NO_x is a major air pollutant and is responsible for environmental problems such as photochemical smog,acid rain,and ozone layer depletion. Exposure to high concentrations of NO_x can cause health problems for humans ^{[1, 2]}. Selective catalytic reduction (SCR) of NO_x,a proven technique,gives higher NO_x control efficiency than do many other NO_x treatment methods ^[3-5]. V₂O₅-WO₃(MoO₃)/TiO₂ is a common commercial SCR catalyst for NO_x removal ^{[2, 3]}. However,there are problems with this catalytic system,e.g.,the toxicity of vanadium species,low N₂ selectivity at high temperatures ^{[6, 7]},and deactivation caused by loading of alkali and alkaline earth metal salts in the ashes ^{[8, 9]}. It is therefore necessary to develop novel environmentally friendly catalysts with high SCR performance. CeTiO_x is a promising catalyst for NH₃-SCR and has high activity and excellent selectivity for N₂ ^[10-13]. Shan et al. ^[12] reported that the NH₃-SCR performance of a CeO₂-TiO₂ catalyst was better than those of V₂O₅-WO₃/TiO₂ and Fe-ZSM-5 catalysts under the same conditions.

Alkali metals poison NH₃-SCR catalysts such as V₂O₅/TiO₂ ^{[14, 15]},V₂O₅-CeO₂/TiO₂ ^[15],and CeTiO_x,especially in biomass combustion plants. Energy generation from biomass contributes to the carbon eco-cycle and helps to maintain the carbon balance on Earth,but there is a high concentration of K species in biomass such as straw and municipal waste ^[16],and this limits the use of CeTiO_x catalysts.

Decreased surface acidity ^[17] and redox reactivity ^{[8, 18]} are two important factors in NH₃-SCR catalyst deactivation. Wang et al. ^[19] proposed that the mechanism of alkali metal poisoning of CeTiO_x catalysts involved enlargement of the CeO₂ nanoparticles,a decreased Ce⁴⁺/Ce³⁺ redox cycle rate,and changes in the surface acidity,which could be major factors in the decline of the SCR activity of the Ce/TiO₂ catalyst after loading with Na⁺ or Ca²⁺ ions. Du et. al. ^[16] conducted a theoretical study of the effects of K⁺ or Na⁺ doping on the NH₃-SCR based on density functional theory calculations. The results showed that the alkali atoms interacted strongly with the oxygens on Ce and Ti,decreasing the reducibility and surface acidity,respectively.

ZrO₂ is an acid-base amphoteric oxide with good redox properties,and it is widely used in catalysis as an inhibitor of crystal transformations and as a cocatalyst ^[20]. The addition of Zr can improve the stability and redox properties of Ce-based oxygen-storage materials,i.e.,three-way catalysis. The addition of Zr to CeTiO_x can improve resistance to alkali metal poisoning,and therefore increase the surface acidity and improve the redox properties.

In this work,CeTiO_x and CeZrTiO_x catalysts were prepared by a coprecipitation method,and alkali metal poisoning over both catalysts was investigated by catalyst impregnation with various contents of K⁺ to determine the effects of Zr addition to CeTiO_x on the K⁺-poisoning resistance. The Brunauer-Emmett-Teller (BET) method,X-ray diffraction (XRD),Raman spectroscopy,transmission electron microscopy (TEM),X-ray photoelectron spectroscopy (XPS) and temperature-programmed desorption of NH₃ (NH₃-TPD) were used to characterize the fresh and poisoned catalysts.

