New flexible sensor enables high-sensitivity pulse monitoring and pattern recognition

Cardiovascular diseases remain a major threat to human health, making reliable physiological signals essential for early warning and long-term management. The radial artery pulse directly encodes information about cardiac contraction, vascular elasticity, blood flow propagation, and peripheral circulation. However, capturing these weak wrist-surface vibrations is challenging, as the recorded waveform can be affected by sensor-skin contact, applied pressure, wrist curvature, motion artifacts, and interindividual physiological differences.

In a study published in ENG. Chem. Eng., researchers at Xiamen University and collaborating institutions report an iontronic pressure sensor designed to address these challenges. The device integrates a magnetron-sputtered Pt interdigitated electrode with a microdome-structured thermoplastic polyurethane (TPU)/spherical boron nitride (SBN)/ionic liquid composite layer. The ionic liquid supplies mobile ions to form electrical double layers at the interface, while the microdome structure converts applied pressure into a staged deformation. Hierarchical SBN interfaces are expected to influence local ion organization and interfacial polarization, and mechanical tests indicate higher dynamic stiffness and lower cyclic peak force variation following the incorporation of SBN under the tested conditions.

The sensor exhibited sensitivities of 67.89 kPa⁻¹ in the 0–100 kPa range, 14.89 kPa⁻¹ in the 100–700 kPa range, and 4.276 kPa⁻¹ in the 700–1500 kPa range. It demonstrated response and recovery times of 136 ms and 30 ms, respectively, and was able to resolve a minimum demonstrated pressure of 173 Pa. Over 7,000 loading-unloading cycles, the maximum signal deviation was 3.72%.

As a proof of concept, the researchers evaluated the sensor’s ability to classify 14 standardized pulse patterns generated by a programmable pulse generator. The best-performing top-20 handcrafted-feature support vector machine achieved a five-fold mean accuracy of 98.23%. A wireless system was also used to record radial artery pulse waves before and after exercise, with frequency analysis showing dominant components corresponding to 58 bpm before exercise and 75 bpm one minute after running.

The authors explicitly note that these pulse patterns and single-participant on-body demonstrations support sensor-based pulse acquisition and analysis, but they do not constitute clinical validation or population-level cardiovascular assessments. Future studies should include repeated sessions, multiple participants, quantified contact pressures, and comparisons with synchronized reference devices. The work establishes the feasibility of controlled waveform discrimination and wearable pulse acquisition, providing a foundation for future non-invasive cardiovascular monitoring technologies.