The procedure for using capped bonds (Cb) is as follows:
A Cb behaves like a monovalent atom, but always has a zero charge. In other words, a Cb has a core charge of +1, and always has an electron population of 1.0. Cbs have one orbital, and so can be regarded as being hydrogen-like.
Capped bonds are different from hydrogen atoms, however, in that they have a large β-value. The β value is used in the calculation of the one-electron two-center integral. Because the β value is so large, the difference in electronegativity of the Cb and the atom it is attached to, A, becomes negligible. Therefore the bonding M.O. consists of (1/2)1/2(Cb+A). From this, it follows that the bonding M.O. contributes 1.0 electrons to the Cb.
In addition to the bonding M.O., there is an antibonding M.O. This M.O. is of form (1/2)1/2(Cb-A), and is of very high energy.
To prevent capped bonds from forming bonds to all nearby atoms, overlaps from capped bonds to all atoms further away than 1.8Å are set to zero. Because of this, the coupling between capped bonds and all atoms, other than A, is zero. Of course, A can interact with other atoms, once it has satisfied the demands of the attached capped bond.
Because of the huge β-value, the energy of a Cb-atom bond is very large. To prevent this interfering with the calculation, when the electronic energy is calculated, contributions from Cb are ignored. For this reason it is important that the bond-length for the Cb should not be optimized.
The M.O. energy levels due to Cb-type bonds are also enormous. Before the M.O. energy levels are printed in the normal output, energy levels arising from Cb-atom bonds are first set to zero.
The electronic behavior of capped bonds can easily be studied by use of
1SCF DENSITY
VECTORS.