Texture memory effect of Nd-Fe-B during hydrogen treatment

Article abstract of DOI:10.1016/j.jmmm.2004.11.423

Highly textured magnet powder can be produced when Nd₂Fe ₁₄B alloy reacts reversibly with hydrogen at high temperature (1073 K). This treatment holds promise as a production method for high-performance magnets. The appearance of anisotropy or isotropy in the magnet powder is determined by the rate of the reaction between Nd₂Fe₁₄B and hydrogen, which has been formulated in the dynamic-hydrogenation disproportionation desorption recombination (d-HDDR) process. The phenomena of the d-HDDR treatment are summarised in this paper including the review of the anisotropy mechanism termed here as the texture memory effect .

Full text of DOI:10.1016/j.jmmm.2004.11.423

ARTICLE IN PRESS
Journal of Magnetism and Magnetic Materials 290–291 (2005) 1282–1285
www.elsevier.com/locate/jmmm
Texture memory effect of Nd–Fe–Bduring hydrogen treatment
a
Y. Honkura , C. Mishima , N. Hamada , G. Drazic , O. Gutfleisch
a
a
b
c,Ã
a
Aichi Steel Corporation, Arao-machi, Tokai-shi, Aichi-ken 476-8666, Japan
b
Jozef Stefan Institute, Jamova 39, 1001 Ljubljana, Slovenia
IFW Dresden, Institute for Metallic Materials, P.O. Box 270016, D-01171 Dresden, Germany
c
Available online 14 December 2004
Abstract
Highly textured magnet powder can be produced when Nd
2
Fe14Balloy reacts reversibly with hydrogen at high
temperature (1073 K). This treatment holds promise as a production method for high-performance magnets. The
appearance of anisotropy or isotropy in the magnet powder is determined by the rate of the reaction between Nd Fe B
2
14
and hydrogen, which has been formulated in the ‘‘dynamic-hydrogenation disproportionation desorption recombina-
tion’’ (d-HDDR) process. The phenomena of the d-HDDR treatment are summarised in this paper including the review
of the anisotropy mechanism termed here as the ‘‘texture memory effect’’.
r 2004 Elsevier B.V. All rights reserved.
Keywords: Permanent magnet; Nd–Fe–B; HDDR; Texture; Electron microscopy
1
. Introduction
In parallel with magnet development, the inducement
of texture was investigated and various anisotropy
mechanisms were proposed. The suggested memory sites
2
Anisotropic Nd Fe14Bmagnet powder with high
performance is produced when NdFeBalloy and
hydrogen react reversibly at high temperature (1073 K)
were, for example, non-decomposed Nd
[10], Fe B [8], localised strain [11], and Fe
the hypothesis of non-decomposed Nd Fe14Bregions,
2
Fe14Bgrains
2
3
B [12,13]. In
[
1,2]. This treatment holds promise as a production
2
method for magnets, which will enable, e.g. the design of
smaller and more economical motors [3,4]. A hydrogen
treatment for magnetic hardening was first reported in
the anisotropy mechanism is constructed on the basis of
the fact that anisotropy appears in the Co-added
NdFeB, reported in 1990 [1]. However, this hypothesis
was disagreed with because texture was also observed in
ternary Nd–Fe–B, i.e. without Co [13,14]. The authors
of the present paper discovered that the appearance of
anisotropy or isotropy is determined by the rate of the
reaction between NdFeBand hydrogen, which led to the
1
985 [5]. The hydrogenation disproportionation deso-
rption recombination (HDDR) process was formulated
in 1990 [6,7]. This treatment was improved to the d-
HDDR treatment by the authors, and the specific
processing conditions for anisotropy were revealed in
1
1
996 [8]. In 2000, an anisotropic-bonded magnet with
À3
d-HDDR process, and then proposed the Fe
thesis for the anisotropy mechanism. We consider Fe B
3
2
Bhypo-
60 kJ m was commercialised [3,9] and came into wide
use.
as an insufficient orientation carrier as it is a metastable
phase and can only be used under specific conditions
3 2
[13]. However, the stable derivative of Fe Bis Fe Band
Ã
we observed the aligned Fe Bphase during the
2
Corresponding author. Tel.: +49 351 4659 664;
fax: +49 351 4659 537.
E-mail address: o.gutfleisch@ifw-dresden.de (O. Gutfleisch).
disproportionation step in the d-HDDR treatment
[15]. In this paper, we summarise the phenomena in
0304-8853/$ - see front matter r 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jmmm.2004.11.423

