Investigation of oxygen vacancies on Pt- or Au-modified CeO₂ materials for CO oxidation

Abstract

Metal (Au or Pt)-modified CeO₂ materials with excellent redox ability for the oxidation of CO to CO₂ were prepared by a redox precipitation method to investigate the role of surface oxygen vacancies. It was found that most of the Au or Pt was encapsulated by CeO₂. Both Au and Pt can effectively improve the reducibility of CeO₂ to form more oxygen vacancies at lower temperatures. The surface oxygen vacancies were thoroughly studied by UV-Vis spectroscopy, CO adsorption FTIR, H₂-TPR and CO pulse oxidation techniques. During the reaction cycle of CO oxidation, the oxygen vacancies were affected by the modified metals but played different roles in the performance of CO oxidation. There were more oxygen vacancies on Pt-CeO₂, which facilitated the formation of more carbonates. A certain amount (4.2%) of oxygen vacancies was necessary to improve the CO oxidation on this material. Comparatively, the rather fewer oxygen vacancies and the lower adsorption strength of carbonates did not affect the final elimination efficiency of CO on Au-CeO₂.

---

1. Introduction

Carbon monoxide (CO) is widespread in the environment from the incomplete combustion of substances containing carbon, and this can have serious toxic effects on human health by binding to cytochrome oxidase and cytochrome P450.^1,2 In fact, the CO interferes with oxygen transport in blood due to its binding to heme iron centers (Fe²⁺) in hemoglobin.^3,4 In addition, CO is also a pollutant in the energy-related field, especially in the clean energy application of H₂ resources in polymer electrolyte membrane fuel cells.⁵ Nowadays, H₂ is mainly produced by the reforming of hydrocarbons; it usually contains small amounts of CO (0.5–2 vol%) even after a series of purification strategies such as high- and low-temperature water gas shift reactions.⁶ This amount of CO can poison the active metal species, leaving them unable to activate O₂. Therefore, a common-sense idea to eliminate the detrimental effect of CO in the fields of the environment and energy is to decrease its binding energy but more importantly to increase the adsorption and activation of oxygen either by providing alternative sites or by promoting the redox ability of the metal oxide.^7,8

Ceria (CeO₂) is an extensively used metal oxide with a good redox ability as an oxygen buffer and supplier.^9–11 It is well known that ceria stores and releases oxygen due to the easy cycle between Ce³⁺ and Ce⁴⁺ via the formation of surface oxygen vacancies.^12–14 Ceria can also serve as a stabilizer, maintaining a high dispersion of the active metal species.¹⁵ Surface science studies have found that Pt encapsulated by CeO₂ can exhibit a higher performance than the metal itself.^16,17 A “normal-support activation” model was established, in which there was a strong interaction between the encapsulated metal and CeO₂, with the activated CeO₂ playing a role in providing active sites.¹⁸ By leaching the surface Au or Pt species while retaining those which interacted strongly with CeO₂, the Au- or Pt-CeO₂ was still effective in CO elimination.¹⁹ The Pt encapsulated in CeO₂ materials was even selective to a water gas shift reaction without CO methanation.²⁰ All of these studies suggested that the activated CeO₂ can act as active sites for CO elimination with no demand for direct participation of metal species. Recently, it was found that, during the preparation of CeO₂-based metal materials, the Ce³⁺ in the precursor could be oxidized to form Ce⁴⁺ hydroxyl species in solution under alkaline environments and O₂, which was usually coupled with the reduction of high valence metal species to metallic ones, such as Pt⁴⁺ to Pt⁰ or Ir⁴⁺ to Ir⁰.^20,21 This provided a driving force to make a structure of a metal encapsulated by CeO₂, which helped in the study of the reducibility of CeO₂ and the role of oxygen vacancies in the oxidation of CO.

In this work, the reducibility of CeO₂ was modified by Pt and Au metals, which were selected due to their having the same 5f orbital and their extensive use in the elimination of CO.^22–24 These samples were prepared by a redox reaction between Ce and metal ions during a co-precipitation process.²⁵ The structures of the Pt- and Au-CeO₂ samples were detected by XRD, XPS, HRTEM and CO adsorption FTIR techniques. The role of these metals in the formation of oxygen vacancies was studied in terms of their metals' nature by UV-Vis spectra and H₂-TPR techniques. Then the dependence of the CO oxidation performance of these samples on the amount of oxygen vacancies was detected by a pulse technique during a sequence of CO + O₂ pulses.

