Abstract
Trace elements within biological tissues are heterogeneously distributed. This complicates the task of deriving concentrations that represent an organ or specimen as a whole. The derivation of representative concentrations is important in the investigation of tissue health status or exposure to the individual from occupational or man-made pollution sources. Therefore, a knowledge of the elemental dispersions in biological tissues is required. Proton induced X-ray emission (PIXE) analysis is employed in the study of the elemental heterogeneity of porcine liver, kidney, heart and lung. Specimens are analysed in two different modes. One method involves the extraction of sub-samples that are dried, homogenised and manufactured into thick target pellets. This approach however limits the spatial resolution on which elemental distributions may be derived and hence thick specimen sections that can be irradiated directly may be preferred. This type of target though suffers in that surfaces are irregular and proton irradiation and X-ray take-off angles are ill defined. The effect of these surface imperfections upon X-ray yield in PIXE are investigated by the development of simple stylised models. The physical parameters of these models are varied and the elements most affected and dominant factors in modifying X-ray yield are identified. The trace element content for like tissues between targets in the form of pellets and freeze-dried sections are compared and mostly excellent correlation is found. The analysis of specimens in either of these modes stresses the high elemental inhomogeneity of biological samples. A quantitative determination of this elemental heterogeneity is made by the derivation of sampling factors, the minimum mass of material required to reduce elemental variations to a given level of precision. Those sampling factors derived by utilising the data from pelletised targets agree well with the limited values from the literature, whereas a large difference is found for those calculated from thick specimen target sections. This disagreement is thought to be due to the failure of sampling factor theory at the small sampling mass employed in the analysis of the latter targets. Photon transmission tomography was investigated for the ability of the technique to provide a measure of biological specimen heterogeneity, differentiate between different composite tissues and identify regions of interest. This may prove useful for the selection of sub-samples for subsequent trace element analysis. Biological specimens were scanned in fresh and dry states to ascertain the most favourable sample preparation technique to best achieve the above aims, the dry sample states were preferred. Freeze-dried specimens are imaged under differing scanning parameters and their data compared to theoretical values derived from PIXE and Rutherford backscattering (RBS) analysis. Good agreement is found. Regions of interest may be identified in tomographs, this being due to density variations rather than elemental variations, tissues of similar but different composition not being differentiatable due to image noise which is a product of finite counts in reconstructed images. However, with the improvement of photon counting statistics in images, these tissues may be more discernible from one another in tomographs thus making photon transmission tomography a viable technique for the selection of representative sub-samples for subsequent elemental analysis.