Dual‐Function Enhancement of PVDF Nanofibers via Bimetallic MOF for Wearable Pressure Sensor and AI‐Driven Non‐Invasive Blood Pressure Prediction

Abstract

Engineering high‐performance wearable piezoelectric sensors requires concurrent optimization of β‐phase crystallinity and conductive pathways, a persistent materials‐level challenge. Here, we introduce a novel two‐stage amplification strategy employing a Zn/Cu bimetallic zeolitic imidazolate framework (Cu‑ZIF‑8) to simultaneously engineer β‑phase alignment and establish percolated conductive networks within electrospun PVDF nanofibers. Incorporating 3 wt% Cu‑ZIF‑8 boosts β‑phase content by 21%, functioning as a dynamic nucleating agent that suppresses α‑phase crystallization and promotes all‑trans chain conformation. Coupled with copper‑ion‑facilitated charge transport, the pressure sensor achieves an exceptional sensitivity of 23.37 mV/kPa, a fast response time of ∼42.89 ms, and robust cyclic stability over 5000 cycles, enabling high‑fidelity detection of multiple physiological signals (including arterial pulse signals). To translate signals into actionable diagnostics, we further integrate a hybrid deep‑learning architecture combining Convolutional Neural Networks with Bidirectional Long Short‑Term Memory (CNN‑BiLSTM) to predict non‐invasive blood pressure (BP) from pulse waveforms. Trained on multi‐subject pulse waveform datasets, the model predicts BP with remarkable accuracy, attaining mean absolute errors of 0.69±3.95 mmHg (systolic) and 0.25±2.21 mmHg (diastolic), surpassing conventional regression models. By unifying tailored material design with embedded artificial intelligence, this work establishes PVDF/Cu‑ZIF‑8 composites as a versatile platform for next‑generation smart wearables capable of continuous health monitoring.