Devices and electrodes

[Kons]

1. I have atomic configuration of a "device", XYZ-file, already optimized from external source. There is no inherent periodicity in the structure.
2. I import XYZ-file into VNL, and start the construction of device.
3. If I will try to use "Device from bulk", I will have something completely wrong, because all interatomic distances are different.

The question is: how to declare part of the structure as electrode without repetions of device's atoms?

As an example, imagine that you have system similar to this http://quantumwise.com/documents/manuals/latest/VNLTutorial/index.html/chap.gnr.html#sect1.gnr.preface, but all atoms of the initial device (between red lines) were randomly displaced or missing.

[Anders Blom]

You can't create a device in this way. The entire concept of the device configuration is based on the fact that the electrodes are a periodic (1d, 2d, or 3d) bulk crystal, so a system without inherent periodicity is only a starting point for the central region, but you also need electrodes. Without knowing anything more about your system we can't help in any details, but possible options are to make a repetition of what you have and call the outer parts electrodes, or you can add perfect electrodes to the edges of your system.

To assist further we would have to know at least a little bit about the system you are trying to construct.

[Kons]

Ok, this is scheme.



"Device" is central region, connected to electrodes with graphene nanoribbons. Nanoribbons itself is a part of device. The atomic structure and geometry of complete device is optimized separately (T=300K), peridodic in Z-direction, so that actaully the atoms of the left nanowire were connected to the atoms of the right nanowire during optimization.

Consequences:
1. There is no short-range periodicity in order of carbon atoms of nanoribbons.
2. Nanoribbons are non-flat, so that I can't extend them with ideal perfect nanoribbons. I can't even smoothly connect them to device.

[Umberto Martinez]

Compare your scheme with the one for a device geometry in ATK reported in the Introduction to transport calculations with ATK tutorial.

In particular, check the figure here http://quantumwise.com/documents/tutorials/latest/ATKTutorialDevice/index.html/chap.input.html

So your scheme would correspond to the scattering region of the figure in the tutorial. You need to extend your GNR at the edges to create the electrodes.
Your GNR in the electrodes does not necessarily need to be flat as you say but a portion of it (the electrode regions in the figure) needs to be periodic.