1. IntroductionMagnetic refrigeration has drawn great attention from researchers because of its advantages over the conventional cooling technique, such as the compact size, low mechanical vibration, environmental friendliness, and high efficiency.[1–3] Materials exhibiting the large magnetic entropy changes (
are potential candidates for magnetic refrigeration applications.[4] Till now, the large
has been widely studied in materials experiencing the magnetic-field-induced first-order transition, e.g., La–Fe–Si,[5–7] Gd–Si–Ge,[8] Mn–Fe–P–As,[9] and MMnX (M= Co or Ni,
Si or Ge) alloys.[10–12] Besides these materials, Heusler-type NiMn-based magnetic shape memory alloys also attract considerable attention for its large and tunable magnetocaloric effect.[13–20]
The magnetocaloric effect in NiMn-based magnetic shape memory alloys is generated from the magnetic-field-induced first-order magnetostructural transformation from martensite to austenite.[21–23] According to Liuʼs report, the adiabatic temperature change with a magnetic field variation of 0–2 T is as large as −6.2 K in Ni–Mn–In–(Co) alloy,[24] which is larger than the magnetocaloric effect of Gd. Additionally, due to the fact that the magnetostructural transformation temperature is sensitive to the valence electron concentration (
,[25] cell volume,[26] grain size,[27] and external parameters (magnetic field or applied hydrostatic pressure),[28] the magnetostructural transformation can be tuned by changing the composition or preparation conditions.[29–35] For instance, the magnetostructural coupling and magnetocaloric effect can be effectively tailored through changing the ratio between Ni and Mn magnetic elements or introducing a fourth magnetic element.[30–32] The doping effect of non-magnetic elements on the magnetostructural transformation and the magnetocaloric properties in Ni–Mn–In alloys has been theoretically studied by Sokolovskiy et al.[33] They predicted that the adiabatic temperature change can be enhanced by a factor of 3 when the Mn atoms are partially replaced by non-magnetic Cu atoms in the Ni50Mn34In16 alloy.[33] Recently, Kaya et al. also reported the improvement of magnetocaloric properties through introducing non-magnetic Al element in the Ni43Mn46In11 alloys.[35] Based on these previous experimental and theoretical reports, it is expected that nonmagnetic V element may also be able to tune the magnetostructural transformation and magnetocaloric properties in NiMn-based Heusler alloys.
In this work, we investigate the magnetostructural transformation and magnetocaloric properties of Mn48−xVxNi42Sn10 (x = 0, 1, 2, and 3) ferromagnetic shape memory alloys. The martensitic transformation temperature decreases when Mn is substituted by V. High field M–T and isothermal magnetization curves confirm the magnetic-field-induced magnetostructural transformation. The tunable martensitic transformation temperature, enhanced magnetic field driving capacity, and large magnetic entropy changes are observed.
3. Results and discussionFigure 1 shows the XRD patterns of Mn48−xVxNi42Sn10 (x = 0, 1, 2, and 3) alloys at room temperature. For x = 0 and 1, a small fraction of martensitic phase is detected in addition to the dominant cubic austenite phase at room temperature. With further increase of x, (022) peak from the martensite shifts to the higher angle and overlaps with the (220) peak from the austenite. It is known that the structure of austenite in Ni–Mn–Sn magnetic shape memory alloy is sensitive to the composition. It could be cubic L21, Hg2CuTi, or B2 structure.[37] In this case, the superlattice (111), (222), (400), and (422) reflections confirm L21 cubic structure. As the size of V (0.134 nm) is larger than that of Mn (r=0.126 nm), with V substitution, the lattice parameter (a) and unit cell volume increase, indicating that V substitution may produce the lattice distortion in these alloys.
Figure 2 shows the DSC curves of Mn48−xVxNi42Sn10 (x = 0, 1, 2, and 3) alloys during the heating and cooling processes. The large endothermic/exothermic peaks represent the occurrence of martensitic and reverse martensitic transformations. The thermal hysteresis, which represents the irreversibility between the heating and cooling processes, identifies the first-order nature of the transformation.[38] According to the DSC data, the start and finish temperature of reverse martensitic transformation (As and Af) and the start and finish temperature of martensitic transformation (Ms and Mf) are listed in Table 1. It can be seen that with the V-substitution, the structural transformation temperature (Tt decreases. Here, Tt in heating (
and cooling (
processes are defined by
and
, respectively. The entropy change accompanied with the complete martensitic transformation (
can be determined from the DSC measurements.[39] The estimated values of
are 32, 30, 29, and
for x = 0, 1, 2, and 3, respectively.
Table 1.
Table 1.
Table 1.
The measured elemental compositions and corresponding electronic concentrations for the present Mn48−xVxNi42Sn10 (x = 0, 1, 2, and 3) samples.
.
x
|
Mn/at.% |
Ni/at.% |
Sn/at.% |
V/at.% |
e/a
|
0 |
47.88 |
42.17 |
9.95 |
00 |
7.961 |
1 |
46.85 |
42.09 |
9.91 |
1.15 |
7.942 |
2 |
45.90 |
42.11 |
10.07 |
1.92 |
7.923 |
3 |
44.88 |
42.16 |
10.11 |
2.85 |
7.904 |
| Table 1.
The measured elemental compositions and corresponding electronic concentrations for the present Mn48−xVxNi42Sn10 (x = 0, 1, 2, and 3) samples.
