The shapes and bending potentials of monomeric MgF 2 and CaF 2

Abstract

Extensive ab initio molecular orbital calculations are reported for monomeric MgF 2 and CaF 2, at both SCF and correlated levels of theory, concentrating on the shapes and bending potentials. In particular, careful attention has been paid to details of the basis sets used, which have been extended until convergence was obtained. Polarization exponents for the molecular environments have been optimized for both d- and f-type functions. They are quite different from those obtained variationally for isolated atoms or ions. They are also different at SCF and MP2 levels of theory. MP2 and QCISD theories were shown to give very similar results for MgF 2. While our final SCF and MP2 results for MgF 2 differ only marginally, we have shown that f-type functions are necessary on Mg if reliable results are to be obtained in correlated calculations. The bending potential for MgF 2 in an argon matrix is shown to be seriously perturbed by matrix effects, but the frequency reported from IR emission at high temperature is found to be reliable. A very large d basis set has been shown to be necessary for Ca; five sets are needed at the SCF level, and at least seven in MP2 calculations. The results for CaF 2 are remarkably sensitive to the particular exponent values used, even for multiple sets. Our best estimate of the bond angle in CaF 2 is 154°, but the linear transition state is only 0.7 kJ mol −1 higher in energy. The factors responsible for the non-linearity of monomeric CaF 2 are discussed, and the importance of core polarization is stressed. It is important to note that the exponents for d-type functions which maximize the polarizability of the isolated Ca 2+ ion are very similar to those optimized variationally for molecular CaF 2.