Deformation of a polymer

Dear LAMMPs developers/users,

I am deforming a single polymer chain with the following command:

fix 3 all deform 1 x vel 15 units box

Now I am not sure why my chain is deforming at all- I am only changing the x direction of my box and that works fine, but because the chain is not squeezed in y and z directions, I cannot understand where the deformation comes from in this situation- the chain is not "attached" to the box in x direction... Also, for some additional information, I am using p f f boundaries in this simulation.

Thank you a lot in advanced!
With best wishes,
Anna

Dear LAMMPs developers/users,

I am deforming a single polymer chain with the following command:

fix 3 all deform 1 x vel 15 units box

Now I am not sure why my chain is deforming at all- I am only changing
the x direction of my box and that works fine, but because the chain is
not squeezed in y and z directions, I cannot understand where the
deformation comes from in this situation- the chain is not "attached" to
the box in x direction... Also, for some additional information, I am
using p f f boundaries in this simulation.

but due to the periodic boundary, the chain will be in contact with
its periodic images, right?

axel.

Hi,

wouldn't it be better to maintain fixed one end of polymer, then apply
a constant speed along the loading direction at the free end? . The
stress can be calculated using stress/atom

A salute
Oscar G.

Hi,

Wait i forgot to mention something else ... (this happens because i'm
a lazy individual that doesn't read carefully all the information in
your email)

Anna Lappala wrote:
Also, for some additional information, I am using p f f boundaries in
this simulation.

COMMENT:
Do not use a periodic box at all, and definitely not in the length
direction of your polymer chain. You want free surfaces to mimic the
surface effects.

A salute
Oscar G.

Thanks a lot Axel and Oscar!
With bet wishes,
Anna

Dear Axel & Oscar,

I am still running into problems with the deformation of a polymer chain. Namely, I have now changed all boundaries to fixed, and the simulation simply crashes, even without deformation. I had a polymer chain initially that I shrink-wrap with s s s boundaries, and when I after writing a new data file from a restart file, use fixed boundaries, the same simulation of a collapsed chain without deformation crashes. I tried to manually increase box size and the simulations runs fine, but still crashes after a while if I use fix deform. The simulation only works if I decrease timestep to 1*10^-8, and still lammps prints out messages about fene bond being too long... I am not sure how to fix this problem as I do not think that by reducing my timestep from 0.01 to 1*10^-8 i am doing the right thing.

I appreciate any suggestions and thank you in advanced!
Anna

Hi,

I’m sure there are many references in the literature that talks about the deformation of polymer chains (please Read) …Ammmm… I have some experience dealing with deformation of Metalic nanowires and I have never suffered a Crash issue . Please check again your initial conditions and your computational box boundaries …It is impossible to know what you might doing be wrong.

ANNA WROTE:
I am not sure how to fix this problem as I do not think that by reducing my timestep from 0.01 to 1*10^-8 i am doing the right thing.

COMMENT:
What time units are you using for the time step ? .LAMMPS uses different time UNITS : http://lammps.sandia.gov/doc/units.html . e.g If you are using SI Units , then your time is measured in seconds and thus your 0.01 timestep is equal to 0.01 seconds …(that’ tooo high) … You have to understand that one of the constraints in MD simulations is the timescales i.e You can only simulate highstrain rates (> 10^8 s^-1) …Unless you have acces to a High Perfomance computers, because trying to run Deformation in the microseconds time scale 10^3 s^-1 would required ALOT of time-steps or either you can use “Accelerate molecular dynamics algorithms”

A salute
Oscar G.

A couple of comments:

1. fix deform can only be applied to periodic dimensions, so you
should have gotten a LAMMPS error message when you tried to deform
with s s s boundaries.

2. In your original simulation, the default is "remap x", so the
entire polymer is rescaled in the deformation direction. If you
specify "remap none," the cell will still shrink in the periodic
direction, but the atom positions will not be adjusted. All atom-atom
distances will be strongly compressed across the periodic boundary.
Neither of these choices seems very physical.

3. You should think about what exactly is the physical situation you
are trying to study and then choose a reasonable simulation protocol
that comes close to that.

Aidan

HI Aidan,

Thanks alot for the info =)
A Salute

Oscar G