Identification of the Dehydrogenated Product of Ca(BH4)2 - The

Mar 13, 2009 - ... Thomas Klassen , Rüdiger Bormann , and Martin Dornheim .... A. Léon , F. Juranyi , J. Wuttke , B. C. Hauback , M. Fichtner , and ...
0 downloads 0 Views 1MB Size
Download PDF
J. Phys. Chem. C 2009, 113, 5865–5871

5865

Identification of the Dehydrogenated Product of Ca(BH4)2 Yoonyoung Kim,† Daniel Reed,‡ Young-Su Lee,*,† Ji Youn Lee,† Jae-Hyeok Shim,† David Book,‡ and Young Whan Cho† Materials Science and Technology Research DiVision, Korea Institute of Science and Technology, Seoul 136-791, Republic of Korea, and School of Metallurgy and Materials, UniVersity of Birmingham, Edgbaston, Birmingham B15 2TT, United Kingdom ReceiVed: October 23, 2008; ReVised Manuscript ReceiVed: January 19, 2009

We have investigated the decomposition path and reversibility of Ca(BH4)2 and Ca(BH4)2 + MgH2 composite using X-ray diffraction, thermogravimetric analysis, differential scanning calorimetry, and Raman spectroscopy. Formation of CaB6 during dehydrogenation of both systems was confirmed for the first time. CaB6 appears as broad peaks in X-ray diffraction data, but Raman spectroscopy unambiguously captures the existence of CaB6. Reversibility of catalyzed Ca(BH4)2 was previously reported, and here we demonstrate reversibility of Ca(BH4)2 + MgH2 composite. Dehydrogenated product of Ca(BH4)2 + MgH2 is composed of CaH2, CaB6, and Mg. About 60% reversibility was achieved after rehydrogenation for 24 h under 90 bar of hydrogen pressure at 350 °C even without the help of catalysts, which makes a good contrast with the case of pure Ca(BH4)2 where almost negligible rehydrogenation occurs under the same conditions. To understand the difference, the role of Mg in rehydrogenation is worth further investigation. Formation of CaB6 seems critical in the reversibility of Ca(BH4)2 containing systems; the case of other borohydrides is compared. 1. Introduction Reversibility of Ti-catalyzed NaAlH4 reported by Bogdanovic et al.1 opened up the possibility of adopting complex metal hydrides as hydrogen storage materials. Since then, numerous candidates such as alanates, borohydrides, amides, and their mixtures have been explored,2-4 yet none of them fully satisfies conditions required for practical on-board applications. Among all the difficulties that need to be overcome to meet the requirements, the most challenging part would be reversible rehydrogenation in a reasonable amount of time. Reversibility is sometimes thermodynamically forbidden when an exothermic process is part of dehydrogenation, but more often is the case where, even without thermodynamic restriction, slow kinetics limits reversibility in practice. For borohydrides, slow kinetics poses a more serious problem compared to alanates since B-B (Tm ) 2076 °C) bonding is much stronger than Al-Al (Tm ) 660 °C) bonding, and one can easily suppose that once B-B bonding is formed it would be very difficult to regenerate the [BH4]- unit. For example, rehydrogenation of LiH + B to reproduce LiBH4 has been carried out at 600 °C and under a hydrogen pressure of higher than 150 bar.5,6 Interestingly, the approach to introduce a second compound to control the thermodynamics of hydrogen absorption and desorption7 offers a way to improve reversibility by changing the dehydrogenated product. The case of LiBH4 is a good example: in its pristine form, LiH and amorphous B (a-B)5,8 is formed after dehydrogenation, and rehydrogenation is only possible under very high temperature and pressure conditions. However, when a second compound, such as MgH2, Al, CaH2, and CeH2, is mixed together, boron combines with the metal in the second compound to form diborides (MgB2 or * Corresponding author. E-mail: [email protected]. † Korea Institute of Science and Technology. ‡ University of Birmingham.

AlB2) or hexaborides (CaB6 or CeB6) instead of a-B, and reversibility improves significantly as a result.9-15 For this reason, it is of great importance to identify the end product of dehydrogenation and to figure out which compound shows relatively faster kinetics during the rehydrogenation process. In view of reversibility, Ca(BH4)2 makes an intriguing case. Kim et al.16,17 reported about 60% rehydrogenation of dehydrogenated of Ca(BH4)2 under 90 bar of H2 pressure at 350 °C. Ro¨nnebro et al.18 obtained Ca(BH4)2 by hydrogenating CaB6 and CaH2 mixed with catalysts under 700 bar of H2 pressure at 400 °C. This temperature and pressure condition is much milder than the case of LiBH4. The two most likely dehydrogenation reaction paths of Ca(BH4)2 are the following:

Ca(BH4)2 T 2/3 CaH2 + 1/3 CaB6 + 10/3 H2 9.6 wt % H2 (1) Ca(BH4)2 T CaH2 + 2B + 3H2

