Hi Everybody!
I would greatly appreciate your comments on the results of the siesta
calculations described below.
1. I did siesta calculations for molecules (label M) adsorbed on a slab of
six Au(111)
layers (label 6L). In the attached figure, the molecule is on the left
side.
In a first step, I computed the slab alone (six Au layers without molecule
adsorbed). From the in-vacuum value E_vac^6L in file 6L.MAV obtained by
using 6L.VH (or 6L.VT because the xc potential vanishes in vacuum) as input
for macroave --- which agrees with the mean(=max. in this case) of the
output file 6L.out (siesta < 6L.fdf > 6L.out) and the Fermi energy E_F_6L
from 6L.out I can deduce the metallic work function W_6L=E_vac^6L - E_F_6L.
2. In the next step I computed molecules (M) adsorbed on six Au(111) layers.
Relaxing M on fixed metal atoms in the slab (may not be a too good
description of reality?). Left and right in-vacuum values
E_left/right^{vac} of VH/VT (M-6L.MAV) computed by switching on dipole
corrections (SlabDipoleCorrection .true.), unlike relaxation of M on the
slab done with SlabDipoleCorrection .false. (; is it OK?) The Fermi
corresponding energy E_F_M_6L reasonably agrees within 16 meV with E_F_6L.
Is it correct to determine work functions for the two sides as W_left/right
= E_left/right^{vac} - E_F_M_6L and to consider the difference delta_W_1 =
W_left - W_right = E_left^{vac} - E_right^{vac} as the change in the
metallic work function caused by the adsorboed molecules? The difference
between the Fermi energies with and without molecules (as said, 16 meV) is
comparable to the difference between E_right^{vac} and E_vac^6L (0.264 meV
vs. 0.240 eV, see attached figure); these differences should vanish taking
(semi)infinite slabs instead of 6 layers.
3. From the output file M-6L.out (siesta < M-6L.fdf > M-6L.out):
dhscf: Vacuum level (max, mean) = 0.588215 0.179863 eV
The maximum value 0.588215 eV is a puzzle for me; it does not coincide with
the
largest (right side, in-vacuum, 0.264 eV) in file M-6L.MAV (0.264 eV, see
attached figure). The mean value above (0.179863 eV) from file M-6L.out
agrees with the values from M-6L.MAV (0.264 + 0.094)/ = 0.179 eV).
4. By the way: if the differences between the potentials at the two
molecular ends could be measured, what is the difference to compare with:
delta_VT or delta_VH?
Many thanks for your help
ticu
