Using Machine Learning to Predict and Understand Complex Self-Assembly Behaviors of a Multicomponent Nanocomposite

Area of Science:

• Materials Science
• Nanotechnology
• Computational Science

Background:

• Multicomponent nanocomposites self-assemble into devices with structure-dependent properties.
• Predicting the relationship between formulation and assembled structure is challenging due to complexity and non-equilibrium states.
• Current design relies on inefficient experimental trial-and-error.

Purpose of the Study:

• To explore the application of machine learning (ML) for predicting nanocomposite structures.
• To investigate correlations between input parameters and output structures in nanocomposite thin films.
• To demonstrate ML's potential in improving nanocomposite device design efficiency.

Main Methods:

• Trained ML models using a dataset of 595 microscopy images of nanocomposite thin films.
• Examined correlations between input and output parameters.
• Utilized the best-performing ML model to predict structures of new compositions.

Main Results:

• Identified correlations providing new insights into nanocomposite systems.
• Successfully predicted structures of novel nanocomposite compositions.
• Demonstrated ML's capability to map high-dimensional spaces for complex systems.

Conclusions:

• ML techniques can significantly enhance the efficiency of nanocomposite device design.
• ML offers a viable alternative to traditional trial-and-error methods for complex materials.
• The study highlights the advantages and challenges of applying ML to complex systems.