In earlier researches (before 2010), porous carbon materials were regarded as a kind of good additives to improve hydrogen storage properties by enhancing the hydrogen diffusion in MgH₂-C systems. Graphite was firstly used as anti-sticking agent to enhance the ball milling efficiency (Imamura et al., 2003; Imamura et al., 2005). Interestingly, comparing with noncarbon additives, all the carbon additives showed positive effects on the hydrogen storage properties of Mg, especially on the increment of hydrogenation capacity (Wu et al., 2006a). Whereafter, carbon materials with three-dimensional structures have attracted more attention. For instance, a C₆₀ fullerene molecule is capable of storing up to 58 H atoms remaining a metastable structure, which is equivalent to a storage capacity of 7.5 wt% (Pupysheva et al., 2008). Owing to the strong adsorption ability for hydrogen and the excellent heat transfer performance, CNTs may enhance the hydrogen storage capacity and lower the hydrogen desorption platform temperature of Mg (Lyu et al., 2019).

The corresponding hydrogenation mechanism was reported to compose of three consecutive steps: 1) the dissociation of hydrogen molecules on Mg or alloy surfaces; 2) the diffusion of atomic hydrogen along grain boundaries; 3) the interaction between metal and atomic hydrogen. CNTs played a dominated role in the second step by forming thousands of 10–50 nm grains in the Mg particle (500–1,000 nm) (Yao et al., 2007), which built hydrogen diffusion channels to drive hydrogenation/dehydrogenation reactions towards complete (Wu et al., 2006b). Besides, absorbed carbon component on freshly exposed Mg surfaces inhibits the reform of oxide layers. In the carbon fragments produced from ball milling during preparation, unhybridized π electrons are delocalized and then interact with hydrogen appropriately (Wu and Cheng, 2010).

However, fullerenes or CNTs have an inherent spatial structure, so that their properties can only be adjusted using physical approaches like ball milling. In order to lower the energy barrier for hydrogen dissociation, the addition of metallic catalyst is supposed to be achieved by alloying with Mg, which demands stringent preparation conditions for melting (Xu et al., 2019). Therefore, a class of more flexible and reproducible carbon-based materials with exposed metal sites is in demand. MOFs, which are assembled by various inorganic nodes and organic linkers through coordination bonds, have become a more competitive option for hydrogen storage in the past decade.

MOFs and MOF-derived carbon nanoarchitectures have been applied in various fields, e.g. separation (Cui et al., 2019), desalination (Wang et al., 2019b; Xu et al., 2020), electrocatalysis (Zhang et al., 2020), chemical sensing (Khan et al., 2019), pollutant adsorption (Wen et al., 2019), supercapacitor (Xu et al., 2017), CO₂ reduction (Xuan et al., 2020), etc. Since hydrogen adsorption in MOFs was innovatively proposed by Yaghi’s group in 2003 (Rosi et al., 2003), extensive researches on MOFs as hydrogen storage materials have been done by functionalization of ligands and modification of textural properties (Langmi et al., 2014). The nanoconfined environments of MOFs provide ideal conditions for hydrogen capture, storage, and release with considerable safety, convenience, and efficiency. Besides, MOFs may also act as precursors to design desired hierarchically porous structures (Guo et al., 2021; Liu et al., 2021). The advantages of using MOFs for hydrogen storage mainly lie in: 1) low density; 2) high specific surface area; 3) tunable porous structure; 4) various exposed metal nodes; 5) reproducible and facile synthesis; 6) controllable chemical functionality; 7) amenability to scale-up.

The interaction between hydrogen and adsorbent surfaces are mainly van der Waals force, the energy of which is as low as 4–8 kJ/mol. Adsorbent pore sizes close to the kinetic diameter of hydrogen molecule (2.89 Å) can promote the interaction with hydrogen, when the potential fields from opposite walls overlap (Perles et al., 2005). Although MOFs are favorable owing to their tunable porosities, high hydrogen storage capacities (up to 4.5–7.5 wt%) are normally achieved at cryogenic temperature (77 K) and high pressure, which is not desirable for practical use (Thomas, 2007; Shet et al., 2021). In order to improve the hydrogen storage capacities of MOFs, several approaches to introduce dissociated adsorption of hydrogen by forming hybrid composites have been come up with.

The hydrogen spillover mechanism has aroused wide interest. Hydrogen molecules are firstly dissociated on the catalyst, and then newly generated hydrogen atoms will more easily migrate to the substrate. The introduction of chemisorption hence enables to store hydrogen at room temperature. The addition of metal dopants, e.g. Pd, Pt, Ni, working as hydrogen spillover receptor can effectively enhance hydrogen storage capacity (Zhou et al., 2015; Guo et al., 2020). This mechanism gives a new perspective on the interaction between MOFs and magnesium for hydrogen storage.

Rowsell and Yaghi (2005) proposed that combining light metals with MOFs was one of the most effective strategies to promote hydrogen adsorption in MOFs. Even substituting metal nodes within MOFs with Mg can enhance hydrogen uptake. According to the theoretical calculation result given by Han et al. (2009), Mg₄O(CO₂)₆H₆ cluster showed a stronger H₂ binding energy than original Zn₄O (CO₂)₆H₆ and Be-substituted one, so that Mg-MOF had the highest hydrogen uptake at low pressure (1 bar).

