Control of conformer population and product selectivity and stereoselectivity in competitive photocyclization/rearrangement of chiral donor-acceptor dyad

Article abstract of DOI:10.1021/ja302768w

The conformer population and the relative excitation efficiency of each conformer, assessed by theoretical and experimental examinations, enabled us to precisely control the chemo- and diastereoselectivities of the competitive photocyclization/rearrangeme

Full text of DOI:10.1021/ja302768w

Communication
pubs.acs.org/JACS
Control of Conformer Population and Product Selectivity and
Stereoselectivity in Competitive Photocyclization/Rearrangement of
Chiral Donor−Acceptor Dyad
Emi Nishiuchi, Tadashi Mori,* and Yoshihisa Inoue*
Department of Applied Chemistry, Osaka University, 2-1 Yamada-oka, Suita 565-0871, Japan
S
* Supporting Information
Scheme 1. Diastereodifferentiating Competitive
ABSTRACT: The conformer population and the relative
excitation efficiency of each conformer, assessed by
theoretical and experimental examinations, enabled us to
precisely control the chemo- and diastereoselectivities of
the competitive photocyclization/rearrangement reaction
of a chiral donor−acceptor (D/A) dyad by irradiation
wavelength, solvent polarity, and reaction temperature.
Manipulating the ground- and excited-state conformer
equilibria is essential for critically controlling the intra-
molecular D/A system, in sharp contrast to the rather
divergent excited species involved in relevant intermolec-
ular systems.
Photocyclization/Rearrangement of 1 to 2 and 3 upon
a
Direct and CT-Band Excitations
a
The stereochemistry of 2 was determined by differential NOE
spectra. See the Supporting Information for details.
analogues.4,5 The diastereomers of 2 were separated by GC
(see Figure S1 in the Supporting Information), and the
stereochemistry was determined by differential NOE (Scheme
1 and Figures S9 and S10). The photorearrangement of 1 to 3
in cyclohexane is known to proceed with only ∼85%
stereoretention, presumably via the aromatic π−π* state.6
The use of an intramolecular dyad facilitated the ground-state
CT interactions, as evidenced by the CT band observed at
longer wavelengths (Figure S2), and permits direct UV−vis and
CD spectroscopic inspection of the ground-state complex. The
cyclization occurs primarily from the energetically favored
gauche conformers (G+ and G−) rather than trans conformer
(T). The two gauche conformers are diastereomeric to each
other, and the re- or si-face of the dicyanoethylene moiety faces,
and is prone to react with, the phenyl ring (Scheme 1).
For a more rational and universal methodology for accurately
controlling the chemo- and diastereoselectivities, we employed
a combined experimental and theoretical approach to elucidate
the conformer geometry, stability, population, and chiroptical
properties relevant to this competitive chiral photocyclodime-
rization/rearrangement reaction. The calculations predict the
relative extinction coefficient of each conformer at a given
wavelength, enabling us to selectively excite the desired
conformer and also to precisely control the product selectivities
and diastereoselectivities in combination with the external
variants, such as solvent and temperature. Indeed, the shift of
equilibrium by varying solvent and temperature, monitored by
CD spectroscopy, can nicely rationalize the photochemical
recisely controlling stereochemistry is still a challenge in
P
chiral photochemistry, despite the recent endeavors
devoted to a variety of molecular, supramolecular, and
biomolecular photochirogenic systems.1 The encounter of
chiral reactant or sensitizer with substrate and the subsequent
chirality transfer should be accomplished within the lifetime of
the excited species involved (∼ns). Entropy-related environ-
mental variants, such as temperature, solvent polarity, and
pressure, are known to function as effective tools for
manipulating the molecular geometry and recognition mode
in the excited state and eventually the stereochemical outcomes
in chiral photoreactions.2 Another inherent advantage unique
to photochemistry is the choice of excitation wavelength,
allowing us to tune the stereochemical consequences beyond
the fetters of ground-state thermodynamics. Indeed, selective
excitation at the charge-transfer (CT) band affords regio- and
diastereoselectivities that significantly differ from those
obtained upon local-band excitation of the donor in the
[2+2] photocycloaddition of stilbene with chiral fumarates and
also in the Paterno−Buchi reaction of 1,1-diphenylethene with
́
̈
chiral p-cyanobenzoates.3 The successful regio-/diastereocon-
trol by wavelength in these intermolecular donor−acceptor (D/
A) systems prompted us to explore the general validity and
elucidate the mechanistic details of wavelength-dependent
photoreactions.
