Download Orientation of perylene derivatives on semiconductor surfaces

Applied Surface Science 212–213 (2003) 501–507
Orientation of perylene derivatives on semiconductor surfaces
T.U. Kampen*, G. Salvan, A. Paraian, C. Himcinschi,
A.Yu. Kobitski, M. Friedrich, D.R.T. Zahn
Institut für Physik, Technische Universität Chemnitz, D-09107 Chemnitz, Germany
Abstract
The orientation of the perylene derivatives 3,4,9,10-perylenetetracarboxylic dianhydride (PTCDA) and N,N0 -dimethyl3,4,9,10-perylenetetracarboxylic diimide (DiMe-PTCDI) on sulphur passivated GaAs(1 0 0) surfaces and its impact on the
optical properties were studied by means of near-edge X-ray absorption fine structure spectroscopy (NEXAFS), Raman
spectroscopy, and variable angle spectroscopic ellipsometry (VASE). NEXAFS shows that PTCDA molecules lie flat on the
substrate with their molecular plane parallel to the substrate surface. DiMe-PTCDI molecules grown on the same type of
substrates are tilted with respect to the substrate surface and are predominantly oriented with their long axis parallel to the [1 1 0]
direction. The optical properties of these films investigated by VASE show that the DiMe-PTCDI films exhibit a much stronger
optical anisotropy than the PTCDA films.
# 2003 Elsevier Science B.V. All rights reserved.
PACS: 78.66.Qn; 78.30.Jw; 68.55.-a
Keywords: Organic films; PTCDA; Structural properties; Optical anisotropy
1. Introduction
The two perylene derivatives 3,4,9,10-perylenetetracarboxylic dianhydride (PTCDA) and N,N0 -dimethyl-3,4,9,10-perylenetetracarboxylic diimide (DiMePTCDI) form monoclinic crystalline structures with a
C2h symmetry. The unit cell contains 2 molecules
almost perpendicular and parallel to each other in
PTCDA and DiMe-PTCDI crystals, respectively. Due
to their symmetry crystals show anisotropies in their
physical properties. On substrates like Ag(1 1 1) [1],
Au(1 1 1) [2], and HOPG [3] PTCDA is adsorbed in a
‘‘herring bone’’ structure with two molecules almost
*
Corresponding author. Tel.: þ49-371-531-3079;
fax: þ49-371-531-3060.
E-mail address: [email protected] (T.U. Kampen).
perpendicular to each other and parallel to the substrate
surface in each unit cell. This molecular arrangement of
the molecules is similar to the one in the (1 0 2) plane of
a PTCDA crystal. On Ag(1 1 0) a phase transition is
observed from a single domain oriented homogenous monolayer with a ‘‘brick wall’’ structure into a
more condensed ‘‘herring bone’’ structure [4]. On
InAs(0 0 1) substrates the monolayer PTCDA interacts
relatively strongly with the substrate and two-dimensional overlayers are formed [5]. For larger coverages a
phase transition occurs where bulk-like three-dimensional PTCDA clusters begin to form. The strong
interaction between PTCDA and semiconductor surfaces is considerably reduced by a passivation of the
surface prior to the growth of the organic film [6,7].
Changing the structure of the molecule, e.g. by an
imide group with different side groups instead of the
0169-4332/03/$ – see front matter # 2003 Elsevier Science B.V. All rights reserved.
doi:10.1016/S0169-4332(03)00390-8
502
T.U. Kampen et al. / Applied Surface Science 212–213 (2003) 501–507
anhydride group, the growth properties are changed
drastically. Depending on the preparation procedure
di(2,6-isopropylphenyl)-3,4,9,10-perylenetetracarboxylic diimide molecules lie parallel or tilted with
their perylene core with respect to the Ag(1 1 1)
substrate [8]. In epitaxial films grown on cleaved
KCl(0 0 1) surfaces the molecular planes of DiMePTCDI molecules are cofacially stacked parallel to the
KCl substrate, whereas dibutyl-3,4,9,10-perylenetetracarboxylic diimide molecules are oriented standing
upright on the surface [9].
In this work we determined the orientation of
PTCDA and DiMe-PTCDI molecules grown on sulfur
passivated GaAS(1 0 0) surfaces by NEXAFS and
Raman spectroscopy. The impact of molecular orientation on the optical properties is investigated by VASE.
