Electroluminescence from Individual Pentacene Nanocrystals
Alexander Kabakchiev,*[a]Klaus Kuhnke,[a]Theresa Lutz,[a]and Klaus Kern[a, b]
Pentacene is a promising candidate for use in organic and op-
toelectronic devices. The electronic properties and structures
of the organic semiconductor have been extensively studied
for thin films and macroscopic crystals. Herein, we report the
first electroluminescence measurements from single pentacene
nanocrystals using the tip of a scanning tunneling microscope
(STM) as a local electrode. For localized charge injection by the
STM tip, strongly red-shifted luminescence at the bulk exciton
energies is observed. The emission from delocalized excitations
and missing features from individual molecules reflects signifi-
cant intermolecular coupling. Excitation and emission mecha-
nism are discussed based on the observed dependence on
electrical and structural parameters.
Organic semiconductors have made a strong entrance in
nanoscale electronics and optoelectronics during the last
decade. Pentacene has become the prime model system to ex-
plore the physical principles and the technological challenges
associated with organic devices.The extended conjugation
and a favorable crystal structure are responsible for its success
as organic semiconductor. On the single-molecule level penta-
cene has also become the gold standard of local scanning
probe imaging and spectroscopy. Impressive results through
the use of STM[2,3]and atomic force microscopy (AFM)dem-
onstrate submolecular imaging capabilities and realize experi-
mental conditions in which the single organic molecule pre-
serves its structural and electronic features. Gas-phase-like
electronic orbitals can, for example, be observed on a molecule
adsorbed on an insulator. On the other hand, these studies
show that the position and width of electronic levels are deter-
mined by their environment. Making the step from the isolated
molecule to a small isolated molecular solid opens up a way to
address interactions between identical molecules. Molecular
crystals are bound together by weak van der Waals forces and
can also preserve properties of their molecular building blocks
to a large extent. They can be described in a first approxima-
tion by an “oriented gas” model.Essential for this study, how-
ever, is the fact that crystalline pentacene exhibits an intermo-
lecular coupling which is comparatively strong for this class of
materials. The resulting electronic dispersion becomes largest
along the direction of overlap between p orbitals of neighbor-
ing molecules.[6–9]Pentacene is thus a material well-suited to
explore consequences of intermolecular coupling on the opti-
cal emission from the low energy electronic excitations.
Herein we employ scanning tunneling microscopy and spec-
troscopy at 4.2 K to study charge carrier excitation and electro-
luminescence of individual pentacene molecules and nanocrys-
tals with submolecular resolution. STM-induced luminescence
has proven its ability to obtain optical spectra from isolated
molecules[10,11]and from thin molecular films.[12–17]We report
on the first STM-induced luminescence from an acene. The
method allows the local characterization of molecular struc-
ture, orientation, and the identification of the excited species
according to their optical spectra. In combination with elec-
tronic spectroscopy and measurements of the electrical param-
eters for luminescence it can provide a detailed scheme of ex-
citation and emission processes.The STM tip injects charges
locally into the molecular top layer. This suggests that lumines-
cence may predominantly be emitted by a single molecule sit-
uated below the tip. This has, in fact, been confirmed in earlier
solids.[13,14,17,19]In contrast, herein the emission line is strongly
red-shifted with respect to the emission of matrix-embedded
single pentacene molecules. This indicates strong intermolecu-
lar coupling and emission from a delocalized excitation. As a
result, the observed luminescence spectrum is very robust and
becomes independent of the point of charge injection. The
pentacene nanocrystal acts already as a molecular crystal, cou-
pled to its top and bottom electrodes through tunnel junc-
tions instead of Schottky contacts. It represents a bulk-like
light source at the limit of small dimensions.
Samples were prepared and studied in ultrahigh vacuum
(UHV). As substrates we used noble metal single crystals. The
crystal surfaces of Au(111), Ag(111) and Cu(111) were cleaned
in situ by Ar+sputtering and annealing. Subsequently, 1–2 ML
(monolayer) of KCl were deposited by thermal evaporation.
