Molcas Forum
Support and discussions for Molcas and OpenMolcas users and developers
You are not logged in.
- Topics: Active | Unanswered
Announcement
Please note: The forum's URL has changed. The new URL is: https://molcasforum.univie.ac.at. Please update your bookmarks!
You can choose an avatar and change the default style by going to "Profile" → "Personality" or "Display".Pages: 1
#1 2025-08-21 19:13:40
- hbschlegel
- Member
- Registered: 2025-08-16
- Posts: 6
CASPT2 and MCPDFT frequencies
Can openMolCAS calculate CASPT2 and MCPDFT frequencies?
If so, can you please provide a sample input? (with symmetry)
Thanks, Berny Schlegel
Offline
#2 2025-08-22 09:01:48
- Ignacio
- Administrator
- From: Uppsala
- Registered: 2015-11-03
- Posts: 1,141
Re: CASPT2 and MCPDFT frequencies
It can, but...
a) The differentiation will be numerical or, at best, seminumerical. (For CASPT2 at least, fully numerical with symmetry, seminumerical without)
b) Only the symmetry-preserving displacements can be performed, so only the fully symmetric modes are reported.
H2O example:
&GATEWAY
Coord = 3
O 0.00 0.00 0.00
H 0.56 0.75 0.00
H 0.56 -0.75 0.00
Basis = cc-pVDZ
Group = Full
RICD
> do while
&SEWARD
&RASSCF
NActEl = 8
Inactive = 1 0 0 0
RAS2 = 3 2 0 1
&CASPT2
IPEA = 0.0
&SLAPAF
> end do
&MCKINLEYSame, without symmetry:
&GATEWAY
Coord = 3
O 0.00 0.00 0.00
H 0.56 0.75 0.00
H 0.56 -0.75 0.00
Basis = cc-pVDZ
Group = NoSym
RICD
> do while
&SEWARD
&RASSCF
NActEl = 8
Charge = 0
RAS2 = 6
&CASPT2
IPEA = 0.0
&SLAPAF
> end do
&MCKINLEYOffline
#3 2025-08-30 20:03:40
- hbschlegel
- Member
- Registered: 2025-08-16
- Posts: 6
Re: CASPT2 and MCPDFT frequencies
After numerous attempts, still trying to get a full set of vibrational frequencies with MC-PDFT (works with Cs symmetry but only yields the symmetric frequencies; with C1 symmetry RASSCF convergence problems during perturbations; trying CISD in the same active space + MC-PDFT)
Let me know what I am missing in the input or if there is a better / proper way of getting the full set of frequencies for a MC-PDFT calculation based on a MC wavefunction with a small active space (8 electrons, 7 orbitals). The goal is to get structures and full set of vibrational frequencies for the four lowest states of the cation using CISD-IP + MC-PDFT.
The following input yields NAN's from MCLR
&GATEWAY
coord=geom.xyz; basis=def2-TZVP; group=C1
&SEWARD
grid input
grid=ultrafine
end of grid input
&RASSCF
FileOrb=Molecule.RasOrb
Spin=1; Symmetry= 1
INACTIVE= 14
NActEl= 8 2 2
RAS1=4
RAS2=0
RAS3=3
CIRoot= 4 4 1
RLXR= 1
&MCPDFT
func=T:PBE
GRAD
&SLAPAF
&MCKINLEY
Perturbation
Hessian
&MCLR
SaLa
The relevant portions of the output include the following:
:: MCPDFT root number 1 Total energy: -265.66188627
:: MCPDFT root number 2 Total energy: -265.46503246
:: MCPDFT root number 3 Total energy: -265.43674587
:: MCPDFT root number 4 Total energy: -265.43198854
--- Stop Module: mcpdft at Sat Aug 30 14:15:03 2025 /rc=_RC_ALL_IS_WELL_ ---
*** files: xmldump
saved to directory /wsu/home/aa/aa00/aa0023/molcas-calc/propiolic/neutral/cisd-pdft
--- Module mcpdft spent 4 minutes 29 seconds ---
*** symbolic link created: INPORB -> Molecule.RasOrb
--- Start Module: slapaf at Sat Aug 30 14:15:03 2025 ---
()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()
&SLAPAF
only a single process is used
available to each process: 4.0 GB of memory, 1 thread?
