Reactions of Dihaloboranes with Electron-Rich 1,4-Bis(trimethylsilyl)-1,4-diaza-2,5-cyclohexadienes

The reactions of electron-rich organosilicon compounds 1,4-bis(trimethylsilyl)-1,4-diaza-2,5-cyclohexadiene (1), 2,3,5,6-tetramethyl-1,4-bis(trimethylsilyl)-1,4-diaza-2,5-cyclohexadiene (2), and 1,1′-bis(trimethylsilyl)-1,1′-dihydro-4,4′-bipyridine (12) with B-amino and B-aryl dihaloboranes afforded a series of novel B=N-bond-containing compounds 3–11 and 13. The B=N rotational barriers of 7 (>71.56 kJ/mol), 10 (58.79 kJ/mol), and 13 (58.65 kJ/mol) were determined by variable-temperature 1H-NMR spectroscopy, thus reflecting different degrees of B=N double bond character in the corresponding compounds. In addition, ring external olefin isomers 11 were obtained by a reaction between 2 and DurBBr2. All obtained B=N-containing products were characterized by multinuclear NMR spectroscopy. Compounds 5, 9, 10a, 11, and 13a were also characterized by single-crystal X-ray diffraction analysis.

Mashima et al. reported a class of electron-rich organosilicon compounds 1, 2, and 12, which can serve as versatile reducing reagents for the group 4-6 metal chloride complexes. The corresponding low-valent metal species were prepared in a salt-free manner [27][28][29][30][31][32][33]. The reducing power mainly derives from the aromatization of the central 1,4-diaza-2,5-cyclohexadiene ring. Deeply inspired by the advantage of the salt-free reduction protocol and easy workup, we decided to examine the ability of the organosilicon compounds 1, 2, and 12 for the reduction of trivalent dihaloboranes. Based on the published results, the disubstituted compounds ArXB(N 2 C 4 R 4 )BXAr are proposed as the reduction products. We hypothesized two possible bonding modes (i.e., A and B in Scheme 1) between the C 4 N 2 ring and the boron atoms. In the first manner, B-N is bound by an electron-precise σ bond (A, Scheme 1) and an additional N-B dative π bond. In the second manner, two nitrogen atoms each

Results and Discussion
First, we examined compound 1 for its ability to reduce ArBX2. The results are summarized in Scheme 2. Compound 1 [34] and ArBX2 (Ar = 2,3,5,6-tetramethylphenyl (Dur), 2,4,6-trimethylphenyl (Mes)) [35] were prepared according to the literature. The reaction of 1 with an equimolar amount of DurBBr2 and MesBCl2 at ambient temperature afforded the expected monosubstituted products 3 and 4, respectively. Adding the second equiv. of dihaloboranes to the reaction mixture led to the disubstituted products 5 and 6. In stark contrast, the reaction of 1 with an equimolar amount of the less sterically demanding PhBCl2 caused precipitation, which is insoluble in all ordinary solvents. This is most likely due to the polymerization of PhClBC4N2H4SiMe3 by chlorosilane elimination. Hence, the stepwise synthetic protocol is unsuitable for the synthesis of 7. Instead, 1 was directly treated with 2 equiv. of PhBCl2 at room temperature (RT), affording 7 in an acceptable yield (48%). Hence, the reaction of the monosubstituted intermediate (i.e., PhClBC4N2H4SiMe3) with PhBCl2 should be much faster than the self-polymerization process. Compounds 3-7 were confirmed by NMR spectroscopic (Figure S1-S15) and HRMS studies. Furthermore, the multinuclear NMR spectroscopic study revealed that the isolated 5-7 all consist of ca. 1:1 cis-trans isomers in the solution phase at ambient temperature due to the nonrotatable B=N double bond (see Electronic Supporting Information (ESI)).

Scheme 1.
Proposed products from the reactions of ArBX 2 with 1 and 2, and two possible bonding modes A and B between the central C 4 N 2 ring and the boron centers.

