For a fundamental understanding of the interaction of electromagnetic radiation and molecular materials, experimental measurements are to be combined with theoretical models. With this combination, materials can be characterized in terms of composition, structure, time-resolved chemical reactions, and other properties. This licentiate thesis deals with the development and evaluation of a theoretical method by which X-ray absorption spectra can be interpreted and predicted.
In X-ray absorption spectroscopy the photon energy is tuned such that core electrons are targeted and excited to bound states. Such core excitations exhibit strong relaxation eects, making theoretical considerations of the processes especially challenging. In order to meet these challenges, a damped formalism of the coupled cluster (CC) linear response function has been developed, and the performance of this approach evaluated. Amongst the quantum chemical methods available, CC stands out as perhaps the most accurate, with a systematic manner by which the correct physical description can be approached. Coupled with response theory, we thus have a reliable theoretical method in which relaxation eects are addressed by means of an accurate treatment of electron correlation.
By use of the hierarchy of CC approximations (CCS, CC2, CCSD, CCSDR(3)), it has been shown that the relaxation eects are accounted for by the inclusion of double and triple excitations in the CC excitation manifold. The performance of the methods for K-edge NEXAFS spectra for water, neon, carbon monoxide, ammonia, acetone, and a number of uorine-substituted ethenes has been investigated, and we observe relaxation eects amounting to 7–21 eV. The discrepancy in absolute energy for the most accurate calculations as compared to experiments are reported as 0.4–1.5 eV, and the means by which this can be decreased further are discussed. For relative energies, it has been demonstrated that CCSD yields excellent spectral features, while CC2 yields good agreement to experiments only for the most intense features. Comparisons have also been made to the more computationally viable method of density functional theory, for which spectral features are in excellent agreement with experiment.
Linköping: Linköping University Electronic Press, 2013. , 55 p.
2013-12-12, Nobel, B-huset, Campus Valla, Linköpings universitet, Linköping, 10:15 (English)
Norman, Patrick, Dr.