A Machine Learning Model for the Prediction of Water Contact Angles on Solid Polymers.

Area of Science:

• Materials Science
• Computational Chemistry
• Surface Science

Background:

• Water-surface interactions are classified as hydrophobic or hydrophilic based on wetting.
• Surface coatings can modify wettability by altering surface chemistry.
• Computational methods are increasingly used alongside experimental approaches for materials development.

Purpose of the Study:

• To develop a supervised machine learning model for predicting water contact angles (WCA) on solid polymers.
• To explore the utility of experimental and computational features in WCA prediction.
• To assess the feasibility of using computational features alone for material design.

Main Methods:

• Utilized the XGBoost algorithm for supervised machine learning.
• Incorporated a range of experimental and computational features to train the model.
• Evaluated model performance using mean absolute error (MAE).

Main Results:

• The machine learning model achieved a mean absolute error (MAE) below 5.0° in predicting water contact angles.
• Models trained solely on computational features also demonstrated high accuracy (MAE < 5.0°).

Conclusions:

• The developed ML model accurately predicts the water contact angle on polymer surfaces.
• Computational features alone are sufficient for accurate WCA prediction, facilitating 'bottom-up' design.
• This approach enables the rational design of novel polymers and coatings with tailored wettability.