Electronic energy levels of organic dyes on silicon: A photoelectron spectroscopy study of ZnPc, F16ZnPc, and ZnTPP on p-Si(111):H

Article abstract of DOI:10.1021/jp0467200

Alignment of energy levels and chemical interaction of the organic pigments ZnPc, substituted F₁₆ZnPc, and ZnTPP on H-terminated Si(111) have been deduced from valence band and core level photoelectron spectroscopy with high surface sensitivity for coverage below and in the monolayer region. Hydrogen-terminated Si(111) was prepared in UHV. The dyes have been purified by zone sublimation and were deposited at room temperature by PVD. SXPS taken at the synchrotron BESSY shows physisorption of the dye molecules; chemical shifts indicating strong interactions with the substrate have not been observed. The maxima of the HOMO photoemission of ZnPc, F₁₆ZnPc, and ZnTPP are measured at 0.45, 1.55, and 1.0 eV below the Si valence band maximum. These experimental values are compared to theoretical values and are related to the position of the optical gap, the transport gap, and the redox potentials of the ground and excited state.

Full text of DOI:10.1021/jp0467200

19398
J. Phys. Chem. B 2004, 108, 19398-19403
Electronic Energy Levels of Organic Dyes on Silicon: A Photoelectron Spectroscopy Study
of ZnPc, F16ZnPc, and ZnTPP on p-Si(111):H
†
,†
†
‡
Ulrich Weiler, Thomas Mayer,* Wolfram Jaegermann, Christian Kelting,
‡
§
§
Derck Schlettwein, Sergey Makarov, and Dieter W o1 hrle
Institute of Materials Science, Darmstadt UniVersity of Technology, D-64287 Darmstadt, Germany, Institute for
Applied Physics, Justus Liebig UniVersity Giessen, D-35392 Giessen, Germany, and Institute of Organic and
Macromolecular Chemistry, UniVersity Bremen, D-28334 Bremen, Germany
ReceiVed: July 23, 2004; In Final Form: September 10, 2004
Alignment of energy levels and chemical interaction of the organic pigments ZnPc, substituted F16ZnPc, and
ZnTPP on H-terminated Si(111) have been deduced from valence band and core level photoelectron
spectroscopy with high surface sensitivity for coverage below and in the monolayer region. Hydrogen-terminated
Si(111) was prepared in UHV. The dyes have been purified by zone sublimation and were deposited at room
temperature by PVD. SXPS taken at the synchrotron BESSY shows physisorption of the dye molecules;
chemical shifts indicating strong interactions with the substrate have not been observed. The maxima of the
HOMO photoemission of ZnPc, F16ZnPc, and ZnTPP are measured at 0.45, 1.55, and 1.0 eV below the Si
valence band maximum. These experimental values are compared to theoretical values and are related to the
position of the optical gap, the transport gap, and the redox potentials of the ground and excited state.
1
. Introduction
example, phthalocyanines, porphyrins, or perylene tetracarboxyl-
ic acid diimides allow for the adjustment to the desired optical
absorption spectrum. Phthalocyanines and porphyrins appear to
Organic semiconducting pigments exhibit interesting opto-
electronic properties as, for example, electroluminescence in
organic light-emitting diodes (OLED) and photovoltaic energy
conversion in organic solar cells (OSC). Due to their high
be the most attractive candidates for photosensitizers as they
1
absorb throughout the visible region and into the near-IR.7
,11
A
2
prerequisite for the optoelectronic integration is the knowledge
of the energy alignment of the organic frontier orbitals HOMO
and LUMO versus the silicon bands as, for example, charge
exchange may be controlled thereby. Theoretical calculations
and electrochemical measurements show that by substitution
of the ligand (unsubstituted, alkyl-substituted, Cl-, or F-
substituted) ionization energies may be varied in a broad range
while the HOMO-LUMO gap of single molecules is hardly
absorptivity, organic chromophores are also used as sensitizer
monolayers on oxide electrodes as TiO23 and ZnO
4-6
in
photoelectrochemical solar cells (PESC). In PESC, nanoporous
oxide electrodes are used to increase the effective surface
covered with a dye monolayer for light absorption. In OSC,
special measures have to be taken for effective charge separation
because the optically generated exciton is strongly bound due
to low dielectric constants of organic semiconductors. For this
1
2
7
8
influenced. Variation of the molecular ionization energy by
substitution is expected to result in a variation of the line-up of
the semiconducting organic film valence and conduction bands
versus the silicon substrate bands. Thus, a systematic engineering
of band line-up by substitution may be possible. Also, changing
the central metal ion results in changes of the ionization energies,
while the optical absorption spectrum of single molecules is
hardly influenced. Variation of the molecular HOMO-LUMO
gap is possible by the change of the molecular structure. In
Figure 1a, the molecular structures of metal-phthalocyanine
and metal-tetraphenylporphyrin are depicted; in Figure 1b,
theoretical ionization energies and affinities for ZnPc and Cl-
purpose, organic-organic or organic-inorganic bulk hetero-
junctions of donor and acceptor phases have been developed.
