19134
J. Phys. Chem. C 2007, 111, 19134-19140
Destabilizing LiBH4 with a Metal (M ) Mg, Al, Ti, V, Cr, or Sc) or Metal Hydride (MH2 ) MgH2, TiH2, or CaH2) Jun Yang,*,† Andrea Sudik,† and C. Wolverton‡ Research and AdVanced Engineering, Ford Motor Company, MD1170/RIC, P. O. Box 2053, Dearborn, Michigan 48121, and Department of Materials Science and Engineering, Northwestern UniVersity, 2220 Campus DriVe, EVanston, Illinois 60208 ReceiVed: August 10, 2007; In Final Form: September 13, 2007
We experimentally investigate several hydrogen storage reactions based on thermodynamic destabilization of LiBH4. The destabilized mixtures include nine M(H2)-LiBH4 compositions, where M(H2) ) Al, Mg, Ti, Sc, V, Cr, MgH2, CaH2, or TiH2, which were selected on the basis of favorable thermodynamics predicted by recent first-principles computational study (Siegel, D. J.; Wolverton, C.; Ozolin¸ sˇ, V. Phys. ReV. B: Condens. Matter, Mater. Phys. 2007, 76, 134102). For all compositions, our measurements reveal significant kinetic barriers for hydrogen release, evidenced by high desorption temperatures (>300 °C) and exceedingly slow hydrogen release rates. Characterization of the desorbed reaction phases indicate that less than half of the mixtures examined (M(H2) ) MgH2, Mg, Al, and CaH2) follow the thermodynamically expected reaction pathway, resulting in the formation of metal boride products (MgB2, AlB2, and CaB6, respectively). Hydrogen release/uptake data for these compositions indicate that the MgH2-(LiBH4)2 and Al-(LiBH4)2 reactions are reversible (10.2 and 6.7 wt %, respectively) at increased desorption pressures of 5 and 3 bar H2, respectively, while CaH2-(LiBH4)6 is irreversible under the conditions tested. For the remaining compositions, M(H2) ) Sc, V, Cr, Ti, and TiH2, we surmise that substantial limitations in kinetics inhibit the expected formation of metal boride products. Finally, we discuss how the relative physical properties, in particular the melting point of the reactants and products, correlate with desorption route and reversibility.
1. Introduction Complex hydrides, one of the primary classes of materials currently under consideration for solid-state on-board hydrogen storage, are ionic compounds often with group 1 or 2 cations (e.g., Li, Na, and Mg), and an anionic “complex” of light metals and hydrogen (e.g., BH4- and AlH4-). Interest in low-Z complex hydrides, such as metal borohydries,1-3 Mn+(BH4)n, and metal alanates,4 Mn+(AlH4)n, stems from their inherent ability to contain large amounts of hydrogen by both weight (up to 18.5 wt % in LiBH4) and volume (up to 150 kg/L H2 in Al(BH4)3), attributes that are essential toward attaining desired system-level hydrogen storage targets (9 wt % H2 and 81 kg/L).5 Unfortunately, the practical use of such traditional complex hydrides in fuel cell (FC) applications is limited due to both thermodynamic and kinetic deficiencies. That is, both the temperaturepressure conditions and rate at which the hydrogen can be incorporated and released from these compounds are generally not suitable for on-board vehicle storage. For example, partial decomposition of LiBH4 according to
LiBH4 f LiH + B + 3/2H2
(1)
(13.7 wt % H2 capacity) has a standard reaction enthalpy (∆H) of ∼67 kJ/mol H2, translating to a thermodynamic desorption temperature of 410 °C at 1 bar [on the basis of the van’t Hoff equation (ln P ) {∆H}/{RT} - {∆S}/{R})].6 Therefore, recent efforts have focused on incorporating additives, such as metals,7,8 * To whom correspondence should be addressed. † Ford Motor Co. ‡ Northwestern University.
