An investigation of the Yang photocyclization reaction in the solid state: Asymmetric induction studies and crystal structure-reactivity relationships

Article abstract of DOI:10.1021/ja9823013

The Norrish/Yang type II photochemistry of 16 ketones having the basic cis-4-tert-butyl-1-benzoylcyclohexane or 2-benzoyladamantane structure has been investigated in the solid state and solution. In both media, ketones bearing methyl substituents α to th

Full text of DOI:10.1021/ja9823013

J. Am. Chem. Soc. 1998, 120, 12755-12769
12755
An Investigation of the Yang Photocyclization Reaction in the Solid
State: Asymmetric Induction Studies and Crystal
Structure-Reactivity Relationships
Mordechai Leibovitch, Gunnar Olovsson, John R. Scheffer,* and James Trotter
Contribution from the Department of Chemistry, UniVersity of British Columbia, 2036 Main Mall,
VancouVer, Canada V6T 1Z1
ReceiVed July 1, 1998
Abstract: The Norrish/Yang type II photochemistry of 16 ketones having the basic cis-4-tert-butyl-1-
benzoylcyclohexane or 2-benzoyladamantane structure has been investigated in the solid state and solution.
In both media, ketones bearing methyl substituents R to the benzoyl group undergo stereoselective Yang
photocyclization to afford endo-arylcyclobutanols. Quantum yield and quenching studies in solution show
that the reactions are efficient triplet-mediated processes. Asymmetric induction studies were carried out by
providing the reactants with carboxylic acid substituents to which “ionic chiral auxiliaries” were attached
through salt formation with optically active amines. Irradiation of the salts (17 in total) in solution gave
racemic cyclobutanols, but in the crystalline state, moderate to near-quantitative enantiomeric excesses were
obtained. Single crystal X-ray diffraction studies were successfully performed on 10 neutral ketones and four
salts. This allowed the reactive γ-hydrogen atoms to be identified and the distance and angular parameters
associated with their abstraction to be tabulated. For the 14 compounds whose crystal structures were
determined, the average value of d, the CdO‚‚‚H abstraction distance, was 2.61 ( 0.07 Å, and the values of
ω (the γ-hydrogen out-of-plane angle), ∆ (the CdO‚‚‚Hγ angle) and θ (the C-Hγ‚‚‚O angle) were 53 ( 11,
84 ( 7, and 115 ( 2°, respectively. In a similar manner, the geometric parameters associated with the ring
closure reactions of the intermediate 1,4-hydroxy biradicals were estimated from the crystallographic data.
This indicates that the biradicals are 66.5 ( 9.8 and 32.7 ( 3.2° out of alignment for cleavage, but that they
are well oriented for cyclization, with radical separations (D) of 3.08 ( 0.09 Å. For one of the salts in the
adamantane series, the solid-state photoreaction was shown to be topotactic; that is, a single crystal of the
reactant salt was transformed quantitatively into a single crystal of the corresponding cyclobutanol salt. Since
the absolute configuration of the ionic chiral auxiliary in this case was known, this permitted the absolute
steric course of the reaction to be mapped by X-ray crystallography and the abstracted γ-hydrogen to be
identified unequivocally. In addition, the crystal structure of a partially reacted crystal containing 60% product
and 40% reactant was successfully determined. This showed that the two compounds in the mixed crystal
have nearly identical shapes and orientations, thus accounting for the single crystal nature of the process.
Finally, it was found that ketones lacking methyl substituents R to the benzoyl group are either photochemically
unreactive or undergo Norrish type II cleavage. Possible reasons for this difference in behavior are presented
and discussed.
Introduction
optically active photosensitizers,1,3 (4) photolysis in the presence
of optically active additives,1,4 (5) attachment of covalently
bound chiral auxiliaries,1 (6) photolysis in the cavities of
optically active host molecules1,5 and (7) solid-state irradiation
of materials that crystallize in chiral space groups.1,6 Of these
methods, the latter two most consistently lead to useful
enantiomeric excesses, and it is with the second of theses
photolysis of chiral crystalssthat we shall be concerned in the
present article.
Organic photochemistry is replete with reactions that trans-
form achiral starting materials into chiral photoproducts, and
one of the challenges facing organic photochemists today is how
to accomplish these reactions enantioselectiVely. While, histori-
cally, much more effort has been put into developing general
methods of asymmetric synthesis for ground-state reactions
compared to those occurring in the excited state, the field of
photochemical asymmetric synthesis has nevertheless seen some
notable advances in recent years.1 Among the approaches that
have been taken are (1) photolysis with circularly polarized
light,1,2 (2) the use of optically active solvents,1 (3) the use of
(4) Pete, J. P. In AdVances in Photochemistry; Neckers, D. C., Volman,
D. H., Von Bu¨nau, G., Eds.; Wiley: New York, 1996; Vol. 21; pp 135-
215.
(5) Toda, F. Acc. Chem. Res. 1995, 28, 480.
(6) (a) Vaida, M.; Popovitz-Biro, R.; Leiserowitz, L.; Lahav, M. In
Photochemistry in Organized and Constrained Media; Ramamurthy, V.,
Ed.; Verlag Chemie: New York, 1991; Chapter 6. (b) Scheffer, J. R.; Garcia-
Garibay, M. In Photochemistry on Solid Surfaces; Anpo, M., Matsuura, T.,
Eds.; Elsevier: Amsterdam, 1989; Chapter 9.3. (c) Caswell, L.; Garcia-
Garibay, M. A.; Scheffer, J. R.; Trotter, J. J. Chem. Educ. 1993, 70, 785.
(d) Leibovitch, M.; Olovsson, G.; Scheffer, J. R.; Trotter, J. Pure Appl.
Chem. 1997, 69, 815.
(1) (a) Inoue, Y. Chem. ReV. 1992, 92, 741. (b) Rau, H. Chem. ReV.
1983, 83, 535.
(2) (a) Kagan, H. B.; Fiaud, J. C. Top. Stereochem. 1988, 18, 249. (b)
Suarez, M.; Schuster, G. B. J. Am. Chem. Soc. 1995, 117, 6732.
(3) (a) Inoue, Y.; Tsuneishi, H.; Hakushi, T.; Tai, A. J. Am. Chem. Soc.
1997, 119, 472. (b) Tsuneishi, H.; Hakushi, T.; Tai, A.; Inoue, Y. J. Chem.
Soc., Perkin Trans. 2 1995, 2057.
10.1021/ja9823013 CCC: $15.00 © 1998 American Chemical Society
Published on Web 12/01/1998

12756 J. Am. Chem. Soc., Vol. 120, No. 49, 1998
LeiboVitch et al.
The major obstacle to be overcome in the solid-state method
of asymmetric synthesis is how to ensure that the reactants adopt
the requisite chiral space groups. Spontaneous crystallization
of achiral compounds in chiral space groups, while well
documented,7 is rare and unpredictable and therefore unsuitable
as a general method of crystal engineering asymmetric environ-
ments for photochemistry. The approach we have taken recently
to solve this problem involves the use of ionic chiral auxiliaries.8
In this procedure, a crystalline salt is formed between a prochiral,
carboxylic acid-containing photoreactant and an optically pure,
nonabsorbing amine. Since the ionic auxiliarysthe cation in
this casesis optically active, such salts are required to crystallize
in chiral space groups, and this provides the asymmetric medium
in which to carry out the photoreaction. Clearly, the opposite
approach, in which the cationic component of the salt is the
photoreactant and the anion is the chiral auxiliary, is equally
valid. One of the limitations of the solid-state ionic chiral
auxiliary technique is that the reactants must contain acidic or
basic functional groups that allow salt formation. On the other
hand, there are several advantages to working with salts, one
being that their high melting points and strong lattice forces
compared to purely molecular crystals permit reactions to be
carried to higher conversions without crystal breakdown. A
second advantage is that the photolyses can be carried out with
polycrystalline samples on a gram scale if desired; there is no
need to use special seeding techniques or carefully grown,
enantiomorphously pure single crystals, as in the case of achiral
compounds that crystallize spontaneously in chiral space groups.
Scheme 1
promising candidate in a paper by Lewis and co-workers.12a In
1974, these authors showed that solution-phase irradiation of
cis-4-tert-butyl-1-methylcyclohexyl phenyl ketone (1c, Scheme
1) afforded excellent chemical yields of cyclobutanol 2c. This
reaction seemed nearly ideal for our purposes since (i) it converts
an achiral starting material into a chiral photoproduct, (ii) only
one of the two possible diastereomers of cyclobutanol 2c is
formed, thus simplifying photoproduct separation and analysis,
(iii) for attachment of the ionic chiral auxiliaries, it appeared
that carboxylic acid-substituted derivatives of ketone 1c (e.g.,
1f) could be readily prepared, and (iv) it seemed likely that
ketone 1c and related compounds would crystallize in confor-
mations favorable for γ-hydrogen atom abstraction. Organic
molecules generally crystallize in their lowest energy conforma-
tions, and MM3 studies showed that ketone 1c has a global
minimum energy conformation in which the ketone oxygen is
well oriented to abstract one of the enantiotopic γ-hydrogens
Hx or Hy.
In developing crystal structure-solid-state reactivity relation-
ships, it is desirable to have a series of closely related
compounds to study, since this permits general trends to be
detected. Accordingly, ketones 1a-i (Scheme 1) were prepared
by standard procedures that are outlined in the Experimental
Section. Of these, compounds, 1a and 1c were previously
known;12a,13a,b the remainder represent new compounds that are
reported here for the first time. We also thought it worthwhile
to extend the study to analogues of ketones 1a-i and chose as
our target molecules the 2-methyl-2-adamantyl phenyl ketones
5a-g (Scheme 1). On the basis of the obvious similarity
between ketones of general structure 1 and 5, we anticipated
that the latter would also undergo Yang photocyclization to
afford cyclobutanols 6a-g, and this proved to be the case. The
synthesis of ketones 5a-g, all of which are new, is outlined in
the Experimental Section. All 16 ketones prepared for the
present study were crystalline, and of these, 10 had their crystal
and molecular structures determined (ketones 1a, 1b, 1c, 1f,
1h, 1i, 5a, 5c, 5d, and 5g); details of the crystal structure
determination for ketone 1c have been published.13c None of
the ketones whose crystal structures were determined crystallized
In this paper, we describe the application of the solid-state
ionic chiral auxiliary approach to asymmetric induction in the
well-known Yang photocyclization reactionsthe photochemical
conversion of γ-hydrogen-containing carbonyl compounds into
cyclobutanols.9,10 As part of these studies, X-ray crystal
structures were obtained for many of the starting salts as well
as for some of the photoproducts. Since the absolute configura-
tions of the ionic chiral auxiliaries are known, this allowed the
absolute enantioselectivities of the photoreactions to be deter-
mined, and this in turn enabled us to identify unequivocally
which γ-hydrogen was abstracted and to correlate the observed
abstraction/cyclization reactions with crystallographically de-
rived geometric data. Such correlations provide important
insights into the stereoelectronic requirements for γ-hydrogen
abstraction and 1,4-biradical closure.11
Results
Choice and Synthesis of Starting Materials. In searching
the photochemical literature for Yang photocyclization reactions
to which the ionic chiral auxiliary approach to solid-state
asymmetric synthesis could be applied, we came across a
(7) Jacques, J.; Collet, A.; Wilen, S. H. Enantiomers, Racemates and
Resolutions; Wiley: New York, 1981; pp 14-18.
(8) Gamlin, J. N.; Jones, R.; Leibovitch, M.; Patrick, B.; Scheffer, J. R.;
Trotter, J. Acc. Chem. Res. 1996, 29, 203.
(9) Cyclobutanol products in type II photochemistry were first reported
by: Yang, N. C.; Yang, D. H. J. Am. Chem. Soc. 1958, 80, 2913.
(10) For general reviews of the Norrish/Yang type II reaction, see: (a)
Wagner, P.; Park, B.-S. In Organic Photochemistry; Padwa, A., Ed.; Marcel
Dekker: New York, 1991; Vol. 11, Chapter 4. (b) Wagner, P. J. In
Molecular Rearrangements in Ground and Excited States; de Mayo, P.,
Ed.; Academic Press: New York, 1980; Chapter 20. (c) Wagner, P. J. Acc.
Chem. Res. 1971, 4, 168.
(11) Portions of this work have appeared in preliminary communication
form: (a) Leibovitch, M.; Olovsson, G.; Sundarababu, G.; Ramamurthy,
V.; Scheffer, J. R.; Trotter, J. J. Am. Chem. Soc. 1996, 118, 1219. (b)
Leibovitch, M.; Olovsson, G.; Scheffer, J. R.; Trotter, J. J. Am. Chem. Soc.
1997, 119, 1462.
(12) (a) Lewis, F. D.; Johnson, R. W.; Johnson, D. E.; J. Am. Chem.
Soc. 1974, 96, 6090. (b) MM3 calculations indicate that, in the adamantane
series, endo-arylcyclobutanols 6 are more stable than their exo isomers 7
by ∼14 kJ/mol. In the tert-butyl series, on the other hand, cyclobutanols 2
and 3 are approximately isoenergetic as a result of an unfavorable endo-
phenyl/tert-butyl steric interaction.
(13) (a) McIntosh, C. L. Can. J. Chem. 1967, 45, 2267. (b) Padwa, A.;
Eastman, D. J. Am. Chem. Soc. 1969, 91, 462. (c) Fu, T.-Y.; Scheffer, J.
R.; Trotter, J. Acta Crystallogr. 1997, C53, 1255.

Yang Photocyclization Reaction in the Solid State
J. Am. Chem. Soc., Vol. 120, No. 49, 1998 12757
Table 1. Results of Photolysis of Ketones 1a-i and 5a-g in the Solid State and Solution
ketone
R
X
medium
cleavage (4 or 8)
endo-arylcyclobutanol (2 or 6)
exo-arylcyclobutanol (3 or 7)
1a
H
H
F
benzenea
crystal
100
no rxn
100
no rxn
0
no rxn
0
no rxn
100
100
94
100
100
100
93b
100
93
0
no rxn
1b
1c
1d
1e
1f
H
benzene
crystal
0
no rxn
Me
Me
Me
Me
Me
Me
Me
H
H
F
benzenea
crystal
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
6
0
benzene
crystal
benzene
crystal
acetone
crystal
acetone
crystal
benzene
crystal
benzene
crystal
benzene
crystal
benzene
crystal
benzene
crystal
benzene
crystal
benzene
crystal
benzene
crystal
CN
0
0
CO2H
CO2Me
Me
7b
0
7
0
0
0
8
0
1g
1h
1i
100
100
100
92
OMe
H
100
no rxn
no rxn
no rxn
no rxn
100
100
100
100
100
100
100b
100
100
100
5a
5b
5c
5d
5e
5f
no rxn
no rxn
no rxn
no rxn
no rxn
no rxn
no rxn
no rxn
H
F
Me
Me
Me
Me
Me
H
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
F
CN
CO2H
CO2Me
5g
benzene
crystal
a Results originally reported by Lewis et al.12a b Carboxylic acid-containing photoproducts 2f and 6f were characterized as their methyl esters
2g and 6g following diazomethane workup of the photolysis mixtures.
in chiral space groups, thus precluding absolute asymmetric
synthesis studies.
Since the endo-phenyl stereochemical assignment by Lewis
et al. in the case of photoproduct 2c was made primarily on the
basis of a 90-MHz 1H NMR spectrum,12a we thought it
worthwhile to establish the stereochemical assignments on a
firmer basis. To this end, we were eventually successful in
growing crystals and obtaining the X-ray crystal structure of
one of the endo-aryl photoproducts, 2i. With this structure
firmly established, it was then relatively straightforward to show
Photolysis of Ketones 1a-i and 5a-g in the Solid State
and Solution; Characterization of the Photoproducts. Table
1 summarizes the results of Pyrex-filtered irradiation of ketones
1a-i and 5a-g in the solid state and solution. The techniques
used for the solid-state irradiations have been described previ-
ously.14 As expected, and in agreement with the findings of
Lewis et al. for ketone 1c,12a endo-arylcyclobutanols of general
structure 2 or 6 were obtained as the major photoproducts for
ketones in which R ) CH3; small amounts (<10%) of the
diastereomeric exo-arylcyclobutanols 3 or 7 were obtained in
the solution photolyses of some of these compounds, but never
in the crystalline state, even at very high conversions. The
situation for ketones in which R ) H, however, was quite
different. Earlier work by Lewis et al.12a had shown that
photolysis of ketone 1a in benzene led exclusively to the type
II cleavage product 1-phenyl-5-tert-butyl-6-hepten-1-one (4a),
and we obtained identical results in the case of ketone 1b.
Interestingly, ketones 1a and 1b were found to be photoinert in
the crystalline state, and in the adamantane series, ketones 5a
and 5b were found to be completely unreactive in solution as
well as in the solid state.15,16
1
by detailed comparisons of the 500-MHz H and 13C NMR
spectra that photoproducts 2c, 2d, 2e, 2g, and 2h also have the
endo-aryl stereochemistry. A key piece of evidence relating to
stereochemistry was the observation of a strong nuclear Over-
hauser effect between the o-aryl hydrogens and the methylene
hydrogens on the 3-carbon bridge. This effect was absent in
the exo-aryl photoproducts of general structure 3 and was
replaced by an NOE interaction between the ortho hydrogens
and the nearer methylene hydrogen on the one carbon bridge.
The NMR data are given in full in the Experimental Section.
A similar procedure was followed for the endo-arylcyclo-
butanol photoproducts 6c, 6d, 6e, and 6g in the adamantane
series. Here, crystals were grown of compound 6d, its X-ray
crystal structure was successfully determined, and the similarity
of the 500-MHz NMR spectra of all four compounds established
their common structure and stereochemistry.12b
Solution-phase irradiation of cis-4-tert-butyl 1-methylcyclo-
hexyl aryl ketones 1d, 1f, 1g, and 1e gave small amounts
(<10%) of a second photoproduct in each case to which the
exo-arylcyclobutanol structure 3 was assigned. The isolation
of these photoproducts in pure form was difficult owing to the
small amounts of material involved and to the fact that they
and their endo epimers have very similar column chromato-
graphic retention times. Repeated chromatography ultimately
(14) Gudmundsdottir, A. D.; Lewis, T. J.; Randall, L. H.; Scheffer, J.
R.; Rettig, S. J.; Trotter, J.; Wu, C.-H. J. Am. Chem. Soc. 1996, 118, 6167.
(15) Ketones 5a and 5b were photoinert even in tert-butyl alcohol, which
Wagner had shown increases type II quantum yields through hydrogen
bonding to the 1,4-hydroxy biradical and retardation of reverse hydrogen
transfer. See: Wagner, P. J.; Kochevar, I. E.; Kemppainen, A. E. J. Am.
Chem. Soc. 1972, 94, 7489.
(16) Attempted trapping of the 1,4-hydroxy biradicals from ketones 5a
and 5b by using high concentrations of deuterated n-heptanethiol according
to the procedure of Wagner was unsuccessful: Wagner, P. J.; Kelso, P. A.;
Zepp, R. G. J. Am. Chem. Soc. 1972, 94, 7480.

