The rise of two-dimensional materials based memtransistors for neuromorphic computing

Conventional von Neumann computing systems are increasingly constrained by latency and energy bottlenecks stemming from the physical separation of memory and processing units. Neuromorphic computing, inspired by biological neural networks, offers a paradigm shift through in-memory computing. Among various hardware candidates, memristive devices have drawn wide attention due to their non-volatile behavior, analog tunability, and compatibility with dense array integration. In particular, two-dimensional (2D) material-based memtransistors integrate resistive switching and gate modulation in a multi-terminal architecture, enabling multi-bit storage and synaptic learning essential for neuromorphic computing. This review systematically surveys recent progress in 2D material-based memtransistors for neuromorphic computing, with a focus on fabrication strategies and the underlying mechanisms that enable non-volatile and multi-level conductance modulation. Two representative switching mechanisms, sulfur vacancy migration and charge trapping, are examined in detail to clarify their roles in resistive switching behavior. The development of memtransistor architectures is then discussed, from early single-device demonstrations to densely integrated arrays, emphasizing optimization strategies such as defect engineering, Schottky barrier modulation, and heterostructure integration. Representative application domains including Kalman filter systems, artificial neural networks (ANNs), hardware security and cryptographic engineering, probabilistic computing, and edge perception and edge artificial intelligence (AI) are reviewed with illustrative examples. Finally, the review outlines current challenges in device scalability, variability control, and system-level integration, and offers perspectives for advancing toward practical and energy-efficient neuromorphic hardware.