Hydrogen Desorption Properties of Bulk and Nanoconfined LiBH 4-NaAlH 4

Nanoconfinement of 2LiBH4-NaAlH4 into a mesoporous carbon aerogel scaffold with a pore size, BET surface area and total pore volume of Dmax = 30 nm, SBET = 689 m2/g and Vtot = 1.21 mL/g, respectively is investigated. Nanoconfinement of 2LiBH4-NaAlH4 facilitates a reduction in the temperature of the hydrogen release by 132  ̋C, compared to that of bulk 2LiBH4-NaAlH4 and the onset of hydrogen release is below 100  ̋C. The reversible hydrogen storage capacity is also significantly improved for the nanoconfined sample, maintaining 83% of the initial hydrogen content after three cycles compared to 47% for that of the bulk sample. During nanoconfinement, LiBH4 and NaAlH4 reacts to form LiAlH4 and NaBH4 and the final dehydrogenation products, obtained at 481  ̋C are LiH, LiAl, AlB2 and Al. After rehydrogenation of the nanoconfined sample at T = 400  ̋C and p(H2) = 126 bar, amorphous NaBH4 is recovered along with unreacted LiH, AlB2 and Al and suggests that NaBH4 is the main compound that can reversibly release and uptake hydrogen.


Introduction
Solid state hydrogen storage has received significant attention during the past few decades and is associated with a future carrier of renewable energy [1,2].Initially, metal hydrides and magnesium hydride were the focus for much research and several useful applications have been developed [3][4][5][6].In particular, lithium borohydride, LiBH 4 , is considered a potential candidate as a hydrogen storage material due to its relatively high gravimetric hydrogen storage capacity of ρ m = 13.6 wt.% based on decomposition reaction (1) taking place above 375 ˝C in vacuum.
LiBH 4 psq Ñ 2LiHpsq `Bpsq `3H 2 pgq (1) However, the kinetics for hydrogen release and uptake is very sluggish i.e., hydrogen uptake requires elevated temperatures and pressures, T = 600 ˝C and p(H 2 ) > 155 bar [7].One way to alter kinetic properties for hydrogen release of reactive composites of LiBH 4 is by additives using transition metals such as Ti, V, Cr or Sc [8][9][10][11].Other metals like Al and Mg have also shown significant destabilization abilities as a reactive hydride composite (RHC) with LiBH 4 , by reducing the enthalpy of reversible dehydrogenation and rehydrogenation reactions [12][13][14].Al nanoparticles have also been considered as an Al-source but the properties are generally inhibited due to an oxide layer, which limits the reactivity [10].Another drawback for the LiBH 4 -Al system is the significant reduction of the hydrogen content, which has pointed attention towards NaAlH 4 as the Al source.Furthermore, NaAlH 4 can reversibly store hydrogen at moderate conditions according to the two-step decomposition reaction shown in reactions (2) and (3).
Nanoconfinement of metal hydrides may change the reaction mechanism for reactive hydride composites, and may also improve the kinetics for chemical reactions significantly [19][20][21].Confinement of NaAlH 4 in carbon aerogel, which was also activated or functionalized, has been investigated.The fastest kinetics are observed for TiCl 3 functionalized scaffolds, while CO 2 activated scaffolds provide slower kinetics but more efficient infiltration and higher reversible hydrogen storage capacity over several cycles of hydrogen release and uptake [22][23][24].
This work presents hydrogen release and uptake properties of nanoconfined 2LiBH 4 -NaAlH 4 into a nanoporous carbon aerogel scaffold, synthesised via melt infiltration for further improvement of the system, with the purpose of destabilizing the borohydride in the solid state.Various techniques are used to investigate hydrogen storage properties of this system.

