Published on Web 02/12/2004

Mechanism of Stereo- and Regioselectivity in the

Paternò-Bu1 chi Reaction of Furan Derivatives with Aromatic
Carbonyl Compounds: Importance of the Conformational
Distribution in the Intermediary Triplet 1,4-Diradicals
Manabu Abe,*,† Takanori Kawakami,† Shinji Ohata,† Koichi Nozaki,‡ and
Masatomo Nojima†
Contribution from the Department of Materials Chemistry, Graduate School of Engineering,
Osaka UniVersity (HANDAI), Suita 565-0871, Osaka, Japan, and Department of Chemistry,
Graduate School of Science, Osaka UniVersity, Toyonaka 560-0043, Osaka, Japan
Received November 7, 2003; E-mail: abe@ap.chem.eng.osaka-u.ac.jp

Abstract: Temperature and substituent effects on the stereo- and regioselectivity have been investigated
in the photochemical [2 + 2] cycloaddition reaction, the so-called Paternò-Büchi (PB) reaction, of
unsymmetrically substituted furans 2a,b (2-methyl- and 3-methylfuran) with aromatic carbonyl compounds
1a,b (benzaldehyde and benzophenone). The regio-random but stereoselective (exo/endo > 97/3) formation
of lower substituted oxetane 3a and higher substituted oxetane 4a is found in the reaction with benzaldehyde
(1a). The exclusive stereoselectivity is not dependent on the position of methyl substituent on the furan
ring and the reaction temperature. The double-bond selection (3a versus 4a) is slightly dependent on the
reaction temperature (3a/4a ) 55/45 to 40/60). The Eyring plots of the regioselectivity are linear.
Contrastively, in the reaction with benzophenone (1b), the double-bond selection (3b versus 4b) largely
depends on the reaction temperature. The Eyring plots are not linear, but the inflection points are observed.
The transient absorption spectroscopic analyses (picosecond time scale) clarify the intervention of triplet
2-oxabutane-1,4-diyls in the photochemical processes. Computational studies reveal the equilibrium
structures of the triplet diradicals, energy barriers between the conformers, and the equilibrium constants.
A rational mechanism is herein proposed by the support of both experimental and computational
investigations to account for not only the exclusive formation of the exo-configured oxetanes 3a and 4a
but also the nonlinear Eyring plots observed in the reaction with 1b.

Introduction Scheme 1. Paternò-Büchi Reaction of Furan Derivatives with

Aromatic Carbonyl Compounds
Schenck1 first reported the exclusive formation of 2-alkoxy-
oxetane, 6,6-diphenyl-2,7-dioxabicyclo[3.2.0]hept-3-ene, in the
Paternò-Büchi (PB) photochemical [2 + 2] cycloaddition2 of
furan with benzophenone. After his finding, several chemists
explored the regioselectivity3 (double-bond selection) for the
reaction of unsymmetrically substituted furans. The stereo-
selectivity4 was also investigated for the reactions with unsym-
metrical carbonyl compounds. For example, the exclusive
formation of higher substituted oxetane 4 was reported in the
reaction of 2-methylfuran (2a) or 3-methylfuran (2b) with
benzophenone (1b) at ambient temperature (Scheme 1).3b,d In
contrast, a regio-random formation of oxetanes 3aa and 4aa was found in the reaction with benzaldehyde 1a,3a,4h although
† the exo-selective formation of the oxetanes was observed
Graduate School of Engineering, Osaka University.
‡ Graduate School of Science, Osaka University. (Scheme 1).4h The mechanism to interpret the regio- and
(1) Schenck, G. O.; Hartmann, W.; Steinmetz, R. Chem. Ber. 1963, 96, 498. stereoselectivity is still unclear, though the PB reaction is one
(2) (a) Paternò, E.; Chieffi, G. Gazz. Chim. Ital. 1909, 39, 341. (b) Büchi, G.;
Inman, C. G.; Lipinsky, E. S. J. Am. Chem. Soc. 1954, 76, 4327. of the fundamental photochemical cycloaddition reactions.4m,5-7
(3) (a) Toki, S.; Shima, K.; Sakurai, H. Bull. Chem. Soc. Jpn. 1965, 38, 760.
(b) Rivas, C.; Payo, E. J. Org. Chem. 1967, 32, 2918. (c) Sakurai, H.; In the present study, we have investigated temperature and
Shima, K.; Toki, S. Nippon Kagaku Zasshi 1968, 89, 537. (d) Nakano, T.; substituent effects on the selectivities in the PB reactions that
Rivas, C.; Perez, C. J. Chem. Soc., Perkin Trans. 1 1973, 2322. (e)
Schreiber, S. L.; Desmaele, D.; Porco, J. A., Jr. Tetrahedron Lett. 1988, have not been reported thus far.8 With help of computational
51, 6689. (f) Carless, H. A.; Halfhide, F. E. J. Chem. Soc., Perkin Trans. calculations, we report here the conclusive mechanism for
1 1992, 1081. (g) D’Auria, M.; Racioppi, R.; Romaniello, G. Eur. J. Org.
Chem. 2000, 3265. rationalizing the regio- and stereoselectivity.
2838 9 J. AM. CHEM. SOC. 2004, 126, 2838-2846 10.1021/ja039491o CCC: $27.50 © 2004 American Chemical Society
Paternò−Büchi Reaction of Furan Derivatives ARTICLES

Figure 1. Structural determination of exo-oxetanes 3 and 4.

