Crystal and geometry‐optimized structure of an anthracene‐based Diels–Alder adduct. Issue 7 (26th June 2020)

• Record Type: Journal Article
• Authors: Hillman, Zachary E.; Tanski, Joseph M.; Roberts, Andrea
• Is Part Of: Acta crystallographica. Volume 76:Issue 7(2020)
• Page Start: 639
• Page End: 646
• Publication Date: 2020-06-26
• Subjects: anthracene -- Diels‐Alder -- computational chemistry -- DFT -- crystal structure -- Boltzmann population analysis; Crystallography -- Periodicals; Crystals -- Periodicals; 548.3
• DOI: 10.1107/S2053229620008128
• Languages: English
• ISSNs: 2053-2296

Abstract

Abstract : Computational calculations of an anthracene‐based Diels–Alder adduct predicting density functional theory (DFT) optimized geometries in the gas phase are compared in terms of accuracy relative to the solid‐state crystal structure and computational cost. Abstract : Computational calculations of an anthracene‐based Diels–Alder adduct, namely, 17‐ethyl‐1‐hydroxymethyl‐17‐azapentacyclo[6.6.5.0 2, 7 .0 9, 14 .0 15, 19 ]nonadeca‐2, 4, 6, 9, 11, 13‐hexaene‐16, 18‐dione, C21 H19 NO3, predicting density functional theory (DFT) optimized geometries in the gas phase are compared in terms of accuracy relative to the solid‐state crystal structure and computational cost. Crystal structure determination and Hirshfeld surface analysis of the racemic product reveal that the molecules are linked by O—H…O=C hydrogen bonds between the hydroxy and carbonyl groups, accounting for 9.5% of the intermolecular contacts, while H…H contacts represent 56.9% of the total. Boltzmann population analysis of computed relative rotamer abundances in the gas phase are based on lower‐level geometry optimization and thermochemical corrections coupled with higher‐level electronic energy calculations using the B2PLYP double‐hybrid functional. As expected, the choice of density functional has a greater effect than the basis set on accuracy for all levels of theory. For any given functional, increasing the basis set size did not always correlate with increasingly accurate structures. The hybrid functional B3LYP without dispersion correction routinely gave the closest approximations to the crystal structure where the B3LYP/aug‐cc‐pVDZ combination afforded the best structure (r.m.s. deviation = 0.1314 Å). However, the B3LYP/6‐31+G(d, p) level of theory represents the best compromise between accuracy (r.m.s. deviation = 0.1388 Å) and cost as it yielded appreciably accurate results in a fraction of the time compared to other method combinations.