Biomimetic halide perovskite synapses for neuromorphic computing

Driven by the exponential surge in computational demands, there is a pressing imperative to pioneer novel information processing paradigms that overcome the inherent von Neumann bottleneck in conventional computing architectures. Neuromorphic computing based on biomimetic synapses aims to emulate the structure and working principles of the brain, offering a transformative alternative approach. Halide perovskites are a promising choice for biomimetic synapses due to their unique intrinsic ability to concurrently manipulate both ions and electronic charges. Their tunable properties across various timescales unlock unprecedented opportunities for advancing neuromorphic computing applications. In this review, we survey recent advances in biomimetic synapses based on halide perovskites in terms of design principles, material properties, device architectures, and mechanisms. An overview of potential applications in neuromorphic computing is provided, and the most critical challenges in hardware implementation are discussed. We also propose a roadmap to guide future innovations in scaling biomimetic synaptic hardware to practical systems. • Fundamental operating mechanisms of halide perovskite synapses are summarized, highlighting the ion-photo/electron carrier synergistic regulation characteristics. • Advances in halide perovskite synapses for near-sensor and in-sensor neuromorphic computing applications is presented. • Key challenges in halide perovskite-based hardware neuromorphic systems are discussed, accompanied by forward-looking perspectives on future development.