WorksPhysics and Astronomy
Nanoscale imaging of spin textures with locally varying altermagnetic response in -FeO
Abstract
Altermagnetism is a recently identified magnetic state in which time-reversal symmetry is broken despite a collinear compensated spin structure. The response of altermagnets is determined not only by their -, -, or -wave spin order, but also the orientation of their Néel vector . Therefore, accessing a response that fundamentally depends on the orientation of , such as the anomalous Hall effect, remains experimentally challenging in particular at the nanoscale. Here, we harness nano-spectroscopic X-ray magnetic circular dichroism (XMCD) to investigate nanoscale modulated altermagnetic responses in -FeO (Hematite). By performing spectroscopy across the temperature-induced -reorientation Morin transition, we observe the on-and-off switching of XMCD, in agreement with our theoretical calculations. Although the bulk XMCD vanishes below the Morin temperature, we confirm the reorientation of by harnessing polarization-independent X-ray absorption spectroscopy. Moreover, we observe a finite XMCD signal in nanoscale domain walls with locally modulated Néel vectors, while the surrounding domains exhibit no XMCD. At room temperature, we instead identify altermagnetic meron spin textures that exhibit XMCD in their planar regions but no XMCD in their nanoscopic cores. Our results establish a pathway to harness complex spin textures with nanoscale functionalities in a broader class of altermagnets with various -orientations and using light, earth-abundant elements.
Altermagnets are collinear magnets which exhibit additional time reversal symmetry breaking associated with the local symmetry of each sublattice [1, 2]. The recent proposal of altermagnetism has triggered extensive research not only into its fundamental physical properties, but also into potential device applications. Following the theoretical characterization of altermagnets by non-relativistic, exchange-driven -, -, or -wave type spin ordering, featuring two, four, or six spin-degenerate nodal surfaces in their non-relativistic band structure [1, 2], altermagnetic spin splitting was experimentally observed by angle-resolved photoemission spectroscopy [3, 4, 5, 6, 7] [8, 9, 10, 11, 12].
Beyond the spin splitting in momentum space, emergent phenomena arising from broken time-reversal symmetry despite antiferromagnetic ordering—such as
the anomalous Hall effect (AHE) [13–23], chiral magnons [24–26], and optical dichroic effects, including the magneto-optical Kerr effect (MOKE) [27–30] and X-ray magnetic circular dichroism (XMCD) [31–35]—have been reported. These phenomena require relativistic spin-orbit coupling (SOC), rendering them dependent not only on the type of altermagnet, but also on the orientation of the Néel vector , where denote the magnetic moments of the two sublattices. This orientation dependence is exemplified by the -wave altermagnets -MnTe and CrSb. Although both are classified as -wave altermagnets, altermagnetic responses such as AHE and XMCD are allowed in -MnTe with in-plane basal -orientation, whereas they are forbidden in CrSb exhibiting a perpendicular to the basal plane -orientation. This dependence of the physical properties of altermagnets on the Néel vector opens the possibility for a spin reorientation-induced on/off switching of emergent effects. Thus, whether the response vanishes or remains finite goes beyond identifying the altermagnetic nature of materials and establishes a platform for controlling altermagnetic functionalities. Experimentally, recent studies demonstrating that a modulation of the emergent AHE can be realized through the reorientation of the Néel vector by application of strain and high magnetic fields [14, 36, 37]. However, while macroscopic spin reorientation enables global control of altermagnetic responses, achieving locally varying, nanoscale responses remains a major challenge.
In this study, we demonstrate the presence of nanoscale spin textures with locally varying altermagnetic properties in α-Fe₂O₃ (Hematite) using nanospectroscopic X-ray imaging. Long known as a naturally occurring antiferromagnet ( K), α-Fe₂O₃ has received recent attention due to its ability to host nanoscale topological textures [38–40], and more recently, due to its prediction as a -wave altermagnetic candidate [1, 2], that has been confirmed with AHE [35], XMCD [35], and MOKE [29, 30]. The magnetic structure is shown in Figure 1a, where the rhombohedral unit cell and -wave spin-polarized electronic structure can be seen in Figure 1b [see the Supplementary Information (SI) for a more detailed discussion and symmetry analysis]. In α-Fe₂O₃, there is a spin reorientation transition, known as the Morin transition, that occurs at K (Figure 1c) [41]. Intriguingly, as one sweeps across this transition, the altermagnetic type and symmetry of the material will remain the same - however as the spins reorient, the emergent effects arising from SOC should disappear, as illustrated in Figure 1d, Figure 1e.
