Beneficial effects of stoichiometry and nanostructure for a LiBH₄–MgH₂ hydrogen storage system

Abstract

The hydrogen storage system [MgH₂–2LiBH₄] shows attractive properties such as favorable thermodynamics, high hydrogen capacity and reversibility. However, there exists an incubation period that amounts up to 10 hours in the dehydrogenation steps, which restricts this system as a practical material. In this study, the influences of stoichiometry and the nanoscale MgH₂ were investigated for the system. Considerably shortened incubation times were achieved with deficit amounts of LiBH₄ or by using nanoscale MgH₂. In addition, the application of nanoscale MgH₂ prevented or suppressed the formation of [B₁₂H₁₂]²⁻ in the dehydrogenation, which is otherwise an issue concerning the re-cyclability.

---

Introduction

Efficient storage of hydrogen as an energy carrier has been proved to be a bottleneck factor in the implementation of this clean energy in the past decades. Mobile application of hydrogen storage materials demands both high capacity and appropriate thermodynamics.^1,2

Borohydrides are a category of compounds that have high H contents. However, the chemical stability of B–H appears to be critical for the application as a hydrogen storage medium. While some metal borohydrides are very unstable and ready to decompose to extremely toxic diborane,³ others are so stable that they only release hydrogen at impractically high temperatures.⁴ For tuning of the thermodynamics there are currently two strategies. One is by forming borohydrides with mixed cations⁵ and the other is to form stabilized desorption products by alloying.⁶

Vajo et al. reported in 2004 an approach for destabilizing LiH and MgH₂ by alloy formation during the dehydrogenation process.⁶ The temperature of 1 bar equilibrium pressure of H₂ could be dramatically lowered in this way. The same research group extended this methodology to lithium borohydride and developed a hydrogen storage system based on solid state reactions (eqn (1)), which were later also investigated by other research groups.^7–10 With a theoretical capacity of 11.4% and thermodynamics superior to each single component in the system, this material has met the DOE requirements in a few critical points and thus attracted widespread interest.

MgH₂ + 2LiBH₄ → MgB₂ + 2LiH + 4H₂ (1)

Nevertheless, the capacities reported by different groups were much lower than the theoretical value.^8,11 In addition, an incubation period longer than a dozen of hours was observed during the hydrogen desorption process, which degrades the applicability of the material.

While transition metal compounds as additives are effective to reduce the incubation time,^11,12 the capacity is, however, lowered as well. To tackle the above-mentioned issues, we investigated the effects of the stoichiometry between MgH₂ and LiBH₄ in eqn (1) on the hydrogen desorption behavior. In addition, magnesium hydride synthesized via an organometallic approach¹³ that showed high reactivity as a single component was used for an enhanced kinetics.

Experimental

Sample preparation: LiBH₄ (95%) and MgH₂ (98%) from Alfa Aesar were stored in a glovebox under an Ar atmosphere and used as received. Nanostructured MgH₂ (denoted as *MgH₂) was synthesized via an organometallic method described in detail elsewhere.¹³ For the ball milling process, the starting chemicals were loaded into a milling vessel inside the glovebox at a certain ratio of MgH₂ to LiBH₄. Ball milling was performed using a Retsch PM 400 at 200 rpm and a ball to powder ratio of 80 : 1 for 20 h.

Fourier Transform Infrared Spectroscopy (FTIR) measurements were carried out using a Perkin Elmer Spectrum GX FT-IR system. The powdery samples were mixed with KBr and pressed into pellets and measured at a resolution of 4 cm⁻¹.

Identification of crystal phases was performed by X-ray diffractometry (XRD) on a Philips X'PERT diffractometer (Cu Kα radiation). In order to protect samples against air, a sample holder consisting of a Kapton foil hood and a silicon single crystal base was used. Powdery samples were spread evenly onto the silicon crystal and then sealed with the Kapton foil hood inside the glovebox. The measurement was conducted in the 2θ range between 10° and 80° at a step length of 0.02°.

