Glass-forming ability and soft magnetic properties of FeCoSiAlGaPCB amorphous alloys

Glass-forming ability and soft magnetic properties of FeCoSiAlGaPCB amorphous alloys J. M. Borrego, A. Conde, S. Roth, and J. Eckert Citation: Journal of Applied Physics 92, 2073 (2002); doi: 10.1063/1.1494848 View online: http://dx.doi.org/10.1063/1.1494848 View Table of Contents: http://scitation.aip.org/content/aip/journal/jap/92/4?ver=pdfcov Published by the AIP Publishing Articles you may be interested in Effects of B and Si contents on glass-forming ability and soft-magnetic properties in ( Co 0.89 Fe 0.057 Nb 0.053 ) 100 − x ( B 0.8 Si 0.2 ) x glassy alloys J. Appl. Phys. 107, 09A319 (2010); 10.1063/1.3356233 Soft magnetic powder-core composites of Fe 90 Zr 7 B 3 and Fe 49 Co 21 Al 5 Ga 2 P 9.65 C 5.75 B 4.6 Si 3 alloys J. Appl. Phys. 99, 08F103 (2006); 10.1063/1.2164413 Mössbauer study of FeCoSiAlGaPCB amorphous alloys J. Appl. Phys. 95, 4151 (2004); 10.1063/1.1682689 Influence of Si addition on thermal stability and soft magnetic properties for Fe–Al–Ga–P–C–B glassy alloys J. Appl. Phys. 83, 6329 (1998); 10.1063/1.367812 Soft magnetic properties of Fe based amorphous thick sheets with large glass forming ability J. Appl. Phys. 81, 4029 (1997); 10.1063/1.364926 [This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP: 150.214.182.17 On: Wed, 27 Jan 2016 10:32:58 Glass-forming ability and soft magnetic properties of FeCoSiAlGaPCB amorphous alloys J. M. Borrego and A. Condea) Departamento de Fı ´sica de la Materia Condensada, Instituto de Ciencia de Materiales, CSIC, Universidad de Sevilla, P.O. Box 1065, 41080 Sevilla, Spain S. Roth and J. Eckert Institut fu ¨r Metalische Werkstoffe, IFW Dresden, Postfach 270016, D-01171 Dresden, Germany 共Received 21 February 2002; accepted for publication 29 May 2002兲 The glass-forming ability of (FexCoyBzCu)80Si3Al5Ga2P10 with x⫽5–70, y⫽0–63, z⫽5–12, and u⫽0–5 amorphous alloys has been analyzed in terms of the width of the supercooled liquid region, the reduced glass transition temperature, and the Vogel–Fulcher–Tammann parameters. Substitution of Fe by Co slightly decreases the glass-forming ability of the studied alloys. The value of the fragility parameter mis discussed in the frame of the general classification scheme of glass-forming liquids. The crystalline phases formed during the first crystallization step are identified. Magnetic moment at low and room temperature, Curie temperature, room temperature magnetostriction, and coercivity decrease with increasing Co content. © 2002 American Institute of Physics. 关DOI: 10.1063/1.1494848兴 I. INTRODUCTION The development of amorphous soft magnetic materials with a wide supercooled liquid region before crystallization has become an important research topic in recent years.1,2 The decrease of the critical cooling rate for glass formation enables the fabrication of bulk amorphous alloys by conventional casting processes and the existence of a wide supercooled liquid region allows for measurements of the thermophysical properties of the undercooled metallic liquid in a broad time and temperature range.3,4 The glass-forming ability 共GFA兲of amorphous alloys can be characterized by their critical cooling rate, but this parameter is usually not easy to measure and several other parameters have been used to predict the glass-forming ability of metallic glasses. One of the most widely used is the width of the supercooled liquid region ⌬Tx(⫽Tx⫺Tg), where Tgis the glass transition temperature and Txthe onset temperature of crystallization of the amorphous alloy. Inoue and Zhang5have stated a close relation between ⌬Txand GFA for a wide variety of