The stability and deactivation of Pd–Cu–Cl_x/Al₂O₃ catalyst for low temperature CO oxidation: an effect of moisture

Abstract

The catalytic stability of a Pd–Cu–Cl_x/Al₂O₃ catalyst for low temperature CO oxidation was investigated in different conditions. Under ∼0.1% moisture, the CO conversion over the Pd–Cu–Cl_x/Al₂O₃ catalyst can be maintained at 100% for 30 h even at 0 °C, but it deactivated reversibly at 25 °C with ∼0.6% moisture and irreversibly deactivated at 25 °C with ∼3.1% moisture or 0 °C with ∼0.6% moisture. The reversible deactivation is resulted from physical capillary condensation in the small pores of the catalysts. The irreversible deactivation is due to a breakage of the close-knit structure of Pd–Cu–Cl leading to an aggregation or transformation of the active copper phase, and the formation of a carbonate species on the catalyst surface, meaning that the inactive Pd⁰ species over the surface of the Pd–Cu–Cl_x/Al₂O₃ catalyst is hard to re-oxidize back to the active Pd⁺ sites by copper species in the high moisture reaction conditions.

---

1. Introduction

The catalytic oxidation of carbon monoxide at low temperature is an important reaction in practical applications and academic studies in the catalysis field, including air purification, respirators, preferential CO oxidation in the presence of hydrogen for pure hydrogen production. Moisture inevitably exists in practical atmospheres, and sometimes the moisture concentration in the feed gas is near-saturated, however, most fundamental studies are usually carried out under relative dry conditions.^1–3 Therefore, whether the presence of moisture results in a deactivation of catalysts and the reasons for the catalyst deactivation made by moisture are unclear, especially for the Pd–Cu–Cl_x/Al₂O₃ catalyst for low temperature CO oxidation.

Hopcalite catalysts (composed of MnO_x and CuO) and Co₃O₄ were reported to be active for CO oxidation, but they all deactivated severely even in the presence of trace moisture because of a blocking of active sites by water adsorption.^4–6 For gold-based catalysts, both the positive and negative effect of H₂O were reported. Date and co-workers claimed that positive or negative effect of H₂O on the CO catalytic oxidation were dependent not only on the moisture concentration in the feed but also on the support nature of gold catalysts.^7,8 Daniells et al. reported that moisture in the feed significantly enhanced CO oxidation over Au/Fe₂O₃ at room temperature because water can promote the decomposition rate of the carbonate intermediate.⁹

A supported PdCl₂–CuCl₂ catalyst has been reported to be a stable and active catalyst for CO oxidation even in the presence of large amounts of moisture in the feed gas.^10–12 Hence, the supported PdCl₂–CuCl₂ catalyst was considered to be a promising catalyst for CO oxidation in the presence of water at low temperature. It has been reported that alumina supported PdCl₂–CuCl₂ catalysts are inferior to carbon supported catalysts, and their catalytic performances are markedly dependent on the partial pressure of water, because alumina is hydrophilic and its pores are filled by water when alumina is exposed to high water concentrations, whereas carbon is more hydrophobic.^13,14 The PdCl₂–CuCl₂ supported on alumina was conventionally prepared by a wet impregnation method. We have reported that alumina supported Pd–Cu–Cl_x catalysts prepared by a coordination-impregnation (CI) method exhibit much higher catalytic activity for CO oxidation than the PdCl₂–CuCl₂ on alumina catalysts prepared by a conventional wet impregnation method,^15,16 and we also found that the active palladium species in the Pd–Cu–Cl_x/Al₂O₃ catalyst is a Pd⁺ species.¹⁶ However, the activity and stability of Pd–Cu–Cl_x/Al₂O₃ catalysts are also influenced by a moisture concentration in the feed gas.

In this paper, we carefully investigated the effect of moisture on the activity and stability of Pd–Cu–Cl_x/Al₂O₃ prepared by a CI method for low temperature CO oxidation and a deactivation of the Pd–Cu–Cl_x/Al₂O₃ catalyst in the feed gas containing different concentrations of water and at different reaction temperatures. Based on the investigation for the physical and chemical properties of fresh and deactivated Pd–Cu–Cl_x/Al₂O₃ catalysts, the reasons for the irreversible or reversible deactivation of PC catalysts are discussed.

