Volume 4 fuel cells and hydrogen technology 4 06 – hydrogen storage liquid and chemical

Volume 4 fuel cells and hydrogen technology 4 06 – hydrogen storage liquid and chemical Volume 4 fuel cells and hydrogen technology 4 06 – hydrogen storage liquid and chemical Volume 4 fuel cells and hydrogen technology 4 06 – hydrogen storage liquid and chemical Volume 4 fuel cells and hydrogen technology 4 06 – hydrogen storage liquid and chemical

P Chen, Dalian Institute of Chemical Physics, Dalian, China

© 2012 Elsevier Ltd All rights reserved

Hydrogen can be stored by either physical or chemical means For physical storage, the conventional options are compressed hydrogen gas and cryogenic adsorption, that is, liquid hydrogen For chemical storage, the hydrogen molecule can be dissociated either homolytically or heterolytically and bonds with other elements to form hydrides The hydrogen uptake and release can be either reversible or irreversible depending on the thermodynamic parameters of the corresponding starting and final materials During the last decade of exploration and study, the scope of candidate hydrogen storage materials has expanded considerably, from conventional metal hydrides, such as LaNi5 and MgH2, to complex and chemical hydrides [3–7] and from activated carbon to carbon nanotubes and

to metal organic frameworks (MOFs) [8–18] The employment of advanced synthetic routes has also allowed the physical state of the storage materials to change from being bulk crystalline to amorphous and to nano structures [16, 19, 20] Advanced t heoretical simulations also have an increasing impact not only on the description of the physical properties of known materials but also on the prediction of novel structures and reaction paths [21–23] A variety of promising storage systems are under intensive investigations Systems of high hydrogen content are the focus of ongoing studies because they allow more space for material modification and optimization A comprehensive survey by Thomas, Sandrock, and Bowman, as shown in Figure 1, lists over 40 material candidates that are being actively investigated Among them, high surface area porous materials and nitrogen- and boron-containing hydrides are the most studied systems [24, 25] In this chapter, a short survey on existing hydrogen storage techniques will be presented Emphasis will be given to chemical and complex hydrides that have been under intensive research since 2005

4.06.2 Physical Hydrogen Storage

Physical storage of hydrogen is normally achieved either under high pressure or at cryogenic temperatures To store high-pressure hydrogen, a compressed hydrogen gas tank made of aluminum or composite materials wrapped with carbon or glass fiber to ensure light weight and high strength is used The hydrogen energy stored in the compressed gas tank increases with an increase in pressure, but not in a directly proportional manner Compressed hydrogen at 350–700 bar has been well developed and adopted in prototype fuel cell vehicles The compression of H2 will consume ∼10–15% of the energy stored, and the size of the tank holding

∼4 kg of H2 is still too large to directly compare with a gasoline tank There have been considerable efforts in the development of lightweight tanks to store hydrogen up to 1000 bar in recent years Even with considerable progress, the tanks are somewhat too expensive (∼$15 kWh−1) to be practically viable [26] A recent demonstration on storing compressed hydrogen at 77 K suggested an alternative method with improved characteristics especially in gravimetric and volumetric storage densities It is, however, of practical importance to take a step forward in reducing the system cost and increasing the energy efficiency (Table 2)

Comprehensive Renewable Energy, Volume 4 doi:10.1016/B978-0-08-087872-0.00414-5 157

Ultimate AB/cat

Li−AB

CH Regen

Required

sorbents IRMOF-177

PCN-12

C aerogel carbide-derived C

M-doped CA PANI

PANI

Na2Zr(BH4)6 NaMn(BH4)4 LiNH2/MgH2

Li3AIH6/Mg(NH2)2 LiMgN LiBH4/MgH2

Li3AIH6/LiNH2

Mg(BH4)(AIH4) Mg−Li−B−N−H Ca(BH4)2/2LiBH4 1,6 naphthyridine M−B−N−H

PANI bridged cat./AX21

B/C MOF-74

New DOE system targets

2015

Table 1 The vehicle performance targets

Cycle life variation % of mean at % confidence 90/90 99/90 99/90 Min H2 delivery pressure atm 4 FC/35 ICE 4 FC/35 ICE 2 FC/35 ICE

Charging/discharging rates

Environmental health and safety

Permeation and leakage scc h−1 Meets or exceeds applicable standards

Table 2 Parameters of compressed and liquid H2 storage

Gravimetric Volumetric Storage

techniques

Storing energy (kJ kg−1)

