Using Ball Milling for Modiﬁcation of the Hydrogenation / Dehydrogenation Process in Magnesium-Based Hydrogen Storage Materials: An Overview

: Magnesium-based hydrogen storage materials are considered to be one of the most promising solid-state hydrogen storage materials due to their large hydrogen storage capacity and low cost. However, slow hydrogen absorption / desorption rate and excessive hydrogen absorption / desorption temperature limit the application of magnesium-based hydrogen storage materials. The present paper reviews recent progress in improving the hydrogen storage properties by element substitution and additives. Ball milling is the promising technology for preparing magnesium-based hydrogen storage materials. The research and development of approaches for modifying magnesium-based hydrogen storage materials prepared by ball milling is systematically expounded. It is concluded that ball milling can signiﬁcantly improve the kinetic and electrochemical properties of magnesium-based hydrogen storage materials and increase the hydrogen storage capacity. In the future, the research of magnesium-based hydrogen storage materials should be developed in terms of hydrogen storage mechanism, computer design of materials and development of a more optimized catalytic system. substituting Ni with Cr. The results showed that after Ti substituted Mg, there were two main phases Mg 2 Ni and Ni and the trace amount of Mg and TiNi phases in the alloy. The TiNi phase


Introduction
As a very important functional material, hydrogen storage materials play an irreplaceable role in the field of secondary energy, especially in the research of fuel cells and rechargeable batteries. The study of hydrogen storage materials is directly related to the application of electric vehicles, and also has an important impact in submarines, spacecraft and other applications. Various hydrogen storage materials have been developed worldwide [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15], and many researchers have focused their research on the study of magnesium-based hydrogen storage materials. Much attention is paid to the study of magnesium-based materials due to the fact that magnesium has many important properties. First of all, magnesium is one of the most abundant elements in the earth's crust, ranking the 8th, accounting for 2.35% of the crustal mass. Secondly, magnesium has large hydrogen storage capacity, theoretical value of which can reach 7.6wt% [16][17][18][19]. However, as a hydrogen storage material pure magnesium has poor kinetic performance for hydrogen absorption/desorption, conditioned by the following: (1) existence of oxide on its surface, (2) slow diffusion rate of hydrogen in the bulk MgH 2 /Mg, (3) with poor decomposition of hydrogen on the Mg surface [20] the rate of hydrogen absorption/desorption is very low, and the hydrogen desorption temperature is relatively high. To achieve an acceptable rate of hydrogen desorption, the desorption temperature must be above 350 • C, which makes the application of magnesium difficult.
The methods of hydrogen storage materials preparation have significant influence on the hydrogen storage properties. Widely used methods include hydriding combustion synthesis, melting method, sintering process, ball milling etc. Ball milling is an important structural modification method developed in the 1980s. Compared with ball milling, other methods mentioned above possess certain limitations to be considered. For example, in spite of the hydrogen storage materials prepared by the hydriding combustion synthesis having high hydrogen absorption/desorption activity and defect density, the hydriding combustion reaction is complicated, the experimental repeatability is poor, and the granularity of reaction products is irregular [21]. Therefore, the hydriding combustion synthesis often needs to be combined with ball milling. The melting method has strict requirements on the melting point of the metal elements of the hydrogen storage alloy [22]. Mechanical alloying based in ball milling has unique advantages for preparation of alloys with large difference in melting points. Mechanical alloying, with no need to heat and completely relying on mechanical action, enables not only easy control of the composition of the alloy, but also allows to directly obtain the materials with metastable structure such as nanocrystalline structure, amorphous structure and supersaturated solid solution. In terms of material preparation methods, ball milling is the main method for preparing magnesium-based hydrogen storage alloys and can cause interdiffusion and alloying reactions between different elements to produce alloys with micron-or even nano-size grains. For the research purposes, ball milling preparation and nanocomposite are one of the main ways to improve the hydrogen absorption/desorption properties of magnesium-based hydrogen storage alloys. This is due to the fact that ball milling introduces a large number of defects and grain boundaries in magnesium-based hydrogen storage alloys. Moreover, magnesium-based hydrogen storage alloys prepared by ball milling are usually nano-size multiphase system, a large number of interfaces in which and the above-mentioned defects and grain boundaries provide significant channels for the diffusion of hydrogen atoms between grains. In addition, ball milling leads to significant reduction of grain size, thereby accelerating the diffusion of hydrogen atoms inside grains [23][24][25][26][27][28][29]. Theoretically ball milling can break the oxide layer on magnesium powder and provide fresh surfaces [18], however, because of the limitation of the vacuum degree of the ball-milled vial, nanocrystalline Mg always interacts with oxygen during ball milling, especially during hydrogen absorption/desorption cycles, which lead to the oxygen layer reform [30]. In this paper, we paid especial attention to the substitution and additive-enhanced MgH 2 and compared the effects of different modification methods and the modification mechanisms in as much detail as possible. Reviewing this data, we presume this paper can assist in better understanding of MgH 2 , and shape perspectives for further research.

