Stereoselectivity of Triplet Photocycloadditions:1 Diene-Carbonyl Reactions and Solvent Effects

Article abstract of DOI:10.1021/jo971767l

The diastereoselectivity of the photocycloaddition of benzaldehyde to furan was determined (exo/ endo = 212:1) and compared with the reaction of other carbonyl compounds and carbonyl analogues. These results were compared with Paterno-Buechi reactions of cycloalkenes and cyclic enol ethers. An increase in steric demand of the α-substituent in benzoyl compounds led to a change in exo/ endo-selectivity for furan cycloadditions that was not observed for cycloalkenes or cyclic enol ethers. Different ISC-reactive conformers with enhanced spin-orbit coupling are postulated as a reasonable explanation for the stereoselectivities observed. Additionally, solvent effects were studied, demonstrating the influence of photoinduced electron-transfer steps on the regio- and diastereoselectivity of Paterno-Buechi reactions with 2,3-dihydrofuran in polar solvents. Two bicyclic oxetanes (8 and 10) were characterized by X-ray structure analysis.

Full text of DOI:10.1021/jo971767l

J . Org. Chem. 1998, 63, 3847-3854
3847
Ster eoselectivity of Tr ip let P h otocycloa d d ition s:1 Dien e-Ca r bon yl
Rea ction s a n d Solven t Effects
Axel G. Griesbeck,* Stefan Buhr, Maren Fiege, Hans Schmickler, and J ohann Lex
Institute of Organic Chemistry, University of Cologne, Greinstr. 4, D-50939 Ko¨ln, Germany
Received September 23, 1997
The diastereoselectivity of the photocycloaddition of benzaldehyde to furan was determined (exo/
endo ) 212:1) and compared with the reaction of other carbonyl compounds and carbonyl analogues.
These results were compared with Paterno`-Bu¨chi reactions of cycloalkenes and cyclic enol ethers.
An increase in steric demand of the R-substituent in benzoyl compounds led to a change in exo/
endo-selectivity for furan cycloadditions that was not observed for cycloalkenes or cyclic enol ethers.
Different ISC-reactive conformers with enhanced spin-orbit coupling are postulated as a reasonable
explanation for the stereoselectivities observed. Additionally, solvent effects were studied,
demonstrating the influence of photoinduced electron-transfer steps on the regio- and diastereo-
selectivity of Paterno`-Bu¨chi reactions with 2,3-dihydrofuran in polar solvents. Two bicyclic
oxetanes (8 and 10) were characterized by X-ray structure analysis.
In tr od u ction
In recent publications, we have tried to combine these
theoretical models with experimental results using the
following concept: if the second bond-forming step in
triplet [2 + 2] cycloaddition reactions is coupled with a
stereochemical probe, the biradical geometry most favor-
able for rapid ISC should be reflected in the product
configuration.7 We have used intermolecular Paterno`-
Bu¨chi reactions8 and investigated a series of cycloalkenes
reacting with prochiral triplet excited carbonyls. In these
cases, triplet 2-oxatetramethylenes are formed during the
course of the reactionsspecies that are known to have
lifetimes of 1-5 ns,9 i.e., relatively short-lived in com-
parison with the 1,4-biradicals without oxygen contribu-
tion.10 For these systems, the degree of stereoselectivity
was dependent on the size of the R-substituent of the
carbonyl component, on the substituent pattern, as well
as on the flexibility of the alkene addend. In most cases,
selectivities not higher than 90:10 resulted. A remark-
able exception was found for furan11 (and also for other
heteroaromatic substrates such as thiophenes, pyrroles,
imidazoles, and pyrazoles)12 and furan derivatives.13 The
photocycloaddition of these cyclo-1,3-dienes with aromatic
Intermolecular photocycloaddition reactions with one
triplet excited component are mostly conducted via triplet
biradical intermediates.1 These species are separated
from their closed-shell successors by an activation barrier
due to the spin-forbidden intersystem crossing (ISC)
process. Thus, the triplet 1,4-biradicals that are formed
in [2 + 2] cycloaddition reactions have relatively long
lifetimes (1 ns to 1 µs).2 Spin-orbit coupling (SOC)
controls the rate of ISC for these type of intermediates,
i.e., strong SOC enhances the ISC rate and lowers the
lifetime of the triplet biradicals. In recent publications
by Michl3 and Adam and co-workers,4 the long-known and
successfully applied Salem-Rowland rules5 for estimat-
ing the magnitude of SOC were modified due to new
experimental and theoretical results. The spatial orien-
tation of the two singly occupied orbitals has been
determined to be highly important for the biradical
lifetimes, whereas the through-space distance between
the radical centers3 plays a subordinate role and the
degree of ionic contribution in the corresponding singlet
state4 often seems to be overestimated. This clearly
indicates that for flexible 1,4-triplet biradicals not only
one conformational arrangement is responsible for fa-
cilitating ISC, but many. After transition from the triplet
to the singlet potential energy surface, immediate product
formation is expected. Thus, the ISC is expected to be
concerted with the formation of a new bond or the
cleavage of the primarily formed single bond. This
picture resembles the concept by Turro of “tight” and
“loose” geometries of biradicals (originally used for sin-
glets and triplets), which are precursors to cycloaddition
and cleavage products, respectively.6
(7) (a) Buhr, S.; Griesbeck, A. G.; Lex, J .; Mattay, J .; Schro¨er, J .
Tetrahedron Lett. 1996, 37, 1195. (b) Griesbeck, A. G.; Mauder, H.;
Stadtmu¨ller, S. Acc. Chem. Res. 1994, 27, 70. (c) Griesbeck, A. G.;
Stadtmu¨ller, S.; Bringmann, G.; Busse, H.; Buddrus, J . Chem. Ber.
1992, 125, 933. (d) Griesbeck, A. G.; Stadtmu¨ller, S. J . Am. Chem. Soc.
1991, 113, 6923. (e) Griesbeck, A. G.; Stadtmu¨ller, S. J . Am. Chem.
Soc. 1990, 112, 1281. (f) Griesbeck, A. G.; Mauder, H.; Peters, K.;
Peters, E.-M.; von Schnering, H. G. Chem. Ber. 1991, 124, 407. (g)
Griesbeck, A. G.; Stadtmu¨ller, S. Chem. Ber. 1990, 123, 357.
