Enhancing the Accuracy of Triboelectric Sensor Based on Triboelectric Material Surface Interface Strain Regulation Strategy.

Triboelectric nanogenerators (TENGs) produce distinct electrical signals upon contact with different objects, enabling their application in tactile sensors for material identification. Current strategies to improve identification accuracy primarily focus on sensor structure, operating mode, and material composition. Here, we introduce an interface strain management strategy to enhance both output performance and recognition accuracy. The proposed TENG device comprises two key components: a copper sheet electrode and a polydimethylsiloxane (PDMS) triboelectric layer with different elastic modulus. Interestingly, the device's output performance does not increase monotonically with enhanced strain capacity. Instead, it exhibits an initial rise followed by saturation and eventual decline. This behavior is attributed to the formation of a high-viscosity surface at elevated PDMS curing ratios, which introduces interfacial adhesion that reduces effective contact stress during the contact-separation cycle. Therefore, optimizing the PDMS curing ratio is essential to balance interfacial strain and surface viscosity, thereby maximizing output performance. Leveraging machine learning, the system achieved a material identification accuracy of 98.6% using a convolutional neural network trained on triboelectric signal features under ambient conditions. Furthermore, an integrated material recognition platform was developed, incorporating the TENG-based sensor, data processing, and display modules, capable of real-time signal acquisition and interpretation. This work offers a promising approach for advancing material perception technologies toward intelligent machine applications.