Materials Science and Engineering A 449–451 (2007) 1079–1082

Ball milling induced crystallization of amorphous Fe90Zr10 P.P. Choi a , I.V. Povstugar b , Y.S. Kwon a,∗ , E.P. Yelsukov b , J.S. Kim a a

Research Center for Machine Parts and Materials Processing, School of Materials Science and Engineering, University of Ulsan, Namgu Mugeo 2-Dong, San 29, Ulsan 680-749, Republic of Korea b Physical-Technical Institute UrB RAS, 132 Kirov Str., 426000 Izhevsk, Russia Received 22 August 2005; received in revised form 29 October 2005; accepted 15 February 2006

Abstract The structural evolution of an amorphous Fe90 Zr10 alloy under high-energy mechanical deformations was investigated by means of X-ray diffraction, Moessbauer spectroscopy and magnetic measurements. Ball milling of melt-spun Fe90 Zr10 ribbons in a planetary ball mill (AGO-2) induces their crystallization into supersaturated ␣-Fe(Zr) grains. The crystallization degree of the amorphous phase increases with increasing milling time, with a fully crystalline state being obtained after 30 min at a milling speed of 1000 rpm. The Zr content of the Fe grains reaches a maximum of about 3 at.% after milling for 30 min. Moessbauer spectra and saturation magnetization measurements suggest that the remaining Zr atoms are most likely segregated at the grain boundaries. Analyses of samples milled at different speeds reveal that the crystallization seems to be a deformation-induced process rather than a thermally induced one. © 2006 Elsevier B.V. All rights reserved. Keywords: Ball milling; Amorphous materials; Crystallization

1. Introduction Mechanical alloying (MA), i.e. ball milling of elemental or intermetallic powder mixtures, is a well known process of producing a wide range of novel materials with unique properties. In comparison with other techniques such as liquid quenching, thermal evaporation or sputtering, etc., it has the advantages of relatively low processing-temperatures, inexpensive equipment and the potential for easy industrial scaling-up. Particular interest has been paid in the past few decades to the formation of amorphous alloys by high-energy ball milling [1–6]. Creation of structural defects [7], a disordering effect [8] or the presence of a high local temperature [9], giving rise to a solid-state reaction, are examples of mechanisms that have been proposed for the milling-induced amorphization but have not yet been completely verified. To open a new route of understanding amorphization during MA, its reverse process, i.e. ball milling induced crystallization, has been studied by a few authors. Trudeau et al. [10] first reported crystallization of melt-spun Fe78 B13 S9 and Fe66 Co18 Si1 B15 metallic glasses upon high-energy ball milling

∗

Corresponding author. Tel.: +82 52 259 2107; fax: +82 52 259 2109. E-mail address: [email protected] (Y.S. Kwon).

0921-5093/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.msea.2006.02.266

and concluded from their results that mechanically induced crystallization is not equivalent to a thermal process. Bansal et al. [11] ascribed the milling-induced crystallization of Fe78 B13 S9 metallic glasses to impurity incorporation during milling and not due to deformation, while Huang et al. [12] suggested heating as a cause for enhanced crystallization kinetics. Xu and Atzmon [13] proposed that mechanically induced crystallization results from diffusivity enhancement due to milling-produced defects. He et al. [14] reported deformation-induced crystallization in several Al-based metallic glasses. The authors observed formation of nanocrystallites in shear bands, induced by bending, and judged heating effects to be unlikely. To date, a full understanding of mechanically driven crystallization has not been reached, and the subject remains controversial. In this work, the structural changes of a model Fe90 Zr10 alloy occurring under high-energy ball milling have been elucidated by means of X-ray diffraction (XRD), Moessbauer spectroscopy and magnetic measurements.

2. Experimental Initial ingots of Fe90 Zr10 composition were prepared by arc melting under high-vacuum conditions and subsequently meltspun to amorphous ribbons of about 50 ␮m in thickness with a

1080

P.P. Choi et al. / Materials Science and Engineering A 449–451 (2007) 1079–1082

single roller unit under Ar atmosphere. The melt-spun ribbons were cut into pieces of less than 1 cm in length and sealed in stainless steel vials together with stainless steel balls under protective Ar atmosphere (5 × 105 Pa). The chosen ribbon to ball weight ratio was 5 g:200 g. Milling was performed in a highenergy planetary ball-mill (AGO-2) at rotation speeds of 300, 400, 600 and 1000 rpm, where the milling time was varied from 5 to 120 min. The vials were water-cooled to avoid excessive heating due to ball-collisions. Structural characterization of the samples was performed by means of XRD with a Rigaku RAD3C, using Cu K␣ radiation. Moessbauer spectra were measured using a spectrometer with a 57 Co(Cr) source operating in constant acceleration mode. A generalizing regular algorithm for the solution of the inverse problem [15] was used to find hyperfine field distribution functions P(H). Magnetic measurements of saturation magnetization were performed using a vibrating sample magnetometer in an external magnetic field of 16 kOe.

