Optically-driven nine-state memory and pattern recognition in a lead-free Tin-doped Cs2AgBiBr6 memristor for neuromorphic visual system
Abstract
Imprinting vision as memory is essential for brain-inspired computing, requiring devices that integrate memory and processing. However, this technique development is hindered by limited memory states and segregated functionalities. In this work, a lead-free Sn-doped Cs 2 AgBiBr 6 (Cs 2 Ag 0.85 Sn 0.15 BiBr 6 ) optoelectronic memristor is proposed to integrate high-density memory and visual perception capabilities. For the first time, the nine-state memory driven by a 450 nm light at different intensities was achieved in the Cs 2 AgBiBr 6 -based device. These states remained stable for over 2000 s and through 620 cycles, demonstrating robust high-density memory capability and guaranteeing the accuracy of visual simulation. Random letter recognition and color image processing were performed by logic operations (AND, XOR) on the optically modulated resistance states. Key synaptic functions were also emulated, enabling a “learning-forgetting-relearning” process. Fundamentally, the optimal 15 % Sn doping at the Ag site narrowed the bandgap of films and increased the bromine vacancy concentration, enhancing the optically controlled memristive effect. The Schottky-like barrier at the Au/Cs 2 Ag 0.85 Sn 0.15 BiBr 6 interface was modulated by photo-induced carriers, which further influenced the multilevel memristive behavior. Our study offers a promising strategy for enabling synergistic sensing-memory-computing in a standalone device for human-machine interaction systems with visual data storage, buffering, and processing capabilities.
Fields of Study & Subjects
Advanced Memory and Neural Computing
Phase-change materials and chalcogenides
Neural Networks and Reservoir Computing
Publication details
- Journal / source
- Materials Today Communications
- Publisher
- Elsevier BV
- Volume
- 51
- Pages
- 114742–114742
- Publication date
- 28 January 2026
- Language
- EN