Novel high voltage polymer insulators using computational and data-driven techniques

Area of Science:

• Materials Science
• Computational Chemistry
• Electrical Engineering

Background:

• Conventional polymer insulators degrade under high voltage due to space charge accumulation, limiting their application.
• Identifying suitable high-voltage insulating materials is a critical bottleneck in electrical system development.

Purpose of the Study:

• To accelerate the discovery of novel polymers for high-voltage insulation.
• To enable rapid prediction of key properties: bandgap (Egap) and electron injection barrier (Φe).

Main Methods:

• Utilized density functional theory (DFT) to generate large datasets of Egap and Φe.
• Employed Bayesian calibration to reconcile computed properties with experimental data.
• Developed machine learning models for rapid property estimation.

Main Results:

• Successfully predicted Egap and Φe for 13,000 polymers.
• Identified and recommended polymers with high Egap and Φe as potential high-voltage insulators.
• Deployed predictive models on www.polymergenome.org for community access.

Conclusions:

• Machine learning models significantly accelerate the identification of advanced polymer insulators.
• The developed methodology and dataset facilitate the design of next-generation high-voltage electrical systems.
• Open-access models empower researchers to explore and discover new insulating materials efficiently.