metal halides,7,9,10 oxides,7,9 sulfides,10 hydrides,7 or more recently, nanoporous scaffolds,10 to thermodynamically destabilize LiBH4 toward optimized (lowered) desorption temperatures (Tdes). While reduced Tdes and partial reversibility have been demonstrated in select cases, the most noteworthy execution of this strategy was achieved by Vajo and Skeith using MgH2.6 Specifically, combining a 2:1 molar ratio of LiBH4: MgH2 results in a decomposition pathway which is distinct from the isolated decomposition of the individual reactants due to the exothermic formation of MgB2 from Mg and B:
2LiBH4 + MgH2 T 2LiH + MgB2 + 4H2
(2)
In particular, this favorable formation of MgB2 effectively stabilizes the dehydrogenated state and results in an overall 27 kJ/mol H2 decrease in ∆H for eq 2 relative to eq 1.6 Additionally, a 8-10 wt % reVersible storage capacity was also demonstrated. Nevertheless, although the extrapolated Tdes at 1 bar for eq 2 was determined to be 225 °C, kinetic barriers limit direct measurement at these temperatures and instead require >300 °C for desorption. Similarly, significant kinetic barriers were also evident for the majority of the other aforementioned destabilizing agents (refs 7-10). Thus, two primary challenges still remain for destabilized-LiBH4 systems: (1) further thermodynamic reduction in ∆H (Tdes) relative to pure LiBH4 and (2) sufficient improvement in reaction kinetics at lower temperatures. The identification of methods to further destabilize LiBH4s and more generally, other strongly bound hydridesshas emerged as an active area of hydrogen storage research. Due to the large number of potential destabilizing agents, one would ideally like
10.1021/jp076434z CCC: $37.00 © 2007 American Chemical Society Published on Web 11/30/2007
Destabilizing LiBH4 with a Metal or Metal Hydride
J. Phys. Chem. C, Vol. 111, No. 51, 2007 19135
to quickly screen a large number of candidate reactions, searching for thermodynamically suitable ones. High-throughput experimentation is one possible method to provide a rapid screen. But, several groups have recently shown that firstprinciples density functional theory (DFT) calculations can also provide a rapid computational screen.11-15 For example, in a recent study, Siegel and co-workers used first-principles calculations to evaluate the thermodynamic properties of a series of reactions aimed at destabilizing LiBH4 and Ca(BH4)2 by mixing with various elemental metals (M) and/or their binary hydrides (MH2).11 They found that a range of destabilization was possible: from strongly destabilized mixtures involving TiH2 (characterized by ∆H less than 5 kJ/mol H2) to more moderately destabilized systems involving ScH2 and Cr, which were found to possess thermodynamics (∆H ∼ 30 kJ/mol H2) enabling more ambient hydrogen storage.16 Motivated by this study, herein we experimentally investigate several of the promising predicted reactions, specifically for LiBH4 mixed with the following nine metals or metal hydrides: M ) Al, Mg, Ti, V, Sc, and Cr; MH2 ) MgH2, TiH2, and CaH2. In ref 11, it is shown that the thermodynamically preferred pathways for these reactions are given by eqs 3-6, respectively:
2LiBH4 + M f MB2 + 2LiH + 3H2 (M ) Al, Cr)
(3)
2LiBH4 + M f 1.5LiBH4 + 0.75MH2 + 0.25MB2 + 0.5LiH f MB2 + 2LiH + 3H2[M ) Mg, Ti, Sc, V* (V* forms V2H)]
(4)
2LiBH4 + MH2 f MB2 + 2LiH + 4H2 (MH2 ) MgH2, TiH2)
(5)
6LiBH4 + MH2 f MB6 + 6LiH +10H2 (MH2 ) CaH2)
(6)
