mandal_h2 storage materials

Upload: luh06002

Post on 06-Apr-2018

218 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/3/2019 Mandal_H2 Storage Materials

    1/35

    Annual Reports on the Progress of Chemistry, Section A Vol 105

    Inorganic Chemistry

    DownloadedbyU

    niversityofConnecticuton07October2011

    Publishedon13Ma

    y2009onhttp://pubs.rsc.org|doi:10.1039/B818951J

    View Online

    http://dx.doi.org/10.1039/b818951j
  • 8/3/2019 Mandal_H2 Storage Materials

    2/35

    Hydrogen storage materials: present scenariosand future directions

    Tapas K. Mandal and Duncan H. Gregory*

    DOI: 10.1039/b818951j

    This review describes the present state of contemporary solid state hydrogen

    storage on the basis of research carried out during the last decade. The

    article focuses on the key aspects of materials based on the physical and

    chemical storage of hydrogen and emerging mechanisms for reversible

    storage. Among chemical storage materials, we consider metal hydrides

    (both light and complex), nitrides-imides-amides and other multi-

    component systems and discuss the emergence of coordination polymers

    (metal organic frameworks; MOFs) among solids exhibiting physicalstorage. Significant challenges remain if we are to meet the practical

    demands required of a solid state storage system, namely high storage

    density together with favourable sorption thermodynamics and kinetics and

    prolonged cycleability and lifetime. This review emphasises both how our

    understanding of the storage mechanism (as a process or phenomenon

    during hydrogen cycling) is evolving and how this understanding impacts on

    future materials design. The prospect of high capacity storage and fast

    kinetics in nanostructured materials is highlighted as is the role of complex,

    multi-component, composite systems in future hydrogen storage research.

    1. Introduction

    At the juncture of the 20th and 21st centuries it was realised that there was an urgent

    requirement for an alternative and sustainable energy vector as reserves of fossil

    fuels faded.1,2 Moreover, if the adverse effects of global warming and consequent

    climate change is to be arrested then the utilisation of green and renewable energy

    sources is imperative. As Grochala and Edwards and later van den Berg and Area n

    note in their articles,3 Jules Verne first brought these issues to the attention of thepublic in a work of fiction over 100 years ago: . . . water will one day be employed as

    a fuel, that hydrogen and oxygen will constitute it, used singly or together, will furnish

    an inexhaustible source of heat and light and water will be the coal of the future

    (The Mysterious Island, Jules Verne, 1874). Vernes work was to presage with

    uncanny pertinence the role of hydrogen as the fuel of the future and predict the

    coming of the so-called hydrogen economy.

    An uninterrupted and secure energy supply for the developed nations and meeting

    the increasing energy demands of the rapidly developing nations are essential. The

    burgeoning need for energy coupled with the rapid depletion of fossil fuels pose

    serious threats for sustainable development. Before the potential for hydrogen as a

    West CHEM, Department of Chemistry, Joseph Black Building, University of Glasgow,Glasgow, UK. E-mail: [email protected]; Fax: +44 (0)141 330 8128;Tel: +44 (0)141 330 4888

    Annu. Rep. Prog. Chem., Sect. A, 2009, 105, 2154 | 21

    This journal is c The Royal Society of Chemistry 2009

    REVIEW www.rsc.org/annrepa | Annual Reports A

    DownloadedbyUniversityofConnecticuton07

    October2011

    Publishedon13

    May2009onhttp://pubs.rsc.org|doi:10.1

    039/B818951J

  • 8/3/2019 Mandal_H2 Storage Materials

    3/35

    future energy carrier can be realised, fundamental research including components of

    invention and discovery, subsequent implementation of new technology and socio-

    economic acceptance must occur. These are the key steps to the hydrogen energy

    transition.

    The greater hydrogen energy picture centres around the production and storage of

    hydrogen, the two most important steps that currently represent a bottle neck to

    utilization of hydrogen more widely in fuel cell systems. It must be realized thatunlike coal or oil, hydrogen is not naturally available. Thus, stable large-scale

    hydrogen production is necessary for the gradual switchover to the hydrogen

    economy. However a safe, efficient and economic storage medium is likely to be a

    crucial prerequisite before hydrogen would be globally acceptable as a fuel,

    particularly in mobile (automotive) applications. It should be emphasised here that

    while gaseous or liquid hydrogen is currently an option for prototype personal

    vehicles (cars) or larger commercial transport, solid state storage of hydrogen is

    potentially superior with regard to its storage capacity (both gravimetric and

    volumetric), energy efficiency and safety.2 Nevertheless compact storage of hydrogen

    in a solid medium is the most demanding and challenging part of realising thehydrogen economy as far as mobile applications are concerned (Fig. 1).2b

    Hydrogen storage is thus a key research area where considerable international

    effort is concentrated. Since this review is written from a materials perspective,

    current chemical research in hydrogen storage materials will be highlighted captur-

    ing the discoveries and developments over the past decade (19982009), the many

    opportunities that could be seized and the future pathways that could be taken.

    Although many of the materials classes meet the majority of the US Department of

    Energy (US-DoE) criteria for vehicular applications (e.g. hydrogen storage capacity;

    the amount of hydrogen stored per unit mass and per unit volume),4 other factors

    such as non-reversibility, slow kinetics and sometimes thermodynamic barriers to

    hydrogen uptake-release limit or prevent their practical use. Holistic and systematic

    research towards understanding mechanism, structure and thermodynamics and

    Fig. 1 Alternatives for storage of 4 kg hydrogen, with volume relative to the size of a car.

    (Reprinted from ref. 2b, with permission from Macmillan Publishers Ltd.)

    22 | Annu. Rep. Prog. Chem., Sect. A, 2009, 105 , 2154

    This journal is c The Royal Society of Chemistry 2009

    DownloadedbyUniversityofConnecticuton07

    October2011

    Publishedon13

    May2009onhttp://pubs.rsc.org|doi:10.1

    039/B818951J

  • 8/3/2019 Mandal_H2 Storage Materials

    4/35

    their inter-relationships are crucial to innovation and materials development. This

    review illustrates the extent of emerging materials discovery and design and

    demonstrates how an improved chemical understanding of storage processes informs

    an evolving materials design strategy.

    2. Chemical storage

    The chemical storage of hydrogen concerns materials where hydrogen is chemi-cally bound in a compound. Therefore, the loading and unloading of hydrogen

    (hydrogenationdehydrogenation) in this class of compound requires a chemisorption

    step prior to the absorption of hydrogen in the bulk and the formation of chemical

    bonds to hydrogen. In contrast, physisorption (see section 3) is responsible in cases

    where molecular hydrogen is only physically adsorbed on the surface of a solid or in

    the internal volume of porous/framework structures. In the chemical storage scenario,

    hydrogen is strongly bound while in physisorption it is so weakly held that often very

    high pressures or low temperatures are required, making physisorbed materials much

    less attractive for practical uses at ambient or elevated temperatures.

    Among chemical storage materials, families of metal alloys and their hydrides

    constitute a core grouping that has been widely investigated for over a century.3,5 In

    general, the storage capacities of these materialswhich often contain f- and/or

    d-block elementsare too low to meet the DOE targets for vehicular applications

    but could find utility in stationary applications due to sometimes excellent reversi-

    bility. For example, Mg2Ni (with a gravimetric capacity of 3.6 wt% hydrogen) and

    LaNi5 (1.28 wt%) are two classic alloy systems that are noteworthy due to their

    extraordinary cycleability.5 Light metal binary hydrides, however, are considered as

    an important class with respect to high hydrogen storage density but are plagued by

    high desorption temperatures, non-reversible thermodynamics and poor kinetics.6

    MgH2 is one of the binary hydrides that still attracts huge attention owing to its light

    weight and high gravimetric density even given its kinetic and thermodynamic

    limitations for storage.7 On the contrary, complex chemical hydrides offer tremen-

    dous opportunities over the binaries in many respects.6,810 Among them, the

    borohydrides,810,11 aluminohydrides (alanates)6,810 and amides810,12,13 are close

    to meeting US-DoE targets in terms of storage capacity, reversibility, cost and

    toxicity. The complex hydrides divide rather neatly into 2 groups depending on their

    composition and bonding: borohydrides and alanates, (for example LiBH4 and

    LiAlH4) are hydridic in nature and contain hydrogen within complex anions

    ([BH4], [AlH4]) that can be represented as Hd. In contrast, the imides andamides, e.g. Li2NH, LiNH2 are protonic and contain hydrogen within complex

    anions ([NH]2, [NH2]) that can be represented as Hd+.These two distinct and

    discrete classes of complex hydride are covered separately in the sections below.

    2.1 Light metal hydrides

    Among the light metal binary hydrides, magnesium hydride (MgH2) is the most

    attractive due to its low atomic weight, high hydrogen storage capacity and low

    cost.7 In Table 1, the gravimetric and volumetric density along with hydrogen

    desorption temperature and reversibility for selected light binary hydrides arecompared with that of gaseous and liquid hydrogen.3a,7,14 Although, MgH2 has a

    high gravimetric capacity of 7.6 wt%, it suffers from slow kinetics and a high

    decomposition temperature (B330 1C). A huge body of research has been devoted to

    improving its performance.7,1527

    Annu. Rep. Prog. Chem., Sect. A, 2009, 105, 2154 | 23

    This journal is c The Royal Society of Chemistry 2009

    DownloadedbyUniversityofConnecticuton07

    October2011

    Publishedon13

    May2009onhttp://pubs.rsc.org|doi:10.1

    039/B818951J

  • 8/3/2019 Mandal_H2 Storage Materials

    5/35

    Particle size reduction by ball milling, alloying or doping are among the most

    common approaches that are being pursued to speed up the kinetics in the Mg-Hsystem.15,16 Catalytically enhanced systems show great promise for improved

    hydrogen storage characteristics and mechanical milling has been employed to

    prepare catalysed MgH2 composites.1719 Various additives include carbon,15f,21

    transition metals,15d,22 oxides23,24 and halides.25 Milled Ni-doped19 MgH2 shows a

    6.5 wt% storage capacity at 150250 1C and a remarkable improvement in sorption

    was reported for 1 mol% Nb2O5 doped milled MgH2 with high desorption kinetics

    at 160 1C under a helium flow.20 Further, a recent investigation of the effect of ZrNi

    activated alloys on milled MgH2 indicated significant improvements in desorption

    kinetics.26 By comparing several compositions in this latter alloy system, it has been

    argued that the alloys are playing a crucial catalytic role distinct from grain size

    effects induced by milling.26 Although, the development of suitable catalysts has

    drastically reduced the charge-discharge times, the thermodynamics of the process

    remain almost unaltered.

