Platinum‐Templated Coupling of B=N Units: Synthesis of BNBN Analogues of 1,3‐Dienes and a Butatriene

Abstract The 1:2 reaction of [μ‐(dmpm)Pt(nbe)]2 (dmpm=bis(dimethylphosphino)methane, nbe=norbornene) with Cl2BNR(SiMe3) (R=tBu, SiMe3) yields unsymmetrical (N‐aminoboryl)aminoboryl PtI 2 complexes by B−N coupling via ClSiMe3 elimination. A subsequent intramolecular ClSiMe3 elimination from the tBu‐derivative leads to cyclization of the BNBN unit, forming a unique 1,3,2,4‐diazadiboretidin‐2‐yl ligand. In contrast, the analogous reaction with Br2BN(SiMe3)2 leads, via a twofold BrSiMe3 elimination, to a PtII 2 A‐frame complex bridged by a linear BNBN isostere of butatriene. Structural and computational data confirm π electron delocalization over the entire BNBN unit.

The replacement of C = C double bonds in organic molecules by isosteric covalent B = N units is not only interesting from a fundamental point of view, but also opens up the exploration of a vast hybrid organic-inorganic chemical space. While the typical B=N double bond (1.39 ) [1] is only marginally longer than a C = C double bond (1.34 , Figure 1), the intrinsic strong polarization of B À N bonds imparts very different electronic properties and stability to the resulting molecules and materials, which can be exploited for new applications in materials science, catalysis, and medicinal chemistry.
Complex 3 SiMe3 could not be fully characterized as it decomposed rapidly in solution into ClSiMe 3 and a number of dmpm-containing platinum complexes, the known complex [m-(dmpm)PtCl] 2 (5-Cl: d( 31 P) = À19.3 ppm, 1 J PÀPt = 2650 Hz) [13a] being the major decomposition product (Scheme 3 b, see Figure S18 in the SI). The fate of the remaining [BNSiMe 3 ] 2 fragment could not be determined as the 11 B NMR spectrum of the final product mixture was silent, and a colorless by-product, insoluble in all common organic solvents, was formed. [15] In contrast, 3 tBu was stable in solution at room temperature but selectively converted to 4 tBu at 80 8C by intramolecular cyclization of the BNBN moiety under  Crystallographically derived molecular structures of (from left to right) 3 tBu (least disordered one of the two molecules of 3 tBu in the asymmetric unit), 4 tBu , and 6. [26] Thermal ellipsoids at 50 % probability. Thermal ellipsoids of ligand periphery and hydrogen atoms omitted for clarity. Only the major part of the disorders in 3 tBu (terminal B(Cl)NtBu(SiMe 3 ) moiety) and 4 tBu (entire (BNtBu) 2 Cl moiety and one dmpm ligand) is shown. Due to the restraints applied to these disorders during refinement, the structural parameters of 3 tBu and 4 tBu may not be fully discussed. ClSiMe 3 elimination (Scheme 3 c). This reaction is analogous to the cyclization of RClB À N(tBu) À B(Cl) À NtBu(SiMe 3 ) (R = NMe 2 , NEt 2 , Et, iBu) to 1,3,2,4-diazadiboretidines by ClSiMe 3 elimination, reported by Paetzold in 1988. [16] The 11 B NMR spectrum of 4 tBu is nearly identical to that of 3 tBu , displaying two broad resonances at 54 (fwmh % 1480 Hz, PtB) and 32 ppm (fwmh % 470 Hz, N 2 BCl). The conversion of 3 tBu to 4 tBu is evidenced more clearly by changes in the 31 P{ 1 H} spectrum, which shows two new 1:1 multiplets with higherorder satellites, both shifted ca. 2 ppm downfield from 3 tBu , at À12.8 ( 1 J PÀPt = 3198 Hz, P 2 PtCl) and À27.6 ppm ( 1 J PÀPt = 2632 Hz, P 2 PtB), the 1 J PÀPt coupling constant of the latter being ca. 100 Hz smaller than in 3 tBu . Crystallization attempts of 4 tBu always yielded pseudo-merohedrally twinned crystals (see solid-state structure in Figure 2), in which the BNBN heterocycle presents a twofold disorder by rotation of about the Pt2 À B1 bond, thus precluding any discussion of bond lengths and angles in this unit. Despite the well-established chemistry of 1,3,2,4-diazadiboretidines as h 4 -ligands for transition metals, [17] 3 tBu represents a hitherto unknown binding mode of this type of ligand as an anionic h 1 -ligand via coordination at boron. In solution at room temperature, compound 4 tBu decomposed very slowly but selectively over a period of several weeks to complex 5-Cl and an unidentified intractable colorless solid, by formal loss of "[BN(tBu)] 2 " (Scheme 3 d). [15] To our surprise the reaction of 1 with Br 2 BN(SiMe 3 ) 2 resulted instead in the formation of the A-frame complex 6, isolated as a yellow solid in 46 % yield (Scheme 4). [18] The 11 B NMR spectrum of 6 displays two broad resonances at ca. 57 (fwmh % 1510 Hz) and 26 ppm (fwmh % 690 Hz), the former being attributed to the platinum-bound boron nucleus by analogy with the 11 B NMR shift of the related dimethylaminoboranediyl-bridged A-frame complex 2-Br NMe2 (d( 11 B) = 52 ppm), [13] the latter to the dicoordinate NBN boron nucleus. The 31 P{ 1 H} NMR spectrum showed a singlet at À7.1 ppm, close to that of 2-Br NMe2 (d( 31 P) = À5.6 ppm), with a higher-order satellite splitting pattern typical for Aframe complexes ( 1 J PÀPt = 3568 Hz, 3 J PÀPt = 272 Hz, 1 J PtÀPt = 1826 Hz). 11 B and 31 P{ 1 H} NMR-spectroscopic monitoring of the reaction showed no sign of formation of the bromide analogue of 3 SiMe3 .
