Graphene nanoribbons by chemists: Nanometer-sized, soluble, and defect-free

Article abstract of DOI:10.1002/anie.201006593

Closing the zipper: A method for the bottom-up organic synthesis of defect-free graphene nanoribbons in solution has been developed. Polyphenylene precursors with a unique kinked backbone enabled full cyclodehydrogenation in a single reaction step by an i

Full text of DOI:10.1002/anie.201006593

Communications
DOI: 10.1002/anie.201006593
Nanoribbon Synthesis
Graphene Nanoribbons by Chemists: Nanometer-Sized, Soluble,
and Defect-Free**
Lukas Dꢀssel, Lileta Gherghel, Xinliang Feng, and Klaus Mꢁllen*
Dedicated to Professor Henning Hopf on the occasion of his 70th birthday
Structural perfection is of key importance in the synthesis of
graphene nanoribbons (GNRs), since there is a fundamental
connection between the electrical properties and the width,
the edge periphery, and the occurrence of defects.[1] Graphene
itself is a zero-band-gap semimetal, whereas GNRs with a
width smaller than 10 nm show semiconducting behavior that
renders them suitable active materials for electronic devices.[2]
A considerable longitudinal extension would enable ready
processing and device fabrication with single ribbons. The
synthetic challenge is the preparation of such defined
graphene structures with high aspect ratios (length/width)
that until now could not be attained by physical methods. Top-
down approaches (Figure 1), such as the reduction of graphite
oxide,[3] lithography,[4] the unzipping of carbon nanotubes,[5] or
the mechanical exfoliation[2] of graphene, have so far lacked
any means of control over the size and edge structure of the
resulting products and have thus led to poorly defined
graphene materials.
Bottom-up organic synthesis enables structural control on
the atomic scale and thus enables chemists to reach their goal
of synthesizing defined carbon materials, such as carbon
nanotubes (CNTs)[6] or nanographenes, with different aspect
ratios.[7] Recently, we fabricated atomically precise GNRs by
employing a surface-assisted coupling reaction followed by a
thermally induced intramolecular cyclodehydrogenation step
to yield nanoribbons with different widths and edges.[8]
However, at present this method is limited to the production
of sub-monolayers on conductive surfaces, and thus process-
ing can only rely on physical methods. In contrast, methods
for the solution synthesis of GNRs have been developed in
the last few years which have enabled the preparation of
nanoribbons with a width of about 1 nm and a length of up to
10 nm.[7c] This route was based on the preparation of tailored
polyphenylene precursors, which were converted into the
corresponding polycyclic aromatic hydrocarbons (PAHs) by a
Scholl reaction (oxidative cyclodehydrogenation). Neverthe-
less, it was found that perfectly shape-defined GNRs could
still not be synthesized. Owing to defects caused by incom-
plete cyclodehydrogenation and side reactions that occurred
during the final reaction step, careful optimization of the
reaction procedure was required.[9] Attempts to extend the
systems in length were hampered by the already low solubility
of the precursors. To further broaden the scope of application
of graphene-type materials produced by bottom-up organic
synthesis, the challenge is to overcome these limitations. In
this study, we investigated whether the Scholl reaction would
enable the synthesis of structurally perfect GNRs.
Herein we present a series of nanoribbons 4, 5, and 6
(Scheme 1) that were synthesized from polyphenylene pre-
cursors with a unique nonrigid kinked backbone to introduce
higher solubility in comparison to that of strictly linear
poly(para-phenylene) systems.[10] We also prepared corre-
sponding model compounds to gain better understanding of
the cyclodehydrogenation in the synthesis of extended con-
jugated nanoribbon systems with a special focus on the degree
of dehydrogenation and the occurrence of rearrangements
within the ribbon backbone. Indeed, full dehydrogenation of
linear polyphenylenes with a length of more than 40 nm is
possible without rearrangement and yields perfectly defined
conjugated nanoribbons that are still soluble in common
organic solvents. The nanoribbons were characterized by
different methods to prove the structural perfection, includ-
ing infrared and Raman spectroscopy, matrix-assisted laser
desorption/ionization time-of-flight (MALDI-TOF) mass
spectrometry (MS), and solution techniques, such as UV/Vis
Figure 1. Schematic overview of the a) top-down and b) bottom-up
fabrication of GNRs.
