Mg-based compounds for hydrogen and energy storage

Abstract

Magnesium-based alloys attract significant interest as cost-efficient hydrogen storage materials allowing the combination of high gravimetric storage capacity of hydrogen with fast rates of hydrogen uptake and release and pronounced destabilization of the metal–hydrogen bonding in comparison with binary Mg–H systems. In this review, various groups of magnesium compounds are considered, including (1) RE–Mg–Ni hydrides (RE = La, Pr, Nd); (2) Mg alloys with p-elements (X = Si, Ge, Sn, and Al); and (3) magnesium alloys with d-elements (Ti, Fe, Co, Ni, Cu, Zn, Pd). The hydrogenation–disproportionation–desorption–recombination process in the Mg-based alloys (LaMg₁₂, LaMg₁₁Ni) and unusually high-pressure hydrides synthesized at pressures exceeding 100 MPa (MgNi₂H₃) and stabilized by Ni–H bonding are also discussed. The paper reviews interrelations between the properties of the Mg-based hydrides and p–T conditions of the metal–hydrogen interactions, chemical composition of the initial alloys, their crystal structures, and microstructural state.

1 Introduction

This work is a review of the recent progress made in the past years, particularly in the framework of the International Energy Agency Task 32 Hydrogen-based Energy Storage, in the area of fundamental and applied studies of Mg-based compounds for hydrogen and energy storage.

In order to find a hydrogen storage system that meets the requirements of low weight (for mobile applications) and adequate capacity, chemical hydrides based on lightweight metals have been extensively investigated [1, 2]. Due to the low cost of extraction from natural chlorides and carbonates which form a significant part of the earth’s crust and oceans, magnesium, a lightweight alkaline earth metal, fits many of these requirements.

The reaction of magnesium with hydrogen yields magnesium dihydride (MgH₂) and is reversible, with a storage capacity of 7.6 wt% of hydrogen. The hydrogenation process, however, suffers from slow kinetics, and the dehydrogenation of MgH₂ requires temperatures exceeding 300 °C, which is impracticable for mobile applications and constitutes a significant net energy loss for any system based on MgH₂.

The kinetics have been shown to improve by crystallite size reduction which allows for shorter hydrogen diffusion paths and the introduction of defects in the material which provides faster pathways for hydrogen to diffuse [3–5]. The introduction of small amounts of additives (typically transition metals and transition metal oxides, but also carbons [6, 7]) has also improved the kinetic rates of absorption and desorption significantly [8]. However, reducing the equilibrium temperature of desorption relies on lowering the thermodynamic enthalpy, and one of the more promising pathways for this is the formation of magnesium alloys with other metals.

Materials suitable for the storage of hydrogen and heat can play an important role for future climate policy goals, because they can compensate for fluctuations in the production of electricity from the renewable energy sources. Of particular interest are metal hydride compounds because these compounds connect energy production and storage with a future hydrogen economy.

Mg-based metal hydrides can be used as solid-state hydrogen storage materials for fuel cell cars, as a hydrogen source for fuel cell auxiliary power units, for the storage of high-temperature heat in industrial processes and in power plants, or for the smoothing of irregular supply of heat and electricity production for fuel cells in domestic energy sectors.

Metal hydrides are produced from different pure metals or metal alloys under the influence of hydrogen. This process of hydrogenation is accompanied by the release of heat (exothermic reaction). The same amount of heat must be supplied to the metal hydride for the decomposition and the release of hydrogen (endothermic reaction, see Eq. 1). These processes are infinitely repeatable, and reversible Mg-based metal hydrides can therefore be used for the storage of heat and hydrogen [9]. Depending on the thermodynamics of the reaction, different temperatures apply to different metal hydrides (Table 1).

Table 1 Selected physical data of Mg-based metal hydrides for hydrogen and heat storage applications

$${\text{Mg}} + {\text{H}}_{2} \mathop{\rightleftarrows}\limits_{\text{heat\,storage}}^{\text{heat\,release}}{\text{MgH}}_{2}$$

(1)

1.1 Thermodynamic tuning

The temperature required to generate 1 bar of hydrogen from a hydride system can be calculated from the change in enthalpy and entropy for dehydrogenation, as shown in Eq. 2. Thermodynamic tuning of a multicomponent hydride system requires the addition of a second phase (Z) to the hydride (YH₂), which reacts with the hydride, releasing hydrogen and generating a new alloy or compound (YZ) [12]. As can be seen from Fig. 1, the reduction in the enthalpy for dehydrogenation is determined by the enthalpy of formation for the alloy or compound formed, as shown in Eq. 3.

Fig. 1

Thermodynamic tuning for multicomponent systems

$$T\left( {1\;{\text{bar}}} \right) \, = \Delta H/\Delta S$$

(2)

$$\Delta H_{b} = \Delta H_{a} + \Delta H_{f}$$

(3)

1.2 Mg alloys with p-elements (X = Al, Si, Ge, Sn)

1.2.1 Silicon

Silicon has a very low solubility in magnesium of 0.003 %. When MgH₂ mixed with Si is heated until the hydrogen is released, it forms the intermetallic Mg₂Si [13]. While the reaction is reversible in a ball-mill environment with hydrogen [14–16], application of pressures of hydrogen to Mg₂Si up to 185 MPa and 350 °C has not yielded a hydride phase [13, 17] except when the Mg₂Si particle size was reduced to 10 nm. Even then, the yield was low and the kinetics very poor [18].

Bystrzycki et al. [19] observed only the metastable Mg₅Si₆ in a study of the destabilization of MgH₂ by Si, where Si was deposited by pulsed laser deposition on a MgH₂/Mg substrate and did not find any destabilization.

However, the Mg–Si–H system has been studied with regard to its hydrogen storage properties. Samples were prepared by reactive ball-milling under hydrogen, and depending on the milling conditions, different paths can be obtained. For high rotation speeds, Mg₂Si is formed, whereas 2MgH₂ + Si are obtained at low rotation speeds. This behaviour can be understood by considering the change in local temperature at the collision impact with milling speed. On this basis, it is observed that the reaction Mg₂Si + 2H₂ ⇌ 2MgH₂ + Si is reversible. The hydrogenation reaction can be achieved by milling Mg₂Si at low rotation speed, whereas the reverse reaction can be triggered by milling at high rotation speed. However, a small quantity of Mg₂Si is necessary to initiate the reaction between magnesium hydride and silicon.

$$2{\text{Mg}}_{{({\text{s}})}} + {\text{Si}}_{{({\text{s}})}} + 2{\text{H}}_{{2({\text{g}})}} \to \raise.5ex\hbox{$\scriptstyle 1$}\kern-.1em/ \kern-.15em\lower.25ex\hbox{$\scriptstyle 2$} {\text{ Mg}}_{2} {\text{Si}}_{{({\text{s}})}} + {\text{MgH}}_{{2({\text{s}})}} + \raise.5ex\hbox{$\scriptstyle 1$}\kern-.1em/ \kern-.15em\lower.25ex\hbox{$\scriptstyle 2$} {\text{ Si}}_{{({\text{s}})}} + {\text{H}}_{{2({\text{g}})}}$$

(4)

$$\raise.5ex\hbox{$\scriptstyle 1$}\kern-.1em/ \kern-.15em\lower.25ex\hbox{$\scriptstyle 2$} {\text{Mg}}_{2} {\text{Si}}_{{({\text{s}})}} + {\text{MgH}}_{{2({\text{s}})}} + \raise.5ex\hbox{$\scriptstyle 1$}\kern-.1em/ \kern-.15em\lower.25ex\hbox{$\scriptstyle 2$} {\text{ Si}}_{{({\text{s}})}} + H_{{2({\text{g}})}} \to {\text{Mg}}_{2} {\text{Si}}_{{({\text{s}})}} + {\text{H}}_{{2({\text{g}})}}$$

(5)

$$\raise.5ex\hbox{$\scriptstyle 1$}\kern-.1em/ \kern-.15em\lower.25ex\hbox{$\scriptstyle 2$} {\text{ Mg}}_{2} {\text{Si}}_{{({\text{s}})}} + {\text{MgH}}_{{2({\text{s}})}} + \raise.5ex\hbox{$\scriptstyle 1$}\kern-.1em/ \kern-.15em\lower.25ex\hbox{$\scriptstyle 2$} {\text{ Si}}_{{({\text{s}})}} + H_{{2({\text{g}})}} \to 2{\text{MgH}}_{{2({\text{s}})}} + {\text{Si}}_{{({\text{s}})}}$$

(6)

$${\text{Mg}}_{2} {\text{Si}}_{{({\text{s}})}} + 2{\text{H}}_{{2({\text{g}})}} \rightleftharpoons 2{\text{MgH}}_{{2({\text{s}})}} + {\text{Si}}_{{({\text{s}})}}$$

(7)

$$2{\text{MgH}}_{{2({\text{s}})}} + {\text{Si}}_{{({\text{s}})}} \to {\text{Mg}}_{2} {\text{Si}}_{{({\text{s}})}} + 2{\text{H}}_{{2({\text{g}})}}$$

(8)

As shown in Fig. 2, starting from the elements, after 4 h of high-energy milling (HEM) under hydrogen, a metastable state is reached where three solid phases coexist (MgH₂, Si, and Mg₂Si). At this stage, further milling keeping to the HEM regime leads to the formation of Mg₂Si without H₂ absorption, whereas low-energy milling (LEM) allows the formation of MgH₂ + Si with full hydrogen absorption. It is therefore possible to tune the reaction by switching from HEM to LEM. The final products are dependent on the milling energy which probably governs the local temperature at milling impact (Eqs. 4–8). Interestingly, the full conversion of Mg₂Si + H₂ into MgH₂ + Si is also possible by LEM. However, the reverse reaction could not be obtained by HEM probably due to kinetic limitations at the MgH₂/Si interfaces. Such limitations are overcome by the presence of Mg₂Si in three-phase mixtures. Contrary to solid/gas reactions, the present work shows that reversible hydrogen storage in Mg₂Si can be achieved by ball-milling at room temperature under moderate hydrogen pressure (9 MPa).

Fig. 2

Free energy diagrams, showing milling reagents and obtained products for two pressure–temperature conditions (100 and 250 °C at p(H₂) = 9 MPa)

The free-energy change, $\Delta_{r} G_{P}^{T}$, of the reactions observed in this study is displayed in Fig. 2. By increasing the local temperature, the energy level of MgH₂ increases as compared to Mg₂Si. Indeed, the large entropy change associated with the absorption of molecular hydrogen from the gas phase into atomic hydrogen in the Mg lattice accounts for the high variation in $\Delta_{r} G_{P}^{T}$ with temperature. As a consequence, while at low temperatures (e.g. 100 °C) $\Delta_{r} G_{P}^{T}$ is lower for MgH₂ than for Mg₂Si, the opposite occurs at high temperatures (e.g. 250 °C). At the imposed pressure conditions (~9 MPa), the crossover temperature is 180 °C. Thus, we argue that under LEM conditions, the reaction temperatures are below 180 °C leading to MgH₂ + Si as stable phases. In contrast, local temperatures above 180 °C occur under HEM conditions, and Mg₂Si is produced.

1.2.2 Germanium

Germanium can also be used to reduce the enthalpy of dehydrogenation from MgH₂, through the formation of magnesium germanide, Mg₂Ge (see Fig. 3) [16]. The enthalpy of formation for Mg₂Ge is 104.6 kJ/mol, which gives an enthalpy of dehydrogenation for 2MgH₂:Ge of 23.0 kJ/mol, yielding a T(1 bar) of −91.6 °C [20]. Interest in the use of Ge was initially as an additive for improving the kinetics of MgH₂ [21]. Ball-milling the elements Mg and Ge together led to the formation of Mg₂Ge dispersed in a Mg matrix. Unfortunately, the Mg₂Ge was not found to have any positive effect on the hydrogenation of the remaining Mg. Ball-milling MgH₂ with Ge was found to be an effective way to limit the formation of Mg₂Ge phase, but still produces an intimate mixture of the two starting phases [20]. It has also been found that ball-milling MgH₂ and Ge under a hydrogen atmosphere reduces the amount of Mg₂Ge formed even further, presumably through suppressing the formation of Mg metal during milling, which reacts more readily with the Ge [22].

Fig. 3

Destabilization of MgH₂ with Ge [19]

High-pressure DSC results gave a dehydrogenation endothermic peak at 235 °C (regardless of the hydrogen back pressure used, from 0 to 25 bar) [20]. Modifying the sample with a Ti-based additive reduced the dehydrogenation even further, starting at 130 °C (also confirmed in Fig. 4). The kinetics of the reaction were investigated, the unmodified system released hydrogen very quickly at 250 °C, taking only 15 min. Dehydrogenation was more sluggish at 200 °C, only reaching 60 % conversion after 60 min (Fig. 4). However, adding a Ti-based additive gave fast kinetics at 200 °C, with complete conversion within 15 min. These are very favourable results in comparison with the silicide analogues where, after 5 h at 250 °C, the reaction was still not complete for Ti-modified 2Mg:Si [23].

Fig. 4

TGA data for the dehydrogenation of 2MgH₂/Ge samples with and without a Ti-based additive. Samples were ball-milled for 140 min; TGA used a flowing Ar carrier

1.2.3 Tin

Mg₂Sn can be prepared either by powder metallurgy or by ball-milling under an argon atmosphere [16]. Powder metallurgy leads to a cubic antifluorite structure as for Ni and Ge, whereas with ball-milling, a metastable rhombohedral phase is obtained. Such a metastable phase is also observed at high pressure (60 kbar, 900 °C) and is possibly stabilized by a magnesium deficiency leading to the Mg₉Sn₅ stoichiometry instead of the 2:1 ratio. The cubic phase obtained by powder metallurgy was ball-milled under hydrogen gas (20 bar). No hydrogen uptake could be observed, but the formation of the rhombohedral phase occurred in a shorter time (3 h) than milling under argon (12 h). The decomposition of Mg₂Sn was not observed, even for long milling times, but the rapid formation of the rhombohedral phase under hydrogen can be understood by the formation of small quantities of MgH₂ leading to the Mg-deficient Mg₉Sn₅ rhombohedral compound [14].

1.2.4 Aluminium

A sample of γ-Mg₁₇Al₁₂, produced by arc-melting and BM, transformed completely to MgH₂ and metallic Al during hydrogenation (p(H₂) = 30 bar, T = 350 °C, t = 17 h) [24]. The apparent activation energy for decomposition of air-exposed MgH₂–Al is estimated as E _A ~ 160 kJ/mol H₂ [24]. This value corresponds to the activation barrier of 160–166 kJ/mol H₂ for the dehydrogenation of pure magnesium hydride [25, 26]. The alloying with Al creates an alloy which is less sensitive towards oxygen contamination and which requires little if any pre-treatment in order to activate [24].

1.3 Magnesium alloys with d-elements (Ni, Cu, Co, Ti, Pd, Fe, Zn)

1.3.1 Nickel

The addition of Ni to the Mg system can be beneficial for different reasons. Nickel is a very good (de)hydrogenation catalyst, which means that it facilitates the absorption and desorption of hydrogen by adsorbing and dissociating hydrogen molecules into adsorbed atomic species and vice versa. It is not necessary to have pure Ni in the system, and Mg–Ni alloys also show these enhanced sorption kinetics [26–28]. For instance, enhanced hydrogen sorption rates, while retaining a high storage capacity of about 6 wt% hydrogen, were obtained for Mg_1−x Ni _x alloys with different compositions (0 ≤ x ≤ 0.33) prepared by melt spinning [27]. A second aspect is that in the processing of Mg materials, such as ball-milling or nanoconfinement, Ni often has an influence on the structural properties of the resulting material [29].

However, the most important aspect is the possibility to form alloys. The thermodynamic properties for hydrogen desorption from Mg₂NiH₄ are considerably more favourable than those of MgH₂, with a formation enthalpy for Mg₂NiH₄ of about −64 kJ/mol (decreasing the equilibrium hydrogen desorption temperature under 1 bar H₂ from 300 to 250 °C). Unfortunately, the addition of nickel also significantly reduces the capacity, with Mg₂NiH₄ only containing 3.6 wt% of hydrogen [30].

Recent developments are mostly related to further insight into the properties of Mg _x Ni alloys by studying the hydrogenation properties of thin films [31, 32] and nanoparticles, prepared either by supporting on a carbon scaffold [33] or by gas-phase techniques [22, 23]. Using an in situ microscope during hydrogenation, it was possible to visualize the nucleation and growth of Mg₂NiH₄ crystals with a lattice spacing of 0.22 nm in Mg-rich domains, followed by crystallization of MgH₂, demonstrating how the presence of Mg₂Ni facilitates the hydrogenation of Mg [32]. In addition, measurements on nanoparticles produced by inert gas condensation clearly connected the enhancement of the hydrogen sorption properties to the formation of Mg₂Ni and Mg₂NiH₄ phases [23]. Mg₂Ni, particularly in a form prepared by rapid solidification nanocomposite with Mg [34], enhances the hydrogenation of the Mg-based alloys allowing higher rates of hydrogen absorption and desorption.

Since Ni destabilizes the magnesium–hydrogen bonding, this results in the reduction in the decomposition temperatures of the formed hydrides, as demonstrated for Mg–Mm–Ni alloy-based hydrides nanostructured by Equal Channel Angular Pressing in [35].The effect of nickel is associated with strong covalent Ni–H bonding which leads to the formation of various types of Ni–H complexes and clusters, [NiH₄]⁴⁻ tetrahedra in the case of Mg₂NiH₄ or buckled nets –Ni–H–Ni–H– penetrating through the structure in the recently studied MgNi₂H₃ [36]. While in both cases the metal sublattice requires a complete rebuilding, for the Laves-type intermetallic MgNi₂, a hydrogen-induced transformation takes place in the metal sublattice when a very high hydrogen pressure of 28 kbar H₂ is applied at a temperature of 300 °C. Surprisingly, instead of a disproportionation to form MgH₂, the hydrogenation leads to a rebuilding of the initial hexagonal structure into the orthorhombic distorted MoSi₂-type sublattice. In the formed MgNi₂H₃, H atoms fill the deformed octahedra Mg₄Ni₂ and the positions within the buckled nets –Ni–H–Ni–H– penetrating through the structure (Fig. 5). Analysis of the electronic structure revealed a charge transfer from Mg to Ni, and to the H atoms. The calculated gross heat of formation for the Cmca phase of MgNi₂H₃ is ~37.3 kJ/mol-H₂, which is consistent with the stability of the hydride at normal conditions.

Fig. 5

Crystal structure of MgNi₂D₃. Two types of sites occupied by deuterium include a Mg₄Ni₂ octahedron for H1(D1) and a chair Ni₄ configuration for H2(D2) located within the buckled Ni–H nets containing bent spirals –Ni–H₂–Ni–H₂–

1.3.2 Copper

No ternary hydrides have been reported in the literature for the Mg–Cu system. However, the role of copper is to substitute the hydrogen in the MgH₂ and form a stable compound, Mg₂Cu. Mg₂Cu disproportionates during hydrogenation as described in Eq. 9.

$$2{\text{Mg}}_{2} {\text{Cu}}\left( {\text{s}} \right) + 3{\text{H}}_{2} \left( {\text{g}} \right) \rightleftharpoons 3{\text{MgH}}_{2} \left( {\text{s}} \right) + {\text{MgCu}}_{2} \left( {\text{s}} \right)$$

(9)

Thus, the alloy Mg₂Cu has a theoretical hydrogen capacity of ρ _m = 2.62 wt% H, and ΔH _f = − 70 kJ/mol H₂ [37, 38].

