Author Topic: Using Egrad to calculate ed.plt density difference map for small compound  (Read 6844 times)

biophys2012

  • Newbie
  • *
  • Posts: 2
  • Karma: +0/-0
Hello!
I'm new to Turbomole, having been a Gaussian user for the last couple years. However,
I need to calculate a density difference map ground -> excited state for a neutral small
molecule (36 atoms) and it seems that Turbomole is the only program that does it. 

I have successfully calculated a td.plt map for the ground state with TD DFT using B3LYP
functional & 6-311++G** basis set. The command I used for this was:

jobex  -level scf  -ri  -c 50 -energy 6 -gcart 3   

and in the control file I had the $pointval keyword that several Turbomole users said was
crucial for the map generation. But this only generated the td.plt file.  Next I tried several
ways to generate the ed.plt file, by getting a TD DFT excited state calculation and then
use command:

egrad -proper > egrad.out

But getting the excited state calculation (jobex -ex) is failing.  The geometry optimization
is not converging and I"m getting errors:

 MODTRACE: no modules on stack
 specification incorrect
 egrad ended abnormally
error in gradient step (1)

I would really appreciate help in calculating the ed.plt file!!!

uwe

  • Global Moderator
  • Sr. Member
  • *****
  • Posts: 486
  • Karma: +0/-0
Re: Using Egrad to calculate ed.plt density difference map for small compound
« Reply #1 on: December 05, 2011, 05:33:58 PM »
Hi,

if you are new to Turbomole: Note that you can write an email to the support (turbomole@cosmologic.de) if something does not work.

Now to your question:

Did you add $scfinstab and $soes in the control file?

The error message

Quote
MODTRACE: no modules on stack
 specification incorrect
 egrad ended abnormally
error in gradient step (1)

looks like egrad does not know what to do. You have to set an option for $scfinstab, namely rpas for TDDFT singlet, rpat for triplet, etc. (see documentation).

Regards,

Uwe

biophys2012

  • Newbie
  • *
  • Posts: 2
  • Karma: +0/-0
Re: Using Egrad to calculate ed.plt density difference map for small compound
« Reply #2 on: December 09, 2011, 06:22:34 PM »
Hi,

Thank you for your response! I was able to calculate the ed.plt map and found that the
problems came from egrad (turbomole-6.3.1) as follows:

1) Egrad implementation does not allow COSMO calculations for implicit solvent effect
2) Egrad does not work with RI options, so those keywords have to be removed if
    RI is used in SCF optimization
3) After a ground state optimization, Egrad needs the following in the control file
    (for a first excited state & ground state difference, c1 molecule symmetry)
$exopt 1
$pointval dens
$scfinstab rpas
$soes
a 1