Facile bucky-bowl synthesis by regiospecific cove-region closure by Hf elimination

Article abstract of DOI:10.1002/anie.201200516

Building bowls: An effective intramolecular aryl-aryl coupling is the key step in rational fullerene synthesis and in synthesis of extended buckybowl structures. Such a process can be embodied very efficiently through quantitative HF elimination on active Al₂O₃. The process is characterized by an unprecedentedly high chemoselectivity and regiospecificity. Copyright

Full text of DOI:10.1002/anie.201200516

.
Angewandte
Communications
DOI: 10.1002/anie.201200516
Aryl–Aryl Coupling
Facile Bucky-Bowl Synthesis by Regiospecific Cove-Region Closure by
HF Elimination**
K. Yu. Amsharov,* M. A. Kabdulov, and Martin Jansen
Rational synthesis of geodesic structures is of great interest, in
particular as a promising method for the synthesis of
individual fullerene isomers[1] and isomerically pure single-
wall carbon nanotubes.[2,3] The strategy generally followed
starts with the synthesis of a polycyclic aromatic hydrocarbon
(PAH) containing the carbon framework required for the
target molecule, which can be rolled up to the desired carbon
nanostructure by aryl–aryl domino coupling. The efficiency of
such a domino reaction has been demonstrated on the
synthesis of Ih-C60 and Td-C84 fullerene cages on the Pt(111)
surface (Figure 1).[4,5]
realized by flash vacuum pyrolysis (FVP).[6] The possibility of
selective fullerene cage formation through FVP has been
demonstrated at the examples of C60, C78, and C84.[7–9]
However, the conversion rates to the target molecules have
remained disappointingly low. The presence of halogen
atoms, such as chlorine or bromine, in the initial precursor
has been found to be essential to initiate such a process. This
approach has proven to be prolific for the synthesis of many
small buckybowl structures, but has reached its limits in the
case of large molecules.[10–13] Fluorine functionality was found
to be a good alternative.[11,14–16] Owing to its low atomic weight
and small size, the fluorine atom can be rather easily
introduced in the sterically hampered cove regions. More-
À
over, high chemical resistance of the C F bond extends the
scope of reactions that can be applied for the precursor
synthesis remarkably.[16] Previously we have demonstrated the
possibility of selective intramolecular ring closure by HF
elimination with FVP.[15] Nevertheless, the conversion rate
was found to be still too low (about 60% for a single
cyclization) to carry out a multifold process effectively. Apart
from modest yield and difficulty in scale-up, the main
drawback of the FVP technique is minimum functional
group tolerance and numerous side reactions caused by the
harsh reaction conditions. To overcome these disadvantages,
several alternative liquid-phase intermolecular aryl–aryl
coupling techniques have been developed. Here the palla-
dium(0)-catalyzed direct arylation utilizing bromo and chloro
derivatives or aryl triflates has been found to be effective for
the synthesis of small buckybowls.[17–20] Unfortunately, this
method can hardly be extended to fluorinated precursors,
which are expected to be significantly less reactive than
chlorinated and brominated analogues. Furthermore, the
main drawback of the approach is aggressive reaction
conditions with respect to the fullerene species.[21] The
Figure 1. Aryl–aryl domino coupling of the Ih-C60 precursor C60H30. The
cove and fjord regions where the direct aryl–aryl coupling can take
place are highlighted in orange. Each subsequent condensation step
leads to formation of new cove or fjord regions. Thus, the reaction can
proceed in a domino fashion, resulting in the formation of the desired
fullerene cage.
À
discovery of C F bond activation under transition-metal-
free conditions utilizing silylium carborane catalysis might be
a solution to this problem.[22] Recently Siegel et al. have
successfully applied this methodology to intramolecular aryl–
aryl coupling.[23] Unfortunately, the remarkable decrease in
yield in the case of nonplanar PAHs, which already drops
from 99% to 50–80% for single condensation, limits its
application for multifold reactions. It is worth mentioning that
all solution approaches might be not effective in the case of
large PAHs, which are typically highly insoluble compounds.
Moreover, the intramolecular aryl–aryl coupling cannot be
realized in a domino fashion by palladium(0)-catalyzed direct
arylation or by silylium carborane catalysis.
Although the platinum-catalyzed cyclodehydrogenation
process was found to occur in a highly selective manner, it is
limited to only picomole amount production. Alternatively,
the consecutive series of intramolecular condensations can be
[*] K. Y. Amsharov, M. A. Kabdulov, Prof. Dr. M. Jansen
Max-Planck-Institut fꢀr Festkçrperforschung, Chemistry III
Heisenbergstrasse 1, 70569 Stuttgart (Germany)
E-mail: k.amsharov@fkf.mpg.de
[**] We are grateful to Patrick Merz for assistance with the syntheses
and Dr. Jꢀrgen Nuss for performing the X-ray measurements.
In the search for a more effective method for ring closure
that is applicable for highly insoluble fullerene precursors, we
have turned our attention to solid-state catalysis. Herein, we
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201200516.
4594
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 4594 –4597

