Structural and magnetic properties of high magnetization FexCo100-x nanoparticles investigated at the nanoscale: Unveiling the origin of the observed anisotropy

Structural and magnetic properties of high magnetization Fe x Co 100-x nanoparticles investigated at the nanoscale: Unveiling the origin of the observed anisotropy Jon Gutierrez a,b,* , Virginia Vadillo b,1 , Ines Puente c,d , Claudia Mondelli e , Marie Capron f , Javier Alonso g , I˜ naki Orue h , Patricia L´ azpita a,b , Joseba S. Garitaonandia a,b , Izaskun Gil de Muro a,b , Tommaso Baroni i , Francesco D’Acapito j , Maite Insausti a,b a Faculty of Science and Technology, Universidad del Pais Vasco/Euskal Herriko Unibertsitatea, Barrio Sarriena s/n, Leioa 48940, Spain b BCMaterials (Basque Center for Materials, Applications &Nanostructures), UPV/EHU Scientific Park, Bldg. Martina Casiano, 3rd. Floor, Barrio Sarriena s/n, Leioa 48940, Spain c Institut Laue Langevin, 71 Avenue des Martyrs, CS 20156, Grenoble Cedex 38042, France d Instituto de Nanociencia y Materiales de Arag´ on, CSIC-Universidad de Zaragoza, Pedro Cerbuna 12, Zaragoza 50009, Spain e CNR-IOM, Institut Laue Langevin, 71, Avenue des Martyrs, Grenoble Cedex 38042, France f European Synchrotron Radiation Facilities (ESRF), 71 Avenue des Martyrs, CEDEX 9, Grenoble 38042, France g CITIMAC Department, Universidad de Cantabria, Santander 39005, Spain h SGIKER, Universidad del Pais Vasco/Euskal Herriko Unibertsitatea, Barrio Sarriena s/n, Leioa 48940, Spain i Department of Earth Sciences, University of Florence, via La Pira 4, Firenze 50121, Italy j CNR-IOM-OGG c/o ESRF-The European Synchrotron, 71 Avenue des Martyrs, CS 40220, Grenoble Cedex 38043, France ARTICLE INFO Keywords: Iron Cobalt alloy Magnetic nanoparticles Magnetic anisotropy Neutron diffraction EXAFS and XANES spectroscopies M¨ ossbauer spectroscopy Shape anisotropy ABSTRACT In this work the authors have performed the synthesis of Fe x Co 100-x (0<x<100) alloy nanoparticles (NPs) with different compositions, as well as pure Fe and Co NPs for comparison, by a chemical reduction technique. The subsequent characterization demonstrated excellent quality NPs with the expected bcc cubic (for Fe and FeCo alloys) and hcp hexagonal (for Co NPs) structures showing a room temperature magnetization as high as 235 emu/g for the Fe 66 Co 34 composition alloy. Nevertheless, this soft magnetic character is accompanied by determined values of effective anisotropy as high as 2 MJ/m 3 . Aiming to deep into the properties of these FeCo alloys as well as to unveil the origin of that observed high anisotropy value, we now present an extensive study at the nanoscale of the synthesized Fe, Co and Fe x Co 100-x alloy nanoparticles by using nuclear techniques as neutron powder diffraction, EXAFS and XANES spectroscopies. M¨ ossbauer spectroscopy revealed that the FeCo alloys are in an A2 disordered solid solution. The obtained results, combined with AFM/MFM images, have demonstrated that despite the cubic bcc structure observed for all FeCo alloys (in excellent concordance with the pure Fe one) the NPs show a "flaky" shape of 50–60 nm size (diameter) but only 3–4 nm thickness, giving rise to a strong shape anisotropy contribution to the observed total effective anisotropy. 1. Introduction Magnetic nanoparticles (MNPs) are nowadays a key class of materials that has contributed very quickly to the improvement and development of applications in the nanotechnology field [1–3]. They can be obtained in the size range from a few to 100 nm, and due to their unique properties they are being widely used in fields like biomedicine (superparamagnetic Fe 3 O 4 and Fe 2 O 3 particles due to their biocompatibility can be applied in magnetic targeting or magnetic resonance imaging, MRI [4]), heat release (like AC excitation magnetic field of a magnetic fluid for hyperthermia [5]) or in catalysis applications (since Fe, Co, Ni and their alloy component metals as Pt and Pd are catalytically active [6]). Apart from those previously mentioned compositions, the FeCo alloy * Corresponding author at: Faculty of Science and Technology, Universidad del Pais Vasco/Euskal Herriko Unibertsitatea, Barrio Sarriena s/n, Leioa 48940, Spain. E-mail address: [email protected] (J. Gutierrez). 