2017 Volume 58 Issue 11 Pages 1511-1518
The phase formation of ferromagnetic L10-MnAl (τ-phase) was accelerated by annealing in magnetic field. With in-field annealing at 623 K, the magnetization of the τ-phase annealed at 15 T was a maximum of 72.2 A･m2/kg at 1.5 T, which is over 4 times larger than that annealed without an external magnetic field for the same annealing time. On the other hand, the transformation from τ-phase to β-phase was suppressed under magnetic field. The obtained bulk sample did not show magnetic anisotropy because of the ε-τ solid-state transformation. Obtained results suggested that magnetic field induced acceleration of transformation was due to the gain of Zeeman energy, which increases the driving force of the ε-τ transformation.
Mn-based ferromagnetic alloys have received attention as magnetic functional materials because they exhibit a high Curie temperature TC, large magnetic moment m, and uniaxial magnetic anisotropy Ku1,2). To date, there have been many reports of Mn-based ferromagnetic materials, including MnBi and MnAlGe compound, Mn-Ga and Mn-Al alloys1–5). Since Kono found that Mn-Al alloy exhibit ferromagnetism4), there have been a lot of reports for improving the magnetic properties of ferromagnetic Mn-Al phase (τ-phase). Koch found that the deformation increased the coercivity field μ0Hc of the τ-phase and was accompanied by a decrease in magnetization6). The magnetic properties, such as TC, m, Ku, and anisotropy field μ0Ha of τ-phase were reported to be TC ∼ 653 K, m = 2.4 μB/f.u, Ku ∼ 1.5 J/m3, and μ0Ha ∼ 5 T, respectively6–8). The τ-phase with a tetragonal CuAu-type (L10) structure is commonly formed by isothermal annealing from a quenched ε-phase with hcp structure or cooling from the high temperature ε-phase6,7,9). The ε-phase was reported to show antiferromagnetism with Néel temperature TN ∼ 90 K10). According to the Al-Mn phase diagram, the τ-phase is not the equilibrium phase11,12). Thus, the single τ-phase is difficult to obtain because of the co-existence of equilibrium non-ferromagnetic β- and γ-phases.
To improve the magnetic properties of Mn-Al, various preparation methods have been performed, such as spark plasma sintering13,14), melt spinning, and mechanical milling15). As it is known that the addition of carbon stabilizes the τ-phase, the magnetic properties of MnAlC magnet have been evaluated9,13–15). However, the addition of carbon decreased the Curie temperature of the τ-phase from 653 to 558 K7).
On the other hand, a magnetic field H stabilizes the ferromagnetic phase because the gain of large Zeeman energy. For example, the equilibrium phase diagrams of Fe-C and Bi-Mn systems were drastically changed by the application of H16–19). The large difference in Zeeman energies between the ferromagnetic phase and non-ferromagnetic phase was largely responsible for the change in phase equilibrium. In addition, the solidification, crystallization, and the reaction process for the materials were influenced by the application of a magnetic field20–22). In-field heat treatment (IFHT) effects on the magnetic materials have been studied for the coercivity field Hc of a Nd-Fe-B permanent magnet, magnetostriction of a TbFe2-based alloy, segregation of Sn in iron, etc23–25). In particular, to control the phase fraction and morphology, IFHT is effective for Mn-based ferromagnetic alloys because of the large magnetic moment of Mn and the large uniaxial magnetic anisotropy of the ferromagnetic phase. In addition, constituent elements of ferromagnetic Mn-based alloys are not ferromagnetic state, so a large difference in Zeeman energy is expected between the synthesized ferromagnetic phase and the raw materials. Previously, we found that the reaction from Mn and Bi to ferromagnetic MnBi was drastically enhanced by in-magnetic-field reactive sintering26,27). Furthermore, the synthesized bulk MnBi exhibits an almost completely uniaxial crystal orientation26–28). On the other hand, IFHT for Mn-Al film also has been carried out by Fischer et al.29) The magnetization for the sample annealed at 1.5 T was 4 times larger than that annealed at 0 T29). In addition, in our previous study, magnetic field effects (MFE) on the ε-τ transformation of bulk Mn-Al alloys were examined with a short annealing time of 4 h30). It was found that the magnetic field also influenced the ε-τ transformation when the annealing temperature was just below TC. For further understanding of MFE on the phase transformation in Mn-Al alloys, it is required to investigate the influence of in-field annealing on the phase formations of Mn-Al in detail.
