$reorder_mo not affecting beta orbitals as expected

Questions
#1

Hi everyone, I am trying to use $reorder_mo to reorder some beta orbitals as specified in following input, but I do not see any corresponding comments related to the changes in the output. Also, when I check the orbitals, it looks like no reordering has taken place. Am I misunderstanding something or doing this incorrectly?

$comment
   Ground state calculation for reference orbitals
$end

$molecule
   0 1
  C   -2.3725346   -1.3927025    0.0463733
  C   -2.7420598   -0.5856619   -1.0346678
  C   -2.7007158    0.8072476   -0.9135538
  C   -2.2904783    1.3926294    0.2886653
  C   -1.9238976    0.5854162    1.3703144
  C   -1.9654128   -0.8071512    1.2493456
  H   -2.4001757   -2.4814486   -0.0495591
  H   -3.0647167   -1.0433275   -1.9738065
  H   -2.9904943    1.4385398   -1.7580672
  F   -2.2544678    2.4813102    0.3824377
  H   -1.6024799    1.0427477    2.3096857
  H   -1.6763489   -1.4381603    2.0940326
  C    2.2904664   -1.3926121   -0.2887337
  C    1.9238658   -0.5854757   -1.3703515
  C    1.9654400    0.8070724   -1.2493161
  C    2.3725478    1.3926780   -0.0463044
  C    2.7420496    0.5857226    1.0347113
  C    2.7007237   -0.8071637    0.9135152
  H    2.2544610   -2.4812995   -0.3825468
  H    1.6024045   -1.0428320   -2.3097467
  H    1.6764283    1.4380483   -2.0939818
  H    2.4002037    2.4813971    0.0496753
  H    3.0646690    1.0434319    1.9738463
  H    2.9905508   -1.4384022    1.7580442
$end

$rem
   METHOD            b3lyp
   BASIS             def2-svp
   ECP               def2-ECP
   SCF_CONVERGENCE   7
   IQMOL_FCHK              true
$end

@@@

$comment
   Actual U-STEP calculation
$end

$molecule
   read
$end

$rem
   jobtype           opt
   METHOD            b3lyp
   BASIS             def2-svp
   ECP               def2-ECP
   SCF_GUESS         read
   STEP              true
   UNRESTRICTED      true
   SCF_CONVERGENCE   5
   MAX_SCF_CYCLES    500
   MEM_TOTAL         100000
   IQMOL_FCHK        true
$end

! default level-shift is 0.1 Hartree
! beta orbital promotion means only betas need constrained
$step
Epsilon 10
Always_Active   beta
$end

$reorder_mo
   1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 
   1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 45 46 50 44 
$end
#2

For one thing, I’m not sure that I understand your $reorder_mos section because this molecule has 46 alpha and 46 beta electrons but you’ve specified 47 entries for beta. Furthermore, the STEP sample job uses $occupied instead of $reorder_mos. Replace that section with this:

$occupied
1:46
1:43 45 46 50
$end

and I think that will give you what you want.

#3

Hi John, thanks a lot for the clarification. I used $reorder_mos to try reordering a virtual orbital as well, but I see now that does not make sense, reordering the occupied orbitals is enough. I also wanted to confirm that different Epsilon values should not affect the final energy, correct?

#4

In principle I think that’s right, assuming convergence thresholds are tight enough, and that epsilon is large enough to consistently shift the virtual levels that you don’t want out of the picture (i.e., too-small epsilon could give different results, I think).

I noticed that the STEP sample that ships with the code uses $occupied, which is what I always use for non-aufbau excited-state Kohn-Sham, but the sample in the manual does use $reorder_mos. Not sure what the difference is and don’t have time to investigate right now, try to look into it later.

#5

Thank you very much, that is very helpful. I will use $occupied and keep epsilon large enough.