Spent energy/

stored energy

energy content (MJ kg−1)

energy content (MJ m−3) Compressed

H2 (350 bar) Compressed

H2 (700 bar) Liquid H2

12 264

14 883

42 600

0.10 0.12 0.36

8.04 7.20 16.81

2492

3599

3999 With reference to http://www.hydrogen.energy.gov/annual_review11_proceedings.html

Liquid hydrogen is another option to store hydrogen onboard Comparatively, the energy density is nearly 2 times higher than that of the 700 bar compressed H2; however, the energy cost to liquefy hydrogen reaches 30% or more of the actual hydrogen energy stored There is also continual hydrogen loss when stored onboard (namely boil-off) due to thermal conduction, convection, and thermal radiation One of the other drawbacks is the use of expensive multilayered vacuum superinsulated vessels Another important practical issue is the ‘cooling-down’ losses during refilling of liquid hydrogen at gas stations The entire transfer line between the liquid H2 source and the vehicle tank system has to be cooled down to about −253.8 °C, and therefore, additional H2 evaporation occurs Clearly, these losses cannot be neglected and remain significant There have been a few comprehensive reviews

on compressed and liquid hydrogen published recently [26, 27], and the commercial employment of these techniques is still an open issue However, currently, while most of the chemical storage systems to be discussed later remain at a research stage, compressed and liquid H2 storage remain as choices for demonstration and evaluation purposes

Porous materials are prone to adsorb hydrogen physically; however, due to the weak interaction of H2 and sorbents, cryogenic conditions typically have to be applied Materials with large surface areas and proper pore size of 2–3 nm are capable of adsorbing up to

7 wt.% of H2 [12–14, 28, 29] Representative sorbent materials include carbon materials [8–11, 30–32], MOFs [12, 16, 17, 33–35], and conjugated polymers [14] A special note on hydrogen storage on MOFs was triggered by breakthroughs in material design and synthesis [12] The interesting bonding nature of the metal and organic link creates a variety of pore structures and active centers, and the nature of the interaction of H2 and the active centers has been a hot topic of study It is obviously of scientific interest to further the research in this area; however, cryogenic adsorption in general is an energy-consuming process and has, in general, relatively low volumetric hydrogen storage density; hence, the remainder of this chapter will focus on alternative solid-state forms of H2 storage

4.06.3 Metal Hydrides

The homolytic dissociation of H2 into atomic H followed by diffusion of the H in the lattice of metals, especially transition metals and alloys, leads to the formation of metal hydrides [2, 36] Table 3 shows a list of conventional metal hydrides that have been extensively studied in the past For example, the H content in terms of the volumetric hydrogen density of LaNi5H6 is 115 kg m−3, which is higher than that of compressed hydrogen and liquid hydrogen However, transition metals and their alloys have relatively low gravimetric hydrogen densities (normally < 3 wt.%) A recent work by Matsunaga et al [37] on hybridizing metal hydrides with a high-pressure tank gives a certain level of promise for the improvement of gravimetric density In addition to the search for new multicomponent metal alloys, current research also focuses on the reduction in size of Mg-based materials to enhance kinetics in dehydrogenation [16, 20]

4.06.4 Chemical Hydrides

There are a vast variety of natural and manmade chemical hydrides including H2O, NH3, alcohol, boranes, and hydrocarbons The

H–X (where X refers to O, N, B, C, etc.) bond is significantly stronger than that of a typical H–M bond of most of the interstitial metal hydrides [1–3] Chemical hydrides that have been investigated for the purpose of hydrogen storage are mainly those

Table 3 Structure and hydrogen storage properties of typical metal hydrides

hydrides Metal Hydrides Structure hydrogen Peq, T

AB5 LaNi5 LaNi5H6 Hexagonal 1.4 2 bar, 298 K

AB3 CaNi3 CaNi3H4.4 Hexagonal 1.8 0.5 bar, 298 K

AB2 ZrV2 ZrV2H5.5 Hexagonal 3.0 10−8 bar, 323 K

A2B Mg2Ni Mg2NiH4 Cubic 3.6 1 bar, 555 K

Figure 2 Molecular structures of AB and lithium amidoborane

LiNH2BH3NH3BH3 → LiN2B2H + 5H2 Solid 50–250 14.7

Ca(NH2BH3)2·2NH3 → Ca(BN2H)2 + 6H2 Solid 75–300 8.9

Mg(NH2BH3)2·NH3 → MgB2N3H + 6H2 Solid 50–300 11.8

a For solid-state dehydrogenation, most of the materials are under molten or semi-molten state

containing H–B, H–N, and H–O bonds [4, 38, 39] A distinctive feature of the most investigated chemical hydrides is the coexistence

