We have investigated the field direction dependence of thermo-magnetic behavior in single crystalline Mn5Ge3. The adiabatic temperature change ΔTad in pulsed fields, the isothermal entropy change ΔSiso calculated from static magnetization measurements, and heat capacity have been determined for fields parallel and perpendicular to the easy magnetic direction [001]. The isothermal magnetization measurements yield, furthermore, the uniaxial anisotropy constants in second and fourth order, K1 and K2. We discuss how the anisotropy affects the magneto-caloric effect (MCE) and compare the results to the related compound MnFe4Si3, which features an enhanced MCE, too, but instead exhibits strong easy plane anisotropy. Our study reveals the importance of magnetic anisotropy and opens new approaches for optimizing the performance of magnetocaloric materials in applications.

The interest in magnetic refrigeration as a new energy efficient and environmentally friendly solid-state cooling technology around room temperature has increased significantly in the last few years due to the concern about global warming and an ever-rising energy consumption. The magnetocaloric effect (MCE), which forms the basis of this refrigeration technology, is defined as the change of temperature (heating or cooling) and entropy of a magnetic material due to a varying magnetic field.1–3 This effect can be characterized quantitatively by the observed adiabatic temperature change (ΔTad) in an adiabatic process and by the entropy change (ΔSiso) in an isothermal process.4 

MCE investigations and applications use typically polycrystalline materials exhibiting large ΔTad or ΔSiso.5–7 However, the magnetocrystalline anisotropy affects magnetic susceptibility and consequently also the magnetocaloric effect, as the magnetic response is different for field along an easy direction or along a hard direction and the overall MCE will be the powder average. In an ideal polycrystalline material, all crystallite orientations occur with identical probability, and the temperature-dependent magnetic susceptibility of an ideal powder can be calculated from the weighted average of magnetic susceptibilities of a single crystal as 1/3 of the value parallel to a certain axis and 2/3 of the value perpendicular to this axis. This relationship changes if there is a stronger tendency for the crystallites in a powder to be oriented more in certain directions (e.g., due to shape anisotropy), and, as a consequence, preferred orientation or texture arises. The presence or absence of preferred orientation should, thus, have a direct influence on overall magnetic susceptibility and might be detrimental or advantageous for the size of the MCE.

The influence of crystal field anisotropy on the MCE has been studied in the context of paramagnetic salts containing rare earth elements.8 Many of the candidate materials for applications close to room temperature in the vicinity of magnetic phase transition crystallize, however, in the hexagonal [e.g., materials related to Fe2P,9,10 La(Fe,Si)13,11,12 or MnAs13,14] or the tetragonal system (e.g., materials related to Mn2Sb15), where anisotropy is inherently important due to the presence of one symmetry-salient direction. Investigations of anisotropy effects in these materials are scarce due to the fact that it is difficult to obtain most of these magnetocaloric materials as single crystals.

We have now succeeded to grow single crystals of the room temperature magnetocaloric compound Mn5Ge3. The compound crystallizes in the hexagonal space group P63/mcm with lattice parameters a = 7.184(2) Å and c = 5.053(2) Å.16 Within the structure, Mn occupies two different Wyckoff positions [Fig. 1 (left)].16–18 The Mn1 atoms form an empty octahedron, while Mn2 is incorporated at the center of [Mn2Ge6]-octahedra.

FIG. 1.

(Left) Projection of the structure of Mn5Ge3 in space group P63/mcm along the [001] direction.16,36,37 Sites occupied by Mn1 are shown in gray (WP4d), and sites occupied by Mn2 are shown in pink (WP6g); Ge atoms are shown in blue (WP6g) and [MnGe6]-octahedra are indicated also in blue. (Right) Schematic diagram illustrating the ferromagnetic structure of Mn5Ge3, projection slightly tilted from the [110] direction. The length of the arrows corresponds to M = 1.96(3) μB and 3.23(2) μB for the WP4d and WP6g sites, respectively.

FIG. 1.

(Left) Projection of the structure of Mn5Ge3 in space group P63/mcm along the [001] direction.16,36,37 Sites occupied by Mn1 are shown in gray (WP4d), and sites occupied by Mn2 are shown in pink (WP6g); Ge atoms are shown in blue (WP6g) and [MnGe6]-octahedra are indicated also in blue. (Right) Schematic diagram illustrating the ferromagnetic structure of Mn5Ge3, projection slightly tilted from the [110] direction. The length of the arrows corresponds to M = 1.96(3) μB and 3.23(2) μB for the WP4d and WP6g sites, respectively.