Differential Scanning Calorimetry (DSC) measurements were performed on a Netzsch DSC 204 HP housed inside a glovebox to detect heat effects accompanying hydrogen desorption. Samples were heated at 5 °C min⁻¹ under 3 bar He or H₂.

For the quantification of H desorption, a home-made Sieverts' system was used to measure the H₂ amount evolved from the materials upon temperature elevation. The sample temperature was raised first at 5 °C min⁻¹ to 400 °C and kept at this temperature for a certain period. To enable a thermodynamically permitted formation of MgB₂,¹⁴ a hydrogen back pressure of 3 bar was applied during the dehydrogenation process.

A TG-DSC-MS (Sensys Evo, Setaram) was housed inside a glovebox with an Ar atmosphere. Approximately 5–8 milligram sample was loaded into the crucible of alumina. Measurement was carried out under He flow of 20 ml min⁻¹ at a heating rate of 5 °C min⁻¹.

A Bruker 600 MHz Avance system with a 4 mm MAS probe head was used for solid state ¹¹B NMR experiments. Echo spectra were recorded at room temperature at a resonance frequency of 192.6 MHz, magic angle spinning speed of 12.5 kHz, and 90/180°-pulses of 2.5/5.0 μs and 6.5/12.0 μs. Na (¹¹BH₄) was used for referencing.

Results and discussion

Influence of stoichiometry on the dehydrogenation

The stoichiometry of the mixture [MgH₂ + xLiBH₄] was varied at x = 0.2, 0.6, 1.0 and 2.0. The results of thermal desorption experiments are shown in Fig. 1. Depending on the ball milling conditions, an incubation period in the range from 10 to 15 h was observed for the composition at x = 2.0 (MgH₂ + 2LiBH₄) by Bösenberg et al.^10,11 An incubation period of more than 10 h for x = 2.0 between the desorption steps of MgH₂ and LiBH₄ was also observed in our experiments. A dramatic decrease in the incubation periods could be achieved at lower values of x or lower concentrations of LiBH₄. At x = 0.2, the incubation almost disappeared. The total amount of desorbed hydrogen, however, decreased at lower x values.

Fig. 1 Dehydrogenation isotherms at 400 °C and 3 bar H₂ for samples [MgH₂ + xLiBH₄] (x = 0.2, 0.6, 1.0 and 2.0).

With varying compositions, the dehydrogenation reaction in eqn (1) can be rewritten as follows:

MgH₂ + xLiBH₄ → x/2MgB₂ + (1 − x/2)Mg + xLiH + (2 + 3x)/2H₂ (2)

The theoretical dehydrogenation amounts were calculated using the above equation and listed in Table 1. Since the H content of LiBH₄ is higher than that of MgH₂, higher H release was expected for LiBH₄-rich compositions, which was proved by the experimental results, i.e., more hydrogen could be released with increased LiBH₄ content. The desorption yield, in contrast, decreased for LiBH₄-rich compositions, indicating that dehydrogenation was more efficient at lower x.

Table 1 Calculated and measured H amount from [MgH₂ + xLiBH₄]

x	Mol. mass	H amount from MgH2 (%)	H wt% cal.	H wt% found	Desorption yield (%)
0.2	30.7	6.51	8.47	7.9	93.3
0.6	39.5	5.06	9.62	8.6	89.4
1.0	48.3	4.14	10.35	9.0	87.0
2.0	70.3	2.84	11.38	9.6	84.4

---

Shown in Fig. 2 are the DSC curves for samples with various x values. Typical thermal events were displayed at x = 2. The weak peak a little bit above 100 °C (T₁) was attributed to the crystal structure transition of LiBH₄, whereas the endothermic event just below 300 °C (T₂) resulted from the melting of LiBH₄. The strong endothermic peak next to T₂ was attributed to the dehydrogenation of MgH₂, followed by a broad peak arising from the decomposition of LiBH₄.^11,15 Therefore, the dehydrogenation of LiBH₄ occurred after the melting of LiBH₄ in the presence of a solid Mg phase, which means that the second dehydrogenation step took place heterogeneously. This phenomenon was observed using SEM/EDX analysis by Deprez et al. that the LiBH₄ phase was highly dispersed and partially covering MgH₂ particles.¹⁶

Fig. 2 DSC curves of samples [MgH₂ + xLiBH₄] (x = 0.2, 0.6, 1.0 and 2.0).

