Competitive Metal Coordination of Hexaaminotriphenylene on Cu(111) by Intrinsic Copper Versus Extrinsic Nickel Adatoms

Article abstract of DOI:10.1002/chem.201803908

The interplay between the self-assembly and surface chemistry of 2,3,6,7,10,11-hexaaminotriphenylene (HATP) on Cu(111) was complementarily studied by high-resolution scanning tunneling microscopy (STM) and X-ray photoelectron spectroscopy (XPS) under ultra-high vacuum conditions. To shed light on the competitive metal coordination, comparative experiments were carried out on pristine and nickel-covered Cu(111). Directly after room-temperature deposition of HATP onto pristine Cu(111), self-assembled aggregates were observed by STM, and XPS results indicated still protonated amino groups. Annealing up to 200 °C activated the progressive single deprotonation of all amino groups as indicated by chemical shifts of both the N 1s and C 1s core levels in the XP spectra. This enabled the formation of topologically diverse π–d conjugated coordination networks with intrinsic copper adatoms. The basic motif of these networks was a metal–organic trimer, in which three HATP molecules were coordinated by Cu₃ clusters, as corroborated by the accompanying density functional theory (DFT) simulations. Additional deposition of more reactive nickel atoms resulted in both chemical and structural changes with deprotonation and formation of bis(diimino)–Ni bonded networks already at room temperature. Even though fused hexagonal metal-coordinated pores were observed, extended honeycomb networks remained elusive, as tentatively explained by the restricted reversibility of these metal–organic bonds.

Full text of DOI:10.1002/chem.201803908

A Journal of
Accepted Article
Title: Competitive metal-coordination of hexaaminotriphenylene on
Cu(111) by intrinsic copper versus extrinsic nickel adatoms
Authors: Matthias Lischka, Renhao Dong, Mingchao Wang, Natalia
Martsinovich, Massimo Fritton, Lukas Grossmann, Wolfgang
M Heckl, Xinliang Feng, and Markus Lackinger
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.201803908
Link to VoR: http://dx.doi.org/10.1002/chem.201803908
Supported by

10.1002/chem.201803908
Chemistry - A European Journal
FULL PAPER
Competitive metal-coordination of hexaaminotriphenylene on
Cu(111) by intrinsic copper versus extrinsic nickel adatoms
Matthias Lischka,[a],[b] Renhao Dong,[c] Mingchao Wang,[c] Natalia Martsinovich,[d] Massimo Fritton,[a],[b]
Lukas Grossmann,[a],[b] Wolfgang M. Heckl,[a],[b],[e] Xinliang Feng,[c] and Markus Lackinger[a],[b],[e],*
Abstract: The interplay between self-assembly and surface
chemistry of 2,3,6,7,10,11-hexaaminotriphenylene (HATP) on
Cu(111) was complementarily studied by high-resolution Scanning-
Tunneling-Microscopy (STM) and X-ray Photoelectron Spectroscopy
(XPS) under ultra-high vacuum conditions. To shed light on
competitive metal-coordination, comparative experiments were
carried out on pristine and nickel-covered Cu(111). Directly after
room temperature deposition of HATP onto pristine Cu(111) self-
assembled aggregates were observed by STM, while XPS indicated
non-deprotonated amino groups. Annealing up to 200 °C activated
the progressive single deprotonation of all amino groups as indicated
by chemical shifts of both N 1s and C 1s core levels in the XP
spectra. This enabled the formation of topologically diverse -d
conjugated coordination networks with intrinsic copper adatoms. The
basic motif of these networks was a metal-organic trimer, where
three HATP molecules were coordinated by Cu3 clusters, as
corroborated by accompanying Density Functional Theory (DFT)
simulations. Additional deposition of more reactive nickel atoms
resulted in both chemical and structural changes with deprotonation
and formation of bis(diimino)-Ni bonded networks already at room
temperature. Even though fused hexagonal metal-coordinated pores
were observed, extended honeycomb networks remained elusive, as
tentatively explained by a restricted reversibility of these metal-
organic bonds.
centimeter large self-supporting sheets was pioneered at the air-
water interface, and has subsequently also been demonstrated
at the interface between two immiscible liquids.[2] For the latter,
typically multilayers are achieved. For these interface-assisted
synthetic approaches triphenylene-based organic ligands with
six-fold hydroxyl, thiol, or amino functionalization at the
2,3,6,7,10,11 positions evolved as important model system.[3]
The targeted binding motif is a planar four-fold coordination
complex, and incorporation of Cu, Ni, and Co coordination
centers yielded isostructural π-d conjugated CONASHs with
honeycomb structure and widely tunable properties.[4] Superior
electrical conductivities up to 103 S cm-1 were demonstrated,
facilitating a variety of applications as electrode materials for
organic field-effect transistor, chemiresistive sensing,[5]
electrocatalytic hydrogen evolution reaction as well as oxygen
reduction,[3a, 6] and supercapacitors.[3c, 3d, 7]
Despite the tremendous progress in the interface-assisted
synthesis of large area CONASH, a molecular level structural
characterization by high-resolution microscopy remains
challenging. Even though recent advances were achieved for
single-layer CONASH with high-resolution transmission electron
microscopy,[8] the high electron beam sensitivity of these
materials imposes severe limitations. On the contrary,
submolecular features are routinely resolved by Scanning
Tunneling Microcopy (STM) both on metal-coordination and
covalent networks with minimal risk of sample damage. Yet, for
networks grown at liquid interfaces, STM characterization
requires a post-synthetic transfer to suitable solid surfaces, with
only a few reported cases of successful imaging.[9] The reasons
for the difficulties in ex-situ imaging of these networks are not
entirely clear, but might be related to an unavoidable waviness
after their transfer to solid surfaces, crucially impairing high
resolution STM imaging. In contrast, both metal-coordination
and covalent networks that were directly grown on solid surfaces
under ultra-high vacuum (UHV) or even under ambient
conditions are routinely imaged with high resolution.[10] Albeit
requiring ultra-high vacuum (UHV) is not desirable in terms of a
low-cost fabrication, fundamental studies can benefit greatly
from these utmost defined conditions. For instance, a thorough
chemical characterization by X-ray Photoelectron Spectroscopy
(XPS) of in-situ prepared samples is not affected by
contaminations that are unavoidable under ambient conditions.
