Thermal Conversion of Unsolvated Mg(B3H8)2 to BH4– in the Presence of MgH2

In the search for energy storage materials, metal octahydrotriborates, M(B3H8)n, n = 1 and 2, are promising candidates for applications such as stationary hydrogen storage and all-solid-state batteries. Therefore, we studied the thermal conversion of unsolvated Mg(B3H8)2 to BH4– as-synthesized and in the presence of MgH2. The conversion of our unsolvated Mg(B3H8)2 starts at ∼100 °C and yields ∼22 wt % of BH4– along with the formation of (closo-hydro)borates and volatile boranes. This loss of boron (B) is a sign of poor cyclability of the system. However, the addition of activated MgH2 to unsolvated Mg(B3H8)2 drastically increases the thermal conversion to 85–88 wt % of BH4– while simultaneously decreasing the amounts of B-losses. Our results strongly indicate that the presence of activated MgH2 substantially decreases the formation of (closo-hydro)borates and provides the necessary H2 for the B3H8-to-BH4 conversion. This is the first report of a metal octahydrotriborate system to selectively convert to BH4– under moderate conditions of temperature (200 °C) in less than 1 h, making the MgB3H8-MgH2 system very promising for energy storage applications.


■ INTRODUCTION
To reach a widespread implementation and integration of renewable energy systems, the development of new energy storage materials is crucial to mitigate the intermittency and unpredictability of renewable energy sources, such as wind or sun, thereby leading to a more reliable power grid. Boron hydrogen (B x H y ) compounds are a unique class of materials due to their three-center two-electron (3c-2e) bonding providing a variety of materials promising for energy storage applications. 1−5 Recently, this class of compounds has been demonstrated to be promising candidates as solid-state electrolytes for all-solid-state batteries. 6−10 For example, KB 3 H 8 was reported to reach a maximum conductivity of 3.4 × 10 −7 S cm −1 at 150°C, 11 and that of unsolvated Mg(B 3 H 8 ) 2 reached 1.4 × 10 −4 S cm −1 at 80°C. 12 Compounds based on the B 3 H 8 − anion have also been studied as nonflammable hydrogen carriers at room temperature in order to address the flammability issues related to the transport of gasoline. Borohydride-based ionic liquids, such as C(NH 2 ) 3 B 3 H 8 and Li(NH 3 )B 3 H 8 , are considered due to their melting point less than −10°C and their liquid nature at room temperature. While C(NH 2 ) 3 B 3 H 8 releases ∼7 mol % of H 2 at 83°C, Li(NH 3 )B 3 H 8 releases ∼95 mol % of H 2 at 133°C along with the liberation of minor B 2 H 6 and B 5 H 9 amounts. 13,14 By adding metal chlorides, MCl x (x = 2−4, M = Al, Zn, Zr), the H 2 evolution of the Li(NH 3 )B 3 H 8 was improved to ∼97, 98, and 96 mol % at 109−119°C. 14 B 3 H 8 − compounds, such as NaB 3 H 8 (13 wt % H) and KB 3 H 8 (5 wt % H), have been also recently studied for long-term solid-state hydrogen storage. 11,15−17 In the case of NaB 3 H 8 , it has a high solubility in water (74 wt %) at room temperature; therefore, possibilities exist for it to be used as a liquid hydrogen carrier in solution. Huang et al. found that the use of CoCl 2 can act as a catalyst to enhance the hydrogen evolution up to ∼7 wt % in the case of NaB 3 H 8 . 18 Compounds based on B 3 H 8 − are also potential materials in preparing metal borides (MB x , x = 1, 2) with chemical vapor deposition methods. For instance, Cr(B 3 H 8 ) 2 , Mg(B 3 H 8 ) 2 (Et 2 O) 2 , and Mg(B 3 H 8 ) 2 (MeO) 2 were used to synthesize CrB 2 and MgB 2 . However, the evolution of pentaborane (B 5 H 9 ) from heating these precursors affects the purity of the deposited metal boride films. 19,20 Mg(B 3 H 8 ) 2 was first observed experimentally as a dehydrogenation intermediate of magnesium borohydride, Mg(BH 4 ) 2 , at 200°C in vacuo after 5 weeks. 21 This result pioneered research initiatives to determine the reversible re-hydrogenation of Mg(B 3 H 8 ) 2 to Mg(BH 4 22 The authors found that the system Mg(B 3 H 8 ) 2 · 2THF-MgH 2 yielded a 9 mol %-conversion to BH 4 − at 200°C under 50 bar of H 2 . However, the role of MgH 2 and THF in the hydrogenation of Mg(B 3 H 8 ) 2 ·2THF-MgH 2 was not clarified. The finding boosted the development of synthetic procedures for B 3 H 8 − compounds and the study of its properties. Initially, synthetic routes of M(B 3 H 8 ) n , n = 1 and 2, involved the use of toxic chemicals (e.g., Hg, B 2 H 6 , B 4 H 10 , and BF 3 O(C 2 H 5 ) 2 ), 23−30 while the more recent synthetic strategies reported are solvent-free for M(B 3 H 8 ) n with n = 1 and 2 and M = Mg, Li, Na, K, Rb, and Cs. 12 12 This finding will open more opportunities in studying unsolvated Mg(B 3 H 8 ) 2 and its chemical and physical properties in view of energy storage materials.
