Epoxy Adhesive Materials as Protective Coatings: Strength Property Analysis Using Machine Learning Algorithms

Area of Science:

• Materials Science
• Polymer Chemistry
• Chemical Engineering

Background:

• Epoxy adhesives are crucial functional coatings.
• Understanding filler and curing agent effects on mechanical properties is vital.
• Optimizing epoxy formulations requires detailed analysis.

Purpose of the Study:

• To investigate the impact of physical modifications on epoxy adhesive mechanical properties.
• To explore the use of mineral, active, and nanostructured fillers.
• To develop predictive machine learning models for epoxy performance.

Main Methods:

• Formulation of epoxy resins (Epidian 5, 53, 57) with various fillers (calcium carbonate, activated carbon, nanoclays).
• Curing epoxy resins using TFF, Z-1, and PAC agents.
• Experimental testing of mechanical properties (tensile, compressive, bending strength).
• Application of machine learning algorithms and Shapley analysis for property prediction.

Main Results:

• Calcium carbonate (10-20 wt%) in Epidian 5/53 resins cured with TFF/Z-1 yielded optimal strengths: up to 64 MPa (tensile), 145 MPa (compressive), and 123 MPa (bending).
• Activated carbon and nanofillers provided moderate improvements, especially in flexible matrices.
• Machine learning models achieved high accuracy (R² 0.93-0.95) for compressive and bending strength prediction.
• Shapley analysis identified curing agents and fillers as significant predictive features.

Conclusions:

• Calcium carbonate is a highly effective filler for enhancing epoxy adhesive mechanical properties.
• Machine learning provides a powerful tool for optimizing epoxy formulations and predicting performance.
• Further research is needed to improve tensile strength prediction models.