Unique Hydrogen Desorption Properties of LiAlH 4 / h-BN Composites

Hexagonal boron nitride (h-BN) is known as an effective additive to improve the hydrogen de/absorption properties of hydrogen storage materials consisting of light elements. Herein, we report the unique hydrogen desorption properties of LiAlH4/h-BN composites, which were prepared by ball-milling. The desorption profiles of the composite indicated the decrease of melting temperature of LiAlH4, the delay of desorption kinetics in the first step, and the enhancement of the kinetics in the second step, compared with milled LiAlH4. Li3AlH6 was also formed in the composite after desorption in the first step, suggesting h-BN would have a catalytic effect on the desorption kinetics of Li3AlH6. Finally, the role of h-BN on the desorption process of LiAlH4 was discussed by comparison with the desorption properties of LiAlH4/X (X = graphite, LiCl and LiI) composites, suggesting the enhancement of Li ion mobility in the LiAlH4/h-BN composite.


Introduction
Hydrogen storage is a key technology for a future hydrogen energy society [1]. However, it is still challenging to develop high performance hydrogen storage materials with high hydrogen density, fast de/absorption kinetics, and high cycle stability under moderate temperature and pressure conditions [2,3]. LiAlH4 is one of the most promising hydrogen storage materials because of its high hydrogen capacity and relatively low desorption temperature [4]. The hydrogen desorption process of LiAlH4 is described as follows: Melting: LiAlH4(s) → LiAlH4(l) Endothermic (150-175 °C) Decomposition in the first step: 3LiAlH4(l) → Li3AlH6(s) + 2Al + 3H2 Exothermic (150-200 °C, 5.3 mass % H2) Decomposition in the second step: Li3AlH6(s) → 3LiH + Al + 3/2H2 Endothermic (200-270 °C, 2.6 mass % H2) The decomposition in the first step is an exothermic reaction with a ∆H of −10 kJ·mol −1 H2, indicating the reversibility of this step is believed to be thermodynamically difficult [4,5]. In the second step, Li3AlH6 decomposes in an endothermic reaction with a ∆H of 25 kJ·mol −1 H2 [4]. Thus, the hydrogenation of LiH/Al to Li3AlH6 is thermodynamically possible.
One of the strategies for improving the properties of hydrogen storage materials is the addition of catalysts/dopants [6]. Ti or its compounds are well-known catalysts for the kinetics of alanate [7][8][9]. Since Bogdanović et al. reported an absence of hysteresis and nearly horizontal pressure plateaus in the TiCl3-doped NaAlH4 [7], many researchers have studied complex hydrides including alanate as potential reversible hydrogen storage materials. In the case of LiAlH4, the improved desorption kinetics was reported by the doping of Ti catalyst using mechanically milling [10][11][12].
Recent study also reported a single step hydrogen release of LiAlH4, which was induced by the synergetic effects of Ti catalytic coating and nanosizing effects [13]. Although the rehydrogenation of the desorbed material was not achieved in the milled sample, the regeneration of LiAlH4 from LiH and Ti-catalyzed Al was possible through the solution synthesis approach using THF and Me2O [14][15][16].
Hexagonal boron nitride (h-BN) is known as an effective additive for chemical hydride and complex hydride systems. For instance, NH3BH3/h-BN composite released hydrogen at low temperature with minimum induction time and less exothermicity [17]. The remarkable hydrogen de/absorption properties were also achieved in the milled LiBH4/h-BN composites [18,19]. The 30 mol % h-BN doped LiBH4 composite started to release hydrogen from 180 °C, which was 100 °C lower than the onset hydrogen desorption temperature of ball-milled LiBH4 [18]. For the 75 mol % h-BN doped LiBH4 composite, the on-set desorption temperature of LiBH4 was reduced to 175 °C and the peak desorption temperature was reduced by 80 °C compared with milled LiBH4 [19]. Furthermore, under moderate rehydrogenation conditions of 400 °C and 10 MPa H2 pressure, the dehydrogenation capacity of the composite maintained 3.1 mass % within three cycles, which was very close to its theoretical capacity. It was assumed that the excellent rehydrogenation property of LiBH4 would be related with the enhanced hydrogen and lithium diffusion capability by the nanoscale h-BN, which was synthesized by ball-milling of h-BN at 490 rpm for 20 h [19]. The enhancement of Li + and/or H − diffusion by adding h-BN was firstly reported in the LiNH2/LiH system [20]. Hydrogen was fully desorbed from the LiNH2/LiH/h-BN composite in less than 7 h, whereas the LiNH2/LiH composite desorbed hydrogen in several days. They proposed that h-BN is an efficient catalyst that improves Li + diffusion and hence the kinetics of the reaction between LiNH2 and LiH [20]. The mobility of Li + ions between LiH and LiNH2 was also enhanced by adding LiTi2O4 catalyst [21].
Thus, h-BN has attracted much attention as an effective additive to improve the hydrogen de/absorption kinetics of hydrogen storage materials, especially for complex hydrides. However, the addition of h-BN to the alanate system has rarely been reported. In the present study, LiAlH4/h-BN composites were synthesized by planetary ball-milling and their hydrogen desorption properties were analyzed. Also, the desorption process of the composite was investigated by using XRD and FT-IR. Finally, the role of h-BN to the desorption properties of LiAlH4 was discussed by comparison with those of LiAlH4/X (X = graphite, LiCl and LiI) composites.