Figure 1. a Magnetic structure of α-Fe₂O₃ in a rhombohedral unit cell. The hexagonal crystallographic axes are also indicated in the figure. b Altermagnetic spin-polarization calculated in momentum space for the plane and an energy of 198.5 meV below the valence band maximum. c Temperature dependence of the magnetization of Sample #1 measured under a magnetic field of 0.01 T along the hexagonal -axis. Néel vectors in a hexagonal setting below and above are schematically illustrated in the inset. Magnetic moments (orange arrows) above are slightly canted within the -plane, resulting in a small uncompensated moment (gray arrow). The uncompensated magnetic moment is oriented by the magnetic field, resulting in the Néel vector becoming perpendicular to the magnetic field. Below , the magnetic moments (green arrows) align parallel to the -axis without canting. d,e Magnetic structure of α-Fe₂O₃ in a rhombohedral unit cell below and above . Under time-reversal symmetry operation , the Néel vector reverses its orientation. Expected X-ray magnetic circular dichroic (XMCD) contrast for normal-incidence X-rays -axis for each spin configuration is schematically illustrated in the bottom.
Here, by probing spectroscopic X-ray magnetic circular dichroism, we observe finite XMCD at room temperature, which we confirm to be of altermagnetic origin with density functional theory (DFT) + dynamical mean-field theory (DMFT) calculations. We combine this altermagnetic XMCD with nanoscale imaging to identify the presence of a variety of altermagnetic spin textures. By exploiting polarization-independent absorption contrast that distinguishes in-plane and out-of-plane L-orientations, we identify the spin-reorientation associated with the Morin transition, and observe the disappearance of the XMCD–thus realizing the reorientation-induced on-off switching of altermagnetic XMCD. Despite the absence of XMCD, we characterize the altermagnetic domain configuration below using this spectroscopic absorption contrast and identify the presence of nanoscale domain walls. Remarkably, these domain walls exhibit non-zero XMCD signal due to their local in-plane spin orientation, resulting in a finite altermagnetic response embedded in zero XMCD domains. Harnessing the L-orientation dependent absorption contrast, together with XMLD and XMCD signals, we are able to identify winding spin textures at room temperature to be topological merons, again with an out-of-plane core that exhibits distinct emergent effects from the surrounding domains. This modulation of emergent properties on the nanoscale in the form of domain walls and merons that exhibit distinct properties from the surrounding domains, demonstrates a key property of altermagnets that is particularly relevant for future information processing using altermagnetic spintronics.
ilar calculation of the XMCD spectrum, with consistent results, has been presented in Ref. [46, 47].
Dichroic altermagnetic spectroscopy
In order to probe the altermagnetic nature of the material, we first perform spectroscopic X-ray magnetic circular dichroism measurements on a bulk single crystal of -FeO (Sample #1). Details of the sample preparation, characterization, and measurements are provided in Methods and Supplementary information. We initially target the high temperature magnetic ordered state, where lies within the -plane. One of the challenges of performing non-spatially resolved spectroscopy is the prospect for multi-domain states that hinder the measurements [31]. In case of -FeO, the Fe moments are slightly canted within this plane due to the Dzyaloshinskii–Moriya interaction [42, 43], resulting in a weak ferromagnetic moment . We exploit this weak to induce a single domain state by applying an in-plane magnetic field during the measurements of the dichroic spectra [44]. As schematically shown in the inset of Figure 2a, we apply an in-plane magnetic field of around 100 mT along a direction close to the -axis to obtain a single-domain state for X-ray spectroscopy in total electron yield (TEY) mode at room temperature.
Figure 2. X-ray magnetic circular and linear dichroism (XMCD and XMLD) of α-Fe₂O₃ across Fe -edges and corresponding calculations. a,b X-ray absorption spectra acquired with circular right/linearly horizontal (red line) and circular left/linearly vertical (blue line) polarizations were measured by integral TEY measurements (Sample #1) above . A magnetic field of approximately 100 mT was applied within the c-plane. c,d Corresponding experimental XMCD and XMLD spectra. e,f Calculated XMCD and XMLD spectra. g,h STXM images of α-Fe₂O₃ (Sample #2) were recorded at the Fe -edge, 724.25 eV for circular polarizations and 724.5 eV for linear polarization (nominal values) indicated by red and blue arrows in c and d. XMCD and XMLD images at room temperature show altermagnetic domains and Néel vector, respectively. i,j Zoomed XMCD and XMLD images of the regions indicated by magenta boxes in g and h exhibit different contrast modulation, consistent with a 180° domain wall. Dependence of XMCD and XMLD contrasts on the in-plane L-orientation as indicated by the contrast wheels. At 50 K, all contrast vanished in the k XMCD and l XMLD images corresponding to g and h.
The X-ray absorption spectra (XAS) measured with left- and right-circularly polarized X-rays are shown in Figure 2a, while the XAS measured with linearly polarized X-rays (horizontal and vertical) are shown in Figure 2b. In the XAS spectra, two distinct peaks are observed at both the Fe (approximately 707–715 eV) and edges (approximately 721–727 eV), consistent with previous studies [45]. We then obtain the Fe -edge XMCD spectrum (Figure 2c), and XMLD spectrum (Figure 2d) from the difference between left and right circular polarizations and between the two orthogonal linear polarizations, respectively. Both spectra exhibit several sign reversals across the and edges, yet with clearly different spectral profiles.
