Synthesis and Surface Behaviour of NDI Chromophores Mounted on a Tripodal Scaffold: Towards Self-Decoupled Chromophores for Single-Molecule Electroluminescence

Article abstract of DOI:10.1002/chem.202101264

This paper reports the efficient synthesis, absorption and emission spectra, and the electrochemical properties of a series of 2,6-disubstituted naphthalene-1,4,5,8-tetracarboxdiimide (NDI) tripodal molecules with thioacetate anchors for their surface inv

Full text of DOI:10.1002/chem.202101264

Full Paper
doi.org/10.1002/chem.202101264
Chemistry—A European Journal
Synthesis and Surface Behaviour of NDI Chromophores
Mounted on a Tripodal Scaffold: Towards Self-Decoupled
Chromophores for Single-Molecule Electroluminescence
Nico Balzer,[a] Jan Lukášek,[a] Michal Valášek,*[a] Vibhuti Rai,[b] Qing Sun,[b] Lukas Gerhard,[b]
Wulf Wulfhekel,*[b, c] and Marcel Mayor*[a, d, e]
Abstract: This paper reports the efficient synthesis, absorption
and emission spectra, and the electrochemical properties of a
series of 2,6-disubstituted naphthalene-1,4,5,8-tetracarboxdi-
imide (NDI) tripodal molecules with thioacetate anchors for
their surface investigations. Our studies showed that, in
particular, the pyrrolidinyl group with its strong electron-
donating properties enhanced the fluorescence of such core-
substituted NDI chromophores and caused a significant
bathochromic shift in the absorption spectrum with a
correspondingly narrowed bandgap of 1.94 eV. Cyclic voltam-
metry showed the redox properties of NDIs to be influenced
by core substituents. The strong electron-donating character
of pyrrolidine substituents results in rather high HOMO and
LUMO levels of -5.31 and -3.37 eV when compared with the
parental unsubstituted NDI. UHV-STM measurements of a
sub-monolayer of the rigid tripodal NDI chromophores spray
deposited on Au(111) show that these molecules mainly tend
to adsorb flat in a pairwise fashion on the surface and form
unordered films. However, the STML experiments also
revealed a few molecular clusters, which might consist of
upright oriented molecules protruding from the molecular
island and show electroluminescence photon spectra with
high electroluminescence yields of up to 6×10À 3. These
results demonstrate the promising potential of the NDI
tripodal chromophores for the fabrication of molecular
devices profiting from optical features of the molecular layer.
Introduction
the attention on NDIs and their core substituted derivatives.[10]
These NDI scaffolds with their large electron deficient aromatic
system show reversible electrochemical reduction properties to
form both, radical anions after one-electron reduction and
dianions after two-electron reduction. Furthermore, physical
and physicochemical properties of NDIs can be chemically
optimised by substituents for particular purposes while the NDI
core structure remains preserved. The electronic properties of
NDIs can be adjusted by either electron withdrawing groups
(EWGs) or/and electron donating groups (EDGs) at the core-
positions and to a lesser extent at the axial positions of the NDI
scaffolds. NDI scaffolds offer six positions for functionalisation
divided mainly in two categories, namely the N-termini at the
axial positions and the naphthalene core. Functionalisation of
With the growing demand of cheap and environmentally
friendly electronic devices like organic light emitting diodes
(OLEDs), the research field of molecular organic electronics
gained increasing attention.[1] Amongst others,[2] various deriva-
tives of naphthalene-1,4,5,8-tetracarboxdiimides (NDIs) are
widely used for organic electronics due to their by core
substituents tuneable electronic and optical properties,[3] their
chemical and photochemical stability,[3,4] and their favourable
electron transport properties.[5,6] Consequently, OLEDs based on
these subunits have been recently reported.[7,8] Although
Vollmann et al. first published this type of molecules already in
the early 1930s,[9] it was the group of Würthner who draw anew
[a] N. Balzer, J. Lukášek, Dr. M. Valášek, Prof. Dr. M. Mayor
Institute of Nanotechnology
[d] Prof. Dr. M. Mayor
Department of Chemistry, University of Basel
St. Johanns-Ring 19, 4056 Basel (Switzerland)
E-mail: marcel.mayor@unibas.ch
Karlsruhe Institute of Technology
P.O. Box 3640, 76021 Karlsruhe (Germany)
E-mail: michal.valasek@kit.edu
[e] Prof. Dr. M. Mayor
[b] V. Rai, Q. Sun, Dr. L. Gerhard, Prof. Dr. W. Wulfhekel
Institute of Quantum Materials and Technologies
Karlsruhe Institute of Technology
Lehn Institute of Functional Materials
School of Chemistry, Sun Yat-Sen University
Guangzhou, Guangdong 510275 (P. R. China)
76021 Karlsruhe (Germany)
E-mail: wulf.wulfhekel@kit.edu
Supporting information for this article is available on the WWW under
https://doi.org/10.1002/chem.202101264
[c] Prof. Dr. W. Wulfhekel
© 2021 The Authors. Chemistry - A European Journal published by Wiley-
VCH GmbH. This is an open access article under the terms of the Creative
Commons Attribution Non-Commercial NoDerivs License, which permits use
and distribution in any medium, provided the original work is properly cited,
the use is non-commercial and no modifications or adaptations are made.
Physikalisches Institut, Karlsruhe Institute of Technology
Wolfgang-Gaede-Straße 1, 76131 Karlsruhe (Germany)
Chem. Eur. J. 2021, 27, 1–13
1
© 2021 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH
��
These are not the final page numbers!

