Excited state optimization with RPA approximation

I am trying to optimize the S1 geometry of a molecule. The rem section of the input file is as below-
$rem
basis = 6-311g**
mem_static = 2000
mem_total = 16000
job_type = optimization
method = wb97x-d
geom_opt_max_cycles = 500
max_cis_cycles = 100
cis_n_roots = 3
cis_state_deriv = 1
cis_triplets = false
cis_singlets = true
rpa = true
scf_convergence = 8
symmetry = false
symmetry_ignore = true
thresh = 14
SCF_MAX_CYCLES = 500
$end

However, after some optimization cycles the calculation terminated with the following error-
Q-Chem fatal error occurred in module setman/rpa_dia.C, line 59:
Unable to diagonalize A-B

I understand one way to deal with the error is using TDA approximation instead of RPA. But I need to perform my calculations with RPA. So any insight about this error message will be appreciated.

That message usually indicates a triplet instability in the reference state, meaning that you fundamentally cannot perform a full TDDFT calculation that geometry.

See additional discussion here:

(If you don’t get the precise error message about w^2 being imaginary, you are probably using a version of Q-Chem that pre-dates that somewhat recent change.)

Also here: