Metal deposition on n-Si(111):H electrodes

Article abstract of DOI:10.1016/S0013-4686(00)00611-3

Pb, Cu and Au were deposited on well-defined n-Si(111):H electrodes. Pb shows a Volmer-Weber growth mode and the Pb crystallites grow epitaxially on the substrate. The Pb(111) planes are aligned parallel to the Si(111) surface and the crystallites show a preferential rotational alignment in the high-symmetry directions of the substrate. The Cu crystallites often exhibit a hexagonal shape. Au is deposited currentless from solution. The counter reactions to Au deposition is the dissolution of Si, which results in a roughening of the surface. Ex-situ-atomic force microscopy images in air reveal that Au forms 3D nuclei on n-Si(111):H with apparently no preferential nucleation sites for the deposit.

Full text of DOI:10.1016/S0013-4686(00)00611-3

Electrochimica Acta 45 (2000) 4599–4605
www.elsevier.nl/locate/electacta
Metal deposition on n-Si(111):H electrodes
a
a
b
c
a,
J.C. Ziegler , A. Reitzle , O. Bunk , J. Zegenhagen , D.M. Kolb *
a
Department of Electrochemistry, Uni6ersity of Ulm, 89069 Ulm, Germany
b
II. Institut f u¨ r Experimentalphysik, Uni6ersity of Hamburg, 22761 Hamburg, Germany
c
European Synchrotron Radiation Facility, 38043 Grenoble, France
Received 4 April 2000
Abstract
Well-defined monohydride-terminated n-Si(111) electrodes were used to study electrodeposition of Pb and Cu as
well as electroless Au deposition on these surfaces. With classical electrochemical methods, in-situ STM, ex-situ AFM,
SEM and in-situ SXRD, the initial stages of metal deposition and the epitaxial properties of the metal deposit were
investigated. © 2000 Elsevier Science Ltd. All rights reserved.
plating on semiconductor substrates [2–7] and most of
the work was done with classical electrochemical meth-
ods. In the following, results of Pb and Cu deposition
as well as electroless Au deposition on n-Si(111):H
electrodes are presented. With the combination of clas-
sical electrochemical methods and new tools for the
structure characterization of electrode surface, i.e. in-
situ STM, AFM and in-situ SXRD, new insights into
the initial stages of metal deposition on n-Si(111):H, the
growth mode and the structural properties of the thin
metal films can be gained.
1
. Introduction
The metallization of semiconductor surfaces is an
important technological process for the formation of
Schottky barriers and ohmic contacts in the fabrication
of integrated circuits. Besides the possibility of deposit-
ing from vapor phase, many metals can be deposited on
semiconductors from solution either by applying a
voltage or simply by dipping the substrate into the
plating solution (electroless deposition). There are sev-
eral important advantages of the electrochemical depo-
sition process over UHV deposition, (1), it is possible to
work under ambient conditions, therefore, this method
is cheaper; (2), by variation of overpotential and elec-
trolyte composition, the growth mechanism and the
structure of the deposit can be influenced; and (3),
electrochemical metal deposition is possible on large
surface areas and on surfaces with complex geometry.
In silicon device technology, there is a trend towards
replacing aluminum metallization by copper [1] and this
has led to an enormous interest in the electrochemical
deposition of metals on silicon. Despite its industrial
importance, there is very little work found in the litera-
ture dealing with the fundamental aspects of electro-
2. Experimental
The in-situ STM and ex-situ AFM experiments were
performed with a Pico-SPM (Molecular Imaging Cor-
poration, Tempe, USA). The STM tips were made from
an 80:20 Pt–Ir wire (0.2 mm diameter) by electrochem-
ical etching in a lamella of 3.4 M NaCN. They were
coated either with Apiezon wax [8,9] or with an elec-
trodeposition paint [10] in order to reduce the area in
contact with the electrolyte and, therefore, minimize the
faradaic current at the tip. The in-situ STM was
equipped with a bipotentiostat, hence the tip and the
sample potentials could be controlled independently
from each other. The AFM was operated in contact
*
Corresponding author.
0
013-4686/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved.
PII: S0013-4686(00)00611-3

4
600
J.C. Ziegler et al. / Electrochimica Acta 45 (2000) 4599–4605
mode, using Si cantilevers (Park Ultralever 0.6 mm)
with a spring constant of 0.31 N m , as specified by
the manufacturer.
The SXRD experiments were performed at the BW2
wiggler beamline at HASYLAB, Hamburg, which is
equipped with a four-circle diffractometer in z-axis
geometry. The X-ray energy was chosen at E=10 keV
All electrolytes were prepared from high purity chem-
icals and Milli-Q-water.
The electrochemical measurements were performed in
a standard electrochemical cell with a three-electrode
configuration. The electrolyte was constantly purged
−
1
with purified N , so that oxygen reduction was below
2
the detection limit. Reference electrodes were either a
Pt wire, a saturated calomel electrode (SCE) or a
miniaturized Ag/AgCl electrode. However, all poten-
tials are quoted with respect to SCE. In all cases, a Pt
wire served as counter electrode.
(
u=1.24 A). Electrochemical cell and Si sample holder
,
for the in-situ SXRD experiments were described in
detail in references [11,12].
In this study one-side polished n-Si(111) samples,
15
−3
phosphorous-doped, 1–20 V cm (ꢀ10 cm
doping
density) with a low miscut angle of about 0.1° were
used (Wacker Siltronic AG, Burghausen, Germany).
For the in-situ SXRD measurements of Pb electrodepo-
sition, samples with a higher doping density were used
3
. Results and discussion
In Fig. 1, the in-situ STM image of a monohydride-
terminated n-Si(111) surface, prepared as described
above, is shown. The typical features of this surface are
,
atomically flat terraces with bilayer steps (3.1 A in
(
n-Si(111), Sb-doped, 0.023 V cm, miscut, 0.6–0.7°).
The samples were first degreased for about 2 h by
refluxing in isopropanol. Then they were cleaned fol-
lowing the RCA method [13], i.e. boiling in a 1:1:5
mixture of NH , H O and H O and afterwards in a
height) and some triangular etch pits aligned along the
main crystallographic directions of the Si(111) sub-
strate. The step density is determined by the miscut of
the wafer (in the case shown in Fig. 1 the miscut is
about 0.1°). The H-termination of the surface was
stable for hours, therefore, no change in surface mor-
phology (e.g. as a result of Si-oxidation) was observed
with in-situ STM. Such well-defined surfaces were used
for all the metal deposition experiments described in the
following.
3
2
2
2
1
:1:5 mixture of HCl, H O and H O at 80°C.
2 2 2
Prior to each measurement, the Si(111) samples were
immersed for 1 min into 50% HF to remove the oxide
and then for 3 min into 40% NH F solution. The
4
procedure is known to result in well-defined monohy-
dride-terminated Si(111) surfaces with atomically flat
terraces [14–16]. Ohmic contact was established on the
back of the Si samples with a Ga/In eutectic.
3.1. Pb deposition on n-Si(111):H
For successfully imaging the n-Si(111):H surface in
acidic solution with in-situ STM a very positive tip
potential was necessary (Etip]300 mV vs. SCE) [17,18].
Hence, when trying to record the initial stages of Pb
deposition on n-Si(111):H, the following problem arises
2
+
[17]. The positive tip potential repels the Pb
ions in
solution from the area underneath the tip and, there-
fore, Pb deposition could not be observed in the
scanned area, although on the rest of the sample, Pb
deposition did take place. A similar observation was
also reported for metal deposition on metal substrates
[19]. However, there was no problem to image ‘thick’
layers of Pb, in which case Pb was deposited with the
tip retracted and the deposit imaged afterwards. Fig. 2a
shows an in-situ STM image of a Pb film with a
thickness of about 100 nm. The Pb crystallites exhibit
atomically flat terraces, which, at first approximation,
are parallel to the underlying Si(111) surface. Some of
the crystallites show a hexagonal shape with 60 and
1
20° angles. The hexagonal symmetry is also reflected
in the atomic-resolution images of the flat terraces (Fig.
b). The distance of the Pb atoms is 3.590.2 A, which
2
,
Fig. 1. STM image of a H-terminated n-Si(111) surface in 0.1
M H SO . Esample= −300 mV, Etip= +700 mV vs. SCE;
is consistent with the bulk (111) face of Pb. Further
experiments with AFM showed that as a result of the
2
4
Itunnel=0.5 nA.

J.C. Ziegler et al. / Electrochimica Acta 45 (2000) 4599–4605
4601
Fig. 3. Specular CTR for n-Si(111):H after deposition of an
equivalent of about 20 monolayers of Pb at Esample= −780
mV vs. SCE. From Ref. [12].
fer normal to the Si(111) surface is shown (this scan is
equivalent to a q–2q scan). From the intensity of the
peaks one can conclude unequivocally that Pb grows
with its (111) planes parallel to Si(111). There are also
peaks for the Pb(200) and Pb(220) orientations, but
with a much smaller intensity. This result confirms the
observation from in-situ STM measurements. From
high-resolution scans the exact q-values can be deter-
mined, which correspond exactly to the q-values one
would expect for bulk Pb. Hence the Pb crystallites
have bulk crystallographic properties. From the
FWHM, Dq, of the Pb(222) peak an average thickness
of the Pb crystallites can be estimated. With N2
:
22
q222/Dq222 (N222, number of (222)-planes in the crystal-
lites) an average thickness of about 150 Pb(111) layers
(
ꢀ40 nm height) is derived. From in-plane scans,
information on the preferred rotation of the Pb(111)
crystallites can be gained. The Pb(111) crystallites are
preferentially aligned in the symmetry directions of the
Si(111) substrate, although there are also crystallites
pointing randomly to all other directions [12]. In con-
clusion, it was possible to proof with in-situ SXRD that
Pb grows epitaxially on n-Si(111):H [12].
Fig. 2. (a) STM image of Pb crystallites deposited on n-Si(111)
from 0.05 M HClO +1 mM Pb(ClO ) . Esample= −860 mV,
4
4 2
Etip= −340 mV vs. SCE; Itunnel=2 nA. (b) STM image with
atomic resolution of a Pb crystallite on n-Si(111):H. Esample
580 mV, Etip= −550 mV vs. SCE; Itunnel=25 nA. From
Ref. [17].
=
−
weak interaction with the substrate Pb grows in a
Volmer–Weber growth mode with no obvious prefer-
ence in nucleation sites on the substrate [20]. To get
more information on the crystallographic properties of
the Pb crystallites, in-situ-SXRD was employed [12]. In
that case, an equivalent of about 20 monolayers of Pb
was deposited and probed in-situ with the X-ray beam.
In Fig. 3, the specular crystal truncation rod (CTR), i.e.
the q-scan with the q-vector, i.e. the momentum trans-
Fig. 4. Cyclic voltammograms for n-Si(111):H in 0.1 M
−1
H SO +0.1 mM CuSO . Scan rate, 20 mV s . Solid line,
2
4
4
first cycle; dashed line, second cycle. The voltammograms
were recorded in the dark.

4
602
J.C. Ziegler et al. / Electrochimica Acta 45 (2000) 4599–4605
first cycle for Cu deposition onto a freshly prepared
sample, whereas the dashed one shows the second
cycle.
In the first cycle, Cu deposition starts at around
−
500 mV versus SCE with quite a large overpoten-
tial. For lower sweep rates, it was found that Cu
deposition starts earlier; however, in all cases, it was
necessary to go negative of the flatband potential Ufb
of Si, in order to deposit Cu (U : −300 to −400
fb
mV vs. SCE [21–23]). Only for UBU , electrons ac-
fb
2
+
cumulate at the surface and can reduce Cu . No sign
for an underpo-tential deposition could be observed.
The voltammogram shows a hysteresis; in the anodic
sweep Cu deposition occurs up to potentials of −150
mV versus SCE. Obviously, a substantial overpotential
for nucleation of Cu is necessary. In contrast to Pb
deposition, Cu deposition is largely irreversible. There
is only a small anodic peak at around 0 mV versus
SCE. The reason for the irreversibility lies in the n-Si/
Cu contact, which forms a Schottky-barrier (barrier
height, :0.6 eV [7]), so that no electrons from the
Fermi level of Cu can reach the conduction band, and
without light no holes in the valence band are avail-
able. The second cycle (dashed curve in Fig. 4) is
different from the first one. Because there are still
some Cu nuclei on the electrode from the first cycle,
Cu deposition starts already at much more positive
potentials at around −150 mV versus SCE.
Fig. 5. (a) Current transient for Cu deposition on a freshly
prepared n-Si(111):H surface in 0.1 M H SO +0.1 mM
2
4
CuSO . The potential was stepped from −50 to −700 mV vs.
4
SCE. (b) Reduced-variable plot of the transient shown in (a).
The calculated curves for instantaneous (solid) and progressive
In order to determine the nucleation behavior of Cu
on n-Si(111):H, potential-step experiments were per-
formed and the resulting current transients analyzed
with the model of Scharifker and Hills [24,25] for
three-dimensional growth under diffusion control. In
Fig. 5a, the current transient for a potential step from
(dashed) nucleation are shown for comparison.
3
.2. Cu deposition on n-Si(111):H
In Fig. 4, two voltammograms for Cu deposition
on n-Si(111):H are shown. The solid curve shows the
−
50 to −700 mV versus SCE is shown. One can see
a sharp spike for double layer charging, followed by a
rising transient. The falling part of the transient can
be described for t]6 s by the Cottrell equation [26]
for diffusion to a planar electrode. From the Cottrell
−
5
2
equation a diffusion coefficient of D=1.6×10
cm
−
1
s
can be calculated. This value is in agreement with
other D values for Cu , derived from transient ex-
2
+
periment, which lie in the range between 0.5 and
−
5
2
−1
2.1×10
cm s
[7].
For analyzing the transient, the reduced-variable
2
plot [24,25] was chosen, i.e. (I/Imax) was plotted ver-
sus (t/tmax) (Fig. 5b). For comparison, the calculated
curves for instantaneous and progressive nucleation
are also shown. The experimental data fit neither to
the progressive nor the instantaneous nucleation
model. This contrasts results from recent investigations
of Cu deposition on n-Si(100), where progressive and
instantaneous nucleation was found, depending on
overpotential and electrolyte composition [7,27].
Fig. 6. SEM image of Cu crystallites on n-Si(111):H. An
equivalent of ten monolayers of Cu was deposited from 0.1 M
H SO +0.1 mM CuSO by a potential step from −250 to
2
4
4
−
450 mV vs. SCE.

J.C. Ziegler et al. / Electrochimica Acta 45 (2000) 4599–4605
.3. Au deposition on n-Si(111):H
4603
3
Different from Pb and Cu, Au deposition on n-
Si(111):H occurs electroless, i.e. just by dipping the
−
substrate into the AuCl -containing plating solution.
4
−
The reason for this is, that the Au/AuCl4 redox couple
has a very positive redox potential and hole injection
into the valence band of Si is readily possible. This is
illustrated in Fig. 7, where the positions of conduction
and valence band of Si as well as those of the redox
−
4
couples are shown. For Au/AuCl , there is already a
distinct overlap of the unoccupied Ox-states in the
−
4
electrolyte (AuCl ) with the occupied states in the
valence band, therefore, hole injection can occur. This
2
+
2+
is different for the Cu/Cu
and Pb/Pb
redox cou-
ples, which are more negative and reduction of the ions
is only possible with electrons from the conduction
band which are only available if the semiconductor is
polarized such that electrons accumulate at the
interface.
Fig. 7. Diagram for the positions of the band edges of
n-Si(111) and of the redox couples under study. The latter
2
+
2+
were calculated for c(Pb )=0.1 mM, c(Cu )=0.1 mM
−
−
and c(AuCl4 )=1 mM at c(Cl )=0.1 M.
The counter reaction of Au deposition is the oxida-
tion of Si due to the holes injected into the valence
band. Because in the experiments a fluoride-containing
plating solution was used, Si was dissolved into the
electrolyte. Assuming a Si dissolution via two holes
In order to characterize the Cu deposit, it was
imaged with SEM. In Fig. 6, an equivalent of about ten
monolayers of Cu was deposited at −450 mV versus
SCE, the electrode then removed from the electrolyte
under potential control and imaged with an SEM. One
can see Cu crystallites of different sizes (between 50 and
+
(h ) [28] the reaction can be summarized as:
−
−
+
AuCl “Au+4Cl +3h
4
7
1
5 nm diameter) with a density of nuclei of about
9
−2
.7×10 cm
as determined from SEM images on
+
2+
Si+2h “Si
larger surface areas. After the deposition of ten
monolayers, the Si surface is by far not fully covered
with Cu, showing that Cu grows in a Volmer–Weber
growth mode. Most of the crystallites show a hexagonal
The Si(111) samples were dipped into the plating
solution (0.1 M HCl+0.5 M HF+1 mM K[AuCl4])
and subsequently imaged with AFM in air. Rutherford
back-scattering experiments [29] revealed for both cases
an Au coverage below 0.01 monolayer. In Fig. 8, two
AFM images are shown after 1 s (Fig. 8a) and 5 s (Fig.
8b) exposure to the solution.
(
or at least a distorted hexagonal) shape with their
edges pointing into the same directions. In order to
determine whether Cu grows epitaxially on n-Si(111):H,
in-situ SXRD measurements are planned.
Fig. 8. AFM images of n-Si(111):H in air after electroless Au deposition from 0.1 M HCl+0.5 M HF+1 mM K[AuCl ] for 1 (a),
4
and 5 s (b).

4
604
J.C. Ziegler et al. / Electrochimica Acta 45 (2000) 4599–4605
can be calculated [32]:
i1/2
|RMS
=
dÞ
1
−i
dÞ being the distance between the Si(111) planes
d =3.14 A).
In Fig. 9, two CTRs are shown, one for a ‘flat’
(
,
Þ
Si(111) surface (squares) without Au, the other one for
a sample which had been dipped for 30 s into a
solution of 0.1 M HCl+3 M HF+0.1 mM K[AuCl4]
(
from RBS experiments [28] we know, that the Au
coverage corresponds to an equivalent of one mono-
layer). The two peaks visible at l=1/3 and 4/3 corre-
spond to the (111) and (220) bulk reflections,
(
Fig. 9. CTRs of a Si(111):H sample without any Au (filled
squares) and after 30 s exposure to 0.1 M HCl+3 M HF+
respectively. In recording the data, a hexagonal coor-
dinate system was used for convenience. A significant
increase in surface roughness after deposition of Au
can be seen, the scattering intensity for the Au-covered
surface in the region between the Bragg reflexes is
much smaller. The i values for the Au-free and the
Au-covered surface are i=0.283 and 0.732, respec-
tively, which means that the |RMS value increases
0
.1 mM K[AuCl ] (open circles). The solid and the dashed
4
curves show the best fits to the experimental data, following
the model of Robinson [31]. The CTRs are recorded in air.
The peaks at l=1/3 and 4/3 correspond to the (111
bulk Bragg reflections.
(
)and (220)
from |RMS=2.3 to 10.0 A
monolayer equivalent of Au.
