Thermally Stable Organic Synaptic Transistors Using a High‐ <i>T</i> <sub>g</sub> Polymer Electret
Polymer electrets are widely used as charge‐trapping layers in organic synaptic transistors (OSTs). However, common polymer electrets such as polystyrene, polyvinyl alcohol, and poly (2‐vinyl naphthalene) have glass‐transition temperatures ( T g ) below 150°C, limiting their use in high temperature processing or operating conditions. Cyclic olefin copolymer (COC) has attracted interest as a high‐ T g polymer electret owing to its excellent thermal robustness ( T g 185°C), yet its inherently low trap density restricts its applicability in polymer electret‐based OSTs. Here, the trap density of COC is effectively enhanced through UV–ozone (UVO) treatment, which introduces polar carbonyl groups and defect sites that act as additional charge‐trapping states, enabling thermally stable synaptic transistor operation. The resulting devices exhibit key synaptic behaviors, including excitatory and inhibitory postsynaptic currents, paired‐pulse facilitation and depression, short‐ to long‐term memory transition, and long‐term potentiation and depression. Optimized devices demonstrate improved linearity in synaptic weight updates with nonlinearity factors of α p = 0.59 and α d = −0.47. When implemented in a multilayer perceptron‐based artificial neural network for handwritten digit recognition, the devices achieve an accuracy of 93.2% after 100 training epochs. These results demonstrate that UVO‐treated COC is a promising high‐ T g polymer electret for thermally robust OSTs and highlight its potential for neuromorphic and artificial intelligence applications in high‐temperature environments.
Authors
- Jaeyeop Lee
- Yeyun Bae
- H. M. Cho
- ChaeGwang Lim
- Kyoungeun Lee (ORCID: https://orcid.org/0000-0002-4671-2033)
- Jiyoon Oh
- Jeongkyun Roh (ORCID: https://orcid.org/0000-0002-0674-572X)
- Moohyun Kim
- Hyeok Kim
- Bombi Kim
Institutions
- University of Seoul (KR)
- Pusan National University (KR)
Publication Details
- Journal
- Advanced Intelligent Systems
- Published
- 2026-05-29
- DOI
- https://doi.org/10.1002/aisy.202501505
- Primary Topic
- Advanced Memory and Neural Computing
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- National Research Foundation of Korea