Close modal
TABLE I.

Comparison between the main magnetic characteristics of Mn5Ge3 and MnFe4Si3 compounds.

Mn5Ge3MnFe4Si338,42
Easy axis c ([001]) Easy plane a,b 
Moment WP6g site: 3.23(2) μB28  WP6g site: 1.5(2) μB 
Moment WP4d site: 1.96(3)28  WP4d site: 1.1(12) μBa 
ΔSiso = 2.5 J/kg K (1 T ‖ [001]) ΔSiso = 1.3 J/kg K (1 T ‖ [100]) 
ΔSiso = 2.15 J/kg K (1 T ⊥ [001]) ΔSiso = 0.47 J/kg K (1 T ⊥ [100]) 
ΔTad = 2.3(1) K (2 T ‖ [001]) ΔTad = 1.38(1) K (2 T ‖ [100]) 
ΔTad = 2.0(1) K (2 T ⊥ [001]) … 
K1 = 3.7(1) × 105 J/m3 (60 K) K1 = − 1.5(1) × 106 J/m3 (60 K)b 
Mn5Ge3MnFe4Si338,42
Easy axis c ([001]) Easy plane a,b 
Moment WP6g site: 3.23(2) μB28  WP6g site: 1.5(2) μB 
Moment WP4d site: 1.96(3)28  WP4d site: 1.1(12) μBa 
ΔSiso = 2.5 J/kg K (1 T ‖ [001]) ΔSiso = 1.3 J/kg K (1 T ‖ [100]) 
ΔSiso = 2.15 J/kg K (1 T ⊥ [001]) ΔSiso = 0.47 J/kg K (1 T ⊥ [100]) 
ΔTad = 2.3(1) K (2 T ‖ [001]) ΔTad = 1.38(1) K (2 T ‖ [100]) 
ΔTad = 2.0(1) K (2 T ⊥ [001]) … 
K1 = 3.7(1) × 105 J/m3 (60 K) K1 = − 1.5(1) × 106 J/m3 (60 K)b 
a

The refined magnetic moment in the M2 sites is not larger than the corresponding standard deviation and, therefore, was not taken into account in the refinement.38 

b

In Figure 4, based on M(H) in Ref. 42 compared to −2.8 × 106 J/m3 at 50 K in Ref. 46, the difference is due to a slightly different Fe content (Mn∼0.86Fe∼4.24Si∼2.90 in Ref. 46).

Mn5Ge3 has a significant saturation magnetization Ms of 2.6(2) μB/Mn at 4 K19,20 and shows a presumably second order phase transition to a ferromagnetically ordered phase with observed Curie temperatures spreading from 290 to 304 K.21–27 The compound exhibits an easy axis anisotropy ‖[001], which was partly explained by magnetic dipolar interactions.23 

Spins on both Mn sites are aligned parallel to the hexagonal c-axis, yet magnetic moments on both sites differ significantly16,27 [Fig. 1 (right)]. The magnetic entropy change of Mn5Ge3 measured on a polycrystalline sample shows a maximum value of 3.8 J/kg K for a field change of 2 T.22 For samples in the shape of ribbons, a maximum value of 4.92 J/kg K was obtained for an external magnetic field of 3 T.21 

In this work, we study the magneto-caloric effect by means of indirect methods to determine ΔSiso and ΔTad in Mn5Ge3 single crystals following the procedures described in the literature.29–33 The direct measurements of the adiabatic temperature change (ΔTad) in pulsed magnetic fields will also be presented. Due to the short pulse duration, this technique provides nearly adiabatic conditions, and the experimental conditions are close to the ones present in real applications.34,35 In addition, they allow extracting information on the response time of the material and provide insight into the stability of a material when repeatedly exposed to a magnetic field. Taking advantage of the existence of large single crystals, we particularly focus our investigations on the elucidation of the direction dependence of the magnetocaloric effect in this compound. To this end, we also provide a detailed comparison to the closely related MnFe4Si3 that exhibits an easy plane anisotropy.

A single crystal of Mn5Ge3 was grown via the Czochralski method from a pre-synthesized polycrystalline material using stoichiometric amounts of the constituent elements (Mn 99.99% purity and Ge 99.9999% purity) according to the procedure described in Ref. 38. The final crystal (diameter ≈1 cm, height ≈4 cm) was oriented with a Laue camera, and individual samples for heat capacity, magnetization measurements, and the direct measurements of the magnetocaloric effect were cut by spark erosion. All samples were cut perpendicular to the hexagonal [001] and [100] crystallographic directions. X-ray powder diffraction on a ground piece of a single crystal confirmed the phase purity (Fig. S1 in the supplementary material).

The heat capacity data were collected using the thermal relaxation calorimeter of the PPMS Dynacool system in the temperature range from 2 to 395 K. For the measurements in the zero field, a low temperature apiezon N grease (cryogenic high vacuum grease; 2 < T < 230 K) and a high temperature H grease (silicone-free high temperature vacuum grease; 210 < T < 395 K) were used. For the measurements in 1 and 2 T, the sample was fixed on the platform using silver paint to prevent the sample from moving and to ensure a good thermal contact between the sample and the platform. Due to the magnetic torque exerted on the sample, the measurement with the field parallel to the hard direction ([100] direction) was restricted to a maximum of 1 T (250 < T < 395 K). For each measurement point, an addenda measurement was performed and subsequently subtracted.

Magnetization measurements were performed in static fields using a vibrating sample magnetometer (VSM) of Quantum Design PPMS. The external magnetic field was oriented perpendicular to the precut faces, and isothermal magnetization curves were measured in the field range of 9 T > μ0H > − 0.1 T with different sweep rates, starting always from the maximum field at each temperature between 20 and 380 K. A demagnetization factor of N = 0.3394 was applied.39 Temperature dependencies were derived from the isothermal magnetization measurements and compared to measurements under isofield conditions (B = 0.01 T and B = 0.5 T) for which the Maxwell relation

SB|T=MT|B

applies. From the data, M(T)B was extracted and the MCE was determined.

For the measurements with the field parallel to the hard [100] direction, we plot H/M vs M2 (Fig. S2 in the supplementary material) to calculate the anisotropy constants K1 and K2 using the method introduced by Sucksmith and Thompson.40 

Direct measurements of the magnetocaloric effect in a pulsed magnetic field were performed at Dresden High magnetic field Laboratory by using their home-built experimental setup.41 The measurements were performed with the field parallel to the [100] and [001] directions up to 20 T and following the procedures described in Ref. 42.

The analysis of Cp(T) at low temperature yields the electronic specific heat coefficient γ = 51.1(1) mJ/mol K and a Debye temperature of ΘD = 509 (1) K. The heat capacity data in the zero field show a well-developed λ-type peak at the magnetic Curie temperature (Fig. 2). In the applied magnetic field, the peak broadens and shifts toward higher temperatures. This observation corroborates the predominance of the ferromagnetic order (see Fig. S3 in the supplementary material for the 2 T measurement in the [001] direction). From the comparison, we can see that the application of a field of 1 T reduces the heat capacity around the transition temperature more if the field is applied‖[001].

FIG. 2.

Temperature-dependent heat capacity data measured at 0 and 1 T with the field parallel to [001] and at 1 T with the field parallel to [100] ([100] measurement—restricted to a max. field of 1 T and temperature range from 250 to 395 K due to the large magnetic torque exerted on the sample).

FIG. 2.

Temperature-dependent heat capacity data measured at 0 and 1 T with the field parallel to [001] and at 1 T with the field parallel to [100] ([100] measurement—restricted to a max. field of 1 T and temperature range from 250 to 395 K due to the large magnetic torque exerted on the sample).

Close modal

We have measured the magnetic response using the high temperature option of the PPMS up to 1000 K. Only in the temperature range T > 550 K, the susceptibility is proportional to 1/T and independent of the direction of the applied field indicating the Curie–Weiss behavior. A fit of the Curie–Weiss law in the region T > 800 K (e.g., Fig. S4 in the supplementary material) yields a Curie constant of C = 1.1(1) × 10−4 m3K/mol, Curie–Weiss temperature of 360(10) K, and effective magnetic moment per transition metal ion μeff = 3.8(2) μB, i.e., the ordered moment as reported in Ref. 28 on the WP6g site at base temperature comes close to the effective paramagnetic moment per Mn, while the moment on the WP4d site is significantly smaller.

The demagnetization corrected magnetization at different initial temperatures [Fig. 3; see also Fig. S5 in the supplementary material for further M(H) curves] shows that along the easy [001] direction, saturation is reached at small fields of <0.3 T, and only close to the transition temperature, the field dependence broadens.

FIG. 3.

Selected magnetization curves M(H) of Mn5Ge3 measured at different temperatures with the magnetic field applied along the [001] direction (solid lines) and along the [100] direction (dashed lines).

