I'm attempting to optimize n-alkane dimers using MP2/DFT in Gaussian09. However the first one or two frequencies are always negative. Can these modes be ignored or do I need to do a tighter optimization?

  • 3
    $\begingroup$ What are the magnitudes of the imaginary frequencies? And, do you want to calculate the Gibbs free energy, or are you interested in other quantities instead? The answer will heavily depend on these two factors. See mattermodeling.stackexchange.com/questions/3942/… $\endgroup$
    – wzkchem5
    Apr 11, 2022 at 8:18
  • $\begingroup$ The magnitude is about -28 wavenumber... I am looking for optimized geometries for further studies. thanks for your comment. $\endgroup$ Apr 11, 2022 at 8:53
  • 2
    $\begingroup$ @Wesley How does BSSE matter for this question? $\endgroup$ Apr 12, 2022 at 1:26
  • 1
    $\begingroup$ You don't mention the size of the alkane. Nevertheless, as @wzkchem5 indicated, for really accurate thermochemistry, you'll want to work on eliminating those imaginary frequencies. IMHO, that's small enough you can ignore for many purposes. (For DFT, it can be due to grid issues, for example.) $\endgroup$ Apr 12, 2022 at 3:54
  • 2
    $\begingroup$ The issues were solved now by taking care of many factors, i. symmetry of the molecules (geometry), ii. DFT-D3 dispersion iii. ultrafine grid iv. read the DFT-D3 Hessians to perform MP2 optimization/Freq. Finally, the negative frequencies have disappeared. Thanks for your comments, however, let the good thoughts keep come. $\endgroup$ Apr 13, 2022 at 2:53

1 Answer 1


Summarizing the OP's comment that describes how they managed to remove negative (imaginary) frequencies for their system:

  1. Ensure correct symmetry of the molecular geometry
  2. Include D3 dispersion for the DFT calculation
  3. Use an ultrafine integration grid for the SCF
  4. Use the DFT-D3 Hessians as a guess to perform the MP2 optimization/Freq.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .