Dual-phase nano-glass-hydrides overcome the strength-ductility trade-off and magnetocaloric bottlenecks of rare earth based amorphous alloys

Abstract

Metal-hydrogen systems have attracted intense interest for diverse energy-related applications. However, metals usually reduce their ductility after hydrogenation. Here, we show that hydrogen can take the form of nano-sized ordered hydrides (NOH) homogeneously dispersed in a stable glassy shell, leading to remarkable enhancement in both strength and ductility. The yield strength is enhanced by 44% and the plastic strain is substantially improved from almost zero to over 70%, which is attributed to the created NOH and their interplay with the glassy shell. Moreover, the hydride-glass composite GdCoAlH possesses a giant magnetic entropy change (−ΔS_M) of 18.7 J kg⁻¹K⁻¹ under a field change of 5 T, which is 105.5% larger than the hydrogen-free sample and is the largest value among amorphous alloys and related composites. The prominent ΔS_M-ductility combination overcomes the bottlenecks of amorphous alloys as magnetic refrigerants. These results provide a promising strategy for property breakthrough of structural-functional alloys.

Introduction

The interaction between hydrogen and metals is a topic of great significance in materials science. With increasing attention on hydrogen, an energy carrier of the future, hydrogen-metal systems have been intensively explored for various energy-related applications, such as hydrogen storage and sensing, rechargeable batteries, catalysis, refrigeration, and so on^1,2,3. On the other hand, many metals and alloys suffer from hydrogen embrittlement^4,5. Hydrogen embrittlement becomes a critical issue in many industrial applications of various alloys, including high-strength steels, aluminum alloys, titanium alloys, and magnesium alloys, etc., and thus has drawn much research attention over 100 years^4,5,6. Although several mechanisms were proposed for the crystalline metals/alloys-hydrogen systems, the nature of hydrogen embrittlement is not well understood except for a few Zr, Ti metals/alloys (where the hydride formation leads to embrittlement)⁶. It was found that amorphous alloys (AMAs) may show better resistance to hydrogen embrittlement than crystalline alloys, but they become rather brittle above a critical hydrogen concentration as well^7,8,9. Thus, it is widely and urgently needed to develop advanced hydrogen-metal systems combining excellent functional and mechanical performances.

Different from conventional metals and alloys, AMAs have a unique disordered atomic structure with no grain boundaries and dislocation. This endues AMAs with high strength, improved wear resistance, excellent soft magnetic and anti-corrosion properties, and so on^10,11,12. Nevertheless, the AMAs have several inherent shortcomings impeding their development and wider applications. Besides the restricted sample size imposed by the critical cooling rate for glass formation and limited working temperature below the glass transition temperature, their poor plasticity at ambient temperature and propensity for catastrophic failure originating from severe plastic-strain localization in shear bands (SBs) are critical drawbacks¹¹, largely undercutting their structural and functional utilization. Moreover, some functional properties cannot be effectively enhanced and even may be degraded through the amorphization of alloys, such as magnetocaloric performances. Although the AMAs take the advantages of magnetic softness, large resistivity leading to smaller eddy current loss, and broad magnetic entropy change (ΔS_M) peak as promising magnetic refrigerant^13,14, they suffer from the drawback of low-to-medium ΔS_M owing to the amorphous nature with second-order magnetic transition. Among various AMA systems, the Gd-based AMAs showing much larger maximum ΔS_M ($-\triangle {S}_{M}^{{pk}}$) than the Pd-, Fe-, Ni- and Co-based AMAs^{13,1f–h, which show homogenous distributions of the Gd, Al, and Co elements and do not correlate apparently with heterogeneity seen from the HAADF-STEM image. Thus, the contrast variation in the STEM image mainly arises from the density fluctuation, which is an intrinsic feature of AMAs.}

Fig. 1: Structural and thermal characterizations of the GdCoAl amorphous powders and isothermal hydrogenation.

a SEM image of the amorphous powders with a diameter of 25–30 μm. b TEM image and c SAED pattern of the powder. d HRTEM image of the powder. e–h High-resolution HAADF image and corresponding EDS map**s of Al, Co, and Gd elements. i XRD patterns of the GdCoAl powders with different diameters less than 70 μm and the GdCoAlH powder with the saturated absorption of hydrogenation. j DSC curves of the GdCoAl and GdCoAlH powders at the heating rate of 20 K/min. The inset displays the curve of hydrogenation dynamics.

