progress in materials sciencematerials has generated an enormous amount of reports. in the last 10...

Progress in Materials Science 58 (2013) 30–75

Contents lists available at SciVerse ScienceDirect

Progress in Materials Science

journa l homepage : www.e lsev ie r .com/ loca te /pmatsc i

Mechanochemical synthesis of hydrogen storage materials

J. Huot a, D.B. Ravnsbæk b, J. Zhang c, F. Cuevas c,⇑, M. Latroche c, T.R. Jensen b

a Université du Québec à Trois-Rivières, 3351 des Forges, Trois-Rivières, Québec, Canada G9A 5H7b Center for Materials Crystallography (CMC), Interdisciplinary Nanoscience Center (iNANO), Department of Chemistry,Aarhus University, Langelandsgade 140, DK-8000 Århus C, Denmarkc ICMPE, CNRS, UMR 7182, 2-8 rue Henri Dunant, 94320 Thiais Cedex, France

a r t i c l e i n f o a b s t r a c t

Article history:Received 13 April 2012Accepted 9 July 2012Available online 27 July 2012

0079-6425/$ - see front matter � 2012 Elsevier Lthttp://dx.doi.org/10.1016/j.pmatsci.2012.07.001

⇑ Corresponding author. Tel.: +33 1 49 78 12 25E-mail address: [email protected] (F. Cueva

New synthesis methods are of utmost importance for most materi-als science research fields. The present review focuses on mecha-nochemical synthesis methods for solid hydrogen storage. Weanticipate that the general methods and techniques are valuablewith a range of other research fields, e.g. the rapidly expandingfields of ‘energy materials science’ and ‘green chemistry’ includingsolvent free synthesis. This review starts with a short historicalreminder on mechanochemistry, followed by a general descriptionof the experimental methods. The use of milling tools for tuningthe microstructure of metals to modify their hydrogenation prop-erties is discussed. A section is devoted to the direct synthesis ofhydrogen storage materials by solid/gas reactions, i.e. by reactiveball milling of metallic constituents in hydrogen, diborane orammonia atmosphere. Then, solid/solid mechano-chemical synthe-sis of hydrogen storage materials with a particular attention to ala-nates and borohydrides is surveyed. Finally, more specialisedtechniques such as solid/liquid based methods are mentioned alongwith the common characteristics of mechanochemistry as a way ofsynthesizing hydrogen storage materials.

� 2012 Elsevier Ltd. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312. Tuning of metal microstructures by mechanical milling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

2.1. BCC alloys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

d. All rights reserved.

; fax: +33 1 49 78 12 03.s).

http://dx.doi.org/10.1016/j.pmatsci.2012.07.001

mailto:[email protected]

http://dx.doi.org/10.1016/j.pmatsci.2012.07.001

http://www.sciencedirect.com/science/journal/00796425

http://www.elsevier.com/locate/pmatsci

J. Huot et al. / Progress in Materials Science 58 (2013) 30–75 31

2.2. Ti-based . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 332.3. Mg-based BCC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 362.4. Amorphization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

3. Synthesis of hydrides by mechanically-induced solid/gas reactions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.1. Binary hydrides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
3.1.1. Magnesium hydride . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 413.1.2. Titanium hydride . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 423.1.3. Vanadium hydride . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

3.2. Ternary hydrides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.2.1. ZrNi hydride . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 433.2.2. TiNi hydride . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 433.2.3. TiFe hydride . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 433.2.4. LaNi5 hydride . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 433.2.5. TiV hydride . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.3. Mg-based complex hydrides. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3.3.1. Mg2Fe hydride . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 453.3.2. Mg2Co hydride . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 483.3.3. Mg2Ni hydride . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.4. Alanates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.4.1. Lithium alanates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 513.4.2. Sodium alanates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 523.4.3. Potassium alanates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 533.4.4. Mixed alkali alanates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 533.4.5. Alkali-earth alanates. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.5. Synthesis of borohydrides by mechanical milling in diborane gas . . . . . . . . . . . . . . . . . . . . . . . . 563.6. Synthesis of metal amides by mechanical milling in ammonia gas. . . . . . . . . . . . . . . . . . . . . . . . 57

4. Synthesis of hydrides by mechanically-induced solid/solid and solid/liquid reactions . . . . . . . . . . . . . . 57
4.1. Mechanochemical synthesis of metal borohydrides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 574.2. Synthesis of novel alane and metal alanates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 614.3. Novel quaternary hydrides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
4.3.1. Metal borohydride amides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 644.3.2. Metal alanate amides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

4.4. Solid–liquid mechanically assisted synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
5. Final remarks and conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

1. Introduction

Synthesis of innovative materials for energy conversion and storage has received increasing focusduring the past decades due to the worlds increasing energy demands and simultaneous needs forenvironmentally friendly energy technologies. Hydrogen is recognized as a possible renewable energycarrier, but its large-scale utilization is mainly hampered by unsatisfactory properties of knownhydrogen storage materials. Hence, preparation and characterization of novel materials are receivingsignificant attention as reviewed elsewhere [1–4]. Traditionally, hydrogen storage materials, such asmetallic or complex hydrides, were prepared by solvent-based synthesis methods or by direct so-lid–gas hydrogenation reactions. However, during the past decade mechanochemical synthesis hasbecome one of the most utilized preparation methods for this class of materials, and is still expectedto hold a significant unexplored potential for development of novel approaches, e.g. for ‘green chem-istry’ including solvent free synthesis methods. In this work, recent progress within the experimentalmethods for preparation of hydrogen storage materials is surveyed.

In the mid-eighties, several research groups initiated the use of mechanical activation methods forthe synthesis of hydrides [5,6]. Mechanical milling (MM) of mixtures of elements under inert gas atmo-sphere was used to synthesize intermetallic compounds, which were subsequently exposed to hydro-

32 J. Huot et al. / Progress in Materials Science 58 (2013) 30–75

gen in an external device to form hydrides. This two-step approach allowed modifying the microstruc-ture of the alloy by milling to study its influence on the hydrogenation thermodynamics and kinetics[6–9]. Mechanical milling was very successful at improving hydrogenation kinetics via the synthesis ofnanocrystalline materials and simultaneous incorporation of selective additives during the millingprocess [10–12].

Recently, it has been shown that Severe Plastic Deformation (SPD) techniques could be used for thesynthesis and processing of metal alloys and their hydrides [13–26]. A few specific cases will be dis-cussed in the forthcoming sections. A possible advantage of SPD techniques over conventional millingis easier scaling up to industrial level. However, some specific nanocrystalline structures and nano-composites may only be synthesized through mechanical milling.

In the early nineties, solid–gas reaction facilitated by mechanical milling in reactive gases (nitro-gen, oxygen and hydrogen) was investigated. This approach was initially designated Reactive Mechan-ical Milling (RMM) and used for preparation of hydrides in hydrogen atmosphere [27]. Theexperiments were performed in small-sized milling vials under moderate hydrogen pressures (below2 MPa), often leading to an incomplete reaction between hydrogen and metals. The extent of the me-tal-hydrogen reaction was determined by ex situ XRD analysis of samples milled during a given periodof time [28,29]. Modern devices for RMM synthesis are now equipped with pressure and temperaturesensors that allow monitoring, e.g., hydrogen absorption during milling [30–32]. Hydrogenation reac-tions can be followed in situ as a function of milling time at working pressures up to ca. 15 MPa.

Today, mechanochemical synthesis of metal hydrides using ball milling has grown to become oneof the most frequently used methods. Typically, planetary ball mills are used, however other typessuch as rotational, vibratory or attritor mills are also operated [33]. The different types of mills differin their milling efficiency and capacity and in some cases additional arrangements for cooling, heating,gas loading etc. can be applied. Typically a few grams of material and balls are placed in the planetaryball mill to give a ball-to-powder weight ratio of 10:1–50:1. This approach offers the advantage thatthe milling vial can be loaded, sealed and unloaded under inert conditions in a glove box, and, ifequipped with valve connections, subsequently filled with reactive gas [30,31]. Thereby, the p,T phasespace for mechanochemistry has expanded significantly.

In some cases, especially for ductile materials, a process control agent (PCA) could be added to in-hibit particle agglomeration [33]. The PCAs can be solids, liquids, or gases. A wide range of PCAs hasbeen used in practice at a level of about 1–5 wt% of the total powder charge. The most common PCAsare stearic acid, hexane, methanol, ethanol, graphite and salts [33].

Several parameters can be varied for the ball-milling synthesis: milling speed, total milling time,vial and ball composition, powder-to-ball weight ratio, vial diameter, ball diameter and density, mill-ing temperature, milling atmosphere and pressure of the selected gas. The latter two parameters re-quire a special high-pressure vial. Most planetary mills only allow controlling the speed of the supportdisk. The speed of the planets, on which the milling vials are mounted, is usually fixed relatively to thespeed of the main disk. However, for special mills, such as the Fritsch Vario-Planetary Mill Pulverisette4, both the speed of the support disk and the planets can be varied freely [34]. Thereby, the trajectoryof the balls within the vial may be controlled at least when the number of balls is low. Ideally the mill-ing can be continuously changed from high-energy mode dominated by high-energy ball–vial impactsto a grinding mode where the balls mainly follow the circumference of the vial [35]. The latter is alsofacilitated by a high number of balls in the vial. High-energy impacts tend to produce high mechanicalpressure in the grain boundaries and in some cases make the high-pressure polymorph of the product.The grinding mode tends to produce more heat by friction and may in some cases lead to thermaldecomposition of the product upon prolonged milling. Heating of the sample may be suppressed byusing short periods of milling intervened by breaks where intrinsic heat produced in the grain bound-aries can be dissipated and the sample can thermally equilibrate. Therefore, not only the total millingtime is important for obtaining the desired compound, but breaks within the period of milling is insome cases crucial, which possibly also reduce agglomeration of the powder on the vial walls and balls[36–42]. Furthermore, the reactant mixture, balls and vial can be placed in a fridge or freezer prior tomilling to lower the temperature further and/or the milling can be intervened by cooling of the vial.Milling at cryogenic conditions, i.e. at liquid nitrogen temperature (77 K), known as cryo-milling, hasproven effective for preparation of some unstable metal hydrides [42].


Within the past two decades, mechanochemistry has expanded widely both within the experimen-tal methods and techniques but also within the variety of materials that can be prepared, e.g. binaryand ternary metallic hydrides [43–47], or complex hydrides such as Mg-based transition metal hy-drides [48–52], alanates [30,53–57], borohydrides [58], amides [59,60], and multi-component systems[61–64] including the broadly studied Reactive Hydride Composites (RHC) [65–68]. These topics arethe focus for further discussion in this review.

2. Tuning of metal microstructures by mechanical milling

The use of mechano-chemical methods for the synthesis and modification of hydrogen storagematerials has generated an enormous amount of reports. In the last 10 years about a thousand papershave been published on the use of ball milling and mechanical alloying for this specific application.Therefore, this review focus on general aspects by discussion selected details and this section focuson the use of mechanochemical methods to tune the microstructure of metal hydride systems in orderto improve their hydrogenation properties.

2.1. BCC alloys

A body-centered cubic (BCC) structure is a coarse packing structure and has more interstitial spacethan face-centered cubic (FCC) and hexagonal close-packed (HCP) structures [69]. Thus, BCC alloys aremore attractive candidates to be explored as possible interstitial hydrogen storage materials. Usually,BCC alloys are synthesized by arc melting or induction melting. However, for some alloys the desiredcomposition is difficult to obtain by using these techniques because the constituting elements mayhave quite different melting temperatures. With mechanical alloying there is in principle no limitationon the nature and number of the raw elements used. For hydrogen storage applications one could dis-tinguish two broad classes of BCC alloys: Ti-based and Mg-based. Each of these classes is discussedbelow.

2.2. Ti-based

BCC alloys of systems Ti–V–Mn and Ti–V–Cr have been intensively studied for hydrogen storage[70–75]. This class of alloys may also catalyse hydrogen release and uptake in magnesium [76].

Fig. 1. X-ray powder diffraction pattern of arc-melted TiV0.9Mn1.1 as a function of milling time (Cu Ka radiation) [77].


Moderate hydrogen capacities as high as 3.6 wt% have been reported for Ti25V40Cr35 alloy, which alsopossess prolific kinetic and thermodynamic properties.

The effect of Severe Plastic Deformation (SPD) on BCC Ti–22Al–27Nb alloy has been investigated byZhang et al. [25,26]. They showed that the first hydrogenation (activation) was much faster for the de-formed alloy compared to the as-quenched sample. The deformed alloy also had faster absorption/desorption kinetics. However, the beneficial effect of deformation was lost after a few hydrogenationcycles. In these studies, SPD was obtained by cold rolling or compression. Cold rolling is a process bywhich a sheet metal or powder is introduced between rollers and then compressed and squeezed. Inthe case of cold rolling one rolling was performed at 10.5% and 80% thickness reduction. Some of the80% rolled specimen were further rolled to 10% thickness reduction in a perpendicular direction withrespect to the first rolling.

Huot et al. have made a systematic study of the effect of milling on TiV0.9Mn1.1 alloy [77]. This com-position is interesting to study because the as-cast alloy is a mixture of BCC and C14 phases. Therefore,it is a good system to test the effect of milling on the crystalline change and the interaction betweenphases. Milling was performed on as-cast alloy as well as on mixtures of elemental powders. Fig. 1shows the effect of milling on as-cast TiV0.9Mn1.1.

The presence of NaCl Bragg peaks is explained by the use of a small amount of this salt as an anti-sticking PCA. It is clear that, with milling time, the C14 phase vanishes and a FCC phase appears. FromRietveld refinement it was found that, for the sample milled 80 h, the crystal structure is a mixture ofcubic (FCC) solid solution phase and a BCC solid solution. The coexistence of FCC and BCC structureswas also observed for the system Fe–Cu and may be due to an enhanced solubility due to the high dis-location density [78]. When milling was performed on the raw elements (Ti, V, and Mn), an identicalresult was obtained, i.e. formation of a nanocrystalline alloy composed of BCC and FCC phases [77].

The BCC alloys need activation, e.g. by cycling hydrogen release and uptake between p(H2) = 5 MPaand vacuum at elevated temperature of 523 K. In Fig. 2 the hydrogen absorption and desorption iso-therm (296 K) for arc-melted TiV0.9Mn1.1 before and after 80 h of milling is presented. The maximumcapacity of the as-melted alloy is 1.9 wt% at 7 MPa which corresponds to an H/M ratio of 0.97. After80 h of milling, the alloy does not absorb hydrogen up to 7 MPa. Because the as-milled materials pres-ent both FCC and BCC phases this means that none of them absorbs hydrogen. In the case of BCC phasethe reason may be reduction of lattice parameters. Iron contamination (even at this low level) mayalso play a role as shown by Santos et al. in the Ti–V–Cr system [79].

Fig. 2. Pressure–composition temperature (PCT) curve, at 313 K, of arc-melted TiV0.9Mn1.1 alloy before and after 80 h of milling[77].

Fig. 3. TEM micrographs of TiV1.6Mn0.4 after arc melting (top), after ball milled for 5 h (middle), and after 150 cold rolls(bottom). Micrographs on the left are bright field images and micrographs on the right are dark field images [82].


Singh et al. studied the effect of milling an arc-melted Ti0.32Cr0.43V0.25 alloy [80]. As they used tung-sten carbide balls, some contamination was observed after long milling time. Ball milling did not affectthe crystal structure of the alloy. Increase of ball milling time resulted in the increase in lattice strainand the decrease in crystallite size, which in turn increased sub-grain boundaries. Contamination frommilling tools and microstructural changes caused an important decrease in the hydrogen storagecapacity [80].

Fig. 4. X-ray powder diffraction patterns of as-cast, milled 5 h and cold rolled 150 times TiV1.6Mn0.4 alloy (Cu Ka radiation) [82].


Amira et al. compared the effect of ball milling and cold rolling for Ti–Cr system [81]. Unlike ballmilling, cold rolling of TiCrx (x = 2, 1.8, 1.5) did not lead to the formation of metastable BCC phase.However, cold rolling was found to be effective to form nanocrystalline C14 Laves phase. Hydrogensorption experiments showed that cold-rolled alloys have similar hydrogen sorption properties toball-milled alloys despite different crystal structures.

The alloy TiV1.6Mn0.4 has been recently investigated by Couillaud et al. [82]. The effect of extendedcold rolling as well as high energy ball milling was a reduction of crystalline size and lattice parameterbut no change in the crystal structure. Fig. 3 shows TEM micrographs of TiV1.6Mn0.4 alloy in the as-cast, milled, and cold-rolled states.

The dark field image shows that all samples are nanocrystalline. The bright field image of the cold-rolled sample clearly shows the pile-up of dislocations. The dark field image shows that, contrary tothe as-cast and ball-milled samples, the crystallites tend to be aligned along dislocations. Fig. 4 showsthe X-ray diffraction patterns of as-cast, milled 5 h and cold rolled 150 times TiV1.6Mn0.4 alloy.

From these X-ray powder diffraction patterns it was determined that the crystallite sizes of as-cast,milled, and cold-rolled samples are respectively 17, 11, and 13 nm [82]. Apart from peak broadeningdue to the reduction of crystallite size, the pattern of the ball-milled sample has the same relativeintensities and lattice parameter as the as-cast sample. For the cold-rolled sample, the lattice param-eter is also the same as the as-cast sample but there is a very strong texture along (200), which is acommon feature of cold-rolled samples. Neither the ball-milled sample nor the cold-rolled samplesabsorb hydrogen even after 10 cycles of hydrogen pressurization (10 MPa) and vacuum at 423 K.The reason for this significant loss of hydrogen capacity is still not clear.

2.3. Mg-based BCC

Recently, Mg-based BCC alloys have been explored in order to achieve higher gravimetric hydrogenstorage capacity. In particular, Akiba’s group has made an extensive study of the synthesis of Mg–Ti[83,84], Mg–Co [85–87], and Mg–Ni [88,89] BCC alloys by means of ball milling. In this review we willlimit our discussion to the Mg–Ti system.

Binary Mg–Ti alloys are being intensively investigated for various applications such as: negativeelectrodes for Ni–MH batteries [90,91], H2 sources for fuel cells [84,92], switchable mirrors for smartsolar collectors [93,94], and optical hydrogen detectors [95]. In the Mg–Ti phase diagram equilibriumsolid solubility of each metal in each other is less than 2 at.% and no intermetallic compound is found.Therefore, non-conventional synthesis methods based on melting or sintering can be used. Metastablesingle-phase Mg–Ti thin films have been successfully synthesized over a large compositional range by

Fig. 5. The powder X-ray diffraction pattern of a magnesium titanium Mg50Ti50 mixture milled for 150 h (Cu Ka radiation) [84].


means of electron-beam and magnetron co-sputter deposition techniques [90,93,94,96,97]. However,these techniques could not be scaled-up to industrial level and other methods have to be investigated.Mechanical alloying has demonstrated its high efficiency for producing metastable Mg–Ti alloys start-ing from elemental Mg and Ti powders [83,91,98–101].

The synthesis of Mg–Ti BCC alloys by mechanical alloying has been extensively studied by Asanoet al. [83,84,92,100,101]. Although both Mg and Ti have a hexagonal closed packed (HCP) structure,during milling of a Mg and Ti mixture they react differently. In the case of magnesium, the deforma-tion is mainly by basal plane slip {0001}h�12�10i while for titanium twinning deformation is moreimportant [100]. In one investigation, Asano et al. had the idea of adding lithium to magnesium in or-der to reduce the yield stress of magnesium and also to decrease its lattice parameter [100]. Theyfound that by adding Li to Mg, the deformation of Mg was easier and the Ti crystallite size was re-duced. This led to a decrease of synthesis time for BCC phase formation.

In a subsequent study, they first synthesized a BCC Mg50Ti50 alloy by ball milling a mixture of50Mg + 50Ti in a Fritsch P5 planetary ball mill for 150 h at a rotation speed of 200 rpm. Fig. 5 confirmsthat a BCC phase was obtained, and from the peaks width a crystallite size of 3 nm was determined[84].

A full hydrogenation at 423 K under 8 MPa of hydrogen and for 122 h resulted in the formation ofMg42Ti58H177 FCC hydride phase and some MgH2.

By controlling milling conditions and Mg:Ti ratio, Asano et al. have also shown that BCC, FCC or HCPphase could be obtained in the Mg–Ti system [83]. In the case of HCP phase it is formed by solution ofTi into Mg while the BCC phase is produced by solution of Mg into Ti and the FCC phase is stabilized byintroduction of stacking faults in Mg and Ti which have a HCP structure [83,84]. If MgH2 is used in-stead of Mg as the starting material then, after ball milling a 50MgH2 + 50Ti mixture, the resultingcompound is FCC Mg33Ti50H94 plus some MgH2 [92]. The importance of mechanical effect during mill-ing is discussed in ref [101]. It shows that during ball milling of Mg and Ti powders in molar ratio of1:1, plate-like particles first stuck on the surface of the milling pot and balls. After these plate-like par-ticles fell off from the surface of the milling pot and balls, spherical particles, in which concentric lay-ers of Mg and Ti are disposed, are formed. These particles have an average diameter of 1 mm. Thesespherical particles are then crushed into spherical particles with a diameter of around 10 lm by intro-duction of cracks along the boundaries between Mg and Ti layers. Finally, the Mg50Ti50 BCC phase witha lattice parameter of 0.342(1) nm and a grain size of 3 nm is formed. During milling, Ti acts as anabrasive for Mg which had stuck on the surface of the milling pot and balls [101].

Recently, Çakmak et al. showed that mechanical milling of Mg–10 vol% Ti yields large Mg agglom-erate, 90–100 lm, with embedded Ti fragments of about 1 lm uniformly distributed within theagglomerates [102]. These Mg agglomerates are made of coherently diffracting volumes (crystallites)of small size. Crystallite size, as determined with X-ray diffraction analysis, can be as small as 26 nmafter 30 h of milling.


In an investigation of high-energy milling of 50Mg–50Ti mixture, Maweja et al. observed twinningin Ti-rich crystallites at intermediate milling time [103]. They attributed the twinning to the deforma-tion of Ti particles. But they also pointed out that in the Mg–Ti system it might also indicate a strain-induced martensitic transformation of the metastable x-FCC into BCC. The crystallite boundariesacted as preferential sites for the heterogeneous nucleation of the twins and for the formation of solidsolution by release of the lattice strain energy [103].

For electrochemical applications, mechanical alloyed Mg–Ti materials must be activated by addingof few at.% of Pd. Rousselot et al. have shown that if a 50Mg–50Ti mixture is pre-milled before addingPd then the alloying of Pd with pre-milled Mg50Ti50 occurs very rapidly (few minutes) and is completeafter 5 h of milling [104]. They also found that the crystalline structure of the Mg50Ti50 alloy (BCC andHCP Mg–Ti phase mixture) does not change significantly with the addition of Pd.

2.4. Amorphization

In amorphization under mechanical driving forces, two categories of alloys could be defined [105].The first category consists of intermetallic compounds induced to undergo polymorphic crystal-to-amorphous transformation by deformation. In this process the introduction of defects by milling in-creases the free energy of the equilibrium alloy such that it goes above the free energy of the amor-phous state. Thus, the amorphous phase becomes the lowest free energy state and the alloybecomes amorphous. The second category of amorphous alloys contains those formed by intermixingof individual elements that have a negative heat of mixing. In this case, deformation plays the role ofenhancing such energy-lowering reactions through deformation-enhanced interdiffusion [105].