Full Paper
doi.org/10.1002/chem.202101264
Chemistry—A European Journal
the axial N-imide positions of NDI mainly influences its
solubility[11] or morphology in thin films,[12] while having
negligible effect on the optoelectronic properties. On the other
hand, the chemical substitution at the 2, 3, 6 and 7 positions of
the NDI core with various groups, such as alkyl or arylamino,
alkoxy, sulfanyl, cyano, thiophene and different heterocycles,
considerably tune their electronic and photophysical properties.
Therefore, molecular engineers use these positions to fine-tune
the HOMO and LUMO energy levels of the NDI core in order to
obtain model compounds with tailor-made optical and elec-
tronic properties.[3,10,13] The ability to change both, colour and
redox properties without global structural changes is the most
attractive characteristics of core-substituted NDIs. Since the
2000s, many groups have developed various strategies to core
substitute NDIs with electron donating heteroatoms like S, O or
N and thus create push-pull systems with the electron poor NDI
core to modulate the absorption and emission properties of the
organic chromophore.[10,13,14] The colours of core-substituted
NDIs originate from a charge transfer band that moves to the
red over the full rainbow spectrum with increasing push-pull
character between the core substituent and the NDI core. While
photoluminescence properties of NDIs have been extensively
studied,[3] their electroluminescence behaviour is barely
studied.[7,8] Given their synthetic accessibility, tuneable photo-
physical properties, reversible electrochemistry,[15,16] ground-
and excited state stability, and structural integrity, NDIs are
particular appealing chromophores for the development of
model compounds to investigate single molecule electrolumi-
nescence. However, to the best of our knowledge, only few
studies have been published so far.[17,18]
Fathoming the underlying light emission processes from
these molecules is thus not only interesting from a scientific
point of view, but also key for the development of new
materials for optoelectronic devices. Scanning tunnelling micro-
scopy induced luminescence (STML) experiments enable to
study electrically induced light emission and the underlying
mechanisms of the optoelectronic response with submolecular
precision.[19–21] Using STML setups, various pathways for single
molecule light emission have been discovered at the single
molecule level, including dipole–dipole coupling, hot electro-
luminescence, plasmon-exciton coupling, and electro-
fluorochromism.[22–25] However, in order to observe electro-
luminescence from single organic chromophores in STML
experiments, the chromophore must be electronically de-
coupled from the underlying metal substrate to compete with
quenching of the molecule’s excited state by the metal’s
electrons.[26,27] Common approaches to achieve such isolation of
individual chromophores are either ultrathin (in)organic insulat-
ing layers with large band gaps of several electron volts
between the metal surface and chromophores,[26–29] or multi-
layered stacks of the chromophore.[30] Alternative, less explored
strategies to electronically decouple organic emitters from the
underlying surface involve either bulky spacer groups within
the tailor-made molecule or large multipodal platforms.[31,32] The
latter has been recently applied for example with a fully
symmetric tetrapodal molecule, from which three functional
arms served as tripod and the fourth arm was perpendicularly
arranged in an upward direction,[33] or by separating the
chromophore subunit with rigid multipodal anchoring
a
scaffold.[34–36] The modular strategy of electronically decoupling
via a rigid multipodal platform not only enables the inves-
tigation of a variety of chromophores, but also allows to align
the chromophore’s dynamic dipole vector perpendicular to the
surface resulting in enhanced luminescence.[37]
Here we report the design, synthesis and characterisation of
the three molecules Tpd-hNDI, Tpd-sNDI and Tpd-nNDI (with
h/s/n representing the (hetero)atom directly connected to the
NDI core) as molecular emitters mounted on a tripodal scaffold
as displayed in Figure 1a), together with first STM and STML
studies of their (sub)monolayers obtained by spray deposition
and subsequent annealing on Au(111) surfaces. While the
molecular design is geared towards upright standing NDI
chromophores fixed by the tripodal platform (Figure 1b), we
learned in the past with model compounds comprising
tetraphenylmethane based tripodal platforms about their hardly
predictable surface behaviour.[38,39] Thus the scanning probe
experiments enable to shed light on whether the three
potential covalent sulfur gold bond of the tripodal platform
dominate the monolayer formation and provide upright stand-
ing chromophores as displayed in Figure 1b), or the van der
Waals attraction between the surface and the chromophore’s
extended π-system dominate the molecule’s surface behaviour
as sketched in Figure 1c). In addition to the parent NDI
structure, amino and sulfanyl groups are considered as core
substituents to cover a broad range of emission wavelengths.
Phenylacetylene serves as a linker to maintain the molecular
rigidity and efficiently decouple the NDI emitter from the
tripodal platform. The tripodal design of the foot structure with
three thiol anchoring groups in the meta-positions was chosen
with the intention to provide a rigid, covalently bonded anchor
to the Au(111) surface with an upright orientation of the
protruding functionality.[40] These tripodal molecules with acetyl
protected anchoring groups can be efficiently deprotected
either in situ[40,41] upon binding to the gold surface or chemically
by cleaving agents.[42] With this strategy, we aim to enable self-
decoupling of NDI chromophores from the substrate by
molecular design. The scope of the present study is not only
the synthesis of the target molecules Tpd-hNDI, Tpd-sNDI and
Tpd-nNDI, but also to study their photophysical and electro-
chemical properties and surface investigations with the STML
setup.
It should be noted that the stabilisation of the functional
tripodal molecules on metallic surfaces is a subtle balance
which is attributed both to the lateral van der Waals
interactions for example between flat delocalised π-systems
and to the chelate effect of the tripodal scaffolds. Therefore,
one has to consider the competing binding energies of the two
potential anchoring sites of functionalised tripods, namely the
functional unit-surface interface and the degree of chelatation
via thiolate anchors of the molecular tripod in a commensurate
order to the gold surface. If the interactions between functional
unit and surface is larger than the energy gained by the foot
structure, the molecules will adsorb flat on the surface.[38] Thus,
the proposed tripodal platforms exposing large NDI chromo-
Chem. Eur. J. 2021, 27, 1–13
www.chemeurj.org
2
© 2021 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH
��
These are not the final page numbers!

Full Paper
doi.org/10.1002/chem.202101264
Chemistry—A European Journal
Figure 1. a) Series of the tripodal platforms with NDI chromophores Tpd-hNDI, Tpd-sNDI and Tpd-nNDI and b) their intended arrangement on a Au(111)
substrate, with the colour code of NDI emitters representing their solution colours. c) Expected arrangement of the model compounds on the substrate in
case of dominating van der Waals attractions.
phores are good candidates to investigate how such extended
molecules arrange on the substrate.
iol anchor groups.[34] In the here presented tripod design, the
aromatic thiol anchors allow direct connections to the gold
substrate avoiding additional flexibility due to CH2 linkers. The
intention is to improve both, the electron transport and the
rigidity of the spatial arrangement, hopefully controlling an
upright orientation of the molecular rods on the gold substrate.
The synthesis of 3,3,3-tris(3-methoxyphenyl)propyne 1 was
previously reported by Garcia-Garibay et al.[44] While the
demethylation of 1 with commonly used Lewis acids such as
BBr3 would require an additional protection of the triple bond
to avoid its borylation,[45] a method originally published by
Feutrill and Mirrington involving sodium ethanethiolate in DMF
as demethylation agent was used to afford the desired product
triol 2 in 83% yield.[46] The following step was triflatation of triol
2 with triflic anhydride to yield triflate 3 in 73%. In the next
step, the thiol anchor groups were introduced in their TMS-
ethyl-protected form. The triflate 3 was converted by the CÀ S
cross-coupling reaction with 2-(trimethylsilyl)ethanethiol in the
presence of palladium catalyst and Xantphos ligand to the
desired product 4 with a yield of 80%. Subsequent trans-
protection of the thiols in 4 was performed using AgBF4 and
acetyl chloride in dichloromethane to afford the desired
thioacetate 5 in very good yield of 92%.[47] Unfortunately, the
thioacetate groups did not provide the stability required for the
Sonogashira coupling reactions disqualifying the modular plan
of decorating the platform in the last step with the NDI
chromophores. As second best alternative the 2-(trimethylsilyl)
ethyl protected derivative 4 was used for the assembly between
NDI chromophore and platform, followed by a final trans-
protection step to get the target molecules as thioacetates.
NDI derivatives hNDI, sNDI and nNDI suitable for their
assembly with the tripodal platform were synthesised as shown
in Scheme 2. 2,6-Dichloronaphthalene-1,4,5,8-tetracarboxylic
Results and Discussion
Design and synthesis of molecules
The synthetic approach towards a rigid tripodal platform 4 is
based on a triphenylmethane derivative and displayed in
Scheme 1.[34,40,43] Introducing a triple bond on the sp3-hybridised
carbon atom of the tripodal platform is expected to provide the
compactness required for STM based investigations and enables
to modularly develop the structure through Sonogashira
coupling chemistry. Note, that a comparable foot structure
based on a tetraphenylmethane scaffold has been recently used
in single molecule light emission experiments by Hou et al., but
significant differences are expected in the sturdiness of the
tripod-Au(111)-interface arrangement due to the used benzylth-
Scheme 1. Synthesis of the tripodal platforms. i) sodium ethanethiolate,
DMF, 83%; ii) Tf2O, Et3N, 73%; iii) 2-(trimethylsilyl)ethanethiol, Xantphos,
Pd2(dba)3, 80%; iv) AgBF4, AcCl, 92%.
Chem. Eur. J. 2021, 27, 1–13
www.chemeurj.org
3
© 2021 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH
��
These are not the final page numbers!

