Inspired by nature: investigating tetrataenite for permanent magnet applications

While increased global connectivity facilitates rapid communication, adoption and production of new technologies, it can also affect supply chain vulnerability—and associated economic impact—if connectivity links are disrupted. Although disruptions are unquestionably detrimental to industries that require the disturbed technology or supplies, at the same time they spark new creative strategies that innovate around the issue. This type of scenario, in part, drove the development of the neodymium-based supermagnets in the 1980s, in response to US Pentagon reports [1] that acknowledged a shortage of critical elements, including cobalt, essential to samarium-based magnets. A similar situation is in play today, with geopolitical shortages of technologically critical rare earth elements such as La, Nd, Pr and Sm, and intrinsic shortages of Y, Eu, Gd, Dy and Tb. In particular, advanced permanent magnets that rely upon rare earth elements for their functionality constitute one of the most important groups of strategic materials; therefore, strong motivation exists to develop superior magnetic materials that are largely free of rare earth elements. Such materials must have a large, temperature-insensitive magnetic remanence B_r matched with a large coercivity H_ci that likewise maintains its integrity at high temperature. These attributes combine into the maximum magnetic energy product (BH)_max, the figure of merit that indicates the maximum energy a given magnetic material can supply to an external magnetic circuit when operating on its demagnetization curve. An ideal next-generation permanent magnet should be inexpensive, easy to process, lightweight and corrosion resistant. Here we present results, analyses and performance projections of a novel magnetic material—tetrataenite—comprised principally of the readily available elements Fe and Ni, that is projected to possess very similar, and in certain circumstances improved, magnetic attributes to the Nd₂Fe₁₄B-based supermagnet. Currently, Nd₂Fe₁₄B-based magnets are an important component of the permanent magnet market, with projected growth to exceed $20 billion by 2020 [2].

Tetrataenite is an intermetallic ferromagnetic compound of nominal equiatomic composition FeNi with the tetragonal L1₀ crystal structure that lends an associated strongly uniaxial magnetic character. Tetrataenite is only found naturally in meteorites that formed over millions of years due to its extremely low atomic Fe and Ni mobilities, with a calculated diffusivity of one atomic jump per 10 000 years at 300 °C [3]. While tetrataenite is anticipated to be extraordinarily challenging to synthesize in the laboratory, it is confirmed in this work to hold good potential as a superior rare-earth-free permanent magnet. To illustrate the potential of L1₀-structured FeNi as a permanent magnetic material we correlate structural and magnetic data obtained from tetrataenite existing in meteorite NWA 6259. These results are informed and extended through computational exploration of the consequences of antisite occupancy on the magnetic moment and magnetocrystalline anisotropy of L1₀ FeNi. Overall, these results inform the potential for production of L1₀ FeNi at time and temperature scales suitable for bulk magnet production.

Among the various sources of magnetic anisotropy, including magnetocrystalline, shape, and stress anisotropies, magnetocrystalline anisotropy provides the largest value and is thus the favored phenomenon to induce coercivity in advanced permanent magnets. Efforts to engineer rare-earth-free permanent magnetic materials are hampered by the fact that the robust magnetocrystalline anisotropy arising from the lanthanide 4f electronic states [4] is no longer available for exploitation. This missing magnetocrystalline anisotropy may be recovered, to various degrees, in materials that adopt a low-symmetry crystal structure, such as hexagonal or tetragonal crystal structures. In such low-symmetry crystal structures, the material's magnetic moment may align perpendicular to the basal plane direction, providing two energy minima for the magnetization that defines the uniaxial magnetic anisotropy state. While the majority of strongly magnetic transition-metal alloys assume a high-symmetry cubic structure that displays low magnetocrystalline anisotropy, certain transition-metal alloys compounds form in the tetragonal L1₀ (P4/mmm space group or AuCu I structure type) structure, with an associated uniaxial magnetic character.

