I'm having some trouble getting accurate values for the heat capacity using M/EAM and NVT. I'm averaging the total energy output from the thermo dump. I thought this might be an issue with the timing of output vs. the timing of the thermostat, as we're looking at the variance in the energy. I ran some simulations staggering the thermo output and the thermostat, but my energy fluctuations are still extremely low; generating cp's of about 0.5 to 2.5 j/mol-K for copper at 298K. Any thoughts?
Thanks,
Michael Sellers
Graduate Student, Chemical Engineering
University at Buffalo
1 Like
Does it work for EAM or any other pair potential?
Are you saying there is something peculiar about MEAM?
Steve
Both EAM and MEAM are showing extremely small energy fluctuations with NVT. I originally ran both to compare the heat capacities. Is this common for Nose-Hoover?
-Mike Sellers
Steve Plimpton wrote:
Here is Aidan's answer:
Effect: calculated heat capacity too small
Possible causes:
-Force field
-Finite size effects
-Thermostat
-LAMMPS error
-User error
Without more specific information, I don't know which of these (if any) is
likely to be the true cause. But here is a suggested approach
1. Reproduce somebody else's calculation of heat capacity.
2. Do it using LAMMPS
If you can do 1, but not 2, it may be be due to different definitions of
energy (in LAMMPS, you can optionally add in the thermostat energy) or
different thermostats
If you can do 1 and 2, but not for MEAM, it is probably a deficiency in
MEAM.