Angewandte
Chemie
present an efficient cove-region closure process (CRC)
realized by regiospecific HF elimination. Apart from the
simplicity and high reproducibility, the described technique
avoids all of the abovementioned limitations. Examination of
the CRC approach for synthesis of buckybowl structures has
demonstrated an unprecedentedly high selectivity and con-
version level.
The conversion of 1-fluorobenzo[c]phenanthrene (1) to
benzo[ghi]fluoranthene (2) was chosen as a model reaction
for the investigation. Benzo[c]phenanthrene is the smallest
molecule that can have a cove region where the new bond
should be established (Scheme 1). It was found that g-Al2O3
decomposition and high enough to achieve the full conversion
in a rather short period of time.
Further investigations have revealed an unexpectedly
high chemoselectivity and regiospecificity of the process.
Thus, if the fluorine atom is not involved directly in the cove
region (2-fluorobenzo[c]phenanthrene), no reaction takes
place and the compound remains completely intact. The
regiospecificity of the process has been demonstrated on the
synthesis of compound 4, in which cove-region closure in 3
was achieved quantitatively (Scheme 1). Surprisingly, chlori-
nated and even brominated analogues (2-chlorobenzo[c]phe-
nanthrene and 2-bromobenzo[c]phenanthrene) remain com-
pletely intact under condensation conditions. The high
chemoselectivity has been demonstrated by the synthesis of
compound 6, for which HF elimination was achieved in the
À
presence of a C Br bond in more than 98% yield. Impor-
tantly, no fluorine-to-hydrogen exchange was observed and
no evidence of any side reactions was obtained during the
CRC process. As monitored by HPLC analysis, the products
form exclusively and require no further purification proce-
dure (Supporting Information). Obviously, the driving force
À
of the reaction is the formation of an Al F bond which is
characterized by substantial bond strength. On the other
À
hand, the results obtained exclude direct C F bond activation
and indicate that no reactive intermediates form during the
CRC process. The most probable mechanism of the CRC
process includes cyclic transition states; that is, the conden-
sation takes place in a concerted fashion (Figure 2). This
Scheme 1. Cove-region closure (CRC) mediated by Al2O3. CRC by HF
elimination in 1-fluorobenzo[c]phenanthrene derivatives, demonstrat-
ing high chemoselectivity and regiospecificity of the process. Com-
pounds 2, 4, and 6 were obtained in quantitative or close to
quantitative yield.
Figure 2. A representation of the most probable mechanism of the
CRC for benzo[c]phenanthrene condensation, showing the key points
of the process: coordination on the aluminum oxide surface, aromatic
À
transition state, and high energy gain owing to Al F bond formation.
displays pronounced activity in ring closure in 1, which starts
above 3008C. In the case of benzo[ghi]fluoranthene, the high
temperature does not cause further decomposition and 1 can
be converted into 2 effectively. Nevertheless, the conditions
are too harsh for buckybowl synthesis. During the optimiza-
tion of the condensation conditions, it was discovered that the
high temperature is only required for g-Al2O3 activation,
whereas the condensation process can proceed smoothly at
lower temperatures. The best results were obtained after
activation of g-Al2O3 by annealing at 5008C for 15 min at
10À3 mbar pressure. Using such activated g-Al2O3, compound
1 can be quantitatively transformed to 2 at remarkably low
temperatures. Thus, quantitative conversion of 1 into 2 was
observed after 20 h at 1508C. Full conversion has even been
achieved at 1008C after 60 h. The same process can be
completed in 1 h or in just several minutes at 2008C and
2508C, respectively. For practical purposes, the temperature
interval of about 150–2008C was found to be the most
appropriate, as it is low enough to prevent possible product
process seems to be an intermolecular variant of the rare
intramolecular dyotropic rearrangement where two sigma
bonds migrate simultaneously through a six-electron transi-
tion state.[24,25] As bond breaking and bond formation takes
place simultaneously no reactive intermediates, which might
cause side reactions, form. An aromatic six-electron transition
state explains the low activation barrier allowing the reaction
to take place under mild conditions. Regiospecificity can also
be understood in terms of the assumed mechanism, as 1,2-HF
elimination is extremely unfavorable owing to the high energy
of benzyne (1,2-didehydrobenzene) species that should be
formed in that case. Interestingly, none of the carbon atoms
changes hybridization during CRC, which is atypical for an
elimination. The formal increase in unsaturation, a necessary
prerequisite of any elimination process, is realized through
the formation of a new ring in the system.
Angew. Chem. Int. Ed. 2012, 51, 4594 –4597
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
4595