The L1₀ structure can form in equiatomic compounds; it features alternating layers of chemically ordered constituent elements stacked parallel to the tetragonal c-axis, creating a natural superlattice. The top monochromatic image of figure 1 illustrates the chemically disordered A1 face-centered-cubic (fcc) structure, with equal probability of atomic site occupation by either of the two component elements. The colored images of figure 1 illustrate the tetragonal L1₀ structure, which forms through a disorder–order transformation from the parent A1 phase. The chemically ordered structure is the thermodynamically stable phase below a critical temperature T_CR; in this structure each crystal lattice site has a different probability of being occupied by one of the two atom types. In ferromagnetic systems, the critical order–disorder temperature can vary rather widely and marks a thermodynamically first-order phase transition in the Ehrenfest sense [5]. For the Ehrenfest designation, the order of a transformation is given by the nth derivative of the free energy that results in a discontinuity with respect to at least one state variable. For a first-order disorder–order transformation, a discontinuous change in the entropy S is expected from the definition of the Gibbs free energy G(T, P):

$$\begin {eqnarray}\label {eq1} \left (\frac {\partial G}{\partial T}\right )_{P,N_i}=-S \end {eqnarray} \tag{ 1 }$$

where T is the temperature, P is the pressure and N_i are the moles of components i. In accordance with the anticipated character of the first-order thermodynamic phase transition, the transformation from the low entropy ordered phase to the high entropy disordered phase is endothermic. The A1 to L1₀ phase transformation proceeds by nucleation and growth of L1₀ chemically ordered regions within the chemically disordered A1 matrix [6]. The local atomic rearrangements necessary for chemical ordering can be facilitated by creating point defects (vacancies) in the structure, or by the addition of ternary alloying elements that enhance diffusion rates. Since the chemical ordering can occur along any one of the three original cube axes (the 〈001〉 directions), three variants of the L1₀ phase may form from the parent A1 phase, each variant possessing a c-axis that is parallel to one of the three original cube axes, as shown in figure 1.

Zoom In Zoom Out Reset image size

Within the magnetic materials and metallurgy communities, the L1₀ form of FeNi remained largely unknown until quite recently [7]; in fact, most published Fe–Ni binary phase diagrams do not explicitly show the tetrataenite phase. Initial work by Paulevé et al in 1962 [8] reported unique properties in FeNi as a result of neutron bombardment under the influence of a magnetic field. This work was followed in 1964 by Néel et al [9], who reported attainment of the L1₀ structure and a coercivity increase in a polycrystalline, nominally equiatomic FeNi metal sample.

The intervening decade saw little published activity on the L1₀ form of FeNi within the physics and metallurgy communities. The activity was picked up within the space sciences community starting in 1977, with tetragonal FeNi identified in large numbers of stony, stony-iron and iron meteorites [10]. In 1980, the chemically ordered tetragonal form of FeNi present in a number of meteorites was optically surveyed and confirmed by Clarke and Scott, who coined the meteoritical phase name, 'tetrataenite' [11]. Tetrataenite is found in the famous Widmanstätten pattern that consists of long interwoven nickel–iron crystals in which the close packed planes of kamacite (body-centered cubic form of low-Ni iron) form on the close packed planes of fcc taenite (face-centered-cubic form of medium-Ni iron). It has been determined that typical cooling rates to facilitate the long-range chemical ordering of tetrataenite are in the range 0.3−6600 K Myr⁻¹, with a typical composition range from 45 to 55 wt% (44–54 at.%) Ni [12]. Based on measured phase and composition relationships in tetrataenite, its diffusivity constant has been calculated to correspond to one atomic jump every 10 000 years at 300 °C [3]. Reuter et al synthesized the FeNi L1₀ phase under high energy electron irradiation in an electron-transparent alloy foil and utilized electron diffraction for the first time to characterize the FeNi L1₀ phase present in different meteorites as well as in the laboratory-synthesized samples [13]. More recently, molecular beam epitaxy was used to fabricate FeNi films that showed some characteristics consistent with the tetragonal form of the compound [14]. While reports on magnetic [15] and microstructural [16] characterization of FeNi films have been published much more recently, these communications do not provide detailed understanding of the correlations between magnetic properties and microstructure to fully realize the coercivity-limited energy product of the FeNi L1₀ phase at room temperature. Similarly, magnetic data have been reported on a wide variety of meteoritic materials containing tetrataenite that exhibit a correspondingly wide distribution of FeNi-based phases, phosphides and microstructures [17]. In particular, in a series of papers Nagata et al reported on the magnetic properties of tetrataenite found in a wide variety of meteorites, including the Olivenza, St Séverin [18] and Santa Catharina meteorites [19], all with significantly different overall metal contents and different thermal histories. These reports offer room-temperature coercivities attributed to tetrataenite in the range of 87.5 kA m⁻¹ (1100 Oe) ∼ 318 kA m⁻¹ (4000 Oe), as quantified by direct measurement in some cases and deduced from so-called remanence coercive force measurement in other cases [20]. Similarly, a range of magnetic remanence values attributed to tetrataenite are likewise provided in these references, but these values are not associated with detailed chemical composition determinations of specific phases, nor are they strongly tied to microstructural characterization. As magnetic parameters that determine the maximum energy product (BH)_max are essentially determined by the microstructure, reported values of remanence and coercivity in the absence of detailed structural and chemical characterization are insufficient to assess the potential of tetrataenite as a permanent magnetic material.