In this paper, the phase transformations from single ε-phase to τ-, β-, γ-phase under magnetic fields were evaluated. Furthermore, for investigating magnetic field effects on τ−β, γ transformation, IFHT for as-melt sample, which initially includes ε-, τ-, β- and γ-phase, are carried out.
An Al-55at.%Mn rod with 10 mm in diameter was prepared by induction melting, following casting into a copper mold. The rod was cut into disks with 2 mm thick. Subsequently, the disks were sealed in a quartz tube with Ar gas. Heat treatment at 1373 K for 1 day was performed, and the samples were then quenched by breaking the quartz tube in ice water. Figure 1 shows the bulk X-ray diffraction pattern for the quenched sample (a) and the magnetization curves for the sample at room temperature (RT) (b). The obtained disks were confirmed to be single ε-phase with an hcp structure. The magnetization curve shows an almost linear relationship and small magnetization (M ≤ 1.1 Am2kg−1 at 2 T), indicating the paramagnetism of the ε-phase. This result is consistent with the reported magnetic properties of the ε-phase10).
Bulk XRD pattern (a) and magnetization curve (b) at room temperature for the sample before IFHT.
The obtained disks and as-melt samples were sealed in the quartz tube with Ar gas again, and IFHT was carried out at magnetic fields up to 15 T. The details of the electric furnace used at high magnetic fields are described in Refs. 17) and 31). A magnetic field Hh was applied parallel to the thickness of the disks. In order to prevent the ferromagnetic sample from moving in the quartz tube in magnetic field, the sample was held by quartz wool. The annealing temperature Ta of the single ε-phase samples was selected to be 573 K or 623 K, and Ta of the as-melt samples was selected to be 623 K. In order to evaluate the MFE on the phase formation of τ-phase, Ta were selected to be below the TC of the τ-phase to obtain a large Zeeman energy. The annealing time t was up to 96 h for Ta = 573 K and 168 h for Ta = 623 K.
The obtained phases in the sample were characterized by bulk XRD measurements. The scattering vector of the X-ray was parallel to the thickness of the disks. The magnetic properties were obtained by using a vibrating sample magnetometer at magnetic fields up to 1.5 T at RT.
Figure 2 presents the typical bulk XRD patterns for the samples annealed at 573 K (a) and 623 K (b). All diffraction peaks can be indexed by the ε-phase with hcp structure, β-phase with cubic structure, γ-phase with trigonal structure, and τ-phase with tetragonal L10 structure, which indicated that there are no other intermediate phases. With increasing t, the fraction of ε-phase decreased due to the formation of τ−, γ, and β-phase. At 623 K, it seemed that the appearance of τ-phase was faster than Ta = 573 K. The coexistence of the β-phase was observed for all samples at 623 K. It should be noted that the diffraction peaks of the coexisting β-phase annealed at Ta = 623 K under a magnetic field were weaker than that in a zero field. These results suggest that the in-field heat treatment suppressed the co-formation of the equilibrium β-phase during the ε-τ transformation.
Bulk XRD patterns for the samples for Ta = 573 K (a) and 623 K (b). Diffraction intensity was normalized by the highest peaks in each pattern.
Figure 3 shows the magnetization curves for Ta = 573 K. For t = 24 h, although the magnetization curve shows a small magnetization and the difference between the IFHT sample and zero-field one was small, the IFHT sample shows ferromagnetism. At t = 48 h, the sample annealed in a zero field shows paramagnetism, indicating it consists almost of ε-phases. Meanwhile, the IFHT sample shows more noticeable magnetic properties, i.e., Ms = 14.2 A･m2/kg at 1.5 T, which was about 5 times larger than the obtained magnetization of Ms = 2.7 A･m2/kg in a zero field.
Magnetization curves for the samples for Ta = 573 K. The annealed conditions are μ0Hh = 0 T for 24 h (a), μ0Hh = 15 T for 24 h (b), μ0Hh = 0 T for 48 h (c), and μ0Hh = 15 T for 48 h (d).
