Wrong density prediction in NpT AIMD

[Abedi, Mostafa]

Hi Everyone,

I've performed NpT AIMD simulations for a system containing 8 CH4 molecules. The system was equilibrated using 2 ps NVT (300K) followed by 2 ps NVE. For some reason, the predicted density keeps decreasing and getting away from the correct value (The correct density is 0.232 g/cm3):

    Time evaluation of density for a box of 8 CH4 at 300 K and 400 bar

The wB97XD/TZVP-MOLOPT-GTH level of theory was used for the electronic structure part. Unfortunately, there is not much in the literature for this system (or similar systems) to get some ideas for the level of theory. To my mind, it could be due to the wrong prediction of the intermolecular repulsive interactions (vdW dispersion forces). But, there might be some other parameters in the simulations that can cause this issue. Any suggestions or advice for fixing this issue would be greatly appreciated. Many thanks.

Best,

Mostafa

Input file:

&GLOBAL
  PROJECT methane
  RUN_TYPE MD  
  WALLTIME 10000000
  IOLEVEL  LOW          
&END GLOBAL

&FORCE_EVAL
 STRESS_TENSOR ANALYTICAL
  METHOD Quickstep
  &DFT
    BASIS_SET_FILE_NAME BASIS_MOLOPT
    POTENTIAL_FILE_NAME GTH_POTENTIALS            
    CHARGE 0
    MULTIPLICITY 1
    &MGRID
       CUTOFF [Ry] 400
    &END
    &QS
       METHOD GPW
       EPS_DEFAULT 1.0E-10
       EXTRAPOLATION ASPC
    &END
    &POISSON
       PERIODIC XYZ
       POISSON_SOLVER PERIODIC
    &END
    &SCF  
      &PRINT
        &RESTART OFF
        &END
      &END        
      SCF_GUESS ATOMIC
      MAX_SCF 300
      EPS_SCF 1.0E-6
      &OT
        PRECONDITIONER FULL_SINGLE_INVERSE
        MINIMIZER CG
      &END OT
      &OUTER_SCF
        MAX_SCF 300
        EPS_SCF 1.0E-6
      &END
    &END SCF
    &XC
      &XC_FUNCTIONAL
        &LIBXC
          FUNCTIONAL XC_HYB_GGA_XC_WB97X_D
        &END LIBXC
      &END XC_FUNCTIONAL
    &END XC
  &END DFT
  &SUBSYS
    &CELL
      ABC [angstrom] 9.71 9.71 9.71
      PERIODIC XYZ
    &END CELL
    &TOPOLOGY
      COORD_FILE_NAME methane.xyz
      COORD_FILE_FORMAT XYZ
    &END
    &KIND H
      ELEMENT H
      BASIS_SET TZVP-MOLOPT-GTH
      POTENTIAL GTH-PBE-q1
    &END KIND
    &KIND C
      ELEMENT C
      BASIS_SET TZVP-MOLOPT-GTH
      POTENTIAL GTH-PBE-q4
    &END KIND
  &END SUBSYS
  &PRINT
    &FORCES ON
     FILENAME forces
    &END
  &END
&END FORCE_EVAL

&MOTION
 &GEO_OPT
   OPTIMIZER BFGS
   MAX_ITER  1000
   MAX_DR    [bohr] 0.003
   &BFGS
   &END
 &END
  &MD
   ENSEMBLE NPT_I
   STEPS 15000
   TIMESTEP 0.5
   TEMPERATURE 300
   &BAROSTAT
    PRESSURE 400
   &END BAROSTAT
   &THERMOSTAT
    REGION MASSIVE
    TYPE CSVR
    &CSVR
    TIMECON 50
    &END CSVR
   &END THERMOSTAT
 &END MD
 &PRINT
   &TRAJECTORY
    FORMAT XYZ
   &END
   &CELL
    FILENAME cell
   &END
 &END
&END MOTION

[hut...@chem.uzh.ch]

Hi

there are a few points to address, some minor, some major:

1) one should use a reference cell (CELL_REF section) in order to
keep the cutoff of the grid consistent.

2) wB97XD is defined using an empirical vdW potential. In your input
this part is missing (it is not in libxc!).

3) You can specify empirical vdW corrections in CP2K in the section
CP2K_INPUT / FORCE_EVAL / DFT / XC / VDW_POTENTIAL / PAIR_POTENTIAL
However, the one used in the definition of wB97XD is not implemented.
The Grimme D2 is close, but uses another damping function, see the original
paper of Chai and Head-Gordon.

regards

Juerg Hutter
--------------------------------------------------------------
Juerg Hutter Phone : ++41 44 635 4491
Institut für Chemie C FAX : ++41 44 635 6838
Universität Zürich E-mail: hut...@chem.uzh.ch
Winterthurerstrasse 190
CH-8057 Zürich, Switzerland
---------------------------------------------------------------

-----cp...@googlegroups.com wrote: -----
To: cp...@googlegroups.com
From: "Abedi, Mostafa"
Sent by: cp...@googlegroups.com
Date: 07/29/2020 02:58PM
Subject: [CP2K:13690] Wrong density prediction in NpT AIMD

Hi Everyone,
I've performed NpT AIMD simulations for a system containing 8 CH4 molecules. The system was equilibrated using 2 ps NVT (300K) followed by 2 ps NVE. For some reason, the predicted density keeps decreasing and getting away from the correct value (The correct density is 0.232 g/cm3):

Time evaluation of density for a box of 8 CH4 at 300 K and 400 bar

--
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/CAG%3DTvs3qvhzF0fpPAbpQ4GipBmEi-DEq2v80SKbFFjtxMtufPA%40mail.gmail.com.

[Thomas Kühne]

Dear Mostafa, 

your simulation is way too short. Equilibrating the density needs many dozens of ps of simulation, 

and much more so since you are using the default 1ps time constant for the barostat. My 

suggestion is to reduce that to 100 fs and rerun your simulation for at least 25 ps. Also, you were 

missing to add the VdW contribution of the wB97X XC functional. 

Cheers, 

Thomas

Am 29.07.2020 um 14:57 schrieb Abedi, Mostafa <mostaf...@brown.edu>:

Hi Everyone,

I've performed NpT AIMD simulations for a system containing 8 CH4 molecules. The system was equilibrated using 2 ps NVT (300K) followed by 2 ps NVE. For some reason, the predicted density keeps decreasing and getting away from the correct value (The correct density is 0.232 g/cm3):

    Time evaluation of density for a box of 8 CH4 at 300 K and 400 bar

--
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/CAG%3DTvs3qvhzF0fpPAbpQ4GipBmEi-DEq2v80SKbFFjtxMtufPA%40mail.gmail.com.

==============================

Thomas D. Kühne

Dynamics of Condensed Matter

Chair of Theoretical Chemistry

University of Paderborn

Warburger Str. 100

D-33098 Paderborn

Germany

tdku...@mail.upb.de

+49/(0)5251/60-5726

[Abedi, Mostafa]

Dear Juerg and Thomas,

Thank you so much for your helpful suggestions. I will try to modify my input file accordingly and repeat the calculation.

Best,

Mostafa

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/66482C44-5D0D-4C5D-9380-5F31D08F4988%40gmail.com.