# Fileset

[Heppell_2025_2D_Mater._12_025001.pdf](https://mdr.nims.go.jp/filesets/3a7aed5e-96c8-4c7d-a515-fad1e79c1b99/download)

## Creator

Emily Heppell, Ryuji Fujita, Gautam Gurung, Jheng-Cyuan Lin, Andrew F May, Michael Foerster, M Waqas Khaliq, Miguel Angel Niño, Manuel Valvidares, Javier Herrero-Martín, Pierluigi Gargiani, [Kenji Watanabe](https://orcid.org/0000-0003-3701-8119), [Takashi Taniguchi](https://orcid.org/0000-0002-1467-3105), Dirk Backes, Gerrit van der Laan, Thorsten Hesjedal

## Rights

[Creative Commons BY Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)

## Other metadata

[Room-temperature in-plane ferromagnetism in Co-substituted Fe<sub>5</sub>GeTe<sub>2</sub> investigated by magnetic x-ray spectroscopy and microscopy](https://mdr.nims.go.jp/datasets/63abb849-c296-45d9-ade5-c77666074b4f)

## Fulltext

Room-temperature in-plane ferromagnetism in Co-substituted Fe5GeTe2 investigated by magnetic x-ray spectroscopy and microscopy2D Materials     PAPER • OPEN ACCESSRoom-temperature in-plane ferromagnetism in Co-substituted Fe5GeTe2 investigated by magnetic x-ray spectroscopy and microscopyTo cite this article: Emily Heppell et al 2025 2D Mater. 12 025001 View the article online for updates and enhancements.You may also likeInterface engineering of van der Waalsheterostructures towards energy-efficientquantum devices operating at hightemperaturesManh-Ha Doan and Peter Bøggild-MXenes: exploiting their unique propertiesfor designing next-generation thermalcatalysts and photocatalystsJoshua O Ighalo, Morgen L Smith, AhmedAl Mayyahi et al.-Recent progress in two-dimensionalpolymer materials: interfacial synthesisand applicationsLili Ma, Wenbo Hou, Anbai Li et al.-This content was downloaded from IP address 144.213.253.16 on 15/07/2025 at 09:24https://doi.org/10.1088/2053-1583/ada040/article/10.1088/2053-1583/ada043/article/10.1088/2053-1583/ada043/article/10.1088/2053-1583/ada043/article/10.1088/2053-1583/ada043/article/10.1088/2053-1583/ada042/article/10.1088/2053-1583/ada042/article/10.1088/2053-1583/ada042/article/10.1088/2053-1583/ada961/article/10.1088/2053-1583/ada961/article/10.1088/2053-1583/ada9612D Mater. 12 (2025) 025001 https://doi.org/10.1088/2053-1583/ada040OPEN ACCESSRECEIVED30 September 2024REVISED11 November 2024ACCEPTED FOR PUBLICATION17 December 2024PUBLISHED6 January 2025Original Content fromthis work may be usedunder the terms of theCreative CommonsAttribution 4.0 licence.Any further distributionof this work mustmaintain attribution tothe author(s) and the titleof the work, journalcitation and DOI.PAPERRoom-temperature in-plane ferromagnetism in Co-substitutedFe5GeTe2 investigated by magnetic x-ray spectroscopy andmicroscopyEmily Heppell1,2,3,9, Ryuji Fujita1,9, Gautam Gurung1,4, Jheng-Cyuan Lin1, Andrew F May5,Michael Foerster6, MWaqas Khaliq6, Miguel Angel Niño6, Manuel Valvidares6,Javier Herrero-Martín6, Pierluigi Gargiani6, Kenji Watanabe7, Takashi Taniguchi8,Dirk Backes2,∗, Gerrit van der Laan2,∗ and Thorsten Hesjedal1,2,∗1 Clarendon Laboratory, Department of Physics, University of Oxford, Parks Road, Oxford OX1 3PU, United Kingdom2 Diamond Light Source, Harwell Science and Innovation Campus, Didcot, Oxfordshire OX11 0DE, United Kingdom3 STFC, ISIS, Rutherford Appleton Laboratory, Didcot OX11 0QX, United Kingdom4 Trinity College, University of Oxford, Oxford OX1 3BH, United Kingdom5 Materials Science and Technology Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831, United States of America6 ALBA Synchrotron, Carrer de la Llum 2-26, 08290 Cerdanyola del Vallès, Barcelona, Spain7 Research Center for Electronic and Optical Materials, National Institute for Materials Science, 1-1 Namiki, Tsukuba 305-0044, Japan8 Research Center for Materials Nanoarchitectonics, National Institute for Materials Science, 1-1 Namiki, Tsukuba 305-0044, Japan∗ Authors to whom any correspondence should be addressed.9 These authors contributed equally.E-mail: Dirk.Backes@diamond.ac.uk, Gerrit.vanderLaan@diamond.ac.uk and Thorsten.Hesjedal@physics.ox.ac.ukKeywords: 2D van der Waals magnets, (Co0.28Fe0.72)5GeTe2, ferromagnetism, x-ray absorption spectroscopy,x-ray magnetic circular dichroism, XPEEMSupplementary material for this article is available onlineAbstractThe exploration of two-dimensional (2D) van der Waals ferromagnets has revealed intriguingmagnetic properties with significant potential for spintronics applications. In this study, weexamine the magnetic properties of Co-doped Fe5GeTe2 using x-ray photoemission electronmicroscopy (XPEEM) and x-ray magnetic circular dichroism (XMCD), complemented by densityfunctional theory calculations. Our XPEEMmeasurements reveal that the Curie temperature (TC)of a bilayer of (CoxFe1−x)5−δGeTe2 (with x= 0.28) reaches∼300 K—a notable enhancement overmost 2D ferromagnets in the ultrathin limit. Interestingly, the TC shows only a small dependenceon film thickness (bulk TC ≈ 340K), in line with the observed in-plane (IP) magnetic anisotropyand robust IP exchange coupling. XMCD measurements indicate that the spin moments for bothFe and Co are significantly reduced compared to the theoretical values. These insights highlight thepotential of Co-doped Fe5GeTe2 for stable, high-temperature ferromagnetic applications in 2Dmaterials.1. IntroductionThe discovery and exploration of two-dimensional(2D) materials have sparked significant interest dueto their unique electronic, optical, and magneticproperties, which differ markedly from their bulkcounterparts [1–4]. Among these, van der Waals(vdW) ferromagnets have emerged as a particularlyexciting class of materials [5, 6] for potential applic-ations in spintronics and data storage [7, 8], due totheir ability to retain magnetic order down to mono-layer thicknesses [9, 10], and their ease of integrationinto functional heterostructures [11–13]. Fe3GeTe2was one of the first materials in which a persistentlong-range magnetic order down to the monolayerlimit was demonstrated, possessing a large perpen-dicular magnetic anisotropy and a high Curie tem-perature, TC, of 230 K in the bulk [14] and 130 Kin the monolayer limit [15]. The recently discovered2D vdW ferromagnets, Fe5GeTe2 and Fe3GaTe2, have© 2025 The Author(s). Published by IOP Publishing Ltdhttps://doi.org/10.1088/2053-1583/ada040https://crossmark.crossref.org/dialog/?doi=10.1088/2053-1583/ada040&domain=pdf&date_stamp=2025-1-6https://creativecommons.org/licenses/by/4.0/https://creativecommons.org/licenses/by/4.0/https://orcid.org/0000-0002-3977-4623https://orcid.org/0000-0001-8557-5812https://orcid.org/0000-0001-7706-6276https://orcid.org/0000-0002-5085-3234https://orcid.org/0000-0003-0777-8539https://orcid.org/0000-0002-4147-6668https://orcid.org/0000-0002-9696-3498https://orcid.org/0000-0003-3692-147Xhttps://orcid.org/0000-0003-4895-8114https://orcid.org/0000-0003-1986-8128https://orcid.org/0000-0002-6649-0538https://orcid.org/0000-0003-3701-8119https://orcid.org/0000-0002-1467-3105https://orcid.org/0000-0002-1019-3323https://orcid.org/0000-0001-6852-2495https://orcid.org/0000-0001-7947-3692mailto:Dirk.Backes@diamond.ac.ukmailto:Gerrit.vanderLaan@diamond.ac.ukmailto:Thorsten.Hesjedal@physics.ox.ac.ukhttps://doi.org/10.1088/2053-1583/ada0402D Mater. 