Study on mechanisms of two-step hydrogen sorption in a MgH₂–TiCrMnFeZr high-entropy alloy composite

Abstract

High-entropy alloys (HEAs) featuring multi-element active sites show exclusive catalytic ability, as well as great potential to store hydrogen (H₂) at room temperature by adjusting the electronic and geometrical factors. Herein, a TiCrMnFeZr HEA was adopted to improve the hydrogen storage properties of magnesium hydride (MgH₂). It was demonstrated that MgH₂ provides substantial cushion protection for the crystal structure and hydrogen storage capacity of the TiCrMnFeZr HEA during ball milling, while the TiCrMnFeZr HEA exhibits superior catalytic ability for the dissociation of H–H and Mg–H bonds. In particular, the MgH₂-40 wt% TiCrMnFeZr composite can absorb and desorb 0.72 wt% H₂ even at room temperature. Moreover, MgH₂ starts to release hydrogen at 162 °C, and 90% of stored H₂ (∼3.7 wt%) can be released at 230 °C within 60 min. There is no capacity fading after 20 cycles at 300 °C for both TiCrMnFeZr HEA and Mg/MgH₂ phases in the composite, showing an outstanding cycling performance. Microstructure investigations reveal that the well-protected TiCrMnFeZr HEA particle surfaces act as the catalytic sites for the dissociation of Mg–H bonds because of their intrinsic multivalent electronic configuration, and also serve as the channels for hydrogen sorption in Mg/MgH₂. Such a method to design and synthesize high-performance Mg-based hydrogen storage composites and to provide H₂ in two steps paves a new way to realize their practical applications in the hydrogen energy field.

---

1. Introduction

Hydrogen (H₂) is perceived as one of the most promising sustainable energy carriers to address the problems caused by excessive consumption of traditional fossil fuels.¹ Effective storage and transportation are the pivotal bottlenecks that connect the integrated H₂ generation plants with dispersive H₂ usage scenarios, limiting its widespread applications in households and industries.² Compared with the 350/700 bar compressed H₂ storage in gas form and 20 K cryogenic H₂ storage in liquid form,³ storing H₂ in the solid state to form covalent/ionic compounds offers the advantages of high safety, fewer engineering challenges and cheap accessory equipment.⁴ Magnesium hydride (MgH₂) exhibits the advantages of high gravimetric and volumetric capacity (7.6 wt% and 110 g L⁻¹), abundant Mg minerals (2.5 wt% on Earth), and environmentally friendly products.⁵ However, the application of MgH₂ is still hampered by its high operating temperature and low desorption rate.⁶

Catalyst doping has been extensively investigated in the past few decades, which can increase active heterostructured interfaces for the dissociation of Mg–H and H–H bonds and accelerate H atom diffusion.^7,8 When catalyzed by different transition metals (TMs, for example Ti, Mn, Fe, Ni, etc.), the onset ab/de-sorption temperature and the kinetics of Mg/MgH₂ were significantly improved.^9–12 Khatabi et al. observed the decrease in the charge density of Mg and H through the doping of TMs by a density functional theory (DFT) study, in which the hybridization of s and d orbitals of Mg and TMs induced a new band in the band gap of MgH₂.¹³ Then, researchers studied the multicomponent catalyst to further optimize the electronic configuration of the composites. Xu et al. illustrated the catalytic effect of PdNi biatomic clusters on MgH₂ decomposition, whose onset desorption temperature was reduced to 154 °C.¹⁴ Zhou et al. introduced a 10 mol% amorphous Ti₄₀Cu₄₇Zr₁₀Sn₃ alloy into MgH₂, and the composite started to dehydrogenate at 150 °C.¹⁵

Generally, ball milling has been widely employed for the doping of catalysts into the MgH₂ matrix. With increasing ball-milling energy, the distribution of catalysts becomes more homogeneous, and the particle size of MgH₂ and catalysts turns finer. Therefore, nano-sized catalysts, with increased surface energy, are also prepared to improve their distribution.¹⁶ However, the delivery temperature and kinetic ab/de-sorption rate gradually decay with the thermal-induced ab/de-sorption processes,¹⁶ attributed to the self-aggregation and growth of the nanoparticles as well as the self-healing of active defects (dislocation, grain boundary, etc.).¹⁷ It is thus important to design new kinds of catalysts with high intrinsic catalytic ability and synthesize Mg-based hydrogen storage composites with superior cycling stability.

Recently, high-entropy alloys (HEAs), with homogeneous composition and adjustable components, have been well studied as a new class of catalysts for Mg-based hydrogen storage materials.^18–21 Li et al. studied the catalytic effect of CrMnFeCoNi, referred to as the “cocktail effect”, which exerted synergistic interactions between Cr, Mn, Fe, Co and Ni elements to improve the overall catalytic efficiency.¹⁸ Liu et al. prepared carbon-supported NiCoFeCuMg HEA nanocatalysts.¹⁹ During the hydrogen ab/de-sorption processes, new interphases Mg₂Ni(Cu)/Mg₂Ni(Cu)H₄, Co₃Fe₇ and carbon are generated, serving as a hydrogen pump, hydrogen gateway, and dispersing material, respectively. However, the intrinsic catalytic mechanism and properties of HEAs remain unclear due to the inevitable alloying reactions.²¹ Although increasing the catalyst content can enhance the formation of active heterostructured interfaces,²² the hydrogen storage capacity of the composites diminishes with excessive catalyst addition. In particular, for metal oxides, MgO is generated inevitably and irreversibly during the synthesis, H₂ absorption and desorption steps.^23,24 Recently, Luo et al. investigated the microstructures and H-storage properties of BCC-structured VTiCrFeM (M = Mn, Co, Sc, or Ni) HEAs, which can absorb hydrogen at room temperature without any pretreatment.²⁵ Ti–V-based alloys, lightweight alloys, and rare earth-based alloys exhibit excellent performance in specific hydrogen storage properties.^26,27 Besides, hexagonal close-packed (HCP) HEAs with a C14 Laves phase structure are a kind of hydrogen storage material that can deliver H₂ at room temperature, and the desorption plateau pressure is adjustable by controlling the components.²⁸ The (Ti, Zr or Nb) (V, Cr, Mn, Fe, Co, Ni, Cu, or Zn)₂ system, resulting in 440 alloys, is prone to solidify in the C14 Laves phase structure,²⁹ which can also serve as the effective high-entropy catalysts. However, their H₂ storage capacities gradually disappear during ball milling due to the destruction of the Laves phase crystal lattice.³⁰