2. Experimental

2.1. Materials

CuCl₂ (AR), 25% ammonia solution and isopropanol (AR) were from Sinoharm Chemical Reagent Co., Ltd. (SCRC); PdCl₂ (AR) from Heraeus Materials Technology Shanghai Ltd.; Al₂O₃ (WHA-204, BET surface area of 194 m² g⁻¹) from Wenzhou Jingjing aluminum Ltd.

2.2. Catalyst preparation

The supported 1.7wt.%Pd–3.3wt.%Cu–Cl_x/Al₂O₃ catalyst was prepared by a NH₃ coordination-impregnation method (CI) with an isopropanol solvent.¹⁵ Weighed PdCl₂ and CuCl₂ were dissolved in 2 ml 25% ammonia solution under ultrasonic stirring at room temperature, and this mixed aqueous solution of PdCl₂ and CuCl₂ was diluted to 8 ml with isopropanol. Then 1 g of Al₂O₃ was impregnated in this solution. After being aged for 24 h, the catalyst was dried at room temperature and calcined at 300 °C for 4 h. The Pd–Cu–Cl_x/Al₂O₃ catalyst prepared by the CI method is denoted as PC.

2.3. Catalyst characterization

The powder X-ray diffraction patterns (XRD) of the samples were performed on a Brook D8 focus diffraction spectrometer using Cu-Kα radiation at room temperature. The nitrogen adsorption and desorption isotherms were measured at −196 °C on an ASAP 2400 system in static measurement mode. The specific surface area was calculated by the BET model. The elements in a catalyst was analysed on an energy dispersive X-ray spectroscopy (EDS, Falcon, EDAX).

H₂-temperature programmed reduction (H₂-TPR) was performed in a conventional flow system with a quartz U–tube reactor and 100 mg catalyst. The fresh catalyst was firstly pretreated in N₂ flow at 300 °C for 1 h, and then cooled to room temperature; the deactivated catalysts were pretreated in a N₂ flow at room temperature for 1 h. The reduction gas consisted of 5% H₂–N₂ (45 ml min⁻¹). The heating rate was 10 °C min⁻¹. The uptake amount of H₂ was measured by a thermal conductivity detector (TCD), which was calibrated by the quantitative reduction of CuO to the metallic copper.

Diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) of CO adsorbed on the catalyst was measured on a Nicolet Nexus 670 spectrometer equipped with a MCT detector, and the sample cell was fitted with ZnSe windows and a heating chamber which can be heated up to 600 °C. The DRIFTS spectra obtained were saved in a Kubelka-Munk unit with a resolution of 4 cm⁻¹ and 64 scans. The deactivated PC sample was pretreated in a He (99.999%) flow of 50 ml min⁻¹ at room temperature for 1 h and the background spectrum in a He flow was acquired. After gas mixture (1), 0.15% CO + O₂ (0% or 20%) + balanced He with 50 ml min⁻¹, flowed through the sample cell for 30 min, the spectra were taken at 25 °C; ∼0.1% H₂O was added in gas mixture (1) above, and the spectra were taken again at 25 °C.

2.4. Testing of catalytic activity

The activities of catalysts for CO oxidation were measured in a (∅5 mm) quartz U–tube reactor, and 0.2 g of catalyst (20–40 mesh) was used and a glass wool was plugged in both sides of catalyst. The feed gas of 1500 ppm CO in air at a flow rate of 50 ml min⁻¹ was directed through a water vapor saturator immersed in an ice–water bath, and then flowed into the reactor. The water concentration in the feed gas was about 0.1–3.1%. Changing the temperature of water bath can adjust the water concentration in the feed gas. The temperature was controlled as follows: below room temperature was obtained by using an ethanol–liquid nitrogen mixture in a vacuum bottle, and above room temperature was achieved by using a warm water bath.

