Concerted two-electron transfer and high selectivity of TiO2 in photocatalyzed deoxygenation of epoxides

Article abstract of DOI:10.1002/anie.201307374

No two ways about it: In the photocatalytic deoxygenation of epoxides, the TiO₂ particle concertedly transfers two stored electrons to generate a carbanion intermediate, which dissociates to the alkene product. This pathway ensures the higher alkene and stereoselectivity of the photocatalytic deoxygenation than those involving a single-electron transfer. Copyright

Full text of DOI:10.1002/anie.201307374

.
Angewandte
Communications
DOI: 10.1002/anie.201307374
Photochemistry
Concerted Two-Electron Transfer and High Selectivity of TiO2 in
Photocatalyzed Deoxygenation of Epoxides**
Yue Li, Hongwei Ji, Chuncheng Chen,* Wanhong Ma, and Jincai Zhao
TiO2 photocatalysis has shown promising potential in the
degradation of pollutants, fixation of CO2, and splitting of
H2O. In recent years, the selective conversions of organic
compounds by TiO2 photocatalysis have been explored
actively.[1] Even though most of the photocatalytic reactions
involve the total gain or loss of multiple equivalents of
electrons, and the term multielectron reaction is frequently
used in documents on TiO2 photocatalysis, these redox
processes are conventionally thought to proceed through
sequential single-electron-transfer (SET) pathways.[2] For
example, the photocatalytic oxidation of an alcohol to an
aldehyde (or ketone) is a typical two-electron oxidation
process through a SET mechanism.[2c] In this process, the
alcohol first reacts with a valence band hole (hvb+) to generate
a radical intermediate, which injects another electron into the
conduction band to complete the two-electron reaction (the
current amplification effect). According to the SET mecha-
nisms radical intermediates are inevitably involved and is
extremely unfavorable for the selective transformation of
organic substrates.
chiometric reductants.[5] For the reduction of epoxides with
E–Z isomerism, metal complexes which can offer single
electron per molecule, such as [Cp2TiCl] (Cp = cyclopenta-
dienyl), give the alkene products with no stereoretention (that
is, the yield distribution of stereoisomeric products is inde-
pendent of substrate configuration) and low overall selectivity
(see Table S1 in the Supporting Information for a literature
survey), because of the formation of labile radical intermedi-
ates (Scheme 1a).[5a,b] In contrast, for W and Mo complexes,
The concerted multielectron-transfer (CMET) pathway,
in which multiple (usually two) electrons are transferred from
or to an adsorbed species in a single kinetic step, could avoid
the formation and side reactions of radical intermediates. As
a result of its high state density, TiO2 particles can store
multiple conduction band electrons (ecbÀ) and hvb+, and is
a character of photocatalysis which distinguishes it from
molecular redox systems. Principally, the storage of multiple
Scheme 1. Sequential SET mechanism (a) and O-atom-transfer mecha-
nism (b) for the reductive deoxygenation of an epoxide.
predominant retention of substrate configuration was
observed in their deoxygenation systems, and was explained
by either the absence of a radical intermediate or by an O-
atom-transfer mechanism (Scheme 1b).[5c,d] Thus, the stereo-
chemistry of alkene products offers a great deal of informa-
tion on the reduction pathway.[6]
Herein, we show our compelling experimental evidence
that CMET is the dominant pathway in the photocatalytic
deoxygenation of epoxides on pristine TiO2 surface. The
CMET mechanism ensures the much higher alkene selectivity
and higher stereoretention of this transformation. This study
provides a clear example of the CMET characteristic of TiO2
photocatalysis, as well as a cheap and green method for the
selective deoxygenation of epoxides.
+
ecbÀ/hvb paves the way for CMET. However, the occurrence
of CMET has rarely been experimentally verified in the
stoichiometric multielectron photocatalytic reactions of pris-
tine TiO2 systems.
The reduction of an epoxide to the corresponding alkene
is an important reaction in organic synthesis[3] and biological
chemistry.[4] Traditionally, the deoxygenation of epoxides is
realized using the complexes of low-valent metals as stoi-
[*] Dr. Y. Li, Prof. H. Ji, Prof. C. Chen, Prof. W. Ma, Prof. J. Zhao
Key Laboratory of Photochemistry
Beijing National Laboratory for Molecular Sciences
Institute of Chemistry, Chinese Academy of Sciences
Beijing 100190 (China)
The photocatalytic reactions were carried out in isopropyl
alcohol (iPrOH), which also serves as a reductive sacrifice to
+ [2c]
scavenge hvb
.
