I have recently started using LAMMPS for investigating the translocation dynamics of polymers. The polymer is considered as a chain of beads. The excluded volume interaction between the beads is modeled using the WCA potential and the connectivity of the beads is modeled using the Finite extension nonlinear elastic (FENE) potential.
I found that the bond_style FENE in Lammps considers both the potential as a single potential. I tried to implement this potential but got the following message in the log file.
Attached please find the geometry and the input file of my set up along with the log file.
The reason the two parts are together in the FENE bond, is b/c
one part is repulsion, the other is attraction. Together they make
what you would normally consider a "bond". You are only getting
WARNINGS in the file, b/c the initial bonds are dangerously long,
near the start of the run, so FENE is telling you your geometry is suspect.
The attractive part does after all blow up at a finite extension.
Then the system appears to run fine.
Thanks Steve for the reply. I still following queries.
1. If I run the simulation long enough, I should be able to get physically
consistent equilibration structure of the polymer right.
2. I would like to know if it is necessary to always specify pair_coeff
between any atom. For instance, in my system there is no interaction
between the wall atoms. Hence, I decided not to give any pair_coeff and
pair_style between the wall atoms. However, I got the following error "Not
all pair_coeff set". I decided to implement the pair_coeff 1 1 1 1
1.122462048 (using lj/cut pair_style) and then set_force on every atom to
be zero using the following command
fix 3 wall setforce 0.0 0.0 0.0. (where wall is group denoting the wall
atoms).
I would like to know if this statement overwrites the previous statement
and ensures that
there is no interaction between the wall atoms.
3. As I am treating the bead-bead interaction using bond_style potential,
should I still have to specify the pair_style and pair_coeff between the
bead atoms. As I found that, if I do not specify them I get the same error
as mentioned above. Hence, I added a the following command lines before the
bond_style.
Thanks Steve for the reply. I still following queries.
1. If I run the simulation long enough, I should be able to get physically
consistent equilibration structure of the polymer right.
2. I would like to know if it is necessary to always specify pair_coeff
between any atom. For instance, in my system there is no interaction
between the wall atoms. Hence, I decided not to give any pair_coeff and
pair_style between the wall atoms. However, I got the following error "Not
all pair_coeff set". I decided to implement the pair_coeff 1 1 1 1
1.122462048 (using lj/cut pair_style) and then set_force on every atom to
be zero using the following command
fix 3 wall setforce 0.0 0.0 0.0. (where wall is group denoting the wall
atoms).
I would like to know if this statement overwrites the previous statement
and ensures that
there is no interaction between the wall atoms.
3. As I am treating the bead-bead interaction using bond_style potential,
should I still have to specify the pair_style and pair_coeff between the
bead atoms. As I found that, if I do not specify them I get the same error
as mentioned above. Hence, I added a the following command lines before the
bond_style.
Thanks Steve for the reply. I still following queries.
1. If I run the simulation long enough, I should be able to get physically
consistent equilibration structure of the polymer right.
yes
2. I would like to know if it is necessary to always specify pair_coeff
between any atom. For instance, in my system there is no interaction
between the wall atoms. Hence, I decided not to give any pair_coeff and
pair_style between the wall atoms. However, I got the following error "Not
all pair_coeff set". I decided to implement the pair_coeff 1 1 1 1
1.122462048 (using lj/cut pair_style) and then set_force on every atom to be
zero using the following command
If you are using a pair style, then you have to specify interactions
between all type pairs. But you can use epsilon = 0.0 if you want
to turn off an interaction. Or the neigh_modify exclude command.
fix 3 wall setforce 0.0 0.0 0.0. (where wall is group denoting the wall
atoms).
I would like to know if this statement overwrites the previous statement and
ensures that
there is no interaction between the wall atoms.
This sets the final force on wall atoms to 0.0. So it effectively wipes
out any previously computed forces.
3. As I am treating the bead-bead interaction using bond_style potential,
should I still have to specify the pair_style and pair_coeff between the
bead atoms. As I found that, if I do not specify them I get the same error
as mentioned above. Hence, I added a the following command lines before the
bond_style.
The pair interaction (LJ term) between 2 bonded atoms in FENE is
computed by the bond style. If you want bead-bead interactions
between bead pairs that are not bonded, then you need the pair style.
I would like to know, how do we specify an external force (F_ext = F.x) acting on the beads, where F is the external force strength exerted on the monomers in the pore and x is the unit vector along the pore axis.
I would like to calculate the average velocity for the group of atoms/beads. I found that dump ID group_ID custom filename vx vy vz would dump the velocity for every atom at a given time step and not the velocity of a group of atoms. Kindly provide your inputs for the same.
I would like to rephrase my previous question. I would like to calculate the average velocity V_avg = (sum(V_every_atom)/Number of atoms) at every time step. Kindly let me know how we could incorporate this in Lammps.
I am using bond_style fene potential to model the dynamics of a polymer chain. The simulated chain consists of 100 beads. The simulation works fine in a serial machine. However, I encounter the following error while running it in parallel.
ERROR on proc 9: Bond atoms 68 69 missing on proc 9 at step 440 (neigh_bond.cpp:55)
application called MPI_Abort(MPI_COMM_WORLD, 1) - process 9
WARNING: FENE bond too long: 440 48 49 1.98461 (bond_fene.cpp:90)
WARNING: FENE bond too long: 440 61 62 2.01992 (bond_fene.cpp:90)
WARNING: FENE bond too long: 440 43 44 1.93595 (bond_fene.cpp:90)
WARNING: FENE bond too long: 440 60 61 1.91492 (bond_fene.cpp:90)
WARNING: FENE bond too long: 440 59 60 1.97461 (bond_fene.cpp:90)
WARNING: FENE bond too long: 440 56 57 1.91859 (bond_fene.cpp:90)
WARNING: FENE bond too long: 440 55 56 1.94351 (bond_fene.cpp:90)
I found that the issue pops out only when I simulate long chains, while the parallel simulations were working fine for a chain of 60 beads.