Melting Metallic Particles on Graphene

Dear All
Hi

I am trying to melt metallic nanoparticles on graphene and compare them with free (without graphene state).
The model is visualized here:
https://ibb.co/b9mXPv

(the graphene is periodic and particles are about 5-7.7 nm)

I did this before (3 years ago) and results were reasonable.
( Attached Here: https://ibb.co/mrF94v )
But recently i tried to produce the same results and i cant, even with the same input.

I Plot the Potential Energy - Temp. for determining melting points and the results are nearly the same. and the curves are very similar: both Free Particle (Without Graphene) and Graphene Supported Supported
I uploaded the new curves here:
https://ibb.co/eD6Gjv

I test this for many inputs like different metals and even core-shells.
I think the problem is with my input
I attached it here:
https://www.4shared.com/office/v_cTPQtAca/Pt-G.html

and it is written bellow:

dimension 3
units metal
atom_style atomic
boundary p p f

read_data final-Pt.data
mass 1 195.084
mass 2 12

pair_style hybrid eam airebo 3.0 lj/cut 6
pair_coeff 1 1 eam Pt_u3.eam
pair_coeff * * airebo CH.airebo NULL C
pair_coeff 1 2 lj/cut 0.1365 1.6 6

group graphene type 2
group particle type 1

velocity all create 298 8927079

neighbor 2 bin

neigh_modify every 20 delay 0 check no

compute 6 particle pe/atom
compute 7 particle reduce sum c_6
compute 8 particle temp
compute 10 all temp
compute 11 graphene temp

thermo_style custom step temp pe c_8 c_7 c_10 c_11

dump 1 all atom 250 dump-allheat1.atom
timestep 0.003

thermo 25

fix 1 all nvt temp 298 298 0.1

run 1000000

unfix 1
fix 2 particle nvt temp 298 2000 0.01
run 5000000

unfix 2
fix 3 particle nvt temp 2000 2000 0.1
run 1000000

for Free state it is simple and i plot the potential energy vs temp.
for the graphene supported state, I compute potential energy and temp of particle (compute 7 & 8) and plot them.
But i think i am making a mistake here.
because when i plot the total potential energy (including grahene atoms) vs temp i got the transition temp very different and i attached the picture here (it is the new blue curve):
https://ibb.co/f0ALBa

Another Important Problem:
I used NVT for All particles just in relaxing in 298K temp.
During heating from 298, i can only heat the particle,
if i use NVT for all (in fix 2) the graphene will explode after some time and i will face LOST ATOMS error…
i really dont know how to heat particle and graphene together

could you please tell me what is the proper way to heat a particle on graphene??
and the correct way of computing potential energy and ploting PE-T curve?
(I tried different Cut offs different damping for NVT and everything… I know i am making a mistake in computing)

Sorry for taking your time and my awful English.

Thank you all

Regards
Rosita

In which sense can you not reproduce the same results? Do the pe-T plots look different?

As for the lost atoms, assuming the dynamics are not unstable, it could be that if you heat both the graphene and metal, some atoms cross the z-boundary, which is “f”, so you will lose them.
To prevent this you could either use fully periodic boundaries in z or put a wall underneath the sheet or something.