pid: 25804
()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()
--- Stop Module: slapaf at Sat Aug 30 14:15:03 2025 /rc=_RC_INVOKED_OTHER_MODULE_ ---
*** files: Molecule.slapaf.h5 xmldump
saved to directory /wsu/home/aa/aa00/aa0023/molcas-calc/propiolic/neutral/cisd-pdft
*** symbolic link created: INPORB -> Molecule.RasOrb
--- Start Module: alaska at Sat Aug 30 14:15:03 2025 ---
()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()
&ALASKA
only a single process is used
available to each process: 4.0 GB of memory, 1 thread?
pid: 25804
()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()
()()
--- Stop Module: slapaf at Sat Aug 30 14:15:03 2025 /rc=_RC_INVOKED_OTHER_MODULE_ ---
*** files: Molecule.slapaf.h5 xmldump
saved to directory /wsu/home/aa/aa00/aa0023/molcas-calc/propiolic/neutral/cisd-pdft
*** symbolic link created: INPORB -> Molecule.RasOrb
--- Start Module: alaska at Sat Aug 30 14:15:03 2025 ---
()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()
()()
&ALASKA
only a single process is used
available to each process: 4.0 GB of memory, 1 thread?
pid: 25807
()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()
()()
--- Stop Module: alaska at Sat Aug 30 14:15:03 2025 /rc=_RC_INVOKED_OTHER_MODULE_ ---
*** files: xmldump
saved to directory /wsu/home/aa/aa00/aa0023/molcas-calc/propiolic/neutral/cisd-pdft
*** symbolic link created: INPORB -> Molecule.RasOrb
--- Start Module: mclr at Sat Aug 30 14:15:03 2025 ---
()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()
()()
&MCLR
only a single process is used
available to each process: 4.0 GB of memory, 1 thread?
pid: 25810
()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()
()()
Ordinary integral handling
No .RLXPT2 or MCKINT , I hope that is OK
Seward mode is assumed, reading perturbation from ONEINT
OrdInt status: non-squared
++ Wave function specifications:
-----------------------------
Number of closed shell electrons 28
Number of electrons in active shells 8
Max number of holes in RAS1 space 2
Max number of electrons in RAS3 space 2
Number of inactive orbitals 14
Number of active orbitals 7
Number of secondary orbitals 146
Spin quantum number 0.0
State symmetry 1
Number of CI roots 4
States considered 1 2 3 4
Weights 0.250 0.250 0.250 0.250
Symmetry species 1
Skipped sym. species 0
Frozen orbitals 0
Inactive orbitals 14
Active orbitals 7
RAS1 orbitals 4
RAS2 orbitals 0
RAS3 orbitals 3
Deleted orbitals 0
Number of basis functions 167
Number of orbitals 167
Number of configurations 91
Number of combinations 109
Natural orbitals are used in the last CI
RASSCF state energy = -264.6058912504
Size of explicit Hamiltonian in PCG: 100
--
Convergence threshold= 1.0000E-04
Max number of iterations in PCG: 200
Lagrangian multipliers are calculated for state no. 1
Linear response function is computed for root no. = 1
Transformation of integrals
Iteration Delta Res(kappa) Res(CI) DeltaK DeltaC
1 NaN NaN NaN NaN NaN
2 NaN NaN NaN NaN NaN
3 NaN NaN NaN NaN NaN
4 NaN NaN NaN NaN NaN
5 NaN NaN NaN NaN NaN
6 NaN NaN NaN NaN NaN
7 NaN NaN NaN NaN NaN
8 NaN NaN NaN NaN NaN
9 NaN NaN NaN NaN NaN
...
Element nr. 97 is NaN
Element nr. 98 is NaN
Element nr. 99 is NaN
Element nr. 100 is NaN
...too many. I give up here.
There were a total of 14028 NANs
--- Stop Module: mclr at Sat Aug 30 14:15:53 2025 /rc=-6 ---
*** files: xmldump
saved to directory /wsu/home/aa/aa00/aa0023/molcas-calc/propiolic/neutral/cisd-pdft
--- Module mclr spent 49 seconds ---
.########################.
