Optical anisotropy of organic layers on GaAs„001…

A. M. Paraiana)
Institut für Physik, TU Chemnitz, Reichenhainerstr. 70, 09107 Chemnitz, Germany
U. Rossow
Institut für Technische Physik, TU Braunschweig, Mendelssohnstr. 2, 38106 Braunschweig, Germany
S. Park, G. Salvan, M. Friedrich, T. U. Kampen, and D. R. T. Zahn
Institut für Physik, TU Chemnitz, Reichenhainerstr. 70, 09107 Chemnitz, Germany
共Received 26 March 2001; accepted 5 June 2001兲
The application of the linear-optical, polarization sensitive methods, in situ reflectance anisotropy
spectroscopy 共RAS兲, and ex situ spectroscopic ellipsometry, for the characterization of organic
layers is discussed and the results of the investigation of 3,4,9,10-perylenetetracarboxylic
dianhydride 共PTCDA兲 layers on sulfur passivated GaAs共001兲 surfaces are presented. The organic
layers were grown via organic molecular beam deposition at room temperature. The RA spectrum
of the sulfur terminated GaAs surface shows a derivative like feature at E 1 gap and a feature in the
higher energy range related to E 2 of bulk GaAs. Upon the PTCDA deposition, additional features
appear in the spectra which can be attributed to PTCDA while the GaAs feature near E 1 remains
unchanged indicating that the surface reconstruction stays intact. The imaginary part of the
pseudo-dielectric function is found to be angular dependent. This dependence also changes as a
function of azimuthal angle. While the first can be well described using existing models for optical
uniaxial layers, the latter is likely to be related to in-plane optical anisotropy. © 2001 American
Vacuum Society. 关DOI: 10.1116/1.1387462兴

I. INTRODUCTION and perpendicular to the plane of incidence, respectively.

Organic materials are used to an increasing extent for the
fabrication of 共opto-兲 electronic devices such as organic light
emitting diodes.1 One of the most widely investigated or-
ganic molecules is 3,4,9,10-perylenetetracarboxylic dianhy-
dride 共PTCDA兲. PTCDA has already been used for the depo-
sition of films on various substrates ranging from van der
Waals crystals2 to metals such as Ag and Au.3 It has recently
been demonstrated that the growth of organic molecules on
Se-passivated GaAs共001兲4 is considerably improved with re-
spect to the growth on the corresponding bare substrate.
However, studies of optical properties involving reflec-
tance anisotropy spectroscopy 共RAS兲 and spectroscopic el-
lipsometry are rather scarce. Unlike the substrate used which
is optically isotropic within the bulk due to its cubic symme-
try, the organic layers represent highly anisotropic systems,
therefore complicating data analysis. Optical anisotropy
can for instance be determined by RAS. Here the value
⌬r/r⫽2(r ␣ ⫺r ␤ )/(r ␣ ⫹r ␤ ) 5 which is the relative difference
in the complex reflectance, r ␣ and r ␤ , for the light polarized
along the ␣ and ␤ directions perpendicular to each other in
the surface plane is measured. As a result RAS predomi-
nantly probes the optical anisotropy in the plane parallel to
the surface 共x⫺y plane兲.
The measured ellipsometric data are expressed in terms
of the effective dielectric function 具 ⑀ 典 ⫽sin2 ⌽
⫹sin2 ⌽ tan2 ⌽关(1⫺␳)/(1⫹␳)兴2 where ␳ ⫽r p /r s tan(⌿)ei⌬ is
the complex reflectance ratio,6 ⌽ is the angle of incidence of
light, ⌿ and ⌬ are so-called ellipsometric angles, and r p and
r s are the Fresnel coefficients of the light polarized parallel
a兲
Electronic mail: andreea.paraian@s1999.tu-chemnitz.de
Here in a very simple picture optical anisotropy may result in
an angular dependent effective dielectric function if the op-
tical constants are different for directions parallel 共x or y
direction兲 and perpendicular 共z兲 to the surface plane. On the
other hand, optical anisotropy within the x⫺y plane may
show up when the effective dielectric function is measured
for different azimuthal directions.
II. EXPERIMENT
In this study, a n-type Te-doped GaAs共001兲 substrate with
a doping concentration of 3⫻1017 cm⫺3 was used for the
deposition growth of the organic films. Prior to growth the
GaAs substrate was sulfur passivated using a wet chemical
etching procedure under atmospheric conditions.7 This etch-
ing procedure consists of a sequence of three steps. In the
first step the substrate was cleaned using an ultrasonic bath,
by acetone, ethanol, and de-ionized water for 5 min each.
