Synthesis of Electrophiles Derived from Dimeric Aminoboranes and Assessing Their Utility in the Borylation of π Nucleophiles.

Dimeric aminoboranes, [H2BNR2]2 (R = Me or CH2CH2) containing B2N2 cores, can be activated by I2, HNTf2 (NTf2 = [N(SO2CF3)2]), or [Ph3C][B(C6F5)4] to form isolable H2B(μ-NR2)2BHX (for X = I or NTf2). For X = [B(C6F5)4]- further reactivity, presumably between [H2B(μ-NMe2)2BH][B(C6F5)4] and aminoborane, forms a B3N3-based monocation containing a three-center two electron B-(μ-H)-B moiety. The structures of H2B(μ-NMe2)2BH(I) and [(μ-NMe2)BH(NTf2)]2 indicated a sterically crowded environment around boron, and this leads to the less common O-bound mode of NTf2 binding. While the iodide congener reacted very slowly with alkynes, the NTf2 analogues were more reactive, with hydroboration of internal alkynes forming (vinyl)2BNR2 species and R2NBH(NTf2) as the major products. Further studies indicated that the B2N2 core is maintained during the first hydroboration, and that it is during subsequent steps that B2N2 dissociation occurs. In the mono-boron systems, for example, i Pr2NBH(NTf2), NTf2 is N-bound; thus, they have less steric crowding around boron relative to the B2N2 systems. Notably, the monoboron systems are much less reactive in alkyne hydroboration than the B2N2-based bis-boranes, despite the former being three coordinate at boron while the latter are four coordinate at boron. Finally, these B2N2 electrophiles are much more prone to dissociate into mono-borane species than pyrazabole [H2B(μ-N2C3H3)]2 analogues, making them less useful for the directed diborylation of a single substrate.

Introduction

Diboron compounds are of significant importance in synthesis, particularly through the use of tetra-alkoxydiboron(4)s, such as B2Pin2 (Figure , top), in transition-metal-catalyzed borylation reactions.[1] Recent years have seen a resurgence in the chemistry of more (relative to B2Pin2) electrophilic diboron(4) compounds. This has expanded on the work of Schlesinger using B2X4 (X = halide),[2] and a number of electrophilic diboron(4) compounds now have been reported, including examples that can borylate π nucleophiles and activate small molecules (e.g., H2 and CO).[3] Parallel to this, there has been significant research into the chemistry of bidentate Lewis acids containing two electrophilic boron centers but no B–B bond, herein termed bis-boranes. While bis-boranes have been widely applied for small-molecule activation [e.g., in frustrated Lewis pairs (FLPs)][4] and in anion sensing,[5] the use of bis-boranes in the double borylation of π nucleophiles is relatively underexplored (vide infra).[6] This is despite the tunable nature of bis-boranes, particularly tailoring the B···B separation to match a specific substrate.

Figure 1

Top, diboron(4) compounds. Middle, bis-boranes able to effect double E–H borylation (E = N or C). Bottom, this work exploring electrophiles derived from dimeric aminoboranes.

One of the most utilized classes of bis-boranes are 1,2-C6H4(BX2)2 (X = Cl or Br) and derivatives (e.g., 9,10-diboraanthracenes). While various groups demonstrated that these bis-borane ditopic Lewis acids can be used for small-molecule activation[7] and anion binding,[5] Wagner et al. demonstrated that they can be used for double electrophilic C–H borylation.[8] Specifically, compound A effected the double vicinal electrophilic C–H borylation of a range of aromatics to form B2-doped polycyclic aromatic hydrocarbons (e.g., B). We recently extended this approach using bis-borane Lewis acids based on pyrazaboles (e.g., C),[9] which enabled the borylation-directed borylation of indoles and indolines (Figure , middle). Notably, the pyrazabole B2N4 core in C is sufficiently robust to persist during both electrophilic borylation steps (N–H and C–H), but it is reactive enough that it can be transformed subsequently into synthetically ubiquitous pinacol boronate esters. The B···B separation in A and C is ca. 3 Å, which is ideal for the vicinal functionalization of aromatics and the N/C7 functionalization of indoles. However, bis-borane Lewis acids with smaller B···B separations will be required for other substrates, for example, for accessing peri diborylated naphthalenes or 1,1-diborylated alkenes [where B···B separations in the double borylation products containing a B-(μ-NMe2)-B unit will be ca. 2.1–2.5 Å].[10,11] Thus, we were interested in accessing electrophiles derived from aminoborane dimers, (H2BNR2)2, which have appropriate B···B separations and are simple to access. Herein, we report the synthesis, characterization, and reactivity studies toward π nucleophiles of a series of boron electrophiles derived from aminoborane dimers.

Results and Discussion

Aminoboranes, R2NBH2, often exist in an equilibrium between dimeric and monomeric forms.[12] As a consequence of this sterically driven equilibrium, only a small number of aminoboranes are accessible as stable dimers in solution. Aminoboranes [Me2NBH2]2, 1, and [(pyrrolidine)BH2]2, 2, were selected due to their inexpensive starting materials (Me2NH/pyrrolidine and L-BH3) and their dimeric form dominating in solution at room temperature (and even on heating). Importantly, solid state data for 1 and 2 show that they exhibit B···B separations of ca. 2.20 Å, which is in the desired region.[13] Furthermore, calculations on the hydride ion affinity (HIA) of a borenium cation derived from 1,[14][D], reveal it to have a high HIA. Indeed, the HIA of [D] is comparable to some of the most reactive (in electrophilic C–H borylation)[15] mono-boron cationic species (e.g., [E], Figure , inset) and greater than the HIA for borocations derived from pyrazabole.[9] Presumably, the high HIA for [D] is due to the absence of any π donors bound to boron and indicates that if borenium cations (or functional borenium equivalents)[14] can be accessed from 1 or 2, they will be highly reactive species.

Figure 2

HIA calculations (relative to BEt3) at M06-2x/6-311G(d,p)//PCM(DCM) (all calculations herein are performed at this level).