In a liquid–solid reaction, the specific area of the solid component is a determining factor for the kinetics. As we can see, the dehydrogenation peak of MgH₂ shifted to the lower temperature with lower x, suggesting that MgH₂ became more reactive or destabilized. Interestingly, the decomposition of LiBH₄ (the broad peak) shifted concurrently to lower temperature. Since the newly formed Mg metal from the decomposition of MgH₂ was immersed in the liquid state LiBH₄, the Mg particles were covered by a liquid layer of LiBH₄,¹⁶ and the coalescence of Mg particles was presumably hindered. As a consequence, reactivity of Mg metal could be preserved, which facilitated its reaction with LiBH₄ to form MgB₂. For Mg-rich compositions (lower x), the lowering of LiBH₄ decomposition suggests a shortened incubation period, which is consistent with the observation by the Sieverts' method (Fig. 1). The reason for a shortened incubation may be twofold. First, in the Mg-rich samples, more Mg particles provide more reactive surface and consequently faster kinetics, given that freshly formed Mg metal particles from different compositions have similar particle size. Using the bulk densities of LiBH₄ and MgH₂ (0.67 and 1.45 g cm⁻³, respectively), the volume ratio of the two starting components varies from 3.65 to 0.365 for the x range from 2.0 to 0.2. If using the density of Mg (1.74 g cm⁻³) instead of MgH₂, the ratio becomes 4.70 to 0.47. Although in reality this volume ratio might shift to a lower value due to the pulverization of MgH₂, this simple calculation enables to envision the substantial change in the reaction conditions in terms of the relationship between liquid reactant LiBH₄ and surface area of Mg metal. Secondly, with Mg-rich samples, ball milling efficiency increased, which can be inferred from the XRD results (ESI†). As a matter of fact, crystal grain sizes of MgH₂ decreased with increased content of MgH₂ in the composition, which is further in favor of faster kinetics in the dehydrogenation from the reaction between liquid LiBH₄ and Mg.

At temperatures above that of LiBH₄ decomposition, several heat events were still detectable, indicating further reactions at higher temperatures, which are illustrated more clearly by TGA curves (Fig. 3). Weight loss proceeded stepwise and went up to the temperature limit of the instrument at 600 °C, which suggests that the dehydrogenation of LiBH₄ is rather complex. For samples with lower x values, the weight loss took place at lower temperature as well. Similar to the Sieverts' results, however, none of the compositions led to a hydrogen amount close to the theoretical values by TGA measurements, which will be clarified by the results of XRD and solid-state ¹¹B NMR in the following sections.

Fig. 3 TGA curves of samples [MgH₂ + xLiBH₄] (x = 0.2, 0.6, 1.0 and 2.0).

Investigations on this two-component system were dominantly performed at a 1 : 2 molar ratio of MgH₂ and LiBH₄. Nevertheless, there have been a few studies dedicated to the stoichiometric effect on the dehydrogenation performance.^17–19 Price et al. found diverse reaction pathways and kinetics under various molar ratios. The compositions equivalent to the x values between 0.11 and 2.0 were studied by Walker et al.¹⁹ They found a decrease in the peak temperature of LiBH₄ decomposition, which agrees partially with our results. Nevertheless, no obvious change in the dehydrogenation temperature of MgH₂ was observed. Although a three-step desorption was observed, the temperature for MgH₂ decomposition and the onset temperature of weight loss during DTG-MS measurements were independent of the stoichiometry, which differs from our observation described above. Comparing the conditions for sample preparation, we noted that the milling duration and energy were much higher in our case than those used by Walker et al. The particle size of MgH₂ induced by ball milling may not be significantly reduced under the applied conditions in ref. 19.

Effect of nanoscale *MgH₂ as a starting component

*MgH₂ synthesized via an organometallic method exhibited much enhanced dehydrogenation. With a specific area of 108 m² g⁻¹, the dehydrogenated material absorbed hydrogen 40 times faster than its commercial counterpart.²⁰ The effectiveness of *MgH₂ in the MgH₂–LiBH₄ system was tested by replacing commercial MgH₂ with nanostructured *MgH₂ at x = 1.0 and 2.0.