The diffusion path of hydrogen can be prominently shortened when Mg nanoparticles are confined inside a scaffold with fine porosity, which promotes the thermodynamics, kinetics, and reversibility of hydrogen storage (Nielsen et al., 2009; Zhang et al., 2017). The imposed nanostructural geometry of Mg/MgH₂ matrix creates a great number of defects, forming specific interactions between Mg-H bonds (Wu, 2008). For hydrogen storage, chemisorption positively correlates to the temperature while physisorption inversely correlates. A balance of high storage capacity and low operation temperature is supposed to be achieved by appropriate combination of associated adsorption and dissociated adsorption. In this way, magnesium showed considerable synergistic effect in the presence of MOFs for hydrogen de/adsorption properties.

Compared with bare Mg, the Mg/ZIF-67 composite lowered dehydrogenation activation energy by 43.2 kJ/mol. When hydrogen adsorbed by Mg/ZIF-67 was completely released at 338°C, only 44% hydrogen within Mg was released. Microstructural characterization revealed the formation of core-shell structure within the Mg/ZIF-67 nanocomposite (Figure 1). The MOF scaffold impeded the growth and agglomeration of Mg/MgH₂ nanocrystals as an “aggregation blocker” in the hydrogen de/adsorption processes and ensured outstanding cyclic stability. Therefore, the Mg/ZIF-67 composite showed no degradation in hydrogenation/dehydrogenation capacity even after 100 cycles (Wang et al., 2019a).

The research of Lim et al. (2012) showed that Mg embedded MOF (Mg@SNU-90′) had lower desorption temperature and higher hydrogen adsorption isosteric heat, owing to significantly reduced chemisorption temperature of Mg nanocrystals within SNU-90′ comparing with pure Mg powder. The estimated hydrogen uptake capacity of Mg alone within the composite reached 7.5 wt% at 473 K and 30 bar, indicating 99% of the Mg nanocrystals were fully utilized, which was very close to the theoretical limit of 7.6 wt%. Desorption experiments showed that the organic ligands of SNU-90′ were not hydrogenated, indicating that Mg@SNU-90’ demonstrated synergistic behavior of physisorption by MOF and chemisorption by Mg. Jia et al. (2013) found that a fraction of hydrogen was desorbed below 80 °C from MgH₂ clusters which were confined to mesoporous carbons, owing to partially transferring electrons from the Mg/MgH₂ matrix. Electron transfer capability between Mg and appropriate hosts hence lowers energy barriers for hydrogenation/dehydrogenation processes.

It is reported that the size effect of catalytic active sites plays an important role in preparing Mg-based nanocomposites (De Jongh and Adelhelm, 2010). MOFs not only adsorb hydrogen through their porosities, but also act as excellent catalyst carriers to enhance hydrogenation thermodynamic properties. Owing to the self-assembled inerratic spatial structures, the metal active sites within MOFs may achieve quite high dispersion. In this way, a fraction of transition metals is introduced into the nanocomposites to modify the high energy barrier for hydrogen dissociation on Mg surfaces with no obvious density increasement.

Ma et al. (2019) used trimasic acid (TMA) as organic linker to prepare TMA-TM MOF (TM = Co., Fe). After ball milled with MgH₂, the activation energy (Ea) for hydrogenation of the dehydrogenated MgH₂-Fe MOF dropped to 66.8 kJ/mol H₂, which was much lower than that of pure Mg (100.7 kJ/mol H₂). The dehydrogenation activation energy (Ed) of MgH₂-Fe MOF composites (142.3 kJ/mol H₂) was also much lower than that of pure MgH₂ (181.4 kJ/mol H₂). MOFs showed considerable catalytic effects in hydrogen storage.

Jia et al. (2015) fabricated ultrafine (2–3 nm) metallic Ni nanoparticles embedded in the crushed MgH₂ matrix by in-situ bottom-up reduction of the Ni-MOF-74 precursor. The composite adsorbed 6.2 wt% hydrogen within only 30 s at a temperature as low as 150 C and fully desorbed within just 20 min at 300 C, which exhibited faster kinetics than conventional MgH₂. Theoretical calculation showed that dissociated hydrogen was capable of shifting quickly from Ni active sites to pure Mg with an energy increment of only 0.07 eV. Hydrogen diffusion on Ni cluster hence became the rate limiting step, so that limited Ni nanoparticle size significantly promoted hydrogenation/dehydrogenation kinetics.

In the Mg embedded Ni-MOF composite, in-situ formation of Mg₂NiH₄ around the MgH₂ nanocrystals (∼3 nm) was observed (Ma et al., 2020). The “hydrogen pump” effect owing to mutual phase transformation between Mg₂Ni and Mg₂NiH₄ made it easier to release and uptake hydrogen for MgH₂ matrix. Huang et al. (2021) further studied the catalytic mechanism of MOF-derived Ni nanoparticles on monolayer Ti₃C₂ MXene carrier. DFT calculations showed that Mg₂Ni (Ti) H₄ had lower formation energy than the neat one, so that Mg₂Ni/Mg₂NiH₄ transformation was facilitated by in-situ formed Ti⁰ from MXene during ball milling, which kept a steady state with no phase or valence change. Comparing with the pristine MgH₂, over 120 C decrease in desorption peak temperature and over 60% reduction in dehydrogenation activation energy were achieved owing to the “facilitated hydrogen pump” effect. MOF-promoted hydrogen storage properties of MgH₂ have been summarized in Table 1.