Steering on-surface polymerization with metal-directed template

Article abstract of DOI:10.1021/ja311890n

On-surface polymerization represents a novel bottom-up approach for producing macromolecular structures. To date, however, most of the structures formed using this method exhibit a broad size distribution and are disorderly adsorbed on the surface. Here we demonstrate a strategy of using metal-directed template to control the on-surface polymerization process. We chose a bifunctional compound which contains pyridyl and bromine end groups as the precursor. Linear template afforded by pyridyl-Cu-pyridyl coordination effectively promoted Ullmann coupling of the monomers on a Au(111) surface. Taking advantage of efficient topochemical enhancement owing to the conformation flexibility of the Cu-pyridyl bonds, macromolecular porphyrin structures that exhibit a narrow size distribution were synthesized. We used scanning tunneling microscopy and kinetic Monte Carlo simulation to gain insights into the metal-directed polymerization at the single molecule level. The results reveal that the polymerization process profited from the rich chemistry of Cu which catalyzed the C-C bond formation, controlled the size of the macromolecular products, and organized the macromolecules in a highly ordered manner on the surface.

Full text of DOI:10.1021/ja311890n

Article
pubs.acs.org/JACS
Steering On-Surface Polymerization with Metal-Directed Template
†
‡
†
,‡
,†
Tao Lin, Xue Song Shang, Jinne Adisoejoso, Pei Nian Liu,* and Nian Lin*
†
Department of Physics, The Hong Kong University of Science and Technology, Clear Water Bay, Hong Kong, China
Shanghai Key Laboratory of Functional Materials Chemistry and Institute of Fine Chemicals, East China University of Science and
‡
Technology, Meilong Road 130, Shanghai, China
*
S Supporting Information
ABSTRACT: On-surface polymerization represents a novel bottom-up approach for producing macromolecular structures. To
date, however, most of the structures formed using this method exhibit a broad size distribution and are disorderly adsorbed on
the surface. Here we demonstrate a strategy of using metal-directed template to control the on-surface polymerization process.
We chose a bifunctional compound which contains pyridyl and bromine end groups as the precursor. Linear template afforded by
pyridyl−Cu−pyridyl coordination effectively promoted Ullmann coupling of the monomers on a Au(111) surface. Taking
advantage of efficient topochemical enhancement owing to the conformation flexibility of the Cu−pyridyl bonds,
macromolecular porphyrin structures that exhibit a narrow size distribution were synthesized. We used scanning tunneling
microscopy and kinetic Monte Carlo simulation to gain insights into the metal-directed polymerization at the single molecule
level. The results reveal that the polymerization process profited from the rich chemistry of Cu which catalyzed the C−C bond
formation, controlled the size of the macromolecular products, and organized the macromolecules in a highly ordered manner on
the surface.
1
2,22−24,26,31
INTRODUCTION
this obstacle, various strategies have been proposed.
■
Linderoth et al. reported that organizational motifs in
preassembled structures formed from one of the reactants
may be retained in the final product structure. Lackinger et al.
showed that polycondensation (dehydration) of boronic acid in
the presence of water led to covalent networks of high
Chi et al. used troughs on the gold
surface to line up the polymer chains formed out of alkane
monomers. More recently, Grill et al. used I- and Br-
functionalized precursors to control the outcome structures via
stepwise coupling. Here we developed an effective approach
of using metal−ligand coordination as template to steer the on-
surface polymerization process. The resulting macromolecular
structures exhibit a very narrow size distribution and are
organized hierarchically through supramolecular assembly.
Template synthesis is an area where “molecular and
On-surface polymerization refers to the process in which
molecular precursors are covalently linked on a surface. It
represents a new strategy to synthesize macromolecular
systems. The earlier successful attempts were demonstrated at
26
1
−5
liquid−solid interfaces.
In 2007, Grill et al. demonstrated
2
2,23
structural quality.
that dimeric, polymeric, and network structures can be formed
on a Au(111) surface under ultrahigh-vacuum conditions
through covalent coupling of bromine-terminated porphyrin
12
6
precursors. In less than six years, on-surface coupling based on
31
various reaction routes has generated a large variety of
macromolecular systems, including polymeric chains,
7
−12
1
3,14
15
hyper-branched oligomers,
graphene ribbons, porous
1
6−24
25
molecular networks,
super honeycomb frameworks,
2
6−31
etc.
