Development of an electron gun design optimisation methodology

This thesis was submitted for the award of Doctor of Philosophy and was awarded by Brunel University LondonThe design of high quality electron generators to meet specific requirements is important in the application of these devices to a variety of materials processing systems (including welding, cutting and additive manufacture), X-ray tubes for medical, scientific and industrial applications, microscopy and lithography. Designs can be analysed by field solvers, and electron trajectories plotted to provide an indication of the beam quality. Incremental improvement of designs has normally been executed by trial and error, and this can be a time consuming activity requiring expert intervention for each iteration of the design process. The unique contribution made to knowledge by this work is the application of optimisation techniques to the design of electron guns to produce beams with the required optical properties. This thesis presents a review of the design of electron guns, including a discussion of thermionic cathode material properties and their suitability for use in electron guns for processing materials, the influence of space-charge on gun design and the derivation of salient beam metrics to characterise the beam. Beam quality metrics have been developed that allow quantification of electron beam characteristics, allowing objectives to be set for the optimisation process. Additionally, a method is presented that enables real world measurements to be directly compared with modelled beams. Various optimisation methods are reviewed. A genetic algorithm was selected, which would use gun modelling and beam characterisation calculations as the objective function, as a suitable method for application to this problem. However, it was recognised that selections for the best evolutionary parameters, the population size, number of parents, the mutation rate and mutation scale, were not readily determined from published work. An investigation is presented where a range of evolutionary parameters was tested for a set of geometrical problems, which had some similarity to electron gun design but could be computed sufficiently quickly to enable an extensive survey, and the most efficient combination of parameters was identified. Detail is given of the customisation of a genetic evolutionary optimisation method for the design of electron guns. Examples are presented of electron gun design optimisation processes to meet specified beam requirements within defined geometric and electrical constraints. The results of this work show that optimum evolutionary parameter settings for the geometric problem vary with the complexity of the problem and trends have been identified. Application of these parameters to an electron gun optimisation has been successful. The derived beam parameter metrics have been applied to electron guns as an objective function. Comparisons of modelled predictions of the beam characteristics with the measured real world values have been shown to be reasonable