记忆电阻器
材料科学
计算机科学
电子工程
纳米技术
能源消耗
电气工程
神经形态工程学
光电子学
记忆晶体管
泄漏(经济)
电阻随机存取存储器
脉搏(音乐)
CMOS芯片
逻辑门
功率消耗
频道(广播)
能量(信号处理)
作者
Bingqi Cai,Chen Wang,Kun Chen,Qingqing Sun,David Wei Zhang,Lin Chen
标识
DOI:10.1109/led.2025.3614142
摘要
To adapt to diverse tasks under varying operational environments, memristor devices are usually required to emulate multiple biological functions. However, integrating such diverse functionalities within a single self-rectifying memristor remains a significant challenge. In this work, we report a self-rectifying memristor based on a Pt/TiOxNy/HfO2/W structure. For the first time, we demonstrate that a single device can emulate both leaky integrate-and-fire (LIF) neuron and nociceptor behaviors. Under pulse stimulation with a 10 μs width, the device functions as an LIF neuron, while increasing the pulse width to 100 μs or longer enables nociceptor-like responses. Notably, the LIF neuron operation exhibits ultra-low energy consumption of only 0.8 fJ/spike. In addition, the device exhibits a high rectifying ratio (>10⁵), large switching ratio (>10⁴), and ultra-low leakage current (~100 pA). This multifunctional self-rectifying memristor offers a promising pathway toward compact neuromorphic systems by enabling simplified device architectures with integrated perceptual and computational capabilities.
科研通智能强力驱动
Strongly Powered by AbleSci AI