In the present study we focus on 2,3,6,7,10,11-
hexaaminotriphenylene (HATP, cf. Fig. 1a). Despite the large
body of work on nitrogen-containing metal-coordination networks
with amines,[11] nitriles,[12] pyridines,[13] or amino-quinone-
imines,[14] HATP or other ligands with ortho-disubstituted amino
groups have not been investigated on solid surfaces so far. The
only example of a related compound intended the on-surface
synthesis of pyrene-fused pyrazaacenes via Schiff-base
reaction.[15] To shed more light on the metal-coordination of this
Introduction
The bottom-up synthesis of two-dimensional (2D) materials with
regular structures and tailored properties is a long-standing goal
in chemistry and material science.[1] A pivotal milestone was the
materialization of 2D coordination nanosheets (CONASH),
where multi-topic organic ligands are cross-linked into regular
networks by metal-coordination bonds. The synthesis of square-
[a]
M. Lischka, M. Fritton, L. Grossmann, Prof. Dr. W.M. Heckl, Prof. Dr.
M. Lackinger
Department of Physics, Technische Universität München
James-Franck-Str. 1, 85748 Garching, Germany
E-mail: markus@lackinger.org
[b]
[c]
Center for NanoScience (CeNS) and Nanosystems-Initiative-Munich
(NIM), Schellingstr. 4, 80799 München, Germany
Dr. R. Dong, M. Wang, Prof. Dr. X. Feng
Center for Advancing Electronics Dresden (cfaed) and Department
of Chemistry and Food Chemistry, Technische Universität Dresden,
Mommsenstr. 4, 01069 Dresden, Germany
Dr. N. Martsinovich
Department of Chemistry, University of Sheffield,
Sheffield, S3 7HF, UK
Deutsches Museum, Museumsinsel 1, 80538 München, Germany
Supporting information for this article is given via a link at the end of
the document.
[d]
[e]
This article is protected by copyright. All rights reserved.

10.1002/chem.201803908
Chemistry - A European Journal
FULL PAPER
important molecular building block on solid surfaces, we present
a detailed study of competitive HATP metal-coordination on
Cu(111) between intrinsic Cu adatoms of the surface[11, 14a] and
co-deposited extrinsic Ni adatoms. This choice is motivated by a
strong interest in comparing coordination centers with different
binding strength and ground state electron configurations. The
utilized combination of STM, XPS, and DFT provides the base
for drawing solid and important conclusions on the interplay
between surface chemistry and structure formation.
Results and Discussion
HATP on pristine Cu(111)
First experiments were conducted on pristine Cu(111). The STM
images in Figure 1b and c acquired directly after room
temperate deposition of HATP reveal self-assembled molecular
aggregates. Cyclic hexamers are most abundant, but incomplete
and fused hexamers as well as more densely packed structures
were similarly observed. Individual HATP molecules and their
respective orientation can be discerned by means of the
triangular footprints of the planar adsorbed triphenylene cores.
In the hexamers all HATP molecules point toward the center and
share edges that are not fully aligned, but slightly offset. To
unveil the stabilizing intermolecular bonds in these hexamers,
acquiring complementary information on the chemical state of
HATP is indispensable. In this respect, it is important to address
the question of a possible deprotonation and formation of metal-
coordination bonds of the amino groups already after room
16]
temperature deposition onto Cu(111).[11,
Therefore, XP
spectra were acquired from N 1s and C 1s core levels, whereby
the spectra of 3 - 4 layer thick films shown in Figures 2a and b
served as references for chemically unaltered HATP molecules.
The sharp N 1s peak indicates a single chemical species,
corroborating the anticipated chemical equivalency of all
nitrogen atoms. The measured binding energy (BE) of 399.8 eV
is in accord with NH2 groups on aromatic moieties, with similar
BEs reported for aniline,[17] 1,3-phenyldiamine,[18] tetraamine,[15]
and melamine[19] adsorbed on various metals surfaces. The C 1s
XP spectrum shown in Figure 2b required fitting with two
chemically shifted components with BEs of 284.7eV and 285.7
eV, resulting in a peak area ratio of 67:33. The C 1s component
at higher BE corresponds to the amino-substituted carbons, the
component at lower BE corresponds to the unsubstituted
carbons of the triphenylene core.[17a, 20] Hence, the deduced ratio
perfectly reflects the HATP stoichiometry.
The submonolayer spectra presented in Figures 2c and d are
essentially similar to the multilayer. Only N 1s exhibits a slight
shift of ~0.39 eV towards higher BE. Yet, core-hole screening
and deprotonation would both result in shifts to lower BE, and
can hence be excluded.[14, 17a, 19, 21] There are no notable shifts in
C 1s, indicating a specific surface-influence on the nitrogen
atoms, such as for instance formation of N-Cu bonds. In
summary, XPS indicates adsorption of HATP molecules with
intact NH2 groups after room temperature deposition onto
Cu(111), in accord with results obtained for a comparable
Figure 1. a) chemical structure of HATP; STM images acquired after b) / c)
room temperature deposition of HATP onto Cu(111), and after subsequent
annealing to d) 100 °C and e) / f) 200 °C; DFT-optimized structures were used
for the overlays. c) depicts a HATP decamer comprised of two fused cyclic
hexamers; d) annealing to 100 °C results in a mixture of van-der-Waals
bonded and metal-coordinated structures; e) overview over predominantly
metal coordinated structures obtained after annealing to 200 °C. The close up
in f) shows that kinked dimers of Cu3-coordinated molecular trimers (examples
highlighted in red) represent the most abundant motif. The insert shows a
close up of such a kinked dimer overlaid by a structure with a metal-
coordination interlink comprised of three copper atoms. Green arrows indicate
straight dimers, red arrows highlight single HATP molecules that interlink
supramolecular trimers. tunneling parameters: a) 0.73 V, 54 pA; b) 0.73 V, 51
pA; c) 0.89 V, 59 pA; d) 0.87 V, 90 pA; (d, inset) 0.33 V, 20 pA)
compound.[22] Based on these XPS results, the formation of
intermolecular metal-coordination bonds after room temperature
deposition can be ruled out.
To address the question as to what other interactions drive the
self-assembly of these aggregates, first DFT simulations of
single HATP molecules on Cu(111) were conducted. Various
starting geometries consistently resulted in adsorption
geometries, where the centers of all aromatic rings of the
triphenylene backbone resided above three-fold hollow sites
This article is protected by copyright. All rights reserved.

10.1002/chem.201803908
Chemistry - A European Journal
FULL PAPER
(either fcc- or hcp-sites). Accordingly, half of the carbon atoms
also resided on three-fold hollow sites, and the remaining carbon
atoms on top-sites of the Cu(111) surface. This preferred
adsorption geometry is well known for benzene on transition
metals,[23] and the close match between the spacing of adjacent
aromatic rings in triphenylene (0.250 nm) and the Cu(111) lattice
parameter (0.255 nm) makes this adsorption geometry also
favorable for the larger aromatic core of HATP. In the next step,
several competing models for the cyclic hexamers comprised of
non-deprotonated HATP molecules as suggested by XPS were
calculated on Cu(111) by DFT. All optimized structures
reproduced the experimentally found molecular arrangement
and intermolecular distances within the experimental error (cf.