8.7 wt % H2

(2)

To our knowledge, the final form of boron has never been identified experimentally. The amount of hydrogen released serves as an indirect evidence of formation of CaB619,20 since 9.6 and 8.7 wt % hydrogen is released through reactions 1 and 2, respectively. Because the formation of CaB6 instead of a-B can be an important factor in reversibility as is the case for LiBH4 + CaH2 versus pure LiBH4, here we focus on identifying the boroncontaining compound in the dehydrogenated product of Ca(BH4)2 using XRD and Raman spectroscopy. As an extension to the investigation of the final form of boron and its effect on reversibility, we also study Ca(BH4)2 + MgH2 composite. Barkhordarian et al.21 reported that this composite can be made from hydrogenation of CaH2 + MgB2. However, dehydroge-

10.1021/jp8094038 CCC: $40.75  2009 American Chemical Society Published on Web 03/13/2009

5866 J. Phys. Chem. C, Vol. 113, No. 14, 2009

Kim et al.

TABLE 1: Compositions and Preparation Conditions for Samples 1-6 sample

composition

temperature

1 2 3 4 5 6

Ca(BH4)2 Ca(BH4)2 Ca(BH4)2 + MgH2 Ca(BH4)2 + MgH2 sample 2 rehydrogenated sample 4 rehydrogenated

330 480 350 400 350 350

pressure duration (bar) time (h) 90 90

2 2 2 2 24 24

nation of Ca(BH4)2 + MgH2 did not produce MgB2; Mg was found instead. Again, the final form of boron is yet unknown, and reversibility has not been tested. Here we identify the dehydrogenated product of Ca(BH4)2 + MgH2 composite and discuss possible decomposition routes based on our experimental results and theoretical calculations. 2. Methods 2.1. Experimental Section. Ca(BH4)2 · 2THF (assay 98%, Sigma-Aldrich) and MgH2 (assay 98%, Alfa Aesar) were used as starting materials. The powders were handled in an argonfilled glovebox (LABstar, MBraun) where water and oxygen levels were kept below 1 ppm. Adduct-free Ca(BH4)2 was prepared by drying the commercial powder at around 100 °C in vacuum for 20 h. This process usually produces a mixture of R- and γ-Ca(BH4)2 with a small fraction of β-Ca(BH4)2. Drying at higher temperature at around 200 °C gives mostly β-Ca(BH4)2. The dehydrogenation of Ca(BH4)2 was carried out in a Sievert’s apparatus. Two samples were prepared; one was dehydrogenated at 330 °C (sample 1) and the other at 480 °C (sample 2). Two hundred milligrams of Ca(BH4)2 was charged into a stainless steel tube reactor. The temperature was raised to the target point in 10 min, and the samples were soaked for 2 h. A static vacuum condition was maintained during dehydrogenation reaction. Rehydrogenation was carried out with sample 2 under 90 bar of hydrogen pressure at 350 °C for 24 h (sample 5). Approximately 2 g of Ca(BH4)2 + MgH2 mixture in a molar ratio of 1:1 was charged with 7 12.7 mm and 14 7.9 mm diameter Cr-steel balls into a hardened steel bowl and sealed with a lid having a Viton O-ring. The mixture was milled using a planetary mill (Fritsch P7) at 600 rpm for 4 h. Again two samples were prepared. Dehydrogenation was done following the above procedure at 350 °C (sample 3) and at 400 °C (sample 4). Rehydrogenation of sample 4 was carried out under 90 bar of hydrogen pressure at 350 °C for 24 h (sample 6). The composition and heat treatment condition applied to each sample are summarized in Table 1. The powder X-ray diffraction (XRD, Bruker D8 with Cu KR radiation, λ ) 1.5418 Å) data were collected for phase composition analysis. All XRD measurements were performed at room temperature. A dome-shaped, vacuum tight, sample holder was used to prevent contact with air during XRD measurement. XRD data were analyzed by Rietveld refinement using the TOPAS software.22 The dehydrogenation path of the samples was characterized using a differential scanning calorimeter (Netzsch DSC 204 F1) and a thermogravimetric analyzer (Netzsch TG 209 F1). For both differential scanning calorimetry (DSC) and thermogravimetric analysis (TGA), ca. 3 mg of sample was heated to 500 °C at a scanning rate of 2 °C/min under flowing argon condition (assay 99.9999%, 50 mL/min).

Figure 1. XRD patterns of the purchased and dried Ca(BH4)2 samples: (a) commercial Ca(BH4)2 · 2THF powder; (b) R-Ca(BH4)2 (f) containing a small amount of γ-Ca(BH4)2 (b) and β-Ca(BH4)2 ([); (c) β-Ca(BH4)2.

