QuantumATK Forum
QuantumATK => General Questions and Answers => Topic started by: Shan on January 4, 2017, 19:43
-
Dear QW Team,
I am using VNL as GUI for Quantum Espresso. First, I implemented the tutorial of Silicon, VNL worked fine for me.
When I went on to Graphene (two atoms) bandstructure calculation, the VNL shows empty band structure plot. However, when I plot the same using GNUPLOT it shows perfect bandstructure and zero band gap.
Why VNL could not show band structure of Graphene, however it is working fine in displaying DOS plot of graphene. I really wasted a lot of time thinking that the problem is with my calculation.
Please clarify.
Thank you very much for your time.
-
That is weird. In the VNL figure the band structure from Gamma to Gamma, that is, no k-points path is selected in the Brillouin zone.
Did you also set up the system using VNL? Which version of QE are you using?
Regards,
Daniele.
-
Yes sir... the system has been setup using VNL only.
I am using latest version QE 6.0.
The VNL version is 2016.3.
-
Try these input files and make sure to update to the last version of the QE plugin (v2.3). Then you should be able to visualize the bands of graphene correctly.
Notice that we have found out a small numerical issue with the way k-points labeling is handled for the K point, so the "K" label will now show up in your band structure plot.
Also, make sure to run the calculations in a clean folder, I manage to reproduce your problem only when I re-run the calculation in a folder that already had some QE output in it.
Regards,
Daniele.
-
Hi Daniele,
I had the same problem as Shan did, and after I corrected values of k-points (G,K,M,G) and do again bands calculations, VNL only shows Gamma,M,Gamma. I do not see where is K. Even I used your setup files, VNL only shows G,M,G points. Did you also see only these three points in your figure? How can VNL show all 4 points G,K,M,G?
-
Yes there is a bug due to a difference between how QE and VNL interprets numerically the values of the k-points: VNL identifies automatically the K-point with a precision of 10-8, but QE prints out the k-points in the band structure only with 4 decimals. Hence, the K-point is not identified as an high symmetry point and it is not printed out.
I have already filed a bug which will be corrected asap.
I will post here when the bug is corrected.
-
Thank you, so we wait for your news.
-
Just to let you know. The bug has been fixed and the fixed version of the plugin should be available soon in the PlugIn repository.
-
Hi,
I also got into the same problem, so I updated the BandstructureAnalyzer to v. 2.6. However, it does not solve the problem.
Btw, I'm using VNL with the following specifications:
Build: 2016.3.7eb51b6
Platform: Linux
Python 2.7.11
What should I do to fix the issue?
-
The problem might be due to the number of digits that you have used to specify the band structure k-points in the QE input file. I suggest you to specify those points with a precision of 10-6, that is with 6 decimals, and report back if the problem persists.
Regards,
Daniele.
-
Hi Daniele,
Following your suggestion, I revised my pw.x input as follows:
K_POINTS {crystal_b}
4
0.00000000 0.00000000 0.00000000 20 !G
0.00000000 0.50000000 0.00000000 20 !M
0.33333333 0.33333333 0.00000000 20 !K
0.00000000 0.00000000 0.00000000 20 !G
The problem still persists: https://drive.google.com/open?id=0B4OPHMMwFxoQZlFkYzU5S0JXcGM
-
Or is there a possible way of manually labelling the high symmetry points?
-
Dear all QE + VNL users.
If you use VNL to plot QE bandstructures, I suggest you upgrade to the latest alpha release of VNL (version 2017.a2): http://www.quantumwise.com/products/download (http://www.quantumwise.com/products/download)
This release features a significantly updated plugin for QE bandstructures, such that e.g. fractional high-symmetry k-points are correctly detected.
-
Thanks Jess,
Installed the 2017.a2 version and it worked!
-
I am very happy to hear that!