Shown in Fig. 4 are the DSC measurements for the compositions at x = 2 with commercial MgH₂ and *MgH₂, under the atmosphere of 3 bar He (Fig. 4A) and 3 bar H₂ (Fig. 4B), respectively. As expected, neither atmosphere nor *MgH₂ induced any effect on the melting of LiBH₄ at about 290 °C, as indicated by the dotted blue line. However, dehydrogenation behaviors differed from each other under various atmospheres. Under a He atmosphere, the peak temperature of *MgH₂ decomposition appeared at a lower position by 25 °C than that of commercial MgH₂ (Fig. 4A), suggesting a higher reactivity of *MgH₂ than that of commercial MgH₂. Dehydrogenation of LiBH₄ took place at lower temperature as well for the sample with *MgH₂ than with commercial MgH₂. Using H₂ as the atmosphere, dehydrogenation reaction was suppressed, and the dehydrogenation shifted to higher temperature. Nevertheless, *MgH₂ decomposed still at lower temperature than commercial MgH₂ by about 5 °C, which could be an indication of improved thermodynamics for *MgH₂, which was not identified by means of van't Hoff plot earlier by Shao et al.²¹

Fig. 4 DSC curves for the samples with commercialized MgH₂ (black) and *MgH₂ (red) at x = 2.

The effectiveness of *MgH₂ is reflected clearly in the desorption isotherms shown in Fig. 5. Both *MgH₂–containing samples showed faster dehydrogenation than the compositions using commercial MgH₂. The incubation times were approximately halved compared to the respective samples with commercial MgH₂.

Fig. 5 Dehydrogenation isotherms at 400 °C and 3 bar H₂: square symbols for x = 2, triangle symbols for x = 1; black for commercial MgH₂ and blue for nanoscale *MgH₂.

Zeng et al. prepared reactive MgH₂ or MgD₂ by doping with Nb₂O₅ for the MgH₂–LiBH₄ system.²¹ It was notable that an H–D exchange was observed in the MgD₂–LiBH₄ mixture below the melting point of LiBH₄. In addition, MgH₂ or MgD₂ doped with Nb₂O₅ exhibited peak desorption temperatures which was over 100 °C lower than *MgH₂ used in our study (ESI†). Nevertheless, the Nb₂O₅ doped components lost their dehydrogenation reactivity in the composites, in contrast to the effect of *MgH₂ observed in this study, which means that the activation method for MgH₂ is also of significance. With the organometallic synthesized MgH₂, the material with nanoscale particle size has both high surface area and interfacial contact with other reactants, and therefore benefits enhanced reaction kinetics.⁸ For the Nb₂O₅ doped MgH₂, however, the increased reactivity originates from the interaction between the oxygen atom of the oxide additive and the hydrogen atom of hydride,^22,23 which may be weakened once in contact with a liquid phase, like LiBH₄ in this case.

The Sieverts' measurements for the composition at x = 2 were interrupted at the 13^th hour and XRD analysis was conducted (Fig. 6). For the sample with commercial MgH₂, both starting chemicals and desorbed products were present after 13 h desorption. In contrast, only targeted desorption products of MgB₂ and LiH were found in the sample comprising *MgH₂, supporting the observation of an enhanced dehydrogenation with the reactive *MgH₂. Obviously, nanoscale MgH₂ played a significant role in enhancing the whole desorption kinetics.

Fig. 6 XRD patterns after 12 h dehydrogenation at 400 °C and 3 bar H₂ for samples with commercial MgH₂ and *MgH₂ at x = 2, respectively.