Pentacene nanocrystals were grown on top of these substrates
in a two-step process. First, highly purified pentacene was
thermally evaporated on the sample held at a low temperature
(T<90 K). The evaporation rate was low to yield a coverage of
about 1 ML during 5 min evaporation time. The result is a dis-
persed arrangement of pentacene molecules adsorbed flat on
the KCl layer. Subsequently, the substrate was annealed to
220 K for 1 min and transferred into the STM operated at 4.2 K.
The annealing step leads to pentacene dewetting on the KCl
layer and to the formation of molecularly ordered three-dimen-
sional pentacene crystallites a few nanometers high with later-
al dimensions of a few tens of nanometers. No pentacene re-
mains in between the nanocrystals. We find that the molecular
nanocrystals nucleate at step edges, which are either due to a
step in the KCl layer or a step of the metal substrate covered
in a carpet-like fashion by the KCl layer. As STM tips we em-
[a] A. Kabakchiev, Dr. K. Kuhnke, T. Lutz, Prof. Dr. K. Kern
Max-Planck Institut f?r Festkçrperforschung
Heisenbergstr. 1, 70569 Stuttgart, (Germany)
[b] Prof. Dr. K. Kern
Institut de Physique de la Mati?re Condens?e
Ecole Polytechnique F?d?rale de Lausanne (EPFL)
1015 Lausanne (Switzerland)
? 2010 Wiley-VCH Verlag GmbH&Co. KGaA, WeinheimChemPhysChem 2010, 11, 3412–3416
ployed electrochemically etched goldand tungsten tips. Ag
tips were prepared inside the UHV chamber by coating of
etched tungsten tips with a thin Ag layer. Prior to the measure-
ments on pentacene the tips were conditioned on a clean
metal substrate and checked for their spectroscopic and imag-
ing performance. Luminescence from the tunnel junction is
collected by a N.A.=0.41 aspheric lens placed a few millime-
ters away from the STM tip. The light propagates as a free col-
limated beam inside the vacuum and exits the UHV chamber
through a glass window. The setup avoids any significant heat
input in the opposite direction from the room-temperature en-
vironment. Spectra were recorded by an Acton Research Spec-
tra Pro 300i Spectrograph using a 150 linesinch?1blazed gra-
ting and a Peltier-cooled CCD camera behind a light intensifier.
Individual pentacene nanocrystals 3–15 layers thick with lat-
eral dimensions of a few tens of namometres were grown on
top of an ultra-thin KCl layer on a metal substrate. When not
stated otherwise herein, the substrate is a single-crystal
Au(111) surface. The KCl layer acts as a tunnel barrier because
its large bulk band gap is higher than 7 eVand its high die-
lectric constant provides only a small voltage drop in the exter-
nal electric field of the tunnel junction. The vacuum gap be-
tween STM tip and crystal surface defines a second (tunable)
tunnel barrier. In this geometry (Figure 1a) we obtained sub-
molecular resolution of pentacene molecules on top of the
crystal and in some cases also at the steep sides of the crystals.
The nanocrystals exhibit a high molecular order with transla-
tional symmetry in which structural defects are sometimes
found. The topography at the nanocrystal surface and the
measured crystal heights allow us to determine the nanocrystal
structure (Figures 1c,d). The lattice parameters with respect to
the surface are listed in Table 1. At negative bias voltage the
surface molecules are imaged through their highest occupied
molecular orbital (HOMO), identified by its five nodes of the
electronic wave function parallel to the intermediate molecular
axis (Figure 1d). For comparison we show the STM topograph
of the HOMO orbital of an isolated pentacene molecule ad-
sorbed flat on the KCl layer (Figure 1b). The pentacene mole-
cules in the nanocrystal layers are orientated with their long
axis parallel to the surface and the molecular plane rotated by
Figure 1. Experiment and pentacene nanocrystal structure. a) Principle of the experiment: The tip of an STM is positioned above a pentacene nanocrystal (col-
ored black–yellow) for imaging and tunneling spectroscopy. A tunnel current is injected into the nanocrystal decoupled from the Au(111) substrate by an ul-
trathin KCl layer. The emitted luminescence light is detected by an optical spectrometer. b) Chemical structure (bottom) of pentacene and its HOMO orbital
(top) imaged by STM for a molecule adsorbed on the KCl layer on Au(111) (U=?2.4 V, I=10 pA). c) STM topographic image in 3D presentation of a penta-
cene nanocrystal on KCl/Cu(111) and detail (right) in top view. The in-plane lattice vectors a and b are drawn in the topograph. The layer height is indicated
in the height profile (green line). d) Detail of the nanocrystal with two visible layers and an admolecule. e) Model of a structure similar to (d) using a cartoon-
like presentation of the HOMO orbitals with different signs of the electronic wavefunction indicated in blue and red. Left: Top view of a flat-lying single mole-
cule. The green arrows indicate the orientation of the transition-dipole vector for an excitation derived from transitions between LUMO and HOMO orbital.
