Unlike the well know and oft studied chemistry of double bonds between carbons, the chemistry of boron-boron double bonds is for the most part unexplored. It is believed that boron should behave similarly to carbon due to its relativity to the element on the periodic table. Anions containing boron double bonds, specifically [R2BBR2]2-, have in the past been predicted to be possible structures of interest to synthesize in the laboratory, however such efforts have failed for the most part.
It was then proposed to explore neutral diborenes, even though they in theory should be highly reactive compounds due to their triplet ground states and two one-electron π-bonds according to molecular orbital theory. The electron deficiency in this structure could however be stabilized by the addition of Lewis base ligands. The stabilizing ability of different ligand groups were assessed, including CO and NHC, which were chosen based on their strong electron donating capabilities. The ligand group that ultimately experimentally produced an actual neutral diborene was :C{N(2,6-PRi2C6H3)CH}2. Previous work from using this ligand group for stabilizing carbenes suggested that this would be a potential stabilizing ligand for a diborene.
This compound, R(H)B=B(H)R, where R is the aforementioned ligand group, was synthesized beginning with RBBr3 and KC8 in diethyl ether. Two products were isolated from this reaction, including the desired diborene R(H)B=B(H)R. It was shown that a ratio of 1:5.4 of RBBr3 to KC8 yielded the highest percentage of R(H)B=B(H)R (12%). Any excess amount of RBBr3 over this ratio resulted in a decrease of R(H)B=B(H)R and thus in increase of the other product, R(H)2B-B(H)2R.
A few methods were utilized in order to determine the chemical makeup of these products. NMR resonances of RBH3, R(H)2B-B(H)2R, and R(H)B=B(H)R were respectively reported to be -35.38, -31.62, and +25.30 ppm. The 11B signal of R(H)B=B(H)R produced a quartet, while the other two compounds elicited singlets. This alone could suggest double bond character between borons.
X-ray structural analysis shows a bond distance of 1.828 Å for R(H)2B-B(H)2R. This number seems to be on point with calculated B-B bond lengths for similar structures such as the CO-ligated analogue (1.819 Å) and an activated m-terphenyl based diborate (1.83 Å). Crystallization of R(H)B=B(H)R reveals B-C bond distances of 1.547 Å, which is marginally shorter than that of the other molecules. In addition to this, it is calculated that the angles between the C3N2 carbene rings and the core are strikingly different than that of the other compounds used and produced. Finally, the B=B bond distance in R(H)B=B(H)R was measured to be much shorter than the B-B distance reported in R(H)2B-B(H)2R, again implying a double bond.
DFT computations were also used to support the nature of R(H)B=B(H)R. The analysis was performed on the simplified model, where R=:C(NHCH)2. The experimental bond lengths for the non-simplified model seem to be in concordance with the computed B-B and B-C bond lengths, and well as the B-B-C bond angle calculated from the simplified model analyzed using DFT. The bond character of these bonds was also delved into via HOMO representations of the compounds among other computational techniques.
In conclusion, the authors of the paper were able to successfully prove that they had synthesized and characterized the first neutral diborene compound. They also ventured into the nature of the elusive boron-boron double bond. Though it was not necessarily expected that this phenomenon could feasibly be synthesized due to the expected reactivity of the boron-boron double bond, these chemists found a way to isolate the compound. In context to the larger field of chemistry, I suppose that the authors could determine other possible ligand groups that would produce a stable neutral diborene. They could also venture into increasing the percent yield, as 12% is on the low side. Finally, they could explore other group 13 elements, such as Al and Ga to see if they can replicate similar double bond properties.