(8) For reviews: (a) Griesbeck, A. G. In Handbook of Organic
Photochemistry and Photobiology; Horspool, W. M., Song, P.-S., Eds.;
CRC Press: Boca Raton, 1995; p 550; (b) p 755. (c) Porco, J . A.;
Schreiber, S. L. In Comprehensive Organic Synthesis; Trost, B. M.,
Fleming, I., Paquette, L. A., Eds.; Plenum Press: New York, 1991;
Vol. 5, p 151. (d) Carless, H. A. J . In Synthetic Organic Photochemistry;
Horspool, W. M., Ed.; Plenum Press: New York, 1984; p 425. (e) J ones,
G., II. In Padwa, A., Ed. Org. Photochem. 1981, 5, 1.
(9) (a) Hu, S.; Neckers, D. C. J . Org. Chem. 1997, 62, 755. (b)
Freilich, S. C.; Peters, K. S. J . Am. Chem. Soc. 1981, 103, 6255. (c)
Freilich, S. C.; Peters, K. S. J . Am. Chem. Soc. 1985, 107, 3819.
(10) (a) Newcomb, M.; J ohnson, C. C.; Manek, M. B.; Varik, T. R. J .
Am. Chem. Soc. 1992, 114, 10915. (b) Rudolph, A.; Weedon A. C. Can.
J . Chem. 1990, 68, 1590. (c) Caldwell, R. A. In Kinetics and Spectros-
copy of Carbenes and Biradicals; Platz, M. S., Ed.; Plenum: New York,
1989; p 77.
(1) Part 8. For part 7 in this series, see ref 7a.
(2) J ohnston, L. J .; Scaiano, J . C. Chem. Rev. 1989, 89, 521.
(3) Michl, J . J . Am. Chem. Soc. 1996, 118, 3568.
(4) Kita, F.; Nau, W. M.; Adam, W.; Wirz, J . J . Am. Chem. Soc. 1995,
117, 8670.
(5) Salem, L.; Rowland, C. Angew. Chem., Int. Ed. Engl. 1972, 11,
92.
(6) Turro, N. J . Modern Molecular Photochemistry; Benjamin/
Cummings Publ. Co.: Menlo Park, 1978; p 434.
S0022-3263(97)01767-2 CCC: $15.00 © 1998 American Chemical Society
Published on Web 05/20/1998

3848 J . Org. Chem., Vol. 63, No. 12, 1998
Griesbeck et al.
1
F igu r e 1. 500 MHz H NMR spectrum of exo- and endo-2.
Ta ble 1. 1H NMR Sh ifts a n d Ca lcu la ted Cou p lin g
and aliphatic aldehydes proceeds with unusually high
stereoselectivity. Exact values, however, cannot be found
in the literature. To understand this effect, we have
investigated solvent and substituent effects as well as
diene modifications.
Con sta n ts for exo- a n d en d o-2
Resu lts a n d Discu ssion
Solven t Effects a n d Regio- a n d Ster eoselectivity
for th e P a r en t System s: F u r a n /Ben za ld eh yd e a n d
2,3-Dih ydr ofu r an /Ben zaldeh yde. For comparison with
the cycloadditions of monoalkenes, we tried to determine
the exact degree of stereo- and regioselectivity of the
benzaldehyde cycloaddition with furan (1): in the 500
δa (ppm)
1-H
3-H
4-H
5-H
6-H
exo-2
endo-2
6.54
6.36
6.71
6.52
5.47
4.68
3.65
4.21
5.57
6.07
J b (Hz)
1
MHz H NMR spectrum of the crude reaction mixture,
1-3 3-4 3-5 1-6 1-5 4-6 5-6 4-5
exo-2
0.78 2.92 1.22 0.74 4.37 0.00 3.12 2.99
(11) (a) Ogata, M.; Watanabe, H.; Kano, H. Tetrahedron Lett. 1967,
533. (b) Toki, S.; Shima, K.; Sakurai, H. Bull. Chem. Soc. J pn. 1965,
38, 760. (c) Toki, S.; Sakurai, H. Tetrahedron Lett. 1967, 4119. (c) Toki,
S.; Sakurai, H. Bull. Chem. Soc. J pn. 1967, 40, 2885. (d) Shima, K.;
Sakurai, H. Bull. Chem. Soc. J pn. 1966, 39, 1806. (e) Zamojski, A.;
Kozluk, T. J . Org. Chem. 1977, 42, 1089. (f) Kitamura, T.; Kawakami,
Y.; Imegawa, T.; Kawanishi, M. Synth. Commun. 1977, 7, 521. (g)
Farid, S.; Hartman, S. E.; DeBoer, C. D. J . Am. Chem. Soc. 1975, 97,
808. (h) Evanega, G. R.; Whipple, E. B. Tetrahedron Lett. 1967, 2163.
(i) Heatchcock, C. H.; Pirrung, M. C.; Montgomery, S. H.; Lampe, J .
Tetrahedron 1981, 37, 4087. (j) J arosz, S.; Zamojski, A. Tetrahedron
1982, 38, 1453. (k) Schreiber, S. L.; Satake, K. J . Am. Chem. Soc. 1983,
105, 6723. (l) Schreiber, S. L.; Satake, K. J . Am. Chem. Soc. 1984, 106,
4186. (m) Schreiber, S. L.; Hoveyda, A. H.; Wu, H.-J . J . Am. Chem.
Soc. 1984, 106, 1148. (n) Schreiber, S. L.; Hoveyda, A. H. J . Am. Chem.
Soc. 1984, 106, 7200. (o) Schreiber, S. L. Science 1985, 227, 857. (p)
Itokawa, H.; Oshima, T.; Yamamura, H.; Tazaki, T.; Matsumoto, H.;
Mihashi, S.; IItaka, Y. Yakugaku Zasshi 1985, 105, 19.
(12) (a) J ones, G., II; Gilow, H. M.; Low, J . J . Org. Chem. 1979, 44,
2949. (b) Matsuura, T.; Banba, A.; Ogura, K. Tetrahedron 1971, 27,
1211. (c) Rivas, C.; Bolivar, R. A. J . Heterocycl. Chem. 1976, 13, 1037.
(d) J ulian, D. R.; Tringham, G. D. J . Chem. Soc., Chem. Commun. 1973,
13. (e) Rivas, C.; Pacheco, D.; Vargas, F.; Ascanio, J . J . Heterocycl.
Chem. 1981, 18, 1065. (f) Nakano, T.; Rodriquez, W.; de Roche, S. Z.;
Larrauri, J . M.; Rivas, C.; Pe´rez, P. J . Heterocycl. Chem. 1980, 17, 1777.
