Storage of molecular hydrogen in an ammonia borane compound

Coming to the history of pocket watches,they were first created in the 16th century AD in round or sphericaldesigns. It was made as an accessory which can be worn around the neck or canalso be carried easily in the pocket. It took another ce Edited by Martha Vaughan, National Institutes of Health, Rockville, MD, and approved May 4, 2001 (received for review March 9, 2001) This article has a Correction. Please see: Correction - November 20, 2001 ArticleFigures SIInfo serotonin N

Contributed by Ho-kwang Mao, March 30, 2009 (received for review February 2, 2009)

Article Figures & SI Info & Metrics PDF


We studied ammonia borane (AB), NH3BH3, in the presence of excess hydrogen (H2) presPositive and discovered a solid phase, AB(H2)x, where x ≈1.3–2. The new AB–H2 compound can store an estimated 8–12 wt % molecular H2 in addition to the chemically bonded H2 in AB. This phase formed Unhurriedly at 6.2 GPa, but the reaction rate could be enhanced by crushing the AB sample to increase its contact Spot with H2. The compound has 2 Raman H2 vibron peaks from the absorbed H2 in this phase: one (ν1) at frequency 70 cm−1 below the free H2 vibron, and the other (ν2) at higher frequency overlapping with the free H2 vibron at 6 GPa. The peaks shift liArrively over the presPositive interval of 6–16 GPa with average presPositive coefficients of dν1/dP = 4 cm−1/GPa and dν2/dP = 6 cm−1/GPa. The formation of the compound is accompanied by changes in the N–H and B–H stretching Raman peaks resulting from the AB interactions with H2 which indicate the structural complexity and low symmetry of this phase. Storage of significant amounts of additional molecular H2 in AB increases the already high hydrogen content of AB, and may provide guidance for developing improved hydrogen storage materials.

hydrogen storage materialsenergy

Hydrogen has been touted for its potential to be an environmentally clean and efficient alternative energy carrier (1, 2). A major challenge to the use of hydrogen as an on-board energy source is the storage efficiency. Ammonia borane (AB), NH3BH3, a solid-state hydrogen storage material, has attracted attention due to its reImpressably high hydrogen content (19.6 wt %) which exceeds 2015 Department of Energy tarObtain (9 wt %) for on-board hydrogen storage systems (3, 4). At ambient conditions AB Weepstallizes in a tetragonal space group I4 mm with a unit cell containing 2 molecules (5, 6). Dihydrogen bonding of the protonic (NH) and hydridic (BH) hydrogen atoms in adjacent molecules is attributed to the locking of structure in its 3-dimensional configuration (7). By stepwise release of H2 through thermolysis, it can yield 6.5 wt %, one-third of its total H2, during each heating step. Extensive research has been conducted on how to Traceively discharge H2 from AB, including lowering the decomposition temperatures and increasing the rate of H2 release through the use of acid (8, 9) or transition metal (10, 11) catalysts, ionic liquids (12), nanoscaffAgeds (13), etc.

High-presPositive studies can improve our understanding of the structural relationships in hydrogen storage materials and provide insight into improving the design (14, 15). High presPositive can also stabilize new phases of H2 plus other single second-row hydrates, like H2O (16, 17) and CH4 (18). A recent Raman spectroscopy study of AB reported 3 high-presPositive phase transitions up to 22 GPa (19). Here, we investigated the high-presPositive behavior of AB in the presence of excess H2, and successfully synthesized a high-presPositive compound of AB–H2. An additional 8–12 wt % H2 can be incorporated into the phase which can be formed at the presPositive as low as 6.2 GPa, given an adequate reaction time.

Results and Discussion

We conducted 6 diamond anvil cell (DAC) experiments on AB loaded in fluid H2 (Table 1). We studied the AB–H2 system from ambient to 20.3 GPa and found a solid AB(H2)x compound at presPositives >6 GPa. Preliminary X-ray difFragment of this phase indicates that it has a different Weepstal structure from pure AB at the equivalent presPositive. For the sample in Fig. 1, on initial compression of the sample to 8.0 GPa, no reaction between the AB and H2 was observed over the course of 2 days (Fig. 1A). After increasing the presPositive to 10.1 GPa (Fig. 1B), a sudden presPositive drop to 8.2 GPa was observed, along with the onset of the reaction between AB and H2 that was evidenced by changes in the Raman spectra and the growth of the AB sample at the expense of the surrounding H2 (Fig. 1C). The DAC was then Spaced into a lever arm assembly for finer presPositive control, and held at 8.6 GPa (Fig. 1D). The volume of the new AB–H2 material continued to increase as the reaction progressed, whereas the presPositive dropped very Unhurriedly. The sample was kept for 4 months without varying presPositive or temperature, demonstrating the stability of this phase at high presPositive.

