Use of the HDDR Process in Preparation of Zirconia Doped Nd-Dy-Fe-B High Coercivity Povvder
HDDR postopek kot metoda za pripravo visoko koercitivnih Nd-Dy-Fe-B prahov
S. Kobe Beseničar,* B. Saje, G. Dražič, Jožef Štefan Institute, Ljubljana, Slovenia
The present paper deals vvith the use of the HDDR process as the preparative method for obtaining zirconia doped high coercive Nd-Dy-Fe-B povvders. The influence of the dopant and of the processing parameters on the HDDR mechanism and the magnetic properties of the povvders obtained, vvas studied. The material vvas characterized by magnetisation measurements at various stages of the HDDR process. Electron microscope studies on the Nd-Dy-Fe-B povvders vvere performed to observe the influence of zirconia addition on the crystallisation during the recombination process. Key vvords: Nd-Fe-B magnets, coercivity, processing.
HDDR postopek (Hidrogenacija Disproporcionacija Desorpcija Rekombmacija) smo uporabili kot metodo za pripravo finih visoko koercitivnih Nd-Dy-Fe-B prahov dopiranih s cirkon oksidom. Študirali smo vpliv dopanta in procesnih parametrov na mehanizem poteka HDDR postopka in na končne morfološke in magnetne lastnosti tako dobljenih prahov. Vzorce smo karakterizirali na različnih stopnjah HDDR postopka z magnetnimi meritvami. Prahove smo opazovali z elektronskim mikroskopom in študirali vpliv cirkonovega oksida na potek kristalizacije med postopkom rekombinacije. Ključne besede: Nd-Fe-B magneti, koercitivnost, procesiranje.
Introduction
One of the vvell knovvn methods for preparation of isotropic Nd-Fe-B coercive povvders is the HDDR process, as reported in several papers14. It vvas also established previously that an anisotropic povvder can be produced by the addition of small amounts of Zr to the initial composition. The intrinsic coercivi-ties of povvders prepared by this method are reported to be up to 1100 kA/m (13.8 kOe)5 6.
In our previous vvork the beneficial influence of zirconia on the microstructure and consequently on the magnetic properties and corrosion resistance of sintered Nd-Dy-Fe-B magnets vvas reported7".
The purpose of the present vvork vvas to prepare high coercive Nd-Dv-Fe-B povvders by the HDDR processing route, using the same composition of the basic alloy, together vvith the addition of 1 wt.% of ZrO,, as previously employed8. High coercivi-ty povvders vvere intended for use as the basic material for the preparation of resin bonded magnets. This paper only deals vvith the povvder processing and its characterization.
Experimental
The basic alloys for the HDDR process vvere prepared by are melting the allovs NdFe, DyFe, FeB and Fe povvder in a pure Ar atmosphere. The alloy composition vvas Nd l„_KDyxFe7f,Bli (0<x<3). In order to prevent oxidation, a Ti sponge vvas used as
* dr Spomenka KOBE BESENIČAR. [nšlilul Jožef Štefan, Jamova 39.61000 Lj..
a getter for oxygen. 1 wt.% of zirconia vvas added before are melting. The melted buttons vvere then subjeeted to HDDR processing. They vvere treated in hydrogen at room temperature (20kPa) first and then heated to different temperatures betvveen 750°C and 850°C and exposed to further hydrogenation for tvvo hours. A thermopiezic (TPA) analysis and DTA analyses vvere performed to follovv the absorption proccss. The procedure vvas follovved by evacuating the system and after exposing the samples to maximum temperature and high vacuum (10 1 Pa) for one hour they vvere furnace cooled6.
Samples vvere lightly crushed and the povvders obtained vvere characterized by magnetic measurements. For the mass magne-tization measurements a DSM8 magnetometer - susceptometer vvas used; the intrinsic coercivity measurements vvere performed using a permeameler. Comparability of the measurements vvas attained by using always the same quantity of examined povvder and binding material (epoxy) (wt.% 85/15). Samples vvere pulsed in a field of 4000 kA/m and demagnetized vvith a field of cca 2000 kA/m. Povvders vvere observed by means of a SEM/EPMA (JEOL JXA 840 A). Phase analyses vvere performed using a TEM (JEOL 2000 FX).
Results and discussion
Figure 1 shovvs the demagnetizing curves of povvders obtained by the HDDR process at different temperatures. Samples vvith and vvithout zirconia addition are compared. The basic composition of those samples is NdlsDy|Fe?(,Bs. Z0 samples are vvithout zirconia addition and Z1 vvith I wt%
-H UOe)
Figure 1: Demagneti/.ing curves of povvders obtained by the 11 DDR process at different temperalures. Samples vvith and vvithout zirconia are compared.
