Introduction

Hydrogen is a potential clean energy carrier for developing a carbon-neutral system and of great significance for realizing a global transition to a sustainable energy economy. Unfortunately, the absence of a safe and efficient hydrogen storage method remains a major bottleneck for the extensive utilization of hydrogen energy1,2,3,4. In comparison with hydrogen storage in pressurized tanks or cryogenic containers, the reversible storage of hydrogen into solid-state light-weight metal hydrides holds the intrinsic advantages of high safety, low-cost, and exceptional hydrogen storage capacity5,6,7,8. Among them, magnesium hydride (MgH2) that has a gravimetric and volumetric hydrogen density of 7.6 wt.% and 110 kg m−3, respectively, and excellent reversibility, is particularly favored9,10,11,12. Large energy input, however, is required to drive hydrogen storage reaction of metal hydrides due to their high thermodynamic stability and activation energy13,14,15. Particularly, a theoretical operating temperature of ~280 °C should be supplied to realize reversible hydrogen storage of MgH2 and the extra high kinetic barrier would lead to even higher operating temperature of 400 °C16,17. As a result, the reversible hydrogen storage of MgH2 mainly relies on external energy input by intricate electric heating apparatus that would unavoidably not only decrease the hydrogen storage capacity of the whole system but also induce high energy consumption and costs for their practical large-scale applications.

In order to solve this critical issue, we have recently proposed the concept of solar-driven reversible hydrogen storage of metal hydrides by using solar energy as the sustainable and unlimited energy source18,19. It is realized by coupling the photothermal effect of light adsorber (i.e., MXene and Cu) that could generate heat for elevating the temperature of MgH2 and catalytic effect of thermochemical catalysts (i.e., Ti/TiHx) that could decrease the operating temperature of MgH2 by reducing kinetic barriers. Undoubtedly, the match of the photothermal effect and the catalytic effect is a prerequisite for realizing solar-driven reversible hydrogen storage of MgH2. The light adsorber and thermal catalysts are physically separated and the phase separation of thermochemical catalysts, light adsorber, and MgH2, however, inevitably leads to the delay of the transfer of thus-generated heat from photothermal materials towards catalytic sites and the loss of heat, which results in poor catalytic activity and light-to-heat conversion efficiency. Moreover, both the photothermal effect of MXene and Cu and the catalytic effect of Ti/TiHx is limited and hence an extremely high light intensity of 4.0 W cm−2 is required for driving reversible hydrogen storage reaction of MgH2.

Theoretically, it is an ideal solution to enhance solar-driven hydrogen storage performance of MgH2 by introducing a single-component phase that simultaneously holds photothermal and catalytic effects. Herein, an in situ alloying reaction between MgH2 and CuNi is developed to achieve the integration of the photothermal effect with full-spectrum light absorption and the catalytic effect that could significantly lower the apparent activation energy of hydrogen storage reaction. Under UV–vis irradiation, the strong localized surface plasmon resonance (LSPR) effect of CuNi alloys would generate localized heat at the catalytic sites of CuNi alloys, leading to superior initial H2 desorption performance than that driven by traditional heating. Together with the initial dehydrogenation process, the in situ alloying reaction between Mg and CuNi results in simultaneous atomic reconstruction, leading to the formation of Mg2Ni(Cu) ternary alloy that has the metallic properties and the combined effect of intra/inter-band transitions. Therefore, an over 85% absorption in the entire spectrum could be achieved for MgH2 catalyzed by CuNi after the initial dehydrogenation process, which, due to the promoted light-to-heat conversion efficiency, results in a high surface temperature of 261.8 °C under 2.6 W cm−2, 15 °C higher than that before cycling under identical condition. On the other hand, in comparison with MgH2, the hydrogen storage reaction of Mg2Ni(Cu) is thermodynamically and kinetically favored resulting from the weakening of Mg–H bonds and the low migration barrier of hydrogen atoms in both Mg2Ni(Cu) and Mg2Ni(Cu)H4, which in turn provides a facile pathway for the spontaneous breaking of Mg–H bonds of MgH2. More importantly, attributed to the smaller work function of Cu than that of Ni and the larger work function of Cu than that of Mg, the hot electrons generated by the LSPR effect of Cu could be transferred to Ni inside of both CuNi and Mg2Ni(Cu) alloys, leading to an increase in electron density around Ni during light irradiation. The uneven distribution of light-induced hot electrons within CuNi and Mg2Ni(Cu) alloy contributes to the effective weakening of Mg–H bonds of MgH2 and hence enhances the “hydrogen pump” effect of Mg2Ni(Cu)/Mg2Ni(Cu)H4, which improves cycling hydrogenation and dehydrogenation of MgH2 and significantly reduces the apparent activation energy required for driving hydrogen storage in MgH2. As a result, the in situ atomic reconstruction of Mg2Ni(Cu) enables the integration of photothermal and catalytic roles in a single-component phase, which allows the direct action of localized photothermal heat on the catalytic sites without any heat loss, resulting in the complete dehydrogenation of MgH2 with a reversible capacity of 6.1 wt.% within 15 min under 3.5 W cm−2. More importantly, the stable reversible generation of Mg2Ni(Cu) during cycling hydrogen storage process results in well-preserved photothermal and catalytic effects in improving repeated hydrogenation and dehydrogenation of MgH2, delivering a reversible H2 storage capacity of 6.1 wt.% with a capacity retention of 95% after ten cycles using solar energy.

Results

Preparation and characterization of CuNi alloys

CuNi bimetallic alloys are first synthesized via a one-step solvothermal calcination method as shown in Supplementary Fig. 1. Nitrate salts containing Cu and Ni ions are added and evenly dispersed in a mixed aqueous solution of glycerol and propanediol, which is then hydrothermally treated at 150 °C for 3 h. After washing and thermal reduction, CuNi alloys encapsulated in carbon layers could be obtained. By changing the appropriate amount of the precursors, the ratio of Cu and Ni atoms in the thus-obtained alloy could be facile controlled, resulting in the synthesis of Cu1Nix (Cu: Ni atomic ratio of 1:x), Cu, and Ni, respectively. Inductively coupled plasma-optical emission spectroscopy (ICP-OES) indicates that the atomic ratios of Cu1Ni1, Cu1Ni2, and Cu1Ni3 are 1:1.03, 1:2.01, and 1:3.01 (Supplementary Table 1), respectively, which demonstrates the successful synthesis of CuNi alloys with controllable atomic ratios by facile adjusting the ratio of precursors.