Full Paper
doi.org/10.1002/chem.202101264
Chemistry—A European Journal
and thus prevent their further substitution under harsh
conditions in the final transprotection step. Initially carbazoles
were considered as amino substituents, however their reactivity
in SNAr reaction with 8 was too low and did not lead to the
desired product. Moreover, primary alkyl amines as NDI core
substituents were avoided, because particular decay mecha-
nisms were suggested for substituents containing an α-proton
next to the heteroatom. In particular their fluorescence might
be quenched either by electron transfer into the SOMO (former
HOMO) of the chromophore, or by the formation of intra-
molecular hydrogen bonds between the α-proton and the
closed carbonyl group, resulting in a radiation less decay of the
excited state.[10,48] Therefore, pyrrolidine was used for the
substitution of 8 in the 2 and 6 positions to obtain nNDI in
almost quantitative yield as a blue solid. As already indicated by
the colour change of all three NDIs, the electron donating
character of substituents is also observed in 1H NMR spectra
which is associated with a general upfield shift of the NDI-core
protons with an increasing number of electron donating
groups.
Comparing the chemical environment of the NDI-core
protons, core-substitution with 2,4,6-trimethylthiophenol results
in a chemical shift of 8.04 ppm and 8.01 ppm for the NDI core
protons, while substitution with pyrrolidine leads to a chemical
shift of the core protons to 8.43 ppm and 8.40 (for NMR spectra
see Figures S22 (sNDI) and S28 (nNDI) in the Supporting
Information). The unsubstituted hNDI core protons show the
most deshielded chemical shift of 8.84 ppm (Figure S14).
The final assembly of the target molecules Tpd-hNDI, Tpd-
sNDI and Tpd-nNDI is outlined in Scheme 3. After successful
isolation of all NDIs, Sonogashira reaction was used to couple
hNDI, sNDI and nNDI respectively with the modular platform 4
to afford precursors 9–11 in 67, 46 and 78% yield, respectively.
THF was used as co-solvent for the Sonogashira cross-coupling
reaction because of poor solubility of the NDI-precursors in
neat amines (Et3N, etc.). Final transprotection of the thiols in 9–
11 was successfully performed using AgBF4 and acetyl chloride
in dichloromethane to afford the desired thioacetate masked
target structures Tpd-hNDI, Tpd-sNDI and Tpd-nNDI in good
yields. Note, that the acetyl protected target molecules in
solution at ambient conditions are only moderately stable.
Therefore, solvent evaporation must be carried out at room
temperature, otherwise thiol deprotection and subsequent
polymerisation to for example disulfides was observed. Also,
long retention times on column chromatography leads to a
significant decrease in isolated yield. The introduction of the
tripodal structure had no impact on the electronic structure of
the NDI chromophore, as very comparable photophysical
properties and chemical shifts of the NDI cores were observed
for precursors and target structures (see Figures S1, S14, S16
and S18 for hNDI, 9, Tpd-hNDI; Figures S2, S22, S24 and S26 for
sNDI, 10, Tpd-sNDI; Figures S3, S28, S30 and S32 for nNDI, 11,
Tpd-nNDI).[49] All target molecules were purified by chromatog-
raphy and fully characterised by conventional analytical and
spectroscopy techniques like NMR spectroscopy, mass spec-
trometry, IR, UV-Vis and florescence spectrometry, as well as by
elemental analysis. More details on the synthesis and purifica-
Scheme 2. Synthesis of NDI chromophores. i) 3,5-di-tert-butylaniline, p-
iodoaniline, AcOH, 43% (39% for compound 8); ii) for sNDI: 2,4,6-trimeth-
ylthiophenol, K2CO3, 69%; for nNDI: pyrrolidine, K2CO3, 99%.
acid dianhydride 7 was synthesised in four reaction steps by
following a literature procedure.[9] The twofold amide condensa-
tion with two different aniline derivatives to obtain asymmetric
naphthalene-1,4,5,8-tetracarboxdiimides turned out to be more
challenging than expected. In the case of unsubstituted hNDI,
the classical step-by-step approach was followed to obtain
hNDI with an overall yield of 43%. Unfortunately, in the case of
2,6-dichlorinated NDI 8, this approach was not successful
because such disubstituted-core derivatives are more soluble in
organic solvents and prone to nucleophilic substitutions at the
2 and 6 positions, resulting in considerable material loss due to
a number of core-substituted NDIs as side products. As
alternative a one-pot reaction with an equimolar stoichiometric
ratio of both anilines was performed, expecting a statistical
distribution of all possible products. To our delight, the desired
NDI derivative 8 was isolated as a yellow powder in 39% yield
with an optimised procedure. In particular, the reaction temper-
°
ature was decreased to room temperature from initially 100 C
and the anilines were deactivated by stirring for 30 minutes in
acetic acid prior their addition to the dianhydride 7. For the
amide condensation, aniline derivatives were chosen carefully
so that one has the iodine moiety attached in the para position
to further interlink the NDI chromophore with the tripodal
platform 4 through a Sonogashira protocol, while the second
one bears two tert-butyl groups in the meta positions providing
both, the solubility required for further processing and the
compactness favouring STM experiments. Although NDI solu-
bility could have also been increased by using (long and
branched) alkyl chains, their flexibility and bulkiness could
handicap the arrangement of the target molecules upright on
the metallic surface. Although benzylic core substituted NDIs
are known to be more fluorescent than their phenyl
counterparts,[48] 2,4,6-trimethylthiophenol and pyrrolidine were
selected as substituents while benzylic substituents might be
more prone to degradation in the final transprotection step.
Core-functionalisation of 8 with 2,4,6-trimethylthiophenol in a
nucleophilic aromatic substitution reaction gave a red sNDI in
69% yield. Methyl groups in the ortho positions of 2,4,6-
trimethylphenyl substituent are increasing its steric hindrance
Chem. Eur. J. 2021, 27, 1–13
www.chemeurj.org
4
© 2021 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH
��
These are not the final page numbers!

Full Paper
doi.org/10.1002/chem.202101264
Chemistry—A European Journal
Scheme 3. Synthesis of the target tripodal molecules. i) 4, CuI, Pd(PPh3)4, Et3N, THF, 67% (9), 46% (10) and 78% (11); ii) AcCl, AgBF4, CH2Cl2, 93% (Tpd-hNDI),
82% (Tpd-sNDI), 69% (Tpd-nNDI).
tion of all new compounds and their characterisation are
provided in the Experimental Section and in the Supporting
Information.
Spectrophotometric studies
UV/Vis absorption and fluorescence spectra were recorded in
dichloromethane with a concentration of about 25 μM and are
shown in Figure 2 and summarised in Table 1. The absorption
and emission properties of Tpd-hNDI, Tpd-sNDI and Tpd-nNDI
noticeably depend on the core substitution. Absorption bands
can be divided into three different parts. The central absorption
bands at around 350 nm, which show the π–π* transition and
their vibrational fine structure of the three different chromo-
phores, barely differ upon substitution, which is in accordance
with the literature.[10] Moreover, the peaks in the UV region
(~300 nm) do not significantly change upon core-substitution,
whereas they are missing in unsubstituted Tpd-hNDI.[13] The rise
of a new charge-transfer band in the visible region, which is
introduced by and strongly depending on the push-pull
character of the core substitution is associated with the colours.
Whereas unsubstituted Tpd-hNDI is pale yellow in CH2Cl2
solution, Tpd-sNDI is red (λmax =524 nm) and Tpd-nNDI is blue
(λmax =602 nm). The remarkable bathochromic shift in the
absorption maxima of the amino substituted NDI Tpd-nNDI can
be related to the stronger electron donating effect (+M effect)
of the pyrrolidine substituent on the frontier molecular orbitals
of the NDI core.[48] In case of Tpd-nNDI, this intramolecular
charge transfer can be hindered by mono protonation with HCl
or even switched off by total protonation and reversibly
restored.[50] The absorption maxima at 524 nm with a shoulder
at 492 nm (Tpd-sNDI) and at 602 nm with a shoulder at 558 nm
(Tpd-nNDI) are quite broad.[50] In contrast to previous reports,
our 2,4,6-trimethylphenylsulfanyl core-substituted NDI chromo-
phore shows fluorescence at 557 nm with a broad emission
band,[17,48] however the quantum yield is low and cannot be
reproducibly measured. For pyrrolidine core-substituted Tpd-
nNDI the emission maximum is observed at 627 nm. The Stoke
Figure 2. UV/Vis absorption spectra (solid line) of Tpd-hNDI (black), Tpd-
sNDI (red), Tpd-nNDI (blue) and normalised emission spectra (dotted line).
Recorded in CH2Cl2 with a concentration of 25 μM at ambient temperature.
A photograph of Tpd-hNDI (left), Tpd-sNDI (middle) and Tpd-nNDI (right)
solutions in CH2Cl2 is shown in the inset.
Table 1. UV/Vis absorption and emission characteristics of Tpd-hNDI, Tpd-
sNDI and Tpd-nNDI in CH2Cl2 at ambient temperature.
Molecule
λabs [nm] (ɛ [LmolÀ 1cmÀ 1])
λem [nm]
Tpd-hNDI 360 (24995), 381 (26976)
n.d.
Tpd-sNDI 297 (50638), 357 (10053), 370 (10620), 524 (19231) 557
Tpd-nNDI 298 (32790), 349 (9600), 365 (11463), 602 (17376) 627
Chem. Eur. J. 2021, 27, 1–13
www.chemeurj.org
5
© 2021 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH
��
These are not the final page numbers!

