Self‐Powered Bidirectional UV Optoelectronic Synapses Based on Competitive Schottky‐Photogating Dynamics Toward Enhanced Neuromorphic Motion Recognition
ABSTRACT Bidirectional photo‐responsive devices, which generate polarity‐switchable photocurrent signals, have attracted increasing attention for applications in neuromorphic vision. However, achieving both self‐powered operation and bidirectional photoresponse in a single device remains a major challenge. In this work, a self‐powered Au/Sr 2 Nb 3 O 10 /MXene optoelectronic synapse is demonstrated, realizing wavelength‐regulated bidirectional synaptic modulation within a single two‐terminal device. Owing to a competition between a fast junction‐associated photoresponse and slower surface oxygen dynamics, the device exhibits bidirectional photocurrent switching behavior. Under 365 nm illumination, a temporal bipolar response is observed, where the photocurrent evolves from an initial negative value to a positive state with prolonged irradiation. In contrast, under 254 nm illumination, a negative transient photoresponse is observed without the temporal polarity reversal found at 365 nm. Benefiting from this self‐powered wavelength‐dependent opposite‐polarity photoresponse, the device suggests potential for denoising‐enhanced motion recognition in neuromorphic vision. This work not only reveals the origin of bidirectional photoresponse governed by the interaction between Schottky barriers and surface oxygen dynamics, but also highlights the importance of single‐device platforms that integrate self‐powered operation and wavelength‐tunable bidirectional photoresponse for efficient neuromorphic vision.
Authors
- Wentao Xu (ORCID: https://orcid.org/0000-0002-1054-6037)
- Jiaxin Chen (ORCID: https://orcid.org/0000-0003-4459-5408)
- Yanhui Li (ORCID: https://orcid.org/0000-0003-0185-0528)
- Jiayi Li
- Yong Zhang
- Linglong Wu (ORCID: https://orcid.org/0009-0009-8653-4946)
Institutions
- Nankai University (CN)
- Suzhou University of Technology (CN)
Publication Details
- Journal
- Advanced Functional Materials
- Published
- 2026-10-09
- DOI
- https://doi.org/10.1002/adfm.78869
- Primary Topic
- Advanced Memory and Neural Computing
- Type
- article
- Field-Weighted Citation Impact
- 0.00