Development of differential evolution algorithms applied to crystal structure solution from powder diffraction data

Bell, Duncan (2012). Development of differential evolution algorithms applied to crystal structure solution from powder diffraction data. University of Birmingham. Ph.D.

Preview

Bell_12_PhD.pdf
PDF - Accepted Version
Download (3MB)

Abstract

An understanding of the crystal structure can aid in the rationalisation of physicochemical properties exhibited by a crystalline material. Advances in the area of direct space crystal structure solution means that it is becoming easier to determine crystal structures from powder diffraction data. However, due to the number of structural models generated during structure solution calculations, direct space methods are computationally demanding.

Work presented in this thesis reports the optimisation of a differential evolution (DE) algorithm and a cultural differential evolution (CDE) algorithm to reduce the computational demands of direct space methods. Characteristics particular to certain crystal structures are identified as having a significant effect on the efficiency and robustness of structure solution calculations by DE and CDE.

The development of a new algorithm that closely mimics the natural evolution of a species is discussed. Results presented in this thesis demonstrate that this new algorithm is significantly more efficient than the DE algorithm.

Despite the complexity of powder diffraction patterns recorded for biphasic crystalline materials, in this thesis, the successful development and application of a direct space method to the simultaneous structure solution of two crystals from a biphasic powder pattern is reported.

Type of Work:

Thesis (Doctorates > Ph.D.)

Award Type:

Doctorates > Ph.D.

Supervisor(s):