Results
Hydrogen desorption properties of LiAlH4/h-BN composites were analyzed by using TG-DTA-MS. Figure 1 shows the DTA and MS (H2, m/z = 2) profiles of the composites. As shown in Figure 1a, ball-milled LiAlH4 (denoted as 0 mass % in Figure 1) started to melt around 150 °C followed by hydrogen desorption in two steps below 250 °C. In the case of LiAlH4/h-BN composites, desorption profiles were clearly changed compared with LiAlH4. First, the melting temperature (Tm) of LiAlH4 was decreased by adding h-BN. For instance, in the 40 mass % h-BN composite, DTA peak value of Tm was 151 °C, which was 11 °C lower than that of milled LiAlH4. Second, the hydrogen desorption temperature (Td) in the first step was slightly increased by h-BN addition. As shown in Figure 1b, Td in the first step became high value with the increasing amount of h-BN. Third, the desorption kinetics in the second step was improved by adding h-BN. As shown in Figure 1b, the desorption peak in the second step became sharp as the amount of h-BN increased up to 14 mass %. However, the peak shape became broad in the 40 mass % h-BN composite, suggesting the addition of too much amount of h-BN could have negative effect on improving the kinetics. Figure 2 shows the TG profiles of LiAlH4/h-BN composites. The total mass loss from TG profile of ball-milled LiAlH4 was 7.7 mass %, which was in good agreement with the theoretical hydrogen desorption amount of LiAlH4 (7.9 mass %) [4]. The hydrogen mass loss of 6.9 mass %, 5.8 mass %, and 3.5 mass % were calculated from the profiles of 4 mass %, 14 mass %, and 40 mass % h-BN composites, respectively. It is noted that theoretical hydrogen desorption capacities of these composites were 7.6 mass %, 6.8 mass %, and 4.7 mass %, respectively, when only considering the hydrogen desorption from LiAlH4. Thus, the experimental values of hydrogen desorption amounts were slightly lower than the theoretical values. This result could originate from the hydrogen desorption during ball-milling or the formation of new H-containing solid compound by the reaction between LiAlH4 and h-BN.  To investigate the interaction between LiAlH4 and h-BN, XRD, and FT-IR measurements were performed for the LiAlH4/h-BN composites. Also, the particle size of composite was observed by using SEM and TEM. Figure 3a shows the XRD profiles of the milled LiAlH4 and LiAlH4/h-BN composites. Only LiAlH4 and h-BN phases were observed in the profiles of the composites. Although the diffraction peaks of LiAlH4 were slightly broadened by h-BN addition, the clear relationship between the broadning and the amount of h-BN was not observed. Figure 3b,c shows the XRD profiles of the milled LiAlH4 and 40 mass % h-BN composite after hydrogen desorption. The 40 mass % h-BN composite formed the similar reaction products compared with LiAlH4. In other words, Li3AlH6 and Al were formed after the hydrogen desorption in the first step, and LiH and Al were formed after the second step. The phase of h-BN was also clearly observed after hydrogen desorption. Broad diffraction peaks around 20° and 27° originate from the polyimide film and grease to prevent the sample oxidation. Figure 4 shows SEM and TEM images of 40 mass % h-BN composite and references. As shown in Figure 4a, the milled LiAlH4 contained a lot of large particles with sizes over 10 µm. On the other hand, the 40 mass % h-BN composite showed the average particle size of a few micrometers, indicating the refinement of LiAlH4 particles occurred in the composite. The submicron particles were also observed in the TEM image of the composite, as shown in Figure 4d. The size of as-received h-BN particle was around 1 µm (Figure 4b). Figure 5 shows the FT-IR spectra of 40 mass % h-BN composite and references. The as-milled composite showed the characteristic Al-H vibrations of LiAlH4 [22] around 1795 cm −1 and 1644 cm −1 . Also, B-N vibrations of h-BN around 1373 cm −1 and 818 cm −1 were observed. Although the Al-H vibrations of LiAlH4 disappeared after heating, the B-N vibrations of h-BN still remained. These results were consistent with the results of XRD. However, new IR absorption peak was clearly observed around 2300 cm −1 after heating up to 183 °C and 300 °C. Also, another new peak appeared around 1100 cm −1 after heating up to 300 °C. Although these peaks were not identified in this work, the peak positions were similar to those of LiBH4 [23], suggesting a such kind of Li-B-H phase exist after heating up to 183 °C and/or 300 °C. The unknown peak around 2300 cm −1 was also observed for the IR spectra of the composite consisting of BN nano Hx (ball-milled h-BN under 1.0 MPa H2 for 80 h) and LiH [24]. Thus, the formation of new H-containing solid compound could result in the slightly low hydrogen desorption amount in Figure 2. Considering the results of Figure 5, the possible new compound could be covalently functionalized h-BN species. The details were explained in the next discussion part.   where Ea is the apparent activation energy for hydrogen desorption, c is the heating rate, Tp is the peak temperature, R is gas constant, and A is the frequency factor. Figure 6 shows the Kissinger plots for the hydrogen desorption in the second step of 4 mass % h-BN composite. The obtained apparent activation energy, Ea was 71.5 kJ·mol −1 . This value is lower than that reported for Li3AlH6 (92 kJ·mol −1 ) [26], indicating the desorption kinetics was improved by adding h-BN.   Figure 7 shows the DTA and MS profiles of LiAlH4/X (X = graphite, LiCl, and LiI) composites. The profiles of LiAlH4 and LiAlH4/h-BN composites are also shown as the references. Graphite was selected as an additive because this compound has a structure similar to h-BN. In spite of its similar structure, hydrogen desorption properties were different from those of h-BN. In the case of LiAlH4/graphite composite, the melting temperature was similar to ball-milled LiAlH4. Also, the desorption kinetics in the second step seemed to be delayed, whereas that of LiAlH4/h-BN was enhanced. Since the both graphite and h-BN are hard materials, the refinement of LiAlH4 particles would occur during the ball-milling process. However, the different desorption profiles were obtained between h-BN and graphite composite, suggesting just the refinement of particles cannot explain this difference. Although the desorption properties of nanoconfined LiAlH4 into graphite with high surface area was reported in the previous study [27], the profiles of graphite composite in this study seemed to be different from those profiles.

