Martensitic Transformation, Magnetic and Mechanical Characteristics in Unidirectional Ni–Mn–Sn Heusler Alloy

Abstract

A textured structure of Ni–Mn–Sn Heusler alloy with [001] preferred orientation has been grown by the directional solidification method. The crystal exhibits a single austenite phase L2₁ cubic structure (a = 5.997 Å) at room temperature. Magnetization and electronic transport measurements reveal the phase transformation characteristics. The maximum values of magnetic entropy change determined by Maxwell’s thermodynamic relation during the structural and magnetic phase transformations are 3.5 J/kg·K and −4.1 J/kg·K, and the total effective refrigerant capacity reaches about 314 J/kg (5 T). The evident reduction in hysteresis loss and broad operating temperature window provide a greater prospect for improving the cyclic stability of refrigeration and optimizing the application of such a magnetic refrigeration material. Both magnetoresistance (−18%, 5 T) and exchange bias field (302 Oe, 2 K) have also been investigated to understand the nature of phase transformations and exchange interactions. Furthermore, as the material exhibits excellent mechanical properties (1068 MPa, 9.0%), our experimental results provide a new reference for the application of Ni–Mn–Sn Heusler alloys.

1. Introduction

The Ni–Mn–(In, Sn, Sb) Heusler alloys have recently been followed with interest owing to the strong coupling between magnetic and structural free degrees [1,2]. These metamagnetic shape memory alloys (MSMAs) would undergo martensitic transformation (MT) at a certain temperature. Such a magneto-structural phase transformation would result in unique functional properties, including magnetic field-induced strain (MFIS) [3], direct (MCE) and inverse magnetocaloric effect (IMCE) [4,5,6], exchange bias (EB) [7,8], magnetoresistance effect (MR) [9,10], etc. These properties are being further developed for potential future applications in magnetic refrigeration, information storage, and sensing drives. Among the MCE, Ni–Mn-based Heusler alloys have a broader application prospect compared to some rare-earth-based compounds because of their high magnetic refrigeration operating temperature (near room temperature) and lower cost compared to rare earth elements [11,12,13]. Many researchers have carried out a series of studies on Ni–Mn-based Heusler alloys and found that MT in such alloys is very sensitive to composition. One widely accepted explanation is the adjustment of valence electron concentration (e/a) to the phase transformation, as the MT temperature would increase linearly with the increase of e/a within a certain range [14,15]. The parent phase of the alloys is usually an L2₁ cubic structure with high symmetry at high temperatures. During the cooling procedure, the alloys usually change into the L1₀ martensite phase with low symmetry or present several modulation structures of martensite variants [16], which give rise to abundant magnetic properties. For instance, the Heusler alloys typically exhibit the coexistence of ferromagnetic (FM), antiferromagnetic (AFM), superferromagnetic (SFM), or superspin glass (SSG) once the martensite transition has occurred [17,18,19,20]. The various potential complex magnetic states arise from the competing interactions between Mn atoms, which is a key technical origin of the EB effect.

At present, researchers mainly explore and improve the physical properties of Heusler materials by adjusting the proportion of elements and substitution doping of the fourth element [21,22,23,24]. Apart from these, there could also be other extrinsic factors for the properties of materials, such as preparation methods and pretreatment conditions [25,26,27]. The microstructure and magnetic characteristics of Heusler alloys prepared by traditional metallurgy, thin film techniques [28,29], or melt spinning [30] have been reported by many investigators. However, information about Heusler single crystal or unidirectional crystal is relatively limited so far. Single crystals of ferromagnetic shape memory alloys are very fascinating and significant from both a foundational and an application viewpoint. For example, Ni–Mn–Ga single crystal exhibits large strain due to a lack of independent crystals and grain boundaries with different orientations [31]. In addition, enhanced MCE was also reported in the Ni–Mn–Sn unidirectional crystal [32]. In the present paper, a highly [001] orientated Ni–Mn–Sn Heusler crystal was prepared by the directional solidification method. The crystal structural, magnetic, electrical, and mechanical characteristics of the material were studied so as to regulate the microstructure and physical properties by crystal growth and extend the application prospects of such alloys.

