7
$\begingroup$

I have some recurrent problem with my DFT calculations [functional: hseh1pbe; basis:6-31+g(d,p), object: porphyrin derivative]. During the MO setting I get:

**** Warning!!: The largest alpha MO coefficient is ***D+03

**** Warning!!: The smallest alpha delta epsilon is ***D-01

I did some first-step research over the internet: the only considerable suggestion I found was to decrease the basis since some MOs got linear-dependent. I already decreased it from 6-311++g(d,p) so I assume that it didn't help much. Moreover, simplifying the basis set would mean increase in inaccuracy, right?

Did someone meet and solve this issue before?

Improve this question
$\endgroup$
4
  • 2
    $\begingroup$ For HD & DFT calculations a near-linear-dependency in a basis is usually not a problem, so you can ignore it. For post-HF treatments this might be an issue. Details. $\endgroup$
    – Wildcat
    Commented Mar 14, 2016 at 8:07
  • $\begingroup$ Thanks for the comment. The link you gave it is what I mentioned. I still feel a bit awkward about those warnings. $\endgroup$
    – D.I. Solonenko
    Commented Mar 14, 2016 at 8:51
  • 1
    $\begingroup$ First, are you sure you actually need diffuse functions (*)? Secondly, try Ahlrichs basis sets instead of Pople ones, say, Def2SVP instead of 6-31g(d,p) or Def2TZVP instead of 6-311g(d,p). $\endgroup$
    – Wildcat
    Commented Mar 14, 2016 at 9:00
  • $\begingroup$ As for diffuse functions, no, I am not sure. I would think that it is good to include them for higher accuracy. It's nice that you mentioned about Ahlrich's basis sets: I initially started with DGTZVP (in Gaussian) but the same problem occurred so then I started to minimize the basis set. $\endgroup$
    – D.I. Solonenko
    Commented Mar 14, 2016 at 14:07

0

Reset to default