A sample of Mg/Mg₂Cu was prepared by arc-melting Mg–Cu (85.5:14.5) [38]. The sample was hydrogenated using two different methods, (a) (p(H₂) = 6 bar and T = 325 °C), and the low hydrogen pressure should facilitate the formation of MgH₂–Mg₂Cu and (b), (p(H₂) = 30 bar and T = 325 °C) in order to facilitate the formation of MgH₂–MgCu₂. The in situ SR-PXD results [39] confirmed that the reversible alloy formed upon dehydrogenation. The apparent activation energy for the dehydrogenation of air-exposed MgH₂ /Mg₂Cu and MgH₂ /MgCu₂ sample was found to be E _A ~ 108 and ~160 kJ/mol, respectively [39]. Furthermore, substantially improved dehydrogenation kinetics of MgH₂ and resistance towards oxidation of Mg due to the presence of Mg₂Cu/MgCu₂ is observed. The apparent activation energy found for dehydrogenation of MgH₂ in MgH₂/Mg₂Cu is in agreement with previous work [38]. The activation energies extracted for the air-exposed samples may be assumed to be maximum values, which may be relevant to know for practical applications. This work also illustrates that in situ synchrotron radiation powder X-ray diffraction (SR-PXD) is a useful technique for investigation of kinetic properties [40, 41]. These results indicate that the presence of either Mg₂Cu or MgCu₂ in MgH₂ has a pronounced positive effect on the dehydrogenation kinetics of air-exposed samples.

In addition to diffusion of hydrogen, hydrogen cycling requires solid-state diffusion of Mg, transforming it from the Mg-poor to the Mg-rich state of the alloy and vice versa. Solid-state diffusion, and hence the hydrogen sorption reaction in this system, is rather slow. Hence, one of the areas of recent study is the effect of reducing particle sizes. Supported nanocrystallites with an average size of 20 nm were prepared by initial deposition of Cu nanoparticles followed by Mg deposition and hydrogenation [42]. The temperature for hydrogen desorption was ~150 °C lower than for micron-sized material, and the kinetic performance was stable upon cycling. An activation energy for the hydrogen desorption of only 97 kJ/mol was found, clearly demonstrating the benefits of a reduction in particle size for the sorption kinetics, however at the expense of hydrogen storage capacity due to the carbon support.

1.3.3 Cobalt

The Mg–Co system contains only one intermetallic compound, a Laves-type MgCo₂. However, sintering of the mixtures of the fine metal powders of Mg and Co at 450–500 °C and hydrogen pressure (90 bar), or reactive ball-milling of the stoichiometric 2:1 mixtures of Mg and Co under hydrogen gas leads to the formation of Mg₂CoH₅.

Substitution of Fe in Mg₂FeH₆ by Co results in the formation of Mg₂CoH₅ with a smaller H/M ratio of 5/3 instead of 2.0, and with a much decreased thermal stability. Hydrogen desorption from Mg₂CoH₅ is a multistep process, which at 500 °C leads to the formation of MgCo and Mg [43]. This process is, however, reversible and allows the synthesis of Mg₂CoH₅ when the mixture MgCo + Mg is heated above 300 °C under a hydrogen pressure of 100 bar.

Hydrogenation of the MgCo₂ intermetallic alloy has been tried at similar high-pressure conditions to those applied to MgNi₂. However, the metal sublattice becomes destabilized by Co which disproportionates to Mg₂CoH₅ and Co. The solid-solution-type ternary alloys MgNi_2−x Co _x (x = 0.25; 0.5, and 0.75) also are unstable compared with disproportionation into Mg₂(Ni,Co)H_5−x , even with the smallest content of Co at x = 0.25 (Antonov, Denys, Yartys, unpublished, 2015).

1.3.4 Titanium

Hydrogen storage behaviour of Mg nanoparticles (size range 100 nm to 1 μm) with metal-oxide core–shell morphology synthesized by inert gas condensation and decorated by titanium clusters via in situ vacuum deposition has been reported to improve only the kinetics [44]. Beside the alloying effect with transition metals, composite materials made of Mg with immiscible metals can also be prepared [36, 45]. Ponthieu et al. [46] investigated the Mg–Ti–H system focusing on MgH₂–TiH₂ nanocomposites. They observed that this mixture leads to materials exhibiting very fast sorption kinetics. The samples were prepared by ball-milling of various mixtures of Mg and Ti under hydrogen pressure. In this way, nanocomposites were prepared in the range x = 0 to 0.5 for the system (1 − x)MgH₂−xTiH₂. They consist of MgH₂ (with the coexistence of α- and γ-polymorphs) and ε-TiH₂ homogenously distributed with crystallite sizes below 15 nm. A study of the sorption properties showed that only the magnesium hydride is reversible under moderate conditions (T < 330 °C, p(H₂) < 1 MPa), and the thermodynamics of the Mg/MgD₂ equilibrium is not affected. However, the presence of the titanium hydride leads to very fast hydrogen sorption kinetics for the magnesium system. Indeed, TiH₂ inclusions with particle sizes of 10 nm are found homogeneously distributed within the MgH₂ matrix, and a coherent coupling between titanium hydride inclusions and the Mg/MgD₂ matrix has been reported. This coupling is expected to enhance hydrogen mobility within the nanocomposite components as TiH₂ inclusions limit grain growth of both Mg and MgH₂ phases preserving short diffusion paths for hydrogen. Furthermore, TiH₂ is expected to favour the diffusivity of hydrogen through the existence of coherent coupling between TiH₂ and Mg/MgH₂. Though TiH₂ does not influence the Mg/MgD₂ thermodynamics, it provides fast sorption kinetics, allowing the nanocomposites to absorb large amounts of hydrogen at room temperature. These kinetic properties are attributed to the limited grain growth of Mg and MgD₂ phases, the coherent coupling between TiH₂ and Mg/MgH₂, and the fast H-diffusion through sub-stoichiometric MgD_2−η and TiD_2−η phases.

1.3.5 Palladium

The Mg–Pd phase diagram is rather complex and comprises several intermetallic phases. The Mg₆Pd phases have attracted most interest for hydrogen storage, as it is the lightest and least costly intermetallic. Mg–Pd composite nanoparticles were synthesized by inert gas condensation of Mg vapours followed by vacuum evaporation of Pd clusters. Formation of the Mg₆Pd intermetallic phase takes place upon vacuum annealing, resulting in Mg/Mg₆Pd, which transforms to different Mg–Pd intermetallic compounds upon hydrogen absorption, depending on temperature and pressure [47].

