Determining Molecular Structures and Conformations Directly from Electron Diffraction using a Genetic Algorithm

• Published in: Chemphyschem 2006-02, Vol.7 (2), p.353-362
• Main Authors: Habershon, Scott, Zewail, Ahmed H.
• Format: Article
• Language: English
• Subjects: Atomic and molecular physics; conformational analysis; electron diffraction; Exact sciences and technology; General molecular conformation and symmetry; stereochemistry; genetic algorithm; molecular modeling; Molecular properties and interactions with photons; Physics; Properties of molecules and molecular ions; reaction dynamics
• ISSN: 1439-4235; 1439-7641
• DOI: 10.1002/cphc.200500532

10.1002/cphc.200500532.absA global optimization strategy, based upon application of a genetic algorithm (GA), is demonstrated as an approach for determining the structures of molecules possessing significant conformational flexibility directly from gas‐phase electron diffraction data. In contrast to the common approach to molecular structure determination, based on trial‐and‐error assessment of structures available from quantum chemical calculations, the GA approach described here does not require expensive quantum mechanical calculations or manual searching of the potential energy surface of the sample molecule, relying instead upon simple comparison between the experimental and calculated diffraction pattern derived from a proposed trial molecular structure. Structures as complex as all‐trans retinal and p‐coumaric acid, both important chromophores in photosensing processes, may be determined by this approach. In the examples presented here, we find that the GA approach can determine the correct conformation of a flexible molecule described by 11 independent torsion angles. We also demonstrate applications to samples comprising a mixture of two distinct molecular conformations. With these results we conclude that applications of this approach are very promising in elucidating the structures of large molecules directly from electron diffraction data. Structure of conformationally flexible molecules (see diagram): Quantum mechanical geometry optimizations, which are expensive and time‐consuming, are bypassed as the authors present a protocol based on the application of a global optimization strategy (a genetic algorithm) which accurately predicts the structure directly from gas‐phase electron diffraction data.