Giant magnetocaloric effect in a rare-earth-free layered coordination polymer at liquid hydrogen temperatures

Magnetic refrigeration exploits the magnetocaloric effect (MCE), offering an environmentally friendly and potentially more efficient alternative to vapor compression at room temperature and to Joule–Thomson expansion for cryogenic applications such as hydrogen liquefaction (Tboil ≈ 20.3 K). Reducing the energy and cost of hydrogen liquefaction motivates the search for magnetocaloric materials composed of earth-abundant elements that perform strongly under moderate magnetic field changes (Δμ0H ≤ 2 T) achievable with permanent magnets. Traditional high-performance MCE materials often rely on rare-earth elements due to their large magnetic moments. This work addresses the need for rare-earth-free alternatives by introducing and investigating a layered organic–inorganic hybrid coordination polymer, Co4(OH)6(SO4)2[enH2], designed to combine strong magnetocaloric performance near the hydrogen liquefaction temperature with minimal hysteresis, structural stability, and operation in moderate magnetic fields.

Prior research on magnetocaloric materials at cryogenic temperatures has focused heavily on rare-earth-containing systems (e.g., Gd5(Si2Ge2), RT2 Laves phases, RTX and RT2X2 families), which maximize magnetic entropy change via large moments. In molecular and coordination polymer materials, rare-earth-based Gd clusters and frameworks (e.g., Gd(HCOO)3, Gd(OH)CO3) exhibit excellent ultra-low-temperature MCE (ΔSM > 50 J kg−1 K−1 for Δμ0H = 7 T). Due to rare-earth scarcity and criticality, transition-metal-based alternatives have gained interest. In the hydrogen liquefaction range, layered transition-metal compounds like β-Co(OH)2 and CrCl3 show large ΔSM ≈ 15 J kg−1 K−1 for Δμ0H = 5 T around 15–20 K; Co3V2O8 reaches ΔSM ≈ 17 J kg−1 K−1 at 11 K for 5 T, and Co2(OH)4−xClx shows ΔSM ≈ 8 J kg−1 K−1 at 10 K for 2 T. These results suggest that layered and anisotropic, rare-earth-free materials can deliver strong MCE at low temperatures. Building on this, the present study targets a layered brucite-type hybrid material to realize strong MCE under low-to-moderate fields, minimal hysteresis, and structural robustness.

Synthesis: Co4(OH)6(SO4)2[enH2] single crystals were synthesized hydrothermally. Starting reagents: CoSO4·7H2O (4 mmol, 1.127 g), ethylenediamine (1.66 mmol, 0.11 mL), H2SO4 (0.5 mmol, 0.027 mL), and deionized water (4 mmol, 0.075 mL) were combined in a 25 mL PTFE liner, sealed in a stainless-steel autoclave, heated to 170 °C over 5 h, held for 5 days, then cooled to room temperature over 10 h. The product was washed with deionized water and ethanol. Exfoliation: Nanosheets were produced by sonication-assisted liquid-phase exfoliation: ~30 mg crystals in 15 mL 96% ethanol, sonicated 60 min. Single-crystal X-ray diffraction (SCXRD): Performed on a Bruker D8 Venture diffractometer with Photon 100 CMOS detector using Mo Kα radiation at 107 K. Twinning identified (180° rotation around [110]) via CELL_NOW; data integration with SAINT; multi-scan absorption correction with TWINABS; twinned structure refinement with SHELXL. FT-IR: PerkinElmer Spectrum 400 (ATR) from 4000 to 650 cm−1 on ground powders. PXRD: Temperature- and field-dependent PXRD on a custom Debye–Scherrer transmission diffractometer with MYTHEN detector and superconducting magnet using Mo radiation (λ1 = 0.70932 Å, λ2 = 0.7134 Å). Samples co-deposited with Si standard on graphite supports. Rietveld refinements via FullProf. AFM: Exfoliated nanosheets deposited on cleaned Si substrates (ethanol, water, acetone baths), dried in air; imaged with Bruker Dimension Icon in tapping mode. (S)TEM: Exfoliated nanosheets deposited on lacey carbon Cu grids; images collected on Thermo Fisher Themis Z S/TEM at 300 kV, probe current <1 pA, with LN2-cooled stage. Samples damaged rapidly in TEM mode. DC magnetization: Quantum Design MPMS XL-7 T SQUID. M(T) measured ZFC/FCC/FCW; M(H) isotherms for entropy analysis measured on warming from +Hmax to 0 T; hysteresis loops cycled +Hmax to −Hmax and back. Measurements performed with field parallel and perpendicular to ab-plane on single crystals. AC susceptibility: Quantum Design PPMS ACMS II option. ZFC conditions; AC drive field 0.43 mT; frequencies including 13, 33, 333 Hz; DC bias fields 0–500 mT; measured with field parallel to ab-plane. Heat capacity: Quantum Design PPMS Heat Capacity option. Sample: 3.585 mg agglomerate of small, approximately aligned crystals, anchored with Apiezon-N grease; high vacuum (≈1.23 mPa). Field-dependent addenda at 0, 2, 5, 10 T. Derived ΔSM and ΔTad from C(T,H) via entropy construction. Direct ΔTad: Dresden High Magnetic Field Laboratory. Two thin single crystals (~100 μm thickness each) glued together with silver epoxy; type E thermocouple (25 μm wires) between crystals; sample fixed with GE varnish; field aligned in ab-plane. Pulsed fields of 2, 5, 10 T; rise time 19 ms. Compared direct ΔTad to indirect values from heat capacity. Structural and morphology characterization: SCXRD identified triclinic P1 brucite-type layers of distorted Co(OH)4(SO4)2 and Co(OH)5(SO4) octahedra forming triangular lattices, separated by hydrogen-bonded enH2 cations. FT-IR confirmed hydroxyl groups. AFM and STEM confirmed interlayer spacing ~1.05 nm via step heights and lattice fringes.