In this study, we prepared a chiral D/A dyad, (S)-1,1-
dicyano-2-methyl-4-phenyl-1-pentene (1), as a tethered sub-
strate for enhanced CT interactions and investigated the
competitive photocyclization/rearrangement behavior (Scheme
1) under a variety of irradiation conditions. Photoirradiation of
1 afforded diastereomeric cyclization product 2 and rearrange-
ment product 3, as was the case with the related achiral
Received: March 21, 2012
Published: April 30, 2012
© 2012 American Chemical Society
8082
dx.doi.org/10.1021/ja302768w | J. Am. Chem. Soc. 2012, 134, 8082−8085

Journal of the American Chemical Society
Communication
behavior of 1. In the intermolecular D/A systems reported
previously, the differences in structure and/or solvation
between the excited CT complex and the conventional exciplex
are the source of the wavelength-dependent diastereoselectivity
observed. In the intramolecular dyad system employed here,
the number of possible conformations is greatly reduced to
only a pro-R/S pair of T, G+, and G− conformers, the properties
and population of which are critical functions of wavelength,
solvent, and temperature to determine the product selectivities
and diastereoselectivities through the conformational equilibria
in the ground and excited states.
Photoirradiations of 1 were run at λ = 260−300 nm through
band-pass filters (fwhm 10 nm) in acetonitrile at 25 °C to give
2 and 3 without any byproduct (Table S1). The 2/3 ratio
gradually increased with increasing irradiation λ, while the
diastereomeric excess (de) of 2 was practically constant at
∼30% with 260−270 nm irradiations (local band excitation),
but then leaped to 56−59% with 280−300 nm irradiations (CT
band excitation). These results permitted us to choose 254 nm
for direct excitation and 280 and 300 nm for CT excitation;
note that the direct excitation is not completely direct, but the
direct/CT excitation ratio is estimated as ∼1.2 at 254 nm from
the absorbance of the CT complex (deconvoluted εCT) relative
to that of a 1:1 mixture of cumene and 1,1-dicyano-2-methyl-1-
propene (estimated εdirect), while the CT excitation was nearly
selective btoh at 280 and 300 nm. Although 1 did not fluoresce
in solution, the presence of oxygen did not affect the yield or
product distribution (Table S2), suggesting that the contribu-
tion of triplet manifold and radical intermediate7 is minor and
the photoreaction proceeds mostly through the singlet
manifold. The reaction was generally clean, and thus the
product ratio was constant throughout the reaction (Table S3).
Table 1 summarizes the results of the photoreaction of (3S)-
1 under a variety of irradiation conditions (see also Table S4).
rearrangement is sigmatropic in nature and hence less
polarized.4,6 In acetonitrile, the 2/3 ratio increased upon
excitation at longer wavelengths and at lower temperatures
(excepting the data obtained at −40 °C upon 280 nm
irradiation). This can be explained by the equilibrium shift from
extended trans to folded gauche with CT character in ground
and/or excited states, facilitating cyclization to 2. The de of 2
obtained in acetonitrile was also temperature-dependent, but
the changing pattern was not very smooth. Upon CT excitation,
(1R,3S)-2 was consistently favored, and the de reached a
maximum at 0 to −20 °C, while (1S,3S)-2 was obtained as the
major product only upon direct excitation in acetonitrile. This
means that the excited CT complex formed upon excitation at
280 or 300 nm affords 2, which is however epimeric to the one
produced from the conventional exciplex generated upon direct
excitation at 254 nm. The de’s obtained upon CT excitation
differed appreciably between 280 and 300 nm; the lower de’s at
280 nm may arise from possible contamination by direct
excitation. The de became much higher in CH2Cl2, but 2 was
not detected in methylcyclohexane. The de of 2 was enhanced
up to 90% in CH2Cl2 upon CT excitation at 300 nm.