Results
Photoreaction of 2-Methylfuran (2a) with Aromatic Car-
bonyl Compounds 1a and 1b. A degassed toluene solution of
freshly distilled 2-methylfuran (2a, 0.50 M) and aromatic
carbonyl compounds 1a or 1b (0.05 M) was irradiated with a
high-pressure Hg lamp (300 W, hν > 300 nm) through a Pyrex
filter (Scheme 1). To avoid the formation of bisoxetanes9 by
the reaction of 1 with the bicyclic monooxetane 3 and/or 4, the
irradiation was stopped at the stage of ca. 30% conversion of
the carbonyl compounds. Under the reaction conditions only Figure 2. Eyring plots of regioselectivity, ln(3/4) against 1/T (K-1), in the
photoreaction of 2-methylfuran (2a) with (a) 9 benzaldehyde (1a), exo/
oxetanes 3 and 4 (total yields > 95%) were detected as endo > 97/3 and (b) b benzophenone (1b).
photoprimary products, which were carefully analyzed by 1H
NMR (600 or 270 MHz) spectroscopy. Thus, it was feasible to As for the photoreaction of 2a with 1b, the regioselectivity
determine accurately the product ratios by comparing the 1H 3ba (R ) Ph, R1 ) Me, R2 ) H) versus 4ba (R ) Ph, R1 )
NMR peak areas of the isomeric oxetanes, error (3%. Me, R2 ) H) was largely dependent on the reaction temperature
In the photoreaction with 1a, two regioisomeric oxetanes 3aa (Figure 2b). Thus, regioisomers 3ba and 4ba (isomer ratio 42/
(R ) R2 ) H, R1 ) Me) and 4aa (R ) R2 ) H, R1 ) Me) 58) were obtained at -77 °C in high yields (>95%, Φ3+4 )
were observed in the photolyzate. The product structures were 0.39). However, the selective formation of higher substituted
confirmed by the 1H and 13C NMR (DEPT) spectroscopic oxetane 4ba was observed at 61 °C, and the product ratio was
analyses. The exo-configuration of the oxetanes was unequivo- 3ba/4ba ) 19/81 (Φ3+4 ) 0.31). It should be noted that the
cally determined by the 1H NMR (600 MHz) NOE measure- Eyring plot for the regioselectivity was not linear but S-shaped
ments (Figure 1). (Figure 2b). The nonlinear Eyring plots clearly indicate that the
In the reactions, we could not detect any trace of the endo- selectivity-determining step varies during the change in the
isomers (<3%) by means of 1H NMR (600 MHz) spectroscopic reaction temperature (vide infra).11
analyses. The highly exo-selective formation of the oxetanes Photoreaction of 3-Methylfuran (2b) with Aromatic Car-
(quantum yield, Φ3+4 ) 0.31-0.45) was observed in the bonyl Compounds 1a and 1b. The temperature effect on the
temperature range (-76 to +52 °C) under investigation (Figure regioselectivity in the photoreaction of 3-methylfuran (2b) was
2a). The Eyring plot10 of the regioselectivity, ln(3aa/4aa) against also investigated under the similar conditions mentioned above
1/T, is shown in Figure 2a. The regioselectivity was linearly for the reaction of 2a. The stereoselective formation (Φ3+4 )
changed with varying reaction temperature. The activation 0.23-0.28) of exo-configured oxetanes 3ab (R ) R1 ) H, R2
parameters, ∆∆Hq ) -0.24 kcal/mol and ∆∆Sq ) -1.1 cal/ ) Me) and 4ab (R ) R1 ) H, R2 ) Me) was found in the
Kmol, were determined from the slope and intercept of the linear reaction with 1a (Figure 1). The diastereoselectivity (>94% de)
plot. was independent of the reaction temperature. As observed in
the reaction of 2a, the Eyring plot for the regioselectivity was
(4) (a) Shima, K.; Sakurai, H. Bull. Chem. Soc. Jpn. 1966, 39, 1806. (b)
Whipple, E. B.; Evanega, G. R. Tetrahedron 1968, 24, 1299. (c) Jarosz, linear (Figure 3a). From the slope and intercept, the activation
S.; Zamojski, A. Tetrahedron 1982, 38, 1447. (d) Jarosz, A.; Zamojski, A. parameters were determined to be ∆∆Hq ) -0.58 kcal/mol and
Tetrahedron 1982, 38, 1453. (e) Schreiber, S. L.; Hoveyda, A. H.; Wu,
H.-J. J. Am. Chem. Soc. 1983, 105, 660. (f) Schreiber, S. L.; Satake, K. J. ∆∆Sq ) -2.7 cal/Kmol.
Am. Chem. Soc. 1983, 105, 6723. (g) Schreiber, S. L.; Hoveyda, A. H. J. In sharp contrast to the temperature effect on the regioselec-
Am. Chem. Soc. 1984, 106, 7200. (h) Schreiber, S. L.; Porco, J. A., Jr. J.
Org. Chem. 1989, 54, 4721. (i) Cantrell, T. S.; Allen, A. C.; Ziffer, H. J. tivity observed in the reaction of 2a with 1b (Figure 2b), the
Org. Chem. 1989, 54, 140. (j) Hambalek, R.; Just, G. Tetrahedron Lett. product ratio of 3bb/4bb in the reaction of 2b nonlinearly
1990, 31, 4693. (k) Hambalek, R.; Just, G. Tetrahedron Lett. 1990, 31,
5445. (l) Griesbeck, A. G.; Mauder, H.; Peters, K.; Peters, E.-M.; Schnering, increased with the increase of the reaction temperature (Figure
H. G. Chem. Ber. 1991, 124, 407. (m) Griesbeck, A. G.; Buhr, S.; Giege, 3b). Thus, at low temperatures (ca. <-20 °C), the regioselec-
M.; Schmickler, H.; Lex, J. J. Org. Chem. 1998, 63, 3847. (n) Hu, S.;
Neckers, D. C. J. Chem. Soc., Perkin Trans. 2 1999, 1771. (o) Zhang, Y.; tivity (3bb/4bb) was around 17/83, whereas the ratio was 34/
Xue, J.; Gao, Y.; Fun, H.-K.; Xu, J.-H. J. Chem. Soc., Perkin Trans. 1 66 at +56 °C. It should be noted that the quantum yield, Φ3+4,
2002, 345.
(5) Toki, S.; Sakurai, H. Bull. Chem. Soc. Jpn. 1967, 40, 2885. for the oxetane formations was largely dependent on the reaction
(6) Abe, M.; Torii, E.; Nojima, M. J. Org. Chem. 2000, 65, 3426. temperature. Thus, the yield drastically decreased with the
(7) Abe, M. In Organic Photochemistry and Photobiology, 2nd ed.; Horspool,
W., Lenci, F., Eds.; CRC Press: Boca Raton, FL, 2004; Chapter 62. increase of the reaction temperature: 0.39 (at -77 °C) f 0.06
(8) In ref 3a,b, the exclusive formation of more-substituted oxetane 4 was
reported, but in this study the regioselectivity has been found to largely
(at +56 °C).
depend on the reaction temperature.
(9) (a) Ogata, M.; Watanabe, H.; Kano, H. Tetrahedron Lett. 1967, 533. (b) (11) (a) Bushmann, H.; Scharf, D. N.; Hoffmann, N. Angew. Chem., Int. Ed.
Letich, J. Tetrahedron Lett. 1967, 1937. (c) Evanega, G. R.; Whipple, E. Engl. 1991, 30, 47. (b) Göbel, T.; Scharpless, K. B. Angew. Chem., Int.
B. Tetrahedron Lett. 1967, 2163. (d) Toki, S.; Sakurai, H. Tetrahedron Ed. Engl. 1993, 32, 1329. (c) Hale, K. J.; Ridd, J. H. J. Chem. Soc., Chem.
Lett. 1967, 4119. Commun. 1995, 357. (d) Gypser, A.; Norrby, P.-O. J. Chem. Soc., Perkin
(10) Eyring, H. J. Chem. Phys. 1935, 3, 107. Trans. 2 1997, 939.