Modification by Element Substitution of Mg 2 Ni
The hydrogen storage capacity of Mg 2 Ni is 3.6%, and it is necessary to release hydrogen gas above 200 • C. Mg 2 Ni belongs to A 2 B type hydrogen storage alloy, A is a hydrogen-absorbing element or an element with a strong affinity for hydrogen. A usually belongs to IA-VB group metal elements, such as Ti, Zr, Ca, Mg, V, Nb, Re (Rare earth element) etc., reaction of which with hydrogen is exothermic (∆H <0). B is the element that absorbs little hydrogen or no hydrogen at all. B is usually Fe, Co, Ni, Cr, Cu, Al etc., reaction of which with hydrogen is endothermic (∆H > 0). A element controls the hydrogen storage capacity, while B element controls the reversibility of hydrogen absorption/desorption and regulates the equilibrium pressure of hydrogen desorption [31][32][33]. Element substitution and changing the proportions of elements are the common method for improving the properties of magnesium-based alloys, the core of which is to adjust the hydrogen absorption capacity and hydrogen absorption/desorption properties of the hydrogen storage alloy by adjusting the proportion and type of A and B element. The purposes of element substitution of Mg 2 Ni are as follows: (1) improving hydrogen desorption pressure; (2) reducing hydrogen desorption temperature; (3) improving hydrogen absorption/desorption rate. The purposes can be realized due to the fact Metals 2019, 9,768 3 of 25 that the endothermic metal in the hydrogen storage alloy is able to control the reversible process of hydrogen absorption/desorption. The existing element substitution methods in the literature can be summarized as follows: (1) Substitution of Mg with an element that is more electronegative than Mg to reduce the stability of the hydride and improve the reversibility; (2) Substitution of Ni with an element, which radius is larger than that of Ni; (3) Substitution of Mg and Ni with various elements.
After substituting part of the Mg elements in the Mg 2 Ni alloy with Ag, it was found that the new alloy had lower hydrogen desorption temperature, absorbing hydrogen at 423 K without any activation, and the hydrogen absorption capacity was also reduced. This can be explained due to the substitution of Mg leading to reduction of the absolute value of the enthalpy [44,45].
The experimental results of substitution of Mg element with Al showed that with increase of the addition of Al the hydrogen storage capacity of the alloy will be reduced. This is due to the substitution of Mg with Al, which reduces the formation of the hydrogen absorbing phase Mg 2 Ni. However, the addition of Al element will increase the hydrogen absorption/desorption rate of Mg 2 Ni hydrogen storage alloy compared with that before substitution. Furthermore, the hydrogen absorption temperature of the alloy decreases with the increase of Al%, and the addition of Al can effectively improve the discharge capacity and cycle stability. Several possibilities explaining the prominent effect of Al additives on the hydrogen storage properties of Mg 2 Ni were considered [46][47][48][49][50]: (1) the addition of Al element leads to formation of non-hydrogen-adsorption phase Al 3 Ni 2 and hydrogen-adsorption phase Mg 3 AlNi 2 in the Mg 2 Ni hydrogen storage alloy, these new phases increase grain boundaries and defects in the alloy, which provide significant hydrogen diffusion channels; (2) The synergistic hydrogen absorption effect of Mg 3 AlNi 2 and Mg 2 Ni can significantly improve the hydrogen absorption/desorption properties of the alloy; (3) After the Al element is added, a dense layer of Al 2 O 3 is formed on the surface of the alloy, and the corrosion resistance of the alloy is also enhanced, thereby the cycle life of the alloy is prolonged. These explanations were also confirmed by the research results of Xue et al. [51] and Yuan et al. [35].
The hydrogen absorption/desorption properties of Mg 2 Ni at low temperatures (below 300 • C) can also be improved by substituting Mg with Ti. The work of Liang et al. [52] has certain representativeness. Mg 1.9 Ti 0.1 Ni nanoalloy was prepared by substituting Mg element with Ti element. The alloy had stable hydrogen absorption/desorption properties and quickly absorbed hydrogen at 200 • C without any activation. Further to that, at 150 • C 3 mass% absorption was achieved within half an hour. Iwakura et al. [53] prepared Mg 1.9 Ti 0.06 V 0.04 Ni by partially substituting Ti element in Mg 1.9 Ti 0.1 Ni with V element. It was found that compared with substitution of Mg with a single Ti element or V element substitution of Mg element simultaneously with Ti and V elements exhibited further improved cycle performance. Improvement of the hydrogen absorption/desorption properties and cycle stability of the alloy by Ti element can be explained by Ti element forming an oxide film on the surface of the alloy, which restrains further oxidation of the Mg element, and at the same time forms a nickel-rich layer on the surface of the alloy, thereby improving the cycle stability of the alloy [54].
Substituting Mg with Zr is also one of the ways to improve the hydrogen storage properties of Mg 2 Ni. Substituting Mg with Zr allowed not only improve the hydrogen desorption properties of the alloy electrode at 30 • C, but also enhanced the hydrogen absorption/desorption capacity of the alloy at low temperatures (30 • C and 200 • C) [55]. Goo and Lee [56] attributed the enhanced hydrogen absorption capacity of the Mg-Ni-Zr alloy to the increase of the atomic spacing in the alloy and the increase of the active point, which promoted the hydrogen dissolve ability of the alloy. Feng et al. [57] believed that the proper addition of Zr can improve the corrosion resistance of the alloy, but maximum discharge capacity of the alloy can be probably reduced.
In addition, the substitution of Mg in Mg 2 Ni with Re elements can also improve the hydrogen storage properties of Mg 2 Ni series alloys to certain extent [55].
Yang et al. [64] prepared Mg 2 Ni 0.75 M 0.25 alloy by substituting Ni element with M (M = Ti, Mn, Co, Cr, Zn, Cu, Fe) and investigated their hydrogen storage properties. The results showed that the substitution of Ni with Mn, Co, Cr leaded to obvious reduction of the decomposition equilibrium pressure of Mg 2 Ni, while the substitution with Ti element and Cu element caused increase in the equilibrium pressure, with regard to Zn element and Fe element, with their little effect on the results observed.
Zhang et al. [65] prepared Mg 2 Ni 1−x Zr x alloy by substituting Ni element with Zr element. It was found that the main phase of the alloy was a hexagonal structure as that of Mg 2 Ni and with the gradual increase of the Zr element content the hydride formation heat was gradually reduced, and the hydrogen desorption temperature was also lowered. Compared with Mg 2 Ni alloy it was also found that the hydrogen desorption capacity of the substituted with Zr alloy increased at 340 • C and under the pressure of 1 atm.
