Efficient regeneration of partially spent ammonia borane fuel

Full text of DOI:10.1002/anie.200900680

Communications
DOI: 10.1002/anie.200900680
Chemical Hydrogen Storage
Efficient Regeneration of Partially Spent Ammonia
Borane Fuel**
Benjamin L. Davis, David A. Dixon,* Edward B. Garner, John C. Gordon,*
Myrna H. Matus, Brian Scott, and Frances H. Stephens
Angewandte
Chemie
6812
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 6812 –6816

Angewandte
Chemie
A necessary target in realizing a hydrogen economy is the
storage of H2 for controlled delivery, presumably to an
energy-producing fuel cell.[1] In this vein, the U.S. Department
of Energyꢀs (DOE) Centers of Excellence (CoE) in Hydrogen
Storage have pursued different methodologies, including
metal hydrides, chemical hydrides, and sorbents, for the
expressed purpose of supplanting gasolineꢀs current driving
range of over 300 miles (ca. 500 km). Chemical hydrogen
storage has been dominated by one appealing material,
Scheme 1. Representative structure of polyborazylene (PB).
ꢀ
ammonia borane (H3B NH3, AB), owing to its high gravi-
metric capacity of hydrogen (19.6 wt%) and low molecular
weight (30.7 gmolꢀ1). AB has both hydridic and protic
moieties, yielding a material from which H2 can be readily
released.[2] As such, a number of publications have described
H2 release from amine boranes, yielding various rates
depending on the method applied.[3–6] The viability of any
storage system is critically dependent on efficient recycla-
bility, but reports on the latter subject are sparse.[1,7–10] For
example, the DOE recently decided to no longer pursue the
use of NaBH4 as a H2 storage material, in part because of
inefficient regeneration. We thus endeavored to find an
energy-efficient regeneration process for the spent fuel from
H2-depleted AB.
of DFT calculations for the gas phase, coupled with exper-
imental data or estimates of the heats of vaporization,
benzenedithiol was predicted to be a better reagent than
thiophenol for the reaction with borazine, a computational
surrogate for PB, where the products are presumed to retain
ꢀ
the B H bond (Table 1, see the Supporting Information for
details).
Table 1: Estimates of reaction energies for digestion.
Reaction
DH (298 K)[a]
Although spent fuel composition depends on the dehy-
drogenation method,[3,5] we have focused our efforts on the
spent fuel resulting from metal-based catalysis, which has to
date shown the most promise to meet the DOE H2 storage
requirements for release rate and extent.[11] Although the first
transition-metal-catalyzed dehydrogenation of AB generated
many products,[12] more recent metal catalysts have produced
single products, the fastest rates for a single equivalent of H2
released from AB,[13] and the greatest extent of H2 release (up
to 2.5 equiv of H2 can be produced within 2 h).[5] While
ongoing work is being carried out to tailor the composition of
spent AB fuel, we have developed a method for regenerating
the predominant product, polyborazylene (PB, Scheme 1),
resulting from dehydrogenation by nickel carbene catalysts.
Our approach utilizes reagents which avoid the formation
42.2/25.1
ꢀ20.4/0.5
[a] Condensed-phase/gas-phase values in kcalmolꢀ1
.
When benzenedithiol and PB were heated at reflux in
THF, 90% of the PB had reacted after 12 h, as judged by the
11B NMR spectrum, which showed two new resonances. The
upfield resonance (d = ꢀ5.6, d, 1JB-H = 128 Hz) was identified
ꢀ
as (C6H4S2)B H·(NH3) (1) by independent synthesis as well
as by comparison to the chemical shift calculated by DFT (see
the Supporting Information). The downfield resonance (d =
10.5 ppm, s) exhibits a similar chemical shift to Li[B-
(C6H4S2)2],[14] suggesting that [NH4][B(C6H4S2)2] is formed
(this assignment is consistent with the calculated NMR
spectrum, see the Supporting Information). Attempts to
ꢀ
of thermodynamically stable B O bonds and the subsequent
need for high-energy reducing agents. Thiols were attractive,
ꢀ
ꢀ
as B S bonds are weaker than analogous B O bonds and the
acidity of the SH moiety could aid the reaction. On the basis
ꢀ
make this product independently from (C6H4S2)B H·(NH3)
[*] Dr. B. L. Davis, Dr. J. C. Gordon, Dr. B. Scott, Dr. F. H. Stephens
Chemistry Division, MS J582, Los Alamos National Laboratory
Laboratory, Los Alamos, NM 87545 (USA)
Fax: (+1)505-667-9905
and benzenedithiol failed to produce a pure material even
under driving conditions (heat and gas removal by freeze–
pump–thaw cycles). When Li[B(C6H4S2)2] was prepared
independently according to the literature procedure and
examined by 11B NMR spectroscopy, the same resonance (d =
10.5 ppm) was observed, in contrast to that reported (d =
12.1 ppm) in the literature.[14] Both resonances (at d = ꢀ5.6
and 10.5 ppm) are also observed in the reaction of borazine
and benzenedithiol, along with concomitant H2 formation.
