Self-Organized Protonic Conductive Nanochannel Arrays for Ultra-High-Density Data Storage

While the highest-performing memristors currently available offer superior storage density and energy efficiency, their large-scale integration is hindered by the random distribution of filaments and nonuniform resistive switching in memory cells. Here, we demonstrate the self-organized synthesis of a type of two-dimensional protonic coordination polymers with high crystallinity and porosity. Hydrogen-bond networks containing proton carriers along its nanochannels enable uniform resistive switching down to the subnanoscale range. Leveraging such nanochannel arrays, we achieve logic operations of graphical gate circuits with negligible leakage and sneak path currents over areas ranging from 0.5 μm × 0.5 μm to 20 nm × 20 nm, providing the smallest building blocks to date for large-scale integration. The nonvolatile resistive switching exhibits high mobility (∼0.309 cm² V^-1 s^-1), a large on/off ratio (∼10³), and ultrahigh-density data storage (∼645 Tbit/in²), even within a trilayer (∼4.01 nm). An ultrahigh-precision artificial retina with integrated convolutional neural network calculations is demonstrated, enabling facial and color recognition capabilities.