This thesis concerns the application of electron diffraction to the problem of structure solution. The technique of precession electron diffraction, in particular, is discussed as a method to improve the applicability of methods borrowed from X-ray crystallography to intensities measured from electron diffraction.

Simulations are presented which demonstrate that diffracted intensities with precession electron diffraction are less sensitive to the phases of structure factors compared to those recorded without, consistent with the intensities becoming “more kinematical” in nature.

The technique of automated diffraction tomography is developed. Transformations are derived which map the positions of diffraction maxima into three-dimensional reciprocal space following an automated peak detection process. The factors which complicate the process of unit cell determination are discussed, and an algorithm designed for autoindexing “obstinate” lists of X-ray reflections was found to be able to determine the unit cells accurately. Intensity measurements from the tomographically acquired diffraction patterns were taken in an automated fashion using this method, and Patterson maps calculated which were comparable with maps calculated from zone-axis precession electron diffraction intensities or kinematical simulation. An algorithm for geometrical refinement of the unit cell and tilting geometry is proposed.

Finally, the application of the technique of precession electron diffraction, combined with techniques developed in the course of this research, is described in the context of the partial solution of an unknown intermediate form of tin oxide. The results support a revised form of a structure recently proposed from ab-initio structure simulations, with a tin sublattice similar to a supercell of the rutile structure of SnO2. Some signs of oxygen atoms were found in the Fourier maps, but the results of a least-squares refinement were unsatisfactory.