Full Paper
doi.org/10.1002/chem.202101264
Chemistry—A European Journal
shift, that is the difference between the absorption and
emission maxima, is 33 nm for the Tpd-sNDI chromophore and
25 nm for Tpd-nNDI chromophore.[3] The emission intensity and
the fluorescence quantum yield significantly decreases from
Tpd-nNDI to Tpd-sNDI and no fluorescence has been observed
for unsubstituted Tpd-hNDI derivative.
electrochemical band gap from peak values is mainly based on
the detailed quantum chemical calculations of Brédas et al. who
found a linear correlation between ionisation potential and
oxidation potential and also between electron affinity and
reduction potentials.[53] Note, that small differences between
optical HOMO/LUMO gap and the electrochemical gap are
expected not only because of the intrinsic experimental
uncertainties but also because charged molecules are com-
pared to neutral ones.[54]
Electrochemical properties
All three NDIs show two reversible reduction processes,
implying the formation of radical anions and dianions, which
are separated by 0.34, 0.43 and 0.48 V for nNDI, hNDI and sNDI
respectively. In case of nNDI, the electron donating core
substitution, namely pyrrolidine, results in destabilising the
radical anion and shifting the reduction to the more negative
potentials. A gradual shift of the reduction potentials to more
negative values for the core substituted NDIs reveals a direct
relationship between the electron-donating properties of the
core substituents and the electrochemical properties of NDIs.
Additional irreversible oxidation waves were observed for nNDI
at around +0.6 and +0.8 V against Fc/Fc+ (Figure S4). To
prevent irreversible oxidations, the CV was carried out only in
the cathodic region. Thus, both hNDI and sNDI chromophores
were also measured between À 2 and 0 V. The amino
substituted nNDI chromophore showed reduction potentials at
À 1.43 and À 1.77 V that are in the same range as previously
observed for the similar amino core-substituted NDIs.[49,50,55] The
LUMO energy estimated from the CV is À 3.37 eV resulting in
the HOMO energy level of À 5.31 eV, which is comparable to
previously reported results.[50] However, each of the two anodic
and cathodic peak potential pairs differ for about 0.3 V.
Electrode films of some newly formed species as an explanation
for these broad peaks can be excluded, since the experiments
were carried out with freshly cleaned electrodes and are
reproducible over several measurements. As seen from the
cyclic voltammograms in Figure 3, the sulfanyl core substitution
in sNDI decreases the electron donating character compared to
the amino derivative nNDI. Hence the reduction potential
values (À 1.14 and À 1.62 V) of sNDI are shifted towards more
positive potentials, corresponding to a LUMO energy level of
À 3.72 eV and a HOMO energy level of À 5.96 eV. The obtained
values are comparable with previously published results on
alkylsulfanyl and arylsulfanyl core-substituted NDIs.[56,57] Within
this series, the core-unsubstituted compound hNDI exhibits the
most positive reduction potentials of À 1.10 and À 1.53 V.
Comparing the sNDI and hNDI chromophores, the donating
character of the arylsulfanyl group is clearly observed in the
Cyclic voltammetry (CV) in CH2Cl2 against ferrocene/ferrocenium
(Fc/Fc+) was conducted to further investigate the influence of
core substitution upon the electrochemical properties of all
chromophores hNDI, sNDI and nNDI (Figures 3 and S4, Table 2).
The half-wave potential values E1/2 were calculated as the
arithmetical mean of the anodic and the cathodic peak
potentials. To calculate the LUMO energy levels (against
red1
onset
vacuum), the onset reduction potentials E
versus the Ag/
Ag+ reference electrode of each NDI were taken and subtracted
from the Fc/Fc+ standard (using À 4.8 eV). It is assumed that the
redox potential of Fc/Fc+ has an absolute energy level of 4.8 eV
to vacuum.[51] The HOMO energy is determined from the LUMO
level minus the HOMO/LUMO gap, which is available from
absorption spectra at the long wavelength absorption edge
opt
(E ¼ 1240=lonset).[52] Such determination of optical as well as
gap
Figure 3. Cyclic voltammograms in the cathodic region of hNDI (black), sNDI
(red) and nNDI (blue) recorded in CH2Cl2 containing a 0.1 M solution of
Bu4NPF6 as a supporting electrolyte. Fc/Fc+ was used as an internal
reference. Scan rate: 100 mVsÀ 1
.
*
À
second reduction NDI /NDI2À but significantly less in the first
Table 2. Electrochemical potentials, energy levels and bandgaps of all NDI
chromophores. Redox potentials E1/2 were obtained from the arithmetical
mean of anodic and cathodic peaks. The LUMO energies (for Fc/Fc+
standard against vacuum) were calculated from the onset of reduction
waves using the empirical equation: ELUMO =(À 4.8 eVÀ Eroends1etÞ: The optical
*
À
reduction NDI/NDI . This suggests that arylsulfanyl as a core
substituent destabilises the dianion more than the radical
anion. The LUMO energy level of hNDI is calculated to be
À 3.80 eV, which results in the HOMO energy of À 6.97 eV. The
HOMO–LUMO optical gaps were calculated from the onset of
the lowest energy absorptions to approximate the HOMO
energies and the obtained values are listed in Table 2.
band gap was calculated from the onset of the lowest energy absorption
opt
as follows: E ¼ 1240=lonset and EHOMO =ELUMOÀ Eogpapt
:
gap
red1
red2
opt
gap
Chromophore
E
E
ELUMO
E
EHOMO
1=2
1=2
hNDI
sNDI
nNDI
À 1.10 V
À 1.14 V
À 1.43 V
À 1.53 V
À 1.62 V
À 1.77 V
À 3.80 eV
À 3.72 eV
À 3.37 eV
3.17 eV
2.24 eV
1.94 eV
À 6.97 eV
À 5.96 eV
À 5.31 eV
Chem. Eur. J. 2021, 27, 1–13
www.chemeurj.org
6
© 2021 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH
��
These are not the final page numbers!

