DESIGN CONSIDERATIONS OF MEMBRANE REACTOR

In the crossflow configuration widely used in the membrane separation industry today, the direction of the feed flow is perpendicular to that of the permeate flow. To increase the driving force for the permeate, a carrier stream to sweep away the permeate is often used. In most of the membrane configurations such as tubes and multichannel monoliths described in Chapter 5, the carrier stream and the feed stream are generally co urrent or counter-current in direction. 

The geometry of the membrane reactor and the relative locations and flow directions of the feed, permeate and reientate streams all play important roles in the reactor performance. The simplest, but not efficient, membrane reactors consist of disk or foil membranes with a flow-through configuration [Mischenko et al., 1979 Fumeaux et al., 1987]. The same type of membrane reactor can also be consu cted and operated in the more common crossflow mode. 

Single tubular membrane reactors are often used in experimental and feasibility studies. Its justification for use in production environments can sometimes be made in small volume applications. As mentioned in Chapters 4 and 5, inorganic composite membranes consist of multiple layers. The inner most layer in a tubular composite membrane reactor does not necessarily possess the finest pores. For example, a two layered tubular ceramic membrane reactor used for enzymatic reactions has an inner layer containing pores larger than those in the outer layer [Lillo, 1986]. The pores of the inner layer are immobilized with enzymes. Under the influence of an applied pressure difference across the membrane matrix, a solution entering the hollow central core of the tube Hows into the inner layer where the solution reacts with the enzyme. The product which is smaller than the enzyme passes through the permselective outer layer membrane which retains the enzyme. Thus the product is removed from the reaction mixture. 

A still higher membrane packing density can be accomplished with some geometries available to certain commercial organic membranes. One such attempt is hollow fibers [Baker et al., 1978 Beaver, 1986]. Sintered monolithic hollow fibers of metals or metal oxides have been developed [Dobo and Graham, 1979] but their burst strengths appear to be critical. Further developments are needed before commercialization. 

In addition to the above conventional configurations, there have been some novel designs proposed for various typxjs of inorganic membrane reactors. They will be discussed here. 

---