2. Experimental

2.1 Sample synthesis

The CeO₂ modified by Au (denoted as Au-CeO₂) material was prepared by a redox precipitation method. Typically, 1.1 g of HAuCl₄ solution and 2.52 g Ce(NO₃)₃·6H₂O were mixed in 100 mL de-ionized water. The mixed solution was added to 100 mL NaOH solution (0.18 mol L⁻¹) at 80 °C to gradually form a purple precipitate with a pH of around 9. During the precipitation, the following reaction occurred: Au³⁺ + 3Ce³⁺ → Au⁰ + 3Ce⁴⁺. After aging for 2 h, the suspension was filtered and washed with deionized water several times until no residual Cl⁻ ions were detected by AgNO₃ solution. Then the sample was dried at 60 °C overnight and finally calcined at 400 °C for 2 h. The Pt-CeO₂ sample was prepared by precipitation from H₂PtCl₆ and Ce(NO₃)₃ under similar conditions. The CeO₂ support was also prepared by precipitation from Ce(NO₃)₃ under similar conditions as a comparison.

2.2 Characterization

The loading amounts of Au and Pt in the as-synthesized materials were determined by inductively coupled plasma spectrometry (ICP-AES) on an IRIS Intrepid II XSP instrument (Thermo Electron Corporation).

X-ray diffraction (XRD) patterns were recorded on a PW3040/60 X′Pert PRO (PANalytical) diffractometer equipped with a Cu Kα radiation source (λ = 0.15432 nm), operating at 40 kV and 40 mA. A continuous mode was used for collecting data in the 2θ range from 10° to 80° at a scanning speed of 10° min⁻¹.

UV-Vis spectra were detected on a Cintra (GBC) apparatus. Before the measurements, the sample was planished in a stainless pot. The spectra were obtained with BaSO₄ as a reference.

The reducibility of the sample CeO₂ modified with Au or Pt was detected by a H₂ temperature programmed reduction technique (H₂-TPR), which was performed on an Auto Chem II 2920 automatic material characterization system. First, a 100 mg sample was loaded into a U-shaped quartz reactor and purged with He at 120 °C for 2 h to remove adsorbed carbonates and hydrates. Then, after cooling to a temperature below 0 °C, the flowing gas was switched to 10 vol% H₂/Ar, and the catalyst was heated to 500 °C at a ramping rate of 10 °C min⁻¹.

CO adsorption Fourier transform infrared (FTIR) spectra were acquired with a BRUKER Equinox 55 spectrometer, equipped with an MCT detector and operated at a resolution of 4 cm⁻¹. Before the measurements, a 40 mg sample was pretreated in 20 mL min⁻¹ H₂ at 400 °C for 1 h, and cooled to room temperature under He; then a background spectrum was recorded. After that, a flow of 1% CO/He was introduced. All FTIR spectra were obtained under steady-state conditions.

The formation of oxygen vacancies and carbonates and their relationship with the activity of CO oxidation were detected by a pulse reactor system. In a typical experiment, 100 mg of a sample (m_cat.) were placed in the U-type sample vessel. The effluents were analyzed by gas chromatography (GC Agilent 6890N). The amount of reactants consumed was determined quantitatively from the change in the peak areas of reactants injected and eluted. During the pretreatment, the quantity of oxygen vacancies was determined by the injection of a certain amount of CO at 400 °C, which was reflected in the released CO₂. The ratio between the amounts of CO₂ produced and the amounts of surface-reducible oxygen species from the H₂-TPR results determined the amount of oxygen vacancies while the ratio of the accumulated carbon species to surface-reducible oxygen species determined the coverage of carbonates (unit: monolayer (ML)). Then the pretreated sample was deposited at 120 °C. Reactants CO and O₂ (molar ratio: 2 : 1) were injected into the pretreated sample. The elimination efficiency of CO to CO₂ under steady state was calculated from the equilibrium of C and O species.

The amount of oxygen vacancies was calculated as:

The quantity of carbonates was calculated as:

nⁱⁿ_CO and were the amounts of CO injected and CO₂ released during CO pretreatment, respectively; m_cat. was the amount of a sample; n_O was the amount of surface reducible oxygen species from the H₂-TPR result.

