Questions regarding Supercell Scaling

[yaz...@gmail.com]

Hi Mit

If I am working on a doped TMD monolayer system, what is the prefered scaling used? Does it strictly need to be "a x a x a " as suggested in the examples or is "a x a x 1.6a" or some other factor also fine? And if I proceed with a "a x a x 1.6a" scaling factor instead of the "a x a x a" one, is there anything additional that I need to consider when finding the correct E^Per,m?

I am confused regarding this because the paper and its Fig 6 did mention and show that the choice of coefficient to the last lattice vector will affect the value of E^Per,m significantly for small values of a; however the optimal coefficient is not mentioned. 

Best wishes

Yaze

[Mit Naik]

Hi Yaze,

Sorry for the delay. This thread surprisingly did not reach my inbox!

The scaling factor is determined by the amount of vacuum in your DFT simulation cell. 

Given some lattice parameter: alat. If your DFT simulation is performed with a cell such that |a1| = |a2| = 5*alat (5x5 in-plane supercell) and you also used 5*alat = |a3| in the out-of-plane direction; then the scaling goes as \alpha x \alpha x \alpha. If the cell dimension in the out-of-plane direction in the DFT simulation (with a 5x5 in-plane supercell) is, say |a3| = 10*alat; then the scaling goes as \alpha x \alpha x 2\alpha. 

I hope this makes sense.

Best,

Mit

[yaz...@gmail.com]

Hi Mit

Thanks a lot for the reply. I assumed the same settings as you mentioned and it is giving me a sensible result. Great to hear that it is actually the right way to do it.

No worries about the not-too-late reply, all is well.

Best wishes

Yaze