One of the greatest challenges encountered by these materials is the ability to de-/rehydrogenate because (i) they lose boron in the form of gas-phase boranes in their thermal decomposition, (ii) they are not selective in their decomposition products as species like B 10 H 10 2− are produced, and (iii) they form solid-state species regarded as thermodynamic sinks, such as B 12 H 12 2− . For example, the thermal decomposition of KB 3 H 8 occurs in the temperature range of 150−250°C with a mass loss of 11 wt %, with the generation of several gaseous by-products such as B 2 H 6 , B 5 H 9 , and B 6 H 10 and solid products B 12 H 12 2− , B 10 H 10 2− , and BH 4 − . 17 To date, the reported hydrogen cycling capacity of Mg(B 3 H 8 ) 2 formed from dehydrogenation of Mg(BH 4 ) 2 is ∼2.5 wt % due to the formation these by-products. 21,36 In solid-state batteries, it has recently been shown that the cation conductivity is enabled by the high mobility of the B 3 H 8 − anion. 12,37 During discharging of the battery, the B 3 H 8 − forms diborane and pentaborane, preventing sufficient discharge/charge cycles to be considered in an application.
With the intent of facilitating the hydrogenation of metal octahydrotriborates, KH was added to KB 3 H 8 , 11  . 40,41 In this work, our approach adopted the addition of MgH 2 to unsolvated Mg(B 3 H 8 ) 2 as a strategy to substantially retain the boron species of unsolvated Mg(B 3 H 8 ) 2 to address the cyclability issues of these materials. Simultaneously, this work shows that the Mg(B 3 H 8 ) 2 -MgH 2 system yields a very high B 3 H 8 − to BH 4 − conversion of 85−88 wt % at 200°C in 1 h, without the supply of excess hydrogen gas. This result is currently the highest conversion to BH 4 − observed in the hydrogenation of all metal-octohydrotriboranes compounds, and it is the first example reported of the conversion to BH 4 − , without the supply of additional hydrogen.
The use of MgH 2 as a reactive hydride composite has been reported earlier. 42,43 In these systems, two hydrides are combined with each other in order to add an exothermic reaction to the endothermic hydrogen desorption reaction to form a new compound, thereby lowering the overall reaction enthalpy to release hydrogen. In the case of MgH 2 and LiBH 4 , MgB 2 (Δ f H°= −91.96 kJ/mol) is formed. 44 In this work, our approach adopted the addition of MgH 2 to unsolvated Mg(B 3 H 8 ) 2 as a strategy to substantially retain the boron species of unsolvated Mg(B 3 H 8 ) 2 to address the cyclability issues of these materials. The new and pertinent result presented here is that MgH 2 , which has been added to keep a constant Mg/B ratio as in Mg(BH 4 ) 2 , does not catalyze the further decomposition of Mg(B 3 H 8 ) 2 but rather favors the rehydrogenation reaction and prevents the release of toxic boranes. In this case, the thermodynamic driving force is the formation of Mg(BH 4 ) 2 , which has a formation enthalpy of Δ f H°= −208 kJ/mol. 45 This raises the possibility that excess MgH 2 may also favor the rehydrogenation of other compounds of the form MgB x H y .