1 on leave from BCMaterials. Contents lists available at ScienceDirect Journal of Alloys and Compounds journal homepage: www.elsevier.com/locate/jalcom https://doi.org/10.1016/j.jallcom.2024.177211 Received 19 July 2024; Received in revised form 21 October 2024; Accepted 22 October 2024 Journal of Alloys and Compounds 1010 (2025) 177211 Available online 28 October 2024 0925-8388/© 2024 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license ( http://creativecommons.org/licenses/bync-nd/4.0/ ). belongs to the group of soft ferromagnetic materials with the highest known magnetization (about 235 emu/g for a 35 %wt of Co content), large permeability, low coercivity, and high Curie temperature (about 1000 K) [7]. The actual trend is also to fabricate this material at the "nano" scale [8–11], where the outstanding magnetic properties of this FeCo alloy make possible to extend applications of MNPs to magnetic recording media [12], exchange-coupled nanocomposite magnets [13] or also smart fluids [14]. In a previous work [15] we have presented the synthesis (by a chemical reduction method) and characterization at the macroscopic scale of three Fe x Co 100-x (x =66, 51 and 33) alloy nanoparticles (NPs). Briefly, all of them show bcc cubic structure like the parent Fe powder also synthesized following the same procedure; they show room temperature magnetization values (at 1,5 T applied magnetic field) over 200 Am 2 /kg and coercive field values in the range 5–6.5 kA/m. But despite this soft magnetic character exhibited by these FeCo compositions, they also show unexpected magnetic anisotropy values as high as 2 MJ/m 3 . This value turns out to be in clear contrast with the magnetic anisotropy for Fe 50 Co 50 bulk alloy, about 15–20 kJ/m 3 [16]. It is already well stablished that magnetic properties shown by nanostructured magnetic materials are different if compared with the same parent composition at the bulk state. In this last, magnetism arises from electronic spin-spin and spin-orbit as well as exchange interactions of free electrons. Reducing material dimensions to the nanoscale leads to a confinement of electronic interactions, and as a first consequence magnetic performance (reflected in magnetic saturation, permeability and remanence) becomes size-dependent. When using MNPs for any particular application, the role of magnetic anisotropy arises as a capital question [17]. Many properties of magnetic nanoparticles like the initial magnetic susceptibility or temperature-dependent magnetic relaxation, when used for applications, depend to a great extent on the magnetic anisotropy [18]. Aiming to deep into the properties of these FeCo alloys as well as to unveil the origin of the observed high anisotropy value, in the following we will report a deep study performed at the nanoscale, by using microscopic techniques as neutron powder diffraction (NPD), transmission electron (TEM) and atomic/magnetic force (AFM/MFM) microscopies, X-ray Absorption (XAS) and M¨ ossbauer spectroscopies. XAS (for our purposes Extended X-ray Absorption Fine Structure, XANES, and X-ray Absorption Near Edge Structure, EXAFS, spectroscopies) is a local-structure sensitive and element specific technique [19] based on the measurement of the variation of the absorption coefficient as a function of the applied X-ray energy. It can provide important information about the phase distribution inside a sample, the relative oxidation state, so as quantitative data on the nature of the ligands and the interatomic distances between a central atom and its neighbors. Due to its properties, XAS perfectly complements the long-range information coming from XRD analysis [19]. The possibility to study separately Fe and Co atoms environment will allow us to determine the correct formation of FeCo alloy nanoparticles, providing evidences on the structural homogeneity of the samples or the segregation of Feor Co-rich regions. The combination of such powerful techniques will give a clear insight mainly in the structure and morphology of the synthesized Fe, Co and FeCo nanoparticles, with emphasis on the assessment of their composition and atomic order/disorder characteristics. 2. Experimental 2.1. Preparation of the Fe x Co 100-x nanoparticles Different composition Fe x Co 100-x nanoparticles were synthesized by employing a chemical reduction route: pure Fe (x=100), Fe x Co 100-x : Fe excess (x/(100-x) >1), FeCo: equiatomic (x/(100-x) =1), Fe x Co 100-x : Co excess (x/(100-x) <1) and pure Co (x =0). Three of the synthesized compositions belong to the FeCo binary alloy, and in the following these NPs will be labelled as Fe, Fe exc Co, FeCo, FeCo exc and Co, respectively. Briefly, in the chemical reduction method employed, aluminium powder was used to reduce Fe (III) and Co (II) precursors in the presence of NH 4 F. By using the same chemical procedure, pure Fe and Co nanoparticles were also synthesized. Extensive information about used products, quantities and the preparation method for all the synthesized compositions can be found in the Supplementary Material (S1 in SM). 2.2. TEM and AFM/MFM microscopy characterization Morphology of the raw powder of all synthesized compositions was first analyzed by Transmission Electron Microscopy (TEM) technique. TEM images were obtained using a Philips CM200 microscope at an acceleration voltage of 200 kV. Observation samples preparation was performed by using Fe, Co and FeCo alloy nanoparticles with different compositions dispersed in hexane and drop-casted onto a copper grid. To estimate the nanoparticles (or their agglomerates) size, ImageJ software was used [20]. A deeper insight to our nanoparticles morphology was achieved from AFM images obtained by using a fast scanning an Asylum Research Cypher-S instrument provided by the ESRF through the Partnership for Soft Condensed Matter (PSCM). The imaging was performed at room temperature (295 K) in Magnetic Force Microscopy (MFM) mode. Magnetic force microscope (MFM) is a powerful non-contact tool for characterizing sub-micrometer magnetic domains [21]. This technique is complementary to TEM observations for the morphological analysis because it is sensitive to the thickness of the nanoparticles, an important characteristic as it will be further discussed. The AFM was equipped with Asylum Research’s ASYMFMHC-R2 probes with a spring constant of 3.3 Nm and a typical resonance frequency of 74.5 kHz, determined via the thermal noise method. The nominal tip radius is 32 nm, while its magnetic tip coating is CoCr. The scan rate was 1.5 Hz. Samples for this technique were prepared by depositing on a clean mica surface one drop of nanoparticles in ethanol suspension, with subsequent drying by spin coating process running at 30,000 rpm for 60 seconds. 2.3. Neutron powder diffraction Neutron Powder Diffraction (NPD) patterns of our NPs were measured at the diffractometer D1B, ILL (Institut Laue-Langevin, France) using an Orange Cryostat. Wavelength was refined to λ= 1.2863 Å. Diffraction data were continuously collected from 10 to 300 K while temperature changed at a rate of 0.5 K/min. The measurements were performed between 20◦and 140◦in 2θwith steps of 0.1◦. The measured peak shape was modeled with a modified Thompson-CoxHasting pseudo-Voigt function with axial divergence asymmetry. The crystal structure was refined using the Rietveld method [22] implemented in the FULLPROF program [23]. Measured data set can be found in [24]. 2.4. X-ray absorption data collection (XANES and EXAFS) Complementary to Neutron Powder Diffraction (NPD) technique that reveals information on the average structure of the alloy, the X-ray absorption spectroscopy (XAS) offers the possibility to study separately Fe and Co atoms environment and allows to prove the correct formation of FeCo alloy nanoparticles as well as the presence of non-desired phases, like Feor Cooxides. Thus, a detailed study on our synthesized Fe, Co and FeCo alloy nanoparticles was performed by means of XAS measurements made in two different regions: the near edge structure (XANES) and the extended fine structure (EXAFS). Room temperature measurements were performed at the BM-08 "LISA" CRG Beamline at the European Synchrotron Radiation Facility (ESRF) in Grenoble, France [25]. Data were collected at the Fe (7112 eV) and Co (7709 eV) K edges using Si (111) crystals in the monochromator. The samples were measured at 10 K through the use of a cold-finger liquid-He cryostat. J. Gutierrez et al. Journal of Alloys and Compounds 1010 (2025) 177211 2 Higher harmonics rejection is obtained through Si coated collimating/focusing mirrors (with Ecutoff ≈15keV). Spectra were acquired in transmission mode with a fixed 5 eV step in the pre-edge region, 0.5 eV step around the edge and a kstep of 0.05 Å −1 up to a maximum value of k max =17 Å −1 for the Co spectra and k max =12.5 Å −1 for the Fe ones due to the co-presence of Co. Up to 4 consecutive scans per sample were acquired in order to improve the signal-to-noise ratio. The samples were prepared by mixing the Fe, Co or FeCo alloy powders with 50 mg of cellulose, in order to keep the total absorption µ<1.5, and afterwards they were pressed in pellets. Together with our study samples, Fe 3 O 4 (magnetite) CoFe 2 O 4 and Co 3 O 4 oxides were also analysed in order to verify the presence of oxides in the FeCo alloys structure. Finally, spectra of Fe and Co foils were also measured simultaneously as references and for data analysis purposes in order to monitor possible energy shifts during consecutive data acquisitions. 2.5. SQUID magnetometry and M¨ ossbauer spectroscopy Magnetic properties were measured by using a superconducting quantum interference device magnetometer (SQUID, Quantum Design MPMS-5) both at room (300 K) and low (5 K) temperatures, by applying a magnetic field up to 5 Tesla (about 4 MA/m). On the other hand, our FeCo alloys order at the atomic scale was analysed by using M¨ ossbauer spectroscopy. Room temperature spectra were measured in transmission geometry by using a 57 Co–Rh source and conventional constantacceleration spectrometer. Isomer shift values were calculated with respect to an α -Fe calibration foil. NORMOS program develop by Brand [26] was used for fitting the spectra. 