Very recently, natural material containing the highest known volume of tetrataenite has been discovered, allowing detailed structural and magnetic exploration of the potential of L1₀-type FeNi as a rare-earth-free permanent magnet. In 2010 a meteorite known as NWA 6259 (named for its location of discovery, Northwest Africa) was discovered and verified to consist of largely single-phase tetrataenite. NWA 6259 is a high-Ni ataxite with an overall composition of 42.6 wt% (41.4 at.%) Ni [21], the second highest Ni content of all known iron meteorites. In this work, L1₀ FeNi extracted from the meteorite NWA 6259 is utilized as a natural source of the chemically ordered tetrataenite phase. The chemical composition of microstructure of the meteorite NWA 6259 was determined with electron microprobe analysis (EPMA), while optical microscopy and electron backscatter diffraction were employed to understand the texture and mosaic structure of the bulk meteorite sample. Transmission electron microscopy was utilized to verify the presence of the L1₀ FeNi structure throughout the sample. Characterization of the chemical disordering transformation of tetrataenite was accomplished using differential scanning calorimetry (DSC); magnetometry was carried out on both highly deformed/polycrystalline and undeformed volumes of the meteorite to assess the structure-sensitive technical magnetic properties as well as intrinsic properties of the material. These results are supported by computational investigations that confirm the influence of the precise L1₀ FeNi stoichiometry on the intrinsic magnetic properties of saturation magnetization and magnetocrystalline anisotropy. Overall, it is confirmed that tetrataenite possesses the requisite magnetic attributes that approximate those of rare-earth-based 'supermagnets'. While bulk synthesis processes and protocols are yet to be established, successful bulk laboratory synthesis of tetrataenite will be a disruptive technological advance.

The crystal structure, microstructure and magnetic attributes of the meteorite NWA 6259 sample were examined using optical microscopy, transmission electron microscopy, electron probe microanalysis and magnetometry. Optical microscopy was carried out on a sample mounted in a 1.25 in (3.175 cm) diameter epoxy mount, metallographically polished and etched with 2% nital solution. Polarized light microscopy was conducted on a sample that was metallographically polished through 0.05 μm colloidal silica to provide a smooth surface free of mechanical deformation. The presence of L1₀-ordered tetrataenite in this meteorite sample was confirmed using electron diffraction studies carried out at Sandia National Laboratories. Transmission electron microscopy (TEM) analysis was conducted on a sample thinned to electron transparency using a dual beam FEI DB-235 focused ion beam scanning electron microscope (FIB/SEM). An FEI Titan G2 80–200 TEM with ChemiSTEM technology, operated at 200 kV and equipped with a high-brightness Schottky electron source, a spherical aberration corrector and a silicon drift detector (SDD) energy-dispersive x-ray detector array, was employed. Cylindrical samples were prepared for microstructural and magnetic characterization. The chemical composition of the NWA 6259 meteorite matrix at the micron level was obtained with a CAMECA SX50 Electron Microprobe. The matrix composition from at least 20 points within the meteorite matrix was measure in each sample, avoiding regions surrounding precipitates in the matrix. The crystallographic orientation of the L1₀-ordered FeNi phase in multiple regions of the meteorite NWA 6259 was probed with electron backscattered diffraction (EBSD). To better understand relationships between the microstructure, magnetic properties and crystallographic orientation, EBSD was conducted on both highly textured and poorly textured regions of the NWA 6259 sample. For these experiments, cylindrical specimens were cold-mounted in conductive epoxy and polished in several steps to a final polish achieved using 0.05 μm colloidal silica (Struers OPS). No etching was needed to obtain clear EBSD patterns. The scans were performed at the center of the mounted specimen at ∼100× magnification using a 6 μm step size. An EDAX-TSL system was used to generate the data sets.