Sample Preparation
The resorcinol formaldehyde carbon aerogel was prepared by mixing 82.87 g resorcinol (Aldrich, 99%), 113.84 mL formaldehyde (37 wt.% stabilized by ~10%-15% methanol, Merck), 113.28 mL deionized water and 0.0674 g Na 2 CO 3 (Aldrich, 99.999%) in a beaker with continuous stirring until complete dissolution was obtained.The pH of the final solution was measured to be 5.91.The preparation and characterization of the aerogel was performed according to previously published procedures [22,25,26].Prior to use, the scaffolds were all degassed at 400 ˝C in a dynamic vacuum for several hours, in order to remove possible adsorbed air and water confined inside the porous structure.This carbon aerogel scaffold is denoted CA.All subsequent handling was performed in a glovebox with a purified argon atmosphere.
An amount of hydride was selected in order to obtain ~33 vol.% pore filling of the carbon scaffold based on the total pore volume, V tot , of the scaffold and the average bulk densities of the hydrides, ρ(LiNa) = 0.8482 g/mL.Melt infiltration was performed in a custom made rig, by heating to T = 310 ˝C (∆T/∆t = 2 ˝C/min) with the temperature kept fixed at 310 ˝C for 30 min, at a hydrogen pressure of 110 bar.Afterwards the sample was cooled naturally to room temperature (RT).

Sample Characterization
Synchrotron radiation powder X-ray diffraction (SR-PXD) data was collected at beamline I711 at MAX-lab, Lund, Sweden.The samples were mounted in a sapphire capillary tube (0.79 mm I.D.), in an airtight sample holder inside an argon filled glovebox [28,29].The sample holder was removed from the glovebox and attached to a gas control system at the synchrotron diffractometer.The data was collected using a MAR165 CCD detector with a selected wavelength of λ = 0.991779 Å. Heating was applied by a tungsten wire placed under the capillary, whereas the temperature was controlled by an external PID regulator and thermocouple inserted into the powder-bed as previously shown [28].
A Perkin Elmer STA 6000 was utilized to conduct thermogravimetric analysis (TGA) coupled with a Hiden Analytical quadrupole mass spectrometer (MS) for differential scanning calorimetry (DSC).Thus, temperature-programmed desorption mass spectroscopy (TPD-MS) data is provided.Data was collected with a constant flow (64 mL/min) of argon (99.99%).A powdered sample (<5 mg), was placed in an Al 2 O 3 crucible with lid and heated in the temperature range of 40 to 550 ˝C (∆T/∆t = 10 ˝C/min).
The MS signals at m/e = 2, 18 and 34 were recorded in order to detect H 2 , H 2 O and B 2 H 6 .
The reversible hydrogen storage capacity of nanoconfined and bulk 2LiBH 4 -NaAlH 4 was studied during multiple hydrogen release and uptake cycles.The samples were sealed in an autoclave under argon and attached to the Sieverts' apparatus (PCTpro 2000).Hydrogen desorption data was collected starting under an initial H 2 pressure of 1 bar and was performed in the temperature range of RT to 500 ˝C, (∆T/∆t = 5 ˝C/min), with the temperature kept constant at 500 ˝C for 10 h.Hydrogen absorption was performed in the pressure range of p(H 2 ) = 140 to 150 bar, at a temperature of 400 ˝C (∆T/∆t = 5 ˝C/min) during 10 h, and then the sample was cooled naturally to RT.
The Fourier transform infrared spectrometry (FTIR) analyses were carried out on a NICOLET 380 FT-IR from Thermo-Electronic Corporation with permanently aligned optics and proprietary diamond-turned pinned-in-place mirror optics.A small amount of sample was placed on the baseplate and subsequently the diamond pin was pressed on to the sample, forming a thin film.The samples were examined within the wave number range of 400-4000 cm ´1.The samples were briefly exposed to air when mounted in the instrument.This may be the metathesis reaction described by reaction (4), however LiAlH 4 is not observed.Lithium tetrahydridoaluminate, LiAlH 4 , may decompose immediately according to reaction (5), which explains observation of Li 3 AlH 6 in the temperature range 116 to 227 ˝C.Since diffraction from Na 3 AlH 6 and NaH is not observed, NaAlH 4 is expected to be fully converted by the metathesis reaction (4) rather than decomposition via reactions ( 2) and (3).Characteristic strong diffraction at 2θ = 25 ˝and 28 ˝from Al occur at T = 115 ˝C and is assumed mainly to be formed via the decomposition of LiAlH 4 .Similarly, diffraction from h-LiBH 4 is not visible in the data possibly due to reaction (4).NaBH 4 is observed in the temperature range from 116 to 435 ˝C.Weak diffraction from LiH, is also observed in the temperature range 116 and 485 ˝C, partially overlapping with the Al peaks and is expected mainly to be formed by decomposition of Li 3 AlH 6 .Furthermore, two unknown peaks are observed at 2θ ~20 ˝and 2θ ~23 ˝, forming simultaneous to Li 3 AlH 6 .A shift of Bragg peaks towards smaller 2θ angles is observed due to thermal expansion and the associated increase in unit cell parameters.During natural cooling of the sample, thermal contraction is observed along with formation of Al, LiAl 3 , and NaBH 4 along with an unknown compound.A difference plot of the in situ SR-PXD data is shown in supporting information to accentuate changes in weaker peaks.A decrease in the background scattering occurs between 160 and 198 °C which coincides with the maximum rate of Li3AlH6 and Al formation.Further work would be required to determine if this relationship is causative or correlative.The Li3AlH6 diffraction decreases at T > 192 °C and disappears at T ~ 227 °C.The decomposition of Li3AlH6 is associated with further production of Al and the observation of weak diffraction from LiH.Following decomposition reactions could take place:

Results and Discussion
Diffraction of NaBH4 decreases in intensity at 412 to 435 °C just prior to a slight increase in the background intensity.Weak peaks of an unknown compound at 2θ ~14°, ~21° and ~22° are also observed to form simultaneously.It should be noted that the disappearance of NaBH4 at 435 °C occurs well below both its melting point of 505 °C and the decomposition temperature of T(1 bar) ~ 515-534 NaH Ñ Na `1{2H 2 (10) Diffraction of NaBH 4 decreases in intensity at 412 to 435 ˝C just prior to a slight increase in the background intensity.Weak peaks of an unknown compound at 2θ ~14 ˝, ~21 ˝and ~22 ˝are also observed to form simultaneously.It should be noted that the disappearance of NaBH 4 at 435 ˝C occurs well below both its melting point of 505 ˝C and the decomposition temperature of T(1 bar) ~515-534 ˝C [30].The simultaneous decrease in diffracted intensity of LiH and NaBH 4 may suggest eutectic melting supported by the reappearance of NaBH 4 and LiH upon cooling at 487 ˝C.We note that eutectic melting of NaBH 4 -NaH at T = 395 ˝C has previously been observed [31].
In situ SR-PXD of nanoconfined LiBH 4 -NaAlH 4 measured during heating to 495 ˝C is shown in Figure 1.At RT diffraction of LiBH 4 and small amounts of NaAlH 4 are observed and partial decomposition due to the temperature used for the melt infiltration would explain the presence of diffraction peaks assigned to Al.Furthermore, the presence of NaBH 4 at RT also indicates metathesis reaction (4) during melt infiltration.A broad unknown peak at 2θ = 27 ˝is observed throughout the measurement.The o-LiBH 4 to h-LiBH 4 polymorphic transition is observed, however only during a short temperature range of 97 to 110 ˝C.The background intensity decrease in the temperature range T ~238 to 250 ˝C but is most pronounced below 2θ = 15 ˝and is associated with the formation of two peaks from an unknown phase at 2θ ~19 ˝and 2θ ~23 ˝.These peaks are similar to those formed in the bulk sample.At T = 318 ˝C, a diffraction peak appears at 2θ ~21.3 ˝which coincides with a decrease in diffraction intensity from NaBH 4 .This peak continues to slowly grow until all of the NaBH 4 decomposes at T ~406 ˝C.At ~487 ˝C, unknown peaks begin to form at 2θ ~19 ˝, 22 ˝and 31.4 and these slowly grow until the heating stops.The growth of these peaks is associated with a slight decrease in the intensity of the Al peaks.The unknown peaks may result from a reaction between the hydride(s) and possible oxygen impurities in the aerogel scaffold.However, due to the relatively weak and overlapping Bragg reflections, indexation remain unsuccessful.