3. Results and discussion In all the synthesized alloy samples, the Fe/Co ratio estimated from different experimental techniques is slightly lower than the initially expected one from the nominal ratio of 1.25:0.5, 1:0.75 and 0.7:1.05 of Fe:Co precursors respectively used for the synthesis (see Supplementary Material). Besides, a small amount of aluminium appears as impurity (in the range 0.3–0.5 %at). This is a direct consequence of an excess of aluminium added to the reaction in order to increase the yield of the chemical reduction process. Nanoparticles compositions were firstly determined by Inductively Coupled Plasma-Mass Spectroscopy, ICP, giving as result Fe x Co 100-x alloys with x =66, 51 and 33 compositions [15] (see Table 1). TEM images (see Fig. 1) reveal that all the obtained Fe, Co and FeCo alloy powders are composed of single nanoparticles ranging in a size from 20 to 80 nm and irregular shape, but that easily agglomerate forming in many cases dendritic-like structures as big as 600 nm. These NPs clusters appear due to the high dipolar interactions among the magnetic nanoparticles. Crystallite sizes calculated from neutron powder diffraction data confirm these estimations (Table S2). AFM/MFM images were obtained for nanoparticles set on a flat, nonmagnetic mica surface. In Fig. 2a negative phase shift is observed because of the attractive magnetic force of the single FeCo nanoparticle. In the corresponding topography image of the same particle (Fig. 2b), the individual magnetic particle can be observed, showing a good correlation between the MFM and AFM response. From the line profile across the white line marked of the single nanoparticle, the domain size has been estimated in 24 nm, which is slightly smaller than that obtained from XRD data (see Table S1). Note that the side shoulders on the figure are an experimental artifact due to a thin layer of water coming from the humidity present in the air. The 2–3 nm thickness observed in the AFM/MFM images reveals the “flaky”shape of the nanoparticle. All the single particles that have been found, have been checked and this planar shape has been corroborated. Due to the strong magnetic moment of each single NP, as it will be showed in the next sections, nanoparticles easily agglomerate forming clusters as big as 500 nm, but with thicknesses not over 12 nm, unveiling for agglomerates of 3–4 layers of single NPs (Fig. 3). The profile along the white line of the topographical image of Fig. 3b shows an agglomerate of magnetic domains ranging from 20 nm to 60 nm size, in agreement with TEM observations. Fig. 4 shows the obtained room temperature neutron diffraction patterns of all the studied samples. For pure Co, the synthesized nanoparticles show a hexagonal structure (P6 3 /mmc) with cell parameters a =2.5098 ±0.0005 Å and c=4.08530 ±0.0003 Å. The pure Fe powder diffraction pattern shows the expected peaks ((110) at 37.06◦, (200) at 53.54◦, (211) at 67.00◦, (220) at 79.23◦, (310) at 90.98◦, (222) at 102.70◦and (321) at 114.45◦) of the α -bcc structure with Im-3m space group and cell parameter a=2.8658 ±0.0002 Å (see Table 1). Like Fe NPs, the structures of the FeCo alloy nanoparticles match also well with the α -bcc structure (S.G. Im-3m), despite the slightly displacement of the diffraction peaks to higher 2θvalues. That is, the cell parameter decreases as the amount of iron decreases, from a=2.8617 (3) Å for the Fe exc Co alloy to a=2.8422(4) Å for the FeCo exc one (see Table 1). On the other hand, the values determined for our cell a parameter are in good agreement with cell parameter values previously obtained by other authors in FeCo alloy nanoparticles synthesized by other chemical routes [27,28] but also to other kind of morphologies as FeCo films where Fe/Co ratio progressively varies [29] or those reported for bulk species [30]. In the case of bulk alloys, despite the tendency is maintained, lattice parameters values are slightly higher than those from nanocrystalline samples [28]. Measured neutron diffraction spectra at 10 K are similar to those observed at room temperature. Fig. 5 shows the temperature dependence of the lattice parameters between 10 and 300 K. Cell parameters obtained at 10 K appear in Supplementary Material, together with calculated crystallite sizes (Table S2). For the Fe and FeCo alloys, there is no evident change up to temperatures around 100 K, followed by a smooth, monotonic linear increase of the lattice parameter with temperature up to 300 K. The lattice parameters of Co nanoparticles show a similar behaviour but with a "plateau" in the values measured up to 120 K. This trend agrees with the one expected for a material without a structural phase transition, as it is our case with all the studied samples. From obtained data, we can estimate the linear thermal expansion coefficient α at high temperatures (from T >100 K for Fe and FeCo alloys Table 1 Samples compositions (expected and estimated from ICP analysis), room temperature lattice parameter values, slope of their change with temperature and calculated linear thermal expansion coefficient obtained by NPD measurements, for Fe, Fe x Co 100-x and Co synthesized nanoparticles. Sample label Sample nominal composition Synthesized powder composition* R.T. lattice parameter a (Å) slope (x10 −5 Å/K) at high temperature α (x10 −6 K −1 ) at high temperature Fe Fe Fe 