A stretchable liquid metal switch for interactive and autonomous soft machines

Soft machines harness material-level physical intelligence to perform adaptive tasks, enabling advancements in biomedical and human-machine interaction fields. Soft switches are the basic building blocks to achieve intelligent functions like autonomous decisions and mechanical computation. However, current soft switches suffer from complex fabrication processes, limited performance, and a lack of multimodal control, which hinder their practical application and the realization of machine intelligence. Herein, by harnessing the unique self-pinch and self-healing effects of the gallium-based liquid metals (LMs), we describe a soft high-performance electric switch composed of an LM line encapsulated within an elastomer. Applying pressure to deform the LM switch can increase local current density, leading to the electromagnetic self-pinch effect for switching off. After releasing pressure, the LM can spontaneously heal with the elastic recovery of the elastomer for switching on. This LM switch shows comprehensive advantages, including a compact design (0.5 mm × 1.5 mm × 10 mm), good stretchability (100%), high on/off ratio (∼10⁹), rapid response time (<100 ms), and excellent durability (>12 000 cycles). Moreover, the LM switches enable multiple control modes, including magnetic and optical stimulation, through the integration of responsive materials. We demonstrate various LM switch-enabled functional soft machines, such as an interactive flexible gripper, a self-oscillating soft crawler, and wearable logic gates. This work will open new avenues for the application of LM in intelligent soft machines and advanced wearable electronics.