Computational Tools for Physical Property Estimation and Solvent Selection

We will present the details of a computational, physical property-based solvent selection methodology that we have used in numerous process development projects. We will use the example of selecting a solvent for a liquid extraction process to explain the methodology’s steps. Specifically, we will discuss:

• Identification of physical property constraints, e.g., the high solubility of the key component, immiscibility with the feed solvent, proper volatility for regeneration by distillation, etc.
• Identification of required or disallowed substructures, e.g., restricting chemically or thermally unstable substructures, requiring reactive substructures, restricting environmentally undesired substructures, etc.
• Estimation of required physical properties, e.g., liquid-liquid equilibrium for extraction, vapor pressures for distillation, surface tensions and densities for settling, etc.
• Screening and ranking of candidate solvents, e.g., selecting a database of candidate solvents, identifying general trends in recommendations, iterating on the entire selection process, etc.

We will demonstrate how this selection methodology has been implemented in a software tool and show how such software tools are essential for conducting a rapid and thorough selection process.