Microstructure and Magnetic Property of Ferromagnetic Fe-Pd-Rh Alloys
Yin-Chih Lin and Hwa-Teng Lee
Department of Mechanical Engineering, National Cheng Kung University, Taiwan (Received 15 July 2007)
This study shows the appearance of an intermediate martensitic structure between the fcc ! L10 martensitic transformation in aged Fe-30Pd-4Rh alloys. The intermediate phase (L1m) has a
monoclinic structure with lattice parameters of a = 3.193 A, b = 3.684 A, c = 3.141 A and = 92.042, as con rmed by transmission electron microscopy (TEM) and X-ray di raction. The crystal
orientation between L1m and L10 can be demonstrated as [101]L10//[100]L1m. This observation
suggests that the observed intermediate L1mmonoclinic phase may be an adaptive martensite. The
magnetization (M) versus magnetic eld (H) M-H curve measured at temperatures of 50, 200 and 350 K of the alloys, which were rst solution-treated (ST) and then thermally aged at 450C for
100 hours, revealed an abrupt drop at the saturation remanence (Mr). This result indicates that two phases exist in the aged alloys and that the saturation remanences of these two phases di er. The two phases, i.e., the adaptive L1mmonoclinic phase and the ordered L10martensitic structure,
were con rmed by TEM and X-ray studies.
PACS numbers: 75.50.Cc
Keywords: Fe-Pd-Rh alloys, Adaptive L1mstructure, TEM, X-ray di raction, Magnetic property
I. INTRODUCTION
The adaptive structure is a metastable phase alter-native to the normal martensitic phase. The adap-tive phase, which cannot exist in the stress-free uncon-strained state, forms only in a conuncon-strained state, when the formation of a more stable normal martensite is sup-pressed by a large transition-induced elastic energy. The crystal lattice of the adaptive phase is derived from that of the normal martensite phase by crystal-plane shuf- ing. The main geometrical feature of the crystal lattice of the adaptive phase is that it is related to the parent phase lattice by an invariant plane crystal-lattice rear-rangement. The smaller the twin surface energy and the greater the crystal-lattice mismatch, the easier it is to form an adaptive martensite. In the Fe-Pd alloy system, the adaptive martensite exists only within a narrow tem-perature range above the temtem-perature of the martensitic transformation. In this case, the formation of the adap-tive structure can be perceived as a pre-martensitic phe-nomenon. The crystal-lattice structure of the adaptive martensite is very sensitive to the applied stress and the appearance of the adaptive phase seems to occur just be-low the second-order transition. Therefore, the adaptive martensitic structure should generate extra di raction spots related to accommodation shuing of the crystal plane. The incommensurate position of its di raction spots is caused by random faulting of the periodic
dis-E-mail: [email protected]; Fax: +886-6208-5103
tribution of the shuing plane; such defects, like the formation of stacking faults in the fcc lattice, lead to a shift of the di raction spots from their regular positions. The value of the shift of the di raction spots depends on the density of defects and, thus, is determined by the crystal-lattice mismatch. The incommensurate positions of the di raction spots of the randomly faulted adaptive phase are not the same in all Brillouin zones [1,2].
The primary purpose of the present study is to provide insight into the phase transformation and the magnetic properties of aged Fe-30Pd-4Rh ferromagnetic alloys by use of selected area di raction patterns (SADP) of transmission electron microscopy (TEM), X-ray di rac-tion and a superconducting quantum interference device (SQUID). There is evidence of an intermediate L1m mon-oclinic structure, which should be formed between the fcc ! L10 martensitic transformation. This observation suggests that the observed intermediate L1m phase is an adaptive martensite. The crystal structure identi -cation and the crystal orientation between the adaptive L1mphase and the normal L10martensite are con rmed by SADP of TEM while the lattice parameters of the L1mand the L10structures are obtained by calculations based on the X-ray di raction d-spacing in association with the measured SADP of the TEM image. After the aging treatment, the magnetic property of the alloys re-vealed an abrupt drop at the saturation remanence (Mr). This phenomenon indicates that two phases exist in the aged alloy and that the saturation remanences of these two phases are di erent.
-1410-II. EXPERIMENTS
The Fe-30Pd-4Rh (at%) alloys were formed by melting pure electrolytic iron (99.9 %), pure palladium (99.95 %) and pure rhodium (99.95 %) in an arc vacuum furnace under a controlled protective argon atmosphere. The cast ingot was sealed in an evacuated quartz capsule, homogenized at 1050 C for 60 hours and subsequently hot and cold forged to a thickness of 2 mm. After forging, the specimens were sliced and sealed in an evacuated quartz capsule again. Then, they were solution-treated at 950C for 1.5 hours and quenched in ice water at room temperature. The aging treatment was performed in the temperature range of 400 { 550 C for various amounts of time.