12758 J. Am. Chem. Soc., Vol. 120, No. 49, 1998
LeiboVitch et al.
Table 2. Quantum Yields and Kinetic Parameters for
Solution-Phase Photolysis of Ketones 1a-i and 5c-ga
in reasonable agreement with the value of 58 ns determined in
benzene via the Stern-Volmer method using 2,5-dimethyl-2,4-
hexadiene as the quencher.
c
ketone
conc, M
Φb
Φmax
kqτ, M-1d
τ, 10-9 se
Preparation of Chiral Salts of Keto Acids 1f and 5f. To
introduce the ionic chiral auxiliaries needed for solid-state
asymmetric induction, keto acids 1f and 5f were treated with a
variety of optically active amines and the resulting salts fully
characterized by conventional methods. The amines were
chosen more or less at random from the chiral pool, major
criteria being that they should not absorb under the photolysis
conditions and that they should be readily available. A total of
17 separate salts was prepared, 11 with keto acid 1f and 6 with
5f. Table 3 lists the amines used along with the melting points
and crystal morphologies of the corresponding salts. The salts
were shown to have 1:1 stoichiometry by elemental analysis.
1a
1b
1c
1d
1e
1f
1g
1h
1I
5c
5d
5e
5f
0.040
0.032
0.043
0.030
0.030
0.030
0.030
0.030
0.030
0.031
0.036
0.023
0.015
0.014
0.0095
0.0022
0.074
0.049
0.020
0.36
0.21
0.29
0.24
0.17
0.22
0.22
0.18
0.31
0.26
0.16
0.69
0.24
0.17
0.42
0.20
8.1
6.1
13
0.084
0.040
1.6
1.2
2.6
0.047
0.067
0.083
0.15
0.10
0.41
9.6
28
292
1.6
1.8
4.5
1.9
5.6
58
0.31
0.36
0.90
5g
0.073
2.1
0.41
1
In addition, the IR and H and 13C NMR spectra were all in
a Valerophenone (0.10 M) in benzene was used as the actinometer
in parallel irradiations at 313 nm. b Determined in purified benzene.
c Determined in 67% tert-butyl alcohol/benzene (w/w). d 2,5-Dimethyl-
2,4-hexadiene quencher. e Calculated by assuming kq ) 5 × 109 M-1
agreement with the proposed structures. X-ray diffraction
analysis was carried out on four of the salts (S9, S14, S15, S17)
to reveal additional structural features that will be discussed
later in the paper.
s-1
.
Photolysis of Chiral Salts in the Crystalline State: Asym-
metric Induction Studies. Each of the salts listed in Table 3
was irradiated in the crystalline state. Photolyses were generally
carried out through Pyrex at room temperature, although some
runs were conducted at -40 °C. No color change or melting
of the crystals was observed during the reaction, and following
irradiation, the mixtures were dissolved in ethyl acetate and
treated with excess ethereal diazomethane to form the corre-
sponding methyl esters 2g and 6g as the only detectable
products. Following diazomethane treatment, the organic layer
containing the esterified photoproduct and starting material was
washed with water and subjected to short-path silica gel
chromatography in order to remove the chiral auxiliary. The
mixtures were then analyzed by GC for extent of conversion
and by chiral HPLC for enantiomeric excess (ee). The salts
were irradiated for varying lengths of time in order to determine
the dependence of the ee values on the extent of conversion.
The results are summarized in Tables 4 and 5; only those salts
that gave ee’s of >50% in low-conversion runs are reported.
In addition to the solid-state runs, each of the salts was
photolyzed in solution (acetonitrile or methanol); in every case,
racemic photoproducts were obtained.
Geometric Data Derived from X-ray Crystallography. To
develop crystal structuressolid-state reactivity relationships, it
is necessary at this point to define some geometric parameters
associated with the Norrish/Yang type II reaction and to tabulate
the value of these parameters for the compounds described
above.
As depicted in Figure 1a, the hydrogen abstraction geometry
can be described by four parameters: d (the CdO‚‚‚Hγ
distancesexpected to be optimal at 2.72 Å, the sum of the van
der Waals radii), ω (the angle by which the γ-hydrogen deviates
from the mean plane of the carbonyl group), ∆ (the CdO‚‚‚Hγ
angle), and θ (the C-Hγ‚‚‚O anglesaccording to theory,20 best
at 180°). Assuming abstraction by the nonbonding orbital on
oxygen,21 the optimum values of ω and ∆ are expected to be 0
and 90-120° respectively.22 Table 6 summarizes the values
gave a pure sample of exo-arylcyclobutanol 3d, but the others
were characterized as mixtures and the structural assignments
in these cases must remain tentative, although we are quite
certain they are correct. Photoproduct 3d is an oil, and it was
therefore characterized primarily by high-resolution NMR
(including 500-MHz COSY, HMQC, HMBC, and NOE experi-
ments) as well as by IR and MS. Details are provided in the
Experimental Section.
Quantum Yield and Stern-Volmer Kinetic Studies. With
the use of a merry-go-round apparatus17 and valerophenone
actinometry,15 quantum yields of photoproduct formation were
measured for ketones 1a-i and 5c-g in both benzene (Φ) and
67% tert-butyl alcohol/benzene (Φmax) (Table 2). Determination
of the quantum yields in the presence of varying amounts of
the triplet quencher 2,5-dimethyl-2,4-hexadiene gave linear
Stern-Volmer plots18 with slopes equal to kqτ, and from these
data, the values of τ, the triplet excited-state lifetimes, could
be estimated. These numbers are also compiled in Table 2.
The kqτ values for ketones 1a and 1c determined by Lewis
et al. in 1974 (0.74 and 7.3 M-1, respectively)12a differ somewhat
from our values of 0.42 and 8.1 M-1. These differences are
probably due to the different quenchers used by Lewis et al.
(trans-1,3-pentadiene for 1a, naphthalene for 1c) as well as to
the use of benzophenone/benzhydrol rather than valerophenone
actinometry. The values of Φmax measured by Lewis et al. for
the same two compounds (0.098 for 1a and 0.06 for 1c) also
differ from ours (0.36 and 0.29, respectively). Lewis numbers
were determined in 8.9 M 1-propanol/benzene, whereas ours
were measured in 67% tert-butyl alcohol/benzene (w/w), the
solvent system recommended by Wagner et al. for such
purposes.19
As an independent check of the accuracy of our steady-state
kinetic data, we determined the triplet lifetime of one of our
ketones by laser flash photolysis. The p-methoxyphenyl deriva-
tive 1i was the only compound whose triplet was sufficiently
long-lived to be measurable with the nanosecond setup available
to us. By Stern-Volmer analysis using 1,3-cyclohexadiene as
the triplet quencher, a lifetime of 77 ns was obtained. This is
(20) Dorigo, A. E.; Houk, K. N. J. Am. Chem. Soc. 1987, 109, 2195 and
references therein.
(17) Moses, F. G.; Liu, R. S. H.; Monroe, B. H. Mol. Photochem. 1969,
1, 245.
(18) Wagner, P. J. In Creation and Detection of the Excited State;
Lamola, A. A., Ed.; Marcel Dekker: New York, 1971; pp 174-212.
(19) Wagner, P. J.; Meador, M. A.; Zhou, B.; Park, B.-S. J. Am. Chem.
Soc. 1991, 113, 9630.
(21) Wagner, P. J. In Rearrangements in Ground and Excited States; de
Mayo, P., Ed.; Academic Press: New York, 1980; Vol. 3; p 405. See also:
Turro, N. J. Modern Molecular Photochemistry; Benjamin-Cummings:
Menlo Park, CA, 1978; Chapter 10 and references therein.
(22) Scheffer, J. R. In Organic Solid State Chemistry; Desiraju, G. R.,
Ed.; Elsevier: Amsterdam, 1987; pp 1-45.

Yang Photocyclization Reaction in the Solid State
J. Am. Chem. Soc., Vol. 120, No. 49, 1998 12759
Table 3. Preparation of Optically Active Salts from Keto Acids 1f and 3f
entry
keto acid
amine
recryst solvent
CH3CN
CH3CN or sublimation
EtOH
CH3CN
CH3CN
CH3CN
CH3CN
CH3CN/EtOH
CH3CN
CH3CN/MeOH
EtOH
CH3CN
CH3CN
CH3CN/EtOH
CH3CN/EtOH
CH3CN
cryst morphol
powder
long, thin plates
powder
long, thin flat needles
long, thin flat needles
long, thin plates
powder
mp (°C)
S1
S2
S3
S4
S5
S6
S7
S8
1f
1f
1f
1f
1f
1f
1f
1f
1f
1f
1f
5f
5f
5f
5f
5f
5f
(S)-(+)-prolinol
(1S,2S)-(+)ψ-ephedrine
(S)-(+)-arginine
(S)-(-)-R-methyl benzylamine
(R)-(+)-R-methylbenzylamine
(1R,2S)-(-)-ephedrine
(S)-(-)-proline tert-butyl ester
(1R,2S)-(-)-norephedrine
(S)-(-)-prolinamide
(1S)-(-)-2,10-camphorsultam
(S)-(+)-lysine
(S)-(+)-prolinol
(1S,2S)-(+)-ψ-ephedrine
(S)-(-)-R-methylbenzylamine
(R)-(+)-R-methylbenzylamine
(1R,2S)-(-)-ephedrine
(1R,2S)-(-)-norephedrine
167-169
151-154
222-224
224-225
224-225
179-181
227-229
171-174
147-148
219-225
204-206
133-135
158-161
210-212
210-212
186-188
147-149
long, flat plates
S9
platesa
S10
S11
S12
S13
S14
S15
S16
S17
flakes
light yellow plates
powder
long, thin plates
long, thin flat needlesa
long, thinflat needlesa
long, thin plates
long, flat platesa
CH3CN/EtOH
a X-ray crystal structure obtained.
Table 4. Asymmetric Induction in the Solid-State Photochemistry of Salts of Keto Acid 1f
salt
amine
photolysis methoda and temp
slides, RTc
irradn time (min)
convrsn (%)
ee (%)b 2g
S2
(1S,2S)-(+)-ψ-ephedrine
5
20
173
2
18
82
3
(-)-99.5
97.3
63.5
tube, RT
7.5
(-)-99.5
73
90
120
480
5
15
42
2
40
61
59
87
12
29
74
0.7
24
44
92
7
24
66
85
0.4
1.0
26
65
3
79.2
72.7
71.0
63.5
S3
S4
(S)-(+)-arginine
slides, RT
slides, RT
(+)-90.5
84.6
83.0
(S)-(-)-R-methylbenzylamine
(+)-97.3
6
95.8
14
52
5
15
25
45
9
21
103
200
69
160
90.4
84.8
S5
S6
(R)-(+)-R-methylbenzylamine
(1R,2S)-(-)-ephedrine
slides, RT
slides, RT
(-)-96.9
96.5
94.3
92.1
(+)-78.5
77.2
67.3
58.7
slides, -40 °C
(+)-81.7
8
78.6
S7
(S)-(-)-proline tert-butyl ester
slides, RT
0.5
2
5
0.6
7
(-)-52.0
51.0
19
53
93
11
32
8
34
58
97
7
51
83
0.5
12
78
46.6
25.5
2.8
15
42
21
69
5
15
42
90
6
30
52
2
6
52
slides, -40 °C
(-)-55.2
50.4
S8
S9
(1R,2S)-(-)-norephedrine
(S)-(-)-prolinamide
slides, RT
(+)-96.5
95.8
86.4
81.1
slides, RT
tube, RT
(+)-86.5
68.2
28.7
(+)-85.2
82.4
33.4
a Slides means that the salt was crushed between two Pyrex microscope slides and irradiated; tube means that the salt was irradiated in an NMR
tube, so that the crystal morphology remained intact. b The first enantiomer of photoproduct 2g eluted from the Chiralcel OD column has a (-)
rotation at the sodium D line. c RT, room temperature.
of these four parameters for the 14 ketones and salts whose
crystal structures were determined.
Since hydrogen abstraction in the solid state is likely to occur
with a minimum of motion in the carbon framework, geometric
data derived from X-ray crystallography can also be used to
analyze the behavior of the 1,4-hydroxy biradical intermediate.
For this purpose, we define two torsion angles with reference
to Figure 1b, which is a depiction of the 1,4-biradical intermedi-