12 (2025) 025001 E Heppell et alFigure 1. Crystal structure and optical micrograph of bilayer flake. Crystal structure of (CoxFe1−x)5GeTe2 in the (a) R3m and (b)P3̄m1 space group. The Fe5GeTe2 monolayer has five Fe-containing Fe layers (termed Fe1–Fe5), while the Te atoms are located inthe two outermost layers and the Ge layer is in the middle of the stack [19, 20]. The purple and golden spheres represent Ge andTe atoms, and the blue-bronze ones Fe or Co atoms (partial occupation) on Fe1, Fe2, Fe3, Fe4 and Fe5 sites, respectively. Note thatfor the common R3̄m space group of (CoxFe1−x)5GeTe2, Fe1 is a split site, i.e. it can occupy up or down positions relative to Ge.Here, we only consider Fe1 in the up position, making the structure belong to the R3m space group instead. The Fe1–Te andFe4–Ge bonds form buckled hexagonal structures akin to graphene. The two are shifted by 1/3 of a diagonal unit cell vector withrespect to each other. In contrast, Fe2, Fe3, and Fe5 are either above or below the hollow sites of these two layers [18, 27]. (c)Optical micrograph showing a triangular CoF5GT bilayer flake exfoliated on PDMS, which was investigated by XPEEM. Thewhite scale bar has a length of 10µm. (d) AFM linescan across the edge of the flake, as indicated in red, revealing a(homogeneous) bilayer thickness of∼2 nm. (e) x-ray diffraction data (00l reflections) of the investigated 28% Co crystal(middle), along with an undoped (above) and a 54% Co crystal (below). Reprinted figure with permission from [27], Copyright(2020) by the American Physical Society.attracted considerable attention owing to their room-temperature Curie temperatures and tunable mag-netic properties, making them promising candid-ates for both fundamental research and practicalapplications [10, 16, 17].Fe5GeTe2 exhibits ferromagnetic (FM) orderingwith a TC of ∼270–310 K, which is relatively highamong 2D ferromagnets [18–20] and close to therivaling class of doped semiconducting transitionmetal dichalcogenides (with a monolayer TC of upto 360 K) [21]. This property opens the door topotential room-temperature applications [8, 22–25].Furthermore, Fe5GeTe2 displays magnetic anisotropyand a layered structure, enabling the exfoliation offew-layer and monolayer samples while maintainingits FM properties [18, 26]. However, the magneticproperties of Fe5GeTe2 can be further enhanced andtailored through chemical doping, which introducesadditional degrees of freedom in tuning its electronicand magnetic behavior [27–29].The substitution of Co atoms for Fe in theFe5GeTe2 lattice is influencing the exchange interac-tions and the density of states near the Fermi level,thereby altering its magnetic characteristics. In fact,(Co,Fe)5GeTe2 hosts both FM and antiferromagnetic(AF) order depending on the degree of Co substitu-tion, whereby a critical Co concentration of 0.4 hasbeen reported, above which AF order emerges [27,28].Moreover, for a concentration of 0.44, FMandAForder coexist at room temperature with odd-layer FMin the monolayer limit [30]. However, owing to therichness of the magnetic phenomena in this mater-ial, the magnetic ground state is complex, and so far,element-specific resonant magnetic x-ray scatteringhad only been carried out in the AF state [29].The properties of the parent compoundFe5GeTe2,which crystallizes in the rhombohedral space groupR3̄m, are governed by Fe vacancies and the thermalcycling history [31], mostly affecting the Fe1 site (seefigure 1(a)). As the exact stoichiometry can differ,Fe5GeTe2 is generally referred to as Fe5−δGeTe2. Notethat (CoxFe1−x)5−δGeTe2 with x< 0.4 crystallizes, inprinciple, in the same space group [18]. However, dueto the large density of stacking faults, there is likelysignificant mixing of the primitive AA and rhombo-hedral ABC stacking orders in these Co-substitutedbulk crystals [32] (compare figures 1(a) and (b)).Here, we present an investigation of themagnetic properties of the FM 2D material(CoxFe1−x)5−δGeTe2 (with x= 0.28; CoF5GT there-after) using a combination of element-specific, x-ray-based magnetic spectroscopy and microscopy,supported by theoretical calculations. We aim toelucidate how Co doping affects the Curie temper-ature, magnetic anisotropy, and coercive field ofFe5GeTe2. Understanding these changes is crucialfor optimizing the material for various technologicalapplications.22D Mater. 12 (2025) 025001 E Heppell et al2. Materials andmethods2.1. Co-doped Fe5GeTe2 sample preparationThe CoF5GT bulk single crystals were synthes-ized via an iodine-assisted chemical vapor transporttechnique, and structurally, chemically and magnet-ically characterized in detail using x-ray diffraction,energy-dispersive spectroscopy (EDS) and SQUIDmagnetometry, respectively, as previously discussedin [27]. For magnetic spectroscopy, bulk crystalswere cleaved in an Ar-filled glovebox connected tothe synchrotron end station. For the magnetic x-ray microscopy study, ultrathin layers of CoF5GTwere exfoliatedwith a polymethyldisiloxane (PDMS)-assisted technique in an Ar-filled glovebox, andencapsulated between thin layers of hBN to preventoxidation.Figure 1(c) shows an optical micrograph of thebilayer (2L) CoF5GT flake exfoliated over PDMS.The flake thickness was determined via atomic forcemicroscopy (AFM), as shown in figure 1(d), scanningacross a step-edge. Figure 1(e) displays x-ray diffrac-tion data (00l reflections) comparing the x= 0.28 Co-substituted sample (which is the