The measurements are presented using the WSxM software.
ChemPhysChem 2010, 11, 3412–3416 ? 2010 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim
approximately 268 with respect to the surface plane. This rota-
tion is evidenced by the fact that the node of the HOMO orbi-
tal along the long axis of the molecule is not seen on the
closed layer, but only for admolecules and step edges (Figur-
es 1c,d). The cartoon in Figure 1e, which models a structure
similar to that shown in Figure 1d, illustrates that the rotation
leads to a short a vector between neighboring molecules and
hides the longitudinal node of the orbital from observation by
the STM tip. We remark that the imaging STM tip is separated
from the molecular top layer through a vacuum tunnel gap of
the order of 0.5–1.0 nm and thus only the lobes of the elec-
tronic wave function extending into the vacuum become ob-
servable. Within experimental accuracy (Figure 1e) the nano-
crystals realize the pentacene bulk lattice structure.The
strong local variations do, however, not allow one to distin-
guish between two distinct lattices obtained by growth from
solution or growth by vapor deposition.The nanocrystal is
markedly different from known structures of thin films on insu-
lators on which layers are formed by molecules standing up-
spectra are plotted in Figure 2.
The upper row presents lumines-
cence spectra recorded with the
STM tip positioned on a penta-
cene nanocrystal for different tip
single-crystal surfaces of differ-
ent metals. The dominating peak
in these spectra is situated at
1.6 eV. Spectra recorded next to
these nanocrystals with the STM
tip positioned above the ultra-
thinKCl insulatorlayer are
shown for comparison in the
lower row. The features in these
spectra are due to tip-induced
metallic tip and metallic sub-
strate and depend on tip shape
and tip and substrate material.
We emphasize that the two cor-
responding spectra ineach
column exhibit no similar fea-
tures. Moreover, we find that the
nanocrystal spectra in the upper
row are well reproduced and the peak positions do not
depend on the underlying plasmonic spectrum. We conclude
that STM-induced luminescence allows to access optical spec-
tra originating from the pentacene nanocrystals. Comparing
the three typical spectra in the upper row of Figure 2 to penta-
cene luminescence spectra in the literature, we find agreement
only with photoluminescence spectra from macroscopic single
crystals measured at 8 K.There is no match with the lowest
singlet transition of individual pentacene molecules, which is
situated at 2.30 eV for the isolated molecule,[27,28]red-shifted
by dielectric screening to 2.03 eV when embedded in p-ter-
phenyland to 1.98 eV for embedding in tetracene.Penta-
cene electroluminescence spectra recorded from organic field-
effect transistor devices exhibit typically broad spectral fea-
tures extending to energies above 2 eV.The low tempera-
ture in our study provides sharp features, and the geometry of
charge transport along the shortest crystal dimension sup-
presses the influence of domain boundaries with respect to
thin-film geometries. Optically induced luminescence from or-
ganic crystals probes bulk properties with surface contribu-
tions being negligible. In contrast, nanocrystals provide a high
surface-to-volume ratio. The agreement of our data with
single-crystal photoluminescence is thus surprisingly good. It
concerns peak energy, peak width, peak asymmetries and the
relative peak heights. Apart from the strong peak at 1.59 eV, a
shoulder around 1.45 eV is found. In addition, a small peak at
1.76 eV can be identified which only sometimes exceeds the
noise level in our spectra. Following the discussions in
refs. [26,32] we can assign the observed features to the free
singlet exciton (1.78 eV), the self-trapped exciton (1.64 eV), and
an impurity related line (ca. 1.4–1.5 eV). The dominance of the
Table 1. Structure parameters of the pentacene nanocrystal. Note that
the value c corresponds to twice the measured layer height as successive
layers have different molecular orientations.
Lattice Vector lengths
angle (long mol.axis,b)=
Figure 2. STM-induced electroluminescence. Top: Drawing of the experiments shown in the color-coded spectra
below. Upper row of spectra: STM-induced luminescence recorded for the STM tip positioned on the pentacene
nanocrystals with the listed crystal height, spectral measuring time and tunneling parameters: bias voltage (U)
and current (I). Lower row of spectra: STM-induced luminescence due to tip-induced plasmons recorded with the
STM tip positioned on the KCl-covered single-crystal surface.
