HAL Id: jpa-00225930
https://hal.archives-ouvertes.fr/jpa-00225930
Submitted on 1 Jan 1986
HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
AN FIM-ATOM PROBE AND TEM STUDY OF ALNICO PERMANENT MAGNETIC MATERIAL
S. Cowley, M. Hetherington, J. Jakubovics, G. Smith
To cite this version:
S. Cowley, M. Hetherington, J. Jakubovics, G. Smith. AN FIM-ATOM PROBE AND TEM STUDY OF ALNICO PERMANENT MAGNETIC MATERIAL. Journal de Physique Colloques, 1986, 47 (C7), pp.C7-211-C7-216. �10.1051/jphyscol:1986737�. �jpa-00225930�
AN FIM-ATOM PROBE AND TEM STUDY OF ALNICO PERMANENT MAGNETIC MATERIAL
S.A. COWLEY, M.G. HETHERINGTON, J.P. JAKUBOVICS and G.D.W. SMITH
Department of Metallurgy and Science of Materials, Parks Road, GB-Oxford OX1 3PH, Great-Britain
Abstract: Alnico 8 specimens have been* studied both in the thermomagnetically aged and further heat treated condition using the field ion microscope (FIM) and high voltage electron microscope. The results are compared with those produced by Hetherington et al., [I] on Alnico 5 and Zhu et al.. 123 on Alnico 8.
Introduction: Alnico alloys commonly contain some 6-12%A1, 14-25%Ni, 0-35%Co, 0- 8%Ti and 0-6%Cu in 40-70%Fe and are distinguished by a number appended to their generic description. Below 845OC [ 3 ] , Alnico alloys are prone to spinodal decomposition resulting in the separation of two crystallographically isomorphous but compositionally dissimilar magnetic phases [4,5,6]; b.c.c. Fe-Co-rich and b.c.c Ni-Al-rich (see for example [2]). Hetherington et al. [I] have been abJe to establish the composition and fine structure of the phases in Alnico 5. The highly interconnected morphology of the phases was cited as having a strong influence on the magnetization properties. In addition .these workers discovered ordering (close to 82 or DO type) in both the Fe-Co and Ni-A1 phases. In contrast to this work, Pfeiffer [lf-used diffraction evidence to deduce that only the Ni-A1 phase was ordered. By examining a homogenised Alnico alloy, Hutten and Grune [8] have recently obtained atom probe results which appear to show that Ni and A1 atoms are ordered even after homogenisation at 900°C for 15 hours (i.e prior to spinodal decomposition). In this paper we explore further the structure/properties relations of the Alnico alloys.
Results and Discussion
Morpholoqv: Alnico 8 (13.7at%A1-11.8at%Ni-33.6at%Co-6.8at%Ti-2.5at%Cu-0.3at%S- balFe) was provided by Swift-Levick Ltd in a thermomagnetically aged condition following: solution treatment at 1250°C. rapid quench, 18 mins at 820°C followed by tempering at 590°C and 560°C respectively for 48 hours each. Figure 1 shows a TEM image of the highly contiguous structure of the as-received Alnico 8 with composition modulations in <loo> directions. An FIM image looking down a [loo] pole (figure 2) illustrates the bright and dark imaging areas corresponding to the Fe-Co- rich and Ni-Al-rich phases respectively [I]. The light area in figure 2 describes a facetted Maltese Cross-like shape which was deduced to be the characteristic of a repeat morphology of the modulated structure. D.C. evaporation of successive planes revealed the depth variation of the Fe-Co phase morphology (figure 3 a-f). By recording the number of planes evaporated between each pair of photographs the repeat distance was found to be -50nm. Figure 4 shows a high voltage electron microscope (HVEM) image of a FIM tip which elegantly illustrates the interconnectivity of the Ni-A1 phase and the isolation of the Fe-Co phase. The
Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1986737
C7-212 JOURNAL DE PHYSIQUE
Alnico 5 investigated by Hetherington et al. [I] displayed contiguous, ellipsoidal particles which were very different in shape to the present morphology.
Composition of Phases: Compositional data from the Fe-Co and Ni-A1 phases was obtained by AP analysis using the Oxford FIM VGlOO atom probe at liquid nitrogen temperatures (-80K) using a pulse fraction of 20%. The average composition of the phases obtained is shown in Table 1.
Table 1:- Average composition of phases in Alnico 8 after standard heat treatment.
Fe-Co 51.620.8 4.920.3 2 . 7 2 0 . 2 0.2f0.1 40.0_+0.7 0.520.1
Using these data and applying a mass balance criterion, the volume fraction of the Fe-Co (bright) phase was calculated to be 0.51 which should be compared with the figure of 0.81 obtained for Alnico 5 [ I ] . After further heat treatment at 550°C (48 hours) and 800°C (0.5 hours), the morpholo@y of the Fe-Co phase changed considerably although no microscopic precipitation reactions as reported by Pfeiffer [ 7 ] were noted. The volume fraction of a plate shaped precipitate did, however, increase.
