Atomic force microscopy reveals bistable configurations of dibenzo[a,h]thianthrene
and their interconversion pathway
Niko Pavliˇ cek,1, ∗Benoit Fleury,2, †Mathias Neu,1Judith Niedenf¨ uhr,1
Coral Herranz-Lancho,2Mario Ruben,2,3and Jascha Repp1
1Institute of Experimental and Applied Physics, University of Regensburg, 93053 Regensburg, Germany
2Institute of Nanotechnology (INT), Karlsruhe Institute of Technology (KIT), 76344 Eggenstein-Leopoldshafen, Germany
3Institut de Physique et Chimie des Mat´ eriaux de Strasbourg (IMCMS),
CNRS-Universit´ e de Strasbourg, 67034 Strasbourg, France
(Dated: January 10, 2012)
We investigated dibenzo[a,h]thianthrene molecules adsorbed on ultrathin layers of NaCl using a
combined low-temperature scanning tunneling and atomic force microscope. Two stable configu-
rations exist corresponding to different isomers of free nonplanar molecules. By means of excita-
tions from inelastic electron tunneling we can switch between both configurations. Atomic force
microscopy with submolecular resolution allows unambiguous determination of the molecular geom-
etry, and the pathway of the interconversion of the isomers. Our investigations also shed new light
on contrast mechanisms in scanning tunneling microscopy.
PACS numbers:68.37.Ef, 68.37.Ps, 68.43.Fg, 82.37.Gk
molecule has been visualized by means of non-contact
atomic force microscopy (AFM) . Shortly after, this
method assisted in identifying the structure of an or-
ganic molecule . In conjunction with the capability of
scanning tunneling microscopy (STM) to perform orbital
imaging on ultrathin insulating films , it is possible to
gain independent and complementary information of the
molecular as well as of the adsorption geometry, but also
of the electronic structure of individual molecules.
Unambiguous identification of configurational changes
of adsorbed molecules is a challenging task by means of
STM alone  probing the local density of states rather
than geometry. Usually, additional techniques such as
near-edge X-ray adsorption fine structure (NEXAFS)
measurements have to be employed [5, 6].
In this Letter, we present combined STM and AFM ex-
periments of dibenzo[a,h]thianthrene (DBTH) molecules
adsorbed on ultrathin layers of sodium chloride.
demonstrate controlled switching between two different
molecular configurations by means of inelastic excita-
tions.AFM images with submolecular resolution di-
rectly reveal the configurational changes. Stereochem-
istry could be utilized to determine their interconversion
pathway in detail.
All AFM measurements were carried out in a home-
built combined STM and AFM operating in ultra-high
vacuum (p < 10−10mbar) at T = 5K. The AFM, based
on the qPlus tuning fork design (spring constant k0 ≈
1.8 × 103Nm−1, resonance frequency f0 = 26057Hz,
quality factor Q ≈ 104) , was operated in the frequency
modulation mode . Sub-˚ Angstrom oscillation ampli-
tudes have been used to maximize the lateral resolution
. Some of the STM measurements (Fig. 1 and Fig. 2)
were performed in a similar modified commercial STM
from SPS-CreaTec. The bias voltage V was applied to
the chemical structure of a pentacene
non-equivalent species of the same individual DBTH molecule
on NaCl(2ML)/Cu(111). (a) Ball and stick model of DBTH
molecules and definition of dihedral angle Θ. Grey, white, and
yellow represent C, H, and S atoms. (b) and (c) Constant-
current images of U and D configuration, respectively, at low
bias voltage (imaging parameters: I = 0.4pA, V = 0.01V).
(d) and (e) Corresponding LUMO images acquired with same
metal tip apex (I = 0.2pA, V = 2.3V).
Bias dependent STM imaging reveals two possible
Sodium chloride was evaporated onto clean Cu(111)
single-crystals at sample temperatures of about 280K
.All experiments were carried out on a dou-
ble layer, and we denote this substrate system as
NaCl(2ML)/Cu(111). The DBTH molecules were syn-
thesized as described previously .
Low coverages of CO (for tip functionalization) and
DBTH molecules were adsorbed at sample temperatures
below 10K. Following a recently developed technique,
the tip had been terminated with a CO molecule for all
AFM measurements to enhance the resolution consider-
DBTH molecules are within the family of thianthrenes,
in which, due to the presence of the lone pairs of the
50 Å 50 Å
by means of inelastic excitations.
shows subsequent STM orbital images (I = 0.2pA, V =
2.35V). Circles indicate previous positions of the center of
the molecule. From (a) to (b) only the adsorption position
changes. In contrast, both, position and in-plane orientation
change from (b) to (c) and subsequently to (d).
