Diagnostic use of compensation phenomena in heterogeneous catalysis−: Reactions of alkanes on platinum and palladium catalysts

Abstract

The linear dependence of apparent activation energy E app upon ln A app ( A represents the pre-exponential factor) as either catalyst or reaction is varied has often been demonstrated in heterogeneous catalysis: the plot of E app versus ln A app is termed a Constable plot. It has been suggested that the basic cause of the correlation is the use of apparent rather than true Arrhenius parameters, the variation in E app arising from differing contributions of adsorption enthalpies to the true activation energy E true. Constable plots are shown for the hydrogenolysis of ethane and for reactions of higher alkanes with H 2 on various Pt and Pd catalysts, and on the standard Pt catalysts EUROPT-1 and -3. The major cause of scatter in these plots is thought to be failure to estimate properly the number of active centres in working catalysts. Constable plots reveal mechanistic information (e.g. concerning a common intermediate for both hydrogenolysis and isomerisation, and a different mechanism for the reactions of neopentane from that used by linear or less branched alkanes). Certain modifications of the active metal are shown in this way not to produce wholly new types of centre (e.g. in PtRe/Al 2O 3 or Al 2O 3 deposited on Pt/SiO 2), but significantly different Constable lines shown by Pt/smectite clay and by PtMo/SiO 2 are indicative of novel, perhaps bimetallic, centres. In the family of reactions considered, the surface coverage by H atoms is the dominant factor in determining the value of E app.