Master of Science (MS)
Physics and Astronomy
First Committee Member
Second Committee Member
Third Committee Member
Fourth Committee Member
Number of Pages
This thesis presents an original open-source Python package called PyXtal (pronounced "pi-crystal") that generates random symmetric crystal structures for use in crystal structure prediction (CSP). The primary advantage of PyXtal over existing structure generation tools is its unique symmetrization method. For molecular structures, PyXtal uses an original algorithm to determine the compatibility of molecular point group symmetry with Wyckoff site symmetry. This allows the molecules in generated structures to occupy special Wyckoff positions without breaking the structure's symmetry. This is a new feature which increases the space of search-able structures and in turn improves CSP performance.
It is shown that using already-symmetric initial structures results in a higher probability of finding the lowest-energy structure. Ultimately, this lowers the computational time needed for CSP. Structures can be generated for any symmetry group of 0, 1, 2, or 3 dimensions of periodicity. Either atoms or rigid molecules may be used as building blocks. The generated structures can be optimized with VASP, LAMMPS, or other computational tools. Additional options are provided for the lattice and inter-atomic distance constraints. Results for carbon and silicon crystals, water ice crystals, and molybdenum clusters are presented as usage examples.
Crystallography; Energy; Materials; Python; Space group; Symmetry
Condensed Matter Physics | Other Physics | Physics
Fredericks, Scott William, "Applied Symmetry for Crystal Structure Prediction" (2019). UNLV Theses, Dissertations, Professional Papers, and Capstones. 3723.