Chemistry Reference
In-Depth Information
OMBE. However, ellipsometric quantities “as measured” are not easily interpreted
and require modeling the sample and its dielectric tensor to extract meaningful
information [
9
,
10
]. This could possibly explain the lack of a really diffuse
application of ellipsometry to monitor the growth and the properties of thin layers
of porphyrins, although some interesting examples have been reported [
11
]. The
interested reader could refer to LoSurdo and Hingerl [
12
] for a detailed and updated
presentation of the technique.
For SDR and RAS, on the contrary, no modeling is strictly necessary to extract
the basic significance from data, that is, (i) some changes have been caused by the
process that the sample has undergone (SDR) or (ii) the sample exhibits an intrinsic
optical anisotropy, intimately connected to its structure or to its electronic proper-
ties (RAS). Also in these cases, for a detailed interpretation of data, a further and
deeper theoretical approach and simulation of the investigated system is needed, as
well as a comparison with results from other techniques.
In SDR, the variation of the sample reflectance due to the surface modification,
e.g., because of contamination (sometimes intentionally produced) or of the depo-
sition of a film, is monitored by
Δ
R
=
R
SDR
¼
ð
R
a
R
b
Þ=
R
,
ð
1
Þ
where
R
a
(
R
b
) is the reflectance of the sample
before
(
after
) its surface is modified,
and R is the average reflectance (
R
a
þ
R
b
)/2. In the well-known application of SDR
to clean semiconductor surfaces, “a” means the clean surface and “b” the oxidized
surface, where the growth of the oxide layer has removed the surface state transi-
tions by saturating the dangling bonds of the clean surface [
4
,
13
].
Under the assumption that the reflectance change is directly a consequence of the
variation in the sample conditions, by Eq. (
1
), we can disentangle the related effect
from
R
/
R
SDR
data and attribute it to the modification that occurred at the outer
layer. Effects due to a change in the substrate however cannot in principle be
discarded and should be carefully considered case by case [
14
].
A more detailed interpretation in terms of the grown film is strongly dependent
also upon the optical properties of the substrate. If this is not absorbing in the
photon energy range used, the detection of new structures in the spectrum can be
directly related to the film. If instead the substrate is absorbing, a more careful and
complex interpretation of the signal is due, as the dielectric functions of the layer
and of the substrate (if we assume - as it normally happens - that the outer medium,
surrounding the sample, is transparent) are profoundly entangled in
Δ
R
/
R
SDR
.In
that case, a detailed deconvolution of spectra is necessary: this usually means an
interpretation of the data within a classical three-layer model plus a Kramers-
Kronig transform [
13
,
15
] that will be discussed in Sect.
3.2
.
A significant example is reported in Fig.
1
, showing SDR spectra that have been
measured
during
the deposition in ultrahigh vacuum by organic molecular beam
epitaxy of a PTCDA (3,4,9,10-perylenetetracarboxylic dianhydride) film on mus-
covite mica(0001) at room temperature [
16
]. The film thickness ranges from 0.1 to
2.7 ML (with a coverage error estimated as 10%). The mica substrate is transparent
Δ
