Thermal desorption of hydrogen from ammonia borane: Unexpected role of homopolar B-H...H-B interactions

Article abstract of DOI:10.1039/c2cc16111g

The solid-state structure of ammonia borane is held together by an intricate N-H...H-B proton-hydride bonding network. These intermolecular interactions have long been considered to mediate the release of hydrogen from this material. Here we reveal the silent but important role played by B-H...H-B interactions in the thermal decomposition of this leading hydrogen storage candidate. The Royal Society of Chemistry 2012.

Full text of DOI:10.1039/c2cc16111g

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
ChemComm
Cite this: Chem. Commun., 2012, 48, 2597–2599
www.rsc.org/chemcomm
COMMUNICATION
Thermal desorption of hydrogen from ammonia borane: unexpected role
of homopolar B–HÁ Á ÁH–B interactionsw
David J. Wolstenholme,* Kyle T. Traboulsee, Ye Hua, Larry A. Calhoun and
G. Sean McGrady*
Received 30th September 2011, Accepted 21st December 2011
DOI: 10.1039/c2cc16111g
The solid-state structure of ammonia borane is held together by
an intricate N–HÁ Á ÁH–B proton–hydride bonding network.
These intermolecular interactions have long been considered to
mediate the release of hydrogen from this material. Here we
reveal the silent but important role played by B–HÁ Á ÁH–B
interactions in the thermal decomposition of this leading hydrogen
storage candidate.
that homopolar B–HÁ Á ÁH–B (hydride-hydride) interactions
contribute to a significant extent in this process.7 Likewise,
Autrey et al. demonstrated that thermolysis of a mixture of
NH3BH3 and ND3BD3 results in the evolution of HD, H2, and
D2, providing further evidence of a bimolecular pathway.8 In
this instance, each of the hydrogen isotopomers can arise through
the intermediacy of homopolar dihydrogen interactions, as well
as through their conventional proton–hydride counterparts.
Our recent discovery of homopolar dihydrogen bonding in
3, in tandem with the report of Autrey et al., has prompted us
to revisit the mechanism of hydrogen desorption from 1. In
this endeavour we have exploited the selectively deuteriated
derivatives ND3BH3 (1a) and NH3BD3 (1b), drawing on
NMR spectroscopy to identify the provenance of each atom
in the hydrogen gas evolved, and we have complemented this
with a thermogravimetric analysis (TGA) study of the mass
loss from 1a and 1b. This combined approach has revealed the
unexpected and surprisingly important role played by
B–HÁ Á ÁH–B interactions in the formation of H2 by this leading
hydrogen storage candidate.
Ammonia borane (NH3BH3; 1) has received a considerable
amount of attention in recent years, owing to its attractive
properties as a hydrogen storage material.1 In the crystalline
state, the respective acidic and basic nature of the N–H and
B–H moieties in 1 results in an elaborate 3D network of
stabilising N–HÁÁÁH–B proton–hydride (dihydrogen) interactions.2
These interactions dominate the physical and chemical properties
of 1, and have long been considered to play a central thermo-
dynamic and mechanistic role in the evolution of hydrogen
from this molecular hydride.1 However, an in situ solid-state
11B NMR study of 1 by Autrey et al. pointed to a more
complex mechanistic pathway.3 This study indicated that
disruption of the proton–hydride bonding during the induction
period leads to the formation of a more mobile phase, which
subsequently converts to the isomeric diammoniate of
diborane ([NH3BH2NH3]+[BH4]À; 2); this ionic material is
then considered to react with remaining 1 to release hydrogen
in a bimolecular process. In this scenario, disruption of the
proton–hydride bonding in 1 is actually a prerequisite for
hydrogen release. Moreover, the significant strength of the
proton–hydride interactions in the solid-state adds to the
thermodynamic irreversibility of 1.4,5
The extended structure of 1 is stabilised primarily by N–HÁÁÁH–B
proton–hydride interactions (dHÁ Á ÁH = 2.02–2.22 A).2 The
conversion of 1 to its isomeric form 2 results in a shortening
and strengthening of these proton–hydride interactions
(dHÁ Á ÁH = 1.80–2.36 A), endowing additional stability to the
solid-state structure.9 In contrast, the amidoborane derivative
3
displays considerably weaker N–HÁ Á ÁH–B bonding
(dHÁ Á ÁH = 2.21 A),7,10 a difference that has been intimated
as the driving force behind the enhanced hydrogen release
properties associated with this material.11 In addition, the
structure of 3 displays a short B–HÁ Á ÁH–B contact of 2.11 A,
illustrating how this system is ideally disposed for the release
of hydrogen through homopolar dihydrogen bonding.7 In
contrast, the much longer analogous B–HÁ Á ÁH–B contacts in
1 (dHÁ Á ÁH = 3.05 A) and 2 (dHÁ Á ÁH = 2.80/2.99 A) preclude any
significant contribution from these interactions in the formation
of hydrogen.
