Defect‐Engineered N‐Doped ZnO as a Dual‐Functional Platform for PENG‐Driven Memristive Sensory Memory

Abstract

Energy‐efficient computing is a critical challenge for artificial intelligence and edge applications. Although self‐powered memory systems that integrate sensing, energy harvesting, and storage are promising, they often suffer from high leakage currents, poor reproducibility, and material incompatibility. Here, we present a dual‐functional N‐doped ZnO (NZO) platform that electrically couples an interface‐type memristor with a piezoelectric nanogenerator (PENG), enabling self‐powered memory programming. Uniform N‐doping via low‐temperature atomic layer deposition passivates oxygen‐related defects, suppresses leakage currents, and stabilizes resistive switching. A thin TiO 2‐x interfacial layer acts as an oxygen reservoir and enables precise barrier engineering for reversible, low‐power multilevel operation. At the same time, the reduced carrier concentration in NZO enhances PENG performance through suppression of electrostatic screening effects. Density functional theory (DFT+U) calculations further predict N‐induced changes in elastic stiffness and compliance, suggesting a possible secondary contribution to the electromechanical response through modified strain accommodation. As a result, the rectified PENG output directly programs the memristor without an external bias. The coupled system exhibits stable multilevel storage, robust endurance, and synaptic plasticity. By co‐designing energy harvesting, sensing, and memory on a unified oxide platform, this work establishes a practical pathway toward energy‐autonomous neuromorphic computing and intelligent tactile memory systems.