The signature of the thermodynamically preferred reaction in each case is the formation of the stable metal boride phase. We note that the compositions involving “unstable” metal additives are predicted to first involve the exothermic formation of a stable hydride which subsequently reacts endothermically with remaining LiBH4 as shown in eq 4.11 However, as we detail below, all reactions may be kinetically constrained, resulting in reaction products of the pure metal or metal hydride. In the present work, we describe the synthesis and characterization of the above nine M(H2)-modified LiBH4 compositions and evaluate their initial desorption properties using temperature-programmed desorption mass spectrometry (TPDMS) and isothermal kinetics measurements (400 °C to 1 bar). The product phases from the latter experiment were analyzed using powder X-ray diffraction (PXRD) and infrared spectroscopy (IR) and used to determine each respective desorption pathway. We show that for select additives, M(H2) ) Al, Mg, MgH2, and CaH2, the expected metal boride product phases are indeed formed (from eqs 3-6), thereby validating the thermodynamically predicted reaction products in ref.11. For these select compositions, we also investigate the impact of desorption pressure conditions toward optimizing the reversible capacity. For the remaining additives, M(H2) ) Cr, V, Sc, Ti, and TiH2, we reveal significant limitations in kinetics that inhibit the expected formation of metal boride products. Finally we discuss potential factors that may explain why the thermodynamically expected desorption pathway is accessible (and reversible) only
in select cases (i.e., accessible for M(H2) ) Al, Mg, MgH2, and CaH2 and reversible for all of these except CaH2). 2. Experimental Section 2.1. Sample Preparation. Lithium borohydride (LiBH4, 95%, Sigma-Aldrich), magnesium hydride (MgH2; 95%, Gelest), titanium hydride (TiH2; 99%, Alfa-Aesar), calcium hydride (CaH2; 98%, Alfa-Aesar), magnesium (Mg; 99.8%, Alfa-Aesar), aluminum (Al; 99.97%, Alfa-Aesar), titanium (Ti; 99.99%, AlfaAesar), vanadium (V; 99.5%, Alfa-Aesar), chromium (Cr; >99%, Sigma-Aldrich), and scandium (Sc; 99.9%, SigmaAldrich) were used as received. All sample handling was performed in an MBraun Labmaster 130 glovebox maintained under an argon atmosphere with 500 °C) are significantly higher than the thermodynamic values predicted using first-principles calculations which range from approximately 25 °C for M ) Cr to 277 °C for M ) Al.11 Thus, to investigate whether these high temperatures are thermodynamic or kinetic in nature and to determine which of the additives are indeed effective for destabilizing LiBH4 via formation of metal boride products, all reactions and product phases were further characterized using isothermal kinetic desorption and PXRD (sections 3.3 and 3.4, respectively). 3.3. Isothermal Kinetic Desorption in Modified LiBH4 Samples. To evaluate the hydrogen storage capacity and relative desorption rate for each M(H2)-LiBH4 mixture, isothermal kinetic desorption data were collected. Experiments were performed using a PCT apparatus, where as-prepared samples were desorbed to a 1 bar hydrogen back-pressure at temperatures ranging from 370 to 400 °C, values employed on the basis of the observed desorption range from TPD-MS screening data (Figure 2). Quantities of desorbed hydrogen (in weight percent) versus time (in hours) were monitored and are depicted for each sample in Figure 3. For all mixtures, the observed amounts of released hydrogen were less than what are expected on the basis of eqs 3-6. Specifically after up to 100 h, the total observed (expected on the basis of eqs 3-6) quantity of hydrogen desorbed from each sample was as follows: TiH2, 1.7 (8.6) wt%; Ti, 2.5 (6.5) wt %; Sc, 2.9 (6.7) wt %; V, 4.4 (6.3) wt %; Cr, 4.4 (6.3) wt %; CaH2, 5.1 (11.7) wt %; Mg, 5.6 (8.8) wt %; Al, 6.8 (8.5) wt %; and MgH2, 10.2 (11.5) wt %. Likewise, the rates of hydrogen release were exceedingly slow for most compositions, with several mixtures (e.g., MgH2, Mg, CaH2, Sc, and TiH2) unable to reach full desorption after being held for 100 h at the isothermal set point (∼400 °C). These extreme kinetic limitations have been previously observed for both neat borohydrides (e.g., Mg(BH4)2 and Ca(BH4)2)19,20 as well as destabilized borohydride mixtures.10 3.4. Product Phase Analysis and Desorption Pathway Determination for M- and MH2-LiBH4 Mixtures. To
