EFFECT OF PREPARATION METHODS OF CeO₂-MnO_x MIXED OXIDES ON PREFERENTIAL OXIDATION OF CO IN H₂-RICH GASES OVER CuO-BASED CATALYSTS

LEI GONG^1,2, LAI-TAO LUO ¹, RUI WANG¹, NING ZHANG¹*

¹ Department of Chemistry, Nanchang University, Nanchang, 330031, P. R. China . *e-mail: nzhang.ncu@163.com
²Department of Chemistry, Jiangxi Agricultural University, Nanchang, 330045, P. R. China

ABSTRACT

Keywords: CuO/CeO₂-MnO_x catalyst, Preparation method of CeO₂-MnO_x, CO oxidation, In H₂-rich gases.

1.       INTRODUCTION

In recent years, polymer electrolyte membrane fuel cell (PEMFC) which utilizes hydrogen as a fuel has been attracting much attention in the applications to electric vehicles or residential power-generations due to its low operation temperature, excellent energy efficiency and zero-emission of air pollutants. Hydrogen for PEMFC is generally generated from steam reforming or partial oxidation of hydrocarbons or methanol followed by the water-gas shift reaction. However, typical hydrogen mixtures from such a process usually contain 0.5-2% CO which can poison the Pt electrode in the PEMFC. Thus, carbon monoxide needs be decreased to a trace-level (below 10 ppm) to avoid poisoning of the Pt electrode. Among the current available methods to remove CO in the H₂-rich gases, preferential oxidation of CO in H₂-rich gases (PROX) was proved to be the most straightforward and economic one ^1-4. So far, extensive studies have focused on improving the activity of the catalysts for CO-PROX reaction. However, developing efficient catalysts for PROX system is still a challenge.

Catalysts for CO PROX are mainly the precious metals catalysts (such as Pt, Ru and Rh), gold-based catalysts and CuO-based catalysts ^1,5. Ofthese catalysts, catalysts based on combinations between copper and cerium oxides have also shown promising properties for CO PROX and constitute a more interesting alternative from an economical point of view. Extensive investigations have been carried out about the preparation methods, characterization, mechanism and testing of CuO/CeO₂ catalysts ^3-8. Some researchers reported the performance of CuO/CeO₂ catalysts was superior to that of Pt-based catalysts because of strong interactions between the copper species and the CeO₂ support ^6,7. However, most CuO/CeO₂ catalysts showed the narrow temperature window with high CO conversions and its catalytic activity strongly depended on the preparation methods ¹. Generally, the mechanism for CO PROX on CuO/CeO₂ can be regarded as a redox reaction process involving lattice oxygen and lattice vacancy, and thus metals modification is important to improve the activity of CuO/CeO₂ catalysts.

Previous investigations in CuO/(CeO, M) catalysts field have revealed that changing the chemical composition of support with ceria-related mixed oxides (Ce-M) instead of pure CeO₂ could greatly improve the catalyst activity, which has mainly been attributed to the enhanced oxygen mobility of CeO₂-MO_x and the synergetic effect through the interaction between the mixed oxides ⁸. CeO₂-MnO_x mixed oxides, as the two-way support, could exhibit much better redox properties than ceria alone due to multi-valence state and strong interaction between ceria and manganese. However, reports on CO-PROX over CeO₂-MnO_x oxides support are still less.

It is well known that catalytic activity strongly depended on not only chemical composition of catalyst, but also the preparation methods. However, to our knowledge, the effect of CeO₂-MnO_x support prepared with different methods on the activity of the CuO-based catalysts for CO-PROX has not been studied. In this work, two CeO₂-MnO_x supports were prepared by deposition-precipitation and surfactant-templated methods, respectively, and the effect of preparation methods on the performance was studied. Characterizations of the CeO₂-MnO_x or CuO/CeO₂-MnO_x were performed with N₂ adsorption, X-ray diffraction (XRD), H₂ temperature- programmed reduction (TPR), CO temperature-programmed desorption (TPD), X-ray photoelectron spectra (XPS). The results of this physicochemical characterization were discussed in relation to the exhibited catalytic performance of the CuO/CeO₂-MnO_x catalysts.