Styrene oxide was consumed almost com-
E-mail: ccchen@iccas.ac.cn
pletely in 15 hours under UV irradiation, and nearly quanti-
tatively deoxygenated to styrene (Table 1, entry 1). Control
experiments showed that both TiO2 and UV irradiation are
necessary for the deoxygenation (see Table S2). Moreover,
when 1-phenyl-1,2-ethanediol or 2-methoxy-2-phenylethanol
was used as substrate, no alkene product was obtained, thus
indicating that photocatalytic deoxygenation does not pro-
ceed via hydrolysis or alcoholysis intermediates. High alkene
Dr. Y. Li
Department of Chemistry, Nankai University
Tianjin 100071 (China)
[**] Financial support from THE 973 project (Nos. 2010CB933503,
2013CB632405), NSFC (Nos. 21137004, 21273245, and 21277147),
and the CAS are gratefully acknowledged.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201307374.
12636
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 12636 –12640

Angewandte
Chemie
Table 1: Comparison of the alkene selectivity and stereochemistry
nation was frequently observed for reductions system in
between TiO2 photocatalytic[a] and [Cp2TiCl][b] deoxygenation systems.
which the reductant can offer more than one electron.[5c,d] The
partial stereoretention of photocatalytic deoxygenation sug-
gests that, instead of separate SETs, this reaction proceeds by
a CMET pathway.
Entry Substrate Product
System
Conv. [%] Sel. [%] E/Z
1
2
TiO2/UV
[Cp2TiCl]
TiO2/UV[c]
97
100
97
57
–
–
The difference between TiO2 photocatalysis and TiIII
systems may be attributed to the ability of TiO2 particles to
3
4
61
100
ca. 100 48:52
37 98:2
[Cp2TiCl]
À
TiO2/UV[d]
[Cp2TiCl]
30
100
ca. 100 97:3
44 99:1
store and simultaneously provide multiple ecb
,
[7] and its high
5
6
À
state density. Therefore, we analyzed the accumulation of ecb
using an in situ UV/Vis-DRS method. Under UV irradiation,
the TiO2 suspension in iPrOH showed a broad absorption
band above l = 380 nm (see Figure S1), which is the typical
absorption of the ecbÀ of TiO2.[8] By trapping the accumulated
[a] 15 mg TiO2 (P25), 0.1 mmol epoxide, 5 mL iPrOH, Ar atmosphere,
100 W Hg lamp cutoff below l=300 nm, irradiation time: 15 h.
Conversion and selectivity determined by GC analysis. [b] 5 mL freshly
prepared [Cp2TiCl] solution of THF (3.4 mmol) was mixed with 5 mL
deaerated iPrOH solution containing 1 mmol epoxide at room temper-
ature, reaction time: 10 min. [c] To avoid the photoisomerization of
alkene products, light with l>350 nm was used, time: 35 h.
[d] l>350 nm, time: 120 h.
ecb with Fe3+ and quantifying the formed Fe2+, the total
À
amount of stored ecbÀ at a photostationary state was estimated
to be 123.8 mm. Assuming that TiO2 (P25) nanoparticles are
spherical with a diameter of 21 nm and their density is
À
4.0 gcmÀ3,[9] an average of 73 ecb are stored in each TiO2
particle (Figure 1a).
selectivity was also observed in the photocatalytic reduction
of (E)- and (Z)-stilbene oxides (Table 1, entries 3 and 5),
although the rates were apparently slower, probably a result
of steric effects. In contrast, the deoxygenation selectivities of
[Cp2TiCl] systems were rather poor (30–60%), even though
the reductive electron of [Cp2TiCl] is also distributed mainly
around the Ti atom.
More intriguingly, in the photocatalytic deoxygenation of
(E)-stilbene oxide, the ratio of the yield of E to Z alkene
(E/Z) was 97:3, whereas a mixture of stilbene stereoisomers
with an E/Z ratio of 48:52 was obtained when (Z)-stilbene
oxide was used as the reactant (Table 1), thus indicating that
a large fraction of the stereochemistry of the epoxide
substrate is retained in photocatalytic deoxygenation. In
control experiments, Z stilbene was dissolved in a TiO2
suspension in iPrOH. After 35 hours of irradiation, only 2%
of Z stilbene was converted into its E isomer. Therefore, the
photoisomerization of alkene products cannot influence
markedly the stereochemistry of products. In contrast, for
the reduction system of [Cp2TiCl], the E alkene was always
the main product (E/Z = 99:1 for (E)-stilbene oxide and 98:2
for its Z isomer), thus indicating that the stereochemistry of
alkene products is independent on the configuration of the
substrate.
