(545d) Simultaneous Solvent and Process Optimization of Continuous Reactive Crystallization Processes Involving Electrolytes and Recycling Using the Electrolyte PC-SAFT

Solution crystallization is one of the most commonly used separation techniques in the pharmaceutical industry. Solvents, which are commonly present in crystallizing systems, have a significant impact on the attainable crystal yield, product quality, waste generation, energy duty, and safety of the process. The large number of different solvent types and their mixtures that are routinely used in the pharmaceutical industry makes solvent selection a challenging task, particularly when an optimal trade-off among different process-related performance criteria (i.e., product quality, yield, waste generation) needs to be made. Traditional solvent selection approaches based on heuristics are generally not sufficient for optimization due to the strongly interdependent nature of these criteria. As a result, the development of computer-aided solvent selection approaches for pharmaceutical crystallization processes has attracted substantial research interest. Early work mainly focused on solvent design for stand-alone crystallization [1]. However, solvent selection approaches for integrated operations (i.e., crystallization embedded within a process including downstream separations) are important when minimizing energy duty and waste generation. In particular, solvent recycling can be useful for reducing waste. Furthermore, the current trend towards continuous manufacturing in the pharmaceutical industry has made recycling more convenient [2]. We have developed an optimization-based thermodynamic approach for simultaneous solvent and process design for anti-solvent crystallization integrated with solvent recycling using the statistical associating fluid theory equation of state (PC-SAFT EoS) [3] in our earlier work [4]–[6].

In addition to downstream solvent separation and recycling, upstream reactions for the synthesis of active pharmaceutical ingredients (APIs) may also need to be considered in an integrated approach, which has received less attention. Furthermore, crystallization of APIs can sometimes be combined with a reaction in the same environment to achieve process intensification. However, thermodynamic modeling of such reactive crystallization can be complicated due to the presence of a reaction equilibrium, various types of chemical species and their intermolecular interactions, and the potential to form multiple solid phases (some of which are undesirable). Thus, the development of novel optimization-based solvent selection and design approaches for processes involving reactive crystallization and downstream separation and recycling is needed. Recycling is likely more important for reactive crystallization processes compared to anti-solvent crystallization processes, as reactants need to be considered for recycling in addition to solvents. Our recent work on an integrated approach for a reactive crystallization process of an API with solvent recycling [7] using the PC-SAFT EoS showed the high potential of such an approach. However, API synthesis reactions likely involve electrolytes, and they have an important impact on the phase and reaction equilibria in the process and, therefore, may affect the optimal solvent choice and process conditions, which has not been addressed yet.

The objective of this work is to develop an optimization-based simultaneous solvent and process design framework suitable for the continuous reactive crystallization processes involving electrolytes and downstream separation and recycling of reactants and solvents. The process configuration consists of a reactor/crystallizer, where the reaction and the product crystal formation occur simultaneously, a filter, and a flash separator, where the mother liquor is concentrated and partially recycled back to the crystallizer. The solvent type to be identified acts both as the reaction and the crystallization medium, thus affecting the extent of the reaction and the final crystal yield as well as the ease of recycling. Furthermore, the optimal choice of solvent type is influenced by the process operating conditions, and those two sub-problems (i.e., solvent selection and optimizing operating conditions) cannot be decoupled effectively [8]. Therefore, mathematical models and effective optimization approaches are needed for optimizing both of these sub-problems simultaneously. The development of a process model to relate process operating conditions with solvent choice is a crucial step. In the case of reactive crystallization, the reaction/crystallization medium is usually a complex mixture consisting of multiple solutes with the potential to form different solid phases. It is also well possible that some of those solutes are electrolytes, which calls for a thermodynamic model that can reliably capture all the different types of interactions in the medium. In this work, the PC-SAFT EoS [3] is used as the basis of a unified thermodynamic model to predict the reaction, solid-liquid and vapor-liquid equilibria in the process. The PC-SAFT EoS has been successfully applied in modeling the phase equilibria of a large number of pure substances and mixtures [3], [9], as well as complex mixtures consisting of API molecules commonly encountered in pharmaceutical processes [10]. Furthermore, the electrolyte PC-SAFT (ePC-SAFT) EoS can model electrolyte systems [11]. Therefore, we have embedded the ePC-SAFT EoS in the process model and optimization framework to capture the impact of electrolytes. The remainder of the process model is constructed from conservation balances and additional constraints to ensure the practical and physical feasibility of the process.