This progress has attracted increasing attention, and
32−34
the field of on-surface polymerization is growing quickly.
3
5,36
supramolecular science meet”.
It uses noncovalent binding
motifs to steer the formation of covalently linked products.
The nonreversibility of the covalent bonds, however,
introduces kinetic trapping in the polymerization processes.
As a result, the size of the macromolecular systems formed by
on-surface polymerization cannot be controlled accurately, and
the products are randomly distributed on surface. To overcome
Received: December 5, 2012
©
XXXX American Chemical Society
A
dx.doi.org/10.1021/ja311890n | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
This approach dates back to the 1960s when Busch studied and
forming one-dimensional metallosupramolecular single-row
(SR) chains at 23 °C (Figure 1a). Since the SR chains emerged
only upon codeposition of Cu and 1 and the nearest-
neighboring molecule distance (1.9 nm) is the same as the
length of the coordination chains formed by 2 and Cu (Figure
S1), we conclude that the SR chains were stabilized by 2-fold
37
classified different modes of templates. Metal−ligand binding,
hydrogen bonding, and π−π interactions have been used as
templates to control the synthesis of molecules with remarkable
36
structural characteristics that are otherwise difficult to achieve.
An elegant example of metal-directed template was demon-
strated by Anderson and his co-workers: belt-like nanorings
consisting of 6, 8, 12, or more porphyrin units were prepared
4
1,42
py−Cu−py coordination.
The py groups at the trans-
positions of 1 give rise to the linear morphology of the model
shown in Figure 1a. The coordinated Cu atoms were not
resolved in the STM data, presumably due to electronic effects
or tip conditions. It is worth noting that many of the SR chains
exhibited a bent or curved shape, implying a high degree of
flexibility of the py−Cu−py coordination. This bond angle
flexibility even allowed a large section of chains to shift on the
surface without fragmentation (cf. Figure 3c,e and Figure S1).
The bond angle flexibility is associated with the two-fold
configuration of the py−Cu−py coordination which is known
3
8
using radial oligo-pyridine templates. Another type of
template is linear template which promotes metal-assisted or
metal-catalyzed solid-state dimerization within a 1D coordina-
3
9
tion polymer. This type of template has the potential to self-
replicate, since the product can in principle be identical to the
40
template, akin to the replication of DNA. In this study, we
used linear template to control the on-surface polymerization.
METHODS
■
(
43
Experiments were performed in an ultrahigh-vacuum system
to exhibit a high degree of conformational flexibility.
Omicron Nanotechnology) with base pressure below 5 × 10−10
After annealing at 180 °C for 30 min, ladder-shaped chains
consisting of double or triple rows (DR or TR, respectively) of
monomers appeared, as shown in Figure 1b. Figure 1c is a high
resolution STM image of a DR chain. Detailed inspection
reveals that there is a gap between the neighboring molecules
along the chain direction (cf. the blue line in Figure 1c),
whereas the two molecules perpendicular to the chain direction
are linked by a bridge (cf. the red line in Figure 1c). The
separation between neighboring molecules (center-to-center)
perpendicular to the chain direction is 1.73 nm, which is shorter
than the separation between neighboring molecules (center-to-
center) along the chain direction (1.95 nm) but the same as the
intermolecular distance in the covalently linked porphyrin
mbar. A single-crystalline Au(111) substrate was prepared by argon
ion sputtering and annealing at ∼630 °C, resulting in terraces of ∼100
nm wide. Copper atoms were deposited by an electron-beam
evaporator on the substrate which was held at room temperature
(
23 °C). The copper deposition rate was determined through
analyzing the area of Cu islands formed on a clean Au substrate.
The crucibles containing the molecules were then heated in an organic
molecular beam deposition source (Dodecon) to a certain evaporating
temperature (350, 320, 370, and 320 °C for compounds 1−4 in
Scheme 1, respectively), and the molecules were deposited on the
Scheme 1. Porphyrin Compounds Used in This Study
6
macromolecules. These structural characteristics clearly
indicate that the monomers that are 1.73 nm apart are
covalently bonded dimeric macromolecules and that these
macromolecules are linked by py−Cu−py coordination along
the chain direction, forming the DR chains as illustrated by the
model in Figure 1b. Another piece of evidence of the proposed
structure is the protrusions at the sides of the porphyrin dimers,
as marked by the arrow in Figure 1c. These protrusions, which
6,31
substrate held at room temperature or higher. Scanning tunneling
microscopy (STM) measurements were performed at 23 °C if not
indicated otherwise in a constant-current mode at 1.0−1.4 V positive
or negative bias and 0.2−0.4 nA tunneling current. The STM scanning
scales were calibrated using Au(111) as a standard sample.
