Material‐Intrinsic Circularly Polarized Light Encryption via Neuromorphic Transistors

ABSTRACT

Brain‐inspired neuromorphic computing provides an efficient paradigm for integrating sensing, memory, and processing, offering new opportunities for secure optical information systems. Circularly polarized light (CPL) is an attractive carrier for information encryption. However, existing CPL‐based systems are fundamentally constrained by the lack of chiral semiconductors with broadband, consistently high absorption dissymmetry factor ( g _abs ), and by encoding strategies that rely on externally defined optical parameters. Here, we propose a material‐centric encryption paradigm enabled by newly designed heterochiral molecules that exhibit uniformly high g _abs (>0.01) across nearly all bisignate circular dichroism bands. This broadband and intrinsic chiroptical response allows CPL to function as a robust, high‐dimensional encoding variable. Leveraging this capability, we develop a multi‐wavelength CPL encryption and decryption framework based on chiral optoelectronic synaptic transistors that integrate sensing, memory, and processing within a single device. Compared to conventional single‐wavelength schemes, our approach significantly enhances encryption performance, achieving a 21% increase in entropy, a 134% improvement in normalized signal‐to‐noise ratio, and a 33% enhancement in signal difference maximization index. An artificial neural network constructed from these devices demonstrates 93.0% accuracy in decrypting dual‐encrypted video. This work establishes a shift from parameter‐defined to material‐intrinsic multidimensional encoding for neuromorphic optical security systems.