We have grown high quality magnetite microcrystals conserving bulk magnetic properties below 20 nm thickness. The observed domain configurations are dictated purely by shape anisotropy thus demonstrating the possibility of designing spin structures in ultrathin, magnetically soft magnetite at will.

In order to build spintronic devices, thin films or smaller dimensionality structures are required. Thus, the growth of magnetite films and nanostructures has been sought through a substantial research effort. Films and nanostructures can now be routinely grown by molecular beam epitaxy, pulsed laser deposition or sputtering on a range of substrates spanning metals, insulators, and semiconductors. A common observation of the thin film materials grown (irrespective of their growth method or substrate employed) is that their magnetic properties differ strongly from those of bulk magnetite. A non-exhaustive list of unexpected properties include high coercivities, high saturation fields, out-of-plane magnetization, superparamagnetism in ultrathin films, or new easy-axes. In many cases, the modified magnetic properties are found in magnetite structures that are virtually indistinguishable from bulk magnetite from a chemical and structural point of view. Furthermore, they cannot be attributed to a reduction in dimensionality as they also appear in films which are hundreds of nanometers thick. The most likely and widely agreed upon explanation for these effects is the presence of growth defects, among which antiphase domain boundaries (APBs) are the most prominent example. The unit cell of magnetite is rather large (0.84 nm) and when nuclei located at non-integer distances of its unit cell coalesce, they form boundaries where the cation lattice is disrupted. APBs introduce magnetic couplings which do not exist in the perfect material. APBs and their evolution have been observed by transmission electron microscopy and their detailed behaviour has been studied, specially by W. Eerenstein and coworkers. Thus, the growth of APB-free structures is of current interest not only for magnetite but for other spinels as well.

In the present work, we study the magnetic domains in flat single-crystal magnetite islands grown on Ru(0001) by reactive molecular beam epitaxy. As every island grows from a single nucleus, they are expected to be APB-free. By means of vectorial X-ray magnetic circular dichroism in photoemission microscopy (XMCD-PEEM) we determine with nanometer resolution the magnetization vector in these islands and study their stability both experimentally and through micromagnetic simulations.