2 Experimental

2.1 Catalyst preparation

CeTiO_x and CeZrTiO_x catalysts were prepared using a conventional coprecipitation method. Ce(NO₃)₃ (AR grade,Jinshan,Chengdu,China),zirconyl nitrate hydrate (AR grade,99%,Yutai,Shangdong,China),and titanylsulfate (AR grade,Dandong,Liaoning,China) were used as precursors. The precursors were mixed at appropriate molar ratios in an aqueous solution and then precipitated with an alkaline buffer solution until the pH reached approximately 9. The obtained precipitates were dried at 90 °C for 24 h,and then calcined in air at 550 °C for 3 h. The obtained CeTiO_x and CeZrTiO_x powders are denoted by CT and ZCT. For both catalysts,the Ce/Ti molar ratio was kept at 0.2:1,and the Zr/Ti molar ratio in CeZrTiO_x was also 0.2:1.

CT and ZCT catalysts were impregnated with KNO₃ at K/Ce molar ratios of 0.1:1 and 0.2:1,and calcined in air at 550 °C for 3 h. The K⁺-poisoned samples are denoted by K0.1-CT,K0.2-CT,K0.1-ZCT,and K0.2-ZCT.

2.2 Catalyst characterization

The surface areas,pore sizes,and pore volumes of the catalysts were determined using the BET method based on N₂ adsorption at -196 °C (Quantachrome automated surface area and pore size analyzer,Autosorb SI). The samples were treated at 300 °C for 3 h prior to measurements.

Powder XRD was performed using a Rigaku D/max-RA diffractometer with Cu K_α (λ = 0.15406 nm) radiation. The tube voltage and current were 40 kV and 100 mA,respectively. The powder XRD patterns were recorded at 0.03°/s intervals in the range 20°-80°. The crystalline phases were identified by reference to the database of the International Center for Diffraction Data (ICDD).

Raman spectroscopy was performed using a Lab-RAM HR laser Raman instrument with a Nd:Yag laser at a 532 nm excitation wavelength. All specimens were illuminated through a 50 × objective and in powder form to prevent diffusion problems. Raman spectra were recorded over the range 100-1000 cm^-1. The grating specification was 600 g/mm.

TEM was performed using a JEM-100CX instrument at an accelerating voltage of 80 kV.

XPS was performed with a British Kratos XSAM-800 instrument at 13 kV and 20 mA electron current,using Mg K_α radiation. Binding energies were calibrated using the C 1s peak (284.8 eV). The pressure in the analytical chamber was about 1× 10^-9 Pa.

NH₃-TPD was performed in a fixed-bed quartz reactor. A sample mass of 100 mg and a gas flow rate of 30 mL/min were used. The experiment included four stages: (1) degasification of the sample in Ar at 450 °C for 1 h,(2) adsorption of 2% NH₃ at 80 °C for 1 h,(3) isothermal desorption in Ar at 80 °C until no NH₃ was detected,and (4) TPD in Ar at 8 °C/min to 600 °C. A thermal conductivity detector was used.

2.3 Catalytic activity tests

Catalytic activity measurements were performed in a fixed-bed quartz flow reactor. The reactor consisted of a quartz tube of diameter 8 mm filled with 0.25 g of catalyst. Temperatures were measured using a thermocouple placed at the center of the bed. The concentrations of the simulated gases were 0.05% NO,0.05% NH₃,5% O₂,and balance N₂. The gas hourly space velocity (GHSV) was about 60000 h^-1. The original and effluent NO concentrations at different temperatures were continuously monitored using an IR detector (ANTARIS IGS analyzer,Thermo Scientific). The data were collected after the reaction had stabilized for 30 min at each temperature. The NO_x conversions were calculated as follows: NO_x conversion = ([NO_x]_inlet - [NO_x]_outlet)/[NO_x]_inlet × 100%.

3 Results and discussion

3.1 SCR performance

The activities of CT and ZCT with different K⁺ loadings were tested to compare their anti-alkali metal abilities; the results are shown in Fig. 1(a). The catalysts without K⁺ gave 90% NO_x conversions over a wide temperature range from 250 to 450 °C; CT performed slightly better than ZCT. Impregnation with K⁺ decreased the NO_x removal activities of CT and ZCT. NO_x conversion declined further when the K/Ce molar ratio was increased to 0.2. The poisoning effect was so distinct that K0.2-CT had nearly no activity or even gave negative NO_x conversions at high temperatures because of non-selective NH₃ oxidation.

