Interfacial Energy calculation

LAMMPS LAMMPS Mailing List Mirror
1

Dear LAMMPS Users,

I wanted to calculate the interfacial energy between two slabs.

Therefore, I made two groups (i.e. top and bottom)

variable zmiddle equal 101.868467157

region topregion block INF INF INF INF ${zmiddle} INF units box

region topregion block INF INF INF INF 101.868467157 INF units box

group top region topregion

4480 atoms in group top

group bottom subtract all top

4480 atoms in group bottom

After that, I tried two methods to calculate interfacial energy

Einterface = Etotal – Etop – Ebottom

(In my system vdw and long range coulombic interactions are the only energy contributions. I used Buck/coul/long)

First one was to use potential energy calculation.

compute topperatom top pe/atom pair kspace

compute toppe top reduce sum c_topperatom

compute bottomperatom bottom pe/atom pair kspace

compute bottompe bottom reduce sum c_bottomperatom

However, when I checked the values

pe = c_toppe + c_bottompe

Therefore, pe – c_toppe – c_bottompe = 0.

Second one was to use group/group command

compute einter bottom group/group top pair yes kspace yes boundary yes

c_inter had finite value, which is different from the first case.

Why do they show different results? Which one is what I want to calculate??

Thank you very much.

Best regards,

Sehun Joo.

2

Dear LAMMPS Users,

I wanted to calculate the interfacial energy between two slabs.

Therefore, I made two groups (i.e. top and bottom)

variable zmiddle equal 101.868467157

region topregion block INF INF INF INF ${zmiddle} INF units box

region topregion block INF INF INF INF 101.868467157 INF units box

group top region topregion

4480 atoms in group top

group bottom subtract all top

4480 atoms in group bottom

After that, I tried two methods to calculate interfacial energy

Einterface = Etotal – Etop – Ebottom

(In my system vdw and long range coulombic interactions are the only
energy contributions. I used Buck/coul/long)

First one was to use potential energy calculation.

compute topperatom top pe/atom pair kspace

compute toppe top reduce sum c_topperatom

compute bottomperatom bottom pe/atom pair kspace

compute bottompe bottom reduce sum c_bottomperatom

However, when I checked the values

pe = c_toppe + c_bottompe

Therefore, pe – c_toppe – c_bottompe = 0.

​yes. this is correct for what you computed, but your Etop and Ebottom are
not what is meant in the formula above. they are to be computed *without*
the other part of the system present. when you do what you do here, Etop
and Ebottom still includes the interactions between Etop and Ebottom.

Second one was to use group/group command

compute einter bottom group/group top pair yes kspace yes boundary
yes

c_inter had finite value, which is different from the first case.

Why do they show different results? Which one is what I want to calculate??

​they have different results, because they compute different things.​
​the situation is a bit more complicated due to using long-range
electrostatics.
so simply doing 3 calculations (only top, only bottom, both slabs) is
likely having some systematic problems.

the simplest way to get the information you want is to *first* do a
simulation with long range electrostatics, so you​ get proper forces and a
decent trajectory. record every 1000 frames or so.

then you change the pair style to replace the long-range coulomb with
cutoff coulomb and increase the coulomb cutoff until it is converged
(remember to watch your boundary conditions. with this setup, you can then
use the "rerun" command, retrace that trajectory and should get the
interface energy directly from using compute group/group or through
manipulating the neighbor list that all interactions *inside* each slab are
removed.

HTH,
     axel.