Magnetic measurements were made on selected samples of the meteorite NWA 6259 to examine intrinsic and extrinsic magnetic attributes. Vibrating sample magnetometry and superconducting quantum interference device (SQUID) magnetometry were carried out on samples extracted from the meteorite. The specimen was ac-demagnetized to provide information concerning the mechanisms involved in the approach to saturation. Hysteresis loops were measured at 5 K in the field range − 5 T ≤ H_appl ≤ 5 T and at 300 K from − 2.85 T ≤ H_appl ≤ +2.85 T, including as a function of angle, to characterize the initial magnetization and coercive behavior. Demagnetizing effects were accounted for in all magnetic measurements. Confirmation of chemical disordering of the meteorite with increased temperature was obtained with differential scanning calorimetry (DSC). Calorimetry was carried out on a sample of the meteorite NWA 6259 in the temperature range 30–700 °C at controlled heating and cooling rates of 10 °C min⁻¹ under 50 ml min⁻¹ flowing UHP Ar. Three heating and cooling cycles were performed to examine the reversibility of the reaction.

As indicated earlier, tetrataenite can accommodate a composition range from 44 to 54 at.% Ni by incorporating a relatively high concentration of antisite defects. It is of interest to computationally investigate the effect of such antisite disorder on the magnetic attributes of the phase, such as the moment per atom, the magnetocrystalline anisotropy constants and the Curie temperature.

First-principle calculations were performed using the frozen-core full-potential projected augmented-wave (PAW) method, as implemented in the Vienna ab initio simulation package (VASP), and the electronic exchange and correlation effects are described within the generalized-gradient approximation (GGA) [22, 23]. The energy cutoff of the plane wave basis set is taken as 450 eV and the total energy of the system is converged to 10⁻⁶ eV for all the structures. To ensure reasonable accuracy, 1183 k-points in the irreducible part of the Brillouin zone are used.

5.1. Crystal structure and microstructure

The meteorite NWA 6259 microstructure consists of a quasi-uniform metallic matrix with dispersed needle-shaped precipitates of (Fe, Ni) phosphides, of the order of 5 vol.%, figure 2. The orientations of needle-shaped phosphide precipitates are generally consistent across the sample, suggesting that they formed when the meteorite was already in single-crystal form. Polarized light optical microscopy allows clear observation of the anisotropic nature of the matrix phase. Figure 3 shows a polarized light microscopy image of a region of the NWA 6259 sample that was obtained under the condition of crossed polarizer/analyzer. This image highlights magnetic domains, where two contrast states can be discerned that correspond to two of the three different crystallographic variants existing in the size range 10−100 μm.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Electron probe microanalysis confirmed the average chemical composition of the meteorite NWA 6259 to be in the range 43.5 wt% (±0.5) (42.3 at.%) to 45 wt% (±0.5) (43.5 at.%) Ni. This overall composition indicates that the tetrataenite phase is Fe rich relative to the equiatomic model compound, with 5–6% antisite defects of iron atoms residing in nickel atomic locations. Figure 4(a) shows a TEM dark field image of the FIB-cut sample of the meteorite NWA 6259 matrix. This dark field image was taken from a superlattice reflection {001} as shown in figure 4(b). The presence of the {001} superlattice reflection, along with allowed A1 (fcc) structure {002} reflection in the electron diffraction pattern confirms the presence of the FeNi L1₀-ordered tetrataenite phase in the meteorite NWA 6259 matrix. The chemically ordered phase appears as a bright area as indicated by the blue arrow in figure 4(a). On the left side of this bright area, a darker area (yellow arrow) also shows a weak {001} superlattice reflection as shown in figure 4(c), which indicates another variant of L1₀-ordered phase. The measured tetragonality of L1₀ FeNi exhibits only a small deviation from cubic symmetry, with a c/a ratio of 1.0026 ± 0.0009 [24].

