Influence of Annealing Conditions on Nanocrystalline and Ultra-Soft Magnetic Properties of Fe75.5Cu1Nb1Si13.5B9 Alloy

J. Mater. Sci. Technol., 2012, 28(1), 21–26.

Influence of Annealing Conditions on Nanocrystalline and Ultra-Soft Magnetic Properties of Fe75.5 Cu1 Nb1 Si13.5 B9 Alloy S.P. Mondal1) , Kazi Hanium Maria2)† , S.S. Sikder1) , Shamima Choudhury2) , D.K. Saha3) and M.A. Hakim3) 1) Department of Physics, Khulna University of Engineering & Technology, Khulna-9203, Bangladesh 2) Department of Physics, University of Dhaka, Dhaka-1000, Bangladesh 3) Materials Science Divisions, Atomic Energy Centre, Dhaka-1000, Dhaka, Bangladesh [Manuscript received May 23, 2011, in revised form December 25, 2011]

The magnetic and structural properties of FINEMET alloy with a composition of Fe75.5 Cu1 Nb1 Si13.5 B9 were investigated after primary and secondary crystallization of amorphous ribbon sample. The crystallization behavior and the nanocrystal formation of the samples were performed by differential thermal analysis (DTA) which in turn was supported by X-ray diffraction (XRD) study. Temperature dependence of initial permeability of amorphous and devitrified toroid shaped samples has been measured. Enhancement of Curie temperature of the amorphous alloy has been observed due to the irreversible structural relaxation. With the appearance of nanocrystalline phase the Curie temperature of the residual amorphous phase gradually decrease with the increase of annealing temperature. Their temperature dependence reflects the characteristic annealing temperature evolution of the basic magnetic parameters in these nanocrystalline systems. Saturation magnetization, Ms , increases with annealing temperature Ta for the samples and finally decreases during annealing at a temperature much higher than peak crystallization temperature. KEY WORDS: Nanostuctured alloy; Curie temperature; Saturation; Magnetization

1. Introduction With the development of nanocrystalline Fe-SiB-Nb-Cu alloys, commercially known as FINEMET, a new approach was established to develop softmagnetic materials with high magnetic flux density. The magnetocrystalline anisotropy can be reduced by refining the grain size in less than a few tens of nanometers[1–5] . Finemet, this name derives from the combination of FINE and METAL which indicates the material s feature of fine crystal grains being formed and having excellent magnetic properties. The precursor of FINEMET is amorphous ribbon (noncrystalline) obtained by rapid quenching at one million ◦ C/second from the molten metal consisting of Fe, Si, B and small amounts of Cu and Nb. In this case, Cu acts as a nucleating agent whereas Nb in† Corresponding author. Tel.: +88 02 01711987595; Fax: +88 02 8615583; E-mail address: [email protected] (K.H. Maria).

hibits the grain growth of the FeSi phase that crystallizes from the amorphous matrix during annealing. These crystallized alloys have grains which are extremely uniform and small, about ten nanometers in size. Amorphous metals which contain certain alloy elements show superior soft magnetic properties through crystallization. The magnetic softness of FINEMET alloy mainly arises from the depression of ultrafine grains in an amorphous matrix which reduces the effective magnetic anisotropy and magnetostriction. This indicates that the exchange interaction between nanocrystalline and amorphous phases plays an important role in achieving good soft magnetic properties[6–8] . It was commonly known that soft magnetic materials are characterized by larger crystal grains. Contrary to this common belief, soft magnetic materials consisting of a small “nano-order”, crystal grains have an excellent soft magnetic property which finds use in electric power applications such as transformer cores and other inductive devices.

S.P. Mondal et al.: J. Mater. Sci. Technol., 2012, 28(1), 21–26.

