Complex Metal Borohydrides: From Laboratory Oddities to Prime Candidates in Energy Storage Applications

Despite being the lightest element in the periodic table, hydrogen poses many risks regarding its production, storage, and transport, but it is also the one element promising pollution-free energy for the planet, energy reliability, and sustainability. Development of such novel materials conveying a hydrogen source face stringent scrutiny from both a scientific and a safety point of view: they are required to have a high hydrogen wt.% storage capacity, must store hydrogen in a safe manner (i.e., by chemically binding it), and should exhibit controlled, and preferably rapid, absorption–desorption kinetics. Even the most advanced composites today face the difficult task of overcoming the harsh re-hydrogenation conditions (elevated temperature, high hydrogen pressure). Traditionally, the most utilized materials have been RMH (reactive metal hydrides) and complex metal borohydrides M(BH4)x (M: main group or transition metal; x: valence of M), often along with metal amides or various additives serving as catalysts (Pd2+, Ti4+ etc.). Through destabilization (kinetic or thermodynamic), M(BH4)x can effectively lower their dehydrogenation enthalpy, providing for a faster reaction occurring at a lower temperature onset. The present review summarizes the recent scientific results on various metal borohydrides, aiming to present the current state-of-the-art on such hydrogen storage materials, while trying to analyze the pros and cons of each material regarding its thermodynamic and kinetic behavior in hydrogenation studies.


Introduction
A world with ever more scarce fossil fuels must transition to alternative energy storing materials. Fossil fuels, such as coal, gas, and oil, are finite, non-renewable resources, and a current estimation based on their ongoing burning rate points to a complete depletion by the end of 2060. While they are more reliable than currently engineered materials, they also pollute the environment and are responsible in great part for the greenhouse effect [1][2][3][4][5][6][7][8][9]. While the risk of running out of fossil fuels has been advocated for decades, the timeline is approaching and humanity needs to make a change. Therefore, alternative energy carriers are long sought-after; among them, metal hydrides and complex borohydrides have a special role [10][11][12][13][14][15][16][17][18]. Reaching or exceeding the U.S. Department of Energy (DOE) target is no easy task; the specific target for 2020 was 1.5 kWh/kg (4.5 wt.% hydrogen) and 1.0 kWh/L (0.03 Kg hydrogen/L), costing more than USD 10 /kWh (or USD 333/Kg of stored hydrogen capacity) [19,20]. A series of options have been investigated for hydrogen storage applications: cryo-compression (7-27 Kg/m 3   A special role is held by complex metal hydrides and borohydrides in particular, given their high gravimetric energy content ( Figure 2).
Computing the theoretical hydrogen storage of a complex metal borohydride yields the following: This relation would imply that, in order to reach a minimum storage capacity set by the DOE of 4.5 wt.%, the following relation should hold: 400 x + 14.842 ≥ 4.5 ⇒ x ≤ 74.406 (2) One could argue that the DOE target could be met, looking at Equations (1) and (2), by any mono-valent metal having AM ≤ 74. 406 , or a divalent metal with AM ≤ 148.812 , or a trivalent metal with AM ≤ 223.218 a.s.o. Indeed, mTHRost known borohydrides fall into this category: they afford a theoretical hydrogen storage capacity superior to that imposed by the DOE (4.5 wt.%). Equation (2) also implies that the hydrogen storage capacity will be the highest for the most lightweight borohydrides, thus cor- A special role is held by complex metal hydrides and borohydrides in particular, given their high gravimetric energy content ( Figure 2).
Computing the theoretical hydrogen storage of a complex metal borohydride yields the following: wt.% hydrogen in M(BH 4 This relation would imply that, in order to reach a minimum storage capacity set by the DOE of 4.5 wt.%, the following relation should hold: One could argue that the DOE target could be met, looking at Equations (1) and (2), by any mono-valent metal having A M ≤ 74.406 g mol , or a divalent metal with A M ≤ 148.812 g mol , or a trivalent metal with A M ≤ 223.218 g mol a.s.o. Indeed, mTHRost known borohydrides fall into this category: they afford a theoretical hydrogen storage capacity superior to that imposed by the DOE (4.5 wt.%). Equation (2) also implies that the hydrogen storage capacity will be the highest for the most lightweight borohydrides, thus corresponding to low A M values: second, third, and fourth period metals are the borohydrides most used to date [21][22][23][24][25][26][27]. A summary of known borohydrides with specific hydrogen storage capacity is presented in Figure 2, which are arranged by metal valency and atomic weight of the metal, in decreasing order of their theoretical hydrogen storage capacity (wt.%). The maximum hydrogen storage capacity (denoted herein as: H, wt.%) is computed for a series of mono-, di-, tri-, and tetravalent metals; in some cases, only hypothetic formulae were used since the actual borohydride's existence is unclear (AuBH 4 , Ni(BH 4 ) 2 , Co(BH 4 ) 2 , Hg(BH 4 ) 2 , Sc(BH 4 ) 3 , Nb(BH 4 ) 3 , U(BH 4 ) 3 , Ge(BH 4 ) 4 , Sn(BH 4 ) 4 ), while for others that exist as adducts with various stabilizing solvents/ligands (like Cr(BH 4 ) 2 ), the capacity is reported with respect to the theoretical, anhydrous form [28,29]. responding to low AM values: second, third, and fourth period metals are the borohydrides most used to date [21][22][23][24][25][26][27]. A summary of known borohydrides with specific hydrogen storage capacity is presented in Figure 2, which are arranged by metal valency and atomic weight of the metal, in decreasing order of their theoretical hydrogen storage capacity (wt.%). The maximum hydrogen storage capacity (denoted herein as: H, wt.%) is computed for a series of mono-, di-, tri-, and tetravalent metals; in some cases, only hypothetic formulae were used since the actual borohydride's existence is unclear (AuBH4, Ni(BH4)2, Co(BH4)2, Hg(BH4)2, Sc(BH4)3, Nb(BH4)3, U(BH4)3, Ge(BH4)4, Sn(BH4)4), while for others that exist as adducts with various stabilizing solvents/ligands (like Cr(BH4)2), the capacity is reported with respect to the theoretical, anhydrous form [28,29]. Many of the formulated borohydrides from Figure 2 comply with the DOE's theoretical minimum hydrogen storage capacity (4.5 wt.%), yet very few of them can actually be used in storage materials. However, there are quite a few aspects plaguing metal borohydrides: high dehydrogenation enthalpies, high dehydrogenation temperature onset, slow dehydrogenation/rehydrogenation kinetics, and side-reactions occurring during dehydrogenation leading to boron-loss and thus incomplete/impossible rehydrogenation, but also very high temperature and H2 pressures required to rehydrogenate them. Many of the formulated borohydrides from Figure 2 comply with the DOE's theoretical minimum hydrogen storage capacity (4.5 wt.%), yet very few of them can actually be used in storage materials. However, there are quite a few aspects plaguing metal borohydrides: high dehydrogenation enthalpies, high dehydrogenation temperature onset, slow dehydrogenation/rehydrogenation kinetics, and side-reactions occurring during dehydrogenation leading to boron-loss and thus incomplete/impossible rehydrogenation, but also very high temperature and H 2 pressures required to rehydrogenate them.
From a structural point of view, borohydrides M(BH 4 ) x are ionic/covalent compounds formed by a metal center (M x+ ) that binds accordingly to the negatively charged borohy-dride tetrahedra BH 4 − . The lightest member of the borohydride family is LiBH 4 , depicted in Figure 3 [30][31][32][33][34][35][36][37][38]. Established by synchrotron XRD at RT, LiBH 4 shows at normal conditions an orthorhombic symmetry (Pnma), with the BH 4  From a structural point of view, borohydrides M(BH4)x are ionic/covalent compounds formed by a metal center (M x+ ) that binds accordingly to the negatively charged borohydride tetrahedra BH4 -. The lightest member of the borohydride family is LiBH4, depicted in Figure 3 [30][31][32][33][34][35][36][37][38]. Established by synchrotron XRD at RT, LiBH4 shows at normal conditions an orthorhombic symmetry (Pnma), with the BH4 tetrahedra aligned along two orthogonal directions. The [BH4] − units are surrounded by four Li + cations, and each Li + by four [BH4] − . More important for hydrogenation studies, the B-H bond is of a covalent nature (rB-H = 119 pm), thus considerable energy is required to break that bond (ΕB-H = 389 kJ/mol). It follows that the dehydrogenation enthalpy ΔHdehydrogenation has a large value, an aspect that must be overcome using thermodynamical considerations. First (IA) and second group (IIA) metal borohydrides are not only known, but also commercially available. Various reviews have tackled different areas regarding complex borohydrides and hydrogen, focusing either on current technologies [1,2,6] and conceptual design of storage materials [3][4][5]15], or on various forms of hydrogen carriers [7,8] utilized to eventually switch the aging and depleting conventional fuel to a "green" fuel [9]. Other reviews evaluate the role of nanostructured metal hydride species for hydrogen storage [10,17], and usually focus on the potential presented by light metal borohydrides [10,16,18]. In the latter, sodium borohydride is actually presented as the fuel for the future [18] and judging by the emerging reports dating from 2021-2022 one could argue that this is true. Synthetic strategies and structure characterization [21] have been complemented by in-depth studies on hydrolysis process of light borohydrides [22,24,25,27] or by investigation on the possible intermediates that dehydrogenation could entail [12]. Destabilization strategies employed to tame the rigid behavior of tetrahydroborates during hydrogenation cycles have been reviewed [30][31][32][33][34], and the field of ionic conductivity has been explored with aplomb especially after the discovery of LT(low temperature)-HT(high temperature) phase transition in lithium borohydride [36][37][38]. The current review aims to gather most areas of interest regarding the fascinating chemistry of complex borohydrides and bring them under one umbrella, considering the historical evolution of the field, important milestones, and current and emerging trends.

Hydrogen Storage Options: Physical vs. Chemical Storage
Hydrogen storage methods can basically be divided in two groups: physical (liquid H2, cryo-compressed and compressed gas, and physically adsorbed as in metal-organic frameworks of MOF-5 type), and chemical storage (organic liquid: BN-methyl cyclopentane; interstitial hydrides: LaNi5H6; complex hydrides: NaAlH4; or in ammonia-borane adducts: NH3.BH3) (Figure 4). More important for hydrogenation studies, the B-H bond is of a covalent nature (r B-H = 119 pm), thus considerable energy is required to break that bond (E B-H = 389 kJ/mol). It follows that the dehydrogenation enthalpy ∆H dehydrogenation has a large value, an aspect that must be overcome using thermodynamical considerations. First (IA) and second group (IIA) metal borohydrides are not only known, but also commercially available. Various reviews have tackled different areas regarding complex borohydrides and hydrogen, focusing either on current technologies [1,2,6] and conceptual design of storage materials [3][4][5]15], or on various forms of hydrogen carriers [7,8] utilized to eventually switch the aging and depleting conventional fuel to a "green" fuel [9]. Other reviews evaluate the role of nanostructured metal hydride species for hydrogen storage [10,17], and usually focus on the potential presented by light metal borohydrides [10,16,18]. In the latter, sodium borohydride is actually presented as the fuel for the future [18] and judging by the emerging reports dating from 2021-2022 one could argue that this is true. Synthetic strategies and structure characterization [21] have been complemented by in-depth studies on hydrolysis process of light borohydrides [22,24,25,27] or by investigation on the possible intermediates that dehydrogenation could entail [12]. Destabilization strategies employed to tame the rigid behavior of tetrahydroborates during hydrogenation cycles have been reviewed [30][31][32][33][34], and the field of ionic conductivity has been explored with aplomb especially after the discovery of LT(low temperature)-HT(high temperature) phase transition in lithium borohydride [36][37][38]. The current review aims to gather most areas of interest regarding the fascinating chemistry of complex borohydrides and bring them under one umbrella, considering the historical evolution of the field, important milestones, and current and emerging trends.

