Messages

1

General Questions and Answers / What do we get with MolecularEnergySpectrum for a two-probe system?

« on: August 28, 2017, 08:14 »

We can set a device configuration for the parameter "configuration".  But what will we get with MolecularEnergySpectrum for a two-probe system? The MPSH? We diagonalize the the Hamiltonian of the central region? Since the Hamiltonian is no Hermitian for a two probe system, what will be the wave functions? The real part?

2

General Questions and Answers / Re: About transmission eigenstates

« on: May 14, 2016, 20:15 »

Many thanks  to your suggestions.

3

General Questions and Answers / Re: About transmission eigenstates

« on: May 13, 2016, 14:53 »

Many thanks for your reply. I still do not understand very well. Several questions:

1. Is the T matrix just the matrix Gamma_LG^RGamma_RG^A?

2. By "propagating the linear combination of the Bloch states" , do you mean  we take the linear combinations of the Bloch states of one lead as  an incoming wave and we calculate the reflected wave in the left lead and the transmitted wave in the right lead and the wave in the central region, by the same procedure as calculating the scattering states in Jeremy's thesis?

3.  How are the transmission eigenstates calculated in detail and do you have any reference for it, please?

4. Do the Bloch states here include all the propagating modes and evanescent modes which are obtained by solving a complex band structure problem? That is, are the incoming waves the linear combinations of all these modes? Do the transmission eigenstates in the left lead include the left moving modes?

4

General Questions and Answers / About transmission eigenstates

« on: May 13, 2016, 05:38 »

From here: http://www.quantumwise.com/documents/manuals/latest/ReferenceManual/index.html/ref.transmissioneigenstate.html

we are told that the transmission eigenstates are linear combinations of Bloch states of the electrodes.  However, I think this definition about transmission eigenstates there is wrong or at least misleading or confusing. Obviously, the Bloch states only belong to one electrode. With the Bloch state of only the left electrode, we can not get the spatial distribution of the transmission eigenstate in the central region and the right electrode since these Bloch states are not defined there.  Thus, I think, the transmission eigenstates are the linear combinations of the scattering states which are distributed in the whole space and span a Hilbert space. The scattering states  are defined by equation (4.1) in Jeremy Talor's PhD thesis.

Please correct me if I am wrong and please give a more detailed description or introduction about it. Many thanks.

5

General Questions and Answers / Re: For some understanding about the gate voltage

« on: September 12, 2013, 16:51 »

Then I have other two questions:

1. Can we add a dielectric region to the left and right electrodes? I mean that, can we add a continuum dielectric region to the whole system, including the electrodes and central region?

2. Shall we always add a metallic region on the top or bottom of a dielectric region? I am wondering about this since I see it from one example in the manual. Is the value of the hartree potential in the metallic region set to a constant, namely, the gate voltage, when solving the Poisson equation?

Many thanks again.

6

General Questions and Answers / Re: For some understanding about the gate voltage

« on: September 12, 2013, 14:18 »

There are several important differences. Indeed the old method only shifted H_ii, and all (selected) atoms were shifted by the same, fixed amount. In the new approach the electrostatic gates are included in the self-consistent procedure, so atoms will have different shifts depending on the positions and also the response of the electron cloud to the field is taken into account.

Many thanks for your reply, Anders. But how is the new gate voltage included in the self-consistent procedure? Do we solve the Poisson equation in another way, namely, by the multigrid method in the real space, and the gate voltages are taken as the boundary values of the Hartree potential? If this is the case, then we will have different dielectric constants in different parts of the central region and we may not use the FFT solver any more, right?

7

General Questions and Answers / For some understanding about the gate voltage

« on: September 12, 2013, 10:33 »

Hello, Dear All,

I am thinking about the gate voltage. Now we can add a metallic region or a dielectric region to the the central region. I want to know, what is the difference between this new method and the old method of shifting the onsite energy (like in version 08.10)? By the way, in the old method, do we just change the diangonal terms H_ii, but not the off-diagonal terms H_ij?

Thanks.

8

General Questions and Answers / about restart from a previous calculation

« on: September 26, 2010, 15:48 »

Dear all,

I have done a calculation with LDA. Now I want to use the converged electron density as an initial electron density for a GGA calculation.  Of course, I have to use restoreSelfConsistentCalculation("twoprobe-scf.nc") to restore the initial density.

My question is:  when I do the GGA calculation, will the lead part be calculated with GGA again or just use the LDA result? I mean, will the lead part be recalculated?

Thanks.

9

General Questions and Answers / Re: Is the convergence considered for the denxity matrix or the charge density?

« on: September 16, 2010, 02:31 »

Thanks!

I want to further ask about the density matrix. For a two-probe system with different electrodes, how is the total charge calculated? By the formula Q=Tr(DS) (where D is the density matrix and S is the basis overlap matrix), just like in Siesta? Or just by integrating the electron density in the real space?

10

General Questions and Answers / Is the convergence considered for the denxity matrix or the charge density?

« on: September 15, 2010, 16:04 »

In siesta, the convergence is considered for the density matrix. In ATK, is it considered for density matrix or the charge density at each point in the real space? Thanks a lot.

11

Scripts, Tutorials and Applications / Re: Function for calculating Projected Density of States

« on: July 6, 2010, 03:27 »

It is great!   Is there any news about the replacement for the version 2008.10, please?

12

Scripts, Tutorials and Applications / Re: Function for calculating Projected Density of States

« on: May 25, 2010, 14:12 »

It is true that for all results we should check  the internal consistency and physical soundness.  Obviously,  in this example of Li-H2-Li system, the two H atoms are exactly equivalent, so they should have the same PDOS. This is physical soundness. However, as I have mentioned above, the PDOS of these two atoms are totally different.  This is the problem. I am not arguing about the negative contributions.

Anyway, it is a very good news to hear that a replacement with the PDOS implemented will be released soon. Let's expect for it and thanks a lot.

13

Scripts, Tutorials and Applications / Re: Function for calculating Projected Density of States

« on: May 24, 2010, 03:15 »

Hello,

PDOS is a very important and very general analysis tool for ab initio calculations. It is very strange that ATK does not provide it.

The results from this script of Nordland are wrong.  However, many people might be using it for scientific research.  It is a serious problem. I hope the development team will work out a correct script to realize it ASAP.

14

Scripts, Tutorials and Applications / Re: Function for calculating Projected Density of States

« on: April 15, 2010, 03:22 »

Any furthr comments on this topic?

15

Scripts, Tutorials and Applications / Re: Function for calculating Projected Density of States

« on: April 7, 2010, 10:50 »

These are the input files I used. In fact, they were simply copied from the ATK "examples" directory.

Whether it can be negative or not is just one thing.  Another thing is that the two equivalent H atoms should have the same DOS. But actually completely not.