Synthesis of Electrophilic “B2N2” Aminoboranes

Aminoboranes 1 and 2 were synthesized via catalyzed dehydrocoupling of the parent amine-borane following a reported method with lithium 2-Bu-pyridine.[16] A series of boron electrophiles derived from these precursors were targeted next. First 1 was treated with 0.5 equiv iodine (Figure , top), resulting in immediate hydrogen evolution and formation of H2B(μ-Me2N)2BH(I), 3. Compound 3 was isolated by sublimation as a crystalline solid in 34% yield (though in situ conversion to 3 is effectively quantitative). The 11B NMR spectrum of 3 revealed two distinct boron signals at δ11B 3.9 (t, 1JBH = 118 Hz, BH2) and 0.4 (d, 1JBH = 142 Hz, BHI), in accordance with the two different boron environments and their respective mono- and di-hydride substitution. The three inequivalent hydrides on the two boron centers were also observed by 1H NMR spectroscopy (see Figure S10). X-ray diffraction (XRD) analysis confirmed the structure of 3 (vide infra), while IR spectroscopy confirmed only terminal B–H units. Notably, attempts to access a bis iodide derivative, [Me2NBH(I)]2, by adding excess iodine to 1 failed even after 3 days at room temperature. While heating did lead to slow further reactivity of 3 with I2, this led to complex mixtures which contained monomeric species, for example, (amine)BI3.

Figure 3

Top, the synthesis of compounds 3–5. Inset bottom, the relative energies (ΔG) of O- and N-bound NTf2 isomers of 4. Bottom right, compound 6, only isolated in our hands as a byproduct from the reaction between alkynes and 4 (vide infra).

Mono and bis-borane systems, involving boron hydrides in combination with HNTf2 (NTf2 = N(S(O)2CF3)2), have been used to generate reactive boron electrophiles[17] including examples which can effect C–B formation.[9,18] Thus, aminoboranes 1 and 2 were treated with one equivalent of HNTf2 (Figure , top). Hydrogen evolution was observed, and H2B(μ-Me2N)2BH(OSOCF3NTf), 4, and H2B(μ-pyrrolidine)2BH(OSOCF3NTf), 5, were isolated in good yields (83 and 65%, respectively). The 11B NMR spectrum of 4 displayed two signals at δ11B 4.9 (d, 1JBH = 142 Hz, (H)BOSOCF3NTf) and 3.3 (t, 1JBH = 119 Hz, BH2). The 11B NMR signals of 5 were overlapped, giving a multiplet (δ11B 5.4–1.1), that turned into two singlets in the 11B{1H} NMR spectrum (δ11B 3.5 and 2.8) (see Figures S22 and S23). The 19F NMR spectra of both 4 and 5 showed two fluorine resonances (4 δ19F = −75.7 and −78.5; 5 δ19F = −75.7 and −78.4), indicating inequivalent CF3 groups in the NTf2 moiety. Similar 19F NMR spectra have been reported by Vedejs and co-workers for R3N–BH2NTf2 compounds where two fluorine resonances for B-NTf2 corresponded to the O-bound NTf2 isomer, in contrast to the more commonly found N-bound.[19] The mode of NTf2 binding to boron is linked closely to the steric environment around boron, for example, Me3N–BH2NTf2 has predominantly N-bound NTf2 (7:1 N/O bound), while with Pr2NEt–BH2NTf2, which contains a much larger amine, it is almost exclusively O-bound NTf2. The existence of only O-bound NTf2 in 4 and 5 suggests a relatively sterically encumbered boron center due to the four flanking methyl/CH2 units. The complete absence (by 19F NMR spectroscopy) of any N-bound isomer is also consistent with calculations which determined that the N-bound isomer of 4 to be 11.5 kcal mol–1 higher in energy than the O-bound isomer (Figure , inset). While multiple crystallization attempts for both 4 and 5 were unsuccessful, all data are consistent with NTf2 being O-bound in both 4 and 5, and IR spectroscopy was consistent with the presence of only terminal B–H units.

In an attempt to access a bis-NTf2 analogue, [(μ-Me2N)BH(OSOCF3NTf)]2, 6 (Figure , bottom right), compound 1 reacted with two equivalents of HNTf2. While this led to rapid formation of 4, further reactivity was only observed on heating. This led to slow consumption of 4 to form a complex mixture; notably, this contained mono-boron species (e.g., δ11B = 27.1, d, 1JBH = 170 Hz—see Figure S80). It was not possible to isolate pure doubly activated diboron aminoboranes from these reactions for use in synthetic studies. However, it should be noted that 6 is accessible as it was isolated in very low quantity by fractional crystallization during reactivity studies between alkynes and 4 (vide infra) and was crystallographically characterized (vide infra for discussion).

Desirous of synthesizing a bis-borane aminoborane electrophile that exists as a separated ion pair with a planar three coordinate borocation, 1 was treated with one equivalent of [Ph3C][B(C6F5)4] (Figure ). The major product from this reaction, 7, could be isolated by crystallization from layering a chlorobenzene solution with hexane. XRD analysis revealed the unexpected formation of the triboron monocation 7 (Figure , vide infra for structural discussion). The solid-state structure showed the formation of a salt, with a [B(C6F5)4]− counter anion, but it was not the desired cation [D]. Instead, the dimeric aminoborane 1 converted into a triboron species, containing a three-center two-electron B–H–B unit. The difference observed between forming 7 (with [B(C6F5)4]−) and 3–5 (with I– or [NTf2]−) is attributed to the very weakly coordinating nature of [B(C6F5)4]−, which provides insufficient stabilization of the borocation [D].[20] Therefore, post hydride abstraction by Ph3C+, borocation [D] reacts further, presumably with remaining 1 through formal transfer of one Me2NBH2 unit from 1 to [D]. The additional Me2NBH2 formally inserts into the N–BH+ bond of [D], ultimately affording 7. With the composition of 7 determined, modification of the stoichiometry to 3 equiv 1:2 equiv [Ph3C][B(C6F5)4] enabled 7 to be formed cleanly (by in situ NMR spectroscopy) and isolated in 70% yield.

Figure 4

Top, synthesis of the cation 7. Inset bottom, reported compounds with a comparable H–B(μ–H)B–H unit.