AMAs have a disordered structure and a wide spectrum of atomic-packing heterogeneities, which provide an infinity of interstitial sites for hydrogen occupation⁸. For the present Gd₅₅Co_17.5Al_27.5 AMA containing Gd element with a strong affinity for hydrogen, a high hydrogen concentration and formation of hydride in the alloy can be expected after hydrogenation. Figure 1i shows the X-ray diffraction (XRD) patterns of the as-prepared Gd₅₅Co_17.5Al_27.5 and hydrogenated powders (GdCoAl and GdCoAlH for short, respectively). The broad diffraction maxima of the pattern shifts to lower 2θ after hydrogenation, indicating a volume expansion as observed in other AMAs^29,30. The differential scanning calorimetry (DSC) curves of the GdCoAl and GdCoAlH powders are presented in Fig. 1j. Distinct endothermal events corresponding to glass transition and exothermal event resulting from crystallization are observed for the as-prepared powders. And the onset temperatures for glass transition and crystallization are determined to be 600 and 670 K, respectively. A large supercooled-liquid region of 70 K, indicating its excellent glass-forming ability. After hydrogenation, the crystallization temperature decreases slightly to around 667 K. Due to the overlap between the glass transition and hydrogen desorption (starting around 600 K), the onset temperature of the glass transition cannot be determined accurately. The existence of a supercooled-liquid region and crystallization peak after hydrogenation indicates that there remains a significant amount of amorphous phase.

To explore the interaction between hydrogen and Gd-based AMA in more details, the atomic structure of the hydrogenated alloy was further investigated by special aberration-corrected high-resolution TEM (HRTEM) and HAADF-STEM. The HRTEM image indicates that the sample possesses a unique structure with some structurally ordered regions with several nanometers in size embedded in the amorphous matrix (Fig. 2a). This can be further seen from the inverse Fourier transform (IFT) image (Fig. 2b) of region A in Fig. 2a, in which lattice stripes (white area) were observed. Moreover, the SAED pattern reveals the presence of the GdH₂ phase (Fig. 2c), which implies that the ordered region may be the GdH₂ clusters. The STEM image indicates more clearly that many nanosized (2–5 nm) ordered hydrides (NOH) are embedded homogeneously in the amorphous matrix (Fig. 2d). The amorphous matrix serves as a shell with a thickness of 1–4 nm, which wraps up the NOH structure. And the concentration profiles of Gd, Al, and Co elements (Fig. 2e) along the diagonal line of the STEM image in Fig. 2d show a random fluctuation without any correlation with the contrast variation observed in Fig. 2d. This means that there is no apparent concentration difference of Gd, Al, and Co between the NOH and amorphous matrix. High-resolution STEM (HRSTEM) has been carried out to further characterize NOH regions (Fig. 2f). The NOH regions can be observed clearly, which comprise several ordered clusters. Element map**s in Fig. 2g–i present a homogenous distribution, which agrees with the results shown in Fig. 2e. This means that hydrogenation does not induce aggregation of elements, and the hydrogen atoms just occupy the interstitial sites and most of them interact with their neighboring Gd atoms to form NOH. Electron energy loss spectroscopy (EELS) was further used to characterize the structurally ordered regions. Figure 2j shows the EELS spectra for the typical ordered region B and disordered region C. A peak at 13.6 eV from the hydrogen K edge is clearly observed for region B (NOH) in Fig. 2f, which is absent for region C. In a word, the Gd-based SNDP-GH material comprises the randomly distributed GdH₂ clusters of around 3.6 nm in diameter and the remaining amorphous shell with a thickness of 1–4 nm. The similar structures were also constructed in the GdNiAl and DyCoAlSi AMAs via hydrogenation (Supplementary Fig. 1).