2.       EXPERIMENTAL

2.1.    Catalysts preparation

Pure CeO₂ and MnO_x were prepared by precipitation method. The appropriate quantities of ammonia solution (5%) were added dropwise into the ammonium cerium(IV) nitrate solution (0.1 mol/L) and manganese(II) nitrate solution (0.1 mol/L), respectively, with vigorous stirring until pH of solution remained approximately at 10.0. The mixture was stirred for 4 h, filtered and washed with deionized water. The obtained sample was dried at 100 °C overnight and then calcined at 500 °C for 4 h under air.

CeO₂-MnO_x (DP) was prepared by deposition-precipitation method. A certain amount of ceria powder was added into the solution of 50% Mn(NO₃)₂ (Mn/(Mn + Ce) = 0.3, molar ratio) under vigorous stirring. Then 5% ammonia solution was added dropwise into the suspension until pH of the suspension solution was maintained at 10.0. The other steps were similar to those for the preparation of CeO₂.

CeO₂-MnO_x (CB) was prepared by surfactant-templated method. A certain amount of cetyltrimethyl ammonium bromide (CTAB, 1 mmol) as template was dissolved in 500 mL of distilled water, followed by the addition of (NH₄)₂Ce(NO₃)₆ and 50% Mn(NO₃)₂ solution (Mn/(Mn + Ce) = 0.3, molar ratio) with vigorous stirring. The other steps were similar to those for the preparation of CeO₂.

2.1.2.    Preparation of CuO-based catalyst

The CuO-based catalysts were prepared by impregnation method. The as-synthesized supports were impregnated with an ethanol solution of copper nitrate for 24 h, followed by drying (100 °C; 12 h) and calcining (400 °C; 4 h). The CuO loading for CuO-based catalysts was 8.0 wt %.

2.2.     Catalytic activity measurements

The CO oxidation reaction in H₂-rich gas was carried out in a continuous micro-reactor. The catalyst (0.1 g) was used. The composition of the mixture gas was 1.9 vol. % CO, 3.3 vol.% O₂, 50 vol.% H₂ and 44.8 vol.% N₂, and the space velocity was 60,000 ml/h-g. The gas composition was analyzed before and after the reaction by on-line gas chromatographs with thermal conductor detector (TCD) and carbon molecular sieve (TDX-01) column. The catalytic activity is expressed as conversion of CO. After the first measurement from 60 °C to 200 °C, the catalysts were kept at 200 °C for 10 h and decreased to room temperature, and then the second measurement was carried out from 60 °C to 200 °C to test the stabilities of the catalysts.

2.3.    Characterization of the catalysts

The BET surface area and porous texture were evaluated by N₂ adsorption isotherms obtained at -196 °C using ASAP 2020 (Micrometrics) equipment. Before each measurement, the samples were degassed at 350 °C in vacuum (0.13 Pa) for 5 h. The surface area and average pore size of the samples were calculated with the BET equation and BJH formula, respectively.

H₂ temperature-programmed reduction (H₂-TPR) measurements were carried out on a Chemisorb 2750 instrument (Micrometrics). A 0.01 g sample was heated from room temperature to 600 °C under a He flow (50 ml/min) at a rate of 10 °C/min in order to remove possible impurities. After cooling to room temperature in He, a gas mixture consisting of H₂ and N₂ (10:90 v/v) was introduced into the system, heated to 600 °C at a rate of 10 °C /min for recording the TPR spectra.

CO temperature-programmed desorption (TPD) measurement was carried out on the same instrument with the H₂-TPR measurement. A 0.01 g sample was heated from room temperature to 400 °C under a He flow (50 ml/min) at a rate of 10 °C/min. After being cooled to room temperature, CO was injected into the reactor until adsorption was saturated. Then the sample was heated to 400 °C at a rate of 10 °C/min in He flow (50 ml/min) for recording the CO-TPD spectra.

X-ray photoelectron spectra (XPS) were determined using a KROTAS AXIS Ultra DLD spectrometer with Al Ka radiation (hv = 1486.6 ev). The spectra of O 1s and Mn 2P_2/3 levels were recorded. Charging effects were corrected by adjusting the binding energy of C 1s peak to 284.6 ev.

