Synthesis of ammonia borane for hydrogen storage applications

Abstract

A new synthetic procedure to make the condensed phase hydrogen storage material, ammonia borane (NH₃BH₃, abbreviated as AB), is described and compared with previous literature procedures. Ammonia borane with a gravimetric density ca. 194 gm H₂ kg⁻¹ and a volumetric density ca. 146 H₂ litre⁻¹, is a promising chemical hydrogen storage material for fuel cell powered applications. The work shows that ammonium borohydride, NH₄BH₄, formed in situ by the metathesis of NH₄X and MBH₄ salts (M = Na, Li; X = Cl, F) in liquid NH₃, can be induced to decompose in an organic ether to yield AB in near quantitative yield. The purity of the AB prepared by this one-pot synthetic strategy is sufficient to meet the thermal stability requirements for on-board hydrogen storage.

---

Introduction

For many years, the synthesis, isolation, and characterization of the simple Lewis acid–base complex AB was an elusive endeavour.^1,2 It was well established that Lewis acids such as CO and Me₃N reacted with diborane to give the monomeric donor–acceptor adducts, (CO)–BH₃ and Me₃N–BH₃, respectively; however, the reaction of ammonia with diborane appeared to proceed by a very different pathway.³ Alfred Stock had shown that the stoichiometry was unusual; two moles of ammonia react with one mole of diborane to form a salt-like compound represented by the empirical formula B₂H₆·2NH₃.⁴ Subsequent molecular weight determinations by others confirmed the dimeric nature of the diammoniate of diborane (DADB);^5,6 however, there was considerable ambiguity in the identity of the molecular nature of the structure ranging from the ammonium cation, I, the borohydride anion, II, to a structure containing both an ammonium cation and a borohydride anion separated by an aminoborane, III.^7–9

In a classic series of papers from Parry's laboratory in the 1950s, the structure of DADB was finally proven to be the species containing the boronium cation and borohydride ion pair, which is species II in the scheme shown above.^10–16 It was during these experimental studies that serendipity played a role in providing AB. Sheldon Shore had added NH₄Cl to the DADB in an attempt to prepare the corresponding chloride salt [NH₃BH₂NH₃]⁺ [Cl]⁻, eqn (1), in analogy with the previous work of Schultz who showed that the addition of NH₄Br to DADB in liquid ammonia yielded the corresponding bromide salt [NH₃BH₂NH₃]⁺ [Br]⁻, eqn (2).¹ However, Shore discovered an unexpected side reaction that formed AB.¹⁷ As it turned out, the choice of the solvent changed the course of the reaction, and he subsequently demonstrated that AB could be formed directly from a mixture of NH₄Cl and LiBH₄ in diethyl ether when a trace of NH₃ was present, eqn (3). Consequently, the metathesis reaction in organic solvents has been one of the most used synthetic strategies to prepare ammonia borane.^18,19

NH₄Cl + II → NH₃BH₃ + [NH₃BH₂NH₃]⁺ [Cl]⁻ + H₂ (1)

NH₄Br + II → NH₄BH₄ + [NH₃BH₂NH₃]⁺ [Br]⁻ (2a)

NH₄BH₄ → II + H₂ (2b)

NH₄Cl + LiBH₄ → NH₃BH₃ (THF with trace NH₃) (3)

A few years later, Shore and Böddeker described reaction conditions that enabled the preparatory scale synthesis of DADB and mixtures of DADB with AB.²⁰ They showed that passing diborane gas into liquid ammonia gave the asymmetric cleavage product DADB, eqn (4), in quantitative yield at temperatures below −78 °C.²¹ On the other hand, when they distilled liquid NH₃ onto a solution of BH₃–THF complex at −78 °C, they observed both DADB and the symmetric cleavage product AB in equimolar amounts, eqn (5).

