Machine learning glass transition temperature of styrenic random copolymers.

Area of Science:

• Materials Science
• Polymer Chemistry
• Computational Chemistry

Background:

• The glass transition temperature (Tg) is a critical thermophysical property for styrenic random copolymers.
• Experimental determination of Tg can be challenging and time-consuming.
• Accurate Tg prediction is essential for material design and application.

Purpose of the Study:

• To develop a data-driven model for predicting the glass transition temperature (Tg) of styrenic random copolymers.
• To establish a statistical relationship between quantum chemical descriptors and Tg.
• To offer a robust and efficient alternative to experimental Tg measurements.

Main Methods:

• Utilized Gaussian process regression (GPR), a machine learning technique.
• Employed quantum chemical descriptors as input features.
• Trained and validated the model using 48 experimentally determined Tg values for styrenic random copolymers.

Main Results:

• The GPR model demonstrated high accuracy and stability in predicting Tg.
• The model successfully captured the complex relationship between molecular descriptors and Tg.
• The predicted Tg values ranged from 246 K to 426 K, covering a wide spectrum of copolymers.

Conclusions:

• Gaussian process regression provides a powerful and reliable tool for estimating Tg in styrenic random copolymers.
• This data-driven approach offers significant advantages in terms of speed and cost-efficiency.
• The developed model serves as a promising tool for accelerating materials discovery and development.