To ascertain the origin of the observed dichroic signal, we performed DFT+DMFT calculations for the altermagnetic configuration. The calculated XMCD/XMLD spectra plotted in Figure 2e,Figure 2f, are consistent with the experimental data, confirming that the measured XMCD is indeed a signature of altermagnetism. Notably, the XMCD spectrum of -FeO closely resembles those reported for another altermagnet -MnTe [31–33]. As XMCD can also originate from ferromagnetic moments, we further evaluated the possible contribution of a net ferromagnetic component. In Supplementary Figure 5, the calculated full optical conductivity tensor and the X-ray Hall vector at the Fe edge are provided, indicating that the contribution from a possible ferromagnetic moment is negligibly small compared with the intrinsic altermagnetic contribution. In addition, as shown in Supplementary Fig. 6, the XMCD spectrum calculated for a hypothetical ferromagnetic configuration with moments aligned along the -axis exhibits a distinct line shape that differs significantly from the measured XMCD. We note that a similar calculation of the XMCD spectrum, with consistent results, has been presented in Ref. [46, 47].
Figure 5. Fe -edge XMCD (top) and XMLD (bottom) spectra of -FeO calculated using the DFT+DMFT method (left) and an Fe atomic model (right). The X-ray geometries used in the calculations are illustrated in the insets of panels b and d.
Dichroic altmermagnetic imaging
We next explore the formation of nanoscale altermagnetic domains by scanning transmission X-ray microscopy (STXM) in the absence of magnetic fields on an x-ray transparent window (average thickness of 150 nm) fabricated from a -FeO single crystal (Sample #2, see Methods and SI for further details on sample fabrication and measurements). The sample geometry is shown in Extended Data Figure 1. As shown there, lithographically patterned metallic Cu and Pt thin-film signal lines were fabricated on the sample for transport measurements; however, they were not used in the present work. For the STXM imaging, the X-ray energy was tuned to values exhibiting sufficient photon transmission ( edge) and strong dichroism, as identified from the spectroscopy (724.25 eV for XMCD and 724.5 eV for XMLD indicated by the arrows in Figure 2c,Figure 2d). As shown in the XMCD and XMLD images of Figure 2g and h, a complex domain configuration was observed at room temperature. First, at the metallic conductor regions indicated in Extended Data Figure 1, the domain morphology is locally modified, implying a strain-induced effect. Second, within this configuration, some domains of similar XMLD contrast exhibit opposite XMCD contrast. The characteristic features are highlighted by the zoomed images in Figure 2i,Figure 2j, indicated by magenta boxes in Figure 2g,Figure 2h. Between these two domains, a domain wall is observed, that exhibits a white-black-white-black contrast in XMCD, and a simpler white-black-white contrast in the XMLD image, consistent with the presence of an in-plane Néel-like 180° domain wall as observed in -MnTe [33]. The time-reversal symmetry breaking and orientations resolved by XMCD and XMLD for the present experimental geometry are indicated by the color wheels in Figure 2, with further details provided in the Supplementary Figure 3.
Having established altermagnetic time-reversal symmetry breaking in the high-temperature phase, we move on to the low-temperature phase below the Morin transition, where reorients along the -axis as illustrated in Figure 1d. XMCD and XMLD images at K are presented in Figure 2k,Figure 2l. The images are acquired under the same experimental conditions as those used for the high-temperature measurements shown in Figure 2g,Figure 2h. Neither the XMCD nor the XMLD image exhibits significant dichroic contrast. This observation is consistent with the fact that XMLD is excluded by geometry and moreover with the prediction that because of relativistic effects, XMCD is not allowed by symmetry in this low-temperature phase/orientation, even though the underlying altermagnetic character of the system is expected to remain unchanged.
L-orientation and dichroic imaging below
To characterize this XMCD-bulk-forbidden low-temperature phase further, we turn, counterintuitively, to a polarization-independent XAS effect that can be used to distinguish the L-orientation (with respect to the crystal axes) in the high- and low-temperature phases [39, 45] as illustrated in Figure 1d. Although the effect even occurs for arbitrary polarizations, we deliberately use unpolarized XAS spectra obtained by the sum of two orthogonal linear polarizations for avoiding regular dichroic effects in the following.
As shown in Figure 3a, we experimentally observed a clear change in the unpolarized XAS spectra extracted from absorption images acquired across . These differences in the unpolarized XAS spectra are in good agreement with corresponding spectra calculated by DFT+DMFT calculations for in-plane and out-of-plane L configurations, displayed in Figure 3b. It is worth pointing out that that the energy dependence and the magnitude of the contrast between the two L-orientations cannot be anticipated from symmetry considerations alone, but arises from many-body interactions in the final state involving the core hole. By exploiting the difference in unpolarized x-ray absorption at certain energies between (L -axis) and (L -axis), it becomes possible to distinguish states of different L-orientations in STXM imaging.