22

In this work, we report the magnetic and structural properties and the primary and secondary crystallization reactions of nano-composite soft magnets with a composition of Fe75.5 Cu1 Nb1 Si13.5 B9 . 2. Experimental The amorphous Fe75.5 Cu1 Nb1 Si13.5 B9 alloy was prepared by rapid solidification of the melt by using the single roller copper wheel melt spinning technique in the form of ribbon. The ribbons were 6 mm wide and 20–25 μm thick. The amorphosity of the ribbons has been confirmed by using a PW 3040-X” Pert Pro (Phillips) X-ray diffractometer with CuK∝ radiation. Crystallization behavior has been performed by differential scanning calorimetry (2960 SDT, USA). Temperature dependence of initial permeability of the as-cast and annealed ribbons is measured by using a laboratory built furnace and a Wayne Kerr 3255 B impedance analyzer. The magnetization measurements of the samples were carried out by using a vibrating sample magnetometer (Model VSM-02, Hirstlab, UK) at room temperature as a function of field.

Fig. 1 DTA trace of the as-cast amorphous ribbon with continuous heating

3. Results and Discussion 3.1 Differential thermal analysis Calorimetric studies of amorphous alloys provide substantial fundamental information concerning the kinetics of crystallization and structural relaxation effects. Luborsky[9,10] studied the kinetics of a variety of amorphous magnetic alloys by means of calorimetric and magnetic techniques. In the present investigation differential thermal analysis (DTA) technique has been used to study the crystallization behavior of nanocrystalline alloys Fe75.5 Cu1 Nb1 Si13.5 B9 . Fig. 1 shows DTA profile of as-cast amorphous ribbons with a heating rate of 20 ◦ C/min in a nitrogen atmosphere. DTA is a direct and effective technique for analyzing the crystallization kinetics of amorphous materials. Two exothermic peaks are distinctly observed which correspond to two different crystallization events at temperatures Tx1 and Tx2 respectively. The soft magnetic properties correspond to the primary crystallization (Tx1 ) of α-Fe (Si) phase. Secondary crystallization (Tx2 ) corresponds to Fe-B phase which causes magnetic hardening of the nanocrystalline alloy. The crystallization onset temperatures (Tx1 and Tx2 ) and peak temperatures (Tp1 and Tp1 ) display exothermic peak, i.e. release of heat during the crystallization of Fe(Si) and Fe-B phases. Since the structure of the beneficial ferromagnetic nanocrystalline phase is comprised of Fe-Si[11,12] , the study of primary and secondary crystallization temperatures are important for the amorphous alloys that are used as a precursor for the production of nanocrystalline FINEMET. Effect of annealing temperatures of the amorphous

Fig. 2 Effects of annealing temperatures on DTA traces of the as-cast amorphous ribbon

ribbons on their crystallization behaviors have been performed by DTA scan at a continuous heating rate of 20 ◦ C/min as shown in Fig. 2. It is observed from the DTA scan that the onset temperature for the sample annealed at Ta =450 ◦ C is almost unchanged with respect to its amorphous precursor which is quite logical, since Ta =450 ◦ C is still lower than it s onset temperature of primary crystallization, Tx1 =470 ◦ C. But when the sample was annealed at Ta = 475 ◦ C which is slightly higher than the onset of crystallization temperature of Tx1 =470 ◦ C, the primary crystallization peak has diminished to large extent and display quite diffused character which signifies that the substantial amount of crystallization has already taken place. Since Cu helps nucleation of Fe(Si) phase and Nb delays the formation of boride phase[13–15] , the observed anomalies of crystallization temperatures in this studied samples are clearly understood from their compositional variation of Cu and Nb. This is to note that the whole process of crystallization takes place over a wide range of temperatures. For example, crystallization starts at Tx1 =470 ◦ C and completed at

S.P. Mondal et al.: J. Mater. Sci. Technol., 2012, 28(1), 21–26. st

23

nd

Table 1 Annealing effects on 1 and 2 crystallization states of the nanocrystalline amorphous ribbon with a composition of Fe75.5 Cu1 Nb1 Si13.5 B9 at a constant heating rate of 20 ◦ C/min Annealing Onset temp. Primary crystal Onset temp. Secondary crystal ΔT (Tp2 − Tp1 ) temperature of primary Peak crystallization of secondary, peak crystallization /◦ C ◦ ◦ ◦ ◦ temperature, Tp1 / C Tx2 / C temperature, Tp1 / C Tx1 / C As-cast 470 482 558 569 87 468 480 557 569 89 450 ◦ C – – 554 566 – 475 ◦ C