Slika I: Demagnetizacijske krivulje prahov, dobljenih po HDDR postopku pri različnih temperaturah. Primerjava vzorcev z in brez dodatka cirkon oksida.
of zirconia additive. A maximum value of the intrinsie coercivi-ty of 1400 kA/tn ( 17.5 kOe) was attained vvhen samples vvith the addition of zirconia vvere processed at 775 C. This value exceeds the coercivity of samples vvithout zirconia and processed under the same conditions for about 13 % of the value obtained vvith samples vvithout additive. At 750 C the poorer magnetic properties obtained vvere attributed to the presence of free iron. Grain grovvth at S00 C causes a decrease in cocrcivitv.
Figure 2 shovvs the differenee in the mass magnetization be-tvveen samples vvith and vvithout ZrO, addition processed at different temperalures. Several experiments in two different labo-ratorics shovved reproduciblc results. Samples vvith zirconia addition revealed ;i higher magnetization at lovver temperature (750°C) than samples free of this additive. In Figure 3 the SEM micrographs of the samples vvithout addition of zirconia (A) and vvith zirconia addition (B) are shovvn. Both samples vvere processed under the same conditions. Thev vvere treated at 750 C. There is obvious differenee in the grain size. vvhieh is around 0.3 um in the samples vvithout ZrO, addition and in the range betvveen 0.3 um and 1 um in the samples with the additive. The most probable reason for the recombination process starting at lovver temperalures and consequently the higher magnetization of doped samples, lies in the different reaction kinetics of zirconia doped samples. At temperalures of the recombination
M 70
E 60 <
^ SO -C o
f| 40-
£3 V
j| ^ "
„ 20 10-
»ithoul Z1O2	750°C
»ilh ZrO i 1 nt %	TStrt:
»ithoul Zr02	800°C
wilhZr02tnt%	S(I0"C
600
800
Magnetic field strcngth 10' (A/m)
Figure 2: Mass magnetization measurements ol samples vvith and
vvithout ZrO, addition, processed at different temperalures. Slika 2: Magnelizacija vzorcev z in brez dodatka cirkon oksida, pripravljenih pri različnih temperaturah, kot funkcija jakosti magnetnega polja.
Figure 3: SEM micrographs of the samples processed under the same conditions (750 C I: A - vvithout zirconia. B - »ith zirconia addition.
Slika 3: SEM posnetki vzorcev pripravIjenih pri enakih pogojih (750 Cl: A - brez cirkon oksida, B - s cirkon oksidom.
process higher than 750 C" the mass magnetization is normallv alvvavs higher in samples free of additive.
Our TEM observations confirmed the presumption about different reaction kinetics of zirconia doped samples. When the samples are exposed to the recombination process at 750 C. nano crvstals of Nd rich phase vvhieh occurs vvithin the origin grains are at least an order of magnitude smaller in the čase of zirconia free samples (2 11111) (Fig. 4A) than in the čase of samples vvith zirconia addition (up to 20 nm) ( Fig. 4B|.
In addition the crvstallites of hard magnetic phase vv ithin the origin grains are at least an order of magnitude bigger in zirconia doped samples (Fig. 5B) comparing to the zirconia free sample (Fig. 5A). vvhere nano crvstals do not exceed 10-20 11111.
In Figure 6 EDX spectra of 2-14-1 phase in samples vv ithout (A) and vvith zirconia (Bi addition are shovvn. The FeK Nd, peak-height ratio is lovver in the čase of sample (A) (cf. Fig. 5l indicating that several small Nd-rich particles vvere also present in the analvsed volume. In the čase of sample B. the particle size of Fe-rich phase is large enough to obtain just the spectrum of this phase vvithout the interference from Nd-rich phase.
TPA analvsis of the absorption process shovved the dispro-portionation starting at higher temperature and procceding slovv cr in zirconia doped samples (Fig. 7). These results vvere confirmed also vvith a DTA analvsis (Fig. S). It is obvious that disproportionation of the samples vvithout zirconia addition starts al lovver temperature.
<j|>P i JF	
	
	
w % nt	a® itn
jAgft


Figure 4: TEM mierographs (dark field) and eorresponding diffraetion patterns of NdDvFeB samples - Nd rieh phase: A - /ireonia Irce. B - /ireonia added. Slika 4: TEM posnetki (temno polje) in odgovarjajoči dilraktogrami vzorcev NdDvFeB - fa/a bogata / Nd: A - brez eirkon oksida. B - / dodatkom eirkon oksida.
v r« ■ T
Figure 5: TEM mierographs (dark field) and eorresponding diffraetion patterns of NdDvFeB samples - Nil,Fel4B phase: A - /ireonia free.
