Print Page - electron-phonon interaction in two probe system

QuantumATK => General Questions and Answers => Topic started by: rebeda on June 8, 2015, 10:50

Title: electron-phonon interaction in two probe system
Post by: rebeda on June 8, 2015, 10:50

Can we calculate electron-phonon interaction for two probe system ?

Title: Re: electron-phonon interaction in two probe system
Post by: Umberto Martinez on June 8, 2015, 15:21

It will be introduce in the next release ATK 2015, coming out soon (next few months).

A preview of this can be found here:
http://quantumwise.com/publications/tutorials/item/837-inelastic-transmission
http://quantumwise.com/publications/tutorials/item/877-inelastic-electron-spectroscopy-of-h-molecule-between-1d-au-chains

Title: electron-phonon coupling analyser
Post by: rebeda on December 14, 2015, 11:58

while I was anlysing my result( *. nc file) on the labfloor for ElectronPhononCoupling calculation I got an error message that goes below. Will you please help me

"Exception message:
Traceback (most recent call last):
File ".\zipdir\NL\GUI\MainWindow\AnalysisBar\AnalysisBar.py", line 58, in widget
File "C:\Program Files (x86)\QuantumWise\VNL-ATK-2015.0\Lib\site-packages\AddOns\ElectronPhononCouplingPlugins\ElectronPhononCouplingPlugin.py", line 47, in createWidget
widget = ElectronPhononCouplingWidget(item, self.labFloor())
File "C:\Program Files (x86)\QuantumWise\VNL-ATK-2015.0\lib\site-packages\AddOns\ElectronPhononCouplingPlugins\ElectronPhononCouplingWidget.py", line 106, in __init__
self.update()
File "C:\Program Files (x86)\QuantumWise\VNL-ATK-2015.0\lib\site-packages\AddOns\ElectronPhononCouplingPlugins\ElectronPhononCouplingWidget.py", line 117, in update
self._plotter_widget.plotData(model)
File "C:\Program Files (x86)\QuantumWise\VNL-ATK-2015.0\lib\site-packages\AddOns\ElectronPhononCouplingPlugins\ElectronPhononCouplingWidget.py", line 395, in plotData
x_dir = 'xyz'[periodic_directions[0]]
IndexError: list index out of range"

Regards
Rebeda

Title: Re: electron-phonon interaction in two probe system
Post by: Umberto Martinez on December 14, 2015, 13:27

can you please provide input/output files?

Also note that 2015.1 has been released: http://quantumwise.com/products/download

Title: Re: electron-phonon interaction in two probe system
Post by: rebeda on December 15, 2015, 05:45

Python file for relaxation for gold nanowire ---

"# -------------------------------------------------------------
# Bulk configuration
# -------------------------------------------------------------

# Set up lattice
vector_a = [22.0, 0.0, 0.0]*Angstrom
vector_b = [ 0.0, 22.0, 0.0]*Angstrom
vector_c = [ 0.0, 0.0, 4.741513]*Angstrom
lattice = UnitCell(vector_a, vector_b, vector_c)

elements = 14*[Gold]

# Define coordinates
cartesian_coordinates = [[ 6.754023, 4.219484, 1.185378],
[ 5.375828, 1.823166, 1.185378],

------

[ 4.097240, 4.219487, 3.556135]]*Angstrom

# Set up configuration
bulk_configuration = BulkConfiguration(
bravais_lattice=lattice,
elements=elements,
cartesian_coordinates=cartesian_coordinates
)

###############################################################
# Calculator
###############################################################
#----------------------------------------
# Numerical Accuracy Settings
#----------------------------------------
numerical_accuracy_parameters = NumericalAccuracyParameters(
k_point_sampling=(1, 1, 800),
)
#----------------------------------------
# Basis Set
#----------------------------------------
basis_set = [
GGABasis.Gold_DoubleZetaPolarized,
]

#----------------------------------------
# Exchange-Correlation
#----------------------------------------
exchange_correlation = GGA.RPBE

