IEEE TRANSACTIONS ON MAGNETICS, VOL. 50, NO. 11, NOVEMBER 2014
2504304
A Comparative Study of Magnetocaloric Properties Between Ni-rich and Mn-rich Ni–Mn–Sn Alloys Arup Ghosh and Kalyan Mandal Magnetism Laboratory, Department of Condensed Matter Physics and Material Sciences, S. N. Bose National Centre for Basic Sciences, Kolkata 700098, India Ni-rich and Mn-rich Ni–Mn–Sn Heusler alloys have been prepared and studied to find out their magnetocaloric properties. The martensitic transition temperature decreases for Ni-rich and increases for Mn-rich alloys with the increase of their respective rich element. The value of isothermal magnetic entropy change is almost double for Mn-rich alloys as compared with the Ni-rich alloys. The refrigerant capacity is also higher for the same alloy series. The origin of such enhancement in magnetocaloric potentials in Mn-rich alloys has been discussed in the context of ferro/antiferro correlations present between the intersite Mn atoms of these alloy families. Index Terms— Heusler alloys, magnetocaloric effect (MCE), magnetostructural transition, refrigerant capacity (RC).
I. I NTRODUCTION
O
FF-STOICHIOMETRIC Ni–Mn–Sn Heusler alloys have become a very effective magnetocaloric material during the last few years as they show large magnetic entropy change (SM ) in the vicinity of their first-order magnetostructural transition (FOMST) [1]–[10]. Ni-rich Ni50 Mn50−x Snx alloys have shown large SM , and thus they have been studied extensively to date [2]–[7]. However, some very recent studies suggest that the SM can be enhanced to a significant extent by taking the Mn near about 50 at% [3], [8]–[10]. Therefore, Mn-rich Ni–Mn–Sn alloys (Mn ∼ 50 at%) have attracted immense attention from the active research community during the last two years. The magnetic response of Mn-rich Ni–Mn–Sn alloys (Mn ∼ 50 at%) is better than that of Ni-rich alloys [8]–[10]. Stoichiometric Ni–Mn–Sn (X2 YZ) Heusler alloys have cubic (L21 ) structure where (0, 0, 0) and (1/2, 1/2, 1/2) sites are occupied by X element’s atoms (either Ni or Mn). (1/4, 1/4, 1/4) and (3/4, 3/4, 3/4) sites are occupied by Y (either Ni or Mn) and Z (Sn) elements’ atoms, respectively [11], [12]. It is well known that Mn atoms carry most of the magnetic moments of these alloys [13], [14]. Therefore, the ferromagnetic (FM)/antiferromagnetic (AFM) interactions between the Mn atoms sitting at the two different sites are expected to be larger for Mn-rich alloys as compared with the Ni-rich one. The off-stoichiometric compositions of these alloys undergo an FOMST from martensite to austenite phase on heating, and this in turn triggers a peculiar change of magnetic state from weakly FM/paramagnetic (PM) like to strongly FM. Therefore, a large discontinuity in magnetization can be achieved across the martensitic transition and this may give rise to a large SM also [10], [15]. The thermal hysteresis (Thys) at FOMST can also be reduced by making these alloys Mn rich [8]–[10]. Manuscript received March 5, 2014; revised April 10, 2014; accepted April 28, 2014. Date of current version November 18, 2014. Corresponding author: A. Ghosh (e-mail:
[email protected]). Color versions of one or more of the figures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identifier 10.1109/TMAG.2014.2321260
