# temperature fluctuation range in npt

Dear lammps users,

I’m calculating the melting point of a metal using the solid-liquid coexistence system method in NPT. There are 31360 atoms in my system and I set the temperature of the system to be 1135K. After reaching equilibrirum, I found that during 1000000 steps,the temperature varies between 1125K and 1145K most of the time, and at some points temperature is even beyond this range, with the minima being 1119K and the maximum being 1152K. At 1135K, the potential energy of the system still decreases. I want to get a more precise melting point by setting more temperature points. Could you tell me what the temperature fluctuation range should be with 31360 atoms in my simulation box? I know that there is a relation between the fluctuations in temperature and other properites with the number of atoms, but after checking lammps mailing list and a statistical mechamics book, I didn’t find an exact equation describing the relation between them.

Rui

Dear lammps users,

I'm calculating the melting point of a metal using the solid-liquid
coexistence system method in NPT. There are 31360 atoms in my system and I
set the temperature of the system to be 1135K. After reaching equilibrirum,
I found that during 1000000 steps,the temperature varies between 1125K and
1145K most of the time, and at some points temperature is even beyond this
range, with the minima being 1119K and the maximum being 1152K. At 1135K,
the potential energy of the system still decreases. I want to get a more
precise melting point by setting more temperature points. Could you tell me
what the temperature fluctuation range should be with 31360 atoms in my
simulation box? I know that there is a relation between the fluctuations in
temperature and other properites with the number of atoms, but after
checking lammps mailing list and a statistical mechamics book, I didn't
find an exact equation describing the relation between them.

Dear Rui,

I tried to calculate the melting point before with solid-liquid coexistence method.
The large fluctuation could also come from the metastable liquid state when the setting temperature is close to the melting point.
For normal fluctuation, you can try time-averaging to get your data when potential tends to be stable.
As for decreasing potential energy, you may try longer time/more steps for equilibration stage.
Hope this helps!

Xiao