Investigation of structure and photocatalytic activity on TiO₂ hybridized with graphene: compared to CNT case

Abstract

The structures of TiO₂ hybridized with graphene by three different methods, i.e., thermal reaction, hydrolyzing synthesis and sol–gel method, were studied and their corresponding photocatalytic activities was compared by degrading RhB under UV and visible light irradiation. It was found that the 2-D graphene endowed an excellent morphology and structure for TiO₂ hybridization using arbitrary synthesis routes, providing well-dispersed TiO₂ particles on the graphene sheets with intimate contact. Their photocatalytic RhB degradation was strongly dependent on original TiO₂ properties. However, the TiO₂ hybridized with CNT via aforementioned three methods determined that their photocatalytic activity was strongly influenced by interfacial structures.

---

1. Introduction

Several strategies have been proposed to increase the efficiency of TiO₂ photocatalyst, among them, the preparation of TiO₂-based composites is an efficient means to enhance the photocatalytic performance via retarding the charge recombination.¹ The composite of TiO₂ and carbon materials, particularly carbon nanotubes (CNTs), has attracted much attention in the past two decades.² It is proposed that the photogenerated electrons in the space-charge regions may be transferred into CNTs, and the holes remain on TiO₂, thus slowing down the recombination of electrons and holes.³ However, TiO₂ hybridized with CNTs proved a much more sophisticated system in which the hydrophobic and inert nature of the CNTs is unfavorable for these applications. For example, Yao et al. have demonstrated that a hydration/dehydration process creating CNT/TiO₂ composites enhanced the photocatalytic activity for degradation of phenol in the liquid phase in comparison to commercial P25.⁴ But the multi-walled carbon nanotube (MWCNT)/TiO₂ composite prepared by this way shows lower photocatalytic activity for the degradation of phenol compared to commercial P25. Yu and co-workers reported that a simple mechanical mixture of TiO₂(P25) and MWCNTs improved the photocatalytic activity of TiO₂ for color removal in an aqueous solution of azo-dyes under UV light.⁵ However, the researchers did not fully explain how a simple mixture of TiO₂ and CNTs would interact to accelerate the degradation of the dye.

Most recently, TiO₂ hybridized with graphene in catalytic applications were shown to be able to compete with CNTs in charge transfer and separation. Yet there have also been reports of TiO₂ hybridized with graphene system with competing arguments. Dai and co-workers reported that a TiO₂–RGO composite prepared by growing TiO₂ nanocrystals on GO, through hydrolysis followed by hydrothermal treatment, efficiently catalyzed the photocatalytic degradation of rhodamine B (RhB), and there was strong visible light activity due to possible formation of Ti–O–C bonds.⁶ Xiong et al. demonstrated that the RGO–TiO₂ composite with high visible-photocatalytic performance for degradation of RhB could be only ascribed to a photosensitization process, rather than the excitation of RGO–TiO₂.⁷ Li and co-workers demonstrated that the TiO₂(P25)–RGO composite was a highly efficient photocatalyst for the degradation of methylene blue (MB) compared to TiO₂(P25)–CNTs composite.⁸ However, a very recent study by Xu and co-workers pointed out that the TiO₂–RGO composite was in essence the same as the TiO₂–CNTs composite on enhancement of photocatalytic performance in the degradation of MB and methyl orange (MO), and gas-phase degradation of benzene.⁹ The remarkable enhancement in the visible light absorption and the photocatalysis of TiO₂ after being hybridized with CNT or graphene is true, but does not determine the photocatalytic process. Therefore, it would be an interesting and important research direction to understand what affect the photocatalytic activity of TiO₂ is by means of CNTs or graphene hybridization.

In this work, we have compared the structure dependent photocatalytic activities of graphene/TiO₂ and CNTs/TiO₂ using three different synthesis methods, and demonstrated the role contributions between interfacial structures and synthetic routes.

