In order to make very sensitive measurements, Jamet and colleagues embedded cobalt clusters into the junctions of a micro-SQUID - a miniature superconducting quantum interference device - that exploits quantum effects to measure extremely weak magnetic signals. The clusters were widely spaced within the junctions so that the signals from individual clusters could be distinguished. The team imposed a strong magnetic field to align the clusters in the same direction, and then applied an opposing magnetic field. The micro-SQUID measured the final state to confirm that the direction of magnetization had switched. The discovery is a success for the new micro-SQUID technique because the signals it measured were a thousand times weaker than any previous measurements.

The clusters are formed using a method known as low-energy cluster beam deposition. The atoms assemble themselves into truncated octahedrons because this shape offers the lowest surface energy. Jamet's team predicted the overall magnetic anisotropy of a single cluster by calculating the interactions of all the atoms within the cluster. But their predictions fell dramatically short of the micro-SQUID measurements of magnetic anisotropy. Using a three-dimensional analysis technique, the team successfully separated the contributions from different regions of the cluster and found that the atoms at the surface dominated this magnetic effect.