JOURNAL OF APPLIED PHYSICS 97, 063518 共2005兲
Optical Properties of 3,4,9,10-perylenetetracarboxylic dianhydride/copper
phthalocyanine superlattices
O. D. Gordan,a兲 S. Hermann, M. Friedrich, and D. R. T. Zahn
Institut für Physik, Technische Universität Chemnitz, D-09107 Chemnitz, Germany
共Received 15 September 2004; accepted 22 December 2004; published online 11 March 2005兲
Organic superlattices consisting of five alternating layers of 3,4,9,10-perylenetetracarboxylic
dianhydride 共PTCDA兲 and copper phthalocyanine 共CuPc兲 were prepared by organic molecular-beam
deposition in high vacuum on hydrogen-passivated, 共111兲-oriented silicon. The substrates were kept
at room temperature during the deposition. The optical response of the multilayered structure was
investigated by means of spectroscopic ellipsometry in the spectral range from 0.73 to 5 eV. While
the infrared spectra show that there is no chemical interaction between the two pigments, the
ellipsometry evaluation suggests an electronic coupling between the ␲ orbitals of the PTCDA and
the ␲ orbitals of the CuPc. This means that the modeling of the optical response requires a more
sophisticated approach than simply superimposing the responses of the individual layers. © 2005
American Institute of Physics. 关DOI: 10.1063/1.1861967兴
I. INTRODUCTION
II. EXPERIMENT
The development of next generation 共opto兲electronic devices involves the application of advanced materials and improved structures. Compared to bulk materials, lowdimensional structures exhibit unusual optical and physical
properties. Since their discovery, inorganic crystal
superlattices1 have attracted a lot of attention, especially in
regard to quantum confinement effects. In contrast organic
superlattices have been much less studied and research on
them has been performed only within the last decade. So and
Forrest2 showed exciton confinement in an organic superlattice consisting of alternative layers of 3,4,9,10perylenetetracarboxylic dianhydride 共PTCDA兲 and 3,4,7,8
naphtalene-tetracarboxylic dianhydride 共NTCDA兲 prepared
by organic molecular-beam deposition 共OMBD兲. Imanishi et
al.3 also proved by means of X-ray and transmission electron
microscopy the formation of an organic superlattice consisting of copper phthalocyanine 共CuPc兲 and NTCDA.
Ellipsometry is a powerful tool commonly employed for
the optical characterization of complex semiconductor
heterostructures.4 The implementation of mathematical
algorithms5 based on a 4 ⫻ 4 transfer-matrix formalism developed by Berreman6 also allows anisotropic layers to be
evaluated using ellipsometry.
We report the optical properties of an organic superlattice consisting of alternative layers of PTCDA and CuPc.
These two materials are red and blue pigments, respectively,
used in dye industry. Both molecules have an intrinsic optical
anisotropy due to their planar structure leading to anisotropic
layer properties. The optical response of a multilayer structure may be of particular interest as the materials exhibit
complementary absorption in the visible range. The absorption spectra are dominated by ␲-␲* transitions with the
shape of the bands depending on the electronic interaction
between the molecules.7–9
The organic superlattices consisting of five alternating
layers of PTCDA and CuPc were prepared by OMBD in high
vacuum 共HV− 8 ⫻ 10−7 mbar兲 on hydrogen-passivated
Si共111兲. The substrates were kept at room temperature during the deposition. In order to evaluate the optical constants
of CuPc and PTCDA several thin films of these materials
were also prepared as single layers. The source material was
␤-phase CuPc supplied by Aldrich and PTCDA powder supplied by Lancaster. The Si共111兲 substrates were cleaned with
isopropanol and de-ionized water. Hydrogen passivation was
performed using a HF共40%兲 etching for 2 min. After passivation the substrates were immediately transferred into the
deposition chamber. The thickness of the organic material
was monitored by a quartz crystal microbalance which was
located in the vicinity of the samples. The change in the
resonant frequency of the quartz is proportional to the film
thickness. The deposition rate was kept constant at approximately 0.3 nm/ min.
Infrared 共IR兲 measurements were performed using a
Bruker Fourier transform infrared 共FTIR兲 spectrometer IFS
66. All samples were measured in reflection at 60° using sand p- polarized light. KBr pellets containing crystallites
with a random orientation of ␤-form CuPc and PTCDA were
also prepared. The pellets were measured in transmission.