2. Experimental
All samples were prepared under ultra-high vacuum
(UHV) conditions on S-passivated Te-doped n-type
GaAs(1 0 0) (Freiberger Compound GmbH, n ¼ 2 1017 cm3). Prior to the sulfur passivation, the substrates were degreased by ultrasonic baths in acetone,
ethanol, and de-ionized water. S-passivation was performed by etching in a diluted solution of S2Cl2 and
CCl4 (1:3), followed by rinsing in CCl4, acetone,
ethanol, and de-ionized water. Immediately after the
chemical treatment the samples were transferred into
the UHV chamber and annealed at 430 8C. This results
in the formation of Ga2S3 like layer terminated by
single S atoms [10]. Onto these surfaces PTCDA and
DiMe-PTCDI films were grown by thermally evaporation of pre-purified material.
NEXAFS spectroscopy was performed at the PM1
beamline of the synchrotron light source BESSY II.
The data were recorded in total yield mode and the
light incidence angle was varied between normal and
near-grazing incidence (08 and 708). For the normalization spectra of a Ag film of 100 nm thickness were
taken and used as reference spectra.
Raman measurements were performed with a Dilor
XY 800 triple monochromator spectrometer equipped
with a CCD detector. As excitation energy the 2.54 eV
emission line of an Arþ laser was chosen which is
resonant with the first maximum of the absorption
spectra of both molecules. The films were measured in
a backscattering geometry with a spectral resolution of
2.5 cm1. The incident beam having a power of
30 mW was focused onto a spot of 300 mm in diameter. For comparison Raman measurements were
taken from a crystal obtained by sublimation using
an Olympus microscope with 100 magnification
objective in backscattering geometry with a spectral
resolution of 1.2 cm1. The spot size in this case was
1 mm in diameter and the power measured in the
focus was 0.08 mW.
Two polarization configurations were used for each
sample position. In the Porto notation zðxxÞz and zðxyÞz
denote the cases where the electric field vector of
scattered light is parallel/perpendicular to that of the
incident light (parallel/crossed polarization configuration). The position of the polarization analyser was
maintained fixed in all experiments, in order to eliminate the systematic error which might be induced by
the different response of the spectrometer to differently
polarized light. For the measurements performed on
the molecular films the laboratory axes coincide with
the substrate axis (x ¼ xs ¼ ½1 1 0
, y ¼ ys ¼ ½1 1 0
,
z ¼ zs ¼ ½100
) when the angle of rotation around the
surface normal is g ¼ 0 . The orientation of molecules
is determined from the depolarization ratio of the
breathing mode at 220 cm1 obtained upon rotation
of the sample around the surface normal.
For a simulation of the depolarisation ratio the
Raman tensor specific for the Ag modes is determined
from density functional calculations using Gaussian98
package at the B3LYP level of theory with a standard
6-31G(d) basis set [11]. The symmetry of the highest
occupied molecular orbital (HOMO) of DiMe-PTCDI
is Au, and that of the lowest unoccupied molecular
orbital (LUMO) is Bg. Therefore, a resonant Raman
effect that involves a HOMO-LUMO transition couples vibrational transitions with an electronic transition dipole of which is orientated along the xm axis
(where xm denotes the long axis within the molecular
plane while ym and zm denote short axis and the
direction perpendicular to the molecular plane, respectively). Under resonant excitation the off-diagonal
components disappear and for the lowest frequency
internal mode (220 cm1) the Raman tensor reads
0
1
1
0
0
@ 0 0:04 0 A
(1)
Am
g ¼
0
0
0
T.U. Kampen et al. / Applied Surface Science 212–213 (2003) 501–507
The spectroscopic ellipsometry measurements were
performed ex situ using a variable angle spectroscopic
ellipsometer (WVASE from J.A. Woollam Co., Inc.)
equipped with vertical sample mounting stage, continuously rotating analyser and auto-retarder and a Xelamp source. The measured ellipsometric data are
expressed in terms of the effective dielectric function
hei ¼ sin2 F þ sin2 Ftan2 F½ð1 rÞ=ð1 þ rÞ
2 where
r ¼ rp =rs tanðCÞeiD is the complex reflectance ratio
[12], F is the angle of incidence of light, C and D are
so-called ellipsometric angles, and rp and rs are the
Fresnel coefficients for light polarized parallel and
perpendicular to the plane of incidence, respectively.
503
3. Results and discussions
In NEXAFS absorption of the incident synchrotron
light takes place due to the excitation of an electron
from a core shell into the lowest unoccupied states. In
perylene derivatives the lowest unoccupied molecular
orbitals are derived from the p-electron system with its
orbitals oriented perpendicular to the molecular plane.