focus of this study)with undoped Fe5GeTe2 and a sample with a veryhigh Co concentration (x= 0.54), revealing the struc-tural influence of Co substitution. The Co concen-tration was determined by EDS to be x= 0.28 [27].The results of the structural study are discussed indetail in [27] and can be briefly summarized as fol-lows: First, the x= 0.28 crystal is strongly affectedby stacking disorder, preventing a refinement of thestructure. In contrast, for the x= 0.54 sample, stack-ing faults present no issue for the refinement, yieldingspace group P3̄m1. Second, Co incorporation leadsto a slight contraction in the ab plane, along withan increase in both the slab thickness and interlayerspacing.2.2. Magnetic domain imaging with x-rayphotoemission electronmicroscopy (XPEEM)XPEEM measurements were conducted at the BL24-CIRCE beamline at the ALBA Synchrotron inBarcelona, Spain [33]. Real-space imaging was con-ducted at the Fe L3 (706.2 eV) and Co L3 (769.0 eV)edges in the temperature range from 250 to 310 K[34]. The fixed angle of incidence of the incident x-rays with respect to the sample surface was 16◦, whichmeans that 28% of the sample’s out-of-plane (OOP)magnetization component is projected along the x-ray propagation direction [35, 36]. Here, the x-raymagnetic circular dichroism (XMCD) asymmetry isdefined as (µ−max −µ+max)/(µ−max +µ+max), where µ−maxand µ+max are the XAS signals at the maximum dif-ference taken with left and right circularly polarizedx-rays, respectively.2.3. Magnetic x-ray spectroscopyX-ray absorption spectroscopy and magnetic circu-lar dichroism measurements were conducted at theHECTOR end station on beamline 29 (BOREAS) atthe ALBA Synchrotron [37]. XAS spectra across theFe (695–745 eV), Co (765–810 eV), and Ge (1200–1290 eV) L2,3, as well as the Te M4,5 (565–595 eV)edges were recorded using right- and left-circularlypolarized (RCP and LCP) x-rays at normal incidence(NI), with a magnetic field of up to 6 T applied alongthe beam direction. The XAS signal, Isum, is defined asIsum = µ− +µ+, while the XMCD signal is defined asIxmcd = µ− −µ+ [38]. The measurements were per-formed in total-electron yield detectionmode, sensit-ive to the upper 3–5 nm of the sample [38], at a tem-perature of 10 K unless otherwise specified.Spin and orbitalmagneticmoments for Fe andCowere determined using sum rule analysis [39]. Usingthe scale shown on the right-hand side of the panelsin figures 3(c) and (d), the quantities p=´L3dEIxmcdand q=´L3,L2dEIxmcd can be determined, and fromfigures 3(e) and (f), the quantity r=´L3,L2dEIsum.Accounting for the number of holes in the 3d shellnh = 10− nd, which is assumed to be 4 for Fe2+ and 3for Co2+, the spin and orbitalmoments per Fe andCoatom were obtained asmorb =−⟨Lz⟩=−(4/3)qnh/rand mspin,eff =−2⟨Seff,z⟩=−(6p− 4q)nh/r [38].Alsomspin,eff =mspin − 7⟨Tz⟩.Note that experimental limitations of the sumrule analysis can arise in multiple ways [40], suchas caused by saturation effects, jj mixing, continuumbackground corrections, canted spinmoments, arbit-rary integration limits, estimation of the hole num-ber, omission of linear dichroism in sum spectrum,and neglect of the magnetic dipole term. All theseeffects, which on their own can either decrease orincrease the moment value, are usually assumed toaccumulate to a total error for Co and Fe of ∼10%[41].2.4. Density functional theory (DFT) calculationsDFT calculations were performed using VASP [42].We applied the generalized gradient approximation[43] using the projector augmented wave method[44]. Spin–orbit coupling was included in all the cal-culations. vdW interactions between the 2D layerswere considered using DFT-D3 method with Becke–Johnson damping [45]. The bulk atomic structure ofFe5GeTe2 with space group R3̄m allows split sites forFe1 (that can occupy up or down positions relative toGe). In this work, we only considered a simple bulkstructure with space group R3m by fixing Fe1 in theup positions. We fixed the lattice parameters to thereported experimental values (a= 4.04 Å, c= 29.25 ÅforR3m [18] and a= 4.02 Å, c= 9.8 Å forP3̄m1 [27]),and all atomic positions were relaxed using a 20×20× 4 Γ-centered k-point grid with a kinetic energy32D Mater. 12 (2025) 025001 E Heppell et alcutoff of 500 eV. For the supercell structure, a 11×11× 2 Γ-centered k-point grid was used. The relax-ation process was repeated until the atomic forceson each atom were less than 0.01 eVÅ−1. The dop-ing effect in (CoxFe1−x)5GeTe2 was simulated usingthe virtual crystal approximation (VCA), and, addi-tionally, by constructing a√3×√3× 1 supercell toaccount for substitutional Co doping at Fe1 positions.Our previous DFT calculations suggest that the Fe1site is the most favorable for Co doping [27]. Holedoping simulations were performed by reducing thetotal number of electrons via a compensating uniformbackground charge. Themagnetic anisotropywas cal-culated using the magnetic force theorem with anelectronic convergence of 10−8 eV. The plots in figure1 were created using VESTA [46].3. Results and discussion3.1. XPEEM study of bilayer CoF5GTXPEEM images of FM domains in bilayer CoF5GTwere acquired as a function of temperature. TheXPEEM image taken at 259 K (figure 2(a)) showslarge, µm-sized domains (white or black contrast). Asthe magnetic contrast is governed by the easy planecharacter of CoF5GT (see magnetic spectroscopy res-ults below), themagneticmoments within these blackand white domains are oriented in-plane (IP) and inopposite directions to one another. In-between, smal-ler domains are found, which disappear first uponheating the sample to 282 K. A general loss of con-trast is observed as the temperature is increased to296 K and above, i.e. closer to TC. However, the largerdomains remainmagnetic up to 302 K. No qualitativechange of themagnetic contrast is observed as a func-tion of temperature. From the loss of magnetic con-trast, the CoF5GT bilayer TC can be estimated to be∼302 K (figure 2(d)), whereas the Co concentration-dependent bulk TC was previously determined to be∼340 K for this Co stoichiometry [27]. Compared toother 2D Ising ferromagnets, including Fe3GeTe2 [15]and CrGeTe3 [10], the relatively high TC of CoF5GT[47] appears to depend less significantly on the num-ber of layers, suggesting a relatively weak intralayerexchange interaction. On the other hand, the