Nanoporous Carbon Aerogel Composite
The synthesized resorcinol formaldehyde carbon aerogel (CA) scaffold has a maximum pore size distribution centred on D max = 30 nm, a specific BET surface area of S BET = 689 m 2 /g and a total pore volume of V tot = 1.21 mL/g (see Table 1).The binary hydride composite 2LiBH 4 -NaAlH 4 (LiNa) has been successfully melt infiltrated into the mesoporous carbon scaffold via melt infiltration upon elevated hydrogen pressures of p(H 2 ) = 110 bar, in order to avoid decomposition of the respective hydrides (please see note about this in experimental section).Confinement of the hydride mixture facilitates the formation of nanoparticles of the respective hydrides, in accordance with previous studies [27].The amount of 2LiBH 4 -NaAlH 4 added to CA is 25.3 wt.% corresponding to 33 vol.% pore filling.The theoretical hydrogen content of the binary hydride system is ρ m (LiNa) = 10.33 wt.% calculated using the bulk densities of the respective hydride according to the molar ratios and reaction (11) The available hydrogen capacity of the CA-LiNa confined sample is determined to be 2.6 wt.% H 2 .However, considering 100 vol.% pore filling would potentially give a 7.8 wt.% hydrogen capacity.

Decomposition of Bulk and Nanocomposites of 2LiBH 4 -NaAlH 4
The of hydrogen desorption from bulk 2LiBH 4 -NaAlH 4 (LiNa), LiNa physically mixed with carbon aerogel scaffold (CA) and LiNa melt infiltrated into CA, was analysed by TG-DSC-MS during heating from 40 to 550 ˝C with the results shown in Figure 2. The TG-DSC-MS measurements can provide information about the endothermic and exothermic nature of any transitions or chemical reactions that take place during heating.In addition, if the same transitions and chemical reactions occur in different samples, TG-DSC-MS can provide information about the relative kinetic differences between samples.Each data set is normalized to the maximum hydrogen desorption rate to allow for qualitative comparison between data sets.Hydrogen emission from bulk 2LiBH 4 -NaAlH 4 is observed as two local maxima at 234 and 496 ˝C.The first hydrogen release peak in the MS spectrum (T = 200-250 ˝C) is associated with the decomposition of Li 3 AlH 6 in accord with the in situ SR-PXD data (see Figure 1).This temperature range is in good agreement with previous reports on decomposition of bulk Li 3 AlH 6 (180-230 ˝C) [32,33].A mass loss of 3.36 wt.% is observed by TGA in the temperature range 200 to 250 ˝C.Differential scanning calorimetry data reveal two thermal events at 112 and 120 ˝C, which may be assigned to the polymorphic oto h-LiBH 4 transition and the metathesis reaction (reaction 4), respectively.Partial amorphisation or melting of the sample may occur, which may be observed as a significant increase in the background of the diffraction data in the temperature range 70 to 150 ˝C (see Figure 1).Several thermal events in the temperature range ~200 to 220 ˝C may be associated with partial melting of the sample and decomposition of Li 3 AlH 6 [32].Eutectic melting of the LiBH 4 -NaBH 4 is reported to occur at ~250 ˝C without release of hydrogen [30].The second MS hydrogen release event (T = 350-500 ˝C) of bulk LiNa is assigned to the decomposition of remaining LiBH 4 and NaBH 4 , which corresponds to a weight loss of 6.16 wt.%.The total TGA mass loss in the temperature range 40 to 550 ˝C is 9.52 wt.%, which corresponds to 92% of the available hydrogen content (ρ m (LiNa) = 10.33 wt.%).
The carbon containing samples (physically mixed and melt infiltrated) both release 2.48 wt.% H 2 when heated from 40 to 550 ˝C, which corresponds to 95% of the theoretical hydrogen content in the sample (see Figure 2).However, the physically mixed sample has a hydrogen release profile similar to that of bulk LiNa at T < ~250 ˝C with a peak value at T = 228 ˝C shifted 6 ˝C towards lower temperatures.The nanoconfined sample LiNa-CA does not show the aforementioned desorption peak.The temperature of the maximum hydrogen release rate of the mixed (T = 356 ˝C) and melt infiltrated samples (T = 364 ˝C) occur at significantly reduced temperatures (∆T ~130 ˝C) as compared to that of bulk LiBH 4 -NaAlH 4 (T = 496 ˝C).This effect has previously been reported for other nanoconfined binary hydride systems that showed a reduction in the temperature of the maximum hydrogen release rate compared to bulk, e.g., LiBH 4 -Mg(BH 4 ) 2 (∆T max ~´60 ˝C), LiBH 4 -Ca(BH 4 ) 2 (∆T max ~´95 ˝C) and LiBH 4 -NaBH 4 (∆T max ~´107 ˝C) [33][34][35].
The improvement of hydrogen release kinetics due to the carbon scaffold, is assigned to the effect induced by nanoconfinement and possible catalytic properties of the carbon surface [36].In addition, the DSC signal assigned to the major hydrogen desorption at T ~340 ˝C is reduced by ~140 ˝C as compared to that of bulk LiNa.
According to Figure S2, after decomposition of the nanoconfined sample AlB 2 , LiAl 3 , LiH and some unknown phases are formed.No diffraction Bragg peaks of NaBH 4 are displayed suggesting possible formation of amorphous metal borides.This is supported by the FTIR data discussed later.