In this work, the TiCrMnFeZr HEA, having no interphase between Mg and each element, was selected to investigate its intrinsic catalytic properties on Mg/MgH₂. We introduced additional quantities (20, 40, 60 wt%) of alloys into MgH₂via ball milling at a low speed of 250 rpm. Surprisingly, the HCP crystal lattice of the HEA was well preserved under the protective effect of MgH₂, as well as the hydrogen storage capacity. Moreover, MgH₂ shows significantly enhanced hydrogen storage properties owing to the pronounced catalytic effect of the TiCrMnFeZr HEA. This synergy induces a distinctive two-stage hydrogen supply behaviour in the composite material. At the room-temperature stage, the HEA phase can ab/de-sorb hydrogen of ∼0.7 wt%. Subsequently, MgH₂ begins to decompose at 162 °C and delivers 4 wt% hydrogen before 275 °C. In particular, the composite shows outstanding cycling stability. There is no capacity decay after 20 cycles, during which MgH₂ exhibits improved kinetics. This strategy of designing composite materials with two-step hydrogen release capability provides a viable pathway toward the practical application of Mg-based hydrogen storage materials.

2. Experimental details

2.1. Materials preparation

Commercial MgH₂ and TiCrMnFeZr HEA were purchased from Shanghai Mg Powder Technology Ltd and Baotou Zhongke Xuanda New Energy Technology Co. Ltd, respectively. MgH₂ was used as received. The TiCrMnFeZr HEA was used after activation, in which the TiCrMnFeZr HEA absorbed H₂ under 50 bar background pressure for 24 h. The initial composition of the TiCrMnFeZr HEA was confirmed by inductively coupled plasma (ICP), with the contents of Ti, Cr, Mn, Fe, and Zr at 27.1, 25.4, 33.6, 3.5, and 10.4 wt%, respectively.

The MgH₂-x wt% activated TiCrMnFeZr samples were prepared using a planetary ball mill (YXQM-1L, MITR Scientific and Technical Corporation, Changsha, China) with a stainless-steel jar (inner diameter: 58 mm; height: 50 mm). Typically, about 1 g of sample was processed with 40 g of stainless-steel balls (diameter: 10 mm; mass per ball: 4 g; quantity: 10 balls) under 15 bar hydrogen pressure for 6 h under 250 rpm at ambient temperature. All the materials were handled in a glove box filled with argon (99.999%, O₂ < 1 ppm, and H₂O < 1 ppm). For comparison, the pure ball-milled MgH₂ (BM MgH₂) and TiCrMnFeZr HEA (BM TiCrMnFeZr HEA) were obtained by milling as-received MgH₂ and activated TiCrMnFeZr under the same conditions.

2.2. Characterization

Inductively coupled plasma (ICP, iCAP 7600, Thermo) was conducted to confirm the composition of composites. X-ray diffraction (XRD, MiniFlex 600, Rigaku, Cu Kα) was performed at 40 kV/15 mA with a scanning speed of 10° min⁻¹ in the range of 10–80°. The samples were prepared in an Ar-filled glovebox and sealed in a custom-designed holder covered with an amorphous tape to avoid air exposure during the test. The morphologies and structures were studied by scanning electron microscopy (SEM, Mira3 and Rise-magna, TESCAN) equipped with an energy-dispersive X-ray spectrometer (EDS), transmission electron microscopy (TEM, FEI Talos F200X G2) and scanning transmission electron microscopy (STEM) with an accelerating voltage of 200 kV, respectively. For the sample preparation for SEM observations, the sample powders were stuck on the SEM stainless holder in a glovebox and sealed in a plastic box, which was rapidly transferred to the SEM instrument. For TEM characterization, the samples were ultrasonically dispersed in tetrahydrofuran (THF), and then the dispersion liquid was dropped on a copper grid in the glovebox, followed by the rapid transfer to the TEM instrument. An air-proof transfer vessel (provided by Kratos Analytical Ltd) was used to transfer the samples from the glovebox to X-ray photoelectron spectroscopy equipment (XPS, Kratos AXIS Ultra DLD, Al K_α/Mg K_α) to characterize the elemental valence and chemical bonding. The C–C peak at 284.8 eV is used to do calibration. DSC curves were obtained using a NETZSCH STA 449 F3 Jupiter with heating rates of 3, 5, 7, 10 °C min⁻¹ under flowing Ar from room temperature to 500 °C (the test durations were 158.3 min (3 °C min⁻¹), 95 min (5 °C min⁻¹), 67.9 min (7 °C min⁻¹), and 47.5 min (10 °C min⁻¹), respectively).

2.3. Tests of hydrogen storage properties

The hydrogen sorption behavior tests (isothermal ab/de-sorption tests and temperature-programmed desorption (TPD) measurements) were carried out in a Sievert-type pressure–composition–isothermal apparatus (PCI, HPSA-auto device, China), which was modified using the Benedict–Webb–Rubin (MBWR) equation. For the isothermal ab/de-sorption tests, ∼100 mg sample is loaded into the testing tube in a glove box and heated to the preset temperature by rapid heating (∼10 °C min⁻¹) and then maintained this temperature during the kinetic tests. The H₂ ab/de-sorption kinetics is measured at various temperatures for an initial H₂ pressure of 30 bar for absorption and ∼0.001 bar (vacuum for 1–2 s) for desorption. The thermodynamic properties of the composites were determined from the pressure–composition–isothermal (PCI) curve and van't Hoff fitting. The hydrogenated samples are heated from room temperature to 350 °C with a heating rate of ∼2 °C min⁻¹ under a primary vacuum to conduct TPD measurements. For cycling tests, the H₂ pressure/duration/temperature parameters were set at 15 bar/15 min/300 °C for re-hydrogenation and 0.25 bar/60 min/300 °C for dehydrogenation using another Sievert-type apparatus from CIQTEK (H-Sorb X, China). The H₂ storage capacity was calculated as wt% of the entire composite, including MgH₂ and the TiCrMnFeZr HEA.