Eyring analyses8 of the de’s obtained at various temperatures
were performed by plotting the ln([(1R,3S)-2]/[(1S,3S)-2])
values against 1/T to give bent lines, irrespective of the
irradiation λ (Table S5), clearly indicating switching of the
diastereodifferentiation mechanism in the temperature range
examined.9 The stereochemistry of cyclization product 2 is a
consequence of the ground- and excited-state equilibria as well
as the relative rate of cyclization (vide infra). If the cyclization
efficiencies are comparable for pro-S/R, the de of 2 directly
reflects the conformer distribution in the ground state at lower
temperatures, where the conformer equilibria become much
slower (relative to the cyclization). Thus, at low temperatures,
the Eyring plot gave a single straight line with the slope and
intercept being obviously dependent on excitation λ (Tables 2
Table 1. Temperature and Solvent Effects on the Product
Ratio and Diastereoselectivity in Photocyclization/
Table 2. Activation Parameters for the Preferential
Formation of (1R,3S)-2 over (1S,3S)-2 upon Direct and CT
Rearrangement of 1 to 2 and 3 upon Direct (254 nm) and
a
a
Excitations of 1
CT Excitations (280 and 300 nm)
excitation mode wavelength/nm ΔΔH⧧/kJ mol−1 ΔΔS⧧/J mol−1 K−1
excitation wavelength
direct
CT
254
280
300
+0.82
−5.9
254 nm
2/3 % de
280 nm
300 nm
−8.1
+42
temp/
b
b
b
b
b
−20
+96
solvent
MeCN
°C
2/3
% de
2/3
% de
a
b
See the Supporting Information for the Eyring plots. Tentative value
derived from the de’s obtained at two different temperatures.
20
0
0.09
0.13
−24
−17
−16
−14
+20
−
0.33
0.44
+28
+66
+52
+45
+82
−
0.86
1.91
+57
+63
+75
+51
+90
−
−20
−40
20
0.27
0.73
2.26
0.35
0.53
2.63
CH2Cl2
0.01
0.02
0.61
and S5), indicating the operation of a single, but distinct,
diastereodifferentiation mechanism for direct (254 nm) and CT
(280 nm) excitation. The activation parameters at 300 nm,
gauged tentatively, gave trends similar to those at 280 nm.
Contrasting trends between the direct and CT excitations are
common to other D/A systems hitherto studied. The
magnitudes of ΔΔH⧧ and ΔΔS⧧ also depend on the mode
of excitation; i.e., the parameters are significantly larger for the
CT rather than direct excitation. This contrasting behavior
upon direct versus CT excitation led us to the efficient
wavelength control in chiral photoreaction.3 The difference in
activation parameter, however, becomes less significant at
higher temperatures (Table S5), where the fast excited-state
equilibrium cancels the effect of selective direct/CT excitation
(vide infra).10
c
MCH
20
<0.01
<0.01
<0.01
a
[1] = 0.1 mM; irradiated under Ar at 254 nm for 1 h or at 280/300
nm for 2 h. Diastereomeric excess of 2; the + and − signs indicate
favored formation of (1R,3S)- and (1S,3S)-2, respectively. Methylcy-
b
c
clohexane.
While (1R,3S)-2, designated by positive de values in Table 1,
was favored under most of the reaction conditions employed,
(1S,3S)-2 became predominant upon irradiation at 254 nm in
acetonitrile. The sample concentration was kept low at 0.1 mM
to expedite the direct comparison of the photobehavior with
the spectral examinations. The 2/3 ratio became smaller in less
polar solvents, indicating that the transition state or
intermediate leading to cyclization is polar. In contrast, the
8083
dx.doi.org/10.1021/ja302768w | J. Am. Chem. Soc. 2012, 134, 8082−8085

Journal of the American Chemical Society
Communication
For better understanding of the photochemical behavior, we
applied the theoretical calculations to this intramolecular D/A
system. The geometrical optimizations were performed at the
SCS-MP2/TZVPP level, which has been shown to provide the
most reliable energies and geometries, comparable to the
CCSD(T) level, for medium-to-large molecules.11 All possible
conformations of 1 were initially considered, and 6 out of 54
conformers were found feasible; see the structures in Table 3
CD spectra are sensitive to conformational changes in
solution.16 Hence, experimental and theoretical CD spectral
studies provided further insights into the mechanism of the
solvent, temperature, and wavelength dependences observed in
the photoreaction of 1. The CD spectra of 1 were recorded in a
variety of solvents (Figure 1a) to exhibit a negative Cotton
Table 3. Relative Energies ΔE and Conformer Populations
(in Parentheses) in Gas Phase and in Acetonitrile Calculated
a
by the SCS-MP2 Methods
ground state
conformation
gas phase
acetonitrile
excited state
gauche-1
(G+)
pro-S, −ac
pro-R, +sc
≡ 0 (55.0)
0.20 (39.0)
≡ 0 (69.7)
0.32 (22.0)
1.5 (6.6)
1.1 (11.7)
gauche-2
(G−)
pro-S, −ac
pro-R, +ac
2.1 (1.7)
2.7 (0.5)
2.4 (1.2)
2.3 (1.4)
1.8 (3.9)
≡ 0 (77.0)
trans
pro-S, +ac
pro-R, −sc
1.8 (2.7)
2.3 (1.1)
1.6 (4.5)
2.4 (1.3)
3.2 (0.4)
3.2 (0.4)
(T)
Figure 1. (a) Experimental CD spectra of (S)-1 in methylcyclohexane,
diethyl ether, dichloromethane, and acetonitrile (from top to bottom
at 260 nm) at 20 °C. (b) Experimental CD spectra of (S)-1 in
acetonitrile at +40, +20, 0, −20, and −40 °C. (c) Theoretical CD
calculated for pro-R- and pro-S-G+ conformers of (S)-1 and the
average; all obtained by Boltzmann population-weighted averaging of
the six conformers (shown in Table 3) at the RI-CC2/TZVPP level.
