NHI- and NHC-Supported Al(III) Hydrides for Amine–Borane Dehydrocoupling Catalysis

: The catalytic dehydrocoupling of amine–boranes has recently received a great deal of attention due to its potential in hydrogen storage applications. The use of aluminum catalysts for this transformation would provide an additional cost-e ﬀ ective and sustainable approach towards the hydrogen economy. Herein, we report the use of both N-heterocyclic imine (NHI)- and carbene (NHC)-supported Al(III) hydrides and their role in the catalytic dehydrocoupling of Me 2 NHBH 3 . Di ﬀ erences in the σ -donating ability of the ligand class resulted in a more stable catalyst for NHI-Al(III) hydrides, whereas a deactivation pathway was found in the case of NHC-Al(III) hydrides.

recent examples of Al hydrides in catalysis have shown that commercially available Al-H-containing reagents such as LiAlH4 and DIBAL-H are active towards the hydrogenation of imines using 1 bar of H2 [15] and hydroboration of alkynes [16], respectively.
Interest in dehydrocoupling reactions has recently risen, as it provides a clean and atom-efficient route to new element-element bonds, with the concomitant formation of dihydrogen, which has numerous applications in organic synthesis and materials chemistry [17][18][19]. The formation of dihydrogen in these types of reactions has gained the most attention due to the potential for hydrogen storage and therefore sustainable energy sources. In this regard, ammonia-borane (NH3BH3) has been earmarked as an ideal candidate for a future "hydrogen economy" due to its high weight percentage of hydrogen [20,21]. As such, catalysts from across the periodic table have shown their potential in the catalytic release of H2 from amine-boranes. Whilst transition metals dominate the field [22], examples from s-block [23][24][25][26], f-block [27,28], and metal-free [29][30][31] based systems have also contributed to furthering our understanding of these complex mechanisms. Central to most of the metal-based systems is the formation of metal hydrides [22] and their subsequent reactivity in enabling turnover. In terms of Al hydrides in dehydrocoupling catalysis, Wright and co-workers have reported the use of Al(III) amide pre-catalysts for amine-borane [32][33][34][35][36] and amine-silane [37] dehydrocoupling, in which the active Al hydride forms in situ.
Recent work within our group has shown the use of N-heterocyclic imines (NHIs) for the stabilization of group-13 metal hydrides [38]. This class of ligand is comparable to that of Nheterocyclic carbenes (NHCs), which are widely established as ligands for many catalytic transformations [39], as they can both be considered to be strong 2σ-electron donors ( Figure 1). The strong donating capabilities of NHI ligands has allowed for the stabilization and subsequent isolation of a variety of aluminum hydrides which have been used for a number of key transformations, such as the activation of elemental sulfur [40] and the dehydrogenative coupling and subsequent formation of a rare aluminum chalcogen double-bonded species [41]. Notably, the NHI-supported aluminum hydrides proved to be efficient catalysts in the hydroboration of terminal alkynes and carbonyl compounds [42]. Thus, our attention turned towards the use of NHI-supported aluminum hydrides for dehydrocoupling catalysis and comparisons with NHC aluminum hydrides.

Results and Discussion
To compare the catalytic potential of NHI-and NHC-stabilized Al(III) hydrides, a direct comparison between the two ligand classes bearing the same substituents was targeted. For ease of synthesis, 2,6-diisopropylphenyl (Dipp) and 2,4,6-trimethylphenyl (Mes) substituents were used for both NHI and NHC ligands. Further comparisons within the NHC ligand class were drawn between aryl and alkyl groups. Synthesis of the six different Al(III) hydrides (1)(2)(3)(4)(5)(6) used in this study ( Figure  2) were prepared according to literature procedures and then subsequently trialed in dehydrocoupling catalysis.

