Abstract by Kiersten Ernstrom
Assembling a database for broad-range bond-valence vs. length relationships
Our group is developing models based on the concept of bond valence to simulate atomic- and molecular-scale chemical structures and reactions. The bond valence method empirically relates bond order to bond length, and has proven remarkably accurate for predicting plausible crystal structures. The first step in our modeling process is to calibrate a bond valence vs. bond length curves for each bond type considered. For this, the standard method is to take known stable crystal structures, assume the bond valences incident to each atom sum to its atomic valence, and fit an exponential equation to the data. The simulations we plan will require accurate bond valence curves over a wide range of bond lengths, but crystal structures exhibit only a narrow range of lengths for any given bond type. Therefore, I am gathering two types of data for the calibrations: thousands of crystal structures from the Inorganic Crystal Structure Database, and both experimental and calculated structures of molecules.