? 2010 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim ChemPhysChem 2010, 11, 3412–3416
1.6 eV peak can be regarded as characteristic for crystals in
contrast to the 1.8 eV peak dominating for pentacene clusters
and thin films.Pentacene-related luminescence could only
be observed at negative bias voltage and for crystallites of a
minimum thickness of 1.5 nm. As the reduced dimension of
the crystallites suggest a spatial exciton confinement we ana-
lyzed the peak position of the 1.6 eV line with respect to crys-
tal width and height. A shift of peak energy could, however,
not be identified within experimental accuracy and is estimat-
ed to be smaller than ca. 10 meV over the range of 1.5–4.0 nm
crystal thickness. While a theoretical estimate on the confine-
ment is not available, the assignment to a self-trapped Frenkel
exciton may rationalize the small effect.
A closer inspection of the features in the STM-induced lumi-
nescence spectra finds consistent red-shifts of approximately
40 meV with respect to a 8 K photoluminescence spectrum.
We find that a significant part of the shift is related to the elec-
tric field between tip and sample which is of the order
1 Vnm?1. The shift of the peak energy of the strongest spectral
feature between ?3 V and ?2 V bias voltage reaches almost
10 meV. This shift is remarkable and challenges the assignment
of the feature to a pure Frenkel-type exciton. A detailed discus-
sion of the Stark shift is, however, beyond the scope of this
paper and will be presented in detail in a forthcoming publica-
The dependence of luminescence intensity of the 1.6 eV
peak on the electric parameters is plotted in Figures 3b,c. The
emission is linear in tunnel current and shows an onset near
1.8 V bias voltage. The small energetic difference between
onset and main emission line indicates only small losses before
exciton formation. An onset even below the bulk transport
threshold of pentacene (2.2 eV) suggests the creation of ex-
citons close to the position of charge injection. Electron tun-
neling spectra measured on single pentacene molecules on
the KCl layer (Figure 3a) yield charge injection potentials of
?2.4 eV for the HOMO and +2.0 eV for the LUMO. The current
between the HOMO and LUMO onsets vanishes already for
pentacene in the first layer on the KCl buffer. This demon-
strates the decoupling of the electronic states of metal and
molecule. Independent of this decoupling the HOMO and
LUMO levels shift only weakly with bias voltage and changing
tip–molecule separationas the KCL layer is thin and exhibits
a high dielectric constant. The nanocrystal tunneling spectra
(Figure 3a) are offset with respect to single molecule spectra
by about 1 eV to more positive voltage. This shift cannot be
due to the STM-related electric field inside the pentacene be-
cause the shift is towards higher energies also at negative bias.
We attribute the offset to band formation in the crystal and a
reduction of work function.
Based on the results presented, we propose a scenario lead-
ing from charge injection to luminescence in the pentacene
nanocrystals: The observed onset bias of luminescence at
?1.8 eV (Figure 3c) suggests that after extraction of an elec-
tron by the tip, an exciton (1.78 eV) is formed. A weak signa-
ture of its decay is observed in the luminescence. At higher
bias voltage additional channels may open up. Following the
assignment of transitions in the literaturethe 1.78 eV exciton
is mobile within the crystal. By self-trapping the exciton be-
comes immobilized, losing about 0.2 eV. This exciton provides
the dominant signal. Figure 3d sketches an energy diagram of
the electronic levels involved in the electroluminescence near
its onset. The Au electrodes inject charge carriers through their
respective tunnel junctions. The electrical conductance re-
quired by STM operation is provided by the valence band. A
bulk-like conduction band cannot line up with the substrate
Fermi energy. Due to the low temperature in the experiment,
the electrons are injected through an extended tunnel barrier
which may favor a direct exciton formation from the continu-
ous current of holes in the valence band. The substrate Fermi
energy is closer to the conduction band edge than to the va-
Figure 3. Electronic properties. a) Tunnel current as a function of bias voltage [Au(111) substrate with respect to STM Au tip] measured on an isolated mole-
cule (blue) and on a nanocrystal (yellow). b) Integrated STM luminescence intensity from a nanocrystal as a function of tunnel current for constant bias volt-
age. c) Integrated-STM-induced luminescence as a function of bias voltage for constant tunnel current, luminescence collection time per point: 250 s. d) Elec-
tronic level diagram for the charge-injection geometry employed for the measurements with the bias voltage near the luminescence onset: Vacuum potential
(top line), electronic levels of the metals and pentacene band structure. The two tunnel barriers are symbolized by the gray (KCl layer) and the white
ChemPhysChem 2010, 11, 3412–3416 ? 2010 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim
lence band edge (Figure 3d). This and the differences of elec-
tron and hole mobilities can account for the fact that light
emission is observed only at negative bias.