J. AM. CHEM. SOC. 9 VOL. 126, NO. 9, 2004 2839

ARTICLES Abe et al.

it is reasonable to conclude that the new species are triplet 1,4-

diradicals T-DR3ba and/or T-DR4ba (Scheme 4). In the
photoreaction of 2b with 1b, the quite similar transient absorp-
tion spectrum was observed (λmax 530 nm, kd ) 3.9 × 108 s-1),
which is also assigned to the triplet diradicals T-DR3bb and/
or T-DR4bb.

Figure 3. Eyring plots of regioselectivity, ln(3/4) against 1/T (K-1), in the

photoreaction of 3-methylfuran (2b) with (a) 9 benzaldehyde (1a), exo/
endo > 97/3 and (b) b benzophenone (1b).

In the following section, we will discuss the mechanism for

the exo-selective formation of oxetanes 3 and 4 in the reaction
with 1a and for the regioselectivity (nonlinear Eyring plots) in
the reaction with 1b.
Discussion
exo-Selective Formation of Oxetanes 3 and 4. As observed
for the PB reaction with benzophenone (1b), triplet 2-oxabutane-
1,4-diyls T-DR3a and T-DR4a are also proposed to be the
intermediates for the formation of exo-oxetanes 3 and 4 in the
reaction with 1a (Scheme 2). The intersystem crossing (ISC)
process to singlet is necessary so that triplet diradicals may give
products. Since the lifetime of singlet diradicals is usually quite
Figure 4. Transient absorption spectra of (a) triplet benzophenone 1b*
short,15 the conformations of their precursor, triplet diradicals,
and (b) triplet 1,4-diradicals T-DR3ba and/or T-DR4ba formed in the play an important role to determine the product selectivity
irradiation of 5 M 2-methylfuran (2a) with 0.025 M benzophenone (1b) in (memory effect).16-19 To get information on the mechanism for
acetonitrile at 1000 ps. the exclusive formation of exo-oxetanes 3 and 4, potential
Transient Absorption Spectroscopy. Nonlinear Eyring plots energy surfaces (PES) around the dihedral angle θ (deg) were
(Figure 2b and 3b) clearly suggest that oxetanes 3 and 4 are calculated for both T-DR3a and T-DR4a at the UB3LYP/6-
formed in the reaction of two or more steps.11 To get information 31G* level of theory20 with GAUSSIAN 9821 suite of programs
of the intermediates in the PB reactions, transient absorption (Figure 5). Among the possible conformers of T-C1a-C3a in
spectra were measured for the reactions with 1b on a picosecond T-DR3a and T-C1a′-C3a′ in T-DR4a, only four conformers,
time scale (λexc ) 355 nm, 20 ps pulse width). The diphenyl inside-gauche conformers T-C1a and T-C1a′ (θ ) ca. 60°) and
chromophore was expected to make it possible to detect transient outside-anti conformers T-C2a and T-C2a′ (θ ) ca. 180°), were
species in visible regions. In the absence of a furan derivative, calculated to be energy minimum structures. The outside-gauche
the triplet state of benzophenone 1b* (τ > 1 µs,12 λmax ) 525 conformers, T-C3a and T-C3a′ (θ ) ca. 300°), did not exist as
nm13) was observed within 20 ps after the laser excitation in equilibrium structures. The energetic preference of the inside-
degassed acetonitrile (Figure 4a). As shown in Figure 4b, upon gauche and outside-anti conformers is reasonably explained by
the addition of 5 M acetonitrile solution of 2-methylfuran (2a),
(15) For unusually long-lived singlet diradicals, see: (a) Adam, W.; Borden,
the transient species 1b* was quenched within 500 ps to generate W. T.; Burda, C.; Foster, H.; Heidenfelder, T.; Jeubes, M.; Hrovat, D. A.;
new species with an absorption maximum (λmax) at 531 nm. Kita, F.; Lewis, S. B.; Scheutzow, D.; Wirz, J. J. Am. Chem. Soc. 1998,
120, 593. (b) Abe, M.; Adam, W.; Heidenfelder, T.; Nau, W. M.; Zhang,
The transient species decayed with almost first-order kinetics X. J. Am. Chem. Soc. 2000, 122, 2019. (c) Abe, M.; Adam, W.; Hara, M.;
(τ ) ca. 4.5 ns, kd ) 2.1 × 108 s-1) at 25 °C. The absorption Hattori, M.; Majima, T.; Nojima, M.; Tachibana, K.; Tojo, S. J. Am. Chem.
Soc. 2002, 124, 6540.
spectrum in the visible region of the new species is similar to (16) Giese, B.; Wetstein, P.; Stahelin, C.; Barnosa, F.; Neuburger, M.; Zehnder,
that of the triplet 1,1-diphenyl-2-oxabutane-1,4-diyl DR (λmax M.; Wessig, P. Angew. Chem., Int. Ed. 1999, 38, 2586.
(17) Scaiano, J. C. Tetrahedron 1982, 38, 819.
) 535 nm, τ ) 1.5 ns in acetonitrile) reported in the (18) (a) Wagner, P. J. Acc. Chem. Res. 1989, 22, 83. (b) Zand, A.; Park, B.-S.;
photoreaction of 1,4-dioxene with benzophenone.14 Therefore, Wagner, P. J. J. Org. Chem. 1997, 62, 2326.
(19) (a) Griesbeck, A. G.; Stadtmüller, S. J. Am. Chem. Soc. 1990, 112, 1281.
(b) Griesbeck, A. G.; Stadtmüller, S. J. Am. Chem. Soc. 1991, 113, 6923.
(12) Poter, G.; Wilkinson, F. Trans. Faraday Soc. 1962, 58, 1686. (c) Griesbeck, A. G.; Mauder, H.; Stadtmüller, S. Acc. Chem. Res. 1994,
(13) Green, B. I.; Hockstrasser, R. M.; Weisman, R. J. Chem. Phys. 1979, 70, 27, 70.
1247. (20) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (b) Lee, C.; Yang, W.;
(14) (a) Freilich, S. C.; Peters, K. S. J. Am. Chem. Soc. 1981, 103, 6255. (b) Parr, R. G. Phys. ReV. B 1988, 37, 785. (c) Hariharan, P. C.; Pople, J. A.
Freilich, S. C.; Peters, K. S. J. Am. Chem. Soc. 1985, 107, 3819. Theor. Chim. Acta 1973, 28, 212.

