Resistive switching (RS) devices are key elements for future neuromorphic computing systems as they can emulate biological synaptic functions as well as short/long term memory effects. However, there are various challenges at the device level because of their operational stability, uniformity, and device-to-device variability, which are often originated by the switching mechanisms. Here, I will discuss on how fully interface controlled memristive devices can solve these issues. I will also discuss an example of Au/Nb:STO model system, which works based on interface charge trapping/detrapping and Schottky barrier modulation mechanism. Such device also shows analog RS switching characteristics with high stability, uniformity, and programmability, making it suitable for artificial synapse and neuromorphic computing.
|