Raman spectroscopy was performed on samples 1-4 at room temperature using a Renishaw InVia reflex spectrometer, with a 633 nm excitation laser, and an Instec HCS621V sample hotstage that allowed inert loading. The Raman spectrum of CaB6 (Sigma-Aldrich) was obtained and used as a reference. 2.2. Theoretical Section. To estimate the thermodynamic stability of compounds of interest, we use both thermodynamic database and first-principles calculations. We perform thermodynamic calculations based on the Scientific Group Thermodata Europe (SGTE)23 substance thermodynamic database incorporated into the Thermo-Calc program.24 More specifically we adopt the parametrization by Shang et al.25 and Liu et al.26 for CaB6 and MgB2, respectively. For the substances without thermodynamic database and also for comparison, total energies and free energies are estimated within the density functional theory level of approximation. For exchange-correlational functional, the generalized gradient approximation by Perdew-Wang27 is adopted. Structural and vibrational properties are calculated using ultrasoft pseudopotentials28 and with plane-wave cutoff of 50 Ry and charge density cutoff of 400 Ry. Phonon density of states are calculated within density functional perturbation theory,29 and then Helmholtz free energies are obtained within harmonic approximation. All calculations have been performed using the QuantumESPRESSO package.30 3. Results and Discussion 3.1. Ca(BH4)2. Figure 1 shows XRD patterns of as-purchased Ca(BH4)2 · 2THF (Figure 1a) and dried Ca(BH4)2 (Figure 1b, c) at room temperature. XRD patterns in Figure 1b, c are from different polymorphs of Ca(BH4)2. Three different polymorphs have been reported up to now.31-34 The structures of lowtemperature R-Ca(BH4)2 (orthorhombic, space group Fddd) and high-temperature β-Ca(BH4)2 (tetragonal, P42/m) phases were resolved, and the crystal system and lattice parameters of the third phase, γ-Ca(BH4)2 (orthorhombic, Pbca), was reported. Relative thermal stability of γ-Ca(BH4)2 with respect to R-Ca(BH4)2 or β-Ca(BH4)2 is not known and needs further investigation. The DSC and TGA curves of Ca(BH4)2 are shown in Figure 2. Earlier studies have found the temperature of the polymorphic transformation from R-Ca(BH4)2 to β-Ca(BH4)2 to be about 130-170 °C.18-21 This polymorphic transformation is not clearly seen in our DSC curve. Part of the reason is that the starting

Dehydrogenated Product of Ca(BH4)2

Figure 2. (a) DSC, (b) TGA, and (c) differential TGA curves of a solvent-free Ca(BH4)2.

Ca(BH4)2 was a mixture of R, β, and γ-Ca(BH4)2 instead of a phase-pure R-Ca(BH4)2 (XRD data not presented here). Both DSC and TGA data indicate that dehydrogenation of Ca(BH4)2 starts at about 320 °C. When the sample was heated up to 550 °C, 9.0 wt % hydrogen was released. Based on the decomposition reaction 1, the theoretical weight loss is 9.6 wt %. The lower-than-ideal weight loss could be due to the purity of the starting material, exposure to air and moisture during the sample loading stage, and reaction 2 with lower hydrogen capacity occurring at the same time. Two major endothermic peaks in the DSC curve and an inflection point between 320 and 550 °C in the TGA curve indicate that Ca(BH4)2 desorbs hydrogen at least in two steps. We note that there is another endothermic peak at around 320 °C that also appears as a minimum in the differential TGA curve in Figure 2c. In previous reports, the dehydrogenation path of Ca(BH4)2 was suggested as follows:19,20,32

Ca(BH4)2 T CaH2 + intermediate compound T CaH2 + a-CaB6 and/or a-B (3) The reason why boron is thought to be in amorphous form is that no boron-containing phase was detected in XRD. For this reason, the final form of boron and the exact decomposition path are not fully understood. In order to elucidate the whole dehydrogenation process, we prepared two samples: one was dehydrogenated at 330 °C (sample 1) to capture the intermediate compound and the other at 480 °C (sample 2) to identify the fully dehydrogenated product. If we refer to Figure 2a, the major dehydrogenation did not even start at 330 °C, but this is because of upshift of the reaction temperature due to the finite heating rate. Under constant temperature of 330 °C, 5.6 wt % hydrogen was released after 2 h, which is consistent with the weight loss after major dehydrogenation in Figure 2b. XRD data of the samples 1 and 2 are presented in Figure 3. The XRD pattern of sample 1 is composed of peaks from CaH2, intermediate compound, and broad background. The peak positions of the intermediate compound agree well with previous reports.19,20 In the case of completely dehydrogenated sample 2, there is no trace of intermediate compound and the background intensity is reduced whereas broad humps from CaB6 become more conspicuous. The dashed blue line in Figure 3b shows contribution from CaB6 obtained by Rietveld refinement where lattice parameter and crystallite size are fitted to 4.12 Å (4.15 Å from CaB6 reference) and 3 nm, respectively. Although it is not as clear as in Figure 3b, we can observe that the