2840 J. AM. CHEM. SOC. 9 VOL. 126, NO. 9, 2004

Paternò−Büchi Reaction of Furan Derivatives ARTICLES

Figure 5. PES analyses around the dihedral angle (θ°) of diradicals (a) T-DR3aa, (b) T-DR4aa, (c) T-DR3ab, and (d) T-DR4ab. The energies, ∆Erel in
kcal/mol, were relative to the most stable conformer.

Scheme 2. Mechanism for the Selective Formation of exo-Oxetanes 3a and 4a in the Photoreaction of Furans 2a,b with Benzaldehyde (1a)

stereoelectronic effect, i.e., gauche effect,22 on the conforma- of T-C3a and T-C3a′ (Scheme 2). The phenyl ring in the
tional stability of the ketals. Thus, the orbital overlap of n orbital optimized structures of intermediary triplet diradicals T-DR3a
of oxygen with σ* orbital of the C-O bond in the furan ring is and T-DR4a is located to be perpendicular to the furan ring, as
quite effective in the four conformers, T-C1a, T-C2a, T-C1a′, exemplified for T-C1aa (Scheme 3). The energetic preference
and T-C2a′. However, the overlap is very poor in the conformers of the perpendicular orientation is reasonably explained by the

J. AM. CHEM. SOC. 9 VOL. 126, NO. 9, 2004 2841

ARTICLES Abe et al.

