The oxidation kinetics of SmCo alloys

The oxidation kinetics of SmCo alloys

Journal of Alloys and Compounds 473 (2009) 389–393 Contents lists available at ScienceDirect Journal of Alloys and Compounds journal homepage: www.e...

899KB Sizes 0 Downloads 4 Views

Journal of Alloys and Compounds 473 (2009) 389–393

Contents lists available at ScienceDirect

Journal of Alloys and Compounds journal homepage:

The oxidation kinetics of SmCo alloys W.M. Pragnell ∗ , H.E. Evans, A.J. Williams Department of Metallurgy and Materials, School of Engineering, The University of Birmingham, Birmingham B15 2TT, UK

a r t i c l e

i n f o

Article history: Received 25 January 2008 Received in revised form 19 May 2008 Accepted 25 May 2008 Available online 11 July 2008 Keywords: Permanent magnets Oxidation Kinetics

a b s t r a c t SmCo alloys form the basis of excellent permanent magnets with potential service applications up to 550 ◦ C. It is suspected, however, that their oxidation behaviour may limit their usefulness but this is a relatively unstudied subject. In this present work, two grades of Sm2 (Co,Fe,Cu,Zr)17 -based alloys were oxidised in air at temperatures between 300 and 600 ◦ C, and the kinetics of the internal oxidation zone (IOZ) measured using SEM examination. The IOZ was found to grow parabolically with time, with an activation energy of ∼143 kJ mol−1 for growth in both the directions perpendicular and parallel to the c-axis alignment. The extent of penetration of the IOZ was always greater in the perpendicular than in the parallel directions, however. © 2008 Published by Elsevier B.V.

1. Introduction Rare-earth permanent magnets based on NdFeB or SmCo have relatively high energy densities, and therefore have applications in efficient motors and generators with high power to weight ratios. While NdFeB-based magnets are cheaper to manufacture than SmCo-based magnets and have superior magnetic properties at room temperature, they suffer from a loss of magnetic properties at elevated temperatures. The highest temperatures at which NdFeB magnets can currently operate usefully is ∼200 ◦ C (according to manufacturer’s specifications). By contrast, recently developed SmCo-based magnetic alloys can retain useful magnetic behaviour up to temperatures of ∼550 ◦ C [1], making them useful for aero-engine applications such as frictionless bearings or co-axial starter-motors/generators. However, at these high temperatures oxidation may become a problem. Oxidation in air of NdFeB magnets results in the formation of an external surface scale consisting of iron-rich oxides and a much thicker sub-surface zone of internal oxidation [2–5]. This internal oxidation zone (IOZ) thickens by the inward diffusion of oxygen and the subsequent oxidation of the Nd2 Fe14 B magnetic phase to create a dispersion of Nd-oxide particles within an iron-rich matrix [5]. The growth in thickness of the IOZ in the NdFeB alloys reasonably follows parabolic kinetics approximately 20–30 times faster than the growth of the surface oxide. Limitations on oxygen supply to the IOZ by diffusion through the external oxide did not arise in these tests because the friable nature of the scale would allow

∗ Corresponding author. Tel.: +44 121 414 5211; fax: +44 121 414 5232. E-mail address: [email protected] (W.M. Pragnell). 0925-8388/$ – see front matter © 2008 Published by Elsevier B.V. doi:10.1016/j.jallcom.2008.05.083