bulk glassy alloys: the larger the width of the supercooled liquid region, the lower the critical cooling rate. However, the importance of a chemical decomposition process in the undercooled liquid should be stressed as a key parameter for very good GFA.6The existence of phase separation in the undercooled liquid state was used to explain the disparity between GFA and thermal stability in the alloy series ZrTiCuNiBe.7 Based on theoretical work on crystal nucleation in undercooled liquid metals, Turnbull8proposed that the glassforming ability should increase with increasing reduced glass transition temperature Trg defined as Trg⫽Tg/Tl共Tlis liquidus temperature兲. This correlation has been confirmed in many experiments 共see Ref. 9 for a summary兲. The study of glass transition kinetics can provide complementary information about the glass-forming ability of the amorphous alloys. It has been found that alloys with high GFA, i.e., low critical cooling rate for glass formation, are stronger metallic glass formers in the ‘‘Angell plot’’10 than thermally less stable metallic liquids.11 There is only one work about the kinetics of the glass transition of Febased alloys with good glass-forming ability.12 Amorphous FeSiAlGaPCB alloys are known to have interesting soft magnetic properties combined with good glassforming ability, promising the formation of bulk soft magnetic materials.1,2 The replacement of iron by other transition metals may further improve these alloys. In this article, the effect of the substitution of Fe by Co on the glass-forming ability and the soft magnetic properties of amorphous (FexCoyBzCu)80Si3Al5Ga2P10 with x⫽5–70, y⫽0–63, z ⫽5–12, and u⫽0–5 alloys is reported. The dependence of the glass transition temperature on the heating rate is analyzed in terms of the Vogel–Fulcher–Tammann 共VFT兲equation, and the value of the fragility parameter mis discussed in the framework of the general classification scheme of glass-forming liquids.10,13 The characterization of the crystalline phases formed in the course of the devitrification process was done by means of x-ray diffraction 共XRD兲measurements. II. EXPERIMENT Multicomponent alloys with compositions (FexCoyBzCu)80Si3Al5Ga2P10 共Table I兲were prepared by arc melting under argon atmosphere. Raw materials of high purity were used; metals: 99.99%, FeC and FeB: 99.5%, and FeP: 97.5%. From these alloys, 10 mm wide and 25 ␮ m thick ribbons were prepared by single-roller melt spinning. a兲Author to whom correspondence should be addressed; electronic mail: [email protected] JOURNAL OF APPLIED PHYSICS VOLUME 92, NUMBER 4 15 AUGUST 2002 20730021-8979/2002/92(4)/2073/6/$19.00 © 2002 American Institute of Physics [This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP: 150.214.182.17 On: Wed, 27 Jan 2016 10:32:58 The samples were proven to be fully amorphous in the asspun state by XRD measurements. The XRD patterns were recorded at room temperature using a Philips PW 1820 diffractometer with CoK ␣ radiation. The values of the onset glass transition temperature Tg, the onset crystallization temperature Tx, and the crystallization peak temperature Tp were determined by differential scanning calorimetry 共DSC兲. The onset glass transition was defined as the point of intersection between the linearly extrapolated curve below the transition with the steepest tangent of the rise in heat flow signal.14 The experiments were performed with a Perkin– Elmer DSC-7 under a continuous argon flow at different heating rates ranging from 2.5 to 160 K/min. The melting behavior was studied with a Netzsch 404 DSC calorimeter at a heating rate of 20 K/min. The liquidus temperature Tlwas determined