Messages

1

General Questions and Answers / Re: Monte Carlo simulation for calculating the exchange couplings

« on: August 3, 2020, 11:55 »

Hello,

there is no Monte-Carlo implementation in QuantumATK to calculate the exchange coupling constants, becasuse QuantumATK implements an efficient Green's function-based method to obtain  the exchange couplings at all orders (nearest neighbor, next-nearest neighbor, etc..). The method is implemented in the Heisenberg Exchange class:

https://docs.quantumatk.com/manual/Types/HeisenbergExchange/HeisenbergExchange.html

Best Regards,
Daniele 

2

General Questions and Answers / Re: Error during phonon transmission calculation

« on: July 16, 2020, 11:59 »

Hello,

Could you please attach the script that you are using ?

Best regards,
Daniele

3

General Questions and Answers / Re: momentum resolved transmission spectrum

« on: July 6, 2020, 08:41 »

Hello,

>>> For instance, if it shows that at a certain energy, certain k points show high transmission, what does that physically mean?

For a periodic material, it means that for that specific value of energy and parallel momentum, the probability of ballistic electron/hole transport through the material will be high. For a device configuration, it means that the probability of source-to-drain ballistic electron transport will be high.

Best regards,
Daniele

4

General Questions and Answers / Re: Getting density matrix, self energies etc.

« on: July 6, 2020, 08:31 »

Hello,

you can access the Hamiltonian and Overlap matrices by using low-level entities, please refer to the following documentation page:

https://docs.quantumatk.com/tutorials/low_level_entities/low_level_entities.html

Best regards,
Daniele

5

General Questions and Answers / Re: Change in space group

« on: July 3, 2020, 17:49 »

Hello,

the Schoenflies notation specifies the crystallographic point group of the system, which depends on the atomic positions in the primitive cell. If you remove part of the atoms, the crystallographic point group will naturally change, because some symmetry operations will be removed:

- p63/mmc: http://img.chem.ucl.ac.uk/sgp/large/194az1.htm
- p6m2: http://img.chem.ucl.ac.uk/sgp/large/187az1.htm

So, the only thing one can do to retain the p63/mmc space group is to keep the extra layer.

Best,
Daniele

6

General Questions and Answers / Re: Projected local density of states

« on: May 19, 2020, 09:23 »

Hi,

the heat maps can be generated by using the projected local density of states (PLDOS) analysis tool:
https://docs.quantumatk.com/manual/Types/ProjectedLocalDensityOfStates/ProjectedLocalDensityOfStates.html

The 2D plots can be generated by using the DeviceDensityOfStates (DDOS) object and by projecting on the atoms of the region of interest, in your specific case the semiconductor region:
https://docs.quantumatk.com/manual/Types/DeviceDensityOfStates/DeviceDensityOfStates.html 

Best regards,
Daniele

7

General Questions and Answers / Re: Projected local density of states

« on: May 13, 2020, 10:12 »

Hello,

Projected Local Density Of States (PLDOS) cannot be used for that, however, one can use the DeviceDensityOfStates to project over, e.g., atoms. In the example below, the DDOS is evaluated for atom 10 in the central region of the device:

ddos = nlread('device.hdf5',DeviceDensityOfStates)[-1]
ddos_projected = ddos.evaluate(projection_list = ProjectionList([10]))

Check also the usage examples in: https://docs.quantumatk.com/manual/Types/DeviceDensityOfStates/DeviceDensityOfStates.html

Best,
Daniele

8

General Questions and Answers / Re: Modelling charged molecules in molecular junctions

« on: April 27, 2020, 11:56 »

Unfortunately, you cannot specify the charge of the molecule in a device calculation. This is because in a DFT-NEGF calculation the charge state of the molecule will be a consequence of the electronic density of the central region reaching equilibrium with the Au electrodes.

There is no unique way of fixing the charge state of the molecule in a DFT-NEGF calculation. If you want to calculate the transport in a charged molecule, that will be in the Coulomb blockade regime, and therefore I would suggest to simulate as documented in the paper linked here below:

https://pubs.acs.org/doi/abs/10.1021/jp104811r

Best,
Daniele

9

General Questions and Answers / Re: Bandstructure of 5 atom Germanene AC Nanoribbon

« on: April 27, 2020, 09:15 »

Hi,

the Extended Huckel method might indeed give incorrect results if the target structure is very different from those used for the original derivation of the Huckel parameters. I would suggest instead to use DFT, either by using the LCAOCalculator or the PlaneWaveCalculator. DFT is more transferable and does not suffer the transferability problem of Extended Huckel, and as such it will give more physically correct trends.

Also, if you target is individual nanoribbons, I would suggest to modify the structure as follow:
1) Use periodic boundary conditions for the two unit cell planes normal to the B direction and increase the length of the B vector by a factor 2-3,  to ensure that the  periodic replicas of the ribbons are effectively decoupled. 
2) Make sure that the length of the A vector is at least 20 Angstrom, and use mixed (Dirichlet+Neumann) boundary conditions for the two unit cell planes normal to the A direction.

Best,
Daniele

10

Scripts, Tutorials and Applications / Re: AttributeError: 'Bandstructure' object has no attribute 'indirectBandGap'

« on: April 20, 2020, 10:59 »

Hello,

the program fails because the function indirectBandGap() is a not a public function in 2018.06-SP1.

In order to make the script work, you should change:
bandstructure.indirectBandGap().inUnitsOf(eV))
to:
bandstructure._indirectBandGap().inUnitsOf(eV))

We have filed an bug for this.

Best regards,
Daniele

11

General Questions and Answers / Re: Negative Mulliken Population

« on: April 6, 2020, 07:55 »

Hi Nayab,

Mulliken population is not supposed to be negative, but I have seen this happening in few circumstances in which the population was supposed to be zero, such as when using ghost atoms. Could you attach a sample of your input file that shows the problem ?

Best
Daniele

12

General Questions and Answers / Re: Partial and projected DOS?

« on: April 6, 2020, 07:42 »

Dear Nayab,

there is no tool in QuantumATK producing a 'partial DOS'. Are you referring to the option of calculating projections in the DensityOfStates analysis object, or to something else ? In that case, one can also project on atoms, etc by specifying a projection_list when evaluating the DOS.

https://docs.quantumatk.com/manual/Types/DensityOfStates/DensityOfStates.html

Best,
Daniele

13

General Questions and Answers / Re: Spectral function

« on: August 12, 2019, 09:05 »

Hello,

You can create plots such as that in Fig. 3b by using the SurfaceBandstructure object: https://docs.quantumatk.com/manual/Types/SurfaceBandstructure/SurfaceBandstructure.html

To create plots such as  Fig. 5a, you can use the ProjectedLocalDensityOfStates object: https://docs.quantumatk.com/manual/Types/ProjectedLocalDensityOfStates/ProjectedLocalDensityOfStates.html

Best regards,
Daniele

14

General Questions and Answers / Re: Noncollinear Mulliken Population

« on: July 31, 2019, 09:52 »

The magnetic moment is expressed in units of Bohr magneton.

15

General Questions and Answers / Re: Noncollinear Mulliken Population

« on: July 30, 2019, 08:58 »

In QuantumATK-P-2019.03, the total spin magnetic moment is actually printed out in the MullikenPopulation object.

If you are using an older version of QuantumATK, you should calculated youself the difference between the spin-up and the spin-down components of the Mulliken population. Notice however that in a non-collinear framework, this difference may not correspond to the total magnetic moment.

Best,
Daniele