doiSerbia

Theoretical study via DFT for prediction of 13C and 1H NMR data of two diterpenoids derived from the root of salvia grandifolia

da Costa Renato A. (Program of Post-Graduation in Chemistry, Federal University of Pará, Guamá, Belém, Pará, Brazil + Federal Institute of Pará, Rodovia PA, Parauapebas, Pará, Brazil)
Silva Sebastião G. (Federal University of Pará, Program of Post-Graduation in Chemistry, Guamá, Belém, Pará, Brazil)
de O. Silva Silvana (Federal University of Pará, Program of Post-Graduation in Chemistry, Guamá, Belém, Pará, Brazil)
Cruz Jorddy N. (Federal University of Pará, Laboratory of Preparation and Computation of Nanomaterials, C.P., Belém, PA, Brazil)
da Costa Wanessa A. (Program of Post-Graduation in Natural Resources Engineering (PRODERNA/ITEC), Federal University of Para, Rua Augusto Corrêa S/N, Guamá, Belém, Pará, Brazil)
dos Santos Brasil Luciane S.N. (Department of Food Technology Center for Natural Sciences and Technology (CCNT) State University of Pará, Travessa Angustura, Pedreira, Belém, Pará, Brazil)
Silva Raí C. (Federal University of Amapá, Laboratory of Modeling and Computational Chemistry, Macapá, Amapá, Brazil)
Santos Cleydson B.R. (Federal University of Amapá, Laboratory of Modeling and Computational Chemistry, Macapá, Amapá, Brazil)
Alves Cláudio N. (Federal University of Pará, Program of Post-Graduation in Chemistry, Guamá, Belém, Pará, Brazil)
Barros Brasil Davi De S. (Federal University of Pará, Program of Post-Graduation in Chemistry, Guamá, Belém, Pará, Brazil)

The DFT method has become a promising alternative in the support of traditional NMR experimental techniques, comparing experimental data with theoretical data, thereby achieving accurate and satisfactory results. In the present study, the experimental data of two diterpenes were compared to the theoretical data obtained by the GIAO method, applying DFT at the B3LYP/cc-pVDZ and B3PW91/DGDZVP levels to verify the degree of correlation, significance and predictability of the models obtained with the purpose of proving which of the computational methods is the most efficient for this class of substances. The theoretical NMR values obtained at B3LYP/cc-pVDZ were the more satisfactory, showing better linear correlations, presenting greater degrees of adjustments, significance and predictability compared to the B3PW91/DGDZVP method. Thus, these values were used to define Hα and Hβ 1, 2, 3, 6 and 7 for diterpene 1 and 1, 2, 3, 6 and 7 for diterpene 2, which was not possible using experimental data. This showed that the quantum method used could help in the structural elucidation of natural products.

Keywords: statistical analysis, B3LYP/cc-pVDZ, linear regression