460397 A Target Upper Bound on Reaction Selectivity Via Feinberg’s Cfstr Equivalence Principle
Mole balances are widely used in the design of chemical plants and processing equipment as well as in conducting techno-economic analyses of potential chemical business ventures . They allow us to calculate unknown flowrates into and out of a system once an appropriate number of specifications are made. On a broader level, mole balances also serve to bound the attainable region of a reaction system. We should not, however, expect that the entire attainable region be accessible for a process on which capacity constraints are imposed. For example, certain areas of the attainable region may be accessible only by permitting extremely low single-pass reactor conversions to achieve high selectivities resulting in impractically large recycle flowrates. It is the underlying kinetics of a chemistry that determine the size and location of these inaccessible regions.
Feinberg and coworkers developed the Continuous Flow Stirred Tank Reaction (CFSTR) Equivalence Principal which allows one to decompose any arbitrary steady-state reactor separator system with total reaction volume “V” into a new system comprising “R+1” CFSTRs (where R is the number of linearly independent reactions) with the same total reaction volume and a perfect separations system [2, 3]. Using this methodology in conjunction with the kinetics of a system of interest, the attainable region given by the mole balances can be refined. Our work aims to further refine the attainable region by introducing flowrate capacity constraints on the CFSTR Equivalence Principle. The constraints on molar flowrates between the “R+1” CFSTRs and the separations system prevent large recycle streams and small reactor conversions. By optimizing this constrained system of CFSTRs, we can determine the maximum possible selectivity of several chemistries completely independent of reactor design given capacity constraints. We have investigated serial reaction networks and more complex reaction networks involving serial and parallel reactions. With the added constraints, the maximum possible selectivity for both types of reaction networks is less than 100%.
We have used this methodology to analyze the maximum possible selectivity of a selection of realistic chemistries and have compared the results to archetypal reactors (e.g., CFSTR, PFR). The results all support the hypothesis that this CFSTR reactor decomposition can be constrained and optimized to obtain a target upper bound on selectivity for chemistries completely independent of reactor-separator design.
 Douglas, J. “Conceptual Design of Chemical Processes”, McGraw-Hill, New York (1988).
 Feinberg, M.; Ellison, P. Ind. Eng. Chem. Res. 2001, 40, 3181-2194.
 Tang, Y.; Feinberg, M. Ind. Eng. Chem. Res. 2007, 46, 5624-5630.