Change in the bulk modulus at the $B1-B2\mathrm{}$ phase transition

Ab initio Hartree-Fock calculations for NaCl are used to derive the changes across the $B1-B2\mathrm{}$ phase transition in both the bulk modulus $\left(\Delta K_T\right)$ and the transverse infrared frequency $\left(\Delta {\nu }_{\mathrm{TO}}\right).\mathrm{}$ $\Delta K_T$ is found to be positive but very small, and $\Delta {\nu }_{\mathrm{TO}}$ negative and relatively large (≈20%). In contrast, using a relationship based on lattice dynamics with rigid ions and short-range repulsion only between nearest neighbors, Hofmeister [Phys. Rev. B 56, 5835 (1997)] has recently deduced from the observed drop in ${\nu }_{\mathrm{TO}},$ itself in good agreement with the Hartree-Fock calculations, that there should be a marked drop in $K_T.$ We have, therefore, also calculated $\Delta K_T$ and $\Delta {\nu }_{\mathrm{TO}}$ for NaCl using lattice dynamics, with a shell model that incorporates short-range potentials between both first and second neighbors, obtaining good agreement with the Hartree-Fock calculations. This calls into question the use at high pressures of a semiempirical relation between infrared frequencies and $K_T,$ based on a model which includes only nearest-neighbor short-range interactions. Similar calculations for CaO give a much larger increase in the bulk modulus across the phase boundary, still accompanying a dramatic drop in ${\nu }_{\mathrm{TO}}.$