2. Experimental

2.1 Materials

Graphite powder (purity 99.9995%) was obtained from Alpha Aesar. Titanium(IV) isopropoxide (TIP) and titanium tetrachloride (TiCl₄) were purchased from Aldrich. Sodium nitrate was obtained from Sigma-Aldrich, potassium permanganate from Yakuri chemicals Japan, P25 from the Degussa Company. The hydrogen peroxide and ethanol were obtained from Junsei chemicals, Japan; hydrochloric acid from LT Baker, USA; Rhodamine B (RhB) and nitric acid from Daejung Chemical & Metals Co., Ltd, (Korean).

2.2 Synthesis procedure

2.2.1 Synthesis of graphene oxide (GO). GO was prepared from graphite powder according to the modified method reported by Hummers and Offeman. The as-prepared GO was reduced to graphene by using hydrazine hydrate.

2.2.2 Synthesis of graphene/TiO₂ composites. Graphene/TiO₂ composites were synthesized by three different methods, i.e., thermal reaction, hydrolyzing synthesis and sol–gel method, and the final products are named GT-T, GT-H and GT-S, respectively.
Thermal reaction. Graphene was first dispersed in EtOH by ultrasonic treatment for 1 h. Then, P25 was added into the graphene suspension. The suspension was stirred and maintained at 393 K for 6 h. The resultant product was collected by centrifugation, washed repeatedly with water, and dried in vacuum at 333 K. GT-T composite was finally obtained by heat treatment at 693 K for 2 h in Ar. CNT/TiO₂ composites was obtained by similar procedure and named as CT-T.
Hydrolyzing synthesis. Graphene was first dispersed in EtOH by ultrasonic treatment for 1 h. Then, TiCl₄ was added into the graphene suspension. The suspension was stirred and maintained at 393 K for 12 h. Meanwhile, TiCl₄ was hydrolyzed by adding 5 mL H₂O. The resultant product was collected by centrifugation, washed repeatedly with water, and dried in vacuum at 333 K. GT-H composite was finally obtained by heat treatment at 693 K for 2 h in Ar. CNT/TiO₂ composites were obtained by a similar procedure and named CT-H.
Sol–gel method. Graphene was first dispersed in EtOH by ultrasonic treatment for 1 h. Then, TIP was added into the graphene suspension, and 2 mL HNO₃ was added simultaneously. The suspension was stirred and maintained at 393 K for 12 h. The resultant xerogel was collected by centrifugation, washed repeatedly with water, and dried in vacuum at 333 K. GT-H composite was finally obtained by heat treatment at 693 K for 2 h in Ar. CNT/TiO₂ composites were obtained by a similar procedure and named CT-S.

2.3 Analysis instruments

XRD patterns were obtained with a diffractometer from Riguka, Japan, RINT 2500 V using Cu Kα radiation. Raman spectra were carried out by a Horiba–Jobin–Yvon LabRAM using a 100× objective lens with a 532 nm laser excitation. Fourier transform-infrared (FT-IR) spectra were recorded in KBr pellets with a Bruker FTIR spectrometer. Diffuse reflectance UV-visible absorption spectra were obtained using a spectrophotometer (Shimadzu UV-2401PC) equipped with a diffuse reflectance accessory, and BaSO₄ was used as the reference. Transmission electron microscopy (TEM) observations were obtained using a JEM-2200FS microscope with Cs correction.

2.4 Photocatalytic activity

In order to analyze the photocatalytic effects, the decomposition reaction of RhB in water was followed. Powdered samples of 0.05 g were dispersed in the RhB solution under ultrasonication for 3 min. Before irradiation, the mixture was magnetically stirred for 1 h in the dark to establish an adsorption–desorption equilibrium of the dye with the catalyst. For the irradiation system, the visible light (λ > 420 nm) was used at a distance of 100 mm from the solution in darkness box. The suspension was irradiated with a light source as a function of irradiation time. Samples were then withdrawn regularly from the reactor and removal of the dispersed powders through centrifugation. The clean transparent solution was analyzed by UV/vis spectroscopy (S-3100, Sainco. Co. Ltd). The concentration of RhB in the solution was determined as a function of irradiation time from the absorbance region at a UV wavelength line of 550 nm.

