region melting and the whole system melting

Dear everyone,
I want to heat a region of the whole system and spread heat around according to gradient.But when I used '“fix nph” to the whole system, the temperature of system increased together. If I use '“fix nph” to the group only,the temperature of the region increase quickly,but the surrounding doesn’t have any significant change.
I wonder how to simulate the situation of Copper local melting not the whole system.
My in file is as follows.

units metal
boundary p p p
atom_style full
lattice fcc 3.61
region box block 0 20 0 20 0 5
create_box 1 box
create_atoms 1 box
region 1 block 31 41 31 41 8.5 18.5 units box
region 2 block 31 41 31 41 8.5 18.5 units box side out
group laser region 1
group copper region 2
compute 1 all temp/region 1
timestep 0.001
thermo 1000
thermo_style custom step temp press vol etotal c_1
pair_style eam
pair_coeff * * Cu_u6.eam
neighbor 0.3 bin
neigh_modify every 5 delay 0 check yes

velocity all create 293 825577 dist gaussian
minimize 1.0e-4 1.0e-6 1000 100000
fix 1 all npt temp 293.0 293.0 10 z 1 1 50
run 100000
unfix 1
#add energy in the region, the energy is divided into 100 parts to simulate a 1 ns energy pulse
fix 2 all heat 10000 0.03 region 1
fix 3 all nph iso 1.0 1.0 50
dump 1 all custom 1000 dump123.lammpstrj id type x y z
run 1000000

Thank you.

Dear everyone,
    I want to heat a region of the whole system and spread heat around
according to gradient.But when I used '"fix nph" to the whole system, the

For this problem, I would just use NVE for all without any other fixes.

Ray

Dear everyone,
    I want to heat a region of the whole system and spread heat around
according to gradient.But when I used '"fix nph" to the whole system, the
temperature of system increased together. If I use '"fix nph" to the group
only,the temperature of the region increase quickly,but the surrounding
doesn't have any significant change.
   I wonder how to simulate the situation of Copper local melting not the
whole system.

you are pumping a constant flow of kinetic energy into your system via
fix heat why should it *not* melt? please have a bit of a read on the
process of melting in MD simulation. you have to regard the impact of
(limited) length and time scales in MD simulations. this topic has
been discussed in the past, so a search through the mailing list
archives might help, too. the best way to get a liquid/solid
co-existence is to prepare and equilibrate a liquid and solid sample
separately, bring them to the temperature of the melting point (nota
bene: the MP for the chosen potential matters, not the real material.
the may differ a bit) and then combine the two. with a single
thermostat right at the melting point, the coexistence should be
preserved. if you want a temperature gradient as well, you need two
thermostats at two different temperatures.

again, please note the hysteresis around the process of melting and
crystallizing and the implications on the total internal energy of the
system. it is not as trivial as you might think it is.

axel.

Thank you.
I need to simulate the situation that nanosecond laser ablates copper. The energy pumps in the system isn’t a single continuous pulse via fix heat. I considered this is the problem. How to add a single continuous pulse?

Fix heat can add energy to a region of atoms every
timestep for as short or long as you like.

How is that different than what you want?

Steve

The manual says,N=add/subtract heat every this many timesteps in fix heat. In my understanding,it adds energy per N timesteps.
Then,how to write a fix heat command to simulate 1ns pulse width,6ev energy adding in region(timestep=0.001ps,units=metal)?
If N=1000000,eflux=6,is it right?

As the fix heat doc page explains you specify a heat flux in units of energy/time.
It’s up to you to make that number match the physical model you have
in mind. Asking a mail list to check your math is not a good idea - why
would you believe the answer you get?

Steve

Sorry, I listed a example to make sure that whether my understand for the fix heat doc is right or not. Is the energy add in instantly or lasted a continuous period of time?

The energy is added continuously (every N timesteps
where you specify N), and at the rate you specify.

If you want to turn it “off”, then simply do a run
with fix heat in place. Then unfix the fix. Then
do another run.

Steve