VBData: A Valence Bond Property Dataset of Single Bonds for Organic Compounds

Abstract

The properties of valence bonds are fundamental to understand and predict the reactivity of chemical reactions. This article presents a comprehensive dataset derived from quantum chemical calculations performed at a consistent computational level for more than 40,000 organic compounds (containing only C, H, N, and O atoms) and over 400,000 non-cyclic single bonds within them. Key bond properties include resonance energies and structural weights. The valence bond computation was performed using XMVB program at the VBSCF level, based on the equilibrium geometries optimized at the B3LYP/def2-SVP level. This dataset is anticipated to serve as a new standard benchmark for bond properties in organic chemistry and to provide a foundation for developing machine learning models.