3. Results and discussion

3.1. Effects of reaction temperature and moisture concentration on the catalytic activity and stability of PC

The catalytic stability of PC for CO oxidation in the presence of ∼0.6% water was studied at different temperatures (Fig. 1A). The CO conversion over PC sharply dropped to 36% at 0 °C after 5 h. At 25 °C, the CO conversion gradually decreased to 77% after 30 h. When the reaction temperature is increased to 60 °C, the CO conversion over PC can remain unchanged for 30 h. It should be noted that PC lost its activity at various rates with different reaction temperatures.

Fig. 1 (A) Catalytic stability of PC for CO oxidation in the feed gas with ∼0.6% moisture at (a) 60 °C, (b) 25 °C, (c) 0 °C, (d) 25 °C (regenerated PC that was deactivated at 25 °C) and (e) 0 °C (regenerated PC that was deactivated at 0 °C). (B) Catalytic stability of PC for CO oxidation at 25 °C (except curve (e)) in the feed gas with a moisture of (a) 0.1%, (b) 0.6%, (c) 3.1%, (d) 3.1% (regenerated PC that was deactivated under 3.1% moisture) and (e) 0.1% at 0 °C. (The feed gas was 1500 ppm CO in air and WHSV was 15 000 ml (g h)⁻¹. The deactivated catalyst was regenerated at 50 °C in air for 10 h).

Fig. 1B shows the effects of moisture concentration on the catalytic stability of PC catalyst for CO oxidation at 25 °C. In the presence of ∼0.1% moisture, the catalytic activity of PC was unchanged for 30 h (Fig. 1B,a), and after continuous reaction at 0 °C for 30 h, its CO conversion was still hardly varied (Fig. 1B,e). The catalytic activity of PC was decreased gradually in 0.6% moisture. In the presence of ∼3.1% moisture, PC lost its activity drastically after 2 h and the CO conversion decreased to ∼50% after 9 h.

For PC deactivated at 25 °C in the feed gas with ∼0.6% moisture, its catalytic activity and stability (Fig. 1A,d) after it was regenerated at 50 °C in air for 10 h, are similar to that of the fresh catalyst (Fig. 1A,b). It has been reported ^13,14 that a relatively high moisture concentration can lead to H₂O capillary condensation in the smaller pores of the hydrophilic alumina support, which can affect the effective contracted area between the reactants and the catalyst surface. Therefore, a reversible deactivation of PC catalyst at 25 °C in the ∼0.6% moisture feed gas is due to the capillary condensation of water inside smaller pores of alumina. However, the regenerated PC catalyst cannot return its original activity after it is used and deactivated at 0 °C in the ∼0.6% water feed or 25 °C in the ∼3.1% water feed gas (Fig. 1A,e and Fig. 1B,d). The phenomena above indicate that an irreversible deactivation of PC may be not resulted from physical capillary condensation. Hence, these deactivated catalysts need to be further characterized in detail, to find out the difference between the deactivated and fresh catalyst.

PC deactivated reversibly at 25 °C in the ∼0.6% water feed is denoted as PC–R(25C/0.6). PC deactivated irreversibly at 0 °C in the ∼0.6% water feed and at 25 °C in the ∼3.1% water feed are denoted as PC–I(0C/0.6) and PC–I(25C/3.1), respectively.

3.2. Surface compositions of PC catalysts before and after being used

It was reported that a deactivation of a PdCl₂–CuCl₂ catalyst is loss of chlorine from the catalyst, which is believed to be essential for the catalytic cycle of Wacker type catalysts for CO oxidation.^11,17 Therefore, an energy dispersive X-ray spectroscopy (EDS) was used to analyze the chlorine amount on the surface of fresh and deactivated PC catalysts, and the results are shown in Table 1.

