High-Performance Organic Photodetectors With Synergistic Trap-State Passivation and Interfacial Engineered ZnO Transport Layer for Deep-Learning-Enabled Pulse Monitoring

Organic photodetectors (OPDs) are emerging as promising candidates for flexible and low-cost sensing applications, but their performance is often limited by trap-induced dark current and unstable interfacial energetics. Here, we systematically studied the defect evolution of classical ZnO electron transport layers and revealed a “resting effect”. Direct and indirect evidence proved that Long-term air aging or strong light activation can effectively mitigate oxygen-vacancy-related shallow trap states and narrow the interfacial energetic disorder. This reduction in energetic disorder is beneficial to charge extraction in the corresponding devices. Furthermore, PDINN interfacial modifier was introduced to passivate residual defects and form a favorable dipole to optimize energy-level alignment. The resulting OPDs deliver a high peak specific detectivity (∼8.30 × 10¹³ Jones) with suppressed dark current density, microsecond-scale response, and stable operation, and the performance-improvement trends remain consistent across several representative non-fullerene active layers beyond BTP-series derivatives (including IDIC, ITIC–4Cl, and PC₆₁BM). Integrated into a wearable self-powered platform, the optimized devices enable high-fidelity pulse acquisition with a signal-to-noise ratio above 25 dB. Coupled with a residual-attention neural network (OPDPulseNet), the system accurately classifies motion-induced pulse patterns with an accuracy of 97.14%. Overall, we connect ZnO defect-state regulation and PDINN interfacial tuning to traps, noise spectra, and noise-limited detectivity, and further demonstrate the benefits for high-SNR wearable pulse sensing.