In this paper, we have prepared Ni-rich (Ni ∼ 50 at%) and Mn-rich (Mn ∼ 50 at%) Ni–Mn–Sn Heusler alloys and studied their magnetic and magnetocaloric properties. The Mn-rich alloys have shown significantly large values of magnetocaloric potentials [SM and refrigerant capacity (RC)], which predicts their usefulness for better magnetic cooling. II. E XPERIMENT Ni-rich Ni48.5 Mn37Sn14.5 (Ni-48.5), Ni50 Mn36.5 Sn13.5 (Ni50) and Mn-rich Ni41 Mn49.5Sn9.5 (Mn-49.5), Ni41 Mn50.5 Sn8.5 (Mn-50.5) alloys were prepared by arc melting technique under a 4 N purity argon atmosphere. The ingots were turned and remelted for several times to ensure homogeneity. Each ingot was wrapped with tantalum foil and sealed in highly evacuated quartz ampoules for annealing. After 24 h of heat treatment at 1173 K, the ampoules were quenched in ice water. The compositions of the prepared samples were confirmed by energy-dispersive spectroscopy. X-ray diffraction patterns have been carried out in Rigaku miniflex II using Cu–Kα radiation to detect the crystallographic parent phase. All the magnetic measurements were performed using a vibrating sample magnetometer (Lake Shore, model-7144) up to 16 kOe magnetic fields with field and temperature sweeping rates of 200 Oe/s and 2 K/min, respectively. III. R ESULTS AND D ISCUSSION Fig. 1(a) and (b) shows the room temperature (at 300 K) XRD patterns for Ni-rich and Mn-rich alloys. The coexistence of both the austenite and martensite peaks in Ni-48.5 predict that the sample is in mixed phase at 300 K, and its structural transition may reside just below the room temperature. In the case of other samples, the cubic austenite (L21 ) phase is predominant at the aforementioned temperature. The superlattice diffraction peak A(111) for Ni-50 and Mn-49.5 ensures the higher level of atomic order in both the alloys [3]. The zero field cooled (ZFC) and field cooled (FC) temperature dependence of magnetization (M–T curves) of these alloys are plotted in Fig. 2(a)–(d). By following the FC curves from 350 K, a sharp rise in magnetization can
0018-9464 © 2014 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
2504304
IEEE TRANSACTIONS ON MAGNETICS, VOL. 50, NO. 11, NOVEMBER 2014
TABLE I S AMPLES ’ N AME AND T HEIR C HARACTERISTIC T RANSITION T EMPERATURES . T HE M AXIMUM P OSSIBLE E RROR I S ±3 K
Fig. 1. Room temperature (at 300 K) XRD patterns for (a) Ni-rich and (b) Mn-rich Ni–Mn–Sn alloys (A: austenite peak and M: martensite peak).
Fig. 3. ZFC M–T curves for Ni-rich (a) Ni-48.5, (b) Ni-50 and Mn-rich, (c) Mn-49.5, and (d) Mn-50.5 alloys in the presence of 5 kOe and 15 kOe fields.
Fig. 2. Temperature dependence of magnetization (M–T curves) for Ni-rich (a) Ni-48.5, (b) Ni-50 and Mn-rich, (c) Mn-49.5, and (d) Mn-50.5 alloys in the presence of 0.1 kOe field.
be observed for Ni-rich alloys near 320 K and for Mn-rich alloys near 280 K, which is due to a para to ferro transition at the Curie temperature of the austenite phase (TCA ). At the temperatures below the TCA of respective samples, another phase transition with thermal hysteresis (between FC and ZFC curves) is present, which is known as the FOMST of these materials. Here, the structure of the samples changes from FM austenite to magnetically complex (weakly FM or PM like) martensite phase on cooling [1]–[10]. The Curie temperature at the martensite phase (TCM ) can be identified from the M–T curves, where the magnetization of the samples starts to
