QuantumATK Forum

QuantumATK => General Questions and Answers => Topic started by: Emilio_bassini on October 26, 2010, 18:29

Title: can a charge located near a molecule modifies its conductivity?
Post by: Emilio_bassini on October 26, 2010, 18:29

I swear this is the last question (for this month ;) )
I got a particular job:
to try to found out if a charge located near the molecule can change its habit to conduct.
With gaussian i made a moecule and i put a charge in a well known position.
can i
1) export the new configuration with the charge in VNL?
2) to study the conduction with the molecular junction system?

if it is possible,i'd finally like to know if i can use directly with vnl without using gaussian or other software...

i hope someone could solve this problem so i'll finally be able to finish my reserch

Title: Re: can a charge located near a molecule modifies its conductivity?
Post by: Anders Blom on October 27, 2010, 00:24

You can certainly transfer the structure from Gaussian to ATK, using XYZ format (use OpenBabel to convert it to XYZ, if needed). You cannot however directly transfer the "charge". It's not possible to introduce a charge in a two-probe system, because it would just run away anyway. The only way to have a charged molecule is to shift the Fermi level such that the central part can accept one more electron.

The SET example is different, it's not set up as a transport system, and then we can place a net charge on the system, but in the transport situation the electrodes are infinite reservoirs so they will just swallow the extra electron.

The Fermi level can be shifted by a gate, or as we did in ATK 2008.10 by artificially moving the Fermi level in the center. That's not possible in 10.8, but it might make a re-appearance later. For now, one should instead introduce real gates (which is a bit costly, memory-wise, so you may want to try using the Huckel method at first).

I didn't get the last point, about not using other software... :)

Title: Re: can a charge located near a molecule modifies its conductivity?
Post by: Emilio_bassini on October 31, 2010, 16:22

what i tried to mean was...
is there a way to build a system made of a molecule and a ion directly on VNL and  subsequently to study its I-V curve...
I said that because i find VNL quite hard to use but gaussian is even more complicated..

Title: Re: can a charge located near a molecule modifies its conductivity?
Post by: Anders Blom on November 5, 2010, 10:49

Sorry for the delay in answering.

It is not really possible to insert an ion in a two-probe system. Any excess charge will be washed away in the self-consistent procedure, since the total charge of an open system is not conserved. In order to obtain a net charge on the system, one should instead tune the Fermi level, by introducing a gate.