The magnetocaloric (MC) properties in many rare-earth (RE)-containing magnetic solids have been intensively investigated, which are aimed to develop suitable candidates for cryogenic magnetic cooling applications and to better understand their intrinsic magnetic characters. We herein fabricated the RE-rich RE6Ni2.3In0.7 (RE = Ho, Er, and Tm) compounds and investigated their structural, magnetic, and MC properties by experimental determination and theoretical calculations. All of these RE6Ni2.3In0.7 compounds crystallize in an Ho6Co2Ga-type structure with an orthogonal Immm space group and order magnetically around the temperatures of 10.6 and 33.0 K for Ho6Ni2.3In0.7, 11.0 K for Er6Ni2.3In0.7, and 7.6 K for Tm6Ni2.3In0.7, respectively. Large cryogenic reversible MC effects were observed in these RE6Ni2.3In0.7 compounds. Moreover, their MC parameters of maximum magnetic entropy changes, relative cooling powers, and temperature-averaged magnetic entropy change are comparable with those of some recently updated cryogenic MC materials.
I. INTRODUCTION
The magnetic cooling1–4 that relied on a magnetocaloric (MC) effect in magnetic solids was recognized as an effective new cooling technology with economic and environmental advantages to replace vapor-compression technology.1–6 The magnetocaloric effect (MCE), known as an inherent thermodynamic response,1–5 can be primarily expressed by the heat absorption or release (entropy or temperature changes) response to the variation of an external magnetic field. Thus, identifying prominent candidate materials with remarkable MC performances3–17 is one of the most important prerequisites for magnetic cooling applications. Moreover, it was found that the magnetocaloric performances of the magnetic solid are closely correlated with their magnetic phase transition (MPT) properties.4–24 As a consequence, various types of magnetic solids, including oxides, alloys, polymers, composites, and intermetallic compounds, have been investigated intensively within around the last 30 years with respect to their MC and MPT properties,8–24 and this study aims to develop suitable MC candidate materials for applications and better elucidate their intrinsic magnetic characters. Among the identified candidate materials featuring remarkable MC performances, the rare-earth (RE)-containing magnetic solids are accounted for large proportions. For example, the famous giant MC materials of La(FeSi)13Hy and Gd5(Si,Ge)4-based compounds1,2 are attractive for room-temperature (RT) cooling applications. Large cryogenic MCE was recently reported in Gd2MgTiO6 and Gd4Al2O9 oxides,18,19 which are correlated with their unstable magnetic ground state. The pseudo-binary RE1−xRExNi2 and pseudo-ternary RE1−xRExCoNi compounds20–22 are found to be attractive candidates for MC hydrogen liquification. A large MCE was very recently reported in the GdCoC and Gd2CuTiO6 compounds,23,24 which are correlated with their unique 4f–3d magnetic interactions. Detailed summaries of the MCE in RE-containing magnetic solids are provided in several review works4–6 published recently. These aforementioned works indicated the highly potentials of RE-containing magnetic solids with remarkable MC performances, warranting further exploration.
The ternary RE-containing (621 phases) compounds25–33 with a general formula of RE6TM2X (TM = 3d transition metals; X = Al, Ga, or In)25–27 that crystallize mainly in an orthorhombic Ho6Co2Ga-type structure (space group Immm) aroused many research interest in recent years.28–35 An important and special feature in the structure of 621 phase compounds25–27 is that a certain mixture range of TM and X can happen at the X site, resulting in the formation of RE6TM2+xX1−x phases.25–35 Previous investigations on the 621 phase compounds28 indicated that most of them display one or more cryogenic MPTs.27–35 Despite their magnetism mainly originated from the RE elements, the MPT order types and temperatures of these RE6TM2+xX1−x compounds are also strongly dependent on the other constituting elements.25–28 For example, three successive MPTs at ∼35, ∼17, and ∼7 K were found in an Er6Ni2Sn compound. The Gd6Co2.2In0.8 compound29 reveals two successive MPTs at ∼66 and ∼34 K. Also, only a single MPT from paramagnetic to antiferromagnetic (AFM) at ∼42 K was observed in a Tb6Co2.35Sn0.65 compound.30 Moreover, the MPT and MC properties in RE6Co2.2Al0.8 compounds were determined by Morozkin et al.,32 in which a large MCE in Gd6Co2.2Al0.8 around ∼75 K with the maximum magnetic entropy change (−ΔSMmax) of 10.89 J/kg K (magnetic entropy change Δμ0H = 0–5 T) was observed. A large cryogenic MCE was also reported in RE6Co2.2In0.8 compounds34 around their MPT temperatures with the −ΔSMmax values of 20.4, 9.6, and 11.8 J/kg K for Ho6Co2.2In0.8, Tb6Co2.2In0.8, and Gd6Co2.2In0.8 under Δμ0H of 0–7 T, respectively. The MPT and MC properties in RE6Co2Ga compounds35 were recently determined by Guo et al., in which the Ho6Co2Ga/Gd6Co2Ga composites show large −ΔSMmax values in a wide temperature range. Given the reported large cryogenic MCE in some RE6TM2+xX1−x compounds, herein, we fabricated the RE6Ni2.3In0.7 compounds and determined their magnetic, MPT, and MC properties by experimental determination and theoretical calculations. The deduced MC parameters of the present RE6Ni2.3In0.7 compounds are comparable with some recently updated RE-containing magnetic solids, making them considerable for applications.