From a systematic study of Mg–Ni system, Rojas et al. proposed the sequence of phase transforma-tions during milling leading to amorphization as [106]:

c-Mgþ c-Ni! nc-Mgþ c-Ni! amorphousþ nc-Niþ nc-Mg

! amorphousþ nc-Niþ nc-Mg2Ni! nc-Mg2Ni

where c and nc denotes crystalline and nanocrystalline state, respectively. The first step shows the factthat grain refinement in nickel is slower than in magnesium. It has been demonstrated that the grainsize attainable by milling depends on the crystal structure of the material being milled [107]. Usually,BCC materials tend to reach the smallest sizes, HCP materials somewhat larger grain sizes, and FCCmaterials tend to produce the largest grain sizes. Since the crystal structure of Mg is HCP and thatof Ni is FCC, such a difference in the grain size after milling is expected.

For Mg–Ni system, amorphous phase could be prepared by ball milling in less than 10 h [108].According to Varin et al., the presence of hard MgNi2 phase helps to reduce crystallite size of Mg2Niphase and thus facilitates amorphization [109,110].

For some compositions and milling parameters, a crystallization–amorphous–crystallization phe-nomenon could appear. One example of this was given by El-Eskandarany et al. for Co75Ti25 [111].A solid-state reaction took place during milling elemental Co and Ti powders and an amorphous phaseof Co75Ti25 was formed after 3 h. They showed that this amorphous phase crystallized into an orderedFCC-Co3Ti phase upon heating to 880 K. Further milling to 24 h also leads to crystallization and theformed phase was a metastable BCC-Co3Ti nanocrystalline phase. They attributed this transformationtaking place in the ball mill to the inability of the formed amorphous phase to withstand the impactand shear forces that are generated by the milling media. When the milling time was further increasedto 100 h, the crystalline phase was subjected to several points and lattice defects that raised the freeenergy from the stable BCC-Co3Ti phase to an amorphous less stable phase. In this case, the crystal-line–amorphous transformation which took place was similar to the mechanical grinding method inwhich the amorphization occurs by relaxing the short-range order without any compositionalchanges. Further milling leads to the formation of crystalline and/or amorphous phases dependingon the milling time. Contamination from milling tools and temperature effect were ruled out as originof this phenomenon [111]. Fig. 6 shows a schematic illustration of this crystallization–amorphization–crystallization process.

Fig. 6. Schematic illustration of amorphous–crystalline–amorphous cyclic phase transformations that took place during ball-milling elemental powders of Co75Ti25, using a rotation speed of 4.2 s�1 [111].


3. Synthesis of hydrides by mechanically-induced solid/gas reactions

Mechanical milling of metal powders under reactive gas, i.e. Reactive Mechanical Milling (RMM), isbecoming a mature and powerful technique for the synthesis of metallic and complex hydrides. Themechanical treatment induces a chemical reaction between the solids and the gas. The synthesis ofseveral metallic and complex hydrides by RMM is surveyed here. RMM under hydrogen gas allowsfor the synthesis of binary and ternary metal hydrides, Mg-based complex hydrides and alanates.More recently, this technique has been extended to other reactive gases such as diborane and ammo-nia for the synthesis of borohydrides and metal amides, respectively. Some particular phenomenasuch as ultra-fast hydride synthesis, reactive-milling induced amorphization, and multi-step reactionsare reported. The obtained hydrides are typically nanocrystalline materials leading to fast kinetic forhydrogen release and uptake reactions useful for hydrogen storage applications.

Significant progress on the understanding of RMM process has been provided by the in situ mon-itoring of the hydrogenation reaction during milling. In 2000, Dunlap et al. connected a ball-millingdevice to a large hydrogen reservoir by means of a rubber tube [112]. They could follow the hydrogenuptake as a function of milling time for several early transition metals (Ti, Zr, Hf, V, Nb and Ta). Exper-iments were conducted near atmospheric pressure (p(H2) � 0.1 MPa). A similar method was used byBellosta et al. to monitor hydrogen release during ball-milling of sodium tetra-alanate with TiCl3 addi-tive [113]. Hydrogen release occurs due to titanium reduction to the zero-valent state on milling. Afurther improvement was reached by using telemetric systems instead of mechanical connectionsto in situ register both hydrogen pressure and vial temperature during milling [30,31]. The sensorswere mounted on the lid of a stainless steel vial which was able to withstand high pressure


(�10 MPa). Under these conditions, the one-step direct synthesis of Ti-doped NaAlH4 using NaH, Aland TiCl3 as starting powders could be monitored.

RMM is generally accomplished in tight stainless steel vials equipped with a connection valve forvacuuming and hydrogen filling. The vial is then placed in a milling device to promote the mechano-chemical reaction leading to the hydride formation. For practical reasons regarding p–T gauges attach-ment to vials, most of current milling devices used for RMM are planetary ball mills. Thus, thispreparation technique is widely named as reactive ball milling. Nonetheless, also shaker, attritor,and vibration mills have been used successfully [51,112,114,115].

Today, compound synthesis can be anticipated from the in situ monitoring of the hydrogenationreaction and subsequently verified by ex situ crystallographic and chemical analyses. Furthermore,if thermodynamic parameters such as vial volume and gas pressure and temperature are accuratelyknown, the quantity of absorbed hydrogen as a function of time can be reliably obtained. Zhanget al. have recently shown that hydrogen uptake can be determined with an accuracy of 95% [116].In situ monitoring of changes in gas pressure is certainly a powerful tool for the study of hydride for-mation kinetics and reaction mechanisms on reactive milling.

3.1. Binary hydrides

Though thermodynamically favourable, the formation of AHx binary hydrides by solid–gas reactionbetween hydrogen gas and a metal (A, a metal with strong affinity for hydrogen, here stands for eitheralkaline earths (Mg) or early transition metals such as Ti and V) is very often hindered by kinetic bar-riers related to the presence of native oxide layers at the metal surface. Then, severe treatments athigh temperature (typically above 700 K) and high pressure (several MPa) are needed in conventionalgas-phase hydrogenation for activation. In the course of these treatments, oxygen at the surface mightreact with the bulk material leading to additional impurities. Such surface limitation can be overcomeby RMM of pure metals in hydrogen atmosphere that allows achieving faster synthesis reactions undermore moderate conditions. Synthesis conditions by RMM under hydrogen gas of representative binaryhydrides are summarized in Table 1.

Table 1Representative binary and ternary metal hydrides synthesised by RMM under hydrogen gas. The employed device, reactants, initialhydrogen pressure, p(H2), total milling time (tmt), milling speed (ms), ball-to-powder weight ratio (BTPWR) and ball diameter (Bd)are given.

Compound Device Reactants p(H2) (MPa) tmt (h) ms (rpm) BTPWR Bd (mm) Ref.

MgH2 Planetary Mg 0.34 25 12 [117]MgH2 Mg + graphite 0.4 1 10:1 [118]MgH2 Fritsch P6a Mg 1–9 8 500 10:1 10 [31]MgH2 Fritsch P4a Mg 8 2 400 60:1 12 [64]TiH1.9 Planetary Ti 0.34 5.5 12 [117]TiH2 Fritsch P4a Ti 8 0.16 400 60:1 12 [64]VHx Fritsch P5 V 1 0.17 400 30:1 7 [47]ZrNiH3 Fritsch P5 ZrNi 2 3 30:1 10 [27]ZrH2 + Ni Fritsch P5 Zr + Ni 2 3 30:1 10 [27]ZrH2 + NiZryHx Fritsch P5 Zr + Ni 2 100 30:1 10 [27]b-ZrNiH Fritsch P7 ZrNi 0.1 0.08 400 30:1 7 [28]c-ZrNiH3 Fritsch P7 ZrNi 1 0.08 400 30:1 7 [28]TiNiH3 Rod-mill Ti + Ni 0.1 200 30:1 10 [119]TiH2 + Ni Ti + Ni 1.1 40 250 10:1 10 [120]TiH2 + Fe Spex 8000 TiFe 0.5 7 8:1 6, 12 [114]a-LaNi5H0.15 Fritsch P7 LaNi5 1 0.08 400 30:1 7 [121]amph-LaNi5�yHx Fritsch P7 LaNi5 1 10 400 30:1 7 [121]BCC TiVH0.9 Fritsch P5 TiV or Ti + V 0.2 100 20:1 10 [43]TiVH2.8 Fritsch P5 TiV or Ti + V 0.4 100 20:1 10 [43]FCC TiVH4.7 Fritsch P5 TiV or Ti + V 1 100 20:1 10 [43]Ti0.20V0.78Fe0.02H2 Fritsch P4* Ti0.20V0.78Fe0.02 8 0.17 400 100:1 12 [122]

a Pressure and temperature measured in situ in the Evico-magnetics vial.


3.1.1. Magnesium hydrideMagnesium hydride is classically prepared by reaction with hydrogen gas and Mg powder at tem-

peratures around 700 K and hydrogen pressures in the range 7–8 MPa for several hours. However,even under these conditions, the presence of Mg is often detected by XRD as the reaction is notcompleted [123]. Indeed, a shell of magnesium hydride is reported to form at the surface of microm-eter-sized magnesium grains, blocking further hydrogenation of the remaining metal core [124,125].Consequently, the hydrogenation rate of bulk magnesium is slow.

First attempts to form magnesium hydride by RMM were conducted by Chen and Williams usingp(H2) = 0.34 MPa and a vertical planetary mill [117]. Complete formation of MgH2 hydride is reportedto occur after long milling time (25 h). Later, similar experiments under 0.5–1 MPa of hydrogen pres-sure were conducted by different groups [46,126,127]. Neither of them could attain hydride formationabove 50 wt%, which was attributed to kinetic effects. This was finally overcome by performing RMMexperiments at high temperature (573 K) with the addition of graphite to obtain complete hydrogena-tion within 1 h [118]. Graphite could act as a PCA to reduce particle agglomeration by cold-welding[33].

Doppiu et al. have however shown that fast MgH2 formation can also be achieved near room tem-perature using high-pressure reactive ball milling [31]. The hydrogenation reaction could be moni-tored by in situ measurements of both pressure, p, and temperature, T, inside the vial during millingby using on-board sensors and radio transmitted data. Syntheses were done with pure Mg powdersball milled with a ball-to-powder weight ratio of 10:1 at 500 rpm and hydrogen pressures of 1, 4and 9 MPa. From the data collection, it was first observed that the temperature of the vial increasesup to 318 K mainly due to mechanical action. Mg absorbs hydrogen in less than 8 h for pressures largerthan 4 MPa. Reaction rate was significantly slower for lower pressure (1 MPa). Moreover, a nucleationtime, strongly dependant of the pressure is also reported; almost undetectable at 9 MPa, it reachesmore than 2 h at 1 MPa. Further investigations by XRD at different milling times show the formationof the metastable orthorhombic c-phase along with the tetragonal b-MgH2 one. The c-phase can alsobe achieved by ball milling of magnesium hydride [128]. With increasing milling time, the crystallitesize decreases to finally stabilizing at 10 nm. Same amounts of hydride phases (>95 wt% for c + b) andidentical crystallite sizes are obtained after 18 h of milling whatever the initial pressure though therate of formation and the size reduction were faster for higher pressures. This was interpreted onthe basis of two different factors. Higher pressures promote a more rapid formation of the hydride thatis in turn known to exhibit higher plastic deformation. Then, for higher amounts of MgH2, the mechan-ical action is more effective than for ductile Mg. However, it is worth noting that at long milling times,all samples reach the same chemical and microstructural states.

Very similar results are also reported by Doppiu et al. who performed reactive milling of an ele-mental Mg87Ni10Al3 powder mixture under hydrogen atmosphere [129]. Milling induces the synthesisof nanocrystalline MgH2 at the first stage followed by the formation Mg2NiH4 when a high degree ofconversion of Mg in the hydride form was reached. A minimum value for the crystallite size of 8 nm

0 60 120 180

5.0

5.5

6.0

6.5

7.0

7.5

8.0

8.5

9.0

300

304

308

312

316

320

P (M

Pa)

t (min)

T (

K)

(A)0 60 120 180

0.0

0.5

1.0

1.5

2.0

H/M

g

t (min)

(B)

Fig. 7. Evolution of the pressure and temperature (A) during RMM of magnesium in hydrogen gas and the calculated H/Mg ratioin the solid state (B) as a function of time [64].


was obtained. Small differences in the hydride stability were observed at different milling times. Inspite of the oxidation of the sample, fast absorption–desorption kinetics were obtained.

A typical example of the evolution of the pressure, the temperature and the H concentration as afunction of time during RMM of Mg is shown in Fig. 7. The starting material was coarse magnesiumpowder to reduce the amount of MgO that may develop at the grain surface of powder material.The milling vial was loaded with p(H2) = 8 MPa and operated at 400 rpm. After initial heating dueto ball friction, the pressure drop related to hydride formation is observed and the reaction is com-pleted after 2 h. The final H/M value reaches 1.9, a value 5% smaller than expected for MgH2. XRD anal-ysis shows that the final product is made of 76 wt% of b-MgH2, 21 wt% of c-MgH2, and 3 wt% of MgO.The mean crystallite size for the hydride phases is close to 6 nm, in good agreement with previousresults.

3.1.2. Titanium hydrideThe formation of titanium hydride by RMM with composition TiHl.9 was first reported by Chen and

Williams in 1995 using a hydrogen filled container p(H2) = 0.34 MPa for 67 h [117]. The hydrogenationreaction was completed in 5.5 h and the TiH1.9 compound was stable during prolonged milling, withonly a reduction of particle size being observed. Very similar results have been published by differentgroups [44–46]. Dunlap’s group has extended this method to other early transition metals such as Zr,Hf, Ta, Nb and V [112].

Short reaction time was explained in terms of clean surface generation and severe reduction of dif-fusion path. Titanium powder is initially passivated by the presence of surface oxides. Upon milling,fresh and highly reactive surfaces are created promoting the formation of near-surface hydride precip-itates. This causes hydrogen embrittlement of the metal, enhances its pulverisation and results in ashorter diffusion path for hydrogen absorption.

The formation of hydride TiH2 during RMM is shown in Fig. 8 [64]. Contrary to magnesium, nonucleation time is observed and the hydride formation proceeds readily and is completed after10 min. XRD analysis confirms the formation of TiH2 though small contamination with iron is ob-served, most probably due to the stainless steel vial abrasion during milling. The mean particle sizedetermined from the diffraction peak widths is around 7 nm.

3.1.3. Vanadium hydrideOrimo et al. reported on the preparation of nanostructured VHx prepared by mechanical milling un-

der H2 atmosphere [47]. Formation of the b2 phase was observed after 5 min of milling at room tem-perature whereas conventional gas-phase hydrogenation would need activation treatments under600 K and 3 MPa. The grain size of the b2 phase decreases from 80 nm at 5 min milling time to10 nm after 60 min. Additional milling time (up to 300 min) does not lead to further decrease ofthe grain size nor the formation of an amorphous state as the b2 phase remains in crystalline state.From the relationship between hydrogen concentration and unit cell, the hydrogen concentration

6.0

6.5

7.0

7.5

8.0

290

295

300

305

310

315

320

P (M

Pa)

t (min)

T (

K)

(A)0 60 120 180 240 0 60 120 180 240

0.0

0.5

1.0

1.5

2.0

H/T

i

t (min)

(B)

Fig. 8. Evolution of the pressure, the temperature (A) and the H concentration (B) as a function of time during RMM of titaniumin hydrogen gas.


was determined as a function of the grain size. It decreases from 0.82 H/M for 80 nm down to 0.72 H/M for 10 nm indicating a modification of the b2–c phase boundary in the V–H system for nanometergrain sizes. Lower concentration, nearly independent of the grain size and higher diffusivity of hydro-gen are also reported in the intergrain domain.

3.2. Ternary hydrides

Ternary metal hydrides of general composition ABHx can be easily synthesized by RMM providingthat the hydrides are stable under the pressure and temperature milling conditions. In this formula-tion, A stands for a metal with strong affinity for hydrogen (an early transition or rare-earth metal) andB stands for a metal with weak hydrogen affinity (a late transition metal). Well-known hydrogen stor-age intermetallic compounds such as ZrNi, TiFe, TiNi and LaNi5 have been used in RMM experiments(see Table 1). As a general behaviour, two effects are observed on prolonged milling: formation ofamorphous ABHx hydrides and compound disproportionation into AHx + B species. Compound amor-phization is driven by the large negative heat of mixing between A and B elements while its dispro-portionation is favoured by the different affinity between both elements for hydrogen.

3.2.1. ZrNi hydrideRMM experiments in the Zr–Ni system have been first reported by Aoki et al. [27]. They performed

RMM (p(H2) = 2 MPa) of both arc-melted ZrNi alloys and equiatomic Zr and Ni powder mixtures. In thefirst case, ZrNiH3 hydride is formed after 3 h of milling. On prolonged milling (over 100 h), the crys-tallite size of the hydride decreases without apparent amorphization. The lack of amorphous phasesis attributed to the difficulty to introduce defects in ZrNiH3 hydride because of its brittleness. ForRMM of elemental powders, a mixture of ZrH2 and elemental Ni is formed at short milling time(<3 h). Further reaction over 100 h leads to the coexistence of ZrH2 and amorphous Zr-poor NiZr1�yHx

phase.Orimo et al. have studied the effect of hydrogen pressure during RMM of ZrNi compound within the

range 0–1 MPa [28]. b-ZrNiH and c-ZrNiH3 hydrides are formed within 5 min of milling. Phase abun-dance depends on hydrogen pressure. Formation of the most stable b-hydride is observed at 0.1 MPa,whereas that of the less stable c-hydride occurs at 1 MPa. At intermediate pressures, 0.3 MPa, bothphases are detected. As observed by Aoki, prolonged milling over 80 h results in the formation ofZrH2 and amorphous Zr-poor NiZr1�yHx phase. Such decomposition reaction seems to be delayed withincreasing hydrogen pressure.

3.2.2. TiNi hydrideRMM experiments (p(H2) = 0.1 MPa) on equiatomic Ti and Ni powder mixture have been conducted

in a rod-milling device [119]. Within the first 3 h of milling metallic Ti transforms to TiH2 and metallicFCC Ni remains unreacted. Further milling up to 200 h leads to the gradual formation of a nanocrys-talline (10 nm) single-phase FCC compound. The compound is described as an FCC TiNiH3 solid solu-tion, though details on the hydrogen content determination are not provided. This result is ratherstriking since TiNi compound only absorbs 1.4 H/f.u. under normal conditions of pressure and temper-ature [130]. In fact, later experiments at higher hydrogen pressure (1.1 MPa) failed to get the solidsolution TiNiH3 phase [120]. Instead, formation of poorly crystallized TiH2 and Ni phases on millingfor 40 h is reported.

3.2.3. TiFe hydrideChiang et al. have performed RMM experiments (p(H2) = 0.5 MPa) in TiFe compound [114]. In situ

manometric measurements reveal that a total hydrogen uptake of 1.6 H/f.u. occurs within 7 h of mill-ing forming TiFeH1.6. Ex situ XRD analysis reveals that the ternary hydride decomposes to TiH2 and Fe.

3.2.4. LaNi5 hydrideFujii et al. have performed RMM experiments (p(H2) = 1 MPa) on LaNi5 alloy and observe formation

of solid solution a-LaNi5H0.15 within the first 5 min of milling [121]. Formation of hydride b-LaNi5H6

phase is not detected, which may indicate that the absorption plateau pressure of this hydride is above


1 MPa at the temperature attained on milling. At longer milling times, 5 min < t < 3 h, the a-phasecoexist with an amorphous phase. This phase forms faster in the presence of hydrogen than in ball-milling experiments performed under inert gas. The a-phase decomposes into Ni and amorphousNi-poor LaNi5�yHx phase upon prolonged milling, 3 h < t < 10 h. Subsequent thermodynamic measure-ments show a significant reduction of total and reversible hydrogen storage capacity for long-timemilled as compared to pristine LaNi5 compound. This is attributed to a lower hydrogenation capacityof both poorly crystallized inter-grain and amorphous regions as compared to the microcrystallinestate. This concurs with the facts that nanocrystalline systems exhibit lower capacity than microcrys-talline ones and that hydrogen binding energies expands over a wide energy range in amorphous sys-tems [131,132].

3.2.5. TiV hydrideOne should notice that the Ti–V system differs from previous ones as concerns the affinity of con-

stituting elements towards hydrogen. Both elements are A-type and exhibit comparable affinity forhydrogen which, in principle, precludes alloy disproportionation by hydrogenation. Furthermore, thissystem exhibits small heat of mixing so that alloy amorphization is expected to be difficult.

RMM (p(H2) = 0.2,0.4 and 2 MPa) of either equiatomic Ti and V powder mixtures or BCC TiV alloyhave been conducted by Aoki et al. [43]. At long milling time (100 h), phase constitution of milledproducts does not depend on the nature of the initial powder. Hydrogen pressure plays, however, amajor role. At low (0.2 MPa) and high (1 MPa) pressures, BCC TiVH0.9 solid solution and FCC TiVH4.7

hydride are formed, respectively. The hydrogen content of FCC TiVH4.7 hydride is probably overesti-mated since maximum hydrogen uptake of both Ti and V is 2 H/M. Nevertheless, both BCC and FCChydrogenated phases are crystalline. In contrast, at intermediate pressure (0.4 MPa), amorphousTiVH2.8 phase is obtained. The formation mechanism of this phase depends on the initial reactants.For Ti + V powder mixture, the amorphous phase is formed by reaction between TiH2 and V, whereasfor the arc-melted alloy it results from gradual amorphization of BCC TiVH2.8 phase on milling. Strik-ingly, for both cases, the amorphization reaction occurs without changing the hydrogen content.

RMM experiments on a BCC Ti–V–Fe alloy of composition Ti0.20V0.78Fe0.02 have been carried out in adevice equipped with p and T sensors at rotation speed of 400 rpm and p(H2) = 8 MPa. The hydroge-nation curve is shown in Fig. 9 [122]. The alloy absorbs 2 H/f.u. in only ten minutes indicating the for-mation of a stoichiometric (Ti,V,Fe)H2 hydride. Further milling produces hydrogen desorption from themilled sample. Its hydrogen content decreases to 1.55 H/f.u. after 8 h of milling. XRD diffraction anal-ysis (Fig. 10) shows that the RMM alloy consists of a mixture of FCC VH2-type hydride and amorphous

0 60 120 180 240 300 360 420 4800.0

0.5

1.0

1.5

2.0

H/f

.u.

t (min)

Fig. 9. Time-evolution of the H concentration in BCC Ti0.20V0.78Fe0.02 alloy during RMM in hydrogen gas.

30 40 50 60 70 80

0

200

400

600

800

I (c

ount

s)

2θ

amorphous phase

Fig. 10. Rietveld analysis of Ti0.20V0.78Fe0.02 alloy after RMM for 480 min. Observed (dots), calculated (top line) and differencecurves (bottom line) are shown. Vertical bars (|) correspond to Bragg positions (Cu K a1,2) for FCC VH2-type hydride. Large dotsstand for the contribution of an amorphous phase to the calculated pattern.


phase. The amorphous phase formation accounts for the spontaneous hydrogen desorption on millingsince it stores less hydrogen.

3.3. Mg-based complex hydrides

Mg-based Mg2THx ternary hydrides (T = Fe, Co and Ni transition metals) are attractive hydrogenstorage materials due to their high specific (5.5, 4.5 and 3.6 wt%) and volumetric (150, 125 and97 g/L) hydrogen contents for Mg2FeH6, Mg2CoH5 and Mg2NiH4, respectively [138]. The synthesis ofthese hydrides is problematic due to the great difference in vapour pressure and melting point be-tween Mg and T and the lack of stable Mg2Fe and Mg2Co intermetallic compounds in their respectivebinary phase diagrams. From these facts, synthesis of Mg2THx ternary hydrides was classicallyachieved by sintering methods from elemental powder mixtures. Temperatures and hydrogen pres-sures as high as 750 K and 9 MPa, respectively, and reaction time of several days are required [139].Synthesis conditions of Mg2THx ternary hydrides by RMM under hydrogen gas are summarized inTable 2.

