(234a) A Methodology for Systematic Design and Selection of Green Solvents for Increased Yield in Organic Reactions

Solvents are widely used in fine chemicals and pharmaceutical industry where they serve to facilitate reaction-based processes by, for example, dissolving reactants and/or bringing them together in suitable concentrations. Reichardt (1988) reports that the reaction between trimethylamine and trimethylsulfonium ion is 119 times faster in nitromethane than in water which is a dramatic influence indicating the need for knowledge-based choice of solvents for reactions. On the other hand, the excessive consumption of solvents leads to disposal of millions of tons of solvents every year which is a pronounced downside of solvent usage. Given this situation, our aim is to design and select solvents that, when used as reaction media for liquid phase organic reactions, lead to an increase in reaction rate and simultaneously minimize the environmental impact. Provided the details of an organic reacting system (identity of reactants and products, and the reaction conditions) are known, our objective is to find solvents that can promote the studied reaction and rank them according to an established scoring/evaluation system. The solvent selection methodology for promotion of organic reactions developed earlier by Gani et al. (2005) has been used as the basis for this work. It incorporates industrial practice knowledge in the currently available computer-aided tools for solvent design and property estimation. A computer-aided molecular design (CAMD) problem is formulated and solved in ICAS software (ICAS Documentation, 2003) generating a list of solvent candidates. These candidates are then ranked according to a specific score assignment system. The needed reaction data is provided by the user while the solvent data is obtained using a hybrid CAMD technique which allows only the chemically stable molecules that satisfy the solvent-property requirements to be generated and then tested. The original method has now been been extended and combined with a methodology for the selection of reaction media in order to quantify the effect of solvents on reaction rate. A simple reaction model which correlates solvent properties (empirical solvatochromic parameters and Hildebrand solubility parameter) with the logarithm of the reaction rate constant has been embedded within the original methodology CAMD framework. It uses as input kinetic data for a small number of solvents for the studied reaction and can then be extrapolated to predict values of reaction rate constants for a large number of solvents based solely on their structural parameters. This therefore enables another scoring criteria to be embedded in the methodology and rank the solvents according to their effect on reaction rate. The new methodology is illustrated through case studies. Results obtained agree well with results previously published in the literature, as well as with industrial processes data supplied by the user, but also provide guide for experimentation generating new solvents that are yet to be tested.

References: 1. Gani, R., Jiménez-González, C., and Constable, D. J. C. (2005). Computers and Chemical Engineering 29. 1661-1676. 2. ICAS Documentation (2003). Internal Report. CAPEC, Department of Chemical Engineering, Techincal University of Denmark. 3. Reichardt, C. (1988). Solvents and Solvent Effects in Organic Chemistry. VCM Publishers (UK).