Feptp With High Coercivity

APPLIED PHYSICS LETTERS

VOLUME 85, NUMBER 5

2 AUGUST 2004

Rapidly solidified „FePt…70P30 alloy with high coercivity A. A. Kündig,a) N. Abe, M. Ohnuma, T. Ohkubo, H. Mamiya, and K. Hono National Institute for Materials Science, 1-2-1 Sengen, Tsukuba 305-0047, Japan

(Received 1 March 2004; accepted 2 June 2004) The alloy Fe35Pt35P30, whose composition is close to a ternary eutectic, was rapidly solidified by melt spinning and the ribbon exhibited a high coercivity, exceeding 20 kOe after annealing. The alloy was mainly composed of L10 ordered FePt and PtP2 with an average grain size of about 50 nm. In the as-cast state, the alloy was comprised of about 20-nm-diam FePt and PtP2 grains supersaturated with P and Fe, respectively, and the coercivity was only 150 Oe. The high coercivity obtained following annealing is discussed on the basis of the microstructural observations. © 2004 American Institute of Physics. [DOI: 10.1063/1.1776333] Ordered FePt with the L10 structure exhibits a large magnetocrystalline anisotropy1 and is therefore considered to be a good candidate for high density magnetic recording applications. The recent study by Shima et al. on FePt islandlike films epitaxially grown on MgO single crystalline substrates reported a large coercivity of 40 kOe from a very thin film 共20 nm兲.2 However, the coercivity drastically drops below 2 kOe when the island-like films are percolated beyond a critical film thickness.3 This is because of the single crystalline nature of the percolated epitaxially grown film, in which no domain wall pinning sites are present. On the other hand, relatively a large coercivity exceeding 13 kOe was recently reported for a polycrystalline 100-nm-thick film.4 This suggests that grain boundaries will be effective in enhancing the coercivity of FePt thin films. A further increase of the coercivity is expected if a nonmagnetic phase is introduced, which can act as a domain wall pinning site. The hard magnetic property of FePt was first reported in the bulk alloy by Watanabe and Masumoto in 1983.5 Although this is an expensive alloy, due to its excellent corrosion resistance, high Curie temperature, and high energy product, certain applications in the dental area as well as in micromachines are expected. However, the largest coercivity reported from bulk FePt was only 8 kOe, in spite of its large anisotropy field of 120 kOe. The low coercivity obtained in the bulk alloy is due to its high order–disorder temperature of 1300° C, because of which as-cast or as-quenched microstructures show coarse ordered domains without many domain wall pinning sites within the L10 phase.6 A recent approach to reducing the grain size of ordered FePt by cold deformation7 extended the coercivity of bulk FePt up to 10 kOe. This result indicates that bulk samples with a coercivity higher than 10 kOe can be obtained by controlling the microstructure of the L10 FePt phase. The aim of this work is to control the microstructure by a metallurgical approach and to prepare cast samples with high coercivity. To obtain a fine microstructure in the as-cast condition, alloy compositions near the ternary Fe– P – Pt eutectic were chosen. Since the melting temperature is reduced at the eutectic composition, the grain size can be refined by rapidly solidifying the melt, often into the nanometer range. If thermodynamic conditions are met, even an amorphous phase can be frozen in near an eutectic composition (e.g., a)

Electronic mail: [email protected]

Zr– Cu– Al8). As FePt alloys are currently used in dentistry9 and have potential biological applications, the exploration of a ternary eutectic alloy containing only biocompatible elements is desirable. For these reasons, phosphorus has been selected as the third element, although no data were available for the ternary Fe– P – Pt system. However, both the Fe– P and Pt– P binary systems show deep eutectics at 17 and 20 at. % P, respectively. Prelimary scanning electron microscopy (SEM) investigations of a series of Fe– Pt– P ternary alloys had indicated that a ternary eutectic is present at about 1180 K close to the composition Fe30Pt40P30; this is comprised of the three phases, Fe45Pt55, Fe2P, and nonmagnetic P2Pt, as determined from compositional analyses by energy dispersive x-ray spectroscopy in the SEM. In the present work, we report the microstructure and magnetic properties of melt-spun Fe35Pt35P30 ribbons, the composition of which is close to the ternary eutectic. Ingots of the alloy were prepared by induction melting Fe, Pt and P with purities of 99.99% or higher in a sealed quartz tube under high vacuum. Rapidly solidified ribbons were prepared from the ingot using a single roller melt spinning unit at a wheel surface speed of 40 m / s in an atmosphere of 6 ⫻ 104 Pa He. The as melt-spun ribbons were annealed at 873 K for 10 s under a high vacuum using an infrared furnace. Figure 1 shows x-ray diffraction (XRD) patterns from the as-quenched and annealed samples obtained using Cr K␣

FIG. 1. X-ray diffraction patterns for Fe35Pt35P30 ribbon in the as-quenched state and following annealing at 873 K for 10 s.

