Phase contrast in x-ray imaging

X-ray phase contrast imaging (PCI) has the potential, over certain energy ranges, to improve conventional x-ray diagnostic practice by utilizing the higher phase than absorption coefficient in the complex index of refraction for low atomic Z materials, thereby potentially enhancing the di®erences between healthy and cancerous tissue. Mammography is a particularly suitable application due to its high biopsy incidence. Of the different types of PCI, the in-line holographic method or propagation-based imaging (PBI) has the ability to utilize polychromatic x-ray sources. In clinical practice, polychromatic x-ray sources require filtering to remove the lower energy rays that have insufficient energy to penetrate the body and provide useful diagnostic information. However, the introduction of filters in PBI is known to cause phase contrast degradation, and this thesis examines both experimentally and theoretically, the various factors affecting phase contrast. Experimental factors using a Feinfocus micro-focus, tungsten x-ray source with 0.5 mm beryllium inherent filtration, included: varying the tube voltage and source size, introducing extra filters of different materials and thicknesses including grain size and surface smoothness, and objects of different shapes and materials and at various positions in the beam. Image detection was via Fuji image plates and BAS-5000 scanner. The most convenient filter material was aluminium, and the object showing the best phase contrast was a thin walled, cylindrical fibre (Cuprophan RC55) normally used in dialysis. The x-ray spectrum of the tube was also measured using a SiLi detector system, with and without various filters. Theoretically, phase contrast was simulated via geometric ray optics, Fresnel and Fresnel-Kirchhoff wave diffraction methods. Inelastic scattering was also included in a simple model but was found to have only a very minor effect. Simulations showed good qualitative concurrence between the ray optics and Fresnel diffraction methods, with the Fresnel method providing greater detail in the line profile shape of the object, but essentially showing the same behaviour. Phase contrast reduction was found to be caused mainly by beam hardening, which narrows the phase peaks obtained from dispersion in the object, and together with source and detector considerations which then smooth the narrower filtered phase peaks more than the unfiltered case. There were some discrepancies with theoretical simulations producing absolute values approximately twice that of experimental ones, but the overall behaviour of phase contrast with the imaging parameters was consistent.