Surface roughness, reflected in local variations in inclination, affects the intensity of the detected X-rays. These changes, caused by alterations in the X-ray emission paths, can lead into inaccuracies in determining elemental concentrations especially during non-destructive analysis of ancient artifacts. In this study, a methodology was developed to rectify the elemental analysis results obtained from micro-PIXE, with consideration given to the surface roughness of the sample. An ancient coin was analyzed using a focused proton beam scanning with an energy of 3MeV and the characteristic X-rays were detected vith a four-segment Silicon Drift Detector (SDD). Here, at first we present an algorithm for reconstructing the 3D surface topography of an ancient coin from micro-PIXE data acquired with a four-segment silicon drift detector (SDD) [1, 2], building on prior elemental characterization. Through this, the local inclination angle was initially determined across various surface regions. This facilitated the correction of the path length traversed by the X-rays within the sample (ξcorr). Subsequently, the yield correction factor was computed in accordance with the relationship pertaining to X-ray yield at depth z as follows: where Ycorr and Ymeas are corrected and measured X-ray yields, respectively. The measured elemental concentrations are quantified from the measured 2D X-ray yield maps using GeoPIXE. ε is detector efficiency, Ω is the solid angle, Q/e is proton flux, and σ(E) is the X-ray production cross-section. C(x, y) is elemental concentration, μ is the absorption coefficient, and ξflat represents the path length traversed by the X-rays within the sample, excluding considerations of surface roughness. At the hitting beam point, the beam encompasses a volume with specific dimensions in the x and y directions, as well as a penetration depth R. Within this confined volume, the sample is presumed to be homogeneous. This assumption of homogeneity at the small volume scale permits the exclusion of the concentration from the integral in the X-ray efficiency equation, thereby establishing a linear relationship between X-ray yield and concentration as: By defining this correction factor based on the comparison of X-ray yield between smooth and rough surfaces, the 2D elemental concentrations were quantitively corrected between ± 10%. The findings underscore the significance of considering the sample surface geometry in augmenting the validity of quantitative data derived from micro-PIXE analyses.
