Controlled synthesis of highly active Au/CeO₂ nanotubes for CO oxidation

Abstract

Uniform CeO₂ nanotubes with smooth thin walls and high porosity were controllably synthesized using a simple well-controlled solvothermal technique. The growth of CeO₂ nanotubes was explored and it was found that it followed the oriented attachment-Ostwald ripening mechanism. Furthermore, through an auto-redox process, gold nanoparticles of ∼5 nm size could be homogeneously generated on these CeO₂ nanotubes. These novel nanocomposites exhibited outstanding performance in terms of both activity and stability for catalytic CO oxidation.

---

1. Introduction

Since Haruta et al.¹ reported the unexpected catalytic activity of nanoparticles of well-known chemically inert Au for CO oxidation, numerous studies have been focused on the experimental² and theoretical³ investigation of noble-metal nanoparticles/metal-oxide catalysts in this field. This can be attributed to two aspects. On the one hand, catalytic CO oxidation is an ideal model reaction for understanding the mechanism of heterogeneous catalysis. On the other hand, it will facilitate the development of new catalysts for many important industrial applications such as automotive-exhaust emission control, electrolyte fuel cells, chemical processing, CO₂ lasers, and sensors.⁴

CeO₂ has been recognized as a promising candidate for advanced heterogeneous catalysis as well as a prevalent supporting material to immobilize noble-metal nanoparticles (NPs), especially for CO catalytic oxidation, owing to their unique physicochemical properties.^5,6 First, it has the ability to easily shift between Ce⁴⁺ and Ce³⁺ oxidation states depending on whether it is present in an oxidizing or reducing atmosphere. Second, it contains numerous oxygen vacancies within its structure, leading to high oxygen mobility.⁷ Furthermore, experimental studies have shown that the surface of CeO₂ with rich oxygen vacancies, in addition to the surface step edges, may strongly bind metal NPs or nanoclusters (NCs).⁸ The oxygen vacancy formation on the CeO₂ surface accompanies the reduction of adjacent Ce⁴⁺ ions to Ce³⁺, and thus, the localized electrons on the occupied 4f-orbital of Ce³⁺ ions contribute to the electronic interaction between the reduced CeO₂ and the supported metal NPs.^9,10 Obviously, CeO₂-supported noble metal catalysts are of considerable interest from both a practical and a theoretical perspective.

Nanocrystals with exposed high-energy facets are known to possess exceptional physiochemical properties because of their unique geometrical and electronic structures, including high densities of atom steps, kinks, dangling bonds, and ledges.¹¹ In the case of CeO₂, surface oxygen vacancy is one of the most important active species for catalysis. Moreover, the formation and migration process of these surface oxygen vacancies are mainly influenced by the morphologies and facets of the CeO₂-based nanomaterials.^7,8,12,13 Thus, in recent years, intensive research attention has been paid to the design and synthesis of CeO₂ nanocrystals of different morphologies. Various well-defined CeO₂ nanostructures including nanoparticles, nanowires,^14,15 nanorods,¹⁶ nanosheets, nanocubes, nanopolyhedrons,¹⁷ nanoflowers,^12,18 nanospheres,^7,11,19 hollow spheres²⁰ or multi-shell hollow spheres,^13,21,22 and nanotubes^23,24 have been widely explored. Among these, the CeO₂ nanotube is of special interest due to its potential high surface area, abundant oxygen vacancy defects, specific exposed facets, and encapsulation of active metal species to prevent NPs from migration and aggregation.^17,18 Note that to date, only a few studies have been reported on the construction of CeO₂ nanotubes.^23–25 Generally, the classical hard-template methods used to prepare nanotubes are both costly and technically sophisticated; moreover, the technique based on the liquid precursor aging process requires 45 days and is thus quite time-consuming.²⁵ Obviously, the synthesis of CeO₂ nanotubes still needs to be largely improved. Moreover, to the best of our knowledge, existing methods invariably generate products with a large amount of mixed nanorods and NPs, and the cavity in each nanotube is always quite small. These disadvantages greatly impair the final performance of the samples. Therefore, achieving a method to appropriately prepare CeO₂ nanotubes with well-controlled morphology still remains a big challenge.