Because we study the thermal decomposition and the products formed upon the decomposition process, the terms "decomposition" and "conversion" will be used interchangeably in this publication. To demonstrate these results, we first characterize the de-/rehydrogenation products of unsolvated Mg(B 3 H 8 ) 2 , with a gravimetric capacity of 15 wt % developed from a novel synthesis involving tetra-alkylammonium salts of B 3 H 8 − . 12 The de-/rehydrogenation of our unsolvated Mg-(B 3 H 8 ) 2 in the presence of MgH 2 is studied with and without pressurizing the system to 6 bar of H 2 . The dominant characterization techniques used in this work are 11 B solid-state magic angle spinning (MAS) NMR to determine the B-species present in the samples, temperature programmed desorption coupled with mass spectrometry (TPD-MS) to follow the evolution of volatile borane species, and X-ray absorption near edge structure (XANES) measurements for the identification and quantification of Mg-and B-species formed during the dehydrogenation.

■ EXPERIMENTAL DETAILS
All preparations of materials/samples for analysis were performed in an argon (Ar) filled glovebox (H 2 O < 0.1 ppm, O 2 < 0.1 ppm). Three stainless steel balls with 10 mm diameters (ball-to-reactant mass ratio ∼40) were placed in a 65 mL airtight milling vessel for the synthesis of unsolvated Mg(B 3 H 8 ) 2 33 The details of the synthesis of NaB 3 H 8 are reported in the Supporting Information. The assynthesized NaB 3 H 8 was mechanochemically converted into unsolvated Mg(B 3 H 8 ) 2 using the corresponding bromide salt as described by the following reaction: 2NaB 3 H 8 + MgBr 2 → Mg(B 3 H 8 ) 2 + 2NaBr.
Unsolvated Mg(B 3 H 8 ) 2 was synthesized by milling 0.2020 g of NaB 3 H 8 with 0.3125 g of MgBr 2 in air-tight milling vessels. The vessels were filled in an Ar glovebox and then milled for 1 h at 200 rpm (15 min milling and 5 min break). Note about the compound unsolvated Mg(B 3 H 8 ) 2 is highly pyrophoric and it must be kept away from water. It is highly recommended to store it in an inert glovebox environment.
The activation of MgH 2 was achieved by ball milling. Before the synthesis of Mg(B 3 H 8 ) 2 -MgH 2 , 0.2000 g of commercial MgH 2 was introduced in an airtight vessel and milled for 2 h at 450 rpm (5 min milling, 1 min break). The activated MgH 2 was stored in the glovebox. Mg(B 3 H 8 ) 2 -4MgH 2 was prepared by mixing 0.0400 g of Mg-(B 3 H 8 ) 2 with 0.0135 g of MgH 2 in an airtight milling vessel according to the following parameters: 35 min at 300 rpm (5 min milling and 1 min break). 0.0400 g of Mg(B 3 H 8 ) 2 included the reaction product NaBr and residual MgBr 2 from the synthesis. The synthetized product was stored in the glovebox. Throughout this manuscript, we will use the nomenclature "Mg(B 3 H 8 ) 2 -MgH 2 " to refer to the material obtained by a physical mixture of Mg(B 3 H 8 ) 2 and 4 weight equivalents of MgH 2 . The details about the cleaning of the jars after the activation of MgH 2 and the synthesis of Mg(B 3 H 8 ) 2 -MgH 2 are provided in Section S1.
