Wavelength control of diastereodifferentiating Paterno-Buechi reaction of chiral cyanobenzoates with diphenylethene through direct versus charge-transfer excitation

Article abstract of DOI:10.1021/ja907156j

(Chemical Equation Presented) In the diastereodifferentiating Patern _o-B_uechi reaction, the excited CT complex was distinctly different in structure and reactivity from the conventional exciplex, and the inherent diastereofacial sele

Full text of DOI:10.1021/ja907156j

Published on Web 11/05/2009
Wavelength Control of Diastereodifferentiating Paterno´-Bu¨chi Reaction of
Chiral Cyanobenzoates with Diphenylethene through Direct versus
Charge-Transfer Excitation
Kazuyuki Matsumura, Tadashi Mori,* and Yoshihisa Inoue*
Department of Applied Chemistry, Osaka UniVersity, 2-1 Yamada-oka, Suita 565-0871, Japan
Received August 30, 2009; E-mail: tmori@chem.eng.osaka-u.ac.jp
The Paterno´-Bu¨chi reaction provides us with a versatile route to a
variety of chiral oxetanes, which are often found in biologically
important molecules, including taxol, preussin, and oxetanocin, and
are recognized as promising drug-discovery modules.1 Thus, the
manipulation of stereoselectivity2 by the use of allylic strain3 and
hydrogen-bonding interactions4 has been a target of intensive study.
Most Paterno´-Bu¨chi reactions are known to proceed through triplet
1,4-biradical intermediates, which either revert to the starting materials
or cyclize to the corresponding oxetanes after spin inversion.5 The
stereochemistry of the product is thus determined in the initial addition6
and/or the second cyclization.7 Because of this mechanistic complexity
in oxetane formation, the Eyring-type treatment, in which the logarithm
of stereoselectivity is plotted against reciprocal temperature, often gives
a bent plot, indicating a switching of the mechanism in the temperature
range employed.8 More recently, the role of the singlet manifold in
the Paterno´-Bu¨chi reaction has been revealed from the stereochemical
viewpoint.9
(K ) 0.04 M-1) in methylcyclohexane at 25 °C (see Figure S1 in the
Supporting Information). Irradiation of a concentrated solution of 1a
(0.2 M) and 2 (1 M) at 254 nm led to the formation of a diastereomeric
mixture of oxetanes 3a in good combined yield (Scheme 1), in
accordance with the previous reports for achiral analogues.12 The yield
and diastereomeric excess (de) of 3a were not affected by the addition
of oxygen (air), in accord with the singlet mechanism (Table S1). The
de of 3a remained constant at 65% upon excitation at 254 and 290
nm but gradually decreased to 43-46% when the wavelength was
shifted to 310-320 nm, and eventually the epimeric product was
obtained in 4% de upon excitation at 330 nm (Table S2).13 These
results clearly indicate that the excited CT complex formed upon
excitation at longer wavelengths affords the same adduct 3a but that
this adduct is epimeric to the one produced from the conventional
exciplex generated upon direct excitation at shorter wavelengths. We
confirmed the absolute configuration of the products by X-ray
crystallography; the minor isomer has the 1′S,2R configuration (Figure
1, left).
In the present study, we report the wavelength-dependent
diastereodifferentiating Paterno´-Bu¨chi reaction of chiral p-cy-
anobenzoates (1a, 1b) with 1,1-diphenylethene (2). In our recent
study of the [2 + 2] photocycloaddition of stilbene with chiral
fumarates, we revealed that direct excitation of stilbene and selective
excitation of the CT band afford the same cyclobutane but with
totally different diastereoselectivities.10 This result indicates that
the excited CT complex differs in structure and reactivity from the
conventional exciplex and that these excited-state complexes do
not interconvert. The effect of excitation wavelength on the
photophysical behavior of the CT complex has also been investi-
gated spectroscopically.11 These observations prompted us to further
explore the general validity and feasibility of the wavelength-
dependent photochemistry based on the CT excitation as a
conventional yet powerful tool for critically manipulating the
product’s stereochemistry. Intriguingly, this photocycloaddition
reaction exhibited the opposite temperature dependence upon CT
versus direct excitation, reflecting the different degrees of freedom
in the excited species involved (see below).
Figure 1. (left) ORTEP drawing of the minor isomer, (1′S,2R)-3a. (right)
Structures of the diastereomeric CT complexes that are precursors to
(1′S,2S)- and (1′S,2R)-3a, optimized at the DFT-D-B-LYP/TZVP level. The
Tg+ conformers in the chiral group are shown.
Mechanistic elucidations of Paterno´-Bu¨chi reactions often rely
on temperature-dependence studies.7,8 Interestingly, the de values
of 3a or 3b obtained upon direct and CT excitation of a
methylcyclohexane solution of 1a or 1b and 2 (Table 1) showed
opposite temperature-dependence profiles. Thus, when the temper-
ature was lowered from 25 to -50 °C, the de decreased from 77 to
34% for 3a and from 74 to 66% for 3b upon direct excitation. In
sharp contrast, the same temperature change improved the de from
3 to 26% for 3a and from 18 to 44% for 3b upon CT excitation.
This contrasting behavior unambiguously confirms that the con-
ventional exciplex and the excited CT complex do not share a
common structure but have distinctly different structures and
reactivities that give cycloadduct 3 with significantly different
diastereoselectivities and temperature dependencies.
Scheme 1. Diastereodifferentiating Photocycloaddition of 1 to 2
upon Direct and CT-Band Excitation
From the Eyring analyses14 of the de’s using plots of ln{[(1′S,2S)-
3]/[(1′S,2R)-3]} versus 1/T (Figure S2), we obtained the differential
activation enthalpy (∆∆Hq) and entropy (∆∆Sq) (Table 2). The
Eyring plots gave nice straight lines that had different slopes and
intercepts with opposite signs, indicating the operation of a single
but distinct diastereodifferentiation mechanism for each substrate
The CT complexation behavior of 1a with 2 was investigated by
UV-vis spectroscopy and found to have a modest binding constant
9
17076 J. AM. CHEM. SOC. 2009, 131, 17076–17077
10.1021/ja907156j CCC: $40.75  2009 American Chemical Society

