Bandstructure with external Electric field

[santanumahapatra]

Hello,
I want to observe the effect of Electric field on the band structure of hexagonal MoTe2, however, the calculation does not converge. Below are my inputs. Kindly suggest.

POSCAR:

Code: Select all

Mo   Te
1.0
   7.0947799683000001   0.0000000000000000   0.0000000000000000
  -1.7736954010999999   3.0721296071999999   0.0000000000000000
   0.0000000000000000   0.0000000000000000  18.6102008820000009
Mo   Te
2  4  
Direct
  0.0000000000000000  1.0000000000000000  0.5000000000000000  
  0.5000000000000000  1.0000000000000000  0.5000000000000000  
  0.3333300000000000  0.3333300000000000  0.5969900000000000  
  0.3333300000000000  0.3333300000000000  0.4030100000000000  
  0.8333300000000000  0.3333300000000000  0.5969900000000000  
  0.8333300000000000  0.3333300000000000  0.4030100000000000

INCAR:

Code: Select all

SYSTEM  = 2H-MoS2
ISTART=0
ENCUT   = 600
ALGO    = Fast
IBRION  = -1
NELM=500
ISIF=2

EDIFF   = 1.0e-06
ISYM    = 0
LREAL   = .FALSE.

ISMEAR  = 0
SIGMA   = 0.05

PREC    = Accurate
ADDGRID = .TRUE.

NWRITE  = 1
LCHARG  = .FALSE.
LWAVE   = .FALSE.
LDIPOL = .TRUE.
IDIPOL = 3   # for Z direction
EFIELD = 1.5   # eV/A**
DIPOL = 0.0 0.0 0.5 
#LNEBCELL = .TRUE

KPOINT:

Code: Select all

MoS2 hexagonal lattice
16
Line-mode
reciprocal
   0.000   0.000   0.000   ! \Gamma
   0.500   0.000   0.000   ! M
 
   0.500   0.000   0.000   ! M
   0.3333  0.3333  0.000   ! K
 
   0.3333  0.3333  0.000   ! K
   0.000   0.000   0.000   ! \Gamma

[alexey.tal]

Dear santanumahapatra,

Thank you for your question.

You should start the band structure calculation with an even k-points mesh before using the linear mesh as outlined in this tutorial.
Also, I see that you are applying a strong field, so it can be challenging to achieve the convergence.
You can try the following:

• switch to the direct optimization ALGO=all
• use linear mixing by setting BMIX=0.0001 and BMIX_MAG=0.0001
• mix slowly, i.e., reduce AMIX and AMIX_MAG

Best wishes,
Alexey

[santanumahapatra]

Thank you for your response.
The enrgy doesn't change however. I varied the EFIELD from 0 to 1.2, but energy remained exactly same. I tried both single point run and full relaxation, but made no difference.
I am using the same INCAR and POSCAR I posted, with changes in NSW accordingly, along with 8*8*1 K point.

KIndly suggest.
Regards.

[alexey.tal]

I'm not sure I understood what you are comparing. It was a bandstructure calculation, right? Could you please provide the OUTCAR files from these calculations where the energy doesn't change?

[santanumahapatra]

I am trying to see if the external electric field has any effect on the POSCAR. For that, I am just doing a SCF calculation with the Efield applied. I have attached 3 OUTCAR files for 0, 0.5V, 1V SCF calculations, the final energies have no difference, which should not be the case.
Below is my POSCAR:

Code: Select all

Mo   Te                                 
    1.000000000000000     
     5.9489048004200002    0.0000000000000000    0.0000000000000000
    -1.1469479507800000    3.9939477239199999    0.0000000000000000
     0.1669877038400000   -0.8921159885600000   18.6027577557199990
  Mo              Te            
     2     4
Direct
 -0.0545595359501158  0.4538629091468335  0.4944781663392859
  0.6488288574526421  0.0452149723303796  0.5017130973036489
  0.8397587526430286  0.3496325971216483  0.6352258398522228
  0.8092479423457553  0.7507896837361355  0.3814740740627136
  0.3839677951204261  0.4083496116320358  0.5812858562671203
  0.3038645627882685  0.2295294524329695  0.4058233785750071
 
  0.00000000E+00  0.00000000E+00  0.00000000E+00
  0.00000000E+00  0.00000000E+00  0.00000000E+00
  0.00000000E+00  0.00000000E+00  0.00000000E+00
  0.00000000E+00  0.00000000E+00  0.00000000E+00
  0.00000000E+00  0.00000000E+00  0.00000000E+00
  0.00000000E+00  0.00000000E+00  0.00000000E+00

INCAR:

Code: Select all

SYSTEM  = 2H-MoS2
ISTART=0
ENCUT   = 600
ALGO    = Normal
#IBRION  = 3
IBRION=-1
NELM=500
ISIF=2
#NSW=200
NSW=0

EDIFF   = 1.0e-06
ISYM    = 0
LREAL   = .FALSE.

ISMEAR  = 0
SIGMA   = 0.05

PREC    = Accurate
ADDGRID = .TRUE.

NWRITE  = 1
LCHARG  = .TRUE.
LWAVE   = .FALSE.
LDIPOL = .TRUE.
IDIPOL = 3   # for Z direction
EFIELD = 1.0   # eV/A**
DIPOL = 0.0 0.0 0.5 
#LNEBCELL = .TRUE

KPOINT:

Code: Select all

Automatic Mesh # Generates Automatically the K-mesh
  0
Monkhorst-pack # Automatic Distribution of the mesh
  8  8  1

[alexey.tal]

I did a quick calculation with the POSCAR you provided in your first post and I see a difference in the total energies.

The structure you used in your last post does not seem to be correct. Also, I see that it is metallic in your OUTCARs, which I don't think is what we would expect for MoTe2.
This is what your new structure looks like:

pic.png

And here is the original structure:

pic1.png

Also, I attach the OUTCAR files for calculations I did with your first POSCAR:
with EFIELD = 1.0: -36.056 eV
without field: -35.676 eV

[santanumahapatra]

Thank you for your detailed response.
Actually, the 2nd POSCAR is a transition image which I got while doing a NEB calculation. I wanted to check if the barrier height is affected by the action of Electric field. What steps can I take to see this effect in NEB? If the endpoints are affected by Efield, but the intermediate images of the NEB are not, then how can I tackle this to succesfully show a barrier change by Efield?

Regards.

[alexey.tal]

It is better to start with a NEB calculation without the field to make sure that you get a meaningful results. I think that the structure you showed looks like something went wrong in your NEB calculation. After you find a sensible barrier without the field it makes sense to proceed with turning the field starting from a small value and gradually increasing the field.

One thing I still don't understand is why you only want to enable field in the end points, but no the intermediate ones.

[santanumahapatra]

Thank you for your response.
I have enabled electric field for all the structures, including the intermediate ones. It's just that the electric field only affects the endpoints but not the intermediate images for some reason. I have used both IDPP and linear, still the NEB barrier height is not getting affected by the presence of the EFIELD parameters.

[alexey.tal]

I see that you have started already a topic about calculating the NEB with the field, so I would suggest that we do not scatter this discussion across multiple threads.
Regarding the band structure calculations with the field, were you able to get a band structure with the original POSCAR?

[santanumahapatra]

Yes, I have got my bandstructure results. Thank you for your help.
Regards