Results and Discussion
To compare the catalytic potential of NHI-and NHC-stabilized Al(III) hydrides, a direct comparison between the two ligand classes bearing the same substituents was targeted. For ease of synthesis, 2,6-diisopropylphenyl (Dipp) and 2,4,6-trimethylphenyl (Mes) substituents were used for both NHI and NHC ligands. Further comparisons within the NHC ligand class were drawn between aryl and alkyl groups. Synthesis of the six different Al(III) hydrides (1-6) used in this study ( Figure 2) were prepared according to literature procedures and then subsequently trialed in dehydrocoupling catalysis. Notable differences between the two ligand classes are apparent upon comparison of the structures. Due to the presence of the imine moiety in NHIs, the steric protection is removed from the metal center, and as the NHI ligands can act as a multifunctional ligand, it forms both σ-and πdonor bonds to Al, resulting in a dimeric Al-H2 complex which prevails in solution and solid state. Initial attempts of using catalysts 1 and 3 in a series of silane-amine (PhSiH3 and HNiPr2) and silaneborane (PhSiH3 and HBpin) dehydrocoupling reactions did not result in turnover, even at elevated temperatures. However, the use of dimethylamine-borane (Me2NHBH3) resulted in turnover under mild conditions.

Catalysis
The initial reaction of 5 mol % of 1 with Me2NHBH3 showed a minor formation of H2 after 24 h at room temperature. Increasing the temperature to 50 °C for the same time period also resulted in minor consumption of Me2NHBH3. Upon increasing the temperature to 80 °C, visible gas evolution was noted to occur, and monitoring by both 1 H and 11 B NMR spectroscopy showed the formation of the cyclic dimer [Me2NBH2]2 (A) and minor formation of the borane species (HB(NMe2)2 (B)). In contrast, upon the use of NHC catalyst 3, turnover was achieved under milder conditions and the formation of A was noted to occur at 50 °C. Further screening using catalysts 1-6 (Table 1) also revealed lower but generally prolonged reaction times for NHC-supported Al(III) hydrides, whilst NHIsupported catalysts required shorter times but higher temperatures. Catalysts 1-6 performed at about the same rate in comparison to previous reports of Al(III) hydride catalysis of Me2NHBH3 [33,34].  Notable differences between the two ligand classes are apparent upon comparison of the structures. Due to the presence of the imine moiety in NHIs, the steric protection is removed from the metal center, and as the NHI ligands can act as a multifunctional ligand, it forms both σand π-donor bonds to Al, resulting in a dimeric Al-H 2 complex which prevails in solution and solid state. Initial attempts of using catalysts 1 and 3 in a series of silane-amine (PhSiH 3 and HNiPr 2 ) and silane-borane (PhSiH 3 and HBpin) dehydrocoupling reactions did not result in turnover, even at elevated temperatures. However, the use of dimethylamine-borane (Me 2 NHBH 3 ) resulted in turnover under mild conditions.

Catalysis
The initial reaction of 5 mol % of 1 with Me 2 NHBH 3 showed a minor formation of H 2 after 24 h at room temperature. Increasing the temperature to 50 • C for the same time period also resulted in minor consumption of Me 2 NHBH 3 . Upon increasing the temperature to 80 • C, visible gas evolution was noted to occur, and monitoring by both 1 H and 11 B NMR spectroscopy showed the formation of the cyclic dimer [Me 2 NBH 2 ] 2 (A) and minor formation of the borane species (HB(NMe 2 ) 2 (B)). In contrast, upon the use of NHC catalyst 3, turnover was achieved under milder conditions and the formation of A was noted to occur at 50 • C. Further screening using catalysts 1-6 (Table 1) also revealed lower but generally prolonged reaction times for NHC-supported Al(III) hydrides, whilst NHI-supported catalysts required shorter times but higher temperatures. Catalysts 1-6 performed at about the same rate in comparison to previous reports of Al(III) hydride catalysis of Me 2 NHBH 3 [33,34].  Notable differences between the two ligand classes are apparent upon comparison of the structures. Due to the presence of the imine moiety in NHIs, the steric protection is removed from the metal center, and as the NHI ligands can act as a multifunctional ligand, it forms both σ-and πdonor bonds to Al, resulting in a dimeric Al-H2 complex which prevails in solution and solid state. Initial attempts of using catalysts 1 and 3 in a series of silane-amine (PhSiH3 and HNiPr2) and silaneborane (PhSiH3 and HBpin) dehydrocoupling reactions did not result in turnover, even at elevated temperatures. However, the use of dimethylamine-borane (Me2NHBH3) resulted in turnover under mild conditions.

Catalysis
The initial reaction of 5 mol % of 1 with Me2NHBH3 showed a minor formation of H2 after 24 h at room temperature. Increasing the temperature to 50 °C for the same time period also resulted in minor consumption of Me2NHBH3. Upon increasing the temperature to 80 °C, visible gas evolution was noted to occur, and monitoring by both 1 H and 11 B NMR spectroscopy showed the formation of the cyclic dimer [Me2NBH2]2 (A) and minor formation of the borane species (HB(NMe2)2 (B)). In contrast, upon the use of NHC catalyst 3, turnover was achieved under milder conditions and the formation of A was noted to occur at 50 °C. Further screening using catalysts 1-6 (Table 1) also revealed lower but generally prolonged reaction times for NHC-supported Al(III) hydrides, whilst NHIsupported catalysts required shorter times but higher temperatures. Catalysts 1-6 performed at about the same rate in comparison to previous reports of Al(III) hydride catalysis of Me2NHBH3 [33,34]. Reaction conditions 5 mol % c 1-6 (0.012 mmol), Me 2 NHBH 3 (0.24 mmol), 0.5 mL C 6 D 6 . a conversion of Me 2 NHBH 3 determined by 11 B NMR relative integrals; b TOF = (conversion/catalyst loading)/time. c Catalyst loading determined per molecule relative to amine-borane concentration.
Differences between the two ligand classes were found upon inspection of the product distribution ( Figure 3 and Table S1), indicating likely subtle changes in the stability of reaction intermediates or different mechanisms. Both ligand classes resulted in the formation of the cyclic dimer (A) as the major species, proceeding with the initial formation of BH 4 − and [Me 2 NBH 2 N(Me) 2 BH 3 ] − anions (D).
However, only NHI-supported Al(III) hydrides (1,2) resulted in the formation of the borane (B) species and a trace amount of dimethylamino-borane, Me 2 N=BH 2 (C), which can be considered an alkene analog. In the case of NHC catalysis, both the free and metal-bound [Me 2 NBH 2 N(Me) 2 BH 3 ] − anions (D and D', respectively) were observed. Differences within the ligand classes were marginal in the case of NHI ligands, with the steric demands being slightly removed from the Al center and thus minor differences in the TOFs were observed. However, in the case of NHC complexes (3)(4)(5)(6), where the steric demands of the ligand substituents were in closer proximity to the Al center, differences in TOFs were apparent. The use of less-sterically demanding Mes (4) or methyl (6) substituents resulted in much higher rates of reaction in comparison to those bearing isopropyl substituents (3 and 5).
Inorganics 2019, 7, x FOR PEER REVIEW 4 of 12 Differences between the two ligand classes were found upon inspection of the product distribution ( Figure 3 and Table S1), indicating likely subtle changes in the stability of reaction intermediates or different mechanisms. Both ligand classes resulted in the formation of the cyclic dimer (A) as the major species, proceeding with the initial formation of BH4 − and [Me2NBH2N(Me)2BH3] − anions (D). However, only NHI-supported Al(III) hydrides (1,2) resulted in the formation of the borane (B) species and a trace amount of dimethylamino-borane, Me2N=BH2 (C), which can be considered an alkene analog. In the case of NHC catalysis, both the free and metalbound [Me2NBH2N(Me)2BH3] − anions (D and D', respectively) were observed. Differences within the ligand classes were marginal in the case of NHI ligands, with the steric demands being slightly removed from the Al center and thus minor differences in the TOFs were observed. However, in the case of NHC complexes (3)(4)(5)(6), where the steric demands of the ligand substituents were in closer proximity to the Al center, differences in TOFs were apparent. The use of less-sterically demanding Mes (4) or methyl (6) substituents resulted in much higher rates of reaction in comparison to those bearing isopropyl substituents (3 and 5).