3.       RESULTS AND DISCUSSION

3.1.      The structural studies of CeO₂-MnO_x supports

The N₂-physisorption isotherms and pore size distributions of CeO₂ and CeO₂-MnO_x supports are shown in Fig. 1. Pure CeO₂ exhibits a typical type-IV isotherm with a pronounced hysteresis loop characteristics which always is connected with pore structure, and CeO₂-MnO_x (DP) exhibits a similar isotherm indicating that the pore structure is maintained. The isotherm of CeO₂-MnO_x (CB) shows a obvious difference as compared with them. The change in the shape of isotherms is probably because the Mn^x+ ions enter ceria bulk lattice resulting in the destruction of pores. The effect of preparation method of CeO₂-MnO_x supports on the surface area (S_BET), pore volume (V_p), the mean pore diameter (r_p), crystallite size and cell parameter of CeO₂-MnO_x are shown in Table 1. The BET surface area of CeO₂-MnO_x (DP) is lower than that of CeO₂, probably resulting from a part ceria being covered with MnO_x during the precipitation processes. Comparing with CeO₂, CeO₂-MnO_x (CB) has larger BET surface area and smaller crystallite size, which might be due to that the CTAB template can effectively inhibit the crystal growth of CeO₂ during precipitation and calcination processes. High surface area and small crystallite size are favorable for the redox properties of the mixed oxides ⁹.

]]>

XRD patterns of the CeO₂, MnO_x and CeO₂-MnO_x mixed oxides are shown in Fig. 1. The peaks observed at 2θ = 28.5°, 33.0°, 47.4° and 56.4° were assigned to the diffraction patterns of CeO₂ (111), (200), (220) and (311) planes, respectively. For pure MnO_x, the intensive and sharp diffractions peaks could be primarily attributed to Mn₃O₄ (PDF#24-0734). Weak diffraction peaks of MnO_x (2θ = 37.1°) is found in CeO₂-MnO_x (DP), indicating the presence of separated MnO_x, but no diffraction of MnO_x is observed in CeO₂-MnO_x (CB). Compared with pure CeO₂, the diffraction peaks of CeO₂ in CeO₂-MnO_x (CB) slightly shift to higher values of the Bragg angles and the cell parameter of that decrease obviously, which indicates the formation of solid solution of CeO₂ and MnO_x ¹⁰. It leads to the increase of surface oxygen vacancies, as confirmed by O₂-TPD ¹¹. And the defect surface of solid solution can enhance the chemisorption of oxygen and accelerate the mobility of surface oxygen species ^{10, 12}.

3.2      The reduction performance of supports and catalysts

H₂-TPR can provide information concerning the reducibility of different chemical species presented in the catalyst as well as the degree of interaction between metal-support. The H₂-TPR profiles of CeO₂, MnO_x, CuO/CeO₂, CeO₂→ MnO_x and CuO/CeO₂→ MnO_x are shown in Fig. 3 and Table 2 lists the reduction temperature and hydrogen consumption of all peaks. Pure CeO₂ exhibits two reduction peaks at 380 and 780 °C, which are ascribed to the reduction of surface and bulk oxygen of CeO₂, respectively ¹³. The amount of H₂ consumption is 342 µmol· g^-1 at temperature less than 400 °C, and is about 12% of the total amount for the complete reduction of ceria from CeO₂ to Ce₂O₃ The H₂-TPR profile of pure MnO_x shows a main reduction peaks at 323 °C with a slight overlapped shoulder peak at 273 °C, indicating a two-step reduction (MnO_x→ Mn₃O₄ and Mn₃O₄ → MnO)¹⁴. The total H₂ consumption is calculated to be 5103 μmol· g^-1, from which the predominant manganese species is estimated to be Mn₃O₄ as well as some higher valence manganese oxides. It is consistent with the result of XRD. For CeO₂-MnO_x mixed oxides, the H₂-TPR profiles also show two reduction peaks (denoted as β and γ, respectively). According to the reduction characteristics of pure CeO₂ and MnO_x associated to the disappearance of the ceria low-temperature peak of CeO₂-MnO_x (CB), peak β is assigned to the reduction of MnO_x2→ Mn₃O₄, and peak γ represents the combined reductions of Mn₃O₄₂→ MnO and the surface Ce⁴+ . Compared with the reduction temperature of pure MnO_x and CeO₂, the reduction temperatures of CeO₂₂→ MnO_x (CB) at 180 and 208 °C greatly shift to lower regions, indicating the presence of interaction between manganese and cerium oxides. Although that of CeO₂₂→ MnO_x (DP) also shift to lower temperature, the magnitude of shift is little. Thus, it can be inferred that the interaction between manganese and cerium oxides in CeO₂-MnO_x (DP) is weaker than that in CeO₂₂→ MnO_x (CB). It is worthy noting that the H₂ consumptions of peak β is more than that of peak γ, which imply that some manganese oxide species turn to higher oxidation states in CeO₂-MnO_x (CB). All these are possibly related to the formation of CeO₂- MnO_x (CB) solid solution.