Supporting Information).
were consistently found to be 0.02 - 0.03 eV per molecule more
favorable than hcp-sites. The mean adsorption height of HATP
molecules in the hexamers (excluding H atoms) above the
Cu(111) surface is (2.900.09) Å, where the standard deviation
was denoted as error bar. The hydrogen atoms point slightly
down and are on average 2.72 Å above the surface. Extremely
small adsorption height differences between fcc- and hcp-sites
of around 0.01 Å provide further evidence for the equivalency of
both adsorption geometries. The DFT-derived center-to-center
distance of two diametrically opposed HATP molecules of 2.41
nm in the structure shown in Figure 3a (cf. Supporting
Information) is in good agreement with the experimental value of
(2.52 ± 0.06) nm. In the modelled hexamer structures half of the
nitrogen atoms reside on top of copper atoms, additionally
allowing N-Cu bond formation. While DFT indicates a rather
strong adsorption of HATP on Cu(111) with adsorption energies
around 4 eV per molecule, the intermolecular interactions in the
hexamers are surprisingly weak, on the order of 0.06 - 0.18 eV
per molecule. The disintegration of aggregates as tracked in
subsequent STM images provides experimental evidence for a
weak stabilization (cf. Supporting Information). Even though
amino groups could in principle also act as hydrogen-bond
acceptors through their nitrogen lone pair, the present DFT
simulations do not provide any evidence for intermolecular
hydrogen bonds. Since intermolecular metal-coordination and
hydrogen bonds can be ruled out, we conclude that solely
comparatively weak van-der-Waals interactions account for the
cohesion of these aggregates, whereas the molecular
orientation is determined by strong and highly site-specific
molecule-surface interactions. It is worth noting that the variety
of aggregates observed after room temperature deposition as
illustrated in Figure 1b, as for instance fused cyclic hexamers
(see Figure 1c), filled hexamers with an additional molecule
inside the central pore, incomplete hexamers, or more densely
packed domains are all related to the cyclic hexamers shown in
Figure 3a with regard to molecular orientations and
intermolecular distances.
To activate deprotonation as a prerequisite for the formation of
metal-coordination bonds, samples were first annealed at
100 °C. The subsequently acquired STM image in Figure 1d
shows the emergence of new cyclic triangular entities, but also
the persistence of the original hexamers with unchanged
dimensions. Occasionally, straight interlinked dimers were also
observed (examples marked by green arrows) with a center-to-
center distance of (1.55 ± 0.08) nm. According to DFT
calculations, the targeted bis(diimino)-Cu motif would result in a
notably smaller center-to-center distance of 1.29 nm. However,
a perfect match is obtained with a dimer, where molecules
coordinate two copper atoms linearly arranged on the axis,
Figure 2. XP spectra of N 1s (left column) and C 1s (right column) acquired for
HATP on Cu(111): a) / b) multilayer; c) / d) submonolayer after room
temperature deposition and subsequent annealing to e) / f) 100 °C, and h) / i)
200 °C. Data are represented by black dots; red lines correspond to the sum
of all fitted components; N 1s and C 1s components are assigned to the
following colour scheme as also shown in the inset: NH2 (green), NH-Cu (blue),
C-C (green), C-NH2 (blue), C-NH-Cu (light blue.)
resulting in
a center-to-center distance of 1.54 nm (cf.
Supporting Information). The emergence of these new
aggregates is also reflected in the XP spectra shown in Figures
2e and f: N 1s is now split into two chemically shifted species,
one at the original BE with a relative peak area of 34% and a
new majority component at a lower BE of 398.1 eV with a
relative peak area of 66%. The lower N 1s BE is characteristic
Figure 3a exemplarily shows one structure, where all aromatic
rings of all six HATP molecules reside above fcc-sites, which
This article is protected by copyright. All rights reserved.

10.1002/chem.201803908
Chemistry - A European Journal
FULL PAPER
for singly deprotonated amino groups in metal-coordination
observed, with a triangular supramolecular HATP trimer as the
bonds.[14a, 18-19, 24] Distinct changes are similarly observed in C 1s, common basic unit. These trimers are interlinked into various
where the previously observed prominent high BE shoulder
appears less pronounced. A satisfying fit is obtained with three
components, corresponding to unsubstituted carbons in the
triphenylene backbone (C-C: 66%) and two different types of
substituted carbons, either binding to fully protonated (C-NH2:
11%) or singly deprotonated and metal-coordinated (C-NH-Cu:
22%) amino groups. The ratio between the latter two is
consistent with the deprotonation ratio deduced from N 1s.
Furthermore, C 1s exhibits a slight integral shift to a lower BE of
284.5 eV for the main triphenylene peak.
configurations: (1) head-to-head; (2) cyclic rings consisting of
four, five, or six trimers; in addition, individual HATP molecules
can be identified attached at the periphery of larger aggregates
either as interlink or termination, respective examples are
marked by red arrows in Figure 1e and f.
For understanding the structural changes, it is pivotal to resolve
the nature of the intermolecular interactions in the trimer. The
threefold-symmetry in STM indicates equivalent orientations of
all molecules in a trimer, whereas XPS suggests chemical
equivalency and deprotonation of all amino groups. Based on
this information, DFT simulations were carried out for two metal-
coordinated models: either with single Cu atoms or Cu3 clusters
as coordination centers. Alternatively, an imine based hydrogen
bonded trimer was considered. The DFT optimized structures of
these trimer models on Cu(111) are presented in Figures 3b - d,
with intermolecular distances of 1.31 nm, 1.34 nm, and 1.20 nm,
for the Cu3-, Cu-, and hydrogen bonded trimer, respectively. All
of these intermolecular distances are consistent with the
experimental value of (1.28 ± 0.03) nm within the experimental
error. Yet, the best agreement is achieved with the Cu3
coordination centers that were similarly proposed for
tetrahydroxy benzene on Cu(111).[25] Additional support for the
Cu3 coordination centers comes from the XPS data: N 1s
indicates a single chemical nitrogen species, whereas the trimer
with single Cu atom coordination features nitrogen atoms in
inequivalent chemical environments. This argument similarly
applies to the hydrogen bonded trimer that contains two types of
nitrogen species: hydrogen-bonded and free imine nitrogen
atoms. In addition, the literature consistently reports copper-
coordination of deprotonated amino groups on copper
14a, 19]
surfaces,[11,
rendering the formation of this hydrogen
Figure 3. DFT-optimized geometries of HATP structures on Cu(111): a) van-
der-Waals bonded hexamer of fully protonated HATP; b) metal-coordinated
trimer with Cu3 coordination centers; c) metal-coordinated trimer with single Cu
atom coordination centers; d) imine based hydrogen bonded trimer. (only the
topmost copper layer is shown, Cu adatoms are depicted in red)
bonded structure improbable. From the DFT calculations HATP
adsorptions heights of (2.64  0.15) Å and (2.57  0.28) Å were
deduced for the Cu3 and single Cu atom coordinated trimers
presented in Figures 3b and c, respectively. These adsorption
heights are approximately 0.3 Å lower than those of the van-der-
Waals bonded hexamers shown in Figure 3a. The closer
proximity to the surface arises from bonding to Cu adatoms
(which themselves reside (1.94 … 2.03) Å above the surface).
