# Fileset

[Manuscript-final submission.docx](https://mdr.nims.go.jp/filesets/1211cd02-52b4-4466-b76b-b332fde86375/download)

## Creator

[Jiawei Lai](https://orcid.org/0000-0002-1351-4268), [Xin Tang](https://orcid.org/0000-0001-6762-6145), [Hossein Sepehri-Amin](https://orcid.org/0000-0002-7856-7897), [Kazuhiro Hono](https://orcid.org/0000-0001-7367-0193)

## Rights

[Creative Commons BY Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)

## Other metadata

[Tuning magnetocaloric effect of Ho1-Gd Ni2 and HoNi2-Co  alloys around hydrogen liquefaction temperature](https://mdr.nims.go.jp/datasets/14ab5db5-6a39-466e-a845-e347137d571d)

## Fulltext

Tuning magnetocaloric effect of Ho1-xGdxNi2 and HoNi2-yCoy alloys around hydrogen liquefaction temperatureJiawei Lai, Xin Tang, Hossein Sepehri-Amin*, Kazuhiro HonoResearch Center for Magnetic and Spintronic Materials, National Institute for Materials Science, 1-2-1 Sengen, Tsukuba, 305-0047, JapanWe show the giant magnetocaloric effect of HoNi2 compound can be tuned around hydrogen liquefaction temperature by substituting Ho with Gd and Ni with Co in Ho1-xGdxNi2-yCoy. While the Curie temperature of Ho1-xGdxNi2-yCoy can be tuned from 15 to 32 K originating from the lattice expansion of the cubic structure, giant entropy change of 22.0 J/kgK is retained near the hydrogen liquefaction temperature. A table-like entropy change observed in a broad temperature range of 20 to 32 K makes them promising as magnetocaloric materials for cryogenic magnetic refrigeration utilizing an Ericsson cycle. Keywords: Table-like magnetocaloric effect, HoNi2 alloys, Curie temperature, Hydrogen liquefaction* Corresponding author: H. Sepehri-Amin (Email address: h.sepehriamin@nims.go.jp ).Cryogenic magnetic refrigeration (MR) using magnetocaloric materials with a giant magnetocaloric effect (MCE) is considered to be promising for a hydrogen liquefaction technology because of its high efficiency and environment friendly features [1, 2]. When applying a magnetic field, a large isothermal magnetic entropy change (|ΔSM|) or adiabatic change of temperature (ΔTad) is induced in the magnetocaloric materials. Giant MCE at the cryogenic temperatures are mainly found in heavy rare-earth-based (RE) alloys such as Gd5(Ge, Si)4 [3], REAl2(RE = Dy, Ho, Er) [4], and RENi2 alloys [5, 6]. Among these, the HoNi2 compound with the MgCu2-type cubic structure shows a large value of |ΔSM| of 26 J/kgK at 14 K due to the existence of crystalline electric field between the RE and 3d-metal [5]. A second order magnetic phase transition occurs in the HoNi2 compound, which provides an advantage of no volume change and no hysteresis [7]. Moreover, rather low melting point (1558K) of HoNi2 makes it feasible to manufacture in a large scale [8]. Hence, HoNi2 alloy has a great potential to be used for magnetic refrigeration for hydrogen liquefaction. However, its ferromagnetic/paramagnetic phase transition temperature, i.e., Curie temperature (TC), is lower than the hydrogen liquefaction temperature (20 K). Thus, tuning TC to above 20 K while maintaining its giant MCE is needed for cryogenic magnetic refrigeration for hydrogen liquefaction. The previous investigations on Fe-doping [9] and Co-doping [10] to HoNi2 have shown the possibility of tuning TC to higher than 20 K. However, the increase of TC is accompanied by the large reduction of |ΔSM| from 26 J/kgK for HoNi2 to 18.9 J/kgK for HoNi1.75Co0.25, suggesting there is room for |ΔSM| to be improved [10]. Hence, tuning TC of the HoNi2 compound to higher than 20 K with minimum reduction of |ΔSM| is required. In addition, finding material that shows a table-like |ΔSM| is of particular interest for cryogenic Ericsson cycle refrigeration, for which a constant change in magnetic entropy within the region of cooling is required [11, 12]. A table-like |ΔSM| of 12 J/kgK from 20 to 30 K had been reported for Gd0.5Er0.5NiAl alloy [13]. Hence, the second aim is to obtain a table-like ΔSM in a broad temperature range by tuning the chemical composition of the HoNi2 compound. Herein, we systematically investigated the effect of Gd and Co in Ho1-xGdxNi2-yCoy with the main motivation of increasing TC to above 20K while maintaining the giant MCE and the second motivation of achieving a table-like |ΔSM|. Polycrystalline Ho1-xGdxNi2-yCoy (x = 0, 0.1, 0.2, and 0.3; y = 0, 0.1, 0.2, and 0.3) alloys were prepared by arc melting. In order to compensate the evaporation of Ho during the melting process, 5 wt.% of excess Ho was added to the raw materials. The crushed ingots were sealed in quartz ampoules under 80 KPa of Ar. In the first batch, the Ho1-xGdxNi2-yCoy alloys were annealed at 1200 K for 3 days. According to the results of TG-DTA, we found that the samples with y = 0.2 and y = 0.3 needed a higher annealing temperature for homogenization. Thus, the samples y = 0.2 and 0.3 were annealed at 1273 K for 1 day. Microstructures were studied by a SEM using the Carl Zeiss CrossBeam1540EsB and a scanning transmission electron microscope (STEM) (Titan G2 80–200 System) with a probe aberration corrector. Specimens for STEM observations were prepared using a FEI Helios G4 dual beam machine using the standard lift-out process. Powder X-ray diffraction with Cr Kα radiation was conducted in the 0.6 kW Rigaku MiniFlex X-ray diffractometer. The temperature and magnetic field dependence of the magnetization was measured by a superconducting quantum interference device vibrating sample magnetometer, SQUID-VSM (Quantum Design MPMS SQUID VSM).Magnetization as a function of temperature for HoNi2-yCoy alloys under 1 T is shown in figure 1 (a). The magnetization under 1 T at 5 K increases from 146.2 to 159.2 A/m2kg-1 when increasing y from 0 to 0.3. An increase of magnetization is observed by Co substitution for Ni, indicating a ferromagnetic moment of Co between Ho is developed in the HoNi2-yCoy alloys. This result is in agreement with the report by Mudryk et al. on the effect of Co on RECo2 (RE = Ho, Er, Gd) alloys [14]. Magnetization as a function of temperature for Ho1-xGdxNi2 alloys under 1 T is shown in figure 1 (b). By substituting Gd for Ho in HoNi2, the magnetization is increased, which is similar to the effect of Co. This is because the spin component of Ni 3d electrons couples antiparallel to that of Gd 4f electrons and it contributes to an angular momentum of 0.23 μB in the GdNi2 alloy [15]. Thus, introducing Gd-Ni magnetic coupling in the HoNi2 alloy is expected to increase the magnetization. ZFC–FC