Tunable Martensitic Transformation and Magnetic Properties of Sm-Doped NiMnSn Ferromagnetic Shape Memory Alloys

Abstract

NiMnSn ferromagnetic shape memory alloys exhibit martensitic transformation at low temperatures, restricting their applications. Therefore, this is a key factor in improving the martensitic transformation temperature, which is effectively carried out by proper element doping. In this research, we investigated the martensitic transformation and magnetic properties of Ni₄₃Mn_46-x Sm_xSn₁₁ (x = 0, 1, 2, 3) alloys on the basis of structural and magnetic measurements. X-ray diffraction showed that the crystal structure transforms from the cubic L2₁ to the orthorhombic martensite and gamma (γ) phases. The reverse martensitic and martensitic transformations were indicated by exothermic and endothermic peaks in differential scanning calorimetry. The martensitic transformation temperature increased considerably with Sm doping and exceeded room temperature for Sm = 3 at. %. The Ni₄₃Mn₄₅SmSn₁₁ alloy exhibited magnetostructural transformation, leading to a large magnetocaloric effect near room temperature. The existence of thermal hysteresis and the metamagnetic behavior of Ni₄₃Mn₄₅SmSn₁₁ confirm the first-order magnetostructural transition. The magnetic entropy change reached 20 J·kg⁻¹·K⁻¹ at 266 K, and the refrigeration capacity reached ~162 J·Kg⁻¹, for Ni₄₃Mn₄₅SmSn₁₁ under a magnetic field variation of 0–5 T.

1. Introduction

The materials showing martensitic transformation (MT) exhibit various multifunctional phenomena, such as the magnetocaloric effect (MCE) [1,2], exchange bias (EB) [3,4,5], magnetothermal conductivity (MC) [6,7] and magnetoresistance (MR) [8,9,10]. This coupled transition is obtained around the MT temperature (T_t). These multifunctional properties show application in vast areas such as magnetic refrigeration, energy harvesting, magneto-mechanical devices and sensors [11,12,13,14]. The MCE is a magneto-thermodynamic phenomenon in which the temperature is changed in the material when exposed to an external non-constant magnetic field [15]. MCE-based magnetic refrigeration has many advantages over conventional gas cooling technology, such as being environmentally friendly, having high refrigerant efficiency and low cast measures, occupying less space, possessing low mechanical vibration and being harmless [16,17,18,19]. MCE materials are utilized in magnetic refrigeration technology. The materials showing first-order magnetic transitions are widely used for magnetic cooling and investigated due to their large MCE, e.g., LeFeSi [20,21,22], MnFePAs [23], GdSiGe [24], MMnX (M = Ni or Co, X = Ge or Si) [25] and NiMnZ (Z = Sn, In, Sb) ferromagnetic shape memory alloys (FMSMAs) [26].

FMSMAs exhibit first-order magnetostructural transformation from high-temperature austenite to low-temperature martensite upon cooling, with considerable thermal hysteresis and other effects such as lattice distortion and abrupt changes in magnetization. The magnetic field induces a first-order magnetostructural transition in these alloys, generating the MCE [27]. It is promising to tune the T_t in a wide temperature range, endowing FMSMAs with the ability to exhibit large multifunctional effects under a low applied magnetic field for magnetic applications. It is reported that the T_t can be tuned by changing the valence electron concentration (e/a) [28], grain size [29], external parameters such as the magnetic field and applied pressure [30], volume of the unit cell [14] and heat treatment [31], while adjustable elemental composition and preparation conditions are very important to achieve the MCE for the desired applications [25,26,27].

NiMnSn-based alloys undergo a reverse martensitic transformation due to magnetostructural coupling and exhibit a variety of multifunctional properties, which indicates a potential for many applications. The utilization of MCE materials strongly depends upon the mechanical properties and magnetostructural coupling. Unfortunately, the highly brittle nature and poor mechanical properties of NiMnSn alloys are major drawbacks. Additionally, NiMnSn alloys show a low working temperature due to coupling between MT and ferromagnetic transition [32].

