Molecular similarity based on atomic electrostatic potential

Area of Science:

• Computational Chemistry
• Quantum Chemistry
• Molecular Modeling

Background:

• Understanding molecular similarity is crucial for predicting chemical properties and reactivity.
• Existing methods may not fully capture nuanced electronic differences between molecules.

Purpose of the Study:

• To introduce a new similarity measure based on atomic potential values in substituted benzene derivatives.
• To evaluate the effectiveness of this measure against reference chemical reactions.

Main Methods:

• Calculating potential values at specific atoms (benzene ring, COOH group) in para-substituted benzoic acids and monosubstituted benzenes.
• Employing Euclidean distance in the potential space to quantify molecular similarity.
• Using benzoic acid dissociation and benzene nitration as reference reactions.

Main Results:

• The developed similarity measure is equivalent to Euclidean distance in the potential space.
• Benzene ring potentials showed high similarity across different molecular contexts (para-substituted acids vs. monosubstituted benzenes).
• Atomic potentials within molecules were found to be highly correlated.

Conclusions:

• The proposed atomic potential-based similarity measure offers a robust way to compare molecules.
• The high correlation of atomic potentials suggests underlying electronic similarities in these benzene derivatives.
• This approach provides insights into structure-property relationships in organic chemistry.