Material Characterization. 11 B Solid-State MAS NMR. 11 B solidstate MAS NMR experiments were performed in an Oxford Instruments 11.7 T widebore magnet, an Agilent VNMRS spectrometer, and a homebuilt 5 mm HX probe (ceramic MAS housing from Revolution NMR) in the EMSL user facility at Pacific Northwest National Laboratory, USA. 1 H and 11 B signals were referred to tetramethylsilane at 0 ppm and boric acid 1 M at 20 ppm and measured at 500.1 and 160.4 MHz, respectively. The RF fields for direct polarization were a π/20 pulse of 0.26 μs for 11 B and a 1 H decoupling field of 42 kHz utilizing a SPINAL-16 scheme. 46 The rotors are 5 mm cavern-style zirconia sleeves (Revolution NMR) that have been modified to accommodate a double O-ring Vespel bushing to seal the sample up to 225 bar at 250°C that were previously described. 47 The samples were packed into the rotor within the glovebox and sealed with the Vespel bushing and needle valve with Viton O-rings. A picture of the rotor is provided in Figure S1. The rotor was spun at 5 kHz at room temperature and heated to achieve a sample temperature of 100, 185, and 200°C with a heating ramp of 10°C/min (calibrated with a rotor of lead nitrate 48 spun at 5 kHz). A spectrum was recorded at each temperature as well as after cooling the sample to 25°C. For the hydrogenation experiment, the sample was pressurized with H 2 utilizing a one-way valve end cap. 47,49 The rotor was loaded in the glovebox and sealed with a one-way valve. The sealed rotor was then placed in a pressurizing chamber. The chamber was evacuated to ≈1 × 10 −3 bar to remove any air and was then repressurized with H 2 at 6 bar for 30 min to equilibrate. The presence of H 2 was verified by 1 H solid-state MAS NMR by comparing the 1 H NMR spectrum before and after introducing H 2 . These results as well as experimental details related to liquid NMR can be found in Figures S2 and S3 and Section S2. The quantification procedure of the species measured by NMR is laid out in the Supporting Information, and the accuracy of the values obtained is within 5 wt %.
TPD-MS. The samples studied for this work were analyzed on a calibrated, custom-built TPD-MS system equipped with the quadrupole mass spectrometer (QMS) Stanford Research Systems RGA 100, with m/z = 1−100 amu, sampling rate of 2−4 s, and 70 eV ionization energy. The samples were heated at a rate of 10°C/min from 26 to 400°C utilizing a Digi-Sense temperature controller. The quantity of material was adjusted to stay within the calibrated mass spectrometer's linear response region. In a typical analysis, 1−2 mg of sample was placed inside a platinum (Pt) foil packet to ensure homogeneous heating of the entire sample during the dehydrogenation process. The sample was then inserted into a quartz tube mounted to the TPD system, as described elsewhere. 50 Care was taken with the ramp rates to ensure that the pumping rate of the system was not altered by the release of hydrogen. All experimental parameters were controlled via a LabView interface with the RGA, heating system, and pressure gauges. Typical initial pressures before heating the sample were at 10 −8 torr, with a flat baseline, i.e., no water, hydrogen, or air signals above background.
The signal for H 2 (m/z = 2) was calibrated on the system using the well-known H 2 capacity of TiH 2 (Alfa Aesar) (4.0 wt % H 2 ). Three different sample sizes were tested, and the capacity value extracted from TPD has an accuracy of ±10%. The TiH 2 calibration samples were heated at a rate of 10°C/min from 25 to 700°C.
XANES. Following the same recipe as described above, a fresh batch of samples was specifically prepared a few days in advance of the XANES measurements. All information related to synthesis and characterization of the second batch is presented in Section S1.5 as well as a comparison of the two sample batches discussed in this work. XANES measurements were acquired in a UHV chamber at Stanford Synchrotron Radiation Light source (SSRL) beam line 10−1.