Only a few studies have reported on the enhancement of the working temperature of these alloys, which is the main point of application. Moreover, it is well known that a large MCE has been reported in rare-earth metal-based materials, i.e., Gd₅Si₂Ge₂ [24]. Meanwhile, it has also been reported that the addition of rare-earth elements could enhance the T_t in these types of alloys. Therefore, it can be anticipated that the addition of proper rare-earth elements in NiMnSn FMSMAs can tune the T_t to a higher temperature, resulting in the MCE near room temperature (RT). There is still a lack of information and knowledge about the rare-earth element Sm after it has been doped. In this research, we doped Sm in Ni₄₃Mn_46-x Sm_xSn₁₁ (x = 0, 1, 2, 3) alloys which increases the T_t considerably, and a large MCE is obtained near RT for the Ni₄₃Mn₄₅SmSn₁₁ alloy sample.

2. Materials and Methods

Polycrystalline Ni₄₃Mn_46-x Sm_xSn₁₁ (x = 0, 1, 2, 3) alloy samples were prepared by arc melting 99.99% pure Ni, Mn, Sn and Sm using a water-cooled copper crucible in an argon atmosphere. The melting process was repeated four times to ensure the purity and uniformity of the alloys. Homogeneity is an important factor to enhance the properties of the materials [31]. Therefore, ingots were annealed at 1123 K for 2 days followed by quenching in ice water. The elemental composition of the samples was determined by X-ray energy-dispersive spectroscopy (EDS, Thermo system 7, Malvern UK), and powder X-ray diffraction (XRD, Bruker, D8 Advance, Madison, WI, US) was used to identify the crystal structure at room temperature. The structural transitions were investigated using differential scanning calorimetry (Mettler Toledo, DSC 3, Nanterre, France), while the magnetic properties were examined by a magnetic property measurement system (MPMS-7, Quantum Design, San Diego, CA, USA) and physical property measurement system (PPMS, Quantum Design, Dynacool, San Diego, CA, USA). The so-called loop method was used to measure isothermal magnetization (MH) curves to avoid irreversibility [33].

3. Results and Discussion

Energy-dispersive X-ray spectroscopy (EDS) was used to check the appropriate amount and homogeneity of the constituent elements in the prepared alloys. The EDS results of the Ni₄₃Mn_46–xSm_xSn₁₁ (x = 0, 1, 2, 3) alloys are shown in

Table 1. Atomic weight percentage and electronic concentration of Ni₄₃Mn_46-xSm_xSn₁₁ (x = 0, 1, 2 and 3) alloys.

Sample	Atomic Weight Percentage (%)				Electronic Concentration (e/a)
	Ni	Mn	Sn	Sm
x = 0	42.96	46.09	11.05	0	79.643
x = 1	42.92	45.11	10.90	1.07	79.713
x = 2	42.90	43.96	11.03	2.11	79.772
x = 3	43.05	43.02	10.90	3.03	79.794

, confirming the presence of all precursor elements. The obtained compositions of the alloys were found to be very close to the nominal composition. Here, a small deviation in the composition of the elements is very common due to various factors such as weighing the precursor elements, melting and annealing. The electronic concentration (e/a) was measured using Equation (1) [34].

e/a = (7 × at. % Mn + 10 × at. % Ni + 4 × at. % Sn + 8 × at. % Sm)/10

(1)

It is well known that martensites with twin boundaries play an important role in the magnetic field-induced strain of FMSMAs, depending upon the twin boundary migration. The X-ray diffraction patterns of the Ni₄₃Mn_46-x Sm_xSn₁₁ (x = 0, 1, 2, 3) alloy samples are shown in Figure 1.

