Open AccessValidation of experimental molecular crystal structures with dispersion‐corrected density functional theory calculations
Author(s) -
Van De Streek Jacco,
Neumann Marcus A.
Publication year - 2010
Publication title -
acta crystallographica section b
Language(s) - English
Resource type - Journals
eISSN - 1600-5740
pISSN - 0108-7681
DOI - 10.1107/s0108768110031873
Subject(s) - density functional theory , crystal structure , cartesian coordinate system , crystal (programming language) , correctness , displacement (psychology) , materials science , dispersion (optics) , crystal structure prediction , crystallography , computational physics , molecular physics , physics , computational chemistry , chemistry , mathematics , computer science , algorithm , geometry , quantum mechanics , psychology , programming language , psychotherapist
This paper describes the validation of a dispersion‐corrected density functional theory (d‐DFT) method for the purpose of assessing the correctness of experimental organic crystal structures and enhancing the information content of purely experimental data. 241 experimental organic crystal structures from the August 2008 issue of Acta Cryst. Section E were energy‐minimized in full, including unit‐cell parameters. The differences between the experimental and the minimized crystal structures were subjected to statistical analysis. The r.m.s. Cartesian displacement excluding H atoms upon energy minimization with flexible unit‐cell parameters is selected as a pertinent indicator of the correctness of a crystal structure. All 241 experimental crystal structures are reproduced very well: the average r.m.s. Cartesian displacement for the 241 crystal structures, including 16 disordered structures, is only 0.095 Å (0.084 Å for the 225 ordered structures). R.m.s. Cartesian displacements above 0.25 Å either indicate incorrect experimental crystal structures or reveal interesting structural features such as exceptionally large temperature effects, incorrectly modelled disorder or symmetry breaking H atoms. After validation, the method is applied to nine examples that are known to be ambiguous or subtly incorrect.