According to X-ray diffraction (XRD) results (Fig. 1a), all the as-synthesized catalysts exhibit three distinct diffraction peaks, indicating the formation of typical face-centered-cubic (FCC) crystal phase with (111), (200), and (220) planes. The diffraction peaks of CuNi alloys show no splitting and only slight shifts in the region between the diffraction peaks of Cu (PDF#00-004-0836) and Ni (PDF#04-003-6597) could be observed for CuNi alloys, which is attributed to the slight change of crystal plane spacing caused by various Cu/Ni molar ratio20. Among them, the diffraction peaks of Cu1Ni1 are located at 43.86°, 51.09°, and 75.17°, respectively, matching the standard diffraction peaks of Cu0.5Ni0.5 (PDF#04-002-1345) ideally, thereby confirming the successful synthesis of Cu1Ni1 alloys. Moreover, the presence of Cu and Ni in CuNi alloys is confirmed in X-ray photoelectron spectroscopy (XPS) survey spectra (Supplementary Fig. 2a). High-resolution Cu 2p and Ni 2p XPS spectra (Fig. 1b, c and Supplementary Fig. 2c, d) demonstrate that Cu and Ni are present mainly in their zero-valent state, consistent with the XRD results. Particularly, the binding energies of Cu0 and Ni0 in CuNi alloys exhibit shifts compared to pure Cu and Ni due to the electronegativity difference, indicating the electronic interactions inside of CuNi alloys21,22. The weak peaks of Cu2+ and Ni2+ could be attributed to the slight oxidation of CuNi alloys during measurement (Supplementary Table 2) and no chemical interaction between the outer carbon layer and CuNi alloys could be observed in the high-resolution C 1 s XPS spectra (Supplementary Fig. 2b).

Fig. 1: Characterization of the properties and structures of CuNi alloys.
figure 1

a XRD patterns of the as-synthesized of CuNi alloys, including Ni and Cu for comparison. The red arrows represent the direction of characteristic peak shifts with increasing Cu content. High-resolution b Cu 2p and c Ni 2p XPS spectra of Cu1Ni1, with Cu and Ni for comparison. d SEM and e HRTEM images with the measured lattice spacing inset of Cu1Ni1. f SAED pattern of Cu1Ni1 with the raw pattern in lightened region on the left side for comparison. g the corresponding particle size distribution of Cu1Ni1 alloys with Gaussian fitting. h–j EDS elemental mapping images with line scanning of Cu1Ni1 alloys. Source data are provided as a source data file.

Full size image

Scanning electron microscopy (SEM) image (Fig. 1d) illustrates that Cu1Ni1 is composed of densely packed spherical particles with an average particle size of approximately 20 nm (Fig. 1g). High-resolution transmission electron microscopy (HRTEM) images (Fig. 1e and Supplementary Fig. 3a–d) confirm that Cu1Ni1 alloys are wrapped by a carbon layer of 3 nm with a weight percent of approximately 9% as evidenced by thermogravimetric analysis (TGA) (Supplementary Fig. 5a). The lattice spacing inside of Cu1Ni1 is calculated to be 0.206 nm, corresponding to the (111) crystal plane of Cu1Ni1. The selected area electron diffraction (SAED) pattern in Fig. 1f confirms the polycrystal structure of Cu1Ni1 alloys and the diffraction rings could be well recorded to (111), (200), and (220) crystal planes. In addition, the lattice spacing measured from HRTEM images (Supplementary Fig. 4a–d) presents a lattice spacing (0.206 nm) of Cu1Ni1 (111) that is between the lattice spacing (0.210 nm) of Cu (111) and the lattice spacing (0.203 nm) of Ni (111) with a periodic pattern, implying the successful synthesis of Cu1Ni1 alloys wrapped with ultrathin carbon layers. Furthermore, energy dispersive X-ray spectroscopy (EDS) elemental mapping and line scanning of two adjacent particles (Fig. 1h–j and Supplementary Fig. 3e–j) results confirm the homogeneous distribution of Cu and Ni, which coincides well with the morphology of particles, further validating the uniform synthesis of CuNi alloys. In addition, the morphology, particle size (Supplementary Fig. 6a–f), and carbon content (Supplementary Fig. 5b, c) of Cu1Ni2, Cu1Ni3, and Ni are similar to those of Cu1Ni1, while the particle size of Cu is approximately 50 nm (Supplementary Fig. 6g, h).

Solar-driven hydrogen storage performance of MgH2

The hydrogen storage performance using conventional electrical heating method demonstrates that, in comparison to MgH2 catalyzed by Ni nanoparticles, MgH2 under the catalysis of CuNi alloys exhibits superior hydrogen storage performance of further decreased operating temperature with lower kinetic barriers and enhanced kinetics of H2 desorption and absorption. The specific experimental results with detailed analysis could be found in the Supplementary Information (Supplementary Figs. 7–10). Attributed to the decrease of H2 desorption activation energy to 80 kJ mol−1, the peak temperature of H2 desorption from MgH2 catalyzed by CuNi is reduced down to approximately 245 °C, and complete H2 desorption could be achieved within 30 min at 250 °C. Moreover, under a hydrogen pressure of 1 MPa at 250 °C, 6.1 wt.% H2 could be recharged within 10 min, and 95% reversible capacity is maintained after ten cycles of H2 absorption and desorption. Although the activation energy of heat-driven reversible H2 storage of MgH2 catalyzed by CuNi has been lowered tremendously, it still requires external thermal energy input towards realizing the operating temperature of over 200 °C. Therefore, attempting to use inexhaustible solar energy to drive H2 absorption and desorption of CuNi-catalyzed MgH2 is a highly promising and environmentally-friendly alternative. To investigate the solar-driven H2 storage performance of MgH2, solar energy serves as the only energy input source without external heating, which aligns with our goal of completely replacing the electric heating with solar energy, via the apparatus shown in Supplementary Fig. 11.

After the ball-milling process of MgH2 with the addition of catalysts with a weight percent of 10%, an obvious color difference could be observed (Fig. 2a). Pure MgH2 is essentially white and it turns into yellowish-brown, and becomes darker and finally dark-black after the addition of Ni, CuNi alloys with the increasing Cu content, and Cu, respectively, indicating the increase of light absorption potential. It could be directly supported by the ultraviolet–visible–near-infrared (UV–vis–NIR) absorption spectra (Fig. 2b), which demonstrates that MgH2 catalyzed by Cu has the highest overall absorbance capability and the absorbance of MgH2 after the addition of CuNi alloys increases as the relative Cu content in these catalysts increases, where Cu1Ni2 falls between those of Cu1Ni3 and Cu1Ni1. Notably, MgH2 under the catalysis of Cu exhibits a characteristic absorption peak from 400 to 800 nm due to the LSPR effect of Cu nanoparticles (Supplementary Fig. 12a)23. By comparison, a slight red-shift of the LSPR characteristic peak of Cu is observed for MgH2 catalyzed by Cu1Ni1 owing to the size changes of nanocrystals24. The FDTD simulations further demonstrate that plasmonic “hot spot” regions could be clearly observed around particles in both Cu and CuNi alloys, which validates the strong LSPR-induced localized electric field of Cu and CuNi (Supplementary Fig. 12e–g). In addition, the relatively weaker absorption observed in the NIR region also increases with higher Cu content, which could be attributed to the scattering effect induced by the LSPR coupling of Cu particles (Fig. 2b and Supplementary Fig. 12b)25,26,27.

