The most stable and fully characterized functionalized heptacene

Article abstract of DOI:10.1002/anie.200803345

(Chemical Equation Presented) Lucky seven: Crystals of heptacene, which consists of seven linearly fused benzene rings, remain stable for at least 21 days in mineral oil and 24 h in degassed solution when the oligoacene is functionalized with phenyl and t

Full text of DOI:10.1002/anie.200803345

Communications
DOI: 10.1002/anie.200803345
Oligoacenes
The Most Stable and Fully Characterized Functionalized
Heptacene**
Doris Chun, Yang Cheng, and Fred Wudl*
Angewandte
Chemie
8380
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 8380 –8385

Angewandte
Chemie
Linear polyacenes 2 are one of the most fascinating classes of
organic molecules.[1,2] Benzene (2, n = 1), the smallest
member of the family, is the most stable known aromatic
hydrocarbon. Since the discovery of benzene in
to afford the desired product.[10–15] We explored other
methodologies to provide a new route to functionalized
higher acenes. To the best of our knowledge, the work
reported herein is the first report of a double Diels–Alder
cycloaddition between a lateral “bisanthracyne” and dienes,
followed by reduction to give heptacene derivatives 1a–c
(Scheme 1).
the fractional distillation of coal extracts by
Faraday in 1825,[3] naphthalene (2, n = 2) and
anthracene (2, n = 3), were also isolated from
coal-tar and petroleum distillates.[1] The higher
homologues (2, n > 3) however, are not found in
nature and can be obtained only through multi-
step syntheses. The electronic coupling of their extensively
delocalized p orbitals in the solid state means that these
linearly fused polycyclic aromatic hydrocarbons (PAHs) are
intrinsic semiconductors.
Scheme 1. The molecular structure of heptacene derivatives 1a–c.
Theoretical calculations predict a narrowing of band gap
energies in the acene family with each successive benzene ring
addition up to hexacene, beyond which the HOMO–LUMO
(HOMO = highest occupied molecular orbital, LUMO =
lowest unoccupied molecular orbital) gap remains constant
because of the emergence of a singlet biradical character in
the ground state.[4,5] Hence, the electronic properties of the
longer acenes have long been a subject of debate.[6] While the
physical properties of the smaller linear PAHs (2 with n ꢀ 5)
are well explored, our knowledge of longer acenes is limited
for two major reasons: 1) the synthetic challenge in their
production and their insolubility and 2) their extreme insta-
bility with increased length, which makes them character-
izable only by electronic spectroscopy in a matrix at low
temperatures. Neckers and co-workers have reported stability
studies of hexacene and heptacene generated photochemi-
cally in PMMA matrix.[7,8] However, these molecules could
not be isolated and they are stable for only a few hours when
protected by the polymer matrix.[7,8] An effective approach to
stabilize and solubilize long PAHs was demonstrated by
Anthony and co-workers who strategically functionalized the
acene backbone with alkylsilylethynyl groups.[9]
The 2,5-diaryl-6-oxo-1,3,4-oxadiazine-6-one 3[16] is an
electron-deficient diene that undergoes an inverse-electron-
demand Diels–Alder reaction twice with benzyne to give the
resulting anthracene derivative 4.[17] Trapping of the “bisan-
thracyne” (generated through sequential dehydrohalogena-
tion of dibromide 4) with diphenylisobenzofuran afforded bis-
endo-oxide 5. Metal reduction of 5 provided the expected
functionalized heptacenes 1a,b (Scheme 2).
The most common synthetic route to functionalized
polyacenes is through the nucleophilic addition of organo-
metallic reagents to an acene quinone, followed by reduction
Scheme 2. a) 1. 5-bromoanthranilic acid, isoamyl nitrite, THF, 08C.
2. DCE, reflux, 30%; b) Diphenylisobenzofuran, LTMP, THF, 08C!RT,
32%; c) Fe metal, AcOH, o-DCB, 1008C. DCE=dichloroethane,
LTMP=lithium tetramethylpiperidide
[*] D. Chun
Department of Chemistry and Biochemistry, University of California
Los Angeles, CA 90095-1569 (USA)
D. Chun, Dr. Y. Cheng,[+] Prof. Dr. F. Wudl
Department of Chemistry and Biochemistry 9510, University of
California
Santa Barbara, CA 93106-9510 (USA)
Fax: (+1)805-893-4120
Heptacene 1a was insoluble even in aromatic solvents. It
readily precipitated as a deep-green solid upon cooling the
reaction mixture in o-dichlorobenzene (o-DCB; Figure 1).