12760 J. Am. Chem. Soc., Vol. 120, No. 49, 1998
LeiboVitch et al.
ee (%)b 6g
Table 5. Asymmetric Induction in the Solid-State Photochemistry of Salts of Keto Acid 5f
entry
amine
photolysis methoda and temp
irradn time (min)
convrsn (%)
S14
(S)-(-)-R-methylbenzylaminec
slides, RT
2
5
1.8
13.5
39.5
64.0
89.4
42.6
99.9
8.4
(+)-87.5
87.5
10
20
47
10
96
30
90
2
86.9
86.0
81.0
tube, RT
(+)-88.5
79.9
slides, -40 °C
slides, RT
(+)-89.4
88.2
19.9
3.5
S15
S17
(R)-(+)-R-methylbenzylaminec
(-)-88.0
4
8
17.9
37.8
60.5
82.3
33.9
99.9
10.5
22.6
1.8
86.9
86.4
86.2
82.4
19
43
10
96
30
90
2
tube, RT
(-)-87.8
78.9
slides, -40 °C
slides, RT
(-)-87.9
87.3
(1R,2S)-(-)-norephedrine
(-)-88.0
4
2.9
87.8
15
41
7
96
210
30
90
11.6
26.8
4.7
43.3
92.4
1.9
81.9
74.8
tube, RT
(-)-87.8
73.7
69.8
slides, -40 °C
(-)-88.2
7.8
87.9
a Slides means that the salt was crushed between two Pyrex microscope slides and irradiated; tube means that the salt was irradiated in an NMR
tube, so that the crystal morphology remained intact. b The first enantiomer of photoproduct 6g eluted from the Chiralcel OD column has a (+)
rotation at the sodium D line. c This salt undergoes a single crystal-to-single crystal (topotactic) reaction. See text for details. d RT, room temperature.
Table 6. Crystallographically Derived Hydrogen Abstraction and
Biradical Parametersa
ketone
or salt
d
ω
(deg)
θ
(deg)
∆
(deg)
æ1
(deg)
æ4
(deg)
D
(Å)
(Å)b
1a
2.61
3.17
2.59
3.17
2.70
2.65
2.61
3.44
2.69
2.63
3.49
2.47
3.13
2.50
3.41
2.54
3.47
2.58
2.75
2.62
2.60
2.55
3.47
32
57
32
57
57
55
52
48
57
56
45
29
57
57
44
61
41
61
60
62
62
60
42
116
113
116
113
114
113
114
108
112
113
108
117
113
117
110
116
109
115
112
114
114
116
109
95
63
96
63
81
85
88
49
83
85
46
98
64
84
44
80
41
79
77
75
75
80
41
85
-30
84
-29
64
68
72
-47
57
68
-51
82
-27
66
-53
61
-58
61
-59
-55
55
61
-57
36
-31
36
-31
33
35
35
-31
35
35
-31
31
-30
29
-28
29
-29
30
-33
-38
27
29
-28
3.23
3.15
3.23
3.15
3.09
3.11
3.12
3.12
3.11
3.10
3.11
3.17
3.12
3.02
3.05
2.99
3.04
3.00
3.06
2.96
2.95
3.00
3.05
1b
1c
Figure 1. (a) Definition of geometric parameters d, ω, ∆, and θ for
intramolecular γ-hydrogen atom abstraction. (b) 1,4-Hydroxy biradical
intermediate derived from ketones of general structures 1 and 5.
1fc
1h
1i
ate derived from ketones of general structure 1 and 5: æ1 is
defined as the angle that the shaded lobe of the p-orbital on C1
makes with respect to the C2-C3 bond), and æ4 is the angle
that the shaded lobe of the p-orbital on C4 makes with respect
to the C2-C3 bond. The p-orbitals at C1 and C4 are assumed
to lie perpendicular to the O1-C1-C2 and C3-C4-C5 planes,
respectively. The basis of this analysis is the generally accepted
idea that cleavage of 1,4-biradicals requires good overlap
between the radical-containing p-orbitals and the C2-C3 bond,
i.e., æ1 ) æ4 ∼ 0°.10 Failing that, and provided that C1 and
C4 are within reasonable bonding distance, cyclization will
predominate. The C1‚‚‚C4 interatomic distance is given the
symbol D. The crystallographically determined values of æ1,
æ4, and D are given in Table 6.
5a
5c
5d
5g
S9
S14
S15
S17
a See text for definition of parameters. b γ-Hydrogen atoms for which
d g 3.50 Å are not included in this table. c Two independent molecules
in the asymmetric unit.
Discussion
concordance with previous investigations, there are still some
questions raised by the present study that need answering. These
include the following: (1) for compounds that undergo Yang
photocylization, why are the endo-arylcyclobutanols 2 and 6
invariably favored over their exo isomers 3 and 7? (2) Why
are ketones 1a and 1b photochemically reactive in solution and
not in the solid state and why are they the only compounds to
undergo cleavage rather than Yang photocyclization? (3) Why
The photochemical and photophysical results with ketones
1a-i and 5a-g (Tables 1 and 2) conform very closely to the
standard picture of the Norrish/Yang type II reaction of aryl
alkyl ketones, viz., hydrogen atom abstraction from a triplet
excited state (presumably n,π* in nature) to produce a triplet
1,4-hydroxy biradical that partitions itself among closure,
cleavage, and return to starting material.10 Despite this general

Yang Photocyclization Reaction in the Solid State
J. Am. Chem. Soc., Vol. 120, No. 49, 1998 12761
Scheme 2
type II cyclization over cleavage are well documented in acyclic
ketones, and this effect has been termed the “R-alkyl group
effect”.24
An additional 90° rotation about the C1-C2 bond leads to
biradical 12, closure of which forms exo-arylcyclobutanols 3
and 7. It is paradoxical that small amounts of these compounds,
but not cleavage products, are formed in solution when R )
CH3: after all, both processes involve unfavorable rotations
about the C1-C2 bond. The explanation could be that the triplet
biradical spends very little time near eclipsed conformation 11,
the result being that intersystem crossing to the singlet manifold
and formation of cleavage products from this geometry cannot
compete with rotation.25 In addition, overlap between the
radical-containing orbitals is less in biradical 11 than 10, and
this may reduce the rate of intersystem crossing (and reaction)
in the former.26 In any event, it is clear that, in the crystalline
state, large-amplitude motions of the aryl and hydroxyl groups
accompanying rotation about the C1-C2 bond would be
topochemically forbidden and have no chance of competing with
closure to endo-arylcyclobutanols. It is therefore perfectly
reasonable that these are the only products formed in this
medium.
But why do ketones lacking R-methyl groups (i.e., 1a, 1b,
5a, and 5b) fail to photocyclize in the solid state? The X-ray
crystal structure data for the first three of these (crystals of 5b
were not suitable for crystallography) suggest that part of the
answer may lie in the distance between the radical centers in
biradical 10. The data in Table 6 show that the average C1-
C4 distance D for compounds 1a, 1b, and 5a is 3.21 Å (for
ketone 5b, the MM3 calculated value for D is 3.15 Å), whereas
the average distance for ketones with R-methyl groups that do
photocyclize is 3.04 Å. This difference is caused by a steric
interaction between the R-methyl group and the aryl group that
“pushes” C1 toward C4. For example, the C1-C2-C3 angle
in the nonmethylated compounds is 114-115°, whereas it is
108-109° for the methylated derivatives.
are ketones 5a and 5b photochemically unreactive in both the
solid state and solution? The answers to these questions clearly
have something to do with the nature of the R-substituent, R,
which is a hydrogen atom for the anomalous compounds 1a,
1b, 5a, and 5b and a methyl group for the ketones that undergo
Yang photocyclization.
To address these points, consider the idealized reaction
pathway shown in Scheme 2. X-ray crystallography and MM3
calculations are in agreement that, regardless of the nature of
the R-substituent, ketones of type 1 and 5 adopt chair conforma-
tions such as that shown by structure 9, in which the aroyl group
is axial and the mean plane of the carbonyl group is roughly
orthogonal to the plane bisecting the cyclohexane ring between
γ-hydrogen atoms Hx and Hy. As a result, one of these
hydrogens (Hx in the case of structure 9) is much more favorably
oriented for abstraction than than the other, a fact that is apparent
from the d values in Table 6, which range from 2.47 to 2.70 Å
for the nearer hydrogens and are g3.13 Å for the more distant
ones. As will become apparent later, we can state with virtually
complete certainty that (1) it is the closer γ-hydrogen that is
abstracted in the case of ketones of type 1 and 5 and (2) the
failure of any of these compounds to form cyclization and/or
cleavage photoproducts is not associated with an unfavorable
hydrogen abstraction geometry.
Orbital overlap is a second factor that may contribute to the
failure of the nonmethylated compounds to form cyclobutanols
in the crystalline state. With only a hydrogen atom in the
R-position to interfere with the aryl group, these ketones adopt
conformations in which the CdO group essentially eclipses the
C2-C3 bond (O1-C1-C2-C3 ≈ 0°). Owing to methyl/aryl
steric repulsion, the same angle in the R-methylated compounds
is ∼-30°. The result, therefore, is that for the nonmethylated
compounds the p-orbital on the carbonyl carbon is rotated away
from the C4 γ-hydrogen and the ensuing p-orbital at this site.
With cyclization thus slowed by increased distance and reduced
orbital overlap between the radical centers, reverse hydrogen
transferswhich is presumably as geometrically favorable as the
With reference to Scheme 2, therefore, 1,4-hydroxy biradical
10 will be the initially formed intermediate. This species is
poorly aligned for cleavage (from Table 6, æ1 ) 66.5 ( 9.8°,
æ4 ) 32.7 ( 3.2°), but with D ) 3.08 ( 0.09 Å, it is well set
up for closure, since this distance is less than the sum of the
van der Waals radii for two carbon atoms (3.40 Å) and there is
good orbital overlap. As a result, cyclization occurs with
“retention of configuration” at C1 and C4 and leads directly to
the experimentally observed major endo-aryl photoproducts 2
and 6.23 Cleavage, on the other hand, requires an approximately
90° rotation about the 1-2 bond of biradical 10 in order to
align the orbital on C1 with the C2-C3 bond (cf. structure 11),
and this motion is slowed when R ) CH3 (Ar/CH3 eclipsing)
but less so when R ) H. As a result, cleavage is seen only
when R ) H. Similar effects of R-substituents in promoting
(24) (a) Lewis, F. D.; Hilliard, T. A. J. Am. Chem. Soc. 1970, 92, 6672.
(b) Lewis, F. D.; Ruden, R. A. Tetrahedron Lett. 1971, 715. (c) Wagner,
P. J.; Kelso, P. A.; Kemppainen, A. E.; McGrath, J. M.; Schott, H. N.;
Zepp, R. G. J. Am. Chem. Soc. 1972, 94, 7506.
(25) For recent experimental evidence that geometry-dependent inter-
system crossing in triplet biradicals occurs dynamically during reaction and
determines product ratios, see: (a) Wagner, P. J.; Zand, A.; Park, B.-S. J.
Am. Chem. Soc. 1996, 118, 12856. For additional contributions to this topic,
see: (b) Griesbeck, A. G.; Mauder, H.; Stadtmu¨ller, S. Acc. Chem. Res.
1994, 27, 70. (c) Scaiano, J. C. Tetrahedron 1982, 38, 819.
(26) For recent updates of the original Salem-Rowland rules relating
T1-S0 spin-orbit coupling in biradicals to molecular structure and
conformation, see: (a) Michl, J. J. Am. Chem. Soc. 1996, 118, 3568. (b)
Kita, F.; Nau, W. M.; Adam, W.; Wirz, J. J. Am. Chem. Soc. 1995, 117,
8670. (c) Zimmerman, H. E.; Kutateladze, A. G. J. Am. Chem. Soc. 1996,
118, 249. For a recent experimental study in which conformation-dependent
enhanced spin-orbit coupling is suggested to control stereoselectivity in
triplet 1,4-biradical reactions, see: (d) Griesbeck, A. G.; Buhr, S.; Fiege,
M.; Schmickler, H.; Lex, J. J. Org. Chem. 1998, 63, 3847.
(23) Very similar structure-reactivity relationships for biradical cycliza-
tion were found in our recent study of the solid-state type II photochemistry
of a series of medium-sized ring and macrocyclic diketones.14 In this work,
stereoselective cyclization with “retention of configuration” at C1 and C4
was observed in biradicals for which æ1 ∼ 70° and æ4 ∼ 25° and D ∼ 3.15
Å.

12762 J. Am. Chem. Soc., Vol. 120, No. 49, 1998
LeiboVitch et al.
forward processspredominates for ketones 1a, 1b, 5a, and 5b,
and cyclization is not observed in the solid state.
Tables 4 and 5 also show that enantiomeric excess declines
with increasing conversion. This is not unexpected, since the
salts react to give products that presumably do not “fit” into
the original crystal lattice, and defect sites are generated at which
the normal topochemical restrictions are relaxed. Although In
some cases the decline in ee is precipitous, the pleasantly
surprising feature of the results is that for many salts ee’s of
80% or higher could be obtained at nearly complete conversion.
This clearly indicates the considerable synthetic potential of the
ionic chiral auxiliary approach to asymmetric induction. How
can high ee be maintained at high conversions? Formation of
the topochemically disallowed enantiomer requires a rotation
of ∼180° about the C1-C2 bond (structure 9, Scheme 2) prior
to hydrogen abstraction, and evidently this process is slow in
the solid state, even when the parent crystal lattice is nearly
destroyed.29 The solid-state medium therefore exerts two main
effects in these reactions: (1) prior to irradiation, it orders the
reactants in a single, homochiral conformation (9) that permits
abstraction of only one of two possible enantiotopic hydrogen
atoms, and (2) as the reaction proceeds and the parent crystal
lattice breaks down, the medium nevertheless remains suf-
ficiently viscous to prevent significant conversion of the original
conformer 9 to its enantiomer via rotation about the C1-C2
bond; specific, anisotropic crystal lattice effects need not be
invoked. An interesting question that remains unresolved,
however, is why some salts give much higher ee’s than others.
Topotactic (Single Crystal-to-Single Crystal) Solid-State
Photoreaction of Salts S14 and S15. During our photochemi-
cal studies, we noticed that the salts of ketoacid 5f with (S)-
(-)- and (R)-(+)-R-methylbenzylamine (S14 and S15, respec-
tively) did not change in appearance upon irradiation, even at
conversions as high as 90%. This suggested the occurrence of
a single crystal-to-single crystal (topotactic) process,30 and this
possibility was therefore investigated further. Accordingly, a
single crystal of salt S15 that had been previously subjected to
X-ray diffraction analysis was photolyzed for a time (2 h)
sufficient to ensure complete conversion to the corresponding
cyclobutanol, whereupon a second crystallographic data set was
collected. Both data sets refined successfully in space group
P212121 to final R values of 4.6 and 4.9%, thus indicating that
the solid-state photorearrangement is indeed a topotactic process.
Parts a and b of Figure 2 show the crystallographically
determined molecular structures of the salts prior to and
following irradiation, respectively. Figure 2a predicts prefer-
ential abstraction of Hx (2.60 Å) over Hy (3.59 Å), and in
agreement with this, the molecular structure of the photoproduct
salt (Figure 2b) shows that the carbon atom to which Hx was
attached is in fact part of the newly formed four-membered ring.
Finally, we address the question of why the adamantyl ketones
5a and 5b fail to to undergo type II cleavage in solution when
their cyclohexyl counterparts 1a and 1b react efficiently in this
manner. The likely answer is that the incipient C3-C4 double
bond cannot easily achieve planarity in the transition state when
it is part of an adamantane ring system. This apparently raises
the barrier for cleavage sufficiently to make reverse hydrogen
transfer the faster process.
Hydrogen Abstraction Geometry. It is interesting to
compare the hydrogen abstraction geometries given in Table 6
with those from previous studies. Our initial work centered
around compounds having the basic R-cycloalkylacetophenone
structure,22 and more recently, we have made a very thorough
study of the crystal structure-solid-state reactivity relationships
in the Norrish/Yang type II reactions of medium-sized ring and
macrocyclic diketones.14 For 17 examples of variously substi-
tuted R-cycloalkyacetophenones, the crystallographically de-
termined abstraction distance d was found to be 2.74 ( 0.16
Å, with ω ) 43 ( 9° and ∆ ) 84 ( 8° (θ was not calculated).22
For 11 photochemically reactive medium and large ring aliphatic
diketones whose crystal structures were determined, the average
value of d was 2.73 ( 0.03 Å, with ω ) 52 ( 5°, ∆ ) 83 (
4°, and θ ) 115 ( 2°.14 For the 14 compounds studied in the
present work, the corresponding values were d ) 2.61 ( 0.07
Å, ω ) 53 ( 11°, ∆ ) 84 ( 7°, and θ ) 115 ( 2°.
The fact that 42 compounds comprising three completely
different classes of ketone react through such similar solid-state
geometries argues very strongly that these data may be taken
as being generally indicative of successful γ-hydrogen atom
abstraction.27 The data obviously do not provide exact informa-
tion on the geometric situation in the excited state, where bond
lengths and angles can differ somewhat from those in the ground
state,28 but there is every reason to believe that the structure-
reactivity relationships developed above are approximately
correct and can now be used in a predictive sense. Molecular
mechanics programs provide rapid and reliable estimates of d,
ω, ∆, and θ, and this means that it should be possible to predict
the outcomes of unknown type II reactions including their
regioselectivity. Such work is under way in our laboratory at
the present time.
Asymmetric Induction Studies. As summarized in Tables
4 and 5, the ionic chiral auxiliary approach to solid-state
asymmetric synthesis works very well in the case of keto acids
1f and 5f. While not all of the optically active amines tested
gave excellent results, most worked reasonably well and some
were outstanding (pseudoephedrine, arginine, R-methylbenzyl-
amine, norephedrine). Enantioselectivity increases with de-
creasing temperature, but significantly longer irradiation times
are required. For both acids, photolysis of the (S)-(-)- and (R)-
(+)-R-methylbenzylamine salts afforded equal optical yields of
the enantiomers of cyclobutanols 2g and 6g (after diazomethane
workup), indicating that the system is well-behaved. None of
the salts gave any detectable asymmetric induction upon
photolysis in methanol or acetonitrile solution.
(29) It is interesting to speculate that the slightly lower ee’s in the case
of salts S14, S15, and S17 (Table 5) derive from the “spherical” shape of
the adamantane ring, which permits freer rotation about the C1-C2 bond
in the crystalline state.
(30) Topotactic solid-state chemical reactions are rare. The following
represent the majority of known examples. (a) Suzuki, T.; Fukushima, T.;
Yamashita, Y.; Miyashi, T. J. Am. Chem. Soc. 1994, 116, 2793. (b)
Enkelmann, V.; Wegner, G.; Novak, K.; Wagener, K. B. J. Am. Chem.
Soc. 1993, 115, 10390. (c) Novak, K.; Enkelmann, V.; Wegner, G.;
Wagener, K. B. Angew. Chem., Int. Ed. Engl. 1993, 32, 1614. (d) Vaida,
M.; Popovitz-Biro, R.; Leiserowitz, L.; Lahav, M. In Photochemistry in
Organized and Constrained Media; Ramamurthy, V., Ed.; Verlag Chemie:
New York, 1991; Chapter 6. (e) Ohashi, Y. Acc. Chem. Res. 1988, 21,
268. (f) Wang, W.-N.; Jones, W. Tetrahedron 1987, 43, 1273. (g) Tieke,
B. J. J. Polym. Sci., Polym. Chem. Ed. 1984, 22, 2895. (h) Miller, E.; Brill,
T. B.; Rheingold, A. L.; Fultz, W. C. J. Am. Chem. Soc. 1983, 105, 7580.
(i) Hasegawa, M. Chem. ReV. 1983, 83, 507. (j) Thomas, J. M. Nature
1981, 289, 633. (k) Thomas, J. M.; Morsi, S. E.; Desvergne, J. P. AdV.
Phys. Org. Chem. 1977, 15, 63. (l) Cheng, K.; Foxman, B. J. Am. Chem.
Soc. 1977, 99, 8102l (m) Wegner, G. Pure Appl. Chem. 1977, 49, 443. (n)
Osaki, K.; Schmidt, G. M. J. Isr. J. Chem. 1972, 10, 189.
(27) Similar geometric data have been published recently for the Norrish
type II reactions of phthalimidocyclohexanes. See: Griesbeck, A. G.; Henz,
A.; Kramer, W.; Wamser, P. Tetrahedron Lett. 1998, 39, 1549. In this work,
the success of hydrogen atom abstraction was found to depend critically
on the angle ∆. For a more thorough discussion of hydrogen atom abstraction
geometry, along with references to the work of others on this topic, see:
Ihmels, H.; Scheffer, J. R. Tetrahedron, in press.
(28) For theoretical treatments that provide insights into the transition-
state geometry of the type II process, see: (a) Dorigo, A. E.; McCarrick,
M. A.; Loncharich, R. J.; Houk, K. N. J. Am. Chem. Soc. 1990, 112, 7508.
(b) Sauers, R. R.; Edberg, L. A. J. Org. Chem. 1994, 59, 7061.