Figure 4 indicates the magnetization curves for Ta = 623 K. All samples, both IFHT and zero-field, show ferromagnetic properties because of the formation of the τ-phase. The difference between the IFHT sample and that annealed in a zero field was observed more clearly than at Ta = 573 K. The maximum values of the magnetic properties were Ms = 72.2 A･m2/kg and Hc = 0.15 T.
Magnetization curves for the samples for Ta = 623 K. The annealed conditions are μ0Hh = 0 T for 24 h (a), μ0Hh = 15 T for 24 h (b), μ0Hh = 0 T for 48 h (c), and μ0Hh = 15 T for 48 h (d).
IFHT effects for the magnetic anisotropy of the bulk sample were evaluated. Figure 5 presents the typical magnetization curves for the sample annealed at 10 T. H⊥ and H// indicate the applied magnetic field direction perpendicular and parallel to H, respectively. The magnetization curves were almost the same for H⊥ and H//, which indicated that a uniaxial oriented phase formation was not observed. This behavior was observed for all in-field samples. IFHT effects for the ε-τ transformation were not the same as that for the reaction of a Bi-Mn system26).
Typical magnetization curves at room temperature for the sample annealed at 10 T. H// and H⊥ indicate the measuring direction of the magnetic field, parallel and perpendicular to Hh, respectively. The solid and broken curves are the magnetization curve of H// and H⊥, respectively. The inset indicates the relationship between H//, H⊥, Hh, and the sample.
Figure 6 indicates the annealing time t dependence of the saturation magnetization Ms at Ta = 573 K (a) and 623 K (b). In this study, Ms was determined as the magnetization at the maximum magnetic fields. For the annealing time t ≥ 48 h at 573 K, enhancement of Ms by IFHT at 15 T was observed. However, it was found that the IFHT at 573 K in 10 T did not influence the magnetic properties. On the other hand, even at μ0H = 10 T, magnetization Ms was drastically enhanced compared with that in a zero field at Ta = 623 K, which was different behavior from that at Ta = 573 K. Moreover, the magnetization of sample for t = 168 h in 15 T was slightly smaller than that for t = 48 h in 15 T. Because τ-phase is only ferromagnetic phase in Mn-Al system alloy, decrease of Ms is due to the transformation from ferromagnetic τ-phase to other non-ferromagnetic equilibrium phase, such as β-phase. The obtained result indicated that τ-phase did not become equilibrium phase even applying 15 T. On the other hand, Ms still increased at t = 168 h in a zero field, and is as high as that for 48 h in 15 T, indicating the same fraction of τ-phase for t = 48 h in 15 T, 168 h in 15 T, and 168 h in 0 T. Thus, it is indicated that the magnetic field accelerated ε−τ transformation. Figure 7 shows the XRD patterns for the IFHT sample at t = 168 h in 0 T and 15 T. Herein, it is noted that the phase fractions of the phases were difficult to determine in detail because the XRD measurements were carried out using bulk samples. However, most diffraction peaks of ε−phase disappeared in both patterns, indicating small fraction of residual ε−phase. Therefore, the almost ε−phase transformed to τ− and β− phase for both patterns. Considering the slight decrease in Ms for 48 ≤ t ≤ 168 h at 15 T, phase formation of β-phase, including τ−β transformation, become dominantly as residual ε-phase vanished.
The annealing time dependence of the magnetization of the samples at 1.5 T. The triangles, and circles indicated the Ms of the sample annealed at each magnetic field, respectively.
Bulk XRD patterns for the samples annealed at 623 K for 168 h in a zero field and 15 T. Diffraction intensity was normalized by the highest peaks of each patterns.
Figure 8 shows the coercivity field Hc as a function of annealing time. With increasing t, Hc increased for both Ta = 573 and 623 K. The increase in Hc was related to the formation of the τ-phase. Thus, at Ta = 623 K, μ0Hc saturated to be 0.15 T, regardless of IFHT or zero-field annealing. This value is as high as the previous report by ageing for a quenched strain-free ε-phase by Crew et al9).
The coercivity field Hc as a function of annealing time at Ta = 573 K (a) and 623 K (b). The squares, triangles, and circles indicated the μ0Hc of the sample annealed at each magnetic field, respectively.
