(586i) Optimal Antisolvent Addition Profile in a Continuous Plug Flow Crystallizer Through Population Balance Based Multi-Objective Optimization

Authors: 
Majumder, A., Loughborough University
Nagy, Z. K., Purdue University



In recent years, the interest in continuous processing in pharmaceutical manufacturing has increased significantly due to its ability to provide several benefits such as improvement in processing time, consistency and reduction in expenses. In this work, we present the simulation study for obtaining optimal antisolvent addition profile in a 4-segment continuous plug flow crystallizer (PFC) with the objectives to maximize the mean size and minimize the coefficient of variation (CV). The system studied here is the crystallization of flufenamic acid in ethanol-water (antisolvent) system with experimentally determined parameters from Alvarez and Myerson [1]. The antisolvent can be introduced at the inlet of each segment. A steady state population balance equation (PBE), coupled with the necessary mass balance equation, is used to describe the evolution of the crystal size distribution (CSD) along the crystallizer. An optimization problem has been formulated which is then solved using nondominated sorting genetic algorithm II (NSGA II) available in Matlab. Due to the competing nature of the two objectives considered, no single optimal point is found, rather a set of optimal points called Pareto optimal points are obtained.  The results are compared with those when 1 to 4 equal additions are used by keeping the total amount of antisolvent the same. It is found that the optimal profile is different from the profiles that distribute the antisolvent equally in 1 to 4 segments and is able to increase the mean size of the crystal significantly while maintaining the similar CV.

[1] A. J. Alvarez, A. S. Myerson: Continuous plug  flow crystallization of pharmaceutical compounds (2010), Cryst. Growth Des. 10 (5), 2219-2228.