Modeling of Conduction Mechanism in Filament-Free Multilayer Bulk RRAM

Filament-free bulk RRAM (resistive-switching random access memory) devices have been proposed to offer multi-level conductance states with less variations and noise and forming-free operation for neuromorphic computing applications. Understanding conduction mechanism and switching dynamics of filament-free bulk RRAM devices is crucial to optimize device characteristics and to build large-scale arrays for compute in memory and neuromorphic computing applications. Here, we first analyze switching characteristics of bulk RRAM by temperature dependent I-V measurements. We then present a quantitative physical model describing the conduction across trilayer stack by a series combination of multiple conduction mechanisms across each layer. Using this model and fitting it to the experimental characteristics of filament-free bulk RRAM devices, we investigate the origin of bulk switching in trilayer stacks. We demonstrate that our model can be used as a guide to design bulk switching RRAM devices from multilayer stacks of metal oxides.