curves and the corresponding dM/dT values as a function of temperature under 0.05 T for Ho1-xGdxNi2-yCoy alloys are shown in the figure 1 (c) and (d). The values of TC in this work are determined by the maximum peak of dM/dT under 0.05 T, as shown in figure 1 (d). The values of TC are listed in Table 1. By increasing Gd substitution x to 0.1 or Co substitution y to 0.2, TC increases to above 20 K. The increase of TC suggests that a stronger exchange interaction of magnetic atoms is developed in the Ho1-xGdxNi2-yCoy alloys. This agrees with the higher magnetization observed in the Gd contained and Co contained alloys, as shown in figures 1 (a) and (b).Table 1: List of TC, ∆SM (J/kgK) at 2 and 5 T applied magnetic field, RC and ∆SM (J/cm3K) at 5 T for the Ho1-xGdxNi2 and HoNi2-yCoy alloys, and compared with that of ErZn/ErZn2 [16] and ErFeSi [17]. Materials TC(K) ∆SM (2T)(J/kgK) ∆SM (5T)(J/kgK) RC (5T)(J/kg) ∆SM (5T) (J/cm3K) Ref. HoNi2 15 14.5 25.9 402 0.27 This work HoCo0.1Ni1.9 18 12.9 24.0 435 0.25 This work HoCo0.2Ni1.8 22 10.5 20.6 453 0.21 This work HoCo0.3Ni1.7 27 8.9 18.6 488 0.19 This work Ho0.9Gd0.1Ni2 21 11.2 22.0 455 0.23 This work Ho0.8Gd0.2Ni2 27 9.1 18.5 465 0.19 This work Ho0.7Gd0.3Ni2 32 8.0 16.8 495 0.17 This work ErZn/ErZn2 9/20 8.1 19.5 365 0.17 [16] ErFeSi 22 14.2 22.5 362 0.18 [17]Figures 2 (a) and (b) show SEM-EDS maps and composition line profiles of constituent elements obtained from Ho0.7Gd0.3Ni2 and HoNi1.7Co0.3 samples, respectively. Microstructures show uniform distribution of Gd and Co in the HoNi2 parent phase in both samples without any segregation. Small amount of HoNi and Ho3Ni2 impurity phases are also observed as indicated in figure 2 (b). Composition line profiles obtained from matrix phase show that Gd with an average content of 10 at. % is substituted for Ho in the Ho0.7Gd0.3Ni2 sample. Co with an average content of ~11 at.% is substituted for Ni in the matrix phase. Fig. 2 (c) shows the high resolution high angle annular dark field (HAADF)-STEM image observed along  zone axis of a HoNi1.7Co0.3 grain. Higher magnification HAADF-STEM images and atomic resolution STEM-EDS maps of Ho (red), Co (blue) and Ni (green) atoms in the crystal are shown in Fig. 2 (c), confirming that Co atoms exist in the Ni atoms sites in the crystal. HoNi2 show cubic C15 crystal structure and the Curie temperature depends on the interatomic distance in the crystal. Hence, substitution of Co for Ni and Gd for Ho can change the lattice parameter of the crystal and accordingly the interatomic exchange in HoNi2 based phase. Figure 3 (a) shows the XRD patterns obtained from Ho1-xGdxNi2 samples. The main phase is found to be MgCu2-type Laves phase (cubic with space group Fdm3) and a very minor impurity phases of HoNi (‎orthorhombic with space group Pnma) and Ho3Ni2 (‎monoclinic with space group C2/m) are also observed in the microstructure. Similar result is also obtained in the HoNi2-yCoy samples (data not shown). Based on the binary phase diagram of Ho-Ni [8], HoNi2, HoNi and a liquid phase occur when cooling down from the stoichiometry composition. The HoNi phase will transform to the Ho3Ni2 phase at the temperature range of 1033 to 1233 K. Therefore, the phases shown in our XRD results are