Koenig T., Zuber M., Trimborn B., Farago T., Meyer P., Kunka D., Albrecht F., Kreuer S., Volk T., Fiederle M., Baumbach T.
in Physics in Medicine and Biology, 61 (2016) 3427-3442, 3427. DOI:10.1088/0031-9155/61/9/3427
© 2016 Institute of Physics and Engineering in Medicine. The x-ray dark-field contrast accessible via grating interferometry is sensitive to features at length scales well below what is resolvable by a detector system. It is commonly explained as arising from small-angle x-ray scattering (SAXS), and can be implemented both at synchrotron beamlines and with low-brilliance sources such as x-ray tubes. Here, we demonstrate that for tube based setups the underlying process of image formation can be fundamentally different. For focal spots or detector pixels that comprise multiple grating periods, we show that dark-field images contain a strong artificial and system-specific component not arising from SAXS. Based on experiments carried out with a nanofocus x-ray tube and the example of an excised rat lung, we demonstrate that the dark-field contrast observed for porous media transforms into a differential phase contrast for large geometric magnifications. Using a photon counting detector with an adjustable point spread function, we confirm that a dark-field image can indeed be formed by an intra-pixel differential phase contrast that cannot be resolved as such due to a dephasing between the periodicities of the absorption grating and the Talbot carpet. Our findings are further corroborated by a link between the strength of this pseudo-dark-field contrast and our x-ray tube’s focal spot size in a three-grating setup. These results must not be ignored when measurements are intended to be reproducible across systems.