2.3 Activity test

Tests of the efficiency of CO elimination were performed with a continuous flow fixed bed reactor under atmospheric pressure. Before evaluation, 100 mg of the sample were flushed by He or pre-reduced in a flow of 20 mL min⁻¹ of 10 vol% H₂/He at 400 °C for 120 min to detect the effect of oxygen vacancies on these samples. A 2.0 vol% CO + 1.0 vol% O₂ + He mixture of 67 mL min⁻¹ was used for the evaluation of the CO oxidation reaction. The amounts of the reactants and products in the inlet and outlet streams were analyzed by an on-line gas chromatograph (Agilent 6890N) using He as carrier gas.

3. Results and discussion

3.1 Structure characterizations by XRD and UV-Vis spectra

The structure of the CeO₂ modified by the metal material was characterized by an XRD technique. As shown in Fig. 1, there was no peak attributed to the modified metal Pt or Au in the diffraction patterns of all samples, indicating that these metal species were highly dispersed or encapsulated by CeO₂. Only the diffractograms corresponding to the CeO₂ fluorite structure (JCPDS 00-034-0394) were observed. According to the Scherrer equations, the CeO₂ crystallite sizes of Pt-CeO₂ and Au-CeO₂ were 12 nm and 11 nm as shown in Table 1, respectively, which were similar to that of CeO₂ (13 nm). On the other hand, the modified metals can lower the lattice parameter of CeO₂. As shown in Table 1, the Pt-CeO₂ presented the lowest value analyzed from XRD results, which indicated that the Pt most modified the CeO₂; thus the Pt-CeO₂ might have the most intrinsic defects. HRTEM results in Fig. S1† showed that no metal particles could be found, probably because the Pt or Au particles were highly distributed and too small to be detected in both samples. On the other hand, we can find the highly crystalline ceria support with particle sizes of around 10 nm, in accordance with the XRD results.

Fig. 1 XRD patterns of CeO₂, Au-CeO₂ and Pt-CeO₂ samples.

Table 1 Some physicochemical properties of the samples

Sample	Metal loading^a (wt%)	dCeO2^b (nm)	Lattice parameter^b (Å)	Work function	Band gap^c		Reducibility^d (μmol gcat^−1)
					λ (nm)	(eV)	Peak 1 (T)
a From ICP detects.b Calculated from XRD results.c From UV-Vis spectra results.d From H2-TPR results.
CeO2	—	13	5.412	—	480	2.583	296 (356 °C)
Au-CeO2	0.95	11	5.408	5.1	412	3.010	388 (146 °C)
Pt-CeO2	1.00	12	5.403	5.7	394	3.147	447 (19 °C)

---

The modification of CeO₂ by the metals can be reflected in the perturbation of the electronic structure of these materials. The electronic band transition of the Ce–O bond was evaluated by UV-Vis diffuse reflectance spectroscopy with the band transition energy extrapolated by a straight line. The results in Fig. 2 clearly show that the blue shift of the absorption edge increased with the modified metal in CeO₂. As listed in Table 1, with the higher work function of the metal, the band transition energy increased. The Pt-CeO₂ exhibited the highest energy. Coupled with the XRD results, this shows that the Ce–O bond was most affected by the Pt metal, which meant that Pt-CeO₂ exhibited the highest redox ability.

Fig. 2 UV-Vis diffuse reflectance spectra of pure CeO₂, Au-CeO₂ and Pt-CeO₂ sample.

3.2 Reducibility detected by H₂-TPR

H₂-TPR is a powerful tool to provide quantitative information on the reducibility of CeO₂ modified by metals.^26,27 As shown in Fig. 3 and Table 1, for pure CeO₂, the peak at a temperature of 356 °C can be attributed to the reduction of surface CeO₂. With the presence of metals, the reducibility of surface oxygen on CeO₂ was greatly promoted. The Pt-CeO₂ exhibited a much lower reduction temperature (19 °C) compared with the Au-CeO₂ (146 °C). Moreover, the H₂ consumption on Pt-CeO₂ (447 μmol g⁻¹) was also much higher than that on Au-CeO₂ (388 μmol g⁻¹). Coupled with the XRD and UV-Vis results, this shows that there were more intrinsic defect sites on Pt-CeO₂ and the Ce–O bond strength was much weakened. It has been found that when CeO₂ was deposited over PGM (Pt group metals), the reducibility was thought to be dependent on the work function of the underlying metals.^28,29 The junction effect between CeO₂ and a metal with a higher work function provided the driving force. This effect led to the higher consumption of H₂ to extract the surface oxygen of CeO₂ and then to more oxygen vacancy sites for Pt-CeO₂. Thus a higher redox ability was found to exist on Pt-CeO₂.