Fig. 2: The initial solar-driven dehydrogenation performance of MgH2 catalyzed by CuNi alloy.
figure 2

a The color, b UV–vis–NIR absorption spectra, c the response of temperature to light irradiation, d the corresponding solar-driven H2 desorption curves of MgH2, and e the heat-driven isothermal dehydrogenation curves at 250 °C of MgH2 under the catalysis of CuNi alloys, Ni, and Cu, respectively, including ball-milled MgH2 (BM MgH2) for comparison. f H2 desorption curves of MgH2 catalyzed by Cu1Ni1 under different light intensities. g Completeness ratio over time using solar energy and direct heating with normalized H2 release capacity of MgH2 under the catalysis of Cu1Ni1 at varying light intensities (or corresponding temperatures). The pentagrams represent H2 capacity. h H2 desorption curves of MgH2 catalyzed by Cu1Ni1 using solar energy (light intensity: 2.8 W cm−2, corresponding to a temperature of 250 °C) and electrical heating (250 °C), respectively. Source data are provided as a source data file.

Full size image

Herein, with the consideration of maximizing the differences brought by photothermal effects under similar catalytic performance, Cu1Ni1 and Cu1Ni3 are selected as representatives in the following. Correspondingly, under the same intensity (2.5 W cm−2 and 3.0 W cm−2) of solar irradiation, the surface temperature of MgH2 also increases with the addition of catalysts that has more relative content of Cu (Fig. 2c and Supplementary Fig. 12a). These results validate that, due to the strong LSPR effect of metal nanoparticles of Cu and Ni, the introduction of CuNi alloys endows MgH2 with overall stronger light absorption and hence better photothermal effect under solar irradiation.

As expected, induced by this photothermal effect, the surface temperature of MgH2 catalyzed by Cu1Ni1 reaches 230.1 °C within 10 min under a solar irradiation intensity of 2.5 W cm−2, which could drive the H2 desorption of MgH2 catalyzed by Cu1Ni1 (Fig. 2c), achieving a rapid H2 desorption of 4.9 wt.% within 30 min (Fig. 2d). In comparison to Cu1Ni1 alloys, MgH2 catalyzed by Cu1Ni3 exhibits the same significant decrease in dehydrogenation temperature and faster kinetics in traditional heat-driven experiments compared to MgH2 catalyzed by Cu1Ni1 (Fig. 2e and Supplementary Fig. 7a). However, under 2.5 W cm−2 illumination, MgH2 catalyzed by Cu1Ni3 only reaches a surface temperature of 224.8 °C, corresponding to a H2 desorption capacity of 4.6 wt.%, which is even 0.3 wt.% lower than that of MgH2 catalyzed by Cu1Ni1 (Fig. 2d). In addition, the difference of the dehydrogenation kinetics of MgH2 catalyzed by Cu1Ni1 and Cu1Ni3 becomes almost negligible under the solar irradiation of 2.5 W cm−2 and 3.0 W cm−2 (Fig. 2d and Supplementary Fig. 13b). This result suggests that the higher content of Cu in CuNi alloys, which brings about a stronger photothermal effect, leads to further improvement of H2 desorption performance under solar irradiation compared to those driven by direct heating, which provides direct evidence to the crucial role of photothermal effect in enhancing solar-driven hydrogen storage of MgH2.

Upon the elevation of the irradiation intensity, the surface temperature of MgH2 catalyzed by Cu1Ni1 would increase, which is capable of promoting the degree and rate of H2 desorption from MgH2 (Fig. 2f and Supplementary Fig. 14a, b). The dehydrogenation instantaneous rate peaks within the initial 100 seconds under solar irradiation, and the relationship between initial H2 release rates and light intensities is exponential (Supplementary Fig. 14c, d), consistent with the Arrhenius empirical equation28,29, which suggests that the photothermal effect primarily operates as a plasma heating. Comparing the amount of H2 desorption at various temperatures and light intensities (Supplementary Fig. 14e), the H2 desorption performance under solar energy is similar to that driven by thermal heating, that is, H2 release from MgH2 through a photothermal pathway induced by the light-to-heat conversion ability of CuNi alloys rather than photocatalytic principle30. In addition, the apparent activation energy under solar irradiation fitted by the Johnson–Mehl–Avrami (JMA) equation joint with the Arrhenius equation is calculated to be 81.23 ± 3.35 kJ mol−1, which is comparable to that driven by thermal heating (i.e., 81.56 ± 2.00 kJ mol−1) (Supplementary Figs. 14f and 15h). These results validate that the solar-driven dehydrogenation of MgH2 is mainly achieved based on the localized heat produced by photothermal effect of CuNi alloys with the combination of its catalytic effect.

To further highlight the advantages of solar-driven dehydrogenation over heat-driven dehydrogenation, the normalized H2 desorption capacities at various light intensities (or corresponding surface temperatures) are calculated to intuitively demonstrate the differences in H2 desorption ratio over time. MgH2 catalyzed by Cu1Ni1 could release 60% of H2 within 10 min under 2.5 W cm−2 irradiation (Fig. 2g), while only 20% of H2 is released from MgH2 in 10 min under electrical heating. Particularly, the superiority of fast response and rate of H2 desorption under solar irradiation over direct heating could be consistently observed for MgH2 in the range of 2.5 W cm−2 to 3.0 W cm−2 (Fig. 2g and Supplementary Fig. 15g). In strong contrast, H2 desorption kinetics of MgH2 using direct heating strongly depends on adopted temperatures induced by the low thermal conductivity of MgH2 (<1 W m−1 K−1)31. The advantages of solar-driven dehydrogenation over heat-driven H2 desorption in realizing fast dehydrogenation are also measured and evaluated using the glass tube with contact-type thermocouple method (Supplementary Figs. 11c, d and 15a–f), and demonstrated in MgH2 catalyzed by Cu1Ni3, Ni, and Cu, respectively (Supplementary Fig. 16a–f).