An important aspect for the understanding of the lumines-
cence mechanism is the absence of observable molecular lumi-
nescence from isolated pentacene molecules on the insulating
layer. Only for nanocrystals reaching a thickness of 1.5 nm
STM-induced intrinsic luminescence from pentacene was
strong enough to become detectable. From the position of
electronic states of the isolated molecule (Figure 3a) the crea-
tion of electron–hole pairs should also be possible for hole in-
jection into the isolated molecule if we assume its HOMO–
LUMO transition at or below 2.1 eV. Two mechanisms may be
responsible for the different behavior of crystal and single mol-
ecule: First, the generation of luminescence inside the nano-
crystal can provide a relatively large spatial separation of the
decaying exciton from the adjacent metal electrodes. Emission
close to metal surfaces is known to lead to efficient nonradia-
tive quenching for example, due to electron–hole pair genera-
tion in the metal (ref.  and references therein). Second,
single molecules adsorb flat on the surface while pentacene
molecules in the crystal are rotated around their long axis. The
transition dipole of HOMO–LUMO-derived transitions is orient-
ed along the intermediate molecular principal axis as indicated
by the double-headed green arrows in Figure 1e for pentacene
on the substrate and in the crystal. The radiation from a flat-
lying molecule is efficiently screened by the metal substrate
and can in addition not couple well to tip-induced plasmonic
modes. The efficiency of radiation into the detected far field
becomes significantly increased through tip enhancement.[17,19]
As the second mechanism, however, would not apply for the
suppression of luminescence for very thin crystals the first
mechanism appears to be decisive herein.
In conclusion, our experiments demonstrate STM-induced lu-
minescence from an acene. Luminescence from delocalized ex-
citons can be obtained even for highly localized injection of
charge carriers from an STM tip. The interaction between
neighboring molecules provides a sufficient delocalization of
the excitation to yield bulk-like emission spectra. The suppres-
sion of luminescence for ultrathin crystals suggests emission
from the nanocrystal at sufficiently large distance (>1 nm)
from the metallic electrodes. The crystal structure can affect
the emission conditions through the orientation of the emit-
ting transition dipole. The experimental setup employed
herein realizes an organic light source made of a homogene-
ous material with bulk properties near its lower size limit. The
results can provide a basis for studies to explore new function-
alities by employing layered organic structures or single mole-
cules embedded in host nanocrystals.
We thank Prof. J. Wrachtrup and Prof. J. Pflaum for valuable dis-
cussions and the supply with purified pentacene powder.
pentacene · scanning probe microscopy
 M. Kitamura, Y. Arakawa, J. Phys. Condens. Matter 2008, 20, 184011.
 J. Repp, G. Meyer, S. M. Stojkovic, A. Gourdon, C. Joachim, Phys. Rev.
Lett. 2005, 94, 026803.
 W.-H. Soe, C. Manzano, A. DeSarkar, N. Chandrasekhar, C. Joachim, Phys.
Rev. Lett. 2009, 102, 176102.
 L. Gross, F. Mohn, N. Moll, P. Liljeroth, G. Meyer, Science 2009, 325, 1110.
 M. Schwoerer, H. C. Wolf, Organic Molecular Solids, Wiley-VCH, 2007.
 C. D. Dimitrakopoulos, A. R. Brown, A. Pomp, J. Appl. Phys. 1996, 80,
 H. Klauk, M. Halik, U. Zschieschang, G. Schmid, W. Radlik, W. Weber, J
Appl. Phys. 2002, 92, 5259.
 H. Kakuta, T. Hirahara, I. Matsuda, T. Nagao, S. Hasegawa, N. Ueno, K. Sa-
kamoto, Phys. Rev. Lett. 2007, 98, 247601.
