CuSe Memristor-Driven Neuromorphic Architecture: Synaptic Dynamics, Logic Reconfigurability, and Unclonable Dual-Key...

Neuromorphic computing leveraging synaptic memristors has emerged as a cutting-edge approach to overcome the limitations of conventional computing architectures and address the bottleneck in information security. However, exploring emerging circuit designs based on memristors and their potential applications in unconventional fields hold critical significance for accomplishing tasks that are challenging for traditional electronic devices. Herein, the CuSe is employed as the functional layer to assemble Ag/CuSe/ITO structured synaptic memristor, focusing on their synaptic behaviors, multi-level storage capabilities, and applications for logic circuit design and information security encryption. The device exhibits a high switching ratio (> 10³), excellent durability (> 10⁴ s), power consumption (∼10^-8 J), and diverse synaptic plasticity simulations (EPSC, LTP/LTD, PPF/PPD, STDP). Furthermore, using a single pair of such CuSe memristors without any reconfiguring connections, the four fundamental multi-functional logic operations (AND, OR, NAND, NOR) are effectively realized. This forms the Minimal Boolean Basis and enables all 16 Boolean operations, establishing a complete set of reconfigurable logic functions. Particularly, a dual-key unclonable cryptosystem is successfully designed for realizing hardware-level encryption/decryption of digital signals, images, and audio data. This work successfully achieves the integration of memristors with logic circuits and widens a novel way for advancements in information security.