0003-6951/2004/85(5)/789/3/$20.00 789 © 2004 American Institute of Physics Downloaded 03 Aug 2004 to 147.46.24.156. Redistribution subject to AIP license or copyright, see http://apl.aip.org/apl/copyright.jsp

790

Appl. Phys. Lett., Vol. 85, No. 5, 2 August 2004

Kündig et al.

FIG. 2. Magnetization curves for Fe35Pt35P30 ribbon in the as-quenched state and following annealing at 873 K for 10 s. The inset shows the initial magnetization curve for the annealed sample up to 20 kOe.

radiation. The two phases PtP2 (cubic) and ordered FePt (tetragonal L10-type) are present in both samples. The asquenched sample shows two additional minor peaks at the shoulders of the main peak of FePt (111), which can be attributed to Fe2P (hexagonal). Although only a small difference is observed in the XRD profiles, the magnetic properties changed drastically by annealing. Figure 2 shows the magnetization curves of the as-quenched and annealed samples of Fe35Pt35P30. The coercivity of the as-quenched sample is only 150 Oe while that of the annealed sample is 20 kOe and the magnetization cannot be fully saturated even in a field of 90 kOe. The saturation magnetization of the as-quenched sample is 180 emu/ cc, while that of the annealed sample is 400 emu/ cc at 90 kOe. Since the saturation magnetization of single phase FePt is 1100 emu/ cc, it is assumed that either the volume fraction of the FePt phase in the annealed sample is less than 50% or the FePt phase is not of stoichiometric composition. Figure 3 shows bright field transmission electron microscopy (TEM) images and selected area electron diffraction (SAED) patterns for the as-quenched and the annealed samples. Coarsening of the microstructure occurs during annealing, i.e., the average grain size increased from 20 to 50 nm. For the annealed sample, all the rings can be indexed with L10-FePt and PtP2 phases. However, as three phases are expected in a ternary eutectic composition, this sample has been further investigated by nano-beam diffraction (not shown). From this more detailed analysis, a third phase was found to be present, which has very similar lattice spacings to the FePt and PtP2 phases so that its main peaks are hidden in the SAED pattern as well as in the XRD profile. However, since no additional peak from this phase can be observed in either the XRD or the SAED patterns, it is assumed that its concentration is low. Figure 4 shows Fe maps for the as-quenched and annealed Fe35Pt35P30 ribbons imaged from the Fe-L3,2 edge using the three-windows technique on a Gatan imaging filter. In the as-quenched sample, the Fe mapping gives a rather uniform contrast. Nano-beam energy dispersive x-ray spectroscopy (EDS) revealed that there were three different concentration areas corresponding to FePt, PtP2, and Fe2P

FIG. 3. TEM images for Fe35Pt35P30 ribbon in the as-quenched state (a) and following annealing at 873 K for 10 s (b).

phases, all of which contain a large amount of the third element in solution (more than 10% each). This indicates that Fe is distributed relatively uniformly, regardless of the phases. Nevertheless, some Fe depleted regions are also observed. In the annealed sample, however, a clear contrast appears in the Fe map. The brightness of the Fe map indicates that there are three levels of Fe concentration, in agreement with the NBD results. Nanobeam EDS on this sample showed that the FePt [gray area in Fig. 4(b)] and PtP2 (dark area) phases are almost free of P and Fe, respectively. The

FIG. 4. Energy filtered TEM images measured by the three windows technique at the Fe edge for Fe35Pt35P30 ribbon in the as-quenched state (a) and following annealing at 873 K for 10 s (b). Downloaded 03 Aug 2004 to 147.46.24.156. Redistribution subject to AIP license or copyright, see http://apl.aip.org/apl/copyright.jsp

Kündig et al.

Appl. Phys. Lett., Vol. 85, No. 5, 2 August 2004

third unidentified phase turns out to be an Fe-rich Fe– P phase containing a few at. % Pt (bright area). Note that the gray area 共FePt兲 in Fig. 4(b) is interconnected, while the other phases are generally isolated. The grain size of FePt is estimated to be approximately 50 nm, which is in the same range as typical islands of FePt observed in highly coercive FePt thin films.2 The critical size for single domain particles 共Dc兲 can be estimated to be approximately 380 nm for isolated spherical FePt particles10 using Dc =