FIG. 3.

Selected magnetization curves M(H) of Mn5Ge3 measured at different temperatures with the magnetic field applied along the [001] direction (solid lines) and along the [100] direction (dashed lines).

Close modal

For the data with the field parallel to the [100] direction, the response at small fields is lower. In the temperature range between TC and 250 K, the slope dM/dH is increasing with temperature, see Fig. S6 in the supplementary material. Below 250 K, dM/dH remains nearly field and temperature independent below the anisotropy field Ha, which we identify as the locus of maximum curvature in the M(H) curves. Above Ha, the magnetization for the field parallel and perpendicular to the easy axis approaches each other.

For the hexagonal system, second (K1) and fourth order (K2) anisotropy constants are considered for anisotropy energy (Ea = K1sin2θ + K2sin4θ). θ is the angle between the field and the easy direction. We applied the method introduced by Sucksmith and Thompson40 to calculate K1 and K2 from a plot of M2 vs μ0H/M. Before saturation is reached, both observables have a linear relation and the slope yields K2, while the y-axis intersection yields K1 + 2K2, as can be seen from a free energy expansion. At 20 K, we find the magnetic anisotropy constant K1 = 3.6(1) × 105 J/m3 (anisotropy field ≈0.8 T), while K2 = 1.3(1) × 104 J/m3. With further temperature increase, the anisotropy slowly decreases until approaching the Curie temperature TC ≈ 296 K (Fig. 4). The results agree well with the ones in Ref. 23.

FIG. 4.

Temperature dependence of the magneto-crystalline anisotropy parameters for Mn5Ge3 and MnFe4Si3. The positive sign in K1 of Mn5Ge3 is due to having an easy axis anisotropy, meanwhile the negative sign for MnFe4Si3 is in line with an easy plane magnetic direction.

FIG. 4.

Temperature dependence of the magneto-crystalline anisotropy parameters for Mn5Ge3 and MnFe4Si3. The positive sign in K1 of Mn5Ge3 is due to having an easy axis anisotropy, meanwhile the negative sign for MnFe4Si3 is in line with an easy plane magnetic direction.

Close modal

Thermal hysteresis loops of Mn5Ge3 showing the magnetization as a function of temperature at applied magnetic fields of 0.01 and 0.5 T parallel to the [001] and [100] directions are presented in Fig. 5. The temperature-dependent magnetic response shows the hysteretic behavior of about 5 K along both directions [see the inset in Fig. 5(a)] independent of the field direction and strength. The comparison between M(T) at 0.5 T from the isofield measurements and the ones extracted from the isothermal magnetization measurements—without demagnetization correction, green curves in Fig. 5—justifies the calculation of ΔSiso from isothermal magnetization measurements, which is presented in Fig. 6 for different magnetic field changes ΔB.

FIG. 5.

Temperature-dependent magnetization of Mn5Ge3 at an applied field of 0.01 and 0.5 T in (a) [001] and (b) [100] directions. Lines are M(T) from isofield measurements and dots are M(T)B extracted from isothermal measurements. Inset shows the magnetic transition region at 0.01 T‖[001].

FIG. 5.

Temperature-dependent magnetization of Mn5Ge3 at an applied field of 0.01 and 0.5 T in (a) [001] and (b) [100] directions. Lines are M(T) from isofield measurements and dots are M(T)B extracted from isothermal measurements. Inset shows the magnetic transition region at 0.01 T‖[001].

Close modal
FIG. 6.

Magnetic entropy change of Mn5Ge3 determined from magnetization data at a field of 0.5, 2, and 3 T parallel to the [001] (black closed symbol) and [100] (red opened symbol) directions.

FIG. 6.

Magnetic entropy change of Mn5Ge3 determined from magnetization data at a field of 0.5, 2, and 3 T parallel to the [001] (black closed symbol) and [100] (red opened symbol) directions.

