Partial Hydrogenation of Soybean Oil in a Piston Oscillating Monolith Reactor Yogesh G. Waghmare, Alan G. Bussard, Robert V. Forest, F. Carl Knopf, and Kerry M. Dooley* Department of Chemical Engineering, Louisiana State UniVersity, Baton Rouge, Louisiana 70803 The partial hydrogenation of soybean oil was carried out in a novel piston oscillating monolith reactor (POMR). POMR performance was studied under the application of low frequency (0-17.5 Hz) and amplitude (2.5 mm) vibrations at 110 °C and 0.41 MPa H 2 using a Pd/Al 2 O 3 monolith catalyst. Results show observed rate improvements of up to 220% for 17.5 Hz, 2.5 mm piston oscillations over low frequency pulsing conditions. For comparison purposes, the reaction was also carried out in a stirred tank reactor using the monolith catalyst. The POMR showed better activity at an equivalent power per unit volume when compared to a stirred tank. With the monolith catalyst, both external and internal mass transfer limitations exist. Using standard diffusion- reaction calculations and measurements over a range of particle sizes it was shown that the vibrations improve external mass transfer rates as well as internal transport within the washcoat. The POMR showed equal or better serial pathway selectivity than a stirred tank, except at the highest frequency, but gave higher trans fatty acid formation. Introduction The heterogeneous catalyzed hydrogenation of edible oils has long been an important process for the food industry, because it provides improved resistance to oxidation and better textural properties (e.g., a higher melting point). This reaction is traditionally carried out in a three-phase agitated tank at 0.1-0.7 MPa and 150-200 °C using a Ni-based catalyst present as a slurry. 1 While Ni-based catalysts are prevalent in industrial applications, supported Pd catalysts have also been investigated because of their higher activity, allowing for lower catalyst loadings and temperatures. 2,3 Several continuous flow laboratory reactors have also been studied for this reaction, such as trickle beds, 4,5 tubular reactors, 6 and bubble columns. 7-9 Winterbottom et al. 9 showed a packed bubble column exhibited less trans product formation than a slurry bubble column, possibly due to a more plug-like flow pattern. Boger et al. 10 noted a similar selectivity effect for soybean oil hydrogenation with a monolithic catalyst, compared to a slurry stirred tank, but they attributed this effect to differences in mass transfer. While a high catalyst activity is desired, the selectivity to the intermediate mono- and diunsaturated triglycerides in the serial hydrogenation pathway is also important. The reaction has historically been operated in the external gas mass transfer- limited regime to avoid excessive hydrogenation. 1 Furthermore, recent health concerns regarding the adverse effects of trans fatty acids (TFA) on LDL/HDL cholesterol ratios has spurred research into minimizing the formation of TFAs. Raw soybean oil has no trans-fat content and so the formation of TFAs is the result of stereoisomerization. Previous work has shown that a higher hydrogen concentration on the catalyst surface lowers the rate of TFA formation. 11 The easiest way to increase the surface H 2 concentration is by increasing the H 2 pressure and the rate of agitation. Unfortunately, this will in turn promote the formation of saturates by serial reactions. 11 The effects of large intraparticle concentration gradients have also been investigated. It has been shown that both the serial pathway selectivity and the stereoselectivity decrease when the reaction is pore diffusion-limited with respect to the triglycerides; 12,13 under these conditions the partially hydrogenated triglycerides diffuse more slowly from the pores, giving them opportunity to react further. 1 Pore diffusion limitations with respect to hydrogen have the opposite effect, improving selectivities. 14 While agitated tank slurry reactors are now common for this process, it would be beneficial if a structured catalyst could be used instead, obviating an additional separation step. The presence of catalyst particles or dissolved transition metal is particularly troublesome for a food product. Previous work on three-phase structured reactors has shown they are a viable alternative for gas mass transfer-limited reactions such as hydrogenations. 15,16 The two most common such systems are monoliths operated in the slug (Taylor) flow regime and trickle beds. Boger et al. 10 did an economic evaluation of a process where a monolith reactor is used for the hydrogenation of edible oil and showed that cost reductions up to 40% can be achieved when compared to the conventional slurry reactor process. Monolith reactors in slug flow show improved surface wetting compared to trickle beds, which suffer from rivulet formation and radial gradients in liquid concentrations at the catalyst external surface. Improvements in trickle bed performance are possible when inducing pulsed flows, through the periodic modulation of the liquid feed flow. 17-23 By alternating between gas- and liquid- rich conditions over the surface of the catalyst, the gradients in the gaseous reactant’s concentration can be reduced and the rate of mass transfer increased. Under these conditions, catalyst monolith reactors exhibit alternating gas and liquid slugs passing over the catalyst surface, with the hydrodynamics approaching plug flow. However the liquid phase reactants used in these systems have typically been characterized by low molecular weights and viscosities. Therefore, the logical extension of soyoil hydrogenation to a structured catalytic system is complicated by the higher viscosity and its effect on the mass transfer. One magnetic resonance imaging study has shown that a sucrose solution with twice the viscosity of water increases the film thickness surrounding gas slugs in monoliths, 24 presumably reducing mass transfer. A piston oscillating monolith reactor (POMR) has previously been used to show activity enhancements of up to 84% and equal or better selectivity for the hydrogenation of alpha-methyl styrene (AMS) to cumene, compared to a stirred tank reactor at the same conditions. 25 These improvements result from low * To whom correspondence should be addressed. E-mail: dooley@ lsu.edu. Ind. Eng. Chem. Res. 2010, 49, 6323–6331 6323 10.1021/ie902000e  2010 American Chemical Society Published on Web 06/25/2010