The use of pattern decomposition methods such as Pawley fitting in decomposing the XRD patterns allows us to reliably reconstruct phase distribution maps in virtual cross sections. Problems caused by poor crystallinity of the materials and Bragg peak overlap are circumvented by this method. The iterative MLEM algorithm was chosen for the reconstruction of the tomographic data because of its robustness, even though it causes a slight blurring of the resulting tomographic images. In cases where FBP is employed,which is an analytic reconstruction method, an effective spatial resolution in the direction perpendicular to the paint layers that is approximately equal to the microbeam size, is obtained.
The combination of microbeam XRF and XRD tomography can be applied to any millimeter-sized crystalline material while obtaining micrometer resolution. In a follow-up article, we intend to use this method to visualize the distribution of pigmented species in historical multilayered materials sampled from 15th to 17th century oil paintings and compare its strong and weak points to other depth-selective methods such as confocal μ-XRF.