Sub‐1 ms Optoelectronic Synapse Based on {ZnCdO/ZnO} Multilayer Structure for High‐Speed Neuromorphic Vision Systems
ABSTRACT The development of versatile, high‐speed optoelectronic devices is crucial for advanced sensing and neuromorphic computing. In this work, we propose highly tunable {ZnCdO/ZnO}/Si heterostructures as a dual‐function platform capable of operating as either self‐powered ultrafast photodetectors with tunable spectral range or high‐speed optoelectronic synapses. We demonstrate that the device functionality is governed by the europium (Eu) doping level. The structure with high Eu doping operates as an ultrafast, self‐powered photodetector under 650 nm illumination, exhibiting rapid rise and fall times of 5 and 7 µs, respectively, while reaching a responsivity of 90 mA W −1 for 5.7 mW cm −2 irradiance. In contrast, lowering the Eu concentration transforms the structure into a fast optoelectronic synapse. Operating in a microsecond regime, with 100 µs pulses and 500 µs intervals, the synapse achieves a paired‐pulse facilitation index of 147% under 405 nm illumination. Furthermore, applying a minor reverse bias (‐0.1 V) enables direct readout of the memory states. Crucially, unlike conventional oxide‐based devices, this synaptic behavior avoids the persistent photocurrent effect, demonstrating short‐term plasticity with a complete recovery to the zero state. These results highlight the immense potential of the proposed platform for sub‐1 ms neuromorphic vision systems and tunable photodetection.
Authors
- E. Przeździecka (ORCID: https://orcid.org/0000-0003-0385-3124)
- E. Zielony (ORCID: https://orcid.org/0000-0003-1676-3638)
- Igor Perlikowski (ORCID: https://orcid.org/0000-0003-1724-942X)
- Adrian Kaim (ORCID: https://orcid.org/0000-0001-8609-8448)
Institutions
- Institute of Physics (PL)
- Institute of Experimental Physics of the Slovak Academy of Sciences (SK)
- AGH University of Krakow (PL)
- Polish Academy of Sciences (PL)
Publication Details
- Journal
- Advanced Functional Materials
- Published
- 2026-09-17
- DOI
- https://doi.org/10.1002/adfm.78523
- Primary Topic
- Advanced Memory and Neural Computing
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- Narodowym Centrum Nauki