Magnetocaloric behavior of Mn rich Ni46Cu2Mn43In11 alloy

In this work, we studied the magnetic entropy change (ΔSM) across the martensite transformation (MT) in Mn-rich Ni46Cu2Mn43ln11 alloy. This compound undergoes a MT and a magnetic phase transition around the temperatures (TM=) 272 K and (TCA=) 325 K, respectively. A large field induced shift (=0.28 K/kOe) of the MT temperatures is observed. An application of magnetic field (H =) of 50 kOe causes a large ΔSM of 20 J/kg-K and -4.4 J/kg-K around TM and TCA, respectively. We also found that the change in magnetic field induced isothermal ΔSM(H)T is mainly depends on the induced austenite phase fraction by the applied magnetic field at that temperature. Possible reasons for the observed behaviours are comprehensively discussed.


Introduction
In recent times, non-stoichiometric Ni-Mn-Z (Z=Sn,In) intermetallic compounds has attracted much attention due to their potential applications in the field of magnetic refrigeration, spintronics and magnetic read head etc [1][2][3][4][5]. It is believed that such novelties are primarily due to the presence of a strong coupling between the structural, the magnetic and the electronic degrees of freedom [2]. It is also well-known that these types of systems are characterized by first-order martensitic transition (MT) at TM as well as a second-order paramagnetic (PM) to ferromagnetic (FM) transition at the Curie point (TCA). The TCA in these systems must be higher than their TM (TCA > TM) to realize the magneto-functionality around the TM, i.e., the material should be in a magnetically well ordered state when MT takes place during cooling [6][7]. The MT in these materials also alters the magnetic state in which material may lose its ferromagnetism or attain a state having antiferromagnetic (AFM)-like correlation below the MT. Since the MT occurs between the magnetically order (austenite) to a magnetically disorder (martensite) state consequently a large change in the magnetization (ΔM) and hence a large magnetic entropy change (ΔSM) is expected to achieve across the MT [8]. It is noteworthy that the high ΔM is one of the necessary conditions for any material to be a good magnetocaloric material (MCM). The presence of large ΔSM produces the magnetocaloric effect (MCE) in these materials has great potential applications in the cooling technology without any harmful impact on the environment unlike the vapour compression based refrigerator. It is also well known that the MCE is a thermal response of a magnetic material against the variation of an external magnetic field and usually expressed in terms of ΔSM in an isothermal process, or adiabatic temperature change (ΔTad). As we already mentioned earlier that a large ΔSM can be achieved across the MT therefore the researchers paid significant efforts in designing Ni-Mn based MCM with a strong magneto-structural coupling at MT. The studies also focused on the improvement of another vital MCE parameter, called the refrigeration capacity (RC) or the relative cooling power (RCP), which quantifies the amount of 1234567890''"" heat transferred from cold to hot sink through the ideal refrigeration cycle [9]. In order to achieve a large RCP, it is important to widen the MT temperature range as this determines the working temperature range of a MCM. Upto now, the expensive rare-earth metal Gd, which experiences a second-order magnetic transition near room temperature, is regarded as a benchmark refrigerant material in magnetic refrigerator prototype [10,11]. Since these rare earth systems are costly, researchers are considering Ni-Mn based intermetallic compounds as an alternative to its rare earth counterpart. But the problems in Ni-Mn based materials are, the TM in these systems is much lower than the room temperature (RT) for most of the Ni-Mn based martensite compounds and also the narrow MT width makes them low cooling efficient materials. In search for a better MCM with improved magneto-functionalities for more useful applications within the family of Ni-Mn we have synthesized the Ni44Cu2Mn43In11 compound and studied its magnetic and magnetocaloric properties.

Experimental details
The ascast sample was prepared by using tri-arc melting furnace from pure metals (99.995%, Alfa Aesar) in argon atmosphere. The sample was then annealed at 1173 K for 24 hrs in vacuum followed by quenching in normal water for homogenization. The crystal structure was determined using room temperature powder X-Ray diffraction (XRD) (Bruker D8 advance diffractometer). The nominal compositions of the annealed alloy was confirmed by energy dispersive X-Ray analysis (EDX) attached with a field emission scanning electron microscope (FESEM) within the error of ±0.1 at.%. Calorimetric measurements were done using Differential Sacnning Calorimetry (DSC). The magnetic measurements were carried out using Physical Property Measurement System (PPMS, Quantum Design) in the temperature range T=5-380 K and the field range H=0-90 kOe.