2. Materials and Methods

Highly textured Ni₅₀Mn₃₆Sn₁₄ Heusler alloy was grown by Bridgman type directional solidification method from a master alloy. The as-cast ingot was prepared by repeated melting of high purity Ni, Mn, and Sn (99.99%, Alfa Aesar, Ward Hill, MA, USA) under argon atmosphere (an extra 2% Mn was added on account of its volatility), and it then was melted and subsequently suctioned into a copper mold to get a precursor rod. The rod was placed into a corundum crucible in a crystal growing furnace and heated to a maximum of about 1723 K in an atmosphere of argon. After holding the temperature for 30 min, the crucible was pulled into the Ga–In–Sn liquid metal at the speed of 50 μm/s to obtain the final sample. Next, the obtained sample rod was annealed at 1173 K for 24 h, followed by quenching by ice water.

The microstructure of the sample was observed by an optical metalloscope (Leica, Wetzlar, Germany); the chemical composition and crystal orientation were characterized by an X-ray energy dispersive spectrometer (EDS) (EDAX, Pleasanton, CA, USA) and Hikari electron backscattering diffraction (EBSD, Macherio, Italy) high-speed probe on a scanning electron microscope (SEM, QUANTA 450) (Hitachi, Ibaraki, Japan), respectively. At room temperature, X-ray diffraction (XRD, D2-Phaser) (Bruker, Karlsruhe, Germany) with Cu-Kα radiation was performed to identify the crystal structure, and Rietveld refinement was carried out according to the XRD result using the Fullprof software package. A small rectangular sample (virgin sample) was cut in the sample rod for magnetic and electrical transmission measurements. Magnetic and electrical transmission data were collected by a physical property measurement system (PPMS-14) (Quantum Design, San Diego, CA, USA) with associated integrated accessories. Applying a magnetic field is parallel to the crystal growth direction (CGD) in each magnetization measurement. For resistance measurement, the electric current and the magnetic field were parallel and perpendicular to CGD, respectively. In the compression test, a sample with a diameter of 3 mm and a length of 7 mm was compressed to fracture with a constant speed of 0.15 mm/min at room temperature using an MTS809 material testing machine (MTS, Eden Prairie, MN, USA).

3. Results and Discussion

3.1. Microstructure and Crystal Structure

The longitudinal and transverse microstructures of Ni–Mn–Sn crystal at room temperature are shown in Figure 1a,b respectively. The images show that the typical microstructure characterized by coarse dendrites with a diameter of about 0.25 mm in the sample, and the crystal growth direction deviates slightly from the solidification direction (SD). The main reason for this deviation can be attributed to the mechanical vibration of the lower driving equipment when the crystal grows in the crucible. In addition, complete austenite phase exists in the sample, which is also confirmed by the EBSD analysis.

Figure 2a shows the orientation map of the longitudinal section of the alloy at room temperature; the EBSD map was indexed in the light of the austenite structure and shows that there is obvious [001]_A preferred orientation in crystal growth. The large temperature gradient along SD during crystal growth would result in such a microstructure feature. According to the (001) pole figure (see Figure 2b) derived from the EBSD map, the high texture in the sample deviates from axis A2 (i.e., SD) by about 20°, which also corresponds to the metallographic figure. It is also worth noting that the microstructure with high texture is very beneficial to improving the mechanical properties and thermal effects of materials [33,34,35]. Furthermore, the elements and distribution of the sample were checked by EDS (as Figure 2c) and the homogeneity was confirmed. In order to observe the variation of the content of individual constituent elements in the sample, compositional distribution was measured by EDS along the CGD, as shown in Figure 2d, which indicates that the sample shows a stable expected composition ratio in a large range. Therefore, the high quality of the [001]-oriented crystal was confirmed.

Figure 3 depicts the XRD pattern of the Ni₅₀Mn₃₆Sn₁₄ crystal measured at room temperature, and the inset on the upper right is the unit cell of the Ni₅₀Mn₃₆Sn₁₄ crystal. Rietveld refinement (Rwp = 12.61%) suggests that the crystal presents a single austenite phase, i.e., L2₁ cubic (space group Fm$\overline{3}$m) structure with the lattice parameters of a = 5.997 Å and α = 90°.

3.2. Magnetic Properties

The phase transformations and magnetic features of the virgin sample can be represented by thermomagnetic curves of 500 Oe magnetic field, as shown in Figure 4. It can be found that the magnetic transformation temperature of the austenite phase (Curie temperature $T_C^A$) is about 325 K. Below $T_C^A$, the abrupt changes of magnetization are observed in the range 230 K ($M_s$)–170 K ($M_f$) and 195 K ($A_s$)–255 K ($A_f$) corresponding to MT and reverse MT, respectively. Furthermore, field cooling (FC) and zero-field cooling (ZFC) curves split at low temperature and there is an inflection point called blocking temperature $T_B$ (~65 K) in the ZFC curve, which indicates the existence of magnetically inhomogeneous states. This is generally related to competition between FM and AFM systems.

After determining the characteristic temperatures of phase transformation from the thermomagnetic curves, the sample was then subjected to isothermal magnetization measurements at different temperatures. The isothermal magnetization measurements around MT and around the magnetic transformation are demonstrated in Figure 5a,b, respectively. Initially, the sample was cooled from 300 K to 100 K to ensure that it was in a complete martensitic phase before each M-H curve was measured. This step is to prevent the possible influence of residual austenite induced by magnetic field during the previous magnetization measurement. This was followed by heating up to the desired temperature to get the isotherms with the magnetic field increased to 5 T. From Figure 5a, the curves show FM behavior over the measured temperature range. It can be seen that there is an inflection point in the curve during the increase of the magnetic field, and this change of magnetization state corresponds to the reverse MT of field-induced martensite to austenite transformation. Due to the thermal hysteresis of first-order phase transformation, the field down curve does not coincide with field up, and hysteresis loss (HL) occurs. In addition, the magnetization intensity of the sample varies little over a large temperature range, increasing from 47.1 to 72.1 emu/g, such that a smaller magnetic entropy change ($\Delta S_M$) can be expected. As the temperature rises further, the austenite undergoes a magnetic transformation, the magnetization intensity decreases, and the HL disappears (Figure 5b).

In order to further investigate the magnetization behavior at low temperatures, the hysteresis loops of the alloy at different temperatures were measured after cooling from 300 K under a 500 Oe field. As shown in Figure 6, the hysteresis loops obviously shift to the negative field at 2 K and 20 K, which confirms the behavior of the EB effect at low temperatures. The exchange bias field $H_E$ shift is almost invisible at 65 K and disappears completely above this temperature, so it is confirmed as the blocking temperature $T_B$ in the ZFC curve. This effect is associated with unidirectional anisotropy caused by the noncollinear coupling between the FM and AFM interfaces in a martensitic phase separation state. As the temperature increases, the pinning effect of AFM on the FM by the microscopic torque decreases, and the exchange bias field disappears at the blocking temperature [7,36].

The magnetocaloric effect was estimated by the values $\Delta S_M$ calculated from M-H curves data based on Maxwell’s thermodynamic relation [6]:

$$\Delta S_M(T,\Delta H)={\mu }_0{{\displaystyle {\int }_0^H\left(\frac{\partial M}{\partial T}\right)}}_{H}dH$$

(1)

the temperature dependence of $\Delta S_M$ under several magnetic fields is described in Figure 7. The inset is the HL with a magnetic field change of 5 T, that is, the area surrounded by isothermal magnetization loops. IMCE and conventional MCE are confirmed by respective peaks, the maximum $\Delta S_M$ attained is 3.5 J/kg·K, attributable to the structural transformation, and −4.1 J/kg·K on account of magnetic transformation, under the magnetic field change of 5 T. Although limited, the most significant feature is the large temperature widths (60 K and more than 100 K) occupied by two characteristic peaks. This brings about a wide operating temperature window for magnetic refrigeration materials, which is important for applications.

