Create / 2019
Changing the interaction opens a new design space.
THE CENTRAL QUESTION
Small permanent magnets behave mainly as dipoles: one end attracts the opposite end of another magnet. That simple rule is excellent for making chains, but it becomes a restriction when the goal is a sheet, a junction or a soft structure with several branches. This project began by asking whether the magnetic interaction itself could be redesigned, so the desired geometry would be natural rather than forced.
A dipole has two poles, so neighbouring units usually settle into head-to-tail chains. More complicated arrangements can be assembled, but many bonds then sit in an unfavourable state and the structure becomes fragile. The problem was therefore deeper than finding a better assembly procedure: the basic building block offered the wrong set of preferred connections.
The new module contains two small magnets placed at an angle. Close to the module, their fields combine into a fourfold quadrupole pattern with four preferred bonding directions. Farther away, a weaker dipole remains, allowing the completed structure to respond to a uniform external field. Bonding and actuation are therefore carried by different parts of the same field.
The team calculated the magnetic energy around a pair of modules and then watched millimetre-scale printed units approach one another. Both the calculation and the experiment showed four stable docking positions. This agreement mattered because it linked the final shapes to a defined energy landscape rather than to trial-and-error assembly.
Once the modules could form stable two-dimensional arrangements, their remaining dipole directions were selected independently. Two soft star-shaped assemblies with the same geometry but different magnetic programs bent into different shapes during the same field sweep. The structure stored one kind of information in its connections and another in its response to the field.
The broader lesson is that changing interaction symmetry can open a new design space before any robot is built. The demonstrated system is millimetre-scale, planar and limited to a small set of magnetization choices, but it shows a route toward modular materials whose shape and motion can be programmed separately.