have been observed before and identified as bromine atoms,
indicate that the Br groups are positioned at right angles to the
DR chain direction. To further verify the proposed DR model,
we added a batch of compound 3 molecules to a sample which
was already covered with the DR chains in advance. We found
that the compound 3 molecules did not attach to the sides of
DR chains but rather formed 2D coordination islands that
attached to the end of the pre-existing DR chains (Figure S2).
It is thus evident that the py groups are positioned along the
chain direction.
The kinetic Monte Carlo (KMC) simulations were performed on a
00 × 100 square lattice, which corresponds to one monolayer of a
1
molecule network containing 33 × 33 molecules. Periodic boundary
conditions were applied. The molecules were deposited randomly
onto the substrate lattice. Desorption of the molecules from the
substrate was not allowed. The molecules could hop to the nearest-
neighboring site or rotate by 90° clockwise or anticlockwise. When
one molecule attached to another molecule in a configuration in which
pyridyl (py) functions of the two molecules were aligned head to head
Cu-Catalyzed Ullmann Coupling. Apparently, 180 °C
annealing activated the C−C coupling reaction of 1, leading to
the formation of the covalently bonded porphyrin
(
py−py), a coordination bond was formed. And when one molecule
6,8,11,15,17,25,31,44
dimers.
The reaction temperature is lower
encountered another in a Br−Br configuration, the rate at which a
covalent bond was formed was determined by an energy barrier of 1.2
eV. The details of the KMC simulation algorithm can be found in ref.
than that of the same coupling reaction which is thermally
6
,31,44
activated on a Au(111) surface,
hinting that Cu acts as a
4
1 and in the Supporting Information (SI).
catalyst. Although the reaction steps were not accessible using
STM, we identified the intermediate Cu organometallic state in
the samples annealed at 120 °C (Figure S3), which evidenced
RESULTS AND DISCUSSION
Identification and Verification of On-Surface Polymer-
ization. Compound 1 (Scheme 1) provides py groups as
coordination sites for metal-directed template and bromine
groups as the Ullmann coupling sites. We found that in the
presence of Cu, the monomers of 1 assembled with Cu atoms
■
45
that Cu atoms were involved in the coupling reactions. To
quantitatively examine the effect of Cu, we varied the Cu
dosage and annealing parameters. First, in a control experiment,
we prepared a sample deposited with 1 but without Cu. We
found that in the absence of Cu, the ratio of C−C bond
B
dx.doi.org/10.1021/ja311890n | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 1. Structures formed from 1 and Cu on a Au(111) surface. STM topographs (100 ×100 nm) showing (a) SR coordination chains formed at
5 °C and (b) DR and TR chains formed after 180 °C annealing. (c) High-resolution STM topograph of a DR chain. The arrow is pointing at Br.
d) Reaction ratios at different Cu dosages. Inset shows the reaction ratios at a fixed Cu dosage but with different annealing durations.
2
(
Figure 2. Comparison of the polymerization of 1 and 4. (a) STM topograph (100 ×100 nm) showing macromolecular structures formed from 4
data acquired at 200 K). (b,c) KMC simulated structures formed out of 1 and 4 (at 450K for 30 min). (d) Reaction ratios of 1 and 4 as a function of
reaction time. Triangles represent simulated results, solid lines represent exponential fitting, and vertical bars represent experimental values. Blue
red) indicates 1(4). (e) Experimental weight distribution of the macromolecular structures formed from 1 and 4 as a function of macromolecule
(
(
size.
formation of 1 was <1.5% after annealing at 180 °C for 10 min.
Figure 1d shows that under the same annealing temperature
but for a shorter duration (180 °C for 3 min), adding Cu at a
dosage of [Cu]/[Mol.] = 3 ([Cu] and [Mol.] refer to the
surface coverage of deposited Cu atoms and that of deposited
molecules of 1, respectively) amplified the reaction ratio
(reaction ratio is defined as the percentage of Br groups that
were replaced by C−C bonds) by more than 10 times. When
the Cu dosage was raised to [Cu]/[Mol.] = 5 and 19, the
reaction ratios were amplified by 20 or 30 times, respectively.