Analysis of 7 by 2D and low-temperature NMR experiments enabled assignment of the 1H and 13C signals, which were in accordance with the solid-state structure, with the presence of the three different hydride environments confirmed by 2D 11B–1H HMQC (see Figure S33). The bridging hydride was significantly more shielded compared to the other hydrides in 7, coming as a broad multiplet at δ 1.72–1.10 (vs 3.45–2.14 and 2.83–1.74 for the terminal hydrides). This chemical shift is in a similar region to that for the bridging hydride in the cation F reported by Himmel and co-workers (δ1H 1.97, in C6D6).[21] Notably, the 11B NMR spectrum of 7 showed a triplet (δ11B 3.4, 1JBH = 120 Hz, BH2) and a doublet (δ11B −6.1, 1JBH = 164 Hz, HB(H)BH), with no coupling between the bridging hydride and the boron atoms observed. The absence of observable coupling with the bridging hydride in the 11B NMR spectrum is consistent with related literature examples (e.g., F and G—Figure , inset bottom)[11,21] and is partly due to the smaller magnitude of 1JB–H coupling involving bridging hydrides (as reported in these related systems). While the additional B–H coupling was observed in higher-temperature NMR experiments for G, coupling with the bridging hydride was not observed at higher temperature for 7, with decomposition of 7 occuring at higher temperatures.

Finally, to enable comparisons during reactivity studies, a strong electrophile derived from a monomeric aminoborane was targeted. Therefore, the more sterically encumbered aminoborane (Pr)2NBH2 (which exists as a monomer at room temperature) was reacted with HNTf2 (Scheme ). This reaction resulted in the immediate formation of the amine-borane adduct 8 instead of the desired product 9, as evidenced by a boron resonance at δ11B −9.5 (br, t, 1JBH = 123 Hz) and was confirmed further by 1H, 13C, and 19F NMR spectroscopies (see Figures S40–S45). Heating 8 in benzene resulted in very slow conversion to 9, which showed a signal in the 11B NMR spectrum at δ 28.9 (d, 1JBH = 170 Hz); however, even after 7 days heating, a significant amount of 8 (ca. 60%) persisted. Therefore, compound 9 was isolated by fractional crystallization and characterized by 1H, 13C, and 19F NMR spectroscopies. The single signal in the 19F NMR spectrum (δ −73.2) indicated an N-bound triflimide, which was later confirmed by XRD analysis (vide infra for discussion). The switch from O-bound NTf2 in the bis-borane structures of 4 and 5 to N-bound NTf2 in 9 is notable and demonstrates the steric crowding present in dimeric aminoborane derivatives. Finally, it should be noted that resonances very close to the δ11B of 8 (−6.5, t) and 9 (27.1, d) are observed in the reaction of 4 with HNTf2, supporting the conclusion that monomeric boranes are being produced from combinations of 4 and two equiv HNTf2.

Scheme 1

Synthesis of Electrophiles by Addition of HNTf2 to a Monomeric Aminoborane

Solid-State Structures

Crystals of 3 were obtained by sublimation and confirmed the formulation from NMR spectroscopy (Figure , left). The B···B distance in 3 (2.142(5) Å) is slightly shorter compared to its precursor 1. Contrary to 1 that exhibits a planar N2B2 four-membered ring (Σ (internal angles of the N2B2 ring) = 360°), the unsymmetric substitution in 3 has distorted the N2B2 ring slightly into a butterfly conformation with the Σ (internal angles of the N2B2 ring) = 355°. A similar minor distortion was observed for an unsymmetric analogue, Br2B(μ-Me2N)2BBr(OEt) (Σ (internal angle of the N2B2 ring) = 357°).[22] The larger iodine atom also causes other distortions in 3, as evidenced by the large B2–B1–I1 angle of 142.9(2)°, a value greater than those in related substituted B2N2 compounds (where angles span the range 128–139°).[22,23] Furthermore, the incorporation of the less coordinating iodine leads to a shortening of the B1–N1 bond length (1.579(3) Å) and elongation of the B2–N1 bond length (1.614(4) Å), compared with its precursor 1, where all the B–N bond lengths are 1.595–1.597 Å. The boron atom substituted with iodine shows a distorted tetrahedral geometry with a B–I bond of 2.252(4) Å, a slightly shorter distance compared with the B–I bond lengths reported for closely related compounds L-BH2I (2.30–2.34 Å, L = N-heterocyclic carbenes or PR3).[24]

Figure 5

Left, the solid-state structure of 3 (50% ellipsoid probability). Hydrogen atoms except H3, H4, and H5 are omitted for clarity. Selected metrics [Å or °] for 3: B···B = 2.142(5), B1–N1 = 1.579(3), B2–N1 = 1.614(4), B1–I1 = 2.252(4), and B2–B1–I1 = 142.9(2). Right, the solid-state structure of 6 (30% ellipsoid probability). Hydrogen atoms except H1 and disorder are omitted for clarity. Selected metrics [Å or °] for 6: B···B = 2.274(2), B1–N1 = 1.589(8), B1–N1 = 1.662(10), B1–O1 = 1.526(7), B1–O1–S1 = 123.3(4), and B1–B1–O1 = 111.8(4).

The bis-NTf2 derivative, [(μ-Me2N)BH(OSOCF3NTf)]2, 6 (Figure , right) was obtained as a byproduct from reactions between 4 and alkynes by fractional crystallization. Its solid-state structure revealed doubly O-bound triflimide moieties, which is in accordance with the O- versus N-bound equilibrium being sterically driven and the expected steric encumbrance around the core N2B2 ring (based on NMR data for 4). 6 crystallizes as the trans isomer, which exhibits a large B1–O1–S1 angle (123.3(4)°) and a B1–O1 bond length of 1.526(7) Å, which is the shortest reported distance for an NTf2 O bound to four coordinate boron.[17c,25] Note, the calculated B–O distance in 4 is comparable at 1.522 Å. Finally, the B···B distance in 6 (2.274(2) Å) is larger compared to 1 and 3, possibly due to steric effects from the two NTf2 units. While the four-membered ring of 6 is planar (Σ(internal angles) = 360°), the O-bound triflimides are impacting the core N2B2 ring with a noticeable difference observed between the B1–N1 bond lengths [1.589(8) and 1.662(10) Å], a difference larger than that in compound 3.

