Author Topic: DFT+D calculations  (Read 9392 times)

bademoor

  • Newbie
  • *
  • Posts: 4
  • Karma: +0/-0
DFT+D calculations
« on: April 07, 2008, 11:32:05 am »
Hello,

I'm currently looking into the possibility of using DFT+D calculations for zeolite systems. When using a pure GGA functional such as BLYP or PBE, the keyword $olddisp (Grimme 2004) or $disp (Grimme 2006) has to be added in the control-file to invoke the dispersion correction as mentioned in the TM-manual.

Here's my question:
If I want to use B97-D (implemented in Turbomole-5.10) as proposed by Grimme (2006), do I still need to add $disp in the control file or has this correction been added by default (without $disp in control file)?

Regards,

Bart

uwe

  • Global Moderator
  • Sr. Member
  • *****
  • Posts: 419
  • Karma: +0/-0
Re: DFT+D calculations
« Reply #1 on: April 08, 2008, 04:05:48 pm »
Hi,

$disp is always needed if you want to switch on the dispersion correction. B97-D without $disp is just B97-1 (but we do not recommend to use B97-1 for non-DFT+D calculations).

Regards,

Uwe

cml

  • Newbie
  • *
  • Posts: 3
  • Karma: +0/-0
Re: DFT+D calculations
« Reply #2 on: April 15, 2008, 09:28:35 pm »
Hello Bart,

 please be careful: in B97-D, the expansion parameters
in the expressions for exchange and correlation energy
have been refitted in the presence of the dispersion
correction (see J.Comput.Chem. 2006, p. 1787).

This means: if you do not turn on dispersion, you get
a modified version of B97-1, that will not yield the same
energies as the original functional ! 
Using B97-D without $disp therefore does not make much
sense and I strongly discourage you from doing so.

Best regards,

Christian Mück-Lichtenfeld


« Last Edit: April 15, 2008, 09:33:06 pm by cml »
Dr. Christian Mück-Lichtenfeld
Westfälische Wilhelms-Universität
Organisch-Chemisches Institut
Corrensstrasse 40
D-48149 Münster, Germany

cml@uni-muenster.de
Tel +49 251 83 33301
Fax +49 251 83 36506
http://www.uni-muenster.de/Chemie/OC/neugebauer/cml.html

bademoor

  • Newbie
  • *
  • Posts: 4
  • Karma: +0/-0
Re: DFT+D calculations
« Reply #3 on: April 18, 2008, 12:12:17 pm »
Hello Uwe and Christian,

Thank you for the very clear comments. I'm happy to use B97-D functional now properly.
I think I was just confused by the name of the functional, B97-D: the "D" in this name wrongly gave me the impression that dispersion correction is already included (which is clearly NOT the case if one doesn't add "$disp" in the control file). Actually, it would be more logic if "$disp" was automatically added in control file if selecting B97-D, since the refitting of the B97-1 functional makes its use without "$disp" not really recommendable (following the comments of Christian).

Best regards,

Bart