Correlations between CH bond stretching frequencies, bond lengths and dissociation energies, previously studied by Bernstein, are reexamined with fresh data of all kinds. The introduction of vCH values from CHD2 substitution is helpful in all cases of CH3 compounds, and vital where the latter contain non-equivalent CH bonds. Predictions of rCH0 are made for a wide range of CH3 and CH2 compounds with a likely error of less than 0·002 Å. The former include CH3NO2, CH3F, PhCH3, (CH3)4M, where MC, Sn, Ge, Sn, Pb, containing symmetrical CH3 groups, and for CH3NH2, (CH3)2NH, (CH3)2N SiF3, CH3OH, CH3OPCl2, HCOOCH3, CH3COOCH3, (CH3)2CH2, (CH3)2CO and (CH3)3CCl where non-equivalent bonds are present. Differences in bond length as small as 0·0005 Å can in principle be distinguished. The existing rCH0 value for CH3CN is seriously wrong. The correlation between vCH and JCH is considered briefly. The correlation between vCH and D0298 is good for most compounds which are likely to dissociate into σ radicals where the stabilization energy is small. In CH3 compounds containing non-equivalent bonds. D0 correlates excellently with the frequency for the weaker bond, as expected. “Radical stabilization energy” is conveniently measured using the correlation graph and vCH value, together with the observed D0. unusually large experimental errors appear to be present in D0 data for c-C3H6, c-C4H8, C2H2 and HCOOH.
Read full abstract