Pd has a large catalytic activity for the hydrogen absorption/desorption of alloys and is a suitable additive for mechanical ball milling. The dispersed in Mg 2 Ni nanosized Pd particles play a catalytic role for the hydrogen absorbing reaction, thereby improving the hydrogen storage properties of the alloy. However, as the Pd content increases, the hydrogen storage capacity of the alloy decreases. The reason for this lower hydrogen capacity is still not clear and may be related to slower hydrogen diffusion through the initially formed layer of Mg 2 NiH x [66,67].
Of all works on partial substitution of element Ni with Cu, the work of Chen et al. stands out. The researchers found that with the increase of Cu content, the alloy formed Mg 2 Cu phase and Cu 11 Mg 10 Ni 9 phase. Besides, the results of their research showed that the addition of Cu allows to increase the hydrogen desorption equilibrium pressure that is consistent with the findings of Yang [64] et al., but leads to reduction of hydrogen storage capacity of the alloy. Two possibilities explaining the effect of Cu additives on the hydrogen storage capacity of Mg were considered: (1) Since the atomic radius of Cu is larger than Ni, the substitution of Ni with Cu leads to decrease in the material density, the volume expansion ratio after hydrogen absorption, and thereby the hydrogen storage capacity; (2) The production of new phase Cu 11 Mg 10 Ni 9 reduces the formation of hydrogen absorption phase Mg 2 Ni. In spite of the negative effect of Cu additive on the hydrogen storage capacity of the alloy, it should be noted that adding Cu element can effectively improve discharge capacity and cycle stability [68,69].

Modification by Substituting Mg and Ni Element
Experiments have also been conducted on the partial substitution of elements on both sides of A and B in Mg-Ni series alloys. Generally, new phase structures are formed. Xie et al. reported microstructure of the Mg 1.8 Al 0.2 Ni 0.8 Cr 0.2 alloy, obtained by simultaneous substituting Mg and Ni with Al and Cr respectively. It was found that there is a synergistic effect between Al and Cr additives: Cr has the ability to inhibit the formation of non-hydrogen-adsorption phase Al 3 Ni 2 , and simultaneously facilitates the formation of hydrogen-adsorption phase Mg 3 AlNi 2 ; similarly, Al can inhibit the formation of non-hydrogen-adsorption phased −[Cr, Ni], such microstructure of Mg 2−x Al x Ni 1−y Cr y alloy is beneficial to hydrogen storage capacity and cyclic stability [47].
Wang et al. [70] prepared the Mg 2−x Ti x Ni 0.8 Cr 0.2 (x = 0.05, 0.10, 0.15, 0.20) alloy by substituting Mg with Ti and substituting Ni with Cr. The results showed that after Ti substituted Mg, there were two main phases Mg 2 Ni and Ni and the trace amount of Mg and TiNi phases in the alloy. The TiNi phase was dispersed in the alloy and had a certain catalytic effect on the hydrogen absorption/desorption properties of the alloy. The more Ti content is in the alloy, the faster the hydrogen desorption rate of the alloy is. However, with the increase of Ti content the hydrogen desorption capacity was reduced. This is due to the reduction of Mg content in the alloy, which leads to the reduction of the absolute enthalpy of the hydride. Duan et al. [71] investigated the structure and properties of Mg 1.5 Al 0.5 Ni 1−y Co y and found that when y = 0.2, the alloy capacity retention rate was optimal and the hydrogen desorption equilibrium pressure was the highest and widest. The hydride decomposition enthalpy of Mg 1.5 Al 0.5 Ni 0.8 Co 0.2 alloy is -39.45 kJ/mol H 2 , i.e., the hydride stability of the alloy increases after Co substitutes Ni.

Mg-Ni Series Alloys
In order to further improve the cycle life of Mg-Ni series hydrogen storage alloys while maintaining their high hydrogen storage capacity, in recent years, scientists have been using mechanical alloying method to prepare stoichiometry or non-stoichiometric Mg-Ni series amorphous alloys and progressed in research of the substitution modification of alloy elements [11,72,73]. Elements, widely used for alloying with Mg-Ni series materials, are V, Ti, Co and rare earth elements.
Zhu et al. [74] used Hydriding Combustion Synthesis (HCS) and Balling Milling technology to prepare Mg 90 Ni 10−x V x (x = 0, 2, 4, 6, 8) alloys. It was observed that the hydrogen absorption properties of Mg-Ni-V alloy were better than that of binary Mg-Ni alloy. This is explained by the fact that the addition of V promotes the hydrogenation of Mg to some extent. While the hydrogen desorption properties of Mg-Ni-V alloy were inferior to that of binary Mg-Ni alloy, which can be explained by the fact that during hydrogen desorption the catalytic effect of Mg 2 Ni is better than V, with the increase of V content in the alloy the content of VH y phase gradually increased, while the content of Mg 2 NiH 4 and Mg 2 NiH 0.3 gradually decreased.
Ti alloying is one of the hot research in recent years. The doping of Ti can weaken the interaction between H atoms and Ni atoms, reduce the hydrogen absorption reaction enthalpy of the alloy system, and improve the hydrogen-decomposing ability of the hydride of Mg 2 Ni [75]. However, the doping of Ti can also lead to the reduction of hydrogen storage capacity of Mg-Ni series alloys, Wang et al. [76] suggested that this may be related to the TiNi 2 phase appearing in the Mg-Ni-Ti alloy, which is not conducive to the gaseous hydrogen absorption of the alloy. Although the doping of Ti leads to decrease in the hydrogen storage capacity of Mg-Ni series alloys, the amount of MgH 2 formed by doping Ti is the highest compared to the doping of Ni and Nb into the Mg-Ni system [77], which suggests that Ti allows to commendably improve hydrogen storage properties of Mg-Ni series alloys.