Herein, we introduced a novel, simple, and efficient solvothermal technique to prepare the uniform CeO₂ nanotubes with narrow size distribution, regular shape, high porosity, and smooth thin walls. Moreover, an oriented attachment-Ostwald ripening mechanism was proposed to reveal the formation of nanotubes, which may bring out a new point of view accounting for other nano/micro-materials with tubular morphology. This is of great significance for exploring the mechanism of materials with a hollow structure.

Subsequently, a well-controlled auto-redox method was used to produce Au NPs on the surface of CeO₂ nanotubes. Note that the as-synthesized catalysts show outstanding performance for CO catalytic oxidation in terms of both activity and stability, which demonstrates the advantage of CeO₂ nanotubes and the superiority of the catalyst preparation procedure. We believe that this study would provide a useful route for the exploration and preparation of highly active catalysts.

2. Experimental

2.1 Chemicals, gases, and materials

The reagents including C₂H₅OH, NH₃·H₂O (25–28 vol%), and NaOH were purchased from Beijing Chemical Works and Shantou Xilong Chemical Industry Incorporated Co., Ltd. HAuCl₄·4H₂O and AgNO₃ were bought from Sinopharm Chemical Reagent Co., Ltd. CeCl₃·7H₂O was purchased from Tianjin Jinke Fine Chemical Research Institute. Poly(ethylene oxide)-b-poly(propyleneoxide)-b-poly(ethylene oxide) Pluronic P123 (EO20PO70EO20, M_av = 5800 g mol⁻¹) was purchased from Sigma-Aldrich, Inc. All the chemicals were analytically pure and used without further purification. Deionized water was used in all the experiments. All the gases such as high purity nitrogen, high purity carbon monoxide, high purity oxygen, and high purity helium gas used in our tests were 99.999% pure and purchased from Beijing BeiWen Gases Factory.

2.2 Preparation of CeO₂ nanotubes

CeO₂ nanotubes were synthesized by the conventional solvothermal method. In a typical synthesis procedure, 5.8 g of P123 and 1.86 g of CeCl₃·6H₂O were dissolved in the mixture of 20 mL of deionized water and 20 mL of ethanol. After vigorous stirring and ultrasonication for 30 min, a clear aqueous colloid was obtained. An appropriate amount of NH₃·H₂O was added dropwise to the mixture to precipitate the salts, and the pH value of the mixture was carefully controlled between 9 and 10. After the violet flocculent precipitation appeared, the mixture was sequentially stirred for 30 min at room temperature and transferred into a 50 mL Teflon-lined autoclave. The autoclave was heated to 160 °C for 72 h before being cooled inside the furnace to room temperature. The precipitate was obtained by centrifugation and then washing several times with ethanol and deionized water. Finally, the CeO₂ powders were obtained by drying the precipitates in an oven at 60 °C for 2 h and then calcination at 400 °C for 5 h.

2.3 Preparation of the Au/CeO₂ nanotubes

Au was deposited on the surface of CeO₂ nanotubes via a modified auto-redox reaction.²⁶ Typically, the CeO₂ nanotubes were dispersed in 25 mL of water to form a homogeneous slurry. Different amounts of HAuCl₄ were added to 20 mL of deionized water to change the Au loading amount, and then, pH of the yellow solution was adjusted to 10 by the addition of a 0.2 M NaOH aqueous solution. Once the pH value was stable, the solution was added to the CeO₂ slurry. After adjusting the pH value of the slurry to 10, the slurry was vigorously stirred for 18 h at room temperature or under heating condition. The solid was isolated by filtration and repeatedly washed with water until there were no chloride ions in the filtrate to be tested by AgNO₃. After this, the products were dried at 80 °C in vacuum overnight.

2.4 Characterization

Powder X-ray diffraction (XRD) patterns were obtained via a Rigaku D/max 2500 X-ray power diffractometer using Cu Kα radiation (λ = 1.54056 Å). Scanning electron microscopy (SEM) was performed using a Hitachi S-4800 electron microscope. Transmission electron microscopy (TEM) was carried out via an FEI Tecnai G² F20 electron microscope operated at 200 kV with the software package for automated electron tomography. Au content was determined by energy dispersive spectroscopy (EDS) using an FEI Tecnai G² F20 electron microscope. X-ray photoelectron spectra (XPS) were obtained by an ESCALAB 250 Xi XPS system of Thermo Scientific, where the analysis chamber was 1.5 × 10⁻⁹ mbar and the X-ray spot was 500 nm.