(13) Carless, H. A. J .; Halfhide, A. F. E. J . Chem. Soc., Perkin Trans.
1 1992, 1081. (b) Schreiber, S. L.; Desmaele, D.; Porco, J . A., J r.
Tetrahedron Lett. 1988, 29, 6689. (c) Schreiber, S. L.; Porco, J . A., J r.
J . Org. Chem. 1989, 54, 4721. (d) Sekreta´r, S.; Ruda´, J .; Stibra´nyi, L.
Collect. Czech. Chem. Commun. 1984, 49, 71.
endo-2 0.72 2.88 1.33 0.00 4.20 0.25 5.69 2.92
a
b
500 MHz NMR (CDCl3). Calculated using PANIC.
0.47% of the minor (endo) diastereoisomer 2 was detected
(relative to the exo diastereoisomer), thus resulting in a
212:1 diastereoisomeric ratio. The relevant signals for
determination of this ratio are shown in Figure 1.
Well-separated signals were obtained for 5-H and 4-H
of the major diastereoisomer at 3.65 and 5.47 ppm. The
corresponding signals of the minor (endo-phenyl) dias-
tereoisomer were detected at 4.21 and 4.68 ppm with the
corresponding multiplicities and similar coupling con-
stants. The other proton signals were detected by
TOCSY, and the coupling pattern was successfully
simulated (see Table 1).
No signals could be detected for the two possible
regioisomeric diastereoisomers, indicating a regioselec-
tivity of >600:1. The latter result is in agreement with
the assumption of regiocontrol exhibited by the radical-
stabilizing effects of the ring substituents, i.e., R-alkoxy
radical vs allylic radical (∆ERS ) ca. 10-12 kJ /mol).14 The
pronounced stereoselectivity of 212:1, however, cannot

Stereoselectivity of Triplet Photocycloadditions
J . Org. Chem., Vol. 63, No. 12, 1998 3849
Sch em e 1
be explained by any model published in the literature.
The SOC-geometry model, which we have postulated for
similar cycloaddition reactions with cyclic enol ethers,7e
predicts only moderate diastereoselectivity (ca. 1:10) and
the reverse direction of stereocontrol. For comparison,
the selectivity values for the benzaldehyde addition with
furan (1) and 2,3-dihydrofuran (4) are listed in eq 1. Both
reactions have been conducted in benzene and acetoni-
trile, respectively, as solvents. The furan cycloaddition
proceeded with identical regio- and diastereoselectivity
in both solvents. The high oxidation potential of furan
(Eox. ca. 2.0 V)15 excludes electron transfer pathways even
in highly polar solvents.
F igu r e 2. Correlation between solvent polarity and oxetane
product ratio 6/5.
of benzaldehyde.18 This lack in driving force can be
compensated by increasing the solvent polarity, which
reduces the coulomb term in the Rehm-Weller equa-
tion.19 In acetonitrile (∆GSSIP[CH3CN, ꢀ: 37.5] ) -0.46
eV), 12% of the regioisomer 520 was formed in addition
to 50% of 6, 18% of the carbinol 7, and ca. 20% of 1,2-
diphenylglycol. The diastereoselectivity of the formation
of product 6 (88:12) did not change when switching from
benzene to acetonitrile. The acetal-type regioisomer 5
also was formed with a similar degree of diastereoselec-
tivity (15:85); in this case, however, the exo product was
the preferred stereoisomer. The carbinol 7 was isolated
as a 1:1 mixture of diastereoisomers.
For a mechanism where the bulk solvent polarity
controls the regioselectivity, a dependence between the
regioisomeric ratio and the polar solvent mole fraction
xp is expected. In ideal solvent mixtures described either
by the Onsager function21 or by the Debye function, the
dielectric constant ꢀ shows only little increase up to xp )
0.4 and the “polarity” goes linear with xp. Consequently,
the relative amount of “polar” products (5 and also 7, vide
infra) should also rise linear with increasing solvent
polarity. We have investigated solvent mixtures consist-
ing of benzene and acetonitrile and found a near-linear
dependency between the dielectric constant ꢀ and the
polar solvent mole fraction xp.22 Consequently, when
describing the solvent polarity by the Onsager or Debye
functions, a strong increase in polarity results for the
region of low xp. The ratio of “polar” products against
“nonpolar” products (i.e., 5 and 7 vs 6) behaved likewise
(Figure 2 for the dependence of the 6/5 ratio vs the Debye
function Φ(D)). Already concentrations of acetonitrile in
(1)
When 2,3-dihydrofuran (4) was irradiated in benzene
in the presence of benzaldehyde, a 12:88 mixture of exo/
endo diastereoisomers 6 resulted (Scheme 1).6g The
regioselectivity of this reaction is high, but only in
benzene: only 3% of the acetal-type regioisomer 5 was
formed. Like for furan, photoinduced electron transfer
(PET)16 with formation of solvent-separated ion pairs
(SSIP) is energetically unfeasible in unpolar solvents
such as benzene or cyclohexane. The oxidation poten-
tial17 for 4 is too high to enable exergonic electron transfer
(∆GSSIP[benzene, ꢀ: 2.3] ) +0.62 eV) from the triplet state
(18) E0red.(benzaldehyde) ) -1.80 V vs SCE in DMF: Letaw, H.,
J r.; Gropp, A. H. J . Phys. Chem. 1953, 57, 964. ET ) 3.17 eV: Kearns,
D. R.; Case, W. A. J . Am. Chem. Soc. 1966, 88, 5087.
(19) (a) Weller, A. Z. Phys. Chem. N. F. 1982, 130, 129. (b) Weller,
A. Z. Phys. Chem. N. F. 1982, 133, 93. (c) Rehm, D.; Weller, A. Isr. J .
Chem. 1970, 8, 259.
(14) Leroy, C.; Peeters, D.; Wilante, C. THEOCHEM 1982, 5, 217.
(15) Yoshida, K.; Fueno, T. Bull. Soc. Chim. J pn. 1987, 55, 229.
(16) (a) Pandey, G. Top. Curr. Chem. 1993, 168, 175-221. (b)
Mariano, P. S.; J . L. Stavinoha, J . L. In Synthetic Organic Photochem-
istry, Horspool, W. A., Ed.; Plenum Press: New York, 1984. (c)
Photoinduced Electron Transfer; Fox, M. A., Chanon, M., Eds.;
Elsevier: Amsterdam, 1988. (d) Kavarnos, G. J . Fundamentals of
Photoinduced Electron Transfer; VCH: Weinheim, 1993. (e) Mattay,
J .; Vondenhof, M. In Mattay, J ., Ed. Top. Curr. Chem. 1991, 159.