Specifically, when the light intensity is increased to 3.0 W cm−2, the surface temperature of MgH2 catalyzed by Cu1Ni1 and Cu1Ni3 could reach 251.6 °C and 248.9 °C, respectively, delivering a H2 desorption capacity of 6.3 wt.% and 6.1 wt.% H2, respectively, corresponding to almost complete dehydrogenation of MgH2 (Supplementary Fig. 13a, b). In contrast, no H2 desorption could be observed from the ball-milled MgH2 without CuNi alloys under the identical condition, which directly demonstrates the vital role of CuNi alloys that act simultaneously as light adsorber and thermochemical catalysts in enabling solar-driven H2 desorption from MgH2. By comparison, although the addition of Cu to MgH2 could obtain a higher surface temperature of up to 260 °C under 3.0 W cm−2 light intensity induced by the strong photothermal effect of Cu nanoparticles, only a capacity of 4.6 wt.% could be achieved at 3.0 W cm−2 due to the negligible catalytic effect of Cu in enhancing H2 desorption performance of MgH2 (Fig. 2c and Supplementary Fig. 7). Nonetheless, it is noteworthy that the H2 desorption capacity of MgH2 catalyzed by Cu at 3.0 W cm−2 is significantly higher than the 0.7 wt.% H2 capacity under the heat-driven temperature of 260 °C (Supplementary Fig. 16e). This result further highlights the significant role of LSPR effect of metal nanoparticles in producing intense localized heat for driving H2 desorption of MgH2 under solar illumination. CuNi alloys, evenly distributed on the MgH2 matrix, could generate localized heat under solar irradiation that could directly act on the catalytic center that is itself while the poor thermal conductivity of “MgH2” could reduce the temperature gradient near CuNi alloys, which is capable of promoting rapid H2 desorption from MgH218. Therefore, MgH2 catalyzed by Cu1Ni1 releases 5.3 wt.% H2 within 10 min under 2.8 W cm−2 irradiation, whereas the time to achieve identical H2 desorption capacity using external heating is doubled (Fig. 2h). Furthermore, upon increasing the loading ratio of Cu1Ni1 to 20%, the temperature of MgH2 could be elevated to 238.1 °C, 7.7 °C higher than that with a loading ratio of Cu1Ni1 of 10% (Supplementary Fig. 17a). The extent of dehydrogenation increases to 93.05% under 2.5 W cm−2, compared with 83.28% under the addition of 10% Cu1Ni1, together with a faster dehydrogenation rate (Supplementary Fig. 17b, c), which could be attributed to the enhanced photo-to-thermal effect combined with the catalytic effect in further reducing the dehydrogenation temperature (Supplementary Fig. 17d). This further validates that, with the increase ratio of CuNi alloys, the simultaneous enhancement of photothermal and catalytic effects of CuNi alloys would mutually contribute to the better solar-driven performance.

Subsequently, the reversible H2 storage performance of MgH2 under the catalysis of CuNi is investigated, which is a critical challenge in practical applications of solid-state hydrogen storage materials. Interestingly, after the initial hydrogenation process, the color of MgH2 catalyzed with Cu1Ni1, Cu1Ni2, and Cu1Ni3 turns deep black, similar to that with Cu, while MgH2 catalyzed by Ni remains yellowish-brown (Fig. 3a, inset). Surprisingly, MgH2 catalyzed by CuNi after the initial hydrogenation process not only maintains strong absorption in the ultraviolet–visible range, like that after ball-milling, but also exhibits the significant increase of overall absorbance in the near-infrared region, with a relative absorbance from ~0.5 to ~0.9 (absorption ratio from ~70 to ~86%) at a wavelength of 2000 nm (Fig. 3b). As a result, the overall absorbance of MgH2 catalyzed by CuNi after cycling even exceeds that of MgH2 catalyzed by Cu (Fig. 3a). In comparison, the overall increase in absorbance of MgH2 catalyzed by Cu and Ni after cycling is limited compared to that of CuNi (Supplementary Fig. 18). This enables MgH2 catalyzed by CuNi to utilize the near-infrared irradiation (from 800 to 2000 nm) that accounts for 52% of the solar spectrum and the enhanced full-spectrum (from 250 to 2000 nm) absorption capability after cycling is achieved32, which is critical for enhancing solar energy utilization efficiency and hence light-to-heat performance.

Fig. 3: The reversible solar-driven H2 storage performance of MgH2 catalyzed by CuNi alloys.
figure 3

a UV–vis–NIR absorption spectra of MgH2 under the catalysis of Cu1Ni1, Cu1Ni2, Cu1Ni3, Ni, and Cu after rehydrogenation. The inset shows the color of their cycled states. b UV–vis–NIR absorption spectra of MgH2 under the catalysis of Cu1Ni1, Cu1Ni2, and Cu1Ni3 after cycling, in comparison with their pristine ball-milled states (before cycling). The inset exhibits solar irradiation distribution of natural sunlight. c The signal data of temperature to light irradiation of MgH2 after cycling under the catalysis of Cu1Ni1, Cu1Ni2, Cu1Ni3, Ni, and Cu, respectively, with their pristine ball-milled states (before cycling) included for comparison. The shaded tan area corresponds to the values under 3.5 W cm−2 and the half-filled triangles represent to the absolute temperature variation. d The response of temperature with corresponding H2 desorption curves of MgH2 under the catalysis of Cu1Ni1 under 2.6 W cm−2 before and after cycling. e Cycling H2 desorption curves of MgH2 under the catalysis of Cu1Ni1 using a light intensity of 3.5 W cm-2. The de1 to de10 represent the 1st to the 10nd dehydrogenation of cycles. Source data are provided as a source data file.

Full size image

Owing to the increased overall absorption in the near-infrared range, the surface temperature of cycled MgH2 catalyzed by Cu1Ni1 is higher than that of the ball-milled state under light irradiation (Fig. 3c). Significantly, under the irradiation intensity of 2.6 W cm−2, the surface temperature of MgH2 catalyzed by Cu1Ni1 after cycling is increased to 261.8 °C, which is 6.1% higher than that after ball-milling process, correspondingly leading to the enhancement of H2 desorption rate of the cycled MgH2. The ball-milled MgH2 catalyzed by Cu1Ni1 releases 5.39 wt.% H2 within 30 min, while the time required to release the same H2 capacity could be shortened to 25 min with more H2 capacity for MgH2 catalyzed by Cu1Ni1 after cycling (Fig. 3d and Supplementary Fig. 19f). Similar phenomena of the increase of the surface temperature (with a 3.7% increase) and the acceleration of H2 desorption could also be observed for MgH2 under the catalysis of Cu1Ni2 and Cu1Ni3 after cycling under the identical light intensity (Supplementary Fig. 19a, b). In contrast, the elevation of the surface temperature of MgH2 catalyzed by Cu or Ni after cycling is only around 2% (Supplementary Fig. 19c, d). These results validate the significant enhancement of photothermal performance for MgH2 with the catalysis of CuNi after cycling, which could further promote the hydrogen storage kinetics of MgH2.

Under the light density of 3.5 W cm−2, the operating temperatures for H2 desorption and absorption of MgH2 are determined to be 277 °C and 265 °C, respectively, in the first cycle, and the corresponding temperatures reach 288 °C and 270 °C in the second cycle (Supplementary Fig. 20a, b). Impressively, a reversible capacity of 6.1 wt.% could be achieved for MgH2 catalyzed by Cu1Ni1 after ten cycles of H2 adsorption and desorption, corresponding to a capacity retention of 95% (Fig. 3e). This result fits well with the stable cycling performance of MgH2 under the catalysis of Cu1Ni1 upon direct heating (Supplementary Fig. 10a, b), which indirectly demonstrates the excellent stability of the catalytic effect under solar irradiation. On the other hand, it could be clearly observed that nearly no change of the absorbance performance could be observed for MgH2 under the catalysis of Cu1Ni1 after ten cycles compared with that after the initial hydrogenation process (Supplementary Fig. 21a), corresponding to the stable enhancement of surface temperature after cycling (Supplementary Fig. 21b), indicating the well-preserved photothermal effect, which is indispensable to realize stable solar-driven reversible hydrogen storage of MgH2. These results indicate that, the stable full-spectrum light absorption capability after cycling, combined with the excellent stability of the catalytic effect, enables MgH2 catalyzed by CuNi to achieve stable hydrogen storage reversibility driven by solar energy.