#----------------------------------------
# Electrode Calculators
#----------------------------------------
calculator = LCAOCalculator(
basis_set=basis_set,
exchange_correlation=exchange_correlation,
numerical_accuracy_parameters=numerical_accuracy_parameters,
)

bulk_configuration.setCalculator(calculator)
bulk_configuration.update()
nlsave('60nw_bulk.nc', bulk_configuration)
nlprint(bulk_configuration)"

python file for electron-phonon-coupling calculation for the same system--

"bulk_configuration = nlread('60nw_bulk.nc', object_id='gID000')[0]

# -------------------------------------------------------------
# Calculator
# -------------------------------------------------------------

# -------------------------------------------------------------
# Hamiltonian derivatives
# -------------------------------------------------------------
hamiltonian_derivatives = HamiltonianDerivatives(
bulk_configuration,
repetitions=(1, 1, 1),
)

nlsave('60nw_bulk.nc', hamiltonian_derivatives)

# -------------------------------------------------------------
# Dynamical matrix
# -------------------------------------------------------------
dynamical_matrix = DynamicalMatrix(
bulk_configuration,
repetitions=(1, 1, 1),
atomic_displacement=0.01*Ang,
max_interaction_range=10*Angstrom,
)

nlsave('60nw_bulk.nc', dynamical_matrix)

# -------------------------------------------------------------
# Electron-phonon coupling
# -------------------------------------------------------------
electron_phonon_coupling = ElectronPhononCoupling(
configuration=bulk_configuration,
hamiltonian_derivatives=hamiltonian_derivatives,
dynamical_matrix=dynamical_matrix,
kpoints_fractional=[[0.0, 0.0, 0.0], [0.0, 0.0, 1.0]],
qpoints_fractional=[[0.0, 0.0, 0.0], [0.0, 0.0, 1.0]],
)

nlsave('60nw_bulk.nc', electron_phonon_coupling)

# -------------------------------------------------------------
# Deformation potential
# -------------------------------------------------------------
deformation_potential = DeformationPotential(
bulk_configuration,
hamiltonian_derivatives,
dynamical_matrix,
kpoint_fractional=[0.0, 0.0, 0.5],
q_path=[[0.0, 0.0, 0.0], [0.0, 0.0, 0.2], [0.0, 0.0, 0.3], \
[0.0, 0.0, 0.4], [0.0, 0.0, 0.45], [0.0, 0.0, 0.5]],
)

nlsave('60nw_bulk.nc', deformation_potential)

# -------------------------------------------------------------
# Mobility
# -------------------------------------------------------------
mobility = Mobility(
configuration=bulk_configuration,
electron_phonon_coupling=electron_phonon_coupling,
)

nlsave('60nw_bulk.nc', mobility)"

Title: Re: electron-phonon interaction in two probe system
Post by: Jess Wellendorff on December 21, 2015, 12:09

I ran your script as you have cited it above, and found no errors with ATK 2015.1

Title: Convergence issue with two-probe SCF run
Post by: rebeda on February 11, 2016, 06:21

Dear Sir,
While working on a two-probe system (spin-polarizd) with QuantumWise, I am struggling with the convergence
issue for the device, while the electrodes as well as the equivalent bulk behave as it should [see below].
Could you please help me out in resolving it? The problem mainly lies in converging dH, though there is no
such issue for dE. The k-points are chosen as "k_point_sampling=(3, 3, 100)", the contour parametrs are
chosen such that "circle_points=60", the iteration control parameters are chosen with "damping_factor=0.1".