Physical Storage of Hydrogen
Hydrogen has an atomic weight of 1.0079 g/mol, and is thus the lightest elemen known. Under normal conditions (standard temperature and pressure), hydrogen is a odorless, nontoxic but combustible gas comprising two hydrogen atoms covalentl bonded by a single σ-bond: H2. Hydrogen ( 1 H, most abundant isotope called protium) ha no neutrons, one proton in the nucleus and one electron in the outer shell, and a covalen radius of only 0.315 Å. The common and perhaps the most important physical propertie of hydrogen related to its storage applications are the melting point (m.p. = −259.14 °C and boiling point (b.p. = −252.87 °C), which, coupled with its low density (0.0898 g/L) an high diffusivity, make hydrogen storage a difficult task from an economic and technolog ical point of view.
Hydrogen has three OS (oxidation states): −1 (hydrides), 0 (molecular dihydrogen and +1 (typical OS for hydrogen, as found in most compounds, as for instance in hydro carbons or the ubiquitous H2O); it is therefore a versatile element, as it can act as both a oxidizing and a reducing reagent. Its single valence electron makes hydrogen a very rea tive element; it is therefore usually found in nature in a variety of inorganic and organ compounds.
Being a fuel and highly flammable (low ignition energy, low explosion limit in air 4% vol.), hydrogen is regulated through a series of EU safety datasheets: S9 (containe must be kept in well-ventilated areas), S16 (no smoking allowed in its proximity, and ig nition sources are forbidden nearby) and S33 (danger due to electrostatic discharge).
Combustion of hydrogen produces the highest gravimetric energy known (33.3-39 Wh/g), depending on the state in which water is produced-gaseous water yields th lower value (33.6 Wh/g), while combustion to liquid water produces 39.4 Wh/g: Hydrogen can be stored either as a liquid (cryogenic temperatures, due to low b.p or as a gas (at high pressures, in the range 350-700 atm). Physically sorbed hydrogen also known: adsorption at surface of solids (MOF-5) or absorption (within solids).

Chemical Storage of Hydrogen
Facing the obvious difficulties of cryogenic temperatures and/or very high pressure

Physical Storage of Hydrogen
Hydrogen has an atomic weight of 1.0079 g/mol, and is thus the lightest element known. Under normal conditions (standard temperature and pressure), hydrogen is an odorless, nontoxic but combustible gas comprising two hydrogen atoms covalently bonded by a single σ-bond: H 2 . Hydrogen ( 1 H, most abundant isotope called protium) has no neutrons, one proton in the nucleus and one electron in the outer shell, and a covalent radius of only 0.315 Å. The common and perhaps the most important physical properties of hydrogen related to its storage applications are the melting point (m.p. = −259.14 • C) and boiling point (b.p. = −252.87 • C), which, coupled with its low density (0.0898 g/L) and high diffusivity, make hydrogen storage a difficult task from an economic and technological point of view.
Hydrogen has three OS (oxidation states): −1 (hydrides), 0 (molecular dihydrogen), and +1 (typical OS for hydrogen, as found in most compounds, as for instance in hydrocarbons or the ubiquitous H 2 O); it is therefore a versatile element, as it can act as both an oxidizing and a reducing reagent. Its single valence electron makes hydrogen a very reactive element; it is therefore usually found in nature in a variety of inorganic and organic compounds.
Being a fuel and highly flammable (low ignition energy, low explosion limit in air at 4% vol.), hydrogen is regulated through a series of EU safety datasheets: S9 (containers must be kept in well-ventilated areas), S16 (no smoking allowed in its proximity, and ignition sources are forbidden nearby) and S33 (danger due to electrostatic discharge).
Combustion of hydrogen produces the highest gravimetric energy known (33.3-39.4 Wh/g), depending on the state in which water is produced-gaseous water yields the lower value (33.6 Wh/g), while combustion to liquid water produces 39.4 Wh/g: Hydrogen can be stored either as a liquid (cryogenic temperatures, due to low b.p.) or as a gas (at high pressures, in the range 350-700 atm). Physically sorbed hydrogen is also known: adsorption at surface of solids (MOF-5) or absorption (within solids).

Chemical Storage of Hydrogen
Facing the obvious difficulties of cryogenic temperatures and/or very high pressures, the physical storage of hydrogen has inherent safety issues that have swayed attention of researchers towards chemical storage of hydrogen. Solid-state hydrogen storage can be achieved in metal hydrides, complex metal hydrides, ammonia-borane adducts, or organic liquids (BN-methylcyclopentane) ( Figure 4).
Among these, complex hydrides have gained more popularity due to their salt-like nature, high hydrogen storage wt.%, and involvement in a variety of advantageous chemical systems (M(BH 4 ) x -MH y , M(BH 4 ) x -M(NH 2 ) y , etc). The general formula of a complex hydride can be written as A x D y H z , with A being usually group IA and IIA elements, and D either boron or aluminum. Complex hydrides contain hydrogen covalently bonded to D, forming the complex anion [D y H z ] r− . These compounds have been known for many years, but they were revisited by Bogdanovic and Schwickardy, who discovered that addition of 2 mol % β-TiCl 3 or Ti(OBu) 4 catalyst to NaAlH 4 would bring down the kinetic barrier of dehydrogenation to technologically feasible levels for hydrogenation studies. The novel material had a tentative formula NaAlH 3.9 Ti 0.02 Cl 0.06 and showed a thermogravimetric curve featuring a reduction in the dehydrogenation temperature of 80-85 • C with respect to undoped NaAlH 4 , while proving the reversible character of the prepared sample up to the 100th cycle. However, the dehydrogenation-rehydrogenation rate was still rather low, requiring high temperatures (>150 • C) and pressures (60-150-bar H 2 ) (5) [39].
This pioneering work opened the door to further investigation of complex hydrides as new motifs for hydrogen storage materials; understanding the role of the catalyst and the exact nature of dehydrogenation and rehydrogenation reactions was the main aim. Other complex hydrides such as Al(BH 4 ) 3 or Mg 2 FeH 6 have almost double volumetric hydrogen density compared to liquid hydrogen (150 kg/m 3 vs. 70 kg/m 3 ), which stimulated research in this intriguing field of hydrogen storage systems.

General Synthesis Strategies for Metal Borohydrides M(BH 4 ) x
There are many known complex metal borohydrides ( Figure 2 summarizes most of them). The first known was aluminum borohydride Al(BH 4 ) 3 , synthesized in 1939 by Schlesinger et al. by mistake, followed by discovery of NaBH 4 in 1943 [40]. Out of all known borohydrides, Be(BH 4 ) 2 remains the one providing the highest gravimetric hydrogen content (20.7 wt.%, Figure 2), but it is also very toxic, which limits its usability as hydrogen storage material. While using them as raw materials for hydrogen storage is impractical due to poor reversibility and very high (or low) decomposition temperature, the high amount of hydrogen bound to boron in complex anion [BH 4 ] − or [B 2 H 7 ] − (the latter is identified only in solution) makes them very interesting starting points for novel energy storage systems.
The general strategies used for metal borohydride synthesis are: solid-state reactions (mechano-chemical), wet chemistry (solvent-assisted), nanoconfined hydrides, and formation of adducts of borohydrides. The mechano-chemical synthesis route is perhaps the most-used and best-known method to obtain metal borohydrides; it is also the only one able to afford mixed-cation borohydrides (like LiZn 2 (BH 4 ) 5 ) [41]. The main drawback of ball-milling / the mechano-chemical route is the impurity of the resulting borohydride with metal halides, which are by-products of the metathesis reaction [42][43][44][45][46]. By contrast, wet-chemistry methods can yield phase-pure metal borohydrides, and they afford far easier separation procedures.

Solid-State-Mechanochemical Synthesis
Ball-milling has a series of advantages: it allows for good homogenization of the sample, reduces the size of the particle, and introduces defects which are highly reactive, inclusive in hydrogenation studies [47]. The mechano-chemical route is a highly energetic process carried out usually in a planetary ball mill [47][48][49]. The milling program typically consists in 1-5 min ball-milling followed by a 1-5 min break [50], such that the chemical activation observed is not due to local temperature increase, which is estimated to remain around 60 • C [51][52][53], but rather due to high pressure caused by collisions between ball, powder, and vial which produce material shear stress and deformations [54]. The mechanochemical approach can increase the reactivity of materials by size-reduction of the particles, which increases their surface area and thereby provides better interaction between particles needed for a complete reaction to occur.
The seemingly first candidate for hydrogen storage (highest gravimetric hydrogen content, 20.7 wt.%), Be(BH 4 ) 2 , was first synthesized by a metathesis reaction, at 145 • C [60]: Schlesinger et al. made a great contribution to the development of novel metal borohydrides, after the discovery of Al(BH 4 ) 3 in 1939, and studied the synthesis of borohydrides by metathetical reactions using alkali metal borohydrides [60], alternative syntheses of NaBH 4 and LiBH 4 [61], synthesis of NaBH 4 from NaH and borate esters [62], and synthesis of U(BH 4 ) 4 [63,64]. The metathetical reaction leading to Be(BH 4 ) 2 is not a proper mechanochemical reaction since it is a solid-state reaction but performed by vigorous shaking of the reagents rather than in a planetary mill [60]. In fact, due to the requirement to continuously remove the volatile (and toxic) Be(BH 4 ) 2 from the system, the reaction cannot be performed in a closed system (like the vial of a planetary mill).
It has also been suggested that, in the case of mechano-chemical synthesis, the need to evacuate the volatile borohydride species (Be(BH 4 ) 2 or Al(BH 4 ) 3 being prime examples) advocates for an equilibrium reaction that is shifted towards products with removal of the volatile product, in line with Le Chatelier's principle [81,82]. Al(BH 4 ) 3 has a theoretical hydrogen storage capacity of 16.8 wt.%, but is a liquid at RT (mp = −64 • C) due to weak intermolecular forces between its molecules; it is fairly unstable and begins to decompose at~40 • C, which is too low for use in a hydrogen production mobile tank. Its crystal structure presents [BH 4 − ] units that are covalently bonded by two H-bridges to the Al 3+ center. A ball-milling reaction of AlCl 3 with NaBH 4 in a stoichiometric molar ratio results in good yields of Al(BH 4 ) 3 [60].
The solid-state Al(BH 4 ) 3 has two polymorphs (transition temperature −93 to −78 • C) [83][84][85][86][87][88]. Due to its liquid nature, toxicity, and moisture and air sensitivity, Al(BH 4 ) 3 rather serves as a starting point for synthesis of other, more stable borohydrides, and it is not currently the focus of research on novel energy storage materials because of these limitations.
Ball-milling is also the only way to produce mixed-cation or mixed-anion borohydrides, with reaction conditions (milling time, temperature, pressure) depending ultimately on the metal electronegativity and d-electron configuration in the case of transition metal systems. For instance, mixing ZnCl 2 and KBH 4 in a 1:1 ratio produces a chloro-borohydride of K + and Zn 2+ . The complexity of the ball-milling process and its high dependence on reagent molar ratio is depicted below, showing a family of products all starting from ZnCl 2 and a metal alkali borohydride source ( Figure 5) [41]. A ball-milling reaction of AlCl3 with NaBH4 in a stoichiometric molar ratio results in good yields of Al(BH4)3 [60].
The solid-state Al(BH4)3 has two polymorphs (transition temperature −93 to −78 °C) [83][84][85][86][87][88]. Due to its liquid nature, toxicity, and moisture and air sensitivity, Al(BH4)3 rather serves as a starting point for synthesis of other, more stable borohydrides, and it is not currently the focus of research on novel energy storage materials because of these limitations.
Ball-milling is also the only way to produce mixed-cation or mixed-anion borohydrides, with reaction conditions (milling time, temperature, pressure) depending ultimately on the metal electronegativity and d-electron configuration in the case of transition metal systems. For instance, mixing ZnCl2 and KBH4 in a 1:1 ratio produces a chloro-borohydride of K + and Zn 2+ . The complexity of the ball-milling process and its high dependence on reagent molar ratio is depicted below, showing a family of products all starting from ZnCl2 and a metal alkali borohydride source ( Figure 5) [41]. Besides the formation of LiCl (or NaCl) as by-products, the metathetic route has another drawback: some amounts of this salt can dissolve in LiBH4, yielding a solid solution of approximate formula Li(BH4)1-pClp (p is an integer number, 0 < p < 1), making further purification even more difficult [89].
Synthesis of borohydrides can also be accomplished from starting elements as raw materials, but under extreme conditions caused by the very low reactivity of boron, as shown by Friedrichs et al. (700 °C, 150 bar) [90,91]: Other high-pressure reaction routes starting from solid metal hydride or metal boride have been investigated. For instance, reacting metal hydride MHx with MgB2 and hydrogen H2 leads to corresponding M(BH4)x and MgH2 [92,93]. Besides the formation of LiCl (or NaCl) as by-products, the metathetic route has another drawback: some amounts of this salt can dissolve in LiBH 4 , yielding a solid solution of approximate formula Li(BH 4 ) 1-p Cl p (p is an integer number, 0 < p < 1), making further purification even more difficult [89].
Synthesis of borohydrides can also be accomplished from starting elements as raw materials, but under extreme conditions caused by the very low reactivity of boron, as shown by Friedrichs et al. (700 • C, 150 bar) [90,91]:  [92,93].
f or x = 1, M = Li, Na; f or x = 2, M = Ca Regeneration of Ca(BH 4 ) 2 during hydrogenation studies was shown to proceed at 350 • C and 150-bar H 2 , with a faster dehydrogenation/rehydrogenation kinetics when a Ti-catalyst was employed [94].
Reaction between sodium hydride NaH and boric oxide have produced sodium borohydride in good yield (60%) after ball-milling at 330-350 • C for 20-48 h [62]: In a similar fashion, but using methyl borate B(OCH 3 ) 3 instead of boric oxide, LiBH 4 was obtained with an overall yield of about 70% [62]: However, the economical aspect of using mechano-chemical synthesis may tip the balance, because large-scale facilities for grinding are known and operable, so that cost-effectiveness is insured for the preparation of light complex hydrides via the ball-milling technique.

