Elastic constant calculation result is not consistent with temperature

LAMMPS LAMMPS General Discussion
1

Dear LAMMPS developers and users,

I am performing just normal elastic constants MD calculation for zinc blende cubic phase silicon carbide (SiC) structure at finite temperatures.

The simulation flow is: NPT equiliibration → NVT equilibration → deformed box NVT averaging in 6 directions.

However, the result showed that C11 is 384 at 300 K, whereas 405 GPa at 600 K.

The elastic constant increased with increasing temperature.

The volume certainly increased at 600 K when compared to 300 K in ‘elastic_constant.dat’ result file, but weirdly, the elastic constant is increased.

Also, it is notable that the pressure is fluctuating a lot during NPT equilibration.

Could you find any error or reason for this absurd elastic constant result?

I attached .zip file of the input and output files.

Thank you for your help.

Best regards,

elastic_constants.zip (134.6 KB)

2

Sorry, but it is unlikely that anybody will have the time to revise your inputs for you.

Some points to note:

  • pressure fluctuations are large for condensed systems since they are not very compressible. that carries over to fix npt where small volume changes will result in large changes in pressure. this requires careful averaging for both the determination of the actual average pressure and volume. please also note that those fluctuations decrease with larger system size.
  • related to the previous issue: you have to make certain that you use a properly converged volume for the second step of your calculation. It would help to do multiple runs with decorrelated starting geometries to get a measure of the error bars in your results.
  • there are commented examples in the LAMMPS distribution for the same kind property calculation. try comparing your workflow with those and detect possible inconsistencies