Replying to H. Rzepa. Nature Communications 10.1038/s41467-021-21433-8 (2021)

We are writing in response to Rzepa’s theoretical analysis of the generation of C₂, which cites our recent paper describing the first chemical synthesis of C₂ (ref. ¹). Rzepa suggests on the basis of several in silico approaches that the formation of free C₂ from alkynyl-λ³-iodane and fluoride ion would be prohibitively endo-energetic. He proposes three possible explanations of the apparent discrepancy between these theoretical calculations and our experimental findings, of which one is that some species other than C₂ is actually formed.

As our original paper was primarily experimental, describing room-temperature chemical synthesis of C₂ and the first bottom-up chemical synthesis of nanocarbons from C₂, we should like to respond to the latter point. We believe that the evidence presented in our paper for the generation of C₂ itself is compelling¹. In particular, the connected-flask, solvent-free experiment clearly supports the generation of free C₂ gas, for the following reasons:

When ¹³C-labeled 1b-¹³C_β was used, a 1:1 mixture of O- ethynyl galvinoxyl 15-¹³C_α and 15-¹³C_β was obtained (Fig. 1a).





Fig. 1

Connected-flask, solvent-free experiment.

a ¹³C-labeling experiment. b APCI mass spectrum (positive ion mode) obtained from the contents in Flask B.

APCI mass spectrum of the contents in Flask B included the peak assignable to acetylene digalvinoxyl ether 16 (Fig. 1b).

It is more difficult to establish conclusively whether free C₂ is generated in solution, but the following experimental facts are relevant:

3.The relative rate of hydrogen abstraction between CH₂Cl₂ and 9,10-dihydroanthracene (12) is calculated to be ca. 1:20 (per 1H).

4.Galvinoxyl radical 14 solely produced O-ethynyl galvinoxyl 15 (not 16).

5.The amount of the mixture of 15-¹³C_α and 15-¹³C_β obtained by the reaction of 1b-¹³C_β with 14 differed in solvents of different viscosities (η).

These results indicated that the radical character of the intermediate is much milder than that of the common unstabilized alkynyl radical, which is consistent with the presence of an interaction between the radicals in C₂, in other words, a singlet biradical (charge-shift bonding) character, as suggested in a recent theoretical study².

At present, we cannot explain the discrepancy between Rzepa’s theoretical evaluation and our experimental results. However, as discussed in our original paper, we believe that all our experimental observations can only be rationalized in terms of the generation of gaseous C₂. Our findings have attracted great interest, and we anticipate that independent experimental findings will emerge in the near future. We ourselves are working on the direct observation of chemically generated C₂ by means of Raman spectroscopy, ESR spectroscopy, and other methods, and we hope that this work will provide definitive experimental evidence for the bond length and electronic state of C₂ (ref. ³ and references therein).