Coercivity Enhancement of Sintered Nd–Fe–B Magnets by Intergranular Adding Micro-Structured Dy–Nd–Pr–Al–Cu Powder

Abstract

In this work, microstructure and magnetic properties of sintered Nd_16.5Fe₇₇B_6.5 magnets with intergranular addition of Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀ were investigated. The additional particles with size in the range of 1–2 µm were homogeneously mixed into micrometer Nd_16.5Fe₇₇B_6.5 master powder with different weight fractions. The investigation shows that coercivity H_c of the added magnets significantly increases, while their remanence B_r decreases with increasing Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀ additional fraction. The coercivity of the magnets increases from 14.5 kOe to 31.5 kOe with increasing the additional fraction from 0 to 9 wt.%. By adding Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀ compound to grain boundary, the Dy concentration can be reduced by 60 wt.% in comparison with the conventional method.

Fig. 4 The dependences of coercivity H_c, remanence B_r and maximum energy product (BH)_max of the added magnets on various fractions of Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀.

1. Introduction

For many applications in practice, such as traction motors of hybrid electric vehicles and wind power generators, sintered Nd–Fe–B magnets with high coercivity, H_c > 25 kOe, are required.^1,2) Because the operating temperature of these devices is high (∼200°C), the coercivity of the magnets rapidly decreases by the thermal demagnetization, resulting in a strong reduction of the power of the machines. Generally, the high coercivity of sintered Nd–Fe–B magnets can be created by partially replacing heavy rare earth Dy for Nd in lattice of the Nd₂Fe₁₄B phase (2:14:1) due to the high magnetocrystalline anisotropy field H_A of Dy₂Fe₁₄B, which is larger than that of Nd₂Fe₁₄B.^3–5) However, Dy has a high price and is scarce. Thus, the enhancement of the coercivity of the sintered Nd–Fe–B magnets without using or using small content of Dy is more and more concerned to study. Previous works^6–13) indicated that the high coercivity sintered Nd–Fe–B magnets could be prepared by the addition of non-ferromagnetic compounds, such as Pr–Cu, Cu–Al, Dy–Nd–Al and Nd–Cu–Al to grain boundary. This way not only creates high coercivity but also reduces content of heavy rare earth for the magnets. Each element in the additional compounds differently affects on structure and magnetic properties of the magnets.^14,15) During sintering and annealing processes, Dy diffuses from the boundaries to the 2:14:1 grains to form (Nd,Dy)₂Fe₁₄B shell with high magnetocrystalline anisotropy.^16,17) In addition, thermal stability, grain size uniformity of the 2:14:1 main phase of the magnets can be improved by Dy addition.^18,19) As for Pr, it not only has more natural abundance than Dy but also contributes to a higher H_A of Pr₂Fe₁₄B than that of Nd₂Fe₁₄B.^4,20,21) As well known that, the coercivity of the Nd–Fe–B magnets is strongly sensitive to their microstructure.²²⁾ Meanwhile, microstructure of the magnets, such as distribution and chemical composition of grain boundary phase, uniformity of the particles, wettability between 2:14:1 phase and intergranular phase, can be improved by adding Cu and Al.¹⁴⁾ Furthermore, Cu and Al can also improve the corrosion resistance of the added magnets because of their relatively high electrode potential.²³⁾

In our previous studies,^24,25) the influence of addition of nanoparticles on the magnetic properties of the sintered Nd–Fe–B magnets was investigated. The coercivity above 20 kOe could be achieved by adding nanoparticles of Dy₄₀Nd₃₀Al₃₀ and Nd₄₀Cu₃₀Al₃₀ to grain boundaries. In general, the nanoparticles are probably distributed on the surface of Nd₂Fe₁₄B hard magnetic grains more homogeneously than microparticles. However, the cost to produce the nanoparticles with large amount is high. Furthermore, fabrication conditions are strict because the nanoparticles are easily oxidized during the fabrication process (milling, mixing and pressing), leading to a decrease of the magnetic properties of the magnets. Therefore, in this work the coercivity of sintered Nd_16.5Fe₇₇B_6.5 magnets is enhanced by the addition of Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀ microparticles. The microstructure and magnetic properties of the added magnets are investigated and discussed in details.

