CHAOS - A Large-scale Database for σ-Profiles and Other Molecular Descriptors

Area of Science:

• Computational Chemistry
• Materials Science
• Chemical Engineering

Background:

• Sigma-profiles are crucial molecular descriptors for applications like solvent selection and thermodynamic modeling.
• Existing sigma-profile libraries are fragmented and inconsistent due to varied computational methods, limiting their utility.
• A need exists for a large-scale, internally consistent database of sigma-profiles and related quantum-chemical data.

Purpose of the Study:

• To introduce CHAOS (Computed High-Accuracy Observables and Sigma Profiles), a novel, extensive, and consistent database of molecular sigma-profiles.
• To provide a comprehensive set of quantum-chemical observables alongside sigma-profiles for a vast number of molecules.
• To facilitate advancements in data-driven molecular design and thermodynamic modeling across scientific disciplines.

Main Methods:

• Generated sigma-profiles and other quantum-chemical observables for 53,091 molecules using a standardized workflow.
• Employed the ωB97X-D/def2-TZVP level of theory for all quantum-chemical calculations.
• Included gas-phase geometries, C-PCM data, IR spectra, thermodynamic properties, and NMR shielding tensors.

Main Results:

• The CHAOS database contains 53,091 molecules with computed sigma-profiles and diverse quantum-chemical data.
• Data generation utilized a consistent, high-accuracy quantum-chemical workflow.
• The database covers a wide range of molecular properties (mass up to 400 amu, dipole moments up to 15 D) and elements.

Conclusions:

• CHAOS significantly expands the availability of public sigma-profile data by over an order of magnitude.
• The database provides a unified and consistent foundation for developing physics-based and machine-learning models.
• CHAOS empowers research in chemistry, chemical engineering, and materials science by offering a rich quantum-chemical data basis.