Catalytic Consequences of a Revised Distribution of Key Elements at the Active Centers of the M1 Phase of the MoVNbTeOx System

In an earlier publication (Grasselli et al., Top Catal 54:595, 2011) [1] we analyzed the distribution of key catalytic elements at the active site of the M1 phase of the MoVNbTeOx catalyst system based on the definition of this center put forth in our original work (Grasselli et al., Top Catal 23:5, 2003) [2]. From that analysis we derived the probabilities of the metal element distributions at the active center and its immediate surroundings, and based on those results, proposed a model for propane ammoxidation on M1. Recently, the occupancies and concomitant charges of the various elements of the M1 phase have been revised (Li et al., Top Catal 54:614, 2011) [3]. Based on this revised structural model we have now recalculated the elemental probabilities at the active center and its immediate surroundings, and describe here the catalytic consequences in the ammoxidation mechanism of propane that these changes portend. Our revised model (REV) predicts more closely the actual experimental results of propane ammoxidation over MoVNbTeOx than does our first model (ORIG). The results obtained are: ORIG Model: 41 % AN (acrylonitrile) concerted, 82 % total possible AN; REV Model: 43 % AN concerted, 59 % total possible AN; experimental: 42 % AN concerted, 62 % total possible AN using both the M1 + M2 phases. Comparing the original (Grasselli et al., Top Catal 23:5, 2003) [2] and current (Li et al., Top Catal 54: 614, 2011) [3] elemental distributions at the active centers of M1 and the ORIG and REV ammoxidation reaction pathways (from the derived models), it is readily apparent that higher concentrations of V5+ at the active centers lead to undesirable overoxidation of propane and thus lower AN selectivity. Therefore, decreasing the surface concentration of vanadium in M1 (to favor more site isolated V5+ sites) should be beneficial and lead to better AN selectivities and yields. Additionally, selective doping or selective isomorphous substitution of M1 (and/or M2) should also be useful approaches towards improved AN yields.