QuantumATK Forum
QuantumATK => General Questions and Answers => Topic started by: sitangshu on July 8, 2011, 09:20
-
How can I make an MoS2 Nanowire of 1 micro meter length??
-
It may be too long for the simulation or the structural visualization.
-
can you let me know for 10nm long MoS2 nanowire?
Thanks
-
I think it is important to understand why you want a 1 micro meter MoS2, since as long as the wire is homogeneous in structure you will just have to simulate a single piece/sample of the wire, that properly is no longer than 5 Angstrom.
All the properties you could derive from the simulation will be equally good for this length, and you will be able to perform the simulation in minutes.
-
Hi,
This 1 micro meter length is needed to put it between 2 electrodes and to measure the electrical properties like resistance. Since, the resistance depends whether the system is diffusive or ballistic, I think the results for MoS2 nanowire will be entirely different from that of 5 Angstrom and 10 nm.
Isn't it?
However, your suggestion may help me to get the band structure and the band gap of the wire.
Hence, can you please let me know how to build such MoS2 nanowire asap ?
Thanks,
-
For a perfect MoS2 nanowire (i.e., defect-free or no disorder), its resistance or conductivity may be ballistic, like the case of CNT. If you wants to study of the effect of defects and disorder on the electronic transport properties of such system, you may need a longer nanowire in the simulations. Otherwise, the use of too long MoS2 wire will be a waste of computing resource.
The generation of the atomic coordinates of a MoS2 nanowire is not unique, and also cannot be obtained automatically using VNL. A possible way for a MoS2 (001) nanowire is outlined as follows:
1) Create the structure of bulk MoS2, which may be selected from the crystal DataBase of VNL;
2) Drag (transfer) it into the "Builder" tool of VNL;
3) Use the "Repetition" in the "Builder" to extend the x and y directions of MoS2, according to the wanted radius of nanowire;
4) According to the wanted shape of nanowire, selectively delete some of atoms;
5) Adjust the lattice vectors of A1 and A2 to make the unit cell orthogonal;
6) According to the wanted length of nanowire, extend the z direction of MoS2 using the "Repetition" in the "Builder".
-
Hi,
Thanks for your quick reply and valuable suggestions.
Can you let me know how to passivate MoS2 nanowire dangling bonds with Hydrogen. Is there any script available??
-
No such script file at current stage.
Since the atomic bondings in MoS2 are quite different from the case of silicon, the passivation of dangling bonds in MoS2 by hydrogen may not make sense for some cases, e.g., the dangling bonds around Mo atoms.
-
How can I make MoS2 monolayer, bilayer, quadrilayer ??
-
It is quite easy.
First, create a bulk system of MoS2 using the "Builder" or find out the crystal structure of bulk MoS2 from the "Crystal DataBase" of VNL.
Then, use the "Reptition" " in the "Builder" to extend the z direction of MoS2 to 4 atomic layers and modify the lattice constant of c (c should be set large enough to get thick vacuum layer). This will give the quardilayer. For the monolayer and bilayer, they can be obtained by deleting the atoms in the 2 or 3 atomic layers.
-
Dear Sir,
I have done the the band structure calculation of the layers according to your previous suggestion but I can't get proper band structure and also get very low band gap value. Attached please find the builder file for the quardilayer. I have checked with both DFT and Huckel method with various k-points. Please let me know if there is any mistake in my model.
I also want to know that if I just repeat in z direction by "Reptition" in the "Builder" to extend the z direction of MoS2 to 4 atomic layers then, will it take care of the Van der Waals interactions between the layers?
Thanks
Ramkrishna
-
The setup for the bulk configuration of 4 atomic layers of MoS2 may be wrong. In particular, the value of "vector_b" for the crystal is not correct. Since MoS2 is a hexagonal crystal, the a and b lattice vectors for the primitive unit cell can be described as follows:
vector_a = [0.5 *a, -0.5*sqrt(3.0)* a, 0]
vector_b = [0.5*a, 0.5*sqrt(3.0)*a, 0]
,where a is one of the lattice constants of bulk MoS2.
The manipulation mentioned by you does not take the Van der Waals interaction between the atomic layers into account. The Van der Waals interaction can be properly described the DFT-D2 method of Grimme or other Vdw-DFT functionals (see http://www.fhi-berlin.mpg.de/th/Meetings/DFT-workshop-Berlin2011/presentations/2011-07-15_Tkatchenko_Alexandre.pdf). However these are implemented in the current version of ATK.
