[lammps_users] pressure fluctuation during the global optimisation

LAMMPS LAMMPS Mailing List Mirror
1

Dear lammps_users and Axel,

I am trying to global optimise a phase bulk structure of ZnO at 0.0 bar constant pressure.

However, I notice that during the optimisation pressure fluctuates in every steps.

I have searched the mailing list which said

(1) pressure will fluctuate for smaller system size for MD simulation,

(2) the large fluctuations may be caused by insufficient number of atoms/molecules

The system that I try to optimise/simulate is periodic system and I assume (1) (2) is not the case for what I am trying to calculate.

Could you please give me advices why the pressure fluctuation happens and if the pressure can be fixed then how to do that?

I appreciate any advices.

Thanking you for your time and advices.

Kind regards,

P.S. Here is the input file that I used:

ZnO bulk phase simulation

units metal

atom_style charge

dimension 3

boundary p p p

#box tilt large # allow lammps to have large tilt facotr (xy yz xz)

read_data lammps_coord

replicate 3 3 3 # supercell

pair_style comb3 polar_off

pair_coeff * * ./lammps_ip Zn O

neighbor 2.0 bin

neigh_modify every 1 delay 0 check yes

set type 1 charge 2.0 # Zn

set type 2 charge -2.0 # O

2

Dear lammps_users and Axel,

I am trying to global optimise a phase bulk structure of ZnO at 0.0 bar constant pressure.

However, I notice that during the optimisation pressure fluctuates in every steps.

I have searched the mailing list which said

(1) pressure will fluctuate for smaller system size for MD simulation,

(2) the large fluctuations may be caused by insufficient number of atoms/molecules

The system that I try to optimise/simulate is periodic system and I assume (1) (2) is not the case for what I am trying to calculate.

your assumption is wrong.

the pressure for a non-periodic should always be zero, since the volume of such a system is infinite. it only shows in LAMMPS as a non-zero value, since LAMMPS (arbitrarily) sets box boundaries (for computational efficiency) and thus making the volume finite.

axel.

3

Dear Axel,

Thank you for the reply.

I appreciate the advice and I understood that periodic boundary causes the pressure.

So just to clarify, you are saying there is no way to is no way to achieve the pressure that I set?

I appreciate your advice.

Thanking you for your time and advice.

T.Kang,

PhD Student.

4

i am saying that what has been explained in the many existing e-mails in the e-mail archive stands:

  • all instantaneous values of bulk properties fluctuate in MD simulations, not just pressure
  • fluctuations depend on system size the larger the system the smaller the fluctuations
  • for systems in equilibrium, the time average is equivalent to the system average
  • pressure fluctuates by its nature more than, e.g. temperature
  • pressure fluctuations also depend on the compressibility, which is usually low for condensed systems

if you have trouble understanding any of those relations or conditions, i suggest you read up on statistical thermodynamics in suitable text books.

axel.