Although hydrogen desorption from binary and ternary
metal hydrides such as AlH3 and LiBH4 is necessarily preceded
by a homopolar HÁ Á ÁH interaction, the structures of 1 and its
derivatives are predisposed to facilitate this process through
heteropolar N–HÁ Á ÁH–B interactions.6 However, in a recent
study of LiNH2BH3, 3, we used selective deuteriation to show
Inferring conclusions about the hydrogen release mechanism
of 1 from its solid-state structure can be misleading, as its
melting point (B104 1C) occurs immediately before the
onset of the first dehydrogenation step (B107–110 1C).1 The
concatenation of these two events emphasises that (i) melting
is required to enable the formation of significant quantities of
Department of Chemistry, University of New Brunswick,
P.O. Box 4400, Fredericton, N.B., Canada E3B 5A3.
E-mail: dwolsten@unb.ca and smcgrady@unb.ca
w Electronic supplementary information (ESI) available: Details of the
synthesis, NMR, and TGA experiments for ND3BH3 and NH3BD3.
See DOI: 10.1039/c2cc16111g
c
This journal is The Royal Society of Chemistry 2012
Chem. Commun., 2012, 48, 2597–2599 2597

View Article Online
hydrogen gas; and (ii) the proton–hydride bonding in the
solid-state structure of 1 actually stabilises the material against
decomposition.4,5 The enhanced mobility of 1 attendant on
melting permits its rapid isomerisation to 2, followed by its
subsequent release of hydrogen.3 Accordingly, we have carried
out a detailed investigation of the decomposition of 1a and 1b
through NMR spectroscopy and TGA measurements. In the
first series of experiments, the gaseous products evolved from
these isotopomers as they are raised through their melting
points were monitored by 1H and 2H NMR spectroscopy,
respectively. Hydrogen gas formed through the intermediacy
of N–HÁ Á ÁH–B or B–HÁ Á ÁH–B pathways will appear as HD or
H2 in the 1H experiments, and as HD or D2 in the corres-
ponding 2H spectra, assuming no hydrogen scrambling occurs
in this process.
experiments shows that the formation of hydrogen at temperatures
close to the melting point of 1 occurs predominantly via a pathway
involving N–HÁÁÁH–B proton–hydride interactions; i.e. those that
dominate the solid-state architecture of this system. However,
once the molecules are freed from this rigid arrangement the
contribution to hydrogen formation from the alternative
B–HÁ Á ÁH–B pathway increases dramatically, to the extent that
these interactions account for almost half the hydrogen
evolved by 120 1C. Although these results provide compelling
evidence for a hydride-hydride pathway, we cannot completely
discount the potential of hydrogen scrambling in these
reactions.12
In order to glean further support for this counterintuitive
pathway, we have characterised the hydrogen evolved from 1a
and 1b through TGA experiments. Fig. 2 presents the mass
loss from each of these isotopomers between 80 and 150 1C.
The isobaric nature of 1a and 1b (each with a molecular mass
of 33.8 g molÀ1), means that a hydrogen release pathway
involving purely N–HÁ Á ÁH–B proton–hydride interactions
would result in an equal loss of mass from both isotopomers,
as HD would be the sole product in each case. However, any
additional contribution from B–HÁ Á ÁH–B interactions will
lead to a greater loss of mass for 1b (D2 vs. H2). Such an
outcome is consistent with the thermal decomposition of the
related ethane 1,2-diamineborane (EDAB) complex, in which
more H2 was released than possible through a N–HÁ Á ÁH–B
pathway alone, hinting at the possibility of a hydride-hydride
pathway.13 The most notable feature of Fig. 2 is an inflection
point at B105 1C. Here the mass loss from 1b begins to exceed
that of 1a, coinciding closely with the melting point of 1. We
have selected two subsequent temperatures (120 and 140 1C) at
which to compare the mass-loss ratio of these two isotopomers.
The former temperature corresponds approximately to completion
of the initial stage in dehydrogenation, at which 1a and 1b
liberate B8.5 and 11.8 wt% hydrogen (1.38 ratio), respectively.