Wet Chemistry-Solvent-Assisted Synthesis
The first borohydride, Al(BH 4 ) 3 , was synthesized by Schlesinger by accident in 1939 [40] (Equation (16)). He was hoping to obtain AlH 3 by mixing Al(CH 3 ) 3 and B 2 H 6 , when the following reaction actually occurred, which has further been expanded to the synthesis of Be(BH 4 ) 2 (Equation (17)) and LiBH 4 (using an alkyl-lithium in the latter case). LiBH 4 is the second most used borohydride today, but it was first synthesized also in the 1940s by Schlesinger et al. according to Equation (18) [102]. 3 At present, the usual procedure for the synthesis of M(BH 4 ) x follows the general reactivity scheme (19) and (20), and it can be successfully applied for divalent borohydride synthesis (M = Be, Mg, Ca, Sr, Ba) or trivalent borohydrides (M = Al): Al(BH 4 ) 3 was obtained by Kollonitsch and Fuchs and reported back in 1955, by reacting AlCl 3 with Ca(BH 4 ) 2 in a 2:3 molar ratio [103].
Synthesis of LiBH 4 according to Equation (19) takes place in NaH-dried isopropylamine i PrNH 2 [60,104], and usage of toxic diborane B 2 H 6 is now avoided as in its first synthesis by Schlesinger and Brown [102]. The metathesis reactions (19) and (20) are welldocumented, and proceed with a reasonable rate for LiBH 4 compared to slower alkali borohydride starting materials (i.e., NaBH 4 , and ever slower with KBH 4 ) [60], and afford the complex borohydride M(BH 4 ) x in near-quantitative yield. NaBH 4 remains the most wide-spread borohydride in terms of synthesis methods and uses, including but not limited to organic synthesis, reduction reactions, and metathesis [105].
Mg(BH 4 ) 2 could be synthesized according to Equation (20), using diethyl ether Et 2 O as solvent for the LiBH 4 , but the product obtained contains impurities of Li and Cl − [106][107][108]. At modest temperatures (34-60 • C), reaction (22) yields some Mg(BH 4 ) 2 (overall yield: 0-11.5%); different conditions must therefore be used if NaBH 4 is used as borohydride source [107,108]. Even when conducted in recommended solvent i PrNH 2 , the final mixture could not surpass 40 wt.% of Mg(BH 4 ) 2 , with the impurities being mainly represented by unreacted NaBH 4 ; this aspect could be proof of an equilibrium reaction being reached under those specific reaction conditions, but it could also originate from poor solubility of the intermediate magnesium amine complex formed in the reaction mixture [106]. Phasepure β-Mg(BH 4 ) 2 was obtained by replacing highly reactive LiBH 4 with NaBH 4 , using a combined mechano-chemical (2 h) double exchange reaction (60 h, refluxing Et 2 O solvent), followed by drying at 240 • C [109].
Zn(BH 4 ) 2 could also be prepared by wet-chemistry at room temperature (RT), using tetrahydrofuran (THF) as solvent and ZnCl 2 and NaBH 4 as starting materials ( Figure 5) [119][120][121][122]. With a hydrogen storage capacity of 8.5 wt.%, Zn(BH 4 ) 2 starts to decompose at~85 • C and eliminates up to 140 • C hydrogen and diborane B 2 H 6 , making it for the time being too unstable for vehicular applications. On the other hand, formation of toxic B 2 H 6(g) should urge researchers to look for alternative decomposition pathways that avoid diborane as by-product, since current dehydrogenation will not be reversible in the absence of the boron source [119][120][121][122].
Potassium borohydride KBH 4 (also RbBH 4 and CsBH 4 ) could be synthesized by a precipitation reaction of NaBH 4 with a concentrated solution of a strong base, KOH; the reaction could be carried out in aqueous or CH 3 OH media, with an overall yield of 75.5% (Equation (24)). The reagents are independently dissolved in MeOH, cooled, and then mixed, when the borohydride of interest precipitates and is obtained in pure form after a subsequent vacuum drying step [123,124].
A similar reaction between the metal hydride MH x (M = Li, Mg, Ca) and B 2 H 6 in a solid-gas mechanochemical process was proposed by Friedrichs et al., who obtained borohydride products in good-to-high yield, but containing some impurities (unreacted starting metal hydride and Fe-Ni phases from vial abrasion during milling): LiBH 4 (η = 94%), Mg(BH 4 ) 2 (η = 91%) and Ca(BH 4 ) 2 (η = 73%). Diborane B 2 H 6 was also proposed as a key intermediate in hydrogenation studies under moderate conditions [93].

Nanoconfined Hydrides
The ball-milling technique is usually the go-to process for obtaining nanosized powders of metal hydrides, to sizes often less than 100 nm [127]. Ball-milling implies milling in a vial using balls of special alloys; however, the milling process could induce impurities in the sample from the materials of the vial and balls. However, these nanoparticles could also sinter into larger particles during hydrogenation-dehydrogenation cycles, which negatively impacts gravimetric hydrogen uptake and release [128][129][130][131]. Using specially engineered nano-scaffolds, complex hydrides can be obtained in the nanometer range, far below the average sizes afforded by the mechano-chemical approach.
Among nanoconfinement methods, the most widely used are the direct synthesis of nanoconfined complex metal (boro)hydride, and infiltration methods (melt-infiltration and wet-infiltration) [132].
Nanoconfinement of metal hydrides synthesized in situ has been extensively studied for magnesium hydrides in carbon aerogels produced by the resorcinol-formaldehyde reaction (RF-CA), using a 1 M solution of Mg( n C 4 H 9 ) 2 in n-heptane [133,134]. Nielsen et al. synthesized by this method MgH 2 @RF-CA of pore sizes 7 and 22 nm [133]. This is a typical case of soft-nanocasting, as the initial porous support is not removed from the final composite material ( Figure 6).
The excess n-butyl magnesium can be removed manually after the reaction from the surface of RF-CA (Equation (27)).
Melt infiltration requires a very inert scaffold, with a melting point higher than that of the respective metal hydride used. The driving force of this method is the lower interfacial energy metal hydride/scaffold, as reviewed by de Gennes [135,136]. This method has the advantage of completely removing the necessity of a solvent, therefore avoiding postsynthesis steps of the hydride@scaffold composite. In addition, the melting point of the hydride must not coincide with the onset of decomposition of the said material. However, sodium tetrahydroaluminate NaAlH 4 (mp = 183 • C) has been utilized for melt infiltration, despite decomposition onset into Al and Na 3 AlH 6 upon melting [137]. The excess n-butyl magnesium can be removed manually after the reaction from the surface of RF-CA (Equation (27)).
Melt infiltration requires a very inert scaffold, with a melting point higher than that of the respective metal hydride used. The driving force of this method is the lower interfacial energy metal hydride/scaffold, as reviewed by de Gennes [135,136]. This method has the advantage of completely removing the necessity of a solvent, therefore avoiding post-synthesis steps of the hydride@scaffold composite. In addition, the melting point of the hydride must not coincide with the onset of decomposition of the said material. However, sodium tetrahydroaluminate NaAlH4 (mp = 183 °C) has been utilized for melt infiltration, despite decomposition onset into Al and Na3AlH6 upon melting [137].
Solvent infiltration requires a suitable solvent that affords complete metal hydride dissolution in order to be successful. A series of physical properties of metal borohydrides (including knowledge of select solubility data) are a prerequisite, as limited solubility in commonly used solvents is the main limitation of this method. Sometimes even weakly coordinating solvents could be used, but these solvents should be removed under moderate conditions (T, p). Following the rather poor solubility of complex metal hydrides, consequent steps of impregnation-evaporation may be required to reach full (or the desired degree of) scaffold pore filling [132]. This technique-incipient wetness method-can be advantageous as it avoids crystallization of active hydride species outside of mesopores; The decomposition of NaAlH 4 must be considered a partial equilibrium (Equations (28) and (29)), because when working around the borohydride melting point and using high H 2 pressure (160-190 bar), the decomposition is avoided and melt infiltration has been successful [137][138][139][140][141][142].
Solvent infiltration requires a suitable solvent that affords complete metal hydride dissolution in order to be successful. A series of physical properties of metal borohydrides (including knowledge of select solubility data) are a prerequisite, as limited solubility in commonly used solvents is the main limitation of this method. Sometimes even weakly coordinating solvents could be used, but these solvents should be removed under moderate conditions (T, p). Following the rather poor solubility of complex metal hydrides, consequent steps of impregnation-evaporation may be required to reach full (or the desired degree of) scaffold pore filling [132]. This technique-incipient wetness method-can be advantageous as it avoids crystallization of active hydride species outside of mesopores; complex metal borohydrides like Ca(BH 4 ) 2 have been infiltrated in chemically activated micro-mesoporous carbons of very high surface area to produce composites exhibiting reversible hydrogen storage under more modest conditions [143].
Various scaffolds can be used; inert, mesoporous silica of 2D-ordered type (SBA-15) could serve such a purpose. Computational studies have shown that molecules of NH x BH x type (x = 1-4) display thermodynamically accessible dehydrogenation steps of less than 40 kJ/mol, and relatively low dehydrogenation temperature onset (100-150 • C); they can thus be considered candidates for hydrogen storage reactions [144]. When a 5.4 M solution of ammonia-borane NH 3 .BH 3 in CH 3 OH infiltrated the~7.5 nm pores of SBA-15 silica, a total loading of roughly 50 wt.% NH 3 BH 3 @SBA-15 was obtained after solvent evaporation [145]. A clear advantage of solvent-mediated impregnation is that the experimental procedure requires less energy, since it can be carried out even at room temperature.