Similarly, Co alloying can well improve hydrogen storage properties of Mg-Ni series alloys. Grigorova et al. [78] reported that the composites 85%Mg-15%Mg 2 Ni 1−x Co x (x = 0, 0.1, 0.3) preserved high absorption capacity with decreasing temperature (4 wt% H 2 at the hydriding temperature of 423 K). Similar results have been observed by Zhao et al. [79].
Sun et al. prepared Mg 22 Y 2 Ni 10 Cu 2 + x%Ni (x = 0, 50, 100, 150; mass fraction) composite materials with different ball milling time (10, 20, 30, 40 h) and analyzed the effects of milling time and nickel addition on the hydrogen storage properties of Mg 22 Y 2 Ni 10 Cu 2 alloy. The results showed that ball milling improved the microstructure of the alloy and promoted the formation of amorphous and nanocrystalline in the alloy, besides, the addition of nickel significantly promoted the progress of the above process, so that the content of amorphous and nanocrystalline in the composite material greatly increased. The improvement of cycle stability was also obvious: S 20 of the ball-milled for 20 h sample with composited nickel x = 150, was up to 80%. In addition, nickel, coated on the surface of the alloy, improved the surface activity of the alloy and effectively improved the overall hydrogen storage properties of the alloy [80].
The influence of elements such as Fe [81] and different Ni content [82] on hydrogen storage properties and electrochemical properties of Mg-Ni series alloys has also been reported in literature. These elements all improve the above properties of the alloy to varying degrees.
Details of the investigation results of the part of the above selected works on modification by element substitution of Mg 2 Ni are presented in Table 1.
The doping of V allows to significantly reduce the hydrogen desorption activation energy and increase the cycle stability of hydrogen absorption/desorption [85,89]. Liang et al. [86], using the ball milling method, produced nanocomposites MgH 2 -5%Tm (molar percentage) (Tm = Ti, V, Mn, Fe, Ni) and conducted the systematic study of their hydrogen absorption/desorption properties. The results showed that these five transition metals catalyzed well the hydrogen absorption/desorption processes in MgH 2, and, in addition, the obtained nanocomposites were better in all properties than the pure ball milled MgH 2. In the range of 30~200 • C with respect to hydrogen absorption properties, the addition of Ti provided the fastest hydrogen absorption rate, and then V, Fe, Mn, and Ni, while with respect to hydrogen desorption properties, the most effective hydrogen desorption reaction was obtained by adding V, and then Ti, Fe, Ni and Mn. In the range of 250~300 • C the MgH 2 -Ti composite best improves the kinetics of hydrogen desorption, while at temperatures below 250 • C MgH 2 -V has the best kinetics of hydrogen desorption, even at 200 • C, 0.015 MPa can desorb hydrogen. Further studies of Liang et al. [90] showed that the good hydrogen desorption properties in MgH 2 -5%V are due to the catalytic effect of vanadium and the powder of small-size vanadium particles and are also associated with the formation of the α-Mg nucleus on the VH 0.81 /MgH 2 phase interfaces as a result of the strong affinity interaction between the vanadium and hydrogen atoms.
During ball milling and hydrogen absorption/desorption in Mg-based alloys the added transition metals sometimes form new metal hydrides, which can provide hydrogen diffusion channels, thereby improving hydrogen absorption/desorption kinetics. With ball milling MgH 2 -5%Nb (molar percentage) for 20 h, followed by dehydrogenation, a metastable new phase NbH 0.6 was formed. This phase was finally decomposed into Nb metal after approximately 200 s. This metastable new phase of Nb-hydride can be considered as an ordered structure with hydrogen vacancies that provide channels for the flow of hydrogen during the dehydrogenation of nanocomposites [91][92][93]. De Castro et al. [94] found that the NbH 2 new phase appeared in the MgH 2 -5%Nb (molar fraction) nanocomposite prepared by ball milling under hydrogen atmosphere, while under the same preparation conditions in the ball-milled pure Nb. NbH 2 phase was not observed. This indicates that magnesium contributes to the hydrogenation of Nb, and Nb also contributes to the hydrogenation of magnesium. De Castro et al. attributed this synergistic behavior to the kinetic catalytic effects of the Mg/Nb nanointerface. In addition, the new phases produced during the ball milling process can also inhibit the reaggregation of powder particles during hydrogen absorption/desorption cycles [95].
Some of the doped metal catalysts can function as "hydrogen pumps" during hydrogen desorption. Janot et al. [96] prepared MgNi + 5wt%Pd using ball milling. It was found that after Pd was added, MgNi alloys began to desorb hydrogen at 150 • C, while the hydrogen desorption capacity increased to 1.5% (mass percent). This is mainly due to the catalytic effect of Pd, which acts as a "hydrogen pump" and accelerates the migration and diffusion of hydrogen across grain boundaries, transporting hydrogen from the inside of the alloy to the surface of the alloy.
The particle size of the doped metal catalysts also significantly influences the catalytic effects. Lee et al. [97] found that different sizes of Co particles had different catalytic effects on MgH 2 . The researchers compared the catalytic effects of Co particles with sizes of 0.5~1.5 microns, 3~7 microns and 44 microns and concluded that with the size of 0.5~1.5 microns the catalytic effect was the best, whereas with the size of 3~7 microns it was the worst.
The lattice type of the alloy also affects its hydrogen storage properties. Luo et al. [98] using ball milling prepared Mg x Ti 100−x (x = 35, 50, 65, 80) alloys. It was observed that ball milled Mg 35 Ti 65 alloy with BCC structure had 3.8% hydrogen storage capacity at 8 MPa hydrogen pressure and 200 • C, and showed good hydrogen absorption kinetics, while ball milled Mg x Ti 100−x (x = 50, 65, 80) alloys with FCC phase or HCP (main phase) + FCC mixed phases had a smaller hydrogen storage capacity, which was less than 1.2%.
