Imine Reduction with Me2S-BH3

Although there exists a variety of different catalysts for hydroboration of organic substrates such as aldehydes, ketones, imines, nitriles etc., recent evidence suggests that tetra-coordinate borohydride species, formed by activation, redistribution, or decomposition of boron reagents, are the true hydride donors. We then proposed that Me2S-BH3 could also act as a hydride donor for the reduction of various imines, as similar compounds have been observed to reduce carbonyl substrates. This boron reagent was shown to be an effective and chemoselective hydroboration reagent for a wide variety of imines.


Introduction
Hydroboration can be considered as one of the most powerful methods for reduction of various organic substrates such as aldehydes, ketones, imines, and nitriles under mild reaction conditions [1][2][3]. Pinacolborane (HBpin) or catecholborane (HBcat) has been predominantly used as the hydroborating agent in these particular transformations, but the boron fragment (i.e., Bcat/Bpin) was normally sacrificed to yield, for example, free alcohols (from carbonyls) or amines (from imines/nitriles). Furthermore, these reduction reactions were mainly performed in the presence of catalytic amounts of a diverse range of (in)organic/organometallic compounds [4][5][6][7][8][9]. Several of the described hydroboration reactions were efficient as catalyst loadings as low as 0.001 mol%, resulting in excellent substrate conversions [7]. Nevertheless, the exact role of these presumed (pre)catalytic species has been divisive, as several reports provided convincing evidence for the existence of hidden boron catalysis (HBC), i.e., the main role of the species that were introduced in "catalytic" amounts was the formation, via activation, redistribution, or decomposition (Scheme 1) of HBcat/HBipn, of boron-based compounds (e.g., hydroborates and boranes) that then acted as the true catalysts [10][11][12].

Introduction
Hydroboration can be considered as one of the most powerful methods for reduction of various organic substrates such as aldehydes, ketones, imines, and nitriles under mild reaction conditions [1][2][3]. Pinacolborane (HBpin) or catecholborane (HBcat) has been predominantly used as the hydroborating agent in these particular transformations, but the boron fragment (i.e., Bcat/Bpin) was normally sacrificed to yield, for example, free alcohols (from carbonyls) or amines (from imines/nitriles). Furthermore, these reduction reactions were mainly performed in the presence of catalytic amounts of a diverse range of (in)organic/organometallic compounds [4][5][6][7][8][9]. Several of the described hydroboration reactions were efficient as catalyst loadings as low as 0.001 mol%, resulting in excellent substrate conversions [7]. Nevertheless, the exact role of these presumed (pre)catalytic species has been divisive, as several reports provided convincing evidence for the existence of hidden boron catalysis (HBC), i.e., the main role of the species that were introduced in "catalytic" amounts was the formation, via activation, redistribution, or decomposition (Scheme 1) of HBcat/HBipn, of boron-based compounds (e.g., hydroborates and boranes) that then acted as the true catalysts [10][11][12].