The spectra were red-shifted by 0.4 eV and scaled by a factor of 10. (d)
Theoretical UV spectra of 1.
a
Relative SCS-MP2/TZVPP energies (in kcal mol−1) and Boltzmann
populations (%) at 25 °C in gas phase or in acetonitrile (COSMO, ε =
36.64). The excited-state energies and populations are based on the
RI-CC2 excitation energies for the ground-state geometries. The
numbers in the figure show the distance between the atoms to be
bonded upon cyclization, while those in the parentheses are for the
diastereomeric product (if formed).
(see also SI for the details of geometry search). These are pro-
R/S pairs of trans and two gauche conformations with respect
to the central C3−C4 bond (Chart S1), and the trans pair
cyclize to antipodal products. As can be seen from the
Boltzmann distributions in Table 3, the pro-R- and pro-S-G+
conformers are of primary importance in considering the
product stereoselectivity. The preference for pro-S-G+ in the
ground state is further emphasized in acetonitrile compared to
the gas phase, as supported by the relative energies and
populations obtained by using a relatively simple PCM-type
solvation model (COSMO).12
The first excitation energies of the six conformers were
calculated by the linear-response resolution-of-identity approx-
imate coupled-cluster theory (RI-CC2).13,14 Although the
vibrational zero-point energies were not taken into account in
our calculations, the conformer preference was switched to
favor the G− conformer, in particular pro-R rather than pro-S,
in the excited state. Thus, the (1S)-product should be favored,
unless the excited-state equilibrium is slower than the
photocyclization. In this relation, it should be noted that our
calculations predicted that the trans conformers are less
populated both in the ground and excited states, despite the
previous suggestion of the trans conformer as a precursor to the
rearrangement product.4 In any case, the ensemble15 of pro-R/
S-T, -G+, and -G− is indispensable for explaining the
diastereodifferentiation mechanism.
effect at ∼260 nm in acetonitrile, which was inverted in sign in
less polar solvents to eventually afford the strongest positive
Cotton effect in nonpolar methylcyclohexane. Such a dramatic
solvent dependence can be readily rationalized if the pro-S-G+
conformer is favored in acetonitrile, while the pro-R-G+
gradually becomes dominant in less polar solvents, since the
theoretical CD calculations predicted negative and positive
Cotton effects at ∼250−270 nm for pro-S- and pro-R-G+
conformers, respectively (Figure 1c). Experimentally, the
negative Cotton effect at ∼260 nm in acetonitrile was further
enhanced by lowering temperature (Figure 1b), indicating the
existence of the equilibrium between the pro-S and the pro-R
forms that gradually shifts to the pro-S side at lower
temperatures.