Mechanistic Studies
A series of stoichiometric reactions were undertaken in order to provide further insight into these reactions and the role of the supporting ligands.

NHI Mechanism, Catalysts 1 and 2
The initial reaction of 1 eq. of Me2NHBH3 with compound 1 resulted in no reaction at room temperature, in line with catalytic reduction. Upon heating to 80 °C, a small emergence of A was noted in the 11 B NMR spectrum after 1 h, whilst the 1 H NMR spectrum appeared unchanged ( Figure  S13 and S14). Further heating overnight led to the complete consumption of Me2NHBH3, as noted by the loss of Me2NHBH3 CH3 signal at δ 1.78 ppm in the 1 H NMR spectrum. Two NHI-containing species were identified via 1 H NMR spectroscopy, with the major species being catalyst 1 (85% by relative integrals). The 11 B NMR spectrum showed the formation of compounds A and B, and a third species which had a minor shift of δ 0.5 ppm in comparison to Me2NHBH3 ( Figure S15). We propose that this new minor NHI-containing species is the first step in the catalytic cycle, whereupon σ-bond metathesis of Al-H with the protic N-H bond of Me2NHBH3 occurs to yield 1 eq. of H2 and compound 7 ( Figure 4). Unfortunately, attempts to isolate compound 7 through varying stoichiometries failed and resulted in the isolation of 1 along with dehydrocoupling products A and B.

Mechanistic Studies
A series of stoichiometric reactions were undertaken in order to provide further insight into these reactions and the role of the supporting ligands.

NHI Mechanism, Catalysts 1 and 2
The initial reaction of 1 eq. of Me 2 NHBH 3 with compound 1 resulted in no reaction at room temperature, in line with catalytic reduction. Upon heating to 80 • C, a small emergence of A was noted in the 11 B NMR spectrum after 1 h, whilst the 1 H NMR spectrum appeared unchanged (Figures S13 and S14). Further heating overnight led to the complete consumption of Me 2 NHBH 3 , as noted by the loss of Me 2 NHBH 3 CH 3 signal at δ 1.78 ppm in the 1 H NMR spectrum. Two NHI-containing species were identified via 1 H NMR spectroscopy, with the major species being catalyst 1 (85% by relative integrals). The 11 B NMR spectrum showed the formation of compounds A and B, and a third species which had a minor shift of δ 0.5 ppm in comparison to Me 2 NHBH 3 ( Figure S15). We propose that this new minor NHI-containing species is the first step in the catalytic cycle, whereupon σ-bond metathesis of Al-H with the protic N-H bond of Me 2 NHBH 3 occurs to yield 1 eq. of H 2 and compound 7 ( Figure 4). Unfortunately, attempts to isolate compound 7 through varying stoichiometries failed and resulted in the isolation of 1 along with dehydrocoupling products A and B. Additional support for the proposed first step (i.e., formation of 7) is found from consideration of the formation of the dehydrocoupling product B. The formation of B is proposed to occur via βhydride elimination from the metal-coordinated [Me2NBH2N(Me)2BH3] − anion along with BH3 elimination; this has been observed in both groups 2 and 3 d 0 metal-based catalysis [23,43,44]. Alternatively, Wright and co-workers proposed that B is also the result of the deprotonation of Me2NHBH3 by the hydridic B-H rather than Al-H [33]. However, in the latter case B was observed to form at the start of the catalysis, with the concentration remaining static throughout the catalytic reaction [32]. On further inspection of the 11 B NMR data obtained from the in situ monitoring of the catalytic reaction ( Figure 3I, Figures S2 and S4), both NHI catalysts showed the initial formation of the [Me2NBH2N(Me)2BH3] − anion with the formation of B occurring in the latter stages of the catalysis. Therefore, it is highly likely that this mechanism proceeds via an intermediate akin to compound 7, as outlined in the proposed mechanism ( Figure 5).

