Many advanced electronic devices – such as OLEDs, batteries, solar cells, and transistors – rely on complex multilayer architectures composed of multiple materials. Optimizing device performance, stability, and efficiency requires precise control over layer composition and arrangement, yet experimental exploration of new designs is costly and time-intensive. Although physics-based simulations offer insight into individual materials, they are often impractical for full device architectures due to computational expense and methodological limitations.

Schrödinger has developed a machine learning (ML) framework that enables users to predict key performance metrics of multilayered electronic devices from simple, intuitive descriptions of their architecture and operating conditions. This approach integrates automated ML workflows with physics-based simulations in the Schrödinger Materials Science suite, leveraging physics-based simulation outputs to improve model accuracy and predictive power. This advancement provides a scalable solution for rapidly exploring novel device design spaces – enabling targeted evaluations such as modifying layer composition, adding or removing layers, and adjusting layer dimensions or morphology. Users can efficiently predict device performance and uncover interpretable relationships between functionality, layer architecture, and materials chemistry. While this webinar focuses on single-unit and tandem OLEDs, the approach is readily adaptable to a wide range of electronic devices.