3. Results and discussion

3.1 Photocatalytic activity

Fig. 1a and b show the degradation efficiencies of RhB over the GT composites with 5% graphene content prepared by different approaches under UV and visible light irradiation, respectively. For comparison, the results obtained for TiO₂ alone and the CT composites with 5% CNT content prepared by three different methods are also listed in Fig. 1a and b. The controlled blank reaction in the absence of any catalyst or in the presence of Gr or CNT alone shows no degradation of RhB. P25 alone shows higher photocatalytic activity than T-H and T-S both under UV and visible light, but T-S processes highest vis-response; this phenomenon will be discussed later. However, TiO₂ hybridized with graphene composites by all three methods exhibit significantly enhanced degradation efficiencies of RhB compared to TiO₂ alone, clearly demonstrating that the presence of graphene can enhance the photocatalytic performance of TiO₂ for the degradation of RhB. The order of photocatalytic activities of GT is consistent with that of original TiO₂, suggesting that the photocatalytic activities of GT is mainly influenced by activities of initial TiO₂. Compared with the CT composites, the distinct difference between the two composites in our case suggests that TiO₂ hybridized with graphene is more suitable for photocatalytic reactions in all of the three methods, with results similar to H₂ evolution reactions.¹⁰ However, the CT composite by sol–gel method shows highest photocatalytic activity among the three CT composites under both UV and visible light irradiation.

Fig. 1 (a) and (b) the degradation efficiencies of RhB over the GT composites with 5% graphene; (c) and (d) the degradation efficiencies of RhB over the GT-T and GT-S with different content of graphene under UV and visible irradiation respectively.

We further investigate the effect of graphene content on photocatalytic activities of GT-T and GT-S under UV and visible light, respectively. As shown in Fig. 1b and c, it is obvious that GT-T with 5 wt% graphene and GT-S with 7 wt% graphene display the highest photocatalytic activity under visible light and UV, respectively. It is slightly different compared with that reported by Zhu et al., who used a similar method for preparing graphene/TiO₂.¹¹ But the photocatalytic efficiencies of RhB over GT-T with different graphene content reveals the same trends, namely, first increasing and then decreasing. In the case of GT-S, a higher content of graphene (7%) in GT-S is required for obtaining the optimum photocatalytic activity under visible light. This may indicate that the graphene has different roles in each reaction system. According to the above results, the sol–gel method seems to be an efficient approach for higher visible response TiO₂, and we further compared the photocatalytic activities of CT-S and GT-S with different nanocarbon content. As shown in Fig. S1,† the graphene provided superior photocatalytic activities of RhB degradation to CT-S under visible light, indicating that graphene is a more promising candidate for modification of TiO₂.

3.2. Characterizations of GT composites

Fig. 2 (a–e) shows the high resolution C 1s XPS spectra for the GT composites as well as GO and Gr. In the C 1s spectra, the main peaks located at 284.6, 285.6, 286.7, 287.7 and 288.4 eV observed from the C 1s deconvolution spectrum of GO correspond to the C–C, C–OH, C–O, C=O, and O–C=O (–COO–) groups, respectively.¹² Compared with GO, the intensity of the C–O peak in graphene together with GT-S and GT-H is decreased significantly, indicating the reduction of GO to graphene sheets and the presence of Gr in GT-S and GT-H. An additional component at 283.7 eV corresponding to Ti–C^13,14 and an increased peak for C–OH at 285.6 eV appear in C 1s XPS spectra of GT-S. The observation of the Ti–C signal in the deconvoluted XPS spectrum of the GT-S can be attributed to formation of a Ti–C bond between the TiO₂ coating and the disorders of the graphene and/or between the TiO₂ coating and the amorphous carbon.¹⁵ The improvement in the visible light photocatalytic activity of the GT-S corresponding to GT-S and GT-H can be related to formation of a Ti–C bond between the graphene and TiO₂ by a sol–gel method. Earlier reports combining theory and experiment indicated that the reduction of the remaining C–OH groups of GO is very difficult.^16,17 Titanium alkoxide (titanium(IV) isopropoxide in this study) can be readily grafted on the reduced graphene surfaces by chemisorption at the molecular level; sonication and stirring in succession assist dispersion of the titanium alkoxide in the graphene suspension and promotes the reaction between the titanium precursor and the –OH groups on the graphene surfaces. The GT colloids are formed via hydrolysis reaction between the adsorbed alkoxide and water molecules. The HNO₃ present in the stirred solution act as chargers, which can render the GT colloids positively charged due to supplying hydroxyl groups by hydrolysis of alkoxide. After the sol–gel process has finished, the TiO₂ particles can be closely coated on to the Gr surface, and combined with the XPS results (Fig. 2e), we deduce that the Ti–C bond is formed between Gr and TiO₂ at their interface. Generally, Ti³⁺ can also be created in TiO₂ through the doping of other elements such as C, N, or F species to form the partially reduced product of TiO_2−x.¹⁸ In the above XPS discussion, it was proven that there is a Ti–C bond in GT-S, which means that if the carbon doping occurred in GT-S, Ti³⁺ might be detected. Herein, we provide the Ti2p XPS of GT-T, GT-S and GT-H in Fig. S2,† where no obvious peaks associated with Ti³⁺ at 456.6 eV can be detected. The results indicate that the Ti–C bond is caused between Gr and TiO₂ at their interface due to surface chemisorptions at the molecular level.

Fig. 2 (a–e) The high resolution C 1s XPS spectra for the GT composites as well as GO and Gr; (f) the Raman spectra of GT.

Fig. 2f shows Raman spectra of GT, the peaks around ∼144, ∼399, ∼513, and ∼639 cm⁻¹ ascribed to anatase TiO₂.¹⁹ A slight difference of peaks width can be seen among the three GT composites. The energy bonds of the anatase phase are perturbed by the hydrolyzing synthesis and sol–gel methods, which suggests some substitutions of the Ti⁴⁺ by carbon atoms forming (Ti–O–C or Ti–C) bonds in the TiO₂ lattice. The insert displays the intensity ratio of D band to G band of graphite, which is gradually enhanced from graphite to graphene, originating from defects associated with vacancies, grain boundaries, and amorphous carbon.²⁰ The above results are further confirmed by XRD and FT-IR studies (Fig. S3, ESI†). In the case of the GT composites, the major peaks for GT-H and GT-S presented in all the diffractograms correspond to the single anatase form of TiO₂. The most widely studied Degussa P25, which at 80% anatase and 20% rutile is also detected in GT-T, produces novel electronic states at anatase–rutile junctions that result in enhanced charge carrier separation, reduced electron–hole recombination, and facilitated charge transfer to adsorbed species.^21,22 This is an essential reason why P25 alone and GT-T have higher photocatalytic activity.

More recently, previous studies have demonstrated the enhanced photo-oxidative degradation of organic dyes for both CNT-based and graphene-based TiO₂ nanocomposites.^23–25 In particular, since these photo-oxidative reactions occur primarily on the unmodified TiO₂ surface, few discernible differences have been reported between nanocomposites based on different carbon polymorphs.⁹ The TEM images are further complicated by the fact that the higher nanocarbons exhibited multiple isomeric forms with shapes ranging from a wall tube for CNT to planar sheets for graphene as shown in Fig. 3. The TEM images of the GT composites show that the composites are composed of graphene sheets and TiO₂ nanoparticles with different size range. The size of these TiO₂ nanoparticles in GT-T, GT-H and GT-S are ∼20 nm, ∼5 nm and ∼7 nm, respectively, and the results are confirmed by their HR-TEM. However, the contact between TiO₂ and graphene is different for the composites prepared by different methods. In the GT-S, TiO₂ nanoparticles are more highly dispersed on the graphene sheets, whereas the segregation of TiO₂ nanoparticles from graphene sheets or the aggregation of TiO₂ nanoparticles is more pronounced in GT-T or GT-H. However, these different morphologies for GT composites are not consistent with their photocatalytic activity. The higher photocatalytic activity of GT-S is attributed to the extended optical absorption, resulting from surface impurity doping (Ti–C bonds). Moreover, the similar morphologies are also found in CT composites. Due to the dense dispersion of TiO₂ on the CNT surface, the CT-S shows higher photocatalytic activity under both UV and visible light irradiation, which is attributed to the increased lifetimes of the TiO₂ confined holes, due to the injection of photoexcited electrons into the CNT.²⁶ These TEM measurements further reveal an easy controlling photocatalytic activity of TiO₂ hybridized with CNT, rather, the TiO₂ hybridized with graphene may make a tunable photocatalytic activity.