Table 1 Surface element concentrations of the fresh and the deactivated PC

Element	Fresh PC		PC-R(25C/0.6)		PC-I(0C/0.6)		PC-I(25C/3.1)
	wt.%	mol.%	wt.%	mol.%	wt.%	mol.%	wt.%	mol.%
Cl	3.7	2.5	3.7	2.5	3.5	2.5	3.4	2.1
Pd	2.2	0.5	2.3	0.5	2.2	0.5	1.8	0.4
Cu	7.7	2.8	7.9	2.9	6.6	2.4	4.1	1.5
Cu/Pd (mol)	5.6		5.8		4.8		3.8
BET surface area (m^2 g^−1)	159		155		148		120

---

It can be seen that the chlorine content of PC–R(25C/0.6) is 3.7 wt.%, in accord with that of the fresh catalyst. The chlorine contents of the irreversible deactivated PC–I(0C/0.6) and PC–I(25C/3.1) catalysts are 3.5 and 3.4 wt.%, which were also similar to the chlorine content of the fresh catalyst. It was reported that chlorine loss of PdCl₂–CuCl₂ catalyst occurs at above 60 °C,^11,17 and herein PC was deactivated at relative lower temperatures (as 0 and 25 °C). Therefore, the irreversible deactivation of PC catalyst is not due to the loss of chlorine.

It should be noted that Pd and Cu concentrations on the surface of the fresh catalyst measured by the EDS method (Table 1) are relatively higher than the total concentrations of Pd and Cu by the Inductively Coupled Plasma (ICP) method (Pd, 1.7 wt.% and Cu, 3.3 wt.%). For the PdCl₂–CuCl₂ catalyst, a relatively higher copper content is essential to re-oxidize Pd⁰ to higher valent palladium to maintain the catalytic cycle.^18,19 Hence, the high molar ratio of Cu/Pd on the surface of the fresh PC catalyst (Cu/Pd = 5.6, mol) and PC–R(25C/0.6) (5.8) is the main reason for two catalysts maintaining a high activity. The molar ratio of Cu/Pd on the surface of two irreversible deactivated catalysts (PC–I(0C/0.6) and PC–I(25C/3.1)) are only 4.8 and 3.8 respectively, which are much lower than that of the fresh and PC–R(25C/0.6) catalysts. This may be due to the transfer of the Cu species into the inner part of alumina support or an aggregation of Cu species during the activity testing.

3.3. X-ray diffraction and BET surface area

Fig. 2 exhibits XRD patterns of PC catalysts before and after the activity testing. The diffraction peaks of palladium species cannot be observed over all catalysts, indicating that palladium species are highly dispersed on alumina or below the detection limit of XRD.^18,19 The diffraction peaks of the copper phase also cannot be discerned in the XRD patterns of fresh PC and PC–R(25C/0.6), indicating the copper species are also highly dispersed on the catalyst, like palladium species. However, the diffraction peaks of the Cu₂Cl(OH)₃ phase can be clearly observed in the XRD patterns of PC–I(0C/0.6) and PC–I(25C/3.1). It was reported that Cu₂Cl(OH)₃ is the active copper phase in conventional PdCl₂–CuCl₂ catalysts and Pd–Cu–Cl_x/Al₂O₃ catalysts prepared by the coordination-impregnation method.¹⁵ A weak diffraction peak of CuCl (111) plane is also shown in the XRD pattern of PC–I(25C/3.1), indicating that higher concentration moisture in the feed may also cause the transformation of active Cu species to CuCl species. Hence, the aggregation or transformation of the active copper phase of Cu₂Cl(OH)₃ would cause irreversible deactivation of the PC catalyst. For the PC–R(25C/0.6) catalyst, its reversible deactivation results from physical capillary condensation in the smaller pores of the catalyst rather than the change of the active copper phase. The regenerated PC catalyst that was reactivated at 50 °C in air has been tested by XRD, and the result (not shown here) shows that its XRD pattern is similar to that of fresh PC catalyst.

Fig. 2 XRD patterns of (a) fresh PC, (b) PC–R(25C/0.6), (c) PC–I(0C/0.6) and (d) PC–I(25C/3.1).

The BET surface area of PC–R(25C/0.6) (155 m² g⁻¹) is near to that of fresh PC catalyst (159 m² g⁻¹), as shown in Table 1. However the BET surface areas of PC–I(0C/0.6) (148 m² g⁻¹) and PC–I(25C/3.1) (120 m² g⁻¹) are much smaller than that of the fresh PC catalyst, especially for PC–I(25C/3.1).

3.4. H₂-temperature programmed reduction

The TPR profiles of the fresh and deactivated PC catalysts are shown in Fig. 3. In the TPR profile of the fresh PC (Fig. 3a), there is a sharp peak at 137 °C due to the co-reduction of Pd and Cu species, and a weak reduction peak at 284 °C due to the reduction of isolated copper species.^15,16 This indicates there is a strong interaction between palladium and copper species on the fresh catalyst.

Fig. 3 H₂-TPR of (a) fresh PC, (b) PC–R(25C/0.6), (c) PC–I(0C/0.6) and (d) PC–I(25C/3.1).

In the TPR profile of PC–I(0C/0.6) (Fig. 3c), the reduction peak of Pd species at 75 °C ²⁰ and a reduction peak of the isolated copper species at 284 °C can be observed, while the reduction peak at 137 °C cannot be observed and the reduction peak at 180–213 °C may be from the Cu species partly interacting with Pd species. The TPR profile (Fig. 3d) of PC–I(25C/3.1) also has three reduction peaks, and their peak positions are the same as those of PC–I(0C/0.6), but their intensities are obviously different with those of PC–I(0C/0.6). For the reduction peak at 75 °C, PC–I(0C/0.6) is stronger than PC–I(25C/3.1); for the peak at ∼213 °C, PC-I(0C/0.6) is weaker than PC–I(25C/3.1), in fact, the peak of PC-I(0C/0.6) is a broad peak at 180–213 °C.

These results show that in the deactivated PC–I(0C/0.6) catalyst, an interaction between palladium and copper species is very weak, which leads that the reduction peak at 137 °C shifting to 180–213 °C with a tremendous decrease of its intensity, while the reduction peak of isolated Pd species at 75 °C and the peak of isolated Cu species at 284 °C appear obviously. Compared with the PC–I(0C/0.6) catalyst, an interaction between Pd and Cu species in the PC–I(25C/3.1) catalyst is stronger, so that the reduction peaks at 75 °C and 284 °C are weaker.

Unlike PC–I(0C/0.6) and PC–I(25C/3.1) catalysts, in the TPR profile (Fig. 3b) of reversibly deactivated PC–R(25C/0.6) catalyst there is the reduction peak at 137 °C besides the peaks at 75 °C and >180 °C. Just this characteristic reduction peak at 137 °C is remarkably like fresh PC and unlike the PC–I(0C/0.6) and PC–I(25C/3.1) catalysts. We can conclude that the co-reduction peak of Pd and Cu species in PC catalyst is associated with its catalytic activity for CO oxidation.

3.5. In situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) for surface reaction

In situ DRIFTS technology was employed to test the surface reaction of CO adsorbed on PC–I(0C/0.6) in different atmospheres, and the results are shown in Fig. 4. In these DRIFTS spectra, the peaks at 1600–1700 cm⁻¹ are due to OH groups^14,21 and the peaks at 1570 and 1460 cm⁻¹ are assigned to carbonate species.²² The weak peaks at 1990–2000 cm⁻¹ are related with bridge-bonded CO on metallic palladium (Pd₂–CO). The peaks at 1934 cm⁻¹ should be ascribed to bridged carbonyl ligands in Pd⁺ complexes (Pd⁺–CO).^14,21 The strong peaks at 1822 cm⁻¹ are related to triply bonded CO on metallic Pd (Pd₃–CO).^23,24

Fig. 4 In situ DRIFTS spectra of PC–I(0C/0.6) under (a) 0.15%CO–He, (b) 0.15%CO–20%O₂–He and (c) 0.15%CO–20%O₂–0.1% H₂O–He at 25 °C.