This pulls the coordinating N atoms and also the nearby C
atoms closer to the Cu(111) surface (as close as 1.91 Å for the
lowest N atom of the single Cu atom-coordinated trimer).
Consequently, the C atoms of the central part of the molecules
also adsorb closer to the surface than in the van-der-Waals
bonded hexamer.
Annealing to 200 °C triggers progressive deprotonation and
formation of metal-coordinated structures. As shown in Figures
2h and i, the chemical changes in XPS previously observed after
annealing at 100 °C are driven to completion: N 1s indicates full
deprotonation, with only one main peak remaining at a lower BE
of 398.1 eV. Accordingly,
C 1s again consists of two
components, corresponding to carbons in the triphenylene core
and carbons substituted with metal-coordinated amino groups,
respectively. The deduced peak area ratio of 68:32 again
reflects the HATP stoichiometry. The integral peak shift of C 1s
to lower BEs is continued with a BE of 284.3 eV, suggesting a
correlation with the amount of metal-coordination sites on the
molecules.[19] The total intensities of both N 1s and C 1s did not
change notably during heating, indicating the absence of
molecular desorption. The chemical changes observed in XPS
are also accompanied by structural changes as illustrated in the
STM image in Figure 1e. Hexamers cannot be discerned
anymore; instead, a large variety of different aggregates are
It is worth noting that N 1s XPS indicates deprotonation and
copper coordination of all amino groups after annealing to
200 °C. Accordingly, in the DFT calculations copper coordination
was also introduced at the peripheral amino groups of the metal-
coordinated trimers. This peripheral group coordination could
explain the formation of larger aggregates and networks. The
STM images in Figures 1e and f reveal the Cu3-coordinated
trimer as the predominating basic motif of the various observed
This article is protected by copyright. All rights reserved.

10.1002/chem.201803908
Chemistry - A European Journal
FULL PAPER
aggregates. Prevalence of this supramolecular building unit also
indicates a hierarchy of bond strengths, where the intra-trimer
bonds are stronger than the inter-trimer bonds. Among the
the amino groups. These also give rise to a third component in C
1s, similar to the thermally activated deprotonation on pristine
Cu(111). A good fit is achieved by applying the deprotonation
ratio obtained from N 1s as a constraint for the relative ratios of
the C 1s components. These striking differences to the van-der-
Waals bonded structures observed on pristine Cu(111) after
room temperature deposition can evidently be attributed to the
presence of Ni, whereby the changes in both N 1s and C 1s
XPS are analogous to the thermally activated deprotonation on
pristine Cu(111). To avoid artefacts known for high molecular
coverages due to steric crowding and to ensure full accessibility
of all amino groups by Ni adatoms, low molecular coverages
were deliberately used for these experiments. This dilution
comes at the cost of an inferior signal-to-noise ratio in the XP
spectra, in particular for N 1s. Nevertheless, the progressive
deprotonation can still unambiguously be verified, in particular
when comparing the initial fully protonated state with the final
completely deprotonated state in Figures 5a and h. Moreover,
the initial deprotonation upon co-deposition of Ni and
subsequent thermally-induced further deprotonation is also
reflected in the corresponding C 1s data presented in Figure 5
through the weakening of the high BE shoulder. These XPS
signatures of the Ni-induced deprotonation are fully consistent
with the deprotonation observed on pristine Cu(111).
To activate further deprotonation and network formation, the
samples were subsequently annealed to 100 °C. XPS confirms
the progressive deprotonation of now approximately 72% of the
amino groups directly in N 1s, and indirectly, but consistently
also in C 1s by both an increase of the metal-organic component
and as an integral peak shift to a lower BE of 284.4 eV. The not
yet complete deprotonation is also reflected in STM by the
occasional observation of van-der-Waals bonded aggregates
similar to those observed at room temperature in the absence of
Ni (cf. Figure 1b). One example for such a cyclic structure is
marked in Figure 4b. Even though expression of the Ni-
coordinated motif is evident this STM image, it does not show
complete pores or networks. To drive deprotonation to
completion, a second annealing step to 200 °C was performed,
resulting in full deprotonation as confirmed by the N 1s and also
consistently in the C 1s XP spectra presented in Figures 5h and
i: N 1s consist of a single peak at lower BE and C 1s shows full
development of the metal-organic component up to the
stoichiometric ratio of 33:67 with a constant amount of carbon.
An analogous experiment with the five-fold amount of co-
deposited Ni similarly resulted only in partial deprotonation at
room temperature, whereas full deprotonation still required
additional annealing (cf. Supporting Information).
various aggregates head-to-head linked trimers with
a
pronounced offset along their bond axis account for the majority
of aggregates, respective examples are highlighted in Figure 1f.
To further elucidate these inter-trimer linkages, DFT simulations
were performed, probing metal-coordination linkages comprised
of one, two, and three copper atoms. Notably, only linking of the
trimers by three copper atoms could reproduce the kinked
geometry, also perfectly matching the experimental structure (cf.
insert to Figure 1f), whereas linking by either one or two copper
atoms results in straight geometries (cf. Supporting Information).
Moreover, the postulated hierarchy of bond strengths is
underpinned by DFT simulations: For a metal-coordination motif
corresponding to the kinked inter-trimer linkage with three Cu
atoms a binding energy of 8.52 eV was found, whereas a single
Cu3-coordination motif corresponding to the intra-trimer linkage
exhibits a notably higher binding energy of 9.82 eV (cf.
Supporting Information).
Additional experiments with three-fold reduced heating rates
(0.50 °C min-1 instead of 1.66 °C min-1) did not increase the
network quality (cf. Supporting Information). Neither did further
annealing to 300 °C, where both STM and XPS studies
consistently indicate molecular degradation: both a low BE
organometallic shoulder at C 1s and the diminished intensity of
N 1s suggest dissociation of amino groups (cf. Supporting
Information).