2. Experimental Procedure

Alloys with nominal composition of Nd_16.5Fe₇₇B_6.5 and Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀ were prepared by using an induction furnace for about 30 minutes under Ar gas to avoid oxidation. Based on the previous research results of our group and other authors,^24–26,35) the additional compound with a chemical composition of Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀ is selected to easily fabricate alloy powder with small particle size and strongly enhance coercivity of the sintered Nd–Fe–B magnets. The resulted ingots were crashed into small pieces with their size smaller than 2 centimeters and coarsely milled before jet-milling to obtain fine particles. The average particle size of Nd_16.5Fe₇₇B_6.5 and Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀ powder was about 3 µm and 1 µm, respectively. These two kinds of powder were mixed together with fraction of Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀ ranging from 1 to 9 wt.% under the protection of Ar gas. The mixing time of powder mixture was 1 h. The mixed powder was compressed, with a pressure of 15 MPa, into rectangular parallelepiped tablets (7 × 5 × 3 cm³) in oriented magnetic field of about 20 kOe yielded by an electromagnet. The pressed magnet tablets were then sintered at 1080°C for 1 h in vacuum and subjected subsequently to two-stage heat treatment at 820°C for 1 h and at 540°C for 1 h.^27,28) For both the stages, the magnets were fast cooled down to room temperature by Ar gas. The structure of the materials was thoroughly analyzed by scanning electron microscope (SEM) and energy dispersive X-ray (EDX) techniques. Based on SEM images, grain size distribution was also determined by ImageJ software. In order to investigate magnetic properties of the sintered magnets on a pulsed high field magnetometer with maximum magnetic field of 90 kOe, the cylinders with 3 mm diameter and 3 mm height were cut from the magnets by means of spark-cutting. Demagnetization factor was estimated based on a semi-experimental data sheet to determine the maximum energy product (BH)_max of the magnets.

3. Results and Discussion

SEM images and the grain size distribution of the Nd_16.5Fe₇₇B_6.5 and Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀ powder are shown in Fig. 1. ImageJ software was used to calculate the grain size distribution by measuring a lot of grains from the SEM image. We can see that the grain size of the Nd_16.5Fe₇₇B_6.5 powder is in range of 1–5 µm with maximum distribution at 2.5 µm. Compared to our previous studies,^24–26) the grains size in this work, which were prepared by jet-milling method, are more uniform than those obtained from ball milling process in industrial white gasoline. As for the Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀ powder, the grain size is in range of 0.2–2 µm with maximum distribution at 1 µm. The size of additional particles should be smaller than that of the master ones to obtain their homogeneous distribution in grain boundaries.

Fig. 1

SEM images and grain size distribution of Nd_16.5Fe₇₇B_6.5 (a), (b) and Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀ (c), (d) powder.

Hysteresis loops of un-added and added (without and with additional Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀ powder) magnets after the heat treatment are presented in Fig. 2. We can see that, the coercivity of the added magnets is significantly enhanced with increasing the additional fractions. From the hysteresis loops, magnetic characteristic curves of the magnets can be deduced. Figure 3 shows the magnetic characteristic curves of the typical magnets added with 0 and 7 wt.% of Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀. Notably, the squareness of the hysteresis loops of the added magnets is better in comparison to that presented in our previous studies.^24,26) This can be explained by the improvement of microstructure of the added magnets such as uniform grains and smooth grain boundaries.

Fig. 2

Hysteresis loops of the magnets added with various fractions of Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀ powder.

Fig. 3

Magnetic characteristic curves of the un-added (a) and 7 wt.%-added (b) magnets.