.# Non-zero return code #.
.########################.
Timing: Wall=392.76 User=358.76 System=22.00Let me know if you need any additional files.
Many thanks for your help,
Berny Schlegel
Offline
#4 2025-09-01 08:52:00
- Ignacio
- Administrator
- From: Uppsala
- Registered: 2015-11-03
- Posts: 1,141
Re: CASPT2 and MCPDFT frequencies
A couple of comments about your input:
No need for SLAPAF if you're not doing an optimization. And if you're doing it, you'll need a loop around it.
I don't think MCLR is needed either. It will be called automatically if needed (and only for analytical 2nd derivatives).
The NaNs could be due to some convergence issues or bug at some specific geometry. I tried this input, and it ran correctly:
&GATEWAY
Coord = 3
O 0.00 0.00 0.00
H 0.56 0.75 0.00
H 0.56 -0.75 0.00
Basis = cc-pVDZ
Group = NoSym
RICD
> do while
&SEWARD
&RASSCF
NActEl = 8
Charge = 0
RAS2 = 6
StAverage = 4
RlxRoot = 1
&MCPDFT
Functional = t:PBE
Grad
&SLAPAF
> end do
&MCKINLEYIf you already have the converged geometry, you can remove the call to SLAPAF and the "do while" loop.
Offline
#5 2025-09-03 01:06:18
- hbschlegel
- Member
- Registered: 2025-08-16
- Posts: 6
Re: CASPT2 and MCPDFT frequencies
Thank you for the working example for H2O with RAS2=6. Your example works fine on our installation of openMolCAS.
For propiolic acid, a nossymm RASSCF frequency calculation fails during the frequency calculation because of convergence problems in RASSCF.
Since we require the full Hessian, we need to use nosymm. A RASSCF calculation with symmetry succeeds, but yields only the symmetric vibrations.
As background information, we are looking for a method for structures and frequencies of cation ground and low lying excited states to study the dynamics of coherent superposition of cation states. There are several molecules of similar size we would like to investigate in connection with our collaborator's experimental work with strong field lasers. MC-PDFT seems like a promising and efficient approach. We are also looking at CCSD-EOM-IP (with partial success). We would welcome other suggestions (we have some indication that perturbative approaches may not work well because of configurational mixing among the excited cation states).
CISD calculations (with RAS2=0 and allowing only 2 holes in RAS1 and 2 particles in RAS3) could be a good black-box alternative to CASSCF to prepare a wavefunction for MC-PDFT calculations.
The CISD frequency calculation for propiolic acid and its cations work fine.
However, the CISD + MC-PDFT calculation for propiolic acid and for H2O runs into NAN's in the frequency calculation.
The input for H2O is listed below along with relevant sections of the .log file.
If this is a reasonable calculation, please let me know if this is a user input error (and let me know how to fix it).
Alternatively, this small example may help in tracking down the source of the NAN's.
Many thanks in advance for your help!
Berny Schlegel
&GATEWAY
Coord = 3
O 0.00 0.00 0.00
H 0.56 0.75 0.00
H 0.56 -0.75 0.00
Basis = cc-pVDZ
Group = NoSym
RICD
> do while
&SEWARD
&RASSCF
NActEl = 8 2 2
Charge = 0
RAS1 = 4
RAS2 = 0
RAS3 = 2
StAverage = 4
RlxRoot = 1
&MCPDFT
Functional = t:PBE
Grad
&SLAPAF
> end do
&MCKINLEY---------
()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()
&MCLR
only a single process is used
available to each process: 4.0 GB of memory, 1 thread?