3. Results and discussion

3.1. Structures of the TiCrMnFeZr HEA and TiCrMnFeZr–MgH₂ composites

First, the element component of the TiCrMnFeZr HEA was confirmed. According to the ICP results (Table S1†), the main elements are Ti, Cr, Mn, and Fe and Zr, which were introduced to the lattice to modify the lattice space to improve the hydrogen storage capacity and cycling stability.²⁹Fig. 1a illustrates the crystal structure of the TiCrMnFeZr HEA in the as-received state, in which all peaks belong to the C14 Laves phase structure. The periodic structure of the crystal identifies the homogeneous doping of Fe and Zr elements. Besides, the hydrogen storage capacity of the TiCrMnFeZr HEA was tested. As the pressure increased, hydrogen was absorbed with a capacity of 1.63 wt% at 50 °C (Fig. 1b). The absorption enthalpy was fitted with the plateau pressures obtained at 40, 60, 70 and 80 °C (Fig. S1†), which was determined to be 26.2 ± 0.1 kJ per mol H₂. The hydrogen absorption process includes interstitial solid solution (metal bulk), hydride nucleation and interstitial solid solution (hydride bulk) procedures, which are consistent with those of the typical C14 Laves phase hydrogen storage alloy.^31,32 The particle size of the HEA showed an obvious reduction after activation due to the expansion and crushing during the hydrogenation (Fig. S2†). The morphologies of the activated TiCrMnFeZr HEA were observed using SEM (Fig. 1c and S3†). The particles show irregular shape and obvious cracks, which may cause further crushing during ball milling. The size distribution is around 2–44 μm, and the average particle size is determined to be 12.6 μm. The corresponding EDS mappings clearly show that the compositing elements of Ti, Cr, Mn, Fe and Zr are well distributed with no segregation (Fig. 1d). The TiCrMnFeZr HEA was doped into MgH₂ with different additive amounts through ball milling, and a relatively low ball-milling speed of 250 rpm was used in this work.^33,34 According to the ICP results (Table S1†), the MgH₂ content in the MgH₂-40 wt% TiCrMnFeZr composite is ∼57.6 wt%. Considering the impurity in the MgH₂ raw materials and its oxidation during the ICP test, the amount of MgH₂ in the MgH₂-40 wt% TiCrMnFeZr composite is ∼60 wt%, indicating a homogeneous mixing of MgH₂ and TiCrMnFeZr in the composite. First, the morphologies and crystal structures of BM MgH₂ and BM TiCrMnFeZr HEA were studied. The particle size of BM MgH₂ and TiCrMnFeZr HEAs became smaller under the mechanical force (Fig. S4†), and the XRD patterns of BM MgH₂ and BM HEAs show amorphous characteristics because of their sensitivity to the compressive and shear stress (Fig. 2a and S5†).³⁵ Surprisingly, with the introduction of MgH₂, the TiCrMnFeZr crystal structure was well retained (Fig. 2a), indicating that MgH₂ served as the buffer and protection to relieve the collision between TiCrMnFeZr and the balls and walls of the jar during the ball milling.³⁶ Even with the high TiCrMnFeZr HEA to MgH₂ weight ratio of 3 : 2, the crystal structure of the HEA is still well retained. Meanwhile, the peak width of MgH₂ became more dilated and the intensity became lower with the increase of TiCrMnFeZr amount, implying that the crystal size of MgH₂ was further reduced. It's worth noting that lower MgH₂ amount means a higher mass ratio of stainless-steel balls to MgH₂; in addition, the TiCrMnFeZr HEA could also serve as the grinding agent to introduce more defects (dislocation and lattice boundary) into MgH₂.³⁷ Only peaks belonging to MgH₂ and TiCrMnFeZr HEAs can be observed, which derives from the absence of interphases between TMs (Ti, Cr, Mn, Fe, Zr) and Mg.³⁸ This unique composition can exclude the alloying reaction and support the exploration of HEA's intrinsic catalytic ability. TEM was applied to analyze the morphology and phase structure of the composite. Due to the heavy atomic mass of Ti, Cr, Mn, Fe, and Zr compared to Mg, the colour shows strong white in the HAADF image (Fig. 2b), which shows the HEA particles with distinct edges, as also seen in the BF image (Fig. 2c). The TiCrMnFeZr HEA was broken in various degrees, in which microparticles with a diameter of 2.6 μm (Fig. S6†) and nanoparticles with a diameter of 36 nm co-existed after ball milling. The insertion of the TiCrMnFeZr HEA into the MgH₂ matrix implies the good protective effect of MgH₂. From the STEM mapping, a clear boundary can be seen between Mg and TM elements (Fig. 2d). The sharp interfaces can also be seen in the HRTEM image (Fig. 2e). The interplanar spacing of 0.242 nm belongs to the C14 Laves phase with dark contrast in the BF image. Correspondingly, the lattice spacings of 0.279 nm, 0.225 nm, and 0.319 nm belong to Mg (100), MgH₂ (200) and MgH₂ (110), respectively. The Mg phase can be seen under electron beam irradiation and a high-vacuum atmosphere, which may be generated under the strong catalysis of the HEA.

Fig. 1 (a) XRD pattern of the as-received TiCrMnFeZr HEA, and (b) its hydrogen absorption curve at 50 °C. (c) SEM images and corresponding (d) elemental mapping of the activated HEA.

Fig. 2 (a) XRD patterns of the ball-milled TiCrMnFeZr HEA and MgH₂-x wt% TiCrMnFeZr composites (x = 20, 40, 60). (b) HAADF image, (c) BF image and (d) corresponding elemental mapping, and (e) HRTEM image of the MgH₂-40 wt% TiCrMnFeZr composite.