3.3.1. Mg2Fe hydrideThe synthesis of Mg2FeH6 by RMM (p(H2) = 1 MPa for 20 h) of Mg and Fe powder was first at-

tempted in 1997 [133]. Mg2FeH6 hydride was not obtained but Mg powder got hydrogenated to formintimate MgH2–Fe mixture. The failure to form the ternary hydride could be due to milling conditionsnot being sufficiently efficient (ball-to-powder weight ratio of 4:1). It was later discovered that thedesired hydride can be obtained in two different ways. The first simply consists in a sintering treat-ment of the reactive milled product MgH2–Fe for one day at 625 K under 5 MPa of hydrogen. The sec-ond, more complex, was reported in a subsequent paper [134]. Mechanical milling of MgH2 and Fepowders in molar ratio 2:1 under argon atmosphere was performed for 60 h in a high-energetic shakermill with ball-to-powder weight ratio of 10:1. The mechanical energy provided under these millingconditions was high enough to promote Mg2FeH6 formation without subsequent sintering. Muchprobably, the following solid-state reaction takes place:

3MgH2ðsÞ þ FeðsÞ !Mg2FeH6ðsÞ þMgðsÞ ð1Þ

The formation of the ternary compound is driven by the fact that the Mg2FeH6 phase is more stablethan MgH2 [140]. In situ SR-PXD patterns measured for a ball milled sample of MgH2–Fe (2:1) revealformation of Mg2FeH6 at �673 K at p(H2) = 10 MPa [141].

Table 2Mg-based complex hydrides synthesised by RMM under hydrogen gas. The employed device, reactants, initial hydrogen pressure,p(H2), total milling time (tmt), milling speed (ms), ball-to-powder weight ratio (BTPWR), ball diameter (Bd), reaction yield andformed side products are given.

Compound Device Reactants p(H2)(MPa)

tmt(h)

ms(rpm)

BTPWR Bd(mm)

Yield(wt%)

Sideproducts

Ref.

Mg2FeH6 Fritsch P5 2Mg + Fe 1 20 325 4:1 10 0 MgH2, Fe [133]Mg2FeH6 Spex 8000 2MgH2 + Fe 0.1 60 10:1 10 0 Mg, MgO,

Fe[134]

Mg2FeH6 Uni-Ball-Mill II 2Mg + Fe 0.5 60 44:1 28 MgO, Fe [49]Mg2FeH6 Szegvari attritor 2Mg + Fe 1 20 400 20:1 6 63 Fe [115]Mg2FeH6 Retsch 2000

vibrating mill2Mg + Fe 0.3 8 32 s�1 16:1 12 90b Fe [51]

Mg2FeH6 Fritsch P4a 2Mg + Fe 7.5 12 400 60:1 12 77 MgO, Fe [116]Mg2CoH5 Kurimoto

planetary mill2MgH2 + Co 0.1 10 700 8:1 4 100 [48]

Mg2CoH5 Uni-Ball-Mill II 2Mg + Co 0.5 90 44:1 50 Co [135]Mg2CoH5 Fritsch P4a 2Mg + Co 7.5 12 400 60:1 12 81 MgO [116]Mg2NiH4 Fritsch P5 2Mg + Ni 0.5 22 325 4:1 10 0 Mg,

MgH2, Ni[29]

Mg2NiH1.8 Fritsch P7 Mg2Ni 1 80 400 30:1 7 100 [136]Mg2NiH4 Kurimoto

planetary millMg2Ni 1 10 885 5:1 7 70b amph-

MgNi[137]

Mg2NiH4 Retsch 2000vibrating mill

2Mg + Ni 0.3 16 32 s�1 16:1 12 100 [51]

Mg2NiH4 Fritsch P4a 2Mg + Ni 7.5 12 400 60:1 12 79 MgO [116]Mg2(FeH6)0.5(CoH5)0.5 Fritsch P6a 4Mg + Fe + Co 5 20 400 40:1 10 95b FeCo [52]

a Pressure and temperature measured in situ in the Evico-magnetics vial.b Estimated values.


In 2002, direct though incomplete synthesis of Mg2FeH6 by RMM of elemental powders was simul-taneously reported [49,115]. Gennari et al. used a Uni-Ball-Mill II device under 0.5 MPa with hydrogenrefilling every 5 h to maintain constant hydrogen pressure in the vial [49]. Mg2FeH6 formation with ayield of 28 wt% was achieved after 60 h of milling. The synthesis was reported to occur in two steps.MgH2 is formed during the first 40 h by mechanically activated solid–gas reaction followed by the so-lid-state reaction between MgH2 and Fe at longer milling times. Raman et al. [115] used a Szegvariattritor device under 1 MPa of hydrogen. The ternary hydride started forming after 14 h of millingand a maximum yield of 63 wt% was achieved at 20 h of milling as determined from XRD analysis.Reaction yield is probably overestimated since a high quantity of Fe (37 wt%) was identified as the un-ique secondary phase, which is not possible from mass-balance considerations. In fact, significantresiduals in the Rietveld analysis likely related to MgO phase can be observed, which explains thepresence of unreacted Fe (similar effects have been later observed [50]). Crystallite sizes of 12 and18 nm are reported for Mg2FeH6 and Fe phases, respectively. Formation on MgH2 as intermediatephase was not detected. Prolonged milling to 30 h is reported to lead to amorphization of the ternaryhydride.

Milling under hydrogen of 2MgH2 + Fe and 2Mg + Fe powder mixtures have also been compared[142]. It was found that a faster reaction and higher yield is achieved for elemental powders as com-pared to 2MgH2 + Fe. The differences were attributed to the dissimilar mechanical properties andmicrostructures of the mixtures. The 2Mg + Fe mixture behaves as a ductile–ductile pair that resultsin a higher contact surface between Mg and Fe, and a better intermixing and size reduction. On thecontrary, the 2MgH2 + Fe mixture performs as a ductile–brittle combination, with less contact area be-tween the reactants and hence lower yield and longer synthesis time.

In 2008, the direct synthesis of Mg2FeH6 by RMM of elemental powders was monitored in situ bymanometric means by Baum et al. [51] RMM experiments were performed in a horizontal vibratingmill operated at 32 s�1 under a hydrogen pressure of 0.3 MPa. In spite of using mild milling conditions(one unique ball and ball-to-powder weight ratio of 16:1), the reaction was completed after only 8 h ofmilling time. The reaction yield is not specified, but judging from the total hydrogen uptake and XRD


data it should be around 90 wt%. This record could be related to the fact that enough hydrogenpressure was kept in the system on milling. The system was refilled with hydrogen when the pressuredecreased below 0.27 MPa. Moreover, from in situ hydrogen uptake curves the authors could infer, inagreement with Gennari et al. a two-step process with formation of MgH2 during the first 2 h of mill-ing followed by the formation of Mg2FeH6 from MgH2 and Fe at longer milling times [49]. These resultshave later been confirmed by Deledda and Hauback by in situ measurements during RMM in an Evico-magnetics vial operated at 5 MPa of hydrogen pressure [52]. The latter authors observed, however,that the first step exceeded the hydrogen capacity of MgH2 and proposed additional hydrogen uptakeat Mg/Fe interfaces. Such additional capacity is doubtful since the authors used the ideal gas law,which is not valid at the imposed pressures, to estimate hydrogen absorption.

The reaction path during RMM synthesis (p(H2) = 7.5 MPa for 12 h) of Mg2FeH6 has been recentlystudied in an Evico-magnetics vial [116]. The evolution of the hydrogen uptake as a function of millingtime is shown in Fig. 11. The result for a similar experiment using only Mg powder is shown in thesame figure for comparison. Hydrogen absorption by 2Mg + Fe powder mixture occurs in two steps.

0

1

2

3

4

5

6

2ndstep

Hyd

roge

n up

take

(H

/f.u

.)

Milling time (min)

2Mg

1st stept = 50 min

3.4 H/f.u.

2Mg+Fe

0 60 120 180 240

Fig. 11. Time-evolution of the H concentration in solid-state during RMM of Mg powder and 2Mg + Fe powder mixture.

Fig. 12. XRD patterns and phase identification of RMM 2Mg + Fe powder mixture after the first (50 min) and the second(720 min) reaction step (Cu Ka radiation).


The first step, with t < 50 min, corresponds to the formation of MgH2 hydride as demonstrated by boththe equivalent amount of absorbed hydrogen in the experiment conducted with elemental Mg and exsitu XRD measurements (Fig. 12). The second step corresponds to the reaction between MgH2 and Fe toform Mg2FeH6. The XRD pattern for 12 h milled product (Fig. 12) shows the presence of three phasesMg2FeH6 (77 wt%), Fe (12 wt%) and MgO (11 wt%) with crystallite sizes of 8, 12 and 4 nm, respectively.In contrast to previous reports no amorphization is observed on prolonged milling [115,143]. MgOcontamination is attributed to undesired surface oxidation of fine Mg powder (36 lm) in glove-boxand accounts for unreacted Fe residual [144]. According to these results, the two-step reaction pathfor Mg2FeH6 synthesis could be described as:

2MgðsÞ þ FeðsÞ þ 3H2ðgÞ ! 2MgH2ðsÞ þ FeðsÞ þH2ðgÞ !Mg2FeH6ðsÞ ð2Þ

It is worth noting that reaction kinetics for the first step, i.e. MgH2 formation, is faster for the 2Mg + Femixture than for pure Mg powder. This striking result may be related either to catalytic effects forhydrogen dissociation at the Fe surface or to nucleation phenomena at Mg/Fe interfaces [145].

The reaction path given by Eq. (2) concurs with recent reports on hydrogen absorption by classicalsolid–gas reaction in nanosized Mg + Fe mixtures [146,147]. Based on DFT calculations, it has beenproposed that the reaction between iron and magnesium hydride may occur through the formationof a (MgFe)H2 solid solution which becomes unstable with increasing Fe content with respect toMg2FeH6 [148].

As we saw in the previous sections, ball milling has been extensively used to synthesize Mg2FeH6.In the case of Severe Plastic Deformation SPD techniques, investigation has only started recently andthe literature is much less abundant. Lima et al. observed substantial improvement in the hydrogensorption kinetics of a Mg–Fe powder mixture processed by high pressure torsion HPT [16]. The authorsnoted that hydrogenation and dehydrogenation of the processed samples did not change the prefer-ential orientation (002) of the Mg phase, i.e. the material retained the microstructure imposed byHPT. In a subsequent investigation, the same authors used a combination of ball milling and extrusionto synthesize Mg2FeH6 [149]. Their results indicate that the iron in the 2Mg–Fe mixture produced abeneficial pinning effect on the Mg grains by hindering grain coarsening even after annealing treat-ments. The desorption kinetics of samples processed by high pressure torsion (HPT) was faster thanthat of extruded samples, probably due to bulk diffusion limitations [149].

3.3.2. Mg2Co hydrideSimilarly to Mg2FeH6, first attempts to produce Mg2CoH5 by RMM (p(H2) = 1 MPa for 20 h) of Mg

and Co powder were unsuccessful. Instead, a mixture of MgH2 and Co phases was obtained [133].Subsequent sintering treatment allowed synthesizing Mg2CoH5 hydride though with a lower yield(26 wt%) than for Mg2FeH6 (65 wt%).

The synthesis of Mg2CoH5 hydride by RMM (p(H2) = 0.1 MPa for 10 h) of MgH2 and Co powders inmolar ration 2:1 under hydrogen pressure was first reported by Chen et al. [48]. Ex situ XRD measure-ments revealed that Mg2CoH5 phase started forming at 1 h milling and became the major phase after10 h milling. Later, the synthesis could be also achieved using Mg and Co powders as reactants under0.5 MPa of hydrogen [135]. The powders were previously milled under Ar atmosphere in the same sys-tem for 200 h. It was supposed that intimate contact and homogeneity between Mg and Co phases isessential to reach hydride formation as occurring for sintering methods [150]. The Mg2CoH5 phasestarts forming after 40 h of milling and a yield of 50 wt% was achieved at 90 h. The formation of anintermediate MgH2 phase occurs from 10 h of milling. The two-step reaction has been also observedby Baum et al. by in situ monitoring of hydrogen uptake during RMM experiments [51].

The reaction path during RMM synthesis (p(H2) = 7.5 MPa for 12 h) of Mg2CoH5 from Fe and Copowders in molar ratio 2:1 has been recently studied in detail [116]. The evolution of the hydrogenuptake as a function of milling time is shown in Fig. 13. Hydrogen absorption occurs in two stepswhich, after the analysis of XRD data (Fig. 14), correspond to the following reactions:

2MgðsÞ þ CoðsÞ þ 5=2H2ðgÞ ! 2MgH2ðsÞ þ CoðsÞ þ 1=2H2ðgÞ !Mg2CoH5ðsÞ ð3Þ

MgH2 hydride is formed as an intermediate phase for t < 50 min. The reaction is again faster as com-pared to Mg milled alone. The second step corresponds to the reaction between MgH2 and Co to form

Fig. 14. XRD patterns and phase identification of RMM 2Mg + Co powder mixture after the first (50 min) and the second(720 min) reaction step (Cu Ka radiation).

0

1

2

3

4

5

3.4 H/f.u. 2ndstep

2Mg+Co

Hyd

roge

n up

take

(H

/f.u

.)

2Mg

1st stept = 50 min

Milling time (min)

0 60 120 180 240

Fig. 13. Time-evolution of the H concentration in solid-state during RMM of Mg powder and 2Mg + Co powder mixture.


Mg2CoH5. The hydrogen uptake corresponding to this reaction is lower but faster than for Mg2FeH6

formation (Fig. 11). The XRD pattern for 12 h milled product (Fig. 14) shows the formation of nano-crystalline (8 nm) Mg2CoH5 phase without significant amorphous contribution.

3.3.3. Mg2Ni hydrideSince Mg2Ni compound exists as stable phase in the binary Mg–Ni phase diagram, the synthesis of

Mg2NiH4 ternary hydride by RMM can be attempted either using 2Mg + Ni elemental mixture orMg2Ni powders as initial reactants.

This synthesis was first tried by RMM (p(H2) = 0.5 MPa for 22 h) of 2Mg + Ni powders [29]. SomeMgH2 phase was formed from 2 h of milling but no ternary hydride could be detected. Either millingenergy (ball-to-powder weight ratio was 4:1 and milling speed 325 rpm) or hydrogen supply wasinsufficient to promote hydride formation.

Orimo et al. later investigated hydrogen absorption in Mg2Ni during reactive milling (p(H2) = 1 MPafor 80 h) using stronger energetic conditions: rotation speed of 400 rpm and ball-to-powder weight


ratio 30:1 [136]. Instead of ternary hydride formation, a gradual solid solution of hydrogen in theMg2Ni powder with a maximum hydrogen content of 1.8 H/f.u. after 80 h is reported. The positionof XRD diffraction lines does not depend however on the hydrogen content. The milled material hasa two-phase microstructure formed by nanocrystalline (15 nm) Mg2Ni regions with low H-content(�0.3 H/f.u.) and disordered intergrain regions that store high amounts of hydrogen. Such a particularmicrostructure has not later been confirmed by other research groups.

Tessier et al. also performed RMM experiments using Mg2Ni powder as starting reactant. RMMexperiments (p(H2) = 1.0 MPa for 10 h) were performed in a Kurimoto planetary mill at a high rotationspeed of 885 rpm [137]. Contrary to Orimo et al., the formation of disordered intergrain regions withhigh H-content is not observed. The synthesis of �70 wt% of Mg2NiH4 ternary hydride, as estimatedfrom total hydrogen content in the milled product, is detected. The hydride crystallizes as a mixtureof low- and high-temperature phases of Mg2NiH4 hydride [151,152]. The high temperature modifica-tion is usually formed above 510 K but it seems to be stabilized by mechanical milling [153].

Baum et al. have monitored the in situ hydrogen uptake during RMM of 2Mg + Ni elemental powdermixture with identical milling parameters as mentioned above for the synthesis of Mg2FeH6 [51]. Suc-cessful formation of Mg2NiH4 ternary hydride is reported after 16 h of milling. The hydrogen uptakecurve exhibits a unique step. This result differs from experiments on Mg2FeH6 and Mg2CoH5 formationand was tentatively attributed to a simpler process for Mg2NiH4 related to the existence of Mg2Nicompound. However, this seems not to be the case. Fig. 15 displays the in situ hydrogenation curvesmonitored during the synthesis of Mg2NiH4 hydride [116]. Indeed, one-hydrogenation step fort < 50 min is observed. As proved by XRD diffraction studies of the sample milled for this time (toppattern in Fig. 16), the reaction is however not completed since a high amount of the starting Ni pow-der remains unreacted. Further milling for 12 h leads to complete formation of Mg2NiH4 hydride asshown by the XRD pattern displayed in bottom Fig. 16. This second step on the formation of the ter-nary hydride is a solid–solid state reaction, i.e. it occurs without any hydrogen absorption. The reac-tion path of Mg2NiH4 formation is then described according to reaction scheme:

Fig.

2MgðsÞ þ NiðsÞ þ 2H2ðgÞ ! 2MgH2ðsÞ þ NiðsÞ !Mg2NiH4ðsÞ ð4Þ

According to this reaction and to the previous experiments reported by Baum et al. [51] and Zhanget al. [116], the synthesis of all Mg-based Mg2THx ternary hydrides (T = Fe, Co and Ni) by RMM of ele-mental pure powders under hydrogen atmosphere occurs through a common reaction path:

2MgðsÞ þ TðsÞ þ x2

H2ðgÞ ! 2MgH2ðsÞ þ TðsÞ þ ðx� 4Þ2

H2ðgÞ !Mg2THxðsÞ ð5Þ

with x = 6, 5 and 4 for T = Fe, Co and Ni, respectively.

0

1

2

3

4

3.4 H/f.u. 2Mg+Ni

Hyd

roge

n up

take

(H

/f.u

.)

2Mg

1st stept = 50 min

Milling time (min)

0 60 120 180 240

15. Time-evolution of the H concentration in solid-state during RMM of Mg powder and 2Mg + Ni powder mixture.

Fig. 16. XRD patterns and phase identification of RMM 2Mg + Ni powder mixture after the first (50 min) and after 720 min ofmilling (Cu Ka radiation).


Baum et al. have also reported the possibility of synthesizing quaternary Mg2T0.5T’0.5Hx hydrideswith T, T’ = Fe, Co and Ni [51]. This result has been confirmed by Deledda and Hauback for T = Feand T’ = Co [52]. The later authors obtained a quaternary hydride of composition Mg2(FeH6)0.5(CoH5)0.5

in which both [FeH6]4� and [CoH5]4� and complex anions coexist in the same compound. The synthe-sized compound desorbed hydrogen in a one-step reaction at temperatures between 500 and 600 Kwhich is between those of Mg2FeH6 and Mg2CoH5. These results open up the possibility of synthesiz-ing other Mg-based transition–metal complex hydrides, in which the hydrogen storage properties canbe tailored by varying the transition metals and the relative content of the complex anion.

3.4. Alanates

Since the pioneering work of Bogdanovic and Schwickardi, who found that Ti-catalyzed NaAlH4 canrelease hydrogen under moderate conditions, the potential interest of metal aluminum complex hy-drides, also known as alanates, for reversible hydrogen storage remains vivid [154]. Alanates areformed by anionic aluminum–hydrogen complexes (typically either [AlH4]� or [AlH6]3�) stabilizedby a cation (typically an alkali or alkaline earth metal) [2,155]. Their gravimetric (up to 10.8 wt%though only 5.6 wt% reversible) and volumetric capacities (�90 g/L) make these compounds attractivefor hydrogen storage.

Conventionally, alanates are prepared by a wet chemical route. For instance, to synthesize sodiumalanate, NaH and Al are diluted in THF and hydrogen pressure up to 20 MPa at 423 K is applied for fewdays. Subsequently, the NaAlH4 in solution is filtered and dried. Since the solubility of NaAlH4 is verylow, a high amount of solvent is necessary. Dymova et al. have shown that the synthesis of NaAlH4 canalso be obtained from Na, Al and H2 at high temperature (553 K), where Na is in liquid state, and highhydrogen pressure (17.5 MPa) [156].

In 1999, mechanochemical synthesis methods started being employed to synthesize Na3AlH6 andNa2LiAlH6 complex hydrides using NaAlH4 and LiAlH4 as reagents [157,158]. Very recently, it has beenshown that NaAlH4 can be formed by milling NaH and Al under a hydrogen atmosphere of 8.3 MPa,4 mol% of TiCl3 was added as a dopant [30]. The alanates synthesized by reactive mechanical millingare surveyed hereafter (Table 3).

3.4.1. Lithium alanatesThe synthesis of LiAlH4 and Li3AlH6 has been attempted by Kojima et al. under 1 MPa of hydrogen

using a planetary ball mill at 400 rpm [165]. After 24 h of milling with addition of TiCl3, the XRD pat-tern shows broad peaks of LiH and Al phases but no clear signature of alanate formation. However, theRaman and 27Al MAS NMR spectra indicate that a small amount of LiAlH4 was formed. Actually, lith-ium alanate LiAlH4 is metastable and it is considered as a non-reversible hydrogen storage compound

Table 3Alanates synthesised by RMM under hydrogen gas. The employed device, reactants, initial hydrogen pressure, p(H2), total millingtime (tmt), milling speed (ms), ball-to-powder weight ratio (BTPWR), ball diameter (Bd), reaction yield and formed side productsare given.

Compound Device Reactants p(H2)(MPa)

tmt(h)

ms(rpm)

BTPWR Bd(mm)

Yield(wt%)

Side products Ref

NaAlH4 Fritsch P7 NaH + Al 8.3 4 500 50:1 NaCl, Al [30]NaAlH4 Fritsch P5 NaH + Al 0.6 120 350 50:1 10 Low Na3AlH6, NaH,

Al[53]

NaAlH4 Fritsch P5 NaH + Al 1.2 240 230 10:1 10 High NaCl, Al [53]NaAlH4 Planetary NaH + Al 3 60 350 30:1 8 86 Na3AlH6, Al [159]Na3AlH6 Fritsch P6 NaH + Al 0.85 20 400 40:1 Al [160]Na3AlH6 Planetary NaH + Al 0.5 30 350 30:1 8 58 Al [159]Na3AlH6 Fritsch

P4a3NaH + Al 10 8 400 90:1 15 92 NaCl, NaH, Al [57]

KAlH4 Fritsch P7 KH + Al 10 10 400 [161]Na2LiAlH6 Fritsch

P4aLiH + 2NaH + Al 10 4 400 90:1 15 97 NaCl [57]

K2NaAlH6 FritschP4a

2KH + NaH + Al 10 2 400 90:1 15 98 NaH [57]

Mg(AlH4)2 Fritsch P7 2AlH3 + MgH2 0.1 4 400 175:1 10 81 AlH3, MgH2, Al [162]Ca(AlH4)2 Fritsch P6 2AlH3 + CaH2 0.1 10 500 13 [163]CaAlH5 Fritsch P7 AlH3 + CaH2 0.1 3 400 80:1 10 91 CaH2, Al [162]Ba2AlH7 Fritsch P5 BaH2 + Al 0.8 10 300 40:1 84 BaH2, Al [164]BaAlH5 Fritsch P5 BaH2 + 2Al 0.8 10 300 40:1 83 Al [164]

a Pressure and temperature measured in situ in the Evico-magnetics vial.


under moderate temperature and pressure conditions [166,167]. Indeed, the desorption of hydrogenfrom LiAlH4 is reported to be an exothermic process [161]. Lithium alanate does not form directly fromLiH, Al and H2 by solid–gas reaction. It is classically produced by using liquid complexing agents fol-lowed by desolvatation [168,169].

3.4.2. Sodium alanatesThe sodium tetra-alanate NaAlH4 is the best performing and most studied among all the alanates.

Bellosta von Colbe et al. have synthesized NaAlH4 by RMM (p(H2) = 8.3 MPa for 4 h) of NaH and Alpowders using 4 mol% TiCl3 as a dopant [30]. The obtained alanate exhibits a reversible capacity of4 wt% and very fast kinetics without activation. In the same year, Wang et al. also tried the synthesisof NaAlH4 by the same means but at a lower pressure of 0.85 MPa and using TiF3 as a dopant [160].Within the first 5 h of milling, alanate formation was not observed. However, further milling up to20 h led to partial synthesis of the more stable Na3AlH6 phase.

Eigen et al. have also tried the synthesis of NaAlH4 at moderate hydrogen pressures (from 0.6 to1.2 MPa) [53]. NaAlH4 could be formed with and without the addition of TiCl4 as catalyst. However,a long milling time (100 h without TiCl4, 20 h with TiCl4) is needed to obtain a small fraction of NaAlH4

under 0.6 MPa of hydrogen. Generated heat due to material plastic deformation and the friction be-tween the milling balls and vial wall during milling could lead to a temperature at which the NaAlH4

is not stable at the imposed hydrogen pressure. A complete formation of NaAlH4 could be achieved byreducing the milling energy (ball-to-powder weight ratio and rotational speed were decreased from50:1 and 350 rpm to 10:1 and 230 rpm) and increasing the hydrogen pressure (1.2 MPa).

Xiao et al. have also actively worked on the synthesis of NaAlH4 by RMM under hydrogen pressureof 0.5–3 MPa [159]. They observed that while Na3AlH6 forms under 0.5 MPa hydrogen pressure,NaAlH4 only forms above 0.8 MPa [159]. These results are similar to those of Eigen et al. and prove thatthe formation of NaAlH4 needs a minimum hydrogen pressure about 1 MPa, which is much higherthan the equilibrium pressure of NaAlH4 at room temperature (about 0.1 MPa) [53]. This is attributedto temperature rising on milling. By increasing the hydrogen pressure to 3 MPa, the synthesis yield ofNaAlH4 reached 87 wt% after milling for 60 h.

0.96

1.00

1.04

1.08

1.12

295

300

305

310

Na3AlH6 Po = 9.6 MPa

Na3AlH6 (TiCl 3 doped) Po = 10.2 MPaP/

P o

Milling time (min)

T (

K)

0 60 120 180 240 300 360 420 480

Fig. 17. Time-evolution of the hydrogen pressure (bottom) and temperature (top) during the synthesis of Na3AlH6 by RMMwith (empty symbols) and without (full symbols) TiCl3 dopant [57].


Sodium hexa-alanate is not available commercially. Some research groups have synthesized thehexa-sodium alanate by ball milling a mixture of the tetra alanate and the sodium hydride throughthe solid-sate reaction 2NaH(s) + NaAlH4(s) ? Na3AlH6(s) without any additives [157,158]. By usingRMM (p(H2) = 10 MPa for 4 h), high yield is obtained through the following reaction [57]:

3NaHðsÞ þ AlðsÞ þ 32

H2ðgÞ ��!2mol%TiCl3 Na3AlH6ðsÞ ð6Þ

This experiment was performed in an Evico-magnetics vial. Fig. 17 shows the temperature and hydro-gen pressure inside the milling vial for the formation of Na3AlH6 with and without TiCl3. In both cases,the temperature first increases for t < 30 min due to the heat generated on milling before reaching aplateau around 310 K due to the balance between internal friction heating and ventilation cooling.The hydrogen pressure increases simultaneously because of the temperature rise in the vial. At longermilling time, t > 30 min, the pressure drops markedly for the doped sample whereas it decreases muchmore slowly for the undoped one.

A further advantage on the synthesis of sodium alanates by RMM is the simultaneous incorporationof dopant. The addition of TiF3[159,160], TiCl4 [170], TiCl3 [57], ScCl3 and CeCl3 [171] dopants duringmilling not only speeds up the formation of NaAlH4 or Na3AlH6 compounds, but also promotes the dis-persion of the dopant in the alanate. The materials synthesized by one-step reactive mechanical mill-ing displayed superior features compared to doping of pre-synthesized NaAlH4 with TiCl3

[30,172,173].

3.4.3. Potassium alanatesPotassium alanate KAlH4 as well as K3AlH6 are stable alanates under normal conditions of pressure

and temperature. They decompose at moderate temperatures (above 550 K) [174]. Up till now no re-ports have been published of the direct synthesis of K3AlH6 by reactive ball milling. KAlH4 can be syn-thesized by RMM of KH + Al powder mixture after 10 h milling under 10 MPa of hydrogen pressure[161].

3.4.4. Mixed alkali alanatesThree mixed alkaline alanates are reported to exist: Na2LiAlH6, K2LiAlH6 and K2NaAlH6 [175–177].

In 1999, Na2LiAlH6 was synthesized by ball milling of NaAlH4 and LiH powder mixtures [157]. Re-cently, Na2LiAlH6 and K2NaAlH6 mixed alanates have been synthesized by RMM (p(H2) = 10 MPa for4 h) [57]. These compounds were prepared according to reaction schemes [56,57]:

LiHðsÞ þ 2NaHðsÞ þ AlðsÞ þ 32

H2ðgÞ ��!2 mol%TiCl3 Na2LiAlH6ðsÞ ð7Þ

Fig. 18.Vertica

Fig. 19.VerticaKa1,2) [


NaHðsÞ þ 2KHðsÞ þ AlðsÞ þ 32

H2ðgÞ ��!2 mol%TiCl3 K2NaAlH6ðsÞ ð8Þ

In situ hydrogen uptake curves proved that the above reactions typically take place in 2 h. Ex situ XRDanalysis after milling demonstrate the completion of the reaction as can be seen, for instance, for Na2-

LiAlH6 synthesis in Fig. 18. In addition, a small amount of NaCl is detected as a result from the reactionof TiCl3 with NaH [113]:

3NaHðsÞ þ TiCl3ðsÞ ! 3NaClðsÞ þ TiðsÞ þ 32

H2ðgÞ ð9Þ

Jeloaica et al. have attempted to synthesize ternary alkaline hexa-alanates of composition Li4/3Na1/3K4/

3AlH6 and LiNaKAlH6 [56]. However these compounds could not be obtained by RMM. The obtained

20 30 40 50 60 70 80

I (a

.u)

2θ

Rietveld analysis of RMM Na2LiAlH6: observed (dots), calculated (solid line) and difference curves (bottom) are shown.l bars (|) correspond to Bragg positions for Na2LiAlH6 (top) and NaCl (bottom) phases (Cu Ka1,2).

I (a

.u)

20 30 40 50 60 70 80

2θ

Rietveld analysis RMM LiNaKAlH6: observed (dots), calculated (solid line) and difference curves (bottom) are shown.l bars (|) correspond to Bragg positions for K2NaAlH6, KAlH4, LiNa2AlH6, NaH, LiH and NaCl phases from top to bottom (Cu56].


materials contain a mixture of alkaline hydrides (LiH, NaH and KH), the tetra-alanate KAlH4 and binaryhexa-alanate (NaK2AlH6 and Na2LiAlH6) phases (Fig. 19). Actually, it has been shown by DFT calcula-tions that the nominal Li4/3Na1/3K4/3AlH6 and LiNaKAlH6 phases are not stable with respect to the mix-ture of LiH, the tetra-alanate KAlH4 and binary hexa-alanate K2NaAlH6 or Na2LiAlH6.

3.4.5. Alkali-earth alanatesConventionally, Mg(AlH4)2 and Ca(AlH4)2 have been synthesized by wet chemistry through the

metathesis reaction between MCl2 (M = Mg and Ca) and NaAlH4 [178,179]. From the same reagentsbut by using mechanochemical milling, several research groups have also synthesized the magnesiumand calcium alanates [180–182]. But all these approaches produce significant amount of magnesiumor calcium chloride as by-products. These salts are ‘‘inert’’ to hydrogen and thus decrease the H-stor-age capacity of the final products.

Reactive mechanical milling of Mg(AlH4)2 alanate has been attempted in 2005 by Varin et al. usinga magneto-ball mill under 0.8 MPa hydrogen [183]. Three stoichiometric Mg + 2Al mixtures, (a) ele-mental Mg and Al powders, (b) elemental Al powder and commercial AZ91 alloy (Mg–Al–Zn alloy)and (c) powder of as-cast Mg + 2Al alloy have been used. No successful synthesis of Mg(AlH4)2 hasbeen achieved. The hydride formed after 270 h of milling is b-MgH2. In 2009, several groups triedthe synthesis of Mg-containing quaternary alkali alanates by RMM under a much higher pressure ofhydrogen (10 MPa). No traces of Mg(AlH4)2 alanate were however detected [184,185]. The lack of suc-cess in the synthesis of Mg(AlH4)2 alanate by RMM could be related to the lower thermodynamic sta-bility of Mg(AlH4)2 as compared to MgH2.

Interestingly, it has been demonstrated that Mg(AlH4)2 can be obtained by RMM (p(H2) = 10 MPafor 4 h) of 2AlH3 and MgH2 powder mixtures [162]. The reaction yield attained 81 wt% of Mg(AlH4)2.Experiments were conducted either in argon or hydrogen atmosphere. Strikingly, the millingatmosphere had minor impact on the yield of Mg(AlH4)2. In contrast, the most important parameterwas the pause used to allow vial cooling (2 min of pause following every 10 min of milling). In fact,it was found that during milling, alane decomposes due to the temperature rise and reduces the yieldof Mg(AlH4)2 formation. This suggests that the alanate formation occurs through an all solid-stateprocess:

MgH2ðsÞ þ 2AlH3ðsÞ ! 2MgðAlH4Þ2ðsÞ ð10Þ

This reaction could be driven by lower stability of alane as compared to magnesium alanate. Althoughthis synthesis route is successful, it should be taken into account that the preparation of alane is still achallenge [186–188].

As for the calcium alanate, progress on its synthesis by mechanical means is quite similar to themagnesium one. The synthesis of Ca–M–Al–H (M = Na and Li) quaternary alanates by RMM of binaryalkali hydrides and Al at 10 MPa of hydrogen pressure was unsuccessful [184]. In contrast, calciumalanate can be formed by mechanical milling using CaH2 and alane as reactants [162,163]. Strikingly,either Ca(AlH4)2 or CaAlH5 compounds are reported to be formed depending on the molar ratio be-tween reactants: 2AlH3 + CaH2 in the former case and AlH3 + CaH2 in the latter.

Other alkaline earth alanates that have been synthesized by RMM under hydrogen atmosphere areBa2AlH7 and BaAlH5 [164]. RMM experiments (p(H2) = 0.8 MPa for 10 h) have been performed usingBaH2 and Al powders as reactants. Three ratios of the reagent powder BaH2/Al 2:1, 1:1 and 1:2 havebeen used. Surprisingly, formation of Ba2AlH7 was favoured for 1:1 ratio as compared to 2:1. 2BaH2/Alpowder mixture forms 42 wt% of Ba2AlH7 (together with 58 wt% of BaH2) whereas BaH2/Al powdermixture forms 84 wt% of Ba2AlH7. Elemental Al is not observed in the former case though it is expectedfrom mass balance considerations. As for the synthesis of BaAlH5, a high yield (83.2 wt%) of this phaseis obtained using the ratio 1:2, i.e. with Al in excess. The reason of these striking results was notdiscussed.

RMM has also been used as intermediary step in the synthesis of Sr2AlH7 [189]. Sr2Al compoundwas milled under hydrogen (0.6 MPa) to get intimate 2SrH2 + Al mixture. Such a mixture could reactwith hydrogen at 553 K and 7 MPa in an external device to form Sr2AlH7. A similar approach has beenused for the synthesis of mixed (Sr,Ca)AlH7 compounds [190].


To summarize, mechanical milling under hydrogen gas is a very effective method for the synthesisof hydrides. Most hydrides formed by this method can be also obtained by solid–gas reaction:metastable hydrides are rarely formed. However, reactive mechanical milling is much faster than con-ventional solid–gas methods even for milder conditions of pressure and temperature. Main reasons forthis are probably clean surface generation and shortening of diffusion paths as result of mechanically-induced particle pulverization. In systems with more than one metal-constituent, extended and cleaninterface formation between reactive powders should also play a major role to speed solid/solidreactions.

Hydrides produced by RMM are generally nanocrystalline. Prolonged RMM may however lead toamorphization and compound disproportionation as it occurs in ternary metallic ABHx hydrides. Thiseffect has not been observed so far in complex ternary hydrides which may indicate that ionic-cova-lent compounds are less prone to amorphisation than metallic compounds.

3.5. Synthesis of borohydrides by mechanical milling in diborane gas

A novel technique for reactive ball-milling in diborane atmosphere was developed at EMPA, Swit-zerland, by Remhof, Friedrichs, Borgschulte, Züttel and co-workers. Solvent-free synthesis of lithiumborohydride, LiBH4 from lithium hydride, LiH in diborane, B2H6 gas at 393 K has recently been dem-onstrated [191]. Diborane gas is produced by heating LiZn2(BH4)5 prepared by ball milling of a2LiBH4:ZnCl2 mixture. This is considered a convenient and relatively safe source of diborane as com-pared to pressurized bottles, since diborane is a poisonous and explosive gas [192]. The formation ofLiBD4 by an addition reaction of LiD and B2D6 was analysed by in situ neutron diffraction, which showsthat nucleation of LiBD4 already starts at temperatures of ca. 375 K, i.e. formation of orthorhombicLiBH4. However, the reaction is incomplete and the yield is only �50% even at elevated temperatures(460 K) and the product contains small amounts of Li2B12H12. A passivation layer of LiBH4 is suggestedto form on the surface of the LiH grains retarding the process [193]. Therefore, a custom-made millconnected to a gas/vacuum supply via a flexible hose made of polyamide was used. The mill wasequipped with two stainless steel milling vials performing horizontal movements based on the see-saw principle. One vial was equipped with ceramic balls. This approach allows continuous removalof the borohydride surface layer, see Fig. 20 [194].

This new mechanochemical method allows solvent-free synthesis of borohydrides at room temper-ature demonstrated by the synthesis of three of the most investigated borohydrides at present: LiBH4,Mg(BH4)2 and Ca(BH4)2. Similarly, Y(BH4)3 was prepared in a reaction with YH3 and B2H6 with thisnew RBM [195]. This new gas–solid mechanochemical synthesis method is based on the reaction ofmetal hydrides with diborane to form the corresponding borohydrides. The synthesis method mayfacilitate preparation of a wide range of different borohydrides in the future. Furthermore, with this

Fig. 20. Schematic presentation of a custom-made mill connected via a flexible hose to a gas/vacuum supply [194]. The gassupply consists of a container that is situated in an oven and is filled with a borane desorbing material, i.e. LiZn2(BH4)5.


technique it was demonstrated that Li2B12H12 and an amorphous Li2B10H10 phase can be prepared by areaction with LiBH4 and diborane (B2H6) at T � 475 K. This tend to suggest that formation of higherboranes, such as Li2B12H12, may occur during thermal decomposition of metal borohydrides [196].

Recently, a metal borohydride was prepared by reactive ball milling of a metal boride in hydrogenatmosphere. The starting materials, MgB2 and hydrogen gas, p(H2) = 10 MPa, was charged into a hard-ened steel high pressure vial (ball to powder ratio of 30:1) and milled at a rotation speed of 600 rpmfor up to 100 h. The product, unsolvated amorphous magnesium borohydride (yield �50%) appearedto form directly but impurities of higher boranes Mg(BnHm)y (�30%) were also identified, which mayhave formed in side reactions [197]. Interestingly, a new polymorph c-Mg(BH4)2 with 30% open spacein the structure transformed from crystalline to semi-amorphous by mechano chemical treatment[198].

The hydrogen absorption mechanism for ball milled samples of 2NaH–MgB2 was also investigatedin detail. At high pressures, p(H2) � 5 MPa, NaBH4 formed after observation of an unknown compound,NaMgH3 and a NaH–NaBH4 molten salt mixture. In contrast, NaBH4 apparently formed directly at low-er hydrogen pressures, i.e. p(H2) = 0.5 MPa. This indicates that the reaction mechanism may be mod-ified by mechano-chemical treatment of the reactants and by the applied hydrogen pressure [199].

This work reveals that reactive ball milling of metal hydrides in diborane, metal boride or mixturesof metal hydrides and borides in elevated hydrogen pressures has a potential for providing new con-venient solvent-free routes for preparation of metal borohydrides.

3.6. Synthesis of metal amides by mechanical milling in ammonia gas

During the past decade lithium amide, LiNH2 has received much interest as a hydrogen storagematerial [200,201]. Recently, LiNH2 and a range of other metal amides i.e. NaNH2, Mg(NH2)2 andCa(NH2)2 have been prepared by mechanochemical methods from metal hydrides, MHx and ammoniagas, NH3 according to reaction scheme Eq. (11) [202,203].

MHxðsÞ þ xNH3ðgÞ ! MðNH2ÞxðsÞ þ xH2ðgÞ ð11Þ

In general, gas loadings of p(NH3) � 0.4–0.5 MPa was used with refilled approximately every secondhour. A well-known problem of utilization of the metal amides as hydrogen storage materials is therelease of NH3 gas besides hydrogen during the thermal decomposition [59,204,205]. In order to avoidthis, a metal hydride can be added to absorb the NH3 according to reaction scheme Eq. (11). The effi-ciency of this approach has been studied by milling LiH in p(NH3) = 0.4 MPa corresponding to LiH–NH3

(1:1). Gas chromatographic analysis revealed that 70% of the initial NH3 reacted with LiH and trans-formed to H2 after 30 min of milling [206].

4. Synthesis of hydrides by mechanically-induced solid/solid and solid/liquid reactions

4.1. Mechanochemical synthesis of metal borohydrides

Metal borohydrides are well known for their properties as reducing agents and neutron absorbers.However, due to their high gravimetric hydrogen density this class of compounds has recently re-ceived increasing interest as potential hydrogen storage materials [2–4]. Schlesinger et al. reportedin 1953 the first mechanochemical synthesis of a metal borohydride i.e. sodium borohydride from so-dium hydride and boric oxide according to reaction scheme [207].

4NaHþ 2B2O3 ! NaBH4 þ 3NaBO2 ð12Þ

Similar methods are still investigated for preparation of alkali borohydrides, e.g. NaBH4 or KBH4

from MgH2 and Na2B4O7 or KBO2, respectively [208,209].Due to the high thermal stability of the alkali borohydrides much attention has been given to syn-

thesis and characterization of novel metal borohydrides e.g. based on alkaline earth and transitionmetals. In 1989 Mal’tseva et al. showed that ball milling of alkali borohydrides, MBH4 (M = Li, Na orK) with zinc chloride, ZnCl2 resulted in a metathesis reaction, i.e. formation of MCl was observed by

Table 4Metal borohydrides synthesized by mechanochemical methods, reactants used for the synthesis, optimal reactant ratio, formedside products, total milling time (tmt) and milling speed (rounds per minute, rpm) used in the synthesis.

Compound Reactants Opt. ratio Side products tmt (min) rpm Ref.

NaBH4 MgH2 + Na2B4O7 4:1 MgO, B2O3 60 2750 [207,209]KBH4 MgH2 + KBO2 2:1 MgO 120 490 [208]Sr(BH4)Cl LiBH4 + SrCl2 1:1 LiCl, Sr(BH4)2 120 400 [39]Sr(BH4)2 LiBH4 + SrCl2 1:1 LiCl, Sr(BH4)Cl 120 400 [39]LiSc(BH4)4 LiBH4 + ScCl3 4:1 LiCl 180 500 [211,212]NaSc(BH4)4 NaBH4 + ScCl3 2:1 Na3ScCl6 120 400 [213]KSc(BH4)4 KBH4 + ScCl3 2:1 K3ScCl6 120 400 [214]Y(BH4)3 LiBH4 + YCl3 3:1 LiCl 120 200 [215,216]NaY(BH4)2Cl2 NaBH4 + YCl3 2:1 Na3YCl6, Na(BH4)1�xClx 120 200 [38]Mn(BH4)2 LiBH4 + MnCl2 2:1 LiCl 350 600 [40]

NaBH4 + MnCl2 2:1 NaCl 350 600 [40]LiZn2(BH4)5 LiBH4 + ZnCl2 5:2 LiCl 120 200 [217]NaZn2(BH4)5 NaBH4 + ZnCl2 5:2 Na2ZnCl4, NaCl 120 200 [217]NaZn(BH4)3 NaBH4 + ZnCl2 3:1 Na2ZnCl4, NaCl 120 200 [217]KZn(BH4)3 KBH4 + ZnCl2 2:1 K2Zn(BH4)xCl4�x, K3Zn(BH4)xCl5�x 350 – [218]KZn(BH4)Cl2 KBH4 + ZnCl2 1:1 – 120 200 [219]Cd(BH4)2 LiBH4 + CdCl2 2:1 LiCl 30 200 [37]

NaBH4 + CdCl2 14:9 NaCl, Na6CdCl8 30 200 [37]KCd(BH4)3 KBH4 + CdCl2 1:1 KCdCl3, K2Cd(BH4)4, Cd(BH4)2 20 200 [37]K2Cd(BH4)4 KBH4 + CdCl2 4:3 KCdCl3 20 200 [37]Li4Al3(BH4)13 LiBH4 + AlCl3 13:3 LiCl 300 500 [220]Li(BH4)0.9Cl0.1 LiBH4 + LiCl – – 120 200 [221,222]Li(BH4)0.47Br0.53 LiBH4 + LiBr – – 120 200 [223]Li(BH4)0.3I0.7 LiBH4 + LiI – – 120 200 [224]Na(BH4)0.9Cl0.1 NaBH4 + NaCl – – 120 200 [225]Ca(BH4)1.6I0.4 Ca(BH4)2 + CaI2 – – 120 250 [226]


PXD [210]. Such mechanochemical methods have become one of the most common methods for prep-aration of novel metal borohydrides during the past decade and a significant increase in the number ofnew compounds and studies of their structural, physical and chemical properties [3,4]. Table 4provides an overview of reactants and conditions for synthesis of metal borohydrides along with sideproducts, which are obtained in some cases.

During the mechanochemical synthesis a variety of reactions can occur and in some cases severalcompeting reactions are observed simultaneously. Metathesis, or double substitution reaction, is awell-known mechanism for chemical reactions during ball milling, here illustrated by the reaction be-tween lithium borohydride LiBH4 and yttrium(III)chloride, YCl3, which results in formation of Y(BH4)3

and LiCl according to reaction scheme [215,216].

ð1 : 3Þ YCl3 þ 3LiBH4 ! YðBH4Þ3 þ 3LiCl ð13Þ

The polymorph a-Y(BH4)3, previously obtained from diethyl ether solutions of LiBH4 and YCl3 at RTand also by MM, can be obtained with varying amounts of a new polymorph, denoted b-Y(BH4)3. a-Y(BH4)3 can be considered as the stable phase at ambient conditions, whereas b-Y(BH4)3, is a high-temperature polymorph formed by an a to b polymorphic phase transition. The high-temperatureb-polymorph can be quenched to ambient conditions [215,216,227,228].

