Chemistry Reference
In-Depth Information
A significant restriction to the application of SDR exists when real-time moni-
toring of a process is essential: the SDR signal raises from a change in the sample
that is necessarily induced via a process able to cause a modification of its
properties, often (not always or necessarily) permanent or irreversible. This
means that SDR cannot detect the state of the sample
as it is
in the moment of
the experiment, but just its variation. In other words, the SDR signal is always
intimately connected to both what the sample has been “before” and what it is
“after” a certain modification.
Differently from SDR, reflectance anisotropy spectroscopy (RAS) [
18
,
19
]
allows to investigate the optical properties of a thin layer (as well as of a surface)
by exploiting the different symmetry of the layer itself with respect to the under-
lying bulk. Linearly polarized light is shined at near-normal incidence onto the
sample surface, detecting the reflected light while polarization is switched between
two orthogonal directions
, usually suggested by the sample symmetry or
structure. The difference of light intensity for different polarizations (i.e., the
reflectance anisotropy), normalized to the average reflectance, is expressed as
ʱ
and
ʲ
=
:
Δ
R
=
R
RAS
¼
2
R
ʱ
R
ʲ
R
ʱ
þ
R
ʲ
ð
2
Þ
R
/
R
RAS
can be directly related to the layer optical properties if the underlying
bulk symmetry allows to cancel out its contribution, as it happens in centrosym-
metric materials. Obviously, the meaningful application of RAS directly depends
upon the anisotropy of the system to be investigated or to the existence of an
unbalanced contribution of domains with well-defined and different symmetry. An
isotropic layer, as well as a layer exhibiting domains equally balancing their
contribution, produces a null RAS signal (Fig.
5
).
The RAS signal, similarly to SDR, can be interpreted within a three-layer model,
expressing
Δ
R
/
R
RAS
in terms of the dielectric properties of the bulk, of the layer
itself, and of the surrounding media (usually vacuum or air, but not necessarily).
Similarly to the case of SDR, if the substrate is transparent, the measured anisotropy
can be directly interpreted in terms of an asymmetry descending from the layer. In
Fig.
6
, we report a significant example from surface physics. The (111) cleavage
surface of diamond, 2
Δ
1 reconstructed, presents a very sharp anisotropy peak at
about 1.5 eV [
20
], well inside the bulk energy gap (~5.5 eV at 300 K), thus clearly
assessing its origin due to the
-bonded chain of the reconstructed surface, although
the state-of-the-art theory for this surface is not able to explain yet its energy
position given by experiments [
21
, and references therein].
If the substrate is absorbing, also for RAS, the presence of peaks cannot be
directly interpreted as the sign of anisotropic optical transitions at certain photon
energies, at least until the imaginary part of the anisotropic dielectric function has
been computed, similarly to the case of SDR. A significant example will be
presented in Sect.
3.2
.
A third case must be taken into account: a substrate with intrinsic anisotropic
optical properties. This situation can be met in inorganic semiconductors, for
example, for vicinal surfaces, due to the directional strain caused by the surface
termination, producing well-defined spectral features at bulk critical points [
22
,
23
],
ˀ