molecular access to the underlying alloy. The growth of the IOZ was then deduced to be due simply to oxygen diffusion within its Ferich matrix. Similar observations have recently been reported [6] in modified NdFeB alloy variants containing Co, Dy, Zr and V. Although the internal oxidation rate and mechanisms are fairly well known in the case of NdFeB-based alloys, little oxidation data exist for the Sm2 Co17 -type magnets. Most of the literature concentrates on characterising the magnetic losses associated with high-temperature degradation of these alloys [7,8], with only a few investigators also focusing on the physical changes. These were thought not to be oxidation-related but instead due to Sm migration and surface vapourisation [e.g. [9,10]]. Recently, a study by the current authors made some detailed observations on the hightemperature degradation of two grades of Sm2 Co17 -based alloys, concluding that rapid internal oxidation is the dominant process [11]. Initial examination indicated that the IOZ penetrated the alloy with parabolic kinetics but more extensive data are presented in this paper. 2. Materials This study examines the oxidation kinetics of two grades of Sm2 (Co,Fe,Cu,Zr)17 alloy: a “standard” grade, Sm(Co0.63 Fe0.27 Cu0.08 Zr0.02 )8.35 , and a “high-temperature” grade, Sm(Co0.74 Fe0.1 Cu0.12 Zr0.04 )8.5 . Both alloys were manufactured by Precision Magnetics Ltd. The samples were supplied in rectangular coupons ∼10 mm × 10 mm × 2 mm in size of two types—firstly, with the c-axis alignment perpendicular to the major faces, and secondly, with the c-axis alignment in-plane with the major faces (see Fig. 1). This facilitated study of the oxidation zone penetration in the two different crystal directions. Throughout this work, the c-axis direction will be indicated on micrographs of alloy cross-sections by a small arrow. The microstructure of Sm2 Co17 -type alloys has been characterised more fully elsewhere [e.g. [7]], but broadly consists of Sm2 Co17 -phase cells of the order of 100 nm in size surrounded by SmCo5 -phase boundaries, both penetrated by Zr-rich lamellae with a thickness of several nm perpendicular to the c-axis. On a larger scale,


W.M. Pragnell et al. / Journal of Alloys and Compounds 473 (2009) 389–393

Fig. 1. Specimen coupons and the c-axis alignment directions (a) perpendicular and (b) in-plane.

this structure is divided into equiaxed grains approximately 40–50 ␮m across and interspersed with 2–5 ␮m particles of Sm2 O3 which form during alloy production. 3. Experimental Interrupted oxidation testing was carried out in laboratory air in a muffle furnace, at temperatures between 300 and 600 ◦ C for up to 1000 h at the lower temperatures and 350 h for the higher temperatures. For each unique combination of temperature, alloy grade and crystal orientation, six specimens were polished to an 800 grit finish, cleaned ultrasonically in ethanol and weighed using a Sartorius balance. The specimens were placed within alumina tubes approximately 15 mm in diameter and then loaded into the furnace, which was brought to temperature within 2–5 min depending on the test temperature. At intervals during the test, single specimens were removed from the furnace within their tubes and allowed to cool to room temperature, before being weighed again. When the test was complete, the furnace was turned off and the final specimen allowed to cool inside it. All specimens were mounted in MetPrep Quickset plastic and ground back by ∼1 mm using wet 120 grit SiC paper on a grinding wheel to obtain a cross-section perpendicular to the major face (sectioning with a slow diamond wheel prior to mounting proved unsuccessful due to the brittle nature of the alloys). The specimens were then polished down to 1200 grit on wet SiC paper and given a final polish with a colloidal silica suspension prior to examination using SEM. Measurements of the IOZ thickness were obtained from SEM micrographs taken from both major faces of the specimens, to ensure that the oxidation zone was thickening evenly on both sides of the specimen. The zone thickness, from the inside of the external scale to the IOZ/substrate interface, was measured at equally spaced points on each oxidation zone micrograph and averaged (Fig. 2); the error on the mean thickness was then calculated as the standard deviation of the individual measurements.

4. Results and discussion As detailed in [11], two oxidation processes are found to occur in Sm2 Co17 -type alloys: the internal oxidation of Sm, and the external oxidation Co, Fe and Cu. Sm is removed from the 2:17 and 1:5 structures by oxidation, releasing the transition elements which can then diffuse to the outer surface. The external oxide (visible as a very dark layer in Fig. 2) is typically ∼15 times thinner than

Fig. 2. Example IOZ thickness measurements on a standard-grade specimen after 75 h at 500 ◦ C.