as the inflection point of the last endotherm of the heating curve 共high temperature side兲. The magnetization as a function of the temperature M(T) was measured with a Faraday magnetometer in a field of 460 kAm⫺1. The Curie temperature TCwas determined from M(T) curves, fitting the data near TCto a critical law of the form M(T)⬀(T⫺TC) ␤ with ␤ ⫽0.36, and extrapolating to M⫽0. The coercive field HCand the saturation magnetostriction constant ␭swere measured at room temperature by a Fo ¨rster Koerzimat and by the small-angle rotation method after Narita15 at 15 kAm⫺1, respectively. The saturation polarization Jsat room temperature was measured with a vibrating sample magnetometer, using a maximum field strength of 1500 kAm⫺1and at 10 K with a superconducting quantum interference device magnetometer using an applied magnetic field of 4 MAm⫺1. III. RESULTS AND DISCUSSION A. Glass-forming ability and crystallization behavior Figure 1 shows the DSC curves for the as-quenched alloys at a heating rate of 20 K/min. All curves exhibit the endothermic event characteristics of the glass transition, followed by a supercooled liquid region and several exothermic crystallization peaks at higher temperatures. Below the glass transition a wide exothermic event can be observed for all the alloys, which is due to structural relaxation. As shown in Fig. 1 a wide supercooled liquid region can be found for all the compositions. The thermal stability of the amorphous alloys does not show a monotonic dependence on Co content: Txis not affected by Co substitution up to about 26 at.% Co and shows a small increase of about 20 K for substitution of more than 52 at.% Co 共Table I兲. It should be noted that the changes in the B and C content 共keeping the total metalloid content constant兲may also affect this behavior. However, a similar tendency was found for Fe73.5⫺xCoxSi15.5B7Cu1Nb3alloys where the B content was kept constant.16 The glass transition temperature increases monotonically as the Co content increases 共see Table I兲. Therefore, the width of the supercooled liquid region ⌬Tx slightly increases from 38 to 54 K as the Fe content of the alloy increases. The melting curves of all the alloys at a heating rate of 20 K/min are shown in Fig. 2. The curve corresponding to the Co-free alloy 共alloy A兲exhibits a sharp single melting event, indicative of a eutectic composition. The alloys with the lowest 共alloy B兲and highest Co content 共alloy F兲exhibit a melting behavior very near to a eutectic point while the FIG. 1. DSC curves of the as-quenched alloys at a heating rate of 20 K/min. TABLE I. Glass transition temperature Tg, crystallization onset temperature Tx, and liquidus temperature Tlat 20 K/min and Curie temperature TCof (FexCoyBzCu)80Si3Al5Ga2P10 alloys. Alloy A B C D E F x70 56 43 29 17 5 y0 1426405263 z56891012 u543210 Tg(K)⫾5 755 760 762 764 782 790 Tx(K)⫾1 809 808 808 812 828 828 Tl(K)⫾5 1280 1300 1330 1330 1320 1340 Structures eutectic neareutectic offeutectic offeutectic offeutectic neareutectic TC(K)⫾3 568 554 526 463 373 328 2074 J. Appl. Phys., Vol. 92, No. 4, 15 August 2002 Borrego et al. [This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP: 150.214.182.17 On: Wed, 27 Jan 2016 10:32:58 curves corresponding to the alloys with intermediate compositions 共alloys C, D, and E兲display two clear melting peaks indicating that they are off eutectic. According to Turnbull’s analysis,8a liquid with Trg⭓2/3 can only crystallize within a very narrow temperature range, thus can be easily undercooled at a low cooling rate to the glass state. The calculated values of Trg , ranging from 0.57 to 0.59, are lower than but close to 2/3, reflecting the good glass-forming ability of the present alloys. The crystallization of the samples occurs in several stages 共Fig. 1兲. The apparent activation energy Ea, the frequency factor Z, and the rate constant of the crystallization process Kcr can be evaluated by using the Kissinger method.17 The dependence of the peak temperature Tpon the heating rate ␤ is described by ␤ /Tp 2⫽共ZkB/Ea兲exp共⫺Ea/kBTp兲,共1兲 where kBis the Boltzman constant and the crystallization rate constant is determined from an Arrhenius law Kcr共T兲⫽Zexp共⫺Ea/kBT兲.