It should be noted that many single-cation metal borohydrides have been reported to be preparedby mechanochemical synthesis e.g. Zn(BH4)2 [229,230], Sc(BH4)3 [229], Ti(BH4)3 [229,231], V(BH4)2

and Cr(BH4)2 [232]. However, no structural data have been reported for these compounds and theirformations are based on observation of alkali chlorides and assumption of a simple metathesis reac-tion similar to reaction scheme Eq. (13). Further investigations of some of these systems have shownthat more complex reactions often take place [213,214,217,219].

The system ZnCl2–MBH4 (M = Li, Na or K) can be used to illustrate the complexity of the mechano-chemical synthesis. Ball milling a mixture of ZnCl2–KBH4 (1:1) leads to an addition reaction and aphase pure product, of KZn(BH4)Cl2 [219].

Fig. 21consistcontain


ZnCl2 þ KBH4 ! KZnðBH4ÞCl2 ð14Þ

Interestingly, the unit cell volume of KZn(BH4)Cl2 (V/Z = 149.5 Å3) is nearly equal to the sum of for-mula volumes for the reactants ZnCl2 (V/Z = 74.3 Å3) and KBH4 (V/Z = 76.2 Å3). However, there are sig-nificant differences between the structure of reactants and the product. The latter contains theheteroleptic complex ion [Zn(BH4)Cl2]� where Zn coordinates to two chloride ions and two hydrogensin g2–BH4, i.e. CN(Zn) = 4. This clearly demonstrates that ball milling induces a complex chemical reac-tion involving bond breaking and bond formation. In this case, the structure is fully ordered. The addi-tion reaction may also provide a solid solution, e.g. Li(BH4)1�xClx, which will be further discussed laterin this section.

The mechanochemical synthesis of other new borohydrides in the system ZnCl2–MBH4 proceedsvia more complex chemical reactions during ball milling [217]

2ZnCl2 þ 5LiBH4 ! LiZn2ðBH4Þ5 þ 4LiCl ð15Þ

2ZnCl2 þ 5NaBH4 ! NaZn2ðBH4Þ5 þ 4NaCl ð16Þ

ZnCl2 þ 3NaBH4 ! NaZnðBH4Þ3 þ 2NaCl ð17Þ

Reaction schemes (16) and (17) illustrates that small deviations in the composition of reactants maylead to significantly different reaction products both in terms of the stoichiometry and the structuraltopology, i.e. the structures of NaZn(BH4)3 and NaZn2(BH4)5 are significantly different, which may sug-gest that the synthesis mechanism for these compounds is also different. The compounds LiZn2(BH4)5

and NaZn2(BH4)5 are isostructural and built from two identical interpenetrated three-dimensionalframeworks consisting of isolated complex anions, [Zn2(BH4)5]�, whereas NaZn(BH4)3 consists of a sin-gle three-dimensional network, containing polymeric anions with the composition ½ZnðBH4Þ3�

n�n (see

Fig. 21) [42,217].These new compounds were prepared by high-energy MM in short intervals (2 min) separated by

pauses (2 min). This procedure suppresses heating of the sample by friction heat and keeps the tem-

. Crystal structure of LiZn2(BH4)5 (left) built from two identical interpenetrated three-dimensional frameworksing of isolated complex anions, [Zn2(BH4)5]�, and NaZn(BH4)3 (right) consisting of a single three-dimensional network,ing polymeric anions with the composition ½ZnðBH4Þ3�

� [42,217].


perature close to room temperature. This suggests that the synthesis is facilitated by high pressurerather than elevated temperature. In fact, this class of M–Zn–BH4 compounds decomposes irreversiblyby heating at �373 K or prolonged MM with release of diborane [36]. Furthermore, these compoundsare metastable, i.e. they decompose when stored at RT within approximately 1 week. These observa-tions suggest that ball milling may lead to a chemical equilibrium state rather than just a statisticaldistribution of reactants.

Reaction scheme Eqs. 13,15,16, and 17 also illustrate a general drawback of the mechanochemicalapproach involving metathesis reactions, namely that the products may be contaminated with ioniccompounds, most often binary metal halides but sometimes also ternary halides. Formation ofLiZn2(BH4)5 (reaction scheme Eq. (15)) proceeds completely, i.e. with formation of LiCl as the only sideproduct, whereas formation of NaZn2(BH4)5 and NaZn(BH4)3 (reaction schemes Eqs. (16) and (17))only proceeds partly due to formation of a ternary metal chloride by a simultaneous and competingreaction described in reaction scheme [42].

ð1 : 2Þ ZnX2 þ 2MX ! M2ZnX4 ð18Þ

Mechanochemical synthesis of M2ZnX4, M = Li or Na, X = Cl or Br from stoichiometric mixtures ofMX and ZnX2 has previously been reported by Solinas et al., who state that reaction times and activa-tion energy decreases as Li2ZnCl4 > Na2ZnCl4 > Na2ZnBr4 [233]. This is in agreement with the resultsfor the zinc-based borohydride systems ZnX2–MBH4 (M = Li, Na, K and X = Cl, Br), where formationof M2ZnX4 has increasing dominance over the heavier elements, i.e. K > Na� Li and Br > Cl. However,reaction Eq. (18) is only weakly coupled with the formation of the borohydrides. In other words, reac-tion Eqs. (16) and (17) are faster than reaction Eq. (18). This contrasts the mechanochemical synthesisof NaSc(BH4)4 and KSc(BH4)4 from ScCl3 and NaBH4 or KBH4, respectively, which is suggested to pro-ceed as described by reaction scheme [213,214].

ð1 : 4Þ ScCl3 þ 4NaBH4 ! NaScðBH4Þ4 þ 3NaCl ð19Þ

ð1 : 4Þ ScCl3 þ 4KBH4 ! KScðBH4Þ4 þ 3KCl ð20Þ

Surprisingly, powder X-ray diffraction data show no presence of NaCl or KCl in any of the ball-milledsamples of ScCl3–MBH4 (M = Na or K) in molar ratios 1:2, 1:3 or 1:4. This indicates that the mechano-chemically induced reactions differ from reaction schemes Eqs. (19) and (20). Two sets of unidentifiedBragg peaks were observed in all the reaction products and one was assigned to MSc(BH4)4 and theother to a new ternary sodium scandium chloride M3ScCl6. Furthermore, the samples with the startingratios ScCl3–MBH4 of 1:3 and 1:4 contain different amounts of MBH4 whereas no diffraction from ScCl3

was observed. The samples with the starting ratio of 1:2 show neither ScCl3 nor MBH4 peaks and ap-pear to contain the largest fraction of the new compounds, NaSc(BH4)4 and KSc(BH4)4 for M = Na or K,respectively [213,214]. This suggests that an addition reaction is responsible for the formation of theternary salts, Na3ScCl6 and K3ScCl6 according to reaction scheme

ð1 : 3Þ ScCl3 þ 3NaCl! Na3ScCl6 ð21Þ

ð1 : 3Þ ScCl3 þ 3KCl! K3ScCl6 ð22Þ

In conclusion, the optimal ratio of reactants ScCl3–MBH4 (M = Na or K) for synthesis of MSc(BH4)4

turns out to be 1:2. The above-mentioned observations can be explained assuming that the formationsof the ternary salts, M3ScCl6 (Eqs. (21) and (22)) are much faster than the formations of the borohy-drides, MSc(BH4)4 (Eqs. (19) and (20)). Hence, the overall reactions for the samples ScCl3–MBH4 in1:2 ratio are described in Eqs. (23) and (24) as a sum of the strongly coupled reactions for formationof MSc(BH4)4 and M3ScCl6. These mechanochemical syntheses lead to maximum borohydride yields of22 wt% and 18 wt% for NaSc(BH4)4 and KSc(BH4)4, respectively [213,214].

ð1 : 2Þ 2ScCl3 þ 4NaBH4 ! NaScðBH4Þ4 þ Na3ScCl6 ð23Þ

ð1 : 2Þ 2ScCl3 þ 4KBH4 ! KScðBH4Þ4 þ K3ScCl6 ð24Þ


In contrast, for the system ScCl3–LiBH4, only one chemical reaction occurs during ball milling,which produces LiSc(BH4)4 and LiCl [211,212].

Ionic-covalent type hydrides, i.e. complex hydrides, may also form solid solutions, e.g. ca. 10 mol%lithium chloride may dissolve in solid lithium borohydride by ball milling as described in reactionscheme [221,222].

xLiClþ 1� xLiBH4 ! LiðBH4Þ1�xClx ð25Þ

On the other hand, Rietveld refinement of powder X-ray diffraction data reveals that Cl� readilysubstitutes for BH�4 in the structure of hexagonal h-LiBH4 upon prolonged heating, e.g. at 498 K for�1 h resulting in formation of h-Li(BH4)1�xClx, x � 0.42 [221]. This observation indicates that forma-tion of such solid solutions is more efficiently facilitated by elevated temperatures rather than by highpressures generated by mechanochemical methods. These types of anion substitution reactions alsotake place for a range of other metal borohydride-metal halide systems both as a result of mechano-chemical treatment and subsequent heating, e.g. LiBH4–LiBr [223], LiBH4–LiI [224], NaBH4–NaCl [225],Ca(BH4)2–CaF2, Ca(BH4)2–CaCl2 and Ca(BH4)2–CaI2 [226,234]. Furthermore, substitution of BH�4 by Cl�

can also take place in the novel metal borohydrides formed during the mechanochemical synthesis,e.g. Rietveld refinement suggests incorporation of �42 mol% of Cl� on one of the two BH�4 sites in Al3-

Li4(BH4)13 [220].Generally, the smaller anion tends to dissolve in the compound containing the larger anion, and the

structure of the latter tends to be preserved in the obtained solid solution in accordance with theobservation of a CaI2-type trigonal solid solution, tri-Ca((BH4)1�xIx)2, i.e. dissolution of Ca(BH4)2 inCaI2. This trend follows the relative size of the anions, I� > BH�4 > Br� > Cl� derived by a comparisonof the unit cell volumes for different inorganic salts [235,236]. This trend in anion substitution reac-tions can be interpreted as an increase in the lattice energy due to the clearly observed decrease in theunit cell volume, i.e. a decrease in the average distance between the ions in the structure.

Furthermore, anion substitution can occur due to structural similarities between the two com-pounds, which may partly explain the observation of two solid solutions of mechanically treatedLiBH4—LiI, i.e. dissolution of LiBH4 in LiI and dissolution of LiI in LiBH4. The two solid solutions ofLiBH4—LiI, b–LiI and h-LiBH4 all adopt isostructural hexagonal structures. Upon heating, the two solidsolutions of LiBH4—LiI merge into one [224]. Similarly, prolonged heating of NaBH4—NaCl producestwo solid solutions while NaBH4 and NaCl both share the same rock salt structure type [225]. This con-trasts the behaviour of LiBH4–LiCl system where there are no indications of any dissolution of LiBH4 inthe alkali halide salts [221,222].

Recently, cation substitution in a borohydride by ball milling was observed for the first time by for-mation of a Mg1�xMnx(BH4)2 solid solution. This substitution is most likely related to the close struc-tural similarity of Mn(BH4)2 to a-Mg(BH4)2 and it is interesting to note that this compound retains theframework structure [237].

Recently, a new series of double-cation double-anion borohydride chlorides based on rare-earthselements, LiM’(BH4)3Cl, M’ = La, Gd, and Ce, were discovered using combined mechano-chemical syn-thesis and heat treatment using M’Cl3–LiBH4 (1:3) mixtures [238,239]. The novel cubic compounds,LiM(BH4)3Cl contains isolated tetranuclear anionic clusters [Ce4Cl4(BH4)12]4� with a distorted cubaneCe4Cl4 core charge-balanced by Li+ cations. The Li+ ions are disordered and occupy 2/3 of the 12dWyckoff sites and DFT calculations indicates that LiCe(BH4)3Cl is stabilized by larger entropy ratherthan smaller energy. The new compound LiM’(BH4)3Cl simultaneously carries moderate amounts ofhydrogen, which is released at relatively low temperatures, and are fast Li-ion conductors.

4.2. Synthesis of novel alane and metal alanates

Among the most attractive materials for reversible hydrogen storage is the light-weight aluminiumbased hydrides due to their relatively high gravimetric hydrogen capacities and, importantly, hydro-gen release and uptake at moderate conditions [2,240]. In particular, Ti-doped NaAlH4 has receivedsignificant attention in the past decade [154]. Hydridoaluminates containing a negatively chargedcomplex anion are denoted alanates in the following whereas polymorphs of aluminium hydride,


AlH3 are denoted alane. Alanates synthesised by RMM in hydrogen gas was discussed in Section 3.4 ofthis review paper.

For the most stable alane polymorph, a-AlH3 the dehydrogenation enthalpy is determined by bombcalorimetry to be 7.6(6) kJ/(mol H2) and the total dehydrogenation entropy is found to be 30.0(4) J/(K�mol a-AlH3) [241]. From these values and the entropies of Al and H2, the hydrogenation pressureof Al at room temperature according to reaction scheme Eq. (26) can be estimated to be above 10 GPa.

Alþ 3=2H2 ! AlH3 ð26Þ

Therefore, the traditional synthesis of Al-based hydrides is based on wet-chemistry methods where acoordinating solvent forms a kinetically-stabilized hydride under mild conditions that can be isolatedfrom the solution [242,243]. Subsequently, the intermediate complex metal hydride solvent complexcan be decomposed to recover the pure hydride. Recently, it was discovered that a reversible hydro-genation reaction using Ti-catalyzed Al powder and triethylenediamine (TEDA) in tetrahydrofuranforms the alane adduct (AlH3�TEDA) at low pressure (p(H2) = 2 MPa, RT) [244].

In contrast, during the past few years mechanochemical techniques at a variety of different condi-tions have been used with increasing frequency for synthesis of alanates or for formation of homoge-neous alanate-additive samples, e.g. in argon or hydrogen atmosphere or at liquid nitrogentemperatures (cryo-milling) [245].

Interestingly, alane, AlH3 can be prepared directly by cryo-milling LiAlH4 and AlCl3 at 77 K resultingin a metathesis reaction according to reaction scheme Eq. (27) [188,246,247].

3LiAlH4 þ AlCl3 ! 4AlH3 þ 3LiCl ð27Þ

The reaction also proceeds at RT, however it results in low yields of alane and a high content of Al, dueto the high milling energy at RT which causes AlH3 to decompose to Al and H2. Furthermore, increasedbrittleness at low temperatures leads to reduced particle sizes, hence shortened diffusion paths. Re-cently it was found that the yield of alane might be further increased by using NaAlH4 instead ofLiAlH4 or AlBr3 instead of AlCl3. The same study also revealed that the relative amount of a- and a’-AlH3 formed in reaction scheme Eq. (27) can be somewhat controlled by adding 2.4 mol% FeF3 tothe reactant mixture [248].

The thermodynamically unstable AlH3, which has a three-dimensional covalent network structure,decomposes slightly above room temperature but may be kinetically stabilized by thin oxide layers[249]. This contrasts the ionic sodium hydride, NaH which decomposes at T � 700 K [1]. Mechano-chemistry reveals an interesting example, namely that reaction between a stable and an unstablecompound may provide a new material with intermediate stability. In this case a two-step additionreaction occurs according to reaction scheme Eqs. (28) and (29) and sodium alanate is formed builtfrom discrete [AlH4]� complex anions and sodium counter cations [65,155].

3NaHþ AlH3 ! Na3AlH6 ð28Þ

Na3AlH6 þ 2AlH3 ! 3NaAlH4 ð29Þ

This approach for tailoring thermodynamic properties is also used for ‘destabilising’ hydrides byformation of reactive hydride composites, e.g. 2LiBH4–MgH2 and LiBH4–Al [66,250–252].

Sodium alanate, NaAlH4 has moderate formation enthalpies, however the rehydrogenation ofNaAlH4 is kinetically hampered and hydrogen release and uptake in this material need to be catalyzedby titanium [154]. The materials resulting from mechanochemical doping NaAlH4 with small amountsof catalytic TiCl3 (<2 wt%) have kinetic and cycling properties that are closer to those required for apractical hydrogen storage medium [172,253]. Doping through mechanical milling of NaAlH4 withdopant precursors is an effective mean of charging the hydride with catalyst but also activates thematerial through reduction of the average particle size [173,253].

Furthermore, mechanochemical treatment can promote decomposition of metal alanates withaddition of catalysts [113,254,255]. Balema et al. and Easton et al. investigated the decompositionof LiAlH4 into Li3AlH6 and Al with release of hydrogen at RT during short-time ball milling with addi-tion of catalytic amounts of TiCl3 (3 and 2 mol%, respectively) [254,255]. Milling times as short as5 min proved effective for this transformation and according to PXD data LiAlH4 was completely trans-

Fig. 22. Ball collision effects on reaction mixtures as a function of ball impact energy.


formed, i.e. only Bragg peaks from Li3AlH6, Al and LiCl were observed [255]. Monitoring of the gaspressure within the milling vial as a function of milling time is shown in Fig. 22. The high catalyticactivity of TiCl3 and other additives may be attributed to the in situ formation of nano/microcrystal-line, e.g. Al–Ti phases, during ball milling [166,256,257]. Similar results was obtained by mechano-chemical treatment of NaAlH4 with TiCl3, ZrCl4, FeCl2 and FeCl3 [113].

Also it has been demonstrated that ball milling alone to some extent improves the dehydrogenat-ing kinetics of undoped metal alanates, e.g. LiAlH4 and NaAlH4 [258,259]. A study of LiAlH4 showedthat milling between 1.5 and 14 h promoted decomposition at 333 K lower than for as-receivedLiAlH4. Furthermore, it was also shown that LiAlH4 milled for less than 14 h decomposes via a two-step reaction pathway (reaction schemes Eqs. (30) and (31)).

3LiAlH4 ! Li3AlH6 þ 2Alþ 3H2 ð30Þ

Li3AlH6 ! 3LiHþ Alþ 1:5H2 ð31Þ

It has also been demonstrated that mechanochemical synthesis is a convenient method for prepa-ration of bialkali metal hydrides from alkali metal hydrides and alkali metal alanates, e.g. Na2LiAlH6, isprepared by ball milling sodium hydride, lithium hydride, and sodium alanate in a stoichiometriccomposition which reacts by an addition reaction according to reaction scheme Eq. (32) [157].

NaHþ LiHþ NaAlH4 ! Na2LiAlH6 ð32Þ

Bialkali alanates such as K2LiAlH6 and K2NaAlH6 have been prepared by similar reactions [176,177].Magnesium and calcium alanate can be readily prepared from sodium or lithium alanate and mag-

nesium or calcium chloride, respectively via a metathesis reaction according to reaction scheme Eq.(33) [161,176,179,260–262].

M0Cl2 þ 2MAlH4 ! M0ðAlH4Þ2 þ 2MClðM0 ¼Mg or Ca; M ¼ Li or NaÞ ð33Þ

Interestingly, if the amount of lithium alanate in the reaction mixture is increased so that a com-position of MgCl2:LiAlH4 � 1:3 is used, lithium magnesium alanate, LiMg(AlH4)3 forms according toreaction scheme Eq. (34) [261].

MgCl2 þ 3LiAlH4 ! LiMgðAlH4Þ3 þ 2LiCl ð34Þ

Recently, also Sr(AlH4)2 and lithium beryllium hydrides LinBemHn+2m have been prepared mechano-chemically [263,264]. Furthermore, rare earth metal chlorides have been ball milled with sodiumalanate and partial decomposition was observed due to a metathesis reaction and release of hydrogento form the more stable hexahydridoaluminate complex ion, AlH3�

6 , according to reaction schemeEq. (35), which is suggested for R = La, Pr, Ce, Nd [265].


RCl3 þ 3NaAlH4 ! RðAlH4Þ3 þ 3NaCl! RAlH6 þ 2AlþH2 ð35Þ

4.3. Novel quaternary hydrides

Recently several novel quaternary hydrides based on the amide anion, NH�2 and alanates, AlH�4 orborohydrides, BH�4 or BH3 have been prepared. This class of materials is of high interest for hydrogenstorage, since it combines both high hydrogen densities and storage properties of two types of storagematerials. A range of novel materials can be synthesized by mechanical milling utilizing several differ-ent types of reactions.

4.3.1. Metal borohydride amidesFormation of a quaternary hydride having the approximate composition Li3BH4(NH2)2, has been

achieved mechanochemically from LiBH4�LiNH2 in molar ratio 2:1 according to the addition reactionshown in reaction scheme Eq. (36) [266,267].

2LiNH2 þ LiBH4 ! Li3BH4ðNH2Þ2 ð36Þ

Pinkerton et al. also reported on the extent of the reaction as a function of the milling time. Thisshowed that a substantial amount of Li3BH4(NH2)2 is formed after 40 min of milling and that the reac-tion is completed after 300 min of milling. Milling for a total of 960 min caused no further changes inthe reaction product suggesting that the milling does not produce an amorphous phase.

Noritake et al. found that varying the LiNH2:LiBH4 molar ratio i.e. using 1:1, 2:1 and 3:1 yieldedproducts of different compositions according to reaction schemes Eqs. (36)–(38) [268].

LiNH2 þ LiBH4 ! Li2ðBH4ÞNH2 ð37Þ

3LiNH2 þ LiBH4 ! Li4ðBH4ÞðNH2Þ3 ð38Þ

The BH�4 and NH�2 anions are preserved during the mechanochemical synthesis and in the structure ofLi4(BH4)(NH2)3 they are positioned on specific crystallographic sites, i.e. an ordered structure is formedduring ball milling as opposed to formation of a solid solution. The structures of Li2(BH4)NH2 andLi3BH4(NH2)2 remain to be solved hence the precise stoichiometry of these compounds is not known.However, the compositions of the ball-milled mixtures and the decomposition reactions suggest thecompositions stated in reaction schemes Eqs. (36) and (37).

Furthermore, a phase diagram study was carried out by Meisner et al. in which samples of LiNH2–LiBH4 in a wide range of molar ratios of x = 0.33–0.80 were prepared by ball milling [269]. The studyshows that samples of x = 0.6–0.8 yield the cubic Li4BH4(NH2)3, while x < 0.5 yield another cubic com-pound, presumably Li3(BH4)(NH2)3. Furthermore, the study also implies existence of two other newcompounds, denoted c and d at x = 0.6 and 0.8, respectively. These compounds might be metastablearising from the highly non-equilibrium mechanochemical process.

Several metal amidoboranes, M(NH2BH3)x have also been prepared by mechanochemical synthesisfrom amidoborane, NH3BH3 and metal hydrides, e.g. LiNH2BH3, NaNH2BH3 and Ca(NH2BH3)2 accordingto reaction scheme Eqs. (39) and (40) [270,271].

NH3BH3ðsÞ þMHðsÞ ! MNH2BH3ðsÞ þH2ðgÞðM ¼ Li;NaÞ ð39Þ

2NH3BH3ðsÞ þ CaH2ðsÞ ! CaðNH2BH3Þ2ðsÞ þ 2H2ðgÞ ð40Þ

Recently, LiNH2BH3 has been used as starting material for synthesis of Y(NH2BH3)3 by a metathesisreaction according to reaction scheme Eq. (41) [272].

YCl3 þ 3LiNH2BH3 ! 3LiClþ YðNH2BH3Þ3 ð41Þ

In the same study similar syntheses were tested utilizing different combinations of MNH2BH3 (M = Li,Na) and YX3 (X = F, Cl), however the reactions seem either to have a very low efficiency or the desiredreaction product forms as an amorphous phase.


4.3.2. Metal alanate amidesIn general, ball milling mixtures of alkali amides, MNH2 and alkali alanates, MAlH4 (M = Li or Na)

result in the release of hydrogen caused by the occurring multistep reactions [273–276]. Furthermore,the presence of amide might facilitate transformation of the tetrahedral AlH�4 into octahedral AlH3�

6 byball milling at ambient conditions. Dolotko et al. observed formation of Li3AlH6 already after 4 min ofmilling of a LiAlH4–LiNH2 sample in molar ratio 1:1 [274]. Xiong et al. report release of approximateone equivalent H2 as all LiAlH4 is transformed to Li3AlH6 according to reaction scheme Eq. (42) [273].