Mg₆Pd nanoparticles of 4 nm were obtained by infiltration of Mg on previously formed Pd nanoparticles dispersed into the pores of porous carbon [48]. Mg infiltration resulted in the in situ formation of intermetallic Mg₆Pd nanoparticles. From the EXAFS-based structural analysis, it was demonstrated that the crystal structure of the nanoparticles is different than bulk Mg₆Pd. The nanoparticles exhibit a simpler crystallographic arrangement and a higher atomic disorder. As for the structural properties, thermodynamic and kinetic sorption properties also show different behaviours than for the bulk material. A change in the plateau pressure (Fig. 6) indicating thermodynamic destabilization of the hydride phase has been reported. In addition, faster kinetics were observed for the Mg₆Pd nanoparticles (Fig. 7) that remain stable for at least 10 sorption cycles though the particle size increased from 4 nm in the as-synthesized compound to 10 nm after hydrogenation. This modification of the sorption properties is related to the small particle size, as well as the difference in the crystal structure compared to that of bulk Mg₆Pd.

Fig. 6

Hydrogen sorption isotherms at 275 °C for the bulk (black line) and nanoconfined (red line) Mg₆Pd. Full and empty symbols stand for absorption and desorption isotherm curves, respectively [48]

Fig. 7

Thermal hydrogen desorption mass spectra of hydrogenated bulk (black line) and nanoconfined (red line) Mg₆Pd (m/e = 2, heating rate 5 °C/min) [48]

1.3.6 Iron

Magnesium iron hydride is known to have one of the highest volumetric hydrogen density of ρ _v  = 150 g H₂/L, moderate gravimetric storage density ρ _m = 5.47 wt% H and can be prepared from inexpensive chemicals [49–51]. The Mg₂Fe alloy is stabilized by hydrogen, i.e. Mg₂FeH₆ forms in a reaction between magnesium hydride and iron at T > T _dec(MgH₂). The reaction (Eq. 10) for a sample of MgH₂–Fe (2:1) in the temperature range 380–470 °C, p(H₂) = 100 bar, was studied using in situ SR-PXD. The hydrogen pressure stabilizes the MgH₂ and facilitates the reaction.

$$2{\text{MgH}}_{2} \left( {\text{s}} \right) + {\text{Fe}}\left( {\text{s}} \right) + {\text{H}}_{2} \left( {\text{g}} \right) \to {\text{Mg}}_{2} {\text{FeH}}_{6} \left( {\text{s}} \right)$$

(10)

Mg₂FeH₆ was cooled to RT where it is stable and then heated in argon (p(Ar) = 1 bar). A series of in situ SR-PXD patterns revealed that Mg₂FeH₆ apparently decomposes directly to the metals according to reaction (Eq. 11) in agreement with previous investigations and the fact that T _dec(Mg₂FeH₆) > T _dec(MgH₂), ΔH _dec(Mg₂FeH₆) = 87 ± 3 kJ/mol H₂ [52].

$${\text{Mg}}_{2} {\text{FeH}}_{6} \left( {\text{s}} \right) \rightleftharpoons 2{\text{Mg}}\left( {\text{s}} \right) + {\text{Fe}}\left( {\text{s}} \right) + 3{\text{H}}_{2} \left( {\text{g}} \right)$$

(11)

1.3.7 Zinc

Mg–Zn alloys have been investigated for hydrogen storage but have not been found to improve the storage properties of Mg [52]. Ex situ X-ray diffraction has been used to show that the Zn forms an intermetallic alloy with Mg, which upon absorption of hydrogen either remains unaffected [52] or forms a higher Zn content intermetallic alloying more Mg to form a hydride [53, 54]. However, it has also been shown by in situ neutron diffraction that a Zn-containing liquid phase forms at the temperatures typically used for hydrogenation [55]. No Zn hydride or Mg–Zn ternary hydride has been observed.

1.4 Mg-containing quasicrystals

Quasicrystals (QCs) are materials with long-range order but which lack three-dimensional translational periodicity. The first of these structures discovered was a rapidly solidified Al–Mn alloy [56]. Over a hundred binary and ternary QC intermetallic systems have been discovered; half of these are metastable and require rapid solidification production techniques, such as splat cooling or melt spinning [57]. The initial QC systems were mostly Al-based; however, Mg-containing QCs in the Mg–Zn–(Y, RE) system were reported by Luo et al. [58]. QC structures present the opportunity to develop new hydride systems, with unusual physical and mechanical properties [59] and high H/M ratios because there are more interstitial sites for hydrogen compared to conventional hydrides. A limited number of studies have explored the hydrogen storage properties of QC systems, most of which focus on Ti-based QCs, for example Ti–Zr–Ni and Ti–Hf–Ni. These achieved relatively high hydrogen storage capacities (H/M = 1.6), but the recovery of the QC phase was difficult, due to the formation of ZrH₂ or TiH₂ and Ti₂Ni alloy after one (de)hydrogenation cycle [60, 61]. Sahlberg and Andersson [62] initially investigated the hydrogenation properties of a cubic Mg₃Y₂Zn₃ ternary alloy with limited success. It only hydrogenated above 400 °C and had a low hydrogen storage capacity (0.3 wt%).

More recently, Luo et al. [63] presented a study on the hydrogenation properties of a Zn–Mg–Y quasicrystal and its ternary alloys. The approximate stoichiometry of the stable Zn–Mg–Y icosahedral QC (i-phase) was confirmed to be Zn₆₀Mg₃₀Y₁₀. Three nominal compositions were prepared: ZMY-1, Zn₅₀Mg₄₂Y₈ by gas atomization in an argon atmosphere; ZMY-2, Zn₆₀Mg₃₀Y₁₀ by induction melting; and ZMY-3, Zn₆₀Mg₃₀Y₁₀ by gas atomization followed by annealing at 585 °C. The as-prepared samples were ball-milled under 1 bar Ar for 2 h. A quasicrystal i-phase was observed in all three as-prepared Zn–Mg–Y alloy samples, (Fig. 8). Despite using different preparation methods and initial nominal compositions, ZMY-1 and ZMY2 showed similar XRD patterns consisting of two major phases, i-phase and a H-phase (hexagonal (Zn,Mg)₅Y alloy). Unlike the other two samples, ZMY-3 exhibited peaks only associated with the i-phase.