Due to selection rules excitation between the core
shell and the p-orbitals takes place if the electric field
vector of the incident light has a component parallel to
the p-orbitals. Fig. 1a presents selected C-K shell
spectra of PTCDA for different angles of incidence.
Fig. 1. NEXAFS spectra of the C-K edge of PTCDA (a) and the O-K edge of DiMe-PTCDI (b) as a function of the angle of incidence of
synchrotron light. Also the intensity ratio max1/max2 is plotted in b) as a function of angle of incidence.
504
T.U. Kampen et al. / Applied Surface Science 212–213 (2003) 501–507
The incident synchrotron light is linear polarized with
the electric field vector lying in the plane of incidence.
For an angle of incidence of 08 only broad features
above 280 eV photon energy are observed. When
changing the angle of incidence to 38 and 708, however, sharp structures below 280 eV photon energy
appear. The features above and below 280 eV are
attributed to unoccupied p- and s-states, respectively.
The intensity increase of the p-resonances with
increasing angle of incidence is clear evidence for
the parallel orientation of the molecular plane with
respect to the substrate surface. For DiMe-PTCDI the
p-resonances depend quite differently on the angle of
incidence. Fig. 1b shows the O-K shell spectra as a
function of the angle of incidence and the intensity of
the p-resonances normalized to the intensity of the
s-resonances. Here, a maximum in intensity of the
p-resonances is obtained for an angle of incidence of
about 408. From this result the angle between the
DiMe-PTCDI molecular plane and the surface is
determined to be about 508. While PTCDA films
grown on S-passivated GaAs show the same molecular
orientation as for the growth on other substrates, e.g.
the (1 0 2) plane of the crystal is parallel to the
substrate surface, the orientation of molecules in
DiMe-PTCDI films is more complicated and will
now be further evaluated by Raman spectroscopy.
In Fig. 2 Raman spectra of a DiMe-PTCDI single
crystal and a DiMe-PTCDI thin film are shown for the
two different polarization geometries. As a consequence of crystal formation with two molecules per
unit cell external vibrational modes can be observed in
the Raman spectra. The vibrational representation for
the external modes is: G ¼ 3Ag þ 3Bg þ 2Au þ Bu .
The six modes having even symmetry will show Raman
activity at frequencies below 125 cm1. As can be seen
in Fig. 2 phonons appear in the Raman spectra of single
phase crystals as well as in those of films exhibiting
similar polarization response. The phonon intensity
relative to that of the internal modes is larger in the
crystal compared to film spectra. The difference can be
explained by the larger size of the crystals compared to
that of the grains in the films. Moreover, the spot size in
the case of films allows to probe more than 105 grains
with various sizes. The strong polarization response for
both external and internal modes is remarkable and
indicates a preferred orientation of the grains with
respect to the substrate axes.
Fig. 2. Comparison between the Raman spectra of DiMe-PTCDI
crystal and film obtained in zðxxÞz (upper curves) and zðxyÞz (lower
curves) polarization configurations.
The packing of the molecules into a crystalline
environment is expected to affect the internal modes
due to in-phase and out-of phase coupling. This
effect called Davydov splitting depends on the
number of molecules in the unit cell and their dipole
and quadrupole interaction. The former determines
the multiplicity and the latter the amount of splitting. For DiMe-PTCDI with two molecules in the
unit cell there should be a two-fold splitting. However, due to the fact that the angle between the
molecules in the unit cell is 368 the dipole interaction which is proportional to the cosine of that
angle is so small that the amount of splitting lies at
the resolution limit of the Raman experiment.
Therefore, the use of a single molecule Raman
tensor is likely to be justified in this particular
case.
For a quantitative analysis of the polarization
response one defines the depolarisation ratio as the
intensity ratio between the Raman signal obtained in
crossed polarization configuration to that obtained in
parallel configuration. The experimental depolarisation
T.U. Kampen et al. / Applied Surface Science 212–213 (2003) 501–507
Fig. 3. Experimental (symbols) and simulated (lines) depolarization ratios of the breathing mode at 220 cm1 obtained upon
rotation around the sample normal with angle g for a DiMe-PTCDI
film.
