Competition for Hydride Between Silicon and Boron: Synthesis and Characterization of a Hydroborane‐Stabilized Silylium Ion

Abstract Potent main‐group Lewis acids are capable of activating element‐hydrogen bonds. To probe the rivalry for hydride between silylium‐ and borenium‐ion centers, a neutral precursor with the hydrosilane and hydroborane units in close proximity on a naphthalene‐1,8‐diyl platform was designed. Abstraction of one hydride leads to a hydroborane‐stabilized silylium ion rather than a hydrosilane‐coordinated borenium ion paired with [B(C6F5)4]− or [HCB11Cl11]− as counteranions. Characterization by multinuclear NMR spectroscopy and X‐ray diffraction supported by DFT calculations reveals a cationic, unsymmetrical open three‐center, two‐electron (3c2e) Si−H−B linkage.

In 1996, Piers reported the ability of the strong boron Lewis acid tris(pentafluoro)phenylborane to catalyze the hydrosilylation of carbonyl compounds. [1] Experimental studies by him [2] and our laboratory [3] along with a subsequent computational analysis [4] indicated that B(C 6 F 5 ) 3 tends to active the SiÀ H bond of the hydrosilane rather than forming a conventional Lewis adduct with the σ-basic carbonyl donor. [5] Yet, the assumed borane/hydrosilane intermediate has remained experimentally elusive. [6] Piers, Tuononen and co-workers eventually achieved the isolation of the related adduct 1 by employing 1,2,3tris(pentafluorophenyl)-4,5,6,7-tetrafluoro-1-boraindene instead of B(C 6 F 5 ) 3 (Scheme 1, top). [7] Since 2014, additional examples of intermolecular SiÀ H bond activation with Al(C 6 F 5 ) 3 as in 2, [8] a borenium ion as in 3[B(C 6 F 5 ) 4 ] [9] as well as a neutral borane as in 4 [10] have been disclosed. The understanding of these intermediates is highly relevant to catalysis, especially in the case of Piers' chemistry. [11] The silicon and boron centers compete for the hydride in these Lewis pairs, resulting in highly interesting bonding situations. Wang's cationic complex 3 + is a previously unprecedented example of an η 2 -coordination of the SiÀ H bond to a Lewis acidic boron atom.
To interrogate this "competition for hydride", we designed the neutral precursor 9 with the SiÀ H and BÀ H bonds in the same molecule in close proximity to arrive at the Si/B hydro-nium ion 10 + after treatment with the trityl cation (Scheme 1, gray box). Such systems based on the naphthalene-1,8-diyl platform have already been utilized by Katz (B/B; K [5]), [12] Gabbaï [13] as well as Suzuki [14] (C/C; 6[BF 4 ]), and Müller (Si/Si; 7[B(C 6 F 5 ) 4 ]) [15] (Scheme 1, bottom). Of note, there is only one example with two different hydride acceptors, that is a Ge/Si hydronium borate 8[B(C 6 F 5 ) 4 ] described by Müller and coworkers for which no crystallographic characterization is available. [16] The key question of our present investigation is whether the Si/B hydronium ion 10 + is a hydrosilane adduct of a borenium ion or a hydroborane-stabilized silylium ion. By this, we are bridging our long-time expertises with Piers-type chemistry [11] and that of silylium ions. [17] The neutral precursor 9 was synthesized in 24 % yield by lithiation of (8-bromonaphthalen-1-yl)diisopropylsilane, followed by the addition of a toluene suspension of IMe·BH 2 I (IMe = 1,3-dimethylimidazol-2-ylidene) [18] at À 78°C (see the Supporting Information for details). The δ( 11 B) NMR resonance of 9 in C 6 D 6 appears as a triplet at À 23.7 ppm with a 1 J B,H coupling constant of 87 Hz. This