3.3     Study of XPS spectra of the catalysts

XPS characterization was performed to obtain information of the chemical state of the surface elements. Fig.4 (a) shows the XPS spectra of Mn 2p. It can be seen that the binding energy (BE) peaks of Mn 2p_3/2 for CuO/CeO₂-MnO_x (DP) and CuO/MnO_x are both centered at 641.8 eV, indicting the similar state of manganese species in the two samples, and that for CuO/CeO₂-MnO_x (CB) shifts to higher binding energy centered at 642.0 eV. According to the literature ¹⁹, the BE of Mn2p_3/2 in pure manganese oxides was found to be at 641.2 eV (Mn²+), 641.8 eV (Mn³+) and 642.1 eV (Mn⁴+), respectively. Then, from deconvolution of the Mn 2p_3/2 peaks, it can be estimated that the relative ratio of Mn³+ is about 83% for CuO/CeO₂-MnO_x (DP) and that of Mn⁴+ is about 62% for CuO/CeO₂-MnO_x (CB). Therefore, the Mn³+ is the dominant species in CuO/CeO₂-MnO_x (DP), while Mn⁴⁺ ions exist in CuO/CeO₂-MnO_x (CB) as main manganese species. Six B.E peaks at 882.3, 888.8, 898.4, 900*8, 907.2 and 916.7 eV are observed in the Ce 3d XPS spectra shown in Fig.4 (b). These components should be assigned to Ce⁴⁺ species by comparison with the data of literatures ²⁰. CuO/CeO₂-MnO_x (CB) catalyst shows an inconspicuous peak at 885.0 eV along with the decrease of BE peak at 898.4 eV(arrowed in figure), which indicates the existence of Ce³⁺ species in addition to Ce⁴⁺. The generation of more Mn⁴⁺ species with existence of partial Ce³⁺ species should be arised from the strong interaction between CeO₂ and MnO_x in CuO/CeO₂-MnO_x (CB). As is well known, high valence state of manganese oxides is easily reduced, so the presence of Mn⁴⁺ may enhance the redox property of the catalyst.

Fig.4 (c) shows the XPS spectra of Cu 2p_3/2 in CuO-based catalysts. It can be observed that these spectra consist of the main peak of Cu 2p_3/2 centered at 932.6-933.8 eV and a shake-up peak around 942.5 eV. The B.E peak of Cu 2p_3/2 for CuO/CeO₂ is found at 933.8 eV with a clear shake-up peak, which is the characteristic of CuO according to the literature ^{21, 22}. Lower B.E at 932.2-933.1 eV and the absence of the shake-up peak are characteristic of Cu₂O ^{22, 23}. The Cu 2p_3/2 peaks shift to lower B.E for CuO/CeO₂ -MnO_x(CB), CuO/CeO₂-MnO_x(DP) and CuO/MnO_x compared with CuO/CeO₂, and CuO/CeO₂-MnO_x(CB) exhibits the lowest BE. The shift indicates some modifications of electronic properties of copper species, suggesting the occurrence of electrons transfer between copper species and manganese species. The valence state of copper species on CuO/CeO₂- MnO_x(CB) are also effected by oxygen vacancy on CeO₂-MnO (CB) solid solution according to U.R. Pillai's reports ²⁴. Simultaneously, the shift is accompanied with a obvious decrease of the intensity of the shake-up peak. Thus, their XPS spectra of Cu 2p_3/2 demonstrate that Cu⁺ species exist in these three catalysts, and the number of Cu⁺ species on CuO/CeO₂-MnO_x (CB) seems to be more than that on CuO/CeO₂ -MnO_x (DP). Cu⁺ species is far likely to be the active site for CO adsorption during the CO PROX reaction ⁷.

Fig. 5 displays the XPS spectra of O1s used to investigate the oxygen species on the surface of CuO/CeO₂-MnO_x catalysts. The O1s spectra of CuO/ CeO₂-MnO_x catalysts contain a main peak at 528.9-529.5 eV could be ascribed to the lattice oxygen (denoted as O ), along with a shoulder at 531.3-532.3 eV assigned to the defect oxides or a mixture of hydroxyl groups on the surface of the catalysts (denoted as O_β) ²⁵. The defect oxides are favorable produce of oxygen vacancy. The comparison of the XPS spectra of O 15 between CuO/CeO₂-MnO_x (DP) and CuO/CeO₂- MnO_x (CB) indicates that the main peak intensity of the two catalysts is similar, indicating their lattice oxygen amount is almost the same. The relative concentrations of O_β listed in table 3 are 23.6 and 41.7 % for CuO/CeO₂-MnO_x (DP) and CuO/CeO₂-MnO_x (CB), respectively. According to the following mechanism of preferential oxidation of CO (PROX) ²⁶: (1) CO + M (adsorption site) → CO-M; (2) CO-M + O*(lattice oxygen) → CO₂-M + O_d (oxygen vacancy); (3) CO₂-M — CO₂ + M; (4) O₂ + 2O_d→ 2O*, the high concentrations of O_β may be favorable for reaction (4). Oxygen molecular in reaction gases can transfer to lattice oxygen via oxygen vacancy. Compared with CuO/CeO₂- MnO_x (DP), O_β peak intensity of CuO/CeO₂-MnO_x (CB) is stronger, which are favorable for the mobility of oxygen species and the generation of lattice oxygen.

3.4.      CO adsorption property of the catalysts

3.5.      Performance of CuO/CeO₂-MnO_x catalysts

The catalytic activity of CeO₂-MnO_x mixed oxides and CuO-based catalysts for CO PROX system is shown in Fig. 7. CeO₂-MnO_x mixed oxides show relative low CO conversion in the temperature range investigated. It can be found that catalytic activity is not only dependent upon active constituents, but also has relations with supports of the catalysts. Obviously, the catalytic activity values of CuO-based catalysts with CeO₂-MnO_x mixed oxides as two-way support are higher than that with single support, especially at low temperature. From Fig.7, the catalytic activity of catalysts can be ranked as: CuO/CeO₂-MnO_x (CB)>CuO/CeO₂-MnO_x (DP)>CuO/CeO₂ >CuO/MnO_x. The activity of CuO/CeO₂-MnO_x (CB) catalyst is higher than that of CuO/ CeO₂-MnO_x (DP). A complete conversion of CO over CuO/CeO₂-MnO_x (CB) achieves at 140 °C, whereas that over CuO/CeO₂-MnO_x (DP) catalyst is at 180 °C. This result indicates that the preparation methods of CeO₂-MnO_x support have a significant influence on the catalytic performance of CuO/CeO₂-MnO_x catalysts. After the first measurement, CuO/CeO₂-MnO_x catalysts are successively under reaction atmosphere at 200 °C for 10 h and cool down to room temperature, and then start the second activity test. The results after the temperature cycle are shown in Fig. 8. It can be observed that the conversion of CO remains almost unchanged, in addition to a slight decrease at 60 °C and 80 °C. The values of CuO/CeO₂-MnO_x (CB) and CuO/CeO₂-MnO_x (DP) decrease by 6% and 10% at 80 °C, respectively. This indicates that CuO/CeO₂-MnO_x(CB) catalyst is more stable for CO oxidation in H₂-rich gases than CuO/CeO₂-MnO_x (DP).

4.       CONCLUSION

ACKNOWLEDGEMENTS

We gratefully acknowledge the financial support from the National Natural Science Foundation of China (No. 21062013).