(17) E0ox.(4) ) 0.97 V vs SCE in CH3CN: Mattay, J . Tetrahedron
1985, 41, 2405.
(20) The exo diastereoisomer of 5 was independently synthesized
from
2 by catalytic reduction with Pd-C in EtOAc, following a
procedure reported in ref 11n.
(21) Onsager, L. J . Am. Chem. Soc. 1936, 58, 1486.
(22) This effect was also reported for other “non-ideal” solvent
mixtures (e.g., n-hexane/propionitrile): Suppan, P. J . Chem. Soc.,
Faraday Trans. 1 1987, 83, 495.

3850 J . Org. Chem., Vol. 63, No. 12, 1998
Sch em e 2. Exo/En d o Ra tios for P a ter n o`-Bu1 ch i Rea ction s w ith F u r a n (1)
Griesbeck et al.
(3)
benzene as low as 10 mol % were sufficient to enable
more than 70% of the polar pathway (i.e., products 5 and
7 and glycol).
Su bstitu en t Effects in F u r a n Cycloa d d ition s. The
extraordinary high stereoselectivity of the benzaldehyde/
furan photocycloaddition is not a result of π-interactions
with the phenyl substituent, i.e., aliphatic aldehydes do
react with exo diastereoselectivities higher than 99:1.7b,23
It was, therefore, of interest to study the influence of
R-substituents in benzoyl compounds on the stereoselec-
tivity of the Paterno`-Bu¨chi reaction with furan. Re-
cently, it was reported by Cantrell and co-workers that
methyl benzoate reacts with furan with a high degree of
exo selectivity.24 The relative configuration of the major
cycloadduct was determined by comparison of the rel-
evant 1H NMR signals. When repeating this reaction,
we found two diastereoisomeric products 8 in a ratio of
95:5 (eq 2). Similar as for the endo isomer 2, which we
F igu r e 3. X-ray structure of product 8.
eq 3). The acetophenone reaction gave only one product,11d
whereas a 77:23 mixture of diastereoisomers resulted
from the addition of benzoyl cyanide.26 Increasing the
size of the aryl group from phenyl to mesityl in aroyl
cyanides led to an increase in exo diastereoselectivity
from 3.9:1 up to 16:1.26 An analogous increase was
reported by us for the cycloaddition of aromatic aldehydes
with 2,3-dihydrofuran; however, in contrast to the furan
case, the endo diastereoisomer is always the major isomer
with the enol ether substrate.7g Scharf and co-workers
have intensively investigated Paterno`-Bu¨chi reactions
of phenyl glyoxylates with furan.27 Whereas the methyl
ester of phenylglyoxylate gave a 10:90 diastereoisomeric
ratio, all other esters (e.g., R ) menthyl)27c reported in
the literature or investigated by ourselves exclusively
resulted in formation of the endo diastereoisomer.
Thus, by changing the R-substituent from H or Me to
CN, COOMe, OMe, and eventually COOR, the direction
(exo-phenyl vs endo-phenyl) of the diastereoselectivity is
completely switched. Again, this cannot be a result of
π-interactions with the phenyl substituent because an
identical sequence is obtained when using the corre-
sponding tert-butyl compounds; i.e., pivaldehyde gave
solely the exo [2 + 2] cycloadduct,7b whereas methyl
trimethylpyruvate exclusively resulted in the formation
of the endo [2 + 2] cycloadduct7a (exo, endo with respect
to the tert-butyl group).
(2)
detected as a minor product in the cycloaddition of
benzaldehyde with furan, the proton 4-H of the major
isomer showed a high-field-shifted resonance line at 4.62
ppm and the corresponding signal of the minor diaste-
reoisomer at 5.36 ppm. This already indicated that the
literature assignment was not correct. Definite proof
resulted from the X-ray analysis of 8 (Figure 3).25 Thus,
the exchange of the hydrogen in benzaldehyde by a
methoxy group completely inverts the diastereoselectivity
in the photocycloaddition with furan.
Further modification of the R-substituent in the ben-
zoyl substrates uncovered a distinct dependence of the
exo/endo ratio on the size of this substituent (Scheme 2,
(23) For the propionic aldehyde photocycloaddition to furan an exo/
endo ratio of 82:1 was determined: Fiege, M. University of Cologne,
unpublished results.
(24) Cantrell, T. S.; Allen, A. C.; Ziffer, H. J . Org. Chem. 1989, 54,
140.
Ster eoselectivity Effects in 1,3-Cyclod ien e Cy-
cloa d d ition s. As already mentioned, not only furan and
substituted furans but also other heteroaromatic sub-
strates show highly stereoselective [2 + 2] photocycload-
(25) The crystals of 8 C12H12O3, M ) 204.22 from CH2Cl2 are
monoclinic, space group P21/n, a ) 1515.7(5) pm, b ) 550.0(1) pm, c )
1242.8(3) pm; â ) 93.15(2)°; V ) 1034.5(5) Å3; Z ) 4; dcalc ) 1.311
g/cm3; µ ) 0.094. Data Collection: Enraf-Nonius-CAD4 diffractometer;
Mo KR; graphite monochromator; Wyckoff-scan; θ range (deg) 2.06-
26.97; crystal dimensions 0.2 × 0.2 × 0.15 mm3; number of reflections
measured, 796; number of unique reflections, 796; number of reflections
I > 2σ(I), 528. Structural Analysis and Refinement: solution by direct
phase determination; method of refinement, full-matrix LSQ; hydrogen
positions of riding model with fixed isotropic U, data-to-parameter
ratio, 4.29; R1, 0.048 (for I > 2σ(I)); R1, 0.048 (all reflexes); weighting
scheme, w ) 1/[σ2(Fo2) + (0.100P)2]; largest differential peak/hole, 0.11
and -0.11 e Å-3; program used, Siemens SHELXTL-93.
(26) Zagar, C.; Scharf, H.-D. Chem. Ber. 1991, 124, 967.
(27) (a) Koch, H.; Scharf, H.-D.; Runsink, J .; Leismann, H. Chem.