3LiAlH4 ! Li3AlH6 þ 2Alþ 3H2 ð42Þ

In comparison negligible amounts of hydrogen were released from LiAlH4 ball milled for 36 h [273].Prolonged milling results in formation of amorphous compounds and from combined MAS NMRand PXD analysis the reactions shown in reaction scheme 43 and 44 are suggested [274].

2LiAlH4 þ LiNH2 ! Li3AlH6 þ AlNþ 2H2 ð43Þ

Li3AlH6 þ LiNH2 ! 4LiHþ AlNþ 2H2 ð44Þ

A similar sample of NaAlH4–NaNH2 in molar ratio 1:1 was found to react at a much slower rate andrequires 60 min of milling to reach its completion. The occurring reactions are similar to the reactionsobserved for LiAlH4–LiNH2 also yielding Na3AlH6 as an intermediate [274].

Similar studies have been performed for samples of LiAlH4–NaNH2 (molar ratio 1:2, 1:1 and 2:1)and NaAlH4–LiNH2 (molar ratio 1:1) [274–276]. These studies all showed release of two equivalentsH2 during milling. For samples containing LiAlH4–NaNH2 the first reaction step is a cation-exchangereaction according to reaction scheme Eq. (45) starting already after 1 min of milling.

3LiAlH4 þ 4NaNH2 ! 3NaAlH4 þ Li3NaðNH2Þ4 ð45Þ

Upon further milling of LiAlH4–NaNH2 and NaAlH4–LiNH2 in molar ratio 1:1, LiNa2AlH6 is formed as anintermediate decomposition product [274]. This formation is facilitated by the presence of the amide,since neither of the pure alanates nor the mixture of the two could transform to AlH3�

6 containing com-pounds by ball milling alone [277,278]. The amides and alanates present in the sample at this stage areslowly consumed leading to formation of Al, LiH, NaH and an unknown compound, possibly AlN orLiAl0.33NH [274,275]. After only 30 min of milling no further changes are observed.

Furthermore, Dolotko et al. reported formation of a solid solution, Na3�xLixAlH6 in the NaAlH4–LiNH2 sample according to reaction scheme Eq. (46).

NaAlH4 þ ð2� xÞNaHþ xLiH! Na3�xLixAlH6 ð46Þ

The samples of LiAlH4–NaNH2 in molar ratio (1:2) and (2:1) exhibit different reaction pathways duringthe prolonged milling due to the excess of NaNH2 or LiAlH4, respectively [276]. The overall reactionssuggested to occur during ball milling for the (1:2) and (2:1) samples are shown in reaction schemesEqs. (47) and (48), respectively.

LiAlH4 þ 2NaNH2 ! 2NaHþ LiAlN2H2 þ 2H2 ð47Þ

2LiAlH4 þ NaNH2 ! NaAlH4 þ Li2AlNH2 þ 2H2 ð48Þ

Apparently, up to now, a mixed borohydride–alanate compound has not been prepared. The systemLiBH4—NaAlH4 was investigated mechanochemically and by hand-mixing, the latter treatment didnot lead to any reactions while a metathesis reaction was observed for the ball milled sample, seeEq. (49) [279,280].

LiBH4ðsÞ þ NaAlH4ðsÞ ! LiAlH4ðsÞ þ NaBH4ðsÞ ð49Þ

4.4. Solid–liquid mechanically assisted synthesis

In some cases addition of a solvent during mechanochemical synthesis is a necessity to promotethe desired reaction, e.g. Ca(AlH4)2 can be prepared by a so-called mechanically assisted synthesis


from NaAlH4 and CaCl2 mediated by tetrahydrofuran (THF) according to reaction scheme Eq. (50)[281].

2NaAlH4 þ CaCl2 þ 4THF! CaðAlH4Þ24THFþ 2NaCl ð50Þ

The synthesis reported by Fichtner et al. was performed in a glass ball mill reactor to create fresh par-ticle surfaces. The mixture was heated under reflux and the grey solid precipitate was milled by theglass balls. The high yield of the solvated product (84%) was attributed to the high dissolution capacityfor NaAlH4 of THF and that the adduct might form more easily compared to CaAlH4. After the synthesisthe solvent can be removed almost completely from freshly prepared Ca(AlH4)2�4THF under vacuumyielding a fine white powder of Ca(AlH4)2.

5. Final remarks and conclusions

Recent progress in the field of hydrogen storage materials has received extensive support frommechanochemistry methods. Mechanical milling has been widely used not only to tune metal micro-structures for modifying their hydrogenation properties but also as an efficient tool for the synthesisof hydrogen storage materials. Solid/solid, solid/liquid and solid/gas reactions can be activated bymechanochemistry.

Mechanical milling acts as a combination of compression and shear on the powder between twocolliding balls and between ball and container wall. Upon impact, the powder particles trapped be-tween them will first experience elastic deformations, which are reversible in the elastic region(Fig. 22). If the load increases, the material enters into plastic region and irreversible deformations oc-cur, which may be followed by breakage of the material. Therefore, collisions result in impulses ofcompression and shear that generate plastic deformation and fracture of the particles. Most of the en-ergy transferred to the powder is mainly used for creating new surfaces and concomitant particlerefining.

Very similar mechanical effects are produced by Severe Plastic Deformation (SPD) techniques suchas Equal Channel Angular Pressing (ECAP), Cold Rolling (CR), and High Pressure Torsion (HPT). There-fore, it could be expected that SPD will have similar impact on hydrogen storage behaviour thanmechanical milling.

In the particular case of solid/solid reactions, while comminution is an important result of milling,agglomeration assisted by cold-welding processes and thereby the formation of active interfaces ismore relevant. Solid-state reactions likely happen at the interface between solid particles of the dif-ferent reactants. The increase of the surface-area to bulk-volume ratio due to mechanically induceddecreasing particle size increases the interfacial contact area between reactive compounds and there-fore also the rate of reaction. In addition, plastic deformations due to mechanical work create defectsproviding fast atomic diffusion pathways and contributing to the increased reaction rate. The intensemixing action provided by the random motion of milled powders and repeated particle fracture andwelding favours the formation of active interfaces and ensures the chemical homogeneity of the finalproduct at long milling time.

As concerns solid/gas reactions, particle comminution and related effects such as fresh surface gen-eration and diffusion path reduction are determinant. Thus, Mechanical Milling under hydrogen gas ischaracterized by fast formation of hydride compounds under moderate pressure and temperature. Forinstance, several days are required for the synthesis of Mg2CoH5 and Mg2FeH6 hydrides by sinteringmethods at temperatures as high as 750 K, whereas the reaction takes place in only 3 h by reactive ballmilling under the same hydrogen pressure (9 MPa) [116,139]. Embrittlement due to hydrogen absorp-tion favours particle refinement and fast synthesis of binary metal hydrides. From this point of view,milling energy (determined by milling process parameters such as rotation speed and ball-to-powdermass ratio) and mechanical properties of reactants (toughness, fracture limit) must play a key role. Forthe synthesis of ternary hydrides starting from elemental powders, the above-mentioned mechanismsof cold-welding and interface diffusion of solid reactants should be also considered.

The feasibility of hydride formation in solid/gas reactions is governed by thermodynamics (i.e. hy-dride stability under external pressure and temperature conditions). Thus, the formation of highly sta-


ble hydrides such as TiH2, ZrH2 and MgH2 is straightforward. In contrast, the synthesis of hydrides thatare reversible near room temperature, such as LaNi5 and NaAlH4 requires high pressure for the hydrideto be stable at the temperatures reached during the milling process [121]. If the pressure is not high en-ough, only the a-solid solution LaNi5H0.15 or the more stable Na3AlH6 phase are observed, respectively.In this context, the failure to obtain the LiAlH4 or ternary alkaline hexa-alanates by reactive ball millingis not surprising since they are not stable at the usual operating pressure and temperature [121].

Therefore, in reactive ball-milling experiments, both macroscopic and local pressures as well astemperatures have to be considered in detail. As concerns pressure, one can distinguish the mechan-ical pressure in the material trapped between two colliding steel balls (internal mechanical pressure)and the gas vial pressure (external isostatic pressure). The gas pressure is not significantly affected bymilling due to the high compressibility of the gas phase. Most modern equipment allows for gas pres-sures up to 15 MPa. In contrast, the internal pressure of the material at mechanical impact may wellreach some GPa and is not isotropic. This would explain for instance the formation of the high pressurec-MgH2 phase by mechanical milling, which otherwise occurs in anvil cells above 2 GPa [282]. Forma-tion of c-MgH2 phase can be obtained by mechanical milling of thermodynamically stable b-MgH2

phase under argon atmosphere, which demonstrates that its formation is related to the internalmechanical pressure.

As concerns the temperature, one can also distinguish between the temperature of the materialtrapped between the milling tools (material temperature at impact) and the macroscopic materialtemperature. For materials with high thermal conductivity (i.e. metals) or by milling under hydrogengas (which also offers high thermal conductivity), the macroscopic material temperature is not ex-pected to differ much from the vial temperature. The temperature increase in the vial does not exceedsome tens of degrees and temperatures in the range 320–350 K have been monitored by temperaturegauges. This temperature increase is however non-negligible for hydrogen storage systems that arereversible near normal conditions of pressure and temperature. Several methods can be followed tocircumvent this problem such as working at higher hydrogen pressures and minimizing temperatureincrease by using short milling times (below �10 min). Another and more elegant alternative ap-proach is working at low temperatures, i.e. cryo-milling, though the temperature of the system hasto be kept high enough to allow for hydrogen mobility in the bulk powder material.

Mechanical milling of powders can be performed under other reactive gases such as diborane andammonia or in a liquid medium such as THF for the synthesis of hydrogen storage materials. Onceagain fresh surface generation induced by mechanical work is likely to be determinant to promote so-lid/gas and solid/liquid reactions. Contrary to reactive mechanical milling under hydrogen gas, thesepreparation methods are only recently explored in the literature. Further progress is still needed tounderstand the involved reaction mechanisms and the feasibility of compound formation by these no-vel routes. A wide research field remains open for the production of new hydrogen storage systemswith undiscovered hydrogenation properties.

Acknowledgments

We thank V. Balema for fruitful discussions regarding this review. The work was supported in partby the Danish National Research Foundation (Center for Materials Crystallography), the Danish Stra-tegic Research Council (Center for Energy Materials and the HyFillFast project), and by the Danish Re-search Council for Nature and Universe (Danscatt). We are grateful to the Carlsberg Foundation. J.H.would like to thank the Natural Science and Engineering Council of Canada and also the ResearchCouncil of Norway for additional funding that permitted a sabbatical leave at the Institute for EnergyTechnology (IFE) in Norway. ML, FC and JZ would like to thank CNRS and the French agency ANR forfinancial supports trough research programs ALHAMO and NANOHYDLI.

References

[1] Grochala W, Edwards PP. Thermal decomposition of the non-interstitial hydrides for the storage and production ofhydrogen. Chem Rev 2004;104:1283–315.


[2] Orimo S, Nakamori Y, Eliseo JR, Zuttel A, Jensen CM. Complex hydrides for hydrogen storage. Chem Rev2007;107:4111–32.

[3] Ravnsbæk DB, Filinchuk Y, Cerny R, Jensen TR. Powder diffraction methods for studies of borohydride-based energystorage materials. Z Kristall 2010;225:557–69.

[4] Rude LH, Nielsen TK, Ravnsbæk DB, Bösenberg U, Ley MB, Richter B, et al. Tailoring properties of borohydrides forhydrogen storage: a review. Phys Status Solidi (a) 2011;208:1754–73.

[5] Song MY, Ivanov EI, Darriet B, Pezat M, Hagenmuller P. Hydriding properties of a mechanically alloyed mixture with acomposition Mg2Ni. Int J Hydrogen Energy 1985;10:169–78.

[6] Harris JH, Curtin WA, Schultz L. Hydrogen storage characteristics of mechanically alloyed amorphous metals. J Mater Res1988;3:872–83.

[7] Aoki K, Aoyagi H, Memezawa A, Masumoto T. Effect of ball milling on the hydrogen absorption rate of FeTi and Mg2Nicompounds. J Alloys Compd 1994;203:L7–9.

[8] Zaluski L, Zaluska A, Tessier P, Strom-Olsen JO, Schulz R. Hydrogen absorption by nanocrystalline and amorphous Fe–Tiwith palladium catalyst, produced by ball milling. J Mater Sci 1996;31:695–8.

[9] Ares JR, Cuevas F, Percheron-Guégan A. Mechanical milling and subsequent annealing effects on microstructural andhydrogenation properrties of multisubstituted LaNi5 alloy. Acta Mater 2005;53:2157–67.

[10] Dornheim M, Eigen N, Barkhordarian G, Klassen T, Bormann R. Tailoring hydrogen storage materials towards application.Adv Eng Mater 2006;8:377–85.

[11] Varin RA, Czujko T, Wronski ZS. Nanomaterials for solid state hydrogen storage. Springer; 2009.[12] Varin RA, Zbroniec L, Polanski M, Bystrzycki J. A review of recent advances on the effects of microstructural refinement

and nano-catalytic additives on the hydrogen storage properties of metal and complex hydrides. Energies 2011;4:1–25.[13] Skripnyuk V, Buchman E, Rabkin E, Estrin Y, Popov M, Jorgensen S. The effect of equal channel angular pressing on

hydrogen storage properties of a eutectic Mg–Ni alloy. J Alloys Compd 2007;436:99–106.[14] Skripnyuk V, Rabkin E, Estrin Y, Lapovok R. The effect of ball milling and equal channel angular pressing on hydrogen

absorption/desorption properties of Mg–4.95 wt% Zn–0.71wt%Zr (ZK60) alloy. Acta Mater 2004;52:405–14.[15] Skripnyuk VM, Rabkin E, Estrin Y, Lapovok R. Improving hydrogen storage properties of magnesium based alloys by equal

channel angular pressing. Int J Hydrogen Energy 2009;34:6320–4.[16] Lima GF, Jorge AM, Leiva DR, Kiminami CS, Bolfarini C, Botta WJ. Severe plastic deformation of Mg–Fe powders to produce

bulk hydrides. In: Schultz L, Eckert J, Battezzati L, Stoica M, editors. 13th International conference on rapidly quenchedand metastable materials. Bristol: Iop Publishing Ltd.; 2009. p. 12015 [art. no. 012015].

[17] Kusadome Y, Ikeda K, Nakamori Y, Orimo S, Horita Z. Hydrogen storage capability of MgNi2 processed by high pressuretorsion. Scr Mater 2007;57:751–3.

[18] Ueda TT, Tsukahara M, Kamiya Y, Kikuchi S. Preparation and hydrogen storage properties of Mg–Ni–Mg2Ni laminatecomposites. J Alloys Compd 2004;386:253–7.

[19] Tanaka K, Takeichi N, Tanaka H, Kuriyama N, Ueda TT, Tsukahara M, et al. Investigation of micro-structural transitionthrough disproportionation and recombination during hydrogenation and dehydrogenation in Mg/Cu super-laminates. JMater Sci 2008;43:3812–6.

[20] Takeichi N, Tanaka K, Tanaka H, Ueda TT, Kamiya Y, Tsukahara M, et al. The hydrogen storage properties of Mg/Cu and Mg/Pd laminate composites and metallographic structure. J Alloys Compd 2007;446–447:543–8.

[21] Mori R, Miyamura H, Kikuchi S, Tanaka K, Takeichi N, Tanaka H, et al. Hydrogenation characteristics of Mg based alloyprepared by super lamination technique. Mater Sci Forum 2007;561–565:1609–12.

[22] Suganuma K, Miyamura H, Kikuchi S, Takeichi N, Tanaka K, Tanaka H, et al. Hydrogen storage properties of Mg–Al alloyprepared by super lamination technique. Adv Mater Res 2007;26–28:857–60.

[23] Dufour J, Huot J. Rapid activation, enhanced hydrogen sorption kinetics and air resistance in laminated Mg–Pd2.5at.%. JAlloys Compd 2007;439:L5–7.

[24] Dufour J, Huot J. Study of Mg6Pd alloy synthesized by cold rolling. J Alloys Compd 2007;446–447:147–51.[25] Zhang LT, Ito K, Vasudevan VK, Yamaguchi M. Hydrogen absorption and desorption in a B2 single-phase Ti–22Al–27Nb

alloy before and after deformation. Acta Mater 2001;49:751–8.[26] Zhang LT, Ito K, Vasudevan VK, Yamaguchi M. Effects of cold-rolling on the hydrogen absorption/desorption behavior of

Ti–22Al–27Nb alloys. Mater Sci Eng A 2002;329–331:362–6.[27] Aoki K, Memezawa A, Masumoto T. Atmosphere effects of the amorphization reaction in NiZr by ball milling. J Mater Res

1993;8:307–13.[28] Orimo S, Fujii H, Yoshino T. Reactive mechanical grinding of ZrNi under various partial pressures of hydrogen. J Alloys

Compd 1995;217:287–94.[29] Huot J, Akiba E, Takada T. Mechanical alloying of Mg–Ni compounds under hydrogen and inert atmosphere. J Alloys

Compd 1995;231:815–9.[30] Bellosta von Colbe JM, Felderhoff M, Bogdanovic B, Schuth F, Weidenthaler C. One-step direct synthesis of a Ti-doped

sodium alanate hydrogen storage material. Chem Commun 2005:4732–4.[31] Doppiu S, Schultz L, Gutfleisch O. In situ pressure and temperature monitoring during the conversion of Mg into MgH2 by

high-pressure reactive ball milling. J Alloys Compd 2007;427:204–8.[32] Evicomagnetics.[33] Suryanarayana C. Mechanical alloying and milling. Prog Mater Sci 2001;46:1–184.[34] Fritsch. Product information ‘‘planetary mills classic line’’.[35] Schilz J. Internal kinematics of tumbler and planetary ball mills: a mathematical model for the parameter setting. Mater

Trans JIM 1998;39:1152–7.[36] Hagemann H, Cerny R. Synthetic approaches to inorganic borohydrides. Dalton Trans 2010;39:6006–12.[37] Ravnsbæk D, Sørensen LH, Filinchuk Y, Besenbacher F, Jensen TR. Screening of metal borohydrides by mechano-chemistry

and diffraction. Angew Chem Int Edit 2012;51:3582–6.[38] Ravnsbæk DB, Ley MB, Lee Y-S, Hagemann H, Cho YW, Filinchuk Y, et al. A mixed-cation mixed-anion borohydride

NaY(BH4)2Cl2. Int J Hydrogen Energy 2012;37:8428–38.


[39] Ravnsbæk DB, Nickels EA, Olesen CH, David WIF, Edwards PP, Filinchuk Y, et al. Novel alkali earth borohydride Sr(BH4)2and borohydride/chloride Sr(BH4)Cl, 2012, submitted for publication.

[40] Cerny R, Penin N, Hagemann H, Filinchuk Y. The first crystallographic and spectroscopic characterization of a 3d-metalborohydride: Mn(BH4)2. J Phys Chem C 2009;113:9003–7.

[41] Sato T, Miwa K, Nakamori Y, Ohoyama K, Li HW, Noritake T, et al. Experimental and computational studies on solvent-freerare-earth metal borohydrides R(BH4)3 (R = Y, Dy, and Gd). Phys Rev B 2008;77.

[42] Ravnsbæk DB, Frommen C, Reed D, Filinchuk Y, Sørby M, Hauback BC, et al. Structural studies of lithium zinc borohydrideby neutron powder diffraction, Raman and NMR spectroscopy. J Alloys Compd 2011;509:S698–704.

[43] Aoki K, Memezawa A, Masumoto T. Amorphization of the TiV system by mechanical alloying and mechanical grinding in ahydrogen and nitrogen atmosphere. J Mater Res 1994;9:39–46.

[44] Zhang H, Kisi EH. Formation of titanium hydride at room temperature by ball milling. J Phys: Condens Matter1997;9:L185. L90.

[45] Cheng ZH, MacKay GR, Small DA, Dunlap RA. Phase development in titanium by mechanical alloying under hydrogenatmosphere. J Phys D – Appl Phys 1999;32:1934–7.

[46] Bobet J-L, Even C, Nakamura Y, Akiba E, Darriet B. Synthesis of magnesium and titanium hydride via reactive mechanicalalloying. Influence of 3d-metal addition on MgH2 synthesize. J Alloys Compd 2000;298:279–84.

[47] Orimo S, Kimmerle F, Majer G. Hydrogen in nanostructured vanadium–hydrogen systems. Phys Rev B 2001;63:094307.[48] Chen J, Takeshita HT, Chartouni D, Kuriyama N, Sakai T. Synthesis and characterization of nanocrystalline Mg2CoH5

obtained by mechanical alloying. J Mater Sci 2001;36:5829–34.[49] Gennari FC, Castro FJ, Gamboa JJA. Synthesis of Mg2FeH6 by reactive mechanical alloying: formation and decomposition

properties. J Alloys Compd 2002;339:261–7.[50] Herrich M, Ismail N, Lyubina J, Handstein A, Pratt A, Gutfleisch O. Synthesis and decomposition of Mg2FeH6 prepared by

reactive milling. Mater Sci Eng B 2004;108:28–32.[51] Baum LA, Meyer M, Mendoza-Zélis L. Complex Mg-based hydrides obtained by mechanosynthesis: characterization and

formation kinetics. Int J Hydrogen Energy 2008;33:3442–6.[52] Deledda S, Hauback BC. The formation mechanism and structural characterization of the mixed transition-metal complex

hydride Mg2(FeH6)0.5(CoH5)0.5 obtained by reactive milling. Nanotechnology 2009;20:7.[53] Eigen N, Kunowsky M, Klassen T, Bormann R. Synthesis of NaAlH4-based hydrogen storage material using milling under

low pressure hydrogen atmosphere. J Alloys Compd 2007;430:350–5.[54] Kang XD, Wang P, Cheng HM. In situ formation of Ti hydride and its catalytic effect in doped NaAlH4 prepared by milling

NaH/Al with metallic Ti powder. Int J Hydrogen Energy 2007;32:2943–8.[55] Wang J, Ebner AD, Ritter JA. Synthesis of metal complex hydrides for hydrogen storage. The J Phys Chem C

2007;111:14917–24.[56] Jeloaica L, Zhang J, Cuevas F, Latroche M, Raybaud P. Thermodynamic properties of trialkali (Li, Na, K) hexa-alanates: a

combined DFT and experimental study. J Phys Chem C 2008;112:18598–607.[57] Zhang JX, Pilette MA, Cuevas F, Charpentier T, Mauri F, Latroche M. X-ray diffraction and NMR studies of Na3�nLinAlH6

(n = 0, 1, 2) alanates synthesized by high-pressure reactive ball milling. J Phys Chem C 2009;113:21242–52.[58] Rongeat C, D’Anna V, Hagemann H, Borgschulte A, Zuttel A, Schultz L, et al. Effect of additives on the synthesis and

reversibility of Ca(BH4)2. J Alloys Compd 2010;493:281–7.[59] Ichikawa T, Isobe S, Hanada N, Fujii H. Lithium nitride for reversible hydrogen storage. J Alloys Compd 2004;365:271–6.[60] Minella CB, Rongeat C, Domenech-Ferrer R, Lindemann I, Dunsch L, Sorbie N, et al. Synthesis of LiNH2 + LiH by reactive

milling of Li3N. Faraday Discuss 2011;151:253–62.[61] Mulana F, Nishimiya N. Application of mechanical milling to synthesize a novel quaterly hydride. J Alloys Compd

2006;413:273–80.[62] Czujko T, Varin RA, Wronski Z, Zaranski Z, Durejko T. Synthesis and hydrogen desorption properties of nanocomposite

magnesium hydride with sodium borohydride (MgH2 + NaBH4). J Alloys Compd 2007;427:291–9.[63] Srinivasan SS, Niemann MU, Hattrick-Simpers JR, McGrath K, Sharma PC, Goswami DY, et al. Effects of nano additives on

hydrogen storage behavior of the multinary complex hydride LiBH4/LiNH2/MgH2. Int J Hydrogen Energy2010;35:9646–52.

