[gmx-users] walls and E-z
Alex
nedomacho at gmail.com
Sat Nov 11 23:53:31 CET 2017
I'll just reply here, since I seem to be unable to respond to Berk's
message on redmine.
> On Fri, Nov 10, 2017 at 1:12 AM, <admin at redmine.gromacs.org
> <mailto:admin at redmine.gromacs.org>> wrote:
> Issue #2291 <http://redmine.gromacs.org/issues/2291#change-15325> has
> been updated by Berk Hess.
> ------------------------------------------------------------------------
>
> * *Status* changed from /New/ to /Rejected/
>
> This is likely not a bug. The wall and electric field code are
> completely independent.
> I guess that what you are seeing is the effect of the ewald-geometry.
> Using geometry=3DC heavily penalizes the build-up of a dipole, so
> alignment of molecules with the electric field decreases by a factor
> about equal to dielectric constant. You can verify this by changing to
> 3D Ewald geometry.
> What Ewald boundary condition you actually want depends on what you
> want to do. This is a critical aspect and you need to think this
> through well.
>
I tried with the default setting for ewald-geometry, which is 3d -- *no
difference* and no ion flux to speak of. My basic solution right now is
to introduce an actual atomistic wall while keeping full PBC and not
using Gromacs implementation of walls. My setup is incredibly simple, as
I am trying to polarize an ionic solution along the Z-axis.
Alex
On 11/9/2017 3:16 PM, Alex wrote:
> Got it -- filed on redmine. With a "real" wall and E-z = 0.1V/nm,
> only five ions cross, which can be confirmed by simple calculations as
> sufficient to counteract the external field, given the membrane's own
> crossing barrier.
>
> Alex
>
> On Thu, Nov 9, 2017 at 2:21 PM, Dan Gil <dan.gil9973 at gmail.com
> <mailto:dan.gil9973 at gmail.com>> wrote:
>
> I simulated ionic liquids with walls and electric field. The field was
> processed in my case.
>
> On Thu, Nov 9, 2017 at 4:13 PM, Alex <nedomacho at gmail.com
> <mailto:nedomacho at gmail.com>> wrote:
>
> > David, not sure this is a bug. For the moment, can someone
> simply tell me
> > if external field directive is processed when walls are used?
> > I fixed the issue by introducing an actual wall as part of the
> system and I
> > think this behavior may be by design. I mean, why apply driving
> fields when
> > the system is impermeable?
> >
> > Alex
> >
> > On Thu, Nov 9, 2017 at 12:03 AM, David van der Spoel
> <spoel at xray.bmc.uu.se <mailto:spoel at xray.bmc.uu.se>
> > >
> > wrote:
> >
> > > On 08/11/17 22:29, Alex wrote:
> > >
> > >> Okay, same thing with 0.5V/nm. I think it's fairly safe to
> say that
> > >> there's
> > >> something wrong here...
> > >>
> > > Haven't followed but if a bug is suspected please file a report at
> > > redmine.gromacs.org <http://redmine.gromacs.org>.
> > >
> > >
> > >> Alex
> > >>
> > >> On Wed, Nov 8, 2017 at 12:25 PM, Alex <nedomacho at gmail.com
> <mailto:nedomacho at gmail.com>> wrote:
> > >>
> > >> Good question. Dielectric breakdown of water is generally poorly
> > >>> understood and the threshold depends on the ionic strength, but
> > >>> 0.4-0.5V/nm
> > >>> is generally where the fun begins. MD modelers working with
> solvated
> > >>> systems casually ignore this, unless they have the great
> misfortune of
> > >>> getting me as a reviewer. :)
> > >>> That aside, I believe your suggestion is sound, at least to
> see if
> > what I
> > >>> observe is an outright bug.
> > >>>
> > >>> Thanks,
> > >>>
> > >>> Alex
> > >>>
> > >>> On Wed, Nov 8, 2017 at 10:39 AM, Dan Gil
> <dan.gil9973 at gmail.com <mailto:dan.gil9973 at gmail.com>>
> > wrote:
> > >>>
> > >>> Yes I saw your plot and it is simply around 0 with walls.
> > >>>>
> > >>>> What is the field required for dielectric breakdown?
> > >>>>
> > >>>> On Wed, Nov 8, 2017 at 12:18 PM, Alex <nedomacho at gmail.com
> <mailto:nedomacho at gmail.com>> wrote:
> > >>>>
> > >>>> Hi Dan,
> > >>>>>
> > >>>>> Yup, periodic, continuous, and electrically neutral. I
> suggested a
> > >>>>>
> > >>>> similar
> > >>>>
> > >>>>> thought in my question, i.e. with walls any transport would
> > definitely
> > >>>>>
> > >>>> be
> > >>>>
> > >>>>> transient and self-limited. However, nothing is
> transported even in
> > the
> > >>>>> perturbative sense, as you can see from the flux. The
> behavior is
> > that
> > >>>>>
> > >>>> of a
> > >>>>
> > >>>>> system without any driving field.
