(214am) Computing Effective Bond Orders in Periodic Materials

While many methods are available for computing effective bond orders (EBOs) in non-periodic (e.g., molecular) systems, only a few methods are available for computing EBOs in periodic materials. The EBO between two atoms can be generically defined as the number of electrons exchanged between them. Different definitions of the exchange projectors lead to different EBO methods. This talk will describe a new method for computing EBOs in materials with any number (0, 1, 2, or 3) of periodic boundary conditions. The method's primary advantage is computational convenience: (a) it requires only the electron density and spin distribution as inputs, (b) its computational cost scales linearly with increasing unit cell size, and (c) it converges robustly. The method's accuracy is assessed by careful comparisons to systems whose bond orders are inferred from experimental measurements and general chemical properties. These assessments show the method is highly accurate. For molecular systems, results are benchmarked against established methods for computing EBOs to further verify its accuracy and reliability. Results will be presented for a variety of systems including: (a) metallic solids, (b) covalently bonded solids, (c) ionic solids, (d) solid surfaces, and (e) molecular systems. These materials include non-magnetic systems, systems with collinear magnetism, and systems with non-collinear magnetism. Of particular note, the method is shown to be accurate for metallic conductors, which other EBO methods have difficulty describing.