Powders were adhered to carbon tape and mounted to an aluminum sample stick oriented at 55°relative to the beam path. X-ray absorption spectra were obtained in the near edge region (−15 to +100 eV) over the Mg K (1 s) edges. Signals were acquired with simultaneous total electron yield (TEY) and total fluorescence yield (TFY). In this type of measurement, TEY is sensitive to the nearsurface phenomena while TFY provides a signal more representative of the bulk-phase. Edge steps were normalized using Athena, and edge positions were located by performing a Gaussian peak fit to the first numerical derivative of the normalized data. Linear combination fits to the data were also performed in Athena using the starting Mg(B 3 H 8 ) 2 2 was prepared according to the procedure recently reported by Moury et al. 12 The compound was analyzed by 11 B solid-state MAS NMR at 25°C, and the chemical shifts are shown in Figure 1. The peak assignments and quantification from the in the starting material will constitute a baseline quantity for the thermal conversion study. Since the synthesis of Mg-(B 3 H 8 ) 2 involves NaB 3 H 8 , the 11 B solid-state MAS NMR was not able to distinguish Mg(BH 4 ) 2 from NaBH 4 ; therefore, we performed 11 B solid-state MAS NMR of the as-synthesized NaB 3 H 8 at 25°C, shown in Figure S4, in the Section S3 revealing the presence of both B 3 H 8 − (80 wt %) and BH 4 − (15 wt %) as summarized in Table S1. Thus, the −42.  Figure  S5 indicates that this reaction was not entirely completed in this case. The presence of NaBr and MgBr 2 can potentially play a role, albeit minor, in the decomposition reaction of Mg(B 3 H 8 ) 2 -MgH 2 . It has been shown that the ionic radii of Br − and BH 4 − are of similar size, 45 making it the formation of mixed salts, such as Mg(BH 4 ) 2−x Br x , possible during heating of Mg(B 3 H 8 ) 2 in the presence of NaBr or MgBr 2 . Furthermore, the ionic substitution of halide for BH 4 was shown to modify the stability of metal borohydrides, specifically of NaBH 4 and LiBH 4 . 60,61 However, with our experimental data available, it is very difficult to appreciate the extent of this mixed salt formation.
The XANES spectra displayed in Figure 2 show the surfacesensitive total electron yield (TEY) and the fluorescence yield (TFY) signals from the sample, recorded at the Mg K-edge at 25°C. The synthesis, characterization, and comparison of this second batch with the main material investigated are presented and discussed in Section S4. Linear combination fits to the sample spectra presented in Figure 2 were performed using the reference spectra from Mg(B 3 H 8 ) 2 , MgH 2 , and Mg(BH 4 ) 2 and are gathered in Table S2. The converged fits supply the contribution of each reference spectrum to the total signal and provide a good measure of the relative abundance. Analysis of the TFY signal shows that the bulk of the starting material is composed of ∼22% of Mg(B 3 H 8 ) 2 and ∼78% of MgH 2 , thereby confirming the expected Mg(B 3 H 8 ) 2 -to-MgH 2 ratio of 1:4. Furthermore, the bulk contribution of Mg(BH 4 ) 2 in the starting material is close to zero, indicating that the BH 4 − species measured in 11 B solid-state MAS NMR are likely to be attributed to NaBH 4 rather than Mg(BH 4 ) 2 . While NaB 3 H 8 was used as the starting material for the synthesis of unsolvated Mg(B 3 H 8 ) 2 making this a plausible hypothesis, this observation requires further future studies. Due to the possibility of coexisting NaBH 4 and Mg(BH 4 ) 2 , we will discuss the conversion of Mg(B 3 H 8 ) 2 to BH 4 − containing species without distinguishing the separate contributions of NaBH 4 from Mg(BH 4 ) 2 after the conversion. Additionally, Figure 2a clearly shows a subtle surface contribution of metallic Mg(0) in the form of a small, broad peak ∼1300 eV, well segregated from the edge step of MgH 2 at ∼1302 eV. This is the first experimental report of a metallic contribution in a ball-milled M(B 3 H 8 ) n , n = 1 and 2 compounds. We hypothesize that the friction-induced heat provided during the ball-milling process is sufficient to partially reduce the Mg-species. Because the Mg(B 3 H 8 ) 2 baseline material, which was also ball-milled, does not show any Mg(0) contribution, we infer that it is either the MgH 2 that is being reduced or the high thermal conductivity of MgH 2 successfully transferring the heat to the Mg(B 3 H 8 ) 2 leading to its partial decomposition. It is known that metallic Mg and H 2 are in equilibrium with MgH 2 , with a H 2 pressure of ∼1.0 × 10 −4 bar at 127°C (see thermodynamic data from NIST in   Figure 4a shows the release of H 2 starting at ∼100°C, confirming that the onset of the decomposition process occurs at ∼100°C. After heating the sample to 200°C, it was rapidly quenched at 25°C to collect the NMR chemical shift, for which the assignments of the species and the quantification are reported in  Figure 4b. The m/z = 26 (B 2 H 4 ) and m/z = 59 (B 5 H 4 ) were monitored as representatives for B 2 H 6 and B 5 H 9 , respectively. These species were identified by comparing the full fragmentation pattern with that from the NIST webbook, as can be seen in Figure S8. From our previous characterization using differential scanning calorimetry coupled with thermal gravimetric analysis of this unsolvated Mg(B 3 H 8 ) 2 , a mass loss of 29 wt % was observed in the temperature range of 80−140°C. 12 Since the theoretical hydrogen content of Mg(B 3 H 8 ) 2 was 15 wt %, we concluded the formation of additional gases, such as B 2 H 6 and B 5 H 9 , together with H 2 . 12 The release of B 2 H 6 and B 5 H 9 was also reported during the thermal decomposition of unsolvated KB 3 H 8 and NaB 3 H 8 . 11,15,17 In the case of solvated Mg(B 3 H 8 ) 2 with diglyme, however, only the formation of B 5 H 9 was reported. 65 Figure 5.