    Chemical destabilisation27 of MgH2 and many other complex hydrides in a multi-

    component mixture with designed reaction paths is a fairly new development that will

    be discussed separately below under multinary systems (section 2.4). However, it must

    be pointed out here that chemical destabilisation of MgH2 by alloying with Cu dates

    back to the work of Reilly and Wiswall in 1967.28 They reversibly hydrogenated Mg2Cu

    to MgH2 and MgCu2 at 2401C (which is

    B

    401C lower than that of pure MgH2) withan equilibrium pressure of 1 bar.28 Similarly, in a recent study, alloying with Si was

    shown to destabilise MgH2 significantly forming Mg2Si upon dehydrogenation.29 But

    the MgH2/Si system suffers from slow kinetics at 150 1C, which is the required

    temperature for hydrogenation at 100 bar.29 Another very recent emerging trend in

    the MgH2 based materials is to chemically nanostructure hydride materials10,30 and

    this approach is highlighted in section 4 under New materials for hydrogen storage.

    These sections in particular emphasise the many opportunities that exist to improve

    reversibility and thermodynamics in light-weight metal hydrides.

    2.2 Complex hydridic hydrides, borohydrides and alanates ([BH4] and [AlH4]

    )

    Hydrides containing complex anionic units with light elements like Al, B and

    achieving charge neutrality through ionic or partially covalent bonding with a more

    electropositive cation, such as, Li, Na, K, Mg and Ca, have been classified as

    Table 1 Hydrogen content of simple binary hydrides in comparison with gaseous and liquid

    hydrogen.3a,7,14 Tdec and kinetic reversibility of the hydrides are also given.

    Compound

    Wt%

    H2

    Volumetric density NH(atoms H/ml 1022)

    Tdec(1C)

    Kinetic

    reversibility

    H2, liquid 100 4.2

    H2, gas

    (100 atm)

    100 0.49

    LiH 12.6 5.3 720 Poor

    NaH 4.2 2.3 425 Good

    CaH2 4.8 5.1 600 Good

    MgH2 7.6 6.7 330 Very poor

    AlH3 10.0 8.84 150 Irreversible

    TiH2 4.0 9.1 380 a

    a No literature data available.

    24 | Annu. Rep. Prog. Chem., Sect. A, 2009, 105 , 2154

    This journal is c The Royal Society of Chemistry 2009

    DownloadedbyUniversityofConnecticuton07

    October2011

    Publishedon13

    May2009onhttp://pubs.rsc.org|doi:10.1

    039/B818951J

  • 8/3/2019 Mandal_H2 Storage Materials

    6/35

    so-called chemical hydrides. In these complex hydrides, the hydrogen within the

    anionic unit resides at the vertices of a tetrahedron with B or Al at the centre to form

    tetrahydroborates e.g. Li(BH4), Mg(BH4)2 or tetrahydroaluminates (also known as

    alanates), e.g. Li(AlH4), Na(AlH4). Although these tetrahydrometallates constitute

    an interesting storage class, the very large gravimetric and volumetric hydrogen

    densities (Table 2) that they exhibit is countered by their thermal stability and high

    desorption temperature which often makes them unsuitable for application. Thedesorption temperature and gravimetric hydrogen densities for some selected

    complex hydrides are shown in Table 2.

    Of these, the lightest borohydride LiBH4 has attracted enormous interest as a

    candidate hydrogen store due to its extremely high gravimetric hydrogen density

    (18.4 wt%),8,3234 although as yet it has not been possible to overcome the profound

    thermodynamic, kinetic and reversibility challenges that would enable it to be

    employed as a practical storage material.

    The first thermal desorption of hydrogen from LiBH4 was reported in 1964 by

    Fedneva et al.35 and four years later Stasinevich and Egorenko36 showed that LiBH4

    releases hydrogen at temperatures greater than 470 1C. According to Fedneva et al.35

    the differential thermal analysis profile of LiBH4 consists of three endothermic

    events. While the first reversible endothermic effect at 108112 1C is attributed to a

    polymorphic transformation of LiBH4, the second peak at 268286 1C corresponds

    to melting with the onset of gradual decomposition of LiBH4.35 The main hydrogen

    release starts at 380 1C liberating 80% of the hydrogen in LiBH4. Moreover, another

    thermal event at around 490 1C was unexplained. However, in their study,

    Stasinevich and Egorenko conducted the thermal analysis under hydrogen pressures

    of 10 bar and according to them, the dehydrogenation reaction of LiBH4 releases

    only 3 of the 4 hydrogen atoms above 268 1C.36 This process can be represented as:

    LiBH4- LiH + B + 3/2 H2 (1)

    Analogous decomposition pathways are observed for other alkali metal boro-hydrides (NaBH4 and KBH4) but with an increase in the decomposition temperaturewith increasing atomic number of the alkali metal. In a later study, Zu ttel et al.demonstrated that three distinctive desorption peaks exist during the dehydro-genation of LiBH4 when performed at a low heating rate.

    37 This suggests a morecomplicated mechanism and the presence of other intermediate phases during the

    Table 2 Desorption temperature and gravimetric storage capacities of complex chemical

    hydrides3a,8

    Compound Tdes (1C)

    Storage

    capacity (wt%)

    LiBH4 380 18.4

    NaBH4 400 10.6

    KBH4 500 7.4

    Mg(BH4)2 260280 14.8

    Ca(BH4)2 320 11.5

    Al(BH4)3 20 16.8

    LiAlH4 125 9.5

    NaAlH4 210 7.4KAlH4 270 5.7

    Mg(AlH4)2 140200 9.3

    Ca(AlH4)2 250a 7.8

    a For complete desorption from neat Ca(AlH4)2.31

    Annu. Rep. Prog. Chem., Sect. A, 2009, 105, 2154 | 25

    This journal is c The Royal Society of Chemistry 2009

    DownloadedbyUniversityofConnecticuton07

    October2011

    Publishedon13

    May2009onhttp://pubs.rsc.org|doi:10.1

    039/B818951J

  • 8/3/2019 Mandal_H2 Storage Materials

    7/35

    decomposition. Only recently, has the involvement of a borohydride cluster,Li2B12H12, been elucidated as an intermediate in the thermal desorption ofLiBH4.

    38,39 A series of systematic theoretical and experimental studies has beencarried out to attempt to rationalise the thermodynamic stabilities of metalborohydrides, M(BH4)n (M = Li, Na, K, Mg, Cu, Zn, Sc, Zr and Hf).

    40 A goodcorrelation has been observed between desorption temperature (Td), the enthalpy ofdesorption (DHdes) and the Pauling electronegativity (wP). Similar studies have been

    performed for Ca(BH4)2 and it was established that the ionic bonding exists betweenCa and the covalent [BH4]

    unit and that there is a linear relationship between theheat of formation and wP of the cation.

    41 The same study also predicted a possibledecomposition pathway as:

    Ca(BH4)2- 2/3 CaH2 + 1/3 CaB6 + 10/3 H2 (2)

    with a calculated gravimetric capacity of 9.6 wt% of hydrogen. Thermogravimetricanalysis leant support to the proposed reaction, yielding an experimental weight lossof 9.2 wt% and suggested that an intermediate species could be involved in thedecomposition.41 Although the CaBH system has a high storage capacity,

    subsequent studies indicated that the dehydrogenation starts only at 640 K andthe kinetics of dehydrogenation is very slow.42,43

    Various approaches have been taken to address the thermodynamics and kinetics

    of hydrogenationdehydrogenation cycling in LiBH4 and the other complex

    borohydrides. These approaches were very much triggered by the chemical activa-

    tion of NaAlH4 by Bogdanovic and Schwickardi in their seminal paper of 1997.44 In

    their pioneering study they showed that 2% Ti-catalyzed NaAlH4 (sodium alanate)

    undergoes reversible dehydriding-rehydriding reactions in the solid state under

    moderate conditions.44 The hydrogen desorption takes place in two steps.

    3 NaAlH42 Na3AlH6 + 2 Al + 3 H2 (3)

    Na3AlH62 3 NaH + Al + 3/2 H2 (4)

    While the first step (3) amounts to 3.7 wt% hydrogen (with DH= 37 kJ mol1 H2),

    the second step (4) releases 1.8 wt% (with DH= 47 kJ mol1 H2) corresponding to a

    total of 5.5 wt% of hydrogen. The enthalpy values correspond to dehydrogenation

    temperatures of about 35 and 100 1C, respectively.45 The theoretical storage capacity

    of 7.4 wt% for the NaAlH4 system can be achieved if the remaining hydrogen stored

    in NaH is released. This, however, requires temperatures greater than 500 1C due to

    the very high stability of NaH. In fact, the two step decomposition mechanism of

    NaAlH4 to NaH, Al and H2 at 210220 and 2501

    C respectively was first demon-strated over 40 years ago.46 Subsequent studies revealed that similar two step

    decomposition reactions also take place for LiAlH4 and KAlH4.4752 Although,

    there have been reports of reversibility in these systems,51 the kinetics are extremely

    slow. Further, relatively harsh conditions are required (200400 1C and 1040 MPa)

    for cycling and the reversibility is essentially only partial.

    The first examples of the use of titanium to catalyse dehydrogenation reactions

    took place in solution and include, for example, the Ti-catalysed dehydrogenation of

    LiAlH4 in ether by Wieberg et al. in 1951.53 However, the above discovery of

    Ti-catalysis for reversible hydrogen storage in the solid state under favorable thermo-

    dynamic conditions, as first represented by the alanates,44 has created a vast amountof fresh research activity in main group complex hydrides. Many other catalytic

    dopants have since been studied, but no other metal-containing phases appear to

    outperform those containing Ti. Use of different additive compositions as a source for

    Ti impregnation/dispersion in NaAlH4 have produced results with some variance in

    26 | Annu. Rep. Prog. Chem., Sect. A, 2009, 105 , 2154

    This journal is c The Royal Society of Chemistry 2009

    DownloadedbyUniversityofConnecticuton07

    October2011

    Publishedon13

    May2009onhttp://pubs.rsc.org|doi:10.1

    039/B818951J

  • 8/3/2019 Mandal_H2 Storage Materials

    8/35

    performance and catalytic enhancement and the doping methods themselves are also

    found to exert an influence. For example, Jensen and Gross have shown that the onset

    of dehydriding reaction differs appreciably (Fig. 2) for solution doped and mechani-

    cally doped NaAlH4 materials.54 A nanocrystalline NaAlH4, synthesised directly by

    ball milling of NaH/Al with TiF3 catalyst under 1525 bar of hydrogen, demonstrated

    fast kinetics with a high reversible capacity of 4.7 wt%.55

    Moreover, recently, Fang et al.56 reported a chemical activation of LiBH4 by

    mechanically milling 3 equivalents of LiBH4 with 1 equivalent of TiF3 that can

    rapidly release 5.7 wt% hydrogen at moderate temperatures (7090 1C) without any

    other gas impurities. However, the isothermal desorption profiles indicated that it

    takes roughly an hour to fully release the hydrogen. Recently, theoretical57,58 and

    experimental5961 investigations have been performed on the Group-II metal

    (Mg and Ca), Group-III metal (Al) and mixed LiMgCa borohydrides owing to

    their potentially high storage capacity (Table 2). While the dehydriding reaction of

    pure Mg(BH4)2 starts at nearly 500 K, the addition of TiCl3 decreased the desorptiontemperature to 361 K,59 whereas the ball-milling had no effect on the dehydriding

    temperature.59a Also, reversible storage in TiCl3-catalysed Ca(BH4)2 has been

    demonstrated with a capacity of 3.8 wt%, when the Ca(BH4)2 was rehydrogenated

    at 623 K under 90 bar for 24 h.60 This improved rehydrogenation condition in

    Ca(BH4)2 suggested further study to exploit its full potential. New reversible

    hydrogen storage reactions have been predicted in the mixed LiMgCa boro-

    hydride systems.57b There remain ample opportunities for further performance

    optimisation and certainly there is a need to understand much more regarding the

    catalytic processes involved in Ti addition to alanates and borohydrides.