Ber. 1985, 118, 1485. (b) Weuthen, M.; Scharf, H.-D.; Runsink, J .;
Vassen, R. Chem. Ber. 1988, 121, 971. (c) Pelzer, R.; Scharf, H.-D.;
Buschmann, H.; Runsink, J . Chem. Ber. 1989, 122, 2, 1187. (d) Pelzer,
R.; J u¨tten, P.; Scharf, H.-D. Chem. Ber. 1989, 122, 487. (e) Buschmann,
H.; Scharf, H.-D.; Hoffmann, N.; Plath, M.; Runsink, J . J . Am. Chem.
Soc. 1989, 111, 5367.

Stereoselectivity of Triplet Photocycloadditions
J . Org. Chem., Vol. 63, No. 12, 1998 3851
Sch em e 3. Exo/En d o Ra tios: Com p a r in g
Cycloa lk en es w ith 1,3-Cyclod ien es
(4)
ditions with benzaldehyde and other triplet excited
aldehydes. Comparison of the cycloaddition stereoselec-
tivity of cycloalkenes with the corresponding cyclo-1,3-
dienes (Scheme 3, eq 4) shows the pertinent effect of an
extra double bond in conjugation with the reactive alkene
moiety. In all these cases, the direction of the diaste-
reoselectivity was inverted when going from the monoalk-
ene (endo selectivity) to the corresponding 1,3-diene (exo
selectivity); however, the magnitude of this effect is
strongly reduced for pure hydrocarbons in comparison
with the dihydrofuran/furan couple.
F igu r e 4. X-ray structure of product 10.
ceed via triplet 1,4-biradicals (2-oxatetramethylenes). To
progress from these intermediates to ground-state singlet
products, spin-inversion is necessary. The spin-imposed
barrier for triplet to singlet intersystem crossing leads
to a dramatic increase in lifetime (1-10 ns) when
compared with the corresponding singlet biradicals.
Because of the enhanced lifetimes, stereochemical infor-
mation originating from the starting materials is ex-
pected to be wiped out during the phototransformation.
The most important mechanism for the interaction
between singlet and triplet biradicals with shorter dis-
tances between the radical centers is spin-orbit coupling
(SOC). In contrast to electron-nuclear hyperfine cou-
pling and spin-lattice relaxation, SOC is strongly de-
pendent on the geometry of the triplet biradical. An
important geometrical prerequisite for strong SOC, as
postulated by Salem and Rowland5 and confirmed by
theoretical3 and spectroscopical4 work, concerns the
orthogonality of the p-orbitals at the spin-bearing carbon
atoms.31 Conservation of the total angular momentum
can be achieved when the axes of the p-orbitals at the
radical centers are oriented perpendicular to each other.
If the distance between the radical centers is less crucial
for facilitating ISC, there is not only one conformational
arrangement responsible for the triplet/singlet transition,
but many. After transition from the triplet to the singlet
potential energy surface, immediate product formation
is expected. Thus, the ISC is expected to be concerted
with the formation of a new bond or the cleavage of the
primarily formed single bond. With other words, the
stereochemistry of the products from [2 + 2] cycloaddition
reactions reflects the molecular ISC geometry.
To test the influence of spiroannulation for these types
of reactions, we also investigated the Paterno`-Bu¨chi
reaction of spiro[4.2]heptadiene 9 (eq 5). In constrast to
(5)
other 1,3-dienes that have been investigated earlier, the
endo side of substrate 9 is additionally shielded by the
spirocyclopropane ring. The exo diastereoselectivity for
the benzaldehyde cycloaddition was substantially re-
duced (from 212:1 for the furan adduct 2 to 3.5:1 for 10),
whereas the regioselectivity remained higher than ca.
500:1. Perfectly in line with the effects observed in the
furan series was the result that we obtained with the
methyl ester of phenyl pyruvate: exclusively the endo
diastereoisomer 11 was formed. To unambiguously
establish the relative configuration of this cycloadduct
we also performed an X-ray analysis (see Figure 4).30 In
both reactions, no products were observed that derive
from a ring-opening process of the spiroannulated cyclo-
propane ring.
We have published this SOC geometry model in order
to explain the stereochemical results of the photocycload-
dtion reactions of aromatic aldehydes with cyclic
monoalkenes.7e These results can be visualized using the
relevant Newman projections A-C of the intermediary
1,4-triplet biradicals (Scheme 4).
Discu ssion of Tr ip let 1,4-Bir a d ica l Geom etr ies. I.
Cyclic Mon oa lk en es. [2 + 2] cycloaddition reactions
of electronically excited triplet carbonyl compounds pro-
The conformers A and B are expected to be similarly
populated; however, ISC from A results in product
formation, whereas ISC from B leads to cleavage of the
singlet biradical and formation of starting material. Spin
inversion is coupled with a torque, which in the case of
conformers A and C leads to an immediate formation of
the new C-C bond. Thus, the torque induced in con-
former A rotates the large substituent (“large” with
respect to H) over the ring plane and results in the
formation of the endo diastereoisomer. From conformer
(28) Hoye, T. R.; Richardson, W. S. J . Org. Chem. 1989, 54, 688.
(29) Shima, K.; Kubota, T.; Sakurai, H. Bull. Chem. Soc. J pn. 1976,
49, 2567.
(30) The crystals of 11 C16H16O3, M ) 256.29 from acetone are
monoclinic, space group P21/c, a ) 880.6(2) pm, b ) 1148.6(2) pm, c )
1345.5(3) pm; â ) 104.34(2)°, V; 1318.5(5) Å3, Z ) 4; dcalc ) 1.291 g/cm3;
µ ) 0.088. Data Collection: Enraf-Nonius-CAD4 diffractometer; Mo
KR; graphite monochromator; Wyckoff-scan; θ range (deg) 2.36-26.97;
crystal dimensions 0.4 × 0.4 × 0.35 mm3; number of reflections
measured, 2319; number of unique reflections, 2143; number of
reflections I > 2σ(I), 1598. Structural analysis and refinement: solution
by direct phase determination; method of refinement, full-matrix LSQ;
hydrogen positions of riding model with fixed isotropic U, data-to-
parameter ratio, 8.78; R1, 0.047 (for I > 2σ(I)); R1, 0.068 (all reflexes);
weighting scheme, w ) 1/[σ2(Fo2) + (0.0575P)2 + 0.3681P]; largest
differential peak/hole, 0.198 and -0.174 e Å-3; program used, Siemens
SHELXTL-93.
(31) (a) Carlacci, L.; Doubleday: C., J r.; Furlani, T. R.; King, H. F.;
McIver, J . W. J . Am. Chem. Soc. 1987, 109, 5323. (b) Doubleday: C.,
J r.; Turro, N. J .; Wang, J .-F. Acc. Chem. Res. 1989, 22, 199.

