Syntheses of N-bridged ferrocene/porphyrin-fullerene dyads and influence of iminofullerene isomers on the attached chromophores

Article abstract of DOI:10.1016/j.tetlet.2013.01.061

N-Ferrocene (Fc)/porphyrin (Por) [5,6]-open azafulleroids and [6,6]-closed aziridinofullerenes have been synthesized for the first time. Electrochemical measurements revealed that chromophores attached to [5,6]-open azafulleroids are more easily oxidizabl

Full text of DOI:10.1016/j.tetlet.2013.01.061

Tetrahedron Letters 54 (2013) 1607–1611
Contents lists available at SciVerse ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Syntheses of N-bridged ferrocene/porphyrin-fullerene dyads and influence
of iminofullerene isomers on the attached chromophores
⇑
Chen Chen, Yi-Zhou Zhu, Hai-Ying Zhao, Jian-Yu Zheng
State Key Laboratory and Institute of Elemento-Organic Chemistry, Nankai University, Tianjin 300071, China
a r t i c l e i n f o
a b s t r a c t
Article history:
N-Ferrocene (Fc)/porphyrin (Por) [5,6]-open azafulleroids and [6,6]-closed aziridinofullerenes have been
synthesized for the first time. Electrochemical measurements revealed that chromophores attached to
[5,6]-open azafulleroids are more easily oxidizable than those attached to [6,6]-closed aziridinofulle-
renes. This was attributed to the delocalization degree of the nitrogen lone-electron pair involved in azir-
idinofullerene that is larger than that in azafulleroid.
Received 22 November 2012
Revised 6 January 2013
Accepted 15 January 2013
Available online 21 January 2013
Ó 2013 Elsevier Ltd. All rights reserved.
Keywords:
[5,6]-Open azafulleroid
[6,6]-Closed aziridinofullerene
Ferrocene
Porphyrin
[60]Fullerenes have attracted wide attention because of their
excellent electronic properties such as high electron affinity, small
reorganization energy, and high singlet oxygen quantum yield, and
have been widely applied in photovoltaics, organic field-effect
transistors, nonlinear optics, and biological area.1 Great efforts
are being continuously devoted to achieving new fullerene deriva-
tives for specific function.2 Many fullerene modification methodol-
ogies have been developed, including hydrogenation, transition
metal complex formation, nucleophilic addition, radical addition,
and cycloaddition.3 Among them, treating fullerene with organic
azide has been recognized as a convenient and practical way to
complete the modification because of its easy preparation, and
good tolerance to water and oxygen.4,5
A series of organic azides have been applied in the syntheses of
iminofullerenes, such as alkyl azides,6 simple aryl azides (phenyl,
naphthyl, pyrenyl),7 sulfonyl azides,8 and tetrafluoroarylazides.9
They can react facilely with fullerene to yield [6,6]-closed aziridi-
nofullerenes or [5,6]-open azafulleroids through a nitrene addition
or a 1,3-dipolar cycloaddition reaction.10 In most cases, a mixture
was obtained simultaneously, and the ratio of [6,6]-closed aziridi-
nofullerenes to [5,6]-open azafulleroids is greatly dependant on
the nature of the introduced substituent.6f Further exploring shows
that the N-substituted group has a significantly remote control
effect on the features of the excited states of fullerene.7 Taking
the mutual interaction into account, the distinction between the
inherent electronic properties of [5,6]-open azafulleroids and
[6,6]-closed aziridinofullerenes is reasonably suggested to result
in a different influence on the electronic features of the N-substi-
tuted groups.11 Shinkai and co-workers had found that a slight dif-
ference in the electron density of the nitrogen atom ([5,6]-open vs
[6,6]-closed) can cause a dramatic change of its guest affinity.12 But
to the best of our knowledge, there is no more example to explore
the influence of the fullerene
p system ([5,6]-open and [6,6]-
closed) on its attached molecule, which may guide us to find a
broadened approach to improve the practical application of
fullerene.
Herein, we report the syntheses of Fc- and Por-iminofullerene
dyads through treating fullerene with the corresponding azides.
Due to the rich electronic active properties of Fc and Por and their
sensitivities to the adjacent substituents,13 the constructed dyads
provide us a convenient visualization tool to estimate the influence
of the iminofullerene.