ARTICLE IN PRESS
Y. Honkura et al. / Journal of Magnetism and Magnetic Materials 290–291 (2005) 1282–1285
1283
the d-HDDR treatment and the anisotropy mechanism,
and then term this anisotropy mechanism as the ‘‘texture
memory effect’’ (TME).
1273
1173
1073
DR
HD
2
. The d-HDDR phenomena
9
73
73
0
In the conventional HDDR treatment, which is based
on the equilibrium state, the formation of fine grains can
be understood, but the formation of texture and the
appearance of anisotropy remained unclear. In the d-
HDDR treatment, which is based on the control of the
reaction rate between NdFeBand hydrogen, the specific
processing conditions for anisotropy were revealed as
shown in Fig. 1 (top). The maximum energy product
8
0-3
0.1
1
10
100 1000
1
H Pressure / kPa
2
Fig. 2. The d-HHDR process in the Nd
diagram.
2
Fe14B–H phase
(
BH)max depends strongly on hydrogen pressure. In
other words, the control of the reaction rate between
NdFeBand hydrogen enables isotropy or anisotropy.
The effect of Nb and Ga on (BH)max is shown by
found that Co is the element, which lowers the reaction
rate. In the past, Co was considered an essential element
for anisotropy, and it can be assumed that the experiments
were performed at atmospheric pressure. Moreover, the
effect of various additive elements was studied, and it was
found that Ga and Nb are effective elements for coercivity
and anisotropy, respectively.
À3
comparing with Nd12.5Febal.B6.2. A (BH)max of 343 kJ m
was obtained for the Nd12.5Febal.B6.4Ga0.3Nb0.2 alloy. The
effect of Co on (BH)max on the reaction kinetics is shown
in Fig. 1 (bottom). As the quantity of Co increases, the
hydrogen pressure needs to be increased in order to
maximise (BH)max. Consistent with Refs. [16,17], it is then
In the HDDR reaction, the process for anisotropy
was formulated as the d-HDDR treatment. In order to
achieve industrial products, the d-HDDR treatment
consists of 4 steps, as shown in Fig. 2. In step 1,
4
3
2
1
00
00
00
00
0
Nd12.5Febal.B6.4Ga0.3Nb0.2
hydrogen is sufficiently absorbed into Nd
disproportionation. In step 2, Nd Fe BH is slowly
2
Fe14Bwithout
2
14
x
disproportionated into 3 phases. In step 3, the 3 phases
are recombined into Nd Fe14BH . Finally, in step 4,
residual hydrogen is removed from Nd Fe BH . Steps 2
2
x
2
14
x
and 3 are the key processes for the appearance of
anisotropy. Steps 1 and 4 are just the absorption process
and desorption process of hydrogen.
Nd12.5Febal.B6.2
40 60
0
20
80
100
H Pressure / kPa
2
3
2
1
00
00
00
0
3
. Hypothesis for anisotropy mechanism
X=10
X=20
X=0
First, the structural change during hydrogen treatment
is such that, in the hydrogenation disproportionation
process, the lamellar structure is formed at the early stage
X=30
(
compare Figs. 3 (left) and 4), and the structure changes
into a spherical morphology (Fig. 3 (right)) as reaction
time increases. In the desorption recombination process,
a rim phase of Nd Fe B-type appears around the Nd-
2
14
hydride phase at the early stage, and finally the re-
crystallisation to Nd Fe14Bis completed. Looking at the
crystal structures, tetragonal Nd Fe Bdecomposes into
2
0
100
200
300
2
14
H Pressure / kPa
2
3
phases, which are cubic NdH
Fe Bin the disproportionation process. In step 3 of the
process, tetragonal Nd Fe14Bis recrystallised from 3
phases. Within the 3 phases, only Fe Bhas an anisotropic
2
, cubic Fe and tetragonal
Fig. 1. Effect of hydrogen pressure on (BH)max at 1093 K in
top) Nd12.5Febal. 6.4Ga0.3Nb0.2 and (bot-
tom) in Nd12.5Febal.
2
(
B
6.2 and Nd12.5Febal.
6.0Ga1.0Co
B
2
B
x
.
2