Attempts to dissolve 1a in o-DCB by sonication and heating
facilitated its oxidation to form an endo-peroxide. The UV/
Vis spectrum of 1a in o-DCB is very broad because of
aggregation; nonetheless, the expected vibronic pattern at
long wavelengths is present. Heptacene 1b was synthesized in
the same manner as 1a, except the aryl hydrazide and aryl
glyoxylic acid needed to be first synthesized. Although the
alkoxy side chains significantly improved its solubility, 1b was
too reactive to obtain an electronic spectrum of the pure
compound. Nonetheless, as shown in Figure 2, the vibronic
pattern of 1a and 1b in the near-IR region is undoubtedly due
to the presence of heptacene. The strong absorption with
E-mail: wudl@chem.ucsb.edu
Homepage: www.chem.ucsb.edu/
[+] Current Address: Core Research & Development, The Dow
Chemical Company
Midland, MI 48674 (USA)
[**] We are indebted to Dr. Guan Wu for his assistance with the single-
crystal structure elucidation. We thank Prof. John Anthony for
providing valuable insight and data that helped facilitate our
characterization. We also thank Prof. Galen Stucky for helpful
suggestions which led to the successful collection of the X-ray data
and Brittnee Veldman for technical assistance with the UV/Vis–NIR
data. This work was funded by the NSF through grant DMR
9796302.
Angew. Chem. Int. Ed. 2008, 47, 8380 –8385
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
8381

Communications
steps (Scheme 3). Unlike the metal reduction of 5, bis-endo-
oxide 8 did not undergo reduction under the same reaction
conditions; a stronger reducing metal, zinc in this case, had to
be used.
Figure 1. A suspension of 5,7,9,14,16,18-hexaphenyl heptacene 1a in
o-DCB (left) and a reaction mixture containing 5,9,14,18-tetraphenyl-
7,16-bis[4-(2-ethylhexyloxy)phenyl]heptacene 1b in toluene (right).
Scheme 3. a) 1. iPr3SiCCH, nBuLi, THF, 08C–RT, 2. SnCl2, 50% AcOH,
86% overall; 3. LTMP, diphenylisobenzofuran, THF, ꢁ108C!RT, 32%;
c) Zn dust, AcOH, toluene, 1008C, 4 days, 73%.
Upon slow cooling in an oil bath, the reaction mixture of
1c produced high-quality crystals for X-ray structural analysis
(Figure 4). These opaque prismatic crystals are stable when
coated with mineral oil even under constant exposure to the
laboratory atmosphere and light for over 21 days. Interest-
ingly, 1c exhibits edge-to-face herringbone packing in the
solid state (Figure 5), however there are no p–p interactions
between the acene backbones. A stability study in toluene was
accomplished by monitoring the disappearance of vibronic
absorptions in the near-IR region (Figure 6). Heptacene 1c is
significantly more stable than any other heptacene derivatives
synthesized to date: Its presence was still detectable by UV/
Vis/NIR spectroscopy in degassed toluene after 41 h of
exposure to air. When 1c is dissolved in a degassed solvent
Figure 2. UV/Vis absorption spectra of 1a (solid line) and 1b (dotted
line).
onset at 540 nm arises from the immediate oxidation of 1b
upon exposure to air and light, despite sample preparation
under strictly anaerobic conditions. After 15 min, the color of
the reaction mixture of 1b changed from dark brown to bright
orange and the UV/Vis spectrum showed a clear vibronic
pattern, which indicated a tetracene-containing compound
(Figure 3).
Among the three heptacene derivatives only 1c was
sufficiently stable to be isolated and characterized. It was
synthesized from 2,6-dibromo-anthraquinone 6[18] in four
Figure 3. UV/Vis spectrum of 1b 15 min after exposure to light and
air. The exposed sample is orange, which indicates the presence of a
tetracene species
Figure 4. X-ray crystal structure and the single crystal of heptacene 1c.