Asymmetric cleavage

Symmetric cleavage:

Subsequently, Mayer reported that the basicity of the solvent had an influence on the competing symmetric and asymmetric reaction pathways.²² He found that AB could be prepared with 68–76% yields in diglyme if gaseous ammonia was added to diborane solutions of ethers. However, when diborane was added to ammonia dissolved in the same solvents the yields were lower and more erratic (i.e., 32–60%). In a paper published the following year, Mayer reported that DADB could be converted to AB, eqn (6), with no apparent H₂ evolution in diglyme containing a trace amount of diborane (yield 80–91%, 40 h at 25 °C).^23,24 He noted that this was in sharp contrast to the fate of DADB in ether solvents containing traces of ammonia. In ether, with a trace of ammonia, Shore observed significant hydrogen evolution with a polymeric product, (NH₂BH₂)_n, and a limited quantity of AB, eqn (7), (ca. <20%).²

In this paper, we describe the details of a modified synthetic strategy for preparing AB for hydrogen storage applications, and we report the surprising result that AB can be prepared in near quantitative yields in a one-pot synthetic strategy. The high yields of isolated AB were surprising for the following two reasons: (1) we found that it was not necessary to remove all traces of ammonia prior to addition of the organic solvent, and (2) we found that it was not necessary to add trace quantities of diborane to get quantitative yields of AB. A synthetic approach to prepare AB in quantitative yields in a single pot will provide researchers with a simple procedure to prepare AB. Furthermore, we envision that this simpler procedure can be scaled up and the solvents can be recycled.²⁵ Efficient routes to the synthesis of material are an important aspect for R&D focused on discovering materials that could be used to store high densities of hydrogen for fuel cell powered applications. Ammonia borane, with a 144 g H₂ L⁻¹ volumetric density and a 194 g H₂ kg⁻¹ gravimetric density, is under investigation by many research groups^26–29 that are looking for hydrogen storage materials to meet system-based hydrogen storage targets (i.e., >82 g H₂ L⁻¹ volumetric density and >90 g H₂ kg⁻¹ gravimetric density) that have been established by the US Department of Energy (DOE).^30–32

Experimental

Sequential addition

Anhydrous NH₃ (25 mL) was condensed in an oven-dried, 100 mL, three-neck, round-bottom flask fitted with a stir bar. The flask was cooled in a dry-ice/isopropanol bath (−78 °C) that was open to a nitrogen atmosphere. Both NH₄Cl (1 g, 18.6 mmol) and NaBH₄ (0.71 g, 18.6 mmol) were added to the liquid ammonia using a solids addition funnel. The mixture was stirred for 2 h under a nitrogen atmosphere at −78 °C. Ammonia then was removed by vacuum, leaving a white polycrystalline powder that was assumed to be NH₄BH₄ and NaCl. Anhydrous THF (100 mL) was cannulated under a nitrogen atmosphere over the residual white solid at −78 °C open to a nitrogen bubbler. After THF addition, the white solid began to foam and release hydrogen. The slurry was stirred at −78 °C for 30 min and then slowly warmed to room temperature. At room temperature, the slurry was stirred for an additional 60 min at which time gas evolution had ceased. The NaCl was filtered and the THF was removed by rotary evaporation to yield 0.57 g of a microcrystalline powder, mp = 110 °C. Analysis of this powder using ¹H NMR in d-glyme showed a triplet at 3.8 ppm (NH₃, J_N–H = 45 Hz) and a 1 : 1 : 1 : 1 quartet centered at 1.5 ppm (BH₃, J_B–H = 94 Hz), and analysis using ¹¹B NMR showed a quartet at −23 ppm (BH₃, J_B–H = 93 Hz), thus providing ca. 99% of the theoretical yield of AB.

Parallel addition I

Anhydrous NH₃ (10 mL) was condensed in an oven-dried, 100 mL, three-neck, round-bottom flask fitted with a stir bar. The flask was cooled in a dry-ice/isopropanol bath (−78 °C) that was open to a nitrogen atmosphere. As before, the NH₄Cl (1.06 g, 18.6 mmol) and NaBH₄ (0.71 g, 18.6 mmol) were added by a solids addition funnel to the three-neck flask, and the reaction was stirred for 2 h under nitrogen at −78 °C. Anhydrous THF (100 mL) was slowly added drop-wise to the flask via an addition funnel, and NH₃ and H₂ were allowed to evaporate while thawing the reaction to room temperature. At room temperature, the flask was stirred for another hour. The precipitated NaCl was removed by filtration, and the THF was removed by rotary evaporation, and the remaining microcrystalline powder then was dried overnight under vacuum. The amount of AB recovered as a microcrystalline powder (mp = 107 °C) was 0.58 g (18.4 mmol, 99% yield). Analysis using ¹H NMR in d-glyme showed a small triplet at 3.8 ppm (NH₃, J_N–H = 45 Hz), and a 1 : 1 : 1 : 1 quartet centered at 1.5 ppm (BH₃, J_B–H = 94 Hz). Analysis using ¹¹B NMR showed a quartet at −23 ppm (BH₃, J_B–H = 93 Hz).

Parallel addition II

Anhydrous NH₃ (25 mL) was condensed in an oven-dried, 100 mL, three-neck, round-bottom flask fitted with a stir bar. The flask was cooled in a dry-ice/isopropanol bath (−78 °C) that was open to a nitrogen atmosphere. A 9.3 mL volume of 2 M LiBH₄ : THF (18.6 mmol) was added to 50 mL anhydrous THF, and then loaded in an addition funnel and slowly added dropwise to the liquid NH₃. NH₄F (0.408 g, 11.0 mmol) was massed in a glove box into a solids addition funnel. The solids addition funnel was connected to the flask, and the NH₄F was slowly added to the THF–ammonia solution and stirred for 1.25 h under nitrogen at −78 °C. After gas evolution had ceased, the flask was warmed to room temperature, and the precipitated NaF was separated from the remaining liquid by filtering. The THF was removed by rotary evaporation and the remaining material then was dried under vacuum overnight to yield 0.420 g of a microcrystalline powder that was composed of a mixture of LiBH₄ and AB. Analysis using ¹¹B NMR in d-glyme showed a quartet at −23 ppm (BH₃, J_B–H = 94 Hz, 87%) and a pentet at −41 ppm (BH₄, J_B–H = 82 Hz).

Results and discussion

The following three general procedures for preparing AB are described in the literature:

1. Condensation of B₂H₆ in organic ethers followed by treatment with gaseous NH₃. This procedure provides AB in yields varying from 45–60%.^20,22

2. Metathesis of NH₄X and MBH₄ in an organic solvent. This procedure provides AB in yields ranging from 30–85%.^{2,12,18,19,33}

3. Dissolving DADB in glyme solvent containing a trace of diborane.²³

The first approach requires making diborane in stoichiometric quantities and results in mixtures of DADB and AB that depend on the reaction temperature and solvent used in the preparation. The second approach does not require diborane; however, to get reasonable yields of AB from the metathesis approach, it has been suggested that the synthesis must be performed under dilute solvent reaction conditions. The third approach, dissolving DADB in glyme containing a trace of diborane, makes one consider if it may be possible to make AB from DADB prepared in situ from decomposition of the NH₄BH₄. There does not appear to be any need to isolate DADB. The judicious observations reported by Parry, Schultz, and Shore in their series of eight sequential papers in J. Am. Chem. Soc. combined with the reports from Mayer's^22,23 and Geanangels'³³ research groups led us to consider alternative approaches to prepare AB for hydrogen storage studies. Some of the important observations are summarized below:

• NH₄BH₄ is stable at −40 °C and in liquid NH₃.

• NH₄BH₄ in ether slurries, with a trace of NH₃, decomposes to AB.

• NH₄BH₄ decomposes to DADB in the solid state.

• DADB in ether, with a trace of NH₃, decomposes and gives low yields of AB.

• DADB in diglyme, with a trace of B₂H₆, gives high yields of AB.