This observation suggests that [NH4][B(C6H4S2)2] might
originate from reaction of benzenedithiol and (C6H4S2)-
E-mail: jgordon@lanl.gov
Prof. D. A. Dixon, E. B. Garner, Dr. M. H. Matus
Department of Chemistry, The University of Alabama
Shelby Hall, Box 870336, Tuscaloosa, Al 35487-0336 (USA)
Fax: (+1)205-348-9104
E-mail: dadixon@bama.ua.edu
[**] This work was funded by the U.S. Department of Energy, Office of
Energy Efficiency and Renewable Energy. We would like to thank Drs.
Troy Semelsberger and Roshan Shrestha for GC characterization of
H2 released during digestion. D.A.D. thanks the Robert Ramsay
Fund at the University of Alabama for partial support.
ꢀ
B H·(NH3) as well as any H2-depleted boron contained
within the spent fuel. The full digestion reaction supported by
our observations is depicted in Scheme 2. A feature of the
first step in this cycle requires highlighting: some of the
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200900680.
Angew. Chem. Int. Ed. 2009, 48, 6812 –6816
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
6813

Communications
Subsequent replacement of the dithiolate ligand in
ꢀ
(C6H4S2)B H·(NH3) with hydride to regenerate AB proved
more difficult. A series of hydride sources were examined, all
of which either over-reduced the boron to borohydride or did
not react at all. We hypothesized that overcoming the chelate
effect of the dithiolate ligand in the first reaction shown in
Table 2 might be a substantial problem and thus screened
suitable hydrides using theory. DFT calculations of the
Scheme 2. Digestion of PB or borazine with benzenedithiol.
nitrogen in the spent fuel is transformed into NH3, which
Table 2: Estimates of reaction energies for reduction.
ꢀ
is subsequently retained by (C6H4S2)B H·(NH3). This
Reaction
DH
process contrasts with other proposed methods, which
(298 K)[a]
+
solely generate NH4 salts (which require thermal
cracking to release and recycle NH3).[8]
ꢀ3.7/2.8
ꢀ9.7/0.8
To transform the products of Scheme 2 into AB, a
reductant is required. Bu3SnH was judged to be a good
starting point, as the reagent is commercially available
and a literature report suggested a possible Bu3SnH
regeneration method by decarboxylation of a tin for-
mate.[15] When Bu3SnH is added to the mixture of
products in Scheme 2, [NH4][B(C6H4S2)2] is fortuitously
[a] Condensed-phase/gas-phase values in kcalmolꢀ1
.
ꢀ
transformed into (C6H4S2)B H·(NH3) (Scheme 3,
reactions of Bu3SnH and Bu2SnH2 (Table 2) were consistent
with this hypothesis and predicted that Bu2SnH2 should react
more favorably on the basis of thermodynamics. Experiments
11B NMR spectrum in Figure 1). The other product is
C6H4SH(S SnBu3) (2), as observed by 119Sn NMR spectros-
ꢀ
ꢀ
copy and verified by an independent synthesis from benze-
nedithiol and Bu3SnH (Figure S1 in the Supporting Informa-
tion). Thus in two steps, we can convert PB into a single new
boron-containing product.
subsequently confirmed that (C6H4S2)B H·(NH3) was trans-
formed into AB using a slight excess of Bu2SnH2 (Figure 2),
Scheme 3. Reduction of [NH4][B(C6H4S2)2] with Bu3SnH.
ꢀ
Figure 2. Reduction of (C6H4S2)B H·(NH3) (1) with Bu2SnH2. Bottom
plot is the 11B NMR spectrum of (C6H4S2)B H·(NH3), top plot is the
ꢀ
spectrum of AB after reduction.
with concomitant formation of Bu2Sn(C6H4S2) (3, Figure 2 in
the Supporting Information). The absence of borohydride in
this hydride transfer makes this reaction very significant.
