discuss about how to calculate the volume change on melting

LAMMPS LAMMPS Mailing List Mirror
1

Hello, everyone!
It’s maybe not a suitable theme for this user lists, but I am confused. Hope someone can give me some suggestions.

What I need to calculate are the melting point and the volume change on melting of bcc-Fe.

At the very beginning, I estimated the melting point by increasing the temperature in NPT ensemble and monitor the volume change or MSD (method1). A sharp increase of the volume change or MSD corresponds to the melting point. The volume change can be obtained simultaneously.

But the referee said it was not an accurate method to obtain the melting point for over-heating. He said the truth. What I calculated is about 2200K, which was higher than the experimental value, say 1811K.

Though he didn’t mention any accurate methods, I found the two-phase method was a better method (method2). I re-calculated the melting point using solid-liquid interface / NVE method and obtained the result 1870K, which was indeed better.

But I don’t know how to obtain the volume change based on method2 because the volume remained unchanged during the simulation. I don’t want to use the volume change based on method1 for its mis-estimating the melting point.

Regards!

2

This is an MD Q, not a LAMMPS Q. If solid/liquid

with NVE works, then there will be no volume
change b/c it is NVE. Since you got what
you wanted (a more accurate estimate of Tmelt),
I don’t see why you care about volume changes.

Steve

3

This is an MD Q, not a LAMMPS Q. If solid/liquid
with NVE works, then there will be no volume
change b/c it is NVE. Since you got what
you wanted (a more accurate estimate of Tmelt),
I don't see why you care about volume changes.

in any case, it should be straightforward to get this from here. it is
only important to keep in mind that simulations often allow for other
approaches than experiments, ...and sometimes doing things in two
steps is much easier and simpler than doing it all at once. so...

since there *is* a hysteresis around Tmelt, and you *do* know the
melting point, you can simply take an all-solid nve system and the
all-liquid nve system and do a simulation at Tmelt for each of them
and average over the volume. take the difference of those two volumes
(and don't forget to consider the error bars!!), and you have what you
ask for.

too obvious?

axel.