Figure 3. Comparison between in-plane and out-of-plane L-configurations. a Unpolarized experimental spectral data ( edge) acquired from sum of linear polarization images (LH+LV) above and below and b correspondingly calculated X-ray absorption spectra ( and edges). c Overview image using a single polarization (LV) at 724.5 eV (indicated by red arrows in a) and corresponding magnified views of the region enclosed by the white box in c. The domain wall with in-plane L-orientation is visible in d an absorption image obtained from the sum of linear horizontal and vertical polarization data. Here, the nonmagnetic absorption contrast was removed to emphasize the domain wall contrast (see Supplementary Figure 4 for details). e XMCD image recorded at 724.25 eV and f XMLD image at 724.5 eV of the same region show dichroic signals only for the area covered by the in-plane domain wall.
Following a reinitialization of the domain configuration by sweeping the sample temperature through the Morin transition, we observed an intriguing contrast near the upper metallic lead in the single-polarization image shown in Figure 3c. The image was acquired at an X-ray energy corresponding to the maximum L-dependent absorption contrast (724.5 eV), at K under an applied magnetic field of 250 mT along the hexagonal -axis. Figure 3d shows an absorption image obtained by the sum of two linear polarizations zoomed into the region where this contrast appears. In this figure, a white, curvilinear feature is clearly visible, with a full width at half maximum of nm. This bright contrast is consistent with a local reorientation of , and corresponds to a domain wall with in-plane , that separates two out-of-plane -distinct altermagnetic domains.
Remarkably, when we perform XMCD (724.25 eV) and XMLD (724.5 eV) imaging of this curvilinear structure, shown in Figure 3e,Figure 3f, a non-zero XMCD and XMLD contrast is observed in the region of the white line. This non-zero XMCD and XMLD is consistent with the in-plane orientation of the domain wall, which has the XMCD-allowed spin orientation typically associated with the high temperature phase, distinct from the surrounding domains in which XMCD is forbidden. The fact that XMCD is locally allowed only within the domain walls implies that other altermagnetic functionalities, such as the AHE, can be selectively activated at the nanoscale, highlighting these domain walls as promising functional elements for high-density, altermagnetic spintronic devices.
With this ability to characterize domain configurations in the altermagnetic phase for which observables such as XMCD and AHE are forbidden, a natural question is whether this contrast is unique to -FeO, or whether it can be exploited in a wider range of altermagnetic candidates. First, we note that the amplitude of the contrast is primarily governed by the interaction strength between the electrons and the core hole, as reflected in the changes of the contrast across the multiplet peaks at the edge in -FeO. In many localized compounds, the edges are known to exhibit similar multiplet structures [48], since this interaction strength does not vary significantly across the elements, and therefore such -dependent XAS contrast can be expected for a wide range of systems. Second, the energy dependence of the contrast, and the energy at which it is maximized, can be predicted using simple ionic multiplet calculations [48], as demonstrated in the Supplementary information - meaning that experimental access to both phases is not necessarily needed to be able to predict the nature of such an effect. As a result, we believe this unpolarized XAS contrast offers a robust and widely implementable route to characterize altermagnetic configurations, even when the primary domains do not exhibit any XMCD.
Topological Altermagnetic Textures
Having identified the presence of an in-plane domain wall in the low temperature regime, we next explore whether similar textures exhibiting a mixture of relativistically distinct -orientations can be observed in the high temperature regime. In particular, when we perform absorption contrast imaging at the region of the cyan box in Figure 2g or h, a distinct nanoscale dark dot is observed in the XAS image in Figure 4a, that is located at the center of rotationally-symmetric XMCD and XMLD patterns. Indeed, as displayed in Figure 4b,Figure 4c, the XMCD and XMLD contrast is consistent with an in-plane winding texture such as a vortex or antivortex, as has been observed previously in -MnTe [32, 33]. However, the dark absorption contrast at the center of the structure indicates that at the core of the vortex, tilts out-of-plane, analogous to what is seen in ferromagnets. Thus, it forms a meron-type topological texture with a skyrmion number of , characterized by a three-dimensionally winding spin configuration. Indeed, when we take line profiles of the unpolarized XAS and XMLD contrasts through the core along the horizontal and vertical spatial directions (Extended Data Fig. 2), one can see that the XMLD line profiles show their maximum and minimum along the horizontal and vertical directions at the same location where there is a maximum in the -dependent absorption contrast, as shown in Figure 4d,Figure 4e. These peaks have a full width half maximum of nm, confirming that the width of the core is significantly larger than our spatial resolution, and thus that the observed vanishing of XMLD/XMCD signal does not occur due to a spatial resolution artefact. This reduction in dichroism, coinciding with the -dependent absorption contrast, confirms that tilts out-of-plane in the core, and indicates that the region in the core is in a relativistically distinct altermagnetic state from the surrounding in-plane configuration.