Fig. 4 Variation of grain size, lattice parameter and Si-content with annealing temperature for the nanocrystalline amorphous ribbon

Fig. 3 XRD patterns for as-cast and heat-treated samples at 450◦ C to 700◦ C for 30 min

T =500 ◦ C. This signifies that the nucleation and growth of crystallites are faster in the initial stage of crystallization, which gradually becomes sluggish with the increase of volume fraction of crystallites. The result of DTA analysis of the samples is displayed in Table 1. 3.2 XRD pattern DTA analysis cannot give information about the phase identification. Phase identification of the samples was performed by X-ray diffraction (XRD). In Fig. 3, the XRD spectra of the samples as-cast and annealed at 475 ◦ C to 700 ◦ C for 30 min have been presented. XRD results indicate that no α-Fe phases are present in the alloys annealed at and below 450 ◦ C for 30 min with the appearance of a broader diffused pattern, which are characteristics of amorphous material[16] . Above 450 ◦ C it is clearly evident that the bcc Fe(Si) peaks become narrower and sharper with higher intensity during increasing annealing temperature, which indicates that the crystallite sizes are growing larger gradually. X-ray pattern of Ta =475 ◦ C, clearly confirms the presence of crystalline phase identified as a bcc α-Fe(Si) solid solution developed in the amorphous matrix. Fig. 4 shows the variation of lattice parameter of

Fe(Si) phase, Si-content and grain size of α-Fe(Si) phase with respect to the annealing temperature of the samples. With the increase of annealing temperature lattice parameter increases gradually. The lattice parameters of α-Fe(Si) phases are smaller than that of pure Fe, the value of which is 0.2866 nm. It is also observed from Fig. 4 and Table 2 that the grain size increases with annealing temperature while Si content decreases with Ta . This is contradictory to original FINEMET alloy. Such a situation may only be explained by assuming that at high temperature (T >500 ◦ C) recrystallization of Fe(Si) grains takes place. Lower Nb content may also be the reason for this deviation. The real cause is not clear and remains an open issue. Increase of grain size with annealing temperature corresponds well with the results by Rubinstein et al.[17] . The formation of this particular nanostructure is ascribed to the combined effects of Cu and Nb and their low solubility in iron. 3.3 Curie temperature measurement In Fig. 5, the temperature dependence of initial permeability of the as-cast amorphous ribbon and the samples annealed at 425–450 ◦ C in the interval of 25 ◦ C has been presented. For as-cast and samples annealed at 425 and 450 ◦ C, permeability passes through a maximum just before a sharp fall with the manifestation of Hopkinsons effect characterizing the ferro-paramagnetic transition of the amorphous phase. However, for the sample annealed at 425 ◦ C

S.P. Mondal et al.: J. Mater. Sci. Technol., 2012, 28(1), 21–26.

24

Table 2 Experimental XRD data of nanocrystalline amorphous ribbon at different annealing temperatures Temp. / ◦ C 475 500 525 550 600 650 700

θ / deg. 22.59 22.56 22.53 22.52 22.50 22.49 22.47

d / nm 0.20066 0.20110 0.20111 0.20121 0.20143 0.20152 0.20162

FWHM / deg. 0.53 0.52 0.46 0.43 0.40 0.38 0.37

a0 / nm 0.28379 0.28396 0.28441 0.28455 0.28486 0.28499 0.28513

Dg / nm 17 18 22 23 25 26 27

Si / at.% 16.96 16.16 14.04 13.39 11.93 11.28 10.67

Table 3 Annealing temperature, Ta dependence of the Curie temperature of amorphous matrix Tc Crystalline state Amorphous state

