Wrap-around errors: Difference between revisions

From VASP Wiki
No edit summary
No edit summary
 
(2 intermediate revisions by 2 users not shown)
Line 2: Line 2:
Wrap around errors arise if the FFT meshes are not
Wrap around errors arise if the FFT meshes are not
sufficiently large. It can be shown that no errors exist
sufficiently large. It can be shown that no errors exist
if the FFT meshes contain all $G$ vectors up to <math>2 G_{\rm cut}</math>.
if the FFT meshes contain all <math>\bold{G}</math> vectors up to <math>2 G_{\rm cut}</math>.


[[File:Wrap errors spheres.png|350px|thumb|Fig. 1: Sphere intersections for <math>G_{\mathrm{cut}}</math>]]
[[File:Wrap errors spheres.png|350px|thumb|Fig. 1: Sphere intersections for <math>G_{\mathrm{cut}}</math>]]
Line 86: Line 86:


----
----
[[Category:Electronic Minimization]][[Category:Electronic Minimization Methods]][[Category:Charge Density]][[Category:Calculational issues]][[Category:Theory]]
[[Category:Electronic minimization]][[Category:Charge density]][[Category:Theory]]

Latest revision as of 16:07, 6 April 2022

In this section we will discuss wrap around errors. Wrap around errors arise if the FFT meshes are not sufficiently large. It can be shown that no errors exist if the FFT meshes contain all vectors up to .

Fig. 1: Sphere intersections for

It can be shown that the charge density contains components up to , where is the "longest" plane wave in the basis set:

The wavefunction is defined as

and in real space it is given by

Using Fast Fourier transformations one can define

Therefore the wavefunction can be written in real space as

The charge density is simply given by

and in the reciprocal mesh it can be written as

Using the above equations for and it is very easy to show that contains Fourier-components up to .

Generally it can be shown that a the convolution of two functions with Fourier-components up to and with Fourier-components up to contains Fourier-components up to .

The property of the convolution comes once again into play, when the action of the Hamiltonian onto a wavefunction is calculated. The action of the local-potential is given by

Only the components with are taken into account (see section ALGO: is added to the wavefunction during the iterative refinement of the wavefunctions , and contains only components up to ). From the previous theorem we see that contains components up to ( contains components up to ).

If the FFT-mesh contains all components up to the resulting wrap-around error is once again 0. This can be easily seen in Fig. 1. Here we see that the small sphere contains all plane waves included in the basis set . The charge density contains components up to (second sphere), and the acceleration components up to , which are reflected in (third sphere) because of the finite size of the FFT-mesh. Nevertheless the components with are correct i.e. the small sphere does not intersect with the third large sphere}