-
I have checked with that lattice vectors but still can't get reasonable band gap. Earlier you had mentioned that I have to modify the lattice constant of b (b should be set large enough to get thick vacuum layer). So if I use the lattice vectors as you say in your previous post then how can I fit a thick vacuum layer? I can understand, something is wrong in my bulk configuration but can't find out properly. If it is possible then, can you please send me builder file?
I am using the current version of atk (11.2.3) but still I can't see no variation if I will increase the layers.
-
Sorry, there is a typo in my previous reply, in which I mentioned the increasing of lattice vector b for a vacuum layer. It should be lattice vector c.
For the coordinates of atoms in the primitive unit cell of bulk 2H-MoS2, they can be referred to the paper:
Ricardo Grau-Crespo and Rafael Lo´pez-Cordero, Phys. Chem. Chem. Phys., 2002, 4, 4078–4079.
The attached scripts for the bulk configuration of 2H-MoS2 and n-layers of MoS2 (n=1, 2, 3, 4) are constructed according to the atomic coordinates listed on the Table 1 of the above paper. The experimental lattice constants of 2H-MoS2(a = 3.16 angstrom, c = 12.296 angstrom) are used.
-
I have done the band structure calculation with that builder scripts using Huckel as well as DFT with k points (11X11X11), but I can't get any significant change between bulk, 1-layer, 2-layer band structure. According to the literatures( like http://pubs.acs.org/doi/abs/10.1021/nl903868w ) there should be a drastic change between these and the 3, 4-layers do not show any band structure at all. Please help me.
-
I am not getting any response. Please help me.
-
I have done the band structure calculation with that builder scripts using Huckel as well as DFT with k points (11X11X11), but I can't get any significant change between bulk, 1-layer, 2-layer band structure. According to the literatures( like http://pubs.acs.org/doi/abs/10.1021/nl903868w ) there should be a drastic change between these and the 3, 4-layers do not show any band structure at all. Please help me.
If possible, please upload the script files for the band structure calculations of 3layers and 2 layers of MoS2.
Since the Huckel method is empirical, the accuracy for the band structure of MoS2 may depends strongly on the used parameters. The meaningful thing is to the DFT calculations for the band structure in order to to compare it with the results in literature.
-
Dear Sir,
I have done the band structure calculation for 2 layers and 3 layers using the following attached scripts.
Regards
Ramkrishna
-
For your 3-layer structure, you have computed the band structure between Gamma and Z, but your system is not periodic in this direction (or rather, there is a big vacuum gap), so naturally there is no dispersion at all. If you want the band structure in the plane of the MoS2 you should keep the structure hexagonal and use the same route as for the other calculations, i.e. between G,K,M.
For the bilayer structure I suspect the results will depend strongly on the distance between the layers. How did you determine that?
-
I don't know exactly how to measure the distance between the layers in ATK. In the builder I just use (Transform -->repetition) to get that builder files. In literature the distance between the layers is shown 6.5 Angstrom and I thought that just doing repletion only it maintains that distances. Can you please let me know how can I measure interlayer distances and how can I change that also?
-
The atomic layer distance can be adjusted easily by modifying the z coordinates of atoms. On the other hand, the atomic layer distance can be measured by numerical operation of the z coordinates of Mo atoms in two neighboring layers.
-
Thank you for your kind reply. The DFT calculation gives proper curvatures of bands for bulk but the indirect band gap (~0.80 eV) is quite lesser than the band gap which is mentioned in the literatures (~1.2eV). On the other hand Huckel method is giving totally wrong band structure. So, to achieve that band gap what modification should I do for the band structure calculation. Attached please find the file for DFT calculation.
-
Some comments to the setup in your script file:
1). The k-mesh is not set properly. The size of k-mesh should be properly scaled according to the length of the lattice vectors. Since the lattice constant c is roughly 4 times larger than the lattice constant a, the size of k-mesh along the kz direction should be smaller than the one along the kx and ky direction. So, "k_point_sampling=(11, 11, 3)," may be more reasonable than "k_point_sampling=(11, 11, 11)".
2). To reproduce the results in literature properly, one had better choose the same exchange-correlation functional with the one used in literature, use the more accurate basis set such as DZP. In addition, the same structural parameters should be used.
-
To be clear, using 11 points in Z for this system is not wrong, it just takes extra time without perhaps improving the accuracy (although this should be checked). The script you attached is using the DoubleZetaPolarized basis set, since that's the default.