Ce-based R2Fe14B (R= rare-earth) nano-structured permanent magnets consisting of (Ce,Nd)2Fe14B core-shell grains separated by a non-magnetic grain boundary phase, in which the relative amount of Nd to Ce is higher in the shell of the magnetic grain than in its core, were fabricated by Nd-Cu infiltration into (Ce,Nd)2Fe14B hot-deformed magnets. The coercivity values of infiltrated core-shell structured magnets are superior to those of as-hot-deformed magnets with the same overall Nd content. This is attributed to the higher value of magnetocrystalline anisotropy of the shell phase in the core-shell structured infiltrated magnets compared to the homogeneous R2Fe14B grains of the as-hot-deformed magnets, and to magnetic isolation of R2Fe14B grains by the infiltrated grain boundary phase. First order reversal curve (FORC) diagrams suggest that the higher anisotropy shell suppresses initial magnetization reversal at the edges and corners of the R2Fe14B grains.

High performance neodymium-based permanent magnets are key components of motors in electrical and hybrid vehicles. These magnets now contain Dy, the Dy partially substituting Nd in the Nd2Fe14B phase, to have enough coercivity (Hc) so as to maintain their remanent magnetization at the motor’s maximum operating temperature of about 473K. However Dy, and to a lesser extent Nd, are considered to be critical elements due to rapidly increasing demand for permanent magnets in motors. From this point of view, Ce-based intermetallic magnets are attracting attention1 because of the greater abundance of Ce compared to Nd and Dy. However, the intrinsic magnetic properties, namely spontaneous magnetization (Js) and anisotropy field (Ha), of Ce2Fe14B are far inferior to those of Nd2Fe14B. The room temperature values of Js and Ha of single crystal Ce2Fe14B are 1.18T and 2.40 MA/m, respectively, while those of single crystal Nd2Fe14B are 1.61T and 5.36 MA/m.2 To use Ce, the partial substitution of another R for Ce in the rare-earth sites of Ce2Fe14B, and an optimization of the microstructure are required to improve extrinsic properties, in particular coercivity.

This study concerns Ce-based nano-structured permanent magnets which consist of core-shell R2Fe14B grains separated by a non-magnetic grain boundary phase. The Ce-based 2-14-1 cores are surrounded by a higher anisotropy shell in which Ce is partially replaced by Nd. The concept of forming core-shell grains, via grain boundary diffusion with Nd-Cu, is expected to improve coercivity due to a suppression of magnetization reversal thanks to the high anisotropy shell and improved magnetic isolation thanks to the non-magnetic Nd-Cu grain boundary phase.3 Here we report on the structural and magnetic properties of such magnets.

The precursor materials were rapidly quenched Ce13.8-XNdXFe75.66Co4.46B5.66Ga0.38 (X is from 0 to 10.35) melt-spun ribbons. The melt-spun ribbons were crushed to a few hundred microns and then sintered at 1020 K under a pressure of 100 MPa. The sintered bulk magnets were hot-deformed to a height reduction of about 80% to develop (001), i.e. easy magnetization axis, texture of the (Ce,Nd)2Fe14B grains. The degree of texture, defined as the ratio of remanence to saturation magnetization, was about 0.9. The hot-deformed magnets were infiltrated with liquid Nd70Cu30 alloy at 970 K for 165 minutes, the amount of eutectic corresponding to 10 or 40 wt% of the magnet. During infiltration, Nd-Cu diffuses between grains and forms high anisotropy shells at the grain surface by partial substitution of Nd for Ce. The as-hot-deformed magnets are hereafter referred to as “alloy magnets” while the infiltrated magnets are referred to as “core-shell magnets”.

Microstructure observations and composition analyses have been performed using scanning transmission electron microscopy (STEM) and energy dispersive X-ray (EDX) spectroscopy, respectively. Magnetization measurements were made on 2 × 2 × 2 mm3 samples in the temperature range 300 K to 473 K, using a vibrating sample magnetometer (VSM), following sample magnetization in a field of 10 T. First order reversal curve (FORC) measurements was made on core-shell magnets following saturation, decreasing the magnetic field from 2.4 MA/m and sweeping it back to 2.4 MA/m in steps of 0.04 MA/m.

An STEM image and EDX elemental maps of a Ce2Fe14B hot-deformed magnet following infiltration with Nd-Cu are shown in Figure 1. Comparing the Ce and Nd maps, we can clearly identify the Nd-rich shell formed on the periphery of the Ce2Fe14B grains (white circled areas in Fig. 1). The compositions of a series of core-shell magnets (samples 1-4) and alloy magnets (samples 5-7), analyzed by STEM-EDX line profiles, are listed in Table I.

