3D X-ray fluorescence (XRF) mapping provides quantitative information on the chemical composition of samples, but requires a focused X-ray beam that scans the sample point by point at each angle of rotation. This sequential acquisition therefore requires a very large number of measurements and very long acquisition times, while also increasing the radiation dose received locally by the sample.
A new approach to 3D XRF tomography using ghost imaging (XRF-GT) employs a wide beam modulated by a transmission mask (e.g., a monolayer of randomly distributed tungsten powder). This mask acts as a filter and generates high-contrast illumination patterns on the sample that are spatially distributed at random. By simultaneously measuring the fluorescence across the entire illuminated area, the elements can be reconstructed without the need for a physical point-by-point scan, thanks to a novel reconstruction method that simultaneously integrates the measurements obtained for all rotation angles. This approach exploits the information and redundancies present in the 3D structure, thereby allowing work to be carried out with far fewer measurements.
On the ESRF's ID19 beamline, this method reduced the number of measurements by a factor of 43, while maintaining a spatial resolution of 20 to 30 µm. It thus paves the way for faster 3D chemical mapping of complex samples.