Effect of Fractal Dimension on CO Selectivity of K-Ni/alpha-Al sub(2)O sub(3) Catalyst for High Pressure Oxidative Reforming of Methane

• Published in: Journal of the Japan Petroleum Institute Vol. 53; no. 3; pp. 191 - 198191
• Main Authors: Omata, Kohji, Kobayashi, Seishiro, Horiguchi, Junpei, Yamazaki, Yuichiro, Itabashi, Daisuke, Nakanishi, Takayuki, Yamada, Muneyoshi
• Format: Journal Article
• Language: English
• Published: 01.01.2010
• Subjects: Carbon monoxide; Catalysts; Diffraction; Fractal analysis; Fractals; Morphology; Nickel; Selectivity
• ISSN: 1349-273X
• DOI: 10.1627/jpi.53.191

To suppress total oxidation and increase the CO selectivity of high-pressure oxidative reforming of methane, preparation parameters of K-Ni/alpha-Al sub(2)O sub(3) catalyst were optimized. The target preparation parameters were the main factors affecting CO selectivity such as calcination temperature of gamma-Al sub(2)O sub(3), and NiO and K loadings. Parameters were designed by L sub(9) orthogonal array containing three factors and three levels, and the catalysts were prepared by the sequential impregnation method with Ni loaded prior to K. Activity test of the L sub(9) catalysts was conducted at 1 MPa and 650[deg]C. The relationship between the preparation conditions and CO selectivity of the L sub(9) catalysts obtained from the experimental results was analyzed using a radial basis function network (RBFN). The CO selectivity was expressed by the RBFN as functions of the catalyst preparation parameters. A grid search, where all combinations of the preparation conditions were input to the RBFN, was conducted to identify the optimum conditions of catalyst preparation for the highest CO selectivity. The predicted catalyst was 1.3 wt% K-14 wt% NiO on alpha-Al sub(2)O sub(3) calcined at 1195[deg]C. Surface NiO morphology was estimated by fractal dimension of Ni distribution in the EPMA image, and X-ray diffraction (XRD) and BET surface area measurements were conducted using the L sub(9) catalysts. These characters were correlated to the catalyst preparation conditions by RBFNs, and the grid data were calculated on the RBFNs. These grid data were analyzed by multivariate analysis to calculate the correlations between CO selectivity, fractal dimension, NiO crystalline parameters, and BET surface area. CO selectivity was expressed by a linear combination of the fractal dimension and NiO crystalline parameters. Another multivariate analysis suggested that fractal dimension was also expressed by primary expression of the NiO diffraction angle, so that the preparation conditions, especially the K loading, affect the NiO diffraction angle. Therefore, K addition was suggested to affect the NiO lattice to modify the surface morphology of NiO resulting in change of CO selectivity.