Nanocrystalline state

Annealing temp., Ta /◦ C As-cast 425 450 465 480 500 525

Fig. 5 Temperature dependence of μ of as-cast and annealed sample

the sharpness of the fall of the real part of permeability μ is relatively weak which might be an indication of initiation of nucleation since no crystalline phase could be detected for this annealing temperature by X-ray diffraction. But for the sample annealed at 465– 525 ◦ C, it is observed that the sharpness of the fall of μ is progressively smeared out with the appearance of a tail in the high temperature region. These results are in good agreement with those previously reported for the FINEMET composition[18–20] . From the graph of μ vs. T for the sample annealed at 465–525 ◦ C (Fig. 5), it can be observed that decrease of μ with T is quite different which is relatively smeared out. It is very difficult to find out a unique value of TC from such diffusion μ vs. T measurements. The more diffusion character of the ferro-paramagnetic transition in the residual amorphous matrix for the samples annealed at higher temperatures may be attributed to the higher degree of compositional and structural dis-

Curie temp., Tc /◦ C 421 427 437 285 268 295 305

order in the residual amorphous phase[21–23] . Table 3 shows the Curie temperature of the samples at amorphous and nanocrystalline state. TC increases when the sample is annealed in the temperature range of 425–450 ◦ C. Enhancement of TC during annealing of the amorphous precursor below Tx1 is caused by the irreversible structural relaxation of microstructural quantities like long range internal stress, topological and chemical short-range order. Increase in packing density of atoms might have significant contribution into the enhancement of TC at amorphous state. But TC of the amorphous matrix decreases significantly when annealed at and above crystallization temperatures. The probable reason for decreasing TC of the amorphous phase when annealing at and above the crystallization temperature is that amorphous matrix is depleted with iron and the relative amount of increase in Nb. This weakens the exchange interaction resulting in a decrease of TC for the amorphous matrix. It is interesting to observe an enhancement of TC again when samples are annealed at higher temperature. This increase cannot be explained in a straight forward way. Probably redistribution of atomic species, local environment of the matrix and the compositional variance of the residual amorphous matrix as well as the procedure used to determine TC are the cause of the behavior. Varga et al.[24] have interpreted this type of reduction of TC in FINEMET alloys after annealing at higher temperature due to compositional gradients within the remaining amorphous phase. At higher measuring temperatures above TCam , permeability value decreases to a very low value. Franco et al [25] have demonstrated that a super paramagnetic behavior is a general feature of these nanocrystalline alloys. The practical requisite for observing superparamagnetic relaxation in the nanocrystalline alloys is the absence of interaction

S.P. Mondal et al.: J. Mater. Sci. Technol., 2012, 28(1), 21–26.

Fig. 6 Temperature dependence of μ after evoluation of Fe2 B in the nanocrystalline state

between nanograins in the residual matrix[26] . In Fig. 6, the variation of real part of initial permeability μ with temperature has been presented for samples annealed at 550 to 575 ◦ C. It has been observed that the value of μ has dropped to a very low value for the sample with annealing temperature of 550 ◦ C. It has been reported earlier that this fall of μ to a very low value might occur due to the evolution of boride phase at higher temperature[27] . In our experiment the presence of boride phase could not be detected by X-ray diffraction. Because annealing at higher temperature above 525 ◦ C leads to the precipitations of small fractions of boride compounds with dimensions of 50 nm to 100 nm, while the ultra fine grain structure of bcc Fe-Si still persists. At this stage Nb and B are excluded from α-Fe(Si) and enriched in the remaining amorphous phase, because they are insoluble in the α-Fe(Si) phase. At the same time the presence of Nb inhibits the formation of FeB compounds[28] . Temperature dependence of μ reveals the presence of Fe2 B[29,30] . Magnetic measurement (μ) being very sensitive, can detect very small amount of boride phase due to its high magnetocrystalline anisotropy (K1 ) and presence of small amount of Fe2 B phase which can cause a substantial reduction of exchange interaction[19] . Since K1 of Fe2 B passes from negative to positive value at 255 ◦ C, a dramatic rise of μ from 760 at room temperature to 3000 at 255 ◦ C for the sample annealed at 550 ◦ C is evidenced. However for the sample annealed at 575 ◦ C the value of μ rises from 700 at room temperature to 2300 at 265 ◦ C. Such a sharp increase of μ at around 260±5 ◦ C is only possible when the anisotropy energy approaches zero at this temperature. Therefore μ vs. T measurement is a powerful technique to confirm the existence of boride phase in the FINEMET alloys, whereas XRD cannot do. 3.4 Saturation magnetization Fig. 7 shows the field dependence of specific mag-