It is well accepted that the deoxygenation of epoxides by
[Cp2TiCl], which can offer only one electron per complex,
proceeds by two consecutive SET processes (Scheme 1a).
The first electron transfer causes the opening of oxirane ring
and the formation of a b-metaloxy radical. This radical has
Figure 1. a) The accumulation and decay of ecbÀ in TiO2 particles in the
À
enough time to rotate around the C1 C2 bond to realize
presence and absence of styrene oxide. b) The comparison between
configuration relaxation before its reduction by the second
equivalent of the TiIII complex. The partial retention of
substrate stereochemistry in the photocatalytic deoxygena-
tion indicates that the long-lived, open-ring radical inter-
mediate is not involved in this reaction. To validate this
argument, ethyl acrylate and methyl methacrylate (radical
traps)[5b,6] were added to the photocatalytic system. No
product of radical capture was detected (see Table S3), thus
confirming the absence of radical intermediates. The partial
retention of substrate configuration during epoxide deoxyge-
À
the fitting of the ecb decay process Ia according to first- and second-
order kinetics, ce: mm.
In the presence of styrene oxide, the accumulated amount
of ecbÀ was significantly decreased because of its reaction with
the epoxide. With 20 and 80 mm styrene oxide, an average of
29 and 17 ecbÀ, respectively, were stored in one TiO2 particle at
the photostationary state. These results mean that during the
photocatalytic reduction of epoxides multiple ecbÀ are always
Angew. Chem. Int. Ed. 2013, 52, 12636 –12640
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 12637

.
Angewandte
Communications
Table 2: The effect of light intensity (I) on the reduction of (Z)-stilbene
available in the TiO2 particle. After the termination of
irradiation, the ecbÀ absorption band underwent a remarkable
decrease in the presence of styrene oxide (Figure 1a). The
decay of ecbÀ provides us an opportunity to study the electron-
transfer mechanism of photocatalytic deoxygenation. The
kinetic distinction between the sequential SET and CMET
pathways originates from the number of electrons transferred
in the rate-determining step. For the SET pathway, the overall
reaction is controlled by the first electron reduction of the
epoxide, so the decay of ecbÀ would exhibit first-order kinetics.
oxide.[a]
Entry
Light source
I [mWcmÀ2
]
t [h]
Conv. [%][b]
E/Z
1
2
3
Hg lamp
Xe lamp
Laser (355 nm)
0.166
3.95
250
35
4
1
61
30
21
48:52
49:51
48:52
[a] 3 gLÀ1 TiO2, 20 mm (Z)-stilbene oxide, 5 mL iPrOH, Ar atmosphere,
light with l>350 nm was used. [b] Conversion determined by GC
analysis.
À
However, in the CMET pathway, since two ecb are simulta-
À
neously transferred, the decay of ecb should obey second-
À
order kinetics. Therefore, through the kinetic analysis of ecb
intermediate if formed) would also be influenced, as proven
by the accelerated deoxygenation rate at enhanced light
intensity. If the loss of substrate stereochemistry were caused
decay, these two pathways could be distinguished (the
detailed kinetic derivation and more discussion are given in
the Supporting Information). As shown in Figure 1b, for
À
by the rotation of the C1 C2 bond of the radical intermediate
À
decay process Ia, the temporal change of ecb concentration
formed in the SET pathway, the enhancement of irradiation
intensity would exaggerate the accumulation of ecbÀ, and
consequently accelerate the transfer of the second ecbÀ. As
a result, the lifetime of the radical intermediate would be
shortened, and thus higher retention of the configuration
would be obtained at an enhanced intensity of irradiation, all
of which is in conflict with our experimental observations.
More reasonably, the independence of stereochemisty on
was much better described by second-order kinetics (1/ce ꢀ t)
2
with the correlation coefficient R2 of 0.997, rather than by
2
first-order kinetics (ln(ce) ꢀ t, R1 = 0.893). The fitting of the
À
other ecb decay processes in Figure 1a (Ib, IIa, and IIb) gave
2
2
similar results (R2 = 0.966–0.997 versus R1 = 0.812–0.905; see
Figure S2). In another experiment, the ecbÀ was first stored in
TiO2 particles by irradiating a TiO2 suspension in iPrOH
under an Ar atmosphere, and then the absorption at l =
800 nm was chronometrically recorded after the addition of
À
light intensity suggests that the transfer of the second ecb
À
occurs prior to the rotation of C1 C2 bond, and is consistent
with the CMET pathway evidenced by the fitting of ecb
2
À
deaerated styrene oxide. Again, second-order kinetics (R2 =
2
0.997–0.998) were superior to the first-order kinetics (R1
=
decay. According to this mechanism, the photocatalytic
deoxygenation is initiated by the two-electron reduction of
À
0.962–0.969) for the ecb decay (see Figure S3). The second-
order kinetic decay of ecb is an indication of the concerted
À
À
the substrate to a carbanion intermediate, which also has C1
C2 bond that can rotate (Scheme 2). The stereochemistry of
two-electron-transfer pathway. However, for the SET-based
photocatalytic reduction of benzaldehyde and acetophe-
alkene products is determined by the competition between
none,[10] first-order kinetics was more suitable for the ecb
C1 C2 bond rotation and C2 O bond cleavage of the formed
carbanion intermediate. Neither of these two processes
involves the redox of the intermediate, so ecbÀ-influencing
factors, such as light intensity, cannot have an effect on the
stereochemistry of the reaction, but can markedly affect the
overall deoxygenation rate. In addition, the E-configured
carbanion intermediate is thermodynamically more favorable
than the Z-configured carbanion. As a result, the loss of
substrate stereochemistry would not happen in the reduction
of (E)-stilbene oxide (Table 1, entry 5).
À
À
À
decay than second-order kinetics (see Figure S4).
We also noted the O-atom-transfer mechanism (Scheme
1b), in which the O atom of the epoxide is transferred directly
to the TiO2 surface in a single step,[5c] would also give the
À
À
second-order decay of ecb because two ecb are involved in
this process. However, according to this mechanism, the
configuration of the epoxide substrate should be predom-
inantly retained in alkene products, and is in conflict with the
partial stereoretention (52%) observed in the photocatalytic
reduction of (Z)-stilbene oxide. Therefore, we do not consider
O-atom-transfer as the dominant pathway.
The CMET character of epoxide deoxygenation is rem-
iniscent of a platinum cocatalyst, which is able to trap ecbÀ and
catalyze multielectron redox reactions on its surface.[7c,12]
However, in our experiments, the platinization of TiO2 was
found to significantly suppress the deoxygentation rate of (Z)-
stilbene oxide (kTiO2/kPt-TiO2 = 1:0.48, see Figure S5), while the
E/Z ratio of the formed stilbene isomers was nearly unaf-
fected (47:53). Obviously, the deposited platinum is inert to
the deoxygenation of epoxide. Otherwise, the overall reaction
rate would be accelerated and the stereochemistry would be
altered. The reaction should predominantly occur on pristine
TiO2 surface, and the supression is due to the competition of
platinum with surface Ti sites for ecbÀ. This experiment
implies that the surface Ti sites of TiO2 are essential for the
deoxygenation. The strong affinity of Ti4+ to oxygen[13] would
facilitate CMET by providing additional thermodynamic
Careful study of the effect of light intensity on the
stereochemistry of the photocatalytic deoxygenation of (Z)-
stilbene oxide provides another line of evidence for the
CMET pathway. When the reduction of (Z)-stilbene oxide
was carried out under a Hg lamp, stilbene products with the
E/Z ratio of 48:52 were obtained (Table 2, entry 1). When
using a stronger Xe lamp (3.95 mWcmÀ2) as the light source,
the stereochemistry of the products was nearly unchanged,
even though the overall deoxygenation was accelerated by 4.3
times (entry 2). Further enhancement of irradiation intensity
with a laser (250 mWcmÀ2) also had little effect on the isomer
distribution of alkene products (E/Z = 48:52, entry 3). Light
intensity is known to be able to greatly affect the accumu-
À
lation and reductive ability of ecb
,
[2b,11] and consequently, the
ecbÀ-transfer rate from TiO2 to the epoxide (or radical
drive with the formation of surface Ti O bond.
