High-capacity hydrogen storage in lithium and sodium amidoboranes

LETTERS

High-capacity hydrogen storage in lithium and sodium amidoboranes ZHITAO XIONG1 , CHAW KEONG YONG1 , GUOTAO WU1 , PING CHEN1,2 *, WENDY SHAW3 , ABHI KARKAMKAR3 , THOMAS AUTREY3 , MARTIN OWEN JONES4 , SIMON R. JOHNSON4 , PETER P. EDWARDS4 AND WILLIAM I. F. DAVID5 1

Department of Physics, National University of Singapore, Singapore 117542, Singapore Department of Chemistry, National University of Singapore, Singapore 117542, Singapore 3 Pacific Northwest National Laboratories, Richland, Washington 99352, USA 4 Inorganic Chemistry Laboratory, University of Oxford, South Parks Road, Oxford OX1 3QR, UK 5 ISIS Facility, Rutherford Appleton Laboratory, Chilton OX11 0QX, UK * e-mail: [email protected] 2

Published online: 23 December 2007; doi:10.1038/nmat2081

The safe and efficient storage of hydrogen is widely recognized as one of the key technological challenges in the transition towards a hydrogen-based energy economy1,2 . Whereas hydrogen for transportation applications is currently stored using cryogenics or high pressure, there is substantial research and development activity in the use of novel condensed-phase hydride materials. However, the multiple-target criteria accepted as necessary for the successful implementation of such stores have not yet been met by any single material. Ammonia borane, NH3 BH3 , is one of a number of condensed-phase compounds that have received significant attention because of its reported release of ∼12 wt% hydrogen at moderate temperatures (∼150 ◦ C). However, the hydrogen purity suffers from the release of trace quantities of borazine. Here, we report that the related alkali-metal amidoboranes, LiNH2 BH3 and NaNH2 BH3 , release ∼10.9 wt% and ∼7.5 wt% hydrogen, respectively, at significantly lower temperatures (∼90 ◦ C) with no borazine emission. The lowtemperature release of a large amount of hydrogen is significant and provides the potential to fulfil many of the principal criteria required for an on-board hydrogen store. One of the most promising materials suggested as a potential solid-state hydrogen store for a future sustainable hydrogen economy is ammonia borane (NH3 BH3 ). This molecular material is a stable solid at room temperature and pressure and is neither flammable nor explosive. Importantly, it contains ∼19.6 wt% hydrogen. Ammonia borane, first synthesized in 1955, is a plastic crystalline solid adopting a tetragonal crystal structure with space group I 4¯ mm and lattice parameters of a = b = 5.240 A˚ and c = 5.028 A˚ at room temperature3 . It thermally decomposes between 70 ◦ C and 112 ◦ C to yield polyaminoborane, [NH2 BH2 ]n , and hydrogen4–8 , reaction (1). In turn, polyaminoborane decomposes over a broad temperature range from about 110 ◦ C to approximately 200 ◦ C with further hydrogen loss, forming polyiminoborane, [NHBH]n , and a small fraction of borazine, [N3 B3 H6 ], reactions (2) and (3) respectively5 . The decomposition of [NHBH]n to BN occurs at temperatures in excess of 500 ◦ C (ref. 5). This final step is not considered practical

for hydrogen storage.

n NH3 BH3 (s) → [NH2 BH2 ]n (s) + n H2 (g)

(1)

[NH2 BH2 ]n (s) → [NHBH]n (s) + n H2 (g)

(2)

[NH2 BH2 ]n (s) → [N3 B3 H6 ]n/3 (l) + n H2 (g)

(3)