C
dx.doi.org/10.1021/ja311890n | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
These trials clearly demonstrated that Cu effectively promoted
the coupling reaction, a process that can be considered as an
Ullmann reaction occurring on the surface. Note that in the
solution-phase reactions, the catalyst loading of the metal, such
as Cu, is usually much smaller than that of the reactant. We
attribute the high catalyst loading in this study to the
immobilization of the reactant as coordinated chains and the
reduced mobility of the Cu catalyst on the surface compared
with the free movement and collision of the catalyst and the
reactant in the solution phase. The inset in Figure 1d shows
that at a constant Cu dosage, extending the annealing duration
to 30 (60) min caused the reaction ratio to go up to 53%
defined as the percentage of monomers that form macro-
molecules of this size out of the total amount of monomers.
Figure 2e shows that the weight distributions of two
compounds exhibit distinctive characteristics. First, monomers
account for 5% in 1 but almost 60% in 4, which reflects the
lower reaction rates of 4. Second, the macromolecular
structures formed from 4 exhibit a wide size distribution. For
instance, structures as long as 15-mers can be found. In
contrast, under the same experimental conditions, the macro-
molecular structures formed from 1 show a narrower size
distribution in which the dimeric macromolecules are
dominantly formed (63%) and the largest structure is the
pentamer. Last but not least, the macromolecules formed out of
1 are regularly organized as chains on the surface through metal
coordination, whereas the products of 4 are randomly
distributed.
(56%). So at this Cu dosage, the reaction at 180 °C for 3 min
was kinetic controlled. Interestingly, the reaction ratio after 60
min of annealing was not enhanced significantly as compared
with that after 30 min of annealing. This finding implies that
the coupling reaction was in a deep kinetic trap. It is worth
noting that the reaction ratio did not reach 100% but was
saturated at about 60% even with high Cu dosage and long
annealing duration. This effect will be discussed in detail later.
Metal-Directed Template Afforded by py−Cu−py
Coordination. To unravel the role of py functions, we
compared the polymerization behavior of 1 and 4. Figure 2a is a
representative STM image of a sample deposited with 4 and Cu
and annealed at 180 °C for 60 min. The figure shows covalently
The reactions, which occurred at 180 °C or higher, could not
be directly monitored using STM. Here we discuss two possible
reaction mechanisms. One scenario is that the metal-
coordinated SR chains decomposed upon annealing and
some molecules formed covalent dimers, trimers or larger
multimers. The coordination chains were formed when the
sample was cooled. However, this process would generate
macromolecules of a broad size distribution like that of
compound 4, since the py function would not play any role in
such a process. Furthermore, the macromolecules of different
sizes would be linked though metal coordination in the cooling
process, which would result in chains consisting of sections of
various width. In other words, the formed chains would not
exhibit homogeneous width. But neither of these features was
observed in the experiments, so we rule out this mechanism.
The second mechanism is based on template-assisted coupling
reaction, as illustrated in Scheme 2, where a DR chain is
developed from a dimeric molecule seed step by step. The
decomposition of the DR chain requires breaking two metal−
ligand bonds simultaneously, which is very unfavorable
energetically. Thus once formed, the DR chain does not
decompose at the annealing temperature. At each growth step,
an intermediate state is formed in which two precursor
monomers are separately anchored through py−Cu−py
coordination to the end dimer of the DR chain, as illustrated
in the upper part of Scheme 2. In the intermediate state, the
anchoring bonds of py−Cu−py coordination must be bent to
avoid spatial conflict between the Br end groups of the two
anchored molecules, as highlighted by the circle in Scheme 2.
As discussed before, the bond angle of py−Cu−py coordination
can deviate from 180°. This bond angle flexibility allows the
intermediate state to be formed. The circles in Figure 3c mark
several examples of the intermediate state. In the intermediate
state, the Br groups of the two anchored molecules are brought
into proximity (see the circled area in the upper part of Scheme
2). Such a configuration facilitates the Ullmann coupling
reaction between the two Br ends as in topochemical
6
bonded polymeric chains. (STM images of the sample
annealed for other durations are shown in Figure S5.) We
quantitatively compared the reaction ratios of 1 and 4 under
the same Cu dosage and annealing treatment and found that
the reaction ratios of 4 were always lower than those of 1: 11%
vs 16%, 38% vs 53%, and 51% vs 57% for 3, 30, and 60 min of
annealing, respectively (represented by the vertical bars in
Figure 2d). We used the KMC method to simulate the
reactions of 1 and 4. Figure 2b,c shows the simulated
polymerization of 1 and 4 at 450K for 30 min, respectively.
