On-demand optical printing of Ag-based neuromorphic devices for brain-inspired learning

Abstract

Neuromorphic computing, where memory and computation occur in the same physical system, promises to overcome the von Neumann bottleneck by mimicking the massively parallel architecture of biological neural networks. Self-formed networks with disordered, nonlinear connectivity are especially promising for synaptic plasticity, but fabrication often demands vacuum deposition or complex chemical processing, limiting broader accessibility and adoption. Here, we present a rapid, on-demand optical printing of Ag nanoparticles assembly to form a continuous film followed by annealing method, which spontaneously transform into disconnected Ag nanostructures with tunable morphologies. By systematically varying printing time and annealing conditions, we realize Ag-based neuromorphic devices showing reliable resistive switching, short-and long-term potentiation, and, importantly, arousal-dependent performance, classically described by the Yerkes-Dodson law, all achieved without external CMOS circuitry. This approach offers a reproducible, cost-effective, and scalable route to self-organized Ag networks, bypassing the constraints of traditional vacuum or chemically intensive methods. These devices, based on emergent architectures, open up pathways toward scalable, adaptive hardware for advanced neuromorphic computing.