Thin foils for TEM were prepared by double jet elec-tropolishing in a solution containing 82 % acetic acid, 9 % perchloric acid and 9 % methanol at a temperature in the range of {7 C to 10C for a current density from 2 A/cm2 to 4 A/cm2. Transmission electron microscopy (TEM), with a double tilt stage, was performed in an analytical-type high-resolution electron microscope (Hi-tachi HF-2000) with a eld emission gun operated at 200 kV and in a JEM-2100F TEM operated at 200 kV. The X-ray di raction patterns were detected at room temperature by using an X-ray di ractometer (Siemens D5000 Karlsruhe) with Cu-K radiation and the di rac-tion angles were in the 2 ranges from 35 to 140. The magnetic property measurements were carried out with a superconducting quantum interference device (SQUID) magnetometer. The magnetization versus magnetic eld (M H) curves for the samples were measured at 50, 200 and 350 K with a maximum applied eld of 30000 Oe.
III. RESULTS AND DISCUSSION
1. fcc Phase Transformation into the Adaptive
L1m Phase+L10 Structure Observed by Using TEM
An essential TEM selected area di raction pattern with a zone axis [101]L10//[100]L1mof the Fe-30Pd-4Rh alloys solution treated (ST) at 950C for 1.5 h, quenched in ice water and then thermally aged at 450 C for 100 h is shown in Figure 1(a) (hkl denotes the ordered L10 phase; hkl denotes the adaptive L1m monoclinic struc-ture). On the basis of the di raction pattern analysis, an extra di raction spot at the (022)L1m, (040)L1mposition can be seen in the SADP micrograph, which is related to accommodation shuing of the crystal planes [1, 2]. This e ect is caused by random faulting of the periodic distribution of the shuing plane, which is similar to the formation of stacking faults in the fcc lattice in that the results will lead to a shift of the di raction spots from their regular positions. The amount of the shift depends on the density of defects and, thus, is determined by us-ing the crystal-lattice mismatch. If there is a certain ran-domness in the fault distribution, then the position of the
Fig. 1. TEM micrographs of the Fe-30Pd-4Rh alloys solu-tion treated (ST) at 950C for 1.5 h, quenched in ice water
and thermally aged at 450 C for 100 h: (a) SADP of the
zone axis [101]L10==[100]L1m (hkl denotes the ordered L10
martensitic structure; hkl denotes the adaptive L1m
mono-clinic phase), (b) DF image of the adaptive (022)L1m
mon-oclinic L1m re ection corresponding to (a), (c) DF image of
the ordered (111)L10L10 re ection corresponding to (a) and
(d) the BF image.
di raction spots is incommensurate [1,2]. By calculating the X-ray di raction d-spacing in association with SADP measurements, we found that the adaptive L1m phase had a monoclinic structure with lattice parameters of a = 3.193 A, b = 3.684 A, c = 3.141 A and = 92.042and that the ordered L10 martensitic structure had lattice parameters of a = 3.876 A, c = 3.684 A and c/a = 0.950. Figure 1(b) is a dark eld (DF) image formed using the adaptive (022)L1m monoclinic re ection corresponding to Figure 1(a). The DF image of Figure 1(b) reveals that the intermediate L1m monoclinic phases are com-prised of antiphase boundaries (APBs) and microtwins. These planar defects strongly support the mechanism of coercivity in the aged Fe-Pd-Rh alloy system; it has tended to favor APBs and microtwins pinning the mag-netic domain wall motion, which has been demonstrated as a possible source of magnetic hardening [3{6]. A dark eld image using the (111)L10 re ection, corresponding to Figure 1(a), is shown in Figure 1(c). Figure 1(d) is a bright eld (BF) image. The (022)L1m adaptive L1m monoclinic phase (gray contrast) shuing the ordered L10 martensitic (111)L10 crystal plane (bright contrast) can be seen in the BF image.
2. X-ray Di raction Pattern Analysis
Figures 2(a)-(c) represent a series of X-ray di raction patterns of the alloys ST and thermally aged at 450
Fig. 2. X-ray di raction patterns for the alloys solution treated (ST) and for alloys ST and then thermally aged at 450 C and 550C for given times (a) 950 C for 1.5 h ST
and quenched in ice water, (b) ST and aged at 450 C for
100 h and (c) ST and aged at 550C for 110 h (
bct denotes
the body-centered tetragonal martensite, L1m denotes the
adaptive L1mmonoclinic phase and L10 denotes the ordered
L10 martensitic structure).