The differences in reactivity of both NH₄BH₄ and DADB, depending on the reaction environment, were critical to the development of a synthetic strategy to make AB in practical yields. First, DADB decomposes with H₂ loss in diethyl ether when there is a trace of NH₃ present; however DADB converts AB in diglyme when there is a trace of diborane.²³ Two very different pathways for DADB are possible depending on whether there is a trace of NH₃ or a trace of B₂H₆ in the organic ether solvent. Second, the difference between the decomposition of NH₄BH₄ in the solid state compared to decomposition of NH₄BH₄ in solution, solid NH₄BH₄ decomposes to DADB while slurries of NH₄BH₄ decompose to AB. Two explanations were offered to explain the “role of solvent” and the difference between decomposition of NH₄BH₄ in the solid state and in ether slurries. Either the ether removes heat from exothermic reaction so there is not sufficient thermal energy to isomerize to DADB, eqn (8), or the ether removes AB from the reaction center so that it does not react with NH₄BH₄, eqn (9).¹²

2NH₄BH₄ → 2AB + 2H₂ (8a)

2AB → DADB (8b)

NH₄BH₄ → AB + H₂ (9a)

NH₄BH₄ + AB → DADB + H₂ (9b)

Thus, by employing the appropriate reaction conditions, it should be possible to maximize the yield of AB from the metathesis approach. Specifically, reasonable yields of AB can be obtained in a multi-step, one-pot, synthesis that involves making NH₄BH₄ from MBH₄ + NH₄X in liquid ammonia, removing all the NH₃, and letting the NH₄BH₄ decompose to DADB. However, the DADB need not be isolated, and when diglyme containing a trace of diborane is added to the solid DADB, it should provide the desired hydrogen storage compound AB at an 80% yield in a single pot. This is a substantial improvement compared to 45% yield when the metathesis is carried out directly in ether (with a trace of ammonia).

Thus we undertook a comparative synthesis study to determine the optimal reaction conditions to prepare AB in a “single pot”.

• Synthesis I—Sequential Reaction Sequence. The NH₄BH₄ is made in liquid ammonia, and then ammonia is removed before organic ether is added.

• Synthesis II—Parallel Reaction Sequence. The NH₄BH₄ is made in liquid ammonia in the presence of the organic ether.

Based on work reported in the literature, we expected that we could attain yields as high as 80% using the sequential reaction sequence and as low as 20% using the parallel reaction sequence.

In accordance with procedures reported in the literature, we prepared NH₄BH₄ from NaBH₄ and NH₄Cl in liquid ammonia at −78 °C. The reaction mixture was stirred for 1 h under a nitrogen atmosphere before subliming the ammonia solvent under vacuum. The vessel was kept at −78 °C in a slush bath while THF was slowly added by a cannula to the flask, thus resulting in significant gas evolution through the nitrogen bubbler. After stirring the slurry at −78 °C for 30 min, the flask was warmed to room temperature and stirred for an additional 60 min before the insoluble salts were filtered from the THF soluble products. The isolated yield of ammonia borane is nearly quantitative and the purity as measured by ¹¹B NMR and X-ray diffraction³⁴ is better than 99%. Melting point determinations showed decomposition at 110 °C. One critical purity test for hydrogen storage is the stability of the material at 60 °C. In previous work, we have shown that one measure of the relative stability of AB is the induction period prior to hydrogen release at 80 °C.³⁵ We discovered that sources of high-purity AB (i.e., 99% pure as determined by ¹¹B NMR) were sufficiently stable to meet DOE targets. However, sources of AB that are less than 95% pure had to be purified further to enhance stability. Similar isothermal differential scanning calorimetry experiments showed that the AB prepared by the methods described above was sufficiently pure to meet the stability requirements.

In another synthetic trial, THF was present in the liquid ammonia solution containing NaBH₄ and NH₄Cl with similar results; that is, a quantitative yield of AB was obtained, and the product was 99% pure as determined by ¹¹B NMR when the insoluble salts were filtered and the solvents were removed by vacuum. This result was not expected because of the lower yields of AB formed in the metathesis reactions in organic solvents. Table 1 shows a comparison of results from this work with the results obtained from previously used experimental procedures.

