Fumiaki Okabe 1; * 1, Hyun Soon Park 1, Daisuke Shindo 1; * 2, Young-Gil Park 2, Ken Ohashi 3 and Yoshio Tawara 3

Transcription

1 Materials Transactions, Vol. 47, No. 1 (2006) pp. 218 to 223 #2006 The Japan Institute of Metals Microstructures and Magnetic Domain Structures of Sintered Sm(Co 0:720 Fe 0:200 Cu 0:055 Zr 0:025 ) 7:5 Permanent Magnet Studied by Transmission Electron Microscopy Fumiaki Okabe 1; * 1, Hyun Soon Park 1, Daisuke Shindo 1; * 2, Young-Gil Park 2, Ken Ohashi 3 and Yoshio Tawara 3 1 Institute of Multidisciplinary Research for Advanced Materials, Tohoku University, Sendai , Japan 2 Institute for Materials Research, Tohoku University, Sendai , Japan 3 Shin-Etsu Chemical Co., Ltd., Takefu-shi, Fukui 915, Japan Microstructures and magnetic domain structures of precipitation-hardened Sm(Co 0:720 Fe 0:200 Cu 0:055 Zr 0:025 ) 7:5 permanent magnets obtained by various heat treatments are investigated by transmission electron microscopy (TEM). It is found that Cu atoms gradually segregate into SmCo 5 phase with the increase in aging time. The domain walls in the solution-treated, 6-h isothermal-aged magnets are straight, while those in the step-aged magnet are zigzag shaped along the cell boundaries of the SmCo 5 phases (1:5 H phases). In the demagnetized state of the step-aged magnet, it is found that the domain wall is located almost on the 1:5 H cell boundary phase containing Cu atoms, where the distribution of lines of magnetic flux strongly deviates from the axis of easy magnetization, particularly near the zigzag domain wall, and the lines of magnetic flux flow symmetrically along the center of the 1:5 H cell boundary phase. In the remanent state of the step-aged magnet, it is confirmed that domain walls are strongly pinned almost to the 1:5 H cell boundary phase containing Cu atoms, eventually resulting in a high coercivity in a Sm(Co 0:720 Fe 0:200 Cu 0:055 Zr 0:025 ) 7:5 permanent magnet. (Received August 25, 2005; Accepted November 22, 2005; Published January 15, 2006) Keywords: energy dispersive X-ray spectroscopy elemental mapping, electron holography, lines of magnetic flux, pinning of domain wall, Sm(Co,Fe,Cu,Zr) z magnet 1. Introduction Precipitation-hardened Sm(Co,Fe,Cu,Zr) z permanent magnets have attracted considerable attention due to their high coercivity and Curie temperatures (T C 1000 K). 1,2) Their magnetic properties depend sensitively on the microstructure and magnetic domain structure controlled by the sophisticated and lengthy heat treatments. 3 5) For example, by employing the step-aging method, the coercivity and maximum energy product increased drastically from 33 to 749 ka/m and from 20 to 192 kj/m 3, respectively, while there was a slight decrease in the remanence from 1.1 to 1.0 T. The high coercivity of Sm(Co,Fe,Cu,Zr) z magnets can be fundamentally explained by the pinning of the domain wall to the nanoscaled cell boundary precipitated in the cellular structure. Based on Lorentz microscope observations and micromagnetic simulation, a few models 6 9) for the pinning mechanism in Sm(Co,Fe,Cu,Zr) z magnets have been reported. Livingston et al. 6) proposed that coercivity is controlled by domain wall pinning by a SmCo 5 barrier phase; this contradicts the model proposed by Thomas or Nagel, 7,8) in which a SmCo 5 phase is an attractive pinning site. Further, it has been pointed out that the transition area from the Cuenriched SmCo 5 phase to the matrix phase (Sm 2 Co 17 ) produces a drastic reduction in the magnetocrystalline anisotropy; this energy well eventually becomes an attractive pinning site. 9) However, the quantitative evaluation of the magnetocrystalline anisotropy in the model has not been carried out in these studies. Moreover, the detailed observation of the magnetic domain structures related to the pinning * 1 Present address: TOYOTA, Hekinan , Japan * 2 Corresponding author, shindo@tagen.tohoku.ac.jp mechanism has not yet been carried out. Thus far, in the Sm(Co,Fe,Cu,Zr) z magnets, the microstructural analyses by transmission electron microscopy (TEM) 10,11) and 3-dimensional atom probe technique 12) as well as the magnetic structural analysis with Lorentz microscopy have been carried out separately. Recent reports have revealed that the Cu enrichment in the cell boundary phase (SmCo 5 phase) reduces the anisotropy field of the SmCo 5 phase in comparison with that of the Sm 2 Co 17 phase; this eventually results in the high coercivity in Sm(Co,Fe,Cu,Zr) z magnets. 