ERROR on proc 0: Particle on or inside fix wall surface initializing fcc

Hello everyone ,

How to avoid placing particles on a wall if I want to use an initial order with an FCC lattice ?

an extract of code:

fcc lattice 0.6
box block region 0 0 and x 0 20
create_box 1 box
create_atoms 1 box
.
.
fix zmuro all wall/lj126 zlo EDGE 1 1 1.5 zhi EDGE 4 4 2.5 units lattice

if I use :
create_atoms 1 random 800 123 NULL

works well , however you wish initial simulation with a crystalline structure. I have to map the position of atoms manually or is there another option ?

regards

Hello everyone ,

How to avoid placing particles on a wall if I want to use an initial order
with an FCC lattice ?

an extract of code:
fcc lattice 0.6
box block region 0 0 and x 0 20
create_box 1 box
create_atoms 1 box
.
.
fix zmuro all wall/lj126 zlo EDGE 1 1 1.5 zhi EDGE 4 4 2.5 units lattice

if I use :
create_atoms 1 random 800 123 NULL

works well , however you wish initial simulation with a crystalline
structure. I have to map the position of atoms manually or is there another
option ?

from your message it is not very clear what you are asking and having
only such a minimal input fragment is not helping either.
however, one thing is obvious. by placing your system box *exactly* on
the lattice points, you are asking for trouble. this has been
discussed on this very mailing list numerous times (and should be
explained in the manual as well). the resolution is very simple. shift
your box by a small increment (e.g. 0.1 units) and all the ambiguity
and problems from having atoms placed on box boundaries should become
non-issues.

axel.

The wall issue is that the force/energy blows
up if the particles are on the wall b/c the
distance from the wall r is 0 and the
lj12/6 is 1/r^12. So even if you put the
particles epsilon away from the wall the
force will be huge. You really should setup
your problem so the particles are roughly sigma
away from the wall. It’s fine to put the wall
outside the simulation box, e.g. by sigma.

Steve

Hello axel and Steve , thanks for the replies .

Have boundary conditions f in the z axis , if the walls 0.3 move from the edges of the simulation box , have lost of atoms .
Lost Atoms : current 7836 8000 Original ( …/thermo.cpp:393 ) .

if I not have lost of atoms then proc 0: Particle fix on wall surface or inside .

What am I doing wrong?

Test 1

units lj
atom_style atomic
dimension 3
boundary p p f

lattice fcc 0.1
region caja block 0 10 0 10 0 20
create_box 1 caja
create_atoms 1 box

mass 1 9
velocity all create 1.44 87287 loop geom

PAIR

pair_style lj/cut 2.5
pair_coeff 1 1 1.0 1.0 2.5
neighbor 0.3 bin
neigh_modify delay 5 every 1

fix 1 all nve
fix zmuro all wall/lj126 zlo -0.3 1 1 1.5 zhi 20.3 4 4 2.5 units lattice

thermo 1000
thermo_style custom step pe etotal press temp

run 10000

Hello axel and Steve , thanks for the replies .

Have boundary conditions f in the z axis , if the walls 0.3 move from the edges of the simulation box , have lost of atoms .
Lost Atoms : current 7836 8000 Original ( …/thermo.cpp:393 ) .

if I not have lost of atoms then proc 0: Particle fix on wall surface or inside .

What am I doing wrong?

Like I already mentioned, you put atoms exactly on the box boundary. That is asking for trouble. Bad setup -> bad dynamics. Shift your box or use a separate, smaller region for creating the atoms, and these trivial issues should be gone.

Axel

hello,

Effectively solve the problem was trivial . use a separate region for positioning atoms .
The code shown below and which highlights the change:

Test 1

units lj
atom_style atomic
dimension 3
boundary p p f

lattice fcc 0.1
region caja_simulacion block 0 10 0 10 0 20
region region_atomos block 0 10 0 10 1 19
create_box 1 caja_simulacion
create_atoms 1 region region_atomos

mass 1 9
velocity all create 100.44 87287 loop geom

PAIR

pair_style lj/cut 2.5
pair_coeff 1 1 1.0 1.0 2.5
neighbor 0.3 bin
neigh_modify delay 5 every 1

fix 1 all nve
fix zmuro all wall/lj126 zlo 0 1 1 3.5 zhi 20 1 1 2.5 units lattice

thermo 1000
thermo_style custom step pe etotal press temp

run 10000

Thanks

Christopher