Cyclic Stability
The reversible hydrogen storage properties of bulk and nanoconfined 2LiBH 4 -NaAlH 4 (LiNa) have been investigated by the Sieverts' method during multiple hydrogen release and uptake cycles and the data is displayed in Figure 3.During heating of bulk LiNa from RT to 500 ˝C, hydrogen is released in two steps with a plateau of slow hydrogen evolution in the temperature range 160 to 190 ˝C.The total amount of hydrogen released from the sample during the first desorption in the temperature range RT to 500 ˝C is 8.37 wt.%, which is 81% of the theoretical available hydrogen content.After subsequent rehydrogenation at p(H 2 ) = 126 bar, T = 400 ˝C for 10 h, the second, third and fourth desorption cycle release 4.65, 3.94 and 3.38 wt.% H 2 , corresponding to 56%, 47% and 40% of the initial hydrogen content, respectively.The slow hydrogen release at 160 to 190 ˝C is only observed during the first desorption, as clearly demonstrated in Figure 4. Thus, the reversible fraction of the sample may be LiBH 4 and/or NaBH 4 formed in a possibly amorphous form.Sodium and lithium alanate are not expected to form at the high temperatures used for rehydrogenation.This fact along with the formation of stable decomposition products LiH, LiAl 3 and Al, may explain the significant difference between the hydrogen storage capacities observed after the first desorption.wt.% H2/LiNa corresponding to 96% and 83% of the initial hydrogen content, and equivalent to 89% and 77% of the available hydrogen content, respectively.These results demonstrate that nanoconfinement has significantly improved hydrogen release kinetics and cyclic stability compared to bulk 2LiBH4-NaAlH4.The nanoconfined sample, LiNa-CA, showed substantially improved cyclic stability and improved hydrogen storage capacity compared to that of bulk LiNa as shown in Figure 3.The hydrogen release of LiNa-CA is shown relative to the amount of 2LiBH 4 -NaAlH 4 in the nanoconfined to facilitate comparison with LiNa.The hydrogen release from the nanoconfined system also has a two-step behaviour for the first H 2 desorption profile similar to bulk LiNa but at higher temperatures (~380 ˝C).The first desorption of LiNa-CA releases 9.6 wt.% H 2 with respect to LiNa and 2.4 wt.% releative to the total mass of the sample.This corresponds to 93% of the theoretical available hydrogen content (ρ m (LiNa-CA) = 2.6 wt.% H 2 ).It is worth noting that the nanoconfined sample is fully desorbed after 2 h at 500 ˝C, whereas bulk 2LiBH 4 -NaAlH 4 , on the contrary continues to release hydrogen throughout the entire measurement.The second and third desorptions release 9.2 and 7.9 wt.% H 2 /LiNa corresponding to 96% and 83% of the initial hydrogen content, and equivalent to 89% and 77% of the available hydrogen content, respectively.These results demonstrate that nanoconfinement has significantly improved hydrogen release kinetics and cyclic stability compared to bulk 2LiBH 4 -NaAlH 4 .
Differentiated Sieverts' data provides the change in rate of H 2 release during each desorption cycle upon heating from RT to 500 ˝C (see Figure 4).Selected desorption profiles from Figure 4 resemble the TPD-MS data shown in Figure 2, despite different measurement conditions and heating rates.During the first hydrogen release cycle of bulk LiNa, the maximum H 2 release rate is observed at 198 ˝C which corresponds to the first plateau observed in Figure 3.A second peak value at 492 ˝C is in good agreement with TPD-MS measurements in Figure 3.During the second, third and fourth desorption, only one hydrogen release event is observed with onset at T ~440 ˝C, which decreases in intensity with increasing desorption cycles.
