Author Topic: NumForce negative frequency  (Read 8224 times)

sv

  • Newbie
  • *
  • Posts: 6
  • Karma: +0/-0
NumForce negative frequency
« on: June 06, 2007, 10:16:09 PM »
Hi,
I calculated a stationary point on the excited state surface by using TD-B3LYP.
But when I calculated frequencies at the excited state surface by using NumForce, I got a negative frequency.
How should I get rid of it? Any comments.

my molecule has c2v symmetry in the ground state and excited state is also optimized in c2v symmetry.

antti_karttunen

  • Sr. Member
  • ****
  • Posts: 227
  • Karma: +1/-0
Re: NumForce negative frequency
« Reply #1 on: June 07, 2007, 07:37:31 AM »
Hi,

if the negative frequency is small, it might be enough if you redo the optimization with tighter criteria. For NumForce it is often necessary to use at least jobex -gcart 4, often something like jobex -gcart 5 -energy 6 is required. Be sure to use high enough $scfconv, too (I suggest 8 ).

If more strict optimization won't help, it might be necessary to lower the symmetry. Use "screwer" tool to distort the structure towards the negative frequency. You will get new coordinates in $newcoord group of control file, redo the optimization using them. It can also be helpful to visualize the negative frequency with TmoleX or some other software.

sv

  • Newbie
  • *
  • Posts: 6
  • Karma: +0/-0
Re: NumForce negative frequency
« Reply #2 on: June 07, 2007, 03:52:28 PM »
Thanks antti,
thant helps, I used only $scfconv 7,
so I will increase it and optimize again.

SV