Understanding substituent effects in noncovalent interactions involving aromatic rings

Noncovalent interactions involving aromatic rings such as π-stacking, cation/π, and anion/π interactions are central to many areas of modern chemistry. Decades of experimental studies have provided key insights into the impact of substituents on these interactions, leading to the development of simple intuitive models. However, gas-phase computational studies have raised some doubts about the physical underpinnings of these widespread models. In this Account we review our recent efforts to unravel the origin of substituent effects in π-stacking and ion/π interactions through computational studies of model noncovalent dimers. First, however, we dispel the notion that so-called aromatic interactions depend on the aromaticity of the interacting rings by studying model π-stacked dimers in which the aromaticity of one of the monomers can be "switched off". Somewhat surprisingly, the results show that not only is aromaticity unnecessary for π-stacking interactions, but it actually hinders these interactions to some extent. Consequently, when thinking about π-stacking interactions, researchers should consider broader classes of planar molecules, not just aromatic systems. Conventional models maintain that substituent effects in π-stacking interactions result from changes in the aryl π-system. This view suggests that π-stacking interactions are maximized when one ring is substituted with electron-withdrawing groups and the other with electron donors. In contrast to these prevailing models, we have shown that substituent effects in π-stacking interactions can be described in terms of direct, local interactions between the substituents and the nearby vertex of the other arene. As a result, in polysubstituted π-stacked dimers the substituents operate independently unless they are in each other's local environment. This means that in π-stacked dimers in which one arene is substituted with electron donors and the other with electron acceptors the interactions will be enhanced only to the extent provided by each substituent on its own, unless the substituents on opposing rings are in close proximity. Overall, this local, direct interaction model predicts that substituent effects in π-stacking interactions will be additive and transferable and will also depend on the relative position of substituents on opposing rings. For cation/π and anion/π interactions, similar π-resonance-based models pervade the literature. Again, computational results indicate that substituent effects in model ion/π complexes can be described primarily in terms of direct interactions between the ion and the substituent. Changes in the aryl π-system do not significantly affect these interactions. We also present a simple electrostatic model that further demonstrates this effect and suggests that the dominant interaction for simple substituents is the interaction of the charged ion with the local dipole associated with the substituents. Finally, we discuss substituent effects in electrostatic potentials (ESPs), which are widely used in discussions of noncovalent interactions. In the past, widespread misconceptions have confused the relationship between changes in ESPs and local changes in the electron density. We have shown that computed ESP plots of diverse substituted arenes can be reproduced without altering the aryl π-density. This is because substituent-induced changes in the ESP above the center of aryl rings result primarily from through-space effects of substituents rather than through changes in the distribution of the π-electron density.