Switching between molecular configurations
thioether groups, the molecules are folded along the S-
S axis [12–17]. To obtain the folding angle for DBTH
molecules, we have performed density functional the-
ory (DFT) calculations [18, 19] for a free molecule (i. e.
without substrate) using the highly optimized CPMD code
.Ab initio norm-conserving pseudopotentials 
were used. In the optimized geometry we find a dihedral
angle [see Fig. 1(a) for definition] of Θ = 133◦. This
value is larger than the one observed for thianthrene by
x-ray crystallography with Θ = 128◦, whereby the
difference can be attributed to the extension of the π-
conjugation in DBTH molecules and thus flattening the
NaCl(2ML)/Cu(111) are presented in Fig. 1.
species of DBTH molecules that are non-equivalent with
respect to translation, mirroring and/or rotations exist
on the surface, which we denote as U and D as indicated
in the caption of Fig. 1. STM images at low bias voltages
[Fig. 1(b),(c)] as well as orbital images [Fig. 1(d),(e)]
show distinct differences between both species. However,
we note that the nodal plane structure of the orbital
images is the same. In addition, STS spectra acquired
on both species are similar, and the only peak in the
accessible voltage range at around 2.45V corresponds to
the negative ion resonance (NIR).
By means of inelastic excitations [23, 24] it is possible
to induce lateral motion of individual molecules, as well
as to switch between the two species. The results are pre-
sented in Fig. 2 . Analysis of many excitations reveals
that whenever a molecule switches from U to D or vice
versa, both in-surface-plane orientation and adsorption
site change (see below for a detailed adsorption geome-
try determination). Thus, every species can be assigned
to a different adsorption site.
STM images at low bias voltages demonstrate that
neither species has an influence on the scattering wave
pattern of the substrate’s interface state. Hence, both
species are neutral [27, 28]. The similarity in the elec-
tronic structure (nodal planes in orbital imaging and the
energy of the NIR) suggests that the molecule is either
in the same or an equivalent configuration, that is, it
has a similar dihedral angle. The different appearance
could be simply due to different adsorption sites leading
to slightly different electronic properties. It must be em-
phasized that we can not draw any conclusions from the
STM data going beyond what has been discussed until
As a next step, we performed constant height ∆f imag-
ing presented in Fig. 3(a) in AFM mode. It is immedi-
ately apparent that the two species correspond to differ-
ent molecular configurations . First, consider the S-S
axis. Whereas U molecules show a characteristic bright
stripe, D molecules do not show any atomic contrast at
the center at the same tip height. In addition, the carbon
rings of U molecules show an apparent distortion.
Before we shed light on the actual structure of both
configurations, we have to discuss the origin of the con-
trast in the frequency shift. For the small amplitudes
used here, the frequency shift well approximates the force
gradient. While attractive long-range forces are responsi-
ble for overall negative ∆f background, repulsive short-
range contributions due to Pauli repulsion are decisive
for the intramolecular contrast as was shown in [1, 26].
These short-range forces are very sensitive to the tip-
molecule distance. Examination of C atoms and C-C
bonds shows a specific gradient for both species. In the D
(U) configuration the contrast increases (decreases) with
the distance to the S-S axis. As discussed above, in this
regime we can attribute a higher contrast with stronger
repulsive contributions, and consequently a smaller dis-
tance to the molecule. Thus, the bright band along the
S-S axis in the U configuration can be attributed to a
small distance to the tip. We believe that the sharp fea-
ture is a fingerprint of the CO molecule at the tip apex
which bends due to the protruding S atoms . Ac-
cordingly, in this configuration the S atoms are pointing
upwards, and the aromatic rings are close to the surface.
In contrast, the S atoms of D molecules are closest to the
surface, whereas its naphthalene units are pointing up-
wards. Both configurations are illustrated in Fig. 3(d).