One can see that the molecules of 1 form ladder-like structures
in which the monomers are linked by py−Cu−py coordination
bonds (marked in blue), while in the orthogonal direction the
monomers are coupled by covalent bonds (marked in black).
Note that Figure 2b shows only one monomer not involved in
the coupling reaction. In contrast, an appreciable amount
(60%) of monomers of 4 is still present in Figure 2c. The
simulated reaction ratios of the two compounds as a function of
annealing time are plotted in Figure 2d. Each set of data can be
−t/τ
fitted using the function A(1 − e ). The time constant τ
characterizes the reaction rate. The time constant of 1 (341 s)
is about one-fifth of that of 4 (1536 s). The prefactor A,
however, is comparable for 1 and 4. Thus the simulation reveals
that the py−Cu−py coordination significantly enhanced the
coupling reaction rates but did not change the reaction
equilibrium. As shown in Figure 3d, the reaction ratios of 4
obtained experimentally are in good agreement with the
simulated ones. The simulation of 1 reproduced qualitatively
the trend of the experimental reaction ratios. However, the
actual experimental values were lower than the simulated ones.
This discrepancy might be associated with the fact that the
parameters used in the simulation did not fully agree with the
experimental ones. For example, surface steps and Cu density
were excluded in the simulation.
46
polymerization processes. As can be seen in the time
sequence frames of the simulation (SI movie), most of the
C−C bonds were formed between the neighboring monomers
that were anchored to the existing chains though coordination
bonds, so the reaction rate of the monomers in the
intermediate state increased substantially in comparison with
that of the reaction between the monomers that moved
randomly on the surface. Hence we propose that Cu played a
dual role in promoting the C−C bond formation. On the one
hand, Cu reduced the reaction barrier as the catalyst in the
We conducted statistical analysis of the experimental weight
distribution of the macromolecular structures formed out of 1
or 4 after 60 min of annealing at 180 °C. The weight of
macromolecules containing a specific number of monomers is
D
dx.doi.org/10.1021/ja311890n | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Scheme 2. Ullmann Coupling Assisted by the py−Cu−py
Coordination Template Owing to Topochemical
Enhancement
Ullmann reaction, and on the other hand, Cu ligated with py as
template which further speeded up the reaction.
Size-Limited Polymerization. We now discuss the size
distribution of the macromolecules. Figures 1b and 2d show
that 180 °C annealing resulted in predominantly dimeric and
trmeric structures. We annealed a series of samples deposited
with 1 and Cu at different annealing temperatures (Figures S6
and S7). Figure 3a shows a representative STM image of a
sample after annealing at 300 °C. One can see that most of the
products are DR chains. Figure 3b shows the weight
distribution of the macromolecular structures generated at
different annealing temperatures. The weight of TR chains
increases slightly at higher temperatures. Nevertheless, the
high-temperature annealing did not significantly enhance the
formation of larger oligomers. For example, in all cases,
tetramers or larger are very rare, and the largest structure
observed is heptamer structures, as marked by the arrow in
Figure 3a. Above 180 °C, DR chains are always the most
popular structure regardless of the annealing parameters
(
temperature and duration). The reaction ratios in all these
trials are <60%. The saturation of the reaction ratio at 60%
corroborates the predominant formation of DR chains (note
that the reaction ratio is 50% when all monomers form dimeric
macromolecules).
Besides the size limitation of the macromolecules, surpris-
ingly, in all the samples, the chains have very smooth edges.
