Micro-Cluster Engineering Enables Hybrid Channel-Network Cracks for Highly Sensitive, Linear, and Robust Strain Sensors

Crack-based strain sensors offer substantial potential for health monitoring, motion detection, and human-machine interaction. Yet their practical use is constrained by inherent performance trade-offs that make it difficult to combine high sensitivity, broad working range, and reliable linearity, as well as by the mechanical instability of brittle conductive layers. This work reports a crack sensor based on an adjustable micron-cluster structure. It is fabricated through screen printing, which enables the production of a structurally tunable carbon nanotube-polydimethylsiloxane (CNT-PDMS) composite film. By controlling the cluster density and size, we successfully guide the formation of high-density, alternating long–short channel-network crack morphology, thereby synergistically optimizing the sensor performance. A high gauge factor (GF) of 149.51 and excellent linearity ( R 2 = 0.979) over a strain range up to 80% were achieved by the fabricated sensor. After 100 cycles of 100% stretching and 360° twisting, the sensor exhibits less than 2% degradation in both sensitivity and linearity. Demonstrations in cardiomyocyte contractile force detection and wearable human-machine interaction confirm its strong potential for applications in biomedical monitoring and intelligent interactive systems.