Results and Discussion
First, we examined compound 1 for its ability to reduce ArBX 2 . The results are summarized in Scheme 2. Compound 1 [34] and ArBX 2 (Ar = 2,3,5,6-tetramethylphenyl (Dur), 2,4,6-trimethylphenyl (Mes)) [35] were prepared according to the literature. The reaction of 1 with an equimolar amount of DurBBr 2 and MesBCl 2 at ambient temperature afforded the expected monosubstituted products 3 and 4, respectively. Adding the second equiv. of dihaloboranes to the reaction mixture led to the disubstituted products 5 and 6. In stark contrast, the reaction of 1 with an equimolar amount of the less sterically demanding PhBCl 2 caused precipitation, which is insoluble in all ordinary solvents. This is most likely due to the polymerization of PhClBC 4 N 2 H 4 SiMe 3 by chlorosilane elimination. Hence, the stepwise synthetic protocol is unsuitable for the synthesis of 7. Instead, 1 was directly treated with 2 equiv. of PhBCl 2 at room temperature (RT), affording 7 in an acceptable yield (48%). Hence, the reaction of the monosubstituted intermediate (i.e., PhClBC 4 N 2 H 4 SiMe 3 ) with PhBCl 2 should be much faster than the self-polymerization process. Compounds 3-7 were confirmed by NMR spectroscopic (Figures S1-S15) and HRMS studies. Furthermore, the multinuclear NMR spectroscopic study revealed that the isolated 5-7 all consist of ca. 1:1 cis-trans isomers in the solution phase at ambient temperature due to the nonrotatable B=N double bond (see Electronic Supporting Information (ESI)).
Suitable single crystals of 5 for X-ray diffraction analysis were obtained by slow evaporation of a saturated hexane solution. Two isomers, 5a and 5b, co-crystallized in the unit cell. The result is depicted in Figure 1 and Figure S37. The central C 4 N 2 ring is nearly planar. The endocyclic N1-C2 (1.385(7) Å), N2-C3 (1.415(7) Å), C1-C2 (1.311(9) Å), and C3-C3* (1.326(8) Å) distances lie in the expected range for N-C single bonds and C=C double bonds. The bond lengths of B1-N1 (1.423(8) Å) and B2-N2 (1.400(8) Å) are shorter than that of a B-N single bond, which is indicative of a significant B=N double bond character. All these geometric parameters suggest the bonding mode A in Scheme 1. Therefore, both boron centers adopt a formal oxidation state of +3. Suitable single crystals of 5 for X-ray diffraction analysis were obtained by slow evaporation of a saturated hexane solution. Two isomers, 5a and 5b, co-crystallized in the unit cell. The result is depicted in Figure 1     Suitable single crystals of 5 for X-ray diffraction analysis were obtained by slow evaporation of a saturated hexane solution. Two isomers, 5a and 5b, co-crystallized in the unit cell. The result is depicted in Figure 1    Compound 2, which features a less-negative redox potential (+0.10 V) with respect to 1 (−0.24 V), was further examined to reduce PhBCl 2 and DurBBr 2 (Scheme 2). Differing from the aforementioned reactions with 1, both monosubstituted products 8 and 9 could be prepared upon a 1:1 ratio reaction of 2 with PhBCl 2 and DurBBr 2 , respectively. Upon the reaction of 2 with two equiv. of PhBCl 2 at RT, the disubstituted compounds 10a and 10b were obtained as 1:1 cis-trans isomers. Surprisingly, treatment of 2 with two equiv. of DurBBr 2 at ambient temperature led to the formation of 11, which can be regarded as the product from an isomerization of 11 [36,37]. Compounds 8-11 were confirmed by NMR spectroscopic (Figures S16-S28) and HRMS studies. There were two sets (intensity ratio of ca. 1:0.3) of 1 H signals between 4 and 6 ppm, each consisting of three multiplets with the integration ratio of 1:1:1, which can be assigned to the migrated H and two remaining olefinic protons. These are the Molecules 2020, 25, 2875 4 of 11 most characteristic signs for the formation of the isomerized product. After assigning each peak (with the help of the NOE spectrum, see the ESI Figure S26 for more details), we could determine that the ratio of isomers 11a and 11b was 77:23. Furthermore, the isomerization was also observed upon the treatment of the isolated 9 with an equimolar amount of DurBBr 2 at RT.