    2.3 Complex protonic hydrides; nitrides, imides and amides

    Nitrides and their conjugate hydrogenated compounds imides and amides form the

    second class of complex hydrides although strictly these materials are not hydrides.

    The discovery of reversible hydrogen storage in Li3N by Chen et al.62 in 2002

    Fig. 2 Temperature programmed desorption of hydrogen from NaAlH4 doped with 2 mol%

    of Ti(OBun)4. (Reprinted from ref. 54 with permission from Springer)

    Annu. Rep. Prog. Chem., Sect. A, 2009, 105, 2154 | 27

    This journal is c The Royal Society of Chemistry 2009

    DownloadedbyUniversityofConnecticuton07

    October2011

    Publishedon13

    May2009onhttp://pubs.rsc.org|doi:10.1

    039/B818951J

  • 8/3/2019 Mandal_H2 Storage Materials

    9/35

    spurred intense activity in investigating the LiNH system and subsequently

    quaternary LiANH (where A = Group I and Group II metals) imide-amide-based

    systems.63101

    The key chemical processes that occur during the hydrogenation and dehydro-

    genation of Li3N are represented by eqn (5).

    Li3N + 2 H22 Li2NH + LiH + H22 LiNH2 + 2 LiH (5)

    Although the theoretical gravimetric hydrogen storage capacity of this system is11.5 wt%, experimentally, the amount of stored hydrogen is less by B2 wt%,effectively storing between 9.310 wt% of hydrogen (Fig. 3).

    In spite of the fact that the above chemical process offers potentially very high

    storage capacity, the system suffers from various drawbacks in terms of high

    operational temperature (especially in desorption), reversibility (only the imide-

    amide step is reversible at practical working temperatures) and poor kinetics.

    Whereas the second step in eqn (5) involving the interconversion of imide and

    amide is readily reversible, the reverse dehydrogenation step to revert to the parent

    lithium nitride and hydrogen occurs at high temperature, under low pressure(4320 1C at 105 mbar). Immediately following the report of the (de)hydrogenation

    reactions in the LiNH system, a major research effort focused on reducing

    the desorption temperature for LiNH2/LiH, optimising performance and latterly

    attempting to understand the storage mechanism.6370

    Pressure-composition (P-C) measurements performed on lithium imide show that

    B6.5 wt% of hydrogen can be reversibly stored at 255 1C according to the following

    reaction.

    Li2NH + H22 LiNH2 + LiH (6)

    The enthalpy change for this reaction is calculated to be 44.5 kJ mol1

    H2 leadingto favourable and reversible reaction thermodynamics.70 However, the temperature

    Fig. 3 Gravimetric hydrogen absorption (Abs) and desorption (Des) curves in the

    lithium nitride-imide-amide system (Reprinted from ref. 62, with permission from Macmillan

    Publishers Ltd.).

    28 | Annu. Rep. Prog. Chem., Sect. A, 2009, 105 , 2154

    This journal is c The Royal Society of Chemistry 2009

    DownloadedbyUniversityofConnecticuton07

    October2011

    Publishedon13

    May2009onhttp://pubs.rsc.org|doi:10.1

    039/B818951J

  • 8/3/2019 Mandal_H2 Storage Materials

    10/35

    of hydrogen cycling remains high for practical on-board storage. Therefore,

    chemical destabilisation of LiNH2 itself was an early target in developing new

    materials with low desorption temperatures (Tdes). The substitution of more

    electronegative elements than Li in LiNH2 was suggested66,68 to weaken the bonding

    between Li+ and [NH2] thereby reducing Tdes. Synthesis of Mg-substituted LiNH2

    from LiMg alloys and investigation of dehydrogenation reactions showed a drastic

    decrease of desorption temperature fromB

    2501

    C in pure LiNH2 to an onsettemperature ofB100 1C in (Li,Mg)NH2.71

    However, a theoretical study of the modified amides, (Li,Mg)NH2, showed that

    the NH bonding is weakened in (Li,Mg)NH2 compared to LiNH2.72 In a later

    study, Zhang and Alavi pointed out that the relative strength of the metalN

    bonding was the key factor in explaining the experimental results of reduced

    desorption temperature in Mg-substituted LiNH2.73

    In an effort to gain mechanistic understanding of the hydrogen storage process in

    the reaction between LiNH2 and LiH (eqn (6)), the work of Hu and Ruckenstein64,65

    and Ichikawa et al.67 has led to the proposition of an ammonia mediated mechanism

    (eqn (7) and (8)).

    2 LiNH2- Li2NH + NH3 (7)

    NH3 + LiH- LiNH2 + H2 (8)

    It is worth pointing out that the above elementary two step pathway was confirmed

    recently by in situ 1H NMR spectroscopy.70c However, others63,74,75 believed the

    dehydrogenation reaction to be redox or acidbase type where protonic hydrogen

    (Hd+) in LiNH2 and hydridic hydrogen (Hd) in LiH combine to give molecular

    hydrogen (H2). The exothermic combination of H

    and H

    +

    probably drives thehydrogen desorption reaction,3a which also successfully explains the chemical

    destabilisation in amide-hydride systems.

    H+ + H- H2 (DH = 17.37 eV) (9)

    This, in practice, was a major driving force in the future materials design of complex

    hydride systems, most notably in multi-component systems where chemical tuning of

    amides through doping and use of hydrides and borohydrides as secondary

    components in composite mixtures have met with some success. (These are discussed

    in section 2.4 in more detail).Among other approaches, the use of catalysts and modification of the micro-

    structure via mechanical milling have been adopted to improve the performance of

    nitrogenous systems. While Ichikawa et al.70 studied the effect of ball milling and

    catalysts on the desorption of hydrogen, Hu and Ruckenstein76 investigated the

    storage performance of amides based on preparative routes. Ichikawa et al.67,70 have

    shown that upon ball milling the hydrogen desorption temperature range decreases

    to 180400 1C. Moreover, when catalysts were used during ball milling, the

    desorption temperature was reduced further and evidence of ammonia evolution

    could be almost negligible. The best performance was obtained when 1% of TiCl 3

    was added as catalyst resulting in 5.5 wt% of hydrogen being released between150250 1C with no ammonia formation.67 Importantly, however, the process by

    which these catalyst additives work and the mechanism by which they influence

    reaction pathways is not yet known. In fact, the connections between processes at the

    solid surface and those that occur in the bulk are vital ones to be made if these and

    Annu. Rep. Prog. Chem., Sect. A, 2009, 105, 2154 | 29

    This journal is c The Royal Society of Chemistry 2009

    DownloadedbyUniversityofConnecticuton07

    October2011

    Publishedon13

    May2009onhttp://pubs.rsc.org|doi:10.1

    039/B818951J

  • 8/3/2019 Mandal_H2 Storage Materials

    11/35

    other storage materials are to be understood and a step change in performance

    effected.

    Among several imide-amide systems, the role of structural modification on

    hydrogen storage kinetics in simple Li2NH itself is not well understood. Recently,

    Yang et al. have studied the crystal structure and phase transitions of a-, b- and

    g-Li2NH.90 It was mentioned that the phase transitions in Li2NH is dependent on the

    synthetic methods. Moreover, among all the modifications, the a-phase showedhigher reversible hydrogen storage capacity (6.8 wt%) and faster kinetics.76

    Fig. 4 Hydrogen absorption at 230 1C and 7 atm initial hydrogen pressure (A) a-Li2NH,

    (B) b-Li2NH and (C) g-Li2NH. In (B): a, b and c corresponds to Li2NH synthesised by LiNH2decomposition under vacuum at 230, 280 and 350 1C, respectively (Reprinted from ref. 76b with

    permission from American Chemical Society).

    30 | Annu. Rep. Prog. Chem., Sect. A, 2009, 105 , 2154

    This journal is c The Royal Society of Chemistry 2009

    DownloadedbyUniversityofConnecticuton07

    October2011

    Publishedon13

    May2009onhttp://pubs.rsc.org|doi:10.1

    039/B818951J

  • 8/3/2019 Mandal_H2 Storage Materials

    12/35

    Volumetric hydrogen desorption studies show that while a-Li2NH absorbs

    5.4 wt% in 10 min, the b- and g-polymorphs display much slower kinetics with less

    than 2 wt% and 4 wt% reversible hydrogen capacity, respectively (Fig. 4).76

    Thehigher storage capacity and faster kinetics have been attributed to the particle size

    that varies with thermal treatments under the synthesis condition.76 Nevertheless, it

    was also found that doping with Mg or Ca can reduce the dehydrogenation

    temperature of hydrogenated Li2NH.77,95,99

    The structural and compositional changes in the LiNH system have been

    investigated by ex situ synchrotron X-ray diffraction to elucidate the processes that

    occur during hydrogen uptake-release cycling in the solid state.93 The hydrogenation

    dehydrogenation of the imide-amide system occurs through non-stoichiometric

    intermediates, Li2xNH1+x (0 r x r 1), based on a cubic anti-fluorite structure

    (Fig. 5). More importantly, the process seems to generate Frenkel-type defects,which in turn creates a mechanism for Li+ and H+ ion mobility. Subsequent

    structural studies by in situ neutron diffraction during deuterationdedeuteration

    of Li3N indicated cell volume changes of Li2NH that was attributed to defect

    formation in cubic Li2NH, implying non-stoichiometric compounds.94

    A recent study98 by synchrotron X-ray and neutron diffraction of the LiMgNH

    system with the amide-hydride mixtures identified a mixed imide a-Li2Mg(NH)2phase that has a superstructure based on the high temperature cubic anti-fluorite

    Li2NH.102,103 The replacement of 2 Li atoms by one Mg atom creates ordered cation

    vacancies and results in a Li2NH supercell. The imide protons are also ordered,

    orienting the NH bonds between two vacant cation sites.98 At higher temperatures

    this a-Li2Mg(NH)2 phase undergoes a structural transition to b- and g-Li2Mg(NH)2(similar to the high temperature Li2NH phase)

    104 where the Li, Mg and vacancies are

    randomly distributed over the tetrahedral sites. The structural transitions are

    believed to be driven by progressive disordering of the cations and vacancies as

    the temperature increases.98

    A dramatic particle size-dependent catalytic enhancement of hydrogen absorption

    desorption has recently been reported in the LiMgNH system.101 A three-

    dimensional diffusion-controlled kinetic mechanism has been identified for the

    dehydrogenation process. The hydrogenation and isothermal dehydrogenationcurves (Fig. 6) reveal that the onset of hydrogen absorption was, remarkably,

    lowered to 80 1C for a sample ball-milled for 36 h and consequently the dehydro-

    genation kinetics at 160 1C were notably improved compared to 3 h milled or hand

    milled samples.101

    Fig. 5 Structural transformation of lithium imide, Li2NH, to lithium amide, LiNH2, via

    non-stoichiometric intermediates (Reprinted from ref. 12b, with permission from The Japan

    Chemical Journal Forum and Wiley Periodicals).