[64] Cuevas F, Korablov D, Latroche M. ynthesis, structural and hydrogenation properties of Mg-rich MgH2–TiH2

nanocomposites prepared by reactive ball milling under hydrogen gas. Phys Chem Chem Phys 2012;14:1200–11.[65] Chaudhuri S, Graetz J, Ignatov A, Reilly JJ, Muckerman JT. Understanding the role of Ti in reversible hydrogen storage as

sodium alanate: a combined experimental and density functional theoretical approach. J Am Chem Soc2006;128:11404–15.

[66] Bösenberg U, Kim JW, Gosslar D, Eigen N, Jensen TR, von Colbe JMB, et al. Role of additives in LiBH4–MgH2 reactivehydride composites for sorption kinetics. Acta Mater 2010;58:3381–9.

[67] Gosalawit-Utke R, von Colbe JMB, Dornheim M, Jensen TR, Cerenius Y, Minella CB, et al. LiF–MgB2 system for reversiblehydrogen storage. J Phys Chem C 2010;114:10291–6.

[68] Dornheim M. Tailoring reaction enthalpies of hydrides. In: Hirscher M, editor. Handbook of hydrogen storage: newmaterials for future energy storage. John Wiley & Sons; 2010.

[69] Fukai Y. The metal-hydrogen system, 2nd ed. Berlin: Springer-Verlag; 2005. p. 497 [Springer Series in Materials Science].[70] Akiba E, Okada M. Metallic hydrides III: body-centered-cubic solid-solution alloys. MRS Bulletin. 2002;27:699–703.[71] Nakamura Y, Akiba E. Hydriding properties and crystal structure of NaCl-type mono-hydrides formed from Ti–V–Mn BCC

solid solutions. J Alloys Compd 2002;345:175–82.[72] Kubo K, Itoh H, Takahashi T, Ebisawa T, Kabutomori T, Nakamura Y, et al. Hydrogen absorbing properties and structures of

Ti–Cr–Mo alloys. J Alloys Compd 2003;356–357:452–5.[73] Iwase K, Nakamura Y, Mori K, Harjo S, Ishigaki T, Kamiyama T, et al. Hydrogen absorption–desorption properties and

crystal structure analysis of Ti–Cr–Mo alloys. J Alloys Compd 2005;404–406:99–102.[74] Shibuya M, Nakamura J, Akiba E. Hydrogenation properties and microstructure of Ti–Mn-based alloys for hybrid

hydrogen storage vessel. J Alloys Compd 2008;466:558–62.


[75] Shibuya M, Nakamura J, Enoki H, Akiba E. High-pressure hydrogenation properties of Ti–V–Mn alloy for hybrid hydrogenstorage vessel. J Alloys Compd 2009;475:543–5.

[76] Yu XB, Wu Z, Xia BJ, Xu NX. Improvement of activation performance of quenched Ti–V-based BCC phase alloys. J AlloysCompd 2004;386:258–60.

[77] Huot J, Enoki H, Akiba E. Synthesis, phase transformation, and hydrogen storage properties of ball-milled TiV0.9Mn1.1. JAlloys Compd 2008;453:203–9.

[78] Eckert J, Holzer III JC, Johnson WL. Mechanically driven alloying and grain size changes in nanocrystalline Fe–Cu powders.J Appl Phys. 1992;73:2794–802.

[79] Santos SF, Costa ALM, Castro JFRd, Santos DSd, Botta WJ, Ishikawa TT. Mechanical and reactive miling of a TiCrV BCC solidsolution. J Metastable Nanocrystal Mater 2004;20–21:291–6.

[80] Singh BK, Shim G, Cho S-W. Effects of mechanical milling on hydrogen storage properties of Ti0.32Cr0.43V0.25 alloy. Int JHydrogen Energy 2007;32:4961–5.

[81] Amira S, Santos SF, Huot J. Hydrogen sorption properties of Ti–Cr alloys synthesized by ball milling and cold rolling.Intermetallics 2010;18:140–4.

[82] Couillaud S, Enoki H, Amira S, Bobet JL, Akiba E, Huot J. Effect of ball milling and cold rolling on hydrogen storageproperties of nanocrystalline TiV1.6Mn0.4 alloy. J Alloys Compd 2009;484:154–8.

[83] Asano K, Enoki H, Akiba E. Synthesis of HCP, FCC and BCC structure alloys in the Mg–Ti binary system by means of ballmilling. J Alloys Compd 2009;480:558–63.

[84] Asano K, Enoki H, Akiba E. Synthesis of Mg–Ti FCC hydrides from Mg–Ti BCC alloys. J Alloys Compd 2009;478:117–20.[85] Zhang Y, Tsushio Y, Enoki H, Akiba E. The hydrogen absorption–desorption performances of Mg–Co–X ternary alloys with

BCC structure. J Alloys Compd 2005;393:185–93.[86] Zhang Y, Tsushio Y, Enoki H, Akiba E. The study on binary Mg–Co hydrogen storage alloys with BCC phase. J Alloys Compd

2005;393:147–53.[87] Shao HY, Asano K, Enoki H, Akiba E. Fabrication, hydrogen storage properties and mechanistic study of nanostructured

Mg50Co50 body-centered cubic alloy. Scr Mater 2009;60:818–21.[88] Shao H, Asano K, Enoki H, Akiba E. Preparation and hydrogen storage properties of nanostructured Mg–Ni BCC alloys. J

Alloys Compd 2009;477:301–6.[89] Shao H, Asano K, Enoki H, Akiba E. Fabrication and hydrogen storage property study of nanostructured Mg–Ni–B ternary

alloys. J Alloys Compd 2009;479:409–13.[90] Niessen RAH, Notten PHL. Electrochemical hydrogen storage characteristics of thin film MbX (X = Sc, Ti, V, Cr) compounds.

Electrochem Solid-State Lett 2005;8:A534. A8.[91] Kalisvaart WP, Notten PHL. Mechanical alloying and electrochemical hydrogen storage of Mg-based systems. J Mater Res

2008;23:2179–87.[92] Asano K, Akiba E. Direct synthesis of Mg–Ti–H FCC hydrides from MgH2 and Ti by means of ball milling. J Alloys Compd

2009;481:L8–L11.[93] Borsa DM, Baldi A, Pasturel M, Schreuders H, Dam B, Griessen R, et al. Mg–Ti–H thin films for smart solar collectors. Appl

Phys Lett 2006;88:241910.[94] Borsa DM, Gremaud R, Baldi A, Schreuders H, Rector JH, Kooi B, et al. Structural, optical, and electrical properties of

MgyTi1�yHx thin films. Phys Rev B 2007;75:205408.[95] Slaman M, Dam B, Schreuders H, Griessen R. Optimization of Mg-based fiber optic hydrogen detectors by alloying the

catalyst. Int J Hydrogen Energy 2008;33:1084–9.[96] Vermeulen P, Niessen RAH, Notten PHL. Hydrogen storage in metastable MgyTi(1�y) thin films. Electrochem Commun

2006;8:27–32.[97] Vermeulen P, Thiel EFMJv, Notten PHL PHL. Ternary MgTiX-alloys: a promising route toward low-temperature, high-

capacity, hydrogen-storage materialsthin films. Chemistry: A Eur J 2007;13:9892–8.[98] Rousselot S, Bichat MP, Guay D, Roué L. Structure and electrochemical behaviour of metastable Mg50Ti50 alloy prepared by

ball milling. J Power Sour 2008;175:621–4.[99] Rousselot S, Guay D, Roué L. Synthesis of fcc Mg–Ti–H alloys by high energy ball milling: Structure and electrochemical

hydrogen storage properties. J Power Sour 2010;195:4370–4.[100] Asano K, Enoki H, Akiba E. Effect of Li addition on synthesis of Mg–Ti BCC alloys by means of ball milling. Mater Trans JIM

2007;48:121–6.[101] Asano K, Enoki H, Akiba E. Synthesis process of Mg–Ti BCC alloys by means of ball milling. J Alloys Compd

2009;486:115–23.[102] Çakmak G, Károly Z, Mohai I, Öztürk T, Szépvölgyi J. The processing of Mg–Ti for hydrogen storage; mechanical milling

and plasma synthesis. Int J Hydrogen Energy 2010;35:10412–8.[103] Maweja K, Phasha M, van der Berg N. Microstructure and crystal structure of an equimolar Mg–Ti alloy processed by

Simoloyer high-energy ball mill. Powder Technol 2010;199:256–63.[104] Rousselot S, Gazeau A, Guay D, Roué L. Influence of Pd on the structure and electrochemical hydrogen storage properties

of Mg50Ti50 alloy prepared by ball milling. Electrochim Acta 2010;55:611–9.[105] Ma E. Amorphization in mechanically driven material systems. Scr Mater 2003;49:941–6.[106] Rojas P, Ordoñez S, Serafini D, Zúñiga A, Lavernia E. Microstructural evolution during mechanical alloying of Mg and Ni. J

Alloys Compd 2005;391:267–76.[107] Fecht H-J, Hellstern E, Fu Z, Johnson WL. Nanocrystalline metals prepared by high-energy ball milling. Metall Trans A

1990;21A:2333–7.[108] Abdellaoui M, Mokbli S, Cuevas F, Latroche M, Percheron-Guégan A, Zarrouk H. Structural and electrochemical properties

of amorphous rich MgxNi100�x nanomaterial obtained by mechanical alloying. J Alloys Compd 2003;356–357:557–61.[109] Varin RA, Czujko T, Mizera J. The effect of MgNi2 intermetallic compound on nanostructurization and amorphization of

Mg–Ni alloys processed by controlled mechanical milling. J Alloys Compd 2003;354:281–95.[110] Varin RA, Czujko T, Mizera J. Microstructural evolution during controlled ball milling of (Mg2Ni + MgNi2) intermetallic

alloy. J Alloys Compd 2003;350:332–9.


[111] Sherif El-Eskandarany M, Aoki K, Sumiyama K, Suzuki K. Cyclic phase transformations of mechanically alloyed Co75Ti25

powders. Acta Mater 2002;50:1113–23.[112] Dunlap RA, Cheng ZH, MacKay GR, O’Brien JW, Small DA. Preparation of nanocrystalline metal hydrides by ball milling.

Hyperfine Inter 2000;130:109–26.[113] Bellosta von Colbe JM, Bogdanovic B, Felderhoff M, Pommerin A, Schüth F. Recording of hydrogen evolution – a way for

controlling the doping process of sodium alanate by ball milling. J Alloys Compd 2004;370:104–9.[114] Chiang C-H, Chin Z-H, Perng T-P. Hydrogenation of TiFe by high-energy ball milling. J Alloys Compd 2000;307:259–65.[115] Raman SSS, Davidson DJ, Bobet J-L, Srivastava ON. Investigations on the synthesis, structural and microstructural

characterizations of Mg-based K2PtCl6 type (Mg2FeH6) hydrogen storage material prepared by mechanical alloying. JAlloys Compd 2002;333:282–90.

[116] Zhang JX, Cuevas F, Zaidi W, Bonnet JP, Aymard L, Bobet JL, et al. Highlighting of a single reaction path during reactive ballmilling of Mg and TM by quantitative H(2) gas sorption analysis to form ternary complex hydrides (TM = Fe, Co., Ni). JPhys Chem C 2011;115:4971–9.

[117] Chen Y, Williams JS. Formation of metal hydrides by mechanical alloying. J Alloys Compd 1995;217:181–4.[118] Huot J, Tremblay ML, Schulz R. Synthesis of nanocrystalline hydrogen storage materials. J Alloys Compd 2003;356–

357:603–7.[119] Eleskandarany MS, Ahmed HA, Sumiyama K, Suzuki K. Mechanically assisted solid-state hydrogenation for formation of

nanocrystalline NiTH3 alloy powder. J Alloys Compd 1995;218:36–43.[120] Bobet J-L, Chevalier B. Reactive mechanical grinding applied to a (Ti+Ni) mixture and to a TiNi compound. Intermetallics

2002;10:597–601.[121] Fujii H, Munehiro S, Fujii K, Orimo S. Effect of mechanical grinding under Ar and H2 atmospheres on structural and

hydriding properties in LaNi5. J Alloys Compd 2002;330–332:747–51.[122] Zhang J, Cuevas F, Latroche M. Unpublished results; 2012.[123] Bogdanovic B, Bohmhammel K, Christ B, Reiser A, Schlichte K, Vehlen R, et al. Thermodynamic investigation of the

magnesium–hydrogen system. J Alloys Compd 1999;282:84–92.[124] Gerard N, Ono S. Hydride formation and decomposition kinetics. In: Schlapbach L, editor. Hydrogen in intermetallic

compounds II. Berlin: Springer Verlag; 1992 [Chapter 4].[125] Sastri MVC, Viswanathan B, Murphy SS. In: Sastri MVC, editor. Metal hydrides-fundamentals and applications. New

Delhi: Narosa Publishing House; 1998. p. 201.[126] Tessier P, Akiba E. Catalysed reactive milling. J Alloys Compd 1999;293–295:400–2.[127] Gennari FC, Castro FJ, Urretavizcaya G. Hydrogen desorption behavior from magnesium hydrides synthesized by reactive

mechanical alloying. J Alloys Compd 2001;321:46–53.[128] Huot J, Swainson I, Schulz R. Phase transformation in magnesium hydride induced by ball milling. Annu Chim Sci Mat

2006;31:135–44.[129] Doppiu S, Solsona P, Spassov T, Barkhordarian G, Dornheim M, Klassen T, et al. Thermodynamic properties and

absorption–desorption kinetics of Mg87Ni10Al3 alloy synthesised by reactive ball milling under H2 atmosphere. J AlloysCompd 2005;404:27–30.

[130] Burch R, Mason NB. Absorption of hydrogen by titanium–cobalt and titanium–nickel intermetallic alloys. 1 Experimentalresults. J Chem Soc – Faraday Trans I 1979;75:561–77.

[131] Pundt A, Kirchheim R. Hydrogen in metals: microstructural aspects. Annu Rev Mater Res 2006;36:555–608.[132] Harris JH, Curtin WA, Tenhover MA. Universal features of hydrogen absorption in amorphous transition-metal alloys.

Phys Rev B 1987;36:5784–97.[133] Huot J, Hayakawa H, Akiba E. Preparation of the hydrides Mg2FeH6 and Mg2CoH5 by mechanical alloying followed by

sintering. J Alloys Compd 1997;248:164–7.[134] Huot J, Boily S, Akiba E, Schulz R. Direct synthesis of Mg2FeH6 by mechanical alloying. J Alloys Compd 1998;280:306–9.[135] Fernández IG, Meyer GO, Gennari FC. Hydriding/dehydriding behavior of Mg2CoH5 produced by reactive mechanical

milling. J Alloys Compd 2008;464:111–7.[136] Orimo S, Fujii H, Ikeda K. Notable hydriding properties of a nanostructured composite material of the Mg2Ni–H system

synthesized by reactive mechanical grinding. Acta Mater 1997;45:331–41.[137] Tessier P, Enoki H, Bououdina M, Akiba E. Ball-milling of Mg2Ni under hydrogen. J Alloys Compd 1998;268:285–9.[138] Yvon K. Complex transition-metal hydrides. Chimia 1998;52:613–9.[139] Selvam P, Yvon K. Synthesis of Mg2FeH6, Mg2CoH5 and Mg2NiH4 by high-pressure sintering of the elements. Int J

Hydrogen Energy 1991;16:615–7.[140] Bogdanovic B, Reiser A, Schlichte K, Spliethoff B, Tesche B. Thermodynamics and dynamics of the Mg–Fe–H system and its

potential for thermochemical thermal energy storage. J Alloys Compd 2002;345:77–89.[141] Polanski M, Nielsen TK, Cerenius Y, Bystrzycki J, Jensen TR. Synthesis and decomposition mechanisms of Mg2FeH6

studied by in situ synchrotron X-ray diffraction and high-pressure DSC. Int J Hydrogen Energy 2010;35:3578–82.[142] Castro JF, Gennari FC. Effect of the nature of the starting materials on the formation of Mg2FeH6. J Alloys Compd

2004;375:292–6.[143] Varin RA, Li S, Calka A, Wexler D. Formation and environmental stability of nanocrystalline and amorphous hydrides in

the 2Mg–Fe mixture processed by controlled reactive mechanical alloying (CRMA). J Alloys Compd 2004;373:270–86.[144] Friedrichs O, Sanchez-Lopez JC, Lopez-Cartes C, Dornheim M, Klassen T, Bormann R, et al. Chemical and microstructural

study of the oxygen passivation behaviour of nanocrystalline Mg and MgH2. Appl Surface Sci 2006;252:2334–45.[145] Liang G, Huot J, Boily S, Neste AV, Schulz R. Catalytic effect of transition metals on hydrogen sorption in nanocrystalline

ball milled MgH2–Tm (Tm = Ti, V, Mn, Fe and Ni) systems. J Alloys Compd 1999;292:247–52.[146] Shao H, Liu T, Wang Y, Xu H, Li X. Preparation of Mg-based hydrogen storage materials from metal nanoparticles. J Alloys

Compd 2008;465:527–33.[147] Polanski M, Plocinski T, Kunce I, Bystrzycki J. Dynamic synthesis of ternary Mg2FeH6. Int J Hydrogen Energy

2010;35:1257–66.


[148] Zhou DW, Li SL, Varin RA, Peng P, Liu JS, Yang F. Mechanical alloying and electronic simulations of 2Mg–Fe mixturepowders for hydrogen storage. Mater Sci Eng: A 2006;427:306–15.

[149] Lima GF, Peres MM, Garroni S, Baró MD, Surinyach S, Kiminami CS, et al. Microstructural characterization andhydrogenation study of extruded MgFe alloy. J Alloys Compd 2010;504:S299–301.

[150] Shao H, Xu H, Wang Y, Li X. Synthesis and hydrogen storage behavior of Mg–Co–H system at nanometer scale. J Solid StateChem 2004;177:3626–32.

[151] Gavra Z, Mintz MH, Kimmel G, Hadari Z. Allotropic transitions of Mg2NiH4. Inorg Chem 1979;18:3595.[152] Zolliker P, Yvon K, Baerlocher C. Low-temperature structure of Mg2NiH4: evidence for microtwinning. J Less-Common

Metals 1986;115:65–78.[153] Rönnebro E, Jensen JO, Noréus D, Bjerrum NJ. Structural studies of disordered Mg2NiH4 formed by mechanical grinding. J

Alloys Compd 1999;293–295:146–9.[154] Bogdanovic B, Schwickardi M. Ti-doped alkali metal aluminium hydrides as potential novel reversible hydrogen storage

materials. J Alloys Compd 1997;253–254:1–9.[155] Graetz J. New approaches to hydrogen storage. Chem Soc Rev 2009;38:73–82.[156] Dymova TN, Eliseeva NG, Bakum SI, Dergachev YM. Direct synthesis of alkali metal aluminum hydrides in the melt. Dokl

Akad Nauk SSSR 1974;215:1369–72 [Engl. 256–259].[157] Huot J, Boily S, Güther V, Schulz R. Synthesis of Na3AlH6 and Na2LiAlH6 by mechanical alloying. J Alloys Compd

1999;283:304–6.[158] Zaluski L, Zaluska A, Strom-Olsen JO. Hydrogenation properties of complex alkali metal hydrides fabricated by mechano-

chemical synthesis. J Alloys Compd 1999;290:71–8.[159] Xiao X, Yu K, Fan X, Wu Z, Wang X, Chen C, et al. Synthesis and hydriding/dehydriding properties of nanosized sodium

alanates prepared by reactive ball-milling. Int J Hydrogen Energy 2011;36:539–48.[160] Wang P, Kang XD, Cheng HM. Direct formation of Na3AlH6 by mechanical milling NaH/Al with TiF3. Appl Phys Lett

2005;87:071911–71913.[161] Mamatha M, Weidenthaler C, Pommerin A, Felderhoff M, Schüth F. Comparative studies of the decomposition of alanates

followed by in situ XRD and DSC methods. J Alloys Compd 2006;416:303–14.[162] Iosub V, Matsunaga T, Tange K, Ishikiriyama M. Direct synthesis of Mg(AlH4)2 and CaAlH5 crystalline compounds by ball

milling and their potential as hydrogen storage materials. Int J Hydrogen Energy 2009;34:906–12.[163] Kabbour H, Ahn CC, Hwang SJ, Bowman RC, Graetz J. Direct synthesis and NMR characterization of calcium alanate. J

Alloys Compd 2007;446:264–6.[164] Liu XF, Asano K, Akiba E. Direct synthesis of BaAlH5 and Ba2AlH7 from BaH2 and Al system and their hydriding/

dehydriding characteristics. J Alloys Compd 2009;477:744–8.[165] Kojima Y, Kawai Y, Haga T, Matsumoto M, Koiwai A. Direct formation of LiAlH4 by a mechanochemical reaction. J Alloys

Compd 2007;441:189–91.[166] Balema VP, Wiench JW, Dennis KW, Pruski M, Pecharsky VK. Titanium catalyzed solid-state transformations in LiAlH4

during high-energy ball-milling. J Alloys Compd 2001;329:108–14.[167] Dymova TN, Konoplev VN, Aleksandrov DP, Sizareva AS, Silina TA. A novel view of the nature of chemical- and phase-

composition modification in lithium hydroaluminates LiAlH4 and Li3AlH6 on heating. Koordinatsionnaya Khimiya1995;21:175–82.

[168] Graetz J, Wegrzyn J, Reilly JJ. Regeneration of lithium aluminum hydride. J Am Chem Soc 2008;130:17790–4.[169] Wang J, Ebner AD, Ritter JA. Physiochemical pathway for cyclic dehydrogenation and rehydrogenation of LiAlH4. J Am

Chem Soc 2006;128:5949–54.[170] Eigen N, Keller C, Dornheim M, Klassen T, Bormann R. Industrial production of light metal hydrides for hydrogen storage.