HATP on Cu(111) + Ni
To achieve the formation of honeycomb networks based on the
four-fold planar bis(diimino)-metal linkages, further experiments
were conducted with additional Ni atoms on the surface. First,
HATP was deposited onto Cu(111) held at room temperature,
followed by co-deposition of Ni. The subsequently acquired STM
image in Figure 4a unveils interlinking of molecules into
hexagonal porous structures that resemble the targeted
honeycomb network. Chemical changes are also corroborated
by XPS shown in Figure 5: direct comparison of N 1s and C 1s
spectra before and after Ni co-deposition clearly confirms a
chemical change in HATP. Both the appearance of a new N 1s
component at lower BE and weakening of the high BE shoulder
of C 1s consistently indicate a similar kind of transformation as
that observed for Cu-coordinated HATP in Figure 2. Fitting of the
N 1s spectra suggests a Ni-induced deprotonation of ~50 % of
This article is protected by copyright. All rights reserved.

10.1002/chem.201803908
Chemistry - A European Journal
FULL PAPER
phase, resulting in a distance of 1.27 nm between directly linked
HATP molecules. This intermolecular distance closely
corresponds to 5-times the Cu(111) lattice parameter (1.28 nm),
hence this size and symmetry match between the Ni-
coordinated honeycomb network and the Cu(111) surface
should foster a nearly perfect epitaxial relation. Indeed, DFT
simulations with periodic boundary conditions of
a fully
reticulated Ni-coordinated honeycomb network adsorbed on
Cu(111) corroborate a perfect match, whereby the two non-
equivalent molecules within the unit cell can adsorb with their
preferred geometry (i.e. with the centres of aromatic rings above
fcc sites, cf. Supporting Information). Accordingly, surface
registry effects are unlikely to account for the absence of
extended Ni-coordinated honeycomb networks on Cu(111).
Figure 4. STM images acquired after deposition of HATP and Ni onto Cu(111)
at a) room temperature, and after annealing to b) 100 °C, and c) / d) 200 °C;
the red arrows highlight terminating Ni atoms; the yellow dashed circle
indicates a van-der-Waals bonded cyclic structure (tunneling parameters: a)
0.90 V, 87 pA; b) 0.95 V, 92 pA; c) 3.27 V, 91 pA; d) 0.88 V, 93 pA)
The STM image in Figure 4c now shows network patches
consisting of up to three hexagonal metal-coordinated pores, but
incomplete pores were still frequently observed. The insert with
an overlaid dimer in Figure 4c and the high resolution image in
Figure 4d clearly confirm formation of the expected bis(diimino)-
Ni binding motif with head-to-head oriented molecules
interlinked by a straight metal-coordination bond. Also the
experimental center-to-center distance of (1.45 ± 0.1) nm is in
accord with the DFT-derived value obtained for a Ni-coordinated
hexamer in the gas phase (cf. Supporting Information).
Furthermore, a clear signature of the Ni coordination center can
now be discerned in the STM contrast as opposed to Cu
coordination centers that are often invisible to the STM.[12a, 14, 26]
Despite the possibility to form a similar bis(diimino)-Cu motif with
intrinsic adatoms of the Cu(111) surface, the Ni motif is strongly
preferred, presumably on energetic grounds: the binding energy
in the bis(diimino)-Ni coordinated dimer with respect to isolated
constituents is calculated to be 9.7 eV, while in the Cu-
coordinated dimer it is only 6.0 eV (cf. Supporting Information).
Interestingly, the termini of the incomplete pores show bright
protrusions, indicating saturation with Ni atoms at the periphery
(examples are marked by red arrows in Figure 4d). While further
annealing to 260 °C did not improve the network quality,
annealing at 300 °C resulted in disordered structures in STM,
while XPS indicated molecular decomposition (cf. Supporting
Information).
Figure 5. XP spectra of N 1s and C 1s core levels acquired after a) / b) room
temperature deposition of HATP on Cu(111), c) / d), after additional deposition
of Ni, and annealing to e) / f) 100 °C, and h) / i) 200 °C. Data is represented by
black dots; red lines correspond to the sum of all fitted components; N 1s and
C 1s components are assigned to the following colour scheme: NH2 (green),
NH-Ni (blue), C-C (green), C-NH2 (blue), C-NH-Ni (light blue).
Additional DFT simulations were performed aiming at elucidating
the role of epitaxial effects of the Cu(111) surface for the Ni-
coordinated structures (cf. Supporting Information). In the first
step, a cyclic Ni-coordinated hexamer was optimized in the gas
To achieve the growth of extended and highly regular metal-
coordination networks, the ratio of molecules to coordination
centers is a further, potentially decisive parameter.[27] Therefore,
This article is protected by copyright. All rights reserved.

10.1002/chem.201803908
Chemistry - A European Journal
FULL PAPER
additional experiments were carried out with lower (~30%) and
higher (~4-fold) amounts of Ni, while the amount of molecules
was kept constant. Similarly, HATP was deposited first onto
Cu(111) held at room temperature, then Ni was co-deposited in
the three different coverages, and finally, the samples were
annealed to 200 °C. The STM image for “low” Ni coverage in
Figure 6a shows two co-existing different metal-coordination
motifs: the trimer based structures as observed on pristine
Cu(111) and the straight motifs as observed in the presence of
Ni. This indicates that the amount of Ni was too small to
suppress Cu coordination. For the previously applied Ni
coverage the results are consistent, exclusively showing the
bis(diimino)-Ni motif. At “high” Ni coverage the onset of cluster
formation on the surface can be seen in Figure 6c. Although
exclusively the Ni coordination motif is observed, the size of
metal-organic networks is markedly diminished.
motif was not observed on Cu(111), given its predominance at
liquid interfaces.[5] This hints towards thermodynamic preference
for
the copper-rich Cu3 coordination on Cu(111).
Notwithstanding that this motif would in principle allow the
extension into fully reticulated and regular 2D networks, only
various aggregates comprised of interlinked supramolecular
trimers were observed. The experimentally observed kinked
structure of two connected supramolecular trimers was
reproduced by DFT simulations with three coordinating copper
atoms in the interlink. According to DFT calculations, the inter-
trimer linkages have a smaller binding energy than the intra-
trimer linkages. This hierarchy of bond strengths can explain the
hierarchical structure formation based on
a
stable
supramolecular Cu3-coordinated trimer as secondary building
unit. A possible explanation for the absence of extended
networks on pristine Cu(111) is offered by pronounced
adsorption site preferences of both the triphenylene backbones
and the Cu3 clusters that cannot be met both at the same time in
larger networks.
.
