Nanosizing Approach—A Case Study on the Thermal Decomposition of Hydrazine Borane

Hydrazine borane (HB) is a chemical hydrogen storage material with high gravimetric hydrogen density of 15.4 wt%, containing both protic and hydridic hydrogen. However, its limitation is the formation of unfavorable gaseous by-products, such as hydrazine (N2H4) and ammonia (NH3), which are poisons to fuel cell catalyst, upon pyrolysis. Previous studies proved that confinement of ammonia borane (AB) greatly improved the dehydrogenation kinetics and thermodynamics. They function by reducing the particle size of AB and establishing bonds between silica functional groups and AB molecules. In current study, we employed the same strategy using MCM-41 and silica aerogel to investigate the effect of nanosizing towards the hydrogen storage properties of HB. Different loading of HB to the porous supports were investigated and optimized. The optimized loading of HB in MCM-41 and silica aerogel was 1:1 and 0.25:1, respectively. Both confined samples demonstrated great suppression of melting induced sample foaming. However, by-products formation was enhanced over dehydrogenation in an open system decomposition owing to the presence of extensive Si-O···BH3(HB) coordination that further promote the B-N bond cleavage to release N2H4. The Si-OH···N(N2H4) hydrogen bonding may further promote N-N bond cleavage in the resulting N2H4, facilitating the formation of NH3. As temperature increases, the remaining N-N-B oligomeric chains in the porous silica, which are lacking the long-range structure may further undergo intramolecular B-N or N-N cleavage to release substantial amount of N2H4 or NH3. Besides open system decomposition, we also reported a closed system decomposition where complete utilization of the N-H from the released N2H4 and NH3 in the secondary reaction can be achieved, releasing mainly hydrogen upon being heated up to high temperatures. Nanosizing of HB particles via PMMA encapsulation was also attempted. Despite the ester functional group that may favor multiple coordination with HB molecules, these interactions did not impart significant change towards the decomposition of HB selectively towards dehydrogenation.


Introduction
Over the past 20 years, chemical hydrides have been widely investigated as potential material for hydrogen storage owing to high hydrogen capacity. For instance, ammonia borane and hydrazine borane, which possess hydrogen capacity of 19.6 wt% and 15.4 wt%, respectively, are the few materials that fulfil the material capacity requirement for vehicular hydrogen storage. The former has received significant research attention in the past decade. In contrast, the latter is comparatively less investigated even though its potential as hydrogen storage material has been highlighted since 2009 by Hugle et al. [1] Hydrazine borane consists of a hydrazine (N 2 H 4 ) and a borane (BH 3 ) group, which are datively bound together via the donation of a lone pair electron from N of N 2 H 4 to B of BH 3 . Furthermore, the covalent molecule of N 2 H 4 BH 3 (HB) stabilizes itself in an orthorhombic structure via dihydrogen bonding network between H δ+ on N and H δ− on B. In 2011, Sutton et al. [2] reported the regeneration of ammonia borane (AB) spent fuel, which utilized hydrazine as the reducing agent. In that process, HB was formed as an intermediate compound, which was later replaced by NH 3 to form ammonia borane. From the energy and cost efficiency point of view, a simpler process indicates better efficient energy and cost saving. Therefore, it attracts us to further focus on HB as an alternative to AB for on-board hydrogen storage application.
Thermolytic investigation has showed that HB melts at~60 • C and starts to decompose over a broad temperature range of 90-140 • C [3]. Similar to AB, dehydrogenation of HB is accompanied by foaming and by-products release, such as N 2 H 4 and NH 3 , preventing direct use of pristine HB as hydrogen carrier in on-board fuel cell application. To overcome these barriers, tremendous effort has been devoted to improving the hydrogen storage properties of HB. Chemical modification via interaction of HB with alkali metal hydrides have been widely investigated, yielding various alkali metal hydridotrihydroborate, (MN 2 H 3 BH 3 , M = Li, Na, K) [4][5][6][7]. In all MN 2 H 3 BH 3 reported, higher extent of dehydrogenation can be achieved at lower temperatures, however, traces of N 2 H 4 and NH 3 are still detectable. A recent study by Moury et al. [8] showed that the by-products formations from the thermolysis of NaN 2 H 3 BH 3 can be effectively suppressed by doping NaN 2 H 3 BH 3 with 5 wt% of NaH, yielding pure hydrogen in the dehydrogenation. In addition, melting oriented dehydrogenation was observed in NaN 2 H 3 BH 3 and KN 2 H 3 BH 3 , therefore, foaming issue remains a critical challenge. Another strategy to improve the hydrogen storage properties of HB is via catalytic hydrolysis or methanolysis, using transition metals or noble metals as metal catalysts [9][10][11][12][13]. However, the use of solvents (H 2 O or alcohol) makes the hydrogen capacity of the system unattractive for application and the exothermic dehydrogenation indicates the regeneration of spent fuel is of thermodynamically unfavorable. Nanosizing is another approach that has been widely used to improve the reaction kinetic of a hydrogen storage material [14]. In 2005, through quantum-chemical calculations, Wagemans et al. [15] proved that when the crystallite size of MgH 2 is reduced to below 2 nm, thermodynamic improvement is made possible. Therefore, parallel tuning the thermodynamic and kinetic properties is possible via nanosizing. Several strategies have been employed to control the size of the particle, including mechanically milling, nanoconfinement and polymer encapsulation. Nanoconfinement appears to be the ultimate approach to retain the size of the particles as the growth of particles is constrained by the pore size of the porous materials. In the same year, Gutowska et al. [16] discovered that AB confined within the channel of mesoporous silica (SBA-15) showed significant improvement in the dehydrogenation kinetic and a change in the dehydrogenation thermodynamic as compared to bulk AB. The dehydrogenation temperatures of neat AB at 110 • C and 155 • C were lowered upon confinement. The apparent activation energy for the hydrogen release was lowered significantly from 184 kJ mol −1 in the bulk ammonia borane to 67 kJ mol −1 in the confined AB. Furthermore, the emission of by-product borazine and the exothermicity of the hydrogen release have been greatly improved as results of the alteration in the decomposition pathway of confined ammonia borane. Lai et al. [17] further investigated the influence of the silica scaffolds on the dehydrogenation from AB/SBA-15 and AB/MCM-41. They found that the confined AB generally had a lower dehydrogenation temperature, released less amount of borazine and diborane but more NH 3 as compared to bulk AB. Polymer encapsulation is another useful approach in retaining the nanosize of the hydride particle. In 2010, Chen and coworkers [18] utilized PMMA to confine ammonia borane to produce a soft hydrogen storage composite. By using this approach, an air stable composite, which demonstrates rapid dehydrogenation kinetic with high density of hydrogen was obtained.
Thus far, no study on the confinement and polymer encapsulation of HB has been reported. In view of the positive impacts of the nanosizing towards the dehydrogenation of ammonia borane, in this study, we investigated the influence of different nanosizing approaches against the thermal decomposition of HB at different loading. Complete nanoconfinement of HB was achieved by tuning the mass ratio of HB to porous materials (MCM-41 or silica aerogel). In conjunction with the small pore sizes and the presence of Si-O and Si-OH functional groups on the porous framework, the dihydrogen bond network of the confined HB molecules was effectively disrupted and resulted in the absence of the melting process of the HB. The dissociation of B-N bonds was promoted and consequence in the substantial release of N 2 H 4 .