The stereochemistry of DBTH molecules can be uti-
lized to determine the pathway of the configurational
change [30, 31]. We have estimated the barrier for flap-
ping to be about 200meV from our DFT calculations
using the relative energy vs. folding angle Θ. For this
purpose we optimized the geometry for fixed Θ in steps
of 10◦. As flapping of molecules is frozen at the temper-
atures of our experiments, there are two enantiomers of
the free molecule possessing C2symmetry, as depicted in
Fig. 3(e) . In general, going from upwards to down-
ward pointing S atoms can be realized either by flipping
from one face to the other, or by flapping the naphtha-
lene units with respect to the S atoms. The molecular
image. Imaging parameters: oscillation amplitude A = 0.5˚ A, V = 0V, ∆z = 0.0˚ A. ∆z corresponds to a distance decrease
with respect to an STM set-point of I = 0.5pA, V = 0.4V above the clean NaCl(2ML)/Cu(111). (b) Image of the same area
as in (a) after both molecules changed their adsorption position (A = 0.5˚ A, V = 0V, ∆z = 0.1˚ A). Insets in (a) and (b) show
constant-current STM images of the same frame. Panel (c) represents the curvature of the image in (a) obtained by calculating
the Laplacian. Molecular models (drawn to scale) for U and D are overlaid as a guide to the eye; the slightly larger appearance
of molecules has been discussed previously . Inset shows atomically resolved NaCl lattice. (d) Model representing molecules
in U and D configuration on a surface. (e) Model depicting chiral enantiomers of the free molecule.
AFM measurements on DBTH on NaCl(2ML)/Cu(111) with a CO-functionalized tip. (a) Constant height AFM
structure projected onto the surface plane would change
for flipping, but not for flapping. As can be seen from the
left molecule in Fig. 3(a) and 3(b), the latter is the case.
From this it follows immediately that switching is caused
by a change of the folding angle, and initial and final con-
figuration correspond to the different chiral enantiomers
of the free molecule as depicted in Fig. 3(e).
To get more insight, we determined the adsorption site
of molecules in both states. Therefore, we used AFM
imaging with STM constant-current feedback. That is,
the frequency shift ∆f is recorded simultaneously with
the STM topography while the STM feedback is used
to control the tip-height .
frequency shift ∆f of such measurements for both con-
figurations. Recently, it has been shown that the faint
maxima correspond to the Cl sites of the NaCl(100) lat-
tice . From the models of the adsorption geometries
shown in Fig. 4(g) and 4(h) we can conclude that the
S atoms are located on top of Na sites for D molecules.
This can be rationalized by the ionic nature of the NaCl
lattice as follows: the downward pointing lone pairs of
S atoms are attracted by Na+ions and repelled by Cl−
ions. In contrast, the S atoms of U molecules are farthest
from the surface, and the naphthalene units align with
a polar direction of NaCl. This results in a geometry in
Fig. 4(c)-(f) shows the
which the center of the molecule is located above a Na
Since we are by now in a position, in which we have
a detailed picture of our system, we would like to re-
visit Fig. 1 particularly with regard to the actual molec-
ular structure. Resonance images resembling the orbital
structure can not be used to make a statement about the
geometry. In contrast, images at low voltages can be ar-
gued to be due to modifications of the tunneling barrier
and should therefore be somewhat related to geometry.
However, these images show a pronounced s-shape for
U and a relatively straight rod for D molecules, whereas
the apparent height above their center is the same. Thus,
they do neither reflect the lateral geometry nor the ver-
tical changes in the structure.
In conclusion, we have revealed the exact geometry
of two different configurations of DBTH molecules on
NaCl(2ML)/Cu(111) using combined STM and AFM ex-
periments. Our experiments demonstrate reproducible
and non-destructive switching between both configura-
tions. We have shown that AFM with submolecular
resolution is capable of directly revealing configurational
changes in real space. Our data visualize displacements
perpendicular to the surface in a nonplanar molecule.
Taking the chirality of DBTH molecules into account,
STM feedback. (a) and (b) Constant current STM images (imaging parameters I = 1.0pA, V = 0.3V). The corresponding
frequency shift ∆f recorded simultaneously is shown in panels (c) and (d) (oscillation amplitude A = 0.5˚ A, and A = 1.0˚ A,
respectively). Panel (e) and (f) show the same data with a different ∆f scale to highlight the faint maxima resembling Cl sites,
for which a lattice was overlaid as a guide to the eye. (g) and (h) Models of the adsorption position as determined from the
experimental data. Big green and small purple balls represent Cl−and Na+ions, respectively.
Determination of the molecular adsorption position for down (top row) and up (bottom) geometry using AFM with
we could unambiguously determine the interconversion
The authors thank L. Gross, N. Moll, G. Meyer, I.
Swart, and F. Albrecht for valuable comments. Financial
support from the Volkswagen Foundation (Lichtenberg
program), the Deutsche Forschungsgemeinschaft (GRK
1570 and SPP 1243), and from the Marie-Curie-ITN
“SMALL” is gratefully acknowledged.
†On leave from Institut Parisien de Chimie Mol´ eculaire,
UPMC and CNRS UMR7201, UPMC-Paris 6, 75252,
Paris Cedex 05, France
 L. Gross, F. Mohn, N. Moll, P. Liljeroth, and G. Meyer,
Science 325, 1110 (2009).
 L. Gross, F. Mohn, N. Moll, G. Meyer, R. Ebel, W. M.
Abdel-Mageed, and M. Jaspars, Nat. Chem. 2, 821
 J. Repp, G. Meyer, S. M. Stojkovi´ c, A. Gourdon, and
C. Joachim, Phys. Rev. Lett. 94, 026803 (2005).
 T. A. Jung, R. R. Schlittler, and J. K. Gimzewski, Na-
ture 386, 696 (1997).
 W. Auw¨ arter, F. Klappenberger, A. Weber-Bargioni,
A. Schiffrin, T. Strunskus, C. W¨ oll, Y. Pennec, A. Rie-
mann, and J. V. Barth, J. Am. Chem. Soc. 129, 11279
 A. Weber-Bargioni,W.
berger, J.Reichert, S.
C. W¨ oll, A. Schiffrin, Y. Pennec,
ChemPhysChem 9, 89 (2008).
 F. J. Giessibl, Appl. Phys. Lett. 76, 1470 (2000).
 T. R. Albrecht, P. Gr¨ utter, D. Horne, and D. Rugar, J.
Appl. Phys. 69, 668 (1991).
 F. J. Giessibl, Rev. Mod. Phys. 75, 949 (2003).
and J. V. Barth,
 R. Bennewitz, V. Barwich, M. Bammerlin, C. Lop-
pacher, M. Guggisberg, A. Baratoff, E. Meyer, and H.-J.
G¨ untherodt, Surf. Sci. 438, 289 (1999).
 A. Spurg, G. Schnakenburg, and S. R. Waldvogel, Chem.
Eur. J. 15, 13313 (2009).
 I. Rowe and B. Post, Acta Crystallogr. 11, 372 (1958).
 H. Lynton and E. G. Cox, J. Chem. Soc. , 4886 (1956).
 J. Stenhouse, Justus Liebigs Annalen der Chemie 149,
 M. J. Aroney, R. J. W. Le Fevre, and J. D. Saxby, J.
Chem. Soc., 571 (1965).
 K. L. Gallaher and S. H. Bauer, J. Chem. Soc., Faraday
Trans. 2 71, 1173 (1975).
 S. Hosoya, Acta Crystallogr. 16, 310 (1963).
 P. Hohenberg and W. Kohn, Phys. Rev. 136, B864
 W. Kohn and L. J. Sham, Phys. Rev. 140, A1133 (1965).
 CPMD V3.15 Copyright IBM Corp 1990-2011, Copyright
MPI fuer Festkoerperforschung Stuttgart 1997-2001.
 J. P. Perdew, K. Burke, and M. Ernzerhof, Phys. Rev.
Lett. 77, 3865 (1996).
 The cell size was 18˚ A×16.2˚ A×9˚ A. A plane-wave cut-off
energy of 130Ry was used.
 T. Komeda, Y. Kim, M. Kawai, B. N. J. Persson, and
H. Ueba, Science 295, 2055 (2002).
 T. Sonnleitner, I. Swart, N. Pavliˇ cek, A. P¨ ollmann, and
J. Repp, Phys. Rev. Lett. 107, 186103 (2011).
 STM and AFM data before and after an excitation show
that molecules and NaCl remain undamaged.
 N. Moll, L. Gross, F. Mohn, A. Curioni, and G. Meyer,
New J. Phys. 12, 125020 (2010).
 J. Repp, G. Meyer, F. E. Olsson, and M. Persson, Science
305, 493 (2004).
 I. Swart, T. Sonnleitner, and J. Repp, Nano Lett. 11,
 C. Loppacher, M. Guggisberg, O. Pfeiffer, E. Meyer,
M. Bammerlin, R. L¨ uthi, R. Schlittler, J. K. Gimzewski,
H. Tang, and C. Joachim, Phys. Rev. Lett. 90, 066107
 R. Raval, Chem. Soc. Rev. 38, 707 (2009).
 M. J. Comstock, D. A. Strubbe, L. Berbil-Bautista,
N. Levy, J. Cho, D. Poulsen, J. M. J. Fr´ echet, S. G.
Louie, and M. F. Crommie, Phys. Rev. Lett. 104, 178301
 In total, 4 configurational isomers exist on the surface:
one pair of enantiomers for molecules in each configura-
tion (U and D).
 STM and AFM signals are convoluted in this mode. How-
ever, this mode enables atomic resolution on NaCl and
stable imaging of the molecules at the same time .
 All possible, independently determined adsorption ge-
ometries are consistent with symmetry considerations.
We estimate the error bars to be ±0.5˚ A for the site de-
termination and ±3◦for the in-plane orientation.