high-TC ferromagnetism in CoF5GT may be stabilized byintralayer exchange interactions confined within theunit cell.To shed light on the microscopic origin of thisbehavior, we carried out XMCD measurements andDFT calculations, which are discussed in detail below.3.2. Orbital and spin magnetic moments of bulkCoF5GTThe orbital and spinmagnetic moments of the Fe andCo sites were obtained from sum rule analysis of theXAS and XMCD spectra [48, 49], measured in totalelectron yield detection mode. In figures 3(a), (c), (e)and (b), (d), (f), XAS and XMCD spectra of Fe andCo, respectively, are shown as measured at 10 K inan applied field of 6 T in grazing incidence (GI). TheXAS spectra shown in figures 3(a) and (b) were takenwith 100% RCP and LCP (labeled µ+ and µ− in thefigure), and the respective XMCD spectra shown inthe panels below calculated by subtracting them fromeach other (figure 3(c) and (d)). A step function witha ratio of 2:1 for the L3 and L2 edges, respectively,was subtracted from the XAS spectra to remove thecontinuum-state background before integration. Theintegrals p and q over the L3 and L2,3 edges, respect-ively, of the XMCD used for the sum rule analysis areindicated in figures 3(c) and (d). The integrated XASintensity over the L2,3 edges provides the normaliza-tion factor r (figures 3(e) and (f)). For further details,we refer to section S1.3.In table 1, the results of the sum rule analysisare summarized. The spectra were taken at temper-atures of 300 and 10 K and in an applied magneticfield of 6 T at both GI and NI to probe the IPand OOP components of the magnetization, respect-ively. (The measurement geometry is illustrated infigure 5(f)). At 6 T, mspin,eff reaches 2.5µB atom−1for Fe and 1.4µB atom−1 for Co at 10 K, which isabout 30% reduced at 300 K [in the case of Fe].These values for the spin moments are not far outof line with those for bcc Fe (1.98µB atom−1) andhcp Co (1.62µB atom−1) [50]. Also, they are compar-able with reports on undoped Fe5GeTe2 bulk crystalsof mspin,eff = 1.8µB/Fe (morb = 0.1µB/Fe, in generalagreement with our findings) [51]. For a further dis-cussion, see section S1.3.Table 1 also lists the orbital moments, morb, andthe ratio of orbital-to-spin moments, morb/mspin,eff.We note that the absolute values of morb are lar-ger for Fe in the NI geometry but larger for Co inthe GI geometry, pointing towards differences in themagnetocrystalline anisotropy energies. This findingwill be discussed in the context of the magnetic hys-teresis loop measurements in the magnetic aniso-tropy section below. Further, the ratio morb/mspin,effis a useful indicator of the importance of the orbitalproperties, as in this ratio, the number of holes andthe varying saturation fields cancel out. Accordingto Hund’s third rule, the spin and orbital momentsfor the more-than-half-filled Fe and Co band arealigned parallel. Indeed, from the sum rule analysis,we find morb/mspin,eff > 0 in both cases. The exper-imental values of the magnetic moments will be fur-ther discussed in the context of the DFT results below.3.3. Magnetic spectroscopy at the Ge L2,3 and TeM4,5 edgesApart from the element-selective analysis of the mag-netic Fe and Co ions at their L2,3 edges, giving XMCDsignals which are proportional to their 3d magnetic42D Mater. 12 (2025) 025001 E Heppell et alFigure 2. XPEEMmeasurements as a function of temperature. The magnetic contrast at (a) 259 K and (b) 282 K is dominated bylarger, µm-sized domains. (c) At 296 K, the magnetic contrast is reduced, and at (d) 302 K, it has almost completely vanished,except for larger domains. Examples of both larger (blue) and smaller (yellow) domains are indicated in panel (a).Figure 3. Right- (red, µ+) and left- (blue, µ−) circularly polarized (a), (b) XAS, (c), (d) XMCD (solid green line) and integratedXMCD (dotted green line), and (e), (f) summed (solid black line) and integrated (dotted black line) XAS spectra over the Fe L2,3(left column) and Co L2,3 (right column) edges. The spectra were measured in total electron yield in grazing incidence at an angleof 20◦ (off the surface plane) at 10 K under a 6 T magnetic field applied along the same grazing incidence direction. The gray stepfunction in (a), (e) and (b), (f) denotes the continuum-state background subtracted before integration.Table 1. Results of the sum rule analysis for the measurements at10 K, in an applied field of 6 T, for grazing and normal incidence(GI/NI). The magnetic momentsmspin,eff andmorb are given inunits of µB/Fe or Co atom. The number of holes was assumed tobe nh = 4 for Fe and nh = 3 for Co. An error on the order of 10%can typically be assumed for the magnetic spin moments of Coand Fe [41].Element Geometry morb mspin,eff morb/mspin,effFe GI 0.148 2.45 0.0604NI 0.167 2.53 0.0661Co GI 0.174 1.39 0.125NI 0.135 1.25 0.108moment, also the ‘non-magnetic’ elements Ge and Tecan be analyzed. This is important since these ele-ments can exhibit a net spin polarization throughhybridization of their p bands with the 3d bandof Fe or Co, giving an XMCD signal. Figures 4(a)and (b) show the Ge L2,3 and TeM4,5 edge XAS (top)and XMCD (bottom) spectra, respectively, meas-ured at 10 K in an applied magnetic field of 6 Tat GI. This observation of spin-polarized Ge andTe states in CoF5GT is consistent with their obser-vation in undoped Fe5GeTe2 [51]. The sizeable Te-XMCD (observed in Fe5GeTe2) has been argued byYamagami et al [51] to be hinting at the major rolespin–orbit coupling of this heavy element could playin the arrangement of the Fe spins, including mag-netocrystalline anisotropy, similar to the role of Pt inthe itinerant ferromagnet FePt [52].For Ge (for which the 3d shell should be fullyoccupied), the L2,3 (2p→ 3d) edges are located at∼1240 eV. Here, the same XMCD sum rule analysis asfor Fe and Co can be applied, using a step height ratio52D Mater. 