BROWSE
Detail
Conference Paper
Accelerated Inverse Design Framework for High Efficiency RGB Tandem OLEDs via Physics Embedded General Artificial Neural Network
Cited - time in 
Share 
Share
- Authors
- Jun Hee Han, Kyu-sung Lee
- Issue Date
- 2026-05
- Citation
- Society for Information Display (SID) International Symposium 2026, pp.1150-1153
- Publisher
- Wiley
- Language
- English
- Type
- Conference Paper
- Abstract
- Conventional grid-search methods for optimizing the thickness of high-efficiency, multi-layered tandem organic light-emitting diodes demand an astronomical amount of computational time. Furthermore, data-driven artificial neural network approaches face an inefficiency problem due to the need for collecting vast amounts of training data. To address these issues, this study proposes an inverse design model that directly integrates a physical formula into the artificial neural network structure. The proposed model entirely bypasses the data training process, allowing for optimization solely through gradient-descent-based iterations. As a key result, the model successfully determined the optimal thicknesses for complex multi-emitting-layer tandem device structures in a time dramatically reduced compared to conventional grid search. Moreover, the model can be flexibly applied to various structures through active modification without requiring additional training, securing the practicality of reflecting manufacturing constraints in the design. The methodology presented in this study offers a groundbreaking approach that maximizes the temporal efficiency, versatility, and practicality essential for the tandem structured OLEDs.
- Keyword
- Tandem, Inverse design, Artificial neural network, Organic light emitting diode, High efficiency