Tuning the Curie Temperature in γ-FeNi Nanoparticles for Magnetocaloric Applications by Controlling the Oxidation Kinetics

Tuning the Curie temperature in c-FeNi nanoparticles for magnetocaloric applications by controlling the oxidation kinetics Huseyin Ucar, 1,a) John J. Ipus, 2 D. E. Laughlin, 1 and M. E. McHenry 1 1 Materials Science and Engineering, Carnegie Mellon University, Pittsburgh, Pennsylvania 15213, USA 2 Dpto. Fısica de la Materia Condensada, ICMSE-CSIC, Universidad de Sevilla, Spain (Presented 16 January 2013; received 10 October 2012; accepted 14 January 2013; published online 20 March 2013; publisher error corrected 3 April 2013) Mechanically alloyed Fe 70 Ni 30 and Fe 72 Ni 28 alloys were characterized in terms of their structural and magnetic properties. Previous studies showed that single phase FCC c-FeNi alloys with 26-30 at. % Ni have Curie temperatures, T c , near room temperature. Having T c near room temperatures along with large magnetization makes c-FeNi alloys attractive for room temperature magnetocaloric cooling technologies. To obtain a single c-phase, particles were solution annealed in the c-phase field and water quenched. The preferential oxidation of Fe during ball milling was used as a means to tune the Curie temperature, T c ,of the alloy. Refrigeration capacities, RC FWHM ,oftheFe 70 Ni 30 and the Fe 72 Ni 28 alloys were calculated to be 470 J/kg and 250 J/kg at 5 T, with peak temperatures 363 K and 333 K, respectively. The RC FWHM for the Fe 70 Ni 30 is higher than the previously reported Nanoperm (Fe 70 Ni 30 ) 89 Zr 7 B 4 type alloy and on the same order of magnitude with other Fe-based alloys. The maximum magnetic entropy change values observed for the Fe 70 Ni 30 and the Fe 72 Ni 28 are 0.65 and 0.5 J kg 1 K 1 , respectively, at a field of 5 T. These are smaller than those of rare earth magnetic refrigerants showing first order transformation behavior. The larger RC FWHM value results mainly from the width of the magnetic entropy curve in these types of materials. We discuss the economic advantage of these rare earth free refrigerants. V C2013 American Institute of Physics.[http://dx.doi.org/10.1063/1.4795012] Soft magnetic c-FeNi nanostructures are attractive candidates for magnetocaloric applications. They have low hysteresis losses and tunable Curie temperatures, T C , with minor compositional changes. They also can be suspended in solutions thus providing versatility in applications. A relatively low magnetic entropy change of these alloys would yield a small adiabatic temperature change at a single temperature but tuning over the potential to span a range of temperatures in a thermodynamic cycle can impact thermodynamic efficiency. To produce c-FeNi nanoparticles, one technique is mechanical alloying (MA) of powders using high energy ball milling followed by solution annealing treatment. In MA, high energy ball milling generates heat due to grinding of steel balls against the powder which causes oxidation of nanoparticles. Here, we demonstrate the tunability of the T C of c-FeNi nanoparticles by controlling the oxidation kinetics during synthesis. In addition, we propose c-FeNi alloys as an alternative to other magnetocaloric refrigerants operating near room temperature. Previous work on controlling oxidation kinetics and interfaces in the FeCo system for RF applications inspired us to tune the T C of c-FeNi metastable phase with similar processes. 1,2 Temperature increases during milling have been reported to be 300 C for the Fe-1.2%C system 3 and 180 C for the Ni-Zr binary system. 4 In another study, it was reported that 80% of the temperature rise during milling comes from the motor and bearings, while 20% is due to the friction between the balls and milling container and the presence of an exothermic reaction. 