12 (2025) 025001 E Heppell et alFigure 4. Right- (red, µ+) and left- (blue, µ−) circularly polarized XAS and XMCD (green) over the (a) Ge L2,3 and (b) TeM4,5edges in total electron yield. The measurements were performed in grazing incidence (at 20◦ off the surface plane) at 10 K under afield of 6 T applied along the incident beam direction.of L3:L2 = 2:1 in the XAS background correction. Onthe other hand, for Te with itsM4,5 edges at∼570 eV,i.e. a 3d→ 5p transition, different coefficients for theXMCDsumrule are used [53]. Sum rule analysis givesat the M4 edge, Ixmcd/Isum = 3/2⟨Sz⟩/nh and at theM5 edge, Ixmcd/Isum =−⟨Sz⟩/nh, where, as usual, theexpectation value of the spin moment is defined asmspin =−2⟨Sz⟩ and nh is the number of holes [38].The step height in the XAS background correctionfor M5:M4 is 3:2. Note that as the XMCD measuresthe moment on the 5p electrons, we can neglect thesmall orbital contribution for Te. Also, note that theTe M4,5 absorption (3d→ 5p transition) reverses theXMCD sign compared to the p→ d transition (L2,3).Therefore, since the TeM5 and Fe L3 XMCD have thesame sign, the Te 5p and Fe 3d moments are alignedantiparallel.As the Te spectra possess a sloping backgroundthat curls significantly at the M4 edge and beyond,we focused on the M5 edge for the sum rule ana-lysis. Linear backgrounds fitted to the M5 pre-edgewere subtracted from the Te spectra. The height of theM5 step for the XAS background correction was setto the lowest intensity of RCP and LCP between theM5 andM4 peaks (with respect to the pre-edge back-ground). This way, from sum rule analysis for Te, spinmoments ofmGIspin,eff = 2 · Ixmcd/Isum = 0.092µB/holeand mNIspin,eff = 0.113µB/hole were determined fromthe GI and NI data at 10 K in a field of 6 T, respect-ively. The presence of a relatively large spin momentms per hole of −0.092µB on the Te site is evidenceof a significant intralayer hybridization between theFe 3d and Te 5p orbitals in bulk CoF5GT. To reachagreement between the DFT calculations (see below)and our experimental value,∼0.5 holes per Te have tobe assumed, giving ms =−0.045µB/Te atom, whichconfirms the strong hybridization between the metal3d and Te 5p bands.Moreover, in the plot of the Fermisurfaces of Fe and Te (figure S1), we find a stronghybridization of the Fe and Te bands near theΓ point.This is further enhanced in Co-substituted Fe5GeTe2,as can be seen in the atom-projected density-of-states(DOS) plots in figure S3.The unique temperature evolution of the mag-netic properties of undoped (and doped) Fe5GeTe2[18], which includes the appearance of non-collinearspin structures and ferrimagnetic behavior [54, 55],was explained by Yamagami et al [51] to be governedby the hybridization between the Fe1 site and Te. Interms of bond distance, Fe1 is the second closest toTe (2.61 Å), however, it forms a buckled hexagonalstructure like graphene, while for the closest Fe site,Fe4 and Te are vertically stacked [20]. This bond con-figuration is absent in the closely related vdW mag-net Fe3GeTe2, which has a much lower transitiontemperature.3.4. Magnetic anisotropyTo determine the magnetic anisotropy of the sample,and to obtain unambiguous experimental proof of thehybridization between Te and Fe/Co states, we car-ried out measurements of the magnetic field depend-ence of the XMCD at 10 and 300 K. Figure 5 showselement-specific hysteresis loops for (a), (b) Fe, (c),(d) Co, and (e) Te in both NI and GI, with the meas-urement geometry illustrated in figure 5(f). Note thattheGe signal was too noisy to carry outmeasurementsof the hysteresis loop (and so was the Te measure-ment at 300 K). All three loops at 10 K show qual-itatively the same behavior, i.e. an OOP hard axis(measured at NI) and an IP easy axis (measured atGI) with a very small coercive field. Importantly, the62D Mater. 12 (2025) 025001 E Heppell et alFigure 5. Element-specific magnetization curves of (Co0.28Fe0.72)5GeTe2 recorded at the (a), (b) Fe L3 (703.8 eV off-edge /706.9 eV on-edge), (c), (d) Co L3 (774.2 eV off-edge / 777.75 eV on-edge), and (e) TeM5 (569.9 eV off-edge / 572.2 eV on-edge)edges in total electron yield. The measurements were carried out in both normal (NI, red) and grazing incidence (GI= 20◦ offthe horizon, blue) at (a), (c), (e) 10 K and (b), (d) 300 K. The curves have been normalized to their saturation values at high field.(f) Illustration of the normal and grazing incidence measurement geometries.field-dependence of the (normalized) Fe, Co, and TeXMCD at 10 K is qualitatively identical, indicatingtheir strong coupling. At 300 K, i.e. ∼30 K below thebulk TC, the anisotropy has markedly reduced withboth the NI and GI loops for Fe and Co (figures 5(b)and (d)) showing a similar easy axis behavior. Thisis consistent with the intralayer exchange couplingbeing stronger than the interlayer coupling. Further,the magnetization does not seem to saturate at themaximum field of 6 T, pointing towards materialsinhomogeneities and the existence of a paramagneticbackground.While the hysteresis curves in figure 5 clearlyshow for both Fe and Co at 10 and 300 K that theeasy axis is IP (GI geometry), the sum rule ana-lysis results in table 1 (measured at the maximumapplied field of 6 T) seems more ambiguous at firstglance. Using perturbation theory, Bruno proposed amodel for the magnetocrystalline anisotropy energy(MAE) under the assumption that the majority spinband is completely filled [56]. Bruno’s model statesthat the absolute value of the orbital moment is lar-ger along the easy magnetization direction, and thatthe difference between the orbital moments alongthe easy and hard directions is proportional to theMAE. Comparing the orbital moments for Fe andCo (table 1), measured in GI and NI, one notes thatfor Fe [∆(mGIorb −mNIorb)=−0.019µB < 0] the MAEprefers a magnetization direction that is OOP, dis-agreeing with the hysteresis behavior, while for Co[∆(mGIorb −mNIorb)= 0.039µB > 0] the MAE prefers amagnetization direction that is IP, in agreement withthe hysteresis.To understand the opposite trend, one has to bearin mind that the hysteresis curves are predominantlydetermined by the behavior of the spin moment.Therefore, while the (local) orbital moment reflectsthe MAE of the ion, the spin moments are affectedby the exchange interaction between them, makingin this case Fe follow the IP behavior of Co. This‘magnetic bully’ effect has been first demonstratedin ultrathin Co/Ni films, in which, using XMCD, itwas found that the stronger anisotropy contributionof a much thinner Co layer redirects the easy mag-netization direction of the entire film [57]. Whilefor 3d transition metals, the orbital moment morb isassumed to be proportional to theMAEof the specificatom, the situation is completely different for the spinmoment. Bruno [56] used second-order perturba-tion theory to derive the effect of the small orbitalmoment on the magnetocrystalline anisotropy, andthis cannot be applied to the spinmoment. Following[58], the spin moment is isotropic, so that mspin =[mOOPspin,eff + 2mIPspin,eff]/3, and the anisotropy in themeasured effective spin moment is due to the mag-netic dipole term ⟨Tz⟩. Using mOOPspin,eff =mNIspin,eff, andapproximating mIPspin,eff by mGIspin,eff, we obtain the fol-lowing (absolute) values for ⟨Tz⟩: −0.0076µB per Featom and −0.0067µB per Co atom, which is ∼0.5%72D Mater. 