Synthesis of MCM-41
The MCM-41 was synthesized according to procedure modified from Ng et al. [19] A solution was prepared by mixing 6.070 g of cetyltrimethylammonium bromide (CTABr, Merck), 0.700 g of NH 4 OH (25 wt% purity, R&M,) in 60.000 g of distilled water at 80 • C. Next, 22.236 g of sodium silicate (Na 2 SiO 3 , Merck) was mixed with the previous clear solution to give new solution composition of 6 SiO 2 : CTABr: 1.5 Na 2 O: 0.15 (NH 4 ) 2 O: 250 H 2 O. The resulting solution was stirred for 15 min before it was hydrothermally treated at 100 • C for 24 h. Then, the pH of the hydrogel was adjusted to 11 by an addition of 25 wt% diluted acetic acid. The hydrothermal process and pH adjustment were repeated twice. The solid product was filtered, washed with distilled water until pH 7, dried overnight at 100 • C and calcined at 550 • C for 10 h to remove the organic template.

Synthesis of Silica Aerogel
A modified method by Rueda et al. [20] was applied in this synthesis of silica aerogel. The synthesis of silica aerogel involved three main steps, which were sol-gel synthesis, aging gel with solvent and removal of organic solvent with supercritical carbon dioxide. Tetraethyl orthosilicate (TEOS, Merck) as a precursor was mixed with absolute ethanol (EtOH, 99.9% purity, Qrec) deionized water and n-hexane (AR grade, Qrec) with a TEOS:EtOH:H 2 O:n-hexane molar ratio of 1:4.4:3.3:4.5. The resulting solution was stirred for 10 min. 0.08 mol of 25 wt% NH 4 OH aqueous solution (R&M) was added as condensation catalyst. Then, the solution was further stirred for 2 h to induce the gelation of silica alcogel particle. The mixture was then left for 2 h to form gel. The formed gel was retrieved and immersed in tetrahydrofuran (THF, Fisher) solvent in a closed vessel to age the gel and strengthen its structure. This process was continued for 7 days, where THF solvent was at least renewed twice to remove traces of ethanol and water. After the aging process, the THF solvent was removed from gel through supercritical CO 2 drying, which was produced by using CO 2 premier (99.995% purity, Air Product).

Synthesis of Hydrazine Borane
Hydrazine borane (HB) was synthesized according to the method used by Wu et al. [6]. Approximately 5.000 g of sodium borohydride (NaBH 4 , >96% purity, Sigma Aldrich) was mixed with 100 mL of THF and 12.000 g of hydrazine hemisulphate salt (N 2 H 4 · 1 2 H 2 SO 4 , 98% purity, Sigma Aldrich) was added later in the 500 mL flask. The flask was capped immediately and brought outside the glove box. The flask was then chilled in an ice bath during the first few hours of synthesis as the reaction proceeded vigorously. The flask was connected to a bubbler to release the gas formed during the reaction. This mixture was vigorously stirred and continued for 3 days. Then, the resulting suspension was centrifuged to separate it from the filtrate. The solvent in the filtrate was then removed under reduced pressure to yield white solid product.

Confinement of HB in MCM-41 and Silica Aerogel
In a glovebox, different amount of HB (0.025, 0.050, 0.100 and 0.200 g) were dissolved in 20 mL methanol to form HB methanol solutions. Next, 0.100 g of MCM-41 was added into these solutions to form mixtures with HB:MCM-41 mass ratios of 0.25:1, 0.5:1, 1:1 and 2:1. Methanol was used as solvent due to smaller molecular size of methanol, which can easily diffuse into the small pores of MCM-41. The mixtures were stirred for 2 h and the solvent was removed via reduced pressures, yielding white solid powders. Similar procedure was repeated for confinement of HB in silica aerogel (denoted as Si-Ae hereafter), but with the use of THF as solvent, at HB:Si-Ae mass ratios of 0.25:1, 0.5:1 and 1:1. The final dry solid HB:MCM-41 and HB:Si-Ae were obtained and kept in glovebox.

Encapsulation of Hydrazine Borane with PMMA
The polymethyl methacrylate (PMMA, Mw = 120,000 g mol −1 , Sigma Aldrich) with different weight were first dissolved with 20 mL of THF. 0.150 g of HB was then added to the solution to form HB:PMMA mass ratios of 0.5:1, 1:1, 2:1 and 4:1. The mixture was then stirred vigorously for 2 h and the solvent was removed via vacuum evaporation.

Characterization
XRD measurements were conducted on a Bruker D8 diffractometer (Cu Kα, 40 kV, 40 mA). A self-made sample cell was used to protect the samples from air contamination. FTIR measurements were conducted on a Perkin Elmer Spectrum 2000 FTIR spectrophotometer. Open system thermal desorption behaviors of the materials were investigated by using a custom-made temperature-programmed desorption-mass spectrometer combined system (TPD-MS) and thermogravimetry and differential thermal analysis (TG-DTA) measurement. In TPD-MS analysis, a sample loading of 10 mg was used and the sample was heated at a heating rate of 5 • C min −1 under a dynamic flow mode with purified argon gas as a carrier gas. MS signal of m/z ratio of 2 (H 2 ), 17 (NH 3 ), 28 (N 2 ) and 32 (N 2 H 4 ) were recorded as a function of temperature. TG-DTA was performed on Perkin Elmer STA8000. A sample loading of~10 mg was used. The analyses were carried out at different heating rates of 2, 5, 10 and 20 • C min −1 in N 2 atmosphere. Since hydrogen release at high temperatures is not practical for application, the kinetic investigations were carried out by treating the TG-DTA data at different heating rates, from room temperature to 200 • C (which corresponds to the main decomposition step), using Kissinger methods. The closed system decomposition was conducted in a sealed reactor with a total volume of 18.6 mL. A sample loading of 150 mg was used and the sample was heated at a heating rate of 5 • C min −1 . The released gaseous product was accumulated in the reactor throughout the experiment.

DTA Analysis
Although porous material is expected to play important role in tuning the dehydrogenation properties of hydrazine borane upon confinement, it is undeniable that the incorporation of porous material, which does not contribute in hydrogen storage will act as a deadweight that inevitably lowers the hydrogen capacity of the composite materials. Therefore, maximizing the loading of HB in the porous material is important to maintain the attractiveness of the materials for practical application. In this study, TG-DTA was utilized for the optimization of HB loading. Since HB melts prior to the decomposition, the absence of the melting phenomena from DTA is a good indication of the successful confinement of HB as it evidences disruption of intermolecular interaction-dihydrogen bond, which present only in the bulk HB. Figure 1 shows the DTA of the MCM-41 and silica aerogel confined HB at different mass ratios. As can be seen, at low loading of HB (0.25:1, 0.5:1 and 1:1 for HB:MCM-41), the HB characteristic melting peak at 60 • C was undetected. These results suggest that all the HB molecules have been successfully infiltrated into the pores of MCM-41. Since the pore size of the MCM-41 was 3.40 nm, the growth of HB particles in the pores is constraint by the pore size, disrupting the intermolecular interactions between HB molecules, thus diminishing the melting process of HB. Furthermore, several exothermic features of HB decomposition in the temperature range of 95-140 • C disappeared and became a broad mild exothermic process upon confinement. These results indicate that nanoconfinement indeed altered the physical and chemical properties of HB. When the loading of HB was increased to 2:1, the melting peak was clearly observed. A possible reason for this observation is that the amount of HB molecules infused into the channels of MCM-41 has exceeded saturation at this loading, which leads to recrystallization and agglomeration of HB on the outside of the MCM-41 that exhibits the properties of bulk HB. Hence, the ratio of 1:1 was determined to be the optimum loading. Similarly in the case of silica aerogel (Figure 1), an optimum ratio of HB:Si-Ae was determined to be 0. 25:1. diminishing the melting process of HB. Furthermore, severa decomposition in the temperature range of 95-140 °C disap mild exothermic process upon confinement. These results in indeed altered the physical and chemical properties of HB. W increased to 2:1, the melting peak was clearly observed. A p vation is that the amount of HB molecules infused into the ceeded saturation at this loading, which leads to recrystalli HB on the outside of the MCM-41 that exhibits the properties of 1:1 was determined to be the optimum loading. Similarly (Figure 1), an optimum ratio of HB:Si-Ae was determined to The IR spectra of MCM-41 an silica absorptions at ~3430 cm −1 , which can be assigned to Si ~1230 cm −1 and ~1080 cm −1 are assigned to asymmetric stret The corresponding symmetric stretching modes of Si-O-Si gr imately 799 cm −1 . The IR spectrum of bulk HB shows broad a 2600-3500 cm −1 , which can be assigned to multiple stretchin sorbance at 2360 cm −1 is attributed to B-H bonds, which is i borane [21]. The two overlapping absorbances at 1600 and asymmetric N-H bending and the corresponding symmetric N at 1261 cm −1 and 1340 cm −1 , respectively. Other absorbances  [19]. The corresponding symmetric stretching modes of Si-O-Si groups are observed at approximately 799 cm −1 . The IR spectrum of bulk HB shows broad absorbances in the range from 2600-3500 cm −1 , which can be assigned to multiple stretching of the N-H bonds. The absorbance at 2360 cm −1 is attributed to B-H bonds, which is identical to that of ammonia borane [21]. The two overlapping absorbances at 1600 and 1621 cm −1 can be ascribed to asymmetric N-H bending and the corresponding symmetric N-H bending can be detected at 1261 cm −1 and 1340 cm −1 , respectively. Other absorbances can also be detected in the fingerprint region, for instances, the broad B-H bending at 1168 cm −1 , N-H rocking at 1045 cm −1 and 984 cm −1 , BN-N asymmetric at 910 cm −1 and B-N stretching at 746 cm −1 , agreeing with that reported by Moury et al. [3]. Upon confinement, the samples clearly display the features from HB and MCM-41 or Si-Ae, suggesting the HB molecules remained intact within the pores without degradation. XRD characterizations on HB:MCM-41 (1:1) in Figure 2b showed the presence of weak HB diffraction pattern with different relative intensity as compared to the bulk, suggesting the remains of small amount of HB molecules that are arranged in a preferred orientation on the external of MCM-41. XRD of HB:Si-Ae (0.25:1), in contrary, showed complete absence of HB diffraction peak, suggesting a complete infiltration of HB molecules into the pores of silica aerogel. aterials 2023, 15, x FOR PEER REVIEW (0.25:1), in contrary, showed complete absence of HB diffraction peak, sug plete infiltration of HB molecules into the pores of silica aerogel.  Figure 3a shows the nitrogen adsorption and desorption isotherms before and after the HB confinement in MCM-41 and silica aerogel. Accord sification of IUPAC, most of the isotherms matched well to class IV isot typical for mesoporous material, except for HB:MCM-41 (1:1), which disp isotherm, indicating a non-porous structure. The hysteresis loop shapes rous structures in MCM-41 and silica aerogel are distinctly different. M strated a H3 type hysteresis loop at pressure (P/Po) sorption between 0.45 pressure of N2 and Po is the saturation pressure of N2), representing the p shaped pores at its surface. Silica aerogel showed H1 type hysteresis loop a between 0.65 to 0.99, suggesting the presence of well-defined cylindricalnels at their surface. The BET specific surface area and the BJH pore volu and silica aerogel were estimated to be 639 m 2 g −1 , 7.22 cm 3 g −1 and 439 m 2 g respectively. Upon confinement, the quantity of gas adsorbed on the mate significantly. The surface area and the free pore volume of HB:MCM-41 ( Ae (0.25:1) decreased to 10 m 2 g −1 , 0.05 cm 3 g −1 and 70 m 2 g −1 , 3.08 cm 3 g suggesting successful infiltration of HB molecules into the pores and even ness in the case of HB:MCM-41 (1:1). SEM micrographs shown in Figu clearly illustrated highly porous network of the empty MCM-41 and silica a tively.  Figure 3a shows the nitrogen adsorption and desorption isotherms of the samples before and after the HB confinement in MCM-41 and silica aerogel. According to the classification of IUPAC, most of the isotherms matched well to class IV isotherm, which is typical for mesoporous material, except for HB:MCM-41 (1:1), which displayed a class II isotherm, indicating a non-porous structure. The hysteresis loop shapes of the mesoporous structures in MCM-41 and silica aerogel are distinctly different. MCM-41 demonstrated a H3 type hysteresis loop at pressure (P/Po) sorption between 0.45 to 0.99 (P is the pressure of N 2 and Po is the saturation pressure of N 2 ), representing the presence of slit-shaped pores at its surface. Silica aerogel showed H1 type hysteresis loop at P/Po sorption between 0.65 to 0.99, suggesting the presence of well-defined cylindrical-like pore channels at their surface. The BET specific surface area and the BJH pore volume of MCM-41 and silica aerogel were estimated to be 639 m 2 g −1 , 7.22 cm 3 g −1 and 439 m 2 g −1 , 8.48 cm 3 g −1 , respectively. Upon confinement, the quantity of gas adsorbed on the materials decreased significantly. The surface area and the free pore volume of HB:MCM-41 (1:1) and HB:Si-Ae (0.25:1) decreased to 10 m 2 g −1 , 0.05 cm 3 g −1 and 70 m 2 g −1 , 3.08 cm 3 g −1 , respectively, suggesting successful infiltration of HB molecules into the pores and even stuffed to fullness in the case of HB:MCM-41 (1:1). SEM micrographs shown in Figure 3b