NHC Mechanism
A similar investigation was carried out for the NHC-supported Al(III) hydrides, again using the diisopropylphenyl substituent (catalyst 3, IDippAlH3). Upon reaction of 3 with 1 eq. of Me2NHBH3, two new NHC-containing species were observed to form along with a notable lack of H2 formation in the 1 H NMR spectrum. One of these species was identified as the dihydroaminal, IDippH2, which has been previously reported from the NHC-induced dehydrocoupling of MeNH2BH3 [30]. The 11 B NMR spectrum again showed a minor shift of approximately δ 0.5 ppm, indicating the formation of the NHC-stabilized [Al]-N(Me)2BH3 complex 8, similar to that proposed for the NHI mechanism (7). Additional heating of this sample led to the increased formation of IDippH2 and a new NHCcontaining compound in the 1 H NMR spectrum. The 11 B NMR spectrum did not show any Additional support for the proposed first step (i.e., formation of 7) is found from consideration of the formation of the dehydrocoupling product B. The formation of B is proposed to occur via β-hydride elimination from the metal-coordinated [Me 2 NBH 2 N(Me) 2 BH 3 ] − anion along with BH 3 elimination; this has been observed in both groups 2 and 3 d 0 metal-based catalysis [23,43,44]. Alternatively, Wright and co-workers proposed that B is also the result of the deprotonation of Me 2 NHBH 3 by the hydridic B-H rather than Al-H [33]. However, in the latter case B was observed to form at the start of the catalysis, with the concentration remaining static throughout the catalytic reaction [32]. On further inspection of the 11 B NMR data obtained from the in situ monitoring of the catalytic reaction ( Figure 3I, Figures S2 and S4), both NHI catalysts showed the initial formation of the [Me 2 NBH 2 N(Me) 2 BH 3 ] − anion with the formation of B occurring in the latter stages of the catalysis. Therefore, it is highly likely that this mechanism proceeds via an intermediate akin to compound 7, as outlined in the proposed mechanism ( Figure 5).  Additional support for the proposed first step (i.e., formation of 7) is found from consideration of the formation of the dehydrocoupling product B. The formation of B is proposed to occur via βhydride elimination from the metal-coordinated [Me2NBH2N(Me)2BH3] − anion along with BH3 elimination; this has been observed in both groups 2 and 3 d 0 metal-based catalysis [23,43,44]. Alternatively, Wright and co-workers proposed that B is also the result of the deprotonation of Me2NHBH3 by the hydridic B-H rather than Al-H [33]. However, in the latter case B was observed to form at the start of the catalysis, with the concentration remaining static throughout the catalytic reaction [32]. On further inspection of the 11 B NMR data obtained from the in situ monitoring of the catalytic reaction ( Figure 3I, Figures S2 and S4), both NHI catalysts showed the initial formation of the [Me2NBH2N(Me)2BH3] − anion with the formation of B occurring in the latter stages of the catalysis. Therefore, it is highly likely that this mechanism proceeds via an intermediate akin to compound 7, as outlined in the proposed mechanism ( Figure 5).

NHC Mechanism
A similar investigation was carried out for the NHC-supported Al(III) hydrides, again using the diisopropylphenyl substituent (catalyst 3, IDippAlH3). Upon reaction of 3 with 1 eq. of Me2NHBH3, two new NHC-containing species were observed to form along with a notable lack of H2 formation in the 1 H NMR spectrum. One of these species was identified as the dihydroaminal, IDippH2, which has been previously reported from the NHC-induced dehydrocoupling of MeNH2BH3 [30]. The 11 B NMR spectrum again showed a minor shift of approximately δ 0.5 ppm, indicating the formation of the NHC-stabilized [Al]-N(Me)2BH3 complex 8, similar to that proposed for the NHI mechanism (7). Additional heating of this sample led to the increased formation of IDippH2 and a new NHCcontaining compound in the 1 H NMR spectrum. The 11 B NMR spectrum did not show any