Although the relatively high light intensity is required to drive the reversible hydrogen storage of MgH2 even under the catalysis of Cu1Ni1 alloys, the low-intensity light (～0.5 W cm−2) could be concentrated to over 3.5 W cm−2 via Fresnel lens (Supplementary Fig. 22a), which makes it possible to drive reversible hydrogen storage reactions by natural light. To further demonstrate the feasibility of H2 desorption using natural solar energy, an outdoor experiment of solar-driven H2 desorption of MgH2 under the catalysis of Cu1Ni1 alloys are performed via a low-cost PMMA Fresnel lens. A striking emergence of hydrogen bubbles from the pipe could be observed (Supplementary Fig. 22b and Supplementary Movie 1), and the predominant phase in the XRD pattern transforms into pure Mg following H2 desorption process (Supplementary Fig. 22c), which offers direct evidence to the potential for driving solid-state hydrogen storage of MgH2 by natural solar energy.

Mechanism of photothermal and catalytic effects

To disclose the evolution of the photothermal and catalytic effects of CuNi, the phase and morphology change of MgH2 catalyzed by CuNi before and after cycling is investigated. Direct observation of the characteristic peaks other than MgH2 in XRD is challenging due to the amorphization of the samples and/or the ultralow content of CuNi (Supplementary Fig. 23b)33. Cu1Ni1 alloys with the characteristic lattice spacing of Cu1Ni1, however, could be found in HRTEM (Supplementary Fig. 24) with uniform distribution of Mg, Ni, Cu, and C elements, indicating that Cu1Ni1 alloys still existsin the original state and are evenly embedded inside of MgH2 after ball-milling. The initial dehydrogenation of MgH2 under the catalysis of Ni and Cu leads to the formation of Mg2Ni (PDF#00-035-1225) and Mg2Cu (PDF#04-008-6025), respectively. By comparison, the characteristic peaks of Mg2Ni(Cu) between Mg2Ni and Mg2Cu could be clearly observed with the presence of a weak peak at around 38~39°, which provides additional evidence for the formation of Mg2Ni(Cu) during the initial dehydrogenation of MgH2 catalyzed by Cu1Ni1 (Supplementary Figs. 23d, e and 25a, b). The refined structure of powder neutron diffraction (PND) of MgH2 catalyzed by CuNi after total dehydrogenation also confirms the presence of the solid solution phase Mg2Ni1-x(Cu)x, where Cu enters Ni1(0.5,0,0) and Ni2(0,0,0) sites within Mg2Ni (Supplementary Fig. 26 and Supplementary Tables 4 and 5). The (200) and (203) planes of Mg2Ni(Cu) are detected in HRTEM with a slight increase compared to the corresponding planes of Mg2Ni (Supplementary Fig. 27). These results indicate that, upon initial dehydrogenation process, CuNi alloys would react with Mg at an atomic ratio of 2:1, resulting in the in situ formation of Mg2Ni with Cu solution, i.e., Mg2Ni(Cu) ternary alloy. After rehydrogenation, the characteristic XRD peaks of Mg2NiH4 and MgCu2 are detected (Fig. 4a and Supplementary Fig. 23c), which agrees with the lattice fringes detected in TEM with the observation of uniform distribution of MgCu2 and Mg2NiH4 within the catalyst region (Supplementary Fig. 28). In addition, during the re-dehydrogenation process, MgCu2 and Mg2NiH4 undergo solid solution at the initial stage (less than 30% reaction progress), and a gradual low-angle shift of Mg2NiH4 lattice towards the characteristic peak of Mg2Ni(Cu) alloy could be observed due to the solid solution of Cu, which demonstrates the presence of an intermediate state of Mg2Ni(Cu)(H) during the hydrogenation process (Fig. 4b and Supplementary Fig. 25c, d). More importantly, the characteristic peaks of Mg2Ni(Cu) could still be clearly detected and remain unchanged in XRD patterns even after ten cycles of H2 absorption and desorption. No phase separation between Cu and Ni is observed in the EDS elemental mapping results with uniform distribution of the characteristic (200) lattice fringes of Mg2Ni(Cu) within Mg matrix. These results provide direct evidence for the reversible formation of Mg2Ni(Cu) upon cycling dehydrogenation, wherein structurally stable phases of Mg2NiH4 and MgCu2 would form after hydrogenation, and they would then rapidly transform into an intermediate state of Mg2Ni(Cu)(H) during the subsequent dehydrogenation process, accompanied by gradual alloying into single phase of Mg2Ni(Cu). Therefore, the reversible formation of Mg2Ni(Cu) with a Cu atomic ratio of x in CuNi during H2 desorption/absorption could be summarized:

$$2{{{{{{\rm{Mg}}}}}}}_{2}{{{{{{\rm{Ni}}}}}}}_{1-x}{{{{{{\rm{Cu}}}}}}}_{x}+ (4-x){{{{{{\rm{H}}}}}}}_{2} \leftrightarrow {2{{{{{\rm{Mg}}}}}}}_{2}{{{{{{\rm{Ni}}}}}}}_{1-x}{({{{{{\rm{Cu}}}}}})}_{x}({{{{{\rm{H}}}}}})\\ \leftrightarrow (2-2x){{{{{{\rm{Mg}}}}}}}_{2}{{{{{{\rm{NiH}}}}}}}_{4} + 3x{{{{{\rm{Mg}}}}}}{{{{{{\rm{H}}}}}}}_{2}+x{{{{{\rm{Mg}}}}}}{{{{{{\rm{Cu}}}}}}}_{2}$$
Fig. 4: Phase changes at different states and the electronic structures of catalytic phases.
figure 4

a XRD patterns of MgH2 under the catalysis of Cu1Ni1 at different states. b Contour plots of XRD patterns during re-dehydrogenation process of MgH2 under the catalysis of Cu1Ni1. c Band structures of Ni-orbitals and d partial densities of states (PDOS) of corresponding atoms for Mg2NiH4. e Band structures of Ni(Cu)-orbitals and f partial densities of states (PDOS) of corresponding atoms for Mg2Ni(Cu)H4. Source data are provided as a source data file.