Fig. 8

X-ray diffraction patterns of as-prepared Zn–Mg–Y samples. ZMY-1 and ZMY-2 consisted of three phases including a QC i-phase, H-phase, and Mg₇Zn₃, while the ZMY-3 sample consisted of a single i-phase [50]

The SEM image of the as-prepared ZMY-1 sample (Fig. 9a) shows that spherical particles of sizes smaller than 200 μm were formed through gas atomization, with a distinctive structure on the surface, while cross-sectional SEM backscattered electron (BSE) images of ZMY-1 and ZMY-2 (Fig. 9b, c) indicated the presence of i-phase and H-phase in both samples, with different microstructures. ZMY-1 was formed of a petal-like i-phase surrounded by small white dots (ca. 1 μm) from the H-phase, while ZMY-2 consisted of dendritic H-phase structures located in the middle of the i-phase. The i-phase dominated ZMY-3, accounting for >95 % of the volume fraction based on cross-sectional BSE imaging.

Fig. 9

SEM backscattered electron (BSE) images of a surface as-prepared ZMY-1, b cross-sectional ZMY-1, and c cross-sectional ZMY-2 [50]

The annealed Zn–Mg–Y QC sample (ZMY-3) had a 0.3 wt% hydrogen storage capacity and a high DSC decomposition temperature (503 °C), while higher hydrogen storage capacity (0.9 wt%) and lower decomposition temperature (445 °C) were achieved with ZMY-1, which had a greater H-phase contribution. The H-phase contains a greater proportion of Y that may cause the expansion of the unit cell and thus facilitate the hosting of more hydrogen atoms, as in ZMY-1. In addition, the lower decomposition temperature may be attributed to a smaller i-phase grain resulting in faster kinetics. TGA results showed that ZMY-1, ZMY-2, and ZMY-3 released 1.0 ± 0.1, 0.3 ± 0.1, and 0.4 ± 0.1 wt% of hydrogen, respectively, which was in close agreement with the PCI curves. A hydrogenation plateau (of ca. 3.5 bar) was observed on its PCI curve at 300 °C which suggests an improvement on the thermodynamics of MgH₂ (1.4 bar at 300 °C). Further work is required to investigate the (de)hydrogenation reaction mechanism and speed up the reaction kinetics.

1.5 RE–Mg–Ni hydrides (RE = La, Pr, Nd)

Despite significant differences in chemistry between the rare earth metals (RE) and magnesium, Mg forms extensive solid solutions in the RENi₃ intermetallic alloys, crystallizing with a PuNi₃-type trigonal structure. Up to 67 % of RE atoms can be replaced by Mg to form REMg₂Ni₉ intermetallic compound. The RENi₃ crystal structures are formed by a stacking of the RENi₅ (Haucke CaCu₅ type) and RENi₂ (Laves type) slabs along the trigonal 00z axis (RENi₅ + 2MgNi₂ = REMg₂Ni₉). The substitution Mg → RE takes place within the Laves-type slabs only leaving RENi₅ layers unmodified (RENi₅ + 2MgNi₂ = REMg₂Ni₉). In fact, a small yet significant replacement of RE by Mg (5 %) was observed in La_0.91Mg_2.09Ni₉ within the AB₅ slabs leading to the formation of the over-stoichiometric compositions RE_1−x Mg_2+x Ni₉ crystallizing with PuNi₃-type structure.

A gradual increase in Mg content in RE_3−x Mg _x Ni₉ is accompanied by a linear decrease in the volumes of the unit cells (Fig. 10). Interestingly, a substantial contraction takes place not only for the (RE,Mg)₂Ni₄ slabs, but also for Mg-free CaCu₅-type RENi₅ slabs [64].

Fig. 10

Relative changes in the volumes of the unit cells and constituent fragments of the La_3−x Mg _x Ni₉ structures during the Mg → La substitution. The data for the individual LaNi₅ and MgNi₂ compounds are given as a reference

Magnesium significantly affects the bonding mechanism of metal–hydrogen interactions studied in detail for La₂MgNi₉-based hydride [65]. A prominent and the most important feature of the structure of La₂MgNi₉D₁₃ is in the local ordering of the hydrogen sublattice built by stacking of the MgH₆ octahedra and NiH₄ tetrahedra (Fig. 11).

Fig. 11

Hydrogen sublattice for the La₂MgNi₉D₁₃ structure. The sublattice is formed by the stacking of MgD₆ octahedra and NiD₄ tetrahedra. Such stacking gives an irregular, buckled framework

Substitution of the largest (in size) atom of the rare earth metals—lanthanum—by smaller atoms like Pr or Nd leads to the contraction of the volumes of the unit cells of (La,Pr,Nd)₂MgNi₉, reaching 2.6 % for Nd₂MgNi₉ [66]. The hydrogen storage capacity of the alloys remains unchanged, 12–13 at.H/f.u.(La,Pr,Nd)₂MgNi₉. However, shrinking of the lattice quite significantly affects the thermodynamic stability of the formed hydrides. As evident from the room-temperature isotherms (Fig. 12), equilibrium pressures of both hydrogen absorption and desorption in the Nd₂MgNi₉–H₂ system are an order of magnitude higher than in the La₂MgNi₉–H₂ system changing from <0.1 bar for the La₂MgNi₉-based hydride to >1 bar for Nd₂MgNi₉H₁₂.

Fig. 12

PCT diagrams for Nd₂MgNi₉–H₂, Pr₂MgNi₉–H₂, and La₂MgNi₉–H₂ systems

The properties of RE_3−x Mg _x Ni₉-based hydrides dramatically change when the content of magnesium extends to the maximum solubility limit, close to two atoms of Mg/f.u. for RE_3−x Mg _x Ni₉. In situ neutron powder diffraction studies of two deuterides La_1.09Mg_1.91Ni₉D_9.5 and La_0.91Mg_2.09Ni₉D_9.4 performed at deuterium pressures reaching 918 bar D₂ showed that the hydrogenation properties of Mg-rich compounds are principally different from those for LaNi₃ [67]. LaNi₅-assisted hydrogenation of MgNi₂ in the hybrid structure of LaMg₂Ni₉ takes place. In the deuterides, the D atoms are accommodated in both Laves- and CaCu₅-type slabs. The overall chemical compositions can be presented as LaNi₅H_5.6/5.0 + 2 * MgNi₂H_1.95/2.2 showing that the hydrogenation of the MgNi₂ slab proceeds at mild H₂/D₂ pressure of just 20 bar. A partial filling by D of the four types of the tetrahedral interstices in the MgNi₂ slab takes place, including [MgNi₃] and [Mg₂Ni₂] tetrahedra. Magnesium causes a significant destabilization of the hydrides; La_1.09Mg_1.91Ni₉ reversibly forms hydrides with ΔH _des = 24.0 kJ/mol H₂ (as compared to ΔH _des = 35.9 kJ/mol H₂ for La₂MgNi₉ [66]) and an equilibrium pressure of H₂ desorption of 18 bar at 20 °C (0.05 bar at 20 °C for La₂MgNi₉ [65]). However, magnesium also introduces a pronounced hysteresis of H₂ absorption and desorption, ~100 bar. The latter is probably caused by a huge difference in the dimensions of the Ni net in the individual crystal lattices of MgNi₂ and LaNi₅.