The major hydrogen release of the nanoconfined sample (Figure 4B) is exhibited as a broad peak with maxima at 305 ˝C during the first desorption.However during the second and third desorption an event is observed with a local maxima at 447 ˝C.The lack of exothermic signal in the DSC measurement suggests that this hydrogen release is not due to a one-time irreversible reaction with oxygen impurities in the carbon framework.Instead it suggests that the hydrogen pressure applied at the temperature used for rehydriding was insufficient to fully reverse all of the hydrogen desorption reactions.In situ SR-PXD of the nanoconfined sample suggests that the first hydrogen release event on the first desorption cycle is due to residual unreacted LiBH 4 from incomplete metathesis (reaction 4).Once the remaining LiBH 4 has decomposed, it is not reformed under the applied hydriding conditions, thus it is not observed during the second desorption.Based on the in situ SR-PXD on the bulk sample, the compounds that are not regenerated under the applied temperatures and hydrogen pressures are possibly LiAlH 4 and/or Li 3 AlH 6 [37].
FT-IR measurements on selected samples are presented in Figure 5. Bulk 2LiBH 4 -NaAlH 4 demonstrates multiple peaks in the fingerprint region (500-1500 cm ´1).The three signals at 1095, 1241 and 1315 cm ´1, corresponds to the characteristic B-H bending bands of LiBH 4 .The signal at 1626 cm ´1 is assigned to the [AlH 4 ] ´stretching band of NaAlH 4 and the large broad signal ranging from 2000 to 2500 cm ´1, with a maximum at 2307 cm ´1, originates from the B-H stretch of LiBH 4 .Furthermore, the small signal at 3444 cm ´1, is assigned to -OH, most likely from exposure to atmospheric moisture moist during measurement.It is worth noting that moisture and air cannot be completely avoided during collection of FT-IR data using the selected apparatus.A completely dehydrogenated sample of melt infiltrated 2LiBH 4 -NaAlH 4 in CA, i.e., heated at 500 ˝C for 10 h shows a distinct signals at 1100 and 2462 cm ´1.These positions are close to those reported for Li 2 B 12 H 12 [38] and are nearly identical to positions reported by Mao et al. [39] after the decomposition of NaBH 4 which suggests that one of the decomposition products may be Na 2 B 12 H 12 .Rehydrogenation of the decomposed sample after three desorption cycles at 400 ˝C for 10 h, show B-H stretching signals at 2113 and 2283 cm ´1 and the respective B-H bending.Furthermore, signals at 1234 and 1080 cm ´1 are observed in the fingerprint region.These results may suggest that the 1234 and 2283 can be assigned to NaBH 4 in combination with another unknown complex boron hydride.It is difficult to make any statement about the possible presence of NaAlH 4 signals due to the high amount of carbon and low total quantity of hydrides in the sample.Based on these results it can be suggested that the partial reversibility of the system is due to the formation and decomposition of NaBH 4 inside the scaffold.