Derivatives-Formation of Adducts of M(BH 4 ) x
Solvent metathesis reactions often lead to the formation of desired borohydrides as solvates: M(BH 4 ) x ·p Et 2 O (p is an integer number), as is the actual case for M = Mg and Y. Both Mg(BH 4 ) 2 ·p Et 2 O [106-108] and Y(BH 4 ) 3 ·p Et 2 O [76] could be obtained as pure, unsolvated solids after heating at 150 • C, as shown by IR and TG data (Equation (30)).
The structural variety of polymorphs obtained in the case of magnesium borohydride (six in total) (Equations (35)-(37)) suggests that novel, nanoporous metal hydrides could be potentially accessible via this desolvation route.
Thermal analysis coupled with IR spectroscopy data has pointed out two possible formulations of solvated complexes of closely related tetrahydroaluminates, when coordinated to a tertiary amine "L" (Figure 7). If a weak coordination is expected, then structural formula I is favorable, and it should resemble that of an uncoordinated complex metal hydride, as pointed out by Schlesinger and Brown. The complex should exhibit high thermal stability-which is actually the case for I: L =⇒ M + [AlH 4 ] − , due to the increased size of the cation and the reduced distortion of the AlH 4 − tetrahedron by alkali metal interaction [102]. By contrast, formula II: H-M. . . AlH 3 L is expected to have a lower thermal stability, resulting in a complex mixture of products upon decomposition due to the complex bond amine-alanate R 3 N: =⇒AlH 3 [159].
The structural variety of polymorphs obtained in the case of magnesium borohydride (six in total) (Equations (35)-(37)) suggests that novel, nanoporous metal hydrides could be potentially accessible via this desolvation route.
Thermal analysis coupled with IR spectroscopy data has pointed out two possible formulations of solvated complexes of closely related tetrahydroaluminates, when coordinated to a tertiary amine "L" (Figure 7). If a weak coordination is expected, then structural formula I is favorable, and it should resemble that of an uncoordinated complex metal hydride, as pointed out by Schlesinger and Brown. The complex should exhibit high thermal stability-which is actually the case for I: L → M + [AlH4] -, due to the increased size of the cation and the reduced distortion of the AlH4 -tetrahedron by alkali metal interaction [102]. By contrast, formula II: H-M…AlH3L is expected to have a lower thermal stability, resulting in a complex mixture of products upon decomposition due to the complex bond amine-alanate R3N: →AlH3 [159]. Full decomposition of solvate LiBH4·L could proceed through either pathway (38) or (39) as evidenced by the ambiguous DTA-TGA curve, with both leading to loss of coordinating molecule L: Mg(BH4)2·THF adducts decompose violently [160] due to the formation of alkoxy species by the THF ring-opening reaction, as is the case with aluminum hydride-THF adducts [161]. However, Ca(BH4)2·THF can be desolvated with relative ease, without such strong exothermic effects, as evidenced by Equation (31) [146][147][148][149][150][151]162].
A metathesis reaction is currently the preferred method of synthesis for metal borohydrides because they can be obtained in pure form. There are recent reports expanding the recognized reducing character of metal borohydrides. For instance, Ca(BH4)2 can be successfully used in advanced organic synthesis, such as polymerization of ε-caprolactine and L-lactide [163].

Structural Considerations of M(BH4)x
There is an intricate connection between structure and solid-state chemistry related to borohydride materials. Among most important physical properties affected by lattice structure are the thermal stability and the accessible decomposition pathways.

Framework and Crystal Structure
Metal borohydrides are typically ionic compounds, but covalent borohydrides are also known. Table 1 summarizes some of the novel metal borohydrides and mixed-cation borohydrides and their structural architecture; many present multiple polymorphism (LT, HT, metastable). Full decomposition of solvate LiBH 4 ·L could proceed through either pathway (38) or (39) as evidenced by the ambiguous DTA-TGA curve, with both leading to loss of coordinating molecule L: 3 Mg(BH 4 ) 2 ·THF adducts decompose violently [160] due to the formation of alkoxy species by the THF ring-opening reaction, as is the case with aluminum hydride-THF adducts [161]. However, Ca(BH 4 ) 2 ·THF can be desolvated with relative ease, without such strong exothermic effects, as evidenced by Equation (31) [146][147][148][149][150][151]162].
A metathesis reaction is currently the preferred method of synthesis for metal borohydrides because they can be obtained in pure form. There are recent reports expanding the recognized reducing character of metal borohydrides. For instance, Ca(BH 4 ) 2 can be successfully used in advanced organic synthesis, such as polymerization of ε-caprolactine and L-lactide [163].

Structural Considerations of M(BH 4 ) x
There is an intricate connection between structure and solid-state chemistry related to borohydride materials. Among most important physical properties affected by lattice structure are the thermal stability and the accessible decomposition pathways.
The o-LiBH 4 to h-LiBH 4 polymorphic transformation taking place at~110 • C can be seen as a contraction along orthorhombic axis a (hexagonal axis c) and an expansion of the [b*c] orthorhombic plane, and it is accompanied by a slight endothermic enthalpy of transition (4.18 J/mol). KBH 4 (3 in Figure 8) in its β-polymorph, crystallizes in the tetragonal P4 2 /nmc space group. The 3D structure shows a 12-coordinate geometry around the alkali metal, in an equivalent H-surrounding. The K-H bond lengths average 2.815 Å, while the B-H bonds in the [BH 4 ] − unit measure 1.23 Å. The structure of β-KBH 4 bears structural similarities to tetragonal NaBH 4 , but it is refined to a higher symmetry group P4 2 /nmc. As the size of the cation increases in the alkali group (from Li to Rb), the cubic MBH 4 (M = alkali metal) unit cell expands and the [BH 4 ]. . . [BH 4 ] repulsions are reduced, such that the cubic-to-tetragonal transition temperature is reduced in heavier alkali metal borohydrides.
RbBH 4 (4 in Figure 8) crystallizes in the cubic Fm3m space group (a = 6.9633 Å at 0.5 GPa). The Rb + cations and BH 4 − anions have a NaCl structural prototype and this roomtemperature polymorph is isostructural to NaBH 4 , with the BH 4 groups orientationally disordered over the two positions around the inversion center, similar to the sodium tetrahydroborate case.
Be(BH 4 ) 2 (5 in Figure 8) presents only one structurally characterized form which crystallizes in the tetragonal I4 1 /cd space group. It contains helical polymeric chains with the independent Be 2+ coordinated in a trigonal-planar geometry to two bridging [ α-Ca(BH 4 ) 2 (7 in Figure 8) can be obtained by desolvation of the THF adduct, and crystallizes in the orthorhombic Fddd space group, being one of the four possible polymorphs (α, α', β and γ) and bearing structural similarities to TiO 2 polymorphs [187,188]. The orthorhombic structure has been further amended to better be represented by a lower symmetry (F2dd or Fdd2). All four polymorphs show Ca 2+ octahedrally coordinated to six equivalent [BH 4 ] tetrahedra, with the Ca-B bond distances falling in the range 2.82-2.97 Å. The divalent calcium cations form in α-Ca(BH 4 ) 2 a close-packed, diamond-type structure where the tetrahydroborate groups exhibit T-shape coordination [188].
Three polymorphs of Mg(BH 4 ) 2 are depicted (9: α-, 10: β-, 11: γ-), showing increased structural complexity, but also polymeric properties [106,158,[171][172][173][189][190][191][192][193][194][195]. The α-Mg(BH 4 ) 2 represents the most stable polymorph and its structure determination was a result of combined synchrotron (most information data) and neutron powder diffraction data (orientation of BH 4 tetrahedra). Single crystal diffraction data collected at 100 K established the space group to be the hexagonal P6 1 22, containing three symmetry-independent Mg 2+ ions and six independent [BH 4 ] units. In this novel 3D arrangement, the Mg-Mg distances lie in the range 4.6-4.9 Å, while the Mg 2+ center is tetrahedrally coordinating four [BH 4 − ] tetrahedra. The Mg-B bond lengths are in the range 2.432(a)-2.434(4) Å, and there is a dispersity of B-Mg-B angles of 92.1(2)-108.80(13) • accounting for a distorted tetrahedral coordination around the metal center. The porous γ-Mg(BH 4 ) 2 has a cubic lattice with inherited porous nature (33% open space), high surface area (in excess of 1100 m 2 /g as deduced from N 2 sorption isotherms by BET method) and a partially covalent framework. The stability of Mg(BH 4 ) 2 can be related to the Mg-BH 4 -Mg linear units, with BH 4 acting as directional ligands, and stable MgH 8 polyhedra. This complex structural behavior resembles that of coordination polymers like MOFs. The γ-polymorph is particularly interesting due to large empty spaces available for gas entrapment [158].
One of the main dehydrogenation borohydride species found during hydrogen desorption studies is the [B 12 H 12 ] 2− anion, as found for instance in BaB 12 H 12 (12 in Figure 8). These dodecahydro-closo-dodecaborate anions [B 12 H 12 ] 2− have a role in the poor recyclability of metal tetrahydroborates from hydrogenation studies; their formation and possible side-reactions are therefore of clear importance. Moreover, it has been hypothesized that these species appear amorphous in XRD diffraction data, and only show short-range ordering. BaB 12  Al(BH 4 ) 3 (1 in Figure 9) is a liquid under normal conditions, but low temperature single-crystal XRD data shows two possible polymorphs with a transition temperature at~190 K [83]. Single-crystal diffraction data by Aldridge et al. from 1997 shows that Al(BH 4 ) 3 crystallizes at 195 K in the orthorhombic Pna2 1 space group with the following unit cell parameters: a = 18.021(3) Å, b = 6.138(2) Å, and c = 6.199(1) Å. It consists of discrete Al(BH 4 ) 3 units, with a trigonal-planar geometry around the Al 3+ center which coordinates the [BH 4 ] tetrahedra via the tetrahedral edges. The Al-B bond lengths are in the range of 2.10(2)-2.14(2) Å, while the longer B-H distances for the B-H. . . Al bridges of 1.12(3)-1.14(4) Å are consistent with Al-H coordination [83].
interesting due to large empty spaces available for gas entrapment [158].
One of the main dehydrogenation borohydride species found during hydrogen desorption studies is the [B12H12] 2-anion, as found for instance in BaB12H12 (12 in Figure 8). These dodecahydro-closo-dodecaborate anions [B12H12] 2-have a role in the poor recyclability of metal tetrahydroborates from hydrogenation studies; their formation and possible side-reactions are therefore of clear importance. Moreover, it has been hypothesized that these species appear amorphous in XRD diffraction data, and only show short-range ordering. BaB12H12 crystallizes in the P31c space group (trigonal symmetry) of wurtzite structural type, and shows Ba 2+ cation surrounded by four [B12H12] 2− anions. Each closoborate [B12H12] 2-anion is tetrahedrally surrounded by four Ba 2+ cations (mutual tetrahedral arrangement of cations and anions), with distances from the center of the anion to the Ba 2+ of 4.40 and 4.54 Å at 295 K. Since each of the four [B12H12] 2− anions is facing the Ba 2+ metal center with three H atoms, the cation is surrounded overall by twelve H atoms (Ba-H bond distances range from 2.77-3.02 Å). The rich boron-hydrogen speciation, open channels, and convenient molecular size (5.8 Å) have made closo-borate salts very interesting candidates for ionic conductivity studies.
This represents the first crystal structure of a 3d-metal tetrahydroborate. The α-Mn(BH 4 ) 2 polymorph has a stability range between 90 to 450 K. The structure bears structural similarities to Mg(BH 4 ) 2 ; the two independent Mn 2+ centers are each tetrahedrally surrounded by four [BH 4 ] units with Mn-B distances of 2.39-2.52 Å.
LiSc(BH 4 ) 4 can best be described as a Li + salt of complex anion [Sc(BH 4 ) 4 ] − that was evidenced by vibrational spectroscopy studies. It crystallizes in the tetragonal P42c space group, with a CuAlCl 4 structural prototype and the following unit cell parameters: a = 6.076 Å and c = 12.034 Å (5 in Figure 9). More specifically, it was obtained by Hageman et al. in 2008 by ball-milling processing of a 4:1 molar ratio LiBH 4 and ScCl 3 [181]. Moreover, the crystal structure contains disordered Li + cations along the z-axis of the tetragonal cell, which in turn could make this mixed borohydride a suitable candidate for ionic conductivity applications [181]. NaZn 2 (BH 4 ) 5 (7 in Figure 9) crystallizes in the monoclinic P2 1 /c space group, similar to the Li + counterpart LiZn 2 (BH 4 ) 5 , and it has no analogues to known inorganic structures. The unit cell parameters are: a = 9.397(2) A, b = 16.635(3)A, c = 9.136(2)A, and β = 112.66(2) • [41]. The structure consists of two independent Zn 2+ cations that have almost a trigonal-planar coordination to three [BH 4 ] tetrahedra (coordination number of Zn is 3), while the Na + cation has a new, saddle-like coordination (coordination number of Na is 4) [41]. Interestingly, NaZn 2 (BH 4 ) 5 consists of two doubly penetrating 3D frameworks, with no covalent bonds between them-a structural motif found in coordination polymers. All [BH 4 ] units are nearly linearly coordinated by two metal atoms, and the coordination occurs via the two opposite tetrahedral edges bridging two Zn 2+ cations, or a Zn 2+ and a Na + cation.
The hexagonal structure observed for many metal borohydrides can be attributed in part to an entropic term brought about by the vibrational amplitudes in BH 4 units and the shortening of B-H bonds from 1.22 to under 1.1 Å [196,197]. Mixed-cation borohydrides give rise to a fascinating family of crystal structures, and have many times exhibited promising results regarding hydrogen release and reversible hydrogen uptake.