Figure 4. Altermagnetic nanotexture at the region indicated by the cyan box in Figure 2g and h. a Unpolarized absorption, b XMCD, and c XMLD images at nominally 724.25, 724.25 and 724.5 eV, respectively. The dark point in the absorption image represents a core with out-of-plane L-orientation. XMCD and XMLD images show dichroic signal due to -symmetry broken domains and the in-plane L orientations surrounding the core, respectively. The core shows a spatial extent of approximately 140 20 nm from the line profile in d (extracted from a), and broad peaks across the core in both the absorption d and XMLD profiles e (extracted from c) indicate a continuous Néel vector rotation from out-of-plane to in-plane L orientations, consistent with the presence of a topological texture with as schematically illustrated in f.
The confirmation of nanoscale complex spin textures—a domain wall in the low temperature regime, and a meron in the high temperature regime—that consist of a mixture of two relativistically-distinct altermagnetic states is not only of interest from a fundamental point of view. Yet the prospect of realizing stable, localized variations in emergent properties, such as XMCD or AHE, whose magnitude and sign are governed by the local orientation of the Néel vector and can be controlled by magnetic fields or electric currents, represents an important step towards developing altermagnetic spintronics.
Conclusions
In conclusion, we have harnessed X-ray spectroscopy to characterize the altermagnetism in -FeO. Using XMCD spectroscopy, we confirm the altermagnetic nature of the material and, by exploiting the temperature-induced reorientation of the Néel vector, identify two distinct spin–orbit-coupling-induced altermagnetic phases above and below the Morin transition. These results confirm that -FeO exhibits different emergent properties dependent on its Néel vector orientation despite being the same type of -wave altermagnet. Moreover, by identifying an -dependent X-ray absorption contrast (independent of the X-ray polarization), we were able to gain insight into the low temperature phase of the material for which XMCD is not allowed, and further identify the presence of nanoscale spin textures including domain walls and merons. The presence of these nanoscale textures composed with different Néel vector orientation is particularly relevant in the context of potential applications. As the same symmetries lead to both XMCD and AHE, the identification of a domain wall with non-zero XMCD embedded between zero-XMCD domains establishes the possibility of isolated nanotextures that alone produce spintronic signals, offering a route towards high-density altermagnetic devices. The underlying on-and-off switching of emergent properties is a general consequence of the Néel vector orientation in altermagnets, and hence such local modulations are expected to be present in various systems with similar magnetic symmetries, paving the way for nanoscale device applications of a broad class of altermagnetic materials.
Methods
Sample preparation
All samples were fabricated from single-crystalline -FeO naturally grown in the Kalahari region of Botswana and supplied by Surfacenet GmbH (Germany). The samples were shaped to sizes of mm (Sample #1, TEY) and mm (Sample #2, STXM) by the supplier, with the -axis being perpendicular to the sample surface. The supplier further polished the samples from one side (Sample #1) and both sides (Sample #2), respectively.
Sample #2 was thinned down to soft x-ray transparent levels ( nm) via plasma focused ion beam (PFIB) lithography using a Xe plasma. The processing follows the protocol outlined in [49]. An acceleration voltage of 30 kV was used for the etching, and the current was varied between 2.5 A at the start of the etching and 4 nA upon reaching the target thickness. The end point detection for the etching procedure was determined by observing secondary electron emission from the “far side” surface of the -FeO membrane with scanning electron microscopy at 10 kV extractor voltage. To protect the etched membrane, a 5 nm thick carbon film was deposited from the back side of the -FeO sample after the etching by means of thermal evaporation using a CCU-010 HV evaporator from Safematic GmbH. Metallic thin film copper conductors for other experimental purposes (transport and dynamic excitation) were fabricated onto these windows by means of electron beam lithography, thin film deposition, and lift-off processing.
Magnetometry
The magnetization of Sample #1 was measured in the temperature range from 200 to 300 K using a commercial superconducting quantum interference device (SQUID) magnetometer in vibrating-sample mode (VSM) (MPMS, Quantum Design). A magnetic field of 0.01 T was applied along the axis. As shown in Figure 1d, a clear drop in the magnetization was observed at K, corresponding to the Morin transition [41].
X-Ray Absorption Spectroscopy
X-ray absorption spectroscopy (XAS) of Sample #1 was carried out at the XUV diffractometer of the UE46-PGM1 undulator beamline with circularly and linearly polarized x-rays by measuring the total electron yield via sample drain current. The photon energy resolution was set to 90 meV at the Fe- resonances. The circular polarization at the 3rd harmonic is not 100% but rather approximately 90%. For measurements with applied field, the sample was placed between two SmCo permanent magnets at a distance of 4 mm. The field in the center of the sample was approx. 100 mT.