Fig. 4a shows the in situDRIFT spectra of CO adsorption over PC–I(0C/0.6). With an increase in the contact time, the peaks of Pd⁺–CO, Pd₃–CO and OH grow obviously, and the broad peak of the carbonate species (HCOO−) appears gradually after CO adsorption for a long time. Our previous research shows that Pd⁺ is the active palladium species on the PC catalyst.¹⁶ Compared with the DRIFT spectra of CO adsorption over fresh PC catalyst,¹⁶ the intensity of Pd₃–CO peak is much stronger but that of Pd⁺–CO peak is weaker, indicating that the Pd⁰ species is the main palladium species on the surface of PC–I(0C/0.6). Therefore, a change of Pd species on the surface of PC-I(0C/0.6) may be the reason for its deactivation.

Fig. 4b shows the in situDRIFT spectra of CO–O₂ co-adsorbed on PC–I(0C/0.6). After 20% O₂ was added in the gas mixture of 0.15%CO–He, the Pd₃–CO peak decreased drastically but the intensity of the Pd⁺–CO peak was almost invariable. With an increase in the adsorption time, the intensity of Pd₃–CO peak further decreased, which is due to the oxidation of CO adsorbed on Pd⁰ to CO₂ whose peak at 2300–2400 cm⁻¹ increases; and the peaks of the carbonate species also enhanced. Compared with the in situDRIFT spectra of CO–O₂ co-adsorbed on fresh PC catalyst,¹⁶ for the PC–I(0C/0.6) catalyst, even in the presence of oxygen the Pd⁰ species is hard to re-oxidize to Pd⁺.

Fig. 4c shows that when ∼0.1% H₂O was introduced into the mixture of gases 0.15%CO–20%O₂–He, the intensity of the Pd⁺–CO and Pd₂–CO peaks decreased quickly and disappeared after 30 min. However, the intensity of the Pd₃–CO peak decreased a little, and the peaks of the carbonate increased obviously and were invariable after 60 min. These results also indicate that the presence of water is in favor of a formation of carbonate species, and the oxidation of CO adsorbed on Pd⁺,¹⁶ resulting in the disappearance of the peak of Pd⁺–CO.

The strong Pd₃–CO peak and relatively weak Pd⁺–CO peak in the DRIFT spectra of CO adsorbed over PC–I(0C/0.6) catalyst (Fig. 4a) imply that a large amount of the Pd⁺ species were reduced to Pd⁰ and the Pd⁰ species were hard to re-oxidize even in the feed gas including oxygen or oxygen and moisture (Fig. 4b and c). It was known that the rate-determining step of CO oxidation over PC catalyst is the re-oxidation of Pd⁰ by Cu²⁺.¹⁶ Hence, the weak interaction of Pd–Cu can cause the slow rate of Pd⁰ re-oxidation by Cu²⁺. If the close-knit structure of Pd–Cu–Cl on alumina was broken, this PC catalyst would be irreversibly deactivated. Furthermore, the presence of large amounts of water is in favor of a formation of a carbonate species, promoting the deactivation of PC catalyst, for instance, PC–R(25C/0.6) is of reversible deactivation and PC–I(25C/3.1) is of irreversible deactivation (Fig. 1), which is attributed to the high water content resulting in a deactivation of PC catalyst. Therefore, it is concluded that an irreversible deactivation of PC catalyst is a variance in the close-knit structure of Pd–Cu–Cl and formation of carbonate species on the catalyst surface. Nevertheless, the close-knit structure of Pd–Cu–Cl is still unclear, and needs to be investigated in future.

4. Conclusions

A Pd–Cu–Cl_x/Al₂O₃ (PC) catalyst can maintain the activity for CO oxidation in the presence of ∼0.1% moisture even at 0 °C, and is reversibly deactivated at 25 °C with 0.6% moisture; but it would be irreversibly deactivated at 25 °C with 3.1% moisture and at 0 °C with 0.6% moisture. The reversible deactivation of PC catalyst is due to the H₂O capillary condensation in the small pores of the alumina support, and hence it can be regenerated by drying. The characterization results for the fresh and deactivated catalysts show that the co-reduction peak of the Pd and Cu species in PC catalyst is associated with its catalytic activity for CO oxidation, indicating a formation of the close-knit structure of Pd–Cu–Cl in PC catalyst and strong interaction between Cu and Pd species; the irreversible deactivation of the PC catalyst are a variance in the close-knit structure of Pd–Cu–Cl, resulting in an aggregation or transformation of the active copper phase and a formation of carbonate species on the catalyst surface. These reasons above may be further causes why the Pd⁰ species is hard to re-oxidize to the active Pd⁺ species in the reaction conditions. Nevertheless, the close-knit structure of Pd–Cu–Cl is still unclear, and needs to be investigated in future.