The resulting mathematical model is a mixed-integer nonlinear programming (MINLP) problem, which is computationally prohibitive for a model of a complicated continuous pharmaceutical process involving multiple unit operations, reaction and phase equilibria, different types of species, and recycling. Therefore, a continuous mapping method [8] is adopted to convert the original MINLP problem into a nonlinear programming (NLP) problem to make the original problem computationally tractable. The approach is successfully demonstrated for a case study involving reactive crystallization of dalfampridine. Electrolytes have a substantial impact on the equilibria of this system. An economic objective function involving the minimization of the total production cost of the API is employed, where the solvent type and process operating conditions are free variables. A solvent is represented by six pure component PC-SAFT parameters. These parameters are treated as continuous variables, and the resulting optimal values are mapped onto a solvent database from our previous work [4] to identify a real solvent. The reactor/crystallizer medium dielectric constant, which depends on the solvent type and the operating conditions, is an important property for the fugacity calculations of electrolytes [11]. For the optimization framework, the medium dielectric constant needs to be determined in a predictive manner. Therefore, a correlation is established to predict the dielectric constant of the medium using the pure component PC-SAFT parameters and process operating conditions.

The case study demonstrates the sensitivity of the optimal solution with respect to the trade-offs for economic performance, such as the extent of the reaction, crystal yield, recycle flow rates, and energy duty. Furthermore, the case demonstrates the importance of taking all those factors into account simultaneously. The proposed approach reveals a more competitive process design compared to a conventional dalfampridine synthesis case from the literature [12]. The inclusion of electrolytes in the thermodynamic model enables more realistic values for the equilibrium constants and the prediction of possible undesired salt precipitation. The latter can be avoided by imposing constraints.

This work was financially supported by the Research Grant Council of Hong Kong under Grant No. 16214418.

References

[1] A. T. Karunanithi, L. E. K. Achenie, and R. Gani, “A computer-aided molecular design framework for crystallization solvent design,” Chem. Eng. Sci., vol. 61, no. 4, pp. 1247–1260, Feb. 2006.

[2] S. Mascia et al., “End-to-End Continuous Manufacturing of Pharmaceuticals: Integrated Synthesis, Purification, and Final Dosage Formation,” Angew. Chemie Int. Ed., vol. 52, no. 47, pp. 12359–12363, Nov. 2013.

[3] J. Gross and G. Sadowski, “Perturbed-Chain SAFT: An Equation of State Based on a Perturbation Theory for Chain Molecules,” Ind. Eng. Chem. Res., vol. 40, no. 4, pp. 1244–1260, 2001.

[4] J. Wang and R. Lakerveld, “Integrated solvent and process design for continuous crystallization and solvent recycling using PC-SAFT,” AIChE J., vol. 64, no. 4, pp. 1205–1216, Apr. 2018.

[5] J. Wang and R. Lakerveld, “Integrated Solvent and Process Optimization Using PC-SAFT for Continuous Crystallization with Energy-intensive Solvent Separation for Recycling,” Comput. Aided Chem. Eng., vol. 44, pp. 1051–1056, Jan. 2018.

[6] J. Wang, L. Zhu, and R. Lakerveld, “A Hybrid Framework for Simultaneous Process and Solvent Optimization of Continuous Anti-Solvent Crystallization with Distillation for Solvent Recycling,” Processes, vol. 8, no. 1, p. 63, Jan. 2020.

[7] N. P. Mendis, J. Wang, and R. Lakerveld, “A Thermodynamic Approach for Simultaneous Solvent and Process Design of Continuous Reactive Crystallization with Recycling,” in 30th European Symposium on Computer Aided Chemical Engineering, 2020 (Accepted).

[8] A. Bardow, K. Steur, and J. Gross, “Continuous-Molecular Targeting for Integrated Solvent and Process Design,” Ind. Eng. Chem. Res., vol. 49, no. 6, pp. 2834–2840, Mar. 2010.

[9] M. Kleiner and G. Sadowski, “Modeling of Polar Systems Using PCP-SAFT: An Approach to Account for Induced-Association Interactions,” J. Phys. Chem., vol. 111, no. 43, pp. 15544–15553, 2007.

[10] F. Ruether and G. Sadowski, “Modeling the Solubility of Pharmaceuticals in Pure Solvents and Solvent Mixtures for Drug Process Design,” J. Pharm. Sci., vol. 98, no. 11, pp. 4205–4215, Nov. 2009.

[11] C. Held, T. Reschke, S. Mohammad, A. Luza, and G. Sadowski, “ePC-SAFT revised,” Chem. Eng. Res. Des., vol. 92, no. 12, pp. 2884–2897, Dec. 2014.

[12] D. S. Rane and M. M. Sharma, “New strategies for the Hofmann reaction,” J. Chem. Technol. Biotechnol., vol. 59, no. 3, pp. 271–277, Mar. 1994.