The conductor-like screening model for real solvents (COSMO-RS) is a powerful method for predicting thermodynamic properties of small molecules, such as solvation energy, vapor pressure, and solubility in mixed solvents. The parameters in the model have been trained by using a dataset of over 200 molecules and it has been proven to be a successful general-purpose method. When we only focus on the prediction of aqueous solubility, the overall performance of COSMO-RS is not satisfactory based on our test on a larger database over 1800 molecules. We explored different factors that could affect the predictions and studied their relations with sigma profile. We found the main challenges is from the electronegative atoms, which will more likely generate errors in statistical calculation from sigma profile and too strong hydrogen bond corrections.

Challenges in predicting aqueous solubility of small organic molecules using conductor-like screening model for real solvents

ACS Fall 2021

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.