This phenomenon reveals that: (1) attachment of individual
monomers to the side of a chain is an unfavorable process and
(
2) in the case where a side attachment does occur, the
attached monomers will quickly develop into a new row to the
side of the chain. Both (1) and (2) can be understood by
referring to the template effect discussed before. A precursor
Figure 3. STM topographs (100 ×100 nm) and weight distributions of 1 with different preparation parameters. (a) Molecules were deposited on the
substrate at room temperature and annealed at 300 °C. (b)Weight distribution of the macromolecular structures formed with different annealing
treatments. Inset shows the corresponding reaction ratios. (c,d) The sample prepared in two steps, where in each step, molecules were deposited on
the substrate at room temperature, annealed at 180 °C, and cooled down to room temperature. (e,f) Molecules were deposited at 240 °C on the
substrate which was covered with Cu in advance.
E
dx.doi.org/10.1021/ja311890n | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
monomer can attach to the side of a chain through a C−C
bond, a process that requires overcoming an energy barrier.
Being anchored to the ends of the existing chains through py−
Cu−py coordination, however, is a process that does not
require overcoming an energy barrier. So the precursor
monomers prefer the latter process which then leads to the
formation of dimeric structures and extend the chain’s length,
as illustrated in Scheme 2. In rare cases, precursor monomers
might undergo covalent attachment to the side of DR chains. In
such a situation, as illustrated in Scheme 3, the covalently
°C) which was covered with Cu in advance. At a high reaction
temperature, more covalently bonded trimers or other large
macromolecules were generated. Figure 3e is a STM image of
this sample. One can see that more TR chains were indeed
formed. The weight distribution (Figure 3f) shows that the
weight of the TR chains reached 50%, exceeding the weight of
the DR chains.
CONCLUSION
■
In summary, the combined STM and KMC study of the
specially designed bifunctional porphyrin compounds has
provided insights into metal-directed template reactions on
surfaces at the molecular level. The results reveal that the metal-
directed template afforded by py−Cu−py coordination
effectively alters the outcome structure of the on-surface
polymerization. First, the dimeric structure is formed with a
very high yield of >60% at 180 °C. Second, the macromolecules
are organized by metal coordination into supramolecular chains
on the surface. We believe these results will inspire the design
and synthesis of size- and shape-controlled macromolecular
systems in the fast emerging field of on-surface synthesis.
Scheme 3. Growth of a New Row to the Side of a DR Chain
Starting from an Initially-Attached Monomer Assisted by
py−Cu−py Coordination Template
ASSOCIATED CONTENT
■
*
S
Supporting Information
Synthesis of compounds 1−4, detailed KMC procedure and
parameters, and supporting figures and movie. This material is
available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
attached monomer can anchor other monomer(s) through py−
Cu−py coordination at its side(s). Owing to the topochemical
enhancement effect discussed before, the anchored mono-
mer(s) will form C−C bonds with the porphyrin unit(s) of the
DR chain. Thus, the attachment of one monomer triggers the
rapid growth of a new row which would become covalently
bonded to the side of the existing DR chain. The simulated
time sequence frames (SI movie) show the proposed process:
Once a monomer is covalently attached to the side of a chain, a
new row develops from it. To summarize, we found that the
coordination template steered the polymerization process in
such a way as to favor the replication of the existing smaller
Corresponding Author
phnlin@ust.hk; liupn@ecust.edu.cn
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported financially by the Hong Kong
Research Grants Council (project no. 603611), the National
Natural Science Foundation of China (project nos. 21172069
and 21190033), the Innovation Program of Shanghai Municipal
Education Commission (project no. 12ZZ050), and the
Fundamental Research Funds for the Central Universities.
(
dimeric and trimeric) structures over the growth of larger
structures, which led to the predominant formation of the
dimeric macromolecules.
REFERENCES
1) Grim, P. C. M.; De Feyter, S.; Gesquier
cker, M.; Valiyaveettil, S.; Moessner, G.; Mu
C. Angew. Chem., Int. Ed. 1997, 36, 2601.
2) Okawa, Y.; Aono, M. Nature 2001, 409, 683.
3) Okawa, Y.; Aono, M. J. Chem. Phys. 2001, 115, 2317.
4) Sakaguchi, H.; Matsumura, H.; Gong, H.; Abouelwafa, A. M.
Science 2005, 310, 1002.
■
To test the proposed mechanism, we carried out the
following control experiment: First we prepared a sample
with DR chains of low surface coverage, then we deposited
more molecules of 1 and Cu on the surface and annealed the
sample at 180 °C for 30 min. One can see in Figure 3c that
sections of several DR chains as marked by the arrows were
widened and became TR chains. However, the STM (Figure
(
̀
e, A.; Vanoppen, P.;
llen, K.; De Schryver, F.