Co-deposition of the more reactive Ni triggers deprotonation
already at room temperature, clearly indicating adatom mediated
surface chemistry. However, full deprotonation at room
temperature was not achieved, even when significantly higher
amounts of Ni were deposited. In the competition with metal-
coordination by Cu adatoms, a distinct preference for Ni-
coordination is expressed. Accordingly, insufficient amounts of
Ni can be held responsible for the observed co-existence of Ni-
coordination and Cu-coordination. Here we found 50%
deprotonation after co-deposition of Ni at room temperature,
whereas the thermally induced further deprotonation showed a
temperature progression comparable to pristine Cu(111). This
suggests that the thermally driven process is Cu-driven without
any further notable contributions from Ni. Even though small
patches of the targeted honeycomb network were formed,
extended regular structures remained elusive. DFT simulations
suggest that a fully reticulated Ni-coordinated honeycomb
network can be commensurate with the Cu(111) surface,
whereby the two non-equivalent molecules of the unit cell adopt
their preferred adsorption geometry. Consequently, these
simulations indicate that a lattice mismatch cannot be held
responsible for the absence of extended Ni-coordinated
honeycomb networks on Cu(111). Yet, the experiments with
variable amounts of co-deposited Ni indicate a limited bond
reversibility of the strong Ni coordination bonds. This prevents
the fusion of the smaller fragments, hence decisively hampers
the growth of extended networks. This is further underpinned by
DFT, suggesting a more than 50% higher binding energy for Ni-
than for Cu-coordination, i.e. 9.7 eV vs. 6.0 eV for the isolated
dimer. The diminished size of molecular aggregates observed
for larger amounts of Ni provides further experimental evidence
for this hypothesis. This restricted bond reversibility is a decisive
difference compared to the situation at liquid interfaces, where
extended self-supporting networks can be obtained.[3d]
Consequently, the aqueous subphase must play a vitally
important role for the reversibility of the metal-organic
coordination bonds. Higher temperatures, that could in principle
Figure 6. STM images of HATP on Cu(111) acquired after deposition of
different amounts of Ni and subsequent annealing to 200 °C; a) “low” Ni
coverage (30%); b) previously applied Ni coverage (100%); c) “high” Ni
coverage (400%) (tunneling parameters: a) 0.88 V, 55 pA; b) 1.80 V, 100 pA;
c) 3.20 V, 89 pA).
Conclusions
The observed structural and chemical changes of HATP on
pristine and Ni-covered Cu(111) surfaces are summarized in
Figure 7. HATP molecules remained chemically unaltered upon
room temperature deposition onto pristine Cu(111). The
absence of deprotonation of amino-functionalized compounds
under these conditions is common,[11, 14a, 22] whereas on more
reactive Ni(111) surfaces partial deprotonation can readily
18]
occur.[17a,
Despite the fact that HATP molecules are not
equipped for strong intermolecular bonds, stable self-assemblies
could be imaged by STM at room temperature. Since
intermolecular metal-coordination and hydrogen bonds were
ruled out, the cohesion of these aggregates must solely arise
from van-der-Waals interactions. According to DFT calculations,
the intermolecular binding energies are extremely small, but a
strong and highly site-specific adsorption of the triphenylene
core on Cu(111) largely accounts for the stabilization.
Subsequent annealing triggered the well-known thermally
14a, 22]
activated deprotonation on pristine Cu(111).[11,
This
provided the basic prerequisite for formation of metal-
coordination bonds, unexpectedly with Cu3 clusters as
coordination centers. Albeit single atom coordination motifs are
most common on surfaces, a similar Cu3 motif was previously
proposed for tetrahydroxy benzene on Cu(111).[25] In this respect,
it appears more surprising that the expected bis(diimino)-Cu
This article is protected by copyright. All rights reserved.

10.1002/chem.201803908
Chemistry - A European Journal
FULL PAPER
Cu(111) single crystals were prepared by cycles of 0.5 (1.0) keV Ar+-ion
sputtering and radiative (e-beam) annealing at 500 °C. HATP (cf.
Supporting Information for synthesis) was deposited by sublimation from
a home-built Knudsen-cell[28] with a crucible temperature of 225 °C at a
pressure of 3 × 10−9 mbar. Nickel (0.5 mm diameter wire, 99.99%,
MaTecK) was deposited from an Omicron EFM 3 e-beam evaporator
using a molybdenum crucible and deposition rates were controlled with
the internal flux monitor. For all sample annealing steps a standard
heating and cooling rate of 1.66 °C min-1 was applied, if not noted
otherwise.
induce bond reversibility also on Cu(111), are prohibitive, as this
leads to molecular decomposition, presumably by dissociation of
the amino groups.
STM data were acquired in-situ at room temperature with a home-built
instrument controlled by an SPM100 controller (RHK Technology Inc.).
XPS measurements were carried out in a SPECS system equipped with
a monochromatic XR50M X-ray source and Phoibos 150 electron
analyzer. XP spectra were acquired using a Al-Kα source with a photon
energy of 1486.3 eV at normal electron emission with a pass energy of
20 eV. For energy calibration a Cu 3p3/2 binding energy of 70.1 eV was
used as internal standard.[29] XP spectra were fitted by Voigt-functions
after subtraction of linear backgrounds. For a given element similar peak
widths were applied for all chemically shifted components.
DFT simulations were carried out with the CP2K software.[30] The PBE
functional was employed[31] with the D3 dispersion correction,[32]
Goedecker-Teter-Hutter pseudopotentials,[33] and double-zeta valence
polarized basis sets.[34] The Brillouin zone was sampled only at the 
point. Isolated aggregates were modelled as non-periodic systems; their
z-coordinates (vertical) were fixed to maintain planarity to mimic the
presence of a surface. Calculations of adsorbates on the Cu(111) surface
were performed using periodic boundary conditions, i.e. the Cu(111)
slabs (and all slab-adsorbate systems) were periodic in two dimensions;
along the z direction the slabs were separated by 10 Å vacuum. Two-
layer thick slabs were used, with atomic positions in the bottom layer
fixed, while the top layer and all adsorbates were allowed to relax.
1616-expanded Cu(111) slabs were used for the van-der-Waals bonded
hexamers and the supramolecular trimers. This lateral size of the slabs
guaranteed a minimum spacing of 6 Å between the periodic repetitions of
the molecular aggregates, i.e. large enough to exclude interactions
between neighbouring hexamers and trimers in image cells; 1515-
expanded Cu(111) slabs were used for the Ni-coordinated honeycomb
networks, resulting in a fully reticulated commensurate 5√3×5√3
R30 ° superstructure. All structure files can be downloaded from:
https://figshare.com/s/e5573906428dfbf16dfa
Figure 7. Summary of thermally and Ni-induced structural and chemical
changes of HATP on pristine and Ni-covered Cu(111)
In summary, studying the formation of metal-coordinated
networks from HATP on Cu(111) under UHV conditions
substantially benefits from the possibility of acquiring high-
resolution STM images and complementary chemical
information from XPS. Important differences to the synthesis at
liquid interfaces become apparent: (1) the high availability of
adatoms as coordination centers on solid metal surfaces
promotes the expression of multi-atomic coordination motifs; (2)
distinct adsorption-site preferences impose further constraints
on crystalline surfaces that may hamper the formation of
extended networks in the present case; (3) the absence of the
liquid subphase substantially affects metal-organic bond
formation, with possible consequences for bond reversibility.
Acknowledgements
The excellence cluster Nanosystems Initiative Munich is
gratefully acknowledged for financial support. The Authors thank
Karl Eberle for support and technical assistance during XPS
measurements. N.M. acknowledges the use of Iceberg and
ShARC high-performance computing clusters at the University of
Sheffield and the use of the THOMAS high-performance cluster
which is partially funded by EPSRC (EP/P020194) and was
accessed via our membership of the UK's HEC Materials
Chemistry Consortium (funded by EPSRC EP/L000202).
Keywords: metal-coordination • hexaaminotriphenylene •
Cu(111) • Scanning Tunneling Microscopy • X-ray Photoelectron
Spectroscopy
Materials and Methods
STM and XPS experiments were carried out under UHV conditions at
base pressures below 3 × 10−10 mbar. For STM (XPS) experiments
This article is protected by copyright. All rights reserved.

10.1002/chem.201803908
Chemistry - A European Journal
FULL PAPER
[1]
a) J. Sakamoto, J. van Heijst, O. Lukin, A. D. Schlüter, Angew. Chem.,
Int. Ed. 2009, 48, 1030-1069; b) D. F. Perepichka, F. Rosei, Science
2009, 323, 216-217; c) J. V. Barth, Annu. Rev. Phys. Chem. 2007, 58,
375-407.
[15] . iang, . C. a ageorgiou, S. C. h, . Sa lam, . eichert, D. A.
Duncan, Y.-Q. Zhang, F. Klappenberger, Y. Guo, F. Allegretti, S. More,
R. Bhosale, A. Mateo-Alonso, J. V. Barth, ACS Nano 2016, 10, 1033-
1041.
[2]
a) R. Sakamoto, K. Takada, X. Sun, T. Pal, T. Tsukamoto, E. J. H.
Phua, A. Rapakousiou, K. Hoshiko, H. Nishihara, Coord. Chem. Rev.
2016, 320-321, 118-128; b) R. Sakamoto, K. Takada, T. Pal, H. Maeda,
T. Kambe, H. Nishihara, Chem. Commun. 2017, 53, 5781-5801; c) T.
Bauer, Z. Zheng, A. Renn, R. Enning, A. Stemmer, J. Sakamoto, A. D.
Schlüter, Angew. Chem. Int. Ed. 2011, 50, 7879-7884; d) L. Chen, J.
Kim, T. Ishizuka, Y. Honsho, A. Saeki, S. Seki, H. Ihee, D. Jiang, J. Am.
Chem. Soc. 2009, 131, 7287-7292.
[16] L. Dong, Z. A. Gao, N. Lin, Prog Surf Sci 2016, 91, 101-135.
[17] a) S. X. Huang, D. A. Fischer, J. L. Gland, J. Phys. Chem. 1996, 100,
10223-10234; b) X. Xu, C. M. Friend, J. Vac. Sci. Technol 1991, 9,
1599-1603.
[18] T. S. Jones, M. R. Ashton, N. V. Richardson, R. G. Mack, W. N. Unertl,
J. Vac. Sci. Technol 1990, 8, 2370-2375.
[19] Y.-P. Lin, O. Ourdjini, L. Giovanelli, S. Clair, T. Faury, Y. Ksari, J.-M.
Themlin, L. Porte, M. Abel, J. Phys. Chem. C 2013, 117, 9895-9902.
[20] a) J. Hwang, Y. Kim, Y. Ahn, Bull. Korean Chem. Soc. 2014, 35, 2831-
2834; b) J. L. Solomon, R. J. Madix, J. Stöhr, Surf. Sci. 1991, 255, 12-
30.
[3]
a) R. Dong, Z. Zheng, D. Tranca, J. Zhang, N. Chandrasekhar, S. Liu,
X. Zhuang, G. Seifert, X. Feng, Chem. Eur. J. 2016, 23, 2255–2260; b)
R. Dong, M. Pfeffermann, H. Liang, Z. Zheng, X. Zhu, J. Zhang, X.
Feng, Angew. Chem. Int. Ed. 2015, 54, 12058-12063; c) D. Sheberla, J.
C. Bachman, J. S. Elias, C.-J. Sun, Y. Shao-Horn, M. Dincă, Nat. Mat.
2016, 16, 220-224; d) D. Sheberla, L. Sun, M. . lood- ors the, S. r,
C. . ade, C. . ro e , . n. s uru- u i , M. Dincă, J. Am. Chem.
Soc. 2014, 136, 8859-8862; e) L. Giovanelli, O. Ourdjini, M. Abel, R.
Pawlak, J. Fujii, L. Porte, J.-M. Themlin, S. Clair, J. Phys. Chem. C
2014, 118, 14899-14904; f) R. Coratger, B. Calmettes, M. Abel, L.
Porte, Surf. Sci. 2011, 605, 831-837.
[21] a) A. C. Papageorgiou, S. Fischer, J. Reichert, K. Diller, F. Blobner, F.
Klappenberger, F. Allegretti, A. P. Seitsonen, J. V. Barth, ACS Nano
2012, 6, 2477-2486; b) K. Diller, F. Klappenberger, M. Marschall, K.
Hermann, A. Nefedov, C. Wöll, J. V. Barth, J. Chem. Phys 2012, 136,
014705; c) T. C. Chiang, G. Kaindl, T. Mandel, Phys. Rev. B 1986, 33,
695-711; d) S. Kohiki, K. Oki, F. Konishi, Anal. Sci. 1985, 1, 115-117.
[22] F. Masini, Y. Ning, Z. Li, E. Lægsgaard, F. Besenbacher, T. R.
Linderoth, Chem. Commun. 2013, 49, 8665.
[4]
[5]
[6]
[7]
D. Herebian, E. Bothe, F. Neese, T. Weyhermüller, K. Wieghardt, J. Am.
Chem. Soc. 2003, 125, 9116-9128.
[23] a) W. Liu, V. G. Ruiz, G.-X. Zhang, B. Santra, X. Ren, M. Scheffler, A.
Tkatchenko, New J. Phys. 2013, 15, 053046; b) X.-Q. Shi, Y. Li, M. A.
V. Hove, R.-Q. Zhang, J. Phys. Chem. C 2012, 116, 23603-23607.
[24] P. R. Davies, D. Edwards, D. Richards, J. Phys. Chem. B 2004, 108,
18630-18639.
M. G. Campbell, D. Sheberla, S. F. Liu, T. M. Swager, M. Dincă, Angew.
Chem., Int. Ed. 2015, 54, 4349-4352.
