A new decomposition-based computer-aided molecular/mixture design methodology for the design of optimal solvents and solvent mixtures

This paper presents a novel computer-aided molecular/mixture design (CAMD) methodology for the design of optimal solvents and solvent mixtures. The molecular/mixture design problem is formulated as a mixed integer nonlinear programming (MINLP) model in which a performance objective is to be optimized subject to structural, property, and process constraints.

The general molecular/mixture design problem is divided into two parts. For optimal single-compound design, the first part is solved. For mixture design, the single-compound design is first carried out to identify candidates and then the second part is solved to determine the optimal mixture. The decomposition of the CAMD MINLP model into relatively easy to solve subproblems is essentially a partitioning of the constraints from the original set.

This approach is illustrated through two case studies. The first case study involves the design of an optimal extractant for the separation of acetic acid from water by liquid-liquid extraction. The results suggest that the new extractant would be able to perform better than the extractant being widely used for this separation.

The second case study is an industrial problem involving the optimal formulation for a pharmaceutical compound. The designed formulation is able to improve the water solubility of the compound by more many fold.