Close modal

ΔSiso features the maximum at 296 K for both field directions. The entropy change for the easy and hard direction differs by 0.4 J/kg K for all three ΔB (see Fig. 6, black closed symbols for μ0H‖[001] and red open symbols for μ0H‖[100]). The difference vanishes for temperatures sufficiently higher than the transition temperature. Below TC, the anisotropy of the effect is quite present, and it is more pronounced at lower fields. Below the transition temperature, the respective average of these values is consistent with the earlier results from a polycrystalline sample22,25 and the results from the sample in the form of ribbons—unfortunately, no information about the orientation of the ribbons is given in the article, so that a more detailed comparison regarding the anisotropy is not feasible.21 

When a small field of <1.2 T is applied along the [100] direction, a small inverse MCE is observed between 150 and 290 K, i.e., the entropy increases with an increasing field (blue area in the right panel of Fig. 7) due to the fact that the small magnetic field cannot overcome the anisotropy. With further increase in temperature, the anisotropy decreases and smaller fields are sufficient to align the moments with the field. As a consequence, the entropy changes the sign and increases stronger than linear with the field. A similar behavior was also observed in the compound MnFe4Si3, which is isostructural to Mn5Ge3, yet exhibits mixed occupancy of Mn and Fe on the WP6g site and a ferromagnetic structure with the spins aligned in the a,b-plane.42 

FIG. 7.

Color plot of the magnetic entropy change of Mn5Ge3 as a function of temperature and magnetic field change parallel to [001] and [100]. Blue colors indicate a positive ΔSiso and hence an inverse MCE; red colors correspond to a negative ΔSiso and hence normal MCE.

FIG. 7.

Color plot of the magnetic entropy change of Mn5Ge3 as a function of temperature and magnetic field change parallel to [001] and [100]. Blue colors indicate a positive ΔSiso and hence an inverse MCE; red colors correspond to a negative ΔSiso and hence normal MCE.

Close modal

To probe the applicability of the material on the time scale close to that of possible applications, we performed direct measurements of ΔTad in the pulsed fields. For that, we record the temperature of the sample as the field is ramped in ∼50 ms to 2 T and 20 T.

The ΔTad values are obtained by processing the signal from the thermocouple, one joint of which is connected to the sample. A lot of effort is put in order to minimize the thermalization time and to cancel the contributions induced in the leads by the time dependent variation of the magnetic field and sample magnetization. Figure 8 presents the as-recorded data measured for the two directions—easy and hard directions—at 297.5 K in 2 T pulses (see Fig. S7 in the supplementary material for the data at 300 K). The upper graphs show the time dependencies of the magnetic field and the thermocouple response. Here, the thermocouple signal resembles the pulse profile rather tightly, indicating a reasonably good coupling to the sample. More thorough test of the coupling is the field dependence of ΔTad. The lower graphs show the temperature changes replotted against the field. Here, the finite response time shows up as an opening of the curve upon up- and down-sweeps. The difference in the shape of the field dependencies between the two directions reflect the thermal contact quality between the particular directions, which is also seen by the lagging of the temperature signal behind the field signal. A similar behavior is seen for 20 T pulses shown in Fig. S8 in the supplementary material.

FIG. 8.

Field and time dependence of ΔTad for a pulsed magnetic field of 2 T applied along the [001] direction (a) and along the [100] direction (b) at 297.5 K.

FIG. 8.

Field and time dependence of ΔTad for a pulsed magnetic field of 2 T applied along the [001] direction (a) and along the [100] direction (b) at 297.5 K.

Close modal

The adiabatic temperature change reaches different maximum values along the easy and the hard directions [Fig. 9(a)]; ≈2.3(1) K for H‖[001] at 295 K as compared to ≈2.0(1) K for H‖[100] at 300 K in pulsed magnetic fields of 2 T. These values are in agreement with the values calculated from the isothermal entropy change and the specific heat measurements (details in Fig. S9 in the supplementary material). The observations show that (i) this material is quite capable of transferring heat on the time scale of about 15 ms and (ii) that at least at the top of the pulse corresponding to the maximum values, the data are quite reliable. With that stated, we have extended the field range up to 20 T [Fig. 9(b); Fig. S8 in the supplementary material] where the the observed peak broadens and shifts to higher temperatures, ΔTad differs also by ∼10% for the two directions; ≈10.8(2) K for H‖[001] at 305 K as compared to ≈9.8(4) K for H‖[100] at 310 K. ΔTad varies roughly as H2/3, as expected for localized ferromagnetism at TC.43 

FIG. 9.

Comparison of ΔTad measured in the pulsed magnetic fields of 2 T (a) and 20 T (b) with the field parallel to [001] and [100]. The lines drawn in the figures are just to guide the eyes. Comparison of ΔTad measured in pulsed magnetic fields and calculated from heat capacity in static magnetic fields of 2 T is given in the supplement (Fig. S9 in the supplementary material).

FIG. 9.

