Abstract
Computer-aided data acquisition, analysis and interpretation have been effectively employed in a number of facets of research recently days generating favorable outcomes. One of the subsets of this technology, the image processing, involves use of images for data extraction and can be implemented for post processing of material characterization results to better understand and predict material features, properties, and behaviors at multiple scales. The current study introduces a unique image processing method based on Python programming language for in-depth analysis of energy dispersive spectroscopy (EDS) results. The method involves an automatic framework developed to use net intensity data from EDS elemental maps and process it to analyze microstructural characteristics such as particle and agglomerate size distribution along with computation of average area occupied by each of the different components and their overlap contained within a single map. The framework developed in this study is implemented to examine the interaction of Cerium (Ce) and Palladium (Pd) particles in membrane electrode assembly (MEA) of an Anion-Exchange Membrane Fuel Cell (AEMFC) and investigate if this approach can be effectively utilized to predict the electrochemical behavior of the fuel cell. The study also discusses how several parameters such as map resolution, map and image filters, brightness, and level of thresholding can influence the computed results and contribute to errors. A detailed evaluation of this method along with its advantages and limitations is also presented. The study concludes that the image processing framework developed in this study is a quick, easy, and reliable method for automatic extraction of underlying data, which is otherwise challenging to compute using other conventional analysis techniques. This method can also be further optimized and tailored to extract more information, which can be valuable for a variety of applications within the clean energy research.