Devitrification of Mechanically Alloyed Fe-Nb System: Mössbauer Study of the Intermetallic Phases

Metallurgical and Materials Transactions A 51 (2020) 1395-1401 1 Devitrification of mechanically alloyed Fe-Nb system: Mössbauer study of the intermetallic phases A.F. Manchón-Gordón1, P. Svec2, J.J. Ipus1, M. Kowalczyk3, J.S. Blázquez*1, C.F. Conde1, A. Conde1, P. Svec Sr.2, T. Kulik3 1Dpto. Física de la Materia Condensada, ICMSE-CSIC, Universidad de Sevilla, P.O. Box 1065, 41080 Sevilla, Spain 2Institute of Physics. Slovak Academy of Sciences. Dúbravská cesta 9, 845 11 Bratislava 45, Slovak Republic. 3 Faculty of Materials Science and Engineering, Warsaw University of Technology, 141 Wołoska st., 02-507 Warsaw, Poland *The corresponding author e-mail: [email protected] ABSTRACT: Intermetallic phases in the Fe-Nb system have been obtained as products of the devitrification of a homogeneous amorphous Fe70Nb30 alloy prepared by mechanical alloying. Besides Fe2Nb Laves and Fe7Nb6 intermetallic phases, 𝛼-Fe phase has been detected by X-ray diffraction (XRD). Hyperfine parameters for both Fe2Nb and Fe7Nb6 intermetallics have been obtained from Mössbauer spectroscopy in correlation with phase identification from XRD results. Thermomagnetic measurements show changes in the Curie temperatures of the amorphous and Fe2Nb phases during the crystallization, due to compositional variations related to the developing of the 𝛼-Fe phase. Kinetics of the crystallization process has been analyzed using the classical Johnson-Mehl-Avrami-Kolmogorov kinetic theory in both isothermal and nonisothermal regimes. KEYWORDS: Fe-Nb intermetallics, amorphous and nanocrystalline alloys, Mössbauer spectroscopy, crystallization kinetics. 1. INTRODUCTION Metallurgical and Materials Transactions A 51 (2020) 1395-1401 2 Fe-Nb systems have received a particular interest in the research community essentially due to their wide range of different applications in industry, mainly as special stainless steels1-4 and as soft magnets5,6. Indeed, the presence of Nb in the Fe matrix limits the formation of 𝛼-Fe grains to the nanometric scale, improving the mechanical properties (through solid solution and/or precipitation strengthening1,2) and the magnetic softness (through averaging out the magnetocrystalline anisotropy7). In spite of the interest on this system, there are some discrepancies in the literature on the complete Fe-Nb phase diagram, particularly on the melt and the homogeneity ranges of the intermetallic phases8-10. In fact, while a previous Fe-Nb phase diagram11 showed four intermetallics (including a possible high temperature phase at 90 at.% Nb), recent results indicate the existence of only two intermetallic phases, the hexagonal Fe2Nb and the trigonal Fe7Nb6 phases8. Many techniques have been used to produce Fe-Nb alloys. However, the poor solubility of niobium in the iron lattice (Fe-Nb mixing enthalpy -16 kJ/mol12) limits the production of homogenous Fe-Nb systems by melting methods to concentrations higher than 33 at.% Nb13, due to the large difference between the melting temperatures of these elements (1811 and 2750 K for Fe and Nb, respectively8). Therefore, the production of samples out of the eutectic compositions requires other ways, such as mechanical alloying. This method often produces non-equilibrium solid solutions and amorphous phases14-16. Consequently, additional thermal treatments are necessary to develop stable crystalline phases. In this sense, the combination of X-ray diffraction (XRD) and Mössbauer spectroscopy (MS) techniques in iron-based systems allows studying in detail the final products of the crystallization