Computational efficiency meets spectroscopic accuracy: an unsupervised workflow for equilibrium geometries and...

Equilibrium molecular geometries are essential for understanding molecular systems, particularly in the gas phase, where intrinsic stereoelectronic effects can be disentangled from environmental influences. High-resolution rotational spectroscopy offers direct structural information and is now applicable to molecules with up to 50 atoms, significantly expanding its scope and demanding more advanced computational support to account for vibrational averaging and spectral complexity. Herein, we present an automated workflow that integrates the Pisa composite schemes (PCS) with efficient vibrational correction models, interfacing seamlessly with Gaussian and MSR programs. The protocol is designed for medium-sized molecules where relativistic and static correlation effects can be neglected, and is demonstrated on a set of prebiotic and biologically relevant compounds. Reliable equilibrium geometries are obtained for both semi-rigid and flexible species, provided that a second-order vibrational perturbation theory (VPT2) treatment is adequate; proline is included as a representative flexible case. Additionally, the phenyl radical is considered as a prototypical open-shell system, supported by extensive isotopic experimental data. This strategy enables the accurate and cost-effective determination of equilibrium geometries for molecules beyond the small-molecule regime, outperforming conventional methods and offering broad applicability in astrochemistry, prebiotic chemistry, and molecular spectroscopy.