AI Approach for Predicting Superhyrophobicity of Thermal Sprayed Copper Coated Aluminum Surfaces

Wettability, characterized by the contact angle of a liquid on a surface, is a critical property that influences numerous natural and industrial applications. In this study, I have developed a CNN-based model to predict the hydrophobicity or super-hydrophobicity of copper-coated aluminum surfaces treated with various reagents or etchants. The data set has been created by analyzing copper-coated aluminum samples with a 3D non-contact profilometer, and contact angle measurements were done to correlate surface properties with the resultant contact angle values. After reagent treatments, the approach had been to preprocess 3D profilometer images to extract surface morphology and structure features. These images and associated contact angle measurements were used as inputs to train the CNN model to classify whether the treated surfaces are hydrophobic or super-hydrophobic. Although the model may initially have limited training accuracy, this study demonstrates the potential of deep learning to predict wettability based on surface characteristics. The results also highlight the need for improvements, such as data set expansion, the inclusion of more varied reagent treatments, and the exploration of hybrid modeling approaches to enhance the model.