Comparison of the room temperature magnetization curves (Fig. 2) of a Ce2Fe14B alloy magnet (sample 5) and a core-shell magnet made from such a magnet (sample 1) reveals that infiltration with 10 wt% Nd-Cu increases coercivity from 0.02 to 0.41 MA/m. The relationship between coercivity and the mean Nd/ (Nd+ Ce) composition of core-shell magnets (samples 1 - 3) and alloy magnets (samples 5-7) is shown in Figure 3. The composition of samples 1-3 and 5-7 were obtained from STEM-EDX line profiles. The coercivity values of all core-shell magnets are superior to those of alloy magnets with the same overall Nd content in the R2Fe14B grains. The temperature dependence of coercivity of a core-shell magnet (sample 2) is compared to that of a Nd2Fe14B hot-deformed magnet in figure 4. While the magnets have the same room temperature value of coercivity, the core-shell magnet has higher values at elevated temperatures, despite having 25at% less Nd.

We attribute the coercivity enhancement demonstrated here to the formation of a high anisotropy shell in the Ce based R2Fe14B grains,4 which suppresses magnetization reversal, and to an improvement in the magnetic isolation of these grains by the formation of a grain boundary phase.3 FORC measurements, which may provide information about magnetization reversal behavior (coercive field distribution, magnetic interactions),5 have been performed to experimentally clarify the difference in reversal behavior between alloy and core-shell samples. FORC measurements and the corresponding FORC diagrams of core-shell magnets (sample 1 and 4) made from Ce2Fe14B magnets with infiltration of 10 wt% and 40 wt% Nd-Cu, respectively, are shown in Figure 5 (coercivity of sample 1 and 4 are 0.41 MA/m and 0.67 MA/m respectively). The presence of two distributions of reversal field along the Hc axis of the sample with the lower Nd-Cu content indicates the presence of heterogeneous microstructures with both a high coercivity portion and a low coercivity portion. On the other hand, the reversal field distribution of the sample with the higher Nd-Cu content (narrow along Hc, broad along interaction field (Hu)) is typical of a high coercivity homogeneous system in which the grains have approximately the same coercive field values and the distribution of Hu is essentially equal to the distribution of demagnetizing field.6,7 Altogether, the change in FORC behavior from low Nd-Cu content to high Nd-Cu content characterizes simply the associated increase in coercivity due to shell formation.

Structural and magnetic characterization has been carried out on Ce-based R-Fe-B based core-shell magnets fabricated by infiltration of Nd-Cu into hot-pressed magnets. The increase in coercivity achieved for a given Nd content of the R2Fe14B phase is attributed to two factors. Firstly, the formation of a shell within the R2Fe14B grains having higher magnetocrystalline anisotropy than the core, thanks to the partial substitution of Nd for Ce. Secondly, improved magnetic decoupling of R2Fe14B grains due to the formation of a non-magnetic grain-boundary phase. The homogeneous character of reversal in Nd-richer infiltrated magnets is illustrated by FORC analysis.

This paper is based on results obtained from the “Development of magnetic material technology for high-efficiency motors” program commissioned by the New Energy and Industrial Technology Development Organization (NEDO) of Japan.

1.
J. F.
Herbst
,
M.S.
Meyer
, and
F.E.
Pinkerton
,
J. Appl. Phys.
111
,
07A718
(
2012
).
2.
S.
Hirosawa
,
K.
Tokuhara
,
H.
Yamamoto
,
S.
Fujimura
, and
M.
Sagawa
,
J. Appl. Phys
61
,
3571
(
1987
).
3.
H.
Sepehri-Amin
,
T.
Ohkubo
,
S.
Nagashima
,
M.
Yano
,
T.
Shoji
,
A.
Kato
,
T.
Schrefl
, and
K.
Hono
,
Acta Materialia
61
,
6622
(
2013
).
4.
S.
Bance
,
J.
Fischbacher
, and
T.
Schrefl
,
J. Appl. Phys.
117
,
17A733
(
2015
).
5.
T.
Schrefl
,
T.
Shoji
,
M.
Winklhofer
,
H.
Oezelt
,
M.
Yano
, and
G.
Zimanyi
,
J. Appl. Phys
111
,
07A728
(
2012
).
6.
C. R.
Pike
,
C. A.
Ross
,
R. T.
Scalettar
, and
G.
Zimanyi
,
Phys. Rev. B
71
,
134407
(
2005
).
7.
A. N.
Dobrynin
,
T. R.
Gao
,
N. M.
Dempsey
, and
D.
Givord
,
Appl. Phys. Lett.
97
,
1
(
2010
).