EUVIMEDEUVIMED entdecken
Crossref

83‐2: Sync‐OLED‐Enabled Physics‐Informed AI for High‐Throughput Screening of OLED Stacks

Sang Ho Jeon, Jisong Chae, Minsoo Kim, Sunmi Choi, Hochul Song, Il Joon Kang, Junghoon Kim, Byunghoon Chun

SID Symposium Digest of Technical Papers · 2026-06 · 57 (1) · 1146-1149

DOI: 10.1002/sdtp.70001_310 ↗

Kurz gefasst

Fachartikel in der Zeitschrift SID Symposium Digest of Technical Papers, online erschienen am 11.09.2026. Verfasst von Sang Ho Jeon und 7 weiteren Autor:innen. Die Arbeit nennt 4 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The development of organic light‐emitting diodes (OLEDs) is increasingly limited by the combinatorial explosion of material stacks: even a five‐layer architecture (HTL, EBL, EML, HBL, ETL) yields more than 10 10 possible combinations, far beyond what conventional trial‐and‐error experiments can evaluate within realistic time and cost. We address this bottleneck with a physics‐informed, simulation‐first workflow that quantitatively links material properties to device performance and is scalable to full‐space screening. At its core is the Sync‐OLED platform, which constructs amorphous bulk morphologies and standardizes solid‐state descriptors such as energetic disorder (σ), carrier‐ and field‐dependent mobility, and optical properties, thereby moving beyond simple gas‐phase HOMO/LUMO alignment. These properties feed a TCAD‐based digital twin that implements OLED‐specific transport and recombination models, distinct from conventional inorganic device physics, capturing hoppinglimited charge transport and exciton dynamics to predict current– voltage–luminance characteristics, efficiency, and lifetime with high fidelity, though still too slowly for exhaustive exploration. To enable practical screening, we build a proxy AI surrogate trained to emulate the physical simulator and refined via active learning, overcoming the usual limitation that “AI cannot predict what it has not seen.” This physics‐informed surrogate reproduces the accuracy and trends of the digital‐twin simulator while providing four to five orders of magnitude acceleration, enabling performance prediction in less than 1 μs per stack. As a result, we can screen more than 10 10 OLED stacks within 24 hours and identify rare regions that simultaneously enhance efficiency and lifetime, offering a generalizable framework for highdimensional optimization of advanced display and optoelectronic devices.

Zitieren

APA

Jeon, S. H., Chae, J., Kim, M., Choi, S., Song, H., Kang, I. J., Kim, J., & Chun, B. (2026). 83‐2: Sync‐OLED‐Enabled Physics‐Informed AI for High‐Throughput Screening of OLED Stacks. SID Symposium Digest of Technical Papers, 57(1), 1146-1149. https://doi.org/10.1002/sdtp.70001_310

Vancouver

Jeon SH, Chae J, Kim M, Choi S, Song H, Kang IJ, et al. 83‐2: Sync‐OLED‐Enabled Physics‐Informed AI for High‐Throughput Screening of OLED Stacks. SID Symposium Digest of Technical Papers. 2026;57(1):1146-1149. doi:10.1002/sdtp.70001_310

BibTeX
@article{Jeon2026,
  author = {Jeon, Sang Ho and Chae, Jisong and Kim, Minsoo and Choi, Sunmi and Song, Hochul and Kang, Il Joon and Kim, Junghoon and Chun, Byunghoon},
  title = {83‐2: Sync‐OLED‐Enabled Physics‐Informed AI for High‐Throughput Screening of OLED Stacks},
  journal = {SID Symposium Digest of Technical Papers},
  year = {2026},
  volume = {57},
  number = {1},
  pages = {1146--1149},
  doi = {10.1002/sdtp.70001_310}
}
Kostenlos starten →