2,13) On the other hand, we have reported that by step aging, the Cu atoms are segregated into the cell boundary phases and the distribution of lines of magnetic flux fluctuates considerably. Some of the experimental results have been published in the previous paper. 14) In this paper, additional microstructural analyses in Sm(Co 0:720 Fe 0:200 Cu 0:055 Zr 0:025 ) 7:5 permanent magnets were systematically performed as a function of aging time by analytical electron microscopy. Further, by Lorentz microscopy and electron holography, the domain wall position and distribution of lines of magnetic flux on the cellular structure are explored in detail. 2. Experimental Procedure Four types of specimens (shown in Fig. 1) were produced by the following procedures. Powders with a nominal composition of Sm(Co 0:720 Fe 0:200 Cu 0:055 Zr 0:025 ) 7:5 were sintered at 1478 K for 2 h, following which they were solution treated at 1363 K for 1 h and rapidly quenched to room temperature (specimen 1). The alloy was heated up to 1093 K (specimen 2) at a rate of 6.83 K/min. Subsequently, isothermal aging was carried out at 1093 K for 6 h (specimen 3).

2 Microstructures and Magnetic Domain Structures of Sintered Sm(Co 0:720 Fe 0:200 Cu 0:055 Zr 0:025 ) 7:5 Permanent Magnet Studied 219 Fig. 1 Magnetic properties of four specimens with various heat treatments. H c and (BH) max indicate the coercivity and maximum energy product, respectively. Illustration of heat treatment processing is inset. This was followed by cooling to 643 K at a cooling rate of 0.5 K/min; finally, the specimen was cooled to room temperature (specimen 4). Specimens 2 and 3 were obtained by rapid quenching from the heat treatment temperature to room temperature. Some of their magnetic properties, i.e., the coercivity H c and maximum energy product (BH) max are also shown in Fig. 1. It is noted that the coercivity H c and maximum energy product (BH) max increase drastically in the step-aged specimen 4. Thin foil specimens for TEM observation were prepared by mechanical polishing and ion milling. Nanobeam diffraction (NBD) and energy dispersive X-ray spectroscopy (EDS) analyses were carried out using a JEM-2100F TEM operated at 200 kv, which was equipped with a field emission gun (FEG) and an EDS system (Noran VoyagerIII). Elemental mapping images were obtained to analyze the distribution of additive elements with an electron probe of 0.6 nm diameter. Magnetic domain structures were investigated by Lorentz microscopy and electron holography using a JEM-3000F TEM operated at 300 kv, which was installed with a field emission gun. The detailed distribution of lines of magnetic flux can be directly visualized by electron holography. 15,16) This microscope also has a special pole piece designed for observing the magnetic domains, that is, the magnetic field at the specimen position can be reduced to less than 2 mt. 17) An ex situ experiment was carried out by using an electromagnet 18) that was designed to apply a magnetic field of up to 2.5 T to the thin films placed in the TEM specimen holder. In this experiment, the thin foil specimen observed was removed from the TEM and a strong magnetic field was applied to it. The remanent state of the specimen was then observed by inserting the specimen back into the TEM. 3. Results and Discussion Figures 2(a) and (b) show the bright-field images and electron diffraction patterns (insets) obtained for specimens 2 and 4 (step-aged), respectively. In the electron diffraction patterns, strong reflections such as indexed ones originate from both SmCo 5 (hexagonal CaCu 5 type, denoted as 1:5 H Fig. 2 (a), (b) Bright-field images and selected-area electron diffraction patterns (insets) obtained from specimens 2 and 4, respectively. (c), (d) Dark-field images obtained with the superlattice reflection of 2:17 R (arrows) in the diffraction patterns and NBD patterns (insets) obtained from the 1:5 H and 2:17 R phases, respectively. The incident electron beam is parallel to the ½1120Š 2:17 direction in both cases.

