Heterogeneous Catalysts for Liquid-Phase Oxidations: Philosophers' Stones or Trojan Horses?

Abstract

Introduction In the early seventies one of us1 was involved in the development of the heterogeneous Ti(IV)/SiO2 catalyst which forms the basis of the Shell process for the epoxidation of propylene with ethylbenzene hydroperoxide (reaction 1).2 Halcon3 and ARCO4,5 workers had previously found, independently, that soluble compounds of early transition metals, e.g., Mo, W, Ti, and V, catalyze reaction 1. The mechanism of catalysis involves withdrawal of electrons from a coordinated alkylperoxo moiety, thereby increasing the electrophilic character of the peroxidic oxygens, i.e., the metal ion acts as a Lewis acid. Hence, effective catalysts are both a strong Lewis acid and a weak oxidant in their highest oxidation state. The latter criterion is necessary in order to minimize competing one-electron oxidation of the ROO ligand leading to homolytic decomposition of ROOH (see Scheme 1). These criteria are best met by molybdenum(VI), and soluble molybdenum compounds exhibit the best combination of activity and selectivity.6,7 Soluble titanium(IV) compounds, on the other hand, are rather mediocre catalysts for reaction 1. In contrast, Ti(IV)/SiO2 exhibits selectivities comparable to homogeneous molybdenum and (for a heterogeneous catalyst) high activities.8 The superior catalytic activity of Ti(IV)/SiO2 was attributed to both an increase in Lewis acidity of the Ti(IV), owing to electron withdrawal by silanoxy ligands, and to site isolation of discrete Ti(IV) centers in the silica lattice preventing oligomerization to unreactive μ-oxo species (which readily occurs with soluble Ti(IV) compounds). Furthermore, it was demonstrated that only the combination of titanium(IV) with silica affords a stable heterogeneous catalyst; all other combinations, e.g., Mo(VI), W(VI), V(V), etc., on silica, gave rapid leaching of the metal ion. One property which soluble Ti(IV) compounds and Ti(IV)/SiO2 share is a marked sensitivity toward deactivation by strongly coordinating ligands, especially water.9 For this reason Ti(IV)/ SiO2 is an ineffective catalyst for epoxidations with aqueous hydrogen peroxide. Hence the appearance in the mid-eighties of Enichem patents10 describing the remarkable catalytic activity of titanium(IV) silicalite (generally known as TS-1) in, inter alia, the selective epoxidation of olefins under very mild conditions with 30% aqueous hydrogen peroxide (Figure 1) was greeted with some scepticism. Thus, two materials, Ti(IV)/SiO2 and TS-1, having roughly the same elemental composition, i.e., 2% Ti in SiO2, exhibited totally different catalytic properties. Initial attempts by various groups to reproduce the Enichem results were largely unsuccessful. However, once it became clear that certain parameters in the synthesis