Zoom In Zoom Out Reset image size

The EBSD measurements reveal the extremely large microstructural variability, and associated technical magnetic property variability, of the tetrataenite phase in meteorite NWA 6259. Figure 5 shows a color-coded inverse pole figure, i.e. a crystal orientation map, along the sample normal direction of a poorly textured region in the sample. Orientation gradients are discernible as color gradients within some mostly uniform regions likely as a result of the presence of dislocations, reflective of the high-strain/high-impact history of the meteorite. This microstructure may be thought of as a mosaic of crystalline blocks with dimensions in the range 50−300 μm, fractionally tilted relative to each other. Each block is separated from the surrounding blocks by faults and dislocations. In contrast, EBSD measurements on two other regions of the sample, such as shown in figure 6, demonstrate a homogeneous grain orientation near the (101) plane normal direction over a macroscopic scale, confirming a highly anisotropic structure. Variations in the crystal mosaic orientation in this region of the meteorite, visible in the figure as changes in color and intensity, are estimated to be about 12°, at most. These results confirm that selected regions of the meteorite NWA 6259 are indeed highly crystallographically textured, and as such are key to the confirmation and quantification of the technical magnetic properties of the tetrataenite phase.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

5.2. Phase transition and magnetic character

The differential scanning calorimetry trace of a portion of meteorite NWA 6259 is shown in figure 7 and features a prominent, broad and irreversible endothermic peak with heating that occurs at an onset temperature of ∼800 K (∼530 °C). These data confirm that the kinetically limited chemical disordering of the L1₀ tetrataenite phase to the disordered A1 phase begins at ∼800 K (∼530 °C) and is most rapid at 830 K (∼560 °C). The influence of microstructure and short-range chemical order on the technical magnetic properties of tetrataenite is remarkable. Figure 8 shows a room-temperature magnetic measurement obtained from the highly defective/polycrystalline volume of meteorite NWA 6259 that was imaged in figure 5. It is noted that the sample exhibits essentially zero remanence and zero coercivity, as might be expected from magnetically isotropic, magnetically soft materials. In contrast, magnetic data collected from a sample originating from the highly uniform region of NWA 6259 possess appreciable coercivity and remanence, attesting to the potential of L1₀-structured FeNi as a rare-earth-free permanent magnetic material. Figure 9 shows hysteresis curves measured at 5 and 295 K from a tetrataenite sample obtained from the highly textured region of meteorite NWA 6259 imaged in figure 6. A number of magnetic features of the sample are highlighted and discussed here. It is observed that the chemically ordered sample has an appreciable coercivity of approximately 95.5 kA m⁻¹ (1200 Oe) that is largely independent of temperature, with an accompanying saturation magnetization value of 1.5 T (15 kG) at 5 K that decreases by only 8% to 1.38 T (13.8 kG) at room temperature. Correcting for the approximately 5 vol% of non-magnetic phosphide and sulfide phases present in the matrix (see above), the corresponding saturation magnetization values for NWA 6259 tetrataenite are 1.58 T at 5 K and 1.45 T at room temperature. As noted earlier, the tetrataenite phase in the meteorite NWA 6259 possesses the second highest Ni content of all known meteoritic manifestations of this phase. It is anticipated (and computationally confirmed, see section 5.3) that antisite nickel occupancy may couple antiferromagnetically to neighboring atoms, reducing the magnetic moment below values reported for tetrataenite derived from other meteorites [19, 20]. The knee in the magnetization curve of the chemically ordered sample that appears at about 1.6 T is associated with the magnetocrystalline anisotropy field that is, like the coercivity, largely independent of temperature. More precise quantitative extraction of the anisotropy from the shape of the demagnetization curve requires accounting for the complex tri-variant magnetic structure of naturally occurring tetrataenite and the angular variations of easy axis directions within the sample, and is beyond the scope of this paper. It is hypothesized that the unusual step-like demagnetization behavior noted in the first quadrant of the hysteresis curves of figure 9 reflects the superposition of the M(H) contributions from the three mutually orthogonal variants of the L1₀ tetrataenite structure, each oriented at a different angle to the applied magnetic field. The easy axis of one variant lies close to the applied field direction, and thus saturates almost immediately; indeed, the extrapolated remanence is roughly one-third of the saturation magnetization. The other two variants (two-thirds of the total moment) contribute hard-magnetic character to the magnetization curves, saturating near 2 T. The inset of figure 9 emphasizes the gradual approach to magnetic saturation that is consistent with magnetization development through a pinning-controlled mechanism. It is conjectured that the variant boundaries may act as a source of pinning in this region.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

After undergoing the phase transformation to the chemically disordered A1 cubic structure above 800 K, the sample exhibits a significant decrease in coercivity to effectively zero while simultaneously demonstrating an increase in the saturation magnetization by 15% to 1.72 T (17.2 kG) at 5 K.

