Development of Tensor Based Building Block Models with Density Functional Theory Calculations and Clustering Algorithms for a Quantitative Description of Molecular Orientation Using Near Edge X-Ray Absorption Fine Structure Spectroscopy of Small Molecule and Polymeric Thin Films

Abstract

Polarized Resonant Soft X-ray Scattering (pRSoXS) is uniquely sensitive to local molecular orientation regardless of crystallinity, making it a powerful tool in characterizing various types of nanostructures. Unfortunately, it is difficult to interpret the anisotropic scattering patterns due to a lack of appropriate optical models. Empirically derived building block models (BBMs) do not have information about the transition dipole moments (TDMs) associated with resonant transitions in NEXAFS and are consequently unable to generate nonuniaxial optical tensors. Density functional theory (DFT) can provide the TDMs but it generates several thousands of transitions for small molecules that prevent their usage into a reasonable optical model. In addition to this, the intensities, energies, and TDM orientations that are outputted from the DFT require refinement in order to get better correspondence with experiment. In this work, an optical model that combines angle-resolved NEXAFS (AR-NEXAFS) measurements and a clustering algorithm applied to DFT to algorithmically create a quantitatively accurate optical tensor for molecules targeted in organic electronic applications is made. First, it is shown how transition potential DFT (TP-DFT) can be used to produce qualitatively accurate simulations of NEXAFS for poly-3-hexylthiophene (P3HT) and Copper (II) Phthalocyanine (CuPc). Then, a clustering algorithm is developed that is capable of quantitatively reproducing the AR-NEXAFS of CuPc using a set of transition clusters that have all the DFT information encoded within them. Lastly, these clusters are paired with results from Mulliken population analysis to assign the chemical character of NEXAFS features. The optical tensor generated from the clustering and refinement process can then be used to aid in the modeling and fitting of resonant X-ray techniques used in structural characterization like pRSoXS and Resonant X-Ray Reflectivity (XRR).