To assess why N-binding of NTf2 to the B2N2 cores in 4 is significantly higher in energy, the calculated structure of the non-observed N-bound isomer of 4, 4-N, was analyzed. The B–NMe2 bonds (1.575 and 1.586 Å) are typical for boron-nitrogen single bonds at tetra-coordinate boron centers; however, the B–NTf2 (1.626 Å) bond length is longer than the B-NTf2 distance in a NTf2 derivative of pyrazabole [1.609(2) Å] and a Ar3P–BH(R)NTf2 species,[17b] suggesting a weaker interaction in 4-N. The N2B2 four-membered ring of 4-N is found in a butterfly conformation and is more distorted than that in the case of 3, with a Σ (internal angles of the N2B2 ring) = 351° (vs 355° in 3). Most notably, a very large B–B-NTf2 angle of 149.8° is found for 4-N, much larger than the B1–B1–O1 angle observed in compound 6 (111.8(4)°) and even higher than that the B2–B1–I1 for 3 (142.9(2)°). These data highlight the steric effects that N-binding of NTf2 to B2N2 cores imparts which is the likely reason why O-bound NTf2 is observed experimentally in 4–6.

Comparison of the calculated structure of 4-N with the solid-state structure of the 9 is also informative as 9 also contains an N-bound NTf2 (Figure , left). This is in accordance with a less hindered tricoordinate boron center. Notably, in 9, the B1–NTf2 exhibits a bond length of 1.586(15) Å, which is much shorter than that calculated for 4-N, further indicating the significant steric crowding around the B2N2 cores. Furthermore, in 9, a short B1–N2 bond length (1.351(17) Å) is observed, which is in the range of B=N double bonds (1.3–1.4 Å),[26] but it is shorter than previously reported R2N=BH2 species (e.g., 1.380(2) Å),[27] possibly due to enhanced electrophilicity at boron on substitution of H for NTf2. The presence of a B=N unit in 9 is also confirmed by the very small C3–N2–B1–N1 torsion angle of −2.5(4)°.

Figure 6

Left, the solid-state structure of 9 (50% ellipsoid probability). Hydrogen atoms except H3 are omitted for clarity. Selected metrics (Å or °) for 9: B1–N1 = 1.566(4), B1–N2 = 1.370(4), and C3–N2–B1–N1 = −2.5(4). Right, the solid-state structure of 7 (50% ellipsoid probability). Hydrogen atoms except H7, H8, H9, H10, and H11 and the anion are omitted for clarity. Selected metrics (Å) for 7: B1–N1 = 1.546(3), B1–N2 = 1.544(3), B2–N1 = 1.545(3), B2–N3 = 1.545(3), B3–N3 = 1.617(3), and B3–N2 = 1.617(3).

Finally, the solid-state structure of 7 (Figure , right for the cationic component) represents a salt derived from the trimer (BH2)3(NMe2)3.[28] Contrary to neutral (BH2)3(NMe2)3, which exhibits a chair conformation, the cationic portion of 7 has a distorted boat conformation, with a planar B1–B2–N3–N2 central part [Σ (angle of N2B2) = 360°] and N1 and B3 orientated on the same face. The boron atoms B1 and B2, involved in the three-center two-electron B–H–B unit, exhibit shorter B–N bond lengths [between 1.544(3) and 1.546(3) Å) than B3 (1.617(3) Å], consistent with B1 and B2 being more electron-deficient boron centers. This is also indicated by the N–B bonds in neutral G being considerably longer [1.614(6) Å]. The B1···B2 distance at 1.912(4) in 7 is also shorter than that in G [1.971(7) Å]; this difference is not due to the peri substitution in G as the B···B separation in (BBN)2(μ-H)(μ-NMe2) is 1.975(4) Å. However, 1.912(4) Å is still longer than a B–B single bond, and we attribute this short B···B separation to the contracted N1–B bonds involving the bridging NMe2 that are short as a result of the electron deficiency at B1 and B2 afforded by the cationic charge.

Reactivity Studies

With a series of bis-borane borocation equivalents isolated, we investigated briefly their reactivity toward simple bases. The significant steric crowding in 4 (indicated by the less common O-bound mode of NTf2 coordination) was confirmed by the fact that 4 does not bind bulky Lewis bases such as PBu3, or even PPh3, in contrast to other reactive (but less hindered) borocations.[14] However, small nucleophilic bases do react with these bis-borane electrophiles, but they lead to cleavage of the B2N2 core. For example, the addition of 4-DMAP to 4 led to the precipitation of a solid that was confirmed by X-ray crystallographic studies to be the mono-boron species [(4-DMAP)2BH2][NTf2] (see Figure S82).

Moving to π nucleophiles, reactions between 3 and naphthalene, N–H-indole, N–Me-indole, or N-methylaniline gave no evidence for C–H borylation at ambient and raised temperatures, although indole hydroboration was observed. Similarly, no C–H borylation was observed for reactions between 4 and N–H-indole, N–Me-indole, or N-methylaniline (with or without 2,6-di-tert-butyl-4-methylpyridine as an exogenous base). As reduction of indoles to indolines is a known reaction achievable with L-BH3,[29] these reactions were not investigated further. We turned our attention next to alkene and alkyne functionalization. When stilbene was combined with 3, this gave no reaction. Hydroboration of stilbene by 4 was observed by 1H NMR spectroscopy; however, the slow rate of the reaction coupled with the formation of multiple species meant we did not pursue this reaction. Terminal alkynes treated with 4 reacted in a faster manner but led to the complex intractable mixtures; however, the reactivity with internal alkynes was cleaner. Reacting 4 with diphenylacetylene led to two new signals in the 11B NMR spectrum at δ11B 40.0 (br, s) and 27.7 (d, 1JBH = 180 Hz), while 1H NMR spectroscopy indicated the formation of a single hydroboration product (a vinylic C–H was observed at δ1H 6.75 ppm). While the reaction is slow at 70 °C, higher temperatures cannot be used due to the limited stability of 4 in solution at raised temperatures. The δ11B = 40 product was also observed using 3, but this reaction was extremely slow (≈5% internal conversion vs an internal standard, after 5 days heating), consistent with the stronger coordination of iodide to boron relative to triflimide. Therefore, only reactivity with NTf2 derivatives is discussed from hereon. The broad signal at δ11B = 40.0 is consistent with the formation of an R2B=NMe2 species,[30] with R in this case presumably vinyl groups formed from the hydroboration of the alkyne to form 10 (Scheme ). Further NMR experiments allowed us to confirm the formation of the divinylaminoborane 10 (see Figures S56–S60). No change in the reaction outcome was observed on repeating in the presence of exogenous base (in an attempt to effect intramolecular aryl C–H borylation after the initial hydroboration).

Scheme 2

Hydroboration of Diphenylacetylene with 4

Internal conversion (vs an internal standard) ≈ 95% (after 4.5 days).