3 220 F. Okabe et al. Fig. 3 (a), (b) Scanning transmission electron microscope (STEM) images and EDS elemental mapping images showing elemental distribution (Cu,Fe,Zr) in specimens 3 and 4, respectively. The 2:17 R, 1:5 H, and Z phases are indicated by arrows in the STEM images in (a) and (b). hereafter) and Sm 2 Co 17 (rhombohedral Th 2 Zn 17 type, denoted as 2:17 R hereafter) phases, 19) while the reflections indicated by the pair of white arrows between the strong reflections correspond to the superlattice reflections of the 2:17 R phase. Figures 2(c) and (d) show the dark-field images obtained by using the superlattice reflections of the 2:17 R phase indicated by white arrows in each pattern. The darkfield images clearly show the cellular structure, that is, the cell interiors and cell boundaries corresponding to the 2:17 R and 1:5 H phases, respectively. A third phase of planar thin plates perpendicular to the c-axis, which has been termed the Z phase because it is enriched with Zr with respect to the matrix, is also observed, and the Z phase is considered to possess a SmCo 3 structure with a 0:5 nm and c 2:4 nm. 19) The crystal structure of the two main phases (1:5 H and 2:17 R) in both specimens was also investigated by the NBD method, as shown in the insets of Figs. 2(c) and (d). Based on the dark-field images and NBD patterns, it is observed that the cell size increases by step aging; however, no changes are observed in the cellular and crystal structures of the two main phases. Figures 3(a) and (b) show the EDS elemental mapping images of specimens 3 (6-h isothermal-aged) and 4 (stepaged). In both specimens, it is observed that the Fe content is lower in the Z phase and 1:5 H phase, while Zr and Cu atoms are segregated into the Z and 1:5 H phases, respectively. Further, it is noted that the content of Cu atoms in the 2:17 R phase of the step-aged specimen (b) is lower than that in the 2:17 R phase of the 6-h isothermal-aged specimen (a). Thus, it is considered that the step-aging treatment leads to Cu segregation into the 1:5 H cell boundary phases without any appreciable microstructural changes. Figure 4 shows the Lorentz microscope images of four specimens obtained in the Fresnel mode. All the images were observed in the demagnetized state. The 180 domain walls nearly parallel to the c-axis or the axis of easy magnetization are clearly visible in all the images. In Figs. 4(b) (c), the magnetic domain walls are straight and tend to pass through the centers of the cellular structures, while the domain walls in Fig. 4(d) appear zigzag along the 1:5 H cell boundary phase. It is observed that the coercivity in specimens 2 and 3 is considerably low, as shown in Fig. 1. Therefore, it can be concluded that the 1:5 H cell boundary phase acts as a repulsive pinning center for the domain walls in specimens 2 and 3, while it acts as an attractive pinning center for the domain walls in specimen 4. Figures 5(a) (c) exhibit the reconstructed phase images of specimens 1, 3, and 4 in the demagnetized state. Although these images include the phase information due to the inner potential with the variation in the specimen thickness, a change in the distribution of lines of magnetic flux with different heat treatments can be investigated by comparing these images because the specimen thickness variation is similar. In the reconstructed phase images, the direction (black arrows) and density of the white lines essentially correspond to the direction and density of lines of magnetic flux projected along the incident electron beam, respectively. In Figs. 5(a) and (b), the distribution of lines of magnetic flux is monotonous, and the magnetic domain wall forming the 180 domain wall is nearly parallel to the axis of easy magnetization (c-axis) of the main 1:5 H and 2:17 R phases. It is noted that despite the presence of the 180 domain wall, the direction of lines of magnetic flux forms about 90 due to the surface magnetic charge at the specimen edge. On the other hand, in the step-aged specimen of Fig. 5(c), the direction of lines of magnetic flux strongly deviates from the axis of easy magnetization. In the region indicated by the asterisk in Fig. 5(c), the density of lines of magnetic flux is considerably low; this is in sharp contrast to the situation in Figs. 5(a) and (b). This difference results from the magnetic charge that may form along the zigzag domain wall, depending on the angle between the magnetization directions. The domain wall position and distribution of lines of magnetic flux on the cellular structure in a step-aged specimen are investigated in detail. Figure 6 shows bright-