5 The temperature rise in Fe-Ni system is expected to be on the same order of magnitude with those cited alloy systems. Fe 70 Ni 30 and Fe 72 Ni 28 alloys were produced by ball milling in a shaker mill (Spex 8000D) from elemental Fe, Ni powders under Ar atmosphere. The initial powder mass was 8 g and the ball to powder ratio was 10:1. After selected milling times as mechanically alloyed powder samples were taken out from the vials to characterize crystal structure by x-ray diffraction (XRD) usingCuK a radiation in an X’Pert PRO MPD diffractometer. In order to obtain a material with a single cFeNi phase, Fe 70 Ni 30 and Fe 72 Ni 28 powders were subsequently sealed in a quartz crucible with Ar atmosphere and annealed in the c-phase region, 700 C, and quenched in water to stabilize the metastable cFeNi phase. Besides the structure using x-ray diffractometer, magnetic properties of the cFeNi phase were studied using a Lakeshore 7407 vibrating sample magnetometer using a maximum applied field of 0.55 T at constant temperatures in the range of room temperature to 523 K. The magnetic entropy change due to the application of a magnetic field has been calculated using a numerical approximation to the equation DSM¼ðHmax 0 @M @T  H dH;(1) where DS M is the magnetic entropy change, M is the magnetization, and T is the temperature. The partial derivative is replaced by finite differences and the integration is performed numerically from zero to the maximum value of the applied magnetic field. In this study, RC FWHM , defined as the product of the peak entropy change times the full width at half maximum (FWHM) of the peak DT, RCFWHM ¼jDSpk MjDTwas used as a) Author to whom correspondence should be addressed. Electronic mail: [email protected]. 0021-8979/2013/113(17)/17A918/3/$30.00 V C2013 American Institute of Physics113, 17A918-1 JOURNAL OF APPLIED PHYSICS 113, 17A918 (2013) 19 May 2025 14:17:22 a figure of merit. For materials that did not cover sufficient area around the peak, e.g., Fe 72 Ni 28 ,RC FWHM was estimated by extrapolating the experimental data to the temperatures required for this calculation. Fig. 1presents x-ray diffraction patterns from powders of both Fe 70 Ni 30 and Fe 72 Ni 28 alloys ball milled for 10, 30, 50 h and powders of these which were annealed in the c-phase field followed by quenching. The diffraction patterns for 10 and 30 h milled Fe 70 Ni 30 and Fe 72 Ni 28 powders reveal both fcc and bcc phases and the fcc ratio is observed to increase with milling time for Fe 70 Ni 30 while it is invariant for Fe 72 Ni 28 . While Fe 70 Ni 30 exhibits only fcc peaks and additional oxide peaks that are indexed to FeO (wustite), Fe 72 Ni 28 alloy is observed to retain both fcc and bcc phases after 50 h of milling. While wustite is observed for the two compositions after solution annealing the 50 h milled alloys, the c-stabilized Fe 70 Ni 30 exhibits NiFe 2 O 4 (NiFe oxide) peaks in addition to the wustite. The lattice parameter of the wustite was found to be a ¼4.290 6003 A ˚which agrees well with the results from JCPDS index of wustite. 6 Even though the equilibrium phase diagram dictates the coexistence of bcc and fcc at low temperatures, the supersaturation of fcc c-FeNi is obvious from 50 h ball milled powders. 7 The two phase region is reported to narrow with milling intensity which is caused by the temperature dependence of the metastable equilibria. 8 Here, we observe the narrowing of the twophase field with milling time at increased temperature attributed to the additional dissolution of Fe in fcc FeNi. The heat generated due to grinding of steel balls against the powder leads to oxidation of particles as is seen in the XRD patterns of 50 h as milled powders of Fe 70 Ni 30 and c-phase stabilized 50 h milled powders of both compositions. In the Fe–Ni system, there is a strong dependence of the Curie temperature, T C ,on composition in the c-phase. 