Another factor reducing further the efficiency of OSC is the
low conductivity of organic semiconductors.
On the other hand, Si shows exceptional good transport
properties for charge carriers but due to the indirect gap lacks
in optical absorptivity. Therefore, integration of pigment
particles into silicon may lead to composite materials which
allow one to combine dye absorptivity and luminescence with
Si charge transport. Composites of organic dyes in a micro-
crystalline silicon matrix may be developed as effective absorber
9
materials for solar cells. Efficient sensitization may be reached
1
2
and F-substituted ZnPc and ZnTPP are displayed.
with dyes that allow one to transfer the absorbed solar energy
from the excited dye molecules to Si by either injection of
electron and hole sequentially or simultaneously as exciton
without energy barriers. Recently, energy transfer from an
antenna system of dye-loaded zeolyte crystals to Si has been
demonstrated.10 A number of stable molecular structures as, for
Systematic studies of organic dye ionization energies have
1
3,14
11
been performed in the gas phase,
condensed on metals.
in solution, and on layers
1
4-16
Systematic changes of the CuPc
ionization energy due to F substitution and partial substitution
17
have been measured on gold. In this paper, we present a
systematic study on the HOMO line-up versus the valence band
of H-passivated Si(111) as a model system. Using literature
values of (i) inverse photoemission, (ii) correction values to
deduce the bulk transport gap from UPS-HOMO and inverse
photoemission LUMO positions, (iii) exciton binding energies,
*
Corresponding author. Phone: +49 6151 165532. E-mail:
mayerth@surface.tu-darmstadt.de.
†
Darmstadt University of Technology.
Justus Liebig University Giessen.
University Bremen.
‡
§
1
0.1021/jp0467200 CCC: $27.50 © 2004 American Chemical Society
Published on Web 11/20/2004

Electronic Energy Levels of Organic Dyes on Silicon
J. Phys. Chem. B, Vol. 108, No. 50, 2004 19399
Figure 1. Molecular structure of metal phthalocyanine and metal tetraphenylporphyrin (a) and ionization energies and electron affinities from
PM3-MO calculations12 (b). Substitution of the ligand RdH with, for example, Cl or F allows systematic variation of the ionization energy.
and (iv) optical absorption, estimates of the line-up of the
photoemission-inverse photoemission HOMO-LUMO gap, of
the bulk transport gap, and of the optical absorption (excitonic)
gap are deduced and compared to predictions from electro-
chemical redox potentials of the ground and excited state of
the dye molecules in solution.
120 °C) to remove inorganic and organic contaminations. The
zinc(II) complex of meso-tetraphenylporphyrin (ZnTpp) was
used as received (Aldrich). The metal complexes were purified
by zone sublimation at 10 -10 mbar: ZnPc at 370 °C, F16-
ZnPc at 350 °C, and ZnTpp at 260 °C.
