S. Annalakshmi et al.
Bull. Chem. Soc. Jpn., 78, No. 11 (2005) 2001
30 min. The tubes were then placed in HEBER multilamp photo-
reactors fitted with (4 ꢂ 8 W) 254 nm lamps and (8 ꢂ 8 W) 365
nm lamps for 1 h. During photolysis, the tubes were rotated peri-
odically to ensure uniform irradiation. All of the solid-state photo-
reactions of CD complexes were carried out by placing the solid
complexes in Quartz/Pyrex tubes, degassed with N2 and then
sealed. They were then photolysed for 48 h with periodic rotation
of the tubes after every 30 min.
motion of the 1,4-biradical. Kinetic studies on the photoreac-
tion of valerophenone have been investigated as a function
of the temperature, pH, and wavelength in an aqueous medi-
um.11 Type-II quantum yields for the photoreaction are close
to unity, and the values for the formation of the photoproducts
are 0.65 (for cleavage to acetophenone and propene) and an
overall yield of 0.32 for cyclization to two cyclobutanols at
20 ꢁC. Using X-ray crystallographic studies, Stezowski et
al.,12 have reasoned the photochemical modification of valero-
phenone in a ꢂ-cyclodextrin cavity by proposing a 2:2 ꢂ-CD/
guest system. These face-to-face, ꢂ-CD dimers contain two in-
cluded aryl alkyl ketone molecules with the ꢂ-CD dimers
packing in a channel. The guest molecules are packed with
their phenyl rings face-to-face, located in the center of the
ꢂ-CD dimer. This leaves the alkyl chains of the ketones ex-
tending to the primary hydroxy ends of the ꢂ-CD dimer. Ke-
tones bearing methyl substituents ꢁ to the benzoyl group, as
in cis-4-tert-butyl-1-methylcyclohexyl phenyl ketone, undergo
photocyclization5 to afford cyclobutanols. Scheffer et al., have
investigated13 the Norrish–Yang type-II photochemistry of 16
such ketones having the basic cis-4-tert-butyl-1-benzoylcyclo-
hexane or 2-benzoyladamantane structure in the solid state and
in solution. Asymmetric induction studies are carried out by
providing the reactants with carboxylic acid substituents to
which ‘‘ionic chiral auxiliaries’’ are attached through salt for-
mation with optically active amines. The irradiation of the
salts in solution gives racemic cyclobutanols, but in the crys-
talline state, moderate to near-quantitative enantiomeric ex-
cesses are obtained. Photochemical reactions of valerophenone
in various surface media,14 such as silica gel, alumina, and wa-
ter-ice, have also been reported. A model in which the short-
lived biradical intermediate interacts with the surface, in addi-
tion to a polar effect on the excited triplet of ketone, is pro-
posed. These interesting features of valerophenone photo-
chemistry, such as the dependence of the E/C ratio and dia-
stereomeric induction on reaction media prompted us to carry
out this photoreaction in various cyclodextrins to gain insight
into the effect of the cavity size and also how the steric con-
straints in CDs can be utilized to control the diastereoselectiv-
ity of the obtained substituted cyclobutanols.
After the completion of photolysis and extraction from the CD
cavity using hot CHCl3, the reaction mixture was analyzed using
a Shimadzu GC-17A, CYDEX-B Chiral capillary column (30 m)
with a FID detector and high-purity nitrogen as the carrier gas.
Only with this column were the two diastereomeric cyclobutanols
well separated, since our attempt to resolve the diastereomeric cy-
clobutanols was unsuccessful with a SE-30 capillary column. Un-
der the conditions used for analysis, low-molecular-weight gases
such as propene, were not detected. Diastereomeric cyclobutanols
were isolated and characterized from their GC-MS and NMR da-
ta18 (for the cis-isomer, a doublet is observed at upfield due to the
anisotropic diamagnetic shielding of methyl protons by the phenyl
ring and for the trans-isomer, the corresponding doublet appears
downfield due to the anisotropic deshielding by phenyl ring). IR
spectra show peaks at 3500 cmꢃ1 (ꢃ{OHstr), and the carbonyl
stretching frequency around 1700 cmꢃ1 is absent. ICD spectra is
recorded using a JASCO J-810 spectropolarimeter, furnished with
a 150 W xenon lamp. The measurements were performed under a
ꢁ
nitrogen flux at 25 ꢄ 1 C, and the samples were contained in a
quartz cuvette of pathlength of 0.1 cm. The acquisition parameters
were: wavelength range, 200–500 nm at steps of 1 nm; bandwidth,
2 nm; time constant, 0.5 s; and sensitivity, 2 mdeg/div. The instru-
ment was calibrated by using a 0.06% aqueous solution of ammo-
nium D-10-camphosulphonate, from JASCO.
Results and Discussion
The results of the photoreaction of valerophenone in various
solvents and also in presence of different cyclodextrins, either
in solution or in the solid state, are presented in Tables 1 and 2.
The reaction mixture irradiated at 254 nm (which excites the
ꢀ
ꢄ, ꢄ state) was analyzed in a SE-30 capillary column, and
a product distribution similar to an earlier report10 was ob-
tained (Table 1). Irradiation at 365 nm (which excites the n,
ꢀ
ꢄ
state directly) did not alter the product distribution upon
Experimental
irradiation in isotropic solvents, but increased the amount of
elimination products when 1 was irradiated as its cyclodextrin
complexes. Also, the isomeric cyclobutanols were not re-
solved, and only a single peak was obtained.
The preparation of 1:1 cyclodextrin complexes was done as
per a reported procedure.15 The existence of an inclusion complex
was evidenced by measuring the dissociation constants using the
Benesi–Hildebrand equation.16 1D and 2D NMR spectra were
recorded in D2O at 25 ꢁC on a Bruker 300 MHz instrument (using
the pulse sequences and standard procedures offered by Bruker).
The upfield chemical shifts of ꢂ-CD protons (H-3 and H-5) in
the presence of valerophenone was indicative of the inclusion of
the phenyl ring into the ꢂ-CD cavity. The formation of only 1:1
stoichiometric complexes with all three cyclodextrins was inferred
from a Job’s plot17 (Fig. 1).
For solution photolysis, 0.082 mL of valerophenone in 5 mL of
respective solvents was irradiated (Table 1) in a N2 atmosphere at
the appropriate wavelength (254/365 nm). The photolysis of cy-
clodextrin complexes of 1 was carried out as follows. Microcrys-
talline CD complexes of 1 (150 mg) were taken in Quartz/Pyrex
tubes (for 254/365 nm irradiations respectively), dissolved in 50
mL of distilled water and degassed with high-purity N2 gas for
When the same reaction mixture (irradiated at 254 nm) was
analyzed in a chiral CYDEX-B capillary column (presented in
Table 2), C1 and C2 were resolved. The ratio C1/C2 is also in
conformity with that reported11 previously. In polar solvents,
the amount of 2 increased compared to the cyclized products.
This is in accordance with reported data,19 and may be due to
greater stabilization and solvation of the newly formed hydrox-
yl group in the biradical intermediate by polar solvents that
suppress cyclization. This biradical consequently undergoes a
preferential rotation to form the transoid biradical, resulting
in larger amount of acetophenone 2.
The reaction, when studied inside the hydrophobic environ-
ment of cyclodextrins, provide very interesting selectivities. In
all of the cyclodextrin catalyzed reactions, the cyclized prod-
Annalakshmi, Subramanian