3852 J . Org. Chem., Vol. 63, No. 12, 1998
Griesbeck et al.
Sch em e 4
F igu r e 5. Calculated triplet 1,4-biradical structures for the
furan/acetaldehyde reaction.
and triplet acetaldehyde. Structure D represents the
global energy minimum (fully optimized); the restricted
conformation F is 9.9 kJ /Mol higher in energy, separated
by an activation barrier of 28 kJ /mol. This activation
energy is obviously too high due to the conformational
restriction for rotation about the exocyclic O-C(H,Me)
bond. The alternative arrangement F ′ with the methyl
group pointing toward the ring plane is much higher in
energy and should be less significant in this reaction. If
structure F (and not D) represents the main contribution
to product formation, a high exo selectivity is expected
due the nearly exclusive formation of the “methyl-up”
conformation F . The torque induced in conformer F
rotates the large substituent (Me or any other substitu-
ent) in the opposite direction with respect to the ring
plane and results in the formation of the major (exo)
diastereoisomer.
Sch em e 5
What happens if the hydrogen in structure F is
exchanged by a substituent larger than methyl; i.e., if
ketones are used as triplet carbonyl substrates? Follow-
ing the arguments from above, both structures F and F ′
are no longer relevant and only structure D remains as
the SOC-controlling arrangement for the ISC process.
Thus, all triplet carbonyl substrates whose substituents
exceed a limiting degree of steric interaction are expected
to give endo-selective C-C bond formation.34 This model
offers an explanation for the dichotomy that some car-
bonyl substrates (e.g., phenyl glyoxylates) give idential
diastereoselectivity with cyclo-1,3-dienes and with the
corresponding cycloalkenes, whereas other carbonyls
(e.g., aldehydes or acetophenone) show a completely
different selectivity pattern. An alternative explanation
for the high exo selectivity in furan-aldehyde photocy-
cloadditions that we have suggested previously was the
extended lifetime of the singlet 1,4-biradical that is
formed after ISC.7g This concept, however, predicts
thermodynamic control for the formation of all cycload-
dition products, whether they are formed from triplet
excited aldehydes or ketones, esters, etc. Obviously, this
is not the case (see eqs 3 (Scheme 2) and 5). Thus, an
interaction between the allylic and the exocyclic radical
in the 1,4-triplet biradical (as depicted in structures F )
must be crucial for the dominance of this biradical
geometry for rapid ISC. This effect can be described as
C, preferentially the exo diastereoisomeric products are
formed. Assuming that A is lower in energy compared
with C for unsubstituted cycloalkenes as substrates, a
pronounced endo selectivity is expected that is confirmed
by the experiments. Increasing steric repulsions between
the substituents at the exo- and endocyclic radical centers
depopulates conformer A, and more exo product is formed
via C. This effect has been described in detail for several
methylated cycloalkenes.7d
II. Cyclic 1,3-Dien es. When comparing these results
with the reactions of triplet excited aldehydes with 1,3-
cyclodienes (furan, cyclopentadiene, and 1,3-cyclohexa-
diene), two important differences become apparent:
besides the complete change in regioselectivity, the di-
astereoselectivity is highly increased and the direction of
stereocontrol is inverted (from endo to exo). Similar as
for cycloalkenes, three 1,4-biradical conformations D-F
are relevantly discussed (Scheme 5).
As already described for the cycloalkene case, ISC from
D and F are expected to lead to endo and exo diastere-
oisomers, respectively. Using density functional theory
(DFT)32-B3LYP/6-31G** from the Gaussian 94
package33swe determined whether structure F (F ′) with
an orthogonal arrangement of the two spin-bearing
orbitals corresponds to an energy minimum (Figure 5).
As a model reaction, we chose the 1,4-biradical from furan
(34) One reviewer remarked that the endo/exo ratios (described, for
example, in eq 3) do not vary in the same order as A values, which are
often used to quantify the steric demand of a specific group. This is
definitely true, and other contributions also have to be considered, e.g.,
spin dilution effects by the (conjugated or notconjugated) substituent.
For cyclic monoalkenes, we have shown that steric effects are truly
decisive (ref 7b-e), and A values are useful parameters to correlate
with stereoselectivities. In our opinion, A values are, however, not
applicable for description of steric interactions in biradicals such as
F ′. The cyano group (A value ca. 1.0) is expected to interact much
stronger in the pseudoperpendicular conformation F ′ than does a
methyl group (A value 7.5).
(32) Bartolotti, L. J .; Flurchick, U. Rev. Comput. Chem. 1996, 7, 187.
(33) Gaussian 94, Revision B.1: Frisch, M. J .; Trucks, G. W.;
Schlegel, H. B.; Gill, P. M. W.; J ohnson, B. G.; Robb, M. A.; Cheeseman,
J . R.; Keith, T.; Petersson, G. A.; Montgomery, J . A.; Raghavachari,
K.; Alaham, M. A.; Zakrzewski, V. G.; Ortiz, J . V.; Foresman, J . B.;
Cioslowski, J .; Stefanov, B. B.; Nanayakkara, A.; Challacombe, M.;
Peng, C. Y.; Ayala, P. Y.; Chen, W.; Wong, M. W.; Andres, J . L.;
Replogle, E. S.; Gomperts, R.; Martin, R. L.; Fox, D. J .; Binkley, J . S.;
Derees, D. J .; Baker, J .; Stewart, J . P.; Head-Gordon, M.; Gonzalez,
C.; Pople, J . A. Gaussian Inc., Pittsburgh, PA, 1995.

Stereoselectivity of Triplet Photocycloadditions
J . Org. Chem., Vol. 63, No. 12, 1998 3853
DEPT. UV-vis: Hitachi U-3200. Column chromatography:
silica gel (Merck) 60-230 mesh; petroleum ether (PE, 40-60
°C), ethyl acetate (EA). All melting points were determined
with a Bu¨chi melting point apparatus (type Nr. 535) and are
uncorrected. Combustion analyses: Institut fu¨r Anorganische
Chemie der Universita¨t zu Ko¨ln. Dielectric constants: dipol-
meter DM 01. Rayonet-chamber photoreactors RPR-208 (8
× 3000 Å lamps, ca. 800 W, λ ) 300 ( 10 nm) and immersion-
well reactors (λ > 280 nm) were used for irradiations. Gas
chromatography: Hewlett-Packard 5890, series II, capillary
column HP-5 (30 m × 0.32 mm × 0.25 µm cross-linked 5% PH
ME silicone), temperature program: 60-250 °C in 20 °C/min
steps (1 min initial time), flow gas, nitrogen.
