Building and testing correlations for the estimation of one‐electron reduction potentials of a diverse set of organic molecules. (29th January 2015)

• Record Type: Journal Article
• Authors: Méndez‐Hernández, Dalvin D.; Gillmore, Jason G.; Montano, Luis A.; Gust, Devens; Moore, Thomas A.; Moore, Ana L.; Mujica, Vladimiro
• Is Part Of: Journal of physical organic chemistry. Volume 28:Number 5(2015:May)
• Page Start: 320
• Page End: 328
• Publication Date: 2015-01-29
• Subjects: computational electrochemistry -- frontier orbitals -- linear correlation -- LUMO energy -- redox potentials; Chemistry, Physical organic -- Periodicals; 547.1
• DOI: 10.1002/poc.3413
• Languages: English
• ISSNs: 0894-3230

Abstract

Abstract : We describe and evaluate a method for computationally predicting reduction potentials of a diverse group of organic molecules by linearly correlating calculated lowest unoccupied molecular orbital energies with ground state reduction potentials measured in acetonitrile. The approach is shown to provide a unique combination of extreme computational simplicity and excellent accuracy across a diverse range of organic structures and a wide window of reduction potentials. A disparate set of molecules (74 compounds belonging to six distinct structural families, comprised of molecules containing C, H, N, O, F, Cl, and Br, with functional groups including esters, ketones, halides, nitriles, quinones, alkenes, arenes, heteroarenes, and pyridinium and higher benzologs, all containing conjugated pi systems, spanning a 3.5‐V range of reduction potentials) was used to build the correlations. This methodology was found to be computationally inexpensive compared with other approaches and to permit the useful prediction of reduction potentials of additional molecules of diverse structural types not included in the families used to determine the correlation parameters. The effects of varying the basis set used in the B3LYP electronic structure calculations and including solvent (compared with calculations in gas phase) were also examined. It was found that the inclusion of a continuum solvent model in the calculations was required for accurate results, particularly when including cationic species in the correlations (although when only neutral molecules were examined, reasonable results could even be obtained in vacuo ). Several of the best correlations were used to predict the reduction potentials of seven much larger and structurally diverse chromophores that were not included in the correlation data set. Strong correlations ( r 2 values > 0.99) with very good predictive abilities (root mean square deviation < 60 mV) were found. This extremely simple and computationally efficient entirely closed‐shell methodology is proven robust and useful for the design of new molecules capable of participating in redox processes, including electron transfer reactions. Copyright © 2015 John Wiley & Sons, Ltd. Abstract : We report on a method for computationally predicting reduction potentials of organic molecules by linearly correlating calculated lowest unoccupied molecular orbital energies with experimental ground state reduction potentials. A total of 74 compounds spanning six distinct structural families and a 3.5‐V range of reduction potentials were used to build the correlations. This approach is shown to provide a unique combination of computational simplicity and excellent accuracy, even for much larger chromophores and structurally diverse chromophores not included in the correlations.