QuantumATK Forum

QuantumATK => General Questions and Answers => Topic started by: Parham on January 9, 2016, 07:01

Title: VNL as Quantum Espresso frontend
Post by: Parham on January 9, 2016, 07:01

Hello

I am trying to use VNL as the frontend to Quantum Espresso. It works fine with the cubic lattices . I successfully reproduced tutorial results on silicon. I did a whole new work with GaAs. I did the same with zincblende CdSe. Unfortunately my experience with wurtzite CdSe failed. The lattice is hexagonal and ibrav=4 for this case while for the all previous was ibrav=2. The Quantum Espresso analyze the inputs but I cannot do the Grid operations with scf output or plolt the bandstructure afte band calculation of QE. The Grid operations pops up an error screen that cannot map (30,30) array into (30,48) array and the bandstructure is completely black.
Am I should change some setting to recover the issue or this VNL feature is not yet complete to cover the hexagonal latticess?

Sincerely Yours
Parham

Title: Re: VNL as Quantum Espresso frontend
Post by: Jess Wellendorff on January 13, 2016, 14:29

Dear Parham.

1) The issue with Grid Operations should be fixed by an upcoming bug-fix to the Quantum ESPRESSO plugin. Release is very close.
2) Could you attach the input/output for the SCF and bandstructure calculations, such that I can see what goes wrong?

Title: Re: VNL as Quantum Espresso frontend
Post by: Parham on January 17, 2016, 10:19

Dear Jess

First, pardon me for the typing errors in my last post.

Second, I am looking for the update to Quantum Espresso plugin impatiently.

Third, I attached scf.in, scf.out, bs.in and bs.out to this  post. You are expert and you already know that the bands calculations should be done around high symmetry points. I should substitue the W point with M point in reciprocal space but I didn't since I am stopped watching density isosurfaces after scf calculations.

I really need your advice on making progress using VNL QE plugin.

Look to hear from you
Sincerely Yours
Parham

Title: Re: VNL as Quantum Espresso frontend
Post by: Jess Wellendorff on January 19, 2016, 11:57

Dear Parham.

The updated ESPRESSO plugin (v1.95) should now be available. Use the AddOn Manager to download it. Fixes your Grid Operations issue.

Regarding the BS calculation: The issue is that ATK cannot match the k-points you have computed with ESPRESSO to paths between high-symmetry points of the hexagonal lattice. The reason is that you use high-symmetry points of the FCC lattice type. I have attached a slightly modified bs.in, which maps out paths along the A-G-L-M path for the hexagonal lattice. The Bandstructure Analyzer now shows the bandstructure (see bs.png).

Best,
Jess

Title: Re: VNL as Quantum Espresso frontend
Post by: Parham on January 20, 2016, 12:35

Dear Jess

Thank you very much. My problem is now solved.

I am new user of VNL. Is it possible to use VNL for finding high symmetry points coordinates? I can go to the Builder and then choose Brillouin Zone viewer. It draws Wigner Seitz cell and shows the high symmetry points in green. I need the coordinate of shown points but I cannot view the numerical valuse of that points in k-space. Some points like X, Gamma and L can be guessed visually but it is impossible to guess W point coordinate in FCC lattice visually.

Sincerely Yours
Parham

Title: Re: VNL as Quantum Espresso frontend
Post by: Jess Wellendorff on January 21, 2016, 08:37

The symmetry points for some crystal type, e.g. FCC, is a property of the Bravais lattice (i.e. encoded into ATK). You can easily use NanoLanguage to query for these symmetry points once you have a bulk configuration:

Code

bulk = nlread('Silver.nc', BulkConfiguration)[0]
lattice = bulk.bravaisLattice()
lattice.symmetryPoints()

Run those lines in a Terminal (perhaps as a script). The symmetry points and their coordinates are printed:

Code

{'G': array([ 0.,  0.,  0.]),
 'K': array([ 0.375,  0.375,  0.75 ]),
 'L': array([ 0.5,  0.5,  0.5]),
 'U': array([ 0.625,  0.25 ,  0.625]),
 'W': array([ 0.5 ,  0.25,  0.75]),
 'X': array([ 0.5,  0. ,  0.5])}

Title: Re: VNL as Quantum Espresso frontend
Post by: Parham on January 25, 2016, 05:53

Thank you very much.