Scheme 2 illustrates the diastereodifferentiation mechanism
and the species involved in the photocyclization of 1. In the
ground state, the pro-R/S-G+ conformers are predominant
(highlighted by boldface) and equilibrated with the other minor
conformers (e.g., G− and T). These conformations are
represented by two ensembles, pro-R- and pro-S-1, and their
equilibrium by one constant K. Upon irradiation, these
conformers are vertically excited to the Franck−Condon states,
which are subsequently subjected to the excited-state
equilibrium with constant K* (Table 3), favoring pro-R
(highlighted in Scheme 2), as well as the competition with
8084
dx.doi.org/10.1021/ja302768w | J. Am. Chem. Soc. 2012, 134, 8082−8085

Journal of the American Chemical Society
Communication
Scheme 2. Factors Controlling the Diastereodifferentiating
Photocyclization of 1
ACKNOWLEDGMENTS
■
a
Financial support of this research by Grant-in-Aid for Scientific
Research (No. 23350018, 24655029, and 21245011) from
JSPS, the Mitsubishi Chemical Corporation Fund, the
Sumitomo Foundation, the Shorai Foundation for Science
and Technology, and the Kurata Memorial Hitachi Science and
Technology Foundation are gratefully acknowledged. We thank
Profs. S. Grimme and C. Bohne for fruitful discussions.
a
K and K* represent the apparent equilibrium constants between the
REFERENCES
pro-R/S conformer ensembles (each composed of T, G+, and G−) in
■
(1) (a) Rau, H. Chem. Rev. 1983, 83, 535. (b) Inoue, Y. Chem. Rev.
1992, 92, 741. (c) Griesbeck, A. G.; Meierhenrich, U. J. Angew. Chem.,
Int. Ed. 2002, 41, 3147. (d) Hoffmann, N. Chem. Rev. 2008, 108, 1052.
(e) Bauer, A.; Westkaemper, F.; Grimme, S.; Bach, T. Nature 2005,
436, 1139.
the ground and excited states, respectively.
cyclization (kR/kS) and deactivation (k−R/k−S). Furthermore,
the extinction coefficients of pro-R- and pro-S-1 are not equal
to each other (εR ≠ εS) and vary with the excitation wavelength.
Hence, the conformer population in the ground state is not
directly reflected in the product distribution. Taking into
account these factors and the mechanism proposed, we can
rationalize the effects of solvent, temperature, and wavelength
as follows: (1) The preference for the (1R)-product in
dichloromethane, than in acetonitrile, is attributable mostly to
the ground-state equilibrium shift to pro-R, as supported by the
experimental and theoretical CD spectra. (2) Since the Eyring
plots of the de values are bent, the diastereodifferentiation
mechanism is likely to be switched in the temperature range
examined, for which the decelerated ground-state equilibrium,
relative to the excited-state equilibrium at lower temperatures,
would be responsible. (3) Irradiation at longer wavelength leads
to the selective excitation of pro-R, as supported by experiment
and theory (Figure 1d).
(2) (a) Inoue, Y.; Wada, T.; Asaoka, S.; Sato, H.; Pete, J. P. Chem.
Commun. 2000, 251. (b) Inoue, Y.; Ikeda, H.; Kaneda, M.; Sumimura,
T.; Everitt, S. R. L.; Wada, T. J. Am. Chem. Soc. 2000, 122, 406.
(c) Inoue, Y.; Matsushima, E.; Wada, T. J. Am. Chem. Soc. 1998, 120,
10687. (d) Inoue, Y.; Yokoyama, T.; Yamasaki, N.; Tai, A. J. Am.
Chem. Soc. 1989, 111, 6480. (e) Inoue, Y.; Yokoyama, T.; Yamasaki,
N.; Tai, A. Nature 1989, 341, 225.
(3) (a) Saito, H.; Mori, T.; Wada, T.; Inoue, Y. J. Am. Chem. Soc.
2004, 126, 1900. (b) Saito, H.; Mori, T.; Wada, T.; Inoue, Y. Org. Lett.
2006, 8, 1909. (c) Matsumura, K.; Mori, T.; Inoue, Y. J. Am. Chem.
Soc. 2009, 131, 17076. (d) Matsumura, K.; Mori, T.; Inoue, Y. J. Org.
Chem. 2010, 75, 5461.
(4) Cookson, R. C.; Sadler, D. E.; Salisbury, K. J. Chem. Soc., Perkin
Trans. 2 1981, 774.
(5) Ito, T.; Nishiuchi, E.; Yang, C.; Fukuhara, G.; Mori, T.; Inoue, Y.
Photochem. Photobiol. Sci. 2011, 10, 1405.
(6) Cookson, R. C.; Kemp, J. E. J. Chem. Soc., Chem. Commun. 1971,
385.