4 Microstructures and Magnetic Domain Structures of Sintered Sm(Co 0:720 Fe 0:200 Cu 0:055 Zr 0:025 ) 7:5 Permanent Magnet Studied 221 Fig. 4 (a), (b), (c), (d) Lorentz microscope images of specimens 1, 2, 3, and 4, respectively. The images were obtained in the Fresnel mode. White arrows and black arrowheads indicate c-axes and domain walls, respectively. Fig. 5 (a), (b), (c) Reconstructed phase images of specimens 1, 3, and 4, respectively. White and black arrows indicate the axes of easy magnetization (c-axes) and the direction of lines of magnetic flux. field, Lorentz microscope, and reconstructed phase images of the same area in the step-aged specimen. The domain wall contrast [bright bands between the white arrows in Fig. 6(b)] appears on the 1:5 H cell boundary phase indicated by white arrows in Fig. 6(a). It can be seen that the direction of lines of magnetic flux strongly deviates from the axis of easy magnetization, especially near the zigzag domain wall, and the lines of magnetic flux flow symmetrically along the 1:5 H cell boundary phase. Taking into account the results of analytical electron microscopy shown in Fig. 3, it is considered to a reasonable extent that the characteristic distribution of lines of magnetic flux results from the segregation of Cu atoms into the 1:5 H cell boundary phase in the step-aged magnet. This feature clearly indicates a strong and attractive domain wall pinning to the 1:5 H cell boundary phase. Furthermore, it is noteworthy that the center of the domain wall is at the center of the 1:5 H cell boundary phase. In previous reports, 6 9) the pinning was considered to result from the decrease in the magnetocrystalline anisotropy with the segregation of Cu atoms into the 1:5 H cell boundary and/or the interface regions between the 1:5 H and 2:17 R phases. Taking into account the present results, it should be pointed out that a single interface region between the 1:5 H and 2:17 R phases does not solely act as a pinning center.

5 222 F. Okabe et al. Fig. 6 Bright-field image (a), Lorentz microscope image (b), and reconstructed phase image (c) of the same area in specimen 4 (stepaged). White arrows indicate the cell boundary and domain wall. Fig. 7 Lorentz microscope images (upper parts) and reconstructed phase images (lower parts) observed in the demagnetized [(a) and (c)] and remanent [(b) and (d)] states of specimen 4 (step-aged). White and black dotted lines indicate the domain walls marked with black arrowheads. Black arrows indicate the direction of lines of magnetic flux. Further domain wall analysis at a higher resolution and quantitative evaluation of the magnetocrystalline anisotropy are necessary to clarify the individual contributions of the cell boundary and the interface regions to the domain wall pinning. In order to confirm the attractive pinning force of the 1:5 H cell boundary phase for domain wall motion, an external magnetic field of 1.0 T was applied to the step-aged specimen (specimen 4), as shown in Fig. 7(b). The black arrowheads and arrows indicate the domain walls and direction of lines of magnetic flux, respectively. In the remanent state shown in Fig. 7(b), the left domain wall appears to be strongly pinned almost to the 1:5 H cell boundary phase, although the right domain wall has slightly changed. It is noted that the left domain wall in Fig. 7(b) is the same as the domain wall in Fig. 6(b). Therefore, it is believed that the 1:5 H cell boundary phase containing Cu atoms acts as a strong pinning center for domain wall motion. It eventually results in a high coercivity in sintered Sm(Co 0:720 Fe 0:200 Cu 0:055 Zr 0:025 ) 7:5 permanent magnets. 