Dear Lammps Users,
I have a water domain confined between identically (both sign and magnitude) charged surfaces which has fixed BC in height. I am using long-range interaction to model between surface-liquid and liquid-liquid interactions.
My question is about the orientation of water molecules near interface regions. Regarding the coulomb rule, we expect negative Oxygen atoms look towards to positively charged surface and positive Hydrogen atoms point away from the positively charged surface. But I see in the results the reverse behavior such as Hydrogen looks towards to positive surface, Oxygen looks away from the positive surface.
I should point out that I also tried the electrical field case by assigning opposite charges to the counter walls. I got the result with expected water molecules orientation which woter molecules aligns through EF directions. Moreover, the behavior of water orientation on a charged single surface (not in a channel which means PPP boundary conditions) also presents expected behavior. I also tried some other cases. Based on these trials, I suspect from long range interaction.
Any suggestions or experience? I thank you in advance for all of your help.
Regards,
Alper C.
Dear Lammps Users,
I have a water domain confined between identically (both sign and magnitude)
charged surfaces which has fixed BC in height. I am using long-range
interaction to model between surface-liquid and liquid-liquid interactions.
My question is about the orientation of water molecules near interface
regions. Regarding the coulomb rule, we expect negative Oxygen atoms look
towards to positively charged surface and positive Hydrogen atoms point away
from the positively charged surface. But I see in the results the reverse
behavior such as Hydrogen looks towards to positive surface, Oxygen looks
away from the positive surface.
I should point out that I also tried the electrical field case by assigning
opposite charges to the counter walls. I got the result with expected water
molecules orientation which woter molecules aligns through EF directions.
Moreover, the behavior of water orientation on a charged single surface
(not in a channel which means PPP boundary conditions) also presents
expected behavior. I also tried some other cases. Based on these trials, I
suspect from long range interaction.
are all of these calculations done with neutralized simulation cells?
axel.
Dear Dr. Axel,
Thank you for the response.
I am using PPPM algorithm as the long range solver. So, I believe that the ewald algorithms ignore the diverging energy term and act as an appointed countercharge around the domain to make it neutralized. But I haven’t done any extra step except that .
Thanks
Best Regards,
Alper C
Dear Dr. Axel,
Thank you for the response.
I am using PPPM algorithm as the long range solver. So, I believe that the
ewald algorithms ignore the diverging energy term and act as an appointed
countercharge around the domain to make it neutralized. But I haven't done
any extra step except that .
then i'd say, your simulation with the two same sign charged walls is
likely bogus.
while PPPM will indeed ignore the divergent term, there is still a
significant residual effect due enforced background neutralization.
this effect is the larger, the more you stray away from having a
neutral system. thus the general rule of the thumb is that the net
charge of a system should have a magnitude of less than one electron
charge.
for comparison, i'd suggest doing a simulation with lj/cut/coul/cut
and non-periodic boundaries in the direction of the walls. the
orientation of molecules near the walls is dominated by short range
interactions, so i doubt that using PPPM or Ewald will mater much.
those matter most in the center of the your liquid slab, where they'll
lead to a non-flat potential.
there should be some relevant papers on that by my PhD adviser Eckhard
Spohr on this written between 1995 and 2000.
axel.
Dear Dr. Axel,
Thank you for all the information. I ll read the relevant papers by your advisor.
Actually, I already tried lj/cut/coul/cut. As you foreseen, it gives me the correct orientation of water molecules as been effective within the cut off distance but becoming ambiguous around the center of channel.
Is there anyway to remove this enforced background neutralization by keeping ewald summation active at the same time?
Thanks
Best regards
Alper
Dear Dr. Axel,
Thank you for all the information. I ll read the relevant papers by your
advisor.
Actually, I already tried lj/cut/coul/cut. As you foreseen, it gives me the
correct orientation of water molecules as been effective within the cut off
distance but becoming ambiguous around the center of channel.
yup. if you want this to work perfectly correct, you'll need to do
2d-ewald summation, i.e. do a lattice sum in x and y and just a
(long?) cutoff, potentially with a switching function or using
something like wolf summation for z. that, however, is quite a tall
order, since this would all need to be programmed.
Is there anyway to remove this enforced background neutralization by keeping
ewald summation active at the same time?
no. the neutralization is not explicit, you get it automatically from
skipping over the divergent term at g=0. if you try to compensate this
in some way, you would end up with having to subtract infinity from
infinity.
axel.
Dear all,
Naively (i.e. from a macrosopic point of view), I would expect the
electric field between two identically charged plates to be zero, so
that I would not expect any preferential orientation of water
molecules (at least resulting from coulombic interactions) at any
location between the plates, if the long range part of the
interactions is correctly taken into account. Did I miss something?
Best regards,
Laurent