Fig. 3 H₂-TPR results of CeO₂, Au-CeO₂ and Pt-CeO₂ samples.

3.3 CO adsorption by FTIR detect

The CO adsorption behaviours on CeO₂ and Au- and Pt-modified CeO₂ samples were investigated by an FTIR technique. As shown in Fig. 4, CO was adsorbed on CeO₂ in the form of carbonates (bands between 1700 cm⁻¹ and 1200 cm⁻¹).³⁰ But when CO was adsorbed on Au-CeO₂ and Pt-CeO₂ catalysts, bands of CO adsorption on metal sites were observed in addition to the carbonates. However, the metal-modified CeO₂ materials exhibited much weaker intensities of the bands for CO adsorption on metal sites than those of the bands attributed to the carbonates, indicating that there were fewer accessible metal species exposed on the surface of CeO₂. These results further confirmed the encapsulation of the metal by CeO₂. On the other hand, the band intensities of the carbonates, especially for Pt-CeO₂, were much higher, possibly due to the creation of more oxygen vacancy sites by Pt than Au in CeO₂ for CO adsorption.

Fig. 4 FTIR results for CO adsorption on CeO₂, Au-CeO₂ and Pt-CeO₂ samples.

3.4 CO elimination efficiency with the oxygen vacancies on CeO₂ modified by metals

For CO oxidation on ceria-based samples, it has been suggested that the redox ability of the Ce–O bond on CeO₂ plays an important role.^14,31,32 In this work, this ability can be reflected in the formation of oxygen vacancies detected in the H₂-TPR results in Fig. 3. We found that the modified metals can improve the reducibility of the CeO₂. On the other hand, there was a large amount of carbonates with CO adsorption, according to the FTIR results in Fig. 4. Both of these factors can affect the elimination efficiency of CO. Based on this consideration, the formation of oxygen vacancies and carbonates was achieved in one step by pretreatment with pulses of CO. The redox precipitation process can result in the encapsulation of Pt or Au by CeO₂ with these metals in metallic states. The XPS results in Fig. S2† showed that before and after H₂ reduction treatment, the Pt or Au exhibited the state of metallic species. Thus the CO₂ was produced only from the CO reaction with the activated CeO₂. With various total pulse numbers, different amounts of oxygen vacancies and carbonates were calculated and are listed in Table 2. Then the yield of CO₂ under steady state was related to these oxygen vacancies by injecting CO + O₂ reactant gases onto these materials. As shown in Fig. 5, it was found that more oxygen vacancies with more carbonates were present on Pt-CeO₂ than on Au-CeO₂, which was in accordance with the H₂-TPR and FTIR results. As for the relationship between CO₂ yields and the pulse number of CO, it was found that the CO₂ yield increased with the CO pretreated pulse number on Pt-CeO₂. However, for pulse numbers higher than 4, this yield reached a constant of around 12 μmol per pulse of CO + O₂. Correspondingly, the coverage of carbonates reached a constant of around 0.1 ML despite the continuously increasing number of oxygen vacancies. The greater coverage of carbonates might prohibit the adsorption and activation of CO and O₂.^21,33 Thus, the further increase in oxygen vacancies did not promote the oxidation of CO. Although it has been found that oxygen vacancies can accelerate the mobility of lattice oxygen and consequently improve the activity of CO oxidation over Pt-CeO₂,^34,35 our work unambiguously and quantitatively demonstrated that the formation of a certain amount of oxygen vacancies was necessary for the high performance of Pt-CeO₂.

Table 2 The quantities of oxygen vacancies (O_v) and carbonates (Carb.) on Au-CeO₂ and Pt-CeO₂ samples during CO pretreatment at 400 °C and the corresponding CO₂ yields (Y_CO2) per pulse of CO + O₂ (No.) under steady state at 120 °C

No.	Au-CeO2			Pt-CeO2
	Ov (%)	Carb. (ML)	YCO2 (μmol)	Ov (%)	Carb. (ML)	YCO2 (μmol)
0	0	0	12.2	0	0	5.3
1	—	—	—	0.8	0.03	9.4
2	—	—	—	1.5	0.05	9.6
4	3.1	0.04	11.7	4.2	0.09	11.4
5	—	—	—	4.4	0.10	11.1
8	4.7	0.05	12.0	5.0	0.11	11.5

---

Fig. 5 The amount of CO₂ yield per pulse of CO + O₂ (2 : 1) under steady state for CO oxidation at 120 °C with the pretreated number of CO pulses on Au-CeO₂ and Pt-CeO₂ samples.

On the other hand, we also studied and compared the oxidation behaviour of CO over Au-CeO₂ to further demonstrate the effects of oxygen vacancies. Similar to Pt-CeO₂, the pretreated number of CO pulses did affect the formation of oxygen vacancies and carbonates. However, these amounts, particularly for the carbonates, were much lower than those for Pt-CeO₂. This lower number of carbonates might not prohibit the adsorption and activation of CO and O₂, and thus the CO elimination efficiency did not change with or without the presence of oxygen vacancies. Therefore, the oxygen vacancies played different roles in CO oxidation on Pt-CeO₂ and Au-CeO₂. The formation of oxygen vacancies can promote the elimination of CO, but also resulted in more carbonates on Pt-CeO₂, while the oxygen vacancies and carbonates did not affect this elimination on Au-CeO₂.

To further prove the effect of oxygen vacancies, we detected the elimination efficiencies of CO with temperature on these materials submitted to different treatments. As seen in Fig. 6, the pure CeO₂ exhibited a very low elimination efficiency of CO. The CO concentrations hardly decreased with temperature even at 200 °C, which might be due to the lower reduction ability of the CeO₂. With metal-modified CeO₂, the CO concentrations decreased significantly. In particular, the reduction treatment greatly improved the elimination efficiency of CO on Pt-CeO₂(r) compared to that just flushed by He (Pt-CeO₂). The temperature for total CO elimination decreased from 200 °C to 100 °C. On the other hand, this efficiency did not change on Au-CeO₂. These results were consistent with those of pulse experimental results shown in Fig. 5, further indicating the importance of oxygen vacancies in CO elimination on Pt-CeO₂ but not on Au-CeO₂.

Fig. 6 CO concentration as a function of temperature over Au-CeO₂ and Pt-CeO₂ submitted to different treatments.

3.5 Discussion

It has been suggested that on CeO₂-encapsulated metal materials, the reducibility of the Ce–O bond is dependent on the work function of the modified metals.²⁹ The UV-Vis spectra results in Fig. 2 demonstrated that with a higher work function of the metals, a greater blue shift of the Ce–O bond occurred on Pt-CeO₂. Thus, the junction effect between CeO₂ and a metal with a higher work function provided the driving force for the reducibility of CeO₂.^28,29 This effect can result in easier reduction of CeO₂ with greater consumption of H₂, as shown from the H₂-TPR results in Fig. 3. Such a high reducibility of CeO₂ promoted the formation of more oxygen vacancies on Pt-CeO₂, favouring the elimination of CO on Pt-CeO_2, as shown in Fig. 6.

For the CeO₂-supported metal materials, it was suggested that the interface between metal and support played an important role, in which the CO adsorbed on metal sites reacted with the O₂ adsorbed on support sites.^31,36 In particular, Cargnello et al. quantitatively controlled the interface length by depositing Pt or Pd metals with different particle sizes onto CeO₂. A model analysis showed that the smaller particles led to an increased boundary length and higher activity, giving direct evidence of the role of interface sites.¹⁵ However, in our study, the FTIR results showed few accessible metal sites for Pt- and Au-CeO₂, while most of the CO can be adsorbed and activated on CeO₂ sites in the form of carbonates, as shown in Fig. 4. The few exposed metal sites may not be favourable for the reaction of CO oxidation at the interface of metal and CeO₂.