Practically, the first step in this regeneration cycle is
difficult to optimize because of the apparent side reaction that
results in [NH4][B(C6H4S2)2] formation. Therefore, to obtain
an estimate of the best possible AB yield, excess Bu3SnH was
used after the digestion step both to consume unreacted
benzenedithiol and to convert any [NH4][B(C6H4S2)2] into
Figure 1. Reduction of [NH4][B(C6H4S2)2] with Bu3SnH. Bottom plot
is the 11B NMR spectrum of (C6H4S2)B H·(NH3) (1) and [NH4]-
ꢀ
[B(C6H4S2)2]; top plot is the spectrum of the resulting (C6H4S2)-
ꢀ
B H·(NH3) after reduction with Bu3SnH.
6814
www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 6812 –6816

Angewandte
Chemie
ꢀ
(C6H4S2)B H·(NH3). The thermal instability of Bu2SnH2
In conclusion, we have shown that one form of spent AB
prompted stepwise addition of excess reductant in the final
step to ensure maximum yield of AB. Using this methodolgy,
an overall yield of 67% isolated AB was obtained on our first
attempt. The complete regeneration cycle is shown in
Scheme 4.
fuel, polyborazylene, can be regenerated efficiently in a one-
pot process by the stepwise addition of appropriate digesting
and reducing agents. A unique feature of the described
process is the formation of NH3, which is retained by some of
the digested species. Future reports will detail a more
complete regeneration cycle, including the recycling of the
tin by-product, use of H2 as the energy source, and improved
overall efficiencies.
Experimental Section
Synthetic details and characterization of compounds 1–3 are given in
the Supporting Information. All reactions were performed in an inert
atmosphere using standard Schlenk line and glovebox techniques.
Ethereal solvents and toluene were distilled from Na/benzophenone
ketyl radical. 1H (400 MHz), 13C (100 MHz), 11B (128 MHz), and 119Sn
(149 MHz) NMR spectra were recorded at room temperature (unless
1
otherwise noted) on a Bruker AVANCE 400 MHz spectrometer. H
spectra were referenced to the signal of residual 1H nuclei in the
deuterated solvent and 13C NMR to solvent 13C signal. 119Sn NMR
spectra were referenced to Me4Sn. 11B NMR spectra were referenced
to an internal BF3·etherate standard placed in a stem coaxial insert
(Wilmad). 1,2-benzenedithiol (96% Acros) was sublimed prior to use
and stored cold (ꢀ208C). Polyborazylene (PB) was formed by the
slow decomposition of borazine, received from Gelest. nBu2SnH2 was
prepared by the literature method, distilled, and stored in the absence
of light at ꢀ208C.[16]
Scheme 4. Demonstrated off-board regeneration scheme for spent
ammonia borane.
Regeneration of AB from polyborazylene: PB (0.049 g,
0.61 mmol) was dissolved in THF (20 mL) and combined with 1,2-
benzenedithiol (0.260 g, 1.83 mmol). This mixture was heated over-
night at 608C. 11B NMR spectroscopy indicated consumption of PB
and formation of two new peaks corresponding to 1 and [NH4]-
[B(C6H4S2)2]. Excess Bu3SnH (240 mL, 0.9 mmol) was added to
reduce [NH4][B(C6H4S2)2] to 1 and convert unreacted benzenedithiol
to 3. This solution was heated to 608C before the addition of
nBu2SnH2 (1020 mL, 4.86 mmol) in four portions over 40 min,
reducing 1 to AB (0.038 g, 67%), which was isolated by washing
with toluene. 11B and 119Sn NMR spectra indicate no undesirable by-
products with the isolated AB.
Reaction of borazine and 1,2-benzenedithiol: Borazine (0.015 g,
0.186 mmol) and 1,2-benzenedithiol (0.080 g, 0.559 mmol) were
combined in THF (0.75 mL) and heated to 658C overnight. The
resulting 11B NMR spectrum revealed two resonances, d = 10.5 ppm
(s) and d = ꢀ5.6 ppm (d, 128 Hz). When the headspace gas was
sampled by GC, more H2 gas was detected than in the control sample
(only borazine and solvent).
As part of the DOE Chemical Hydrogen Storage CoE, we
would like to have a method for differentiating (to a first
approximation) the various regeneration schemes in the
literature and estimating which processes might be inefficient
if thoroughly evaluated with all reagents recycled. To estimate
the optimum energetic efficiency for AB regeneration
schemes, Equation (1) has been developed on the basis of
the energy of oxidation for one mole of H2 to form H2O
(57.8 kcal).
efficiency ¼
ðequiv H2 storedÞð57:8Þ
P
P
ðequiv H2 usedÞð57:8Þ þ ðDHendoÞꢀð% heat recoveryÞ ðꢀDHexo
Þ
ð1Þ
Computational details: The calculations were performed at the
DFT level as described in the Supporting Information. Heats of
formation in the gas-phase compounds were obtained from isodesmic
reactions, and COSMO-RS predictions[17] of the boiling points were
used to estimate vaporization enthalpies to obtain heats of formation
of the liquid.
