PulseExploreJournal ClubDebatesTrendingResearchersJournals
Instagram
HomeExploreJournal ClubTrending
Synapse
⌘+K
Synapse
October 11, 2025Applied Physics Letters2 citations

Polarization-oriented enhancement of ferroelectric photovoltaic performance in BiFeO3 synapses for neuromorphic vision systems

View Full Paper
RCRui ChenJLJianquan LiuYKY. Kan

Key Points

  • The [111]-oriented BiFeO3 device achieves a photocurrent and power conversion efficiency two orders of magnitude greater than [001]-oriented devices.
  • With a remnant polarization of approximately 120 μC/cm2, the [111]-oriented BFO has the highest reported among ferroelectric photovoltaic devices.
  • The device demonstrates optoelectronic long-term potentiation, enabling effective synaptic plasticity for neuromorphic computing applications.
  • Simulated arrays of the [111]-oriented BFO synapse achieve over 94% accuracy in in-sensor image recognition tasks.

Abstract

Ferroelectric photovoltaic (PV) synapses, which utilize polarization to precisely tune the photocurrent as synaptic weight, have attracted increasing interest due to their excellent polarization controllability, fast response speed, and low energy consumption. Herein, we identify a 111 polarization-oriented strategy that significantly enhances the ferroelectric PV performance of BiFeO3 (BFO) epitaxial-film devices without compromising their ferroelectricity. The photocurrent and power conversion efficiency of the 111-oriented BFO device are two orders of magnitude larger than that of 001-oriented BFO. The 111-oriented BFO device exhibits the highest remnant polarization (∼120 μC/cm2) among those reported ferroelectric PV devices, as well as excellent fatigue resistance. The high electric-field tunability of photocurrent enables 111-oriented BFO device to function as a ferroelectric PV synapse, showing optoelectronic long-term potentiation/depression synaptic plasticity. Simulated 111-oriented BFO synapse arrays achieve various neuromorphic tasks such as in-sensor image recognition with a high accuracy of more than 94% and edge feature extraction by using them as an in-sensor convolution kernel. This work demonstrates the superiority of the 111-oriented BFO synapse in neuromorphic vision systems.

Ask AI
Helpful
Bookmark
Share
View Full Paper

Cite This Study

Chen et al. (2025) studied this question.

synapsesocial.com/papers/68e9b1c1ba7d64b6fc1321b3https://doi.org/10.1063/5.0297842
Ask AI
Helpful
Bookmark
Share
View Full Paper