-------------------
| Left Electrode Calculation [Started ------ ] |
Calculating Eigenvalues : ==================================================
| 24 E = -143.906 dE = 3.632179e-04 dH = 4.283244e-05 |
Calculating Eigenvalues : ==================================================
| 25 E = -143.906 dE = 3.514783e-05 dH = 1.171219e-05 |
| Left Electrode Calculation [Finished ----- ] |
--------------------
| Right Electrode Calculation [Started ----- ] |
--------------------
Calculating Eigenvalues : ==================================================
| 24 E = -143.906 dE = 3.680349e-04 dH = 3.973195e-05 |
Calculating Eigenvalues : ==================================================
| 25 E = -143.906 dE = 3.113314e-05 dH = 8.242738e-06 |
| Right Electrode Calculation [Finished ] |
--------------------
| Equivalent Bulk [Started ------ ] |
Calculating Eigenvalues : ==================================================
| 40 E = -604.546 dE = 1.674462e-02 dH = 7.109732e-03 |
Calculating Eigenvalues : ==================================================
| 41 E = -604.55 dE = 3.249493e-03 dH = 3.351605e-03 |
| Equivalent Bulk [Finished ----- ] |
--------------------
------
| 210 E = -631.902 dE = 6.540553e-05 dH = 1.848707e-03 |
| 211 E = -631.903 dE = 6.446038e-04 dH = 1.820017e-03 |
| 212 E = -631.907 dE = 3.662838e-03 dH = 1.753000e-03 |
| 213 E = -631.908 dE = 1.013363e-03 dH = 2.456666e-03 |
| 214 E = -631.906 dE = 1.372071e-03 dH = 2.714449e-03 |
| 215 E = -631.892 dE = 1.380006e-02 dH = 2.288700e-03 |
| 216 E = -631.893 dE = 6.937208e-04 dH = 5.274790e-03 |
| 217 E = -631.893 dE = 4.849841e-05 dH = 5.752056e-03 |
| 218 E = -631.894 dE = 1.074899e-03 dH = 5.493987e-03 |
| 219 E = -631.898 dE = 3.471558e-03 dH = 4.989637e-03 |
| 220 E = -631.9 dE = 2.690475e-03 dH = 3.936320e-03 |
| 221 E = -631.898 dE = 2.843382e-03 dH = 3.412088e-03 |
| 222 E = -631.897 dE = 1.992910e-04 dH = 4.081402e-03 |
| 223 E = -631.897 dE = 7.049964e-05 dH = 4.022176e-03 |
| 224 E = -631.898 dE = 6.790056e-04 dH = 3.998828e-03 |
| 225 E = -631.898 dE = 1.639834e-04 dH = 4.121747e-03 |
| 226 E = -631.897 dE = 6.953735e-04 dH = 4.528531e-03 |
| 227 E = -631.897 dE = 6.417012e-04 dH = 4.451000e-03 |
| 228 E = -631.897 dE = 2.927810e-04 dH = 3.333894e-03 |
| 229 E = -631.897 dE = 3.350161e-04 dH = 3.087998e-03 |
| 230 E = -631.896 dE = 1.572252e-03 dH = 3.273902e-03 |
| 231 E = -631.893 dE = 2.436914e-03 dH = 3.563415e-03 |
| 232 E = -631.894 dE = 6.613918e-04 dH = 4.290060e-03 |
| 233 E = -631.898 dE = 3.723214e-03 dH = 4.320211e-03 |
| 234 E = -631.898 dE = 4.412318e-05 dH = 3.644584e-03 |
| 235 E = -631.897 dE = 4.934228e-04 dH = 2.729455e-03 |
| 236 E = -631.899 dE = 2.232635e-03 dH = 2.961352e-03 |
| 237 E = -631.901 dE = 1.345609e-03 dH = 2.987998e-03 |
| 238 E = -631.901 dE = 6.869148e-04 dH = 2.617813e-03 |
| 239 E = -631.9 dE = 1.389278e-03 dH = 2.480923e-03 |
| 240 E = -631.9 dE = 4.429134e-04 dH = 2.187557e-03 |
| 241 E = -631.9 dE = 6.041444e-04 dH = 2.788400e-03 |
| 242 E = -631.899 dE = 4.840091e-04 dH = 2.678021e-03 |
| 243 E = -631.899 dE = 1.118098e-04 dH = 2.643998e-03 |
| 244 E = -631.899 dE = 2.690704e-04 dH = 2.394330e-03 |
| 245 E = -631.9 dE = 6.676876e-04 dH = 2.401902e-03 |
| 246 E = -631.899 dE = 1.190354e-03 dH = 2.648432e-03 |
| 247 E = -631.898 dE = 1.446554e-04 dH = 2.502319e-03 |
| 248 E = -631.899 dE = 2.109255e-04 dH = 2.519078e-03 |
| 249 E = -631.899 dE = 8.053678e-04 dH = 2.355521e-03 |
-------------------
Awaiting your advice and useful inputs, :(

