Ryan Marson1,Chris Roth1,Wanglin Yu1,Michael Tate1
Dow1
Ryan Marson1,Chris Roth1,Wanglin Yu1,Michael Tate1
Dow1
Crop protection is a multi-million-dollar industry requiring complex formulations of multiple surfactants in a solvent, which must form stable emulsions prior to being dispersed. We outline a joint computational and experimental study undertaken within Dow to predict the stability of a given formulation. Data from thousands of high-throughput formulation experiments were compiled and analyzed for emulsion stability. This experimental data set was then used in combination with physical and chemical descriptors of the formulation components to train an ensemble of cluster-based machine learning models (e.g., Decision Trees, k-means, etc.) to predict whether a given formulation would be stable, some of which demonstrated cross-validation accuracies as high as 75%. We outline the details of this effort and discuss opportunities for continuous improvement of the models.