共2兲 The values of the apparent activation energy and the frequency factor found for the first crystallization stage of the studied alloys rise in the same order as the crystallization onset temperature 共Table II兲. The frequency factor Zcan be considered as a measure of the probability that an atom having energy Eaparticipates in a crystallization reaction.17 The high variation in Zbetween the alloys with the lowest and the highest Co content could be explained from differences in the crystallization process, i.e., the appearance of new crystalline phases, as will be reported below. Big differences in the value of Zhave been also reported for other alloys systems. As an example, in Zr41Ti14Cu12.5Ni10⫺xFexBe22.5 共x⫽0, 2, and 5兲alloys18 a substitution of only 5 at.% of Ni by Fe causes a difference in Eaof 0.8 eV and in Zof 6 orders of magnitude. A correlation between the value of the rate constant and the width of the supercooled liquid region was found in ZrTiCuNiFeBe 共Ref. 18兲and FeNbAlGaPCB 共Ref. 12兲amorphous alloys: the smaller Kcr , the larger ⌬Tx. The opposite tendency is found in the present alloys, for which the rate constant evaluated at the peak temperature Tpat 20 K/min decreases with increasing Co content, in the same order as ⌬Tx共Table II兲. The study of the relaxation dynamics of supercooled liquids can be discussed in terms of the fragility, which is the degree of departure from an Arrhenius law of the temperature dependence of a characteristic relaxation time.13 The fragility concept is used as the basis for a classification of liquids, to estimate the sensitivity of the liquid structure to temperature changes.10 Since viscosity relaxation and the glass transition measured by calorimetric methods occur on the same time scale, the heating rate dependence of the glass transition can be used as a way to determine the fragility of the material.11 Prior to the DSC measurements the samples were fully relaxed, as complete relaxation leads to a state that is equivalent to a supercooled liquid,11 and which is, therefore, independent of the history of the sample. The conditions of the thermal treatment were chosen in order to get the lowest coercive field: the samples were isothermally annealed at 733 K 共alloys A, B, C, and D兲,743K共alloy E兲, and 773 K 共alloy F兲for 30 min. With increasing heating rate, the glass transition temperature shifts to higher temperatures. As an example Fig. 3 shows the curves corresponding to the Fe70B5C5Si3Al5Ga2P10 alloy 共alloy A兲. The dependence of Tgon the heating rate ␤ given by the VFT equation can be written in the form14 ␤ 共Tg兲⫽BexpbDTg 0/共Tg 0⫺Tg兲c,共3兲 where Tg 0is the asymptotic value of Tg, usually approximated as the onset of the glass transition in the limit of infinitely slow cooling and heating rate, Bhas the dimension of a heating rate, and Dis the strength parameter. The fitting of the experimental data was performed using the equation ln ␤ 共Tg兲⫽lnB⫺DTg 0 共Tg⫺Tg 0兲,共4兲 FIG. 2. Melting curves of all the alloys at a heating rate of 20 K/min. TABLE II. Kissinger parameters. The rate constant Kcr was evaluated at T⫽Tpat 20 K/min. Alloy A B C D E F Ea(eV) 