9 Since wustite and nickel-iron oxides are Fe-rich, their formation depletes Fe from the core. This in turn changes the T C of the system. Tuning the RC with the T C can be accomplished with the formation of an oxide layer. In tuning the RC, exchange coupling of the 2-phase core/shell structure is exploited. This is analogous to disorder induced broadening of the transition due to distributed exchange interactions. 10,11 Magnetocaloric response is calculated using Eq. (1) using isothermal magnetization curves. Fig. 2illustrates the magnetic entropy change, DS m , for annealed c-Fe 70 Ni 30 and c-Fe 72 Ni 28 powders for a maximum applied field of 0.55 T. The maximum magnetic entropy change values observed for annealed c-Fe 70 Ni 30 are 0.32, 0.22, and 0.17 J kg 1 K 1 with peak entropy change temperatures of 333, 363, and 443 K for 10, 30, and 50 h of ball milling, respectively. From the Fe-Ni phase diagram 7 and the composition dependence of the Curie temperatures of each powder, the Fe content in the nanoparticles is estimated. The Fe content was 71 at. % Fe after 10 h of milling and systematically decreases to 69 and 67 at. % Fe for 30 and 50 h of milling. The maximum entropy change is 0.19 J kg 1 K 1 with a peak entropy change around 333 K for 50 h milled c-Fe 72 Ni 28 powder. Since the Curie temperatures of the 10 and 30 h milled c-Fe 72 Ni 28 powders were slightly below the room temperature, their peak entropy and other magnetocaloric properties could not be assessed. The DS m and RC FWHM can be expressed as a power law for materials with a second order phase transition 12 DSm¼AHn;(2) where A is a prefactor. With this equation, DS m and RC FWHM can be extrapolated to higher fields, which is necessary for comparing the magnetocaloric response with some of the benchmark refrigerants. The magnetic entropy was calculated using the experimental results obtained for a range of fields from 100 Oe to 5500 Oe in 200 Oe field increments and fitted to Eq. (2). The fitting parameters obtained from the experimental data were used to extrapolate the DS m and RC FWHM .In calculating the RC FWHM , the aforementioned relationship, RCFWHM ¼jDSpk MjDT, was used. The extrapolated RC FWHM at 5T for c-Fe 70 Ni 30 are 342, 470, and 324 J K 1 for 10, 30, and 50 h milled powders, respectively. The fitting parameters from Eq. (2) are A ¼0.0320, 0.0283, 0.0134 and n ¼0.858, 0.898, and 0.933 for 10, 30, and 50 h milled powders, respectively. The decrease in RC after 50 h of milling is due mainly to the increased oxide formation as it lowers the magnetic moment. The extrapolated RC FWHM value for 50 h milled Fe 72 Ni 28 is 250 J K 1 with A ¼0.0141 and n ¼0.905. The peak magnetic entropy change and RC values of the studied alloys with their extrapolations to 5 T are shown in Table I. Other magnetocaloric materials are listed for comparison. FIG. 1. X-ray diffraction patterns of 10, 30, 50 h as-milled powders and solution annealed powders at 700 C for 1 h followed by quenching for (a) Fe 70 Ni 30 and (b) Fe 72 Ni 28 . 