Key points are not available for this paper at this time.
Electrospinning is a versatile technique for fabricating triboelectric materials. Although carbonization can significantly improve the electrical conductivity of fiber membranes, their application as triboelectric layers remains underexplored. In this work, a copper–silver (CuAg) alloy-doped carbon nanofibers (CuAg@CNFs) membrane was designed and employed as a triboelectric layer to enhance the accuracy and sensitivity of self-powered tactile sensors based on triboelectric nanogenerators (TENGs) for material identification. The CuAg@CNFs membrane, obtained by electrospinning and subsequent thermal conversion, exhibits a three-dimensional conductive network with in situ–generated alloy nanoparticles. Such a structural configuration promotes efficient charge migration across the interface, thereby enhancing electrical output. The fabricated CuAg@CNFs based-TENG delivers an output voltage of 4.3 V and a charge density of 40 nC m –2, maintaining stable operation over repeated cycles. By correlating triboelectric signals with material surface characteristics, an intelligent sensing framework integrating a convolutional neural network (CNN) was constructed. The system achieved a recognition accuracy of 99.6% across ten materials and maintained high discriminative capability in complex environments. During real-time operation, the recognition accuracy for four types of material categories (aluminum, board, glass, and plastic) reached 100, 93, 99, and 97.5%, respectively. This study demonstrates a feasible strategy for constructing high-performance self-powered sensing systems through the synergistic combination of material design and machine learning algorithms.
Li et al. (Sat,) studied this question.