C O M M U N I C A T I O N S
and excitation mode. As a consequence of the opposite temperature
dependence of the de, oppositely signed ∆∆Hq and ∆∆Sq values
were obtained for the direct and CT excitations. It should be noted
that the positive de, or the preferred formation of (1′S,2S)-3, is an
incidental phenomenon only in the temperature range employed
and that the epimeric product should be obtained at lower and higher
temperatures upon direct and CT excitation, respectively.
in the exciplex, where the interaction occurs at the carbonyl (to
give oxetane). The diastereoselectivities at low temperatures, where
the enthalpy dominates the steric consequence of the peripheral
modification (from 1a to 1b, Table 2), and the substantial solvent
effect (Figure S4) jointly support this conclusion. The stereoselec-
tivities in acetonitrile were low for both excitation modes, presum-
ably because of dissociation of the radical ionic species, keeping
distinctly different linear correlations.
In this study, we have demonstrated for the first time that the
conventional exciplex and the excited CT complex are independent,
noninterconvertible excited species that possess distinctly different
structures and reactivities, occasionally leading to opposite stere-
ochemistries. The magnitude of ∆∆Sq is appreciably larger for the
direct excitation than for the CT excitation. This seems reasonable,
as the exciplex is relatively flexible and susceptible to environmental
variants such as temperature, while the CT complex is better stacked
to each other in the ground state and seems to be less flexible in
the excited state, giving the significantly smaller differential
activation parameters.
Table 1. Temperature Effects on the Diastereoselective
Paterno´-Bu¨chi Reaction of 1a or 1b with 2 upon Direct and CT
Excitationa
% de of 3a
direct
% de of 3b
direct
temperature (°C)
CT
CT
50
25
0
-25
-50
77
64
63
43
34
3
4
3
24
26
74
71
-
18
22
-
41
44
-
68
a Using a methylcyclohexane solution of 1 (0.2 M) and 2 (1.0 M)
irradiated at 290 nm for direct excitation (3 h for 3a or 5 h for 3b) and
at 330 nm for CT-band excitation (10 h for 3a or 13 h for 3b).
Acknowledgment. This work is dedicated to the memory of
Professor Jay Kazuo Kochi. The support of this work by a Grant-
in-Aid for Scientific Research, JSPS (T.M. and Y.I.), the Mitsubishi
Chemical Corporation Fund (T.M.), and the Sumitomo Foundation
(T.M.) are gratefully acknowledged. We thank Prof. S. Grimme
for the fruitful discussion of theoretical aspects.
Table 2. Activation Parameters for the Formation of
Diastereomeric Oxetanes 3a and 3b in the Paterno´-Bu¨chi
Reaction of 1a or 1b with 2 upon Direct and CT Excitationa
substrate
excitation mode
∆∆Hq (kJ mol-1
)
∆∆Sq (J mol-1 K-1
)
1a
direct
CT
direct
CT
+6.4
-3.3
+1.6
-3.8
+39
-10
+21
-9
Supporting Information Available: Experimental procedures,
theoretical calculations, spectroscopic data, and a CIF file. This material
is available free of charge via the Internet at http://pubs.acs.org.
1b
a See Figure S2 in the Supporting Information for the Eyring plots.
References
The magnitudes of ∆∆Hq and ∆∆Sq also depend on the mode
of excitation, i.e., the activation parameters are significantly larger
for direct excitation than for CT excitation, except for the ∆∆Hq
for 1b. In this Paterno´-Bu¨chi reaction occurring in the singlet
manifold, the photocyclization likely proceeds through a concerted
mechanism or a short-lived singlet 1,4-biradical, and hence, the
diastereofacial-differentiating complexation of chiral cyanobenzoate
1 with alkene 2 plays a decisive role in determining the product’s
de (unless the efficiency of the subsequent cycloaddition consider-
ably differs between the diastereomeric complexes). If this is the
case, the activation parameters gained upon direct and CT excitation
directly reflect the enthalpic and entropic contributions to the