-
7
-6
The dyes have been evaporated from Al O crucibles onto
2
3
the Si samples at room temperature. For each dye a separate
source was used, one source mounted to the UHV preparation
chamber at a time. ZnPc, F16ZnPc, and ZnTPP have been
evaporated at 280, 300, and 190 °C at a deposition rate of 1-10
Å/min after degassing at lower temperature (150 °C). The
temperature was measured with a thermocouple mounted to the
crucible. At room temperature, ZnPc adsorbs onto Si(111):H
in the R modification as detected with differential reflection
spectroscopy in a separate experiment not displayed here. Flash
annealing and hydrogen termination have been carried out in
one UHV chamber and dye deposition was carried out in a
separate chamber. Both chambers are integrated to the analysis
chamber. The nominal thickness of the adsorbed dye layers has
been deduced from the attenuation of the Si 2p emission
measured at 130 eV excitation energy. Exponential damping
and a mean free path of λ ) 6 Å were assumed for the Si 2p
electrons with a kinetic energy of 30 eV in the organic
2
. Experimental Section
The SXPS experiments have been performed at the undulator
beam line U49/2 of the BESSY storage ring, which provides
photons in the energy range between hν ) 90 and 1400 eV. In
addition, the used setup is equipped with a HeI UPS source
which was used for the measurement of the valence band
spectra. The spectra were obtained using a hemispherical
Phoibos 150 (SPECS) energy analyzer of the experimental
system SoLiAS (Solid/Liquid Analysis System) permanently
operated at BESSY. The emitted photoelectrons were detected
in normal emission with an acceptance angle of (9°. The base
-
10
pressure of the analysis chamber is in the lower 10
mbar
range. The Si 2p core levels were measured with 130 eV
excitation energy. Binding energies are given relative to the
Fermi edge of an Au sample. For the fitting of the core level
lines, a polynomial background was removed and mixed
Gaussian-Lorentzian line shapes were used. HOMO positions
values are given for the peak maxima, not the onset of the
emission.
21
material. Energy levels of dye molecules in the gas phase have
been calculated using the Hyperchem program. The calculation
method PM3, next lowest state, RHF approximation was
employed.
Si(111) wafer samples p-doped with 2 × 10 B atoms/cm3
16
(5-10 Ω cm) were used. A thermal oxide was etched away by
3
. Experimental Results
a 40% HF/sat. NH4F/H2O solution (1:7:70). Resistive direct flash
annealing in UHV was performed for formation of the 7 × 7
reconstruction. Hydrogen was admitted at 400 °C to the 7 × 7
reconstructed sample via a hot wire source inducing a 1 × 1
structure.18 A reverse view LEED system (Omicron) was used
for monitoring the surface reconstructions.
As at higher coverage the substrate valence band maximum
cannot be identified in the valence band spectra, it has to be
deduced from the binding energy difference between the Si 2 p
core level and the valence band maximum of the clean
2
1
H-terminated substrate using 98.8 eV for EVBM - ESi2p3/2
. A
The zinc(II) complex of phthalocyanine (ZnPc) was prepared
by the reaction of sublimed phthalonitrile and zinc(II) acetate
good fit to the Si 2p core levels of the clean substrate is achieved
with two doublets shifted by 0.23 eV assigned to the bulk
species and the H-terminated surface species at higher binding
energy. Binding energies of the Si 2p 3/2 bulk component ESi
2p 3/2 have been measured at 99.5, 99.3, and 99.45 eV for the
clean p-doped samples used for ZnPc, F16ZnPc, and ZnTPP
adsorption, respectively. The line widths of the bulk and
H-terminated surface component were 0.3 and 0.2 eV. Because
the Si 2p 3/2 binding energy can vary for extremely p- to
extremely n-doped samples between 98.8 and 99.9 eV, the
measured values correspond to mid-gap or slight n positions of
(
99,99%, Aldrich) in dry n-pentanol in the presence of 1,8-
1
9
diazabicyclo[5,4,0] undec-7-ene (DBU) under reflux. The
isolated product was washed with water and then treated with
methanol in a Soxhilet apparatus. The zinc(II) complex of
hexadecafluorophthalocyanine (F16ZnPc) was obtained by heat-
ing a 4:1 mixture of sublimed tetrafluorophthalonitrile in zinc-
(
II) acetate (99.99%, Aldrich) in a vacuum-sealed glass tube
2
0
for 1 h at 180 °C. The resulting dark blue powder was treated
in a Soxhilet apparatus with water and light petroleum (bp 100-

19400 J. Phys. Chem. B, Vol. 108, No. 50, 2004
Weiler et al.
without forming chemical bonds with the substrate surface
atoms. Similar spectra are obtained in the case of F16ZnPc,
indicating that substitution by F does not result in increased
reactivity. Also, ZnTPP was found to be physisorbed.
