Heterostructure and interlayer distance

[abhishek77492652]

Hello sir,
I want to make the heterostructure of MoS2 and MoSe2 and also i want to calculate the optimum interlayer distance .
I have also  followed the tutorial for bilayer graphene  but i am not getting the desired information (correct interlayer distance )

[Petr Khomyakov]

Given the amount of information provided in your post, it is hard to say anything specific on this issue. You would need to describe what you did in more detail.

[krahil]

Hello Sir,
We are trying to make heterostructures and we started with the MoS2 and MoSe2. The experimental value of the interlayer distance between the two is 6.65A. I have 2 questions-

1. I have first increased the 'c' vector for mono layer MoS2 to 23A  and then added MoSe2 to it. By default it aligned itself to almost ~6.65A but when I shifted MoSe2 a bit and then optimized using PBE-D2, then the resultant interlayer distance doesn't come out to be 6.65A rather it remains at the same position it was kept. So how to proceed further in a case where new structure is being made with van der waal forces in between them.
2. Whether interface should be done to avoid the above situation? If yes, then whether lattice vector 'c' for mos2  and should be increased  and then interfaced or directly without increasing?

[Petr Khomyakov]

Could you enclose your python script related to this calculation? Another thing is that interlayer distance of 6.65 Angs seems to be about twice larger than I would expected for this kind of 2D material structures. So, could you post the reference to the experimental work you mentioned? 

[krahil]

Sir,
I have attached the py file.
DOI of reference : 10.1103/PhysRevB.88.085318 .
In the 4th section, in the last line of 1st para, it has mentioned the distance. Also the paper mention distance for mos2/ws2 .

[Petr Khomyakov]

The first thing I would do is to significantly decrease the max_forces parameter, as the van der Waals forces are very small.  Also, you have done everything with default ATK computational settings, but defaults need to be checked to ensure convergence of physical quantities of your interest with respect to these settings, e.g., density mesh cut-off and so on.

Another suggestion would be to do a "manual" scan of the total energy vs. separation distance (e.g., from 4.5 to 7.5 Angstrom) between the monolayers to roughly locate the energy minimum, and then do optimization at that approximately-calculated equilibrium separation distance, using strict convergence criteria.

[krahil]

Thank you for the suggestion sir.

I tried manual scanning as well. But the paper results are different.
Taking example of a paper DOI: 10.1088/1361-6528/aa68d8,
there in the graph shown for binding ergy. V/s interlayer distance, the minimum energy is at at 2.8A. But when I had tried to check, it was ~8.89eV and when divided by total atoms it was around -88meV but they are reporting  -144meV.
Also i tried with one more heterostructure and it was revealing -160meV B.E. and paper was mentioning ~40meV.
Is it because of another tool they are using and could I go with my results.
Please suggest me.

[Petr Khomyakov]

I do not get what kind of calculations you have done, and what you are comparing to.  You originally quoted the 6.65 Angs separation distance as a reference; now it turned out to be 2.5 Angs.

[krahil]

Sir,
With reference to your earlier reply, I had tried for different heterostructures so that I can be sure of the results. Also, in my last reply  I had mentioned the paper for the '2.5A' thing. This 2.5A distance is for MoS2/stanene heterostructure.
I was trying the method for different structures so that before asking you I may be sure of the method I am performing.

Thank you for your time and suggestion.
Any  advice will be really helpful.

[Petr Khomyakov]

Going through the following tutorial might be helpful for setting up accurate geometry optimization when using DFT-D within the framework of LCAO basis sets, see https://docs.quantumwise.com/tutorials/dispersion_corrections_and_bsse/dispersion_corrections_and_bsse.html. You really have to produce a plot of E vs separation distance using strict-enough computational settings (proper mesh cut-off, k-point sampling, basis set) for total energy calculations. 

[krahil]

Thank you again sir.