24冑AKu NM s2

with Ku the uniaxial anisotropy energy and M s the saturation magnetization for FePt and N = 4␲ / 3 the demagnetization factor for spherical particles. A = AexS2 / alattice is an intrinsic constant with Aex the exchange stiffness constant, S the spin quantum number, and alattice the lattice constant. Hence, the FePt grains in the annealed sample are thought to be interactive single domain magnetic particles. The grain boundaries of randomly oriented FePt single domain particles will act as effective pinning sites for the magnetic domain wall motion. In addition, PtP2 nonmagnetic particles will also work as pinning sites, as its diameter 共⬃50 nm兲 is much lager than that of the estimated magnetic domain wall width 共⬃4 nm兲 for L10-FePt. Compared to the as-quenched sample, the PtP2 phase is almost Fe-free, which suggests that the PtP2 phase is not ferromagnetic after annealing. It is noted that the initial magnetization curve for the annealed sample is not of the typical pinning type, which might be due to the effect of the Fe-rich phase. In spite of the presence of the L10-FePt phase, the as-quenched sample shows only weak contrast in Fe concentration and the PtP2 phase does not appear to act as a strong pinning site. This may be the reason why the as-quenched sample shows a low coercivity. In FePt epitaxially grown thin films,2 a coercivity of 40 kOe was obtained at room temperature. The present coercivity is approximately half of this high coercivity reported in ideally isolated FePt single domain particles which are perfectly aligned in the direction normal to the film. It should be noted that the coercivity of 20 kOe 共1600 kA m−1兲 is almost the same as that obtained from randomly oriented FePt isolated particles that were grown on amorphous SiO2 substrates.4 This value is much higher than that 共13 kOe兲 obtained from interconnected polycrystalline FePt thin films. This indicates that the pinning effect of the nonmagnetic

791

phase 共PtP2兲 is more effective than that of the FePt polycrystalline grain boundaries. This suggests that the introduction of strong domain wall pinning sites is important for enhancing the coercivity of the FePt alloy. The present results show that the concept of reducing the grain size and introducing magnetic domain wall pinning sites in a casting process is effective in inducing a high coercivity in bulk FePt alloy (bulk in respect to thin films). The coercivity of 20 kOe is the largest ever reported for a bulk FePt-alloy sample. A high coercivity of approximately 5 kOe was also achieved from a 1-mm-thick plate processed by injection casting, indicating that the high coercivity can be obtained with the same principle not only from melt-spun ribbons but also from real bulk samples. Unfortunately, the magnetization is not sufficiently large to yield a high energy product. However, by controlling the volume fraction of the third, Fe-rich phase, the magnetization and the energy product will be improved by making use of its contribution to the magnetization in the alloy. The three phases present in this alloy all have a beneficial effect on the magnetic properties, i.e., FePt for the large anisotropy, PtP2 for pinning the magnetic domain wall, and the Fe-rich Fe– P phase for enhancing the magnetization, and their optimal combination can be a useful design tool for high performance magnets. This work was partly supported by the Special Coordination Funds for Promoting Science and Technology on “Nanohetero Metallic Materials” from the Ministry of Education, Culture, Sports, Science and Technology. A.K. greatly acknowledges support by the Japanese Society for the Promotion of Science through a JSPS fellowship. 1

O. A. Ivanov, L. V. Solina, V. A. Demishima, and L. M. Magat, Phys. Met. Metallogr. 35, 81 (1973) T. Shima, K. Takanashi, Y. K. Takahashi, and K. Hono, Appl. Phys. Lett. 81, 1050 (2002). 3 K. Watanabe and H. Masumoto, J. Jpn. Inst. Met. 48, 930 (1984). 4 Y. K. Takahashi, T. O. Seki, K. Hono, T. Shima, and K. Takanashi, J. Appl. Phys. 96, 475 (2004). 5 K. Watanabe and H. Masumoto, Trans. Jpn. Inst. Met. 24, 627 (1983). 6 Y. Tanaka, N. Kimura, K. Hono, K. Yasuda, and T. Sakurai, J. Magn. Magn. Mater. 170, 289 (1997). 7 N. H. Hai, N. M. Dempsey, and D. Givord, J. Magn. Magn. Mater. 262, 353 (2003). 8 A. Inoue, D. Kawase, A. P. Tsai, T. Zhang, and T. Masumoto, Mater. Sci. Eng., A 178, 255 (1994). 9 T. Nakayama, M. Watanabe, M. Homma, T. Kanno, K. Kimura, and O. Okuno, J. Magn. Soc. Jpn. 21, 377 (1997) (in Japanese). 10 G. Q. Li, H. Takahoshi, H. Ito, H. Saito, S. Ishia, T. Shima, and K. Takanashi, J. Appl. Phys. 94, 5672 (2003). 2

Downloaded 03 Aug 2004 to 147.46.24.156. Redistribution subject to AIP license or copyright, see http://apl.aip.org/apl/copyright.jsp