Fig. 3 The TEM images of the CT-T (a) GT-T (b), CT-H (d), GT-H (e), CT-S (g) and GT-S (h) composites; HR-TEM images of GT-T (c), GT-H (f) and GT-S (i).

Diffuse reflectance UV-vis spectroscopy was applied to further study the interaction between TiO₂ and graphene in the composites prepared by three different methods, which are shown in Fig. 4. All GT composites show a red shift of the absorption edge compared to pure TiO₂, suggesting that the carbon may exist in TiO₂ matrix. However, the considerable red shift for GT-S relates to the formation of Ti–C bonds in the presence of the C–Ti interfacial structure by attaching titanium alkoxide onto functionalized graphene sheets by chemisorption. Therefore, the photocatalytic activity of TiO₂ hybridized with graphene can be tuned by various synthetic routes. According to the adsorption edge, T-S has a markedly higher visible absorption tailing off by more than 420 nm due to carbon dopant, which has been confirmed by Choi’s work.²⁷

Fig. 4 Diffuse reflectance UV-vis spectroscopy of different samples.

In summary, the photocatalytic activity of TiO₂ hybridized with CNT strongly depends on the structure, a high degree dispersion of TiO₂ on the CNT surface by a sol–gel route processes the highest photocatalytic activity for degradation of RhB under both UV and visible light irradiation compared to using thermal reaction and hydrolyzing synthesis routes, although pure P25 alone shows higher photocatalytic activity. However, there is no evidence that the photocatalytic activity of TiO₂ hybridized with graphene is dependent on their structure. As shown in the TEM and HR-TEM images, the GT composites prepared by three different methods also show greater distribution features, supporting the creation of an interphase interaction between TiO₂ and graphene. Fig. 5a displays the photographs of the GT suspensions obtained by three different methods. The suspensions of GT are similar homogeneous dispersions, suggesting a strong interaction between the two components in GT composites. Moreover, the gradual darkening color from CT-T to CT-S can be explained by the increasing amount of impurities as the amorphous carbon is introduced in CT-S, which is related to use of different Ti precursor. The absorption intensity changes in the UV and visible regions over CT composites prepared by three different methods indicate that the chemical causing the reduction is graphene in accordance with the UV-vis spectrum of plain graphene, shown in Fig. 5b.

Fig. 5 (a) The photographs of the GT suspensions with GT-T(1), GT-H(2) and GT-S(3); (b) UV-vis spectrum of different samples.

Therefore, we deduce that the defect of graphene is a major driving force for an ideal distribution state of TiO₂ (as shown in the Raman data), the defect of graphene with dangling bonds on the basal plane expand their application in different method. As known in the graphene/TiO₂ field, the synthetic approaches include photoreduction method,²⁸ hydrothermal method,²⁹ sol–gel method,³⁰ hydrazine reduction method,¹¹ ionic liquid method,³¹ molecule graft method,³² water-phase method,³³ sonochemical method,³⁴ electrochemical deposition method³⁵ and liquid phase deposition method.³⁶ Accelerating the distribution of TiO₂ on the graphene surface is mainly ascribed to physisorption, electrostatic binding, or through charge transfer interactions. However, some special properties or morphologies can be observed using different synthetic methods, such as the liquid phase deposition method can form morphology of TiO₂ nanoparticles distributed on both sides of the graphene sheets.³⁷ A self-assembly method can construct p/n heterojunction between TiO₂ and graphene interface.³⁸ In this work, we not only provide an example of graphene/TiO₂ composites using three different methods and demonstrate that graphene is a very promising candidate for regulation of photocatalytic activity of TiO₂, but also opens new possibilities to provide some insight into the design of graphene-based photocatalysts with high photocatalytic activity in further applications.