II. EXPERIMENTAL DETAILS
Polycrystalline RE6Ni2.3In0.7 (RE = Ho, Er, and Tm) samples were prepared by a typical arc-melting route from a high-purity metal of RE (99.9%), Ni (99.99%), and In (99.99%). The alloy ingots were all melted four times followed by flipping over after each melting to achieve better homogeneity. The obtained alloy ingots were then directly sealed in the evacuated quartz crucible. Finally, the quartz crucibles with the alloy ingots were annealed at 1013 K for one week. The structural characters of RE6Ni2.3In0.7 compounds were determined by powder x-ray diffraction (XRD, Rigaku, SmartLab-9KW) at RT. The magnetization (M) data of RE6Ni2.3In0.7 compounds were collected by a magnetic property measurement system (MPMS, QD, Model3) with temperature ranges of 3–300 K (sweeping rate of 3 K/min) and the magnetic field up to 7 T (sweeping rate of 250 Oe/s).
The atomic-level first-principles DFT calculations for RE6Ni2.3In0.7 compounds were performed using commercialized VASP software. The electronic exchange and correlation of RE6Ni2.3In0.7 compounds were modeled by Perdew–Burke–Ernzerhof functional36–38 within spin-polarized generalized gradient approximation. The valence electron contributions in pseudopotentials are [4f11 6s2] for Ho, [4f12 6s2] for Er, [4f13 6s2] for Tm, [3d8 4s2] for Ni, and [5s2 5p1] for In, respectively. Structural optimizations were conducted using the conjugate gradient algorithm until the Hellman–Feynman forces and self-consistent electronic loops were converged to lower than 0.01 eV/Å and 10−4 eV, respectively. Also, the k-point grid using the Monkhorst–Pack method is set as 4 × 4 × 3 for calculations.
III. RESULTS AND DISCUSSION
The experimental obtained XRD patterns at RT of the RE6Ni2.3ln0.7 (RE = Ho, Er, and Tm) compounds and corresponding Rietveld refinement profiles fitted by FullProf39 are presented in Figs. 1(a)–1(c). Evidently, all of these RE6Ni2.3ln0.7 compounds exhibit similar behaviors, apart from a small amount of secondary phase of Er3Ni2In (∼3.19 wt. %) and Tm3Ni2In (∼7.15 wt. %) in the corresponding RE6Ni2.3ln0.7 compounds, and the main diffraction peaks of these RE6Ni2.3ln0.7 compounds can be all indexed and well-fitted by a Ho6Co2Ga-type structure with an orthogonal Immm space group. The obtained reliability factors, including the Rp, Rwp, and Rexp, are 2.06%, 2.85%, and 1.2% for Ho6Ni2.3ln0.7; 3.79%, 5.42%, and 1.07% for Er6Ni2.3ln0.7; and 3.52%, 4.8%, and 1.06% for Tm6Ni2.3ln0.7, respectively. These low values of all reliability factors verify the reliability of the fittings. The determined lattice constants of a, b, c, and V are 9.321, 9.497, 9.888 Å, and 875.30 Å3 for Ho6Ni2.3ln0.7; 9.246, 9.413, 9.832 Å, and 855.70 Å3 for Er6Ni2.3ln0.7; and 9.225, 9.401, 9.830 Å, and 852.50 Å3 for Tm6Ni2.3ln0.7, respectively. All of these lattice constants show decrease trends with an increase in the RE atomic number, adhering the contraction law of RE elements. The schematic crystal structure of RE6Ni2.3ln0.7 compounds is presented in Fig. 1(d). The RE atoms occupied at three different Wyckoff sites of 8n, 8m, and 8l, as shown by different colors. The RE1 atom forms grid locating within xy planes. Similarly, the RE2 and RE3 atoms form grids locating within xz and yz planes, respectively. The Ni atom occupied at two different Wyckoff sites of 4j and 4g, while the Al atom occupied at the 2c site. Additionally, the remaining Al and Ni atoms share the 2a Wyckoff site. Additionally, six RE atoms form an octahedron [see Fig. 1(e)], and there are eight such octahedrons in the unit cell. These octahedrons connect via RE1, RE2, and RE3 atoms along the z axis, y axis, and x axis, respectively, forming a typical three-dimensional network structure.