1.6 Magnesium compounds for high-temperature energy storage

Thermochemical gas–solid reactions are suitable as long-term or short-term heat storage systems. Important for short-term heat storage systems is good kinetics of the absorption and desorption of the gaseous compound combined with good heat conductivity of the reaction bed. After separation of both compounds during the decomposition process, all of this energy is trapped in the system and can be used at any time without energy losses (long-term storage). Reversible Mg-based metal hydride systems show a rather high range of energy densities (62–77 kJ/mol H₂; see Table 1) making their use convenient in thermochemical heat storage systems, even though the values 62–77 kJ/mol H₂ are significantly lower as compared to the highest observed values for the metal hydrides, YH_1.8 (220 kJ/mol H₂), ZrH_1.6 (212 kJ/mol H₂) and CaH₂ (207.7 kJ/mol H₂) [68]).

While research in past years has been focused on the pure Mg/MgH₂ system, which can be used for temperatures up to 400 °C in solar power parabolic trough systems, new developments now focus on the 2 Mg + Fe/Mg₂FeH₆ system. This material can be used at temperatures up to 550 °C as a possible heat storage material for the solar tower power plants and shows excellent stability over hundreds of hydrogenation and dehydrogenation cycles (Eq. 12).

$$2{\text{Mg}} + {\text{Fe}} + 3{\text{H}}_{2} \mathop{\rightleftarrows}\limits_{\text{heat\,storage}}^{\text{heat\,release}}{\text{Mg}}_{ 2} {\text{FeH}}_{6}$$

(12)

Current research is focused on the identification of cost-effective materials undergoing hydrogenation–dehydrogenation reactions (heat storage–heat release) with a high degree of reversibility upon multiple cycles and fast and controllable kinetics [69]. A wide range of temperatures with different Mg-based metal hydrides between 400 and 600 °C can be covered.

1.7 The HDDR processes in the Mg-based alloys (LaMg₁₂, LaMg₁₁Ni)

The hydrogenation–disproportionation–desorption–recombination (HDDR) process includes two stages of hydrogen absorption, normally at relatively lower and higher temperatures promoting, respectively, the formation of interstitial hydrides with expanded cells and the disproportionation of the metallic matrix of this hydride accompanied by the formation of binary hydrides and other phases with reduced (zero) content of the hydride-forming element, e.g. the products of disproportionation.

The high-temperature desorption of hydrogen is accompanied by the recombination of the source compound but already with sub-micrometre grain sizes. For the magnesium-rich compounds of rare earth metals, because of a very large enthalpy of formation of both MgH₂ (−74 kJ/mol H₂) and rare earth metals (−208 kJ/mol H₂ for LaH₂), the formation of the interstitial-type hydrides becomes thermodynamically unfavourable as compared to the formation of the hydrides of Mg and rare earth metals. However, even though the decomposition of the rare earth hydrides needs application of extremely high temperatures, more than 800 °C (in vacuum) for LaH₂, it becomes destabilized by magnesium metal formed during hydrogen desorption from MgH₂. This makes possible a recombination of the intermetallics of rare earths and magnesium below 450 °C. The latter temperature is 400 °C lower than the temperature required to decompose individual LaH₂.

In the case of LaMg₁₂, hydrogen absorption resulted in a two-step disproportionation process: LaMg₁₂ + H₂ → LaH₃ + Mg → LaH₃ + MgH₂ [33]. It is interesting that instead of immediate formation of MgH₂, the magnesium metal was formed first, before interacting with hydrogen to yield a magnesium dihydride.

A decrease in the grain size of the initial alloy was achieved by applying a rapid solidification (RS) process, and this was found to assist in improving the hydrogenation kinetics. It is interesting to note that ribbons of the RS samples retained their initial shapes after the hydrogenation despite the fact that formation of LaH₃ and MgH₂ hydrides is accompanied by a significant volume expansion.

Hydrogen thermal desorption in vacuum occurred in different ways for the alloys solidified at different quenching rates. A major peak of hydrogen evolution was located at ∼370 °C; however, for the alloys synthesized by rapid solidification and having the smallest grain size, an extra desorption event was observed at 415 °C (Fig. 13). SR XRD study showed that this extra peak is associated with Mg-assisted low-temperature hydrogen desorption from LaH₂ proceeding below 450 °C and leading to a recombination process to form the initial intermetallic alloy LaMg₁₂.

Fig. 13

Hydrogen thermal desorption mass spectra of the hydrogenated RS alloys measured in vacuum

The chemical equations taking place in these two processes can be described as follows:

• 1. Desorption without recombination (RS samples prepared at low solidification rate and as-cast samples)

  $${\text{MgH}}_{2} + {\text{LaH}}_{3} \to {\text{Mg}} + {\text{LaH}}_{2} + 3/2{\text{H}}_{2}$$

• 2. Desorption followed by recombination of LaMg₁₂ (RS samples prepared at high solidification rate)

  $$12{\text{MgH}}_{2} + {\text{LaH}}_{3} \to 12{\text{Mg}} + {\text{LaH}}_{2} + 25/2{\text{H}}_{2} \to {\text{LaMg}}_{12} + {\text{H}}_{2}$$

SEM studies revealed significant differences between the samples after completion of the thermal desorption of hydrogen. Samples for which the recombination stage was observed were found to have extremely porous morphology (pore sizes 0.5–2 μm; see Fig. 14). Such dramatic microstructural changes can be attributed to the recombination of the LaMg_12−x intermetallic during H desorption and associated with volume shrinkage during the Mg + LaH₂ → LaMg_12−x  + H₂ transformation. No such transformation took place in the RS samples prepared at low solidification rate and the as-cast samples, which is attributed to their coarse initial microstructures. These observations show that fine nanostructuring of the material leads to a complete recombination process and to the enhancement of hydrogen exchange.

Fig. 14

Backscattered SEM images of the RS LaMg₁₂ samples (a RS300; b RS1000; c RS2000) and RS2000 sample at different stages of hydrogen processing (d after TDS from the hydride; e, after five hydrogen absorption–desorption cycles; f hydrogenated sample). For the inserted magnified square areas, the sides are 5 μm

From comparison of the data for LaMg₁₂ [33] with the results for the rapidly solidified LaMg₁₁Ni alloy where 1/12 of Mg was substituted by Ni [34, 35], it has been concluded that Ni substitution in combination with rapid solidification significantly improves the hydrogenation absorption–desorption properties of the alloy leading to a dramatic increase in the rates of hydrogenation–dehydrogenation processes as a result of the nanostructuring of the samples and because of the catalytic influence of the formed Mg₂Ni intermetallics. On hydrogen desorption, the presence of mobile hydrogen in the metal lattice allows a low-temperature diffusion of both lanthanum and magnesium to proceed. Because of the thermodynamic advantage of the formation of La–Mg intermetallics compared to the mixture of Mg metal and lanthanum dihydride, this causes a recombination reaction to proceed below 450 °C, 400 °C lower than the decomposition temperature of LaH₂, to sequentially form LaMg_12−x , and later, La₂Mg₁₇ intermetallics (Fig. 15).

