechanism for the enhancement of the photocatalytic activity is summarized and illustrated in Fig. 6c. A recyclability test was carried out for the WO_3−x/C nanocomposite up to four cycles, as shown in Fig. 6d. The degradation rate decreased only by 1.51% in activity after being used 4 times, which demonstrates that this catalyst is quite stable during the photocatalytic degradation of RhB.

4. Conclusions

Pancake Rocks-like WO_3−x/C nanocomposite was synthesized via the in situ solid state thermolysis of organic dedocylamine–intercalated H₂W₂O₇. Pancake Rocks-like WO_3−x/C nanocomposite can provide more active sites and adsorb more reactive species. The interlayered carbon acts as an electron transfer channel, which can promote electron shuttling and suppress charge recombination. In addition, the electronic interaction and charge equilibration between carbon and lead to the shift of the Fermi level and decrease the conduction band potential, which can be attributed to the unique Pancake Rocks-like structure of WO_3−x/C nanocomposite. This study not only demonstrated that the Pancake Rocks-like WO_3−x/C nanocomposite is a very promising candidate for the development of high performance photocatalysts, but also provides some insights into the design of new photocatalysts with high activity for environmental purification and other applications.

Acknowledgements

This study was supported by the National Natural Science Foundation of China (21401093), Program for Liaoning Excellent Talents in University (LNET LR2015036), the Opening Funds of State Key Lab of Chemical Resource Engineering.

Notes and references

1. M. G. Walter, E. L. Warren, J. R. Mckone, S. W. Boettcher, Q. Mi, E. A. Santori and N. S. Lewis, Chem. Rev., 2010, 110, 645.
2. N. Serpone and A. V. Emeline, J. Phys. Chem. Lett., 2012, 3, 673.
3. S. Kuriakose, B. Satpati and S. Mohapatra, Phys. Chem. Chem. Phys., 2015, 17, 25172.
4. J. Li, X. Liu, Q. Han, X. Yao and X. Wang, J. Mater. Chem. A, 2013, 1, 1246.
5. S. Y. Chae, M. K. Park, S. K. Lee, T. Y. Kim and W. I. Lee, Chem. Mater., 2003, 15, 3326.
6. A. A. Ashkarran, M. M. Ahadian and S. A. M. Ardakani, Nanotechnology, 2008, 19, 195709.
7. S. Kuriakose, B. Satpati and S. Mohapatra, Phys. Chem. Chem. Phys., 2015, 17, 25172.
8. J. Livage and D. Ganguli, Sol. Energy Mater. Sol. Cells, 2001, 68, 365.
9. T. Tao and J. N. Yao, Mater. Chem., 2007, 17, 4547.
10. C. S. Rout, A. Govindaraj and C. N. R. Rao, J. Mater. Chem., 2006, 16, 3936.
11. Q. Xiang, G. F. Meng, H. B. Zhao, Y. Zhang, H. Li, W. J. Ma and J. Q. Xu, J. Phys. Chem. C, 2010, 114, 2049.
12. J. Li, X. Liu, Q. Han, X. Yao and X. Wang, J. Mater. Chem. A, 2013, 1, 1246.
13. X. C. Song, E. Yang, G. Liu, Y. Zhang, Z. S. Liu, H. F. Chen and Y. Wang, J. Nanopart. Res., 2010, 12, 2813.
14. D. S. Martínez, M. D. L. Cruz and E. L. Cuéllar, Appl. Catal., A, 2011, 398, 179.
15. C. Liu, L. Wu, J. Chen, J. Y. Liang, C. S. Li, H. M. Ji and W. H. Hou, Phys. Chem. Chem. Phys., 2014, 16, 13409.
16. C. Hu, F. Chen, T. Lu, C. Lian, S. Zheng, Q. Hu and R. Zhang, Appl. Surf. Sci., 2014, 317, 648.
17. F. Kishimoto, D. Mochizuki, M. M. Maitani, E. Suzuki and Y. Wada, RSC Adv., 2015, 5, 77839.
18. B. S. Wang, Chin. J. Inorg. Chem., 2007, 23, 200.
19. Z. Chen, Q. Wang, H. Wang, L. Zhang, G. Song, L. Song, J. Hu, H. Wang, J. Liu and M. Zhu, Adv. Mater., 2013, 25, 2095.
20. W. Xu, Q. Tian, Z. Chen, M. Xia, D. K. Macharia, B. Sun, L. Tian, Y. Wang and M. Zhu, J. Mater. Chem. B, 2014, 2, 5594.
21. C. Guo, S. Yin, Q. Dong and T. Sato, RSC Adv., 2012, 2, 5041.
22. T. Zhang, Z. Zhu, H. Chen, Y. Bai, S. Xiao, X. Zheng and S. Yang, Nanoscale, 2015, 7, 2933.
23. G. Wang, Y. Ling, H. Wang, X. Yang, C. Wang, J. Z. Zhang and Y. Li, Energy Environ. Sci., 2012, 5, 6180.
24. C. Guo, Y. Shu, Y. Mei, M. Kobayashi and T. Sato, Inorg. Chem., 2012, 51, 4763.
25. G. Leftheriotis, S. Papaefthimiou, P. Yianoulis and A. Siokou, Thin Solid Films, 2001, 384, 298.
26. R. M. Silverstein, F. X. Webster, D. Kiemle and D. L. Bryce, Spectrometric identification of organic compounds, John Wiley & Sons, 2014.
27. S. Zhu, X. Liu and Z. Chen, J. Mater. Chem., 2010, 20, 9126.
28. M. Niederberger, G. Garnweitner, F. Krumeich, R. Nesper, H. Cölfen and M. Antonietti, Chem. Mater., 2004, 16, 1202.
29. Q. Xiang, J. Yu and M. Jaroniec, J. Phys. Chem. C, 2011, 115, 7355.
30. X. Quanjun, Y. Jiaguo and J. Mietek, Nanoscale, 2011, 3, 3670.
31. E. Gao, W. Wang, M. Shang and J. Xu, Phys. Chem. Chem. Phys., 2011, 13, 2887.
32. J. Xu, J. Li, F. Wu and Y. Zhang, Environ. Sci. Technol., 2013, 48, 272.
33. D. Zhou, L. Chen, C. Zhang, Y. Yu, L. Zhang and F. Wu, Water Res., 2014, 57, 87.
34. Y. Gao, Y. Wang and H. Zhang, Appl. Catal., B, 2015, 178, 29.
35. Y. G. Lin, Y. K. Hsu, Y. C. Chen, L. C. Chen, S. Y. Chen and K. H. Chen, Nanoscale, 2012, 4, 6515.
36. S. Chen, Y. Hu, S. Meng and X. Fu, Appl. Catal., B, 2014, 150, 564.
37. Y. Du, D. Tang, G. Zhang and X. Wu, Chin. J. Catal., 2015, 36, 2219.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra00984k

---