Like the patterns shown in Figure 4, the signal of B 3 H 8 − starkly decreases already at 100°C while the BH 4 − -related peaks increase up to 185°C, after which no significant change occurs suggesting that the conversion is complete at that temperature. Unlike the conversion of unsolvated Mg(B 3 H 8 ) 2 , the peaks related to B 10 H 10 2− , B 12 H 12 2− , and borates do not visibly arise during the heating of the MgH 2 -containing material. After cooling the heated sample, the NMR spectrum is collected (upper red spectrum in Figure 5), and the peaks are assigned to species that are quantified, as summarized in Table 3. The net conversion of BH 4 − is ∼88 wt %. Because the 22 wt % of BH 4 − from the starting material is subtracted to obtain 88 wt % and there is a clear absence of the B 3 H 8 − signal from Figure 5, we conclude that this 88 wt % is equivalent to a complete conversion to BH 4 − in this case. Not only is this number 4 times as much as in the absence of MgH 2 , but it is also the highest conversion to BH 4 − occurring at ≤200°C in this study. Indeed, the wt % of B 12 H 12 2− was decreased from 16 wt % without MgH 2 to 2 wt % in the presence of MgH 2 . Therefore, the presence of MgH 2 selectively leads to BH 4 − as the main decomposition product of the thermal decomposition of Mg(B 3 H 8 ) 2 . To provide additional insight into the conversion of BH 4 − as a function of temperature, we quantified the species at 100°C and the values are reported in Table S5. The net formation of BH 4 − (i.e., after subtraction of the BH 4 − wt % from the starting material) is ∼59 wt %, indicating that BH 4 − is the major product at a temperature as low as 100°C and within 1 h (from the ramping rate of 10°C/min used in the NMR experiments).
Interestingly, the presence of borates species (BO x ) was not observed after heating Mg(B 3 H 8 ) 2 -MgH 2 (see Table 3) while a content of 15 wt % was found for the heat treatment of Mg(B 3 H 8 ) 2 as shown in Table 2. While 15 wt % is higher than expected (probably due to accidental exposure to some air/ moisture), the "absence" of BO x after heating of Mg(B 3 H 8 ) 2 -MgH 2 suggests that the BO x is reduced by MgH 2 following the reaction Mg(B 3 H 8 ) 2 + MgH 2 + BO x → Mg(BH 4 ) 2 + Mg(OH) 2 . The fact that B is not "lost" through the formation of BO x when the Mg(B 3 H 8 ) 2 -MgH 2 system is heated increases the overall BH 4 − yield. The reduction of BO x with MgH 2 has been recently demonstrated at room temperature and even exploited in the synthesis of NaBH 4 from NaBO x ball-milled with MgH 2 . 67−70 For the first time, the combination of this high conversion percentage occurring at ≤200°C, within 1 h, and without the external supply of H 2 is an unprecedented performance for M(B 3 H 8 ) n -MH n , n = 1 and 2 systems. Indeed, the KB 3 H 8  Figure 6, the decomposition starts at ∼110°C , which is similar to that of Mg(B 3 H 8 ) 2 . The H 2 signal recorded in Figure 6a confirms that in the temperature range considered, the dominant portion of H 2 originates from Mg(B 3 H 8 ) 2 rather than MgH 2 . The H 2 quantification from the areal integration of the signal was performed for Mg(B 3 H 8 ) 2 and Mg(B 3 H 8 ) 2 -MgH 2 and is presented in Table 4. Compared to the Mg(B 3 H 8 ) 2 sample, the amount of H 2 released in the presence of MgH 2 is almost halved. While the quantification of B 2 H 6 and B 5 H 9 was not performed, their associated m/z signal is significantly attenuated in the Mg(B 3 H 8 ) 2 -MgH 2 system, as shown in Figure 6b . It is further noteworthy and puzzling to see the release of B 5 H 9 at lower temperatures than the release of B 2 H 6 , as it is widely accepted that B 5 H 9 is formed from B 2 H 6 . 