Thermal‐Conductivity‐Driven Dual‐Wedge Thermoelectric Architecture for Intelligent Thermal Perception and Material Identification

ABSTRACT Maintaining a stable temperature difference within ultrathin thermoelectric sensors remains a key challenge for achieving skin‐conformal and integrated thermal perception. Here, we present a dual‐wedge thermoelectric (DWTE) architecture composed of two complementary wedge‐shaped layers with distinct thermal conductivities. This structural configuration suppresses vertical heat diffusion while facilitating lateral heat flow, thereby establishing a stable lateral temperature gradient in thermoelectric films. This dual‐wedge design not only enhances thermal response speed and signal stability but also significantly amplifies thermoelectric output under transient stimuli. Based on the thermal‐conductivity‐dominated thermal diffusion differences and an intelligent classification algorithm, the DWTE device integrated on an artificial hand enables rapid identification and classification of seven materials with an accuracy of 99.7%. Meanwhile, a 32‐channel flexible thermoelectric array provides conformal real‐time thermal mapping during object grasping. This work provides an effective solution to the long‐standing trade‐off between device thickness and thermal gradient stability, offering a universal strategy for fast‐response thermoelectric sensing in advanced wearable and human‐machine interactive systems.