Resistive Switching Effect in Ag-poly(ethylene Glycol) Nanofluids: Novel Avenue Toward Neuromorphic Materials

Conventional computation techniques face challenges of deviations in Moore's law and the high-power consumption of data-centric computation tasks. Neuromorphic engineering attempts to overcome these issues by taking inspiration from neuron assemblies, ranging from distributed synaptic plasticity through orchestration of oscillator-like action potential toward avalanche dynamics. Although solid networks of nanoparticles (NPs) are proven to replicate fingerprints of criticality and brain-like dynamics, the aspect of dynamic spatial reconfigurations in the connectivity of networks remains unexplored. In this work, Ag/poly(ethylene glycol) (PEG) nanofluids are demonstrated as potential systems to mimic the spatio-temporal reconfiguration of network connections. The nanofluids are prepared by directly loading Ag NPs from the gas aggregation cluster source into liquid PEG. The NPs exhibit a negative zeta potential in PEG; if the potential difference is applied between two electrodes submerged in this nanofluid, the NPs migrate toward the anode, accumulate in its vicinity, and form a conductive path. Spikes of electric current passing through the path are detected, accompanied by resistive switching phenomena, similar to the random switching dynamics in solid NPs networks. The unique behavior of Ag/PEG nanofluids makes them promising for the realization of spatio-temporal reconfigurations in network topologies with the potential to transition to 3D. Surfactant-free metal nanofluids are prepared by direct loading of gas-phase aggregated Ag nanoparticles into poly(ethylene glycol). When biased, the nanoparticles undergo electrophoresis and form conductive paths, which pass current in a spiking manner, resembling firing processes in biological neural systems. The dynamic attachment/detachment of nanoparticles to/from the conductive path leads to a random resistive switching effect.image

Automated summary

Neuromorphic engineering attempts to overcome challenges in conventional computation techniques by taking inspiration from neuron assemblies. In this study, Ag/poly(ethylene glycol) (PEG) nanofluids are used to mimic the spatio-temporal reconfiguration of network connections. The unique behavior of Ag/PEG nanofluids shows promise for transitioning to 3D network topologies.