Low‐Temperature Processed Optoelectronic Synapses With Enhanced Responsivity for Edge‐Oriented In‐Sensor Reservoir Computing
Fuente:
Wiley Open Access
Gespeichert in:
| Hauptverfasser: | , , , |
|---|---|
| Format: | Artículo Open Access |
| Veröffentlicht: |
Wiley
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1867021134697857024 |
|---|---|
| author | Jo‐Lin Chen Shi‐Qian Guo Yu‐Chuan Chiu Po‐Tsun Liu |
| author_facet | Jo‐Lin Chen Shi‐Qian Guo Yu‐Chuan Chiu Po‐Tsun Liu Jo‐Lin Chen Shi‐Qian Guo Yu‐Chuan Chiu Po‐Tsun Liu |
| collection | Wiley Open Access |
| contents | Low‐Temperature Processed Optoelectronic Synapses With Enhanced Responsivity for Edge‐Oriented In‐Sensor Reservoir Computing Jo‐Lin Chen Shi‐Qian Guo Yu‐Chuan Chiu Po‐Tsun Liu Advanced Materials Technologies ABSTRACT Neuromorphic systems that integrate sensing and computation are essential for enabling energy‐efficient artificial intelligence. In this study, we proposed a hybrid supercritical fluid (HSCF) treatment that integrates supercritical fluid and thermal annealing methods to fabricate high‐performance InGaZnO (IGZO) synaptic transistors for in‐sensor reservoir computing (ISRC) applications. The HSCF method enables low‐temperature processing, thereby enhancing its compatibility with flexible substrates while simultaneously passivating oxygen vacancies and hydroxyl‐related defects. Under stimulation with 400 nm light, the devices demonstrate a superior signal‐to‐noise ratio of 4.3 × 10 4 , optical responsivity of 4.7 × 10 3 A W −1 , and a range of synaptic plasticity, including paired‐pulse facilitation, long‐term potentiation/depression, and spike‐dependent plasticity. These characteristics suggest their strong potential for integrating into artificial neural network architectures. Moreover, the ISRC network trained on HSCF‐treated devices achieves 93.9% recognition accuracy in handwritten digit classification. In object recognition tasks, it achieves 89.4% accuracy by leveraging dynamically adjustable noise‐filtering thresholds. These results demonstrate that HSCF‐treated IGZO synapses serve as a potential candidate for integrating edge AI technologies into next‐generation flexible electronic systems. 10.1002/admt.70950 http://creativecommons.org/licenses/by/4.0/ |
| doi_str_mv | 10.1002/admt.70950 |
| format | Artículo Open Access |
| id | wiley_oa_10_1002_admt_70950 |
| institution | Wiley Open Access |
| license_str_mv | http://creativecommons.org/licenses/by/4.0/ |
| publishDate | 2026 |
| publisher | Wiley |
| record_format | wiley_oa |
| spellingShingle | Low‐Temperature Processed Optoelectronic Synapses With Enhanced Responsivity for Edge‐Oriented In‐Sensor Reservoir Computing Jo‐Lin Chen Shi‐Qian Guo Yu‐Chuan Chiu Po‐Tsun Liu Advanced Materials Technologies Low‐Temperature Processed Optoelectronic Synapses With Enhanced Responsivity for Edge‐Oriented In‐Sensor Reservoir Computing Jo‐Lin Chen Shi‐Qian Guo Yu‐Chuan Chiu Po‐Tsun Liu Advanced Materials Technologies ABSTRACT Neuromorphic systems that integrate sensing and computation are essential for enabling energy‐efficient artificial intelligence. In this study, we proposed a hybrid supercritical fluid (HSCF) treatment that integrates supercritical fluid and thermal annealing methods to fabricate high‐performance InGaZnO (IGZO) synaptic transistors for in‐sensor reservoir computing (ISRC) applications. The HSCF method enables low‐temperature processing, thereby enhancing its compatibility with flexible substrates while simultaneously passivating oxygen vacancies and hydroxyl‐related defects. Under stimulation with 400 nm light, the devices demonstrate a superior signal‐to‐noise ratio of 4.3 × 10 4 , optical responsivity of 4.7 × 10 3 A W −1 , and a range of synaptic plasticity, including paired‐pulse facilitation, long‐term potentiation/depression, and spike‐dependent plasticity. These characteristics suggest their strong potential for integrating into artificial neural network architectures. Moreover, the ISRC network trained on HSCF‐treated devices achieves 93.9% recognition accuracy in handwritten digit classification. In object recognition tasks, it achieves 89.4% accuracy by leveraging dynamically adjustable noise‐filtering thresholds. These results demonstrate that HSCF‐treated IGZO synapses serve as a potential candidate for integrating edge AI technologies into next‐generation flexible electronic systems. 10.1002/admt.70950 http://creativecommons.org/licenses/by/4.0/ |
| title | Low‐Temperature Processed Optoelectronic Synapses With Enhanced Responsivity for Edge‐Oriented In‐Sensor Reservoir Computing |
| topic | Advanced Materials Technologies |
| url | https://advanced.onlinelibrary.wiley.com/doi/10.1002/admt.70950 |