Raising the temperature to 140 1C results in the evolution of a
total of B10.9 and 15.8 wt% hydrogen (1.44 ratio). Using
these ratios, and assuming that zero-point energy differences
do not significantly perturb the relative involvement of the
N–HÁÁÁH–B and B–HÁÁÁH–B pathways for these two iso
topomers, we have estimated the individual contributions of HD,
H2, and D2 at each stage in the decomposition of these materials.14
Staged heating of samples of 1a and 1b in contact with
toluene (in which these materials are insoluble but can absorb
significant quantities of hydrogen gas), resulted in the appearance
of comparable amounts of HD and H2 for 1a or HD and D2 for
1b (Fig. 1). The onset of hydrogen release from both isotopomers
occurs at B100 1C, at which temperature these systems initially
liberate HD and H2(D2) in an approximate 2 : 1 and 4 : 3 ratio,
respectively, after accounting for the silent para-H2 in the
1H NMR experiment. Further heating for 20 min at 120 1C
resulted in an increase in the proportion of H2(D2), with the
ratios changing to B1 : 1 and 3 : 4, respectively. This series of
Fig. 1 (a) 1H and (b) 2H NMR spectra (in toluene) of the gaseous
products evolved from the thermal decomposition of 1a and 1b,
respectively. Spectra were collected sequentially after holding each
sample at the appropriate temperature for 20 mins.
Fig. 2 TGA plots for the thermal decomposition of 1a (blue) and
1b (red).
c
2598 Chem. Commun., 2012, 48, 2597–2599
This journal is The Royal Society of Chemistry 2012

View Article Online
This analysis reveals that immediately following the melting
of 1, the two pathways contribute comparable amounts of
hydrogen, in good agreement with the conclusions drawn from
the NMR experiments presented in Fig. 1. In addition, the
mass difference between these isotopomers at the end of the
first decomposition stage (B120 1C) is approximately one
atomic mass unit, a difference which is unlikely to occur
if these variations in the solid-state chemistry were attributed
to hydrogen scrambling.
but also by counterintuitive B–HÁ Á ÁH–B interactions, which are
implicated to a comparable extent in the formation of hydrogen
gas. The long B–HÁ Á ÁH–B distances in the crystal structures of
1 and 2 militate against the involvement of these homopolar
interactions in the solid state: it is only upon melting of 1 that
the B–H moieties can approach each other to facilitate such a
pathway. This study not only sheds new light on the hydrogen
release mechanism of this important hydrogen storage material;
it also represents the first evidence of such B–HÁ Á ÁH–B inter-
actions occurring outside the solid state. These novel and
intriguing homopolar interactions appear to play a silent but
central role in this process, and we expect that they will
prove to have wider significance in the fundamental and
applied chemistry of a broad range of molecular materials.
We are currently exploring in detail the nature and extent of
hydrogen scrambling in the polymeric [NH2BH2]n and [NHBH]n
decomposition products that result from the desorption of
hydrogen from 1.
The substantial participation of a B–HÁ Á ÁH–B pathway is a
counterintuitive and remarkable scenario, given the homopolar
nature of the two B–H moieties involved, and the absence of
any such interactions in the crystal structures of 1 and 2. It is
noteworthy that a recent solution 11B NMR study of the
formation of 1 and 2 through reaction of NH3 and THFÁBH3
proposed the involvement of an ammonia diborane inter-
mediate NH3BH2(m–H)BH3; 4.15 The bridging hydride in this
putative intermediate is likely to be significantly more acidic
than the terminal B–H moieties, so that a B–HÁ Á ÁH–B inter-
action involving 4 may actually be proton–hydride in nature.
However, the solution decomposition of 1 is known to give rise
to different products (i.e. cyclic oligomers) then the solid-state
pathway.16 Moreover, a solid-state 11B NMR study of this
process has produced no evidence of 4 as an intermediate, and
we conclude that it is unlikely to be a participant in the
pathway adopted in the solid-state.3
Notes and references
1 (a) F. H. Stephens, V. Pons and R. T. Baker, Dalton Trans., 2007,
2613–2626; (b) C. W. Hamilton, R. T. Baker, A. Staubitz and
I. Manners, Chem. Soc. Rev., 2009, 38, 279–293; (c) A. Staubitz,
A. P. M. Robertson and I. Manners, Chem. Rev., 2010, 110,
4079–4124; (d) M. Bowden and T. Autrey, Curr. Opin. Solid State
Mater. Sci., 2011, 15, 73–79.