2.8658 ±0.0002 2.489 8.84 Fe exc Co Fe 71 Co 29 Fe 66 Co 34 2.8617 ±0.0003 2.453 8.85 FeCo Fe 57 Co 43 Fe 51 Co 49 2.8527 ±0.0004 2.608 8.18 FeCo exc Fe 40 Co 60 Fe 33 Co 67 2.8422 ±0.0004 2.559 7.98 Co Co Co a=2.5098 ±0.0005 1.549 5.84 c=4.0853 ±0.0002 2.559 21.9 (*) estimated by using ICP technique. J. Gutierrez et al. Journal of Alloys and Compounds 1010 (2025) 177211 3 or from 120 K for the Co nanoparticles) as α =1 a0 Δa ΔT(where a0is the lattice parameter at T 0 =300 K). Thus, the estimated thermal expansion coefficients α range from 8.85 (Fe and Fe-rich FeCo alloy) to 7.98 ×10 −6 K −1 for the Co-rich alloy (see Table 1). These values are slightly lower than coefficients for pure bulk Fe and Co, with values about 10–12 × 10 −6 K −1 , but this decrease in these coefficients between bulk and nanocrystalline phase of the material has been already observed, for example for Fe [31]. Moreover, our observations are in good agreement with theoretical predictions done in FeCo alloys about those slopes with values of 9.6 ×10 −6 K −1 [32] and 10.17 ×10 −6 K −1 [33]. In a solid, thermal expansion is caused by anharmonic terms in the interaction / restoring potential between individual atoms / molecules. For metals (as our Fe, FeCo alloys, and Co NPs are) at low temperatures, the dominant temperature dependence of the linear expansion coefficient turns out to be that of the specific heat, that is ≈a.T +b.T 3 , being the first term the electronic contribution and the second one the phonon contribution. As a consequence, at low temperatures the thermal expansion coefficient α deviates from linearity (see for example Fig. 2.5a in [34]) and in fact it goes to zero slope as temperature approaches absolute zero. Actually, the observed plateau in the range from 5 to some 70–80 K could be considered, at a sight, a trend of the appearance of the Invar effect. Invar and anti-Invar effects are governed by electronic properties and related to anharmonic enhacement of the total energy in the ground state. In the case of FeCo alloys, that anharmonicity is caused by the simultaneous presence of states with different magnetic character (ferromagnetism and antiferromagnetism or spin disorder), and associated to the simultaneous presence, with different percentages, of Fig. 1. TEM images of the raw powder of Fe, Co and FeCo alloys synthesized by the chemical reduction method. Fig. 2. a) MFM phase shift image of a single nanoparticle of the FeCo exc alloy and b) its corresponding topographical image; c) nanoparticle profile along the white line marked in b). J. Gutierrez et al. Journal of Alloys and Compounds 1010 (2025) 177211 4 different crystalographic structures, cubic bcc and fcc (extensive information can be found in [35]). This is not the case of our synthesized FeCo NPs: on the one hand, neutron diffraction patterns have shown a unique crystallographic structure (cubic bcc) for the three studied FeCo different composition alloys. On the other hand and as it will be further discussed, room temperature M¨ ossbauer spectra show undoubtedly one magnetic state for Fe leading to nice fits by using an only broad sextet. Moreover, well established Invar alloys like Fe 64 Ni 36 , exhibit a very low thermal expansion coefficient α below 2 ×10 −6 K −1 around room temperature compared to most metallic materials which have α values about 10–20 ×10 −6 K −1 [36]. The synthesized Fe and FeCo NPs show thermal expansion coefficient values much more close to that of common metallic materials. The procedure to extract the structural EXAFS signal (k• χ (k)) followed the standard steps of pre-edge background removal followed by a spline modeling of bare atomic background, edge step normalization through a polynomial function interpolated far above the edge region and energy calibration using the software ATHENA [37,38]. The modelling of atomic clusters centered on the absorber atom was obtained by ATOMS using atomic coordinates from Crystallographic Information Files (CIFs) [39]. Theoretical amplitude and phase functions were generated using the FEFF6 code [40]. Finally, EXAFS spectra were fitted through the ARTEMIS software [38] in the Fourier-Transform (FT) space; the analysis have been extended to the range of 1.5 –5.7 Å for all the samples, showing signals from further coordination shells. Fig. 6 shows the XAS spectra in the XANES region of FeCo exc (lowest Fe/Co ratio) Fe exc Co (highest Fe/Co ratio) nanoparticles at the Co and Fe K-edges (Fig. 6a and b, respectively), each one compared with the two pure Co and Fe nanoparticle samples (Fe NPs and Co NPs) and Feand Co-based standard compounds (Fe and Co pure elements foils as Fig. 3. a) MFM phase shift image for the FeCo composition nanoparticles, showing both single nanoparticles (red arrows) and one agglomerate of them (white arrow); b) its corresponding topographical image; c) nanoparticles agglomerate profile along the white line marked in b). Fig. 4. Room temperature ND patterns of the raw