3.2. Hydrogen storage properties of MgH₂–TiCrMnFeZr HEA composites

According to the XRD results, the crystal structure of the TiCrMnFeZr HEA was well retained, which makes it possible to absorb H₂ after ball milling. Fig. 3a shows the PCI curves of composites with different doping amounts of TiCrMnFeZr at 50 °C. The BM TiCrMnFeZr HEA showed H₂ solubility properties as the pressure increased, taking ∼0.5 wt% at a pressure of 70 bar. This is in line with the amorphous H₂ storage alloy.³⁹ In contrast, TiCrMnFeZr HEAs can store ∼0.35 wt%, 0.72 wt%, and 0.95 wt% under 60 bar in the composites with 20 wt%, 40 wt%, and 60 wt% HEA contents, respectively, and exhibit three-stage H₂ absorption behavior. The TiCrMnFeZr HEA in the composite with 60 wt% doping amount shows a slight capacity fading (∼1 wt% with no fading), which indicates that the partial surface structure is destroyed due to the decrease in the protective MgH₂ matrix. The TPD curves of the composites were measured to characterize the catalytic ability of the TiCrMnFeZr HEA to MgH₂ (Fig. 3b). With the increasing temperature, MgH₂ doped with a 40 wt% TiCrMnFeZr HEA starts to release H₂ first at 162 °C, 125 °C lower than pure BM MgH₂ (Fig. 3c and d). MgH₂ with 20 wt% and 40 wt% TiCrMnFeZr composites start to decompose at 184 and 174 °C, respectively. Besides, it can be observed that the TPD curve of the MgH₂–60 wt% TiCrMnFeZr composite exhibits a sloped stage compared with other samples, deriving from the H atoms encapsulated in the amorphous crystal structure, which may affect the catalytic ability of TiCrMnFeZr.⁴⁰ It is consistent with the TiCrMnFeZr PCI absorption capacity shown in Fig. 3a.

Fig. 3 (a) PCI and (b) TPD curves of the ball-milled TiCrMnFeZr HEA and MgH₂–x wt% TiCrMnFeZr composites (x = 20, 40, 60), and (c and d) corresponding enlarged parts.

To further illuminate the catalytic performances of TiCrMnFeZr, the isothermal hydrogen storage properties of MgH₂–40 wt% TiCrMnFeZr HEA and BM MgH₂ were compared. Fig. 4a shows the isothermal dehydrogenation curves of the MgH₂–40 wt% TiCrMnFeZr composite at temperatures of 170, 190, 210, 230, and 250 °C. It can be seen that the composite can fully desorb ∼4.1 wt% H₂ within 20 min at 250 °C and the desorption finishes at a final H₂ pressure of ∼0.2 bar. Compared with the TPD curve, the measurable capacity loss (∼0.3 wt%) is derived from the irreversible side-reaction of MgH₂ with residual impurities (e.g., oxides in the raw materials) to form MgO. Besides, isothermal desorption measurements and data acquisition were preceded by a brief evacuation (1–2 s) at the target temperature, which will also result in differences in hydrogen sorption capacity. At an isothermal temperature of 210 °C, ∼2 wt% H₂ is released within 60 min. Even at the low temperatures of 170 and 190 °C, ∼0.3 and 0.9 wt% H₂ is released within 60 min, respectively. In contrast, 40 min was needed for the BM MgH₂ to desorb ∼6.2 wt% H₂ at 350 °C (Fig. 4b). There is almost no dehydrogenation from BM MgH₂ at 290 °C within 60 min, while MgH₂–40 wt% TiCrMnFeZr can fully desorb at 290 °C within 4 min (Fig. S7†). The Mg in the desorbed MgH₂–40 wt% TiCrMnFeZr composite can absorb H₂ at 50 °C, and absorb ∼3 wt% H₂ within 60 min at 100 °C, accounting for 75% of the capacity (Fig. 4c). Hydrogenation can be finished within 5 min at 200 °C, but pure BM MgH₂ can only absorb 3.4 wt% H₂ in 5 min at 200 °C, accounting 52% of its reversible capacity, and is fully hydrogenated in 10 min at 290 °C (Fig. 4d). The results indicate that TiCrMnFeZr significantly improves the hydrogen desorption and absorption kinetics of MgH₂. The kinetic barrier energy and enthalpy were calculated to further investigate the dehydrogenation process of the MgH₂–40 wt% TiCrMnFeZr composite based on the DSC analyses. The DSC curves of the MgH₂–40 wt% TiCrMnFeZr composite and BM MgH₂ with different heating rates were compared (Fig. S8†). The curves of the MgH₂–40 wt% TiCrMnFeZr composite show two peaks, which may originate from the big particle size and uneven distribution of the TiCrMnFeZr HEA. The peak desorption temperatures of the MgH₂-40 wt% TiCrMnFeZr composite decreased by ∼98.7 °C compared to that of the BM MgH₂ at a heating rate of 3 °C min⁻¹. According to Kissinger's method, the apparent activation energies (E_a) of the MgH₂–40 wt% TiCrMnFeZr composite and BM MgH₂ were obtained. The results show that the E_a of the MgH₂–40 wt% TiCrMnFeZr composite is 69.2 ± 1.6 and 104.7 ± 6.2 kJ per mol H₂ for the two peaks, respectively, much lower than that of the BM MgH₂ (142.4 ± 21.4 kJ per mol H₂). In order to evaluate the apparent activation energy of whole composites, dehydrogenation kinetic models of composites were fitted with nine different kinetic models proposed by Sharp and Jones,^41,42 and the activation energy was calculated according to the suitable kinetic model (more details are given in Table S2†). As shown in Fig. S9,† nine curves using different kinetic models were obtained based on the isothermal dehydrogenation curves of the MgH₂–40 wt% TiCrMnFeZr composite at 250, 270, 290 °C. The dehydrogenation kinetic model should be the R2 or R3 model, which showed a very close value of R² to 1. The R2 model is further verified through plotting g(α) (where α is the fraction reacted in time and g(α) is the function of α) related to R2 (g(α) = 1 − (1 − α)^1/2) against the reaction time at 230, 250, 270 and 290 °C when α ranges from 0.2 to 0.7, which exhibits a strong linear relationship (R² is more than 0.998) as shown in Fig. 5a. The desorption activation energy (E_d) was calculated to determine the energy barrier for the dehydrogenation process of MgH₂–40 wt% TiCrMnFeZr according to the Arrhenius equation. According to isothermal desorption profiles, E_d was determined to be 76.1 ± 3.2 kJ per mol H₂ through linear fitting of ln k and 1000/T (R² > 0.99) (Fig. 5b). Similarly, the E_d of BM MgH₂ was calculated to be 154.1 ± 5.8 kJ per mol H₂ (Fig. S10†). Thus, the addition of TiCrMnFeZr HEAs can significantly reduce the apparent activation energy of dehydrogenation for MgH₂ and improve the kinetics of the composites. Subsequently, the PCI data were collected at 270, 290, and 310 °C, showing that the MgH₂–40 wt% TiCrMnFeZr composite can fully absorb and desorb H₂ without capacity loss (Fig. 5c). Even at 270 °C, the hysteresis between absorption and desorption is only 0.1 bar, implying the good hydrogen sorption kinetics. The ab/de-sorption enthalpy was fitted with the plateau pressure, which was 75.6 ± 0.9 and 78.9 ± 2.3 kJ per mol H₂, respectively (Fig. 5d). The enthalpy values are very close to the theoretical and test results in the literature,^43,44 meaning that the TiCrMnFeZr HEA showed no thermodynamic improvement on MgH₂. This is consistent with the structural characterization studies indicating that the TiCrMnFeZr HEA has no effect on the structure and composition of the MgH₂ phase.