aforementioned reactions with 1, both monosubstituted products 8 and 9 could be prepared upon a 1:1 ratio reaction of 2 with PhBCl2 and DurBBr2, respectively. Upon the reaction of 2 with two equiv. of PhBCl2 at RT, the disubstituted compounds 10a and 10b were obtained as 1:1 cis-trans isomers. Surprisingly, treatment of 2 with two equiv. of DurBBr2 at ambient temperature led to the formation of 11, which can be regarded as the product from an isomerization of 11′ [36,37]. Compounds 8-11 were confirmed by NMR spectroscopic (Figure S16-S28) and HRMS studies. There were two sets (intensity ratio of ca. 1:0.3) of 1 H signals between 4 and 6 ppm, each consisting of three multiplets with the integration ratio of 1:1:1, which can be assigned to the migrated H and two remaining olefinic protons. These are the most characteristic signs for the formation of the isomerized product. After assigning each peak (with the help of the NOE spectrum, see the ESI Figure S26 for more details), we could determine that the ratio of isomers 11a and 11b was 77:23. Furthermore, the isomerization was also observed upon the treatment of the isolated 9 with an equimolar amount of DurBBr2 at RT.
The structures of 9, 10a, 11a, and 11b were confirmed by single-crystal X-ray diffraction analysis (Figure 2 and S38-S40). All four compounds adopt a boat conformation, which could be explained by the small energy difference between the planar and nonplanar geometry of the C4N2 ring, and the steric congestion between the central exocyclic methyl groups and the bulky boron substituents. Bond lengths (Å) of 9 (B1-N1 1.376 (7), N1-C1 1.451(6), C1-C2 1.343 (7), N2-C2 1.434(6), N2-Si1 1.759(5)), 10 (B1-N1 1.401(2), N1-C2 1.4494 (17), C1-C2 1.328(2), N1-C1* 1.4513 (18)) are all as expected. The overall structures of 11a and 11b resemble that of 10a. However, since the C3 position in 11a and 11b accepted one H atom from the methyl group at the C2 position, respectively, and thus became sp 3 -hybridized, the torsion angles C4-C3-N2-B2 (11a: 94.25°; 11b: 94.69°) are notably greater than those at the other three carbon positions (57-63°) in the central six-membered ring. Due to the disordered nature of the crystal, the bond lengths of 11a and 11b cannot be further discussed.  The reaction of (SiMe 3 ) 2 NBCl 2 with 12 [38] of greater reducing power (redox potential of −0.40 V) [39] was performed at ambient temperature in C 6 D 6 . After the removal of the solvent and extraction with hexane, an NMR spectroscopically pure product 13 was obtained with a yield of 75%. Compound 13 was confirmed by NMR spectroscopic (Figures S29-S31) and HRMS studies. Suitable single crystals of 13a for X-ray diffraction analysis were obtained upon storage of the reaction mixture overnight at RT (Figure 3 and Figure S41). The N1-C1/N1-C5 (1.402 (8) with hexane, an NMR spectroscopically pure product 13 was obtained with a yield of 75%. Compound 13 was confirmed by NMR spectroscopic (Figure S29-S31) and HRMS studies. Suitable single crystals of 13a for X-ray diffraction analysis were obtained upon storage of the reaction mixture overnight at RT (Figure 3 and S41). The N1-C1/N1-C5 (1.402 (8)   Apparently, both the RT-NMR spectroscopic and crystallographic studies failed to prove any successful reduction of the trivalent borane to divalent boron radical. Since the rotational barrier around an N-B dative bond should be lower than that of a B=N double bond, we assumed that any contribution from the bonding mode B (Scheme 1) should slightly lower the rotational barrier around the exocyclic B-N bond. In this context, we conducted a variable-temperature 1 H-NMR experiment to provide further insight. Toluene-d 8 was selected as the solvent with a temperature ranging from −60 °C to 80 °C. In general, the exocyclic H or Me as marked in Figure 4 (top right) should display two signals if the B=N bond is nonrotatable. The separated signals will coalesce at an elevated temperature when the B-N bond overcomes the rotational barrier and begins to rotate. Determination of the separation (Hz) of two signals and the coalescent temperature allows calculation of the B-N rotational barrier. The results of the VT-NMR experiments and assignment of the signals of interest are depicted in Figure 4 and S32-S36. The obtained ΔG ≠ values are summarized in Table 1. Analysis of the VT-NMR spectra revealed 7 with a strong B=N bond, and 10 and 13 with weak B=N bonds, as reflected by their rotational barriers >71.56 kJ/mol (7), 58.79 kJ/mol (10), 58.65 kJ/mol (13) when compared with ordinary B=N double bonds (71-100 kJ/mol) [40]. When taking the aforementioned assumption into account, the remarkably lower B-N rotational barrier of 10 compared to that of 7 is not in line with the fact that the reducing power of 2 is slightly weaker than that of 1, according to   Apparently, both the RT-NMR spectroscopic and crystallographic studies failed to prove any successful reduction of the trivalent borane to divalent boron radical. Since the rotational barrier around an N-B dative bond should be lower than that of a B=N double bond, we assumed that any contribution from the bonding mode B (Scheme 1) should slightly lower the rotational barrier around the exocyclic B-N bond. In this context, we conducted a variable-temperature 1 H-NMR experiment to provide further insight. Toluene-d 8 was selected as the solvent with a temperature ranging from −60 °C to 80 °C. In general, the exocyclic H or Me as marked in Figure 4 (top right) should display two signals if the B=N bond is nonrotatable. The separated signals will coalesce at an elevated temperature when the B-N bond overcomes the rotational barrier and begins to rotate. Determination of the separation (Hz) of two signals and the coalescent temperature allows calculation of the B-N rotational barrier. The results of the VT-NMR experiments and assignment of the signals of interest are depicted in Figure 4 and S32-S36. The obtained ΔG ≠ values are summarized in Table 1. Analysis of the VT-NMR spectra revealed 7 with a strong B=N bond, and 10 and 13 with weak B=N bonds, as reflected by their rotational barriers >71.56 kJ/mol (7), 58.79 kJ/mol (10), 58.65 kJ/mol (13) when compared with ordinary B=N double bonds (71-100 kJ/mol) [40]. When taking the aforementioned assumption into account, the remarkably lower B-N rotational barrier of 10 compared to that of 7 is not in line with the fact that the reducing power of 2 is slightly weaker than that of 1, according to Apparently, both the RT-NMR spectroscopic and crystallographic studies failed to prove any successful reduction of the trivalent borane to divalent boron radical. Since the rotational barrier around an N-B dative bond should be lower than that of a B=N double bond, we assumed that any contribution from the bonding mode B (Scheme 1) should slightly lower the rotational barrier around the exocyclic B-N bond. In this context, we conducted a variable-temperature 1 H-NMR experiment to provide further insight. Toluene-d 8 was selected as the solvent with a temperature ranging from −60 • C to 80 • C. In general, the exocyclic H or Me as marked in Figure 4 (top right) should display two signals if the B=N bond is nonrotatable. The separated signals will coalesce at an elevated temperature when the B-N bond overcomes the rotational barrier and begins to rotate. Determination of the separation (Hz) of two signals and the coalescent temperature allows calculation of the B-N rotational barrier. The results of the VT-NMR experiments and assignment of the signals of interest are depicted in Figure 4 and Figures S32-S36. The obtained ∆G values are summarized in Table 1. Analysis of the VT-NMR spectra revealed 7 with a strong B=N bond, and 10 and 13 with weak B=N bonds, as reflected by their rotational barriers >71.56 kJ/mol (7), 58.79 kJ/mol (10), 58.65 kJ/mol (13) when compared with ordinary B=N double bonds (71-100 kJ/mol) [40]. When taking the aforementioned assumption into account, the remarkably lower B-N rotational barrier of 10 compared to that of 7 is not in line with the fact that the reducing power of 2 is slightly weaker than that of 1, according to the CV data. Therefore, the lower B-N rotational barrier in 10 should be mainly due to its boat conformation, which allows for less steric hindrance. Furthermore, although the central C 4 N 2 rings of 7 and 13 both adopt a planar structure, the B-N rotational barrier in 13 is significantly lower than that of 7. This finding could be explained by the competition in π donation from another B-amino function (-N(SiMe 3 ) 2 ) in 13.