    Annu. Rep. Prog. Chem., Sect. A, 2009, 105, 2154 | 31

    This journal is c The Royal Society of Chemistry 2009

    DownloadedbyUniversityofConnecticuton07

    October2011

    Publishedon13

    May2009onhttp://pubs.rsc.org|doi:10.1

    039/B818951J

  • 8/3/2019 Mandal_H2 Storage Materials

    13/35

    From the close similarity of the mechanisms in the LiNH and LiMgNHsystems, it is apparent that the vacancy distribution, defect concentration and the

    ionic mobility of Li+ and H+ play an important role in the storage processes.

    Recently, the crystal structure of a ternary imide, Li2Ca(ND)2, was determined by

    NPD.100a The structure of Li2Ca(ND)2 resembled a combined imide structure

    with an ordered arrangement of Ca[NH]6 octahedra and Li[NH]4 tetrahedra

    alternating along the c-axis, distinctly different from any of the polymorphs of

    Li2Mg(NH)2.100a The Li vacancy was attributed to defect reactions at elevated

    temperatures and the dehydrogenation mechanism was proposed to be mediated by

    small mobile ion migration in both amide and hydride.27b,100 Similar factors may

    well be fundamental to other complex hydride systems and could eventually lead tothe improved understanding and better design of materials. Further work is required

    with respect to understanding and exploiting the defect chemistry, complex

    phase behaviour and the involvement of chemical intermediates in these and similar

    systems.

    Fig. 6 (A) Gravimetric hydrogen uptake versus temperature for Li2MgN2H2 as a function of

    milling time and (B) Isothermal dehydrogenation curves (at 160 1C) of hydrogenated

    Li2MgN2H2 samples (Reprinted from ref. 101, with permission from American Chemical

    Society).

    32 | Annu. Rep. Prog. Chem., Sect. A, 2009, 105 , 2154

    This journal is c The Royal Society of Chemistry 2009

    DownloadedbyUniversityofConnecticuton07

    October2011

    Publishedon13

    May2009onhttp://pubs.rsc.org|doi:10.1

    039/B818951J

  • 8/3/2019 Mandal_H2 Storage Materials

    14/35

    2.4 Other multi-component systems: amide-hydride-borohydrides

    The underlying strategy behind employing more than one component is to alter the

    overall thermodynamics of hydrogen release, lowering Tdes, in addition to improving

    the kinetics of sorption. Indeed, it has been observed that in multinary systems the

    hydrogen uptake-release temperatures are lower and the kinetics often better than

    for the individual component systems. Performance is highly dependent on the

    reaction pathway of hydrogenation and dehydrogenation, together with the stabilityof the product formed in the dehydrogenated state.

    The report of reversible hydrogen storage by Chen et al.62,63 in a mixture of LiH

    and LiNH2 initiated a flurry of research activity, which later opened up a number of

    diverse routes for complex hydride destabilization in the form of binary, ternary or

    multinary reactive mixtures involving amides, hydrides, borohydrides and alanates

    with substantial improvements in (de)hydrogenation thermodynamics.105137 The

    various multinary combinations studied can be classified mainly as binary

    (two component) and ternary (three component) systems. Moreover, the binary

    systems can be considered as simple when the combination involves a single metal

    hydride and an imide/amide, as compared to complex binaries where a hydride

    containing a complex anion, such as a borohydride or aluminohydride, is involved.

    As the number of components increases, the system becomes more and more

    complex often due to the increased number of intermediates (and/or reaction steps)

    involved in the hydrogenationdehydrogenation cycles and the non-trivial interplay

    between structures and reactivities.

    Simple binary component systems: Amides-hydrides. The lithium amidelithium

    hydride [LiNH2 + LiH] combination, containing both protonic and hydridic

    components was perhaps the first step towards a reversible and light-weight multi-component hydrogen store.62,63 The reactions with varying amide:hydride ratios and

    their mechanisms have been widely investigated. With the realisation that the

    LiNH2LiH system was restricted in its ability to release hydrogen under favourable

    and practical conditions, several alternative binary combinations of amides and

    hydrides have been investigated.74,105133 In addition to reacting Group-II hydrides

    such as MgH2 or CaH2 with LiNH2, another option was to react binary hydrides

    (such as LiH) with Mg(NH2)2 with the reasoning that the latter amide has a lower

    decomposition temperature than the lithium counterpart, LiNH2. The initial studies

    by Chen and co-workers reacted LiNH2 with MgH2 or CaH2 with the aim of

    destabilising the ternary imides.74 Their results showed a remarkable drop in thehydrogen uptake and release temperatures as revealed by temperature programmed

    absorption and desorption studies (Fig. 7).

    Based on the starting composition, the chemical formula of the compound

    produced in the LiNH2:MgH2 2:1 reaction was proposed to be Li2MgN2H2.

    Interestingly, after the rehydrogenation step, the initial components (LiNH2 and

    MgH2) were not regenerated and instead Mg(NH2)2 and LiH were the products.74

    Subsequent studies of the LiMgNH system with varying ratios of Mg(NH2)2:LiH

    (or LiNH2:MgH2)105110 demonstrated that an onset of hydrogen desorption as low

    as 150 1C was achievable while maintaining up to B5 wt% of hydrogen in these

    systems.74,106108

    Recently, Alapati et al.111 predicted a 1:1 LiNH2-MgH2 hydrogen storage system

    with 8.19 wt% capacity by density functional theory (DFT) calculations:

    LiNH2 + MgH22 LiMgN + 2 H2 (10)

    Annu. Rep. Prog. Chem., Sect. A, 2009, 105, 2154 | 33

    This journal is c The Royal Society of Chemistry 2009

    DownloadedbyUniversityofConnecticuton07

    October2011

    Publishedon13

    May2009onhttp://pubs.rsc.org|doi:10.1

    039/B818951J

  • 8/3/2019 Mandal_H2 Storage Materials

    15/35

    First-principles calculations of total energies and vibrational free energies for thesame binary composition have suggested the following sequential reactions:112

    LiNH2 + MgH2- 1/2 Mg(NH2)2 + LiH + 1/2 MgH2

    - 1/4 Mg(NH2)2 + 1/4 Mg3N2 + LiH + H2

    - 1/4 Li2MgN2H2 + 1/4 Mg3N2 + 1/2 LiH + 3/2 H2

    - LiMgN + 2 H2 (11)

    However, experimentally a new intermediate compound with compositionLi2Mg2(NH)3 was identified113 on dehydrogenation of the initial mixture at

    210 1C although this mixed imide phase was known long ago.114

    LiNH2 + MgH2- 1/3 Li2Mg2(NH)3 + 1/3 MgH2 + 1/3 LiH + H2 (12)

    A recent XRD and IR investigation showed that the 1:1 LiNH2MgH2 systemreleased 6.1 wt% H2 first during ball milling and then by subsequent heating thatfollowed a four step reaction.115

    LiNH2 + MgH2- 1/2 Mg(NH2)2 + LiH + 1/2 MgH2 (12 h ball milled)

    -1/2 MgNH + 1/4 Mg(NH2)2 + LiH + 1/4 MgH2 + 1/2 H2 (36 h ball milled)

    -1/2 Li2MgN2H2 + 1/2 MgNH + 1/2 LiH + 1/4 MgH2 + H2 (at 220 1C)

    -1/4 Mg3N2 + 1/4 Li2MgN2H2 + 1/2 LiH + 3/2 H2 (at 390 1C) (13)

    Fig. 7 Temperature-programmed absorption (a) and desorption (b) curves for (I) 2LiNH2 +

    CaH2, (II) 2LiNH2 + MgH2 and (III) Li2NH samples (Reprinted from ref. 74, with permission

    from Wiley-VCH).

    34 | Annu. Rep. Prog. Chem., Sect. A, 2009, 105 , 2154

    This journal is c The Royal Society of Chemistry 2009

    DownloadedbyUniversityofConnecticuton07

    October2011

    Publishedon13

    May2009onhttp://pubs.rsc.org|doi:10.1

    039/B818951J

  • 8/3/2019 Mandal_H2 Storage Materials

    16/35

    Nevertheless, the predicted product lithium magnesium nitride, LiMgN, was notrealized in this study and only 1 equivalent of hydrogen was rechargeable under highhydrogen pressure.115

    Given that in the initial studies of the reactions of LiNH2 with MgH2 an initial

    exchange reaction occurred to produce Mg(NH2)2 on dehydrogenation and subsequent

    rehydrogenation, further explorations employed various ratios of Mg(NH2)2 and

    LiH, as starting materials. According to Nakamori et al.,116 the following overall

    reaction occurs for a 1:4 Mg(NH2)2LiH ratio:

    Mg(NH2)2 + 4 LiH2 1/3 Mg3N2 + 4/3 Li3N + 4 H2 (14)

    Although a calculated value of 9.1 wt% of hydrogen was expected from the above

    reaction, the experimental weight loss at 527 1C was only B7 wt%. However,

    importantly, the major weight loss occurred below 230 1C which is lower than the

    equivalent desorption in the LiNH2 + 2 LiH system. Amide:hydride ratios of 1:2

    and 3:8 for the Mg(NH2)2 LiH system were also investigated, viz:74,107,109,117121

    Mg(NH2)2 + 2 LiH2 Li2Mg(NH)2 + 2 H2 (15)

    3 Mg(NH2)2 + 8 LiH2 4 Li2NH + Mg3N2 + 8H2 (16)

    Reactions (15) and (16) were expected to release 5.6 and 6.9 wt% of hydrogen. The

    above combinations were systematically investigated by thermogravimetry, PCT and

    XRD studies.120,121 While the pressure-composition isotherms were similar at

    250 1C, the quantity of released hydrogen was more with higher LiH content.

    (5.4, 4.5 and 5.1 wt% for reaction 14, 15 and 16, respectively).120 Moreover, the mass

    spectrometry profiles atB190 1C indicated an increase of ammonia production for

    compositions with smaller equivalents of LiH.