The ferrocenylazide was first introduced to react with fullerene
in chlorobenzene at 130 °C according to the literature method,6–9
but only insolubles were obtained probably because of the instabil-
ity of ferrocenylazide.14 The reaction temperature was then re-
duced to 60 °C, a main product with 16% yield was got and
identified to be [6,6]-closed aziridinofullerene 1b (Scheme 1). A
molecule ion peak [M]+ was observed at 919.013 m/z (MALDI),
and the 13C NMR spectrum reveals 17 signals between 146 and
124 ppm and one sp3 resonance at 85.4 ppm (Fig. S9, 1b) indicating
a C2v-symmetry of fullerene and a [6,6]-closed form.6g,8 This struc-
ture might come from a [2+1] nitrene addition procedure instead
of a [3+2] cycloaddition because the thermal instability of ferroce-
nylazide and a [5,6]-open form is usually the major product in a
[3+2] cycloaddition reaction according to the reported work.10
⇑
Corresponding author. Tel./fax: +86 022 2350 5572.
E-mail address: jyzheng@nankai.edu.cn (J.-Y. Zheng).
0040-4039/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.tetlet.2013.01.061

1608
C. Chen et al. / Tetrahedron Letters 54 (2013) 1607–1611
The UV/vis absorption spectra of
1
(1.5 Â 10À5 M) and
2 (2 Â 10À6 M) in toluene are shown in Figures S22 and S23, and
the absorption data are listed in Table 1. The characteristic absorp-
tion for aziridinofullerene derivatives was observed at 427 nm for
1b. The absorption of fullerene moiety in both azafulleroids 1a and
2a is bathochromically shifted compared to the corresponding azir-
idinofullerenes 1b and 2b. Meanwhile, the absorption of porphyrin
moiety in 2a shows slight but unambiguous 1 nm red shift than that
of porphyrin in 2b. This difference should result from the distinction
of two iminofullerene types, since there is only a slight difference of
structure between 2a and 2b.
N
Fe
N3
Fe
3
ClPh
60oC
1a
trace
+
N
Fe
Further examination was done by cyclic voltammetry to esti-
mate the ground state electronic interaction between fullerene
and the attached chromophore (Fc or Por). The cyclic voltammo-
grams of porphyrin-fullerene dyads are shown in Figure 1. The
experiments were carried out in benzonitrile at room temperature
and the data are summarized in Table 2. The structures of reference
compounds are shown in Scheme 3.
main product
1b
Scheme 1. Preparation of N-ferrocene [5,6]-open azafulleroid 1a and [6,6]-closed
aziridinofullerene 1b.
For Fc-fullerene dyads, three quasi reversible one-electron
reduction waves were observed for 1a (À0.931, À1.355,
À1.805 V) and 1b (À0.932, À1.368, À1.805 V) corresponding to
the first three reduction processes of C60. The porphyrin-fullerene
dyads were found to have the similar fullerene reduction waves
at À0.935, À1.337, À1.832 V for 2a and À0.937, À1.303,
À1.818 V for 2b, and two additional reduction waves at À1.701,
À2.076 V for 2a and À1.685, À2.066 V for 2b which were assigned
to the reduction of porphyrin moiety by comparing with the refer-
ence compound 6. The similar reduction potentials between aza-
fulleroids and aziridinofullerenes can be rationalized by Wuld’s
result, the reduction potentials of functionalized C60 had been as-
cribed to come from a combinational effect of the nature of the
bridge nitrogen and the binding mode of fullerene.6f
Interestingly, the oxidation potentials of the isomers show a
dramatic change between [5,6]-open and [6,6]-closed forms. The
one-electron oxidation potential of 1a is observed at À0.056 V,
64 mV negative shift than that of 1b at 0.008 V. Same tendency
was also found in dyads 2a and 2b, and the oxidation potential
of 2a is evaluated to be 50 mV cathodic shift than that of 2b. The
electrochemical test reveals that chromophores attached to [5,6]-
open azafulleroids are more easily oxidizable.
Besides aziridinofullerene 1b, a trace amount of compound 1a
with a less polarity was also obtained. It has the same molecular
ion signal [M]+ at 919.009 as 1b. Comparing to 1b, there is no ali-
phatic carbon signal observed, and a total of 32 signals of fullerene
appear at 148–124 ppm (Fig. S9, 1a), indicating that product 1a has
a Cs-symmetry and should be a [5,6]-open azafulleroid. The [5,6]-
open product may be generated through a direct addition of singlet
nitrene to [5,6]-ring junctions proposed by the previous theoretical
calculations.10,15
As shown in Scheme 2, the porphyrin-fullerene dyads 2a and 2b
were obtained by the reaction of azidophenylporphyrin 4 and C60 in
chlorobenzene at 130 °C for 9 h. The isomers 2a and 2b were both
separated in 12% yield. The MALDI mass spectrometry gave
[M+H]+ peak of 1685.6266 for 2a and 1685.6237 for 2b. The 13C
NMR spectra reveal that 2a and 2b have a different C60-symmetry,
Cs-symmetry for 2a and C2v-symmetry for 2b. A sharp peak at
82.9 ppm found in the aliphatic region of 2b demonstrates that com-
pound 2b stands as a [6,6]-closed form. Unlike the [2+1] nitrene
addition of Fc–N3, Por–N3 might prefer a 1,3-dipolar cycloaddition
of azide to the 6,6 double bond of the fullerene to form the interme-
diate triazoline, subsequently thermal cleavage of N2 to afford the
[5,6]-open azafulleroid and [6,6]-closed aziridinofullerene.10
The oxidation potential of Fc and Por is strongly sensitive to the
adjacent substituent, and has been broadly used as a detection
NH N
N
NH N
N3
N HN
N HN
2a
ClPh
4
reflux
+
NH
N
N
N HN
2b
Scheme 2. Preparation of N-porphyrin [5,6]-open azafulleroid 2a and [6,6]-closed aziridinofullerene 2b.

C. Chen et al. / Tetrahedron Letters 54 (2013) 1607–1611
1609
Table 1
Absorption data for compounds 1a, 1b, and 2a, 2b
Compound
Absorption (nm)
1a
1b
2a
2b
334
332, 427
334, 423, 515, 554, 594, 650
327, 422, 515, 550, 593, 649
NH
N
N
N
NH2
Fe
HN
5
6
Scheme 3. Reference compounds 5 and 6 for electrochemical test.
the different degree of the bridge nitrogen involved in the fullerene
delocalization, a 64 mV difference of the oxidation potential be-
tween [5,6]-open azafulleroid 1a and [6,6]-closed aziridinofulle-
rene 1b was observed. Guldi et al. accounted that the electron
pair on the nitrogen can engage in the delocalized
fullerene, as a consequence the [5,6]-open azafulleroid comprises
a quasi 62 -electron and the [6,6]-closed aziridinofullerene has
a 60
-electron system,11 while Wudl and co-workers considered
p-electrons of
p
p
that the nature of the bridge heteroatom may not engage with
the n-type electronic properties of C60, and [5,6]-open isomer has
60
bridge nitrogen atom is engaged with the delocalized
p
-electron nature.6f We thought the lone-pair electron of the
p
-electrons
of fullerene in both a and b isomers, and the electron donating abil-
ity to the attached chromophore of the bridge nitrogen is signifi-
cantly related with how much it participates in the delocalization
of fullerene
p-electrons. A lesser degree involvement of the lone-
pair electron on the bridge nitrogen in [5,6]-open azafulleroid
delocalization results in retaining the electronic nature of nitrogen
atom, and makes Fc attached to [5,6]-open iminofullerene more
readily oxidizable.
Figure 1. Cyclic voltammograms of (a) 6, (b) 2a, and (c) 2b.
In the dyads 2a and 2b, a more than 140 mV anodic shift of the
porphyrin oxidation potential compared to reference 6 has also
been obtained, and the porphyrin unit attached to the [5,6]-open
isomer shows a 50 mV negative shift than that of [6,6]-closed azir-
idinofullerene because of the same reason mentioned above.
