effective potential

[xiaolight]

in http://quantumwise.com/forum/index.php?topic=5114.msg22147#msg22147
It tells me I can use effective potential to calculate the electron affinity but the data is really large than I thought. I use the MoS2 supercell and make the vacuum level really large.

[Anders Blom]

I'm not sure you can calculate the workfunction/affinity of a 2D sheet this way. The method essentially assumes that you can define a "bulk-like" region and compare the potential there to vacuum. In this case, it would mean cutting a surface in the 2D plane. So, in your case, rather, you would repeat the structure many times in B or C (not both), then add vacuum in that direction, and then study the potential in this direction, using the averaged value across the perpendicular plane (and ideally also average in the projection direction to get rid of the atomic variation).

Alternatively, and maybe better, you might try to use Dirichlet/Neumann boundary conditions and use the fact that then the Fermi level represents the work function (for metal); for semiconductors you need to also factor in the band gap.

[Kim_W]

By comparing the values reported in the tutorials for the (100), (110) and (111) crystalline faces of Ag, it appears clear that the work function depends on the crystalline face considered. Thus, the method you posted may be not reasonable. Can we use two layers MoS2 for this calculation? And turning the first layer MoS2 into gost atoms.

[Petr Khomyakov]

By comparing the values reported in the tutorials for the (100), (110) and (111) crystalline faces of Ag, it appears clear that the work function depends on the crystalline face considered.

This is a well-known experimentally-confirmed fact, and not for Ag only.

Thus, the method you posted may be not reasonable.

Which method? And why not reasonable?

Can we use two layers MoS2 for this calculation? And turning the first layer MoS2 into gost atoms.

You can certainly do it. Whether this will give an improvement for accuracy of work function or affinity calculation, it is another question.

[Kim_W]

Thus, the method you posted may be not reasonable.

Which method? And why not reasonable?

So, in your case, rather, you would repeat the structure many times in B or C (not both), then add vacuum in that direction, and then study the potential in this direction, using the averaged value across the perpendicular plane (and ideally also average in the projection direction to get rid of the atomic variation).

This one.

[Petr Khomyakov]

Why is it not a "reasonable" method for work function calculation?

[Jess Wellendorff]

Dear Wang, the work function does in general depend on surface facet.