J. Phys. Chem. C, Vol. 113, No. 14, 2009 5867

Figure 3. XRD patterns of Ca(BH4)2 after dehydrogenation under static vacuum (a) at 330 °C and (b) at 480 °C. The dashed blue line indicates the broad background coming from nanocrystalline CaB6.

background is raised around the main peak position of CaB6 also in Figure 3a. Raman spectroscopy outperforms XRD in detecting nanosized crystallites. For both samples 1 and 2, three distinct peaks from CaB6 are seen in their Raman spectrum. They are T2g, Eg, and A1g Raman active modes of CaB6 in ascending frequency order. We compare Raman spectrum obtained from pure CaB6 in Figure 4c. Combining XRD and Raman data, we may draw a conclusion that a portion of Ca(BH4)2 decomposes directly into CaH2 and CaB6, while the remainder passes through the intermediate compound. Since decomposition of the intermediate compound starts at about 420 °C, CaH2 and CaB6 found in sample 1 are not likely from the dissociation of the intermediate compound. 3.2. Ca(BH4)2 + MgH2. The thermal decomposition process of Ca(BH4)2 + MgH2 composite was analyzed by DSC and TGA; the results are shown in Figure 5. The DSC curve shows two endothermic peaks from polymorphic transformation of Ca(BH4)2 at around 170 °C. The third peak between 320 and 400 °C accompanies weight loss as shown in the TGA curve. The Ca(BH4)2 + MgH2 composite starts to decompose at 320 °C and releases 8.1 wt % hydrogen when heated up to 400 °C. Three possible dehydrogenation reactions can be considered: Ca(BH4)2 + MgH2 T 2/3 CaH2 + 1/3 CaB6 + Mg + 13/3 H2 9.1 wt % H2 (4)

Ca(BH4)2 + MgH2 T CaH2 + MgB2 + 4H2 8.4 wt % H2 (5) Ca(BH4)2 + MgH2 T CaH2 + 2B + Mg + 4H2 8.4 wt % H2 (6) Boron takes a different form in all three cases, and through which reaction hydrogen is released can have a meaningful effect on the reversibility of this composite system. The first question to answer is which reaction is thermodynamically most favorable. Reaction 5 will always be more favorable than reaction 6 since the standard free energy of formation of MgB2 is negative. Between reactions 4 and 5, there could be a subtle competition depending on hydrogen pressure and temperature. The inspection of the number of hydrogen molecules released tells us that higher temperature and low hydrogen pressure will favor reaction 4. The free energy

5868 J. Phys. Chem. C, Vol. 113, No. 14, 2009

Kim et al.

Figure 4. Raman spectra of Ca(BH4)2 after dehydrogenation under static vacuum (a) at 330 °C and (b) at 480 °C. The Raman spectrum of CaB6 reference is compared in (c).

Figure 5. DSC curves of the first dehydrogenation of (a) Ca(BH4)2 + MgH2 and (b) Ca(BH4)2 + 0.2 Mg composite. TGA curves of (c) first and (d) second dehydrogenation of Ca(BH4)2 + MgH2 composite; rehydrogenation was done under 90 bar H2 pressure at 350 °C for 24 h.

change of the following reaction will determine which reaction pathway would be taken:

CaH2 + 3MgB2 T CaB6 + 3Mg + H2

(7)

Figure 6shows van’t Hoff plot for reaction 7, predicted by first-principles calculations and by SGTE thermodynamic database. First-principles calculation predicts that reaction 4 is favored under 1 bar of hydrogen pressure at 350 °C, i.e., Ca(BH4)2 and MgH2 would dissociate separately,35 but SGTE predicts the opposite. The deviation between the two predictions is appreciable, and to find out the source of discrepancy we summarize in Table 2 ∆H and ∆S of reaction 7 and of each constituent. It can be immediately seen that the enthalpy of formation of MgB2 differs by 28 kJ/mol between ab initio calculation and the experimental result by Liu et al.26 Since the formation energy of MgB2 enters three times in reaction 7, this is the main source of disagreement. It is difficult to say which one is more reliable. Experimental formation energies of borides often show a rather large scattering, and here the Gibbs energy function of MgB2 was obtained by fitting to a data set that varies as large as 30 kJ/mol,26 which cannot guarantee the required accuracy. Firstprinciples calculation has its own limitation: for example, the dissociation temperature of Ca(BH4)2 is predicted to be

Figure 6. Stability region of 2/3 CaH2 + 1/3 CaB6 + 13/3 H2 + Mg and CaH2 + MgB2 + 4H2 as a function of hydrogen pressure and temperature from (a) first-principles calculations (solid black line) and (b) SGTE thermodynamic database (dashed blue line). The lines indicate the temperature and hydrogen pressure condition where the free energies of the two dehydrogenated products are the same, i.e., ∆G ) 0 in reaction 7. The red lines show possible change in the equilibrium pressure when an error of (10 kJ/mol H2 in ∆G is considered in DFT calculations.