Result and discussions
The temperature dependent dc susceptibility χ(T) measurements were conducted using the zero field cooled warming (ZFCW) and the field cooled cooling (FCC) protocols and the results are shown in Fig.1. In the measurement of ZFCW χ(T), we first cooled the sample under zero magnetic field down to T= 5 K and then applying a magnetic field (H =) 100 Oe we recorded the χ (T) during warming the sample upto 395 K whereas the FCC χ(T) was measured in subsequent cooling of the sample. The χ(T) data show multiple features depending on the change in magnetic phases due to the thermal variations. At T= 330 K, the sample shows a sharp upturn with the decreasing temperature, which corresponds to the paramagnetic (PM) to ferromagnetic (FM) transition within the austenite phase, called Curie temperature of austenite phase (TCA). On further cooling, χ (T) starts to decrease and a sharp fall is observed around Tms. The fall in χ (T) is possibly due to the loss of ferromagnetism and the sample presumably becomes PM or AFM below the MT [12,13]. On the other hand χ(T) starts to increase around Tas during warming the material and a thermal hysteresis between the χZFCW (T) and χFCC (T) curve appears in the temperature range of 245-310 K. Such hysteresis signifies the first order nature of MT. All the characteristic transition temperatures, i.e., the austenite start and finish temperatures (Tas and Taf) and the martensite start and finish temperatures (Tms and Tmf) are labelled in Fig.1. Now below the MT, the compound continues to be in a low magnetic phase and eventually shows a spin glass type magnetic ground state below T=65 K. To validate the transformations characteristics we have carried out differential scanning calorimetry (DSC) in the absence of any magnetic field. The thermodynamic heat profiles pertaining to endothermic and exothermic peak are obtained during the heating and cooling cycles, respectively (inset of Fig.1). The exothermic and endothermic transformation peaks represent forward (austenite phase to martensite phase trasition) and reverse (martensite phase to austenite phase transition) MT, respectively. The values of the characteristic MT temperatures obtained from the magnetic measurement are in good agreement with our calorimetric measurements. Small discrepancy between the results of magnetic and calorimentric measurement can be ascribed to the influence of some factors such as the different thermal contacts of the samples with the thermocouples in the DSC and PPMS instruments and the two tangent procedures used for the determination of the characteristic temperatures etc. After the baseline correction of our calorimetric data we have quantified the transition entropy change using the following equation: where T1 and T2 denote the temperatures well below and above the phase transition, respectively and ( ௗொ ௗ் ) is the heat capacity. Using this equation we have numerically calculated the transition entropy change (ΔStr) with the value of 17.7 J/kg-K. It is also important to mention that in the measurement of χ(T), χ(T) starts to drop below the TCA and the thermal hysteresis even exists in temperature range Tms/Taf to TCA but we did not observed any trace of this hysteric region in our calorimetric and resistivity measurements. We have further measured the thermo-magnetization M(T) cuves using field cooled cooling (FCC) and field cooled warming (FCW) protocols in the temperature range T=150-395 K under different applied fields (see the Fig.2). It can be seen from the Fig.2 that the magnetization difference (ΔM= M|Tmf-M|Tms) between the austenite and martensite phases keeps on increasing with the increase of the magnetic field. Also the M(T) below Tmf remains nearly temperature independent for all the applied fields which signifies the low magnetic behaviour of this phase. It is also evident from the Fig.2 that the characteristic martensite transformation temperatures shift towards the lower temperatures side with the application of magnetic field. The change in Tms with magnetic field H, i.e., ௗ்௦ ௗு is found to be 0.28 K/kOe. This observed value of ௗ்௦ ௗு has good agreement with the one we obtained using the Clausius-Clapeyaron (C.C) equation 2. If we consider the FCC M(T) curve for H=70 kOe then the ΔM comes out to be 55 emu/g and from our DSC measurement we found the ΔStr is 17.7 J/kg-K. (2)