J. Phys. Chem. C, Vol. 111, No. 51, 2007 19137 determine if the hydrogen release reactions proceed according to their thermodynamically predicted pathways (i.e., forming a metal boride as in eqs 3-6), the identities of the desorbed species must be determined. Therefore, postreaction products from the isothermal desorption experiments described in section 3.3 (desorption at ∼400 °C to 1 bar H2 for e100 h) were analyzed using PXRD and are shown in Figure 4. Along with each pattern, ICDD PDF card data for the assigned product phases are compared as lines within each respective panel. Based on these results, the nine M- or MH2-modified LiBH4 reactions can be classified into three categories: case I, reactions involving pure metal products; case II, reactions involving metal hydride products; and case III, reactions involving metal boride products. Case III represents the thermodynamically preferred pathway (i.e., eqs 3-6), and hence cases I and II represent kinetically controlled decomposition pathways. We next elaborate on each of these categories. 3.4.1. Cases I and II: Reactions InVolVing Kinetically Controlled Metal or Metal Hydride Products (M ) Cr, V, Sc, and Ti and MH2 ) TiH2). PXRD analysis for desorbed (LiBH4)2-M and (LiBH4)2-MH2 samples where M ) Cr, V, Sc, and Ti or MH2 ) TiH2 are not consistent with the thermodynamically expected formation of a metal boride product. Instead, the observed hydrogen desorption likely corresponds to decomposition of LiBH4 (into LiH, B, and H2) as reflected by reduced B-H stretches (2570 to 2000 cm-1) in the IR spectra for desorbed samples (see Figure S1 in the Supporting Information for a select example). The M or MH2 starting material remains either unreacted (for Cr and TiH2) or, when thermodynamically favored, forms a stable hydride (for V, Sc, and Ti). In particular, the phases present in the PXRD patterns (Figure 4) for desorbed (LiBH4)2-Cr and (LiBH4)2TiH2 are simply Cr and TiH2, respectively. For (LiBH4)2-V, (LiBH4)2-Sc, and (LiBH4)2-Ti, the PXRD patterns show V2H, ScH2, and TiH2 phases, respectively. However, since metal boride phases do not accompany these newly formed metal hydride phases as predicted by eq 4, instead we hypothesize that a portion of hydrogen either from desorbed LiBH4 or from the 1 bar H2 back-pressure facilitates the favorable formation of these observed hydrides (see ∆Hf° values in Table 1). Interestingly, on the basis of the TPD-MS screening data from Figure 2 (section 3.2), all of the non-boride forming reactions do release hydrogen at temperatures slightly less than that of milled LiBH4 (peak Tdes ∼ 450 °C). This suggests that while the Cr, TiH2, V, Sc, and Ti additives do not thermodynamically destabilize LiBH4, they do result in modest improvements in kinetic properties (i.e., reduced desorption temperatures and accelerated desorption rates). Kinetic enhancements have also been observed for metal oxide and chloride additives.7,9 To confirm that the above M or MH2 additives are simply acting as LiBH4 dehydrogenation “catalysts”, we prepared two LiBH4 samples containing 5 wt % V or TiH2, respectively (milled under conditions identical to those previously described), and repeated TPD-MS collections. These data, found in the Supporting Information (Figures S2 and S3), indeed reveal desorption behavior virtually identical to the original (LiBH4)2-V or (LiBH4)2-TiH2 compositions. Thus, we conclude that for M ) Cr, V, Sc, and Ti and MH2 ) TiH2, the observed desorption temperature enhancements are kinetic, but not thermodynamic, in nature. 3.4.2. Case III: Reactions InVolVing Thermodynamically Preferred Metal Boride Products (M ) Al and Mg and MH2 ) CaH2 and MgH2). In contrast to the previous mixtures, PXRD data for desorbed (LiBH4)2-Al, (LiBH4)2-Mg, (LiBH4)2-
19138 J. Phys. Chem. C, Vol. 111, No. 51, 2007
Yang et al.
Figure 4. Room-temperature powder X-ray diffraction patterns (PXRD) for desorbed (LiBH4)2-M and (LiBH4)x-MH2 samples, where M ) Mg (pink), Al (dark green), Ti (light blue), scandium (dark blue), or vanadium (green); MH2 ) MgH2 (gray), CaH2 (purple), or TiH2 (orange); and x ) 2 for MgH2 and TiH2 or 6 for CaH2. Samples were desorbed at ∼400 °C to 1 bar hydrogen (unless otherwise indicated) for up to 100 h (conditions for individual samples identical to those in Figure 3). Phase assignments, which were made on the basis of comparisons with PDF data, include metal borides (blue lines), or M or MH2 products (line color same as pattern) and are included within each individual panel.
TABLE 1: Physical Properties of Reactant/Product Species for (LiBH4)2-M and (LiBH4)x-MH2 Mixtures Tm b (°C) reaction 2LiBH4 + Cr 2LiBH4 + V 2LiBH4 + Sc 2LiBH4 + Ti 2LiBH4 + TiH2 6LiBH4 + CaH2 2LiBH4 + Al 2LiBH4 + Mg 2LiBH4 + MgH2 a
∆Hf°(solid) (kJ/(mol‚H2)) (metal hydride) a
N/A (CrH2 metastable) -143 (V2H) -201 (ScH2) -131 (TiH2) -183 (CaH2) -7 (AlH3) -75 (MgH2)
metal
metal boride
1863 1910 1541 1670
2200 (CrB2) 2750 (VB2) 2250 (ScB2) 3225 (TiB2)
842 660 650
2400 (CaB6) 1030 (AlB2) 800 (MgB2)
Reference 24. b Reference 25.
MgH2, and (LiBH4)6-CaH2 reveals the presence of the desired metal boride phases (AlB2, MgB2, and CaB6, respectively), products consistent with the thermodynamically predicted reaction pathways (eqs 3-6). Specifically, for (LiBH4)2-Mg and (LiBH4)2-MgH2 (Figure 4), diffraction peaks corresponding to MgB2 are apparent from PXRD data. As previously observed, to maximize the extent of the boride-forming reaction and avoid self-decomposition of reactant LiBH4, a hydrogen back-pressure of 5 bar has been recommended.6 In agreement with previous measurements,6 we obtain full conversion to MgB2 for desorption to a 5 bar hydrogen back-pressure, while desorption to a 1 bar back-pressure yields a mixture of MgB2 and Mg phases (Figure 4). For this latter case, the conditions are such that the MgB2 formation and LiBH4 and MgH2 self-decomposition reactions can occur simultaneously. A similar observation is found for the (LiBH4)2-Al composition, where a mixture of product phasessAlB2, Al, and an unknown phase (major peak ∼50°)sis identified for desorption to 1 bar (Figure 4). The concurrent existence of both the AlB2-forming and LiBH4 decomposition reaction pathways is not unexpected as the pressure-temperature conditions (1 bar; ∼400 °C) are conducive for both reactions to occur on the basis of thermodynamics
(for example, at 1 bar Tdes ≈ 275 °C for 2LiBH4 + Al f AlB2 + 2LiH + 3H2 and Tdes ≈ 320 °C for 2LiBH4 f 2LiH + 2B + 3H2).11 However, by increasing the desorption backppressure to 3 bar at ∼400 °C, a relative enhancement in the contribution of the desired boride-forming reaction can be achieved (for example at 3 bar Tdes ≈ 340 °C for 2LiBH4 + Al f AlB2 + 2LiH + 3H2 and Tdes ≈ 400 °C for 2LiBH4 f 2LiH + 2B + 3H2). This enhancement is evidenced by a relative increase in diffraction peaks corresponding to AlB2 and simultaneous decrease in the unknown product phase (see PXRD data in Supporting Information Figure S4). Experiments involving higher desorption back-pressures were only examined for M(H2) ) MgH2 and Al, compositions where evidence of boride formation was initially observed. On the basis of the PXRD data (Figure 4), MH2 ) CaH2 also appears to be effective for destabilizing LiBH4 as evidenced by the formation of CaB6. In addition to this product (of eq 6), other unidentifiable phases are also present. While persistence of starting CaH2 is not observed in the desorbed pattern, completion of the boride-forming reaction is not achieved after desorption at ∼400 °C for 100 h on the basis of the significant discrepancy between the observed (5.1 wt %) and theoretical
