The electronic structure of tetrahedral III–V compounds: The ground state of boron nitride

The bond‐orbital theory of III–V compounds, previously described by Coulson, Redei and Stocker, is used to calculate the effective atomic charges and the binding energy per bond in boron nitride. The theory is reformulated in a manner which is convenient for performing both ab initio and semiempirical calculations. Two different choices for the atomic‐orbital exponents are considered and, in both cases, the results obtained from the ab initio method are at variance with the earlier calculations in predicting an electronic charge displacement from nitrogen to boron. The magnitude of the effective charges is found to vary according to the method of partitioning the overlap charge between the nitrogen and boron atoms. The use of orthogonalized Slater 2 s functions is also examined. The semiempirical calculations are performed with an explicit inclusion of the Madelung energy from the outset. The ionicity in the bond is shown to be determined by the competition between the difference in orbital electronegativities and the difference in Madelung potential across the ends of the bond. Unfortunately, the semiempirical theory breaks down because the energy per bond passes through a maximum at the optimum value of the polarity parameter. The reasons for this behaviour are examined and discussed.