Full size image

A similar reversible transformation between Mg2Ni(Cu), Mg2NiH4, and MgCu2 during H2 desorption/absorption is also observed in MgH2 catalyzed by equimolar ratios of Cu and Ni, further validating the reversible formation of Mg2Ni(Cu) (Supplementary Fig. 23f, g). In addition, a loss in H2 capacity of the above system could nearly be neglected under the catalysts of CuNi alloys with a loading ratio less than 10% (Supplementary Table 6). After the initial hydrogenation process, numerous interfaces and lattice distortions are observed at the area between MgCu2, Mg2NiH4, and MgH2. In addition, the presence of embedded MgCu2 and atomic disorder are even found within MgH2 lattice (Supplementary Fig. 28). After ten cycles of H2 desorption, more lattice distortions and dislocations exist both at the interfaces and within the alloys, and specifically, dislocation analysis at region (4) reveals that these regions exhibit a higher degree of disorder than Mg matrix (Supplementary Fig. 29). This is attributed to the solid-state diffusion of Mg required for the transition between Mg2Ni(Cu) and MgCu234, and the lattice distortions and disorders generated during H2 storage reactions could reduce local thermal conductivity35,36, which also helps to enhance the local thermal field.

To reveal the mechanism behind the enhancement of solar absorption of MgH2 under the catalysis of CuNi alloys before and after cycling, the electronic density of states (DOS) is investigated. In terms of pure MgH2, the valence band is mainly contributed by H(s) (79.3%) with strong hybridization with Mg(sp) (20.7%) and the conduction band is composed of 63.1% Mg(sp) and 36.9% H(s) (Supplementary Fig. 30a). Strong ionic bonding occurs between Mg2+ and H−, resulting in electrons primarily distributed around H atoms (Supplementary Fig. 31a), while the electron distribution near the Fermi level approaches zero, indicating distinct insulating properties (Supplementary Figs. 30b and 32a, b). After adding CuNi alloys into MgH2, strong absorption peaks in the visible light range could be ascribed to the LSPR effect due to intra-band absorption (Supplementary Fig. 12a), which occurs by the excitation of free electrons of conduction electrons near the Fermi surface from sp-d hybridized atomic orbitals of Cu (Supplementary Fig. 33), implying that the initial dehydrogenation process is driven by the LSPR effect of CuNi. As mentioned above, CuNi alloys would react with Mg during the dehydrogenation process, leading to the formation of Mg2Ni(Cu) alloy and then the transformation of Mg2NiH4, MgCu2, and Mg2Ni(Cu)H4 upon reversible hydrogenation. In comparison to the ionic bonding of MgH2, covalent bonds between Ni(Cu) metal atoms and their neighbor H atoms exist in Mg2NiH4 and Mg2Ni(Cu)H4 with clear electrons sharing (Supplementary Fig. 31b, c), leading to the weakening of Mg–H bonds and uniform distribution of s-p band. Near the Fermi level, a relatively weak distribution of electrons could be observed in Mg2NiH4 with semiconductor properties (Fig. 4c, d and Supplementary Fig. 32c–e), while, due to the solubility of Cu, a more flattened hybridized orbital, resembling a “flat band”, appears, which is mainly Ni(d) orbitals hybridized with weaker Mg(sp) and H(s) (Fig. 4e, f and Supplementary Fig. 32f–h). As a result, a continuous band structure fills the vacant Fermi level in Mg2Ni(Cu)H4 with metallic properties, together with MgCu2 (Supplementary Fig. 30c), which induces the facile excitation of electrons under light irradiation. Attributed to the robust hybridization of Cu(d) and Ni(d) orbitals within the valence band, the d-band is extended, which facilitates hybridization of them with Mg(sp) and H(s) orbitals, thereby providing facile pathway for inter-/intra-band transitions of electrons (including s-p and d orbitals) and leading to a broad absorption spectrum within the visible to near-infrared range37,38,39. In addition, the d electrons may undergo transitions to a higher energy state when absorbing photons with more probability due to the overlap of the localization of d orbitals, thereby inducing stronger inter-band absorption40,41. More importantly, the localization of d electrons introduced by the presence of Ni and Cu could be well-preserved even after the formation of Mg2Ni(Cu) alloy after dehydrogenation. These results indicate that the observed broad absorption spectrum of the hydrogenated state after cycling results from the synergistic effect of intra-band and inter-band transitions, and the in situ formation of Mg2Ni(Cu) alloy expands the range of photothermal effects.

On the other hand, although it is well known that Mg2Ni/Mg2NiH4 could function as the “hydrogen pump” to enhance hydrogen storage performance of MgH242,43,44 as evidenced by the significant decrease of H2 desorption temperature of MgH2 induced by the catalysis of Ni, the catalytic effect of CuNi in enhancing hydrogen storage performance of MgH2 is even superior (Supplementary Fig. 7). To unveil the catalytic mechanism of Mg2Ni(Cu)/Mg2Ni(Cu)H4 on MgH2, theoretical calculations are conducted based on density functional theory (DFT). The average length of Mg–H bonds in Mg2Ni(Cu)H4 is calculated to be 2.03 Å, much longer than that in  both MgH2 (1.93 Å) and Mg2NiH4 (1.95 Å) (Supplementary Fig. 34). This result demonstrates that, in comparison to MgH2 and Mg2NiH4, the dehydrogenation of Mg2Ni(Cu)H4 to form Mg2Ni(Cu) is thermodynamically favored, which corresponds to the preferred formation of Mg2Ni(Cu) alloy at the initial stage of re-dehydrogenation process (Fig. 4b). After the initial dehydrogenation of Mg2Ni(Cu)H4, under the presence of thus-formed Mg2Ni(Cu) alloy, the barrier of H2 desorption from MgH2 is significantly reduced down to 1.13 eV, while this value reaches 3.12 eV for pristine MgH2, validating the catalytic role of Mg2Ni(Cu) alloy in improving the dehydrogenation performance of MgH2 (Fig. 5a, b and Supplementary Fig. 35). More importantly, H atoms released from MgH2 are spontaneously transferred into Mg2Ni(Cu) alloy, which provides direct evidence for the “hydrogen pump” role of Mg2Ni(Cu) in improving the dehydrogenation of MgH2. In addition, the chemical disorders introduced by the interfaces and lattice distortions in the Mg–Cu lattice during cycling could enhance the reactivity of Mg and Mg2Ni(Cu) due to the formation of the diffusion pathways of H atoms, and serve as preferential catalytic sites for H2 dissociation/recombination45,46, which corresponds well with the theoretical calculation results that the diffusion energy barriers of H atoms inside of Mg2Ni(Cu) is slightly lower than that in the bulk phases of Mg2Ni. After hydrogenation, the diffusion energy barrier of H atoms inside of Mg2Ni(Cu)H4 (i.e., 0.62 eV) is significantly lower than that of Mg2NiH4 (i.e., 1.48 eV) (Supplementary Fig. 36), which endows Mg2Ni(Cu)/Mg2Ni(Cu)H4 superior “hydrogen pump” effect in improving hydrogen storage performance of MgH2 than traditional Mg2Ni/Mg2NiH4. Hence, the in situ stable formation of Mg2Ni(Cu) offers a kinetically favored pathway for repeated dehydrogenation an hydrogenation of MgH2.