Discussion
Also, LiCl and LiI were selected as additives. Aguey-Zinsou et al. reported that h-BN would enhance the Li ion mobility across the interface of LiNH2 and LiH [20]. Thus, the enhancement of Li ion mobility could be one of the reasons for the unique hydrogen desorption properties of LiAlH4/h-BN composites. Oguchi et al. reported the Li ion conductivity of Li3AlH6/LiI composite at 120 °C was much higher than that of Li3AlH6, but that of Li3AlH6/LiCl at 120 °C was the similar value compared with Li3AlH6 [28]. Their results suggest that LiI additive would be effective for increasing the Li ion mobility of LiAlH4, but LiCl would not be effective near the decomposition temperature range of LiAlH4. As shown in Figure 7, the desorption profiles of LiCl composite was similar to those of milled LiAlH4. On the other hand, those of LiI composite showed the decrease of melting temperature and the delay of the first desorption reaction, which was consistent with the results of h-BN composite. This comparative result suggests that Li ion conductivity would increase in the LiAlH4/h-BN composite. The kinetics in the second step wasdelayed in the LiI composite. The possible origin of the high conductivity of LiAlH4 is the anion substitution from complex anion to I − [28], which partially took place in the case of LiBH4 [29]. Thus, the high decomposition temperature in the second step could originate from the formation of stable solid solution similar to LiBH4-LiI(LiCl) system [29][30][31][32][33]. For clarifying the total desorption process of these composites, the detailed mechanistic study is needed. The analysis of Li ion conductivity of LiAlH4/h-BN composite is currently in progress.
As shown in Figure 1, all the LiAlH4/h-BN composites showed the melting of LiAlH4. This phenomenon was different from the case of LiAlH4 catalyzed by transition metal (Ti, Fe, Co, Nb, etc. [34][35][36]), which can release hydrogen below the melting temperature. Thus, h-BN would have little interaction with complex anion of [AlH4] − , whereas transition metal like Ti would destabilize the covalent Al-H bond of complex anion as Sandrock et al. suggested [8]. According to the proposed mechanism by Atakli et al. [37], Li3AlH6 is formed by transferring the alkali cation and a hydrogen anion from the two neighboring alanate molecules to central one. In this context, the diffusion distance of Li + seems to be very short to form Li3AlH6, suggesting Li + diffusion would not be the ratelimiting step in the desorption of LiAlH4 in the first step. On the contrary, the desorption properties of h-BN and LiI composite suggested the excess enhancement of Li + ion mobility might be related to delaying the formation of Li3AlH6. First-principle DFT studies also suggested the formation and migration of [AlH4] − vacancy would be the rate-limiting step in the decomposition of LiAlH4 [38]. In the second decomposition step, it was proposed that three LiH are formed from Li3AlH6, leaving AlH3, which spontaneously desorbs hydrogen [37]. Thus, the enhanced mobility of Li + may help the destabilization of complex anions to improve the hydrogen desorption kinetics.
As shown in Figure 5, the presence of new bonds similar to those of LiBH4 or ball-milled h-BN with LiH suggests that covalently functionalized h-BN species could be formed after heating the LiAlH4/h-BN composites. It is known that ball-milling of h-BN with a lot of different materials can generate functionalized h-BN nanosheets [39]. For instance, a one-step method for the preparation and functionalization of few-layer BN was developed based on urea-assisted solid exfoliation of commercially available h-BN [40]. Such kind of functionalized BN nanosheets are attractive for a lot of applications such as polymer matrix composites [41], ion conductors [42], and hydrogen storage [43]. Further investigations of LiAlH4/h-BN composites could pave the way for the covalent functionalization of h-BN nanosheets by interaction with LiAlH4.