Si 2p spectra of samples covered with 16 Å of ZnPc, 12 Å
of F16ZnPc, and 14 Å of ZnTPP are displayed in Figure 3a; the
respective HeI valence band spectra are displayed in Figure 3b.
By referring to the energetic position of EF - (EVBM - ESi2p3/2)
)
EF - 98.8 eV in Figure 3a, the substrate valence band
maximum position below the Fermi level can be directly
deduced from the bulk peak position as indicated. Due to
adsorption of dye molecules, the binding energy of the Si 2p
core levels is changed, indicating charge transfer. This translates
to a Fermi level shift induced in the Si-surface. In the case of
ZnPc, the surface Fermi level is shifted toward the conduction
band by 150 meV and by 200 meV in the case of ZnTPP, while
F16ZnPc induces a shift toward the valence band by 100 meV.
The values are given for the respective highest measured
coverages, while in Figure 4 the development of the Si 2p 3/2
binding energy in the course of ZnPc, F16ZnPc, and ZnTpp
deposition is displayed. Clearly the electron-donating character
of ZnPc has been changed to electron-accepting character by F
substitution.
Figure 2. Si 2p core levels measured with 130 eV excitation energy
of a H-terminated substrate and samples covered with 16 Å ZnPc. Minor
changes of the line shape indicate physisorption of the dye molecules.
Similar spectra have been obtained for F16ZnPc and ZnTPP.
the surface Fermi levels, indicating that the H-termination has
not been perfect and the surface Fermi level is pinned by some
remaining Si dangling bonds.
The spectral shapes of the HeI valence band spectra of ZnPc
and F16ZnPc (Figure 3b) agree well with spectra measured in
1
4
the gas phase and on Au. A polynomial background has been
subtracted from the displayed valence band spectra, taking
account of the secondary electrons. The respective HOMO
maxima are found at 1.15, 2.05, and 1.65 eV below the Fermi
level. Taking the valence band maximum at bulk Si 2p 3/2 as
98.8 eV and the HOMO position from the valence band spectra
displayed in Figure 3b, the Si(111):H-HOMO offsets as
measured with photoelectron spectroscopy for ZnPc, F16ZnPc,
and ZnTPP are 0.45, 1.55, and 1.0 eV, respectively.
In the course of dye molecule deposition, the substrate
emission is increasingly damped. Only minor changes occurred
in the line shapes of the substrate Si 2p emission as demonstrated
for ZnPc in Figure 2. The Si 2p core level emission of the freshly
H-terminated substrate and the substrate covered with a nominal
coverage of 16 Å of ZnPc are displayed. The spectra have been
normalized in intensity and in energy to the fresh sample
emission. For the Si 2p emission of the ZnPc covered sample,
a good fit is still achieved with the same components; only the
line width of the surface component has been increased by 100
meV. The absence of chemically shifted components is a clear
indication that the deposited dye molecules are physisorbed
4
. Discussion
The main objective of this contribution is the experimental
determination of the alignment of energy levels and the
Figure 3. (a) Si 2p core levels of H-terminated Si(111) covered with 16 Å of ZnPc, 12 Å of F16ZnPc, and 14 Å of ZnTPP. The energetic difference
of the Si 2p 3/2 bulk component to the mark at E
F
- (EVBM - ESi2p3/2) ) E - 98.8 eV gives a direct reading of the valence band maximum position
F
below the Fermi level as indicated. (b) He I valence band spectra of ZnPc, F16ZnPc, and ZnTPP on Si(111):H. The HOMO peak positions are
indicated. The secondary electron background has been removed.

Electronic Energy Levels of Organic Dyes on Silicon
J. Phys. Chem. B, Vol. 108, No. 50, 2004 19401
depicted the line-up of the UPS-IPES gap EUPS/IPES, the
transport gap Et, and the optical absorption HOMO-LUMO
gap Eopt of ZnPc, F16ZnPc, and ZnTPP. For lack of detailed
UPS/IPES
data, Et and Eopt are placed symmetrically versus E
,
assuming the same correction values for deducing bulk HOMO/
LUMO positions from UPS/IPES for electrons and holes23 and
the same stabilization energy for each exciton partner due to
coulomb interaction. While the Q-band absorption of dissolved
ZnPc and F16ZnPc molecules is narrow (fwhm ) 20 meV) with
a maximum at almost the same energy (1.82 and 1.81 eV), the
absorption of deposited films is broad (fwhm ) 450 and 670
meV) with two maxima at 1.75 and 2.00 eV for ZnPc and 1.52
Figure 4. Si 2p 3/2 binding energy in the course of ZnPc, ZnTpp,
and F16ZnPc adsorption. Increasing binding energy corresponds to
raising of the substrate surface Fermi level due to electron donation
by the adsorbate. By F substitution, the electron-donating character of
ZnPc has been changed to electron-accepting character.