 S. Berkebile, P. Puschnig, G. Koller, M. Oehzelt, F. P. Netzer, C. Ambrosch-
Draxl, M. G. Ramsey, Phys. Rev. B 2008, 77, 115312.
 W. Deng, D. Fujita, T. Ohgi, J. Chem. Phys. 2002, 117, 4995.
 X. H. Qiu, G. V. Nazin, W. Ho, Science 2003, 299, 542.
 R. Berndt, R. Gaisch, W. D. Schneider, J. K. Gimzewski, B. Reihl, R. R.
Schlittler, M. Tschudy, Appl. Phys. A 1993, 57, 513.
 Z. C. Dong, X. L. Guo, A. S. Trifonov, P. S. Dorozhkin, K. Miki, K. Kimura, S.
Yokoyama, S. Mashiko, Phys. Rev. Lett. 2004, 92, 086801.
 E. C´avar, M.-C. Bl?m, M. Pivetta, F. Patthey, M. Chergui, W.-D. Schneider,
Phys. Rev. Lett. 2005, 95, 196102.
 H. W. Liu, Y. Ie, R. Nishitani, Y. Aso, H. Iwasaki, Phys. Rev. B 2007, 75,
 D. Ino, T. Yamada, M. Kawai, J. Chem. Phys. 2008, 129, 014701.
 Z. C. Dong, X. L. Zhang, H. Y. Gao, Y. Luo, C. Zhang, L. G. Chen, R. Zhang,
X. Tao, Y. Zhang, J. L. Yang, J. G. Hou, Nat. Photonics 2010, 4, 50.
 S. W. Wu, G. V. Nazin, W. Ho, Phys. Rev. B 2008, 77, 205430.
 F. Rossel, M. Pivetta, F. Patthey, W.-D. Schneider, Optics Express 2009, 17,
 B. Ren, G. Picardi, B. Pettinger, Rev. Sci. Instrum. 2004, 75, 837.
 T. Tomiki, J. Phys. Soc. Jpn 1967, 22, 463.
 C. C. Mattheus, A. B. Dros, J. Baas, A. Meetsma, J. L. de Boer, T. T. M. Pal-
stra, Acta Crystallogr. C 2001, 57, 939.
 M. L. Tiago, J. E. Northrup, S. G. Louie, Phys. Rev. B 2003, 67, 115212.
 L. Ruppel, A. Birkner, G. Witte, C. Busse, T. Lindner, G. Paasch, C. Woll, J.
Appl. Phys. 2007, 102, 033708.
 R. Berndt, J. K. Gimzewski, P. Johansson, Phys. Rev. Lett. 1991, 67, 3796.
 R. He, X. Chi, A. Pinczuk, D. V. Lang, A. P. Ramirez, Appl. Phys. Lett. 2005,
 E. Heinecke, D. Hartmann, A. Hese, J. Chem. Phys. 2003, 118, 113.
 R. Lehnig, A. Slenczka, J. Chem. Phys. 2005, 122, 244317.
 U. P. Wild, F. Guttler, M. Pirotta, A. Renn, Chem. Phys. Lett. 1992, 193,
 W. E. Geacintov, J. Burgos, I. M. Pope, C. Strom, Chem. Phys. Lett. 1971,
 M. Schidleja, C. Melzer, H. von Seggern, Appl. Phys. Lett. 2009, 94,
 T. Aoki-Matsumoto, K. Furuta, T. Yamada, H. Moriya, K. Mizuno, Int. J.
Mod. Phys. B 2001, 15, 3753.
 R. He, N. G. Tassi, G. B. Blanchet, A. Pinczuk, Appl. Phys. Lett. 2005, 87,
 D. V. Lang, X. Chi, T. Siegrist, A. M. Sergent, A. P. Ramirez, Phys. Rev. Lett.
2004, 93, 086802.
 K. Kuhnke, R. Becker, M. Epple, K. Kern, Phys. Rev. Lett. 1997, 79, 3246.
 I. Horcas, R. Fernandez, J. M. Gomez-Rodriguez, J. Colchero, J. Gomez-
Herrero, A. M. Baro, Rev. Sci. Instrum. 2007, 78, 013705.
Received: July 1, 2010
Published online on October 28, 2010
? 2010 Wiley-VCH Verlag GmbH&Co. KGaA, WeinheimChemPhysChem 2010, 11, 3412–3416