Regards,
Rebeda

Title: Re: electron-phonon interaction in two probe system
Post by: Jess Wellendorff on February 11, 2016, 08:10

Could you attach the script you use?

Title: Re: electron-phonon interaction in two probe system
Post by: rebeda on February 11, 2016, 10:49

Dear Sir,
Please find the attachment. Kindly help me out in setting the correct parameters for the smooth SCF run.
Regards,
Rebeda

Title: Re: electron-phonon interaction in two probe system
Post by: rebeda on February 12, 2016, 02:11

Dear Sir,
Finally the device calculation ends up with no convergence even after 300 iterations as this:
+------------------------------------------------------------------------------+
| 299 E = -631.898 dE = 1.208148e-02 dH = 3.087065e-02 |
+------------------------------------------------------------------------------+
################################################################################
# #
# Warning: The calculation did not converge to the requested tolerance! #
# #
###############################################################################
--------------------------------------------------------------------------------
Total : 350125.23 s (4d01h15m25.23s)

What parameters I should tune to get better dH convergence.
(i) Should I modify the K-points, which I did set as (3, 3, 100)?
(ii) Should I increase the circle_points, which I did set as 60?
(iii) Should I further reduce the damping_factor, which I took as the default value of 0.1?
(iv) Should I set the initial scaled spins reduced from 1.0 to say, 0.6?
I did try some of these already, but to no avail, as I continue to struggle for about a month to get it converged. Hence, I am really looking forward to your advice and kind cooperation in resolving this convergence issue, which I am sure, would help many others like me.
With warm regards,
Rebeda

Title: Re: electron-phonon interaction in two probe system
Post by: Jess Wellendorff on February 12, 2016, 08:44

That script cannot be the one you are really using. Lines 28-30, 62-64, and 130-132 are not valid Python code ([ ... ] is invalid):

Code

left_electrode_coordinates = [[  0.936693371666,   0.758276943974,   0.78499999998 ],
                              [ ... ],
                              [ 10.806693371665,  10.628276943974,   5.01500000002 ]]*Angstrom

Please upload a valid script so I can do a test run.

Title: Re: electron-phonon interaction in two probe system
Post by: rebeda on February 16, 2016, 17:31

Dear Dr. Wellendorff,
I am curious to know if you have done the test run successfully by now. Awaiting then your reply with the tips and tricks that has been needed to be adopted here, in particular, for the device convergence.
Thanks and warm regards,
Rebeda

Title: Re: electron-phonon interaction in two probe system
Post by: rebeda on February 19, 2016, 09:07

Dear Sir,
Even after a month or so, I am still struggling with the convergence issue for the above system. Please help me with some ideas so that I can make it converged at last.
With thanks and regards,
Rebeda

Title: Re: electron-phonon interaction in two probe system
Post by: Anders Blom on February 19, 2016, 12:22

What did you try? You can increase the damping factor a bit, and reduce the history steps perhaps.
I would also skip the equivalent bulk, it takes a lot of time for a system like this, but might give a bad starting guess because there is a stacking fault in the central region (your two electrodes are not identical).

Title: Re: electron-phonon interaction in two probe system
Post by: Jess Wellendorff on February 22, 2016, 10:51

Hi Rebeda. I am trying to run the calculation with a few changes: Thicker electrodes, more aggressive Pulay mixing, and NeutralAtoms instead of EquivalentBulk. Hope it helps, will let you know...

