2
$\begingroup$

I'm trying to optimize a structure that contains Pd atom with the gen function on gaussian16

    %chk=pd3cgen.chk
    # opt freq b3lyp/gen pseudo=read scrf=(solvent=ch2cl2,read,pcm)

        Zmatrix coordinates
               .....

pd 0
 LANL2DZ
 ****
 S 0
 6-31G(d)
 ****
 O 0
 6-31G(d)
 ****
 Br 0
 6-31G(d)
 ****
 C 0
 6-31G(d)
 ****
 H 0
 6-31G(d)
 ****
 N 0
 6-31G(d)
 ****
 Cl 0
 6-31G(d)
 ****
 blank space
  radii=uaks

I have done before the calculation with only lanl2dz alone it worked perfectly but when I used a mixing basis set(this case) I got this error message:

Maximum number of bonds 12 exceeded for atom  13.
 Error termination via Lnk1e 
$\endgroup$
4
  • 1
    $\begingroup$ You may be able to avoid this error by manually entering the connectivity with geom=connectivity. $\endgroup$
    – Tyberius
    Jul 21, 2023 at 18:00
  • $\begingroup$ @Tyberius it still displays the same problem $\endgroup$
    – diamond999
    Jul 21, 2023 at 19:19
  • $\begingroup$ This message surely does not have anything to do with a basis set and most certainly points to some issue with Z-matrix. Try to replace Z-matrix with plain xyz coordinates to check if that's the case. $\endgroup$ Jul 22, 2023 at 23:01
  • 1
    $\begingroup$ Ah, I think the problem is that you need to specify both pseudo-potential and basis set for Pd, please see the example in the answer $\endgroup$ Jul 22, 2023 at 23:04

1 Answer 1

2
$\begingroup$

I think the problem is that for atoms with pseudo-potential you need to define both localized and pseudo-potential parts. Here's example for your input:

END-OF-GEOM-BLOCK

H C N O S Cl Br 0
6-31G(d)
****
Pd 0
LANL2DZ
****

Pd 0
LANL2DZ

radii=uaks
$\endgroup$
3
  • 1
    $\begingroup$ thank you so much that's what I actually did this morning and it worked but I still don't understand how it worked , is it because of the frozen core electrons? $\endgroup$
    – diamond999
    Jul 22, 2023 at 23:41
  • $\begingroup$ exactly, and I'm not sure why is that an option since basis set and pseudo-potential are usually optimized together forming one "pseudopotential". Seems like it's some kind of legacy of quantum chemistry 30 years ago :-) $\endgroup$ Jul 22, 2023 at 23:53
  • $\begingroup$ although, when using only lanl2dz without the gen keyword It seems like I got better results when compared to experimental + the computational cost is cheaper . is it ok to just keep a one basis? $\endgroup$
    – diamond999
    Jul 23, 2023 at 8:32

You must log in to answer this question.

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