Using Equation (1), our regeneration process in which all
reagents and by-products are recycled has a theoretical peak
energy efficiency that is higher than that of the only other
peer-reviewed account of AB regeneration (Table 3, see the
Supporting Information for complete equations). We realize
that other factors (upscale yield, engineering) will play a
significant role in the implementation of any regeneration
process on a large scale.
Received: February 4, 2009
Published online: June 9, 2009
Keywords: amine–boranes · density functional calculations ·
.
hydrogen storage · NMR spectroscopy · sustainable chemistry
Table 3: Efficiency estimates for AB regeneration.
Regeneration process
Efficiency
(0% heat recovery)
[1] F. H. Stephens, V. Pons, R. T. Baker, Dalton Trans. 2007, 2613 –
2626.
[2] M. H. Matus, K. D. Anderson, D. M. Camaioni, S. T. Autrey,
D. A. Dixon, J. Phys. Chem. A 2007, 111, 4411 – 4421.
[10]
ꢀ
NH4B(OMe)4 +3H2!H3N BH3 +4MeOH
46%
65%
1
ꢀ
= B3N3H6 +2H2!H3N BH3
3
Angew. Chem. Int. Ed. 2009, 48, 6812 –6816
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
6815

Communications
[3] M. E. Bluhm, M. G. Bradley, R. Butterick III, U. Kusari, L. G.
[11] Targets for On-Board Hydrogen Storage Systems: Current R&D
Focus is on 2010 Targets, Department of Energy, Energy
Efficiency and Renewable Resources, Accessed online 01/28/
09 (http://www1.eere.energy.gov/hydrogenandfuelcells/storage/
pdfs/targets onboard hydro storage.pdf).
[12] C. A. Jaska, K. Temple, A. J. Lough, I. Manners, J. Am. Chem.
Soc. 2003, 125, 9424 – 9434.
[13] N. Blaquiere, S. Diallo-Garcia, S. L. Gorelsky, D. A. Black, K.
Fagnou, J. Am. Chem. Soc. 2008, 130, 14034 – 14035.
[14] J. Knizek, H. Noth, J. Organomet. Chem. 2000, 614–615, 168 –
187.
[15] R. J. Klingler, I. Bloom, J. W. Rathke, Organometallics 1985, 4,
1893 – 1894.
[16] A. G. Hernꢁn, P. N. Horton, M. B. Hursthouse, J. D. Kilburn, J.
Organomet. Chem. 2006, 691, 1466 – 1475.
[17] a) A. Klamt, Quantum Chemistry to Fluid Phase Thermody-
namics and Drug Design, Elsevier, Amsterdam, 2005; b) A.
Klamt, V. Jonas, T. Bꢂrger, J. C. W. Lohrenz, J. Phys. Chem. A
1998, 102, 5074 – 5085.
Sneddon, J. Am. Chem. Soc. 2006, 128, 7748 – 7749.
[4] M. C. Denney, V. Pons, T. J. Hebden, D. M. Heinekey, K. I.
Goldberg, J. Am. Chem. Soc. 2006, 128, 12048 – 12049.
[5] R. J. Keaton, J. M. Blacquiere, R. T. Baker, J. Am. Chem. Soc.
2007, 129, 1844 – 1845.
[6] F. H. Stephens, R. T. Baker, M. H. Matus, D. J. Grant, D. A.
Dixon, Angew. Chem. 2007, 119, 760 – 763; Angew. Chem. Int.
Ed. 2007, 46, 746 – 749.
[7] L. G. Sneddon, Amineborane Hydrogen Storage, Department of
Energy Hydrogen Program Review, 2006 (http://www.hydroge-
n.energy.gov/pdfs/review06/st_3_sneddon.pdf).
[8] L. G. Sneddon, Amineborane-Based Chemical Hydrogen Stor-
age, Department of Energy Hydrogen Program Review, 2007
(http://www.hydrogen.energy.gov/pdfs/review07/st_27_sned-
don.pdf).
[9] S. Hausdorf, F. Baitalow, G. Wolf, F. O. R. L. Mertens, Int. J.
Hydrogen Energy 2008, 33, 608 – 614.
[10] P. V. Ramachandran, P. D. Gagare, Inorg. Chem. 2007, 46, 7810 –
7817.
6816
www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 6812 –6816