Fig. 1

Fig. 1. (a) NO conversion as a function of temperature; (b) NO conversion at 350 °C as a function of K/Ce molar ratio. Reaction conditions: [NO] = [NH3] = 0.05%,[O2] = 5%,balance N2,GHSV = 60000 h-1.

The mechanism of the decrease in activity with increasing K⁺ loading was clarified by investigating NO_x conversion at 350 °C as a function of the K/Ce molar ratio; the results are shown in Fig. 1(b). CT and ZCT achieved nearly 100% NO_x conversions,and there was little difference between the performance of the two catalysts. ZCT gave better NO_x conversion than the catalyst without Zr addition after impregnation with K⁺. The NO_x conversions by K0.1-CT and K0.1-ZCT were 90% and 62%,respectively; for K0.2-CT and K0.2-ZCT,the conversions were 48% and 13%,respectively. The catalytic activity of ZCT therefore decreased more slowly than that of CT with increasing K⁺ loading.

The NO_x removal activity of ZCT remained higher than that of CT after K⁺ poisoning,indicating that Zr addition improved the resistance of CT to K⁺ poisoning.

3.2 Characterization

3.2.1 BET analysis

The textural properties of the two catalysts before and after poisoning are shown in Table 1. Previous studies have shown that a high surface area and large pore volume promote effective mass transfer by providing more active sites for adsorption of reactant molecules and intermediates,consequently improving the catalytic performance ^[21]. The specific surface area of CT was 101.7 m²/g and the pore volume was 0.34 mL/g. After poisoning by K⁺,the specific surface areas dropped to 67.9 and 64.6 m²/g and the pore volume also decreased slightly. Similar changes were observed for ZCT. The specific surface areas of ZCT,K0.1-ZCT,and K0.2-ZCT were 142.6,88.8,and 80.4 m²/g,respectively. The specific surface areas and pore volumes of the K⁺-poisoned catalysts decreased significantly compared with those of the fresh catalyst,showing that K⁺ damaged the textural properties of CT and ZCT. The sharp decrease in the specific surface area was in line with the significant decline in NH₃-SCR activities.

表 1 (Table 1)

Table 1 Textural properties of catalysts. Sample Specific surface area (m2/g) Pore volume (mL/g) Average pore radius (nm) CT 101.7 0.34 6.6 K0.1-CT 67.9 0.33 9.7 K0.2-CT 64.6 0.30 9.3 ZCT 142.6 0.42 5.9 K0.1-ZCT 88.8 0.36 8.0 K0.2-ZCT 80.4 0.34 8.3

Table 1 Textural properties of catalysts.

The specific surface area and pore volume of ZCT were higher than those of the CT catalyst without K⁺ loading. After poisoning by K⁺,ZCT still showed better textural properties. It is therefore concluded that Zr addition improved the textural properties of the CT catalyst even after poisoning by K⁺; this might be why the catalytic activity of ZCT was better than that of CT.

3.2.2 XRD

The effects of Zr addition on the structural properties of the fresh and poisoned catalysts were clarified by examining their crystal structures using powder XRD. The crystalline phases were identified by comparison with ICDD files. The XRD patterns of the CT catalysts are shown in Fig. 2(a). Only anatase TiO₂ was detected and no diffraction peaks from CeO₂ were observed for any of the catalysts. This indicates that Ce was highly dispersed or amorphous. The diffraction peaks of anatase TiO₂ grew sharper with increasing K⁺ loading,showing that amorphous TiO₂ was gradually transformed into the crystalline phase. Crystallite growth decreases the surface area,as shown by the data in Table 1.

Fig. 2

Fig. 2. XRD patterns for CT (a) and ZCT (b) catalysts.