process17,18. Metallurgical and Materials Transactions A 51 (2020) 1395-1401 3 Concerning the devitrification process, crystallization kinetics can be characterized by Johnson and Mehl19, Avrami20 and Kolmogorov21 classical crystallization theory (JMAK theory). Although JMAK theory was developed to be applied to isothermal polymorphic transformations, it has been successfully extended to transformations implying compositional changes22,23 and to non-isothermal processes24-29. Compared to isothermal experiments, non-isothermal processes have the advantages of an easier operation, lower time cost and higher signal-to-noise ratio28. In this work, an amorphous Fe alloy with 30 at.% Nb has been synthesized by means of mechanical alloying in order to obtain a homogeneous sample. Different thermal treatments have been applied in order to develop stable intermetallic Fe-Nb phases. The products of the crystallization process have been studied by X-ray diffraction (XRD), Mössbauer spectroscopy (MS) and magnetometry. JMAK theory has been used to describe the crystallization process for non-isothermal and isothermal regimes. 2. EXPERIMENTAL Amorphous alloy with Fe70Nb30 at.% composition was prepared by high energy ball milling, starting from high purity elemental powders of iron and niobium (>99.9 %). The milling was carried out for 50 hours at 350 rpm in a planetary ball mill Fritsch Pulverisette Vario 4. Further milling parameters and a detailed microstructural evolution of the sample with milling time can be found elsewhere16. Differential thermal analysis (DTA) was carried out in a Perkin-Elmer DTA7 unit under Ar flow for both isothermal and non-isothermal experiments. Different heating rates (β=5, 10, 20, 40 and 60 K/min) were used for non-isothermal treatments up to 1300 K without evidencing any changes (see supplementary material). In order to follow the Metallurgical and Materials Transactions A 51 (2020) 1395-1401 4 progress of the transformations, heating at 10 K/min up to different temperatures (from 908 to 1300 K) was also performed. In the case of isothermal experiments, the amorphous sample was annealed for 5 hours at 954 K, 10 K below the onset temperature obtained for 𝛽=10 K/min (see Fig. 1a). The corresponding baselines were obtained repeating the same treatment after the complete crystallization of the sample. XRD patterns were recorded in a Bruker D8 Advance diffractometer (Cr-K𝛼, 𝜆=2.28970 Å) in a Bragg-Brentano geometry. The local environment of Fe atoms was analyzed at room temperature by Mössbauer spectrometry in transmission geometry using a 57Co(Rh) source. Values of the hyperfine parameters were obtained by fitting the measured spectra with the NORMOS program30 and the isomer shift (IS) was quoted relative to that of an α-Fe foil at room temperature. Magnetization measurements were performed in a vibrating sample magnetometer (standard option of a Quantum Design Physical Properties Measurement System, PPMS) applying a magnetic field of 100 Oe. Curie temperatures 𝑇𝐶 have been determined by identifying the temperature at which the temperature derivative of magnetization shows a minimum. 