We have shown that both the product selectivity and
diastereoselectivity can be critically controlled by irradiation
wavelength in this intramolecular D/A dyad system. The
stereochemical outcomes of photocyclization are significantly
affected also by solvent polarity and temperature, mostly due to
the shift of the ground-state conformer equilibrium. The
excited CT state generated upon excitation at the CT band
behaves differently from the conventional exciplex formed upon
local-band excitation. At higher temperatures, the excited-state
conformer equilibrium competes with the deactivation (k−R and
k−S) and cyclization (kR and kS) to achieve a population
different from that in the ground state. Interestingly, the 2/3
ratios and de values obtained upon CT excitation at 280 and
300 nm differed significantly, as a result of the different spectral
features of the G+ and G− conformers, enabling us to fine-tune
the stereochemical outcomes. We may conclude that
manipulating the excited-state ensemble is essential for critically
controlling the stereochemical outcomes of photoreaction by
temperature, solvent polarity, and most effectively by irradiation
wavelength.
(7) (a) Gorman, A. A.; Rodgers, M. A. J. J. Am. Chem. Soc. 1986, 108,
5074. (b) Sajimon, M. C.; Ramaiah, D.; Thomas, K. G.; George, M. V.
J. Org. Chem. 2001, 66, 3182.
(8) (a) Eyring, H. J. Chem. Phys. 1935, 3, 107. (b) Eyring, H. Chem.
Rev. 1935, 17, 65.
(9) (a) Abe, M.; Kawakami, T.; Ohata, S.; Nozaki, K.; Nojima, M. J.
Am. Chem. Soc. 2004, 126, 2838. (b) Adam, W.; Stegmann, V. R. J. Am.
Chem. Soc. 2002, 124, 3600.
(10) More details on the temperature effects remain to be elucidated
in future studies.
(11) (a) Grimme, S. J. Phys. Chem. A 2005, 109, 3067. (b) Grimme,
S. J. Chem. Phys. 2003, 118, 9095. (c) Grimme, S. J. Comput. Chem.
2003, 24, 1529.
(12) (a) Tomasi, J.; Persico, M. Chem. Rev. 1994, 94, 2027.
(b) Sinnecker, S.; Rajendran, A.; Klamt, A.; Diedenhofen, M.; Neese,
F. J. Phys. Chem. A 2006, 110, 2235.
(13) (a) Christiansen, O.; Koch, H.; Jorgensen, P. Chem. Phys. Lett.
1995, 243, 409. (b) Hattig, C.; Kohn, A. J. Chem. Phys. 2002, 117,
6939.
(14) For a recent report on the calculation of CD spectra by the RI-
CC2 method, see: Wakai, A.; Fukasawa, H.; Yang, C.; Mori, T.; Inoue,
Y. J. Am. Chem. Soc. 2012, 134, 4990.
(15) (a) van der Auweraer, M.; Gilbert, A.; de Schryver, F. C. J. Am.
Chem. Soc. 1980, 102, 4007. (b) Hirata, Y.; Kanda, Y.; Matag, N. J.
Phys. Chem. 1983, 87, 1659. (c) Patel, A. D.; Nocek, M.; Hoffman, B.
M. J. Am. Chem. Soc. 2005, 127, 16766.
(16) (a) Hembury, G. A.; Borovkov, V. V.; Inoue, Y. Chem. Rev.
2008, 108, 1. (b) Berova, N.; Di Bari, L.; Pescitelli, G. Chem. Soc. Rev.
2007, 36, 914. (c) Superchi, S.; Giorgio, E.; Rosini, C. Chirality 2004,
16, 422. See also: (d) Mori, T.; Inoue, Y. Angew. Chem., Int. Ed. 2005,
44, 2582. (e) Furo, T.; Mori, T.; Wada, T.; Inoue, Y. J. Am. Chem. Soc.
2005, 127, 8242; 2005, 127, 16338 (erratum).
ASSOCIATED CONTENT
* Supporting Information
Experimental procedures, details of theoretical calculation, and
spectroscopic data. This material is available free of charge via
the Internet at http://pubs.acs.org.
■
S
AUTHOR INFORMATION
Corresponding Author
tmori@chem.eng.osaka-u.ac.jp
■
Notes
The authors declare no competing financial interest.
8085
dx.doi.org/10.1021/ja302768w | J. Am. Chem. Soc. 2012, 134, 8082−8085