Combinação de Cálculos DFT e algoritmos de Machine Learning para o estudo da reatividade de complexos quadráticos plano de Platina (II)

Abstract

Ligand exchange reactions in quadratic-plane Pt(II) complex represent one of the most well-established mechanisms in inorganic chemistry, where characteristics such as stereochemistry of the products generated as well as the activity of these compounds can be rationalized based n the concepts of trans influence and trans effect. The trans effect is an effect of essentially kinetic nature, in which the effect that a given L-type ligand coordinated to the metal has on the rate of substitution of the ligand in the trans position to it is measured. Much of the knowledge about these effects is based mainly on a large set of empirical measurements obtained over the years that gave rise to constructions of some chemical series of ligands ordered according to their trans effects and influences. However, from the atomic/molecular scale point o view, there are still open discussions about the real interpretation of these effects, that is, about how the electronic structures of these systems is related to the properties observed experimentally on a macroscopic scale. In this work, after the application of DFT calculations together with calculations of local reactivity descriptors based on Fukui functions and the model of chemical bond overlay region analysis, it was found that the analysis of the electronic repulsion integral of coating J in the Pt-T binding region in the trans-Pt[(NH3)2T(H2O)]n+ (T=H2O, NH3, F-, Cl-, CN-, CO, CH3-), may present another interpretation of the trans effect and trans influence from an electronic parameter. Additionally, it was found that the application of supervised machine learning algorithm for the construction of predictive models, such as the one developed through a linear regression model, brought approximations in very good agreement of the Gibbs free energy of activation as a function of the structural parameter, as the Metal-Binding bond distances.