Comparison of ΔTad measured in the pulsed magnetic fields of 2 T (a) and 20 T (b) with the field parallel to [001] and [100]. The lines drawn in the figures are just to guide the eyes. Comparison of ΔTad measured in pulsed magnetic fields and calculated from heat capacity in static magnetic fields of 2 T is given in the supplement (Fig. S9 in the supplementary material).

Close modal

It is intriguing to compare Mn5Ge3 with isostructural Mn5Si344,45 that has also been widely studied for its magnetic and magnetocaloric properties. However, Mn5Si3 orders antiferromagnetically at much lower temperature and exhibits an inverse MCE related to a phase transition between two antiferromagnetic structures. Yet, the isostructural compound MnFe4Si3 where the WP4d site is mainly occupied by Fe while the WP6g site has a mixed occupancy of Mn and Fe provides an ideal point of comparison: it orders ferromagnetically with a TC close to 300 K and a normal MCE.38,42 The main magnetic characteristics of both compounds are given in Table I.

In difference to Mn5Ge3 where the spins are aligned along c ([001]), the spins in MnFe4Si3 are aligned in the a,b plane.38,42 In both compounds, the magnetic moment on the WP6g site is larger than that on the WP4d site; in MnFe4Si3, it was even not possible to refine any ordered moment for the WP4d site2.

Mn5Ge3 has small magnetic anisotropy [K1 = 3.7(1) × 105 J/m3 at 60 K] with c being the easy axis, while MnFe4Si3 shows a much larger magnetic anisotropy (see Fig. 4). Provided an ideal powder is used in applications, the large anisotropy of MnFe4Si3 could, therefore, limit the size of the MCE in this compound (and other compounds with similar characteristics) when compared to the MCE in a single crystal, while for Mn5Ge3 (and similar compounds), the reduction of the MCE will be comparatively weak. On the other hand, the targeted introduction of the preferred orientation might also be beneficial for increasing the MCE in materials with large anisotropy and might represent a new approach for optimizing their performance.

The anisotropy of the magnetocaloric properties in Mn5Ge3 was studied in static and pulsed magnetic fields. The uniaxial magnetic anisotropy decreases with temperature and can be overcome by applied fields μ0H > 0.8 T; the anisotropy constants are calculated over a broad temperature range up to second order. The comparison with MnFe4Si3, which exhibits an easy plane anisotropy, shows that in Mn5Ge3, the dependence of the size of the MCE on the field direction is less pronounced. However, despite the fact that anisotropy constants vanish toward TC, the MCE in Mn5Ge3 features also a significant anisotropy that is seen in the adiabatic temperature change in the pulsed field and also in the isothermal entropy change.

This study suggests that the magnetic anisotropy should be taken into account when trying to optimize the performance of magnetocaloric materials. In applications, the control of the preferred orientation and texture, depending on the specific anisotropic characteristics of the candidate materials, could be beneficial for increasing the size of the magnetocaloric effect.

See the supplementary material for more information about the anisotropy calculation, heat capacity measurements, Curie Weiss fit, and isothermal magnetization curves with dM/dH calculation and for more details about the pulsed field measurements with a comparison of the results (ΔTad) with the ones obtained from the static field measurements.

This work was part of a collaborative agreement between the Forschungszentrum Jülich and Al-Quds University and was supported by the BMBF under the programme Zusammenarbeit mit Entwicklungs- und Schwellenländern im Nahen Osten, Nordafrika, Türkei (Project MagCal, No. 01DH17013) and under the Joint Research and Education Programme Palestinian German Science Bridge PGSB. The work is based upon experiments performed at HLD at HZDR, member of the European Magnetic Field Laboratory (EMFL). We would like to thank Lukas Berners for the XRD measurements.

The data that support the findings of this study are available from the corresponding author upon reasonable request.

