Messages

1

General Questions and Answers / Re: Questions about the total energy of a twoprobe system!

« on: February 4, 2011, 11:01 »

Yes, you are correct. Do note however that ATK 10.8 doesn't actually compute the total energy for devices (although the function name would indicate so); it returns the free energy. There is a correction term which plays a role in particular at finite bias. The total energy is not a well-defined quantity unless there is charge conservation, which is not the case for open systems (device configurations).

Anders

2

General Questions and Answers / Questions about the total energy of a twoprobe system!

« on: January 17, 2011, 16:03 »

Hello, evergyone
    For ATK-Version 2008.10 (DFT edition), the manual shows that the total energy of a twoprobe system can be calculated and it mainly include three parts: Electron kinetic energy, Electrostatic energy, and Exchange-corr. energy. However, for ATK 10.8 (DFT edition) I find that for a twoprobe system or device configuration the total energy of it is divided into four parts: Electron kinetic energy, Exchange-corr. energy, Hartree energy, and Interaction energy with the pseudo potential ions and other electro-static external potentials.
    I want to know whether Electrostatic energy of ATK-Version 2008.10 equals to the sum of Hartree energy and Interaction energy of ATK 10.8? Is that true even when bias voltage is applied?

3

General Questions and Answers / How to choose MultigridForElectrostatics method under homogeneous electrodes

« on: May 20, 2010, 09:14 »

Hi, everyone.
     I want to know how to choose MultigridForElectrostatics method to conduct calculations for a two-probe system with homogeneous electrodes. The following is my choice and is that right?
     Take Li-H2_Li as an example, I just use the same ElectrodeParameters but define the two parameters separately, that is,
==================================================================     
# Set k-points for electrodes
bz_int_param = brillouinZoneIntegrationParameters( (1,1,100) )

# Create parameters for electrodes
electrode_params_1 = ElectrodeParameters(
    brillouin_zone_integration_parameters = bz_int_param,
    iteration_control_parameters=iterationControlParameters(tolerance=1e-5,max_steps=300)
    )

electrode_params_2 = ElectrodeParameters(
    brillouin_zone_integration_parameters = bz_int_param,
    iteration_control_parameters=iterationControlParameters(tolerance=1e-5,max_steps=300)
    )   
......
method = TwoProbeMethod(
    (electrode_params_1,electrode_params_2),
    basis_set_parameters = basis_set_params,
    exchange_correlation_type = LDA.PZ,
    iteration_control_parameters = iteration_control_params,
    energy_contour_integral_parameters=energy_contour,
    algorithm_parameters=two_probe_params
    )
=====================================================================

Does it achieve the goal of using the MultigridForElectrostatics method?

4

General Questions and Answers / Re: How to choose the IterationControl criteria??

« on: January 8, 2010, 02:00 »

I'm afraid that I can't understand your explanation very well. So let us take an example.

The default criterion is TotalEnergy (Case 1.) and its corresponding tolerance is 1.0e-5. If we choose TotalEnergyAndDensityMatrix (Case 2.) or even Strict (Case 3.) as our criterion and the tolerance is still 1.0e-5. Of course, we test them in the same system. Could you tell me which one will give a more accurate result, Case 1., Case 2. or Case 3? My intuition tells me Case 3. does it and am I right?

In the mannual, the statement appears for the IterationControl criteria that "When performing calculations using a different criterion than the default one, an accuracy which is equivalent to the one provided by the default choice can be often obtained by increasing the tolerance of one (or even more, in some cases) order of magnitude."
Does that mean the accuracy provided by the tolerance 1.0e-7 or even lower for Case 1 could be equivalent to that provided by the tolerance 1.0e-5 for Case 2. or Case 3.?  

5

General Questions and Answers / How to choose the IterationControl criteria??

« on: January 7, 2010, 14:34 »

Hello,
         I am confused about the iterationControl criteria and want to know their differences.
         Especially I want to know which one to choose in order to execute an efficient and reliable self-consistent calculation. 

6

General Questions and Answers / Re: Pulay mixing inversion failed. Using only last step

« on: December 20, 2009, 13:34 »

Yes, I check the result for the calculation with "initial_scaled_spin" and it just gives the obtained total magnetic moment of the whole system as 5 hbar. But a self-consistent calculation with "initial_spin = 5 hbar"  still meets the above problem.

7

General Questions and Answers / Re: Pulay mixing inversion failed. Using only last step

« on: December 20, 2009, 08:07 »

Well done! I solve this problem just using the initial_scaled_spin, and the result gives the spin taken as 5 hbar. But I really don't know why the initial_spin=5 hbar meet the problem.

8

General Questions and Answers / Re: Pulay mixing inversion failed. Using only last step

« on: December 20, 2009, 02:43 »

hi, Nordland. I just caltulate a magnetic molecule that has four Iron atoms. the problem happens just when I set the initial_spin = 5*hbar. However, when the initial_spin =1*hbar, everything is OK. So I think the key lies in the set of the initial_spin for this system.

9

General Questions and Answers / Pulay mixing inversion failed. Using only last step

« on: December 19, 2009, 13:12 »

    hello, everyone. When I calculate a isolated system more than 200 atoms, a problem happens and the message is given that "Pulay mixing inversion failed. Using only last step".
    Besides, the runtimeParameters gives:
sc 1 : Fermi Energy = 0.00000 Ry  Ebs =nan Ry dRho =0.0000E+00  dEbs = NAN Ry  dH =  0.0000E+00 Ry
......
sc 14: Fermi Energy = 0.00000 Ry  Ebs =nan Ry dRho =0.0000E+00  dEbs = NAN Ry  dH =  0.0000E+00 Ry
......

    How to deal with this problem? Thank you for any relies.

10

General Questions and Answers / Re: About the energy resolution in Transmission Coefficients

« on: April 18, 2009, 10:52 »

Thank you for your help! 

11

General Questions and Answers / Re: About the energy resolution in Transmission Coefficients

« on: April 18, 2009, 02:20 »

Actually, for my two-probe system the feature of meV was found in experiment, and I obtained the same result in my calculation, but was not sure that my calculation is physical or not! And my green_function_infinitesimal was 1.0e-5*eV

12

General Questions and Answers / About the energy resolution in Transmission Coefficients

« on: April 17, 2009, 16:35 »

Hello, everyone. I want to ask some questions about the energy resolution in Transmission Coefficients. In my calculation, Transmission Coefficients were calculated from -3eV to 3eV with the interval of 10E-4eV. Here is my puzzle, is such little interval of 10E-4eV or the  results calculated with the interval reliable ?

13

General Questions and Answers / About the molecular unit cell!!

« on: March 4, 2009, 04:00 »

Hello everyone,
       I have some questions about the the molecular unit cell. The manual tells us that the molecular unit cell is automatically chosen such that there are no matrix elements between atoms in adjacent cells. Here I want to kown what "automatically" mean and whether it is related to the basis range of elements. If it is, how do you choose the distance to ensure no matrix elements between atoms in adjacent cells?!  And could you offer data about the basis range of elements?  Thank you very much.

14

General Questions and Answers / Re: About the parallel calculation using ATK

« on: January 13, 2009, 08:50 »

Yes, we have eight CPUs available in the server. And I also did a test, however, the parallel (two cpus)and the serial calculation for the LiH2Li twoprobe systerm both took five minutes.

15

General Questions and Answers / Re: About the parallel calculation using ATK

« on: January 13, 2009, 02:28 »

Yes, of course. Yesterday, I had a test to time a self-consistent calculation for the LiH2Li twoprobe systerm, and both the  parallel and the serial calculation took five minutes. Does that mean someting?