Ellipsometric measurements were carried out using a
variable angle spectrometric ellipsometer 共VASE, J. A. Woollam Co., Inc.兲. In order to determine the film thicknesses and
the energy dependence of the optical constants, ellipsometric
spectra were recorded at different angles of incidence 共55°,
65°, and 75°兲 in the range of 0.73– 4.5 eV with a 0.02-eV
step for each sample. A detailed description of the ellipsometry principles and theory can be found in Refs. 5 and 10.
a兲
The anisotropic dielectric functions of CuPc and
PTCDA are presented in Figs. 1 and 2, respectively. The
Author to whom correspondence should be addressed; electronic mail:
[email protected]
0021-8979/2005/97共6兲/063518/5/$22.50
III. RESULTS
97, 063518-1
© 2005 American Institute of Physics
Downloaded 27 May 2005 to 134.109.68.45. Redistribution subject to AIP license or copyright, see http://jap.aip.org/jap/copyright.jsp
063518-2
Gordan et al.
FIG. 1. Anisotropic dielectric function for CuPc. a-real part, b-imaginary
part.
ellipsometric model applied for the determination of the dielectric function is identical to the one described in Ref. 11.
The optical anisotropy and the shape of the absorption for
PTCDA are similar to the ones reported in Refs. 12 and 13.
The IR spectra confirm the anisotropy of the CuPc and
PTCDA layers derived from ellipsometry. Figure 3 shows the
IR spectra of the CuPc layers compared with the bulk one.
The spectra were normalized with respect to the reflection of
the Si共111兲 substrate and the 1100-cm−1 peak. The peak positions for all samples indicate that the films consist of
␣-phase CuPc while the CuPc in the pellet exhibits ␤-form
characteristics.14,15 According to Debe16 the 722 and
750 cm−1 peaks correspond to the out-of-plane vibrations of
the CuPc molecule while the peak at 753 cm−1 and all bands
above 800 cm−1 are due to molecular in-plane vibrations.
Considering p polarization, the 722 and 750 cm−1 peaks have
derivative like shapes or are pointing downwards. This indicates that the displacement of atoms for these oscillations is
in the z direction of the films 共perpendicular on the film
surface兲. Consequently, the average tilt angle of the molecular plane with respect to the substrate surface is less than 45°.
This can explain the higher coupling of the p polarization
with the out-of-plane vibrations of the molecule, while for s
polarization the IR spectra are similar to that of the bulk.
A similar evaluation of the IR spectra was performed for
J. Appl. Phys. 97, 063518 共2005兲
FIG. 2. Anisotropic dielectric function for PTCDA. a-real part, b-imaginary
part.
PTCDA 共Fig. 4兲. In this case the molecule exhibits out-ofplane vibrations below 900 cm−1. As it can be seen in Fig. 4
these bands appear just for the pellet spectra and for the p
polarization while they disappear completely for the s polarization. This indicates that the PTCDA molecules lie flat with
respect to the Si共111兲 substrate surface. This is in agreement
with Ref. 13.
FIG. 3. IR reflection spectra of the CuPc layers using s- and p-polarized
lights at 60° angle of incidence.
Downloaded 27 May 2005 to 134.109.68.45. Redistribution subject to AIP license or copyright, see http://jap.aip.org/jap/copyright.jsp
063518-3
J. Appl. Phys. 97, 063518 共2005兲
Gordan et al.
FIG. 4. IR reflection spectra of the PTCDA layers using s- and p-polarized
lights at 60° angle of incidence.
The ellipsometry spectra of the multilayered structure
were simulated, taking into account the previously determined optical constants of CuPc and PTCDA. A superlattice
model was built using the WVASE software and the layer
thicknesses were determined via a numerical fit. In this
model the thicknesses of all five PTCDA layers are considered to be the same. A similar condition is imposed also for
the CuPc layers. The differences between the model and the
experimental points are evaluated through the values of the
mean-square error 共MSE兲.
MSE =
冑
1
2N − M
N
兺
i=1
冋冉
⌿mod
− ⌿exp
i
i
exp
␴⌿,i
冊 冉
2
+
⌬mod
− ⌬exp
i
i
exp
␴⌬,i
冊册
2
,
where N is the number of the experimental points, M is the
number of fit parameters, and ␴ is the standard deviation for
each point.
Very good agreement between the simulated and experimental data was achieved considering the PTCDA layer as
being anisotropic while only small changes in the MSE are
noticed when the CuPc layer is considered as anisotropic or
isotropic. This can be explained by the markedly smaller
anisotropy of the CuPc layer compared to the strong anisotropy of the PTCDA layer.