As for CO oxidation on these Pt- and Au-encapsulating CeO₂ materials, a redox cycle between Ce³⁺ and Ce⁴⁺ coupled with the oxygen vacancies can be proposed. As shown in Fig. 7, the CO reacted with the surface oxygen atoms of CeO₂, during which the modified metals promoted the reducibility of the Ce–O bond and the easy extraction of these surface oxygen atoms. Then the CO₂ and the carbonates were produced on the surface, which resulted in the corresponding production of oxygen vacancies and the formation of Ce³⁺. It has been suggested that the presence of oxygen vacancies was helpful for the activation of O₂ in the form of superoxide or peroxide, which may further react with CO to form carbonates or dissociate to form atomic oxygen.^37–39 With the filling of these oxygen vacancies by atomic oxygen, the Ce³⁺ can be oxidized to Ce⁴⁺. Some of the adsorbed carbonates may become easily desorbed and decomposed to release CO₂.^40,41 The modified metals can determine the formation of oxygen vacancies by promoting the removal and the deposition of surface oxygen. Different amounts of oxygen vacancies can be found with modified Pt and Au metals with different work functions, which further affect the accumulation of carbonates. The calorimetric results in Fig. S3† indicated that the adsorption strength of carbonates on Pt-CeO₂ was 126 kJ mol⁻¹, while that on Au-CeO₂ was lower at 92 kJ mol⁻¹. Therefore, on Pt-CeO₂, the much easier formation of oxygen vacancies improved the CO elimination efficiency. Comparatively, the lower intensity of carbonates on Au-CeO₂ can be facilely desorbed and decomposed to form CO₂ at 120 °C according to the rule that the temperature (K) for desorption of adsorbed species equates to around four times the adsorption heat value,⁴² which might lead to an unchanged efficiency of CO elimination with or without the oxygen vacancies. This result was in accordance with a temporal analysis of CO oxidation on Au-CeO₂, which showed that CO conversion under steady state was independent of oxygen removal mode.⁴³

Fig. 7 A redox process during CO oxidation on M-CeO₂. O_v: oxygen vacancies.

4. Conclusion

The formation of oxygen vacancies on the Pt- and Au-modified CeO₂ materials and their effects on the elimination efficiency of CO were thoroughly studied. Most of the Pt or Au metal was encapsulated by CeO₂, which can effectively improve the reducibility of the Ce–O bond and subsequently facilitate the formation of oxygen vacancies, particularly for the Pt metal, which has a higher work function. It has been demonstrated that these oxygen vacancies played different roles in the CO elimination efficiency for Pt-CeO₂ and Au-CeO₂. The presence of 4.2% oxygen vacancies was favourable, but greater amounts did not further increase the CO elimination efficiency on Pt-CeO₂. Comparatively, the CO elimination efficiency did not change on Au-CeO₂ with or without the oxygen vacancies.

Acknowledgements

This work was supported by the Excellent Talents Plan Project for Universities of Liaoning Province (LJQ2014055), the Open Project of Key Laboratory for Micro/Nano Technology and System of Liaoning Province (20140401) and the Doctoral Starting up Foundation of Dalian Polytechnic University (61020726). We also gratefully acknowledge the helpful discussion and technical support of Dr Jian Lin from Dalian Institute of Chemical Physics, Chinese Academy of Sciences.

