I-V curve dependence of device length along transport direction

Asked by Pengfei Ou on 2017-11-07

Hi,

I am using the Transiesta and TbTrans to calculate the I-V curve under the finite voltage (0-2 eV), normally I run the Transiesta calculations first and followed by the TbTrans code for each applied voltage. But the results show that if I change the length of the devices along the transport directions, it will significantly affect the values of current (36.5 Angstrom of the length corresponds to 1.05*10^-10 A, 54.7 Angstrom to 1.46*10^-6 A). I am wondering why there will be such a huge variations? Is it normal?

Below is my input files for Electrodes and Transport calculations:

Thanks for your kindness.

Elec:

SystemName elec
SystemLabel elec

# ---------------------------------------------------------------------------
# Lattice
# ---------------------------------------------------------------------------

LatticeConstant 1.00 Ang

%block LatticeVectors
    24.323716 -0.131002 -0.182012
     0.000000 9.934577 0.000000
     0.000000 0.000468 9.082148
%endblock LatticeVectors

# ---------------------------------------------------------------------------
# Species and Atoms
# ---------------------------------------------------------------------------

NumberOfSpecies 1
NumberOfAtoms 24

%block ChemicalSpeciesLabel
  1 15 P
%endblock ChemicalSpeciesLabel

%block PAO.BasisSizes
 P DZP
%endblock PAO.BasisSizes

# ---------------------------------------------------------------------------
# Exchange-Correlation Functionals
# ---------------------------------------------------------------------------
XC.functional GGA
XC.authors PBE

# ---------------------------------------------------------------------------
# Atomic Coordinates
# ---------------------------------------------------------------------------

AtomicCoordinatesFormat Ang

%block AtomicCoordinatesAndAtomicSpecies
        9.0008696 4.0865698 0.2964436 1
        8.9926968 0.773647 0.299635 1
        8.993913 7.3984385 0.3001742 1
        8.9983399 2.4317411 1.7799078 1
        8.9992399 5.7410405 1.779924 1
        8.9992885 9.0527092 1.7837393 1
        11.1027792 5.7419701 2.5831458 1
        11.1019765 2.4297956 2.5833803 1
        11.1028765 9.0531747 2.5841275 1
        11.1006874 4.0860958 4.0660813 1
        11.1009549 7.3975081 4.0664905 1
        11.0987658 0.7745301 4.0665069 1
        8.995494 7.39688 4.8638732 1
        8.9935238 0.7747817 4.8648705 1
        8.9952995 4.08658 4.86504 1
        8.991432 9.0531023 6.3457953 1
        8.9935968 5.7423956 6.3477897 1
        8.9925022 2.4309645 6.3478893 1
        11.0942416 9.0524364 7.1508733 1
        11.0954578 2.4311091 7.1521208 1
        11.0961388 5.7418007 7.1529154 1
        11.0932443 0.7743777 8.633275 1
        11.0926119 7.3980433 8.6344908 1
        11.0950443 4.0861296 8.6353408 1
%endblock AtomicCoordinatesAndAtomicSpecies

# ---------------------------------------------------------------------------
# Calculation Details
# ---------------------------------------------------------------------------
%block kgrid_Monkhorst_Pack
   1 0 0 0.0
   0 3 0 0.0
   0 0 60 0.0
%endblock kgrid_Monkhorst_Pack

MeshCutoff 150.0 Ry
MaxSCFIterations 500
DM.MixingWeight 0.2
DM.NumberPulay 5
DM.Tolerance 1.d-4
WriteCoorXmol .true.
Diag.DivideAndConquer False

ElectronicTemperature 300 K
OccupationFunction FD

SolutionMethod diagon
TS.WriteHS .true.

