Effects of switching layer morphology on resistive switching behavior: A case study of electrochemically synthesized mixed-phase copper oxide memristive devices

Resistive switching (RS) behavior can serve as a building block in the development of non-volatile memory and neuromorphic computing applications. Thus far, various device parameters have been modulated appropriately to achieve the desired characteristics from RS devices. However, no clear guideline for device parameter modulation has been reported. Herein, we systematically investigate the effect of the switching layer morphology and thickness, as well as the choice of the bottom electrode, on RS properties by using electrochemically synthesized mixed-phase copper oxide (Cu_xO) as a model material for memristive devices. By controlling various electrochemical parameters, we have fabricated Cu_xO switching layers with various morphologies (microcrystal, microcubic, compact thin film, short dendritic nanowire, granular, and nanoparticle). Interestingly, the Cu_xO-based RS devices fabricated by means of electrodeposition (Cu_xO/FTO) exhibit the forming-free digital RS property, which is suitable for non-volatile memory, whereas those fabricated by utilizing anodization (Cu_xO/Cu) exhibit the analog RS property, which makes them suitable for use in synaptic learning applications. A convolutional neural network (CNN) was implemented using experimental synaptic weights of the anodized Cu_xO RS device for image edge detection application. In addition, conduction mechanisms and possible RS mechanisms are suggested for both types of devices. These results indicate that the electrochemically synthesized switching layers are promising for use in both non-volatile memory and neuromorphic computing applications.