consistent with the reported binary phase diagram. The Rietveld refinements are calculated using Fullprof software, as shown in the figure 3 (b). The volume fraction of the main phase is 92.5, 92.2, 93.3, 95.4 % for y = 0, 0.1, 0.2, 0.3 of HoNi2-yCoy alloys annealed at 1273 K while it is 93.2, 93.3, 93.6, 92.8 % for x = 0, 0.1, 0.2, 0.3 of Ho1-xGdxNi2 alloys annealed at 1200 K. Figure 3 (c) shows measured lattice constant a of the cubic MgCu2-type phase from XRD results for Ho1-xGdxNi2 and HoNi2-yCoy samples. By increasing Gd from 0 to 0.3, the value of a increases from 7.1407 to 7.1586 Å by 0.25%. When increasing Co from 0 to 0.3 in the HoNi2-yCoy alloys, the value of a increases from 7.1335 to 7.1471 Å by 0.19%. Replacing even smaller atoms of Gd and Co for Ho and Ni atoms increases the volume of the unit cell in the MgCu2-type structure. Isothermal magnetic entropy change, |ΔSM|, is calculated by using Maxwell equation from the measured data of iso-field magnetizations as a function of cooling temperature under a magnetic field of 0.5-5.0 T by a field step of 0.5 T. |ΔSM| measured under a magnetic field of 5 T is 25.9, 24.0, 20.6, and 18.6 J/kgK for HoNi2-yCoy (y = 0, 0.1, 0.2, 0.3) alloys, respectively. |ΔSM| is measured as 22.0, 18.5 and 16.8 J/kgK for Ho1-xGdxNi2 (x = 0.1, 0.2, 0.3) alloys, respectively. By tuning TC to 20 K, we obtained a similar value of |ΔSM| in the Ho0.9Gd0.1Ni2 alloy (|ΔSM|=22.0 J/kgK, TC = 21 K) and the HoNi1.8Co0.2 alloy (|ΔSM|=20.6 J/kgK, TC = 22 K). Obtained giant MCE close to 20K in these alloys shows that these materials are promising candidates for magnetic refrigeration application for H2 liquefaction compared with the other existing alloys such as Er5Pd2 (|ΔSM| = 13.5 J/kgK, TC = 20 K) [18], ErFeSi (|ΔSM|= 23.1 J/kgK, TC = 20 K) [17] , and DyNi2 (|ΔSM| = 21.3 J/kgK, TC = 20.3 K) [5, 6] as shown in figure 4 (c). It was found that in the Ho0.7Gd0.3Ni2 alloy, the |ΔSM| -T curves under 1 to 5 T show a rather broad peak at a similar height around TC i.e. a table-like behavior of |ΔSM|, as shown in the figure 4 (b). The table-like |ΔSM| of 16.6 J/kgK from 23 to 32 K is obtained in the Ho0.7Gd0.3Ni2 alloy. In addition, a table-like |ΔSM| of 18.5 J/kgK from 20 to 25 K is also shown in the Ho0.8Gd0.2Ni2 alloy, as plotted in the orange curve of figure 4 (a). Table-like |ΔSM| was previously observed in the (Gd1-xErx)NiAl alloys, which possessed multiple transition from heat capacity measurement [13]. In the Gd(1-x)ErxNiAl system with multiple magnetic ordering process, the origin of table-like |ΔSM| was attributed to the frustrated spin lattice in their hexagonal crystal structure and the competing effects of exchange and CEF interactions [13]. In this work, the underlying physics for the observed table-like behavior of |ΔSM| in the cubic Ho1-xGdxNi2 alloys is still an open question which needs to be clarified in future.Since the gap generated by the superconducting magnets is limited, the density of material determines the total amount of materials that can be used in the MR system. Thus, |ΔSM| values need to be compared in the unit of J/(cm3.K) for practical application in the cryogenic magnetic refrigeration systems. Figure 4 (d) shows |ΔSM| values measured under 5 T in the unit of