I am trying to optimize a transition metal complex with a few (3-4) solvent (H2O) molecules around it. What are the usual input control that will not give me a RMS Displacement of ~ 0.00000. My initial guess is taken from a snap shot of an MD simulation.
What is the problem with RMS displacement of 0? This usually means that you are close to the minimum, which is probably what you're looking for.
If this is DFT, there's a usual set of commands which are necessary and/or helpful in converging the geometry and SCF:
opt(tight)
grid(ultrafine)
opt or opt=z-matrix or opt=cartesian
scf or scf(qc) or scf(xqc)
Try different combinations of these.
i mean that at the first iteration, it cannot proceed because the initial guess is way way far from the equilibrium. Thank you.
Ah, OK.
One keyword that may help here is opt=cartesian, since a multi-molecule system is likely to have some degrees close to 180, which always pose problems for gaussian. Othwerwise, you have to change the starting geometry or (better), preoptimize it in a low-level method like semiempirical PM6/PM7 (use free mopac, if gaussian does not have parameters for your transition metal) or something like HF/3-21G*/LANL2DZ or B3LYP/3-21G*/LANL2DZ or even UFF (gaussian or free avogadro), perhaps with freezing some coordinates so the system stays together.
You do not mention the spin of your system. For unrestricted wavefunction definitely try Guess=(Mix,Only) and check the initial guess for sanity. As for opt=cartesian, use with care and to start up only. Optimization in redundant internals is nearly always much faster.
- Badges
- Science topic