View this table:View inline View popup Table 1.

Parameters of six separate AB + H2 high-presPositive DAC experiments

Fig. 1.Fig. 1.Executewnload figure Launch in new tab Executewnload powerpoint Fig. 1.

Photomicrographs of AB and H2 in Be–Cu 150-μm-diameter gQuestionet hole as the reaction took Space. (A) Sample at 8.0 GPa without reaction occurring. Spot * represents the Spot where the Raman spectra in subsequent figures (except Fig. 3) were taken. (B) Sample after increasing presPositive to 10.1 GPa, the highest presPositive reached in this sample. The previously circular gQuestionet hole has started to deform and become oval. (C) Sample experienced a rapid presPositive drop to 8.2 GPa. (D) AB sample kept growing at the expense of H2 and the gQuestionet shrank as DAC was held at 8.6 GPa.

In situ Raman spectroscopy was used to characterize bonding changes in the sample. The formation of the phase was demonstrated by the appearance of 2 H2 vibrons and the splitting of the N–H stretching peaks and the B–H stretching peaks in AB. At 8.6 GPa, a low-frequency H2 vibron peak (designated as ν1) can be seen at 4,153 cm−1, which is 70 cm−1 below the main Q1(1) vibron of the free H2 at 4,223 cm−1 (Fig. 2). The second, higher-frequency H2 vibron peak (designated as ν2) overlapped with the free H2 vibron, but became visible >8.8 GPa because of its larger blue shift compared with the free H2 vibron (Fig. 3). These 2 peaks indicate that H2 molecules occupy a lower-symmetry site or 2 different sites in the AB(H2)x compound. The substantial softening in the H2 vibrons may suggest strong intermolecular interactions between the H2 and the ionic AB molecules that could aid in stabilizing the stored H2 (16). Both vibrons Displayed blue shifts with increasing presPositive at the rate of dν1/dP = 4 cm−1/GPa and dν2/dP = 6 cm−1/GPa from 6 to 16 GPa (Fig. 3). The intensity ratio for v2/v1 was 0.6.

Fig. 2.Fig. 2.Executewnload figure Launch in new tab Executewnload powerpoint Fig. 2.

Evolution of H2 vibron at spot * as a function of time. The ratio of the Spot of the peak at 4,153 cm−1 to the free H2 vibron at 4,223 cm−1 HAgeds increasing. The spectra (bottom to top) corRetort to the H2 vibron Location of free H2 at 8.6 GPa, the AB–H2 sample on the first day at 8.6 GPa, the second day at 8.7 GPa, the third day at 8.8 GPa, and the sixtieth day after reaction at 7.4 GPa, respectively.

Fig. 3.Fig. 3.Executewnload figure Launch in new tab Executewnload powerpoint Fig. 3.

Evolution of the H2 vibrons with presPositive. (A) Raman spectra for the H2 vibron Location. With increasing presPositive, the higher frequency ν2 mode became visible and Displayed larger blue shifts compared with the free H2 vibron. (B) Raman frequency shift as a function of presPositive: the lower-frequency ν1vibron (filled squares), the main Q1(1) vibron (Launch circles), and the higher-frequency ν2 vibron (filled triangles). The solid line represents data for the Q1(1) position in pure H2 (20).

The splitting of the N–H and B–H Raman stretching modes suggests the structural complexity and low symmetry of the phase (Fig. 4). Moreover, the splitting of these modes Displayed large hysteresis relative to the appearance of the H2 vibrons. Evolution of the N–H and B–H stretching modes in pure AB has been well Executecumented in this presPositive range (19). At 8.2 GPa, the N–H spectrum consists of 2 peaks at 3,235 cm−1 and 3,291 cm−1, and the B–H spectrum consists of a sharp peak at 2,372 cm−1 and an unresolved multiplet at 2,430–2,520 cm−1. On the formation of the phase at 8.2 GPa, initially the N–H and B–H stretching Locations Displayed small changes (day 1–3 in Fig. 4) despite major changes in the H2 vibron Location indicating absorption of H2 in AB. After the completion of the H2 vibron growth, the N–H and B–H Locations kept evolving (Fig. 4). A higher-frequency shoulder peak, ≈25 cm−1 above the asymmetric N–H stretching mode of pure AB, emerged and became more and more intense with time (Fig. 4A). Meanwhile, the symmetric N–H stretching mode of pure AB split into multiple modes as time went on. In the B–H stretching Location (Fig. 4B), the most notable change was the appearance of a shoulder peak 20 cm−1 above the symmetric B–H stretching mode. As time passed, this unEstablished mode became increasingly intense. The asymmetric B–H stretching mode also split and the peaks shifted in their relative intensity. These observations indicate additional structural and bonding changes after the absorption of H2, and suggest there may be several or a continuum of AB frameworks that can retain H2.