Gen er a l P r oced u r e for P h otocycloa d d ition Rea ction s.
A solution of 5 mmol of carbonyl substrate in 100 mL of solvent
in the presence of 50 mmol of 1,3-diene (furan, spiro-
[2,4]heptadiene36) was irradiated until quantitative conversion
of the carbonyl addend was detected by TLC and/or GC. After
evaporation of the solvent, the crude product mixture was
analyzed by high-field NMR. Purification was performed by
bulb-to-bulb distillation, by flash chromatography, or by
repeated crystallization.
Deter m in a tion of Solven t P ola r ity Effects on th e
P a ter n o`-Bu1 ch i Rea ction of 4 w ith P h CHO. A solution
of 1.0 g (9.43 mmol) of benzaldehyde and 7.0 g (100 mmol) of
2,3-dihydrofuran in a mixture of benzene and acetonitrile was
irradiated until complete conversion of the aldehyde. The
product composition was determined directly from the crude
mixture by gas chromatogrphy. GC retention times (min): 6
(7.3), 7 (7.4), 5 (7.6), 1,2-diphenylglycol (9.8).
exo-6-P h en yl-2,7-d ioxa bicyclo[3.2.0]h ep ta n e (exo-5).20
1H NMR (300 MHz, CDCl3): δ 1.82 (m, 1H, 4-H), 2.08 (m, 1H,
4′-H), 3.24 (ddd, J ) 3.8, 4.0, 4.0 Hz, 1H, 5-H), 4.35 (m, 2H,
3,3′-H), 5.05 (d, J ) 4.0 Hz, 1H, 6-H), 6.08 (d, J ) 3.8 Hz, 1H,
1-H), 7.10-7.38 (m, 5H, ArH). 13C NMR (75 MHz, CDCl3): δ
28.8 (t), 49.7 (d), 67.7 (t), 82.5 (d), 106.3 (d), 125.2 (d), 128.4
(d), 128.7 (d), 142.1 (s). Anal. Calcd for C11H12O2: C, 75.00;
H, 6.87. Found: C, 75.10; H, 7.40.
a secondary orbital interaction that facilitates intersys-
tem crossing by means of an increase in spin-orbit
coupling.
III. Solven t Effects. The solvent effect simulta-
neously on the regio- and stereoselectivity of the 2,3-
dihydrofuran photocycloaddition with benzaldehyde de-
scribed above can be explained by assuming intermediate
contact or solvent-separated radical ion pairs (CIP, SSIP)
of the aldehyde radical anion and the enol ether radical
cation formed via photoinduced electron transfer (PET).
In unpolar solvents, this pathway is unfavorable, and an
increase in solvent polarity leads to effective competition
3
3
between biradical (via BR) and PET (via CIP) pathway
(see Scheme 1). The nonlinear correlation between PET-
product formation and xp resembles that of excited-state
properties such as fluorescence lifetimes, quantum yields,
or solvatochromic shifts in solvent mixtures.35 The
radical coupling product 7, which was observed under all
conditions as the major side product, could be formed by
a hydrogen abstraction reaction (which is often observed
in unpolar solvents) and subsequent combination of the
geminate radical pair. An alternative way involves the
contact ion pair (CIP ) and leads to 7 by a sequence of
proton transfer and radical combination. Compound 7
was already present in minor quantities in unpolar
solvents, and its relative appearance shows a similar
nonlinear behavior in solvent mixtures. Thus, product
7 is also preferentially formed by a photoinduced electron-
transfer mechanism. The product stereoselectivity for 7
is determined at the second reaction step and is only
marginal, as expected from the relatively symmetric
structure of the allyloxy radical formed from 4. Another
hint for the assumption of contact or solvent separated
radical ion pair formation in polar solvents was the
increasing amount of pinacol formation (ca. 20% in
acetonitrile), which normally accompanies radical-cou-
pling processes via the hydroxybenzyl radical. To form
diphenyl glycol via the PET pathway, the radical anion
has to escape from the radical ion cage, which is an
important competing process. In summary, competing
biradical and PET processes can be differentiated using
solvent effects on the product ratio. The stereoselectivity
of the photocycloaddition with 4 is controlled by different
parameters for both pathways: SOC geometries in the
case of the biradical path and radical combination
geometries for the PET path. Combination of the two
dominantly spin-bearing carbon centers results in a 1,4-
zwitterion that subsequently could undergo the second
bonding event to give 5. Whereas for 6 the diastereose-
lectivity is determined at the second reaction step (C-C
bond formation between two prostereogenic carbon radi-
cals), the stereochemistry of product 5 corresponds to the
relative configuration of the 1,4-zwitterion (see Scheme
1). Obviously, Coulombic interaction must already exist
for the radical ion pair, allowing the radical anion and
radical cation to orientate in such a way that the product
configuration is matched.
en d o-6-P h en yl-2,7-d ioxa bicyclo[3.2.0]h ep ta n e (en d o-5).
1H NMR (300 MHz, CDCl3): δ 6.06 (d, J ) 3.7 Hz, 1H, 1-H),
no other signals were detectable.
3-(2,3-Dih yd r ofu r a n yl)-1-p h en ylm eth a n ol (7). Diaster-
eomeric ratio 71:29 after preparative thick-layer chromatog-
raphy (cyclohexane/ethyl acetate 3:1, silica). Major diastere-
oisomer. 1H NMR (300 MHz, CDCl3): δ 1.55 (br s, 1H, OH),
3.37 (m, 1H, 3-H), 4.26 (m, nonresolved AB system, 2H, 2-H),
4.54 (d, J ) 6.9 Hz, 1H, CHOH), 4.58 (dd, J ) 2.5, 2.5 Hz, 1H,
4-H), 6.39 (dd, J ) 2.1, 2.5 Hz, 1H, 5-H), 7.21-7.45 (m, 5H,
ArH). 13C NMR (75 MHz, CDCl3): δ 50.5 (d), 72.0 (t), 77.2
(d), 100.4 (d), 126.2 (d), 127.8 (d), 128.5 (d), 142.5 (s), 147.9
(d). Minor diastereoisomer. 1H NMR (300 MHz, CDCl3): δ
1.55 (br. s, 1H, OH), 3.37 (m, 1H, 3-H), 4.43 (m, nonresolved
AB system), 4.54 (d, J ) 6.9 Hz, 1H, CHOH), 4.92 (dd, J )
2.5, 2.5 Hz, 1H, 4-H), 6.43 (dd, J ) 2.1, 2.5 Hz, 1H, 5-H), 7.21-
7.45 (m, 5H, ArH). 13C NMR (75 MHz, CDCl3): δ 50.1 (d),
72.2 (t), 76.1 (d), 99.2 (d), 126.1 (d), 127.9 (d), 128.7 (d), 142.5
(s), 148.3 (d). Anal. Calcd for C11H12O2: C, 75.00; H, 6.87.
Found: C, 74.84; H, 7.16.
6-Meth oxy-en d o-6-p h en yl-2,7-d ioxa bicyclo[3.2.0]h ep t-
3-en e (en d o-8).25 Irradiation time: 43 h, rt. Yield: 62%. dr
) 95:5. 1H NMR (300 MHz, CDCl3): δ 3.18 (s, 3H, OMe), 4.01
(ddd, J ) 1.2, 2.9, 3.9 Hz, 1H, 5-H), 4.64 (dd, J ) 2.9, 2.9 Hz,
1H, 4-H), 6.41 (ddd, J ) 0.8, 1.2, 2.9 Hz, 1H, 3-H), 6.43 (dd, J
) 0.8, 3.9 Hz, 1H, 1-H), 7.32-7.43 (m, 5H, ArH). 13C NMR
(75 MHz, CDCl3): δ 50.2 (q), 56.4 (d), 101.5 (d), 105.0 (d), 114.5
(s), 126.7 (d), 128.0 (d), 128.4 (d), 136.8 (s), 148.3 (d).
6-Meth oxy-exo-6-p h en yl-2,7-d ioxa bicyclo[3.2.0]h ep t-3-
en e (exo-8). 1H NMR (300 MHz, CDCl3): δ 3.85 (m, 1H, 5-H),
5.36 (dd, J ) 2.9, 2.9 Hz, 1H, 4-H), no other signals were
detectable.
Exp er im en ta l Section
Gen er a l Meth od s. 1H NMR: AC 200 (200 MHz), Bruker
AC 250 (250 MHz), Bruker AC 300 (300 MHz) Bruker AC 500
(500 MHz). 13C NMR: Bruker AC 200 (50.3 MHz), Bruker
AC 250 (63.4 MHz), carbon multiplicities were determined by
Sp ir o[cyclop r op a n e-1,2′-(exo-6-p h en yl-7-oxa b icyclo-
[3.2.0]h ep t -3-en e)] (exo-10). Irradiation time: 20 h, rt.
(35) (a) Ghoneim, N.; Suppan, P. Pure Appl. Chem. 1993, 65, 1739.
(b) Nitsche, K.- S.; Suppan, P. Chimia 1982, 36, 346.
(36) Wilcox, C. F.; Craig, R. A. J . Am. Chem. Soc. 1961, 83, 3866.

3854 J . Org. Chem., Vol. 63, No. 12, 1998
Griesbeck et al.
Conversion: 20%. Due to the low degree of conversion, the
two diastereoisomeric oxetanes were not isolated; the dr value
(78:22) was determined from the crude product mixture. 1H
NMR (300 MHz, CDCl3): δ 0.58-0.92 (m, 2H), 1.02-1.20 (m,
2H), 3.55 (m, 1H, 5-H), 4.84 (dd, J ) 1.0, 5.5 Hz, 1H, 1-H),
5.25 (d, J ) 3.3 Hz, 1H, 6-H), 5.53 (ddd, J ) 1.0, 1.0, 5.6 Hz,
1H, 3-H), 5.91 (dd, J ) 2.5, 5.6 Hz, 1H, 4-H), 7.13-7.86 (m,
5H, ArH). 13C NMR (75 MHz, CDCl3): δ 6.8 (t), 15.0 (t), 29.6
(s), 35.3 (d), 89.2 (d), 90.1 (s), 125.2 (d), 128.5 (d), 128.9 (d),
129.7 (d), 134.4 (s), 139.9 (d).
5.27 (dd, J ) 2.5, 5.7 Hz, 1H, 4-H), 5.37 (ddd, J ) 1.0, 1.0, 5.7
Hz, 1H, 4-H), 7.26-7.46 (m, 5H, ArH). 13C NMR (50 MHz,
CDCl3): δ 6.6 (t), 14.2 (t), 34.8 (s), 52.8 (s), 54.5 (d), 86.6 (d),
90.3 (s), 125.2 (d), 125.4 (d), 127.6 (d), 127.9 (d), 137.9 (s), 140.9
(d), 175.2 (s). Anal. Calcd for C16H16O3: C, 74.98; H, 6.29.
Found: C, 74.72; H, 6.55.
Ack n ow led gm en t. Financial support came from the
Deutsche Forschungsgemeinschaft and the Fonds der
Chemischen Industrie. The DFT calculations were
performed by K. Malsch from the research group of Prof.
G. Hohlneicher, University of Cologne.
S p ir o [c y c lo p r o p a n e -1,2′-(en d o-6-p h e n y l-7-o x a b i-
cyclo[3.2.0]h ep t -3-en e)] (en d o-10). 1H NMR (300 MHz,
CDCl3): δ 5.47 (d, J ) 5.4 Hz, 1H, 1-H), 5.96 (dd, J ) 2.5, 5.6
Hz, 1H, 4-H), no other signals were detectable.
S p ir o [c y c lo p r o p a n e -1,2′-(6-p h e n y l-7-o x a b ic y c lo -
[3.2.0]h ep t-3-en -6-ca r boxylic a cid m eth yl ester )] (11).
Irradiation time: 65 h, rt. Yield: 50%. dr ) >95:5. Purifica-
tion by column chromatography (silica, EtOAc-PE 1:5), color-
less needles. Mp: 120-121 °C. 1H NMR (200 MHz, CDCl3):
δ 0.70-0.85 (m, 2H), 1.25-1.32 (m, 2H), 3.80 (s, 3H), 4.42 (ddd,
J ) 1.0, 2.5, 5.2 Hz, 1H, 5-H), 4.82 (d, J ) 5.2 Hz, 1H, 1-H),
Su p p or tin g In for m a tion Ava ila ble: X-ray data for com-
pounds 8 and 11 (6 pages). This material is contained in
libraries on microfiche, immediately follows this article in the
microfilm version of the journal, and can be ordered from the
ACS; see any current masthead page for ordering information.
J O971767L