Recent work has investigated various approaches to lower the decomposition temperature of ammonia borane through the use of nanoscaffolds, iridium or base-metal catalyst, carbon cryogel and ionic liquids9–14 and so on. We have adopted an approach that has been applied in the manipulation of the thermodynamic property of compounds through chemical alteration2 to modify ammonia borane, that is, through substituting one H in the NH3 group in BH3 NH3 with a more electron-donating element. The rationale behind this approach is to alter the polarity and intermolecular interactions (specifically the dihydrogen bonding) of ammonia borane to produce a substantially improved dehydrogenation profile. Here, we report that lithium amidoborane (LiNH2 BH3 ) and sodium amidoborane (NaNH2 BH3 ) show substantially different and improved dehydrogenation characteristics with respect to ammonia borane itself. These alkali-metal amidoborane materials are formed through the interactions of alkali-metal hydrides (LiH and NaH) with ammonia borane, which lead to the replacement of a single ammonia borane hydrogen atom by a lithium or sodium atom. We observed that more than 10 wt% and 7 wt% of hydrogen desorbs from LiNH2 BH3 and NaNH2 BH3 , respectively, at around 90 ◦ C. Lithium and sodium amidoboranes were prepared by ballmilling 1:1 molar ratios of ammonia borane (NH3 BH3 ) and the corresponding alkali-metal hydrides (see the Supplementary Information). A gradual pressure increase within the ball-mill vessel was observed and the gaseous product was identified as H2 by mass spectrometry. The amount of hydrogen evolved from the starting chemicals was calculated from this pressure increase

138

nature materials VOL 7 FEBRUARY 2008 www.nature.com/naturematerials

© 2008 Nature Publishing Group

LETTERS 26.5 29.8

–22.8 –19.7

(iv) (iii) 40

20 p.p.m.

0

(ii) (i) 0

–10

–20

–30 p.p.m.

–40

–50

–60

Figure 2 High-field 289.2 MHz (21.2 T) 11 B NMR of LiNH2 BH3 and NH3 BH3 samples. (i), As-prepared LiNH2 BH3 sample (−19.7) p.p.m.; (ii), untreated NH3 BH3 (−22.8 p.p.m.); (iii), LiNH2 BH3 sample after dehydrogenation to 140 ◦C (+29.8 p.p.m.); (iv), polyborazylene (+26.5 p.p.m.).

Figure 1 Schematic diagram of the crystal structure of LiNH2 BH3 and NaNH2 BH3 determined from high-resolution X-ray powder diffraction data at room temperature. Boron is represented by orange spheres, nitrogen by green spheres, hydrogen by white spheres and lithium by red spheres.

and it was observed that about 1.0 molar equivalent of H2 was released from both LiH–NH3 BH3 and NaH–NH3 BH3 mixtures within 3 h. In contrast, no gaseous products were detected when pure ammonia borane was ball-milled under the same conditions (see the Supplementary Information). The release of one mole of H2 per mole of ammonia borane and alkali-metal hydride during the milling process suggests that reactions (4) and (5) have occurred: NH3 BH3 (s) + LiH (s) → LiNH2 BH3 (s) + H2 (g)

(4)

NH3 BH3 (s) + NaH (s) → NaNH2 BH3 (s) + H2 (g),

(5)

leading to the formation of alkali-metal amidoboranes, LiNH2 BH3 and NaNH2 BH3 . It is likely that one of the driving forces for the reactions above may come from the high potential for the Hδ+ in NH3 and Hδ− in alkali-metal hydrides to combine and form H2 (ref. 2). The formation of LiNH2 BH3 and NaNH2 BH3 in solution was reported in 1996 (ref. 15) and 1938 (ref. 16), respectively. LiNH2 BH3 , termed lithium amidotrihydroborate and used as a reducing agent for organic synthesis, was produced from a reaction between ammonia borane and n-butyllithium in tetrahydrofuran15 . The identity of the substituted ammonia borane species in solution