The dependences of the coercivity H_c, remanence B_r and maximum energy product (BH)_max of the magnets on various weight fractions of the Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀ microparticles are represented in Fig. 4. Compared to the starting magnet with H_c = 14.5 kOe, B_r = 13.5 kG and (BH)_max = 41.5 MGOe, all the Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀-added magnets exhibit a strong enhancement of H_c and a small reduction of B_r and (BH)_max. We can see that, the coercivity H_c do not linearly depend on the fraction of Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀. The H_c increases rapidly when the fraction of the additional compound is low (in range from 0% to 3%) and tends to increase slowly with high fraction (above 3%). This variation is different from the common way (almost linearly) when Dy is partially substituted for Nd in the lattice of the entire magnet. In this work, the substitution of Dy for Nd is based on the addition of Dy-containing compound to the grain boundary of the magnets and the diffusion of the additional elements does not linearly depend on the fraction of the additional compound. The decrease of the saturation magnetization and remanence is due to the anti-ferromagnetic coupling between Fe and Dy. On the other hand, the reduction of fraction of the 2:14:1 ferromagnetic phase due to the addition of the non-ferromagnetic phase at grain boundary is another reason for the decrease of B_r. By increasing the fraction of Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀ to 7 wt.%, the H_c is enhanced to 30.5 kOe while B_r and (BH)_max are reduced to 11.4 kG and 30.8 MGOe, respectively. Further increase of fraction of Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀ to 9 wt.% makes H_c increase to 31.5 kOe, but reduces (BH)_max to 28.3 MGOe. It implies that an appropriate adding fraction of Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀ should be selected in order to obtain both the high coercivity and large enough maximum energy product for the magnets. Thus, the fraction of Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀ of 7 wt.% can be considered as an optimal one.

Fig. 4

The dependences of coercivity H_c, remanence B_r and maximum energy product (BH)_max of the added magnets on various fractions of Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀.

In order to understand the mechanism of the coercivity enhancement by the addition of Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀, SEM images and EDX spectroscopies of the magnets were performed. The SEM images of the magnets added with 0 (un-added) and 7 wt.% of Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀ are shown in Fig. 5. The obtained results reveal that the grain size of the magnet after sintering and heat treatment, which is in range of 6–10 µm, is larger than that of the as-milled powder. This can be explained as follows. During the sintering process, the grain surfaces and small grains are melted to form the liquid phase. This liquid phase binds and envelops the crystalline grains when it is solidified, leading to the increase of the grain size. According to our obtained results, the magnets with the optimal magnetic properties are created from powder with the grain size smaller than 5 µm. For the un-added magnets, the grain boundaries are not clear (blear boundary lines) and some grains of the 2:14:1 ferromagnetic phase are directly connected to each other (Fig. 5(a)). After adding 7 wt.% of Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀, the grain boundaries are finer and sharper, leading to a decrease of exchange interaction of the 2:14:1 grains (Fig. 5(b)). As reported in the previous publications,^18,24,29) the intergranular region of the added magnets mainly consists of rare earth-rich non-ferromagnetic phases after the heat treatment process. The formation of a rare earth-rich phase at the grain boundary makes Nd₂Fe₁₄B grains isolated and prevents the magnetic interaction between them. When the 2:14:1 grains are well isolated, the magnetic reversal process is difficult to propagate from one to another. Instead, the magnetization reversal needs a nucleation of a reverse domain in each grain. As a result, this process requires a larger magnetic field (higher coercivity). The uniformity of the 2:14:1 grains probably is reason for improving squareness of the hysteresis loops of the added magnets. While, the homogeneous distribution of the additional particles at grain boundaries can prevent the formation of nucleation center of reversal domains (the soft magnetic α-Fe phase) causing higher demagnetization filed of the added magnets. When adding to grain boundary, a part of Dy and Pr atoms exists in the matrix of grain boundary phase (Fig. 5(c)) and the other diffuses into grain lattice (Fig. 5(d)). However, Dy and Pr were not detected at the grain center (Fig. 5(e)). That means, Dy and Pr only diffuse into grain outer. Consequently, magnetocrystalline anisotropy of the outer layer is higher than that of the interior of the grains. This leads to a stronger inhibition of the formation and propagation of a reverse domain in the added grains than that in the un-added ones. Then strength of external magnetic field must be large enough for the formation and growth of reverse nucleations, meaning that the added magnets have high coercivity. Thus, the diffusion of Dy and Pr created a high anisotropic shell for the 2:14:1 grains. This shell is sufficient to prevent the formation and development of reverse nucleations that normally occur at the grain boundary. The heavy rare earth atoms in the core of the particle are believed to have little effect on the coercivity, while reducing the saturation magnetization M_s. The concentrated distribution of the heavy rare earth atoms at the grain shell strongly increases H_c but not much reduces M_s, i.e. not much reduces (BH)_max, in comparison with the whole grain distribution case. This is also a reason for reducing concentration of the heavy rare earth for the magnets. Unlike Dy and Pr, the non-ferromagnetic elements of Cu and Al are mostly excluded in the 2:14:1 phase and located at the intergranular regions (Fig. 5(c)), leading to an improvement of the grain boundaries for increasing the coercivity of the magnets. None of Cu or Al were founded by EDX spectroscopy at both the outer and center of the 2:14:1 grains (Fig. 5(d), Fig. 5(e)).

