Fermi Surface plot

[ramkrishna]

Dear Sir,
      Can you please suggest/ provide a script to plot the Fermi surface of an orthorhombic system??

Regards
Ramkrishna

[ramkrishna]

Dear Sir,
      Is it possible to plot the Fermi surface or generate data for the Fermi surface (2D or 3D) in ATK, then it will be helpful for me.

Regards
Ramkrishna

[zh]

1. generate the k-points of a mesh M x N in the first BZ;
2. calculate the eigenvalues of these k-points and print out the data in the XCRYSEN format (i.e., .bxsf. For the details, you can refer to the http://www.xcrysden.org/doc/fermi.html) or the 3D format (kx, ky,  eigenvalue).
3. visualize the data in the format of .bxsf by XCYRSDEN; or visualize the data in the 3D format by using gnuplot or matplotlib in python to draw the contour plot.

[ramkrishna]

Dear Sir,
     Can you please provide any python script to calculate the eigenvalues for each k-points of a mesh MxN in the 1st BZ from ATK so that we can print in XCRYSDEN format or any other 3-D format?? It will be very much helpful then.

Regards
Ramkrishna

[Anders Blom]

I'm on Windows and my Cygwin doesn't have X-Windows, so I couldn't test it yet, but this is an attempt at such a script. Please share your experiences, and cool pictures if it works!

Edit: Newer version of script farther down

[ramkrishna]

I will try this and let you know 

[ramkrishna]

Dear Sir,
     I am trying to find the Fermi surface in xcrysden through generating bxsf file from the script (fermi_surface_XCrySDen) that you have given. But it is showing an error when I am going to load the file gold.bxsf in xcrysden (please find the attached bxsf file) whereas the xcrysden can easily detect its example bxsf file. The error massage is "ERROR: an error occured while executing fsReadBXSF program". Is it related to an error in saving a bxsf file? If you have time, can you please check this one once? it will be then very much helpful.     

Thanks
Ramkrishna

[Anders Blom]

I have an idea what this is due to. Let me see if I can get XCrysDen up and running so I can test it, before sending an updated script version.

[Anders Blom]

XCrysDen 1.6 (which however gets installed in a dir called 1.5.18...) segfaults on my Cygwin when I try even the example, so I still can't test this 100%. However, it doesn't complain any more about the file at least with the newly attached version of the code PLUS - very importantly in case you are doing this on Windows - you must convert the file to Unix format using dos2unix! Windows Python writes DOS line endings (CR+LF) but XCrysDen only accepts Unix-style files.

If you get some nice plots - please share them!

[ramkrishna]

Dear Sir,
      The script runs well and it produces the bxsf file that xcrysden can accept . I have generated the bxsf file to check the Fermi surface of copper. Please find the attached figure (Fig. 1, cu.png). However, it does not reproduce the existing results of copper (Fig. 2, Copper_ref.png). I have attached the bxsf file with this also. Please check once.

Thanks
Ramkrishna

[Anders Blom]

The file format may be acceptable to the program, but there is actually a linebreak missing (my fault) before the second, third etc "BAND" entry. Not sure if that has anything to do with your plot, but you can try to change it manually by just editing the file. I'll amend the code a bit.

Which band did you select for the plot? BAND 5, I would assume?

I don't know, I have no experience with this (yet!). Unfortunately XCrySDen doesn't run on my Windows machine in Cygwin, it just segfaults... So, it's hard to do any testing, really. I'll look at the code I wrote, and the file format, to see if there was something I missed.

[Anders Blom]

Good news! It's simply because you have too few points. See figures attached for gold with 4x4x4 (look odd, like yours) and 20x20x20

[ramkrishna]

Dear Sir,
      Its really great. Thanks a lot.

Regards
Ramkrishna

[ramkrishna]

Dear Sir,
     I have one more request here. XCrySDen gives a very nice 3D plot of the Fermi surface but it does not have the facility to plot the 2D Fermi surface in the kx-ky plane. If you help to plot the 2D Fermi surface also, it will be very much helpful then.

Thanks
Ramkrishna

[Anders Blom]

That would be a "Fermi line" - i.e. just a simple contour plot at the value E=0 would do it. Do you have a particular example in mind?