Scheme 3. Selective Formation of exo-Oxetane 3aa by the ISC processes from the triplet to the singlet diradicals, which
Outward Rotation of the Phenyl Ring in the Intermediary Triplet
Diradical T-C1aa
are spin-forbidden process,25 are estimated to be around 108 s-1,
as observed for diradicals T-DR3b and T-DR4b (R ) Ph)
(Figure 4). Thus, the conformational changes between the inside-
gauche and outside-anti conformers are expected to be faster
than the ISC processes. According to the Curtin-Hammett
principle,26 the ratios of the productive conformers S-C1a and
S-C1a′ are determined not only by the populations of T-C1a
and T-C1a′ but also by the relative rate constant of the ISC
processes, kI1a/kI2a and kI1a′/kI2a′. Since in general the rate constant
of ISC processes are insensitive to temperature,18a,27 the ratios
of kI1a/kI2a and kI1a′/kI2a′ would be independent of the reaction
temperature.28 When the difference between the two ratios is
orbital interaction between 2p AO of oxygen and the radical negligible (kI1a/kI2a ≈ kI1a′/kI2a′), the regioselectivity (3a/4a) can
p-orbital of the benzyl radical part.23 According to the memory be simply determined by the difference of the rate constants
effect, the inside-gauche conformers, T-C1a and T-C1a′, are between the initial double-bond selection (kD3a versus kD4a) and
expected to give exo-3a and exo-4a by the outward rotation of the equilibrium constants (Ka and Ka′) (eq 1). The equation is
the phenyl group; see the curved arrow in T-C1a and T-C1a′ applied under the condition of k3a/k1a ) k4a/k1a′. When the ratio
(Schemes 2 and 3).24 The outside-gauche conformers T-C3a of the two rate constants is largely different, the regioselectivity
and T-C3a′ may afford endo-isomers endo-3a and endo-4a by is determined by eq 2. Even under such conditions, the Eyring
the inward rotation of the phenyl group. The anti-conformers plots should be linear as long as the conformational changes
T-C2a and T-C2a′ go back to the starting compounds 1a and are faster than the ISC processes (kcon > kI). This mechanism
2 after the ISC process to the singlet states. The PES analyses rationalizes the linear Eyring plots observed in the PB reaction
(Figure 5) clearly suggest that among the productive conformers of benzaldehyde, despite going through the intermediary triplet
only inside-gauche conformers T-C1a and T-C1a′ exist as diradicals T-DR3a and T-TD4a.
energy minimum structures. Thus, the exclusive formation of
the exo-oxetanes is expected in the PB reactions, and the kcon > kI
stereoselectivity should be temperature-independent. In fact, we Ka
did observe the highly exo-selective formation of the oxetanes
kD3a Ka + 1
3 and 4, and the stereoselectivities were temperature-indepen- 3a/4a ) × (1)
dent. kD4a K′a
As for regioselectivity (3 versus 4), the Eyring plots of the K′a + 1
reaction with 1a were linear (Figures 2a and 3a). There are four
factors that determine the regioselectivity (Scheme 2): (i) Initial Ka k3a
double bond selection by triplet benzaldehyde, kD3a versus kD4a, kD3a Ka + 1 k3a + k1a
(ii) equilibrium constants, Ka ) [T-C1a]/[T-C2a] versus Ka′ ) 3a/4a ) × × (2)
kD4a K′a k4a
[T-C1a′]/[T-C2a′], (iii) relative rate constant of the ISC
K′a + 1 k4a + k1a′
processes in the triplet 1,4-diradicals, kI1a/kI2a versus kI1a′/kI2a′,
and (iv) relative rate constant of the bond-forming and bond-
breaking step from the singlet 1,4-diradicals S-C1a and S-C1a′, Nonlinear Eyring Plots in the PB Reaction of Benzophe-
k3a/k1a versus k4a/k1a′. As shown in Figure 5, the energy barriers none (1b). Scharf and Hoffmann first reported nonlinear Eyring
on the triplet energy surfaces between the inside-gauche (T- plots in enantioselectivity in PB reactions.29 Griesbeck and co-
C1a and T-C1a′) and outside-anti conformers (T-C2a and workers have also found nonlinear Eyring plots in diastereo-
T-C2a′) are calculated to be around 3 kcal/mol. Thus, the selectivity of PB reactions of aliphatic aldehydes. The nonlin-
conformational changes between the two conformers are earity found by Griesbeck was reasonably explained by the
expected to be very fast in the temperature range under temperature-dependent change of spin state of excited aldehydes
investigations, kcon > 10∼9-10 s-1. The rate constant (kI) of the (singlet versus triplet) that react with alkenes.28,30 However, the
mechanism of nonlinear Eyring plots is still unclear in PB
(21) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. reactions initiated by pure triplet state of carbonyl compounds.
A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.; Stratmann, In the present study, we have found nonlinear Eyring plots of
R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin,
K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi,
R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.; (25) (a) Johnston, L. J.; Scaiano, J. C. Chem. ReV. 1989, 89, 521. (b) Adam,
Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.; W.; Grabowski, S.; Wilson, R. M. Acc. Chem. Res. 1990, 23, 165.
Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz, (26) (a) Curtin, D. Y. Rec. Chem. Prog. 1954, 15, 111. (b) Seeman, J. I. Chem.
J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; ReV. 1983, 83, 83.
Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, (27) (a) Turro, N. J.; Renner, C. A.; Waddell, W. H.; Katz, T. J. J. Am. Chem.
C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.; Soc. 1976, 98, 4320. (b) Chang, M. H.; Dougherty, D. A. J. Am. Chem.
Johnson, B. G.; Chen, W.; Wong, M. W.; Andres, J. L.; Head-Gordon, Soc. 1982, 104, 2333. (c) Jain, R.; McElwee-White, L.; Dougherty, D. A.
M.; Replogle, E. S.; Pople, J. A. Gaussian 98, revision A 11.3; Gaussian, J. Am. Chem. Soc. 1988, 110, 552.
Inc.: Pittsburgh, PA, 1998. (28) Griesbeck and co-workers have recently reported a temperature-independent
(22) The Anomeric Effect and Associated Stereoelectronic Effects; Thatcher, G. diastereoselectivity: Griesbeck, A. G.; Bondock, S.; Gudipati, M. S. Angew.
R. J., Ed.; ACS Symposium Series 539; American Chemical Society: Chem., Int. Ed. 2001, 40, 4684.
Washington, DC, 1993. (29) Buschmann, H.; Scharf, H.-D.; Hoffmann, N.; Plath, M. W.; Runsink, J. J.
(23) Palmer, I. J.; Ragazos, I. N.; Bernardi, F.; Olivucci, M.; Robb, M. A. J. Am. Chem. Soc. 1989, 111, 5367.
Am. Chem. Soc. 1994, 116, 2121. (30) Griesbeck, A. G.; Fiege, M.; Bondock, S.; Gudipati, M. S. Org. Lett. 2000,
(24) Kutateladze, A. G. J. Am. Chem. Soc. 2001, 123, 9279. 2, 3623.

