Unsupervised real-world knowledge extraction via disentangled variational autoencoders for photon diagnostics.

We present real-world data processing on measured electron time-of-flight data via neural networks. Specifically, the use of disentangled variational autoencoders on data from a diagnostic instrument for online wavelength monitoring at the free electron laser FLASH in Hamburg. Without a-priori knowledge the network is able to find representations of single-shot FEL spectra, which have a low signal-to-noise ratio. This reveals, in a directly human-interpretable way, crucial information about the photon properties. The central photon energy and the intensity as well as very detector-specific features are identified. The network is also capable of data cleaning, i.e. denoising, as well as the removal of artefacts. In the reconstruction, this allows for identification of signatures with very low intensity which are hardly recognisable in the raw data. In this particular case, the network enhances the quality of the diagnostic analysis at FLASH. However, this unsupervised method also has the potential to improve the analysis of other similar types of spectroscopy data.

Scan-Free GEXRF in the Soft X-ray Range for the Investigation of Structured Nanosamples.

Scan-free grazing-emission X-ray fluorescence spectroscopy (GEXRF) is an established technique for the investigation of the elemental depth-profiles of various samples. Recently it has been applied to investigating structured nanosamples in the tender X-ray range. However, lighter elements such as oxygen, nitrogen or carbon cannot be efficiently investigated in this energy range, because of the ineffective excitation. Moreover, common CCD detectors are not able to discriminate between fluorescence lines below 1 keV. Oxygen and nitrogen are important components of insulation and passivation layers, for example, in silicon oxide or silicon nitride. In this work, scan-free GEXRF is applied in proof-of-concept measurements for the investigation of lateral ordered 2D nanostructures in the soft X-ray range. The sample investigated is a Si3N4 lamellar grating, which represents 2D periodic nanostructures as used in the semiconductor industry. The emerging two-dimensional fluorescence patterns are recorded with a CMOS detector. To this end, energy-dispersive spectra are obtained via single-photon event evaluation. In this way, spatial and therefore angular information is obtained, while discrimination between different photon energies is enabled. The results are compared to calculations of the sample model performed by a Maxwell solver based on the finite-elements method. A first measurement is carried out at the UE56-2 PGM-2 beamline at the BESSY II synchrotron radiation facility to demonstrate the feasibility of the method in the soft X-ray range. Furthermore, a laser-produced plasma source (LPP) is utilized to investigate the feasibility of this technique in the laboratory. The results from the BESSY II measurements are in good agreement with the simulations and prove the applicability of scan-free GEXRF in the soft X-ray range for quality control and process engineering of 2D nanostructures. The LPP results illustrate the chances and challenges concerning a transfer of the methodology to the laboratory.