Unreal cell parameters obtained after CELL_OPT

[Diego López]

Dear all,

I have been trying to optimize a unit cell of the 2D semiconductor CrSBr using GPW method implemented in CP2K. The initial cell parameters were a = 3.6 A, b = 4.8 A and c = 25.0 A (obtained from previous calculations with Quantum Espresso and in good agreement with experimental data). After 128 optimization cycles the geometry is optimized and the obtained cell parameters were a = 3.64 A and b = 6.60 A, which leads to a meaningless structure since there is no longer a 2D structure. Also the geometry of the unit cell is really distorted. Any suggestions to improve the accuracy of my calculation? Thank you in advanced.

I used this input:

&GLOBAL
  PROJECT test2
  RUN_TYPE CELL_OPT
  PRINT_LEVEL MEDIUM
&END GLOBAL

&FORCE_EVAL
  METHOD Quickstep              ! Electronic structure method (DFT,...)
  STRESS_TENSOR ANALYTICAL
  &DFT
    BASIS_SET_FILE_NAME  /BASIS_MOLOPT
    POTENTIAL_FILE_NAME  /GTH_POTENTIALS
    MULTIPLICITY 7
    &POISSON                    ! Solver requested for non periodic calculations
      PERIODIC XYZ
    &END POISSON
     &XC
      &XC_FUNCTIONAL PBE
      &END XC_FUNCTIONAL
     &END XC
    &SCF
      SCF_GUESS SPARSE
      MAX_SCF 100
      &OT
      PRECONDITIONER FULL_KINETIC
      MINIMIZER CG
      &END OT
      &OUTER_SCF
       MAX_SCF 50
      &END
    &END SCF
    &MGRID
      NGRIDS 4
      CUTOFF 550
      REL_CUTOFF 90
    &END MGRID
    &QS
       METHOD GPW
    &END
  &END DFT
  &SUBSYS
    &CELL
      ABC 3.601367151 4.818258107 25.0
      ALPHA_BETA_GAMMA 90 90 90
      PERIODIC XYZ
      SYMMETRY ORTHORHOMBIC
    &END CELL
    &TOPOLOGY                    ! Section used to center the atomic coordinates in the given box. Useful for big molecules
      COORD_FILE_FORMAT xyz
      COORD_FILE_NAME  ./crsbr.xyz
    &END
    &KIND Cr
      ELEMENT Cr
      BASIS_SET DZVP-MOLOPT-SR-GTH
      POTENTIAL GTH-PBE-q14
      MAGNETIZATION 3.0
    &END KIND
    &KIND S
      ELEMENT S
      BASIS_SET TZV2P-MOLOPT-GTH
      POTENTIAL GTH-PBE-q6
    &END KIND
    &KIND Br
      ELEMENT Br
      BASIS_SET DZVP-MOLOPT-SR-GTH
      POTENTIAL GTH-PBE-q7
    &END KIND
  &END SUBSYS
&END FORCE_EVAL
&MOTION
  &CELL_OPT
    TYPE DIRECT_CELL_OPT
    MAX_DR    1.0E-03
    MAX_FORCE 1.0E-03
    RMS_DR    1.0E-03
    RMS_FORCE 1.0E-03
    MAX_ITER 200
    OPTIMIZER BFGS
    KEEP_SYMMETRY
  &END CELL_OPT
 &GEO_OPT
    MAX_DR    1.0E-03
    MAX_FORCE 1.0E-03
    RMS_DR    1.0E-03
    RMS_FORCE 1.0E-03
    OPTIMIZER BFGS
  &END GEO_OPT
&END MOTION

[Krack Matthias]

Without k points, I suggest to try MULTIPLE_UNIT_CELL 3 3 1, i.e. a 3x3x1 supercell which has to be defined consistently in the &CELL and TOPOLOGY section.

--
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/ffb8824d-4222-4ba7-8a85-f0a99c49ee46n%40googlegroups.com.

[Diego López]

Thanks a lot for your suggestion Krack. Then, should I use k-points in the calculations or this  MULTIPLE_UNIT_CELL option?

Best,

Diego

[Krack Matthias]

That depends on your goals. I guess that you used a k point mesh with QE which you could apply likewise with CP2K to check if you get similar results. On the other hand, with k points less features are available and if you want to study larger unit cells (surface areas) later anyway then a supercell approach might be the better choice.

 

Matthias

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/ae820a14-e3e5-4be2-8dd6-5e60c6abc282n%40googlegroups.com.

[Diego López]

Thank you for your suggestions! I will take it into account.

Best,

Diego