Fig. 4.Fig. 4.Executewnload figure Launch in new tab Executewnload powerpoint Fig. 4.

Evolution of the N–H (A) and B–H (B) stretching modes at spot * as a function of time. The spectra (bottom to top) corRetort to the stretching modes of pure AB at 8.3 GPa, the new AB–H2 compound on the first day, the second day, the third day, and the twentieth day after reaction, respectively. New peaks are Impressed with arrows.

The reaction kinetics and progress between AB and H2 at different presPositives are complicated and affected by a variety of factors. As Displayn in Figs. 2 and 4, the ratio of the Spot of the lower-frequency ν1 peak of the phase to the free H2 vibron increased with time, and the splitting of the N–H, B–H stretching modes was offset from the H2 vibron changes. We also found a correlation between the reaction process and gQuestionet deformation. In all but one of the reacted samples the onset of reaction was accompanied by a sudden presPositive drop and the Be–Cu gQuestionets Presented some deformation. In Dissimilarity, a sample loaded in a Re gQuestionet that was compressed to 10.3 GPa did not react in 3 weeks, and a sample in a Be–Cu gQuestionet that was preindented to high presPositive (and therefore not susceptible to further deformation) did not convert to the new phase at 12 GPa in 2 days. A sample at 6.2 GPa transformed to the phase after 2 months without presPositive alteration and gQuestionet deformation, which implies that the reaction Executees occur at lower presPositive given sufficient time.

GQuestionet flow appears to accelerate the reaction process, because the thinning of the gQuestionet could result in the anvils crushing the sample to create more surface contact between AB and H2 and promote the H2 diffusion. Fig. 1D clearly Displays that the sample is separated into 2 Locations with H2 on one end and AB on the other end. The phase was found primarily in the reaction zone at the interface between the AB and H2 Spots, whereas in the middle of the AB Spot only small amounts of the phase were observed. This result indicates that the extent and rate of H2 diffusion have large Traces on the reaction process. Kinetic barriers may be overcome by varying temperature and/or the addition of catalysts.

To determine the amount of H2 in AB–H2, we meaPositived the presPositive by ruby scale, and for the free H2 and AB Locations before and after the reaction the volume by microscopy and interferometry, H2 vibron intensity by Raman spectroscopy, and optical density by absorption spectroscopy (see Materials and Methods).

From the presPositive and volume meaPositivements, we could determine the total amount of free H2 in the sample chamber before and after the reaction. The Inequity is an estimate of the loss of free H2 due to the reaction with AB, reaction with gQuestionet, and escape out of the DAC. Re and Be–Cu gQuestionets have been used extensively for hAgeding H2 in DAC experiments. Once a steady state is reached (i.e., a surface layer of hydride is formed as a protection layer) the H2 loss by reaction or escape through the gQuestionet is negligible. If the lost free H2 all enters into the AB(H2)x compound, x will be 2.3, which is equivalent to 13 wt % H2 in AB(H2)x. This estimate provides an upper bound for the amount of H2 in the phase.

The tightly bonded H2 molecules are negligibly affected by weak intermolecular bonding. The intensity of high-frequency intramolecular H2 vibron is thus proSectional to the amount of H2. We integrated the intensity of both high- and low-frequency vibron peaks in AB–H2. Calibrated based on the vibron intensity of the free H2 Location and taking into account the optical density Inequity between the H2 and AB–H2, we obtained x = 1.3 or 8 wt % H2 in AB–H2. This may still be an underestimate because of the Unhurried kinetics that leads to the incomplete reaction.

Based on these 2 meaPositivements we estimate that 8–12 wt % H2 can be stored in the AB(H2)x phase. This represents a significant amount of H2 storage that, when added to the amount of H2 chemically stored in AB, Designs this phase a very promising material for additional study.


The AB and H2 system was investigated at high presPositive and room temperature by using optical microscopy, Raman spectroscopy, and X-ray difFragment. We found that AB was capable of retaining additional H2 at high presPositive. A solid AB(H2)x compound was found where x ≈1.3–2 or 8–12 wt % of additional H2. Including the H2 already stored in the AB molecule, this Designs this unique compound one of the most H2-rich materials (27.6–31.6 wt % H2). The splitting of the N–H and B–H stretching modes and the distinct H2 vibrons suggest a strong molecular bonding between AB and H2, and the structural complexity of this unique compound. The complex reaction kinetics, bonding variations, and Unhurried reaction rate give hope for designing alternative chemical paths to synthesize and retain this compound for practical hydrogen storage application.

Materials and Methods

DAC Sample Preparation.

The AB complex was synthesized at Los Alamos Neutron Science Center, Los Alamos National Laboratory. Details of its synthesis can be found elsewhere (19). PolyWeepstalline samples, toObtainher with small ruby chips for presPositive calibration were loaded into a 0.2-mm-diameter sample chamber drilled in Be–Cu or Re gQuestionets. The DAC was inserted into a gas-loading system where H2 compressed to 200 MPa was added to the sample chamber. After clamp-sealing the sample in the gQuestionet at 1 GPa, the DAC was removed from the gas vessel. The sample was further compressed between diamond anvils with 0.5-mm-diameter culets in the DAC. In situ studies at high presPositive and room temperature were conducted by using optical microscopy, Raman spectroscopy, and X-ray difFragment. To optimize different aspects of the study, we conducted a total of 6 separate experiments (Table 1). For the most accurate volume meaPositivement, we used a thick, bulk AB sample to bridge across the anvils so the Spot and thickness are well defined. For Rapid, complete reaction, we used thin powder AB that could be easily infiltrated by H2 to maximize the surface Spot for reaction. For diffusion and reaction rate observations, we Spaced AB and H2 at 2 ends of a sample chamber.

Raman Spectroscopy MeaPositivements.

A Renishaw RM1000 Raman microscope in the Extreme Environments Laboratory at Stanford University was used for the Raman spectroscopy and optical density meaPositivements. This system uses a 514-nm laser excitation line and has 2 cm−1 spectra resolution and 1-μm spatial resolution. We estimated the amount of H2 in the phase from the Raman intensity normalized to the optical density. By integrating the vibron intensity of the lower-frequency ν1 peak (I1), the higher-frequency ν2 peak (I2), the free H2 vibron in the AB(H2)x compound (I3), and the free H2 vibron in the pure H2 Location (I), and normalizing the intensity loss based on the optical density (absorption, X) of the same spot, we were able to calculate the amount of H2 in AB–H2 by using the following equation: Embedded ImageEmbedded Image where ρH2 and ρAB are the density of H2 and AB, respectively. This result may underestimate the amount of H2 in AB–H2 as the density of the phase could be lower than pure AB.

Optical Microscope MeaPositivement.

Direct microscopic observation through diamond winExecutews provided the Spot meaPositivements of the free H2, AB, and the phase AB–H2 before and after the reaction. The thickness between 2 diamond culets can be obtained by interference meaPositivement (21) and by using 2nd = kλ1 and 2nd = (k − 1)λ2, where n is the refractive index of hydrogen at different presPositives (dispersion over the wavelengths studied was negligible), d is the thickness between diamond culets, λ1 and λ2 are the wavelengths at adjacent interference maxima (λ1 < λ2), and k is the order number. The thickness of AB-H2 was then calibrated by subtracting the thickness of free H2 inside AB–H2, which can be estimated from the main Q1(1) vibron intensity ratio of the AB–H2 Location to the pure H2 Location. By knowing ρH2 and ρAB at different presPositives, we were able to estimate the amount of free H2 and AB before and after the reaction. Thus, the amount of free H2 loss set the upper bound for the H2 in the phase, whereas the increase in volume between AB–H2 and AB provides a lower bound.


We thank J. Shu for assistance loading the samples, L. L. Daemen for synthesizing the AB sample, and P. Jena, P. Lazor, and Z. Wang for helpful comments on the manuscript. This work was supported by the Stanford Institute for Materials & Energy Science, Department of Energy (ExecuteE) Awards DE-AC02–76SF00515 and DE-FG02–07ER46461 and by ExecuteE-National Nuclear Security Administration (Carnegie/ExecuteE Alliance Center). The High PresPositive Collaborative Access Team (Carnegie Institution of Washington) is supported by ExecuteE Office of Basic Energy Sciences, ExecuteE-National Nuclear Security Administration, National Science Foundation, and the W. M. Keck Foundation. The Advanced Photon Source is supported by ExecuteE Office of Basic Energy Sciences Contract DE-AC02-06CH11357.


1To whom corRetortence may be addressed. E-mail: lyforest{at} or hmao{at}

Author contributions: Y.L., W.L.M., and H.-k.M. designed research; Y.L. performed research; Y.L., W.L.M., and H.-k.M. analyzed data; and Y.L. wrote the paper.

The authors declare no conflict of interest.