about 100 °C in our calculation and in previous study,35 while the experimental value is about 300 °C.19,20,32 However, even after taking into account an error of (10 kJ/mol H236 in DFT results, the deviation still remains significant. To better understand the thermodynamics of this composite system, reevaluation of the Gibbs energy function of MgB2 is necessary. In the case of CaB6, Shang et al.25 recently argued that experimental stability of CaB6 is largely underestimated compared to other divalent alkaline-earth hexaborides and provided a new Gibbs energy function obtained by ab initio quasiharmonic calculation. We adopted their Gibbs energy function instead of the one in the SGTE database. Although it is not at all conclusive, there could be a chance of close competition between reactions 4 and 5 under usual experimental conditions. For example, at 350 °C the SGTE result predicts that dehydrogenation under 1 bar of hydrogen would see the formation of MgB2 whereas dynamic vacuum condition would find CaB6. From the DSC and TGA results in Figure 5a, c, Ca(BH4)2 + MgH2 seems to undergo a single-step dehydrogenation reaction. In order to find out details of the decomposition path, two samples were prepared by dehydrogenating Ca(BH4)2 + MgH2 at 350 °C (sample 3) and 400 °C (sample 4). The XRD data of sample 3 is plotted in Figure 7a. Ca4Mg3H14, Mg, and unknown intermediate compound(s) were detected. The peak positions of the unknown phases agree well with the previous result by Barkhordarian et al.21 XRD data hints that the dehydrogenation would not occur in a single step and at least two steps are involved, which will be explained later. It is likely that several endothermic peaks overlap in the temperature range of 350-400 °C. The XRD peaks of sample 4 in Figure 7b are mainly from CaH2 and Mg. Raman spectroscopy is employed to find out boron-containing compound in samples 3 and 4. The results are shown in Figure 8. For both samples, CaB6 is found instead of MgB2. We again point out that because of uncertainty in thermodynamic data, it is not clear whether our experimental conditions would favor the formation of CaB6 or not. If MgB2 is a more stable specie at the given condition, formation of CaB6 can be explained by

Dehydrogenated Product of Ca(BH4)2

J. Phys. Chem. C, Vol. 113, No. 14, 2009 5869

TABLE 2: Standard Enthalpy and Entropy Change under 1 Bar of Hydrogen Pressure at 25 °C and the Temperature at Which the Equilibrium Hydrogen Pressure Becomes 1 Bar reaction

∆HoDFT (kJ/mol H2)

∆SoDFT (J/K · mol H2)

TDFT (° C)

∆HoSGTE (kJ/mol H2)

∆SoSGTE (J/K · mol H2)

TSGTE (° C)

Ca(BH4)2 T 2/3 CaH2 + 1/3 CaB6 + 10/3 H2 Ca(BH4)2 T CaH2 + 2B + 3H2 Ca(BH4)2 + MgH2 T 2/3 CaH2 + 1/3 CaB6 + Mg + 13/3 H2 Ca(BH4)2 + MgH2 T CaH2 + MgB2 + 4H2 Ca(BH4)2 + MgH2 T CaH2 + 2B + Mg + 4H2 MgH2 T Mg + H2 CaH2 + 3MgB2 T CaB6 + 3Mg + H2 CaH2 T Ca + H2 Mg + 2B T MgB2 Ca + 6B T CaB6 4CaH2 + 3MgH2 T Ca4Mg3H14

40.6 57.3 45.0 46.9 57.9 59.7 22.2 173.1 -43.9 -282.8 1.3

109.3 105.7 114.4 110.0 112.0 131.2 166.3 136.0 -8.0 6.2 14.5

98 263 120 151 238 179 -148 (1003) -

75.7 104.9 177.0 -72.1 -288.4 -

132.3 128.8 130.9 1.0 1.3 -

288 508 1017 -

short-range mass transport required when forming Ca-B compound since both Ca and B are from Ca(BH4)2, whereas formation of MgB2 would require a longer-range diffusion of Mg or B. Reasoning from what we have found in XRD and Raman data of samples 3 and 4, the decomposition path can be summarized as follows:

Ca(BH4)2 + MgH2 T Ca4Mg3H14 + CaB6 + Mg + unknown + H2 T CaH2 + CaB6 + Mg + H2 (8) The initial stage would probably involve dissociation of MgH2. The highest-temperature decomposition peak in Figure 2a is again not seen in the DSC curve of Ca(BH4)2 + 0.2Mg in Figure 5b, which implies that MgH2 would already have transformed to Mg when Ca(BH4)2 starts to dissociate.

MgH2 T Mg + H2

(9)

Dissociation of Ca(BH4)2 (reaction 1) and formation of Ca4Mg3H14 and the unknown compound will follow. It is not clear whether Mg directly reacts with Ca(BH4)2 to form Ca4Mg3H14 or reaction 1 precedes and the product, CaH2, combines with Mg to form Ca4Mg3H14. The last stage is dissociation of Ca4Mg3H14 and the unknown compound:

Ca4Mg3H14 T 4CaH2 + 3Mg + 3H2

(10)

Ca4Mg3H14 dissociates at a higher temperature than MgH2 since the Helmholtz free energy change of the following reaction is predicted to be slightly negative by our first-principles calculations

4CaH2 + 3MgH2 f Ca4Mg3H14 ∆F ) -7.9 kJ at 350 °C (11) implying that the activity of MgH2 is lowered by forming Ca4Mg3H14. Now we turn our attention to the role of MgH2. If we simply compare the dehydrogenated product of Ca(BH4)2 + MgH2 composite with that of the individual constituents, Ca(BH4)2 and MgH2, we may conclude that there is no interaction between Ca(BH4)2 and MgH2 because CaB6 was formed instead of MgB2, which is better described by reaction 4. Note that reaction 4 is a simple sum of reactions 1 and 9. Apart from the formation of Ca4Mg3H14, which seems not to make much difference in the

decomposition process, the most important change made by MgH2 is that the intermediate compound appearing in decomposition of pure Ca(BH4)2 was not found in the case of Ca(BH4)2 + MgH2, which removes the highest-temperature endothermic peak in Figure 2a, thereby lowering the end point of decomposition. The different unknown compound(s) found here seems not as stable as the intermediate compound found in pure Ca(BH4)2. In short, MgH2 (or Mg) changes the decomposition path instead of the decomposition product, like a catalyst does. A 1:1 ratio of Ca(BH4)2 to MgH2 derived from reaction 5 might not be so meaningful if MgH2 really behaves like a catalyst, and a small amount of MgH2 is desirable since it will increase the overall hydrogen capacity. 3.3. CaB6 and Reversibility. Since the reversibility of pure Ca(BH4)2 has already been shown,16-18 here we tested the reversibility of Ca(BH4)2 + MgH2 composite. Starting from sample 4, we tried rehydrogenation reaction without catalyst under 90 bar of hydrogen pressure at 350 °C for 24 h. The amount of hydrogen absorbed was measured by TGA. The result is shown in Figure 5d together with TGA curve of as-milled Ca(BH4)2 + MgH2 sample for comparison. As can be seen in Figure 5d, 60% hydrogen was reversibly absorbed. The result is very encouraging since 60% reversibility was achieved without catalyst, and we retested whether the same thing can be found for pure Ca(BH4)2, but the amount of rehydrogenation was negligible even after 48 h under the same temperature and pressure condition. Our tentative conclusion from this result is that Mg (or MgH2) brings about some catalytic effects also during the rehydrogenation reaction. Turning to the common features of Ca(BH4)2 and Ca(BH4)2 + MgH2 that boron ends up with CaB6 and that higher than 50% reversibility is achieved at relatively milder conditions than LiBH4, the formation of CaB6 is certainly beneficial for reversibility. CaB6 may improve reversibility because each B atom is connected to five other B atoms via covalent bond while in R-B each B atom has more than five covalent bondings. In this respect, MgB2 would show the fastest kinetics since each B atom has only three B-B bondings.14 The number of B-B bondings in the structures is closely related to the average bond strength. The DFT-calculated energies required to break these compounds into isolated atoms are 4.76, 6.00, and 6.20 eV/atom for MgB2, CaB6, and R-B, respectively. As discussed before, reversibility of LiBH4 is very difficult to achieve. Friedrichs et al. demonstrated that it can be improved by using an LiB3-like alloy as a starting material instead of Li-B mixture.37 Mg(BH4)2 can be an another example. Dehydrogenation proceeds largely through a two-step reaction, Mg(BH4)2 T MgH2 + 2B + 3H2 T Mg + 2B + 4H2, and only the second step has shown

5870 J. Phys. Chem. C, Vol. 113, No. 14, 2009

Kim et al. 4. Conclusions We have identified that boron exists as CaB6 in the dehydrogenated product of Ca(BH4)2 and Ca(BH4)2 + MgH2composite using Raman spectroscopy and X-ray diffraction. The dehydrogenation path of each case was summarized based on this finding. Reversibility shown in both compounds under relatively mild condition could be due to the formation of CaB6 instead of a-B. Finally, MgH2 seems to act as a catalyst in Ca(BH4)2+ MgH2 composite during both dehydrogenation and rehydrogenation processes. The detailed mechanism of how MgH2 changes the reaction path is worth further investigation.

