The field of bio-separations, like many familiar aspects of chemical engineering, remains a mystery: even if one can describe the feed and desired products in detail no unique or truly optimal process can be found. This is in part because there is no single sufficient objective function and in part because the pertinent parameter space is so large. As a result designs tend to cluster closely around familiar prototypes even though these these may not appear very promising.

A heuristic approach for seeking alternates is suggested here, and a promising novel process obtained in this way is presented. Moreover, it is shown that, once found, this process may be described by a simple algorithm suitable for process design.

Extended Abstract Status: Not Uploaded