Afterwards the GaAs was dipped into a solution consisting
of a 3:1 mixture of CCl4 and S2Cl2 for 10 s. In the last step
the substrate was rinsed by dipping it successively into CCl4,
acetone, ethanol, and de-ionized water for 5 s each. After this
treatment the substrate was transferred into an UHV system
with a base pressure lower than 2⫻10⫺8 Pa and annealed at
430 °C. This procedure leads to a well ordered, (2⫻1) re-
constructed surface as revealed by low energy electron dif-
fraction. The sulfur dimers are likely oriented in 关110兴
direction.8
The PTCDA film was deposited at room temperature onto
the sulfur passivated GaAs substrate with an evaporation rate
of about 0.36 nm/min under the ultrahigh vacuum condi-

1658 J. Vac. Sci. Technol. B 19„4…, JulÕAug 2001 1071-1023Õ2001Õ19„4…Õ1658Õ4Õ$18.00 ©2001 American Vacuum Society 1658
1659 Paraian et al.: Optical anisotropy of organic layers on GaAs„001…
FIG. 1. RAS spectrum of sulfur-passivated GaAs共001兲.
tions. The film thickness was measured with a quartz mi-
crobalance. The deposition was interrupted for RAS mea-
surements at certain thicknesses.
For the in situ RAS measurements a spectrometer similar
in design to that described by Aspnes5 was employed. The
spectrometer was operated at near normal incidence in the
photon energy range from 1.8 to 5.3 eV. The spectra were
recorded by measuring the relative reflectance difference of
the sample in two directions ␣ and ␤ corresponding, respec-
tively, to the 关⫺110兴 and 关110兴 directions in the surface
plane of GaAs共001兲, respectively.
The following ellipsometric measurements were per-
formed ex situ using a spectroscopic ellipsometer from J. A.
Woollam Co., Inc. with the capability of measuring under
various angles of incidence from the near ultraviolet and vis-
ible spectral range. The incidence angle was varied in steps
of 5° from 60° to 75°. The spectra were measured in two
azimuthal directions by changing the azimuthal sample ori-
entation, i.e., with the 关110兴 or 关⫺110兴 direction of the
sample surface, respectively, parallel to the plane of incident
light.
III. RESULTS AND DISCUSSION
A typical reflectance anisotropy 共RA兲 spectrum measured
for a sulfur terminated GaAs共001兲 surface is shown in Fig. 1
and represents a RA fingerprint for the sulfur induced
(2⫻1) reconstruction of the GaAs surface. The sharp
derivative-like feature around 3 eV is attributed to the E 1 and
E 1 ⫹⌬ 1 critical points of GaAs,9 while the broad dip at 4.7
eV is associated with the E 0⬘ and E 2 critical points of GaAs.
The feature near 3.5 eV is related to the sulfur passivation.
Its origin is, however, still under discussion. While Hughes
et al. attributed its occurrence to a transition involving sulfur
dimers,10 other explanations like transitions within Ga–S
bonds are also possible.11
The evolution of RA spectra with increasing PTCDA cov-
erage is demonstrated in Fig. 2 for layers with nominal thick-
JVST B - Microelectronics and Nanometer Structures
1659
FIG. 2. RAS spectra of chalcogen-passivated GaAs upon PTCDA deposi-
tion.
nesses of 1, 3, 5, and 10 nm. Note that the features in the RA
spectrum of clean GaAs appear as very weak in this figure
clearly illustrating the pronounced optical anisotropy in-
duced by the presence of the PTCDA layer. Features indi-
cated by arrows can clearly be assigned to PTCDA and their
energy positions are listed in Table I. It is important to note
that these positions do not necessarily coincide with transi-
tion energies observed in optical absorption or ellipsometry
(Im具⑀典) spectra due to the low symmetry of PTCDA. Still a
comparison of the region enlarged in the inset of Fig. 2 with
optical absorption spectra12 reveals good correspondence
corroborating the assignment to PTCDA. The observation of
a finite RAS response in the energy range 2–2.7 eV where
no GaAs related features are present and the anisotropy of
the starting surface is small rules out that crystallites are
completely random in their azimuthal orientation. The RA
line shape of the GaAs related feature at 3 eV hardly changes
by deposition of PTCDA thus suggesting that the underlying
GaAs surface is quite likely to remain intact and chemical
interaction between organic molecules and the sulfur passi-
vated substrate is weak.