    According to the report of Aoki et al. all the reactions, Mg(NH2

    )2

    + 2 LiH,

    3 Mg(NH2)2 + 8 LiH and Mg(NH2)2 + 4 LiH, showed similar desorption PC

    isotherms with the formation of Li2Mg(NH)2 as the final product at 250 1C and the

    equilibrium dehydriding reactions had no serious effect on the quantity of LiH.121

    The structure of the intermediate imide Li2Mg(NH)2 was previously described in

    section 2.3. One fascinating outcome is the identification of a Li-rich and H-poor

    imide, Li2+xMgN2H2x, in the 3 Mg(NH2)2 + 8 LiH system studied by Nakamura

    and co-workers.119b The intermediate phase was synthesized and characterized as

    Li2.6MgN2D1.4 (using LiND2 and D2 gas)119b by combined synchrotron X-ray and

    neutron diffraction experiments. The structure of Li2.6MgN2D1.4 has the same

    symmetry as the previously described Li2Mg(NH)2 but with a smaller cell dimension(Fig. 8). In addition, in the non-stoichiometric imide structure 30% of the D-sites are

    vacant [as compared to fully occupied sites in Li2Mg(NH)2] and the adjacent cation

    Fig. 8 Structure of LiMgND intermediate (left) and formation scheme for

    Li2+xMgN2H2x (right) (Reprinted from ref. 119b, with permission from Elsevier B. V.).

    Annu. Rep. Prog. Chem., Sect. A, 2009, 105, 2154 | 35

    This journal is c The Royal Society of Chemistry 2009

    DownloadedbyUniversityofConnecticuton07

    October2011

    Publishedon13

    May2009onhttp://pubs.rsc.org|doi:10.1

    039/B818951J

  • 8/3/2019 Mandal_H2 Storage Materials

    17/35

    sites are 20% occupied (by contrast, vacant in Li2Mg(NH)2). Based on this, the

    following scheme has been proposed for the formation of the intermediate defect

    imide phase mediated by interfacial diffusion of hydrogen (H from LiH and H+

    from LiMgNH phase) (Fig. 8).

    More in situ studies by synchrotron and neutron diffraction would enhance the

    understanding of the hydrogen uptake-release processes that occur in specific

    systems and the role of intermediates. A deeper understanding of intermediatestructures and their formation mechanisms are key elements in the future design of

    materials.

    Complex binary component systems: Amides-complex hydrides. Similar to the

    simple binary mixtures, complex binary composite systems are considered here as

    reactive mixtures of complex hydrides, such as borohydrides or aluminohydrides

    with amides or binary hydrides. Thermodynamic destabilisation with combined

    chemical systems occurs either through chemical destabilisation of the hydrogenated

    state or by formation of stable products on dehydrogenation.3b An illustration of the

    chemical destabilisation of LiBH4 was provided by Vajo and co-workers with areduced decomposition temperature achieved in a 2:1 mixture of LiBH4 and MgH2as compared to pure LiBH4.

    122 The formation of stable MgB2 (reaction (17)) in

    the dehydrogenated state seems to drive the reaction with an enthalpy gain of

    25 kJ mol1 of H2.122

    2 LiBH4 + MgH22 2 LiH + MgB2 + 4 H2 (17)

    A reversible storage of more than 9 wt% hydrogen (theoretical capacity 11.9 wt%)

    has been demonstrated for reaction (17) when catalysed with 23 mole% of TiCl3.122

    It is worth noting here that a very small amount of LiBH4 mixed with MgH2 alsoenhanced the dehydrogenation/hydrogenation kinetics of MgH2.

    123 The requirement

    of a reduced temperature (168 1C) for an equilibrium hydrogen pressure of 1 bar in

    the 2LiBH4MgH2 system indicates that both LiBH4 and MgH2 are destabilised.27

    Subsequently, similar destabilisation strategies have been employed for many

    other complex hydride combinations.124 With the successful demonstration of the

    hydrogenation reactions (18) and (19), it has been suggested that MgB2 facilitates

    production of metal borohydrides by promoting the formation of the [BH4]

    complex anion.124b

    2 NaH + MgB2 + 4 H2- 2 NaBH4 + MgH2 (18)

    CaH2 + MgB2 + 4 H2- Ca(BH4)2 + MgH2 (19)

    Very recently, another new coupled complex hydride system, LiBH4/CaH2, was

    reported with a theoretical capacity of 11.7 wt% via the following reaction:124f

    6 LiBH4 + CaH22 6 LiH + CaB6 + 10 H2 (20)

    Interestingly, the above reaction, with TiCl3 added, reproducibly stores 9.1 wt%

    hydrogen (95% of the theoretical total accounting for the added wt. of catalyst in the

    calculation) and cycles between LiBH4CaH2 and LiHCaB6 in the hydrogenatedand dehydrogenated state respectively.124f Despite many advantages, this system is

    highly unlikely to be suitable for onboard storage due to the high equilibrium

    temperature at 1 bar H2 (418 1C) predicted from thermodynamic calculations based

    on experimental data under differing conditions.124f

    36 | Annu. Rep. Prog. Chem., Sect. A, 2009, 105 , 2154

    This journal is c The Royal Society of Chemistry 2009

    DownloadedbyUniversityofConnecticuton07

    October2011

    Publishedon13

    May2009onhttp://pubs.rsc.org|doi:10.1

    039/B818951J

  • 8/3/2019 Mandal_H2 Storage Materials

    18/35

    Another destabilised complex binary mixture, which has been investigated

    simultaneously by Aoki et al. and Pinkerton et al., is that of LiBH4 and

    LiNH2.125,126 A dehydrogenation enthalpy of 23 kJ mol1 of H2 was predicted for

    a (2:1) LiNH2 + LiBH4 mixture, which is lower than either of the individual LiNH2and LiBH4 components, according to the following dehydrogenation reaction (21)

    by first-principle calculations.125a

    LiBH4 + 2 LiNH2- Li3BN2 + 4 H2 (21)

    Experimental investigations of this mixture were carried out by ball milling followed by

    XRD analysis and by p-c isothermal measurements. A new phase was identified upon

    ball milling and was indexed in a cubic cell (10.67 A ).125a In a separate study Pinkerton

    et al. reported formation of a quaternary hydride, Li3BN2H8, that formed by either

    ball-milling or heating (above 95 1C) the same binary mixture.126a They also noticed

    melting of Li3BN2H8 at B190 1C followed by Z10 wt% hydrogen release above

    B250 1C.126a Soon after, the compound was identified as a lithium borohydride amide,

    Li4(NH2)3BH4, with a bcc structure (lattice parameter 10.66 A ).127 This phase, however,

    did not match with the starting stoichiometry indicating a non-stoichiometric reactionof LiBH4 with 2 LiNH2. The ideal reaction stoichiometry was determined to be a 3:1

    ratio of LiNH2 and LiBH4 to produce the LiNBH quaternary phase. Although

    Li4(NH2)3BH4 had a high hydrogen content of 11.9 wt% and the major decomposition

    product of this phase above 260 1C was hydrogen, it is non-reversible.127 Recently,

    many new destabilised systems with enhanced hydrogen capacity and thermodynamics

    have been recognised by DFT based computational studies.111,128 However, it is often

    observed that the introduction of a non-hydride second species diminishes the overall

    gravimetric capacity of the mixed system.

    Chemical destabilization of LiAlH4 by LiNH2 was originally shown by Lu and Fang

    employing a thermogravimetric study of a 2LiAlH4 + LiNH2 mixture.129

    The systemreleasedB8.1 wt% of hydrogen between 85 and 320 1C in three dehydrogenation steps

    without any catalysts.129 The reaction involved a two step decomposition of LiAlH4similar to NaAlH4 (eqns (3) and (4)) followed by a third step as per the dehydrogenation

    in eqn (6). The resulting overall reaction can be represented as in eqn (22).129

    2 LiAlH4 + LiNH2- LiH + Li2NH + 2 Al + 4 H2 (22)

    Ball-milling of other molar ratios, (1:1) and (1:2), of the LiAlH4LiNH2 system have

    also been investigated. The (1:1) LiAlH4LiNH2 system has emerged as an inter-

    esting case that releases 6.1 wt% hydrogen under ball-milling as reported by Xiong

    et al. and the enthalpy effect of the H2 release reaction was very low.130 Afterwards,reversible storage was reported by other groups in the LiAlNH system.131,132 For

    example, a Li3AlH63LiNH2 mixture with 4 wt% TiCl31/3AlCl3 releases 7.1 wt%

    of hydrogen in two dehydrogenation steps between 150 and 300 1C.132 It was

    demonstrated that in short-cycle experiments the system is 100% reversible with the

    following overall reaction (23):132

    Li3AlH6 + 3 LiNH22 Al + 3 Li2NH + 9/2 H2 (23)

    Recently, Liu et al. have reported a new chemical system LiAlH4Mg(NH2)2 (1:1)with a hydrogen content of 8.5 wt% and releasing 6.2 wt% under ball-milling near

    ambient temperature.133 The overall reaction was expressed as:

    Mg(NH2)2 + LiAlH4- 1/3 Mg3N2 + 1/3 Li3AlN2 + 2/3 AlN + 4 H2 (24)

    Further investigations are suggested to explore its potential for on-board storageconcentrating on the capacity and reversibility issues.

    Annu. Rep. Prog. Chem., Sect. A, 2009, 105, 2154 | 37

    This journal is c The Royal Society of Chemistry 2009

    DownloadedbyUniversityofConnecticuton07

    October2011

    Publishedon13

    May2009onhttp://pubs.rsc.org|doi:10.1

    039/B818951J

  • 8/3/2019 Mandal_H2 Storage Materials

    19/35

    Ternary composite systems: Amides-hydrides-borohydrides. In addition to simple

    and complex binaries the strategy of hydride destabilisation was extended to ternary

    mixtures.134137 Recently Yang et al. have reported a self-catalysing ternary

    composite LiNH2/LiBH4/MgH2 (in a 2:1:1 molar ratio) that has resulted in

    enhanced low temperature desorption kinetics with reduction in ammonia liberation

    as compared to the state-of-the-art binary composites (Fig. 9).134 It is believed that

    the self-catalysing effect is due to a set of coupled reactions,LiNH2 + LiBH4- Li4BN3H10 (25-1)

    2 Li4BN3H10 + 3 MgH2- 3 Mg(NH2)2 + 2 LiBH4 + 6 LiH (25-2)

    2 Li4BN3H10 + 3 MgH2- 3 Li2Mg(NH)2 + 2 LiBH4 + 6 H2 (25-3)

    Mg(NH2)2 + 2 LiH2 Li2Mg(NH)2 + 2 H2 (25-4)

    where the first step (251) occurs during the milling and the produced Li4BN3H10melts and reacts with MgH2 when the mixture is heated to 100 1C (reaction (252)).