powder of Fe, Co and FeCo alloys. Fig. 5. Variation of the lattice parameters obtained from NPD patterns with the temperature of the raw powders of Fe, Co and FeCo alloys. J. Gutierrez et al. Journal of Alloys and Compounds 1010 (2025) 177211 5 references and oxides like Fe 3 O 4 , CoFe 2 O 4 and Co 3 O 4 ). The Co absorption edge values for FeCo exc , the metallic foil (Co foil) and the pure Co NPs are identical (7709 eV) while the shape of the XANES spectra are similar only for Co foil and Co NPs because of the hcp structure shared between them (FeCo exc presents a bcc structure instead, see EXAFS analysis in the following), even if a marked shift in the normalized intensity points out the inability of describing the nanoparticle sample as bulk-based material. Another difference is the lower amplitude of the oscillations above the edge, which is also typical of a nanostructured material [41]. The higher is the oxidation state of the studied element the higher is the energy of its edge position as well, so as the presence of pre-edge features which is distinctive for non-centrosymmetric sites. These characteristics can be seen in the CoFe 2 O 4 . For the FeCo exc alloy, the similarities with the Co foil spectrum, the position of the edge and the absence of pre-peak features indicate that Co is in the metallic state, without detectable amounts of Co oxides. Similar behavior is observed for FeCo and Fe exc Co nanoparticles, which only exhibit a slight decrease in the normalized intensities proportional to the decrease of Co content. A clearer trend is depicted for the Fe samples at the Fe K edge, due to the same bcc structure shared among Fe x Co 100-x alloys, Fe NPs and Fe foil. These samples present the same edge value (7712.3 eV), noticeably lower than the Fe 2+/3+-based standards, as well as similar overall XANES trends, with differences in the normalized intensities and absence of pre-edge peaks. No evidence of Fe-oxides presence has been detected all over the Fe x Co 100-x samples or in the Fe NPs. Structural refinement of the Fe and Co local environments (up to 5–6 Å from the absorber atom) was performed by least square minimization of the EXAFS signals (from ~150 eV above the absorption edge) of the Fe, Co and Fe x Co 100-x alloy nanoparticle samples (see Figs. 7 and 8). A full description of the refinement procedure can be found in the Supplementary Material (S2 in SM). After using ATOMS to generate a list of all the atomic coordinates in a cluster of 6 Å size and the FEFF6 software to calculate all the photo-electron scattering functions, the resulting paths were manually selected in order to obtain the best possible fits. Particular attention was given to the first five Fe-Fe and CoCo single backscattering paths used as reference parameters for refining the Fe-Fe and Co-Co distances. The Fe NPs and Co NPs EXAFS analysis highlighted an excellent agreement between the crystallographic distances of the starting CIF files and the found ones (see Fig. 7 and Table 2). The Co NPs local structure is perfectly described by the standard hexagonal hcp Co lattice, with only a 0.01 Å first shell contraction between experimental data and CIF model, both in-plane and out-of-plane (that is in path 1 and 2, see S3.1 in SM) while the Fe NPs spectrum matches the standard cubic α -bcc iron lattice, with a more pronounced average contraction of 0.02 Å all over the investigated shells (see Table 2). This light contraction can arise from the difference of temperature between the used CIF data, acquired at 298 K, and the samples data temperature acquired at 10 K. These results confirm the XANES findings (see Fig. 6) in terms of absence of FeO and Co-O signals and good structure correspondence between CIF models and our alloy samples spectra. Regarding the FeCo alloys, similar conclusions can be drawn, independently from the Fe/Co ratio, such as the absence of Feand Co-oxides evidences and a good correspondence with the crystallographic model used. The reference model resulting from all the best fits is exclusively the bcc, both at the Fe and the Co K edges, since the hcp Co turns out to be not suitable for describing any of the above samples, in particular the considerable splitting of the first shell distances. The use of a Fe-based standard to fit Co K edge signals is appropriate due to the close similarity between the Fe and Co photoionization cross-sections [42], while the adoption of a bcc crystal model attests the interchangeability of the two metals inside the FeCo alloy structure, since there is no evidence of relevant Co hcp structure presence. Despite the low-temperature measurements, it was not possible to discriminate between Co-Co, Co-Fe and Fe-Fe signals due to the close similarity between the photoelectron backscattering functions and typical bond distances of the two elements. The Fe and Co local structures, studied at the Fe and Co K edges respectively, match with good agreement the α -bcc iron CIF, with bond