Stability of MBH
The most ionic and stable borohydrides have been studied since the 1940s, namely NaBH 4 , KBH 4 , RbBH 4 , and CsBH 4 ; they are often stable in basic aqueous solution [165][166][167][168][169]198]. Closo-boranes are stable under neutral and acidic solutions. Being typically air-and moisture-sensitive compounds, their synthesis should be performed using Schlenk techniques or in the glove box, under inert (N 2 , Ar) atmosphere. Metal borohydrides are prone to adsorbing water, with which they react upon heating to release hydrogen. Some borohydrides of low electronegative elements are stable at RT, and even more so in the form of their hydrates. High electronegative metals forming metal borohydrides are highly reactive even below RT, and produce a highly exothermic reaction with water (some even explode upon contact with moisture). The layered structure of LiBH 4 is non-trivial and shows local distortion of the polymorphs, which defines their stability and plays a role in corresponding phase transitions [199]. Theoretical computations based on first-principles study shed new light on the stability of the family of possible borohydride intermediates from LiBH 4 , namely Li 2 B n H n (n = 5-12) [200].
Transition-metal borohydrides prepared by mechano-chemical methods are only stable when the ion itself is stable (d 0 , d 5 and d 10 ), that is, when it has no electrons in the d-orbital, at half-occupation, or when the d-shell is completely occupied [201]. However, lowering synthesis temperature to below −30 • C and using ammonia as stabilizing ligand, stable M(BH 4 ) 2 (NH 3 ) 6 have been prepared in the case of M = Cr 2+ (d 4 ), Fe 2+ (d 6 ), and Co 2+ (d 7 ). These borohydrides are only stable in solution at low temperature, and they have not been isolated in pure, solid form, with a d 1 exception to the rule coming from a molecular borohydride of Ti(BH 4 ) 3 , which has limited stability below 0 • C [202]. Some borohydrides like Zr(BH 4 ) 4 (sublimes at 29 • C) form quality single-crystals by CVD when stored in a closed vessel, due to its molecular structure [203][204][205][206][207][208][209][210].
Bi-or tri-metallic borohydrides can usually be prepared by a mechano-chemical route; however, some diethyl ether Et 2 O solvates have been prepared and are stable in solvated form, such as NaMn(BH 4 ) 3 ·0.5 Et 2 O [211]. Dimethyl sulfide S(CH 3 ) 2 has recently been used as a weak-coordinating solvent which eases borohydride removal from the products, avoiding formation of ternary chlorides/bimetallic borohydrides. This strategy allowed for the synthesis and stabilization of some rare earth metal borohydrides M(BH 4 ) 3 ·S(CH 3 ) 2 , where M is Y, La, Ce, or Gd [72,[212][213][214][215]. Formation of ternary chlorides leads to incomplete reaction in mechanochemical synthesis; for instance, if Zn(BH 4 ) 2 is targeted, a mixture of NaBH 4 and ZnCl 2 used as reagents can often lead to the formation of Na 2 ZnCl 4 by-product, which blocks Zn 2+ cation availability for the metathesis reaction (Equation (40)).

Multi-Cationic Borohydrides
The first reported mixed alkali metal borohydride, LiK(BH 4 ) 2 , was observed to form when a mixture of LiBH 4 and KBH 4 was mixed and heated to 125 • C [218]. This intermediate compound was fully characterized, including by crystal structure determination: Li + coordinates to four [BH 4 ] units, while K + coordinates seven [BH 4 ] tetrahedral unit. The decomposition temperature of LiK(BH 4 ) 2 is almost half of that of individual parent alkali-metal borohydrides.
Other mixed cation borohydrides are metastable, such as NaK(BH 4 ) 2 ; upon cooling to RT, the mixed metal borohydride separates into the monomeric borohydrides within one day [219]. The certainty of its formation by heating the system NaBH 4 -KBH 4 was confirmed by an increase in the lattice parameter a, and the higher potassium content [220]. However, it was later proved that the mixed borohydride NaK(BH 4 ) 2 is not stoichiometric, but a solid solution of approximate formula Na p K 1-p BH 4 (0 < p < 1), and only stable at elevated temperatures (200-450 • C) [221].

Anion Substitution of MBH: From Light to Heavy Halides and Pseudo-Halide Substitution
Substitution of BH 4 − was reported for LiBH 4 -LiCl [228], with this mixture actually forming a partial solid solution of anticipated formula Li(BH 4 ) 1-p Cl p (0 < p < 1), an as-pect which hindered further purification of the borohydride. The superionic HT-LiBH 4 is formed at~110 • C (108 • C), but this phase transition could occur at even lower temperatures possibly due to the formation of Li(BH 4 ) 1-p Cl p during ball-milling [89,228]. Easier substitution has been confirmed for halide anions (F − , Cl − , Br − , I − ) and it allowed for fine-tuning of the unit cell parameters and eventually to physical properties related to the crystallographic system and strain [229]. The shrinkage or enlargement of the unit cell volume mirrors the size of the anions: F − (1.33 Å) < Cl − (1.81 Å) < Br − (1.96 Å) < BH 4 − (2.05 Å) < I − (2.20 Å) [153,230].
Depending on the crystallographic system and local environment, some polymorphs undergo anion substitution more easily than others, as shown by HT-LiBH 4 and β-Mg(BH 4 ) 2 [231,232]. If both components of the system MBH 4 -MX (M-alkali metal, X-halide) are isostructural, anion substitution may occur both ways, such that two solid solutions can form: LiBH 4 -LiBr, LiBH 4 -LiI, or NaBH 4 -NaCl. While heavier halides (Br − , I − ) stabilize the hydride, hindering hydrogen release at lower temperature, they lower the energy of hydrogen uptake during rehydrogenation [233][234][235][236]. While o-LiBH 4 (orthorhombic) shows no conduction at RT, its HT-polymorph, HT-LiBH 4 (hexagonal, stable at t > 110 • C), exhibits fast lithium-ion conductivity that is enhanced upon iodide anion I − substitution of 25% with respect to starting BH 4 − units, in Li[BH 4 ] 0.75 I 0.25 (Equation (41)) [237,238]. J. Rude et al. showed through usage of synchrotron radiation powder X-ray diffraction (SR-PXD) and attenuated total reflectance infrared spectroscopy (ATR-IR) that two solid solutions of hexagonal structure (just like HT-LiBH 4 and β-LiI) form in the system LiBH 4 -LiI, which combine into one upon heating. The solid solutions of composition Li(BH 4 ) 1-p I p (0 < p < 0.62) have been found to be stable over a 300-degree temperature range, and after four hydrogen release/uptake cycles they retained 68% of expected hydrogen capacity, in contrast to only 25% retained by pristine LiBH 4 .
Anion substitution has also been expanded to include amide -NH 2 -anions, and these boron-nitrogen species are promising from a hydrogen storage viewpoint and for properties of fast ion conductivities, as seen above in the case of halide substitution. A dihydrate adduct has been isolated in the case of NaBH 4 ·2H 2 O crystallohydrate, where Na+ has octahedral coordination to four [BH 4 ] units and two H 2 O molecules bridging two Na + (space group: orthorhombic, Pbca) [153,239]. NaBH 4 ·2H 2 O loses the two wa-ter molecules upon mild heating at 40 • C, and then slowly starts to release hydrogen (Equation (42)).
Depending on the number of coordinating ligands, the thermal decomposition of ammine metal borohydrides can follow different pathways, as shown in Equation (44) [245].
Reactions (45)-(47) were performed under Ar flow, when NH 3 was released. Running the reaction in a closed vessel, no NH 3 evolution was detected and instead a sudden dehydrogenation reaction with onset at 190 • C took place (Equation (48)). The final dehydrogenated species was assigned based on 11 B NMR spectra, consistent with an overall hydrogen storage capacity of Ca(BH 4 ) 2 ·2NH 3 of 12.3 wt.%. This diammoniate system was not reversible, as no rehydrogenation occurred during attempts at 50-bar H 2 and 20-300 • C [249].
The closo-borates are currently under active research due to their rapid ion conduction, due in part to the open structure with channels for ionic conduction [281]. Both Li 2 B 12 H 12 and Na 2 B 12 H 12 exhibit high ionic conductivities of the order of magnitude of 10 −1 S cm −1 which are beneficial for Na-battery electrolytes [257][258][259]. The superionicity found in these closo-salts could also originate from the high rotational mobility of the large B 12 H 12 2− anion [259].

Optical and Magnetic Properties
As expected for RE (RE = rare earth) complexes, Eu(BH 4 ) 2 ·2THF has blue luminescence and an impressive quantum yield of~75%, due to charge separation of Eu 2+ centers by borohydride BH 4 − anions, which avoid quenching. Similar favorable lumines-cent properties were found in Y(BH 4 ) 2 ·2THF [282]. The ionic bond Eu 2+ -BH 4 − allows d =⇒ f emission in the blue region of the spectrum. A series of mixed metal borohydrides of perovskite-type structure, such as KYb(BH 4 ) 3 and CsEu(BH 4 ) 3 , have also been investigated, with a red shift in the latter compared to parent Eu(BH 4 ) 2 ·2THF [225,226]. Avoidance of quenching has been proven by addition of a 5% Eu 2+ in CsEu(BH 4 ) 3 , which showed similar luminescence properties regardless of [Eu 2+ ] doping concentration [225,226].

M(BH 4 ) x as Semi-and Superconductors
Borohydrides of perovskite structure could be fine-tuned for photovoltaics, as their typical band-gap (>5 eV) could be lowered to~1.5 eV at RT in the particular case of CsPb(BH 4 ) 3 , which has semiconductor behavior. The reasoning for this seems to lie in the partial covalent bonding in the Pb(BH 4 ) 3 framework, supported by the p(Pb) states present in the conduction band edge [225].
Superconductivity has been achieved for metal hydrides at very high pressures, exceeding 100 GPa, but providing a solid proof-of-concept. While the lanthanum-based borohydride LaBH 8 is not yet confirmed experimentally, computations predict it to be a superconductor and stable at a considerably lower pressure of about 40 GPa [283][284][285]. However, recent reports (2021) suggest that under pressure (12 GPa) a ternary alkali borohydride KB 2 H 8 shows phonon-mediated high-temperature superconductivity [226].

CO 2 Capture
Reaction of MBH with CO 2 (g) can be used for carbon dioxide sequestration/removal, or to produce derivatives like formates (Equation (53)) [286,287].

