• • Ferroelectric tunnel junctions (FTJs) based on Hf0.5Zr0.5O2 (HZO) achieve sub-nA operating currents (e.g., <1 nA) and fast switching speeds, enabling high-density, low-power artificial synapses for deep learning acceleration (Ref. 191).
• • Ultrathin (3.5 nm) ferroelectric tunnel junctions demonstrate energy-efficient reservoir computing for temporal data learning, with robust performance metrics suitable for edge AI applications (Ref. 201).
• • In-plane charged domain walls in ferroelectric films exhibit memristive behavior, offering a novel mechanism for analog synaptic weight modulation with potential for high-density crossbar arrays (Ref. 195).
• • Ferroelectric domain wall memories and FTJs show enhanced tunneling electroresistance (TER) ratios, with values exceeding 1000% in optimized heterostructures, crucial for reliable multi-state storage in neuromorphic systems (Ref. 187).