Fig. 4 Isothermal, (a and b) desorption and (c and d) absorption curves of the (a and c) MgH₂–40 wt% TiCrMnFeZr composite and (b and d) pure BM MgH₂ at different temperatures.

Fig. 5 (a) Time dependence of the kinetic modelling equation g(α) = 1 − (1 − α)^1/2 for the MgH₂–40 wt% TiCrMnFeZr composite with 0.2 < α < 0.7 at different temperatures, and (b) the Arrhenius plot for the dehydriding kinetics of the MgH₂–40 wt% TiCrMnFeZr composite. (c) PCI curves of the MgH₂–40 wt% TiCrMnFeZr composite at 270, 290 and 350 °C, and the corresponding (d) van't Hoff fitting plot.

An automated cycling test was performed to evaluate the reversible properties of the MgH₂–40 wt% TiCrMnFeZr composite. Due to the limitation of the instrument's chamber volume, the initial H₂ pressures for absorption and desorption were set at 15 and 0.25 bar, respectively. The composite could achieve reversible de/re-hydrogenation with a hydrogen capacity of up to 3.85 wt%, and showed outstanding cycling stability with no capacity decay during the first 20 hydrogen absorption and desorption cycles (Fig. 6a and b). The final working H₂ pressures for absorption and desorption are 11.7 and 0.54 bar, respectively. As stated, the cycling tests were performed under isothermal conditions with fixed pressure parameters and cycle durations to evaluate the stability, and hydrogen cannot be fully absorbed and desorbed due to the kinetic limitations. More interestingly, an enhancement of the desorption kinetics can be clearly observed in the first 20 cycles. In view of the stable interfaces and weak alloying reactions between elements in the TiCrMnFeZr HEA and MgH₂, we assume that the HEA with good crystallinity shows stronger catalytic ability. During the ball milling process, the surface structure of the HEA underwent slight but inevitable damage, which was gradually healed during thermal-induced cycling. As the cycling test progressed, the kinetics showed a decay caused by the grain growth.⁴⁵ Besides, the cycling performance of the TiCrMnFeZr HEA was also tested (Fig. 6c). The hydrogen storage content and plateau pressure of the HEA exhibit good stability, and there is no toxification from Mg under thermal driving force. The superior hydrogen storage performances, including high initial desorption rate, low onset desorption temperature and stable cycling performance of the MgH₂–40 wt% TiCrMnFeZr composite, remarkably exceed those of the state-of-the-art MgH₂-alloy systems reported in the literature (Fig. 6d). Other hydrogen storage properties are also compared in details and listed in Table S3.†

Fig. 6 (a) MgH₂ cycling profiles of the MgH₂–40 wt% TiCrMnFeZr HEA composite at 300 °C and (b) its kinetic desorption rate at different cycles (1st, 10th, 20th, 30th, and 50th), and (c) TiCrMnFeZr HEA cycling stability tests of the MgH₂–40 wt% TiCrMnFeZr HEA composite at room temperature, (d) comparison of the onset desorption temperature and initial desorption rates of the MgH₂–40 wt% TiCrMnFeZr HEA composite with those of other alloy-catalyzed MgH₂ systems reported in the literature (*: the initial desorption rates summarized at different temperatures, a: 230 °C, b: 250 °C, and c: 275 °C).