5.4共2兲5.4共2兲5.5共2兲6.0共2兲6.4共2兲7.1共2兲 Z(s⫺1)1(3)⫻1032 1(3)⫻1032 4(3)⫻1032 5(5)⫻1035 1(6)⫻1037 3(9)⫻1041 Kcr(s⫺1)4(2)⫻10⫺24(2)⫻10⫺24(2)⫻10⫺27(2)⫻10⫺32(2)⫻10⫺31(2)⫻10⫺3 2075J. Appl. Phys., Vol. 92, No. 4, 15 August 2002 Borrego et al. [This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP: 150.214.182.17 On: Wed, 27 Jan 2016 10:32:58 with three adjustable VFT parameters: B,D, and Tg 0. The calculated values are given in Table III and the best fits are shown by lines in Fig. 4. The fragility can be quantified by the strength parameter Din Eq. 共3兲, or by the fragility parameter defined as19 m⫽dlog10 具 ␶ 典 d共Tg/T兲 冏 T⫽Tg ,共5兲 where Tis the temperature, Tgthe glass transition, and 具 ␶ 典is the average relaxation time. From the VFT fits the fragility parameter at a particular Tgcan be calculated from13 m⫽DTg 0Tg 共Tg⫺Tg 0兲2ln10.共6兲 The larger the deviation from an Arrhenius behavior, the larger the value of m. The fragility index can be used to classify glass-forming liquids into three general categories: strong, intermediate, and fragile.10 Strong liquids with approximately Arrhenius temperature dependence of relaxation times have values of mlower than 30 with an estimated lower limit of m⬇16. In contrast, fragile liquids such as polymers and ionic melts display values of mabove 100. It should be noted that although the uncertainty of the values of the parameters Tg 0,B, and Dis quite high 共also pointed out in Ref. 20兲, changes in the value of Tg 0result in changes in Dthat keep the value of mreasonably constant. The obtained values of Dare close to those found for FeNbAlGaPCB alloys.12 The increase of the lower limit for the glass transition Tg 0and the decrease of the strength parameter Dwith increasing Co content point in the same direction as the decrease in the width of the supercooled liquid region. The fragility parameter m, evaluated at Tgcorresponding to a heating rate of 20 K/min, was found to be around 35 for alloys A, B, C, D, and E, and about 63 for the alloy with the highest Co content 共alloy F兲. This result indicates that these alloys lie in the intermediate category according to Angell’s classification scheme, i.e., between the strong and the fragile extremes. A fragility mof about 35 would suggest a very good glass-forming ability, a finding that it is corroborated by the fact that Fe–共Al, Ga兲–共P, C, B, Si兲bulk glassy samples with thickness of 1–15 mm have been obtained.1 Similar values of mhave been found in several multicomponent amorphous systems such as ZrTiCuNiBe and MgCuY alloys,21 which are reported to be excellent metallic glass formers.22,23 A fragility value of about 60 for the Co richest alloy indicates a lower glass-forming ability for this alloy. In order to clarify the reason for the good glass-forming ability of these alloys, the crystalline phases formed in the first crystallization stage were analyzed. Figure 5 shows the XRD patterns corresponding to samples of all the alloys annealed at 800 K for 1 h. The annealed samples A, B, C, and D present the same crystalline phases: 共i兲a bcc Fe solid solution containing Co and M, where M stands for P, C, Al, B, and Ga elements, with a lattice parameter that decreases as the Fe content of the amorphous alloy decreases, indicating changes in the composition 共a⫽0.2868 nm alloy A, a ⫽0.2857 nm alloy B; a⫽0.2853 nm alloy C, and a ⫽0.2852 nm alloy D兲;共ii兲a phase with a Ni3P-like tetragonal structure 共space group: I4兲and lattice parameters a ⫽0.893 nm and c⫽0.441 nm 共Ref. 24兲and a composition close to Fe3(M), and 共iii兲a phase with an Fe3(NiN)2-like cubic structure 共space group Pm3m兲, with a lattice parameter of about 0.378 nm 共Ref. 24兲and a composition close to Fe3(M)2-like cubic structure. The last two phases were previously found in alloys with similar compositions.25 The simultaneous crystallization of these three phases causes the crystallization to be retarded, and thus promote a wide supercooled liquid region. The slightly higher crystallization FIG. 3. DSC curves for the alloy Fe70B5C5Si3Al5Ga2P10 at different heating rates. The arrows indicate Tg. FIG. 4. Glass transition temperature as a function of the heating rate ␤ and VFT fit of the data 关Eq. 共4兲兴. TABLE III. Vogel–Fulcher–Tammann parameters for the best fit of the DSC data according to Eq. 共4兲. Alloy ln B共K/s兲 ⫾2D ⫾2.0 Tg 0共K兲 ⫾30 Tg/Tg 0 ⫾0.1 m ⫾10 A 13 3.2 620 1.2 34 B 10 1.6 660 1.2 35 C 8 1.3 670 1.1 35 D 7 1.1 680 1.1 35 E 6 0.8 710 1.1 37 F 5 0.5 745 1.1 63 2076 J. Appl. Phys., Vol. 92, No. 4, 15 August 2002 Borrego et al. [This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP: 150.214.182.17 On: Wed, 27 Jan 2016 10:32:58 temperature of the Co-richest alloys 共E and F兲might be due to the appearance of new Co-rich phases: a phase Co2M with an orthorhombic structure 共space group Pnam兲with lattice parameters a⫽0.565 nm; b⫽0.660 nm, and c⫽0.351 nm 共Ref. 24兲is found in the two alloys; and 共v兲a phase M5Co2 in alloy F.24 However, this increase in Txin alloys E and F is accompanied by an increase in Tgresulting in a final decrease of the width of the supercooled liquid region. B. Magnetic properties The saturation magnetization at 300 and 10 K of as-cast samples shows a linear dependence on the Fe content 关Fig. 6共a兲兴 which reflects the substitution of Fe by Co moments. Similar results are obtained for Co75⫺xFexSi15B10 amorphous alloys.26 The average magnetic moment per magnetic atom 具 ␮ 典decreases from 1.8 to 0.7 ␮ Bas the Fe content decreases 关Fig. 6共b兲兴. These values are much lower than those reported for crystalline FeCo alloys27 and amorphous FeCoSiB alloys28 with lower metalloid concentration 共⬃20 at.%兲. The low values of 具 ␮ 典can be attributed to the decrease of the number of nearest-neighbor magnetic atoms.29 Similar results are found for alloys with metalloid content close to 30 at.%.29,30 The Curie temperature TCof the as-quenched alloys decreases monotonically as the Co content in the alloy increases 共Table I兲. This result can be mainly ascribed to the decrease of the exchange interaction between Co–Co pairs provoked by the presence of metalloid atoms.29 The Curie temperature and the coercive field are affected by structural relaxation processes:31 TCincreases as the annealing temperature increases and HCdecreases monotonically as the relaxation progresses. The lowest coercivity, HC(min), obtained after annealing the samples for 30 min at temperatures between 713 and 733 K, decreases as the Co and B content of the alloy increases 关Fig. 6共c兲兴. This dependence correlates well with that observed for the saturation magnetostriction 关Fig. 6共d兲兴. IV. CONCLUSIONS The GFA of (FexCoyBzCu)80Si3Al5Ga2P10 amorphous alloys has been analyzed in terms of different parameters, such as the width of the supercooled liquid region, the reduced glass transition temperature, and the VFT fitting parameters. Substitution of Fe by Co slightly decreases the GFA. The value of the fragility parameter mindicates that these alloys lie in the intermediate category according to Angell’s classification scheme, i.e., between the strong and the fragile extremes. Curie temperature, saturation magnetization, coercivity, and magnetostriction decrease as the Co content increases, the last reaching a nearly zero value. ACKNOWLEDGMENTS The authors thank H. Grahl for the preparation of the samples and N. Mattern and A. Ostwald from IFW Dresden for the XRD measurements and their valuable help for the identification of the crystalline phases. This work was partially supported by the Spanish Ministry of Science and Technology and EU-FEDER 共Project Nos. PB97-1119C02-01 and MAT 2001-3175兲and by the PAI of the Junta de Andalucı ´a. J.M.B. acknowledges the IFW Dresden and the Junta de Andalucı ´a for research fellowships. 