QuantumATK Forum
QuantumATK => General Questions and Answers => Topic started by: spin on June 30, 2011, 08:46
-
Dear all
I have a problem with drag-and-drop the molecule configuration into costum. I want to have a two electrode system with a molecule in central region but it has not any reaction and the only thing in it is the default molecule in VNL!!!!
tnx
-
can you post the molecule that you are trying to drag, such that we can use it for testing
-
can you post the molecule that you are trying to drag, such that we can use it for testing
the molecule is Adamantane (C10H16).
-
Try the following:
import the file adamantane.xyz into the builder,
next drag the molecule from the builder into the molecular junction custom builder
set repeat to 4x4
and atoms to 19 and 31,
this works for me, see attached
-
Thank you very much
But the shape that is desired is different, please take a look on it.
-
Your problem is quite possibly related not to the molecule itself, but the way you are trying to use it with VNL. So please be careful to specify all details, it's not enough to know which molecule, but the problem might lie in the file you are importing, for example. If it's a PDB file, VNL will not know how to handle it. Etc. The devil is in the details ;)
-
I made the Adamantane molecule in VNL and I got the coordinates of all atoms in VNL by myself and the shape of molecule has the specific format adjusted with VNL, is there any way to make the molecular junction without using costum?
-
Fine, but as the post above by kstokbro shows, provided the molecule is in the correct format, there is no problem to set up the system. I suggest you use his approach.
-
Try the following:
import the file adamantane.xyz into the builder,
next drag the molecule from the builder into the molecular junction custom builder
set repeat to 4x4
and atoms to 19 and 31,
this works for me, see attached
t
should I depict the molecule in 2 dimension? whether VNL can not calculate the 3 dimension molecules?
-
Of course ATK can handle 3D molecules :) I didn't note the actual molecule was a bit different in your case, however at any rate, as I have said many times, what seems to be the matter is what precisely you are dropping on the "molecule dropzone". Is it a file, and if so which file format, or do you drag from another Builder, or an Editor? If you try the approach described in kstokbro's post, then just adapt the file to represent your molecule instead of his, let's see where that gets you. Or post your molecule (coordinates, the file, not a picture from Wikipedia) so we can try to help with the actual problem, instead of this kind of abstract discussion which I think doesn't really lead anywhere.
-
The Adamantane configuration made by VNL and also picture of the 2 electrode system are attached. ???
-
and I tested the xyz file sent by kstokbro, it doesn't work for me. I think the custom molecular junction works only for the default molecule.
-
All works fine. I took your Python file, dropped it into the Molecule Device Builder, set atoms to 14 and 25, and right absorption height to 2.5 Å (just for something), and all looks ok.
You will need to adjust the exact distances, of course, and perhaps also set the repetitions to 4x4.
-
Thank you very much
It is nice now and there are no problem with it.
for havin the linear electrode ( only a chain) the configuration of adamantane is changed, do you have any proposal for having a linear electrode?
-
Why changed?
The electrodes should generally speaking be chosen to reflect the problem you are trying to simulate. The 1D density of states and band structure of an atomic chain is usually quite different from an electrode made of e.g. a gold 111 surface, so the results will be different. Which you prefer, again, depends on what you are trying to achieve.
-
I want to have a two electrode system like attached figure. In molecular junction/custom, the width of central region can not be determined by user! I am making effort to put the Adamantane molecule in the central region between two electrodes system ( 2*2 and 1*1) , then compare the conductance of molecule for this two state. on the other hand I have two value of central region, 9 and 11 Ang.
Can I do it in custom?
-
The Custom Builder (the one for Molecular Junction) is not intended to solve all problems :) It's function is to easily be able to insert the molecule between the electrodes. For further manipulations, you are relatively easily extend the unit cell by hand, in the Python code, or send the structure to the more generic Builder, where you can
1) convert it to bulk, by clicking the Bulk icon on the left
2) change the unit cell
3) convert back to device by clicking the Device button (same place)
-
Dear Dr. Blom, thank you for your excellent responses. yes, I've tried the builder in VNL, in builder I can not have the molecule in central region with gold electrode in two side (device form). First we put the gold, then we define the lattice and number of repetitions for central region and afterward we can transmit it to device mode. when I put Adamantane as my molecule, the electrodes also changes from gold to Adamantane. The final device would be two electrode (Adamantane) and an ADM molecule between them!!!
I think it would be solved by python code, I haven't already written a python code but I start it and I'll come back ;)
Best
-
I think I know what the problem is, but it's hard to explain (easy to show...). The issue is probably that your surface has only 3 gold layers, so the switching from bulk to device can't find any periodicity. But you can instruct it manually to use the same value as before. Then there should be no molecule in the electrode.
-
Dear Dr. Blom
Is it possible to have an non-alternating device? in a way that atoms in electrodes be independent than central region. some times when the central region width be small, a portion of the central molecule would be appear in the electrodes!!! :'(
-
Yes, that is inevitable, and it's crucial to understand why. In ATK, there must be a copy of the electrode inside the central region. If not, not only does the algorithm break down, but you almost certainly do not have enough screening to adiabatically go from the highly non bulk-like electronic structure around the molecule to the bulk like electrode.
See http://quantumwise.com/documents/tutorials/latest/UpgradeGuide/index.html/chap.upgrade.html#sect1.upgrade.twoprobe for some more details.
-
Thank you, I got it. And how the distance between electrodes in ATK can be defined?
whether that is the C vector measure when we define the central region?
-
Yes, the length of the central region in C becomes the distance over which the bias is applied, if we by this mean the distance between the electrodes. It is obviously not, however, the distance between the metal surfaces.
-
In the custom/molecular junction for example we define a 2*2 electrode [Au(100)] and minimum length of 4, now; if the molecule's vertical length be more than the vertical length of electrode, the configuration of molecule would be changed. In this case, if I calculate the quantum conductance(G) of the mentioned molecule, whether the results would be correct or not?
-
The surface, or rather the unit cell in the transverse plane, must be large enough that the molecule fits, otherwise you will get completely wrong results. And if the cell is only "just big enough", there may be electrostatic interactions between the repeated copies in X/Y.