Formation of [B₁₂H₁₂]²⁻ in the desorption products

Metal borohydrides, LiBH₄ and Mg(BH₄)₂ in particular, are viewed as promising compounds for hydrogen storage because of their high H contents in the molecules. However, closo-borane structures of [B₁₂H₁₂]²⁻ were identified in the desorption products from compounds like LiBH₄ or Mg(BH₄)₂, as well as from the mixture MgH₂–2LiBH₄. Due to the high stability, [B₁₂H₁₂]²⁻ is recognized as a thermodynamic sink, which impairs the reversibility of hydrogen storage materials.^24,25 Whereas Yan et al. observed the formation of the closo-borane structure via a gradual evolution starting from [BH₄]⁻via [B₂H₆]²⁻, [B₅H₉]²⁻ and [B₅H₈]⁻ to [B₁₂H₁₂]²⁻,²⁶ Hwang et al. found that the formation of a closo-borane structure was a universal pathway in the MgH₂–2LiBH₄ system, regardless of the kinds of additives added to the system.²⁷

In order to reveal the stoichiometric effect on the dehydrogenation reaction, the desorbed products of the samples with commercial MgH₂ at x = 1.0 and 2.0 were analyzed by the solid state ¹¹B NMR after isotherm desorption at 3 bar H₂ and 400 °C for 20 h (Fig. 7). The resonance at 104 ppm appeared in both samples, indicating the formation of the targeted desorption product MgB₂. In addition, the typical resonance at about −12 ppm suggests the presence of boron nuclei of [B₁₂H₁₂]²⁻. A big difference appears at −40 ppm, where the boron nuclei in the anions of tetrahydroboride [BH₄]⁻ was detected for x = 2.0 but absent for x = 1.0 after dehydrogenation. The consumption of the starting component LiBH₄ was obviously not complete for x = 2.0, which resulted in the lower H yield for the LiBH₄-rich samples in Table 1.

Fig. 7 ¹¹B NMR for samples with x = 1.0 (red) and 2.0 (black).

Substituting commercial MgH₂ with *MgH₂ at x = 2, it is interesting to note that no [B₁₂H₁₂]²⁻ was found in the desorbed products from the sample prepared with highly reactive *MgH₂, though unconsumed LiBH₄ and target compound MgB₂ still co-existed after 20 h desorption time (Fig. 8). Theoretical studies show that the formation of [B₁₂H₁₂]²⁻ is thermodynamically favored in the metal borohydride systems during dehydrogenation.²⁸ Pinkerton et al. found that a better reversibility could be achieved by applying a minimal H₂ backpressure during the dehydrogenation process.¹⁴ They recognized that a suppressed decomposition of LiBH₄ is crucial for reversibility. Yan found later on that even at a H₂ pressure up to 10 bar,²⁹ formation of Li₂B₁₂H₁₂ still took place. This is obviously not the case in this study using *MgH₂. The formation of thermodynamically stable species was effectively avoided or at least suppressed, which may hint a different reaction mechanism between the two components under kinetic control, which is worth further investigating. A report by Yu et al. revealed a new mechanism for the MgH₂–LiBH₄ composition that differs from the mechanism of MgB₂ alloying.¹⁸ At temperatures higher than 440 °C, dehydrogenation under vacuum resulted in the formation of Li–Mg alloys, which could reversibly absorb H₂ quantitatively. The reaction pathway with *MgH₂ in this study seems different from the mechanism proposed in ref. 18.

Fig. 8 ¹¹B NMR for samples with x = 2.0 with commercial MgH₂ (black) and *MgH₂ (red).

Dehydrogenation sequence

Typically, hydrogen desorption of the LiBH₄–MgH₂ system takes place in two steps, as is shown in Fig. 9 for x = 1.0. The first step concerning dehydrogenation of MgH₂ is fast, followed by a sluggish hydrogen release step. In order to understand the chemical progression in the dehydrogenation process, sampling of MgH₂–LiBH₄ (x = 1) at three dehydrogenation stages was conducted for analyses using FTIR and XRD. At the respective points, heating was terminated, and at the same time, the hydrogen back pressure was released from the reactor. The first sample was chosen between the two major desorption steps (s1). The second point was in the middle of the 2^nd desorption step (s2), followed by the third point at the virtual end of the 2^nd desorption (s3), as indicated in Fig. 9.

Fig. 9 Sampling positions for the composition of MgH₂–LiBH₄.

Attempts were made by Bösenberg et al. to investigate the reaction pathways by performing the dehydrogenation at various temperatures and H₂ pressures.¹⁵ The results excluded a simultaneous decomposition of the two components and formation of LiH/MgB₂.

As shown above in this study, the system at x = 1 exhibited a faster desorption kinetics than that at x = 2. According to our XRD results (Fig. 10), Mg, MgH₂ and LiBH₄ were available in all the three samples s1, s2 and s3. However, a very limited amount of MgB₂ could be detected as a target desorption product even at s3 in the stage 3, despite the fact that hydrogen release ceased substantially at this point. The unambiguous presence of MgB₂ along with the disappearance of the LiBH₄ phase was observed only afterwards under the isothermal conditions for a certain duration (20 h in this test), which agrees with the observation in ref. 16 and indicates that most MgB₂ was formed after dehydrogenation of LiBH₄. This means that the formation of MgB₂ may simply be from the alloying between Mg and B metals, instead of via the reaction suggested in eqn (3) by Vajo et al.⁷

LiBH₄ + 1/2Mg → LiH + 1/2MgB₂ + 3/2H₂ (3)

Fig. 10 XRD patterns of the sample at x = 1.0 at different desorption stages.

Another interpretation might be that MgB₂ was present at first in the amorphous state, crystallizing fairly slowly and thus undetectable by XRD.

This finding is endorsed by the FTIR investigation on the same samples (Fig. 11). The absorbance originating from deformation vibration of H–B–H in the ranges of 1050–1150 cm⁻¹ and the bridging B–H stretching in the range 2160–2460 cm⁻¹, respectively, were detectable for all the 3 stages, indicating the availability of the [BH₄]⁻ moiety. At the end of the measurement, no bridging absorbance from B–H stretching vibration in the range 2160–2460 cm⁻¹ was observed, which agrees with the NMR and XRD observations that [BH₄]⁻ was absent after desorption measurement.

Fig. 11 FTIR of the sample at x = 1.0 at different desorption stages.

Conclusions

Reduced content of LiBH₄ in relation to [MgH₂–2LiBH₄] favored an enhanced dehydrogenation and increased the yield of H desorption. The introduction of nanostructured magnesium hydride *MgH₂ in place of commercial MgH₂ not only led to a faster kinetics, but also altered the reaction pathway of dehydrogenation.

Acknowledgements

Financial supports by the German Ministry of Economy in the framework of the German-Chinese Stainable Fuel Partnership (grant # 03BV108A) and European project Fuel Cell Coupled Solid State Hydrogen Storage Tank (SSH2S) (grant agreement # 256653) are gratefully acknowledged.