3. RESULTS and DISCUSSION 3.1. Characterization of crystallization products Figure 1a shows the non-isothermal DTA plot taken at 10 K/min for the as-milled amorphous sample. All the features found in the curve vanish in a second heating, which was used as the corresponding baseline. Deviations from baseline at lower temperatures are ascribed to relaxation phenomena. A well-defined exothermic peak can be observed around 950 K, corresponding to the crystallization of the amorphous phase. The measured peak shape in Fig. 1a suggests that at least two transformations are taking Metallurgical and Materials Transactions A 51 (2020) 1395-1401 5 place. However, no experimental deconvolution could be obtained in the heating rate from 5 to 60 K/min (see figure 1b). A second and weaker peak is observed at ~1200 K in Fig. 1a. The effective value of the activation energy 𝐸𝑎=2.48±0.07 eV/at, corresponding to the first peak temperature, has been calculated by Kissinger’s method31 (see inset of Fig. 1b). This value is similar to that reported by El-Eskandarany et al.32 for the crystallization of the amorphous Fe52Nb48 alloy powder obtained by mechanical alloying. 800 900 1000 1100 1200 1300 900 1000 1100 (b) exo 0.05 W/g Heat flow Tonset (a) 0.96 1.00 1.04 -12 -11 -10 ln(/T2 p) 1/Tp (10-3 K-1) 60 40 20 Temperature (K) exo 0.3 W/g 5 10 Figure 1. (a) Non-isothermal DTA scan at 10 K/min for the as-milled sample. Arrows indicate the temperatures up to which different samples have been heated. The onset temperature is indicated. (b) Non-isothermal DTA scans for as-milled sample at the different indicated heating rates (in K/min) in the region of the first DTA exotherm. The inset shows the Kissinger’s plot. Metallurgical and Materials Transactions A 51 (2020) 1395-1401 6 Figure 2 shows the XRD patterns of as-milled sample and of those samples after heating at 10K/min up to the indicated temperatures to evidence the formation of the different crystalline phases. The as-milled powders only show a broad halo distinctive of an amorphous phase. This broad halo is the only feature observed in the patterns collected for samples heated up to 907 K. Above this temperature, several diffraction peaks appear superimposed to the amorphous halo, corresponding to three phases: hcp Fe2Nb (after heating up to 968 K, with space group P63/mmc), bcc-Fe (after heating up to 1003 K, with space group Im3m) and trigonal Fe7Nb6 (space group R-3m, after heating up to 1273 K). 60 80 100 120 140 Fe2Nb Fe Fe7Nb6 1300 K 1003 K 983 K 968 K Intensity (a.u.) 2(degrees) as-milled 907 K Figure 2. XRD patterns taken at room temperature of as-milled amorphous powders and of samples heated at 10 K/min up to the indicated temperatures. These temperatures correspond to those indicated by arrows in Figure 1a. Figure 3 shows the room temperature Mössbauer spectra of as-milled and heat treated samples up to the different indicated temperatures at 𝛽=10 K/min. A quadrupolar distribution has been used to describe the paramagnetic contribution, assigned to the amorphous phase. The quadrupolar distribution is the unique contribution that exhibits Metallurgical and Materials Transactions A 51 (2020) 1395-1401 7 the alloy heated up to 907 K, in agreement with the XRD results, showing only an amorphous phase. Above this temperature, MS spectra show a decrease of the relative area of the quadrupolar distribution along with the appearance of a doublet, indicating the migration of Fe atoms to an ordered structure which corresponds to the Fe2Nb intermetallic detected by XRD. After heating up to 983 K, besides the quadrupolar distribution and the doublet, a sextet appears with an hyperfine magnetic field of 32.84±0.09 T corresponding to the 𝛼-Fe phase, detected also by XRD but after heating up to 1003 K. In fact, MS is a more sensitive technique than XRD in order to detect small fractions of Fe-phases when paramagnetic and ferromagnetic phases coexist16. After heating up to 1300 K, the alloy is totally crystalline, evidenced by the absence of any quadrupolar distribution. A new doublet was used to fit the spectrum showing the appearance of a new ordered phase, in agreement with the