We propose that the formation of complexes 3 R and 6 proceeds via a same intermediate h 1 -(silylamino)haloboryl complex Int-X R formed by the oxidative addition of X 2 BNR-(SiMe 3 ) to 1 (Scheme 5). [19] This step can be followed either by B À N coupling with a second equivalent X 2 BNR(SiMe 3 ) via XSiMe 3 elimination (reaction rate constant k a ) to form an h 1 -(N-aminoboryl)aminoboryl complex analogous to 3 R , or by the oxidative addition of the second B À X bond of the silylamino(halo)boryl ligand to platinum to form the (silylamino)boranediyl A-frame complex 2-X NR(SiMe3) (reaction rate constant k b ). For R = SiMe 3 , the latter then undergoes twofold XSiMe 3 elimination with a second equivalent of X 2 BN(SiMe 3 ) 2 to form complex 6. The selectivity of the reaction is therefore determined by the relative values of the reaction rate constants k a and k b : for X = Cl the rate of B À N coupling outperforms that of oxidative addition of B À Cl to Pt, leading to the exclusive formation of 3 R , the opposite being the case for X = Br, leading to the exclusive formation of 6.
The electronic structure of 6 was further investigated using DFT and intrinsic bond orbital (IBO) [20] calculations. The BNBN motif in the optimized structure of 6, obtained at the M06 [21] -D3 [22] /cc-pVDZ [23] ,aug-cc-pVDZ-PP{Pt} [24] level of theory, shows a larger deviation from linearity (B1-N1-B2 161.38, N1-B2-N2 176.28) than that of the solid-state structure. Similar results were obtained with other density functionals (see details in the SI). In order to investigate the origin of this deviation, we performed computations on four truncated model systems, in which the PMe 2 and SiMe 3 groups were successively replaced with PH 2 and SiH 3 or H, respectively (see Figure S19 in the SI). In all of these cases, the BNBN moiety was found to be linear (B1-N1-B2 and N1-B2-N2 178.8-180.08). The distortion from linearity therefore seems to arise from the steric repulsion between the PMe 2 and SiMe 3 substituents, although the additional influence of crystal packing forces in the solid-state structure cannot be discounted. Furthermore, the calculated Mayer bond orders (MBOs) [25] of the BNBN motif in 4 (B1ÀN1: 1.38, N1ÀB2: 2.11, B2ÀN2: 1.32) are very similar to those obtained for the parent H 2 BNBNH 2 system (B1 À N1: 1.51, N1 À B2: 2.13, B2 À N2: 1.43), these values suggesting strong cumulenic character in both cases. Indeed, inspection of the IBOs of 6 (Figure 3 a) reveals that IBO-1 and IBO-3, which are orthogonal to the (Pt1-B1-Pt2) plane, are partially delocalized to the neighboring B2 and B1 atoms, evidencing deviation from the 1-boryl-2-(amino)iminoborane picture. This view is also supported by inspection of the canonical Kohn-Sham molecular orbitals (MOs) of 6 and H 2 BNBNH 2 (Figure 3 b and S20 in the SI), where p electron delocalization over the entire BNBN unit is observed. The description of 6 as a BNBN analogue of butatriene is, therefore, fully supported by quantum chemical investigations.
To conclude, we have shown that the [m-(dmpm)Pt] 2 framework acts as an effective template for the coupling of B = N units obtained by the intermolecular B À N coupling of dihalo(silylamino)boranes via halosilane elimination. For Cl 2 BNR(SiMe 3 ) precursors BN chain growth occurs at a side-on Pt I 2 complex, whereas for Br 2 BN(SiMe 3 ) 2 an Aframe Pt II 2 complex bridged by a linear BNBN unit is formed. Structural and computational analyses confirm a cumulenic motif isosteric with butatriene.