The reflection peaks corresponding to the cubic austenite phase are indicated by green squares. On comparing the NiMnSn alloys reported by Krenke et al. [35], the main reflection peaks corresponding to the martensitic phase are indicated by red circles, and extra peaks which indicate the presence of an extra γ phase are represented by blue triangles. The XRD pattern for x = 0 shows typical diffraction peaks from the austenite phase, indicating a single cubic L2₁ crystal structure. As MT is closely linked to the crystal structure, the presence of the L2₁ cubic structure for x = 0 indicates that its T_t is below the RT. For the x = 1 and 2 Sm-doped samples, a mixed complicated structure of the cubic L2₁ phase, orthorhombic martensite phase and γ phase is exhibited. The coexistence of austenite and martensite phases indicates that the T_t lies near RT for these samples. Ni₄₃Mn₄₃Sn₁₁Sm₃ presents a two-phase structure, i.e., orthorhombic martensite phase and γ phase. The disappearance of the austenite phase in Ni₄₃Mn₄₃Sm₃Sn₁₁ indicates that the T_t is increased from RT. This is in good agreement with the DSC measurement, which is discussed in the next session. For x = 0, the alloy sample crystallizes in the cubic L2₁ structure with lattice parameters a = b = c = 5.986 Å. In the case of x = 1 and 2, there is the appearance of an orthorhombic martensite phase with dominating cubic austenite with lattice parameters a = b = c = 5.784 Å. It is well known that the crystal structure of NiMnSn alloys is sensitive to the composition. With Sm doping, the 220 peak shifts minutely towards a higher angle, and the cell volume decreases. When Sm is increased to 3 at. %, the crystal structure changes completely to the 10 M orthorhombic phase with lattice parameters a = 476 Å, b = 5.64 Å and c = 26 Å.

Furthermore, it can also be seen that as the Sm content is increased, the γ peak becomes stronger. Therefore, it can be speculated that the appearance of unknown peaks is due to the presence of Sm, as reported earlier for rear-earth metals [36]. A detailed investigation about the γ phase may be obtained from the temperature-dependent XRD and TEM analyses. Therefore, XRD patterns show that with increasing Sm contents, the structural transformation from austenite to martensite takes place, which is responsible for an increase in the T_t.

The stability of the martensite temperature is a key factor in the application temperature of FMSMAs based on twin boundary migration [34]. Therefore, it is very important to tune the T_t temperature in a wide stable temperature range. The thermal-induced structural transformations in the Ni₄₃Mn_46-x Sm_xSn₁₁ (x = 0, 1, 2, 3) alloys were studied by DSC analysis from 160 to 360 K at a ramp rate of 10 K/min during heating and cooling cycles, as shown in Figure 2a. The existence of large exothermic/endothermic peaks during the heating/cooling process is assigned to the occurrence of reverse martensitic/martensitic transformation in the alloys. These peaks confirm the structural transformation between the austenite and martensite phases.

The first-order nature of the transformation is evident by the presence of thermal hysteresis among the heating and cooling processes [37]. The characteristic MT temperatures are the austenite start temperature A_s, austenite finish temperature A_f, martensite start temperature M_s and martensite finish temperature M_f. The T_t in heating and cooling processes can be calculated using the equations T_tH = (A_s + A_f)/2 and T_tC = (M_s + M_f)/2. The characteristic temperatures of the MT of the Ni₄₃Mn_46-xSm_xSn₁₁ (x = 0, 1 and 3) alloys are listed in

Table 2. The measured values of characteristic temperatures for Ni₄₃Mn_46-xSm_xSn₁₁ (x = 0, 1, 2 and 3) alloys.

x	As (K)	Af (K)	Ms (K)	Mf (K)	Tt (Heating) (K)	Tt (Cooling) (K)
0	200	214	197	184	207	190
1	269	276	261	247	272	254
2	294	305	292	278	299	285
3	314	327	311	294	320	302

. The sample without Sm doping (x = 0) is found far below the RT (A_s = 200 K, A_f = 214 K, M_s = 197 K and M_f = 184 K). With a small amount of Sm doping (x = 1), the T_t increases, with a range of 62 to 69 K, i.e., the A_s, A_f, M_s and M_f are increased by 69, 62, 64 and 65 K, respectively. An increase in the T_t has already been reported in FMSMAs by Gd (rare element substitution) [36,37,38]. With the further increase, the M_t shows an overall increasing tendency.

