Molcas Forum

Anugraha

Member

Registered: 2025-12-17

Posts: 22

Non-equilibrium PCM optimization error

Dear OpenMolcas developers,

I am trying to optimize the first excited πσ* state of my system using CASSCF with the PCM solvent model in methanol. The πσ* state corresponds to the first root of symmetry 2. I have included the NONEquilibrium keyword together with RFRoot=1 in the RASSCF module.

The input is :

& GATEWAY
 RICD
 coord = tuca_1meth_cc_pvdz_mp2_s0_opt_pcm.Opt.xyz
 Group = z
Basis set = aug-cc-pvdz
RF-Input
PCM-Model
Solvent = Methanol
End of RF-Input
 Constraints
 r1 = bond H13 N12
 r2 = bond H13 O17
Values
 1.0 r1 - 1.0 r2 = -0.72051 angstrom
End of Constraints

>>> EXPORT MOLCAS_MAXITER = 2000

>>> DO WHILE

  &SEWARD
  &SCF
  &RASSCF
   symmetry = 2
   Levshift = 1.0
   Iterations = 200 100
   Spin = 1
    nActEl = 8 0 0
   Inactive = 37 4
   Ras2 = 1 8
   ciroot = 1 1 1
   Rlxroot = 1
   RFroot=1
   NONEquilibrium

   &SLAPAF

>>> END DO

But the calculation terminated with an error message :

()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()

                                              &SLAPAF

             launched 40 MPI processes, running in PARALLEL mode (work-sharing enabled)
                       available to each process: 8.0 GB of memory, 1 thread
                                         master pid: 208368
()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()

--- Stop Module: slapaf at Fri Jul  3 14:48:24 2026 /rc=_RC_INVOKED_OTHER_MODULE_ ---
--- Module slapaf spent 7 seconds ---

*** symbolic link created: INPORB -> nh_oh__0.72051_uca_1meth_pcm_pisig_cas89_opt_noneq.RasOrb
--- Start Module: alaska at Fri Jul  3 14:48:33 2026 ---

()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()

                                              &ALASKA

             launched 40 MPI processes, running in PARALLEL mode (work-sharing enabled)
                       available to each process: 8.0 GB of memory, 1 thread
                                         master pid: 209572
()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()

 ###############################################################################
 ###############################################################################
 ###                                                                         ###
 ###                                                                         ###
 ###    ERROR: Error in Alaska                                               ###
 ###                                                                         ###
 ###                                                                         ###
 ###############################################################################
 ###############################################################################
 NonEq=.True., invalid option
--- Stop Module: alaska at Fri Jul  3 14:48:41 2026 /rc=_RC_INTERNAL_ERROR_ ---
--- Module alaska spent 8 seconds ---

.########################################################.
.# Some internal inconsistency of the code was detected #.
.########################################################.

    Timing: Wall=1016.17 User=34811.73 System=704.29

I would be very thankful if anyone could help me understand the cause of this issue or suggest a possible solution. Any guidance would be greatly appreciated.

Thank You
Anugraha

Ignacio

Administrator

From: Uppsala

Registered: 2015-11-03

Posts: 1,204

Re: Non-equilibrium PCM optimization error

I don't think gradients are supported in nonequilibrium conditions. Even conceptually it's not clear what you're trying to do. The "slow" part is not in equilibrium with the current state, but it should be in equilibrium with something: with what? What happens when the geometry changes? The PCM cavity shape will change so the slow part will need to be recomputed... but now it cannot be in equilibrium with whatever it was earlier.

Operationally, you first need to do a RASSCF calculation without NONEq and with RFROot (or a single root) to compute the equilibrium field(s). Then you do a second RASSCF with NONEq. See e.g. https://gitlab.com/Molcas/OpenMolcas/-/ … type=heads