,
when depositing one
In the images, one can see the typical features of the
H-terminated Si(111):H surface and some three-dimen-
sional Au nuclei. The very same 3D growth mode was
also found for electroless Au deposition on an atomi-
cally rough, only HF-etched Si(111) surface [30], and
for vapor deposition of Au on Si(111) (1×1)-H in
UHV [31]. From the AFM images, we are led to
conclude that there are no preferential nucleation sites
for Au deposition, because Au clusters can be seen
both on terraces and at steps. For a longer period of
exposure to the electrolyte, the number of Au nuclei
The Au deposit has further been characterized by
SXRD and XSW (X-ray standing wave). These results
will be presented in a forthcoming publication [33],
showing that Au grows also epitaxial on the H-termi-
nated Si(111) surface.
4. Summary
−
2
−2
Pb, Cu and Au were deposited on well-defined n-
Si(111):H electrodes. Pb shows
(
Fig. 8a, 350 mm ; Fig. 8b, 430 mm ) and the size
of the nuclei (Fig. 8a, average height 7.5 A; Fig. 8b,
average height 9.5 A) increases. The real width of the
a Volmer–Weber
,
growth mode and the Pb crystallites grow epitaxially
on the substrate. The Pb(111) planes are aligned paral-
lel to the Si(111) surface and the crystallites show a
preferential rotational alignment in the high-symmetry
directions of the substrate. Different from Pb, Cu elec-
trodeposition is irreversible. The current transients
show neither instantaneous nor progressive nucleation,
and SEM images reveal a 3D growth of Cu. The Cu
crystallites often exhibit a hexagonal shape, but their
alignment with respect to the substrate still has to be
investigated by in-situ SXRD experiments.
Au is deposited currentless from solution. The coun-
ter reaction to Au deposition is the dissolution of Si,
which results in a roughening of the surface. This can
clearly be seen in the SXRD experiments, when
recording crystal truncation rods, scattering appears to
be significantly decreased in the region between the
Bragg reflections. Ex-situ-AFM images in air reveal
that Au forms 3D nuclei on n-Si(111):H with appar-
ently no preferential nucleation sites for the deposit.
,
Au clusters, however, is difficult to determine from
AFM measurements because of the limited resolution
of the AFM tip. No significant change of the Si(111)
surface due to dissolution of Si could be observed in
the AFM images. Possibly, the changes are too small
(
less than 0.01 monolayers of Au are deposited, so
only 0.015 monolayers of Si are dissolved) to be im-
aged with AFM.
In the SXRD measurements, a roughening of the
Si surface was observed when depositing more Au.
For the characterization of the Si(111) surface mor-
phology by SXRD, the (1,0,l)-crystal truncation rods
(
CTRs) were recorded. In the region between the
Bragg-reflections at l=1/3 and 4/3, the scattered in-
tensity sensitively depends on surface roughness. By
fitting the experimental data with
a model of
Robinson [32], a i value can be derived, from which a
root-mean-square value of surface roughness |RMS

J.C. Ziegler et al. / Electrochimica Acta 45 (2000) 4599–4605
[13] W. Kern, D.A. Puotinen, RCA Rev. 31 (1970) 187.
4605
Acknowledgements
[
[
[
[
[
[
[
14] Y.J. Chabal, G.S. Higashi, K. Raghavachari, V.A. Bur-
rows, J. Vac. Sci. Technol. A7 (1989) 2104.
15] G.S. Higashi, R.S. Becker, Y.J. Chabal, A.J. Becker,
Appl. Phys. Lett. 58 (1991) 1656.
16] H.E. Hessel, A. Feltz, M. Reiter, U. Memmert, R.J.
Behm, Chem. Phys. Lett. 186 (1991) 275.
17] D.M. Kolb, R.J. Randler, R.I. Wielgosz, J.C. Ziegler,
Mater. Res. Soc. Symp. Proc. 451 (1997) 19.
18] J.C. Ziegler, G.E. Engelmann, D.M. Kolb, Z. Phys.
Chem. 208 (1999) 151.
19] U. Stimming, R. Vogel, D.M. Kolb, T. Will, J. Power
Sources 43–44 (1993) 169.
The synchrotron X-ray diffraction experiments are
supported by the German BMBF under contract
05SE8GUA5. We like to thank D. Carstanjen and D.
Plachke for the RBS measurements. One of us (A.R.)
gratefully acknowledges receipt of a stipend from the
Fonds der Chemischen Industrie.
References
[
[
[
[
[
[
[
[
[
1] P.C. Andricacos, The Electrochemical Society Interface,
Spring 1998, 23.
2] P. Bindra, H. Gerischer, D.M. Kolb, J. Electrochem. Soc.
20] J.C. Ziegler, R.I. Wielgosz, D.M. Kolb, Electrochim.
Acta 45 (1999) 827.
21] J.-N. Chazalviel, Surf. Sci. 88 (1979) 204.
22] G. Schlichth o¨ rl, L.M. Peter, J. Electrochem. Soc. 141
[
[
1
24 (1977) 1012.
3] P. Allongue, E. Souteyrand, J. Electroanal. Chem. 286
1990) 217.
4] P. Allongue, E. Souteyrand, J. Electroanal. Chem. 362
1993) 79.
(
1994) L171.
(
[
[
23] J.-N. Chazalviel, in: J.-C. Vial, J. Derien (Eds.), Porous
Silicon Science and Technology, Springer, Berlin, 1995, p.
(
1
7.
5] P. Allongue, E. Souteyrand, L. Allemand, J. Electroanal.
Chem. 362 (1993) 89.
6] G. Scherb, D.M. Kolb, J. Electroanal. Chem. 396 (1995)
24] G. Gunawardena, G. Hills, T. Montenegro, B. Scharifker,
J. Electroanal. Chem. 138 (1982) 225.
25] B. Scharifker, G. Hills, Electrochim. Acta 28 (1983) 879.
26] Southampton Electrochemistry Group, Instrumental
Methods in Electrochemistry, Ellis Horwood, Chichester,
[
[
1
51.
7] G. Oskam, J.G. Long, A. Natarajan, P.C. Searson, J.
Phys. D: Appl. Phys. 31 (1998) 1927.
8] J. Wiechers, T. Twomey, D.M. Kolb, R.J. Behm, J.
Electroanal. Chem. 248 (1988) 451.
9] L.A. Nagahara, T. Thundat, S.M. Lindsay, Rev. Sci.
Instrum. 60 (1989) 3128.
1
985.
27] A.A. Pasa, W. Schwarzacher, Phys. Stat. Sol. (a) 173
1999) 73.
28] P. Allongue, V. Kieling, H. Gerischer, Electrochim. Acta
0 (1995) 1353.
[
[
[
(
4
[
[
[
10] C.E. Bach, R.J. Nichols, W. Beckmann, H. Meyer, A.
Schulte, J.O. Besenhard, P.D. Jannakoudakis, J. Elec-
trochem. Soc. 140 (1993) 1281.
11] J. Zegenhagen, A. Kazimirov, G. Scherb, D.M. Kolb,
D.-M. Smilgies, R. Feidenhans‘l, Surf. Sci. 352–354
29] J.C. Ziegler, A. Reitzle, D.M. Kolb, J. Zegenhagen, un-
published results.
[30] C. Rossiter, I.I. Suni, Surf. Sci. 430 (1999) L553.
[31] C. Grupp, A. Taleb-Ibrahimi, Phys. Rev. B 57 (1998)
6258.
[32] I.K. Robinson, Phys. Rev. B 33 (1986) 3830.
[33] J.C. Ziegler, A. Reitzle, A. Kazimirov, O. Bunk, D.M.
Kolb, J. Zegenhagen, in preparation.
(1996) 346.
12] J.C. Ziegler, G. Scherb, O. Bunk, A. Kazimirov, L. Cao,
D.M. Kolb, R.L. Johnson, J. Zegenhagen, Surf. Sci. 452
(2000) 150.
.