lengths values which approach the Fe NPs sample as the Fe/Co increases (see Fig. 8,Table 3 and Table 4). The multiple-edge fits, where the two edges are used simultaneously, lead to the same results (see Table 5). Again, the observed slight contraction between bcc Fe and our samples bond lengths can be explained from the difference of temperature between the used CIF data, acquired at 298 K, and the samples data acquired at 10 K. Regarding the S 0 2 value, which describes the amplitude of the EXAFS signal a considerable reduction is reported (Co foil: S 0 2 =0.79 +/- 0.03 fitted with hcp Co CIF; Fe foil: S 0 2 =0.80 +/- 0.06 fitted with bcc Fe CIF) that would point to a reduced dimensionality of the particles [41,43,44]. However, and applying the data for bcc Fe published by Beale &Weckhuysen [45], it would correspond to particles of 1–2 nm of diameter, a size that is not in agreement with our previous microscopy observations. Considering that the fits could include also the Fig. 6. Normalized XANES spectra of (a) Fe x Co 100-x alloys, Co NPs and related standards at the Co K edge, and (b) Fe x Co 100-x alloys, Fe NPs and related standards at the Fe K edge. J. Gutierrez et al. Journal of Alloys and Compounds 1010 (2025) 177211 6 higher coordination shells (that would be strongly depleted in case of reduced size), we conclude that the amplitude parameters cannot be exploited for an estimation of the particle size. Thus, the FeCo alloy nanoparticles appear homogeneous despite the Fe/Co ratio, without apparent nano-segregation of different phases. A clear trend validated by the three different fit approaches (see Fig. S3.2 in SM) seems to point to a slight increase of the bond lengths all over the five paths considered, which is directly proportional to the Fe content (0.013 and 0.026 Å for path 1 and path 10 respectively for the ≈ 30 % increase in Fe content from the FeCo exc composition to the Fe exc Co one). Finally, calculated Fe x Co 1-x alloys nanoparticles compositions from the ratio of the normalized edge steps at the Fe and Co edges can be seen in Table 6. It can be observed that compositions calculated from EXAFS measurements are slightly different than those obtained from ICP analysis and more similar to the expected ones from the stoichiometry of the precursors. This discrepancy is more related to the treatment of the sample before the analysis than to the technique itself. EXAFS measurements have been performed directly in the NPs powders, so the relation between the normalized edge steps is proportional to the Fe/Co content in the sample. Nevertheless, when performing ICP analysis, it could happen that solution of both metals was not very efficient and that small amounts of metals were left undissolved. This could be the reason of the discrepancies in ICP for the samples. Similar discrepancies have been also observed in other FeCo alloy nanoparticles synthesized by chemical methods [46,47]. Fig. 9 shows the room temperature hysteresis loop measured for the Fe, Co and Fe x Co 100-x nanoparticles (hysteresis loops measured at 5 K can be found in Fig. S4.1 of the SM). Table 6 summarizes the different magnetic characteristics obtained for each sample. In all samples magnetic saturation is clearly observed at applied magnetic field values above 2 MA/m (or above 2.5 T). The measured values of the room temperature saturation magnetization of our synthesized Fe and Co NPs well agree (within the experimental error) with the accepted values in the bulk state for pure Fe and Co, that is M S = 221.9 Am 2 /kg and 162.5 Am 2 /kg at R.T., respectively [48]. For the Fe x Co 100-x alloys, measured saturation magnetization follows previously well-known observations (see for example Figure 4.21(b) in [48]): the addition of cobalt increases the magnetization of the FeCo alloy, and the 30 % cobalt-content composition has the highest measured M S value at room temperature than any other known material. All our measured hysteresis curves show also low coercive field values, confirming the soft magnetic character of all the synthesized nanopowders. As expected, the anisotropic character of pure Co reflects in a higher coercivity (about 17 kA/m at room temperature) than the Fig. 7. Fourier transform module (transformation window in the range k=2–11.6 Å −1 for the Fe K edge spectrum and k=2–15.6 Å −1 for the Co K edge one) of the EXAFS spectra for the Fe NPs (a) and Co NPs (c) samples, analysed at the Fe and Co K edges respectively, together with the correspondent k-space signals ((b) and (d) respectively) (k 2 -weighted). Black dots are experimental data, solid red lines are the best obtained fit. J. Gutierrez et al. Journal of Alloys and Compounds 1010 (2025) 177211 7 pure Fe one (only 2 kA/m). For the Fe x Co 1-x alloy nanoparticles, this room temperature coercivity ranges between 4.8 and 5.7 kA/m, close to the desired good soft magnetic behavior [49]. At low temperature, 5 K, we have measured a slightly increase both in