MBH(HCOO) 3 (DME)
The reaction with CO 2 can follow different pathways, depending on temperature and solvent used (DME, dimethoxyethane), which adds to the growing interest in CO 2 capture globally [288,289]. LiBH 4 and NaBH 4 reaction with CO 2 was also monitored in non-catalytic, solvent-free conditions, when IRGA (infrared gas analysis), FTIR, 11 B NMR, and 13 C NMR confirmed the presence, respectively, of borate, formate, and methoxy species [290].
Reaction of KBH 4 under solvent-free ball-milling with solid CO 2 leads to a formylhydroborate species K[H x B(OCHO) 4-x ] (x = 1-3) [291]. The synthesis of this interesting substituted borate species follows a three-step, consecutive nucleophilic attachment of the borohydride H − on the electron-deficient carbon of CO 2 molecules (Equation (54)).
While reaction of MBH with H 2 O usually follows the pattern seen in Equations (42) and (55), variations including running the reactions under small CO 2 amounts have proven fruitful in synthesizing novel substituted metal borohydrides [292].
Contrary to the Li + and Na + counterparts, KBH 4 borohydride produces under CO 2 -promoted hydrolysis a metal complex still retaining the [BH 4 − ] moiety: K 9 [B 4 O 5 (OH) 4 ] 3 (CO 3 )(BH 4 )·7H 2 O. This complex was investigated by XRD, DSC, TGA, Raman, and IR spectroscopies and fully characterized by solving its single crystal structure, showing a hexagonal crystal structure belonging to the P-62c space group, and high hydrogen gravimetric content. The reaction proposed by Filinchuk et al. is complex and leads to potassium carbonate as by-product (Equation (56)), along with a large amount of hydrogen, so that it can be regarded as having a double utility: removal of CO 2 from car exhaust gases, and efficient conversion of CO 2 to a clean fuel, H 2 [286]: The porous polymorph of γ-Mg(BH 4 ) 2 adsorbs CO 2 under very mid, near-RT conditions and also yields formate and methoxy products, proving suitable for CO 2 sequestration. The highest CO 2 uptake of 12 mol/kg was recorded after 7 days of exposure at 1-bar CO 2 and 30 • C, and was attributed to the very porous nature of magnesium borohydride polymorph, which features a very large specific surface area [293].

Thermodynamic Properties of MBH
Thermal decomposition of metal borohydrides is directly related to the electronegativity of the M metal, which roughly displays a clear decrease in decomposition temperature with the increase in M electronegativity [294].
Pauling hypothesized that the energy E(A-B) of a heteroatomic bond (A = B) relates to the difference in electron affinities of A and B by Equation (57), where the second term can be approximated in the case of compounds MR n (R-anion) by −∆H f /n.
The decomposition temperature (T d , K, and t d , • C; T d [K] = t d ( • C) + 273.15) can be correlated to the Pauling electronegativity of the cation [295][296][297][298][299]. The empirical formula (Equation (58)) used to determine T d was found to be [300]: The van't Hoff equation adapted for decomposition of metal borohydrides (Equation (58)) contains the factor 3.11, which can be regarded as the electron attractive coefficient of borohydride anion [BH 4 ] − , while 175.914 is a coefficient containing entropy values of decomposition (oversimplified as being constant for all compounds, in a first approximation) ( Figure 12).
A representative plot for metal borohydrides presented in Figure 2 is given in Figure   Figure 12. Decomposition temperature of M(BH 4 ) x dependent on cation (*Pauling) electronegativity following a van't Hoff equation plot. In green are highlighted elements forming borohydrides with predicted decomposition temperature 50-150 • C, suitable for mobile application. Very toxic elements are in red [295][296][297][298][299].
∆H boro = 248.7 χ P − 390.8 ; T d = 175.914(χ P − 3.11) 2 (59) A representative plot for metal borohydrides presented in Figure 2 is given in Figure 13, where a linear correlation between enthalpy and electronegativity χ P is confirmed. Pauling electronegativities are preferred because when inferring Mulliken scaling ( ), the values are fitted for ΔHboro with a higher error (13.1 kJ/mol) and moreover the enthalpies for Cu + and Zn 2+ borohydrides are almost the same (Mulliken electronegativities: 4.48 eV and 4.45 eV, respectively), which cannot explain the differences observed experimentally between the two [294,301,302]. Based on Equation (59), it can be inferred that for metal borohydrides having ΔH 0 f ≥ 0, that is for ≥ 1.57, the borohydride would have an endothermic formation enthalpy and could be considered unstable (Table 3). For selected M(BH4)x the free Gibbs energy ΔG has also been computed. Group I and II metal borohydrides have been the most investigated regarding thermodynamic data, and the following suggestive formulae were found (in kJ/mol formula unit); ∆ = −222466. 273    Pauling electronegativities are preferred because when inferring Mulliken scaling (χ M ), the values are fitted for ∆H boro with a higher error (13.1 kJ/mol) and moreover the enthalpies for Cu + and Zn 2+ borohydrides are almost the same (Mulliken electronegativities: 4.48 eV and 4.45 eV, respectively), which cannot explain the differences observed experimentally between the two [294,301,302]. Based on Equation (59), it can be inferred that for metal borohydrides having ∆H 0 f ≥ 0, that is for χ P ≥ 1.57, the borohydride would have an endothermic formation enthalpy and could be considered unstable (Table 3). For selected M(BH 4 ) x the free Gibbs energy ∆G has also been computed. Group I and II metal borohydrides have been the most investigated regarding thermodynamic data, and the following suggestive formulae were found (in kJ/mol formula unit); ∆G LiBH 4 = −222466. 273 [205] Screening the known borohydrides, one might expect the following elements to form unstable borohydrides: Cu, Ag, (Be), Cr, Fe, Ni, Co, Zn, Cd, Hg, Al, V, Ga, Nb, In, Ge, and Sn. Many of these are known, but either they are stable at very low temperatures or they need other techniques for stabilizations, i.e., formation of adducts with neutral molecules ( Table 2).
Many borohydrides follow a decomposition scheme leading to production of B 2 H 6 and H 2 in a 1:1 molar ratio (Equation (60)). These pathways that lead to production of diborane (toxic gas) will be, unfortunately, not useful for expanding the class of hydrogen storage material because they lead to a loss of the boron source [300].

Destabilizing Methods for Complex Metal Borohydrides
Nanoconfinement is a valuable method for the synthesis of nano-sized MBH. This route can be utilized to lower energy barriers and speed up kinetics. Nanoscaffolds can be engineered to have highly advantageous textural properties: tunable pore size (D p ), as well as high pore volume (V p ) and specific surface area (SSA), which in turn will greatly enhance the kinetics of hydrogen release/uptake. These recyclability enhancements will be described for representative metal borohydride systems and their composites. Anion substitution in complex metal hydrides can lead to system destabilization: TiF 3 and TiCl 3 were investigated in NaBH 4 and showed a 6.6 kJ/mol decrease in decomposition enthalpy, by altering hydrogen release and uptake [303]. The replacement of H − by F − was investigated in LiBH 4 as well, showing significant theoretical thermodynamic improvements [304]. Nanoconfinement improvements brought about by size-restriction are discussed in relation to their bulk state in the following chapter (Section 7). Some volatile borohydrides such as Zr(BH 4 ) 4 could be embedded into mesoporous silica (MCM-48 type) to show enhanced stability towards air and moisture, and more importantly, to retain Zr(BH 4 ) 4 inside siloxanic pores, thus mitigating the high volatility issue of neat Zr(IV) borohydride.
Thermal destabilization by utilizing mixtures of metal borohydrides has proven to be an effective technique [305]. Coupled dehydrogenation of a physical mixture (1 − x)LiBH 4 + xCa(BH 4 ) 2 comprised of o-LiBH 4 and α-/γ-Ca(BH 4 ) 2 (synchrotron XRD patterns) also led to enhanced ionic conductivity of the resulting cubic-CaH 2 . Moreover, it seems that the ionic conductivity may be solely due to Li + and irrespective of Ca 2+ content, as shown by samples (1 − x)LiBH 4 +xCa(BH 4 ) 2 with similar ionic conductivity to that of pure ball-milled LiBH 4 [305]. Further investigations by first-principles computations of a similarly coupled system by Ozolins et al. have shown that the decomposition reaction yields the complex intermediate Li 2 B 12 H 12 alongside the anticipated CaH 2 and hydrogen H 2 , with an endothermic enthalpy of 37.9 kJ/mol and 6.7 wt.% H 2 (Equation (61) (66) and (67)). It is thermodynamically more favored the addition of CaH 2 rather than MgH 2 to CaB 12 H 12 , as the switch from MgH 2 to CaH 2 used for destabilization could lower the operating temperature of the system even more, from 144 • C to 86 • C [306].
The 11 B NMR data showed B 2 H 6 to be an impurity during dehydrogenation processes. Therefore, running the reaction of LiBH 4 with B 2 H 6 confirmed the formation of the hypothesized Li 2 B 12 H 12 during the dehydrogenation pathway (Equation (82)) [330].
Attempts to lower the hydrogen desorption to temperatures below 400 • C have been successful when using SiO 2 powder addition (LiBH 4 :SiO 2 = 1:3) [331]. Three important peaks in TPD spectra of LiBH 4 corroborated the NMR studies, Raman spectroscopy, and XRD data, all of which point to the formation of dodecahydro-closo-dodecaborate species [B 12 H 12 ] 2− , and possibly other boron hydrides as intermediates [328,329].
Nanoconfinement of LiBH4 was investigated in a variety of supports (2-25 nm), such as mesoporous silica (MSU-H, SBA-15 etc.) or mesoporous carbon (carbon replica of MSU-H silica: C-MSU-H, or of SBA-15 silica: C-SBA-15 i.e., CMK-3) (Figure 15) [334][335][336]. A pertinent overview of C-replica synthesis of mesoporous silicas comes from M. Kruk et al. [337] and R. Ryoo et al. [338]. Mesoporous SiO2, however, is to be avoided, since an irreversible reaction with LiBH4 during dehydrogenation could lead to lithium silicates (Figure 16) [335,339]. Attempts to improve hydrogen uptake by using Mo-decorated mesoporous silica (MSU-H type) increased the reversible hydrogen capacity from 3.2 wt.% (pristine LiBH4) to 5.2 wt.% H2 in the nanocomposite LiBH4-Mo:MSU-H [335]. Formation of lithium silicates (Li2SiO3, Li4SiO4) was confirmed by XRD and synchrotron powder XRD, but only after LiBH4 decomposed in the as-synthesized composite LiBH4@meso-SiO2; the nanocomposite seems otherwise stable at temperatures t < 275 °C (Equation (84) Nanoconfinement can be performed by melt infiltration or solvent infiltration, both of which lead to a decrease in the crystallite size of the LiBH4 [341,342]. Incipient wetness technique has been used to infiltrate MTBE (methyl tert-butyl ether) solution of LiBH4 in mesoporous carbon (C-MSU-H). The maximum loading attempted was about 40 wt.% LiBH4, but the XRD data confirms that even for a 20 wt.% loading some diffraction peaks of LiBH4 appear. This is indicative of the borohydride species crystallizing outside of the mesopores [334].  Formation of lithium silicates (Li 2 SiO 3 , Li 4 SiO 4 ) was confirmed by XRD and synchrotron powder XRD, but only after LiBH 4 decomposed in the as-synthesized composite LiBH 4 @meso-SiO 2 ; the nanocomposite seems otherwise stable at temperatures t < 275 • C (Equation (84)) [89,335,340].
Nanoconfinement can be performed by melt infiltration or solvent infiltration, both of which lead to a decrease in the crystallite size of the LiBH 4 [341,342]. Incipient wetness technique has been used to infiltrate MTBE (methyl tert-butyl ether) solution of LiBH 4 in mesoporous carbon (C-MSU-H). The maximum loading attempted was about 40 wt.% LiBH 4 , but the XRD data confirms that even for a 20 wt.% loading some diffraction peaks of LiBH 4 appear. This is indicative of the borohydride species crystallizing outside of the mesopores [334].

