I am using the delete_atoms command in the input file to create a SiGe nanowire with random placement of Si/Ge atoms.
I would like to have a 110 axially oriented wire (in y-direction) with a rectangular cross-section as specified in the input file. When I run this file with LAMMPS, the code generates about 6802 atoms of Germanium but while executing the delete_atoms command, it delete only 6466 Si atoms (which overlapped the Ge atoms created earlier). this leads to thermodynamic quantities to be printed as NaNs… log is copied below the input commands.
any help to solve this behavior is greatly appreciated.
Thank you.
Anurag
portion of input script is copied below (I am using LAMMPS 21Dec2010 build)
log log.lammps.1
units metal
boundary s s s
processors 2 * 2
atom_style atomic #lattice diamond 5.5466
lattice diamond 5.5466 orient x 1 0 0 orient y 0 1 1 orient z 0 -1 1
region box block 0 5 0 30 0 5
create_box 2 box
create_atoms 2 box
delete_atoms porosity box 0.5 79301682
group sige_ge region box
create_atoms 1 box
group sige_si region box
pair_style sw
pair_coeff * * SiGe.sw Si Ge
neighbor 0.3 bin
neigh_modify delay 2
mass 1 28.0855
mass 2 72.641
deletes overlapping Si and Ge atoms
delete_atoms overlap 0.1 sige_si sige_ge
group sige union sige_si sige_ge
---------log ----------
LAMMPS (21 Dec 2010)
Lattice spacing in x,y,z = 5.5466 7.84408 7.84408
Created orthogonal box = (0 0 0) to (27.733 235.322 39.2204)
1 by 1 by 1 processor grid
Created 13376 atoms
Deleted 6574 atoms, new total = 6802
6466 atoms in group sige_ge
Created 13376 atoms
19506 atoms in group sige_si
System init for delete_atoms …
Deleted 6466 atoms, new total = 13712
13040 atoms in group sige
…
Setting up run …
Memory usage per processor = 5.14022 Mbytes
Step Temp TotEng PotEng Volume Pxx Pyy Pzz Pxy Pxz Pyz Lx Ly Lz
0 276.3037 nan nan 255963.19 nan nan nan nan nan nan 27.7332 235.32251 39.220585
The only way you will debug this is
to visualize what is produced at each
step and verify that the geometries and
atom counts are what you think they should
be. Dump file snapshots with coords
may also help.
There are various parameters to tweak
with respect to creating atoms and deleting
them in such a geometry and you will
need to verify that LAMMPS is doing what
you expect.
I am interested in simulating a nanowire subjected to compressive strain in the axial direction.
I would like to have non-periodic boundary conditions in xyz but I am not able to find a suitable LAMMPS command which would enable me to strain the wire.
To my understanding commands like displace_box or fix box/relax or fix deform can only be used for alter the structure in periodic dimension.
any suggestions/corrections are welcome.
regards,
Anurag
PS: I am sorry of this question is posted a second time. I am having slight problems with my email.
I am interested in simulating a nanowire subjected to compressive strain in
the axial direction.
I would like to have non-periodic boundary conditions in xyz but I am not
able to find a suitable LAMMPS command which would enable me to strain the
wire.
so what you want to do is basically a "tug-o-war" on your nanowire?
same as in that setup, you can apply a force to both ends of
the wire and that can be done using fix setforce or you can impose
movement with fix move.
I would like to get stress-strain curves for different aspect ratio of the nanowire. moving atoms with a certain velocity however, affects the stress-strain curves which change with the applied velocity (strain rate).
I would like to somehow avoid this dependence and have been thinking of using molecular statics. but I am not sure how to setup the problem at present.
I would like to get stress-strain curves for different aspect ratio of the
nanowire. moving atoms with a certain velocity however, affects the
stress-strain curves which change with the applied velocity (strain rate).
I would like to somehow avoid this dependence and have been thinking of
using molecular statics. but I am not sure how to setup the problem at
present.
i honestly don't understand what you are asking about here
and particularly i don't get why you cannot do this with a 1d periodic cell.
to me that would be the perfect way of doing this.
I would like to setup a nanowire simulation – with a “mobile” region and a “boundary” region wherein the boundary atoms are fixed (I use setforce 0 0 0 for boundary region). mobile atoms are allowed to move.
however, I would like to compute the force atoms in the mobile region exert on the boundary atoms. is there a way to compute this force ?
I would like to setup a nanowire simulation -- with a "mobile" region and a
"boundary" region wherein the boundary atoms are fixed (I use setforce 0 0 0
for boundary region). mobile atoms are allowed to move.
however, I would like to compute the force atoms in the mobile region exert
on the boundary atoms. is there a way to compute this force ?
fix setforce should provide this information. you may need to
use two or more fixes depending on how you geometry is set up.
please have a good look at its documentation.
I would like to run a series of fix/unfix commands until one of the thermo outputs becomes less than pre-defined number. I have been trying to use the if-then statements to sort of loop around these fix/unfix statements but LAMMPS (version 25 Aug 2011) reports an illegal IF command error.
I am missing something here or probably if-then construct will not work in this case. I have tried with double/single quotes with similar result.
I used LAMMPS (25 Aug 2011) to simulate Silicon nanowires under free boundary conditions with MEAM and Tersoff potentials, I see that “Tersoff wire” tends to elongate and expand in cross-section while “MEAM wire” tends to contract in length and expands in cross-section. these simulatations are @ very low temperature (< 1K). one axial end of the wire is kept fixed while the system is relaxed.
I am wondering which of these is a physical behavior ? can anyone share some insights or if I can do further checks to choose one potential over another ?
for bulk Si, MEAM gives better properties (compared to expt.) vs. tersoff.
I would like to simulate a situation where two surfaces are progressively brought closer to each other till the separation distance becomes zero.
similar to an AFM experiment, I would like to extract the force as a function of separation distance. I am not sure how else to express this situation another way and would appreciate any pointers to published papers.
any ideas on how to setup this problem in LAMMPS ?
I would like to simulate a situation where two surfaces are progressively brought closer to each other till the separation distance becomes zero.
similar to an AFM experiment, I would like to extract the force as a function of separation distance. I am not sure how else to express this situation another way and would appreciate any pointers to published papers.
any ideas on how to setup this problem in LAMMPS ?
use steered MD in constant velocity mode.
details and references are on the manual page for the fix smd command.
We employed lately LAMMPS for dealing with an adhesion problem between
nanoparticles (I did not update LAMMPS website on my LAMMPS-related
publications, which may be a good idea). You can find some information in the
supplemental material. Mordehai et al. PRL 107, 096101 (2011).