GC-SCF calculations using quantum environ

[Johanna W.]

Hello,

has anyone experience on quantum espresso calculations with environ at a fixed electrode potential in grand canonical systems? If I start my calculation including lgscf=.true. in quantum espresso the calculation breaks and asks me to activate ESM and RISM despite including environ. I wonder if this is correct?

Best wishes

Johanna

[Brad Ayers]

Hi Johanna,

I have run a fair few GC-DFT calculations with environ enabled with no such issue.
My best guess is that the addition of environ support is a somewhat recent addition (a year or so) so you'll need a newer version of QE & Environ - i.e. QE 7.3 and Environ (3.1) 

see this commit on the QE main repo:  https://gitlab.com/QEF/q-e/-/commit/1f372311cd943634484181999b38033becdd01cc

I hope this helps, 
if not share your environ.in and pw.in and I will do my best to help.

kind regards,
Brad

[Johanna W.]

Hi Brad,

thank you for your answer! Actually that was the point - I was using QE 7.2 and seemingly that led to issues. With QE 7.3 it works now.

One remaining issue I have is that the Gaussian potential shift I need to add to the Fermi energy when using parabolic correction seemingly does not appear in my Output file as it seems to be in the example https://github.com/environ-developers/Environ/tree/master/examples/qe/pw/slab . I tried to increase the verbosity but seemingly it did not solve the issue. Is there another parameter that is needed to be activated for that?

I attach my environ and QE input files (not yet including the GC part, only a Pt slab with 0.25 ML oxygen in vacuum as example for a simpler system)

Best wishes and thank you very much

Johanna

[Brad Ayers]

Dear Johanna,

I am by no means an expert/developer, this is just what I have gathered from other discussions on here and the code itself, so take what I say with a pinch of salt:

The PBC correction in Environ is designed to handle electrostatics in systems with 2D periodicity, like slabs or surfaces. Here's what it does:

1. It modifies the long-range electrostatic potential to decay correctly in the non-periodic direction (typically perpendicular to the surface).
2. This ensures that the potential goes to zero far from the quantum mechanical system, providing a consistent reference point across different calculations.
3. It's particularly important for proper treatment of dipoles in these systems and for achieving consistent results regardless of the amount of vacuum in your simulation cell.

Now, regarding the shift you're not seeing in your output:

I don't believe you particularly need this shift value in most of your calculations. It is used more for band alignment, etc., where you need a non-zero electrostatic potential plateau, i.e., what you get in normal vacuum calculations.

It's not typically output directly, but can possibly be inferred from your results.

To find this shift, you'd typically:
1. Look at the planar-averaged electrostatic potential
2. Find the plateau value in the vacuum region
3. Use the negative of this plateau as your shift

I'm not aware of a specific parameter to output this shift directly, but increasing verbosity or checking post-processing data might help.

It might be worth reaching out to the Environ developers for more specific guidance, as once again, I am not entirely sure myself. It is not particularly well explained anywhere, to be honest haha!

Lastly, one more thing of note is that the parabolic correction is only designed for orthogonal cells, or at least orthorhombic where c is orthogonal to a & b, as the area perpendicular to z is computed by taking the volume and dividing it by the length of a3 along z; so it may not work perfectly otherwise.


I hope that all made sense, to summarise : 
I don't think you need to worry about it , unless you wish to do some form of vacuum/band alignment; furthermore, as you wish to use GC-DFT , you should be able to get your PZC and it will work as expected due to the potential at far away being set to 0! 

Best regards,
Brad