Microprobe analysis showed this to be a titanium sulphide (probably TiS). Figure 5 shows a FIM image of the spinodal obtained from a sample after 30 mins at 800°C. The interface has adopted a rounded cross-section reminiscent of the morphology noted by Hetherington et al. for Alnico 5. In addition, TEM (figure 6) and field evaporation sequences revealed that the longitudinal morphology of the Fe-Co was in the form of elongated rods. It has been suggested that the large Ti atom size introduces lattice strains which affects the morphology. The removal of Ti via TiS precipitatjon would, therefore, reduce the overall lattice strain and this may have contributed to the change in morphology.
Comparing the compositional data from the as-received and heat treated samples (Table 2) we fidd that Ni has diffused out of the Ni-A1 phase whilst Fe, Co, Al, Ti and Cu have stayed within the statistics of the experiment.
Table 2:- Average phases in Alnico 8 after further heat treatment at
8 0 0 ° C (see text).
Fe-Co 51.522.6 5.820.1 4.3f0.1 0.1+0.1 33.652.6 1.320.1
The volume fraction of bright phase calculated from the above data was 0.46 which is comparable with the result obtained from the "as-received" sample. This indicates that although a morphological change has been effected by the heat treatment, the volume fractions of the phases changes little. Figure 7 shows a high magnification FIM picture of the spinodal interface after heat treatement at 800°C.
It can be seen that the bright rings of Fe/Al atoms are continuous across the interface. The association of like-atom pairs in the interface leads to a lower interfacial energy than the coupling of unlike atom pairs. It is suggested that the magnetostatic energy during thermomagnetic aging and strain energy are probably more important than surface energy (or magnetostriction) in dictating the morphology of the Alnico 8 spinodal.
De Vos 693 has suggested that the low diffusivity of Ti in Alnico 8 hinders the procession of the interface and thus accounts for the sluggish spinodal reaction kinetics. Figure 8 shows the analysis across the interface in terms of the Fe, A1 and Ti content. It can be seen that there is an A1 "spike" on the Ni-A1 side of the
and Ti.
Orderinq in Fe-Co and Ni-Al: For comparison with the work of Pfeiffer, diffraction evidence for the existence of ordering in the Fe-Co phase was sought but none was found. However, close examination of the micrographs obtained from the FIM image revealed ordering in the Fe-Co phase identical to that observed by Hetherington et ai. [I] in Alnico 5; bright (Fe + A1 atoms) and dark (Co + Ni) rings were visible when imaging at a <loo> superlattice pole. In addition, the layers evaporated at different rates indicating a difference in composition. The ordering was confirmed after heat treating specimens at 800°C for 30 mins and performing compositional analysis in the atom probe. The probe hole in the channel plate was set at the maximum distance fom the specimen so that atoms from only one plane were seen by the analyser. Figure 9 shows auto-correlation curves for Fe and Co and the periodic arrangement of Fe and Co is clear. The auto-correlation curves have a periodicity of 60 atoms which is the average number of atoms caught as the atoms from two planes pass through the probe hole. Ordering was also detected in the Ni-A1 phase. In addition, auto-correlation curves from the Fe and Co in the Ni-A1 phase indicated ordering of these elements within the dark phase itself.
The existence of such a high degree of Fe-Co order at 800'C is, at first sight, surprising. The Fe-Co binary equilibrium diagram [lo] indicates that the maximum previously detected Fe-Co ordering temperature was some 733°C at 50/50at%
Fe/Co. The presence of Ni, Ti or A1 has therefore stabilised the ordering reaction to much higher temperatures which confirms the conclusions of Mal'tsev et al. [Ill.
These workers demostrated that A1 substitutes for Fe on the Fe-Co ordered lattice and that the A1-Co interaction is much stronger than that of A1-Fe. If the relative interaction between A1-Co is greater than that between Fe-Co it is then easy to understand how the ordering kinetics may be stabilised to higher temperatures by the presence of Al. It is, of course, possible that what is observed is an as yet unreported type of Fe-Co ordering. However, the proportions of iron and cobalt present appear to indicate 82 ordering.
The Onset of Order: To test the stability of the ordering, samples of Alnico 5 were homogenised for 2 hours at 1300°C in evacuated capsules followed by rapid quenching into brine. During AP analysis the collapse of an atom plane past the probe hole was marked in the data. This allowed a plane by plane analysis of the ions arriving at the analyser. It was found that A1 resided on Fe-rich planes whilst Ni was associated with the Co-rich planes (figure 10). In addition, statistical significance for the associations was noted in the auto-correlation curves; at the 5% level the standard deviation was 0.065 f 0.022 which should be compared with the theoretical standard deviation of 0.040. In contrast to the results of Hutten and Grune 181, these results illustrate the affinity of A1 and Ni for these preferred sites and the stabilisation of Fe-Co ordering even at very high temperatures. It should be mentioned, however, that quenching specimens whilst still sealed in encapsulating tubes decreases the quenching rate by some forty-fold as compared with specimens in direct contact with the quenching medium and it is possible that both in the present investigation and that of Hutten and Grune [8] the observed affinity between species is as much due to a quenching effect as it is to do with any specific ordering reactions.