Open System Decomposition (TPD-MS and TGA)
To identify the gaseous species released during the decomposition, the confined samples were subjected for TPD-MS analyses and signal of H2 (m/z = 2), NH3 (m/z = 17), N2H4 (m/z = 32), N2 (m/z = 28) and B2H6 (m/z = 26) were monitored simultaneously. Figure 4 shows the relative intensity of the MS signals of the gaseous products formed from the decomposition of bulk HB, HB loaded in MCM-41 and silica aerogel at various loading, respectively. For bulk HB, small amount of B2H6 was detected during the melting process at 74 °C. Since no significant weight loss was detected in Figure 5, the release of B2H6 maybe originated from the side reaction (e.g., decomposition of hydrazine diborane, N2H4BH2(μ-H)BH3-a side product formed from the intermolecular interaction between HB molecule and BH3 generated from B-N bond cleavage, analogous to ammonia diborane, NH3BH2(μ-H)BH3, that was observed in the decomposition of AB) [22]. H2 only began to release at 94 °C and was released primarily up to 300 °C with NH3 as the major byproduct, which was detected simultaneously at ~100 °C and peaked at 138 °C. Other byproducts, such as B2H6 and N2 were present in only a minimum amount at high temperatures. However, N2H4 was not detected during the decomposition. This may be due to the condensation of N2H4 on the cold wall of the tubing prior to entering the mass spectrometer. Previous studies had confirmed the release of N2H4 in the decomposition of HB [3].
For all the confined samples except HB:Si-Ae (1:1), no B2H6 was detected at 74 °C, indicating that the formation of B2H6 is melting dependent and the confinement of HB in the porous silica framework impedes the intermolecular interaction to form N2H4BH2(μ-H)BH3 intermediate. These samples exhibited a similar dehydrogenation profile where H2 was released in stepwise manner. The first step involved rapid dehydrogenation, which peaked at ca. 145 °C and followed subsequently by a sluggish dehydrogenation. This result indicates an identical dehydrogenation mechanism in the confined samples. For the release of NH3 and N2, a similar trend was detected in MCM-41 and silica aerogel confined samples. In HB:MCM-41 (1:1), NH3 and N2 were released in a broad temperature range, which began at 130 °C and peaked at 240 °C. As the loading of HB in MCM-41 decreases (HB:MCM-41 (0.5:1) and (0.25:1)), two-step NH3 release was detected with the additional step began at temperatures as low as 60 °C. In the case of HB:Si-Ae (0.25:1), which has low loading of HB, a 2-step NH3 release profile was also detected, but at much higher relative To identify the gaseous species released during the decomposition, the confined samples were subjected for TPD-MS analyses and signal of H 2 (m/z = 2), NH 3 (m/z = 17), N 2 H 4 (m/z = 32), N 2 (m/z = 28) and B 2 H 6 (m/z = 26) were monitored simultaneously. Figure 4 shows the relative intensity of the MS signals of the gaseous products formed from the decomposition of bulk HB, HB loaded in MCM-41 and silica aerogel at various loading, respectively. For bulk HB, small amount of B 2 H 6 was detected during the melting process at 74 • C. Since no significant weight loss was detected in Figure 5, the release of B 2 H 6 maybe originated from the side reaction (e.g., decomposition of hydrazine diborane, N 2 H 4 BH 2 (µ-H)BH 3 -a side product formed from the intermolecular interaction between HB molecule and BH 3 generated from B-N bond cleavage, analogous to ammonia diborane, NH 3 BH 2 (µ-H)BH 3 , that was observed in the decomposition of AB) [22]. H 2 only began to release at 94 • C and was released primarily up to 300 • C with NH 3 as the major by-product, which was detected simultaneously at~100 • C and peaked at 138 • C. Other by-products, such as B 2 H 6 and N 2 were present in only a minimum amount at high temperatures. However, N 2 H 4 was not detected during the decomposition. This may be due to the condensation of N 2 H 4 on the cold wall of the tubing prior to entering the mass spectrometer. Previous studies had confirmed the release of N 2 H 4 in the decomposition of HB [3].
For all the confined samples except HB:Si-Ae (1:1), no B 2 H 6 was detected at 74 • C, indicating that the formation of B 2 H 6 is melting dependent and the confinement of HB in the porous silica framework impedes the intermolecular interaction to form N 2 H 4 BH 2 (µ-H)BH 3 intermediate. These samples exhibited a similar dehydrogenation profile where H 2 was released in stepwise manner. The first step involved rapid dehydrogenation, which peaked at ca. 145 • C and followed subsequently by a sluggish dehydrogenation. This result indicates an identical dehydrogenation mechanism in the confined samples. For the release of NH 3 and N 2 , a similar trend was detected in MCM-41 and silica aerogel confined samples. In HB:MCM-41 (1:1), NH 3 and N 2 were released in a broad temperature range, which began at 130 • C and peaked at 240 • C. As the loading of HB in MCM-41 decreases (HB:MCM-41 (0.5:1) and (0.25:1)), two-step NH 3 release was detected with the additional step began at temperatures as low as 60 • C. In the case of HB:Si-Ae (0.25:1), which has low loading of HB, a 2-step NH 3 release profile was also detected, but at much higher relative intensity as compared to H 2 signal. Therefore, it is plausible to ascribe the release of NH 3 to a similar deammoniation pathway (i.e., disproportionation of N 2 H 4 ), which may be attributed to the higher extent of Si-O···BH 3 (HB) interaction when less HB is encaged in the pores. In additions, the substantial release of NH 3 suggests that the release of N 2 H 4 is preferred over dehydrogenation. As the loading of HB in silica aerogel increases (HB:Si-Ae (0.5:1) and (1:1)), the H 2 release was enhanced. In HB:Si-Ae (1:1), the decomposition profile resembles that of bulk HB. This result is in agreement with the DTA result (Figure 1), which demonstrate the presence of sharp melting peak as results of the presence of excessive HB in the bulk. As can be seen, N 2 has a similar pattern as that of NH 3 , hence, it is plausible to ascribe the concurrent release of NH 3 and N 2 to the disproportionation of N 2 H 4 [23]. N 2 H 4 was not detected in all confined samples owing to the possible condensation of N 2 H 4 vapor during the TPD-MS analyses.
in the temperature range of 80-150 °C to release N2H4 and NH3 in the decomposition of bulk HB. The third step involves a grad wide temperature range of 150-300 °C, which can be attributed to of NH3. In both confined samples, their weight losses in the tem °C are much lower than that in bulk HB as results of the dead w the porous materials. Without considering the dead weight, the tributed by decomposition of HB in HB:MCM-41 (1:1) and HB beyond the expected value of 35.7 wt% from bulk HB, suggest heavy by-products still remains upon confinement. However, it that the excessive weight loss observed in the confined samples by the entrapped solvent, which requires higher temperatures ment.  For the ease of illustration, the intensity of NH 3 , N 2 H 4 , N 2 signals in TPD-MS analyses were intensified by a factor of 10, respectively, while B 2 H 6 signal was intensified by 5000. Figure 5 shows the TGA analyses of the optimal loading of HB:MCM-41 (1:1) and HB:Si-Ae (0.25:1) as compared to bulk HB. The decomposition of HB occurred in two steps with a major mass loss detectable in the temperature range of 80-150 • C, achieving 35.7 wt% upon heated to 300 • C. Such a huge weight loss indicates the substantial emission of heavy by-products, i.e., N 2 H 4 during the decomposition of HB, complementing the results obtained from TPD-MS. In the confined samples, a 3-steps weight loss was detected, achieving 21.2 wt% and 28.5 wt% from HB:Si-Ae (0.25:1) and HB:MCM-41(1:1), respectively, upon heated to 300 • C. The first step of weight loss, which occurred at temperatures < 80 • C corresponds to the detachment of solvent. The second step of weight loss occurred in the temperature range of 80-150 • C to release N 2 H 4 and NH 3 , similar to that observed in the decomposition of bulk HB. The third step involves a gradual weight loss over the wide temperature range of 150-300 • C, which can be attributed to the predominant release of NH 3 . In both confined samples, their weight losses in the temperature range of 80-300 • C are much lower than that in bulk HB as results of the dead weight contribution from the porous materials. Without considering the dead weight, the weight losses solely contributed by decomposition of HB in HB:MCM-41 (1:1) and HB:Si-Ae (0.25:1) are much beyond the expected value of 35.7 wt% from bulk HB, suggesting substantial release of heavy by-products still remains upon confinement. However, it is also worth to highlight that the excessive weight loss observed in the confined samples may also be contributed by the entrapped solvent, which requires higher temperatures for its complete detachment. where β is the heating rate, Tp is the peak temperature, Ea is the a pre-exponential factor and R the gas constant. The activation ene sition process can be obtained from the slope of the linear regres shown in Figure 7, demonstrate only minimum change in the main decomposition process, from 104.6 mol −1 in bulk HB to 91.3 (1:1) and 98.6 kJ mol −1 in HB:Si-Ae (0.25:1), respectively, indicatin tion barrier upon confinement. These results are somewhat expec  Figure 6. The kinetic of the HB decomposition, which corresponds to the main weight loss in the confined samples was investigated by treating the corresponding TGA/DTG results using Kissinger method based on the following equation: where β is the heating rate, T p is the peak temperature, E a is the activation energy, A is the pre-exponential factor and R the gas constant. The activation energy (E a ) of the decomposition process can be obtained from the slope of the linear regression. The Kissinger plots shown in Figure 7, demonstrate only minimum change in the activation energy of the main decomposition process, from 104.

