Smart phosphor with neuromorphic behaviors enabling full-photoluminescent Write and Read for all-optical physical reservoir computing

exhibits adaptive photoluminescence variations with neuromorphic characteristics. Theoretical simulations reveal interface-mediated halogen migration processes with progressively evolving energy barriers, underpinning the neuron-like photoluminescence property variations. The system enables full photoluminescence-based Write and Read functionalities for all-optical neuromorphic computing, achieving 4-bit binary sequence discrimination as physical reservoirs. It further demonstrates potential in photoluminescence-based fingerprint authentication with 94.4% accuracy. This work advances smart phosphor as an alternative approach to neuromorphic computing with optical-stimuli and optical-output. It also opens avenues for designing function-oriented phosphor materials with tailored properties for information science and artificial intelligence applications.