NHC Mechanism
A similar investigation was carried out for the NHC-supported Al(III) hydrides, again using the diisopropylphenyl substituent (catalyst 3, IDippAlH 3 ). Upon reaction of 3 with 1 eq. of Me 2 NHBH 3 , two new NHC-containing species were observed to form along with a notable lack of H 2 formation in the 1 H NMR spectrum. One of these species was identified as the dihydroaminal, IDippH 2 , which has been previously reported from the NHC-induced dehydrocoupling of MeNH 2 BH 3 [30]. The 11 B NMR spectrum again showed a minor shift of approximately δ 0.5 ppm, indicating the formation of the NHC-stabilized [Al]-N(Me) 2 BH 3 complex 8, similar to that proposed for the NHI mechanism (7). Additional heating of this sample led to the increased formation of IDippH 2 and a new NHC-containing compound in the 1 H NMR spectrum. The 11 B NMR spectrum did not show any dehydrocoupling products, but the formation of two broad signals at δ −13 and −17 ppm was evident. Subsequent scale-up of this stoichiometric reaction and XRD-analysis revealed the unexpected formation of compound 9 ( Figure 6).
Inorganics 2019, 7, x FOR PEER REVIEW 6 of 12 dehydrocoupling products, but the formation of two broad signals at δ −13 and −17 ppm was evident. Subsequent scale-up of this stoichiometric reaction and XRD-analysis revealed the unexpected formation of compound 9 ( Figure 6). We propose that the formation of compound 9 (Figure 7) occurs in a similar sequence to that reported by Rivard and co-workers for the formation of NHC-BH2-N(H)(Me)-BH3 [30]. Firstly, upon reaction of 3 with Me2NHBH3, the reaction occurs at the NHC-Al bond rather than Al-H (Figure 7a). Here the NHC acts as a base, resulting in the formation of IDippH2 (confirmed by 1 H NMR), dimethylamino-borane (C), and AlH3 species. In the second step (Figure 7b Compound 9 was identified as NHC-BH 2 -N(Me) 2 -AlH 2 -N(Me) 2 -BH 3 complex, wherein one NHC-AlH 3 reacted with two molecules of Me 2 NHBH 3 . The AlH 2 unit is pseudo-tetrahedral, with the N(3)-Al(1) bond longer than N(4)-Al(1), suggesting some dative bonding character from the lone pair on N(3) to the empty orbital of Al(1). The NHC-BH 2 bond lengths (1.628(3) Å) were in line with previously reported NHC-BH 2 -containing compounds, and also fit with the 11 B NMR data matching the broad signal at δ −17 ppm [30,45,46]. Interestingly, the 27 Al NMR spectrum showed a triplet (δ 85 ppm, J Al-H = 354 Hz), which has a comparable shift to that observed by Wright and co-workers for their Al-H-containing dehydrocoupling catalyst (δ 82.8 ppm, d, J Al-H = 288 Hz).
We propose that the formation of compound 9 (Figure 7) occurs in a similar sequence to that reported by Rivard and co-workers for the formation of NHC-BH 2 -N(H)(Me)-BH 3 [30]. Firstly, upon reaction of 3 with Me 2 NHBH 3 , the reaction occurs at the NHC-Al bond rather than Al-H (Figure 7a). Here the NHC acts as a base, resulting in the formation of IDippH 2 (confirmed by 1 H NMR), dimethylamino-borane (C), and AlH 3 species. In the second step (Figure 7b), C inserts into the NHC-AlH 3 bond, resulting in NHC-BH 2 formation. Finally, the second molecule of Me 2 NHBH 3 undergoes σ-bond metathesis of its N-H with the terminal Al-H bond, resulting in the formation of compound 9 (Figure 7c).