2842 J. AM. CHEM. SOC. 9 VOL. 126, NO. 9, 2004

Paternò−Büchi Reaction of Furan Derivatives ARTICLES

the regioselectivities (3ba/4ba and 3bb/4bb) in the pure triplet When the ratio of k3b/k1b is equal to k4b/k1b′, the regioselectivity
PB reaction with 1b (Figures 2b and 3b). Furthermore, the can be simply determined by the initial double-bond selection
temperature effect on the regioselectivities (3ba/4ba versus 3bb/ (kD3b versus kD4b) by triplet benzophenone (eq 4). If the
4bb) largely depends on the position of the methyl substituent conformational changes are faster than the ISC processes at high
on the furan ring. Thus, the product ratio of 3ba/4ba in the temperatures, kI < kcon (mechanism II), the regioselectivity may
reaction of 2-methylfuran (2a) nonlinearly decreased with the be determined by eq 5. Equation 6 can be derived from eq 5
increase of the reaction temperature (Figure 2b). Alternatively, under the condition that the relative rate constant (kD3b/kD4b) is
in the reaction of 3-methylfuran (2b) the product ratio of 3bb/ proportional to the total double-bond selection (kD3b + kD3b′)/
4bb nonlinearly increased with the increase of the reaction (kD4b + kDb′). When the switching between the two mechanisms
temperature (Figure 3b). (mechanisms I and II) occurs in the temperature range under
The proposed mechanism for the formation of oxetanes 3b investigation, the Eyring plots of the regioselectivity (3b/4b)
and 4b is shown in Scheme 4. The transient absorption spectra should be nonlinear.
mentioned before (Figure 4) strongly support the generation of To assess the temperature-dependent change of the mecha-
triplet diradicals T-DR3b and T-DR4b in the photochemical nism, the potential energy surfaces around the dihedral angle θ
processes. As discussed in the regioselectivity (3a versus 4a) (deg) of the triplet diradicals, T-DR3b and T-DR4b, were
of the reaction with 1a (Scheme 2), if the conformational change calculated at the UB3LYP/3-21G level of theory (Figure 6).
between the productive conformer (T-C1b or T-C1b′) and the Again, the outside-gauche conformers (T-C3b and T-C3b′, θ
nonproductive conformer (T-C2b or T-C2b′) is faster than the ) ca. 300°) were not calculated to be energy minimum
ISC process to the singlet diradicals, the Eyring plots of the structures, as predicted for T-DR3a and T-DR4a (Scheme 3).
regioselectivities (3ba/4ba and 3bb/4bb) should be linear. The energy barriers (∆Erel ) ca. 4-8 kcal/mol) between the
Although temperature effect on the ISC process (kI) is possible conformers in the triplet diradicals were calculated to
small,17,27 the rate constants (kcon) of conformational changes be higher than those of T-DR3a and T-DR4a. The second
are temperature-sensitive. The regioselectivity (3b/4b) can be phenyl substituent derived from benzophenone is responsible
determined by eq 3 under the conditions that the ISC processes for the increase in the energy barriers. The rate constants (kcon)
from the triplet diradicals T-DR3b and T-DR4b to the singlet for the conformational changes between the equilibrium struc-
states overcome the conformational changes of the triplet tures can be estimated to be 108 to 105 s-1 at -70 °C, 109 to
diradicals at low temperatures, kI > kcon (mechanism I). 106 s-1 at 25 °C, and 1010 to 107 s-1 at +50 °C. From the
Mechanism I: At low temperatures: lifetime (τ) of the triplet diradicals T-DR3b and T-DR4b
determined by the transient absorption spectra (Figure 4), the
kI > kcon ISC rate constant (kI) of the triplet diradicals (τ ) 1/kI) is

( )
kD3b k3b k4b determined to be ca. 108 s-1 at 25 °C. Thus, it is possible that
3b/4b ) × / (3) the ISC processes in the triplet diradicals overcome the
kD4b k3b + k1b k4b + k1b′
conformational change between the equilibrium structures at
if low temperatures, kI > kcon (mechanism I). At high temperatures,
mechanism II is operative, since the conformational changes
k3b/k1b ) k4b/k1b′ are expected to be faster than the ISC processes (kI < kcon).
To assess our hypothetical mechanism, the theoretical ratios
then of ln(3b/4b) calculated by eq 4 (green line, mechanism I) and
eq 6 (blue line, mechanism II) were plotted against the inverse
3b/4b ) kD3b/kD4b (4)
of reaction temperatures (Figure 7). The activation parameters
Mechanism II: At high temperatures: ∆∆Hq and ∆∆Sq (Figure 7a,d) for the double-bond selection,
kD3b/kD4b, were calculated from the product ratios of 3b/4b
kI < kcon, k3b/k1b ) k4b/k1b′ observed below the temperature of -55 °C for 2a and -50 °C
for 2b. The entropy-controlled selections, i.e., kD3b/kD4b, are
Kb
consistent with the early transition state of bond-forming steps
3b kD3b + kD3b′ Kb + 1 of the energy-rich nπ*-1b with furans (Figure 7a,d).31 The
) × (5)
4b kD4b + kD4b′ K′b equilibrium constants Kb in T-DR3b and Kb′ in T-DR4b were
K′b + 1 calculated from the computed energy differences, ∆G ) ∆H
- T∆S, between the equilibrium conformers. As for the
if diradicals formed in the reaction with 2-methylfuran (2a), the
population of the productive conformer T-C1ba′ in T-DR4ba
kD3b kD3b + kD3b′ is calculated to be higher than that of the productive conformer
) T-C1ba in T-DR3ba, i.e., Kb′ > Kb > 1 (Figure 6a,b, Scheme
kD4b kD4b + kD4b′
5). Thus, under the conditions of mechanism II (kcon > kI), higher
then substituted oxetane 4ba is expected to form more than the
chemical yield anticipated from the initial double-bond selection
Kb (T-DR3ba/T-DR4ba, mechanism I). Consistent with this pre-
3b k D3b Kb+1 diction, the Eyring plot (Figure 7b, red line) that was experi-
) × (6)
4b kD4b K′b (31) (a) Skell, P. S.; Cholod, M. S. J. Am. Chem. Soc. 1969, 91, 7131. (b) Giese,
B. Angew. Chem., Int. Ed. 1977, 16, 125. (c) Bach, T.; Jodicke, K.; Kather,
K′b + 1 K.; Frohlich, R. J. Am. Chem. Soc. 1997, 119, 2437.

J. AM. CHEM. SOC. 9 VOL. 126, NO. 9, 2004 2843

ARTICLES Abe et al.

