Can I use split basis set system in Gaussian09 for Gold with other atoms than gold modelled at PM3 level of theory? I have a system which has C, N, O, H, and Au. I want to model C,N,H and O using PM3 semi-empirical model and Au using LANL2DZ basis sets. Is it possible? If yes, what will be the format? If I use psuedo=read in command line, it gives error with PM3, but works fine at B3LYP level.
This is impossible. To combine two different theoretical levels in Gaussian, you have to use ONIOM feature. However, even if you use ONIOM(B3LYP:PM3), namely using B3LYP as high level while PM3 as low level, according to the algorithm of ONIOM, as an intermediate step, PM3 will still be employed to calculate Au, which is a high layer atom. Since PM3 was not parameterized for Au, the task must be failed.
If you want to save computational cost, you may consider to use relatively small basis set for C, N, H in a proper way.
This is impossible. To combine two different theoretical levels in Gaussian, you have to use ONIOM feature. However, even if you use ONIOM(B3LYP:PM3), namely using B3LYP as high level while PM3 as low level, according to the algorithm of ONIOM, as an intermediate step, PM3 will still be employed to calculate Au, which is a high layer atom. Since PM3 was not parameterized for Au, the task must be failed.
If you want to save computational cost, you may consider to use relatively small basis set for C, N, H in a proper way.
Not only it is technically impossible, but you probably want to model the interaction between the gold ion and the ligand atoms. So from a chemist’s point of view this would not make much sense.
You could anyway use different basis sets for different type of atoms, adopting a more complex bs for gold (lanl2dz or lanl08 would be ok) and lighter bs (def2svp for example) for the rest of the molecule, provided you do not have any other heavy atoms.
Hope this may help.
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Atomic, Molecular and Optical Physics
- Atoms