Characterization
Hydrogen desorption properties of the composites were examined by a thermogravimetry and differential thermal analysis equipment (TG-DTA, Bruker, 2000SA, Yokohama, Japan) connected to a mass spectrometer (MS, ULVAC, BGM-102, Chitose, Japan). The desorbed gases were carried from TG-DTA to MS through a capillary by 300 mL·min −1 stream of high purity He as a carrier gas. The samples were heated from room temperature to 300 °C with a heating rate of 5 °C·min −1 . Structural properties were investigated by powder X-ray diffraction (XRD) measurements (Philips, X'Pert Pro with Cu Kα radiation, Amsterdam, The Netherlands), where all the samples were covered with a polyimide sheet (Kapton, The Nilaco Co., Ltd., Tokyo, Japan) in the glovebox to avoid oxidation during the measurement. Morphology of the composites was observed using scanning electron microscope (SEM, JEOL, JSM-6510LA, Tokyo, Japan) and transmission electron microscope (TEM, JEOL, JEM-2010, Tokyo, Japan). For TEM observations, the samples were dispersed on a molybdenum micro-mesh grid. Fourier transform infrared spectrometer (FT-IR, JASCO, FT/IR 660Plus, Tokyo, Japan) was operated to investigate chemical bonds in the composites. Each sample was put between KBr plates and pressed for measurement.

Conclusions
Hydrogen desorption properties of the ball-milled LiAlH4/h-BN composites were investigated. Compared with milled LiAlH4, the composites showed the different desorption profiles, where the decrease of melting temperature (Tm), the delay of desorption kinetics in the first step and the enhancement of the kinetics in the second step were observed. In the 40 mass % h-BN composite, the DTA peak value of Tm was 151 °C, which was 11 °C lower than that of milled LiAlH4. The LiAlH4/h-BN composite formed Li3AlH6 after desorption in the first step similar to LiAlH4. Thus, h-BN would have a catalytic effect on the desorption kinetics of Li3AlH6. The apparent activation energy in the second step desorption was 71.5 kJ·mol −1 for the 4 mass % h-BN composite. From SEM and TEM observations, the refinement of LiAlH4 particle was confirmed in the 40 mass % h-BN composite. The particle size of the composite was around a few micrometers. The hydrogen mass loss of the composite was slightly lower than the theoretical value. The new chemical bond similar to Li-B-H species was observed in the FT-IR spectra of the 40 mass % h-BN composite after the desorption. This result suggested covalently functionalized h-BN nanosheets could be formed in the composite. Finally, the desorption properties of LiAlH4/h-BN composite were compared with those of LiAlH4/X (X = graphite, LiCl, and LiI) composites, suggesting the enhancement of Li ion mobility in the LiAlH4/h-BN composite compared with LiAlH4. The present work first demonstrates the effect of h-BN addition on the hydrogen desorption properties of alanate. the "Material Analysis and Structure Analysis Open Unit (MASAOU)" and "Nanotechnology Platform" program of the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan.
Author Contributions: Yuki Nakagawa was involved in all stages of the work, including designing the work, conducting experiments, and analyzing the data; Shigehito Isobe acted as co-supervisor and was involved in the discussion of the results and work planning; Takao Ohki performed the TG-DTA-MS and XRD experiments; Naoyuki Hashimoto acted as the main supervisor and was involved in the discussion of the results; Yuki Nakagawa wrote the paper; and all the authors contributed to the revision of the paper.

Conflicts of Interest:
The authors declare no conflict of interest.