- Structure: Co4(OH)6(SO4)2[enH2] crystallizes in triclinic P1 with brucite-type layered Co–O frameworks (distorted Co(OH)4(SO4)2 and Co(OH)5(SO4) octahedra) forming triangular lattices; interlayer spacing ~10.538 Å, layers separated by enH2 via H-bonding. Co–O distances 2.028(2)–2.338(2) Å. - Magnetic ordering: Ferromagnetic second-order transition with 10 K < Tc < 15 K; negligible thermal hysteresis in FCC/FCW; small ZFC–FC difference below ~11 K indicates dynamic character. - Anisotropy: Pronounced easy-plane (ab-plane) anisotropy. M||(T) in 200 mT nearly saturates at low T with small apparent moment in that field; isothermal M(H) below Tc saturates near Msat ≈ 2.80 μB/Co atom for μ0H < 1 T (ab-plane). Coercive field μ0Hc ≈ 5 mT at T < 12 K (much smaller than related layered cobalt compounds). - Curie–Weiss analysis (100–400 K): Modified Curie–Weiss fits yield θab = 21.2 ± 1.1 K (in-plane FM) and θc = −50.1 ± 1.6 K (out-of-plane AFM). Temperature-independent susceptibilities χ0ab = 1.59(6)×10−7 m3/kg, χ0c = 3.89(8)×10−7 m3/kg. Effective moments μeffab = 2.86 ± 0.02 μB/Co and μeffc = 2.88 ± 0.02 μB/Co (lower than spin-only S=3/2, attributed to deviations from Curie–Weiss due to low dimensionality/short-range correlations); low-T behavior confirms S=3/2 ground state. - AC susceptibility: χ′ shows a peak at 10.5 K with concurrent large χ″; below 10.5 K both become frequency dependent. With DC bias field, χ′ peak splits: a higher-T peak emerges around 11 K and shifts to higher T with field, while lower-T frequency-dependent peaks shift to lower T and diminish, vanishing for μ0H ≥ 500 mT. Indicates dynamic ferromagnetic ordering without evidence of spin-glass or superparamagnetism. - Magnetocaloric effect (single crystal, ab-plane): ΔSM(T) shows broad asymmetric peaks shifting with field. Peak temperature ~13 K for Δμ0H = 0.5 T, moving to ~17 K for 7 T. Max ΔSM values: −6.3 J kg−1 K−1 (1 T), −11.4 J kg−1 K−1 (2 T), −18.3 J kg−1 K−1 (5 T). Easy-plane anisotropy leads to much smaller ΔSM for H ⟂ ab-plane. - Heat capacity (polycrystalline textured sample): Zero-field lambda-like anomaly at 10.2 K; with field, the peak shifts to higher T and decreases. Indirect ΔSM maxima: −6.2 (1 T), −8.7 (2 T), −15.3 J kg−1 K−1 (5 T). Reduced values vs single crystal due to misalignment and anisotropy. - Adiabatic temperature change: Indirect ΔTad maxima from heat capacity: 1.98 K (1 T), 3.22 K (2 T), 5.88 K (5 T). Direct pulsed-field ΔTad (2, 5, 10 T) agrees well for T > 25 K; near Tc, direct values are lower due to thermal losses from mounting media; hysteresis observed in ΔTad vs field cycle at low T supports heat-loss explanation. - Critical behavior: Arrott and scaling analyses classify the transition as second order. Critical exponents: η = 0.489, δ = 12.585, β = 0.134, γ = 1.53, close to 2D XY values (β = 1/8, γ = 7/4, δ = 15). Modified Arrott–Noakes plots and universal scaling confirm consistency. - Application relevance: At hydrogen liquefaction temperatures, performance in 1–2 T fields is exceptional for rare-earth-free materials and competitive with many rare-earth alloys. Material contains ~30% less Co by weight for similar performance (at 2 T) compared to Co3V2O8; structural flexibility suggests potential for further optimization.