Many investigations on FMSMA alloys show that the composition of the constituents influences the T_t. The results reveal that the valance electron concentration (e/a) is the most significant factor affecting the martensitic transformation, and that there is a linear relationship between the T_t and e/a [34]. The second important factor that affects the MT is the change in the matrix composition. It is agreed that the substitution with atoms having different sizes from their parent atoms results in the expansion or contraction of the unit cell, and hence lattice distortion. Figure 2b shows that as the e/a is increased, the characteristic temperatures of martensitic transformation increase linearly, which is in accordance with previous reports [32].

The temperature dependence of the magnetization (M–T) curves of the Ni₄₃Mn_46-xSm_xSn₁₁ (x = 0, 1, 2 and 3) alloys were measured to investigate the magnetic phase transition. M–T curves were measured under the applied magnetic field of 0.01 T, with a temperature range of 150–400 K, during heating and cooling processes, as shown in Figure 3a–d. It can be seen that for x = 0, the alloy sample experiences a magnetic order–disorder transition of martensite at about ≈175 K, followed by reverse martensitic transformation to ferromagnetic austenite at ≈200 K during heating.

With a further increase in the temperature to about ≈310 K, the magnetization gradually decreases to zero, which corresponds to the Curie temperature of austenite. During the cooling process, however, a sudden drop in magnetization can be observed at about ≈197 K, corresponding to the occurrence of MT from the martensite to the austenite phase. Irreversibility during the heating and cooling processes is observed, which is the signature of the first-order transition and is frequently seen in these alloys [16]. It is seen that the T_t increases remarkably with the Sm addition. For x = 1, martensitic transformation occurs between the paramagnetic austenite and ferromagnetic martensite, and the T_t increases by ≈ 65 K, but the Curie temperature of austenite almost remains the same. As shown in Figure 3c,d, the behavior is different for x = 2 and 3: here, the structural transition occurs in the paramagnetic state, and magnetic transition disappears, indicating that the magnetic field cannot induce the transition in x = 2 and 3. The characteristic temperatures measured in the M–T curves agree well with the DSC measurements.

It can be seen from the DSC and M–T measurements that the Ni₄₃Mn₄₅S_mSn₁₁ alloy undergoes structural as well as magnetic transition near the T_t. Therefore, isothermal magnetization (M–B) curves under the applied magnetic field of 5 T around MT were measured to investigate the field-induced transition, as shown in Figure 4a. Initially, the sample was cooled down to complete the martensitic state and then slowly heated to the target temperature before starting each M–B measurement, known as the loop method [33]. This alloy exhibited a weak magnetic behavior at temperatures below the T_t (248 and 257 K), while near the T_t (260, 263 and 266 K), it showed a metamagnetic behavior, which indicates that the magnetic field can induce the transition and then transform to a strong magnetic state at 269 K. Hence, it can be concluded from the M–B curves that the magnetic field induced MT between the weak magnetic austenite and strong magnetic martensite state. A magnetization difference (ΔM) of about 38 Am²/kg under a field of 5 T was observed in this alloy. This ΔM is attributed to the field driving capacity [39].

