We present a new compact instrument designed for scanning transmission X-ray microscopy. It has piezo-driven linear stages, making it small and light. Optical components from the virtual source point to the detector are located on a single optical table, resulting in a portable instrument that can be operated at a general-purpose spectroscopy beamline without requiring any major reconstruction. Careful consideration has been given to solving the vibration problem common to high-resolution microscopy, so as not to affect the spatial resolution determined by the Fresnel zone plate. Results on bacteriogenic iron oxides, single particle aerosols, and rare-earth permanent magnets are presented as examples of its performance under diverse applications.

We have studied the magnetic layer thickness dependence of the orbital magnetic moment in magnetic heterostructures to identify contributions from interfaces. Three different heterostructures, Ta/CoFeB/MgO, Pt/Co/AlO_x and Pt/Co/Pt, which possess significant interface contribution to the perpendicular magnetic anisotropy, are studied as model systems. X-ray magnetic circular dichroism spectroscopy is used to evaluate the relative orbital moment, i.e. the ratio of the orbital to spin moments, of the magnetic elements constituting the heterostructures. We find that the relative orbital moment of Co in Pt/Co/Pt remains constant against its thickness whereas the moment increases with decreasing Co layer thickness for Pt/Co/AlO_x, suggesting that a non-zero interface orbital moment exists for the latter system. For Ta/CoFeB/MgO, a non-zero interface orbital moment is found only for Fe. X-ray absorption spectra shows that a particular oxidized Co state in Pt/Co/AlO_x, absent in other heterosturctures, may give rise to the interface orbital moment in this system. These results show element specific contributions to the interface orbital magnetic moments in ultrathin magnetic heterostructures.

Spectroscopy is a widely used experimental technique, and enhancing its efficiency can have a strong impact on materials research. We propose an adaptive design for spectroscopy experiments that uses a machine learning technique to improve efficiency. We examined X-ray magnetic circular dichroism (XMCD) spectroscopy for the applicability of a machine learning technique to spectroscopy. An XMCD spectrum was predicted by Gaussian process modelling with learning of an experimental spectrum using a limited number of observed data points. Adaptive sampling of data points with maximum variance of the predicted spectrum successfully reduced the total data points for the evaluation of magnetic moments while providing the required accuracy. The present method reduces the time and cost for XMCD spectroscopy and has potential applicability to various spectroscopies.

The electronic structure of spinel FeV${}_2$O${}_4$, which contains two Jahn-Teller active Fe and V ions, has been investigated by employing soft x-ray absorption spectroscopy (XAS), soft x-ray magnetic circular dichroism (XMCD), and nuclear magnetic resonance (NMR). XAS indicates that V ions are trivalent and Fe ions are nearly divalent. The signs of V and Fe $2p$ XMCD spectra are opposite to each other. It is found that the effect of the V $3d$ spin-orbit interaction on the V $2p$ XMCD spectrum is negligible, indicating that the orbital ordering of V $t_{2g}$ states occurs from the real orbital states and that the orbital moment of a V${}^{3+}$ ion is mostly quenched. NMR shows that V spins are canted to have a Yafet-Kittel-type triangular spin configuration.

Spatial inhomogeneity on the electronic structure is one of the vital keys to provide a better understanding of the emergent quantum phenomenon. Given the recent developments on spatially resolved ARPES (ARPES: angle-resolved photoemission spectroscopy), the information on the spatial inhomogeneity on the local electronic structure is now accessible. However, the next challenge becomes apparent as the conventional analysis encounters difficulty handling a large volume of a spatial mapping dataset, typically generated in the spatially resolved ARPES experiments. Here, we propose a machine-learning-based approach using unsupervised clustering algorithms (K-means and fuzzy-c-means) to examine the spatial mapping dataset. Our analysis methods enable automated categorization of the spatial mapping dataset with a much-reduced human intervention and workload, thereby allowing quick identification and visualization of the spatial inhomogeneity on the local electronic structures.

The surface magnetism of Fe(001) was studied in an atomic layer-by-layer fashion by using the in situ iron-57 probe layer method with a synchrotron Mössbauer source. The observed internal hyperfine field $H_{\mathrm{int}}$ exhibits a marked decrease at the surface and an oscillatory behavior with increasing depth in the individual upper four layers below the surface. The calculated layer-depth dependencies of the effective hyperfine field $|H_{\mathrm{eff}}|$, isomer shift $\delta$, and quadrupole shift $2ϵ$ agree well with the observed experimental parameters. These results provide the first experimental evidence for the magnetic Friedel oscillations, which penetrate several layers from the Fe(001) surface.