the internal oxidation zone, and is not thought to have a significant impact on the alloy’s magnetic properties. The oxides in this layer are very porous and allow oxygen through to the substrate quite easily; subsequent oxygen diffusion through the IOZ to the unoxidised alloy bulk is therefore the rate-controlling process for the internal reaction. Fig. 3 shows an example of the mass gain kinetics measured for both grades of alloy (parallel c-axis) at 600 ◦ C. The kinetics look parabolic at first glance (3a) but when mass gain squared is plotted (3b) it can be seen that the kinetics follow a slightly sub-parabolic law. This is typical of both grades at all temperatures tested, and is undoubtedly due to the fact that there are two separate oxidation processes at work (described above). Each occurs at different rates and the observed mass gain kinetics will therefore be a combination of the two. Since this work is mainly concerned with the loss of magnetic properties following the growth of an internal reaction zone, upon which the external scale has little or no effect, mass gain is not the best way to quantify the oxidation process. Fig. 3 does, however, clearly show that the oxidation behaviour of the two alloy grades is almost identical. It was found that the IOZ extended inwards at different rates depending upon the local crystallographic c-axis orientation. Generally, internal oxidation was found to progress more quickly in the direction perpendicular to the c-axis than in the parallel direction, with a more convoluted internal interface. Co/Cu precipitates in the IOZ were also found to form with different morphologies depending on c-axis orientation. This is undoubtedly linked to the anisotropy of the microstructure, causing oxygen to diffuse more slowly in the parallel direction than in the perpendicular direction. A more detailed examination of these differences, and possible explanations for them, can be found in [11]. Fig. 4 shows two typical sets of IOZ thickness kinetics measured both perpendicular and parallel to the c-axis alignment; the error bars represent the standard deviation of the measurements. The size of the error bars reflects the convoluted nature of the IOZ inner interface, especially when measured perpendicular to the c-axis and at longer times. Fig. 5 shows the curves of  2 vs time, where  is the thickness of the IOZ, at the temperatures of 300–600 ◦ C. Note that there are only three curves for the 300 ◦ C tests due to sample preparation problems as a consequence of the brittle nature of the high-temperature grade, together with the relatively thin IOZ at this temperature. It was not considered necessary to repeat these experiments. Fig. 5 shows that although in most cases the IOZ thickening is broadly parabolic in nature, in some cases it exhibits very slight subparabolic trends similar to those shown in Fig. 3 above. This could be due to the IOZ microstructure changing during oxidation—the flake-like Co/Cu precipitates observed throughout the IOZ appear to increase in size over time and depth. This could, in turn, reduce the oxygen diffusion paths through the IOZ, which would reduce the instantaneous oxidation rate constant. Evidence that these precipitates can limit, locally, the ingress of oxygen is shown in Fig. 6

W.M. Pragnell et al. / Journal of Alloys and Compounds 473 (2009) 389–393

Fig. 3. Example mass gain kinetics for samples with c-axes in the parallel direction exposed at 600 ◦ C, (a) as measured and (b) showing mass gain squared against time to illustrate the slight sub-parabolic behaviour.

where it can be seen that IOZ development under a flake-like particle is inhibited. Nevertheless, for the exposure times considered here, the errors associated with these deviations are considered to be second order and the IOZ growth kinetics can be adequately treated as parabolic. Fig. 7 shows an Arrhenius plot of the parabolic growth rate constants (defined as  2 /t) of the IOZs formed in the two grades of Sm2 Co17 alloy (the standard errors for these data obtained from their linear regressions are listed in Table 1). The solid data points fitted by the unbroken regression line denote zone growth in the direction perpendicular to the c-axis (the faster moving, jagged interface), and the open data points fitted by the broken regression line denote zone growth in the direction parallel to the c-axis. Each line was fitted to data from both alloy grades because they exhibited such close agreement in growth rate; examination of Fig. 6 shows that the perpendicular values are consistently higher


Fig. 4. Example IOZ thickness measurements at 400 ◦ C for (a) the standard grade and (b) the high-temperature grade.

than the parallel values, and neither grade dominates the other in either orientation. The lines are quite similar in slope, indicating very similar activation energies for each orientation. Also plotted are several data points taken from the work of Chen et al. [9,12]. From [12], the point was obtained by taking the measured IOZ thickness on the relevant micrograph and assuming parabolic behaviour. Despite the differences between the various alloys studied, there is a strong agreement in internal oxidation Table 1 standard errors on measured kp values plotted in Fig. 7

Standard (⊥) Standard (| |) High-T (⊥) High-T (| |)

300 ◦ C

400 ◦ C

500 ◦ C

600 ◦ C

3.5% 5.3%

6.7% 6.2% 8.2% 3.2%

6.1% 3.8% 4.3% 2.7%

5.0% 9.8% 3.8% 1.1%



W.M. Pragnell et al. / Journal of Alloys and Compounds 473 (2009) 389–393

Fig. 5. Plots of IOZ thickness squared against time, with linear regression lines.

