How Can Melting Metal Make a Soft Robot Stronger?

Soft robots, also called soft manipulators, are like silicone arms that bend and adapt to complex shapes. Their strength lies in flexibility, but that makes them unstable when lifting heavy objects. To tackle this, a research team designed a new system that uses low‑melting‑point alloys (LMPA). These alloys solidify just above 30 °C and melt around 60 °C. Placed in tiny chambers inside the arm, the LMPA can switch from solid to liquid with a small heater. When solid, the arm becomes stiff; when liquid, it returns soft. By using alloys with different melting points, the robot can select three stiffness levels: low, medium, high. The manipulator’s body is 3D‑printed with silicone molds, while sensors track temperature and pressure. Researchers modeled heat distribution with Fourier’s law and validated it with FEM simulations. Tests showed that heating the LMPA lets the arm lift heavier loads without losing its ability to bend. This method paves the way for robots that blend the delicacy of a rubber hand with the strength of a metal clamp, useful for medical, industrial, or exploration tasks.

The core of the new manipulator consists of chambers filled with low‑melting‑point alloys (LMPA) such as bismuth‑indium‑tin (Bi‑In‑Sn) or gallium‑indium‑tin (Ga‑In‑Sn). These alloys exhibit sharp phase transitions, with elastic modulus changes that can exceed an order of magnitude between solid and liquid states. The thermal model was derived from Fourier’s integral equation, \(\int_V \rho c_p \frac{\partial T}{\partial t} dV = -\int_S \mathbf{q}\cdot d\mathbf{S} + Q_{gen}\), where \(\rho\) is density, \(c_p\) specific heat, and \(Q_{gen}\) the heat generated by the electric heater. FEM simulations accounted for conduction through silicone, LMPA, and surrounding air, showing temperature gradients within 2 °C of the target. Stiffness characterization employed a cantilever bending test, measuring the bending moment \(M = EI\theta/L\) where \(E\) is the elastic modulus, \(I\) the second moment of area, and \(\theta\) the deflection angle. Results indicated that raising the temperature from 30 °C to 65 °C reduces \(E\) from roughly 1.2 GPa to 0.02 GPa, enabling three distinct stiffness regimes. Dynamic response was assessed with a temperature step test, revealing an average transition time of 1.8 s from soft to rigid, fast enough for load‑changing scenarios. Integrated pressure and temperature sensors feed a real‑time PID controller, keeping temperature within ±0.5 °C of the setpoint and preventing overshoot that could jeopardize mechanical stability. This combination of phase‑change materials, advanced thermal modeling, and feedback control makes the manipulator a versatile platform for next‑generation soft robotics.