Ru
̈
̈
(
(
(
3
c) and the weight distribution (Figure 3d) reveal that the DR
(5) Sakaguchi, H.; Matsumura, H.; Gong, H. Nat. Mater. 2004, 3,
551.
chains are still the dominant structures. Thus it is obvious that
the majority of the late-arriving molecules formed new DR
chains or extended the pre-existing DR chains, which is
consistent with the proposed mechanism. This reaction
mechanism suggests that to enhance the yield of larger
macromolecules, some larger macromolecular structures must
be formed at the early stage of the reaction as seeds. These
structures may trigger the formation of structures of the same
size at the later stage of the reaction, i.e., a self-replicating
process. We deposited molecules of 1 on a hot surface (240
(
6) Grill, L.; Dyer, M.; Lafferentz, L.; Persson, M.; Peters, M. V.;
Hecht, S. Nat. Nanotechnol. 2007, 2, 687.
7) Matena, M.; Riehm, T.; Stohr, M.; Jung, T. A.; Gade, L. H.
Angew. Chem., Int. Ed. 2008, 47, 2414.
8) Lipton-Duffin, J. A.; Ivasenko, O.; Perepichka, D. F.; Rosei, F.
Small 2009, 5, 592.
9) Schmitz, C. H.; Ikonomov, J.; Sokolowski, M. J. Phys. Chem. C
009, 113, 11984.
(10) Jensen, S.; Fru
131, 16706.
(
̈
(
(
2
̈
chtl, H.; Baddeley, C. J. J. Am. Chem. Soc. 2009,
F
dx.doi.org/10.1021/ja311890n | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
́
(
11) Lipton-Duffin, J. A.; Miwa, J. A.; Kondratenko, M.; Cicoira, F.;
(43) Heim, D.; Ecija, D.; Seufert, K.; Auwar
̈
ter, W.; Aurisicchio, C.;
Sumpter, B. G.; Meunier, V.; Perepichka, D. F.; Rosei, F. Proc. Natl.
Acad. Sci. U.S.A. 2010, 107, 11200.
Fabbro, C.; Bonifazi, D.; Barth, J. V. J. Am. Chem. Soc. 2010, 132, 6783.
(44) Lafferentz, L.; Ample, F.; Yu, H.; Hecht, S.; Joachim, C.; Grill, L.
Science 2009, 323, 1193.
(45) Wang, W.; Shi, X.; Wang, S.; Van Hove, M. A.; Lin, N. J. Am.
Chem. Soc. 2011, 133, 13264.
(
12) Zhong, D.; Franke, J. -H.; Podiyanachari, S. K.; Blo
Zhang, H.; Kehr, G.; Erker, G.; Fuchs, H.; Chi, L. Science 2011, 334,
13.
13) Weigelt, S.; Busse, C.; Bombis, C.; Knudsen, M. M.; Gothelf, K.
̈
mker, T.;
2
(
(
46) Nagahama, S.; Matsumoto, A. J. Poly. Sci. A: Poly. Chem. 2004,
4
2, 3922.
V.; Lægsgaard, E.; Besenbacher, F.; Linderoth, T. R. Angew. Chem., Int.
Ed. 2008, 47, 4406.
(
14) Wang, S.; Wang, W.; Lin, N. Phys. Rev. B. 2012, 86, 045428.
(
15) Cai, J.; Ruffieux, P.; Jaafar, R.; Bieri, M.; Braun, T.; Blankenburg,
S.; Muoth, M.; Seitsonen, A. P.; Saleh, M.; Feng, X.; Mu
R. Nature 2010, 466, 470.
16) Zwaneveld, N. A. A.; Pawlak, R.; Abel, M.; Catalin, D.; Gigmes,
D.; Bertin, D.; Porte, L. J. Am. Chem. Soc. 2008, 130, 6678.
17) Gutzler, R.; Walch, H.; Eder, G.; Kloft, S.; Heckl, W. M.;
Lackinger, M. Chem. Commun. 2009, 29, 4456.
18) Tanoue, R.; Higuchi, R.; Enoki, N.; Miyasato, Y.; Uemura, S.;
Kimizuka, N.; Stieg, A. Z.; Gimzewski, J. K.; Kunitake, M. ACS Nano
011, 5, 3923.