Fig. 2: Nanoscale structure and elemental analyses of the GdCoAlH powder.

a HRTEM image and b the further magnification of region A (after inverse Fourier transform) in a, showing the local ordering structures induced by hydrogenation. c Corresponding SAED pattern. d, e HAADF image and element fluctuation along the diagonal. f–i HRSTEM image and corresponding EDS map**s of Gd, Co, and Al elements. j EELS curves of the regions B (local ordering structure) and C (adjacent amorphous structure) in (f).

Mechanical properties

The brittleness of AMAs arising from a lack of crystalline defects is one of the critical problems hindering their wider practical applications as structural and/or functional materials, including magnetic refrigerant. It is well-known that many crystalline alloys are subjected to hydrogen embrittlement^4,5. Hence, it is of interest to explore the mechanical performance of this hydrogenated alloy with a dual-phase structure. Since powders were investigated in the present study, we carried micro-compression tests on the micropillars fabricated from the powders. Figure 3a compares the compressive engineering stress-strain curves of the pillars with a diameter of 500 nm and a height of 1 μm for the as-prepared and hydrogenated alloys. The GdCoAl sample has a yield strength of 2.5 GPa without any plasticity, indicating its intrinsic brittleness. In contrast, the hydrogenated sample shows a yield strength of 3.6 GPa and a plastic strain of over 70%, reversing the hydrogen embrittlement in many metals and alloys. Larger size sample with 5 μm in diameter was also investigated and the result is shown in the inset of Fig. 3a. A large strain of over 55% (without fracturing) is still observed, illustrating that such a brittle-to-ductile transition does not arise from the size effect. Figure 3b,c shows the morphology images after compression of the GdCoAl and GdCoAlH pillars with a diameter of 500 nm. The GdCoAl pillar shows the features of brittle fracture. In contrast, the GdCoAlH pillar was compressed into a flake without fracturing and obvious SBs are observed. The multiple SBs can be seen more clearly for the GdCoAlH pillar with 5 μm in diameter (Fig. 3e). Furthermore, we investigated different systems (GdNiAl and DyCoAlSi), and similar brittle-to-ductile transition induced by hydrogenation is observed (Supplementary Fig. 2). Figure 3f shows the excellent mechanical performances of the several hydrogenated rare earth (RE)-based AMAs compared with their hydrogen-free counterparts and other typical RE-based bulk metallic glasses (BMGs). It can be seen clearly that the SNDP-GH nanostructure enhances greatly both the strength and plasticity, overcoming the strength-ductility trade-off in AMAs. Meanwhile, these results imply that creating an SNDP structure could be an effective strategy to overcome the hydrogen embrittlement of many metal-hydrogen systems.

Fig. 3: Mechanical properties of the GdCoAl and GdCoAlH micropillars.

a Engineering stress-strain curves of the pillars with a diameter (Φ) of 500 nm and height of 1 μm. The inset shows the stress-strain curves of the GdCoAlH pillar with a diameter of 5 μm and height of 10 μm. b, c Morphology images after compression of GdCoAl and GdCoAlH pillars with a diameter of 500 nm. d, e Morphology images of the GdCoAlH pillar with a diameter of 5 μm before and after compression. f Summary of the stress versus strain for the typical rare earth-based amorphous micropillars and BMGs, the data of this study are also added.

Magnetic transition and magnetocaloric properties

Next, we investigated the effect of hydrogenation on the magnetic transition and magnetocaloric properties. Figure 4a presents the temperature dependence of the zero-field-cooling (ZFC) and field-cooling (FC) magnetization of the GdCoAl and GdCoAlH alloys. The same temperature dependence of the ZFC and FC curves is observed for GdCoAl, indicating a typical ferromagnetic transition. In contrast, for the GdCoAlH alloy, both the ZFC and FC magnetization curves show a peak around the Néel temperature T_N, below which bifurcation appears between the FC and ZFC branches. From the Curie-Weiss fit of the magnetic susceptibility far above the magnetic ordering temperature (Supplementary Fig. 4a), the paramagnetic Curie temperatures are determined to be 107 K and −27 K for the GdCoAl and GdCoAlH alloys, respectively. This verifies that ferromagnetic and antiferromagnetic exchange interactions dominate the magnetic structure in GdCoAl and GdCoAlH, respectively. The AC susceptibility χ (real part) curves under different frequencies from 13 to 9673 Hz are almost the same (Supplementary Fig. 4b), indicating further the antiferromagnetic transition nature, which is different from the obvious change of the peak position and intensity with increasing frequency in the Ho- and Dy-based AMAs showing spin-glass-like behavior¹⁹ (Supplementary Fig. 7a). The antiferromagnetic-to-paramagnetic transition arises from the GdH₂ hydride³¹. Besides, hydrogenation does not change the magnetic soft nature of the alloy, as indicated in Fig. 4b, which is beneficial for magnetic cooling applications.

Fig. 4: Magnetic transition and magnetocaloric effect of the GdCoAl and GdCoAlH powders.

a Temperature dependence of magnetization (M). b Magnetic hysteresis loops measured at 5 K. c Temperature dependence of magnetic entropy change (ΔS_M) under the maximum applied field of 2 and 5 T. d Summary of maximum magnetic entropy change (${-\triangle S}_{M}^{{pk}}$) (magnetic field change ΔH = 5 T) versus the peak tem_perature (T_p) for the typical Gd-based amorphous and dual-phase alloys, as well as the value of GdCoAlH. e Comparison of the mechanical (ultimate strength σ_u, yield strength σ_y, and strain ε) and magnetic pro_perties ($-\triangle {S}_{M}^{{PK}}$, T_p, coercivity H_c, and relative cooling power RCP) of the GdCoAl and GdCoAlH alloys and representative crystalline Gd and GdSiGe alloys.

From the isothermal magnetization curves (Supplementary Fig. 4c, d), the ΔS_M of the alloys can be evaluated from the Maxwell relations. As presented in Fig. 4c, the hydronated alloy shows a giant $-\triangle {S}_{M}^{{pk}}$ of 18.7 J kg⁻¹ K⁻¹ under a field change of 5 T, which is comparable to that of the prototype giant-MCE material Gd₅Si₂Ge₂^32,33 and is twice as large as that (9.1 J kg⁻¹ K⁻¹) of the GdCoAl AMA. A rescaled temperature $(\theta=(T-{T}_{p})/({T}_{r}-{T}_{p}))$ dependence of the normalized magnetic entropy change ${\varDelta S}_{M}/\triangle {S}_{M}^{{PK}}$ was analyzed to check the scaling behavior and illustrate the second-order nature (which is revealed from the Arrott-plot in Supplementary Fig. 4e, f)¹⁴, where T_p is the peak temperature of ΔS_M, and T_r is the reference temperature (${{\varDelta S}_{M}({T}_{r})/\triangle {S}_{M}^{{PK}}=}$ 0.7). As shown in the Supplementary Fig. 4i, all the MCE curves under different fields are collapsed into a single master curve for the GdCoAlH alloy with a dual-phase structure. To highlight the giant-MCE of the hydrogenated alloy, we compare present data with the $-\triangle {S}_{M}^{{pk}}$ values under a field change of 5 T for various Gd-based AMAs and related dual-phase alloys

Data availability

All data needed to evaluate the conclusions in the paper are present in the paper and/or the supplementary information. Data were also available from the corresponding author upon request.