was confirmed by 11 B NMR. The reaction of (CH3 )2 OBH3 and Na in liquid ammonia produces NaNH2 BH3 (ref. 16). However, no crystal structure information or decomposition properties of LiNH2 BH3 or NaNH2 BH3 have previously been reported. Hence, to confirm our assumption that LiNH2 BH3 and NaNH2 BH3 had been formed, high-resolution X-ray powder diffraction data were collected on the post-ball-milled LiH–NH3 BH3 and NaH–NH3 BH3 samples on the high-resolution powder diffractometer ID31 at the ESRF, Grenoble. The structures (Fig. 1) were solved using the computer program DASH17 , which confirmed the stoichiometry to be LiNH2 BH3 and NaNH2 BH3 with Li/Na directly substituting for one of the ammonia hydrogen atoms. LiNH2 BH3 and NaNH2 BH3 are isostructural, crystallizing in the orthorhombic space group Pbca. For LiNH2 BH3 , the ˚ b = 13.94877(14) A, ˚ lattice constants are a = 7.11274(6) A, ˚ V = 510.970(15) A˚ 3 , whereas for NaNH2 BH3 , c = 5.15018(6) A, ˚ b = 14.65483(16) A, ˚ the lattice constants are a = 7.46931(7) A, ˚ V = 618.764(20) A˚ 3 . Both the Li–N (1.98 A) ˚ and c = 5.65280(8) A, ˚ are substantially longer than the H–N bond in Na–N bonds (2.35 A) ammonia borane, which is indicative of substantial M+ (NH2 BH3 )− ˚ is ionic character. As expected, the B–N bond length (1.56 A) slightly shorter than in ammonia borane, revealing stronger bonding between B and N in the alkali-metal amidoboranes, which is in agreement with the predictions of Armstrong et al.18 . The substitution of H by a stronger electron-donating alkali metal will induce considerable changes in the electronic state of N with concomitant modification of the chemical bonding between B and N. As a consequence, the chemical bonding of B–N, B–H and N–H will be affected. LiNH2 BH3 was characterized to illustrate these changes. High-field 11 B NMR studies revealed a boron species with a chemical shift of −19.7 p.p.m. (Fig. 2), which is consistent with the previously published NMR data for the LiNH2 BH3 species15 . The subtle downfield shift of the 11 B resonance in the LiNH2 BH3 complex compared with that in ammonia borane (−22.8 p.p.m.) suggests that the lithium amido group—NH2 (Li)—may form a stronger donor complex with borane, which is consistent with the crystallographically observed shortening of the B–N bond length for LiNH2 BH3 . Dehydrogenation of the LiNH2 BH3 and the NaNH2 BH3 samples was investigated using temperature-programmeddesorption (TPD) and volumetric-release techniques. The postmilled ammonia borane was also tested for comparison. On



139

LETTERS DSC

12

LiNH2BH3

(i) 10

(ii) H content (wt%)

(iii) 92

MS-H2 Intensity (a.u.)

108 89

154

(i) (ii)

8

NaNH2BH3

6

BH3NH3

4

(iii) 2 MS-Borazine (i) 0 (ii)

0

(iii) 50

100 150 Temperature (°C)

5

10 Time (h)

15

20

200

Figure 4 Time dependences of hydrogen desorption from alkali amidoboranes and post-milled BH3 NH3 samples at about 91 ◦C.

Figure 3 TPD and DSC spectra. (i), Post-milled ammonia borane; (ii), Li amidoborane sample; (iii) Na amidoborane sample. MS: mass spectrometry.

decomposition, the milled ammonia borane released hydrogen with a small amount of borazine; maximum decomposition rates occurred at 108 ◦ C and 154 ◦ C, respectively (Fig. 3). In contrast, LiNH2 BH3 decomposed directly to hydrogen on heating, with vigorous hydrogen release at around 92 ◦ C; no borazine was detected (Fig. 3). The thermal dehydrogenation of NaNH2 BH3 resembles that of LiNH2 BH3 , but at a slightly lower peak temperature of 89 ◦ C. Volumetric-release measurements showed that about 8 wt% and 6 wt% of hydrogen are released from the LiNH2 BH3 and the NaNH2 BH3 samples, respectively, within the first hour at ∼91 ◦ C. Extending the reaction period to 19 h enables about 11 wt% and 7.4 wt% or about 2.0 equivalent moles of H2 to fully desorb (Fig. 4). In contrast, milled ammonia borane released only about 5.3 wt% or about 0.82 equivalent moles of H2 under similar conditions. Differential scanning calorimetry (DSC) measurements (Fig. 3) indicated that the heat of hydrogen desorption from LiNH2 BH3 and NaNH2 BH3 is about −3 to −5 kJ mol−1 -H2 , which is significantly less exothermic than from the pristine NH3 BH3 (about −20 kJ mol−1 -H2 (ref. 7)). The nearly thermally neutral dehydrogenation of alkali-metal amidoboranes indicates that rapid near-room-temperature hydrogen release may be feasible if kinetic barriers can be overcome by catalytic modification. As approximately two molar equivalents of H2 were released from LiNH2 BH3 and NaNH2 BH3 , respectively, the final solid products have the chemical composition of LiNBH and NaNBH, as shown by the dehydrogenation reactions (6) and (7).

