Catalysis Letters Vol. 77, No. 1–3, 2001 107 A hydrophobic gel for epoxidation of olefins with organic peroxides Alfonsus Karli and Gustavo Larsen ∗ Department of Chemical Engineering, 228 Avery Hall, University of Nebraska Lincoln, NE 68588-0126, USA Received 8 June 2001; accepted 8 August 2001 The epoxidation of cyclooctene by tert-butyl hydroperoxide (t -BuOOH) over a TiO 2 –SiO 2 xerogel made from a permethylated cy- clooligosiloxane, tetraethylorthosilicate, and a Ti chloroalkoxide, was carried out in acetonitrile as the reaction solvent. The organic moi- eties of the hybrid gel (in this case, –CH 3 groups) and the Ti content of the catalyst appear to be stable on prolonged exposure to the reaction medium. Besides very good stability, the hybrid catalyst in this study displays 100% selectivity toward cyclooctene epoxide production. Solid state 29 Si magic angle spinning NMR ( 29 Si MASNMR), diffuse reflectance Fourier transform infrared spectroscopy (DRIFTS), and chemical analysis were used to monitor the stability of the hybrid material. The kinetics of olefin epoxidation was studied in a batch reactor in the 313–343 K range. KEY WORDS: hybrid gels; epoxidation; Ti; tert-butyl hydroperoxide 1. Introduction For the heterogeneously catalyzed epoxidation of olefins, the pioneering work of Taramasso et al. [1] on the synthesis of Ti-substituted zeolites still plays a pivotal role in the field. Amorphous TiO 2 –SiO 2 gels on the other hand, have sparked some interest, but it is generally difficult to synthesize such catalysts in a state where Ti leaching is not a serious con- cern [2]. On the other hand, the inherent pore limitations of zeolites are perhaps the primary motivation for research on TiO 2 –SiO 2 gels as epoxidation catalysts. The latter have average pore sizes in the 20 Å range even in their “xerogel” state [2]. Another approach that has also proven very attrac- tive is that presented by Corma [3,4], and later by van der Waal [5], who overcame pore size limitations by synthesiz- ing Ti-containing mesoporous sieves. Within the field of sol–gel derived amorphous materials, hybrid (organic–inorganic) gels are a sub-class that exhibits unique properties [6–8]. A most salient feature of this fam- ily of porous solids is that their hydrophobicity and me- chanical properties can be fine-tuned rather easily [6]. For their use as catalysts, as for all other sol–gel derived mate- rials, one must remain on guard against liquid-phase leach- ing of the active function [9]. In some cases, as Vanoppen et al. [10] also observed in the autoxidation of cyclohexane with a Co-containing molecular sieve as a catalyst, zeolitic materials are also subject to active-function leaching. It is generally believed that leaching is the result of solvolysis of metal–oxygen bonds [9]. Another consideration for uti- lizing hybrid gels as catalysts is that of cost of materials processing [9]. Unless a sol–gel catalyst displaying long- term stability is obtained, such material would not be able to compete with a Ti-zeolite in the epoxidation of small size olefins. ∗ To whom correspondence should be addressed. In this paper, we examine the behavior of the same hy- brid gel catalyst we reported on before for cyclooctene epoxidation [2], this time using tert-butyl hydroperoxide (t -BuOOH) and acetonitrile as the oxidant and solvent, re- spectively, instead of H 2 O 2 and tert-butanol (t -BuOH) as the oxidant/reaction medium. Based on our previous expe- rience, the titanium content in hybrid gels becomes stable after 48 h under reaction conditions, regardless of whether t -BuOOH or H 2 O 2 play the oxidant role. However, further studies done in our laboratory pointed to the fact that methyl groups from the “organic silicon” centers (the material is made from tetraethylorthosilicate and [(CH 3 ) 2 SiO] 3 rings as sol–gel precursors) are not entirely immune to long-term hy- drogen peroxide (H 2 O 2 ) attack, in light of solid state 29 Si nuclear magnetic resonance data. Furthermore, even though one would tend to favor the use of H 2 O 2 from a cost view- point [11], the unimolecular decomposition of H 2 O 2 is an undesirable competing reaction that can hardly be avoided with this type of hybrid Ti-based gels [12]. For example, we have found that that nearly 20% of the H 2 O 2 was lost to this side reaction during cyclooctene epoxidation with H 2 O 2 and t -BuOH as the solvent [13]. In addition, Holmes et al. [12] indicated that the selectivity of t -BuOOH based epoxida- tions by Ti/silica catalysts is typically improved relative to those observed in the same reactions using H 2 O 2 . Thorough spectroscopic, gas adsorption, and thermal be- havior characterization work on these hybrid gels has been reported in previous papers [8,15,16]. In brief, we found that that the Ti function in a stable catalyst, upon proper thermal and gaseous treatments [15], exists as small TiO x aggregates. This was counter-intuitive at first sight, since it has long been argued that isolated, tetrahedral Ti is re- sponsible for epoxidation catalysis. Our materials prepa- ration protocol is identical to that of Neuman and Levin- Elad [16], except for the replacement of about one third of the prescribed amount of tetraethylorthosilicate (TEOS) by 1011-372X/01/1100-0107$19.50/0  2001 Plenum Publishing Corporation