ratio of the 220 cm1 mode for a film is depicted as a
function of rotation angle g in Fig. 3. In order to extract
the geometrical arrangement of the molecules from the
depolarisation ratios there are three coordinate systems
to take into account: molecular (xm, ym, zm), substrate
(xs, ys, zs) and laboratory (x, y, z). Two consecutive
transformations are required to transform the molecular Raman tensor to the Raman tensor Ag,l in the
laboratory reference frame. The first orthogonal transformation can be applied using the Eulerian angles
(j,y,c) which were previously applied by Aroca et al.
to Raman study of orientations in a highly symmetric
molecular system [13]. The second transformation is
from the substrate to the laboratory coordinate system
and implies a clock-wise rotation around the substrate
normal (zs) with the angle g. The Raman intensity is
then calculated as: I ¼ ðes Ag;l ei Þ2 , where ei and es
are the electric field vectors for the incident and the
scattered light.
A least square fit of the experimental ratio was
performed using a Levenberg-Marquard algorithm.
The angles are determined with an of 68. The best
match between the calculated depolarisation ratios
using a one-molecule approximation is represented
in Fig. 3 by the dashed line. Here, the molecular plane
is tilted with respect to that of the substrate by 538, and
the angle between the projection of xm onto substrate
plane and [1 1 0] substrate axis is 78. However, even
though the main maxima of the experimental data are
reproduced well, the steep minima and the same
height of all maxima indicate that a more complex
505
model is required. For the molecular planes of the two
molecules tilted by þ5 and 58 with respect to the
(xuyu) plane, and the long axis of the two molecules are
rotated by þ18 and 188 with respect to the xu axis,
several solutions provide a reasonable fit to the experimental data. The one which approaches mostly the
relative orientation of the two molecules in the crystal
unit cell is presented in Fig. 3 (full line) and provides
the following Euler angles: j ¼ 116 ; y ¼ 58 ,
c ¼ 25 .
This means, that the crystal plane (1 0 2) forms an
angle of 588 with the substrate plane, and the projections of the long molecular axis are deviating from the
[1 1 0] direction of the substrate by 9 and 488
respectively.
The impact of the different orientation of molecules in PTCDA and DiMe-PTCDI films on the
optical properties is demonstrated in Fig. 4. Here
the imaginary part of the effective dielectric function
(Imhei) for two azimuthal orientations is presented.
The PTCDA has already been discussed in detail
and is now compared with data obtained from
DiMe-PTCDI films [14]. For the PTCDA film the
ellipsometry spectra show characteristic absorption
features of PTCDA at 2.21 eV and also between
2.47–2.57 eV in agreement with the results of absorption spectroscopy [15,16]. The spectra measured
in both azimuthal directions differ slightly as can be
seen in the spectral region around 3.75 eV. At this
spectral position a simple shift is observed for the
[1 1 0] direction while the shift is accompanied by a
change in magnitude of Imhei in the case of [110].
The DiMe-PTCDI films, on the other hand, show a
very strong angular dependence on the angle of
incidence. This is accompanied by large differences
in Imhei between different azimuthal orientations of
the sample. As interference induced feature below 2
eV can be seen for the [110] direction, but not for
the [1 1 0] direction. On the other hand, the feature
around 3.2 eV attributed to the E1 gap in GaAs is
enhanced due to interference in the [1 1 0] direction.
The presence of negative values Imhei is also
induced by interference. With the help of absorption
spectroscopy one can also identify electronic transitions characteristic for DiMe-PTCDI in the spectra
of Imhei. The feature around 2.14 eV corresponds to
the first absorption maximum in the optical absorption spectrum. Its Imhei values are larger for the
506
T.U. Kampen et al. / Applied Surface Science 212–213 (2003) 501–507
Fig. 4. Imginary part of the effective dielectric constant for PTCDA (150 nm) and DiMe-PTCDI (120 nm) films grown on S-passivated
GaAs(1 0 0) surfaces. Spectra are taken for two different azimuths and different angles of icidence.
[110] azimuth. The higher transitions in the DiMePTCDI film observed in the absorption spectrum
are not very well resolved in the ellipsometry spectra
for this film also due to the interference effects
which may cover some of the transitions of the film
itself. These results clearly indicate that the optical
anisotropy is larger in the case of DiMe-PTCDI
films.
be modelled considering the Raman tensors of two
molecules arranged in orientation similar to the single
crystal. The angle between the (1 0 2) plane of crystalline domains are tilted with respect to the substrate
plane by 588 and the long molecular axis are deviating
from the [1 1 0] direction of the substrate by 9 and
488 respectively. The macroscopically extremely
high order of molecules in DiMe-PTCDI films leads
to a larger optical anisotropy compared to PTCDA
films.