C and 550 C for various times. The X-ray di rac-tion pattern of the alloys ST at 950 C for 1.5 h and quenched in ice water is shown in Figure 2(a), where the re ections of the two phases (101)L1m and (111)L10 are the main di raction peaks. Meanwhile, the (111) peak also appears, but there is no observable bct peak in the X-ray di raction pattern. The X-ray experi-ment results show that the addition of Rh to the Fe-Pd alloy system will enhance formation of the L1mand the L10 phases. In the alloy aged at 450 C for 100 h, many adaptive L1m monoclinic phases and ordered L10 martensitic structures appear in the X-ray di rac-tion patterns in addidi rac-tion to the (111) phase separa-tion to the (110) bct+L1m+L10 structures, as shown in Figure 2(b). It is also interesting to note that the X-ray di raction peak for the plane (002)L10, (202)L10, (113)L10 re ections corresponding to the ST specimen of Figure 2(a) have been transformed into a tetragonal splitting peak (200)L10; (002)L10, (220)L10; (202)L10and (131)L10; (113)L10re ections in the specimen aged at 550 C for 110 h, as shown in Figure 2(c). These splitting peaks are caused by the formation of completely ordered tetragonal L10 martensitic structures attendant on the
L1m! L10phase transformation reaction.
When the L10tetragonal splitting peaks of Figure 2(b) are compared with those of Figure 2(c), it is apparent that the axial ratio c/a of the ordered L10 martensite gradually decreased as the aging temperature was low-ered to 450C. On the other hand, a degree of tetragonal-ity developed in the specimen aged at temperatures lower than 450 C, which gave rise to an increased anisotropy and led to an enhanced magnetic coercivity in the aged Fe-Pd-Rh alloys [7{16]. Careful analysis of Figures 2(a)-(c) reveals a small shift in the di raction angles, demon-strating the lattice change due to the movement of the Pd-Rh atoms in association with the phase transforma-tion. The fcc phase separation into L1m+L10 phases, which was con rmed by TEM, was discussed in the pre-vious section. The two phases existing in the aged alloys, as investigated by TEM and X-ray di raction analyses, are exactly the same.
3. Magnetic Property
The mass magnetization (M) versus magnetic eld (H) curves, measured at temperatures of 50, 200 and 350 K for the alloys ST at 950 C for 1.5 h and quenched in ice water are shown in Figure 3(a). The hysteresis loops reveal that the saturation magnetization values of the magnetic moment per unit mass are 137.4 { 99.5 (emu/g) for the ST sample measured at temperatures of 50 { 350 K. As shown in Figure 3(b), we see the M-H curves mea-sured at the same temperatures of the same (ST) alloys after thermal aging at 450C for 100 h. The hysteresis loops in Figure 3(b) display a high saturation magneti-zation (Ms = 112.9 { 120.7 emu/g), as well as a high coercivity Hc = 800 { 1000 Oe, throughout the temper-atures of 50 { 350 K. Figure 3(b) also reveals that the specimen aged at 450C for 100 h possesses a superior coercivity to that of the ST specimen (Figure 3(a)). This is due to the much greater amount of ferromagnetic pre-cipitates of the L1m and the L10 phases appearing in the latter specimen. The magnetic test suggests that in the aged specimen, the increase in the saturation mag-netization (Ms) and the coercivity (Hc) can be ascribed to the precipitation of the L1mand the L10 structures. In addition, it is also interesting to note that the mag-netization (M) versus magnetic eld (H) curve of the aged specimen exhibits an abrupt drop at the saturation remanence (Mr), as indicated by the arrow in Figure 3(b). This characteristic magnetic property illustrates that two phases exist in the aged alloys and that the sat-uration remanences in the two phases are di erent. The two phases existing in the aged alloys have been demon-strated by TEM and X-ray di raction observation and were discussed in the above sections. The characteristic magnetic property of the magnetic test exactly coincides with that of the microstructure analysis.
Fig. 3. Magnetization (M) versus magnetic eld (H) curves measured at temperatures of 50, 200 and 350 K for the Fe-30Pd-4Rh alloys: (a) solution-treated (ST) at 950C for 1.5
h and ice water quenched and (b) ST and then thermally aged at 450C for 100 h.
the alloys aged at 450 C for 100 h is shown in Figure 3(b), in which some of the intrinsic magnetic properties and domain parameters of the aged Fe-Pd-Rh alloys are obtained from the experimental results and some are cal-culated from equation [3,17]: (1) anisotropy constant K1 = MsHa/2 = 1.45 107 ergs/cm3, (2) Ms (saturation magnetization) at R.T. = 970 emu/cc, (3) Ha(anisotropy eld) = 30 kOe, (4) Tc (Curie temperature) = 485C (758 K), (5) B (domain wall thickness) = (A1/K1)1=2 = 82A (A1 = 10 6 ergs/cm; A1: exchange parameter), (6) (domain wall energy) = 4(A1K1)1=2= 15 ergs/cm2 and (7) energy product (BH)maxtheor= (4Ms/2)2 = 37 MGOe.
IV. CONCLUSIONS
1. For solution-treated the alloys and quenched in ice water, the fcc phase separates into fcc+L1m+ rst-order L10 structures. Then, during thermal
aging of the alloys, the fcc+L1m+ rst-order L10 phases decompose into bct+L1m+second-order L10 structures, as veri ed by both the TEM and the X-ray di raction measurements.
2. TEM and X-ray di raction studies con rm that an intermediate martensitic structure exists be-tween the fcc ! L10martensitic transformation in the aged Fe-Pd-Rh alloys. The intermediate phase (L1m) has a monoclinic structure with lattice pa-rameters of a = 3.193 A, b = 3.684 A, c = 3.141
A and = 92.042. The observed intermediate L1m monoclinic phase appears to have an adap-tive martensitic structure.
3. TEM selected area di raction pattern (SADP) analysis reveals that the orientation relationships between the L10 and the L1m phases can be demonstrated as [101]L10==[100]L1m.
4. The observed APBs and microtwins in the TEM image suggest that the primary mechanism of co-ercivity in the aged Fe-Pd-Rh alloy system is con-trolled by pinning and that planar defects, such as APBs and microtwins, are the primary obstacles to domain wall motion.
5. The hysteresis loops of the aged alloys exhibit an abrupt drop at the saturation remanence (Mr), in-dicating two phases with di erent saturation re-manences exist in the aged alloys. The two phases, i.e., the adaptive L1mmonoclinic phase and the or-dered L10 martensitic structure, are demonstrated by the TEM and the X-ray di raction studies.
ACKNOWLEDGMENTS
The authors would like to express their sincere ap-preciation to the National Science Council Republic of China for supporting the work (under Grant-in-Aid for the project No: NSC-94-2212-E-151-008).
REFERENCES
[1] A. G. Khachaturyan, S. M. Shapiro and S. Semen-ovskaya, Phys. Rev. B 43, 10832 (1991).
[2] K. Shimizu and T. Tadaki, Materials Trans. Japan Inst. of Metals 33, 165 (1992).
[3] T. J. Klemmer, D. Hoydick, H. Okumura, B. Zhang and W. A. So a, Scripta Metallurgica et Materialia 33, 1793 (1995).
[4] T. J. Klemmer, C. Liu, N. Shukla, X. W. Wu, D. Weller, M. Tanase, D. E. Laughlin and W. A. So a, J. of Mag. Magn. Mat. 266, 79 (2003).
[5] B. Zhang and W. A. So a, Phys. Stat. Sol. (a) 131, 707 (1992).
[6] T. Kakeshita, T. Fukuda and T. Takeuchi, Mater. Sci. and Engin. A 431, 12 (2006).
[7] J. Cui, T. W. Shield and R. D. James, Acta Materialia 52, 35 (2004).
[8] K. Tsuchiya, T. Nojiri, H. Ohtsuka and M. Umemoto, Materials Trans. The Japan Inst. Metals 44, 2499 (2003). [9] T. Okazaki, H. Nakajima and Y. Furuya, Materials
Transactions. The Japan Inst. Metals 44, 665 (2003). [10] T. Sakamoto, T. Fukuda, T. Kakeshita, T. Takeuchi and
K. Kishio, Mater. Trans. The Japan Inst. Metals 44, 2495 (2003).
[11] E. C. Oliver, T. Mori, M. R. Daymond and P. J. Withers, Acta Materialia 51, 6464 (2003).
[12] T. Wada, T. Tagawa and M. Taya, Scripta Materialia
48, 207 (2003).
[13] S. Inoue, K. Inoue, K. Koterazawa and K. Mizuuchi, Mater. Sci. and Engin. A 339, 29 (2003).
[14] C. T. Hu, T. Goryczka and D. Vokoun, Scripta Materi-alia 50, 539 (2004).
[15] T. Wada, Y. Liang, H. Kato, T. Tagawa, M. Taya and T. Mori, Mater. Sci. and Engin. A 361, 75 (2003). [16] T. Kubota, T. Okazaki, Y. Furuya and T. Watanabe, J.
Mag. Magn. Mater. 239, 551 (2002).
[17] B. D. Cullity, Introduction to Magnetic Materials, edit-ted by M. Cohen (Addison-Wesley, Reading, Mas-sachusetts, 1972), Chap. 1-2.
![Fig. 1. TEM micrographs of the Fe-30Pd-4Rh alloys solu- solu-tion treated (ST) at 950 C for 1.5 h, quenched in ice water and thermally aged at 450 C for 100 h: (a) SADP of the zone axis [101] L10 ==[100] L1m (hkl denotes the ordered L1 0](https://thumb-ap.123doks.com/thumbv2/9libinfo/8822229.232598/2.892.485.810.145.464/micrographs-alloys-treated-quenched-water-thermally-denotes-ordered.webp)