25

Fig. 7 Specific magnetization versus magnetic field of ascast and annealed samples

netization for amorphous as-quenched and thermally treated samples measured by using a vibrating sample magnetometer (VSM). From the curves it is clearly evidenced that the magnetization is saturated for all the samples at the amorphous and annealed states within an applied field of 2000 Oe. It can be seen that with increasing annealing temperature magnetization increases until Ta =525 ◦ C. The maximum saturation magnetization is reached at Ta =525 ◦ C for the samples. Aranda et al.[31] have studied the approach to saturation in nanocrystalline FINEMET materials. The magnetization prior to saturation is associated with reversible rotation and has been fitted to the law  a2  a1 − 2 + bH 1/2 M (H) = Ms 1 − H H a2 where the term H 2 was described as a direct consequence of the random anisotropy model, and attributable to Fe-Si grains. The co-efficient a2 reflects the Herzer s predicted effective magnetic anisotropy of the nanocrystalline material, where as in amorphous alloys it is postulated as being caused by local stress and magneto elastic coupling. Saturation magnetization Ms , has been observed to increase from 140 to 163 emu/gm for Fe75.5 Cu1 Nb1 Si13.5 B9 with the increase of annealing temperature. An increase of Ms for the annealed samples at 450 to 525 ◦ C compared with the amorphous state is due to the irreversible structural relaxation, changing the degree of chemical disorder of the amorphous state[32,33] and enhanced volume fraction of Fe(Si) nanocrystals that are exchange coupled. The saturation magnetizations (Ms ) are shown in Table 4. It is to be noted that an increase in Ms due to structural relaxation has also been detected in Febased glasses[34–36] . A rapid decrease in Ms has been observed with increasing annealing temperature at 550 ◦ C. The decreasing of Ms may be connected with the enrichment of the residual amorphous phase with Nb that weakens the coupling between ferromagnetic nanograins. Also the role of Si diffusion

S.P. Mondal et al.: J. Mater. Sci. Technol., 2012, 28(1), 21–26.

26

Table 4 Saturation magnetization at different annealing temperatures with a constant annealing time period of 30 min Sample

Fe75.5 Cu1 Nb1 Si13.5 B9

Annealing temperature, Ta /◦ C As-cast 450 500 525 550

into Fe (Si) nanograins and these local environments also may have effects on decreasing Ms . 4. Conclusion DTA reveals the primary, Fe(Si) and secondary (Fe-B) crystallization temperatures with the manifestation of two well-defined exothermic peaks. Curie temperature of the as-cast amorphous ribbon is 421 ◦ C. Enhancement of Curie temperature was observed for the samples annealed below the crystallization temperature due to structural relaxation. Curie temperature of the interfacial amorphous phase has decreased for samples annealed above the crystallization temperature. Temperature dependence of real part of permeability of the annealed samples between the annealing temperatures of 550 and 575 ◦ C exhibits superparamagnetic/superferromagnetic behavior at T >Tcam . The saturation magnetization for nanocrystalline samples has slightly increased for annealing at temperature around the onset of crystallization. When annealing at higher temperature at which complete crystallization takes place, magnetization decreases again. REFERENCES [1 ] Y. Yoshizawa, S. Oguma and K. Yamauchi: J. Appl. Phys., 1988, 64, 6044. [2 ] T.H. Noh, M.B. Lee, H.J. Kim and I.K. Kang: J. Appl. Phy., 1990,.67, 5568. [3 ] H. Gleiter: Prog. Mater. Sci., 1989, 33, 223. [4 ] R.M. Bozorth: Ferrimagnetism, D. Van Nostrand Company, Inc., Princeton, NJ, 1964, 74. [5 ] M. Manivel Raja, K. Chattopadhyay, B. Majumder and A. Narayanasamy: J. Alloy. Compd., 2000, 297, 199. [6 ] T. Kulik, A. Hernando and M. Vasquez: J. Magn. Magn.. Mater., 1994, 133, 310. [7 ] M.A. Hakim and S. Manjura Hoque: J. Magn. Magn. Mater., 2004, 284, 395. [8 ] J. Bigot, N. Lecaude, J.C. Perron, C. Millan, C. Ramiarinjaona and J.F. Rialland: J. Magn. Magn. Mater., 1994, 133, 299. [9 ] F.E. Luborsky: Mater. Sci. Eng., 1977, 28, 139. [10] F.E. Luborsky, J.J. Decker and R.O. McCarv: IEEE Trans. Mag., 1975, 11, 1644. [11] G. Herzer and H. Warlimont: Nanostruct. Mater., 1992, 1, 263.