À
12638 www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 12636 –12640

Angewandte
Chemie
applicability of this reaction.
As shown in Table 3, similar
to styrene oxide, the alkene
yields of the para-halogen-
substituted derivatives of sty-
rene oxide were all greater
than 90% (entries 1–3). The
photocatalytic deoxygenation
of E-b-methylstyrene oxide,
a
1,2-disubstituted oxirane,
was also nearly quantitative.
However, the deoxygenation
of 1,1-disubstituted aromatic
oxiranes exhibited lower
selectivities (entries 5 and 6).
The poor selectivities could
be explained by the compet-
ing hydrolysis and alcoholysis
Scheme 2. Concerted two-electron-transfer pathway for the photocatalytic deoxygenation of epoxides.
Table 3: Photocatalytic deoxygenation of epoxides.[a]
Entry Substrate
Product
t [h] Conv. [%][b] Sel. [%][b]
1
15
15
15
100
100
100
99
96
91
2
3
4
60
80
15
68
46
72
ca. 100
65
Figure 2. Hammett plots of log (k/kH) versus the substituent constant
À
sÀ for the photocatalytic deoxygenation of para-substituted styrene
p
5
oxide.
6
21
7
8
25
25
8
0
–
The formation of a carbanion intermediate is supported
by the linear free energy relationship of photocatalytic
deoxygenation. It was observed that para substitution with
F, Cl, and Br all facilitated the photocatalytic deoxygenation
(the selectivities were all > 90%; see Table 3). Hammett plots
of the logarithm of relative rates (log(k/kH)) against corre-
<1
[a] Reaction conditions: 15 mg TiO2 (P25), 0.1 mmol epoxide, 5 mL
iPrOH, 100 W Hg lamp cutoff below l=300 nm, Ar atmosphere.
[b] Conversion and selectivity determined by GC analysis.
À
sponding sp substituent constants gave good correlation
(R2 = 0.998, Figure 2), while the correlation with sp was poor
(R2 = 0.791, see Figure S6), thus illustrating that an electro-
negative reaction center is formed and the aromatic ring
stabilizes it through conjugation.[14] The Hammett constant
reactions, since such substrates are highly acid sensitive.[5e]
This argument was supported by the GC-MS analysis of the
reaction solution. For example, the hydrolysis product, 2-
phenyl-1,2-propanediol, was the main byproduct in the
photocatalytic reduction of a-methylstyrene oxide (see Fig-
ure S7). Aliphatic epoxides were inert to photocatalytic
reduction (entries 7 and 8), and may result from the absence
of the stabilizing effect of the aromatic ring on the carbanion
intermediate.
In summary, we have achieved selective deoxygenation of
epoxides on TiO2. The high selectivity of this reaction is
attributed to a CMET pathway. Through this pathway, highly
reactive radical intermediates are avoided, and is beneficial to
À
(1) obtained was 0.66. The better correlation with sp and
a positive value of 1 imply that the reductive dissociation of
[15]
À
the C O bond is involved in the rate-determining step. In
our mechanism, the charge accumulation in the carbanion
À
intermediate would be weakened by the C1 O bond cleav-
À
age, and the formation of surface Ti O bond also has an effect
of charge compensation, and makes the CMET pathway more
feasible. These effects make the 1 value for photocatalytic
deoxygenation a relatively small absolute value.
Inspired by the CMET nature of photocatalytic deoxyge-
nation, we extended the scope of substrates to test the
Angew. Chem. Int. Ed. 2013, 52, 12636 –12640
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 12639

.
Angewandte
Communications
874; d) K. G. Moloy, Inorg. Chem. 1988, 27, 677 – 681; e) J. E.
McMurry, M. G. Silvestri, M. P. Fleming, T. Hoz, M. W. Grays-
ton, J. Org. Chem. 1978, 43, 3249 – 3255.
the selectivity of organic transformation. Our work also
suggests that CMET might also proceed in other photo-
catalytic reactions, such as the reduction of CO2 and O2, as
a result of the high state density of the semiconductor.
[6] T. Mitsudome, A. Noujima, Y. Mikami, T. Mizugaki, K.
Jitsukawa, K. Kaneda, Angew. Chem. 2010, 122, 5677 – 5680;
Angew. Chem. Int. Ed. 2010, 49, 5545 – 5548.
