Cell Optimization Not Converged for Copper

186 views
Skip to first unread message

GENG YUAN

unread,
Oct 26, 2021, 5:02:47 PM10/26/21
to cp2k
Dear CP2K Users,

I am attempting to run the cell optimization for a single bulk Cu (with 72 atoms). However, the SCF calculation can not converge: It reached the maximum number of optimization steps (500) and therefore exit the optimization. 

I also tried changing the original lattice in my simulation (with associated coordinate) but it didn't work either.

So far, the cell optimization only converged when I use a smaller bulk Cu (with 32 atoms). But we'd like to run a larger system.

I've attached my input, output, the original Cu coordinate, as well as the trajectory during cell optimization here for your convenience. Any ideas are welcome! 

Below is the input:
-----------------------------
&GLOBAL
  PROJECT Cu_cellopt
  RUN_TYPE CELL_OPT
  PRINT_LEVEL LOW
&END GLOBAL
&FORCE_EVAL
  METHOD QS
  STRESS_TENSOR ANALYTICAL
  &SUBSYS
    &CELL
      ABC 10.83 10.83 7.22
      SYMMETRY TETRAGONAL
    &END CELL
    &TOPOLOGY
     COORD_FILE_NAME ./Cu332.xyz
     COORDINATE xyz
    &END
    &KIND Cu
      BASIS_SET DZVP-MOLOPT-SR-GTH
      POTENTIAL GTH-PBE-q11
    &END KIND
  &END SUBSYS
  &DFT
    BASIS_SET_FILE_NAME ./BASIS_MOLOPT
    POTENTIAL_FILE_NAME ./GTH_POTENTIALS
    &QS
      EPS_DEFAULT 1.0E-12
    &END QS
    &MGRID
      CUTOFF 400
      NGRIDS 5
      REL_CUTOFF 60
    &END MGRID
    &SCF
      SCF_GUESS ATOMIC
      EPS_SCF 1.0E-06
      MAX_SCF 200
      ADDED_MOS 100
      &OUTER_SCF
        MAX_SCF 50
        EPS_SCF 1.0E-6
      &END
      &DIAGONALIZATION T
        ALGORITHM STANDARD
      &END DIAGONALIZATION
      &MIXING T
        METHOD BROYDEN_MIXING
        ALPHA 0.1
        NBROYDEN 8
      &END MIXING
      &SMEAR ON
        METHOD FERMI_DIRAC
        ELECTRONIC_TEMPERATURE [K] 500
      &END SMEAR
      &PRINT
        &RESTART ON
        &END RESTART
      &END PRINT
    &END SCF
    &XC
      &XC_FUNCTIONAL
         &LIBXC
           FUNCTIONAL XC_GGA_X_PBE_R
         &END
         &LIBXC
           FUNCTIONAL XC_GGA_C_PBE
         &END
      &END XC_FUNCTIONAL
      &VDW_POTENTIAL
         POTENTIAL_TYPE PAIR_POTENTIAL
         &PAIR_POTENTIAL
            LONG_RANGE_CORRECTION
            PARAMETER_FILE_NAME dftd3.dat
            TYPE DFTD3
            REFERENCE_FUNCTIONAL revPBE
         &END PAIR_POTENTIAL
       &END VDW_POTENTIAL
       &XC_GRID
        XC_DERIV NN50_SMOOTH
       &END XC_GRID
    &END XC
  &END DFT
&END FORCE_EVAL
&MOTION
  &CELL_OPT
    EXTERNAL_PRESSURE 0
    TYPE DIRECT_CELL_OPT
    KEEP_SYMMETRY TRUE
    MAX_DR    1.0E-05
    MAX_FORCE 1.0E-05
    RMS_DR    1.0E-05    
    RMS_FORCE 1.0E-05
    MAX_ITER 500
    OPTIMIZER BFGS
  &END CELL_OPT
&END MOTION 
------------------------------------
Many thanks,
Geng.

trajectory.xyz
Cu332.xyz
heacell.out
heacell.inp

Fabian Ducry

unread,
Oct 27, 2021, 3:42:53 AM10/27/21
to cp...@googlegroups.com

Hi Geng,

Add a &CELL_REF section (https://manual.cp2k.org/cp2k-8_2-branch/CP2K_INPUT/FORCE_EVAL/SUBSYS/CELL/CELL_REF.html). With

&CELL_REF
  ABC 10.83*1.5 10.83*1.5 7.22*1.5
&END

the optimization converges within 35 steps. You could decrease EPS_SCF to reduce the noise in the forces.

Cheers,

Fabian

--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/6e75296d-eecd-4691-8e6b-560464b59bebn%40googlegroups.com.

GENG YUAN

unread,
Oct 27, 2021, 2:50:42 PM10/27/21
to cp2k
Dear Fabian,

Many thanks for your reply. I added CELL_REF in my simulation and it works!

Is it because that CELL_REF set a larger cell dimension, so the change of my real cell volume during the cell_opt is always under this reference cell volume, therefore the SCF could converge rather than keep fluctuating? (please correct me if I'm wrong). I found some information saying that CELL_REF is to fix the plane-wave basis whilst doing a cell_opt, but I don't know what does this means (I will try searching more).

About the EPS_SCF, do you have any suggestions about what criteria could I apply instead of the one I used in my simulation (1.0E-06)?

Many thanks for your help!
Sincerely,
Geng.

Fabian Ducry

unread,
Oct 28, 2021, 5:25:31 AM10/28/21
to cp...@googlegroups.com

Hi Geng,

The plane wave basis depends on the size of the unit cell. Therefore, a change of the cell dimensions corresponds to a change of basis. If the energy difference that comes from this change of basis is large the optimizer cannot converge. With CELL_REF you prevent this issue by fixing the basis (which is now independent of the volume of the simulation domain).

By the way, you are confusing the cell optimization and SCF: The SCF is the energy minimization at each step of the cell optimization (and this always converged). Only the cell optimization did not converge.

Regarding the EPS_SCF: a tighter value can lower the number of cell optimization steps, but each step will be more expensive. You have to test it yourself.

Cheers,

Fabian

GENG YUAN

unread,
Oct 28, 2021, 1:22:05 PM10/28/21
to cp2k
Hi Fabian,

Many thanks for your explanations!

Sincerely,
Geng

Reply all
Reply to author
Forward
0 new messages