Notably, the second observed signal at δ11B = 27.7 while close to that reported for (Me2N)2BH (δ11B 26–29, 1JBH = 135–139 Hz),[31] exhibited a 1JBH = 180 Hz. This is more consistent with the formation of 11 (Scheme ), which is closely comparable in spectroscopic data to 9. The B2N2 core of 4 has split during formation of 10 and 11, with 10 the product derived from the formal hydroboration of two equivalents of diphenylacetylene with Me2N=BH2, while 11 is then the expected byproduct to maintain mass balance. Stoichiometry studies confirmed the full consumption of all diphenylacetylene, and 4 occurs only at a ratio of 2:1. Crystallization of a reaction mixture post hydroboration enabled isolation (in a very small quantity) and structural characterization of the dimer of 11, compound 6. As discussed above 6 contains two O-bound NTf2, whereas in the monomeric derivatives 11 and 9, NTf2 is N-bound (by 19F NMR spectroscopy and by crystallographic analysis for 9). A sufficient quantity of pure 6 for NMR analysis was not obtainable by fractional crystallization, precluding full analysis. Nevertheless, the observation of O-bound NTf2 in the solid-state structure of 6 is consistent with the NMR data for 4 and calculations (the O-bound isomer of 4 is more stable than the N-bound by 11.5 kcal mol–1), in contrast calculations on the isomers of 11 (11-O and 11, respectively) show that the N-bound NTf2 isomer is more stable by 1.6 kcal mol–1 than 11-O (Figure , top inset), consistent with a less sterically encumbered boron center in 11 (relative to 4).

Figure 7

Inset, relative stability of 11 and 11-O. Bottom, absence of reactivity from 1, Pr2NBH2, and 9 toward diphenyl-acetylene.

The mechanism for the hydroboration of diphenylacetylene starting from 4 at raised temperature could proceed via bis-borane 4, or via mono-boron species derived from the dissociation of 4 at raised temperatures. No significant hydroboration of diphenylacetylene was observed when it was treated with 1 or the mono-boron species Pr2NBH2 indicating the need for an electrophilic borane (Figure , bottom—see Figures S68–S75). More notably, Pr2NBH(NTf2), 9, also did not affect hydroboration (<5% after 4 days at 70 °C). This indicates the necessity of an electrophilic bis-borane (as in 4) for hydroboration to occur in these aminoborane systems. This is supported by the observation of 11 at the end of the reactions starting from 4 (even when using excess diphenylacetylene), indicating that 11 also does not hydroborate diphenylacetylene under these conditions. Therefore, the hydroboration of the first equivalent of diphenylacetylene occurs via the bis-borane 4 and not via mono-borane species from the dissociation of 4. No intermediates between 4 and 10 are observed by NMR spectroscopy; thus, the subsequent steps occur rapidly relative to the first step of the reaction. Decreasing the temperature and the polarity of the solvent slowed the reaction rate; however, the reaction was cleanest in benzene (relative to reactions in haloarenes), therefore only reactions in benzene are discussed.

The disparity in reactivity towards diphenylacetylene of 4 versus 9/11 is notable as it indicates that a four coordinate at boron B-NTf2 species (in 4) is more reactive in hydroboration than a three coordinate at boron B-NTf2 species (in 9/11). However, it should be noted that in the mono-boron R2N=BH(NTf2) compounds there is significant B=NR2 multiple bond character; thus, the reaction of 4 and 9/11 with alkynes are all expected to proceed via displacement of NTf2 (by an SN1 or an SN2 at boron mechanism). Significant B=N character precluding hydroboration is supported by Pr2N=BH2, not effecting alkyne hydroboration under these conditions. If the hydroboration reaction with 4 or 9/11 proceeds via an SN1-type mechanism, then this would require dissociation of NTf2 from boron and formation of borenium (e.g., [D] from 4) or borinium (e.g., [R2N=B–H]+ from 9/11) cations. Borenium cations are much more energetically accessible than borinium cations, and we have recently shown that [R2N=BY]+ borinium cations are extremely high in energy and not feasible intermediates in C–H borylation.[32] If an SN2-type mechanism is operative, the absence of any diphenylacetylene hydroboration starting from 9/11 can be attributed to the stronger binding of NTf2 to boron via nitrogen relative to O-bound NTf2 (as found in bis-borane 4), leading to a higher reaction barrier.

With a better understanding of the reaction, the regioselectivity in the hydroboration of an unsymmetric alkyne was probed. Hydroboration of 1-phenyl-1-propyne afforded a mixture of isomers, 12a, 12b, and 12c (Scheme ), in a ratio of 29:54:17 (see Figure S63) in a conversion of 81%. The Markovnikov/anti-Markovnikov regioselectivity therefore is low using 4 and is comparable to that previously reported for the hydroboration of the same substrate with (2,6-lutidine)BH2I.[18] Attempts to alter the regioselectivity by varying the reaction conditions did not lead to any significant improvement. Furthermore, the use of 5 also led to a mixture of hydroboration isomers but the complexity of the in situ1H NMR spectra (on switching NMe2 for pyrrolidine) precluded determination of the exact ratios in this case. It should be noted that while 10 and 12 proved unstable to silica gel or distillation, fractional crystallization from pentane at low temperature afforded sufficiently pure material to enable the full characterization (see Figures S56–60 and S64–66).

Scheme 3

Synthesis of 12a–c

Internal conversion (vs an internal standard) ≈ 81% (after 5 days).

Finally, we were interested in understanding the greater propensity of these dimeric aminoborane derivatives to cleave to form mono-boron species compared to pyrazabole derivatives as the dissociation of 3 and 4 into mono-boron species is undesirable for their use in borylation-directed borylation.[9] It was found that the dissociation was much more endergonic in the case of pyrazabole than that for [Me2NBH2]2, 1 (Scheme ). This is in accordance with the lower stability observed in reactivity studies with the bis-borane aminoborane systems (e.g., 3 and 4) relative to the pyrazaboles. We attribute this in part to ring strain in dimeric aminoboranes and the greater π donor ability of a NMe2 group relative to a pyrazole unit (where the N lone pair is part of the aromatic system) that helps stabilize the BH2 center in the monomeric form.

