This paper summarizes the effects of irregular shape on the results of a quantitative X-ray fluorescence (XRF) micro-analysis. These effects become relevant when an XRF analysis is performed directly on an investigated material. A typical example is XRF analyses of valuable and historical objects whose measurements should be performed non-destructively and non-invasively, without taking samples. Several measurements and computer simulations were performed for selected metallic materials and shapes to evaluate the accuracy and precision of XRF. The described experiments and the corresponding Monte Carlo simulations were related to the XRF device designed and utilized at the Czech Technical University. It was found that the relative uncertainty was typically about 5-10% or even higher in quantitative analyses of minor elements due to irregular shapes of surfaces. This must be considered in cases of the interpretation of XRF results, especially in the cultural heritage sciences. The conclusions also contain several recommendations on how to measure objects under hard-to-define geometric conditions with respect to reduction in the surface effect in quantitative or semi-quantitative XRF analyses.
Keywords: Monte Carlo method; X-ray fluorescence; metal; quantitative analysis.