Ultra‐Large Modulation Range of Synaptic Plasticity in 2D Homologous Ti ₃ C ₂ T <i> _x </i> ‐TiO <i> _x </i> Heterosystem to Realize Diverse Synapses

Abstract With intelligent science advancing rapidly, von Neumann's architectural limits are evident, and memristor‐based synapses offer a key breakthrough. One of the core functions of biomimetic neural networks is to realize the forgetting function (synaptic plasticity regulation) by accurately simulating diverse synaptic behaviors, such as short‐term plasticity, long‐term plasticity, and spike‐timing‐dependent plasticity, observed in biological systems. In this study, volatile synaptic memristors (FTO/Ti 3 C 2 T x :TiO x /Ag) are fabricated and non‐volatile synaptic memristors (FTO/TiO x /Ag) based on the homologous Ti 3 C 2 T x (MXene) aerogel material. The volatile device exhibited fast dynamic response behavior similar to electrical synapses in the I–V characteristic curve sweeps, with a tunable conductance change rate reaching up to 2269% (potentiation) and 280% (depression), respectively. Additionally, under pulse conditions, the device demonstrates excellent Spike‐Timing‐Dependent Plasticity (STDP) characteristics, achieving significant weight change within the frequency range of 2–10 kHz. The non‐volatile device exhibits behavior resembling chemical synapses and maintains good stability over multiple sweeps. Furthermore, pulse relaxation tests of both devices demonstrate the transition from short‐term memory (STM) to long‐term memory (LTM), with the relaxation time change ratio of the volatile device reaching 2.35 × 10 4 . This study successfully simulates both chemical synapses and electrical synapses with an ultra‐large plasticity tuning range using MXene‐based devices.