20
and 1.90 eV for F16ZnPc. Because due to instrumental
broadening and averaging over different surroundings the line
width of UPS and especially IPES is broad as compared to
optical spectra, we use in Figure 5 a mean value for the optical
absorption gap of 1.8 eV for both phthalocyanine films, ignoring
any details of the absorption spectra. Similarly, we use an optical
absorption gap of 2.2 eV for ZnTPP films as measured for
comparison to calculated ionization and affinity potentials and
electrochemical redox potentials. After the HOMO positions
have been experimentally determined, we can deduce the LUMO
positions using results from inverse photoemission spectroscopy
on metallic samples. Yet the HOMO-LUMO gap as measured
with surface sensitive UPS and inverse photoemission IPES on
organic films appears larger than the electronically relevant
transport gap Et. A correction value for deducing bulk HOMO/
LUMO positions from UPS/IPES has to be applied. Following
the arguments given by Hill et al.,23 the difference is due to the
smaller polarization at the surface versus the bulk (approximately
2
4
dissolved molecules. The phthalocyanine UPS-HOMO peak
offsets relative to the Si VB show a 1.1 eV shift to higher
binding energy due to F substitution. This value agrees well
with the PM3-MO calculated ionization shift of 1.07 eV.1
While in the gas phase 0.9 eV and on gold a value of 0.56 eV
for the stabilization of the HOMO due to F substitution was
measured with UPS, a difference of 0.7 eV was found for the
2,14
redox potential of the first oxidation of ZnPc in DMF and F -
16
1
4
ZnPc in CH2Cl2 solution. As in the MO calculations, the
UPS-HOMO position of adsorbed ZnTPP is found below the
HOMO of ZnPc. Yet while the experimental difference is 0.55
eV, the calculated value for the free molecules is 0.1 eV only.
UPS
0
.6 eV) and to vibrational excitation in UPS/IPES which tend
to shift both of the measured HOMO and LUMO peaks away
from the Fermi level to the Frank Condon maxima (ap-
proximately 0.2 eV).23 For comparison with electrochemically
and photoelectrochemically determined redox potentials of the
ground and excited state of the dye molecules in solution, the
With the experimental ZnPc HOMO
below the VBM, we find the LUMO
value of 0.45 eV
IPES
at 1.45 eV above the
UPS/IPES
25
CBM by using E
) 3.0 eV as measured on gold. With
UPS/IPES
positions of the transport gap Et and the optical (excitonic)
Eopt ) 1.8 eV and placed symmetrical to E
the HOMO is found 0.15 eV above the VBM and the
LUMO is 0.85 eV above the CBM. Using an exciton binding
in Figure 5,
UPS/IPES
opt
absorption gap Eopt are important. While E
and Et are
opt
given by states of positive and negative ions, Eopt is given by
states of the neutral molecule. In Figure 5, we have tentatively
23
energy value of 0.6 eV as determined for CuPc, we tentatively
Figure 5. HOMO position of ZnPc, F16ZnPc, and ZnTPP on Si(111):H as determined experimentally (hatched HOMO peak). In addition, the ZnPc
23
LUMO is sketched, using the HOMO-LUMO gap as measured with photoemission and inverse photoemission on gold. The bulk transport gap
and the optical (excitonic) gap are indicated, where 0.3 eV both for electron and for hole relaxation and a 0.6 eV exciton binding energy have been
2
5
20
24
taken as determined for CuPc. Using the optical gap of F16ZnPc and ZnTpp and the same values for exciton binding energy and correction
UPS/IPES
values for deducing bulk HOMO/LUMO positions from UPS/IPES as for ZnPc, the HOMO-LUMO gaps E
, E
t
, and Eopt have been sketched
UPS/IPES
for these dyes also. While E
and E are given by states of positive and negative ions, Eopt is given by states of the neutral molecule.
t

19402 J. Phys. Chem. B, Vol. 108, No. 50, 2004
Weiler et al.