pid: 9554
()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()
No .RLXPT2 or MCKINT , I hope that is OK
Seward mode is assumed, reading perturbation from ONEINT
OrdInt status: non-squared
++ Wave function specifications:
-----------------------------
Number of closed shell electrons 2
Number of electrons in active shells 8
Max number of holes in RAS1 space 2
Max number of electrons in RAS3 space 2
Number of inactive orbitals 1
Number of active orbitals 6
Number of secondary orbitals 17
Spin quantum number 0.0
State symmetry 1
Number of CI roots 4
States considered 1 2 3 4
Weights 0.250 0.250 0.250 0.250
Symmetry species 1
Skipped sym. species 0
Frozen orbitals 0
Inactive orbitals 1
Active orbitals 6
RAS1 orbitals 4
RAS2 orbitals 0
RAS3 orbitals 2
Deleted orbitals 0
Number of basis functions 24
Number of orbitals 24
Number of configurations 45
--
Convergence threshold= 1.0000E-04
Max number of iterations in PCG: 200
Using the Cho-MO Algorithm
Lagrangian multipliers are calculated for state no. 1
Linear response function is computed for root no. = 1
Transformation of integrals
Reordering of the Cholesky vectors to full storage.
Elapsed time for the reordering section: 1.0000000000000009E-002
CPU time for the reordering section: 1.0000000000000009E-002
TIMING INFORMATION: CPU(s) %CPU Elapsed(s)
Transformation 0.00 0.0 0.00
Generation 0.00 0.0 0.01
TOTAL 0.02 2.0 0.03
Iteration Delta Res(kappa) Res(CI) DeltaK DeltaC
1 NaN NaN NaN NaN NaN
2 NaN NaN NaN NaN NaN
3 NaN NaN NaN NaN NaN
4 NaN NaN NaN NaN NaN
5 NaN NaN NaN NaN NaN
...
197 NaN NaN NaN NaN NaN
198 NaN NaN NaN NaN NaN
199 NaN NaN NaN NaN NaN
No convergence for perturbation no: 1. Increase Iter.
!!! WARNING !!!
NANs encountered
The numbers in the array will now be checked.
There are 300 elements.
Element nr. 1 is NaN
Element nr. 2 is NaN
Element nr. 3 is NaN
Element nr. 4 is NaN
Element nr. 5 is NaN
Element nr. 6 is NaN
Element nr. 7 is NaN
...
Element nr. 97 is NaN
Element nr. 98 is NaN
Element nr. 99 is NaN
Element nr. 100 is NaN
...too many. I give up here.
There were a total of 300 NANs
--- Stop Module: mclr at Tue Sep 2 19:19:07 2025 /rc=-6 ---
*** files: xmldump
saved to directory /wsu/home/aa/aa00/aa0023/molcas-calc/propiolic/neutral/h2o
--- Module mclr spent 6 seconds ---
.########################.
.# Non-zero return code #.
.########################.
Timing: Wall=12.23 User=5.25 System=0.43Offline
#6 2025-09-03 09:51:12
- Ignacio
- Administrator
- From: Uppsala
- Registered: 2015-11-03
- Posts: 1,141
Re: CASPT2 and MCPDFT frequencies
Hmm... I'm not sure MCPDFT has really been tested with RASSCF wave functions, and I would bet analytical gradients are not implemented (and that's why there are NaNs). Even plain RASSCF analytical gradients have not been well tested (they're only available as a side effect of CASPT2 gradients).
You may try RASPT2 with some multistate variant (XMS or RMS). In this particular water case, I get a negative frequency for the asymmetric stretch, but I can track this to a discontinuity in the RASSCF surface, and it can be avoided by reducing the displacement for numerical differentiation (unfortunately that requires manually writing the second SLAPAF loop, instead of the MCKINLEY call):
&GATEWAY
Coord = 3
O 0.00 0.00 0.00
H 0.56 0.75 0.00
H 0.56 -0.75 0.00
Basis = cc-pVDZ
Group = NoSym
RICD
> do while
&SEWARD
&RASSCF
NActEl = 8 2 2
Charge = 0
RAS1 = 4
RAS2 = 0
RAS3 = 2
StAverage = 4
RlxRoot = 1
&CASPT2
Imaginary = 0.1
RMult = all
IPEA = 0.0
&SLAPAF
> end do
> export MOLCAS_MAXITER=500
> do while
&SEWARD
&RASSCF
NActEl = 8 2 2
Charge = 0
RAS1 = 4
RAS2 = 0
RAS3 = 2
StAverage = 4
RlxRoot = 1
&CASPT2
Imaginary = 0.1
RMult = all
IPEA = 0.0
&SLAPAF
Numerical
Iterations = 0
Delta = 1.0e-3
> end doOffline
Pages: 1