question about melting point

LAMMPS LAMMPS Mailing List Mirror
1

Dear all friend

I want to calculate Gold melting point and I use EAM potential and this script :

units metal
dimension 3
boundary p p p
atom_style atomic
atom_modify map array

# ---------- Create Atoms ---------------------

read_data Au256.data

# ---------- Define Interatomic Potential ---------------------
pair_style eam
pair_coeff * * Au_u3.eam
neighbor 2.0 bin
neigh_modify delay 10 check yes

# ---------- Define Settings ---------------------

compute eng all pe/atom
compute eatoms all reduce sum c_eng

# ---------- Run Minimization ---------------------

reset_timestep 0
fix 1 all box/relax iso 0.0 vmax 0.001
thermo 10000
min_style cg
minimize 1e-25 1e-25 5000 10000

#-----------------dynamic run in new temperature

dump npt all xyz 5000 ionized-npt.xyz

velocity all create 100 45827325 mom yes rot yes dist uniform

fix npt1 all npt temp 100.0 2000.0 0.01 y 1.0 1.0 500
fix 2 all momentum 10 linear 1 1 1 angular
fix 3 all recenter INIT INIT INIT units box

thermo_style multi

timestep 0.001

run 1000000
unfix npt1

undump npt

the gold structure is FCC and it has 256 atom

after run I draw potential energy v.s temperature curve but I see no
any jump in curve that show transition phase

I search in mailing list but i no find any thing

what is your suggestion?

best

Ehsan

2

How does your sample look like? Typically you need half box FCC and half
box random arrangement with the same density - i.e. amorphous liquid like.

Then you run for long time at a constant temperature and observe if the
FCC is slowly growing or shrinking. It tells you, if the temperature is
below or above the melting point, you are looking for.

3

Dear all friend

[...]
[

the gold structure is FCC and it has 256 atom

after run I draw potential energy v.s temperature curve but I see no
any jump in curve that show transition phase

yes, because you are not accounting for finite size effects and the
fact, that melting is an activated process. the method you are trying
to use is seriously flawed. it only works in the macroscopic world
(and even there, you'll have to account for hysteresis effects in some
cases).

I search in mailing list but i no find any thing

what is your suggestion?

search again! and search smarter!!
the topic of determining the melting point of materials from MD
simulations has been discussed repeatedly on this very list and it
also has been explained that coexistence simulations are the proper
way to determine phase transition conditions on the atomic scale.

axel.

4

There is a script in examples/user for calculation of the melting point by two phase method from the work by Starikov S.V., Stegailov V.V. Premelting of iron and aluminum: implication for high pressure melting curve measurements // Phys. Rev. B., 2009, V. 80, P. 220104®.