4. Summary
The orientation of molecules in PTCDA and DiMePTCDI films is found to be parallel and tilted with
respect to the sample substrate surface, respectively.
Raman-active internal modes of DiMe-PTCDI exhibit
a strong polarization dependence and their depolarisation ratio varies with a period of 1808 as the sample
is rotated around substrate normal. This behaviour can
Acknowledgements
The authors gratefully acknowledge the financial
support provided by the Graduiertenkolleg ‘‘Dünne
Schichten und nichtkristalline Materialien’’ at Technical University Chemnitz and the EU DIODE network (HPRN-CT-1999-00164).
T.U. Kampen et al. / Applied Surface Science 212–213 (2003) 501–507
References
[1] E. Umbach, K. Glöckner, M. Sokolowski, Surf. Sci. 402–404
(1998) 20.
[2] T. Schmitz-Hübsch, T. Fritz, F. Sellam, R. Straub, K. Leo,
Phys. Rev. B 55 (1997) 7972.
[3] C. Kendrick, A. Kahn, S.R. Forrest, Appl. Surf. Sci. 104/105
(1996) 586.
[4] C. Seidel, J. Poppensieker, H. Fuchs, Surf. Sci. 408 (1998)
223.
[5] C. Kendrick, A. Kahn, J. Cryst. Growth 181 (1997) 181.
[6] Y. Hirose, S.R. Forrest, A. Kahn, Phys. Rev. B 52 (1995)
14040.
[7] T.U. Kampen, G. Salvan, D. Tenne, D.R.T. Zahn, Appl. Surf.
Sci. 175–176 (2001) 326.
[8] R. Nowakowski, C. Seidel, H. Fuchs, Phys. Rev. B 63 (2001)
195418.
[9] H. Yanagi, Y. Toda, T. Noguchi, Jpn. J. Appl. Phys. 34 (1995)
3808.
[10] T.U. Kampen, D.R.T. Zahn, W. Braun, C. Gonzáles, I. Benito,
J. Ortega, L. Jurczyszyn, J.M. Blanco, R. Pérez, F. Flores,
Appl. Surf. Sci., this volume.
[11] Gaussian 98 (Revision A.1), M.J. Frisch, G.W. Trucks, H.B.
Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman, V.G.
[12]
[13]
[14]
[15]
[16]
507
Zakrzewski, J.A. Montgomery, R.E. Stratmann, J.C. Burant,
S. Dapprich, J.M. Millam, A.D. Daniels, K.N. Kudin, M.C.
Strain, O. Farkas, J. Tomasi, V. Barone, M. Cossi, R. Cammi,
B. Mennucci, C. Pomelli, C. Adamo, S. Clifford, J.
Ochterski, G.A. Petersson, P.Y. Ayala, Q. Cui, K. Morokuma,
D.K. Malick, A.D. Rabuck, K. Raghavachari, J.B. Foresman,
J. Cioslowski, J.V. Ortiz, B.B. Stefanov, G. Liu, A.
Liashenko, P. Piskorz, I. Komaromi, R. Gomperts, R.L.
Martin, D.J. Fox, T. Keith, M.A. Al-Laham, C.Y. Peng, A.
Nanayakkara, C. Gonzalez, M. Challacombe, P.M.W. Gill,
B.G. Johnson, W. Chen, M.W. Wong, J.L. Andres, M. HeadGordon, E.S. Replogle, J.A. Pople, Gaussian, Inc., Pittsburgh
PA, 1998.
R.M.A. Azzam, N.M. Bashara, Ellipsometry and Polarized
Light, North - Holland Amsterdam, 1997.
R. Aroca, C. Jennings, R.O. Loutfy, A.M. Hor, J. Phys. Chem.
90 (1986) 5255.
T.U. Kampen, A.M. Paraian, U. Rossow, S. Park, G. Salvan,
M. Friedrich, D.R.T. Zahn, Phys. Stat. Sol. 181 (2001) 1307.
R. Kaiser, M. Friedrich, T. Schmitz-Hübsch, F. Sellam, T.U.
Kampen, K. Leo, D.R.T. Zahn, Fresenius J. Anal. Chem. 363
(1999) 189.
V. Bulovic, P.E. Burrows, S.R. Forrest, J.A. Cronin, M.E.
Thompson, Chem. Phys. 210 (1996) 1.