Full Paper
doi.org/10.1002/chem.202101264
Chemistry—A European Journal
Scanning tunnelling microscopy induced luminescence
experiments
tigation. In addition, there might be different charge states of
the molecule involved which drastically influence the emission
energy.[25,37] Still, at fixed positions, the junction is stable enough
to perform extended measurements.
The surface behaviour of all three tripodal molecules bearing
NDI chromophores was investigated using STML at 5 K in UHV.
Spray deposition of about 1 μL of Tpd-hNDI solution (c~1 mg/
mL, for details see the Experimental Section) with a post
This allows to record photon spectra as a function of the
applied sample bias which defines the energy of the tunnelling
electrons. Figure 4d shows such spectra recorded at the
position marked by the red cross in Figure 4b. In this experi-
ment, light emission sets in for a sample bias above +1.8 V and
below À 2 V with photon energies of less than 1.7 eV, reflecting
the asymmetry of the STM junction under reversion of the
polarity of the applied voltage.[22,63] For low currents and low
temperatures, the high energy edge of a photon spectrum of
plasmonic origin is expected to be limited by the energy
provided by the tunnelling electrons: eV=hν. However, the
threshold values for light emission observed in Figure 4d are
clearly higher than the high energy edge of the photon spectra,
which is a clear sign of molecular contribution to the light
emission. In addition, both peaks emerge at the same bias
voltage, which indicates that the redshifted peak can be
attributed to the same molecular transition with an additional
excitation of a molecular vibration,[26,60] eV=hν+�hω. Indeed,
most photon spectra recorded on the unordered clusters of
Tpd-hNDI show a series of peaks with a typical spacing of
about 160 meV (see also Figure S34c), in full agreement with a
recent study on tetrapodal perylene diimides (PDIs) on Au-
(111).[33] Rarely, light emission above the quantum threshold
has been observed, even at low currents, which is an indication
for intramolecular up-conversion (Figure S34b).[61]
°
annealing temperature of 100 C leads to the formation of
unordered islands of about 0.3 nm in height (Figure 4a). This
height indicates that the molecules are not adsorbed as
intended but rather adsorb flat on the metal surface. While the
islands seem to be composed of identical objects with clear
substructure, the lack of any long-range order makes an
assignment of the molecular adsorption impossible. When the
STM tip is placed above these islands, plasmonic light emission
which is characteristic for noble metal nanogaps,[21] is partially
or fully suppressed and the molecular layer merely acts as a
dielectric layer that lowers the emission intensity.[58] At some
positions, molecular clusters of more than 1 nm in height
protrude from the molecular islands (Figure 4b). STM images of
these clusters typically show white streaks and abrupt changes
in the apparent height from one horizontal scan line to the
next. These are signs of instability of the tunnel junction and
can be explained by multilayers of Tpd-hNDI molecules of
unknown orientation or by standing molecules in a less stable
configuration as compared to the ordered monolayer. On these
clusters and in the close vicinity, some light emission spectra
are clearly different from a plasmonic light emission. We
recorded 2500 photon spectra at different positions across the
two clusters as depicted in Figure 4b. At most positions, the
photon spectra are similar to the grey spectrum shown in
Figure 4c recorded at the position indicated by the grey cross in
Figure 4b, with a broad peak centred around 1.5 eV. While the
precise energy of the plasmonic resonance might shift depend-
ing on the exact position and the shape of the tip apex,[19,59] the
FWHM of about 300 meV indicates a plasmonic origin. About
200 out of 2500 spectra recorded on this area have a clearly
different appearance and exhibit one or more sharper peaks of
30 to 100 meV FWHM to which we therefore tentatively assign
a molecular origin. Six examples of such photon spectra are
shown in Figure 4c (coloured photon spectra). The peak
position ranges from 1.5 to 2.05 eV within the measurement on
the area shown in Figure 4b. The quantum yield, that is the
number of photons emitted per tunnelling electron, ranges
from 1×10À 4 to 1×10À 3 for an applied sample bias of À 2.5 V.
Neither the spectral shape nor the quantum yield can be
definitely related to the position where the spectrum has been
recorded (that is the topography) and small displacements of
the STM tip may lead to drastic variations of the photon
spectrum (compare, e.g., the grey and dark blue spectra in
Figure 4c). This can be explained by the large number of
possible adsorption configurations and the absence of long-
range order which lead to a number of different electronic
environments for each molecule. The orientation with respect
to the substrate, the orientation with respect to the electric
field present in the STM junction and the coupling to
neighbouring molecules varies for each molecule under inves-
The combination of our highly efficient photon collection
setup[20] and an efficient conversion process in the molecular
junction results in unusually high count rates observed for Tpd-
hNDI. The spectrum recorded at 2.5 V and 5.2 pA shown in
Figure 4d corresponds to 1.3×10À 3 photons counted per
tunnelling electron, which is, to best of our knowledge, about
one order of magnitude higher than the highest external
quantum yields reported so far.[19,62] After correction by the
detector efficiency, this results in an internal quantum yield of
about 6×10À 3.
After annealing, the Tpd-hNDI sample for a second time, at
°
180 C, ordered island structures form. Figure 5a shows a typical
island which exhibits a monoclinic lattice structure with a unit
2
°
cell (red lines) of 2.7 nm×4.04 nm (α=53 )=8.6 nm . The
apparent height of about 0.3 nm of these islands (see cross
section along the white line in Figure 5a) is similar to that of the
unordered ones discussed above and indicates that the Tpd-
hNDI molecules adsorb in
a horizontal configuration. In
comparison, at similar voltages, a naphthalene diimide cyclo-
phane “double decker” has been reported to show an apparent
height of about 0.7 nm.[18] Although the precise adsorption
configuration is difficult to infer from the topographic image,
we suggest a pairwise alignment as outlined in Figure 5b so
that each bright blob corresponds to one molecule. This would
result in an area of 4.3 nm2 covered per molecule.
Similar to the unordered flat islands, plasmonic light
emission is strongly suppressed when the tip is placed on top
of the ordered islands. Figure 5c shows examples of photon
Chem. Eur. J. 2021, 27, 1–13
www.chemeurj.org
7
© 2021 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH
��
These are not the final page numbers!

Full Paper
doi.org/10.1002/chem.202101264
Chemistry—A European Journal
Figure 4. Light emission from unordered clusters of Tpd-hNDI. a) Typical island structure of Tpd-hNDI with clean (black) areas of Au(111) in between. The
cross section taken along the white line reveals an apparent height of the island of about 0.3 nm. The molecular model in the proposed orientation is not to
scale. Sample bias: 1.8 V, tunnelling current: 5 pA. The scan frame is 30 nm wide. b) Two clusters of Tpd-hNDI (bright areas) with an apparent height of 1.2 nm
(see cross section along the white line) above the island (darker areas). Sample bias: À 2.3 V, tunnelling current: 5 pA. The scan frame is 50 nm wide. c), d)
Photon spectra recorded at the positions indicated by the corresponding crosses in (b). The applied sample bias, the internal quantum yield (count rate
corrected for detector efficiency), the vertical z-position of the tip and average current during the acquisition are indicated in each panel. Integration time is
2 s. Bias voltage, yield and average current during the acquisition as indicated in the graphs c), d).
spectra with the tip placed above Au(111) (red curve) and
placed above the molecular island (black curve). While there are
expected small intensity variations superimposed to a broad
plasmonic spectrum in both curves, sharp peaks of molecular
origin as described in Figure 4c are absent.
Chem. Eur. J. 2021, 27, 1–13
www.chemeurj.org
8
© 2021 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH
��
These are not the final page numbers!

Full Paper
doi.org/10.1002/chem.202101264
Chemistry—A European Journal
Figure 5. Ordered structure of Tpd-hNDI. a) Ordered island of Tpd-hNDI with a unit cell indicated in red. Black areas correspond to empty Au(111) surface.
The cross section taken along the white line shows an apparent height of about 0.3 nm. Molecular model not to scale. Sample bias: 2.1 V, tunnelling current:
9 pA, scan width: 22 nm. b) Close-up scan of an ordered island with the unit cell indicated in red and molecular models of Tpd-hNDI superimposed to scale in
the suggested adsorption configuration. Sample bias: À 2.1 V, tunnelling current: 5 pA, scan width: 8 nm. c) Typical photon spectra recorded on the ordered
islands (black) and the nearby Au(111) surface (red). Sample bias: 3 V, tunnelling current: 140 pA, integration time: 2 s.
Similar to Tpd-hNDI, the core-substituted variant Tpd-sNDI
forms unordered islands of about 0.3 nm in apparent height as
Figure 6d reveals a rich sub-molecular structure. We propose an
adsorption configuration similar to the case of Tpd-hNDI
(Figure 5b): Molecular pairs arrange in chains with the foot
group pointing towards the next pair (Figure 6e). Molecules of
adjacent chains are mirrored. A closer look at the edges of the
molecular island shows one quarter of the unit cell as the
smallest unit, which confirms the hypothesis of pairwise
adsorption (Figure S34d). In this rather tightly packed config-
°
is shown in Figure 6a at a post annealing temperature of 100 C.
In addition, these islands act as a dielectric that reduces light
emission from the decay of local gap plasmons, similar to
Figure 5c. While we cannot exclude formation of clusters at
°
100 C, we only observed such clusters of increased height after
°
annealing a second time, to 180 C. Figure 6b shows an example
of a cluster that exceeds the surrounding island by 0.2 to
0.4 nm in apparent height. With the STM tip placed above these
clusters, we observe light emission with two pronounced peaks
that are typically separated by about 160 to 170 meV. We have
seen the same spacing for the spectra of Tpd-hNDI (Figures 4c
and S34c) which corresponds to the energy of the CÀ N stretch
mode. Therefore, we assume that also in the case of Tpd-sNDI
the red shifted peak originates from a decay from the first
electronically excited state to the vibrationally excited electronic
ground state. Apparently, the local surrounding has a significant
influence on both the electroluminescence efficiency and the
energy of the emitted light. In agreement with the recent study
on PDI tripodal molecules,[33] the molecular configuration seems
to be easily influenced by the nearby tip which hampers
uration, the unit cell with an area of 3.3 nm × 5.0 nm (α=
2
47.6 )=12.2 nm contains 4 molecules (=3.05 nm2 per mole-
cule).
°
The third NDI tripod, Tpd-nNDI, also forms unordered
islands of about 0.3 nm in apparent height at a post annealing
°
temperature of 100 C, thus suggesting that also for these
molecular complexes the horizontal arrangement is preferred
(Figure S34e). In this horizontal arrangement, the optical activity
of the chromophore is expected to be quenched and indeed no
molecular features in the photon spectra were observed.
Formation of clusters consisting of molecules in more upright
configurations were not observed.
reproducibility of STML experiments. The wavelength shift of Conclusions
the substituted NDI molecules compared to the unsubstituted
chromophore as it is observed in our PL measurements, does
not translate to a shift of the electroluminescence photon
spectra of the molecules deposited on the Au(111) surface.
Similar to Tpd-hNDI, Tpd-sNDI forms well-ordered islands at
In summary, three tripodal model compounds consisting of a
triphenylmethane platform with three acetyl-protected thiol
anchors in the meta position substituted with different core-
substituted NDI chromophores linked via alkyne linker have
°
elevated post-annealing temperatures of 180 C (Figure 6d). The
been synthesised and fully characterised. In a modular
apparent height of 0.3 nm suggests that these islands are
composed of flat-lying molecules, which is in agreement with
the absence of any molecular signature in the photon spectra
on these islands. A close-up scan of an ordered island shown in
approach, the alkyne-decorated rigid tripodal platform was
fused with various NDI chromophores by Sonogashira coupling.
For this purpose, asymmetric NDI subunits were developed, and
their electronic properties were tuned by their core substitu-
Chem. Eur. J. 2021, 27, 1–13
www.chemeurj.org
9
© 2021 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH
��
These are not the final page numbers!

