Is Tetrahedron method of EPH module suitable for 2D system?

Running Abinit Response calculations
1

Hi there,

The new EPH module has two kinds of Electron Phonon integration method, the tetrahedron and Gaussian technique. It seems that tetrahedron is originally design for bulk phase, is it suitable to be used in 2D system or monolayer?

Regards,
Kongfer

2

Hi Kongfer,

good question. Indeed it’s not great for 2D systems, you should be using triangles but this is not implemented yet. The 3D mode can work, but you’ll get a spiky DOS, it’s better to add another layer of k (n x n x 2 kpts) to avoid exact degeneracies, but if you have a lot of vacuum this will not change things much.

3

Hi mverstra,

Thanks for your quick response!

And what if I choose Gaussian technique? Could it be used in 2D system accompany by a manually set dense dos mesh and a proper broadening parameter?

I also noticed that when we choose tetrahedron method, the automatic generated dos mesh grid is not very dense. Is it enough to make the integration over energy converge?

4

Gaussians are fine. Need to converge the width and density of course.

The DOS grid is up to you, it is not automatically imposed with the integration technique. in general the tetrahedra converge faster, but this is empirical.