Fig. 5

SEM images of the un-added (a) 7 wt.%-added (b) magnets, and EDX spectroscopies at boundary (c), outer (d) and center (e) of grain of the 7 wt.%-added magnet.

Magnetic parameters of the added magnets with various fractions of Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀ and corresponding content of Dy are determined and listed in Table 1. We can see that the highest additional fraction of 9 wt.% of Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀ corresponds to about 4.6 wt.% of heavy rare earth element of Dy. This content of Dy is quite low in comparison with that of the commercial sintered (Nd, Dy)–Fe–B magnets (up to 15 wt.%). Although the content of Dy is low, the magnetic parameters of our added magnets are high enough to fulfill the requirement in the high temperature application fields such as motors, generators.

Table 1 Magnetic parameters of the magnets added with various fractions of Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀ and corresponding content of Dy.

A comparison of obtained magnetic parameters in this work with others is presented in Table 2. With the same additional method (grain boundary addition - GBA), the coercivity of our magnet is much higher than that of magnets added with other Dy-containing compounds. Although the maximum energy product (BH)_max of our magnet lower than that of others but it is still above 30 MGOe. With grain boundary diffusion (GBD) method,^37–41) the concentration of heavy rare earth elements is reduced to a low level (only 0.9%) and still creates a magnet with high coercivity (over 25 kOe). However, this technology is only applied to enhance the coercivity for bulk magnets with small size (requirement for diffusion process). This limits practical applications of magnets. As mentioned in the introduction section, for applications in high temperature environment such as motors and generators, the coercivity of the magnets should be higher than 25 kOe. Thus, our added magnets are more applicable in practice. Furthermore, the heavy rare earth (Dy) content in our magnet is much lower than the magnet with the same H_c fabricated by conventional method, i.e. replacing Nd by Dy in the lattice. For example, to have H_c to reach above 30 kOe, the Dy concentration in commercial magnets is usually about 10 wt.%, which is equivalent to Nd₁₀Dy₄Fe₈₀B₆ composition. As for our method, in order to get such H_c for the magnet, the Dy content only needs 3.6 wt.%, equivalent to the additional fraction of 7 wt.% of Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀. Thus, the Dy content has been reduced by about 60 wt.%, while the magnet still has the similar hard magnetic parameters. This result has an important significance in reducing the cost of the high coercivity sintered Nd–Fe–B magnets. Besides, it also reduces the need for increasingly scarce heavy rare earth and limits the exploitation of rare earth causing environmental pollution.

Table 2 Magnetic properties of the sintered Nd–Fe–B magnets un-added and added with various Dy-containing compounds.

4. Conclusion

The influence of concentration of the additional microparticles of Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀ on microstructure and magnetic properties of the sintered Nd_16.5Fe₇₇B_6.5 magnets has been investigated. The coercivity H_c is considerably enhanced by adding the microparticles to grain boundaries of the magnets. The strong increase of the coercivity is suggested due to the diffusion of Dy and Pr atoms to the Nd₂Fe₁₄B grains to form the phases with high magnetocrystalline anisotropy and the improvement of the grain boundaries by Cu and Al. The optimal H_c of 30.5 kOe to retain high enough (BH)_max of 30.8 MGOe is achieved for the magnets contained 3.6 wt.% of Dy. This work suggests that the intergranular addition of Dy₃₀Nd₁₇Pr₃Al₄₀Cu₁₀ is promising to manufacture high coercivity sintered Nd–Fe–B magnets with using low content of the heavy rare earth elements.

Acknowledgments

This research is funded by the National Foundation for Science and Technology Development (NAFOSTED) of Viet Nam under grant number of 103.02-2018.339. A part of the work was done in the Key Laboratory for Electronic Materials and Devices, Institute of Materials Science, Viet Nam.

REFERENCES

• 1)   O.  Gutfleisch,  M.A.  Willard,  E.  Bruck,  C.H.  Chen,  S.G.  Sankar and  J.L.  Liu: Adv. Mater. 23 (2011) 821–842. doi:10.1002/adma.201002180
• 2)   K.  Hono and  H.  Sepehri-Amin: Scr. Mater. 67 (2012) 530–535. doi:10.1016/j.scriptamat.2012.06.038
• 3)   M.H.  Ghandehari: Appl. Phys. Lett. 48 (1986) 548–550. doi:10.1063/1.96503
• 4)   S.  Hirosawa,  Y.  Matsuura,  H.  Yamamoto,  S.  Fujimura,  M.  Sagawa and  H.  Yamauchi: J. Appl. Phys. 59 (1986) 873–879. doi:10.1063/1.336611
• 5)   E.  Boltich,  E.  Oswald,  M.  Huang,  S.  Hirosawa,  W.  Wallace and  E.  Burzo: J. Appl. Phys. 57 (1985) 4106–4108. doi:10.1063/1.334633
• 6)   Z.  Zhang,  J.  Jin,  T.  Ma,  L.  Liang and  M.  Yan: J. Magn. Magn. Mater. 498 (2020) 166162. doi:10.1016/j.jmmm.2019.166162
• 7)   L.  Liang,  T.  Ma,  C.  Wu,  P.  Zhang,  X.  Liu and  M.  Yan: J. Magn. Magn. Mater. 397 (2016) 139–144. doi:10.1016/j.jmmm.2015.08.091
• 8)   Y.L.  Liu,  J.  Liang,  Y.C.  He,  Y.F.  Li,  G.F.  Wang,  Q.  Ma,  F.  Liu,  Y.  Zhang and  X.F.  Zhang: AIP Adv. 8 (2018) 056227. doi:10.1063/1.5008766