is lowfield relative to δ( 11 B) À 31.8 ppm for IMe·BH 2 I and in the range of arylated NHCboranes. [19] The δ( 29 Si) NMR signal is observed at 18.9 ppm, and the 1 J Si,H coupling constant is 182 Hz. Colorless crystals of precursor 9 suitable for X-ray diffraction were obtained from a concentrated CH 2 Cl 2 /n-hexane solution (2 : 1) at À 30°C overnight ( Figure 1). [20] The SiÀ H bond length of 1.42(2) Å is in the typical range of SiÀ H bonds (ca. 1.425 Å) [21] and heading away from the boron atom. The distance between the silicon and the boron atoms is 3. 19(2) Å, which is longer than the typical range of SiÀ B single bonds (1.91 Å-2.12 Å) [22] but still within the sum of their van der Waals radii as a result of the steric congestion imposed by the rigid, peri-substituted naphthalene backbone. The repulsion of the silyl and NHC-boryl moieties can be seen from the deviation of C6À C1À Si1 (130.7(1)°) and C6À C10À B1 (123.0(1)°) angles from the ideal value 120°. Those tight steric constraints likely account for the moderate chemical stability of compound 9 which slowly decomposes within weeks even when kept in the glovebox.
Treatment of precursor 9 with 1.0 equiv. of [Ph 3 C][B(C 6 F 5 ) 4 ] in C 6 D 6 led to a biphasic mixture (Scheme 2). The phases were allowed to separate, and the upper phase was removed. The lower phase containing the cationic product 10[B(C 6 F 5 ) 4 ] was washed three times with a few drops of C 6 D 6 and then dissolved in 1,2-Cl 2 C 6 D 4 for NMR spectroscopic characterization. The chemical shift of the silicon atom in the 29 Si NMR spectrum of 10[B(C 6 F 5 ) 4 ] is significantly low-field shifted compared to the precursor 9 [δ( 29 Si) 56.0 ppm versus 18.9 ppm]. Moreover, this value is close to the bissilylhydronium ions with a naphthalene-1,8-diyl platform reported by Müller [δ( 29 Si) 54.4 ppm for 7 + ], [15a] clearly indicating the development of silylium ion character. The broad 1 H NMR signal at δ( 1 H) 2.65 ppm of the bridging hydrogen atom in 10[B(C 6 F 5 ) 4 ] is remarkably shifted to high field compared to the SiÀ H resonance value of 4.82 ppm in 9. An integration to two protons corroborates that the two boronbound hydrides in 10 + are equivalent due to fast hydrogen exchange process. [6b] This is consistent with the computed very low free energy barrier of only 8 kJ mol À 1 for this process in solution (at standard conditions; Scheme S1). Due to the line width of the signal, the J SiÀ HÀ B was not detected in the 1 H NMR spectrum in 1,2-Cl 2 C 6 D 4 at 298 K. The VT NMR showed that the width of signal narrows with decreasing temperature. Thus, the average coupling constant of 1 J SiÀ HÀ B(H) and 3 J SiÀ HÀ B(H) = 28 Hz was determined by a 1 H/ 29 Si-1D-CLIP-HSQMBC NMR experiment in ClC 6 D 5 at 240 K, which is significantly reduced compared to the 1 J Si,H = 182 Hz for 9. The broad signal in 11 B NMR spectrum shows a lowfield shift to À 8.

Scheme 2. Generation and key 1 H and 29
Si NMR resonance signals of the hydroborane-stabilized silylium ion 10 + with different counteranions. All NMR data were recorded in 1,2-Cl 2 C 6 D 4 .
Both silylium ions and NHC-stabilized borenium ions [26] can be generated by hydride abstraction with trityl salts from hydrosilanes and -boranes, respectively. To probe whether 10 + is a hydroborane-stabilized silylium ion or a borenium-ionactivated hydrosilane, quantum chemical calculations using DFT methods were performed (see the Supporting Information for the computational details).
