"NAN Convergence criterion met"

Questions
#1

Hello Q-Chem community,

I am performing SCF calculations with “BASIS2”, used to set the small basis set to use in basis projection (more information in Q-Chem 4.3 User’s Manual : Dual basis sets), as:

“JOBTYPE sp
EXCHANGE HF
BASIS def2-TZVP
BASIS2 STO-3G”

Although in the small basis set the convergence criterion is met, in the large basis set the SCF reaches a point where it says that the convergence criterion is met but the result is “Not a number”:

“177 -3492.4983546993 3.56E-08 Line search: overstep
178 -nan NAN Convergence criterion met”

Do you know what is going on in this situation? Thank you in advance for your time and help.

Regards,

Antonio

#2

I you post the whole input, someone will take a look. Given the number of SCF cycles, something is clearly problematic about convergence in this case, and that’s usually the result of a poor guess (for whatever reason), so the solution is usually to try something different with the guess.

#3

I’ve also faced this problem, and I’m conducting QM/MM simulation.

(below are the lines when NaN first occured.)
197 -5135.9647964100 5.48E-05 Line search: overstep
198 -5135.9647964100 5.48E-05 Line search: overstep
199 -5135.9647964100 5.48E-05 Line search: overstep
200 nan NAN Convergence criterion met

<S^2> = -nan
SCF time: CPU 74849.57 s wall 3563.80 s
E(QM + QM/MM): nan E(MM): 162509.0349613 E(Tot): nan
SCF energy in the final basis set = nan
Total energy in the final basis set = nan

#4

The fact that E(MM) is so large and positive suggests that you may have poorly positioned atoms in the MM region (steric clashes).