except for ketone 1a, where low temperatures were required.
The photochemical results for ketones 1a-c are summarized
in Table 1.
Scheme 2. Proposed Mechanism for Solid-State Oxetane
Formation
Table 1. Photochemical Results for Ketones 1a-c
reaction temp reaction conva
ketone medium (°C) time (h)
(%) 2 (%)b 3 (%)b 4 (%)b
1a
CH3CNc
crystal
rt
-20
15
2
10
22.5
2
100
38
95
100
39
53
100
100
100
48
43
0
0
0
0
0
60
0
47
0
0
0
91
59
63
92
83
0
13
0
19
20
0
6
13
12
4
8
0
28
0
12
11
-30
7
1b
1c
CH3CNc
crystal
CH3CNc
crystal
rt
rt
rt
rt
5.5
46
3
24
51.5
73
oriented for oxetane formation. Under the reaction conditions,
benzaldehyde is able to absorb a photon and undergo a
Paterno`-Bu¨chi reaction,6 presumably through its n-π* triplet
excited state, with the ground-state alkene 5 to form the
observed oxetanes 3 and 4. The regioselectivity of the process
is what one would expect (in solution) on the basis of
formation of the more stable biradical intermediate. This
mechanism is supported by the detection of trace amounts
of intermediates 5 and 6 in the reaction mixture by GC-
MS. It is likely that these two species do not accumulate
owing to the efficient formation of the oxetanes in the crystal
lattice.
For purposes of comparison, the solution-state photore-
action of benzaldehyde (6) with 1-phenylcyclopentene (5)
was studied in benzene. The reaction was very slow and
produced a number of products as evidenced from the GC
trace of the reaction mixture (Figure 1a). After 69 h of
irradiation, benzaldehyde 6 was almost entirely consumed,
but the product mixture contained only 6% of exo-oxetane
3 and 13% of endo-oxetane 4. The preference for formation
of endo-oxetanes in the solution-phase Paterno`-Bu¨chi reac-
tions of benzaldehyde with cyclic olefins has been noted and
explained on the basis of faster spin-orbit coupling in the
1,4-triplet biradical leading to the endo product.7 This makes
the preferred exo stereochemistry observed in the solid state
a Based on total GC integral due to remaining starting material.
b Percentage of total GC integral due to a given product in the product
mixture. c Typical concentrations for solution-phase photolyses were 30-
100 mg of ketone in 5 mL of acetonitrile.
Solution photolysis of phenyl ketone 1a afforded com-
pound 2, which was formed via a typical Norrish type I
R-cleavage pathway, as the only isolable product. In the solid
state, ketone 1a showed different reactivity and the oxetanes
were formed very efficiently, with exo-isomer 3a as the major
component and endo-isomer 4a as the minor component. The
structures of products 3a and 4a were determined on the
basis of spectroscopic measurements (IR, LRMS, HRMS,
1H NMR, 13C NMR, COSY, HMQC, and NOE difference)
and elemental analysis. Similar behavior was observed in
the solid-state photolysis of ketones 1b and 1c (Table 1),
although the yield of oxetanes was considerably lower owing
to the formation of numerous minor (<5% each) side
products.
Solid-state oxetane formation is unprecedented and a
possible mechanism is presented in Scheme 2. In this
mechanism, irradiation of crystals of 1a results in Norrish
type I cleavage of the R carbon-carbon bond to form a
benzoyl radical and a 1-phenylcyclopentyl radical. Subse-
quent hydrogen atom transfer to the benzoyl radical from
the C2 position of the adjacent alkyl radical affords two
neutral molecules, 1-phenylcyclopentene (5) and benzalde-
hyde (6). Because the process takes place in the crystalline
state, the radical pair and molecules 5 and 6 are formed
within a crystal cage, with the relative orientation between
5 and 6 resembling the starting ketone 1a.5 It therefore seems
reasonable to suggest that the CdO double bond of benzal-
dehyde and the CdC double bond of alkene 5 are favorably
(6) (a) Paterno`, E.; Chieffi, G. Gazz. Chim. Ital. 1909, 341. (b) Bu¨chi,
G.; Inman, C. G.; Lipinsky, E. S. J. Am. Chem. Soc. 1954, 76, 4327. For
recent reviews about the Paterno`-Bu¨chi reaction, see: (c) Bach, T. Synthesis
1998, 683. (d) Griesbeck, A. G. In CRC Handbook of Organic Photochem-
istry and Photobiology; Horspool, W. M., Song, P.-S., Eds.; CRC Press:
New York, 1995; p 522. (e) Rivas, C. In CRC Handbook of Organic
Photochemistry and Photobiology; Horspool, W. M., Song, P.-S., Eds.; CRC
Press: New York, 1995; p 536. (f) Griesbeck, A. G. In CRC Handbook of
Organic Photochemistry and Photobiology; Horspool, W. M., Song, P.-S.,
Eds.; CRC Press: New York, 1995; p 550. (g) Carless, H. A. J. In CRC
Handbook of Organic Photochemistry and Photobiology; Horspool, W. M.,
Song, P.-S., Eds.; CRC Press: New York, 1995; p 560. (h) Inoue, Y. Chem.
ReV. 1992, 92, 741. (i) Porco, J. A., Jr.; Schreiber, S. L. In ComprehensiVe
Organic Synthesis; Trost, B. M., Fleming, I., Paquette, L. A., Eds.; Pergamon
Press: Oxford, 1991; Vol. 5, p 151. (j) Demuth, M.; Mikhail, G. Synthesis
1989, 145. (k) Carless, H. A. J. In Photochemistry in Organic Synthesis;
Coyle, J. D., Ed.; Royal Society of Chemistry: London, 1986; Vol. 57, pp
95. (l) Carless, H. A. J. In Synthetic Organic Photochemistry; Horspool,
W. M., Ed.; Plenum Press: New York, 1984; pp 425. (m) Jones, G., II.
Org. Photochem. 1981, 5, 1.
(4) Ketone 1b was obtained by heating fluoro-substituted ketone 1 (R
) F) with potassium cyanide in dimethyl sulfoxide (DMSO). When R )
F, ketone 1 is a liquid and its solid-state photochemistry could not be studied.
The details of the preparation of compounds 1a-g will be given in a
subsequent full paper.
(5) For reviews about topochemical rules, see: (a) Ramamurthy, V.;
Venkatesan, K. Chem. ReV. 1987, 87, 433. (b) Weiss, R. G.; Ramamurthy,
V.; Hammond, G. S. Acc. Chem. Res. 1993, 26, 530.
3362
Org. Lett., Vol. 3, No. 21, 2001