Coercivity and Microstructure Dependence on B Content in ThMn₁₂-Type Sm-Zr-Fe-Co-Ti-Cu-B Alloys

Abstract

The effect of B content was studied on the crystalline structures, magnetic properties, and microstructures of (Sm_0.80Zr_0.20)_8.6(Fe_0.71Co_0.20Ti_0.08Cu_0.01)_91.4−xB_x (x = 0.8, 2.0, 5.0, and 8.0) melt-spun alloys with a ThMn₁₂-type structure (1–12 phase) to achieve high saturation magnetization (μ₀M_s) and large intrinsic coercivity (H_cj). With increasing x, the amorphous forming ability improved in the as-quenched alloys, and the average crystallite size decreased in the annealed alloys. The x = 5.0 alloy annealed at 1173 K for 60 min exhibited the largest H_cj of 454 kA/m and the μ₀M_s was estimated to be 1.29 T. In addition, the x = 5.0 alloy exhibited demagnetization curve without a two-step shape, which indicates a low volume fraction of soft magnetic phases. The scanning transmission electron microscopy revealed that 1–12 phase grains have sizes in the range of 30–100 nm in the x = 5.0 alloy after annealing at 1173 K. Fine Ti-B precipitates were also observed. The reduction in the 1–12 phase grain size is attributed to the suppression of crystal nucleation and the crystal growth caused by the amorphous formation in the as-quenched alloy, and the grain boundary pinning by the Ti-B inhibiting the grain growth. The grain sizes in the x = 5.0 alloy were smaller than the estimated critical single-domain diameter of 190 nm, which contributes to the increase in H_cj.

1. Introduction

In recent years, the demand for high-efficiency motors and generators to reduce CO₂ emissions has increased. Consequently, the demand for manufacturing of Nd-Fe-B magnets has also increased as these are primarily used in electric devices. However, scarcity of resources poses a risk due to the uneven distribution of production areas for rare-earth elements such as Nd, Dy, and Tb, which are used in magnets. Therefore, high-performance magnets with a lower rare-earth element content are crucial.

From the late 1980s to the 1990s, RFe_12−xT_x (R = rare-earth elements, T = stabilizing elements such as Ti, V, Cr, Mo, and W) compounds with a ThMn₁₂-type structure (1–12 phase) garnered considerable attention as candidates for next-generation permanent magnets owing to their high magnetocrystalline anisotropies and high curie temperatures [1–5]. However, achieving bulk magnets with a large maximum energy product is challenging since the stabilizing elements required to form the 1–12 phase degrade the saturation magnetization (μ₀M_s). This issue has resulted in declining interest in these compounds.

Recently, the formation of the 1–12 phase without stabilizing elements has been successfully achieved in thin films. For instance, Sm(Fe_0.8Co_0.2)₁₂ thin film [6] exhibited a μ₀M_s of 1.78 T and an anisotropy field of 8.2 T. Furthermore, Sm(Fe_0.8Co_0.2)₁₂B_0.5 thin film [7] exhibited a high μ₀M_s of 1.5 T and large intrinsic coercivity (H_cj) of 1.2 T, which is attributed to B-rich grain boundary phases. These magnetic properties, which are superior to or comparable with those of the Nd₂Fe₁₄B compound, have stimulated a renewed interest in RFe_12−xT_x compounds and their study. Nevertheless, such properties have still not been achieved in the bulk materials.

To improve μ₀M_s in the bulk, reduction of Ti content by Zr substitution for rare-earth elements has been studied [8–12]. The (Sm_0.8Zr_0.2)_1.1(Fe_0.9Co_0.1)_11.3Ti_0.7 strip-casted alloys [13] exhibit a relatively high μ₀M_s of 1.38–1.48 T. However, its maximum H_cj is 434 kA/m, which is insufficient to achieve an alternative material to Nd-Fe-B magnets.