12 (2025) 025001 E Heppell et alof mspin for Co. The magnetic dipole term has noinfluence on the exchange interaction, but it gives asmall contribution to the magnetic anisotropy (seeequation (28) in [58]). This is because ⟨Tz⟩ accountsfor a spin quadrupolemoment [59] aswell as for spin-flip excitations between the exchange split majorityand minority spin bands. Furthermore, the Brunomodel is strictly only valid if the minority band iscompletely empty [58], which is of course not reallythe case for the Fe andCo atoms. Thus, there are smalladditional terms besides the orbital moment aniso-tropy, which can be neglected in our case.On the other hand, the hysteresis loops are dom-inated by the dichroism due to the spin moment,which is much larger than that of the orbital moment.As mentioned, the spin moment is isotropic andits absolute value is not expected to change as afunction of applied field. However, below a certainfield threshold, the spin moments, when measuredalong the hard axis, will tilt away towards the easyaxis. Then, the projection of the spin moments ontothe beam direction (which is taken along the field)becomes smaller, thereby reducing the dichroism sig-nal. This explains the behavior of the magnetizationloops in figure 5, andmakes it a more direct indicatorfor the easy direction than the orbital moment values.The observed IP magnetic anisotropy is in con-trast with other vdW ferromagnets, such as CrI3 andFe3GeTe2, which typically exhibit OOP (perpendicu-lar) magnetic anisotropy due to stronger spin–orbitcoupling effects that stabilize perpendicular magneticorientations. In Co-doped F5GT, however, Co sub-stitution shifts the anisotropy to an IP configuration,a characteristic that enhances the magnetic stabilitywithin the atomic layers. This IP alignment is signi-ficant, as it improves the stability of magnetic ordereven in ultrathin layers, making this material par-ticularly promising for spintronic applications thatbenefit from IP magnetization. This unique aniso-tropy in Co-doped F5GT demonstrates the potentialto tune vdWmaterials to fit specific application needsthrough chemical doping.3.5. DFT calculationsThe magnetic anisotropy in Fe5−δGeTe2 bulk crystalsdepends on Fe vacancies, which introduce hole dop-ing. In our DFT calculations for bulk Fe5GeTe2 (spacegroup R3̄m), we identified IP anisotropy with a valueof 0.41meV/f.u. and an average magnetic moment of1.68µB/Fe atom. Upon Co doping (x= 0.25) usingthe VCA, the anisotropy remained IP but reducedto 0.18meV/f.u. The Fe (average) and Co momentswere 1.75 and 1.97µB atom−1, respectively, thoughdiscrepancies with experimental XMCD values sug-gest that Fe vacancies and model simplicity mightbe influencing these results. Increased hole dopingenhanced the magnetic moment by approximately21% (see figure S1).In the primitive P3̄m1 unit cell, Fe5GeTe2 exhib-ited IP anisotropy (0.40 meV/f.u.), but Co dop-ing (x= 0.25) switched the easy axis to OOP with0.60meV/f.u., contradicting experimental findings.Potential explanations include disorder, Fe1 site vari-ations, or strain effects in 2D materials. We alsoexplored the√3×√3× 1 supercell of theR3m struc-ture, motivated by the supercell observed in the x-ray diffraction data [27], which showed that Co dop-ing reduces IP anisotropy and increases hybridizationbetween Co 3d and Te 5p states (see section S5 fordetails and the results summarized in table S1). Thisbehavior is consistent with the atom-projected DOSfindings presented in figure S3.4. ConclusionIn this study, we investigated the magnetic proper-ties of Co-doped Fe5GeTe2 using a combination ofXPEEM, XMCD measurements, and DFT calcula-tions. Our XPEEM measurements reveal a remark-ably high Curie temperature of ∼300 K for a bilayerof CoF5GT, which is significantly higher than thatof other materials in that class of 2D ferromagnetsin the ultrathin limit. Notably, the TC shows only asmall dependence on film thickness, consistent withthe observed IP magnetic anisotropy and strongerIP exchange coupling. The XMCD measurementsprovided insights into the spinmoments of Fe andCo,showing an increase and reduction, respectively, com-pared to theoretical predictions from DFT calcula-tions. The XMCDmeasurements also highlighted theorbital moment anisotropy of the elemental constitu-ents, demonstrating that the introduction ofCo redir-ects the easy magnetization direction due to its pref-erence for IP anisotropy. This finding is critical forthe development of stable, high-temperature 2D fer-romagnets and underscores the potential of chemicaldoping to tune the magnetic properties of 2D vdWferromagnets.Data availability statementAll data that support the findings of this study areincluded within the article (and any supplementaryfiles).AcknowledgmentsThe XMCD experiments were performed at theBOREAS beamline, and the XPEEM experiments atthe CIRCE beamline, both at the ALBA Synchrotronunder proposals 2023027573 and 2022097164,respectively. E H acknowledges a STFC-Diamond-EPSRC studentship (2604894, EP/R513295/1,EP/T517811/1). K W and T T acknowledge supportfrom the JSPS KAKENHI (Grant Numbers 21H05233and 23H02052) and World Premier InternationalResearch Center Initiative (WPI), MEXT, Japan. R82D Mater. 