[7] a) J. Tang, J. R. Durrant, D. R. Klug, J. Am. Chem. Soc. 2008,
130, 13885 – 13891; b) N. M. Dimitrijevic, B. K. Vijayan, O. G.
Poluektov, T. Rajh, K. A. Gray, H. He, P. Zapol, J. Am. Chem.
Soc. 2011, 133, 3964 – 3971; c) C. M. Sꢀnchez-Sꢀnchez, A. J.
Bard, Anal. Chem. 2009, 81, 8094 – 8100.
[8] T. Yoshihara, R. Katoh, A. Furube, Y. Tamaki, M. Murai, K.
Hara, S. Murata, H. Arakawa, M. Tachiya, J. Phys. Chem. B
2004, 108, 3817 – 3823.
[9] M. R. Hoffmann, S. T. Martin, W. Choi, D. W. Bahnemann,
Chem. Rev. 1995, 95, 69 – 96.
[10] a) C. Joyce-Pruden, J. K. Pross, Y. Li, J. Org. Chem. 1992, 57,
5087 – 5091; b) S. Kohtani, E. Yoshioka, K. Saito, A. Kudo, H.
Miyabe, J. Phys. Chem. C 2012, 116, 17705 – 17713.
[11] J. N. Schrauben, R. Hayoun, C. N. Valdez, M. Braten, L. Fridley,
J. M. Mayer, Science 2012, 336, 1298 – 1301.
[12] a) R. Abe, H. Takami, N. Murakami, B. Ohtani, J. Am. Chem.
Soc. 2008, 130, 7780 – 7781; b) W. Zhao, C. C. Chen, X. Z. Li,
J. C. Zhao, H. Hidaka, N. Serpone, J. Phys. Chem. B 2002, 106,
5022 – 5028.
[13] P. M. Burkinshaw, C. T. Mortimer, Coord. Chem. Rev. 1983, 48,
101 – 155.
[14] C. Hansch, A. Leo, R. W. Taft, Chem. Rev. 1991, 91, 165 – 195.
[15] a) D. D. Tanner, J. A. Plambeck, D. W. Reed, T. W. Mojelsky, J.
Org. Chem. 1980, 45, 5177 – 5183; b) E. V. Blackburn, D. D.
Tanner, J. Am. Chem. Soc. 1980, 102, 692 – 697.
Received: August 21, 2013
Published online: October 7, 2013
Keywords: alkenes · electron transfer · photochemistry ·
.
small ring systems · titanium
[1] a) M. Zhang, C. Chen, W. Ma, J. Zhao, Angew. Chem. 2008, 120,
9876 – 9879; Angew. Chem. Int. Ed. 2008, 47, 9730 – 9733; b) G.
Palmisano, E. Garcia-Lopez, G. Marci, V. Loddo, S. Yurdakal, V.
Augugliaro, L. Palmisano, Chem. Commun. 2010, 46, 7074 –
7089; c) X. Lang, H. Ji, C. Chen, W. Ma, J. Zhao, Angew.
Chem. 2011, 123, 4020 – 4023; Angew. Chem. Int. Ed. 2011, 50,
3934 – 3937; d) H. Kisch, Angew. Chem. 2013, 125, 842 – 879;
Angew. Chem. Int. Ed. 2013, 52, 812 – 847.
[2] a) S. O. Obare, T. Ito, G. J. Meyer, J. Am. Chem. Soc. 2006, 128,
712 – 713; b) W. Choi, M. R. Hoffmann, J. Phys. Chem. 1996, 100,
2161 – 2169; c) M. Miyake, H. Yoneyama, H. Tamura, Chem.
Lett. 1976, 635 – 640.
[3] W. S. Johnson, M. S. Plummer, S. P. Reddy, W. R. Bartlett, J. Am.
Chem. Soc. 1993, 115, 515 – 521.
[4] R. Teufel, T. Friedrich, G. Fuchs, Nature 2012, 483, 359 – 362.
[5] a) T. V. RajanBabu, W. A. Nugent, M. S. Beattie, J. Am. Chem.
Soc. 1990, 112, 6408 – 6409; b) T. V. RajanBabu, W. A. Nugent, J.
Am. Chem. Soc. 1994, 116, 986 – 997; c) L. M. Atagi, D. E. Over,
D. R. McAlister, J. M. Mayer, J. Am. Chem. Soc. 1991, 113, 870 –
12640 www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 12636 –12640