• 9)   Y.  Wang,  Y.  Luo,  Z.  Wang,  G.  Wu,  J.  Xie,  W.  Yan and  D.  Yu: J. Magn. Magn. Mater. 458 (2018) 85–89. doi:10.1016/j.jmmm.2018.02.082
• 10)   M.  Tang,  X.  Bao,  Y.  Zhou,  K.  Lu,  J.  Li and  X.  Gao: J. Magn. Magn. Mater. 505 (2020) 166749. doi:10.1016/j.jmmm.2020.166749
• 11)   W.Q.  Liu,  H.  Sun,  X.F.  Yi,  X.C.  Liu,  D.T.  Zhang,  M.  Yue and  J.X.  Zhang: J. Alloy. Compd. 501 (2010) 67–69. doi:10.1016/j.jallcom.2010.04.030
• 12)   W.Q.  Liu,  C.  Li,  M.  Zakotnik,  M.  Yue,  D.T.  Zhang and  X.L.  Huang: J. Rare Eart. 33 (2015) 846–849. doi:10.1016/S1002-0721(14)60494-4
• 13)   T.  Ma,  X.  Wang,  C.  Gao,  X.  Liu,  L.  Liang,  C.  Wu and  M.  Yan: Mater. Express 6 (2016) 93–99. doi:10.1166/mex.2016.1275
• 14)   S.  Pandian,  V.  Chandrasekaran,  G.  Markandeyulu,  K.J.L.  Iyer and  K.V.S.  Rao Rama: J. Appl. Phys. 92 (2002) 6082–6086. doi:10.1063/1.1513879
• 15)   S.  Pandian,  V.  Chandrasekaran,  G.  Markandeyulu,  K.J.L.  Iyer and  K.V.S.  Rao Rama: J. Alloy. Compd. 364 (2004) 295–303. doi:10.1016/S0925-8388(03)00541-3
• 16)   N.J.  Yu,  M.X.  Pan,  P.Y.  Zhang and  H.L.  Ge: J. Magn. 18 (2013) 235–239. doi:10.4283/JMAG.2013.18.3.235
• 17)   W.F.  Li,  H.  Sepehri-Amin,  T.  Ohkubo,  N.  Hase and  K.  Hono: Acta Mater. 59 (2011) 3061–3069. doi:10.1016/j.actamat.2011.01.046
• 18)   L.  Liang,  T.  Ma,  P.  Zhang,  J.  Jin and  M.  Yan: J. Magn. Magn. Mater. 355 (2014) 131–135. doi:10.1016/j.jmmm.2013.11.007
• 19)   X.F.  Zhang,  S.  Guo,  C.J.  Yan,  L.W.  Cai,  R.J.  Chen,  D.  Lee and  A.R.  Yan: J. Appl. Phys. 115 (2014) 17A757. doi:10.1063/1.4868491
• 20)   B.C.  Chen,  X.M.  Liu,  R.J.  Chen,  S.  Guo,  D.  Lee and  A.R.  Yan: J. Alloy. Compd. 516 (2012) 73–77. doi:10.1016/j.jallcom.2011.11.132
• 21)   Y.M.  Rabinovich,  S.  Szymura,  J.J.  Wysocki and  H.  Bala: J. Magn. Magn. Mater. 140–144 (1995) 1065–1066. doi:10.1016/0304-8853(94)01259-8
• 22)   U.M.R.  Seelam,  T.  Ohkubo,  T.  Abe,  S.  Hirosawa and  K.  Hono: J. Alloy. Compd. 617 (2014) 884–892. doi:10.1016/j.jallcom.2014.07.166
• 23)   J.J.  Ni,  T.Y.  Ma,  Y.R.  Wu and  M.  Yan: J. Magn. Magn. Mater. 322 (2010) 3710–3713. doi:10.1016/j.jmmm.2010.07.031
• 24)   N.H.  Dan,  P.T.  Thanh,  N.H.  Yen and  L.T.  Hung: IEEE Trans. Magn. 50 (2014) 210204. doi:10.1109/TMAG.2014.2301294