Results & Discussion
Instead of generating an "optimized" reaction condition with one of the examined imines and then implementing this procedure for the rest of the substrates, we decided to optimize each transformation in order to maximize the substrate conversions. Thus, the reactions were screened by varying the amount of Me 2 S-BH 3 and the reaction temperature while the reactants were mixed in about 1 mL of CDCl 3 in a sealed J. Young NMR tube. The most important outcomes and observations are summarized in Table 1. In a vast majority of examined transformations, heating to 60 • C was necessary to obtain quantitative substrates conversions with Me 2 S-BH 3 loadings varying between 0.75 and 1.50 equiv. For example, most of the reaction mixtures showed negligible reactivity at room temperature, while imine substrates with enhanced steric properties (entries 5 and 6, Table 1) required, in general, 1.50 equiv loadings of Me 2 S-BH 3 with respect to the imine. Reductions of imines that contain a 2,6-disopropylaniline fragment (e.g., entry 5) have been rarely examined, presumably due to low conversions of these particular substrates under the reported reaction conditions [32,33]. Furthermore, reduction of the imines that contained N-aryl substituents (entries 4 and 12) required not only a lower Me 2 S-BH 3 loading (e.g., 0.75 vs. 1.10 mmol) but also a shorter reaction time (e.g., 6 vs. 12 h) in comparison to their N-alkyl containing analogues (e.g., entries 2 and 3 vs. entry 12). This can be potentially explained by the presence of the resonance structures involving the N-aryl fragment puling the electron density away from the N=C fragment and hence allowing hydride transfer to the carbon atom of this fragment (see below). It was then not surprising to observe that the presence of an electron withdrawing group (CF 3 ) had a rate-enhancing effect (entry 7) while an electron donating group (OMe) had an opposite effect (entry 8). These observations suggested that the rate limiting step for the examined reactions was nucleophilic in nature (i.e., hydride transfer from a B-H fragment to the imine substrate; see below) and not electrophilic (i.e., formation of a imine-BH 3 adduct) [20].
More importantly, according to 1 H-NMR spectroscopy, all reactions resulted exclusively or solely in the anticipated reduction of the C=N double bond. This was particularly important for the reduction of the imine substrates that also contained an alkenyl group (i.e., α,β-unsaturated imines; entries 13 and 14, Table 1). Quantitative substrate conversions with excellent chemoselectivities (>98%) were achieved with these particular imines, while, at the same time, generating the fastest reaction rates among the examined substrates, despite the transformations performing at −78 • C (for the selectivity purposes). Lastly, according to the results summarized in Table 1, it appeared that this reduction protocol favored, in terms of reaction rates, ketimines over aldimines, which is not typically observed in the literature [10,[32][33][34][35][36][37][38][39][40][41][42][43][44][45]. At the moment, the precise reason(s) for this observation is(are) not known but it may suggest that the electrophilic step i.e., coordination of imine to BH 3 (see below) was rate determining, as one would expect that the hydride transfer (i.e., the nucleophilic step) would be less favored for ketimines over aldimines. coordination of imine to BH3 (see below) was rate determining, as one would expect that the hydride transfer (i.e., the nucleophilic step) would be less favored for ketimines over aldimines.  1 Reactions were performed using 1.0 mmol of imines. 2 This reaction was also preformed using 1.0 g (8.4 mmol), resulting in 66% product yield. 3 These reactions were performed at least three times. 4 Reaction performed at −78 °C in DCM.
As mentioned in the introduction, catalytic hydroboration of unsaturated C=X fragments (X=O, N, etc., but X≠C) has been a controversial topic. However, there is a significant body of evidence suggesting that four-coordinate B-H containing compound(s) (usually anionic), generated by activation, decomposition, or redistribution of boron reagents, act as initial hydride donors and hence as initiators of catalytically active species [11,12]. Clark and co-workers suggested a mechanism (Scheme 2a) that involved "activation" of HBpin (or HBcat) by coordination of a nucleophile (the electrophilic step), followed by hydride transfer (the nucleophilic step) from boron to the substrate (e.g., aldehyde), to yield the corresponding anion (e.g., alkoxide) [46]. This anion would then bind to another molecule of H-Bpin to generate the catalytically active species (e.g., [HBpin(alkoxy)] -). Thus, we propose that for reduction of imines with Me2S-BH3, initial hydride transfer occurs from either Me2S-BH3 or imine-BH3 (A, Scheme 2b) to produce amide anion B. This anion then displaces Me2S from Me2S-BH3 to generate the catalytically active species C, which acts as a hydride donor to another imine completing the cycle while also yielding reduced species D. As mentioned in the introduction, catalytic hydroboration of unsaturated C=X fragments (X=O, N, etc., but X =C) has been a controversial topic. However, there is a significant body of evidence suggesting that four-coordinate B-H containing compound(s) (usually anionic), generated by activation, decomposition, or redistribution of boron reagents, act as initial hydride donors and hence as initiators of catalytically active species [11,12]. Clark and co-workers suggested a mechanism (Scheme 2a) that involved "activation" of HBpin (or HBcat) by coordination of a nucleophile (the electrophilic step), followed by hydride transfer (the nucleophilic step) from boron to the substrate (e.g., aldehyde), to yield the corresponding anion (e.g., alkoxide) [46]. This anion would then bind to another molecule of H-Bpin to generate the catalytically active species (e.g., [HBpin(alkoxy)] − ). Thus, we propose that for reduction of imines with Me 2 S-BH 3 , initial hydride transfer occurs from either Me 2 S-BH 3 or imine-BH 3 (A, Scheme 2b) to produce amide anion B. This anion then displaces Me 2 S from Me 2 S-BH 3 to generate the catalytically active species C, which acts as a hydride donor to another imine completing the cycle while also yielding reduced species D.
Recently, Abe and Yamataka proposed that reduction of carbonyl compounds using BH 3 (in THF), the first step was H 3 B-carbonyl adduct formation (similar to A, Scheme 2b), followed by a hydride transfer step via a BH 3 -assisted transition state (Scheme 2c) [20]. However, although a majority of our examined hydroboration reactions require excess Me 2 S-BH 3 , it was still possible to achieve quantitative imine reduction with sub-stoichiometric amounts (0.75 mol%) of this boron reagent for several transformations (entries 1, 4, 13 and 14; Table 1). This suggested that, at least in certain instances, it was not only possible to reduce more than one imine substrate with one equivalent of Me 2 S-BH 3 but also that the BH 3 -catalysed hydride transfer step (in our case going from A to D) step was less likely to occur. Furthermore, it was also suggested that the hydride transfer step (e.g., A → D in our case) occurred via a bimolecular transition state (Scheme 2d) [19]. This would help explain our observation that more than one equivalent of imine was reduced by MeS 2 -BH 3 but a recent theoretical study indicated that a similar transition was high in energy [47]. Regardless of the nature of the hydride transfer step(s), it is still important to mention that we identified, via 11 B[ 1 H]-NMR spectroscopy, several proposed intermediates described in Scheme 2b. After mixing Me 2 S-BH 3 and N-benzylideneaniline in a 1:1 mol ratio at room temperature for 6 h, it was possible to detect respective intermediates A (δ B~− 9 (cis) and −14 (trans) ppm, Figure 1; [48]), D (δ B~4 1 ppm; [49]) and E (δ B~3 1 ppm; [50,51]). The fact that unreacted Me 2 S-BH 3 (δ B~− 20 ppm) was also present strongly suggested the existence of an equilibrium process between this reagent and intermediate A as indicated in Scheme 2b. Recently, Abe and Yamataka proposed that reduction of carbonyl compounds using BH3 (in THF), the first step was H3B-carbonyl adduct formation (similar to A, Scheme 2b), followed by a hydride transfer step via a BH3-assisted transition state (Scheme 2c) [20]. However, although a majority of our examined hydroboration reactions require excess Me2S-BH3, it was still possible to achieve quantitative imine reduction with sub-stoichiometric amounts (0.75 mol%) of this boron reagent for several transformations (entries 1, 4, 13 and 14; Table 1). This suggested that, at least in certain instances, it was not only possible to reduce more than one imine substrate with one equivalent of Me2S-BH3 but also that the BH3-catalysed hydride transfer step (in our case going from A to D) step was less likely to occur. Furthermore, it was also suggested that the hydride transfer step (e.g., A→D in our case) occurred via a bimolecular transition state (Scheme 2d) [19]. This would help explain our observation that more than one equivalent of imine was reduced by MeS2-BH3 but a recent theoretical study indicated that a similar transition was high in energy [47]. Regardless of the nature of the hydride transfer step(s), it is still important to mention that we identified, via 11 B[ 1 H]-NMR spectroscopy, several proposed intermediates described in Scheme 2b. After mixing Me2S-BH3 and N-benzylideneaniline in a 1:1 mol ratio at room temperature for 6 h, it was possible to detect respective intermediates A (δB~−9 (cis) and −14 (trans) ppm, Figure 1; [48]), D (δB~41 ppm; [49]) and E (δB~31 ppm; [50,51]). The fact that unreacted Me2S-BH3 (δB~−20 ppm) was also present strongly suggested the existence of an equilibrium process between this reagent and intermediate A as indicated in Scheme 2b. In conclusion, we have shown that Me2S-BH3 could also be used for reduction of a number of imines under mild reaction conditions and excellent chemoselectivity control. We have also managed to detect several key intermediates in the overall reaction pathway, which should aid in a better understanding of the overall hydroboration mechanism. In conclusion, we have shown that Me 2 S-BH 3 could also be used for reduction of a number of imines under mild reaction conditions and excellent chemoselectivity control. We have also managed to detect several key intermediates in the overall reaction pathway, which should aid in a better understanding of the overall hydroboration mechanism.

