I want to use pp.x for calculation of work function for a semiconductor slab,
I have calculated the difference between the fermi energy and vacuum level using the plot_num=(1 and 11), for investigate the effect of including xc term.
Including xc term give rise to significant ( about 1 eV) decrease in total potential in vacuum region.
which one is physically meaningful?--considering xc potential or not?-
furthermore, the fermi energy in output is strongly sensitive to degauss value?
I think if we set occupation~fixed--without any smearing-- we should consider the VB edge as the Fermi energy at 0K.
In my opinion the electrostatic potential has to be considered without XC contribution for the WF. I'm using VASP for metals, and gives perfect WF results. For semiconductors it can indeed be more tricky by setting the Fermi level correctly, the VB edge sounds reasonable to me.
XC is zero for zero density so it should decay in the vacuum but since your vacuum is finite, you have nonzero density which results in non zero XC. Omitting this XC is, therefore, a better approximation for the true vacuum.
If XC contribution to the WF is large, the error due to slab-slab interaction should be large too, since it means you have significant density inbetween. Consider running an scf with a larger vacuum after your system has relaxed in the tight one.
Dear All
I am practicing this calculation based on the example given for Al slab and bulk, in Quantum Espresso package. But I couldn't find clear relationship between the unit cell size (in the scf calculation) and the Length in the plot of the final Energy vs. length plot (in the Al100.wf.eps file).
In fact, this length (in the plot) is obtained after Re-scaling the out put of the average.x by a factor of 0.52918 (see the run_example file at line 290). What is the importance of this scaling?
How does the package obtained the original range of values of this length (0 to about 34.491)? Has this length any relationship with the input in the scf calculation?
Is there any method to estimate the possible value of npt for the input of the average.x? Or how can I know the possible value of the nfft?
Please your response has great significance for my progress.
Regards
- Badges
- Science topic
- Similar topics
- Filing