Value of chemical bonding models for prediction of properties of new classes of materials

With the increase in number and complexity of facilities needed to solve the many problems which threaten life on earth, there is a great need for new materials with novel properties to provide the desired performance. Many multicomponent phases will be required. Just mixing elements together randomly is not practical when one considers the tremendous number of combinations of over one hundred elements. If specific properties are needed, chemical bonding models such as the Brewer-Engel Theory can be very powerful in predicting crystal structures and properties that one can obtain for given mixtures. However, one must properly characterize the type of bonding. To characterize all interactions would be very complicated, but I will show how it can be done in a simple manner. For NaCl one must recognize the electron transfer from Na to Cl and the use of an ionic bonding model to calculate the stability of the compound. If we take the simple example of the gaseous diatomic molecule NaCl, point charge interactions provide bonding energy values close to the experimentally determined values. However, there are deviations from the point charge model due to quadrupole, octapole, and other distortions of the symmetrical point charges. The additional parameters provided small values but their uncertainties added up to a substantial increase in overall uncertainty. The conclusion is to not calculate minor contributions. Compare the results of the point charge model with the experimental values and represent the difference by a simple parameter that will generally vary smoothly enough with position in the Periodic Table to serve as a parameter for the minor contributions to the bonding energy. This is used as a general procedure. Only major contributions to the energy are calculated, and minor contributions are combined to be treated by a single parameter.