3.3. Catalytic mechanism of TiCrMnFeZr on the hydrogen storage properties of MgH₂

Due to the strong catalytic effect of TiCrMnFeZr on Mg/MgH₂, the composites exhibit improved hydrogen ab/de-sorption kinetics. The electronic configuration of TiCrMnFeZr played an important role in the dissociation of H–H and Mg–H bonds. First, XPS signals of the TiCrMnFeZr HEA and Mg–HEA composites were collected to explore the catalytic mechanism. As shown in Fig. 7a and S11,† the photoelectron signals from five metals (Ti, Cr, Mn, Fe, and Zr) were observed in the activated HEA sample, consistent with the EDS and STEM mapping results. Different from pure Ti metal, the Ti element in the HEA exhibits a multivalent state, which can serve as the electron transfer bridge between H₂ and Mg and thereby accelerate the dissociation of the H–H bond and Mg–H bond.⁴⁶ Similarly, Cr, Fe and Zr also show multivalent states due to the homogeneous mixing and interaction in the HEA bulk. Unfortunately, we cannot detect any signal for the catalytic alloy in the composites (Fig. S12†), even though the addition amount is as high as 40 wt%. Given that XPS is a surface-sensitive technique with a detection depth of around 10 nm, the absence of the HEA alloy signal means that the surface of the HEA is fully covered by MgH₂, which well protects the crystal structure of the alloy. The phase change of the composites at different states is further detected (Fig. 7b). The XRD pattern of the as-synthesized composite shows two sets of peaks belonging to MgH₂ and the C14 Laves phase, respectively. After H₂ de/ab-sorption, the peaks from MgH₂ become sharper, implying its crystallinity gets improved due to the healing effect during the thermal treatment.⁴⁷ Then, the MgH₂ phase transformed into Mg after dehydrogenation. There is no new phase generated during the processes similar to the ball milling process, verifying the stability between Mg and HEA alloy elements at temperatures up to 300 °C. Meanwhile, the interfacial structure of the MgH₂–40 wt% TiCrMnFeZr composite after 20 sorption cycles was examined by TEM. As shown in Fig. 7c and d, the clear interface still remained, and there is no obvious TM element diffusion into the MgH₂ matrix. In the HRTEM images (Fig. 7e and f), the clear crystal lattices belonging to MgH₂, HEA and MgO can be found near the interfaces. MgO is generated in the interfaces between the TiCrMnFeZr HEA and MgH₂, which means that (1) MgO is easy to form under the catalysis of the TiCrMnFeZr HEA;⁴⁸ or (2) MgH₂ is prone to desorb under the catalysis of the TiCrMnFeZr HEA, then the in situ formed Mg is oxidized immediately.⁴⁹ Besides, the TiCrMnFeZr HEA particles' microstructure in the composite remains stable during the cycling under the protection and confinement of Mg/MgH₂ (Fig. S13 and S14†). In summary, the multivalent electronic configuration of the HEA delivers its high intrinsic catalytic ability on Mg/MgH₂. The TiCrMnFeZr HEA with stable interfaces ensures the continuous catalysis to the formation and decomposition of MgH₂. Simultaneously, the utilization of the HEA also provides a simplified procedure to synthesize multivalent catalysts.^46,50 Based on the isothermal de/ab-sorption data (Fig. S15†), we confirmed the most suitable kinetic models for the de/ab-sorption phases to imply the catalytic mechanism of the HEA and de/ab-sorption processes of the MgH₂–40 wt% TiCrMnFeZr composite. As shown in Fig. 8a and b, D2 and D4 are two-dimensional diffusion and three-dimensional diffusion models, respectively.⁴¹ It means that, in the beginning of desorption, the TiCrMnFeZr HEA exhibits a strong catalytic ability to break Mg–H bonds and to promote the formation of H–H bonds. Consequently, a concentration gradient appears near the interfaces, which greatly drives the further decomposition of MgH₂ controlled by the H atom diffusion to the interfaces.^51,52 In contrast, for BM MgH₂ (Fig. S10†), the A3, three-dimensional nucleation and growth models, indicates that Mg can only form inside the MgH₂ particles at temperature of 290 °C.⁴⁴ For the hydrogen absorption at 50 °C and 100 °C, the hydrogenation is controlled by the H atoms' three-dimensional diffusion (D3/D4) (Fig. 8c and d), indicating that TiCrMnFeZr HEA can break the H–H bonds at low temperature and play the role of “spillover” to promote the H atom diffusion in three dimensions.⁵³ In particular, the plateau pressure of the TiCrMnFeZr HEA is ∼18 bar at 50 °C (Fig. 1b), indicating that it can also serve as a “hydrogen pump” to improve the H diffusion into the Mg bulk. The dehydrogenation kinetic model for BM MgH₂ should be D1/D2 (one/two-dimensional diffusion) at 250 °C (Fig. S16†), indicating that the H₂ molecules dissociate at the surface of Mg particles under the effect of high temperature and the H atoms diffuse along the grain boundaries.⁵⁴

Fig. 7 (a) High-resolution Ti 2p XPS spectra of the TiCrMnFeZr HEA after activation. (b) XRD patterns of the MgH₂–40 wt% TiCrMnFeZr composite in different states (as-synthesized, absorbed and desorbed). (c) HAADF image and corresponding (d) STEM mapping, and (e and f) HRTEM images of the MgH₂–40 wt% TiCrMnFeZr composite after 50 cycles.

Fig. 8 (t/t_0.5)_theo.vs. (t/t_0.5)_exp. of the MgH₂–40 wt% TiCrMnFeZr composite at (a) 170 °C and (b) 210 °C for desorption, and (c) 50 °C and (d) 100 °C for absorption based on various kinetic models.

Based on the analyses above, the hydrogen sorption mechanism of the composite is schematically displayed in Fig. 9. Initially, MgH₂ and TiCrMnFeZr HEAs establish sufficient contact through ball milling. The crystal structure of the HEA was well protected with MgH₂ serving as a buffer layer, enabling the HEA to deliver H₂ at room temperature. Besides, due to multivalent elemental states, the HEA showed an outstanding catalytic effect on the hydrogen sorption in Mg/MgH₂. As a result, MgH₂ started to release hydrogen at 162 °C and completely desorbed within 20 min at 250 °C. The whole composite showed superior cycling stability with no capacity decay in 20 cycles for the two-step hydrogen supply.

Fig. 9 Schematic diagram of the structure and hydrogen sorption mechanisms in the MgH₂–TiCrMnFeZr HEA composite.

4. Conclusions

In this work, the TiCrMnFeZr HEA was introduced into MgH₂ through ball milling to obtain a composite with improved hydrogen storage performance in a wide temperature range. With the cushioning protection effect of MgH₂, the crystal structure of HEA was well protected during ball milling, as along with its hydrogen storage capacity. In particular, the MgH₂–40 wt% TiCrMnFeZr composite can reversibly absorb and desorb 0.72 wt% H₂ at room temperature. Meanwhile, the TiCrMnFeZr HEA also improved the hydrogen ab/de-sorption performance of MgH₂, which started to release hydrogen at 162 °C, 125 °C lower than pure MgH₂ under the same ball milling conditions. 90% of the stored hydrogen (∼3.7 wt%) can be released at 230 °C within 60 min, and complete dehydrogenation can be achieved within 20 min at 250 °C. There is no capacity fading after 20 cycles at 300 °C for both the TiCrMnFeZr HEA and MgH₂, during which the desorption of MgH₂ exhibited an improved kinetics. The intrinsic multivalent electronic configuration and structural stability of the TiCrMnFeZr HEA provide active interfacial sites for the dissociation of Mg–H and H–H bonds, and it also acts as the hydrogen channels to improve H atom diffusion in three dimensions during the hydrogenation. In a possible real application scenario, the H₂ absorbed in the TiCrMnFeZr HEA can be easily released and supplied to the fuel cell at room temperature, while the waste heat generated from the fuel cell (∼180–220 °C for the high temperature proton exchange membrane fuel cell, HT-PEMFC) can be used as the heat resource for MgH₂ decomposition. This method offers another promising route for the application of Mg-based hydrogen storage materials as the safe and reliable hydrogen supply source.