Fig. 1 show the XRD patterns of the initial melt-spun ribbons and samples milled for 5, 30 and 120 min at 1000 rpm, respectively. Beside a broad halo around 30◦ , which stems from the sample holder, the XRD pattern of the as-melt-spun ribbon only exhibits a halo peak at 45◦ , indicating a fully amorphous structure within the resolution of the XRD method. Upon ball milling, crystallization of the amorphous phase sets in, and the formation of a bcc crystalline phase with lines close to ␣-Fe can be detected (see Fig. 1). The XRD pattern does not reveal any pronounced

peaks belonging to other phases such as intermetallics. The crystallization proceeds with increasing milling time, as can be seen by increasing intensities of the bcc peaks and peak sharpening. The bcc peaks of the milled samples are substantially shifted to lower angles compared to those of pure ␣-Fe powder [16]. Hence, the milled samples exhibit solution of Zr into the ␣-Fe grains. Applying Vegard’s law, the ␣-Fe(Zr) grains are found to be supersaturated with 1.8, 2.8 and 0.8 at.% Zr for the 5, 30 and 120 min samples, respectively. This change in supersaturation as a function of milling time cannot be presently explained. Weak peaks can be detected around 2Θ angles of 34◦ , 40◦ and 56.5◦ , which cannot be assigned to ␣-Fe(Zr) nor to any oxide or nitride phases and are assumed to stem from a some metastable phase, possibly from the disordered ␾-FeZr2 phase [17]. Further detailed information on structure and phase composition of the samples was gained from Moessbauer spectroscopy analyses. Fig. 2 shows the room temperature Moessbauer spectra of as-melt-spun and milled samples and the corresponding hyperfine field distribution functions P(H). The spectrum of the as-melt-spun state only shows a poorly resolved doublet, stemming from the amorphous phase, which is paramagnetic at room temperature [18]. After 5 min of milling a sextet with asymmetrical lines is superposed to the non-split component in the Moessbauer spectrum. The contribution of this component increases, and after milling for 60 min no doublet is observed in the spectra. The P(H) functions of all milled samples reveal a corresponding component occurring at fields larger than 250 kOe, indicating the formation of bcc ␣-Fe(Zr) solid solution. From the relative intensities of the low-field components of the P(H)

Fig. 1. XRD patterns of as-melt-spun ribbons and samples milled for various milling time.

Fig. 2. Room temperature Moessbauer spectra and corresponding hyperfine field distribution functions of samples milled at 1000 rpm for various milling time.

3. Results and discussion

P.P. Choi et al. / Materials Science and Engineering A 449–451 (2007) 1079–1082

Fig. 3. Moessbauer spectra taken at 77 K and corresponding hyperfine field distribution functions of samples milled under various conditions.

functions, the fraction of the residual amorphous phase was estimated to be 30, 8 and 0% for samples milled for 5, 15 and 30 min, respectively. A fully crystalline state is reached after 30 min of milling. An additional broad component between 50 and 250 kOe can be clearly seen in the P(H) functions of the milled samples (see Fig. 2), that can neither be assigned to the amorphous phase nor to the solid solution. This component may either result from a strongly disordered Fe3 Zr phase, which appears under thermally induced crystallization of amorphous Fe90 Zr10 [19,20], or, more likely, arise from Fe atoms, which have Zr atoms enriched in their nearest neighborhood. These Zr enrichments are likely to be grain boundary segregation, as it was found by XRD and Moessbauer spectra that not more than one third of the Zr atoms pass into the Fe grains. Fig. 3 shows Moessbauer spectra and P(H) functions of initial and milled samples, taken at a temperature of 77 K. Paramagnetic doublet becomes split at this temperature, and the spectrum as well as the corresponding broad P(H) function of the initial sample are typical of an amorphous Fe-Zr sample. The mean hyperfine field amounts to 200 ± 5 kOe, which corresponds to a composition of the amorphous phase close to Fe90 Zr10 according to [19,21]. Upon milling the fraction of the amorphous phase decreases, as already found from room temperature spectra. However, the mean hyperfine field of the amorphous phase component remains about 200 kOe, indicating that ball milling does not induce a compositional change of the residual amorphous phase. The results of magnetic measurements are presented in the Fig. 4. The increase in saturation magnetization ␴S indicates that part of the Zr atoms becomes chemically unbound with Fe atoms. Those atoms can be located at iron grain boundaries, which is consistent with XRD and Moessbauer results mentioned above. Fig. 5 shows the Moessbauer spectra of samples milled at different milling speeds (intensities) for the same duration (30 min). It is seen that the crystallization degree of Fe90 Zr10 increases with increasing milling intensity. The fractions of the residual amorphous phase after milling for 30 min at 300 and 400 rpm

1081

Fig. 4. Saturation magnetization of the samples plotted against the milling time (at 1000 rpm).

amount to 30 and 16%, respectively. After milling at 600 rpm the amorphous phase is not observed in the spectrum. Crystallization of amorphous Fe90 Zr10 under ball milling may be either a deformation-induced process or linked to local temperature rises during ball collisions. At low milling speeds such as 300 and 400 rpm the temperature rise under milling is presumably not more than 200 ◦ C following the results of ref. [22]. Hence, thermally induced crystallization should not occur in this case. Moreover, the crystallization product after ball milling (supersaturated solid solution ␣-Fe(Zr)) strongly differs from those obtained under thermally induced crystallization of Fe90 Zr10 (pure ␣-Fe + Fe3 Zr [19]). An oxygen content ≤100 ppm as well as a Fe impurity concentration ≤0.5 at.%, stemming from the debris of the milling tools, was detected by means of carrier-gas hot-extraction and EDX, respectively. Hence, impurity incorporation during milling can be rather excluded as a major cause of the crystallization process. Therefore, it appears that crystallization of amorphous Fe90 Zr10 under ball milling is predominantly a deformation-induced process.

