Large Elastocaloric Effect Driven by Low Stress Induced in [001]-Oriented Polycrystalline Co_51.6V_31.4Ga₁₇ Alloy

Abstract

In this work, we have studied the elastocaloric effect in directionally solidified Co_51.6V_31.4Ga₁₇ alloys with a strong [001] preferred orientation. The entropy change for thermal-induced martensitic transformation is determined as 19.6 J kg⁻¹ K⁻¹. The sample exhibits stress-induced martensitic transformation with a hysteresis of 46 MPa, and the superelasticity is also verified by the in situ X-ray diffraction method. According to the elastocaloric effect tests, a noticeable change in adiabatic temperature up to 12.2 K has been achieved at the strain of 6%. The specific temperature change upon the critical stress loading can be attained as 132 K MPa⁻¹. In addition, the difference in the loading–unloading temperature change can be ascribed to the imperfect adiabatic environment.

1. Introduction

Considerable attention among researchers has been paid to the development of advanced refrigeration technology in recent years. This is because of the expected enhanced efficiency of traditional gas compression refrigeration and the greenhouse gas emissions of the refrigerant. By searching 600 million organic compounds in the PubChem database, it was concluded that there are currently no unknown compound refrigerants [1]. Being one of the most promising new refrigeration technologies, solid-state refrigeration has aroused great interest as an environmentally friendly and energy-saving technique. In particular, elastocaloric refrigeration based on the elastocaloric effect (eCE) has recently become a popular topic due to its simple uniaxial loading style and wide refrigeration temperature range. At present, the reported elastocaloric materials include shape memory alloys (SMAs), oxide ceramics [2], natural rubber [3], and graphene [4]. Specifically, the eCE in SMAs has been studied extensively and in depth over the past few decades. Among the mainstream elastocaloric alloys, there are Ni-Ti alloys [5,6,7], Cu-based alloys [8,9], Heusler-type alloys [7,10,11], Fe-based alloys [12,13], and so on.

The external factors, including applied stress field, working temperature, and adiabatic environment, can directly affect the eCE of materials [14,15,16]. Moreover, the chemical composition, crystal structure, and grain orientation of the alloy may impact the eCE associated with the stress-induced martensitic transformation [17,18,19,20,21,22]. For anisotropic SMAs, the mechanical properties, transformation strain, and critical stress may vary with grain orientation. In particular, a large amount of research has been performed on the eCE of polycrystalline, single-crystal, and textured polycrystalline Ni-Mn-based Heusler alloys [12,17,18,19,20,21,22,23,24]. According to various reports, the eCE properties of Ni-Mn based alloys are directly related to the temperature sensitivity of the transformation strain and critical stress, and strongly refer to the crystallographic orientation. For example, the adiabatic temperature change values of the [001] oriented Ni₅₀Mn₃₂In₁₆Cr₂ alloys are 1.9 K and 2.6 K, respectively, larger than those of the [111] and [110] oriented Ni₅₀Mn₃₂In₁₆Cr₂ alloys [16]. For polycrystalline Ni_50.4Mn_27.3Ga_22.3 alloys with different orientations, the adiabatic temperature changes depend on whether the loading direction is parallel or perpendicular to the [001] direction [14]. In Ni₄₅Mn₄₄Sn₁₁ alloy, the thermal responses of samples with [111] orientation vary from those corresponding to other orientations under the same stress conditions [15].

Therefore, the grain orientation of the alloy has a significant influence on its eCE. Apart from the aforementioned Ni-Mn based alloys, Co-Ni-Ga alloys, as a representative of Co-based Heusler alloys, have also attracted much attention in recent years. It is noteworthy that in the case of Co-based alloys with grain orientation, most researchers mainly focus on the microstructure, magnetic properties, and martensitic transformation of Co-Ni-Ga textured polycrystalline alloys prepared by directional solidification or single-crystal growth [25,26,27,28]. Meanwhile, the elastocaloric effect in textured polycrystalline alloys has less been studied systematically. A pronounced eCE has been found in Co-V-Ga alloys undergoing a stress-induced phase transformation from the L2₁ atomic ordered austenite to the DO₂₂ tetragonal martensite [29,30,31,32]. However, the reported eCE was observed in texture-free polycrystalline alloys. Therefore, this study aims to investigate the eCE in textured polycrystalline Co-V-Ga alloys.

