Large number of imaginary frequencies with ROKS and delta SCF

Questions
#1

Hello, I performed a singlet excited-state (S1) geometry optimization using ROKS and del SCF approach. The optimization converged smoothly in a few cycles with small gradients, However, when I performed a frequency calculation at the optimized geometry, I obtained many imaginary frequencies. I previously did this with TDDFT approach but did not face this issue. How can I resolve it?
Input file with ROKS:
$molecule
0 1
H 1.3292676 -1.6836893 -3.0809952
C 0.7356205 -0.9500791 -3.6169987
C 0.0000000 0.0000000 -2.8999730
C -0.7356205 0.9500791 -3.6169987
H -1.3292676 1.6836893 -3.0809952
C -0.7357354 0.9503581 -5.0068087
H -1.3188870 1.6920649 -5.5435948
C 0.0000000 0.0000000 -5.7078619
H 0.0000000 0.0000000 -6.7931470
C 0.7357354 -0.9503581 -5.0068087
H 1.3188870 -1.6920649 -5.5435948
C 0.0000000 0.0000000 -1.4156563
C -0.0005202 1.1978303 -0.6937517
C 0.0005202 1.1978303 0.6937517
C 0.0000000 0.0000000 1.4156563
C -0.0005202 -1.1978303 0.6937517
C 0.0005202 -1.1978303 -0.6937517
C 0.0000000 0.0000000 2.8999730
C 0.7356205 0.9500791 3.6169987
C 0.7357354 0.9503581 5.0068087
C 0.0000000 0.0000000 5.7078619
C -0.7357354 -0.9503581 5.0068087
C -0.7356205 -0.9500791 3.6169987
H 0.0109345 2.1444854 -1.2243742
H -0.0109345 2.1444854 1.2243742
H 0.0109345 -2.1444854 1.2243742
H -0.0109345 -2.1444854 -1.2243742
H 1.3292676 1.6836893 3.0809952
H 1.3188870 1.6920649 5.5435948
H 0.0000000 0.0000000 6.7931470
H -1.3188870 -1.6920649 5.5435948
H -1.3292676 -1.6836893 3.0809952
$end

$rem
JOB_TYPE SP
METHOD CAMB3LYP
BASIS 6-31G(d,p)
SCF_CONVERGENCE 8
MAX_SCF_CYCLES 200
SYM_IGNORE TRUE
SOLVENT_METHOD PCM
MEM_TOTAL 248000
MEM_STATIC 4000
$end

$pcm
THEORY CPCM
method swig
radii bondi
solver inversion
$end

$solvent
DIELECTRIC 37.5
OPTICALDIELECTRIC 1.8068
$end

@@@

$molecule
read
$end

$comment
OPP S1 Delta-SCF (HOMO->LUMO) in MeCN PCM
$end

$rem
JOB_TYPE OPTIMIZATION
METHOD CAMB3LYP
BASIS 6-31G(d,p)
UNRESTRICTED FALSE
ROKS TRUE
SCF_GUESS READ
SCF_CONVERGENCE 6
MAX_SCF_CYCLES 200
SYM_IGNORE TRUE
SOLVENT_METHOD PCM
MEM_TOTAL 248000
MEM_STATIC 4000
$end

$pcm
THEORY CPCM
method swig
radii bondi
solver inversion
$end

$solvent
DIELECTRIC 37.5
OPTICALDIELECTRIC 1.8068
$end

$occupied
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 47 48 49 50 51 52 53 54 55 56 57 58 59 60 62
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 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61
$end

@@@

$molecule
read
$end

$comment
OPP S1 Delta-SCF (HOMO->LUMO) in MeCN PCM
$end

$rem
JOB_TYPE Frequency
METHOD CAMB3LYP
BASIS 6-31G(d,p)
UNRESTRICTED FALSE
ROKS TRUE
SCF_GUESS READ
SCF_CONVERGENCE 6
MAX_SCF_CYCLES 200
SYM_IGNORE TRUE
SOLVENT_METHOD PCM
MEM_TOTAL 248000
MEM_STATIC 4000
$end