Closed System Decomposition (Volumetric Release & FTIR)
Different from the dynamic flow mode (open system) of TG-DTA and TPD urements, a closed system volumetric release measurement enables the gaseo to be trapped in a confined space and allow it to further participate in the seco tion [24]. Therefore, the decomposition behavior of a closed system may diffe to that of an open system. In the presence of hydrogen rich B-H and N-H grou close system, the H δ+ on N2H4 (or NH3) is likely to pair up with all available H release H2, leaving behind N and B elements in the solid residue as results of t neous B-N bond formation. A closed system volumetric release measurement

Closed System Decomposition (Volumetric Release & FTIR)
Different from the dynamic flow mode (open system) of TG-DTA and T urements, a closed system volumetric release measurement enables the gas to be trapped in a confined space and allow it to further participate in the se tion [24]. Therefore, the decomposition behavior of a closed system may dif to that of an open system. In the presence of hydrogen rich B-H and N-H gr close system, the H δ+ on N2H4 (or NH3) is likely to pair up with all available release H2, leaving behind N and B elements in the solid residue as results o neous B-N bond formation. A closed system volumetric release measureme

Closed System Decomposition (Volumetric Release & FTIR)
Different from the dynamic flow mode (open system) of TG-DTA and TPD-MS measurements, a closed system volumetric release measurement enables the gaseous product to be trapped in a confined space and allow it to further participate in the secondary reaction [24]. Therefore, the decomposition behavior of a closed system may differ compared to that of an open system. In the presence of hydrogen rich B-H and N-H groups under a close system, the H δ+ on N 2 H 4 (or NH 3 ) is likely to pair up with all available H δ− on BH 3 to release H 2 , leaving behind N and B elements in the solid residue as results of the simultaneous B-N bond formation. A closed system volumetric release measurement of HB ( Figure 8) showed that decomposition of HB involves two major steps, with the first step begins gradually at~75 • C, peaked at~120 • C and followed by gradual release at 150 • C, releasing~1.5 mole of gaseous product upon heated to 225 • C. The second step followed subsequently and peaked at 261 • C, releasing additional~0.9 mole of gaseous product upon heated to 350 • C. FTIR characterization on the decomposed HB showed weak absorptions in the range of 2600-3500 cm −1 and at 2362 cm −1 , suggesting that the hydrogen remains in the solid residue appear in the form of N-H and B-H. For HB:MCM-41 (1:1), the gaseous product began to release at lower temperature of ca. 55 • C and liberated 2.3 mole of gaseous product in the first step. Further increasing in temperatures to 350 • C results in the release of additional 0.6 mole of gaseous product, giving a total of 2.9 moles. FTIR characterization on the solid residue showed the remains of weak absorptions correspond to N-H stretching and bending region, suggesting that all B-H has been fully utilized in the dehydrogenation. For HB:Si-Ae (0.25:1), the decomposition started even at 30 • C. As proven by TPD-MS (Figure 4), emission of by-product, such as NH 3 was detected. Therefore, it is plausible to deduce that the NH 3 that was released at low temperature may participate in the secondary reaction to release hydrogen. The rate of decomposition in the first step is significantly faster than those of HB:MCM-41 (1:1) and bulk HB owing to a lower barrier reaction involving gas (NH 3 )-solid reaction. Since NH 3 was also predominantly detected from TPD-MS in the temperature range of 150-225 • C, the improved rate of decomposition observed in the volumetric release under the same temperature range can be unambiguously ascribed to the secondary reaction involving NH 3 . Further increasing the temperature to 350 • C yielded a total of 3.4 mole of gaseous product. The FTIR spectra of the decomposed product resembles to that of silica aerogel, indicating a complete utilization of H atoms in the sample in a close system dehydrogenation since no N-H and B-H bond can be detected in the solid residue. aterials 2023, 15, x FOR PEER REVIEW remains in the solid residue appear in the form of N-H and B-H. For H the gaseous product began to release at lower temperature of ca. 55 °C mole of gaseous product in the first step. Further increasing in tempe results in the release of additional 0.6 mole of gaseous product, giving a FTIR characterization on the solid residue showed the remains of weak respond to N-H stretching and bending region, suggesting that all B-H h lized in the dehydrogenation. For HB:Si-Ae (0.25:1), the decomposition °C. As proven by TPD-MS (Figure 4), emission of by-product, such as N Therefore, it is plausible to deduce that the NH3 that was released at low participate in the secondary reaction to release hydrogen. The rate of the first step is significantly faster than those of HB:MCM-41 (1:1) and b a lower barrier reaction involving gas (NH3)-solid reaction. Since NH3 w nantly detected from TPD-MS in the temperature range of 150-225 °C, t of decomposition observed in the volumetric release under the same te can be unambiguously ascribed to the