rate constants, indicating that the oxidation mechanism seems to be largely independent of the Fe, Cu and Zr content. The activation energies for the internal oxidation processes are 143.1 kJ mol−1 for the perpendicular direction and 143.8 kJ mol−1 for the parallel direction, allowing the rate constants to be calculated at a given temperature using kp = 3.43 × 10−5 exp(−143103/RT ) m2 s−1


and kp = 7.17 × 10−5 exp(−143742/RT ) m2 s−1

Fig. 6. SEM micrograph of the high-temperature grade alloy oxidised in the perpendicular crystallographic direction for 5 h at 600 ◦ C. The formation of the IOZ can be seen to be inhibited locally under a transverse platelet of a (Cu,Co)-rich precipitate.


for the perpendicular and parallel directions, respectively. Here T is the absolute temperature and R is 8.314 J K−1 . In practice, these activation energies are indistinguishable and confirm that the ratecontrolling process is the same in each crystallographic direction. The nature of this process is thought to be the inward diffusion of oxygen through the CoFe matrix of the IOZ, and is likely to be dominated by grain-boundary diffusion as is the case with NdFeB magnets [5]. It is interesting to note that the value of activation energy found here for the SmCo system of ∼143 kJ mol−1 is appreciably higher than those of ∼100 kJ mol−1 reported [5,6] for the

W.M. Pragnell et al. / Journal of Alloys and Compounds 473 (2009) 389–393


kinetics are reasonably parabolic, however, indicating a diffusioncontrolled process. Internal oxidation speeds vary between the different crystallographic directions, but the oxidation activation energies are found to have the same value of ∼143 kJ mol−1 for both alloy grades. Acknowledgements This work was funded by the UK’s Engineering and Physical Sciences Research Council (EPSRC) under grant reference EP/C518411/1. The authors would also like to thank Dr. Gerhard Martinek of Precision Magnetics Ltd. for supplying the magnet coupons. References

Fig. 7. Arrhenius plot of rate kinetics.

NdFeB-based system. The difference reflects the greater difficulty of oxygen diffusion within the IOZ of the SmCo alloys. 5. Conclusions Two grades of Sm2 (Co,Fe,Cu,Zr)17 -based alloys were oxidised in air at temperatures between 300 and 600 ◦ C. The resulting mass gain kinetics are sub-parabolic, due to simultaneous internal and external oxidation reactions. The internal oxidation layer thickness

¨ [1] W. Tang, Y. Zhang, A.M. Gabay, H. Kronmuller, G.C. Hadjipanayis, in: G.C. Hadjipanayis, M.J. Bonder (Eds.), Proceedings of the 17th International Workshop on Rare-Earth Magnets and their Applications, Rinton Press, 2002, p. 685. [2] R.B. Blanc, E. Adler, in: C. Herget, R. Poerschke (Eds.), Proceedings of the 9th International Workshop on Rare Earth Magnets and their Applications, Bad Soden, Germany, 1987, pp. 537–544. [3] D.S. Edgley, J.-M. Le Breton, D. Lemarchand, I.R. Harris, J. Teillet, J. Magn. Magn. Mater. 128 (1993) L1–L7. [4] D.S. Edgley, J.-M. Le Breton, S. Steyaert, F.M. Ahmed, I.R. Harris, J. Teillet, J. Magn. Magn. Mater. 173 (1997) 29–42. [5] Y. Li, H.E. Evans, I.R. Harris, I.P. Jones, Oxid. Met. 59 (2003) 167–182. [6] I. Skulj, H.E. Evans, I.R. Harris, J. Mater. Sci. 43 (2008) 1324–1333. [7] R. Gopalan, K. Muraleedharan, T.S.R.K. Sastry, A.K. Singh, V. Joshi, D.V. Sridhara Rao, V. Chandrasekaran, J. Mater. Sci. 36 (2001) 4117–4123. [8] M.S. Walmer, C.H. Chen, M.H. Walmer, IEEE T. Magn. 36 (2000) 3376–3381. [9] C.H. Chen, M.H. Walmer, E.H. Kottcamp, W. Gong, IEEE T. Magn. 37 (2001) 2531–2533. [10] S. Kardelky, A. Gebert, O. Gutfleisch, V. Hoffmann, L. Schultz, J. Magn. Magn. Mater. 290–291 (2005) 1226–1229. [11] W.M. Pragnell, A.J. Williams, H.E. Evans, J. Appl. Phys. 103 (2008) 07E127. [12] C.H. Chen, M.S. Walmer, M.H. Walmer, S. Liu, G.E. Kuhl, IEEE T. Magn. 36 (2000) 3291–3293.