Control of White Light Emission for Illumination and Imaging

Abstract

The bulk of the work in this thesis addresses a well defined problem in the field of white solid-state lighting. The problem was defined by this project’s industry sponsors and consequentially this thesis is to be covered by an embargo period in which it is not publicly accessible, to protect commercial sensitivities and intellectual property. The title of this thesis and this abstract are not protected by these restrictions and so must be selectively descriptive. There are two bodies of work that can be freely described here. The first, presented in Chapter 2 concerns the performance of parabolic antennas as collimators when their size is reduced to of the order of the wavelength of emission. Using finite-element modelling we calculate the dependence of the amount of power directed into a 20° half-angle cone on the emitter’s position within the paraboloid and compare with results obtained using geometrical optics. It is demonstrated that due to variations in the local density of optical states, changing the characteristic size of the reflector within the range from 0.5 to 1.5 times the emission wavelength has a strong bearing on the optimum emitter position, a position that does not in general coincide with the paraboloid’s focus. We calculate that the optimal antenna size and emitter position allow for the maximum directed power to exceed that obtained in the geometrical optics regime. In the other section that can be described here, Chapter 6, we present a novel technique in the field of ghost imaging. In our method we show that one can modify the illumination basis patterns to perform common image post-processing steps directly during the reconstruction process. This can remove the amplification of detector noise caused by post processing with a filter and yields favourable noise spatial statistics. This technique can be applied to any operation which could be written as a matrix multiplication with the image, which includes the application of image filters by convolution.