Scheme 4

Comparison of the Energy Change during Dissociation of Pyrazabole, 1, and 4 into Their Respective Mono-Boron Species

It should also be noted that substitution of H for NTf2 only alters the energy involved in dissociation of the bis-borane by a small amount (δΔG = 3.8 kcal mol–1, Scheme bottom); thus, it can also can be expected to split into mono-boron species to some degree, particularly on heating. Notably, heating 4 for 3 days in benzene at 100 °C (in a sealed tube) led to the observation (by 11B NMR spectroscopy) of small quantities of 11. The fact that 11 is observed at room temperature from this reaction, that is, it is not consumed to reform 4 on cooling, was surprising based on the DFT calculations. This observation is attributed tentatively to a significant kinetic barrier associated with the linkage isomerism that has to occur to convert from N-bound NTf2 in 11 to form O-bound NTf2 in 4.

Conclusions

In conclusion, a series of dimeric aminoborane-derived electrophiles were synthesized using readily accessible starting material. The optimal coordinating ability of the anion X to boron in H2B(μ-NR2)2BHX electrophiles is crucial, with iodide proving too coordinating (inhibiting reactivity with π nucleophiles), while [B(C6F5)4]− is insufficiently coordinating which leads to further reactivity to form a B3N3-based cation. Triflimide (NTf2) proved to be optimal, enabling bis-borane electrophiles with B2N2 cores to be accessed that do react relatively cleanly with certain π nucleophiles. Thus, hydroboration of internal alkynes was achieved using H2B(μ-NR2)2BH(NTf2) with the B2N2 core maintained during the first hydroboration but subsequently splitting to ultimately produce R2N=B(vinyl)2. Notably, mono-boron analogues, for example, R2NBH(NTf2) exist with N-bound NTf2 in contrast to the bis-boranes and are much less reactive in alkyne hydroboration. This represents an unusual case where the four coordinate at boron species, H2B(μ-NR2)2BH(NTf2), is more reactive than a three coordinate at boron analogue R2NBH(NTf2). Finally, the stability of dimeric bis-boranes with respect to dissociation into monomers needs to be carefully considered for use in borylation-directed borylation, with the B2N2 core in dimeric aminoboranes too weakly bound for that particular application.

Experimental Section

General Materials and Methods

All reactions were performed under inert conditions using standard Schlenk techniques or in an MBraun Unilab glovebox (<0.1 ppm H2O/O2). Unless otherwise stated, solvents were degassed with nitrogen, dried over activated aluminum oxide (Solvent Purification System: Inert PureSolv MD5 SPS), and stored over 3 Å molecular sieves in ampules equipped with Young’s valves. Chlorobenzene, 1,2-difluorobenzene, and 1,2-dichlorobenzene were dried over calcium hydride, distilled, and stored over 3 Å molecular sieves. Deuterated solvents [CDCl3, C6D6, and C6D5Br (99.6% D, Sigma-Aldrich)] were dried and stored over 3 Å molecular sieves. All chemicals were, unless stated otherwise, purchased from commercial sources and used as received. BH3·SMe2 was transferred to an ampule fitted with Young’s valve prior to use. [Ph3C][B(C6F5)4] and lithium 2-Bu-dihydropyridine were synthesized following the literature procedure.[33,34] NMR spectra [1H, 1H{11B}, 11B, 11B{1H}, 13C{1H}, and 19F] were recorded on Bruker Avance III 400 MHz, Bruker Avance III 500 MHz, Bruker Avance III 600 MHz, or Bruker PRO 500 MHz spectrometers. Chemical shifts (δ) are quoted in parts per million (ppm), and coupling constants (J) are given in hertz (Hz) to the nearest 0.5 Hz and as positive values regardless of their real individual signs. 1H and 13C shifts are referenced to the appropriate residual solvent peak, while 11B and 19F shifts are referenced relative to external BF3·Et2O and C6F6, respectively. Abbreviations used are s (singlet), d (doublet), t (triplet), q (quartet), sep (septet), m (multiplet), and br (broad). Background signals in 11B NMR spectra arise to a significant degree from glass components of the probes used in our spectrometers. Unless otherwise stated, all NMR spectra were recorded at 20 °C. Mass spectrometry was performed at the Scottish Instrumentation and Resource Centre for Advanced Mass Spectrometry (SIRCAMS) at the University of Edinburgh using electron impact (EI) or electrospray ionization (ESI) techniques. CHN elemental analyses were carried out by Elemental Microanalysis Ltd. FTIR spectra were recorded on Shimadzu IRAffinity-1S FTIR.

Synthesis of Boron Electrophiles and Precursors

Preparation of [Me2NBH2]2, 1

Borane dimethylamine complex (3.00 g, 51.00 mmol, 1.00 equiv) and lithium 2-Bu-dihydropyridine (0.18 g, 1.30 mmol, 2.5 mol %) were added to a Schlenk flask. Using a bend adapter tube, the flask was connected to a second Schlenk flask, immersed in an ice bath. The mixture was heated overnight at 100 °C with the collection Schlenk flask being left opened to the nitrogen atmosphere. The product started subliming during overnight heating. More product 1 was isolated by further sublimation (heat-gun under N2) as colorless crystals in 55% yield (1.58 g, 13.92 mmol). Analytical data were in accordance with literature values.[35]1H NMR (500 MHz, CDCl3, 300 K): δ 2.64 (1:1:1:1, q, 1JHB = 112 Hz, 4H, BH), 2.45 (s, 12H, N(CH3)2); 11B NMR (160 MHz, CDCl3, 300 K): δ 5.2 (t, 1JBH = 112 Hz, BH); 11B {1H} NMR (160 MHz, CDCl3, 300 K): δ 5.2 (s, BH); 13C {1H} NMR (126 MHz, CDCl3, 300 K): δ 51.93 (s, N(CH3)2).