Fig. 5: The mechanism of the integration of photothermal and catalytic effects.
figure 5

a Schematic illustration of H2 desorption and H diffusion pathway of MgH2 on the surface of Mg2Ni(Cu) alloy and b the corresponding energy profiles. c Schematic diagram of the ideal integration of photothermal and catalytic effects via continuous in situ atomic reconstruction upon repeated dehydrogenation process. Source data are provided as a source data file.

Full size image

Furthermore, upon light irradiation, the presence of hot electrons induces internal disparities in electron distribution within CuNi and Mg2Ni(Cu) alloys as observed via in situ irradiated XPS (ISI-XPS) (Supplementary Fig. 37a–e). Attributed to the work function difference, the hot electrons generated by the LSPR effect of Cu transferred to Ni, leading to an increase in electron density around Ni (Supplementary Fig. 38a, b), which enables Ni sites to contribute more electrons to H bonded to Mg atoms of MgH2 (Supplementary Fig. 39a–d), and hence leads to the weakening Mg–H bonds (Supplementary Fig. 39e–h). As a result, compared to pure Ni with uniform electron distribution, the electron enrichment at the Ni sites in CuNi and Mg2Ni(Cu) alloys facilitates the elongation of Mg–H bonds from 1.87 Å to 1.91 Å and 2.00 Å, respectively, while the impact of Cu sites remains unchanged. The internal uneven electron distribution within CuNi and Mg2Ni(Cu) alloys is conducive to promoting the dissociation of Mg–H bonds and hence improving the “hydrogen pump” effect of Mg2Ni(Cu)/Mg2Ni(Cu)H4. In addition, Mg does not affect the electron transfer between CuNi under light irradiation due to its positive charge, ensuring the promotion of Mg–H bonds dissociation by the in situ well-preserved Mg2Ni(Cu) during cycling. These hot electrons could be generated continuously under light irradiation and, however, cannot be directly participate in the MgH2 dehydrogenation process, which subsequently would decay into localized thermal energy47, acting on catalytic sites in the form of heat, thereby elevating the local temperature of catalytic sites and enhancing their catalytic effects.

Therefore, the in situ formation of Mg2Ni(Cu) alloys induces not only photothermal effect via broadening the light absorption range to cover the entire solar spectrum but also a light-enhanced “hydrogen pump” role in improving reversible hydrogen storage performance of MgH2 (Fig. 5c). More importantly, the in situ atomic reconstruction of Mg2Ni(Cu) single phase upon repeated dehydrogenation process realizes the continuous integration of photothermal and catalytic roles to ensure the direct action of localized heat on the catalytic sites without any heat loss, thereby achieving the ideal coupling of photothermal and catalytic effects for driving reversible hydrogen storage of MgH2 using solar energy.

Discussion

In summary, a single-component phase of Mg2Ni(Cu) ternary alloy via atomic reconstruction is designed to achieve the ideal integration of photothermal and catalytic effects to enhance the solar-driven hydrogen storage performance of MgH2. Taking advantage of the strong LSPR effect of CuNi alloys, the formation of Mg2Ni(Cu) ternary alloy is realized through in situ alloying reaction between MgH2 and CuNi alloys during the initial dehydrogenation process of MgH2 under solar irradiation. Interestingly, Mg2Ni(Cu) alloy and its hydrogenated state (i.e., Mg2Ni(Cu)H4) exhibit metallic properties with strengthened intra/inter-band transitions, resulting in over 85% absorption in the full-spectrum range, which effectively elevate the surface temperature of MgH2 to 261.8 °C under 2.6 W cm−2 due to the promoted light-to-heat conversion efficiency. More importantly, in comparison with MgH2, the hydrogen storage reaction of Mg2Ni(Cu) is thermodynamically and kinetically favored resulting from the weakening of Mg–H bonds and the low migration barrier of hydrogen atoms in both Mg2Ni(Cu) and Mg2Ni(Cu)H4, which in turn provides a facile pathway for the spontaneous breaking of Mg–H bonds of MgH2. The uneven distribution of the light-induced hot electrons within CuNi and Mg2Ni(Cu) alloys also contributes to the effective weakening of Mg–H bonds of MgH2 and the accelerated formation of Mg2Ni(Cu)/Mg2Ni(Cu)H4, hence resulting in a light-enhanced “hydrogen pump” effect in improving cycling hydrogenation and dehydrogenation of MgH2 and the significant decrease of the apparent activation energy required for driving hydrogen storage of MgH2. As a result, the reversible formation of Mg2Ni(Cu) ideally integrates the photothermal and catalytic effects to ensure that the direct action of localized heat on the catalytic sites without any heat loss, which leads to the complete dehydrogenation of MgH2 within 15 min under the irradiation intensity of 3.5 W cm−2. Moreover, the stable reversibility between Mg2Ni(Cu) could be well-preserved during cycling H2 desorption and adsorption, which results in the excellent stability of photothermal effect with full-spectrum light absorption and catalytic effect, leading to a reversible hydrogen storage capacity of 6.1 wt.% with a capacity retention of 95%. Our work expands the functionality of the alloying strategy to modify metal hydrides under solar irradiation and may enlighten the design of efficient photothermal-catalytic solar-driven reversible hydrogen storage systems and solar-chemical energy reactions.

Methods

Materials

Copper nitrate hydrate (Cu(NO3)2·3H2O), nickel nitrate hexahydrate (Ni(NO3)2·6H2O), tartaric acid (C4H6O6), glycerol (C3H8O3), and polyethylene glycol were all purchased from Sinopharm chemical reagent co., Ltd. Magnesium hydride (MgH2) was purchased from Alfa Aesar. All chemicals were used without further purification.

Synthesis of CuNi alloys

CuNi alloys were synthesized by a one-step solvothermal calcination method. Firstly, 386.6 mg of Cu(NO3)2·3H2O, 465.5 mg of Ni(NO3)2·6H2O, and 1200 mg of tartaric acid were added to 8 mL of deionized water and stirred until completely dissolved, referred to as Solution 1. Subsequently, 2 g of polyethylene glycol, 20 ml of glycerol, and 5 ml of deionized water were mixed and sonicated for 10 min, referred to as Solution 2. The two obtained solutions were mixed and stirred for 1 h to achieve uniform dispersion and transferred to a Teflon stainless-steel autoclave and heated at 150 °C for 3 h. The resulting product was centrifuged at about 3500×g and washed three times with anhydrous ethanol to remove excess reactants and impurities. The centrifuged product was then dried completely in an 80 °C oven for 12 h. The dried product was placed in a tube furnace and heated at a rate of 3 °C min−1 in an H2/Ar atmosphere (with 5% H2) up to 800 °C for 2 h. CuNi alloys with encapsulated carbon layer were obtained through the gas reduction method, with a Cu to Ni atomic ratio of 1:1, denoted as Cu1Ni1. By varying the relative ratios of Cu and Ni salts added to precursor Solution 1, the ratio of CuNi alloys could be adjusted. Under the same reaction conditions, adding 184.74 mg Cu(NO3)2·3H2O with 667.1 mg Ni(NO3)2·6H2O yielded Cu1Ni3 with a Cu-to-Ni atomic ratio of 1:3. Without adding Cu(NO3)2·3H2O or Ni(NO3)2·6H2O, Ni and Cu were obtained.