71−73 Moreover, the concomitant presence of B 2 H 6 and B 5 H 9 was also observed by in situ variable-temperature 11 B MAS NMR of Mg(B 3 H 8 ) 2 at 100°C (red spectrum in Figure 3), and their quantification is provided in Table S3. Therefore, it is plausible that the conversion of B 2 H 6 into B 5 H 9 occurs at a time scale faster than seconds as the refreshing time of the MS and of the 11 B MAS NMR each takes seconds and these techniques are not able to differentiate from B 2 H 6 formation and its conversion process. This observation indicates that MgH 2 not only promotes the selective B 3 H 8 -to-BH 4 − conversion but also potentially affects the decomposition mechanism, either by trapping/inhibiting reaction intermediates or by   Figure 7 shows the 11 B solid-state MAS NMR spectrum collected at 25°C (black line) and after heating the sample using the same procedure used for the two previous samples. The chemical shifts collected at 100, 185, and 200°C as well as after cooling the sample to 25°C after heating are also shown in Figure 7. The quantification of the species of the starting material is reported in Table S4. The quantification from the integrated intensity from the sample cooled after heating is reported in  The signal was normalized to the mass of the sample used and normalized to the mass of the "active" sample (i.e., H 2 -containing species) using the molecular weights of Mg(B 3 H 8 ) 2 , MgH 2 , and NaBr used in the synthesis, i.e., normalized to the molar mass of the active sample (see Table S3 for more details). Values apply for the desorption temperatures within 25 and 200°C. The values are accurate within a 10% error.  In this case, this back conversion to BH 4 is not complete. Therefore, an excess of MgH 2 is necessary to provide the hydrogen atoms necessary to form BH 4 − : This implies here that Mg 2+ has been reduced to Mg 0 . While XAS has not confirmed the presence of Mg 0 after heating Mg(B 3 H 8 ) 2 with MgH 2 , additional analysis needs to be performed to be conclusive. A possible "trapping" mechanism for the volatile products diborane and pentaborane is The addition of activated MgH 2 was added to the solventfree Mg(B 3 H 8 ) 2 , which led to a complete conversion to BH 4 − occurring at ∼100°C within 1 h, which is the first system to be reported of this achievement. This work demonstrates that the addition of activated MgH 2 leads to the selective conversion of BH 4 − , with only traces of closoboranes formed. The result is of significant importance as it is the first to be reported in the literature and expands the use of M(B 3 H 8 ) n -MH n , n = 1 and 2 to a series of applications. This work has shown that MgH 2 acts (i) as an inhibitor for the volatile B 2 H 6 and B 5 H 9 , thereby maximizing the mass retention of the Mg(B 3 H 8 ) 2 -MgH 2 system, and (ii) as the hydrogen donor for the B 3 H 8 -to-BH 4 − conversion to occur. This was confirmed when the thermal conversion of Mg(B 3 H 8 ) 2 -MgH 2 remained ∼85 wt % when exposed to 6 bar of H 2 , compared to ∼88 wt % without excess H 2 available. The exact role of MgH 2 in the lowtemperature B 3 H 8 -to-BH 4 − conversion process, or more generally its role as a H 2 donor, will be scrutinized in future studies, especially in terms of MgH 2 activation, which seems to play a paramount part in achieving the high conversion percentage.
Overall, these findings can increase the cyclability of B 3 H 8 −based electrolytes for batteries, liquid hydrogen carriers, and H 2 long-term storage. Furthermore, by inhibiting toxic volatile by-products, such as B 2 H 6 and B 5 H 9 , the addition of MgH 2 provides a green chemistry approach to using B 3  ; and hydrogen pressure obtained for the decomposition of MgH 2 (PDF)