2.5 Catalytic evaluation test

The experiments for catalytic CO oxidation were carried out using a continuous fixed-bed reactor coupled online with a gas chromatograph (GC-2014C, Shimadzu), a molecular sieve column, and a thermal conductivity detector (TCD). The operation conditions were as follows: 200 mg of catalysts (40–60 Meshes), a gas mixture of 1% CO, 1.6% O₂, and 97.4% He (balance gas) with a total flow rate of 50 mL min⁻¹ corresponding to a gas hourly space velocity (GHSV) of 15 000 mL h⁻¹ g_cat⁻¹, and the programmed-temperature rate was 0.5 °C min⁻¹.

3. Results and discussion

3.1 Synthesis and characterization of the CeO₂ nanotubes

The morphologies and structures of the CeO₂ nanotubes were investigated by SEM and TEM (Fig. 1a and b). It was found that the CeO₂ samples are hollow tubular in shape with a narrow size distribution, and the average dimension of the CeO₂ nanotubes is about 35 ± 5 nm in inner diameter, 70 ± 10 nm in outer diameter, and 2.5 ± 0.5 μm in length. As observed from the XRD analysis, all the diffraction peaks can be assigned to a cubic fluorite-type Fm3m crystal structure of CeO₂ (a = 0.5411 nm, JCPDS No. 34-0394 or PDF card No. 01-081-0792) (Fig. 1d), with no other crystalline impurities being detected. Moreover, the high resolution TEM (HRTEM) images indicate that the shell of the tubes is about 15 nm in diameter, and two sets of atomic lattice fringes with the interplanar spacing of 0.27 nm and 0.31 nm correspond to the (200) and (111) crystal planes of the cubic CeO₂, respectively (Fig. 1c). Note that the obtained material is highly crystalline, which exhibits a single crystal-like pattern with ordered bright matrix diffraction spots (inset in Fig. 1c). There are two sets of diffraction spots possibly because the electron beam penetrates the thin wall of the tube and projects on the other side of the wall, and the walls at the opposite sides have a small difference in crystal orientation. However, metal oxides with high crystallinity are ideal support materials for a catalytic reaction due to their excellent chemical stability, high-temperature resistivity, and good repeatability. Moreover, good crystallinity means fewer defects and traps for electron transfer, which is also favourable for the catalytic reaction.

Fig. 1 Morphologies and structures of CeO₂ nanotubes. (a) TEM image of the CeO₂ nanotubes, (b) SEM image of the CeO₂ nanotubes, (c) HRTEM images of the CeO₂ shell of the nanotubes, and inset is the corresponding selected area electron diffraction (SAED) pattern. (d) XRD pattern of the CeO₂ nanotubes.

Note that this synthesis method is simple, effective, and controllable. Impressively, the product was obtained in high yield (about 80% of the product was in tubular morphology) and had good reproducibility. These advantages are critical for practical application of advanced materials.

3.2 Study of the CeO₂ nanotube formation process

A complete understanding of the CeO₂ nanotube formation and growth in solution is essential for the optimization of the synthesis method and provides guidance for the design and fabrication of CeO₂ and other nanocrystals of well-defined morphologies. Therefore, the crystal growth mechanism of the CeO₂ nanotube was elucidated according to the analysis of the crystal structures and morphologies obtained after reaction for different time periods.

