# difference in transition temperature

I am doing the melting of Al by using npt and nvt fix.In these two cases i have found the transition temperature is different. In npt condition it is around 1080K which is approximately near to experimental melting point but in nvt condition it is around 1500K.Why this much difference is coming and which is the best condition to consider.

I am doing the melting of Al by using npt and nvt fix.In these two cases i have found the transition temperature is different. In npt condition it is around 1080K which is approximately near to experimental melting point but in nvt condition it is around 1500K.Why this much difference is coming and which is the best condition to consider.

you should first explain, which method exactly you have been
using to determine the melting point. depending on that,
people may have some comments. in general, finite size
MD simulations are (logically) plagued by finite size effects.
however, those apply differently to different properties and
how you determine them.

cheers,
axel.

I am using the Energy vs Temperature plot to determine the Transition point.In npt condition the transition point is around 1080K which is approximately near to experimental melting point but in nvt condition it is around 1500K The remaining condition are same for both the cases .Why this much difference is coming and which is the best condition to consider.

I am using the Energy vs Temperature plot to determine the Transition point.In npt condition the transition point is around 1080K which is approximately near to experimental melting point but in nvt condition it is around 1500K The remaining condition are same for both the cases .Why this much difference is coming and which is the best condition to consider.

it is well documented, that in MD simulations it is very difficult to
determine the melting point accurately, due to large hysteresis resulting
from finite size effect. those are _much_ more prevalent in constant
volume simulations than in variable cell simulations as the latter allow
fluctuations and reorientations more easily to occur.

a recommended way to determine a melting point in a finite size system
is by setting up a two phase system (half solid, half liquid) and then run
at different temperatures around the melting point. the melting point is
reached, if the phase boundary doesn't move. if you are over, the liquid
phase will grow, if you are under the solid one will grow.

cheers,
axel.