Title: Re: mode assignment for negative bias for IETS of hydrogen molecule
Post by: rebeda on June 7, 2016, 14:38

In the tutorial given in the link below to calculate the IETS of hydrogen molecule-
http://quantumwise.com/publications/tutorials/item/877-inelastic-electron-spectroscopy-of-h-molecule-between-1d-au-chains
1. why vibrational modes were not assign for the negative bias?
2. I calculated the same for my system[Au(111)-single molecular magnet-Au(111)] within the bias -0.1 to 0.1 volt and I got asymmetric IETS. How can I find the modes for the negative bias?
In my vibrational analyser, mode zero lie at -53.99 meV which is -0.05399 volt, what about the IETS peaks within -0.1 to -0.05399 volt? How can we identify them?

Please help me. I will be waiting for your kind reply.
Regards,
Rebeda

Title: Re: Negative inelastic transmission
Post by: rebeda on June 9, 2016, 07:10

Dear Sir,

According to my knowledge transmission, being a probability, can have minimum value of 0 and maximum of 1 (though it depends on the no. of channels available for conduction) and hence, is always positive, but refering to one of your quantumwise tutorial-
link: http://quantumwise.com/support/tutorials/item/877-inelastic-electron-spectroscopy-of-h-molecule-between-1d-au-chains
I observe negative inelastic transmission at mode 0 and mode 5 in the IETS study of H molecule between 1D Au chains. I am not able to interpret this "negative inelastic transmission". In other words, how transmission can have negative values in this case, though for IETS, one can understand to have both positive and negative values.
Will you please help me?

Regards,
Rebeda :( :( :(

Title: thermoelectric transport properties in vnl window - reg.
Post by: rebeda on June 27, 2017, 07:18

Dear sir,
I have some confusion on the thermoelectric transport properties in vnl window. What does T really stand for???

1. does it stand for average temperature (an average of hot and cold electrodes temperature of the two probe device)?
2. does it stand for the environment temperature of the device?

looking forward to your reply.

regards,
Rebead

Title: Re: electron-phonon interaction in two probe system
Post by: Daniele Stradi on June 27, 2017, 09:02

Hello,

- the IETS should be symmetric, so the modes at negative bias are the same as those at positive bias. Notice that the IETS is the normalized second derivative of the transmission spectrum, hence it is not required to have values between 0 and 1.

- The fact that you get an asymmetric IETS could indicate that the system is not converged, however, I cannot say anything more, without looking at the actual scripts.

- I suggest you to read the papers about the underlying methodology to the IETS calculations, which will help you in understand how the spectra should be understood and analysed:
https://journals.aps.org/prb/abstract/10.1103/PhysRevB.75.205413

- Regarding the thermoelectric transport properties, if I understood your question correctly, T should be the temperature which enters the the formula of the thermoelectric figure of merit: http://docs.quantumwise.com/tutorials/phonon_calcs/phonon_calcs.html

Regards,
Daniele.

Title: thermoelectric transport properties in vnl window - reg.
Post by: rebeda on June 27, 2017, 10:28

Thank you Daniele Stradi for your quick response.

The T in the formula is introduced to make it a dimensionless value figure of merit "ZT". Usually thermoelectric behaviour is achieved when we make a temperature gradient/difference. In our case (two probe junction) the two electrodes are maintained one at cold (T_C) and the other at hot (T_H) temperature.

So, when we enter the temperature values to calculate the thermoelectic coefficients (in the vnl window) is -

1. T = T_H-T_C ?
or
2. T=(T_H-T_C)/2 ?

Lastly, in atk calculations what are the values of T_Cand T_H ?

Awaiting for your kind reply.

Regards,
Rebeda

Title: Re: electron-phonon interaction in two probe system
Post by: Petr Khomyakov on June 27, 2017, 10:42

The temperature in the formula for ZT is given as an arithmetical average, T=(T_H + T_C)/2.