ARTICLE IN PRESS
Y. Honkura et al. / Journal of Magnetism and Magnetic Materials 290–291 (2005) 1282–1285
1284
PH2
PH2
30 kPa
5 kPa
c-axis
tetragonal
Fe B
tetragonal
Nd Fe B
cubic
NdH2
2
tetragonal
Nd Fe B
cubic Fe
2
14
2
14
Fig. 5. Schematic structural changes with Fe
hypothetical orientation memory site: (left) before d-HDDR,
2
Bas the
(
middle) disproportionated structure with aligned Fe
right) fine grained, textured material after d-HDDR.
2
Band
(
random if the reaction is higher due to a higher hydrogen
partial pressure, and Nd Fe14Bwill be isotropic.
2
Fig. 3. (left) SEM micrograph in the backscattered mode
showing the reaction front upon disproportionation; (right)
4
. Structural changes during d-HDDR
Experimental evidence for the Fe Bhypothesis was
TEM bright field image showing well-aligned Fe
within the disproportionated structure (A—Nd Fe14B, B—Fe/
Fe(B), C—NdH and D—Fe B).
2
Bgrains
2
2
2
2
obtained using scanning and transmission electron
microscopy (SEM, TEM). The conditions of d-HDDR
treatment are 1093K for 3 h in hydrogen pressure of 30
kPa for disproportionation (step 2), and 5 kPa upon
recombination (step 3). For details of the experiment,
please refer to Ref. [15]. Fig. 4 shows the reaction front
moving from the parent phase (A) to the disproportio-
nated phases (B,C,D) in more detail. A lamellar structure
of NdH , and Fe is observed at the wave-shaped
2
interface. Electron energy loss spectroscopy (EELS)
shows that some boron is found in the Fe, indicating a
possible initial oversaturation of boron in Fe. In the
progress of the reaction, the Fe(B) phase is transformed
into Fe
lamella structure is the key for precipitated and well-
aligned Fe B. There must be uni-axial stress along the
2 2
Band Fe, and well-aligned Fe Bprecipitates. The
2
interface because of the lamellar structure. Well-aligned
tetragonal Fe Bmay precipitate along the uni-axial stress
in order to relieve the stress. Fig. 3 (right) shows the
2
completion of disproportionation into NdH
2
, Fe and
Fig. 4. TEM bright field image showing the reaction front upon
disproportionation in more detail (A—Nd Fe14B, B—Fe/
Fe(B), and C—NdH ).
Fe B . FeBgrains have stacking faults, indicated by the
2 2
2
hatched lines in the TEM bright field image. Stacking
faults could be induced during the rapid cooling. The
hatched lines are all well aligned in one direction and
more detailed conventional beam electron diffraction
2
2
crystal structure indicating that the Fe Bphase could be
crucial for the memory effect of anisotropy. This
hypothesis is shown schematically in Fig. 5. After
(
2
CBED) analysis reveals that Fe Bcrystallites are aligned
within 201 during all processing steps. Therefore, it is
decomposition, tetragonal Fe
to tetragonal Nd Fe14B, and the Fe
the crystal orientation of Nd Fe Bduring phase
2
Bphases precipitates along
thought that this well-aligned Fe Bis a memory phase for
anisotropy after recombination. We term this mechanism
2
2
Bcrystallites keep
2
2
14
‘‘texture memory effect (TME)’’.
transformation. Subsequently, fine-grained Nd
phases precipitates in step 3. We assume that the degree
of alignment, that is, the fluctuation of Fe B, depends on
the reaction rate. Fe Bmemorises the direction of
2
Fe14B
2
5. Projections
2
2
Nd Fe14Bsufficiently if the reaction rate between
À3
hydrogen and NdFeBis low, and as a result, recombined
Nd Fe Bwill be anisotropic. Fe 2Bprecipitates at
In 1997, bonded NdFeBmagnets with 136 kJ m
were first commercialised into market, and in 2002, the
2
14