2. Materials and Methods

Co_51.6V_31.4Ga₁₇ polycrystalline alloy ingots were prepared by arc-melting pure Co (99.99%), V (99.99%), and Ga (99.99%) in an argon atmosphere, and then casting them into a 7-mm-diameter cold copper mold. The rods were subsequently re-melted at 1773 K and grown in alumina crucibles via the liquid–metal cooling directional solidification method, with a pulling rate of 150 μm s⁻¹. The compressive test specimens were cut from the steady growing zone of the solidified sample into rectangular parallelepipeds with nominal dimensions of 3 mm × 3 mm × 6 mm, so that their long axis was along the growth direction. The specimens were afterward annealed at 1473 K for 24 h in an Ar atmosphere and quenched in ice water.

The temperature dependence of magnetization M-T and magnetic field dependence of magnetization M-H were measured by a vibrating sample magnetometer attached to a superconducting quantum interference device (SQUID-VSM, Quantum Design, San Diego, CA, USA). The calorimetric measurements were performed using differential scanning calorimetry (DSC, NETZSCH DSC 214) at the heating/cooling rate of 10 K min⁻¹. The texture analysis was conducted via electron backscattering diffraction (EBSD, OXFORD Nordlys Max2). Prior to the EBSD experiments, specimens were exposed to ion beam polishing (LEICA EM TIC3X) to reduce surface residual stress. Crystal structure testing was implemented on a two-dimensional X-ray diffraction system (VANTEC500 XRD4). Before the XRD analysis, the samples were compressed with a self-designed micro-compressive device to maintain their compression state. The room-temperature compressive experiments were conducted on a universal compression testing machine (SUNS-UTM5000) at 303 K. The surface temperature changes were monitored via infrared (IR) thermography on samples pre-sprayed with black matte paint. For compressive tests, the loading was applied along the direction of solidification growth. For the eCE measurements, the samples were rapidly loaded–unloaded at a strain rate of 2.58% s⁻¹ for each curve. The high strain rate provided a quasi-adiabatic environment, and a holding step of 30 s was applied after loading to ensure that the sample returned to the thermal equilibrium state before unloading started.

3. Results and Discussion

Figure 1a displays the DSC and M-T curves acquired under a magnetic field of 5 T during heating and cooling. Applying the tangent extrapolation method to the DSC curves enabled us to establish the start and finish temperatures of the forward martensitic transformation (M_s and M_f) and the reverse transformation (A_s and A_f) as 270 K, 255 K, 272 K, and 285 K, respectively. The transformation equilibrium temperature T₀, determined as a function of T₀ = (A_f + M_s)/2 [33], was 271 K. The average transformation enthalpy change ΔH_tr found from the area under the exothermic and endothermic peaks was 5.3 J g⁻¹. The transformation entropy change ΔS_tr associated with the martensitic transformation was determined to be 19.6 J kg⁻¹ K⁻¹ according to the following formula: ΔS_tr = ΔH_tr/T₀. In addition, the characteristic martensitic transformation temperatures confirmed from the M-T curve were almost consistent with those obtained from the DSC curve. The magnetization difference ΔM in the two phases crossing the martensitic transformation was only 1.9 emu g⁻¹ under the magnetic field of 5 T (Figure 1a), indicating that it is difficult to drive the phase transition by applying a magnetic field.

Figure 1b depicts the EBSD orientation micrograph of the longitudinal section for the Co_51.6V_31.4Ga₁₇ alloy at room temperature, where the alloy was in the austenite phase state. The EBSD phase calibration was performed assuming the lattice parameter a of 5.80 Å, the space group Fm3m (225), and the calibration rate of 93.7%. The inverse pole figure (see the inset) reveals that the sample had a strong [001]_A preferred orientation along the solidification growth direction. Generally, the solid–liquid interface energy required for the [001] orientation growth of the cubic phase was lower than that in other orientations, so the [001] orientation tended to be the most common growth direction. Such a microstructural feature in the directionally solidified alloy also plays a significant role in decreasing the transformation hysteresis due to the reduced amount of grain boundaries [34,35]. Hence, the [001] orientation with low critical stress is the preferred uniaxial stress loading direction, leading to a maximum transformation strain [36,37], being similar to previously reported SMAs, including Cu-Al-Mn [38], Ni-Mn-In-Co [39], Ni-Mn-Ga [40], and Ni-Co-Mn-Ti [41].