Title: thermoelectric transport properties in vnl window - reg.
Post by: rebeda on June 27, 2017, 11:26

Thank you so much for your reply it cleared my doubt. :D :D :D

T=(T_H + T_C)/2.
In the above formula either T_C or T_H are kept constant in the entire atk calculations . If so, what is the constant temperature value?
I need these values to calculate the maximum effeciency.

Is T_C = 4.2K (liquid helium temperature) or T_C = 0K ?????

Awaiting for your useful comments.

Regards,
Rebeda

Title: Re: electron-phonon interaction in two probe system
Post by: Petr Khomyakov on June 27, 2017, 14:52

I am not sure I understand your problem. The T temperature is for ZT-calculation only; ZT is calculated in a post-processing manner after all the physical quantities in the ZT formula have been computed. What are the actual ATK calculations you are referring to?

Title: thermoelectric transport properties in vnl window - reg.
Post by: rebeda on June 28, 2017, 07:12

As I have mentioned before seebeck effect is realized when there is a temperature difference (in our two probe system, the electrodes are maintained one at cold T_Cand other at hot T_Htemperature ). And ZT is directly proportional to seebeck coefficient by formula.

So, My question is, when ATK perform thermoelectric coefficient calculations , what T_C /T_H value it takes ?

As an example, when we want to find ZT experimentally using STM, the STM tip and the sample are maintained at different temperatures, one at cold (T_C) and other at hot (T_H) temperature. In this case the tip is (T_C) and the sample temperature is (T_H).

Similarly, in atk theoretical calculations T_C and T_H must be assigned with some temperature values in which we keep on changing one of the temperature (say T_H ) to calculate thermoelectric coefficients at variable temperature.

I hope you might understand my question. I have added some link of articles to help undestand the problem.

http://www.pnas.org/content/112/27/8205.full.pdf
http://dx.doi.org.sci-hub.ac/10.1063/1.4958999

regards,
Rebeda

Title: Re: electron-phonon interaction in two probe system
Post by: Petr Khomyakov on June 28, 2017, 10:09

You may see one example of how the thermoelectrics calculations can be done in ATK in the following tutorials at http://docs.quantumwise.com/tutorials/thermoelectrics_cnt_isotope/thermoelectrics_cnt_isotope.html and http://docs.quantumwise.com/tutorials/phonon_calcs/phonon_calcs.html.

The calculations are done in the linear response regime, and no T_c and T_h are explicitly introduced in the ATK calculations. To compute the thermoelectric coefficients and ZT in a post-processing manner, one can use a thermoelectrics plugin as discussed in the tutorials, and the temperature T=(T_c+T_h)/2 is to be manually set in the widget.

In ATK, you may also calculate interfacial thermal conductance, using a molecular dynamics (MD) approach as described in http://docs.quantumwise.com/tutorials/interfacial_thermal_conductance/interfacial_thermal_conductance.html. In this case, the temperature gradient across the interface is explicitly taken into account in the non-equlibrium MD simulations.

Title: Re: Problem arising with the arrival of new atk-vnl versions - reg.
Post by: rebeda on July 12, 2017, 13:53

Dear Sir,

It is great that atk-vnl is upgraded everytime with the upcoming of every new versions. But one big problem is that, we are not able to access the new versions with the previous xxx.nc files calculated from lower versions. Some of the device calculations are time consuming and we rather not repeat the same calculations in the new versions. Due to it we are not able to utilize some of the new upgraded features which may help in better analysis of our calculated datas. Would you suggest a way to solve this issue??????

Waiting for your response.

Regards,
Rebeda

Title: Re: Problem arising with the arrival of new atk-vnl versions - reg.
Post by: Anders Blom on July 12, 2017, 19:42

Quote from: rebeda on July 12, 2017, 13:53

we are not able to access the new versions with the previous xxx.nc files calculated from lower versions

We make a lot of effort to ensure new versions can read old NC files. I you have any other experience, please contact support with the specifics, and we will help you out.

QuantumATK Forum

QuantumATK => General Questions and Answers => Topic started by: rebeda on June 8, 2015, 10:50