Related Experiment Video
Updated: Mar 19, 2026
09:49
In Situ Transmission Electron Microscopy with Biasing and Fabrication of Asymmetric Crossbars Based on Mixed-Phased a-VOx
Published on: May 13, 2020
4.4K
Programmable ferroelectric rectifier for reliable and efficient neuromorphic crossbar array
Youngmin Kim1, Yoon Jung Lee1,2,3, Jiwoong Yang4
1Department of Materials Science and Engineering, Research Institute of Advanced Materials, Seoul National University, Seoul, Republic of Korea.
Nature Communications
|March 18, 2026
Summary
Researchers developed novel memristor devices using barium-substituted bismuth ferrite (BBFO) for energy-efficient neuromorphic computing. These devices exhibit excellent rectification and synaptic behavior, paving the way for advanced computing hardware.
Area of Science:
- Materials Science
- Solid State Physics
- Computer Engineering
Background:
- Neuromorphic computing hardware demands energy efficiency, scalability, and reliable conduction.
- Ferroelectric materials in memristors offer a promising solution through electrostatically modulated conduction.
Purpose of the Study:
- To develop and characterize memristor devices using functional ferroelectric materials for neuromorphic applications.
- To investigate the role of material composition in achieving high-performance memristive behavior.
Main Methods:
- Fabrication of epitaxial heterostructures: Pt/Ba₀.₂Bi₀.₈FeO₃ (BBFO)/SrRuO₃/SrTiO₃.
- Characterization of memristor devices for electrical properties, including rectifying ratio and off-state current.
- Analysis of ferroelectric polarization and oxygen vacancy migration under pulsed bias.
Main Results:
- Memristor devices achieved a rectifying ratio > 10⁶ and off-state current < 10⁻¹² A.
- 20% Ba substitution in BiFeO₃ coupled ferroelectric polarization with oxygen vacancy migration, enabling reliable synaptic behavior (>10⁷ cycles, near-zero nonlinearity).
- Selector-free crossbar arrays (CBAs) were implemented, mitigating sneak currents and cell variability.
Conclusions:
- BBFO-based memristors offer a robust material platform for high-performance neuromorphic systems.
- The developed devices address key challenges in energy-efficient and scalable neuromorphic hardware.
- BBFO's unique properties enable selector-free CBA implementation for advanced computing.