[*] Dipl.-Chem. L. Dꢀssel, Dipl.-Chem. L. Gherghel, Dr. X. Feng,
Prof. Dr. K. Mꢁllen
Max-Planck-Institut fꢁr Polymerforschung
Ackermannweg 10, 55128 Mainz (Germany)
Fax: (+49)6131-379-350
E-mail: muellen@mpip-mainz.mpg.de
[**] This research was supported financially by the Max Planck Society
through the program ENERCHEM, the German Science Foundation
(Korean-German IRTG), the DFG Priority Program SPP 1355, the
DFG (1459 Graphene Priority Program), One-P (FP7, no. 212311),
and GOSPEL.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201006593.
2540
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 2540 –2543

The polymer precursors 1–3 were
synthesized by a microwave-assisted
Suzuki polycondensation of an equi-
molar mixture of ortho-dibromoben-
zenes and benzene-1,4-diboronic
esters (see the Supporting Informa-
tion) and subsequent end capping of
the bromo end with a phenylboronic
ester and of the boron end with
bromobenzene. Finally, the polymers
were converted into the correspond-
ing ribbons 4–6 through an intramo-
lecular cyclodehydrogenation reac-
tion with FeCl3 as an oxidant in
dichloromethane at room tempera-
ture for 3 days. In contrast to previous
approaches in which limited solubil-
ity and steric hindrance provided
problems during the polymerization
step to form polyphenylene precur-
sors,[7] it was found that the kinked
backbone enabled ready synthesis of
polymers 1–3 with high molecular
weights. Molar masses of up to
16000 gmolÀ1 for polymer 1 were
found by MALDI-TOF MS with a
signal pattern typical for A–B-type
polymerization (see the Supporting
Scheme 1. Polyphenylene precursors 1–3, conjugated ribbons 4–6, and model compounds 7 and 8;
repeat units are highlighted in red.
absorption, photoluminescence, and gel permeation chroma-
tography (GPC) analysis.
Oligomer 7, a defined segment of polymer precursor 2
with a kinked backbone, can have two conformational
isomers, 7 and 7* (Scheme 1). If rearrangement reactions
are suppressed, full cyclodehydrogenation of these isomers
would lead to the symmetrical products 8a and 8b. Cyclo-
dehydrogenation of 7 was carried out with FeCl3 as an oxidant
in dichloromethane at room temperature for 1 h to yield
solely PAH 8a as a waxy material. The solubility of 8a in
common organic solvents enabled full structural character-
ization. 1H NMR spectroscopy of PAH 8a in [D8]THF clearly
confirmed that 8a rather than 8b was formed with high
specificity (Figure 2a). High-resolution MALDI-TOF MS
gave further proof, as the main peak found corresponded to
8a with a calculated mass of 1048.88 gmolÀ1 (Figure 2b)—any
other product would have a different mass as a result of a
different number of newly formed bonds. The factors
controlling this reaction are still undetermined.
On the basis of the concept of a kinked backbone, we
further enlarged our system towards polymeric model com-
pounds 1–3 (Scheme 1) with the aim of synthesizing ribbon-
type compounds 4–6 with extended conjugation. Polymer 1 is
a linear polyphenylene precursor that can be subjected to
cyclodehydrogenation to yield conjugated nanoribbon 4. For
comparison, model systems 2 and 3 were developed with a
large number of dodecyl alkyl chains for better solubility. The
Figure 2. Structural elucidation of the dehydrogenated model com-
pound 8a by a) H NMR spectroscopy ([D8]THF, room temperature)
and b) MALDI-TOF MS.
synthesis of polymer 3 demonstrates how easily the width of
the resulting system can be extended.
Angew. Chem. Int. Ed. 2011, 50, 2540 –2543
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
2541

Communications
Information). This molecular weight corresponds to approx-
imately 105 repeat units, and a polymer-chain length of more
than 40 nm was calculated. As MALDI-TOF MS has
significant limitations for the detection of high-molecular-
weight species with a high polydispersity, it cannot be used to
determine the actual maximum molecular weights reached.[11]
Also, GPC analysis can only give approximate relative values,
which are calculated by using a polystyrene (PS) standard. For
this reason, we only discuss a few characteristic results
herein.[12] For polymer 2, a molecular weight of up to
20000 gmolÀ1 was detected by MALDI-TOF MS, and GPC
analysis with PS standards indicated an average molecular
weight of Mn = 9900 gmolÀ1 and a polydispersity index as low
as 1.40. After dehydrogenation, Mn = 20000 gmolÀ1 was
found for 5 by GPC analysis, in full agreement with our
expectations, owing to a considerable increase in the stiffness
during the conversion from a flexible polymer into a ribbon
with a higher hydrodynamic radius. Ribbon 4 was obtained as
an insoluble powder, whereas the cyclodehydrogenation of
alkyl-substituted polymers 2 and 3 yielded 5 and 6 as waxy
materials that could be dissolved in common organic solvents,
such as toluene, THF, and dichloromethane. This solubility
enabled detailed structural characterization in solution. The
good solubility of more than 40 mgmLÀ1 at room temperature
can be attributed to the steric demand of the peripheral alkyl
chains, which hinder aggregation in solution.
For the first time, ribbons 5 and 6 enabled characterization
by UV/Vis and photoluminescence spectroscopy in solution.
Strong broadening of the peaks in the absorption spectra
hindered analysis. However, in the photoluminescence spec-
tra, a red-shift of the emission maximum of almost 200 nm
between polymer 2 and ribbon 5 clearly indicated an
expanded chromophore, which leads to a smaller band gap
(see the Supporting Information).
By way of example, an oligomer 1* of polymer precursor 1
with seven repeat units (n = 6 with end capping at both ends)
was isolated and dehydrogenated (Figure 3a). It was possible
to monitor the loss of hydrogen atoms during the cyclo-
dehydrogenation by MALDI-TOF mass spectrometry and
compare this experimental result with a calculated result.
Figure 3a presents an enlarged segment of superimposed
mass spectra of the oligomeric precursor 1* and the corre-
sponding dehydrogenated ribbon 4*. It could be expected that
the terminal ortho-biphenyl group in the polymer would
rotate freely (see the red arrow in Figure 3a) and enable the
formation of two isomers after dehydrogenation. However, as
shown for model compound 7 (Figure 2a), only one specific
ribbon 4* was formed. The mass spectrum clearly shows that
exactly 16 hydrogen atoms were removed, in full accordance
with the expected result presented in the reaction scheme, to
yield the desired GNR with predictable structural perfection.
This result was further validated by Raman spectroscopy,
which provided relevant information about the extended
p conjugation of ribbon 4*.[13] The sample was subjected to
Raman spectroscopy in a KBr pellet with laser excitation at a
wavelength of 532 nm. The resulting first-order Raman
spectrum shows two important lines: a first band located at
1335 cmÀ1 with a shoulder at 1248 cmÀ1 and a second band
located at 1605 cmÀ1 (Figure 3b). This pattern is characteristic
Figure 3. a) Superimposed MALDI-TOF mass spectra of an oligomer
1* (7 repeat units, n=7) of polymer 1 (blue) and the corresponding
ribbon 4* (green). The spectra show the loss of exactly 16 hydrogen
atoms during cyclodehydrogenation. b) Raman spectrum of the dehy-
drogenated oligomer 4* showing that it is a ribbon with a length of
approximately 2.9 nm. c) Relevant spectral region of the IR spectra of
polymer 1* (red) and ribbon 4* (black) showing full disappearance of
the band at 4050 cmÀ1 (free phenyl rotation) upon cyclodehydrogena-
tion.
for all-benzenoid PAHs, with a unique fine structure at
1300 cmÀ1. The relative intensities and positions of the signals
correlate to the size and symmetry of the molecules. On the
2542
www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 2540 –2543

J. W. Lyding, Nat. Mater. 2009, 8, 235; d) Y.-W. Son, M. L. Cohen,
S. G. Louie, Nature 2006, 444, 347.
[2] X. Li, X. Wang, L. Zhang, S. Lee, H. Dai, Science 2008, 319, 1229.
[3] S. Stankovich, D. A. Dikin, R. D. Piner, K. A. Kohlhaas, A.
Kleinhammes, Y. Jia, Y. Wu, S. T. Nguyen, R. S. Ruoff, Carbon
2007, 45, 1558.
[4] a) M. Y. Han, B. ꢀzyilmaz, Y. Zhang, P. Kim, Phys. Rev. Lett.
2007, 98, 206805; b) Z. Chen, Y.-M. Lin, M. J. Rooks, P. Avouris,
Phys. E 2007, 40, 228.
[5] a) L. Jiao, X. Wang, G. Diankov, H. Wang, H. Dai, Nat.
Nanotechnol. 2010, 5, 321; b) D. V. Kosynkin, A. L. Higginbo-
tham, A. Sinitskii, J. R. Lomeda, A. Dimiev, B. K. Price, J. M.
Tour, Nature 2009, 458, 872.
basis of the model of so-called disordered graphite, it was
possible to calculate the La dimension of the conjugated
nanoribbon 4* to be on the order of 2.8–2.9 nm.[14] This value
shows remarkable agreement with the dimensions of 2.9 nm
computed on the basis of theoretical calculations (Spartan ’04,
minimized energy) of the heptamer shown in Figure 3a.
These results unambiguously validate that all possible bonds
within polymer 1* were constructed as indicated, and
conjugation is extended over the whole molecule. They are
further supported by comparison of the IR spectra of
precursor 1* and dehydrogenated ribbon 4* (Figure 3c). Of
particular diagnostic importance is the combination peak at
4050 cmÀ1. This band can be taken as a “marker” of the
presence of free rotating phenyl rings in the molecule and is
clearly observed in the spectrum of 1*. The disappearance of
this peak in the spectra of 4* and 4–6 indicates the absence of
noncondensed benzene rings in the molecule,[15] and thus
confirms the full cyclodehydrogenation of polymers 1–3.
In conclusion, it has been shown for the first time that
structural perfection in the classical bottom-up synthesis of
defined graphene nanoribbons 4 and 5 is possible through the
use of an oxidative cyclodehydrogenation with FeCl3. Rib-
bons with lengths of 40 nm were synthesized from polyphen-
ylene precursors 1 and 3 with a flexible kinked backbone. The
straightforward introduction of alkyl chains made the result-
ing ribbons soluble and thus enabled solution processability:
an important requirement for the large-scale preparation of
electronic devices. By following the above concept, one can
readily design other oligophenylene building blocks, which
upon (Suzuki- or Yamamoto-type) polymerization should
afford suitable precursor polymers for the fabrication of
GNRs with different aspect ratios and edge structures.
Examples documenting the power of chemical GNR synthesis
will follow.
[6] E. H. Fort, L. T. Scott, Angew. Chem. Int. Ed. 2010, 122, 6776 –
6778; Angew. Chem. Int. Ed. 2010, 49, 6626 – 6628.
[7] a) Y. Fogel, L. Zhi, A. Rouhanipour, D. Andrienko, H. J. Rꢁder,
K. Mꢂllen, Macromolecules 2009, 42, 6878; b) J. Wu, L. Gher-
ghel, M. D. Watson, J. Li, Z. Wang, C. D. Simpson, U. Kolb, K.
Mꢂllen, Macromolecules 2003, 36, 7082; c) X. Yang, X. Dou, A.
Rouhanipour, L. Zhi, H. J. Rꢁder, K. Mꢂllen, J. Am. Chem. Soc.
2008, 130, 4216.
[8] J. Cai, P. Ruffieux, R. Jaafar, M. Bieri, T. Braun, S. Blankenburg,
M. Muoth, A. P. Seitsonen, M. Saleh, X. Feng, K. Mꢂllen, R.
Fasel, Nature 2010, 466, 470.
[9] a) B. T. King, J. Kroulꢃk, C. R. Robertson, P. Rempala, C. L.
Hilton, J. D. Korinek, L. M. Gortari, J. Org. Chem. 2007, 72,
2279; b) C. D. Simpson, G. Mattersteig, K. Martin, L. Gherghel,
R. E. Bauer, H. J. Rꢁder, K. Mꢂllen, J. Am. Chem. Soc. 2004,
126, 3139; c) K. Yoshimura, L. Przybilla, S. Ito, J. D. Brand, M.
Wehmeir, H. J. Rꢁder, K. Mꢂllen, Macromol. Chem. Phys. 2001,
202, 215.
[10] a) M. Rehahn, A.-D. Schlꢂter, G. Wegner, W. J. Feast, Polymer
1989, 30, 1054; b) R. Kandre, K. Feldman, H. E. H. Meijer, P.
Smith, A. D. Schlꢂter, Angew. Chem. 2007, 119, 5044; Angew.
Chem. Int. Ed. 2007, 46, 4956.
[11] K. Martin, J. Spickermann, H. J. Rꢁder, K. Mꢂllen, Rapid
Commun. Mass Spectrom. 1996, 10, 1471.
[12] For GPC analysis, a PS standard instead of a poly(para-
phenylene) standard was used, as polymers 1–3 have a nonrigid
backbone.
[13] a) C. Mapelli, C. Castiglioni, E. Meroni, G. Zerbi, J. Mol. Struct.
1999, 480, 615; b) C. Mapelli, C. Castiglioni, G. Zerbi, K. Mꢂllen,
Phys. Rev. B 1999, 60, 12710; c) C. Castiglioni, C. Mapelli, F.
Negri, G. Zerbi, J. Chem. Phys. 2001, 114, 963; d) C. Castiglioni,
F. Negri, M. Rigolio, G. Zerbi, J. Chem. Phys. 2001, 115, 3769;
e) M. Rigolio, C. Castiglioni, G. Zerbi, F. Negri, J. Mol. Struct.
2001, 563, 79; f) F. Negri, C. Castiglioni, M. Tommasini, G. Zerbi,
J. Phys. Chem. A 2002, 106, 3306.
Received: October 20, 2010
Published online: January 12, 2011
Keywords: dehydrogenation · graphene · nanoribbons ·
.
Scholl reaction · Suzuki polymerization
[1] a) D. A. Areshkin, D. Gunlycke, C. T. White, Nano Lett. 2007, 7,
204; b) A. Castro Neto, F. Guinea, N. M. R. Peres, K. S. Novo-
selov, A. K. Geim, Rev. Mod. Phys. 2009, 81, 109; c) K. A. Ritter,
[14] A. C. Ferrari, J. Robertson, Phys. Rev. B 2000, 61, 14095.
[15] A. Centrone, L. Brambilla, T. Renouard, L. Gherghel, C. Mathis,
K. Mꢂllen, G. Zerbi, Carbon 2005, 43, 1593.
Angew. Chem. Int. Ed. 2011, 50, 2540 –2543
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
2543