19) Schmitz, C. H.; Ikonomov, J.; Sokolowski, M. J. Phys. Chem. C
011, 115, 7270.
20) Schlogl, S.; Sirtl, T.; Eichhorn, J.; Heckl, W. M.; Lackinger, M.
Chem. Commun. 2011, 47, 12355.
21) Faury, T.; Clair, S.; Abel, M.; Dumur, F.; Gigmes, D.; Porte, L. J.
Phys. Chem. C 2012, 116, 4819.
22) Dienstmaier, J. F.; Gigler, A. M.; Goetz, A. J.; Knochel, P.; Bein,
T.; Lyapin, A.; Reichlmaier, S.; Heckl, W. M.; Lackinger, M. ACS Nano
011, 5, 9737.
23) Dienstmaier, J. F.; Medina, D. D.; Dogru, M.; Knochel, P.; Bein,
T.; Heckl, W. M.; Lackinger, M. ACS Nano 2012, 6, 7234.
24) Guan, C. -Z.; Wang, D.; Wan, L. -J. Chem. Commun. 2012, 48,
943.
25) Bieri, M.; Treier, M.; Cai, J.; Aït-Mansour, K.; Ruffieux, P.;
Groning, O.; Groning, P.; Kastler, M.; Rieger, R.; Feng, X.; Mullen, K.;
Fasel, R. Chem. Commun. 2009, 45, 6919.
26) Weigelt, S.; Bombis, C.; Busse, C.; Knudsen, M. M.; Gothelf, K.
V.; Lægsgaard, E.; Besenbacher, F.; Linderoth, T. R. ACS Nano 2008,
, 651.
27) Treier, M.; Fasel, R.; Champness, N. R.; Argent, S.; Richardson,
N. V. Phys. Chem. Chem. Phys. 2009, 11, 1209.
28) Coratger, R.; Calmettes, B.; Abel, M.; Porte, L. Surf. Sci. 2011,
05, 831.
29) Men
011, 40, 4578.
30) Lackinger, M.; Heckl, W. M. J. Phys. D: Appl. Phys. 2011, 44,
64011.
31) Lafferentz, L.; Eberhardt, V.; Dri, C.; Africh, C.; Comelli, G.;
Esch, F.; Hecht, S.; Grill, L. Nat. Chem. 2012, 4, 215.
̈
llen, K.; Fasel,
(
(
(
2
(
2
(
̈
(
(
2
(
(
2
(
̈
̈
̈
(
2
(
(
6
(
́ ́
dez, J.; Lopez, M. F.; Martín-Gago, J. A. Chem. Soc. Rev.
2
(
4
(
(32) Champness, N. R. Nat. Nanotechnol. 2008, 3, 324.
(33) Gourdon, A. Angew. Chem., Int. Ed. 2008, 47, 6950.
(34) Perepichka, D. F.; Rosei, F. Science 2009, 323, 216.
(35) Anderson, S.; Anderson, H. L.; Sanders, J. K. M. Acc. Chem. Res.
1
(
993, 26, 469.
36) Diederich, F.; Stang, P. J. Templated Organic Synthesis, Wiley-
VCH: Weinheim, Germany, 2000.
(
37) Thompson, M. C.; Busch, D. H. J. Am. Chem. Soc. 1964, 86, 213.
(
38) Sprafke, J. K.; Kondratuk, D. V.; Wykes, M.; Thompson, A. L.;
̈
Hoffmann, M.; Drevinskas, R.; Chen, W. -H.; Yong, C. K.; Karnbratt,
J.; Bullock, J. E.; Malfois, M.; Wasielewski, M. R.; Albinsson, B.; Herz,
L. M.; Zigmantas, D.; Beljonne, D.; Anderson, H. L. J. Am. Chem. Soc.
2
(
2
011, 133, 17262.
39) Toh, N. L.; Nagarathinam, M.; Vittal, J. J. Angew. Chem., Int. Ed.
005, 117, 2277.
(
(
(
40) Hoffmann, S. Angew. Chem., Int. Ed. 1992, 31, 1013.
41) Li, Y.; Lin, N. Phys. Rev. B 2011, 84, 125418.
42) Adisoejoso, J.; Li, Y.; Liu, J.; Liu, P. N.; Lin, N. J. Am. Chem. Soc.
2
012, 134, 18526.
G
dx.doi.org/10.1021/ja311890n | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX