Only CELL_OPT with the given structure of the molecule

[arobendo mondal]

Hi,

I am new to CELL_OPT and not have any idea how to use it.

attached is the structure of the molecule having four iron dimers in an unit cell. This data is from XRD experiments. Cell size is  "ABC [angstrom]      14.0995    14.0995     34.5177" . Is it possible to keep the Fe-Fe distance as it is in the XRD structure and optimizing the cell_size. Or keeping the structure of the entire molecule as it is and optimizing the Cell_size.

Could you please suggest some cp2k input files with reasonable setups which can do calculate this.


Best wishes
Arobendo Mondal

TU- Berlin

[Samuel Andermatt]

I would start from CP2K%MOTION%CONSTRAINT to add the constraints you might want to your simulation.

[Florian Schiffmann]

Hi,

CONSTRAINTS will only work as restraints in geometry and cell optimizations. There is one exception and that is the FIXED_ATOMS part of cp2k. For what you want to do, I would either recommend a restraint on the Fe-Fe distance via a COLVAR or using fixed atoms on your whole system and run a cell optimization.In the latter case use the Cell_opt method GEO_OPT and CG as minimizer. Not 100% sure CP2K will be happy about such drastic measures but it will either crash or do what you want.

Cheers
Flo

[arobendo mondal]

Thank you Samuel and Florian.

I already tried something like this. I don't know why but CONSTRAINTS didn't worked. Below is the setting I am using. I have also attached the cell_opt-pos file.

Could you please suggest me what else can be done.

Best wishes

Arobendo

Input file :

###############################################################

&GLOBAL
  PROJECT Result-fe-dimer-cell_opt
  RUN_TYPE CELL_OPT
&END GLOBAL

###############################################################

&FORCE_EVAL
  METHOD QS
  STRESS_TENSOR ANALYTICAL
  #---------------------------------------------------------
  &DFT
    BASIS_SET_FILE_NAME ../../../../Project_CP2K/pot-set/BASIS_MOLOPT
    POTENTIAL_FILE_NAME ../../../../Project_CP2K/pot-set/POTENTIAL
    UKS             
    CHARGE                0
    MULTIPLICITY          33
    &MGRID
      CUTOFF 300
    &END MGRID
    &QS
      METHOD GPW
    &END QS
    &SCF
      SCF_GUESS           ATOMIC  
      EPS_SCF             1.E-6      
      MAX_SCF             100       
      &OUTER_SCF                  
        EPS_SCF           1.E-6        
        MAX_SCF           20          
      &END OUTER_SCF
      &OT
        ALGORITHM IRAC
        PRECONDITIONER FULL_ALL
      &END OT                               
    &END SCF
    &XC
      &XC_FUNCTIONAL       PBE        
      &END XC_FUNCTIONAL
      &XC_GRID                       
        XC_SMOOTH_RHO     NN10         
        XC_DERIV          SPLINE2_SMOOTH 
      &END XC_GRID
    &END XC
  &END DFT
  #-----------------------------------------------------------
  &SUBSYS
    &CELL

      ABC [angstrom]      14.0995    14.0995     34.5177

      PERIODIC            XYZ          # default= XYZ
    &END CELL
    &COORD
      @INCLUDE      './fe-4dimer.xyz'
    &END COORD
    &KIND           H
      BASIS_SET    SZV-MOLOPT-SR-GTH-q1
      POTENTIAL    GTH-PADE-q1
    &END KIND
    &KIND           C
      BASIS_SET    SZV-MOLOPT-SR-GTH-q4
      POTENTIAL    GTH-PADE-q4
    &END KIND
    &KIND           O
      BASIS_SET    SZV-MOLOPT-SR-GTH-q6
      POTENTIAL    GTH-PADE-q6
    &END KIND
    &KIND           Si
      BASIS_SET    SZV-MOLOPT-SR-GTH-q4
      POTENTIAL    GTH-PADE-q4
    &END KIND
    &KIND           Fe
      BASIS_SET    SZV-MOLOPT-SR-GTH-q16
      POTENTIAL    GTH-PADE-q16
    &END KIND
  &END SUBSYS
&END FORCE_EVAL
############################################################
&MOTION
  &CONSTRAINT
    &FIXED_ATOMS
      COMPONENTS_TO_FIX XYZ
      LIST 1 90 179 268 357 446 535 624   #
    &END FIXED_ATOMS
  &END CONSTRAINT
  &CELL_OPT
    EXTERNAL_PRESSURE 0.0
    TYPE DIRECT_CELL_OPT
    OPTIMIZER CG
    MAX_ITER 20
    KEEP_ANGLES
    &CG
      &LINE_SEARCH
        TYPE 2PNT
        &2PNT
        &END 2PNT
      &END LINE_SEARCH
    &END CG
  &END CELL_OPT

# &CELL_OPT
#    TYPE                  DIRECT_CELL_OPT  # default= DIRECT_CELL_OPT
#    MAX_ITER              200          # default= 200
#    OPTIMIZER             BFGS         # default= BFGS; for large systems: CG
#    EXTERNAL_PRESSURE [bar] 1.0        # default= 1.0 [bar]
#    KEEP_ANGLES                        # default= .FALSE.
# &END
 &GEO_OPT
    MAX_ITER              200          # default= 200
    OPTIMIZER             BFGS         # default= BFGS ; for large systems: CG
 &END GEO_OPT
&END MOTION
############################################################

--
You received this message because you are subscribed to a topic in the Google Groups "cp2k" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/cp2k/X_IGmrmBRKE/unsubscribe.
To unsubscribe from this group and all its topics, send an email to cp2k+uns...@googlegroups.com.
To post to this group, send email to cp...@googlegroups.com.
Visit this group at http://groups.google.com/group/cp2k.
For more options, visit https://groups.google.com/d/optout.

[Florian Schiffmann]

Hi,

I haven't cheged the output file, but in general there is a problem defining what fixed means if you change scale your coordinate system (cell). The famous example of painting two dots on a ballon and blowing it up. The dots do not move but their distance increases as the coordinate systems (surface of the baloon expands). The motion of atoms you see should only be in absolute cartesians but not in fractional coordinates (which is the only meaningful way to fix them).
If you want a given absolute distance preserved you will have to specify a COLVAR (e.g. distance) in TOPOLOGY, and then use it in the CONSTRAINT section with the RESTRAINT section set and a reasonably large K to add a spring term which drives your system in the desired state.
However be careful. The Total energy now contains the contribution from the spring and is thus an artificial energy. If you want to compare energies in that case you will have to look at the SCF energies.

Cheers
Flo

[arobendo mondal]

Thank you Florian.

I am new to CP2K and have never played with these keywords. I was looking into regtest files but didn't get anything reasonable for this purpose. Could you please help me with the input file.

Best wishes
Arobendo