Full Paper
doi.org/10.1002/chem.202101264
Chemistry—A European Journal
Figure 6. Adsorption and electroluminescence of Tpd-sNDI. a) Unordered island structure of Tpd-sNDI. The cross section taken along the white line shows a
typical height of 0.3 to 0.4 nm. Molecular model not to scale. Sample bias: À 1.4 V, tunnelling current: 2 pA, scan width: 15 nm. b) Unordered cluster of Tpd-
sNDI with a maximal apparent height of 0.4 nm above the surrounding island (see cross section along the white line). Sample bias: 2.4 V, tunnelling current:
5 pA, scan width: 4 nm. c) Typical photon spectra recorded at the positions indicated by the blue and red crosses in (b). Integration times of 4 s were used.
Bias voltage, yield and average current during the acquisition as indicated in the graph c). d) Molecular model not to scale. Sample bias: 1.8 V, tunneling
current 3 pA. (e) Close up scan of an ordered island with molecular models superimposed to scale. Adjacent rows are mirrored and translated along the
dashed white line. The red line indicates a possible unit cell. Sample bias: 2.6 V, tunneling current 8 pA.
ents. Structure-property correlations revealed the influence of
the core substituents on optical and electrochemical properties
of these model compounds.
configuration as suggested by their arrangement in flat islands.
Both Tpd-hNDI and Tpd-sNDI molecules arrange in a close-
packed pairwise fashion and form islands with a long-range
°
Moreover, the self-assembly features of all three tripodal
chromophores were analysed after spray deposition on Au
(111) surfaces by low-temperature ultra-high-vacuum STM with
the electroluminescence setup. These surface studies demon-
strated that all tripodal molecules can be deposited in sub-
monolayer concentration by “spraying” techniques, and after an
annealing procedure, tend to mainly adsorb in a horizontal
lateral order upon annealing at 180 C. At some positions, these
molecules formed clusters with an apparent height clearly
exceeding that of the surrounding islands. According to the
apparent height, these clusters are composed of either multi-
layers of horizontally aligned molecules, or a monolayer of
upright oriented molecules in a less stable configuration. The
lateral size clearly suggests that these clusters are composed of
Chem. Eur. J. 2021, 27, 1–13
www.chemeurj.org
10
© 2021 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH
��
These are not the final page numbers!

Full Paper
doi.org/10.1002/chem.202101264
Chemistry—A European Journal
each series of measurements. The junction potential was corrected
to the reference electrode afterwards. The molecules (0.1 mM) were
dissolved in 0.1 M solution of Bu4NPF6 as an electrolyte in CH2Cl2,
and the solution was purged with argon for 10 min before the
measurement was taken. Scan rate of the CV measurements was
100 mVsÀ 1 and all experiments were performed under ambient
conditions at room temperature. All STM measurements were
performed in a homebuilt STM at about 4.4 K in UHV (10À 10 mbar).
This STM is equipped with a photon collection system that allows
to analyse light that is emitted from the tunnel junction.[20] All
photon spectra were corrected for the energy-dependent detector
efficiency (Figure S34a). Electroluminescence count rates are nor-
malised by energy resolution and charge injected (labelled as
counts per meV and nC). A clean Au(111) substrate was prepared
by argon sputtering at 1.5 kV and subsequent annealing to about
several molecules. In photon spectra, these clusters of mole-
cules showed sharp peaks that are clearly distinct from the
plasmonic light of the same tip. Extensive low-temperature
UHV-STML studies of these molecules deposited on Au(111) by
spray deposition revealed that, at most, a minority of the
tripodal chromophores are standing upright and show lumines-
cence; the majority are lying flat on the gold substrate. Even so,
only a small minority of the spray-deposited molecules arrange
in luminescent clusters, their recorded high electrolumines-
cence yields of up to 6×10À 3 in the STML experiment not only
support the hypothesis of upright standing and decoupled NDI
chromophores, but also showcase their potential in highly
efficient OLED devices.
°
500 C. Molecules were dissolved in CH2Cl2 (c ~1 mg/mL) and an
Our current work is thus geared towards increasing control
over the surface geometry of these chromophores by improving
the ordering power of the tripodal architecture.
amount of about 1 μL was being sucked into a vacuum chamber (p
~10À 3 mbar) through an aperture of about 0.15 mm. The clean gold
surface placed inside the chamber was exposed to the spray
formed at the aperture. After deposition, samples were transferred
in situ into the ultra-high vacuum (UHV) and were annealed at
°
100/180 C in order to promote deprotection of the thiol anchoring
Experimental Section
groups and to facilitate ordered arrangements of the molecules.[47]
This deposition technique has two major advantages for STM
studies: First, the amount of impurities from the solvent is
minimised by reducing the amount of solvent deposited onto the
sample. Second, deposition of the solution by spraying results in a
number of tiny droplets landing randomly across the sample. This
leads to an inhomogeneous local coverage, higher in the centre of
those former droplets, lower towards the borders. By moving the
scan area of our STM, we are able to study different local coverages
within one sample (sample size of 9×3 mm). Post annealing leads
to a more uniform adsorption configuration on the local scale, but
still the coverage is highly variable across the sample. The classical
incubation method by immersing gold substrates in thiol solutions
for a few hours turned out not to be suitable for our UHV-STM
experiments.[40]
Materials: All starting materials and reagents were purchased from
commercial suppliers (Alfa Aesar, Sigma–Aldrich, TCI Chemicals
Europe (Zwijndrecht, Belgium), Merck and used without further
purification. Solvents utilised for crystallisation, chromatography
and extraction were used in technical grade. Anhydrous tetrahy-
drofuran and CD2Cl2 were taken from MBraun Solvent Purification
System equipped with drying columns. Triethylamine was dried
and distilled from CaH2 under nitrogen. TLC was performed on silica
gel 60 F254 plates, spots were detected by fluorescence quenching
under UV light at 254 and 366 nm. Column chromatography was
performed on silica gel 60 (particle size 0.040–0.063 mm).
Equipment and measurements: All NMR spectra were recorded on
a Bruker Avance 500 spectrometer at 25 C in CDCl3, CD2Cl2 or
°
MeOD. 1H NMR (500 MHz) spectra were referred to the solvent
residual proton signal (CDCl3, δH =7.26 ppm; CD2Cl2, δH =5.32 ppm;
MeOD, δH =3.31 ppm). 13C NMR (126 MHz) with total decoupling of
protons were referred to the solvent (CDCl3, δC =77.16 ppm; CD2Cl2,
δC =53.84 ppm; MeOD, δC =49.00 ppm). UV-Vis-NIR absorption
spectra were recorded with a Cary 5000 Scan spectrophotometer in
a 1 cm quartz cell at ambient temperature (excitation coefficient ɛ
[LmolÀ 1 cmÀ 1] is given below). Fluorescence spectra were measured
with a Varian Cary Eclipse Fluorescence spectrometer at room
temperature in a 1 cm quartz cell. The molecules sNDI, 10 and Tpd-
sNDI were excited at 370 nm and both excitation and emission slits
were set to 20 nm. The molecules nNDI, 11 and Tpd-nNDI were
excited at 366 nm and both excitation and emission slits were set
to 10 nm. EI MS spectra were recorded with a Thermo Trace 1300-
ISQ GC/MS instrument (samples were dissolved in dichloromethane
or introduced directly using direct injection probes DIP, DEP) and
m/z values are given along with their relative intensities [%] at an
ionisation voltage of 70 eV. High-resolution mass spectra were
recorded with a Bruker Daltonics (ESI microTOF-QII) mass spectrom-
eter. IR spectra were recorded with a Nicolet iS50 FTIR spectrometer
Experimental details and synthetic procedures: The synthesis of
3,3,3-tris(3-methoxyphenyl)propyne 1 was previously reported by
Garcia-Garibay et al.[44] 2,6-Dichloronaphthalene-1,4,5,8-tetracarbox-
ylic acid dianhydride 7 was synthesised in four reaction steps by
following a literature procedure.[9] Detailed synthetic procedures
along with complete characterisation data and copies of NMR
spectra for all new compounds are given in the Supporting
Information.
Acknowledgements
We acknowledge financial support by the Helmholtz Research
Program STN (“Science and Technology of Nanosystems”), DFG
grant (MA 2605/6-1) and the DAAD grant (57299294). The
authors also thank the Institute of Nanotechnology (INT),
Karlsruhe Institute of Technology (KIT) for ongoing support.
M.M. acknowledges support by the 111 Project (90002-
18011002). Open access funding enabled and organized by
Projekt DEAL.
°
under ATP mode. Analytical samples were dried at 40–100 C under
reduced pressure (10À 2 mbar). Melting points were measured with a
Büchi Melting point M-560 apparatus and are uncorrected.
Elemental analyses were obtained with a Vario MicroCube CHNS
analyser. The values are expressed in mass percentage. Cyclic
voltammetry (CV) experiments were carried out with a Gamry
potentiostat, connected to a C3 standard electrochemical cell.
Glassy carbon was used as the working electrode, a platinum wire
as the counter electrode and Ag/AgNO3 as the reference electrode.
Ferrocene was used as an internal standard and was added after
Conflict of Interest
The authors declare no conflict of interest.
Chem. Eur. J. 2021, 27, 1–13
www.chemeurj.org
11
© 2021 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH
��
These are not the final page numbers!