(Received: August 5, 2011 - Accepted: October 3, 2011)

REFERENCES

1. E. D Park, D. Lee, H. C. Lee. Catal. Today. 139, 280, (2009).         [  ]

2. Eun-Yong Ko, Eun Duck Park, Kyung Won Seo. Catal.Toda .116, 377, (2006).         [  ]

3. G. Avgouropoulos, T. Loannides, H.K. Matralis. Catal. Lett. 73, 33, (2001).         [  ]

4. D.H. Kim, J.E. Cha. Catal.Lett. 86, 107, (2003).         [  ]

5. Hui Wang, Huaqing Zhu, Zhangfeng Qin. J. Catal. 264, 154, (2009).         [  ]

6. A.M.D. deFarias, A.P.M.G. Barandas, R.F. Perez, M.A. Fraga. J. Power. Sources. 165, 854, (2007).         [  ]

7. W. Liu, MF. Stephanopoulos. J. Catal. 153, 304, (1995).         [  ]

8. A. Martinez-Arías, A.B. Hungría, M. Fernández-García. J. Power. Sources. 151, 32, (2005).         [  ]

9. Li Lei, Zhan Ying-Ying, Chen Chong-Qi. ActaPhys.-Chim. Sin. 25, 1397, (2009).         [  ]

10. X.F. Tang, J.L. Chen, X.M. Huang, Y.D. Xu, W.J. Shen. Appl. Cata. B: environ. 81, 115, (2008).         [  ]

11. Qing Liang, Xiaodong Wu,Duan Weng, Haibo Xu. Catal. Today. 139, 113, (2008).         [  ]

12. Gregor Sedmak, Stanko Hocevar, Janez Levec. J. Catal. 213, 135, (2003).         [  ]

13. Tiziana Caputo, Luciana Lisi, Raffaele Pirone, Gennaro Russo. Appl. Cata. A: Genl. 348, 42 (2008)        [  ]

14. X.F. Tang, J.L. Chen, Y.G. Li, Y. Li,Y.D. Xu, W.J. Shen. Chem. Eng. J. 118, 119, (2006).         [  ]

15. Zhigang Liu, Renxian Zhou, Xiaoming Zheng. J. Mol. Catal. A:Chem. 267, 137, (2007).         [  ]

16. Meng-Fei Luo, Jing-Meng Ma, Ji-Qing Lu, Yu-Peng Song, Yue-Juan Wang. J. Catal. 246, 52, (2007).         [  ]

17. Liu Y, Fu Q, M.F. Stephanopoulos. Catal. Today. 9, 241, (2004).         [  ]

18. A.Martinez-Arias, M.Fernandez-Garcia, J. Soria, J.C. Conesa. J. Catal. 182, 367, (1999).         [  ]

19. Kanaparthi Ramesh, Luwei Chen, Fengxi Chen, Yan Liu, Zhan Wang, Yi-Fan Han. Catal. Today. 131, 477, (2008).         [  ]

20. Hanbo Zou, Xingfa Dong, Weiming Lin. Appl. Surf. Sci. 253, 2893, (2006)          [  ].

21. D. Gamarra, A. Hornés, Zs. Koppány, Z. Schay, G. Munuera J. Soria, A. Martinez-Arias. J. Power. Sou rces. 169, 110, (2007).         [  ]

22. G. Avgouropoulos, T. Ioannides, H. Matralis. Appl. Cata. B: environ. 56, 87, (2005).         [  ]

23. W. Liu , M.F. Stephanopoulos. J. Catal. 153, 317, (1995).         [  ]

24. U.R. Pillai, S. Deevi. Appl. Cata. B: environ. 65,110, (2006).         [  ]

25. Z.Y. Pu, J.Q. Lu, M.F. Luo, Y.L. Xie. J. Phys. Chem. C. 111, 18695, (2007)          [  ].

26. Christopher S. Polster, Hari Nair, Chelsey D. Baertsch. J. Catal. 266, 308, (2009).         [  ]

27. Xuesong Liu, Jiqing Lu, Xiaoxia Wang, Mengfei Luo. Chin. J. Catal. 31, 181, (2010).         [  ]

28. Meng-Fei Luo, Yijun Zhong, Xian-xin Yuan, Xiao-Ming Zheng. Appl. Catal. A: Gen. 162, 121, (1997).         [  ]