To investigate the magnetic and MPT properties of RE6Ni2.3ln0.7 compounds, temperature-dependent M(T) curves were determined. The resulting M(T) and inverse susceptibility 1/χ(T) curves of RE6Ni2.3ln0.7 compounds under an external magnetic field (μ0H) of 1 T are presented in Fig. 2, and the corresponding M(T) curves under μ0H of 0.1 T for RE6Ni2.3ln0.7 compounds in field-cooled (FC) and zero-field-cooled (ZFC) modes are presented in the inset. All of these RE6Ni2.3ln0.7 compounds evidently exhibit cryogenic magnetic orderings. Specifically, two successive low-temperature MPTs at TM of ∼10.6 and ∼33.0 K can be noted for Ho6Ni2.3ln0.7. Only a single MPT from a paramagnetic (PM) to an AFM state was identified at TM of ∼11.0 K for Er6Ni2.3ln0.7, whereas the Tm6Ni2.3ln0.7 shows the MPT around TM of ∼7.6 K from a PM to ferromagnetic (FM) state. The small deviations that observed below TM in present RE6Ni2.3ln0.7 compounds are probably attributed to the existence of domain-wall pinning effects and/or the indication spin-freezing. Moreover, the 1/χ(T) curves of present Ho6Ni2.3ln0.7 and Er6Ni2.3ln0.7 compounds above 120 K, as presented in Fig. 2, generally show linear behavior, i.e., obeying the Curie–Weiss law of and , in which θP and μeff represent the paramagnetic Curie temperature and the effective magnetic moment, respectively. The linear fitting of 1/χ(T) curves resulted in the μeff values of 10.40, 9.46, and 7.33 μB/f.u. for Ho6Ni2.3ln0.7, Er6Ni2.3ln0.7, and Tm6Ni2.3ln0.7, respectively, and the slight difference for Tm6Ni2.3ln0.7 with those of theoretical values by only considering the free RE3+ ions is probably due to the existence of an impurity phase.
In an attempt to shed some light on the ground state properties of RE6Ni2.3ln0.7 compounds, the spin-polarized DFT calculations36–38 were conducted. The stability of the crystal structure was first checked, and the orthorhombic Ho6Co2Ga-type structure (space group Immm) assumed from the XRD refinement was selected as the initial structure for optimization. The differences between optimized and experimental results of all the RE6Ni2.3In0.7 compounds are less than 1.0%, suggesting that the optimized crystal structure is logical and accessible. Then, the total energy (Etot) with different potential magnetic structures of RE6Ni2.3ln0.7 compounds was calculated by using the optimized structural parameters to decide their magnetic ground state. The arrangements of RE atoms by setting the spin directions (spin-up or spin-down) and, therefore, four magnetic configurations are constructed, including FM, A-AFM, C-AFM, and G-AFM states, as presented in Fig. 3. The corresponding calculated Etot values are −276.9681, −276.0709, −277.3717, and −277.3759 eV/f.u. for Ho6Ni2.3In0.7; −249.0665, −249.93017, −249.6143, and −250.2607 eV/f.u. for Er6Ni2.3In0.7; and −228.51133, −227.1848, −227.8451, and −227.6089 eV/f.u. for Tm6Ni2.3In0.7, respectively. Apparently, the lowest Etot of Ho6Ni2.3In0.7, Er6Ni2.3In0.7, and Tm6Ni2.3In0.7 corresponds to G-AFM, G-AFM, and FM state, respectively. Moreover, the total density of states (DOS) and partial DOS of each element in RE6Ni2.3ln0.7 compounds were also calculated, as presented in Figs. 4(a)–4(c), to further understand their electronic and magnetic ground state properties. The DOS of these RE6Ni2.3ln0.7 compounds crosses the Fermi level without creating a bandgap, illustrating the metallic nature of them. The DOS of Ho6Ni2.3In0.7 and Er6Ni2.3In0.7 have the symmetric splitting, indicating the AFM ground state, whereas polarization splitting for Tm6Ni2.3In0.7 indicates the FM ground state. Obviously, the partial DOS of RE-4f orbits show obvious splitting behavior and mainly controlled the total DOS of RE6Ni2.3ln0.7 compounds, suggesting the spontaneous polarization of RE3+ ions and the realization of large magnetic moments. These calculation results are well-consistent with those of aforementioned magnetization determination, which further proved vital roles of RE3+ ions in magnetism for these RE6Ni2.3ln0.7 compounds.