8382
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 8380 –8385

Angewandte
Chemie
Scheme 4. Proposed structure of oxidation product.
oxidation, namely the 6,17 or 8,15 positions. By following the
evolution of absorption patterns in the UV/Vis–NIR spectra,
it is apparent that the reaction did not terminate with the
addition of one molecule of oxygen. The absorptions at
535 nm decrease in intensity while the absorptions in the UV
region intensify. It is very likely that both 6, 17 and 8, 15
positions underwent cycloaddition. The fact that no dimer was
isolated, even after recrystallization for three weeks, indicates
the effectiveness of phenyl groups in preventing self-reaction
better of 1c.
Figure 5. The herringbone packing motif of 1c.
It appears that besides the steric hindrance imposed by
the TIPS-acetylene groups (TIPS = triisopropylsilyl), it is also
possible that electronic effects may have helped stabilize 1c.
This is a conclusion drawn from two observations. Firstly.
according to the current proposed empirical model for
engineering crystalline derivatives using an alkylsilylacety-
lene group, the solubilizing group needs to have a diameter of
35–50% of the length of the acene to have a significant
stabilizing effect.[12] However, the TIPS groups in 1c are
significantly smaller than the suggested range, yet the acene is
the most stable in the group 1a–c, as well as the reported
bisTTMSS heptacene (TTMSS = tris(trimethylsilyl)silane).[9]
Secondly, the onset of the NIR absorption of 1c is red-shifted
by approximately 20 nm compared to compounds 1a and 1b,
which indicates an electronic effect of the TIPS acetylene
groups on the acene backbone (Figure 7).
Figure 6. Degradation study of 1c in toluene under ambient condi-
tions. The absorption of 1c at 0 min is shown as an inset. Over the
course of 41 h, with one spectrum taken every six hours, the NIR
absorption decreases while a new vibronic absorption appears (onset
at 550 nm).
The bandgap width was further confirmed by cyclic
voltammetry (Figure 8). Heptacene 1c has one reversible
reduction at ꢁ1.13 eVand one reversible oxidation at 0.25 eV
(vs. Ag/Ag+ reference electrode). This results in a bandgap of
1.38 eV, which is in close agreement with the bandgap
extrapolated from the Vis/NIR onset at 917 nm (1.35 eV).
Although small molecules are expected to have a significant
discrepancy between the electrochemical bandgap energies
onset of electronic absorption, this heptacene derivative, with
a molecular weight of 1042 Da, is large enough to be
considered an equivalent of two oligoacetylenes, hence the
general trend observed in small molecules is unlikely to be
applicable, which explains the small discrepancy between the
two methods for band gap determination. When the cyclic
voltammetry (CV) experiment of 1c was extended to 1 eV, a
strong oxidation wave that was only partially reversible was
observed. The HOMO and LUMO levels were calculated
with ferrocene as internal standard to give a HOMO and
LUMO levels of approximately ꢁ4.8 eVand ꢁ3.5 eV, respec-
tively.
and kept in a capped NMR tube, only minor decomposition
1
was observed in the H NMR spectrum after 24 h. Based on
1
UV/Vis absorption, H NMR spectroscopy, and FAB mass
spectrometry results, we believe that 1c undergoes photo-
oxygenation to give a tetracene-containing structure 10
(Scheme 4). The growth of a new vibronic pattern with
onset at 550 nm and the appearance of a signal with m/z 1075
[M+H] in the FAB mass spectrum, corresponding to the
dioxygen adduct of 1c, suggested that a new oxidized species
had formed (Figure 6). The formation of 10 was also shown by
the change in solution color from brown to orange, which
supports the presence of a tetracene moiety. Although the
signal at m/z 1075 has the same m/z value as the starting
material 8, it arises from addition of an oxygen molecule to
heptacene 1c during analysis. The reason for this oxidation
pattern can be deduced from analysis of the chemical
structure of 1c. Since the most reactive 7,16 positions are
protected by the triisopropylsilylethynyl groups, the next
reactive positions that are not functionalized can undergo
Angew. Chem. Int. Ed. 2008, 47, 8380 –8385
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
8383

Communications
immediately. nBuLi was titrated with 2,2’ dipyridyl in anhydrous THF
to determine the correct concentration prior to use. 1H and 13C NMR
spectra were recorded on a Bruker AVANCE500 spectrometer in
CDCl3 with tetramethylsilane as the internal standard, unless other-
wise specified. Mass spectrometry data were obtained from a VG70
Magnetic Sector instrument in FAB mode. Electrochemical experi-
ments were performed with a Princeton Applied Research Potentio-
stat/Gavanostat model 263Awith Bu4NPF6 (0.1m) in dichloromethane
containing ferrocene as the internal standard. X-ray crystallographic
analysis was carried out on a Bruker 3-axis platform diffractometer.
