Mimicking the Biological Synapse Functions of Analog Memory, Synaptic Weights, and Forgetting with ZnO-Based Memristive Devices

The present work demonstrates a simple and cost effective way to fabricate ZnO memristive device for a neuromorphic application using the single-step hydrothermal method. The structural, morphological, and electrical characterizations were carried out using X-ray diffraction, scanning electron microscopy, and programmable electrochemical workstation respectively. The effect of active layer thickness on the analog memory, synaptic weights, and resistive switching voltages of Al/ZnO/SS memristive devices were investigated. It is observed that ZnO memristive device possess thickness dependent synaptic property. The asymmetric resistive switching voltages are observed for all devices and magnitude of V _SET is dependent on the active layer thickness. The mixed blend of Ohmic and trap-controlled space charge limited conduction mechanisms are observed in developed devices. The notable ‘memristor type-II’ like characteristics and coexistence of memristance and meminductance memory effects are observed in some devices, which shows poor performance for the neuromorphic application. The forgetting curve of the ZnO memristive device is in good agreement with the human forgetting curve and show rapid decay in initial stage followed by long and slow decay. The results indicate that the hydrothermally grown ZnO memristive device is a potential candidate for future neuromorphic devices.