J/(cm3.K) for Ho1-xGdxNi2-yCoy alloys developed in this work and compared with the other promising candidates in the temperature range of 14 to 34 K. The density values are the theoretical densities obtained from the Pearson database: HoNi2 (10.3 g/cm3), HoN (10.3 g/cm3) [19], ErZn2/ErZn (8.6 g/cm3) [9], Er5Pd2 (10.0 g/cm3) [18], DyNi2 (10.1 g/cm3) [5], ErFeSi (8.0 g/cm3) [17], GdCo2B2 (8.0 g/cm3) [20], DyCuAl (7.4 g/cm3) [21], HoAl2 (6.1 g/cm3) [4]. The values of |ΔSM| for Ho1-xGdxNi2 and HoNi2-yCoy alloys are shown in table 1. Due to the high density of HoNi2, the Ho1-xGdxNi2 and HoNi2-yCoy alloys, a superior performance to the previously reported alloys in the range of 20 to 32 K is observed making these materials excellent candidates for the cryogenic MR of hydrogen liquefaction. Another criterion to evaluate refrigeration efficiency is the refrigeration capacity (RC) [22] which represents the amount of heat transferred during one thermodynamic cycle. By introducing Gd and Co atoms in the HoNi2 alloy, the value of |ΔSM| is reduced while the sharpness of |ΔSM| is broadened. Thus, it increases the value of RC and enhances the refrigeration efficiency. The value of RC here is calculated by integrating the |ΔSM|-T curves between Thot and Tcold, where Thot and Tcold correspond to the temperature range at which the |ΔSM| value is half of the peak value. The calculated values of RC at 5 T for Ho1-xGdxNi2-yCoy are shown in Table 1. The value of RC for HoNi2 alloy is 405 J/kg. By 10 at% of Gd and Co substitution, the RC are enhanced to 495 and 488 J/kg by 23% and 20%, which are higher than ErFeSi (362 J/kg) [17] and Ho0.4Er0.6Ni (418 J/kg). [23]   In conclusion, TC of the Ho1-xGdxNi2-yCoy alloys is successfully tuned to the temperatures above 20K by the substitution of Gd for Ho and the substitution of Co for Ni. The TC increase is ascribed to an expansion of the lattice constant of the crystal originating from the substation of Gd atoms and Co atoms for Ho and Ni atoms, respectively, in the MgCu2-type structure. A giant MCE of 22.0 J/kgK is obtained in the Ho0.9Gd0.1Ni2 sample at 21 K and 20.6 J/kgK is obtained in the HoNi1.8Co0.2 sample at 22 K, making them excellent candidates for the cryogenic MR of hydrogen liquefaction. The table-like |ΔSM| of 18.5 and 16.8 J/kgK for the Ho0.8Gd0.2Ni2 and Ho0.7Gd0.3Ni2 alloys are another remarkable finding in this work raising the merits of these materials for cryogenic MR application utilizing an Ericsson cycle. AcknowledgementThis work was supported by the JST-Mirai Program Grant Number JPMJMI18A3, Japan.References[1] V. Franco, J.S. Blázquez, J.J. Ipus, J.Y. Law, L.M. Moreno-Ramírez, A. Conde, Progress in Materials Science 93 (2018) 112-232.[2] K.A. GschneidnerJr, V.K. Pecharsky, A.O. Tsokol, Reports on Progress in Physics 68(6) (2005) 1479-1539.[3] A.O. Pecharsky, K.A. Gschneidner, V.K. Pecharsky, Journal of Magnetism and Magnetic Materials 267(1) (2003) 60-68.[4] J.C.B. Monteiro, F.G. Gandra, Journal of Applied Physics 121(21) (2017) 213904.[5] P.J. von Ranke, D.F. Grangeia, A. Caldas, N.A. de Oliveira, Journal of Applied Physics 93(7) (2003) 4055-4059.[6] P.J. von Ranke, V.K. Pecharsky, K.A. Gschneidner, Physical Review B 58(18) (1998) 12110-12116.