How to obtain the value for "Dielectric_Parameters": Width section?

[yaz...@gmail.com]

Hi

I am looking into the MoS2 example in the Coffee package and am confused by the value given in the Dielectric_Parameters: Width option of the following file: ..\CoFFEE_1.1\Examples\2D\MoS2\Model_Scaling\alpha.6\in

What I understand is that this Width corresponds to the thickness of the dielectric profile, described by equation 10 of the "CoFFEE" paper. I am confused because the "in" file's value of "Width = 11.40454 bohr" is much larger than the thickness of a monolayer MoS2 described by ..\CoFFEE_1.1\Examples\2D\MoS2\QE_input\charged\in, which I found to be about 6 bohr using the coordinates of the Se atoms on the top and bottom sides of MoS2.

May I know what additional considerations do I need to take into account to obtain this Width of 11.40454 bohr as used in the example?

Thank you very much for your help!

Best wishes

Yaze

[Mit Naik]

Hi Yaze,

The screening is determined by the extent of electronic charge density in the out-of-plane direction. This would be larger than the distance between the top and bottom S layers in MoS2. Hence, the right way to estimate the width of the dielectric profile would be to plot the planar averaged charge density of the material along the out-of-plane direction and fit the profile to that. 

For the example in the CoFFEE paper, we have arbitrarily chosen the width to be that of the thickness of the monolayer + the interlayer spacing (S of top to S of bottom) following Ref. https://journals.aps.org/prb/pdf/10.1103/PhysRevB.96.041411.

Best,

Mit

[yaz...@gmail.com]

Thank you very much Mit.

I've proceeded to get the value of sigma for the MoS2 example but the Sigma fit from g_fit.py using the .cube file I generated was 3.894565 instead of the presented 1.89035.

May I double check if the defect wavefunction charge density should be generated from the Gamma point of the highest valence band of the charged defect supercell calculation? Or would it be something else, or there is some details I missed?

Attached below is the fit.png generated by g_fit.py using the partial charge density from gamma point of the highest valence band (band 465) of the charged supercell system, for your reference.

best wishes

Yaze

[Mit Naik]

Hi Yaze,

I'm not sure why this is not working out. The defect band must be flat so I don't think it will make a big difference if you plot it at a different k-point in the BZ. I believe the cube file you generated would be too large to share. I would suggest you go ahead and fit the Gaussian by hand to roughly mimic the defect wavefunction, since g_fit is not giving you the right result. This should be fairly simple since the Gaussian has just two parameters. The potential correction term will be incorporating the difference arising from the model Gaussian not exactly matching the defect wavefunction.

Best,

Mit