12 (2025) 025001 E Heppell et alF and T H acknowledge financial support from theOxford-ShanghaiTech collaboration project. R F, TH, and G L were supported by the Engineering andPhysical Sciences Research Council under GrantsEP/X015556/1 and EP/X015599/1, respectively. Wethank Diamond Light Source for access to the facilit-ies of theMaterials Characterisation Laboratory. Bulkcrystal synthesis and characterization (A.F.M.) weresupported by the US Department of Energy, Officeof Science, Basic Energy Sciences, Materials Sciencesand Engineering Division.Conflict of interestThe authors declare no competing financial interest.ORCID iDsEmily Heppell https://orcid.org/0000-0002-3977-4623Ryuji Fujita https://orcid.org/0000-0001-8557-5812Gautam Gurung https://orcid.org/0000-0001-7706-6276Jheng-Cyuan Lin https://orcid.org/0000-0002-5085-3234Andrew F May https://orcid.org/0000-0003-0777-8539Michael Foerster https://orcid.org/0000-0002-4147-6668MWaqas Khaliq https://orcid.org/0000-0002-9696-3498Miguel Angel Niño https://orcid.org/0000-0003-3692-147XManuel Valvidares https://orcid.org/0000-0003-4895-8114Javier Herrero-Martín https://orcid.org/0000-0003-1986-8128Pierluigi Gargiani https://orcid.org/0000-0002-6649-0538Kenji Watanabe https://orcid.org/0000-0003-3701-8119Takashi Taniguchi https://orcid.org/0000-0002-1467-3105Dirk Backes https://orcid.org/0000-0002-1019-3323Gerrit van der Laan https://orcid.org/0000-0001-6852-2495Thorsten Hesjedal https://orcid.org/0000-0001-7947-3692References[1] Novoselov K S, Geim A K, Morozov S V, Jiang D, Zhang Y,Dubonos S V, Grigorieva I V and Firsov A A 2004 Science306 666–9[2] Qian X, Liu J, Fu L and Li J 2014 Science 346 1344–7[3] Yu Y et al 2015 Nat. Nanotechnol. 10 270–6[4] Tsen AW, Hunt B, Kim Y D, Yuan Z J, Jia S, Cava R J, Hone J,Kim P, Dean C R and Pasupathy A N 2016 Nat. Phys.12 208–12[5] Park J G 2016 J. Phys.: Condens. Matter 28 301001[6] Burch K S, Mandrus D and Park J G 2018 Nature 563 47–52[7] Kim H H, Yang B, Patel T, Sfigakis F, Li C, Tian S, Lei H andTsen A W 2018 Nano Lett. 18 4885–90[8] Wang X et al 2019 Sci. Adv. 5 eaaw8904[9] Huang B et al 2017 Nature 546 270–3[10] Gong C et al 2017 Nature 546 265–9[11] Geim A K and Grigorieva I V 2013 Nature 499 419–25[12] Zhong D et al 2017 Sci. Adv. 3 e1603113[13] Fujita R et al 2024 Adv. Funct. Mater. 34 2400552[14] Chen B, Yang J, Wang H D, Imai M, Ohta H, Michioka C,Yoshimura K and Fang M H 2013 J. Phys. Soc. Jpn. 82 124711[15] Fei Z et al 2018 Nat. Mater. 17 778–82[16] Lee J U, Lee S, Ryoo J H, Kang S, Kim T Y, Kim P, Park C H,Park J G and Cheong H 2016 Nano Lett. 16 7433–8[17] Zha H, Li W, Zhang G, Liu W, Deng L, Jiang Q, Ye M, Wu H,Chang H and Qiao S 2023 Chin. Phys. Lett. 40 087501[18] May A F, Ovchinnikov D, Zheng Q, Hermann R, Calder S,Huang B, Fei Z, Liu Y, Xu X and McGuire M A 2019 ACSNano 13 4436–42[19] Stahl J, Shlaen E and Johrendt D 2018 Z. Anorg. Allg. Chem.644 1923–9[20] May A F, Bridges C A and McGuire M A 2019 Phys. Rev.Mater. 3 104401[21] Yun S J, Duong D L, Ha D M, Singh K, Phan T L, Choi W,Kim Y M and Lee Y H 2020 Adv. Sci. 7 1903076[22] Wang Z, Sapkota D, Taniguchi T, Watanabe K, Mandrus Dand Morpurgo A F 2018 Nano Lett. 18 4303–8[23] Gao Y, Yin Q, Wang Q, Li Z, Cai J, Zhao T, Lei H, Wang S,Zhang Y and Shen B 2020 Adv. Mater. 32 2005228[24] Park T E et al 2021 Phys. Rev. B 103 104410[25] Gao Y et al 2022 Phys. Rev. B 105 014426[26] McGuire M A 2017 Crystals 7 121[27] May A F, Du M H, Cooper V R and McGuire M A 2020 Phys.Rev. Mater. 4 074008[28] Tian C, Pan F, Xu S, Ai K, Xia T and Cheng P 2020 Appl.Phys. Lett. 116 202402[29] Chen X, Schierle E, He Y, Vranas M, Freeland J W,McChesney J L, Ramesh R, Birgeneau R J and Frano A 2022Phys. Rev. Mater. 6 094404[30] Lu L et al 2024 Nano Lett. 24 5984–92[31] Chen X, Tian W, He Y, Zhang H, Werner T L, Lapidus S,Ruff J P C, Ramesh R and Birgeneau R J 2023 Phys. Rev.Mater. 7 044411[32] May A F, Yan J and McGuire M A 2020 J. Appl. Phys.128 051101[33] Aballe L, Foerster M, Pellegrin E, Nicolas J and Ferrer S 2015J. Synchrotron Radiat. 22 745–52[34] Foerster M, Prat J, Massana V, Gonzalez N, Fontsere A,Molas B, Matilla O, Pellegrin E and Aballe L 2016Ultramicroscopy 171 63–69[35] Kronast F and Molina S V 2016 J. Large-Scale Res. Facilit2 A90[36] Sander A et al 2021 Sci. Rep. 11 20788[37] Barla A, Nicolás J, Cocco D, Valvidares S M,Herrero-Martín J, Gargiani P, Moldes J, Ruget C, Pellegrin Eand Ferrer S 2016 J. Synchrotron Rad. 23 1507–17[38] van der Laan G and Figueroa A I 2014 Coord. Chem. Rev.277–278 95–129[39] van der Laan G 1998 Phys. Rev. B 57 112–5[40] van der Laan G 1999 J. Synchrotron Rad. 6 694–5[41] Piamonteze C, Miedema P and de Groot F M F 2009 Phys.Rev. B 80 184410[42] Kresse G and Joubert D 1999 Phys. Rev. B 59 1758–75[43] Perdew J P, Burke K and Ernzerhof M 1996 Phys. Rev. Lett.77 3865–8[44] Blöchl P E 1994 Phys. Rev. B 50 17953–79[45] Grimme S, Ehrlich S and Goerigk L 2011 J. Comput. Chem.32 1456–65[46] Momma K and Izumi F 2008 J. Appl. Crystal. 41 653–89https://orcid.org/0000-0002-3977-4623https://orcid.org/0000-0002-3977-4623https://orcid.org/0000-0002-3977-4623https://orcid.org/0000-0001-8557-5812https://orcid.org/0000-0001-8557-5812https://orcid.org/0000-0001-8557-5812https://orcid.org/0000-0001-7706-6276https://orcid.org/0000-0001-7706-6276https://orcid.org/0000-0001-7706-6276https://orcid.org/0000-0002-5085-3234https://orcid.org/0000-0002-5085-3234https://orcid.org/0000-0002-5085-3234https://orcid.org/0000-0003-0777-8539https://orcid.org/0000-0003-0777-8539https://orcid.org/0000-0003-0777-8539https://orcid.org/0000-0002-4147-6668https://orcid.org/0000-0002-4147-6668https://orcid.org/0000-0002-4147-6668https://orcid.org/0000-0002-9696-3498https://orcid.org/0000-0002-9696-3498https://orcid.org/0000-0002-9696-3498https://orcid.org/0000-0003-3692-147Xhttps://orcid.org/0000-0003-3692-147Xhttps://orcid.org/0000-0003-3692-147Xhttps://orcid.org/0000-0003-4895-8114https://orcid.org/0000-0003-4895-8114https://orcid.org/0000-0003-4895-8114https://orcid.org/0000-0003-1986-8128https://orcid.org/0000-0003-1986-8128https://orcid.org/0000-0003-1986-8128https://orcid.org/0000-0002-6649-0538https://orcid.org/0000-0002-6649-0538https://orcid.org/0000-0002-6649-0538https://orcid.org/0000-0003-3701-8119https://orcid.org/0000-0003-3701-8119https://orcid.org/0000-0003-3701-8119https://orcid.org/0000-0002-1467-3105https://orcid.org/0000-0002-1467-3105https://orcid.org/0000-0002-1467-3105https://orcid.org/0000-0002-1019-3323https://orcid.org/0000-0002-1019-3323https://orcid.org/0000-0002-1019-3323https://orcid.org/0000-0001-6852-2495https://orcid.org/0000-0001-6852-2495https://orcid.org/0000-0001-6852-2495https://orcid.org/0000-0001-7947-3692https://orcid.org/0000-0001-7947-3692https://orcid.org/0000-0001-7947-3692https://doi.org/10.1126/science.1102896https://doi.org/10.1126/science.1102896https://doi.org/10.1126/science.1256815https://doi.org/10.1126/science.1256