of both protonic and hydridic H atoms Representative entities are NH3BH3 (ammonia borane, AB) [7, 24, 40–45], metal amidoboranes (MAB) [46–52], and H2O–borohydride systems Figure 2 shows the molecular structures of AB and lithium amidoborane H bonded with N has positive charge, which is opposite to H bonded with B When a crystal is formed, the shortest distance between these two H atoms is found to be less than twice the van der Waals radius of H (see Figure 3) The exceptionally high chemical potentials for the combination of H− and H+ to molecular H2 and the formation of strong B–N or B–O bond are likely

to be the driving forces for the dehydrogenation For most of the chemical hydrides, dehydrogenation is an exothermic process The increase in stability of the dehydrogenated product is the result of the formation of strong B–N or B–O covalent bond In this section, the research activities on the development of AB and amidoboranes for hydrogen storage will be reviewed Table 4 presents the conditions applied in the dehydrogenation of AB, alkali and alkaline earth amidoboranes, and their derivatives

4.06.4.1 Ammonia Borane

AB, first synthesized in 1955 [53], is a plastic crystalline solid adopting a tetragonal crystal structure with space group I�4 mm and lattice parameters of a = b = 5.240 Å and c = 5.028 Å at room temperature [54, 55] As shown in Figure 2, this molecular crystal is

stabilized by dihydrogen bonding between H(B) and H(N) The crystal melts at ∼100 °C and decomposes to hydrogen (1 equiv.) between 70 and 112 °C to yield polyaminoborane (PAB, [NH2BH2]n) according to eqn [1] Subsequently, [NH2BH2]n decomposes with an additional 1 equiv hydrogen loss, over a broad temperature range around 150 °C, forming amorphous polyiminoborane (PIB, [NHBH]n) and a small fraction of borazine ([N3B3H6]), according to eqns [2] and [3], respectively The decomposition of [NHBH] to boron nitride (BN) occurs at temperatures in excess of 500 °C This final step is not considered practical for hydrogen storage due to high temperatures needed for hydrogen release Thermodynamic analyses and theoretical calculations show that hydrogen release from either AB or PAB or PIB is an exothermic process, revealing the irreversibility of hydrogen desorption from these materials

nNH3BH3ðsÞ →½NH2BH2 nðsÞ þ nH 2ðgÞ ½1

½NH2BH2 nðsÞ →½N3B3H6 n = 3ðlÞ þ nH2ðgÞ ½3 Figure 4 presents some of the likely forms of products from releasing the first and second equivalent molecules of H2 from AB The structure and composition of the product vary with the conditions applied during dehydrogenation As an example, on catalytic dehydrogenation of AB by iridium (Ir) catalyst in tetrahydrofuran (THF), crystalline PAB is formed [43] A recent report from He

et al [68] demonstrated the formation of crystalline linear PAB in the FeB-catalyzed solid-state dehydrogenation of AB at ∼60 °C However, in most cases, the solid product is essentially amorphous and is a mixture of linear, cyclic, branched, and cross-linked B–N structures

Although it has an exceptionally high hydrogen content, AB has to overcome a few drawbacks to be practically viable The first two challenges are the mass production of AB and energy-efficient regeneration of the used fuel [56, 57], while the kinetic-borne dehydrogena­tion of AB and the coproduction of unwanted gaseous products (such as borazine and NH3) are to be alleviated Moreover, severe material foaming in the dehydrogenation is also problematic Tremendous efforts have been devoted to these issues since the first report on the dehydrogenation of AB for hydrogen storage by Wolf et al [40], among which investigations on catalytic modification of AB or dispersing

AB into porous materials attract significant attention [43–45, 58–71] Dehydrogenation of AB in ionic liquids shows improved kinetics in comparison with neat AB [42] Moreover, a number of intermediates and products have been identified by in situ nuclear magnetic resonance (NMR) and the density functional theory–gauge including/invariant atomic orbital (DFT–GIAO) calculations As shown in Figure 5, different states of hybridization (sp2

or sp3) and bonding environments (with H or N) have chemical shifts ranging from −26.9 to +39.3 ppm There are a few comprehensive reviews in this area to which the readers may like to further refer [24, 25, 56, 57]

Figure 4 Molecular structures of possible dehydrogenation products [56]

H

H HH

Figure 5 DFT–GIAO calculated 11

B NMR chemical shift for possible structures arising from the dehydropolymerization of AB [42]