General procedure for the synthesis of 5a, 5b, and 8: nBuLi
(1.30 mmol) was added dropwise to a separate solution of 2,2,6,6-
tetramethypiperdine (1.30 mmol) in dry THF (10 mL) at 08C. The
mixture was allowed to stir at 08C for 20 min. The resulting lithium
tetramethylpiperidide was slowly added by cannula (ca. 20 min) to a
solution of 2,6-dibromoanthracene (0.216 mmol)and diphenylisoben-
zofuran (0.648 mmol) in dry THF (10 mL) at 08C. This solution had
been degassed by bubbling argon, which had been passed through
activated oxygen-scavenger RIDOX and CaSO4, for 30 min with
protection from light. The reaction was allowed to proceed in the dark
and to reach room temperature overnight, it was then quenched by
pouring into 20 mL of saturated NH4Cl, the aqueous layer was then
extracted with ethyl acetate (3 ꢀ 10 mL). The combined organic layers
were washed with water (2 ꢀ 15 mL), brine (1 ꢀ 15 mL) and then dried
over MgSO4. Upon evaporation of all solvents under reduced
pressure on a rotary evaporator, the crude product was chromato-
graphed on a SiO2 column with an eluent gradient starting with
hexanes, followed by 1–5% ethyl acetate in hexanes. Yield: 32–34%.
5a (syn and anti isomers): 1H NMR (500 MHz, CDCl3): d =
7.805–7.770 (m, 8H), 7.562–7.399 (m, 22H), 7.304–7.226 (m, 8H),
7.049 (dd, J1 = 3 Hz, J2 = 6 Hz, 2H), 7.018 ppm (dd, J1 = 3 Hz, J2 =
6 Hz, 2H) MS(FAB+): m/z 867.24 [M+H].
Figure 7. Bathochromic shift in the NIR absorption that results from
the presence of TIPS groups in heptacene 1c (solid line) compared to
heptacene 1b (dotted line).
5b (syn and anti isomers): 1H NMR (500 MHz, CDCl3): d =
7.839–7.801 (m, 8H), 7.573, 7.537 (2 ꢀ s, 4H), 7.491–7.406 (m, 12H),
7.301 (dd, J1 = 3 Hz, J2 = 5.5 Hz, 3H), 7.179–7.125 (m, 4H), 7.087–
7.010 (m, 7H), 6.957 (dd, J1 = 2 Hz, J2 = 8.5 Hz, 2H), 3.993 (dd, J1 =
3 Hz, J2 = 6 Hz, 4H), 1.576 (m, 2H), 1.445 (m, 16H), 1.056 (t, J =
7.5 Hz, 6H), 1.004 ppm (t, J = 7.5 Hz, 6H). MS(MALDI CH2Cl2/a-
cyano-4-hydroxycinnamic acid) m/z 1123.7 [M+H].
Figure 8. Cyclic voltammogram (vs. Ag/Ag+) of 1c with Bu4NPF6 in
DCM solution (0.1m), cycled 10 times.
