Re: langevinHydrogen off with rigidBonds all?

From: Peter Freddolino (petefred_at_ks.uiuc.edu)
Date: Sun Mar 16 2008 - 08:55:42 CDT

Hi Soo Mei,
yes, if you're using rigidBonds you should also turn langevinHydrogen
off. The summary you give below is correct; this means that friction and
random forces will not be applied to the hydrogens, but that the
hydrogens will still move according to forces from the other atoms,
subject to their bond length constraints. If you're interested more on
how this works, you should look up the original RATTLE and SETTLE
papers, listed below in bibtex form.
Best,
Peter

@ARTICLE{Andersen1983,
  author = {Andersen, Hans C.},
  title = {Rattle: A "velocity" version of the shake algorithm for molecular
    dynamics calculations},
  journal = {Journal of Computational Physics},
  year = {1983},
  volume = {52},
  pages = {24--34},
  number = {1},
  month = oct,
  url =
{http://www.sciencedirect.com/science/article/B6WHY-4DD1MVH-6R/2/e7d2ac9700e994cee6eb0cbc18698e5d}
}

@ARTICLE{MIYA93,
  author = {Shuichi Miyamoto and Peter A. Kollman},
  title = {SETTLE: An Analytical Version of the SHAKE and RATTLE Algorithm
    for Rigid Water Models},
  journal = {Journal of Computational Chemistry},
  year = {1993},
  volume = {13},
  pages = {952-962},
  file = {MIYA93.pdf:MIYA93.pdf:PDF}
}

Low Soo Mei wrote:
> Dear NAMD users,
>
> The question has been asked before, here,
> http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l/3693.html ,
> but I could not find a reply on the mailing list. I wonder if someone
> with better physics than me could help :)
>
> My timestep is 2 fs and I am using Langevin dynamics, NVT for early
> equilibration and NPT for the later part of equilibration and also the
> actual runs. For a 2 fs timestep, I read that rigidBonds "all" is
> recommended, to freeze all hydrogen-related bond lengths and bond
> angles. Under these conditions, does it necessarily follow that
> langevinHydrogen must be "off"?
>
> >From what little I understand of Langevin dynamics, it controls the
> temperature and pressure by introducing frictional damping force and
> random forces on each particle in the system. So, langevinHydrogen
> "off" means that these friction and random forces are not applied to
> the H atoms, but yet the H atoms will still be able to move according
> to the forces on other atoms? Even though the hydrogen-linked bonds
> have been 'frozen' by rigidBonds.
>
> Thanks for your advice in advance, and I am most grateful. References
> for additional (basic) reading are welcome as well.
>
> Cheers,
> Soo Mei

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:49:19 CST