Preparation of [Pyrrolidine-BH2]2, 2

Neat BH3·Me2S (2.63 mL, 27.70 mmol, 1.00 equiv) was slowly added to a solution of pyrrolidine (2.00 g, 27.70 mmol, 1.00 equiv) in pentane (30 mL) at room temperature. The solution was stirred for 1 h. Volatiles were removed in vacuo affording the pyrrolidine·BH3 adduct as a white solid. Lithium 2-Bu-dihydropyridine (0.10 g, 0.69 mmol, 2.5 mol %) was added to the crude pyrrolidine·BH3 adduct, and the mixture was heated overnight at 100 °C in an unpressurized system. Product 2 was isolated by distillation (100 °C, 10–3 to 10–2 mbar) as a colorless waxy solid in 80% yield (1.84 g, 11.09 mmol). Analytical data were in accordance with literature values.[36]1H NMR (500 MHz, CDCl3, 300 K): δ 2.86 (br, t, 8H, NCH2CH2), 2.61 (1:1:1:1, q, 1JHB = 112 Hz, 4H, BH), 1.76–1.66 (m, 8H, NCH2CH2); 11B NMR (160 MHz, CDCl3, 300 K): δ 3.0 (t, 1JBH = 112 Hz, BH); 11B {1H} NMR (160 MHz, CDCl3, 300 K): δ 3.0 (s, BH); 13C {1H} NMR (126 MHz, CDCl3, 300 K): δ 60.08 (s, NCH2CH2), 23.62 (s, NCH2CH2).

Preparation of H2B(μ-Me2N)2BH(I), 3

Iodine (0.54 g, 2.11 mmol, 0.48 equiv) was dissolved in benzene (10 mL) and slowly added to a solution of [Me2NBH2]2 (0.50 g, 4.39 mmol, 1.00 equiv) in benzene (10 mL) at room temperature. The resulting solution was stirred at room temperature for 30 min. The volatiles were removed under vacuum with care, to avoid the solid product subliming during the process. The product was extracted with pentane (5 mL). After vacuuming the volatiles, the product 3 was isolated by sublimation under vacuum (heat-gun, 10–3 to 10–2 mbar) as a colorless crystalline solid in 34% yield (0.35 g, 1.47 mmol). 1H NMR (500 MHz, C6D6, 300 K): δ 4.37 (1:1:1:1, q, 1JHB = 142 Hz, 1H, BHI), 2.88 (1:1:1:1, q, 1JHB = 119 Hz, 1H, BH), 2.69 (1:1:1:1, q, 1JHB = 115 Hz, 1H, BH), 2.22 (s, 6H, (NCH3)2), 2.00 (s, 6H, (NCH3)2); 11B NMR (160 MHz, C6D6, 300 K): δ 3.9 (t, 1JBH = 118 Hz, BH2), 0.4 (d, 1JBH = 142 Hz, BHI); 11B {1H} NMR (160 MHz, C6D6, 300 K): δ 3.9 (s, BH2), 0.4 (s, BHI); 13C {1H} NMR (126 MHz, C6D6, 300 K): δ 51.24 (s, (NCH3)2), 49.86 (s, (NCH3)2) (see Figure S1). Elemental analysis: calculated for C4H15B2N2I: C 20.04%, H 6.31%, N 11.69%; observed: C 20.67%, H 6.46%, N 11.53%. IR: (νmax (neat)/cm–1) 2499 (B–H), 2432 (B–H), 2358 (B–H).

Preparation of H2B(μ-Me2N)2BH(OSOCF3NTf), 4

HNTf2 (1.18 g, 4.18 mmol, 1.00 equiv) was dissolved in benzene (10 mL) and slowly added to a solution of [Me2NBH2]2 (0.50 g, 4.39 mmol, 1.05 equiv) in benzene (10 mL) at room temperature. The solution was stirred at room temperature for 24 h. The volatiles were removed under vacuum. The product was extracted with pentane (10 mL). Drying under vacuum for 1 h afforded the product 4 as a colorless oil in 83% yield (1.37 g, 3.48 mmol). 1H NMR (500 MHz, C6D6, 300 K): δ 3.14 (br, q, 1JHB = 138 Hz, 1H, B(H)OS(O)(CF3)NTf), 2.33 (br, q, 1JHB = 123 Hz, 2H, BH2), 1.96 (s, 3H, CH3), 1.88 (s, 3H, CH3), 1.87 (s, 3H, CH3), 1.79 (s, 3H, CH3); 1H {11B} NMR (500 MHz, C6D6, 300 K): δ 3.14 (br, s, 1H, B(H)OS(O)(CF3)NTf), 2.33 (br, s, 2H, BH2), 1.97 (s, 3H, CH3), 1.88 (s, 3H, CH3), 1.87 (s, 3H, CH3), 1.80 (s, 3H, CH3); 11B NMR (160 MHz, C6D6, 300 K): δ 4.9 (d, 1JBH = 142 Hz, B(H)OS(O)(CF3)NTf), 3.3 (t, 1JBH = 119 Hz, BH2); 11B {1H} NMR (160 MHz, C6D6, 300 K): δ 4.9 (s, B(H)OS(O)(CF3)NTf), 3.3 (s, BH2); 13C {1H} NMR (126 MHz, C6D6, 300 K): δ 119.99 (q, 1JCF = 320 Hz, CF3), 119.43 (q, 1JCF = 321 Hz, CF3), 49.40 (s, CH3), 49.32 (s, CH3), 44.48 (s, CH3), 44.43 (s, CH3); 19F NMR (471 MHz, C6D6, 300 K): δ −75.7 (s, 3F, CF3), −78.5 (s, 3F, CF3). Elemental analysis: calculated for C6H15B2F6N3O4S2: C 18.34%, H 3.85%, N 10.69%; observed: C 18.40%, H 3.62%, N 10.50%. IR: (νmax (neat)/cm–1) 2515 (B–H), 2474 (B–H), 2372 (B–H).