1
8 (syn and anti isomers): H NMR (500 MHz, CDCl3) d = 8.399,
Of the three heptacene derivatives synthesized, only the
bis(alkylsilylethynyl)tetraphenyl derivative 1c was suffi-
ciently stable to allow characterization beyond UV/Vis–NIR
spectroscopy. While phenyl groups are effective in preventing
dimerization and polymerization of the longer acene, the
central ring must be functionalized with alkylsilylethynyl
groups in order to provide sufficient stability for isolation and
characterization. Like pentacene, 1c packs edge-to-face in
single crystals. The band gap determined from CV and NIR
onset supports the theoretical prediction of narrow but
nonzero band gap;[4,5] however, judging from the sharp
NMR splittings and narrow line widths, if there is a singlet
diradical character in the ground state,[4,5] it is detectable only
by the high reactivity of the molecules and not spectroscopi-
cally. As long as heptacene 1c is protected from oxygen, it can
be exposed to light, and will remain stable for extended
periods. We are currently exploring the potential electronic
applications of this exciting oligoacene.
8.388 (2 ꢀ s, 4H), 8.030–8.006 (m, 8H), 7.603–7.565 (m, 8H), 7.515–
7.493 (m, 4H), 7.456 (dd, J = 3 Hz, 2H), 7.432 (dd, J = 3 Hz, 2H),
7.085 (m, 4H), 1.162–1.118 ppm (m, 42H). 13C NMR (500 MHz,
CDCl3): d = 149.31, 149.28, 147.33, 135.09, 135.07, 131.93, 129.05,
128.45, 126.90, 126.73, 126.71, 120.87, 120.80, 119.63, 118.05, 104.52,
104.50, 103.51, 90.22, 19.09, 19.07, 11.61 ppm. MS (FAB+): m/z 1075
[M + ] 1076 [M+H].
1c: Glacial acetic acid (0.5 mL) and zinc dust (< 10 micron, ca.
50 mg, 0.76 mmol) was added to a suspension of bis-endo-oxide 8
(28 mg, 0.026 mmol) in toluene (10 mL) in a 25 mL Schlenk tube
equipped with a stirrer bar. The mixture was degassed by three freeze-
pump-thaw cycles and then heated to 1008C under argon for 4 days
with protection from light. The reaction was monitored by the
changes in the color of the solution over time; it was considered
complete when the color was dark brown and exhibited no visible
fluorescence. The hot reaction mixture was allowed to cool slowly in
the oil bath in the dark. High quality opaque prismatic crystals
(20 mg, 73% yield) were obtained after 3 weeks. 1H NMR (500 MHz,
CD2Cl2): d = 9.136 (s, 4H), 7.645 (t, J = 1.4 Hz, 8H), 7.575 (tt, J1 =
1.3 Hz, J2 = 7.4 Hz, 4H), 7.532 (dd, J1 = 1.5 Hz, J2 = 6.7 Hz, 8H),
7.3771 (dd, J1 = 3.3 Hz, J2 = 7 Hz, 4H), 7.107 (dd, J1 = 3.3 Hz, J2 =
7 Hz, 4H), 1.025 (d, J = 7.3 Hz, 36H), 0.874 ppm (m, 6H). 13C-apt-
NMR (500 MHz, CDCl3): d = 131.81, 129.24, 128.21, 127.63, 126.80,
125.68 (all aromatic CH), 19.14 (iPr CH3), 12.12 ppm (iPr CH). MS
(FAB): m/z 1042 [M + ], 1043 [M+H], 1044 [M+2H]. UV/Vis/NIR
(toluene): labs = 355, 383, 437, 463, 696, 772, 863 nm.
Experimental Section
All general reagents were purchased from Aldrich and were used
without further purification unless otherwise specified. Tetrahydro-
furan was distilled from sodium and benzophenone and was used
8384
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 8380 –8385

Angewandte
Chemie
1a and 1b: The procedure for the synthesis of 1a and 1b was the
same as for 1c. The only characterization data obtained for 1a and 1b
are the UV/Vis absorption spectra described above.
[4] M. Bendikov, H. M. Duong, K. Starkey, K. N. Houk, E. A.
Carter, F. Wudl, J. Am. Chem. Soc. 2004, 126, 10493.
[5] M. Bendikov, H. M. Duong, K. Starkey, K. N. Houk, E. A.
Carter, F. Wudl, J. Am. Chem. Soc. 2004, 126, 7416.
[6] D. E. Jiang, S. Dai, J. Phys. Chem. A 2008, 112, 332.
[7] R. Mondal, R. M. Adhikari, B. K. Shah, D. C. Neckers, Org. Lett.
2007, 9, 2505.
[8] R. Mondal, B. K. Shah, D. C. Neckers, J. Am. Chem. Soc. 2006,
128, 9612.
[9] M. M. Payne, S. R. Parkin, J. E. Anthony, J. Am. Chem. Soc.
2005, 127, 8028.
[10] M. S. Taylor, T. M. Swager, Angew. Chem. 2007, 119, 8632;
Angew. Chem. Int. Ed. 2007, 46, 8480.
[11] J. E. Anthony, J. S. Brooks, D. L. Eaton, S. R. Parkin, J. Am.
Chem. Soc. 2001, 123, 9482.
[12] J. E. Anthony, D. L. Eaton, S. R. Parkin, Org. Lett. 2002, 4, 15.
[13] J. H. Chen, S. Subramanian, S. R. Parkin, M. Siegler, K. Gallup,
C. Haughn, D. C. Martin, J. E. Anthony, J. Mater. Chem. 2008,
18, 1961.
[14] S. K. Park, T. N. Jackson, J. E. Anthony, D. A. Mourey, Appl.
Phys. Lett. 2007, 91, 063514.
X-ray single crystal structure analysis: The dark-brown heptacene
crystal 1c of approximate dimensions 0.25 ꢀ 0.25 ꢀ 0.2 mm was
mounted on a glass fiber and transferred to a Bruker CCD platform
diffractometer. The SMART[19] program was used to determine the
unit cell parameters and data collection (15 sec/frame, 0.3 deg./frame
for a sphere of diffraction data). The data were collected at 150 K
using an Oxford nitrogen gas cryostream system. The raw frame data
were processed using the SAINT[20] program. Empirical absorption
corrections were applied based on psi-scan. Subsequent calculations
were carried out using SHELXTL[21] program. The structure was
solved by direct method and refined on F2 by full-matrix least-squares
techniques. Hydrogen atoms were located from the difference map.
At convergence, R1 = 0.0641 for 3207 reflections with I > 2s(I) and
GOF = 0.996. Rint = 0.0710, wR = 0.1267. The orthorhombic crystals
belong to space group Pccn, with a = 28.481(8) ꢁ, a = 908, b =
14.140(5) ꢁ, b = 908, c = 14.869(5) ꢁ, g = 908, Z = 4. CCDC 692899
contains the supplementary 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
[15] C. D. Sheraw, T. N. Jackson, D. L. Eaton, J. E. Anthony, Adv.
Mater. 2003, 15, 2009.
[16] W. Steglich, E. Buschmann, G. Gansen, L. Wilschowitz, Syn-
thesis 1977, 252.
Received: July 10, 2008
[17] This dibromoanthracene derivative was synthesized under the
same reaction condition as reported in Ref. [16], except instead
of anthranilic acid, 5-bromoanthranilic acid was used.
[18] J. H. Park, D. S. Chung, J. W. Park, T. Ahn, H. Kong, Y. K. Jung,
J. Lee, M. H. Yi, C. E. Park, S. K. Kwon, H. K. Shim, Org. Lett.
2007, 9, 2573.
[19] SMART Software Users Guide, Version 5.1, Bruker Analytical
X-ray Systems, Inc., Madison, WI, 1999.
[20] SAINT Software Users Guide, Version 5.1, Bruker Analytical
X-ray Systems, Inc., Madison, WI, 1999.
Published online: September 29, 2008
Keywords: arynes · heptacenes · pericyclic reaction · polycycles ·
.
semiconductors
[1] J. E. Anthony, Angew. Chem. 2008, 120, 460; Angew. Chem. Int.
Ed. 2008, 47, 452.
[2] M. Bendikov, F. Wudl, D. F. Perepichka, Chem. Rev. 2004, 104,
4891.
[21] Sheldrick, G. M. SHELXTL, Version 6.12, Bruker Analytical X-
ray Systems, Inc., Madison, WI, 2001.
[3] M. Faraday, Philos. Trans. R. Soc. London 1825, 440.
Angew. Chem. Int. Ed. 2008, 47, 8380 –8385
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
8385