= (85)
This approximation was performed assuming the same pre-exponential Arrhenius factor A (an oversimplification), but this shows that nonporous carbon barely has any benefit from confinement, whereas the rate in nanoconfined composites is ~10 6 times faster. Moreover, reducing the nanoscaffold from 24 nm to 13 nm affords a rate that is 23.8 The influence of carbon porosity was investigated showing that a proper impregnation of LiBH 4 inside 1.8-3.2 nm activated carbon (S BET = 1341 m 2 /g; V p = 0.87 cm 3 /g) is accompanied by reduction of both surface area (to 245 m 2 /g) and pore volume (to 0.2 cm 3 /g) [336,343]. The size of the scaffold alters the activation energy E a of dehydrogenation: non-porous graphite (146 kJ/mol LiBH 4 ), 24 nm RF-CA (111 kJ/mol LiBH 4 ), and 13 nm RF-CA (103 kJ/mol LiBH 4 ), consistent with inverse proportionality between nanoscaffold size and activation energy E a [344,345]. The Arrhenius Equation (85) shows that, taking solely into account the Ea of these three supports, one might expect the dehydrogenation rates to follow this order: k graphite :k 24nm :k 13nm = 1:e 14.03 :e 17.20 (at 298.15 K), and k graphite :k 24nm :k 13nm = 1:e 11.28 :e 13.86 (at 373.15 K).
This approximation was performed assuming the same pre-exponential Arrhenius factor A (an oversimplification), but this shows that nonporous carbon barely has any benefit from confinement, whereas the rate in nanoconfined composites is~10 6 times faster. Moreover, reducing the nanoscaffold from 24 nm to 13 nm affords a rate that is 23.8 times higher (at 25 • C) or 13.2 times higher (at 100 • C, a temperature considered optimal for operation of fuel cell operation). Nanoconfined LiBH 4 shows a decrease in melting temperature of 30 • C (in 13 nm RF-CA) or 23 • C (in 24 nm RF-CA), but only when melt infiltration was employed; a modest 1 • C decrease was recorded for the solvent infiltration technique [341,346].
When a 33.67 wt.% loading of LiBH 4 was achieved in C-SBA-15 (CMK-3 carbon, average diameter 4 nm) prepared by the concentrated solution infiltration technique, a clear reduction in temperature onset of dehydrogenation was recorded, and the composite LiBH 4 @CMK-3 exhibited complete dehydrogenation up to 400 • C [336]. The solvent infiltration approach can alter the dehydrogenation pathway, as well as the reaction kinetics; for comparison, a ball-milled sample containing LiBH 4 affords full dehydrogenation above 450 • C [347].
While bulk dehydrogenated LiBH 4 (LiH and B) can be rehydrogenated under harsh conditions (p, t, Equation (83)), the nanoconfined borohydride can be regenerated using 100-bar H 2 at below 400 • C within 2 h when microporous carbon was used as scaffold. The improved recyclability of such a composite points to better reversibility when pore size decreases to a microporous range (D pore < 2 nm) [341]. A comparison between nanoconfinement of LiBH 4 in carbon and silica supports of pore radii 8-20 Å showed improvement in desorption kinetics, reversibility, and Li + ionic conductivity with decreasing pore size. It was shown that similarly sized silica and carbon scaffolds produce different effects upon the melt infiltration of LiBH 4 , with silica-based support showing obvious improvements compared to carbon scaffolds, an aspect attributed to increased LiBH 4 interfacial layer thickness "t" for SiO 2 (1.94 ± 0.13 nm) compared to C-based support (1.41 ± 0.16 nm). In this regard, the Gibbs-Thomson equation correlates with the shift (depression) in melting temperature (∆T) to interfacial effects, namely, it increases with the interface energy (∆γ: 0.053 J/m 2 for LiBH 4 /SiO 2 and 0.033 J/m 2 for LiBH 4 /C system) and decreases with pore radius (r p ) (Equation (86)) [348,349].
The phase transition shift for LiBH 4 (~114 • C for bulk) was observed at 65 • C for LiBH 4 /SiO 2 (3.3 nm) and at 19 • C for LiBH 4 /C m-4.9 (of high microporosity). P. Ngene et al. assumed a core-shell model and found a direct relation for computing the enthalpy for the confined material, which could allow fine-tuning of thermodynamic parameter by adjusting pore size and interfacial thickness (Equation (87)) [350].
Pore size distribution (PSD) and pore geometry should always play a decisive role in nanoconfinement results, especially for C-nanoscaffolds where microporosity percentage could be dominant [351]. The nano-synergy effect has recently made possible a reversible hydrogen capacity of 9.2 wt.% [352]. Investigation of possible side-pathways in a dehydrogenation reaction and particularly establishing the nature of intermediate species (such as Li 2 B 12 H 12 ) could trigger new directions in nanoconfined borohydride materials [41,[323][324][325].

LiBH 4 + MgH 2
Magnesium-based materials (alloys, hydrides, alanates, borohydrides, etc.) have sparked interest as promising hydrogen storage materials due to their low cost, high wt.% hydrogen storage, and interesting properties, including the much sought-after recyclability/reversibility condition for any future energy storage material [353]. Among all hydrides, MgH 2 is very promising due to high abundance of magnesium, low cost, and reversibility behavior in hydrogenation studies, along with high hydrogen uptake capacity (7.7 wt.%) and very high energy density (9 kJ/g Mg) [354][355][356][357][358][359]. However, the desorption kinetics are slow at 300 • C and 1-bar H 2 , while the reactivity towards air (oxygen in particular) and the hydrogen desorption has a high enthalpy-thus requiring operation at high temperature [355,360]. Lowering desorption enthalpy can be achieved by introducing high-reactivity defects in the structure (high-energy ball-milling) reaction with a destabilizing element or with a catalyst [361].
Given the advantages of using the Mg-based hydrides mentioned above, inclusion of MgH 2 in various reactive hydride composites (RHCs) has been attempted. The most studied such reactive composite is 2LiBH 4 + MgH 2 (11.6 wt.% hydrogen storage capacity, according to the decomposition pathway in Equation (88). While taken individually, both LiBH 4 and MgH 2 have unfavorably high thermal stability and slow kinetics, and their 2:1 mixture is found to produce MgB 2 instead of the rock-stable boron B, which brought new hopes in pursuit of a viable energy-storage material (Equation (88)) [92,306,362].
Nanoconfinement of the (2LiBH 4 + MgH 2 ) system was attempted in 21 nm RF-CA porous scaffolds showing 4 wt.% reversible hydrogen storage capacity (92% of theoretical storage of the composite RHC@RF-CA, after three cycles), and it was found that decomposition in the nanoconfined state mirrors that from the bulk (Equation (88)) [346]. The hydrogen capacity is higher than that of nanoconfined MgH 2 @RF-CA 22nm (1.4 wt.%) [133].
DSC analysis of the composite 2LiBH 4 -MgH 2 shows four endothermic peaks: 113, 267, 332, and 351 • C for the nanoconfined state in 21 nm RF-CA vs. 117, 290, 364, and 462 • C in bulk (Table 6) [346]. The hydrogen desorption kinetics are clearly improved by nanoconfinement, with MgH 2 registering a 32 • C lowering of dehydrogenation temperature, while dehydrogenation of LiBH 4 is lowered by a significant 111 • C. Regardless of the kinetic improvements during recyclability studies, the high thermal stability of bulk metal borohydrides remains an obstacle in their use as hydrogen storage materials.

NaBH 4
With a hydrogen storage capacity of 10.6 wt.%, NaBH 4 is a valuable reducing reagent used in organic synthesis and decomposes at a high temperature typical of ionic compounds like metal salts (534 • C) [370][371][372][373][374]. The synthesis was reported first by Schlesinger, who reacted trimethyl borate with sodium hydride in hydrocarbon oil, at boiling point (250 • C) (Equation (90)) [375]. The sodium methoxide is hydrolyzed (NaOCH 3 + H 2 O =⇒ NaOH + CH 3 OH) to form CH 3 OH, which allows separation of the reactive mixture, and a sodium hydroxide solution of NaBH 4 that allows borohydride extraction using isopropylamine.
An alternative method (Bayer process) utilizes Na 2 B 4 O 7 ·7SiO 2 with Na and 3 atm H 2 at 400-500 • C in a one-pot synthesis [376,377]. This process allows for obtaining large quantities of NaBH 4 , but it does present the risk of explosion (the working temperature is above decomposition of NaBH 4 ) and it produces large quantities of Na 2 SiO 3 of lower market demand or value (Equation (91)).
The decomposition is different to that of LiBH 4 analogue, due to the lower stability of NaH compared to LiH, and it leads to the formation of the parent elements (Equation (93)). The decomposition process of NaBH 4 starts at~240 • C and releases the greatest amount of hydrogen above 450 • C.
In fact, an alternative reaction mechanism has been proposed concerning formation of the MgB 12 H 12 species in a first step (−15 ppm, singlet, in 11 B-MAS NMR spectrum), dehydrogenation of MgH 2 in a second step, and a final reaction of Mg and MgB 12 H 12 forming MgB 2 -which we know is the final dehydrogenation "resting-state" of magnesium ( Figure 18). The decomposition temperature could be lowered to 100°C by the addition of TiCl3 catalyst [109]. Rehydrogenation of MgB2 under 100-bar H2 at 350 °C showed ~3 wt.% hydrogen capacity, confirming the reversibility of this dehydrogenation step [156,385]. The intermediacy of MgB12H12 was investigated, as in the case of other metal borohydrides, and is considered plausible [387].
The first decomposition step (277 °C) has an enthalpy of −40 …57 kJ/mol, depending on experimental data [385,387], and a computed value of 38 kJ/mol (DFT computations) The decomposition temperature could be lowered to 100 • C by the addition of TiCl 3 catalyst [109]. Rehydrogenation of MgB 2 under 100-bar H 2 at 350 • C showed~3 wt.% hydrogen capacity, confirming the reversibility of this dehydrogenation step [156,385]. The intermediacy of MgB 12 H 12 was investigated, as in the case of other metal borohydrides, and is considered plausible [387].
The first decomposition step (277 • C) has an enthalpy of −40. . . 57 kJ/mol, depending on experimental data [385,387], and a computed value of 38 kJ/mol (DFT computations) [388]. In situ synchrotron radiation SR-PXD data show as final dehydrogenation products a mixture of Mg, MgO, and MgH 2 , but these results should be considered with care given the high affinity of Mg for oxygen and proneness to subsequent oxidation [118].
Regeneration of Mg(BH 4 ) 2 by hydrogenation of MgB 2 requires high H 2 pressure and temperature (950-bar H 2 , 400 • C), but also prolonged time (108 h), which makes the bulk material impractical for use in a tank for mobile applications [386]. When starting from nano-sized MgB 2 synthesized at RT, a partial rehydrogenation to Mg(BH 4 ) 2 occurred at 300 • C and 30-bar H 2 , leading to a material with 2.9 wt.% hydrogen storage, but quite far off the theoretical value for magnesium borohydride (14.8 wt.%) [389,390].
Mg-based materials greatly benefit from nanoconfinement in carbon nanotubes, which reduces the enthalpy of dehydrogenation even by 50% (for MgH 2 , a reduction from 36 kJ/mol H in bulk, to 17 kJ/mol H in 9-unit cell thick MgH 2 thin film slab) [393]. Nanoconfinement of Mg(BH 4 ) 2 in microporous carbon AC resulted in nanocomposites with loadings as high as 44 wt.%, albeit at the cost of partially crystallizing on the outer surface of AC (as seen by powder-XRD and EDS data analysis) [382,383]. The total weight loss of 6 wt.% corresponds to 44% borohydride loading in the nanocomposite [382]. Small angle neutron scattering (SANS) is a technique allowing for the study of particle size and distribution. Using isotopically labeled magnesium borohydride, Mg( 11 BD 4 ) 2 @AC, Sartori et al. observed nanoconfined borohydride particles of less than 4 nm [383]. Improved hydrogen release is obvious from the shifting of the endothermic peak events from 269 • C to 257 • C (−12 • C decrease) and from 376 • C to 317 • C (−50 • C decrease), and it can be ascribed to the nanoconfinement effect.