.
Angewandte
Communications
Considering the multifold CRC, it is easy to see that each
new pentagon formation imposes additional strain to the
system as a result of perturbation of the aromatic system
owing to the curvature introduced (Figure 1). Normally, the
efficiency of each subsequent condensation step drops
remarkably or the condensation process even stops. The
suitability of the g-Al2O3-initiated CRC method for construc-
tion of nonplanar PAHs has been examined on the synthesis
of indacenopicene (9), which is known to be stable com-
pound.[26] Despite of the strain imposed, benzopicene 7 was
converted into the desired buckybowl 9 quantitatively under
mild conditions (1508C, 60 h). At 2508C, the same process can
be fully completed in 1 h. Interestingly, the product of mono
condensation (compound 8) cannot be detected in the
reaction even if the starting material was not fully converted,
indicating that the HF elimination in 8 proceeds even faster
than in benzopicene 7. Indacenopicene 9 was obtained in pure
form by simple extraction with hot toluene (yield of isolated
product 99%). Single-crystal X-ray analysis unambiguously
confirms the structure and reveals the geodesic shape of 9
(Figure 3).
been demonstrated by quantitative transformation of pre-
cursor molecules to the desired buckybowl structures. Taking
into consideration that fluorine can promote the desired ring
closure only if a hydrogen atom is neighboring in space in the
precursor structure, it seems to be possible to fully control the
process and carry out the condensation in domino fashion.
The high conversion level into extended buckybowl 11, which
formally represents more than 75% of the C60 fullerene
connectivity, demonstrates high potential of the CRC tech-
nique for generating extended nonplanar carbon based
nanostructures, including higher fullerenes, giant buckybowls,
and nanotubes. The high tolerance to chlorine and bromine
functionalities makes the CRC approach a very powerful
synthetic method for synthesis of functionalized buckybowls,
which are useful building blocks for the construction of
complex carbon-based nanostructures. In general, our finding
À
shows that the C F bond, which is widely believed to be the
most passive functionality, can be reconsidered as a useful
À
functional group allowing very effective C C bond formation.
Experimental Section
Cove-region closure experiments were carried out using commer-
cially available aluminum oxide (activated, neutral, 50–200 micron,
Acros). Typically, g-Al2O3 (2–3 g) was placed in glass ampule and
activated by annealing at 5008C for 10–15 min in vacuum (10À3 mbar).
The respective fluoroarene (20–50 mg) was mixed with activated
aluminum oxide under argon atmosphere. The ampule was evacuated
(10À3 mbar) and sealed. The condensation was carried out at 150–
2008C for 2–10 h. The corresponding product was obtained after
extraction with toluene (xylene). Synthesis procedures, HPLC
chromatograms, MS, NMR, and UV/Vis spectra of all of the new
compounds and crystallographic data for compound 9 can be found in
Supporting Information. CCDC 860433 (9) contains the supplemen-
tary crystallographic data for this paper. These data can be obtained
free of charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
Received: January 18, 2012
Published online: April 3, 2012
Keywords: alumina · aryl–aryl coupling · buckybowls ·
.
À
C F activation · HF elimination
Figure 3. Synthesis of buckybowls by the CRC approach. Top: Bucky-
bowl synthesis by two- and fourfold cove-region closure. Bottom:
ORTEP of compound 9 (top and side view) and DFT optimized
structure of 11 (top and side view), demonstrating the geodesic
shapes. Compounds 9 and 11 were obtained in quantitative yield.
[1] L. T. Scott, Angew. Chem. 2004, 116, 5102 – 5116; Angew. Chem.
Int. Ed. 2004, 43, 4994 – 5007, and references therein.
[2] T. J. Hill, R. K. Hughes, L. T. Scott, Tetrahedron 2008, 64, 11360 –
11369.
To address the solubility issue we investigated the CRC
approach for the badly soluble precursor 10. It has been found
that the fourfold HF elimination can be realized very
effectively. After deposition of 10 on an activated g-Al2O3
the mixture was annealed for 10 h at 2508C. Subsequent
Soxhlet extraction with o-xylene gave the desired buckybowl
11 in pure form as a hardly soluble red powder (yield of
isolated product 98%). No side products and no products of
partial condensation were found in the sample as indicated by
HPLC, MS, and NMR analysis (Supporting Information).
In summary, a highly effective solid-state strategy for
regiospecific aryl–aryl coupling (cove-region closure) by HF
elimination is presented. The efficiency of the approach has
[3] A. Mueller, K. Y. Amsharov, M. Jansen, Tetrahedron Lett. 2010,
51, 3221 – 3225.
[4] G. Otero, G. Biddau, C. Sanchez-Sanchez, R. Caillard, M. F.
Lopez, C. Rogero, F. J. Palomares, N. Cabello, M. A. Basanta, J.
Ortega, J. Mendez, A. M. Echavarren, R. Perez, B. Gomez-Lor,
J. A. Martin-Gago, Nature 2008, 454, 865 – 868.
[5] K. Y. Amsharov, N. Abdurakhmanova, S. Stepanow, S. Rau-
schenbach, M. Jansen, K. Kern, Angew. Chem. 2010, 122, 9582 –
9586; Angew. Chem. Int. Ed. 2010, 49, 9392 – 9396.
[6] R. F. C. Brown, Pyrolytic Methods in Organic Chemistry:
Application of flow and flash vacuum pyrolytic techniques,
Academic Press, New York, 1980, p. 101.
[7] L. T. Scott, M. M. Boorum, B. J. McMahon, S. Hagen, J. Mack, J.
Blank, H. Wegner, A. de Meijere, Science 2002, 295, 1500 – 1503.
4596 www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 4594 –4597

