1
$\begingroup$

I was under the impression that differences between the various long range electrostatic algorithms, ewald, PME, PPPM, to name a few, lied solely in their treatment of the reciprocal space term. Specifically, I thought all methods looped over a particles neighbor and summed q(i)*q(j)erfc(alphar)/r for the real space calculation. However, this is my impression, which I have yet to validate with others. I became confused when I recently read "Ewald summation techniques in perspective: a survey" whose equation 22 seems to have suggested a different way to calculate the short range potential. Could anyone else chime in here, and either tell me where I am wrong, or I am correct?

Thanks!

$\endgroup$
  • $\begingroup$ You can beautify your posts with $\LaTeX$ syntax using the Math Jax implementation, please have a look here and here. Please do not use markup in the title field, see here for details. $\endgroup$ – Martin - マーチン Jul 3 '15 at 3:48
1
$\begingroup$

There are a number of ways of looking at Ewald's summation, but one is that instead of trying to solve the original problem where the potential is due to a set of point charges you, instead, solve the system for a set of smeared charges and add a correction. Now there are limitations on what are good ways to do the smearing, but there are still quite a lot of possible choices. The most common is to replace the original delta function charge with a Gaussian, the width of which is determined by alpha in your question. This leads to the error function and exponentials you are used to. However it's not the only possible choice. In equation 22 in the paper you refer to a different way of smearing the charges has been chosen, a linear ramp centred on the particle. This is described in equation 21. As the smearing is different the individual terms in the Ewald summation have different forms when compared to Gaussian smearing, but it's no less valid a way of proceeding.

If you want to learn more about this look into Fourier Transforms and the Convolution theorem. A bit of reading on that and the Ewald Summation in all its forms becomes pretty straightforward.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.