Our BeBe distances turn out to be very slightly smaller (by about 0.01 Å) than the already small values found by Zhao et al. .[13] We notice that the BeBe separations in the D 3h[NgBeH3BeNg]+ cations areca . 0.04 Å shorter than those found for the rhombic Be2O2 cluster.[8-9, 11] We are not aware of any previously reported geometries for the magnesium systems. We find that the MgMg separation in the D 3h[HeMgH3MgHe]+ cation is shorter byca . 0.03 Å than was found for the rhombic Mg2O2 cluster.[9, 11]
The total energies from calculations at various levels of theory are reported in Table S1 in the Supporting Information. For each system, the energy difference between the SCGVB(6) and the corresponding CASSCF(6,6) descriptions was very small (ca . 3 kJ/mol); the SCGVB(6) wavefunctions for the various Be and Mg systems incorporate more than 96.8% and 99.4%, respectively, of the non-dynamical correlation energy that is retrieved with the CASSCF(6,6) calculations. Increasing the number of active orbitals to 11, as in CASSCF(6,11), retrieved further energy lowerings in each case of ca . 120 kJ/mol. Much larger energy improvements are of course associated with the dynamical correlation incorporated in the all-electron CCSD(T) calculations. It remains to be seen, however, whether our various modes of analysis reveal any significant differences in the nature of the bonding when we switch to different levels of theory.