PTCDA features discussed so far are weak compared to
the RA features observed above 4 eV, which show a dra-
matic increase with coverage. It was already shown that in-
TABLE I. Spectral position of PTCDA features in the case of both reflectance
anisotropy spectroscopy 共RAS兲 and spectroscopic ellipsometry 共SE兲.
Energy/eV
RAS SE
2.15 2.21
2.29
2.37 ¯
2.42 2.47
2.54 2.57
2.63 ¯
¯ 3.38
¯ 3.52
1660 Paraian et al.: Optical anisotropy of organic layers on GaAs„001…
FIG. 3. Imaginary part of the pseudo-dielectric function of 100-nm-thick
PTCDA layer on sulfur-passivated GaAs共001兲.
terference effects are predominantly contributing to these
high energy range features.13 Unfortunately, optical absorp-
tion data for PTCDA are not available for this energy range.
It is known from other experimental results, e.g., obtained
by Raman spectroscopy14 that PTCDA layers deposited at
room temperature are not single crystal but are polycrystal-
line in nature. A future more detailed RAS analysis may help
to understand the structural order.
After PTCDA deposition ellipsometric measurements
were carried out at a different angle of incidence ex situ for a
PTCDA layer of 100 nm nominal thickness. Figure 3 pre-
sents the corresponding imaginary part of the effective di-
electric function (Im具⑀典) for both azimuthal orientations.
In the ellipsometry spectra characteristic absorption fea-
tures of PTCDA at 2.21 eV and also between 2.47 and 2.57
eV occur in agreement with the results of absorption
spectroscopy.12,15 Even two weak features at 3.38 and 3.52
eV appearing in the absorption spectra are also present in the
spectra of Im具⑀典. The features assigned to PTCDA are also
listed in Table I.
At the first view the spectra measured in both azimuthal
directions look very similar. There are, however, differences
as can, for instance, be seen in the spectral region around
3.75 eV. At this spectral position a simple spectral shift is
observed for the 关110兴 direction while the shift is accompa-
nied by a change in the magnitude of Im具⑀典 in the case of
关⫺110兴. Attempts were made to simulate the line shape. The
results are shown in Fig. 4. These were obtained under the
J. Vac. Sci. Technol. B, Vol. 19, No. 4, JulÕAug 2001
1660
FIG. 4. Simulated Im具⑀典 of effective dielectric function and the imaginary
parts of the PTCDA dielectric function in the substrate plane used for simu-
lation.
following assumptions: The planar PTCDA molecules pref-
erentially lie almost flat on the substrate surface when depos-
ited at room temperature and with low deposition rate. This
is supported by our own unpublished infrared investigations.
Consequently the films are supposed to be uniaxial thus ex-
hibiting different optical constants in the x⫺y plane and per-
pendicular to it. The high number of unknown parameters
and their strong correlation makes it impossible to determine
optical constants from the measurements of a single film.
Therefore optical constants previously determined by a
multi-sample analysis of PTCDA films in the energy range
up to 4.2 eV were used to simulate the effective dielectric
function. The in-plane imaginary part of the dielectric func-
tion ( ⑀ 2 ) of PTCDA is also shown in Fig. 4 by a thin solid
line. The corresponding in-plane optical constants together
with the out-of-plane ones served as input for the simula-
tions. The starting point was the simplest model of an
uniaxial layer of PTCDA on GaAs. For the best match with
experimental data a thickness inhomogeneity of about 8%–
10% and an additional surface roughness of about 2.5% of
the film thickness had to be included. The latter was deter-
mined to be about 150 nm which is higher but still in fairly
good agreement with the nominal thickness of 100 nm.
Comparing the simulated spectra with the experimental
results, it is apparent that satisfying agreement in line shape
considering all angles is achieved only for the spectra with
关110兴 azimuthal direction in the photon energy range from
1.5 to 4 eV. This also holds for a comparison of measured
and simulated real parts Re具⑀典 of the dielectric function.
Looking, e.g., at the dip at 3.75 eV and its angular depen-
dence the model is able to describe the spectra measured in
关110兴 direction but not the upward shift in magnitude with
increasing angle of incidence in the 关⫺110兴 direction. In
principle, such upward shift can be modeled for one angle by
an increased inhomogeneity of film thickness, but the angu-
lar variation should then also only induce a horizontal shift
of the dip rather than the observed upward shift. The current
model can thus not account for the observed changes of
Im具⑀典 with angle of incidence at 关⫺110兴 azimuth and re-
quires further refinement.
