Assessing the Limits of the "Lego-Brick" Approach: Equilibrium Structures of Strained and Flexible Cyclic Molecules

An accurate description of molecular structures is essential in several fields of chemistry and, in particular, in high-resolution molecular spectroscopy. The so-called "Lego-brick" approach has proven to provide near-spectroscopic accuracy at a fraction of the computational cost of high-level composite schemes, but its applicability has so far been mainly assessed for rather rigid systems. In this work, we systematically investigate the performance of the "Lego-brick" approach for strained and conformationally flexible cyclic molecules. A chemically diverse benchmark set of three-, four-, and five-membered rings, including heterocycles and species with multiple conformers, is considered. By comparing template-molecule (TM) and full "Lego-brick" (TM+LR) rotational constants with the experimental counterparts, the accuracy of corresponding equilibrium structures is analyzed. The results show that the "Lego-brick" approach retains good accuracy for small cyclic systems, although the data set turned out to be a challenging test case. Linear-regression (LR) corrections are found to be fundamental to achieve the aimed precision. Interestingly, the TM+LR geometries are so accurate that can be employed in the framework of the semiexperimental approach, thus allowing one to obtain equilibrium structures of experimental quality also when there is a lack of isotopic data. Overall, this study delineates the applicability limits of the "Lego-brick" approach for flexible systems, pointing out the ability of significantly improving the initial density functional theory results.