Dear All users,
I want to calculate the solvation energy of cations in water.For this purpose I should have a non-
neutralsystem. In some cases I have small net charge(+1e) and another case(+40e). I get a
warning by using ewald summation. I’ve checked the mailing list and I came across Paul said there
is a correction for energy
http://lammps.sandia.gov/threads/msg19784.html
When should we make this correction? and why?
Dear All users,
I want to calculate the solvation energy of cations in water.For this
purpose I should have a non-neutralsystem. In some cases I have small net charge(+1e) and another
case(+40e). I get a
a +40 charged periodic system is highly unphysical and whatever
numbers you compute from that will be pretty bogus. without any
screening from counter charges, those like charges will repel each
other and thus change the structure of your system and thus will
result in numbers that i would not attribute much meaning to.
Thank you very much for your reply.
Sorry for late response.
To Stan: I think it does make sense use of artifact. I have another good reference too: http://www.mpibpc.mpg.de/14063977/Hub_2014_JCTC.pdf
I can not figure out how it can be right for my system too. In that article we read(as one of their methods which they’ve proposed)
“mimicking the uniform background charge by adding
small partial charges on all hexadecane and water atoms
in the system, compensating the total net charge and thus
creating a neutral system. For this purpose, the partial
charges of all hexadecane and water atoms were changed
by δqhex = −Q/(Vboxρhex) and δqw = −Q/(Vboxρw),
respectively, where ρhex and ρw denote the atom number
density in the hexadecane and water phase, respectively.”
Do you mean we can use the method of uniform background charge as they used in the article?
I can not figure how I can do distributing uniformly charge. If there is another option please let me know.
In addition, in fact I am reproducing the results of an article which the authors used this procedure:
“Charge balance is
achieved by a neutralizing background of approximately
−0.0001 e/Å3 which is relatively small and is expected to
have minimal impact on the cation distributions.”
Again, I do not know how many methods there are in neutralizing my box with charge background.
Thank you very much for your reply.
Sorry for late response.
To Stan: I think it does make sense use of artifact. I have another good reference too: http://www.mpibpc.mpg.de/14063977/Hub_2014_JCTC.pdf
I can not figure out how it can be right for my system too. In that article we read(as one of their methods which they’ve proposed)
“mimicking the uniform background charge by adding
small partial charges on all hexadecane and water atoms
in the system, compensating the total net charge and thus
creating a neutral system. For this purpose, the partial
charges of all hexadecane and water atoms were changed
by δqhex = −Q/(Vboxρhex) and δqw = −Q/(Vboxρw),
respectively, where ρhex and ρw denote the atom number
density in the hexadecane and water phase, respectively.”Do you mean we can use the method of uniform background charge as they used in the article?
They don’t. You quoted what they did instead.
The uniform background that Stan is referring to happens automatically. You just ignore the g=0 term in reciprocal space (which is 0 for neutral systems)
I can not figure how I can do distributing uniformly charge. If there is another option please let me know.
What?? You have quoted the procedure (which is questionable in itself). What else do you need?
In addition, in fact I am reproducing the results of an article which the authors used this procedure:
“Charge balance is
achieved by a neutralizing background of approximately
−0.0001 e/Å3 which is relatively small and is expected to
have minimal impact on the cation distributions.”Again, I do not know how many methods there are in neutralizing my box with charge background.
To Axel: Yes I know there is large amount of positive charge. That’s why I wanted to know it does make sense or not in addition to a successful run.
There is a difference between what makes sense and what people have done. More importantly, if you mess with electrostatics like this you better know what you are doing, and that means reading a lot more textbooks and papers than what you obviously have done. I worked in part on such matters when i was a grad student and it was tough. There are many subtleties that need to be considered.
Axel.
In other words, if you just ignore the warning that LAMMPS gives you and use the Ewald sum for a non-neutrally charged system, you are implicitly assuming that there is a uniform background charge, and a correction will automatically be applied to the energy (which is equal to equation 9 of the paper you referenced: http://www.mpibpc.mpg.de/14063977/Hub_2014_JCTC.pdf ).
Dear Stan and Axel,
I am sorry again to late response I had no access to a computer for checking email.
Thank you very much for your reply and time.
To Stan: I really appreciate for checking that paper and defining what lammps does to uniform background charge.