[7] J. Ćwik, Y. Koshkid'ko, K. Nenkov, E.A. Tereshina, K. Rogacki, Journal of Alloys and Compounds 735 (2018) 1088-1095.[8] Z. Huaiying, O. Xiangli, Z. Xiaping, Journal of Alloys and Compounds 177(1) (1991) 101-106.[9] N.K. Singh, S. Agarwal, K.G. Suresh, R. Nirmala, A.K. Nigam, S.K. Malik, Physical Review B 72(1) (2005) 014452.[10] R. Mondal, R. Nirmala, J. Arout Chelvane, S.K. Malik, Journal of Magnetism and Magnetic Materials 393 (2015) 376-379.[11] J.W. Lai, Z.G. Zheng, X.C. Zhong, V. Franco, R. Montemayor, Z.W. Liu, D.C. Zeng, Journal of Magnetism and Magnetic Materials 390 (2015) 87-90.[12] L. Li, C. Xu, Y. Yuan, S. Zhou, Materials Research Letters 6(8) (2018) 413-418.[13] B.J. Korte, V.K. Pecharsky, K.A. Gschneidner, Journal of Applied Physics 84(10) (1998) 5677.[14] Y. Mudryk, D. Paudyal, A.K. Pathak, V.K. Pecharsky, K.A. Gschneidner, Journal of Materials Chemistry C 4(20) (2016) 4521-4531.[15] K. Yano, Y. Tanaka, I. Matsumoto, I. Umehara, K. Sato, H. Adachi, H. Kawata, Journal of Physics: Condensed Matter 18(29) (2006) 6891-6895.[16] L. Li, Y. Yuan, Y. Qi, Q. Wang, S. Zhou, Materials Research Letters 6(1) (2018) 67-71.[17] H. Zhang, B.G. Shen, Z.Y. Xu, J. Shen, F.X. Hu, J.R. Sun, Y. Long, Applied Physics Letters 102(9) (2013) 092401.[18] M.K. Sharma, K. Yadav, K. Mukherjee, Journal of Physics: Condensed Matter 30(21) (2018) 215803.[19] T.A. Yamamoto, T. Nakagawa, K. Sako, T. Arakawa, H. Nitani, Journal of Alloys and Compounds 376(1) (2004) 17-22.[20] L. Li, K. Nishimura, H. Yamane, Applied Physics Letters 94(10) (2009) 102509.[21] Q.Y. Dong, B.G. Shen, J. Chen, J. Shen, J.R. Sun, Journal of Applied Physics 105(11) (2009) 113902.[22] M.E. Wood, W.H. Potter, Cryogenics 25(12) (1985) 667-683.[23] X.Q. Zheng, B. Zhang, H. Wu, F.X. Hu, Q.Z. Huang, B.G. Shen, Journal of Applied Physics 120(16) (2016) 163907.Figure 1: Magnetization as a function of temperature for (a) HoNi2-yCoy and (b) Ho1-xGdxNi2 alloys under a magnetic field of 1 T; (c) Zero field cooling (ZFC) - field cooling (FC) and (d) dM/dT values as a function of temperature for Ho1-xGdxNi2-yCoy alloys under 0.05 T showing change of their Curie temperature.Figure 2: SEM-EDS maps and composition line profiles of constituent elements obtained from matrix phase of (a) (Ho0.7Gd0.3)Ni2 and (b) Ho(Ni1.7Co0.3) alloys. (c) High resolution HAADF-STEM images and atomic resolution STEM-EDS maps of Ho (red), Co (blue), and Ni (green) obtained along  zone axis in HoNi1.7Co0.3 sample.Figure 3 (a) XRD patterns Ho1-xGdxNi2 alloys and (b) the refinement result for Ho0.7Gd0.3Ni2; and (c) lattice parameter a of HoNi2 phase in the Ho1-xGdxNi2 and HoNi2-yCoy samples.Figure 4: (a) Isothermal magnetic entropy change (∆SM) as a function of temperature for the Ho1-xGdxNi2 and HoNi2-yCoy alloys under a magnetic field of 5 T; (b) Table like ∆SM as a function of temperature is observed for the Ho0.7Gd0.3Ni2 alloy under 1 to 5 T applied magnetic field. (c) and (d) ∆SM with different units of J/KgK and J/cm3K under 5 T applied magnetic field as a function of temperature for Ho1-xGdxNi2 and HoNi2-yCoy alloys and comparison of the ∆SM with the other promising candidates in the temperature range of 14 to 34 K [5-6, 17-18].Figure 1Figure 2Figure 3Figure 411image3.pngimage4.pngimage1.pngimage2.jpeg