SPICE Workshop on Nanomagnetism in 3D, April 30th - May 2nd 2024
Kai Liu
Interconnected magnetic nanowire networks offer a promising platform for 3-dimensional (3D) information storage and integrated neuromorphic computing. Previously, in interconnected quasi-ordered Co nanowire networks, we have demonstrated interesting magnetization reversal mechanisms as well as discrete propagation of magnetic states driven by magnetic field and current, manifested in distinct magnetoresistance (MR) features [1,2]. In a complex network with many intersections, sequential switching of nanowire sections separated by interconnects was observed, along with stochastic characteristics. The pinning/depinning of the domain walls can be further controlled by the driving current density. More recently, we have also investigated the feasibility of utilizing random magnetic nanowire networks connected by multiple electrodes as neuromorphic computing elements [3]. Multiple discrete jumps (i.e. step-by-step switching) with each electrode pair showing unique MR feature are found. Utilizing this design, diverse programming of synaptic weights may be achieved by assigning different electrode pairs as inputs/outputs and controllably switching a certain subsection of the networks by applying current pulses of varying magnitudes, pulse-widths, or repetitions. Hence, these results illustrate the promise of such interconnected networks for non-Boolean computing devices such as spintronic memristors and synaptic devices.
This work has been supported in part by the NSF (DMR-2005108 and ECCS- 2151809).
[1] E. Burks et al, Nano Lett. 21, 716 (2021).[2] D. Bhattacharya et al, Nano. Lett. 22, 10010 (2022).
[3] D. Bhattacharya et al, submitted.