In order to investigate MFE on the τ-β, γ transformation, we performed IFHT for as-melt samples. As as-melt sample initially contained ε−, τ−, β−, and γ-phase, as-melt sample is suitable to observe the proceeds of τ−β or τ−γ transformation in shorter annealing time than that with single ε-phase.
Figure 9 shows the bulk XRD patterns of the as-melt samples. The diffraction peaks of the τ-phase in the sample annealed for 12 h in 15 T were same as that in 0 T. The peaks of the β-phase annealed in 15 T were smaller than that in 0 T for both annealing times, suggesting the suppression of the phase formation of β-phase by IFHT. Herein, it is difficult to determine whether the β-phase was formed by τ−β or ε−β transformation.
Bulk XRD patterns for as-melt samples for Ta = 623 K. Diffraction intensity was normalized by the highest peak of each patterns.
Figure 10 shows the annealing time t dependence of the magnetization of the as-melt samples. The Ms for t = 12 h in 15 T was same as that in 0 T. With increasing t, the Ms decreased for the sample annealed in a zero field, which was due to the transformation from τ-phase to β-phase. This indicates that the IFHT suppressed the τ-β transformation. On the other hand, Ms slightly increased for the IFHT sample with increasing t. Thus, it is also indicated the suppression of the τ-β transformation.
The annealing time dependence of Ms of the as-melt samples. The triangles, and circles indicated the MsT of the sample annealed at each magnetic field, respectively.
The magnetic field effects for phase transformation in Mn-Al are discussed with respect to the thermodynamics. The magnetic contribution for the free energy of ε-, β-, and τ-phase Mn55Al45 under a magnetic field is expressed by
| \[ E_{\rm M} = \int_{0}^{B} M(T,H) \quad {\rm d}H, \] | (1) |
Figure 11 shows the calculated thermomagnetization curves (a) and the temperature dependence of EM at 10 and 15 T (b). The EM values at 573 K and 623 K, evaluated from the curves, were 97 and 70 J/mol, respectively, at 15 T and 63 and 44 J/mol, respectively, at 10 T.
Thermomagnetization curves obtained by mean-field calculation (a) and the Zeeman energy EM as a function of temperature at 10 and 15 T (b).
Figure 12 shows a schematic diagram of the energy states of the phases at annealing temperatures of 623 K (a) and 573 K (b). In a zero field, the schematic image of Gε, Gγ, Gβ, and Gτ, which are the free energies of the quenched ε-phase, equilibrium γ- and β-phase, and the metastable τ-phase, respectively, are shown in solid curves. At both temperatures, with an applied magnetic field, the gain in EM for the ferromagnetic τ-phase is much larger than those for non-ferromagnetic β- and γ-phases, resulting in a large decrease in Gτ, shown by the broken curves. As a result, the dip in the Gτ under magnetic fields brings it close to the common tangent for Gβ and Gγ. In other words, the τ-phase becomes close to the equilibrium phase under magnetic fields. Therefore, the driving force of ε-τ and τ-β transformation is enhanced and decreased by IFHT, respectively, resulting in acceleration of the ε-τ phase transformation and suppression of the τ-β phase transformation. Furthermore, based on the magnetic contribution on G, magnetic field is ineffective for ε-β, γ transformation because ε−, β− and γ−phase do not have large M. On the other hand, although EM for the τ-phase at 573 K was larger than that at 623 K, Gβ and Gγ become more stable than those at 623 K. Therefore, the gain in the Zeeman energy becomes ineffective for enhancement of the ε-τ transformation. In addition, thermal activation of the transformation decreased because of the low Ta, and the transformation becomes sluggish. For these reasons, in-field annealing-induced enhancement of the ε-τ phase transformation was smaller than that for Ta = 623 K. It is noted that the gain in the Zeeman energy is less than 0.1 kJ/mol. This value is quite small and may only have a small effect on the phase equilibrium. Thus, the τ-phase is metastable but probably close to the equilibrium.
Schematic diagrams of the energy states of the phases at Ta = 623 K (a) and 573 K (b). Gε, Gγ, Gβ, and Gτ are the free energies of the ε-, γ-, β-, and τ-phases, respectively. The solid curve indicates the free energy of the phases at a zero field. The broken curve is the Gτ in magnetic fields. The solid line indicates the common tangent of Gβ and Gγ.