Additionally, the M(μ0H) look at 2 K and series M(μ0H) curves of RE6Ni2.3ln0.7 compounds were determined, which are given in Figs. 5(a) and 5(b)–5(d), respectively. Generally, only very tiny differences during M(μ0H) loops at 2 K, as displayed in the inset of Fig. 5(a), can be observed. Specifically, the values of M have linear increasing behavior with an increase in μ0H of Ho6Ni2.3ln0.7 and Er6Ni2.3ln0.7 in the curves below TN under low-μ0H, then show quick rises in M under medium-μ0H, and exhibit saturation-like behavior under high-μ0H regions. These behaviors in Ho6Ni2.3ln0.7 and Er6Ni2.3ln0.7 compounds are quite similar with those of recently reported RE2Ga2Mg40 and RE2O2SO441 compounds, in which a typical μ0H-induced metamagnetic transition arisen from AFM to FM-like state occurred, whereas the M values increased abruptly with a rise in μ0H of Tm6Ni2.3ln0.7 under low-μ0H regions and then show saturation-like characters under high-μ0H regions at low temperatures, illustrating a typical FM state. Moreover, for the curves well above TM, nearly linear relations between M and μ0H can be noted in all of the present RE6Ni2.3ln0.7 compounds, illustrating the PM states. Moreover, it was found that the magnetocaloric performances of the magnetic solid are closely correlated with their MPT properties,4–10 which could be determined primarily from the slopes42 in the curves of the Arrott plot (M2 vs μ0H/M) based on the Banerjee criterion.42 The exhibition of positive or negative values in the slope, in principle, corresponds to the second-order or first-order type MPT.42 Thus, we directly transferred the M(μ0H) data to Arrott plot curves, which are given in Figs. 6(a)–6(c), in which clearly negative slopes can be identified for Ho6Ni2.3ln0.7 and Er6Ni2.3ln0.7, illustrating first-order type MPT(s) at low temperatures, whereas only positive slopes can be found for Tm6Ni2.3ln0.7, illustrating second-order type MPT.
The resultant −ΔSM(T) curves of present RE6Ni2.3ln0.7 compounds are given in Figs. 7(a)–7(c). Generally, the MCE can be divided into two different types, i.e., the direct-type (or conventional) MCE and the inverse-type MCE.4–6,40,41 The material heats up when μ0H is applied for the direct MCE, which can be usually observed in the traditional FM-ordered materials,20–23 whereas the material cools down when μ0H is applied for the inverse MCE, which has been reported for some materials with special AFM ordered.4–6,40,41 An obvious crossover in −ΔSM(T) curves for Ho6Ni2.3ln0.7 and Er6Ni2.3ln0.7 under low Δμ0H can be noted; e.g., it changes from negative −ΔSM values (inverse MCE) to positive −ΔSM values (direct MCE) with the rise in temperature. The observed inverse MCE of Ho6Ni2.3ln0.7 and Er6Ni2.3ln0.7 further confirms the AFM states in these compounds, and the direct MCE is believed to be related to their μ0H-induced FM-like states. The maximum value of negative −ΔSM (−ΔSMmax) referring the inverse MCE is −1.54 J/kg K for Ho6Ni2.3ln0.7 under Δμ0H of 0–3 T and −1.03 J/kg K for Er6Ni2.3ln0.7 under Δμ0H of 0–2 T, respectively. Moreover, only positive –ΔSM values can be observed within the present tested μ0H and T regions for Tm6Ni2.3ln0.7, which are consistent well with its FM state. Moreover, the positive −ΔSMmax values of these RE6Ni2.3ln0.7 compounds that refer to direct MCE are deduced to be 2.54 and 9.66 J/kg K for Ho6Ni2.3ln0.7, 4.13 and 13.93 J/kg K for Er6Ni2.3ln0.7, and 9.60 and 17.59 J/kg K for Tm6Ni2.3ln0.7, under Δμ0H of 0–3 and 0–7 T, respectively. Evidently, the Tm6Ni2.3ln0.7 exhibits the largest direct MCE, which is probably related to its FM ground state and larger magnetic moments. Additionally, the MPT order types can also