The calculated 29 Si and 11 B chemical shifts for the DFToptimized structure of 10 + in 1,2-Cl 2 C 6 H 4 (using a continuum solvent model) are in excellent agreement with the experimental values (see Table S4): δ( 29 Si) 55.3 ppm and δ( 11 B) À 8.5 ppm (computed) versus δ( 29 Si) 56.0 ppm and δ( 11 B) À 8.2 ppm for 10[B(C 6 F 5 ) 4 ] (experimental). This indicates a correct description of electronic structure details at the chosen computational levels. Of note, the 11 B chemical shift is particularly sensitive to geometrical distortions in the present case. This provided a good basis for closer analyses of bonding. Natural bond orbital (NBO) and natural resonance theory (NRT) analyses confirm the identification of the BÀ HÀ Si moiety in 10 + as a delocalized 3c2e bond, and the obtained natural bond orders (BOs) are consistent with asymmetrical multicenter σ-bonding (Figure 3, left and Table S5). Specifically, both a higher total bond order (0.47 versus 0.37) as well as a larger covalent character (0.29 versus 0.19) for the BÀ H1 bond compared to the SiÀ H1 bond are found. We also note that the computed Wiberg bond indices show a similar picture (Table S5)

Chemistry-A European Journal
Research Article doi.org/10.1002/chem.202104464 Based on their results, Wang and co-workers classified 3 + as borenium-ion-activated hydrosilane. A closer analysis of 10 + by means of two NBO Lewis structures (LS) featuring either an explicit BÀ H1 (LS BH1 ) or SiÀ H1 (LS SiH1 ) σ-bond with otherwise identical bonding setups (Figure 4) reveals that LS BH1 provides a moderately but notably better fit (e. g. a smaller residual non-Lewis density) of the total density matrix than LS SiH1 (3.989e versus 4.125e; Table S7). Compared to the occupancies of the corresponding NBOs in the precursor 9, substantial charge delocalization from the Si/BÀ H1 σ-bond takes place in both cases but significantly more so in LS SiH1 (0.47e) than in LS BH1 (0.31e). As expected, the predominant acceptor is the (formally) vacant p-type atomic orbital on the opposite center in each case, which is consequently populated significantly (LS SiH1 : 0.53e; LS BH1 : 0.38e). Back-donation of charge density into the Si/ BÀ H1 σ*-antibonding orbital is negligible in both cases, which was also observed by Wang and co-workers for 3 + .
We also estimated the relative Lewis acidity of the silicon and boron centers in 10 + by computing their fluoride-ion affinities (FIAs) using F 2 CO as a standard for the appropriate isodesmic reactions (Scheme S2). The results clearly indicate a larger electrophilicity of the silicon atom (644 kJ mol À 1 ) compared to the boron atom (609 kJ mol À 1 ). Together with essentially all other bonding indicators (see above), this also is consistent with the picture of an open 3c2e SiÀ HÀ B bond that tends to be somewhat closer to a hydroborane-stabilized silylium ion than to a hydrosilane-stabilized borenium ion.
In conclusion, we presented herein the synthesis of naphthalene-1,8-diyl-based Si/B hydronium ion 10 + paired with [B(C 6 F 5 ) 4 ] À and [HCB 11 Cl 11 ] À by hydride abstraction from neutral precursor 9. Ion pairs 10 + were fully characterized by NMR spectroscopy and X-ray diffraction. X-ray crystallography analysis and DFT calculations provide strong evidence for a delocalized 3c2e BÀ HÀ Si bond with more pronounced silyliumion than borenium-ion character. The high activation degree of the SiÀ H bond in 10 + and the structure of 10[B(C 6 F 5 ) 4 ] can be viewed as a snapshot of the "competition for hydride" between two different main-group element Lewis acid centers, an important feature in Piers-type chemistry. With an appropriate tether, it may even be possible to synthesize a silylium/ borenium dication. [27]