4. Conclusion

We have prepared graphene/TiO₂ composites by three different methods, i.e., thermal reaction, hydrolyzing synthesis and sol–gel methods. These composites have shown better photocatalytic performances for degradation of RhB than each original TiO₂ alone. The photocatalytic activity of these composites is strictly arranged in photocatalytic activity order of original TiO₂ under both UV and visible light irradiation, suggesting that the photocatalytic activity of graphene/TiO₂ composites can be tuned by modifying original TiO₂. However, the CNT/TiO₂ composites have shown best photocatalytic performance for degradation of RhB by sol–gel route under both UV and visible light irradiation, suggesting that the photocatalytic activity of graphene/TiO₂ composites can be controlled by the synthetic method. By investigation of structures of graphene/TiO₂ and CNT/TiO₂ composites, the structures of CNT/TiO₂ composites may cause a difference in their photocatalytic activity due to interphase interaction between TiO₂ and CNT, however, the intimate contact between TiO₂ and graphene can be observed by three different methods, which weakens the influence of the transfer of photogenerated electrons.

Acknowledgements

This work was supported by the National Science Foundation of China (NSFC) (Grant no. 21271010) and Shanghai Municipal Education Commission (No: 15ZZ088) and Science and Technology Commission of Shanghai Municipality (No: 14DZ2261000).

