Switching Perspectives on Hirshfeld Partitioning Schemes

We explore the foundations and applications of Hirshfeld partitioning, a fundamental technique used in the analysis of atoms in molecules, and the related Hirshfeld surface used to study molecules in crystals. Hirshfeld partitioning has direct connections to information theory, machine learning, and statistical mechanics, and this work aims to deepen our understanding of why these techniques are successful and so generally applicable, highlighting the connections of such methods to our physical understanding of the systems to which they are applied. We further investigate practical considerations in the implementation and interpretation of Hirshfeld surfaces and Voronoi partitioning.