Prediction of Relative Kinetic Solvent Effects for SN2 Reactions Using COSMO-RS

Abstract Although ionic reactions comprise a major component of industrially relevant reactions, the prediction of their rate coefficients remains a challenge. Furthermore, existing ab initio kinetic modeling methods are primarily suited for gas-phase conformers and can be computationally expensive. In this work, we highlight a method for predicting relative rate coefficients (i.e., ratio of rates between two solvents) of bimolecular nucleophilic substitution (SN2) reactions by using the COSMO-RS solvation model and a simple relationship derived from transition state theory. To benchmark this quantum-chemical method, we digitized a data set of room-temperature rates for ionic reactions, comprised of 200 data points. Among these, 111 are SN2 rate data, spanning 55 different reactions across various solvents. We used these SN2 data to test a variety of different geometry optimization methods as inputs to the COSMO-RS method. The errors in relative rate constants at 298 K are relatively low (RMSE ≈1 log10 unit) and are not very sensitive to geometry, demonstrating good accuracy and flexibility of this method across different workflows and solvent systems, with higher errors generally observed for systems with some halide nucleophiles. The data and calculations are available at doi:10.5281/zenodo.17260106.