In contrast, H_cj improvement in various compounds has been achieved through melt-spinning method that can reduce the grain size [14–18]. There are mainly two types of alloys that exhibit large H_cj. The first is alloys based on Sm-Fe-V [19–22]. The H_cj for Sm-Fe-V based alloys is 637–876 kA/m; however, their magnetizations are low because a large amount of V is required to form the 1–12 phase. The second is Sm-Fe-Ti-B alloys. SmFe_10.8Ti_1.2B_0.2 [23] and Sm_1.3Fe_10.4TiB_0.3 [24] alloys exhibit H_cj of approximately 478 and 613 kA/m, respectively. However, the effect of B on H_cj and microstructures in the alloys has not been extensively discussed in the literature, unlike the research conducted on B-doped thin film [6]. Therefore, clarifying the dependence of H_cj and microstructures on B content can lead to H_cj enhancement in the bulk. Furthermore, their low magnetization can be improved by reducing the Ti content. In addition, precipitation of α-Fe that can reduces the H_cj has been observed in the Sm_1.3Fe_10.4TiB_0.3 alloy [24]. The H_cj can be further improved by suppression of the α-Fe precipitation as reported by Dirba et al. [25] where the α-Fe content was reduced in the Cu-doped Sm(Fe_0.8Co_0.2)Ti alloy.

In this study, the effect of B doping on Sm-Zr-Fe-Co-Ti-Cu-based compounds was investigated. The study focused on the dependence of H_cj and microstructures on B content in the melt-spun alloys to achieve both high μ₀M_s and large H_cj.

2. Experimental Procedure

The (Sm_0.80Zr_0.20)_8.6(Fe_0.71Co_0.20Ti_0.08Cu_0.01)_91.4−xB_x (x = 0.8, 2.0, 5.0, and 8.0) alloys were prepared using a single roller melt-spinning method. The rotational speed of the Cu roller was set at 20 m/s. The melt-spun flakes have a thickness of approximately 60 µm and a width of approximately 8 mm. The alloys were annealed at 973–1173 K for 60 min in an Ar atmosphere. Compositional analyses of the alloys were performed using inductively coupled plasma atomic emission spectroscopy (ICP-AES). Phase identification was conducted using X-ray diffraction (XRD) analysis (Bruker D8 Advance) with Cu-Kα radiation. All the XRD patterns were measured on the alloy surface. The magnetic properties were measured using a vibrating sample magnetometer with high-temperature superconducting magnets at room temperature under a maximum applied field of 4.7 MA/m. Specimens for scanning transmission electron microscopy (STEM) observations were extracted from the center of the flakes in the thickness direction by focused-ion beam nanofabrication using an FEI Versa 3D Dual Beam. STEM observations of the microstructure were performed using an FEI Talos F200X. Elemental mapping analyses were conducted using energy-dispersive X-ray spectrometry (EDX) with an FEI Super-X equipped with windowless Si drift detectors capable of detecting B.

3. Results and Discussions

The compositions of the melt-spun alloys were quantified from the ICP-AES analysis, as summarized in Table 1. All the compositions were successfully controlled to (Sm_0.80Zr_0.20)_8.6(Fe_0.71Co_0.20Ti_0.08Cu_0.01)_91.4−xB_x.

Table 1 Chemical compositions of the melt-spun alloys.

Figures 1(a) and (b) show the XRD patterns of the as-quenched and annealed alloys, respectively. In the as-quenched alloys, the 1–12 phase was observed at x = 0.8. The XRD peaks of the 1–12 phase weakened as x increased and no distinct peaks were observed for 5.0 ≤ x. This decrease in peak intensity indicates an enhancement in the ability to form an amorphous phase with higher B content. After annealing at 1173 K, the XRD peaks of the 1–12 phase, bcc (e.g., α-Fe), and MgZn₂-type structures (e.g., Fe₂Ti) were observed. The 1–12 phase peaks broadened with increasing x. The 1–12 phase peaks of all the samples did not show any discernible crystal orientation, which indicates that the alloys exhibit magnetic isotropy.

Fig. 1

XRD patterns of (Sm_0.8Zr_0.2)_8.6(Fe_0.7Co_0.2Ti_0.08Cu_0.01)_91.4−xB_x alloys (x = 0.8, 2.0, 5.0 and 8.0): (a) as-quenched and (b) annealed at 1173 K for 60 min. (online color)

The average crystallite size of the 1–12 phase after annealing as functions of x, was investigated to gauge the relationship between grain size and B content, as shown in Fig. 2. The crystallite sizes were determined from the full width at half maximum of the 1–12 phase peaks using the Scherrer equation. The sizes are underestimated compared to the grains observed in the TEM images mentioned later. The crystallite size decreased as x increased at each annealing temperature. Furthermore, at x = 0.8, the crystallite size increased with increasing annealing temperature while at higher values of x, the size reduced even at high temperatures. This suggests the suppression of the crystal growth of the 1–12 phase with an increase in B content.