5.3. Computational insight

A particular feature of the L1₀-ordered FeNi phase, tetrataenite, is a certain degree of antisite disorder and off-stoichiometry. For example, the tetrataenite phase present in the meteorite NWA 6259 investigated here has the approximate composition Fe₅₆Ni₄₄. It is therefore necessary to consider the effect of chemical disorder on the magnetic properties, such as atomic moment per atom, net magnetization, Curie temperature and magnetocrystalline anisotropy of this compound. This task becomes particularly important due to competing phases, such as disordered A1-structured (fcc) FeNi. These chemically disordered cubic phases, which include technologically important soft-magnetic phases (permalloys), have been investigated for decades [25], but even today still provide various experimental and theoretical challenges, including the influence of site disorder on the technical magnetic properties.

Figure 10 depicts the structures considered in the calculations of the antisite and off-stoichiometric chemical disorder in L1₀-structured FeNi. The following crystallographic configurations and defect types displayed in figure 10 were examined and compared: (a), perfect L1₀-type chemical order; (b), antisite defects, where two Ni and Fe nearest neighbors in the supercell interchange their positions, and (c), which provides the addition of one Fe atom per unit cell. The Fe and Ni atoms occupy regular lattice sites in the supercell, and optimized lattice parameters of a = 3.557 Å and c = 3.57 Å have been employed in the calculations. Table 1 summarizes the results of these calculations, by showing moments per atom. When these numerical results are recast in terms of moment changes per defect, it is demonstrated that each antisite defect reduces the total moment by 0.14 μ_B, whereas each extra Fe atom enhances the total moment by about 1.56 μ_B. These values describe the limit of small defect concentrations. Once the concentration of iron-rich and/or antisite defects becomes too high, the defects start to interact with each other, and the moment is no longer a linear function of the defect density.

Zoom In Zoom Out Reset image size

Table 1.  Calculated Fe, Ni and total moments for the FeNi structures of figure 10.

	Perfect L10	Antisite L10	Fe17Ni15
Total moment (μB/atom)	1.625	1.620	1.673
Average Fe moment (μB)	2.630	2.616	2.604
Average Ni moment (μB)	0.619	0.625	0.619

It is noted that the supercell approach employed here is better adapted to the present problem than the widely used coherent potential approximation (CPA). The CPA is a single-site approximation which replaces the real atomic environment by a self-consistently obtained nonrandom effective medium. In a binary AB alloy, where atoms of one type are denoted 'A' and atoms of a second type are denoted 'B', this approximation fails to properly describe cluster-localization effects where electrons preferentially occupy sites A or B [26, 27], for example in the A or B planes in L1₀ alloys. Note that the values in table 1 contain an inaccuracy of the order of 0.02 μB, because the atomic positions slightly relax, in spite of the very similar atomic radii.

A mean-field Curie-temperature estimate for FeNi is obtained by considering figure 11 with M  = Ni. The corresponding energies per unit cell (figure 11), using the optimized lattice constants, are as follows: −55.186 921 eV (ferromagnetic (FM) configuration), −53.638 711 eV (checkerboard antiferromagnetic (AFM) configuration), and −54.640 657 eV (layered AFM configuration). These values show that the ferromagnetic state has a much lower energy than the competing antiferromagnetic states and yields a Curie temperature T_c = 1000 ± 200 K. For chemically disordered Fe₅₀Ni₅₀, the Curie temperature is reported to be about 800 K, but there remains some uncertainty in the literature, both experimentally and theoretically, and the T_c of chemically ordered FeNi, tetrataenite, is difficult to determine experimentally due to the kinetically limited nature of the coupled magnetic and structural phase transition. The high Curie temperature noted for L1₀-type FeNi is an effect of the strong Fe–Fe interatomic exchange. Note that the relatively stable ferromagnetic interactions and the quite high Curie temperature are a firm indication of a weak temperature dependence of the magnetocrystalline anisotropy. Since the temperature dependence of the anisotropy is largely determined by the Curie temperature [28, 29], the strong ferromagnetic coupling in tetrataenite indicates that the tetrataenite anisotropy energy K₁(T) is anticipated to be fairly temperature independent. This is an advantage over Nd₂Fe₁₄B, which has a higher room-temperature anisotropy but a very unfavorable temperature coefficient dK₁/dT = −2.16 × 10⁻² MJ m⁻³ K⁻¹ (−0.45% K⁻¹) [30].

Zoom In Zoom Out Reset image size

The results reported here present, for the first time, correlations between crystal structure, microstructure and magnetic properties of naturally occurring chemically ordered L1₀-type FeNi, also known as tetrataenite, and thereby provide fundamental information concerning its promise as a rare-earth-free permanent magnetic material. It is recognized that laboratory synthesis of tetrataenite in bulk quantities is an extremely demanding challenge; however, once this goal is achieved, these results strongly suggest that tetrataenite will be an important magnetic materials option for technological application. The formation of the chemically ordered L1₀ phase from the chemically disordered FeNi parent phase requires short-range atomic rearrangements to take place without change of composition. Two mechanisms that encourage short-range diffusion, rather than long-range diffusion, are (i) the generation and utilization of a large concentration of vacancies to facilitate chemical order and (ii) manipulation of thermodynamic phase stability. It is anticipated that the introduction of vacancies and chemical disorder may be accomplished both by conventional metallurgical-based techniques and by non-equilibrium processing techniques. Guidance for the selection of appropriate substitutional alloying additions that may stabilize the L1₀ FeNi structure may be found on the basis of first-principle calculations of L1₀ compound stability carried out by the Center for Atomic-Scale Materials Physics at the Technical University of Denmark (CAMP-DTU, [31])

The measured and calculated saturation magnetization of FeNi with the tetragonal L1₀ structure is very close to that of Nd₂Fe₁₄B, but with much smaller reduction in magnitude with increasing temperature. The Curie temperature of tetrataenite, approximately 830 K (560 °C), is coincident with the kinetic chemical order–disorder temperature and is higher than that of Nd₂Fe₁₄B-based magnets by 100°–200°. Calculations emphasize that low levels of Fe antisite substitution on the Ni site in L1₀-ordered FeNi are very favorable for fostering a high moment, with a strongly ferromagnetic Fe–Ni interlayer exchange coupling. However, higher levels of Fe antisite defects are detrimental to the magnetic exchange, causing the Curie temperature and the magnetocrystalline anisotropy to drop drastically. Similarly, the temperature dependence of the coercivity, and by implication the magnetocrystalline anisotropy, is low. By way of comparison, the saturation magnetization of Nd₂Fe₁₄B drops by 11% from its maximum value of 1.8 T at 135 K (just above the low-temperature spin reorientation) to 1.6 T at room temperature, while that of tetrataenite is reduced by only 8% in a similar temperature range. The microstructure of the tetrataenite sample in meteorite NWA 6259 is far from optimal; the presence of the three mutually orthogonal crystallographic variants of the L1₀ structure complicates the approach to saturation and indicate that microstructural engineering to favor a single variant may be necessary. Even so, the naturally occurring coercivity of tetrataenite found in the meteorite NWA 6259 is already 1200 Oe. An optimized microstructure will allow full manifestation of the intrinsic magnetic properties for maximum coercivity and remanence to realize a high energy product. These preliminary results point to a tetrataenite theoretical magnetic energy product $(\mbox {\textit {BH}})_{{\rm max}}= 1/4 B_{{\rm r}}^{2}= 420\ {\rm kJ}\ {\rm m}^{-3}$ (53 MG Oe) in the highly textured region of meteorite NWA 6259. The estimated anisotropy field (1.6 T) is comparable to the saturation magnetization (1.45 T); in this circumstance the energy product may be coercivity limited (i.e., H_ci < 1/2B_r). Even so, a theoretical coercivity-limited energy product of at least 335 kJ m⁻³ (42 MG Oe) is anticipated, making L1₀ FeNi magnets competitive with today's best rare-earth-based magnets [32].

This research is funded under a cooperative agreement with the US Department of Energy's Advanced Research Project Agency—Energy (ARPA-E). The authors would like to thank Drs Joseph Michael and Paul Katula at Sandia National Laboratory for the assistance with the TEM, Dr Michael J Jercinovic at the University of Massachusetts for the assistance with electron probe microanalysis andDr Michael P Balogh and Nicole Ellison for assistance with x-ray diffraction. A special thank you to Dr Rob Reisener, Cave Creek, AZ, USA, for inspiration and information regarding NWA 6259.