Author Topic: No convergence with COSMO-RS  (Read 5196 times)

mwoch

  • Newbie
  • *
  • Posts: 5
  • Karma: +0/-0
No convergence with COSMO-RS
« on: December 04, 2013, 02:28:04 PM »
Hello,

I'm running ri-dft geometry optimisations of iron(II)-complexes with one or two aqua ligands (amongst others).
I obtain the best results, if I'm using the COSMO-RS model, however these calculations always tend to kind of oscillate and never converge.
Usually I'm using the jobex script with statpt but also tried using relax and optimising in cartesian coordinates.

Do you have any idea of how I could speed up optimisation?

A typical calculation looks like this (Turbomole 6.3):

def2-tzvp
tpssh
scfconv 8
denconv 1d-7
weight derivatives
grid m4
disp3
cosmo (water)
d-cosmo-rs (water)

jobex -ri -c 500 -gcart 4 -energy 8

Attached is a plot of how such an optimisation looks like (with zoom for the last 40 steps).

I'm thankful for any help!