Two Molecule System Doesn't Optimize


Just Got Here
I am optimizing a system of two molecules, and each molecule has 70 atoms total of C, H, and F. When I run a DFT geometry optimization using m06-L and 6-31G** or 3-21G, the two molecules optimize and change shape due to interactions but do not change their positions relative to each other. For example, in 31 optimization cycles using 3-21G, the centers of the molecules move by less than 0.01 Angstroms relative to each other, even for rather poor initial guesses. Increasing step size using TRUST has no effect. When I run this optimization in Gaussian and increase the step size, the two molecules move up to 1.8 angstroms in relation to each other in the same number of optimization steps, and this greatly reduces the interaction energy between the two molecules. Is there a way in NWChem to cause the molecules to move more quickly in relation to each other?

Here is the command section of my code:

driver
maxiter 300
TRUST 1000
end

set int:txs:limxmem 20077924

dft
xc m06-L
grid xfine
iterations 300
end

task dft optimize

Forum Regular
Hi Lkbalhorn,

I am a bit confused as I am not sure what this calculation means. With C, H, and F atoms there is going to be some polarization of the molecules and that will create some inter molecule interaction. However, beyond that DFT does not really provide a sensible interaction potential for non-bonded molecules (except hydrogen bonding). So I am not sure what the molecules are supposed to do.

Huub


Forum >> NWChem's corner >> General Topics