Using the previously determined dielectric functions a fit
was performed in order to obtain the thicknesses. The results
are summarized in Table I. In a first model sharp interfaces
between PTCDA and CuPc were assumed. The MSE value
was 35.4 while taking into account an additional intermixed
interface layer the MSE was lowered in a second model to
19.4. The intermixed layer is simulated using an effective
medium approximation consisting of 64% PTCDA and 46%
TABLE I. Thickness of the PTCDA and CuPc layers in the superlattice
determined from ellipsometry.
Model 1
Thickness/ nm
Model 2
Thickness/ nm
PTCDA
¯
CuPc
9.5± 0.12
¯
3.7± 0.11
PTCDA
Intermix
CuPc
5 ± 0.19
2.9± 0.21
2.2± 0.43
FIG. 5. Ellipsometric ␺ and ⌬ spectra at 55°, 65°, and 75° angles of incidence for a superlattice with five alternating layers of PTCDA and five
alternating layers of CuPc on a Si共111兲 substrate. The open circles are the
experimental points and continuous lines the fits.
CuPc. The percentage is also a parameter of the fit. The
results are again summarized in Table I. The atomic force
microscopy 共AFM兲 investigation on single layers of PTCDA
and, respectively, CuPc revealed a surface roughness between 2.5 and 3.5 nm for the PTCDA and between 2 and
3 nm for the CuPc. This indicates that the mixing of the
materials at interfaces has to be taken into consideration.
Downloaded 27 May 2005 to 134.109.68.45. Redistribution subject to AIP license or copyright, see http://jap.aip.org/jap/copyright.jsp
063518-4
J. Appl. Phys. 97, 063518 共2005兲
Gordan et al.
FIG. 6. The in-plane imaginary part of the dielectric function of CuPc and
PTCDA compared with the optical response of the superlattice.
Figure 5 shows the experimental ellipsometric spectra
and the simulated data with the two models. As can be seen
in the figure, the second model provides an almost perfect
match with the experimental data in the range from 0.73 to
1.5 eV. The corresponding MSE value is 2, while in the
same region the first model gives 8.7 MSE value. In this
range the PTCDA layer has a very small absorption and the
CuPc is transparent. Consequently, the experimental spectra
are dominated by the real part of the refractive index. The
small MSE value for the second model proves that the optical constants of PTCDA and CuPc for this range are similar
to the previously determined ones 共Figs. 1 and 2兲. However,
in the 1.5– 4.5-eV range the same model reveals small deviations when compared to the experimental data thus giving
a higher value for the MSE. In this range both materials
exhibit strong absorption bands due to their corresponding
␲-␲* transitions.7–9 The interaction of the ␲ orbitals of
PTCDA with the ␲ orbitals of CuPc at the interfaces obviously seems to affect the absorption shape. This explains the
higher deviation for the model which considers sharp optical
interfaces.
Figure 6 shows the imaginary part of the in-plane dielectric function for CuPc and PTCDA compared with the imaginary part of the effective dielectric function 具␧典 of the superlattice. The effective dielectric function is calculated from
the ellipsometric experimental data as if the sample would be
an isotropic bulk.17
具␧典 = sin2 ⌽ + sin2 ⌽ tan2 ⌽
冉 冊
1−␳
1+␳
FIG. 7. Simulation of the imaginary part of the effective dielectric function
for different superlattice periods.
Figure 8 shows the IR spectra of the superlattice compared with the ones of single layers of PTCDA and CuPc. As
can be seen in the figure, the IR spectra of the superlattice
can be reconstructed using a linear superposition of the IR
spectra of PTCDA and CuPc. As the IR peak positions in the
superlattice are identical to the ones in the single layers we
can conclude that no chemical interaction occurs between
2
,
where ⌽ is the angle of incidence and ␳ = r pp / rss with r pp and
rss being the Fresnel reflection coefficients.
The optical response of such a multilayer is given by the
optical properties of each individual layer and also by the
optical interferences. As a result the optical properties of the
superlattice can be tuned within a large range using different
layer thicknesses and different periods. For example, the
imaginary part of the effective dielectric function 具␧2典 is presented in Fig. 7 from a simulation using the model 2 with
different periods for the superlattice. For clarity only 具␧2典 for
the 55° angle of incidence is presented.
FIG. 8. 共a兲 The IR spectrum of the superlattice measured in s polarization at
60° angle of incidence compared with the IR spectra of CuPc and PTCDA.
共b兲 The IR spectra of the superlattice in p polarization at 60° angle of
incidence compared with the IR spectra of CuPc and PTCDA.