Destabilizing LiBH4 with a Metal or Metal Hydride (11.7 wt %) hydrogen release amounts in Figure 3. It is interesting to note that in contrast to the MgH2-(LiBH4)2 whose mechanism involves initial decomposition of MgH2 resulting in a two-step desorption process (Figure 2), the CaH2-(LiBH4)6 composition provides the first example of a LiBH4-destabilization reaction involving a MH2 that occurs at conditions below the decomposition of the metal hydride itself. This suggests that low-decomposition-temperature hydrides are not an essential component for this class of LiBH4 destabilization reactions. Overall, it is clear that for these select M and MH2 additives (i.e., Mg, MgH2, Al, and CaH2) the thermodynamically predicted reaction products can be experimentally realized. It is also evident that the contribution of these reactions is dependent on the temperature-pressure conditions and that in some cases (e.g., M ) Al and MH2 ) CaH2) there are additional unknown side reactions which are observed. 3.5. Reversibility. To evaluate the reversible properties of the M(H2)-LiBH4 compositions which formed the desired metal boride products, PCT cycling data were collected. Specifically, hydrogen desorption and uptake was monitored for M(H2) ) MgH2, Al, and CaH2 and the resulting desorbed and rehydrided product phases characterized using PXRD and IR (included as Supporting Information). Desorption conditions were performed at temperatures between 370 and 400 °C and at hydrogen pressures of 1 bar and either 3 (Al) or 5 (MgH2) bar until equilibrium was reached (up to >300 h); rehydriding conditions were performed at 350 °C and 150 bar hydrogen pressure.21 As previously observed by Vajo and co-workers and as discussed above, the MgH2-(LiBH4)2 composition fully forms the expected MgB2 product phase and releases >10 wt % hydrogen when desorbed to a 5 bar hydrogen back-pressure at 385 °C (Figure 4).6 This reaction is able to be fully reversed to the original MgH2 and LiBH4 starting components upon rehydriding as evidenced by the presence of these phases in PXRD and IR data, respectively (Figures S5 and S6 in Supporting Information). For the unexplored Al-(LiBH4)2 composition, PCT data were collected for three desorption cycles at 395 °C to a 1 bar hydrogen back-pressure and reveals a continuously degrading capacity (Figure S7 of the Supporting Information). In particular, 6.3, 4.2, and 3.8 wt % are observed for the first, second, and third desorption cycles, respectively. However, as previously observed for the MgH2-(LiBH4)2 composition, a higher desorption back-pressure can be used to try to further maximize the formation of the desired metal boride product.6 Therefore, PCT data were also collected for two desorption cycles of Al-(LiBH4)2 at 395 °C to a 3 bar hydrogen back-pressure and is shown in Figure 5. These data indicate that the reaction process is indeed reversible and is also kinetically improved in the subsequent cycle. Here, the first desorption cycle (red) reaches 5.2 wt % after 320 h, whereas the second desorption cycle (blue) achieves 6.7 wt % after 170 h. The IR data for desorbed (to 3 bar) and rehydrided samples also confirm the reversible nature of this system by the respective disappearance and reappearance of B-H stretches corresponding to LiBH4 (Figure S8 in the Supporting Information). Finally, the CaH2-(LiBH4)6 composition was found to form the expected CaB6 product phase upon the first desorption cycle but was not capable of being reversed under the pressuretemperature conditions that were examined.