Fig. 2(b) shows that K⁺ poisoning did not significantly change the ZCT phase; this indicates that Zr prevented the formation of crystalline TiO₂ and inhibited crystal nucleus growth after doping with K⁺. These results show that Zr improved the structural stability of CT,even when it was poisoned by K⁺,and this enabled ZCT to retain a higher NO_x conversion ability compare with that of CT.

3.2.3 Raman spectroscopy

Raman spectroscopy was used to obtain information on the surfaces of the CT and ZCT catalysts with different K⁺ contents. Fig. 3(a) shows the Raman spectra from 100 to 1000 cm^-1 for the CT catalyst. The peaks at 147 and 642 cm^-1 are attributed to anatase TiO₂ ^[22]; these peaks gradually became more intense and sharper with increasing K⁺ loading; this is in accordance with the XRD results. No Raman bands were observed for the fresh and poisoned ZCT catalysts (Fig. 3(b)),showing that the catalysts were amorphous. The Raman results confirmed that Zr addition stabilized the structure of the CT catalyst.

Fig. 3

Fig. 3. Raman spectra of CT (a) and ZCT (b) catalysts.

3.2.4 TEM

The morphologies,structures,and particle sizes of the catalysts were further examined using TEM. The images in Fig. 4 were obtained at ×140000 magnification. The CT particle size increased with increasing K⁺ content; this is in accordance with the XRD results. The average grain sizes of CT,K0.1-CT,and K0.2-CT were about 7,13,and 15 nm,respectively. In contrast,the ZCT crystal size did not change. The average particle sizes of ZCT,K0.1-ZCT,and K0.2-ZCT were about 5,8,and 10 nm,respectively. This clearly shows that Zr addition inhibited particle aggregation and therefore enhanced the structural stability of CT.

Fig. 4

Fig. 4. TEM images of the catalysts.

The XRD,Raman,and TEM results indicate that the size of the crystalline anatase TiO₂ particles increased after K⁺ loading. The changes are mainly ascribed to the molten salt flux effect ^[19]. Under real flue gas conditions or in a simulated loading-calcination process,the loaded KNO₃ forms a molten salt flux on the surface of the CT catalyst and covers the TiO₂ nanoparticles. Under these conditions,the crystal size of the TiO₂ nanoparticles increases and their crystallinity improves ^[19].

3.2.5 XPS

The Ce 3d XPS spectra of the fresh and K⁺-poisoned catalysts are shown in Fig. 5. The XPS peaks labeled u₁ and v₁ represent Ce³⁺,and those labeled u,u₂,u₃,v,v₂,and v₃ are assigned to Ce⁴⁺.

Fig. 5

Fig. 5. Ce 3d XPS spectra of CT (a) and ZCT (b) catalysts.

There is an oxidation-reduction cycle between Ce³⁺ and Ce⁴⁺. Most of the Ce in the catalyst was in the Ce⁴⁺ state ^[16]. A larger amount of Ce³⁺ gives a catalyst with higher oxygen-storage capacity and better redox properties ^[16]. Ce³⁺ on the catalyst surface is crucial in the SCR reaction because Ce³⁺ can create a charge imbalance,resulting in the formation of more vacancies and unsaturated chemical bonds on the catalyst surface,leading to an increased amount of chemisorbed oxygen on the catalyst surface ^{[23, 24]}. The Ce³⁺/Ce ratio therefore reflects the catalyst redox properties or catalytic activity to some extent. The Ce³⁺/Ce ratio was calculated from the total area of u₁ and v₁,which are attributed to Ce³⁺,divided by the total area of the ten peaks assigned to Ce. The Ce³⁺/Ce ratios of various catalysts are listed in Table 2. The Ce³⁺/Ce ratio was always less than 0.5,indicating that Ce⁴⁺ was the main Ce valence state in the two catalysts.