815https://doi.org/10.1038/nnano.2014.323https://doi.org/10.1038/nnano.2014.323https://doi.org/10.1038/nphys3579https://doi.org/10.1038/nphys3579https://doi.org/10.1088/0953-8984/28/30/301001https://doi.org/10.1088/0953-8984/28/30/301001https://doi.org/10.1038/s41586-018-0631-zhttps://doi.org/10.1038/s41586-018-0631-zhttps://doi.org/10.1021/acs.nanolett.8b01552https://doi.org/10.1021/acs.nanolett.8b01552https://doi.org/10.1126/sciadv.aaw8904https://doi.org/10.1126/sciadv.aaw8904https://doi.org/10.1038/nature22391https://doi.org/10.1038/nature22391https://doi.org/10.1038/nature22060https://doi.org/10.1038/nature22060https://doi.org/10.1038/nature12385https://doi.org/10.1038/nature12385https://doi.org/10.1126/sciadv.1603113https://doi.org/10.1126/sciadv.1603113https://doi.org/10.1002/adfm.202400552https://doi.org/10.1002/adfm.202400552https://doi.org/10.7566/JPSJ.82.124711https://doi.org/10.7566/JPSJ.82.124711https://doi.org/10.1038/s41563-018-0149-7https://doi.org/10.1038/s41563-018-0149-7https://doi.org/10.1021/acs.nanolett.6b03052https://doi.org/10.1021/acs.nanolett.6b03052https://doi.org/10.1088/0256-307X/40/8/087501https://doi.org/10.1088/0256-307X/40/8/087501https://doi.org/10.1021/acsnano.8b09660https://doi.org/10.1021/acsnano.8b09660https://doi.org/10.1002/zaac.201800456https://doi.org/10.1002/zaac.201800456https://doi.org/10.1103/PhysRevMaterials.3.104401https://doi.org/10.1103/PhysRevMaterials.3.104401https://doi.org/10.1002/advs.201903076https://doi.org/10.1002/advs.201903076https://doi.org/10.1021/acs.nanolett.8b01278https://doi.org/10.1021/acs.nanolett.8b01278https://doi.org/10.1002/adma.202005228https://doi.org/10.1002/adma.202005228https://doi.org/10.1103/PhysRevB.103.104410https://doi.org/10.1103/PhysRevB.103.104410https://doi.org/10.1103/PhysRevB.105.014426https://doi.org/10.1103/PhysRevB.105.014426https://doi.org/10.3390/cryst7050121https://doi.org/10.3390/cryst7050121https://doi.org/10.1103/PhysRevMaterials.4.074008https://doi.org/10.1103/PhysRevMaterials.4.074008https://doi.org/10.1063/5.0006337https://doi.org/10.1063/5.0006337https://doi.org/10.1103/PhysRevMaterials.6.094404https://doi.org/10.1103/PhysRevMaterials.6.094404https://doi.org/10.1021/acs.nanolett.4c00472https://doi.org/10.1021/acs.nanolett.4c00472https://doi.org/10.1103/PhysRevMaterials.7.044411https://doi.org/10.1103/PhysRevMaterials.7.044411https://doi.org/10.1063/5.0015971https://doi.org/10.1063/5.0015971https://doi.org/10.1107/S1600577515003537https://doi.org/10.1107/S1600577515003537https://doi.org/10.1016/j.ultramic.2016.08.016https://doi.org/10.1016/j.ultramic.2016.08.016https://doi.org/10.17815/jlsrf-2-86https://doi.org/10.17815/jlsrf-2-86https://doi.org/10.1038/s41598-021-99963-whttps://doi.org/10.1038/s41598-021-99963-whttps://doi.org/10.1107/S1600577516013461https://doi.org/10.1107/S1600577516013461https://doi.org/10.1016/j.ccr.2014.03.018https://doi.org/10.1016/j.ccr.2014.03.018https://doi.org/10.1103/PhysRevB.57.112https://doi.org/10.1103/PhysRevB.57.112https://doi.org/10.1107/S0909049599000412https://doi.org/10.1107/S0909049599000412https://doi.org/10.1103/PhysRevB.80.184410https://doi.org/10.1103/PhysRevB.80.184410https://doi.org/10.1103/PhysRevB.59.1758https://doi.org/10.1103/PhysRevB.59.1758https://doi.org/10.1103/PhysRevLett.77.3865https://doi.org/10.1103/PhysRevLett.77.3865https://doi.org/10.1103/PhysRevB.50.17953https://doi.org/10.1103/PhysRevB.50.17953https://doi.org/10.1002/jcc.21759https://doi.org/10.1002/jcc.21759https://doi.org/10.1107/S0021889808012016https://doi.org/10.1107/S00218898080120162D Mater. 12 (2025) 025001 E Heppell et al[47] Chen H et al 2023 ACS Appl. Mater. Interfaces15 3287–96[48] Thole B T, Carra P, Sette F and van der Laan G 1992 Phys.Rev. Lett. 68 1943–6[49] Carra P, Thole B T, Altarelli M and Wang X 1993 Phys. Rev.Lett. 70 694–7[50] Chen C T, Idzerda Y U, Lin H J, Smith N V, Meigs G,Chaban E, Ho G H, Pellegrin E and Sette F 1995 Phys. Rev.Lett. 75 152–5[51] Yamagami K et al 2021 Phys. Rev. B 103 L060403[52] Solovyev I V, Dederichs P H and Mertig I 1995 Phys. Rev. B52 13419–28[53] Figueroa A I, Bonell F, Cuxart M G, Valvidares M,Gargiani P, van der Laan G, Mugarza A and Valenzuela S O2020 Phys. Rev. Lett. 125 226801[54] Ohta T, Sakai K, Taniguchi H, Driesen B, Okada Y,Kobayashi K and Niimi Y 2020 Appl. Phys. Express 13 043005[55] Zhang H et al 2020 Phys. Rev. B 102 064417[56] Bruno P 1989 Phys. Rev. B 39 865–8[57] Dürr H A, Guo G Y, van der Laan G, Lee J, Lauhoff G andBland J A C 1997 Science 277 213–5[58] van der Laan G 1998 J. Phys.: Condens. Matter 10 3239[59] Miura Y and Okabayashi J 2022 J. Phys.: Cond. Matter34 47300110https://doi.org/10.1021/acsami.2c18028https://doi.org/10.1021/acsami.2c18028https://doi.org/10.1103/PhysRevLett.68.1943https://doi.org/10.1103/PhysRevLett.68.1943https://doi.org/10.1103/PhysRevLett.70.694https://doi.org/10.1103/PhysRevLett.70.694https://doi.org/10.1103/PhysRevLett.75.152https://doi.org/10.1103/PhysRevLett.75.152https://doi.org/10.1103/PhysRevB.103.L060403https://doi.org/10.1103/PhysRevB.103.L060403https://doi.org/10.1103/PhysRevB.52.13419https://doi.org/10.1103/PhysRevB.52.13419https://doi.org/10.1103/PhysRevLett.125.226801https://doi.org/10.1103/PhysRevLett.125.226801https://doi.org/10.35848/1882-0786/ab7f18https://doi.org/10.35848/1882-0786/ab7f18https://doi.org/10.1103/PhysRevB.102.064417https://doi.org/10.1103/PhysRevB.102.064417https://doi.org/10.1103/PhysRevB.39.865https://doi.org/10.1103/PhysRevB.39.865https://doi.org/10.1126/science.277.5323.213https://doi.org/10.1126/science.277.5323.213https://doi.org/10.1088/0953-8984/10/14/012https://doi.org/10.1088/0953-8984/10/14/012https://doi.org/10.1088/1361-648X/ac943fhttps://doi.org/10.1088/1361-648X/ac943f Room-temperature in-plane ferromagnetism in Co-substituted Fe5GeTe2 investigated by magnetic x-ray spectroscopy and microscopy 1. Introduction 2. Materials and methods 2.1. Co-doped Fe5GeTe2 sample preparation 2.2. Magnetic domain imaging with x-ray photoemission electron microscopy (XPEEM) 2.3. Magnetic x-ray spectroscopy 2.4. Density functional theory (DFT) calculations 3. Results and discussion 3.1. XPEEM study of bilayer CoF5GT 3.2. Orbital and spin magnetic moments of bulk CoF5GT 3.3. Magnetic spectroscopy at the Ge L2,3 and Te M4,5 edges 3.4. Magnetic anisotropy 3.5. DFT calculations 4. Conclusion References