1661 Paraian et al.: Optical anisotropy of organic layers on GaAs„001…
For a first simulation in the higher energy range⬎4 eV the
complex dielectric function was approximated up to 5 eV
assuming an almost absorption free film on GaAs. The cal-
culated strong peak at about 4.5 eV does not reflect the mea-
sured Im具⑀典. Therefore a point to point fit of the measured
real and imaginary part of the effective dielectric function
共关110兴 azimuthal direction兲 was performed keeping all pa-
rameters constant and varying only the in-plane dielectric
function. This is allowed in a first approximation since the
in-plane optical constants are dominating the spectral re-
sponse. The result of the fit is shown as a dashed line in Fig.
4 revealing an absorption peak at about 4.7 eV superimposed
on a background absorption increasing towards higher ener-
gies. This peak is likely to be related to the ␲␲ * transition
highest occupied molecular orbital (HOMO)→lowest unoc-
cupied molecular orbital (LUMO)⫹4 of PTCDA as calcu-
lated by Scholz.16 Going back to the RA spectra in Fig. 2, the
strong feature evolving at the same energy as this absorption
peak may therefore be composed of contributions due to this
PTCDA absorption and additional interference enhanced
GaAs related features.
Future work will involve correlation of the observed dif-
ferences in ellipsometry spectra for different azimuthal ori-
entations with the RA spectra.
IV. SUMMARY
PTCDA films grown at room temperature on sulfur-
passivated GaAs共001兲 were investigated by means of in situ
JVST B - Microelectronics and Nanometer Structures
1661
RAS and ex situ spectroscopic ellipsometry. The RA spec-
trum of passivated GaAs shows bulk-like features induced
by the presence of a reconstructed surface. During film
growth the development of anisotropy was detected by RAS
and after growth confirmed by spectral ellipsometry. The re-
sults indicate a preferential in-plane orientation of organic
crystallites.
1
S. R. Forrest, Chem. Rev. 97, 1793 共1997兲.
2
P. E. Burrows, Y. Zhang, E. I. Haskal, and S. R. Forrest, Appl. Phys. Lett.
61, 2417 共1992兲.
3
K. Glöckler, C. Seidel, A. Soukopp, M. Sokolowski, E. Umbach, M.
Böhringer, R. Berndt, and W. D. Schneider, Surf. Sci. 405, 1 共1998兲.
4
C. Kendrick and A. Kahn, Appl. Surf. Sci. 123Õ124, 405 共1998兲.
5
D. E. Aspnes, J. P. Harbison, A. A. Studna, L. T. Florez, and M. K. Kelly,
J. Vac. Sci. Technol. A 6, 1327 共1988兲.
6
R. M. A. Azzam and N. M. Bashara, Ellipsometry and Polarized Light
共North-Holland, Amsterdam, 1997兲.
7
Z. S. Li, W. Z. Cai, R. Z. Su, G. S. Dong, D. M. Huang, X. M. Ding, X.
Y. Hou, and X. Wang, Appl. Phys. Lett. 64, 3425 共1994兲.
8
M. D. Pashley and D. Li, J. Vac. Sci. Technol. A 12, 1848 共1994兲.
9
D. E. Aspnes and A. A. Studna, Phys. Rev. B 27, 985 共1983兲.
10
G. Hughes, C. Springer, U. Resch, N. Esser, and W. Richter, J. Appl.
Phys. 78, 1948 共1995兲.
11
V. L. Berkovits and D. Paget, Appl. Phys. Lett. 61, 1835 共1992兲.
12
R. Kaiser, M. Friedrich, T. Schmitz-Hübsch, F. Sellam, T. U. Kampen, K.
Leo, and D. R. T. Zahn, Fresenius J. Anal. Chem. 363, 189 共1999兲.
13
T. U. Kampen, U. Rossow, M. Schumann, S. Park, and D. R. T. Zahn, J.
Vac. Sci. Technol. B 18, 2077 共2000兲.
14
T. U. Kampen, G. Salvan, D. Tenne, R. Scholz, and D. R. T. Zahn, Appl.
Surf. Sci. 175–176, 326 共2001兲.
15
V. Bulovic, P. E. Burrows, S. R. Forrest, J. A. Cronin, and M. E. Thomp-
son, Chem. Phys. 210, 1 共1996兲.
16
R. Scholz 共private communication兲.