Potenciais de redução em meio aquoso: aplicação e desenvolvimento de protocolos de calculo baseados em modelos implícitos e métodos híbridos.

• Ano de defesa: 2020
• Autor(a) principal: Silva, Daniel Gabriel da
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal da Paraíba; Brasil; Química; Programa de Pós-Graduação em Química; UFPB
• Palavras-chave em Português: Benchmarking; Exatidão química; Método híbrido; Chemical accuracy; Hybrid method; CNPQ::CIENCIAS EXATAS E DA TERRA::QUIMICA
• Link de acesso: https://repositorio.ufpb.br/jspui/handle/123456789/19724

The implicit solvent models were parameterized to predict solvation free energies of neutral and ionic solutes, which can be combined with experimental energies or ab initio of the respective gas-phase systems to estimate the energy variations in the solution. Such procedures have been widely applied in the calculation of several properties, such as the reduction potentials and binding energies in various solvents. The main objective of the present work is to describe the solvation process in water, using implicit models and hybrid method, having as an evaluative character the calculation of the reduction potential of 42 redox reactions (14 amines, 4 nitroxides, and 24 phenols). In the present study, the methodological approaches were investigated: (i) systematic scanning and simulated annealing to search for structures with minimum energy; (ii) use of the direct reduction potential calculation protocol; (iii) use of the electronic structure methods G3(MP2)-RAD(+) and M06-2X (with the basis sets 6-31+G(d) and cc-pVDZ, with the addition of the remaining parameters of the composite method); and (iv) use of the C-PCM and SMD solvation models with and without the inclusion of explicit solvent molecules in the calculation protocol. The research has shown that the use of the composite method, without including explicit solvent molecules, provides results similar to those obtained with the M06-2X method, the latter at a much lower computational cost, but with mean absolute errors greater than 200 mV (or 4.6 kcal·mol-1). The use of a hybrid method, in which explicit molecules are included in the direct calculation protocol, can lead to the prediction of reduction potentials with an absolute error of less than 40 mV (less than 1.0 kcal·mol-1).