Figure 7. XRD patterns of Ca(BH4)2 + MgH2 composite after dehydrogenation under static vacuum (a) at 350 °C and (b) at 400 °C. The dashed blue line indicates the broad background coming from nanocrystalline CaB6.

Acknowledgment. The authors at KIST acknowledge financial support from the Hydrogen Energy R&D Center, one of the 21st Century Frontier R&D Programs funded by the Ministry of Education, Science, and Technology of Korea. D.R. and D.B. acknowledge support from the AWM Science City Hydrogen Energy project and from EPSRC UK Sustainable Hydrogen Energy Consortium. References and Notes

Figure 8. Raman spectra of Ca(BH4)2 + MgH2 composite after dehydrogenation under static vacuum (a) at 350 °C and (b) at 400 °C. The Raman spectrum of CaB6 reference is compared in (c).

to be fully reversible.38-40 Since most of the metal borohydrides dissociate in multiple steps, intermediate compounds could be as important as the final products in determining the overall reversibility. Hwang et al.41 recently confirmed formation of [B12H12]2- during the dehydrogenation of LiBH4 and Mg(BH4)2 using NMR spectroscopy. Li et al.42 proposed that MgB12H12 could be an intermediate state of the first dehydrogenation step of Mg(BH4)2 and the rehydrogenation up to MgB12H12 might be reversible based on the amount of hydrogen absorbed. In the case of Ca(BH4)2, it is yet unknown whether CaB12H12 would be formed and the intermediate compound in Figure 3a would be CaB12H12. In our Raman experiment on samples 1 and 3, we were not able to find vibration frequencies of [B12H12]2-, and therefore CaB12H12 might not be a major intermediate compound at least under our experimental conditions. Still, a question remains as to why full reversibility could not be achieved. It could be because kinetics is too slow or because part of B goes in a-B, which is hardly reversible. We think the former is the case since Raman spectroscopy confirmed that CaB6 still exists in sample 6 (not shown here) and reversibility of pure Ca(BH4)2 is almost none. If the latter is the case, one should find a way to promote the formation of CaB6 and to suppress the competing process of forming a-B, to achieve better reversibility.

(1) Bogdanovic, B.; Schwickardi, M. J. Alloys Compd. 1997, 253, 1. (2) Orimo, S.; Nakamori, Y.; Eliseo, J. R.; Zu¨ttel, A.; Jensen, C. M. Chem. ReV. 2007, 107, 4111. (3) Meisner, G. P.; Scullin, M. L.; Balogh, M. P.; Pinkerton, F. E.; Meyer, M. S. J. Phys. Chem. B 2006, 110, 4186. (4) Miwa, K.; Ohba, N.; Towata, S.; Nakamori, Y.; Orimo, S. J. Alloys Compd. 2005, 404, 140. (5) Orimo, S.; Nakamori, Y.; Kitahara, G.; Miwa, K.; Ohba, N.; Towata, S.; Zu¨ttel, A. J. Alloys Compd. 2005, 404-406, 427. (6) Mauron, P.; Buchter, F.; Friedrichs, O.; Remhof, A.; Bielmann, M.; Zwicky, C. N.; Zu¨ttel, A. J. Phys. Chem. B 2008, 112, 906. (7) Vajo, J. J.; Mertens, F.; Ahn, C. C.; Bowman, R. C.; Fultz, B. J. Phys. Chem. B 2004, 108, 13977. (8) Orimo, S.; Nakamori, Y.; Ohba, N.; Miwa, K.; Aoki, M.; Towata, S.; Zu¨ttel, A. Appl. Phys. Lett. 2006, 89, 3. (9) Vajo, J. J.; Skeith, S. L.; Mertens, F. J. Phys. Chem. B 2005, 109, 3719. (10) Jin, S.-A.; Shim, J.-H.; Cho, Y. W.; Yi, K.-W.; Zabara, O.; Fichtner, M. Scr. Mater. 2008, 58, 963. (11) Pinkerton, F. E.; Meyer, M. S. J. Alloys Compd. 2008, 464, L1. (12) Jin, S.-A.; Lee, Y.-S.; Shim, J.-H.; Cho, Y. W. J. Phys. Chem. C 2008, 112, 9520. (13) Yang, J.; Sudik, A.; Wolverton, C. J. Phys. Chem. C 2007, 111, 19134. (14) Barkhordarian, G.; Klassen, T.; Dornheim, M.; Bormann, R. J. Alloys Compd. 2007, 440, L18. (15) Bo¨senberg, U.; Doppiu, S.; Mosegaard, L.; Barkhordarian, G.; Eigen, N.; Borgschulte, A.; Jensen, T. R.; Cerenius, Y.; Gutfleisch, O.; Klassen, T.; Dornheim, M.; Bormann, R. Acta Mater. 2007, 55, 3951. (16) Kim, J.-H.; Jin, S.-A.; Shim, J.-H.; Cho, Y. W. Scr. Mater. 2008, 58, 481. (17) Kim, J.-H.; Shim, J.-H.; Cho, Y. W. J. Power Sources 2008, 181, 140. (18) Ro¨nnebro, E.; Majzoub, E. H. J. Phys. Chem. B 2007, 111, 12045. (19) Kim, J.-H.; Jin, S.-A.; Shim, J.-H.; Cho, Y. W. J. Alloys Compd. 2008, 461, L20. (20) Aoki, M.; Miwa, K.; Noritake, T.; Ohba, N.; Matsumoto, M.; Li, H.-W.; Nakamori, Y.; Towata, S.; Orimo, S. Appl. Phys. A 2008, 92, 601. (21) Barkhordarian, G.; Jensen, T. R.; Doppiu, S.; Bo¨senberg, U.; Borgschulte, A.; Gremaud, R.; Cerenius, Y.; Dornheim, M.; Klassen, T.; Bormann, R. J. Phys. Chem. C 2008, 112, 2743. (22) TOPAS version 4.1, Bruker AXS GmbH, Karlsruhe, Germany, 2003-2008. (23) http://www.sgte.org. (24) Sundman, B.; Jansson, B.; Anderson, J.-O. Calphad-Comput. Coupling Ph. Diagrams Thermochem. 1985, 9, 153. (25) Shang, S. L.; Wang, T.; Liu, Z.-K. Calphad-Comput. Coupling Ph. Diagrams Thermochem. 2007, 31, 286. (26) Liu, Z.-K.; Zhong, Y.; Schlom, D. G.; Xi, X. X.; Li, Q. CalphadComput. Coupling Ph. Diagrams Thermochem. 2001, 25, 299. (27) Perdew, J. P.; Wang, Y. Phys. ReV. B 1992, 45, 13244. (28) Vanderbilt, D. Phys. ReV. B 1990, 41, 7892. (29) Baroni, S.; de Gironcoli, S.; Dal Corso, A.; Giannozzi, P. ReV. Mod. Phys. 2001, 73, 515.