Ca(BH 4 ) 2
Ca(BH 4 ) 2 has a hydrogen storage capacity of 11.5 wt.% (maximum theoretical capacity), but the real capacity is only 9.6 wt.% considering the actual dehydrogenation reaction (Equations (12) and (95)). Equation (95) is essentially the same as Equation (12), but written for one mol Ca(BH 4 ) 2 , as the reaction is reversible and decomposition and rehydrogenation processes follow the same reaction, having a ∆H = 32 kJ/mol [188].
The orthorhombic (α-phase) can be obtained free of THF solvent by evacuation under vacuum of the commercially available THF adduct. There is another low-temperature orthorhombic phase proposed for Ca(BH 4 ) 2 (γ-phase polymorph) [118]. The decomposition temperatures are higher than those anticipated based on thermodynamic data, partly due to sluggish kinetics [394]. Heating α-Ca(BH 4 ) 2 to 170 • C leads to a phase transition to β-Ca(BH 4 ) 2 and upon further heating the DSC scan data shows two endothermic peaks at 360 • C and 500 • C [395,396]. The rehydrogenation of this reversible system can proceed at 400 • C and 690 atm H 2 , but these conditions are too harsh to warrant the use of bulk Ca(BH 4 ) 2 in vehicular applications [95]. The influence of using catalysts (Nb, Ti) showed beneficial reduction in hydrogen desorption temperature, but the rehydrogenation conditions are far from ideal (350 • C, 150-bar H 2 , 12-24 h) and the re-hydrogenated material exhibits 4.5 wt.% H 2 capacity [397]. Using additives like TiCl 3 could lower even further re-hydrogenation kinetics, making it possible when dehydrogenated material was subjected at 350 • C for 24 h at a lower pressure of 90-bar H 2 [394]. Ball-milling Ca(BH 4 ) 2 with 5 mol%TiCl 3 (Ca(BH 4 ) 2 :TiCl 3 = 1:0.05) leads to hydrogen evolution and formation of CaCl 2 , according to Equation (96) [394].
This is also an example where the additive (TiCl 3 ) alters the dehydrogenation pathway (commonly accepted as described by Equation (95)), as the proposed weight-loss based on DSC and TGA curves conform to about 7.1 to 8.7% reduction. This is consistent with the following decomposition pathway, producing CaH 2 , B, and H 2 (7.2 wt.% hydrogen storage capacity based on Equation (97)) [394]. The hydrogen storage capacity after rehydrogenation is only 3.8 wt.%; complete rehydrogenation was thus not achieved. It is, however, apparent that TiCl 3 additive plays an important role in lowering activation energy for the hydrogenation reaction.
A proof-of-concept process for complete hydrogen release/uptake was shown for a ball-milled mixture CaB 6 + 2 CaH 2 + 8 wt.% TiCl 3 /Pd, but the reaction conditions were harsh (700-bar H 2 , 400-440 • C); however, almost complete recyclability with 9.6 wt.% hydrogen storage has been demonstrated [95]. A key role in this achievement is thought to be played by the deposition of TiCl 3 on porous Pd, which is itself known for absorbing record amounts of hydrogen and giving interstitial hydrides, while other TM halides (RuCl 3 ) led to polymorphs of Ca(BH 4 ) 2 and other unknown phases, confirming the alteration in the rehydrogenation pathway [95,398].
Nanoconfined Ca(BH 4 ) 2 in microporous carbon showed tangible improvements in the recyclability behavior of as-prepared nanocomposites [143]. Using carefully engineered chemically activated micro-mesoporous carbon as scaffolds (specific surface area SSA = 1780 m 2 /g, V pore~1 cm 3 /g and 60% microporosity), Comanescu et al. obtained by the incipient wetness method Ca(BH 4 ) 2 @MC-a nanocomposites with 36 wt.% borohydride loading ( Figure 19).  These Ca(BH 4 ) 2 @MC-a composites started the hydrogen release step at about 100 • C and managed to retain 2.4 wt.% hydrogen capacity even after 18 cycles of hydrogen release/uptake. The rehydrogenation of the desorbed composites needed rather mild conditions (25-40 atm H 2 , 6.5 h), confirming the important role of the nanoporous scaffold of choice (Figure 20) [143]. Figure 19. TPD curves of Ca(BH4)2@MC-a nanocomposites with a heating rate of 2.5 °C/min. Re printed with permission from [143].
These Ca(BH4)2@MC-a composites started the hydrogen release step at about 100 °C and managed to retain 2.4 wt.% hydrogen capacity even after 18 cycles of hydrogen re lease/uptake. The rehydrogenation of the desorbed composites needed rather mild cond tions (25-40 atm H2, 6.5 h), confirming the important role of the nanoporous scaffold o choice (Figure 20) [143].

Ammonia Borane NH3BH3
A special place is reserved to ammonia-borane, an interesting adduct with 19.6 wt.% hydrogen. NH3BH3 (ammonia-borane, AB) and ammonium borohydride (NH4BH4) hav been reported from 1958, but NH4BH4 is far too unstable releasing H2 at even −40 °C t form AB [399]. NH3BH3 is a stable solid, water-sensitive, and susceptible to oxidation o attack by acidic reagents, having a high solubility in water (33.6g/100 g H2O). The A molecules present dipole-dipole interactions due to the dihydrogen bonding, i.e., interac tion between a hydridic hydrogen (-NH3) and an acidic one (BH3) [400], and organizes i an orthorhombic phase at low temperature (<225 K) due to these interactions. While it i is an improvement over previously reported data on reversibility studies on Ca(BH 4 ) 2 , confirming that about 68% of Ca(BH 4 ) 2 behaves reversibly [143].

Ammonia Borane NH 3 BH 3
A special place is reserved to ammonia-borane, an interesting adduct with 19.6 wt.% hydrogen. NH 3 BH 3 (ammonia-borane, AB) and ammonium borohydride (NH 4 BH 4 ) have been reported from 1958, but NH 4 BH 4 is far too unstable releasing H 2 at even −40 • C to form AB [399]. NH 3 BH 3 is a stable solid, water-sensitive, and susceptible to oxidation or attack by acidic reagents, having a high solubility in water (33.6 g/100 g H 2 O). The AB molecules present dipole-dipole interactions due to the dihydrogen bonding, i.e., interaction between a hydridic hydrogen (-NH 3 ) and an acidic one (BH 3 ) [400], and organizes in an orthorhombic phase at low temperature (<225 K) due to these interactions. While it is not a proper metal borohydride, NH 3 BH 3 has the advantageous property of being RT-stable, exhibiting a two-step dehydrogenation at very reasonable temperatures: 114 • C (melting point, 6.5 wt.% H 2 release, forming polyaminoboranes (NH 2 BH 2 ) n ), and 150 • C (remaining H 2 release, formation of polyiminoborane (NHBH) n ), with the downsides of irreversibility, slow kinetics, and formation of "inorganic benzene" as by-product during dehydrogenation (borazine, c-(NHBH) 3 ) (Equation (98)). The N-B (nitrogen-boron) units are isoelectronic to C-C (carbon-carbon) moieties, and they can therefore be regarded as hydrocarbon analogues [145,401,402].
Mesoporous silica or carbon have served as scaffolds for NH 3 BH 3 impregnation, and this nano-reduction strategy bypassed borazine formation [145,402,403]. Even NH 3 release is suppressed when Li + -based dopants are used [402]. Some key thermodynamic aspects connected to NH 3 BH 3 confinement in nanoporous substrates are summarized in Table 7.
The reduction in decomposition temperature of AB@MOF was found to be linearly proportional to the reciprocal of AB size; in other words, the expectations are that a lower pore size nano-restriction would yield the best destabilization in ammonia borane nanocomposites [411].
Another advantage of utilizing SiO 2 -based scaffolds in AB nanoconfinement is that this does not lead to a reaction with the framework, as was the case for typical M(BH 4 ) x . It is also apparent from Table 7 that 1D-structured mesosilica (MCM-41) are less attractive as scaffolds than 2D-mesostructured silica (SBA-15) of hexagonal pore arrangement, although the pore size might suggest otherwise (4 nm for MCM-41 vs. 7.5 nm for SBA-15). These conclusions could be extended to other borohydride@silica nanosystems as well, and could provide a starting point for further nanoscaffold engineering research. Nanoconfinement alters the dehydrogenation pathway in AB, as utilizing mesoporous carbons like CMK-3 manages to avoid toxic gas evolution; borazine or ammonia are not detected in that case. It can be hypothesized that nanoconfinement might help rehydrogenation by ammonia borane restructuring inside scaffold nanopores, and by lowering activation energies to improve the thermodynamics of hydrogen release/uptake. Some advances have been made by suggesting a methanolysis cycle, where [NH 4 ][B(OCH 3 ) 4 ] and B(OCH 3 ) 3 are key intermediates [412]. Complete elucidation of a yet-unclear dehydrogenation mechanism, combined with collaborative theoretical/experimental research efforts, could pave the way to the ultimate hydrogen-storage material.

Conclusions and Outlook
It has been almost half a century since hydrogen was first considered a renewable energy carrier, aiming at replacing fossil fuel and building up towards a "hydrogen economy" goal. While not without its pitfalls, solid-state hydrogen storage exhibits obvious advantages compared to the physical storage of H 2 gas: it is safer, requires less extreme conditions for storage and regeneration, and occupies less space.
Among various metal hydrides, metal borohydrides have inherited structural flexibility, accommodating both mixed cations and anion substitutions, and can be tailored towards multi-functional materials. Borohydrides can be engineered to form novel composites that reversibly store hydrogen chemically, reaching very good hydrogen storage capacities, in line with the DOE's expectations for the near future. Recent studies conducted under high temperature and pressure (hydrogen, diborane) have shed new light on mechanistic aspects regarding formation and decomposition pathways of complex metal borohydrides.
Very recent progress on alkali metal borohydrides is directed towards controlling the kinetics of hydrogen production by utilizing nanoporous silica and carbon, Co-based catalysts, bimetallic catalysts (NiPt, NiCo 9 , RuPd, Co-Ru) and nano-catalyst composites, or harnessing the latent potential in alcoholysis (methanolysis, in particular) of said metal borohydrides. Kinetic studies on metal borohydrides currently target PEM fuel cell applications, and have shown that the hydrolysis mechanism has as a rate-limiting step the splitting of the H-OH bond. Transition metal nanoparticles (Ni, Pt, Zr, Co, Mn, Cu), mixedvalence oxides (Co 3 O 4 ), transition metal complexes (of Ru 3+ , for instance), and MOFs have shown controlled hydrolysis of alkali metal borohydrides. Engineered catalysts (Mg 95 Ni 5 ) have managed to suppress ammonia release from borohydride ammoniates by improving hydrogen release at very reasonable temperatures (70-80 • C). Exploring new decomposition pathways of such ligand-stabilized borohydrides has also gained new momentum. Divalent metal borohydrides stabilized by neutral molecules show high ionic conductivity and a real promise for use in solid-state batteries. The rich boron-hydrogen chemistry, together with the open channels of closo-structured borohydrides (B 10 H 10 2− , B 12 H 12 2− ), currently make research on closo-borate salts of alkali and alkali-earth metals a hot topic in ionic conductivity for solid-state electrolytes. Direct borohydride fuel cell (DBFC) and direct borohydride-hydrogen peroxide fuel cell (DBHPFC) systems research has expanded recently towards high-performance applications aimed at space exploration, by harnessing the power of nanoparticle catalysis (Ni, Cu, or bimetallic PdFe, PdAg, PdAu) to control the resultant power density. Various destabilization strategies of borohydride compounds have now included novel catalyst systems and complementary in-depth mechanistic insights. Nanostructuring of employed catalysts for hydrolysis reactions or de-/redrydrogenation studies has confirmed a positive effect by lowering activation energies associated with individual mechanistic steps. The porous polymorph of Mg(BH 4 ) 2 has gained attention recently by involvement in a core-shell, oxidation-resistant nanostructure (γ-Mg(BH 4 ) 2 @MgCl 2 ) to afford dehydrogenation onset as low as 100 • C. Theoretical computations on ternary borohydride materials (K 2 B 2 H 8 ) predict superconducting properties at high pressures and around 140 K. Fundamental research is constantly expanding the class of structurally characterized metal borohydrides with new members (Th(BH 4 ) 4 is a new addition from 2021), while more cost-effective and high-yielding synthesis routes emerge-the synthesis of NaBH 4 was reported by milling sodium tetraborate with Al and NaH as additive.
Meeting targets related to performance range, refueling time, car performance, and passenger space and safety are at the forefront of development of novel hydrogen storage materials and material-based storage system technologies. Mobile applications and vehicular applications in particular could receive a new breath of life when a proper hydride-based material is engineered by scientists, in line with safety, performance, and cost estimation goals.