As shown in Fig. 2, the product at the initial stage of reaction demonstrates an amorphous state (Fig. 2a). Subsequently, the synthesized nanocrystals changed to the regular hexagon sheets with an average size of about 5 nm. HRTEM images show that most particles are uniform, with a (111) crystal face and a (100) crystal face exposed (Fig. 2b). The following growth process reveals severe agglomeration of the sheets. The most crystals appear to have the irregular stereo shapes, with an average size of around 10–20 nm. These particles tend to further agglomerate together. The TEM image shows that elongated nanocrystals originate from the fusion of some neighbouring nuclei via sharing of the common crystallographic orientation, resulting in the elongated or assembled structure (Fig. 2c). After this, the CeO₂ nanorods with a rather rough surface were formed, and a close observation of these nanorods disclosed that they were assemblies and composed of small NPs²⁹ (Fig. 2d). After this, the inner structure of the rods tended to be loosened (Fig. 2e), leading to the occurrence of some cavities with rapidly increasing diameter (Fig. 2f). Simultaneously, gain boundaries began to coalesce and disappear (Fig. 2h and g). Finally, nanotubes with uniform and smooth walls were formed (Fig. 2i). Note that the phase of the samples did not change with the prolongation of the reaction time (Fig. S1, ESI†). The samples that reacted in a very short time (about 5 min) already displayed highly crystalline structures, and the crystal pattern was retained regardless of the changes in the morphology and size.

Fig. 2 Investigation of the growth process of CeO₂ nanotubes. (a–i) TEM images of the sample obtained after reaction for different time periods. (a) 0 min, (b) 5 min, (c) 30 min, (d) 1 h, (e) 4 h, (f) 12 h, (g) 24 h, (h) 36 h, and (i) 72 h.

Unlike the most common nanotube synthesis via micelle templating of the self-assembled surfactant molecules, the product obtained herein exhibited a transformation from ceria nanoparticles to nanorods and finally to nanotubes. According to the study on the intermediate state of CeO₂ nanotubes, we proposed an oriented attachment-Ostwald ripening mechanism (shown in Scheme 1). At the beginning, CeO₂ nucleated, which involved precipitation of Ce³⁺ cations by OH⁻ ions to form Ce(OH)₃²⁸ and transition from Ce(OH)₃ to CeO₂ nuclei (hexagon sheets, 4–5 nm) through an oxidation and dehydration process. The oxidation and dehydration process rapidly and simultaneously occurred because no crystalline Ce(OH)₃ or Ce(OH)₄ phase was found in the product. After the initial formation of CeO₂ nanosheets was finished, the oriented CeO₂ attachment nanosheets were formed. Under the alkaline conditions (pH ≈ 9–10), the dehydration process occurred among hydroxyl groups on the surface of CeO₂ to form Ce–O–Ce bonds, giving rise to aggregation between adjacent CeO₂ nanosheets. As the CeO₂ nanosheets possess both low energy facets (mainly (111)) and relatively higher energy facets (the edges of nanosheets, probably (100) and (110)),²¹ they tend to orientedly aggregate by sharing the high energy facets at the edges of the nanosheets through coalescing and fusion. Instead of separately growing larger laterally, the sheets cylindrically curved into a closed surface to reduce the energy of the system further (taking the kinetic speed factors into consideration as well). Then, the subsequent nanosheets tended to settle down on the established cylinder surface, and the radius dimension of the nanostructure gradually increased. Finally, the nanosheets stacked to form a 1D elongated nanorod morphology with the low energy facets exposed. The following stage of nanocrystal growth was dominated by the Ostwald ripening process. When the energetically unstable structure made of many small nanocrystals was formed, it tended to form a more energetically favoured regular well-defined larger crystalline structure at the expense of small crystallites to minimize the total surface area. Thus, the Ostwald ripening process occurred, and the cavity was generated in a location where crystallites were smaller and/or less compact.²⁷ The smaller CeO₂ grains dissolved and left cavities at the center of the rods. The cavity within the nanorods became bigger and bigger as the ripening process proceeded further. Finally, the dissolution and precipitation occurred and converted the oriented aggregated assemblies into regular crystalline nanotubes. In this specific case, we can observe that the rough rod-like aggregates consisting of small CeO₂ nanocrystals finally turned into uniform single-crystalline-like nanotubes. Note that the surfactant P123 plays an important role in the hollow cavity formation process, and the control experiments indicate that there is barely any transformation from nanorods to nanotubes without the addition of P123 (Fig. S2, ESI†). We conjectured that the roles of P123 were as follows: (1) P123 hindered the growth and recrystallization of the inner crystallites via partially blocking the initial CeO₂ nuclei to restrain small size or loose morphology; (2) P123 weakened the aggregation and oriented attachment of the nanosheets by capping the surfaces of the primary crystals, such that the fusion between adjacent crystals via elimination of surface hydroxyls and formation of the Ce–O–Ce bond was hindered as well. Apparently, an appropriate amount of P123 that could provide an ideal concentration gradient from the center to the outer part of the nanorods is the key factor for formation of highly qualified nanotubes.