Yang Photocyclization Reaction in the Solid State
J. Am. Chem. Soc., Vol. 120, No. 49, 1998 12763
Figure 2. (a) Molecular structure of a crystal of salt S15 prior to irradiation. (b) Molecular structure of the same crystal following complete
conversion to the corresponding cyclobutanol salt. (c) Molecular structure of the mixed crystal composed of 40% of salt S15 and 60% of its
photoproduct. The thermal displacement ellipsoids in views (a) and (b) are drawn at the 30% probability level, 20% in view (c).
It was of interest to determine the structure of a partially
reacted crystal, and to this end, a single crystal of salt S15 was
photolyzed for 15 min and then subjected to X-ray diffraction
analysis. Once again the data set refined successfully in space
group P212121, but this time the final R value was significantly
higher (6.8%). GC analysis of the reaction mixture following
diazomethane workup indicated a conversion of ∼60%, and this
was supported by refinement of the atom occupancy factors in
the X-ray study.
Figure 2c shows an ORTEP diagram of the superposition of
salt S15 and its photoproduct in the mixed crystal. Although
the slightly larger values of the anisotropic displacement
parameters in the adamantane ring indicate a less than perfect
match in location of reactant and product, there is still a very
close atom-for-atom correspondence between the two, and this
is undoubtedly the source of the topotactic behavior: the
photoproduct fits nearly perfectly into the lattice site of the
reactant. The major difference between the two structures is
an upward movement of the γ-carbon atom to accommodate
the four-membered ring and a slight change in orientation of
the carbon-oxygen bond. These differences are indicated by
the shaded bonds in Figure 2c. Interestingly, the unit cell
volume is slightly greater for the mixed crystal (2319 Å3) than
for either the pure starting material (2301 Å3) or the pure
photoproduct (2310 Å3). Similar behavior has been seen in
other topotactic solid-state reactions.30b,k Recrystallization of
the “as-formed” photoproduct salt failed to give crystals suitable
for X-ray diffraction analysis. There were, however, slight
differences in the solid-state IR spectra and DSC traces before
and after recrystallization.
without loss of crystal quality, which raises the intriguing
possibility that salts in general will prove to be more topotac-
tically inclined than their nonionic counterparts. Further studies
with salts are required before this point can be settled. In the
meantime, the solid-state photochemistry of salts in which one
of the ions is a chiral auxiliary promises to become a viable
and general method of asymmetric synthesis. The present paper
contributes to the growing list of successful applications of this
method in organic photochemistry,8 and there is every reason
to believe that a similar approach will work for ground-state
asymmetric syntheses as well.31 Strategic planning along these
lines is underway in our laboratory at the present time.
In addition to the points raised above, the present work
provides further penetrating insight into the details of the
Norrish/Yang type II reaction. A general and consistent picture
is beginning to emerge as to the geometric requirements for
both stages of the reactionshydrogen atom abstraction as well
as the partitioning of the 1,4-hydroxy biradical intermediate.
Through molecular modeling, this now makes it possible to
predict the feasibility and outcome of hypothetical Norrish/Yang
type II reactions.
Experimental Section
General Information. Commercial spectral-grade solvents were
used for photochemical experiments unless otherwise noted. Infrared
spectra were recorded on a Perkin-Elmer 1710 Fourier transform
spectrometer. Solid samples were ground in KBr (1-5%) and pelleted
in an evacuated die (Perkin-Elmer 186-0002) at 17 000 psi. Liquid
samples were run neat as thin films between two salt plates. Melting
points were determined on a Fisher-Johns hot-stage apparatus and are
uncorrected. Nuclear magnetic resonance spectra were recorded in
deuterated solvents from Aldrich on 200-, 400-, and 500-MHz (Bru¨ker)
and 300-MHz (Varian) instruments. 13C spectra were run under broad-
band 13C{1H} decoupling and assignments, where given, were supported
by the attached proton test. For APT, (+) signifies C or CH2 and (-)
signifies CH or CH3. Low- and high-resolution mass spectra were
obtained from a Kratos MS 50 instrument at 70 eV. FAB mass spectra
were recorded on an AEI MS-9 spectrometer with xenon bombardment
Conclusions
In contrast to solid-state chemical reactions of nonionic
(molecular) crystals, which generally have to be carried out at
low temperatures and/or to low conversions owing to the
problem of crystal melting, the present work indicates that the
analogous processes of ionic crystals can be carried to much
higher conversions without significant loss of topochemical
control. Such reactions thus offer genuine prospects for
synthetic utility. In the case of the single crystal-to-single crystal
photoreaction of salt S15, quantitative conversions were possible
(31) The use of optically active quaternary ammonium salts to bring about
the enantioselective alkylation of achiral carbonyl compounds under basic
phase transfer catalysis conditions has been reported. See: Corey, E. J.;
Xu, F.; Noe, M. C. J. Am. Chem. Soc. 1997, 119, 12414 and references
therein.

12764 J. Am. Chem. Soc., Vol. 120, No. 49, 1998
LeiboVitch et al.
of an alcohol matrix as noted. Ultraviolet spectra were recorded on a
Perkin-Elmer Lambda-4B spectrometer in the solvents indicated.
Elemental analyses were performed by the departmental microanalyst,
Mr. P. Borda. Gas chromatographic analyses were performed on a
Hewlett-Packard 5890 instrument fitted with a flame ionization detector
and a Hewlett-Packard 3392 A integrator. The following 15 m × 0.25
mm fused-silica capillary columns from J & W Scientific, Inc. were
used: DB-1, DB-5, DB-17, and HP-5. A 20 m × 0.21 mm Carbowax
20M capillary column from Hewlett-Packard was also used on occasion.
Analytical thin-layer chromatography was carried out on commercial
precoated silica gel plates (E. Merck, type 5554). Column chroma-
tography was performed by using 230-400-mesh silica gel (Merck
9385) slurry packed with the eluting solvent. High-pressure liquid
chromatography was performed on a Waters 600E system controller
connected to a tunable UV detector (Waters 486) or programmable
photodiode array detector (Waters 994). For the determination of
enantiomeric excesses, a Chiralcel OD column (250 mm × 4.6 mm)
from Chiral Technologies, Inc. was used.
(CDCl3) δ 24.48 (+), 27.39 (-), 28.45 (-), 32.28 (+), 37.49(+), 47.57
(-), 48.45 (+), 114.83 and 115.27 (2JC-F ) 22 Hz, -ve), 130.01 and
130.18 (3JC-F ) 8 Hz, -ve), 135.39 and 135.45 (4JC-F ) 3 Hz, +ve),
161.60 and 166.63 (1JC-F ) 252 Hz, +ve), 207.69 (+); IR (KBr) 2952,
2864, 1667 (CdO), 1599, 1504, 1451, 1364, 1293, 1230, 1195, 1157,
1101, 965, 846, 767 cm-1; UV (hexane) 241 (8500), 311 (sh 200) nm;
MS (EI) m/z 277, 276 (M+), 262, 247, 219, 205, 187, 165, 152, 138,
123 (100), 109, 97, 83, 81, 71, 69, 57, 55, 43, 41; HRMS 276.1897
(calcd for C18H25OF, 276.1889). Anal. Calcd for C18H25OF: C, 78.22;
H, 9.12. Found: C, 78.45; H, 9.18.
cis-4-tert-Butyl-1-methylcyclohexyl p-Cyanophenyl Ketone (1e).
In a flame-dried 100-mL flask equipped with a stirrer and condenser,
2.90 g (10.5 mmol) of ketone 1d and 1.4 g (21.6 mmol) of potassium
cyanide were dissolved in 70 mL of anhydrous dimethyl sulfoxide.
The mixture was refluxed until no starting material remained as
indicated by TLC (2% ether/hexanes). The dark orange solution was
poured into 100 mL of water and extracted with ether (3 × 50 mL).
Standard workup afforded a yellow solid which was recrystallized from
5% ether/petroleum ether to give ketone 1e (2.9 g, 98%) as a faintly
yellow powder: mp 115-117 °C; 1H NMR (CDCl3) δ 7.78-7.65 (4H,
m, ArH), 2.52-2.30 (2H, m), 1.63-1.50 (2H, m), 1.34 (3H, s, CH3),
1.25-1.10 (2H, m), 1.05-0.80 (3H, m), 0.76 (9H, s, C(CH3)3); 13C
(CDCl3) δ 25.54 (+), 27.38 (-), 28.29 (-), 32.31 (+), 36.91 (+), 47.44
(-), 48.85 (+), 114.16 (+), 118.07 (+), 127.73 (-), 131.97 (-), 143.58
(+), 208.74 (+); IR (KBr) 2963, 2942, 2857, 2232 (CN), 1674 (Cd
O), 1602, 1453, 1393, 1364, 1291, 1274, 1242, 1219, 1196, 1167, 1142,
983, 958, 931, 861, 844, 771, 563 cm-1; UV (methanol) 241 (17 100),
280 (1500), 328 (sh 240) nm; MS (EI) m/z 284, 283 (M+), 268, 255,
226, 172, 153, 131, 130, 102, 97, 95, 83, 69, 57 (100), 55, 41; HRMS
(EI) m/z 283.1942 (calcd for C19H25NO, 283.1936). Anal. Calcd for
C19H25NO: C, 80.52; H, 8.89; N, 4.94. Found: C, 80.67; H, 8.85; N,
4.84.
cis-4-tert-Butyl-1-methylcyclohexyl p-Carboxyphenyl Ketone (1f).
To a 100-mL flask equipped with a condenser was added 2.0 g (7.1
mmol) of ketone 1e, 50 g of potassium hydroxide, 75 mL of water,
and 15 mL of ethanol. The mixture was refluxed for 72 h, cooled to
room temperature, and washed with ether. The aqueous layer was
acidified with concentrated HCl and the resulting white precipitate
collected by vacuum fitration and recrystallized from ethanol/water to
afford 2.1 g (98%) of keto acid 1f as long white needles: mp 232-
233 °C; 1H NMR (CDCl3) δ 8.26-8.00 (2H, d, J ) 8 Hz, ArH), 7.81-
7.53 (2H, d, J ) 8 Hz, ArH), 2.56-2.31 (2H, m), 1.75-1.45 (2H, m),
1.38 (3H, s, CH3), 1.42-1.12 (2H, m), 1.12-0.83 (3H, m), 0.78 (9H,
s, C(CH3)3), no COOH resonance observed; 13C NMR (CDCl3) δ 24.55
(+), 27.40 (-), 28.37 (-), 32.32 (+), 36.97 (+), 47.49 (-), 48.83
(+), 127.17 (-), 130.01 (-), 130.68 (+), 144.74 (+), 170.10 (+),
209.88 (+); IR (KBr) 3433 (OH), 2952, 2905, 2869, 2667, 2544, 2349,
2303, 1956, 1690 (CdO), 1672 (CdO), 1568, 1504, 1452, 1423, 1405,
1364, 1317, 1287, 1220, 1194, 1126, 961, 948, 931, 867, 735, 556,
534 cm-1; UV (methanol) 244 (12 700), 284 (sh 1800), 322 (200) nm;
MS (EI) m/z 303, 302 (M+), 287, 278, 269, 257, 245, 227, 206, 191,
165, 153, 152, 150, 149, 136, 121, 109, 97, 95, 94, 83, 81, 71, 69, 65,
57 (100), 55, 41; HRMS (EI) m/z 302.1887 (calcd for C19H26O3,
302.1882). Anal.Calcd for C19H26O3: C, 75.46; H, 8.67. Found: C,
75.20; H, 8.84.
cis-4-tert-Butyl-1-methylcyclohexyl p-Carbomethoxyphenyl Ke-
tone (1g). A solution of 0.7 g (2.3 mmol) of keto acid 1f and 0.5 g of
p-toluenesulfonic acid in 35 mL of methanol was refluxed for 24 h.
The reaction mixture was cooled, poured into 30 mL of ice/water, and
extracted with five 30-mL portions of ether. Standard workup afforded
a light yellow solid, and recrystallization from petroleum ether gave
0.7 g (96%) of keto ester 1g as a white powder: mp 140-141 °C; 1H
NMR (CDCl3) δ 8.12-7.99 (2H, d, J ) 8 Hz, ArH), 7.71-7.55 (2H,
d, J ) 8 Hz, ArH), 3.94 (3H, s, COOCH3), 2.54-2.33 (2H, m), 1.66-
1.48 (2H, m), 1.37 (3H, s, CH3), 1.36-1.12 (2H, m), 1.09-0.81 (3H,
m), 0.67 (9H, s, C(CH3)3); 13C NMR (CDCl3) δ 24.53 (+), 27.40 (-),
28.38 (-), 32.31 (+), 37.02 (+), 47.49 (-), 52.33 (-), 127.07 (-),
129.32 (-), 131.68 (+), 143.85 (+), 166.37 (+), 209.85 (+); IR (KBr)
2963, 2948, 2865, 1722 (CdO), 1670 (CdO), 1471, 1463, 1449, 1404,
1365, 1311, 1282, 1223, 1195, 1110, 1019, 961, 865, 754, 737, 716
cm-1; UV (methanol) 244 (17 000), 281 (1200), 327 (180) nm; MS
Preparation of Starting Materials. Ketone 1a was prepared as
described by Padwa and Eastman,13b and ketone 1c was prepared
according to the procedure of Lewis et al.12a As far as we are aware,
the remainder of the starting materials described below are new.
cis-4-tert-Butylcyclohexyl p-fluorophenyl ketone (1b) was prepared
as described for ketone 1a,13b except that fluorobenzene was used instead
of benzene in the Friedel-Crafts reaction. Recrystallization of the crude
product from methanol afforded colorless plates: mp 82-84 °C, in a
1
yield of ∼57%; H NMR (CDCl3) δ 7.99-7.82 (2H, m, ArH), 7.18-
7.02 (3H, m, ArH), 3.52-3.41 (1H, m, R-CH) 2.22-2.09 (2H, m),
1.72-1.50 (4H, m), 1.36-1.11 (3H, m), 0.80 (9H, s, C(CH3)3); 13C
NMR (CDCl3) δ 23.34 (+), 27.46 (-), 28.27 (+), 32.52 (+), 40.32
(-), 48.06 (-), 115.28 and 115.70 (2JC-F ) 21 Hz, -ve), 130.66 and
130.84 (3JC-F ) 9 Hz, -ve), 133.11 and 133.17 (4JC-F ) +ve), 162.71
and 167.74 (1JC-F ) 253 Hz, +ve), 202.60 (+); IR (KBr) 2923, 2856,
1675 (CdO), 1599, 1508, 1467, 1453, 1407, 1393, 1365, 1345, 1302,
1217, 1171, 1158, 1142, 974, 908, 856, 828, 791, 661, 607, 559, 522,
484 cm-1; UV (hexane) 242 (15 500), 269 (3320), 325 (110) nm; MS
(EI) m/z 263, 262 (M+), 247, 205, 165, 151, 138, 124, 123 (100), 112,
109, 95, 83, 75, 69, 57, 55, 41, 32; HRMS (EI) m/z 262.1734 (calcd
for C17H23OF, 262.1733). Anal. Calcd for C17H23OF: C, 77.82; H,
8.84. Found: C, 78.05; H, 9.10.
cis-4-tert-Butyl-1-methylcyclohexyl p-Fluorophenyl Ketone (1d).
A solution of 7.4 g (34.5 mmol) of an approximately 85:15 cis/trans
mixture of methyl 4-tert-butyl-1-methylcyclohexanecarboxylate32 was
dissolved in 100 mL of anhydrous THF and refluxed for 1 h with 4.0
g (105 mmol) of lithium aluminum hydride. After cooling to room
temperature, the solution was quenched with 1 N sulfuric acid and
extracted with ether, and the ether layers were washed and dried.
Removal of the ether in vacuo afforded a quantitative yield (6.3 g) of
a mixture of cis- and trans-4-tert-butyl-1-methylcyclohexylmethanol.
Separation of the isomers at this stage was readily accomplished via
silica gel column chromatography using 7% ethyl acetate in hexanes
as the eluent. The cis-alcohol so obtained (5.3 g) was converted directly
into the corresponding aldehyde (5.1 g, 97%, light yellow oil) by
standard pyridinium chlorochromate oxidation,33 and to a portion of
this material (3.0 g, 16.5 mmol) in 75 mL of dry THF was added 50
mL of a 1.6 M hexane solution of p-fluorophenylmagnesium bromide
in hexane (Aldrich) over 5 min at 0 °C. The solution was stirred at
room temperature for 5 h and then treated with saturated ammonium
chloride solution. Standard workup (extraction with ether, washing,
drying, and removal of solvent in vacuo) afforded 4.1 g (89%) of the
secondary alcohol as a yellow liquid. Without further purification, this
material was subjected to Jones oxidation (10.0 g of chromium trioxide,
5.0 mL of water, and 2.0 mL of concentrated sulfuric acid in 50 mL of
acetone), and standard workup afforded ketone 1d as a light yellow
solid (3.9 g, 96%). Recrystallization from methanol gave colorless
flakes: mp 76-78 °C; 1H NMR (CDCl3) δ 7.60-7.50 (2H, m, ArH),
6.80-6.60 (2H, m, ArH), 2.48-2.42 (2H, m), 1.48-1.42 (2H, m), 1.16
(3H, s, CH3), 1.11-0.90 (5H, m), 0.78 (9H, s, C(CH3)3); 13C NMR
(32) Krapcho, A. P.; Dundulis, E. A. J. Org. Chem. 1980, 45, 3236.
(33) Corey, E. J.; Suggs, J. W. Tetrahedron Lett. 1975, 2647.