High Voltage Lorentz MicroscoDy: High voltage Lorentz microscopy was performed using the Oxford A.E.1 EM7 HVEM. Figure 11 shows an image of the domains obtained from a thin foil of the 800°C heat treatment. The black/white contrast represents the magnetic domains. These domains are much smaller than those observed in Alnico 5 [I] but larger than those in optimally heat treated Alnico 8. This is probably due to the larger degree of interconnectivity of the Fe-Co phase in overaged Alnico 8. In optimally heat treated Alnico 8, the movement of Bloch walls is considerably restricted. This may, in part, explain the superior coercivity and remanence of these alloys. Figure 12 shows the magnetic domains close to the interface of a TiS precipitate. It should be noted that there is little change in
C7-214 JOURNAL DE PHYSIQUE
the domain structure close to the spinodal/precipitate interface as compared with an area remote from the precipitate illustrating that this decomposition product probably has little, if any, effect on the magnetic properties.
Conclusions: The morphology of the spinodaj phases in Alnico 8 is quite different from Alnico 5 . This may account for the higher coercivity of the former. Ordering reactions in Alnico alloys may occur at temperatures well in excess of the maximum temperatures predicted by the binary phase diagrams of the individual chemical species.
Acknowledgements. The authors are grateful to Swif t-Levick Ltd for the provision of specimens. SAC and MGH are grateful to the Science and Engineering Research Council (SERC) Central Electricity Generating Board (CEGB) respectively for the provision of research fellowships.
References
1 Hetherington, M. G., Cerezo, A., Jakubovics, J. P. and Smith, G. D. W., J.
Physique, (1984). C9-429.
2. Zhu, F., Wendt, H., Alvensleben, L. V. and Haasen, P., Map-20 (1984), pp. 1619.
3 McCurrie. R. A. in "Ferromagnetic Materials", (Ed. E. P. Wohlfarth), Noth-Holland pub1 . , 3, pp. 138
4 Sergeyev, V. V. and Bulgina, T. I. IEE Trans. Magn., Mag-6, (1970), pp. 194.
5 Berkowitz, A. E., "Magnetism and Metallurgy" (Ed. A. E. Berkowitz and E. Kneller), Academic Press (1969). pp. 331.
6 Yermolenko, A. S. and Shur Ya, S., Fiz. Met. Mealloved, 1& (19691, pp. 31.
7 Pfeiffer, I., Cobalt, 44. (1969), pp. 115.
8 Hutten, A. and Grude, R. Scripta Met. (Manuscript submitted).
9 De Vos, K. J. "Maghetism and Metallurgy" (Ed. A. E. Berkowitz and E. Kneller), Academic Press (1969). pp. 473.
10 Kubachewski, "Fe-Co, Iron Cobalt", Springer-Verlag. (1982). pp. 27.
11 Mal'tsev, Ye. I., Goman'kov, V. I., Mokhov, B. N., Puzey, I. M. and Nogin, N. I., The Physics of Metal and Metallography, 40, (1975), pp. 190.
Figure 1. Figure 2.
TEM micrograph of Alnico 8 morphology FIM micrograph of Alnico 8
morphology: Bright areas = Fe-Co phase, dark areas = Ni-A1 phase.
showing the highly interconnected spinodal after further heat treatment at morphol~gy of the Ni-A1 phases and 800'C for 30 mins. Note the rounded cross- the isolation of the Pe-Co ~hase. section.
Figure 6. TEM image of the Alnico 8 Figure 7 . High magnification FIM spinodal after further heat treatment picture of the spinodal interface after at 800QC for 30 mins. Note the further heat treatment. Note the morphology is in the form of long rods. continuity of the rings across the
interface.
JOURNAL DE PHYSIQUE
"I
-1I
Number of Samples (k) 158Figure 9a. Auto-correlation curve for Fe in Fe-Co phase showing the periodic ordered arrangement of the iron atoms.
Auto-correlatlm Co Mean=-O.0024 (-0.0019) ' 0 0 ~
Variance= 0.0150 ( 0.0019)
-
-fi.-- ,A
-&+- --J!. . - - - . - --
0 Number of ions 3000 I 8
Number of samples (k) 158
Figure 8. The composition at the spinodal Figure 9b. Auto-correlation curve for interface. Note the existence of an Co in Pe-Co phase showing the periodic
A 1 "spike". arrangement of the cobalt atoms.
-1 !
8 Number of S@SS (k) i!8
a. b. w of Ganples (k) 28
Figure 10. Auto-correlation curve illustrating Ni-Co association after homogenisation at 1300°C. a.) Ni-Co, b) Fe-Al. (theoretical mean and variance in
Pigure 11. Magnetic domain structure of Alnico 8. Figure 12. Domain structure close to a TiS precipitate.
