Vienna team reconstructs an ultrathin magnet's inner pattern and sensor height from field maps

The underlying problem is non-uniqueness. Setescak, a doctoral researcher at the University of Vienna and first author of the study, described it with an analogy: "Trying to reconstruct the magnet from this external field is like trying to guess the shape of a light bulb from the beam of light on a wall without knowing how far away it is held," said Alexander Setescak ^[8].

The team's answer is a selection rule. The algorithm weighs candidate magnetic patterns and keeps the one that minimizes the material's physical energy while still matching the measured field ^[3]. "We let the physical energy of the magnet decide which solution is realistic," said Claas Abert, who led the study ^[9]. To compute that energy, the software calls NeuralMag, an existing open-source micromagnetic package developed in Vienna ^[12].

In our view, that is a narrower result than "the ambiguity is gone." The field maps still admit many internal patterns ^[1]. The method adds a criterion that chooses among them, and the criterion is only as good as the assumption that the sample sits in the state the physics predicts. The thing this doesn't tell you is how the reconstruction behaves on a sample that is not in its lowest-energy state. Phys.org's account of the paper does not describe such a test.

The height result follows from the same logic. The diamond sensor sits tens of nanometers above the magnetic layer, and because the field weakens rapidly with distance, a small error in the assumed height distorts the reconstructed pattern ^[6]. Since the decay obeys strict physical laws, the software solves for pattern and height together ^[7]. For the Dresden measurements it found a distance of 81 nanometers ^[7].

"In experiments, the exact sensor height is usually unknown," said Uri Vool, a group leader at the Max Planck Institute for Chemical Physics of Solids in Dresden ^[10]. He added: "Extracting the distance directly from the magnetic measurement removes a major source of error and makes the reconstructed map far more reliable" ^[11]. The source does not say whether the 81-nanometer figure was checked against an independent height measurement, so for now it is the method's own output on one set of samples ^[13].

Setescak said the group hopes other laboratories "will pick up the framework and apply it to their own measurements" ^[16]. The reconstruction code is freely available ^[14]. Moving it to a different instrument is the authors' stated expectation, not something the paper demonstrates ^[15].