• 25)   P.T.  Thanh,  N.V.  Duong,  N.H.  Yen,  N.H.  Ngoc,  N.M.  Lam,  K.X.  Hau,  S.C.  Yu and  N.H.  Dan: Curr. Appl. Phys. 18 (2018) 329–334. doi:10.1016/j.cap.2017.12.014
• 26)   P.T.  Thanh,  D.T.K.  Oanh,  N.H.  Yen and  N.H.  Dan: J. Nanosci. Nanotechnol. 21 (2021) 2558–2562. doi:10.1166/jnn.2021.19100
• 27)   V.P.  Menushenkov and  A.G.  Savchenko: J. Magn. Magn. Mater. 258–259 (2003) 558–560. doi:10.1016/S0304-8853(02)01144-7
• 28)   P.T.  Thanh,  N.H.  Yen and  N.H.  Dan: Vietnam J. Sci. Technol. 52 (2014) 465–472. doi:10.15625/0866-708X/52/4/3151
• 29)   Y.  Zhang,  T.  Ma,  X.  Liu,  P.  Liu,  J.  Jin,  J.  Zou and  M.  Yan: J. Magn. Magn. Mater. 399 (2016) 159–163. doi:10.1016/j.jmmm.2015.09.071
• 30)   B.  Zhou,  X.  Li,  X.  Cao,  G.  Yan and  A.  Yan: Chin. Phys. B 25 (2016) 117504. doi:10.1088/1674-1056/25/11/117504
• 31)   X.  Liang and  G.  Yan: Wuhan Univ. J. Nat. Sci. 21 (2016) 339–343. doi:10.1007/s11859-016-1179-0
• 32)   X.L.  Liu,  X.J.  Wang and  L.P.  Liang: J. Magn. Magn. Mater. 370 (2014) 76–80. doi:10.1016/j.jmmm.2014.06.051
• 33)   X.  Li,  S.  Liu,  X.  Cao,  B.  Zhou,  L.  Chen,  A.  Yan and  G.  Yan: J. Magn. Magn. Mater. 407 (2016) 247–251. doi:10.1016/j.jmmm.2016.01.085
• 34)   B.  Zhou,  X.  Li,  X.  Liang,  G.  Yan,  K.  Chen and  A.  Yan: J. Magn. Magn. Mater. 429 (2017) 257–262. doi:10.1016/j.jmmm.2017.01.025
• 35)   Y.  Itou,  Y.  Fukui,  T.  Kitayama,  Y.  Iwama,  T.  Kobayashi and  Y.  Kato: J. Japan Inst. Met. Mater. 59 (1995) 103–107. doi:10.2320/jinstmet1952.59.1_103
• 36)   A.M.  Gabay,  M.  Marinescu,  W.F.  Li,  J.F.  Liu and  G.C.  Hadjipanayis: J. Appl. Phys. 109 (2011) 083916. doi:10.1063/1.3574444
• 37)   F.  Li,  J.  Li,  S.U.  Rehman,  L.  Zhang,  Y.  Hu,  M.  Yang,  X.  Yu,  S.  Zhong and  T.  Liang: J. Alloy. Compd. 899 (2022) 163270. doi:10.1016/j.jallcom.2021.163270
• 38)   Z.  Wang,  T.T.  Sasaki,  Y.  Une,  T.  Ohkubo and  K.  Hono: Acta Mater. 248 (2023) 118774. doi:10.1016/j.actamat.2023.118774
• 39)   D.  Zhao,  F.  Liu,  Y.  Gao,  P.  Jiang,  L.  Liu,  M.  Zhao,  S.  Ren and  W.  Pei: J. Magn. Magn. Mater. 563 (2022) 169943. doi:10.1016/j.jmmm.2022.169943
• 40)   M.  Uenohara,  Z.  Hanlin,  H.  Nishio and  K.  Machida: Mater. Trans. 61 (2020) 782–786. doi:10.2320/matertrans.MT-MBW2019006
• 41)   M.  Uenohara,  H.  Nishio,  K.  Toyoki,  Y.  Shiratsuchi,  R.  Nakatani and  K.  Machida: Mater. Trans. 62 (2021) 1216–1224. doi:10.2320/matertrans.MT-M2020389