Scheme 1 Schematic of the formation of CeO₂ nanotubes using the solvothermal method.

3.3 Synthesis and characterization of the Au/CeO₂ nanotubes

Different from previous reports, in this study, Au NPs could be formed on CeO₂ nanotubes by simply adjusting pH of the reaction system without the addition of any other reductant. As shown in the TEM image (Fig. 3a), the Au NPs of less than 5 nm with darker contrast were deposited on the surface of the CeO₂ nanotubes, and the composite can be confirmed by the EDS analysis. The XRD pattern further shows that the product is composed of fluorite-type CeO₂, and some diffraction peaks of Au can be discerned although they are very weak, owing to the low loading amount and small particle size. Moreover, the XPS spectrum of Au 4f displays typical duplet peaks with the binding energy of 87.6 and 83.9 eV, revealing that Au existed in the elementary metallic form.^22,24 Moreover, the O 1s XPS spectrum shows an apparent difference between CeO₂ and Au/CeO₂ nanotubes (Fig. S3, ESI†), which is another indication of the successful loading of Au NPs.

Fig. 3 Morphology and structure of the Au/CeO₂ nanotubes. (a) TEM image of the Au/CeO₂ nanotubes. (b) XRD pattern of the Au/CeO₂ nanotubes. (c) EDS analysis of the Au/CeO₂ nanotubes. (d) XPS spectrum of the Au 4f of Au/CeO₂ nanotubes.

With respect to the loading process of the Au NPs, we proposed that its reduction mechanism is as follows. Under aqueous conditions, chloroauric acid ionized into protons and chloroauric ions. Chloroauric ions are known to be extremely unstable and tend to decompose. At higher pH (about 10.0), the bound chlorine ions are partially released and replaced by the hydroxyl groups, and the reaction results in the formation of [Au(OH)_xCl_4−x]⁻ anions.²⁹ After the CeO₂ nanotube slurry is added into the solution, many active sites are provided for the gold precursors. For instance, the surrounded hydroxyl groups on CeO₂ would absorb Au³⁺ to form Ce–O–Au bonds. By maintaining the pH value of the system at 10, abundant OH⁻ would act as reductant and donate the electrons to [Au(OH)_xCl_4−x]⁻ anions, promoting the reduction of Au(III). In addition to the abovementioned favourable kinetic conditions, the semiconductor property of CeO₂ may contribute to the thermodynamic driving force for Au NP formation.³⁰

3.4 Catalytic test of Au/CeO₂ nanotubes

To evaluate the performance of Au/CeO₂ nanotubes for catalytic oxidation of CO, Au/CeO₂ nanotubes with different Au loading amounts and different particle sizes were investigated along with a contrast sample of bare CeO₂ nanotubes. Fig. 4 summarizes the typical conversion ratio of CO as a function of reaction temperature over five catalysts under the same conditions. It can be seen that the Au/CeO₂ samples are already active at room temperature and the CO conversion ratio is about 20%. By comparison, pure CeO₂ nanotubes only show activity above 200 °C. With the increasing reaction temperature, the complete CO conversion for the Au/CeO₂ sample was achieved at about 73 °C. As for the CO oxidation over pure CeO₂, a temperature of 338 °C was needed to reach a conversion ratio of 100%. Note that to date, only slight improvement in the catalytic activity over other CeO₂-based materials has been reported in the literature.³¹ Evidently, the presence of Au NPs, regardless of the fact that whether they are inside or outside the CeO₂ tubes, could efficiently enhance the catalytic activity of CeO₂, demonstrating the synergetic effect of Au and CeO₂ during the catalytic processes. On the one hand, when Au was present, more oxygen vacancies formed and the oxygen species nearby were activated.³² On the other hand, CeO₂ contributed to the dispersion and anchoring of Au nanoparticles, prevented them from sintering together and maintained the high activity during the catalytic reaction process.³³ Thus, the catalytic performance showed strong mutual reinforcement.

Fig. 4 Catalytic activities of CeO₂ nanotubes with different gold loading amount and different particle sizes.