Fig. 2

Average crystallite size of the 1–12 phase annealed at 973–1173 K for 60 min as functions of x. (online color)

Figure 3 shows H_cj as functions of x. The values of H_cj were maximized at x = 5.0 for all annealing temperatures. In particular, H_cj increased from 127 kA/m at x = 0.8 to 454 kA/m with annealing at 1173 K.

Fig. 3

H_cj of the alloys annealed at 973–1173 K for 60 min as functions of x. (online color)

Figure 4 compares the magnetization curves of the x = 0.8 alloy—characterized by a large crystallite size and small H_cj—with those of the x = 5.0 alloy, featuring a small crystallite size and large H_cj after annealing at 1173 K. The magnetizations were not saturated, as shown in the inset. The values of μ₀M_s were estimated from the intercept of μ₀M vs. 1/H² plot, based on the law of approach to saturation described as eq. (1).

  

\begin{equation} \mu_{0}M = \mu_{0}M_{\text{s}}(1-a/H^{2}), \end{equation} (1)

where a is a constant and H is the applied field. The estimated μ₀M_s were 1.34 and 1.29 T for the x = 0.8 and x = 5.0 alloys, respectively. In addition, the x = 5.0 alloy demonstrated a demagnetization curve without a two-step shape near the zero field, indicating a low volume fraction of soft magnetic phases.

Fig. 4

Magnetization curves of the x = 0.8 (dashed line) and x = 5.0 (solid line) alloys, annealed at 1173 K for 60 min. The inset shows the first quadrant with the horizontal axis extended to the maximum applied field of 4.7 MA/m. (online color)

Microstructures of the x = 0.8 and 5.0 alloys annealed at 1173 K were observed to understand the reasons for the increase in H_cj at x = 5.0. Figure 5 shows bright-field STEM (BF-STEM) and EDX mapping images for the x = 0.8 alloy magnified at 68 k. Figure 6 shows the corresponding images for the x = 5.0 alloy magnified at 270 k. In the x = 0.8 alloy, the 1–12 phase grains with size of 100–1000 nm were observed along with sparse grains of Ti-B, Sm-Cu, and Ti-Zr coexisting with Fe and Co. The XRD peaks of MgZn₂-type structure shown in Fig. 1(b) may correspond to the Ti-Zr grains with Fe and Co since the formation of MgZn₂-type structure is not associated with Ti-B and Sm-Cu compounds [26, 27] unlike Ti-Zr-Fe-Co alloys [28].

Fig. 5

(a) BF-STEM and (b)–(h) EDX mapping images captured from the x = 0.8 alloy annealed at 1173 K for 60 min. (online color)

Fig. 6

(a) BF-STEM and (b)–(i) EDX mapping images captured from the x = 5.0 alloy annealed at 1173 K for 60 min. (online color)

Conversely, it was found that the 1–12 phase grain size in the x = 5.0 alloy is 30–100 nm, which is smaller than that in the x = 0.8 alloy. This decrease in the grain size with increasing B content is consistent with the decrease in the average crystallite size of the 1–12 phase shown in Fig. 2. Additionally, numerous Ti-B grains surrounded by Zr were observed from overlapping EDX mapping image of Zr and Ti shown in Fig. 6(i). The Ti-B precipitation is attributed to increased B content, which has a large negative mixing enthalpy with Ti and Zr [29]. The majority of Ti-B grains were found to be in the range of 10–50 nm in size, which is smaller than that for the x = 0.8 alloy having size in the range of 50–100 nm. The reduction in grain size of the 1–12 phase and the Ti-B precipitates is attributed to the suppression of crystal nucleation and grain growth during the annealing, due to amorphous phase formation in the as-quenched state. Furthermore, grain boundary pinning by the Ti-B grains may contribute to the inhibition of grain growth of the 1–12 phase.

The bcc structural phases were not observed in the STEM and EDX images despite the occurrence of bcc peaks in the XRD patterns for the x = 0.8 and x = 5.0 alloys. In addition, demagnetization curve of the x = 5.0 alloy indicated a low volume fraction of soft magnetic phase, as shown in Fig. 3. These results imply that the α-Fe precipitation occurs only on the alloy surface, possibly induced by Sm evaporation during annealing [9, 21].

Assuming that the 1–12 phase grains are spherical with Bloch domain walls, the critical single-domain diameter (D_c) can be estimated using eq. (2).