Figure 2a presents the isothermal compressive stress–strain curve for the textured Co_51.6V_31.4Ga₁₇ alloy at 303 K, which is above A_f = 285 K. The strain rate of 0.03% s⁻¹ was applied to provide sufficient heat exchange between the sample and the surrounding environment. The sample was then loaded slowly in an isothermal process so that the strain was further increased up to 5.0%. It should be mentioned that the critical stress (σ_cr) required for the alloy is approximately 109 MPa. The critical stress is an important parameter closely related to the eCE of the material, and is usually affected by the testing temperature, microstructure, and deformation conditions. As shown in Figure 2a, once the applied uniaxial stress reached 109 MPa, the alloy started to undergo martensitic transformation, indicating that the critical stress of this alloy was significantly lower than that of the texture-free polycrystalline alloy with the same composition [29]. Moreover, it was inferior to that of the Ni-Ti based alloy system [6,7,42,43,44] and several Ni-Mn based alloys [14,16,45,46]. In addition, the alloy under consideration exhibited very good superelastic behavior, where a nearly reversible strain up to 5.0% could be obtained through the reversible stress-induced martensitic transformation under a relatively low stress of 188 MPa. Under the same stress conditions, the stress-induced transformation fraction of the textured polycrystalline alloy was significantly larger than that of the texture-free polycrystalline alloy [29]. The stress hysteresis of the shape memory alloy is defined as: Δσ_hy = σ_Ms − σ_Af. For our textured alloy, the calculated Δσ_hy hysteresis was only 45 MPa, being lower than those of some Ni-Mn based alloys. The low stress hysteresis reduces the energy dissipation and enhances the repetitive stability of stress cycles [15,47].

A self-designed micro-compressive device was utilized to compress the textured alloy at room temperature and to hold the sample in a compressed state for further testing of its crystal structure. Figure 2b displays the one-dimensional XRD patterns within the angular range of 40–50° under various strains. It was evident that the uncompressed alloy was in the austenite phase. As the amount of applied strain increased, the main (220)_A peak began to split into the peaks of the martensitic phase, indicating the stress-induced martensitic transformation. The diffraction peak at 43° corresponds to the (112)_M. Once the applied strain increased to 3.33%, the alloy transformed into the state of martensite–austenite phase coexistence, which was consistent with the stress–strain curve shown in Figure 2a.

The eCE of the [001] oriented Co_51.6V_31.4Ga₁₇ polycrystal was studied at room temperature, and the temperature change was characterized by infrared thermography (IR). According to Figure 3a, the temperature change value gradually increased with the applied strain. The maximum adiabatic temperature change of 12.2 K was achieved under the applied strain of 6%. Taking the compressive curve and the corresponding infrared spectrum at ε_appl = 5.5% as an example, as shown in Figure 3a, at the critical stress of 91 MPa, the corresponding average temperature changes during loading and unloading, obtained from the rectangular area in the center of the sample (highlighted in Figure 3b), were 12.0 K and 10.9 K, respectively. Moreover, as seen in Figure 3b, the temperature at the edge of the sample was slightly lower due to part of the heat loss originating from the heat exchange between the sample and the clamps [48]. Thus, the IR images revealed that the temperature space evolution of the samples was homogeneous.