B - /ireonia added. Slika 5: TEM posnetki (temno polje) in odgovarjajoči difraktoerami vzorcev NdDvFeB: laza - Nd,Fe,',B A - bre/ eirkon oksida. B - / dodatkom eirkon oksida.
	F	I	A
/FC D S2 J&JLJU-------	■ ^ . ^.jlhtJ	lis)	C A
	4.760 keU		9.9 >
# 0p C D 5 z AO. ir	j K d « .	C \b"l	B w
4.760 keU
9.9 >
Figure 6: EDX spectra of 2-14-1 phase in samples vvilhout A and vvilh B /ireonia addition. Slika 6: EDX spektri 2-14-1 faze v vzorcih bre/ A in / dodatkom eirkon oksida B.
A tentative explanation is that the statement about the local stabilization of 2-14-1 phase vvitlt rcspect to disproportionation. vvhen elentents sueh as Zr are added to the basic alloy" could be also transmitted in the čase of zirconia addition. With the addition of zirconia the so called cells remain unaffected by the disproportionation and thev act as nueleation centres in the recom-bination process. This model which explains anisotropv of sueh povvders can tentatively also explain our TEM observations. A certain texture strueture of the Nd-rich crystals formed during the recombination process in zirconia doped samples vvas detected (Fig. 4B. diffraetion pattern). In the samples vvithout Zr02 addi-tion the strueture consists of randomly oriented nano crystallites (Fig. 4A). A detailed analvsis vvith TEM on sulici samples vvill give us more information about phase composition. There are
1 1 00 1000 900 800 700 600 500 400
S3
-O
B
D
v?
aj D. cs X
0	200 400 600 800 1000
Temperature (°C)
Figure 7: T PA traees for /ireonia free (Z0) in zirconia added (XI) Nd|,Dy|FeT„BK samples.
Slika 7: TPA anali/a vzorcev Nd,5Dy,Fe „BS: 7.0 - bre/ eirkon oksida, /.I - / dodatkom eirkon oksida.
Figure 8: DTA of Ndl5Dy,Fe7f,Bs in 1 barhydrogen (note: graphs are not calibrated to the mass)
Slika 8: DTA analiza vzorca Nd15Dy1Fe7„B,( pri I bar vodika (grafi niso kalibrirani na maso).
some indications about the presence of a phase based on Fe and Zr and this will also help lis to explain the results of our first TEM analyses.
On the basis of these results, one can conclude that the HDDR processing route for the preparation of high coercive zirconia doped Nd Dy-Fe-B povvders reaches its optimum at 775°C for the present processing conditions (temperatures tested vvere: 750°C, 775°C, 800 'C. 850°C). Further vvork will study the preparation of anisotropic magnets by different established tech-niques.
Acknowledgment
The financial assistance of the Ministry of Science and Technology and the British Council. UK are gratefully ac-
knovvledged. Part of the vvork vvas performed at the Universitv of Birmingham. School of Metallurgv and Materials. Authors (S. Kobe Beseničar and B. Saje) vvould like to thank Prof. I. Rex llairis for his guidance and fruitful discussions. Prof. 1. Rex Harris is also gratefully acknovvledged for his critical revievv of the manuscript. The assistance of Mr. E. Dime for the magnetic measurements and Mr. Z. Samardžija for SEM observations are gratefully acknovvledged.
References
1 T. Takeshita, R. Nakavama, Proeeedings of the 10. International Workshop on Rare Earth Magnets and their Applications. Kvoto, (1989) 551 : P. J. McGuiness, X. J. Zhang. X.J. Yin, I.R. Harris. J. Less
Common Metals, 158 (1990) 359 1 I. R. Harris, P. J. McGuiness, J. Less Common Metals. 172
(1991)1273
4	X. J. Zhang, P. J. McGuiness, 1. R. Harris../. Appl. Phvsies. 68(8) (1991) 5838
5	T. Takeshita, R. Nakayama, Proeeedings of the 11. International Workslwp on Rare Earth Magnets and their Applications. Pittsburgh. PA. (1990) 7
(1 P. J. McGuiness. C. L. Short. 1. R. Harris. IEEE Trans, on
Magn.. 28, (1992) 5, 2160-2162 7 S. Beseničar. B. Saje. G. Dražič. J. Holc. JMMM 104-107.
(1992)1175
s S. Kobe Beseničar, J. Holc, G. Dražič. B. Saje, IEEE Trans.
on Magn.. 30. 2. (1994) 693-696 " 1. R. Harris. Proeeedings of the 12. International workshop on rare-earth magnets and their applications. Canberra (1992) 347