Enhancing the recognition accuracy of tactile sensor through electrode and triboelectric material interface structure management strategy

Abstract

As each material has various capabilities to lose or gain electrons, unique triboelectric signals are generated when the Triboelectric nanogenerator (TENG) device touches different objects. Thus, the device is widely used in tactile sensors for material identification. At present, the improvement of identification accuracy is solved from the perspective of designing sensor structure and working mode. Here, we explore an electrode/triboelectric material interface structure management strategy that can improve sensor output performance and identification accuracy. It would facilitate the formation of [Cu(NH3)4]2+ ligands along {111} crystal plane when the copper sheet was immersed in ammonia solution, resulting in the establishment of rough surface, corresponding surface roughness is enhanced from 1.506 µm to 3.441 µm. The as-fabricated TENG device is mainly composed of two parts: an electrode material (ammonia-etched copper sheet) and a triboelectric layer (polydimethylsiloxane (PDMS) layer) embedded in etched holes. Significantly, the triboelectric signal relates strongly to the specific electrode/PDMS interface structure and target material surface charge condition, establishing a rough PDMS/electrode interface can increase the maximum voltage by 40–80% and the transferred charges by 3–180% under different stresses. With the help of machine learning, the construction of a electrode/PDMS interface structure could further enhance accuracy rate from 98.3% to 99.6%. Meanwhile, a material perception system integrated with a TENG-based sensor, data processing and display modules was developed, which can read the real-time triboelectric signal, and an acceptable recognition result can be achieved. This work could extend material perception system to machine intelligence field.