References

1. K. Rajeshwar, N. R. de Tacconi and C. R. Chenthamarakshan, Chem. Mater., 2001, 13, 2765–2782.
2. K. Woan, G. Pyrgiotakis and W. Sigmund, Adv. Mater., 2009, 21, 2233–2239.
3. M. R. Hoffman, S. T. Martin, W. Choi and W. Bahnemann, Chem. Rev., 1995, 95, 69–96.
4. Y. Yao, G. Li, S. Ciston, R. M. Lueptow and K. A. Gray, Environ. Sci. Technol., 2008, 42, 4952–4957.
5. Y. Yu, J. C. Yu, C. Y. Chan, Y. K. Che, J. C. Zhao, L. Ding, W. K. Ge and P. K. Wong, Appl. Catal., B, 2005, 61, 1–11.
6. Y. Liang, H. Wang, H. S. Casalongue, Z. Chen and H. Dai, Nano Res., 2010, 3, 701–705.
7. J. T. Zhang, Z. G. Xiong and X. S. Zhao, J. Mater. Chem., 2011, 21, 3634–3640.
8. H. Zhang, X. Lv, Y. Li, Y. Wang and J. Li, ACS Nano, 2010, 4, 380–386.
9. Y. Zhang, Z. R. Tang, X. Fu and Y. J. Xu, ACS Nano, 2010, 4, 7303–7314.
10. W. Q. Fan, Q. H. Lai, Q. H. Zhang and Y. Wang, J. Phys. Chem. C, 2011, 115, 10694–10701.
11. Y. J. Wang, R. Shi, J. Lin and Y. F. Zhu, Appl. Catal., B, 2010, 100, 179–183.
12. Y. Zhou, Q. Bao, L. A. L. Tang, Y. Zhong and K. P. Loh, Chem. Mater., 2009, 21, 2950–2956.
13. Y. Huang, W. Ho, S. Lee, L. Zhang, G. Li and J. C. Jimmy, Langmuir, 2008, 24, 3510–3516.
14. H. Sun, Y. Bai, Y. Cheng, W. Jin and N. Xu, Ind. Eng. Chem. Res., 2006, 45, 4971–4976.
15. O. Akhavan, M. Abdolahad, Y. Abdi and S. Mohajerzadeh, Carbon, 2009, 47, 3280–3287.
16. D. Boukhvalov and M. Katsnelson, Phys. Rev. B: Condens. Matter Mater. Phys., 2008, 78(8), 1–5.
17. D. X. Yang, A. Velamakanni, G. Bozoklu, S. J. Park, M. Stoller, R. D. Piner, S. Stankovich, I. Jung, D. A. Field, C. A. Ventrice Jr and R. S. Ruoff, Carbon, 2009, 47, 145–152.
18. J. Zhang, Z. Xiong and X. S. Zhao, J. Mater. Chem., 2011, 21, 3634–3640.
19. C. Chen, W. M. Cai, M. C. Long, B. X. Zhou, Y. H. Wu, D. Y. Wu and Y. J. Feng, ACS Nano, 2010, 4, 6425.
20. Y. Zhou, Q. Bao, L. A. L. Tang, Y. Zhong and K. P. Loh, Chem. Mater., 2009, 21, 2950–2956.
21. A. L. Linsebigler, G. Lu and J. T. Yates, Chem. Rev., 1995, 95, 735–758.
22. D. C. Hurum, A. G. Agrios, K. A. Gray, T. Rajh and M. C. Thurnauer, J. Phys. Chem. B, 2003, 107, 4545–4549.
23. Y. Yao, G. Li, S. Ciston, R. M. Lueptow and K. A. Gray, Environ. Sci. Technol., 2008, 42, 4952–4957.
24. K. Woan, G. Pyrgiotakis and W. Sigmund, Adv. Mater., 2009, 21, 2233–2239.
25. Y. H. Ng, I. V. Lightcap, K. Goodwin, M. Matsumura and P. V. Kamat, J. Phys. Chem. Lett., 2010, 1, 2222–2227.
26. K. Zhang, F. J. Zhang, M. L. Chen and W. C. Oh, Ultrason. Sonochem., 2011, 18, 765–772.
27. Y. S. Park, W. Y. Kim, H. W. Park, T. Tachikawa, T. Majima and W. Y. Choi, Appl. Catal., B, 2009, 91, 355–361.
28. G. Williams, B. Seger and P. V. Kamat, ACS Nano, 2008, 2, 1487–1491.
29. F. Wang and K. Zhang, J. Mol. Catal. A: Chem., 2011, 345, 101–107.
30. X. Y. Zhang, H. P. Li, X. L. Cui and Y. Lin, J. Mater. Chem., 2010, 20, 2801–2806.
31. J. F. Shen, M. Shi, B. Yan, H. W. Ma, N. Li and M. X. Ye, Nano Res., 2011, 4, 795–806.
32. Y. B. Tang, C. S. Lee, J. Xu, Z. T. Liu, Z. H. Chen, Z. B. He, Y. L. Cao, G. D. Yuan, H. S. Song, L. M. Chen, L. B. Luo, H. M. Cheng, W. J. Zhang, I. Bello and S. T. Lee, ACS Nano, 2010, 4, 3482–3488.
33. C. Z. Zhu, S. J. Guo, P. Wang, L. Xing, Y. X. Fang, Y. M. Zhai and S. J. Dong, Chem. Commun., 2010, 46, 7148–7150.
34. J. J. Guo, S. M. Zhu, Z. X. Chen, Y. Li, Z. Y. Yu, Q. L. Liu, J. B. Li, C. L. Feng and D. Zhang, Ultrason. Sonochem., 2011, 18, 1082–1090.
35. N. L. Yang, J. Zhai, D. Wang, Y. S. Chen and L. Jiang, ACS Nano, 2010, 4, 887–894.
36. H. J. Zhang, P. P. Xu, G. D. Du, Z. W. Chen, K. Oh, D. Y. Pan and Z. Jiao, Nano Res., 2011, 4, 274–283.
37. G. D. Jiang, Z. F. Lin, C. Chen, L. H. Zhu, Q. Chang, N. Wang, W. Wei and H. Q. Tang, Carbon, 2011, 49, 2693–2701.
38. C. Chen, W. M. Cai, M. C. Long, B. X. Zhou, Y. H. Wu, D. Y. Wu and Y. J. Feng, ACS Nano, 2010, 4, 6425–6432.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra09773h

---