Yang Photocyclization Reaction in the Solid State
J. Am. Chem. Soc., Vol. 120, No. 49, 1998 12765
(EI) m/z 317, 316 (M+), 301, 285, 257, 245, 218, 192, 179, 165, 164
(100), 163, 153, 152, 136, 135, 120, 104, 97, 95, 94, 83, 81, 69, 57,
55, 41; HRMS (EI) m/z 316.2036 (calcd for C20H28O3, 316.2039). Anal.
Calcd for C20H28O3: C, 75.91; H, 8.92. Found: C, 75.84; H, 8.91.
was stirred at room temperature for 5 h and then carefully decomposed
using saturated aqueous ammonium chloride. Extraction with ether
and standard workup afforded the secondary alcohol, which was
subjected directly to Jones oxidation (10 g of chromium trioxide, 5
mL of water, and 2.0 mL of concentrated sulfuric acid in 50 mL of
acetone). Standard workup afforded a light yellow solid, which was
recrystallized from methanol to yield 3.9 g (86%) of colorless plates:
cis-4-tert-Butyl-1-methylcyclohexyl p-tolyl ketone (1h) was pre-
pared as described for 1d, except that p-tolylmagnesium bromide was
used in the Grignard step instead of (p-fluorophenyl)magnesium
bromide. Recrystallization of the crude solid from petroleum ether
afforded colorless plates of ketone 1h: mp 96-98 °C; 1H NMR (CDCl3)
δ 7.64-7.58 (2H, d, J ) 8 Hz, ArH), 7.20-7.10 (2H, d, J ) 8 Hz,
ArH), 2.60-2.42 (2H, m), 2.38 (3H, s, ArCH3), 1.60-1.48 (2H, m),
1.38 (3H, s, CH3), 1.30-1.10 (2H, m), 1.08-0.82 (3H, m), 0.78 (9H,
s, C(CH3)3); 13C NMR (CDCl3) δ 21.42 (-), 24.53 (+), 27.42 (-),
28.54 (-), 32.32 (+), 37.54 (+), 47.65 (-), 48.43 (+), 127.76 (-),
128.65 (-), 136.67 (+), 141.14 (+), 209.00 (+); IR (KBr) 2965, 2938,
2860, 2837, 1666 (CdO), 1608, 1450, 1433, 1364, 1293, 1273, 1228,
1184, 965, 829, 760, 747 cm-1; UV (methanol) 245 (10 800), 270 (sh
860), 324 (120) nm; MS (EI) m/z 273, 272 (M+), 257, 239, 215, 161,
152, 135, 121, 120, 119 (100), 97, 95, 91, 83, 67, 65, 57, 55, 41; HRMS
(EI) m/z 272.2138 (calcd for C19H28O, 272.2140). Anal. Calcd for
C19H28O: C, 83.77; H, 10.36. Found: C, 83.53; H, 10.22.
1
mp 93-94 °C; H NMR (CDCl3) δ 7.92-7.80 (2H, m, ArH), 7.18-
7.02 (2H, m, ArH), 3.40 (1H, m, R-CH), 2.25 (2H, m), 2.16-1.46 (12H,
m); 13C NMR (CDCl3) δ 27.47 (-), 27.92 (-), 30.40 (-), 32.70 (+),
37.38 (+), 38.80 (+), 52.05 (-), 115.24 and 115.67 (2JC-F ) 22 Hz,
-ve), 130.52 and 130.70 (3JC-F ) 9 Hz, -ve), 133.36 and 133.42 (4JC-F
) 3 Hz, +ve), 162.61 and 167.64 (1JC-F ) 253 Hz, +ve), 202.32 (+);
IR (KBr) 2902, 2851, 1678 (CdO), 1593, 1504, 1451, 1407, 1344,
1336, 1297, 1268, 1237, 1206, 1157, 1101, 1010, 948, 861, 846, 829,
813, 795, 619, 505 cm-1; UV (methanol) 241 (10 700), 274 (sh 300),
342 (60) nm; MS (EI) m/z 259, 258 (M+), 257, 240, 215, 197, 177,
163, 151, 138, 136, 135, 124, 123 (100), 107, 95, 93, 91, 81, 79, 67,
55, 41; HRMS (EI) m/z 258.1413 (calcd for C17H19OF, 258.1420). Anal.
Calcd for C17H19OF: C, 79.04; H, 7.41. Found: 79.09; H, 7.57.
2-Benzoyl-2-methyladamantane (5c). Into a 100-mL flame-dried
flask fitted with a condenser was placed a suspension of 2.0 g (8.2
mmol) of 5a and 1.0 g (41.7 mmol) of powdered sodium hydride in
50 mL of anhydrous 1,2-dimethoxyethane. The mixture was refluxed
for 4 h and cooled in an ice/water bath, and 2.3 g (16.2 mmol) of
methyl iodide was added by syringe. After stirring for 30 min, 0.5 g
more of sodium hydride was added and the mixture was then refluxed
for 90 min. After cooling, the reaction mixture was treated with 30
mL of water and extracted with pentane, and the organic layer was
washed, dried, and concentrated in vacuo to afford a solid residue, which
was chromatographed over silica gel using 5% ether/hexanes as the
eluent. The second compound to elute from the column was collected
and recrystallized from hexanes to afford 1.1 g (52%) of colorless
plates: mp 56-57 °C; 1H NMR (CDCl3) δ 7.78-7.64 (2H, m, ArH),
7.50-7.30 (3H, m, ArH), 2.38 (2H), 2.18-2.07 (2H, m), 1.90-1.62
(10H, m), 1.60 (3H, s, CH3); 13C NMR (CDCl3) δ 23.80 (-), 27.04
(-), 27.24 (-), 32.54 (+), 34.11 (-), 35.04 (+), 38.10 (+), 53.10
(+), 127.85 (-), 127.92 (-), 130.79 (-), 139.20 (+), 210.80 (+); IR
(KBr) 2950, 2913, 2862, 1666 (CdO), 1596, 1462, 1446, 1256, 1214,
1163, 1138, 1104, 1081, 1002, 968, 953, 942, 928, 888, 795, 726, 697
cm-1; UV (methanol) 239 (9500), 270 (sh 795), 322 (143) nm; MS
(EI) m/z 254 (M+), 150, 149 (100), 148, 121, 107, 105, 93, 91, 81, 79,
77, 67, 55, 41; HRMS (EI) m/z 254.1668 (calcd for C18H22O, 254.1671).
Anal. Calcd for C18H22O: C, 84.99; H, 8.72. Found: C, 84.78; H,
8.75.
cis-4-tert-Butyl-1-methylcyclohexyl p-methoxyphenyl ketone (1i)
was prepared as described for 1d, except that (p-methoxyphenyl)-
magnesium bromide was used in the Grignard step instead of (p-
fluorophenyl)magnesium bromide. Recrystallization of the crude solid
from 95:5 (v/v) petroleum ether/ethanol afforded ketone 1i as white
1
needles: mp 106-107 °C; H NMR (CDCl3) δ 7.80-7.70 (2H, m,
ArH), 6.93-6.82 (2H, m, ArH), 3.83 (3H, s, ArOCH3), 2.60-2.42 (2H,
m), 1.60-1.50 (2H, m), 1.38 (3H, s, CH3), 1.30-1.10 (2H, m), 1.05-
0.80 (3H, m), 0.75 (9H, s, C(CH3)3); 13C NMR (CDCl3) δ 24.52 (+),
27.42 (-), 28.59 (-), 32.31 (+), 37.85 (+), 47.71 (-), 48.24 (+),
55.32 (-), 113.18 (-), 130.19 (-), 131.50 (+), 161.73 (+), 208.84
(+); IR (KBr) 2962, 2937, 2867, 2841, 1661(CdO), 1604, 1575, 1507,
1455, 1444, 1413, 1366, 1309, 1253, 1230, 1176, 1143, 1033, 964,
836, 767, 648 cm-1; UV (methanol) 268 (18 700), 319 (sh 700) nm;
MS (EI) m/z 289, 288 (M+), 273, 257, 231, 206, 163, 150, 137, 135
(100), 107, 97, 92, 77, 64, 57, 41; HRMS (EI) m/z 288.2089 (calcd for
C19H28O2, 288.2089). Anal. Calcd for C19H28O2: C, 79.12; H, 9.78.
Found: C, 79.24; H, 9.78.
2-Benzoyladamantane (5a). A mixture of 5.0 g (27 mmol) of
2-adamantane carboxylic acid34 and 10.0 g (84 mmol) of thionyl
chloride was placed in a flame-dried 50-mL flask equipped with a
condenser and stirred for 30 min under nitrogen and then for 2 h at
reflux. Excess thionyl chloride was removed by distillation, and the
remaining acid chloride was used in the next step without purification.
A solution of 3.0 g (14.8 mmol) of acid chloride in 25 mL of anhydrous
benzene was added over 30 min to a suspension of 7.0 g (52.5 mmol)
of aluminum chloride in 150 mL of anhydrous benzene. After stirring
for 5 h at room temperature, 100 mL of water was cautiously added
and the resulting mixture was extracted with 2 × 75 mL of ether.
Standard workup afforded an oil that crystallized upon standing
overnight under vacuum. Recrystallization from methanol yielded 1.2
g (33%) of colorless plates: mp 91-93 °C; 1H NMR (CDCl3) δ 7.90-
7.80 (2H, m, ArH), 7.60-7.40 (3H, m, ArH), 3.50-3.40 (1H, m,
R-CH), 2.30 (2H, m), 2.10-1.40 (12 H, m); 13C NMR (CDCl3) δ 27.52
(-), 27.96 (-), 30.31 (-), 32.77 (+), 37.44 (+), 38.82 (+), 52.14
(-), 128.04 (-), 128.41 (-), 132.12 (-), 137.23 (+), 204.08 (+); IR
(KBr) 2950, 1676 (CdO), 1447, 1364, 1221, 973, 766, 698 cm-1; UV
(methanol) 238 (17,800), 278 (sh 1500), 339 (80) nm; MS (EI) m/z
241, 240 (M+), 239, 222, 197, 179, 145, 135, 120, 105, 93, 91, 79, 77,
67, 51, 41; HRMS (EI) m/z 240.1509 (calcd for C17H20O, 240.1514).
Anal. Calcd for C17H20O: C, 84.96; H, 8.39. Found: C, 85.12; H,
8.43.
2-(p-Fluorobenzoyl)-2-methyladamantane
(5d).
Methyl-
ation of 2-(p-fluorobenzoyl)adamantane was carried out as described
above to afford a 52% yield of ketone 5d as colorless plates (hexane):
1
mp 86-87 °C; H NMR (CDCl3) δ 7.84-7.70 (2H, m, ArH), 7.14-
6.99 (2H, m, ArH), 2.34 (2H, m), 2.19-2.05 (2H, m), 1.90-1.72 (3H,
m), 1.72-1.55 (7H, m), 1.50 (3H, s, CH3); 13C NMR (CDCl3) δ 23.67
(-), 26.97 (-), 27.19 (-), 32.50 (+), 34.21 (-), 34.96 (+), 38.01
(+), 53.00 (+), 114.74 and 115.17 (2JC-F ) 21 Hz, -ve), 130.30 and
130.47 (3JC-F ) 8 Hz, -ve), 135.01 and 135.08 (4JC-F ) 3 Hz, +ve),
161.64 and 166.65 (1JC-F ) 232 Hz, +ve), 208.47 (+); IR (KBr) 2921,
2858, 1668 (CdO), 1598, 1504, 1454, 1354, 1259, 1226, 1160, 1142,
1103, 1084, 970, 954, 940, 851, 821, 769, 759, 608, 543, 492 cm-1
;
UV (methanol) 241 (8,900), 333 (70) nm; EI (MS) m/z 273, 272 (M+),
189, 181, 169, 150, 149 (100), 148, 133, 123, 121, 107, 95, 93, 81, 79,
67, 55, 41; HRMS (EI) m/z 272.1571 (calcd for C18H21OF, 272.1577).
Anal. Calcd for C18H21OF: C, 79.38; H, 7.77. Found: C, 79.31; H,
7.86.
2-(p-Cyanobenzoyl)-2-methyladamantane (5e). This compound
was synthesized from 5d in 98% yield by a procedure identical to that
used to convert ketone 1d to 1e. Recrystallization of the crude product
from 95:5 (v/v) petroleum ether/ether gave pale yellow plates: mp 134-
2-(p-Fluorobenzoyl)adamantane (5b). A solution of 3.0 g (16.5
mmol) of 2-adamantanecarboxaldehyde34 in 75 mL of anhydrous THF
was cooled in an ice bath, and to this was added (p-fluorophenyl)-
magnesium bromide (Aldrich, 1.6 M in hexane solution, 50 mL, 80
mmol) via syringe over a period of 5 min under nitrogen. The solution
1
135 °C; H NMR (CDCl3) δ 7.81-7.65 (4H, m, ArH), 2.28 (2H, m),
2.18-2.05 (2H, m), 1.90-1.72 (2H, m), 1.72-1.55 (8H, m), 1.50 (3H,
s, CH3); 13C NMR (CDCl3) δ 23.75 (-), 26.87 (-), 27.11 (-), 32.34
(+), 33.92 (-), 35.03 (+), 37.89 (+), 53.42 (+), 114.31 (+), 118.10
(+), 128.33 (-), 131.85 (-), 142.86 (+), 209.09 (+); IR (KBr) 2904,
(34) Alberts, A. H.; Wynberg, H.; Strating, J. Synth. Commun. 1972, 2,
79.