increase noticeably as the temperature decreases further below the FOMST. A noticeable amount of bifurcation between the ZFC and FC curve can be identified below the TCM of all the samples, which predicts the existence of magnetically inhomogeneous phase consisting of the mixture of FM, AFM, and re-entrant spin glass phases, as reported in [4] and [10]. The characteristic transition temperatures for magnetic and structural transition of these alloys are given in Table I and indicated in an M–T curve also. It is interesting to note that the FOMST temperature for Ni-rich alloys decreases with increasing the Ni content, whereas the same increases for Mn-rich alloys as the Mn content increases. The thermal hysteresis (Thys) at the FOMST has the opposite dependence on Ni and Mn contents, respectively, for Ni- and Mn-rich samples. Fig. 3(a)–(d) shows the ZFC M–T curves of these alloys in the vicinity of martensitic transition in the presence of 5 and 15 kOe fields. The field-induced shift of magnetostructural transition is observed in all the samples. The values of field sensitivity of martensitic critical temperature (|T /H |), width of the martensitic transition (Ttran = (A f − AS )), saturation magnetization near austenite start A AS (Msat ), finish (Msatf ), and the corresponding jump in the magnetization due to martensite–austenite phase transition A AS ) are measured from these M–T curves (Msat = Msatf − Msat and given in Table II. |T /H | and Ttran increase with increasing Ni content in Ni-rich alloys and decrease in Mn-rich alloys as their rich element increases. The Msat decreases for
GHOSH AND MANDAL: COMPARATIVE STUDY OF MAGNETOCALORIC PROPERTIES
2504304
TABLE II VALUES OF S OME I MPORTANT PARAMETERS U SEFUL FOR M AGNETOCALORIC C HARACTERIZATION
Fig. 5. Temperature dependence of magnetic entropy change (SM ) of these alloys near their structural phase transition due to 16 kOe field change.
Fig. 4. Magnetization versus magnetic field (M–H curves) for Ni-rich (a) Ni-48.5, (b) Ni-50 and Mn-rich, (c) Mn-49.5, and (d) Mn-50.5 alloys in the vicinity of their martensitic transition.
Ni-rich samples and increases for Mn-rich samples with the increase of their respective rich elements. The importance of M/T in showing large magnetocaloric effect (MCE) is discussed in detail in the following. Fig. 4(a)–(d) shows the isothermal magnetization versus magnetic field curves (M–H curves) of these alloys in the vicinity of their structural phase transition. The M–H curves of Mn-rich samples near their austenite start temperature (AS ) are PM like [in Fig. 4(c) and (d), M–H curves at 227 and 230 K for Mn-49.5 and at 260 and 263 K for Mn-50.5]. It is weakly FM like for the Ni-rich alloys [in Fig. 4(a) and (b), M–H curves at 291 K for Ni-48.5 and at 279 K for Ni-50]. This can be explained by considering the FM/AFM interactions present in these alloy systems between the intersite Mn atoms in the austenite/martensite phase [15]. The FM correlation is dominated at the austenite phase of these samples. As the structure transforms from austenite to martensite on cooling, the separation between the Mn atoms sitting at the two different sites decreases. The Mn–Mn intersite interaction becomes strongly AFM when their separation reduces beyond a critical value, and as a result, the net magnetization of the system decreases enormously during the austenite to martensite phase transition. The jump in magnetization near the FOMST in the M–T curves [as shown in Figs. 2(a)–(d) and 3(a)–(d)] supports this argument very well. The Mn-rich alloys have the maximum number of intersite
Mn–Mn pair, and this is why their AFM correlation in the martensite phase is much stronger as compared with that of Ni-rich alloys and it makes Mn-rich alloys PM like within the temperatures TCM and AS H ∂M SM (T, H ) = μ0 dH (1) ∂T H 0 T2 |SM (T, H )|d T. (2) RC = T1