The refrigeration capacity (RC) is another essential parameter for appraising the MCE. It represents the amount of thermal energy transmitted between cold and hot sources in the ideal thermodynamic cycle, and can be defined with the following formula [37]:

$$RC={\displaystyle {\int }_{Tc}^{T_h}{\Delta S}_M}(T)dT$$

(2)

where $T_c$ and $T_h$ are determined as the full-width half the maximum peak (FWHM) of the $\Delta S_M\text{-}T$ curve, which corresponds to temperatures of cold and hot sources. So, the interval $\delta T_{FWHM}=T_h-T_c$ is also defined as the useful operating temperature range of MCE materials. The calculated values of RC for IMCE and MCE are 94 J/kg and 234 J/kg under the magnetic field change of 5 T, as shown in the shaded parts of Figure 7. After subtracting the average HL (14 J/kg for structural transformation and 0 for magnetic transformation, respectively), the total effective refrigerant capacity $R{C}_{eff}^{total}$ amounts to 314 J/kg. This is comparable to Ni₄₃Mn₄₆Sn₁₁ ($R{C}_{eff}^{IMCE}$= 82.9 J/kg, 5 T) [38], Ni_49.8Co_1.2Mn_33.5In_15.5 ($R{C}_{eff}^{IMCE}$= 76.6 J/kg, 5 T) [39], Ni₄₅Mn₃₉Sb₁₂Co₄ ($R{C}_{eff}^{IMCE}$= 118.86 J/kg, 7 T) [40], Ni₄₇Mn₄₀Sn₁₃ ($R{C}_{eff}^{total}$= 269 J/kg, 5 T) [41], Ni_44.9Fe_4.3Mn_38.3Sn_12.5 ($R{C}_{eff}^{total}$= 310.5 J/kg, 5 T) [42], and much greater than Ni₅₀Mn₃₇Sn₁₃ ($R{C}_{}^{total}$= 105 J/kg, 5 T) [43]. High effective refrigeration capacity benefits from small HL and large temperature ranges, and the resulting potential improvement in magnetic refrigeration cyclic stability and broad operating temperature window are advantageous for applications of magnetic refrigeration materials.

In addition, to reveal the effect of magnetic crystal orientation, the magnetic properties of the sample were measured under a magnetic field perpendicular to CGD (not shown here) and all the results were not significantly different compared to the measurements with a parallel magnetic field. This indicates that the Ni–Mn–Sn unidirectional crystal prepared in this study has a low magnetic crystal anisotropy.

3.3. Electrical Properties

To further explore the effect of phase transformations on electronic transport characteristics, the temperature dependence of electrical resistivity has been measured under the action of 0 and 5 T magnetic fields, as shown in Figure 8. Obviously, the resistivity shows a jump in the temperature range of 170 K to 250 K. This results from the fact that the structural phase transformation would change the electrical transport characteristics of the sample, and the parameters for MT are basically consistent with the magnetization measurement in Figure 4. One reason for the resistance change is the difference in the intensity of scattering between electrons and phonons inside the crystal in two phases [44]. On the other hand, the variation of density of states on the Fermi surface during phase transformation is another factor [45,46]. In addition, the characteristic temperatures of phase transformation shift to low temperature by about 10 K under the condition of a 5 T applied magnetic field, which can be ascribed to the fact that the ferromagnetic stability of high-temperature parent phase would be improved with the introduction of a magnetic field, leading to the decrease in phase transformation temperatures. MR was calculated using the formula $MR=\left[\rho (H)-\rho (0)\right]/\rho (0)$, where $\rho (H)$ and $\rho (0)$ are the resistivity under a 5 T magnetic field and zero field respectively, as shown in the inset of Figure 8. A maximum of about −18% MR (205 K) and hysteresis was observed during both cooling and heating procedures. Negative MR is on account of the transformation from the high resistance martensite to low resistance austenite phase induced by magnetic field. Due to a transition from a less ordered state to more ordered state during heating reducing residual resistivity associated with the disorder, the maximum MR obtained during the heating procedure is smaller than that of the cooling procedure [47].