# ---------------------------------------------------------------------------
# Parallelized Running
# ---------------------------------------------------------------------------
BLOCKSIZE 8
Diag.Memory 1.3

Transiesta and TbTrans:

SystemName scat
SystemLabel scat

# ---------------------------------------------------------------------------
# Lattice
# ---------------------------------------------------------------------------

LatticeConstant 1.00 Ang

%block LatticeVectors
    24.323716 -0.131002 -0.182012
     0.000000 9.932301 0.000000
     0.000000 0.001880 36.523614
%endblock LatticeVectors

# ---------------------------------------------------------------------------
# Species and Atoms
# ---------------------------------------------------------------------------

NumberOfSpecies 3
NumberOfAtoms 116

%block ChemicalSpeciesLabel
  1 1 H
  2 6 C
  3 15 P
%endblock ChemicalSpeciesLabel

%block PAO.BasisSizes
 H DZP
 C DZP
 P DZP
%endblock PAO.BasisSizes

# ---------------------------------------------------------------------------
# Exchange-Correlation Functionals
# ---------------------------------------------------------------------------
XC.functional GGA
XC.authors PBE

# ---------------------------------------------------------------------------
# Atomic Coordinates
# ---------------------------------------------------------------------------

AtomicCoordinatesFormat Ang

%block AtomicCoordinatesAndAtomicSpecies
        9.0008696 4.071713 0.2883808 3
        8.9926968 0.7587903 0.2915722 3
        8.993913 7.3835817 0.2921115 3
        8.9983399 2.4168843 1.7718451 3
        8.9992399 5.7261837 1.7718612 3
        8.9992885 9.0378524 1.7756766 3
        11.1027792 5.7271133 2.575083 3
        11.1019765 2.4149389 2.5753175 3
        11.1028765 9.038318 2.5760648 3
        11.1006874 4.0712391 4.0580185 3
        11.1009549 7.3826514 4.0584278 3
        11.0987658 0.7596734 4.0584442 3
        8.995494 7.3820232 4.8558105 3
        8.9935238 0.759925 4.8568077 3
        8.9952995 4.0717233 4.8569772 3
        8.991432 9.0382455 6.3377325 3
        8.9935968 5.7275388 6.339727 3
        8.9925022 2.4161078 6.3398266 3
        11.0942416 9.0375796 7.1428105 3
        11.0954578 2.4162524 7.1440581 3
        11.0961388 5.7269439 7.1448527 3
        11.0932443 0.759521 8.6252123 3
        11.0926119 7.3831865 8.626428 3
        11.0950443 4.0712729 8.6272781 3
        8.9894861 0.7595561 9.4271291 3
        8.9911158 4.0720739 9.4286478 3
        8.988951 7.3824562 9.4295551 3
        8.9920401 2.4154183 10.90955 3
        8.9933779 5.7274509 10.9117563 3
        8.991359 9.0382728 10.9120913 3
        11.0967956 9.0373606 11.711209 3
        11.0958226 2.4177143 11.7130671 3
        11.0975982 5.7250154 11.7137849 3
        11.0945091 0.7617199 13.1949685 3
        11.0961145 7.3818961 13.1954363 3
        11.0873093 4.0709472 13.198701 3
        15.4151552 4.0590903 13.621666 1
        8.9882699 7.3838536 13.9872161 3
        8.9869078 0.7590398 13.9879803 3
        8.9780783 4.0722862 13.9900114 3
        14.0924072 3.1594682 14.406954 1