表 2 (Table 2)

Table 2 Ce3+/Ce ratios of the catalysts. Sample Ratio of Ce3+/Ce (%) K/Ce = 0 K/Ce = 0.1 K/Ce = 0.2 CT 31 26 17 ZCT 30 29 27

Table 2 Ce3+/Ce ratios of the catalysts.

The Ce³⁺/Ce ratios for CT,K0.1-CT,and K0.2-CT were 31%,26%,and 17%,respectively,i.e.,they decreased with increasing K⁺ content. These changes show that K⁺ significantly affected Ce oxidation. This is attributed to the interactions between K⁺ and Ce species inhibiting Ce⁴⁺ reduction to Ce³⁺,resulting in a sharp decrease in the Ce³⁺ concentration ^[19]. Because Ce³⁺ is formed from CeO₂ defects,the amount of Ce³⁺ decreased with increasing particle size ^{[25, 26]}. This is in accordance with the XRD and Raman results.

For the ZCT,K0.1-ZCT and K0.2-ZCT catalysts,the Ce³⁺/Ce values were 30%,29%,and 27%,separately. There was no obvious decrease in the amount of Ce³⁺ with increasing K⁺ impregnation of the ZCT catalyst,showing that Zr prevented the transformation of Ce³⁺ to Ce⁴⁺. For the fresh catalyst,the Ce³⁺/Ce values for CT and ZCT were nearly the same. The proportions of Ce³⁺ for the two catalysts both gradually decreased with increasing K⁺ content,and Ce³⁺/Ce ratios for the ZCT catalysts were always clearly higher than those for the CT catalysts. This is why the K⁺-poisoned ZCT performed better than the K⁺-poisoned CT catalysts.

The XRD and XPS results show that K⁺ impregnated on the CT surface can lead to the crystallization of CeO₂ nanoparticles and a reduction in the amounts of CeO₂ defects and Ce³⁺. These combined results suggest that the Ce³⁺/Ce values for the ZCT catalysts were higher than those for the CT catalysts after poisoning,therefore,the ZCT catalysts had better oxygen-storage capacities and redox properties,and more surface defects; these factors cocontributed to their higher NH₃-SCR activities.

3.2.6 NH₃-TPD

According to the mechanism proposed by Topse et al. ^[27],NH₃ adsorption on the catalyst surface is crucial for the SCR reaction. The acid sites,on which NH₃ is adsorbed,are therefore a key factor in the catalytic performance. Previous reports focusing on alkali metal poisoning of catalysts for NO_x removal attributed deactivation to a decrease in the number of acid sites ^{[28, 29]},which would result in low NH₃ adsorption.

Fig. 6 shows the NH₃-TPD curves of the fresh and K⁺-containing catalysts in the range 120-600 °C. The peak areas for the catalysts are listed in Table 3. On K⁺ addition,these peak areas decreased significantly. The calculated peak areas for CT,K0.1-CT,and K0.2-CT were 1071,104,and 54,respectively,i.e.,poisoning strongly affected NH₃ adsorption. The sharp decrease in surface acidity was in keeping with the decreased NH₃-SCR activity. This is also in accordance with the conclusion that decreased surface acidity was one of the main reasons for the decreased activity after doping with K^{+ [9, 30]}.

Fig. 6

Fig. 6. NH3-TPD profiles of catalysts.

表 3 (Table 3)

Table 3 Peak areas in NH3-TPD profiles of the catalysts. Sample Peak area K/Ce = 0 K/Ce = 0.1 K/Ce = 0.2 CT 1071 104 54 ZCT 854 320 168

Table 3 Peak areas in NH3-TPD profiles of the catalysts.

The peak areas for ZCT,K0.1-ZCT,and K0.2-ZCT were 854,320,and 168,respectively.After Zr addition,the catalyst acidity was better retained after poisoning by K⁺,showing thatthe decrease in the surface acidity was inhibited by Zr doping. The ZCT catalyst therefore gave higher NO_x conversion than did the CT catalyst after K⁺ poisoning.