To gain a deep insight into the nature of the isomers, quantum
chemistry calculations were done with GAUSSIAN 03 (Page S16). The
optimized geometries were shown in Figures 2 and 3 and the se-
lected calculated data are listed in Table 3. In the case of the
[5,6]-open isomer, the so-called [5,6]/5 and the [5,6]/6 structures
exist. In the [5,6]/5 structure, the chromophore tends to face the
pentagonal ring, whereas in the [5,6]/6 structure the chromophore
faces the hexagonal ring.10 Two possible structures of [5,6]-open
window. In our case, the oxidation potentials of 1a and 1b were ob-
served to have a more than 250 mV anodic shift compared to the
reference compound 5, which cannot be merely ascribed to the
electron withdrawing property of the fullerene sphere. Comparing
with the only 50 mV anodic shift of the oxidation potential which
has been found by Maurizio Prato when Fc was introduced directly
through a carbon atom to N-methyl fulleropyrrolidine,18 more
remarkable change of the oxidation potential observed in our case
should come from a combinational effect of the iminofullerene
structure and the weaker electron donating ability of the bridge
nitrogen. What’s more, the changed electron donating ability of
the bridge nitrogen may contribute to a crucial effect. Because of
Table 2
Redox potentials of 1, 2 and reference compounds 5, 6
E1ox
E2ox
E1red
E2red
E3red
E4red
E5red
E1ox À E1red
Compound
1a
À0.056
0.008
À0.314
0.485
0.535
0.345
À0.931
À0.932
À1.355
À1.368
À1.805
À1.805
0.875 eV
0.940 eV
1b
516
2a
0.742
0.799
0.450
À0.935
À0.937
À1.337
À1.303
À1.701
À1.685
À1.678
À1.832
À1.818
À2.076
À2.066
À2.044
1.420 eV
1.472 eV
2b
617
Experimental conditions: values for 0.5 (Epa + Epc) in V versus Fc/Fc+ in benzonitrile with Bu4NPF6 (0.1 M) as the supporting electrolyte and Ag/Ag+ electrode as the reference
electrode.

1610
C. Chen et al. / Tetrahedron Letters 54 (2013) 1607–1611
Figure 2. Optimized geometry of [5,6]/5-1a, [5,6]/6-1a, and 1b.
Figure 3. Optimized geometry of [5,6]/5-2a, [5,6]/6-2a, and 2b.
isomer show nearly the same DC–C (distance between the center of
Fc/Por and center of fullerene) with the [6,6]-closed isomer, and
this indicates the electron withdrawing effect of fullerene in the
[5,6]-open and [6,6]-closed isomer should have little impact on
the oxidation ability of chromophores (Fc or Por).
The HOMO energy levels of 1a and 2a are calculated to be high-
er than those of 1b and 2b, and this supports that [5,6]-open 1a
and 2a are more easily oxidizable than 1b and 2b, which is in
agreement with the electrochemical results. Meanwhile, the
charge density at the bridge N-atom in 1a and 2a is evaluated to
be larger than that in 1b and 2b. The different electron density of
the bridged N-atom in the isomers should come from the different
delocalization extent of the nitrogen lone-electron pair to fuller-
ene. The greater electron density around the bridged N in [5,6]-
open isomers than in [6,6]-closed isomers leads to the larger elec-
tron-donating ability of the bridged N-atom in [5,6]-open isomers.
In summary, electronic active compounds ferrocene and por-
phyrin were successfully introduced to the fullerene in order to
estimate the different influence of the iminofullerene isomers on
the attached chromophores. Results revealed that chromophores
connected to [5,6]-open isomer are more easily oxidizable than
those attached to [6,6]-closed isomer. This tendency was also sup-
ported by DFT calculation. We suppose that the engagement be-
tween nitrogen lone-electron pair and the delocalized
p-system
of fullerene should exist in both [5,6]-open and [6,6]-closed forms,
but the extent of engagement should be larger in [6,6]-closed iso-
mer. This work suggests that a perturbation of the
p-electrons of
Table 3
Center to center distances (DC–C) (Å), HOMO energy levels, HOMO/LUMO gaps (eV),
and NBO charge (e) of the isomers
the fullerene can exactly affect electronic properties of N-substi-
tuted aryl groups through the bridge nitrogen atom, and will make
fullerene-based materials an extensive application in functional
materials.19
Isomers
DC–C HOMO energy level HOMO/LUMO gaps NBO charge
[5,6]/6-1a
[5,6]/5-1a
1b
7.5 À5.050
7.4 À5.038
7.6 À5.250
1.812
1.786
2.031
1.624
1.586
1.672
À0.469
À0.486
À0.418
À0.459
À0.476
À0.418
Acknowledgment
[5,6]/6-2a 13.5 À4.898
[5,6]/5-2a 13.7 À4.873
We thank the 973 Program (2011CB932902) and NSFC (Nos.