Notes and references

1. J. R. Alonso, F. Cardellach, S. López, J. Casademont and O. Miró, Pharmacol. Toxicol., 2003, 93, 142.
2. B. S. Zuckerbraun, B. Y. Chin, M. Bilban, J. C. d'Avila, J. Rao, T. R. Billiar and L. E. Otterbein, FASEB J., 2007, 21, 1099.
3. P. Haab, Experientia, 1990, 46, 1202.
4. A. Pamplona, A. Ferreira, J. Balla, V. Jeney, G. Balla, S. Epiphanio, A. Chora, C. D. Rodrigues, I. P. Gregoire and M. Cunha-Rodrigues, Nat. Med., 2007, 13, 703.
5. R. Navarro, M. Peňa and J. Fierro, Chem. Rev., 2007, 107, 3952.
6. C. Ratnasamy and J. Wagner, Catal. Rev.: Sci. Eng., 2009, 51, 325.
7. J. Lin, L. Li, Y. Q. Huang, W. S. Zhang, X. D. Wang, A. Q. Wang and T. Zhang, J. Phys. Chem. C, 2011, 115, 16509.
8. D. Widmann and R. J. Behm, Acc. Chem. Res., 2014, 47, 740.
9. A. Trovarelli, Catal. Rev.: Sci. Eng., 1996, 38, 439.
10. J. Yu, G. S. Wu, G. Z. Lu, D. S Mao and Y. Guo, RSC Adv., 2014, 4, 16985.
11. W. X. Zou, C. Y. Ge, M. Y. Lu, S. G. Wu, Y. Z. Wang, J. F. Sun, Y. Pu, C. J. Tang, F. Gao and L. Dong, RSC Adv., 2015, 5, 98335.
12. Y. Zhang, J. Wang and T. Zhang, Chem. Commun., 2011, 47, 5307.
13. C. T. Campbell and C. H. F. Peden, Science, 2005, 309, 713.
14. V. Shapovalov and H. Metiu, J. Catal., 2007, 245, 205.
15. M. Cargnello, V. Doan-Nguyen, T. Gordon, R. Diaz, E. Stach, R. Gorte, P. Fornasiero and C. Murray, Science, 2013, 341, 771.
16. C. Hardacre, R. Ormerod and R. Lambert, J. Phys. Chem., 1994, 98, 10901.
17. J. Rodriguez, S. Ma, P. Liu, J. Hrbek, J. Evans and M. Pérez, Science, 2007, 318, 1757.
18. S. Golunski, R. Rajaram, N. Hodge, G. J. Hutchings and C. J. Kiely, Catal. Today, 2002, 72, 107.
19. Q. Fu, H. Saltsburg and M. Flytzani-Stephanopoulos, Science, 2003, 301, 935.
20. C. M. Y. Yeung, K. M. K. Yu, Q. J. Fu, D. Thompsett, M. I. Petch and S. C. J. Tsang, J. Am. Chem. Soc., 2005, 127, 18010.
21. J. Lin, Y. Huang, L. Li, A. Wang, W. Zhang, X. Wang and T. Zhang, Catal. Today, 2012, 180, 155.
22. Q. Fu and T. Wagnera, Surf. Sci. Rep., 2007, 62, 431.
23. K. Liu, A. Wang and T. Zhang, ACS Catal., 2012, 2, 1165.
24. T. Montini, M. Melchionna, M. Monai and P. Fornasiero, Chem. Rev., 2016, 116, 5987.
25. Y. Q. Huang, A. Q. Wang, L. Li, X. D. Wang, D. S. Su and T. Zhang, J. Catal., 2008, 255, 144.
26. S. Carabineiro, S. Bastos, J. Órfão, M. Pereira, J. Delgado and J. Figueiredo, Appl. Catal., A, 2010, 381, 150.
27. B. C. Zhang, X. L. Tang, Y. Li, W. J. Cai, Y. D. Xu and W. J. Shen, Catal. Commun., 2006, 7, 367.
28. J. C. Frost, Nature, 1988, 334, 577.
29. N. Acerbi, S. C. Tsang, S. Golunski and P. Coller, Chem. Commun., 2008, 1578.
30. P. Bera, A. Cámara, A. Hornés and A. Martínez-Arias, J. Phys. Chem. C, 2009, 113, 10689.
31. S. O. Shekhtman, A. Goguet, R. Burch, C. Hardacre and N. Maguire, J. Catal., 2008, 253, 303.
32. A. M. Venezia, G. Pantaleo, A. Longo, G. Di Carlo, M. P. Casaletto, F. L. Liotta and G. Deganello, J. Phys. Chem. B, 2005, 109, 2821.
33. O. Pozdnyakova, D. Teschner, A. Wootsch, J. Kröhnert, B. Steinhauer, H. Sauer, L. Toth, F. C. Jentoft, A. Knop-Gericke, Z. Paál and R. Schlögl, J. Catal., 2006, 237, 1.
34. A. Holmgren, B. Andersson and D. Duprez, Appl. Catal., B, 1999, 22, 215.
35. H. H. Liu, Y. Wang, A. P. Jia, S. Y. Wang, M. F. Luo and J. Q. Lu, Appl. Surf. Sci., 2014, 314, 725.
36. M. Boaro, F. Giordano, S. Reccia, V. Santo, M. Giona and A. Trovarelli, Appl. Catal., B, 2004, 52, 225.
37. J. Guzman, S. Carrettin and A. Corma, J. Am. Chem. Soc., 2005, 127, 3286.
38. H. Liu, A. Kozlov, A. Kozlova, T. Shido, K. Asakura and Y. Iwasawa, J. Catal., 1999, 185, 252.
39. H. Guan, J. Lin, B. Qiao, X. Yang, L. Li, S. Miao, J. Liu, A. Wang, X. Wang and T. Zhang, Angew. Chem., Int. Ed., 2016, 55, 2820.
40. Z. Song, W. Liu and H. Nishiguchi, Catal. Commun., 2007, 8, 725.
41. C. Binet, M. Daturi and J. Lavalley, Catal. Today, 1999, 50, 207.
42. M. Li and J. Shen, J. Catal., 2002, 205, 248.
43. D. Widmann, R. Leppelt and R. Behm, J. Catal., 2007, 251, 437.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra12049k

---