  

\begin{equation} D_{\text{c}} = 36(AK_{\text{u}})^{1/2}/\mu_{0}M_{\text{s}}{}^{2}, \end{equation} (2)

where μ₀ is the magnetic constant, A is the exchange stiffness constant, and K_u is the uniaxial magnetic anisotropy constant. Using A = 9.9 pJ/m and K_u = 4.8 MJ/m³ for SmFe₁₁Ti [30, 31], along with M_s = 1.03 MA/m obtained from the estimated μ₀M_s = 1.29 T for the x = 5.0 alloy, its D_c was estimated to be 190 nm. The grain sizes smaller than D_c contribute to the increase of H_cj. In the case of the x = 8.0 alloy, although the grain size of the 1–12 phase is estimated to be smaller than that of the x = 5.0 alloy from the decrease in average crystallite size as shown in Fig. 2, the H_cj decreased as shown in Fig. 3. The H_cj decrease at x = 8.0 can be attributed to decline of crystallinity in the 1–12 phase by excessive reduction in the grain size. The decline in crystallinity in the x = 8.0 alloy is also suggested by the weaker and broader XRD peaks compared to the x = 5.0 alloy.

No grain boundary phases were observed in the x = 5.0 alloy. The formation of non-magnetic grain boundary phases is expected to further improve in the H_cj. In the future, improvement of the magnetic properties in sintered magnets is necessary to realize the commercially available alternatives to conventional Nd-Fe-B magnets. In future studies, the inhibition of grain growth in the 1–12 phase by Ti-B precipitation will be utilized in the sintered magnets.

4. Conclusion

(Sm_0.80Zr_0.20)_8.6(Fe_0.71Co_0.20Ti_0.08Cu_0.01)_91.4−xB_x (x = 0.8, 2.0, 5.0, and 8.0) melt-spun alloys were investigated to achieve both high μ₀M_s and large H_cj. As x increased, the ability to form an amorphous phase improved in the as-quenched alloys, while the average crystallite size decreased in the annealed alloys. The x = 5.0 alloy annealed at 1173 K exhibited the largest H_cj of 454 kA/m and the μ₀M_s in this condition was estimated to be 1.29 T. In addition, the x = 5.0 alloy demonstrated a demagnetization curve without two-step shape suggesting a low volume fraction of soft magnetic phases. The results of microstructural observations revealed that the sizes of the 1–12 phase grains are in the range of 100–1000 nm in the x = 0.8 alloy and in the range of 30–100 nm in the x = 5.0 alloy, after annealing at 1173 K. Furthermore, numerous fine Ti-B precipitates were also observed. The reduction in the 1–12 phase grain size is attributed to the suppression of crystal nucleation and the crystal growth caused by amorphous formation in the as-quenched alloy, and the grain boundary pinning by Ti-B, which inhibits the grain growth. The grain sizes in the x = 5.0 alloy were smaller than estimated D_c of 190 nm, contributing to an increase in H_cj.

Acknowledgments

We would like to thank the Sugimoto Laboratory, Department of Material Science, Graduate School of Engineering, Tohoku University, for their cooperation in the measurement of the magnetic properties.

REFERENCES

• 1)   K.  Ohashi,  Y.  Tawara,  R.  Osugi and  M.  Shimao: Magnetic properties of Fe-rich rare-earth intermetallic compounds with a ThMn₁₂ structure, J. Appl. Phys. 64 (1988) 5714–5716. doi:10.1063/1.342235
• 2)   K.  Ohashi,  T.  Yokoyama,  R.  Osugi and  Y.  Tawara: The magnetic and structural properties of R-Ti-Fe ternary compounds, IEEE Trans. Magn. 23 (1987) 3101–3103. doi:10.1109/TMAG.1987.1065256
• 3)   D.B.  De Mooij and  K.H.J.  Buschow: Some novel ternary ThMn₁₂-type compounds, J. Less-Common Met. 136 (1988) 207–215. doi:10.1016/0022-5088(88)90424-9
• 4)   K.H.J.  Buschow: Permanent magnet materials based on tetragonal rare earth compounds of the type RFe_12−xM_x, J. Magn. Magn. Mater. 100 (1991) 79–89. doi:10.1016/0304-8853(91)90813-P
• 5)   G.C.  Hadjipanayis,  S.H.  Aly and  S.-F.  Cheng: Hard magnetic properties of R-Fe-Ti alloys, Appl. Phys. Lett. 51 (1987) 2048–2050. doi:10.1063/1.98288
• 6)   Y.  Hirayama,  Y.K.  Takahashi,  S.  Hirosawa and  K.  Hono: Intrinsic hard magnetic properties of Sm(Fe_1−xCo_x)₁₂ compound with the ThMn₁₂ structure, Scr. Mater. 138 (2017) 62–65. doi:10.1016/j.scriptamat.2017.05.029
• 7)   H.  Sepehri-Amin,  Y.  Tamazawa,  M.  Kambayashi,  G.  Saito,  Y.K.  Takahashi,  D.  Ogawa,  T.  Ohkubo,  S.  Hirosawa,  M.  Doi,  T.  Shima and  K.  Hono: Achievement of high coercivity in Sm(Fe_0.8Co_0.2)₁₂ anisotropic magnetic thin film by boron doping, Acta Mater. 194 (2020) 337–342. doi:10.1016/j.actamat.2020.05.026
• 8)   S.  Suzuki,  T.  Kuno,  K.  Urushibata,  K.  Kobayashi,  N.  Sakuma,  K.  Washio,  M.  Yano,  A.  Kato and  A.  Manabe: A new magnet material with ThMn₁₂ structure: (Nd_1−xZr_x)(Fe_1−yCo_y)_11+zTi_1−zN_α (α = 0.6–1.3), J. Magn. Magn. Mater. 401 (2016) 259–268. doi:10.1016/j.jmmm.2015.10.042
• 9)   K.  Kobayashi,  S.  Suzuki,  T.  Kuno,  K.  Urushibata,  N.  Sakuma,  M.  Yano,  T.  Shouji,  A.  Kato and  A.  Manabe: The stability of newly developed (R,Zr)(Fe,Co)_12−xTi_x alloys for permanent magnets, J. Alloy. Compd. 694 (2017) 914–920. doi:10.1016/j.jallcom.2016.09.311
• 10)   T.  Kuno,  S.  Suzuki,  K.  Urushibata,  K.  Kobayashi,  N.  Sakuma,  M.  Yano,  A.  Kato and  A.  Manabe: (Sm,Zr)(Fe,Co)_11.0-11.5Ti_1.0-0.5 compounds as new permanent magnet materials, AIP Adv. 6 (2016) 025221. doi:10.1063/1.4943051
• 11)   P.  Tozman,  H.  Sepehri-Amin,  Y.K.  Takahashi,  S.  Hirosawa and  K.  Hono: Intrinsic magnetic properties of Sm(Fe_1−xCo_x)₁₁Ti and Zr-substituted Sm_1−yZr_y(Fe_0.8Co_0.2)_11.5Ti_0.5 compounds with ThMn₁₂ structure toward the development of permanent magnets, Acta Mater. 153 (2018) 354–363. doi:10.1016/j.actamat.2018.05.008
• 12)   A.G.  Popov,  A.V.  Protasov,  V.S.  Gaviko,  D.A.  Kolodkin,  P.B.  Terentev,  E.G.  Gerasimov,  T.  Zhang and  C.  Jiang: Magnetic properties of melt-spun ribbons (Sm_1−xZr_x)(Fe_0.92Ti_0.08)₁₀ with ThMn₁₂ structure and their hydrides, J. Rare Earths 37 (2019) 1066–1071. doi:10.1016/j.jre.2019.04.007
• 13)   T.  Kuno,  T.  Yamamoto,  K.  Urushibata,  K.  Kobayashi and  S.  Sugimoto: Preparation of High-Coercivity Magnetic Powder via Heat Treatment of a Rapidly Quenched Amorphous Starting Compound with a ThMn₁₂ Structure, Mater. Trans. 61 (2020) 657–662. doi:10.2320/matertrans.MT-M2019356
• 14)   S.  Sugimoto,  A.  Kojima,  M.  Okada and  M.  Homma: Enhancement of Magnetic Properties of Sm(Fe, Co, Ti)₁₂ Melt-Spun Ribbons by Refining Crystallized Grains, Mater. Trans. JIM 32 (1991) 1180–1183. doi:10.2320/matertrans1989.32.1180