Preparation of H2B(μ-pyrrolidine)2BH(OSOCF3NTf), 5

HNTf2 (1.21 g, 4.30 mmol, 1.00 equiv) was dissolved in benzene (10 mL) and slowly added to a solution of [pyrrolidine-BH2]2 (0.75 g, 4.52 mmol, 1.05 equiv) in benzene (10 mL) at room temperature. The solution was stirred at room temperature for 20 h. The volatiles were removed under vacuum. The product was extracted with pentane (10 mL). Drying under vacuum overnight at 40 °C afforded the product 5 as a colorless oil in 65% yield (1.24 g, 2.79 mmol). 1H NMR (500 MHz, C6D6, 300 K): δ 4.00–2.90 (br, m, 1H, B(H)OS(O)(CF3)NTf), 2.88–1.96 (m, 10H, NCH2CH2 and BH2), 1.35–1.11 (m, 8H, NCH2CH2); 1H {11B} NMR (500 MHz, C6D6, 300 K): δ 3.43 (br, s, 1H, B(H)OS(O)(CF3)NTf), 2.88–1.96 (m, 10H, NCH2CH2 and BH2), 1.35–1.11 (m, 8H, NCH2CH2); 11B NMR (160 MHz, C6D6, 300 K): δ 5.4–1.1 (m, B(H)OS(O)(CF3)NTf, BH2); 11B {1H} NMR (160 MHz, C6D6, 300 K): δ 3.5 (s, B(H)OS(O)(CF3)NTf), 2.8 (s, BH2); 13C {1H} NMR (126 MHz, C6D6, 300 K): δ 120.05 (q, 1JCF = 320 Hz, CF3), 119.49 (q, 1JCF = 322 Hz, CF3), 58.62 (s, NCH2CH2), 58.55 (s, NCH2CH2), 53.85 (s, NCH2CH2), 22.89 (s, NCH2CH2), 22.88 (s, NCH2CH2), 22.79 (s, NCH2CH2), 22.78 (s, NCH2CH2); 19F NMR (471 MHz, C6D6, 300 K): δ −75.7 (s, 3F, CF3), −78.4 (s, 3F, CF3). Elemental analysis: calculated for C10H19B2F6N3O4S2: C 26.99%, H 4.30%, N 9.44%; observed: C 26.94%, H 4.18%, N 9.39%. IR: (νmax (neat)/cm–1) 2499 (B–H), 2451 (B–H), 2374 (B–H).

Preparation of [(Me2N)3B3H5][B(C6F5)4], 7

(Me2NBH2)2 (0.03 g, 0.22 mmol, 3.00 equiv) and [Ph3C][B(C6F5)4] (0.14 g, 0.15 mmol, 2 equiv) were dissolved in PhCl (3 mL) and heated to 60 °C until the solution turned colorless (30 min). The solution was carefully layered with hexane (5 mL). After 21 days, the formed colorless crystals were washed with hexane (2 × 2 mL) and dried in vacuo. The product 7 was isolated as colorless crystals in 70% yield (0.09 g, 0.10 mmol). 1H NMRa,b (500 MHz, C6H4F2, 300 K): δ 3.45–2.14 (br, m, 2H, HB(H)BH), 3.45–1.74 (br, m, 18H, N(CH3)2), 2.83–1.74 (br, m, 2H, BH2), 1.72–1.10 (br, m, 1H, HB(H)BH); 1H NMRa (500 MHz, C6H4F2, 278 K): δ 3.45–2.14 (br, m, 2H, HB(H)BH), 3.13 (s, 3H, N(CH3)), 2.83–1.74 (br, m, 2H, BH2), 2.55 (s, 6H, N(CH3)), 2.47 (s, 6H, N(CH3)), 2.36 (s, 3H, N(CH3)), 1.72–1.10 (br, m, 1H, HB(H)BH); 1H {11B} NMRa,b (500 MHz, C6H4F2, 300 K): δ 3.45–2.14 (br, m, 2H, HB(H)BH), 3.45–1.74 (br, m, 18H, N(CH3)2), 2.83–1.74 (br, m, 2H, BH2), 1.52 (br, s, 1H, HB(H)BH); 11B NMRa (160 MHz, C6H4F2, 300 K): δ 3.4 (t, 1JBH = 120 Hz, BH2), −6.1 (br, d, 1JBH = 164 Hz, HB(H)BH), −16.2 (s, B(C6F5)4); 11B {1H} NMRa (160 MHz, C6H4F2, 300 K): δ 3.4 (s, BH2), −6.1 (s, HB(H)BH), −16.2 (s, B(C6F5)4); 13C {1H} NMRa (126 MHz, C6H4F2, 300 K): δ 148.8 (br, m, CF), 140.6 (br, t, CF), 138.6 (br, m, CF), 136.7 (br, m, CF), 54.3 (br, s, N(CH3)), 49.8 (br, s, N(CH3)), 46.4 (br, s, N(CH3)), 40.1 (br, s, N(CH3)); 19F NMRa (471 MHz, C6H4F2, 300 K): δ −132.5 (br, s, CF), −163.9 (t, 3JFF = 20 Hz, CF), −167.7 (br, t, 3JFF = 17 Hz, CF).a Due to solubility and stability issues, NMR data of 7 were recorded in 1,2-difluorobenzene (C6H4F2). Reference NMR experiments using SiMe4 were carried out to determine the 1H and 13C shifts of 1,2-difluorobenzene. Hydrogen atoms located on boron centers [BH2, HB(H)BH, and HB(H)BH] were identified by the 2D 11B–1H HMQC experiment. Mass spectrum: HRMS (ESI+) m/z: calcd for C6H23B3N3+: 170.21657; found: 170.21576. IR: (νmax (neat)/cm–1) 2538 (B–H), 2465 (B–H), 2403 (B–H).

Preparation of 10

A solution of H2B(μ-Me2N)2BH(OSOCF3NTf), 4 (0.22 g, 0.56 mmol, 1.00 equiv) in benzene (1 mL) was added to a solution of diphenylacetylene (0.20 g, 1.12 mmol, 2.00 equiv) in benzene (1 mL). The resulting solution was heated at 70 °C for 1 week. While heating, the solution turned slowly from colorless to dark orange. The volatiles were removed under vacuum, affording an oil. The oil was extracted with pentane (2 mL), giving a clear orange solution. The solution was cooled to −35 °C and filtered at this temperature. Removal of volatiles in vacuo afforded the product 10 as an orange oil (0.11 g), contaminated with remaining trace of −NTf2 side-products. 1H NMR (500 MHz, C6D6, 300 K): δ 7.23–7.19 (m, 4H), 7.15–7.12 (m, 4H), 7.12–7.08 (m, 4H), 7.03–6.99 (m, 2H), 6.99–6.94 (m, 4H), 6.92–6.88 (m, 2H), 6.75 (s, 2H), 2.68 (s, 6H); 11B NMR (160 MHz, C6D6, 300 K): δ 40.0 (br, s, BN(Me)2); 13C {1H} NMR (126 MHz, C6D6, 300 K): δ 147.43 (br), 142.95, 138.50, 134.72, 129.77, 129.03, 128.82, 128.27a, 126.90, 126.29, 40.99. Mass spectrum: HRMS (ESI+) m/z: calcd for C30H28BN: 413.23093; found: 413.23142.