Details of the investigation results of the part of the above selected works on modification using transition metal additives are presented in Table 2.  Metal oxides, as a catalyst, on the one hand, can play the role of hydrogen absorption/desorption catalyzing, on the other hand, since the oxide is brittle, they can play the role of grinding aid refining particles during ball milling which helps to form nanomaterials. The reason for improved kinetics by doping transition metal oxides lies in the following: transition metal oxides are sensitive to hydrogen molecules thus even at lower temperatures a large amount of hydrogen molecules can be adsorbed on the surface, so that the hydrogen molecules are in an activated state and the bond energy between the hydrogen atoms is weakened, which makes the hydrogen molecules be decomposed easier into hydrogen atoms. Since the catalysis is usually bidirectional, the addition of a transition metal oxide is also advantageous for hydrogen desorption [99][100][101][102][103][104][105][106][107].
Widely ) had excellent catalytic effect. This is due to the fact that oxides of polyvalent metals allow increase the probability of electron exchange reaction between oxides and hydrogen molecules so that the gas-solid reaction is accelerated. Doping metal oxides with variable valence also allows improve the hydrogen absorption/desorption cycle stability of the alloy [111,117] V 2 O 5 has a good catalytic effect. It can faster chemically adsorb hydrogen molecules on its surface. This adsorption effect can even directly decompose hydrogen molecules into hydrogen atoms, which can make hydrogen atoms enter the metal lattice faster, thereby greatly accelerating the hydrogen absorption process. For instance, nanocrystalline composites Mg-Ni-V 2 O 5 can adsorb hydrogen up to more than 6.2% (mass percent) at 200 • C in 60 s; at 300 • C, 0.1 MPa can desorb hydrogen up to 6.0% (mass percent) in 700 s. doping V 2 O 5 not only improves hydrogen absorption/desorption rate, but also decreases the hydrogen desorption temperature by about 30 • C [118].
For multivalent metal oxide catalysts, the hydrogen absorption/desorption cycles are usually accompanied by oxidation/reduction of the oxide. For example, for nanocrystalline composites Mg-3Ni-2Cr 2 O 3 it is believed that the two-position mechanism where two Cr 3+ function as an activation center for hydrogen absorption/desorption may serve as the catalytic mechanism of Cr 2 O 3 [119]. According to the crystal field theory, the d 3 configuration of Cr 3+ ions in Cr 2 O 3 is the high-spin octahedral configuration. Under the effect of hydrogen molecules or hydrogen atoms Cr 3+ in Cr 2 O 3 can be reduced to Cr 2+ , obtaining an electron from a hydrogen atom. Due to the fact that in a high-spin octahedral field, the stabilization energy of the d 3 configuration of Cr 2+ ions is less than that of the d 3 configuration of the Cr 3+ ions (i.e., the stability of Cr 3+ is greater than the stability of Cr 2+ ), an unstable state is formed. Cr 2+ ion again loses the electron and turns from an unstable state into Cr 3+ . At the same time H + turns to hydrogen atom, obtaining an electron from Cr 2+ ion. Thus, the kinetic balance is built. Similar results have been observed by Song et al. [120]. Song et al. has also found that the addition of an oxide catalyst such as Cr 2 O 3 contributes to the refinement of the Mg powder during the ball milling process.
Barkhordarian et al. [112,121] investigated and compared the average hydrogen desorption rate of MgH 2 from 20% to 80% of the maximum hydrogen storage capacity using various metal oxide catalysts and concluded that Nb 2 O 5 is the most catalytically active transition metal oxide. Figure 1 showed the hydrogen desorption rate of MgH 2 with different doped metal oxides at 573K in a vacuum state. Using the kinetic rate constants at 300 • C and 250 • C. Barkhordarian et al. reported that the desorption activation energy varies exponentially with Nb 2 O 5 concentration and reaches the limit of 61 kJ/mol at 1 mol%. Further, the researchers also presumed that much poorer catalytic effect of the pure metals than that of the respective oxide is possibly due to pure metal itself not having the catalytic effect, as most pure metals exhibit the catalytic effect because there are oxides in the samples. However, this kind of explanation has not been confirmed, as in order to study the catalytic effect, the samples are not to be exposed to ambient air at any time. Oliver et al. [122] investigated hydrogen absorption mechanism of MgH 2 /Nb 2 O 5 and using the channel model interpreted the reason why Nb 2 O 5 can greatly improve the hydrogen desorption kinetics of magnesium. In the first hydrogen desorption process, a redox reaction occurs between Nb 2 O 5 and magnesium. The resulting magnesium-niobium oxide is dispersed in the sample and appears on the surface of the sample. Since Nb has different valence states, binary magnesium-niobium oxide channels having different stoichiometry are formed. These channels are distributed in the sample to facilitate the diffusion of hydrogen and enhance hydrogen desorption kinetics of the sample.
the d 3 configuration of the Cr 3+ ions (i.e., the stability of Cr 3+ is greater than the stability of Cr 2+ ), an unstable state is formed. Cr 2+ ion again loses the electron and turns from an unstable state into Cr 3+ . At the same time H + turns to hydrogen atom, obtaining an electron from Cr 2+ ion. Thus, the kinetic balance is built. Similar results have been observed by Song et al. [120]. Song et al. has also found that the addition of an oxide catalyst such as Cr2O3 contributes to the refinement of the Mg powder during the ball milling process.