Scr Mater 2007;56:847–51.[171] Bogdanovic B, Felderhoff M, Pommerin A, Schüth F, Spielkamp N, Stark A. Cycling properties of Sc- and Ce-doped NaAlH4

hydrogen storage materials prepared by the one-step direct synthesis method. J Alloys Compd 2009;471:383–6.[172] Zidan RA, Takara S, Hee AG, Jensen CM. Hydrogen cycling behavior of zirconium and titanium–zirconium-doped sodium

aluminum hydride. J Alloys Compd 1999;285:119–22.[173] Bogdanovic B, Brand RA, Marjanovic A, Schwickardi M, Tölle J. Metal-doped sodium aluminium hydrides as potential new

hydrogen storage materials. J Alloys Compd 2000;302:36–58.[174] Ares JR, Aguey-Zinsou KF, Leardini F, Ferrer IJ, Fernandez JF, Guo ZX, et al. Hydrogen absorption/desorption mechanism in

potassium alanate (KAlH4) and enhancement by TiCl3 doping. J Phys Chem C 2009;113:6845–51.[175] Wang FH, Liu YF, Gao MX, Luo K, Pan HG, Wang QD. formation reactions and the thermodynamics and kinetics of

dehydrogenation reaction of mixed alanate Na2LiAlH6. J Phys Chem C 2009;113:7978–84.[176] Sorby MH, Brinks HW, Fossdal A, Thorshaug K, Hauback BC. The crystal structure and stability of K2NaAlH6. J Alloys

Compd 2006;415:284–7.[177] Ronnebro E, Majzoub EH. Crystal structure, Raman spectroscopy, and ab initio calculations of a new bialkali alanate

K2LiAlH6. J Phys Chem B 2006;110:25686–91.[178] Komiya K, Morisaku N, Shinzato Y, Ikeda K, Orimo S, Ohki Y, et al. Synthesis and dehydrogenation of M(AlH4)2 (M = Mg,

Ca). J Alloys Compd 2007;446–447:237–41.[179] Fichtner M, Fuhr O, Kircher O. Magnesium alanate – a material for reversible hydrogen storage? J Alloys Compd

2003;356–357:418–22.[180] Varin RA, Chiu C, Czujko T, Wronski Z. Mechano-chemical activation synthesis (MCAS) of nanocristalline magnesium

alanate hydride [Mg(AlH)4] and its hydrogen desorption properties. J Alloys Compd 2007;439:302–11.[181] Li C, Xiao X, Chen L, Jiang K, Li S, Wang Q. Synthesis of calcium alanate and its dehydriding performance enhanced by FeF3

doping. J Alloys Compd 2011;509:590–5.[182] Xiao X-z, Li C-x, Chen L-x, Fan X-l, Kou H-q, Wang Q-d. Synthesis and dehydrogenation of CeAl4-doped calcium alanate. J

Alloys Compd 2011;509(Supplement 2):S743. S6.[183] Varin RA, Chiu C, Czujko T, Wronski Z. Feasibility study of the direct mechano-chemical synthesis of nanostructured

magnesium tetrahydroaluminate (alanate) Mg(AlH4)2 complex hydride. Nanotechnology. 2005;16:2261–74.


[184] Sartori S, Léon A, Zabara O, Muller J, Fichtner M, Hauback BC. Studies of mixed hydrides based on Mg and Ca by reactiveball milling. J Alloys Compd 2009;476:639–43.

[185] Léon A, Zabara O, Sartori S, Eigen N, Dornheim M, Klassen T, et al. Investigation of (Mg, Al, Li, H)-based hydride andalanate mixtures produced by reactive ball milling. J Alloys Compd 2009;476:425–8.

[186] Vajeeston P, Ravindran P, Fjellvag H. Stability enhancement by particle size reduction in AlH3. J Alloys Compd2011;509:S662. S6.

[187] Graetz J, Reilly JJ, Yartys VA, Maehlen JP, Bulychev BM, Antonov VE, et al. Aluminum hydride as a hydrogen and energystorage material: past, present and future. J Alloys Compd 2011;509:S517. S28.

[188] Paskevicius M, Sheppard DA, Buckley CE. Characterisation of mechanochemically synthesised alane (AlH3) nanoparticles.J Alloys Compd 2009;487:370–6.

[189] Zhang QA, Akiba E. Reactions in the SrH2–Al–H2 system. J Alloys Compd 2008;460:272–5.[190] Liu DM, Fang CH, Zhang QA. Synthesis of the (Sr, Ca)2AlH7 hydride. J Alloys Compd 2009;477:337–40.[191] Friedrichs O, Borgschulte A, Kato S, Buchter F, Gremaud R, Remhof A, et al. Low-temperature synthesis of LiBH4 by gas–

solid reaction. Chem – A Eur J 2009;15:5531–4.[192] Remhof A, Yan Y, Friedrichs O, Kim JW, Mauron P, Borgschulte A, et al. Towards room temperature, direct, solvent free

synthesis of tetraborohydrides. In: 5th European conference on neutron scattering. Bristol: Iop Publishing Ltd.; 2012.[193] Friedrichs O, Kim JW, Remhof A, Wallacher D, Hoser A, Cho YW, et al. Core shell structure for solid gas synthesis of LiBD4.

Phys Chem Chem Phys 2010;12:4600–3.[194] Friedrichs O, Remhof A, Borgschulte A, Buchter F, Orimo SI, Zuttel A. Breaking the passivation – the road to a solvent free

borohydride synthesis. Phys Chem Chem Phys 2010;12:10919–22.[195] Remhof A, Borgschulte A, Friedrichs O, Mauron P, Yan Y, Zuttel A. Solvent-free synthesis and decomposition of Y(BH4)3.

Scr Mater 2012;66:280–3.[196] Friedrichs O, Remhof A, Hwang SJ, Zuttel A. Role of Li2B12H12 for the formation and decomposition of LiBH4. Chem Mater

2010;22:3265–8.[197] Pistidda C, Garroni S, Dolci F, Bardají EG, Khandelwal A, Nolis P, et al. Synthesis of amorphous Mg(BH4)2 from MgB2 and H2

at room temperature. J Alloys Compd 2010;508:212–5.[198] Filinchuk Y, Richter B, Jensen TR, Dmitriev V, Chernyshov D, Hagemann H. Porous and dense magnesium borohydride

frameworks: synthesis, stability, and reversible absorption of guest species. Angew Chem Int Edit 2011;50:11162–6.[199] Pistidda C, Garroni S, Minella CB, Dolci F, Jensen TR, Nolis P, et al. Pressure effect on the 2NaH+MgB2 hydrogen absorption

reaction. J Phys Chem C 2010;114:21816–23.[200] Chen P, Xiong Z, Luo J, Lin L, Tan KL. Interaction of hydrogen with metal nitrides and imides. Nature 2002;420:302–4.[201] Hu YH, Ruckenstein E. H2 storage in Li3N. Temperature-programmed hydrogenation and dehydrogenation. Ind Eng Chem

Res 2003;42:5135–9.[202] Hino S, Ichikawa T, Leng H, Fujii H. Hydrogen desorption properties of the Ca–N–H system. J Alloys Compd

2005;398:62–6.[203] Leng HY, Ichikawa T, Hino S, Hanada N, Isobe S, Fujii H. Synthesis and decomposition reactions of metal amides in metal–

N–H hydrogen storage system. J Power Sour 2006;156:166–70.[204] Chen P, Xiong Z, Luo J, Lin J, Tan KL. Interaction between lithium amide and lithium hydride. J Phys Chem B.

2003;107:10967–1070.[205] Hu YH, Ruckenstein E. Ultrafast reaction between LiH and NH3 during H2 storage in Li3N. J Phys Chem A

2003;107:9737–9.[206] Ichikawa T, Hanada N, Isobe S, Leng H, Fujii H. Mechanism of novel reaction from LiNH2 and LiH to Li2NH and H2 as a

promising hydrogen storage system. J Phys Chem B 2004;B108:7887–92.[207] Schlesinger HI, Brown HC, Finholl AE. The preparation of sodium borohydride by the high temperature reaction of sodium

hydride with borate esters. J Am Chem Soc 1953;75:205–9.[208] Li ZP, Liu BH, Morigazaki N, Suda S. Preparation of potassium borohydride by a mechanico-chemical reaction of saline

hydrides with dehydrated borate through ball milling. J Alloys Compd 2003;354:243–7.[209] Li ZP, Morigazaki N, Liu BH, Suda S. Preparation of sodium borohydride by the reaction of MgH2 with dehydrated borax

through ball milling at room temperature. J Alloys Compd 2003;349:232–6.[210] Mal’tseva NN, Myakishev RG, Saidov BI, Kedrova NS, Gorbacheva II, Patapova OG, et al. Russ J Inorg Chem 1989:806–8.[211] Hagemann H, Longhini M, Kaminski JW, Wesolowski TA, Cerny R, Penin N, et al. LiSc(BH4)4: a novel salt of Li and discrete

Sc(BH4)4– complex anions. J Phys Chem A 2008;112:7551–5.

[212] Hwang SJ, Bowman RC, Reiter JW, Rijssenbeek J, Soloveichik GL, Zhao JC, et al. NMR confirmation for formation ofB12H12

2� complexes during hydrogen desorption from metal borohydrides. J Phys Chem C 2008;112:3164–9.[213] Cerny R, Severa G, Ravnsbæk DB, Filinchuk Y, D’Anna V, Hagemann H, et al. NaSc(BH4)4: a novel scandium-based

borohydride. J Phys Chem C 2010;114:1357–64.[214] Cerny R, Ravnsbæk DB, Severa G, Filinchuk Y, D’Anna V, Hagemann H, et al. Structure and characterization of KSc(BH4)4. J

Phys Chem C 2010;114:19540–9.[215] Ravnsbaek DB, Filinchuk Y, Cerny R, Ley MB, Haase D, Jakobsen HJ, et al. Thermal polymorphism and decomposition of

Y(BH4)3. Inorg Chem. 2010;49:3801–9.[216] Frommen C, Aliouane N, Deledda S, Fonneløp JE, Grove H, Lieutenant K, et al. Crystal structure, polymorphism, and

thermal properties of yttrium borohydride Y(BH4)3. J Alloys Compd 2010;496:710–6.[217] Ravnsbæk D, Filinchuk Y, Cerenius Y, Jakobsen Hans J, Besenbacher F, Skibsted J, et al. A series of mixed-metal

borohydrides. Angew Chem Int Ed 2009;48:6659–63.[218] Cerny R, Ravnsbaek DB, Schouwink P, Filinchuk Y, Penin N, Teyssier J, et al. Potassium zinc borohydrides containing

triangular ZnðBH4Þ�3 and tetrahedral ZnðBH4ÞxCl2�4�x anions. J Phys Chem C 2012;116:1563–71.[219] Ravnsbæk DB, Sørensen LH, Filinchuk Y, Reed D, Book D, Jakobsen HJ, et al. Mixed-anion and mixed-cation borohydride

KZn(BH4)Cl2: synthesis, structure and thermal decomposition. Eur J Inorga Chem 2010:1608–12.[220] Lindemann I, Domènech Ferrer R, Dunsch L, Filinchuk Y, Cerny R, Hagemann H, et al. Al3Li4(BH4)13: a complex double-

cation borohydride with a new structure. Chem – A Eur J 2010;16:8707–12.


[221] Arnbjerg LM, Ravnsbæk DB, Filinchuk Y, Vang RT, Cerenius Y, Besenbacher F, et al. Structure and dynamics for LiBH4�LiClsolid solutions. Chemi Mater 2009;21:5772–82.

[222] Mosegaard L, Moller B, Jorgensen J-E, Filinchuk Y, Cerenius Y, Hanson JC, et al. Reactivity of LiBH4; in situ synchrotronradiation powder X-ray diffraction study. The J Phys Chem C 2008;112:1299–303.

[223] Rude LH, Zavorotynska O, Arnbjerg LM, Ravnsbæk DB, Malmkjær RA, Grove H, et al. Bromide substitution in lithiumborohydride, LiBH4–LiBr. Int J Hydrogen Energy 2011;36:15664–72.

[224] Rude LH, Groppo E, Arnbjerg LM, Ravnsbæk DB, Malmkjær RA, Filinchuk Y, et al. Iodide substitution in lithiumborohydride, LiBH4–LiI. J Alloys Compd 2011;509:8299–305.

[225] Ravnsbæk DB, Rude LH, Jensen TR. Chloride substitution in sodium borohydride. J Solid State Chem 2011;184:1858–66.[226] Rude LH, Filinchuk Y, Sørby MH, Hauback BC, Besenbacher F, Jensen TR. Anion substitution in Ca(BH4)2–CaI2: synthesis,

structure and stability of three new compounds. J Phys Chem C 2011;115:7768–77.[227] Lee YS, Shim JH, Cho YW. Polymorphism and thermodynamics of Y(BH4)3 from first principles. J Phys Chem C

2010;114:12833–7.[228] Jaron T, Grochala W. Y(BH(4))(3) – an old–new ternary hydrogen store aka learning from a multitude of failures. Dalton

Trans 2010;39:160–6.[229] Nakamori Y, Li HW, Miwa K, Towata S, Orimo S. Syntheses and hydrogen desorption properties of metal-borohydrides

M(BH4)(n) (M = Mg, Sc, Zr, Ti, and Zn; n = 2–4) as advanced hydrogen storage materials. Mater Trans. 2006;47:1898–901.[230] Jeon E, Cho YW. Mechanochemical synthesis and thermal decomposition of zinc borohydride. J Alloys Compd

2006;422:273–5.[231] Balema VP. Mechanical processing in hydrogen storage research and development. Mater Matter 2007;2:16–8.[232] Nakamori Y, Li H-W, Kikuchi K, Aoki M, Miwa K, Towata T, et al. Thermodynamical stabilities of metal-borohydrides. J

Alloys Compd 2007;446–447:296–300.[233] Solinas I, Lutz HD. Nonceramic preparation techniques for ternary halides AB2X4 with A = Mg, Mn, Zn B = Li, Na X = Cl, Br-

1. J Solid State Chem 1995;117:34–8.[234] Lee JY, Lee YS, Suh JY, Shim JH, Cho YW. Metal halide doped metal borohydrides for hydrogen storage: The case of

Ca(BH4)2–CaX2 (X = F, Cl) mixture. J Alloys Compd 2010;506:721–7.[235] Shannon RD. Revised effective ionic-radii and systematic studies of interatomic distances in halides and chalcogenides.

Acta Crystallogr Sect A 1976;32:751–67.[236] Pistoriu Cw. Melting and polymorphism of LiBH4 to 45 KBar. Z Phys Chem – Frank 1974;88:253–63.[237] Cerny R, Penin N, D’Anna V, Hagemann H, Durand E, Ruzicka J. MgxMn(1�x)(BH4)2 (x = 0–0.8), a cation solid solution in a

bimetallic borohydride. Acta Mater 2011;59:5171–80.[238] Ley MB, Ravnsbæk DB, Filinchuk Y, Lee Y-S, Cheptiakov CD, Janot R, et al. A new lithium ion conductor and hydrogen

storage material with isolated tetranuclear anionic clusters. LiCe(BH4)3Cl. Chem Mater 2012;24:1654–63.[239] Ley MB, Boulineau S, Janot R, Filinchuk Y, Jensen TR. New Li ion conductors and solid state hydrogen storage materials

[LiM(BH4)3Cl (M = La, Gd)]. Inorg. Chem 2012, submitted for publication.[240] Hauback BC. Structures of aluminium-based light weight hydrides. Z Kristal 2008;223:636–48.[241] Ma E. Alloys created between immiscible elements. Prog Mater Sci 2005;50:413–509.[242] Bastide JP, Elhajri J, Claudy P, Elhajbi A. A new route to alkali-metal aluminum hydrides MAlH4 with M = Na, K, Rb, Cs and

structural features for the whole family with M = Li TO Cs. Synth React Inorg Met – Org Chem 1995;25:1037–47.[243] Brower FM, Matzek NE, Reigler PF, Rinn HW, Roberts CB, Schmidt DL, et al. Preparation and properties of aluminum

hydride. J Am Chem Soc 1976;98:2450–3.[244] Graetz J, Chaudhuri S, Wegrzyn J, Celebi Y, Johnson JR, Zhou W, et al. Direct and reversible synthesis of AIH3-

triethylenediamine from Al and H2. J Phys Chem C 2007;111:19148–52.[245] Schüth F, Bogdanovic B, Felderhoff M. Light metal hydrides and complex hydrides for hydrogen storage. Chem Commun

2004;2004:2249–58.[246] Brinks HW, Istad-Lem A, Hauback BC. Mechanochemical synthesis and crystal structure of alpha-AlD3 and alpha-AlD3. J

Phys Chem B 2006;110:25833–7.[247] Brinks HW, Hauback BC, Jensen CM, Zidan R. Synthesis and crystal structure of Na2LiAlD6. J Alloys Compd

2005;392:27–30.[248] Sartori S, Istad-Lem A, Brinks HW, Hauback BC. Mechanochemical synthesis of alane. Int J Hydrogen Energy

2009;34:6350–6.[249] Sandrock G, Reilly J, Graetz J, Zhou W-M, Johnson J, Wegrzyn J. Accelerated thermal decomposition of AlH3 for hydrogen-

fueled vehicles. Appl Phys A 2005;80:687–90.[250] Bösenberg U, Doppiu S, Mosegaard L, Barkhordarian G, Eigen N, Borgschulte A, et al. Hydrogen sorption properties of

MgH2–LiBH4 composites. Acta Mater 2007;55:3951–8.[251] Bösenberg U, Ravnsbæk DB, Hagemann H, D’Anna V, Minella CB, Pistidda C, et al. Pressure and temperature influence on

the desorption pathway of the LiBH4�MgH2 composite system. The J Phys Chem C 2010;114:15212–7.[252] Ravnsbæk DB, Jensen TR. Mechanism for reversible hydrogen storage in LiBH4–Al. J Appl Phys 2012;111:112621.[253] Jensen CM, Zidan RA, Mariels N, Hee A, Hagen C. Advanced titanium doping of sodium aluminum hydride: segue to a

practical hydrogen storage material? Int J Hydrogen Energy 1999;24:461–5.[254] Easton DS, Schneibel JH, Speakman SA. Factors affecting hydrogen release from lithium alanate (LiAlH4). J Alloys Compd

2005;398:245–8.[255] Balema VP, Dennis KW, Pecharsky VK. Rapid solid-state transformation of tetrahedral [AlH4]� into octahedral [AlH6]3� in

lithium aluminohydride. Chem Commun 2000;2000:1665–6.[256] Balema VP, Balema L. Missing pieces of the puzzle or about some unresolved issues in solid state chemistry of alkali metal

aluminohydrides. Phys Chem Chem Phys 2005;7:1310–4.[257] Pitt MP, Paskevicius M, Webb CJ, Sorby MH, Delleda S, Jensen TR, et al. Nanoscopic Al1�xCex phases in the

NaH + Al + 0.02CeCl3 system. Int J Hydrogen Energy 2011;36:8403–11.[258] Ares JR, Aguey-Zinsou KF, Porcu M, Sykes JM, Dornheim M, Klassen T, et al. Thermal and mechanically activated

decomposition of LiAlH4. Mater Res Bull 2008;43:1263–75.


[259] Zaluska A, Zaluski L, Ström-Olsen JO. Sodium alanates for reversible hydrogen storage. J Alloys Compd 2000;298:125–34.[260] Dymova TN, Mal’tseva NN, Konoplev VN, Golovanova AI, Aleksandrov DP, Sizareva AS. Solid-phase solvate-free formation

of magnesium hydroaluminates Mg(AlH4)2 and MgAlH5 upon mechanochemical activation or heating of magnesiumhydride and aluminum chloride mixtures. Russ J Coord Chem 2003;29:385–9.

[261] Mamatha M, Bogdanovic B, Felderhoff M, Pommerin A, Schmidt W, Schüth F, et al. Mechanochemical preparation andinvestigation of properties of magnesium, calcium and lithium–magnesium alanates. J Alloys Compd 2006;407:78–86.

[262] Mal’tseva NN, Golovanova AI, Dymova TN, Aleksandrov DP. Solid-phase formation of calcium hydridoaluminatesCa(AlH4)2 and CaHAlH4 upon mechanochemical activation or heating of mixtures of calcium hydride with aluminumchloride. Russ J Inorg Chem 2001;46:1793–7.

[263] Dymova TN, Konoplev VN, Sizareva AS, Aleksandrov DP. Peculiarities of the solid-phase formation of strontiumpentahydroaluminate from binary hydrides by mechanochemical activation and modeling of the process on the basis ofthermogasovolumetry data. Russ J Coord Chem 2000;26:531–7.

[264] Zaluska A, Zaluski L, Ström-Olsen JO. Lithium–beryllium hydrides: the lightest reversible metal hydrides. J Alloys Compd2000;307:157–66.

[265] Weidenthaler C, Pommerin A, Felderhoff M, Sun WH, Wolverton C, Bogdanovic B, et al. Complex rare-earth aluminumhydrides: mechanochemical preparation, crystal structure and potential for hydrogen storage. J Am Chem Soc2009;131:16735–43.

[266] Aoki M, Miwa K, Noritake T, Kitahara G, Nakamori Y, Orimo S, et al. Destabilization of LiBH4 by mixing with LiNH2. ApplPhys A 2005;80:1409–12.

[267] Pinkerton FE, Meisner GP, Meyer MS, Balogh MP, Kundrat MD. Hydrogen desorption exceeding ten weight percent fromthe new quaternary hydride Li3BN2H8. J Phys Chem B 2005;109:6–8.

[268] Noritake T, Aoki M, Towata S, Miwa K, Ninomiya A, Nakamori Y, et al. Crystal structure analysis of novel complex hydridesformed by the combination of LiBH4 and LiNH2. Appl Phys A 2006;83:277–9.

[269] Meisner GP, Scullin ML, Balogh MP, Pinkerton FE, Meyer MS. Hydrogen release from mixtures of lithium borohydride andlithium amide: a phase diagram study. J Phys Chem B 2006;110:4186–92.

[270] Wu H, Zhou W, Yildirim T. Alkali and alkaline-earth metal amidoboranes: structure, crystal chemistry, and hydrogenstorage properties. J Am Chem Soc 2008;130:14834–9.

[271] Xiong ZT, Yong CK, Wu GT, Chen P, Shaw W, Karkamkar A, et al. High-capacity hydrogen storage in lithium and sodiumamidoboranes. Nat Mater 2008;7:138–41.

[272] Genova RV, Fijalkowski KJ, Budzianowski A, Grochala W. Towards Y(NH2BH3)3: probing hydrogen storage properties ofYX3/MNH2BH3 (X = F, Cl; M = Li, Na) and YH(x similar to 3)/NH3BH3 composites. J Alloys Compd 2010;499:144–8.

[273] Xiong ZT, Wu GT, Hu JJ, Chen P. Investigation on chemical reaction between LiAlH4 and LiNH2. J Power Sour2006;159:167–70.

[274] Dolotko O, Zhang HQ, Ugurlu O, Wiench JW, Pruski M, Chumbley LS, et al. Mechanochemical transformations inLi(Na)AlH4–Li(Na)NH2 systems. Acta Mater 2007;55:3121–30.

[275] Xiong ZT, Hu JJ, Wu GT, Liu YF, Chen P. Large amount of hydrogen desorption and stepwise phase transition in thechemical reaction of NaNH2 and LiAlH4. Catal Today 2007;120:287–91.

[276] Chua YS, Wu GT, Xiong ZT, Chen P. Investigations on the solid state interaction between LiAlH4 and NaNH2. J Solid StateChem 2010;183:2040–4.

[277] Lovvik OM, Opalka SM, Brinks HW, Hauback BC. Crystal structure and thermodynamic stability of the lithium alanatesLiAlH4 and Li3AlH6. Phys Rev B 2004;69.

[278] Chung S-C, Morioka H. Thermochemistry and crystal structures of lithium, sodium and potassium alanates as determinedby ab initio simulations. J Alloys Compd 2004;372:92–6.

[279] Ravnsbæk DB, Jensen TR. Tuning hydrogen storage properties and reactivity: investigation of the LiBH4NaAlH4 system. JPhys Chem Solids 2010;71:1144–9.

[280] Shi Q, Yu X, Feidenhans’l R, Vegge T. Destabilized LiBH4–NaAlH4 mixtures doped with titanium based catalysts. J PhysChem C 2008;112:18244–8.

[281] Fichtner M, Frommen C, Fuhr O. Synthesis and properties of calcium alanate and two solvent adducts. Inorg Chem2005;44:3479–84.

[282] Bortz M, Bertheville B, Böttger G, Yvon K. Structure of the high pressure phase g-MgH2 by neutron powder diffraction. JAlloys Compd 1999;287:L4–6.

progress in materials sciencematerials has generated an enormous amount of reports. in the last 10...

Documents