Different bend frequencies for a D_{infty h} tri-atomic

Questions
#1

For a linear triatomic of the type O-X-O in this calculation:

$molecule
  1  1
  Eu
  O  1 rXO1
  O  1 rXO2  2 aOXO

  rXO1 =   1.6615601209
  rXO2 =   1.6615601209
  aOXO = 180.0000000000
$end

$rem
  JOBTYPE                                  FREQ
  METHOD                                  wB97M-V
  BASIS                                       def2-SVP
  SYMMETRY                             TRUE
  SYM_IGNORE                         FALSE
  GEN_SCFMAN                        TRUE
  SCF_ALGORITHM                   GDM
  SCF_CONVERGENCE            9
  MAX_SCF_CYCLES                400
  MEM_TOTAL                            8000
  MEM_STATIC                           250
  INTERNAL_STABILITY            TRUE
  INTERNAL_STABILITY_ITER  250
  VIBMAN_PRINT                       6
  THRESH                                  14
  MOLDEN_FORMAT                TRUE
$end

I’m getting bend frequencies different by about 15 cm-1. There’s no spatial symmetry-breaking I can perceive when I look at the orbitals and symmetry is not the culprit here (calc. without symmetry gives the same result). Do you have any ideas as to why the bend frequencies (which ought to be degenerate, right?) are different?

 **********************************************************************
 **                                                                  **
 **                       VIBRATIONAL ANALYSIS                       **
 **                       --------------------                       **
 **                                                                  **
 **        VIBRATIONAL FREQUENCIES (CM**-1) AND NORMAL MODES         **
 **     FORCE CONSTANTS (mDYN/ANGSTROM) AND REDUCED MASSES (AMU)     **
 **                  INFRARED INTENSITIES (KM/MOL)                   **
 **                                                                  **
 **********************************************************************


 Mode:                 1                      2                      3
 Frequency:       353.69                 367.63                 883.94
 Force Cnst:      1.3950                 1.5071                 7.3634
 Red. Mass:      18.9268                18.9268                15.9949
 IR Active:          YES                    YES                    YES
 IR Intens:       33.201                 33.177                  0.000
 Raman Active:       YES                    YES                    YES
               X      Y      Z        X      Y      Z        X      Y      Z
 Eu        -0.103 -0.103 -0.001   -0.103  0.103  0.000   -0.000 -0.000 -0.000
 O          0.495  0.495  0.007    0.495 -0.495  0.000    0.000  0.000  0.707
 O          0.495  0.495  0.007    0.495 -0.495 -0.000    0.000  0.000 -0.707
 TransDip  -0.130 -0.130 -0.006   -0.130  0.130 -0.000   -0.000 -0.000 -0.000

 Mode:                 4
 Frequency:      1079.15
 Force Cnst:     12.9865
 Red. Mass:      18.9268 
 IR Active:          YES
 IR Intens:      203.838
 Raman Active:       YES
               X      Y      Z        X      Y      Z        X      Y      Z
 Eu         0.001  0.001 -0.146
 O         -0.005 -0.005  0.699  
 O         -0.005 -0.005  0.699
 TransDip   0.001  0.001 -0.457
#2

It’s a grid issue. Use EML(99,590) to recover the degeneracy. (XC_GRID = 000099000590)

 Mode:                 1                      2                      3
 Frequency:       346.93                 346.93                 841.68
 Force Cnst:      1.3422                 1.3422                 6.6761
1 Like
#3

Actually SG-3 grid (XC_GRID = 3) is dense enough to recover the degeneracy, and will be a bit faster than EML(99,590).

 Mode:                 1                      2                      3
 Frequency:       347.38                 347.38                 842.43
 Force Cnst:      1.3457                 1.3457                 6.6880
#4

Thank you John. Now that I know it’s a grid issue I’m not too worried. However, I am curious that even with the denser (99, 590) grid, I still see a small difference a little over 3 cm-1. I am now using

$molecule
  1  1
  Eu
  O  1 rXO1
  O  1 rXO2  2 aOXO

  rXO1 =   1.6615581344
  rXO2 =   1.6615581344
  aOXO = 180.0000000000
$end

$rem
  METHOD                   wB97M-V
  BASIS                    def2-SVP
  SYMMETRY                 FALSE
  SYM_IGNORE               TRUE
  GEN_SCFMAN               TRUE
  SCF_ALGORITHM            GDM
  SCF_CONVERGENCE          9
  XC_GRID                  000099000590
  MAX_SCF_CYCLES           400
  MEM_TOTAL                8000
  MEM_STATIC               250
  INTERNAL_STABILITY       TRUE
  INTERNAL_STABILITY_ITER  250
  THRESH                   14
$end

@@@

$molecule
  read
$end

$rem
  JOBTYPE         FREQ
  METHOD          wB97M-V
  BASIS           def2-SVP
  SYMMETRY        FALSE
  SYM_IGNORE      TRUE
  GEN_SCFMAN      TRUE
  SCF_GUESS       READ
  SCF_ALGORITHM   GDM
  SCF_CONVERGENCE 9
  XC_GRID         000099000590
  MAX_SCF_CYCLES  400
  MEM_TOTAL       8000
  MEM_STATIC      250
  THRESH          14
$end

which produces

 Mode:                 1                      2                      3
 Frequency:       319.13                 322.94                 860.45
 Force Cnst:      1.1357                 1.1630                 6.9772
 Red. Mass:      18.9268                18.9268                15.9949
 IR Active:          YES                    YES                    YES
 IR Intens:       33.122                 33.125                  0.000
 Raman Active:       YES                    YES                    YES
               X      Y      Z        X      Y      Z        X      Y      Z
 Eu        -0.105  0.102 -0.000   -0.102 -0.105  0.000   -0.000 -0.000 -0.000
 O          0.503 -0.486  0.000    0.486  0.503 -0.000    0.000  0.000 -0.707
 O          0.503 -0.486  0.000    0.486  0.503 -0.000    0.000  0.000  0.707
 TransDip  -0.133  0.128 -0.000   -0.128 -0.133  0.000   -0.000 -0.000 -0.000

For reference, I am using a 6.2.2 version of QChem that is about a few months old.

Compilation information
-----------------------

 SVN revision             | 45062
 SVN branch               | trunk
 Last SVN revision        | 45059
 Last SVN author          | linus.dittmer
 Hostname                 | 
 User                     | 
 System                   | Linux-5.4.0-176-generic
 Processor                | x86_64
 Compiler                 | GNU
 C compiler version       | 9.4.0
 CXX compiler version     | 9.4.0
 Fortran compiler version | 9.4.0
 Configuration time       | 2025-01-18 11:50:05

What $rem variables are you using to achieve this < 0.01 cm-1 degree of agreement?

#5

THRESH=14 but nothing else special.

$molecule
  1  1
  Eu
  O  1 rXO1
  O  1 rXO2  2 aOXO

  rXO1 =   1.6615601209
  rXO2 =   1.6615601209
  aOXO = 180.0000000000
$end

$rem
  JOBTYPE                                  opt
xc_grid 3
  METHOD                                  wB97M-V
  BASIS                                       def2-SVP
  SYMMETRY                             TRUE
  SYM_IGNORE                         FALSE
  THRESH                                  14
$end

followed by the corresponding FREQ calculation. This is with the current Q-Chem trunk.

#6

I am able to reproduce your results now. However, there is a lower SCF solution that needs to be reached with stability analysis. For that solution the lack of degeneracy is still there with the SG-3 or (99, 590) grids.

Upon further inspection, there does seem to be something funny going on with the f orbitals. I go back on my word, spatial symmetry breaking is probably what it is.