Figure 6. Band line-up as predicted from electrochemical measurements of dye oxidation potentials of ground and excited state versus electron
affinity and ionization potentials of Si(111):H.
t
place the transport gap Et and find the HOMO at 0.15 eV below
gaps referring to the energy band positions of H-terminated
silicon. Considering energetic barriers from the obtained line-
up, electron and hole injection from the dye to the semiconductor
is energetically favored for ZnPc and ZnTPP, while electron
injection from the Si conduction band to F16ZnPc is favored.
For ZnTPP, also exciton injection from the dye to the
semiconductor is barrier-free, while for ZnPc, the exciton bound
hole has to overcome a small barrier of 0.15 eV.
t
the VBM and the LUMO at 1.15 eV above the CBM. For lack
of specific correction values for deducing bulk HOMO/LUMO
positions from UPS/IPES for electrons and holes and exciton
binding energies in the literature, we use the same values we
used for ZnPc also for F16ZnPc and ZnTPP. The obtained line-
UPS/IPES
up positions of E
, Et, and Eopt are summarized in Figure
5
.
For comparison, we show in Figure 6 a prediction of the line-
For organic semiconductor materials, the position of the
HOMO and LUMO depends strongly on the respective environ-
ment (bulk, surface, interface, solution, gas phase). Different
measurement methods (UPS/IPES, I-V, electrochemical) will
result in different values. Specific measurements and consid-
erations will be needed to conclude from experimentally
determined values on the specific HOMO/LUMO positions in
a specific application. It is evident that a more detailed
experimental determination of the dye energy level positions
and their alignment to the substrate levels is needed for
systematically changed environments (adsorbed monomolecular
dyes, dye films, dyes in solvation shells).
up as given by electrochemical redox potentials of the ground
and excited singlet state and Si(111):H affinity and ionization
potential on a common vacuum scale using 4.5 eV for adjusting
the NHE redox scale to vacuum. The oxidation potentials of
the ground and excited states Vox and Vox* of ZnPc have been
measured at +0.92 and -0.91 V, and those of ZnTPP have
11
been measured at +0.95 and -1.1 V. For F16ZnPc, only Vox
is known at +1.6 V.26 In Figure 6, we used the same difference
of 1.83 V between Vox and Vox* for F16ZnPc as is known for
ZnPc. Assuming that in Vox the stabilization due to solvation is
included and assuming that this stabilization is similar to the
extramolecular polarization in the organic film, it is expected
t
that Vox coincides with HOMO . This coincidence is nearly
Acknowledgment. This contribution was conducted in the
framework of a cooperation project funded by the Volkswagen-
Stiftung under contract numbers I/78203, 78204, and 78205.
Financial support of BESSY beam time by the German Ministry
of Education and Research (BMBF) under contract number
05ES3XBA/5 is gratefully acknowledged.
found for ZnPc, but for F16ZnPc and ZnTPP our deduced
t
HOMO is found at an approximately 0.3 eV lower position.
The reason may be found in the many estimates and approxima-
t
tions we had to make to deduce the HOMO energies and, in
addition, specific interface dipoles due to electrochemical double
layers not considered in Figure 6.
References and Notes
5
. Summary and Conclusion
(
(
1) Hung, L. S.; Chen, C. H. Mater. Sci. Eng., R 2002, 39, 143.
2) Peumans, P.; Yakimov, A.; Forrest, S. R. J. Appl. Phys. 2003, 93,
Three dyes, ZnPc, F16ZnPc, and ZnTPP, have been deposited
3
693.
in UHV onto UHV prepared p-Si(111):H at room temperature
and analyzed with 130 eV PES and He I UPS. Si(111):H has
proven to be an ideal nonreactive substrate as indicated by
almost nonaffected line shapes of the Si 2p core level emission.