12766 J. Am. Chem. Soc., Vol. 120, No. 49, 1998
LeiboVitch et al.
2855, 2230, 1669 (CdO), 1460, 1450, 1401, 1354, 1286, 1251, 1221,
1140, 1105, 1083, 973, 954, 941, 845, 768, 707, 594, 545, 528, 512
cm-1; UV (methanol) 243 (15 600), 323 (200) nm; MS (EI) m/z 280,
279 (M+), 181, 169, 150, 149 (100), 131, 130, 121, 107, 102, 95, 93,
91, 81, 79, 77, 69, 67, 55, 41; HRMS (EI) m/z 279.1621 (calcd for
C19H21NO, 279.1623). Anal. Calcd for C19H21NO: C, 81.68; H, 7.58;
N, 5.01. Found: C, 81.66; H, 7.67; N, 4.97.
(S)-(+)-Arginine Salt of Keto Acid 1f (S3). Keto acid 1f (197
mg, 0.65 mmol) and 113 mg (0.65 mmol) of (S)-(+)-arginine in 2 mL
of water were dissolved in 25 mL of ethanol. After standing for several
hours at room temperature, the solution was concentrated to dryness
in vacuo and the resulting white solid washed with ether, leaving 301
mg (97%) of the desired salt: MS (FAB, -LSIMS) m/z 475 (M- - 1),
445, 409, 407, 383, 317, 302, 301 (100), 285, 257, 229, 209, 173, 121,
120 (matrix, thioglycerol). Anal. Calcd for C25H40O5N4: C, 63.00;
H, 8.46; N, 11.76. Found: C, 62.92; H, 8.52; N, 11.63.
(S)-(-)- and (R)-(+)-r-Methylbenzylamine Salts of Keto Acid
1f (S4 and S5). A solution of 13.2 mg (0.11 mmol) of (S)-(-)-R-
methylbenzylamine in 10 mL of ether was added to a solution of 32.6
mg (0.11 mmol) of keto acid 1f in 20 mL of ether. The resulting
precipitate was filtered and washed with ether to give 33 mg (71%) of
salt: MS (FAB, +LSIMS) m/z 424 (M+ + 1), 395, 377, 333, 304,
303, 285, 277, 241, 229, 214, 165, 149, 122 (100), 105 (matrix: glycerol
+ methanol). Anal. Calcd for C27H37O3N: C, 76.55; H, 8.81; N, 3.31.
Found: C, 76.69; H, 8.84; N, 3.44.
2-(p-Carboxybenzoyl)-2-methyladamantane (5f). This compound
was synthesized from 5e in 98% yield by a procedure identical to that
used to convert ketone 1e to 1f. Recrystallization of the crude product
1
from aqueous ethanol gave long white needles: mp 216-218 °C; H
NMR (CDCl3) δ 8.20-8.10 (2H, d, J ) 8 Hz, ArH), 7.85-7.75 (2H,
d, J ) 8 Hz, ArH), 2.32 (2H, m), 2.20-2.08 (2H, m), 1.92-1.74 (2H,
m), 1.73-1.59 (8H, m), 1.53 (3H, s, CH3), no COOH signal was
detectable; 13C NMR (CDCl3) δ 23.79 (-), 26.96 (-), 27.17 (-), 32.43
(+), 33.94 (-), 35.07 (+), 37.99 (+), 53.40 (+), 127.81 (-), 129.89
(-), 130.92 (+), 143.97 (+), 170.99 (+), 210.17 (+); IR (KBr) 3420
(OH), 2910, 2882, 2855, 1694 (CdO), 1681(CdO), 1506, 1456, 1435,
1404, 1322, 1298, 1255, 1214, 1127, 1104, 974, 955, 940, 866, 789,
766, 735, 703, 547,531, 459 cm-1; UV (methanol) 247 (16 000), 275
(sh, 1300), 322 (200) nm; MS (EI) m/z 299, 298 (M+), 281, 253, 178,
159, 150, 149 (100), 148, 131, 121, 107, 93, 81, 79, 65, 55, 41; HRMS
(EI) m/z 298.1562 (calcd for C19H22O3, 298.1569). Anal. Calcd for
C19H22O3: C, 76.47; H, 7.44. Found: C, 76.35; H, 7.39.
The (R)-(+)-R-methylbenzylamine salt was prepared analogously
(78% yield) and also gave satisfactory MS and elemental analysis
results.
(1R,2S)-(-)-Ephedrine Salt of Keto Acid 1f (S6). To a solution
of 14.4 mg (0.09 mmol) of (1R,2S)-(-)-ephedrine in 10 mL of ether
was added a solution of 26.4 mg (0.09 mmol) of keto acid 1f in 20 mL
of ether. Precipitate started to form within 1 min, and after 1 h, filtration
and washing with ether afforded 33 mg (80%) of salt: MS (FAB,
+LSIMS) m/z 468 (M+ + 1), 452, 433, 395, 361, 331, 304, 303, 285,
258, 229, 167, 166 (100), 149, 148, 133 (matrix: glycerol + methanol).
Anal. Calcd for C29H41O4N: C, 74.48; H, 8.84; N, 3.00. Found: C,
74.12; H, 8.79; N, 2.97.
(S)-(-)-Proline-tert-butyl Ester Salt of Keto Acid 1f (S7). This
salt was prepared in 87% yield by a procedure identical to that described
for salts S4-S6: MS (FAB, +LSIMS) m/z 474 (M+ + 1), 441, 409,
379, 371, 343, 304, 303, 301, 285, 229, 173, 172, 149, 133, 116 (100),
83, 70 (matrix, thioglycerol). Anal. Calcd for C28H43O5N: C, 71.00;
H, 9.15; N, 2.96. Found: C, 70.98; H, 8.91; N, 2.80.
(1R,2S)-(-)-Norephedrine Salt of Keto Acid 1f (S8). This salt
was prepared in 97% yield by a procedure identical to that described
for salts S4-S6: MS (FAB, +LSIMS) m/z 454 (M+ + 1), 433, 395,
336, 304, 303, 285, 258, 244, 229, 166, 152 (100), 134, 118 (matrix:
glycerol + methanol). Anal. Calcd for C28H39O4N: C, 74.14; H, 8.67;
N, 3.09. Found: C, 74.24; H, 8.54; N, 3.08.
(S)-(-)-Prolinamide Salt of Keto Acid 1f (S9). To a solution of
7.6 mg (0.07 mmol) of (S)-(-)-prolinamide in 2 mL of ethanol was
added a solution of 20.2 mg (0.07 mmol) of keto acid 1f in 5 mL of
acetonitrile. The solution was allowed to stand until almost all of the
solvent had evaporated. The remaining solvent was decanted and the
solid material was washed with diethyl ether to give 27 mg (96%) of
salt: MS (FAB, +LSIMS) m/z 417 (M+ + 1), 395, 317, 304, 303,
285, 258, 244, 229, 215, 217, 166, 152, 134, 115 (100) (matrix: glycerol
+ methanol). Anal. Calcd for C24H36O4N2: C, 69.20; H, 8.71; N,
6.72. Found: C, 68.91; H, 8.80; N, 6.66.
2-(p-Carbomethoxybenzoyl)-2-methyladamantane (5g). This com-
pound was synthesized from 5f in 96% yield by a procedure identical
to that used to convert ketone 1f to 1g. Recrystallization of the crude
1
product from petroleum ether gave white plates: mp 90-92 °C; H
NMR (CDCl3) δ 8.10-8.00 (2H, d, J ) 8 Hz, ArH), 7.80-7.70 (2H,
d, J ) 8 Hz, ArH), 3.94 (3H, s, CO2CH3), 2.32 (2H, m), 2.21-2.07
(2H, m), 1.93-1.73 (2H, m), 1.73-1.50 (8H, m), 1.52 (3H, s, CH3);
13C NMR (CDCl3) δ 23.78 (-), 26.95 (-), 27.16 (-), 32.44 (+), 33.95
(-), 35.04 (+), 37.99 (+), 52.33 (-), 53.33 (+), 127.71 (-), 129.23
(-), 131.85 (+), 143.10 (+), 166.37 (+), 210.10 (+); IR (KBr) 2911,
2859, 1727 (CdO), 1673 (CdO), 1457, 1438, 1402, 1285, 1255, 1218,
1107, 1017, 972, 954, 860, 791, 743, 709 cm-1; UV (methanol) 247
(15 800), 275 (sh 1360), 323 (200) nm; MS (EI) m/e 312 (M+), 297,
281, 253, 164, 163, 150, 149 (100), 135, 121, 107, 93, 79; HRMS (EI)
m/z 312.1720 (calcd for C20H24O3, 312.1725). Anal. Calcd for
C20H24O3: C, 76.88; H, 7.75. Found: C, 76.91; H, 7.84.
Preparation of Salts. Salts were prepared by dissolving equimolar
amounts of keto acid (1f or 5f) and amine in the solvents indicated.
The initial precipitates were collected and recrystallized from the
solvents given in Table 3; melting points and crystal morphologies are
also given in Table 3. Satisfactory NMR, IR, and UV spectra were
recorded for all salts. In the IR, salt formation resulted in a
characteristic change in the carboxylic acid OH stretching band at
∼3500-2400 cm-1, which was replaced by multiple combination bands
for ammonium in the 3200-2200-cm-1 region. In addition, the
carboxylic acid carbonyl stretch at ∼1690 cm-1 was replaced by two
bands for the carboxylate anion: a strong absorption in the 1650-
1550-cm-1 region and a weaker one near 1400 cm-1
.
(S)-(+)-Prolinol Salt of Keto Acid 1f (S1). Keto acid 1f (246 mg,
0.8 mmol) and 82 mg (0.8 mmol) of (S)-(+)-prolinol were mixed in
30 mL of acetone. The solution was heated until the solid dissolved,
and after standing for 1 h at room temperature, the precipitated salt
was collected by suction filtration (252 mg, 77%). MS (FAB,
+LSIMS) m/z 404 (M+ + 1), 391, 333, 305, 304, 303, 301, 286, 285,
229, 149 (100), 133, 102 (matrix: thioglycerol + methanol). Anal.
Calcd for C24H37O4N: C, 71.42; H, 9.25; N, 3.47. Found: C, 71.27;
H, 9.26; N, 3.43.
(1S,2S)-(+)-Pseudoephedrine Salt of Keto Acid 1f (S2). Keto acid
1f (202 mg, 0.67 mmol) and 111 mg (0.67 mmol) of (1S,2S)-(+)-
pseudoephedrine were dissolved in 30 mL of ethanol. The solution
was concentrated in vacuo to a volume of ∼5 mL and ether added
until a white precipitate was observed. The precipitated salt was
collected by suction filtration (293 mg, 94%). MS (FAB, +LSIMS)
m/z 468 (M+ + 1), 393, 380, 347, 304, 303, 285, 229, 167, 166 (100),
165, 149, 133 (matrix: thioglycerol + methanol). Anal. Calcd for
C29H41O4N‚2H2O: C, 69.15; H, 9.01; N, 2.78. Found: C, 69.16; H,
8.86; N, 2.76.
(1S)-(-)-2,10-Camphorsultam Salt of Keto Acid 1f (S10). This
salt was prepared in 84% yield by a procedure identical to that described
for salts S4-S6: MS not obtained. Anal. Calcd for C29H43O5N2S:
C, 67.28; H, 8.37; N, 2.71. Found: C, 67.18; H, 8.47; N, 2.69.
(S)-(+)-Lysine Salt of Keto Acid 1f (S11). This salt was prepared
by combining a solution of 10.8 mg (0.07 mmol) of (S)-(+)-lysine in
0.5 mL of water and a solution of 22.3 mg (0.07 mmol) of keto acid
1f in 6 mL of ethanol. After standing for several hours at room
temperature, the solution was concentrated to dryness in vacuo and
the resulting light yellow solid washed with ether, leaving 33 mg (100%)
of the desired salt: MS (FAB, +LSIMS) m/z 449 (M+ + 1), 433, 417,
383, 347, 341, 331, 315, 303, 293, 285, 261, 229, 215, 201, 169, 147
(100), 130, 91 (matrix, thioglycerol). Anal. Calcd for C25H40O5N2:
C, 66.94; H, 8.99; N, 6.24. Found: C, 66.74; H, 9.08; N, 6.16.
(S)-(+)-Prolinol Salt of Keto Acid 5f (S12). This salt was prepared
in 77% yield by a procedure identical to that described for salt S1:
MS (FAB, +LSIMS) m/z 400 (M+ + 1), 371, 358, 343, 316, 300,
299, 297, 281, 237, 203, 181, 149, 133, 131, 124, 103, 102 (100)

