As a prelude to commissioning a reactive distillation column (RDC) for the continuous epoxidation of cyclohexene using t-buty1hydroperoxide and a heterogeneous polymer-supported Mo(VI) catalyst, the long-term stability of three candidate polymer-supported Mo catalysts has been evaluated. The polymer catalysts are a polybenzimidazole-supported species, PBI.Mo, and two poly(styrene-divinylbenzene) resin-based species, Ps.AMP.Mo1 and 2, prepared in-house by amination of vinyl benzyl chloride-containing resins using 2-aminomethyl pyridine, followed by loading with Mo. The stability of each polymer catalyst was assessed by recycling a sample 10 times in small batch reactions using conditions that will form the basis of the continuous process. At the same time the loss of Mo from each support has been investigated by isolating any residue from reaction supernatant solutions, following removal of the heterogeneous polymer catalyst, and then using the residues as potential catalysts in epoxidation reactions. This is a powerful technique for identifying unambiguously loss of catalytically active Mo from the support and is not dependent on Mo analytical procedures. All three polymer catalysts are highly active and selective in 10 consecutive reactions but the supernatant solutions also contain catalytically active Mo residues. In the case of PBI.Mo and Ps.AMP.Mo2 the contribution to catalysis from homogeneous Mo species lost from the supports is all but eliminated after aging through 10 cycles, but in the case of Ps.AMP.Mo1 the contribution from leached Mo species remains significant even after similar aging. The major factor determining this differential behaviour seems to be the mole ratio of polymer-bound ligand to Mo which must be significantly above unity to provide long-term retention of Mo. (c) 2007 Elsevier Ltd. All rights reserved.