the saturation magnetization value as well as the coercive field for all samples, but the soft magnetic character observed at room temperature remains. Close attention should be payed to the remanence-to-saturation (M r / M S ) magnetization value at R.T., of the order of 10 −2 in all the studied alloys. This fact will be subject of further discussion. In Fe x Co 100-x alloys, the measured saturation magnetization values clearly correlate with the atomic order and level mixing of Fe and Co atoms and subsequent redistribution of spin states in the Fe sites. This fact arises from the dependence of the Fe and Co magnetic moment values on their environment: while magnetic moment of Co atoms does Fig. 8. Fourier transform module (transformation window in the range k=2–11.6 Å −1 for Fe K edge spectra and k=2–15.6 Å −1 for Co K edge ones) of the EXAFS spectra for the three Fe x Co 100-x alloy samples, analysed at the Fe (a) and Co K edge (c), together with the correspondent k-space signals ((b) and (d) respectively) (k 2 - weighted). Black dots are experimental data, solid red lines are the best fit. All the spectra are shifted vertically for better visualization. Table 2 structural parameters of the best fits from Fig. 7 obtained by multiple-shell fitting of the Fe NPs and Co NPs EXAFS spectra using Fe bcc and Co hcp base models respectively (see Supplementary Material). Fe and Co nanoparticles Sample Path 1 Path 2 Path 5 Path 8 Path 10 Fe NPs S2 00.42 ±0,06 R(Å)2.465 ±0.007 2.84 ±0.01 4.037 ±0.008 4.733 ±0.009 4.944 ±0.01 σ 2(Å2)0.0028±0.0012 0.0029±0.0029 0.0049±0.0012 0.0048±0.0012 0.0047±0.0029 Sample Path 1 Path 2 Path 3 Path 7 Path 10 Co NPs S2 00.39 ±0.01 R(Å)2.486 ±0.002 2.497 ±0.002 3.539 ±0.003 4.069 ±0.004 4.337 ±0.004 σ 2(Å2)0.0042±0.0002 0.0042±0.0002 0.0066±0.0002 0.0065±0.0002 0.0065±0.0002 J. Gutierrez et al. Journal of Alloys and Compounds 1010 (2025) 177211 8 Table 3 Structural parameters of the best fits from Fig. 8a and b obtained by multiple-shell fitting of the Fe x Co 100-x alloys EXAFS signals acquired at the Fe K edge, employing Fe bcc as base model. Fe x Co y alloy nanoparticles (Fe K edge) Sample Path 1 Path 2 Path 5 Path 8 Path 10 FeCo excess S2 00,55 ±0,08 R(Å)2.455 ±0.006 2.83 ±0.01 4.021 ±0.008 4.715 ±0.009 4.924 ±0.01 σ 2(Å2)0.0035±0.0011 0.0043±0.0015 0.0058±0.0011 0.0057±0.0011 0.0056±0.00215 FeCo S2 00.53 ±0.08 R(Å)2.465 ±0.006 2.834 ±0.01 4.035 ±0.008 4.731 ±0.009 4.94 ±0.01 σ 2(Å2)0.0036±0.0011 0.0042±0.0014 0.0058±0.0012 0.0056±0.0014 0.0056±0.0012 Fe excess Co S2 00.46 ±0.05 R(Å)2.470 ±0.005 2.847 ±0.009 4.044 ±0.008 4.74 ±0.010 4.95 ±0.010 σ 2(Å2)0.0026±0.0012 0.003±0.0012 0.0053±0.0012 0.005 ±0.0012 0.0051±0.0012 Table 4 Structural parameters of the best fits from Fig. 8c and d obtained by multiple-shell fitting of the Fe x Co 100-x alloys EXAFS signals acquired at the Co K edge, employing Fe bcc as base model. Fe x Co y alloy nanoparticles (Co K edge) Sample Path 1 Path 2 Path 5 Path 8 Path 10 FeCo excess S2 00,48 ±0,03 R(Å)2.459 ±0.003 2.832 ±0.006 4.022 ±0.004 4.716 ±0.005 4.925 ±0.005 σ 2(Å2)0.0031±0.0004 0.0046±0.0008 0.0059±0.0004 0.0057±0.0006 0.0057±0.0004 FeCo S2 00.56 ±0.03 R(Å)2.463 ±0.003 2.840 ±0.005 4.033 ±0.004 4.729 ±0.005 4.939 ±0.005 σ 2(Å2)0.0033±0.0004 0.0042±0.0007 0.0055±0.0004 0.0053±0.0007 0.0053±0.0004 Fe excess Co S2 00.63 ±0.04 R(Å)2.468 ±0.003 2.853 ±0.005 4.040 ±0.004 4.738 ±0.005 4.948 ±0.005 σ 2(Å2)0.0032±0.0004 0.0040±0.0007 0.0052±0.0004 0.0051±0.0004 0.0051±0.0007 Table 5 Structural parameters obtained by multiple-shell fitting using concurrently both of the Fe x Co 100-x alloy EXAFS signals acquired at the Fe and Co K edge, employing Fe bcc as base model. Fe x Co y alloy nanoparticles (Fe and Co K edges) Sample Path 1 Path 2 Path 5 Path 8 Path 10 FeCo excess S2 0(Fe K) 0.53 ±0.03 S2 0(Co K) 0.48 ±0.03 R(Å)2.455 ±0.003 2.830 ±0.004 4.021 ±0.004 4.715 ±0.005 4.925 ±0.005 σ 2(Å2)0.0032±0.0004 0.0043±0.0006 0.0060±0.0004 0.0058±0.0005 0.0058±0.0005 FeCo S2 0(Fe K) 0.53 ±0.03 S2 0(Co K) 0.57 ±0.03 R(Å)2.465 ±0.003 2.839 ±0.004 4.036 ±0.004 4.733 ±0.005 4.943 ±0.005 σ 2(Å2)0.0034±0.0004 0.0043±0.0006 0.0056±.0005 0.0055±.0004 0.0055±.0005 Fe excess Co S2 0(Fe K) 0.47 ±0.03 S2 0(Co K) 0.63 ±0.04 R(Å)2.469 ±0.003 2.848 ±0.004 4.042 ±0.004 4.740 ±0.005 4.951 ±0.005 σ 2(Å2)0.0029±0.0004 0.0037±0.0006 0.0052±.0004 0.0050±.0005 0.0050±.0006 Table 6 Magnetic properties obtained for Fe, Co and Fe x Co 100-x alloys NPs. Sample composition (ICP) Sample composition (EXAFS) M S (Am 2 /kg) *1 H c (kA/m) *2 M r / M S 5 K 300 K 5 K 300 K 5 K 300 K Fe Fe 224.6 220 3 2 0.008 0.012 Fe 66 Co 34 Fe 74 Co 26 232.5 228.3 8 4.8 0.038 0.023 Fe 51 Co 49 Fe 62 Co 38 224.3 220 10 5.7 0.034 0.022 Fe 33 Co 67 Fe 45 Co 55 213.6 210.7 8.8 5.4 0.036 0.023 Co Co 161.2 158.2 31.8 17 0.101 0.070 *1 Error in magnetization measurement: 0.5 % of measured value. *2 Error in magnetic field determination: ±1 kA/m. J. Gutierrez et al. Journal of Alloys and Compounds 1010 (2025) 177211 9