21172126 and 21272123) for their generous financial support.
2b
13.3 À4.926

C. Chen et al. / Tetrahedron Letters 54 (2013) 1607–1611
1611
Schuster, D. I. J. Org. Chem. 2001, 66, 5449–5455; (f) Yang, C. D.; Cho, S.; Heeger,
A. J.; Wudl, F. Angew. Chem., Int. Ed. 2009, 48, 1592–1595; (g) Hachiya, H.;
Kakuta, T.; Takami, M.; Kabe, Y. J. Organomet. Chem. 2009, 694, 630–636; (h)
Wu, R.; Lu, X. Y.; Zhang, Y.; Zhang, J. M.; Xiong, W. T.; Zhu, S. Z. Tetrahedron
2008, 64, 10694–10698.
Supplementary data
Supplementary data associated with this article can be found, in
the
online
version,
at
http://dx.doi.org/10.1016/j.tet-
7. (a) Ouchi, A.; Hatsuda, R.; Awen, B. Z. S.; Sakuragi, M.; Ogura, R.; Ishii, T.; Araki,
Y.; Ito, O. J. Am. Chem. Soc. 2002, 124, 13364–13365; (b) Ouchi, A.; Awen, B. Z. S.;
Hatsuda, R.; Ogura, R.; Ishii, T.; Araki, Y.; Ito, O. J. Phys. Chem. A 2004, 108, 9584–
9592; (c) Ouchi, A.; Awen, B. Z. S.; Luo, H. X.; Araki, Y.; Ito, O. Tetrahedron Lett.
2005, 46, 6713–6716.
let.2013.01.061. These data include MOL files and InChiKeys of
the most important compounds described in this article.
References and notes
8. Ulmer, L.; Mattay, J. Eur. J. Org. Chem. 2003, 2933–2940.
9. Yan, M. D.; Cai, S. X.; Keana, J. F. W. J. Org. Chem. 1994, 59, 5951–5954.
10. Cases, M.; Duran, M.; Mestres, J.; Martín, N.; Solà, M. J. Org. Chem. 2001, 66,
433–442.
11. Guldi, D. M.; Hungerbühler, H.; Carmichael, I.; Asmus, K. D.; Maggini, M. J. Phys.
Chem. A 2000, 104, 8601–8608.
12. (a) Ikeda, A.; Fukuhara, C.; Kawaguchi, M.; Numata, M.; Shinkai, S.; Liu, S. G.;
Echegoyen, L. J. Chem. Soc., Perkin Trans. 2 2000, 307–310; (b) Ikeda, A.;
Fukuhara, C.; Shinkai, S. Chem. Lett. 1998, 415–416; (c) Yamazaki, M.; Fujitsuka,
M.; Ito, O.; Ikeda, A.; Fukuhara, C.; Kawaguchi, M.; Shinkai, S. J. J. Photochem.
Photobiol., A Chem. 2001, 140, 139–146.
13. (a) Otón, F.; Tárraga, A.; Espinosa, A.; Velasco, M. D.; Molina, P. J. Org. Chem.
2006, 71, 4590–4598; (b) Beer, P. D.; Cadman, J. Coord. Chem. Rev. 2000, 205,
131–155; (c) LeCours, S. M.; DiMagno, S. G.; Therien, M. J. J. Am. Chem. Soc.
1996, 118, 11854–11864; (d) Cao, Q. Y.; Pradhan, T.; Lee, M. H.; Choi, D. H.;
Kim, J. S. Tetrahedron Lett. 2012, 53, 4917–4920.
1. (a) Prato, M. Top. Curr. Chem. 1999, 199, 173–187; (b) Guldi, D. M. Chem.
Commun. 2000, 321–327; (c) Gust, D.; Moore, T. A.; Moore, L. A. Acc. Chem. Res.