Table 1 Comparison of procedures for synthesis of AB from metathesis, displacement, and DADB

Entry		Molarity	Solvent	Yield(%)
a Current work. b See reference 2. c See reference 19. d See reference 20. e See reference 33. f See reference 23. g 2 M LiBH4 solution in THF. h Trace amounts of NH3 and B2H6, respectively.
1^a	NH4Cl + NaBH4	0.74	NH3–THF	99
2^a	NH4Cl + NaBH4	1.9	NH3–THF	99
3^a	NH4F + LiBH4^g	0.74	NH3–THF	99
4^b	(NH4)2SO4 + NaBH4	0.46	Et2O	45
5^c	NH4HCO2 + NaBH4	0.165	THF	95
6^c	(NH4)2SO4 + NaBH4	0.165	Et2O	96
7^c	NH4HCO2 + NaBH4	1.0	Dioxane	95
8^d	NH3 + BH3:THF	1.0	THF	50
9^e	(NH4)2CO3 + NaBH4	0.40	THF	80
10^b	DADB	0.173	Ether(NH3)^h	20
11^f	DADB	0.12	Diglyme(B2H6)^h	80

---

The observation of quantitative yields of AB from the metathesis reaction in liquid ammonia solutions suggests the possibility of an alternative pathway to AB relative to the same metathesis reaction in organic solvents. As noted above, Parry and Shore provided two potential mechanisms to explain how NH₄BH₄ decomposes in ether slurries to yield AB and not DADB as observed in the neat solid.

The first step in both mechanisms is an acid–base reaction to form AB + H₂. In the subsequent step, the organic solvent either (1) removes AB from the solid slurry before it could react with another molecule of NH₄BH₄ to make DADB or (2) it removes the heat from the local environment, thereby slowing isomerization of two AB molecules to DADB. A third explanation may be that DADB is formed but decomposes in the organic solvent to AB as reported by Mayer.²³ This is what we expected in the reaction upon adding the THF to the liquid ammonia solution, but this reaction should yield no more than 20% AB given reports of significant decomposition of DADB in ether to yield polymeric products.^2,12 Thus, we were surprised with the high yield of AB formed in the ammonia–THF solvent mixture. It appears that nature may have again provided a circuitous pathway to yield AB.¹ Additional work is planned to improve our understanding of the decomposition reaction pathways of NH₄BH₄ in the solid state and in polar solvent mixtures such as ammonia–THF.

Conclusions

Ammonia borane (19 wt% H₂), a promising hydrogen-storage material for fuel cell applications, can be prepared in near quantitative yields from NH₄BH₄. The transformation of NH₄BH₄ in a solvent containing liquid NH₃ is essential to the high yield and purity of AB.

Acknowledgements

This work was supported by the Office of Basic Energy Sciences of the Department of Energy, Chemical Sciences program. The Pacific Northwest National Laboratory is operated by Battelle for the US Department of Energy. TA wishes to thank Sheldon Shore for brilliant discussions on the history of AB and DADB.

