Molcas Forum

debu_molcas

Member

Registered: 2016-11-13

Posts: 9

Optimization in presence of external electric field

Dear All,
             I need to optimize a system in presence of external electric field (EEF). I got informed that "FFPT — A Finite Field Perturbation Program" module can do the job. So, I used FFPT module to incorporate EEF during optimization. Please see the below for my input

> Export  MOLCAS_MAXITER=500
> Export  MOLCAS_PRINT=2

 &GATEWAY
Coord
2
Coordinates in Angstrom units
 H                 -0.19937304    1.26959246    0.00000000
 H                 -0.79937304    1.26959246    0.00000000
Basis=H.6-31G*...., 
Title=GateWay data By Gabedit
Group=Full
End of input

>>> Do while

 &SEWARD
Title=Seward data
End of input
 &FFPT
Title= Finite Perturbation with a dipole in the x negative of strength 0.1 au
FFPT
Dipole
 Z 0.1
End of input

 &SCF
Title=SCF data
KSDFT=B3LYP
End of input

 &ALASKA 
 &SLAPAF

>>> EndDo

 &MCKINLEY

 &GRID_IT
ASCII
ALL
End of input

I have seen that the output printed the following section

                                                 &FFPT

                                   only a single process is used
                       available to each process: 2.0 GB of memory, 1 thread
()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()


      **************************************************************************************************************************
      *                                                                                                                        *
      *                                                        Project:                                                        *
      *                         Finite Perturbation with a dipole in the x negative of strength 0.1 au                         *
      *                                                          FFPT                                                          *
      *                                                                                                                        *
      **************************************************************************************************************************

So, the FFPT module is switched on properly. Though I have some confusion about this calculation. Is FFPT is really do the optimization in presence of external electric field (EEF)?
Because in presence of EEF on bond axis for H2 (Here Z-axis) the ionicity of H2 increase and the H atoms get some partial charge and final geometry/energy should change at least little bit. But in this calculation no effects are observed. Here everything is same as in with of without FFTP.

Could you please that molcas has the possibility to do the calculation in presence of EEF?

One more Question about input: For a CASSCF calculation which files (from previous SCF calculations) required for the initial guess and how to call them in present calculation? If possible then please provide a sample input. 
   
Thanks & Regards

Debasish

Ignacio

Administrator

From: Uppsala

Registered: 2015-11-03

Posts: 1,085

Re: Optimization in presence of external electric field

You requested a calculation with full symmetry, and it will be impossible to get any asymmetry. Also, with Group=Full the molecule is translated and rotated, so the z axis may not be what you expect (check the actual geometry in the output); and if it is what you expect, you may be expecting the wrong thing: you have the molecule in the x direction an are applying a perpendicular field, and you have only s-type basis functions (no polarization on hydrogens with 6-31G*)

Turn off symmetry (Group=NoSym), add a star to your basis set (Basis=6-31G**), apply the field in the x direction (X 0.1), and you will see some effect. Unfortunately the effect seems to be that the optimization does not converge, and even when it does (with Hartree-Fock and smaller field) it takes much longer, I think somewhere along the line (maybe in alaska) the effect of the field is not fully considered.

debu_molcas

Member

Registered: 2016-11-13

Posts: 9

Re: Optimization in presence of external electric field

Thank you very much for the explanation. The single point calculation in presence of electric filed (EEF) is fine but the optimization is not successful. It is not that the job always terminates with an error but in the final output the influence of EEF is completely neglected. The two inputs are given below. Could you please go through and let me know that if there is any fault from my side.

Single point input

 &GATEWAY
Coord
2
Coordinates in Angstrom units
H     0.37139221     0.00000000     0.00000000
H    -0.37139221     0.00000000     0.00000000
Basis=H.6-31G*...., 
Title=GateWay data By Gabedit
Group=NoSym
End of input


 &SEWARD
Title=Seward data
End of input
 &FFPT
Title= Finite Perturbation with a dipole in the x negative of strength 0.1 au
FFPT
Dipole
 X 0.01
End of input

 &SCF
Title=SCF data
KSDFT=B3LYP
End of input

Optimization input

> Export  MOLCAS_MAXITER=500
> Export  MOLCAS_PRINT=2

 &GATEWAY
Coord
2
Coordinates in Angstrom units
H     0.37139221     0.00000000     0.00000000
H    -0.37139221     0.00000000     0.00000000
Basis=H.6-31G*...., 
Title=GateWay data By Gabedit
Group=NoSym
End of input

>>> Do while
 
 &SEWARD
Title=Seward data
End of input
 &FFPT
Title= Finite Perturbation with a dipole in the x negative of strength 0.1 au
FFPT
Dipole
 X 0.01
End of input

 &SCF
Title=SCF data
KSDFT=B3LYP
End of input

 &ALASKA 
 &SLAPAF

>>> EndDo

 &MCKINLEY
 
 &GRID_IT
ASCII
ALL
End of input

I have a query about the form of coordinate used in MOLCAS. Is it only Cartesian XYZ or any otehr form also possible? How to fix the direction of axis throughout the calculation? For example,
If the following initial geometry is used for optimization where Z-axis orients through C-Cl bond.
 

 C                  0.00000000    0.00000000   -1.13936200
 Cl                 0.00000000    0.00000000    0.66401000
 H                  0.00000000    1.03319700   -1.48399800
 H                  0.89477500   -0.51659800   -1.48399800
 H                 -0.89477500   -0.51659800   -1.48399800

After optimization the axis should be same but it changes the direction though NoSym keywords was used (see below the coordinate).

 C    -0.00000000    -0.00003231    -1.13853878
Cl    -0.00000000     0.00004549     0.66496435
 H     0.00000000     1.03373882    -1.48453464
 H     0.89527815    -0.51687551    -1.48461847
 H    -0.89527815    -0.51687551    -1.48461847

Could you please let me know how to fix it during calculation?

One more question about &FEPT: I have seen the input for field in one direction. Is it possible to provide a field with more than one direction simultaneously?

Ignacio

Administrator

From: Uppsala

Registered: 2015-11-03

Posts: 1,085

Re: Optimization in presence of external electric field

The single point calculation in presence of electric filed (EEF) is fine but the optimization is not successful. It is not that the job always terminates with an error but in the final output the influence of EEF is completely neglected.

Yes, that's what I saw. I think you'll see a change in the energy, but not in the geometry, that's why think &ALASKA may not be taking the field into account. If you can simulate your external field with external point multipoles (charges, dipoles...), that should work better (see XField keyword in &GATEWAY).

I have a query about the form of coordinate used in MOLCAS. Is it only Cartesian XYZ or any otehr form also possible? How to fix the direction of axis throughout the calculation? For example,
If the following initial geometry is used for optimization where Z-axis orients through C-Cl bond.

You could try with this: http://www.molcas.org/documentation/man … 0000000000
but I have no experience. In general, Molcas works in internal coordinates, assuming that the energy is invariant to translations/rotations. If there is an external perturbation, that's usually not the case and Molcas should detect it, if it doesn't, it's an indication that the combination is not fully tested/supported, and is probably not working. If, regardless, you want to fix some atoms/coordinates in space, you may do it through constraints or custom internal coordinates, see: http://www.molcas.org/documentation/man … 0000000000