Hello all,
I have a question about the ESP fitting to the quantum electrostatic potential module in nwchem. After my ESP runs for various systems (I work specifically ionic systems like LiH and NaF) it reports to me these kinds of numbers. For an example this is LiH at 1.4 angstrom separation.
Atom Coordinates Charge
ESP
1 Li 0.000 0.000 0.000 0.738920
2 H 0.000 0.000 0.140 -0.738920
------------
0.000000
Dipole moment 1.954898
Quadrupole moment Qxx 0.000000
Qyy 0.000000
Qzz 0.000000
RMS deviation kJ/mol 0.831466
ABS deviation % 30.520279
I simply use the default parameters since I found that increasing grid points started to give completely nonsensical charges. Should I be concerned about these numbers? Does convergence of these charges need to be checked by raising or lowering some parameter? I realize that definition of charge (mulliken, lowdin, ESP) is arbitrary in nature due to the fact that atomic charges are not observable. However, I have seen instances where Mulliken is considered to be "bad" in the context of DFT and Lowdin was a suggested alternative. Just want to make sure the ESP is correct or "ok" to use in the instances of ionic systems. Thanks.
|