Conclusions
Nanoconfinement of 2LiBH4-NaAlH4 was obtained via melt infiltration under H2 pressure and the reversibility and the reaction mechanisms during infiltration and decomposition are investigated.Nanoconfinement of 2LiBH4-NaAlH4 into a mesoporous carbon aerogel scaffold significantly enhances the kinetics for hydrogen desorption, the hydrogen storage capacity and reversibility during hydrogen release and uptake cycling of 2LiBH4-NaAlH4, compared to that of bulk.The temperature of the maximum hydrogen desorption rate is reduced by 132 °C when employing the carbon scaffold, which is assigned to the effect induced by nanoconfinement and the carbon surface acting as a catalyst for hydrogen release on 2LiBH4-NaAlH4.The stability during cycling is significantly enhanced by melt infiltration of the hydride.Bulk 2LiBH4-NaAlH4 releases 81% of the available hydrogen content during the first desorption, compared to 96% when nanoconfined.During the third desorption cycle of the bulk sample, only 47% of the initial hydrogen content is retained compared to 83% for that of the nanoconfined sample.These results demonstrate significant stabilization of the cyclic hydrogen storage capacity of metal borohydrides compared to those previously presented.Fu et al. [40] reports elemental analysis of carbon aerogel prepared using alternative reagents but similar pyrolysis conditions.Analysis of carbon and oxygen concentration of the carbon aerogels after pyrolysis reveal C and O are 94.00 wt.% and 5.87 wt.%, respectively (re-calculated by excluding K).Assuming similar oxygen levels in our scaffolds and a 1:1 reaction with LiH would reduce the released amount of hydrogen to 2.066 wt.% H 2 for the nanoconfined sample.

Conclusions
Nanoconfinement of 2LiBH 4 -NaAlH 4 was obtained via melt infiltration under H 2 pressure and the reversibility and the reaction mechanisms during infiltration and decomposition are investigated.Nanoconfinement of 2LiBH 4 -NaAlH 4 into a mesoporous carbon aerogel scaffold significantly enhances the kinetics for hydrogen desorption, the hydrogen storage capacity and reversibility during hydrogen release and uptake cycling of 2LiBH 4 -NaAlH 4 , compared to that of bulk.The temperature of the maximum hydrogen desorption rate is reduced by 132 ˝C when employing the carbon scaffold, which is assigned to the effect induced by nanoconfinement and the carbon surface acting as a catalyst for hydrogen release on 2LiBH 4 -NaAlH 4 .The stability during cycling is significantly enhanced by melt infiltration of the hydride.Bulk 2LiBH 4 -NaAlH 4 releases 81% of the available hydrogen content during the first desorption, compared to 96% when nanoconfined.During the third desorption cycle of the bulk sample, only 47% of the initial hydrogen content is retained compared to 83% for that of the nanoconfined sample.These results demonstrate significant stabilization of the cyclic hydrogen storage capacity of metal borohydrides compared to those previously presented.