4.06.4.1.1 Homogeneous catalytic dehydrogenation of AB

As summarized by Hamilton et al [57] and Smythe and Gordon [56], a few homogeneous catalysts including Ru-, Ir-, and Ni-based complexes and Lewis acid (B(C6F5)3) have been developed, which are effective in removing 1–2.5 equiv H2 from AB under mild conditions Figure 6 shows the time dependence of H2 evolution from a AB/THF solution with different concentrations of (POCOPf)Ir(H)2 With 1 mol.% of catalyst, ∼1 equiv H2 can be released within 5 min A quantitative yield of crystalline PAB was observed [43]

The dehydrogenation of AB by the Ni-N heterocyclic carbene (Ni-NHC) complex in a molar ratio of 10:1 shows unprecedented evolution of 2.5 equiv H2 at 60 °C [45] The formation of Ni-NHC is by the reaction of biscyclooctadiene nickel (Ni(cod)2) with Enders’ NHC Theoretical investigation suggests that the formation of the first transition state (highest energy barrier) is through transferring H(N) of AB to ligand carbene, which is different from the β-H elimination of AB in some other cases

One way to activate AB by Lewis or Brønsted acid is by attracting a hydridic H(B) from AB to form the initiative cation [H2BNH3]+ [63] As shown in Figure 7, the overall process after the formation of the initiative cation resembles cationic polymerization and dehydrogeneration One equivalent H2 can be removed from AB at ambient temperature

H BH2 NH3

+

‡ + HN2

Figure 7 Reaction of NH3BH3 with Lewis or Brønsted acid (A) results in the formation of borenium cation 2 Subsequent reaction with another equivalent

of NH3BH3 results in the formation of 3 with subsequent expulsion of H2 and concomitant formation of 4 [63]

As shown above, the chemistry involved in the catalytic dehydrogenation of AB in solvent is considerably rich and worthy of detailed experimental and theoretical investigations The use of the solvent allows sufficient mobility of both reactant and catalyst but will decrease the energy density of the system It is a subject of system engineering to minimize the side effect of the solvent but retain the efficiency of homogeneous catalysis

4.06.4.1.2 Solid-state dehydrogenation of AB

As mentioned in Section 4.06.4.1, the thermal decomposition of solid-state AB is a stepwise process having considerable kinetic barriers at each step Efforts in alleviating the barrier in solid-state dehydrogenation are through dispersing AB into porous substrates [7] and introducing a catalyst in the material (solid form) [68]

An introductory work by Gutowska et al demonstrated that, when dispersing AB onto porous SBA-15 nanoscaffold, hydrogen started to release at temperatures just above 50 °C and peaked at ∼100 °C, which is considerably lower than that of neat AB Moreover, the formation of borazine was largely depressed [7] Further isothermal testing showed that the dispersed AB presented a significantly shortened induction period and reduced kinetic barrier As shown in Figure 8, 1 equiv H2 can be released at 50 °C within 150 min However, for pristine AB, it has to go through a ∼100-min induction period and another 400–500 min to remove the same amount of H2 A few successful attempts in using carbon cryogel, lithium catalysis and mesoporous carbon (Li-CMK-3), nano-BN, poly(methyl acrylate) (PMA), and so on, to improve the dehydrogenation properties of solid AB have been reported in the past 5 years [70, 72, 73] Another approach in improving dehydrogenation of AB is via solid-state catalysis through the use of transition metals or alloys

He et al reported that, upon the introduction of nano-sized Co- or Ni- or Fe-based catalyst to solid AB via the so-called coprecipitation method, ∼1.0 equiv or 6 wt.% of H2 can be released at 59 °C (shown in Figure 9) [68] It was observed that the presence of the nano-sized catalyst largely depressed the sample foaming and the coproduction of borazine In the meantime, crystalline rather than amorphous PAB was formed (Figure 10), which should be derived from the catalyst-oriented growth of aminoborane and is significantly different from the ion-initiated dehydrogenation

Figure 8 Scaled exotherms (solid lines) from isothermal differential scanning calorimetry (DSC) experiments that show the time-dependent release of H2 from AB and AB:SBA-15 (1:1 w/w) The area under the curve for neat AB corresponds to ΔHrxn = −21 kJ mol−1, and the area under the curve for AB:SBA-15 corresponds to ΔHrxn = −1 kJ mol−1 The release of hydrogen from AB proceeds at a more rapid rate and at lower temperatures in SBA-15 The dashed line (-) is the integrated signal intensity; (•) is the point at which the reaction is 50% complete [7]

Intensity (a.u.) PABFeB PAB

d (Å)

0.0 0.2 0.4 0.6 0.8 1.0

Figure 10 XRD patterns of the postdehydrogenated neat AB (80 °C) and 2.0 mol.% Fe-doped AB samples (60 °C) ▼, crystalline linear PAB