n LiNH2 BH3 (s) → (LiNBH) n (s) + 2n H2 (g)

10.9 wt%

(6)

n NaNH2 BH3 (s) → (NaNBH) n (s) + 2n H2 (g)

7.5 wt%

(7)

High-resolution X-ray powder diffraction data indicated that the post-TPD samples are very poorly crystalline (see the Supplementary Information). A chemical shift of 29.8 p.p.m. was observed in LiNBH in 11 B NMR measurements (Fig. 2), consistent with a trigonal planar N–BH–N environment for boron19 , which strongly suggests the final product is a borazine-like or polyborazine-like compound. Similar phenomena were observed for the NaNBH sample (see the Supplementary Information).

One of the key features of the alkali-metal amidoboranes is the presence of hydrogen atoms with both positive (that is, H in –NH2 ) and negative (that is, H in –BH3 ) charges. This is reminiscent of the amide–hydride combination20,21 and also of borohydride– amide materials22 . One of the advantages of the dehydrogenation of alkali amidoboranes is that it may not necessarily involve an interface reaction and mass transport through different phases as for the amide–hydride combination23,24 but rather involve the local combination of Hδ+ and Hδ− as in Li4 BH4 (NH2 )3 (ref. 22), which more readily produces H2 . LiNH2 BH3 and NaNH2 BH3 are the first two examples of a potentially large class of metal amidoboranes. The [NH2 BH3 ]− anion is likely to bond with other metal or metalloid elements to form a variety of corresponding amidoboranes, which will not only open up significant opportunities for hydrogen-storage materials development, but will also bring considerable insights to B–N chemistry. The search for a safe, economical, hydrogen-rich store is a key materials challenge in the move towards a hydrogen-based energy economy. We have discovered that alkali-metal amidoboranes, LiNH2 BH3 and NaNH2 BH3 , satisfy a number of the principal criteria demanded for hydrogen-storage media. LiNH2 BH3 and NaNH2 BH3 provide high storage capacity (10.9 wt% and 7.5 wt%, respectively) of hydrogen at easily accessible dehydrogenation temperatures (about 90 ◦ C) without the unwanted by-product borazine. These materials offer significant advantages over their parent compound, ammonia borane. Alkali-metal amidoboranes are also environmentally harmless, non-flammable, non-explosive and stable solids at room temperature and pressure. At present, the only disadvantageous aspect of these materials is their lack of facile reversibility. However, the almost thermal neutral dehydrogenation of alkali-metal amidoboranes will greatly facilitate their offboard regeneration.

METHODS SAMPLE PREPARATION LiH, NaH and NH3 BH3 were purchased from Sigma-Aldrich with stated purities of 95%, 95% and above 90%, respectively, and were used without further purification. The LiH–NH3 BH3 and NaH–NH3 BH3 mixtures in a hydride/NH3 BH3 molar ratio of 1:1 and pristine NH3 BH3 were ball-milled on a Retsch PM400 planetary mill at 200 r.p.m., respectively. Graphite was added to the mixture to enhance ball-milling efficiency. After milling the samples, gases released were identified by a mass spectrometer and the pressure was measured