        14.0758914 4.9326082 14.4082201 1
        14.7337019 4.0520717 14.4854613 2
        8.9852294 9.0385869 15.47057 3
        8.976108 5.7294245 15.4711409 3
        8.9750135 2.4148872 15.4712176 3
        15.4950829 4.0586567 15.8121801 2
        16.1467882 4.9475171 15.8667783 1
        16.1623068 3.1814065 15.8671649 1
        11.0876985 9.035767 16.2770215 3
        11.0765582 2.4218084 16.2775161 3
        11.0775555 5.7222289 16.2780113 3
        13.8864583 4.9274079 16.9568452 1
        13.9027552 3.1624059 16.959328 1
        14.558182 4.051061 17.0185595 2
        11.0870904 0.7642434 17.7584835 3
        11.0862877 7.3793559 17.7594263 3
        11.0570263 4.0718092 17.7683506 3
        15.2716209 4.0599686 18.3688512 2
        15.9210155 4.9512297 18.4349825 1
        15.9414961 3.1839796 18.4354233 1
        8.9526357 4.0716156 18.5640802 3
        8.9844268 0.7588948 18.5649619 3
        8.9832835 7.3832817 18.565793 3
        13.6479156 3.1619836 19.470194 1
        13.6285539 4.9209983 19.4710015 1
        14.300983 4.0487283 19.5495728 2
        8.973627 2.4147825 20.0473228 3
        8.9881483 9.0401282 20.0479681 3
        8.9732865 5.7281454 20.048582 3
        11.0834418 5.7280652 20.8401244 3
        11.0833932 2.4155389 20.8403761 3
        11.0923443 9.0389446 20.8486945 3
        14.9972981 4.0564537 20.9090864 2
        15.6477386 3.1750778 21.0236488 1
        15.6266256 4.9530179 21.0240353 1
        14.2696784 4.0478968 21.7360296 1
        11.1016603 4.0720815 22.3228393 3
        11.0934146 7.3803409 22.3275621 3
        11.0930011 0.7641455 22.3285934 3
        9.0008696 4.0728894 23.1369808 3
        8.9926968 0.7599667 23.1401722 3
        8.993913 7.3847581 23.1407115 3
        8.9983399 2.4180607 24.6204451 3
        8.9992399 5.7273601 24.6204612 3
        8.9992885 9.0390288 24.6242765 3
        11.1027792 5.7282897 25.423683 3
        11.1019765 2.4161153 25.4239175 3
        11.1028765 9.0394944 25.4246648 3
        11.1006874 4.0724155 26.9066185 3
        11.1009549 7.3838278 26.9070278 3
        11.0987658 0.7608498 26.9070441 3
        8.995494 7.3831996 27.7044105 3
        8.9935238 0.7611014 27.7054077 3
        8.9952995 4.0728997 27.7055772 3
        8.991432 9.039422 29.1863325 3
        8.9935968 5.7287152 29.188327 3
        8.9925022 2.4172842 29.1884266 3
        11.0942416 9.038756 29.9914105 3
        11.0954578 2.4174288 29.9926581 3
        11.0961388 5.7281203 29.9934527 3
        11.0932443 0.7606974 31.4738123 3
        11.0926119 7.3843629 31.475028 3
        11.0950443 4.0724493 31.475878 3
        8.9894861 0.7607325 32.2757291 3
        8.9911158 4.0732503 32.2772478 3
        8.988951 7.3836326 32.2781551 3
        8.9920401 2.4165947 33.75815 3
        8.9933779 5.7286273 33.7603563 3
        8.991359 9.0394492 33.7606913 3
        11.0967956 9.0385371 34.559809 3
        11.0958226 2.4188907 34.561667 3
        11.0975982 5.7261918 34.5623849 3
        11.0945091 0.7628963 36.0435685 3
        11.0961145 7.3830725 36.0440363 3
        11.0873093 4.0721236 36.047301 3
%endblock AtomicCoordinatesAndAtomicSpecies