. |
The temperature dependence of magnetization (M–T) curves for Mn48−xVxNi42Sn10 (x = 0, 1, and 2) alloys were measured to investigate the magnetic phase transitions. Figure 3 shows the M–T curves of Mn48−xVxNi42Sn10 (x = 0, 1, and 2) alloys under an applied field of 0.01 T. With increasing the temperature, all the samples transform from a weak magnetic martensite to a ferromagnetic austenite, as shown in Fig. 3. A thermal hysteresis of about 11 K is observed for all the three samples accompanied with the martensitic transformation, indicating the first-order nature for the magnetostructural transformation. The estimated values of As, Af, Ms, Mf,
, and
from M–T curves are listed in Table 2. These characteristic temperatures for the transitions from the M–T curves agree well with the DSC measurement (Fig. 2). It is known that Tt is related to the value of e/a.[31] In this case, e/a decreases with the substitution of V for Mn, as Tt decreases. Additionally, Tt in NiMn-based magnetic shape memory alloy is also sensitive to the cell volume. The partial replacement of Mn by V atoms induces the expansion of the unit cell, which also shifts the magnetostructural transformation to lower temperatures.
In order to investigate the magnetic-field-induced structural transformation, the isothermal magnetization (M–B) curves for Mn48−xVxNi42Sn10 (x = 0, 1, and 2) alloys in the field of 0–5 T were measured by the so-called loop method.[36] Initially the samples were cooled down to complete the martensitic state and then slowly heated to the target temperature before starting each M–B measurement. As shown in Figs. 4(a)–4(c), the curves show an obvious metamagnetic behavior around Tt, which corresponds to the magnetic-field-induced structural transformation from a week magnetic martensite to a ferromagnetic austenite. The magnetic hysteresis reflects the first-order nature of the transition. The values of magnetization difference (
under the field of 5 T are 41.6, 43.9, and 46.7 Am2/kg for x = 0, 1, and 2, respectively. The increase of
in Mn46V2Ni42Sn10 is attributed to the enhanced field driving capacity.
Owing to the magnetization difference between the martensite and austenite, the structural transformation can be driven by the magnetic field. With increasing the magnetic field from 3 to 5 T, the M–T curves of Mn48−xVxNi42Sn10 (x = 0, 1, and 2) alloys shift to the lower temperature, as shown in Fig. 4(d). The rate of As shifted by magnetic field, i.e.,
dB, is −1.4, −1.7, and
for the x = 0, 1, and 2 samples, respectively, confirming the field-induced magnetostructural transformation. The enhanced magnetic field driving capacity with the V-substitution in Mn48−xVxNi42Sn10 (x = 0, 1, and 2) alloys can be attributed to the larger magnetization difference. It is also well known that the maximum value of magnetic entropy change (
can be estimated by M–T curves measured at different constant fields. We calculated the
value for Mn46V2Ni42Sn10 alloy using Clausius–Clapeyron equation (
from M–T curves at 3 and 5 T, as shown in Fig. 4(d). The calculated maximum value of
is
(using
Am2/kg,
T, and
K), and it is in a good agreement with that obtained from M–B curves using Maxwell relations.
According to the M–B curves, the
during the magnetic-field-induced structural transformation is calculated by Maxwell relation
The temperature dependence of
with a magnetic field variation of 0–5 T for Mn
48−xV
xNi
42Sn
10 (
x = 0, 1, and 2) alloy is shown in Fig.
5. The maximum values for
are 15.2, 18.8, and
for
x = 0, 1, and 2, respectively, which are smaller than
derived from the DSC measurements. The low
derived from Maxwell relation originates from the incomplete phase transition driven by a magnetic field of 5 T, as shown in Figs.
4(a)–
4(c). The large values of
for Mn
48−xV
xNi
42Sn
10 are comparable with other magnetocaloric materials experiencing a first-order transition, such as Ni
48−xCo
2Mn
38+xSn
12, Mn
50Ni
40−xCo
xSn
10, Mn
50Ni
40Sn
10−xSb
x, Mn
1−xAl
xCoGe, and Mn
1+xCo
1−xGe alloys.
[39–44] It is known that the value of
in a material that experiences the magnetic-field-induced magnetostructural transformation is highly related to the sharp variation in magnetization between austenite and martensite. In NiMn-based magnetic shape memory alloys, the magnetization is mainly induced by the Mn atoms and very sensitive to the Mn–Mn distance. As a result, a small variation in Mn atoms may cause lattice distortion in these alloys. The large values of
are attributed to two different interaction mechanisms: one is the strong spin–phonon interaction which leads the different magnetic states between two phases, while the other one is magnetostructural coupling, resulting in the large magnetization jump around martensitic transformation.
The effective refrigeration capacity (RCeff) of Mn48−xVxNi42Sn10 (x= 0, 1, and 2) alloys is calculated by subtracting the average hysteresis loss (HL) from the refrigeration capacity (RC) values. Here, the value of RC is obtained by integrating the area under
–T curves using the temperature at half maximum of the peak as the integration limits.[38] The average HL is calculated by the method mentioned in Ref. [45]. The temperature dependence of HL for Mn46V2Ni42Sn10 alloy is shown as the inset of Fig. 5. The values of RCeff around the temperature of martensitic transformation under the field change of 0–5 T are 37.21, 75.4, and
for x = 0, 1, and 2, respectively. Consequently, the substitution of V for Mn leads to an increase in the magnetization jump at the magnetostructural transition, the enhanced magnetic field driving capacity (d
dB), and effective refrigeration capacity (RCeff).