Scheme 4. Mechanism for the Formation of Oxetanes 3 and 4 in the PB Reaction of Furans 2 with Benzophenone (1b)

mentally obtained departs from the green line (Figure 7a, The contrastive difference between the population of produc-
mechanism I) at ca. -55 °C with the increase in the reaction tive conformers T-C2ba′ and T-C2bb′ in T-DR4b can be
temperature and approaches to the blue line (Figure 7c, reasonably explained by steric effect (Scheme 5). Since the steric
mechanism II). As for the regioselectivity in the reaction with repulsion between the methyl and the diphenylmethyl group is
3-methylfuran (2b) (Figure 7d-f), since the population of the significant in T-C2ba′, the conformer T-C1ba′ is more stable
productive conformer T-C1bb′ in T-DR4bb is calculated to in T-DR4ba (∆Erel ) -3.1 kcal/mol). In contrast, the conformer
be lower than that of the conformer T-C1bb in T-DR3bb, Kb
T-C2bb′ is a more favorable structure than T-C1bb′ (∆Erel )
> 1, Kb′ < 1 (Scheme 4), at high temperature lower substituted
+1.1 kcal/mol), since the steric repulsion between the methyl
oxetane 3bb is expected to form more than the chemical yield
anticipated from mechanism I (T-DR3bb/T-DR4bb). Thus, the and diphenylmethyl group is severer than that between the
Eyring plot (Figure 7e, red line) departs from the green line hydrogen and the diphenylmethyl group. The equilibrium
(Figure 7d) with increase in the temperature, and approaches constant (Kb) between T-C1b and T-C2b (∆Erel ) -1.0 kcal/
the blue line (Figure 7f) plotted according to eq 6 (mechanism mol) is not dependent on the position of the methyl group, since
II). The decrease in the quantum yield at high temperature, Φ3+4 the methyl substituent is located far from the diphenylmethyl
) 0.39 at -77 °C f 0.06 at +56 °C, supports the mechanism. group.

Figure 6. PES analyses around the dihedral angle (θ°) in diradicals (a) T-DR3ba, (b) T-DR4ba, (c) T-DR3bb, and (d) T-DR4bb. The total electronic
energies, ∆Erel in kcal/mol, were relative to the most stable conformer.
2844 J. AM. CHEM. SOC. 9 VOL. 126, NO. 9, 2004
Paternò−Büchi Reaction of Furan Derivatives ARTICLES

Figure 7. Theoretical and experimental plots of ln(3/4) against 1/T in the PB reaction of benzophenone (1b) with (i) 2-methylfurane (2a) and (ii) 3-methylfuran
(2b).

Scheme 5. Steric Effect on the Equilibrium Constants Kb and Kb′ none is reasonably explained by the switching between mech-
in Triplet 1,4-Diradicals T-DR4ba and T-DR-4bb
anisms I and II in the temperature range under investigations.
These fundamental findings on the reaction mechanism provide
valuable information to predict the selectivities in other Pa-
ternò-Büchi reactions.
Experimental Section
General Procedure for the Photoreactions of Aromatic Carbonyl
Compounds with Furan Derivatives. A degassed toluene solution of
freshly distilled furan 2 (0.5 M, 5 mmol) and aromatic carbonyl
compound 1 (0.05 M, 0.5 mmol) was irradiated for 30 min with a high-
pressure Hg lamp through a Pyrex filter at various temperatures. After
the solvent was removed under the reduced pressure by using a vacuum
pump (0.2 mmHg, <10 °C), the photolyzate was directly analyzed by
1H NMR (270 and 600 MHz) spectroscopy. The spectra revealed the

quantitative formation of the bicyclic oxetanes 3 and 4 together with

excess amount of aromatic carbonyls. The product ratios were
determined by the comparison of the peak areas, error <3%. The
photoproducts 3 and 4 were isolated by using alumina column
chromatography. Some decomposition of the oxetanes was observed
during the column chromatography. The oxetanes were decomposed
to unidentified products on silica gel column chromatography. Although
bicyclic oxetanes 3aa,3a 4aa,3a 4ba,3b,d 3bb3b,d are known compound,
the high-resolution spectra were not available. 1H NMR spectra (270
or 600 MHz) for the compounds are listed below, together with the
new compounds 3ab, 4ab, 3ba, and 4bb assigned in this study.
3-Methyl-6-phenyl-2,7-dioxabicyclo[3.2.0]hept-3ene (exo-3aa).3a
1H NMR (600 MHz, CDCl ): δ 2.00 (s, 3 H), 3.62-3.65 (m, 1 H),
3
In summary, we found unusual temperature32 and substituent 5.08-5.10 (m, 1 H), 5.54 (d, J ) 3.0 Hz, 1 H), 6.50 (d, J ) 4.6 Hz,
effects on the stereo- and regioselectivity in the PB reaction of 1 H), 7.24-7.43 (m, 5 H).
furan derivatives. The experimental and computational results 1-Methyl-6-phenyl-2,7-dioxabicyclo[3.2.0]hept-3ene (exo-4aa).3a
have provided convincing evidence regarding the mechanism 1H NMR (600 MHz, CDCl ): δ 1.73 (s, 3 H), 3.42-3.43 (m, 1 H),
3
of these reactions, which has been debated until now. The 5.38-5.40 (m, 1 H), 5.51 (d, J ) 3.6 Hz, 1 H), 6.63 (d, J ) 3.0 Hz,
conformational distributions of the intermediary triplet 1,4- 1 H), 7.21-7.42 (m, 5 H).
diradicals, which are controlled by the gauche effect and the 3-Methyl-6,6-diphenyl-2,7-dioxabicyclo[3.2.0]hept-3ene (3ba). 1H
steric effect, play a crucial role to determine the stereo- and NMR (270 MHz, CDCl3): δ 2.29 (s, 3 H), 4.30-4.32 (m, 1 H), 4.51-
regioselectivity. Thus, the energetic preference of the inside- 4.52 (m, 1 H), 6.22 (d, J ) 4.6 Hz, 1 H), 7.11-7.76 (m, 10 H); 13C
NMR (67.8 MHz, CDCl3): δ 13.7, 57.1, 90.7, 97.4, 105.2, 125.3, 125.4,
gauche conformers in intermediary triplet diradicals was pro-
126.7, 127.1, 127.8, 143.0, 145.3; IR (liquid film): 3106-2926, 1612,
posed to be responsible for the exclusive formation of exo- 1491, 1452, 1396, 1136 cm-1; Anal. Calcd for C18H16O2: C, 81.79; H,
oxetanes 3a and 4a in the reactions with benzaldehyde. The 6.10. Found: C, 81.60; H, 6.40.
nonlinear Eyring plots observed in the reactions with benzophe- 1-Methyl-6,6-diphenyl-2,7-dioxabicyclo[3.2.0]hept-3ene (4ba).3b,d
1H NMR (270 MHz, CDCl ): δ 1.57 (s, 3 H), 4.16 (d, J ) 1.0 and 3.0
3
(32) Unusual temperature effect on the diastereoselectivity was observed in the
PB reaction of cyclooctene with ketones. See: Adam, W.; Stegmann, V. Hz, 1 H), 4.90 (t, J ) 3.0 Hz, 1 H), 6.39 (dd, J ) 1.0 and 3.0 Hz, 1
R. J. Am. Chem. Soc. 2002, 124, 3600. H), 7.15-7.51 (m, 10 H).

