New Developments in Advanced X-ray Spectroscopy and Their Applications

Abstract

X-ray spectroscopy is a non-destructive, element-specific probe that gives information on electronic structure and local coordination geometry. It describes a family of X-ray spectral techniques including X-ray absorption fine structure (XAFS), X-ray emission spectroscopy (XES), High Energy Resolution Fluorescence Detected (HERFD), and non-resonant X-ray Raman Scattering (XRS), which are utilized across multiple fields of science, including chemistry, physics, and materials science. However, the full diagnostic potential of these techniques is often limited by optical constraints, particularly in spherically bent crystal spectrometers. Additionally, persistent analytical bottlenecks exist in unmixing multi-phase experimental spectra when no extensive reference libraries are available. To address the optical limitation, this dissertation presents a comprehensive theoretical and experimental development of asymmetric operation for spherically bent crystal analyzers (SBCAs) under Rowland geometry. By accessing different crystal planes through hkl hopping, a single SBCA can cover a broad energy range that includes the K-edge absorption and emission spectra of 3d transition metals. Furthermore, through extensive geometric ray-tracing simulations and experiments, we show that operating near the Johann Normal Alignment (JNA) condition systematically suppresses or eliminates Johann error, the major source of energy-resolution compromise in a Rowland-geometry SBCA spectrometer. The new operation method of a Rowland spectrometer preserves high energy resolution and large collection solid angles while enhancing sample clearance for specialized environments in XAFS, XES, HERFD, and XRS imaging applications. Complementing these hardware advances, this work introduces "spectral harvesting", a novel geometric data analysis framework for non-reference-based spectral decomposition. Operating within latent space generated with the aid of Principal Component Analysis (PCA), the algorithm enforces a suite of physics- and chemistry-motivated constraints to systematically filter unphysical latent space regions and establish an Area of Feasible Solutions (AFS). Pure phase candidates are selected through subsequent maximum-volume simplex fitting within the reduced AFS, effectively reducing sensitivity to rotational ambiguity that is commonly faced by standard matrix-based decomposition algorithms such as Multivariate Curve Resolution-Alternating Least Squares (MCR-ALS). Together, the instrumental and analytical developments established in this dissertation significantly expand the operational scope, spectral resolution, and spectral decomposition accuracy of modern X-ray spectroscopy, providing robust methodologies for both laboratory bench-top instruments and synchrotron beamline spectrometers.