Barkhordarian et al. [112,121] investigated and compared the average hydrogen desorption rate of MgH2 from 20% to 80% of the maximum hydrogen storage capacity using various metal oxide catalysts and concluded that Nb2O5 is the most catalytically active transition metal oxide. Figure 1 showed the hydrogen desorption rate of MgH2 with different doped metal oxides at 573K in a vacuum state. Using the kinetic rate constants at 300 °C and 250 °C. Barkhordarian et al. reported that the desorption activation energy varies exponentially with Nb2O5 concentration and reaches the limit of 61 kJ/mol at 1 mol%. Further, the researchers also presumed that much poorer catalytic effect of the pure metals than that of the respective oxide is possibly due to pure metal itself not having the catalytic effect, as most pure metals exhibit the catalytic effect because there are oxides in the samples. However, this kind of explanation has not been confirmed, as in order to study the catalytic effect, the samples are not to be exposed to ambient air at any time. Oliver et al. [122] investigated hydrogen absorption mechanism of MgH2/Nb2O5 and using the channel model interpreted the reason why Nb2O5 can greatly improve the hydrogen desorption kinetics of magnesium. In the first hydrogen desorption process, a redox reaction occurs between Nb2O5 and magnesium. The resulting magnesium-niobium oxide is dispersed in the sample and appears on the surface of the sample. Since Nb has different valence states, binary magnesium-niobium oxide channels having different stoichiometry are formed. These channels are distributed in the sample to facilitate the diffusion of hydrogen and enhance hydrogen desorption kinetics of the sample. Details of the investigation results of the part of the above selected works on modification using metal oxide additives are presented in Table 3. Table 3. Hydrogen storage properties of selected Mg modified by metal oxide additives. Details of the investigation results of the part of the above selected works on modification using metal oxide additives are presented in Table 3.

Transition Metal Halide Additives
Many metal halides have a good catalytic effect on the hydrogen desorption temperature and hydrogen absorption kinetics of magnesium-based hydrogen storage materials [123][124][125][126][127][128]. Jin et al. [129,130] investigated the effect of transition metal fluorides on the hydrogen absorption/desorption properties of MgH 2 . As shown in Figure 2, many fluorides had a positive catalytic effect on the hydrogen absorption kinetics of MgH 2. Among them, NbF 5 and TiF 3 significantly improved the hydrogen absorption kinetics of MgH 2 . Jin et al. believed that the key to improving the hydrogen absorption kinetics of MgH 2 is the single-phase metal hydrides, formed in the reaction of MgH 2 with these transition metal fluorides during ball milling or hydrogen absorption. These unique single-phase metal hydrides are believed to effectively inhibit the growth of the MgH 2 nanoparticles during balling milling and hydrogen absorption/desorption cycles and there by maintaining a better catalytic activity in the hydrogen absorption/desorption cycles.  It should be noted that different amounts of additives NbF5 and different ball milling times have different catalytic effects on the hydrogen absorption/desorption properties of MgH2 [131]. Luo et al. [132] investigated the influence of different amounts of additives NbF5 on catalytic effect. Results showed that when adding 2% (molar percentage) NbF5, the most rapid kinetics was achieved. After ball milling of MgH2 + 2%NbF5 (molar percentage) for 5 h, at 300 °C 5 mass% absorption was achieved within 12 min and 6 mass% absorption was achieved within 60 min. The addition of NbF5 leaded to It should be noted that different amounts of additives NbF 5 and different ball milling times have different catalytic effects on the hydrogen absorption/desorption properties of MgH 2 [131]. Luo et al. [132] investigated the influence of different amounts of additives NbF 5 on catalytic effect. Results showed that when adding 2% (molar percentage) NbF 5 , the most rapid kinetics was achieved. After ball milling of MgH 2 + 2%NbF 5 (molar percentage) for 5 h, at 300 • C 5 mass% absorption was achieved within 12 min and 6 mass% absorption was achieved within 60 min. The addition of NbF 5 leaded to decrease in the hydrogen desorption temperature by 64 • C and the hydrogen desorption activation energy was reduced to 90 kJ/mol, decreasing by 30 kJ/mol compared to pure MgH 2 . Recham et al. [133] also believed that 2% NbF 5 (molar percentage) is the best doping amount, which leads to significant reduction in hydrogen desorption temperature of MgH 2.
It is worth noting that the anion species of the halides have a great influence on their catalytic effects: usually the catalytic effect of chloride is worse than that of fluoride. Malka et al. [134] compared the influence of different chlorides and fluorides on the decomposition temperature of MgH 2 as measured by temperature programmed desorption (TPD). As shown in Figure 3, the fluorides, as compared to the chlorides, had much stronger influence on the MgH 2 decomposition process. In all chlorides only TiCl 3 lead to the reduction of the desorption temperature down to below 300 • C. In the case of the fluorides ZrF 4, NbF 5 , TaF 5 , TiF 3 and FeF 2 a strong catalytic influence on MgH 2 decomposition was exhibited. The largest reduction of hydrogen desorption temperature was observed for the ZrF 4 additive, which allowed hydrogen desorption start below 250 • C. This data is in agreement with the results presented by Jin et al. [129], de Castro et al. [135] and Deledda et al. [136]. The fact that fluorides can significantly influence the desorption properties of MgH 2 may be caused by the presence of the F anions, which weaken the Mg-H bonding and led to the formation of MgF 2 , providing electron-rich centers to trap the transition metal atoms. In addition, the F anion may tailor the electronic structure of the transition metal atom to influence its activity for hydrogen dissociation/recombination. The ZrF 4 additive exhibits a different mechanism and is unchanged after milling and desorption/absorption cycle. It is possible that the catalytic function of ZrF 4 is due to the presence of both Zr atoms and F anions in the sample that facilitated the recombination of H 2 molecules on the surface of ZrF 4 [137,138].
