X-ray crystallography is a widely used analytical technique for the structure solution of small molecules. Since the determination of the rock salt structure in 1913 by Henry and Lawrence Bragg the technique has developed allowing the solution of larger and more complex structures. The information that can be determined about these structures has increased as X-ray sources, detectors, and computational methods have improved. However, certain properties of molecules cannot always be directly determined from single wavelength X-ray diffraction. These include, inter alia: the site specific oxidation or spin state of an element in compounds where more than one state of the same element exist; discrimination between consecutive heavy elements in the periodic table. As the size of molecules being studied increases, reduced data resolution also becomes a problem. The aim of this research was to determine whether these problems can be addressed by measuring the changing anomalous scattering contribution of heavy atoms within structure through careful selection of the X-ray energy. Firstly, I report an investigation into the problemof discriminating oxidation state, spin state and elements of near identical scattering by exploiting their anomalous signal. I first present DetOx, a program written during the course of the project to deconvolute the fluorescence signal from materials containing more than one state of the same element into their respective spectra. This allows the calculation of anomalous scattering factors for both atomic states of an atom, which can subsequently be used to refine the occupancy of the different states at ambiguous sites within the crystal structure. The approach taken here, to determine differences due to relatively small anomalous signals, is analogous to the refinement of the Flack parameter whereby small changes in many hundreds or thousands of observations can be used to fit a parameter with a high degree of precision and accuracy. I show the application of this technique to the mixed oxidation state compound, GaCl<sub>2</sub>, and the two-step spin crossover material, Fe(btr)<sub>3</sub>(ClO<sub>4</sub>)<sub>2</sub>. Refinement of the occupancy of charged ions on multiple sites using data at a single, carefully selected wavelength proved successful for these compounds, although upon extension to materials containing a larger number of anomalous scatterers, the absorption became a major issue in the data along with problems associated with simultaneously refining occupancies at more sites in the structure. We have demonstrated that calculations can be made to select specific experimental data to collect in order to improve the measured signal. However, due to limitation of the current collision model on the diffractometer used we have not yet been able to construct data collection strategies to take advantage of this. I next present a new ratio refinement technique to overcome this absorption problem due to the increased number of scatterers. By using ratios between datasets close in energy, but below the absorption edge, we were able to exploit small changes in f' without encountering absorption problems associated with the increase in f''. These ratio values were then refined against a lab structure using a modified version of CRYSTALS to reveal the site specific occupancies of different atomic species within a given structure. For mixed-valence compounds, e.g. Mn<sub>3</sub> and Mn<sub>6</sub> clusters, the difference in anomalous signal between the different states proved too small for a stable least-squares refinement solution. However, we have shown that using a simulated annealing algorithm (to refine only occupancies), we can consistently obtain the expected structure. For mixed-metal structures e.g. the Mn<sub>5</sub>Co<sub>4</sub> cluster, there was enough contrast in the data to refine occupancies with a least-squares approach, and these results were supported using simulated annealing. Lastly, I describe the application of structure solution techniques based on methods used in macromolecular crystallography to 'large' small molecules. Traditionally these have been reserved for non-centrosymmetric protein structures, however with the trend of synthesising larger and larger small molecules, problems encountered in macromolecular crystallography leading to low resolution datasets are becoming increasingly common. I have shown that it is possible to solve the structure of centrosymmetric structures by exploiting the anomalous signal in multiple wavelength diffraction experiments. The technique is applied successfully to two relatively small molecules, however the results are promising for moving to larger structures in the future.
Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:647701 |
Date | January 2015 |
Creators | Sutton, Karim J. |
Contributors | Cooper, Richard |
Publisher | University of Oxford |
Source Sets | Ethos UK |
Detected Language | English |
Type | Electronic Thesis or Dissertation |
Source | http://ora.ox.ac.uk/objects/uuid:3944a985-9339-4c8c-970b-4b460848f200 |
Page generated in 0.0022 seconds