4 Conclusions

Catalytic activity results showed that a CeTiO_x catalyst was seriously poisoned by KNO₃; but it performed distinctly better by Zr doping. The BET data showed that Zr improved the textural properties of the CT catalyst. The XRD,Raman,and TEM results showed that crystal grain growth caused by the molten salt flux effect was inhibited by Zr addition,suggesting that the structural properties of the CT catalysts were stabilized by the addition of Zr. XPS results showed that K⁺ loading led to the transformation of Ce³⁺ to Ce⁴⁺,resulting in decreased concentrations of oxygen vacancies and Ce³⁺ in CeO₂ particles. The larger amounts of Ce³⁺ and surface defects on the ZCT catalysts than on the CT catalysts suggest that the redox properties of the CT catalysts were stabilized by Zr addition. The acidity of the catalyst with added Zr was better retained after K⁺ poisoning,showing that destruction of the surface acidity was inhibited by Zr doping.

References

[1]	K. Skalska, J. S. Miller, S. Ledakowicz, Sci. Total Environ.,2010, 408 :3976–3989. doi:http://dx.doi.org/10.1016/j.scitotenv.2010.06.001
[2]	G. Busca, L. Lietti, G. Ramis, F. Berti, Appl. Catal. B,1998, 18 :1–36.
[3]	X. X. Cheng, X. T. Bi, Particuology,2014, 16 :1–18. doi:http://dx.doi.org/10.1016/j.partic.2014.01.006
[4]	Z. M. Liu, S. I. Woo, Catal. Rev.-Sci. Eng,2006, 48 :43–89. doi:http://dx.doi.org/10.1080/01614940500439891
[5]	B. Guan, R. Zhan, H. Lin, Z. Huang, Appl. Therm. Eng.,2014, 66 :395–414. doi:http://dx.doi.org/10.1016/j.applthermaleng.2014.02.021
[6]	S. Roy, M. S. Hegde, G. Madras, Appl. Energy,2009, 86 :2283–2297. doi:http://dx.doi.org/10.1016/j.apenergy.2009.03.022
[7]	L. Chen, Z. C. Si, X. D. Wu, D. Weng, R. Ran, J. Yu, J. Rare Earth.,2014, 32 :907–917. doi:http://dx.doi.org/10.1016/S1002-0721(14)60162-9
[8]	D. Nicosia, I. Czekaj, O. Kröcher, Appl. Catal. B,2008, 77 :228–236. doi:http://dx.doi.org/10.1016/j.apcatb.2007.07.032
[9]	F. S. Tang, B. L. Xu, H. H. Shi, J. H. Qiu, Y. N. Fan, Appl. Catal. B,2010, 94 :71–76. doi:http://dx.doi.org/10.1016/j.apcatb.2009.10.022
[10]	W. Q. Xu, Y. B. Yu, C. B. Zhang, H. He, Catal. Commun.,2008, 9 :1453–1457. doi:http://dx.doi.org/10.1016/j.catcom.2007.12.012
[11]	X. Gao, Y. Jiang, Y. Zhong, Z. Y. Luo, K. F. Cen, J. Hazard. Mater.,2010, 174 :734–739. doi:http://dx.doi.org/10.1016/j.jhazmat.2009.09.112
[12]	W. P. Shan, F. D. Liu, H. He, X. Y. Shi, C. B. Zhang, Catal. Today,2012, 184 :160–165. doi:http://dx.doi.org/10.1016/j.cattod.2011.11.013