detection of Fe7Nb6 in the corresponding XRD patterns. The presence of doublets is frequently reported in the Mössbauer spectra of Fe-Nb alloys32-34. -8 -6 -4 -2 0 2 4 6 8 -8 -6 -4 -2 0 2 4 6 8 -8 -6 -4 -2 0 2 4 6 8 amorphous Intensity (a.u.) 907 K 968 K 983 K Intensity (a.u.) Velocity (mm/s) 1003 K Velocity (mm/s) 1300 K Velocity (mm/s) Metallurgical and Materials Transactions A 51 (2020) 1395-1401 8 Figure 3. Experimental Mössbauer spectra taken at room temperature (symbols) and model fitting (lines) of as-milled and heat treated samples at 10 K/min up to the indicated temperatures. These temperatures correspond to those indicated by arrows in Figure 1a. Green, black, red and blue subspectra correspond to Fe atoms in the amorphous, Fe2Nb, α-Fe and Fe7Nb6 phases, respectively. Figure 4a shows the evolution of the fraction areas of different contributions to Mössbauer spectra as a function of the final heating temperature at 𝛽=10 K/min (values at 300 K correspond to the as-milled sample with 100% amorphous phase). Doublet 1 (ascribed to Fe sites in Fe2Nb) is the first contribution detected and its fraction shows a continuous increase up to ~50 % as the final heating temperature increases. The ferromagnetic site (ascribed to 𝛼-Fe) is clearly detected after heating up to 983 K. On the other hand, doublet 2 (ascribed to Fe sites in Fe7Nb6) is not detected until the sample is heated up to 1300 K, when the area fraction of this contribution reaches ~20 %. The different fraction areas of the samples are practically independent of the heating rate, as is shown in Fig. 4b. Only in the case of 𝛽=5 K/min, the decrease of the doublet 1 fraction is compensated by an increase of the fraction of the ferromagnetic site and the doublet 2. 300 900 1000 1100 1200 1300 0 20 40 60 80 100 010 20 30 40 50 60 Phase fraction (%) T (K) (b) quadrupolar distribution sextet doublet 2 Heating rate (K/min) doublet 1 (a) Figure 4. Evolution of the phase fraction obtained by MS as a function of a) the maximum heating temperature reached at 10 K/min and b) the heating rate up to 1300 K. Lines are guides to the eye. Metallurgical and Materials Transactions A 51 (2020) 1395-1401 9 The fitted hyperfine values for doublet 1, attributed to the Fe2Nb crystalline phase, are a quadrupole splitting Q=0.40±0.02 mm/s and an isomer shift 𝛿=-0.30±0.01 mm/s. Doublet 2, assigned to the Fe7Nb6 intermetallic, exhibits a Q=0.45±0.07 mm/s and 𝛿=- 0.03±0.04 mm/s. These values and the average values found for the as-milled amorphous sample are shown in Table 1 and can be compared to those reported in the literature for samples prepared by different methods35-38. Metallurgical and Materials Transactions A 51 (2020) 1395-1401 16 show the evolution of the Curie temperatures of the amorphous and Fe2Nb phases during the crystallization, as well as the developing of the bcc-Fe crystalline phase. JMAK theory has been used to describe the crystallization process in both isothermal and non-isothermal regimes. A stronger growth impingement in the isothermal regime can explain the lower Avrami exponent found with respect to those obtained for nonisothermal treatments. Acknowledgements This work was supported by AEI/FEDER-UE (Project MAT 2016-77265-R) and the PAI of the Regional Government of Andalucía. Support of the project APVV-15-0621 is also acknowledged. A.F. Manchón-Gordón acknowledges his contract to the VPPIUS of the University of Sevilla. Metallurgical and Materials Transactions A 51 (2020) 1395-1401 17 References 1 N. Fujita, K. Ohmura, A. Yamamoto: Mater. Sci. Eng., A. 2003, vol. 351, pp. 272-281. 2 G.M. Sim, J.C. Ahn, S.C. Hong, K.J. Lee, K. S. Lee: Mater. Sci. Eng., A. 2005, vol. 396, pp. 159-165. 3 A. Malfliet, W. Van den Broek, F. Chassagne, J.D. Mithieux, B. Blanpain, P. Wollants: J. Alloy Comp. 2011, vol. 509, pp. 9583-9588. 4 Y. Kang, W.M. Mao, Y.J. Chen, J. Jing, M. Cheng: Mater. Sci. Eng., A. 2016, vol. 677, pp. 453-464. 5 M.E. McHenry, M.A. Willard, D.E. Laughlin: Prog. Mater. Sci. 1999, vol. 44, pp. 291-433. 6 L.E. Zamora, G.A.P. Alcazar, J.A. Tabares, J.D. Betancur, F.R. Sives, J. Jaen, J.M. Greneche, J.F. Marco, J.M. Gonzalez: J. Phys. D: Appl. Phys. 2008, vol. 41, p. 155010. 7 G. Herzer: Scr. Metall. Mater. 1995, vol. 33, pp. 1741-1756. 8 S. Voss, M. Palm, F. Stein, D. Raabe: J. Phase Equilib. Diffus. 2011, vol. 32, pp. 97-104. 9 K.W. Li, X.B. Wang, S.M. Li, W.X. Wang, S.P. Chen, D.Q. Gong, J.L. Cui: High. Temp. Mater. Processes (Berlin, Ger.) 2015, vol. 34, pp. 479-485. 10 C.G. Schon, J.A.S. Tenorio: Intermetallics 1996, vol. 4, pp. 211-216. 11 H.J. Goldschmidt: Research (London) 1957, vol. 10, pp. 289-291. 12 A. Takeuchi, A. Inoue: Mater. Trans., JIM 2000, vol. 41, pp. 1372-1378. 13 E.P. Abrahamson, S.L. Lopata: Trans. Metall. Soc. AIME 1966, vol. 236, pp. 76-87. 14 J.J. Ipus, J.S. Blázquez, V. Franco, A. Conde, L.F. Kiss: Intermetallics 2010, vol. 18, pp. 565-568. 15 J.S. Blázquez, J.J. Ipus, S. Lozano-Perez, A. Conde: JOM 2013, vol. 65, pp. 870-882. 16 A.F. Manchón-Gordón, J.J. Ipus, J.S.; Blázquez, C.F. Conde, A. Conde: J. Non-Cryst. Solids 2018, vol. 494, pp. 78-85. 17 J. Balogh, L. Bujdoso, D. Kaptas, T. Kemeny, I. Vincze, S. Szabo, D.L. Beke: Phys. Rev. B 2000, vol. 61, pp. 4109-4116. 18 J.S. Blázquez, J.J. Ipus, V. Franco, C.F. Conde, A. Conde: J. Alloy Comp. 2014, vol. 610, pp. 92-99. 19 W.A. Johnson, R.F. Mehl: Tech. Publ. - Am. Inst. Min. Metall. Eng. 1939, 135, 416–442. 20 M. Avrami: J. Chem. Phys. 1941, vol. 9, pp. 177-184. 21 A.N. Kolmogorov, Bull. Russ. Acad. Sci. 1937, vol. 1, pp. 355–359. 22 P. Bruna, D. Crespo, R. Gonzalez-Cinca, E. Pineda: J. Appl. Phys. 2006, vol. 100, p. 054907. 23 J.S. Blázquez, M. Millán, C.F. Conde, A. Conde: J. Alloy Comp. 2010, vol. 505, pp. 91-95. 24 T. Ozawa, Polymer 1971, vol. 12, pp. 150-158. 25 K. Nakamura, T. Watanabe, K. Katayama: J. Appl. Polym. Sci. 1972, vol. 16, pp. 1077- 1091. 26 K. Nakamura, T. Watanabe, T. Amano: J. Appl. Polym. Sci. 1974, vol. 18, pp. 615-623. 27 M.J. Starink, A.M. Zahra: Thermochim. Acta 1997, vol. 298, pp. 179-189. 28 J.S. Blázquez, C.F. Conde, A. Conde: Acta Mater. 2005, vol. 53, pp. 2305-2311. 29 J. Torrens-Serra, P. Bruna, S. Roth, J. Rodriguez-Viejo, M.T. Clavaguera-Mora: J. Alloy Comp. 2010, vol. 496, pp. 202-207. 30 R.A. Brand, J. Lauer, D.M. Herlach: J. Phys. D: Appl. Phys. 1983, vol. 13, pp. 675-683. 31 H.E. Kissinger: Anal. Chem. 1957, vol. 29, pp. 1702-1706. 32 M.S. El-Eskandarany, A.A. Bahgat, N.S. Gomaa, N.A. Eissa: J. Alloy Comp. 1999, vol. 290, pp. 181-190. 33 E. Jartych, D. Oleszak, J.K. Zurawicz: Hyperfine Interact. 2001, vol. 136, pp.25-33. 34 G.Y. Velez, G.A. Perez Alcazar, L.E. Zamora: J. Magn. Magn. Mater. 2014, vol. 354, pp. 333-335. 35 M.T. Raposo, J.D. Ardisson, A.I.C. Persiano, R.A. Mansur: Hyperfine Interact. 1994, vol. 83, pp. 235-238. Metallurgical and Materials Transactions A 51 (2020) 1395-1401 18 36 G.Y. Velez, G.A. Perez Alcazar, L.E. Zamora, J.A. Tabares: J. Supercond. Novel Magn. 2014, vol. 27, pp. 1279-1283. 37 A. Blachowski, K. Ruebenbauer, J. Zukrowski: Phys. Status Solidi B 2005, vol. 242, pp. 3201-3208. 38 M.J.M. Pires, W.A.A. Macedo, L.P. Cavalcanti, A.M.G. Carvalho, J. Phys. Chem. Solids 2015, vol. 86, pp. 36-41. 39 J.S. Blázquez, J.M. Borrego, C.F. Conde, A. Conde, S. Lozano-Perez: J. Alloy Comp. 2012, vol. 544, pp. 73-81.