Data availability

The data supporting this article have been included as part of the ESI.†

Author contributions

Yinghui Li: conceptualization, methodology, data curation of XRD, SEM, TEM, XPS, DSC, and H₂ storage tests, formal analysis, validation, visualization, and writing – original draft; Yingyan Zhao: conceptualization, methodology, formal analysis, and validation; Shusheng Cao: data curation of H₂ storage tests; Zi Li: data curation of H₂ storage and TEM tests; Yueqing Shen: data curation of SEM tests; Yasemen Kuddusi: data curation of XPS tests; Chong Lu: data curation of TEM tests; Xi Lin: supervision and funding acquisition; Prof. Andreas Züttel: validation, supervision, project administration, and funding acquisition; Chongnan Ye: validation, supervision, funding acquisition and writing – review & editing; Prof. Jianxin Zou: conceptualization, methodology, validation, supervision, project administration, funding acquisition and writing – review & editing.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the National Key R&D Program of China (2022YFB3803700), National Natural Science Foundation of China (52201266, 52171186), Young Elite Scientists Sponsorship Program by CAST (2023QNRC001), and Science and Technology Projects of China Minmetals Corporation (2021ZXA03). Y. L. acknowledges the support from the China Scholarship Council (CSC) and Shanghai Jiao Tong University “Zhiyuan Honor Program” for doctoral students. C. Y. acknowledges the support from SNSF Fellowships (TMPFP2-216980). The authors appreciate the support from the Center of Hydrogen Science, Shanghai Jiao Tong University, China ENFI Engineering Corporation, and the Laboratory of Materials for Renewable Energy (LMER), École Polytechnique Fédérale de Lausanne Valais/Wallis.