The structural transformation between the austenite and martensite phases in the Ni₄₃Mn₄₅SmSn₁₁ alloy can be driven by the magnetic field, indicated by the magnetization difference. The M–T curves of the Ni₄₃Mn₄₅SmSn₁₁ alloy were measured under a field of 5 T, as shown in Figure 4b. If a comparison is made with M–T at 0.01 T, the T_t decreased by about ≈9 K, and ΔM increased by ≈37 Am²/kg, with the rate of A_s shifted by the magnetic field, i.e., dAs/dB, being −1.8 K·T⁻¹, confirming the field-induced MT. This magnetic field driving capacity for the Ni₄₃Mn₄₅SmSn₁₁ alloy can be attributed to the larger value of ΔM. Commonly, M–T curves at different magnetic fields can be used to obtain the values of magnetic entropy change (ΔS_M). The transformation hysteresis defined as A_f–M_s [39] was 18 and 16 K under the applied magnetic fields of 5 and 0.01 T. Here, we calculated the ΔS_M of the Ni₄₃Mn₄₅SmSn₁₁ alloy using the Clausius–Clapeyron equation (ΔS = −ΔM·dB/dT) from the M–T curves at 0.01 and 5 T. The obtained value of ΔS was 20.51 J·K⁻¹·kg⁻¹ (using ΔM = 37 Am²/kg, change in magnetic field (ΔB) = 4.99 T and temperature change (ΔT) = 9 K), which is in good agreement with the ΔS_M obtained from the M–B curves using Maxwell relations.

The temperature-dependent ΔS_M of the Ni₄₃Mn₄₅SmSn₁₁ alloy was calculated by the M–B curves using the Maxwell relation given in Equation (2). The measured ΔS_M for the Ni₄₃Mn₄₅SmSn₁₁ alloy was 20 J·kg⁻¹·K⁻¹ at a temperature of 266 K, with a magnetic field variation of 0–5 T, as shown in Figure 5. The ΔS_M originates from a magnetic field-induced magnetostructural transition. It can be seen that with 1% doping of Sm, the working temperature of the NiMnSn alloy increased to 266 from ~205 K, showing potential for magnetic refrigeration applications near RT. Moreover, the ΔS_M obtained for Ni₄₃Mn₄₅S_mSn₁₁ at 266 K is comparable to many other first-order magnetocaloric materials [16,34,39,40,41].

$$\Delta {\mathrm{S}}_{\mathrm{M}}={\displaystyle \underset{0}{\overset{\mathrm{B}}{\int }}\frac{\mathsf{\delta }\mathrm{M}(\mathrm{B},\mathrm{T})}{\mathsf{\delta }\mathrm{T}}}\mathrm{dB}$$

(2)

The maximum value of the ΔS_M derived from Equation (2) is smaller than the ΔS_T calculated from the DSC measurements because the magnetic field of 5 T is unable to undergo a complete phase transition.

The refrigeration capacity (RC) is a very important factor to assess the MCE and can be calculated by integration of the area under the ΔS_M–T curves, using the temperature at half maximum of the peak as the upper and lower integration limits [34]. The calculated RC for the Ni₄₃Mn₄₅S_mSn₁₁ alloy under the 5 T field was 262 J·Kg^–1. The high MCE and RC values near RT show that the Ni₄₃Mn₄₅S_mSn₁₁ alloy can be used in magnetic refrigeration.

4. Conclusions

The effect of Sm doping on martensitic transformation and the magnetic properties of Ni₄₃Mn_46-xSm_xSn₁₁ ferromagnetic shape memory alloys was investigated. The results show that Sm doping changes the crystal structure from cubic L2₁ to orthorhombic martensite, with a fraction of the γ phase. The DSC measurements showed endothermic/exothermic peaks, indicating the martensitic and reverse martensitic transformation. The T_t increased significantly with the Sm content and exceeded RT for Sm = 3 at. %. The Ni₄₃Mn₄₅SmSn₁₁ alloy exhibited a magnetostructural transition, and a magnetic field-induced transition was confirmed by the high-field M–T and M–B measurements. Moreover, a magnetic entropy change (ΔS_M) value of 20 J·kg⁻¹·K⁻¹ was reached at a temperature of 266 K, and the RC was about 166 J·kg⁻¹, obtained for the Ni₄₃Mn₄₅SmSn₁₁ alloy under an applied magnetic field of 5 T. The large ΔS_M value near RT for the Ni₄₃Mn₄₅SmSn₁₁ alloy shows its potential for magnetic refrigeration application.