Angewandte
Chemie
À
together with the demonstrated selectivity of C C bond
adjacent molecules stabilizes the random aggregates. Thus, a
formation, the only possibility left is the formation of a single
product, the structure of which is defined by the precursor
molecules; thus, isomer-specific species are very likely
formed.
low coverage of precursor molecules is necessary also on the
Pt(111) surface to avoid the interlinking of neighboring
molecules. On the other hand, the strong molecule–substrate
interaction is required for the successful completion of the
SCCDH process. Indeed, our STM observations of immobile
molecules and time-of-flight secondary-ion mass spectrome-
try (ToF-SIMS; see the Supporting Information) data indicate
that this is the case for Pt(111) substrate.
To conclude, we have shown that the SCCDH process is
highly selective in nature. Therefore, the final structure of any
kind of carbon-based nanostructure, such as a fullerene or
nanotube, can be programmed or built in at the precursor-
synthesis stage. Our findings enabled us to produce C84
fullerene for the first time. Although the quantities of the
final products are too small to permit direct confirmation of
the structural uniformity of the fullerene formed, we believe
that SCCDH is an efficient approach to the production of
various isomerically pure higher fullerenes and open-cage
fullerenes once the respective precursors have been synthe-
sized. SCCDH has been successfully conducted so far on
Pt(111) and Ru(0001)[13] substrates only. Therefore, other
substrates have to be explored with regard to whether the
SCCDH process can be carried out with optimal efficiency for
the desired carbon nanostructure. Furthermore, these find-
ings go even beyond SCCDH reactions. They prove the
principle suitability of prefabricated planar precursors for the
synthesis of bulk fullerenes.
Our STM experiments show that the conversion ratio of
the planar precursor molecules 1, 2, and 3 into the corre-
sponding fullerene and open-cage structures is nearly 100%
(that is, no desorption occurs during the annealing) and that
all precursors are transformed into nonplanar structures. The
variation in the shape and size distribution of the products is
related to many different possible adsorption geometries and
potential surface modification below the products. Our
findings demonstrate that the SCCDH method is a very
efficient approach towards the synthesis of fullerenes, open-
cage structures, heterofullerenes, and endofullerenes. There is
essentially no limitation to the variety of organic nonplanar
target molecules that could be accessed on the basis of this
retrosynthetic strategy.[21] The procedure could be extended
to the fabrication of carbon nanotubes of well-defined
diameter and chirality, since the buckybowls can be viewed
as a seed for carbon-nanotube growth by chemical vapor
deposition (CVD).
Besides the clear advantages of the SCCDH synthesis
method, there are some limitations. First, it can only be
applied in a well-defined temperature interval. Our results
show that the annealing temperature required for the
completion of SCCDH increases with the number of
C atoms in the precursor molecule (4808C for 1 and 2, and
5508C for 3). On the other hand, annealing at higher
À
temperatures of about 7208C leads to complete C C rear-
Experimental Section
The C60H30 fullerene precursor 1 was prepared according to a
previously described procedure.[25] The synthesis of the C84H42
precursor 3 is described elsewhere[11] (see the Supporting Information
rangement and subsequent decomposition of all the carbon
structures into a planar adlayer similar to that grown by CVD
on different substrates.[22] These results indicate that only a
small temperature window exists for each specimen for the
application of SCCDH, and that the SCCDH process on
Pt(111) for the synthesis of high-mass fullerenes might be
difficult. Therefore, appropriate substrates have to be
explored for the efficient conversion of precursor molecules
into the desired carbon nanostructures. Such substrates could
be based on platinum-group metals (Ru, Rh, Pd, Os, Ir, and
Pt), since these metals are good catalysts for hydrogenation or
dehydrogenation reactions.[23] Control measurements by
Otero et al.[14] for the C60 cyclization on an Au(111) surface
showed only very little fullerene formation.
for details). Precursor
2 was obtained by the reaction of
1-methylbromo-2-bromonaphthalene with truxene trianione. Subse-
quent palladium-catalyzed intramolecular arylation gives compound
2 (see the Supporting Information). All precursors were purified by
gradient sublimation.
Samples were prepared in an ultrahigh-vacuum (UHV) chamber
with a base pressure of 3 ꢀ 10À10 mbar. The single-crystal Pt(111)
substrate was cleaned by repeated cycles of sputtering with Ar+ ions
at a sample temperature of 6008C and annealing at 9608C. Precursor
molecules were deposited on the clean Pt(111) surface by organic
molecular beam epitaxy from a Knudsen cell type evaporator with the
sample kept at room temperature. Precursors 1 and 2 were sublimed
at 5408C; precursor 3 was sublimed at 5908C.
Evaporators were degassed for at least 14 h before deposition at
208C below the indicated evaporation temperature to avoid any
contamination. After deposition, the sample was imaged by variable-
temperature STM (Omicron GmbH). After the molecular coverage
and integrity of the precursor had been confirmed, the sample was
annealed for 2–15 min and then imaged with STM at room temper-
ature. Tunneling parameters of 0.1–0.5 nA and 0.5–1.1 V were used
for imaging. To determine the required annealing temperatures for
complete SCCDH, the temperature was increased stepwise in the
interval 420–7208C, and the surface was imaged with STM after each
annealing step. STM images were analyzed by using WSxM soft-
ware.[26]
To elucidate the role of the catalytic properties of the
substrate, we performed measurements on Cu(100), on which
the molecule–substrate interaction is stronger than on
Au(111) but significantly weaker than on Pt(111).[17,24] The
STM measurements demonstrated that the molecules were
mobile on the Cu(100) surface. After annealing at 3808C,
stable islands of randomly merged triangular-shaped mole-
cules evolved (see Figure S6 in the Supporting Information).
Annealing at 4808C yielded fuzzy mobile features, but no
fullerene formation was observed. These observations con-
trast with the results on the Pt(111) surface, on which the
molecules did not form islands at similar temperatures. They
indicate that thermally induced diffusion prevents closing of
Received: August 10, 2010
Revised: September 27, 2010
Published online: October 28, 2010
À
the cage: it is likely that the formation of C C bonds between
Angew. Chem. Int. Ed. 2010, 49, 9392 –9396
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
9395