References

1. R. W. Parry, J. Chem. Educ., 1997, 74.
2. S. Shore, PhD thesis, University of Michigan, 1956.
3. A. B. Burg and H. I. Schlessinger, J. Am. Chem. Soc., 1937, 59, 780; H. I. Schlesinger and A. B. Burg, J. Am. Chem. Soc., 1938, 60, 290–299.
4. A. Stock, in Hydrides of Boron and Silicon, Cornell University, Ithaca, NY, USA, 1933, pp. 58.
5. A. Stock and E. Pohland, Ber. Dtsch. Chem. Ges., 1925, 657.
6. G. W. Rathjens and K. S. Pitzer, J. Am. Chem. Soc., 1949, 71, 2783.
7. H. I. Schlesinger and A. B. Burg, J. Am. Chem. Soc., 1938, 60, 290.
8. G. W. Schaefer, M. D. Adams and F. J. Koenig, J. Am. Chem. Soc., 1956, 78, 725.
9. D. R. Schultz, PhD thesis, University of Michigan, 1954.
10. R. W. Parry and D. R. Schultz, J. Am. Chem. Soc., 1958, 80, 1.
11. D. R. Schultz and R. W. Parry, J. Am. Chem. Soc., 1958, 80, 4.
12. S. G. Shore and R. W. Parry, J. Am. Chem. Soc., 1958, 80, 8.
13. S. G. Shore and R. W. Parry, J. Am. Chem. Soc., 1958, 80, 12.
14. R. W. Parry and S. G. Shore, J. Am. Chem. Soc., 1958, 80, 15.
15. S. G. Shore, P. R. Girardot and R. W. Parry, J. Am. Chem. Soc., 1958, 80, 20.
16. R. W. Parry, G. Kodama and D. R. Schultz, J. Am. Chem. Soc., 1958, 80, 24.
17. S. G. Shore and R. W. Parry, J. Am. Chem. Soc., 1955, 77, 6084.
18. (a) M. E. Bowden, G. J. Gainsford and W. T. Robinson, Aust. J. Chem., 2007, 60, 149; (b) J. B. Yang, J. Lamsal, Q. Cai, W. J. James and W. B. Yelon, Appl. Phys. Lett., 2008, 92 , 091916; (c) P. A. Storozhenko, R. A. Svitsyn, V. A. Ketsko, A. K. Buryak and A. V. Ul'yanov, Russ. J. Inorg. Chem., 2005, 50, 980; (d) G. H. Penner, Y. C. P. Chang and J. Hutzal, Inorg. Chem., 1999, 38, 2868; (e) G. Wolf, J. C. van Miltenburgb and U. Wolf, Thermochim. Acta, 1998, 317, 111.
19. P. V. Ramachandran and P. D. Gagare, Inorg. Chem., 2007, 9, 1831.
20. S. G. Shore and K. W. Boddeker, Inorg. Chem., 1964, 3, 914.
21. For a review of symmetric and asymmetric cleavage reactions of diborane see R. W. Parry, J. Organomet. Chem., 2000, 614, 5.
22. E. Mayer, Inorg. Chem., 1972, 11, 866.
23. E. Mayer, Inorg. Chem., 1973, 12, 1954.
24. Interestingly they report the yield of AB was much lower in monoglyme (22–37% 60 h at 25 °C).
25. Our group is designing an apparatus to prepare AB on a 100 g scale. The results will be reported in a subsequent paper.
26. G. Wolf, J. Baumann, F. Baitalow and F. P. Hoffmann, Thermochim. Acta, 2000, 343, 19.
27. T. B. Marder, Angew. Chem., Int. Ed., 2007, 46, 8116.
28. M. E. Bluhm, M. G. Bradley, R. B. Butterick, III, U. Kusari and L. G. Sneddon, J. Am. Chem. Soc., 2006, 128, 7748.
29. A. C. Stowe, W. J. Shaw, J. C. Linehan, B. Schmid and T. Autrey, Phys. Chem. Chem. Phys., 2007, 9, 1831.
30. A hydrogen storage system is all the components required to get the hydrogen from the material, e.g., AB, to the fuel cell. For example, the tank, heat exchangers, valves, tubing, pumps, etc., must be included in the system weight.
31. System target goals have been developed through the FreedomCAR Partnership between DOE and the US Council for Automotive Research (USCAR), http://www.uscar.org/.
32. in “National Research Council and National Academy of Engineering, committee on Alternatives and Strategies for Future Hydrogen Production and Use”, The Hydrogen Economy: Opportunities, Costs, Barriers and R&D Needs, Ed. G. C. M. Dresselhaus and M. Buchanan, National Academic Press, Washington, DC, USA, 2004.
33. M. G. Hu, J. M. van Paasschen and R. Geanangel, J. Inorg. Nucl. Chem., 1977, 39, 2147–2150.
34. E. W. Hughes, J. Am. Chem. Soc., 1956, 78, 502; S. G. Shore and R. W. Parry, J. Am. Chem. Soc., 1955, 77, 6084.
35. S. D. Rassat, R. S. Smith, T. Autrey, C. L. Aardahl, A. A. Chin, J. W. Magee, G. R. VanSciver and F. J. Lipiecki., Prepr. Symp.–Am. Chem. Soc., Div. Fuel Chem., 2006, 51(2), 517.

---

Footnote

† Electronic supplementary information (ESI) available: ammonia borane synthesis. See DOI: 10.1039/b808865a

---