An Electronic Synapse Device Based on TiO2 Thin Film Memristor

In order to harness and explore the power of an artificial intelligence, it is obligatory to develop a functional electronic device which can mimic the basic properties of a biological synapse. The memristor possesses passivity, nonlinearity, and in-memory computing capabilities. These properties can be used to develop a fundamental building block for the neuromorphic computing architecture. In view of this, the present paper reports the development and characterization of low-cost Ag/TiO2/FTO thin film memristor for the neuromorphic application. The developed memristor shows pinched hysteresis loop in I-V plane with +/- 2.1 V resistive switching voltage. Similar to synaptic weights of biological neurons, the continuous change in the current was observed in the developed memristor. Furthermore, non-symmetric synaptic weights are present in the device and such kinds of weights are useful for the unidirectional data flow. Moreover, the synaptic weights are incrementally increased and decreased by applying programmable potentiating and depressing pulses. The charge transportation reveals the Ohmic and space charge limited conduction (SCLC) mechanisms dominant in the hydrothermally grown TiO2 memristor device. The possible bipolar resistive switching mechanism is also discussed. In the nutshell, the results suggested that the developed device is capable of mimicking the biological synapse-like characteristics and useful for the development of synaptic devices.