# ---------------------------------------------------------------------------
# Calculation Details
# ---------------------------------------------------------------------------
%block kgrid_Monkhorst_Pack
   1 0 0 0.0
   0 3 0 0.0
   0 0 1 0.0
%endblock kgrid_Monkhorst_Pack

MeshCutoff 150.0 Ry
MaxSCFIterations 500
DM.MixingWeight 0.2
DM.NumberPulay 5
DM.Tolerance 1.d-4
WriteCoorXmol .true.
Diag.DivideAndConquer False
DM.UseSaveDM .true.

ElectronicTemperature 300 K
OccupationFunction FD

SolutionMethod transiesta
TS.WriteHS .true.

# Left Electrode
TS.HSFileLeft ../Elec/elec.TSHS
TS.ReplicateA1Left 1
TS.ReplicateA2Left 1
TS.NumUsedAtomsLeft 24
TS.BufferAtomsLeft 0

# Right Electrode
TS.HSFileRight ../Elec/elec.TSHS
TS.ReplicateA1Right 1
TS.ReplicateA2Right 1
TS.NumUsedAtomsRight 24
TS.BufferAtomsRight 0

%block TBT_kgrid_Monkhorst_Pack
   1 0 0 0.0
   0 10 0 0.0
   0 0 1 0.0
%endblock TBT_kgrid_Monkhorst_Pack

TS.Voltage 2.0 eV
TS.BiasContour.Eta 0.001 Ry
TS.BiasContour.NumPoints 100
TS.TBT.HSFile './scat.TSHS'
TS.TBT.Emin -2.0 eV
TS.TBT.Emax 2.0 eV
TS.TBT.NPoints 101

# ---------------------------------------------------------------------------
# Parallelized Running
# ---------------------------------------------------------------------------
BLOCKSIZE 8
Diag.Memory 1.3

Question information

Language:
English Edit question
Status:
Solved
For:
Siesta Edit question
Assignee:
No assignee Edit question
Solved by:
Nick Papior
Solved:
2017-12-03
Last query:
2017-12-03
Last reply:
2017-11-08
Nick Papior (nickpapior) said : #1

It may very well happen for your system.

But 1) you haven't clarified how you elongated your system (if you only increase the electrode parts then it means that your screening region is too short), and
2) consider how the device region influences the scattering potential and whether it makes sense.

Pengfei Ou (barneycavs) said : #2

Hi Nick,

I would like to check a few things with you, because I am not sure about the definitions and settings even though I read through the manual,

1) In my calculations, c-axis is for transport direction, a-axis is for vacuum, so the b-axis is the semi-infinite direction according to the definition, right? So, I should set a large value of k-points along the b-axis, let's say 50.

2) Do I need to set a large k-points for c-axis along the transport calculations for the electrode calculations? I know the Transiesta will automatically reduce it to 1.

3) I only increase the length of scattering region in my calculations, I keep the electrodes the same length. Could it be possible that it is due to the low k-points I set for the semi-infinite direction (b-axis, only 3 herein).

Thanks a lot.

Nick Papior (nickpapior) said : #3

1) I think you have written something wrong, please carefully examine your system and figure out how the semi-infinite and transport direction are related.
The electrode k-points should be very high in the semi-infinite direction (say > 50). So on that part you are correct.

2) Please search the siesta mailing list for this answer, this has been asked multiple times. Please consider what it means to attach semi-infinite electrodes to a system. This should provide you with the answer.

3) You still haven't clarified what you increase, a) do you increase number of atoms?, if so are the added atoms additional electrode layers or additional interface atoms, b) do you increase vacuum along semi-infinite direction, c) or both.

Pengfei Ou (barneycavs) said : #4

3) I increase the number of atoms for the substrate in the scattering region (additional interface atoms), with the all the other setting the same. (Same electrodes, same vacuum, same adsorbed molecule position).

Best Nick Papior (nickpapior) said : #5

When you say substrate is it the same as the electrode region? If this changes the bias then most probably you haven't reached a converged, so if your system is this:
Elec-left - Elec-left - interface - molecule - interface - Elec-right - Elec-right
Elec-left - Elec-left - Elec-left - interface - molecule - interface - Elec-right - Elec-right - Elec-right

If this is the way you changed the length and it changes the results, then definitely your potential is not bulk like in your electrode regions and then you really need to converge the number of electrode layers.

Pengfei Ou (barneycavs) said : #6

Thanks Nick Papior, that solved my question.

Jianjun Mao (jjmao) said : #7

Dear Nick,

For this question, I have some doubts that....if the scattering region and the electrode are the same-graphene, except that the scattering region with a defect, when we change the width of the scattering region......the density of the defect is changed so the resistance should change. I think the IV curve should change at the same time.

Looking forward to your answer.

Nick Papior (nickpapior) said : #8

There is no discussion of width in this question, only length.

Jianjun Mao (jjmao) said : #9

Dear Nick,

OK, but if we change the length, I think it is similar with the width....I think the IV curve should change due to the change of the density of the defect.

Nick Papior (nickpapior) said : #10

No, that is not the case. The density of the defect does not change if you simply increase the electrode regions (i.e. length).

If this is not immediately clear to you, I would suggest to read up on the NEGF scheme, and importantly what an electrode is in such simulations.

Jianjun Mao (jjmao) said : #11

Yeah, I am a little confused with the length of electrode part of the scattering regions....OK, thank you Nick~~