Implementation of Bayesian optimization on grazing-exit XANES: applications on compositionally complex alloys

Abstract

Compositionally complex alloys (CCAs) are a new class of alloys containing at least 5 elements with concentrations between 5 and 35 atomic percent. Due to their adjustable composition, which enables modifications of mechanical properties and their stability at high temperatures, CCAs have been the focus of various studies [1,2]. Especially the corrosion behavior of CCAs has been a wide research interest. However, there are only few studies that deals with the degradation process on such materials, which is highly relevant for the safety aspect for future component design. To thoroughly investigate the corrosion processes and to determine oxidation states of metal components within the reaction products, we need special analytical tools. Grazing exit X-ray fluorescence (GEXRF) offers a non-destructive way to collect this information and it as a useful method to investigate how CCAs behave in corrosive environments. With GEXRF enhance the detection of fluorescence signal in the upper few nanometers. [3]. When compared to a conventional CCD-based camera, the advantage and most important feature of the detector system (Color X-Ray Camera (CXC)) is that each pixel is an energy sensitive detector. This position and area sensitive detector, with 264x264 pixel detector area, provides information regarding the signal emitted from the sample as a function of the emission angle and thus allows depth-sensitive analysis. Furthermore, the data collected from samples of an incidence energy which can be controlled with a resolution of 0.5 eV provides XANES data to determine oxidation states. We address the feasibility of our setup and provide a new optimization procedure (Bayesian Optimization and Gaussian Regression) to decrease measuring time. The results settle on a conceptual study on a reference sample (Cr-Oxide layer (300nm) on Cr layer (500nm) on Si wafer) and on a medium entropy alloy CrCoNi (Cr-Oxide (>1.5μm) layer on CrCoNi substrate). [1] [2] [3] Figure 1 The XANES of the layers (reference sample). Orange represents the Cr-oxide layer and blue is Cr layer. R. Kozak, A. Sologubenko,W. Steurer (2015). Zeitschrift für Kristallographie. 230. M.Tsai & J. Yeh (2014). Materials Research Letters. 2:3. 107-123. Y. Kayser, J. Szlachetko, D. Banaś, W. Cao, J.-Cl. Dousse, J. Hoszowska, A. Kubala-Kukuś, M. Pajek (2013). Spectrochimica Acta Part B: Atomic Spectroscopy. 88.