Li, NingNingLiWang, ZhaosuZhaosuWangNiu, YuYuNiuLi, YuYuLiWen, SuyiSuyiWenZhang, HulinHulinZhangZONG-HONG LIN2025-07-302025-07-302025-06-29https://www.scopus.com/record/display.uri?eid=2-s2.0-105009405656&origin=resultslisthttps://scholars.lib.ntu.edu.tw/handle/123456789/730816The emergence of an e-skin receptor is an optimal solution for restoring the hand function of patients with sensation disorder, while constructing an e-skin receptor with high sensitivity, self-supervised capability, and open-environmental stability remains challenging. Here, inspired by the human skin perception mechanism, an ultrasensitive self-powered multimodal fingertip receptor that integrates thermogalvanic hydrogels as active mechanoreceptors and thermoreceptors for entropy-stabilized material fingerprint perception is proposed. A micropatterned and gradient structure strategy is introduced to improve the sensitivity to 53.6 kPa−1 with a low detection limit of 1.9 Pa. By exploiting static thermovoltage and dynamic differential signals to visualize the unsteady interfacial heat conduction, different materials can be determined in 80 ms based on the fast and slow adaptive sensations of the receptor. The self-supervised thermovoltage compensation is realized by self-decoupling contact pressure and thermal contact coefficients of materials, accommodating variations in applied forces. Benefiting from the robust interfacial heat transfer process and thermoelectric conversion, the tactile perception mechanism demonstrates universality under various external surroundings and contact conditions. With the assistance of deep learning, the fingertip receptor can function as an augmented somatosensory receptor to accurately perceive cutaneous cues of objects with an accuracy of 95.5%, which provides the potential of intelligent haptic perception to human-machine interfaces and prosthetics.falsedeep learningfingertip receptorskin-inspiredtactile perceptionthermogalvanic hydrogeljournal article10.1002/advs.2025058732-s2.0-105009405656