140



LETTERS by a pressure gauge. All samples were handled in a MBRAUN glove box filled with purified argon gas. CHARACTERIZATION TPD was used over a gas chromatography–mass spectrometry combined system to investigate the thermal desorption properties of the post-milled samples. Detailed operation procedures are described elsewhere20 . Quantitative measurement of hydrogen desorption at elevated temperatures was carried out on a commercial gas reactor controller provided by the Advanced Materials Corporation. A Netzsch DSC 204 HP unit was applied in detecting the heat flow in the dehydrogenation process. Structural identifications were carried out on both a Bruker X-ray diffractometer equipped with an in situ cell at NUS and the high-resolution synchrotron X-ray powder diffractometer on ID31 at the ESRF, Grenoble. 11 1 B{ H} magic-angle-spinning NMR experiments were carried out at room temperature on a Bruker Advance 400 NMR spectrometer operating at 9.7 T on 128.3 MHz 11 B frequency at NUS and on a Varian Unity Inova console operating at 900 MHz 1 H frequency at PNNL. Spectra were obtained using a two-channel custom-built probe with a 3.2 mm rotor, a 4 µs 90◦ pulse, a 20 s pulse delay and 15 kHz spinning speed. Two hundred to three hundred scans were taken for the ball-milled and post-TPD LiH–NH3 BH3 samples, whereas model compounds were obtained with 20–40 scans.

Received 5 May 2007; accepted 8 November 2007; published 23 December 2007. References 1. Schlapbach, L. & Zuttel, A. Hydrogen-storage materials for mobile applications. Nature 414, 353–358 (2001). 2. Grochala, W. & Edwards, P. P. Thermal decomposition of the non-interstitial hydrides for the storage and production of hydrogen. Chem. Rev. 104, 1283–1315 (2004). 3. Shore, S. G. & Parry, R. W. The crystalline compound ammonia-borane, H3 NBH3 . J. Am. Chem. Soc. 77, 6084–6085 (1955). 4. Stowe, A. C., Shaw, W. J., Linehan, J. C., Schmid, B. & Autrey, T. In situ solid state 11 B MAS-NMR studies of the thermal decomposition of ammonia borane: Mechanistic studies of the hydrogen release pathways from a solid state hydrogen storage material. Phys. Chem. Chem. Phys. 9, 1831–1836 (2007). 5. Hu, M. G., Geanangel, R. A. & Wendlandt, W. W. The thermal decomposition of ammonia borane. Thermochim. Acta 23, 249–255 (1978). 6. Sit, V., Geanangel, R. A. & Wendlandt, W. W. The thermal dissociation of NH3 BH3 . Thermochim. Acta 113, 379–382 (1987). 7. Wolf, G., Baumann, J., Baitalow, F. & Hoffmann, F. P. Calorimetric process monitoring of thermal decomposition of B–N–H compounds. Thermochim. Acta 343, 19–25 (2000). 8. Baitalow, F., Baumann, J., Wolf, G., Jaenicke-R¨oßler, K. & Leitner, G. Thermal decomposition of B–N–H compounds investigated by using combined thermoanalytical methods. Thermochim. Acta 391, 159–168 (2002).

