Dear Q-Chem Group,
We are studying a cluster containing N water molecules that forms an interior cavity capable of trapping one or two excess electrons. We would like to treat the water molecules using EFP, while describing only the excess electron(s) in the QM region without including any real atoms.
However, when we include only ghost atoms in the QM region, the calculation fails with the following error message:
“Only ghost atoms are found in $molecule section!”
This suggests that it may not be possible to define a QM region containing only one or two electrons without real atoms.
In both GAMESS and Gaussian, it is possible to perform calculations with only ghost atoms in the QM region. Therefore, I was wondering whether a similar setup is also possible in Q-Chem, or if there is a specific keyword or option that I may have overlooked.
Best regards,
Tae Hoon Choi
The following is a bit of a hack so you should definitely do some testing to see if this works. Haven’t tried it with EFP, but for ordinary QM-only calculation (which is also not allowed when $molecule only contains ghost atoms), you can set the nuclear charges to whatever you want, including (apparently) zero:
$molecule
0 1
He 0.0 0.0 0.0
$end
$rem
method hf
basis cc-pvdz
charge_stabilize true
$end
$nuclear_charges
1
2 0.0
$end
This functionality is described in the manual under charge stabilization,
https://manual.q-chem.com/latest/sec_CAGE.html
Q-Chem still won’t let you call the atom a ghost, but you should be able to combine this “fake” He atom with BASIS = GEN and a $basis input section to select the basis functions that you want to put on the ghost atom.
You should test this to make sure it’s working as it looks like it is. I would suggest trying to implement counterpoise correction (for something simple) in this way, and see if you get the same answer that you do with normal ghost atoms. Do this with Hartree-Fock theory because for DFT, the ghost atoms have grids that may not be right.