Messages

5086

General Questions and Answers / Re: fixed spin

« on: June 2, 2009, 21:07 »

Interesting idea. Although, I'm not sure the results will be different compared to if you just allow the magnetization to be free. ATK will adjust the magnetization to its minimum energy configuration for each lattice constant, so in the end you should end up in the same global minimum (provided it's not blocked by very high energy barriers separating local minima, but that's a risk you always face with any kind of optimization).

In the end, the magnetization is just another degree of freedom, so the situation would be quite similar if you try to optimize, say, the a and c lattice constants of a hexagonal structure by keeping a fixed (lattice constant) and scanning c (magnetization), or keeping c fixed and scanning a. Running the self-consistent loop is pretty much the same as scanning the magnetization.

5087

General Questions and Answers / Re: fixed spin

« on: June 2, 2009, 15:52 »

No, it can only be used for molecular calculations.

If you have a particular application for this functionality in mind, it would be interesting if you could share it with us!

5088

General Questions and Answers / Re: a question about the source code to define the broadening function in ATK

« on: June 2, 2009, 13:18 »

These expressions are actually equivalent (except the factor 2, which probably is a spin degeneracy) since GA = (GR)+ in the non-interacting case.

5089

General Questions and Answers / Re: a picture

« on: June 2, 2009, 11:06 »

The pictures show the "local density of states" as computed with the function calculateLocalDensityOfStates(), available in VNL under the analysis options for two-probe systems. What you do to get the different plots is to choose the energy at the Fermi level, at the top of the valence band, and bottom of the conduction band, respectively.

Store the results in a VNL file and use VNL to plot the results (Nanoscope).

5090

General Questions and Answers / Re: ATKError

« on: June 2, 2009, 10:56 »

It's hard to know exactly what goes wrong in your case, but generally speaking this error arises for 3 reasons:

* The use of the constraint "DensityMatrix" when the system converges to zero charge
* Atoms placed on top of each other
* Converging to very low tolerance, like 1e-16

In general the final NC file produced by a the optimization is indeed a properly converged calculation, so you don't actually need to rerun the calculation, just restore it and perform the analysis tasks. Also take care to extract the optimized geometry from the VNL file!

Perhaps, when I think about it, your problem is a simple matter of removing the checkmark for "Only use initial density" on the self-consistent tab in the NanoLanguage Scripter in VNL?

5091

General Questions and Answers / Re: About MPSH

« on: June 2, 2009, 10:51 »

Nanowires are often semiconducting, but the smaller they get the more metallic they become. Thus the wires people typically measure on are almost always semiconducting, but for narrower diameter tubes you should really compute the band structure to see how it looks.

The MPSH concept is designed for molecules between metal surface, put simply. Therefore it doesn't really make sense in your case, unless there is a specific impurity for instance, then you can project onto that atom (and it's neighbors, typically).

5092

General Questions and Answers / Re: About MPSH

« on: May 31, 2009, 22:28 »

I suppose the system is metallic in the central region.

5093

General Questions and Answers / Re: How to calculate the density of states (DOS) of isolated region in graphene?

« on: May 29, 2009, 15:29 »

Yes and no

You approach will give the molecular energy levels of the isolated segment corresponding to the central region. This does not take into account the possible shifts of these levels induced by the coupling to the electrodes. Moreover, if you just take out a piece of graphene ribbon like that without considering the edge termination etc, you may get the wrong results. Also, the alignment of the level w.r.t. the Fermi level of the transmission spectrum will not be correct.

A more proper approach is to use the functionality in ATK called MPSH spectrum. Some details can be found in the manual, and you can also search the Forum for "MPSH". This way you get is the molecular spectrum taking into account the electrode coupling, normalized to a common Fermi level.

I have not made the bilayer structures, because as I said there it depends on how the edges should look like. If you have an idea about how you would like them, then I can look into the Python code to make it.

5094

General Questions and Answers / Re: How to calculate the density of states (DOS) of isolated region in graphene?

« on: May 29, 2009, 08:49 »

They don't explicitly write how they compute the "scattering region DOS" shown in the APL paper you mention (which is a very recent and nice application of ATK to graphene), but my guess is they perform a molecular calculation for the central region (more specifically, I think they use the MPSH functionality in ATK for this), and then plot the energy levels broadened by some Gaussian distribution. It doesn't really matter, they could just as well just have plotted the DOS as sharp molecular level, since the peaks are so narrow anyway.

5095

General Questions and Answers / Re: How to set up bilayer graphene model?

« on: May 29, 2009, 08:37 »

Yes, we sure do use ATK for the simulations

I guess some care has to be taken when setting up a bilayer structure, when it comes to the edge termination. It's probably a good idea to see how others have tackled this (e.g. APL 92, 223106 (2008)).

5096

General Questions and Answers / Re: How to set up bilayer graphene model?

« on: May 28, 2009, 10:28 »

You can use graphite for this, by just changing the c lattice constant while keeping the layer separation fixed. Are you considering ribbons or infinite graphene? Or rather, where do you take the graphene model from, from the Crystal Cupboard or the graphene tutorial scripts?

5097

General Questions and Answers / Re: How to calculate the density of states (DOS) of isolated region in graphene?

« on: May 28, 2009, 10:26 »

The question is a bit ambiguous. The concept of density of states applies to a system which is infinite and periodic in at least some directions. Therefore, there is no DOS for an isolated region... Or, well, you can of course say that a molecule has a spectrum of delta function-like spikes, which is like a DOS.

But perhaps you mean more something like the DOS for an infinite 2D sheet of graphene, or something like that?

5098

General Questions and Answers / Re: Speed issue in Two probe configuration

« on: May 26, 2009, 10:43 »

The 8 and 16 CPUs, how are the distributed w.r.t. RAM? That is, for 8 CPUs, for instance, are they located in 8 separate boxes, or in 4 machines with two sockets, or 2 machines with 4 sockets, for instance? Same for the 16.

Competition for RAM, cache and communication bandwidth will depend critically on these factors.

5099

General Questions and Answers / Re: Speed issue in Two probe configuration

« on: May 26, 2009, 09:20 »

I was just about the suggest running without MPI...

Cache size can certainly play a role when you are shuffling huge amounts of data around.

Another thing, do you have the same OS on both machines?

When you measure the time per SCF cycle, is that done by taking the total compute time divided by number of steps, or timing of each cycle (using verbosity=20)? If you use the total time, note that the calculation may take different number of steps on the two machines in the MPI case, so take care to normalize by each respective number of steps...

5100

General Questions and Answers / Re: question about calculate the local_density_of_states

« on: May 26, 2009, 09:11 »

Did you figure it out? I got a notification of a reply to this topic, but it seems you deleted you post? Hope it means all is working