ARTICLE IN PRESS
Y. Honkura et al. / Journal of Magnetism and Magnetic Materials 290–291 (2005) 1282–1285
1285
performance of bonded magnets increased up to
[6] T. Takeshita, R. Nakayama, Proceedings of the 10th
International Workshop on RE Magnets and their
Applications, 1989, p. 551.
À3
2
2
00 kJ m . In the laboratory, a bonded magnet with
16 kJ m has been produced [18]. The accomplished
À3
[
[
[
7] P.J. McGuiness, X.J. Zhang, X.J. Yin, I.R. Harris, J. Less-
Common Metals 158 (1990) 379.
8] H. Honkura, C. Mishima, H. Mitarai, Tokaihei 10-135019
detailed knowledge of the ‘‘texture memory effect
TME)’’ should enable the realisation of further
improved magnetic properties. In the near future, an
(
(
Japanese patent).
anisotropic bonded NdFeBmagnet with about
À3
9] C. Mishima, N. Hamada, H. Mitarai, Y. Honkura, IEEE
Trans. Magn. 37 (2001) 2467.
2
40 kJ m could be produced on an industrial scale.
[
[
[
10] M. Uehara, H. Tomizawa, S. Hirosawa, T. Tomida, Y.
Maehara, IEEE Trans. Magn. 29 (1993) 2770.
11] H. Nakamura, K. Kato, D. Book, S. Sugimoto, M. Okada,
M. Homma, J. Magn. Soc. Jpn. 23 (1999) 300.
References
12] T. Tomida, N. Sano, K. Hanafusa, H. Taomizawa,
S. Hirosawa, Acta Mater. 47 (1999) 875.
[
[
[
1] T. Takeshita, N. Nakayama, Proceedings of the 11th
International Workshop on RE Magnets and their
Applications, Pittsburgh, 1990, p. 49.
[13] O. Gutfleisch, B. Gebel, N. Mattern, J. Magn. Magn.
Mater. 210 (2000) 5.
2] O. Gutfleisch, I.R. Harris, Proceedings of the 15th
International Workshop on RE Magnets and their
Applications, 1998, p. 487.
[14] H. Nakamura, R. Suefuji, S. Sugimoto, M. Okada, M.
Homma, J. Appl. Phys. 76 (1994) 6828.
[15] O. Gutfleisch, K. Khlopkov, A. Teresiak, K.H. Muller, G.
Drazic, C. Mishima, Y. Honkura, IEEE Trans. Magn. 39
(2003) 2926.
3] Y. Honkura, Proceedings of the 7th International Work-
shop on RE Magnets and their Applications, Delaware,
2
002, p. 52.
4] Y. Hayashi, H. Mitarai, Y. Honkura, IEEE Trans. Magn.
9 (2003) 2893.
[16] A. Fujita, I.R. Harris, IEEE Trans. Magn. 30 (1994) 860.
[17] S. Sugimoto, O. Gutfleisch, I.R. Harris, J. Alloys
Compound. 260 (1997) 284.
[
[
3
5] Y. Inoue, N. Imaizumi, Tokukaisho 62-23903 (Japanese
patent).
[18] N. Hamada, C. Mishima, H. Mitarai, Y. Honkura, IEEE
Trans. Magn. 39 (2003) 2953.