2001, 34, 40–48; (d) Echegoyen, L.; Echegoyen, L. E. Acc. Chem. Res. 1998, 31,
593–601; (e) Thompson, B. C.; Fréchet, J. M. J. Angew. Chem., Int. Ed. 2008, 47,
58–77; (f) Martín, N.; Sánchez, L.; Illescas, B.; Pérez, I. Chem. Rev. 1998, 98,
2527–2547; (g) Nakamura, E.; Isobe, H. Acc. Chem. Res. 2003, 36, 807–815; (h)
Guldi, D. M.; Prato, M. Acc. Chem. Res. 2000, 33, 695–703; (i) Matsuo, Y.; Zhang,
Y.; Soga, I.; Sato, Y.; Nakamura, E. Tetrahedron Lett. 2011, 52, 2240–2242.
2. (a) Nakahodo, T.; Okada, M.; Morita, H.; Yoshimura, T.; Ishitsuka, M. O.;
Tsuchiya, T.; Maeda, Y.; Fujihara, H.; Akasaka, T.; Gao, X. F.; Nagase, S. Angew.
Chem., Int. Ed. 2008, 47, 1298–1300; (b) Tsuruoka, R.; Nagamachi, T.; Murakami,
Y.; Komatsu, M.; Minakata, S. J. Org. Chem. 2009, 74, 1691–1697; (c) Lafleur-
Lambert, A.; Rondeau-Gagné, S.; Soldera, A.; Morin, J. Tetrahedron Lett. 2011, 52,
5008–5011; (d) Rotas, G.; Tagmatarchis, N. Tetrahedron Lett. 2009, 50, 398–401.
3. (a) Hirsch, A. Top. Curr. Chem. 1999, 199, 1–65; (b) López, A. M.; Mateo-Alonso,
A.; Prato, M. J. Mater. Chem. 2011, 21, 1305–1318.
4. (a) Prate, M.; Li, Q. C.; Wudl, F. J. Am. Chem. Soc. 1993, 115, 1148–1150; (b) Liu,
T. X.; Wei, T.; Zhu, S. E.; Wang, G. W.; Jiao, M. Z.; Yang, S. F.; Bowles, F. L.;
Olmstead, M. M.; Balch, A. L. J. Am. Chem. Soc. 2012, 134, 11956–11959; (c)
Strom, T. A.; Barron, A. R. Chem. Commun. 2010, 46, 4764–4766.
5. Rostovtsev, V. V.; Green, L. G.; Fokin, V. V.; Sharpless, K. B. Angew. Chem., Int. Ed.
2002, 41, 2596–2599.
6. (a) Yashiro, A.; Nishida, Y.; Ohno, M.; Eguchi, S.; Kobayashi, K. Tetrahedron Lett.
1998, 39, 9031–9034; (b) González, S.; Martín, N.; Swartz, A.; Guldi, D. M. Org.
Lett. 2003, 5, 557–560; (c) Xiao, S. Q.; Li, Y. L.; Fang, H. J.; Li, H. M.; Liu, H. B.; Shi,
Z. Q.; Jiang, L.; Zhu, D. B. Org. Lett. 2002, 4, 3063–3066; (d) Guldi, D. M.;
González, S.; Martín, N.; Antón, A.; Garín, J.; Orduna, J. J. Org. Chem. 2000, 65,
1978–1983; (e) MacMahon, S.; Fong, R., II; Baran, P. S.; Safonov, I.; Wilson, S. R.;
14. Shafir, A.; Power, M. P.; Whitener, G. D.; Arnold, J. Organometallics 2000, 19,
3978–3982.
15. Cases, M.; Duran, M.; Solà, M. J. Mol. Model. 2000, 6, 205–212.
16. According to the literature, aminoferrocene has nearly the same redox
potential with dimethylaminoferrocene. See: Britton, W. E.; Kashyap, R.; El-
Hashash, M.; El-Kady, M. Organometallics 1986, 5, 1029–1031.
17. N,N-Dimethylbenznylporphyrin 6 gave the third oxidation potential at 1.151 V,
which was considered to be the oxidation of N,N-dimethyl group. See:
Macdonald, T. L.; Gutheim, W. G.; Martin, R. B.; Guengerich, F. P. Biochemistry
1989, 28, 2071–2077.
18. Prato, M.; Maggini, M.; Giacometti, C.; Scorrano, G.; Sandonh, G.; Farnia, G.
Tetrahedron 1996, 52, 5221–5234.
19. Fukuda, T.; Masuda, S.; Kobayashi, N. J. Am. Chem. Soc. 2007, 129, 5472–5479.