1A. Inoue, A. Makino, and T. Mizushima, J. Magn. Magn. Mater. 215–216, 246 共2000兲. 2A. Inoue, A. Takeuchi, and B. Shen, Mater. Trans., JIM 42,970共2001兲. FIG. 5. XRD patterns corresponding to samples of all the alloys annealed at 800Kfor1h. FIG. 6. Saturation magnetization Jsat 300 and 10 K 共a兲; magnetic moment per (Fe⫹Co) atom 具 ␮ 典at 10 K 共b兲; minimum coercivity after annealing HC(min) 共c兲; and saturation magnetostriction constant ␭sas a function of the Fe content 共d兲. 2077J. Appl. Phys., Vol. 92, No. 4, 15 August 2002 Borrego et al. [This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP: 150.214.182.17 On: Wed, 27 Jan 2016 10:32:58 3R. Busch, Y. J. Kim, and W. L. Johnson, J. Appl. Phys. 77, 4039 共1995兲. 4U. Geyer, S. Schneider, W. L. Johnson, Y. Qiu, T. A. Tombello, and M. P. Macht, Phys. Rev. Lett. 75, 2364 共1995兲. 5A. Inoue and T. Zhang, J. Non-Cryst. Solids 250–252, 552 共1999兲. 6W. L. Johnson, MRS Bull. 24,42共1999兲. 7T. A. Waniuk, J. Schroers, and W. L. Johnson, Appl. Phys. Lett. 78, 1213 共2001兲. 8D. Turnbull, Contemp. Phys. 10,473共1969兲. 9Z. P. Lu, H. Tan, Y. Li, and S. C. Ng, Scr. Mater. 42, 667 共2000兲. 10C. A. Angell, Science 267, 1924 共1995兲. 11 R. Busch, E. Bakke, and W. L. Johnson, Acta Mater. 46, 4725 共1998兲. 12N. Mitrovic, S. Roth, and J. Eckert, Appl. Phys. Lett. 78, 2145 共2001兲. 13R. Bo ¨hmer, K. L. Ngai, C. A. Angell, and D. J. Plazek, J. Chem. Phys. 99, 4201 共1993兲. 14R. Bru ¨ning and K. Samwer, Phys. Rev. B 46, 11318 共1992兲. 15K. Narita, J. Yamasaki, and H. Fukunaga, IEEE Trans. Magn. 16, 435 共1980兲. 16M. Mu ¨ller, H. Grahl, N. Mattern, U. Ku ¨hn, and B. Schnell, J. Magn. Magn. Mater. 160, 284 共1996兲. 17H. E. Kissinger, Anal. Chem. 29, 1702 共1957兲. 18Y. X. Zhuang, W. H. Wang, Y. Zhang, M. X. Pan, and D. Q. Zhao, Appl. Phys. Lett. 75,2392共1999兲. 19R. Bo ¨hmer and C. A. Angell, Phys. Rev. B 45, 10091 共1992兲. 20S. Linderoth, N. H. Pryds, M. Ohnuma, A. S. Pedersen, M. Eldrup, N. Nishiyama, and A. Inoue, Mater. Sci. Eng., A 304–306, 656 共2001兲. 21D. N. Perera, J. Phys.: Condens. Matter 11, 3807 共1999兲. 22Y. J. Kim, R. Busch, W. L. Johnson, A. J. Rulison, W. K. Rhim, and D. Isheim, Appl. Phys. Lett. 65, 2136 共1994兲. 23A. Inoue, T. Nakamura, N. Nishiyama, and T. Masumoto, Mater. Trans., JIM 33, 937 共1992兲. 24JCPDS cards for the crystalline phases are: 14-212 for the Ni3P structure, 9-318 for the Fe3(NiN)2structure, 32-306 for Co2M structure, and 481301 for M5Co2structure. International Center for Diffraction Data, 1601 Park Lane Swarthmore, PA 19081. 25J. Z. Jiang, J. S. Olsen, L. Gerward, S. Abdali, J. Eckert, N. Schlorke-de Boer, L. Schultz, J. Truckendrodt, and P. X. Shi, J. Appl. Phys. 87, 2664 共2000兲. 26M. Knobel, R. Sato Turtelli, and R. Gro ¨ssinger, J. Magn. Magn. Mater. 116, 154 共1992兲. 27D. I. Bardos, J. Appl. Phys. 40, 1371 共1969兲. 28M. Goto, H. Tange, and T. Tokunaga, Jpn. J. Appl. Phys. 17,1877共1978兲. 29B. G. Shen, L. Cao, and H. Q. Guo, J. Appl. Phys. 73, 5730 共1993兲. 30K. Narita, J. Yamasaki, and H. Fukunaga, IEEE Trans. Magn. 14,1016 共1978兲. 31J. M. Borrego, A. Conde, S. Roth, and J. Eckert, J. Magn. Magn. Mater. 共in press兲. 2078 J. Appl. Phys., Vol. 92, No. 4, 15 August 2002 Borrego et al. [This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP: 150.214.182.17 On: Wed, 27 Jan 2016 10:32:58