secondary reaction involving NH ing the temperature to 350 °C yielded a total of 3.4 mole of gaseous p spectra of the decomposed product resembles to that of silica aerogel, plete utilization of H atoms in the sample in a close system dehydrogen H and B-H bond can be detected in the solid residue. Decomposition of bulk HB begins with the disruption of intermole the melting process, which is accompanied by the release of B2H6. Howe B2H6 was not reported in previous study [3]. Its detection suggests poss side product hydrazine diborane during the melting process, which is p composition via B-N cleavage to release B2H6 and N2H4. In fact, this resu Decomposition of bulk HB begins with the disruption of intermolecular bonding in the melting process, which is accompanied by the release of B 2 H 6 . However, the release of B 2 H 6 was not reported in previous study [3]. Its detection suggests possible formation of side product hydrazine diborane during the melting process, which is prone towards decomposition via B-N cleavage to release B 2 H 6 and N 2 H 4 . In fact, this result agrees with the theoretical calculation reported by Nguyen et al., which proved that the B-N bond cleavage in monomer HB is favored over the hydrogen release to form N 2 H 4 + BH 3 (Equation (1)) [25]. The resulting BH 3 then reacts with the neighboring HB (BH 3 + HB) to yield hydrazine diborane (N 2 H 4 BH 2 (µ-H)BH 3 ) with single bridge B-H-B bond (Equation (2)). However, in the opposite direction of hydrazine diborane formation, B-N cleavage of hydrazine diborane to form B 2 H 6 + N 2 H 4 (Equation (3)) is preferred over B-H-B bond cleavage to form BH 3 + HB (Equation (2)), in agreement with our experimental observation on the detection of B 2 H 6 during the melting of HB.
Further increasing the temperatures slightly above the melting point of HB results in the release of H 2 , which may involve several pathways derived from HB or HB + BH 3 . As suggested by Moury and coworkers [3], bimolecular interaction of N 2 H 4 BH 3 to form N 2 H 4 BH 2 N 2 H 3 BH 3 , cyclization of two N 2 H 4 BH 3 to form N 2 H 3 BH 2 N 2 H 3 BH 2 , propagation of N-N-B oligomers by reacting with unreacted N 2 H 4 BH 3 , dehydrogenation of hydrazine diborane (N 2 H 4 BH 2 (µ-H)BH 3 ) via BH 3 -catalyzed H 2 -elimination are possible pathways for dehydrogenation. As the major mass loss of 35.7 wt% was only detectable in the temperature range of 80-150 • C, it is thus plausible to ascribe the mass loss to the release of N 2 H 4 resulting from B-N bond cleavage of N 2 H 4 BH 2 N 2 H 3 BH 3 . The theoretical mass loss for Equation (4) is 34 wt%, close to that of the experimental value. However, other possible pathways, such as B-N bond cleavages of hydrazine diborane (Equation (3)) and oligomers, which result in the release of N 2 H 4 are also possible considering myriad of potential structures that may form resulting from N-N-B chains propagation during dehydrogenation.
As evidenced by TPD-MS (Figure 4), H 2 was simultaneously released over the broad temperature range from 90 to 300 • C, thenceforth, the remaining B-H rich moieties upon B-N cleavage, i.e., BH 2 N 2 H 3 BH 3 or BH 3 may further react with neighboring N-H bearing compounds (i.e., unreacted N 2 H 4 BH 3 , any oligomers or cyclized rings) to release 1 mole of H 2 simultaneously. Intramolecular interaction between N-H and B-H of the resulting products is also possible to release H 2 at high temperatures.
When HB was confined in the framework of MCM-41 and silica aerogel, the intermolecular interactions were disrupted and the melting induced release of diborane were not detected, suggesting that the formation of hydrazine diborane is unlikely. With the presence of Si-O-Si and Si-OH surface functional groups, formation of multiple interactions with HB molecules (Figure 9a-c) similar to those observed in the confined AB [17], significantly altered the decomposition of HB. The formation of extensive Si-O···BH 3 (HB) coordination ( Figure 9a) may further promote the B-N cleavage, making the formation of N 2 H 4 even more barrierless as compared to that in bulk HB. The resulting N 2 H 4 is likely to undergo three possible reaction paths: (1) Similar to that of bulk HB, N 2 H 4 leaches out from the sample as evidenced by the huge weight loss observed in the temperature range of 80-150 • C in TGA. (2) N 2 H 4 interacts with the surface functional groups of MCM-41 and silica aerogel via hydrogen bonding Si-OH···N(N 2 H 4 ) (Figure 9d), facilitating ammonia formation via N-N bond cleavage; (3) The confined N 2 H 4 may react with the BH 3 group of the neighboring HB molecules to release H 2 . As NH 3 was still detectable at temperatures > 150 • C, it is thus plausible to ascribe it to the intramolecular B-N and N-N bond cleavage of the oligomers with short-range structure to release N 2 H 4 and NH 3 , respectively. dynamic flow mode, (i.e., TPD-MS and TGA analyses), the composition of the prod formed are supposed to be N-deficient. However, it is worth mentioning that both N and N2H4 are hydrogen rich compounds and they contribute to hydrogen release via s ondary reaction with B-H groups in the sample. In the case when the decomposition p cess was conducted in the closed system, more H2 can be released and N-rich produc likely to form.