Yang Photocyclization Reaction in the Solid State
J. Am. Chem. Soc., Vol. 120, No. 49, 1998 12767
(matrix, thioglycerol). Anal. Calcd for C24H33O4N: C, 72.15; H, 8.33;
N, 3.51. Found: C, 72.32; H, 8.36; N, 3.55.
7.15-7.06 (2H, m, ArH), 5.60-5.49 (1H, m, vinyl CH), 5.08-4.88
(2H, m, vinyl CH2), 2.97-2.80 (2H, m), 1.70-1.60 (1H, m), 1.83-
1.70 (1H, m), 1.60-1.45 (2H, m), 1.25-1.10 (1H, m), 0.88 (9H, s,
C(CH3)3); 13C NMR (CDCl3) δ 23.20 (+), 27.82 (-), 28.60 (+), 32.60
(+), 38.36 (+), 55.40 (-), 115.25 and 115.69 (2JC-F ) 22 Hz, -ve),
116.42 (+), 130.66 and 130.84 (3JC-F ) 9 Hz, -ve), 134.07 and 134.00
(4JC-F ) 4 Hz, +ve), 140.25 (-), 168.50 and 163.50 (1JC-F ) 252 Hz,
+ve), 198.88 (+); IR (KBr) 3080, 2970, 2912, 2870, 1682 (CdO),
1597, 1504, 1477, 1459, 1407, 1364, 1345, 1318, 1291, 1269, 1233,
(1S,2S)-(+)-Pseudoephedrine Salt of Keto Acid 5f (S13). This
salt was prepared in 94% yield by a procedure identical to that described
for salt S2: MS (FAB, +LSIMS) m/z 464 (M+ + 1), 448, 413, 388,
348, 331, 314, 299, 297, 281, 258, 225, 201, 167, 166 (100), 148, 133
(matrix, glycerol). Anal. Calcd for C29H37O4N: C, 75.13; H, 8.04;
N, 3.02. Found: C, 74.80; H, 8.02; N, 2.96.
(S)-(-)- and (R)-(+)-r-Methylbenzylamine Salts of Keto Acid
5f (S14 and S15). The (S)-(-)-salt was prepared in 71% yield by a
procedure identical to that described for salts S4-S6: MS (FAB,
+LSIMS) m/z 421, 420 (M+ + 1), 391, 327, 300, 299, 297, 282, 281,
253, 214, 201, 165, 152, 149, 122 (100), 105 (matrix: glycerol +
methanol). Anal. Calcd for C27H33O3N: C, 77.29; H, 7.93; N, 3.34.
Found: C, 77.35; H, 7.89; N, 3.23.
1197, 1155, 1095, 1008, 998, 985, 917, 851, 819, 745, 601, 568 cm-1
;
UV (hexane) 241 (15 600), 268 (1285), 273 (1040) nm; MS (EI) m/z
263, 262 (M+), 247, 229, 207, 206, 205, 191, 177, 165, 152, 151, 139,
138 (100), 123, 109, 95, 75, 67, 58, 57, 56, 41; HRMS (EI) m/z
262.1726 (calcd for C17H23OF, 262.1733). Anal. Calcd for C17H23-
OF: C, 77.82; H, 8.84. Found: C, 77.91; H, 9.00.
Prolonged irradiation of a benzene solution of ketone 1b resulted in
a complex mixture of products and a decrease in the amount of
photoproduct 4b. p-Fluoroacetophenone could be detected in this
mixture (GC comparison with an authentic sample). Photolysis of
ketone 1b In the solid state led to no detectable photoproducts.
Photolysis of Ketone 1c. In agreement with the report of Lewis et
al.,12 photolysis of ketone 1c in benzene was found to afford
cyclobutanol 2c (an oil) in >95% yield. The same compound was the
only detectable product when ketone 1c was irradiated in the crystalline
state (sample melting at higher conversions).
The (R)-(+)-R-methylbenzylamine salt was prepared analogously
(78% yield) and also gave satisfactory MS and elemental analysis
results.
(1R,2S)-(-)-Ephedrine Salt of Keto Acid 5f (S16). This salt was
prepared in 80% yield by a procedure identical to that described for
salts S4-S6: MS (FAB, +LSIMS) m/z 465, 464 (M+ + 1), 429, 391,
350, 331, 314, 300, 299, 297, 282, 281, 258, 225, 167, 166 (100), 149,
148, 133, 105 (matrix: glycerol + methanol). Anal. Calcd for
C29H37O4N: C, 75.13; H, 8.04; N, 3.02. Found: C, 75.17; H, 8.07; N,
3.19.
Photolysis of Ketone 1d. A solution of 208 mg (0.75 mmol) of
ketone 1d in 20 mL of benzene was photolyzed through Pyrex for 3 h
(solution turns light red). GC analysis indicated the complete consump-
tion of starting material and the presence of two products in a 94:6
ratio. After three Chromatotron separations (3% ether/hexanes), 152
mg (73%) of cyclobutanol 2d, 7 mg (3%) of cyclobutanol 3d, and 46
mg (22%) of a mixture of 2d and 3d were obtained.
endo-Arylcyclobutanol 2d: oil; 1H NMR (C6D6) δ 7.16-7.11 (2H,
m, ArH), 6.76-6.71 (2H, m, ArH), 2.85-2.79 (2H, m), 2.02-1.95
(1H, m), 1.85-1.78 (1H, m), 1.67-1.61 (1H, m), 1.39-1.23 (2H, m),
1.22 (1H, s, OH), 1.03 (3H, s, CH3), 1.00-0.97 (1H, d, J ) 8 Hz),
0.63 (9H, s, C(CH3)3); 13C NMR (C6D6) δ 19.73 (-), 20.41 (+), 28.72
(-), 32.70 (+), 37.09 (+), 46.66 (+), 47.13 (+), 47.49 (-), 54.48
(-), 80.39 (+), 114.58 and 114.41 (2JC-F ) 21 Hz, -ve), 129.63 and
129.57 (3JC-F ) 8 Hz, -ve), 142.09 and 142.06 (4JC-F ) 4 Hz, +ve),
163.17 and 161.21 (1JC-F ) 247 Hz, +ve); IR (film) 3424 (OH), 2951,
2866, 1605, 1510, 1472, 1365, 1229, 1157, 1052, 1012, 909, 839, 814
cm-1; MS (EI) m/z 277, 276 (M+), 275, 274, 258, 243, 219, 201, 187,
178, 177, 165, 163, 161, 153, 152, 151, 147, 138, 137, 123 (100), 109,
97, 95, 84, 83, 82, 81, 69, 57, 55, 43, 41; HRMS (EI) m/z 276.1893
(calcd for C18H25OF, 176.1889). Anal. Calcd for C18H25OF: C, 78.22;
H, 9.12. Found: C, 78.48; H, 9.18.
(1R,2S)-(-)-Norephedrine Salt of Keto Acid 5f (S17). This salt
was prepared in 97% yield by a procedure identical to that described
for salts S4-S6: MS (FAB, +LSIMS) m/z 451, 450 (M+ + 1), 429,
391, 334, 304, 303, 300, 299, 297, 281, 244, 225, 153, 152 (100), 149,
134, 118 (matrix, glycerol). Anal. Calcd for C28H35O4N: C, 74.80;
H, 7.85; N, 3.12. Found: C, 74.76; H, 7.89; N, 3.21.
Photochemical Procedures. All photolyses were performed using
a 450-W Hanovia medium-pressure mercury lamp placed in a water-
cooled immersion well. For small-scale solution-phase and solid-state
photolyses, the sample in a Pyrex tube or an NMR tube was degassed
by several freeze-thaw-pump cycles, sealed under nitrogen, and
suspended ∼10 cm from the immersion well. A minimum of three
samples was irradiated, and for each photolysis, at least three GC runs
were performed. GC detector responses were calibrated with appropri-
ate internal standards and compounds to be detected. Average results
are reported. For preparative-scale liquid-phase photolyses, the solution
was purged with nitrogen for at least 30 min prior to photolysis.
Workup consisted of solvent removal in vacuo followed by silica gel
column or radial chromatograpy. For larger scale solid-state photolyses,
the sample was placed between two Pyrex microscope slides, and by
sliding the top and bottom plates back and forth, the sample was
distributed over the surface in a thin, even layer. The sample plates
were then Scotch-taped together at the top and bottom ends, placed in
a polyethylene bag, and thoroughly degassed with nitrogen. The bag
was then sealed under a positive pressure of nitrogen with a heat-sealing
device and suspended ∼10 cm from the immersion well. Low-
temperature solid-state photolyses were performed by placing the bags
as close to the immersion well as possible in a cooled ethanol/water
bath (Neslab Cryocool CC-100 II). Temperatures were maintained
within (2 °C of the designated value.
Photoproduct Identification. X-ray crystal structures were obtained
for photoproducts 2i and 6d. A detailed comparison (including 2D-
COSY, NOE, HMQC, and HMBC correlations) of the 400- and/or 500-
MHz NMR spectra of these compounds with the corresponding spectra
of the other photoproducts allowed unambiguous structural assignments
to be made.
Photolysis of Ketone 1a. The solution-phase photochemistry of
this compound has been reported by Lewis et al.;12 our results agree
with theirs. In the solid state, ketone 1a was found to be photochemi-
cally inert.
exo-Arylcyclobutanol 3d: oil; 1H NMR (C6D6) δ 7.17-7.12 (2H,
m, ArH), 6.80-6.74 (2H, m, ArH), 1.76-1.65 (1H, m), 1.59-1.52
(1H, m), 1.28-1.20 (2H, m), 1.10-0.99 (2H, m), 1.08 (3H, s, CH3),
0.95 (9H, s, C(CH3)3), 0.89-0.80 (2H, m), 0.19-0.15 (1H, t, J ) 7
Hz); 13C NMR (C6D6) δ 18.86 (+), 19.88 (+), 23.52 (-), 27.37 (-),
28.68 (-), 33.85 (+), 35.04 (+), 35.20 (+), 46.20 (-), 68.65 (+),
114.88 and 114.71 (2JC-F ) 21 Hz, -ve), 133.73 and 133.66 (3JC-F
)
9 Hz, -ve), 140.18 and 140.16 (4JC-F ) 3 Hz, +ve), 163.08 and 161.04
(1JC-F ) 244 Hz, +ve); IR (film) 3467 (OH), 2955, 2866, 1636, 1510,
1472, 1457, 1397, 1364, 1229, 1157, 1052, 1012, 987, 949, 909, 840,
816, 733, 602, 546, 474 cm-1; MS (EI) m/z 277, 276 (M+), 261, 259,
258, 243, 233, 221, 220, 219, 205, 202, 201, 191, 187, 178, 177, 166,
165, 163, 161, 154, 153, 152, 151, 148, 147, 139, 138, 137, 133, 124,
123, 122, 109, 97, 96, 95, 94, 83, 81, 71, 69, 57 (100), 55, 43, 41;
HRMS (EI) m/z 276.1892 (calcd for C18H25OF, 176.1889). Anal. Calcd
for C18H25OF: C, 78.22; H, 9.12. Found: C, 78.33; H, 8.99.
Regardless of the extent of conversion, endo-arylcyclobutanol 2d
was shown by GC and 1H NMR to be the sole product resulting from
irradiation of crystals of ketone 1d.
Photolysis of Ketone 1b. Ketone 1b (108 mg, 0.4 mmol) in 10
mL of benzene was irradiated through Pyrex for 2 h, during which
time the solution turned orange. GC indicated a 24% conversion, and
chromatography of the crude reaction mixture (5% ether/hexanes)
afforded 23 mg of 1-(p-Fluorophenyl)-5-tert-butyl-6-hepten-1-one
Photolysis of Ketone 1e. A solution of 79 mg (0.28 mmol) of
ketone 1e in 10 mL of benzene was photolyzed through Pyrex for 4 h.
GC indicated the presence of a single photoproduct, and column
chromatography of the crude photolyzate using 5% ether/hexanes
afforded 74 mg (94%) of endo-arylcyclobutanol 2e as a white powder.
1
(4b): mp 31-34 °C; H NMR (CDCl3) δ 7.98-7.92 (2H, m, ArH)),

