In general, SCF calculation involving halogen atoms that are far from neutral organic molecules give rise to electronic excited states. An example would be a fluorine atom about 10 Angstroms from a methane molecule. In this system (see data set below), the fluorine atom should have one unpaired extra alpha electron, and methane should be neutral. Instead, the SCF results in the fluorine atom having one unpaired beta electron and methane has two unpaired alpha electrons. What went wrong? The problem was traced to the initial guess for the density matrix. In the first few iterations of the SCF calculation, fluorine has a unit negative charge. The odd alpha electron goes on to the methane. As the SCF forms, a second alpha electron goes on to the methane; this results in the excited state forming.
To correct errors of this type, add PULAY to the keyword list.
This error occurs so rarely that a bug-fix has not been made.
Data set that shows this fault:
1scf CH4 plus F F -0.13005907 +1 0.05852702 +1 0.00195838 +1 0.0000 C 11.69496462 +1 -5.34383705 +1 -1.18161572 +1 -0.4405 H 10.09784740 +1 -4.44034777 +1 1.15461967 +1 -0.0002 H 12.10011696 +1 -6.26901783 +1 -0.86283618 +1 0.1468 H 10.80424362 +1 -5.32647020 +1 -1.75399119 +1 0.1468 H 12.17920911 +1 -4.43724484 +1 -0.92583745 +1 0.1470