$pcm
THEORY CPCM
method swig
radii bondi
solver inversion
$end

$solvent
DIELECTRIC 37.5
OPTICALDIELECTRIC 1.8068
$end

Input file with del SCF:
$molecule
0 1
H 1.3292676 -1.6836893 -3.0809952
C 0.7356205 -0.9500791 -3.6169987
C 0.0000000 0.0000000 -2.8999730
C -0.7356205 0.9500791 -3.6169987
H -1.3292676 1.6836893 -3.0809952
C -0.7357354 0.9503581 -5.0068087
H -1.3188870 1.6920649 -5.5435948
C 0.0000000 0.0000000 -5.7078619
H 0.0000000 0.0000000 -6.7931470
C 0.7357354 -0.9503581 -5.0068087
H 1.3188870 -1.6920649 -5.5435948
C 0.0000000 0.0000000 -1.4156563
C -0.0005202 1.1978303 -0.6937517
C 0.0005202 1.1978303 0.6937517
C 0.0000000 0.0000000 1.4156563
C -0.0005202 -1.1978303 0.6937517
C 0.0005202 -1.1978303 -0.6937517
C 0.0000000 0.0000000 2.8999730
C 0.7356205 0.9500791 3.6169987
C 0.7357354 0.9503581 5.0068087
C 0.0000000 0.0000000 5.7078619
C -0.7357354 -0.9503581 5.0068087
C -0.7356205 -0.9500791 3.6169987
H 0.0109345 2.1444854 -1.2243742
H -0.0109345 2.1444854 1.2243742
H 0.0109345 -2.1444854 1.2243742
H -0.0109345 -2.1444854 -1.2243742
H 1.3292676 1.6836893 3.0809952
H 1.3188870 1.6920649 5.5435948
H 0.0000000 0.0000000 6.7931470
H -1.3188870 -1.6920649 5.5435948
H -1.3292676 -1.6836893 3.0809952
$end

$rem
JOB_TYPE SP
METHOD CAMB3LYP
BASIS 6-31G(d,p)
SCF_ALGORITHM DIIS
SCF_CONVERGENCE 8
MAX_SCF_CYCLES 200
SYM_IGNORE TRUE
UNRESTRICTED TRUE
SOLVENT_METHOD PCM
MEM_TOTAL 248000
MEM_STATIC 4000
$end

$pcm
THEORY CPCM
method swig
radii bondi
solver inversion
$end

$solvent
DIELECTRIC 37.5
OPTICALDIELECTRIC 1.8068
$end

@@@

$molecule
read
$end

$rem
JOB_TYPE OPTIMIZATION
METHOD CAMB3LYP
BASIS 6-31G(d,p)
UNRESTRICTED TRUE
OPSING TRUE
MOM_START TRUE
MOM_METHOD MOM
SCF_ALGORITHM DIIS
SCF_GUESS READ
SCF_CONVERGENCE 6
MAX_SCF_CYCLES 500
SOLVENT_METHOD PCM
SYM_IGNORE TRUE
MEM_TOTAL 248000
MEM_STATIC 4000
$end

$pcm
THEORY CPCM
method swig
radii bondi
solver inversion
$end

$solvent
DIELECTRIC 37.5
OPTICALDIELECTRIC 1.8068
$end

$occupied
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 47 48 49 50 51 52 53 54 55 56 57 58 59 60 62
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 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61
$end

@@@

$molecule
read
$end

$rem
JOB_TYPE Frequency
METHOD CAMB3LYP
BASIS 6-31G(d,p)
UNRESTRICTED TRUE
OPSING TRUE
MOM_START TRUE
MOM_METHOD MOM
SCF_ALGORITHM DIIS
SCF_GUESS READ
SCF_CONVERGENCE 6
MAX_SCF_CYCLES 500
SOLVENT_METHOD PCM
SYM_IGNORE TRUE
MEM_TOTAL 248000
MEM_STATIC 4000
$end

$pcm
THEORY CPCM
method swig
radii bondi
solver inversion
$end

$solvent
DIELECTRIC 37.5
OPTICALDIELECTRIC 1.8068
$end

#2

I’m not sure that OPSING has analytic Hessians implemented; manual isn’t clear on that point and the code looks like it’s trying to run them but I’m suspicious. You could test against finite-difference (IDERIV=1) on a small job.

For ROKS, manual is clear that the frequencies are computed by finite difference, it which case you need to be concerned about variational collapse of your displaced SCF solutions.