1234567890''""
The Therefore, if we use these values in C.C equation, ௗ்௦ ௗு comes out to be 0.31 K/kOe which is very close to that obtained from direct determination of Tms shift. The shift in Tms is easy to understand by considering the change in the free energy during the MT. We know that the reverse martensitic transformation (RMT) is motivated by the chemical free energy difference between the martensite and austenite phases and the elastic energy stored in the martensite phase. Therefore, one can express the Gibbs free energy difference between the martensite and austenite phases during the RMT (martensite to austenite transformation) as where ‫ܩ߂‬ ெ→ represent the total Gibbs free energy dieerence between the two phases and it should be < 0 to trigger the RMT [14,15]. While ‫ܩ߂‬ ெ→ represent the chemical free energy difference, ‫ܩ߂‬ ெ→ is the elastic energy stored in the martensite phase and ‫ܧ߂‬ ெ→ is the dissipation energy generated by defects in the system. Now when we apply a magnetic field, an additional and magnetic contribution to the free energy comes into play a crucial role in the total Gibbs free energy equation. Then the modified Gibbs free energy difference can be rewritten as follows [16] ‫ܩ߂‬ where ‫ܩ߂‬ ெ→ is the Zeeman energy difference between the martensite and austenite phases and can written as Here ‫ܯ‬ ௌ and ‫ܯ‬ ௌ ெ are the saturation magnetizations of the austenite and martensite phases, respectively. In our studied compound the martensite phase is weak-magnetic, and the austenite is a strong FM ‫ܯ(‬ ௌ ‫ܯ>‬ ௌ ெ ), which results in ‫ܩ߂‬ ெ→ < 0. Therefore, the magnetic field favours the RMT, leading to a decrease in transformation temperatures. Using the same ideas one can also explain the increase of the thermal hysteresis width (ΔT) across the MT with measuring H (see the inset of Fig.2). For a given M(T) curve (say H=70 kOe), the ΔM is little higher for FMT (austenite to martensite) than that for RMT. Therefore, the shifts of Tms and Tmf to the lower temperature side are larger than the shifts of Tas and Taf which eventually leads to the increase in ΔT with magnetic fields. Similar hysteresis loops were obtained at 50 K and 100 K.
To find out the effect of magnetic fields on the ΔSM we have conducted the isothermal M-H measurements in the temperature range of T=220-275 K. up to a magnetic field of H= 90 kOe. where Tav and ΔT represents the average temperature and temperature difference between two ZFC M-H curves, respectively. ΔH is the changing step in the applied field and ΔMk is the difference in magnetization at the field k×H between two isothermal curves, where k is the step number of applied field, k=1 and k = n are the step correspond to H=0 and the last step of maximum field, respectively. ΔSM versus H at several temperatures is shown in Fig.3. From the Fig.3 we can see that when the system is far below of its MT region i.e. below 240 K, only a small part of martensite phase is transformed into the austenite by the available magnetic field and thus the induced ΔSM remains low and unsaturated. In this region even a magnetic field as large as 90 kOe is not enough to produce a complete field induced martensite to austenite transformation. Now as we move towards the MT region, more fractions of the martensite phase is now start to convert into the austenite phase by the available field and ΔSM(H)T starts to increase up to T=250 K. Whereas for the temperatures just below of martensite temperature TM (= ଵ ଶ (Tms + Taf )), a field of a lower values than H=90 kOe is sufficient to produce saturated ΔSM. With the increase of temperatures, the austenite phase fraction increases and ΔSM starts to the decrease again due to the decrease of the induced austenite phase fraction. So, one can conclude from Fig.3 that the induced ΔSM is proportional to the induced austenite phase fraction. Apart from the magnetic entropy change ΔSM, the relative cooling power (RCP) of a magnetic material is another important parameter as it measure the efficiency of the material to be used as a refrigerant in magnetic refrigerator. The RCP is defined as: RCP=ΔSM(max)×δTFWHM, where δTFWHM is the full-width-at-half-maximum of the ΔSM(T) curve. There are several approaches to determine RCP, though all of them target the range of working temperature [17]. Therefore there is further need to correlate MT with ΔSM as well as with the working temperature range. Amorphous materials with second order transition favour large RCP and low magnetic entropy change opposite to the crystalline materials with first order transition [18]. The key concern is to develop magnetic refrigerant with high magnetic entropy change without compromising the working temperature range. Since this system shows two different types of transition i.e., MT (at   Fig.4. A strong field dependence of both the ΔSM and RCP across the MT is observed. We have obtained a RCP value of 215 J/kg across the MT for the field change of 50 kOe which is comparatively very high considering Ni-Mn based systems. Moreover, in contrary to what was expected in this kind of material, the temperature of the negative peak shifts towards higher temperatures than TCA. The ΔSM has the characteristic shape of a first and second-order phase transitions both expanded over a large temperature range.

Conclusions:
In summary, the magneto-structural martensite transitions and magnetic entropy change of Mnrich Ni44Cu2Mn43In11 were studied. A strong field dependence of ΔSM and RCP across the MT is observed. From the field induced behaviour of ΔSM it is evident that the magnetic field induced austenites phase contributing more in the total ΔSM. The observed large RCP in this system can be useful for practical applications as magnetic refrigerant material.