The x-ray magnetic circular and linear dichroism spectra measured were processed in the following fashion. The measured intensity of x-rays is calibrated using the incident photon flux . The linear background is removed from the spectra (). The intensity is normalized by the difference between pre-edge and post-edge. The dichroic spectra were then calculated in the following way.
Here, and correspond to the normalized absorption spectra measured using right- and left-circularly polarized X-rays, respectively, and are used for the XMCD calculation. Similarly, and denote the normalized absorption spectra measured using linearly polarized X-rays with horizontal (LH) and vertical (LV) polarization, respectively, and are used for the XMLD calculation. The difference of opposite polarization is divided by which is the maximum absorption of the material at the edge. Calculating the XMCD/XMLD as a percentage of this value allows us to obtain thickness independent values. These values can then be quantitatively compared to different energies, and furthermore with predicted XMCD/XMLD values from the LDA+DMFT AIM calculations.
Scanning Transmission X-Ray Microscopy
Sample #2 was imaged using scanning transmission x-ray microscopy (STXM) at at the MAXYMUS [50] end station of the UE46 beamline of the BESSY II synchrotron radiation facility of Helmholtz-Zentrum Berlin. X-rays with defined polarization (linear horizontal, linear vertical, right circular, or left circular) and at energies corresponding to the Fe- absorption edges are provided by an undulator insertion device, operating in the third harmonics. The photon energy is further confined in the beamline to meV by a subsequent monochromator. For STXM imaging, the X-rays are focused onto the sample by a diffractive Fresnel zone plate to a spot of nm diameter and both, undiffracted light and higher diffraction orders are blocked by an order selective aperture (OSA) in front of the sample. Photons transmitted through the sample are detected by a point detector. Raster scanning the sample through the beam yields a two-dimensional image. Images were collected at different polarizations and energies for achieving nanoscopic spectral dichroism. The transmission images recorded with right-/left-circular and horizontal-/vertical-linear polarized x-rays are denoted as and , respectively. In order to correct for differences in incident flux or beamline intensity drift between the two polarization states, a normalization coefficient
was applied, where represents the spatial average over the image. The image with right-circular or horizontal-linear polarized x-ray was normalized as
This procedure ensures that the average transmission level is consistent between the two polarization states. The dichroic image in absorption geometry was then computed from the normalized transmission intensities as
In addition to the dichroic signal, the polarization-averaged intensity
and the absorption contrast
were calculated.
X-Ray Magnetic Dichroism Calculations
We simulate the Fe -edge spectra shown in Figure 2 and Figure 3 using density functional theory combined with dynamical mean-field theory (DFT+DMFT), following the implementation of Refs. [51–53]. First, we performed a standard DFT+DMFT calculation in the magnetically ordered phase for the experimental crystal structure of -FeO, as described in Ref. [53]. The Fe -edge spectral intensities were then computed using an Anderson impurity model (AIM), incorporating the spin-dependent hybridization densities obtained from DFT+DMFT, which represent the Weiss field in the antiferromagnetic ordered state. The same parameters for the – and – interactions, as well as the charge-transfer energy , were adopted as in the previous study on Fe -edge resonant inelastic x-ray scattering of -FeO [53]. Since the core hole created by the x-rays is fully localized at the excited Fe site, the total -edge spectrum is obtained by summing the contributions from the AIMs corresponding to the four Fe sites in the unit cell, as successfully applied in previous works for -MnTe [31].
Acknowledgments
We thank Aleš Hrabec for the Pt deposition during the sample fabrication as well as Benny Böhm and Fabian Ganss for initial SQUID measurements on similar samples. Part of the sample fabrication was carried out at the cleanroom facility of the Paul Scherrer Institute. We acknowledge the Helmholtz-Zentrum Berlin für Materialien und Energie for the allocation of synchrotron radiation beamtime. S.M. was funded by the Swiss National Science Foundation (Grant Agreement No. 172517). A.H. was supported by JSPS KAKENHI Grant Numbers 25K00961, 25K07211, the 2025 Osaka Metropolitan University (OMU) Strategic Research Promotion Project (Young Researcher). J.K. was supported by the project Quantum materials for applications in sustainable technologies (QM4ST), funded as project No. CZ.02.01.01/00/22 008/0004572 by Programme Johannes Amos Commenius, call Excellent Research, and by the Ministry of Education, Youth and Sports of the Czech Republic through the e-INFRA CZ (ID:90254). L.Š. acknowledges funding from the ERC Starting Grant No. 101165122 and Deutsche Forschungs-gemeinschaft (DFG) grant no. TRR 288-7422213477 (Projects A09 and B05).
