I’d like to constrain the polymer bond length and calculate the vapor pressure. I know the fix shake isn’t work for this, so I plan to increase the bond coefficient 3-4 times to constrain the bond length and reduce the time step. Indeed, the bond length changes a little with high bond coefficient. However, the vapor pressure is always overestimated and the density is smaller than the reference paper which uses the shake algorithm. Could you give me some suggestions about this?
Hi Lammps users!
I'd like to constrain the polymer bond length and calculate the vapor pressure. I know the fix shake isn't work for this, so I plan to increase the bond coefficient 3-4 times to constrain the bond length and reduce the time step.
This should work.
Keep in mind that as you increase the stiffness of the bond coefficient
Indeed, the bond length changes a little with high bond coefficient.
However, the vapor pressure is always overestimated and the density is
smaller than the reference paper which uses the shake algorithm. Could
you give me some suggestions about this?
What do you mean by vapor pressure?
(I'm guessing you have many polymers in your simulation. Some of them
condense into an aggregate, and some do not. The polymers that don't
contribute to the vapor pressure. Is this correct?)
There could be so many different reasons. It's also possible that
there was a mistake made by the researchers who wrote the reference
paper. (Sadly, I don't think reviewers typically rerun the
calculation to verify the conclusions.) I suggest
1) scrutinize the pair_coeffs you are using. (Make sure your pairwise
non-bonded interactions (eg. Lennard-Jones interactions) are identical
to the interactions they are using, including the cutoff distance.)
2) scrutinize the angle_coeffs (and dihedral_coeffs, if you use
those). Make sure your bond-angles are correct because this will have
a large effect on the effective surface area of each polymer, which
will effect the energy needed to pull a polymer away from the
aggregate.
Hi Andrew,
Thanks for your help. I also test the density of ethane in NPT with fix shake method or large bond coefficient method. The density from the large bond coefficient is always lower than the fix shake method, even I reduce my timestep to 0.1fs. Although the large bond coefficient can constrain the bond length, it will bring some uncertaintys.
Best wishes.
Ren
Hi Andrew,
Thanks for your help. I also test the density of ethane in NPT with fix shake method or large bond coefficient method.
Comment: I think you already realize that the “fix shake” in LAMMPS can only be used to immobilize molecules with 3 atoms or less (such as H2O or CO2). I think this is the reason why you are not using it with the polymer.
How are you using “fix shake” with ethane? Are you constraining only one of the bonds in ethane? (For example, the C-C bond?)
The density from the large bond coefficient is always lower than the fix shake method, even I reduce my timestep to 0.1fs.
It’s good that you were able to reproduce your problem using a smaller simpler molecule (ethane). That is the correct path! Keep simplifying and see if you can find the source of the problem. (For example try using fix shake on H2 or H2O.)
If you cannot find the source of the error, try posting a small test example containing nothing but H2 (or H2O or ethane) to the LAMMPS mailing list, including a short input script and data file. (If you do this, change the name of the subject so that it gets new attention. I worry that people might ignore this thread because they might assume I have answered your question.)
I can’t promise anything, but somebody might test your files if they are really simple.
Cheers
Andrew
P.S. I would trust the simulation with the large bond coefficient more than simulation using “fix shake”.