The SM of these alloys has been estimated from the M–H data using the Maxwell’s thermodynamic relation (1) and plotted in Fig. 5 as a function of temperature [2]. Although (1) may give overestimated values of SM for these materials undergoing a first-order phase transition, it is quite appreciable for comparing these results with the literature values. During the M–H measurements, after the completion of each curve, the samples were cooled below their respective M f and heated back to the next targeted temperature to avoid any possibility of occurrence of the ghost peaks in the SM –T curves. One can notice that the value of SM (due to 16 kOe field change) is double for Mn-rich samples (∼12 J/kg K at 270 K for Mn-50.5) as compared with the Ni-rich alloys (∼6 J/kg K at 300 K for Ni-48.5). To explain these results in a rigorous way, d M/d T has been calculated from Fig. 3 data and plotted as a function of T in Fig. 6 for all the samples in the presence of 15 kOe fields. One can notice that the peak value of d M/d T is the maximum for Mn-rich alloys at their FOMST, and it follows the similar trajectory as the SM –T curves does. As discussed earlier in this paper, the large values of Msat and smaller values of Ttran in Mn-rich samples enhance M/T (data are given in Table II). This in turn boosts their d M/d T to a significant extent (in other words, it improves the values of SM to an enormous magnitude). Our results reveal that the value of M/T becomes larger as the strength of the intersite Mn–Mn AFM correlations increases, and it is found to be the maximum for Mn-rich Mn-50.5 alloy. The same
2504304
IEEE TRANSACTIONS ON MAGNETICS, VOL. 50, NO. 11, NOVEMBER 2014
martensitic transition on the concentration of the rich element has opposite nature between these two alloy species (Ni-rich and Mn-rich). The Mn-rich Ni–Mn–Sn alloys show very large magnetic entropy change and RC capacity as compared with Ni-rich similar alloys. This comparative study concludes that the Mn-rich Ni–Mn–Sn alloys are better magnetic refrigerant than the Ni-rich one. ACKNOWLEDGMENT This work was supported by the Defence Research and Development Organisation, Government of India, under Project ERIPR/ER/0902182/M/01/1296. R EFERENCES
Fig. 6. Temperature dependence of d M/d T of these alloys near their structural phase transition in the presence of 15 kOe fields (as calculated from Fig. 3 data). TABLE III C OMPARISON OF SM VALUES B ETWEEN O UR S AMPLES AND THE O THER R EPORTED S IMILAR Ni–Mn–Sn H EUSLER A LLOYS
sample has shown maximum SM also. Therefore, a valid relation may present between the SM of these alloys and the aforementioned interaction. SM decreases for Ni-rich samples and increases for Mn-rich alloys with increasing their respective rich elements. We have compared the values of SM obtained from our samples with the literature values for similar alloys in Table III, and found that the SM in Mn-rich samples are quite larger as compared with the other Ni–Mn–Sn Heusler alloys reported to date [16]. The RC of these alloys has been calculated from the SM versus T curves [8] using (2). In this equation, T1 and T2 are the two temperatures at the full with at half maxima (FWHM) of the SM –T peak. RC estimates the amount of heat that can be extracted from a colder body (at T1 ) and transferred it to a hotter one (at T2 ) in one complete thermodynamic cycle. RC ∼53 J/kg is obtained for Mn-rich Mn-50.5 sample, and it is larger than the other samples (∼40 J/kg for Ni-rich Ni-50 alloy). It is evident from here that the Mn-rich Ni–Mn–Sn heusler alloys can provide a better magnetic cooling as compared with the Ni-rich one. IV. C ONCLUSION In summary, we have systematically investigated Ni-rich (Ni ∼ 50 at%) and Mn-rich (Mn ∼ 50 at%) off-stoichiometric Ni–Mn–Sn heusler alloys for their MCE. The dependence of
[1] A. Planes, L. Manosa, and M. Acet, “Magnetocaloric effect and its relation to shape-memory properties in ferromagnetic Heusler alloys,” J. Phys., Condensed Matter, vol. 21, no. 23, p. 233201, 2009. [2] T. Krenke et al., “Inverse magnetocaloric effect in ferromagnetic Ni-Mn-Sn alloys,” Nature Mater., vol. 4, pp. 450–454, May 2005. [3] S. E. Muthu, N. V. R. Rao, M. M. Raja, D. M. R. Kumar, D. M. Radheep, and S. Arumugam, “Influence of Ni/Mn concentration on the structural, magnetic and magnetocaloric properties in Ni50−x Mn37+x Sn13 Heusler alloys,” J. Phys. D, Appl. Phys., vol. 43, no. 42, p. 425002, 2010. [4] A. Ghosh and K. Mandal, “Large inverse magnetocaloric effect in Ni48.5−x Cox Mn37 Sn14.5 (x = 0, 1 and 2) with negligible hysteresis,” J. Alloys Compounds, vol. 579, pp. 295–299, Dec. 2013. [5] C. Jinga et al., “Martensitic transition and inverse magnetocaloric effect in Co doping Ni-Mn-Sn Heulser alloy,” Eur. Phys. J. B, Condensed Matter Complex Syst., vol. 67, no. 2, pp. 193–196, 2009. [6] T. Krenke, E. Duman, M. Acet, X. Moya, L. Manosa, and A. Planes, “Effect of Co and Fe on the inverse magnetocaloric properties of Ni-Mn-Sn,” J. Appl. Phys., vol. 102, no. 3, p. 033903, 2007. [7] P. J. Shamberger and F. S. Ohuchi, “Hysteresis of the martensitic phase transition in magnetocaloric-effect Ni-Mn-Sn alloys,” Phys. Rev. B, vol. 79, no. 14, p. 144407, 2009. [8] L. Ma et al., “Martensitic and magnetic transformation in Mn50 Ni50−x Snx ferromagnetic shape memory alloys,” J. Appl. Phys., vol. 112, no. 8, pp. 083902-1–083902-4, 2012. [9] Q. Tao et al., “Phase stability and magnetic-field-induced martensitic transformation in Mn-rich NiMnSn alloys,” AIP Adv., vol. 2, no. 4, p. 042181, 2012. [10] A. Ghosh and K. Mandal, “Large magnetic entropy change and magnetoresistance associated with a martensitic transition of Mn-rich Mn50.5−x Ni41 Sn8.5+x alloys,” J. Phys. D, Appl. Phys., vol. 46, no. 43, p. 435001, 2013. [11] A. Ayuela, J. Enkovaara, K. Ullakko, and R. M. Nieminen, “Structural properties of magnetic Heusler alloys,” J. Phys., Condensed Matter, vol. 11, no. 8, pp. 2017–2026, 1999. [12] R. B. Helmholdt and K. H. J. Buschow, “Crystallographic and magnetic structure of Ni2 MnSn and NiMn2 Sn,” J. Less-Common Metals, vol. 128, pp. 167–171, Feb. 1987. [13] P. J. Brown, A. Y. Bargawi, J. Crangle, K. U. Neumann, and K. R. A. Ziebeck, “Direct observation of a band Jahn–Teller effect in the martensitic phase transition of Ni2 MnGa,” J. Phys., Condensed Matter, vol. 11, no. 24, pp. 4715–4722, 1999. [14] T. Krenke, M. Acet, E. F. Wassermann, X. Moya, L. Mañosa, and A. Planes, “Ferromagnetism in the austenitic and martensitic states of Ni-Mn-In alloys,” Phys. Rev. B, vol. 73, no. 17, p. 174413, 2006. [15] S. Aksoy, M. Acet, P. P. Deen, L. Manosa, and A. Planes, “Magnetic correlations in martensitic Ni-Mn-based Heusler shape-memory alloys: Neutron polarization analysis,” Phys. Rev. B, vol. 79, no. 21, p. 212401 2009. [16] A. G. Varzaneh, P. Kameli, F. Karimzadeh, B. Aslibeiki, G. Varvaro, and H. Salamati, “Magnetocaloric effect in Ni47 Mn40 Sn13 alloy prepared by mechanical alloying,” J. Alloys Compounds, vol. 598, pp. 6–10, Jun. 2014. [17] Z. Han et al., “Phase diagram and magnetocaloric effect in Mn2 Ni1.64−x Cox Sn0.36 alloys,” Scripta Mater., vol. 66, no. 2, pp. 121–124, 2012. [18] B. Gao, F. X. Hu, J. Shen, J. Wang, J. R. Sun, and B. G. Shen, “Fieldinduced structural transition and the related magnetic entropy change in Ni43 Mn43 Co3 Sn11 alloy,” J. Magn. Magn. Mater., vol. 321, no. 17, pp. 2571–2574, 2009.
