Benchmarking the DFT methodology for assessing antioxidant-related properties: quercetin and edaravone as case studies.

The overall objective was to identify an accurate computational electronic method to virtually screen phenolic compounds through their antioxidant and free-radical scavenging activity. The impact of a key parameter of the density functional theory (DFT) approach was studied. Performances of the 21 most commonly used exchange-correlation functionals are thus detailed in the evaluation of the main energetic parameters related to the activities of two prototype antioxidants, namely quercetin and edaravone, is reported. These functionals have been chosen among those belonging to three different families of hybrid functionals, namely global, range separated, and double hybrids. Other computational parameters have also been considered, such as basis set and solvent effects. The selected parameters, namely bond dissociation enthalpy (BDE), ionization potential (IP), and proton dissociation enthalpy (PDE) allow a mechanistic evaluation of the antioxidant activities of free radical scavengers. Our results show that all the selected functionals provide a coherent picture of these properties, predicting the same order of BDEs and PDEs. However, with respect to the reference values, the errors found at CBS-Q3 level significantly vary with the functional. Although it is difficult to evidence a global trend from the reported data, it clearly appears that LC-ωPBE, M05-2X, and M06-2X are the most suitable approaches for the considered properties, giving the lowest cumulative mean absolute errors. These methods are therefore suggested for an accurate and fast evaluation of energetic parameters related to an antioxidant activity via free radical scavenging.