Search

The limitations are addressed that emerge from the computer aided molecular design (CAMD) applications in the optimal synthesis of solvents through a novel approach of designing solvents built upon multiple objective optimization technology. This technology has certain advantages over single objective optimization approaches that have been used for the design of optimal solvent molecules in the past. The different synthesis drives, described by the various objective functions, are treated independently without being submitted to unnecessary fixed assumptions that bias the designs. As a result, all the underlying trends and trade-offs among the properties of the candidate optimal molecules, as well as the structure – property relations are allowed to be revealed. The ability of the presented synthesis framework to embed highly inclusive solutions is elaborated, and the trends are revealed that dominate the relations between the utilized objective functions. It is also shown that solvent molecules, designed in the event that process performance feedback is accounted for, are existent in the Pareto optimal front. The proposed methodology is illustrated with two comprehensive examples in the design of solvents for liquid-liquid extraction and gas-absorption processes. © 2005 American Institute of Chemical Engineers AIChE J, 2006