4.06.4.2 Amidoboranes and Derivatives

4.06.4.2.1 Alkali and alkaline earth amidoboranes

As shown in the previous section, various methods have been employed to lower the decomposition temperature of AB through the use of additives and catalysts A different approach has been applied recently in the manipulation of the thermodynamic properties of compounds through chemical alteration to AB, that is, through substituting one H in the NH3 group in BH3NH3 with a more electron-donating element [46, 49] The rationale behind this approach is to alter the polarity and intermolecular interactions (specifically the dihydrogen bonding) of AB to produce a substantially improved dehydrogenation profile Lithium amidoborane (LiNH2BH3) [47, 49, 50, 52, 74, 75], sodium amidoborane (NaNH2BH3) [48, 49, 76], calcium amidoborane [46, 50], and strontium amidoborane [77] have been synthesized, which show substantially different dehydrogenation characteristics with respect to AB itself These alkali and alkaline earth amidoboranes (MABs) were synthesized mainly through the interactions of alkali or alkaline earth metal hydrides (LiH, NaH, CaH2, and SrH2) with AB (see eqn [4]), which lead to the replacement of hydrogen atom of AB by alkali or alkaline earth metals

where x = 1 when M is an alkali metal and x = 2 when M is an alkaline earth metal

The replacement of the H of the NH3 group in AB by alkali or alkaline earth element results in the alteration of the crystal structure and dehydrogenation property As shown in Figure 11, LiNH2BH3 crystallizes in the orthorhombic space group Pbca with the lattice constants a = 7.112 74(6) Å, b = 13.948 77(14) Å, c = 5.150 18(6) Å, and V = 510.970(15) Å3 The Li–N bond is essentially ionic and the B–H bond length is slightly longer than that in neat NH3BH3 NaNH2BH3 is of identical structure to LiNH2BH3 [49, 50] Ca(NH2BH3)2, on the other hand, is a monoclinic structure with a = 9.100(2) Å, b = 4.371(1) Å, c = 6.441(2) Å, and

β = 93.19°(see Figure 12) [50] Li or Na is coordinated with four NH2BH3 groups Ca, on the other hand, sits in the center of an octahedron made of NH2BH3 groups The THF adduct of Ca(NH2BH3)2 was also determined [46] It is interesting to note that, unlike Ca and Sr, no report on the formation of Mg(NH2BH3)2 has appeared in the literature so far

The experimental results show that more than 10 and 7 wt.% of hydrogen desorbs exothermically from LiNH2BH3 and NaNH2BH3, respectively, at around 91 °C (Figure 13) [49] Ca(NH2BH3)2, on the other hand, releases ∼8 wt.% H2 in the temperature range of 100–300 °C In all the cases, borazine production is beyond the detection limit of mass spectrometry (MS) The induction period that is associated with the dehydrogenation of pristine NH3BH3 is absent from these amidoboranes, indicating that a different dehydrogenation mechanism is occurring There are a few theoretical and experimental investigations on the dehydrogenation mechanism of these amidoboranes, especially of LiNH2BH3 [52, 75] Kim et al reported that the dehydrogenation of LiNH2BH3 is via abstracting H from the BH3 group by Li [52] An isotopic investigation also evidenced the bimolecular dehydrogenation

Figure 11 Crystal structure of LiNH2BH3 Li, B, N, and H are represented by red, orange, green, and white spheres, respectively [49]

H2 B1

H11 H9

Figure 14 Molecular packing and network of N–H⋯H–B dihydrogen bonding in MgAB·NH3 (a) and close contacts around the Mg2+ center (b)

Figure 15 Temperature-programmed desorption (TPD)-MS spectra (a) and volumetric release (b) measurements on Mg(NH2BH3)2·NH3 at the heating rate of 2 (a) and 0.5 °C min−1 (b)

with nearby H(B) The shortest (N)H⋯H(B) distance in Mg(NH2BH3)2·NH3, for example, is around 1.92 Å (Figure 14) The difficulty in forming Mg(NH2BH3)2 and the existence of its ammoniate indicate that the unstable crystal of Mg(NH2BH3)2 (probably due to small but dense charged cation (Mg2+) and big anion) can be stabilized by NH3 through the establishment of coordination of a lone pair of N to Mg2+ The thermal decomposition of Mg(NH2BH3)2·NH3 performed in a closed vessel demonstrated a stoichiometric conversion of NH3 and desorption of ∼11.2 wt.% H2 (shown in Figure 15)