Exploring and Understanding Photovoltaic Materials Using Atomistic DFT Modelling
Abstract
At the Solar Cell Technology division within the Department of Materials Science and Engineering, Uppsala University, we are doing research and development on photovoltaic thin film materials based on Cu(In,Ga)Se2 (CIGS), Cu2ZnSn(S,Se)4 (CZTS), and other emerging chalcogenides (like chalcogenide perovskites, with BaZrS3 as the key representative). Increasingly important in our work is to combine experimental studies with theory to understand the underlying physics and to help choose among materials synthesis strategies. The work started within the SSF framework project “Gradient control in Thin Film Solar Cells” (Prof. Charlotte Platzer-Björkman) in collaboration between Uppsala University and KTH Stockholm (Prof. Clas Persson). The work for the next year will primarily be associated with a research project funded by the WISE program from Knut och Alice Wallenbergs Stiftelse, although it will also include aspects from other smaller projects funded by, for example, Vetenskapsrådet (VR).
In this project, the crystalline structure, optoelectronic properties, phase stability, defect/cluster formations and diffusion barrier heights in various solar energy materials will be determined computationally. We will utilize the regular GGA, meta-GGA, and hybrid HSE06 functionals depending on the system size, property of interest, and material system. For the ground-state stability assessments, we will first analyse thousands of different, relatively small structures, generating exhaustive convex hulls and identifying the most stable atomistic arrangements. Such structure will include all known literature phases along with thousands of polytypes generated by our custom-built scripts. The majority of such calculations for the stability assessment will be performed using one 32-core node/job running in parallel to converge total energies for different structures. Then, a more accurate analysis will be needed to refine the thermodynamic stability assessment and predict measurable material properties, including Raman vibration modes, detailed electronic structures, and optical absorption coefficients. We also plan to carry out computational assessments of the entropy contributions to stability within the quasi-harmonic approximation and molecular dynamics (AIMD) simulations. In addition, point defect and alloy stability calculations will be performed using the supercell approach, in which model systems containing hundreds of atoms will be simulated. These jobs will need hundreds (up to two thousands) of cores to be completed within a reasonable time frame (several days), and the number of such calculations is expected to be large because point defects in semiconductors exhibit different ionization states and geometries depending on the Fermi level. Luckily, the parallelization implemented in the first-principles calculation package (VASP) is highly efficient, enabling rational utilization of computational resources. The obtained results will reveal the fundamental properties of the solar energy materials and explain the experimentally observed phenomena, helping to outline the optimum materials processing conditions and ultimately overcoming the limitations of the modern thin-film solar cells.