2 (a) E. W. Hughes, J. Am. Chem. Soc., 1956, 78, 502–503;
(b) W. T. Klooster, T. F. Koetzler, P. E. M. Siegbahn,
T. B. Richardson and R. H. Crabtree, J. Am. Chem. Soc., 1999,
121, 6337–6343; (c) M. E. Bowden, G. J. Gainsford and
W. T. Robinson, Aust. J. Chem., 2007, 60, 149–153.
3 A. C. Stowe, W. J. Shaw, J. C. Linehan, B. Schmid and T. Autrey,
Phys. Chem. Chem. Phys., 2007, 9, 1831–1836.
4 T. Autrey, M. Bowden and A. Karkamkar, Faraday Discuss., 2011,
151, 157–169.
5 W. I. F. David, Faraday Discuss., 2011, 151, 399–414.
6 S. Orimo, Y. Nakamori, J. R. Eliseo, A. Zuttel and C. M. Jensen,
¨
Chem. Rev., 2007, 107, 4111–4131.
7 D. J. Wolstenholme, J. T. Titah, F. N. Che, K. T. Traboulsee,
J. Flogeras and G. S. McGrady, J. Am. Chem. Soc., 2011, 133,
16598–16604.
8 R. S. Smith, B. D. Kay, B. Schmid, L. Li, N. Hess, M. Gutowski
and T. Autrey, Prep. Pap. Am. Chem. Soc, Div. Fuel Chem., 2005,
50, 112–113.
9 M. Bowden, D. J. Heldebrandt, A. Karkamkar, T. Proffen,
G. K. Schenter and T. Autrey, Chem. Commun., 2010, 46,
8564–8566.
10 Z. Xiong, C. K. Yong, G. Wu, P. Chen, W. Shaw, A. Karkamkar,
T. Autrey, M. O. Jones, S. R. Johnson, P. P. Edwards and
W. I. F. David, Nat. Mater., 2008, 7, 138–141.
11 Y. S. Chua, P. Chen, G. Wu and Z. Xiong, Chem. Commun., 2011,
47, 5116–5129.
12 Raman spectra of 1a and 1b display no signs of hydrogen scrambling
prior to decomposition; T. Autrey, private communication.
13 D. Neiner, A. Karkamkar, M. Bowden, Y. J. Choi, A. Luedtke,
J. Holladay, A. Fisher, N. Szymczak and T. Autrey, Energy
Environ. Sci., 2011, 4, 4187–4193.
14 The contribution from HD, H2, and D2 in the thermal decomposition
of 1a and 1b was determined by solving the simultaneous equations:
x MH2 + y MHD = 1.0, x MD2 + y MHD = ND3BH3/NH3BD3 %wt
ratio - 120 1C: y = 0.21, x = 0.19; 140 1C: y = 0.19, x = 0.22.
15 X. Chen, X. Bao, J.-C. Zhao and S. G. Shore, J. Am. Chem. Soc.,
2011, 133, 14172–14175.
The NMR and TGA results presented in Fig. 1 and 2 shed
important new light on the hydrogen release process of 1. It is
well established that this material isomerises to 2 at temperatures
close to its melting point, at which stage hydrogen release is
accelerated.3 This process has been previously considered to
occur exclusively through the intermediacy of N–HÁ Á ÁH–B
proton–hydride interactions,1 with the sequential formation of
polyamidoborane [NH2BH2]n and polyimidoborane [NHBH]n
products. However, the results presented here reveal a
significant additional B–HÁ Á ÁH–B pathway, which plays an
important role in this dehydrogenation process. The origin of
the hydrogen released from the melt containing 1 and 2 can be
traced back to nearly equal contributions from N–HÁ Á ÁH–B
and B–HÁ Á ÁH–B interactions. Autrey et al. proposed that the
initial loss of hydrogen occurs through direct reaction between
the N–H moieties of the cation in 2 and the B–H bonds of 1.3
We concur with this initial step, but propose that the majority
of subsequent hydrogen formation arises through B–HÁ Á ÁH–B
interactions in the molten phase, accompanied presumably by
hydrogen transfer between neighboring amine and borane moieties
that preserve the integrity of the incompletely characterised
[NH2BH2]n polyamidoborane product. Whatever the process
that delivers this end product, it is clear that the majority of
the hydrogen atoms in the H2 desorbed from 1 originate from
B–H, rather than from N–H, moieties; the availability of a
B–HÁ Á ÁH–B pathway is responsible for this surfeit.
In summary, the results presented here show that hydrogen
evolution from 1 is significantly more complex than has
hitherto been appreciated. The process is facilitated not
only by ubiquitous N–HÁ Á ÁH–B proton–hydride interactions,
16 J. S. Wang and R. A. Geanangel, Inorg. Chim. Acta, 1988, 148,
185–190.
c
This journal is The Royal Society of Chemistry 2012
Chem. Commun., 2012, 48, 2597–2599 2599