The viability of compound 9 as a catalyst in the dehydrocoupling reaction was trialed, however only further decomposition to IDippH 2 was noted to occur, suggesting that this is a deactivation pathway in the active catalyst cycle. This would also account for the increased reaction times with 3. To confirm the viability of the active cycle, that is, via compound 8 (NHC-Al(H) 2 NMe 2 BH 3 ), the use of a more sterically demanding secondary amine-borane was tested. Reaction of 3 with 1 eq. of diisopropylamine-borane (iPr 2 NHBH 3 ) resulted in the formation of 8 iPr , with a reduced amount of IDippH 2 , thus confirming that this is the likely first step of the active catalytic cycle. It is further proposed that due to the increased steric congestion at the Al center with NHC ligands, the δ-hydride elimination is kinetically preferred over β-hydride elimination from the Al-coordinated The viability of compound 9 as a catalyst in the dehydrocoupling reaction was trialed, however only further decomposition to IDippH2 was noted to occur, suggesting that this is a deactivation pathway in the active catalyst cycle. This would also account for the increased reaction times with 3.
To confirm the viability of the active cycle, that is, via compound 8 (NHC-Al(H)2NMe2BH3), the use of a more sterically demanding secondary amine-borane was tested. Reaction of 3 with 1 eq. of diisopropylamine-borane (iPr2NHBH3) resulted in the formation of 8 iPr , with a reduced amount of IDippH2, thus confirming that this is the likely first step of the active catalytic cycle. It is further proposed that due to the increased steric congestion at the Al center with NHC ligands, the δ-hydride elimination is kinetically preferred over β-hydride elimination from the Al-coordinated [Me2NBH2N(Me)2BH3] − anion, therefore accounting for the absence of borane product (B) in the case of 3-6.
A proposed mechanism is outlined in Figure 8, taking both the active and deactivation pathways into account. Dimethylamino-borane (C) could either (i) react with compound 8 to form the metalbound [Me2NBH2N(Me)2BH3] − anion (D') (which was observed to form during the course of the catalysis) or (ii) insert into the NHC-AlH3 of the catalyst, leading to deactivation and formation of 9. Therefore, over the course of the catalysis, the remaining "active" species likely diminishes and results in prolonged reaction times. The differences in the reaction times observed with catalysts 3-6 were therefore likely due to the strength of the NHC-Al bond with strong donors such as IMe4 (6) resulting in the higher TOF due to the decreased likelihood of deactivation. This was also reflected in the use of the stronger σ-donating NHI ligands, as no deactivation pathways were observed. A proposed mechanism is outlined in Figure 8, taking both the active and deactivation pathways into account. Dimethylamino-borane (C) could either (i) react with compound 8 to form the metal-bound [Me 2 NBH 2 N(Me) 2 BH 3 ] − anion (D') (which was observed to form during the course of the catalysis) or (ii) insert into the NHC-AlH 3 of the catalyst, leading to deactivation and formation of 9. Therefore, over the course of the catalysis, the remaining "active" species likely diminishes and results in prolonged reaction times. The differences in the reaction times observed with catalysts 3-6 were therefore likely due to the strength of the NHC-Al bond with strong donors such as IMe 4 (6) resulting in the higher TOF due to the decreased likelihood of deactivation. This was also reflected in the use of the stronger σ-donating NHI ligands, as no deactivation pathways were observed.