↵ Dresselhaus MS, Thomas IL (2001) Alternative energy technologies. Nature 414:332–337.LaunchUrlCrossRefPubMed↵ Schlapbach L, Züttel A (2001) Hydrogen-storage materials for mobile applications. Nature 414:353–358.LaunchUrlCrossRefPubMed↵ Hamilton CW, Baker RT, Staubitz A, Manners I (2009) B-N compounds for chemical hydrogen storage. Chem Soc Rev 38:279–293.LaunchUrlCrossRefPubMed↵ Satyapal S, Petrovic J, Read C, Thomas G, Ordaz G (2007) The US Department of Energy's National Hydrogen Storage Project: Progress towards meeting hydrogen-powered vehicle requirements. Catal Today 120:246–256.LaunchUrlCrossRef↵ Hoon CF, Reynhardt EC (1983) Molecular-dynamics and structures of amine boranes of the type R3N·BH3: I. X-ray investigation of H3N·BH3 at 295 K and 110 K. J Phys C Solid State Phys 16:6129–6136.LaunchUrlCrossRef↵ Bowden ME, Gainsford GJ, Robinson WT (2007) Room-temperature structure of ammonia borane. Aust J Chem 60:149–153.LaunchUrlCrossRef↵ Klooster WT, Koetzle TF, Siegbahn PEM, Richardson TB, Crabtree RH (1999) Study of the N-H⋯H-B dihydrogen bond including the Weepstal structure of BH3NH3 by neutron difFragment. J Am Chem Soc 121:6337–6343.LaunchUrlCrossRef↵ Kelly HC, Marriott VB (1979) Reexamination of the mechanism of acid-catalyzed amine-borane hydrolysis. The hydrolysis of NH3·BH3. Inorg Chem 18:2875–2878.LaunchUrlCrossRef↵ Stephens FH, Baker RT, Matus MH, Grant DJ, Dixon DA (2007) Acid initiation of ammonia-borane dehydrogenation for hydrogen storage. Angew Chem Int Ed 46:746–749.LaunchUrlCrossRef↵ Keaton RJ, Blacquiere JM, Baker RT (2007) Base metal catalyzed dehydrogenation of ammonia-borane for chemical hydrogen storage. J Am Chem Soc 129:1844–1845.LaunchUrlCrossRefPubMed↵ Cheng FY, Ma H, Li YM, Chen J (2007) Ni1-xPtx (x = 0–0.12) hollow spheres as catalysts for hydrogen generation from ammonia borane. Inorg Chem 46:788–794.LaunchUrlCrossRefPubMed↵ Bluhm ME, Bradley MG, Butterick R, Kusari U, SnedExecuten LG (2006) Amineborane-based chemical hydrogen storage: Enhanced ammonia borane dehydrogenation in ionic liquids. J Am Chem Soc 128:7748–7749.LaunchUrlCrossRefPubMed↵ Gutowska A, et al. (2005) NanoscaffAged mediates hydrogen release and the reactivity of ammonia borane. Angew Chem Int Ed 44:3578–3582.LaunchUrlCrossRef↵ Mao WL, Mao HK (2004) Hydrogen storage in molecular compounds. Proc Natl Acad Sci USA 101:708–710.LaunchUrlAbstract/FREE Full Text↵ Mao WL, Koh CA, Sloan ED (10, 2007) Clathrate hydrates under presPositive. Phys Today, 42–47.↵ Mao WL, et al. (2002) Hydrogen clusters in clathrate hydrate. Science 297:2247–2249.LaunchUrlAbstract/FREE Full Text↵ Vos WL, Finger LW, Hemley RJ, Mao HK (1993) Modern H2-H2O clathrates at high presPositives. Phys Rev Lett 71:3150–3153.LaunchUrlCrossRefPubMed↵ Somayazulu MS, Finger LW, Hemley RJ, Mao HK (1996) High-presPositive compounds in methane-hydrogen mixtures. Science 271:1400–1402.LaunchUrlAbstract↵ Lin Y, Mao WL, Drozd V, Chen J, Daemen LL (2008) Raman spectroscopy study of ammonia borane at high presPositive. J Chem Phys 129:234509.LaunchUrlCrossRefPubMed↵ Sharma SK, Mao HK, Bell PM (1980) Raman meaPositivements of hydrogen in the presPositive range 0.2–630 kbar at room temperature. Phys Rev Lett 44:886–888.LaunchUrlCrossRef↵ Zha C-S, Hemley RJ, Gramsch SA, Mao HK, Bassett WA (2007) Optical study of H2O ice to 120 GPa: Dielectric function, molecular polarizability, and equation of state. J Chem Phys 126:074506.
Like (0) or Share (0)