Figure 3c depicts the loading–unloading temperature change values under different applied strains. It can be seen that when the strain amount was below 4.5%, the loading temperature changed more slowly than the unloading one. However, with the gradual increase in the applied strain, the values of ΔT_loading and ΔT_unloading gradually decreased. This temperature asymmetry might be related to the imperfect adiabatic environment. In other words, during the initial loading stage, the contact of the compressive machine with the top of the sample was incomplete, resulting in heat leakage. Meanwhile, with the increase in testing time, the loading temperature change gradually exceeded the unloading one. This could be attributed to the fact that the temperature between the compressive machine and the sample tended to be closer, which resulted in the improvement of the adiabaticity, and consequently, the phenomenon of temperature change was reduced due to heat leakage during loading. In addition, according to the stress–strain curve in Figure 3d, as the residual strain after the rapid loading–unloading cycle was approximately 1.4%, we might exclude the possibility of temperature asymmetry being responsible for the generation of the unrecovered martensite phase upon plastic deformation. Instead, the temporary residual strain in the texture-free Co_51.6V_31.4Ga₁₇ polycrystal [29] was not observed in the present work.

The temperature interval between the testing temperature T_test and A_f has a great influence on the eCE of the alloy. In view of this, the typical elastocaloric alloy systems with |T_test − A_f| approximately 20 K were selected for comparison with our textured Co_51.6V_31.4Ga₁₇ alloy (the relevant data taken at ε_appl = 5.5% are listed in

Table 1. The values of $\left|\Delta T_{\mathrm{ad}}/\Delta {{\sigma }}_{\mathrm{cr}}\right|$ for the present textured Co_51.6V_31.4Ga₁₇ alloy and some typical elastocaloric alloys.

Alloys	Af (K)	Ttest (K)	$\Delta$Tad (K)	σcr (MPa)	|$\Delta$Tad/σcr| (K GPa−1)
This work	285	303	12.0	91	132
Co50.7V33.3Ga16 [31]	249	273	17.3	180	96
Co50V35Ga14Ni1 [32]	280	300	12.1	129	94
Co50Ni20Ga30 [28]	264	291	3.7	75	49
Ni45Mn36.4In13.6Co5 [49]	268	293	2.4	96	25
Ni55Mn18Ga26Ti1 [50]	265	295	3.1	85	36
Ni50Mn31.5In16Cu2.5 [48]	280	300	6.0	145	41
Ni55Mn18Ga27 [47]	281	298	10.7	150	71
Ni45Mn44Sn11 [51]	273	298	7.5	250	30
Ni51.4Ti48.6 [52]	318	333	23.2	500	46
(Ni42.5Ti50Cu7.5)99Co1 [53]	286	304	14.4	320	45

). It can be seen that the adiabatic temperature change of the textured alloy under consideration was at a relatively high level, whereas the testing was also near room temperature. In turn, the critical stress exceeded those of Ni-Mn based alloys [47,48,49,50,51], but was significantly lower than that of Ni-Ti based alloys [52,53]. It is worth noting that our textured alloy exhibited a relatively high $\left|\Delta T_{\mathrm{ad}}/\Delta {{\sigma }}_{\mathrm{cr}}\right|\approx$ 132 K GPa⁻¹, which was lower than other elastocaloric SMAs and higher than that of other components in the Co-V-Ga system [31,32]. The low specific adiabatic temperature change per stress would be beneficial for saving the driven component space and minimizing the size of the elastocaloric cooling device. This indicates that polycrystalline Co_51.6V_31.4Ga₁₇ with the [001] orientation is a promising elastocaloric material with high efficiency and great energy-saving potential for next-generation refrigeration technology.

4. Conclusions

The Co_51.6V_31.4Ga₁₇ polycrystalline alloy prepared by directional solidification exhibited a strong [001] preferred orientation and a large entropy change of 19.6 J kg⁻¹ K⁻¹. The alloy demonstrated good superelasticity. The stress-induced martensitic transformation in the alloy was confirmed from the XRD patterns under stressing, revealing an obvious austenite–martensite coexistence state under the strain of 3.33%. Moreover, a large ΔT_ad of 12.0 K was achieved under the applied strain of 5.5%, resulting in the large value of |ΔT_ad/σ_cr| ≈ 132 K MPa⁻¹. Thus, the [001]-oriented Co_51.6V_31.4Ga₁₇ polycrystalline alloy has application potential in the field of solid-state refrigeration based on eCE.