Materials and Methods
All imines were synthesized according to the literature reports (Table 2), while Me 2 S-BH 3 was purchased from a commercial source and used as received. CDCl 3 was dried by distilling it over CaSO 4 , while CH 2 Cl 2 was dried by distilling over CaH 2 . Reduction of imines was performed using standard Schlenk techniques, while subsequent work-up steps (with MeOH) were performed in on a benchtop. Table 2. Literature references for the synthesis of the examined imines and spectroscopic data for the corresponding amines. and Me2S-BH3 (amounts according to Table 1) were mixed in a sealed J. Young NMR tube using about 1 mL of CDCl3, the reaction mixture was left at 60 °C for the time duration indicated in Table 1. For α,β-unsaturated imines (entries 13 and 14), the reactants were mixed in CH2Cl2 at −78 °C. After the reaction was completed (via 1 H-NMR spectroscopy), it was quenched with 5 mL of MeOH, followed by removal of all volatiles under reduced pressure. The crude product mixture was then dissolved in 10 mL ethyl acetate, washed three times with 10 mL of water/brine, and dried with MgSO4. All amine samples were collected as oils after removal of solvent apart from benzylmethylamine (entry 1) and Nbenzylaniline (entry 4), which were obtained as solids. The spectroscopic data for all amines matched those reported (  General procedure for reduction of imines: After 1.0 mmol of an imine (entries 1-12) and Me 2 S-BH 3 (amounts according to Table 1) were mixed in a sealed J. Young NMR tube using about 1 mL of CDCl 3 , the reaction mixture was left at 60 • C for the time duration indicated in Table 1. For α,β-unsaturated imines (entries 13 and 14), the reactants were mixed in CH 2 Cl 2 at −78 • C. After the reaction was completed (via 1 H-NMR spectroscopy), it was quenched with 5 mL of MeOH, followed by removal of all volatiles under reduced pressure. The crude product mixture was then dissolved in 10 mL ethyl acetate, washed three times with 10 mL of water/brine, and dried with MgSO 4 . All amine samples were collected as oils after removal of solvent apart from benzylmethylamine (entry 1) and N-benzylaniline (entry 4), which were obtained as solids. The spectroscopic data for all amines matched those reported (