Optimization for CH4+ can not converge

[Eric]

Hi, This is my input file

start CH4_ion

 title "CH4_ion in aug-cc-pvdz basis set " 

 charge 1.0

geometry units angstrom

C 0.0000 0.0000 0.0000

H 0.6276 0.6276 0.6276

H 0.6276 -0.6276 -0.6276

H -0.6276 0.6276 -0.6276

H -0.6276 -0.6276 0.6276

 end

basis 

 * library aug-cc-pvdz 

 end 

dft 

 odft

 MULT 2 

  XC slater vwn_5 

 decomp 

end 

task dft optimize 

 task dft frequencies

May I ask is there anything wrong with my input file?

[Edoardo Aprà]

Two changes to the original input

1. Since the output of the original input was showing symmetry breaking issues, I have lowered the symmetry from Td (detected by automsym) to C2
2. Lowered the level shifter default value from 0.5 to 0.25 https://nwchemgit.github.io/Density-Functional-Theory-for-Molecules.html#convergence-scf-convergence-control
   

start CH4_ion
title "CH4_ion in aug-cc-pvdz basis set "
charge 1.0
geometry units angstrom

 symmetry c2

 C     0.0000    0.0000    0.0000
 H1     0.6276    0.6276    0.6276
 H2     0.6276    -0.6276    -0.6276
 H2    -0.6276    0.6276    -0.6276
 H1    -0.6276    -0.6276    0.6276

end
basis
 * library aug-cc-pvdz
end
dft
 odft
 MULT 2
 XC slater vwn_5

 convergence lshift 0.25
 decomp
 maxiter 99

[Eric]

Thank you so much!

[Eric]

I am calculating the ionization energy for a lot of molecules. By following your suggestion about lowering the symmetry, I successfully make systems like CH4, NH3, C2H2 converged. But I failed on a bunch of diatomic systems, like OH, OH+, HF+, PH, PH+, SH, SH+, HCl+, O2, O2+, S2, S2+, Cl2+, ClF+. (obviously, I can not lower the symmetry this time)

Most of them are failed due to   

"dft optimize failed:Received an Error in Communication".

Only PH has trouble on energy converge

hess_anal: energy failure:Received an Error in Communication

I attached the i/o files of OH+, PH for example. 

Thanks,

Eric

[Edoardo Aprà]

Same symmetry problem as earlier. You can clearly see from the following lines in the output file

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
 Warning - the HOMO and LUMO are degenerate and you are using symmetry.
 This can lead to non-variational energies and poor convergence.
 Modify the initial guess, or use an open-shell wavefunction, or turn
 off symmetry.
 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

The difference for these two new cases is that the geometry change does not help.

The only solution is to add the two line "sym off" and "adapt off" in the dft input section

start PH
 title "PH in aug-cc-pvdz basis set "
geometry units angstorm
H       0.0000  0.0000  0.0000
P       0.0000  0.0000  1.4223

end
basis
 * library aug-cc-pvdz
end
dft

 sym off
 adapt off
 XC slater vwn_5
end
task dft optimize