QuantumATK Forum
QuantumATK => General Questions and Answers => Topic started by: NW on February 9, 2018, 10:09
-
Hello,
I tried to build a Bismuth nanowire but it is not possible in VNL due to lattice structure of Bi (Rehombohedral). Is there any solution for this problem in VNL?
Thanks
-
Thanks Berna but as I know to build a nanowire in VNL we need a structure in cubic form (simple, Fcc, Bcc).
-
ATK expert,
Could you please help me how I can build a bismuth nanowire in [110] and [111] directions?
Many thanks
-
I think you can transform the primitive unit cell to a cubic cell, using Supercell plugin (choose face-centered cubic cell) in the Builder, and then use NW plugin as usual.
Another option would be to cleave the NW in the desired crystallographic direction (which should coincide with the NW axis) using the Cleave tool in the Builder. Choose proper in-plane lattice vectors and then create a NW by increasing the lattice parameters (i.e. adding vacuum), and then manually delete some atoms to have a proper shape (faces) of the NW.
-
To add to Petr's approach, you will also need to manually specify the Lattice type of the structure as, for example, Simple Cubic, under Bulk Tools -> Lattice Parameters. The you can create a nanowire. However, note that the last step may introduce a little strain in the structure, so you should make sure to relax the nanowire if needed.
-
Hi, I followed the steps and build a BiNW with 1.1nm diameter in [1 1 1] direction. I calculated the band structure of the NW but the value of the band gap is quite different with the value that mentioned in the [10.1088/1361-648X/aa4e63, Electronic and structural properties of rhombohedral [111] and [110] oriented ultra-thin bismuth nanowires].
I attached the script file and the band structure image.
Please help me what should I do?
-
Did you use the same computational parameters as in the paper? if yes, please provide the script so we can have a look at it.
-
Dear Ulrik,
In the given paper authors used OpenMX package with numerical atomic orbital (NAO) basis sets and DFT/LDA method. I used OMX medium basis set in ATK with DFT/LDA method. I have changed my method to DFT/GGA and DFT/mGGA with and without spin orbit interaction (SOC) but the band gap value did not change noticeably. Beside that I repeated my calculation with Tier 4 basis set. I attached the script in my previous message.
-
Dear ATK experts please tell me what should I do?
Thanks
-
A few points:
1) It does not look like you have geometry optimized the nanowire. In the paper you mention, they clearly relax the structure before doing band structure analysis.
2) If you want to consider the band path G -> K, you probably want to convert the lattice from UnitCell to Hexagonal (do that using the Lattice Parameter widget).
3) I think you should use SG15 pseudopotentials instead of OMX, and probably PBE instead of LDA.
4) Also compute the DOS: should clearly indicate if the hydrogen passivation worked as it should (removing spurious states around the Fermi level)
5) As a starting point, don't worry about spin-orbit coupling.
I attach a script that basically does this. Hope it is useful.
BR,
Jess
-
Dear Jess,
Thank you so much for your helpful suggestions. I will run your script.
-
Hi,
The structure has been relaxed. The attached image shows the structure after relaxation. As you can see the structure has a lot of distortions! I attached the script file of the structure after fully relaxation.
-
Bismuth has a fairly open crystal structure, so some re-arrangement at the surface is to be expected. Do you have a reference for how such a nanowire is supposed to look?
-
Yes, this ref. [10.1088/1361-648X/aa4e63, Electronic and structural properties of rhombohedral [111] and [110] oriented ultra-thin bismuth nanowires]. Fig. 1a in that paper shows the BiNW in [1 1 1] orientation after relaxation.
-
Thank you, we will take a look at it.
-
Thank you Ulrik.
I am completely confused about this issue ??? ???
-
I have looked at the reference, and I do not think it is surprising that the structures are a little different. Whether that difference in the structure is important for what you want to model is up to you to decide.
As mentioned above, we introduce some strain into the structure when it is forced to a cubic lattice, and it is probably the relief of that strain you see. If you wish to reproduce exactly the calculations in the paper, you need to construct the nanowire directly from the rhombohedral crystal structure, which is unfortunately currently not possible in VNL.
-
Thanks a lot dear Ulirk. Do you think if I construct a BiNW in other software like Material Studio it is possible to continue my calculation in ATK?
-
Certainly, we support importing structures in other formats, such as .xyz. See these links for more information:
https://docs.quantumwise.com/tutorials/builder_manual/builder_import/builder_import.html
https://docs.quantumwise.com/tutorials/vnl_io_import/vnl_io_import.html
https://docs.quantumwise.com/tutorials/vnl_import_export_files/vnl_import_export_files.html