# Question of Fix efield

Dear All,

I’m trying to simulate polarized water based on drude oscillator model under external electric field. I found a package of drude oscillator model at https://github.com/jdevemy/lammps-drude.

My current problem is about fix efield. According to the theory of drude oscillator, the position of drude particle is updated instantaneously to make the relative velocity to the drude center be zero, which is achieved by thermostat the the relative motion to 1K. I need to include a fictitious potential energy like E=-qrEfield in to the total potential energy. I check the manual of fix efield. It seems that I need to create a variable and include the energy keyword. Am I correct? Could this fictions potential energy be included in the thermostat?

Han

Dear All,

I'm trying to simulate polarized water based on drude oscillator model under
external electric field. I found a package of drude oscillator model at
https://github.com/jdevemy/lammps-drude.

My current problem is about fix efield. According to the theory of drude
oscillator, the position of drude particle is updated instantaneously to
make the relative velocity to the drude center be zero, which is achieved by
thermostat the the relative motion to 1K. I need to include a fictitious
potential energy like E=-q*r*Efield in to the total potential energy. I
check the manual of fix efield. It seems that I need to create a variable
and include the energy keyword. Am I correct? Could this fictions potential
energy be included in the thermostat?

well, first and foremost this is a question that you should ask the
author(s) of the package.

fix efield is a *very* crude representation of an external field, it
is essentially the same as fix setforce, if it was scaled by the
charge of each particle. it is correct only in vacuum. i would
recommend to consider using a more realistic model, e.g. having
oppositely charged particles that are kept apart through confinement
potentials.

axel.

Dear Axel,

Thanks for your reply. Actually, I have done some literature review of the polarized water molecule and the way to deal with water under electric field. The three typical ways are polarizable point dipole method ( add imposed dipole directly and solve extend lagrangian equation for the magnitude of dipole), distributed charge sites method( as you said, having 4 charged particles constrained to molecule and changing the charge by the polarization energy, still need to solve extend Lagrangian equation) and drude oscillator method. Personally, I prefer the first two methods but I don’t know which part of the code I should modify to achieve variable charges or dipoles. I know that you have contribute to LAMMPS for a long time. Could you please give me some advice? Thank you!

Han

Dear Axel,

Thanks for your reply. Actually, I have done some literature review of the
polarized water molecule and the way to deal with water under electric
field. The three typical ways are polarizable point dipole method ( add
imposed dipole directly and solve extend lagrangian equation for the
magnitude of dipole), distributed charge sites method( as you said, having 4
charged particles constrained to molecule and changing the charge by the
polarization energy, still need to solve extend Lagrangian equation) and
drude oscillator method. Personally, I prefer the first two methods but I
don't know which part of the code I should modify to achieve variable
charges or dipoles. I know that you have contribute to LAMMPS for a long

you are describing methods how to implement a polarizable water
potential. this has next to *nothing* to do with your question of how
an external field is modeled, and also not with what i said. you have

what i said about fix efield is true in general and applies to
non-polarizable and polarizable potentials all the same. however, it
is my assertion that if you do care enough about how polarization of
your system is treated so that you want to use a polarizable water
potential, it would matter much more how well you do model the impact
of your external field. as i already said, what fix efield does is
extremely crude and really only a good model for an individual
molecule in an electric field in vacuum. it also appears to me, that
you have not yet understood the implications of how fix efield works.

i have no advice on which type polarizable model you should use. the