Acknowledgements

This project was supported financially by the National Basic Research Program of China (2010CB732300), the Fundamental Research Funds for the Central Universities, and the “Shu Guang” Project (10GG23) of Shanghai Municipal Education Commission and Shanghai Education Development Foundation.

Notes and references

1. H. Q. Zhu, Z. F. Qin, W. J. Shan, W. J. Shen and J. G. Wang, Catal. Today, 2007, 126, 382.
2. J. Y. Luo, M. Meng, X. Li, X. G. Li, Y. Q. Zha, T. D. Hu, Y. N. Xie and J. Zhang, J. Catal., 2008, 254, 310.
3. Y. B. Yu, T. Takei, H. Ohashi, H. He, X. L. Zhang and M. Haruta, J. Catal., 2009, 267, 121.
4. X. W. Xie, Y. Li, Z. Q. Liu, M. Haruta and W. J. Shen, Nature, 2009, 458, 746.
5. F. Grillo, M. M. Natile and A. Glisenti, Appl. Catal., B, 2004, 48, 267.
6. D. H. Cunningham, T. Kobayashi, N. K. Haruta and M. Haruta, Catal. Lett., 1994, 25, 257.
7. M. Date and M. Haruta, J. Catal., 2001, 201, 221.
8. M. Date, M. Okumura, S. Tsubota and M. Haruta, Angew. Chem., Int. Ed., 2004, 43, 2129.
9. S. T. Daniells, M. Makkee and J. A. Moulijn, Catal. Lett., 2005, 100, 39.
10. J. S. Lee, E. D. Park and B. J. Song, Catal. Today, 1999, 54, 57.
11. K. D. Kim, I. S. Nam, J. S. Chung, J. S. Lee, S. G. Ryu and Y. S. Yang, Appl. Catal., B, 1994, 5, 103.
12. E. D. Park and J. S. Lee, J. Catal., 2000, 193, 5.
13. C. W. Lee, S. J. Park, Y. S. Kim and P. J. Chong, Bull. Korean Chem. Soc., 1995, 16, 296.
14. K. I. Choi and M. A. Vannice, J. Catal., 1991, 127, 489.
15. Y. X. Shen, G. Z. Lu, Y. Guo and Y. Q. Wang, Chem. Commun., 2010, 46, 8433.
16. Y. X. Shen, G. Z. Lu, Y. Guo, Y. Q. Wang, Y. L. Guo and X. Q. Gong, Catal. Today, 2011 DOI:10.1016/j.cattod.2011.03.042.
17. M. N. Desai, J. B. Butt and J. S. Dranoff, J. Catal., 1983, 79, 95.
18. E. D. Park, S. H. Choi and J. S. Lee, J. Phys. Chem. B, 2000, 104, 5586.
19. J. S. Lee, S. H. Choi, K. D. Kim and M. Nomura, Appl. Catal., B, 1996, 7, 199.
20. A. B. Gaspar and L. C. Dieguez, Appl. Catal., A, 2000, 201, 241.
21. K. I. Choi and M. A. Vannice, J. Catal., 1991, 127, 465.
22. K. I. Tanaka, M. Shou, H. He, X. Y. Shi and X. L. Zhang, J. Phys. Chem. C, 2009, 113, 12427.
23. S. D. Ebbesen, B. L. Mojet and L. Lefferts, Phys. Chem. Chem. Phys., 2009, 11, 641.
24. G. Rupprechter, H. Unterhalt, M. Morkel, P. Galletto, L. J. Hu and H. J. Freund, Surf. Sci., 2002, 502, 109.

---