Angewandte
Chemie
[8] K. Y. Amsharov, M. Jansen, J. Org. Chem. 2008, 73, 2931 – 2934.
[9] K. Y. Amsharov, M. Jansen, Chem. Commun. 2009, 2691 – 2693.
[10] L. T. Scott, Pure Appl. Chem. 1996, 68, 291 – 300.
[11] M. J. Plater, M. Praveen, D. M. Schmidt, Fullerene Sci. Technol.
1997, 5, 781 – 800.
[12] G. Mehta, H. S. P. Rao, Tetrahedron Lett. 1998, 54, 13325 – 13370.
[13] V. M. Tsefrikas, L. T. Scott, Chem. Rev. 2006, 106, 4868 – 4884.
[14] K. Y. Amsharov, M. Kabdulov, M. Jansen, Eur. J. Org. Chem.
2009, 6328 – 6335.
[20] B. D. Steinberg, E. A. Jackson, A. S. Filatov, A. Wakamiya,
M. A. Petrukhina, L. T. Scott, J. Am. Chem. Soc. 2009, 131,
10537 – 10545.
[21] Our experiments (unpublished results) have revealed that C60
fullerene in particular does not survive under Pd0-catalyzed
coupling conditions reported in Ref. [20].
[22] C. Douvris, O. V. Ozerov, Science 2008, 321, 1188 – 1190.
[23] O. Allemann, S. Duttwyler, P. Romanato, K. K. Baldridge, J. S.
Siegel, Science 2011, 332, 574 – 577.
[15] K. Y. Amsharov, M. Kabdulov, M. Jansen, Chem. Eur. J. 2010, 16,
5868 – 5871.
[24] M. T. Reetz, Angew. Chem. 1972, 84, 161 – 162; Angew. Chem.
Int. Ed. Engl. 1972, 11, 129 – 130.
[16] M. Kabdulov, K. Y. Amsharov, M. Jansen, Tetrahedron 2010, 66,
8587 – 8593.
[25] L. A. Paquette, M. A. Kesselmayer, R. D. Rogers, J. Am. Chem.
Soc. 1990, 112, 284 – 291.
[17] A. M. Echavarren, B. Gomez-Lor, J. J. Gonzalez, O. De Frutos,
Synlett 2003, 585 – 597.
[18] D. Alberico, M. E. Scott, M. Lautens, Chem. Rev. 2007, 107, 174 –
238.
[26] Two indacenopicene derivatives were obtained previously by
double palladium-catalyzed cyclization from corresponding
chloroderivatives: L. Wang, P. B. Shevlin, Org. Lett. 2000, 2,
3703 – 3705.
[19] S. Pascual, P. de Mendoza, A. M. Echavarren, Org. Biomol.
Chem. 2007, 5, 2727 – 2734.
Angew. Chem. Int. Ed. 2012, 51, 4594 –4597
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
4597