be simply determined from the signs of exponent n, which was proposed recently by Franco et al.,43,44 and the values of n were evaluated by . The resultant n(T) curves of RE6Ni2.3ln0.7 compounds are presented in Figs. 8(a)–8(c). Generally, n > 2 is a first-order type MPT; otherwise, it will be of second-order type MPT.43,44 Obvious overshoots with n above 2 can be observed for Ho6Ni2.3ln0.7 and Er6Ni2.3ln0.7, further proving the first-order type MPT, whereas no overshoot with n above 2 can be found for Tm6Ni2.3ln0.7, further proving the second-order type MPT. The relative cooling power ( )4–6 and the temperature-averaged entropy change ( )45 with proper lifted temperatures (ΔTlift) of present RE6Ni2.3ln0.7 compounds were also determined. The Tmid denotes the central temperature where TEC(ΔTlift) exhibits its maximum. The resultant RCP values are deduced to be 67.85 and 344.09 J/kg for Ho6Ni2.3ln0.7, 86.87 and 388.71 J/kg for Er6Ni2.3ln0.7, and 119.24 and 371.17 J/kg for Tm6Ni2.3ln0.7 under Δμ0H of 0–3 and 0–7 T, respectively. The TEC(5) and TEC(10) values (Δμ0H = 0–7 T) are 9.48 and 9.27 J/kg K for Ho6Ni2.3ln0.7, 13.45 and 13.01 J/kg K for Er6Ni2.3ln0.7, and 17.07 and 16.17 J/kg K for Tm6Ni2.3ln0.7, respectively. Interestingly, the deduced MC parameters of these RE6Ni2.3ln0.7 compounds, especially for Tm6Ni2.3ln0.7, are evidently better than the reported RE6TM2+xX1−x compounds32–35 and also comparable to some recently updated RE-containing magnetic solids, such as the RE2Ga2Mg,40, RE2RuIn,46, RE2Cr2C3,47 and RE5Ni2Sb48 compounds, as well as the RE2MgTiO6,49, RE2O2SO4,41, REOCl,14, RE2Ti2O7,50 and RE2BaZnO5 oxides,51 with remarkable cryogenic performances, making the Tm6Ni2.3ln0.7 compound considerable for applications.
IV. CONCLUSION
In summary, three polycrystalline RE6Ni2.3ln0.7 (RE = Ho, Er, and Tm) samples were fabricated and investigated their structure, magnetic, and magnetocaloric properties by experimental determination and theoretical calculations. All of these RE6Ni2.3ln0.7 compounds crystallize in a Ho6Co2Ga-type structure with an orthogonal Immm space group and magnetically ordered at low temperatures. Large cryogenic MCE and remarkable performances were realized in these RE6Ni2.3ln0.7 compounds. Moreover, their MC parameters under Δμ0H of 0–7 T, including the –ΔSMmax, RCP, and TEC(5), reach 9.66 J/kg K, 344.09 J/kg, and 9.48 J/kg K for Ho6Ni2.3In0.7; 13.93 J/kg K, 388.71 J/kg, and 13.46 J/kg K for Er6Ni2.3In0.7; and 17.59 J/kg K, 371.17 J/kg, and 17.07 J/kg K for Tm6Ni2.3In0.7, respectively, which are comparable to some recently updated RE-containing magnetic solids with remarkable cryogenic performances.
ACKNOWLEDGMENTS
The present work was supported by the National Natural Science Foundation of China (No. 52071197). The authors also acknowledge the Supercomputing Center of Hangzhou Dianzi University for providing computing resources.
AUTHOR DECLARATIONS
Conflict of Interest
The authors have no conflicts to disclose.
Author Contributions
Yang Xie: Data curation (equal); Investigation (equal); Validation (equal); Writing – original draft (equal). Jinyi Wang: Data curation (equal); Investigation (equal); Visualization (equal). Fuyu Yang: Data curation (equal); Investigation (equal); Validation (equal). Jiayu Ying: Data curation (equal); Investigation (equal); Validation (equal); Visualization (equal). Yikun Zhang: Conceptualization (equal); Funding acquisition (equal); Project administration (equal); Supervision (equal); Writing – original draft (equal); Writing – review & editing (equal).
DATA AVAILABILITY
The data that support the findings of this study are available from the corresponding author upon reasonable request.