> > >>>>>
> > >>>>> The electric field is already quite high (0.1 V/nm) and of
> course I
> > >>>>>
> > >>>> could
> > >>>>
> > >>>>> go completely nuts and exceed the experimental dielectric
> breakdown
> > >>>>> threshold values for water, but the question remains, no?
> > >>>>>
> > >>>>> Thanks,
> > >>>>>
> > >>>>> Alex
> > >>>>>
> > >>>>>
> > >>>>>
> > >>>>> On 11/8/2017 9:58 AM, Dan Gil wrote:
> > >>>>>
> > >>>>> Hi Alex,
> > >>>>>>
> > >>>>>> Is your system without walls periodic and continuous in all
> > >>>>>>
> > >>>>> directions? I
> > >>>>
> > >>>>> can see a scenario where this sort of system will maintain
> charge
> > >>>>>> neutrality in the different reservoirs separated by the
> semi-porous
> > >>>>>> membrane. While cations will be transported, the charge
> in each
> > >>>>>>
> > >>>>> reservoir
> > >>>>
> > >>>>> will be maintained constant because as one cation leaves, its
> > periodic
> > >>>>>> image enters the same reservoir. It is a steady-state
> process.
> > >>>>>>
> > >>>>>> In the system with walls, charge neutrality will be broken if
> > cations
> > >>>>>>
> > >>>>> are
> > >>>>
> > >>>>> transported across the membrane because it won't have a
> periodic
> > image
> > >>>>>> that
> > >>>>>> enters the same reservoir as it leaves. I think that the
> cation
> > >>>>>>
> > >>>>> transport
> > >>>>
> > >>>>> would be more like capacitance since a constant electric
> field will
> > >>>>>>
> > >>>>> only
> > >>>>
> > >>>>> be
> > >>>>>> able to hold a finite number of cations across the
> membrane. This is
> > >>>>>> an
> > >>>>>> equilibrium process.
> > >>>>>>
> > >>>>>> Maybe try higher electric field?
> > >>>>>>
> > >>>>>> Dan
> > >>>>>>
> > >>>>>> On Fri, Nov 3, 2017 at 2:43 AM, Alex <nedomacho at gmail.com
> <mailto:nedomacho at gmail.com>> wrote:
> > >>>>>>
> > >>>>>> Hi all,
> > >>>>>>
> > >>>>>>>
> > >>>>>>> It appears that the external field is refusing to move
> the ions
> > when
> > >>>>>>> walls
> > >>>>>>> are present. I am comparing two setups of a system that
> has an
> > >>>>>>> aqueous
> > >>>>>>> bath
> > >>>>>>> (1M KCl) split by a semi-porous (infinitely selective
> for cations)
> > >>>>>>> membrane
> > >>>>>>> in XY. The only difference between them is that one is
> periodic in
> > >>>>>>> XYZ
> > >>>>>>> and
> > >>>>>>> the other has two walls. The difference isn't minor --
> consider K+
> > >>>>>>>
> > >>>>>> fluxes
> > >>>>
> > >>>>> with and without walls: https://www.dropbox.com/s/jve0
> > >>>>>>> hqqpfkn4ui6/flux.jpg?dl=0
> > >>>>>>>
> > >>>>>>> Initially, ionic populations in each case are homogeneous. I
> > realize
> > >>>>>>>
> > >>>>>> that
> > >>>>
> > >>>>> with walls the process will stop when all cations end up
> at the top
> > of
> > >>>>>>> the
> > >>>>>>> box (and that's the goal). However, there is no flux
> right from the
> > >>>>>>> start.
> > >>>>>>> Relevant portion of the mdp with walls below (not sure
> if this is
> > >>>>>>> important, but 'ewald-geometry' directive isn't in the
> mdp without
> > >>>>>>> walls):
> > >>>>>>>
> > >>>>>>> pbc = xy
> > >>>>>>> nwall = 2
> > >>>>>>> wall-type = 12-6
> > >>>>>>> wall-r-linpot = 0.25
> > >>>>>>> wall_atomtype = opls_996 opls_996
> > >>>>>>> wall-ewald-zfac = 3
> > >>>>>>> periodic_molecules = yes
> > >>>>>>> ns_type = grid
> > >>>>>>> rlist = 1.0
> > >>>>>>> coulombtype = pme
> > >>>>>>> ewald-geometry = 3dc
> > >>>>>>> fourierspacing = 0.135
> > >>>>>>> rcoulomb = 1.0
> > >>>>>>> rvdw = 1.0
> > >>>>>>> vdwtype = cut-off
> > >>>>>>> cutoff-scheme = Verlet
> > >>>>>>>
> > >>>>>>> Any ideas?
> > >>>>>>>
> > >>>>>>> Thanks,
> > >>>>>>>
> > >>>>>>> Alex
> > >>>>>>>
> > >>>>>>> --
> > >>>>>>> Gromacs Users mailing list
> > >>>>>>>
> > >>>>>>> * Please search the archive at
> http://www.gromacs.org/Support
> > >>>>>>> /Mailing_Lists/GMX-Users_List before posting!
> > >>>>>>>
> > >>>>>>> * Can't post? Read
> http://www.gromacs.org/Support/Mailing_Lists
> <http://www.gromacs.org/Support/Mailing_Lists>
> > >>>>>>>
> > >>>>>>> * For (un)subscribe requests visit
> > >>>>>>>
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users
> <https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users>
> > >>>>>>> or
> > >>>>>>> send a mail to gmx-users-request at gromacs.org
> <mailto:gmx-users-request at gromacs.org>.
> > >>>>>>>
> > >>>>>>>
> > >>>>>>> --
> > >>>>> Gromacs Users mailing list
> > >>>>>
> > >>>>> * Please search the archive at http://www.gromacs.org/Support
> > >>>>> /Mailing_Lists/GMX-Users_List before posting!
> > >>>>>
> > >>>>> * Can't post? Read
> http://www.gromacs.org/Support/Mailing_Lists
> <http://www.gromacs.org/Support/Mailing_Lists>
> > >>>>>
> > >>>>> * For (un)subscribe requests visit
> > >>>>>
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users
> <https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users>
> > or
> > >>>>> send a mail to gmx-users-request at gromacs.org
> <mailto:gmx-users-request at gromacs.org>.
> > >>>>>
> > >>>>> --
> > >>>> Gromacs Users mailing list
> > >>>>
> > >>>> * Please search the archive at http://www.gromacs.org/Support
> > >>>> /Mailing_Lists/GMX-Users_List before posting!
> > >>>>
> > >>>> * Can't post? Read
> http://www.gromacs.org/Support/Mailing_Lists
> <http://www.gromacs.org/Support/Mailing_Lists>
> > >>>>
> > >>>> * For (un)subscribe requests visit
> > >>>>
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users
> <https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users>
> or
> > >>>> send a mail to gmx-users-request at gromacs.org
> <mailto:gmx-users-request at gromacs.org>.
> > >>>>
> > >>>>
> > >>>
> > >>>
> > >
> > > --
> > > David van der Spoel, Ph.D., Professor of Biology
> > > Head of Department, Cell & Molecular Biology, Uppsala University.
> > > Box 596, SE-75124 Uppsala, Sweden. Phone: +46184714205
> <tel:%2B46184714205>.
> > > http://www.icm.uu.se
> > >
> > > --
> > > Gromacs Users mailing list
> > >
> > > * Please search the archive at http://www.gromacs.org/Support
> > > /Mailing_Lists/GMX-Users_List before posting!
> > >
> > > * Can't post? Read
> http://www.gromacs.org/Support/Mailing_Lists
> <http://www.gromacs.org/Support/Mailing_Lists>
> > >
> > > * For (un)subscribe requests visit
> > >
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users
> <https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users>
> or
> > > send a mail to gmx-users-request at gromacs.org
> <mailto:gmx-users-request at gromacs.org>.
> > >
> > --
> > Gromacs Users mailing list
> >
> > * Please search the archive at http://www.gromacs.org/
> > Support/Mailing_Lists/GMX-Users_List before posting!
> >
> > * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
> <http://www.gromacs.org/Support/Mailing_Lists>
> >
> > * For (un)subscribe requests visit
> >
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users
> <https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users>
> or
> > send a mail to gmx-users-request at gromacs.org
> <mailto:gmx-users-request at gromacs.org>.
> >
> --
> Gromacs Users mailing list
>
> * Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List
> <http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List>
> before posting!
>
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
> <http://www.gromacs.org/Support/Mailing_Lists>
>
> * For (un)subscribe requests visit
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users
> <https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users>
> or send a mail to gmx-users-request at gromacs.org
> <mailto:gmx-users-request at gromacs.org>.
>
>
More information about the gromacs.org_gmx-users
mailing list