Surfactants play a crucial role in stabilizing interfaces in detergency, pharmaceuticals, and emulsification. A key thermodynamic quantity governing their performance is the standard Gibbs free energy of adsorption (ΔGads), which reflects how readily surfactant molecules accumulate at interfaces. This work presents an integrated machine learning (ML) and molecular dynamics (MD) framework to predict and validate ΔGads for surfactants at aqueous interfaces. A dataset of experimental ΔGads values for various surfactants at air–water and hydrocarbon–water interfaces was curated from literature sources. Using structure-based molecular descriptors, we trained predictive models employing regression algorithms and graph-based neural networks. Selected molecules were simulated using GROMACS to calculate ΔGads via free energy perturbation (FEP) and validate the ML predictions. The ML models demonstrate strong correlation with MD results, and our pipeline enables fast, accurate screening of surfactants prior to computationally intensive simulations. This approach offers a scalable pathway for surfactant design and optimization based on interfacial thermodynamics.
