Clarification on RASSCF Calculation of Multiple Spin States in Q-Chem

Questions
#1

Dear Q-Chem Developers,

I am currently performing RASSCF calculations and would like to obtain the 1TT, 3TT, and 5TT states within a single output file. At present, I have only been able to achieve this by running three separate jobs with different values of the ras_spin_mult keyword. Could you please advise if there is a way to calculate all of these spin states within a single output file?

Additionally, I am using the crystal structure of an anthracene dimer, which shows three imaginary frequencies around the –30 cm⁻¹ region. Could you let me know whether these low-frequency modes are likely to cause any issues in the present calculations?

My input file is given below:

$rem
JOBTYPE SP
EXCHANGE HF
BASIS 6-31G(d,p)
UNRESTRICTED FALSE
SYM_IGNORE TRUE
SYMMETRY FALSE
SCF_ALGORITHM DIIS
MAX_SCF_CYCLES 512
SCF_CONVERGENCE 8
THRESH 14
BASIS_LIN_DEP_THRESH 14
MEM_TOTAL 6000
MEM_STATIC 1000
$end

@@@

$molecule
0 5
read
$end

$rem
scf_guess read
jobtype sp
exchange HF
basis 6-31G(d,p)
SCF_ALGORITHM DIIS
MAX_SCF_CYCLES 512
SCF_CONVERGENCE 8
THRESH 14
BASIS_LIN_DEP_THRESH 14
MEM_TOTAL 6000
MEM_STATIC 1000
$end

@@@

$molecule
0 5
read
$end

$rem
jobtype sp
exchange HF
basis 6-31G(d,p)
unrestricted false
SCF_ALGORITHM DIIS
MAX_SCF_CYCLES 512
SCF_CONVERGENCE 8
BASIS_LIN_DEP_THRESH 14
SCF_GUESS READ
gui 2
correlation rasci
ras_roots 10
ras_act 4
ras_elec 4
ras_act_orb [185,186,187,188]
ras_occ 184
max_cis_cycles 400
ras_spin_mult 1
ras_nfrag 2
ras_nfrag_atoms [46,46]
ras_frag_mo 1
ras_frag_sets [184,2,2]
n_frozen_core fc
ras_print 2
molden_format true
print_orbitals 10
mem_total 6000
mem_static 1000
thresh 14
$end

Thank you in advance.

#2

I’m not a RASSCF expert, but in general Q-Chem is not set up to compute different multiplicities within a single calculation except in some specialized cases (e.g., some versions of energy transfer couplings).

Regarding the imaginary frequencies, whether they matter depends on what you want to do with these calculations. If it’s just to get relative spin-state (electronic) energies then probably not, whereas if you’re trying to use a harmonic partition function to add an entropy correction and get free energy differences, then you need to worry about the fact that such a partition function is divergent for imaginary frequencies.

If you think they are artifacts, you can try re-optimizing the geometry with tighter stopping criteria. These keywords in $rem,

geom_opt_tol_gradient 150
geom_opt_tol_displacement 600
geom_opt_tol_energy 50

are 2x tighter than the default values, for example. That’s the easiest thing to try first; an alternative is to displace the geometry along the largest negative-curvature mode but there’s not an automated way to do that.