Compared Fig. 6 with Fig. 10, the behaviors of Ms seems to be different each other. As seen in Fig. 6, decrease in Ms appeared for IFHT sample, whereas monotonically increase in Ms was observed for the sample annealed in 0 T. Meanwhile, IFHT sample showed the increase in Ms for as-melt one. The annealing of as-melt sample in 0 T led the decrease in Ms for 12 ≤ t ≤ 24 h. These differences were explained by the residual fraction of ε-phase. Based on the above transformation model of Mn-Al alloys under magnetic fields, the kinetics of the in-field transformation for single ε-phase and as-melt sample at 623 K is discussed. In this model, magnetic-field-induced stabilization of τ-phase led the acceleration of ε−τ transformation and suppression of τ−β transformation. As seen in Fig. 6, Ms in 15 T increased for t ≤ 48 h and decreased for 48 ≤ t ≤ 168 h, whereas Ms in 0 T increased for t ≤ 168 h. Here, as seen in Fig. 2 (b), small fraction of ε-phase was remained at 48 h in 15 T. Residual ε-phase was almost vanished for 48 ≤ t ≤ 168 h, and τ-phase was no longer synthesized. Thus, the decrease of Ms under 15 T was due to the disappearance of ε-phase and following τ−β transformation. Meanwhile, the ε−τ transformation in 0 T was slower than 15 T, Ms increased for 48 ≤ t ≤ 168 h, due to ε-τ transformation. According to the XRD pattern (see Fig. 7), although the fraction of β-phase seemed to be larger than that in 15 T, residual ε-phase was also larger than 15 T. As a result, increase of Ms due to ε-τ transformation was more effective than decrease of Ms due to τ-β transformation even t = 168 h. As ε- and β-phase did not show ferromagnetism, ε-β transformation did not affected Ms. On the other hand, as-melt sample contained ε−, τ−, and β-phase at initial condition. As seen in Fig. 9, the fraction of residual ε-phase was small, suggesting the slight occurrence of ε−τ transformation in following heat treatment. Ms in 15 T did not decreased for t ≤ 24 h, whereas Ms in 0 T decreased for 12 ≤ t ≤ 24 h. Monotonically increase of Ms for IFHT sample was due to the ε−τ transformation and the suppression of τ−β transformation. Meanwhile in 0 T, decrease of Ms was due to the decrease of τ-phase. As residual ε-phase was small and the τ-phase was enough large that τ-β transformation became dominantly for 12 ≤ t ≤ 24 h, leading the decrease of Ms in 0 T.
In this study, no crystal orientation was induced by IFHT for Mn-Al alloys. This was a result of the ε-τ solid-phase transformation. The ε-τ transformation was reported to be a massive transformation33). During the massive transformation, ε-phase with disordered hcp structure became ordered hcp structure, following the hcp-L10 transformation. The relationship of the axis between the ε- and τ-phase is shown in Fig. 13. When there is an ε-τ transformation, the c-plane of the ε-phase with hcp structure is compressed and a L10 structure is formed. Therefore, the axis of the formed L10 structure was determined by that of the ε-phase. Consequently the solid-state ε-τ transformation caused the no IFHT-induced crystal orientation.
The relationship between the ε-phase with a hcp structure (broken lines) and the τ-phase with a L10 structure (solid lines).
In-magnetic-field heat treatment (IFHT) effects for the phase formation of τ-phase Mn-Al were investigated. The drastic improvement of the ferromagnetic properties was observed by IFHT because of the acceleration and the suppression of the ε-τ and τ-β phase transformation, respectively. These IFHT effects on the selected phase formation of τ-phase can be explained by the stabilization of the ferromagnetic τ-phase by contribution of EM for the free energy under magnetic fields.
This work was partly supported by KAKENHI 26820281, 16H04547, and 16K14374. In-field heat treatment and magnetization measurements were carried out at the High Field Laboratory for Superconducting Materials (HFLSM), Institute for Materials Research (IMR), Tohoku University, and Cooperative Research and Development Center for Advanced Materials (CRDAM), IMR, Tohoku University, respectively. The authors are deeply grateful to Prof. K. Takanashi from IMR for permission to use the equipment of CRDAM.