Metals 2019, 9, x FOR PEER REVIEW 12 of 28 may tailor the electronic structure of the transition metal atom to influence its activity for hydrogen dissociation/recombination. The ZrF4 additive exhibits a different mechanism and is unchanged after milling and desorption/absorption cycle. It is possible that the catalytic function of ZrF4 is due to the presence of both Zr atoms and F anions in the sample that facilitated the recombination of H2 molecules on the surface of ZrF4 [137,138]. There are also studies showing that mechanical alloying of halides and alloys of Mg or Mg-Ni both reduces the size of the metal powder and modifies the surface of metals (destroy the oxide layer on the metal surface), thereby stimulating the hydrogen absorption/desorption, in particular, stimulating the first hydrogen absorption/desorption [139]. There are also studies showing that mechanical alloying of halides and alloys of Mg or Mg-Ni both reduces the size of the metal powder and modifies the surface of metals (destroy the oxide layer on the metal surface), thereby stimulating the hydrogen absorption/desorption, in particular, stimulating the first hydrogen absorption/desorption [139].
Details of the investigation results of the part of the above selected works on modification using transition metal halide additives are presented in Table 4.
Yu et al. [149] prepared Mg-3Ni-2MnO 2 -0.25CNTs using mechanical ball milling. The results showed that hydrogen capacity of this material reached 7.0%, the hydrogen absorption reaction was quickly completed within 100 s, the hydrogen desorption reaction was completed within 600 s at 0.1 MPa, and the temperature of the hydrogen desorption platform was approximately 280 • C. Yu et al. considered that the excellent hydrogen storage properties of the CNTs-doped alloy are related to the strong adsorption ability of carbon nanotubes for hydrogen molecules and the excellent heat transfer performance of carbon nanotubes.
It is believed that there is the synergistic effect between ball milling and carbon material. On the one hand ball milling can shorten the length of carbon nanotubes, open the tube port, increase the surface area, and increase the hydrogen adsorption capacity of carbon nanotubes [150] On the other hand, with increase of absorbed carbon component on freshly exposed Mg surfaces, the impermeable surface oxide/hydroxide layer is not reformed and the composite is readily accessible to hydrogen, which enhances the ball milling effect [151]. The inhibition effect of carbon materials on reformed oxides can be explained by two different mechanisms: one is that in the initial stage of the milling process, a stack of carbon layers is adsorbed on freshly created Mg surfaces, which lower the surface tension of the compound, thereby reducing the driving force responsible for the diffusion of oxygen atoms to the surface, therefore to certain extent altering the formation of the impermeable surface oxide layer. Besides, the breaking of C-C bonds in the carbon layer and the presence of radicals yield to the formation of highly reactive C species which might be oxidizing enough to react with oxygen-containing species that are either present at the surface of Mg or diffuse to the surface [152]. It is worth noting that unhybridized electrons of carbon fragments produced from the ball-milling process can not only absorb oxygen-containing species, but also interact with hydrogen molecules, which increases the hydrogen diffusion driving force [153][154][155].
Zhou et al. [156,157] found that microlitic carbon has a graphite-like crystal structure, and like graphite and graphene, has a good grinding effect on magnesium, the former can prevent the powder from agglomerating, and can quickly refine the magnesium particles. The particle size of magnesium powder ball milled for 2 h was down to 30-50 nm when doped by 40% microlitic carbon. In addition, it was found that between microlitic carbon and magnesium there is a synergistic effect on hydrogen storage. This synergistic effect is concluded in the following two aspects: (1) Under the action of magnesium, hydrogen molecules decompose into hydrogen atoms, which can either form ≡CH, =CH 2 , -CH 3 as a result of combination of hydrogen atoms with broken bonds on the surface of carbon particles (chemical absorption), or enter into the space between carbon microcrystalline layers, thereby hydrogen is stored in carbon (physical absorption); (2) Microlitic carbon can significantly reduce hydrogen desorption temperature of MgH 2 . This is due to the function of microlitic carbon as a grinding aid, which allows reduce the size of magnesium grains, increase the specific surface, increase structural defects and increase the activity of dehydrogenation. Unlike Zhou et al., based on crystal structure characteristics of microlitic carbon and magnesium Lu and Tang [158] explained the synergistic effect between microlitic carbon and magnesium: (1) The disordered structure of long chains in combination with the ordered structure of short chains in microscopic carbon determines their good grindability, greater brittleness, good dispersion and lubricity, which facilitates pulverize and disperse magnesium particles during ball milling, thereby fastening nanocrystallization; (2) Mutual peritectics are formed between the atoms of carbon and magnesium, which on the one hand, provides for the catalyst with carriers, so that the catalyst can be distributed more homogeneously in the hydrogen storage material, and on the other hand, provides the channels for the diffusion of hydrogen atoms.
Zhang et al. [159] studied the effect of carbon material doping on the hydrogen desorption properties of MgH 2 and its microscopic mechanism according to the first principle. The calculations showed that the intrinsic reason for the improved hydrogen desorption properties of MgH 2 by the carbon material doping is that its doping reduces the dissociation energy of H atoms from the MgH 2 matrix, weakening the bond strength of Mg-H.
Wu et al. compared the effect of single CNTs, the single metal catalyst and combination of CNTs and metal catalyst on hydrogen storage properties of MgH 2 [160][161][162][163]. They found that CNTs allowed to significantly increase hydrogen storage capacity, however, it was not as good as a metal catalyst in terms of increasing the hydrogen absorption/desorption rate. In addition, it was found that the synergistic effect of the metal catalyst and carbon enabled higher hydrogen storage capacity and faster hydrogen absorption rate. Research of Li et al. proved this point [164]. The co-catalytic effect for hydrogenation may be understood as a spillover mechanism through the interaction of facile p-electrons of carbon with metallic catalyst [165].
Details of the investigation results of the part of the above selected works on modification using carbon additives are presented in Table 5.