While ZnPc and ZnTPP show electron-donating character
moving the pinning position of the substrate by 150 and 200
meV toward the conduction band, F16ZnPc shows electron-
accepting character moving the pinning position by 100 meV
toward the valence band. From the measurements, the UPS-
HOMO maximum line-up can directly be found at 0.45, 1.55,
and 1.0 eV below the Si valence band. Using estimates from
literature values for the optical gaps, the exciton binding
energies, and correction values for deducing bulk HOMO/
LUMO positions from UPS/IPES energies, we conclude on the
positions of the optical gap and the electron and hole transport
(
3) Oregan, B.; Gratzel, M. Nature 1991, 353, 737.
(4) Fessenden, R. W.; Kamat, P. V. J. Phys. Chem. 1995, 99,
12902.
(5) Oekermann, T.; Yoshida, T.; Schlettwein, D.; Sugiura, T.; Minoura,
H. Phys. Chem. Chem. Phys. 2001, 3, 3387.
(
6) Yoshida, T.; Iwaya, M.; Ando, H.; Oekermann, T.; Nonomura, K.;
Schlettwein, D.; Wohrle, D.; Minoura, H. Chem. Commun. 2004,
400.
(
7) Wohrle, D.; Kreienhoop, L.; Schnurpfeil, G.; Elbe, J.; Tennigkeit,
B.; Hiller, S.; Schlettwein, D. J. Mater. Chem. 1995, 5, 1819.
8) Trombach, N.; Tada, H.; Hiller, S.; Schlettwein, D.; Wohrle, D.
(
Thin Solid Films 2001, 396, 109.
(9) Mayer, T.; Weiler, U.; Jaegermann, W., submitted.
10) Huber, S.; Calzaferri, G. ChemPhysChem 2004, 5, 239.
11) Darwent, J. R.; Douglas, P.; Harriman, A.; Porter, G.; Richoux,
(
(
M.-C. Coord. Chem. ReV. 1982, 44, 83.
12) Schnurpfeil, G.; Sobbi, A. K.; Spiller, W.; Kliesch, H.; W o¨ hrle, D.
J. Porphyrines Phthalocyanines 1997, 1, 159.
(

Electronic Energy Levels of Organic Dyes on Silicon
J. Phys. Chem. B, Vol. 108, No. 50, 2004 19403
13) Berkowitz, J. J. Chem. Phys. 1979, 70, 2819.
(21) Zangwill, A. Physics at Interfaces; Cambridge University Press:
Cambridge, 1987.
W.; Armstrong, N. R. J. Phys. Chem. B 2001, 105, 4791.
(22) Chiang, T. C. Band structure and core levels of tetrahedrally bonded
semiconductors. Landolt-B o¨ rnsteinsGroup III Condensed Matter; Springer-
Verlag: Heidelberg, 1989; Vol. 23A.
(
15) Koch, E.; Grobman, W. D. J. Chem. Phys. 1997, 67, 837.
16) Peisert, H.; Knupfer, M.; Schwieger, T.; Auerhammer, J. M.;
(
Golden, M. S.; Fink, J. J. Appl. Phys. 2002, 91, 4872.
17) Peisert, H.; Knupfer, M.; Schwieger, T.; Fuentes, G. G.; Olligs,
D.; Fink, J.; Schmidt, T. J. Appl. Phys. 2003, 93, 9683.
(23) Hill, I. G.; Kahn, A.; Soos, Z. G.; Pascal, R. A. Chem. Phys. Lett.
2000, 327, 181.
(24) Harriman, A.; Porter, G.; Searle, N. J. Chem. Soc., Faraday Trans.
2 1979, 75, 1515.
(25) Kahn, A.; Koch, N.; Gao, W. Y. J. Polym. Sci., Part B: Polym.
Phys. 2003, 41, 2529.
(
(
(
18) Grupp, C.; Taleb-Ibrahimi, A. Phys. ReV. B 1998, 57, 6258.
19) Wohrle, D.; Schnurpfeil, G.; Knothe, G. Dyes Pigm. 1992, 18,
9
1.
(
20) Hiller, S.; Schlettwein, D.; Armstrong, N. R.; Wohrle, D. J. Mater.
(26) Schlettwein, D. In Supamolecular PhotosensitiVe and ElectroactiVe
Materials; Nalwa, H. S., Ed.; Academic Press: New York, 2001.
Chem. 1998, 8, 945.