Preliminary investigation of the effect of artificial sweat on a wearable textile sensing system

Abstract

Introduction Textile wearable systems for human movement monitoring are increasingly popular. However, few examples report on robustness to sweat, which is relevant for use in real life. Some reported the effect of artificial sweat like phosphate buffered saline (PBS; Lin et al., 2022) or simply moisture (Xu et al., 2020) on custom materials. We previously developed an all-textile wireless sensing platform with commercial conductive yarns and fabrics containing silver. There is no study on the effect of sweat on such materials, therefore we performed a preliminary study to account for moisture and potential oxydation of silver. Methods The textile sensing system is resonating RLC circuit, where the sensing part is a capacitive parallel plate strain sensor (C) located on a joint (knee). All components are textile based and contain silver. As the capacitive sensor stretches, capacitance increases and the resonance of the circuit fres decreases. This information is transmitted wirelessly via inductive coupling (Galli et al., 2023). We sprayed 1 ml of 0.1 M PBS solution on the textile capacitive sensor to simulate sweating, and applied mechanical strain before (damp state) and after air drying (dry state). The unmodified sensor (before the addition of any PBS) was also used as a baseline measure. First, we applied fixed strain (10%) with a universal testing machine; then, we tested the response of the sensorized pants when bending the knee. Results The resonance frequency of the textile sensing (RLC) circuit in the damp state was much lower than the baseline (14.85 ± 0.11 MHz vs 22.70 ± 0.12 MHz) as expected from the higher dielectric constant of water that increases the baseline capacitance of the sensor. As for the change in Δfres upon 10% strain (Δfres = fres,baseline - fres,stretch), interestingly a larger change was observed for the damp configuration as compared to the baseline and dried (1.08 ± 0.08 vs 0.79 ± 0.06 vs 0.66 ± 0.03 MHz). A similar behaviour was observed in the test with pants, where the response for flexion was Δfres = 1.58 MHz for the damp sensor and Δfres = 1.28 MHz for the dried sensor. Discussion/Conclusion This preliminary investigation showed promising results in terms of robustness of our system to artificial sweat, as there was a measurable response both in the damp and dried configurations. Further tests with different sweat amounts and rate are needed to determine the full functioning range, e.g., how much sweat is tolerated.