    The first hydrogen desorption begins at 150 1C by reaction (253). The total capacity

    of 8.2 wt% can only be achieved at higher temperatures (up to 375 1C) with furtherreaction of Li2Mg(NH)2 and LiBH4 forming Mg3N2, Li3BN2, traces of LiH plus an

    unknown phase.134c It is thought that pre-formation of Li2Mg(NH)2 in step (25-3)

    provides nuclei for subsequent reaction (step (254)) in a Li4BN3H10 melt (acting as

    Fig. 9 Hydrogen and ammonia desorption kinetics data as a function of temperature for 2:1:1

    LiNH2/LiBH4/MgH2 ternary mixtures and comparison with other unary and binary systems

    (Reprinted from ref. 134c, with permission from Wiley-VCH).

    38 | Annu. Rep. Prog. Chem., Sect. A, 2009, 105 , 2154

    This journal is c The Royal Society of Chemistry 2009

    DownloadedbyUniversityofConnecticuton07

    October2011

    Publishedon13

    May2009onhttp://pubs.rsc.org|doi:10.1

    039/B818951J

  • 8/3/2019 Mandal_H2 Storage Materials

    20/35

    an effective mass transfer medium) and is self-catalysed.134c Another point to note is

    that step (254) is the only reversible one in the entire reaction sequence.

    A subsequent study systematically investigated the impact of MgH2 on the ternary

    system LiNH2LiBH4MgH2 with 2:1:x molar ratios and 0 r x r 1. This study

    suggested that the addition of MgH2 facilitates milling-induced reaction and reduces

    the ammonia release.135 Conversely, however, with the addition of MgH2 the total

    theoretical capacity of the system decreased. This work confirmed the directinvolvement of MgH2 in the first desorption event as x is varied in the LiNH2

    LiBH4MgH2 ternary mixture. It corroborated the previous study for the (2:1:1)

    system (i.e., when x = 1) in demonstrating that this was the optimum composition

    for low temperature hydrogen release (B4 wt%) and indicated future opportunities

    for further kinetic enhancements.135

    In a study of another ternary mixture, Mg(NH2)22LiHLiBH4, and with

    comparison with the LiNH2LiBH4MgH2 ternary system above, the former

    showed very similar hydrogen desorption, thermal effects and structural changes.136

    However, Mg(NH2)22LiH0.67LiBH4 gave 9.1 wt% hydrogen at 300 1C, higher

    than the total theoretical capacity of the 2 LiNH2MgH2LiBH4 composite.136

    Theeffects of varying the reaction stoichiometry in the (LiNH2)X(LiBH4)Y(MgH2)Zternary system (ratio x:y:z) have been recently investigated by Sudik et al.137 Among

    the 2:1:2, 1:1:1, 2:0.5:1 and 2:1:1 ratios tested (in ref. 134), only 2:1:1 and 2:0.5:1 were

    reversible. The 2:0.5:1 stoichiometry had a maximum capacity of 8.6 wt% of which

    only 3.5 wt% was reversible at low temperature (180 1C).

    2.5 Ammonia-borane and amido-boranes

    Ammonia-borane (NH3BH3) is considered to be an attractive candidate within the

    chemical storage category largely due to its safety features and a high hydrogencontent of 19.6 wt% that exceeds that of gasoline.138141 NH3BH3 is a colourless

    solid that is stable at ambient temperature and was first synthesised in 1955.142 Many

    other preparative methods, both in solution and in the solid state, have since been

    described.143146 In the context of amide-hydride destabilization combining the

    protonic and hydridic hydrogens, NH3BH3 is a remarkable compound in that both

    hydridic BH and protonic NH bonds exist within the same molecule.147 Moreover,

    the strong BN bond prevents ammonia and borane emission prior to hydrogen

    release under most thermolysis conditions. Although the detailed mechanism for the

    decomposition of NH3BH3 is not clear yet, significant progress has been made

    towards understanding its hydrogen release reactions.141,148156During thermal decomposition several intermediates have been reported and the

    desorption products vary depending on the thermolysis conditions (temperature

    range, ramping rate). Intermediate species such as polyaminoborane, [NH2BH2]n,

    polyiminoborane, [NHBH]n, and even a small fraction of borazine, [N3B3H6], have

    been identified as part of the dehydrogenation process. According to Wolf and

    co-workers149,150 the stepwise solid state decomposition of NH3BH3 to release

    12 wt% H2 can be represented as:

    n NH3BH3 (s)- [NH2BH2]n (s) + n H2 (g) (26-1)

    [NH2BH2]n (s)- [NHBH]n (s) + n H2 (g) (26-2)

    [NH2BH2]n (s)- [N3B3H6]n/3 (l) + n H2 (g) (26-3)

    NH3BH3 decomposition often requires a long induction period. Use of ionic solventshas been shown to reduce the induction period and speed up the kinetics of hydrogen

    Annu. Rep. Prog. Chem., Sect. A, 2009, 105, 2154 | 39

    This journal is c The Royal Society of Chemistry 2009

    DownloadedbyUniversityofConnecticuton07

    October2011

    Publishedon13

    May2009onhttp://pubs.rsc.org|doi:10.1

    039/B818951J

  • 8/3/2019 Mandal_H2 Storage Materials

    21/35

    release.152 Several approaches have been employed to reduce dehydrogenationtemperatures,139 increase the kinetics of H2 release

    152,157168 and minimize borazineemission.152 Use of transition metals,157,158 base-metal catalysts,162 acid catalysis,163

    nanoscaffolds,139 ionic liquids152 and carbon cryogels164 are reported to improvedehydrogenation properties.

    Recently, a significant development in the hydrogen release temperature has been

    reported by Xiong et al. in alkali-metal amidoboranes, LiNH2BH3 and NaNH2BH3,

    releasing 10.9 and 7.5 wt% of hydrogen (Fig. 10), respectively, at B90 1C via:169

    n LiNH2BH3 (s)- (LiNBH)n (s) + 2n H2 (g) 10.9 wt% (27)

    or

    n NaNH2BH3 (s)- (NaNBH)n (s) + 2n H2 (g) 7.5 wt% (28)

    In addition, the compounds do not release unwanted borazine by-product and the

    reactions are devoid of lengthy induction periods and foaming. Moreover, the

    dehydrogenation process is less exothermic than NH3BH3 (the enthalpies of

    dehydrogenation from Li- and Na-amidoborane are 3 and 5 kJ mol

    1 vs. 5 kJ mol

    1

    for NH3BH3), which would facilitate its off-board materials regeneration.169,172

    Another derivative of NH3BH3, solvent-containing calcium amidoborane,

    Ca(NH2BH3)2 2THF,170,171 has also been reported with enhanced dehydrogenation

    kinetics and suppressed borazine release. Contrary to previous reports, a recent

    study indicated a significant amount of ammonia emission from the sodium

    amidoborane, thereby decreasing the practical hydrogen content and suitability

    for use in fuel cells.173

    Despite the several advantages of ammonia-borane and amidoborane systems, the

    most serious drawback is the lack of facile reversibility which makes them unsuitable

    for on-board storage applications. The thermal decomposition of ammonia-borane

    and amidoboranes is less exothermic in comparison to their hydrolysis, making them

    nearly thermoneutral which would greatly facilitate their off-board regeneration

    Fig. 10 Time evolution of hydrogen release from alkali amidoboranes and post milled

    ammonia-borane at about 91 1C (Reprinted from ref. 169, with permission from Macmillan

    Publishers Ltd.).

    40 | Annu. Rep. Prog. Chem., Sect. A, 2009, 105 , 2154

    This journal is c The Royal Society of Chemistry 2009

    DownloadedbyUniversityofConnecticuton07

    October2011

    Publishedon13

    May2009onhttp://pubs.rsc.org|doi:10.1

    039/B818951J

  • 8/3/2019 Mandal_H2 Storage Materials

    22/35

    from solid decomposition products (BNHx). In spite of major progress, significant

    challenges remain for the development of efficient off-board regeneration of ammonia-

    borane. With an efficient regeneration procedure, these systems could possibly serve

    as an alternative to other complex chemical hydrides as a sustainable hydrogen

    storage medium.

    3. Physical storage

    Unlike chemical storage where the hydrogen is held chemically to component atoms

    within a solid (formation of chemical bonds), in physical storage systems the

    hydrogen is attached/included to/in a host-a surface or porous solid-by much

    weaker interactions/forces commonly defined as a process of physical adsorption

    or physisorption. Coordination polymers, metal organic frameworks (MOFs),

    polymers with intrinsic microporosity (PIMs), activated carbons and zeolites are

    among the widely investigated examples of materials for hydrogen storage that

    operate via physisorption.3b In physisorption the forces of attraction between the

    hydrogen and the host originate mainly from weak van der Waals interactions,

    thereby restricting any significant hydrogen adsorption to that at very low tempera-

    tures and/or at very high pressures. Notable progress has been made during the past

    few years with gas storage in MOFs using innovative materials design to improve

    their performance in terms of increased storage capacity.

    Metal-organic frameworks (MOFs)

    MOFs (also known as coordination frameworks, coordination network materials or

    coordination polymers) can broadly be thought of as synthetic analogs of naturally

    occurring zeolitic materials, mostly with a three dimensional framework. These

    frameworks are interconnected in such a fashion that ordered networks of channels

    and pores are generated within the structure. Due to their porous structures, various

    small organic molecules, solvent molecules and gas molecules can diffuse in and out

    of the channel structures, sometimes selectively. The primary building blocks that

    form the MOFs are multidentate organic ligands (linkers) coordinating to a metal

    ion or a cluster of metal atoms. MOF materials often form with solvent molecules

    trapped within the pores and following removal of these solvent guests, the host

    materials are typically high in surface area and of low density.

    The first report of a MOF as a potential hydrogen storage medium was published

    by Yaghi and co-workers in 2003.

