Abstract

Determining the maximum concentrations dissolvable in a given solvent (solubilities) of organic molecules is critical in various fields such as pharmaceuticals, agrochemicals, and environmental science. Knowing how a solute will behave in different solvents and under different temperatures is essential for drug formulation, synthesis, purification, and crystallization. Unknown solubility limits currently hinder the design of new processes, making them difficult or expensive. We propose a fast and general method for predicting the solubilities of neutral organic molecules in a wide range of solvents and temperatures. Our method uses a thermodynamic fusion cycle to combine machine learning predictions of the activity coefficient, enthalpy of fusion, and fusion temperature. This method was tested on a combined dataset with more than ~70,000 experimental solubility labels, showing better or comparable performance on many solubility benchmarks. We also introduce reference ensembling to leverage all available experimental solubilities for a given solute in estimating its solubility in an unseen solvent. Progress on extending this method to solvation properties in mixtures of solvents will also be discussed.

Biography

Emad Al Ibrahim obtained his B.S. (2017) in Mechanical Engineering from the University of Washington. He gained practical experience through internships at Saudi Aramco and Lucid Motors. Then, he earned his M.S. (2018) and Ph.D. (2023) in Mechanical Engineering from King Abdullah University of Science and Technology (KAUST). At KAUST, he worked on Machine learning-enabled spectroscopic sensing with Prof. Farooq. He is conducting postdoctoral research at the Massachusetts Institute of Technology (MIT), focusing on solvation properties and techno-economic studies with Prof. Green.