E. M. Miner, T. Fukushima, D. Sheberla, L. Sun, Y. Surendranath, M.
Dincă, Nat. Commun. 2016, 7, 10942.
[25] F. Bebensee, K. Svane, C. Bombis, F. Masini, S. Klyatskaya, F.
Besenbacher, M. Ruben, B. Hammer, T. R. Linderoth, Angew. Chem.
Int. Ed. 2014, 53, 12955-12959.
C. Yang, K. S. Schellhammer, F. Ortmann, S. Sun, R. Dong, M.
Karakus, Z. Mics, M. Löffler, F. Zhang, X. Zhuang, E. Cánovas, G.
Cuniberti, M. Bonn, X. Feng, Angew. Chem. Int. Ed. 2017, 56 3920-
3924.
[26] a H. alch, . Dienstmaier, . der, . ut ler, S. Schl gl, T. Sirtl, .
Das, M. Schmittel, M. Lackinger, J. Am. Chem. Soc. 2011, 133, 7909-
7915; b) J. Björk, M. Matena, M. S. Dyer, M. Enache, J. Lobo-Checa, L.
H. Gade, T. A. Jung, M. Stöhr, M. Persson, Phys. Chem. Chem. Phys.
2010, 12, 8815; c) S. L. Tait, A. Langner, N. Lin, S. Stepanow, C.
Rajadurai, M. Ruben, K. Kern, J. Phys. Chem. C 2007, 111, 10982-
10987.
[8]
[9]
X. Huang, S. Zhang, L. Liu, L. Yu, G. Chen, W. Xu, D. Zhu, Angew.
Chem. Int. Ed. 2017, 57, 146-150.
D. J. Murray, D. D. Patterson, P. Payamyar, R. Bhola, W. Song, M.
Lackinger, A. D. Schlüter, B. T. King, J. Am. Chem. Soc. 2015, 137,
3450-3453.
[10] a) N. A. A. Zwaneveld, R. Pawlak, M. Abel, D. Catalin, D. Gigmes, D.
Bertin, L. Porte, J. Am. Chem. Soc. 2008, 130, 6678-6679; b) J. F.
Dienstmaier, D. D. Medina, M. Dogru, P. Knochel, T. Bein, W. M. Heckl,
M. Lackinger, ACS Nano 2012, 6, 7234-7242; c) S. Spitzer, A.
Rastgoo-Lahrood, K. Macknapp, V. Ritter, S. Sotier, W. M. Heckl, M.
Lackinger, Chem. Commun. 2017, 53, 5147-5150.
[27] a) X. Zhang, G. Gu, N. Li, H. Wang, H. Tang, Y. Zhang, S. Hou, Y.
Wang, RSC Adv. 2018, 8, 1852-1856; b) U. Schlickum, F.
Klappenberger, R. Decker, G. Zoppellaro, S. Klyatskaya, M. Ruben, K.
Kern, H. Brune, J. V. Barth, J. Phys. Chem. C 2010, 114, 15602-15606.
[28] R. Gutzler, W. M. Heckl, M. Lackinger, Rev. Sci. Instrum. 2010, 81,
015108.
[11] M. Knor, H.-Y. Gao, S. Amirjalayer, A. Studer, H. Gao, S. Du, H. Fuchs,
Chem. Commun. 2015, 51, 10854-10857.
[29] N. S. McIntyre, M. G. Cook, Anal. Chem 1975, 47, 2208-2213.
[30] J. VandeVondele, M. Krack, F. Mohamed, M. Parrinello, T. Chassaing,
J. Hutter, Comput. Phys. Commun. 2005, 167, 103-128.
[12] a) T. Sirtl, S. Schlögl, A. Rastgoo-Lahrood, J. Jelic, S. Neogi, M.
Schmittel, W. M. Heckl, K. Reuter, M. Lackinger, J. Am. Chem. Soc.
2013, 135, 691-695; b) S. Stepanow, N. Lin, D. Payer, U. Schlickum, F.
Klappenberger, G. Zoppellaro, M. Ruben, H. Brune, J. V. Barth, K. Kern,
Angew. Chem., Int. Ed. 2007, 119, 724-727; c) K. Müller, J. C. Moreno-
López, S. Gottardi, U. Meinhardt, H. Yildirim, A. Kara, M. Kivala, M.
Stöhr, Chem. Eur. J. 2015, 22, - d . Schlic um, . Dec er, .
la en erger, . o ellaro, S. l ats a a, . uw rter, S. Ne l, .
Kern, H. Brune, M. Ruben, J. V. Barth, J. Am. Chem. Soc. 2008, 130,
11778-11782.
[31] J. P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 1996, 77, 3865-
3868.
[32] S. Grimme, J. Antony, S. Ehrlich, H. Krieg, J. Chem. Phys 2010, 132,
154104.
[33] S. Goedecker, M. Teter, J. Hutter, Phys. Rev. B 1996, 54, 1703-1710.
[34] J. VandeVondele, J. Hutter, J. Chem. Phys 2007, 127, 114105.
[13] Y. Li, J. Xiao, T. E. Shubina, M. Chen, Z. Shi, M. Schmid, H.-P.
Steinrück, J. M. Gottfried, N. Lin, J. Am. Chem. Soc. 2012, 134, 6401-
6408.
[14] a) A. Shchyrba, C. Wäckerlin, J. Nowakowski, S. Nowakowska, J. Björk,
S. Fatayer, J. Girovsky, T. Nijs, S. C. Martens, A. Kleibert, M. Stöhr, N.
Ballav, T. A. Jung, L. H. Gade, J. Am. Chem. Soc. 2014, 136, 9355-
9363; b) M. Matena, J. Björk, M. Wahl, T.-L. Lee, J. Zegenhagen, L. H.
Gade, T. A. Jung, M. Persson, M. Stöhr, Phys. Rev. B 2014, 90,
125408.
This article is protected by copyright. All rights reserved.

10.1002/chem.201803908
Chemistry - A European Journal
FULL PAPER
FULL PAPER
The interplay between self-
Matthias Lischka, Renhao Dong,
assembly and surface chemistry of
hexaaminotriphenylene was
Mingchao Wang, Natalia Martsinovich,
Massimo Fritton, Lukas Grossmann,
Wolfgang M. Heckl, Xinliang Feng,
Markus Lackinger*
comparatively studied on pristine vs.
Ni-covered Cu(111) surfaces.
Without Ni, coordination by Cu3
clusters was observed, whereas in
the presence of Ni the bis(diimino)-
Ni motif was dominating.
Page No.– Page No.
Competitive metal-coordination of
hexaaminotriphenylene on Cu(111)
by intrinsic copper versus extrinsic
nickel adatoms
This article is protected by copyright. All rights reserved.