How can I calculate thermal expansion coefficient from cross-correlations between volume and enthalpy?

Sorry, I made a typo.

The initial simulation box is 2.552^3 nm^3, not 2.552 nm^3. This equals to a density of 0.9 g/cm^3.

In this system, the calculations of isothermal compressibility and heat capasity at constant pressure are in good agreement with GROMACS.

And to make sure of your comment, I changed the system to 1000 water molecules in 3.2153^3 nm^3 box. (0.9 g/cm^3)

However, the AlphaP is still negative values of -0.003/K.

Perhaps this system is still too small to calculate the fluctuations?


Hmm. 1000 is still not very big and volume fluctuations could still be very big but there might be something else systematically going wrong. Try increasing the size once more, if you keep getting the same value something else might be wrong.
Do the dynamics of the system look good? No drift in any thermodynamic properties or anything? I’m afraid I cannot be of much help beyond the obvious things, I’m no expert in determining expansion coefficients. :frowning:

Dear Stephan,

Thank you for your reply.

I tried simulations of 10,000 water molecules in 6.9271^3 nm^3 box.(The version is lammps-30Jul16)

It seems that the thermodynamic properties I checked was well conserved during 2ns simulations, but unfortunately, the values of AlphaP were not improved and converged at -0.0029/K.(Just in case, I attached the log file.)

I don’t know what to do, but thanks anyway!

Best regards,


lmp_tmp.log (193 KB)