J. Phys. Chem. C, Vol. 111, No. 51, 2007 19139
Figure 5. Isothermal kinetic desorption data for two desorption cycles (first, red; second, blue) for (LiBH4)2-Al at 395 °C and 3 bar hydrogen back-pressure. Rehydriding between desorption cycles was performed at 350 °C and 150 bar hydrogen pressure for ∼50 h.
relative physical properties of the reactants and products (Table 1). We note that further experiments are necessary to concretely validate these suggested trends. While thermodynamics suggest that all of the M(H2)-LiBH4 compositions should proceed according to eqs 3-6 and (at high enough temperatures) involve the favorable formation of a metal boride product, kinetic attributes of the reacting M or MH2 component appear to determine whether or not this desired reaction pathway is accessible. The reaction pathways described by eqs 3-6 require mass transport of M-, B-, or H-containing species, and one possible kinetic limitation could be in the atomic diffusivities of the additive metal, hydride, or boride species at temperatures at which the desorption reaction occurs. The melting point of a compound is often used as a rough indicator of diffusivity in the material.22 We give the melting points of the reactant metals and metal borides in Table 1. From this table, we note that the ability for the two starting materials (LiBH4 and M or MH2) to react appears to be correlated with the melting point of the metal.23 This observation is evidenced by the disparity in melting point temperatures for those reactions that formed metal borides and those that did not (see Table 1). For example, for Mg, Al, and Ca, the melting points are 650, 660, and 842 °C, respectively. Conversely, the melting points of the metals for the remaining M(H2) additives are all >1650 °C, approximately double the value of those that successfully formed metal boride products. Thus, when diffusion of the M(H2) additives are inadequate (i.e., have high melting points), they either remain unreacted (e.g., Cr and TiH2), or when favorable (see ∆Hf° values in Table 1), form a stable metal hydride (e.g., V, Sc, and Ti). We also considered the relative physical properties of the product metal borides to gain an understanding of why the CaH2-(LiBH4)6 composition was practically incapable of being reversed in our experiments. Data in Table 1 reveal that it may be due to the mobility of the metal boride product phase. That is, sufficient diffusion of the metal boride may be required to promote the rehydriding process. Indeed, there is a notable disparity between the melting points of AlB2 (1030 °C) and MgB2 (800 °C) relative to CaB6 (2400 °C), which supports this hypothesis.
4. Discussion To gain an understanding of why only select M or MH2 additives result in the thermodynamically preferred boride formation (under our experimental conditions), we refer to the
5. Conclusions The use of reactant destabilization (or product stabilization) continues to serve as a promising strategy for tuning the
19140 J. Phys. Chem. C, Vol. 111, No. 51, 2007 thermodynamic properties of high-density hydrogen storage materials. In this way, suitable temperature-pressure conditions for hydrogen uptake and release can be rationally targeted. Firstprinciples computational analysis has drastically aided in this process by “virtually” screening potential destabilization additives and identifying candidates having the most desirable thermodynamic properties which can subsequently be examined experimentally. In particular, we have shown here that for select LiBH4 destabilization additives, M(H2) ) Mg, MgH2, Al, and CaH2, the thermodynamically predicted reaction products are indeed observed experimentally. However, as with many promising multicomponent hydrogen storage compositions, extreme kinetic limitations prohibit the remainder of the predicted reactions from occurring (i.e., for M(H2) ) Ti, TiH2, Sc, V, and Cr). We hypothesize these kinetic barriers stem from poor atomic diffusion (i.e., high-temperature melting) of the reactant M(H2) component. Instead, these additives appear to merely serve as catalysts for the self-decomposition of LiBH4. Therefore, perhaps the most difficult future challenge for destabilized-LiBH4 systemssand the majority of other hydrogen storage compositionssis to identify rational strategies to predict and improve reaction kinetics. Acknowledgment. The authors thank Dr. Don Siegel for his valuable comments and discussions. Supporting Information Available: Infrared (IR) spectra and powder X-ray diffraction (PXRD) characterization data for various select desorbed and rehydrided M(H2)-LiBH4 compositions, stoichiometry-dependent temperature-programmed desorption mass spectrometry (TPD-MS) data for V-LiBH4 and TiH2-LiBH4 compositions, and pressure-composition-temperature (PCT) cycling data for Al-(LiBH4)2 desorption to a 1 bar hydrogen back-pressure. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Zu¨ttel, A.; Rentsch, S.; Fisher, P.; Wenger, P.; Sudan, P.; Mauron, Ph.; Emmenegger, Ch. J. Alloys Compd. 2003, 356, 515.