12768 J. Am. Chem. Soc., Vol. 120, No. 49, 1998
LeiboVitch et al.
Recrystallization from acetonitrile gave colorless prisms: mp 181-
183 °C; 1H NMR (C6D6) δ 7.01-6.96 (4H, m, ArH), 2.76-2.69 (2H,
m), 1.89-1.82 (1H, m), 1.78-1.69 (1H, m), 1.60-1.54 (1H, m), 1.37-
1.25 (1H, m), 1.16 (1H, br s, OH), 1.10-1.00 (1H, m), 0.93 (3H, s,
CH3), 0.93-0.91 (1H, d, J ) 8 Hz), 0.53 (9H, s, C(CH3)3); 13C NMR
(C6D6) δ 19.59 (-), 20.49 (+), 28.59 (-), 32.62 (+), 36.78 (+), 46.44
(+), 46.88 (+), 47.86 (-), 54.30 (-), 80.33 (+), 111.40 (+), 118.82
(+), 128.36 (-), 131.33 (-), 150.47 (+); IR (KBr) 3482 (OH), 2968,
2950, 2906, 2863, 2227, 1605, 1475, 1456, 1393, 1361, 1057, 1016,
986, 977, 950, 910, 853, 835, 576, 551, 521, 502 cm-1; MS (EI) m/z
284, 283 (M+), 282, 268, 266, 265, 250, 228, 227, 226, 222, 214, 213,
212, 210, 209, 208, 204, 199, 198, 196, 195, 194, 193, 186, 185, 184,
173, 172, 171, 170, 169, 168, 154, 153, 145, 144, 131, 130 (100), 116,
109, 103, 102, 97, 95, 83, 81, 69, 57, 55, 41; HRMS (EI) m/z 283.1936
(calcd for C19H25NO, 283.1936). Anal. Calcd for C19H25NO: C, 80.52;
H, 8.89; N, 4.94. Found: C, 80.46; H, 8.89; N, 4.86.
(+), 46.78 (+), 47.29 (+), 47.62 (-), 54.71 (-), 80.92 (+), 127.90
(-), 128.46 (-), 136.68 (+), 143.45 (+); IR (KBr) 3432 (OH), 3600-
3200, 2948, 2864, 1514, 1470, 1396, 1365, 1327, 1304, 1217, 1052,
1011, 984, 948, 908, 820, 760 cm-1; MS (EI) m/z 273, 272 (M+), 254,
239, 226, 215, 198, 197, 184, 183, 182, 181, 171, 170, 169, 167, 165,
161, 157, 156, 155, 153, 152, 143, 142, 141, 129, 128, 121, 119, 117,
115, 106, 105 (100), 91, 83, 81, 79, 77, 69, 67, 57, 55; HRMS (EI)
m/z 272.2141 (calcd for C19H28O, 272.2140). Anal. Calcd for
C19H28O: C, 83.77; H, 10.36. Found: C, 83.84; H, 10.25.
endo-Arylcyclobutanol 2h was also found to be the sole product
1
resulting from irradiation of crystals of ketone 1h (GC, H NMR).
Photolysis of Ketone 1i. Photolysis of 90 mg (0.31 mmol) of ketone
1i in 10 mL of benzene through Pyrex for 4 h led to the formation of
two products in a 94:6 ratio as determined by GC. Silica gel
chromatography using 5% ether/hexanes as eluent afforded 10 mg
(11%) of recovered starting material, 60 mg (67%) of photoproduct
2i, and 19 mg (21%) of a mixture of photoproducts 2i and 3i; despite
repeated chromatography, a pure sample of compound 2i could not be
obtained. Recrystallization of endo-arylcyclobutanol 2i from 5% ether/
hexanes afforded colorless prisms: mp 73-76 °C; 1H NMR (C6D6) δ
7.32-7.28 (2H, d, J ) 8 Hz, ArH), 6.73-6.69 (2H, d, J ) 8 Hz, ArH),
3.32 (3H, s, OCH3), 2.95-2.89 (2H, m), 2.19-2.12 (1H, m), 1.93-
1.86 (1H, m), 1.74-1.68 (1H, m), 1.56-1.47 (1H, m), 1.47-1.39 (1H,
m), 1.31 (1H, s, OH), 1.15 (3H, s, CH3), 1.05-1.02 (1H, d, J ) 7 Hz),
0.69 (9H, s, C(CH3)3); 13C NMR (C6D6) δ 19.91 (-), 20.55 (+), 28.85
(-), 32.78 (+), 37.39 (+), 46.79 (+), 47.34 (+), 47.53 (-), 54.67
(-), 54.68 (-), 80.76 (+), 113.19 (-), 129.04 (-), 138.53 (+), 159.17
(+); IR (KBr) 3516 (OH), 2942, 2863, 1608, 1511, 1473, 1456, 1362,
1242, 1177, 1033, 1013, 985, 908, 834, 817, 792, 605, 555 cm-1; MS
(EI) m/z 289, 288 (M+), 287, 271, 270, 257, 232, 231, 213, 199, 186,
177, 175, 173, 171, 164, 150, 137, 136, 135 (100), 121, 115, 109, 107,
105, 95, 92, 91, 81, 79, 77, 69, 57, 55, 43, 41; HRMS (EI) m/z 288.2091
(calcd for C19H28O2, 288.2089). Anal. Calcd for C19H28O2: C, 79.12;
H, 9.78. Found: C, 79.33; H, 9.85.
endo-Arylcyclobutanol 2e was also found to be the sole product
1
resulting from irradiation of crystals of ketone 1e (GC, H NMR).
Photolysis of Ketone 1f. A solution of 5 mg of keto acid 1f in 3
mL of acetone was irradiated through Pyrex for 2 h. The solvent was
removed in vacuo and the residue dissolved in ether and treated with
excess diazomethane. GC analysis showed that keto acid 1f was
completely absent and that two photoproducts were present in a 94:6
ratio. The GC retention times of these peaks were the same as those
of the products formed by photolysis of keto ester 1g (i.e., the endo-
and exo-arylcyclobutanols 2g and 3g, respectively). Irradiation of
crystals of keto acid 1f followed by diazomethane workup gave a single
photoproduct which had the same GC retention time as endo-
arylcyclobutanol 2g. The characterization of this material is described
below.
Photolysis of Ketone 1g. A solution of 125 mg (0.40 mmol) of
keto ester 1g in 20 mL of acetone was photolyzed through Pyrex for
2 h. GC showed the presence of a small amount of starting material
plus two photoproducts in a 94:6 ratio. Silica gel column chromatog-
raphy using 5% ether/hexanes as eluent afforded 15 mg (12%) of
recovered starting material plus a mixture of photoproducts 2g and 3g.
Further column and radial chromatography (3% ether/hexanes) gave
81 mg (65%) of pure 2g (white powder) and 30 mg (23%) of a mixture
of 2g and 3g; a pure sample of photoproduct 3g could not be obtained.
Recrystallization of 2g from various solvents returned only white
powder.
Irradiation of crystals of ketone 1i (74 mg, 0.26 mmol) for 4 h
through Pyrex gave a 58% conversion to a single product, shown by
1
GC and H NMR to be endo-arylcyclobutanol 2i.
Photolysis of Ketone 5c. A solution of 88 mg (0.35 mmol) of
ketone 1h in 10 mL of benzene was photolyzed through Pyrex for 1.5
h. GC indicated the presence of a single photoproduct. Column
chromatography of the crude photolyzate using 5% ether/hexanes
afforded 2 mg of recovered starting material plus 85 mg (97%) of endo-
arylcyclobutanol 6c: oil; 1H NMR (C6D6) δ 7.18-7.11 (3H, m, ArH),
7.09-7.01 (2H, m, ArH), 2.94-2.90 (1H, m), 2.65-2.61 (1H, t, J )
6 Hz), 2.02-1.97 (1H, m), 1.85-1.80 (1H, m), 1.80-1.72 (2H, m),
1.70-1.60 (2H, m), 1.60-1.53 (3H, m), 1.53-1.49 (1H, m), 1.32 (3H,
s, CH3), 1.04 (1H, s, OH), 0.80-0.75 (1H, d, J ) 11 Hz); 13C NMR
(C6D6) δ 18.96 (-), 25.88 (-), 29.40 (-), 31.16 (+), 34.22 (+), 35.15
(+), 35.95 (+), 37.25 (-), 37.54 (-), 46.54 (-), 49.99 (+), 83.20
(+), 124.52 (-), 125.11 (-), 126.88 (-), 129.21 (-), 129.43 (-),
147.36 (+); IR (film) 3413 (OH), 2911, 2850, 1617, 1485, 1448, 1369,
1335, 1056, 1015, 976, 950, 904, 787, 772, 701, 580 cm-1; MS (EI)
m/z 255, 254 (M+, 100), 253, 236, 221, 211, 197, 193, 179, 167, 159,
150, 149, 145, 134, 120, 115, 105, 93, 91, 77, 67, 55, 41; HRMS (EI)
m/z 254.1670 (calcd for C18H22O, 254.1671). Anal. Calcd for
C18H22O: C, 84.99; H, 8.72. Found: C, 84.79; H, 8.72.
1
endo-Arylcyclobutanol 2g: mp 114-115 °C; H NMR (C6D6) δ
8.07-8.03 (2H, m, ArH), 7.34-7.30 (2H, m, ArH), 3.52 (3H, s, OCH3),
2.86-2.80 (2H, m), 2.06-2.00 (1H, m), 1.84-1.76 (1H, m), 1.67-
1.60 (1H, m), 1.41-1.22 (2H, m), 1.40 (1H, br s, OH), 1.05 (3H, s,
CH3), 1.00-0.96 (1H, m), 0.61 (9H, s, C(CH3)3); 13C NMR (C6D6) δ
19.79 (-), 20.62 (+), 28.71 (-), 32.69 (+), 37.04 (+), 46.63 (+),
47.11 (+), 47.94 (-), 51.55 (-), 54.54 (-), 80.67 (+), 127.94 (-),
129.27 (-), 129.48 (+), 151.02 (+), 166.54 (+); IR (KBr) 3492 (OH),
2983, 2946, 2866, 1703 (CdO), 1608, 1568, 1472, 1456, 1440, 1407,
1363, 1332, 1282, 1225, 1197, 1182, 1111, 1052, 1015, 986, 964, 946,
911, 860, 830, 778, 720, 530, 509 cm-1; MS (EI) m/z 317, 316 (M+),
301, 285, 260, 259, 258, 257, 241, 227, 205, 203, 201, 192, 173, 169,
165, 164, 163 (100), 159, 153, 152, 149, 146, 145, 143, 141, 136, 135,
131, 129, 128, 115, 105, 104, 103, 97, 95, 94, 91, 83, 81, 77, 69, 59,
57, 55; HRMS (EI) m/z 316.2039 (calcd for C20H28O3, 316.2038). Anal.
Calcd for C20H28O3: C, 75.91; H, 8.92. Found: C, 75.72; H, 8.88.
endo-Arylcyclobutanol 2g was also found to be the sole product
endo-Arylcyclobutanol 6c was also found to be the sole product
1
resulting from irradiation of crystals of ketone 5c (GC, H NMR).
1
resulting from irradiation of crystals of ketone 1g (GC, H NMR).
Photolysis of Ketone 5d. A solution of 127 mg (0.47 mmol) of
ketone 5d in 10 mL of benzene was photolyzed through Pyrex for 2 h.
GC indicated the presence of a single photoproduct. Column chro-
matography of the crude photolyzate using 5% ether/hexanes afforded
9 mg of recovered starting material plus 118 mg (93%) of endo-
arylcyclobutanol 6d. Recrystallization from petroleum ether gave
Photolysis of Ketone 1h. A solution of 69 mg (0.25 mmol) of
ketone 1h in 10 mL of benzene was photolyzed through Pyrex for 2 h.
GC indicated an 85% conversion and the presence of a single
photoproduct. Column chromatography of the crude photolyzate using
5% ether/hexanes afforded 54 mg of endo-arylcyclobutanol 2h as a
white powder. Recrystallization from various solvents returned only
1
colorless prisms: mp 96-98 °C; H NMR (C6D6) δ 6.95-6.90 (1H,
1
white powder: mp 59-61 °C; H NMR (C6D6) δ 7.30-7.26 (2H, m,
m, ArH)), 6.84-6.76 (3H, m, ArH), 2.87-2.82 (1H, m), 2.54-2.50
(1H, t, J ) 6 Hz), 1.93-1.88 (1H, m), 1.78-1.73 (1H, m), 1.70-1.45
(8H, m), 1.23 (3H, s, CH3), 1.09 (1H, br s, OH), 0.78-0.73 (1H, d, J
) 11 Hz); 13C NMR (C6D6) δ 18.79 (-), 25.81 (-), 29.29 (-), 31.11
(+), 34.19 (+), 35.03 (+), 35.91 (+), 37.14 (-), 37.58 (-), 46.56
(-), 49.92 (+), 82.55 (+), 115.83 and 116.00 (2JC-F ) 21 Hz, -ve),
ArH), 6.95-6.91 (2H, d, J ) 8 Hz, ArH), 2.93-2.88 (2H, m), 2.20-
2.13 (1H, m), 2.10 (3H, s, ArCH3), 1.92-1.84 (1H, m), 1.73-1.67
(1H, m), 1.55-1.46 (1H, m), 1.46-1.38 (1H, m), 1.36 (1H, br s, OH),
1.14 (3H, s, CH3), 1.06-1.01 (1H, m), 0.68 (9H, s, C(CH3)3); 13C NMR
(C6D6) δ 19.94 (-), 20.61 (+), 21.07 (-), 28.86 (-), 32.77 (+), 37.34

Yang Photocyclization Reaction in the Solid State
J. Am. Chem. Soc., Vol. 120, No. 49, 1998 12769
116.16 and 116.33 (2JC-F ) 21 Hz, -ve), 125.99 and 126.05 (3JC-F
)
1439, 1310, 1294, 1248, 1174, 1112, 1058, 1015, 1004, 951, 860, 791,
721 cm-1; MS (EI) m/z 313, 312 (M+), 297, 294, 281, 269, 255, 254,
253, 235, 228, 214, 192, 178, 164, 163, 150, 149 (100), 135, 119, 107,
105, 93, 91, 79, 67, 55, 41; HRMS (EI) m/z 312.1718 (calcd for
C20H24O3, 312.1725). Anal. Calcd for C20H24O3: C, 76.88; H, 7.75.
Found: C, 76.72; H, 7.74.
8 Hz, -ve), 126.81 and 126.87 (3JC-F ) 8 Hz, -ve), 143.19 and 143.22
(4JC-F ) 4 Hz, +ve), 161.60 and 163.01 (1JC-F ) 245 Hz, +ve); IR
(KBr) 3596 (OH), 2906, 2868, 2847, 1600, 1504, 1483, 1445, 1392,
1368, 1354, 1338, 1210, 1149, 1095, 1056, 1014, 998, 975, 953, 911,
839, 810, 582 cm-1; MS (EI) m/z 273, 272 (M+), 271, 254, 239, 229,
215, 211, 197, 185, 177, 163, 150, 149 (100), 138, 134, 123, 107, 105,
95, 93, 79, 67, 55, 41; HRMS (EI) m/z 272.1571 (calcd for C18H21OF,
272.1577). Anal. Calcd for C18H21OF: C, 79.38; H, 7.77. Found:
C, 79.55; H, 7.77.
endo-Arylcyclobutanol 6g was also found to be the sole product
1
resulting from irradiation of crystals of ketone 5g (GC, H NMR).
Photolysis of Salts S1-S17. Crystals of salts S1-S17 (1-2 mg)
were photolyzed between Pyrex microscope slides for varying lengths
of time and at different temperatures (see Table 4) as described under
Photochemical Procedures. The solid mixtures were removed from
the slides with ethyl acetate and treated with excess ethereal diaz-
omethane to convert the acids to the corresponding methyl esters. After
washing with water, drying, and removal of solvent in vacuo, the
mixtures were chromatographed on silica gel using 3% ether/hexanes
as eluent. The solid materials obtained after evaporation of the eluent
were analyzed by GC and chiral HPLC to determine conversions and
enantiomeric excesses of the only photoproducts observed, endo-
arylcyclobutanols 2g and 6g.
endo-Arylcyclobutanol 6d was also found to be the sole product
1
resulting from irradiation of crystals of ketone 5d (GC, H NMR).
Photolysis of Ketone 5e. A solution of 60 mg (0.21 mmol) of
ketone 5e in 10 mL of benzene was photolyzed through Pyrex for 3 h.
GC indicated the presence of a single photoproduct. Column chro-
matography of the crude photolyzate using 5% ether/hexanes afforded
2 mg of recovered starting material plus 58 mg (97%) of endo-
arylcyclobutanol 6e. Recrystallization from various solvents gave only
a white powder: mp 168-170 °C; 1H NMR (C6D6) δ 7.10-7.00 (2H,
m, ArH), 6.84-6.75 (1H, m, ArH), 6.69-6.59 (1H, m, ArH), 2.82-
2.71 (1H, m), 2.45-2.37 (1H, t, J ) 6 Hz), 1.85-1.65 (2H, m), 1.65-
1.27 (8H, m), 1.14 (3H, s, CH3), 0.95 (1H, br s, OH), 0.76-0.65 (1H,
d, J ) 13 Hz); 13C NMR (C6D6) δ 18.66 (-), 25.61 (-), 28.95 (-),
30.85 (+), 33.91 (+), 35.27 (+), 35.61 (+), 36.80 (-), 37.40 (-),
46.26 (-), 49.69 (+), 82.49 (+), 111.02 (+), 116.69 (+), 124.90 (-),
125.51 (-), 132.87 (-), 133.00 (-), 151.36 (+); IR (KBr) 3495 (OH),
2911, 2855, 2232, 1683, 1638, 1456, 1407, 1375, 1345, 1281, 1250,
1211, 1191, 1171, 984, 868, 547 cm-1; MS (EI) m/z 280, 279 (M+),
278, 261, 251, 236, 218, 206, 204, 181, 170, 159, 150, 149 (100), 134,
130, 107, 102, 93, 79, 67, 55, 41; HRMS (EI) m/z 279.1617 (calcd for
C19H21NO, 279.1623). Anal. Calcd for C19H21NO: C, 81.67; H, 7.58;
N, 5.02. Found: C, 81.55; H, 7.59; N, 4.99
Quantum yields were performed on thiophene-free benzene (purified
by successive washes with concentrated sulfuric acid, drying, and
distillation over sodium metal) or tert-butyl alcohol solutions containing
0.014-0.040 M ketone (Table 2), ∼0.003 M internal standard
(docosane, eicosane, heneicosane, or nonadecane), and 0 to ∼1.5 M
2,5-dimethyl-2,4-hexadiene as quencher. Equal volumes (2.8 mL) of
each solution (including actinometer, see below) were placed in
specially cleaned 13 × 100 mm Pyrex culture tubes, degassed by four
freeze-pump-thaw cycles, and sealed. Irradiations were conducted
at 20 ( 3 °C in a merry-go-round apparatus17 with a 450-W Hanovia
medium-pressure mercury lamp housed in an immersion well. The
313-nm mercury line was isolated by a filter combination of 7-54
Corning glass plates and an aqueous solution of 0.002 M K2CrO4
containing 5% K2CO3 by weight circulated through a Pyrex cooling
jacket. Quantum yields of photoproduct formation were determined
by quantitative GC relative to acetophenone formation (<10%) from
parallel runs on 0.1 M valerophenone in benzene containing 1.000 g/mL
of tetradecane as internal standard (Φ ) 0.33).15 GC detector response
factors were calculated from three injections of accurately prepared
solutions containing starting ketone, photoproduct, and internal standard.
Internal standards were chosen so that their peaks would not overlap
with any other peaks and yet have retention times close to those of the
photoproducts.
endo-Arylcyclobutanol 6e was also found to be the sole product
1
resulting from irradiation of crystals of ketone 5e (GC, H NMR).
Photolysis of Ketone 5f. A solution of 8 mg of keto acid 5f in 3
mL of benzene was irradiated through Pyrex for 2 h. The solvent was
removed in vacuo and the residue dissolved in ether and treated with
excess diazomethane. GC analysis showed the absence of starting
material and the presence of a single photoproduct, whose retention
time was the same as that of the product formed by photolysis of keto
ester 5g (i.e., the endo-arylcyclobutanol 6g). Irradiation of crystals of
keto acid 5f followed by diazomethane workup gave a single photo-
product which had the same GC retention time as endo-arylcyclobutanol
6g. The characterization of this material is described below.
Photolysis of Ketone 5g. A solution of 96 mg (0.31 mmol) of
ketone 5e in 10 mL of benzene was photolyzed through Pyrex for 3 h.
GC indicated the presence of a single photoproduct. Column chro-
matography of the crude photolyzate using 5% ether/hexanes afforded
9 mg of recovered starting material plus 87 mg (91%) of endo-
arylcyclobutanol 6g. Recrystallization from petroleum ether gave
colorless prisms: mp 141-143 °C; 1H NMR (C6D6) δ 8.15-8.07 (2H,
d of q, J ) 2 and 9 Hz, ArH), 7.13-7.06 (1H, d of d, J ) 2 and 9 Hz,
ArH), 7.00-6.93 (1H, d of d, J ) 2 & 9 Hz, ArH), 3.53 (3H, s, OCH3),
2.87-2.82 (1H, m), 2.55-2.50 (1H, t, J ) 6 Hz), 1.94-1.88 (1H, m),
1.76-1.70 (1H, m), 1.70-1.45 (8H, m), 1.22 (3H, s, CH3), 1.06 (1H,
s, OH), 0.79-0.72 (1H, d, J ) 12 Hz); 13C NMR (C6D6) δ 18.79 (-),
25.77 (-), 29.17 (-), 31.02 (+), 34.07 (+), 35.31 (+), 35.79 (+),
37.01 (-), 37.50 (-), 46.49 (-), 50.02 (+), 51.58 (-), 82.80 (+),
124.47 (-), 125.11 (-), 129.17 (+), 130.86 (-), 130.96 (-), 151.95
(+), 166.59 (+); IR (KBr) 3458 (OH), 2912, 2849, 1693 (CdO), 1609,
Acknowledgment is made to the donors of the Petroleum
Research Fund, administered by the American Chemical Society,
for partial support of this research. Financial support by the
Natural Sciences and Engineering Research Council of Canada
is also gratefully acknowledged. J.R.S. acknowledges funding
by NATO, in the form of a Senior Guest Fellowship, and
hospitality by Jean-Marie Lehn, Universite´ Louis Pasteur,
Strasbourg, France, where the manuscript was prepared. Finally,
we thank Professor Cornelia Bohne and Mr. Mark Kleinman
of the University of Victoria for carrying out the triplet lifetime
measurements, and Mr. Eugene Cheung of UBC for help with
the ORTEP drawings.
JA9823013