Conclusions
In conclusion, differences in the use of NHI and NHC ligands for the stabilization of Al(III) hydrides and influence on catalytic activity were found. Whilst NHI-supported complexes required increased thermal activation, these systems were more stable in comparison to their NHC

Conclusions
In conclusion, differences in the use of NHI and NHC ligands for the stabilization of Al(III) hydrides and influence on catalytic activity were found. Whilst NHI-supported complexes required increased thermal activation, these systems were more stable in comparison to their NHC counterparts. Through a series of stoichiometric reactions, the isolation of compound 9 revealed a catalyst deactivation pathway that is prevalent in the NHC systems.

General Methods and Instruments
All manipulations were carried out under argon atmosphere using standard Schlenk or glovebox techniques. Glassware was heat-dried under vacuum prior to use. Unless otherwise stated, all chemicals were purchased from Sigma-Aldrich (Steinheim, Germany) and used as received. Me 2 NHBH 3 was sublimed three times prior to use. Benzene, toluene, and n-hexane were refluxed over standard drying agents (benzene/hexane over sodium and benzophenone), distilled, and deoxygenated prior to use. Deuterated benzene (C 6 D 6 ) was dried by storage over activated 4 Å molecular sieves. All NMR samples were prepared under argon in J. Young PTFE tubes. Catalysts 1-6 were synthesized according to procedures described in the literature [47]. NMR spectra were recorded on a Bruker AV-400 spectrometer (Rheinstetten, Germany) at ambient temperature (300 K). 1 H, 11 B, 13 C, and 27 Al NMR spectroscopic chemical shifts δ are reported in ppm relative to tetramethylsilane. δ( 1 H) and δ( 13 C) were referenced internally to the relevant residual solvent resonances. Details on XRD data are given in the Supplementary Materials.

General
Catalytic Procedure for Catalysis of Me 2 NHBH 3 5 mol % aluminum hydride catalyst (0.012 mmol) and Me 2 NHBH 3 (0.24 mmol) in 0.5 mL of C 6 D 6 were added to a J. Young PTFE NMR tube. The 1 H and 11 B NMR spectra were recorded and then placed in an oil bath at either 50 or 80 • C. Reaction progress was monitored by 1 H/ 11 B NMR spectroscopy until the consumption of Me 2 NHBH 3 .

General Procedure of Stoichiometric Reactions
Aluminum hydride catalyst (1: 40 mg, 0.046 mmol; 3: 40 mg, 0.095 mmol) and 1 eq. of Me 2 NHBH 3 (1: 3 mg, 0.046 mmol; 3: 5.6 mg, 0.095 mmol) in 0.5 mL of C 6 D 6 were added to a J. Young PTFE NMR tube. The 1 H and 11 B NMR spectra were recorded and then they were placed in an oil bath at either 50 or 80 • C. Reaction progress was monitored by 1 H/ 11 B NMR spectroscopy until consumption of Me 2 NHBH 3 .

Synthesis of Diisopropylamineborane, iPr 2 NHBH 3
This synthesis was adapted from general literature procedures [48], with the exception that this was solely carried out in hexanes rather than THF. Diisopropyl amine (1.00 g, 1.3 mL, 9.88 mmol) was placed in a Schlenk flask followed by 10 mL of hexanes. The reaction mixture was cooled in an ice bath to approximately 0 • C. BH 3 ·dms (dms = dimethylsulfide) (0.75g, 0.94 mL, 9.88 mmol) was added dropwise via syringe over 5 min. The reaction was allowed to stir for 10 min at 0 • C, before warming to room temperature and stirring overnight. Volatiles were removed under reduced pressure to leave a colorless oil (0.83 g, 73%). 1

Synthesis of Compound 8 iPr
IDippAlH 3 (3: 100 mg, 0.24 mmol) was dissolved in 3 mL of toluene and transferred to a Schlenk flask, followed by the addition of 1 eq. of iPr 2 NHBH 3 (35 µL, 0.24 mmol). This was allowed to stir at room temperature for 3 h. Solvent and volatiles were removed under reduced pressure to leave a pale-yellow gummy solid. This was subsequently washed with 3 × 10 mL of n-hexane to yield a colorless solid, 80 mg, 67% yield. 1

Synthesis of Compound 9
IDippAlH 3 (3: 200 mg, 0.48 mmol) and 2 eq. of Me 2 NHBH 3 (56 mg, 0.96 mmol) were weighed in a Schlenk tube. To this, 5 mL of toluene was added, and the Schlenk tube was placed in an oil bath at 50 • C for 24 h. Solvent and volatiles were removed under reduced pressure to leave a pale-yellow gummy solid. This was subsequently washed with 3 × 10 mL of n-hexane to yield a colorless solid, 102 mg, 40% yield. Crystals suitable for single-crystal X-ray diffraction were grown from a concentrated toluene solution at −30 • C. 1