Catalysis Today, Vol.159, No.1, 12-24, 2011
Development of dry reforming catalysts at elevated pressure: D-optimal vs. full factorial design
In this paper high-throughput and combinatorial strategies are used to investigate the development of dry reforming catalysts at elevated pressure (7 bar). Via a design of experiment (DoE) two straightforward search strategies, D-optimal and full factorial, were evaluated in their ability to identify highly active and selective catalysts. An extensive high-throughput screening has pointed out that Ni and MgAl(2)O(4) were the active element and support, respectively, with the highest potential for the dry reforming reaction at elevated pressure. A promoter screening of alkaline, alkaline-earth, transition and lanthanide metal oxides on NiMgAl(2)O(4) catalysts showed that promoter effects, reported at atmospheric pressure, not always hold at elevated pressure. La(2)O(3), Y(2)O(3) and ZrO(2) were selected as promoters for further use in DoE. A cubic space, constructed with elements with high potential, was explored by the two search strategies. ANOVA analyses of the experimental data resulted in linear equations for the conversions of CH(4) (X(CH4)), CO(2) (X(CO2)) and the H(2)/CO ratio. Both strategies allowed predicting comparable catalytic results in the major part of the cubic space and deducing the same effects on activity and selectivity for the three selected promoter elements. La(2)O(3) affected activity and selectivity in a positive way, while the other two promoters, ZrO(2) and Y(2)O(3), resulted in a decreased catalyst performance. Based on the linear equations an objective function R was created to determine the most active and selective catalysts. Maximum R values were obtained for the same catalyst composition in case of the two different search strategies, hereby demonstrating the advantage of the D-optimal strategy over the full factorial strategy, since in the former case comparable results are obtained with only half of the number of experiments. (C) 2010 Elsevier B.V. All rights reserved.
Keywords:CO(2) reforming;Dry reforming;Design of experiment;High-throughput experimentation;D-optimal;Full factorial