    174

    They demonstrated an uptake of 4.5 wt%of hydrogen at 78 K and 1 wt% at ambient temperature and a pressure of 20 bar

    in a Zn-containing framework material (MOF-5) of composition Zn4O(BDC)3(BDC = 1,4-benzenedicarboxylate).174 They also reported up to 2 wt% of hydrogen

    uptake in topologically similar (replacing BDC with naphthalene linkers) isoreticular-

    MOF-8 (IRMOF-8) at room temperature and 10 bar.174 After these remarkable

    reports, the hydrogen uptake in MOF-5 was established independently at about

    4.55.2 wt% at 77 K and B50 bar.175

    Realising their high potential as new materials for hydrogen storage, Rowsell and

    Yaghi outlined several strategies to improve the storage capacity of MOFs under

    more favourable working conditions.176 A variety of parameters, such as pore size,adsorption energy, impregnation, catenation and incorporation of open metal sites

    and lighter metals in MOFs were proposed that can be tuned for performance

    optimisation.176 Many of these parameters are interrelated so, for example, pore size

    can be modified by changing organic linkers or by catenation that occurs with

    Annu. Rep. Prog. Chem., Sect. A, 2009, 105, 2154 | 41

    This journal is c The Royal Society of Chemistry 2009

    DownloadedbyUniversityofConnecticuton07

    October2011

    Publishedon13

    May2009onhttp://pubs.rsc.org|doi:10.1

    039/B818951J

  • 8/3/2019 Mandal_H2 Storage Materials

    23/35

    interpenetration or interweaving of networks. Moreover, impregnation of another

    adsorbent within large-pore MOFs can lead to reduced pores. Pore size, in turn, has

    a direct influence on the hydrogen adsorption enthalpy. A pore size comparable to

    the diameter of molecular hydrogen should increase the van der Waals interaction

    thereby increasing the enthalpy.176 Similarly, higher enthalpy would also result from

    interaction of hydrogen with frameworks having coordinatively unsaturated metal

    centres or open metal sites.Subsequently, a variety of framework materials have been synthesized and

    investigated for hydrogen storage applications.175189 Hydrogen uptake studies on

    a series of MOFs (IRMOF-1, -6, -11, -20, -74, MOF-177 and HKUST-1) reveal that

    saturation uptake correlates with the surface area.175a,d At 77 K the saturation

    uptake and the saturation pressure varied widely in the series reaching a strikingly

    high capacity of 6.7 and 7.5 wt% in IRMOF-20 and MOF-177 at saturation

    pressures of 70 and 80 bar, respectively.175a Many of the initial studies aimed at

    highly porous frameworks with high surface areas. Another recent report by Lin

    et al. showed high hydrogen uptake of 6.07 wt% at 20 bar (at 77 K) and

    projected 7.01 wt% from Langmuir plots for MOFs with a BET surface area of2932 m2 g1.181 It has been demonstrated in MOFs that high surface area and pore

    volume are approximately proportional to the amount of stored hydrogen.175a,181,190

    Very high storage capacities have also been reported in a series of MIL (Mate riaux

    de lInstitut Lavoisier) MOFs possessing extremely high surface areas and concen-

    trations of unsaturated metal sites. It is noteworthy that MIL-101 (based on

    Cr-carboxylates) adsorbs large amounts of hydrogen reaching nearly 6.1 wt% at

    77 K.188 Its high heat of adsorption (10 kJ mol1) at low pressures was believed to be

    related to small pore sizes.188

    It must be pointed out that the maximum hydrogen uptake (up to the plateau of

    isotherm at 77 K) in the MOFs varies almost linearly with the specific surface area.

    However, a conflicting scenario appears with large pores. While a large pore volume is

    required to enhance storage capacity, the interaction strength with hydrogen becomes

    weak with wider pores. In a Ni-based nanoporous framework with flexible linkers, the

    hydrogen was shown to confine within the pores where the entry to the pores was

    controlled by dynamical opening of the pore windows.191 This implies that once hydrogen

    is adsorbed at high pressure it can be stored at low pressure. Moreover, the metal centre

    and the organic linker units were identified as the two major hydrogen binding sites in the

    MOFs by inelastic neutron scattering experiments and theoretical studies.192,193

    MOFs with unsaturated coordination positions or open metal sites have beenexplored and shown to be attractive for hydrogen storage.177,178,184,192194 For

    example, HKUST-1 or Cu-BTC based on Cu(COO)4 unit and benzene-1,3,5-

    tricarboxylate linkers possessing smaller pores (as compared to MOF-5) and Cu(II)

    open metal sites show enhanced hydrogenadsorbent interactions and consequently

    higher heats of adsorption.177,184,194196 Many possibilities for creating MOFs with

    exposed metal sites have been reviewed recently.197 It is proposed that increasing the

    hydrogen binding affinity and accordingly imparting a high adsorption enthalpy can

    be achieved by systematic incorporation of unsaturated metal ions. Bhatia and

    Myers calculated the optimal enthalpy of adsorption to be 15.1 kJ mol1 for storage

    at 298 K with operating pressures between 1.5 and 30 bar.198 Limitations due todecreased H2 adsorption with increasing temperature have been improved in

    some MOFs by increasing H2-adsorbent surface interactions.184,199201 Recently,

    strong interaction between hydrogen and exposed metal sites within MOFs

    have been demonstrated with very high heat of adsorption values of 12.3 and

    42 | Annu. Rep. Prog. Chem., Sect. A, 2009, 105 , 2154

    This journal is c The Royal Society of Chemistry 2009

    DownloadedbyUniversityofConnecticuton07

    October2011

    Publishedon13

    May2009onhttp://pubs.rsc.org|doi:10.1

    039/B818951J

  • 8/3/2019 Mandal_H2 Storage Materials

    24/35

    13.5 kJ mol1.202,203 The total stored amount was dependent on surface area and

    pore volume which led to the conclusion that, a high surface density of strong

    adsorption sites is required for efficient storage under workable conditions.203

    Recently, many so-called covalent organic frameworks (COFs) have been syn-

    thesised and structurally characterised.204208 By contrast to MOFs, within COFs

    the organic structural building units are held together by strong covalent bonds

    (for example, CC, CO, BO, SiC) as opposed to metals. The resulting solidspossess a very low density and an ultrahigh porosity. These properties of COFs make

    them excellent candidates for hydrogen storage. Theoretical studies on prototypical

    COFs predict that storage of up to 10 wt% H2 in COF-5 (at 80 bar) and COF-8

    (at 100 bar) is possible.204,207 In COF-8 an excess hydrogen uptake of 10 wt% has

    been found experimentally at 77 K and 100 bar (Fig. 11).208 These values are

    significantly higher than the highest values reported in MOFs (7 wt% in MOF-177

    and 7.1 wt% in MOF-5) although the pressure employed is quite high at 77 K. 175a,209

    Also, the 3-D COFs show 2.5 to 3 times higher storage capacity than the 2-D COFs.

    The associative hydrogen storage in COF-8 is the highest value reported for a

    framework material.208

    Another new development for increasing the hydrogen uptake of MOFs is the

    hydrogen spillover technique.210 By employing this technique Li et al. reported an

    eightfold enhancement in hydrogen uptake in MOF-5 and IRMOF-8.210 The

    enhancement was attributed to the secondary spillover of hydrogen atoms from a

    Pt/activated carbon (AC) catalyst to the surface of MOFs in a composite system.

    Further, significant improvement in H2 uptake capacity at near ambient temperature

    (30 1C) has been predicted in Li-doped MOF-C30 with a gravimetric uptake of

    6 wt% at 100 bar and Fig. 12 shows how enhanced gravimetric capacity for an array

    Fig. 11 Hydrogen adsorption isotherm of COF-108 in wt% at 77 K (Reprinted from ref. 208,

    with permission from American Chemical Society).

    Annu. Rep. Prog. Chem., Sect. A, 2009, 105, 2154 | 43

    This journal is c The Royal Society of Chemistry 2009

    DownloadedbyUniversityofConnecticuton07

    October2011

    Publishedon13

    May2009onhttp://pubs.rsc.org|doi:10.1

    039/B818951J

  • 8/3/2019 Mandal_H2 Storage Materials

    25/35

    of Li-doped MOFs compares theoretically to the undoped materials.211 Hence

    despite a linear relationship with surface area in both doped and undoped materials,

    it is the lithium concentration that is believed to be the dominant factor for high

    uptake near room temperature.211 The introduction of Li-atoms within the

    framework has been proposed to functionalise the organic linkers211,212 and a

    gravimetric uptake of 4.5 wt% at room temperature and 10 wt% at 77 K at

    100 bar has been predicted in MOFs by following this Li-doping strategy.213

    Among the various MOFs reported, after independent authentication of 7.5 wt%

    hydrogen uptake at 77 K and 70 bar, MOF-177 is considered as an excellent material

    to serve as a benchmark adsorber.214 A very recent study by Schro der and

    co-workers showed promising uptake of 7.3 wt% in a Cu(II)-tetracarboxylate based

    framework (NOTT-103) at 77 K and 60 bar.

    215

    This drop in pressure can beconsidered significant in comparison to most of the previous studies conducted at

    pressures up to 100 bar.

    In summary, the small enthalpy change (heat of adsorption) in physically

    adsorbed MOFs offers a huge advantage for faster kinetics and reversibility over

    the chemisorbed materials. Heats of adsorption vary with the type of metal,

    framework topology, surface area and pore size and hence the extent of hydrogen

    adsorption in MOF materials can be influenced by manipulating these parameters.216

    Therefore, the exploration of new MOFs and developing an understanding of

    adsorption mechanisms are very important steps towards optimising the hydrogen

    storage properties of MOFs at near ambient conditions. Incorporation of moieties orfunctionalities that increase the interaction of the framework with hydrogen

    may increase the temperatures at which large quantities of hydrogen can be adsorbed.

    In addition, optimization of pore size without compromising the specific surface area

    and total pore volume will play a key role in developing viable materials.

    Fig. 12 Gravimetric uptake of hydrogen as a function of BET surface area for MOF and

    Li-doped MOF at 300 K and 100 bar (Reprinted from ref. 211, with permission from American

    Chemical Society).

    44 | Annu. Rep. Prog. Chem., Sect. A, 2009, 105 , 2154

    This journal is c The Royal Society of Chemistry 2009

    DownloadedbyUniversityofConnecticuton07

    October2011

    Publishedon13

    May2009onhttp://pubs.rsc.org|doi:10.1

    039/B818951J

  • 8/3/2019 Mandal_H2 Storage Materials

    26/35

    4. New materials for hydrogen storage

    In this section, novel emerging materials for hydrogen storage will be considered.

    This will mainly encompass nanosized and nanostructured materials spanning the

    classes of solids described above. Both experimental studies and promising theo-

    retical predictions of improved hydrogen storage will be highlighted. In parallel to

    common concepts in catalysis, it has been shown that decreasing the particle size to

    the nanoscale drastically improves the kinetics of hydrogenationdehydrogenationin metal hydrides and alloys. In this regard simple ball-milling can have a significant

    effect in altering the grain size. Moreover, ball-milling can be also employed to

    disperse catalysts into chemical hydrides thus generating nanosized catalysts.