Synthesis of MgH2 catalyzed by CuNi alloys

CuNi–MgH2 composite was synthesized via mechanical milling. Commercially available MgH2 with 10 wt.% catalysts were added in a 100-mL stainless-steel container under an argon atmosphere. The ball-to-powder weight ratio was controlled to be 100:1, and the rotation speed of the miller was maintained at 500 rpm (alternate clockwise and counterclockwise rotation) for 12 h (5-min-stop every 30 min). Ball-milled MgH2, Cu–MgH2, and Ni–MgH2 composites were also prepared under identical processing conditions. All samples were handled in an argon-filled glovebox with O2 and H2O <0.1 ppm.

Materials characterization

X-ray diffraction (XRD, D8 advance, Bruker AXS, USA) with Cu Kα radiation (1.542 Å) at 40 kV and 40 mA was used to determine the phase compositions. The materials were protected against oxidation from the air by adhesive tape (amorphous with a broad peak around 2θ ≈ 20°). X-ray photoelectron spectroscopy (XPS) results were obtained on a Thermo Scientific K-Alpha system equipped with a dual X-ray source, adopting an Al Kα (1486.6 eV) anode with a hemispherical energy analyzer. In situ irradiated XPS (ISI-XPS) were performed on a Thermofisher ESCALAB 250Xi under Xe lamp irradiation. Time-of-flight (TOF) powder neutron diffraction (NPD) data were collected from a general-purpose powder diffractometer (GPPD) (90° bank) at the China Spallation Neutron Source (CSNS), Dongguan, China. Neutron diffraction patterns were recorded at room temperature. Rietveld refinements of the NPD patterns were performed using the GSAS software. The morphologies of the synthesized catalysts and MgH2 catalyzed by CuNi with different states were observed by scanning electron microscopy (SEM, JEOL 7500FA, Japan) and transmission electron microscopy (TEM, JEOL JEM-2100F, Japan). The Cu and Ni contents of samples were measured by an inductively coupled plasma source mass spectrometer (ICP-MS, Agilent ICPOES730, USA). The absorption spectra of samples in the wavelength range of 200~2200 nm were determined using a UV–vis–NIR Spectrophotometer (Lambda 750, Perkin-Elmer, USA). The content of carbon and alloy were analyzed by thermogravimetric analysis (TGA, Discovery TGA 550, USA) from room temperature to 800 °C at a rate of 10 °C min−1 in the air.

The light source was a 300 W Xenon lamp (Perfect Light Co., Ltd., China), and the light intensity was adjusted by tuning the current of the light source and the distance between the light source and the sample. The light intensity of the 300 W Xenon lamp was measured by an optical power meter (PL-MW2000, Perfect Light Co., Ltd., China). A short-wave infrared thermometer (DGE44N, range: 75~50 °C, DIAS, Germany) was used to measure the surface temperature of the composites during the reaction process under vacuum conditions through a sapphire window.

Hydrogen storage measurements

Temperature-programmed desorption (TPD), isothermal dehydrogenation tests, and corresponding hydrogenation process of the as-prepared samples were conducted on a home-built high-pressure gas sorption apparatus (HPSA-auto), which was carefully calibrated by adopting LaNi5 as a reference sample in terms of hydrogen storage capacity and guaranteed accuracy of ± 1%. All samples’ hydrogen capacity (wt.%) is calculated based on the total mass of MgH2 and the additives.

As for electric heating, for each volumetric temperature-programmed desorption measurement, approximately 30 mg of ball-milled sample were conducted under vacuum from room temperature to preset temperature (400 °C) at a heating rate of 5 °C min−1 under an initial pressure lower than 0.0001 bar, with an electric heater to determine the hydrogen release properties of the MgH2 materials briefly. The isothermal H2 desorption and absorption test were handled by rapidly heating up to the target temperature followed by keeping at the preset temperature. In the isothermal H2 absorption experiment, ~80 mg of samples were used for hydrogenation under the hydrogen pressure of 10 atm.

As for light irradiation, customized solar-driven H2 absorption and desorption equipment (Supplementary Fig. 11) was used to perform experiments under solar energy. The adjustment of the light intensity was realized by changing the current intensity. The light was incident through the above sapphire window, and the outlet of the reactor was connected to the PCT equipment. The temperature of the material during the H2 absorption and desorption was measured by a short-wave infrared thermometer. About 30 mg samples were taken each time and pressed into pieces with a diameter of 15 mm under a pressure of 5 tons to ensure the stability of the test and the comparability of the results. Glass fiber sheets made up of mainly SiO2 with low thermal conductivity, which have limited absorption of solar energy, were placed at the bottom of the reactor to reduce the heat dissipation of the material to the outside. The sample was placed on an Al2O3 ceramic sheet, which could be isolated from the glass fiber to prevent the sample from oxidizing by reacting with the insulation layer due to the high temperature during the reaction. The expansion of internal gases within the reactor due to solar irradiation has a negligible impact on the H2 capacity results, as a large chamber, 25 times the volume of the reactor, was opened during the dehydrogenation tests to minimize testing errors. The hydrogen pressure used for H2 absorption under solar irradiation was consistent with that using electric heating.

Theoretical calculations

Density functional theory (DFT) calculations were carried out using the Vienna ab initio simulation package (VASP)48,49. Core-valence electron interaction was described by the projector-augmented wave (PAW) method50, and electron exchange-correlation was processed by generalized gradient approximation (GGA) of Perdew-Burke-Ernzerhof (PBE) functional51,52,53. The van der Waals (vdW) correction DFT-D3 proposed by Grimme was employed to include the dispersion interaction54. A plane wave energy cutoff of 520 eV and Gamma-centered k-point meshes (3 × 3 × 1) with same density were applied to surface calculations. A 7 × 7 × 7 Gamma-centered k-point meshes is used for the dos calculation of Cu, MgH2, Mg2NiH4, Mg2Ni(Cu)H4, and 10 × 10 × 10 is used for the dos calculation of CuNi. For all slab structures, a 20 Å vacuum layer in the z-direction is added to prevent surface interactions between the upper and lower surfaces. In energy band calculation, the number of k-points is uniformly 20, and the k-path and energy band information are exported by the VASP post-processing tool VASPKIT. The Gaussian smearing method was used with a width of 0.05 eV, and spin polarization was considered in all calculations. The structures were relaxed until the forces and total energy on all atoms were converged to less than 0.05 eV Å−1 and 1 × 10−5 eV. The MgH2 dissociation and diffusions of H in various crystal lattices are simulated using the climbing-image nudged elastic band (CI-NEB) method55. The climbing-image nudged elastic band (Cl-NEB) method developed by Henkelman et al. was employed to locate the transition states of the MgH2 dissociation and diffusions of H in various crystal lattice56,57.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.