Impact of the substrate thermal conductivity on switching events in VO₂multibridge structures.

Thermal management is a key factor in understanding and developing biomimetic Artificial Neural Networks (ANN) in condensed matter systems. To make ANNs as viable as possible, it is essential to work as closely as possible to the resistive transition to induce the insulator to metal transition by minimizing the switching power necessary to mimic biological neurons. Therefore, to avoid spontaneous switching and to reduce the working power, it is important develop a thorough understanding of thermal effects and heat propagation in these systems.

In this work, we study the switching of electronically and thermally coupled VO₂ multibridge devices grown by magnetron sputtering on various substrates including TiO₂(100), Al₂O₃ (1-102), Al₂O₃ (0001) and Si(100). This work aims to understand the role different substrates play in switching power and thermal behavior of multibridge devices. We observe multiple resistive states on all samples with a clear dependance on substrate thermal conductivity. Mid-Wave InfraRed temperature mapping microscopy is used to directly observe multiple switching events. Thermal measurements of heat wave propagation of multibridge devices will be presented and discussed.