exciplex and CT complex formation, respectively. The first excited
singlet state of cyanobenzoates is mostly π-π* in nature but
borrows some intensity from the n-π* state, as indicated by the
theoretical calculations (Figure S3). When such an excited state of
cyanobenzoate interacts with an electron-rich alkene to form an
exciplex, the interaction is expected to occur near the carbonyl
group. The larger ∆∆Sq values for the direct excitation may be
related to the relatively flexible structure of the exciplex thus
formed. On the other hand, the CT complexation is achieved in
general by maximizing the π overlap between the donor and
acceptor to give a conformationally rigid structure,15 which is also
supported by the dispersion-corrected DFT calculations16 (Figure
1, right). Because of the strong π-π stacking, the excited CT
complex is not be able to gain conformational freedom upon
excitation, and the entropic difference between the members of the
diastereomeric pair should be much smaller than that for the
exciplex pair, in nice agreement with the experimental results (Table
2). The lower selectivities upon CT excitation may be attributed to
the fact that the chiral auxiliary is located at a more distant position
from the point of interaction (arene-arene) in the CT complex than
(1) (a) Wang, M.; Cornett, B.; Nettles, J.; Liotta, D. C.; Snyder, J. P. J. Org.
Chem. 2000, 65, 1059. (b) Wuitschik, G.; Rogers-Evans, M.; Mu¨ller, K.;
Fischer, H.; Wagner, B.; Schuler, F.; Polonchuk, L.; Carrieira, E. M. Angew.
Chem., Int. Ed. 2006, 45, 7736.
(2) (a) Griesbeck, A. G.; Fiege, M. In Molecular and Supramolecular
Photochemistry; Ramamurthy, V., Schanze, K. S., Eds.; Marcel Dekker:
New York, 2000; Vol. 6, p 33. (b) Fleming, S. A.; Bradford, C. L.; Gao,
J. J. In Molecular and Supramolecular Photochemistry; Ramamurthy, V.,
Schanze, K. S., Eds.; Marcel Dekker: New York, 1997; Vol. 1, p 187.
(3) (a) Bach, T.; Jo¨dicke, K.; Kather, K.; Fro¨hlich, R. J. Am. Chem. Soc. 1997,
119, 2437. (b) Bach, T.; Jo¨dicke, K.; Kather, K.; Hecht, J. Angew. Chem.,
Int. Ed. Engl. 1995, 34, 2271.
(4) (a) Griesbeck, A. G.; Bondock, S. J. Am. Chem. Soc. 2001, 123, 6191. (b)
Adam, W.; Peters, K.; Peters, E. M.; Stegmann, V. R. J. Am. Chem. Soc.
2000, 122, 2958. (c) Bach, T.; Bergmann, H.; Harms, K. J. Am. Chem.
Soc. 1999, 121, 10650.
(5) (a) Freilich, S. C.; Peters, K. S. J. Am. Chem. Soc. 1985, 107, 3819. (b)
Freilich, S. C.; Peters, K. S. J. Am. Chem. Soc. 1981, 103, 6255.
(6) Lhiaubet-Vallet, V.; Encinas, S.; Miranda, M. A. J. Am. Chem. Soc. 2005,
127, 12774.
(7) Buschmann, H.; Scharf, H.-D.; Hoffmannn, N.; Esser, P. Angew. Chem.,
Int. Ed. Engl. 1991, 30, 477.
(8) (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. (c) Adam, W.; Stegmann, V. R.; Weikno¨tz, S. J. Am.
Chem. Soc. 2001, 123, 2452.
(9) (a) Griesbeck, A. G.; Fiege, M.; Bondock, S.; Gudipati, M. S. Org. Lett.
2000, 2, 3623. (b) Griesbeck, A. G.; Bondock, S.; Cygon, P. J. Am. Chem.
Soc. 2003, 125, 9016.
(10) (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.
(11) (a) Nicolet, O.; Banerji, N.; Page`s, S.; Vauthey, E. J. Phys. Chem. A 2005,
109, 8236. (b) Levy, D.; Arnold, B. R. J. Phys. Chem. A 2005, 109, 2113.
(12) Shigemitsu, Y.; Katsuhara, Y.; Odaira, Y. Tetrahedron Lett. 1971, 12, 2887.
(13) We chose 330 and 290 nm for CT and direct excitation, respectively.
Although the irradiation at 290 nm may excite a hidden CT band, its effect
seems to be small, as the de was distinctly different or even opposite in
sign.
(14) Eyring, H. J. Chem. Phys. 1935, 3, 107.
(15) (a) Rosokha, S. V.; Kochi, J. K. J. Org. Chem. 2002, 67, 1727. Also
see: (b) Kobayashi, T.; Yoshihara, K.; Nagakura, S. Bull. Chem. Soc. Jpn.
1971, 44, 2603.
(16) Grimme, S.; Antony, J.; Schwabe, T.; Mu¨uck-Lichtenfeld, C. Org. Biomol.
Chem. 2007, 5, 741.
JA907156J
9
J. AM. CHEM. SOC. VOL. 131, NO. 47, 2009 17077