Dislocation Minimization

LAMMPS LAMMPS Mailing List Mirror
1

Hi LAMMPS Users,

I am trying to construct a dislocation by positioning the atomic planes such that a dislocation will be generated after the minimization at the centre of the box. I can create the geometry which looks good in visualization as well. I need to minimize the energy while allowing the box to expand.
I am using box/relax along with minimize. I intend to impart strain in successive quasistatic steps after the minimization. I am also fixing the top and bottom slab of few atoms each and constraining them by fix setforce 0 0 0.

When I run the minimization, the temperature and pressure in thermodynamic output doesnot corrrespond to 0 K and 0 pressure. I was wondering whether it is a consequence of constraining by setforce, or it is something else? I doubt it as a consequence of setforce. If it is, does it make sense to use different compute id of temperature and pressure corresponding to nonfrozen atoms? What does that signify physically? And while straining it, temperature and pressure corresponding to which group of atoms would be more sensible.

Attaching the input file for your perusal

variable TEMP equal 0.0
variable LATSCALE equal 1.0
variable NINIT equal 9

variable XTOP1 equal 1
variable XTOP2 equal 1
variable XTOP3 equal 1

variable YTOP1 equal 1
variable YTOP2 equal -1
variable YTOP3 equal 0

variable ZTOP1 equal 1
variable ZTOP2 equal 1
variable ZTOP3 equal -2

variable unitX equal 1.73205*0.5
variable unitY equal 1.41421
variable unitZ equal 2.44950

variable SIZEX equal 80
variable SIZEY equal 80
variable SIZEZ equal 12

variable HALFX equal {SIZEX}*0.5 variable HALFY equal {SIZEY}*0.5
variable LENGTHX equal ${SIZEX}*1.0001

variable LENGTHY equal {SIZEY}*1.00 variable LENGTHZ equal {SIZEZ}*1.0001
variable UPPERX equal ${SIZEX}*1.000001

variable LOWERX equal ({SIZEX}-1)*1.0001 variable FINALX equal ({SIZEX}-0.5)*1.000001

variable sclUPP equal {FINALX}/{UPPERX}
variable sclLOW equal {FINALX}/{LOWERX}

variable HALFLX equal 0.5*{LENGTHX} variable HALFLY equal 0.5*{LENGTHY}*1.0000001

variable frozenslab equal 10

variable latparam equal 2.85531344*${LATSCALE}

variable cutoffpctg equal 0.7
variable delcutoff equal 0.5sqrt(3){latparam}*{cutoffpctg}

variable NCLUST equal 1
variable DISTY equal 0.0
variable DISTX equal 30.0
variable helrad equal 0.5
variable SIZECLUST equal 3

log LOG.create.CLUSTER.EDGE.dislocation.SHEAR.INIT.{NINIT}.{SIZEX}.{SIZEY}.{SIZEZ}.txt

clear
units metal
dimension 3
boundary p p p
atom_style atomic
box tilt large
lattice bcc {latparam} spacing {unitX} {unitY} {unitZ} orient x {XTOP1} {XTOP2} {XTOP3} orient y {YTOP1} {YTOP2} {YTOP3} orient z {ZTOP1} {ZTOP2} ${ZTOP3}

region box prism 0 {LENGTHX} 0 {LENGTHY} 0 ${LENGTHZ} 0 0 0 units lattice
create_box 2 box

region uppbox prism 0 {UPPERX} 0 {HALFLY} 0 {LENGTHZ} 0 0 0 units lattice region lowbox prism 0 {LOWERX} {HALFLY} {LENGTHY} 0 ${LENGTHZ} 0 0 0 units lattice

create_atoms 1 region uppbox
create_atoms 1 region lowbox
group upper region uppbox
group lower region lowbox

pair_style eam/fs
pair_coeff * * FeHe-with_radcrit_0.50.eam.fs Fe He

neighbor 2.0 bin
neigh_modify delay 10 check yes

delete_atoms overlap ${delcutoff} all all

compute poteng all pe/atom

thermo 1000
thermo_style custom step pe ke etotal temp pxx pyy pzz press vol lx ly lz

#dump dump2 all xyz 1 XYZ.create.EDGE.dislocation.*.xyz

variable xupp atom x*({sclUPP}-1) variable xlow atom x*({sclLOW}-1)

fix 1 upper move variable v_xupp NULL NULL NULL NULL NULL
fix 2 lower move variable v_xlow NULL NULL NULL NULL NULL
run 1
unfix 1
unfix 2

change_box all x final 0 {FINALX} units lattice delete_atoms overlap {delcutoff} all all

variable LENGTHXX1 equal lx
variable LENGTHYY1 equal ly
variable LENGTHZZ1 equal lz

variable lowery equal ylo
variable uppery equal yhi

variable freezeLO equal {lowery}+{frozenslab}
variable freezeUP equal {uppery}-{frozenslab}

region freeze1 prism EDGE EDGE {lowery} {freezeLO} EDGE EDGE 0 0 0 units box
region freeze2 prism EDGE EDGE {freezeUP} {uppery} EDGE EDGE 0 0 0 units box
group frozenlo region freeze1
group frozenup region freeze2
region freeze union 2 freeze1 freeze2
group frozen region freeze

include adding.clusters.N${NCLUST}.size${SIZECLUST}.init${NINIT}.y[{DISTY}].x[{DISTX}].in

group helium type 2
group nonfrozenhel subtract all frozen
group nonfrozen subtract nonfrozenhel helium

fix freezeatoms frozen setforce 0.0 0.0 0.0

dump dump1 nonfrozenhel cfg 10000 DUMP/CFG.create.CLUSTER.EDGE.INIT${NINIT}.NCLUST${NCLUST}.DISTY${DISTY}.DISTX${DISTX}.*.cfg mass type xs ys zs id
dump_modify dump1 element Fe He

thermo 100
thermo_style custom step pe ke etotal temp pxx pyy pzz press vol lx ly lz
fix 1 all move variable NULL NULL NULL NULL NULL NULL
run 1
unfix 1

fix 1 all box/relax aniso 0.0
min_style cg
minimize 1e-15 1e-15 50000 100000

2

I don’t understand why you are referring to temperature.
Minimization does not read or write atom velocities.
It ignores them. If the velocities are 0.0, it won’t
change the result of a minimization and the printed
temp/ke will be 0.0. And they all not contribute
(a kinetic term) to the pressure. If you set
them to non-zero for some reason (why?) then
they will contribute to the pressure.

Steve

3

Steve and Shafat,

Thanks for replying. I would also assume the temperature to be 0, but there is a constant term of kinetic energy and thus temperature as well, which is the main spurce of confusion. I am not giving the system any kinetic energy, which is why i wonder if this constant temperature of kinetic energy is a consequence of fixed atom slabs at the top and bottom of the box?

Thanks
Ayush

4

Steve and Shafat,

Thanks for replying. I would also assume the temperature to be 0, but there is a constant term of kinetic energy and thus temperature as well, which is the main spurce of confusion. I am not giving the system any kinetic energy, which is why i wonder if this constant temperature of kinetic energy is a consequence of fixed atom slabs at the top and bottom of the box?

I don’t see how frozen atoms could be able to generate temperature?

From your simulation parameters, I can see that you do at least two “run 1” before the minimize. These run compute forces AND then attribute the corresponding velocity to the atoms. So, that is normal to have a velocity in your system… after 2 steps of dynamics.

To be sure to not have any temperature BEFORE a minimize with CG, you should set the velocity to 0 with “velocity”

Julien

Thanks
Ayush

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! [http://sdm.link/slashdot](http://sdm.link/slashdot)

_______________________________________________
lammps-users mailing list
[[email protected]](mailto:[email protected])
[https://lists.sourceforge.net/lists/listinfo/lammps-users](https://lists.sourceforge.net/lists/listinfo/lammps-users)
5

Hi Folks,

Can anyone shed some light into the excess temperature in thermodynamic output. Is it due to internal forces in the system as a consequence of fix setforce 0,0,0 during minimization?

Regards
Ayush

6

Your input script is complicated. I suggest you dump
atoms (and their velocities) at various points in your
script or runs and see which atoms have velocities.
If LAMMPS is reporting a finite temperature or
kinetic energy, than assuredly there are atoms
with non-zero velocities.

Steve

7

Thanks Steve for the reply.

Changing the box was resulting in assignment of initial velocity to some atoms. Since I wanted zero velocity, It worked by initializing the velocity of all the atoms at zero temperature before minimization as suggested in http://lammps.sandia.gov/doc/fix_box_relax.html.

Regards
Ayush