Fig. 5. Room temperature Moessbauer spectra and corresponding hyperfine field distribution functions of samples milled at 300, 400 and 600 rpm for 30 min.

1082

P.P. Choi et al. / Materials Science and Engineering A 449–451 (2007) 1079–1082

Nevertheless, for higher milling speeds it cannot be completely ruled out that crystallization results from both mechanical and thermal activation. 4. Conclusion Amorphous Fe90 Zr10 ribbons crystallize into supersaturated ␣-Fe(Zr) grains under high-energy ball milling, while intermetallic phases are not formed. The crystallization degree increases with increasing milling time and speed, and a fully crystalline state is reached after milling for 30 min at 1000 rpm. The supersaturation of the Fe(Zr) grains does not exceed 3 at.%. Moessbauer and magnetization measurements strongly suggest that the remaining Zr atoms are segregated at the Fe grain boundaries. As ␣-Fe(Zr) is also formed at low milling speeds, local temperature rises and impurity incorporation can be excluded as major causes of the crystallization process. It can be concluded that the crystallization of amorphous Fe90 Zr10 under ball milling is a deformation-induced process. Acknowledgements This work was supported (in part) by the Ministry of Commerce, Industry and Energy (MOCIE) of the Republic of Korea, through the Research Center for Machine Parts and Materials Processing (ReMM) at the University of Ulsan. We are thankful to Prof. Kim Do-Hyang and Mr. Park Jin-Man from Yonsei University (Korea) for the preparation of melt-spun samples.

References [1] C.C. Koch, O.B. Cavin, G. McKamey, J.O. Scarborough, Appl. Phys. Lett. 43 (1983) 1017. [2] L. Schultz, Mater. Sci. Eng. 97 (1988) 15. [3] E. Hellstern, L. Schultz, Mater. Sci. Eng. 97 (1988) 39. [4] J. Eckert, L. Schultz, K. Urban, Mater. Sci Eng. A 134 (1991) 1389. [5] R.B. Schwarz, C.C. Koch, Appl. Phys. Lett. 49 (1986) 146. [6] K. Yamada, C.C. Koch, J. Mater. Res. 8 (1993) 1317. [7] R.B. Schwarz, R.R. Petrich, C.K. Saw, J. Non-Cryst. Solids 76 (1985) 281. [8] H. Bakker, G.F. Zhou, H. Yang, Mater. Sci. Forum 179–181 (1995) 47. [9] J. Eckert, L. Schultz, E. Hellstern, K. Urban, J. Appl. Phys. 64 (1988) 3224. [10] M.L. Trudeau, R. Schulz, D. Dussault, A. Van Neste, Phys. Rev. Lett. 64 (1990) 99. [11] C. Bansal, B. Fultz, W.L. Johnson, Nanostruct. Mater. 4 (1994) 919. [12] B. Huang, R.J. Perez, P.J. Crawford, A.A. Sharif, S.R. Nutt, E.J. Lavernia, Nanostruct. Mater. 5 (1995) 545. [13] J. Xu, M. Atzmon, Appl. Phys. Lett. 73 (1998) 1805. [14] Y. He, G.J. Shiftlet, S.J. Poon, Acta Metall. Mater. 43 (1995) 83. [15] E.V. Voronina, N.V. Ershov, A.L. Ageev, Yu.A. Babanov, Phys. Status Solidi (b) 160 (1990) 625. [16] J.D. Hanawalt, H.W. Rinn, L.K. Frevel, Anal. Chem. 10 (1938) 475. [17] S.E. Lee, H.Y. Ra, T.H. Yim, W.T. Kim, Mater. Sci. Forum 179–181 (1995) 121. [18] H. Tange, Y. Tanaka, M. Goto, K. Fukamichi, J. Magn. Magn. Mater. 81 (1989) L243. [19] M. Grafari, U. Gonser, H.G. Wagner, M. Naka, Nucl. Instrum. Methods 199 (1982) 197. [20] L.F. Kiss, G. Huhn, T. Kemeny, J. Balogh, D. Kaptas, J. Magn. Magn. Mater. 160 (1996) 229. [21] K.M. Unruh, C.L. Chien, Phys. Rev. B 30 (9) (1984) 4968. [22] E.V. Shelekhov, T.A. Sviridova, Materialovedenie 10 (1999) 13 (in Russian).