J. AM. CHEM. SOC. 9 VOL. 126, NO. 9, 2004 2845

ARTICLES Abe et al.

4-Methyl-6-phenyl-2,7-dioxabicyclo[3.2.0]hept-3ene (exo-3ab). 1H sample remained above 1.5 during the irradiation, and thus, no
NMR (270 MHz, CDCl3): δ 1.79 (s, 3 H), 3.47-3.49 (m, 1 H), 5.57 correction for transmitted light was necessary. The intensity of the light
(d, J ) 3.0 Hz, 1 H), 6.37 (s, 1 H), 6.48 (d, J ) 4.3 Hz, 1 H), 7.31- source was determined on the basis of the known quantum yield for
7.46 (m, 5 H); 13C NMR (67.8 MHz, CDCl3): δ 9.5, 55.9, 91.1, 108.6, the Norrish type-II photochemical decomposition of valerophenone
113.9, 125.0, 128.1, 128.8, 141.3, 141.9; IR (liquid film, with 4ab): (ΦPhCOMe ) 0.33 at 313 nm).33 The obtained values in the text are the
3111-2912, 1634, 1453, 1387 cm-1; Anal. Calcd for C12H12O2 (with average after three trials, error <10%.
4ab): C, 76.57; H, 6.43. Found: C, 76.22; H, 6.45. Transient Absorption Spectroscopic Measurement. Picosecond34
5-Methyl-6-phenyl-2,7-dioxabicyclo[3.2.0]hept-3ene (exo-4ab). 1H time-resolved difference spectra were obtained by using the third
NMR (270 MHz, CDCl3): δ 0.81 (s, 3 H), 5.37 (d, J ) 2.7 Hz, 1 H), harmonic of a mode-locked Nd3+:YAG laser (Continuum PY61C-10,
5.66 (s, 1 H), 6.11 (s, 1 H), 6.63 (d, J ) 2.7 Hz, 1 H), 7.31-7.46 (m, λexc ) 355 nm) for excitation. The transient absorption spectra in the
5 H); 13C NMR (67.8 MHz, CDCl3): δ 15.8, 56.1, 94.6, 110.9, 111.6, time range from 20 ps to 6 ns were acquired by using continuum pulses
125.2, 128.7, 128.9, 139.2, 146.8. with delays of 20-6000 ps. The latter was generated by focusing the
4-Methyl-6,6-diphenyl-2,7-dioxabicyclo[3.2.0]hept-3ene (3bb). 1H fundamental laser pulse into a flowing H2O/D2O mixture (1:1 by
NMR (270 MHz, CDCl3): δ 1.95 (s, 3 H), 4.10 (d, J ) 4.3 Hz, 1 H), volume).
6.05 (d, J ) 4.3 Hz, 1 H), 6.12 (s, 1 H), 7.21-7.87 (m, 10 H); 13C
NMR (67.8 MHz, CDCl3): δ 18.6, 61.1, 96.6, 108.3, 109.9, 125.3, Acknowledgment. We thank Mrs. T. Muneishi at the
125.9, 127.2, 127.7, 128.3, 128.4, 142.9, 143.8, 145.4; IR (KBr, with Analytical Center of Faculty of Engineering, Osaka University,
4bb): 2867-3225, 1657, 1560, 1396 cm-1; Anal. Calcd for for measuring the 600 MHz NMR spectra.
C18H16O2: C, 81.79; H, 6.10. Found: C, 81.60; H, 6.08.
5-Methyl-6,6-diphenyl-2,7-dioxabicyclo[3.2.0]hept-3ene (4bb).3b,d Supporting Information Available: Computational details
1H NMR (270 MHz, CDCl ): δ 1.10 (s, 3 H), 5.01 (d, J ) 2.7 Hz, 1
3
(PDF). This material is available free of charge via the Internet
H), 5.96 (s, 1 H), 6.15 (d, J ) 2.7 Hz, 1 H), 7.21-7.87 (m, 10 H). at http://pubs.acs.org.
Determination of Quantum Yields of Oxetane Formations. The
JA039491O
quantum yields of the oxetane formations in the photoreaction of
aromatic carbonyl with furan derivatives were measured by Xenon lamp (33) Wagner, P. J.; Kemppainen, A. E. J. Am. Chem. Soc. 1968, 90, 5896.
focused at 313 nm. The absorbance at 313 nm of a solution of the (34) Ohno, T.; Nozaki, K.; Haga, M. Inorg. Chem. 1992, 31, 548.

2846 J. AM. CHEM. SOC. 9 VOL. 126, NO. 9, 2004