Fig. 15

Evolution of the in situ SR XRD pattern from the hydrogenated LaMg₁₁Ni alloy during hydrogen vacuum thermal desorption in the temperature range 100–500 °C (heating rate 2 °C/min; SNBL BM01A; λ = 0.72085 Å). Peaks marked FCC correspond to a LaHO oxyhydride

Ribbons of the RS samples retained their initial shapes despite the fact that formation of LaH₃ and MgH₂ hydrides is accompanied by a significant volume expansion.

2 First-principle study on Mg-based hydrides

Between interstitial metal hydrides, such as LaNi₅H₆, and ionic complexes, such as Mg₂NiH₄, a large family of (La,Mg)-Ni hydrides exists, exhibiting intermediate electronic properties. The host intermetallics belong to the (La,Mg)Ni _y family, with y = (5n + 4)/(n + 2). It has been shown [70] that such intermetallics are stabilized by the stacking along the c-axis of n * [LaNi₅] + m * [LaMgNi₄] layers of Haucke and Laves phases, respectively, with Mg atoms occupying only half La sites of the latter. In this context, a systematic investigation of the electronic properties of the (La,Mg)₂Ni₇H _x hydrides has been recently done [71]. To understand the preferential interstitial H sites, a systematic comparison of 13 different sites showed that both geometrical and chemical factors govern the stability of the hydrides of the RE–Mg–Ni family. It has been shown that the calculated site preferences are in agreement with the experimental data, for which H occupies [La₂Ni₄] slab in La₂Ni₇ compound mainly for geometrical reasons, whereas H goes in both [(La,Mg)₂Ni₄] and [LaNi₅] slabs for the Mg-substituted compound driven by both geometrical and chemical factors. The La_1.5Mg_0.5Ni₇ hydrides show electronic charge transfer from La and Mg to H. Depending on the H site environment, the Bader charge analysis gives a value of about −0.31e on H, which classifies these hydrides as conventional metallic ones, such as LaNi₅H _x .

Another interesting study on Mg-based hydrides has been recently published on the trihydride MgNi₂H₃ [36, 72]. The DFT calculations showed the formation of an electronic structure around −10 to −6 eV caused by the chemical bonds of hydrogen and its 1s states mainly via interaction with the Ni 3d states. The bonding mechanism is dominated by the formation of covalent bonds between Ni and H and a charge transfer from Mg to the two different H atoms which leads to a significant degree of ionic bonding between Mg (+1.59e) and H (−0.55e, −0.31e). The calculated enthalpy of formation of MgNi₂H₃ is approximately -30 kJ/mol H₂, which is consistent with the stability of a hydride at normal conditions. Moreover, the analysis of calculated phonon dispersion curves has confirmed that the structure is mechanically stable in Cmca symmetry (Fig. 16), the most stable one among all considered alternatives. In addition, the thermodynamic properties of MgNi₂H₃ with temperature and volume dependence have been estimated in the frame of the quasiharmonic approximation.

Fig. 16

Partial phonon density of state (DOS) of MgNi₂H₃ in the Cmca space group

The variety of chemical bondings within Mg-based hydrides is illustrated in Fig. 17, showing the electron localization function (ELF) representations for several compounds. The probability function ranges between 0 (no probability) and 1 (perfect localization) for the valence electrons, 0.5 indicating probability similar to electron gas. Whereas MgH₂ and Mg₂NiH₄ show isolated anions around H atoms with a partial negative charge, and tetrahedral [NiH₄]²⁻complex, respectively, the hydrides La_1.5Mg_0.5Ni₇H _x and MgNi₂H₃ exhibit a non-null ELF region (>0.25) between Mg and H. This behaviour fully shows that the partial covalent interaction between Mg and H is typical for a variety of magnesium-containing hydrides.

Fig. 17

Electron localization function (ELF) images of selected planes containing Mg and H atoms into MgH₂ (P4₂/mnm), Mg₂NiH₄ (C12), MgNi₂H₃ (Cmca), and La_1.5Mg_0.5Ni₇H₈ (P6₃/mmc)

3 Concluding remarks

Magnesium-based compounds continue to be the subject of intensive research for hydrogen storage, due to their light weight and inexpensive availability of magnesium. Practical limitations on the application for the pure metal, such as slow kinetics and high temperatures of desorption, require the development of compound materials to overcome these limitations. While improvements in kinetics have been achieved with nanostructuring processes, such as ball-milling, rapid solidification and other processes that introduce defects, and combination with relatively small amounts (<10 wt%) of additives [8], improvements in the thermodynamics are necessary to substantially reduce dehydrogenation temperatures. Alloys and intermetallic compounds of magnesium provide a way of reducing the enthalpy of the reaction compared to the pure metal hydride.

A large number of materials, mostly metals, have been used to form magnesium compounds and the thermodynamics of their interaction with hydrogen has been characterized. In many cases, the new hydrogenated compounds have reduced enthalpies; however, problems such as slower kinetics, instability, and disproportionation upon hydrogen cycling often arise. Sometimes the preparation process introduces metastable phases that, although beneficial to the interaction with hydrogen, are lost after the initial hydrogen cycling. In addition, the introduction of another material to the magnesium typically results in a reduction in hydrogen capacity.

Binary compounds of magnesium have been investigated quite thoroughly, and it may be that the feasible alloys for effective, low-temperature hydrogen storage materials have been exhausted. Ternary and higher alloys may offer new possibilities in this area.

Finally, a new area of use of the Mg-based hydrogenated compounds appears to be heat storage. Here, for the entire range of temperatures, future research will include the identification of the most promising candidates together with the synthesis and the full thermodynamic/thermophysical and kinetics characterization [73]. Moreover, hydrides working at relatively low equilibrium pressure (between 1 and 50 bar) and at a certain selected temperature will be preferred in order to minimize the complexity and the cost of the thermal energy storage system.

The present special issue contains two other closely related papers which could be of interest for the readers entitled “Review of magnesium hydride based materials: development and optimisation” [74] and “Integrated with FC H storage systems utilising magnesium hydride: experimental Studies and modelling” [75].