Machine Learning-Validated Silk Fibroin Composite Hydrogels: Robust, Antifreezing, and Conductive Sensors

Owing to their renewable nature and biocompatibility, silk fibroin composite hydrogels have been widely employed as flexible strain sensors in recent studies. However, three major limitations hindering the application of flexible devices include prolonged dialysis-induced undesirable ion loss, elevated interfacial impedance due to spatially heterogeneous distribution of exogenous ions, and dysfunction caused by low temperature. This study proposes a simple in situ ion preloading method for fabricating the PZS composite hydrogel composed of silk fibroin (SF), poly(vinyl alcohol) (PVA), zinc chloride (ZnCl2), and ethylene glycol (EG). This strategy significantly increases the cross-linking sites within hydrogel networks while retaining uniform ion distribution. The resulting hydrogel architecture exhibits excellent mechanical properties (tensile stress exceeding 1.92 MPa at 500% strain), outstanding conductivity (up to 12 S/m), and retainable functionality at low temperature (without freezing at −120 °C). The hydrogel can serve as an effective strain sensor with the ability to detect multiple types of human movements. Notably, the strain sensor exhibits 99.06% accuracy in motion digit recognition through machine learning-assisted verification. This research presents a significant advancement by circumventing the intricate procedures of conventional approaches. Through strategically optimized material compositions and fabrication techniques, the synergistic enhancement of antifreezing, mechanical, and conductive properties is accomplished, providing an idea for designing future flexible electronics.