Full Paper
doi.org/10.1002/chem.202101264
Chemistry—A European Journal
[34] S. E. Zhu, Y. M. Kuang, F. Geng, J. Z. Zhu, C. Z. Wang, Y. J. Yu, Y. Luo, Y.
Xiao, K. Q. Liu, Q. S. Meng, L. Zhang, S. Jiang, Y. Zhang, G. W. Wang, Z. C.
Dong, J. G. Hou, J. Am. Chem. Soc. 2013, 135, 15794–15800.
[35] F. Matino, G. Schull, F. Köhler, S. Gabutti, M. Mayor, R. Berndt, Proc. Natl.
Acad. Sci. USA 2011, 108, 961–964.
Keywords: cyclic voltammetry
naphthalene diimides · scanning probe microscopy · tripodal
platforms
·
electroluminescence
·
[36] F. L. Otte, S. Lemke, C. Schütt, N. R. Krekiehn, U. Jung, O. M. Magnussen,
R. Herges, J. Am. Chem. Soc. 2014, 136, 11248–11251.
[37] V. Rai, L. Gerhard, Q. Sun, C. Holzer, T. Repän, M. Krstić, L. Yang, M.
Wegener, C. Rockstuhl, W. Wulfhekel, Nano Lett. 2020, 20, 7600–7605.
[38] M. Lindner, M. Valášek, M. Mayor, T. Frauhammer, W. Wulfhekel, L.
Gerhard, Angew. Chem. Int. Ed. 2017, 56, 8290–8294; Angew. Chem.
2017, 129, 8405–8410.
[39] T. Frauhammer, L. Gerhard, K. Edelmann, M. Lindner, M. Valášek, M.
Mayor, W. Wulfhekel, Phys. Chem. Chem. Phys. 2021, 23, 4874–4881.
[40] M. Lindner, M. Valášek, J. Homberg, K. Edelmann, L. Gerhard, W.
Wulfhekel, O. Fuhr, T. Wächter, M. Zharnikov, V. Kolivoška, L. Pospíšil, G.
Mészáros, M. Hromadová, M. Mayor, Chem. Eur. J. 2016, 22, 13218–
13235.
[41] J. Homberg, M. Lindner, L. Gerhard, K. Edelmann, T. Frauhammer, Y.
Nahas, M. Valášek, M. Mayor, W. Wulfhekel, Nanoscale 2019, 11, 9015–
9022.
[42] H. Valkenier, E. H. Huisman, P. A. Van Hal, D. M. De Leeuw, R. C. Chiechi,
J. C. Hummelen, J. Am. Chem. Soc. 2011, 133, 4930–4939.
[43] T. Sakata, S. Maruyama, A. Ueda, H. Otsuka, Y. Miyahara, Langmuir 2007,
23, 2269–2272.
[44] J. E. Nuñez, T. A. V. Khuong, L. M. Campos, N. Farfán, H. Dang, S. D.
Karlen, M. A. Garcia-Garibay, Cryst. Growth Des. 2006, 6, 866–873.
[45] J. E. Nuñez, A. Natarajan, S. I. Khan, M. A. Garcia-Garibay, Org. Lett. 2007,
9, 3559–3561.
[46] G. I. Feutrill, R. N. Mirrington, Tetrahedron Lett. 1970, 1327–1328.
[47] M. Valášek, K. Edelmann, L. Gerhard, O. Fuhr, M. Lukas, M. Mayor, J. Org.
Chem. 2014, 79, 7342–7357.
[48] A. Błaszczyk, M. Fischer, C. Von Hänisch, M. Mayor, Helv. Chim. Acta
2006, 89, 1986–2005.
[1] G. Hong, X. Gan, C. Leonhardt, Z. Zhang, J. Seibert, J. M. Busch, S. Bräse,
Adv. Mater. 2021, 33, 2005630.
[2] U. Mitschke, P. Bäuerle, J. Mater. Chem. 2000, 10, 1471–1507.
[3] N. Sakai, J. Mareda, E. Vauthey, S. Matile, Chem. Commun. 2010, 46,
4225–4237.
[4] M. Al Kobaisi, S. V. Bhosale, K. Latham, A. M. Raynor, S. V. Bhosale, Chem.
Rev. 2016, 116, 11685–11796.
[5] J. Hak, S. Sabin-Lucian, W. Y. Lee, M. Könemann, H. W. Höffken, C. Röger,
R. Schmidt, Y. Chung, W. C. Chen, F. Würthner, Z. Bao, Adv. Funct. Mater.
2010, 20, 2148–2156.
[6] F. Würthner, M. Stolte, Chem. Commun. 2011, 47, 5109–5115.
[7] M. Korzec, S. Kotowicz, K. Łaba, M. Łapkowski, J. G. Małecki, K. Smolarek,
S. Maćkowski, E. Schab-Balcerzak, Eur. J. Org. Chem. 2018, 1756–1760.
[8] H. F. Higginbotham, P. Pander, R. Rybakiewicz, M. K. Etherington, S.
Maniam, M. Zagorska, A. Pron, A. P. Monkman, P. Data, J. Mater. Chem. C
2018, 6, 8219–8225.
[9] H. Vollmann, H. Becker, M. Corell, H. Streeck, Liebigs Ann. 1937, 531, 159.
[10] F. Würthner, S. Ahmed, C. Thalacker, T. Debaerdemaeker, Chem. Eur. J.
2002, 8, 4742–4750.
[11] S. Erten, Y. Posokhov, S. Alp, S. Içli, Dyes Pigm. 2005, 64, 171–178.
[12] D. Shukla, S. F. Nelson, D. C. Freeman, M. Rajeswaran, W. G. Ahearn,
D. M. Meyer, J. T. Carey, Chem. Mater. 2008, 20, 7486–7491.
[13] C. Thalacker, C. Röger, F. Würthner, J. Org. Chem. 2006, 71, 8098–8105.
[14] S. L. Suraru, F. Würthner, Angew. Chem. Int. Ed. 2014, 53, 7428–7448;
Angew. Chem. 2014, 126, 7558–7578.
[15] R. E. Dawson, A. Hennig, D. P. Weimann, D. Emery, V. Ravikumar, J.
Montenegro, T. Takeuchi, S. Gabutti, M. Mayor, J. Mareda, C. A. Schalley,
S. Matile, Nat. Chem. 2010, 2, 533–538.
[49] R. Fernando, Z. Mao, E. Muller, F. Ruan, G. Sauvé, J. Phys. Chem. C 2014,
[16] S. Guha, F. S. Goodson, L. J. Corson, S. Saha, J. Am. Chem. Soc. 2012, 134,
118, 3433–3442.
13679–13691.
[50] S. Chopin, F. Chaignon, E. Blart, F. Odobel, J. Mater. Chem. 2007, 17,
4139–4146.
[51] J. Pommerehne, H. Vestweber, W. Guss, R. F. Mahrt, H. Bässler, M.