Moreover, we compared the catalytic activity of the Au/CeO₂ nanotubes with different Au loading contents. As shown in Fig. 4, the order of catalytic activity was obtained as follows: 1.5% Au/CeO₂ > 1% Au/CeO₂ > 0.5% Au/CeO₂ ≫ CeO₂. The difference in the catalytic activity mainly originates from different active sites in the samples. The sizes of Au NPs in all the samples were around 3–5 nm (Fig. 3a and Fig. S4a, ESI†), and the increase of Au content was favourable for the improvement of the catalytic activity. This is because the Au species, which activated and enriched the CO and O₂ molecules and greatly decreased the absorption energy of both reactants, were the active sites for CO oxidation in Au/CeO₂ systems. Increase in the number of active sites with the increasing Au content was favorable for catalysis. The investigation on Au particle size demonstrates that the sample of 1% Au/CeO₂ with the Au NP sizes of about 5 nm (Fig. S4b, ESI†) possesses much higher catalytic activity than that with Au NPs of 10 nm diameter. We attributed this to the increase of low coordination Au sites and the stronger interaction between the Au NPs and supports for smaller Au NPs compared with those for larger Au NPs.

Stability is another important index to assess the performance of the catalysts. We set the temperature of the catalytic reaction at 100 °C, at which the CO conversion ratio was 100%. As displayed in Fig. S5 (ESI†), no deactivation occurred even when the catalytic reaction was performed for as long as 100 h. The XPS spectrum further reveals that there are no obvious changes in the chemical state of the active sites (Fig. S6, ESI†) after the catalytic reaction. Evidently, encapsulation of Au NPs inside the CeO₂ matrix or the strong attachment of Au NPs on the CeO₂ surface can effectively improve their stability during the catalytic process.

4. Conclusions

In summary, by developing a precisely controlled surfactant-assisted solvothermal method, CeO₂ nanotubes with high quality and yield were constructed. Compared with the previously reported results, the CeO₂ nanotubes could be obtained in much higher yield, better uniformity, and higher crystallinity. Detailed analysis of the synthetic procedure discloses that the formation of the CeO₂ nanotubes follows the oriented attachment-Ostwald ripening mechanism. After the loading of Au NPs, the Au/CeO₂ nanotubes show excellent performance in the CO oxidation reaction with the full CO conversion temperature as low as 73 °C. Furthermore, these Au/CeO₂ nanotubes exhibit a high stability, and no deactivation happens after 100 h of reaction. This solvothermal strategy is expected to open the avenue towards the synthesis of nanomaterials with different morphologies as well as rational design and fabrication of the structure-controlled catalysts.

Acknowledgements

We are grateful for the financial support received from the National Natural Science Foundation of China (No. 21271021, 51472025, 51572261) and Youth Innovation Promotion Association of CAS (No. 2017070).