References

1. T. He, P. Pachfule, H. Wu, Q. Xu and P. Chen, Nat. Rev. Mater., 2016, 1, 16059.
2. I. Staffell, D. Scamman, A. Velazquez Abad, P. Balcombe, P. E. Dodds, P. Ekins, N. Shah and K. R. Ward, Energy Environ. Sci., 2019, 12, 463–491.
3. M. R. Usman, Renewable Sustainable Energy Rev., 2022, 167, 112743.
4. M. D. Allendorf, V. Stavila, J. L. Snider, M. Witman, M. E. Bowden, K. Brooks, B. L. Tran and T. Autrey, Nat. Chem., 2022, 14, 1214–1223.
5. Y. Sun, C. Shen, Q. Lai, W. Liu, D. W. Wang and K. F. Aguey-Zinsou, Energy Storage Mater., 2018, 10, 168–198.
6. H. Wang, J. Li, X. Wei, Y. Zheng, S. Yang, Y. Lu, Z. Ding, Q. Luo, Q. Li and F. Pan, Adv. Funct. Mater., 2024, 34, 2406639.
7. J. Cui, J. Liu, H. Wang, L. Ouyang, D. Sun, M. Zhu and X. Yao, J. Mater. Chem. A, 2014, 2, 9645–9655.
8. B. Han, J. Wang, J. Tan, Y. Ouyang, Y. Du and L. Sun, J. Energy Storage, 2024, 83, 110612.
9. H. Guan, Y. Lu, J. Liu, Y. Ye, Q. Li and F. Pan, ACS Catal., 2024, 14, 17159–17170.
10. Y. Chen, H. Zhang, F. Wu, Z. Sun, J. Zheng, L. Zhang and L. Chen, Trans. Nonferrous Met. Soc. China, 2021, 31, 3469–3477.
11. H. Yang, X. Sun, Q. Luo, Y. Lu, Q. Li and F. Pan, Scr. Mater., 2024, 239, 115782.
12. Z. Lan, Z. Liu, H. Liang, W. Shi, R. Zhao, R. Li, Y. Fan, H. Liu and J. Guo, J. Mater. Sci. Technol., 2024, 196, 12–24.
13. M. El Khatabi, M. Bhihi, S. Naji, H. Labrim, A. Benyoussef, A. El Kenz and M. Loulidi, Int. J. Hydrogen Energy, 2016, 41, 4712–4718.
14. N. Xu, K. Wang, Y. Zhu and Y. Zhang, Adv. Mater., 2023, 35, 2303173.
15. C. Zhou, R. C. Bowman Jr, Z. Z. Fang, J. Lu, L. Xu, P. Sun, H. Liu, H. Wu and Y. Liu, ACS Appl. Mater. Interfaces, 2019, 11, 38868–38879.
16. X. L. Zhang, Y. F. Liu, X. Zhang, J. J. Hu, M. X. Gao and H. G. Pan, Mater. Today Nano, 2020, 9, 100064.
17. F. Bu, A. Wajid, N. Yang, M. Gu, X. Zhao, L. Huang, X. Ji, S. Ding, Y. Cheng and J. Zhang, J. Mater. Chem. A, 2024, 12, 12190–12197.
18. L. Wang, L. Zhang, X. Lu, F. Wu, X. Sun, H. Zhao and Q. Li, Chem. Eng. J., 2023, 465, 142766.
19. Y. Liu, M. Yue, Y. Guo, Y. Jiang, Y. Sun, L. Feng and Y. Wang, J. Magnesium Alloys, 2025, 13, 1232–1242.
20. J. Zhang, H. Liu, C. Zhou, P. Sun, X. Guo and Z. Z. Fang, J. Mater. Chem. A, 2023, 11, 4789–4800.
21. H. Wan, X. Yang, S. Zhou, L. Ran, Y. Lu, Y. a. Chen, J. Wang and F. Pan, J. Mater. Sci. Technol., 2023, 149, 88–98.
22. L. Xie, Y. Liu, X. Zhang, J. Qu, Y. Wang and X. Li, J. Alloys Compd., 2009, 482, 388–392.
23. Z. G. Huang, Z. P. Guo, A. Calka, D. Wexler, C. Lukey and H. K. Liu, J. Alloys Compd., 2006, 422, 299–304.
24. R. Zou, J. Adedeji Bolarin, G. Lei, W. Gao, Z. Li, H. Cao and P. Chen, Chem. Eng. J., 2022, 450, 138072.
25. L. Luo, H. Han, D. Feng, W. Lv, L. Chen, L. Li, T. Zhai, S. Liu, S. Sun, Y. Li, W. Pei, J. Cui and Y. Li, Renewables, 2024, 2, 138–149.
26. Z. Ding, Y. Li, H. Jiang, Y. Zhou, H. Wan, J. Qiu, F. Jiang, J. Tan, W. Du, Y. Chen, L. Shaw and F. Pan, Interdiscip. Mater., 2025, 4, 75–108.
27. X. Zhang, Z. Zhang, C. Xu, X. Xing, M. Wei, B. Cao and T. Liu, Chem. Eng. J., 2024, 499, 156643.
28. R. Li, A. Penmathsa, T. Sun, N. Gallandat, J. Li, J. Park, H.-J. Kim, P. Kim, N. Yoon, J.-H. Jang and A. Züttel, Int. J. Hydrogen Energy, 2024, 72, 687–693.
29. J. B. Ponsoni, V. Aranda, T. d. S. Nascimento, R. B. Strozi, W. J. Botta and G. Zepon, Acta Mater., 2022, 240, 118317.
30. J. Huot, H. Enoki and E. Akiba, J. Alloys Compd., 2008, 453, 203–209.
31. K. Young, T. Ouchi, B. Huang, B. Reichman and M. A. Fetcenko, Int. J. Hydrogen Energy, 2011, 36, 12296–12304.
32. G. Andrade, B. H. Silva, G. Zepon and R. Floriano, Int. J. Hydrogen Energy, 2024, 51, 246–254.
33. Y. Zheng, L. Zhang, H. Zheng, J. Chen, Z. Chen, Y. Liu, J. Tu and C. Gu, J. Colloid Interface Sci., 2025, 685, 65–72.
34. A. Baran, T. R. Jensen and M. Polański, J. Energy Storage, 2024, 103, 114271.
35. M. S. El-Eskandarany, F. Al-Ajmi, M. Banyan and A. Al-Duweesh, Int. J. Hydrogen Energy, 2019, 44, 26428–26443.
36. S. Nachev, P. de Rango, N. Skryabina, A. Skachkov, V. Aptukov, D. Fruchart and P. Marty, Int. J. Hydrogen Energy, 2015, 40, 17065–17074.
37. F. Yin, Z. Chen, T. Si, D. Liu, Y. Li, H. Li and Q. Zhang, J. Alloys Compd., 2024, 997, 174822.
38. H. Okamoto and T. Massalski, Binary Alloy Phase Diagrams, ASM International, Materials Park, OH, USA, 1990, vol. 12, pp. 3528–3531.
39. L. J. Huang, H. J. Lin, H. Wang, L. Z. Ouyang and M. Zhu, J. Alloys Compd., 2023, 941, 168945.
40. Z. Yuan, Y. Sui, Q. Yuan, Z. Qi, T. Zhai, X. Li and T. Li, Int. J. Hydrogen Energy, 2023, 48, 11340–11351.
41. J. H. Sharp and S. A. Wentworth, Anal. Chem., 1969, 41, 2060–2062.
42. L. F. Jones, D. Dollimore and T. Nicklin, Thermochim. Acta, 1975, 13, 240–245.
43. J. F. Stampfer Jr, C. Holley Jr and J. Suttle, J. Am. Chem. Soc., 1960, 82, 3504–3508.
44. Y. Li, L. Ren, Y. Yao, Y. Zhao, H. Xu, Z. Li, Z. Li, X. dai, Y. Tian, S. Cao, X. Lin, C. Ye, A. Züttel and J. Zou, Adv. Funct. Mater., 2025, 35, 2417915.
45. P. Rizo-Acosta, F. Cuevas and M. Latroche, J. Mater. Chem. A, 2019, 7, 23064–23075.
46. J. Cui, H. Wang, J. Liu, L. Ouyang, Q. Zhang, D. Sun, X. Yao and M. Zhu, J. Mater. Chem. A, 2013, 1, 5603–5611.
47. N. Lobo, A. Takasaki, K. Mineo, A. Klimkowicz and K. Goc, Int. J. Hydrogen Energy, 2019, 44, 29179–29188.
48. C. Peng, Y. Li and Q. Zhang, Scr. Mater., 2024, 248, 116149.
49. Y. Li, L. Ren, Y. Yao, Y. Zhao, H. Xu, Z. Li, Z. Li, X. Dai, Y. Tian, S. Cao, X. Lin, C. Ye, A. Züttel and J. Zou, Adv. Funct. Mater., 2025, 35, 2417915.
50. L. Zhang, X. Xiao, C. Xu, J. Zheng, X. Fan, J. Shao, S. Li, H. Ge, Q. Wang and L. Chen, J. Phys. Chem. C, 2015, 119, 8554–8562.
51. D. Korablov, T. K. Nielsen, F. Besenbacher and T. R. Jensen, Powder Diffr., 2015, 30, S9–S15.
52. S. Dong, C. Li, J. Wang, H. Liu, Z. Ding, Z. Gao, W. Yang, W. Lv, L. Wei, Y. Wu and H. Li, J. Mater. Chem. A, 2022, 10, 22363–22372.
53. Y. Qi, Z. Zhang, Q. Tang, J. Liu, R. Shi, J. Zhang, Y. Liu, J. Wang, J. Zhang, S. Chen and Y. Zhu, Chem. Mater., 2024, 36, 6288–6298.
54. B. Rathi, S. Agarwal, K. Shrivastava, H. Miyaoka, T. Ichikawa, M. Kumar and A. Jain, Int. J. Hydrogen Energy, 2025, 120, 213–224.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5ta03497c

---