AIMLDM: A program to generate and analyze electron localization–delocalization matrices (LDMs)

• Published in: Computational and theoretical chemistry 2015-10, Vol.1070, p.55-67
• Main Authors: Sumar, Ismat, Cook, Ronald, Ayers, Paul W., Matta, Chérif F.
• Format: Article
• Language: English
• Subjects: AIMAll; Multidimensional scaling (MDS); Principal component analysis (PCA); Quantitative structure–activity relationships (QSAR); Quantitative structure–property relationships (QSPR); Quantum Theory of Atoms in Molecules (QTAIM)
• ISSN: 2210-271X
• DOI: 10.1016/j.comptc.2015.07.014

[Display omitted] •A merging of chemical graph theory and the Quantum Theory of Atoms in Molecules is proposed.•A programme to construct and use “localization–delocalization matrices” (LDMs) is presented.•An LDM contains a wealth of information about the electronic structure and distribution. A merging of Chemical Graph Theory (CGT) with the Quantum Theory of Atoms in Molecules (QTAIM) has recently been proposed through the usage of the “localization–delocalization matrices” (LDMs). An LDM of a molecule composed of n atoms lists the complete set of localization indices as its diagonal elements while the delocalization indices divided by two are entered as the off diagonal elements. In this manner the matrix summarizes the electron distribution of every atom in the molecule in terms of an atomic (self) and (n−1) diatomic (interaction) terms that sum to yield its electron population. LDMs are rich in information and have several interesting properties but their direct usage as a molecular fingerprinting tool can also have limitations, limitations resolved by following the lead of CGT. A Python 3.4.1 programme is presented (AIMLDM) that automates the extraction, manipulation, and preliminary analysis of these LDMs given the QTAIM integration and analysis output of the software AIMAll. The programme is shown to be stable with some numerical test cases. The LDM condenses information about the molecular electron distribution at the atomic and diatomic level and shows promise as a tool for measuring molecular similarity distances and in assessing the quality of basis sets. The main anticipated use of the LDM-analysis is in the construction of predictive robust quantitative structure–activity relationships (QSAR) models especially when used in conjunction with principal component analysis (PCA) and multidimensional scaling (MDS).