3. 1 .
In Situ Synchrotron Radiation-Powder X-Ray Diffraction of Bulk and Nanoconfined 2LiBH 4 -NaAlH 4 The decomposition of physically mixed bulk 2LiBH 4 -NaAlH 4 (LiNa) has been investigated using in situ SR-PXD during heating up to 487 ˝C, see Figure 1.At RT a physical mixture of the low temperature polymorph orthorhombic LiBH 4 , o-LiBH 4 , and NaAlH 4 are present in the sample.The polymorphic transition from o-LiBH 4 to hexagonal h-LiBH 4 takes place in the temperature range 98 to 105 ˝C.A metathesis reaction between LiBH 4 and NaAlH 4 occurs at 116 ˝C observed by formation of NaBH 4 .
cell parameters.During natural cooling of the sample, thermal contraction is observed along with formation of Al, LiAl3, and NaBH4 along with an unknown compound.A difference plot of the in situ SR-PXD data is shown in supporting information to accentuate changes in weaker peaks.

Figure 2 .
Figure 2. (A) Thermal programmed desorption mass spectroscopy (TPD-MS) profiles of the hydrogen release (H 2 + ions, m/e = 2); (B) thermogravimetric analysis (TGA); and (C) differential scanning calorimetry (DSC) measurements of bulk 2LiBH 4 -NaAlH 4 (LiNa) (red), LiNa mixed with carbon aerogel scaffold (CA) (green) and LiNa melt infiltrated into CA (blue).The data is measured in the temperature range 40 to 550 ˝C and TPD-MS for each data set is normalized by its maximum hydrogen release rate to allow for qualitative comparison (∆T/∆t = 10 ˝C/min).

Figure 3 .
Figure 3. Sieverts' measurement showing four and three hydrogen release cycles for (A) bulk 2LiBH 4 -NaAlH 4 and (B) infiltrated into CA, LiNa-CA.Hydrogen desorption was performed at a fixed temperature of 500 ˝C (∆T/∆t = 5 ˝C) for 10 h under an initial hydrogen pressure of 1 bar.Hydrogen absorption was performed at 400 ˝C, p(H 2 ) = 126 bar for 10 h.

Figure 3 .
Figure 3. Sieverts' measurement showing four and three hydrogen release cycles for (A) bulk 2LiBH4-NaAlH4 and (B) infiltrated into CA, LiNa-CA.Hydrogen desorption was performed at a fixed temperature of 500 °C (ΔT/Δt = 5 °C) for 10 h under an initial hydrogen pressure of 1 bar.Hydrogen absorption was performed at 400 °C, p(H2) = 126 bar for 10 h.

Crystals 2016, 6 , 70 10 of 12 Figure 5 .
Figure 5. FTIR spectra of bulk LiBH4, 2LiBH4-NaAlH4 (LiNa), and nanoconfined 2LiBH4-NaAlH4 CA-LiNa after dehydrogenation at 500 °C for the third time and finally CA-LiNa rehydrogenated at p(H2) = 140 bar, T = 400 °C for 10 h after three desorption cycles.The intensity of the spectra with the carbon containing samples was scaled corresponding to their quantity of LiNa to allow for better comparison.

Figure 5 .
Figure 5. FTIR spectra of bulk LiBH 4 , 2LiBH 4 -NaAlH 4 (LiNa), and nanoconfined 2LiBH 4 -NaAlH 4 CA-LiNa after dehydrogenation at 500 ˝C for the third time and finally CA-LiNa rehydrogenated at p(H 2 ) = 140 bar, T = 400 ˝C for 10 h after three desorption cycles.The intensity of the spectra with the carbon containing samples was scaled corresponding to their quantity of LiNa to allow for better comparison.

Table 1 .
Texture parameters of as prepared carbon aerogel scaffold (CA).The gravimetric hydrogen content of the confined sample is also provided.