    As described in section 2.1, bulk MgH2 offers a high storage capacity of 7.6 wt%

    but its use is limited by the poor reversibility originating from the high thermal

    stability of the hydride and slow desorption kinetics.7 Although considerable effort

    has been devoted to modify the kinetics and lower Tdes of the bulk hydride, these

    approaches have afforded only relatively small improvements in the sorption kinetics

    and exerted little influence over the desorption temperature (300 1C for bulk

    MgH2).7

    However, in a combined quantum chemical (HartreeFock) and DFT study,

    Wagemans et al. have shown that both Mg and MgH2 become less stable with

    decreasing cluster size (below 20 Mg atoms), thereby decreasing the hydrogen

    desorption energy.217 The study showed significant decreases in desorption energy

    below a crystal grain size ofB1.3 nm. Moreover, another interesting outcome was

    the fact that the destabilization of MgH2 was stronger than Mg (Fig. 13).217 Similar

    destabilisation effects have been predicted in the case of Mg and MgH2 nanowires as

    well by DFT calculations.218 The calculations reveal the possibility of a room

    temperature hydrogen storage in MgH2 nanowires of 0.85 nm diameter.218

    Recently, Li et al. have successfully synthesized nanowires of Mg and experimen-

    tally demonstrated the enhanced kinetics of hydrogen absorptiondesorption of

    nanowires as compared to bulk MgH2 and, moreover, the thinner the nanowire the

    lower the desorption energy (Fig. 14).219 Interestingly, despite the complete

    Fig. 13 Calculated energies for Mg and MgH2 clusters as a function of cluster size (Reprinted

    from ref. 217, with permission from American Chemical Society).

    Annu. Rep. Prog. Chem., Sect. A, 2009, 105, 2154 | 45

    This journal is c The Royal Society of Chemistry 2009

    DownloadedbyUniversityofConnecticuton07

    October2011

    Publishedon13

    May2009onhttp://pubs.rsc.org|doi:10.1

    039/B818951J

  • 8/3/2019 Mandal_H2 Storage Materials

    27/35

    disappearance of nanowires (and their collapse to nanoparticles) after only

    10 cycles, there was no change in capacity even after 50 hydridingdehydriding

    cycles.219 Also, Aguey-Zinsou and Ares-Ferna ndez have reported 7.6 wt% reversible

    storage near room temperature for Mg nanoparticles of 5 nm diameter.220 The size

    dependent performances even for a complex hydride, NaAlH4, have been reported

    recently by Balde et al.221 Hydrogen desorption temperatures of below 70 1C were

    achievable for 210 nm particles of NaAlH4 followed by reabsorption under lowhydrogen pressures (beginning at 20 bar).221 Furthermore, recent reports describe

    the hydrogen desorption of hollow nanospheres of Mg(NH2)2 mixed with 2 moles of

    MgH2 nanoparticles starting at 375 K under hydrogen with enhanced desorption

    kinetics compared to bulk materials.222 Recently, the prospects of metal/metal

    hydride nanostructures were reviewed by Be rube et al. considering both their

    advantages and disadvantages.223 Peng et al. have described some recent develop-

    ments in Mg-nanostructures.30 There are advantages of low temperatue hydrogen

    cycling and fast kinetics but poor cycleability, degradation of sorption properties

    and a reduction in thermal conductivity would seem to suggest practical challenges

    that would need to be overcome.223

    Therefore, the concept of hydride destabilization

    Fig. 14 Hydrogen absorption (top) and desorption (bottom) curves of Mg nanowires at

    different temperatures; triangle, 3050 nm; circle, 80100 nm and square, 150170 nm diameters

    (Reprinted from ref. 219, with permission from American Chemical Society).

    46 | Annu. Rep. Prog. Chem., Sect. A, 2009, 105 , 2154

    This journal is c The Royal Society of Chemistry 2009

    DownloadedbyUniversityofConnecticuton07

    October2011

    Publishedon13

    May2009onhttp://pubs.rsc.org|doi:10.1

    039/B818951J

  • 8/3/2019 Mandal_H2 Storage Materials

    28/35

    via nanowire, nanotube or nanoparticle formation presents some immense

    opportunities for further development. More research is now required towards the

    development of synthesis methods for nanostructured materials together with much

    needed insight into every aspect of structure. Further, it is also worth noting that

    other structural factors may be utilised to modify and optimise sorption perfor-

    mance. For example, different polymorphs of MgH2 (a-, b- and g-phase) appear to

    have intrinsically different desorption temperatures.

    7c,224

    5. Future directions and challenges for materials design

    Despite significant progress made in the development of solid state hydrogen storage

    materials, there remains immense challenges ahead before we enter into a genuine

    hydrogen energy era. Considering the two major storage types discussed, namely

    chemical and physical, there are advantages and disadvantages for mobile applica-

    tions in both categories. For example, many of the complex chemical hydrides have

    high storage capacity with favourable kinetics or thermodynamics (to some extent),but mostly lack reversibility. Conversely, MOFs have very good reversibility, but the

    higher storage capacity is realized only at low temperatures (77 K) which is far

    beyond the operational temperature of PEM fuel cells.

    Apart from those discussed above, there are many other classes of materials that

    are also being investigated for potential hydrogen stores. These include PIMs,225227

    carbon nanostructures,228,229 clathrates,230233 and zeolites.3b,234 Clathrates are

    compounds that are formed by association of two or more molecular entities and

    not by strong chemical bonds wherein one set of molecules encloses the other.

    Hydrogen is stored by a mode of encapsulation in a clathrate and can be released by

    changing the pressure or temperature.233 It is only recently that these materials have

    been suggested for hydrogen storage after the discovery of hydrogen clathrates.230233

    More future work is necessary with clathrates before their potential can be

    judged.

    PIMs are formed either by polymerisation of monomers with rigid units (having

    restricted conformation or rotation) or by cross-linking of polymeric gels.225,226

    They have been investigated for hydrogen storage due to their microporosity which

    gives rise to a large surface area and low density material. Storage capacities up to

    2.17 wt% at 77 K and 10 bar have been reported in PIMs.227 However, due to the

    weak nature of interaction with hydrogen, similar to that of MOFs, their ambienttemperature storage capacities are significantly low. Strategies similar to those

    adopted for MOFs can be applied to PIMs towards improving their performance.

    Carbon materials including carbon nanotubes have been intensively invesitgated

    for hydrogen storage applications.228 Yang et al. have recently reported a reversible

    hydrogen uptake of 6.9 wt% at 77 K and 20 bar in an ordered porous carbon

    material.229 This is very a encouraging result and requires further attention. Besides

    carbon, hydrogen storage properties of inorganic nanowires and nanotubes of boron

    nitride (BN), titanium disulfide (TiS2) and molybdenum disulfide (MoS2) have been

    investigated.235 It seems apparent from the very limited studies of such systems that

    those involving light elements are more attractive than heavier ones to maximise thegravimetric capacity. For example, in collapsed BN nanotubes hydrogen uptake up

    to 4.2 wt% was reported at a pressure of 10 MPa.236 These diverse examples of very

    different materials show that there is still plenty of room for further invention and

    discovery in the bid to produce new solid state hydrogen storage solutions.

    Annu. Rep. Prog. Chem., Sect. A, 2009, 105, 2154 | 47

    This journal is c The Royal Society of Chemistry 2009

    DownloadedbyUniversityofConnecticuton07

    October2011

    Publishedon13

    May2009onhttp://pubs.rsc.org|doi:10.1

    039/B818951J

  • 8/3/2019 Mandal_H2 Storage Materials

    29/35

    References

    1 J. K. Dowson, Nature, 1974, 249, 724.2 (a) M. S. Dresselhaus and I. L. Thomas, Nature, 2001, 414, 332; (b) L. Schlapbach and

    A. Zu ttel, Nature, 2001, 414, 353.3 (a) W. Grochala and P. P. Edwards, Chem. Rev., 2004, 104, 1238; (b) A. W. C. van den

    Berg and C. O. Area n, Chem. Commun., 2008, 668.4 (a) DOE. http://www1.eere.energy.gov/hydrogenandfuelcells/storage/current_technology.

    html, targets for on-board hydrogen storage systems. Also look for other related linksand updates in http://www1.eere.energy.gov/hydrogenandfuelcells; (b) S. Satyapal, J.Petrovic, C. Read, G. Thomas and G. Ordaz, Catal. Today, 2007, 120, 246.

    5 G. Sandrock, J. Alloys Compd., 1999, 295, 877.6 F. Schu th, B. Bogdanovic and M. Felderhoff, Chem. Commun., 2004, 2249.7 (a) B. Sakintuna, F. Lamari-Darkrim and M. Hirscher, Int. J. Hydrogen Energy, 2007, 32,

    1121; (b) M. Dornheim, S. Doppiu, G. Barkhordarian, U. Boesenberg, T. Klassen,O. Gutfleisch and R. Bormann, Scr. Mater., 2007, 56, 841; (c) D. Grant, in Solid StateHydrogen Storage-Materials and Chemistry, ed. G. Waker, Woodhead Publishing,Cambridge, 2008, pp. 357380.

    8 S. Orimo, Y. Nakamori, J. R. Eliseo, A. Zu ttel and C. M. Jensen, Chem. Rev., 2007, 107,4111.

    9 J. Graetz, Chem. Soc. Rev., 2009, 38, 73.

    10 P. Chen and M. Zhu, Mater. Today, 2008, 11, 36.11 P. Wang and X.-D. Kang, Dalton Trans., 2008, 5400.12 (a) D. H. Gregory, J. Mater. Chem., 2008, 18, 2321; (b) D. H. Gregory, Chem. Rec., 2008,

    8, 229; (c) D. H. Gregory, in Solid State Hydrogen Storage-Materials and Chemistry, ed.G. Waker, Woodhead Publishing, Cambridge, 2008, pp. 450477.

    13 H. W. Langmi and G. S. McGrady, Coord. Chem. Rev., 2007, 251, 925.14 K. C. Hoffman, J. J. Reilly, F. J. Salzano, C. H. Waide, R. H. Wiswall and

    W. E. Winsche, Int. J. Hydrogen Energy, 1976, 1, 133.15 (a) R. Schulz, J. Huot, G. Liang, S. Boily, G. Lalande, M. C. Denis and J. P. Dodelet,

    Mater. Sci. Eng. A, 1999, 267, 240; (b) A. Zaluska, L. Zaluski and J. O. Stro m-Olsen,J. Alloys Compd., 1999, 288, 217; (c) J. Huot, G. Liang, S. Boily, A. Van Neste andR. Schulz, J. Alloys Compd., 1999, 295, 495; (d) G. Liang, J. Huot, S. Boily and R. Schulz,

    J. Alloys Compd., 2000, 305, 239; (e) H. Reule, M. Hirscher, A. Weihardt and H.Kronmu ller, J. Alloys Compd., 2000, 305, 246; (f) J. Huot, M.-L. Tremblay and R. Schulz,J. Alloys Compd., 2003, 356357, 603; (g) J. Huot, G. Liang and R. Schulz, Appl. Phys. A,2001, 72, 187.

    16 T. Vegge, L. S. Hedegaard-Jensen, J. Bonde, T. R. Munter and J. K. Norskov, J. AlloysCompd., 2005, 386, 1.

    17 W. Oelerich, T. Klassen and R. Bormann, Adv. Eng. Mater., 2001, 3, 487.18 V. M. Skripnyuk, E. Rabkin, Y. Estrin and R. Lapovok, Acta Mater., 2004, 52, 405.19 N. Hanada, T. Ichikawa and H. Fujii, J. Phys. Chem. B, 2005, 109, 7188.20 N. Hanada, T. Ichikawa, S. Hino and