The Nature of Chemical Bonds in the Age of Artificial Intelligence: Revisiting Electronegativity of Organic Molecules

Electronegativity can be considered a data-driven concept that is widely used since Pauling proposed this property. However, updating the electronegativity based on the vast amount of high quality experimental and computational data has been overlooked. Thus, advances in artificial intelligence (AI), with its ability to manage large datasets and identify underlying patterns, necessitate reconsid ering data-driven concepts such as electronegativity. In this work, we present a data-driven method to generate more informative multidimensional electronegativity of organic molecules using graph neural networks. Focusing on two-dimensional electronegativity, we were able to do a more detailed classification of the atoms and their covalent bonds. By replacing the conventional electronegativity with the newly proposed one, we demonstrated the performance improvement in molecular machine learning tasks. We believe that our findings will be useful in understanding of electronegativity and chemical bonds by judiciously applying AI-driven methods to chemical studies.