TY - CONF A1 - Radtke, Martin T1 - Machine learning for direct quantification of XRF measurements N2 - In this talk I’ll describe the use of artificial neural networks (ANN) for quantifying X-ray fluorescence (XRF) measurements. The main idea of this talk is to give an overview of the process needed to generate a model that can then be applied to a specific problem. In XRF, a sample is excited with X-rays and the resulting characteristic radiation is detected to determine elements quantitatively and qualitatively. This is traditionally done in several time-consuming steps. I’ll show the possibilities and problems of using a neural network to realise a "one-click" quantification. This includes generating training data using Monte Carlo simulation and augmenting the existing data set with an ANN to generate more data. The search for the optimal hyperparameters, manually and automatically, is also described. For the case presented, we were able to train a network with a mean absolute error of 0.1% by weight for the synthetic data and 0.7% by weight for a set of experimental data obtained with certified reference materials. T2 - Seminar series: Artificial Intelligence applied to X-ray / synchrotron techniques CY - Online meeting DA - 24.06.2021 KW - Artificial intelligence KW - Machine learning KW - Synchrotron KW - XRF PY - 2021 AN - OPUS4-54140 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - JOUR A1 - Lutz, C. A1 - Hampel, S. A1 - Beuermann, S. A1 - Turek, T. A1 - Kunz, U. A1 - de Oliveira Guilherme Buzanich, Ana A1 - Radtke, Martin A1 - Fittschen, U. E. A. T1 - Investigation on Vanadium Species Distribution in Nafion™ 117 after Cyclization in a Vanadium Redox Flow Battery N2 - The vanadium redox flow battery (VRFB) is currently a potential candidate for stationary energy storage. A major challenge is the unintended vanadium transport through the separator, which results in a fade of capacity. To overcome this issue, it is necessary to understand the transport processes in the membrane on a more fundamental level. In this work, the vanadium species distribution in Nafion™ 117 after cyclization was investigated. Two membranes, one from a charged VRFB and another from a discharged VRFB, were analyzed using ultraviolet–visible spectroscopy (UV/VIS) and X-ray absorption near edge structure spectroscopy (XANES). Little difference between the two membranes was recognizable according to the UV/VIS results. In comparison, the XANES results showed that the membrane from the charged VRFB contains more V3+ than VO2+, whereas for the discharged case, more VO2+ is present in the membrane. KW - Synchrotron KW - BAMline KW - XANES KW - Vanadium redox flow battery PY - 2021 VL - 64 SP - 1 EP - 8 AN - OPUS4-54144 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - CONF A1 - Iro, M. A1 - Ingerle, D. A1 - Radtke, Martin A1 - Tsuji, K. A1 - Hampel, S. A1 - Fittschen, U. A1 - Streli, C. T1 - Investigation of the transmission properties of a polycapillary half-lens and their effect on quantification for confocal μXRF N2 - Confocal Micro X-ray fluorescence analysis (μXRF) is a powerful method for the destruction-free investigation of the elemental composition of various samples in three dimensions. Quantitative results rely heavily on the transmission properties of polycapillary optics, used in confocal setups. Different models for the functional correlation between energy and transmission of a polycapillary half-lens are compared with experimental data, established with a monochromatic confocal laboratory setup [1], as well as Monte-Carlo-Simulations [2]. The effect of the different approaches on quantitative results are compared using well known SRMs, ranging from thick samples to thin layers and structured materials. Along with this the basic principles for quantitative evaluations of XRF spectra taken with confocal setups are presented, and compared with experimental results. T2 - Denver X-Ray Conference CY - Online meeting DA - 02.08.2021 KW - Synchrotron KW - BAMline KW - Polycapillaries PY - 2021 AN - OPUS4-54143 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - JOUR A1 - De Samber, B. A1 - Scharf, Oliver A1 - Buzanich, Günter A1 - Garrevoet, J. A1 - Tack, P. A1 - Radtke, Martin A1 - Riesemeier, Heinrich A1 - Reinholz, Uwe A1 - Evens, R. A1 - De Schamphelaere, K. A1 - Falkenberg, G. A1 - Janssen, C. A1 - Vincze, L. T1 - Three-dimensional X-ray fluorescence imaging modes for biological specimens using a full-field energy dispersive CCD camera N2 - Besides conventional scanning X-ray fluorescence imaging at synchrotron sources, full-field X-ray fluorescence (FF-XRF) imaging techniques that do not implicitly require spatial scanning of the sample have become available. FF-XRF has become achievable thanks to the development of a new type of energy dispersive CCD-based 2D detector, also referred to as a 'color X-ray camera (CXC)' or 'SLcam'. We report on different imaging schemes for biological samples using FF-XRF imaging: (a) 2D 'zoom' imaging with pinhole optics using the 'camera obscura' principle; (b) 2D 'fixed magnification' imaging using magnifying polycapillary optics; and (c) 3D-FF-XRF imaging using an X-ray sheet beam or computed tomography (CT). The different FF-XRF imaging modes are illustrated using the crustacean Daphnia magna, a model organism for investigating the effects of metals on organism/ecosystem health, and foraminifera, a class of amoeboid protist. Detailed analytical characterization of the set-up is performed through analyzing various reference materials in order to determine limits of detection (LODs) and sensitivities. Experiments were performed using the BAMline at the BESSY synchrotron (Berlin, Germany) and using the P06 Hard X-ray Microprobe at the PETRAIII synchrotron (Hamburg, Germany). KW - CXC KW - BAMline KW - Maia detector KW - Synchrotron PY - 2019 DO - https://doi.org/10.1039/c9ja00198k VL - 34 IS - 10 SP - 2083 EP - 2093 PB - Royal Society of Chemistry CY - Cambridge, United Kingdom AN - OPUS4-49359 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER - TY - JOUR A1 - Radtke, Martin A1 - Rakotondrajoa, A. T1 - Machine learning based quantification of synchrotron radiation-induced X-ray fluorescence measurements - a case study N2 - In this work, we describe the use of artificial neural networks for the quantification of X-ray fluorescence measurements. The training data were generated using Monte Carlo simulation, which avoided the use of adapted reference materials. The extension of the available data set by means of an ANN to generate additional data was demonstrated. Particular emphasis was put on the comparability of simulated and experimental data and how the influence of deviations can be reduced. The search for the optimal hyperparameter, manual and automatic, is also described. For the presented case, we were able to train a network with a mean absolute error of 0.1 weight percent for the synthetic data and 0.7 weight percent for a set of experimental data obtained with certified reference materials. KW - Machine learning KW - BAMline KW - XRF KW - Synchrotron KW - Neural network PY - 2020 UR - https://nbn-resolving.org/urn:nbn:de:kobv:b43-519008 DO - https://doi.org/10.1088/2632-2153/abc9fb SP - 1 EP - 16 PB - IOP Publishing AN - OPUS4-51900 LA - eng AD - Bundesanstalt fuer Materialforschung und -pruefung (BAM), Berlin, Germany ER -