The work targets the development of rare-earth-free magnetocaloric materials effective near 20 K under moderate magnetic fields. Co4(OH)6(SO4)2[enH2] achieves this via a second-order ferromagnetic transition in the 10–15 K range, combining minimal hysteresis (enabling cyclic operation) with large magnetic entropy and adiabatic temperature changes under low fields (1–2 T). The layered brucite-derived structure induces strong easy-plane anisotropy and XY-like critical behavior, allowing rapid alignment of spins in small fields, thus maximizing ΔSM without structural or volumetric discontinuities. Curie–Weiss analysis indicates competing exchange interactions: ferromagnetic in-plane superexchange (near 90° Co–O–Co) and weak interlayer interactions near the critical distance (~10 Å) where AFM superexchange and FM dipolar interactions compete, plausibly accounting for the very small coercive fields. AC susceptibility reveals a dynamic ordered state but, together with structural order and other ferromagnetic signatures, excludes spin-glass or superparamagnetic origins. Compared to rare-earth-based and other transition-metal layered materials (β-Co(OH)2, CrCl3, Co3V2O8), the present system offers competitive ΔSM and ΔTad in practical fields while employing earth-abundant chemistry. The material’s structural tunability (via metal and interlayer substitutions) and potentially favorable thermal transport properties (by analogy to other coordination polymers) further enhance its promise for magnetic hydrogen liquefaction technologies.

This study introduces a rare-earth-free layered hybrid coordination polymer, Co4(OH)6(SO4)2[enH2], with a second-order ferromagnetic transition in the 10–15 K range and a giant magnetocaloric effect suitable for hydrogen liquefaction. Under moderate fields accessible with permanent magnets, the material exhibits large ΔSM (e.g., −6.3 J kg−1 K−1 at 1 T; −11.4 J kg−1 K−1 at 2 T) and ΔTad (up to 1.98 K at 1 T and 3.22 K at 2 T indirectly), minimal hysteresis, and no detectable magnetostructural coupling. Critical scaling indicates 2D XY-like behavior linked to the layered structure and easy-plane anisotropy. The compound can be synthesized via a simple, low-temperature hydrothermal route (170 °C). Future work should exploit the structural flexibility of brucite-type layers to optimize magnetocaloric performance and sustainability: substituting Co with more abundant high-moment ions (e.g., Fe2+, Mn2+), tuning interlayer spacing/chemistry, and engineering texturing to maximize anisotropy benefits. Measuring and optimizing thermal conductivity and device-level regenerator performance will be valuable next steps.

- Orientation effects: Heat capacity and indirect ΔSM/ΔTad were obtained on a textured, polycrystalline agglomerate with imperfect ab-plane alignment, leading to reduced apparent MCE values compared to single crystals. - Direct ΔTad underestimation near Tc: In pulsed-field measurements, thermal coupling to the holder (silver epoxy and GE varnish) and small sample heat capacity at low T acted as heat sinks, reducing measured ΔTad and introducing hysteresis in temperature during field cycling. - No thermal conductivity data: Thermal transport, important for device performance, was not measured; anisotropic conductivity is anticipated but unverified. - High-temperature magnetic analysis: Deviations from Curie–Weiss behavior above Tc lead to anomalous μeff values; origin likely tied to low dimensionality and short-range correlations, but not fully resolved. - Anisotropy: Strong magnetic anisotropy implies device designs must control texture/orientation for optimal performance; bulk polycrystalline averages will underperform single crystals or textured materials.