(141a) Membrane Reactor for the Partial Oxidation of Propylene: Increasing the Yield of Acrolein

Catalytic oxidation of hydrocarbons is an important process in the chemical industry. Partially oxygenated hydrocarbons provide the building blocks for plastics and synthetic fibers as well as precursors for specialty chemicals. However, the oxidation of hydrocarbons with O2 results in the formation of undesired byproducts. Selectivity of partial oxidation products is often low as the desired products are often more readily oxidized than the hydrocarbon feed. In addition, complete combustion to CO2 is thermodynamically favored. The partial oxidation (POx) of hydrocarbons has seen significant interest in distributive membrane reactors (MRs). [1-2] The results from these works demonstrate the ability of MRs to inhibit the formation of full oxidation products and promote the selectivity of partial oxidation products. Acrolein is one of these partial oxidation products of interest as it is a precursor to a number of polymers and an intermediate for various specialty chemicals. Current production of acrolein is attained by the partial oxidation of propylene over Bi-Mo oxide catalysts. The current investigation looks at the advantage of implementing a highly porous, inert, ceramic membrane in the POx of propylene to acrolein (POA). The inert membrane reactor (IMR) distributes oxygen along the catalyst bed thereby reducing the partial pressure of oxygen, as opposed to a traditional fixed bed reactor (FRB) feed that saturates the catalyst bed with a high concentration of oxygen at the entrance to the catalyst bed. The IMR demonstrates a higher selectivity and yield towards acrolein than a conventional FBR (Figure 1). The distribution of oxygen in the IMR takes advantage of the difference in the oxygen reaction order between the full (first order) and partial (zero order) oxidation reaction. The IMR promotes the formation of acrolein and inhibits the production of unwanted CO and CO2 byproducts. In addition to the reactor studies, the Bi-Mo oxide catalyst is examined through XRD, XPS, and IR spectroscopy. The use of these techniques aid in the understanding of the structure of the catalyst and the nature of the reactive species. Further studies look at the production of propylene through the oxidative dehydrogenation of propane over a supported vanadia catalyst. This ODH reaction has oxygen reaction orders identical to that of the POA reaction with the ODH reaction zero order and the combustion reaction first order. Due to this similarity, the ODH reaction is expected to behave similar to the POA reaction with the IMR increasing the selectivity and yield of the desired product over that of a FBR. Modeling of these two reaction networks has been performed and reported earlier [3] using a fourth order Runge-Kutta method in MatLab. The model is being refined and improved with additional experimental data to better simulate the reaction in both the FBR and IMR.

[1] Capannelli, G., E. Carosini, et al. (1996). Chemical Engineering Science 51(10): 1817-1826. [2] Ramos, R., M. Menendez, et al. (2000). Catalysis Today 56: 239-245. [3] O'Neill, C. and E.E. Wolf (2006). Ind. Eng. Chem. Res. 45:2697-2706