[13]	W. P. Shan, F. D. Liu, Y. B. Yu, H. He, Chin. J. Catal.,,2014, 35 :1251–1259. doi:http://dx.doi.org/10.1016/S1872-2067(14)60155-8
[14]	Q. C. Li, S. F. Chen, Z. Y. Liu, Q. Y. Liu, Appl. Catal. B,2015, 164 :475–482. doi:http://dx.doi.org/10.1016/j.apcatb.2014.09.036
[15]	Z. M. Liu, S. X. Zhang, J. H. Li, J. Z. Zhu, L. L. Ma, Appl. Catal. B,2014, 158-159 :11–19. doi:http://dx.doi.org/10.1016/j.apcatb.2014.03.049
[16]	X. S. Du, X. Gao, R. Y. Qu, P. D. Ji, Z. Y. Luo, K. F. Cen, ChemCatChem,2012, 4 :2075–2081. doi:http://dx.doi.org/10.1002/cctc.201200316
[17]	L. Chen, J. H. Li, M. F. Ge, Chem. Eng. J.,2011, 170 :531–537. doi:http://dx.doi.org/10.1016/j.cej.2010.11.020
[18]	J. Due-Hansen, S. Boghosian, A. Kustov, P. Fristrup, G. Tsilomelekis, K. Stahl, C. H. Christensen, R. Fehrmann, J. Catal.,2007, 251 :459–473. doi:http://dx.doi.org/10.1016/j.jcat.2007.07.016
[19]	H. Q. Wang, X. B. Chen, S. Gao, Z. B. Wu, Y. Liu, X. L. Weng, Catal. Sci. Technol.,2013, 3 :715–722. doi:http://dx.doi.org/10.1039/C2CY20568H
[20]	S. Verdier, E. Rohart, H. Bradshaw, D. Harris, Ph. BIchon, G. Delahay, SAE Technical Paper, 2008-01-1022, 2008.
[21]	M. Machida, D. Kurogi, T. Kijima, Chem. Mater.,2000, 12 :3165–3170. doi:http://dx.doi.org/10.1021/cm000219c
[22]	Z. M. Liu, Y. Yi, J. H. Li, S. I. Woo, B. Y. Wang, X. Z. Cao, Z. X. Li, Chem. Commun.,2013, 49 :7726–7728. doi:http://dx.doi.org/10.1039/c3cc43041c
[23]	Y. Jiang, Z. M. Xing, X. C. Wang, S. B. Huang, X. W. Wang, Q. Y. Liu, Fuel,2015, 151 :124–129. doi:http://dx.doi.org/10.1016/j.fuel.2015.01.061
[24]	S. X. Yang, W. P. Zhu, Z. P. Jiang, Z. X. Chen, J. B. Wang, Appl. Surf. Sci.,2006, 252 :8499–8505. doi:http://dx.doi.org/10.1016/j.apsusc.2005.11.067
[25]	P. Dutta, S. Pal, M. S. Seehra, Y. Shi, E. M. Eyring, R. D. Ernst, Chem. Mater.,2006, 18 :5144–5146. doi:http://dx.doi.org/10.1021/cm061580n
[26]	X. W. Liu, K. B. Zhou, L. Wang, B. Y. Wang, Y. D. Li, J. Am. Chem. Soc.,2009, 131 :3140–3141. doi:http://dx.doi.org/10.1021/ja808433d
[27]	N. Y. Topsoe, Science,1994, 265 :1217–1219. doi:http://dx.doi.org/10.1126/science.265.5176.1217
[28]	H. Kamata, K. Takahashi, C. U. I. Odenbrand, J. Mol. Catal. A,1999, 139 :189–198. doi:http://dx.doi.org/10.1016/S1381-1169(98)00177-0
[29]	M. Klimczak, P. Kern, T. Heinzelmann, M. Lucas, P. Claus, Appl. Catal. B,2010, 95 :39–47. doi:http://dx.doi.org/10.1016/j.apcatb.2009.12.007
[30]	R. T. Guo, Q. S. Wang, W. G. Pan, W. L. Zhen, Q. L. Chen, H. L. Ding, N. Z. Yang, C. Z. Lu, Appl. Surf. Sci.,2014, 317 :111–116. doi:http://dx.doi.org/10.1016/j.apsusc.2014.08.082