pair_write with pppm

Hi

I have a doubt regarding pair_write command in LAMMPS along with charge interactions. I am using LAMMPS version: 5Apr2014.

Q: When I use pair_style lj/cut/coul/cut after giving explicit charges used in the data file for each type, the output I am getting matches with the analytical expression for r<r_c. However when I use pair_style lj/cut/coul/long with k_space style as pppm, even the short range values explicitly output by the pair_write command is very different from analytical values. Given is a plot obtained from using both pair styles as mentioned above obtained from pair_write.

command used:
pair_write 3 3 5000 r 0.01 12.0 coulong_33.txt LJ_Coul -0.348 -0.348

compare.png

Hi

I have a doubt regarding pair_write command in LAMMPS along with charge
interactions. I am using LAMMPS version: 5Apr2014.

Q: When I use pair_style lj/cut/coul/cut after giving explicit charges
used in the data file for each type, the output I am getting matches with
the analytical expression for r<r_c. However when I use pair_style
lj/cut/coul/long with k_space style as pppm, even the short range values
explicitly output by the pair_write command is very different from
analytical values. Given is a plot obtained from using both pair styles as
mentioned above obtained from pair_write.

​you are comparing apples to oranges. they *must* be different since the
coulomb interaction for the real space of a coul/long style is very
different from the 1/r of a coul/cut style due to the applied damping.
please have a look at your favorite textbook on MD where it explains the
"trick" behind handling long-range coulomb efficiently, e.g. via ewald
summation (pppm differs only in how the reciprocal space part is computed).

axel.

compare.png