In this study, we demonstrate both of digital and analog memory operations in InGaZnO (IGZO) memristor devices by controlling the electrode materials for neuromorphic application. The switching properties of the devices are determined by the initial energy barrier characteristics between the metal electrodes and the IGZO switching layer. Digital switching characteristics are obtained after the forming process when Schottky junction occurs at both of top and bottom electrodes. On the other hands, analog resistive switching is achieved when Schottky and Ohmic junctions exist at each side because the applied voltage modulates the Schottky barrier height through the Ohmic contact. In addition, the weightupdate properties of the devices are verified depending on identical and incremental pulse schemes. The incremental pulse trains improve the linearity and variation of weight modulation, leading to the stable learning characteristics of neuromorphic system in terms of pattern recognition with MNIST handwritten digit images.

No abstract

Non-von-Neumann computer architecture is gaining a great deal of interest for eliminating the speed bottleneck in transferring data between the processing and memory units by improving the processing parallelism. Hardware-driven neuromorphic systems are pursued actively for this goal, and they should accompany the innovations in the hardware components for higher energy efficiency. In this work, an indium gallium zinc oxide (IGZO)-based synaptic device was developed, and its synaptic behaviors were closely characterized. Processing simplicity has been improved in the structure of the device using a p+-Si bottom electrode (BE) in the Si substrate, and gradual switching characteristics have been obtained using a Pd top electrode (TE) with self-graded oxygen concentrations. By controlling the amount of oxygen atoms in depositing the switching layer, both highly linear weight adjustability and low-energy operation capability have been accomplished. In the end, the visual recognition of the IGZO synaptic device is evaluated with the Modified National Institute of Standards and Technology (MNIST) patterns in the neural network.

No abstract