Downloaded 27 May 2005 to 134.109.68.45. Redistribution subject to AIP license or copyright, see http://jap.aip.org/jap/copyright.jsp
063518-5
J. Appl. Phys. 97, 063518 共2005兲
Gordan et al.
PTCDA and CuPc. Moreover, the relative peak intensities
show that the molecular orientation remains the same for
PTCDA in the superlattice. For CuPc the analysis is more
complex as the refraction index in the IR region also has to
be taken into account. We cannot exclude the possibility that
the CuPc molecules have a slightly different orientation in
the multilayer compared to the CuPc in the single layers.
However, from the IR peak positions we can conclude that
the crystal structure of the CuPc is the same as the one in the
CuPc single layers. As CuPc exhibits small differences in the
effective values of the in-plane, respectively, out-of-plane dielectric functions just small deviations in the values of the
dielectric function for the CuPc in the superlattice are expected.
The IR spectra in Fig. 8 are normalized with respect to
the reflection of the Si共111兲 substrate. The PTCDA IR spectrum of a 70-nm layer is divided by a factor of 3. The CuPc
IR spectra were recorded for a 50-nm layer. If we consider
that the intensities of the peaks in the IR spectra are proportional to the film thickness, the thickness ratio of the layers
in the superlattice layers can be calculated. Comparing the
1594-cm−1 PTCDA peak and the 1092-cm−1 CuPc peak in s
polarization a PTCDA/ CuPc thickness ratio of 2.17 was deduced. This is in very good agreement with the thickness
ratio determined by ellipsometry.
IV. SUMMARY
The optical response of an organic superlattice consisting of five alternating layers of PTCDA and CuPc was presented. As the IR spectra are unaffected we can exclude a
chemical interaction between these two materials. The differences between the experimental data and simulated data in
the absorbing range of the materials indicate that an electronic interaction between the ␲ orbitals of the PTCDA with
the ␲ orbitals of the CuPc takes place at the interfaces. Ellipsometry is a powerful tool in evaluating the optical response of the organic superlattices with high accuracy.
ACKNOWLEDGMENT
This work was financially supported by Deutsche
Forschungsgemeinschaft, Graduiertenkolleg 829 “Akkumulation von einzelnen Molekülen zu Nanostrukturen.”
L. Esaki and R. Tsu, IBM J. Res. Dev. 14, 61 共1970兲.
F. F. So and S. R. Forrest, Phys. Rev. Lett. 66, 2649 共1991兲.
3
Y. Imanishi, S. Hattori, A. Kakuta, and S. Numata, Phys. Rev. Lett. 71,
2098 共1993兲.
4
R. W. Collins, D. E. Aspnes, and E. A. Irene, Spectroscopic Ellipsometry
共Elsevier, Lausanne, 1998兲.
5
M. Schubert, Phys. Rev. B 53, 4265 共1996兲.
6
D. W. Berreman, J. Opt. Soc. Am. 62, 502 共1972兲.
7
N. B. McKeown, Phthalocyanine Materials 共Cambrige University Press,
Cambrige, 1998兲.
8
E. Hädicke and F. Graser, Acta Crystallogr., Sect. C: Cryst. Struct.
Commun. 42, 189 共1986兲.
9
E. Hädicke and F. Graser, Acta Crystallogr., Sect. C: Cryst. Struct.
Commun. 42, 195 共1986兲.
10
R. M.A. Azzam and N. M. Bashara, Ellipsometry and Polarized Light
共Elsevier, Amsterdam, 1992兲.
11
O. Gordan, M. Friedrich, and D. R. T. Zahn, Thin Solid Films 455–456,
551 共2004兲.
12
M. Friedrich, Th. Wagner, G. Salvan, S. Park, T. U. Kampen, and D. R. T.
Zahn, Appl. Phys. A: Mater. Sci. Process. 75, 501 共2002兲.
13
M. Friedrich et al., J. Phys.: Condens. Matter 15, 2699 共2003兲.
14
A. N. Sidorov and I. P. Kotlyar, Opt. Spektrosk. 11, 175 共1961兲.
15
J. M. Assour, J. Phys. Chem. 69, 2295 共1965兲.
16
M. K. Debe, J. Vac. Sci. Technol. A 10, 2816 共1992兲.
17
G. Bauer and W. Richter, Optical Characterization of Epitaxial Semiconductor Layers 共Springer, Berlin, 1996兲.
1
2
Downloaded 27 May 2005 to 134.109.68.45. Redistribution subject to AIP license or copyright, see http://jap.aip.org/jap/copyright.jsp