Abstract:
Soft X-ray Fourier transform holography (FTH) is a lensless imaging technique. In the past, it has been mostly used to image magnetic samples as well as biological specimens with a lateral resolution up to 40 nm. Through the interference of the object's scattered wave with a known reference wave the relative phase information is encoded as intensity modulation in the hologram. This allows for the obtaining of a real space image of the sample by a simple Fourier transform of the hologram. Usually, FTH with soft X-rays is realized by using a fixed mask-sample design where the sample is rigidly connected to the holographic mask and the reference wave is produced by a small pinhole. With this setup it is not possible to image samples which are larger than the field of view as it is fixed. Furthermore, this design doesn't allow for the recording of a three dimensional tomographic dataset. This is due to the high aspect ratio of the reference hole (typically 1:10) which blocks X-rays already at small rotation angles. This thesis deals with the development of FTH by increasing the field of view as well as making it possible to record tomograms. The field of view restriction can be overcome by the separation of the sample from the holographic mask. To this end, the sample is prepared on a X-ray transparent silicon nitride membrane. The holographic mask and the sample membrane can now be moved with respect to each other. This allows the imaging of larger samples by taking holograms at several sample positions. A small gap between mask and membrane usually exists. If this gap is larger than the depth of field, the holographic image reconstruction is out of focus. A wave field propagation based approach to correct for this effect is designed and successfully tested. The limitations of this approach are investigated. In particular for high resolution FTH with variable field of view, the approach turns out to be crucial. A small gold sphere on a transparent silicon nitride membrane can be used to produce the reference wave instead of a small aperture. This geometry makes it possible to rotate the sample reference structure. A complete dataset of two dimensional projections can be recorded to compute a three dimensional tomogram of the sample. The approach is successfully established using a diatom test sample where the 3D structure is reconstructed. This constitutes the first X-ray tomogram recorded via FTH. Furthermore, the gold sphere is replaced by extended reference structures (coded platinum dot array and a thin slit milled trough a gold membrane), which also allows to rotate the sample to record a tomographic dataset. Both reference structures are more robust against noise than a single gold sphere. This is advantageous when the photon flux is limited or the sample has to be protected from radiation damage. Both alternative reference structures are successfully used to record tomographic datasets of a test sample. On this basis two 3D tomograms are calculated.