References
Extended Data Figure 2. XMLD image at 724.5 eV. The lines used for the line profiles in Figure 4d and Figure 4e are indicated.
si_v4.pdf
Supplementary Information: Nanoscale imaging of spin textures with locally varying altermagnetic response in -FeO
Characterization of single crystals
The single crystallinity and lateral crystallographic orientation of Sample #1 was confirmed by Laue measurements. The Laue pattern of Sample #1 is displayed in Supplementary Figure 1.

Figure S1. Supplementary Extended Data Figure 1. Laue diffraction pattern of single crystalline -FeO (Sample #1).
Extended Data Figure 1. Overview absorption images from the STXM experiments. The metallic conductor regions are highlighted in yellow as a guide to the eye.
Symmetry analysis
According to our spin group theory analysis, -FeO below the Néel temperature K is described by the spin point group and thus belongs to the -wave symmetry altermagnets [1]. As illustrated in Supplementary Figure 2, the four spin sublattices in a rhombohedral unit cell exhibit, in addition to antiparallel dipolar order, ferroically ordered -wave spin densities [2, 3, 4]. This ferroic ordering gives rise to a -wave spin-polarized electronic structure, shown in Figure 1b. Three vertical spin-degenerate nodal surfaces are symmetry-enforced to be flat. In contrast, the fourth in-plane nodal plane is not symmetry constrained and, owing to its curved shape, forms additional accidental crossings in the plane [5]. This nodal topology [6] is therefore distinct from that of the -wave altermagnet -MnTe, which exhibits a flat horizontal nodal surface.
Supplementary Figure 2. Magnetic structure of -FeO in a rhombohedral unit cell. Antiferromagnetic ordering on the four sublattices (A-D) breaks space-time inversion () symmetry with respect to the inversion center highlighted by the red cross. The collinear spin order, together with the oxygen octahedra, stabilizes a -wave altermagnetic order that exhibits ferroic alignment across the four spin sublattices, as illustrated in the left column.
Magnetic structures of -FeO and expected XMCD/XMLD contrast
Supplementary Figure 3 schematically summarizes the magnetic structure of -FeO and the XMCD and XMLD contrast expected for the experimental geometry used in this work. For temperatures below the Morin transition (), the spins are collinearly aligned along the -axis. In this phase, XMCD measured with the X-ray wave vector -axis vanishes. In the temperature range , the spins rotate into the basal plane, leading to a finite XMCD signal, depending on the Néel vector azimuthal orientation and sign ().
In contrast, XMLD is sensitive to the orientation of the Néel vector with respect to the linear polarization direction of the incident X-rays, enabling the identification of antiferromagnetic domains.

Figure S3. Crystal and magnetic structures of -FeO and the expected XMCD and XMLD contrast for the experimental geometry used in this work. a. Magnetic structure below the Morin transition temperature (), where magnetic moments are aligned along the -axis. XMCD measured with the X-ray wave vector is expected to vanish. b. Spin configuration in the range , where spins lie in the basal plane, resulting in a finite XMCD contrast, depending on . c,d Expected XMLD contrast for different in-plane orientations of the Néel vector.
Analysis of scanning transmission X-ray microscopy data
Calibrating thickness variation of absorption image
To improve visibility of the magnetic contrast, the contribution from the metallic lead stripline in the absorption image shown in Supplementary Figure 4 was calibrated using the following procedure and obtained the image displayed in Figure 3e. In this procedure, an absorption image acquired at low temperature, where no magnetic features are present, was used as reference. The reference image was rescaled and spatially aligned to the corresponding target image, and the intensities were normalized using the signal at the central region of the field of view, as shown in Supplementary Figure 4. The normalized target absorption image was then calibrated by the reference image according to
Supplementary Figure 4. a Absorption image prior to background removal. The white region is the metallic lead stripline. b Reference absorption image acquired at low temperature, which is rescaled, spatially aligned, and intensity-normalized using the signal at the central region of the field of view.
where the minus sign is introduced to ensure consistency with the contrast convention used for the calculations displayed in Figure 3c. The absorption image obtained using this procedure is shown in Figure 3e.
Quantitative analysis of dichroic signal
To quantitatively evaluate the XMCD obtained from the transmission experiment, we use the following definition,
Here, denotes the absorption peak value at the edge. Transmission images for left- and right-circularly polarized X-rays, , were acquired as a function of photon energy, and was obtained by the normalization procedure described in the Methods section.
In the transmission experiment performed at the MAXYMUS end station, the large sample thickness prevents a reliable determination of the -edge absorption peak from the transmission data. Instead, is estimated using X-ray absorption spectra measured in total electron yield (TEY) mode.
To relate the imaging data to the spectroscopy measurements, the peak intensities at the edge obtained from both datasets are compared by defining the peak ratio
Using this ratio, the absorption maximum expected for the transmission imaging experiment is estimated as
The resulting is then used in Eq. (2) to quantitatively evaluate the XMCD obtained from the transmission experiment. The XMCD signal obtained by this procedure reaches approximately 1.5% at the edge.