Nanosizing via Polymer Encapsulation
In view of the fact that polymer encapsulation is also an efficient nanosizing proach [18,26,27], we compared the efficiency of polymethyl methacrylate (PMMA) capsulation in improving the decomposition of HB with that of nanoconfinement us porous silica. As shown in the DTA analysis ( Figure 10), polymer encapsulation of HB not impart significant change to HB particles as evidenced by the existence of melt phenomena regardless of the HB loading despite good dispersion of HB molecules PMMA (0.5:1) as shown in SEM-EDX mapping ( Figure 11). The main weight loss still curred in the temperature range of 80-150 °C. Without considering the dead weight c tribution and solvent release at temperatures < 80 °C, the weight loss solely contribu by decomposition of HB in HB:PMMA (0.5:1) is ca. 37.5 wt%, close to that observed bulk HB ( Figure 5). This result suggests that PMMA coating, despite with ester functio group that may favor multiple coordination with HB molecules, is equally inefficien altering the decomposition of HB.
From the results above, it is clearly evidenced that nanosizing of HB particle has be successful via confinement and polymer encapsulation, respectively. However, the composition of the nanosized HB particles were enhanced towards by-products f mation, suggesting that silicious porous materials and polymer with carbonyl (or est group are not appropriate hosts in tailoring the decomposition of HB selectively towa dehydrogenation for hydrogen storage application. To release H2 predominantly in decomposition of HB, a change in the reaction pathway is needed primarily. Therefo one should consider a host that favors the B-H and/or N-H bond cleavage over the B bond cleavage during the decomposition Since MCM-41 and silica aerogel have similar surface functional groups, i.e., -Si-O-Siand Si-OH, it is thus not surprising to observe similar alteration in the bonding chemistry of HB. However, the difference in the pore size and shape of both porous silica may pose slight change in the decomposition profile of the confined HB. MCM-41 has slit-shaped pores with small pore size, thus it results in smaller HB particles with extensive interaction with the surface functional groups, limiting intermolecular interaction to occur. In contrast, silica aerogel, which has larger cylindrical-like pore channels is likely to permit HB particles to grow inside the channel, enabling more intermolecular interactions to occur and thus demonstrates a different decomposition profile. This is clearly evidenced in the TPD-MS (Figure 4) of the confined HB at low loading of 0.25:1 in which HB:Si-Ae, which was at optimum HB loading released relatively more NH 3 at temperatures below 150 • C as compared to that in HB:MCM-41, which was underloaded. As significant amount of N 2 H 4 and NH 3 were detached during the decomposition process in the dynamic flow mode, (i.e., TPD-MS and TGA analyses), the composition of the product formed are supposed to be N-deficient. However, it is worth mentioning that both NH 3 and N 2 H 4 are hydrogen rich compounds and they contribute to hydrogen release via secondary reaction with B-H groups in the sample. In the case when the decomposition process was conducted in the closed system, more H 2 can be released and N-rich product is likely to form.

Nanosizing via Polymer Encapsulation
In view of the fact that polymer encapsulation is also an efficient nanosizing approach [18,26,27], we compared the efficiency of polymethyl methacrylate (PMMA) encapsulation in improving the decomposition of HB with that of nanoconfinement using porous silica. As shown in the DTA analysis (Figure 10), polymer encapsulation of HB did not impart significant change to HB particles as evidenced by the existence of melting phenomena regardless of the HB loading despite good dispersion of HB molecules in PMMA (0.5:1) as shown in SEM-EDX mapping ( Figure 11). The main weight loss still occurred in the temperature range of 80-150 • C. Without considering the dead weight contribution and solvent release at temperatures < 80 • C, the weight loss solely contributed by decomposition of HB in HB:PMMA (0.5:1) is ca. 37.5 wt%, close to that observed in bulk HB ( Figure 5). This result suggests that PMMA coating, despite with ester functional group that may favor multiple coordination with HB molecules, is equally inefficient in altering the decomposition of HB.

Conclusions
Nanoconfinement is known to be a time-tested approach that has b tive in altering both thermodynamic and kinetic properties of hydrides case of HB, despite successful infiltration into the pores of porous sil

Conclusions
Nanoconfinement is known to be a time-tested approach that has been proven effective in altering both thermodynamic and kinetic properties of hydrides. However, in the case of HB, despite successful infiltration into the pores of porous silica, no significant enhancement in the dehydrogenation selectivity of HB has been observed. The presence of Si-O and Si-OH surface functional groups tends to form multiple interactions with the HB molecules on its surface. Previous calculation has proven that B-N bond cleavage in bulk HB is favored over dehydrogenation. As N2H4 was still released predominantly in the confined HB, it is therefore believed that the SiO···BH3 (HB) coordination makes the B-N bond cleavage even more barrierless. The resulting N2H4 may interact with the surface functional groups via hydrogen bonding, which promote the formation of NH3 via disproportionation. Since N2H4 and NH3 were detached from the sample in an open system Figure 11. SEM images and EDX mapping of HB: PMMA (0.5:1) (green is referred to boron, orange to nitrogen and white to oxygen).
From the results above, it is clearly evidenced that nanosizing of HB particle has been successful via confinement and polymer encapsulation, respectively. However, the decomposition of the nanosized HB particles were enhanced towards by-products formation, suggesting that silicious porous materials and polymer with carbonyl (or ester) group are not appropriate hosts in tailoring the decomposition of HB selectively towards dehydrogenation for hydrogen storage application. To release H 2 predominantly in the decomposition of HB, a change in the reaction pathway is needed primarily. Therefore, one should consider a host that favors the B-H and/or N-H bond cleavage over the B-N bond cleavage during the decomposition.

Conclusions
Nanoconfinement is known to be a time-tested approach that has been proven effective in altering both thermodynamic and kinetic properties of hydrides. However, in the case of HB, despite successful infiltration into the pores of porous silica, no significant enhancement in the dehydrogenation selectivity of HB has been observed. The presence of Si-O and Si-OH surface functional groups tends to form multiple interactions with the HB molecules on its surface. Previous calculation has proven that B-N bond cleavage in bulk HB is favored over dehydrogenation. As N 2 H 4 was still released predominantly in the confined HB, it is therefore believed that the SiO···BH 3 (HB) coordination makes the B-N bond cleavage even more barrierless. The resulting N 2 H 4 may interact with the surface functional groups via hydrogen bonding, which promote the formation of NH 3 via disproportionation. Since N 2 H 4 and NH 3 were detached from the sample in an open system decomposition, the solid residue remains in the product are likely to be boron rich. However, in a closed system decomposition, N 2 H 4 and NH 3 were retained for the secondary reaction to release more hydrogen, yielding nitrogen rich solid residue. Nanosizing of HB via polymer encapsulation using PMMA was found to be equally inefficient in altering the decomposition of HB selectively towards dehydrogenation as substantial weight loss, which corresponds to the release of N 2 H 4 remains upon encapsulation.