[QE-users] some queries about mono layer optimization and vacuum

[Lucas Nicolás Lodeiro Moraga]

Hello Rekha!
I am not an expert in monolayers, I work with slabs, but the problems
should be similar.
First, the ecutrho value that you use (52) it is very low... I do not know
wich type of pseudopotential do you use (NormConserving, UltraSoft, PAW),
but the ecutrho must be 4 times the ecutwfc (For US and PAW form 6 to 12
times). Other thing, the convergence criteria of 1meV its good, but I
prefer a criteria independent of number of atoms, as 1meV/atom... If you
have the same system with more atoms, with your criteria you need a higher
values to reach ir, but if use an independent atoms criteria, the values
are the same for small and big systems.

MONOLAYER:
kpoints: I use 1 point in the vacuum direction, and I do not found any
difference when I use 2 or more kpoints in this direction. The bands
dispersion in this direction is negligible.
cell: The 2Dshape flag does not change your cell because it does not change
the xy area, how you have an orthogonal a,b vectors and a=b. The broken of
this symmetry is energetically not allowed. For this, 2Dxy flag works and
minimize the cell in xy directions, and reduce the stress.
If is it okay the parameters deviates from the bulk? I do not know for the
case of Monolayers... but for me makes sense, in your calculation the xy
area reduces to maximize the interaction with the neighborhood. If you want
do a vc-relax the 2Dxy is the correct flag.
Vacuum: 30 bohr should be enough... In my systems I do a similar vacuum
test, and 30 bohr gives the same result as 200 bohr.
I guess your system have charged surfaces, dipoles or something like that.
If is a bynary semiconductor AB... a monolayer gives you a "A type" and a
"B Type" surfaces, how A and B have different electronegativities, exists a
dipole between both surfaces, that could explain the energy behaviour with
vacuum.
I recommend, do a post-processing calculation for the monolayer, calculate
the electrostatic potential (V) along the vacuum direction... If V is not
plane in the vaccum region, do you have a dipole on the monolayer. You can
fix it using a dip-field on the vacuum... with it you add a "V step" on the
vacuum, giving you two V planes on vaccum, "two types of vaccum".
Other way, is making a monolayer with equal surfaces, like A-A or B-B
(adding a plane of atoms)... this fix the dipole in most cases... but I do
not know if this is still a monolayer or what you want.

Regards

Lucas Lodeiro
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20200102/87678470/attachment.html>