Notes and references

1. M. Haruta, N. Yamada, T. Kobayashi and S. Iijima, J. Catal., 1989, 115, 301–309.
2. G. X. Chen, Y. Zhao, G. Fu, P. N. Duchesne, L. Gu, Y. P. Zheng, X. F. Weng, M. S. Chen, P. Zhang, C. W. Pao, J. F. Lee and N. F. Zheng, Science, 2014, 344, 495–499.
3. H. Y. Kim and G. Henkelman, J. Phys. Chem. Lett., 2012, 4, 216–221.
4. H. J. Freund, G. Meijer, M. Scheffler, R. Schlogl and M. Wolf, Angew. Chem., Int. Ed., 2011, 50, 10064–10094.
5. W. L. Deng and M. Flytzani-Stephanopoulos, Angew. Chem., Int. Ed., 2006, 45, 2285–2289.
6. J. Li, Z. Zhang, W. Gao, S. Zhang, Y. Ma and Y. Qu, ACS Appl. Mater. Interfaces, 2016, 8, 22988–22996.
7. C. W. Sun, H. Li and L. Q. Chen, Energy Environ. Sci., 2012, 5, 8475–8505.
8. S. Zhang, C. R. Chang, Z. Q. Huang, J. Li, Z. Wu, Y. Ma, Z. Zhang, Y. Wang and Y. Qu, J. Am. Chem. Soc., 2016, 138, 2629–2637.
9. Y. Liu, H. Li, J. Yu, D. Mao and G. Lu, Phys. Chem. Chem. Phys., 2015, 17, 27758–27768.
10. M. M. Branda, C. Loschen, K. M. Neyman and F. Illas, J. Phys. Chem. C, 2008, 112, 17643–17651.
11. W. X. Huang, Top. Catal., 2013, 56, 1363–1376.
12. L. S. Zhong, J. S. Hu, A. M. Cao, Q. Liu, W. G. Song and L. J. Wan, Chem. Mater., 2007, 19, 1648–1655.
13. J. Qi, K. Zhao, G. Li, Y. Gao, H. Zhao, R. Yu and Z. Tang, Nanoscale, 2014, 6, 4072–4077.
14. L. Yan, X. R. Xing, R. B. Yu, J. X. Deng, J. Chen and G. R. Liu, Phys. B, 2007, 390, 59–64.
15. P. Xu, R. Yu, C. Wang, X. Yan, D. Wang, J. Deng, J. Chen and X. Xing, J. Nanosci. Nanotechnol., 2013, 13, 1498–1502.
16. L. Yan, X. R. Xing, R. B. Yu, L. J. Qiao, J. Chen, J. X. Deng and G. R. Liu, Scr. Mater., 2007, 56, 301–304.
17. L. Yan, R. Yu, J. Chen and X. Xing, Cryst. Growth Des., 2008, 8, 1474–1477.
18. R. Yu, L. Yan, P. Zheng, J. Chen and X. R. Xing, J. Phys. Chem. C, 2008, 112, 19896–19900.
19. X. Yao, X. Yang, R. Yu, P. Xu, J. Chen and X. Xing, Mater. Res. Bull., 2015, 61, 22–25.
20. K. Zhao, J. Qi, S. Zhao, H. Tang, H. Yin, L. Zong, L. Chang, Y. Gao, R. Yu and Z. Tang, Chin. J. Catal., 2015, 36, 261–267.
21. J. Qi, X. Y. Lai, J. Y. Wang, H. J. Tang, H. Ren, Y. Yang, Q. Jin, L. J. Zhang, R. B. Yu, G. H. Ma, Z. G. Su, H. J. Zhao and D. Wang, Chem. Soc. Rev., 2015, 44, 6749–6773.
22. X. Y. Lai, J. E. Halperta and D. Wang, Energy Environ. Sci., 2012, 5, 5604–5618.
23. K. Yoon, Y. Yang, P. Lu, D. H. Wan, H. C. Peng, K. S. Masias, P. T. Fanson, C. T. Campbell and Y. N. Xia, Angew. Chem., Int. Ed., 2012, 51, 9543–9546.
24. Z. Tang, X. Yin, Y. Zhang, N. Zhang and Y. Xu, Chin. J. Catal., 2013, 34, 1123–1127.
25. W. Q. Han, L. J. Wu and Y. M. Zhu, J. Am. Chem. Soc., 2005, 127, 12814–12815.
26. S. Carrettin, P. Concepción, A. Corma, J. M. López Nieto and V. F. Puntes, Angew. Chem., Int. Ed., 2004, 43, 2538–2540.
27. H. C. Zeng, Curr. Nanosci., 2007, 3, 177–181.
28. M. Lin, Z. Y. Fu, H. R. Tan, J. P. Y. Tan, S. C. Ng and E. Teo, Cryst. Growth Des., 2012, 12, 3296–3303.
29. S. Wang, K. Qian, X. Z. Bi and W. X. Huang, J. Phys. Chem. C, 2009, 113, 6505–6510.
30. B. Schumacher, V. Plzak, M. Kinne and R. J. Behm, Catal. Lett., 2003, 89, 109–114.
31. J. Qi, J. Chen, G. Li, S. Li, Y. Gao and Z. Tang, Energy Environ. Sci., 2012, 5, 8937–8941.
32. V. Shapovalov and H. Metiu, J. Catal., 2007, 245, 205–214.
33. J. Shi, Chem. Rev., 2012, 113, 2139–2181.

---

Footnotes

† Electronic supplementary information (ESI) available: Fig. S1–S6. See DOI: 10.1039/c7qm00134g

‡ These authors contributed equally to this work.

---