Bridging scales: machine learning for the rational design and modelling of shape memory polymers

Area of Science:

• Materials Science
• Polymer Science
• Computational Science

Background:

• Shape memory polymers (SMPs) are stimuli-responsive materials with broad applications.
• Traditional SMP design relies on empirical methods, limiting discovery speed.
• Predicting SMP behavior often requires complex theoretical mechanics.

Purpose of the Study:

• To critically review recent advances in applying machine learning (ML) to SMPs.
• To discuss ML-assisted SMP design, chemistry identification, and property prediction.
• To outline future directions for ML in SMP research.

Main Methods:

• Review of current literature on ML applications in SMPs.
• Discussion of ML techniques for SMP chemistry identification.
• Analysis of ML for predicting thermo-mechanical and shape memory properties.

Main Results:

• ML shows promise in accelerating SMP discovery and property prediction.
• Various ML tools have been employed to identify new SMP chemistries.
• ML aids in predicting thermo-mechanical and shape memory properties of SMPs.

Conclusions:

• ML-assisted SMP discovery and modeling are in early stages.
• Current limitations include incomplete structural representations and challenges integrating thermal/temporal effects.
• Future work should focus on advanced ML tools for complex SMP topologies and polymer-specific networks.