Yang et al. (2) Zu¨ttel, A.; Wenger, P.; Rentsch, S.; Sudan, P. J. Power Sources 2003, 118, 1. (3) Nakamori, Y.; Li, H.; Miwa, K.; Towata, S.; Orimo, S. Mater. Trans. 2006, 47, 1898. (4) Bogdanovic´, B.; Schwickardi, M. J. Alloys Compd. 1997, 253, 1. (5) Satyapal, S.; Petrovic, J.; Read, C.; Thomas, G.; Ordaz, G. Catal. Today 2007, 120, 246. (6) Vajo, J. J.; Skeith, S. L. J. Phys. Chem. B 2005, 109, 3719. (7) Reilly, J. J.; Wiswall, R. H. Inorg. Chem. 1968, 7, 2254. (8) Au, M.; Jurgensen, A.; Zeigler, K. J. Phys. Chem. B 2006, 110, 26482. (9) Au, M.; Jurgensen, A. J. Phys. Chem. B 2006, 110, 7062. (10) Vajo, J. J.; Salguero, T. T.; Gross, A. F.; Skeith, S. L.; Olson, G. L. J. Alloys Compd. 2007, 446-447, 409. (11) Siegel, D. J.; Wolverton, C.; Ozolin¸ sˇ, V. Phys. ReV. B: Condens. Matter, Mater. Phys. 2007, 76, 134102. (12) Alapati, S. V.; Johnson, J. K.; Sholl, D. S. J. Phys. Chem. B 2006, 110, 8769. (13) Alapati, S. V.; Johnson, J. K.; Sholl, D. S. J. Phys. Chem. C 2007, 111, 1584. (14) Wolverton, C. U.S. Patent Appl. US2006455468A, 2006. (15) Wolverton, C.; Lewis, G.; Low, J. U.S. Patent Appl. US2006471443A, 2006. (16) This is the approximate target enthalpy for hydrogen storage materials assuming ∆G is zero. (17) Soulie´, J-Ph.; Renaudin, G.; C ˇ erny´, R.; Yvon, K. J. Alloys Compd. 2002, 346, 200. (18) Determined using Derby-Scherrer formula based on the half-width at maximum peak of the metal or metal hydride. (19) Matsunaga, T.; Buchter, F.; Mauron, P.; Bielman, M.; Nakamori, Y.; Orimo, S.; Ohba, N.; Miwa, K.; Towata, S.; Zu¨ttel, A. J. Alloys Compd., in press (doi: 10.1016/j.jallcom.2007.05.054). (20) Ro¨nnebro, E.; Majzoub, E.; McDaniel, T. Discovery and Development of Metal Hydrides for Reversible On-board Storage. DOE Hydrogen Program Merit Review, Arlington, VA, May 2007. (21) Pressure conditions used (150 bar) are higher than those predicted (87 bar) for M ) Al based on its predicted ∆H ) 57.9 kJ/mol H2 11 and ∆S ) 130 J/(K‚mol) H2. (22) Borg, R. J.; Dienes, G. J. The Physical Chemistry of Solids; Academic Press: San Diego, CA, 1991; Chapter 13. (23) The stability of a binary metal hydride (MH2) can be represented by its formation enthalpy. On the other hand, their phase transformation temperatures (e.g. melting) are related to the melting points of pure metal (M). See ref 24. (24) Fukai, Y. The Metal-Hydrogen System, 2nd ed.; Springer: Berlin, Germany, 2005. (25) Massalski, T. B., Okamoto, H., Subramanian, P. R., Kacprzak, L., Eds. Binary Alloy Phase Diagrams, 2nd ed.; ASM International: Materials Park, OH, 1990.