9. Gutowska, A. et al. Nanoscaffold mediates hydrogen release and the reactivity of ammonia borane. Angew. Chem. Int. Edn 44, 3578–3582 (2005). 10. Denney, M. C., Pons, V., Hebden, T. J., Heinekey, M. & Goldberg, K. I. Efficient catalysis of ammonia borane dehydrogenation. J. Am. Chem. Soc. 128, 12048–12049 (2006). 11. Bluhm, M. E., Bradley, M. G., Butterick, R., Kusari, U. & Sneddon, L. G. Amineborane-based chemical hydrogen storage: Enhanced ammonia borane dehydrogenation in ionic liquids. J. Am. Chem. Soc. 128, 7748–7749 (2006). 12. Keaton, R. J., Blacquiere, J. M. & Baker, R. T. Base metal catalyzed dehydrogenation of ammonia-borane for chemical hydrogen storage. J. Am. Chem. Soc. 129, 1844–1845 (2007). 13. Clark, T. J., Lee, K. & Manners, I. Transition-metal-catalyzed dehydrocoupling: A convenient route to bonds between main-group elements. Chem. Eur. J. 12, 8634–8648 (2006). 14. Feaver, A. et al. Coherent carbon cryogel-ammonia borane nanocomposities for H-2 storage. J. Phys. Chem. B 111, 7469–7472 (2007). 15. Myers, A. G., Yang, B. H. & Kopecky, D. J. Lithium amidotrihydroborate, a powerful new reductant. Transformation of tertiary amides to primary alcohols. Tetrahedron Lett. 37, 3623–3626 (1996). 16. Schlesinger, H. I. & Burg, A. B. Hydrides of boron. VIII. The structure of the diammoniate of diborane and its relation to the structure of diborane. J. Am. Chem. Soc. 60, 290–299 (1938). 17. David, W. I. F., Shankland, K. & Shankland, N. Routine determination of molecular crystal structures from powder diffraction data. Chem. Commun. 8, 931–932 (1998). 18. Armstrong, D. R., Perkins, P. G. & Walker, G. T. The electronic structure of the monomers, dimers, a trimer, the oxides and the borane complexes of the lithiated ammonias. THEOCHEM-J. Mol. Struct. 122, 189–203 (1985). 19. Gervais, C. et al. B-11 and N-15 solid state NMR investigation of a boron nitride preceramic polymer prepared by ammonolysis of borazine. J. Eur. Ceram. Soc. 25, 129–135 (2005). 20. Chen, P., Xiong, Z. T., Luo, J. Z., Lin, J. Y. & Tan, K. L. Interaction of hydrogen with metal nitrides and imides. Nature 420, 302–304 (2002). 21. Chen, P., Xiong, Z. T., Luo, J. Z., Lin, J. Y. & Tan, K. L. Interaction between lithium amide and lithium hydride. J. Phys. Chem. B 107, 10967–10970 (2003). 22. Chater, P. A., David, W. I. F., Johnson, S. R., Edwards, P. P. & Anderson, P. A. Synthesis and crystal structure of Li4 BH4 (NH2 )3 . Chem. Commun. 23, 2439–2441 (2006). 23. David, W. I. F. et al. A Mechanism for non-stoichiometry in the lithium amide/lithium imide hydrogen storage reaction. J. Am. Chem. Soc. 129, 1594–1601 (2007). 24. Chen, P., Xiong, Z. T., Yang, L. F., Wu, G. T. & Luo, W. F. Mechanistic investigations on the heterogeneous solid-state reaction of magnesium amides and lithium hydrides. J. Phys. Chem. B 110, 14221–14225 (2006).

Acknowledgements P.C., Z.X. and G.W. are grateful for the financial support from A∗ STAR, Singapore, and helpful discussions with T. Kemmitt and M. Bowden from IRL and L. Sneddon from Univ. Pennsylvania. T.A., W.S. and A.K. would like to thank the DOE Center of Excellence in Chemical Hydrogen Storage for support. M.O.J., S.J. and P.E. would like to thank the EPSRC (SUPERGEN) for financial support. We wish to thank A. Fitch, M. Brunelli and I. Margiolaki for assistance in using the high-resolution beamline ID31 at the ESRF, Grenoble (France). A portion of the research described here was carried out in the Environmental Molecular Sciences Laboratory, a national scientific user facility sponsored by the US Department of Energy’s Office of Biological and Environmental Research and located at Pacific Northwest National Laboratory. This work was carried out as a collaboration established by the IPHE project ‘Combination of Amine Boranes with MgH2 & LiNH2 for High Capacity Reversible Hydrogen Storage’. Correspondence and requests for materials should be addressed to P.C. Supplementary Information accompanies this paper on www.nature.com/naturematerials. Reprints and permission information is available online at http://npg.nature.com/reprintsandpermissions/



141