Li et al. [175,176] investigated the influence of various La-Ni intermetallic catalysts on hydrogenation behavior of Mg-based alloys. The results showed that Mg-60%LaNi 5 had relatively high activity, so that 2.37 mass% absorption was achieved within 15 min at the room temperature, and at 553 K, the maximum hydrogen absorption capacity was 4.23 mass%. For Mg-30%LaNi 2.28 the maximum hydrogen absorption capacity at 553 K was 5.4 mass%. Liang et al. [177] found that the ternary Mg-Ni-La alloy exhibited much better absorption/desorption kinetics than that of the binary alloys Mg-La and Mg-Ni, besides, lanthanum hydride had strong catalytic effect on hydrogen absorption of Mg, but weak effect on hydrogen desorption. Similar results have been observed by the other research group [178].
Zhou et al. [179,180] investigated the effect of various Ti, V intermetallic catalysts on hydrogen storage properties of magnesium hydride. As shown in Figures 4 and 5, TiMn 2 doped MgH 2 had the lowest hydrogen desorption temperature and exhibited drastically improved hydrogen absorption kinetics at the room temperature. As shown in Figure 6, V 75 Ti 5 Cr 20 provided the largest hydrogen storage capacity at the room temperature, which was comparable to TiMn 2 doped MgH 2.      Another fast-developing intermetallic compound is the Zr-based alloy. 10% ZrNi-MgH2 (mass fraction) completed 5% hydrogen absorption at 300 °C within 3 min and it took only 7 min to release the same amount of hydrogen, which suggested the excellent catalytic effect of Zr-based intermetallic compounds [187]. Makihara   Another fast-developing intermetallic compound is the Zr-based alloy. 10% ZrNi-MgH 2 (mass fraction) completed 5% hydrogen absorption at 300 • C within 3 min and it took only 7 min to release the same amount of hydrogen, which suggested the excellent catalytic effect of Zr-based intermetallic compounds [181]. Makihara et al. [182] reported that Mg-50wt%ZrMn 2 completed 2.6% hydrogen absorption at 275 • C, while maximum hydrogen capacity of Mg-50wt%ZrFe 1.4 Cr 0.5 was 3.2% at 300 • C and hydrogen desorption was completed at 0.1 MPa within 15 min [183].
Wang et al. [184] prepared Mg x /Mg 2 Ni (x = 10%, 30%, 50%) composite alloys by mixing and ball milling of the powder of Mg 2 Ni alloys and magnesium. In contrast to the fact that with separate ball milling of Mg 2 Ni alloys Mg 2 Ni are decomposed into Mg and Ni, these mixed powders do not form new phases during ball milling due to the effective inhibition of decomposition of the Mg 2 Ni phase by adding magnesium powder.
The intermetallic alloys have even more pronounced catalytic effects. It is not clear at present time, however, why the intermetallic alloys would have more significant effects than their elemental counter parts. One observation is that these alloys form less stable hydrides than their corresponding elemental metals.
Details of the investigation results of the part of the above selected works on modification using intermetallic compound additives are presented in Table 6.

Summary
In this paper, different modification methods by ball milling were reviewed. It is clear that ball milling is the promising method for preparing hydrogen storage alloys, especially hydrogen storage alloys with amorphous and nanocrystalline structures, as it significantly improves their hydrogen storage properties, including kinetic properties, electrochemical properties and cycle stability. Ball milling has opened up new avenues for the field of synthetic magnesium-based hydrogen storage materials. However, ball milling still possesses certain limitations. For example, in hydrogen absorption/desorption cycles defects caused by ball milling will disappear, nanograin will agglomerate, and MgO will reform.
Adding catalysts such as transition metals, metal oxides, intermetallic compounds, transition metal halides, and carbon materials to magnesium-based hydrogen storage materials can not only improve the hydrogen absorption/desorption kinetics, but also appropriately reduce the hydrogen absorption/desorption temperature. Moreover, there is probability for the synergistic effect between catalysts and ball milling process, effectively improving the limitations of ball milling. The catalytic effects of different types of catalysts are quite different. In general, transition metal halides, multivalent transition metal oxides and carbon materials have better catalytic effects. These types of catalysts not only promote the refinement of the metal powder during ball milling, but also modify the metal as well as inhibit the reformation of MgO and the agglomeration of crystal grains in hydrogen absorption/desorption cycles, thus further improving the hydrogenation-dehydrogenation kinetics and the cycle stability. However, all current modification approaches only improve one or two aspect of hydrogen storage properties, the comprehensive properties have not been significantly improved, i.e., to date no single catalyst can significantly lower the hydrogen desorption temperature, increase the hydrogen absorption/desorption rate, or increase the cycle stability of hydrogen absorption/desorption while maintaining the high hydrogen storage capacity of pure magnesium. From this review it can be concluded that each type of additive has its own advantages and limitations. In general, carbon additives are recommended if an increase in hydrogen storage capacity is desired, while niobium-containing fluoride or oxide additives are recommended if the kinetics of hydrogen absorption/desorption (lowering the hydrogen absorption/desorption temperature, increasing the rate of hydrogen absorption/desorption) are to be well improved. Furthermore, the vanadium additive is a good choice for improving the stability of hydride absorption/desorption cycles. Characteristics of the different additives mentioned above should be taken into account when new materials are designed.
The action mechanism of various catalyst systems is still unclear, and many conclusions still differ. In view of the current research status and development trends, we believe that there are two key research directions in the future: (1) The research of the hydrogen storage mechanism of various magnesium-based hydrogen storage materials, especially magnesium-based alloy materials and magnesium-based composite materials. Through the calculation of materials based on atomic scale (computer design of materials), selecting the appropriate alloy composition and ball milling parameters (ball milling time, ball-to-powder weight ratio, Ball mill speed, etc.) will be an effective way to realize large-scale practical use of magnesium-based hydrogen storage materials; (2) Development of a more optimized catalytic system by combining different types of catalysts based on current catalytic systems.