A MRCI PS AND CASSCF STUDY OF THE GROUND-STATE MGO DISSOCIATION-ENERGY - COMMENT

We demonstrate that the ''systematic'' procedure used by Garcia Cuesta, Sanchez de Meras, and Nebot Gil to select a CASSCF active space for MgO is flawed by local minima in the optimization. We show that their correlation treatment underestimates D(e) by 0.2 eV. In addition, the small basis set employed resulted in at least another 0.6 eV underestimation of the D(e). Thus their computed D(e) value is in error by at least 0.8 eV. Our calculations demonstrate that D(e) (to Mg 1S and O 3P) is larger than 2.70 eV and strongly support our previous estimate of 2.80 +/- 0.1 eV. Hence, this value is not in error by 1 eV as asserted by Nebot Gil and co-workers.