Dehydrogenated Product of Ca(BH4)2 (30) Baroni, S.; et al. http://www.quantum-espresso.org. (31) Miwa, K.; Aoki, M.; Noritake, T.; Ohba, N.; Nakamori, Y.; Towata, S.; Zu¨ttel, A.; Orimo, S. Phys. ReV. B 2006, 74, 155122. (32) Riktor, M. D.; Sørby, M. H.; Chłopek, K.; Fichtner, M.; Buchter, F.; Zu¨ttel, A.; Hauback, B. C. J. Mater. Chem. 2007, 17, 4939. (33) Buchter, F.; Łodziana, Z.; Rernhof, A.; Friedrichs, O.; Borgschulte, A.; Mauron, P.; Zu¨ttel, A.; Sheptyakov, D.; Barkhordarian, G.; Bormann, R.; Chłopek, K.; Fichtner, M.; Sørby, M.; Riktor, M.; Hauback, B.; Orimo, S. J. Phys. Chem. B 2008, 112, 8042. (34) Fichtner, M.; Chłopek, K.; Longhini, M.; Hagemann, H. J. Phys. Chem. C 2008, 112, 11575. (35) Siegel, D. J.; Wolverton, C.; Ozolinsˇ, V. Phys. ReV. B 2007, 76, 134102. (36) Alapati, S. V.; Johnson, J. K.; Sholl, D. S. J. Phys. Chem. C 2007, 111, 1584.

J. Phys. Chem. C, Vol. 113, No. 14, 2009 5871 (37) Friedrichs, O.; Buchter, F.; Borgschulte, A.; Remhof, A.; Zwicky, C. N.; Mauron, P.; Bielmann, M.; Zu¨ttel, A. Acta Mater. 2008, 56, 949. (38) Li, H.-W.; Kikuchi, K.; Nakamori, Y.; Miwa, K.; Towata, S.; Orimo, S. Scr. Mater. 2007, 57, 679. (39) Chłopek, K.; Frommen, C.; Le´on, A.; Zabara, O.; Fichtner, M. J. Mater. Chem. 2007, 17, 3496. (40) Matsunaga, T.; Buchter, F.; Mauron, P.; Bielman, A.; Nakamori, Y.; Orimo, S.; Ohba, N.; Miwa, K.; Towata, S.; Zu¨ttel, A. J. Alloys Compd. 2008, 459, 583. (41) Hwang, S. J.; Bowman, R. C.; Reiter, J. W.; Rijssenbeek, J.; Soloveichik, G. L.; Zhao, J. C.; Kabbour, H.; Ahn, C. C. J. Phys. Chem. C 2008, 112, 3164. (42) Li, H.-W.; Kikuchi, K.; Nakamori, Y.; Ohba, N.; Miwa, K.; Towata, S.; Orimo, S. Acta Mater. 2008, 56, 1342.

JP8094038