Threshold Voltage and Memory Window Engineering of
Surface Modified Ti₃C₂T_x Mxene Memristor for Memory and Neuromorphic Computing

Mxenes, a new type of two-dimensional material, have attracted extensive interest in memory and artificial synaptic device application because of its excellent electrical and chemical properties. In this work, the conductivity of Mxenes is engineered to enhance the switching of the device with addition of functional groups, via the etching and oxidation processes. The exceptional properties of partially oxidized MXene (Ti₃C₂T_x) memristors have shown a large memory window and decreased threshold bias; on the nanosecond scale, we emulated the complex spike time-dependent plasticity-dependent synaptic rules using electrical pulses. The low threshold voltage, steady retention time (10⁴ s), clearly distinguishable resistance states and high ON/OFF rate (>10⁶) are the main memory-related attributes of this device. Furthermore, the physical origin of electronic transport in this device is found to occur by a filamentary resistive switching mechanism, as determined by analyzing the atomic/kelvin probe force microscopy and modelled electrical fittings. The image edge detection ability of the synaptic device is also revealed by using a convolutional neural network. Therefore, the present functionalization method for Mxene memristors can facilitate the essential manufacturing complexity of high-density non-volatile memory storage and artificial synapse system.