Saturation magnetization, Ms /(emu/gm) 140 148 155 163 130

[12] K. Hono, A. Inoue and T. Sakurai: Appl. Phys. Lett., 1991, 58, 2180. [13] H.L. Chatelier: Bull. Soc. France Mineral, 1887, 10, 204. [14] G. Herzer: IEEE Trans. Magn., 1989, 25, 3327. [15] J. Zhi, K.Y. He, L.Z. Cheng and Y.J. Fu: J. Magn. Magn. Mater., 1996, 153, 315. [16] S.N. Kane, S. Sarabani, A. Gupta, L.K. Varga and T. Kulit: J. Magn. Magn. Mater., 2000, 372, 215. [17] M. Rubinstein, V.G. Harris and P. Lubitz: J. Magn. Magn. Mater., 2001, 234, 306. [18] A. Olszewski, J. Zbroszczyk, J. Olszewski, W.H. Ciurzynka, H. Fukunaga, K. Narita, B. Wyslockc and M. Hasiak: J. Magn. Magn. Mater., 2000, 215–216, 422. [19] G. Herzer: Hand Book of Magnetic Materials, eds K.H.J. Buschow, Vol.10, Elsevier Science B.V., Amsterdam, 1997, 427. [20] R. Alben, J.J. Becker and M.C. Chi: J. Appl. Phy., 1978, 49, 1653. [21] S.M. Hoque and M.A. Hakim: J. Mater. Chem. Phys., 2007, 101, 112. [22] H. Okumura , D.E. Laughlin and M.E. McHenry: J. Magn. Magn. Mater., 2003, 267, 347. [23] I. Skorvanek, J. Kovac and J. Kotzler: physia. status. solidi (b), 2003, 236, 303. [24] L.K. Varga and K.V. Rao: Nanostruct. Mater., 1999, 12, 1157. [25] V. Franco, C.F. Conde and A. Conde: J. Magn. Magn. Mater., 1998, 185, 353. [26] M.M. Uddin, S.M. Hoque, M.S. Mahmud, M.A. Hakim and F.U.Z. Chowdhury: Indian J. Pure Appl. Phys., 2009, l47, 289. [27] X.Y. Zhang, J.W. Zhang, Y.Z. Zheng, P.R. Liu and J.H. Zhao: J. Magn. Magn. Mater., 1997, 172, 301. [28] M. E. McHenry, M.A. Willard and D.E. Laughlin: Prog. Mater. Sci., 1999, 44, 291. [29] K. Hono and T. Sakuria: Appl. Surf. Sci, 1995, 87, 166. [30] K. Hono, K. Hiraga, Q. Wang, A. Inoue and T. Sakurai: Acta Metall. Mater., 1992, 40, 2137. [31] G. R, Aranda, J. Gonzalez and K. Kulakowski: J. Appl. Phys., 1998, 83, 6341. [32] I. Skorvanek, J. Kovac and J.M. Greneche: J. Phys. Condens. Mater., 2000, 12, 9085. [33] A. Lovas, L. F. Kiss and I. Balong: J. Magn. Magn. Mater., 2000, 215–216, 463. [34] A.E. Berkowitz, J.L. Walter and K.F. Wall: Phys. Rev. Lett., 1981, 46, 1484. [35] K. Suzuki, A. Makino, N. Kataoka, A. Iniue and T. Masumoto: J. Appl. Phys., 1991, 70, 6232. [36] K. Suzuki, A. Makino, A. Inoue and T. Masumoto: J. Appl. Phys., 1993, 74, 3316.