the k-sampling understanding of NiSi2–Si interface documment in the tutorial

[njuxyh]

Hi all:

i am reading the "NiSi2–Si interface" in  http://docs.quantumwise.com/tutorials/nisi2-si/nisi2-si.html， accrording my undstanding the interface Nisi2-si is 2D slabs, so the k sampling in the x-drection(vacu) should be setted to 1(see figure), why in the documnet the k point in the x is setting to 25, which is the same with that in the y periodic direction.(see figure)


is there anyone can help me undstand this issue

thanks very much!

[Jess Wellendorff]

The NiSi2-Si interface is an interface between two different bulk materials, and is therefore not composed of 2D slabs. It is composed of 2 bulk configurations that meet at an interface. The bulks configurations are on purpose kept fairly small in the x and y directions, in order to maximize computational speed, but a proper k-point sampling along x and y, e.g. 5x5, takes care or the bulk periodicity along those directions. For the subsequent DOS calculation, the k-point grid is on purpose more dense than what was used for the SCF calculation, 25x25.

[Anders Blom]

Indeed, as Jess writes, the entire point of this study and the models implemented in ATK is that we don't treat systems like this in the slab approximation, but are able to use a much more accurate description where the two materials are semi-infinite.

Note that the interface spans the XY plane.