3.4. Mechanical Properties

The inherent brittleness of Mn-rich Ni–Mn-based Heusler alloys is the key factor hindering their engineering applications. Figure 9 shows the stress–strain curves of the materials compressed to fracture at room temperature; the SEM fracture cross-section is shown in the insets. A small cylinder with a diameter of 3 mm and a length of 7 mm was cut from the sample rod for a compression experiment. The direction of stress applied during compression of [001]-oriented sample is parallel to the SD with a constant speed of 0.15 mm/min. As a contrast, we performed a compression test under the same conditions on a non-oriented polycrystalline sample. The result shows that the maximum compressive strength of the [001]-oriented sample is 1068 MPa and the fracture strain is 9.0%, which is increased by 63.8% and 114%, respectively, compared with 652 MPa and 4.2% of the polycrystalline sample. This mechanical property is excellent compared with previous studies on polycrystalline Ni–Mn-based alloys, and it can be compared with some Co-based alloys [48,49]. The reason can be attributed to the fact that directional solidification eliminates the transverse grain boundaries sensitive to void and crack in the crystallization process, leaving a negligible amount of longitudinal grain boundaries, and making the grain boundaries parallel to the stress axis, thus improving the unidirectional mechanical properties of the material. The microstructure texture can help encourage strain coordination between grain boundaries and decrease the stress concentrations at triple junctions [50]. Furthermore, Chen et al. found unprecedented hyperelasticity in [001]-oriented Ni–Co–Fe–Ga crystal and explained that it is probably because of the lack of fluidity of the habit plane interfaces along with its remarkable lattice anti-slip resistance [51], which may also apply to our material. The good mechanical properties also provide the basis for its possible elastocaloric effect (eCE) [52,53], which deserves further experimental study. Based on the fracture morphologies, the non-oriented sample presents an intergranular crack of intermetallics with fracture flake and smooth crystal interface, and the fracture mechanism of the [001]-oriented sample is a mixture of intergranular fracture and transgranular fracture. Thus, the ductility of the sample is enhanced. It is unambiguously certain that this unidirectional crystal growth method can improve the mechanical properties of Ni–Mn-based alloys and broaden their application prospects.

4. Conclusions

In the present work, a Ni₅₀Mn₃₆Sn₁₄ Heusler alloy with unidirectional crystal was grown by the directional solidification method. The crystal exhibits a single austenite phase L2₁ cubic structure at room temperature, and the lattice parameter is a = 5.997 Å. The total effective magnetic refrigerant capacity amounts to about 314 J/kg under the magnetic field change of 5 T, and the broad operating temperature window and small magnetic hysteresis loss are significant conditions for improving the refrigerant performance. The exchange bias field of 302 Oe was observed (2 K, 500 Oe field cooling) as a result of the coexistence of FM and AFM interactions at low temperatures. It is noteworthy that the magnetic properties of [001]-oriented Ni–Mn–Sn unidirectional crystal do not exhibit magnetic anisotropy. Measurements of magnetization and resistivity characterize the phase transformation of the sample and show −18% magnetoresistance (205 K, 5 T) during the cooling procedure. Furthermore, the ultimate compression test shows that the mechanical properties of the [001]-oriented sample (fracture strength and strain are 1068 MPa and 9.0%, respectively) are significantly better than the polycrystalline sample, which could extend the engineering applications of such alloys.