Question about calculating Work function of metal

[zmisha]

Hello Sir,
   
   http://quantumwise.com/publications/tutorials/mini-tutorials/134 is the tutorial that how to calculation the Work function.
   I followed the tutorial and tried to calculate the work function of Ag-111.
   What I don't understand is the choice of top vacuum length and bottom vacuum length. When i use Top vacuum = 20 Å and Bottom vacuum = 5 Å, which is the same as the tutorial, the Work function is 4.775822 eV. But when i use Top vacuum = 35 Å and Bottom vacuum = 30 Å, 4.758183 eV.
   Could you tell me why different vacuum length and bottom vacuum length give different result?

   Thanks for your time.

[zh]

This is a normal phenomenon. The calculated work function may be affected by the vacuum thickness. Usually one has to do a convergence test for the vacuum thickness in order to avoid the interaction between repeated images in the slab supercell calculations.  Alternatively, your test calculations suggest the vacuum thickness of 25 angstrom may be sufficient because the obtained workfunction is only  0.017639 eV smaller than that obtained by the use of 65 angstrom.

[zmisha]

Thank you sir.

[Subhban15]

How does one perform a convergence test for vacuum thickness for vacuum on the sides (along the width) of a nanoribbon? Does the procedure vary with calculator or material? I plan to use DFT GGA.RPBE calculator for Silicene. How must I check for the best vacuum for the sheet 10 atoms wide, 10 atoms long sheet?

[Umberto Martinez]

plot the WF against vacuum spacer anche check for convergence, for example.
in ATK (LCAO method) 10 Å should be enough.

It is also a good idea to plot the effective potential along the direction perpendicular to your surfaces.
See here for more details: http://quantumwise.com/publications/tutorials/item/519-tuning-the-work-function-of-silver-by-deposition-of-ultrathin-oxide-layers

[zh]

Additional comment:

The effective potential printed out by ATK contains three contributions, i.e., Hartree potential, XC potential,  other electrostatic interactions (e.g., ionic potential (pseudoptential)). The convergence of  planar average of effective potential with the vacuum thickness is much slower than the Hartree potential. This behavior has been already demonstrated in the following paper:
Deriving accurate work functions from thin-slab calculations, J. Phys.: Condens. Matter 11 (1999) 2689–2696, http://iopscience.iop.org/0953-8984/11/13/006

The plot of planar average of ElectrostaticDifferencePotential could give fast convergence of vacuum level with respect to the vacuum thickness.

[Subhban15]

Thank you Umberto and zh.

The subject probably misdirected my question.

I am not trying to find the work function. I am simply unable to decide how much vacuum around the nanoribbon is necessary (specifications mentioned in my previous post). I read here that one must check for convergence of vacuum thickness. I want to know the procedure for checking the convergence.

[zh]

Draw the plot of the planar average of ElectrostaticDifferencePotential and see whether the curve at the vacuum region is flat or not.
If the vacuum thickness is too thin, the curve at the center of vacuum region may not flat. In this case, you have to increase the vacuum thickness.

[zh]

Hello Sir,
   
   http://quantumwise.com/publications/tutorials/mini-tutorials/134 is the tutorial that how to calculation the Work function.
   I followed the tutorial and tried to calculate the work function of Ag-111.
   What I don't understand is the choice of top vacuum length and bottom vacuum length. When i use Top vacuum = 20 Å and Bottom vacuum = 5 Å, which is the same as the tutorial, the Work function is 4.775822 eV. But when i use Top vacuum = 35 Å and Bottom vacuum = 30 Å, 4.758183 eV.
   Could you tell me why different vacuum length and bottom vacuum length give different result?

   Thanks for your time.

Please show the plot of planar average of ElectrostaticDifferencePotential (or effective potential) which used to estimate the work function in your calculations. 

[Subhban15]

Oh okay! Thank you very much!

Merry Christmas!

[Subhban15]

Draw the plot of the planar average of ElectrostaticDifferencePotential and see whether the curve at the vacuum region is flat or not.
If the vacuum thickness is too thin, the curve at the center of vacuum region may not flat. In this case, you have to increase the vacuum thickness.

I have added vacuum by changing the lattice parameters and then created a device (simply from bulk to device) and then calculated the ElectronDifferencePotential.

I cannot find much difference, i.e. the curve at the centre of the vacuum region seems to be flat in all the three cases (5A, 10A and 20A vacuum on sides of the width of the ribbon as well as above and below the surface of the ribbon)

C direction is the transport direction where the electrodes are located.
B implies the width of the ribbon.
A is direction perpendicular to the surface.

Could you please tell me how much amount of vacuum must be incorporated for further calculations?

Thanks in advance!

[zh]

From your plot, the vacuum thickness in the range from 5 angstrom 10 angstrom seems sufficient.  The use of too large value for vacuum thickness may waste the computational time.

[Subhban15]

Okay got it! Thank you very much!