Porsch, J. Daub, Adv. Mater. 1995, 7, 551–554.
[52] R. S. K. Kishore, O. Kel, N. Banerji, D. Emery, G. Bollot, J. Mareda, A.
Gomez-Casado, P. Jonkheijm, J. Huskens, P. Maroni, M. Borkovec, E.
Vauthey, N. Sakai, S. Matile, J. Am. Chem. Soc. 2009, 131, 11106–11116.
[53] J. L. Bredas, R. Silbey, D. S. Boudreaux, R. R. Chance, J. Am. Chem. Soc.
1983, 105, 6555–6559.
[54] C. Q. Ma, M. Fonrodona, M. C. Schikora, M. M. Wienk, R. A. J. Janssen, P.
Bäuerle, Adv. Funct. Mater. 2008, 18, 3323–3331.
[17] C. W. Marquardt, S. Grunder, A. Błaszczyk, S. Dehm, F. Hennrich, H. V.
Löhneysen, M. Mayor, R. Krupke, Nat. Nanotechnol. 2010, 5, 863–867.
[18] N. L. Schneider, F. Matino, G. Schull, S. Gabutti, M. Mayor, R. Berndt,
Phys. Rev. B 2011, 84, 153403.
[19] K. Kuhnke, C. Große, P. Merino, K. Kern, Chem. Rev. 2017, 117, 5174–
5222.
[20] K. Edelmann, L. Gerhard, M. Winkler, L. Wilmes, V. Rai, M. Schumann, C.
Kern, M. Meyer, M. Wegener, W. Wulfhekel, Rev. Sci. Instrum. 2018, 89,
123107.
[21] R. Berndt, J. K. Gimzewski, P. Johansson, Phys. Rev. Lett. 1991, 67, 3796–
3799.
[55] R. Fernando, F. Etheridge, E. Muller, G. Sauvé, New J. Chem. 2015, 39,
2506–2514.
[56] Y. Zhao, G. Huang, C. Besnard, J. Mareda, N. Sakai, S. Matile, Chem. Eur.
[22] Y. Zhang, Y. Luo, Y. Zhang, Y. J. Yu, Y. M. Kuang, L. Zhang, Q. S. Meng, Y.
Luo, J. L. Yang, Z. C. Dong, J. G. Hou, Nature 2016, 531, 623–627.
[23] H. Imada, K. Miwa, M. Imai-Imada, S. Kawahara, K. Kimura, Y. Kim, Phys.
Rev. Lett. 2017, 119, 1–6.
[24] 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, 50–54.
[25] B. Doppagne, M. C. Chong, H. Bulou, A. Boeglin, F. Scheurer, G. Schull,
Science 2018, 361, 251–255.
[26] X. H. Qiu, G. V. Nazin, W. Ho, Science 2003, 299, 542–546.
[27] R. Berndt, R. Gaisch, J. K. Gimzewski, B. Reihl, R. R. Schlittler, W. D.
Schneider, M. Tschudy, Science 1993, 262, 1425–1427.
[28] Y. M. Kuang, Y. J. Yu, Y. Luo, J. Z. Zhu, Y. Liao, Y. Zhang, Z. C. Dong, Chin.
J. Chem. Phys. 2016, 29, 157–160.
[29] J. Repp, G. Meyer, S. M. Stojković, A. Gourdon, C. Joachim, Phys. Rev.
Lett. 2005, 94, 1–4.
J. 2015, 21, 6202–6207.
[57] C. Röger, F. Würthner, J. Org. Chem. 2007, 72, 8070–8075.
[58] X. Tao, Z. C. Dong, J. L. Yang, Y. Luo, J. G. Hou, J. Aizpurua, J. Chem. Phys.
2009, 130, 084706.
[59] J. Aizpurua, S. P. Apell, R. Berndt, Phys. Rev. B 2000, 62, 2065–2073.
[60] B. Doppagne, M. C. Chong, E. Lorchat, S. Berciaud, M. Romeo, H. Bulou,
A. Boeglin, F. Scheurer, G. Schull, Phys. Rev. Lett. 2017, 118, 127401.
[61] G. Chen, Y. Luo, H. Gao, J. Jiang, Y. Yu, L. Zhang, Y. Zhang, X. Li, Z.
Zhang, Z. Dong, Phys. Rev. Lett. 2019, 122, 177401.
[62] L. Zhang, Y. J. Yu, L. G. Chen, Y. Luo, B. Yang, F. F. Kong, G. Chen, Y.
Zhang, Q. Zhang, Y. Luo, J. L. Yang, Z. C. Dong, J. G. Hou, Nat. Commun.
2017, 8, 1–7.
[63] G. Reecht, F. Scheurer, V. Speisser, Y. J. Dappe, F. Mathevet, G. Schull,
Phys. Rev. Lett. 2014, 112, 047403.
[30] 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, 1–4.
[31] M. Valášek, M. Mayor, Chem. Eur. J. 2017, 23, 13538–13548.
[32] M. Valášek, M. Lindner, M. Mayor, Beilstein J. Nanotechnol. 2016, 7, 374–
405.
[33] T. Ijaz, B. Yang, R. Wang, J. Zhu, A. Farrukh, G. Chen, G. Franc, Y. Zhang,
A. Gourdon, Z. Dong, Appl. Phys. Lett. 2019, 115, 1–5.
Manuscript received: April 8, 2021
Accepted manuscript online: June 21, 2021
Version of record online: ■■■, ■■■■
Chem. Eur. J. 2021, 27, 1–13
www.chemeurj.org
12
© 2021 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH
��
These are not the final page numbers!

FULL PAPER
Three-legged race: The efficient
synthesis, optical and electrochemical
properties of a series of 2,6-disubsti-
tuted naphthalene-1,4,5,8-tetracarbox-
diimides (NDIs) mounted on a tripodal
platform are reported. Moreover, their
surface behaviour after spraying on
Au(111) was studied. In a STML setup,
such NDI tripods show electrolumi-
nescence photon spectra with high
electroluminescence yields and thus
demonstrate their potential for optoe-
lectronic devices.
N. Balzer, J. Lukášek, Dr. M. Valášek*, V.
Rai, Q. Sun, Dr. L. Gerhard, Prof. Dr. W.
Wulfhekel*, Prof. Dr. M. Mayor*
1 – 13
Synthesis and Surface Behaviour of
NDI Chromophores Mounted on a
Tripodal Scaffold: Towards Self-
Decoupled Chromophores for
Single-Molecule Electroluminescence