17A918-2 Ucar et al. J. Appl. Phys. 113, 17A918 (2013) 19 May 2025 14:17:22 From Table I,Gd 5 Ge 1.9 Si 2 Fe 0.1 and Fe 88 Zr 7 B 4 Cu 1 have slightly higher refrigeration capacities than c-Fe 70 Ni 30 and c-Fe 72 Ni 28 alloys. However, as previous study on Fe based alloys showed, Fe-Ni system is economically more advantageous which would facilitate its industrial scale-up. 21 Fe based alloys suffer from relatively low magnetic entropy change, which would result in low adiabatic temperature changes but may be attractive in cycles designed to span large temperature ranges, i.e., between the freezing and boiling points of typical fluid carriers. Microstructure and magnetocaloric response of mechanically alloyed Fe 70 Ni 30 and Fe 72 Ni 28 compositions have been studied with the following observations: (1) With milling time, we found that the bcc phase was destabilized with respect to fcc phase. This also agrees well with previous studies, which state that the composition ranges of the fcc and bcc single phase regions are greatly extended with respect to their equilibrium ranges by mechanical alloying. 8,22 (2) Extended milling times induce oxidation of particles due to high temperatures achieved during operation. This can be utilized to tune the T C of c-FeNi alloy since preferential oxidation of Fe atoms changes the composition of the particles. (3) Fe 70 Ni 30 and Fe 72 Ni 28 are introduced as an alternative to other rare-earth containing magnetocaloric refrigerants operating near room temperature. H.U., M.E.M., and D.E.L. acknowledge support of the NSF through Grant No. DMR #0804020. 1 N. J. Jones et al., J. Appl. Phys. 107, 09A304 (2010) 2 K. N. Collier et al., J. Appl. Phys. 105, 07A328 (2009). 3 R. M. Davis et al., Metall Trans. A 19, 2867 (1988). 4 R. Schulz et al., Phys. Rev. Lett. 62, 2849 (1989). 5 C. Koch, J. Mechanochem Mech. Alloy. 1, 56, (1994). 6 Joint Committee on Powder Diffraction Standards, The International Center for Diffraction Data, 1999, File No. 89-0686. 7 Binary Alloy Phase Diagrams, 2nd ed., edited by T. B. Massalski, H. Okamoto, P. R. Subramanian, and L. Kacprzak (ASM International, 1990), Vol. 2, p. 1736. 8 C. Kuhrt and L. Schultz, J. Appl. Phys. 73, 1975 (1993). 9 K. J. Miller et al., J. Appl. Phys. 107, 09A305 (2010). 10 N. J. Jones et al., “The effect of distributed exchange parameters on magnetocaloric refrigeration capacity in amorphous and nanocomposite materials,” J. Appl. Phys. 111, 07A334 (2012). 11 K. A. Gallagher et al., “Distributed exchange interactions and temperature dependent magnetization in amorphous Fe 88x Co x Zr 7 B 4 Cu 1 Alloys,” J. Appl. Phys. 85, 5130–5132 (1999). 12 V. Franco et al., Appl. Phys. Lett. 90, 052509 (2007). 13 J. J. Ipus et al., J. Appl. Phys. 105, 123922 (2009). 14 W. Dagula and O. Tegus, IEEE Trans. Magn. 41, 10 (2005). 15 S. Fujieda et al., Appl. Phys. Lett. 81, 1276 (2002). 16 P. Gorrıa et al., J. Phys. D: Appl. Phys. 41, 192003 (2008). 17 V. Provenzano et al., Nature 429, 853 (2004). 18 R. Caballero-Flores et al., Appl. Phys. Lett. 96, 182506 (2010). 19 J. J. Ipus et al., IEEE Trans. Magn. 47, 10 (2011). 20 J. Y. Law et al., J. Alloys Compd. 508, 14 (2010). 21 H. Ucar et al., JOM 64, 782 (2012). 22 L. B. Hong and B. Fultz, J. Appl. Phys. 79(8), 3946 (1996). FIG. 2. Temperature dependence of the magnetic entropy change, DS M , corresponding to a magnetic field change DH M ¼0.55 T for (a) cFe 70 Ni 30 and (b) cFe 72 Ni 28 at different milling times followed by an additional solution annealing treatment to obtain the cphase. TABLE I. Peak temperature, peak entropy change, RC FWHM ,RC AREA , and RC WP values of promising magnetocaloric refrigerants are presented. T pk jDS pk M j(1.5 T) RC FWHM (1.5 T) RC FWHM (2 T) RC FWHM (5 T) RC AREA (5 T) RC WP (5 T) Nominal composition (K) J kg 1 K 1 Jkg 1 Jkg 1 Jkg 1 Jkg 1 Jkg 1 Ref. Fe 75 Nb 10 B 15 250 0.6 115 445 283 13 Mn 1.1 Fe 0.9 Pe 0.78 Ge 0.22 280 10 400 14 La(Fe 0.89 Si 0.11 ) 13 H 1.3 291 165.6 15 Pr 2 Fe 17 300 573 16 Gd 5 Fe 1.9 Si 2 Fe 0.1 300 630 355 240 17 Fe 88 Zr 7 B 4 Cu 1 300 1.3 166 654 496 18 Fe 72 Ni 28 333 0.5 84 110 250 This study (Fe 70 Ni 30 ) 89 Zr 7 B 4 342 0.7 330 19 Fe 79 Gd 1 Cr 8 B 12 355 1.42 153 627 459 20 Fe 70 Ni 30 363 0.65 158 204 470 This study 17A918-3 Ucar et al. J. Appl. Phys. 113, 17A918 (2013) 19 May 2025 14:17:22