1.
K. A.
Gschneidner
, Jr.
, and
V. K.
Pecharsky
,
Int. J. Refrig.
31
,
945
(
2008
).
2.
C. B.
Zimm
,
A.
Jastrab
,
A.
Sternberg
,
V.
Pecharsky
,
K. A.
Geschneidner
, Jr.
,
M.
Osborne
, and
I.
Anderson
,
Adv. Cryog. Eng.
43
, 1759 (
1998
).
3.
B. F.
Yu
,
Q.
Gao
,
B.
Zhang
,
X. Z.
Meng
, and
Z.
Chen
,
Int. J. Refrig.
26
,
622
(
2003
).
4.
A.
Kitanovski
and
P. W.
Egolf
,
Int. J. Refrig.
29
,
3
(
2006
).
5.
C. F.
Sánchez-Valdés
,
R. R.
Gimaev
,
M.
López-Cruz
,
J. L.
Sánchez Llamazares
,
V. I.
Zverev
,
A. M.
Tishinc
,
A. M. G.
Carvalho
,
D. J. M.
Aguiar
,
Y.
Mudryk
, and
V. K.
Pecharsky
,
J. Magn. Magn. Mater.
498
,
166130
(
2020
).
6.
A. O.
Guimarães
,
M. E.
Soffner
,
M. A.
Mansanares
,
A. A.
Coelho
,
A.
Magnus
,
G.
Carvalho
,
M. J. M.
Pires
,
S.
Gama
, and
E. C.
da Silva
,
J. Appl. Phys.
107
,
073524
(
2010
).
7.
K.
Fukamichi
,
A.
Fujita
, and
S.
Fujieda
,
J. Alloys Compds.
408-412
,
307
(
2006
).
8.
M. D.
Kuz'min
and
A. M.
Tishin
,
J. Phys. D Appl. Phys.
24
,
2039
(
1991
).
9.
N. T.
Trung
,
Z. Q.
Ou
,
T. J.
Gortenmulder
,
O.
Tegus
,
K. H. J.
Buschow
, and
E.
Brück
,
Appl. Phys. Lett.
94
,
102513
(
2009
).
10.
O.
Tegus
,
E.
Brück
,
K. H. J.
Buschow
, and
F. R.
de Boer
,
Nature (London)
415
,
150
(
2002
).
11.
F. X.
Hu
,
B. G.
Shen
,
J. R.
Sun
,
Z. H.
Cheng
,
G. H.
Rao
, and
X. X.
Zhang
,
Appl. Phys. Lett.
78
,
3675
(
2001
).
12.
A.
Fujita
,
S.
Fujieda
,
Y.
Hasegawa
, and
K.
Fukamichi
,
Phys. Rev. B
67
,
104416
(
2003
).
13.
H.
Wada
and
Y.
Tanabe
,
Appl. Phys. Lett.
79
,
3302
(
2001
).
14.
P. F.
Xu
,
S. H.
Nie
,
K. K.
Meng
,
S. L.
Wang
,
L.
Chen
, and
J. H.
Zhao
,
Appl. Phys. Lett.
97
,
042502
(
2010
).
15.
S. C.
Ma
,
D.
Hou
,
Y. Y.
Gong
,
L. Y.
Wang
,
Y. L.
Huang
,
Z. C.
Zhong
,
D. H.
Wang
, and
Y. W.
Du
,
Appl. Phys. Lett.
104
,
022410
(
2014
).
16.
R.
Ciszewski
,
Phys. Status Solidi B
3
,
1999
(
1963
).
17.
G.
Kappel
,
G.
Fischer
, and
A.
Jaegle
,
Phys. Status Solidi A
34
,
691
(
1976
).
18.
D. H.
Day
and
R. N.
Sinclam
,
Acta Crystallogr. B
26
,
2079
(
1970
).
19.
G.
Kappel
,
G.
Fischer
, and
A.
Jaegle
,
Phys. Lett.
45
,
267
(
1973
).
20.
X. B.
Liu
and
Z.
Altounian
,
J. Appl. Phys.
99
,
08Q101
(
2006
).
21.
T. F.
Zheng
,
Y. G.
Shi
,
C. C.
Hu
,
J. Y.
Fan
,
D. N.
Shi
,
S. L.
Tang
, and
Y. W.
Du
,
J. Magn. Magn. Mater.
324
,
4102
(
2012
).
22.
D.
Songlin
,
O.
Tegus
,
E.
Brück
,
F. R.
de Boer
, and
K. H. J.
Buschow
,
J. Alloys Compd.
337
,
269
(
2002
).
23.
Y.
Tawara
and
K.
Sato
,
J. Phys. Soc. Jpn.
18
,
773
(
1963
).
24.
F. Q.
Zhao
,
W.
Dagula
,
O.
Tegus
, and
K. H. J.
Buschow
,
J. Alloys Compd.
416
,
43
(
2006
).
25.
T.
Tolinski
and
K.
Synoradzki
,
Intermetallics
47
,
1
(
2014
).
26.
K. H.
Kang
,
J. H.
Kim
,
J. W.
Kim
,
K. C.
Chung
, and
C. S.
Yoon
,
J. Alloys Compd.
729
,
603
(
2017
).
27.
Y.
Kim
,
K. H.
Kang
,
J. H.
Kim
,
E. J.
Kim
,
K.
Choi
,
W. B.
Han
,
H. S.
Kim
,
Y.
Oh
, and
C. S.
Yoon
,
J. Alloys Compd.
644
,
464
(
2015
).
28.
J. B.
Forsytht
and
P. J.
Brown
,
J. Phys. Condens. Matter
2
,
2713
(
1990
).
29.
V. K.
Pecharsky
and
K. A.
Gschneidner
, Jr.
,
J. Appl. Phys.
85
,
5365
(
1999
).
30.
M.
Foldeaki
,
R.
Chahine
, and
T. K.
Bose
,
J. Appl. Phys.
77
,
3528
(
1995
).
31.
K. A.
Gschneidner
, Jr.
, and
V. K.
Pecharsky
,
Annu. Rev. Mater. Sci.
30
,
387
(
2000
).
32.
V. K.
Pecharsky
and
K. A.
Gschneidner,
 Jr.
,
J. Appl. Phys.
86
,
565
(
1999
).
33.
V. K.
Pecharsky
,
K. A.
Gschneidner,
 Jr.
,
A. O.
Pecharsky
, and
A. M.
Tishin
,
Phys. Rev. B
64
,
144406
(
2001
).
34.
J.
Liu
,
T.
Gottschall
,
K. P.
Skokov
,
J. D.
Moore
, and
O.
Gutfleisch
,
Nat. Mater.
11
,
620
(
2012
).
35.
F.
Cugini
,
G.
Porcari
,
C.
Viappiani
,
L.
Caron
,
A. O.
dos Santos
,
L. P.
Cardoso
,
E. C.
Passamani
,
J. R. C.
Proveti
,
S.
Gama
,
E.
Brück
, and
M.
Solzi
,
Appl. Phys. Lett.
108
,
012407
(
2016
).
36.
C.
Salazar-Mejía
,
V.
Kumar
,
C.
Felser
,
Y.
Skourski
,
J.
Wosnitza
, and
A. K.
Nayak
,
Phys. Rev. Appl.
11
,
054006
(
2019
).
37.
P.
Devi
,
M.
Ghorbani Zavareh
,
C.
Salazar Mejía
,
K.
Hofmann
,
B.
Albert
,
C.
Felser
,
M.
Nicklas
, and
S.
Singh
,
Phys. Rev. Mater.
2
,
122401
(
2018
).
38.
P.
Hering
,
K.
Friese
,
J.
Voigt
,
J.
Perßon
,
N.
Aliouane
,
A.
Grzechnik
,
A.
Senyshyn
, and
T.
Brückel
,
Chem. Mater.
27
,
7128
(
2015
).
39.
D. X.
Chen
,
E.
Pardo
, and
A.
Sanchez
,
IEEE Trans. Magn.
38
,
4
(
2002
).
40.
W.
Sucksmith
and
J. E.
Thompson
,
Proc. R. Soc. A
225
,
362
(
1954
).
41.
M.
Ghorbani Zavareh
,
C.
Salazar Mejía
,
A. K.
Nayak
,
Y.
Skourski
,
J.
Wosnitza
,
C.
Felser
, and
M.
Nicklas
,
Appl. Phys. Lett.
106
,
071904
(
2015
).
42.
N.
Maraytta
,
Y.
Skourski
,
J.
Voigt
,
K.
Friese
,
M. G.
Herrmann
,
J.
Perßon
,
J.
Wosnitza
,
S. M.
Salman
, and
T.
Brückel
,
J. Alloys Compd.
805
,
1161
(
2019
).
43.
H.
Oesterreicher
and
F. T.
Parker
,
J. Appl. Phys.
55
,
4334
(
1984
).
44.
M.
Gottschilch
,
O.
Gourdon
,
J.
Perßon
,
C.
de la Cruz
,
V.
Petricek
, and
T.
Brückel
,
J. Mater. Chem.
22
,
15275
(
2012
).
45.
G.
Kappel
,
G.
Fischer
, and
A.
Jaegle
,
Phys. Status Solidi B
34
(
2
),
691
696
(
1976
).
46.
H.
Yibolea
,
W.
Hanggaia
,
Z. Q.
Oua
,
R.
Hamaneb
,
V.
Hardyb
, and
F.
Guilloua
,
J. Magn. Magn. Mater
504
,
166597
(
2020
).

Supplementary Material