• 15)   F.E.  Pinkerton and  D.J.  Van Wingerden: Magnetic hardening of SmFe/sub 10/V/sub 2/ by melt-spinning, IEEE Trans. Magn. 25 (1989) 3306–3308. doi:10.1109/20.42285
• 16)   H.  Sun,  Y.  Otani,  J.M.D.  Coey,  C.D.  Meekison and  J.P.  Jakubovics: Coercivity and microstructure of melt-spun Sm(Fe₁₁Ti), J. Appl. Phys. 67 (1990) 4659–4661. doi:10.1063/1.344845
• 17)   L.  Schultz and  J.  Wecker: Coercivity in ThMn₁₂ -type magnets, J. Appl. Phys. 64 (1988) 5711–5713. doi:10.1063/1.342234
• 18)   Y.Z.  Wang and  G.C.  Hadjipanayis: Magnetic properties of Sm-Fe-Ti-V alloys, J. Magn. Magn. Mater. 87 (1990) 375–378. doi:10.1016/0304-8853(90)90774-K
• 19)   M.  Okada,  A.  Kojima,  K.  Yamagishi and  M.  Homma: High coercivity in melt-spun SmFe/sub 10/(Ti,M)/sub 2/ ribbons (M=V/Cr/Mn/Mo), IEEE Trans. Magn. 26 (1990) 1376–1378. doi:10.1109/20.104383
• 20)   J.  Ding and  M.  Rosenberg: Magnetic hardening of melt-spun and crystallized Sm-Fe-V and Sm-(Fe, Co)-V alloys, J. Less-Common Met. 161 (1990) 369–374. doi:10.1016/0022-5088(90)90049-P
• 21)   Y.  Wang,  G.C.  Hadjipanayis,  A.  Kim,  N.C.  Liu and  D.J.  Sellmyer: Magnetic and structural studies in Sm-Fe-Ti magnets, J. Appl. Phys. 67 (1990) 4954–4956. doi:10.1063/1.344745
• 22)   X.  Tang,  J.  Li,  A.K.  Srinithi,  H.  Sepehri-Amin,  T.  Ohkubo and  K.  Hono: Role of V on the coercivity of SmFe₁₂-based melt-spun ribbons revealed by machine learning and microstructure characterizations, Scr. Mater. 200 (2021) 113925. doi:10.1016/j.scriptamat.2021.113925
• 23)   M.S.  Anagnostou and  D.  Niarchos: Large coercivities i melt-spun Sm-Fe-Ti type magnets, J. Magn. Magn. Mater. 88 (1990) 100–104. doi:10.1016/S0304-8853(97)90018-4
• 24)   E.W.  Singleton,  J.  Strzeszewski and  G.C.  Hadjipanayis: High coercivity in rapidly quenched Sm(Fe, T)₁₂-type magnets, Appl. Phys. Lett. 54 (1989) 1934–1936. doi:10.1063/1.101501
• 25)   I.  Dirba,  Y.  Harashima,  H.  Sepehri-Amin,  T.  Ohkubo,  T.  Miyake,  S.  Hirosawa and  K.  Hono: Thermal decomposition of ThMn₁₂-type phase and its optimum stabilizing elements in SmFe₁₂-based alloys, J. Alloy. Compd. 813 (2020) 152224. doi:10.1016/j.jallcom.2019.152224
• 26)   H.  Okamoto: Cu-Sm (copper-samarium), J. Phase Equilib. 19 (1998) 183. doi:10.1361/105497198770342698
• 27)   H.  Okamoto: B-Ti (Boron-Titanium), J. Phase Equilibria Diffus. 27 (2006) 303. doi:10.1361/154770306X109953
• 28)   S.  Miura,  S.  Takizawa,  K.  Ohkubo and  T.  Mohri: An Approach for a Broad Search of New Compound Using the First Principles Calculation; AB Type Binary and Quaternary Intermetallic-Alloy, J. Japan Inst. Metals 71 (2007) 860–868. doi:10.2320/jinstmet.71.860
• 29)   A.  Takeuchi and  A.  Inoue: Calculations of Mixing Enthalpy and Mismatch Entropy for Ternary Amorphous Alloys, Mater. Trans. JIM 41 (2000) 1372–1378. doi:10.2320/matertrans1989.41.1372
• 30)   T.  Fukazawa,  H.  Akai,  Y.  Harashima and  T.  Miyake: Spin-wave dispersion and exchange stiffness in Nd₂Fe₁₄B and RFe₁₁Ti (R = Y, Nd, Sm) from first-principles calculations, Phys. Rev. B 103 (2021) 024418. doi:10.1103/PhysRevB.103.024418
• 31)   B.-P.  Hu,  H.-S.  Li,  J.P.  Gavigan and  J.M.D.  Coey: Intrinsic magnetic properties of the iron-rich ThMn₁₂-structure alloys R(Fe₁₁Ti); R=Y, Nd, Sm, Gd, Tb, Dy, Ho, Er, Tm and Lu, J. Phys. Condens. Matter 1 (1989) 755–770. doi:10.1088/0953-8984/1/4/009