Indirect carbon–carbon couplings across one, two and three bonds in substituted benzenes: Experiment and theory

Abstract

An excellent linear correlation is found between a large body of experimental spin–spin carbon–carbon couplings, J(CC), and B3PW91/6-311++G(d,p)//B3PW91/6-311++G(d,p) calculated estimates. The correlation does not differ significantly from the simplest relationship possible, J(CC) exp. = J(CC) calcd., within a small and random spread of about 1 Hz. There are 257 experimental values considered, and 126 out of these are new and come from the present work. The carbon–carbon couplings which include those across one, two and three bonds, span a considerable range of about −3 through +83 Hz. This is the first report of such a fine correspondence between experiment and theory in the field of spin–spin couplings where a large set of experimental data and a broad range of couplings is confronted with the relevant density functional theory (DFT) computations where the latter are preceded by molecular geometry optimizations using the same combination of the density functional and the basis set of wavefunctions. The quantum mechanical computations on the level of sophistication employed can reliably predict the signs of aromatic couplings J(CC) and this turned out to be important for 2 J(CC)′s which, according to the present work, span a range of about −3 through about +7 Hz. It is demonstrated for the first time that aromatic 2 J(CC)′s can bear either sign, and that substituent effects on the latter are significant. As can be inferred from the present report, substituent effects on aromatic spin–spin couplings between carbon–carbon nuclei result in the following ranges of the coupling constants: 1 J(CC)′s, +48.0 through +82.8 Hz; 2 J(CC)′s, −2.9 through +6.6 Hz; 3 J(CC)′s, +4.7 through +11.3 Hz. These can be compared with the literature experimental data for benzene: 1 J(CC) = (+)56.0 Hz; 2 J(CC) = (−)2.5 Hz; 3 J(CC) = (+)10.1 Hz. It is fairly obvious from the present work that rovibronic effects on aromatic J(CC)′s and those of nuclear motions at zero K are negligible.