Supplementary Simulations
Supplementary Figures 5 and 6 compare the simulated Fe -edge XMCD, XMLD and XAS spectra in the high-temperature altermagnetic phase calculated using the DFT+DMFT method with those obtained from an Fe atomic model. The crystal-field parameters in the atomic model are taken from a DFT calculation by constructing an effective Fe -space model in which the O states are projected out. Since -FeO is a good Mott insulator with localized Fe electrons, it is not surprising that the DFT+DMFT results and the atomic model yield similar absorption spectra. The broader high-energy feature around 715 eV in the DFT+DMFT spectrum, which shows improved agreement with the experimental data, is provided by the explicit inclusion of hybridization with the O bands. In the atomic model, a small molecular field is introduced to simulate the Néel state of this phase. A small in-plane component is included to account for the canting of the magnetic moment, whose amplitude is adjusted to reproduce the experimental magnetization of /Fe. Supplementary Figure 7(a) shows the optical conductivity tensor at the Fe edge calculated using the atomic model, and the corresponding Hall vector is shown in Supplementary Fig. 7(b). The XMCD is given by , where is the X-ray wave vector [7]. In the adopted coordinate system, the , , and axes correspond to the , , and axes shown in Figure 1 of the main text, respectively. The in-plane components of the Hall vector are symmetry-allowed but negligibly small compared with the out-of-plane component . We have also checked that the presence or absence of the in-plane canted moment does not alter , confirming that it represents the altermagnetic contribution. Finally, Supplementary Fig. 8 shows the XAS and XMCD spectra calculated using the DFT+DMFT method for a ferromagnetic configuration with magnetic moments aligned parallel to the axis. A large XMCD signal appears, as expected, but its frequency dependence is significantly different from that of the altermagnetic case. This is another demonstration of the fact that the observed XMCD spectrum is not due to any net magnetic moment parallel to the axis, which is not observed but is symmetry allowed.
Figure S6. Dependence of the XAS spectra on the orientation (IP: in-plane, OOP: out-of-plane), calculated by (a) the DFT+DMFT method (same as Figure 3c in the main text) and b the Fe atomic model.
Figure 7. Supplementary Fig. 7. a Real (red) and imaginary (blue) parts of the optical conductivity tensor at the Fe edge. b Imaginary part of the Hall vector . The spectra are calculated using the Fe atomic model.
Figure 8. Supplementary Fig. 8. a Fe -edge XAS spectra calculated for circularly polarized light and b the corresponding XMCD spectra for a ferromagnetic configuration along the axis. Here the DFT+DMFT method is employed.
References
References
- [1]L. Šmejkal, J. Sinova, and T. Jungwirth, Beyond Conventional Ferromagnetism and Antiferromagnetism: A Phase with Nonrelativistic Spin and Crystal Rotation Symmetry, Phys. Rev. X 12, 031042 (2022).DOI
- [2]S. Bhowal and N. A. Spaldin, Ferroically ordered magnetic octupoles in d-wave altermagnets, Physical Review X 14, 011019 (2024).DOI
- [3]X. Verbeek, D. Voderholzer, S. Schären, Y. Gachnang, N. Spaldin, and S. Bhowal, Nonrelativistic ferromagnetotriakontadipolar order and spin splitting in hematite, Physical Review Research 6, 043157 (2024).
- [4]R. Jaeschke-Ubiergo, V.-K. Bharadwaj, W. Campos, R. Zarzuela, N. Biniskos, R. M. Fernandes, T. Jungwirth, J. Sinova, and L. Šmejkal, Atomic altermagnetism, arXiv preprint arXiv:2503.10797 (2025).
- [5]R. Hoyer, P. P. Stavropoulos, A. Razpopov, R. Valentí, L. Šmejkal, and A. Mook, Altermagnetic splitting of magnons in hematite α-Fe₂O₃, Phys. Rev. B 112, 064425 (2025).DOI
- [6]I. Mazin, R. González-Hernández, and L. Šmejkal, Induced Monolayer Altermagnetism in MnP(S,Se)₃ and FeSe (2023), arXiv:2309.02355 [cond-mat.mes-hall].arxiv.org/abs/2309.02355
- [7]A. Hariki, A. Dal Din, O. J. Amin, T. Yamaguchi, A. Badura, D. Kriegner, K. W. Edmonds, R. P. Campion, P. Wadley, D. Backes, L. S. I. Veiga, S. S. Dhesi, G. Springholz, L. Šmejkal, K. Výborný, T. Jungwirth, and J. Kuneš, X-Ray Magnetic Circular Dichroism in Altermagnetic α-MnTe, Phys. Rev. Lett. 132, 176701 (2024).DOI