Notes and references

1. L. Schlapbach and A. Züttel, Nature, 2001, 414, 353–358.
2. B. Bogdanovic and M. Schwickardi, J. Alloys Compd., 1997, 253–254, 1–9.
3. O. Friedrichs, A. Borgschulte, S. Kato, F. Buchter, R. Gremaud, A. Remhof and A. Züttel, Chem.–Eur. J., 2009, 15, 5531–5534.
4. W. Grochala and P. P. Edwards, Chem. Rev., 2004, 104, 1283–1316.
5. D. Ravnsbaek, Y. Filinchuk, Y. Cerenius, H. J. Jakobsen, F. Besenbacher, J. Skibsted and T. R. Jensen, Angew. Chem., Int. Ed., 2009, 48, 6659–6663.
6. J. J. Vajo, F. Mertens, C. C. Ahn, R. C. Bowman and B. Fultz, J. Phys. Chem. B, 2004, 108, 13977–13983.
7. J. J. Vajo, S. L. Skeith and F. Mertens, J. Phys. Chem. B, 2005, 109, 3719–3722.
8. J. J. Vajo, T. T. Salguero, A. F. Gross, S. L. Skeith and G. L. Olson, J. Alloys Compd., 2007, 446–447, 409–414.
9. M. Dornheim, S. Doppiu, G. Barkhordarian, U. Boesenberg, T. Klassen, O. Gutfleisch and R. Bormann, Scr. Mater., 2007, 56, 841–846.
10. U. Bösenberg, S. Doppiu, L. Mosegaard, G. Barkhordarian, N. Eigen, A. Borgschulte, T. Jensen, Y. Cerenius, O. Gutfleisch and T. Klassen, Acta Mater., 2007, 55, 3951–3958.
11. U. Bösenberg, S. Doppiu, L. Mosegaard, G. Barkhordarian, N. Eigen, A. Borgschulte, T. R. Jensen, Y. Cerenius, O. Gutfleisch, T. Klassen, M. Dornheim and R. Bormann, Acta Mater., 2010, 58, 3381–3389.
12. N. Boucharat, D. Wang, E. G. Bardají, M. Fichtner and W. Lohstroh, J. Phys. Chem. C, 2012, 116, 11877–11885.
13. B. Bogdanovic, S. Liao, M. Schwickardi, P. Sikorsky and B. Spliethoff, Angew. Chem., Int. Ed. Engl., 1980, 19, 818–819.
14. F. E. Pinkerton, M. S. Meyer, G. P. Meisner, M. P. Balogh and J. J. Vajo, J. Phys. Chem. C, 2007, 111, 12881–12885.
15. U. Bosenberg, D. B. Ravnsbæk, H. Hagemann, V. D. Anna, C. B. Minella, C. Pistidda, W. Van Beek, T. R. Jensen and M. Dornheim, J. Phys. Chem. C, 2010, 114, 15212–15217.
16. E. Deprez, M. A. Muñoz-Márquez, M. C. Jimenez de Haro, F. J. Palomares, F. Soria, M. Dornheim, R. Bormann and A. Fernández, J. Appl. Phys., 2011, 109, 014913.
17. T. E. C. Price, D. M. Grant, V. Legrand and G. S. Walker, Int. J. Hydrogen Energy, 2010, 35, 4154–4161.
18. X. B. Yu, D. M. Grant and G. S. Walker, Chem. Commun., 2006, 3906–3908.
19. G. S. Walker, D. M. Grant, T. C. Price, X. Yu and V. Legrand, J. Power Sources, 2009, 194, 1128–1134.
20. H. Shao, M. Felderhoff, F. Schüth and C. Weidenthaler, Nanotechnology, 2011, 22, 235401.
21. L. Zeng, H. Miyaoka, T. Ichikawa and Y. Kojima, J. Phys. Chem. C, 2010, 114, 13132–13135.
22. W. Oelerich, T. Klassen and R. Bormann, J. Alloys Compd., 2001, 315, 237–242.
23. W. Oelerich, T. Klassen and R. Bormann, Adv. Eng. Mater., 2001, 3, 487–490.
24. N. Ohba, K. Miwa, M. Aoki, T. Noritake, S. Towata and A. Züttel, Phys. Rev. B: Condens. Matter Mater. Phys., 2006, 74, 1–7.
25. H.-W. Li, K. Miwa, N. Ohba, T. Fujita, T. Sato, Y. Yan, S. Towata, M. W. Chen and S. Orimo, Nanotechnology, 2009, 20, 204013.
26. Y. Yan, H.-W. Li, H. Maekawa, M. Aoki, T. Noritake, M. Matsumoto, K. Miwa, S. Towata and S. Orimo, Mater. Trans., 2011, 52, 1443–1446.
27. S. Hwang, R. C. Bowman, J. J. W. Reiter, J. Rijssenbeek, G. L. Soloveichik, J. Zhao, H. Kabbour and C. C. Ahn, J. Phys. Chem. C, 2008, 112, 3164–3169.
28. V. Ozolins, E. H. Majzoub and C. Wolverton, J. Am. Chem. Soc., 2009, 131, 230–237.
29. Y. Yan, H. Li, H. Maekawa, K. Miwa, S. Towata and S. Orimo, J. Phys. Chem. C, 2011, 115, 19419–19423.

---

Footnote

† Electronic supplementary information (ESI) available: DSC curves of nanoscale MgH₂, XRD profiles of as-milled samples from x = 0.2–2.0. See DOI: 10.1039/c3ta13775a

---