the CV data. Therefore, the lower B-N rotational barrier in 10 should be mainly due to its boat conformation, which allows for less steric hindrance. Furthermore, although the central C4N2 rings of 7 and 13 both adopt a planar structure, the B-N rotational barrier in 13 is significantly lower than that of 7. This finding could be explained by the competition in π donation from another B-amino function (-N(SiMe3)2) in 13.

General Information
All manipulations were performed under dry argon using standard Schlenk line or glovebox techniques. Solvents were purified by distillation from Na under dry argon. C 6 D 6 was dried over an Na/K alloy and then degassed by freeze-pump-thaw cycles. PhBCl 2 was purchased from Beijing MREDA Technologie Co., Ltd., without any special treatment before use. The NMR spectra were acquired on a Bruker AVANCE 400 ( 1 H: 400 MHz, 13 C{ 1 H}: 101 MHz, 11 B: 128 MHz) NMR spectrometer at 298 K. Variable-temperature NMR experiments were conducted on a Bruker AVANCE 400 NMR spectrometer ( 1 H: 400 MHz, 213-353 K). Chemical shifts are given in ppm. 1 H and 13 C{ 1 H} NMR spectra were referenced to an external tetramethylsilane (TMS) via the residual protons of the solvent ( 1 H) or the solvent itself ( 13 C{ 1 H}). 11 B NMR spectra were referenced to the external BF 3 ·OEt 2 . High-resolution mass spectrometry (HMRS) was performed with a Thermo Fisher Scientific Q Exactive Mass Spectrometer (MS) system.

Synthesis of 11a and 11b
In the glove box, 9 (37.6 mg, 0.1 mmol, 1 equiv.) and DurBBr 2 (30.3 mg, 1 mmol, 1.0 equiv.) were added into C 6 D 6 (0.6 mL) in a J. Young NMR tube. The mixture was rested overnight prior to the removal of the volatiles under vacuum to get 11 as a yellow oil with a yield of 76%. Mixture 11 contains two olefin isomers, 11a and 11b, with a ratio of about 1:0.3

Conclusions
In summary, the reactions of electron-rich organosilicon compounds 1, 2, and 12 with various B-amino and B-aryl dihaloboranes were comprehensively studied. No direct evidence for the presence of divalent boron radical character could be obtained from NMR spectra and single-crystal structures. The rotational barrier around the exocyclic B-N bonds was studied by VT 1 H-NMR spectroscopy, which revealed relatively small barriers for 10 and 13. The steric hindrance as well as the competition from additional B-amino functions were the main factors affecting the B-N rotational barrier. In addition, the reaction between 2 and DurBBr 2 resulted in 11 via an isomerization process. Although this study does not access the desired biradial species, we believe that the novel B=N-containing products could act as an RXB• source upon the liberation of the aromatic linker (i.e., pyrazine and 4,4 -bipyridine). Studies of the mechanism of the isomerization reaction, as well as the application of 10 and 13 as RXB• transfer reagents to unsaturated organic substrates, are currently underway in our laboratory, and will be reported in due course.