Last updated April 26, 2000
(1) Technical University Munich, Physics Department E19
James-Franck-Str. 1, D-85748 Garching
(2) Johannes Kepler University Linz, Physical Chemistry
Altenberger Str. 69, A-4040 Linz, Austria
With the invention of the scanning tunneling microscope (STM) not only the
investigation of structural properties of surfaces down to atomic scale became possible. Also
new techniques were developed to study a variety of physical, e.g. electronic, properties. Using
semiconductor samples one approach is to investigate the local electronic properties by using
the magnitude of the photocurrent at a semiconductor surface. Here, a resolution of 1 nm in the
photocurrent image can be achieved [1]. The photocurrent is generated at the semiconductor
surface where a large band bending is caused by the water layer present at the semiconductor
surface in ambient air. The photogenerated charge carriers can be measured as a tunneling
current through the STM tip. Since the local magnitude of the photocurrent is influenced by the
position of the band edges at the surface, local space charge regions connected with surface
structures like steps can be detected [2]. With this method it has also been possible to directly
image the particle size dependent width of the space charge regions connected to nanosized
metal particles on a semiconductor surface [3] which is of interest e. g. for the catalytic influence
of nanosized metal particles (multiple nano contacts: MNCs) on the electron transfer in
electrochemistry or photoelectrochemistry [4]. In ambient air the contact formed by the STM tip,
the tunneling gap, and the semiconductor surface behaves like a
MIS-(metal-insulator-semiconductor) contact [5]. The most important property of this contact is
that the short circuit photocurrent is in a certain range independent of the distance between the
tunneling tip and the semiconductor surface, which means that it is not determined by tunneling
but by the transport in the semiconductor.
Although measurements in the ambient have been proven to be useful for studying the
electronic properties of semiconductor surfaces, more insight is expected from in-situ
measurements in an electrochemical cell. Here, the potential of the semiconductor and the
tunneling tip can be controlled independently of the bias of the tip. The potential of the
semiconductor determines the barrier height at the semiconductor surface and thereby e.g. also
the forward current over this barrier. The tunneling current is therefore determined by two
barriers: the tunneling barrier and the barrier at the semiconductor surface. The properties of
the contact between the tip and the semiconductor formed under these controlled conditions can
be studied by measuring the tunneling current for different potentials of the tip and the
semiconductor.
These spectroscopic measurements have been performed in sulfuric acid at a tungsten
diselenide semiconductor sample. Measurements of current-potential curves for different
tunneling gaps and different potentials of the tip and the semiconductor are recorded and results
are shown. The differences in the tunneling current measured at different tip potentials and
tunneling gaps are attributed to an electronic influence of the tip potential on the potential of the
semiconductor surface connected with a change in the barrier heights.
EXPERIMENTAL
The setup is based on a commercial STM (Nanoscope III, Digital Instruments, Santa
Barbara, CA.) which is equipped with a bipotentiostat. The electrochemical cell is made from
Teflon. A platinum wire is used as a counter electrode and as a reference electrode Ag/AgCl
wire is used. The potential of the tip and the semiconductor sample are controlled by the
bipotentiostat with respect to the reference electrode. The bias voltage for tunneling is defined
by the difference of the two potentials. All measurements have been performed in 0.01 M
H2SO4. The whole STM is working in a closed chamber where a controlled oxygen free gas
atmosphere can be maintained. As tunneling tips electrochemically etched Pt:Ir (80:20) have
been used which have been insulated with an electrophoretically applied lacquer for minimizing
the faradaic currents. As a model semiconductor n-WSe2 monocrystalline samples with doping
levels of 4.1016 cm-3 have been used. These layered semiconductor can be cleaved with
adhesive tape exposing large terraces of non-reactive van-der-Waals surfaces separated by
only few monolayer high steps.
The measurements have been performed with different tip and semiconductor potentials.
The distance between tip and semiconductor has been varied by changing the setpoint current
during feedback operation biasing the semiconductor far into the forward direction. The
measuring time per point and the time for the feedback operation are both set to 0.3 s to
establish a stationary tunneling current after the potential switch. The distance change due to
thermal drift during this period has been checked separately and is negligibly small. The current
through the STM tip during the measuring time and the potential for every set of parameters are
then measured by two external voltmeters (Keithley 404) and stored in a separate computer.
Measurements of current-potential curves for different tunneling gaps and different potentials
of the tip and the semiconductor have been recorded. The differences in the curves for different
distances can not be explained by the different tunneling transmission factors. In addition, for
small distances between the tip and the semiconductor an influence of the potential of the tip
on the barrier height in the semiconductor has been found. This can easily be understood since
the tip in all these measurements is located well inside the electrochemical double layer
consisting of a first starre Helmholtz and the more extended Gouy Chapman layer. Thereby the
tip is changing the potential change between the semiconductor and the electrolyte depending
on the tip potential. As a result the structure of the electrochemical double layer can directly be
investigated by its influence on the band edge potential of the semiconductor surface using
these spectroscopic measurements.
ACKNOWLEDGEMENTS
Financial support by the Volkswagen Foundation under contracts No. I/72 365, I/71 902 and
by the Deutsche Forschungsgemeinschaft under contracts No. ME 855/3-1 and STI 74/9-2 is
gratefully acknowledged. We are very grateful to Prof. Levy, EPFL Lausanne, Prof. Lux-Steiner,
Dr. Alonso-Vante and Y. Tomm, HMI Berlin, for providing semiconductor samples.
REFERENCES