Imaginary RPA root detected

Questions
#1

Hi, I encountered a problem when perform a TDDFT calculation.
¨ Q-Chem error occurred in module setman/rpa_dia.C, line 81:

Imaginary RPA root detected (w^2 = -3.087e-04):

  • if w is small, tighter thresholds may help
  • if w is large, SCF may not be at a stable solution¨

Input as below:
$rem
!QSYS wt=20h
jobtype sp
method wb97xd
basis def2-TZVP
BASIS2 STO-3G
CIS_N_ROOTS 50
CIS_SINGLETS false
CIS_TRIPLETS true
SYM_IGNORE true
SYMMETRY false
RPA TRUE
STS_MOM TRUE
mem_total 960000
INTERNAL_STABILITY true
threads 16
THRESH 14
SOLVENT_METHOD PCM
$end

$pcm
theory iefpcm
$end

$solvent
Dielectric 4.7113
$end

$molecule
0 3
C -3.5561036188 2.1415153680 -0.7128350067
C -2.3816174748 2.8346609961 -0.4044571484
C -3.5791676590 0.7526757995 -0.5595906092
C -1.2235084933 2.1696051284 -0.0417318989
C -2.4479172800 0.0398910865 -0.1898075562
C -1.2162095328 0.7326219316 -0.0188346211
N 0.0000765621 0.0465940462 0.0995333811
C 1.2155291027 0.7342218067 -0.0182316981
C 0.0009729736 -1.3509293161 0.3137632585
C 1.2211143636 2.1712154518 -0.0405911690
C 2.3787386586 2.8378042417 -0.4021185182
C 3.5542500412 2.1462419980 -0.7101270559
C 3.5787511263 0.7573649550 -0.5579280461
C 2.4481334981 0.0430865389 -0.1890857482
C -1.2297939856 -2.0361359263 0.4404776052
C -1.2032424686 -3.4103254005 0.8056088041
C 0.0027785479 -4.0687483694 0.9964765352
C 1.2325687272 -2.0346380536 0.4400919935
C 1.2079234613 -3.4089080378 0.8048505851
H -4.4510313590 2.6806724516 -1.0276954289
H -2.3887326776 3.9233982812 -0.4415792916
H -4.5164062887 0.2157370420 -0.7137991153
H 2.3846097116 3.9265548697 -0.4387174177
H 4.4487587754 2.6866512053 -1.0240247363
H 4.5166091351 0.2215609505 -0.7123651406
H -2.1394221161 -3.9565518879 0.9182651380
H 0.0034949449 -5.1253988094 1.2788019671
H 2.1448862515 -3.9539848888 0.9167893538
C -2.5659024590 -1.4365846321 0.0836272704
C -3.1046016237 -2.1587449223 -1.1760076759
C -3.5589822475 -1.6152017652 1.2582505182
C -0.0019362862 2.9029016918 0.4649665725
C -0.0028987200 4.3878864533 0.0969933800
C -0.0026497827 2.7870205457 2.0164342279
C 2.5676862048 -1.4335649743 0.0824111527
C 3.1051838488 -2.1533718979 -1.1790801866
C 3.5624508285 -1.6131398034 1.2553970373
H -4.1071142043 -1.7987762686 -1.4491074687
H -3.1709852469 -3.2414602546 -0.9960478188
H -2.4303318613 -1.9936057725 -2.0294412077
H -3.6631880607 -2.6763758685 1.5247540856
H -4.5569257500 -1.2374170491 0.9947989025
H -3.2022217325 -1.0717718683 2.1457578572
H -0.8824550263 4.8897902605 0.5216099512
H -0.0023005658 4.5393972419 -0.9925848444
H 0.8752521560 4.8912557966 0.5227694575
H 0.8920934895 3.2748805338 2.4314505805
H -0.0022114712 1.7363726647 2.3427391788
H -0.8983664925 3.2738598473 2.4305440710
H 3.1720708265 -3.2363805443 -1.0009049707
H 4.1072404577 -1.7924647379 -1.4527082102
H 2.4298472232 -1.9870325600 -2.0314394773
H 4.5596072954 -1.2336691127 0.9914313552
H 3.6684086042 -2.6746659620 1.5197598223
H 3.2061408718 -1.0718414704 2.1443871020
$end

Excited state optimization with RPA approximation
#2

Run stability analysis on your SCF solution, as the imaginary root typically implies an instability, or else give up on “full” TDDFT and use the Tamm-Dancoff approximation where this is not an issue.

This question has been dealt with before:

#3

Thanks. But I have already used the keyword ¨INTERNAL_STABILITY true¨. I do not know how to solve it without using TDA.

#4

My advice is to simply use TDA.

However, if the stability analysis does indicate an instability, I believe that the code perturbs the MOs along the unstable eigenmode and leaves them on disk, so that you can use SCF_GUESS = READ in a subsequent job, in order to try to locate a stable solution.