4. Conclusion The precipitation-hardened Sm(Co 0:720 Fe 0:200 Cu 0:055 Zr 0:025 ) 7:5 magnets obtained by various heat treatments were characterized by analytical electron microscopy, Lorentz microscopy, and electron holography. The results are summarized as follows. (1) From microstructural analysis, it is found that the cell boundary phase forms at an early stage, and microstructures, i.e., cellular structures and crystal structures of the cell boundary and matrix, essentially remain unchanged during the various heat treatments. (2) Through EDS elemental mapping by analytical electron microscopy, the segregation of Cu atoms into the 1:5 H cell boundary phase proceeds gradually with the increase in aging time; however, the coercivity increases drastically only after the step aging process. (3) In the solution-treated, 6-h isothermal-aged magnets, the magnetic domain walls are straight and pass through the cell interiors of the 2:17 R phases, while domain walls in the step-aged magnet appear to be zigzag along the 1:5 H cell boundary phase. (4) In the demagnetized state of the step-aged magnet, it is found that the domain wall is located almost on the 1:5 H cell boundary phase containing Cu atoms, where the distribution of lines of magnetic flux strongly deviates from the axis of easy magnetization, especially near the zigzag domain wall, and the lines of magnetic flux flow symmetrically along the cell boundary. (5) In the remanent state of the step-aged magnet, it is confirmed that the domain walls are strongly pinned to the 1:5 H cell boundary phase containing Cu atoms, resulting in a high coercivity in a Sm(Co 0:720 Fe 0:200 - Cu 0:055 Zr 0:025 ) 7:5 permanent magnet. Acknowledgments The authors thank Dr. Y. Gao, IMRAM, Tohoku University for his helpful discussion. This work was partly supported by financial assistance from the Special Coordination Funds for Promoting Science and Technology on

6 Microstructures and Magnetic Domain Structures of Sintered Sm(Co 0:720 Fe 0:200 Cu 0:055 Zr 0:025 ) 7:5 Permanent Magnet Studied 223 Nanohetero Metallic Materials from the Science and Technology Agency and also from a Grant-in-Aid for Scientific Research (A) from the Japan Society for the Promotion of Science. REFERENCES 1) J. Fidler, P. Skalicky and F. Rothwarf: IEEE Trans. Magn. 19 (1983) ) G. C. Hadjipanayis, W. Tang, Y. Zhang, S. T. Chui, J. F. Liu, C. Chen and H. Kronmüller: IEEE Trans. Magn. 36 (2000) ) J. Fidler and P. Skalicky: IEEE Trans. Magn. 17 (1981) ) G. C. Hadjipanayis, R. C. Hazelton, K. R. Lawless and L. S. Horton: IEEE Trans. Magn. 18 (1982) ) K. J. Strnat: Ferromagnetic Materials, ed. by E. P. Wohlfarth and K. H. J. Buschow (Elsevier Science, Amsterdam, 1988), Vol. 4, pp ) J. D. Livingston and D. L. Martin: J. Appl. Phys. 48 (1977) ) R. K. Mishra and G. Thomas: Proc. of the 4th International Workshop on Rare Earth-Cobalt Permanent Magnets and Their Applications. Hakone (Japan) (2002), ) H. Nagel: J. Appl. Phys. 50 (1979) ) D. Goll, H. Kronmüller and G. C. Hadjipanayis: Proc. of the 17th International Workshop on Rare-Earth Magnets and Their Applications. ed. by G. C. Hadjipanayis and M. J. Bonder, (Delaware (USA), Rinton Press, 2002) pp ) K. N. Melton and R. S. Perkins: J. Appl. Phys. 47 (1976) ) K. Hiraga, M. Hirabayashi and N. Ishigaki: J. Microsc. 142 (1986) ) X. Y. Xiong, T. Ohkubo, T. Koyama, K. Ohashi, Y. Tawara and K. Hono: Acta Mater. 52 (2004) ) H. Kronmuller and D. Goll: Scr. Mater. 47 (2002) ) Y.-G. Park, J. M. Yang and D. Shindo: J. Electron Microsc. 53 (2004) ) A. Tonomura: Electron Holography, 2nd Ed., (Springer-Verlag, 1999) pp ) D. Shindo and T. Oikawa: Analytical Electron Microscopy for Materials Science, (Springer-Verlag, 2002) pp ) D. Shindo, Y.-G. Park, Y. Murakami, Y. Gao, H. Kanekiyo and S. Hirosawa: Scr. Mater. 48 (2003) ) D. Shindo, Y.-G. Park, Y. Gao and H. S. Park: J. Appl. Phys. 95 (2004) ) L. Rabenberg, R. K. Mishra and G. Thomas: J. Appl. Phys. 53 (1982)